Foreword to the First Edition
might be replaced by "the study of those features of life that appear to be common to al...
343 downloads
1561 Views
172MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Foreword to the First Edition
might be replaced by "the study of those features of life that appear to be common to all forms." Whatever definition we choose, however, it is of immense satisfaction to me that this new book, essentially, has been fitted into the same sectional headings that I employed in my own book. If I may be permitted to reminisce further, I have wondered frequently how a single scientist, actively engaged in research, could write a new book of such wide scope. The answer is that I wrote the book during the 2 years immediately following the end of World War II. Thus, for several yearsmin Great Britain for as many as 10 yearsmvery little original academic physiology and new research had been published. This made it possible to survey the original literature of a lengthy period without being overwhelmed by a rapid succession of new discoveries that would have rendered my task nearly impossible, a fate similar to that of Sisyphus. Today, this task would be impossible, and it only surprises me that Nicholas has been able to produce this magnificent book with so few collaborators.
It was kind and generous of my friend Nicholas Sperelakis to relate this excellent book so closely to my own, A Textbook of General Physiology. In the preface to the first edition of my book, I had expressed the hope that it might be compared with Bayliss' Principles of General Physiology. If this comparison is valid, it is very appropriate that the present book be organized and partly written by one who is, along with Sir William Bayliss and myself, associated with University College London (Professor Sperelakis having spent a sabbatical year there). It is a pleasure to recall that it was there that I first met Nicholas, and I remember discussing his pioneering study on the potentials across the crystalline lens of the eye. For reasons that I think bear no relation to its scope, this book has a different title; I presume it is because the distinction between "ordinary" and "general" physiology has become sufficiently blurred to demand something more appropriate. The def'mition of general physiology that I had proposed in the preface to the first edition of my textbook was "the study of those aspects of living material that show some immediate prospect of being described in terms of the known laws of physics and chemistry." Later, I had misgivings as to the narrowness of this definition, and I then suggested that it
Hugh Davson 1995
XV
Foreword to the Second Edition
In his Foreword to the first edition of the Cell Physiology Source Book, Hugh Davson established it as
extensive revisions and the new material in the second edition raise it to a new level. Cell physiology, an area of central importance in biology, has grown out of a number of more traditional fields, and as a result, the literature continues to be widely dispersed. The great value of the Cell Physiology Source Book is that it gathers together under a single cover a broad range of up-to-date chapters that, taken together, define the field. The various chapters exhibit a uniformity of style and level of presentation that are a credit to the editor. Because of this and the scope and clarity of the presentations, this book can serve exceptionally well as an advanced undergraduate- or graduatelevel text for cell physiology courses. The broad coverage of this second edition also makes it very attractive for use in cell biophysics, membrane biology, and biomedical engineering courses. It can serve equally well as a textbook for introductory courses in ion channel structure and physiology. I was pleased, and indeed proud, to be asked by my colleague Nicholas Sperelakis to contribute the Foreword to the second edition of the Cell Physiology Source Book. This book clearly sets a new standard of excellence.
the lineal descendent of his own well-known and highly respected work The Textbook of General Physiology. The second edition of the Cell Physiology Source Book, again edited by Nicholas Sperelakis, continues in this same tradition. Although the first edition was enthusiastically received by the cell physiology community because of its depth and breadth of coverage, considerable important progress has been made in this rapidly developing area since its publication. The second edition deals with these new developments by a thorough reworking of topics and by the inclusion of new chapters in all sections covered in the first edition. The new topics introduced into the various sections include lipid structure, mitochondrial physiology, cell responses to hormones, red blood cell transport, neuron physiology, developmental changes in ion channels, sonotransduction, excitation-contraction coupling, and electroplax cells. In addition, the scope of the new edition has been valuably broadened by the inclusion of two entirely new sections. One titled Protozoa and Bacteria covers the physiology of these organisms in two chapters. In the other, Cell Division and Programmed Cell Death, there are chapters on the regulation of cell division, the cancer cell, apoptosis, and the effects of ionizing radiation. The
ThomasE. Thompson 1997
oo
XVll
Foreword to the T hird Edition
discovered, fashioned, combined, and interwoven into the tangled skein of processes that form the function of even the simplest cell is a process full of surprises as we discover with awe the audacity with which nature's molecular mechanisms are reused again and again in different contexts. One might call this audacity the "dance of the genes" were it not for the inanimate and unthinking nature of these bits of code that transmit the logic from one generation to another. The logic is physiological, not genetic. The molecular biologist Sydney Brenner put the point succinctly when he wrote recently (in a Novartis symposium titled The Limits of Reductionism in Biology) that "Genes can only specify the properties of the proteins they code for, and any integrative properties of the system must be 'computed' by their interactions." We are approaching the point at which our understanding of those interactions is deep enough for physiology to aspire to become the quantitative analytical discipline it needs to be to solve the problems it tackles. Brenner went on to remark, significantly, "This provides a framework for analysis by simulation." It is a sign of the maturity of our science that simulation is indeed also becoming an essential tool of analysis. It is further such sign that the scope of cell physiology is now so wide. The breadth of this book is therefore one of its greatest strengths. The chapters encompass the great majority of important molecular and cell systems, so that it does indeed justify its title as a source book.
When Nicholas Sperelakis kindly invited me to write this foreword, I wondered how I could possibly follow in the footsteps of Hugh Davson. I studied physiology at University College London (UCL) at the time when his monumental Textbook of General Physiology was published. UCL was an extraordinary place in which to be a student at that time. Not only was Hugh Davson laying his particular cornerstone of the subject, but Leonard Bayliss was also reworking his father's famous Principles of General Physiology. These were the books that convinced me to become a research physiologist and that cell physiology was the place to begin. Between them, Davson and Bayliss were responsible for seducing generations of medical students to discover the challenge and delights of physiological research. Later, as a young lecturer at UCL, I remember asking Hugh Davson why he didn't lecture very much to the students. He simply replied: "Denis, I've written it all. Tell them to read!" Actually, that remark inspired me to reread his books. My copies are still in my library, and they are well and truly thumbed. It will be the greatest tribute to the work of Nicholas Sperelakis and his colleagues that the pages of their book will also become the well-thumbed bible of a new generation of physiologists. They will be entering the discipline at an exciting time, for cell physiology is the base on which our understanding of all aspects of integrative and systems physiology must rest. This book, therefore, will be an essential source for all of us who aspire to understand the "logic of life," which is, after all, the meaning and origin of the word "physiology." Life has no single "logic," of course. Unraveling the physicochemical mechanisms that nature has
Denis Noble 1999
xix
Preface to the First Edition
This book is intended primarily for graduate and advanced undergraduate students in the life sciences, including those taking courses in cell physiology, cell biophysics, and cell biology. Selected parts of this book can be used fo~ courses in neurobiology, electrobiology, electrophysiology, secretory biology, biological transport, and muscle contraction. Students majoring in engineering, biomedical engineering, physics, and chemistry should find this book useful to help them understand the living state of matter. Postdoctoral scholars and faculty engaged in biological research would also find this text of immense value for obtaining a better understanding of cell function and as a reference source. Medical, dental, and allied health students could use this book as a valuable comparison text to complement their other textbooks in medical/mammalian physiology. The latter texts often strongly emphasize organ system physiology to the virtual exclusion (in some cases) ot cell physiology, membrane biophysics, and mechanisms underlying homeostasis of the entire organism. The chapter authors were asked to include the following aspects wherevel appropriate: comparative physiology, developmental changes, pathophysiology, membrane diseases, and molecular biology. Considering the tremendous amount of new informatior gained during the past two decades using sophisticatec new instruments and techniques, I have recruited a numbel of outstanding researchers, who are leaders in their respective fields, to contribute to this undertaking. I am delightec that they agreed to participate. It has been my great pleasure and honor to work with them on this important anc timely project.
Hugh Davson's original textbook on cell physiology, A
Textbook of General Physiology, was a true classic and a huge success. The first edition was published in 1959, and the fourth and final edition in 1970. In the past two decades, cell physiology has advanced by leaps and bounds, as, for example, in the area of ion channels and their regulation. The present book attempts to fill the void left by the discontinuation of Davson's monumental book. This book is dedicated to Professor Davson in recognition of the great impact he has had on the dissemination of the principles of cell physiology. At present, there is a need for a good text on cell physiology. Several good texts on cell biology are available, but these generally treat cell physiology in a superficial and incomplete manner. Therefore, it was our intent to prepare a new work on cell physiology that is high quality, is comprehensive, and covers recent developments (the chapter authors were asked to limit their bibliography to no more than 40 key papers and review-type articles). We hope that the reader will find this book clearly written, thorough, up to date, and worthy of being the successor to Professor Davson's book. This book focuses on physiology and biophysics at the cellular level. Organ systems and whole organisms are essentially not covered, except for unicellular organisms. The major topics covered include ultrastructure, molecular structure and properties of membranes, transport of ions and nonelectrolytes, ion channels and their regulation, membrane excitability, sensory transduction mechanisms, synaptic transmission, membrane receptors, intracellular messengers, metabolism, energy transduction, secretion, excitation-secretion coupling, excitation-contraction coupling, contraction of muscles, cilia and flagellae, photosynthesis, and bioluminescence. These topics are covered in a comprehensive, but didactic, manner, with each topic beginning in an elementary fashion and ending in a sophisticated and quantitative treatise.
Nicholas Sperelaki: 199~
xxi
Preface to the Second Edition
The first edition of the Cell Physiology Source Book was conceived to serve as the replacement for Davson's classic textbook of cell physiology, which was discontinued after the fourth edition was published in 1970. There was a pressing need for a comprehensive and authoritative textbook on cell physiology because the available cell biology textbooks treated cell physiology in a superficial and incomplete manner. To accomplish this Herculean task, I recruited a number of outstanding researchers who were leaders in their respective fields to contribute to the undertaking. The result was a comprehensive source book (738 pages, hardcover) that appeared in February 1995 and rapidly became successful. Within a year, the book had to be reprinted in hardcover, and a softcover version was printed. Book reviews were complimentary to the book, and I received numerous comments from contributors and other scientists about the excellence of the Source Book. One of my favorite comments was made by an Israeli physical chemist at a conference in Prague where the book was on display. He said, "Nick, this is the type of book that speaks to me." The first edition was selected as one of the outstanding academic books for 1996 by CHOICE (current reviews for academic libraries published by the Association of College & Research Libraries, a division of the American Library Association). I began to formulate plans for the second edition almost immediately after the appearance of the first edition. I believed that it was important for the second edition to appear within 3 years after the first edition. It was especially important to have a relatively short interval because the field is advancing so rapidly. I also wanted to take the opportunity to include some new chapters on topics quite relevant to the discipline of cell physiology. In addition, several of the original chapters were expanded, reorganized, and restructured, and several other chapters were shifted in location, with the goal of making the Source Book more useful as a textbook. A total of 18 new chapters were added, and two new sections were organized: Section VII, Protozoa and Bacteria; and Section IX, Cell Division and Programmed Cell Death. The new chapters include lipid structure, physiology of mitochondria, responses to hormones, transport in erythro-
xxiii
cytes, developmental changes of ion channels, calcium receptors, cytoskeletal modulation of ion channels, neuronal cells, sonotransduction, excitation-contraction coupling, electroplax cells, electroreceptors, protozoa, bacteria, cell division, cancer cells, apoptosis, and ionizing radiations. These additional chapters should make the Source Book more complete and more useful to graduate students in neuroscience, biology, cell biology, transport physiology, and electrophysiology. Selected chapters in the Source Book can be used for courses in cell physiology, cell biophysics, electrophysiology, electrobiology, transport physiology, secretory biology, sensory physiology, neurobiology, and contractile systems. Students majoring in the physical sciences should find this book helpful in understanding the living state of matter. This book is intended primarily for graduate students and advanced undergraduate students majoring in the life sciences. A new edition also allows the individual contributor the opportunity of adjusting the level at which his or her chapter is pitched (i.e., to aim it at the median level at which the majority of the chapters are pitched). Thus, parts of some chapters were expanded, whereas parts of others were constricted. Certain figures were improved. The subject index was improved by limiting entries to first order and second order. I sincerely hope that the student will find this book to be clearly written, thorough, up to date, and worthy of being the successor to Davson's monumental book. The book contains several opening chapters and an appendix to help the student review the physical chemistry of solutions, protein structure, lipid structure, and electricity. Hence, within the covers of this one book, the student should have all the information needed to understand the various topics that constitute the discipline of cell physiology. The contributors, publisher, and I want to establish this book as the leading textbook of cell physiology.
Nicholas Sperelakis 1997
Preface to the Third Edition
These changes, along with the updating of all chapters, should make the Sourcebook more complete and more useful to students and faculty. Graduate students and advanced undergraduate students majoring in the life sciences should find this book very useful. Students majoring in physics, chemistry, and biomedical engineering will find this book helpful in understanding the living state of matter. Selected chapters in the Sourcebook can be used for academic courses in cell physiology and biophysics, electrophysiology, transport physiology, sensory physiology, and neurobiology. The Sourcebook can be used as an adjunct text for courses in neuroscience, cell biology, and membrane biophysics. The contributors and I hope that the student will find this Sourcebook to be clearly written, comprehensive, up to date, and worthy of being the successor to Hugh Davson's classic book on cell physiology. The book contains several opening chapters to help the student review the physical chemistry of solutions, protein structure, and lipid structure. The Appendix to the book provides a review of electricity relevant to biology. It is my pleasure, as editor, to work with a distinguished group of scientists who are experts in the various topics included in this Sourcebook. The contributors and I have put much effort into this enormous task, and we are very proud of our accomplishment.
The first edition and second edition of Cell Physiology Source Book have done very well in terms of sales and usage. In addition, each edition was selected as one of the outstanding academic books [for 1996 (lst ed.) and 1998 (2nd ed.)] by CHOICE of the American Library Association. This is indeed an honor for the editor, contributors, and publisher. Therefore, this textbook/sourcebook is now well established, and it is probably the leading book in the field of cell physiology and biophysics. The second edition was substantially expanded from the first edition, and the third edition is somewhat expanded from the second edition. The new chapters include Membrane Structure (Chapter 4) and Intracellular Chloride Regulation (Chapter 20), an appendix on infrared detectors (Chapter 49), and an appendix on mechanism of electroreceptors (Chapter 50). Several chapters have new authors and new focus, including Transport of Ions and Nonelectrolytes (Chapter 16), Direct Regulation of Ion Channels by G Proteins (Chapter 34), Sensory Receptors and Mechanotransduction (Chapter 45), and Physiological Effects of Pressure on Cell Function (Chapter 59). In addition, some new authors were brought in to help update several chapters, including Diffusion and Permeability (Chapter 13), Regulation of Ion Channels by Phosphorylation (Chapter 33), Ion Channels as Targets for Disease (Chapter 39), Skeletal Muscle Action Potentials (Chapter 51), Smooth Muscle Action Potentials (Chapter 53), Amoeboid Movement, Cilia, and Flagella (Chapter 57), and Plant Cell Physiology (Chapter 63). One chapter was dropped due to time constraints (Polarity of Cells and Membrane Regions; old Chapter 21).
Nicholas Sperelakis 2000
xxv
Contributors
Numbers in parentheses indicate the pages on which the authors' contributions begin.
Hugu~s Abriel (643), Department of Pharmacology, College of Medicine, Columbia University, 630 W 168th Street, New York, New York 10032
Nacional Aut6noma de M6xico, Apartado Postal 510-3, 62250 Cuernavaca, Morelos, Mexico John R. Dedman (167), Department of Molecular and Cellular Physiology, College of Medicine, University of Cincinnati, Cincinnati, Ohio 45267-0576 Louis J. DeFelice (539), Department of Pharmacology, School of Medicine, Vanderbilt University, Nashville, Tennessee 37232-6600 Guy Droogmans (485), Department of Physiology, KU Leuven, Campus Gasthuisberg, Herestraat, B-3000 Leuven, Belgium
Francisco J. Alvarez-Leefmans (301), Department of Physiology and Biophysics, Wright State University School of Medicine, Room BS 158, 3640 Colonel Glenn Highway, Dayton, Ohio 45435 Juan Manuel Arias (1171), Department de Bioqufmica, Centro de Investigaci6n y de Estudios Avanzados del I.P.N., 07000 M6xico, D.E, Mexico David M. Baishaw (261), Department of Biochemistry, University of North Carolina, Chapel Hill, North Carolina 275997260
Istvan Edes (271), Department of Cardiology and Pulmonary, Medical University of Debrecen, 4004 Debrecen, Hungary
David W. Barnett (725), Departments of Medicine and Cell Biology/Physiology, Jewish Hospital, Washington University Medical Center, St. Louis, Missouri 63110
Joseph J. Feher (319), Department of Physiology, Box 551, Medical College of Virginia, Virginia Commonwealth University, Richmond, Virginia 23298-0551
Clive M. Baumgarten (319), Department of Physiology, Medical College of Virginia, Virginia Commonwealth University, Box 551 MCV Station, Richmond, Virginia 23298-0551
Donald G. Ferguson (95), Department of Anatomy, College of Medicine, Case Western Reserve University, Cleveland, Ohio 44106-4930 Harvey M. Fishman (857), Department of Physiology and Biophysics, University of Texas Medical Branch, Galveston, Texas 77555-0641 Darrell E. Fleischman (1097), Department of Biochemistry and Molecular Biology, Wright State University, 3640 Colonel Glenn Highway, Dayton, Ohio 45435 Michael S. Forbes (95), Merry Oaks, Rt. 1, Box 245, Troy, Virginia 22974-9741 Jeffrey C. Freedman (3), Department of Physiology, SUNY Health Science Center at Syracuse, 766 Irving Avenue, Syracuse, New York 13210 Andrew S. French (761), Department of Physiology and Biophysics, Dalhousie University, Halifax, Nova Scotia B3H 4H7, Canada Akikazu Fujita (601), Department of Pharmacology II, Faculty of Medicine, Osaka University, 2-2, Yamadoaka, Suita, Osaka 565, 0871 Japan
Michael M. Behbehani (479), Department of Molecular and Cellular Physiology, University of Cincinnati College of Medicine, Cincinnati, Ohio 45267-0576 Kenneth M. Blumenthai (625), Department of Molecular Genetics, University of Cincinnati College of Medicine, Cincinnati, Ohio 45267-0524 John H. B. Bridge (283), Department of Physiology, University of Utah, Nora Eccles Harrison Building, Salt Lake City, Utah 84112 Shirley H. Bryant (653), Department of Pharmacology and Cell Biophysics, Universitiy of Cincinnati College of Medicine, Cincinnati, Ohio 45267-0575 Alma D. Ch~ivez (1135), Departamento de Bioenerg6tica, Universidad Nacional Aut6noma de M6xico, Instituto de Fisiologia Celular, Apartado Postal 70-243, 04510 M6xico, D.E, Mexico
Nicole Gallo-Payet (705), Endocrine Service, Department of Medicine, University of Sherbrooke, Sherbrooke, Quebec, J1H 5N4, Canada
Alberto Darszon (509), Departmento Gen6tica y Fisiologia Molecular, Instituto de B iotechnologia, Universidad xi
xii Keith D. Garlid (139), Department of Biochemistry and Molecular Biology, Oregon Graduate Institute of Science and Technology, 20,000 NW Walker Road, Beaverton, Oregon 97006-8921, or P.O. Box 91000, Portland, Oregon 97291-1000
W. Gibson Wood (81), Department of Physiology and Pharmacology, Texas A&M University, College Station, Texas 77843-4466 Diana Gonz~ilez (1135), Departamento de Bioenergrtica, Universidad Nacional Aut6noma de Mrxico, Instituto de Fisiologia Celular, Apartado Postal 70-243, 04510 Mrxico, D.F., Mexico Hugo Gonzalez-Serratos (865), Department of Physiology, University of Maryland at Baltimore School of Medicine, 655 West Baltimore Street, Baltimore, Maryland 21201-1559
Thomas Gorrell (1041), Haskins Laboratories, Pace University, 41 Park Row (at Pace Plaza), New York, New York 10038-1598 Kathleen Heppner Goss (1161), Department of Biochemistry and Molecular Genetics, College of Medicine, University of Cincinnati, Cincinnati, Ohio 45267-0524
Contributors Judith A. Heiny (911), Department of Molecular and Cellular Physiology, College of Medicine, University of Cincinnati, Cincinnati, Ohio 45267-0576 Kent R. Hermsmeyer (899), Dimera LLC, 2525 NW Lovejoy, Suite 401, Portland, Oregon 97210
Nelson Horseman (191), Department of Molecular and Cellular Physiology, University of Cincinnati College of Medicine, Cincinnati, Ohio 45267-0576 Ching-hsien Huang (43), Department of Biochemistry and Molecular Genetics, University of Virginia School of Medicine, Box 440, Charlottesville, Virginia 22908-0001
Atsushi Inanobe (573), Department of Pharmacology II, Faculty of Medicine, Osaka University, 2-2, Yamadoaka, Suita, Osaka 565,0871 Japan Marcia A. Kaetzei (167), Department of Molecular and Cellular Physiology, College of Medicine, University of Cincinnati, Cincinnati, Ohio 45267-0576 Edna S. Kaneshiro (959), Department of Biological Sciences, University of Cincinnati, Cincinnati, Ohio 45221-0006
Anthony L. Gotter (1025), Laboratory of Developmental Chronobiology, Children's Service GRJ 1226, Massachusetts General Hospital, Boston, Massachusetts 02114
Robert S. Kass (643), Department of Pharmacology, College of Medicine, Columbia University, 630 W. 168th Street, New York, New York 10032
Michael S. Grace (832), Department of Biological Sciences, Florida Institute of Technology, Melbourne, Florida 32901
James G. Kereiakes (1185), Department of RadiologyMedical Physics, College of Medicine, University of Cincinnati, Cincinnati, Ohio 45267-0579
Steven M. Grassl (249), Department of Pharmacology, SUNY Health Science Center, 750 East Adams Street, Syracuse, New York 13210 Lawrence Griffing (1079), Department of Biology, Texas A & M University, College Station, Texas 77843 Joanna Groden (1161), Department of Biochemistry and Molecular Genetics, College of Medicine, University of Cincinnati, Cincinnati, Ohio 45267-0524
Ann B. Kier (81), Department of Physiology and Pharmacology, Texas A&M University, College Station, Texas 77843-4466 Evangelia G. Kranias (271), Department of Pharmacology and Cell Biophysics, College of Medicine, University of Cincinnati, Cincinnati, Ohio 45267-0575
Dennis W. Grogan (1063), Department of Biological Sciences, University of Cincinnati, Cincinnati, Ohio 45267-0006
Yoshihisa Kurachi (573), Department of Pharmacology II, Faculty of Medicine, Osaka University, 2-2, Yamadoaka, Suita, Osaka 565, 0871 Japan
Augustin Guerrero (1171), Department de Bioqufmica, Centro de Investigacion y de Estudios Avanzados del I.P.N., 07000 Mrxico, D.E, Mexico
William J. Larsen (523), Department of Anatomy and Cell Biology, College of Medicine, University of Cincinnati, Cincinnati, Ohio 45267-0521
Andrrs A. Gutirrrez (1135), Departamento de Bioenergrtica, Universidad Nacional Aut6noma de Mrxico, Instituto de Fisiologia Celular, Apartado Postal 70-243, 04510 Mrxico, D.E, Mexico
Harold Lecar (441), Department of Molecular/Cell Biology, 102 Donner Lab, University of California at Berkeley, Berkeley, California 94720
Eric J. Hall (1185), Center for Radiological Research, College of Physicians and Surgeons, Columbia University, New York, New York 10032 J. Woodland Hastings (1115), Biological Laboratories, Harvard University, 16 Divinity Avenue, Cambridge, Massachusetts 02138
Hiroshi Hibino (601), Department of Pharmacology II, Faculty of Medicine, Osaka University, 2-2, Yamadoaka, Suita, Osaka 565, 0871 Japan Christopher D. Heinen (1161), Department of Biochemistry and Molecular Genetics, College of Medicine, University of Cincinnati, Cincinnati, Ohio 45267-0524
Michael Levandowsky (1041), Haskins Laboratories, Pace University, 41 Park Row (at Pace Plaza), New York, New York 10038-1598 Simon Rock Levinson (455), Department of Physiology, Health Science Center, University of Colorado, 4200 East Ninth Avenue, Denver, Colorado 80262 Michael A. Lieberman (119), Department of Molecular Genetics, College of Medicine, University of Cincinnati, Cincinnati, Ohio 45267-0524
Arturo Lirvano (509), Departmento de Bioqufmicia, Instituto de Biotechnologfa, Universidad Nacional Autrnoma de Mrxico, Apartado Postal 510-3, 62271 Cuernavaca Morelos, Mexico
Contributors
xiii
Alister G. Macdonald (1003), Department of Biomedical
Robert W. Putnam (357), Department of Physiology and
Sciences, c/o Zoology Building, University of Aberdeen, Tillydrone Avenue, Aberdeen AB24 2TZ, United Kingdom
Biophysics, School of Medicine, Wright State University, 3640 Colonel Glenn Highway, Dayton, Ohio 45435
Daniel C. Marcus (775), Departments of Anatomy and Physiology, Kansas State University, 1600 Denison Avenue, Manhattan, Kansas 66506
llaria Rivolta (643), Department of Pharmacology, College of Medicine, Columbia University, 630 W 168th Street, New York, New York 10032
Martinez (1135), Departamento de B ioenerg6tica, Universidad Nacional Aut6noma de M6xico, Instituto de Fisiologia Celular, Apartado Postal 70-243, 04510 Mexico, D.E, Mexico
Stephen D. Roper (815), Department of Physiology and Biology, University of Miami School of Medicine, R430, P.O. Box 016430, Miami, Florida 33101
James Maylie (653), Oregon Health Sciences University,
3181 SW Sam Jackson Park Road, Portland, Oregon 97201
College Wisconsin, 8701 Watertown Plank Road, Milwaukee, Wisconsin 53226-0509
Michele Mazzanti (539), Dipartimento di Fisiologia e Biochimica Generali, Laboratorie di Erettrofisiologia, Universita degli Studi di Milano, Via Celoria 26, 1-20133 Milan, Italy
Michael Sanderson (959), Department of Physiology, University of Massachusetts Medical Center, 55 Lake Avenue, Worcester, Massachusetts 01655-0127
Gerhard Meisner (927), Department of Biochemistry, Biophysics, and Physiology, University of North Carolina, Chapel Hill, North Carolina 27599
William T. Sather (455), Department of Pharmacology, University of Colorado Health Sciences Center, Denver, Colorado 80262
Lauren A. Millette (261), Department of Pharmacology
Friedhelm Schroeder (81), Department of Physiology and
and Cell Biophysics, College of Medicine, University of Cincinnati, Cincinnati, Ohio 45267-0575
Pharmacology, Texas A&M University, College Station, Texas 77843
Stanley Misler (725), Departments of Medicine and Cell
Richard G. Sleight (119), Yale University Graduate
Biology/Physiology, Jewish Hospital, Washington University Medical Center, St. Louis, Missouri 63110
School, P.O. Box 208236, 320 York Street, New Haven, Connecticut 06520-8236
Catherine E. Morris (745), Departments of Biology and Medicine, Loeb Institute, University of Ottawa, Ottawa Civic Hospital, 1053 Carling Avenue, Ottawa, Ontario K1Y 4E9 Canada
Nicholas Sperelakis (209), Department of Molecular and
Francisco
Nancy J. Rusch (899), Department of Physiology, Medical
Cellular Physiology, College of Medicine, University of Cincinnati, Cincinnati, Ohio 45267-0576
Chipeta Way, Salt Lake City, Utah 84108-1256
Kira Steigerwald (1161), Department of Biochemistry and Molecular Genetics, College of Medicine, University of Cincinnati, Cincinnati, Ohio 45267-0524
John New (839), Biology Department, Loyola University,
Richard G. Stout (1079); Department of Biology, College
6525 North Sheridan Road, Chicago, Illinois 60626 Bernd Nilius (485), Department of Physiology, KU Leuven,
of Letters and Science, Montana State University, Bozeman, Montana 59717-0001
Campus Gasthuisberg, Herestraat, B-3000 Leuven, Belgium
Janusz B. Suszkiw (689), Department of Molecular and
Denis Noble (xix), Department of Physiology, University of
Oxford, Parks Road, Oxford OX1 3PT, United Kingdom
Cellular Physiology, College of Medicine, University of Cincinnati, Cincinnati, Ohio 45267-0576
Richard J. Paul (941), Department of Molecular and
Nortisugu Tohse (585), Department of Physiology, College
Edward E Nemeth (179), NPS Pharmaceuticals, Inc., 420
Cellular Physiology, College of Medicine, University of Cincinnati, Cincinnati, Ohio 45267-0576
of Medicine, Sapporo Medical University, South 1 West 17, Chuo-ku Sapporo 060, Japan
Marcel D. Payet (705), Department of Molecular and
Paivi H. Torkkeli (761), Department of Physiology and Biophysics, Dalhousie University, Halifax, Nova Scotia B3H 4H7, Canada
Cellular Physiology, Faculty of Medicine, University of Sherbrooke, Sherbrooke, Quebec J1H-5N4, Canada
C. Tricas (839), Department of Zoology, University of Hawaii at Maroa, 2538 McCarthy Mall, Edmonson Hall, Honolulu, Hawaii 96822
J. Wesley Pike (191), Department of Molecular and Cellular Physiology, College of Medicine, Universiity of Cincinnati, Cincinnati, Ohio 45267-0576
Timothy
Matthew R. Pincus (19), Department of Pathology and
Richard D. Veenstra (412), Department of Pharmacology,
Laboratory Medicine, Veterans Administration Medical Center, Brooklyn, New York 11209
College of Medicine, State University of New York, 750 East Adams Street, Syracuse, New York 13210
David M. Pressel (725), Departments of Medicine and Cell
Gordon M. Wahler (559), Department of Physiology, Mid-
Biology/Physiology, Jewish Hospital, Washington University Medical Center, Street Louis, Missouri 63110
western University, 555 31st Street, Downers Grove, Illinois 60515
Raymund Y. K. Pun (441), Department of Molecular and Cellular Physiology, College of Medicine, University of Cincinnati, Cincinnati, Ohio 45267-0576
Earl T. Wallick (261), Department of Pharmacology and Cell Biophysics, College of Medicine, University of Cincinnati, Cincinnati, Ohio 45267-0575
XiV
Contributors
Gary L. Westbrook (675), Vollum Institute and Department of Neurology, Oregon Health Sciences University, 3181 SW Sam Jackson Park Road, L474, Portland, Oregon 97201-3098
Medicine, Hokkaido University School of Medicine, Kita-15, Nishi-7, Kita-ku, Sapporo 060 Japan
George Witman (959), Department of Cell Biology, University of Massachusetts Medical School, 55 Lake Avenue N., Worcester, Massachusetts 01655
Anita L. Zimmerman (807), Department of Physiology, Brown University, Box G-B329, Providence, Rhode Island 02912
Hisashi Yokoshiki (585), Department of Cardiovascular
Jeffrey C. Freedman
Biophysical Chemistry of Physiological Solutions
Crothers (1979) and by Tinoco et al. (1994). During the past fifty years, outstanding monographs on biophysical chemistry have also been available: Hrber (1945), Edsall and Wyman (1958), Tanford (1961), Cantor and Schimmel (1980), Silver (1985), van Holde (1985), and Bergethon and Simons (1990). To help the reader begin to understand how cellular homeostasis is achieved, this chapter and the following one will introduce some of the conceptual underpinnings of cell physiology by describing certain physicochemical properties of water, electrolytes, and proteins that are relevant for understanding the structure and function of living cells.
!. Introduction All living cells contain proteins, salts, and water enclosed in membrane-bounded compartments. These biochemical and ionic cellular constituents, along with a set of genes, enzymes, substrates, and metabolic intermediates, function to maintain cellialar homeostasis and enable cells to replicate and to perform chemical, mechanical, and electrical work. Homeostasis, a term introduced by the physiologist Walter Cannon in The Wisdom of the Body (1932), means that certain parameters, including cellular volume, intracellular pH, the transmembrane electrical potential, and intracellular concentrations of salts are maintained relatively constant in resting cells. Cellular homeostasis depends on a relative constancy of the extracellular fluids that bathe cells. The extracellular fluid compartment was termed the milieu intdrieur, or internal environment, by Claude Bernard, who recognized around 1865 that "La fixitd du milieu intdrieur est la condition de la vie libre," that the constancy of the internal environment is the condition for independent life. In An Introduction to the Study of Experimental Medicine (see 1949 translation), Bernard wrote that "only in the physicochemical conditions of the inner environment can we find the causation of the external phenomena of life." The development of cell physiology was greatly influenced by the Bernard-Cannon theory of physical-chemical homeostasis. Biophysical chemistry concerns the application of the concepts and methods of physical chemistry to the study of biological systems. Physical chemistry includes such physiologically relevant subjects as thermodynamics, chemical equilibria and reaction kinetics, solutions and electrochemistry, properties and kinetic theory of gases, transport processes, surface phenomena, and molecular structure and spectroscopy. Throughout this book, it is seen that many cellular physiological phenomena are best understood with a rigorous and comprehensive understanding of physical chemistry. Physical chemistry texts that specifically emphasize biological applications include those by Eisenberg and Cell Physiology Sourcebook: A Molecular Approach, Third Edition
II. Structure and Properties of Water Biological cells contain a large amount of water, ranging from 0.66 g H20/g cells in human red blood cells to around 0.8 g H20/g tissue in skeletal muscle. The amount of water in cells is determined by osmosis (for review, see Dick, 1959, and Chapter 21). Liquid water is a highly polar solvent, with a structure stabilized by extensive intermolecular hydrogen bonds (Fig. 1). The hydrogen bonds between adjacent water molecules are linear (O-H"-O) but are able to bend by about 10 ~ From x-ray diffraction studies of ice crystals, it is known that each of the two covalent O-H bonds in water is 1 ~ in length, that each of the two hydrogen bonds is about 1.8 /~ in length, and that these occur around each oxygen atom at angles of 104o30 " in a tetrahedral array. The energy required to break hydrogen bonds in liquid water is 5 to 7 kcal/mol, much less than the 109.7 kcal/mol required to break the covalent O-H bond. A calorie is a unit of work equal to 4.18 J in the SI system of units, where 1 joule (J) equals 1 Newton.meter. The newton (N = kg.m/s2) is the unit of force in tile International System of Units (SI); the dyne (dyn = g-cm/s 2) is the corresponding unit of force in the older cgs system (1 N = 105 dyn). Recall that, according to Newton's first 3
Copyright 9 2001 by Academic Press. All rights of reproduction in any form reserved.
4
SECTION I Biophysical Chemistry, Metabolism, Second Messengers, and Ultrastructure
law, the force (F) is mass times acceleration, and that work is defined as force times distance. Because of the strength of hydrogen bonds, liquid water has unique physicochemical properties, including a high boiling point (100 ~ a high molar heat capacity (18 cal/mol.K), a high molar heat of vaporization (9.7 kcal/mol at 1 atm), and a high surface tension (72.75 dyn/cm). As discussed by Bergethon and Simons (1990), all of these thermodynamic parameters are considerably higher for H20 than for the analogous compound I-I2S, which boils at -59.6 ~ The outer electrons of the S atom shield its nuclear charge and reduce its electronegativity, making the S-H bond weaker, longer, and less polar than the O-H bond. Also, the bond angle of H2S is only 92~ ", an angle that does not form a tetrahedral array, and thus I-I2S does not form a hydrogen bonded network like water. The extensive intermolecular association in liquid water is due to the geometry of the water molecule, and also to its strong permanent dipole moment. Water molecules have no net charge, yet possess permanent charge separation along each of the two O-H bonds. Since oxygen is more electronegative than hydrogen, its electron density is preferentially greater. Oxygen thus acquires a partial negative charge (8-), leaving the hydrogens with a partial positive charge (6 +) (Fig. 1C). The magnitude of the dipole moment (/z) of a chemical bond is computed as the product of the separated charge (q) at either end, and the distance (d) between the centers of separated charge (/z = q 9d). The dipole moment of a molecule is the vector sum of the dipole moments of each bond. For water in the gaseous phase, the dipole moment is 1.85 debye, where 1 debye = 10-18 esu.cm. One esu (electrostatic unit), or statcoulomb, equals 3.336 • 10-10 C , the coulomb (C) being a unit of electric charge. In liquid water, the dipole moment becomes even larger (2.5 debye) due to association with other water molecules. In comparison, the linear molecule carbon dioxide (O=C=O) also has a separation of charge, with a partial negative charge on each oxygen atom and a partial positive charge on the carbon atom. In this case, the oppositely directed dipole moments of the two C=O bonds sum to a zero dipole moment for the CO 2 molecule. In H2S, the S-H bond has a smaller dipole moment (1.1 debye) than the O-H bond in water because sulfur is less electronegative than oxygen. Electrostatic attractions between opposite charges in adjacent water dipoles stabilize the hydrogen-bonded structure of liquid water. Water strongly affects the forces between ions in solution by virtue of its high dielectric constant. The dielectric constant (e) of a medium known as a dielectric is defined as the ratio of the coulombic force (Fcoul) between two charges in a vacuum to the actual force (F) between the same two charges in the dielectric medium. F oul
F In a vacuum, the coulombic force, Fcoul (newtons, N), between two ions with charges q+ and q- (C) separated by a distance d (meters, m) is given by Coulomb's law: q+q-
Fcoul - 47re 0 d2 where e0 is the permittivity constant (8.854 • 10--12C2/N-m2). Coulomb's law states that the force between two point
FIGURE 1. Hydrogenbonds in ice and liquid water. (A) Crystal structure of ice, showing the puckered hexagonal rings of oxygen atoms. (B) Each oxygen atom is connected to four others in a tetrahedron. (C) The linear hydrogen bonds are indicated by dotted lines between adjacent water molecules (O-H'"O). Two covalent bonds (O-H), indicated by solid lines, and two hydrogen bonds occur around each oxygen atom in a tetrahedral array. The partial charge separation of the O-H dipoles is indicated on the upper water molecule.
charges is directly proportional to the magnitude of each charge, and inversely proportional to the square of the distance between the charges. Substituting Fcoul into the above expression and rearranging yields the following. q+qF=
4"rre oe d 2
The dielectric constant of a vacuum is unity, and that of air is close to unity ( e a i r - - 1.00054). Increasing the dielectric constant of the solvent decreases the attractive force between oppositely charged ions. The force felt by a distant ion is reduced in a dielectric medium such as water, as compared with a vacuum, because part of the interaction energy is spent aligning the intervening water dipoles and distorting their polarizable electron clouds. The dielectric constant of water at 25 ~ is 78.5, much greater than methanol (e = 32.6), ethanol (24.0), or methane (1.7). In liquid water, the high dielectric constant weakens the coulombic attractive forces between oppositely charged particles, and thus promotes dissociation and ionization of salts. Although the thermodynamic properties of liquid water are explicable in terms of extensive hydrogen-bonding, the
5
1. Biophysical Chemistry of Physiological Solutions
actual structure of water is still unknown (see Eisenberg and Kauzmann, 1969). In the flickering cluster model, groups of 50-70 water molecules, resembling a slightly expanded broken piece of the ice lattice (icebergs), are continuously associating and dissociating on a picosecond time scale. This dynamic model contrasts with the extended order of solid ice. At a given instant, some water molecules are unattached to the clusters and are located in the interstitial regions of the network, but may attach and detach as the clusters continuously form and break down. Other theories treat water as a mixture of distinct states, or as a continuum of states, with considerable short-range order, characteristic of the crystalline lattice of ice. Despite these uncertainties regarding the structure of liquid water and the influence of macromolecules on its properties in cells (see Cooke and Kuntz, 1975), the ability of water to act as a solvent inside cells closely resembles that of extracellular water. Thus, a variety of permeant, hydrophilic, nonmetabolized nonelectrolytes distribute at equilibrium across human red blood cell membranes with ratios of intracellular to extracellular concentrations that deviate from unity by less than 10% (Gary-Bobo, 1967). In mouse diaphragm muscle, C. Miller (1974) found that several alcohols, diols, and monosaccharides exhibit distribution ratios within 2% of unity, whereas certain other sugars appear to be excluded from membrane-bounded intracellular compartments. The amount of solute that dissolves at equilibrium is also nearly normal in water that is constrained in gels, which are cross-linked networks of fibrous macromolecules. The diffusion of solutes within gels, however, may be hindered by collisions and interactions with the macromolecules, and by the tortuosity of the diffusion paths. Viscosity, another property of water, contributes to the resistance to flow. A fluid with a greater viscosity exhibits less flow under the influence of a given pressure gradient than a fluid with a lesser viscosity. For example, molasses, a concentrated sugar solution, is much more viscous than pure water. During laminar flow, a frictional force develops between adjacent layers (laminae) in the fluid, and this force impedes the sliding of one lamina past its neighbor. In a Newtonian fluid, the frictional force per unit area, or shear stress (z, dyn/cma), is proportional to the velocity gradient, or rate of strain ( d v / d y , s -1) between laminae,
and Foster, 1977; Horowitz and Miller, 1984) set limits on the extent to which the physicochemical properties of intracellular water differ from those of extracellular water.
III. Interactions B e t w e e n Water a n d Ions Ions in solution behave as charged, hard spheres that interact with and orient water dipoles. When crystals of sodium chloride are dissolved in water, the electrostatic attractive forces between water dipoles and ions in the crystal lattice overcome the interionic attractive forces between oppositely charged ions in the crystal. The dissociated ions then acquire the freedom of translational motion as they diffuse into the solution accompanied by a layer of tightly associated hydration water (Fig. 2). Ion is the Greek word for "wanderer." The strength of the attraction between ions and water dipoles depends to a large extent on the ionic charge and radius. For the alkali metal cations, the force of attraction, and the energy of interaction of ions with water, decreases according to the following series: Li § > Na § > K § > Rb § > Cs § Li § being the smallest of the alkali metal cations, has the strongest interaction with water because its positively charged nucleus can approach most closely to the negative side of neighboring water dipoles. As the ionic radius increases with increasing atomic number in the alkali metal series, the filled outer shells of electrons effectively shield the cationic charge and reduce the distance of closest approach to water molecules. The smallest ion thus acquires the greatest degree of hydration and has the largest hydrated ionic radius. According to Frank and Wen (1957), the orienting influence of ions on water dipoles results in three regions of water structure (Fig. 3). The electric field of an ion is sufficiently strong to remove water dipoles from the bulk water clusters and to attract to itself the oppositely charged ends of from one to five water dipoles. A certain number of these water dipoles, called the hydration number, then become
"r = rl" ( d v / d y )
where the viscosity (r/) is the proportionality constant with units of poise (= 1 dyn-s/cm2), named after Poiseuille. The viscosity of H20 at 20.3 ~ is 0.01 poise, or 1 centipoise (cp). In cultured fibroblasts, the fluid-phase cytoplasmic viscosity, as determined from rotational motions of fluorescent probes on a picosecond time scale, is only 1.2-1.4 times that of pure water (Fushimi and Verkman, 1991). Fluorescence studies also show that the viscosity is the same in the cytoplasm and nucleoplasm, and is unaffected by large decreases in cell volume or by disruption of the cytoskeleton with cytochalasin B. The fluid-phase viscosity, as determined from fluorophore rotational motions, is not nearly as affected by macromolecules as is the bulk viscosity, and it thus provides a more accurate view of the physical state of the aqueous domain of the cytoplasm. These and other studies (e.g. Schwan
FIGURE 2. Dissociationof NaC1 in water into hydrated Na§ and C1ions.
6
SECTION I Biophysical Chemistry, Metabolism, Second Messengers, and Ultrastructure
TABLE 1 Radii, E n t h a l p i e s of Hydration, a n d Mobilities of S e l e c t e d Ions mobility c 0 b A Hhydratio n
Ion
Nonhydrated radius a (A)
H§
(kcal/mol)
10 -4
cm/s
V/cm
-269
36.25
Li §
0.60
-131
4.01
Na§
0.95
-105
5.19
K+
1.33
-85
7.62
Rb§
1.48
-79
8.06
Cs +
1.69
-71
8.01
Mg 2+
0.65
-476
2.75
C a 2+
0.99
-397
3.08
S r 2+
1.13
-362
3.08
Ba2+
1.35
-328
3.30
C1-
1.81
-82
7.92
Radii are from Pauling (1960). b Standard enthalpies of hydration at 25 ~ are from Edsall and McKenzie (1978). c Mobilities in water at 25 ~ are from Hille (1992, p. 268). a
FIGURE 3. Three regions of water structure near an ion. A central cation is shown with a primary hydration shell of four oriented water dipoles, surrounded by a partially oriented secondary sheath. The gray region indicates the hydrogen-bonded structure of the bulk water. trapped and oriented in the ion's electric field. The inner hydration shell includes water molecules that are aligned by the force field and in direct contact with the ion, 5 + 1 for Li +, 4 + 1 for Na +, 3 + 2 for K + and Rb +, 4 + 1 for F-, 2 + 1 for C1- and Br-, and 1 + 1 for I-. Thus in physiological saline at 0.15 M NaC1 in water (55.5 M), about 1.6% (= 100 x 0.15 x 6/55.5) of the water is located in the inner hydration shells of Na + and of C1-. Water in the inner hydration shells of Na +, K +, and Ca 2+ rapidly exchanges with bulk water on a nanosecond time scale, but in contrast, the smaller divalent cation Mg 2+, with its high charge density, is some four orders of magnitude slower in exchanging its inner hydration water. The inner hydration sheath of water molecules moves together with the ion as a distinct and single kinetic entity. The mobilities of the alkali cations in water decrease as the nonhydrated ionic radius decreases, but as the hydrated ionic radius increases (Table 1; for discussion, see Hille, 1992). The ionic mobility (u) is defined as the proportionality constant that relates the velocity (v) of ionic migration to the force exerted by an external electric field E (v = u 9 E); in other words, mobility is the velocity per unit electric field. Farther away from the ion, where the ion's electric field falls towards zero, water retains the structure of bulk water. In the region between the inner hydration sheath and the bulk water, the orienting influences of the ion and the bulk water network tend to compete and to disrupt the structure of water. In this intermediate structure-breaking region, the water dipoles are partially oriented toward the central ion, yet do not migrate with the ion, and, they join only infrequently with the hydrogen-bonded clusters of the bulk water. If the net effect of an ion is to disorganize more water in the intermediate region than is found in the primary hydration shell, then the ion is termed a structure-breaker. Conversely, water may form a
variety of hydrogen-bonded clathrate structures around apolar protein sidechains, whose action resembles that of the class of solutes termed structure-makers (see Klotz, 1970, for review). The e n t h a l p y of h y d r a t i o n of an ion is a measure of the strength of the interaction between ions and water, and is defined as the increase in enthalpy when one mole of free ion in a vacuum is dissolved in a large quantity of water. The enthalpy of hydration may be estimated from the heat released upon dissolving salts in water, usually less than 10 kcal/mol, taking into account the energy needed to dissociate the salt crystal and then to hydrate the ions. The enthalpy of hydration may be calculated by using either the Born charging method, which estimates the energy needed to transfer a rigid charged sphere into a structureless continuum, or, alternatively, by the Bernal-Fowler structural method, which takes into account the dipolar structure of water (see Bockris and Reddy, 1970). More accurate results are obtained if water is considered to be an electric quadrupole with four centers of charge, two partial positive charges near the hydrogen nuclei and two partial negative charges on the nonbonded electron orbitals near the oxygen nucleus. A further improvement in the calculated enthalpy of hydration is obtained when the polarizability (c~) of the water dipole by the ion is taken into account. The electric field of the ion distorts the electron cloud of the hydration water along its permanent dipole axis, thus inducing an additional increment of charge separation and increasing the dipole moment. For small fields, this i n d u c e d dipole mom e n t (/zind) is proportional to the electric field strength (E), and the constant of proportionality is the polarizability (/.Lin d = O~ 9 E ) . In Fig. 4, the measured enthalpies of hydration for alkali metal cations and halides are compared with
1. Biophysical Chemistry of Physiological Solutions
7
account the formation of molecular groups with g nearestneighbors linked to the central molecule. These groups, which could be a tetrahedral group of water molecules, orient as a unit in the specific local electric fields, as distinct from an externally applied field. The Kirkwood equation is
-40 "o i
o
-~
o >:,
i
.
Rb K+I,
-80
.,
CI
Br-
(e-1)(2e+ 9e
1)
41rn
- ~
3
OL-k-
/~2( 1 +gc-b--~y) 2 3kT
-
cc-
§
Na
"0 - 1 2 0
F
"1-
llLi - 1 6 0 v__ -160
+
-120 -80 Hydration enthalpy, measured
-40
FIGURE 4. Comparison of measured enthalpies of hydration of alkali metal cations and halides in water (abscissa), with those calculated (ordinate) taking into account the Born charging energy, iondipole interactions, ion-quadrupole interactions, and ion-induced dipole interactions (data from Bockris and Reddy, 1970, p. 107).
calculated values; the impressive agreement demonstrates the primary importance of electrostatic forces in determining the solvation of ions in water. The enthalpy of hydration for the smallest alkali metal cation Li + is quite large at -131 kcal/mol. With increasing ionic radius, the enthalpy falls progressively for Na +, K +, and Rb +, reaching -71 kcal/mol for Cs +, the largest nonhydrated but smallest hydrated ion in the series (see Table 1). For the divalent, alkaline earth metal ions Mg 2+ and Ca 2+, the enthalpies of hydration also follow the ionic radius but are considerably larger a t - 4 7 6 kcal/mol and -397 kcal/mol, respectively. As pointed out by Hille (1992), the magnitude of ionic hydration enthalpies approximates the cohesive strength of the ionic bonds in a crystalline salt lattice. In 1929, the physical chemist Peter Debye explained the dielectric constant, a macroscopic property of water, in terms of its molecular properties--its dipole moment and polarizability. In a gas consisting of dipoles at concentration n, the macroscopic dielectric constant (e) is related to the polarizability (a) and the molecular dipole moment (/~) at temperature T by the Debye equation: 47rn/~ 2 e-
1 - 4~rna + 3kT
where k is Boltzmann's constant (1.38 x 10 -23 J j ~ ) . In this equation, the first term represents the effect of the polarizability (a) of the electron clouds, and the second term represents the effect of reorienting the permanent dipoles, with moments/~, in an electric field, as opposed by the randomizing influence of thermal energy (kT). Thus, increasing the temperature reduces the dielectric constant by disorienting the dipoles. In a condensed polar dielectric such as water, the Kirkwood equation for the dielectric constant takes into
where cos y is the average of the cosines of the angles between the dipole moment of the central water molecule and those of its bonded neighbors. In small spaces that restrict the formation of tetrahedral water clusters in such a way that g is reduced, the Kirkwood equation predicts that the dielectric constant will also be reduced, and electrostatic forces between ions will then be correspondingly increased. In the primary hydration shell of ions, the oriented water is polarizable but cannot be reoriented by applied fields, and so the dielectric constant is reduced from its bulk value of 78 to a value of about 6, with intermediate values in the partially oriented structure-breaking region between the primary hydration shell and the bulk water. For this reason, the dielectric constant of a solution decreases with increasing salt concentration, but in protein-free solutions at physiological salt concentrations, the extent of the decrease is small due to the small fraction of hydration water.
IV. P r o t o n s in S o l u t i o n
The hydrogen-bonded structure of liquid water also contributes to high proton mobility by a mechanism that differs fundamentally from the migration of other hydrated ions. The proton is a highly reactive, positively charged hydrogen nucleus, devoid of electrons. Whereas ions with electron shells typically have diameters in angstroms, the diameter of the proton is only about 10-5 ,~. With such a small size, the proton has a strong attraction to electrons, as indicated by the high ionization energy of 323 kcal/mol needed to remove an electron from a hydrogen atom to form a proton. By comparison, the ionization energies for the alkali cations decrease with increasing atomic number in the series as follows: Li+(124 kcal/mol) > Na + (118 kcal/mol) > K + (100 kcal/mol) > Rb + (96 kcal/mol) > Cs + (90 kcal/mol) As more filled electron shells separate and shield the outer shell from the positively charged atomic nucleus, it becomes easier to form a cation by removing an electron. The high affinity of the proton for electrons, and its small size, explain its tendency to form hydrogen bonds with the unshared electrons of oxygen in water. Infrared spectra indicate that protons in solution exist predominantly in the form of hydronium ions H3O+. Nuclear magnetic resonance (NMR) data are consistent with a flattened trigonal pyramidal structure, with O-H bond lengths of 1.02 A and H - O - H bond angles of 115 ~ a structure resembling that of NH 3 but with a radius similar to that of K +. The free energy of formation of H3O+ from H20 and H + is about-170 kcal/mol, which corresponds to a hypothetical concentration of free protons in solution at room temperature of about 10-150 M--"about as zero as one can get"
8
SECTION I Biophysical Chemistry, Metabolism, Second Messengers, and Ultrastructure
say Bockris and Reddy (1970). The heat of hydration of H3O+ is an additional -90 kcal/mol. The change in the density of water with temperature is consistent with the hydration of H3O§ with an additional three water molecules, forming the tetrahedral cluster H904 + (Bockris and Reddy, 1970). The proton mobility in water is 36 x 10-4 cm2/s 9V, much slower than expected on the basis of hydrodynamic theory, if free protons were to carry the current, but curiously about seven times as fast as expected if H3O§ migrates like K +. The abnormally high proton mobility in water (and also in ice) is consistent with a Grotthus chain mechanism (Fig. 5) in which the successive breakage and formation of hydrogen bonds, accompanied by proton jumps between neighboring water molecules, effectively results in the passage of H30+ along a chain. The transport process is rate-limited by the time needed for each successive acceptor water molecule to rotate and reorient its nonbonded orbital into a suitable position to accept the donated proton. The Grotthus mechanism accounts for about 80% of proton mobility, with the remaining 20% due to H3O§ itself, as a single kinetic entity, undergoing a translational migratory movement through the solvent like other ions.
chemical activity increase the chemical potential. The chemical potential of a solute in an ideal dilute solution, where the activity equals the chemical concentration (c), is
-/x 0
ideal
~ ' P) + R T In c
In order to describe the properties of more concentrated nonideal solutions, G. N. Lewis introduced the activity coefficient (y), such that the activity is the product of the concentration (c) and the activity coefficient (a = y - c ) . The chemical potential of an ion in a nonideal solution is tx - IX~~, P) + R T ln c + R T ln y
where the term R T In 3' includes the effect of ion-ion interactions on the chemical potential. When a salt (c+c_) is added to water, the cations and anions, at concentrations c+ and c_ respectively, both contribute to the free energy of the solution. The chemical potentials of the cations and anions (/x+ a n d / x , respectively) are given by _
0
IX+ I~+ + RTlnc+ + R T l n y+ o I~_ - IX_+ R T ln c_ + R T ln y_
Adding these two expressions, and taking the average, gives
0
V. I n t e r a c t i o n s B e t w e e n Ions 2 Interactions between ions may be weak or highly selective. Ions in solution attract ions of the opposite charge, in accordance with Coulomb's law. Since the long-range attractive forces are inversely proportional to the square of the distance between charges, the interaction energies are greater in more concentrated solutions. In a uni-univalent salt solution, defined as having monovalent cations and anions, of concentration c (mol/L), the average distance (d, angstroms or A) between any two ions is given by d - (1027/2NA c) 1/3
where N a is Avogadro's number (6.023 x 1023). The factor 2 is included because the anions and cations are both counted, and the factor 1027 is A3/L. For 0.15 M NaC1, the average distance between Na + and the nearest C1- is 17.7 * , at which range coulombic forces are highly significant. At pH 7.4, where the concentration of "protons" (really hydronium ions) is only 40 nM, the average distance between "protons" is 0.28 Ixm. Attractive ion-ion interactions are stabilizing, and they lower the chemical potential of an ion from its value in an ideal, infinitely-dilute solution. The activity (a) of an ion is defined in terms of its chemical potential (/x) as follows: tx - tx~
P)+ R T l n a
where/z~ P) is the standard-state chemical potential, or the chemical potential when the activity is 1. Increases in
0 2
The mean ionic activity coefficient y + is defined as
3/+- (,}/+ 3/_) 1/2 Mean ionic activity coefficients of salts have been estimated by the Debye-Hiickel theory, which supposes that central ions attract oppositely-charged ions (or counterions) in a diffuse and structureless ion cloud (or atmosphere). The forces of coulombic attraction between the central ion and its cloud of counterions are opposed by the randomizing influence of the thermal motion of the ions. In dilute solutions, in which the theory treats central ions as point charges relative to the size of the ion cloud, the mean ionic activity coefficient (y+) for salt ions with charges z+ and z_, is given by the limiting law as follows: log y + - - A lz+ z_l ll/2
where I is the ionic strength, defined as
1 2 I - -~ E Ciz i i The constant A, which equals 0.5108 kgl/2, mo1-1/2 is given in terms of B, which equals 0.3287 x 108 kg 1/2. mo1-1/2- cm -1, both at 25 ~ as follows"
2
A-
1 NAeo B 2.303 2 e R T
l H+,,
H\ O-H---O-H""
/H "0
H\ ~
O "
/H H--O-- .H-
+H \
~ ~-"' \ H H H/ H/ FIGURE 5. Grotthuschain mechanism for proton mobility in water.
B-
2
8"n'Na e o l O00e.kT
where N a is Avogadro's number (6.023 x 1023 ions/mol), e 0 is the charge on an electron (4.80 • 10-1~ statcoul = 1.60 • 10-19 C), E is the dielectric constant, R is the gas
1. Biophysical Chemistry of Physiological Solutions
9
constant (8.32 J/mol.K), T is the absolute temperature (K), and k (= R/NA) is Boltzmann's constant (1.38 • 10 -23 J/K). The derivation of the Debye-Htickel limiting law (see Bockris and Reddy, 1970) begins with Gauss's law, one of the four fundamental Maxwell equations of electromagnetic theory, and computes the spherically symmetric electric field around an ion, leading to Poisson's differential equation relating the charge density (Pr) of the ionic cloud to the electrostatic potential (~0r) at a distance r from the central point charge. The counterions, at concentration n i for ionic species i at a distance r from the central ion, are considered to distribute in the electric field according to a B o l t z m a n n distribution 11i
-- tl?
e-Zi eo I~tr/kT
where n lo is the bulk ion concentration (ions/L). By linearizing the Boltzmann equation, using a Taylor series expansion and retaining only the first two terms, which assumes that Zieo~tr gauche isomerizations. Since the trans --->gauche isomerization proceeds with rotations of carbon atoms about the C-C single bond, different rotational isomers (t, g+, and g-) resuiting from the rotated bond are also called rotarners or rotomers. In this section, some spectroscopic studies will be described, suggesting the simultaneous occurrence of coupled and isolated trans --->gauche isomerizations about C-C bonds in lipid acyl chains at the phase transition temperature. When a single g+ or g- rotamer is formed, the chain will make a 120 ~ bend. If this pronounced bend is introduced into the center of an acyl chain in the bilayer, it will be inhibited sterically by the neighboring chains. Hence, the single g rotamer is most likely to occur, at T > T , near the methyl ends of the two acyl chains in a C(16):C(16)PC molecule where the steric hindrance due to neighboring chains is small. If, on the other hand, a sequence of two gauche bonds separated by a trans bond (g§ or g-tg § is formed, the resulting acyl chain is then kinked in the shape of a crankshaft. The presence of such a kink leads to a largely parallel packing of the acyl chains in the bilayer. Consequently, the coupled g+tg- or g-tg + kink is favored in the center of the acyl chain at T > T . In comparison with an all-trans chain, the overall length of a kinked chain is shortened by 1.27 A along the long-chain axis. In Fig. 11D, this type of kink is graphically illustrated in each of the two acyl chains of a C(16):C(16)PC molecule.
One spectroscopic technique that has been elegantly used to examine the lipid chain order in terms of the trans ---> gauche isomerizations about the C-C bond is 2H-NMR (Seelig, 1977; Davis, 1983). Specifically, the methylene and the terminal methyl groups in the acyl chain are substituted by the CD 2 and CD 3 groups, respectively; the congested powder patterns and the corresponding de-packed spectrum with a larger number of resolvable quadrupolar splittings can be obtained from a single 2H-NMR experiment at T > T (Sternin et al., 1983). The experimentally determined quadrupolar splittings can be converted to the order parameter (Sco) of the C-D bond (Seelig, 1977). Since the order parameter, Sco = 0.5, is a time-averaged parameter of the angular fluctuation (/3) of the C-D bond relative to the bilayer normal on the 2H-NMR time scale (< 10 -5 s), the quantity IScDI can be used to infer the possible presence of gauche rotamers in the acyl chain at T > T . For instance, if the acyl chain is assumed to adopt an all-trans configuration and is simply reorienting about the long-chain axis, then/3 = 90 ~ and IScDI - 0.5. If, on the other hand, the chain is assumed to undergo an isotropic motion over all space, this motional averaging will then lead to IScDI - 0. The trans --->gauche isomerizations of the CD 2 groups around the C-C bonds will undoubtedly result in a partially disordered orientation of the acyl chain, which can be expected to give rise to IScDI values in between the two limiting cases (0.5 and 0). The IScDI values of the C-D bonds along the acyl chain of C(16):C(16)PC at T > T are found experimentally to be approximately constant and are, indeed, equal to about 0.2 over the initial segment near the interface; this pan is usually referred to as the plateau region. A rapid decrease in IScDI is observed near the methyl terminus, indicating that the motional averaging of the quadrupolar interactions increases towards the middle of the bilayer. The shape of this orientational order profile can be rationalized by'a structural model of the acyl chain at T > T as follows: the coupled g+-tgTkinks are allowed in the plateau region, whereas the isolated g- rotamers exist predominantly near the terminal methyl end. The relative ratios of trans and gauche rotamers in the lipid acyl chains at various temperatures can be detected directly by Raman spectroscopy (Levin, 1984). This method has indeed demonstrated that the gel ---> liquid-crystalline phase transition exhibited by the bilayer of C(16):C(16)PC is accompanied by an abrupt increase in the ratio of gauche/trans at the onset of Tm (Gaber and Peticolas, 1977). Moreover, single g• rotamers as well as g + t g - a n d g-tg + kinks have been detected by FT-infrared spectroscopy in the bilayers of C(16):C(16)PC at T > T (Mendelsohn and Senak, 1993; Casal and McElheney, 1990). Most interestingly, these single g• rotamers and coupled g+-tg -v- kinks are also observed in the computer-simulated C(14):C(14)PC bilayer, in excess water, at T > T using the molecular dynamics method (Chiu et al., 1995). Moreover, the g+-tgu kinks shown by molecular dynamics simulations are not fixed at any given preferred positions along the sn-1 and sn2 acyl chains of C(14):C(14)PC molecules in the liquidcrystalline bilayer. Based on 2H-NMR data, it has been suggested earlier that the coupled g• -v- kinks migrate up and down along the plateau regions of lipid acyl chains in bilayers at T > T (Seelig and Seelig, 1980).
55
3. Structural Organization and Properties of M e m b r a n e L i p i d s
Based on the spectroscopic and computational results just discussed, it is obvious that the thickness of the lipid bilayer composed of saturated identical-chain phosphatidylcholines must be reduced at T > T due to the presence of g-+ rotamers and g+-tgu kinks along the acyl chain. In addition, the lateral chain-chain van der Waals interactions must also be diminished at T > T due to the random distribution and the dynamic nature of the g+-tgu kinks. Consequently, the average rate of lateral diffusion for saturated identical-chain phosphatidylcholine molecules in each of the two monolayers within the two-dimensional plane of the lipid bilayer can be expected to increase appreciably as the bilayer undergoes the thermally induced gel-to-liquid-crystalline phase transition. Indeed, it has been shown by using fluorescent lipid molecules that an averaged lateral diffusion coefficient of about 10-8 cm2/s is observed in the two-dimensional plane of the liquid-crystalline bilayer, and a much smaller value on the order of 10-11 cm2/s is estimated for the same fluorescent lipid molecules in the gel-state bilayer (Clegg and Vaz, 1985).
of the longer of the two acyl chains in C-C bonds, and is defined as CL = X - 1 if the sn-1 acyl chain is the longer chain or as CL = Y - 2.5 if the sn-2 acyl chain is the longer chain. In the latter case, the effective chain length extends from the point corresponding to the position of carbonyl carbon of the sn-1 acyl chain to the methyl terminus of the sn-2 acyl chain. This structural parameter, CL, is also illustrated schematically in Fig. 12A. For C(18):C(14)PC, the CL value is 17 C-C bond lengths. The normalized chain-length difference or asymmetry is the ratio of AC/CL. For C(18):C(14)PC, the AC/CL value is 0.32. Another structural parameter, N, is taken to be the distance, in C-C bond lengths, between the two carbonyl oxygens of the sn-1 acyl
A C(18):C(14)PC [
C. Factors Affecting the Gel-to-Liquid-Crystalline Phase Transition Temperature The phase transition behavior of lipid bilayers composed of a given species of pure diacyl phospholipids can be influenced by many internal and external factors. Internal factors refer to changes in chain length, chain asymmetry, unsaturation, and headgroup structure. External factors include pressure, pH, and chemical composition of the medium. Of the two thermodynamic parameters (Tm and ~ that are associated directly with the main phase transition of the lipid bilayer, the value of Tm can be determined with greater accuracy by high-resolution DSC. We shall examine how some of the internal and external factors affect the Tm value associated with the thermally induced main phase transition of diacyl phospholipids in excess water. Before we discuss the effects of chain length and chain asymmetry on T , let us first define some structural parameters (AC, AC/CL, and N) underlying saturated diacyl phosphatidylcholines or C(X):C(Y)PC packed in the gelstate bilayer. To proceed, we return for a moment to Fig. 5C in which the effective chain-length difference between the sn-1 and the sn-2 acyl chains of an identical-chain phosphatidylcholine in the crystalline state, ACcry, can be clearly identified. This value of AC c r y is seen in Fig " 5C to be 3 "68 C-C bond lengths along the chain. For the same identicalchain phosphatidylcholine molecule packed in the gel-state bilayer, the effective chain-length difference is reduced to about 1.5 C-C bond lengths, and we designate it as the chain-length difference of the reference s t a t e (ACref). For a saturated mixed-chain C(X):C(Y)PC such as C(18):C(14)PC in the gel-state bilayer, the effective chain-length difference AC is related to X and Y as follows (Huang et al., 1994): AC - I X - Y+ ACref - 18 - 14 + 1.51 = 5.5 C-C bond lengths. Schematically, this structural parameter is presented in Fig. 12A. Here, a crystalline structure is conveniently drawn for C(18):C(14)PC; actually, all structural parameters of C(X):C(Y)PC are defined for lipid molecules packed in the gel-state bilayer with shorter chain lengths. A second structural parameter, CL, refers to the effective chain length
CL = X - I
[
I
I
J TI|
~ I
I I I1 I~-------
AC
ac = Ix-Y + 1.51
L
B Trans-bilayer dimer of C(18):C(14)PC I ]
N -- X + Y - 0 . 5
i
I
0
9 ~LJI,.~.K..K,..~.,t..,L.K.
,,.,•
i, J
C(18):C(18:IA'2)pc
I
FIGURE 12. Molecular graphics representations of phospholipids with various structural parameters. (A) A monomer of saturated C(18):C(14)PC. CL is the effective chain length of the longer of the two acyl chains, and AC is the effective chain-length difference between the two acyl chains along the long-chain axis. For C(X):C(Y)PC in the gel state, the relations between CL, AC, and X, Y are indicated. In the case of gel-state C(18):C(14)PC, CL and AC are 17 and 5.5 carbon-carbon bond lengths, respectively. (B) A molecular graphics drawing to illustrate a transbilayer dimer of C(18):C(14)PC with a partially interdigitated packing motif. N is the hydrophobic thickness of the dimer, corresponding to the distance separating the two carbonyl oxygens of the two opposing sn1 acyl chains. VDW is the van der Waals contact distance between two opposing methyl groups in the bilayer interior. (C) The monomeric C(18):C(18:lA12)pc. The sn-2 acyl chain is shown to adopt the crankshaftlike motif; hence it consists of two segments. The chain-chain lateral interaction is assumed to be energetically more favorable between the sn-1 acyl chain and the longer segment of the sn-2 acyl chain.
56
SECTION I Biophysical Chemistry, Metabolism, Second Messengers, and Ultrastructure
chains in a trans-bilayer dimer of C(X):C(Y)PC along the long molecular axis as shown schematically in Fig. 12B. The N value of a C(X):C(Y)PC dimer is related to X and Y as follows" N = ( X - 1) + VDW + ( Y - 2.5) = X + Y - 0.5, where VDW is the van der Waals separation distance between the two opposing terminal methyl groups in the transbilayer dimer (Fig. 12B), which is taken to be 3 C-C bond lengths in the gel-state bilayer. When the normalized chain asymmetry ratio, AC/CL, is less than 0.42, saturated C(X):C(Y)PC molecules in excess water at T < T can self-assemble into the partially interdigitated bilayer, in which the methyl end of the longer acyl chain of one lipid molecule from one leaflet packs end-toend with the methyl end of the shorter acyl chain of another lipid molecule from the opposing leaflet and vice versa (Li et al., 1993). This partially interdigitated packing motif is diagrammatically shown in Fig. 12B. If the AC/CL value of a phosphatidylcholine molecule is greater than 0.42, this highly asymmetric lipid tends to self-assemble into the mixed interdigitated bilayer at T < T (Li et al., 1993). Phosphatidylcholines with AC/CL values greater than 0.42 are rare in biological membranes; hence, they are not discussed in this chapter. At AC/CL value less than 0.42, DSC measurements on saturated C(X):C(Y)PC have demonstrated that the Tm value of the main phase transition is related to the structural parameters AC and N according to the following equations (Huang et al., 1994; Huang and Li, 1999): T - 1 6 1 . 7 5 - 3 7 0 6 . 0 6 - ~1- - 2 7 8 . 7 5 AC+239.94 N A-------~ N +AC (1) for lipids with a longer effective sn-1 acyl chain, and T - 1 5 5 . 1 1 - 3 5 3 4 . 3 1 ~1- - 2 4 5 . 7 8 AC+199.04 N A'-------~ N +AC (2) for lipids with a longer effective sn-2 acyl chain. Implicitly, these two equations indicate that if AC is constant, the T value increases with increasing N; however, the increase tends to level off towards an asymptote. In contrast, the T value decreases with increasing AC if N is constant; as a resuit, AC counteracts the N-promoted increase in Tm. Thus, it is the delicate balance between the two structural parameters (N and AC) underlying the dimeric C(X):C(Y)PC molecules in the gel-state bilayer that determines the unique T m value of the fully hydrated lipids. Equations 1 and 2 not only show the fundamentally antagonistic effect between N and AC in determining the T value, but also can be employed to predict the unknown value of T based on the chemical formula of C(X):C(Y)PC (Huang et al., 1994; Huang and Li, 1999). Table 3 lists some representative T values obtained calorimetrically with multilamellar vesicles containing diacyl phospholipids. The values of AC and N calculated according to AC - ] X - Y + 1.51 and N = X + Y - 0 . 5 are also listed for those saturated lipid species. For the homologous series of saturated identical-chain phosphatidylcholines from C(14):C(14)PC to C(20):C(20)PC with a common AC value of 1.5 C-C bond lengths, the T m value is seen to increase progressively with a step-wise increase in N which, in fact, is directly related to the chain length. This raises the in-
teresting question of why the T m value increases progressively with an increase in the chain length of the saturated lipid molecule. The answer lies in the strength of the lateral chain-chain van der Waals interaction in the gel-state bilayer. This energetic term is directly proportional to the chain length. Specifically, the lateral van der Waals interaction between two all-trans saturated chains (EvDw) is related to the number of methylene units in the chain (Nm) according to the following equation (Salem, 1962):
3~Nm
EVDw-A ~
8lD 5
(3)
where l is the distance between two methylene units along the long-chain axis (1.27 ~), D is the lateral distance separating the two all-trans chains, and A is a constant (-1340 kcal/A -6 mol). If D is taken to be 4.8 ~, then Evo w = - 0 . 4 9 kcal/mol per methylene unit. Clearly, the longer the acyl chain or the larger the number of the methylene units (Nm), the greater is the attractive van der Waals interaction between the two saturated acyl chains. As a result, a larger thermal energy and a higher Tm are required to promote the trans ---) gauche isomerizations within the longer chains as the phospholipid molecules in the bilayer undergo the thermally induced gelto-liquid-crystalline phase transition. In Table 3, C(16):C(16)PC, C(14):C(18)PC, and C(18): C(14)PC are shown to have a common N value of 31.5 C-C bond lengths. The AC and Tm values of these saturated lipids are also shown to be 1.5, 2.5, and 5.5 C-C bond lengths and 41.5, 39.2, and 31.2 ~ respectively. This set of saturated C(X):C(Y)PC thus serves to demonstrate that the Tmvalue decreases with increasing AC when the N value is held constant. Similarly, another pair of the position isomers, C(16):C(18)PC/C(18):C(16)PC, exhibits the same inverse relationship between Tm and AC (Table 3). Fundamentally, the structural parameter AC represents the chain asymmetry. It is also a perturbation term that affects the closest lateral chain-chain interaction in the gel-state bilayer. This perturbation stems from the bulky methyl terminal groups. Most of the naturally occurring phospholipids in eukaryotic cell membranes have a saturated sn-1 acyl chain and an unsaturated sn-2 acyl chain with different numbers and positions of cis double bonds. It is very convenient to abbreviate the names for these enormously complicated sn-1 saturated/sn-2 unsaturated phospholipid species according to the chemical natures of their two acyl chains. These abbreviated names, as specified in Table 3, are used throughout the rest of this chapter. Now, we can proceed to discuss the effects of unsaturation of the sn-2 acyl chain on the gel-to-liquidcrystalline phase transition temperature of phospholipid bilayers as monitored by high-resolution DSC. First, the introduction of a single cis double bond into the sn-2 acyl chain markedly reduces the T m value of lamellar phosphatidylcholines. A good example is demonstrated by the bilayer composed of 1-stearoyl-2-oleoyl-phosphatidylcholines or C(18):C(18" 1A9)pc, which exhibits calorimetrically the T value of 5.6 ~ In contrast, the T value of 55.3 ~ is detected for the bilayer composed of the saturated counterparts or C(18):C(18)PC. Second, the T-lowering effect of acyl chain monounsaturation depends critically on the position of
3. Structural Organization and Properties of Membrane Lipids TABLE 3
57
The Main Phase Transition Temperature for Some Diacyl Phospholipids
Lipida
AC
N
1.5 1.5 1.5 1.5 2.5 5.5 0.5 3.5 3.5
27.5 31.5 35.5 39.5 31.5 31.5 33.5 33.5 37.5
Tm (~
Lipida
Tm (~
Saturated phospholipids C(14):C(14)PC C( 16):C( 16)PC C(18):C(18)PC C(20):C(20)PC C(14):C(18)PC C(18):C(14)PC C(16):C(18)PC C(18):C(16)PC C(20):C( 18)PC
24.1 41.5 55.3 66.4 39.2 31.2 48.8 44.4 57.5
C(14):C(14)PE C(16):C( 16)PE C(18):C(18)PE C(20):C(20)PE C(14):C(18)PE C(18):C(14)PE C(16):C(18)PE C(18):C(16)PE C(20) :C(18)PE
49.6 63.2 74.0 82.5 61.6 54.9 70.8 66.0 76.3
C(16):C( 18:1A6)pE C( 16):C( 18:1A9)pE C( 18):C( 18:1A6)pE C( 18):C(18:1Av)PE C( 18):C( 18:1A9)pE C(18):C(18:1A11)pE C(18):C(18:1A 12)pE C(18):C(18:1A 13)pE C( 18):C(18:2A9,12)pE C(18):C(18:2A9,12,15)p E
39.1 26.1 42.6 35.4 31.5 29.8 32.3 35.4 4.4 m
Unsaturated phospholipids C( 16):C( 18:1A6)pc C( 16):C( 18:1A9)pc C( 18):C( 18:1A6)pc C( 18):C( 18:1A7)pc C(18):C( 18:1A9)pc C(18):C(18:1All)pc C( 18):C( 18:1A12)pc C( 18):C( 18:1A13)pc C( 18):C( 18:2 A9,12)pc C(18):C(18:2A9,12,15)PC
18.8 -2.6 24.8 16.7 5.6 3.8 9.1 15.9 -14.6 -12.2
aSaturated phosphatidylcholines and phosphatidylethanolamines are abbreviated as C(X):C(Y)PC and C(X):C(Y)PE, respectively. X denotes the total number of carbons in the sn-1 acyl chain and Y denotes the total number of carbons in the sn-2 acyl chain, sn-1 saturated and sn-2 monounsaturated acyl chains with the cis-double bond (A) at the position n from the carboxyl end abbreviated as C(X):C(Y:IAn); similarly, lipids with sn-1 saturated and sn-2 diunsaturated acyl chains are abbreviated as C(X):C(Y:2An'm),where m is the second cis-double bond position at the mth carbon atom from the carboxyl end.
the cis carbon-carbon double bond ( A n) in the sn-2 acyl chain. As shown in Table 3, the minimal Tm for a homologous series of C(18):C(18"IAn)pc occurs at A n "- m 11 or when the cis double bond is positioned near the center of the linear segment of the sn-2 acyl chain. Moreover, the T value increases progressively as the single cis double bond moves from A 11 toward either end of the acyl chain. Third, the progressive increase in the level of acyl chain unsaturation has a nonadditive effect on the Tm. Based on the Tm values of C(18):C(18)PC, C(18):C(18:IA9)pc, C(18):C(18: 2A9,12)pc, and C(18):C(18:2A9,12,15)pc shown in Table 3, it is evident that the introduction of the first cis double bond near the chain center gives rise to a drastic decrease of 49.7 ~ in Tm. Although the introduction of a second cis double bond produces a substantial decrease of 20.2 ~ in T , this decrease is not nearly as large as the first one. Most interestingly, the successive incorporation of a third cis double bond at ~15 results in a slight increase of 2.4 ~ in T . Let us now consider a molecular model that has been invoked to interpret the large T - l o w e r i n g effect of acyl chain monounsaturation and the characteristic dependency of T m on the position of the cis double bond along the acyl chain (Wang et al., 1995). This molecular model makes three basic assumptions. (1) The monoenoic sn-2 acyl chain in the sn-1 saturated/sn-2 monounsaturated phospholipid molecule is as-
sumed to adopt, at T < Tm, an energy-minimized crank-shaftlike kink motif as shown in Fig. 12C. The sn-2 acyl chain thus consists of a longer segment and a shorter segment separated by the An-containing kink. (2) The longer segment and the neighboring sn-1 acyl chain run in a nearly parallel manner with favorable van der Waals attractive distance between them. (3) The shorter segment is considered to be partially disordered at T < Tm, thus playing a perturbing role in the lateral chain-chain interactions in the gel-state bilayer. With this molecular model, we are now in a position to interpret first the drastic Tm-lowering effect of a single cis double bond. Because the shorter segment is a perturbing element, it behaves similar to the structural parameter AC to counteract the effect of N, thus lowering Tm. Moreover, since this short segment is already partially disordered at T < Tm, it cannot contribute significantly to the conformational disordering process of trans --~ gauche isomerizations involved in the phase transition. As a result, the total number of trans ~ gauche isomerizations during the phase transition is decreased appreciably for the bilayer composed of cis-monoenoic phosphatidylcholines relative to that of the saturated counterparts, leading to a significantly lower T value. We can use the same molecular model to see how the T m value varies as the single cis bond migrates along the sn2 acyl chain. Since the longer segment of the kinked chain
58
SECTION I B i o p h y s i c a l Chemistry, M e t a b o l i s m , S e c o n d Messengers, and Ultrastructure
is assumed to undergo a favorable van der Waals contact interaction with the sn-1 acyl chain, this contact interaction energy must then depend on the length of the longer segment. When the single cis double bond is positioned at the center of the sn-2 acyl chain, the longer segment of the kinked chain has a minimal length, which is almost equal to that of the shorter segment. Hence, the van der Waals interaction with the sn-1 acyl chain is also minimal. As the single cis double bond migrates away successively from the chain center towards either end, the length of the longer segment is progressively increased, leading to a proportionally increased van der Waals interaction and hence a gradual increase in Tm. The characteristic change of T as a function of the location of the cis double bond, shown in Table 3, can thus be explained by the structural model. The phospholipid headgroup can also exert a profound effect on the phase transition behavior of the lipid bilayer. For instance, when the headgroup's choline moiety in a saturated identical-chain phosphatidylcholine molecule is replaced by ethanolamine, the resulting phosphatidylethanolamine can self-assemble into the gel-state bilayer in excess water after incubating at T < T for a brief time. Upon heating, the gelstate PE bilayer transforms directly into the liquid-crystalline ( L ) phase without going through the intermediate Pt~' phase that is commonly observed for fully hydrated PC with identical chains. Hence, the main phase transition of saturated identical-chain PE corresponds to the L/3 --->L phase transition, where Lt~ denotes the gel phase with the saturated acyl chains orienting perpendicular to the bilayer surface. For comparison, the second DSC heating scans obtained with aqueous dispersions of C(16):C(16)PC and C(16):C(16)PE are shown in Fig. 13. Clearly, the Tm value is significantly higher for PE than for PC with the same saturated fatty acyl chain composition, although their AH values are comparable. In Table 3, the T values for nine species of saturated diacyl PE as well as PC are given. The T values for saturated PE are consistently higher than those for PC with the same fatty acyl chain composition. However, for the homologous series of identical-chain phospholipids from C(14):C(14)PE/PC to C(20):C(20)PE/PC, the difference in T between PE and PC (ATm) decreases with increasing acyl chain length (Table 3). Figure 13 also shows the second DSC heating scans for aqueous dispersions of C(16):C(18"IA6)pc and C(16):C(18: 1A6)pE; here, the T value of PE is 20.3 ~ higher than that of PC. In Table 3, the T values of a large number of sn- 1 saturated/sn-2 unsaturated PC/PE are presented. Clearly, the T values of unsaturated PE are also consistently higher than those of unsaturated PC with the same fatty acyl chain composition. These higher T values can thus be taken as evidence to suggest that, in general, the lateral chain-chain interactions in the PE bilayer are stronger than those in the PC bilayer at T < T . This stronger lateral chain-chain interaction may be attributed to many factors such as the smaller PE headgroup, the interlipid H-bond networks on the two surfaces of the gel-state PE bilayer, the smaller hydration of the PE bilayer, and so on. At the present time, however, the quantitative contribution of each factor to the overall difference in Tm between PE and PC is not well understood and deserves further investigation.
C(16):C(16)PC
C(16):C(16)PE
Tm= 41.5~ AH = 8.5 kcal/mol
T,, = 63.2~ AH = 8.3 kcal/mol
A Tv2 = 0.3~
' 20
3'0
' 40
ATl/2 = 0.7~
l
]
I
I
50
60
40
50
C(16):C(18:1z~ 6 )PC
|
|
60
|
70
80
C(16):C(18:1AS )PE
T,, = 18.8~ ~ H = 6.0 kcal/mol
AH =
z~ T1,2 = 0.5~
Z~L/2 = 0.8 ~
"I'm = 39.1~ 6.3 kcal/mol
______
,'0
2'0
3'0
2'0
3'0
,'o
5'0
Temperature (~ RepresentativeDSC second heating scans for aqueous dispersions of PC and PE with the same saturated and monounsaturated acyl chain compositions.The T values of PE bilayers are observed to be consistently higher than those of PC bilayers with the same fatty acyl chain compositions irrespective of acyl chain unsaturation. These DSC experiments were performed with a Microcal Model MC-2 differential scanning microcalorimeter (Northampton, MA). Scan rate: 15 ~ FIGURE 13.
Based on experimental data obtained with 33 molecular species of C(X):C(Y)PE, two Tm equations similar to Eqs. 1 and 2 for C(X):C(Y)PC have also been derived. As a result, for PC and PE with the same sn-1 and sn-2 acyl chains, the Tm value of one lipid can be converted to the other via the common structural parameters. Specifically, the difference in T between PE and PC ( A T ) can be calculated based on their common values of N and AC according to the following equations (Huang and Li, 1999): ATPE-PC- TPE_ T PC m
m
m
AC (4) 1 AC =-2.05 + 719.03 ~- - 39.49--~- + 62.22 A--------~ U+ for lipids with an effective longer sn-1 acyl chain, and ATPE-PC- TPE_ T PC m
m
m
= 2.77 +584.4
AC AC -l10.46-~+151.01N+AC
(5)
for lipids with an effective longer sn-2 acyl chain. Using Eqs. 4 and 5, the Tm value of a given C(X):C(IOPC bilayer, Tm Pc, can be calculated from the experimental Tm ~
3. Structural Organization and Properties of Membrane Lipids
value of the corresponding C(X):C(Y)PE bilayer, TPmz. Likewise, the unknown TPmE value can also be calculated, provided that the value of TPc is known experimentally. The two structural parameters, N and AC, associated with C(15):C(15)PC or C(15):C(15)PE packed in the gel-state bilayer, for example, are 29.5 and 1.5 C-C bond lengths, respectively. With these numbers, the ATmPE-PC term in Eq. 4 can be calculated to be 23.3 ~ The Tm value of 34.0 ~ is known calorimetrically for the C(15):C(15)PC bilayer; consequently, the phase transition temperature underlying the gel-to-liquid-crystalline phase transition for lamellar C(15):C(15)PE can be estimated as follows: TPE= ATmPE-PC+ Tm Pc= 57.3 ~ This calculated value has indeed been verified by DSC experiments. One external factor that has been extensively studied is the pressure (Wong, 1986). Take the C(16):C(16)PC dispersions at 47 ~ as an example. The lamellar PC in the L phase can undergo two isothermal phase transitions upon elevation of pressure up to 1.4 kbar. The first pressure-induced isothermal transition occurs at the critical pressure of 0.17 kbar, and the second occurs at 0.60 kbar. These two phase transitions correspond to the cooling-induced main and pretransitions, respectively. Clearly, an increase in pressure acts in a manner similar to a decrease in temperature by exerting an ordering effect on the acyl chain packing in the bilayer.
59
DSC heating curves for a given binary phospholipid system at various relative compositions. Basically, the onset and completion temperatures for each transition peak, after appropriate correction, are plotted as a function of the mole fraction of the higher melting component lipid (XB). These corrected onset and completion temperature points form the bases for defining the solidus and liquidus, respectively, of the temperature-composition phase diagram (Mabrey and Sturtevant, 1976). Of all the known phase diagrams constructed on the basis of calorimetric data obtained with binary phospholipids, the mixing behavior of the two lipid components in the bilayer plane is found to be always nonideal. This implies that the phase transition of the component A (or B) is affected by the
A
~ ,
V. Binary P h o s p h o l i p i d Mixtures
I
G |
i
v
v
....
B :3
Thus far the structure and the phase behavior of a limited number of one-component phospholipid systems have been discussed. Attention is now directed to the mixing behavior of two-component (A and B) phospholipid systems at ambient pressure. The mixing behavior of two component lipids in the bilayer is expected to depend on the difference in the pair-interaction energies between the mixed pairs (A-B) and the like pairs (A-A and B-B) which, in turn, depends on the extent of structural similarity between the two component lipids. Furthermore, the structural characteristics of A and B in the bilayer may change abruptly and significantly as the two-component bilayer undergoes the respective phase transition of components A and B. The relative pair-interaction energy may thus be quite different in the gel and liquid-crystalline states. For instance, the lateral phase separation of a given two-component system may occur in the gel phase due to their structural dissimilarity, whereas a complete miscibility between the same two components may take place in the L phase as a result of conformational changes of the component lipids occurring at the T . The mixing behavior and the interaction energies related to the component lipids in the gel and liquid-crystalline phases are fully contained in the temperature-composition phase diagram for binary lipid systems. In this section, we consider three of the most commonly observed temperature-composition phase diagrams for binary lipid systems. These phase diagrams are shown diagrammatically in Fig. 14A-C. In particular, each phase diagram is characterized by its position and shape, which are determined by the phase boundaries called the solidus and liquidus as shown in Fig. 14A. Experimentally, these phase boundaries are commonly constructed from the
L
It. E
i
C
S
I
|"
L
L+G
B
..I Gd 3+ > Be 2+ > Ca 2+ = Ba 2+ > Sr 2+ > Mg 2+. Of these, Mg 2+ is the only cation likely to have any impact on Ca 2+ receptor activity under physiological conditions. In addition to inorganic polycations, many organic polycations also activate the Ca 2+ receptor; these include polyamines (spermine), aminoglycoside antibiotics (neomycin), polyamino acids (polylysine), and proteins (protamine). In general, there is some correlation between net positive charge of these compounds and their potency in activating the Ca 2+ receptor. Nonetheless, there are exceptions to this rule that indicate that the Ca 2+ receptor is sensing not just net positive charge. Like Ca 2+, all these polycations are believed to act largely, if not exclusively, in the extracellular domain of the Ca 2+ receptor. It is conceivable that regions of the receptor containing a high density of acidic amino acids, and therefore a high density of negative charge, are involved in the binding of these inorganic and organic polycations. The organic polycations are not useful tools for investigating the cellular physiology of parathyroid cells or other extracellular Ca2+-sensing cells because they lack the potency and specificity necessary for many in vitro studies and all in vivo studies. A structurally distinct class of compounds, typified by NPS R-467, is used in superior pharmacological probes that overcome these problems. These phenylalkylamine derivatives carry only one positive charge at physiological pH and, at nanomolar concentrations, act selectively on the Ca 2+ receptor to inhibit PTH secretion (Nemeth et al., 1998). These compounds act in a stereoselective manner at the Ca 2+ receptor, and the R-enantiomers are 10- to 100-fold more potent than the corresponding S-enantiomers. The mechanism of action of the phenylalkylamine compounds differs from those of the inorganic and organic polycations. Unlike the
184
SECTION I B i o p h y s i c a l Chemistry, M e t a b o l i s m , S e c o n d Messengers, and Ultrastructure
polycations, these phenylalkylamine compounds require extracellular Ca 2§ for their activity. Thus these compounds fail to mobilize intracellular Ca 2+ in the absence of extracellular Ca 2§ although a polycation such as neomycin is fully competent in doing so. These phenylalkylamines behave as positive allosteric modulators to increase the sensitivity of the Ca 2§ receptor to activation by extracellular Ca 2§ thereby shifting the concentrationresponse curve to the left. These different mechanisms are also manifest in the concentration-response curves to polycations and to phenylalkylamines. The curve for polycations is extremely steep and apparently reflects cooperative binding of the ligand to the receptor, whereas that for phenylalkylamines is more conventional and suggests a 1:1 ligand:receptor complex. The phenylalkylamines bind in the transmembrane domain of the Ca 2+ receptor (Hammerland et al., 1999). Compounds that mimic or potentiate the actions of extracellular Ca 2§ at the Ca 2§ receptor have been termed calcimimetics. As such, they behave as receptor agonists to mobilize intracellular Ca 2§ and inhibit PTH secretion. Two types of calcimimetics can be distinguished: Type I, exemplified by the inorganic and organic polycations, which bind largely in the extracellular domain and mimic extracellular Ca 2+, and Type II, typified by the phenylalkylamines, which bind within the transmembrane domain of the Ca 2§ receptor and potentiate the actions of extracellular Ca 2§ Type II calcimimetics have been used to activate the Ca 2+ receptor in vivo and, when administered to animals or humans, cause time- and dose-dependent decreases in
plasma levels of PTH and Ca 2+. Circulating levels of PTH start to fall sooner than those of Ca 2+, as expected, since it is this decrease in PTH that leads to hypocalcemia (Fig. 4). These findings, together with those from molecular genetic studies, provide solid evidence that the Ca 2+ receptor is the essential regulator of systemic Ca 2+ homeostasis. More recently, potent and selective small organic antagonists of the Ca 2+ receptor have been developed. These compounds have been termed calcilytics and are the only substances known that antagonize the Ca 2§ receptor. Calcilytic compounds stimulate secretion of PTH in vitro and in vivo (Fig. 5) and, together with the Type II calcimimetics, are pharmacological tools possessing the requisite potency and selectivity necessary to explore the role of the Ca 2§ receptor in physiological processes beyond systemic Ca 2§ homeostasis.
IV. Calcium R e c e p t o r - D e p e n d e n t of Cellular F u n c t i o n s
Regulation
There are a number of cells throughout the body whose activity is regulated by changes in the concentration of extracellular Ca 2§ For many of these cells, it is not certain that the effects of extracellular Ca 2§ are mediated by a parathyroid-like Ca 2§ receptor. However, there is clear evidence for the expression of this receptor in certain cells and some understanding of how this receptor regulates cellular responses. Discussed next are the three cell types that have furnished most of the information about how the Ca 2§ receptor regulates cellular activity.
A. Parathyroid Cells These are the classic cells long known to be responsive to small changes in the concentration of extracellular Ca 2§ For -.[ v CH 3
-OCH3
25 150 --0--
PTH
9..O..- Ca2*
"8
400
110
2
c 100
100 * 8
~E
C
8e
200 -~
a.
of) (0 ~.~
so
90
tO
O.
lOO o
g
g o
~
.....
i ,2o
1do
~o
8o
"I-
'.T."
2.
0
,
o
o'.i
i
~ "
o ~2
1'o
16o
0
NPS R-568, mg/kg Time (rain)
FIGURE 4.
Activation of the Ca 2+ receptor in vivo lowers p l a s m a lev-
of PTH and Ca2§ Normal rats were given 10 mg/kg NPS R-467 orally at time 0. There is a rapid fall in plasma levels of PTH followed by a decrease in plasma Ca2§ levels. Both parameters return to normal over several hours. Insert: The structure of NPS R-467, the R-enantiomer. els
FIGURE 5. Preferential effects of a Type II calcimimetic compound on plasma levels of PTH. Orally administered NPS R-568 (the chloro-analog of NPS R-467) was about 40-fold more potent at depressing secretion of PTH than at stimulating secretion of calcitonin. (From Fox et al., 1999, Journal of Pharmacology and Experimental Therapeutics, 290: 480)
11. Regulation of Cellular Functions b y Extracellular Calcium
80 -~
185
Maximum
oint"
E Q. v
"1IX.
o
40-
......................................
E
r
n
Minimum O-
I
1.0
T I
1.2
1.14
Plasma Ca 2+ (mM)
FIGURE b. Smallchanges in the plasma levels of Ca 2+ cause large changes in the plasma levels of PTH. In the normal adult human, plasma levels vary between 10 and 65 pg/ml and the plasma Ca 2+ levels are maintained between 1.12 and 1.23 mM. The concentration of plasma Ca 2+ causing a 50% decrease in the maximal levels of plasma PTH (achieved by lowering extracellular Ca 2+ until no further increase occurs) is called the "set point" for extracellular Ca 2+. The Ca2+receptor determines the set point and thereby functions as a "calciostat" to maintain systemic Ca 2+ homeostasis.
these cells, extracellular Ca 2+ is the primary physiological stimulus regulating secretion of PTH. The sensitivity of the parathyroid cell to the ambient Ca 2+ concentration is impressive: minimal and maximal rates of PTH secretion are obtained over a concentration difference of only 0.5 mM (range of 1.0 to 1.5 mM). The concentration of serum Ca 2+ that suppresses maximal PTH levels by 50% is called the "set point" for extracellular Ca 2+ (Brown, 1983), and it is usually around the normal serum level of 1.2 mM (Fig. 6). Because of the steep concentration-response relationship between extracellular Ca 2+ and PTH, very small changes in the concentration of extracellular Ca 2+ cause large changes in the secretion of PTH (see Fig. 6). The Ca 2+ receptor on parathyroid cells is coupled to the activation of phospholipase C through a G protein (Fig. 7). The identity of this G protein is not yet certain, although it is believed to be either G ll or G . The G protein linking the Ca 2+ receptor to phospholipase ~ is not inhibited by pertussis toxin. Activation of phospholipase C results in the rapid formation of IP 3, leading to the mobilization of Ca 2+ from some nonmitochondrial intracellular store, presumably some component of the endoplasmic reticulum. There is additionally an influx of extracellular Ca 2+ through voltage-insensitive channels. Thus when the concentration of extracellular Ca 2+ is increased, there is a rapid and transient increase that is followed by a lower yet sustained increase in [Ca2+]i (Fig. 8). It is not at all clear how the influx of extracellular Ca 2+ in parathyroid cells is regulated. Many Type I calcimimetics are just as efficacious as extracellular Ca 2+ in evoking the mobilization of intracellular Ca 2+, yet they fail to promote
the influx of extracellular Ca 2+ (see Fig. 8). Thus, activating the Ca 2+ receptor is not sufficient to trigger influx of extracellular Ca 2+. Presumably, the Ca 2+ receptor is not directly linked to an influx channel. The influx of extracellular Ca 2+ might be regulated by an intracellular signal or by depletion of intracellular stores ("capacitive" Ca 2+ influx; Petersen, 1996; see Chapter 55). The Ca 2+ receptor is also coupled to adenylate cyclase through a pertussis toxin-sensitive Gi-like protein. Type II calcimimetics potentiate the effects of extracellular Ca 2+ on cyclic AMP levels, [Ca2+]i, and PTH secretion. These findings suggest that the Ca 2+ receptor is the only molecular entity involved in sensing changes in the concentration of extracellular Ca 2+, and that the Ca 2+ receptor can couple to two distinct transmembrane signaling systems in parathyroid cells. Curiously, C a 2+ receptor-mediated increases i n [Ca2+]i are associated with an inhibition of PTH secretion (see Figs. 4 and 8). This is in contrast to most other secretory cells in which increases in [Ca2+]i stimulate secretion (see Chapter 42) and it suggests that cytoplasmic Ca 2+ may have an inhibitory effect on exocytotic secretion in parathyroid cells. At present, however, it is uncertain whether cytoplasmic Ca 2+ acts as the principal intracellular signal regulating PTH secretion. Although increases in [Ca2+]i in parathyroid cells almost invariably inhibit secretion, changes in PTH secretion can be dissociated from changes i n [Ca2+]i, and this suggests that additional or alternative intracellular signals play an important role in regulating PTH secretion. There is much evidence showing that increases in [Ca2+]i arising from influx of extracellular Ca 2+ are not importantly involved in the regulation of PTH secretion (Nemeth and Scarpa, 1987). Some of this evidence derives from the use of Type I calcimimetics that, as mentioned earlier, mobilize intracellular Ca 2+ but do not allow influx of extracellular Ca 2+. Moreover, blocking the influx of extracellular Ca 2+ (with channel blockers) does not affect the ability of extracellular Ca 2+ to inhibit secretion of PTH. Nonetheless, most of these polycations inhibit secretion of PTH just as well as extracellular Ca 2+, which mobilizes intracellular Ca 2+ and allows the influx of extracellular Ca 2+ (see Fig. 8). These finding show that transient increases in [Ca2+]i arising from mobilization of intracellular C a 2+ a r e associated with the inhibition of PTH secretion. Yet there is still no consensus that cytoplasmic Ca 2+ plays the dominant role in regulating secretion of PTH. Various other intracellular signaling mechanisms have been postulated to play a role in controlling PTH secretion, but no compelling molecular model has yet emerged. One signaling molecule that has been studied extensively is protein kinase C (PKC). When PKC is inhibited in parathyroid cells, or when cellular levels are depleted by downregulation of PKC activity, extracellular Ca 2+ is still capable of regulating PTH secretion in a normal manner. However, PKC does play a role in modulating the sensitivity of the Ca 2+ receptor to extracellular Ca 2+. PKC, probably by phosphorylation of the Ca 2+ receptor, depresses the sensitivity of parathyroid cells to regulation by extracellular Ca 2+. Thus, PKC seems to act in a negative-feedback manner to dampen
186
SECTION I Biophysical Chemistry, Metabolism, Second Messengers, and Ultrastructure
F I G U R E 7. Schematic representation of some of the mechanisms known to be coupled to the Ca2§ receptor in parathyroid cells. The G proteins linking the Ca2§ receptor to adenylate cyclase (AC) and phospholipase C (PLC) are still uncertain, as discussed in the text. Activation of the Ca2§ receptor by extracellular Ca2§ results in the rapid formation of inositol 1,4,5-trisphosphate [Ins(1,4,5,)P3] and mobilization of intracellular Ca2§ There is additionally an influx of extracellular Ca2§ through voltageinsensitive channels. Inhibitory pathways are shown by dotted lines. Pertussis toxin inhibits the G protein coupling the Ca2§ receptor to adenylate cyclase, but not that coupling the receptor to phospholipase C. Protein kinase C can phosphorylate the Ca2§ receptor at five potential sites (filled circles on cytoplasmic regions), and it also increases influx of extracellular Ca2§
transmembrane signaling mechanisms linked to the Ca 2§ receptor (see Fig. 7; Racke and Nemeth, 1993a, b and 1994). In addition to the rapid effects of extracellular Ca 2§ on PTH secretion, there are other cellular responses apparently regulated by extracellular Ca 2§ that become manifest over a day or two. Prolonged hypocalcemic conditions increase the synthesis of preproPTH, whereas hypercalcemic conditions decrease expression of the gene for PTH (Martin and Slatopolsky, 1994). Although it is reasonable to suppose that the Ca 2§ receptor also mediates these long-term effects of extracellular Ca 2§ on parathyroid cells, this additional function is not yet proven.
Neomycin
Calcium 100
0
0
t-
~ o
O (D
0
"1-
0
50
03 ~..
B. Parafollicular Cells Scattered throughout the thyroid gland are parafollicular or C cells that secrete the h o r m o n e caicitonin, a 32-aminoacid peptide (Azria, 1989). The C cell, like the parathyroid cell, has long been k n o w n to respond to changes in the ambient level of Ca 2§ but unlike P T H secretion, calcitonin secretion is stimulated by elevated levels of extracellular Ca 2+. Calcitonin acts mostly on bone to inhibit b o n e resorption and therefore diminishes the flux of Ca 2§ from bone into the general circulation. Calcitonin therefore lowers serum levels of Ca 2§ and thus acts in opposition to PTH. In some species, calcitonin plays an important role in maintaining the serum level of Ca 2§ The physiological sig-
0
I
0.5
.....
(mM)
I
I
2
1
I
(mM)
50
Concentration of type I calcimimetic
FIGURE 8. Different effects on Type I calcimimetics o n [Ca2+]i, but not PTH secretion, in parathyroid cells. The traces show increases in [Ca2+]i when the concentration of extracellular Ca2§ is increased from 0.5 to 2 mM (left panel) or when neomycin is added (50 mM; right panel). Extracellular Ca2§ causes a transient and sustained increase in [Ca2§ whereas neomycin causes only a transient increase. Nonetheless, both calcimimetics inhibit t~H secretion to the same degree.
11. Regulation of Cellular Functions by Extracellular Calcium
nificance of calcitonin in systemic Ca 2+ homeostasis in humans, however, is still a matter of debate. When administered in pharmacological doses, calcitonin can lower plasma Ca 2+ levels, but physiological levels of this hormone do not figure prominently in the moment-to-moment regulation of systemic Ca 2+ homeostasis in humans (McDermott and Kidd, 1987). Most of our knowledge regarding the cellular physiology of C cells is based on studies using cells derived from rat or human medullary thyroid carcinoma (MTC) tumors. Because MTC cells might not express the normal C-cell phenotype, the accumulated data might not always accurately reflect the cellular physiology or normal parafollicular cells. Increasing the concentration of extracellular Ca 2+ evokes rapid increases in [Ca2+]i in a number of different rat and human MTC cell lines (Raue and SheriJbl, 1995). In contrast to PTH secretion, increases in [Ca2+]i in C cells are associated with a stimulation of calcitonin secretion. The C cell, therefore, behaves like most other secretory cells, where cytoplasmic Ca 2+ acts to stimulate exocytotic secretion. The C cell also differs from the parathyroid cell in the mechanisms used by extracellular Ca 2+ to increase [Ca2+]i. C cells arise from the neuroectoderm and thus display a phenotype more characteristic of neurons. C cells express a number of voltage-sensitive ion channels, including L- and T-type Ca 2+ channels. Increases in [Ca2+]/and calcitonin secretion elicited by extracellular Ca 2+ in MTC cells are blocked by inhibitors of voltage-sensitive Ca 2+ channels such as dihydropyridines. Conversely, activators of voltage-sensitive Ca 2+ channels increase [Ca2+]i and stimulate calcitonin secretion. These compounds are without effect on [Ca2+]i or PTH secretion in parathyroid cells (Muff et al., 1988). It has been difficult to demonstrate mobilization of intracellular Ca 2§ in C cells. It seems that influx of extracellular Ca 2§ through voltage-sensitive, L-type Ca 2§ channels accounts for most of the increase in [Ca2+]i elicited by extracellular Ca 2§ Unknown at present is how the Ca 2§ receptor couples to these voltage-sensitive Ca 2§ channels to permit influx of extracellular Ca 2§ It seems that extracellular Ca 2§ acts initially on the Ca 2§ receptor to allow influx of extracellular Ca 2+, because the human MTC cell line TT, which fails to respond to extracellular Ca 2+, does not express the Ca 2+ receptor (Garrett et al., 1995b). However, these TT cells also do not express certain subtypes of voltage-sensitive Ca 2+ channels, and this confounds the interpretation. Perhaps the most compelling evidence so far derives from the use of selective Type II calcimimetics. Such compounds increase[Ca2+]i and stimulate calcitonin secretion in various MTC cell lines, but not in TT cells, which lack the Ca 2§ receptor. The comparative pharmacology of the parathyroid and the parafollicular Ca 2§ receptors is a textbook example showing the influence of cellular systems on receptor pharmacology. Despite identical receptors on both cell types there are a number of Type I calcimimetics (Mg 2§ for example) which fail to activate the Ca 2+ receptor on MTC cells. Although these differential effects of Mg 2§ might be peculiar to MTC cells, other studies performed in
1[17
vivo (therefore assessing Ca 2+ receptor activity in authentic parafollicular cells) clearly show that the pharmacology of the Ca 2+ receptor differs in these two cell types. Thus, Type II calcimimetic compounds are 10- to 40-fold more potent in suppressing plasma levels of PTH than they are in increasing those of calcitonin (Fox et al., 1999). One possible explanation mentioned above, for these differences involves postreceptor mechanisms that are peculiar to the cell. Another possibility is that the density of receptors on parathyroid cells is much greater than that on parafollicular cells, and this possibly makes the parathyroid cells more sensitive to these compounds. In either case, the functional properties of the Ca 2+ receptor are dependent on the cell in which it is expressed.
C. Renal Epithelial Cells After the parathyroid gland, the Ca 2+ receptor is most densely expressed in the kidney (Riccardi et al., 1995). mRNA encoding the Ca 2+ receptor can be detected along most of the nephron, but highest levels are expressed in the thick ascending limb (TAL) of the loop of Henle and in the inner medullary collecting duct (IMCD). In the TAL, where much of the Ca 2+ (and Mg 2+) reabsorption from the nephron takes place, the Ca 2+ receptor is located largely on the basolateral side of the tubule and is thus positioned to sense changes in the plasma concentration of Ca 2+. The Ca 2+ receptor in the TAL couples to mechanisms that control the lumen-positive transepithelial voltage, which, in turn, drives Ca 2+ (and Mg 2+) reabsorption by a paracellular pathway (Hebert, 1996). Thus, in hypercalcemic conditions, the TAL Ca 2+ receptor on the basolateral surface of the epithelial cell is activated and Ca 2+ reabsorption is diminished. Conversely, hypocalcemic conditions and consequent low Ca 2+ activity will increase Ca 2+ reabsorption in the TAL. In this segment of the nephron, the Ca 2+ receptor acts to control the amount of Ca 2+ that will be removed in the urine. One of the more convincing pieces of evidence supporting this role of the Ca 2+ receptor in the TAL derives from molecular genetic studies. Recall that patients suffering from FBHH are hypocalciuric despite their hypercalcemia. These patients continue to reabsorb an inappropriate amount of Ca 2+. Ca 2+ receptors in the TAL of patients with FBHH are less sensitive to extracellular Ca 2+. These mutant receptors do not perceive the hypercalcemia so Ca 2+ reabsorption is not diminished as it normally would be. In the IMCD, the Ca 2+ receptor is localized mostly in vesicles within the epithelial cell and also on the luminal membrane. When luminal Ca 2+ increases, there is a compensatory decrease in the amount of water reabsorption, and this effect is seemingly mediated by the Ca 2+ receptor. The model proposed holds that activation of Ca 2+ receptors on the luminal surface of collecting duct cells inhibits the exocytotic insertion of aquaporin (water) channels in the luminal membrane (Brown and Hebert, 1995). This action results in less reabsorption of water, thereby diluting the urine with respect to Ca 2+. This effect makes sense in the context of systemic Ca 2+ homeostasis because if the plasma levels rise, then the kidney is called on to remove excess Ca 2+ from the
188
SECTION i Biophysical Chemistry, Metabolism, Second Messengers, and Ultrastructure
body. The increased levels of Ca 2+ appearing in the urine must be diluted to avoid the formation of kidney stones.
D. Other ExtraceUular Ca2§
Cells
The number of different cell types expressing the Ca 2+ receptor continues to grow and one must consider the possibility that extracellular Ca 2+ regulates a variety of bodily functions besides those involved in systemic Ca 2+ homeostasis. Indeed, it now seems that many kinds of cells are sensitive to extracellular Ca 2+. Some caution in interpreting the physiological significance of these fmdings is warranted. Many of the studies claiming cellular sensitivity to extracellular Ca 2+ used flawed experimental designs. Typically, the studies involved exposing the cell to a buffer which lacked divalent cations, followed by a rapid and large increase in the concentration of extracellular Ca 2+. Bathing cells in buffers lacking divalent cations tends to increase membrane permeability so that the subsequent addition of extracellular Ca 2+ will increase [Ca2+]i and this, in turn, will affect cellular responses. Another feature of many studies which limits theft significance is the use of transformed cell lines, which are far easier to study than the authentic cell type from which they are derived. This convenience, however, is not without cost, and the phenotype of the transformed cell line often lacks one or more characteristics of the parental cell. Moreover, the cellular phenotype can change during culture. Yet even when more rigorous criteria are applied, a wide variety of cell types still emerges which express the Ca 2+ receptor and seem to respond to extracellular Ca 2+ under physiological and/or pathophysiological conditions (Table 1). It seems difficult to escape the conclusion that extracellular Ca 2+ plays a general role as an extracellular signal affecting the activity of numerous cells throughout the body.
V. S u m m a r y A biological phenomenon that is gaining recognition is the peculiar ability of extracellular Ca 2+ to regulate the activity of certain cells in the body. Many of these cells, such as parathyroid cells, parafollicular cells in the thy-
TABLE 1
Extracellular Ca2+-Sensing Cells Cell Type
Function
Parathyroid
PTH secretion and synthesis; cellular proliferation
Parafollicular
Calcitonin secretion
Kidney TAL IMCD
Ca 2+ and Mg2+ reabsorption Water transport
Pancreas Endocrine Exocrine Epithelial ducts G cells Osteoblasts
Unknown Gastrin secretion Proliferation
roid (C cells), and certain renal tubule cells, are involved in maintaining systemic Ca 2+ homeostasis. These cell types, unlike most body cells, alter their activity in response to small, physiological changes in the concentration of Ca 2+ in the plasma or the extracellular fluids. These "extracellular CaZ+-sensing cells" express a Ca 2+ receptor in the plasma membrane that enables them to detect and respond to small changes in the level of extracellular Ca 2+. The Ca 2+ receptor is structurally and functionally similar to other plasma membrane receptors that sense changes in the concentration of extracellular ligands, and translate these ligand-receptor interactions into intracellular signals that alter cellular activity. The major difference is that the Ca 2+ receptor binds to an inorganic ion rather than an organic molecule; the Ca 2+ receptor is the first example of an "inorganic ion receptor." Extracellular Ca 2+ can thus act in a messenger capacity much like any other extracellular signal, such as a neurotransmitter or hormone. This newly discovered function of extracellular Ca 2+ complements the well-known ability of cytoplasmic Ca 2+ to act as an intracellular signal. Thus both extracellular Ca 2+ and intracellular Ca 2+ convey information and do so by interacting with CaZ+-binding proteins. Intracellular Ca 2+ receptors, such as calmodulin, bind Ca 2+ with high affinity (nanomolar to low micromolar levels), whereas the extracellular Ca 2+ receptor binds Ca 2+ with much lower affinity (millimolar levels), consistent with the very different concentrations of Ca 2+ in the cytoplasm and extracellular space. The Ca 2+ receptor is a G-protein-coupled receptor and is structurally homologous to metabotropic glutamate receptors. Both these receptor types possess an unusually large extracellular domain that binds the respective physiological ligand, extracellular Ca 2+ or glutamate. Although there is no conclusive evidence at present for subtypes of the Ca 2+ receptor, it seems likely that homologous yet distinct receptors will be discovered. The Ca 2+ receptor couples to a number of different transmembrane signaling mechanisms, depending on the particular cell type where it is expressed. In parathyroid cells, the Ca 2+ receptor couples to phospholipase C and adenylate cyclase to i n c r e a s e [Ca2+]i and decrease cAMP levels and secretion of PTH. The coupling to adenylate cyclase is mediated by a Gi-like protein, whereas that to phospholipase C is mediated either by G~l or G_. q In C cells, activation of the Ca 2+ receptor allows the influx of extracellular Ca 2+ through voltage-sensitive Ca 2+ channels, leading to an increase in [Ca2+]i and the secretion of calcitonin. In renal epithelial cells, entirely different intracellular signaling mechanisms are coupled to the Ca 2+ receptor, and these intracellular signals regulate Ca 2+ and Mg 2+ reabsorption and water transport. Both genetic and pharmacological studies show that the Ca 2+ receptor is the essential molecular entity maintaining systemic Ca 2+ homeostasis. Mutations in the Ca 2+ receptor that decrease its sensitivity to extracellular Ca 2+ result in hypercalcemic disorders, such as FBHH and NSHPT, and the latter is often fatal if not treated. Mutations in the Ca 2+ receptor that increase its sensitivity to extracellular Ca 2+ result in hypocalcemic disorders such as ADH. Moreover,
11. Regulation of Cellular Functions by Extracellular Calcium
calcimimetic c o m p o u n d s that activate the Ca 2+ receptor cause decreases in circulating levels of P T H and Ca 2+. The last studies (Nemeth and Fox, 1999) also suggest that the Ca 2+ receptor is a novel molecular target for drugs useful in treating a variety of bone and mineral disorders such as hyperparathyroidism and osteoporosis. Most of what is known about how the Ca 2+ receptor regulates cellular activity derives from the study of cells involved in systemic Ca 2+ homeostasis. Yet it is clear that many other types of cells, which have little if any role in systemic Ca 2+ homeostasis, also express the Ca 2+ receptor. The function of the Ca 2+ receptor in most of these cells is far from understood. The widespread distribution of the Ca 2+ receptor, however, suggests that extracellular Ca 2+ has an important messenger function throughout the body. These recent discoveries add a new dimension to physiology and show that extracellular Ca 2+ (and perhaps other inorganic ions) must now be considered a variation on the theme of chemical communication in the body.
Bibliography Azria, M. (1989). "The Calcitonins. Physiology and Pharmacology." Karger, New York. Bai, M., Quinn, S., Trivedi, S., Kifor, O., Pearce, S. H. S., Pollak, M. R., Krapcho, K., Hebert, S. C., and Brown, E. M. (1996). Expression and characterization of inactivating and activating mutations in the human CaZ+-sensing receptor. J. Biol. Chem. 271, 19537-19545. Bai, M. (1999). Structure and function of the extracellular calciumsensing receptor (Review). Int. J. Molec. Med. 4, 115-125. Beck-Sickinger, A. G. (1996). Structural characterization and binding sites of G-protein-coupled receptors. Drug Disc. Today 1, 502-513. B ikle, D. D., Ratnam, A., Mauro, T., Harris, J., and Pillai, S. (1996). Changes in calcium responsiveness and handling during keratinocyte differentiation. Potential role of the calcium receptor. J. Clin. Invest. 97, 1085-1093. Brown, E. M. (1983). Four parameter model of the sigmoidal relationship between parathyroid hormone release and extracellular calcium concentration in normal and abnormal parathyroid tissue. J. Clin. Endocrinol. Metab. 56, 572-581. Brown, E. M. (1991). Extracellular Ca2+-sensing, regulation of parathyroid cell function, and role of Ca2+ and other ions as extracellular (first) messengers. Physiol. Rev. 11, 371-411. Brown, E. M. (1994). Homeostatic mechanisms regulating extracellular and intracellular calcium metabolism. In "The Parathyroids" (J.E Bilezikian, R. Marcus, and M.A. Levine, Eds.), Raven Press, New York, pp. 15-54. Brown, E. M., and Hebert, S. C. (1995). A cloned Ca2+-sensing receptor: A mediator of direct effects of extracellular Ca2+ on renal function? J. Am. Soc. Nephrol. 6, 1530-1540. Brown, E. M., Gamba, G., Riccardi, D., Lombardi, M., Butters, R., Kifor, O., Sun, A., Hediger, M. A., Lytton, J., and Hebert, S. C. (1993). Cloning and characterization of an extracellular Ca 2+sensing receptor from bovine parathyroid. Nature 366, 575-580. Clark, E. A., and Brugge, J. S. (1995). Integrins and signal transduction pathways: The road taken. Science 268, 233-239. Drtieke, T. B. (1995). The pathogenesis of parathyroid gland hyperplasia in chronic renal failure. Kidney Intl. 48, 259-272. Fox, J., Lowe, S. H., Conklin, R. L., Petty, B. A., and Nemeth, E. E (1999). Calcimimetic compound NPS R-568 stimulates calcitonin secretion but selectively targets parathyroid gland Ca2+ receptor in rats. J. Pharm. and Exper. Therapeutics 290(2), 480-486.
189
Garrett, J. E., Capuano, I. V., Hammerland, L. G., Hung, B. C. R, Brown, E. M., Hebert, S. C., Nemeth, E. F., and Fuller, F. (1995a). Molecular cloning and functional expression of human parathyroid calcium receptor cDNAs. J. Biol. Chem. 270, 12,919-12,925. Garrett, J. E., Tamir, H., Kifor, O., Simin, R. T., Rogers, K. V., Mithal, A., Gage, R. E, and Brown, E. M. (1995b). Calcitonin-secreting cells of the thyroid express an extracellular calcium receptor gene. Endocrinology 136, 5202-5211. Hammerland, L. G., Krapcho, K. J., Garrett, J. E., Alasti, N., Hung, B. C. E, Simin, R. T., Levinthal, C., Nemeth, E. E, and Fuller, E H. (1999). Domains determining ligand specificity for Ca 2+ receptors. Molec. Pharm. 77, 642-648. Hebert, S. C. (1996). Extracellular calcium-sensing receptor: Implications for calcium and magnesium handling in the kidney. Kidney Intl. 50, 2129-2139. H6bert, T. E., and Bouvier, M. (1998). Structural and functional aspects of G protein-coupled receptors oligomerization. Biochem. Cell Biol. 76, 1-11. King, E D., and Wilson, S. (1999). Recent advances in 7-transmembrane receptor research. Curr. Opin. Drug Disc. Devel. 2(2), 83-95. Kolakowski, L. E (1994). GCRDb: A G-protein-coupled receptor database. Recep. Channels 2, 1-7. Lundgren, S., Hj~ilm, G., Hellman, E, Ek, B., Juhlin, C., Rastad, J., Klareskog, L., Akerstr6m, G., and Rask, L. (1994). A protein involved in calcium sensing of the human parathyroid and placental cytotrophoblast cells belongs to the LDL-receptor protein superfamily. Exp. Cell Res. 212, 344-350. Martin, K. J., and Slatopolsky, E. (1994). The parathyroids in renal disease. In "The Parathyroids" (J.E Bilezikian, R. Marcus, and M.A. Levine, Eds.), Raven Press, New York, pp. 711-719. Mayer, G. E, and Hurst, J. G. (1978). Sigmoidal relationship between parathyroid hormone secretion rate and plasma calcium concentration in calves. Endocrinology 102, 1036-1042. McDermott, M. T., and Kidd, G. S. (1987). The role of calcitonin in the development and treatment of osteoporosis. Endocr. Rev. 8, 377-390. Muff, R., Nemeth, E. E, Haller-brem, S., and Fischer, J. A. (1988). Regulation of hormone secretion and cytosolic Ca2+by extracellular Ca2+ in parathyroid cells and C-cells: Role of voltage-sensitive Ca2+ channels. Arch. Biochem. Biophys. 265, 128-135. Mundy, G. R., (1989). "Calcium Homeostasis: Hypercalcemia and Hypocalcemia." Martin Dunitz, London. Nemeth, E. E, and Scarpa, A. (1986). Cytosolic Ca2+ and the regulation of secretion in parathyroid cells. FEBS Lett. 203 (1), 15-19. Nemeth, E. E, and Scarpa. A. (1987). Are changes in intracellular free calcium necessary for regulating secretion in parathyroid cells? Ann. NY Acad. Sci. 493, 542-551. Nemeth, E. E (1990). Regulation of cytosolic calcium by extracellular divalent cations in C-cells and parathyroid cells. Cell Calcium 11, 323-327. Nemeth, E. E, and Heath, H., III. (1995). The calcium receptor and familial benign hypocalciuric hypercalcemia. Curr. Opin. Endo. Diabetes 2, 556-561. Nemeth, E. E (1996). Calcium receptors as novel drug targets. In "Principles of Bone Biology" (J. E Bilezikian, L. G. Raisz, and G. A. Rodan, Eds.), Academic Press, New York, pp. 1019-1035. Nemeth, E. E, Steffey, M. E., Hammerland, L. G., Hung, B. C. E, Van Wagenen, B. C., DelMar, R. G., and Balandrin, M. E (1998). Calcimimetic with potent and selective activity on the parathyroid calcium receptor. Proc. Natl. Acad. Sci. USA 95, 4040--4045. Nemeth, E. E, and Fox, J. (1999). Calcimimetic compounds: a direct approach to controlling plasma levels of parathyroid hormone in hyperparathyroidism. TEM 10(2), 66-71. Pearce, S. H., S., and Brown E. M., (1996). The genetic basis of endocrine disease. Disorders of calcium ion sensing. J. Clin. Endo. Metab. 81, 2030-2035.
190
SECTION ! Biophysical Chemistry, Metabolism, Second Messengers, and Ultrastructure
Petersen, C. C. H. (1996). Store operated calcium entry. Semin. Neurosci. 8, 293-300. Pollak, M. R., Brown, E. M., Chou. Y. H. W., Hebert, S. C., Marx, S. J., Steinmann, B., Levi, T., Seidman., C. E., and Seidman, J. G. (1993). Mutations in the human Ca2+-sensing receptor gene cause familial hypocalciuric hypercalcemia and neonatal severe hyperparathyroidism. Cell 75, 1297-1303. Pollak, M. R., Brown, E. M., Estep, H. L., McLaine P. N., Kifor, O., Park J., Hebert, S. C., Seidman, C. E. and Seidman, J. G. (1994). Autosomal dominant hypercalcaemia caused by a Ca2+-sensing receptor gene mutation. Nat. Genet. 8, 303-307. Racke, E K., and Nemeth E. E (1993a). Cytosolic calcium homeostasis in bovine parathyroid cells and its modulation by protein kinase. C.J. Physiol. 468. 163-176. Racke, E K., and Nemeth, E. E (1993b). Protein kinase C modulates hormone secretion regulated by extracellular polycations in bovine parathyroid cells. J. Physiol. 468, 163-176. Racke, E K., and Nemeth E. E (1994). Stimulus-secretion coupling in parathyroid cells deficient in protein kinase C activity. Am. J. Physiol. 267, E429-E438.
Raue, E, and Scherubl, H. (1995). Extracellular calcium sensitivity and voltage-dependent calcium channels in C cells. Endocr. Rev. 16, 752-764. Ray, K., Clapp, P., Goldsmith, P. K., and Spiegel, A. M. (1998). Identification of the sites of N-linked glycosylation on the human calcium receptor and assessment of their role in cell surface expression and signal transduction. J. Biol. Chem. 273(51), 34558-34567. Riccardi, D., Park, J., Lee, W. -S., Gamba, G., Brown, E. M., and Hebert, S. C. (1995). Cloning and functional expression of a rat kidney extracellular calcium-sensing receptor. Proc. Natl. Acad. Sci. USA 92, 131-135. Takahashi, K., Tsuchida, K., Tanabe, Y., Masu, M., and Nakanishi, S. (1993). Role of the large extracellular domain of metabotropic glutamate receptors in agonist selectivity determination. J. Biol. Chem. 268, 19341-19345. Zaidi, M., Alam, A. S. M. T., Huang, C. L. -H., Pazianas. M., Bax, C. M. R., Bax. B. E., Moonga, B. S., Bevis, P. J. R., and Shankar, V. S. (1993). Extracellular Ca 2§ sensing by the osteoclast. Cell Calcium 14, 271-277.
Nelson D. Horseman and J. Wesley Pike
Cellular Responses to Hormones
i. Introduction Hormones are secreted biochemical substances that affect the function of cells within an individual by binding to receptors and eliciting specific reactions from their target cells. Hormones affect cellular functions by altering rates of gene expression and metabolism, but are not themselves substrates for the processes that they regulate. Therefore, hormones are primarily information carriers. The first hormones were identified based on their secretion from specialized endocrine (internal secretion) glands. In recent years a large number of non-endocrine hormones have been discovered using advanced methods of cell culture, biochemical purification, and recombinant DNA technology. These paracrine (locally secreted) hormones have many novel features. Homeostasis and development are controlled through the actions of endocrine and paracrine hormones on target cells, which each express receptors for a limited array of hormones. Because hormones do not directly participate in the processes they control, we conceive of them as initiating signal transduction protocols, that lead to the appropriate cellular outcomes. In this chapter we will cover examples of signal transduction mechanisms for two broad classes of hormones: the lipophilic hormones such as steroids, thyroxine, retinoids and vitamin D, which diffuse into cells and bind to intracellular receptors; and the hydrophilic peptide and amine hormones, which interact with receptors that are exposed on the surface of cells.
il. Actions of Lipophilic Hormones via IntraceUular Receptors A. Signal Transduction Through lntracellular Receptors: Basic Principles The sex and adrenal steroids and the thyroid, retinoic acid, and vitamin D hormones are members of a chemically diverse set of endocrine signaling molecules that are produced and regulated in response to both internal and enCell PhysiologySourcebook:A MolecularApproach,Third Edition
| 9 1
vironmental cues. Acting on both distant and local tissue targets, they exert regulatory control over a myriad of specific cellular functions associated with virtually every vertebrate organ system. These actions impact, for example, cellular metabolism, reproductive function, nutrient homeostasis, and behavior. A common biological feature of each of these hormones is their capacity to regulate cellular differentiation. The mechanism through which lipophilic hormones dictate specific biological responses involves the ability of these hormones to enter the nucleus and modulate the expression of single genes and gene networks. This action is selective by virtue of the presence in cells of individual receptors for each of these hormones, and the presence of these receptors represents the primary determinant of tissue response to the cognate hormone. Nuclear receptors belong to a large gene family of transcription factors that recognize their respective endocrine signals and respond accordingly. In this section, we describe the progress over the past decade that has enhanced our understanding of how these signaling pathways modify gene expression. In contrast to peptide hormones, growth factors, and cytokines, the lipophilic hormones are not limited in their ability to gain entry into the cell (Beato, 1989; Katzenellenbogen et al., 1996). Due to their solubility in the membrane and their small size, they are believed to enter the cell by diffusion through the cell membrane; this energy-independent process apparently permits them to ultimately reach the cell nucleus where they associate though a high-affinity interaction with a specific nuclear receptor (Fig. 1). While the bulk of the members of the nuclear receptor family are believed to reside in the nucleus prior to ligand activation, the receptors for the glucocorticoids (GR) and the mineralocorticoids (MR) appear to be the single exceptions in that they are found in the cytoplasm (Evans, 1988; Beato et al., 1995; O'Malley, 1990). In this specific case, interaction with glucocorticoid or mineralocorticoid ligand induces translocation of the receptors to the nucleus. The interaction of hormonal ligand with its receptor triggers a series of events which culminate in alterations in gene expression. Allosteric changes in the receptor induced by ligand lean to Copyright 9 2001 by Academic Press. All rights of reproduction in any form reserved.
192
SECTION ! Biophysical Chemistry, Metabolism, Second Messengers, Ultrastructure
gen receptor (AR), thyroid hormone receptors (TRee, /3), retinoic acid receptors (RARce,/3, 30, 9-cis-retinoic acid receptors (RXRoz,/3, 3i) and the vitamin D receptor (VDR) (Gronemeyer, 1993; Truss and Beato, 1993; Mangelsdorf and Evans, 1995; Mangelsdorf et al., 1995; Beato et al., 1995). Single genes encode certain of the receptors whereas other receptors are derived from multiple genes. In the latter case, these receptors may exhibit differences in tissue distribution and in gene regulatory activity as a result of slight but significant differences in their primary structure. During transcription, receptor genes can also undergo alternative splicing events that often lead to multiple mRNA transcripts that encode receptor gene products that exhibit different functional activities. The existence of multiple genes for receptors, alternative splicing events within single genes, and differential use of translation start sites within the mRNA transcripts result in a heterogeneous mix of receptor gene products, as well as the possibility for highly diverse responses to ligand within cells, depending upon the particular receptor gene product being expressed. FIGURE 1. Modelfor the molecular mechanism of action of steroid, thyroid, retinoid, and vitamin D hormones. Hormonalligand enters the cell by diffusion and interacts with its cognate receptor. Activation by the ligand leads to the interaction of receptor with responsive genes and the modulation of gene expression. Active receptors are comprised of monomers, homodimers, or heterodimers (see text). (a) Model for glucocorticoid and mineralocorticoid receptors, wherein the receptor is cytoplasmic in the absence of ligand. Upon ligand activation the receptor undergoes nuclear translocation and eventuallybinds to the regulatoryregion of a modulated gene. (b) Model for most of the nuclear receptors wherein the receptor is located in the nucleus and following ligand activation becomes bound to the regulatoryregion of hormone-responsivegenes.
dissociation or associated proteins which function as inhibitors of receptor DNA-binding ability and transcriptional capacity, formation of functional protein units, binding of the units to specific DNA sequence elements located adjacent to hormone sensitive gene promoters, and activation or repression of transcription (Fig. 1). The capacity of a hormone to regulate transcription is dependent upon the presence of the receptor for that hormone in a cell, but the quantitative and qualitative nature of the response can be both gene-specific and cell-specific depending upon the presence and activity of additional cellular protein factors. Cellular mechanisms which serve to inactivate receptor signals within the nucleus are largely unknown, although the induction of enzymes by the hormone-activated receptor, that degrade the corresponding hormonal signal, is not uncommon.
B. The Nuclear Receptor Gene Family 1. Ligand-Activated Nuclear Receptors
The genes for all of the receptors which mediate the actions of traditional nonpeptide endocrine hormones have been cloned (Fig. 2). These include the receptor for the glucocorticoids (GR), mineralocorticoid receptor (MR), estrogen receptors (ERa,/3), progesterone receptor (PR), andro-
2. Orphan Receptors
The cloning of the receptors for known endocrine hormones and the observation that all belonged to a structurally related family of genes precipitated a search for additional gene members in vertebrate tissues and in nonvertebrate species. This search has resulted in the identification of numerous additional genes which are clearly members of the nuclear receptor gene family (Fig. 2) (O'Malley, 1990; Mangelsdorf and Evans, 1995). While none of these orphan receptors appear currently to mediate the actions of known endocrine hormones, several appear to facilitate the actions of local cellular factors such as 9-cis-retinoic acid and prostaglandin J2 (Forman et al., 1995) as well as cellular metabolic intermediates such as farnesol, arachidonic acid, and perhaps certain cholesterol derivatives. Current efforts are focused on determining whether novel ligands indeed exist for these receptors, although activation pathways independent of ligand are clearly possible. Moreover, in addition to functioning as gene activators, there is evidence that certain of the orphans may play largely negative roles in transcription, acting not only to repress the transcriptional activity of specific genes, but also to function on other genes as repressors of the inducing actions of ligand-activated nuclear receptors. Novel biological roles for several of these orphan receptors have been defined. 3. Invertebrate Receptor G e n e s A large number of genes from invertebrates have been identified as belonging to the steroid receptor gene family (Fig. 2) (Mangelsdorf et al., 1995; Thummel, 1995). The genes include those from Drosophila, Caenorhabditis elegans, and other metazoan species. Perhaps the most interesting member of the nuclear receptor family found in Drosophila melanogaster is the receptor for ecdysone, the metamorphic or molting hormone of insects. This hormone has long been associated with its capacity to induce chromosomal puffing at specific sites on chromatin and to induce complex networks of genes. Interestingly, as will be described below, the
12. Cellular Responses to Hormones
193
Genes
Species*
Ligand
GR MR PR AR EcR FXR LXR ~,13 VDR xONR M B 6 7 ~,13 CeF11A1.3 TR ~,13 CEB0280.8 CeC2B4.2 C O U P ~,13,~( RXR c~,13,y,(8,~) TR2-11 ~,13 HNF-4 TLL CeF21D12.4 GCNF P P A R ot,13,),,8? RA R c~,13,y CNR14 E78 R e v - E r b ~,13 E75A R O R m,13,~/ DHR3 N G F I - B ~,13,7 FTZ-F 1 ~,13 CeFllC1.6 ERR tt,13 ER et,13 K N I R P S ~,13,y CeE2H1.6 CEF43C1.4 CeKO6A1.4 CeODR7 CeZK418.1 CeF16H9.1
1 1 1 1 2 1 1 1 1 1 2 1 2 2 1,2 1,2 1,2 1,2 1,2 2 1 1 1 2 2 1 2 1 2 1,2 1,2 2 1 1 2 2 2 2 2 2 2
GLUCOCORTICOIDS MINERALOCORTICOIDS PROGESTERONE TESTOSTERONE ECDYSONE FARNESOIDS 1,25(OH)2Ds
THYROID HORMONE
9-cis-RA
EICOSANOIDS, PROSTAGLANDINS RETINOIC ACID
ESTROGEN
FIGURE 2. Membersof the nuclear receptor superfamily of genes. The figure documents cloned nuclear receptor family members. Known ligands are indicated on the right.
ecdysone receptor functions as a heterodimer with a nuclear protein partner called ultraspiracle. Ultraspiracle represents the insect homolog of vertebrate retinoid X receptor (RXR) genes, which in turn function as permissive heterodimer partners with the thyroid receptor (TR), the retinoic acid receptor (RAR), the vitamin D receptor (VDR), and several others. The presence of nuclear receptors in insects suggests that this highly successful receptor family evolved prior to the divergence of invertebrates and vertebrates.
C. The Highly Conserved Structural Domains of the Nuclear Receptors The nuclear receptors display a highly modular domain structure composed of a DNA-binding domain of 60 to 70 amino acids and an extended carboxy-terminal ligand-binding
domain of 300 amino acids separated by a highly flexible hinge of 100 to 150 amino acids, as seen in Fig. 3 (Evans, 1988). Many of the individual functions of each domain are retained following their separation through enzymatic cleavage. The DNA-binding domain is highly conserved among members of the nuclear receptor superfamily and represents the hallmark of this transcription factor family. In addition to containing determinants of specific DNA binding, this domain also contains subregions which are involved in directing the protein to the nucleus following synthesis and stabilizing the protein's interaction with DNA through dimerization with a protein partner. The carboxy-terminal domain of the nuclear receptors, which is over 300 amino acids in length, exhibits subregions of homology across members of the family and contains hydrophobic residues that create a high affinity lipophilic pocket exquisitely selective for the receptor's cognate
194
SECTION ! Biophysical Chemistry, Metabolism, Second Messengers, Ultrastructure AIB
N-[ ' ' ,
,
c
D
m,,
I m,,, ~
[]
I-c e
~r
E Hormone binding
I
Heat shock association
9
Dimerization Nuclear localization DNA binding
9
9
Ill
I I
m
Transactivation
FIGURE 3. Functionaldomain structure of the nuclear receptor superfamily. The nuclear receptors (NR) are separated into four regions designated A/B, C, D, and E. Regions of homology within the nuclear receptor family are shaded. Regions involved in known specific functions of the receptor are indicated below the diagram.
ligand. Like that of the DNA-binding domain, this region also exhibits additional functional activities, including a dimer interface for interaction with the receptor's protein parmer, and additional protein/protein interfaces that are integrally involved in the receptor's ability to regulate transcription. Finally, receptors for estrogen, progesterone, androgens, and the glucocorticoids contain an extensive amino-terminal domain that exhibits additional functions associated with transcriptional activation. This complexity in structural organization of the nuclear receptor family implies diverse, highly flexible, and specific activities during the regulation of gene expression.
1. The DNA-Binding Domain of the Nuclear Receptors The DNA-binding domain of the nuclear receptor family is comprised of two highly conserved zinc finger structures. These structures contain several invariant amino acids that include two sets of four cysteine residues that have been shown to coordinate two zinc ions (Berg, 1989). The two zinc ions hold the DNA-binding domain in an active conformation structurally; in their absence the finger motifs are nonfunctional and sensitive to proteolytic degradation. The three-dimensional structures of the DNA-binding domains of the receptors for the glucocorticoids, estrogen and 9-cisretinoic acid have been established (Hard et al., 1990; Schwabe et al., 1990; Luisi et al., 1991; Lee et al., 1993; Schwabe et al., 1993) In addition, the three-dimensional structure of the TR and RXR DNA-binding domains bound to DNA as heterodimers has also been determined. The observed structures of these regions of the nuclear receptors confirm and extend earlier predictions which were made based upon biochemical evaluations. The two zinc finger modules are known to serve very different functional roles during association with DNA (Mader et al., 1989). The first functions to dictate the specificity of DNA interaction in the major groove on a core DNA sequence element, whereas the second functions to establish contacts with the receptor's protein partner. These contacts stabilize and strengthen the binding of the two proteins on the DNA sequence. Different regions located within the second zinc module, however, participate in this dimerization function depending upon the organization of the DNA-binding site and the orientation of the two receptor molecules which become bound to that site.
2. The Ligand-Binding Domain of the Nuclear Receptors In contrast to the DNA-binding domain, the region of the receptor that is required for ligand binding extends to over several hundred amino acids. Since the ligand itself is rather small relative to this region, it is clear that the protein pocket which makes direct contact with the ligand is determined through proper folding of different components of the protein domain itself. Indeed, several regions of homology have been defined within this domain that exist among the nuclear receptor family. These regions of homology, however, participate not only in creating a ligand-binding pocket, but also serve several additional functional roles that involve protein/protein interactions; two such interactions are those that lead to homodimer or heterodimer formation with the receptor's protein partner, and those which mediate contact with the general transcriptional apparatus. Additional interactions include receptor associations with either inhibitor or repressor proteins that function to prevent receptor DNAbinding and transcriptional activity in the absence of ligand, or that facilitate transcriptional repression while bound to DNA in the presence of the activating ligand (Horlein et al., 1995). Recent studies have led to the elucidation of the crystal structure of ligand binding domains of TR, RAR, and RXR (Bourguet et al., 1995; Renaud et al., 1995; Wagner et al., 1995; Rastinejad et al., 1995). Crystallization of the first two was achieved in the presence of the corresponding ligands whereas the structure of the RXR ligand-binding domain was determined in the absence of its cognate ligand. These studies have revealed the three-dimensional structure of the nuclear receptors to be that of an antiparallel a-helical sandwich comprised of as many as twelve helices. Importantly, the binding of hormone leads to a significant repositioning of several of these a-helices, including the most carboxy-terminal one that is known to be involved in mediating contacts with the general transcription apparatus. Helices 9 and 10 participate in the formation of receptor homo- or heterodimers. The ligand-binding pocket is comprised of extended clusters of residues that include helices 1, 3, 5, a/3 turn between 5 and 6, a loop between 6 and 7, 11 and 12, and the loop between 11 and 12. While the crystal structure of a complete receptor has yet to be determined, these studies with both the DNA-binding domain and the
12. Cellular Responses to Hormones
195
ligand-binding domain have provided significant insight into structure-function relationships which exist for this class of transcription factors.
3. Additional Nuclear Receptor Domains Two additional domains are found within the nuclear receptor gene family. The first is a region common to all the receptors that serves to link the DNA-binding domain with the ligand-binding domain. It is not conserved within the receptor gene family. One characteristic of this region is flexibility; depending upon the orientation of the DNA sequence elements to which the individual receptors bind, the hinge region allows 180 ~ swiveling of the two domains which surround it. It is likely that this region also subserves other as yet unknown functions. The second domain is a large amino-terminal region found predominantly in receptors for estrogen, progesterone, androgens, and the glucocorticoids. This domain contains additional regions that function to make direct or indirect protein-protein contacts with the general transcriptional apparatus. The three-dimensional structure of this region has not been determined, either independently or in association with the DNA-binding domain or other portions of the receptor molecule.
D. DNA-Binding Motifs within Nuclear Receptors Nuclear receptors interact with unique sequences of DNA located within the promoters for hormone-regulated genes. These associations represent the first in a series of steps which result in transcriptional modulation. Research over the past decade on both naturally regulated genes as well as synthetically derived DNA sequences has revealed the nature of these elements which serve as selective binding sites for members of the nuclear receptor family. As seen in Fig. 4, three classes of DNA-binding sites have emerged which reflect the three modes of DNA-binding characteristic of members of this gene family--homodimeric DNA binding, heterodimeric binding, and monomeric binding (Umesono and Evans, 1989; Umesono et al., 1991; Perlmann et al., 1993). The core DNA-binding site within promoters is comprised of a hexanucleotide sequence. In the case of nuclear receptors which bind to DNA as monomers, this core sequence together with a relatively well-defined short stretch
Orphan receptors
BIB
Steroid receptors
I
Thyroid, retinoid, vitamin D receptors
I
of nucleotides located immediately 5', comprise the DNAbinding site. The sequence of the core as well as the adjacent nucleotides represent determinants of specificity. Numerous orphan receptors interact with DNA as monomers. The nuclear receptors which bind to DNA as dimers interact with core elements similar to those just described. In contrast, however, these core elements are repeated to provide the two similar and adjacent binding sites necessary to accommodate the association of a dimeric nuclear receptor. Two dissimilar motifs arise as a result of this repeating structure: one wherein the core sequence is repeated in a direct fashion, and the second in which the core sequence is inversely repeated in palindromic fashion. Those receptors which function as homodimers such as the GR, ER, PR, and AR, bind to palindromically repeated core sequences. In contrast, receptors which function as heterodimers in combination with a dissimilar nuclear receptor protein partner, such as TR, RAR, and VDR, bind to directly repeated motifs. The importance of the highly flexible hinge region is highlighted here. Clearly, dimeric protein binding to palindromic DNA sequences requires a symmetric interface between the two protein partners in a head-to-head configuration, whereas dimeric binding to directly repeated DNA sequences requires an asymmetric interface wherein the proteins are arranged head to tail. Research has revealed that the DNAbinding domain of the nuclear receptors configure in a head-to-head or head-to-tail arrangement depending upon the structure of the DNA binding site. The carboxy-terminal ligand-binding domains, in contrast, retain their symmetric head-to-head configurations irrespective of whether the DNA-binding site is palindromic on direct in orientation. These two possibilities are accomplished through the hinge region of the nuclear receptors, which allows for an independent swiveling of the DNA-binding domain relative to the ligand-binding domain. Finally, the individual sequences of the core elements determine the ability and therefore specificity of the interaction by homodimeric receptors such as ER, PR, and GR. In contrast, directly repeated core sequences are generally similar; specificity is determined not through a difference in the sequence, but rather through the number of nucleotides located between the two core halfsites. Thus, binding sites for the VDR, TR, and RAR are determined largely by nucleotide spacings of three, four, and five, respectively; additional arrangements have also been determined.
- >(~)
~>(~)others
Gi2
Ubiquitous
Gi3
Others>neural
GO Gz
Neural, endocrine Neural, platelets
Adenylyl cyclase K§ channel Adenylyl cyclase K§ channel Adenylyl cyclase K§ channel Ca2+channel 9
Gq Gll
Gin
Ubiquitous Ubiquitous Liver, lung, kidney
PLC/3 PLC/3 PLC/3
G15/16 G12
Blood cells Ubiquitous
PLCfl ?
G13
Ubiquitous
?
cent years (Table 1) and the details of effector modulation by these isoforms is only partly understood. Signaling by G proteins to the nucleus converges on the transcription factor CREB (cAMP response element binding protein), which is a 341 amino acid protein that is activated by phosphorylation of serine 133 (Fig. 9). cAMP-dependent protein kinase (PKA) and calcium-calmodulin-dependent protein kinase II (CAM kinase II) each phosphorylate this site. CREB is able to integrate numerous G protein signals because it responds to changes in both cAMP and calcium, each of which is controlled by multiple G protein isoforms. Activation of G ~q leads to calcium release from intracellular stores in response to the second messenger IP 3. IP 3 is a product of the breakdown of the membrane phospholipid phosphatidylinositol 4,5-bisphosphate (PIP2) by phospholipase C, and diacylglycerol is generated simultaneously. IP 3 provokes calcium release from the endoplasmic reticulum and calcium activates numerous targets, including calmodulin and calcium-calmodulin-dependent proteins, and protein kinase C (PKC). PKC is localized to membrane compartments because it binds to diacylglycerol, as well as to calcium. Recent identification of multiple PKC isoforms has allowed a renewed focus on the downstream effectors that are regulated by PKC. It is unclear how PKC participates in nuclear gene regulation, but the presence of hormone-activated PKC in the nucleus suggests that it may be an important regulator of transcription or RNA processing (Goss et al., 1994).
The a subunits of G proteins are the most heterogeneous in structure and function. However, the/37 complex is also represented by multiple isoforms. The primary functions of the/3 y complexes are to bind the inactive o~ subunits and to localize the G protein complex to the membrane. However, it has recently been shown that the liberated/37 complexes can directly activate the MAP kinase pathway (see Section III.C) by interacting with the Ras protein (Luttrell et al., 1995).
2. Transcriptional Regulation by G Protein-Activated Pathways Genes that are regulated by cAMP share a regulatory DNA sequence termed the cAMP response element (CRE) (Pestell and Jameson, 1995). The CRE consists of an octanucleotide sequence that is identical, or closely related, to the palindromic consensus 5'-TGACGTCA. The transcription factor that binds to the CRE is called CRE Binding protein (CREB). CREB is a member of a family of transcription factors termed basic-leucine zipper (bZIP) factors, which bind to related response elements. The bZIP proteins form homodimers and heterodimers by virtue of interactions between their leucine zipper motifs, and they interact with DNA through basic residues located in the C-terminal DNA binding domain (Fig. 9) (Meyer and Habener, 1993). CREB can form homodimers that activate transcription. It can also form heterodimers that either activate or inhibit transcription by interacting with isoforms of two closely related bZIP proteins, CREM (CRE modifier), and ATF-1 (activating transcription factor-I) (Habener, 1990). At least seven CREM isoforms are generated by alternative splicing, leading to a rich potential for selective modulation of transcriptional activity. CREB is phosphorylated on a critical regulatory site (Ser 133) by cAMP-dependent protein kinase, calciumcalmodulin (CAM) kinase II and CaM kinase IV. Other phosphorylation sites in the molecule can downregulate CREB activity in response to a variety of kinases (Meyer and Habener, 1993). The analysis of transcriptional activation by CREB led to the identification of a large (molecular weight 300000) CREB binding protein (CBP) that has a high affinity for the phosphorylated CREB dimer. CBP enhances the transcriptional activation driven by CREB (Kwok et al., 1994). CBP was the first identified representative of a wide variety of coactivator proteins that form complexes with hormoneinducible transcription factors. These factors enhance, or in some cases inhibit, the transcriptional activity of regulated genes by altering the interactions of upstream transcription factors with the basal transcription machinery. The mechanisms for most of these coactivator proteins are not yet understood. One of the mechanisms by which CBP increases transcriptional activity is by acetylating histones that are bound to the DNA, thereby increasing the ability of proteins to gain access to specific regions of genes (Ogryzko et al., 1996). CREB is a focal point of regulation by virtue of its ability to be phosphorylated by multiple hormone-regulated kinases, and to interact with other proteins such as CREM and CBP. The
200
SECTION I Biophysical Chemistry, Metabolism, Second Messengers, Ultrastructure
FIGURE 9. (A) Regulation of nuclear gene transcription by G-protein-mediated signaling. Both cAMP, generated in response to Gas coupling, or calcium-calmodulin, generated in response to G~q, causes phosphorylation of the transcription factor cAMP response element binding protein (CREB*). Protein kinase A catalytic subunits (C) are released from the regulatory subunits (R) in response to cAMP binding to the R subunits. CREB interacts with DNA and other nuclear proteins, such as CBP and TFIIB both directly and indirectly. (B) The domain structure of CREB protein. CREB contains a large N-terminal domain that is involved in transcriptional activation and is regulated by phosphorylation on Ser 133, as well as other sites. The C-terminal DNAbinding domain includes both a region rich in basic amino acids and a leucine zipper motif, which is involved in homo- and heterodimerization.
commonness of CRE and CRE-like DNA sequences among inducibly regulated genes results in ensembles of coregulated genes in individual cells, as well as diverse cell-specific responses among cell types.
C. Signaling via Tyrosine Phosphorylation Tyrosine kinase activity is a property that is shared among certain virus-encoded oncogenes and with proteins associated with normal cell signal transduction (Hunter, 1987). This discovery has driven a major advance in our understanding of cellular signal transduction for hormones and other molecules. Two types of hormone-regulated tyrosine kinase pathways are used by vertebrate cells. In some pathways hormones bind to receptors that have intrinsic tyrosine kinase activity in their intracellular domains, allowing direct activation of tyrosine phosphorylation. Examples of receptors that mediate this type of signaling include those for insulin, insulin-like growth factor I (IGF I), epidermal growth factor (EGF), platelet-derived growth factor (PDGF), and fibroblast growth factor (FGF). Other types of receptors, including those for growth hormone (GH), prolactin, erythropoietin, and a variety of hematopoietic cytokines have tyrosine kinases associated with them noncovalently, and the
kinase is activated indirectly following ligand binding. Tyrosine phosphorylation leads to association of SH2 domaincontaining proteins in the receptor signaling complex (Fig. 10). Tyrosine phosphorylation accounts for only 1% of the protein-bound phosphate in a typical cell, and serine/ threonine phosphorylation accounts for the remaining 99%. Phosphorylation on serine and threonine residues is used for many purposes that are not associated with signal transduction, such as creating a highly charged protein that can chelate cations. In contrast, all known tyrosine kinase reactions are directly associated with signal transduction events. Tyrosine kinases operate in part by stimulating a cascade of serine/threonine kinase reactions that impact numerous intracellular targets. The prototypic kinase cascade is initiated when tyrosine phosphorylation of a receptor recruits growth factor receptor-binding protein-2 (Grb-2) and a guanine nucleotide exchange factor (GEF) to the receptor signaling complex. GEFs (of which there are several types) convert inactive GDP-bound Ras to active GTP-Ras by catalyzing the exchange of GDP for GTP. Ras-GTP is anchored to the membrane by a C-terminal lipid moiety and binds the serine/threonine kinase Raf-1. Raf-1 is the first of a series of serine/threonine kinases that phosphorylate, in stepwise fashion, other kinases, culminating in phosphorylation of a
12. Cellular Responses to Hormones
FIGURE 10. Insulin-regulatedsignaling pathways. The insulin receptor is a tetramer consisting of extracellular a subunits and transmembrane/3 subunits that are linked by disulfide bridges. The intracellular domains include a tyrosine kinase region (crosshatched) and several phosphorylatable tyrosines. Phosphorylated tyrosine residues serve as docking sites for IRS proteins and GRB-2. Each of these is phosphorylated on tyrosine residues also, and recruits SH2 domain-containing effector proteins. The known effector proteins include the guanine nucleotide exchange factor mSoS, the SH2 phosphatase SYP, phosphatidylinositol-3' kinase (P1-3K), and phospholipase C7 (PLCT).
mitogen-activated protein kinase (MAP kinase) (Campbell et al., 1995). MAP kinases are unusual bifunctional kinases in that they phosphorylate proteins simultaneously on both serine/threonine residues and on tyrosines. Each of the members of this kinase-activating cascade is represented by multiple isoforms that have different substrate specificities and intraceUular localizations. A major goal in signal transduction research is to resolve the mechanisms that allow selectivity amid an array of closely related protein kinases and kinase substrates.
1. Hormone Receptors That Have Intrinsic Tyrosine Kinase Activity The insulin and epidermal growth factor receptors are archetypes for receptors that have intrinsic tyrosine kinase activity. The signal transduction pathways of these receptors share many features. The insulin receptor (INS-R) is a heterotetrameric protein that contains two identical ce subunits that are extracellular and bind directly to insulin, and two identical/3 subunits that traverse the plasma membrane. The /3 subunits are covalently linked to the ce subunit, and to each other, by disulfide bridges (Fig. 10). The intracellular portion of the INS-R/3 subunit has a tyrosine kinase domain and several phosphorylatable tyrosine residues. Binding of insulin to the extracellular domain causes a conformational change that results in autophosphorylation of several tyrosine residues in the/3 subunit. The main SH2 domain protein that docks to the INS-R is called insulin receptor substrate, which is represented by two isoforms (IRS-1 and IRS-2). The function of IRS proteins is to serve as a scaffold on which the INS-R signaling complex is constructed from nu-
201 merous effector proteins. IRS-proteins are large (MW 185 000) proteins that contain several potentially phosphorylatable tyrosines, at least eight of which are phosphorylated by the INS-R (Myers and White, 1996). Association of IRS proteins with the INS-R is mediated by a particular phosphotyrosine residue on the INS-R, but the IRS proteins do not have conventional SH2 domains, so the binding to INSR is by some alternative, as yet unclear, phosphotyrosinerecognition motif. IRS proteins recruit several SH2 domain proteins that have specific effector activities (Myers and White, 1996). These include the regulatory p85 subunit of phosphotidylinositol-3' kinase (PI-3 kinase), SYP (src-homology phosphotyrosine phosphatase), c-src (and related src-family kinases) and Grb2. Grb2 serves as a docking protein for mSoS (mammalian homolog of drosophila son-ofsevenless). The mSoS protein represents one of the types of GEF, which in this case is employed by the INS-R to activate Ras and a kinase cascade that culminates in MAP kinase activation. The most well-studied effector pathway for receptor tyrosine kinases has been the MAP kinase cascade. There are two main branches of the MAP kinase family: the ERKs (extracellular signal regulated kinases), which activate a ribosomal protein kinase, $6 kinase, and at least two nuclear transcription factors, Elk and Ets; and the JNKs (Jun Nterminus kinases), which phosphorylate the nuclear transcription factor c-Jun, and related proteins. These substrates are believed to represent the "business end" of the signaling pathway, at which specific changes in the activity of the translation and transcription machinery are imposed. Less well studied is another important effector pathway, the PI-3 kinase pathway. PI-3 kinase is capable of generating, in collaboration with phospholipases, unique second messengers that mediate some of the rapid metabolic effects of insulin, such as accelerated glucose uptake (Saltiel, 1996). Phospholipid products activated by PI-3 kinase bind to, and stimulate the activity of protein kinase B (PKB), which is related to cAMP-dependent protein kinase and protein kinase C. PKB promotes cell survival by preventing the activation of the apoptotic pathway (Frank et al., 1997). Epidermal growth factor receptor (EGF-R) also contains intrinsic tyrosine kinase activity in its intracellular domain. But unlike INS-R, the EGF-R is a single polypeptide chain that includes both an extracellular ligand-binding domain and an intracellular kinase domain: Binding of EGF to its receptors causes receptor dimerization, bringing the kinase domains of the receptor pair into close proximity where they transphosphorylate tyrosines in the intracellular domain. SH2 domain proteins associate with the receptor through the phosphotyrosine residues (Fig. 11). Some of the proteins that associate with the EGF-R are identical to those that associate with the INS-R, while others appear to be unique. For instance, whereas both EGF-R and INS-R recruit Grb-2and PI-3 kinase, the IRS proteins that are important for insulin signaling do not appear to be involved in EGF signaling. Receptor tyrosine kinases exert some of their most important actions through activating phospholipid turnover, and generating low molecular weight second messengers. The
202
SECTION I BiophysicalChemistry,Metabolism, Second Messengers,Ultrastructure
FIGURE I 1. Epidermalgrowth factor (EGF) signal transduction to the nucleus. EGF binding to its receptor causes dimerization and activates the tyrosine kinase function of the intracellular domain. Binding of GRB-2 to the receptor results in activation of membrane-bound Ras through the action of mSoS. Ras recruits Raf-family ser/thr kinases, which phosphorylate one of several MAP kinase kinases (MAPKK), leading to phosphorylationand activation of one or more MAP kinases (ERK or JNK). MAP kinases phosphorylate several nuclear transcription factors, as well as cytoplasmic target proteins that are also important for gene expression.
SH2 domain protein phospholipase Cy (PLCy) is activated as a consequence of recruitment to the signaling complexes, and tyrosine phosphorylation in its regulatory domain. PLCy catalyzes the release of IP 3, which generates calcium release from intracellular stores. In this regard the receptor tyrosine kinases converge with the G-protein coupled receptors, which generate I P 3 release by activating the PLCfl isoform. IP3-stimulated calcium release may activate numerous effector proteins, such as PKC, calmodulin, CaM kinase II, CREB, and others. However, it is unclear which, if any, of these known calcium-sensitive proteins participate in physiological effects of the receptor tyrosine kinases.
2. The lak-Stat Pathway and Cytokine Receptor Signaling Growth hormone, prolactin, erythropoietin, interferons, and a variety of interleukins and hematopoietic cytokines bind to receptors that lack intrinsic tyrosine kinase activity and bind to a unique family of nonreceptor kinases called the Janus kinases (Jak). The Jak proteins were named with the Roman gatekeeper god Janus in mind, but the discovery of new tyrosine kinases has become routine enough that many have interpreted Jak to mean "Just another kinase." However, although the receptor tyrosine kinases and src family kinases are closely related in structure and function, the Jaks have
distinctive structures and unique functions in cytokine receptor signal transduction. The primary substrates for Jak kinases are a family of transcription factors named Stats (signal transducers and activators of transcription). The currently known members of the Stat family consist of seven paralogous proteins in vertebrates, and one insect Stat (Damell, 1996). Tyrosine phosphorylation of Stats promotes their dimerization and translocation to the nucleus, where they bind to specific DNA sequences on cytokine-responsive genes. The cytokine receptor superfamily includes two main branches that are distantly related. The receptors for GH and interleukin-6 (IL-6) are representatives of one family, which is large and diverse, and is referred to as the Type I cytokine receptor family; interferon (IFN) receptors comprise the other family, which is less diverse. Similar to the INS-R, IFN receptors are made up of covalently coupled subunits that contain extracellular and intracellular portions. Many of the details of Jak-Stat signaling were initially discovered for interferon signaling, and later found to be widely applicable to other hormones (Damell, 1996). Type I cytokine receptors are single transmembrane polypeptides that are characterized by a distinctive signature in the extracellular domain consisting of two pairs of cysteines and a tryptophan-serine-X-tryptophan-serine (WSXWS) motif. The ligands for the Type I cytokine receptors are collectively referred to as "helix bundle peptide hormones" and they share a similar 3-D structure that consists of four a-helices arranged as two antiparallel pairs. The basic features of signaling in the cytokine receptor superfamily are the following: binding of the hormone causes activation of a Jak-family kinase, which tyrosine-phosphorylates a latent Stat transcription factor. Signaling via Type I cytokine receptors is initiated by the formation of either homodimers (in the case of GH) or heterodimers (in the case of IL-6) (Horseman and YuLee, 1994). To initiate GH-R signaling, a single GH molecule draws together two GH-R molecules. The two binding sites on GH, referred to as sites 1 and 2, have completely different sequences, and the affinity of site 1 is much higher than that of site 2. Although the sequences of the ligand at site 1 and site 2 are different, there is a single binding surface on the GH-R so that the receptors are drawn together in a mirror image conformation (Fig. 12). Homodimerization of the GH-R brings the intracellular domains in close proximity, and receptor-bound Jak2 phosphorylates sites on both the receptor and on Jak2 itself. This creates docking sites for SH2 domain proteins. Stat proteins bind to the receptor via the SH2 domains in their N-terminus, and are then phosphorylated by Jak2 on a tyrosine in their Cterminus. Stat5 is the primary GH activated Stat protein, although Stat 1 and Stat3 are also activated. Activation of Jak2 by GH creates docking sites for SH2 domain proteins other than Stats. As with the insulin or EGF receptor, this results in the recruitment of Grb-2, Shc, mSos, and the activation of a MAP kinase cascade. GH-stimulated Stat proteins and MAP kinase substrates participate in transcriptional activation of genes that are involved in GH actions, including insulin-like growth factor 1, c-fos, and the pro-
12. Cellular Responses to Hormones
203 strate for the gp 130-Jak complex. Activation of the IL-6 receptor in the liver results in the transcription of several acute-phase inflammatory mediators that have Stat3 binding sites in their promoters. Cytokine receptors induce both terminally differentiated cell functions and cell proliferation during hematopoiesis and organ development. The Stat proteins are connected to the induction of differentiated cell functions, but Jak activation is necessary for both proliferation and differentiation in response to cytokines. Bifurcation of the cytokine signaling pathways downstream of Jak activation leads to both generalized signaling pathways that regulate trophic functions such as growth and cell survival (i.e., MAP kinase, PKC, PKB), and to the differentiation-associated activation of Stat proteins (Horseman and Yu-Lee, 1994; Ihle et al., 1995) (Fig. 13). It is not yet clear how the relative activity in each of these paths is established to allow cells
FIGURE 12. Growthhutmu,,c ~i~.al tiatisdtictioii thi-otigh;he JakStat pathway. GH binding to its receptor draws together a receptor pair into a homodimerwith which the tyrosinekinase Jak2 is associated. Jak2 phosphorylates tyrosine residues on itself, the GH receptor, and on Stat proteins, represented here by Stat5. Stat5 phosphorylationresults in nuclear translocation and dimerization, followed by binding to sites on DNA that accelerate the transcription of GH-inducible genes.
tease inhibitor Spi 2.1 (Carter-Su et al., 1996). Like GH, signal transduction for prolactin and erythropoietin is initiated by homodimerization of their receptors, and activation of Jak2. IL-6 is representative of many cytokines that bind to Type I cytokine receptors that heterodimerize after ligand binding (Hibi et al., 1996). The c~ subunit of the IL-6 receptor is the primary ligand binding moiety, but it has only a short intracellular domain and does not have any capability to transduce a signal to the cell's interior. The/3 subunit, which is also known as gpl30, has no significant affinity for IL-6, but it associates with the liganded re subunit, and the long intracellular domain of gpl30 binds and activates both Jakl and Jak2 (Ihle et al., 1995; Hibi et al., 1996). The gp 130 signaling subunit is shared by several other cytokine receptors, including those for leukemia inhibitory factor (LIF) and oncostatin M (OSM). Stat3 is the primary sub-
FIGURE 13. Cytokinereceptors activate multiple signaling pathways. Binding to cytokine receptors activate Jak kinases, which in turn phosphorylate both Stat proteins and Grb-2 which recruits multiple SH2 domain proteins. Stat signaling is primarily associated with tissuespecific expression of differentiation-associated genes. Multiple effector molecules (Ras, MAP kinase, PKC, PKB, and others) are synthesized in response to cytokines and may contribute to generalized trophic actions, cell proliferation, and cell survival.
204
SECTION I Biophysical Chemistry, Metabolism, Second Messengers, Ultrastructure
to proliferate or terminally differentiate in response to a cytokine signal.
D. Signaling via Hormone Receptors That Are Serine/Threonine Kinases Transforming growth factor /3 (TGF/3), bone morphogenetic protein, activin, and inhibin are members of a family of hormones called cystine knot proteins, which are vital developmental regulators. The receptors for this family of polypeptide hormones were recently discovered to consist of an extracellular ligand-binding domain and an intracellular ser/thr kinase domain (Massagur, 1992; ten Dijke et al., 1996). The kinase activity of the receptors is essential for signal transduction by these receptors. These ligands cause their receptors to aggregate into tetrameric complexes that phosphorylate sites on the receptors and effector proteins. A group of proteins called SMADs, named for related genes in C. elegans (small) and drosophila (mothers against decapentaplegic), are phosphorylated upon ligand binding, and translocate to the nucleus where they bind to DNA elements on hormone-regulated genes (Massagur, 1998). Although the nuclear targets of this family of receptors are poorly understood at this time, several genes that are regulated by TGF/3 have been characterized.
E. Signaling via Protein Tyrosine Phosphatases The pervasive involvement of protein phosphorylation in signal transduction is complemented by a variety of phosphatase mechanisms that are both positive and negative regulators of cell functions. Protein tyrosine phosphatases (PTPases), in particular, appear to include members that are highly specific signal transducing molecules. PTPases generally have highly specific "addressing domains" that target them to cell compartments, or to associations with other proteins. These include membrane anchors, cytoskeletal association motifs, SH2 domains, transmembrane helices, and fibronectin-like domains (Dixon, 1996; Tonks and Neel, 1996). Two types of PTPases, the SH2 domain phosphatases and the receptor-like PTPases, are positive signal transducers in mammalian cells. The SH2 domain phosphatases, for example, Syp in mammals and the corkscrew gene product in drosophila, are recruited to phosphotyrosine residues in receptor signaling complexes. These phosphatases are essential for the activation of some, but not all, downstream events following receptor activation. Two mechanisms may account for the involvement of SH2 domain phosphatases in signal transduction. First, the SH2 phosphatase can recruit unique proteins to the complex which, when dephosphorylated, are active effector proteins. Secondly, SH2 phosphatases can dephosphorylate inhibitory isoforms of signaling proteins, thereby increasing the effectiveness of positive effector proteins. The receptor-like PTPases have a single transmembrane domain with the catalytic region oriented to the inside of the cell, and a presumptive ligand-binding domain region outward. The CD45 protein is a T-cell surface antigen that encodes a receptor-like PTPase for which the ligand is unknown. Some of the receptor-like PTPases have extracellular
domains that are similar to known cell-adhesion molecules. These contain fibronectin Type III repeats or immunoglobulinlike repeats that are known to associate with extracellular matrix factors (Dixon, 1996). Presumably, the activity of these PTPases is regulated by their binding to matrix factors and through direct cell-to-cell interactions. These proteins may, therefore, play important roles in differentiation and density-dependent regulation of cell growth.
IV. Integration of Signal Transduction Pathways in Health and Disease Regulation of cell functions by hormones is necessary for both homeostasis and vectoral differentiation. The actions of endocrine hormones drive changes in cell functions in collaboration with a host of locally secreted hormones and cell matrix factors, providing for the execution of appropriate responses to constantly changing environmental conditions. The widespread use of common signaling reactions, such as phosphorylation-dephosphorylation, and the sharing of signaling proteins, as with the nuclear retinoid X-receptor and various kinases, assures that cellular responses are highly coordinated and efficient. The responses of cells are determined by the combination of signaling molecules that are activated at a given time, and by the kinetics of their activation. In the case of MAP kinases, transient activation appears to favor cell differentiation, whereas sustained activation favors proliferation (Marshall, 1995). Human disease states occur when signal pathways are inappropriately activated or disrupted by mutations or infectious agents. A multitude of human diseases are associated with defects in hormone signaling. Diseases arise at every level of signaling; defects are found in the control of hormone production, synthesis, secretion, transport, and in the hormonal signal in target tissues. Defects occur at various levels within the signaling pathways from receptor proteins themselves through downstream effectors (Latchman, 1996). Cholera toxin and pertussis (whooping cough) toxin cause disease by catalyzing ADP-ribosylation of Gas and G i, respectively. In the case of cholera toxin, the inappropriate cAMP synthesis in the gut leads to dysregulated water secretion, uncontrollable diarrhea, and death from dehydration. Many virus-encoded oncogenes, such as v-src and vras, are constitutively active forms of normal cellular signaling proteins. Each of the nuclear hormone receptor genes is represented by at least one genetic disease caused by mutation of the receptor. For example, testicular feminization results from mutations that inactivate the androgen receptor, and vitamin D-resistant tickets results from mutations in the DNA-binding region of the VDR (Hughes et al., 1991). Elucidation of the details of signaling pathways will continue to lead to advances in diagnosis, prevention, and therapy for numerous diseases. It is important to recognize that the phrase "signal tranduction" is merely a metaphor for the processes that are initiated by hormones interacting with cells. There are many ways in which this metaphor is inadequate. For example, most everyday signals, such as a red light or a green light,
12. Cellular Responses to Hormones
can be interpreted unambiguously. In contrast, the meanings of hormone signals are always very ambiguous except in the context of a myriad of prior events and coregulatory factors. Cells continuously integrate multiple signals by coupling each of the signaling components to multiple receptors. Remarkably, these integrated, combinatorial regulatory systems produce all of the organism's normal physiological responses and disease states through a highly conserved repertoire of intracellular signaling molecules.
V. Summary Hormones are regulatory molecules that are secreted from either circumscribed endocrine glands or cells of nonendocrine organs. Hormones activate signal transduction protocols that are highly conserved throughout eukaryotic organisms. Lipophilic hormones, such as steroids, thyroxine, retinoids, and vitamin D enter cells passively and bind to intracellular receptors. Many of these receptors, which are part of a large family of transcription factors, bind known hormones, and others, collectively referred to as orphan receptors, may bind unknown ligands, or function independently of ligand binding. The nuclear receptors are characterized by a conserved DNA binding domain flanked by gene activation and ligand-binding domains. Upon interaction with the cognate hormone, or a hormone analog, the nuclear receptor undergoes significant conformation changes that convert it to an active transcription factor. Hydrophilic protein and amine hormones interact with receptors that are integral membrane proteins. Ligand binding to membrane receptors induces both conformational changes and oligomerization of collaborating signal transduction proteins. A large family of membranebound receptors share a structure that includes seven membrane-spanning a helices (heptahelical receptors). These receptors interact with heterotrimeric GTP-binding proteins, which elicit changes in numerous effector pathways, such as cyclic nucleotide synthesis, phospholipid mobilization, and Ca 2+ influx. Other membrane-associated receptors activate tyrosine phosphorylation of specialized signaling proteins. The protein-tyrosine kinase enzymatic activity may be either an intrinsic property of the intracellular domain of the hormone receptor, or a property of a receptor-associated protein. Tyrosine phosphorylation creates docking sites for effector proteins that include several families of kinases and latent transcription factors. Physiological integration of signal transduction is essential to normal physiology, and defects in signaling are a common feature of disease states.
Bibliography Beato, M. (1989). Gene regulation by steroid hormones. Cell 56, 335-344. Beato, M., Herrliche, P., and Schutz, G. (1995). Steroid hormone receptors: Many actors in search of a plot. Cell 83, 851-857. Berg, J. M. (1989). DNA-binding specificity and steroid receptors. Cell 57, 1065-1068.
205
Bourguet, W., Ruff, D., Chambon, E, Gronemeyer, H., and Moras, D. (1995). Crystal structure of the ligand-binding domain of the human nuclear receptor RXRce. Nature 375, 377-382. Campbell, J. S., Seger, R., Graves, J. D., Graves, L. M., Jensen, A. M., and Krebs, E. G. (1995). The MAP kinase cascade. Recent Prog. Horm. Res. 50, 131-159. Carter-Su, C., Schwartz, J., and Smit, L. S. (1996). Molecular mechanisms of growth hormone action. Annu. Rev. Physiol. 58, 187-207. Danielian, E S., White, R., Lees, J. A., and Parker, M. G. (1992). Identification of a conserved region required for hormone-dependent transcriptional activation by steroid hormone receptors. EMBO J. 11, 1025-1033. Darnell, J. E., Jr. (1996). The JAK-STAT pathway: Summary of initial studies and recent advances. Recent Prog. Horm. Res. 51, 391-404. Dixon, J. E. (1996). Protein tyrosine phosphatases: their roles in signal transduction. Recent Prog. Horm. Res. 51, 405-415. Evans, R. M. (1988). The steroid and thyroid hormone receptor superfamily. Science 240, 889-895. Forman, B. M., Tontonoz, E, Chen, J., Brun, R. E, Spiegelman, B. M., and Evans, R. M. (1995). 15-Deoxy-deltalZ'14-prostaglandinJ2 is a ligand for the adipocyte determination factor PPAR% Cell 83, 803-812. Frank, T. E, Kaplan, D. R., and Cantley, L. C. (1997). P13K: Downstream AKTion blocks apoptosis. Cell 88, 435-437. Glass, C. (1994). Differential recognition of target genes by nuclear receptor monomers, dimers, and heterodimers. Endocr. Rev. 15, 391-407. Goss, V. L., Hocevar, B. A., Thomson, L. J., Stratton, C. A., Bums, D. J., and Fields, A. E (1994). Identification of nuclear beta II protein kinase C as a mitotic lamin kinase. J. Biol. Chem. 269, 19074-19080. Gronemeyer, H. (1993). Transcriptional activation by nuclear receptors. J. Receptor Res. 13, 667-691 Habener, J. E (1990). Cyclic AMP response element binding proteins: A cornucopia of transcription factors. Mol. Endocrinol. 4, 1087-1094. Hard, T., Kellenbach, E., Boelens, R., Maler, B. A., Dahlman, K., Freedman, L. E, Carlstedt-Duke, J., Yamamoto K. R., Gustafsson, J. A., and Kaptein, R. (1990). Solution structure of the glucocorticoid receptor DNA-binding domain. Science 249, 157-160. Herskowitz, I. (1995). MAP kinase pathways in yeast: For mating and more. Cell 80, 187-197. Hibi, M., Nakajima, K., and Hirano, T. (1996). IL-6 cytokine family and signal transduction: A model of the cytokine system. J. Mol. Med. 74, 1-12. Horlein, A. J., Naar, A. M., Heinsel, T., Torchia, J., Gloss, B., Kurokawa, R., Ryan, A., Kamei,Y., Soderstrom, M., Glass, C. K., and Rosenfeld, M. G. (I 995). Ligand-independent repression by the thyroid hormone receptor mediated by a nuclear receptor co-repressor. Nature 377, 397-404. Horseman, N. D. and Yu-Lee, L. - Y. (1994). Transcriptional regulation by the helix bundle peptides: Growth hormone, prolactin and hematopoietic cytokines. Endocr. Rev. 15, 627-649. Hughes, M. R., Malloy, E J., O'Malley, B. W., Pike, J. W., and Feldman, D. (1991). Genetic defects of the 1,25-dihydroxyvitamin D receptor. J. Receptor Res. 11, 699-716. Hunter, T. (1987). A thousand and one protein kinases. Cell 50, 823-829. Ihle, J. N., Witthun, B. A., Quelle, E W., Yamamoto, K., and Silvennoinen, O. (1995). Signaling through the hematopoietic cytokine receptors. Annu. Rev. Immunol. 13, 369-398. Katzenellenbogen, J. A., O'Malley, B. W., and Katzenellenbogen, B. S. (1996). Tripartite steroid hormone receptor pharmacology: Interaction with multiple effector sites as a basis for the cell- and promoter-specific action of these hormones. Mol. Endocrinol. 10, 119-131. Kleiwer, S. A., Umesono, K., Mangelsdorf, D. J., and Evans, R. M. (1992). Retinoid X receptor interacts with nuclear receptors in retinoic acid, thyroid, and vitamin D signaling. Nature 355, 446--449.
206
SECTION 1 Biophysical Chemistry, Metabolism, Second Messengers, Ultrastructure
Kwok, R. E S., Lundblad, J. R., Chrivia, J. D., Richards, J. E, Bachinger, H. E, Brennan, R. G., Roberts, S. G. E., Green, M. R., and Goodman, R. H. (1994). Nuclear protein CBP is a coactivator for the transcription factor CREB. Nature 370, 223-226. Latchman, D.S. (1996). Transcription factor mutations and disease. N. Engl. J. Med. 334, 28-33. Lee, M. S., Kliewer, S. A., Provencal, J., Wright, P. E., and Evans, R. M. (1993). Structure of the retinoid X receptor, a DNA-binding domain: ct helix required for homodimeric DNA binding. Science 260, 1117-1121. Luisi, B. E, Xu, W., Otwinowski, Z., Freedman, L. P., Yamamoto, K. R., and Sigler, P. B. (1991). Crystallographic analysis of the interaction of the glucocorticoid receptor with DNA. Nature 352, 497-505. Luttrell, L. M., van Biesen, T., Hawes, B. E., Koch, W. J., Touhara, K., and Lefkowitz, R. J. (1995). Gt3v subunits mediate mitogen-activated protein kinase activation by the tyrosine kinase insulin-like growth factor 1 receptor. J. Biol. Chem. 270, 16495-16498. Mader, S., Kumar, V., deVereneuil, H., and Chambon, P. (1989). Three amino acids of the oestrogen receptor are essential to its ability to distinguish an oestrogen from a glucocorticoid-responsive receptor. Nature 338, 271-274. Mangelsdorf, D. J. and Evans, R. M. (1995). The RXR heterodimer and orphan receptors. Cell 83, 841-850. Mangelsdorf, D. J., Thummel, C., Beato, M., Herrliche, P., Schutz, G., Umesono, K., Blumberg, B., Kastner, P., Mark, M., Chambon, P., and Evans, R. M. (1995). The nuclear receptor superfamily: The second decade. Cell 83, 835-839. Marshall, C. J. (1995). Specificity of receptor tyrosine kinase signaling: Transient versus sustained extracellular signal-regulated kinase activation. Cell 80, 179-185. Massagur, J. (1992). Receptors for the TGFfl family. Cell 69, 1067-1070. Massagur, J. (1998). TGF-fl signal transduction. Annu. Rev. Biochem. 67, 753-791. Meyer, T. E., and Habener, J. E (1993). Cyclic adenosine 3', 5'monophosphate response element-binding protein (CREB) and related transcription-activating deoxyribonucleic acid-binding proteins. Endocr. Rev. 14, 269-290. Myers, M. G., Jr., and White, M. E (1996). Insulin signal transduction and the IRS proteins. Annu. Rev. Pharmcol. Toxicol. 36, 615--658. O'Malley, B. W. (1990). The steroid receptor superfamily: More excitement predicted for the future. Mol. Endocrinol. 4, 363-369. Ogryzko, V. V., Schiltz, R. L., Russanova, V., Howard, B. H., and Nakatani, Y. (1996) The transcriptional coactivators p300 and CBP are histone acetyltransferases. Cell 87, 953-959. Pawson, T. (1995a). Protein modules and signalling networks. Nature 373, 573-580. Pawson, T. (1995b). SH2 and SH3 domains in signal transduction. Adv. Cancer Res. 64, 87-110. Perlmann, T., Rangarajan, P. N., Umesono, K., and Evans, R. M. (1993). Determinants for selective RAR and TR recognition of direct repeat HREs. Genes Dev. 7, 1411-1422. Pestell, R. G. and Jameson, J. L. (1995). Transcriptional regulation of endocrine genes by second-messenger signaling pathways. In
"Molecular Endocrinology" (B. D. Weintraub, Ed.) pp. 59-76. Raven Press, New York. Pratt, W., Jolly, D. J., Pratt, D. V., Hollenberg, S. M., Giguere, V., Cadepon, E M., Schweizer-Groyer, G., Cartelli, M. G., Evans, R. M., and Baulieu, E. E. (1988). A region in the steroid-binding domain determines formation of the non-DNA-binding glucocorticoid receptor complex. J. Biol. Chem. 263, 267-273. Rastinejad, E, Perlmann, T., Evans, R. M., and Sigler, P. B. (1995). Structural determinants of nuclear receptor assembly on DNA direct repeats. Nature 375, 203-211. Renaud, .J-P., Natacha, R., Ruff, M., Vivat, V., Chambon, E, Gronemyer, H., and Moras, D. (1995). Crystal structure of the RARy ligand-binding domain botmd to all-tram retinoic acid. Nature 378, 681--689. Saltiel, A. R. (1996). Diverse signaling pathways in the cellular actions of insulin. Am. J. Physiol. 270, E375-E385. Schwabe, J. W. R., Chapman, L., Finch, J. T., and Rhodes, D. (1993). The crystal structure of the oestrogen receptor DNA-binding domain bound to DNA: how receptors discriminate between their response elements. Cell 75, 567-578. Schwabe, J. W. R., Neuhaus, D., and Rhodes, D. (1990). Solution structure of the DNA-binding domain of the oestrogen receptor. Nature 348, 458--461. Speigel, A. M., Shenker, A., Simonds, W. E, Weinstein, L. S. (1995). G-protein dysfunction in disease. In "Molecular Endocrinology" (B. D. Weintraub, Ed.) pp. 297-318. Raven Press, New York. ten Dijke, P., Miyazono, K., and Helden, C. H. (1996). Signaling via hetero-oligomeric complexes of type I and type II serine/threonine kinase receptors. Curr. Opinion Cell Biol. 8, 139-145. Thummel, C. S. (1995). From embryogenesis to metamorphosis: The regulation and function of Drosophila nuclear receptor superfamily members. Cell 83, 871-877. Tonks, N. K., and Neel, B. G. (1996). From form to function: Signaling by protein tyrosine phosphatases. Cell 87, 365-368. Truss, M., and Beato, M. (1993). Steroid hormone receptors: Interaction with deoxyribonucleic acid and transcription factors. Endocr. Rev. 14, 459-479. Umesono, K., and Evans, R. M. (1989). Determinants of target gene specificity for steroid/thyroid hormone receptors. Cell 57, 1139-1146.
Umesono, K., Murikami, K. K., Thompson, C. C., and Evans, R. M. (1991). Direct repeats as selective response elements for the thyroid hormone, retinoic acid, and vitamin D 3 receptors. Cell 65, 1255-1266. Wagner, R. L., Apriletti, J. W., McGrath, M. E., West, B. L., Baxter, J. D., and Fletterick, R. J. (1995). A structural role for hormone in the thyroid hormone receptor. Nature 378, 690-697. Yu, V., Delsert, C., Andersen, B., Holloway, J. M., Devary, O. V., Naar, A. M., Kim, S. Y., Boutin, J.-M., Glass, C. K., and Rosenfeld, M. G. (1991). RXR: A coregulator that enhances binding of retinoic acid, thyroid hormone, and vitamin D receptors to their cognate response elements. Cell 67, 1251-1266. Zhang, E L., and Casey, P. J. (1996). Protein prenylation: Molecular mechanisms and functional consequences. Annu. Rev. Biochem. 65, 241-269.
Jeffrey C. Freedman and Nicholas Sperelakis
Diffusion and Permeability
1. I n t r o d u c t i o n In order to understand the fundamentals of membrane transport, as well as the mechanisms for development of the resting electrical potential of cells, it is first necessary to consider the basic processes of diffusion and membrane permeability. Therefore, this chapter discusses fundamental principles that are utilized in subsequent chapters in this book. Molecules of gases and liquids, and of all dissolved solutes, are continuously in motion. The velocities of individual molecules vary tremendously, as do their kinetic energies, in accordance with the Maxwell-Boltzmann distribution. Diffusion is the process whereby particles in a gas, liquid, or solid tend to intermingle due to their spontaneous motion caused by thermal agitation. Any diffusing substance tends to move from regions of higher concentration to regions of lower concentration, until the substance is uniformly distributed at equilibrium. At any given temperature above absolute zero, the molecules of a substance continue to move, even at equilibrium, but the net movement is zero. In the absence of convection, which refers to bulk flow of solvent containing solutes such as that caused by stirring, the movement of the molecules is by diffusion only. Diffusion occurs because of the random thermal motions of the molecules. Molecular collisions that transfer momentum are more probable in a region of higher concentration. Particles flow from a region of high concentration to one of low concentration whenever the concentration is greater behind than in front of any particle; that is, whenever a gradient of concentration exists, such a particle has a greater probability of being bumped from behind than from in front. The net result of these collisions imparts a forward movement to the center of mass of the particles, thus creating a flow of particles down the concentration gradient. Consequently, if a region of high solute concentration is adjacent to one of low solute concentration, separated by an imaginary plane, it is probable that more molecules per unit of time will be crossing the plane from the side of higher concentration to the side of lower concentration than in the opposite direction. Thus there are fluxes, or movements of molecules, in both directions (known as unidirectional fluxes), but the net
Cell Physiology Sourcebook: A Molecular Approach, Third Edition
2 09
flux is from the side of higher concentration to the side of lower concentration. Now if the imaginary plane were replaced with a thin membrane permeable to the molecules, then the same situation would apply; the particles would diffuse from the side of higher concentration to the side of lower concentration across the membrane. In a living cell it is often assumed that the inside and outside solutions are relatively well stirred, and diffusion of most substances through the cell membrane is much slower than that through a free solution. Therefore our primary concern is about the diffusion of substances through the cell membrane, the rate-limiting step. For simplicity we assume that the solutions on either side are well stirred and that there are no concentration gradients within the bulk solutions on either side of the membrane (although in reality there probably are unstirred layers near the membrane). We confine ourselves here to the diffusion of small molecules or ions across membranes. For thin membranes, we first consider the case of diffusion without partitioning, as illustrated on the left side of Fig. 1, in which the diffusing molecule is freely able to enter the membrane. We will then introduce the effect of partitioning into the membrane, a process that involves a change of chemical potential upon
DIFFUSION WITHOUT PARTITIONING
WITH PARTITIONING
,,cl c2
c2
_.,.,.......
I
inl
I
out
in
out
FIGURE 1. Concentration gradients across the membrane during diffusion with and without partitioning.
Copyright 9 200~ by Academic Press. All rights of reproduction in any form reserved.
210
SECTION 11 M e m b r a n e Potential, Transport Physiology, Pumps, a n d Exchangers
entrance into the membrane phase, as illustrated on the right side of Fig. 1. In either case, we disregard structural effects that were illustrated by Danielli (1943) as a series of potential energy barriers within the membrane.
Ii. Fick's Law of Diffusion The fundamental law describing diffusion was discovered empirically by Adolph E. Fick (1829-1901), Demonstrator of Anatomy in ZUrich, who in 1855 noted an analogy between diffusional transport and Fourier's law describing the flow of heat. Fick's law of diffusion was deduced theoretically in 1860 by James Clerk Maxwell (1831-1879) from the kinetic theory of gases, and it should be understood from the outset that the derivation of Fick's law makes the following assumptions: (1) the continuum hypothesis applies, that is, the molecular concentration is sufficiently high to ensure the validity of statistical laws; (2) the average duration of a collision is small compared to the average time between collisions so that all collisions are binary, a condition that would pertain to dilute solutions; (3) the distribution of molecular properties varies only slowly in time and space; (4) the particles move independently so that there is no correlation between the properties of any two particles; (5) quantum effects are negligible so that classical mechanics can be used to describe molecular collisions; (6) energy, momentum, and mass are conserved in every molecular collision; and (7) in liquids, the diffusing particles are much larger than the solvent molecules. These assumptions necessarily make the quantitative description of diffusion an approximation. A more accurate and complete description, especially for diffusion in liquids, depends upon knowledge of the structure of the medium in which diffusion occurs. Consider two compartments separated by a membrane of thickness l, in which a substance diffuses from a cell (denoted as compartment 2 with volume V2 and high concentration c2) across the cell membrane into the medium (denoted as compartment 1 with volume V 1 and low concentration c l). We assume that the temperature and pressure are the same in both compartments. For such a system, Fick's first law of diffusion states that the flux density (J), given by dN/Adt mol/(cm2-s), or the number (dN) of moles that diffuse across a unit area (A, cm 2) of the membrane during an interval of time (dt, s), in a direction (x) perpendicular to the membrane, is directly proportional to the concentration gradient dc/dx (mol/cm3/cm) across the membrane, the constant of proportionality being the diffusion coefficient D (cm2/s). J-
dN2 dc =-O~ Adt dx
(1)
The diffusion coefficient D, which is the flux density when the concentration gradient is unity, indicates how much flux will occur for a given concentration gradient. For a thin membrane, when the solution is well stirred on either side, the concentration gradient in the steady state is equal to the difference in the concentration (Ac = c 2 - cl) of
the solute on both sides of the membrane divided by the thickness of the membrane (Ax), so that Ac J =-D ~ Ax
(2)
Assume that the concentration gradient across the membrane is linear over the thickness (1) of the membrane, which is a reasonable assumption for a thin membrane. Then AclAx = (c 2 - Cl)/l, and the flux density (J) is J --D
C2 -- C1
l
(3)
Note that when the concentration gradient is defined as the internal minus the external concentration, then the minus sign in Fick's law connotes that an effiux is negative and an influx is positive. Thus the net flux is equal to the concentration gradient times the diffusion coefficient. If c 1 is zero, then according to Fick's law, the flux density is linearly related to the concentration of the solute on the side from which the flux originates. Consequently, if one experimentally finds a linear dependence of flux on concentration, then the mechanism of transport is consistent with, but not proof of, simple diffusion. Finding a linear dependence of flux on concentration over a particular range of concentrations does not prove that diffusion is the mechanism of transport because the flux could deviate from linearity at a higher concentration. To examine the rate of change of concentration (dc2/dt), divide both sides of Eq. 1 by the intracellular volume (V2) and multiply both sides by A, giving dc 2
dt
dN______L =
V2 dt
=-
\ D___~Ad__s DA / =/ c2 - c l) V 2 dx V2 l
(4)
Integration of Eq. 4 shows that the time course of diffusion, or the change in intracellular concentration of a diffusing substance as a function of time, is exponential (see the Appendix following this chapter). For two-dimensional diffusion, the time required for diffusion to become 50, 63, or 90% complete varies directly with the square of the distance, and inversely with the diffusion coefficient. Thus, diffusion is extremely fast over short distances (e.g., 10-1000 nm), but is exceedingly slow over long distances (e.g., 1 cm). For a small particle like K § or acetylcholine (ACh+), with a D value of about 1 • 10-5 cm2/s, the time required for 90% equilibration to be reached over a distance of 1 Ixm is about 1 ms. Table 1 shows how the time required changes as a function of distance. The chef knows that the time required for a meat roast to cook in the middle is dependent on the thickness of the roast, and the physiologist knows that there is a critical thickness (e.g., 0.5-1.0 mm) of a muscle bundle in an incubation bath that allows adequate diffusion of oxygen to the core of the strip, no matter how vigorously the bath is oxygenated.
ili. Diffusion Coefficient The diffusion coefficient (D) is a constant for a given substance and membrane under a given set of conditions. The dif-
13. Diffusion and Permeability
21 1
TABLE 1 Calculations of Time R e q u i r e d for Diffusion to Become 90% Complete Distance, A
Time, t
100 A
0.1 lxs
0.1 Ixm
0.01 ms
1 Ixm
1 ms
10 Ixm
100 ms
100 Ixm
10 s
1 mm
16.7 min
1 cm
28 h
D M 1/2 = constant
fusion coefficient of substances in free solution is dependent upon molecular size (and shape, for large molecules). For ions, the smaller the unhydrated radius, the greater the charge density, which means that more water molecules are held in the hydration shells giving a larger hydrated radius; the larger water shell causes diffusion to be slower (see Chapter 1). The hydrated and unhydrated radii of some relevant ions are given in Table 2. Thus, although Na + has a lower atomic weight than K +, Na + attracts and holds a larger hydration shell because of its greater charge density (same charge as K +, but smaller unhydrated ion size), and therefore Na + diffuses in water considerably slower than K + (ratio of about 1:2).
"It
0"2 pM1/2
.
.
.
.
.
Singly
equivalent conductivity (cm2/S.eq)
Activity coefficienta
Crystallographic radius (~)
(6)
where k is Boltzmann's constant, T is the absolute temperature, N A is Avogadro's number, o- is the radius of the particle, p is the pressure, and M is the molecular mass (see Cunningham and Williams, 1980). The random molecular motions due to thermal energy are counteracted by intermolecular attractive forces, but in dilute gases these forces are small because the molecules are relatively far apart. The kinetic energy of gas molecules is given by nMv2/2, where n is the number of moles, M is their molecular weight, and v is their mean velocity. At a given temperature, the average velocities of the molecules are inversely proportional to the square root of their masses. Because the velocity of the colliding molecules influences the speed of diffusion, the diffusion coefficient is also inversely proportional to the square root of the molecular weight, as stated in Graham's law. The Einstein-Stokes e q u a t i o n applies to diffusion in liquids when spherical solute molecules are much larger than the solvent molecules, in which case the diffusion coefficient
Physicochemical Properties of Selected Ions
Limiting Ion
(5)
Graham's law was originally discovered by experiment, and was later derived theoretically by Maxwell. From the kinetic theory of gases, the diffusion coefficient is related to molecular parameters as follows:
Note: Diffusion time varies with square of diffusion distance. These values are for a substance having a diffusion coefficient (D) of about 1 • 10-5 cm2/s in free solution. The diffusion time is inversely proportional to D. Einstein's approximation equation is A = v/D--/, where A is the mean displacement of the diffusing particles along the x-axis at any time (t).
TABLE 2
The diffusion coefficient is inversely related to the resistance to free diffusion and is very much less in water than in a gas. G r a h a m ' s law states that the diffusion coefficient of a gas is inversely proportional to the square root of the molecular weight, and is often given in the following form:
hydrated radius b (,~)
Hydrated volumeC (,k3)
Heat of hydration (kcal/g ion)
Number of water molecules ~
Na§
50.10
0.778
0.96
3.67
150
-115
(4-6)
K+
73.50
0.770
1.33
4.05
(150)
-90
(3-6)
C1-
76.35
--
1.81
3.92
90
-59
(0-6)
Mg2+
53.05
0.528
0.65
3.60
360
-501
12
Ca2+
59.50
0.518
0.99
3.70
310
--428
10
Sr2+
59.45
0.515
1.13
3.85
310
-381
10
Ba2+
63.63
0.508
1.35
4.08
290
-347
9-10
Note: Values taken from R. A. Robinson and R. H. Stokes, Electrolyte Solutions, London, Butterworth & Co. (Publishers), Ltd., 1959, and from J. O. M. Bockris and B. E. Conway, Modern Aspects of Electrochemistry, London, Butterworth & Co. (Publishers), Ltd., 1954. a Activity coefficient for each cation given for the C1- salt at 0.1 M and at 25 ~ b Singly hydrated radius equals crystallographic radius plus 2.72 ,~ (diameter of water molecule) for cations with crystal radii between 0.9 and 1.7/~, and crystal radius plus 2.23 ~ for anions. Values for singly hydrated radii taken from Mullins (1961). c Values in parentheses are estimated.
212
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
of the solute was found by Einstein (1926) to have the following relationship to the viscosity (rl) of the solvent and to the radius (r) of the solute particle: O=~ 6r
RT A
(7)
where R is the gas constant, T is the absolute temperature, and N a is Avogadro's number. This equation is more applicable to the diffusion of solutes of colloidal size than to molecules that approximate the size of the solvent. Temperature also affects the rate of diffusion. The relative increase in the diffusion coefficient when the temperature (T) is raised by 10 ~ is known as Q10" From the EinsteinStokes equation, DT+IO _ T+I 0 tit DT T ~T+lO
(8)
In dilute solutions for which the viscosity doesn't vary much over a range of 10 ~ Q10 should be the same for all molecules. Neglecting changes in r/, Q~0 is 1.03 between 25 ~ and 35 ~ A value of Q10 much greater than 1 is often taken to indicate that some process other than diffusion is probably responsible for transport. In liquids, the solute and solvent molecules are close together and within each others' spheres of mutual attraction. Hence, for any solute to move, it must first break away from its surrounding molecules. Diffusion in liquids consequently occurs in a discontinuous manner; a molecule is able to move only when it has acquired sufficient energy by collision to break away from its neighbors and push aside other molecules. Thus, the molecule diffuses in a series of jumps, each jump requiring a certain critical activation energy. When the attractive forces between molecules are weak, diffusion is rapid. Danielli (1943) derived from kinetic theory a relationship for D that includes the temperature coefficient (Q10) raised to a power as follows: DM1/2 =10 o(T+10)/10 - constant
(9a)
or, D~
1 M1/2 t3(T+10)/lo
(9b)
~-10
This relationship is applicable at constant temperature. The Q10 for diffusion of a substance in water depends on the activation energy necessary for a jump. If a large amount of energy is required (high Q l0), only a few molecules have the necessary energy to diffuse at any moment; raising the temperature increases the number of molecules with the required energy and thereby appreciably speeds up the net rate of diffusion. If only a small amount of activation energy is required (low Q10), a greater fraction of the molecules possesses this minimal energy at any moment, and so raising the temperature has less of an effect on the net rate of diffusion. In either case, elevation of temperature increases the speed of the molecules and causes an increase in the rate of diffusion. When Q10 is close to 1.0, then Eqs. 9a and 9b reduce to Graham's law (Eq. 5). The Q10 for diffusion of Na § or K § in water is only about 1.22. Equations 9a and 9b also apply
when the permeability coefficient (K), as defined in Section IV, is substituted for the diffusion coefficient (D) of the solute in the membrane. The cell membrane constitutes a barrier to diffusion. In penetrating through the lipid bilayer matrix, a noncharged small solute molecule must move as follows: (1) detach itself from its surrounding solvent molecules and jump into the membrane phase; (2) move through the thickness of the membrane, perhaps by a series of small jumps over energy barriers; and (3) detach itself from the membrane environment and jump into the solvent phase again on the opposite side of the membrane. The permeability of the membrane to the solute depends greatly on the activation energy required for the molecule to jump into and out of the membrane. If the activation energy is high, the permeability is low and the Q~0 is high; if the activation energy is low, then the permeability is great and the Q~0 is low.
IV. D i f f u s i o n A c r o s s a M e m b r a n e with Partitioning When a solute partitions into the membrane, or is excluded from it, the concentration gradient across the membrane phase itself differs from that calculated from the internal and external concentrations in the bulk solution. For example, suppose that a solute is at 10 mM in the internal solution and at 1 mM in the external solution; the concentration gradient would be c 2 - c 1 = 10 - 1 = 9 mM. Now further suppose that the solute dissolves into the membrane phase such that the concentration just within the membrane at the internal boundary is 10-fold greater than in the internal solution, and the same ten-fold partitioning occurs at the external boundary. With such partitioning, the concentrations in the membrane would be 100 mM at the internal boundary, and 10 mM at the external boundary, constituting a gradient of 100 - 10 = 90 mM, or a tenfold higher concentration gradient within the membrane than the gradient based on the concentrations in the internal and external bulk solutions. We can now apply Fick's first law to the case of diffusion across a membrane with partitioning of the solute into the membrane in a partition/diffusion model of transport. Letting the superscript m denote the concentrations in the membrane phase, we have from Eqs. 3 and 4,
j--7 dc2 _ dt -
O
m
- c;
m_ Cl ) 1V2 ( C2
DA
(lO) (11)
The partition coefficient (13) relates the concentration just within the membrane at the boundary to the external or internal concentrations. In a homogeneous membrane, 13 is assumed equal at the internal and external boundaries.
_ mc___,_c7 C1
(12)
C2
Substituting the expressions for/3 into Eq. 11 results in the following transport equation:
13. Diffusion and Permeability
dc2_
213
\
flD A
(13)
For a partition/diffusion process, the permeability coefficient (K, cm/s) is defined as the product of the partition coefficient (fl) and the diffusion coefficient (D) divided by the thickness (l) of the membrane:
/30
s: - ~
l
(1.4)
Note that although diffusion coefficients apply to diffusion in free solution or in membranes, permeability coefficients apply only to membranes. The higher the diffusion coefficient for movement of a substance across a membrane, the higher the permeability coefficient. The permeability (P, s-1) itself is defined as the permeability coefficient (s:) times the area (A) of the membrane divided by the volume (V) of the cell:
A
~OA
P=K . . . . V 1 V
(15)
so that from Eq. 13,
dc 2
d, Therefore the permeability (P) is the rate of change of concentration with time (dc 2/dt) for a concentration gradient of unity. In general, the permeability is the proportionality constant that relates the rate of change of concentration with time to the concentration gradient. Note that the unit of P in Eqs. 15 and 16 is s-1, whereas the permeability coefficient (so) defined in Eq. 15 has units of cm/s. In electro-
physiology, however, the symbol P is often also used to represent the permeability coefficient with units of cm/s. We shall see that the permeability depends on the model chosen for the mechanism of membrane transport. Now we can understand the basis for Overton's classical rule that the rate of diffusion of many substances across cell membranes correlates with the oil/water partition coefficient. The rate of permeation of a water-soluble substance is primarily determined by how easily the substance passes from water to lipid; its oil/water partition coefficient is a measure of this ease. The more a substance dissolves in the membrane, the greater its concentration gradient within the membrane, and the faster the rate of diffusion across the membrane. The classical data for permeation of nonelectrolytes including alcohols, amides, and various substituted ureas into algae, red blood cells, and other cells is a striking example of this correlation (see Lieb and Stein, 1986, for a particularly insightful discussion). The diffusion coefficient across cell membranes for various substances is generally greater when the molecular size of the substance is small and when the lipid solubility is high; that is, small molecules of high lipid solubility (i.e., less polar and nonpolar molecules) penetrate most quickly through the membrane. Most nonpolar molecules pass directly through the lipid bilayer matrix of the membrane, that is, not through special sites (e.g., water-filled pores or channels). Small charged ions (e.g., Na +, K +, CI-, Ca 2§ apparently pass through water-filled channels, some of which
have a voltage-dependent gating mechanism; these channels can exhibit a high degree of selectivity for specific ions, with the selectivity orders not being solely based on the hydrated or unhydrated sizes of the ions. One type of ionophore, valinomycin, forms a hydrophobic cage around K +. The K+-ionophore complex is lipid-soluble and passes rapidly through the lipid bilayer matrix. Transport proteins are normally present in the cell membrane for net uptake of nutrients such as glucose and amino acids. Some of these are for downhill transport only, that is, down the electrochemical gradient. One type of proteinmediated transport in cells (e.g., for glucose) is known as facilitated diffusion, a type of transport that contrasts with the so-called simple diffusion across the lipid bilayer matrix as described in the preceding paragraphs. Another type of mediated transport is known as exchange diffusion, in which one molecule of substance (or ion) inside the cell is exchanged for one molecule outside the cell, in which case there is no net movement. In either type of mediated transport, the rate of downhill movement of the substance across the membrane is enhanced by the transport protein. Mediated transport systems differ from simple diffusion in that they exhibit saturation kinetics; that is, the rate of transport increases with an increase in the substrate concentration up to a maximum, after which the rate of transport levels off due to a finite number of available sites. Other characteristics of mediated transport systems include competitive inhibition, in which two or more substances may compete at the same site for binding to the transporter, and noncompetitive inhibition, in which a nontransported substance may bind to the transporter at a site different from the transport site and thereby alter or prevent binding of the usual substrate. Mediated transport exhibits greater specificity than simple diffusion; for example the rate of mediated transport of D-glucose is much greater than L-glucose, a difference that could not occur with simple diffusion. Because mediated transport resembles an enzymatic process involving protein conformational changes, the temperature coefficient (Q10) is also higher than for simple diffusion.
V. Electrodiffusion Combining Eqs. 11, 12, and 14, the flux density (J) across the membrane for a partition/diffusion process is given by
J - - K ( c 2 - Cl)
(17)
Hence the net flux density of a nonelectrolyte across a membrane is equal to the permeability coefficient (K) times the difference in concentration across the membrane. The permeability coefficient for K + (KK) across resting striated muscle membrane is about 1 x 10-6 cm/s. Some values for KI~ (represented by the symbol PK in cm/s) are given in Table 3 for some cardiac tissues. The unidirectional influx (Ji) and efflux (Jo) are defined by
Ji=-Kc1
(18a)
Jo = -/ 1.0, there is a net inward current carried by Na + ion, coupled with net Ca 2+ effiux from the cell. This represents forward mode of operation of the exchanger. For AG ratios < 1.0, there is a net outward current carried by Na + ion, coupled to a net Ca 2+ influx into the cell. This represents reverse mode of operation of the exchanger. The equilibrium potential or reversal potential for the Ca2+-Na § exchanger (ENa_Ca) , for an exchange ratio of 3 Na+: 1 Ca 2+ is ENa_Ca = 3ENa- 2Eca
(7)
where ENa and ECa are the equilibrium potentials for Na + and Ca 2§ respectively, as calculated from the Nemst equation. Thus, the ENa_Ca varies with the [Na+]i level and during changes in [Ca2+]i levels that occur with contraction. ENa_Ca is less negative (more positive) when [Ca2+]i is elevated, and shifts to more negative potentials when [Na+]i is elevated (Fig. 4B). Therefore, when a myocardial cell changes from the resting potential ( c a . - 8 0 mV) to the AP plateau (ca. +20 mV), simultaneous with [Ca2+]i being elevated from about 0.1 to 3 pM, the exchanger switches to reverse mode of operation, with Ca 2+ influx. As stated previously, this Ca 2+ influx can be a significant source of the total Ca 2§ influx during excitation-contraction coupling. If the [Na+]i level were 25 or 30 mM, the exchanger would always operate in the reverse mode at the physiological E m levels.
a parallel resistance (em) and capacitance (Cm). As discussed previously, the capacitance results from the lipid bilayer matrix of the membrane, and the resistance or conductance (G m) resuits from the protein ion channels floating in the lipid bilayer and spanning it. The resting potential is depicted as a battery in series with Rm, with the negative pole facing inside the cell. In Fig. 5B, the equivalent circuit for the resting membrane is shown further expanded, by breaking up the membrane conductance (Gm)into its four component parts, one for each ion of major importance electrophysiologically, namely K +, CI-, Na § and Ca 2§ The respective conductances are gK, gcl, gNa' and gca" The equilibrium potential (E) for each ion is placed in series with the conductance pathway for that ion, as depicted. The polarity of each battery is as shown; namely, the pole facing inwards is negative for K + and C1-and positive for Na § and Ca 2§ These polarities are based on the directions of the concentration gradients and charge on the ions. The reasons and mechanisms are discussed as follows. B. Nernst Equation
For each ionic species distributed unequally across the cell membrane, an equilibrium potential (Ei) or battery can be calculated for that ion from the Nernst equation (for 37 ~
Ei Ei
- R T In Ci
_
(8)
- 2 . 3 0 3 R T log Ci
_
Co
(8a)
J
IV. E q u i l i b r i u m P o t e n t i a l s A. Equivalent Electrical Circuit
The electrical equivalent circuit of the cell membrane at rest is depicted in Fig. 5. In Fig. 5A, the membrane is indicated as
Ei
_
- 6 1 m______VV log
z
C
i
(8b)
CO
where C i is the intemal concentration of the ion, C Ois the extracellular concentration, R is the gas constant (8.3 J/mol.K), T is the absolute temperature in kelvins (K = 273 + ~
2,27
14. Origin of Resting Membrane Potentials
is the Faraday constant (96 500 C/eq), and z is the valence (with sign). Thus, z ~ = C/mol. Taking the R T / ~ constants and the factor of 2.303 for conversion of natural log (In) to log to the base of 10 (log10) gives 2
-RT - 2 . 3 0 3 RT - - ~ In N ~ l o g N - - 6 1 mV log N
(8c)
Therefore, the Nemst equation (Eq. 8) becomes
Ei_ - 2 . 3 0 3 RT C z~ log ~oo - 6 1 mV Ei- ~ l o g
C. ~
(8d) (8e)
z
The --61 mV constant (2.303 RT/~) becomes -59 mV at 22 ~ A derivation of the Nernst equation is given in the Appendix (part I) to this chapter. The Nemst equation gives the
potential difference (PD) (electrical force) that would exactly oppose the concentration gradient (diffusion force). Only very small charge separation (Q, in coulombs) is required to build a very large PD. E - -Qm
Cm
of energy. In many cell types, only C1- ion appears to be at or near equilibrium, whereas Na +, Ca 2+, and K + are actively transported. Even H + ion is off equilibrium, E H being closer to zero potential (see Table 2). If H + were passively distributed, the negative intracellular potential would pull in more H + ions, causing [H+]i to increase, making the cell interior more acidic. The mechanism for development of the equilibrium potential is depicted in Fig. 6. To show the development of an equilibrium potential, we can use an artificial membrane (e.g., one made of celloidin) to separate two solutions, that is, to form a concentration cell. This membrane contains negatively charged pores, which therefore allows cations (like K +) to pass through, but prevents anions (like C1-) from passing. This is because like charges repel one another, and unlike charges attract one another. Therefore, in this particular membrane, K + is permeant and C1- is impermeant. If one side (side 1) contains a salt like KC1 at a concentration higher (e.g., 0.10 M) than that in the other side (side 2) (e.g., 0.01 M), then a steady PD is very quickly built up across the membrane. As can be calculated from the Nernst equation,
(9)
where C m is the membrane capacitance. This is further discussed in the next section. For the ion distributions given previously (see Table 2), the approximate equilibrium potentials are ENa = +60 mV
Eca = +130 mV E K = - 9 3 mV E c l = - 8 0 mV The sign of the equilibrium potential represents the inside of the cell with reference to the outside (see Fig. 5). Because Na + is higher outside (ca. 145 mM) than inside (ca. 15 mM), the positive pole of the Na + battery (ENa) is inside the cell. The concentration gradient for Ca 2+ is in the same direction as for Na + (1.8 mM [Ca2+]o and about 1 x 10-7 M [Ca2+]i), and so the positive pole of Eca is inside. K + is higher inside (ca. 150 mM) than outside (ca. 4.5 mM), and so the negative pole is inside. Because C1- is higher outside (ca. 100 mM) than inside (ca. 5 mM), the negative pole is inside. Voltages are, by convention, given for the inside with respect to the outside.
C. Concentration Cell In a concentration cell (essentially a two-compartment system separated by a membrane), the side of higher concentration becomes negative for cations (positive ions) and positive for anions (negative ions). Any ion whose equilibrium potential is different from the resting potential (e.g.,-80 mV for a myocardial cell or skeletal muscle fiber) is off equilibrium and therefore must effectively be pumped at the expense 2 In (or 1Oge, or 10g2.717) is the log to the base of e (2.717), and loga0 is the log to the base of 10.
FIGURE 6. Upper diagram: Concentration cell diffusion potential developed across artificial membrane containing negatively-charged pores. The membrane is impermeable to C1- ions, but permeable to cations such as K§ Concentration gradient for K§ causes a potential to be generated, the side of higher K+ concentration becoming negative. Lower diagram: Expanded diagram of water-filled pore in the membrane, showing the permeability to K+ ions, but lack of penetration of C1- ions. Potential difference is generated by charge separation, a slight excess of K§ ions being held close to the right-hand surface of the membrane; a slight excess of C1- ions is aggregated close to the left surface.
228
SECTION II M e m b r a n e Potential, Transport Physiology, Pumps, and Exchangers
for a tenfold difference in concentration of the permeant monovalent cation (K+), the PD would be -61 mV at 37 ~ ( o r - 5 9 mV at a room temperature of 22 ~ This PD is between the two solutions and expressed across the membrane. Side 1 (side of highest K + concentration) becomes negative with respect to side 2. The PD is developed because of the tendency for diffusion (diffusion force) from high concentration to low concentration. This is based on the random thermal motion of the ions (particles), somewhat related to Brownian motion of larger particles. That is, the side of higher concentration has a greater probability of K + ions moving from side 1 to side 2 than in the reverse direction, based on the greater number of particles, all moving in random directions. Therefore, there will be a loss of positive charges (K § ions) from side 1 and a gain of positive charges in side 2. Because negative charges (C1- ion) cannot accompany the positive charges, as the membrane was made impermeable to anions, a charge separation is built up across the membrane. It can now be readily understood why side 1, the side of higher cation concentration, becomes negatively charged (due to loss of positive charges) and why side 2, the side of lower cation concentration, becomes positively charged (due to gain of positive charges). The charge separation is very tiny, and they stay plastered very close to the membrane. That is, for the example depicted in Fig. 6, side 1 will have the small excess of CI- ions held very close to the membrane, and side 2 will have the small excess of K § ions held very close to the membrane. The force holding them there is called the electrostatic or Coulombic force, based on the attraction between unlike charges. This is related to Coulomb's law and the nature of capacitors discussed in Chapter 70. In the two bulk solutions, the law of electroneutrality is upheld; that is, for every cation (K +) there is a nearby anion (C1-). Thus, the charge separation occurs only directly across the membrane and is very tiny with respect to the total number of charges in the two solutions. In fact, after equilibrium is reached (within a few seconds), the most sensitive chemical analyses would fail to detect the very slight decrease in K + in side 1 or gain of K + on side 2. Thus, the system comes to equilibrium quickly and with very little charge (K § and C1-) separation. That is, K + does not continue to have a net movement from side 1 to side 2 until the concentrations become equal. Why not? The answer is that the small charge separation produces a large PD across the membrane, and this PD is in such a polarity that it antagonizes further net movement of K § from side 1 to side 2. That is, the positive voltage that is developing on side 2 repels the positively charged K + ions because like charges repel. At equilibrium, these two forces become equal and opposite, and there is no further net movement of ions. Unidirectional fluxes of K + ions would still occur (because of their random thermal motion), but these would be equal and opposite and so no further net flux would occur. In the example selected, KCI was used. However, any salt, such as NaC1, CaC12, or N_a2SO4, could have been illustrated. If a divalent cation like Ca 2§ were used, then from the Nernst equation, the same tenfold concentration gradient would develop a potential of only half, namely 30.5 mV (37 ~ or 29.5 mV (22 ~ The reason that this factor is half, rather
than double as one might guess from the fact that the charge is double, is that the Nernst equation gives the PD that exactly opposes the diffusion force due to the concentration gradient, as stated in the paragraph above. Therefore, because the charge is double, only half the voltage is necessary to effectively oppose the concentration force. If the cation in question were trivalent, such as La 3§ then the 2.303 RT/z~ factor would be one-third of 61 mV, or about 20.3 mV. It is for this reason that it is convenient to give the Nernst equation in the form shown in Eq. 8b, namely with the factor (at 37 ~ being --61 mV/z. This allows easy calculation of the equilibrium potential for an ion of any charge and sign (polarity). That is, the sign and charge should be used, for example, + 1 for K + or Na +, +2 for Ca 2+, and -1 for C1-. When the ion in question is an anion like CI-, t h e n - 1 / - 1 gives a plus (+). Because [C1-]o>[C1-] i, this concentration ratio can be inverted by changing the sign of the 2.30 RT/z~ factor back to negative (-). That is, changing the sign of the factor in front of a log ratio simply inverts the ratio. Finally, if the concentration cell depicted in Fig. 6 were made with a membrane that had positively charged pores, then everything would be reversed. The membrane would be permeable to C1- and impermeable to K +, and Ecl would be +59 mV (at 22 ~ and +61 mV (at 37 ~ Again, the voltage is given for side 1 with respect to side 2. Thus, in dealing with an anion, the side of higher concentration becomes positive (due to a small loss of C1-) and the side of lower concentration becomes negative (due to a small gain of C1-). The separated charges again are plastered very close to the membrane: K + ions on side 1 and C1- ions on side 2.
D. Activity Coefficient Thus, an equilibrium potential can be calculated for any species of ion that is distributed unequally across a membrane. All that one needs to know are the concentrations in the two solutions and the charge (and temperature). Actually, we should use the activity (a) of the ion in question in the two solutions, instead of concentration. Thus the Nernst equation given in Eq. 8b becomes (for K +, for example) i
E K=
-61 mV aK +-----]--~log aK
The activity of an ion (in molar) can be obtained by multiplying the concentration of the ion (in molar) by the activity coefficient (7) for the ion: a=c.y (10) In the biological case, the activity coefficients are relatively close (0.7 to 0.9) to 1.0 for Na +, K +, and C1- in both the extracellular and the intracellular solutions. Therefore, in these cases, using the concentrations gives a good approximation. However, in the case of Ca 2+, the activity coefficient in the intracellular solution especially is substantially lower, and so this would affect the calculated value of Eca.
E. Nernst-Planck Equation The basic Nernst equation has been modified in several ways for special situations. For example, in the concentra-
14. Origin of Resting Membrane Potentials
229
tion cell depicted in Fig. 6, a cell in which a single salt (both ions of same valence) is distributed across the membrane at two different concentrations, the PD developed across the membrane ( E m ) can be calculated from the equation
Em
_ U c - U a - 6 1 mV log [salt] 1 Uc+ Ua z [salt] 2
force is the sum of two forces: an electrical force (the negative potential in a cell at rest tends to pull in positively charged ions, because unlike charges attract) and a diffusion force (based on the concentration gradient) (Fig. 7); that is,
(11)
where [salt] 1 and [salt] 2 are the concentrations of the salt on side 1 and side 2, and U c a n d U a are the mobilities of the cations and anions, respectively, through the membrane. Thus, when the mobilities (or permeabilities) of the cation and anion are equal (U c - Ua), E m is zero, regardless of the equilibrium potentials. When the anion is impermeable ( U a - 0), the mobility fraction in Eq. 11 becomes 1.0, and the equation reduces to the simple Nernst equation. When the cation is impermeable (U C = 0), the fraction becomes -1.0, and the same numerical value of E m is produced, but of the opposite sign. Equation 11 can be used to calculate E m for any combination of U a and U c. For example, if U a = 0.5 U c, then for the problem illustrated in Fig. 6, E m is a b o u t - 2 0 mV (side 1 negative). Thus, the membrane potential is related to the relative mobilities.
driving force =
Thus, in a resting cell, the driving force for Na § is (Em-ENa)=-8OmW - (+60mV)---140mV
G. Half-Cell Potentials In the measurement of biological potentials, care must be taken not to introduce artifacts, such as reversible electrode half-cell potentials. See the Appendix (part B) to this chapter for a brief discussion of this topic.
V. E l e c t r o c h e m i c a l Driving F o r c e s a n d M e m b r a n e Ionic Currents A. Electrochemical Driving Forces The electrochemical driving force for each species of ion is the algebraic difference between its equilibrium potential, Ei, and the membrane potential, E m. The total driving
(12a)
The negative sign means that the driving force is directed to bring about net movement of Na § inward. The driving force for Ca 2+ is very large and is directed inward: (E m-
Eca)=-80 m V -
(+129 mV) = - 2 0 9 mV (12b)
The driving force for K + is m V - ( - 9 4 mV) : + 1 4 mV
(E m- E K):-80
(12c)
Hence, the driving force for K + is small and directed outward. The driving force for C1- is nearly zero for a cell at
,3o_
E Energy Wells Ions do not just "fall" through a water-filled pore in the membrane (protein ion channel) down an electrochemical gradient. Instead, an ion may bind to several charged sites on its journey through the channel pore. The K + ion depicted in the bottom of Fig. 6, for example, is shown as binding to three negatively charged sites within the pore. These may be considered energy wells, and the ion must gain kinetic energy to become dislodged from this energy well to pass over the next energy barrier and into the next energy well. This energy comes from the ion being hit by another ion just entering the pore, producing a billiard ball effect. Some evidence for this model was presented in Chapter 13, which discussed the Ussing flux ratio equation. From the measured value of the conductance of single ion channels (e.g., 20 pS, range of 10-300 pS), how many ions that pass through a single channel per second can be estimated. This number is about 6 000 000 ions/s. Therefore, the average transit time for a single ion to cross the membrane (50-70 A thick) is about 0.17 laS.
(12)
Em- Ei
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Eca
,=
60---
DIFFUSION FORCE
E~ (my)
0
"=
"
"'
ELECTRICAL FORCE
RP-8o
- -
-
2
m,
m
. . . . . . . .
-94 . . . . . . . . . . . . . . . . . .
ECl
EK
FIGURE 7. Representationof the electrochemical driving forces for Na+, Ca2+, K+, and C1-. Equilibrium potentials for each ion (e.g., ENa)are positioned vertically according to their magnitude and sign: they were calculated from the Nernst equation for a given set of extracellular and intracellular ion concentrations. Measured resting potential (RP) is assumed to be -80 mV. Electrochemical driving force for an ion is the difference between its equilibrium potential (Ei) and the membrane potential (Era), that is (Em- Ei). Thus, at rest, the driving force for Na§ is the difference between ENa and the resting Em; if ENa is +60 mV and resting Emis -80 mV, the driving force is 140 mV; that is, the driving force is the algebraic sum of the diffusion force and the electrical force, and is represented by the length of the arrows in the diagram. Driving force for Ca2§ (about 210 mV) is even greater than that for Na+, whereas that for K§ is much less (about 14 mV). Direction of the arrows indicates the direction of the net electrochemical driving force, namely, the direction for K+ is outward, whereas that for Na§ and Ca2§ is inward. If C1- is passively distributed, then its distribution across the cell membrane can only be determined by the net membrane potential; for a cell sitting a long time at rest, Ecj = E m and there is no net driving force.
230
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
rest in which C1- is passively distributed (e.g., neuron, myocardial cell, skeletal muscle fiber); that is, (Em
-
Ec1 )= --80 m V -
( - 8 0 mV) = 0
(13)
However, during the AP, when E m is changing, the driving force for Ci- becomes large, and there is a net driving force for inward C1- movement (C1- influx is an outward C1- current). Similarly, the driving force for K + outward movement increases during the AP, whereas those for Na + and Ca 2+ decrease.
B. Membrane Ionic Currents The net current for each ionic species (li) is equal to its driving force times its conductance (gi' reciprocal of the resistance) through the membrane. This is essentially Ohm's law, V l-~-g.V
(14)
modified to reflect the fact that, in an electrolytic system, the total force tending to drive net movement of a charged particle must take into account both the electrical force and the concentration (or chemical) force. Thus, for the four ions, the net current can be expressed as INa = gNa (Era - ENa )
(15)
Ica-gca(Em-Eca)
(16)
IK = gK ( Em - EK )
(17)
I Cl =gcl ( Em - E Cl )
(18)
In a resting cell, C1- and Ca 2+ can be neglected, and the Na + current (inward) must be equal and opposite to the K + current (outward) to maintain a steady resting potential: IK = --/Na
(19)
gK ( E m - EK ) = --gNa (Era- ENa )
(19a)
Thus, although in the resting membrane the driving force for Na + is much greater than that for K +, gK is much larger than gNa' SO the currents are equal. Hence, there is a continuous leakage of Na + inward and K + outward, even in a resting cell, and the system would run down if active pumping were blocked. Because the ratio of the Na + to K + driving forces (-140 mV/-14 mV) is 10, the ratio of conductances (gNa/gK) will be about 1:10. The fact that gK is much greater than gNa accounts for the resting potential being close to E K and far from EN.
VI. D e t e r m i n a t i o n of Resting Potential a n d N e t Diffusion Potential (Ed|ff) A. Determining Factors For given ion distributions, which normally remain nearly constant under usual steady-state conditions, the resting potential is determined by the relative membrane conductances (g) or permeabilities (P) for Na + and K + ions. That is, the resting potential (of about -80 mV in cardiac muscle or skeletal muscle) is close to E K (about-94 mV) because gK >> gNa or PK >> PNa" There is a direct proportionality between P and g at constant E m and concentrations. From sim-
pie circuit analysis (using Ohm's law and Kirchhoff's laws), one can prove that the membrane potential will always be closer to the battery (equilibrium potential) having the lowest resistance (highest conductance) in series with it (see Figs. 5 and 7). In the resting membrane, this battery is E K, whereas in the excited membrane it will be ENa (or Eca), because there is a large increase in gNa and/or gca during the AP. Any ion that is passively distributed cannot determine the resting potential; instead, the resting potential determines the distribution of that ion. Therefore, C1- is not considered for myocardial cells, skeletal muscle fibers, and neurons because it seems to be passively distributed. However, transient net movements of C1- across the membrane do influence Em; for example, washout of C1- (in C1--free solution) produces a transient depolarization, and reintroduction of C1-produces a transient hyperpolarization. C1-movement is also involved in the production of inhibitory postsynaptic potentials (IPSPs) (see Chapter 41). Because of its relatively low concentration, coupled with its relatively low resting conductance, the Ca 2§ distribution has only a relatively small effect on the resting E m and can be ignored.
B. Constant-Field Equation A simplified, but most useful, version of the GoldmanHodgkin-Katz constant-field equation can be given (for 37 ~ g m -
-61 mV log
[ K + ] i + - ~ K [Na+] i PNa
(20)
[K+]o + - ~ K [Na+] o This equation shows that for a given ion distribution, the resting E m is determined by the PNa/PK ratio, the relative permeability of the membrane to Na § and K § For myocardial cells and skeletal muscle fibers, the PNa/PK ratio is about 0.04, whereas for nodal cells of the heart and smooth muscle cells, this ratio is closer to 0.10 or 0.20. Inspection of the constant-field equation shows that the numerator of the log term will be dominated by the [K+]i term [since the (PNa/PK) [Na+]/term will be very small], whereas the denominator will be affected by both the [K+]o and (PNa/PK) [Na§ terms. This relationship thus accounts for the deviation of the E m versus log [K+]o curve from a straight line (having a slope of 61 mV/decade) in normal Ringer solution (Fig. 8). When [K+]o is elevated ([Na+]o being reduced by an equimolar amount), the denominator becomes more and more dominated by the [K+]o term, and less and less by the (PNa/PK) [Na+]o term. Therefore, in bathing solution containing high K § the constant-field equation approaches the simple Nernst equation for K § and E m approaches E K. As [K+]~ is raised stepwise, E K becomes correspondingly reduced, because [K+]i stays relatively constant; therefore, the membrane becomes more and more depolarized (see Fig. 8). A more detailed discussion of the constant-field equation and its other variants is given in Section III of the Appendix to this chapter. When [K+]o is elevated (e.g., to 8 mM) in some types of cells, a hyperpolarization of up to about 10 mV may be
231
14. Origin of Resting Membrane Potentials
[K] o ( l o g s c a l e ) [mV] 1
4
5
10
20
[mM] 5O
100
O,
-20
\
I I I I
-10
]so 200
PNa/PK , I
-30 0,2
i
-40 -50
0.1
-60 .05 -70 -80 -90
.01
-100
SLOPE = 60 m Y / d e c a d e t
-110
I I I
-120
, oir
15 m M [Na]i 150 m M [K]i
-'EK
I I I
FIGURE 8. Theoreticalcurves calculated from the Goldman constant-field equation for resting potential (Em) as a function of [K+]o. Family of curves is given for various PNa/PKratios (0.001, 0.01, 0.05, 0.1, and 0.2). K§ equilibrium potential (EI0 calculated from the Nernst equation (broken straight line). Curves calculated for a [K+]i of 150 mM and a [Na+]i of 15 mM. Calculations made holding [K+]o+ [Na+]oconstant at 154 mM; that is, as [K+]owas elevated, [Na+]owas lowered by an equimolar amount. Change in PI( as a function of [K+]owas not taken into account for these calculations. Point at which E m is zero gives [K+]i. The potential reverses in sign when [K+]oexceeds [K+]i.
produced. Such behavior is often observed in cells with a high PNa/PI(ratio (due to low PI~) and therefore a low resting E m, such as in young embryonic hearts. This hyperpolarization could be explained by several factors: (a) stimulation of the electrogenic Na + pump current (I.), (b) an increase in PK (and therefore gK) due to [K+]o ~ffect on P~:, and (c) an increase in gK (but not PI0 due to the concentration effect. A similar explanation may apply to the fall-over in the E m v e r s u s log [K+]o curve when [K o] is lowered to 1 mM and less, hence depolarizing the cells. This effect is prominent in rat skeletal muscle, for example (see Fig. 7 of Chapter 51).
gK EK W-~g gNa ENa+-~g gca Eca Em- --~g
C. Chord Conductance Equation An alternative method of approximating the membrane resting potential (Em) is the chord conductance equation. The word chord means a straight line connecting two points on a curve, and here specifically refers to the average slope of a nonlinear steady-state voltage-current curve, that is, a straight line from any point on the curve through the origin (zero applied current). (In contrast, slope conductance is the tangent at any point on the curve.) Thus, gK gNa gc1 gca Em- --~g EK + -~g ENa+ --~g Ecl + -~g Eca
where gK, gNa' gcl' ~ d gca are the membrane conductances for K § Na § CI-, and c a z§ respectively, and Eg is the total conductance (sum of all ionic partial conductances). The ratio of gK/Eg, for example, is the relative or fractional conductance for K +. The chord conductance equation can conveniently take into account all ions, including divalent cations, that are distributed unequally across the membrane. The ions important to membrane potentials (including action potentials, postsynaptic potentials, and receptor potentials) are K +, Na +, CI-, and Ca 2+. As discussed previously, C1-cannot help in determining the resting potential if it is passively distributed. Thus, Eq. 21 can be rewritten, omitting the C1- term, as
(21)
(21a)
For simplicity, we can ignore the Ca 2+ term also, giving E m --
gi~ EK + ~ gNa ENa gK + gNa gI( + gNa
(21 b)
where Zg is now equal to gK + gNa" The chord conductance equation can be derived simply from Ohm's law and from circuit analysis for the condition when net current is zero (INa + IK = 0). (See part IV of the Appendix to this chapter for the derivation.) The equation holds true whenever the net current across the membrane is zero, as for the resting potential.
232
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
The chord conductance equation is useful for giving the membrane potential when the ion conductances and distributions are known. For example, at the neuromuscular junction, the neurotransmitter acetylcholine opens the gates of many ionic channels that allow both Na + and K + to pass through equally well (that is, gNa=gK). Hence, the potential that the postsynaptic membrane tends to seek when maximally activated (i.e., the equilibrium potential or so-called reversal potential for the end-plate potential, EPP) is
1
1
EEp P - "~" ( - 9 4 m V ) + -~-(+60 mV)
(21c) = - 1 7 mV A disadvantage of the chord conductance equation is that it gives nearly a straight line for the E m versus log [K+]o plot (actually a slight bend in the opposite direction at low [K+]o). In contrast, the constant-field equation gives the complete bending of the curves (for different PNa/PK ratios) (see Section VI.B). The chord conductance equation again illustrates the important fact that the gK/gNa ratio determines the resting potential. When grr ~ gNa' then E m is close to EK; conversely, when gNa >>gK (as during the spike part of the AP), E m shifts to close to ENa or to Eca (in the case of many types of smooth muscle cells). The chord conductance equation can be rewritten, using resistances instead of conductances, and may then be called the chord resistance equation, E m-
RK RNa E K RNa ENa+ RK + RNa
(21d)
RK +
where R K and RNa are the K + and Na + resistances, which are the reciprocals of the conductances (R K = 1/g K and RNa = l/gNa ). Note that in this equation the positions of the two batteries are interchanged. This equation can be derived by simply substituting the two reciprocals given above into the chord conductance equation. It can also be derived by circuit analysis, as discussed in Section V of the Appendix to this chapter. This Appendix section also shows how simple circuit analysis can be used to determine what the resting potential should be, without using either the Goldman constant-field equation or the chord conductance equation.
D. Net Diffusion Potential, Ediff In the presence of ouabain (short-term exposure only) to inhibit the Na+-K + pump and V , the resting potential that remains reflects the net diffusion potential, Ediff. Ediff is determined by the ion concentration gradients for K § and Na § and by the relative permeability for K § and Na +. When the Na+-K + pump is operating, there is normally a small additional contribution of V to the resting E m of about 2-16 mV, depending on cell type (discussed in following section). Inhibition of the Na+-K + pump for long periods will gradually run down the ion concentration gradients. The cells lose K § and gain Na § and therefore E K and ENa become smaller. The cells thus become depolarized (even if the relative permeabilities are unaffected), which causes them to
gain C1- (because [C1]i was held low by the large resting potential) and to therefore also gain water (cells swell).
VII. Electrogenic Sodium Pump Potentials A brief summary of the previous principles is as follows: The Na+-K + pump is responsible for maintaining the cation concentration gradients. The equilibrium potentials for K § (EK) and Na + (ENa) are a b o u t - 9 4 mV and +60 mV, respectively. The resting potential value is usually near E K, because the K + permeability (PK) is much greater than PNa in a resting membrane. The exact resting membrane potential (Era) depends on the PNa/PK ratio, myocardial cells and skeletal muscle fibers having PNa/PK ratios of 0.01-0.05, whereas smooth muscle or nodal cells of the heart have a ratio closer to 0.10-0.15. In the various types of cells, the resting E m has a smaller magnitude (i.e., is less negative) than E K by 10-40 mV. If there were no electrogenic pump potential contribution to the resting potential (that is, as though the Na+-K + pump was only indirectly responsible for the resting potential by its role in producing the ionic gradients), E m would equal Edift. However, a direct contribution of the pump to the resting E m can be demonstrated. For example, if the Na+-K § pump is blocked by the addition of ouabain, there usually is an immediate depolarization of 2-16 mV, depending on the type of cell. Thus the direct contribution of the electrogenic Na § K § pump to the measured resting E m is small under physiologic conditions (but very important). However, under conditions in which the pump is stimulated to pump at a high rate (e.g., when [Na+]i or [K+]o is abnormally high) the direct electrogenic contribution of the pump to the resting potential can be much greater, and E m can actually exceed E K by as much as 20 mV or more. For example, if the ionic concentration gradients are allowed to run down (e.g., by storing the tissues in zero [K+]o and at low temperatures for several hours), then after the tissues are allowed to restart pumping, the measured E m can exceed the calculated E K (e.g., by 10-20 mV) for a time (Fig. 9). The Na + loading of the cells is facilitated by placing them in cold low or zero [K+]o solutions, because external K + is necessary for the Na+-K+-linked pump to operate; K m of the Na+,K § ATPase for K + is about 2 mM. After several hours in such a solution, the internal concentrations of Na +, K § and C1- approach the concentrations in the bathing Ringer solution, and the resting potential is very low ( Rb § > Cs § > Na § > Li § For example, the relative permeabilities (assigning PK = 1) in squid giant axon (Baker et al., 1968) are PK > eRb > Pcs > eNa > eLi
1.0
0.69
0.19
0.17
0.12
(A-8)
larization produced by taking into account the Ca 2+ ion is only +0.4 mV for a PNa/PK ratio of 0.1. (2) Even when Pca is set equal to 10 times PNa' the effect of Ca 2§ on the resting potential is still relatively small (e.g., +3.5 mV for a PNa/PKratio of 0.01, and +8.0 mV for a PNa/PKratio of 0.1). (3) The effect of Ca 2§ is somewhat greater when the PNa/PKratio is higher (as in cardiac nodal cells or smooth muscle). (4) The effect of taking into account Ca 2§ is considerably less at high [K+]o values. Thus, these calculations support the view that Ca 2§ can be virtually ignored in discussion of the ionic basis of the resting potential. This agrees with the well-known fact that a variation in [Ca 2+] throughout a relatively wide range has a negligible effect on the resting potential (see, for example, Sperelakis, 1972). Further, these conclusions have implications about the relative importance of Na § versus Ca 2§ background currents (inward) during genesis of the pacemaker potential (concomitant with the decrease in gK and IK)" Finally, it should be emphasized that, for example, when Pca = PK' gfa does not equal gK, because of the concentration differences. To calculate gK from a given PK, one must use the appropriate equation that takes into account the concentrations and membrane potential. If gca = gK' then the membrane potential is halfway between E K and Eca, and Ca 2+ would have a much greater effect on the resting potential.
In frog sartorius (Mullins, 1961) the values are
PK > eRb > ecs > eNa 1.0
0.54
0.11
0.04
(A-9)
The PNa/eca ratio in frog sartorius is about 3 (Mullins, 1961). So far in our discussion, Ca 2+ has been ignored. It can be demonstrated that Ca 2§ has only a negligible effect on the resting potential, even if it has a permeability equal to, or 10 times greater than, Na § This is because of the relatively low extracellular and intracellular concentration of free Ca 2§ ion compared with those of the K § and Na § ions. A modified version of the Goldman constant-field equation, which includes a Ca 2+ term, is 3
E m--- - 6 0 mV log
(B - A) + v/y 2 ( A - 4 P c a [Ca2+] i )
IV. Derivation of Chord Conductance Equation Ohm's law states that the current (/) is equal to the voltage (E) either divided by the resistance (R) or multiplied by the conductance (g = l/R) E I- -R - gE (A-11) When dealing with solutions, the voltage or driving force must take into account both the concentration force and the electrical force. In this case, the ionic current (L) is a product of the conductance for a given ion times the total driving force on that ion (Em - E i) and may be expressed as
Ii- gi(Em-Ei)
(A- 10)
where
A - PK[K+]i + PNa[Na+]i B - PK [K+]o + PNa[Na+]o
(A-12)
In a resting cell membrane (stable resting potential), the total ionic current must be zero; otherwise, the membrane potential would change. Therefore, the K + current in the outward direction (IK) must be equal and opposite to the Na + current (INa) entering the cell (neglecting Ca 2+, Cl-, and minor ions) expressed as
y = (B - A ) 2 + 4 ( A + 4Pca ['-'t_;a2+'li) (B + 4Pca [Ca2+]o)
IK = --INa
(A-13)
I K + INa = 0
(A-14)
Therefore, For simplification, the analogous C1- terms (+Pcl[Ca2+] o in the definition of A and +PcI[C1-]i term in B) have been omitted, assuming C1- to be passively distributed. Calculations made from Eq. A-10 demonstrate some interesting points: (1) For the same permeabilities (Pc~ = PNa)' Ca 2+ has much less effect on E than does Na +, because of the lower Ca 2§ concentrations and because of the square root function for the Ca 2§ concentrations. For example, the depo-
Substituting the equations for ionic currents from Eq. A-12,
gK (Era- EK ) + gNa (Era- ENa) = 0 Algebraic manipulations give
0 = gK Em- gK EK + gNa Em- gNa ENa
gK Em § gNa Em = gK EK - gNa ENa 3This equation was kindly provided by Professor D. E. Goldman.
(A-14)
Em (gK + gNa ) = gK EK + gNa ENa
(A- 15)
240
S E C T I O N I1 M e m b r a n e
Rearrangement gives Em-
gI,: EK + ~ gNa ENa (A- 16) gK + gNa gK + gNa Equation A-16 is the chord conductance equation. The ratios gK/(gK + gNa) and gNa/(gK + gNa) are the fractional conductances (relative) and are dimensionless. If C1- were to be included, the same steps in the derivation would give the chord conductance equation containing a C1term, _
Em
~
grr
gNa
gcl
(A-17)
--~gEK+'~gENa+--~gECI
Zg = gK + gNa +gcl" However, if C1- is passively distributed (in equilibrium at the resting Era), then CI- cannot be involved in determining the resting potential (although transient movements of C1- can affect the membrane potential when CI- is shifted off equilibrium during an AP or postsynaptic potential). On the other hand, the Ca 2+ ion is actively transported and is off equilibrium, so its conductance influences the resting potential. Thus, the chord conductance equation containing the Ca 2+ term is gK gNa gca E m _ ~ggEK'4r'-~gENa"l--~gEca
Potential, Transport Physiology, Pumps, and Exchangers
lower terminal), that is, halfway between both batteries because both should be equally expressed, (2) if R 2 is made infinite (e.g., open circuit in branch 2) and R 1 is finite, then the PD is exactly + 100 V, since the right battery (E 2) cannot be expressed at all, and (3) if R 1 is much less than R 2, then the PD approaches + 100 V because E 1 is dominant. The circuit in Fig. A2 can also be analyzed quantitatively. For example, if R 1 = 10 1~, and R 2 = 990 1~, the exact PD may be reasoned from the following analysis. The current (I) has one magnitude; that is, it is constant throughout this simple closed circuit, but the current flows upward in branch 2 and downward in branch 1. This occurs because the right battery (E2) is larger than E~, and so the net driving force for the net current is in the direction as indicated in the figure. Therefore, the voltage drops produced across R 1 and R 2 are in opposite polarities, as shown in the figure. The voltage drop across R 1 adds to E1 to make a greater PD across branch 1, like two batteries in series ( + - , + - ) . In contrast, the voltage drop across R 2 subtracts from E 2 to make a smaller PD, like two batteries back to back ( - +, + -). Therefore, the following two equations can be written for the PD across branch 1 [(PD) 1] and across branch 2 [(PD)2]:
(PD)I :
(a-18)
E 1- I R 1
(A-20)
(PD)2 : E 2 - IR2
(A-21)
Zg = g~: + gNa + gca The chord conductance equation, of course, can be written using resistances rather than conductances. For Eq. A-16 using only K + and Na + terms, substitution of R = 1/g in the equation and algebraic manipulation gives the equation RK
RNa
EK
9-
+
150V
(A-19)
E m - RK + RNa ENa + RK + RNa
Note that the ENa and E K terms are interchanged from the chord conductance equation. This form of the equation might be termed the chord resistance equation. The chord conductance equation applies only to those situations in which the net ionic current is zero, such as when the membrane is at rest. This equation is derived simply from Ohm's law, and one advantage it has over the constantfield equation is that it can more easily include divalent cations such as Ca 2§
(Z)
/?2 100
+ "
200V E2
,1 p.d.
V. Circuit Analysis Applicable to Cell Membrane Using Ohm's and Kirchhoff's laws and logic, it is possible to see why in the nerve or muscle cell, the K + battery dominates the resting potential, whereas the Na § or Ca 2+ batteries, or both, dominate the peak of the AP. The circuit in Fig. A2 will be used to show that the battery having the lowest resistance in series with it is the battery that is the most expressed across the network. Before analyzing the circuit rigorously, we can consider three conditions and make some qualitative judgments: (1) ff the left resistor (R 1) equals the fight resistor (R2) (regardless of their absolute values), the PD across the network is + 150 V (upper terminal positive with respect to the
(1)
-
(2)
R2-oo
(3)
R2
- 'IOD, R2 -
+ 150V + 100V + 101V
9 9 0 Q,
FIGURE A2. Circuitdiagram of the circuit analysis applicable to the cell membrane, showing why the resting potential of a cell is determined by the relative permeabilities (or conductances). Battery having the lowest resistance in series with it is the battery most expressed across such a network.
14. Origin of Resting Membrane Potentials
241
To solve these equations, we must first calculate the current (/). The net driving force for I is equal to E 2 - El; hence from O h m ' s law, the net current is equal to (E 2 - El) divided by the total resistance (R 1 + R2)" I = E2 - E1 R 2 + R~
(A-22)
I = 200 V - 100 V 990 f~ + 10 El =
100 V ~ =O.1A
1000 f~
Now we can enter this value for I into Eq. A-20 ( p D ) l - E 1- IR 1 : 100 V - ( - 0 . 1 A ) ( I O t ~ ) : 100 V - (-1 V) --- 1 0 0 V + 1V = 101V The negative sign in the current is because the current in branch 1 produces a voltage drop that adds to E 1. Because the two branches are connected by zero resistances, that is, they are effectively the same points, (PD)~ must equal (PD)2: (PD)~- (PD)2
(A-23)
Therefore, we can also calculate the PD by substituting into Eq. A-21. (PD)2- E 2 - IR 2 = 200 V - (+0.1 A)(99012) - 200 V - (99 V) =101V Thus, the two methods check. To summarize, it has been quantitatively demonstrated that the battery with the lowest series resistance is the battery most expressed across this network. This analysis holds true regardless of the absolute values of the resistances or batteries or the polarity of each battery. Other methods of circuit analysis can be used to calculate the PD across such a network, but this method is one of the simplest. The following chord resistance equation, analogous to the chord conductance equation, can also be derived and used to calculate the PD across the network: R1
PD-
R 1+ R 2 _
E2+
R2 E1 R 1+ R------~
(A-24)
1012 99012 200 V + 100 V 101"~ + 990 El 101~ + 9901"~ 10 2 0 0 V + 990 1 0 0 V 1000 1000
=2V+99V =101V This equation again emphasizes the point that it is the relative resistances that determine which battery is most expressed.
Bibliography Baker, E E (1968). Nervous conduction: Some properties of the ionselective channels which appear during the action potential. Br. Med. Bull. 24, 179-182. Carmeliet, E., and Vereecke, J. (1979). Electrogenesis of the action potential and automaticity. In "Handbook of Physiology" (R. M. Berne, and N. Sperelakis, Eds.) pp. 269-334. American Physiological Society, Bethesda, MD. Cole, K. S. (1968). "Membranes, Ions and Impulses: A Chapter of Classical Biophysics." University of California, Berkeley. Daniel, E. E., Kwan, C. Y., Matlib, M. A., Crankshaw, D., and Kidwai, A. (1977). Characterization and Ca2+-accumulation by membrane fractions from myometrium and artery. In "Excitation-Contraction Coupling in Smooth Muscle" (R. Casteels, T. Godfraind, and J. C. Ruegg, Eds.), pp. 181-188. Elsevier-North-Holland, Amsterdam. Dhalla, N. S., Ziegelhoffer, A., and Hazzow, J. A. (1977). Regulatory role of membrane systems in heart function. Can. J. Physiol. Pharmacol. 55, 1211-1234. Gadsby, D. C., and Nakao, M. (1989). Steady-state current-voltage relationship of the Na-K pump in guinea pig ventricular myocytes. J. Gen. Physiol. 94, 511-537. Glitsch, H. G. (1972). Activation of the electrogenic sodium pump in guinea-pig auricles by internal sodium ions. J. Physiol. (London) 220, 565-582. Goldman, D. E. (1943). Potential, impedance, and rectification in membranes. J. Gen. Physiol. 27, 37-60. Henn, F. A., and Sperelakis, N. (1968). Stimulative and protective action of Sr2§ and Ba 2§ on (Na§ K+)-ATPase from cultured heart cells. Biochim. Biophys. Acta 163, 415-417. Hermsmeyer, K., and Sperelakis, N. (1970). Decrease in K § conductance and depolarization of frog cardiac muscle produced by Baa§ Am. J. Physiol. 219, 1108-1114. Hodgkin, A. L., and Katz, B. (1949). The effect on sodium ions in electrical activity of the giant axon of the squid. J. Physiol. (London) 108, 37-77. Irisawa, H. (1978). Comparative physiology of the cardiac pacemaker mechanism. Physiol. Rev. 58, 461--498. Jain, M. K. (1972). "The Bimolecular Lipid Membrane: A System." Van Nostrand, New York. Jones, I. R., Maddock, S. W., and Besch, H. R., Jr. (1980). Unmasking effect of alamethicin on the (Na§ K§ beta-adrenergic receptorcoupled adenylate cyclase, and cAMP-dependent protein kinase activities of cardiac sarcolemmal vesicles. J. Biol. Chem. 255, 9971-9980. McDonald, T. E, and MacLeond, D. P. (1971). Maintenance of resting potential in anoxic guinea pig ventricular muscle: Electrogenic sodium pumping. Science 172, 570-572. Meech, R. W. (1972). Intracellular calcium injection causes increased potassium conductance in Aplysia nerve cells. Comp. Biochem. Physiol. 42A, 493-499. Mullins, L. J. (1961). The macromolecular properties of excitable membranes. Ann. NY Acad. Sci. 94, 390-404. New, W., and Trautwein, W. (1972). Inward membrane currents in mammalian myocardium. Pflugers Arch. 334, 1-23. Noble, D. (1975). "Initiation of the Heartbeat." Oxford University Press (Clarendon), London. Noma, A., and Irisawa, H. (1974). Electrogenic sodium pump in rabbit sinoatrial node cell. Pflugers Arch. 351, 177-182. Pelleg, A. Vogel, S., Belardinelli, L., and Sperelakis, N. (1980). Overdrive suppression of automaticity in cultured chick myocardial cells. Am. J. Physiol. 238, H24-H30. Sperelakis, N. (1972). (Na§ K+)-ATPase activity of embryonic chick heart and skeletal muscles as a function of age. Biochim. Biophys. Acta 266, 230-237. Sperelakis, N. (1979). Origin of the cardiac resting potential. In "Handbook of Physiology, Vol. 1, The Cardiovascular System" (R. M. Beme, and N.
242
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
Sperelakis, Eds.), pp. 187-267. American Physiological Society, Bethesda, MD. Sperelakis, N. (1980). Changes in membrane electrical properties during development of the heart. In "The Slow Inward Current and Cardiac Arrhythmias" (D. P. Zipes, J. C. Bailey, and V. Elharrar, Eds.), pp. 221-262. Martinus Nijhoff, The Hague. Sperelakis, N. (1993). Origin of the resting membrane potential. In "Physiology" (N. Sperelakis and R. Banks, Eds.), pp. 29--48. Little, Brown, Boston. Sperelakis, N. (1993). Basis of the resting potential. In "Physiology and Pathophysiology of the Heart," 3rd ed. Kluwer Academic Publishers, New York. Sperelakis, N. (1995). "Electrogenesis of Biopotentials." Kluwer Academic Publishers, Boston.
Sperelakis, N., and Fabiato, A. (1995). Electrophysiology and excitationcontraction coupling in skeletal muscle. In "The Thorax: Vital Pump" (C. Roussos, Ed.), second edition, pp. 33-95. Marcel Dekker. New York. Sperelakis, N., and Lehmkuhl, D. (1966). Ionic interconversion of pacemaker and nonpacemaker cultured chick heart cells. J. Gen Physiol. 49, 867-895. Sperelakis, N., Schneider, M., and Harris, E. J. (1967). Decreased K § conductance produced by Ba 2§ in frog sartorius fibers. J. Gen. Physiol. 50, 1565-1583. Trautwein, W., and Kassebaum, D. G. (1961). On the mechanism of spontaneous impulse generation in the pacemaker of the heart. J. Gen. Physiol. 45, 317-330. Vassalle, M. (1970). Electrogenic suppression of automaticity in sheep and dog Purkinje fibers, Circ. Res. 27, 361-377.
Nicholas Sperelakis
Gibbs-Donnan Equilibrium Potentials
i. Introduction Because intracellular cytoplasm contains many colloids, including large nondiffusible polyvalent electrolytes, a Donnan equilibrium can be established across the cell membrane with an accompanying transmembrane GibbsDonnan (G-D) potential. The resting potential of most cells in the body, including nerve and muscle cells, however, is not due to a Donnan equilibrium, and the normal resting potential is not a Gibbs-Donnan potential, as is erroneously stated in some textbooks. In the true Donnan equilibrium,
all diffusible ions are in equilibrium across the membrane. But many ionsmlike Na § K § Ca 2§ and H+---in nerve and muscle cells are not in equilibrium; that is, ENa :r E m E K =1 t: E m
Eca 4= E m and EH =/= E m
On the other hand, C1- is at equilibrium (i.e., passively distributed) in many vertebrate cells; namely, Ec1 = E m In addition, a large internal pressure and concomitant swelling of animal cells would occur if a Donnan equilibrium were allowed to become established. The action of two types of cation pumps keeps the Donnan osmotic pressure from developing and keeps certain cations out of equilibrium. Thus, a second important function of the Na+-K + pump is the regulation of cell volume. The Na+-K + pump actively pumps 3 Na + ions out (to 2 K + ions pumped in) with each cycle. The pump action decreases the osmotic pressure of the cytoplasm and prevents cell swelling. Inhibition of active ion transport by any means leads to osmotic swelling because of the establishment of the Donnan equilibrium.
Cell Physiology Sourcebook: A Molecular Approach, Third Edition
243
Under such conditions, the cells gain Na +, CI-, Ca 2+, and H20, and they lose K +. Thus, the Gibbs-Donnan potential is passive; that is, energy is not necessary to its establishment. In contrast, the resting potential is actively generated (indirectly or directly) by the action of the Na+-K + pump. The Gibbs-Donnan potential is usually less t h a n - 2 0 mV, whereas the resting potential i s - 4 0 t o - 1 0 0 mV, depending on the cell type (and its ratio of PNa to PK)"
11. M e c h a n i s m for D e v e l o p m e n t of the G i b b s - D o n n a n Potential In the Gibbs-Donnan equilibrium, a small membrane potential is established even though the biological membrane involved, or the artificial membrane used in a laboratory experiment, may be equally permeable to the small diffusible ions used. For the example illustrated in Fig. 1, where aqueous solutions of 0.1 M (Na+)n-proteinate n- (side 1) and 0.1 M NaC1 (side 2) are initially placed on the two sides of a two-compartment chamber separated by a membrane, and if gNa = gc1 in this membrane, then at equilibrium, ENa = Ec1 = - 1 8 mV. The side containing the protein anion becomes negative with respect to the other side. Thus, because both diffusion potentials have the same polarity (as well as magnitude), a potential difference (PD) occurs across the membrane, even though conductances for Na § and C1- across the membrane may be equal. The osmotic pressure of the solution on side 1 containing the nonpermeant protein is greater than that on side 2. In the G-D equilibrium, all permeant ions are in electrochemical equilibrium across the membrane, that is, they are passively distributed, and there is no net electrochemical driving force: (Em-ENa)= (Em -
O
ECl )= 0
Copyright 9 2001 by Academic Press, All rights of reproduction in any form reserved.
244
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
FIGURE I. Gibbs-Donnan potential. (A) Gibbs-Donnan experiment. Diagram depicts the experimental arrangement for obtaining Gibbs-Donnan potential. A membrane freely permeable to all small ions, but impermeable to the large protein molecules, is used to separate two solutions, only one of which (side 1) contains protein. Side containing the protein becomes negative, with respect to the other side, by a small voltage (-18 mV in example). This membrane potential (Era) does not depend on active ion transport or on selective permeability properties of the membrane, as normal cell resting potential does. The diffusible ions (Na § and CI- in example), however, become unequally distributed across the membrane, and it is their diffusion potentials (ENa = EcI) that produce the Gibbs-Donnan potential. (B) Equivalent circuit for Gibbs-Donnan experiment (depicted in A) demonstrating that E m = ENa = Ecp the Na+ and CI- batteries being of equal magnitude and of the same sign. Therefore, the relative conductances of the membrane to Na + and CI-, whether equal or not, are irrelevant to the potential (gcl and gNa are conductances for C1- and Na+, respectively).
A more complete explanation for the development of the Gibbs-Donnan potential follows. The Gibbs-Donnan potential (which is an equilibrium PD) does not depend on metabolic energy. Therefore, this discussion applies to a cell that either has no ATP for pumping ions against electrochemical gradients or has had its Na+-K + pump completely blocked by either ouabain or another agent. The Gibbs-Donnan potential is passively produced by the concentration gradients for diffusible electrolytes (e.g., Na + and C1-) across a membrane. These ion gradients are caused by the presence of one or more large nondiffusible (with respect to the membrane) polyvalent electrolytes (e.g., nega-
tively charged proteins) on one side of the membrane, as is present in all biological cells. In essence, the negatively charged protein molecules (at pH 7) inside the cell attract cations (e.g., Na + or K +) and repel anions (e.g., C1-). Therefore, in the Gibbs-Donnan situation, the inside of the cell has a higher concentration of Na + (or K § and a lower concentration of C1- than has the solution bathing the cell. The equilibrium potentials for Na + (ENa) and for C1- (Ecj) are equal in magnitude and are of the same sign, thereby producing a PD across the membrane. The PD is negative on the inside (side containing the protein) and usually is a b o u t - 2 0 mV or less.
15. Gibbs-Donnan Equilibrium Potentials !Ii. G i b b s - D o n n a n
245
Equilibrium
[Na+]l _ [ C l - ] 2
To quantitate the ion distributions produced at equilibrium and the PD developed, let us examine the artificial system shown in Fig. 1A. In this system, a chamber is separated into two compartments by a collodion membrane, which has small uncharged pores that allow Na + and C1- ions, but not large protein molecules, to diffuse through. A 100 mM solution of Na + proteinate is added to one side (compartment 1), and a 100 mM solution of NaC1 to the other side (compartment 2). An electrode is positioned on each side so that the PD across the membrane can be recorded (37 ~ Let us assume that the Na + proteinate is completely ionized and, for simplicity, that the protein has a net negative charge of only one. Thus there is, at the first instant, no diffusion force for Na +, but there is diffusion force for CI-, because C1- is 100 mM in compartment 2, and 0 mM in compartment 1. Na + must accompany the diffusion of C1- from side 2 to side 1, because the principle of electroneutrality in the bulk solution cannot be violated (i.e, there must be an equal number of cations and anions). So one relation that must be true when the system comes to equilibrium is that [Na+]2 - [C1-]2
(2)
[Na+]2 [Cl-]~
Equation 3 indicates that, at equilibrium, the product of the diffusible ions on side 1 must be equal to the product of the diffusible ions on side 2. From the Nernst equation, the relationships -R__.~Tin ~[Na+]l [Na+]2
(6)
- 6 1 m V log [Na+]~ +1 [Na+]2
(7)
ENa- 2~
and -61 mV log
[Cl-ll
can be given because C1- is negative (z = - 1 ) , whereas Na + is positive (z = +1). Equation 8 is the same as (note that a negative sign in front of a log inverts the ratio)
(1) E c ~ - - 6 1 mV log [C1-]2
In actuality, there is a small charge separation directly across the membrane to account for the PD; that is, side 2 of the membrane has a small excess of Na + ions, and side 1 has a small excess of C1- ions. Such a charge separation is very small, but is necessary to develop a PD across the membrane (V = Q/C), and is discussed in the preceding chapter on the resting potential. The principle of electroneutrality also requires that the increase in Na + on side 1 must be exactly equal to the increase in C1- on side 1. Thus, the concentration difference of Na + that is built up at equilibrium must be exactly equal to the final concentration difference for C1-. This is because the large initial gradient for C1- is what drives the Na + to make its gradient. Therefore, it must also be true that [Na+]l _ [ C l - ] 2
(2)
[Na+]2 [Cl-]~ Cross-multiplying gives
(9)
[cl-]l
because [Na+]l/[Na+] 2 = [C1-]2/[C1-] 1, as Eq. 2 indicates, and from Eqs. 7 and 9, it is clear that Eq. 4 holds true; that is, ENa = Ec1.
IV. Q u a n t i t a t i o n
of the Gibbs-Donnan
Potential
For quantitation, let us use x to indicate the amount (in mM) of C1- or Na + that shifted from side 2 to side 1 at equilibrium. Then the amount of Na + on side 2 is 100 m M - x (the original amount minus the amount lost); C1-on side 2 is also 100 m M - x, because [Na+]2 = [C1-] 2. The Na § on side 1 at equilibrium is 100 mM + x (the original amount plus the amount gained), and the C1- on side 1 is simply x. These parameters may be listed as follows: [Na+]2 - 1 0 0 m M - x
[Na+]l [ C l - ] 1 - [Na+12 [Cl-]2
(3)
Another way of considering this is that ENa must equal Ecl, and therefore, using the respective Nernst equations (see Chapter 14), we can write
ENa = EC1 -61 mV log [Na+]l = - 6 m V
-1 -61 m V +1
(4)
[Na+]l - l O O m M + x
[cl-]1The value for x can be obtained by substituting these values into Eq. 3:
[Na+]l [C1-]I- [Na+]2 [C1-]2
log [C1-]1
[c1-] [Cl-] 2 log ~
[C1-12- l O O m M - x
(1 O0 + x ) x - (1 O0 - x)( 1 O0 - x)
100x + x 2= 10 000 - 200x + x 2 (5)
[C1-]1
Dividing both sides b y - 6 1 mV and removing the log gives
3 0 0 x : 10 000 x - 33.3
(3)
246
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
Thus, at equilibrium
The osmotic pressure (II, in atm) of each solution is equal to the osmotic concentration in osmol/L (C) times the osmotic coefficient (i) times the gas constant (R, 0.082 L-atm/mol-K) times the absolute temperature (T, in K)
[CI-]I - 33mM [ N a + ] l - (100 + 3 3 ) - 133 mM
[Na+]2 - (100 - 3 3 ) - 67 mM These values are also given in Fig. 1A. Note that all the equations and conditions are obeyed. The Gibbs-Donnan potential produced then may be calculated by substituting into Eqs. 7 and 8"
ENa=
(14)
II = iCRT
[C1-]2 - (100 - 3 3 ) - 67 mM
-61 mV 133 mM +1 log 67 mM
(10)
= - 1 8 mV
where C is the number of osmoles per liter of solution. In the example depicted in Fig. 1, a hydrostatic pressure of 3.17 atm would need to be applied to side 1 to prevent this compartment from gaining water from side 2 (at 20 ~ and assuming i = 1.0)
=(1.0)(266 mM - 134 mM)(0.082
:(0
and
(15)
II - i ACRT
L.atm ] (273 + 20)K mol.K /
mol'atm
/ (293 K) g
Ecl=
= 3.17 atm
-61 mV 33 mM -1 log 6 7 ~
(11 )
=-18mV Hence, ENa =Ecl (Eq. 4). That is, the two diffusion potentials are equal in magnitude and of the same sign. The PD across the membrane i s - 1 8 mV; side 1 containing the protein is negative. Therefore, relative permeability of the membrane to Na § and C1- is irrelevant. The equivalent circuit for this example at equilibrium is given in Fig. lB.
V. O s m o t i c C o n s i d e r a t i o n s We should note that, at equilibrium, the sum of Na § and C1-on side 1 (166 mM) is greater than that on side 2 (134 mM). In addition, there is 100 mM protein on side 1. Thus, the total osmotic concentration on side 1 is 266 mOsm (milliosmolar), compared to 134 mOsm on side 2. Therefore, there is a large osmotic gradient between the two sides. Water moves from side 2 to side 1 (i.e., water accompanies the net movement of Na § and CI-) until the hydrostatic pressure head buildup is sufficient to oppose further net movement of water. As expected, the biological cell swells when a Gibbs-Donnan equilibrium is allowed to develop following a blockade of active ion transport for long periods. The total osmotic concentration, [osm], on each side, at equilibrium, may be summarized as follows: [osm] 1 -[Na+]x + [C1-]I + [protein]
1
[osm] 2 - [ N a + ] a + [C1-]2 Substitution gives [osm]l - 133 mM + 33 mM + 100mM = 266 mM [osm]2 - 67 mM + 67 mM =134mM
The situation illustrated in Fig. 1 is actually more complex because the net water movement into side 1 acts to dilute the ion concentrations building up there, and therefore a true Gibbs-Donnan equilibrium can become established only if the net water movement is stopped, that is, by allowing an osmotic pressure gradient to develop by making side 1 a closed, or rigid, system. Otherwise, theoretically all of the water and NaC1 eventually would move out of side 2. The example of a G-D equilibrium in Fig. 1 could have been illustrated using another salt, such as KC1, instead of NaC1, or two or more salts. The extra osmotic pressure in side 1 (or inside a cell) produced by the presence of the negatively charged proteins and other impermeant large charged molecules is known as the colloid osmotic pressure (COP). The COP is also important for water movement across the capillary wall, which separates the blood plasma (containing impermeant proteins) and the interstitial fluid (ISF). At the arterial end of the capillary, the intracapillary hydrostatic blood pressure exceeds the COP, so water moves out of the capillary into ISF space; at the venous end, the COP exceeds the capillary hydrostatic pressure, so water moves into the capillary. In the midregion of the capillary, the two pressures are about equal, and there is no net water flow. Thus, there is a circulation of fluid distributed along the length of the capillary, and this idea is generally known as the Starling hypothesis.
(12)
VI. S u m m a r y (13) In summary, a Gibbs-Donnan equilibrium becomes established and a Gibbs-Donnan potential is developed across the cell membrane of cells under conditions in which metabolism and energy production have been inhibited or the Na+-K + pump has been inhibited by digitalis. The G-D equilibrium occurs because of the large impermeant charged macromolecules, such as proteins, inside the cell. The G-D equilibrium does not require energy for its establishment; that is, it is passive. This contrasts with the normal resting
15. Gibbs-Donnan Equilibrium Potentials potential of the cell, which requires active ion transport and use of metabolic energy to establish large ionic electrochemical gradients. The G-D potential is usually less t h a n - 2 0 mV, whereas the resting potential is considerably greater. In the G-D equilibrium, all permeable ions are in equilibrium across the membrane, whereas this is not true for the normal resting potential. The equilibrium potentials for all permeant ions (e.g., E K, Ecl) are of equal magnitude and polarity. The G-D potential is developed even if the cell membrane had equal permeability or conductance for all small ions, whereas the normal resting potential requires different permeabilities for Na + and K +, namely a low PNa/PK ratio. In the G-D equilibrium, the osmolarity of the cell becomes higher than the interstitial fluid bathing the cell, and so the cell tends to gain water and swell (unless prevented from doing so by a rigid cell wall, such as in plant cells). When equilibrium is established, the product of the concentrations of the permeant ions inside the cell is equal to
247 that outside the cell, and the concentrations of the anions and cations outside the cell must be equal (law of electroneutrality). The gains in cations and anions inside the cell also must be equal to each other. From these required conditions, the equation can be solved algebraically to give the final concentrations at equilibrium, and from this, the calculated potential difference across the membrane.
Bibliography Davson, H. (1964). "A Textbook of General Physiology," 3rd ed. Little, Brown, Boston. Sperelakis, N. (1979). Origin of the cardiac resting potential. In "Handbook of Physiology, Vol. 1, The Cardiovascular System" (R. M. Berne and N. Sperelakis, Eds.), pp. 187-267. American Physiological Society, Bethesda, MD. Sperelakis, N. (1995). "Electrogenesis of Biopotentials." Kluwer Publishing Co., New York.
Steven M. Grassl
Mechanisms of Carrier-Mediated Transport: Facilitated Diffusion, Cotransport, and Countertransport
Passive transport is defined as movement of a solute from
1. Introduction The selective and regulated passage of ions and nonelectrolytes across the cell membrane is an essential component of cellular homeostasis. The maintenance of cell pH and volume and the accumulation of nutrients for protein synthesis and cell metabolism are physiological processes that depend on membrane transport for cells to thrive. Cell membranes are composed of phospholipids organized as a bilayer (5 nm) of two closely opposed leaflets separating the intracellular from the extracellular space. The hydrophobic properties of phospholipids make the cell membrane an impermeable barrier excluding the transfer of hydrophilic solutes that are either charged (anions and cations) or uncharged (nonelectrolytes). The selective passage of hydrophilic solutes across the hydrophobic barrier, a physiological property known as membrane permeability, is mediated by the presence of membrane transport proteins that span the phospholipid bilayer. Transport proteins may be functionally subdivided into channels, pumps, and carriers according to differences in the mechanism mediating ion and nonelectrolyte transport. This chapter describes the mechanisms of carrier-mediated transport, which include facilitated diffusion, cotransport, and counter-
transport. II. Electrochemical Potential Transport mechanisms may be distinguished thermodynamically according to their ability to mediate active or passive transport. Active transport is defined as movement of a solute from a region of low electrochemical potential on one side of the cell membrane to a region of higher electrochemical potential on the opposite side.
Cell Physiology Sourcebook: A Molecular Approach, Third Edition
2 49
a region of high electrochemical potential on one side of the cell membrane to a region of lower electrochemical potential on the opposite side. The electrochemical potential of a solute is the partial molar free energy of the solute or the potential to do work when a difference in electrochemical potential exists across the cell membrane. The electrochemical potential of a solute on either side of the cell membrane is a function of the solute activity (or concentration in dilute solution), the solute charge and valence, and the electrical potential. The difference in electrochemical potential, therefore, reflects the magnitude of the difference in transmembrane solute concentration and the difference in transmembrane voltage factored by the charge and valence of the solute. Notably, for solutes without charge, such as nonelectrolytes, solute free energy is neither increased nor decreased by electrical potential, and only the chemical potential of the solute is considered. Thus, the electrochemical or chemical potential difference of a solute across the cell membrane may be considered a driving force acting on solute transport. In the absence of an electrochemical potential difference or a driving force for solute transport, transport mechanisms that are passive mediate equal solute transport in the forward and reverse direction across the membrane resulting in no net transport. For anions and cations, this would occur when the chemical and electrical driving forces acting on solute transport are equal and opposite in direction across the membrane such that the net driving force is zero. For nonelectrolytes, this would occur in the absence of a solute concentration gradient where transmembrane solute concentrations are equal. In both instances where no net transport occurs, the ion and nonelectrolyte are at electrochemical equilibrium with the driving forces acting on the transported solutes. Where an electrochemical potential
Copyright 9 2001 by Academic Press. All rights of reproduction in any form reserved.
250
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
difference for a solute exists across the cell membrane, the direction of net solute transport by a passive transport mechanism will depend on the direction and magnitude of the chemical and/or electrical driving forces acting on the solute. The chemical and electrical potential difference of a charged solute or ion may occur as opposing driving forces of unequal magnitude, with the direction of net solute transport determined by the direction of the larger driving force. In no instance would a passive transport mechanism mediate net transport of a charged solute in a direction across the cell membrane that opposed both the chemical and electrical driving forces acting on the ion. The same limitation holds for net transport of nonelectrolytes in a direction that opposes the chemical driving force acting on the solute. Active transport mechanisms may be distinguished from passive transport mechanisms by the ability to generate and maintain an electrochemical or chemical potential difference for ions and nonelectrolytes across the cell membrane. This requires net transfer of ions or nonelectrolytes across the membrane in a direction that is opposed by the prevailing electrical gradient and/or chemical concentration gradients as driving forces acting on the transported solutes. To perform the work of moving solutes "uphill" against an electrical gradient and/or a chemical concentration gradient, active transport mechanisms require energy. The source of energy driving active transport is the hydrolysis of ATP. The direct or indirect coupling of active transport to ATP hydrolysis distinguishes primary active transport from secondary active transport. Primary active transport mechanisms such as the ion translocating ATPases or pumps are directly coupled to ATP hydrolysis and thermodynamically transduce the energy released upon ATP hydrolysis to the energy stored in the formation of an ion electrochemical potential difference. Secondary active transport mechanisms such as cotransporters and countertransporters are indirectly coupled to ATP hydrolysis and thermodynamically transduce the energy from one solute electrochemical potential difference to the energy stored in the formation of a second solute electrochemical potential difference. The indirect coupling of secondary active transport to ATP hydrolysis arises from the intermediate formation and ATP dependence of the solute electrochemical potential difference that drives secondary active transport.
until the solute is at equilibrium with the electrical and/or chemical driving forces acting on the solute. At equilibrium, facilitated diffusion of a solute occurs equally in both directions resulting in no net transport. The transport mechanism mediating facilitated diffusion may be modeled as shown in Fig. 1. The two main features of the transport mechanism are an association and dissociation of the transported solute with the transport protein, and a change in the conformation of the transport protein which makes the occupied or unoccupied site of solute interaction accessible from either side of the membrane (Fig. 1A). The transport mechanism may be considered in greater detail as a four-step process (Fig. 1B). First, solute S associates with the transport protein C facing side 1 (labeled "outside" in the figure) to form a solute-carrier complex SC. Second, the solute-carrier complex undergoes a conformational change that reorients the solute-carrier complex to face side 2 (labeled "inside" in the figure). Third, the solute dissociates with the transport protein on side 2. Fourth, the unoccupied carrier undergoes a second conformational change to reorient the solute association site to face side 1. Net solute transport from side 1 to side 2 occurs when an unoccupied solute association site is reoriented from side 2 to side 1. However, as the solute concentration on side 2 increases, net solute transport from side 1 to side 2 decreases because a greater proportion of the unoccupied carrier becomes associated with solute on
solute |
cell
~1 ~ OUTSID ~Ep ~
I~@ | |
membrane ~ - -
TTTTTIB ~ ~ IITTTTTB ~ ~ JTTTTT ~ el0= ec;:(~h:nTical =~.,~,o,,/,o~,~,=,o0 '""~
|
facilitateddilfusion
|
STEP 4
1,.-
Cside I ...r
"--- Cside 2
i11. Carrier-Mediated Transport M e c h a n i s m s A. Facilitated Diffusion Facilitated diffusion or uniport is the simplest form of carrier-mediated transport and results in the transfer of large hydrophilic molecules (sugars, amino acids, nucleotides, and organic acids and bases) across the cell membrane. Transport by facilitated diffusion is passive and reversible, with the direction of net transport into or out of the cell determined by the direction of the electrochemical potential difference of the transported solute. Net transport by facilitated diffusion may continue in either direction
STEP 3
STEP 1 Sside I
Cell M e m b r a n e
Sside 2
r
SCside I
~7
FIGURE 3. Consensus structure of the mammalian D-glucose carriers. (Reprinted with permission from Bell, G. I. (1991), Molecular defects in diabetes mellitus, Diabetes 40: 413-422. Copyright 9 1991 American Diabetes Association.)
254
TABLE 1 Designation
SECTION !! Membrane Potential, Transport Physiology, Pumps, and Exchangers D-Glucose Carrier Isoforms
Function
of solute translocation across the membrane. The minor tissuespecific differences in sequence distinguishing isoforms of the same protein presumably reflect a unique variation in the functional properties and/or regulation of the transporter serving the specialized physiology of the tissue.
GLUT1
Basal uptake in placenta, brain, kidney, colon
GLUT2
Primarily Found in transport in liver cells
GLUT3
Basal uptake in the brain
B. Cotransport
GLUT4
Insulin-stimulated uptake in the skeletal muscle, cardiac muscle, and adipose tissue
GLUT5
Absorption in small intestines
Cotransport, or symport, is a form of secondary active transport that mediates net transfer of a solute across the cell membrane from a place of low solute electrochemical potential to a place of higher solute electrochemical potential. The source of energy driving secondary active transport of a solute against its electrochemical potential difference arises from a coupling to the transport of a second solute from a place of higher electrochemical potential to a place of lower electrochemical potential. The coupling of the driving solute to the driven solute results in the cotransport of both solutes in the same direction across the cell membrane. Cotransport mechanisms are reversible and will mediate net transport either into or out of the cell until the opposing electrochemical potential differences of the driven solute and the driving solute become equal across the cell membrane. The direction of net transport mediated by a cotransport mechanism may be determined by either solute, depending on which solute has the larger driving force and by the direction of the larger driving force across the membrane. The mechanism mediating cotransport may be modeled as a six-step process as shown in Fig. 4: (1) An
glycosylation of the transport protein. Likewise, the presence of serine, threonine, and tyrosine in consensus phosphorylation sites in intracellular peptide loops suggests that the phosphorylation state of the transporter may regulate its activity. The structure-function relationship of the protein mediating facilitated diffusion may be investigated upon identification of the primary structure or amino acid sequence of the protein. Mutant transport proteins may be engineered with single or multiple amino acid deletions, or with substitutions in the transmembrane domains, the intra- and extracellular peptide loops between transmembrane domains, and in the amino and carboxy termini. The functional properties of the engineered transport proteins may be assessed by heterologous expression in frog oocytes or cell cultures and compared to the native protein. Thus, the individual amino acids or amino acid sequences important for solute recognition by, and interaction with, the transport protein may be identified as well as those involved in the conformational change reorienting the solute-associated and -dissociated transporter. The mutational analysis of transporter function is limited by null mutations which result in the expression of nonfunctional transport proteins or by the absence of expression. The information obtained from identifying and cloning the gene coding for a facilitated diffusion mechanism may be used to survey the distribution of the same or related transport proteins in different tissues and in different cells within the same tissue. Oligonucleotide probes may be made from short sequences of the transporter gene and used to detect the presence of complementary transporter sequences in cDNA from tissue libraries, in mRNA prepared from multiple tissues, or by in situ hybridization of mRNA in thin tissue slices. The positively identified mRNA or cDNA may be isolated and sequenced for comparison with the deduced amino acid sequences present in different tissues. Such a comparison will indicate either complete sequence identity, suggesting the presence of the same transport protein in different tissues, or will indicate minor tissue-specific differences in sequence, suggesting that different isoforms of the same protein exist in different tissues. Thus, the facilitated diffusion of glucose in different tissues is mediated by at least five different isoforms of the same transport protein as shown in Table 1. The high degree of sequence conservation among different isoforms of the same transport protein is consistent with a common solute specificity and mechanism
STEP 6 .Jl
Cside I
~"
Cside 2
.=
STEP 5
STEP 1 /~
Sside 1"~
Sside 2
Cell Membrane CSside 2
SCside I
L
STEP 4
STEP 2
//~Aside 2
Aside 1"~
' SCAside 2
SCAside I STEP 3 FIGURE 4.
Kineticmodel of cotransport.
16. Mechanisms of Carrier-Mediated Transport: Facilitated Diffusion, Cotransport, and Countertransport
association of solutes S and (2) A with the carrier facing side 1 to form the SCA solute-carrier complex, (3) a conformational change reorienting the SCA carrier complex to face side 2, (4) dissociation of A and (5) S with the carrier facing side 2, and (6) a second conformational change reorienting the solute-free carrier to face side 1. The cotransport mechanism may be further modeled in three different variations depending on the solute interaction with the carrier. Solute interaction with the carrier may be random, with either solute associating and dissociating first with the carrier, or solute association-dissociation may be ordered, in which either one of the solutes must associate or dissociate first with the carrier before the second solute may associate or dissociate. Note, as shown in Fig. 4 for an ordered interaction of solute with the carrier, that the second solute associating with the carrier at one side of the membrane is the first solute to dissociate with the carrier at the opposite side. Neither form of the partially associated carrier SC or AC may undergo a conformational change reorienting the sidedness of the partially associated carrier and therefore the cotransport mechanism mediates only the coupled transport of both solutes but does not mediate the uncoupled transport of either solute. Each step in the cotransport process is reversible and is linked to the preceding and succeeding step by rate constants defining the rate of solute association and dissociation on either side of the membrane and by the rate of conformational change for carrier reorientation in either direction when associated with or without both solutes. Similar to facilitated diffusion, solute association-dissociation with the carrier occurs faster than the conformational change reorienting the sidedness of the carrier and the rate-limiting step in the cotransport process is the conformational change reorienting the solute-dissociated carrier. The unidirectional rate of solute transport across the cell membrane mediated by a cotransporter is dependent on the concentration of both solutes at the side where cotransport originates. When either solute concentration is constant, the cotransport rate of the second solute will approach a maximum value with increasing solute concentration and the saturation kinetics of the cotransport process may be analyzed using the Michaelis-Menten equation to describe the kinetic parameters Km and Vmax for each solute. The random or ordered association of cotransported solutes with the transporter may be determined from the mutual dependence of kinetic parameters on cotransported solute concentrations. As shown in Fig. 4, where the association of solutes S and A with the cotransporter at side 1 is ordered and S associates first, solute A will have the same maximal rate of transport (Vmax) when determined at different concentrations of solute S. However, the K m value for solute A will decrease when determined at increasing concentrations of solute S, indicating the initial association of S with the cotransporter effectively increases its affinity for association with solute A. In contrast, the Vmax and K m values characterizing transport of solute S, the first solute to associate with the cotransporter, will both vary when determined at different solute A concentrations. An increase in solute A concentration will increase the maximal rate of solute S through the cotransport mechanism and will also effectively
255
increase its affinity for solute S as reflected by a decreased value. The effect of solute A on the maximal rate of solute S transport arises from its ordered association, as the second solute, to form the SCA solute-carrier complex, which may then transfer both solutes across the membrane. Thus, where a mutual solute concentration dependence of the kinetic parameters characterizing the transport of both solutes is determined for a cotransport mechanism, an ordered association-dissociation of solutes with the cotransporter may be indicated and a random associationdissociation of solutes may be excluded. The cotransport of a solute may be inhibited by the presence of other solutes that interact with, but are not necessarily transported by, the same cotransport mechanism. Similar to facilitated diffusion, the competitive or noncompetitive nature of inhibitor interaction with the cotransport mechanism may be assessed by determining the effect of the inhibitor on the kinetic parameters characterizing transport of both solutes by the cotransport mechanism. The same cotransport mechanism may mediate transport of multiple solutes that share common chemical and physical determinants. Similar to facilitated diffusion, the solute specificity of a cotransport mechanism may be characterized by determining the kinetic parameters of different solutes transported by the same cotransport mechanism. A comparison of the relative affinities and maximal velocities of solute transport determined experimentally will suggest the relative magnitudes of solute transport mediated by the cotransporter in v i v o . The cotransport mechanism modeled in Fig. 4 will be at a steady state and mediate equal transport of S and A in both directions across the membrane when the driving forces acting on the cotransported solutes are equal and opposite. This would occur when the transmembrane concentration difference or gradient of S in one direction is equal to the transmembrane concentration difference or gradient of A in the opposition direction. At the steady state, where no net solute transport occurs, the solute driving forces will be in thermodynamic equilibrium with each other and the transmembrane concentration ratios of S and A may be described by
Km
S---L= A---~2 (2) $2 A1 Thus, in the presence of a ten-fold concentration gradient of S, a cotransport mechanism coupling the transport of S and A across the membrane may generate up to, but not exceeding, a ten-fold concentration gradient of A. Because the cotransport mechanism effectively transduces the energy stored in the concentration gradient of S to the energy stored in the concentration gradient of A, it would be thermodynamically impossible for the cotransport mechanism to generate more than a ten-fold concentration gradient of A. The cotransport process modeled in Fig. 4 describes the coupled transport of one molecule of S together with one molecule of A, or an S to A coupling ratio of 1" 1. However, the stoichiometric coupling of solutes mediated by a cotransport mechanism may exceed unity, resulting in cotransport of unequal amounts of solute. This may be
256
SECTION 11 Membrane Potential, Transport Physiology, Pumps, and Exchangers
modeled kinetically as an extension of Fig. 4 to include an extra step of solute association on side 1 prior to carrier reorientation to face side 2 and an extra step of solute dissociation on side 2 prior to carrier reorientation to face side 1. The kinetic parameter (Km) characterizing the solute affinity of a cotransport mechanism must account for each additional solute association or dissociation at mutually exclusive sites in the cotransport protein. The thermodynamic equilibrium of solute driving forces defining the steady state, where no net solute transport occurs, is profoundly affected by the stoichiometric coupling of solutes by the cotransport mechanism. Where the stoichiometric coupling is other than 1:1, the solute driving forces will be in thermodynamic equilibrium at transmembrane solute concentration ratios raised to the power of their respective coupling coefficient as shown. S
S1
A1J
(3)
The coupling coefficients s and a are the number of S and A molecules that must associate with the carrier before a conformational change may occur reorienting the sidedness of the solute-carrier complex. The stoichiometric coupling of solutes by a cotransport mechanism is significant because in the presence of a ten-fold concentration gradient of S, a cotransport mechanism coupling the transport of two molecules of S to one molecule of A may generate a onehundred-fold concentration gradient of A! Similar to the facilitated diffusion of charged solutes, cotransport mechanisms may also be electrogenic and mediate net charge transfer across the membrane in the direction of net solute transport. In general, the electrogenicity of cotransport mechanisms arises from the transport of the inorganic cation sodium coupled to the transport of organic and inorganic anions and cations such as negatively charged amino acids, mono- and dicarboxylic acids, sulfate, and phosphate or positively charged amino acids and amines as well as nonelectrolytes such as sugars. In most instances, the coupled transport of charged or uncharged solutes with the transport of sodium results in the net transfer of positive charge in the direction of net solute transport into the cell. The positive electrogenicity of solute transport mediated by a sodium-coupled cotransport mechanism will tend to depolarize the cell membrane potential difference and results from at least one voltagesensitive step in the process accounting for electrogenic cotransport. A voltage-dependent increase or decrease in the kinetic parameters characterizing an electrogenic cotransport mechanism may indicate an effect of voltage on the affinity of the cotransporter for either or both solutes ( K ) and/or on the conformational change reorienting the solute-associated or-dissociated transporter (Vmax). However, because the ratelimiting step in the process accounting for cotransport is the conformational change reorienting the solute-dissociated transporter, a voltage-dependent increase in the rate of cotransport must result from the voltage sensitivity and acceleration of at least this step in the cotransport process. The voltage sensitivity of an electrogenic cotransport mechanism makes the voltage difference across the cell membrane a driving force acting on solute cotransport. Thus, the rate, magni-
tude, and direction of solute transport mediated by an electrogenic sodium cotransport mechanism is determined by two driving forces including the transmembrane voltage difference as well as the transmembrane sodium and solute concentration gradients. An electrogenic cotransport mechanism will be at steady state and mediate equal solute transport in both directions across the membrane when the electrical and chemical forces driving cotransport are at equilibrium across the membrane. For an uncharged solute S, and a charged solute A, the driving forces may be related by
/
- - R T In
A2} ~1 +zY(q'2-q'~)
(4)
where R is the universal gas constant, T is absolute temperature, z is the valence of A, ~ is the Faraday constant, and (We - ~1) is the transmembrane voltage difference. Thus, in the presence of a ten-fold sodium concentration gradient and a transmembrane voltage difference of 60 mV, a cotransport mechanism mediating 1:1 Na-to-solute transport may generate a one-hundred-fold uncharged solute gradient. The functional properties of solute transport mediated by sodium-dependent cotransport mechanisms have been characterized in detail for many different cell- and solutespecific cotransporters. Like facilitated diffusion, a major question remaining in the study of sodium-dependent cotransport mechanisms is the structure-function relationship of the cotransport protein. Only recently have several sodium-dependent cotransport proteins been identified using a molecular biological approach to clone the genes coding for these membrane proteins. The primary structure of sodium-dependent cotransport proteins suggests topological features similar to membrane proteins mediating facilitated diffusion. These include the presence of 10-12 putative transmembrane domains, the location of extracellular Nlinked glycosylation sites, and the location of intracellular phosphorylation sites. The structure-function relationship of the sodium-glucose cotransport protein SGLT1 has been investigated in mutagenesis studies of the cotransport protein. Intestinal glucose-galactose malabsorption is a genetic disorder resulting from multiple mutations localized in two putative transmembrane domains of SGLT1. Individual functional analysis of the different mutant SGLT1 proteins suggested the identity and location of specific amino acids that participate in the association of glucose and the inhibitor phloridzin with the cotransport protein as well as the amino acids involved in the conformational changes reorienting the solute-dissociated cotransporter. In addition to identifying the amino acids affecting the functional properties of the cotransport protein, mutagenesis studies of SGLT1 further suggest the identity and location of amino acids that participate in the membrane trafficking of cotransport proteins from its site of synthesis in the cell interior to the cell surface. Three isoforms of the sodium-glucose cotransporter have been identified (SGLT1, SGLT2, SGLT3) and are distinguished by a high (SGLT1) or low (SGLT2, SGLT3) affinity for glucose, by differences in substrate specificity, and by tissue distribution in small intestine (SGLT1) and/or renal proximal tubule (SGLT1, SGLT2). The structural basis for isoform-specific differences in function may be assessed by
16. Mechanisms of Carrier-Mediated Transport: Facilitated Diffusion, Cotransport, and Countertransport
functional studies of chimeric proteins composed of different peptide sequence combinations from different isoforms. Thus, the chimera composed of the amino terminus including the initial seven transmembrane domains of SGLT3 and the final five transmembrane domains including the carboxy terminus of SGLT1 is observed to have the substrate specificity of the SGLT1 isoform. Accordingly, this finding indicates that the topological location of the glucose association site in the sodium-glucose cotransport protein is in a region of the peptide sequence extending from the eighth transmembrane domain to the carboxy terminus.
C. Countertransport Countertransport, or antiport, is a form of secondary active transport that, like cotransport, may also mediate net transfer of a solute across the cell membrane in a direction against the electrochemical potential gradient of the solute. The energy driving countertransport of a solute against its electrochemical potential difference arises from a coupling to the transport of a second solute moving across the cell membrane from a place of higher electrochemical potential to a place of lower electrochemical potential. In contrast to cotransport, the coupling of the driving solute to the driven solute in countertransport results in transport of solutes in the opposite direction across the cell membrane. Thus, countertransporters mediate the exchange of intracellular and extracellular solutes across the cell membrane. Countertransport mechanisms are reversible and will mediate net transport either into or out of the cell until the opposing electrochemical potential differences of the driven solute and the driving sol-
STEP 5 ,d
= SCside 2
SCside 1-
STEP 4
STEP
~
Sside I
~X'~ Sside 2
Cell Membrane Cside 2
CsiqJe I
Aside I
Aside 2
STEP 1
STEP 3
= ACside 2
ACside I STEP 2 F I G U R E 5.
Kinetic model of countertransport.
2 57
ute become equal across the cell membrane. The direction of net transport mediated by a countertransport mechanism may be determined by either solute, depending on which solute has the larger driving force and the direction of the larger driving force across the membrane. The mechanism mediating countertransport may be modeled as a six-step process as shown in Fig. 5. The principle features of the countertransport mechanism are (1, 4) a mutually exclusive association of either solute S or A with the carrier facing side 1 or side 2 to form the SC or AC carrier complex, (2, 5) a conformational change reorienting the sidedness of the solute-cartier complex, and, (3, 6) a dissociation of solute with carrier on the opposite side of the membrane. In contrast to the process of facilitated diffusion or cotransport, the solute-free carrier does not undergo a conformational change in countertransport and only the SC or AC carrier complex may undergo a conformational change reorienting the sided access of carrier to solute. It is the absence of a conformational change reorienting the sidedness of the solute-free carrier that explicitly distinguishes countertransport from facilitated diffusion and confers the ability to mediate net solute transport against an electrochemical potential difference. Countertransport mechanisms may mediate the exchange of the same solutes (homoexchange) or different solutes (heteroexchange) and will mediate net solute transport only when exchanging different solutes. Each step describing the process of countertransport is reversible and is linked to the preceding and succeeding step by rate constants defining the rate of solute association and dissociation on either side of the membrane and by the rate of conformational change reorienting the sidedness of the solute-associated cartier. The conformational change reorienting the sidedness of the solute-associated carrier occurs more slowly than solute association-dissociation with the carrier and is considered the rate-limiting step in the countertransport process. The unidirectional rate of solute transport across the cell membrane mediated by a countertransport mechanism depends on the concentration of both solutes at opposite sides of the membrane. When either solute concentration is constant, the countertransport rate of the second solute will approach a maximum value with increasing solute concentration and the saturation kinetics of the countertransport process may be analyzed using the Michaelis-Menten equation to obtain the kinetic parameters K m and Vmax for each of the exchanged solutes. The solute concentration dependence of the rate of solute transport by countertransport may be different for transport occurring into or out of the cell. A kinetic asymmetry of solute transport mediated by countertransport results from a difference in the apparent affinity (Km) for solute at either side of the membrane and not from a difference in the maximal rate of transport (Vmax). This limitation in kinetic asymmetry is a unique property of countertransport arising from the inability of the solute-dissociated carrier to reorient across the membrane. While kinetic asymmetry is not a property of all countertransport mechanisms, the asymmetry in solute transport mediated by a countertransporter is observed to occur for all possible combinations of solutes exchanged by the countertransport mechanism. The countertransport of a solute may be inhibited by the presence of other solutes that interact with, but are not necessarily transported by, the same countertransport mechanism.
258
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
The competitive or noncompetitive nature of interaction of the inhibitor with the countertransport mechanism may be assessed by determining the effect of the inhibitor on the kinetic parameters characterizing transport of exchanged solutes by the countertransport mechanism. The same countertransport mechanism may mediate the exchange of multiple pairs of solutes that share common chemical and physical determinants. Countertransport mechanisms may be broadly categorized as either anion or cation exchange mechanisms such as the chloride-bicarbonate countertransporter present in kidney, intestine, and red blood cells or the ubiquitous pHregulating sodium-proton countertransporter. The solute specificity of a countertransport mechanism may be characterized by determining the kinetic parameters of different solutes proven to be countertransported by the same mechanism. A comparison of the relative affinities and maximal velocities of solute transport determined experimentally will suggest the relative magnitudes of solute transport mediated by the countertransporter in vivo. The countertransport mechanism modeled in Fig. 5 will be at a steady state and mediate equal transport of S and A from side 1 to side 2, and from side 2 to side 1, when the driving forces acting on solutes S and A are equal. For countertransport this will occur when the transmembrane concentration differences of S and A are equal and in the same direction across the cell membrane. At the steady state, where no net solute transport occurs, the solute driving forces will be in thermodynamic equilibrium with each other and the transmembrane concentration ratios of S and A resulting from countertransport may be described by $1 - -
$2
At =
~
(5)
A2
Thus, in the presence of a ten-fold concentration gradient of S, a countertransport mechanism coupling exchange of S for A may generate up to, but not exceeding, a ten-fold concentration gradient of A. Because the countertransport mechanism effectively transduces the energy stored in a ten-fold concentration gradient of S to the energy stored in a concentration gradient of A, it would be thermodynamically impossible for the countertransport mechanism to generate more than a ten-fold concentration gradient of A. The important difference distinguishing cotransport and countertransport at the steady state is the sidedness or direction of solute concentration gradients across the membrane. Cotransport is at steady state when solute concentration gradients are of equal magnitude but in opposite direction across the membrane, whereas countertransport is at steady state when solute concentration gradients are of equal magnitude and in the same direction across the membrane. The countertransport process modeled in Fig. 5 describes the coupled exchange of one molecule of S for one molecule of A, or an S-to-A coupling ratio of 1"1. Whereas most countertransport mechanisms mediate the coupled exchange of single solutes, the stoichiometric coupling of solutes may exceed unity, resulting in unequal amounts of solute transported to opposite sides of the membrane. This may be modeled kinetically as an exten-
sion of Fig. 5 to include an extra step of solute association on side 1 prior to carrier reorientation to face side 2 and by an extra step of solute dissociation on side 2. The kinetic parameter (Km) characterizing the solute affinity of a countertransport mechanism must account for each additional solute association-dissociation at mutually exclusive sites in the countertransport protein. Similar to cotransport, the thermodynamic equilibrium of solute driving forces coupled by the countertransport mechanism and defining the steady state is profoundly affected by the stoichiometric coupling of solutes. Where the stoichiometric coupling of solutes by a countertransport mechanism is other than 1:1, the solute driving force will be in thermodynamic equilibrium at transmembrane solute concentration ratios raised to the power of their respective coupling coefficient as shown. $1 i s _ A1 /~) [~2
a
(6)
Thus, in the presence of a ten-fold concentration gradient of S, a countertransport mechanism coupling the exchange of two molecules of S for one molecule of A may generate a one-hundred-fold concentration gradient of A in the same direction across the membrane! Most countertransport mechanisms mediate the coupled exchange of single solutes with the same positive or negative charge and same valence and are therefore electroneutral and do not mediate net charge transfer across the membrane. However, when countertransport mechanisms mediate the coupled exchange of solutes of the same charge but different valence or mediate an unequal stoichiometric coupling of solutes of the same charge, a net charge transfer will occur across the cell membrane. Thus, the countertransport mechanism mediating sodium-calcium exchange across the cell membrane couples the influx of three sodium ions to the efflux of one calcium ion, resulting in a net transfer of positive charge into the cell. The positive electrogenicity of the sodium-calcium countertransport mechanism will tend to depolarize the cell membrane potential and results from the voltage sensitivity of at least one step in the process describing a complete sodium-calcium countertransport cycle. A voltage-dependent increase or decrease in the kinetic parameters characterizing an electrogenic countertransport mechanism would suggest an effect of membrane potential on the solute affinities of the countertransporter (Km) and/or on the conformational change reorienting the sidedness of the solute-associated countertransporter (Vmax). Because the conformational change reorienting the countertransporter is the rate-limiting step in the process accounting for electrogenic countertransport, a voltage-dependent increase in the rate of countertransport must result from an acceleration of at least this step. The voltage sensitivity of an electrogenic countertransport mechanism makes the cell membrane potential difference a driving force acting on solute countertransport. Thus, the rate, magnitude, and direction of electrogenic countertransport of solutes across the membrane is determined by the transmembrane voltage difference in addition to the transmembrane differences in solute concentration for both solutes coupled by the countertransport mechanism. An
16. Mechanisms of Carrier-Mediated Transport: Facilitated Diffusion, Cotransport, and Countertransport
electrogenic countertransport mechanism will be at steady state, and mediate equal solute transport in both directions across the membrane, when the electrical and chemical forces driving countertransport are at equilibrium. For the 1:1 exchange of charged solutes S and A with valences zs and zA, respectively, the driving forces may be related by R T In
- R T In
The electrogenicity of a countertransport mechanism may arise from the exchange of solutes with the same charge and valence, but with an unequal stoichiometric coupling. In this instance the driving forces acting on electrogenic countertransport must be factored by the stoichiometric coupling coefficients s and a for solutes S and A, respectively, and may be related by
Rlnllsts2,RlnlAlla
Thus, in the presence of a ten-fold sodium concentration gradient and a transmembrane voltage difference of 60 mV, an electrogenic countertransport mechanism mediating the 3:1 exchange of sodium and calcium ions, respectively, may generate a maximum calcium concentration gradient of ten-thousand-fold across the cell membrane! The functional properties of solute transport mediated by countertransport mechanisms have been characterized in detail for many different cell- and solute-specific countertransporters. Like facilitated diffusion and cotransport, the structure-function relation of membrane proteins mediating countertransport continues to be an important question under study. The techniques of modem molecular biology have resulted in the recent cloning of several genes coding for countertransport proteins. Analysis of the deduced amino acid sequences of countertransport proteins suggests topological features similar to membrane proteins mediating facilitated diffusion and cotransport, including the presence of 10-12 putative transmembrane domains, the location of extracellular N-linked glycosylation sites, and the location of intracellular phosphorylation sites. Recently, low-resolution electron density maps of the red blood cell chloride-bicarbonate exchange protein have been obtained from two-dimensional crystals of the countertransport protein studied by electron microscopy. The membrane domain of the countertransport protein exists as a dimeric structure with three electron-dense regions in each monomer. Within each monomer the configuration of electron densities suggests two closely apposed peptide structures which are spatially separated from a third structure by a peptide domain existing in two different conformations. The structure-function relationship of the anion exchange protein AE1 has been investigated in mutagenesis studies of the countertransport protein. The functional significance of single amino acids or domains of multiple amino acids on the interaction with transport inhibitors at the extracellular side of AE1 has been studied to determine the nature and topological location of amino acids involved in substrate interaction and translocation. These mutagenesis studies involving both the
259
deletion and substitution of amino acids indicate that several positively charged lysines at the surface of AE1 influence, and are necessary for, substrate recognition by the countertransport protein. The genes coding for three anion exchange proteins (AE1, AE2, AE3) have been identified and their gene products may be functionally distinguished by differences in inhibitor sensitivity and by the intracellular pH dependence of anion exchange activity. Multiple species-specific isoforms of the AE1 gene have been identified and the predominant sites of expression are the erythropoietic tissues and the acidsecreting cells of the renal cortical collecting duct. In contrast, AE2 gene expression occurs more widely in epithelial and nonepithelial tissues and AE3 gene expression is predominant in brain and heart. The putative transmembrane amino acid sequences of AE1, AE2, and AE3 are approximately 70% identical. The aligning portion of the N-terminal cytoplasmic domains of AE1 and AE2 are only 30% identical, with little or no corresponding sequence identity among AE1, AE2, or AE3 at the N-terminus. The observed amino acid sequence identity in the putative transmembrane domains of AE 1, AE2, and AE3 is consistent with the common functional properties of the three anion exchange proteins. The divergent amino acid sequences noted in the N-terminal regions suggest possible cell- and tissue-specific differences in the regulation of anion exchange activity and/or in the intracellular sorting and membrane targeting of anion exchange proteins. Mutations in the human AE1 gene in the form of nucleotide deletions, insertions, and substitutions result in inheritable diseases of the red blood cell which manifest multiple clinical phenotypes ranging from mild to severe. The single amino acid substitution characterizing mutant AE1 in Band 3 Memphis has essentially no effect on red blood cell anion exchange activity. However, the nine-amino-acid deletion characterizing AE1 in Southeast Asian ovalocytosis (SAO) results in a heterozygous genetic disorder of red blood cells with only half the normal complement of functional anion exchangers. Interestingly, the red blood cell membrane of SAO patients is more resistant to plasmodial invasion, a property that may confer a selective survival advantage against malarial infection and thereby perpetuate this genetic disorder.
Bibliography Alper, S. L. (1994). The band 3-related AE anion exchanger gene family. Cell Physiol. Biochem. 4, 265-281. Aronson, E S. (1981). Identifying secondary active solute transport in epithelia. Am. J. Physiol. 240 (Renal Fluid Electrolyte Physiol. 9), F1-F11. Heinz, E. (1978). "Mechanics and Energetics of Biological Transport." Springer-Verlag, New York. Heinz, E. (1981). "Electrical Potentials in Biological Membrane Transport." Springer-Verlag, New York. Neame, K. D., and Richards, T. G. (1972). "Elementary Kinetics of Membrane Carrier Transport." Blackwell Scientific, Oxford. Steel, A., and Hediger, M. (1998). The molecular physiology of sodium- and proton-coupled solute transporters. News Physiol. Sci. 13, 123-131. Stein, W. D. (1990). "Channels, Carriers and Pumps."Academic Press, New York.
David M. Balshaw, Lauren A. Millette, and Earl T. Wallick
Sodium Pump Function
!. Introduction As primordial life forms were evolving, it became necessary to separate the internal milieu of developing cells from their varied, sometimes harsh, external surroundings. For this purpose, biological membranes developed (see Chapters 4 and 5). This compartmentalization enabled the evolution of other processes that allowed for functional specialization of cells. One such process that evolved was the active pumping of ions through the phospholipid membrane barrier. Early forms of biological pumps that evolved in prokaryotes were proton pumps that regulated intracellular pH (Gogarten et al., 1989). As biological complexity increased, more elaborate transport activities evolved to improve the control of the organism over its environment. As early life forms became exposed to various ionic environments, a need to develop some means of regulating cellular volume and of maintaining cytosolic homeostasis arose. Since water follows the movement of ions as they are transported between compartments, it became imperative to develop mechanisms to control the flow of ions, such as Na + and K +, through the plasma membrane. In eukaryotic cells, these processes are controlled by the sodium pump. The enzyme that fuels the sodium pump is the Na+,K+-ATPase. Na+,K+-ATPase is a member of the P-type ATPase superfamily, which differs structurally and functionally from both the F-type ATPases (ATP-synthases present in prokaryotes, chloroplasts, and mitochondria) and V-type ATPases (e.g., the H + pump located in vacuolar membranes of eukaryotic cells). The mechanism of action of the P-type ATPases involves formation of a transient, covalently phosphorylated intermediate during their reaction cycle. Structurally, they include at least one catalytic subunit with a mass ranging from 70 to 130 kDa. There are P-type ATPases in both prokaryotes (e.g., copper and cadmium ion pumps) and eukaryotes. The eukaryotic P-type ATPases can be further subdivided into two groups. One group of eukaryotic P-type ATPases consists only of a single subunit, designated ce, and includes the sarco-endoplasmic reticulum CaZ+-ATPase (SERCA), the plasma membrane CaZ+-ATPase, and the H+-ATPase found in yeast and plants. Another grouping of the eukaryotic PCell Physiology Sourcebook: A Molecular Approach, Third Edition
2 6 |
type ATPase family contains an additional subunit (/3), and includes the gastric H+,K+-ATPase and the Na+,K+-ATPase. The ce-subunit of the H+,K+-ATPase and the Na+,K+-ATPase show approximately 60% homology in amino acid sequence, and the structure of the genes encoding the two enzymes is very similar. The ce-subunit of SERCA is approximately 30% homologous in amino acid sequence to the ce-subunit of the Na+,K+-ATPase. Not surprisingly, the function and mechanism of action of this family of enzymes are closely related.
II. N a + - K + T r a n s p o r t
Figure 1 shows the distribution of Na +, K +, and C1- across the plasma membrane of a typical mammalian cell with a resting membrane potential o f - 7 0 mV. There is little driving force for chloride ions, since their Nernst potential (see Chapter 1) i s - 6 9 mV, near the resting membrane potential. Potassium ions will tend to flow out of the cell, since their equilibrium potential (-91 mV) is more negative than the transmembrane potential: Conversely, Na + will have a very strong force driving it into the cell, since both the chemical and electrical gradients (equilibrium potential of +64 mV) favor Na + uptake. The pumping of Na + out of the cell and K + into the cell, and hence regulation of cytosolic ionic concentrations and cell volume, therefore requires sufficient energy (derived from the hydrolysis of ATP) to overcome these driving forces. This, by definition, is the sodium pump activity. The nonequivalent transport of three Na + for two K + across the membrane maintains transmembrane gradients for these ions, and this became a convenient driving force for the secondary transport of metabolic substrates. In addition, the nonequivalent transport is electrogenic and leads to the generation of a transmembrane electrical potential, allowing cells to become excitable. The presence of the ionic gradients provided a mechanism for the regulation of many of the processes essential for the viability and stability of the cell. These include maintenance of intracellular pH via the activity of the Na+-H+ antiporter, contractility by means of K + activation of the actomyosin ATPase as well as Na+-Ca2+ exchange, and Copyright 9 2001 by Academic Press. All rights of reproduction in any form reserved.
262
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
FIGURE I. Distribution of sodium, potassium, and chloride ions in a typical mammalian cell.
metabolism through symport of Na + with metabolites such as sugars, amino acids, and neurotransmitters. The internal environment has a high concentration of K + and a low concentration of Na +. Intemal biological functions (such as the contractility of muscle cells) have, therefore, evolved to be activated by K + and inhibited by Na § ff a cell is damaged or becomes starved for energy, the concentration of K § will decrease and the concentration of Na + will increase because of either an increased leak of the ions with their gradients or a decreased pumping against the gradient. This decrease in K + and increase in Na + leads to inhibition of energy-utilizing processes and allows other compensatory mechanisms to attempt to repair any damage.
!11. Transport Mechanism Early studies of the pump used intact systems such as giant squid axons and erythrocyte ghosts, enabling both the direct measurement of ion fluxes using radioactive tracers and indirect measurements using electrophysiological techniques. Initial studies found that the appearance of 22Na in the bath solution following microinjection into the squid axon required that K + be present in the extracellular bath solution. This suggested that Na + effiux (transport out of the cell) was coupled to K + influx (transport into the cell). Experiments performed in parallel showed that, under physiological conditions, the effiux of 22Na and influx of 42K occurred at a coupling ratio of 3 Na+:2 K+:I ATP. More recent studies have demonstrated that this ratio is not fixed, but can vary depending on the concentrations of the ions both within and outside the cell (reviewed in Lauger, 1991; Tepperman et al., 1997). Transport studies using erythrocytes allowed researchers to determine the ion selectivity of the sodium pump that intraceUular activation of the pump explicitly requires. These sites have an intrinsic affinity 1of 0.2 mM for Na +. Potassium competes (but does not activate transport) with the Na + sites with an intrinsic affinity 1 Intrinsic affinity is the affinity of the ion in the absence of competing ions and is obtained by extrapolation to a concentration of zero competitor.
of 9 mM (Garay-Garrahan, 1973). In the presence of normal concentrations of internal K + (see Fig. 1), the apparent affinity for Na + at the intracellular site is approximately 3 mM. The extracellular transport sites are much less specific for potassium since Rb +, Cs +, Li +, T1+, NH~, and even Na + can activate the ATPase (Bell et al., 1977). The intrinsic affinities of the extracellular transport sites for K + and Na + are approximately 0.2 and 9-16 mM, respectively (Sachs, 1977). In the presence of normal concentrations of external Na + (see Fig. 1), the affinity for the extracellular K + is approximately 2 mM. Affinities for cations derived from studies of Na + and K + activation of hydrolysis of ATP in fragment enzyme preparations (Lindenmayer et al., 1974) are in agreement with the studies of intact erythrocytes. Studies with axons and erythrocytes also demonstrated that the activation of the pump is dependent on the presence of intracellular ATP and that, in particular, it is specific for MgATE The products of the hydrolysis, ATP and Pi (inorganic phosphate), remain within the cell. The dependence of the pump activity on metabolically derived ATP was confirmed by the finding that metabolic inhibitors such as azide, cyanide, and dinitrophenol greatly reduce the pump activity. Schatzmann (1953) reported that the sodium pump activity was specifically inhibited by cardiac glycosides, plant-derived steroids that have been used for 200 years to treat heart failure (Withering, 1785). Skou isolated an enzyme found in the plasma membrane whose hydrolysis of ATP was activated by both Na + and K + (Skou, 1957). This is in contrast to the majority of cytosolic enzymes, which are activated by K + and inhibited by Na +. This Na+,K+-ATPase is specifically inhibited by cardiac glycosides in the same concentration range that inhibits transport in intact cells, demonstrating that this enzyme is responsible for the sodium pump activity. For his work, Skou was awarded the 1997 Nobel Prize in Chemistry. Studies using partially purified, fragmented membrane preparations of Na+,K+-ATPase found that, in the presence of Na +, Mg 2+, and ATP-T32E the TP was covalently transferred to the enzyme, forming an acid-stable acyl phosphate bond. This established the Na+,K+-ATPase as a member of the P-type family of ATPases. When K + is added to the enzyme that has been phosphorylated in the presence of Na +, Mg 2+, and ATE the phosphate group is rapidly cleaved and the enzyme is allowed to cycle, continuing to hydrolyze ATP (Fig. 2). Both N-ethylmaleimide and oligomycin decrease the sensitivity of the enzyme to K+-induced dephosphorylation and increase the sensitivity of phosphorylated enzyme to ADP (rev. in Glynn, 1993). This led to the concept of two primary enzyme conformations, both of which can exist in either a phosphorylated or an unphosphorylated form. These are designated E 1, which has a high affinity for intracellular Na +, Na+n, and E 2, which has a high affinity for extracellular K +, (Ke+x).The cycling of the enzyme between the E 1 forms, binding cations at the intracellular face, and E 2 forms, binding cations at the extracellular face, results in the transport of the ions through the membrane. This concept, illustrated in Fig. 2, is referred to as the Post-Albers scheme in recognition of the contribution of these two scientists. In the E 1 conformation, the enzyme has high affinity for both intraceUular Na + and MgATE Following the binding of both of these ligands, the y phosphate group of ATP is transferred to an aspartyl residue on the enzyme, forming a "high-
17. Sodium Pump Function
3Na ++ATP in
ADP
2K+ x
FIGURE 2.
263
3Na+ ex
Post-Albersreaction mechanism for the sodium pump.
energy" phosphoenzyme. In this high-energy state, the phosphate group can be readily returned to ADP, resynthesizing ATE In the forward cycle of the enzyme, the "high energy" residing in the phosphate bond is used to drive a change in the conformation of the enzyme, resulting in the exposure of Na + to the extracellular surface. This "low-energy" phosphoenzyme, with Na + bound, is now said to be in an E 2 conformation. The phosphate group in this state cannot be easily transferred back to ADP to form ATE The E2P conformation has low affinity for extracellular Na+; therefore, it is released. This conformation has high affinity for extracellular K § After K + binds at the extracellular surface, the phosphatase activity of the enzyme is activated, the phosphate group is cleaved, and K § is translocated and released on the intracellular surface. The enzyme is once again in the E 1 conformation and capable of binding Na § and ATP, completing the cycle. Because of the cycling between these two conformations, the P-type ATPases are also referred to as E1E2ATPases. Additional evidence in support of the two-conformation hypothesis came from studies using controlled proteolysis. Jorgensen found that in the presence of Na § specific sequences for trypsin proteolysis were exposed, designated T2 and T3 (see Section IV and Fig. 4). In addition, a site for chymotrypsin, C3, was observed. In the presence of K +, the T2 site is still cleaved, but an additional site, T1, is now exposed, while the T3 and C3 site are now hidden (Jorgensen, 1977). Fluorescence measurements of enzyme labeled at K501 with fluorescein isothiocyanate (FITC) also supported the hypothesis that the enzyme undergoes structural rearrangements dependent on ligand conditions. FITC labeling is competitive with ATP; however, mutation of the labeled residue did not affect the activity of the enzyme. It is likely, therefore, that the inhibition of ATP binding following FITC modification is due to an indirect, steric effect. Nonetheless, a change in the fluorescence properties of the FITC following the addition of Na + or K + indicates that the conformation of the active site is changed (Hegyvary and Jorgensen, 1981). These two types of studies have clearly demonstrated that the enzyme is undergoing structural rearrangement when exposed to different ionic conditions. It is known that the Post-Albers scheme (see Fig. 2) is overly simplistic. Perhaps a more realistic view of the reaction mechanism is that it consists of a series of steps whereby the ions are bound, occluded, and released from the enzyme (Forbush, 1987a). The occluded state refers to a form of the enzyme where the ions are trapped within the
membrane and are not accessible from either the extracellular or intracellular compartments. This becomes intuitively obvious if one considers the movement of a 2-A molecule being moved through a 50- to 70-/~ thick membrane against its concentration gradient. Figure 3 illustrates a modified version of the Post-Albers scheme that is more consistent with recent findings. First, in the E1 conformation, it has been proposed that the Na § ions bind in a multistep process, with the rate of binding for the first two sodium ions being so fast as to be indistinguishable, followed by the third Na + binding at a slower rate (Heyse et al., 1994). In the presence of Na +, MgATP is bound with high affinity and the 7 phosphate is transferred to the enzyme. This occurs with the simultaneous occlusion of either two or three Na + ions. The sodium ions are then translocated and released on the extracellular surface; it is believed that the release of sodium is an ordered process. One ion is released first, followed by the remaining two. The release of the sodium ions allows for the binding of the potassium ions, again in an ordered fashion, with positive cooperativity, such that the binding of the first potassium to the extracellular surface increases the affinity for the second (Tepperman et al., 1997). If the concentration of extracellular K + is low, it is possible that only one K + will bind and be transported. The binding of potassium results in a rapid dephosphorylation of the enzyme along with occlusion of the ions. In the absence of intracellular ATE the rate of K + release to the intracellular surface is very slow. Following the binding of ATP the rate of deocclusion is greatly increased. Occlusion studies by Forbush (1987b) have demonstrated that the release of K + in the absence of ATP following transport is a multistep, ordered process in which one ion is released at a rapid rate while the second ion is released more slowly. This process is consistent with a flickering gate model, in which the exposure of the cation binding sites limits the release to a single ion at a time. The release of K + allows Na + to bind and increases the affinity of the enzyme for ATE thus completing the enzyme cycle. Figure 3 does not, however, take into account that internal K + binds to (competes with) the Na + transport sites, or that extracellular Na + is able to bind to (compete with) the external K + transport sites. Extracellular Na + has been implicated additionally as an allosteric effector of K + transport (Sachs, 1977). From the mechanism depicted in Fig. 3, it is clear that the coupling ratio does not have to be fixed at 3 Na+:2 K + under all conditions. The study by Tepperman et al. demonstrated that the only models that fit their data were models in which translocation could take place when either one or two K + were bound to the enzyme (Tepperman et al., 1997).
IV. Na§
Structure
The Na+,K+-ATPase is composed of two subunits (Fig. 4) in a one-to-one stoichiometry: the alpha (a) or catalytic subunit and the beta (/3) or glycoprotein subunit (Lane et al., 1973). The a-subunit contains all of the known binding sites for ligands that regulate pump function. The complete primary structure of the sheep a- and fl-subunits has been determined by the cloning of cDNAs corresponding to the
264
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
FIGURE 3. ModifiedPost-Albers scheme showing binding, occlusion, and release of cations (Na+ and K+) from both the intracellular and extracellular surfaces, and possible noncanonical flux modes. Enzyme forms that contain occluded cations are shown with the cations in parenthesis (i.e., E1P(Na)3). Enzyme forms with cations bound are shown with a dot (i.e., EI-Na3). Na+ is designated by filled circles and K§ by open circles. For simplicity, the binding and hydrolysis of ATP are omitted.
mRNA for both subunits (Shull et al., 1985, 1986a). The sheep kidney a-subunit consists of a protein of M r 112 117 (1016 amino acid residues) and the fl-subunit consists of 302 amino acid residues with a core protein of M r of 34 939 (Shull et al., 1986b). The cel-subunit exists as multiple isoforms, and the primary sequences of these isoforms from a variety of species, including the a 1, a2, a3, and a4 from the human and rat, have been determined (rev. in Lingrel et al., 1990, and see Shamraj and Lingrel, 1994). The distribution of the four a isoforms is tissue specific (demonstrated in Fig. 5) and is both hormonally and developmentally regulated. The a l isoform has been found in all tissues and cell types examined and is considered to be the housekeeping form of the enzyme. The a 2 isoform has been found primarily in skeletal muscle and brain, with smaller
amounts in the adult hearts of most species, including human. The a 3 isoform has been found primarily in brain, but also in neonatal rat heart and adult human heart. The a 4 isoform has been found only in rat and human testis, at an mRNA level comparable to the level of c~1 in the brain. The sequence of the four rat isoforms shows 80-85% homology. The sequence of the ~ 1 isoforms is also highly conserved between species. Based on hydropathy plots and analogy with the H+,K+-ATPase and the Ca2§ the a-subunit of Na+,K+-ATPase is predicted to have 10 membrane-spanning regions labeled HI to H10, as shown in Fig. 4. The fl-subunit of sheep kidney contains three N-linked oligosaccharide sites and seven cyst(e)ine residues. The one free sulfhydryl, C44, is postulated to be in the membranespanning region. The other six residues participate in three
17. Sodium Pump Function
265
FIGURE 4. The predicted transmembrane model of the sheep a 1- and/31-subunits of the Na+,K+-ATPase. The membranespanning helices are represented by H 1-H 10. Asterisks (*) identify the three glycosylationsites, and the open circles denote the position of the disulfide bridges on the/3-subunit. The sites of tryptic digestion have been designated as T1 (R438), T2 (K30), T3 (R262), and chymotrypticdigestion, C3 (L266). The FITC binding site is at residue K501 and the phosphorylationsite (P) is at D369.
disulfide bonds (Kirley, 1989). Based on hydropathy plots, the /31-subunit possesses only one transmembrane region (see Fig. 4). Three isoforms of the/3-subunit have been described (Shull et al., 1988). The/31 isoform described previously is the major form expressed in virtually all tissues, with /32 occurring primarily in brain. Although the two mammalian isoforms exhibit only 35% homology, the location of the glycosylation sites, the conserved cysteine residues, and similarity of the hydropathy plots suggest that these two isoforms possess high structural homology. Interestingly, the 132
FIGURE 5. Northemblot analysis of the mRNA expression levels of the four ~-subunit isoforms in various tissues of rats. (From Shamraj, O. I., and Lingrel, J. B. (1994).A putative fourth Na§ alpha-subunit gene is expressed in testis. Proc. Natl. Acad. Sci. USA 91, 12952-12956. Copyright 1994 National Academy of Sciences, U.S.A.)
isoform appears to be identical to the adhesion molecule on glia (AMOG). A third isoform,/33, has only been observed during early development in X e n o p u s laevis. The physiological function of the/3-subunit is unknown. Expression of functional Na+,K+-ATPase activity in yeast, which do not contain a Na+,K+-ATPase, requires that both the a- and/3-subunits be cotransfected (Horowitz et al., 1990). This indicates that the/3-subunit is required for the proper assembly of the a-subunit into the membrane and for the exit of this a/3 complex from the endoplasmic reticulum and proper insertion into the cell membrane (rev. by Lingrel et al., 1990; McDonough et al., 1990; Sweadner, 1989). The/3-subunit of the enzyme also alters the susceptibility of the a-subunit to proteolytic enzymes, further showing its involvement in stabilizing the enzyme's structure. Furthermore, this subunit has been postulated to affect cation affinity, suggesting that it may undergo conformational changes associated with enzyme function (Chow and Forte, 1995). The/3-subunit of Na+,K +ATPase is believed to be in close proximity to the cardiac glycoside binding site on the ~-subunit because it can be labeled with photoaffinity derivatives of digitoxin. The/3-subunit of H+,K+-ATPase and Na+,K+-ATPase are highly homologous and, in fact, the/3-subunit of the H+,K+-ATPase can replace the native Na+,K+-ATPase/3-subunit and allow functional expression of ~ (Horisberger et al., 1991). The c~-subunit of the transport ATPases have in common the phosphorylation of an aspartic acid residue (D369, 2 see Fig. 4) as a part of their enzymatic cycle. The large intracellular loop between H4 and H5 has been shown, by chemical modification and ATP analog labeling, to contain the ATPbinding site (in the region of K501 where FITC is covalently 2The numberingsystemused in this chapteris for the sheep a 1 isoform.
266
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
attached; Kirley et al., 1984) and the phosphorylation site (D369). Immunological data suggest that both the Nterminal and C-terminal regions are on the inside of the cell. Since the ATP site is also located intracellularly, these findings orient the 10 transmembrane regions (see Fig. 4). The N-terminal region contains a large number of both positively and negatively charged residues, including a lysine-rich sequence. It has been argued that this region may act as a cation-selective gate and/or participate in transport via saltbridge formation. Studies investigating the contribution of the N-terminal domain to cation affinity have shown that this region, which varies greatly between the a isoforms, does reduce the apparent affinity for cations, albeit only by a factor of 2 to 3 (Horisberger et al., 1991). It has also been suggested that the sequence between the phosphorylation site (D369) and the H4 transmembrane region may act as an energy-transduction region (Shull et al., 1985). Another area of research pertains to both the internal and external cation binding sites. It has been difficult to probe the internal Na § sites because of the difficulty of controlling internal Na + and the plethora of compensatory Na § transport mechanisms. Information about the interactions of external Na § and K + with the Na§247 has come from 86Rb uptake in whole cells and [3H]ouabain experiments in fragmented membrane preparations. These techniques have yielded apparent affinities for K + and Na + and detailed information about their interactions through the analysis of complex mechanistic models. The c~3 isoform has a higher affinity for extracellular K + and lower affinity for intracellular Na + than do a 1 and ce2 (Munzer et al., 1994). In addition, site-directed mutagenesis has been used in an effort to determine the amino acids involved in cation binding. Although the initial hypothesis was that negatively charged residues in the predicted transmembrane region would be involved in cation binding and transport, these studies have suggested that the binding site does not consist of a single residue, but rather is a binding pocket composed of a number of residues. The most significant effects have been observed by replacement of the Asp residue at position 804 in sheep ce1, where every replacement that has been made resulted in a non-functional enzyme. At residue 808, the only surviving mutation is the conservative replacement of aspartate with glutamate. Other residues (shown in Fig. 6) that affect cation affinity, possibly through indirect mechanisms,
are E327, E779, $775, and D926 (summarized in Van Huysse et al., 1996). This has led to the hypothesis that the H5-H6 region, which includes residues $775, E779, D804, and D808, is directly involved in cation binding and transport. It is conceivable that this entire domain may shift in the membrane, facilitating ion transport, similar to what has been postulated for the $4 segment of the voltage-gated ion channels (Sigworth, 1993). The most convincing evidence regarding the nature of the cardiac glycoside binding site has arisen from the elegant experiments first performed by Price and Lingrel (1988). These experiments took advantage of the fact that the rodent a 1 isoform is one thousand-fold less sensitive to glycoside inhibition than other isoforms. Expressing an "insensitive" enzyme in HeLa cells, whose endogenous enzyme is glycoside sensitive, allowed growth in media with a high concentration of ouabain, a water-soluble cardiac glycoside. If the cells do not express an insensitive transfected enzyme, they die in 0.2 txM ouabain. Expression, however, of mutant enzyme with decreased sensitivity to glycosides will allow the cells to live in 0.2 IxM ouabain. The amino acid sequence of the H1-H2 region differs substantially between sheep and rat a l-subunits. Using chimeric proteins and site-directed mutagenesis, it was determined that the presence of the charged residues, R111 and D 122, in the rat ce1 isoform accounts for the decreased ouabain sensitivity observed for the rat (Price and Lingrel, 1988). Most species have uncharged residues in this position (e.g., sheep a l has Q l l l and N 122), and these species are sensitive to cardiac glycosides. It has been hypothesized that, following the binding of the cardiac glycosides, these noncharged residues partition into the membrane, resulting in a conformation of the enzyme from which the rate of glycoside release is very slow, and thus the affinity very high. In addition to these two residues, D121, also in the H1H2 extracellular loop, has been shown to affect the glycoside affinity. Mutagenesis of other regions has revealed that the glycoside binding pocket spans the majority of the extracellular and transmembrane topology (see Fig. 6) (summarized in Palasis et al., 1996). These diverse regions include the H 1 transmembrane domain (C104 and Y108), the H5 transmembrane segment (F786), the H5H6 extracellular "hairpin" loop (L793), the H6 transmembrane segment (T797), the H7 transmembrane domain (F863), and the H7H8 extra-
FIGURE 6. The transmembrane amino acid residues of the sheep c~l sequence. The boxed residues represent amino acids implicated in affecting glycoside sensitivity, and encircled residues are those implicated in altering cation affinity.
17. S o d i u m P u m p F u n c t i o n
cellular loop (R880) (see Fig. 6). It is interesting to note that the H5H6 region has been implicated in glycoside binding and also appears to be central in the cation-transport processes of the enzyme. Ouabain or K + is able to protect the enzyme from cleavage under extensive tryptic digestion at the borders of the H5H6 region, suggesting that this region is involved in the conformational transitions of the enzyme (Lutsenko and Kaplan, 1993).
V. Cardiac Glycosides Plant extracts containing cardiac glycosides have been used for many years as medicines (for their therapeutic effect) and as poisons (for their toxic effect) by diverse groups that include ancient Egyptians, Romans, Chinese, and African tribesmen (Hoffman and Bigger, 1985). Withering (1785) noted the use of foxglove (Digitalis purpurea) extracts (which contain cardiac glycosides) in treatment of dropsy (pedal edema secondary to heart failure). For the next hundred years, digitalis (plant extracts of cardiac glycosides, containing primarily digoxin) was used, frequently at excessive doses, for the treatment of a range of diseases. By the early 20th century, digitalis was known to increase the force of myocardial contractility (positive inotropic effect) and thus to be useful in the treatment of congestive heart failure. The toxic effects of digitalis, however, are frequent and can be severe, even fatal. In spite of this, digoxin (the drug used clinically) is still one of the most widely used drugs, being the sixth most commonly prescribed drug in the United States in 1995 (Anon., 1996). The mechanism of action now generally accepted is termed the sodium pump lag hypothesis and postulates that therapeutic concentrations of digoxin inhibit a moderate (30-40%) fraction of the membrane Na+,K+-ATPase. This, in turn, causes a slight increase in internal Na § that has two significant effects: (1) it inhibits the forward operation of the Na+-Ca 2+ exchanger (Chapter 19), slowing down the rate at which Ca 2§ exits the cell (note that if Na § concentration is raised above approximately 20 mM, the Na+-Ca 2+ exchanger can be forced into reverse mode and drive Ca 2§ influx); (2) it stimulates the 60-70% fraction of remaining uninhibited pumps, which tend to return the internal Na + concentration to homeostasis. Thus, therapeutic levels of digoxin produce transient increases in internal Na § in heart muscle, which lead to transient increases in internal Ca 2§ which in turn stimulate the release of a larger amount of Ca 2§ from internal stores in the sarcoplasmic reticulum (Chapter 55). The released calcium interacts with contractile proteins to increase the force of contraction. Toxic concentrations of digoxin lead to extensive (>60%) inhibition of Na+-K § transport, so that restoration of normal diastolic levels of Na § and Ca 2+ is not possible. Digoxin also indirectly produces sympathetic and parasympathetic effects which can result in cardiac arrhythmias. The first step in both the therapeutic and toxic effects is clearly mediated via inhibition of Na+,K § ATPase (see review by Thomas et al., 1989). Three factors determine the affinity of cardiac glycosides and their analogs for Na+,K+-ATPase: (1) the nature (and concentration) of the physiological ligands (Mg 2§ Na § ATP,
267
Pi, and K +) bound to the enzyme, (2) the structure of the drug, and (3) the structure of the receptor. In the following scheme, LRD represents the ternary complex of drug (D), receptor (R), and physiological ligand(s) (L). The dissociation constant for the rapid binding of the physiological ligand to receptor is represented by K L, and the association and dissociation rate constants for binding of drug to receptor are represented by k 1 and k_1, respectively. The dissociation constant K o is equal to k_l/k 1. KL
R + L ~-~ LR + D
k_ 1 kl
LRD
For a given cardiac glycoside, the on-rate, and consequently the K o, is very sensitive to the nature of the ligand(s) present in the medium. Inorganic phosphate, Mg 2§ Na +, and ATP stimulate and K § inhibits the rate of binding, leading many to hypothesize that phosphorylation is necessary for ouabain binding. By varying the concentration of the ligand, the dissociation constant for the binding of the ligand to the enzyme can be obtained. The dissociation constant (KL) for the Mg ~+ site is 0.2 mM, and when saturated, increases the affinity for ouabain by a factor of 300 over that in presence of buffer alone. The stimulation by Na § (KL = 14 mM) is competitively inhibited by K + (KL = 0.2 mM). The similarity of these K L values to those derived from transport studies suggests that these monovalent cation regulatory sites for cardiac glycosides are equivalent to the transport sites. In contrast to the on-rate, the off-rate of a specific cardiac glycoside from a specific isoform of the enzyme is not influenced to the same degree by the presence of physiological regulatory ligands. This could be the result of a ligandinduced conformational change, such that the enzymedrug complex initially formed is transformed into a second, more stable complex whose rate of dissociation is independent of the regulatory ligands present. Mutation of the phosphorylation site, D369N, produces an enzyme that cannot be phosphorylated, cannot hydrolyze ATP, and cannot translocate Na § and K § Expression of the sheep ~1 mutant in NIH 3T3 cells, whose endogenous Na§ is insensitive, produces a membrane form of the Na§247 that can still bind ouabain and can be used as a probe for the physiological ligands (Kuntzweiler et al., 1995). The mutant enzyme still binds ouabain with high affinity, demonstrating that phosphorylation is not necessary for ouabain binding. The stimulatory effect of phosphate on the binding of ouabain, in the presence of Mg 2§ is absent in the mutant, showing that the effect of phosphorylation occurs at D369. The fact that ATP and MgATP still bind to the mutant enzyme but inhibit ouabain bindingmin contrast to the wild type, in which they stimulate ouabain binding-demonstrates that the Mg-nucleotide site is still present. Three Na § and two K § still bind to the mutant enzyme with only slight changes in affinity and still affect ouabain binding (in the presence of Mg2+). This indicates that Na § and K § can induce conformational changes in the enzyme that affect the affinity for ouabain independent of whether the enzyme is phosphorylated. This is consistent with the hypothesis that ouabain prefers to bind to the E 2 conformation. This study also demonstrated that there are distinct sites for Mg 2§ and MgATP that can be simultaneously occupied.
268
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
The structure of the cardiac glycosides also affects their affinity. Classic structure-activity relationship studies of cardiac glycosides, reviewed by Thomas et al. (1989), concluded that three primary domains of the cardiac glycosides contribute to their biological activity. The first of these is a steroid ring system with a cis configuration of A/B and C/D rings, unique to the cardiac glycosides, as well as 3/3- and 14/3-hydroxyl groups (Fig. 7). The cardiac glycosides also require the presence of an unsaturated/3-1actone on carbon 17. The second domain is a five-membered a/3 unsaturated lactone ring in the cardenolide family of the cardiac glycosides, and a six-membered diunsaturated lactone ring in the bufadienolides, a related family of compounds (found primarily in toads) with similar pharmacological action, but significantly higher affinity. Reduction of the double bond(s) in the lactone ring reduces the biological activity. The third domain is a c a r b o h y d r a t e moiety. Replacement of the 3/3-hydroxyl group of the aglycone with a sugar residue increases affinity by dramatically slowing the rate of dissociation of the cardiac glycoside from the receptor. The presence of the rhamnose moiety on ouabain, for example, increases the affinity ten- to thirty-fold. Cardiac glycosides lacking a carbohydrate group are referred to as genins (i.e., digoxigenin is digoxin without the tridigitoxose moiety on C3, but with a 3/3-hydroxyl group; see Fig 7).
VI. Summary The Na+,K+-ATPase is one of the most vital proteins present in animal cells, as demonstrated by the remarkable degree of conservation across evolutionary lines and the fact that it has been found in every animal cell type studied. This enzyme places a high metabolic burden on the cell, using a substantial portion of the total cellular energy production in maintaining the osmotic balance and ionic gradients across the plasma membrane. This enzyme is crucial for the survival of animal cells, as demonstrated by the fact that transgenic knockout (either a 1 or a 2 isoform) mice do not survive (Lingrel, personal communication, 1997). In addition to its role in maintaining homeostasis, inhibition of the pump has significant therapeutic relevance by virtue of the secondary increase in Ca 2+ concentration leading to an increase in cardiac contractility. The pump activity is due to an integral membrane protein consisting of two subunits, a and/3, and belonging to the Ptype family of ATPases. This enzyme hydrolyzes ATP in the
Me
O
~
O tone
tridigitoxose- O
FIGURE 7. side, digoxin.
\
H The structure of the most widely used cardiac glyco-
presence of Na + and K +. The a-subunit has at least four isoforms and the/3-subunit three, each of which is differentially expressed. The cation affinities for the activation of ion transport, ATP hydrolysis, and effects on glycoside binding agree well with each other, indicating that the same binding sites are involved in all three activities. Although the enzyme has been studied in great detail over the past 40 years, there remain many unanswered questions that can be placed into three basic categories: (1) What is the three-dimensional structure of the enzyme? (2) What is the mechanism of cation transport? (3) What are the structural domains that make up ligandbinding sites?
Bibliography Anon. (1996). Top 200 drugs of 1995. Pharmacy Times, April, p. 29. Bell, M. V., Tonduer, F., and Sargent, J. R. (1977). The activation of sodium-plus-potassium ion-dependent adenosine triphosphatase from marine teleost gills by univalent cations. Biochem. J. 163, 185-187. Chow, D. C., and Forte, J. G. (1995). Functional significance of the betasubunit for heterodimeric P-type ATPases. J. Exp. Biol. 198, 1-17. Forbush, B. (1987a). Na+, K+, and Rb+ Movements in a single turnover of the Na/K pump. Curr. Top. Membr. Transp. 28, 19-39. Forbush, B. (1987b). Rapid release of 42K or 86Rb from two distinct transport sites on the Na,K-Pump in the presence of Pi or vanadate. J. Biol. Chem. 262, 11116-11127. Garay, R. P., and Garrahan, P. J. (1973). The interaction of sodium and potassium with the sodium pump in red cells. J. Physiol. (London) 231, 297-325. Glynn, I. M. (1993). All hands to the sodium pump. J. Physiol. 462, 1-30. Gogarten, J. P., Rausch, T., Bemasconi, P., Kibak, H., and Taiz, L. (1989). Molecular evolution of H+-ATPase I. Methanococcus and Solfolobus are monophyletic with respect to eukaryotes and eubacteria. Z. Naturforsch. 44, 641-650. Hegyvary, C., and Jorgensen, P. L. (1981). Conformational changes of renal sodium plus potassium ion-transport adenosine triphosphatase labeled with fluorescein. J. Biol. Chem. 256, 6296-6303. Heyse, S., Wuddell, I., Apell, H.-J., and Sturmer, W. (1994). Partial reactions of the Na,K-ATPase: determination of rate constants. J. Gen. Physiol. 104, 197-240. Hoffman, B. E, and Bigger, J. T. (1985). Digitalis and allied cardiac glycosides. In "Pharmacological Basis of Therapeutics" (A. G. Gilman, W. S. Goodman, T. W. Rail, and E Murad, Eds.), pp. 716-747. Macmillan, New York. Horisberger, J. D., Jaunin, P., Reuben, M. A., Lasater, L. S., Chow, D. C., Forte, J. G., Sachs, G., Rossier, B. C., and Geering, K. (1991). The H,K-ATPase beta-subunit can act as a surrogate for the betasubunit of Na,K-pumps. J. Biol. Chem. 266, 19131-19134. Horowitz, B., Eakle, K. A., Scheiner-Bobis, G., Randolph, G. R., Chen, C. Y., Hitzeman, R. A., and Farley, R. A. (1990). Synthesis and assembly of functional mammalian Na,K-ATPase in yeast. J. Biol. Chem. 265, 4189-4192. Jorgensen, P. L. (1977). Purification and characterization of (Na§ + K+)-ATPase. VI. Differential tryptic modification of catalytic functions of the purified enzyme in presence of NaCI and KC1. Biochim. Biophys. Acta 466, 97-108. Kirley, T. L. (1989). Determination of three disulfide bonds and one free sulfhydryl in the/3 subunit of (Na,K)-ATPase J. Biol. Chem. 264, 7185-7192. Kirley, T. L., WaUick, E. T., and Lane, L. K. (1984). The amino acid sequence of the fluorescein isothiocyanate reactive site of lamb and rat
17. Sodium Pump Function kidney Na § and K+-dependent ATPase. Biochem. Biophys. Res. Commun. 125, 767-773. Kuntzweiler, T. A., Wallick, E. T., Johnson, C. L., and Lingrel, J. B. (1995). Amino acid replacement of D369 in the sheep al isoform eliminates ATP and phosphate stimulation of [H3]ouabain binding to the Na+,K+-ATPase without altering the cation binding properties of the enzyme. J. Biol. Chem. 270, 16206-16212. Lane, L. K., Copenhaver, G. H., Lindenmayer, G. E., and Schwartz, A. (1973). Purification and characterization of and [3H]ouabain binding to the transport adenosine triphosphatase from outer medulla of canine kidney. J. Biol. Chem. 248, 7197-7200. Lauger, P. (1991). "Electrogenic Ion Pumps," Vol. 5. Sinauer Associates, Sunderland, MA. Lindenmayer, G. E., Schwartz, A., and Thompson, H. K., Jr. (1974). A kinetic description for sodium and potassium effects on (Na + + K+)adenosine triphosphatase: a model for a two-nonequivalent site potassium activation and an analysis of multiequivalent site models for sodium activation. J. Physiol. (London) 236, 1-28. Lingrel, J. B., Orlowski, J., Shull, M. M., and Price, E. M. (1990). Molecular genetics of Na,K-ATPase. In "Progress in Nucleic Acids Research and Molecular Biology," Vol. 38, pp. 37-89. Academic Press, London. Lutsenko, S., and Kaplan, J. H. (1993). An essential role for the extracellular domain of the Na,K-ATPase/3-subunit in cation occlusion. Biochemistry 32, 6737-6743. McDonough, A. A., Geering, K., and Farley, R. A. (1990). The sodium pump needs its/3 subunit. FASEB J 4, 1598-1605. Munzer, J. S., Daly, S. E., Jewell-Motz, E. A., Lingrel, J. B., and Blostein, R. (1994). Tissue- and isoform-specific kinetic behavior of the Na,K-ATPase. J. Biol. Chem. 269, 16668-16676. Palasis, M., Kuntzweiler, T. A., Arguello, J. M., and Lingrel, J. B. (1996). Ouabain interactions with the H5-H6 hairpin of the Na,KATPase reveal a possible inhibition mechanism via the cation binding domain. J. Biol. Chem. 271, 14176-14182. Price, E. M., and Lingrel, J. B. (1988). Structure-function relationships in the Na,K-ATPase a-subunit: site-directed mutagenesis of glutamine-111 to arginine and asparagine-122 to aspartic acid generates a ouabain-resistant enzyme. Biochemistry 27, 8400-8408. Sachs, J. R. (1977). Inhibition of the Na-K pump by external sodium. J. Physiol. (London) 264, 449-470.
269 Schatzmann, H. J. (1953). Herzglykoside als Hemmstoffe fur den activen Kalium und Natrium-transport durch die Erythrocytenmembran. Helv. Physiol. Pharmacol. Acta 11, 346-354. Shamraj, O. I., and Lingrel, J. B. (1994). A putative fourth Na+,K~+)ATPase alpha-subunit gene is expressed in testis. Proc. Natl. Acad. Sci. USA 91, 12952-12956. Shull, G. E., Schwartz, A., and Lingrel, J. B. (1985). Amino-acid sequence of the catalytic subunit of the (Na+,K+)-ATPase deduced from a complementary DNA. Nature 316, 691-695. Shull, G. E., Greeb, J., and Lingrel, J. B. (1986a). Molecular cloning of three distinct forms of the Na,K-ATPase alpha subunit from rat brain. Biochemistry 25, 8129-8132. Shull, G. E., Lane, L. K., and Lingrel, J. B. (1986b). Amino-acid sequence of the/3-subunit of the (Na§ deduced from a cDNA. Nature 321, 429--431. Shull, G. E., Young, R. M., Greeb, J., and Lingrel, J. B. (1988). Amino acid sequences of the c~ and/3 subunits of the Na,K-ATPase. Prog. Clin. Biol. Res. 268A, 3-18. Sigworth, F. J. (1993). Voltage gating of ion channels. Quart. Rev. Biophys. 27, 1-40. Skou, J. C. (1957). The influence of some cations on an adenosine triphosphatase from peripheral nerves. Biochim. Biophys. Acta 23, 394--401. Sweadner, K. J. (1989). Isozymes of the Na§ Biochim. Biophys. Acta 988, 185-220. Tepperman, K., Millette, L. A., Johnson, C. L., Jewell-Motz, E. A., Lingrel, J. B., and Wallick, E. T. (1997). Mutational analysis of glutamate 327 of Na§ reveals stimulation of 86Rb uptake by external K. Am. J. Physiol., 273 (cell physiol. 42), C2065-C2079. Thomas, R., Gray, P., and Andrews, J. (1989). Digitalis: its mode of action, receptor and structure-activity relationship. In "Advances in Drug Research" (B. Testa, Ed.), Vol. 19, pp. 311-562. Academic Press, New York. Van Huysse, J. W., Kuntzweiler, T. A., and Lingrel, J. B. (1996). Critical effects on catalytic function produced by amino acid substitutions at Asp804 and Asp808 of the ce1 isoform of Na,K-ATPase. FEBS Lett. 389, 179-185. Withering, W. (1785). "An Account of the Foxglove and Some of Its Medicinal Uses: With Practical Remarks on Dropsy and Other Diseases" C. G. J. and J. Robinson, London.
Istvan Edes and Evangelia G. Kranias
Ca2+.ATPases
|. Introduction An important role of Ca 2+ in muscle contraction was first indicated a century ago by Ringer (1883), who demonstrated that the frog's heart would not contract in the absence of extracellular Ca 2+. Since then, it has been shown that Ca 2+ is a physiological regulator for the contractile proteins and several other enzymes and processes in muscle. This chapter will focus on the role of the various CaZ+-ATPases in maintaining Ca 2+ homeostasis in the cell, with special emphasis on the sarcoplasmic reticular (SR) Ca2+-ATPase(s), which is the primary regulator of the Ca 2+ levels and thus contractility in muscle. During the cardiac action potential, Ca 2§ enters the cell via Ca 2+ channels, which also act as dihydropyridine receptors (Fig. 1). This Ca 2+ can either activate the myofilaments directly or produce the release of additional Ca 2+ from the SR. The SR Ca2+-release channel in cardiac and skeletal muscle also acts as a ryanodine receptor and spans the gap between the transverse tubule and the SR ("foot" protein). Furthermore, it has been shown that the outer cell membrane Ca 2§ channel is located close to the SR Ca 2§ channel. Thus, the excitationcontraction coupling apparently involves the sarcolemmal Ca 2+ channel and the SR Ca2+-release channel with the Ca 2+ current through the sarcolemmal channel being responsible for the initiation of Ca 2§ release from the SR (Fig. 1). In skeletal muscle, the sarcolemmal membrane depolarization itself apparently is responsible for the induction of SR Ca 2§ release. The relative importance of release from the SR in activation of the cardiac muscle contraction varies from preparation to preparation, but in the heart of mammals it usually accounts for 40-70% of the Ca 2+ required (Bers, 1991). The rising cytosolic Ca 2§ concentration induces contraction through binding to troponin C, which activates a chain of conformational changes, allowing the thin and thick filaments to interact. Subsequently, Ca 2§ is dissociated from troponin C and is rapidly removed from the cytosol by various systems, resulting in relaxation. At least three processes are responsible for the removal of Ca 2§ to end contraction (Fig. 1): (a) the SR Ca 2§ pump, which actively translocates
Cell Physiology Sourcebook: A Molecular Approach, Third Edition
2 71
Ca 2+ at the cost of ATP into the SR system; this is thought to be the most important process in mediating relaxation; (b) the Na+-Ca 2+ exchanger, which transports Ca 2§ out of the cell during diastole; and (c) the sarcolemmal Ca2+-ATPase, which also extrudes Ca 2+ from the cell. The SR is a tubular network, which serves as a sink for Ca 2§ ions during relaxation and as a Ca 2§ source during contraction. In cardiac muscle, about 60-70% of the intracellular Ca 2§ released during systole is taken up by the SR (Bers, 1991), and the remaining amount is extruded from the cell by the Na+-Ca 2§ exchanger and the sarcolemmal Ca2+-ATPase.
FIGURE 1. Schematicdiagram of Ca2+ fluxes in cardiac cell. Na-CaX, Na+-Ca2§ exchanger; Calseq., calsequestrin;/Ca' slow inward Ca2§ current; SR, sarcoplasmic reticulum.
Copyright 9 2001 by Academic Press. All rights of reproduction in any form reserved.
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
272
Structure and Distribution of the Sarcoplasmic Reticular Ca2+-ATPase (SERCA) Isoforms
TABLE 1
Gene
Splice
Tissue
SERCA1a
ao
Adult fast skeletal muscle
SERCA1
b
Neonatal fast skeletal muscle
SERCA2
a
Cardiac/slow skeletal muscle
SERCA2
b
Smooth muscle/nonmuscle
SERCA3
9
Various tissues
aThe SERCA numbers identify different gene products. bThe letters a and b indicate spliced isoforms.
The SR in both skeletal and cardiac muscles contains an acidic protein, calsequestrin (Fig. 1), which binds 40-50 mol of Ca 2+ per mol of protein. The binding and release of Ca 2+ by calsequestrin is believed to be an integral step of excitation-contraction coupling, but the details of this process are still not fully understood. Mitochondria can also accumulate large amounts of Ca 2+ under pathological conditions (ischemia, Ca 2+ overload, etc.) (Bers, 1991).
!!. Sarcoplasmic Reticular (SR) Ca2§
ratio has been generally found to be lower (0.4-1.0 mol Ca2+/mol ATP). Ca 2+ has been shown to bind to a region involving several of the membrane-spanning a-helices (M4, M5, and M6) on the cytoplasmic side. The important amino acid residues constituting the Ca 2+ binding sites are Glu 309 on the M4 transmembrane segment; Glu 771 on the M5 transmembrane segment; and Asn 796, Thr 799, and Asp 800 on the M6 transmembrane segment. It has been proposed that Glu 771 and Thr 799 are associated with the first Ca 2+ binding site (site 1), while Glu 309 and Asn 796 are associated with the binding of the second Ca 2+ ion (site 2). Asp 800 was suggested to donate ligands to both Ca 2§ binding sites (MacLennan et al., 1997; Andersen and Vilsen, 1998). During the Ca 2§ transport cycle, the enzyme undergoes a transition from a high-affinity state to a low-affinity state for Ca 2§ and the ions are translocated from the binding sites into the lumen of the SR (Fig. 2). This reaction pathway is characterized by the covalent phosphorylated Ca2+-ATPase form (El~P), when the energy of ATP is transferred to an acylphosphoprotein intermediate (Fig. 3). El~P rapidly becomes E2-P when the energy contained originally in the acylphosphoprotein is transduced into the translocation of bound Ca 2§ into the SR ("mario-nette" model, see Fig. 2). Subsequently, the acid-labile intermediate (E2-P) decomposes to enzyme (E 2) and inorganic phosphate. In this model it is assumed that the phosphorylation of the enzyme at Asp 351 triggers a series of conformational changes in which the high affinity Ca 2§ binding sites are disrupted, access to the
A. Properties of SR Ca2§ The major protein in the SR membrane is the Ca2+-ATPase (Mr 100 000), representing about 40% of the total protein in cardiac SR. The cardiac SR Ca2+-ATPase can create intraluminal Ca 2+ concentrations of 5-10 mM. Recombinant DNA studies revealed that the SR or endoplasmic reticulum (ER) Ca2+-ATPase family (SERCA) is the product of at least three alternatively spliced genes, producing a minimum of five different proteins (Burk et al., 1989) (Table 1). SERCA1 is expressed in fast skeletal muscle, and alternative splicing of the 3' end of the primary transcript gives rise to two mRNA forms, which are expressed at different stages of development (Brandl et al., 1986). Alternatively, spliced forms of SERCA2 have been detected in cardiac muscle and slow skeletal muscle (SERCA2a) and in adult smooth muscle and nonmuscle tissues (SERCA2b). SERCA3 is expressed in a selective manner, with the highest mRNA levels in intestine, spleen, lung, uterus, and brain. SERCA2 is about 85% identical to SERCA1, whereas SERCA3 is about 75% identical to either SERCA1 or SERCA2. The human SERCA2 gene is localized on chromosome 12 and maps to position 12q23-q24.1. The proposed general model of the enzyme has three cytoplasmic domains joined to a set of 10 transmembrane helices by a narrow extramembrane pentahelical stalk (Brandl et al., 1986). The cytoplasmic region includes a nucleotide-binding site, or a domain to which the MgATP substrate binds, and a phosphorylation domain, which contains an aspartic acid residue (Asp 351) to which the phosphate is covalently bound (Fig. 2). The third cytoplasmic domain is the/3-strand domain, whose function is still not fully understood. In skeletal muscle, 2 mol of Ca 2§ is transported per mole of ATP hydrolyzed. In cardiac muscle, a similar stoichiometry is expected, but this
FIGURE 2. Model illustrating C a 2+ translocation by SERCA-type Ca2+pumps. In EI---Pconformation, Ca2+binds to the high-affinity binding sites in the cytosol. The energy of the hydrolyzedATP triggers a series of conformational changes and transforms the EI~P intermediate to the E2-P intermediate. These conformational changes are directly coupled to alterations in the orientation of the transmembrane regions, leading to Ca2§ release into the lumen of the sarcoplasmic reticulum.
273
18. Ca2+-ATPases
sites by cytoplasmic Ca 2+ is closed off, and access to the sites from the luminal surface is gained (E 2 conformation) (MacLennan et al., 1992; MacLennan et al., 1997). The CaZ+-free form of the enzyme exists in two different conformational states: one with low affinity for Ca 2+ (E2) and one with high affinity for Ca 2+ (El) (Fig. 3). The conversion of E 2 to E 1 is proposed to be the rate-limiting step in the cycle. Thapsigargin (a plant sesquiterpene lactone) has been shown to interact specifically with the M3 transmembrane segment of the E 2 form of all members of the SR CaZ+-ATPase family and to inhibit enzyme activity even at subnanomolar concentrations (Lytton et al., 1991). The E 1 form of the enzyme has been stabilized and crystallized in the presence of lanthanide (La 3+) or Ca 2+ ions (Dux et al., 1985). On the other hand, vanadate ions in the absence of Ca 2+ induced the formation of Ez-type crystals. El-type crystals consist of single chains of CaZ+-ATPase molecules evenly spaced on the surface of the SR. E2-type crystals consist of dimer chains of ATPase molecules forming an oblique surface lattice. The transition between E 1 and E 2 conformation may involve a shift in the monomer-oligomer equilibrium (Dux et al., 1985). Recently, it has been shown that the cardiac SR CaZ+-ATPase (SERCA2) can be phosphorylated by the CaZ+/CAM-dependent protein kinase at Ser 38 (Toyofuku et al., 1994a). However, the physiological role of this phosphorylation is still not fully understood.
B. Regulation of SR Ca2+-ATPase by Phospholamban 1. Structure of phospholamban In cardiac muscle, slow-twitch skeletal muscle, and smooth muscle, the SR contains the low-molecular-weight protein phospholamban, which can be phosphorylated by various protein kinases. The phosphorylation and dephosphorylation of phospholamban, which comprises 3-4% of the SR membrane protein, regulate the Ca2+-ATPase activity in the SR
Cytosol 2Ca ~§ out
/
V
ADP
E2-P[-~ ~
E2-P~-~
2Ca 2§ in
Lumen of SR FIGURE 3.
Reaction scheme of sarcoplasmic reticular CaZ+-ATPase.
membrane. However, the exact stoichiometry of phospholamban to SR CaZ+-ATPase is not known. In early studies, a stoichiometric relationship of 1 or 2 mol phospholamban per 1 mol of Ca2+-ATPase was proposed for cardiac SR membranes (Colyer and Wang, 1991). In reconstituted systems, a molar ratio of 3:1 of phospholamban-Ca2+-ATPase was necessary to obtain the maximal regulatory effects (Reddy et al., 1995). Moreover, the "functional unit" of phospholamban is not currently clear. Experiments using electron paramagnetic resonance spectroscopy revealed that phospholamban is present primarily as pentamers in the SR membrane (Cornea et al., 1996). In the dephosphorylated form, a substantial fraction of phospholamban monomers (20%) exists, but upon phosphorylation, phospholamban appears to form mainly pentamers, which is due to changes in the isoelectric point (from 10 to 6.7) of the protein (Simmerman and Jones, 1998). The complete amino acid sequence of phospholamban has been determined for various tissues and species. There is currently no evidence for the existence of any isoforms for this protein and the phospholamban gene has been mapped to human chromosome 6 (Fujii et al., 1991). The calculated molecular weight of phospholamban is 6080 (Fujii et al., 1987), and the protein has been proposed to contain two major domains (Fig. 4): a hydrophilic domain (domain I) with two unique phosphorylatable sites (Ser 16 and Thr 17), and a hydrophobic C-terminal domain (domain II) anchored into the SR membrane. The hydrophylic domain (amino acids 1-30) has been further divided into two subdomains: domain Ia (amino acids 1-20) and Ib (amino acids 21-30). Domain Ia has a net positive charge in the dephosphorylated form and consists of an a-helix followed by a Pro residue at position 21 (stalk region). Domain Ib has been suggested to be relatively unstructured (Simmerman and Jones, 1998; Mortishire-Smith et. al., 1995). The hydrophobic domain (amino acids 31-52) forms an a-helix in the SR membrane (Fig. 4). Phospholamban migrates as a 24- to 28-kDa pentamer on SDS gels and dissociates into dimers and monomers upon boiling in SDS before electrophoresis. Spontaneous aggregation of phospholamban into pentamers was also observed upon expression of this protein in bacteria or in mammalian cells. Site-specific mutagenesis experiments identified Cys (Cys 36, Cys 41, and Cys 46), Leu (Leu 37, Leu 44, and Leu 51), and Ile (Ile 40 and Ile 47) residues in the hydrophobic transmembrane domain as essential amino acids for phospholamban pentamer formation (Fujii et al., 1989; Simmerman et al., 1996). The leucine and isoleucine amino acids are suggested to form five zippers in the membrane that stabilize the pentameric form of the protein with a central pore (Fig. 5), defined by the surface of the hydrophobic amino acids (Simmerman et al., 1996). Based on this pentameric selfassociation of phospholamban, a channel function for this protein has been proposed (Wegener et al., 1986). Monoclonal antibodies raised against phospholamban stimulate SR Ca 2+ uptake (Morris et al., 1991). Furthermore, removal of phospholamban from the SR or uncoupling phospholamban from the CaZ+-ATPase (using detergents, high-ionic-strength solutions, or polyanions such as heparin sulfate) markedly increases the affinity of the SR Ca 2+ pump for Ca 2+. These findings suggest that the dephosphorylated form of phospholamban is an inhibitor of the SR CaZ+-ATPase.
274
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
Cytosol NH2 m
,,,,
Ser-16 and Thr-17 - - Pro-21 ------
.. .......
...
.......
.............)!
.
___=~
Leu-37 o
Leu-44 Leu-51
COOH
Lumen of SR FIGURE 4. Molecularmodel of the structure of phospholamban. The cytoplasmic c~-helix (domain Ia, residues 8-20) is interrupted by Pro 21 (heavy circle). Residues 22-32 (domain Ib) are relatively unstructured and may interconvert between transient conformations; residues 33-52 constitute the transmembrane domain II (ce-helix). Ser 16 and Thr 17 (black circles) are the adjacent phosphorylation sites. The shaded circles indicate the leucines (Leu 37, Leu 44, and Leu 51), which are important for the phospholamban subunit interactions (pentamer formation).
This "depression hypothesis" has been confirmed by studies using purified Ca2+-ATPase and purified or recombinant phospholamban in reconstituted systems. Inclusion of phospholamban resulted in inhibition of the SR Ca2+-ATPase activity in reconstituted vesicles or cells (Kim et al., 1990; Reddy et al., 1995; Reddy et al., 1996). Cyclic AMP phosphorylation of phospholamban reversed its inhibitory effect on the Ca 2§ pump. Recently, the inhibitory role of phospholamban on SR and cardiac function has been directly confirmed using transgenic animal models. Overexpression of the protein (phospholamban-overexpressing mice) was associated with inhibition of SR Ca 2§ transport, Ca 2§ transient, and depression of basal left ventricular function (Kadambi et al., 1996). On the other hand, partial (phospholamban-heterozygous mice) or complete ablation of the protein (phospholamban-deficient mice) in mouse models was associated with increases in SR Ca 2§ transport and cardiac function (Luo et al., 1994; Luo et al., 1996). Actually, a close linear correlation between the levels of ph0spholamban and cardiac contractile parameters was observed, indicating that phospholamban is a prominent regulator of myocardial contractility. These findings suggest that changes in the level of this protein will result in parallel changes in SR function and cardiac contraction.
The region of phospholamban interacting with the Ca 2+ATPase may involve amino acids 2-18 (Morris et al., 1991; Toyofuku et al., 1994b). Based on these reports, the simplest model for the interaction between the phospholamban cytoplasmic domain and the SR Ca2+-ATPase is one in which the highly positively charged region of phospholamban (residues 7-16) interacts directly with a negatively charged region on the surface of the Ca2+-ATPase (Lys-Asp-Asp-Lys-Pro-Val 402) to modulate the inhibitory interactions between the two proteins (Fig. 6) (Toyofuku et al., 1994c; Kimura et al., 1997; MacLennan et al., 1998). This association is disrupted by phosphorylation of Ser 16 or Thr 17 in phospholamban, because the positive charges of the phospholamban cytosolic domain are partially neutralized by the phosphate moiety in this vicinity. Phosphorylation of phospholamban by the cAMP-dependent protein kinase at Ser 16 is associated with local unwinding of the a-helix at position 12-16 resulting in conformational changes in the recognition unit of the protein (Mortishire-Smith et al., 1995). Interestingly, phospholamban peptides, corresponding to the hydrophobic membrane-spanning domain, also affect Ca2+-ATPase activity by lowering its affinity for Ca 2+ (Sasaki et al., 1992; Kimura et al., 1996). The importance of the membrane-spanning region of phospholamban in inhibiting SR Ca2+-ATPase activity has been clarified recently by Kimura et al., (1997). It was shown that substitution of the pentamer-stabilizing residues (Leu 37, Leu 44, Leu 51, Ile 40, and lie 47) in the membrane-spanning region (domain II) by Ala resulted in monomeric mutants, which were more effective inhibitors of the SR Ca2+-ATPase activity than wild-type phospholamban. These phospholamban monomeric mutants were called "supershifters" because they decreased the apparent affinity of SR Ca2+-ATPase more effectively than wild-
FIGURE 5. Heptadrepeat model of the transmembrane domain of phospholamban monomer (A) and pentamer (B). (A) Residues 31-52 of monomeric phospholamban are configured as a 3.5 residues/360~ turn helix with positions from a to g of the heptad repeat-squared. The Leu and lie residues constituting zippers are localized to positions a and d, respectively. Darkly shaded circles represent mutations that enhance the inhibitory function of phospholamban by enhanced monomer formation (destabilization of pentamer structure). Lightly shaded circles represent mutations that reduce inhibitory function. These "loss-of-function" mutants are all located on the exterior face of each helix in a pentamer (positions b, e, and f). (B) The phospholamban pentamer model shows the interaction between monomers at positions a and d to form the leucine zipper (heavy squares). (Adapted from Simmerman et al., (1996) and MacLennan et al., (1998).)
18. Ca2+-ATPases
type phospholamban. Thus, it was proposed that monomeric phospholamban is the active form, which is involved in the interaction with SR Ca2+-ATPase. Furthermore, the scanning alanine-mutagenesis studies (Kimura et al., 1997) have identified the amino acid residues in the transmembrane domain of phospholamban (Leu 31, Asn 34, Phe 35, lie 38, Leu 42, lie 48, Va149, and Leu 52), which are associated with loss of function. These amino acids are located on the exterior face of each helix in the pentameric assembly of phospholamban (opposite from the pentamer-stabilizing face) (see Fig. 5). A schematic representation of interaction of phospholamban with SR Ca2+-ATPase is shown in Fig. 6. It has been proposed that the phospholamban monomer is the active species for interaction with the SR Ca2+-ATPase and the pentamers are regarded as functionally inactive forms of phospholamban (Kimura et al., 1997; MacLennan et al., 1998). Phosphorylation of phospholamban monomers promotes association into inactive pentamers. Thus, two important steps for SR Ca2+-ATPase inhibition have been suggested: (1) dissociation of monomeric phospholamban from dephosphorylated pentamers (Kal); and (2) binding of phospholamban monomers to the SR CaZ+-ATPase (Ka2). These dissociation constants (Kal and Ka2) will control both the concentration of phospholamban monomers and the concentration of units in which monomers are associated with the SR Ca2+-ATPase (Kimura et al., 1997; and MacLennan et al., 1998). There are at least two interaction sites between phospholamban and the SR CaZ+-ATPase (Fig. 6): one in the cytoplasmic domains of the two proteins and another one within the transmembrane sequences. The interaction between the hydrophobic membrane-spanning regions is associated with inhibition of the apparent affinity of SR Ca 2+ ATPase for Ca 2+ (Kca). The interaction between the cytosolic phospholamban domain Ia and the SR CaZ+-ATPase modulates the inhibitory interaction in the transmembrane region (domain II) through long-range coupling. Disruption of the cytosolic interactions (domain Ia) by phosphorylation of phospholamban or binding of a phospholamban antibody re-
FIGURE 6. Modelfor regulation of SR Ca2+-ATPaseby phosphorylated and nonphosphorylated phospholamban. Phosphorylation of phospholamban disrupts the interaction between the two proteins so that the inhibition of the Ca2+-ATPaseis relieved. Note that both the cytosolic domain and the membrane-spanningregion of phospholamban are involved in the phosphorylation-mediatedconformational change to relieve the inhibition. Phosphorylationof phospholambanmonomers promotes association into inactive phosphorylatedpentamers.
275
suits in disruption of the inhibitory intramembrane interactions. However, resolution of the exact molecular mechanism by which phospholamban inhibits the SR Ca2+-ATPase Ca 2+ affinity and the concomitant regulation of SR Ca 2§ transport will have to await the development of a new methodology that allows detection of protein-protein interactions in a membrane environment. 2. In Vitro Studies on Regulation of SR Ca2+-ATPase
In the early 1970s, it was suggested that the effects of various catecholamines on cardiac function may be partly attributed to phosphorylation of the SR by cAMP-dependent protein kinase(s). It soon became clear that the substrate for the protein kinase (PK) was not the SR Ca2+-ATPase but phospholamban. Various other high- and low-molecularweight SR proteins were also identified as minor substrates for cAMP-dependent PK, but only the changes in the phosphorylation of phospholamban were associated with functional alterations of the cardiac SR. Cardiac SR membranes contain an endogenous cAMPdependent PK and a Ca2+-CaM-dependent PK that have been shown to phosphorylate phospholamban independently of each other (Kranias,1985a). Phosphorylation by cAMPdependent PK occurred on Ser 16, whereas CaZ+-CaMdependent PK catalyzed exclusively the phosphorylation of Thr 17 (Simmerman et al., 1986). Phosphorylation by either kinase was shown to result in stimulation of the SR Ca 2+ATPase activity and the initial rates of SR Ca 2+ transport. Stimulation was associated with an increase in the apparent affinity of the SR CaZ+-ATPase for Ca 2+ (Kca). In vitro, phospholamban is phosphorylated by two additional PKs: PK-C and a cGMP-dependent PK. Protein kinase C (Ca2+/phospholipid-dependent PK) phosphorylated the protein at a site distinct from those phosphorylated by either cAMP-dependent PK or CaZ+-CaM-dependent PK (Movsesian et a/.,1984). Phosphorylation stimulated the SR CaZ+-ATPase activity and it was postulated that this activity played a role in the action of agents known to stimulate phosphoinositide (PI) hydrolysis, since one product of PI hydrolysis, diacylglycerol, is an activator of PK-C. Cyclic GMP-dependent PK was shown to phosphorylate phospholamban on the same residue (Ser 16) as that phosphorylated by cAMP-dependent PK (Raeymakers et al., 1988). This phosphorylation stimulated cardiac SR Ca 2+ transport, similar to the effects of cAMP-dependent PK. Furthermore, the stimulatory effects on Ca 2+ transport, mediated by cGMPdependent phosphorylation of phospholamban, were also observed in smooth muscle, and this may be of particular interest because some vasodilators act by increasing cGMP levels in vascular smooth muscle. The presence of endogenous PKs in cardiac SR necessitates the presence of phosphoprotein phosphatase(s) for the reversible regulation of the Ca 2+ pump. Protein phosphatases have been generally classified into type 1 and type 2. Type 1 phosphatase is inhibited by nanomolar concentrations of the protein inhibitor-1 and inhibitor-2, whereas type 2 phosphatases are unaffected. In heart muscle, both types of phosphatase have been reported to be present and both can dephosphorylate phospholamban (Kranias and Di Salvo, 1986;
276
SECTION I1 Membrane Potential, Transport Physiology, Pumps, and Exchangers
MacDougall et al., 1991). A type 1 protein phosphatase has been shown to be associated with cardiac SR membranes, and this activity could catalyze the dephosphorylation of both the cAMP-dependent PK and the CaZ+-CaM-dependent PK phosphorylated sites (Ser 16 and Thr 17) on phospholamban (Steenaart et al., 1992). Dephosphorylation was associated with a reduction in the stimulatory effects of PKs on the SR Ca 2+pump (Kranias, 1985b). The SR phosphatase is similar to the skeletal muscle protein phosphatase IG (PPIc), which is composed of a catalytic (C) subunit and a G subunit (MacDougall et al., 1991). The G subunit may become phosphorylated by cAMP-dependent PK, and this causes release of the C subunit from the SR vesicles or glycogen particles into the cytosol, rendering phospholamban in the phosphorylated state and thus capable of stimulating SR Ca 2+ transport. In vivo studies have also shown that the phosphatase activity associated with cardiac SR membranes may be regulated by cAMP-dependent processes. /3Adrenergic stimulation of intact beating hearts was associated with inhibition of the SR phosphatase activity, and this inhibition correlated with increases in the phosphorylation status of inhibitor-1 (Neumann et al.,1991). Thus, regulation of the SR phosphatase activity may be one of the mechanisms by which cells achieve amplification of the cAMP-dependent processes. 3. In Vivo Studies on Regulation of SR Ca2§
The phosphorylation of SR proteins and their regulatory effects on the SR Ca2+-ATPase activity have been studied in perfused hearts from various animal species whose ATP pool was labeled with [32p]orthophosphate. Microsomal fractions enriched in SR were prepared from hearts freeze-clamped during stimulation with different agonists (catecholamines, forskolin, phosphodiesterase inhibitors, phorbol esters) and analyzed by gel electrophoresis and autoradiography for 32p incorporation. /3-adrenergic agonist (isoproterenol) stimulation of the perfused hearts produced an increase in 32p incorporation into phospholamban (Kranias and Solaro, 1982; Lindemann et al., 1983). The stimulation of 32p incorporation into phospholamban was associated with an increased rate of Ca 2+ uptake into SR membrane vesicles and an increased SR Ca2+-ATPase activity (Lindemann et al., 1983; Kranias et al., 1985). These biochemical changes were associated with increases in left ventricular functional parameters (contractility and relaxation). The in vivo phosphorylation of phospholamban was specific only for inotropic agents that increased the cAMP content of the myocardium (fl-adrenergic agonists, forskolin, and phosphodiesterase inhibitors). On the other hand, positive inotropic interventions, which increased the intracellular Ca 2+ level by cAMP-independent mechanisms (a-adrenergic agonists, ouabain, and elevated [Ca2+]), failed to stimulate phospholamban phosphorylation and relaxation. Calmodulin inhibitors (fluphenazine) attenuated the isoproterenol-induced phosphorylation of phospholamban (Lindemann and Watanabe, 1985a), and it was shown that at steady-state isoproterenol exposure, phospholamban contains equimolar amounts of phosphoserine (pSer 16) and phosphothreonine (pThr 17). Phosphorylation of Ser 16 however, correlated most closely with changes in cardiac function in beating hearts (Talosi et al., 1993). Based on these results and recent
findings in transgenic animals (Luo et al., 1997) it is proposed that prevention of Ser 16 phosphorylation (Ser 16 ---)Ala mutation) results in attenuation of the/3-adrenergic response in mammalian hearts, and that phosphorylation of Ser 16 is a prerequisite for Thr 17 phosphorylation. The muscarinic agonist acetylcholine attenuated the increases in cAMP levels, phosphorylation of phospholamban, and the SR Ca2+-ATPase activity produced either by/3adrenergic stimulation or by phosphodiesterase inhibition (using isobutylmethylxanthine) (Lindemann and Watanabe, 1985b). Protein kinase C and cGMP-dependent PK, which have been shown to phosphorylate phospholamban in vitro, failed to demonstrate similar effects in beating guinea pig hearts in response to stimuli that activate PK-C or elevate the cGMP levels (Edes and Kranias, 1990; Huggins et a/.,1989). Thus, the physiological relevance of PK-C and PK-G in beating hearts is not clear at present. The functional alterations in the SR Ca2+-ATPase activity may explain, at least partly, the activating and relaxing effects of/3-adrenergic agents in cardiac muscle (Figs. 7 and 8). The cAMP-dependent phosphorylation of phospholamban under either in vitro or in vivo conditions increases the rate of SR Ca 2+ transport and SR Ca2+-ATPase activity. Such an increase in Ca 2+ transport is expected to contribute primarily to the relaxing effects of catecholamines (Fig. 7). An additional mechanism, which contributes to the increased phosphorylation of phospholamban upon/3-adrenergic stimulation, is the phosphorylation of the phosphatase inhibitor protein by the stimulated cAMP-dependent kinase. This phosphorylation results in inactivation of protein phosphatase 1 and, thus, inhibition of dephosphorylation of phospholamban during the action of catecholamines (Fig. 7). The increased phosphorylation of phospholamban and the increased Ca 2+ levels accumulated by the SR would lead to the availability of higher levels of Ca 2+ to be subsequently released for binding to the contractile proteins (Fig. 7). The critical and prominent role of phospholamban in the mediation of/3-adrenergic functional responses was also confumed in transgenic animal studies. Cardiac myocytes or work-performing heart preparations from phospholamban-deficient mice exhibited largely attenuated responses to/3-adrenergic agonist stimulation (Luo et al., 1996; Wolska et al., 1996), indicating that phospholamban is a key phosphoprotein in the heart's responses to/3-adrenergic agonists. Phosphorylation of other myocardial phosphoproteins has also been suggested to be involved in the mediation of positive inotropic and lusitropic effects of/3-adrenergic agonists. Cyclic AMP-dependent protein kinase-mediated phosphorylation of the al-subunit of the Ca 2+ channel (Fig. 8) is associated with an increase in the voltage-dependent Ca2+ current (/Ca)' which enhances the Ca2+ levels available in the cytosol during/3adrenergic agonist stimulation. Phosphorylation of troponin I has been shown to decrease the sensitivity of myofilaments for Ca ~+ both in intact myocardium and skinned fibers (Kranias et al., 1985). The desensitization of myofibrils is accompanied by an increased off-rate of Ca 2§ from troponin C, which could contribute to faster relaxation (Fig. 8). In addition, phosphorylation of the SR Ca2+-release channel (ryanodine receptor) by Ca2+CaM-dependent protein kinase may stimulate Ca2+release from the SR vesicles and contribute to the elevation of intracellular Ca2+ levels during systole. Thus, the enhanced Ca2§ influx
277
18. Ca2+-ATPases g-adrenergic agent
Protein kinase-A activated
Ic~ channel phosphorylated
Phosphorylation
of inhibitor-1
+
Increased Ca2+ levels in cytosol; Ca>/CAM protein kinase activated
Inhibition of PP1 in SR
I
"-
Phosphorylation of phospholamban
I
..,
Phosphorylation of ryanodine receptor ~ c r e a s ~ d Ca2+ levels in SR
Relaxing effects
I,
Increase in Increased amount " ~ Ca2+available of Ca2§ released by SR for contraction
l
of numerous investigations. It is assumed that in hypo- and hyperthyroid hearts the altered gene expression of the cardiac SR proteins, and hence the changes in the intracellular Ca 2§ transients, is the most important determinant of the altered myocardial function. It was shown that the velocity of ATPdependent Ca 2+ transport and the Ca2+-ATPase activity are specifically increased in SR vesicle preparations from hyperthyroid compared with euthyroid hearts (Beekman et al., 1989). Opposite changes were noted for hypothyroid animals compared with euthyroid ones (Beekman et al., 1989). Examination of the steady-state mRNA levels of the cardiac SR Ca2+-ATPase and the ryanodine receptor revealed a significant increase (140-190%) in hyperthyroid and a marked decline (40-50%) in hypothyroid animals (Arai et al., 1991). The changes in mRNA levels for the Ca2+-ATPase in hypothyroid and hyperthyroid conditions also reflected changes in the protein amounts of the enzyme in these hearts (Kiss et al., 1994; Kiss et al., 1998). Interestingly, in the case of phospholamban, the regulator of the Ca2+-ATPase, and calsequestrin, there was no coordinated regulation with respect to the Ca2+-ATPase. In fact, both the relative mRNA level and the protein content of phospholamban were reported to decrease in hyperthyroid animals, whereas there was no change noted in the calsequestrin mRNA level upon L-thyroxine treatment. In hypothyroid hearts an opposite trend
Activating effects FIGURE 7. Schematicdiagram of possible relaxing and activating effects of fl-adrenergic agents in the heart. PP1, protein phosphatase 1.
across the sarcolemma, together with the increased Ca 2+ levels to be released from the SR, may result in an elevation of the Ca 2+ available for the contractile machinery, leading to an increase in the amplitude of contraction (Fig. 8). C. SR Ca2§
in Cardiac D i s e a s e s
The complex regulation of the SR function clearly indicates that even small disturbances in SR Ca 2§ handling may result in profound changes and deterioration of normal myocardial function. The fast removal of Ca 2§ by the SR Ca2+-ATPase during diastole and the subsequent rapid release through the SR Ca 2§ channel (ryanodine receptor) at the beginning of contraction are prerequisites for normal diastolic and systolic function. We briefly outline in the next section the alterations in the SR Cae§ in the major cardiac diseases.
1. SR Ca2+ -ATPase in Hyperthyroidism and Hypothyroidism Thyroid hormones are important regulators of myocardial contractility and relaxation. Chronic increases in thyroid hormone levels lead to cardiac hypertrophy, with increases in the heart rate and cardiac output as well as left ventricular contractility and velocity of relaxation. On the other hand, opposite effects are associated with a hypothyroid condition. The mechanisms underlying these changes have been the subject
FIGURE 8. Effectsof fl-adrenergic agents on protein phosphorylation in cardiac cells. Increased intracellular cAMP levels activate the cAMP-dependent protein kinase(s), which phosphorylates various proteins (phospholamban, inhibitor-1, Ca2+-channel and myofibrillar proteins) and increases the rates of SR Ca2+uptake and release.
278
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
was noted since the protein amount of phospholamban was found to be increased as compared to the euthyroid or hyperthyroid animals (Kiss et al., 1994; Kiss et al., 1998). Consequently, the phospholamban-CaZ+-ATPase protein ratio was highest in the hypothyroid animals, followed by euthyroid and hyperthyroid animals. These changes in the phospholamban-CaZ+-ATPase ratio were associated with coordinate alterations in the SR Ca 2+ uptake, affinity of the SERCA2 for Ca 2+, and myocardial function (Kiss et al., 1994; Kimura et al., 1994). These changes indicate that the SR proteins responsible for Ca 2+ uptake and release (CaZ+-ATPase and ryanodine receptor) are coordinately regulated in hypothyroid and hyperthyroid hearts and provide a simple explanation for the altered Ca 2+ release and reuptake capacity and hence the myocardial function under these conditions.
2. SR Ca2+-ATPase in Cardiomyopathies Dilated cardiomyopathy is a frequent form of cardiac muscle disease and is characterized by an impaired systolic function and dilatation of both ventricles (systolic pump failure). In various animal models of primary and secondary dilated cardiomyopathy, it was shown that both SR Ca 2+ binding capacity and uptake were depressed because of the decreased activity and protein level of the SR CaZ+-ATPase (Edes et al., 1991). In some studies of human idiopathic dilated cardiomyopathy, decreases were noted for both SR Ca 2+ uptake rates and Ca 2+ATPase activity (Limas et al., 1987; Unverferth et al., 1988) as well as myocardial Ca 2+ handling (Gwathmey et al., 1987; Beuckelmann, et al., 1992). Examination of the mRNA levels in left ventricular biopsies from patients with dilated cardiomyopathy revealed a significant decrease in mRNA content for the SR Ca2+-ATPase relative to other mRNA forms (Mercadier et al., 1990; Arai et al., 1993). In contrast, other authors were unable to detect a decrease in SR Ca 2+ uptake activity (Movsesian et al., 1989) or the immunodetectable levels of the SR Ca2+-ATPase protein (Schwinger et al., 1995) in the left ventricular myocardium from patients with idiopathic dilated cardiomyopathy. Furthermore, the gating mechanism of the SR CaZ+-release channel was recently reported to be abnormal in dilated cardiomyopathy (D'Agnolo et al., 1992), and it was suggested that defective excitation-contraction coupling is involved in the pathogenesis of this disease. Another type of cardiomyopathy, hypertrophic cardiomyopathy, has only been recognized in clinical practice for the last three decades. The characteristics of this disease are asymmetric interventricular septal hypertrophy and narrowing of the left ventricular outflow tract, with or without outflow obstruction (outflow tract pressure gradient). In the familial form of hypertrophic cardiomyopathy, which accounts for about 60% of all cases, mutations in myofibrillar protein genes (/3-myosin heavy chain, troponin T, atropomyosin, and C protein) are associated with the disease (Schwartz and Mercadier 1996). Additionally, prolongation of the Ca 2+ transient, abnormal Ca 2+ handling, and a decline in SR CaZ+-ATPase mRNA levels are reported to be characteristic for human hypertrophic cardiomyopathy (Gwathmey et al., 1987; Mercadier et a/.,1990), which may explain the diastolic function impairment in this disease.
In chronic heart failure due to hemodynamic overload, irrespective of the specific etiology (valvular heart disease, cardiomyopathy, chronic ischemic heart disease, or hypertension), a reduction was observed in both the number and the activity of the SR Ca2+ pump (Limas et al., 1987; Unverferth et al., 1988). Furthermore, a close correlation was obtained between the SR CaZ+-ATPase mRNA or protein levels and the myocardial function (Mercadier et al., 1990; Hasenfuss et al., 1994). Interestingly, the Na+-Ca2+ exchanger gene expression was reported to be increased in failing human hearts and it was hypothesized that the upregulation of this protein may compensate for the depressed SR function (Studer et al., 1994).
3. SR Ca2+-ATPase in Ischemia A brief period of ischemia (10-20 rain) induces reversible tissue damage in cardiac muscle, resulting in a "stunned" myocardium. This condition is characterized by regional contractile abnormalities (declines in both systolic and diastolic function) that persist for several hours despite the absence of necrosis. These hemodynamic changes are associated with a reduction in SR Ca 2+ transport (Limbruno et al., 1989; Krause et al., 1989). The maximal activity of the SR Ca2+-ATPase was found to be depressed, and the Ca 2+ sensitivity of this enzyme was decreased (Krause et al., 1989). Furthermore, a decrease in the coupling ratio (tool Ca2+/mol ATP) was observed in the SR membranes isolated from the stunned myocardium, which was suggested to be the result of an increase in the Ca 2+ permeability of the SR membrane. The SR CaZ+-release process was also found to be impaired in the stunned myocardium due to a reduction of the number of ryanodine receptors (Zucchi et al., 1994). These data suggest that complex modifications of the SR function occur in the stunned myocardium, which are at least partly responsible for the contractile impairment found in this condition. In long-lasting myocardial ischemia, gradual declines in both SR CaZ+-ATPase activity and Ca 2+ uptake were found, which may be due to degradation of the SR CaZ+-ATPase (Akiyama et al., 1986; Schoutsen et al., 1989). Ischemia was also shown to result in a gradual decrease in the phosphorylation status of phospholamban under both in vitro (Schoutsen et al., 1989; Lamers et al., 1986) and in vivo conditions (Bartel et al., 1989), and this correlated with a decrease in the SR Ca2+-ATPase activity. Thus, it has been postulated that the long-lasting ischemia-induced progressive inactivation of the SR Ca 2+pump not only is a consequence of the specific loss of enzyme activity, but may also be related to the altered characteristics of phospholamban (Schoutsen et al., 1989). A combination of various pathogenic factors has been suggested to be responsible for the reduced SR function and the final tissue necrosis in the ischemic myocardium. These pathological factors include pH reduction (acidosis), activation of intracellular proteolytic enzymes, and increased generation of free radicals.
!!!. Other ATPases The Ca 2+ regulation in eukaryotic cells involves a complex mechanism that maintains a low background Ca 2+ concentration (usually 0.1-0.2 I~M) in the cell interior. Eukaryotic cells generally satisfy their Ca 2+ demands by ex-
18. Ca2+-ATPases tracting Ca 2+ from their own internal stores, but it is also evident that the long-term regulation of the Ca 2+ gradient across the plasma membrane is a result of the concerted operation of the importing (Ca 2+ channel) and exporting (SR Ca 2+ pump; Na+-CaZ+exchanger and Ca 2+ pump of the surface membrane) Ca 2+ systems (see Section IIA). The plasma membrane CaZ+-ATPase is a low-capacity system possessing a very high Ca 2+ affinity, which enables the enzyme to interact with Ca 2+ at low intracellular concentrations. Consequently, its function is continuous and presumably satisfies the fine tuning of Ca 2+ homeostasis.
A. General Properties of Plasma Membrane Ca2+-ATPase(s) The plasma membrane Ca2+-ATPase (molecular mass of 140 kDa) general kinetic mechanism follows the pattern of the SR Ca2§ ATP phosphorylates an Asp residue to yield an acid-stable phosphorylated intermediate. The elementary steps of the cycle are probably similar in both SR and plasma membrane Ca2+-ATPases (Schatzmann, 1989). The stoichiometry between transported Ca 2+ and hydrolyzed ATP is only 1.0 for the plasma membrane Ca 2+ pump. The administration of La 3+ under various experimental conditions has been associated with an increase in the steady-state phosphoenzyme level of the plasma membrane CaZ+-ATPase, and this increase possibly results from stabilization of the aspartyl phosphate (inhibition of hydrolysis of the phosphate group). The other classic inhibitor of Ca 2+ pumps, vanadate, has been found to be a potent inhibitor of the plasma membrane CaZ+-ATPase even at low concentrations (Bond and Hudgins, 1979). Calmodulin stimulates the plasma membrane CaZ+-ATPase by direct interaction with the enzyme. It has been shown that the stimulation results from a combined effect on the affinity for Ca 2+ (Kin) and the maximal transport rate (Vm~) (Carafoli, 1992). The calmodulin-binding domain of the Ca~+ pump has been suggested to function as a repressor of the enzymatic activity (autoinhibitory function), and calmodulin may relieve this inhibition (Carafoli, 1992). A common mechanism in the autoinhibition of plasma membrane CaZ+-ATPase and phospholamban inhibition of SR CaZ+-ATPase has been suggested. In both proteins, the interacting sites are amphiphilic and located in the cytoplasmic region. The interaction occurs with homologous regions in the SR and plasma membrane Ca 2+ATPases close to the phosphorylation sites. In the absence of calmodulin, the plasma membrane Ca 2+pump can be activated by several other compounds. Polyunsaturated fatty acids and acidic phospholipids (phosphatidylinositol, phosphatidylinositol 4-phosphate, and phosphatidylinositol 4,5-diphosphate) have been reported to be good activators and, since they are present in the plasma membrane, they may be important regulators of the CaZ+-ATPase under in vivo conditions (Carafoli, 1991). Phosphorylation of the enzyme by cAMP-dependent PK or PK-C has also been reported to stimulate plasma membrane CaZ+-ATPase activity. The cAMP-dependent phosphorylation occurs C-terminally to the calmodulin-binding domain and the phosphorylation-mediated activation may likewise be significant in vivo. The PK-C phosphorylation occurs in the calmodulin-binding domain, inhibiting the binding of calmodulin to the plasma membrane CaZ+-ATPase and lowering the
279 autoinhibitory potential of this domain (Hofmann et al., 1994). Cyclic GMP--dependent PK has also been reported to stimulate the plasma membrane Ca 2+ pump in vascular smooth muscle, but the Ca2+-ATPase enzyme was not found to be the substrate for this kinase.
B. Primary Structure and Topography of Plasma Membrane Ca2+-ATPase(s) The complete amino acid sequence of the plasma membrane Ca 2+ pump has been deduced from rat and human cells (Shull and Greeb, 1988; Verma et al., 1988). It appears that the plasma membrane CaZ+-ATPase (PMCA) isoforms are encoded by a multigene family, and additional variability is produced by alternative RNA splicing of each gene transcript (Table 2). The regions important for the catalytic function and the transmembrane domains are highly conserved, with no observed diversity. The isoform diversity seems to alter primarily the regulatory characteristics of the enzyme and it can be regarded as an adaptation to tissue specificity. The secondary structure of the plasma membrane Ca 2+ATPase is similar to that of the SR CaZ+-ATPase (Shull and Greeb, 1988). The enzyme contains 10 putative transmembrane helices, which are connected on the outside of the plasma membrane by short loops. Three primary domains (about 80% of the pump protein) protrude into the cytoplasm. The first domain corresponds to the transducing unit, which couples ATP hydrolysis to Ca 2§ translocation. The second protruding domain contains the aspartyl phosphate site (phosphorylation domain). The C-terminal portion of this domain can also be labeled by ATP analogs and contains a "hinge" region that permits the movement of aspartyl phosphate and the ATP-binding site. The third C-terminal protruding domain contains the calmodulin-binding sequence and the phosphorylation sites for protein kinase C and cAMP-dependent protein kinase. The latter is not present in all isoforms. As in the SR Ca2+-ATPase, selective mutations in the M4 and M6 transmembrane segments in the plasma membrane Ca2+-ATPase have been associated with loss of its ability to form the ATP- and Ca2+-dependent phosphorylated intermediate and to transport Ca 2+ (Guerini et al., 1998).
IV. S u m m a r y The Ca 2+ levels in muscle are primarily regulated by the sarcoplasmic reticulum (SR) network, which serves as a sink for Ca 2+ ions during relaxation and as a Ca 2+ source during contraction. In cardiac muscle, most of the intracellular Ca 2+ released during systole is taken up by the SR through its Ca2+-ATPase. This translocation of Ca 2+ from the cytosol into the SR lumen uses ATP as the energy source, and is characterized by the formation of a phosphorylated intermediate (E1NP) for the CaZ+-ATPase. In cardiac muscle, slow-twitch skeletal muscle, and smooth muscle, the CaZ+-ATPase is regulated by a lowmolecular-weight phosphoprotein called phospholamban. In its dephosphorylated form, phospholamban is an inhibitor of the CaZ+-ATPase and phosphorylation relieves this inhibition. Phosphorylation of phospholamban occurs by cAMPdependent, cGMP-dependent, CaZ+-calmodulin-dependent,
280
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
TABLE 2 Distribution of the Human Plasma Membrane Ca2+-ATPase (PMCA) Isoforms Gene
Tissue distribution
Level of expression
Inhibition by calmodulin
PMCA 1a
Ubiquitous
High
Medium sensitive
PMCA2
Restricted (brain high)
High
Highly sensitive
PMCA3
Restricted (brain low)
Low
N/Ab
PMCA4
Ubiquitous
Medium
Medium sensitive
apMCA numbers identify different gene products. bN/A = data not available.
and Ca2+-phospholipid--dependent protein kinases in vitro. However, in vivo studies have indicated that phospholamban is phosphorylated only by cAMP-dependent and Ca 2+calmodulin-dependent protein kinases in intact beating hearts. A phospholamban phosphatase activity has been reported to be present in SR membranes, which can dephosphorylate this regulatory protein and reverse its stimulatory effects on the CaZ+-ATPase. Alterations in the SR CaZ+-ATPase activity and its regulation by phospholamban have been shown to occur in cardiac diseases such as hypothyroidism, hyperthyroidism, hypertrophy, heart failure, and ischemia. In most instances, alterations in CaZ+-ATPase activity correlated with alterations in its mRNA levels and ventricular function. Another CaZ+-ATPase, which is also important for maintaining Ca 2+ homeostasis in muscle, is the plasma membrane Ca2+-ATPase. This enzyme transports Ca 2+ to the extracellular space and uses ATP as its energy source, similar to the SR CaZ+-ATPase. The plasmalemmal CaZ+-ATPase may be distinguished from the SR CaZ+-ATPase primarily by its distinct sensitivity to La 3+, vanadate, and calmodulin. The primary structure of the various Ca 2+ pumps has been published, and there is a growing interest in further use of molecular biological approaches and specifically sitedirected mutagenesis for these enzymes to obtain more information about their structural-functional relationships. The ultimate question is, what is the precise mechanism by which Ca 2+ is transported across the ATPases? Site-directed mutagenesis studies and construction of molecular models have already given some information along these lines and hopefully will provide further data that will finally answer this question. Furthermore, in the absence of appropriate crystallographic data, a deeper understanding of the mechanisms that regulate the Ca2+-ATPases under normal and pathological conditions may elucidate the structuralfunctional relationships in these enzymes and their role in maintaining Ca 2§ homeostasis in the cell.
Acknowledgments This study was supported by National Institutes of Health Grants HL-26057 and P40RR12358 and the Hungarian Academy of Sciences Grant OTKA T 020691.
Bibliography Akiyama, K., Konno, N., Yanagishita, T., Tanno, F., and Katagiri, T. (1986). Ultrastructural changes in the sarcoplasmic reticulum in acute myocardial ischemia. Jpn. Circ. J. 50, 829-838. Andersen J. P., and Vilsen, B. (1998). Structure-function relationships of the calcium binding sites of the sarcoplasmic reticulum Ca2§ ATPase. Acta Physiol. Scan. 163, 45-54. Arai, M., Otsu, K., MacLennan, D. H., Alpert, N. R., and Periasamy, M. (1991). Effect of thyroid hormone on the expression of mRNA encoding sarcoplasmic reticular proteins. Circ. Res. 69, 266-276. Arai, M., Alpert, N. R., MacLennan, D. H., Barton, P., and Periasamy, M. (1993). Alterations in sarcoplasmic reticulum gene expression in human heart failure. A possible mechanism for alterations in systolic and diastolic properties of the failing myocardium. Circ. Res. 72, 463-469. Bartel, S., Karczewski, P., and Krause, E-G. (1989). Phosphorylation of phospholamban and troponin I in the ischemic and reperfused heart: attenuation and restoration of isoprenaline responsiveness. Biomed. Biochim. Acta 48, 108-113. Beekman, R. I., Hardeveld, C., and Simonides, W. S. (1989). On the mechanism of the reduction by thyroid hormone of/3-adrenergic relaxation rate stimulation in rat heart. Biochem. J. 259, 229-236. Bers, D. M. (1991). Ca regulation in cardiac muscle. Med. Sci. Sports Exer. 23, 1157-1162. Beuckelmann, D. J., N~ibauer, Erdmann, E. (1992). Intracellular calcium handling in isolated ventricular myocytes from patients with terminal heart failure. Circulation 85, 1046-1055. Bond, G. H., and Hudgins, P. (1979). Kinetics of inhibition of Na§ K§ ATPase by Mg2§ K§ and vanadate. Biochemistry 18, 325-331. Brandl, C. J., Green, N. M., Korczak, B., and MacLennan, D. H. (1986). Two ATPase genes: homologies and mechanistic implications of deduced amino acid sequences. Cell 44, 597-607. Burk, S. E., Lytton, J., MacLennan, D. H., and Shull, G. E. (1989). cDNA cloning, functional expression and mRNA tissue distribution of a third organellar Ca2§ pump. J. Biol. Chem. 264, 18 561-18 568. Carafoli, E. (1991). The calcium pumping ATPase of the plasma membrane. Annu. Rev. Physiol. 53, 531-547. Carafoli, E. (1992). The Ca2+ pump of the plasma membrane. J. Biol. Chem. 267, 2115-2118. Colyer, J., and Wang, J. H. (1991). Dependence of cardiac sarcoplasmic reticulum calcium pump activity on the phosphorylation status of phospholamban. J. Biol. Chem. 266, 17 486-17 493. Cornea, R. L., Jones, L. R., Autry, J. M., and Thomas, D. D. (1997). Mutation and phosphorylation change the oligomeric structure of phospholamban in lipid bilayers. Biochemistry 36, 2960-2967. D'Agnolo, A., Luciani, G. B., Mazzucco, A., Gallucci, V., and Salviati, G. (1992). Contractile properties and Ca2+release activity of the sarcoplasmic reticulum in dilated cardiomyopathy. Circulation 85, 518-525.
18. C
a
2
+
-
A
T
P
Dux, L., Taylor, K. A., Tin-Beall, H. P., and Martonosi, A. (1985). Crystallization of the CaZ+-ATPase of sarcoplasmic reticulum by calcium and lanthanide ions. J. Biol. Chem. 260, 11 730-11 743. Edes, I., and Kranias, E. G. (1990). Phospholamban and troponin I are substrates for protein kinase C in vitro but not in intact beating guinea pig hearts. Circ. Res. 67, 394-400. Edes, I., Talosi, L., and Kranias, E. G. (1991). Sarcoplasmic reticulum function in normal heart and in cardiac disease. Heart Failure 6 (6) 221-237. Fujii, J., Ueno, A., Kitano, K., Tanaka, S., Kadoma, M., and Tada, M. (1987). Complete complementary DNA-derived amino acid sequence of canine cardiac phospholamban. J. Clin. Invest. 79, 301-304. Fujii, J., Maruyama, K., Tada, M., and MacLennan, D. H. (1989). Expression and site-specific mutagenesis of phospholamban. Studies of residues involved in phosphorylation and pentameric formation. J. Biol. Chem. 264, 12 950-12 955. Fujii, J., Zarain-Herzberg, A., Willard, H. F., Tada, M., and MacLennan, D. H. (1991). Structure of the rabbit phospholamban gene, cloning of the human cDNA, and assignment of the gene to human chromosome 6. J. Biol. Chem. 266, 11 669-11 675. Guerini, D., Garcia-Martin, E., Zecca, A., Guidi, E, Carafoli, and E. (1998). The calcium pump of the plasma membrane: membrane targeting, calcium binding sites, tissue-specific isoform expression. Acta Physiol. Scand. Suppl. 643, 265-273. Gwathmey, J. K., Copelas, L., MacKinnon, R., Schoen, E J., Feldman, M. D., Grossman, W., and Morgan, J. P. (1987). Abnormal intracellular calcium handling in myocardium from patients with end-stage heart failure. Circ. Res. 61, 70-76. Hasenfuss, G., Reinecke, H., Studer, R., Meyer, M., Pieske, B., Holtz, J., Holubarsch, C., Posival, H., Just, H., Drexler, H. (1994). Relation between myocardial function and expression of sarcoplasmic reticulum Ca2+-ATPase in failing and nonfailing human myocardium. Circ. Res. 75, 434--442. Hofmann, F., Anagli, J., and Carafoli, E. (1994). Phosphorylation of the calmodulin binding domain of the plasma membrane Ca 2+ pump by protein kinase C reduces its interaction with calmodulin and with its pump receptor site. J. Biol. Chem. 269, 24 298-24 303. Huggins, J. P., Cook, E. A., Pigott, J. R., Mattinsley, T. J., and England, P. J. (1989). Phospholamban is a good substrate for cGMP-dependent protein kinase in vitro, but not in intact cardiac or smooth muscle. Biochem. J. 260, 829-835. Kadambi, V. J., Ponniah, S., Harrer, J. M., Hoit, B. D., Dorn, G. W., Walsh, R. A., and Kranias, E. G. (1996). Cardiac-specific overexpression of phospholamban alters calcium kinetics and resultant cardiomyocyte mechanics in transgenic mice. Clin. Invest. 97, 533-539. Kim, H. W., Steenaart, N. A. E., Ferguson, D. G., and Kranias, E. G. (1990). Functional reconstitution of the cardiac sarcoplasmic reticulum Ca2+-ATPase with phospholamban in phospholipid vesicles. J. Biol. Chem. 265, 1702-1709. Kimura, Y., Otsu, K., Nishida, K., Kuzuya, T., and Tada, M. (1994). Thyroid hormone enhances Ca2+ pumping activity of the cardiac sarcoplasmic reticulum by increasing Ca2+ ATPase and decreasing phospholamban expression. J. Mol. Cell. Cardiol. 26, 1145-1154. Kimura, Y., Kurzydlowski, K., Tada, M., and MacLennan, D. H. (1996). Phospholamban regulates the Ca2+-ATPase through intramembrane interaction. J. Biol. Chem. 271, 21 726-21 731. Kimura, Y., Kurzydlowski, K., Tada, M., and MacLennan, D. H. (1997). Phospholamban inhibitory function is activated by depolymerization. J. Biol. Chem. 272, 15 061-15 064. Kiss, E., Jakab, G., Kranias, E. G., and Edes, I. (1994). Thyroid hormone-induced alterations in phospholamban protein expression. Regulatory effects on sarcoplasmic reticulum Ca 2§ transport and myocardial relaxation. Circ. Res. 75, 245-251. Kiss, E., Brittsan, A. E., Edes, I., Grupp, I. L., Grupp, G., and Kranias, E. G. (1998). Thyroid hormone-induced alterations in phospholambandeficient mouse hearts. Circ. Res. 83, 608--613.
a
s
e
s
2
8
1
Kranias, E. G., and Solaro, R. J. (1982). Phosphorylation of troponin I and phospholamban during catecholamine stimulation of rabbit heart. Nature 298, 182-184. Kranias, E. G. (I985a). Regulation of Ca2+transport by cyclic 3',5'-AMPdependent and calcium-calmodulin-dependent phosphorylation of cardiac sarcoplasmic reticulum. Biochim. Biophys. Acta 844, 193-199. Kranias, E. G. (1985b). Regulation of calcium transport by protein phosphatase activity associated with cardiac sarcoplasmic reticulum. J. Biol. Chem. 260, 11 006-11 010. Kranias, E. G., Garvey, J. L., Srivastava, R. D., and Solaro, R. J. (1985). Phosphorylation and functional modifications of sarcoplasmic reticulum and myofibrils in isolated rabbit hearts stimulated with isoprenaline. Biochem. J. 226, 113-121. Kranias, E. G., and Di Salvo, J. (1986). A phospholamban protein phosphatase activity associated with cardiac sarcoplasmic reticulum. J. Biol. Chem. 261, 10 029-10 032. Krause, S. M., Jacobus, W. E., and Becker, L. C. (1989). Alterations in cardiac sarcoplasmic reticulum calcium transport in the postischemic "stunned" myocardium. Circ. Res. 65, 526-530. Lamers, J. M., De Jonge-Stinis, J. T., Hiilsman, W. C., and Verdouw, E D. (1986). Reduced in vitro 32 P-incorporation into phospholambanlike protein of sarcolemma due to myocardial ischaemia in anaesthetized pigs. J. Mol. Cell. Cardiol. 18, 115-125. Limas, C. J., Olivari, M. T., Goldenberg. I. E, Levine, T. B., Benditt, D. G., and Simon, A. (1987). Calcium uptake by cardiac sarcoplasmic reticulum in human dilated cardiomyopathy. Cardiovasc. Res. 21, 601-605. Limbruno, U., Zucchi, R., Ronca-Testoni, S., Galbani, E, Ronca, G., and Mariani, M. (1989). Sarcoplasmic reticulum function in the "stunned" myocardium. J. Mol. Cell. Cardiol. 21, 1063-1072. Lindemann, J. E, Jones, L. R., Hathaway, D. R., Henry, B. G., and Watanabe, A. M. (1983)./3-Adrenergic stimulation of phospholamban phosphorylation and CaZ+-ATPase activity in guinea pig ventricles. J. Biol. Chem. 260, 4516--4525. Lindemann, J. E, and Watanabe, A. M. (1985a). Phosphorylation of phospholamban in intact myocardium. Role of CaZ+-calmodulindependent mechanisms. J. Biol. Chem. 260, 4516-4525. Lindemann, J. P., and Watanabe, A. M. (1985b). Muscarinic cholinergic inhibition of/3-adrenergic stimulation of phospholamban phosphorylation and Ca 2+ transport in guinea pig ventricles. J. Biol. Chem. 260, 122-133. Luo, W., Grupp, I. L., Harrer, J., Ponniah, S., Grupp, G., Duffy J. J., Doetschman, T., and Kranias, E. G. (1994). Targeted ablation of phospholamban gene is associated with markedly enhanced myocardial contractility and loss of/3-adrenergic stimulation. Circ. Res. 75, 401--409. Luo, W., Wolska, B. M., Grupp, I. L., Harrer, J. M., Haghighi, K., Ferguson, D. G., Slack, J. P., Grupp, G., Doetschman, T., Solaro, R. J., and Kranias, E. G. (1996). Phospholamban gene dosage effect in the mammalian heart. Circ. Res. 78, 839-847. Luo, W., Chu, G., Sato, Y., Zhou, Z., Kadambi, V., and Kranias, E. G. (1997). Transgenic approaches to define the functional role of dual site phosphorylation of phospholamban. J. Biol. Chem. 273, 4734-4739. Lytton, J., Westlin, M., and Hanley, M. R. (1991). Thapsigargin inhibits the sarcoplasmic or endoplasmic reticulum Ca-ATPase family of calcium pump. J. Biol. Chem. 266, 17 067-17 071. MacDougall, L. K., Jones, J. R., and Cohen, P. (1991). Identification of the major protein phosphatases in mammalian cardiac muscle which dephosphorylate phospholamban. Eur. J. Biochem. 196, 725-734. MacLennan, D. H., Clarke, D. M., Loo, T. W., and Skerjanc, I. S. (1992). Site-directed mutagenesis of the Ca2+-ATPase of sarcoplasmic reticulum. Acta Physiol. Scand. 146, 141-150. MacLennan, D. H., Rice, W. J., and Green N. M. (1997). The mechanism of Ca 2+ transport by sarco(endo)plasmic reticulum Ca 2+ATPases. J. Biol. Chem. 272, 28 815-28 818.
2112
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
MacLennan, D. H., Kimura, Y., and Toyofuku, T. (1998). Sites of regulatory interaction between Ca2§ and phospholamban. Ann. NY Acad. Sci. 175, 31-42. Mercadier, J. J., Lompre, A. M., Duc, P., Boheler, K. R., Fraysse, J. B., Wisnewsky, P., Allen, P. D., Komajda, M., and Schwartz, K. (1990). Altered sarcoplasmic reticulum Ca2+-ATPase gene expression in the human ventricle during end-stage heart failure. J. Clin. Invest. 85, 305-309. Morris, G. L., Cheng, H., Colyer, J., and Wang, J. H. (1991). Phospholamban regulation of cardiac sarcoplasmic reticulum (Ca 2§ Mg2+)-ATPase. Mechanism of regulation and site of monoclonal antibody interaction. J. Biol. Chem. 266, 11 270-11 275. Mortishire-Smith, R. J., Broughton, H., Garsky, V. M., Mayer, E. J., and Johnson, R. G. (1998). Structural studies on phospholamban and implications for regulation of the CaZ+-ATPase. Ann. NY Acad. Sci. 853, 63-78. Movsesian, M. A., Nishikawa, M., and Adelstein, R. S. (1984). Phosphorylation of phospholamban by calcium-activated, phospholipid-dependent protein kinase. J. Biol. Chem. 259, 8029-8032. Movsesian, M. A., Bristow, M. R., and Krall, J. (1989). Ca 2+ uptake by cardiac sarcoplasmic reticulum from patients with idiopathic dilated cardiomyopathy. Circ. Res. 65, 1141-1144. Neumann, J., Gupta, R. C., Schmitz, W., Scholz, H., Nairn, A. C., and Watanabe, A. M. (1991). Evidence for isoproterenol-induced phosphorylation of phosphatase inhibitor- 1 in the intact heart. Circ. Res. 69, 1450-1457. Raeymakers, L., Hofmann, E, and Casteels, R. (1988). Cyclic GMPdependent protein kinase phosphorylates phospholamban in isolated sarcoplasmic reticulum from cardiac and smooth muscle. Biochem. J. 252, 269-273. Reddy, L. G., Jones, L. R., Cala, S. E., O'Brian, J. J., Tatulian, S A., and Stokes, D. L. (1995). Functional reconstitution of recombinant phospholamban with rabbit skeletal Ca2+-ATPase. J. Biol. Chem. 270, 9390-9397. Reddy, L. G., Jones, L. R., Pace, R. C., and Stokes, D. L. (1996). Purified, reconstituted cardiac Ca2+-ATPase is regulated by phospholamban but not by direct phosphorylation with Ca2§ ulin-dependent protein kinase. J. Biol. Chem. 271, 14 964-14 970. Ringer, S. A. (1883). A further contribution regarding the influence of different constituents of the blood on the contraction of the heart. J. Physiol. 4, 29-42. Sasaki, T., Inui, M., Kimura, Y., Kuzuya, T., and Tada, M. (1992). Molecular mechanism of regulation of Ca2§ by phospholamban in cardiac sarcoplasmic reticulum. J. Biol. Chem. 267, 1674-1679. Schatzmann, H. J. (1989). The calcium pump of the surface membrane and of the sarcoplasmic reticulum. Annu. Rev. Physiol. 51, 473-485. Schoutsen, B., Blom. J. J., Verdouw, P. D., and Lamers, J. M. (1989). Calcium transport and phospholamban in sarcoplasmic reticulum of ischemic myocardium. J. Mol. Cell. Cardiol. 21, 719-727. Schwartz, K., and Mercadier, J-J. (1996). Molecular and cellular biology of heart failure. Curr. Opinion Cardiol. 11, 227-236. Schwinger, R. H. G., Brhm, M., Schmidt, U., Karczewski, P., Bavendiek, U., Flesch, M., Krause, E-G., and Erdmann, E. (1995). Unchanged protein levels of SERCA II and phospholamban but reduced Ca 2+ uptake and Ca2+-ATPase activity of cardiac sarcoplasmic reticulum from dilated cardiomyopathy patients compared with patients with nonfailing hearts. Circulation 92, 3220-3228.
Shull, G. E., and Greeb, J. (1988). Molecular cloning of two isoforms of the plasma membrane Ca 2§ transporting ATPase from rat brain. Structural and functional domains exhibit similarity to Na+,K§ other cation transport ATPases. J. Biol. Chem. 263, 8646-8657. Simmerman, H. K. B., Collins, J. H., Theibert, J. L., Wegener, A. D., and Jones, L. R. (1986). Sequence analysis of phospholamban. Identification of phosphorylation sites and two major structural domains. J. Biol. Chem. 261, 13 333-13 341. Simmerman, H. K. B., Kobayashi, Y. M., Autry, J. M., and Jones, L. R. (1996). A leucine zipper stabilizes the pentameric membrane domain of phospholamban and forms a coiled-coil pore structure. J. Biol. Chem. 271, 5941-5946. Simmerman, H. K. B., and Jones, L. R. (1998) Phospholamban: protein structure, mechanism of action, and role in cardiac function. Physiol. Rev. 78, 921-947. Steenaart, N. A. E., Ganim, J. R., DiSalvo, J., and Kranias, E. G. (1992). The phospholamban phosphatase associated with cardiac sarcoplasmic reticulum is a type 1 enzyme. Arch. Biochem. Biophys. 293, 17-24. Studer, R., Reinecke, H., Bilger, J., Eschenhagen, T., B6hm, M., Hasenfuss, G., Just, H., Holtz, J., and Drexler, H. (1994). Gene expression of the cardiac Na+-Ca 2+ exchanger in end-stage human heart failure. Circ. Res. 75, 443-453. Talosi, L., Edes, I., and Kranias, E. G. (1993). Intracellular mechanisms mediating the reversal of/3-adrenergic stimulation in intact beating hearts. Am. J. Physiol. 264, H791-H797. Toyofuku, T., Kurzydlowski, K., Narayanan, N., and MacLennan, D. H. (1994a). Identification of Ser 38 as the site of cardiac sarcoplasmic reticulum Ca2+-ATPase that is phosphorylated by Ca2+/calmod ulin-dependent protein kinase. J. Biol. Chem. 269, 26 492-26 496. Toyofuku, T., Kurzydlowski, K., Tada, M., and MacLennan, D. H. (1994b). Amino acids Glu 2 to Ile 18 in the cytoplasmic domain of phospholamban are essential for functional association with the Ca2+-ATPase of sarcoplasmic reticulum. J. Biol. Chem. 269, 3088-3094. Toyofuku, T., Kurzydlowski, K., Tada, M., and MacLennan, D. H. (1994c). Amino acids Lys-Asp-Asp-Lys-Pro-Val 4~ in the Ca 2§ ATPase of cardiac reticulum are critical for functional association with phospholamban. J. Biol. Chem. 269, 22 929-22 932. Unverferth, D. V., Lee, S. W., and Wallick, E. T. (1988). Human myocardial adenosine triphosphate activities in health and heart failure. Am. Heart J. 115, 139-146. Verma, A. K., Filoteo, A. G., Stanford, D. R., Wieben, E. D., Penniston, J. T., Strehler, E. E., Fischer, R., Heim, R., Vogel, G., Mathews, S., Strehler-Page, M. A., James, P., Vorherr, T., Krebs, J., and Carafoli, E. (1988). Complete primary structure of a human plasma membrane Ca 2+ pump. J. Biol. Chem. 263, 14 152-14 159. Wegener, A. D., Simmerman, H. K. B., Liepnieks, J., and Jones, L. R. (1986). Proteolytic cleavage of phospholamban purified from canine cardiac sarcoplasmic reticulum. Generation of a low resolution model of phospholamban structure. J. Biol. Chem. 261, 5154-5159. Wolska, B. M., Stojanovic, M. O., Luo, W., Kranias, E. G., and Solaro, R. J. (1996). Effect of ablation of phospholamban on dynamics of cardiac myocyte contraction and intracellular Ca 2+. Am. J. Physiol. 271, C391--C397. Zucchi, R., Ronca-Testoni, S., Yu, G., Galbani, P., Ronca, G., and Mariani, M. (1994). Effect of ischemia and reperfusion on cardiac ryanodine receptors--sarcoplasmic reticulum Ca 2§ channels. Circ. Res. 74, 271-280.
John H.B. Bridge
Na+-Ca 2+ Exchange Currents
I. Introduction The existence of Na+-Ca 2+ exchange was first postulated in 1964 by both Repke (1964) and Langer (1964) as a consequence of their studies on the contractility of heart muscle. Three years later Baker and colleagues (1967) provided the first report documenting Na+-Ca 2+ exchange in giant squid axons. Shortly after this Reuter and Seitz (1968) presented the first complete study describing Na+-Ca 2+ countertransport in heart. Based on studies of isotopic fluxes, these authors proposed that two Na + ions were coupled to the extrusion of a single Ca 2+ ion in a modified exchange diffusion process. Blaustein and Hodgkin (1969) then published the results of their studies on squid axons. They recognized that the distribution of free Ca 2+ could not be predicted on simple electrochemical principles. However, cyanide (which was expected to block metabolic processes) failed to prevent the effiux of 45Ca, so that it seemed unlikely that a metabolic pump was involved in Ca 2+ extrusion. However, this effiux of Ca 2+ was (among other things) dependent upon external Na +. Blaustein and Hodgkin therefore concluded that, in unpoisoned axons, some or possibly all of the energy for extruding Ca 2+ ions came from the inward movement of Na + down its electrochemical gradient. These early studies were seminal and provided impetus for an enormous number of subsequent investigations that have led not only to a study of Na+-Ca 2+ exchange currents, but to the molecular cloning and elucidation of the structure of the exchanger molecule itself. This in turn introduced the possibility of studying the relationship between molecular structure and function. This chapter provides a brief description of current knowledge of Na+-Ca 2+ exchange currents. The ease with which heart cells can be patch-clamped, together with the presence of a vigorous exchange activity, doubtless explains the fact that most of our information on exchange current comes from this tissue. Na+-Ca 2+ exchange currents have been measured in other cell types including the squid giant axon (Matsuoka et al., 1997). In this chapter we will not deal with Na+-Ca 2+ exchange in the vertebrate rod outer segment Cell Physiology Sourcebook,: A Molecular Approach, Third Edition
283
(ROS). It is now known that this exchange process is different from that found in heart muscle. For example, under physiological conditions the exchange involves not only the Na + and Ca 2+ gradients but the K + gradient as well. In addition, the structure of the ROS exchanger is quite different from the heart exchanger (Achilles et al., 1991). There have been extensive studies of the currents associated with the ROS Na+-Ca 2+ exchange. The reader interested in the electrogenicity of this exchange can refer to the brief review by Lagnado and McNaughton (1990) and the more recent and extensive review on Na+-Ca 2+ exchange by Blaustein and Lederer (1999). Although measurements of exchange current are difficult in many tissues, it is now possible to express Na+-Ca 2+ exchangers in frog oocytes. This together with the development of giant excised patches that can be voltageclamped has made studies of the relationship between structure of the exchanger and function much easier.
II. Structure, Topology, and Distribution of the Na+-Ca 2§ Exchanger The cardiac Na+-Ca 2+ exchanger was the first to be isolated and cloned (Philipson et al., 1988; Nicoll et al., 1990). This exchanger is now referred to as NCX1. The functional exchanger is known to be 980 amino acids long (this excludes a signal peptide of 32 amino acids) with nine transmembrane segments (Fig. 1) (Nicoll et al., 1999). A large intracellular loop separates two sets of hydrophobic domains and it is these hydrophobic domains that contain the transmembrane segments. It is now known that the intracellular loop is not necessary for transport (Matsuoka et al., 1993). However two regions--the c~1 and a2 repeats, which are the sequences that span repeats 2 and 3 and 8 and 9mare similar and appear to be important in transport. There are also repeats in the large cytoplasmic loop. These are known as the /31 and /32 repeats. The function of these repeats is unknown. The cytoplasmic loop is, however, of great functional importance, particularly in the regulation of exchange (see Section IX.B). Copyright 9 2001 by Academic Press. All rights of reproduction in any form reserved.
284
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
F I G U R E 1. A proposed model of the Na+-Ca 2+ exchanger. Membrane-spanning segments are labeled 1 to 9. A large hydrophilic cytoplasmic domain contains a receptor that binds exchanger inhibitory peptide (XIP) and this site is thought to be associated with the phenomenon of Na+-dependent inactivation. In addition, this cytoplasmic domain contains a secondary Ca 2§ regulatory site and a region encoded by six small exons from which splice variants arise. For further details on the structure of the exchanger the reader is referred to Philipson and Nicoll (2000). (This figure was kindly provided by Dr. Kenneth Philipson.)
The topology of almost the entire exchanger has been mapped (Nicoll et al., 1999). A couple of features of this topology are particularly interesting. The ninth hy'drophobic segment has been proposed as a reentrant membrane loop that is similar to the pore-forming loop region of ion channels. The a repeats are currently modeled to be on the opposite side of the membrane (Fig. 1), which resembles that found in aquaporin water channels.
I!!. The Phylogeny of the Na§
2+ Exchanger
The number of proteins belonging to the exchange family is increasing. These proteins are defined by their similarity of structure including the presence of a repeats and multiple transmembrane segments. Recent analysis indicates four families of exchangers, which constitute a "superfamily." This is explained in some detail in the review by Philipson and Nicoll (2000). The first is the NCX family. These are all mammalian exchangers and include NCX1, the first exchanger to be isolated, as well as NCX2 and NCX3. It is on these exchangers, particularly NCX 1, that most studies of currents that are generated by exchange have been performed. Other exchangers in the NCX family (of which there are eight) are found, for example, in the nematode C. elegans and also in the fruit fly Drosophila. The NCX family is found in a wide variety of mammalian species including human, dog, rabbit, cow, rat, and guinea pig as well as in various nonmammalian species. A second family of exchangers is designated NCKX and consists of proteins similar to the Na+-(Ca2++K+) exchanger initially cloned from retinal tissue. NCKX 1 and NCKX2 are mammalian. An additional eight exchangers have been cloned from various invertebrates. These exchangers are structurally quite different from the NCX family and there have been considerable studies of the electrogenic properties of the NCKX ROS exchanger.
Two other families of proteins, one of which is bacterial and the other of which functions as a yeast Ca2+-H§ exchanger, are known to exist. However, discussion of these is outside the scope of this chapter since currents have not been measured in these exchangers.
IV. lsoforms of the Na§
2§ Exchanger
The large intracellular loop of NCX1 can be spliced in different ways to produce a variety of splice variants. A region of the loop (see Fig. 1) contains six small exons that are used in different combinations to produce tissue-specific splice variants which produce isoforms of the exchanger. The physiological significance of tissue-specific isoforms is unclear. Presumably they evolved to perform functions required only by those tissues in which they are found. Recent studies support this (Omelchenko et al., 1998). These workers observed, for example, that Na § and Ca 2§ regulated two splice variants of the exchanger found in Drosophila quite differently.
V. Energetics of Na§
2+ Exchange
Reuter and Seitz in their classic study proposed that two Na + ions exchanged for a single Ca 2+ ion in a modified exchange diffusion process. Based upon this proposal, one would not expect steady-state exchange activity to produce a measurable electric current. However, the existence of exchange currents is a well-established fact, and one that can be appreciated by consideration of the energetic and stoichiometric properties of the exchange as they are currently understood. It is convenient (but not necessarily correct) to represent the transmembranous exchange reaction as a sequential (simultaneous) process,
19. Na+-Ca2+Exchange Currents
285
nNa + + Ca 2+ ,--, nNa + + Ca 2+
o
o
i
where n is the stoichiometric coefficient of the exchange reaction. If the forward and reverse reaction rates are equal, the exchange reaction is at equilibrium. Even if n is greater than 2, there can be no net charge movement and, hence, no electric current generated at equilibrium. An electric current can only be measured when the exchange is displaced from equilibrium. While the net reaction rate at equilibrium is zero, the unidirectional rates might be substantial (Axelsen and Bridge, 1985). As soon as the forward and reverse exchange rates differ from one another, net ion translocation takes place; provided that the stoichiometric coefficient is appropriate (i.e., n greater than 2) and no ions of opposite charge are cotransported, one can expect to measure an electric current as a consequence of exchange. If electrochemical forces solely determine net movement of ions through the Na+-Ca 2+ exchanger, classical thermodynamics may be used to calculate both the direction of exchange and the conditions under which we may expect equilibrium to occur. Having the capacity to do this is of enormous value when designing experiments to measure exchange currents. The electrochemical potential difference or driving force (nA/2) producing exchange is the difference between n times the electrochemical potential difference or force producing sodium movement (nA/7~a) and calcium movement (A/2ca). Driving force can be expressed in terms of membrane potential (Em), Na + equilibrium potential (ENa), and Ca 2+ equilibrium potential (Eca). Thus, we may write
A/~ = nA/~Na- A/~ca
(1)
At equilibrium, A/2 = 0, so that nA/2Na- A/2ca = 0 RT
Rr
m -
ENa + E
m
m _
2Eca +
2E
[ ca~+]
[Ca2+]
In . _ _ _ _ _ _ q _ o
+ 2E
(3)
m
m =
n ENa
n E N a - 2 E c a - (n - 2) Em= 0
By nulling Ca 2+ movement over a range of membrane potentials, the value for n that these authors obtained was 2.97 + 0.03, which is close to the currently accepted value of 3.0. Equipped with a value for the stoichiometric coefficient of exchange, some useful parameters can be calculated. Heart muscle is used as an illustrative example to calculate expected reversal potentials for the exchange. It is assumed that a resting ventricular cell maintains intracellular Na + and Ca 2+ at concentrations of 10 mM and 100 nM respectively, and extracellular Na + and Ca 2+ at 140 mM and 2.0 mM. For n = 3, the reversal potential is given by (see Eq. 5)
exp - ( n - 2)
(8)
For the foregoing conditions Erev may be calculated to be -50 mV.
Vi. Methods and Problems Associated with the Measurement of Na§ 2§ Exchange Current (5)
The exponential form of Eq. 5 is
i2+1o {la+lo/ (
(7)
(4)
Substitution yields
[--C~a~ = Na+] i
(n - 2 ) E
Erev- 3ENa- 2 E c a - - 5 0 mV
[Na+]o
A/~Na(mV)- --~ In [Na+]i + E
A/~ca(mV) _ ~
(2)
An elegant (though somewhat indirect) demonstration of exchange stoichiometry was provided by Reeves and Hale (1984). Their study not only produced a value for the exchange stoichiometry, but provided an excellent example of the way that the foregoing energetic principles may guide experimental design. These authors took advantage of the fact that exchange equilibrium can be achieved simply by appropriate adjustment of the Na + and Ca 2+ electrochemical gradients. Bovine sarcolemmal vesicles containing Na +Ca 2+ exchanger were equilibrated with solutions of both Na + and 45Ca. Under these circumstances, the equilibrium may be described by Eq. 5. After treating the membrane with valinomycin in the presence of KC1, known membrane potentials were established that caused disequilibrium of the exchanger and either Ca 2+ entry or exit. By adjusting the Na + gradient, it was possible to precisely null the tendency of membrane potential to produce Ca 2+ movement. Thus, if any of the quantifies E m, ENa, and Eca are held constant, the relationship between the other two may be found. The point at which Ca 2+ movement was nulled by Na + gradient is given by
(6)
These are the equations that can be used to predict the equilibrium conditions of the exchanger. Before doing so, one needs to know the stoichiometric coefficient (n) of the exchanger. This issue has been the subject of a lengthy debate and numerous investigations. However, there now appears to be broad agreement that three Na + ions are exchanged for a single Ca 2+ ion in most mammalian systems studied (Bridge and Bassingthwaighte, 1983; Kimura et al., 1987; Bridge et al., 1990; Crespo et al., 1990). There is also good evidence that exchange stoichiometry is 3:1 in barnacle fibers (Rasgado-Flores and Blaustein, 1987).
Several recent techniques have greatly facilitated the isolation of exchange current. First, the whole-cell ruptured patch voltage clamp in conjunction with intracellular dialysis (Hamill et al., 1981) has probably contributed most to the initial isolation of exchange currents. To record whole-cell current with the ruptured patch technique, a microelectrode or patch pipette is pressed onto the cell membrane and suction is applied, usually with a syringe or directly by the experimentalist. This often results in a high-resistance seal that forms between the pipette and the cell membrane. The resistance of this seal is ideally > 101~ ohms and it is therefore referred to as a gigaseal. This is usually accompanied by the formation of an ~) structure as the membrane is pulled into the pipette (see Chapter 26). Continued suction ruptures the 1), and this creates continuity between the cell interior and the pipette solution. By pulling the sealed pipette from the cell one may isolate an inside-out patch. It is also possible
286
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
Rf
Rs
vo E ................ -
........ : - : ..... ............................................
Ag/AgCl t'-: ......
Ca 2*
l
FIGUI1E 2. An arrangement suitable for measuring Na+-Ca2+ exchange current in a ventricular heart cell. Under these circumstances the exchanger is producing a net outward current across the cell membrane by exchanging three internal Na+ ions with a single external Ca2+ ion. This produces outward movement of one net charge. The voltage-clamp circuit consists of a highgain amplifier A. Any deviations from the command potential Vc produced by the exchanger are compensated by feedback from the output through the feedback resistor Rf. The amplifier is connected to the cell through the electrode E, which is sealed to a ruptured patch. The output of this amplifier is proportional to the current flowing through the feedback resistor Rf and hence the cell C. The success of this method depends upon minimizing voltage drops across the series resistance R s. The bath in which the cell is maintained is grounded by a Ag/AgC1 electrode connected to a salt bridge (not shown).
with the methods illustrated in Chapter 26 to create an outside-in patch. Hilgemann (1989) has refmed this technique so that very large membrane patches can be isolated from appropriately treated cell membranes. A simplified arrangement for voltage-clamping heart cells and recording exchange current through a ruptured patch is depicted in Fig. 2. Essentially the same arrangement applies to the voltage-clamping of an isolated patch. The micropipette attached to the cell is connected via a (nonpolarizable) electrode to the front end of a high-gain amplifier known as a current-to-voltage converter. The most significant feature of this high amplifier is the feedback loop Rf and its branch point (the summing junction). In operational firnplifiers (of which this current-to-voltage converter is an example), this branch point has special significance. Negative feedback through the feedback loop tends to cancel the input signal at S. A command potential VC is applied to the noninverting input (+). A property of this amplifier is that as a result of negative feedback through Rf, current always flows so that the voltage at the summing junction is identical to the voltage Vc (the command potential). The potential difference between the summing junction and ground is therefore ideally equal to V c. This potential difference therefore falls across the electrode resistance (and any other resistances in series with the membrane resistance Rm). Provided that the product of this series resistance (Rs) and the exchange current INa-ca (RsXlNa-Ca) is small, most of the potential difference V occurs across the membrane. Feedback through Rf always compensates varying input at the summing junction, thus ensuring that the voltage across the membrane is held constant (i.e., is clamped).
Exchange current INa-Ca passing across the membrane produces a potential difference across the membrane. As this potential difference causes the membrane potential to depart from Vc, current (equal in magnitude to the exchange current) flows into the summing junction. In reality additional circuitry is required to reduce noise in the recording system while increasing its fidelity. This is dealt with in greater detail in Chapter 26 by Pun and Lecar. The current flowing through a heart cell membrane,/m' consists of two components: a capacity current and an ionic c u r r e n t / r The capacity current I c = C dV/dt and the membrane current is denoted I m. Therefore dV
I m - C --~ + I i +
INa-Ca
(9)
The ionic current consists of ions flowing through ion channels as well as those translocated by ion exchange mechanisms (principally the Na+-Ca 2+ exchange and the Na + pump). For purposes of illustration, the exchange current has been separated from the ionic current in Eq. 9. Under voltage-clamp dV/dt = O, In = l i +
INa-Ca
( ] 0)
If one can remove or m i n i m i z e / / ( w h i c h includes the Na + pump current and all contaminating channel currents) without affecting exchange current /Na-Ca' one may measure a membrane current that is comprised mainly of L a Ca" How will the activity of an electrogenic Na -Ca -exchange affect membrane potential? A convenient equivalent circuit to
19.
Na+-Ca 2+
Exchange Currents
287
explain this is depicted in Fig. 3. This simplified equivalent resting membrane circuit consists of a membrane capacitance that is in parallel with a battery Em, which is largely responsible for producing the membrane potential. The membrane resistance R m is modeled as the internal resistance of this battery. Under resting conditions the reversal potential for the Na +Ca 2+ exchange in heart is approximately -40 mV. A t - 8 0 mV the exchange will produce a net inward current, which can be modeled as a current generator that is in parallel with the membrane battery and capacitance (Fig. 3). At the steady state, when the resting membrane potential is not changing, C dV/dt = 0. Thus the net membrane potential is composed of contributions from ionic currents (mainly through K § channels), exchange currents, and the Na + pump current (Eq. 10). The exchange will produce an IR drop across the membrane resistance so that the contribution of Na+-Ca 2§ current to the membrane potential (WNa_Ca) at rest is given by WNa-Ca - Rm INa-Ca
(11)
A large inward exchange current produced by elevated intracellular [Ca 2§ may be of the order 200 pA. If we assume that the resting exchange current is one-tenth of this value and given a resting R m of 50 mO, the exchange will contribute about 1 mV to the resting membrane potential. In reality the value is likely to be less than this but cannot be specified with certainty because reliable measurements of resting current are not available.
Cm
Inside
Em Rm , I i+'vvvv~,
Outside
F I G U R E 3. A simple equivalent circuit for a resting heart cell. The resting membrane potential E m is typically about-80 mV. This resting membrane potential is modeled as a battery in series with an internal resistance of about 50 M~. The membrane capacitance is in series with the battery. The Na+-Ca 2§ exchanger may be considered to be a current generator in parallel with the battery and membrane resistance (symbolized by an open circle containing an arrow). At rest, with the ex-
changer steadily extruding Ca2§ a small depolarizing current INa-Ca causes a small voltage drop across the membrane resistance R m. The solid arrows indicate the direction of current. The circle containing an arrow indicates a current generator (the Na+-Ca2§ exchange). Thus for each cycle of the exchanger a single net charge is conveyed from the cell exterior to its interior.
The first observations of exchange current were not published until nearly 20 years after Reuter and Seitz originally reported Na+-Ca 2+ exchange activity in heart tissue. It is worth considering some of the problems associated with the measurement of exchange current that might account for this lengthy hiatus. To stimulate exchange so that it can be measured, one can either change electrochemical gradients (and hence the exchange driving forces) by using a voltage-clamp step or one can abruptly change external ion composition with a rapid switching device. In the case of the voltage-clamp step, the experimentalist must have a complete understanding of what interfering currents will be activated, as well as some means of inhibiting them. If exchange is to be activated by changing ionic gradients, then these changes must be sufficiently rapid that they are not immediately dissipated by the exchange activity. It is difficult to rapidly change or control extracellular ionic composition in multicellular preparations because diffusion distances are large and diffusion may be hindered by a surface structure. In multicellular preparations, voltage-clamp is usually established with small microelectrodes, which are unsuitable for cell dialysis. It is, therefore, very difficult to control the intracellular ionic composition. This doubtless contributed to the difficulty of isolating exchange currents in multicellular preparations. Despite these difficulties a measurement of exchange currents has been reported in multicellular preparations (Horackova and Vassort, 1979). Recently, it has proved feasible to use the patch-clamp technique in conjunction with wide-tipped microelectrodes (10-1xm diameter) to obtain an isolated patch whose surface area is approximately 75 lxm2 (Fig. 2). A conventional isolated patch is of the order 2-3 ixm2. These giant patches are sufficiently large to permit the detection of exchange currents (Hilgemann, 1989). Their great virtue is that they permit relatively easy access to either side of the membrane. Thus, solution adjacent to the external surface of the sarcolemma can be changed by changing the pipette solution. The internal surface of the sarcolemma can be changed rapidly by changing the bathing solution. Clearly this method of voltageclamp allows the experimentalist considerable control over the forces that drive exchange. Several other techniques have recently become available that improve exchange current isolation. For example, forward exchange current in intact cells can be activated by abruptly elevating intracellular Ca 2§ which is then transported to the cell exterior. This has recently been accomplished in an elegant fashion by Niggli and Lederer (1991 b). These workers used the compound DM-dinitrophen that is commonly referred to as caged Ca 2§ This can be introduced into the cell interior through a patch pipette. Upon appropriate irradiation with UV light, Ca 2§ is released from the caged Ca 2§ with extremely rapid kinetics. This abruptly elevates cytosolic Ca 2§ and transiently stimulates forward Na § Ca 2§ exchange as the released Ca 2+ is pumped to the cell exterior. As we have indicated, to activate exchange it is desirable to change external ionic composition (and the electrochemical gradients driving exchange) extremely rapidly in comparison with the time required for the exchange to dissipate these
288
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
changes. It is now possible to change external solutions surrounding heart cells or isolated patches with a half-time of about 20 ms (this includes exchange of the unstirred layer). One method (Spitzer and Bridge, 1989) consists of placing a cell with attached microelectrodes in one of two adjacent microstreams of solution. The boundary separating these streams is abruptly moved across the cell so that it is placed in the adjacent stream. This rapid switching method has proved valuable in stimulating both inward and outward exchange currents (Chin et al., 1993).
A
,,I min
,,,,l~li~: -160mV "J-120
- . , ~ , ~ ~ I ~ , I . ~ 1 1 ~ I ~ :~, l l l l l i ~ l ' l l l ~ ' " l ~ I1
a
b
r
d
1Cao B
1.1
e~
TIq.I .
1Ca o inAt
0 Na i
~
I t,
0.1mMLa ,,..~l!~i
0hA
1Cao C
inA/e 30 Na i
-,r.V .oo. ./ol
VII. Isolation of Na+-Ca 2+ E x c h a n g e Current A. Whole-Cell Patch-Clamp Studies The first clear evidence that an electric current was generated by Na+-Ca 2+ exchange was provided simultaneously by Kimura and coworkers (1986) and by Mechmann and Pott (1986). Given the probable stoichiometry of the exchange, electrophysiologists expected to be able to measure a current generated by the exchange. By using the whole-cell patchclamp technique together with intracellular dialysis, Junko Kimura and her associates were able to isolate the Na+-Ca 2+ exchange current. These investigators voltage-clamped guinea pig ventricular myocytes with single microelectrodes. The pipette contained a dialyzing solution completely deficient in Na + and in which Ca 2+ was buffered to a value of 73 nM with EGTA. The superfusing external solution contained no Ca 2+. Under these circumstances no exchange could take place. Very little change in current was observed when external Ca 2+ was reapplied and subsequently removed. However, an outward current was generated after changing the pipette solution for one containing 30 mM Na + and then applying 1.0 mM external Ca 2+. Under these circumstances Ca 2+ entered the cell in exchange for internal Na +, which was extruded. This current (Fig. 4) is clearly attributable to the operation of electrogenic Na§ 2+ exchange. This conclusion was strengthened by several additional observations. Upon removal of extemal Ca 2§ the current was turned off. The current was reduced when lower concentrations of dialyzing Na + were used and enlarged by increasing external Ca 2+. Thus, the current exhibited the expected dependency on both Na + and Ca 2+. The current was blocked by La 3§ which is known to block Na+-Ca 2+ exchange. The current also exhibited voltage dependence (see Section X). One might have expected sustained current under these circumstances, but this was not observed (Fig. 4) and the current exhibited a tendency to decay. One plausible explanation for this is that excessive Ca 2+ entry displaces protons when it is buffered by the dialyzing EGTA. The resulting acidification in the vicinity of the exchanger should inhibit exchange activity. An alternative possibility is that EGTA fails to buffer incoming Ca 2+, which accumulates in the vicinity of the exchanger. The resulting collapse of the Ca 2§ gradient would also slow exchange. Kimura and colleagues reported the measurement of an outward current corresponding to the entry of Ca 2+ and extrusion of Na + from the cell. Mechmann and Pott (1986) demonstrated the existence of an inward exchange current corresponding to Ca 2+ extrusion and Na + entry into a cardiac cell. In one experiment, these authors voltage-clamped spherical atrial cells
FIGURE 4.
(A) Voltage-clamp record from a single myocyte showing voltage (upper trace) and current (lower trace). A ramp pulse from +60 mV to-120 mV was given from a holding potential of-30 mV every 10 s at a ramp speed of about 0.2 V/s. Current-voltage relationships were then constructed. An application of 1 mM Ca2+ in the absence of intracellular Na§ (zero pipette Na § produced a small outward current. To exchange the pipette solution, a piece of tapered polyethylene tube was inserted into the pipette. The arrow indicates when the solution change was started. The holding current shifted outward slightly on loading Na+. The middle current record shows 1 mM Ca2§ superfusion in the presence of 30 mM pipette Na+. The Ca2+-induced outward current appears to decay after reaching a peak. Current undershoot is seen after washing off Ca2+. These phenomena may be due to Ca2+ accumulation immediately below the membrane of acid induced by binding of Ca2§ to EGTA. (B) I-V relations measured before (a) and during (b) lmM Ca2+ superfusion in the absence of Na+. (C) I-V curves before (c) and after (d) loading Na+ and during subsequent lmM Ca~ application (e). (Reprinted from Mechmann and Pott (1986) with kind permission of the authors and Nature, copyright 1986, Macmillan Magazines Ltd.)
from an adult guinea pig. Transient inward current occurred spontaneously in this preparation. These transient inward currents seemed to depend on intracellular Ca 2+ because they were abolished if the cell was dialyzed with 1.0 m M EGTA. Moreover, they could not be elicited in the presence of caffeine, which depletes the SR of Ca 2§ Presumably the transient release of SR Ca 2+ activated these currents. The currents also showed a dependence on extracellular Na + and voltage as expected of a Na+-Ca 2+ exchange current. The more negative the voltage, the larger the inward current. This is reasonable and simply indicates that at more negative potentials the exchange rate is greater so that Ca 2+ released from the SR is removed rapidly from the cell. Shortly after these early demonstrations of exchange current, Hume and Uehara (1986a, b) were able to demonstrate Na+-Ca 2+ exchange currents in isolated frog atrial cells by using the whole-cell patch-clamp technique in combination with intracellular dialysis. It is fair to say that since this time measurement of exchange current has become both reliable and routine. An example of an inward exchange current measured in a ventricular cell is displayed in Fig. 5. The development of the giant patch technique has also facilitated measurement of the exchange current produced by the squid exchanger NCX-SQ1 (He et al., 1998) (see Section X).
19. Na+-Ca2+ Exchange Currents
289
0.16
,~
0.08
1
0 mM
extremely sensitive to mutations in the ce repeat region (see Fig. 1). For example, a nonfunctional exchanger can be produced if Ser 110 is mutated to Ala. It appears therefore that the gene duplication event that produced the ce repeat regions is of special functional importance. In general, exchange activity is not affected by mutations in the intracellular loop; however, mutations of transmembrane segment 2 dramatically affect ion selectivity. Normally Li § is not transported by the exchanger. However, when Thr 103 is mutated to a Val, Li+-Ca 2§ exchange c~n occur (Doering et al., 1998). Mutations of this nature could prove extremely valuable as probes of the function of the exchanger, as in cardiac excitation contraction coupling.
Na(Li)
-0.08 -0.16 -0 . 2 4 t
I.
0
1
1
2 Time (s)
i
3
__l
4
FIGURE 5. Transient inward Na+-Ca2+ exchange. To activate this current, a guinea pig ventricular cell was tetanized with voltage-clamp pulses in the presence of ryanodine and in the absence of external Na§ This resulted in an elevation of cytosolic Ca2§ and a sustained contraction (not shown). Abrupt application of 145 mM extracellular Na§ while the cell was held at --40mV produced mechanical relaxation and activated a transient inward Na+-Ca2§ exchange current. As intracellular Ca2+ declined, the current decayed. (Reprinted with permission of the Annals of the New York Academy of Science.)
B. Na+-Ca 2+ Exchange Current Reversal Potential It is desirable when studying any ionic current to be able to demonstrate a reversal potential for that current. The existence of reversal potentials is a thermodynamic necessity and provides one of the least ambiguous ways of identifying an ionic current. However, it is not always a straightforward matter to demonstrate reversal potentials. For example, if exchange current were extremely small over a large voltage range in the vicinity of the reversal potential, then the precise potential at which zero current occurred could be extremely difficult to specify. Moreover, subsarcolemmal spaces from which diffusion is restricted can produce changes in ionic concentration that confound measurement of the exchange reversal potential. For the Na+-Ca 2+ exchange in heart the reversal potential is given by Eq. 8. For the conditions of their experiment, Kimura and coworkers (1986) (see Fig. 4) calculated that the reversal potential of exchange ought to be -131 mV. Their results indicate that exchange current becomes zero at a potential close to this value. Ehara and colleagues (1989) extended studies of the exchange reversal potential. With the fairly specific exchange inhibitor Ni, these authors were able to show that over a wide range of external Na § values the measured exchange reversal potential conformed to theoretical expectation. Moreover, in view of the wide range over which the reversal potential remains constant, the stoichiometry is unlikely to vary.
VIII.
Ionic D e p e n d e n c i e s of N a + - C a 2+ Exchange Current
It is of considerable physiological significance that the Na+-Ca 2+ exchange current is extremely sensitive to internal Na +. There are two aspects of this sensitivity to consider, one purely thermodynamic and the other kinetic. If we assume that a resting heart cell contains 100 nM free cytosolic Ca 2+ and is bathed in 2.0 m M Ca 2+ and 140 m M Na +, we can use Eq. 8 to calculate the way the exchange reversal potential varies with internal Na +. The results are displayed in Fig. 6. It is apparent that with 10 m M intracellular Na +, 140 m M extracellular Na +, and 100 nM intracellular free Ca 2+, the reversal potential for exchange is - 5 0 mV. Since the resting membrane potential is at least 30 mV negative to this, at rest the exchanger will extrude Ca 2+. However, a modest increase in intracellular Na + to 12 mM would change the reversal potential by 14 mV t o - 6 4 mV. Were the Na + to accumulate to 15 mM, then the reversal potential would be - 8 0 mV. Since the resting potential is likely to be close to this value the exchange would be close to, or at, equilibrium and would be incapable of extruding intracellular Ca 2+. Further Na + accumulation
Erev (mV}
- 150
- 100
I
I
- 50 .
.
.
.
0 l
50 9
5
C. Na+-Ca 2+ Exchange Current (Transport) and Structure Some information is now available on the relationship between molecular structure and transport function. It appears that exchange activity (and presumably exchange current) is
FIGURE 6. Calculated variation of exchange reversal potential with intraceUular Na+. Intracellular Ca2+ was set at 100 nM and extracellular Ca2+ was 2.0 mM. Extracellular Na+ was 140 mM.
290
SECTION
would cause the exchange to reverse and transmit Ca 2+ into the cell. Na + accumulation will tend to bring the exchange closer to its equilibrium potential and, therefore, increase the likelihood of Ca 2+ entry via the exchanger upon membrane depolarization. Cardiac glycosides, which, depending on dose, partially or completely block the sodium pump, tend to produce an accumulation of intracellular Na + (Sheu and Fozzard, 1982). Part of the basis of their inotropic effect resides in their shifting the reversal potential of the Na+-Ca 2+ exchange to more negative values. In extreme cases, this will prevent Ca 2+ extrusion which, in the face of continued Ca 2+ leak, will lead to Ca 2+ accumulation. The Na+-Ca 2+ exchange current also exhibits a physiologically important kinetic dependence upon internal Na +. This issue has been investigated directly by Miura and Kimura (1989), who studied outward exchange current in guinea pig ventricular cells under voltage-clamp. Intracellular Ca 2+ was buffered to 100 nM and intracellular Na + was varied by varying the pipette Na + concentration. External Na + was reduced to zero and a unidirectional exchange reaction was activated by applying 0.2 mM Ca 2+ as rapidly as possible to the cell exterior. The results demonstrate that exchange current density exhibits a steep and somewhat sigmoid dependence on internal Na +. The K m for this dependency is approximately 20 mM and Hill plots reveal a Hill coefficient of 1.9. It is notable that an application of rate theory revealed that the voltage dependence of the exchange did not depend on intracellular Na + (see Section X). It is clear from these results that at least Ca 2+ influx on the exchange will be extremely sensitive to fluctuations in intracellular Na + produced, for example, by increases in stimulation rate or cardiac glycosides. It is unfortunate that we have no information on the intracellular Na + dependence of inward exchange current which corresponds to Ca 2+ efflux. However, studies with isotopic fluxes reveal that Na + and Ca 2+ compete for transport on the exchanger. For example, Na + on the same side of the cardiac sarcolemmal membrane as Ca 2+ inhibits Ca 2+ flux through the exchanger with a K / o f about 15 mM (Reeves and Sutko, 1983). Finally, the net inward exchange current does depend on external Na +. This has been investigated by Kimura and coworkers (1986), who showed that inward current exhibited a sigmoidal dependence on external Na + with a K1/2 of 87.5 mM and a Hill coefficient of 2.9. Net inward exchange current (i.e., Ca 2+ extrusion in exchange for Na + entry) is also dependent on intracellular Ca 2+. The relationship between Ca 2+ concentration and current has been measured in two different ways. Miura and Kimura (1989) used whole-cell patch-clamp on guinea pig myocytes to determine K m (Ca2+). They controlled intracellular Ca 2+ by dialyzing the cell with various EGTA-buffered solutions and determined a K m (Ca 2+) to be 0.6 JaM. With a different approach, Barcenas-Ruiz and colleagues (1987) used guinea pig ventricular myocytes to measure exchange current and intracellular Ca 2+ simultaneously. Intracellular Ca 2+ was measured with the CaZ+-sensitive indicator Fura-2, introduced through the dialyzing pipette used to voltage-clamp the cells. Heart cells can be tetanized or caused to go into contracture in the presence of ryanodine. Various clamp pulses up to 120 mV in amplitude (holding potential o f - 8 0 mV) produced slow and sustained in-
!I
Membrane Potential, Transport Physiology, Pumps, and Exchangers
creases in Ca 2+. This Ca 2+ declined upon repolarization and the decline was accompanied by a Na+-Ca 2+ exchange current tail. Plots of current against [Ca2+]i revealed a linear relationship and no sign of saturation (Fig. 7). It is not possible to extract a K value from this data since the relationship between Ca 2+ an~ current does not saturate. However, it is reasonable to conclude that the K m value is well above the 0.6 IxM obtained by Kimura and colleagues. Discrepancies of this nature are not easy to resolve. It is, however, worth noting that EGTA has the rather curious effect of increasing the affinity of the exchanger for Ca 2+ (Trosper and Philipson, 1984).
IX. Regulation of Na§
2§ E x c h a n g e Current
The regulation of Na+-Ca 2+ exchange current has largely been studied in NCX1. We will consider four types of regulation. First, intracellular Ca 2+ regulates the exchange current. The exchange current also exhibits a regulatory phenomenon known as Na+-dependent inactivation. Thus the exchanger not only transports Na + and Ca 2+, but these ions are capable of regulating its function as well. Exchange current is also known to be regulated by phosphatidylinositol-4,5-bisphosphate (PIP2), and finally, by phosphorylation. In addition, a great deal has been learned about regulatory function and its relationship to the structure of the exchanger.
A. Regulation by Na § and Ca 2§ There are two Ca 2+ binding sites on the intracellular surface of the cell membrane: one is the primary transport site, and the other is a high-affinity Ca 2+ site required for secondary ionic regulation. It is now apparent that Ca 2+ influx in exchange for internal Na + cannot occur if intracellular Ca 2+ is reduced below a certain critical level. The first description of this secondary regulation of Na+-Ca 2+ exchange by internal Ca 2+ was provided by Baker and McNaughton (1976) in their study of squid axons. In their more recent studies of heart cell membrane, both Kimura and colleagues (1986) and Hilgemann (1990) have shown that when internal Ca 2+ is sufficiently reduced, exchange current is inactivated. It appears that in dialyzed myocytes the K o for the regulatory site is about 50 nM. However, in isolated patches, the K o appears to be from 0.1 to 0.3 I~M. If this is true, the regulatory sites in heart muscle would always be saturated and serve little function. If, on the other hand, the lower value obtained in dialyzed cells is correct, then Ca 2+ regulation would acquire significance as a regulator of both Ca 2+ influx and effiux on the exchanger. Why the regulatory site exists remains somewhat unclear. Na+-dependent inactivation of exchange current is a rather curious regulatory phenomenon first described by Hilgemann and coworkers (1992). Outward Na+-Ca 2+ exchange currents were activated in giant sarcolemmal patches by increasing Na + on the cytoplasmic side of the patch. Ca 2+ was then transported from the pipette to the cytoplasmic side of the patch in exchange for Na +. With 60 mM Na + on the pipette side of the patch, application of 60 mM
19.
Na+-Ca 2+
291
Exchange Currents
4o]
!
1600
-8oJ
!
,._. 1 2 0 0 i
.- 800 1.2
o
'-"
O 400
O
0.8
~0.4 8:8
0|'
i..
-80
0.01
Ill
-;~0
.
v (mv)
'.
6
',l'
"
,o
C
-0.1
-0.1
~,-0.2
"" - 0 . 2
-0.3 ;
1
1
s
-0.4
t t
0 oo
.
.
.
0 40
.
.
c-
0 80 [Ca2+lj
1 20
9
1.~0
(I~M)
F I G U R E 7. Changes in [Ca2+]i and membrane current attributed to Na+-Ca2§ exchange in a single guinea pig ventricular myocyte under voltage-clamp. Micropipette [Na+] was 7.5 mM. (A) Simultaneous recordings of [Ca2+]i and membrane current. The holding potential was -80 mV, and depolarizing pulses lasting for 1.5 s were given from -60 mV to +40 mV. Outward currents during the depolarizing pulse are off-scale. (B) Voltage dependence of the change in [Ca2+]i. Values plotted are average [Ca2+]i over the last 200 ms of the depolarizing pulse. (C) The relationship between [Ca2+]i and membrane current after repolarization. The current, I, has been plotted as a function of [Ca2+]i at 2 ms intervals during the first second after repolarization from +40 mV (circles), 0 mV (triangles), -20 mV (diamonds), and -40 mV (crosses). [Reprinted with permission from Bridge, J. H. B., Smolley, J. R., and Spitzer, K. W. (1990). The relationship between charge movements associated with/Ca and INa-Cain cardiac myocytes. Science 248, 376-378. Copyright 1990 American Association for the Advancement of Science.]
Na + on the cytoplasmic side of the patch activated outward current in the presence of a large Ca 2+ gradient (at 0 mV membrane potential). This outward current declined with a time constant of approximately 1 s. This decay rate is far too low to represent an initial turnover of exchangers and it appears to be a true regulatory phenomenon. An example of this sort of behavior is displayed in Fig. 8. Here, application of 100 mM Na § to the cytoplasmic surface activated a transient outward current. B. R e l a t i o n s h i p b e t w e e n Ca 2+ a n d Na + R e g u l a t i o n a n d Structure
Advances in the study of the molecular biology of the Na +Ca 2+ exchanger have led to a study of the relationship between ionic regulation and the structure of the exchange molecule. Both Ca 2+ secondary regulation and Na§ inactivation can be removed if the cytoplasmic surface of an excised patch is treated with chymotrypsin (Hilgemann, 1990). From this it may be inferred that these regulatory properties are related in some way to a cytoplasmic domain of the exchange protein. Recent evidence suggests that the sarcolemmal Na+-Ca 2§ exchanger is regulated by intracellular Ca 2§ at a high-affinity binding site separate from the Ca 2§ transport site. It was first suggested by Matsuoka and colleagues (1993) that the large hydrophilic loop of the exchanger was the binding site for regulatory Ca 2+. Levitsky and colleagues (1994) have identified a high-affinity Ca2§
region comprising amino acids 371-521. This is a large region of the cytoplasmic loop and Levitsky and colleagues have inferred that substantial secondary structure is necessary for high-affinity Ca 2§ binding. Detailed functional studies to measure exchange currents on giant patches from oocytes expressing both wild-type and mutant exchangers have revealed additional information on the Ca 2§ regulation of Na+-Ca 2+ exchange. Mutation of certain Asp residues within two acidic
20
10
pA
L_ I
100 mM Nai +
FIGURE 8. Secondary modulation of outward Na+-Ca2+ exchange current in giant excised inside-out patches (37 ~ 0 mV). Current transients activated with 100 mM Na + (substituted for 100 mM Cs+). (This figure was kindly provided by Dr. Donald Hilgemann.)
292
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
segments markedly decreased Ca 2+ binding. These mutations exhibited parallel effects on functional Ca 2+ regulation. Until recently it has only been possible to study Ca 2+ regulation when the exchanger operates in the reverse mode. This is because regulatory and transported Ca 2+ are on opposite sides of the membrane so that the concentration of regulatory Ca 2+ can be manipulated without affecting the concentration of transported Ca 2+. Matsuoka and coworkers (1995) have been able to exploit a class of mutants with low affinity for Ca 2+ at the regulatory site to demonstrate regulatory effects of Ca 2+ on forward mode exchange. Finally, the affinity of both wild and mutant Na+-Ca2+ exchangers for transported Na + declines with declining regulatory Ca 2+ concentration. This further indicates that Ca 2+ regulation modifies transport properties and does not simply control the fraction of exchangers that are available in an active state. Some information is now available on the part of the exchange molecule that is involved in Na+-dependent inactivation (Matsuoka et al., 1997). By studying mutations in the XIP region of the exchanger molecule, these authors have concluded that the endogenous XIP region is primarily involved in movement of the exchanger into and out of the Na+-induced inactive state. The name XIP refers to exchanger inhibitory peptide, which has the same sequence as the XIP region and is a potent inhibitor of exchange activity.
C. PIP2 Hilgemann has been able to show that outward exchange current in isolated giant sarcolemmal patches is stimulated by MgATP (Hilgemann, 1990). With this preparation it is possible to control MgATP on the cytoplasmic side of the cell membrane. This stimulation by MgATP appears to be a different process than that described in squid axon. For example, it does not appear to involve a protein-dependent kinase (Collins et al., 1992). Hilgemann has also shown that brief treatment of membrane patches with chymotrypsin abolished modulation of exchange current by MgATE However, recent evidence has been obtained for a mechanism that seems likely to explain regulation by ATP (Hilgemann and Ball, 1996). It now appears that phosphatidylinositol-4,5-bisphosphate (PIP2) can strongly activate the cardiac Na+-Ca 2+ exchanger. Apparently ATP generates PIP 2 from phosphatidylinositol (PI). A number of observations suggest that the generation of PIP 2 can explain the regulation of Na+-Ca 2+ exchange by ATP: (1) PI-specific phospholipase C can abolish the action of ATP, which can in turn be restored by the addition of exogenous PI. (2) The effect of ATP can be reversed by a PIP2-specific phospholipase C. However, the stimulatory effect of ATP can be mimicked by exogenous PIP 2. (3) Aluminum, which binds with high affinity to PIP 2, reverses the effect of ATP on the exchanger. It therefore appears that the generation of PIP 2 is an important regulator of Na+-Ca 2+ exchange activity in heart.
D. Phosphorylation It has been known for some time that exchange activity in squid giant axons could be regulated by ATE The main ef-
fect of ATP is to increase the affinity of the exchanger for its substrates Na + and Ca 2+. Moreover, recent evidence suggests the involvement of a CaZ+-dependent protein kinase. It appears that, at least in the squid, Na+-Ca 2+ exchange that is stimulated by ATP requires intracellular Ca 2+ and can be mimicked by hydrolyzable ATP analogs (DiPolo and Beauge, 1987a, b). It has been difficult to demonstrate regulatory effects of ATP in cardiac tissues. The whole-cell isolated patch technique does not lend itself to studies of this nature because it is extremely difficult if not impossible to control intracellular ATP concentrations. However, recently Iwamoto and Shigekawa (1998) directly detected phosphorylation in cardiac and aortic smooth muscle using immunoprecipitation and autoradiography. It is also now known that exchanger from amphibian heart is down-regulated by isoproterenol via the protein kinase.
X. Current-Voltage Relationships and Voltage D e p e n d e n c e of Na+-Ca 2+ Exchange Current The origin of the voltage dependence of Na+-Ca 2+ exchange has yet to be explained. However, current-voltage relationships obtained under a variety of ionic conditions can provide a great deal of information that forms a basis for discussing possible mechanisms of voltage dependence. Here we will try and show what can and cannot be concluded about voltage-dependent mechanisms from available current-voltage data. Before interpreting current-voltage relationships, the experimentalist should be confident that he or she is dealing with a pure current. Thus, considerable care in eliminating contaminating currents with appropriate inhibitors and an understanding of which currents (besides exchange current) are activated over the voltage range of interest is essential. Most Na+-Ca 2+ exchange current-voltage relationships that have been measured so far do not show a region of negative slope. This is consistent with the idea that only a single rate-limiting charge translocation step exists in the reaction pathway. Were a second voltage-dependent step to exist in which charge moved in the opposite direction, then one might expect a region of slight negative slope. This is because the increasing voltages that stimulated forward exchange would necessarily begin to retard the movement of exchange in the opposite direction, with resulting decline in net transport and hence current. However, conditions under which a region of negative slope in an exchange currentvoltage relationship have been detected recently. Using the giant patch technique, Hilgemann and coworkers (199 lb) have measured outward exchange current-voltage relationships as a function of extracellular Ca 2+ (Fig. 9). When extracellular Ca 2+ is 0.1 mM, voltage dependence is diminished and at extreme depolarization the slope of the current-voltage relationship becomes discernibly negative. It is currently believed that most of the exchange voltage dependence is associated with Na + translocation. Hilgemann has suggested that the negative slope could be explained if Ca 2+ passes through a small fraction of the membrane field before binding to the carrier. This would constitute a second
19. Na+-Ca2+ Exchange Currents
293
0.1 mM
100
50
8 mM [Ca]o
i
-100
I
-50
,.,
I
I
0
50
E m (mY)
FIGURE 9. Ion and voltage dependencies of cardiac N a + - C a 2+ excurrent conform to consecutive exchange model with voltage dependencies at extracellular Na§ release and binding. Flattening and saturation of outward INa_ca-VOltagerelations as extracellular [Ca2§ is lowered from 8 mM (closed circles) to 2 mM (open circles) to 0.4 mM (closed squares ) to 0.1 mM (open squares). All results are normalized to the largest current occurring in the current-voltage relation. Note the negative slope with strong depolarization at 0.1 mM. (Reprinted with kind permission of the author and Nature.) change
voltage-dependent step associated with Ca 2+ translocation in the transport pathway. A second property expected of current-voltage relationships for electrogenic exchange reactions is that they should saturate at extreme voltages. This is because reaction steps whose rate increases with voltage should become sufficiently rapid that they cease to be rate limiting for the entire reaction sequence. The reaction may then be rate limited by a voltage-independent step, at which point reaction rate will cease to be responsive to voltage. Many current-voltage relationships do show clear evidence of saturation. For example, the measurements of outward current depicted in Fig. 9 clearly exhibit saturation. Though not so obvious, measurements of inward current by Bridge and colleagues (1991) and by Miura and Kimura (1989) also show signs of saturation at extremely negative voltages. In contrast, recent results (from giant patches) under zero trans conditions suggest little saturation of inward exchange current at very negative potentials (Matsuoka and Hilgemann, 1992). The origin of these discrepancies is not clear, but may be related to whether or not net or unidirectional exchange is producing the inward current. Moreover, current-voltage relationships obtained in whole cells may be complicated by other partial reactions (e.g., Ca 2+ dissociation from intracellular buffers) not present in the giant patch experiments. The way that exchange current-voltage relationships depend on ionic conditions can also yield information about voltage-dependent steps in the reaction sequence that leads to exchange. With voltage-clamped giant patches under zero trans conditions--that is, Ca 2+ in the pipette (extra-
cellular) side and Na § in the bath solution--that superfuse the cytoplasmic surface, outward current-voltage relationships showed a striking dependence on pipette (extracellular) Ca e§ (Fig 9). However, when the extracellular (pipette) Ca 2§ is reduced to very low values, voltage dependence tends to disappear. An attractive explanation for this result is that as extracellular Ca 2§ is reduced, a voltage-independent step or a step with very modest voltage dependence (presumably associated with Ca 2§ translocation) starts to become rate limiting, with the result that exchange loses its voltage dependence. However, Ca e§ translocation is not independent of voltage in all species. In the squid Na+-Ca 2§ exchanger N C X - S Q 1, charge movement accompanies Ca 2§ translocation rather than Na § translocation, as has been shown rather elegantly by He and coworkers (1998). These authors examined giant patches from oocytes expressing either NCX1 or NCX1-SQ1. Na § and Ca 2§ were respectively inside the pipette. This forced Na § and Cae+-bind ing sites to the intracellular surface. Rapid application of Na § or Ca 2§ to the intracellular (outside) surface of the patch produces a transient half-reaction, which in turn was expected to produce a transient charge m o v e m e n t if the half reaction is electrogenic. It appears that in NCX1 transient current is produced by application of Na § but in the case of NCX1-SQ1 application of Ca 2§ produces an electrogenic half-reaction. Thus it seems that in the squid, Ca 2§ translocation is electrogenic (Fig. 10). The dependence of outward current-voltage relationships on Na § is also interesting. Hilgemann and colleagues (1991 b), using the giant patch technique, have measured outward current-voltage relationships when extracellular Na § was 100
A
B
l pA :7_: ".: _:_~ :;-:~ jfaPd*'az~Nlz,iNuo~ '
I
V
14o mM Na+l
_ 140 mi Na;'i
C l.,..t
. . . . . .
D 11~ . . . . . .
,,, ,,
| 5 pM Ca=§
IS pM ca'+i
FIGURE 10. Identification of electrogenic reactions of NCX1 and NCX-SQ1 using concentration jumps. (A) Current transients recorded from an NCXl-expressing patch when 40 mM cytoplasmic Na+ was applied and removed in the presence of 20 mM extracellular Na+. (B) Typical lack of current transients recorded from an NCX-SQ1expressing patch when 40 mM cytoplasmic Na+ was applied and removed in the presence of 20 mM extracellular Na+. (C) Inward NCX1 current activated when a solution with 5 ~M free C a 2+ w a s applied as in A. (D) Inward NCX-SQ 1 current activated when a solution of 5 puM free Ca2+was applied as in B. (Modified from He et al., (1998) and reprinted with kind permission from the Journal of General Physiology.)
294
SECTION I1 Membrane Potential, Transport Physiology, Pumps, and Exchangers
and 400 mM (Fig. 11). At 400 mM there was a pronounced flattening of the current-voltage relationship. There are two possible (and not necessarily unrelated) explanations for this. If we assume that some step in the Na § translocation pathway is both rate limiting and voltage dependent, it is possible that as extracellular Na § increases, this step ceases to be rate limiting because it is accelerated. Were some other voltage-independent step to become rate limiting, one would expect a flattening of the current-voltage relationship as it became dominated by the voltage-independent step, as is in fact observed. Alternatively, if Na § binding to some external site were voltage dependent, one would expect voltage dependence of this binding to be lost as the site becomes saturated (Lagnado and McNaughton, 1990). These results seem to suggest that somewhere in the Na § translocation pathway a voltage-dependent (charge translocation) step may be involved. In this regard the data obtained with giant patches, where ideal zero trans conditions are most likely to be achieved, are particularly compelling.
XI. M e c h a n i s m of Na+-Ca 2+ E x c h a n g e A discussion of the evidence for and against various kinetic schemes to account for Na§ 2§ exchange is beyond the scope of this chapter. The interested reader is referred to Khananshvili (1991). Here we will simply show the way 'in which measurement of exchange current can be used to infer mechanism. It should, however, be clearly understood that the measurement of isotopic fluxes and rapid mixing techniques provide indispensable tools to investigate mechanism but will not be discussed here.
0
-
~ f I
~
For the Na§ 2+ exchange, two different basic mechanisms for ion translocation can be considered: the PingPong or consecutive mechanism and the sequential or simultaneous mechanism. Although it has been difficult to distinguish between these mechanisms, recent data using the patch-clamp technique provide strong evidence in favor of the Ping-Pong mechanism, diagrammed in its simplest form in Fig. 12. There is only one set of binding sites that binds either Ca 2§ or Na § and the translocation of Na § and Ca 2§ are separate events. For example, Ca 2§ binds to the carrier E' at the inner (cytoplasmic) surface and is translocated to the exterior, where it is released. The carrier E " then can either bind Na + or Ca 2+ at the external face, which is then transported to the internal surface where it is in turn released. A basic property of the Ping-Pong mechanism is that Na§ § and Ca2+-Ca2§ exchange are reversible partial reactions of the exchange. As such, it should be possible to isolate them (see Fig. 10). These partial reactions do not exist as elementary steps in the simultaneous mechanism. Partial reactions of the exchanger have been observed (Hilgemann et al., 1991 b). Using giant sarcolemmal patches, rapid application of Na § to the intracellular surface caused binding and translocation of intracellular Na § to the pipette side. This produced a transient current which could be blocked by exchange blockers including the exchanger inhibitory peptide (XIP) (Li et al., 1991). Similar currents were measured in giant oocyte patches in which the cloned exchanger hadbeen expressed. C~ntrol oocytes did notexhibit this current. It appears that charge translocation is associated with Na § translocation and that the Na § translocation step can be isolated, consistent with the idea that a ping-pong mechanism is operative. Further support for a consecutive model of exchange comes from work by Niggli and Lederer (1991a). These authors measured very small transient Ca 2+ currents induced by the photo release of caged Ca 2+ DM-dinitrophen. Insofar as these currents could be inhibited by known blockers of the Na § Ca 2+ exchange, they appear to be associated with exchange. Since they are unaffected by Na + they are presumably associated with a partial reaction of the exchange. These authors have speculated that this current represents a charge move-
50 Out
E'
9
E' .Ca
Na 3
k_1 i
1 O0 I
-100
j
I
,,
-50
A
1
0
i
I
k-4
50
k2
k-2
i
i
fr~ CmV) F I G U R E 1 1. Flattening and saturation of inward INa_Ca-VOltagerelationships when extracellular [Na+] is increased from 100 mM to 400 mM. Na+-MES was replaced by Ca2+-MES in the pipette, as osmolarity for all extracellular solutions was equal. 20 txM free [Ca 2§ on the cytoplasmic side. INa-Ca magnitudes in each current-voltage relationship were normalized to values at-100 mV.
E" Ca
E"- N a 3
k-a In
F I G U R E 12. cation.
Diagram of the Ping-Pong mechanism of ion translo-
295
19. Na+-Ca2+ Exchange Currents
ment associated with Ca 2t binding and the associated conformational change of the exchange molecule. This suggests that (as we have already seen) a voltage-dependent step might be associated in some way with the partial reactions leading to Ca 2t translocation. It is becoming clear that we as yet do not know exactly how many charge-translocating steps are actually in the exchange pathway. However, evidence is accruing that the mechanism is consecutive in nature. Two additional pieces of evidence suggest that a consecutive mechanism could account for Nat-Ca 2t exchange. First, Hilgemann and colleagues have demonstrated that the apparent affinity of one ion for the exchanger is a function of the concentration of the other (1991 a). This is a requirement of a sequential reaction scheme. Moreover, Kimura has employed classic enzyme kinetics to outward exchange currents under assumed zero trans conditions. Her most recent results suggest that while it is difficult to discriminate between a simultaneous and consecutive reaction, the available evidence favors a consecutive reaction (Li and Kimura, 1991).
145 mM
Na o
I
A ~
,
~
:7_- ._ :--_
O--
:7 ......
~
B
""-'-------------"---: :
ItflA .5
f 3.0~
Xli. Na+-Ca 2+ Exchange Currents during the Cardiac Action Potential
0
-,o__J During every action potential, a Ca 2+ current is activated and a modest quantity of Ca 2+ enters the cell. As we shall discuss, it now seems likely that reverse exchange takes place, which also causes some Ca 2+ entry. During a steady-state train of action potentials, this continual Ca 2+ entry must in some way be compensated. It is therefore likely that forward exchange during the repolarizing phase of the action potential extrudes the Ca 2+ entering during the initial phase of the action potential and thus maintains a beat-to-beat homeostasis. Direct evidence that Ca 2t entering the cell through Ca 2t channels can be extruded by the Nat-Ca 2+ exchange has been obtained by Bridge and colleagues (1990). Guinea pig ventricular cells treated with caffeine can be tetanized in the absence of extracellular Na t with a voltage-clamp pulse from -40 mV to +10 mV. During this clamp an inward Ca 2+ current can be measured (Fig. 13). Although the SR is depleted by caffeine, the enlarged Ca 2+ current can apparently produce sufficient Ca 2+ entry to cause contractures. It should be emphasized that in these experiments the pipette contained no Na + so that in the absence of extracellular Na +, Ca 2+ entry by reverse exchange was unlikely. At the peak of the contracture, rapid application of extracellular Na + produced prompt mechanical relaxation and activated an inward transient current. This current was most likely due to forward Na+-Ca2+ exchange because it could not be activated when intracellular Ca 2+ was buffered with EGTA. ff the Na+-Ca2+ exchange current extruded all the entering Ca 2+, and if we further assume that three Na + ions exchange with a single Ca 2+ ion, it follows that the integral of the exchange current is one-half that of the Ca 2+ current. The relationship between the integrals of exchange current and Na+-Ca2+ exchange current was best explained by assuming that three Na + ions exchange with a single Ca 2+ ion and that the exchange extruded all the entering Ca 2+. It seems therefore that the Na+-Ca2+ exchange does have the capacity to extrude all Ca 2+ entering during the duty cycle. It now seems likely that the activity of Na+-Ca2+ exchange is profoundly modified by the cardiac action potential. It is also
o
l
mV
-,o I s
F I G U R E 13. Two-second voltage-clamp pulses in the absence of Na~247 and presence of 10.0 mM caffeine cause contraction. Rapid application of Na~ 500 ms after repolarization causes relaxation. (A) /Ca elicited by membrane depolarization. (B) The application of Na§ produces putative transient inward INa-Ca"(C) This current is displayed on an expanded scale. (D) Contraction (cell shortening) activated by /Ca
recorded in A. (E) After repolarization to -40 mV, relaxation does not occur until 500 ms after the clamp pulse when Na+ is suddenly applied. [Reprinted with permission from Barcenas-Ruiz, L., Beuckelmann, D. J. and Wier, W. G. (1987). Sodium-calcium exchange in heart: membrane currents and changes in (Ca2+)i. Science 238, 1720-1722. Copyright 1987 American Association for the Advancement of Science] likely that the exchange current in part determines the duration of the cardiac action potential. The first comprehensive discussion of this topic was made by Mullins (1979). We will consider the behavior of the Nat-Ca 2+ exchange during the ventricular action potential. The guinea pig ventricular action potential is approximately 300 ms in duration. During the initial part of this action potential Ca 2t is released from the sarcoplasmic reticulum and this causes a rise of Ca 2+ in the cytosol. Peak values for cytosolic free Ca 2t are probably 1-2 IxM. The Ca 2+ transient rises to a peak in approximately 50 ms, whereas the peak of the upstroke of the action potential occurs in approximately 2 ms. After the upstroke of the action potential, but before intracellular Ca 2+has risen appreciably, the membrane potential becomes positive to the exchange reversal potential, which is about -50 mV. Therefore, an outward exchange current is generated, and this will be accompanied by Ca 2t entry and Na + exit. This current has recently been implicated in the "triggering" of SR Ca 2t release (Leblanc and Hume, 1990). As the Ca 2t released from the SR begins to rise and the membrane begins to repolarize, the membrane potential will become negative to the reversal potential, the exchange current will reverse its direction, and Ca 2t will be extruded in exchange for Na t. As the
296
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
intracellular Ca 2+ begins to decline, inward exchange current will also decline to resting values. The same principles presumably govern the behavior of the atrial cell. Earm and Noble (1990) have modeled the time course of Ca 2+ current, Na+-Ca2+ exchange current, and the Ca 2+ transient during the rabbit atrial action potential (Fig. 14). The Na+-Ca2+ exchange current is the main depolarizing current during the plateau. Moreover, the exchange activity required to maintain the late plateau is precisely sufficient to balance Ca 2+ influx (Ca 2+ current + Na+-Ca2+ exchange) during the early part of the action potential. It should be appreciated that, regardless of the value of intracellular Ca 2+, membrane repolarization will tend to stimulate inward exchange current. On the other hand, the decline of the Ca 2+ transient will tend to reduce exchange current. Therefore the relationship between the Ca 2+ transient and membrane repolarization will largely determine the time course of inward exchange current and therefore the pattern of Ca 2+ extrusion.
Xlll. N a + - C a 2. E x c h a n g e Currents a n d Excitation-Contraction Coupling Contractions occur in heart ceils when Ca 2+ released from intracellular stores known as the sarcoplasmic reticulum (SR) activates the contractile elements (for a more detailed account of excitation-contraction the reader is referred to Bers (1991)). This release is coupled to electrical excitation at the surface membrane and the whole process is often referred to as excitation-
5o. A
>0
E
O >
=
~16o
z~o
"~ ~
'KI
-400 ' iCa 0.5 C
9- ~ -
O
o.
28o
.
0
80
Time (ms)
280
FIGURE 14. The Earm-Noble model of the single rabbit atrial cell, based on the multicellular model of Hilgemann and Noble (1987). (A) Computed action potential. (B) Computed currents. (C) [Ca2+]iand contraction (Earm and Noble, 1990). (Reprinted with kind permission of the author and the Royal Society.)
contraction coupling. The pioneering studies on skinned fibers by Alexander Fabiato lead to at least two fundamental findings (for a discussion of these see Stem and Lakatta (1992)). The first was that Ca e+ can be released from the SR when the concentration of Ca 2+ in the vicinity of the SR is abruptly increased. This small increase in the concentration of Ca 2+ leads to a much larger release of SR Ca 2+, so the system is one of inherently high gain. This process is usually referred to as CaZ+-induced Ca 2+ release or CICR. A second and important property of CICR is that, under normal circumstances, the release is graded with the size of the Ca 2+ increase that induces the release. A priori such a system might be expected to be regenerative since the Ca 2+ that is released from the SR ought to stimulate further release. A discussion of why SR Ca 2+ release in heart is not normally regenerative is beyond the scope of this chapter, which nevertheless will discuss some of the evidence that suggests that Na +Ca 2+ exchange currents are involved in CICR. It is now known that large tetrameric CaZ+-release channels (sometimes referred to as ryanodine receptors) are embedded in the sarcoplasmic reticulum (Saito et al., 1988). These apparently respond to elevations of Ca 2+ in their vicinity by gating the release of Ca 2+ from the SR (Stem and Lakatta, 1992). Therefore a molecular basis for the early observations by Fabiato has been established. A central question for those studying excitation-contraction in heart is, What in intact cells produces the rise in intracellular Ca 2+ that gates the SR release channel and triggers SR Ca 2+ release? It is now well established that the L-type Ca 2+ current is principally involved in triggering SR Ca 2+ release in mammalian ventricular cells (Lopez-Lopez et al., 1994; London and Krueger, 1986; Beuckelmann and Wier, 1988; Cheng et al., 1993). Because SR Ca 2+ release is graded with the size of the increase in Ca 2+ concentration that induces the release, it follows that if L-type Ca 2+ channels trigger or induce the release of Ca 2+, the extent of SR Ca 2+ release also should be graded with the size of the L-type Ca 2+ current. Since the size of the Ca 2+ current has a bell-shaped dependence on voltage, the finding that the rate or extent of SR Ca 2+ release (or the magnitude of triggered contractions) is (under appropriate conditions) also bell-shaped lends strong support to the idea that the Ca 2+ current is a trigger for SR Ca 2+ release. For example, a detailed study by Beuckelmann and Wier (1988) has clearly established the bell-shaped relationship between triggered Ca 2+ transients and voltage. However, a number of studies in ventricular cells have revealed, under certain circumstances, a more complex relationship between the voltage dependence of triggered contractions and Ca 2+ current (Nuss and Houser, 1992; Litwin et al., 1998; Vomanen et al., 1994). In particular, tension measurements did not follow a simple bell-shaped relationship with voltage but rather showed a sigmoid relationship. At positive potentials shortening did not decline steeply with voltage. Studies by Litwin and coworkers (1998) on guinea pig cells under voltage-clamp and dialyzed with various Na + solutions indicate that while the shape of the Ca 2+ current/voltage relationship was independent of pipette Na +, the triggered shortening-voltage relationship showed a striking dependence on pipette Na + (Fig. 15). At positive potentials the shortening-voltage relationship departed from a simple bell shape and the extent of this departure depended on the concentration of dialyzing Na +.
297
19. Na+-Ca 2+ Exchange Currents
If L-type Ca 2+ current is blocked extremely rapidly, triggered contractions are reduced but not abolished (Levi et al., 1996). Thus if Ca 2+ currents are elicited under voltage-clamp it is possible to record what could be a triggered contraction or a triggered Ca 2§ transient. It therefore seems likely that another process besides the Ca 2+ current is involved, triggering the release of SR Ca 2+. The most likely process is in fact the Na +Ca 2+ exchange. Recently Grantham and Cannell (1996) used voltage-clamp pulses shaped like an action potential to infer the magnitude and trajectory of Na+-Ca 2+ exchange currents during the initial part of an action potential. They found that the magnitude of the exchange current was somewhat less than 30% of the magnitude of the Ca 2§ current that occurred during the initial part of the action potential (Fig 16). As the authors point out, this is not a negligible current and is consistent with the idea that at least some of the triggered SR Ca 2+ release could be due to the activity of the Na§ 2§ exchange. It is worth mentioning that if the Na+-Ca 2§ exchange is capable of contributing part of the trigger for SR Ca 2+ release, then this component of the trigger will be extremely sensitive to intracellular Na +. In this regard it has been proposed that the initial Na + current might provide enough Na + accumulation (provided the accumulation takes place in a restricted space) in the vicinity of the Na+-Ca 2+ exchangers to enhance triggering by reverse Na+-Ca 2+ exchange current (Leblanc and Hume, 1990; Lipp and Niggli, 1994). A difficulty with the idea that exchange can under physiological circumstances trigger significant Ca 2§ release is that it is extremely small. Litwin and colleagues (1998) have suggested, for example, that the exchange may sum its effects with the Ca 2§ current in a highly nonlinear way to augment triggering by Ca 2§ current. It is possible that one function of the exchange is to increase the probability with which a Ca 2+ current triggers exchange in a highly nonlinear system. It is therefore not yet clear what contribution Na+-Ca 2+ exchange currents make to triggering SR Ca 2+ release under physiological circumstances. However, since the exchanger is regulated by intracellular Na § Ca 2§ and voltage, it seems
likely that its contribution to triggering (either direct or indirect) will be both complex and variable.
XIV. Summary At first, study of the Na+-Ca2+ exchange current was impeded because multicellular preparations did not permit adequate control of the driving forces producing exchange. However, methods for isolating single cells together with the patch-clamp technique made it possible to isolate exchange current and to provide reliable measurements of its properties. Now that the exchange molecule has been cloned, it is possible to express the mammalian exchanger in other cell types including frog oocytes. Currents may now be measured in these cell types with the giant
A
Vm ~]
9 Nifedlpine ~
Control
50 mV
. . . .
im
o u c o o
aO=
3.0 O
L. LO
2.0
200 ms
B 9# bumetanide > piretanide > furosemide > amino-piretanide. Besides their clinical importance, these substances, particularly bumetanide, have played a crucial role in the identification of the NKCC proteins and their subsequent cloning (Xu et al., 1994). Moreover, they have provided information that has been used to interpret ion binding to the NKCC. Bumetanide is by far the most commonly used loop diuretic for research on NKCC. It reversibly inhibits the NKCC in a variety of preparations by a concentration-dependent mechanism, with a half inhibitory constant (IC50) of NI• -7 M (Russell, 2000). NKCC2 is more sensitive to bumetanide than NKCC 1 (Isenring et al., 1998). However, none of these compounds are specific toward the NKCC. Thus, used alone and/or in relatively high concentrations, not even bumetanide can provide positive proof that a given function is mediated by the NKCC. Other C1- transport systems that can be inhibited by loop diuretics include the C1--HCO 3 exchanger, Ca 2§ activated and GABA-gated C1- channels, and KCC (for references see Alvarez-Leefmans, 1990 and Russell, 2000). Loop diuretics are anionic in the physiological range with very high lipid solubility and hence they do cross cell membranes (Alvarez-Leefmans et al., 1990). This raises the question of their site of action. It has been suggested but not proven that bumetanide binds exclusively to the external face of the NKCC (Haas, 1994), but whether it acts also from the inside remains to be determined.
B. M o l e c u l a r S t r u c t u r e a n d Distribution of Na§247 - Cotransport Proteins
Presently, two distinct Na+-K+-C1- cotransporter isoforms have been identified by cDNA cloning and expression: NKCC1 (also referred to as BSC2 or CCC1) and NKCC2 (also known as BSC1 or CCC2). A cDNA encoding NKCC1 was first cloned from the shark rectal gland, a well-characterized secretory epithelium with high density of Na+-K+-C1- cotransporters in its basolateral membrane (Xu et al., 1994). Subsequently, cDNAs encoding NKCC1 have been cloned from other secretory epithelia, including
309
20. Intracellular Chloride Regulation
COOH A
COOH
NH2SO2~
NH2SO
NH(CH2)3CH3
O
FurosemideZ~ COOH
CI
@ Bumetanide@
NH2SO~
N
COOH CONH2
~
Piretanide9
NH2SO2~~L NHCH2- ~ O
@ Benzmetanide A NH2SO2~ amino-PiretanideO
B 100 -
o 80m n
O t_ 6 0 O
L)
40-
20I
I
I
I
I
I
I
I
I
"10
"9
"8
"7
"B
"5
"4
"3
"2
Log Ligand Concentration FIGURE 7. (A) Structures of several 5-sulfamoyl benzoic acid derivatives that affect NKCC in flounder intestine. (B) Doseresponse curves for these derivatives on short-circuit current measured in flounder intestine. Order of potency is benzmetanide (5 x 10-8 M; n = 4), bumetanide (3 x 10-7 M; n = 7), piretanide (3 x 10-6 M; n = 4), furosemide (7 x 10-6 M; n = 4), aminopiretanide (1 x 10-5 M; n = 4). (Modified from O'Grady et al., 1987.)
T84 cells, a human colonic epithelial cell line (Payne et al., 1995), and cultured inner medullary collecting duct cells (mIMCD-3) from mouse kidney (Delpire et al., 1994). Both of these mammalian NKCC1 proteins share a high degree of homology with the shark NKCC1 (71-74%) and even higher identity with each other (91%). More recently, a cDNA encoding the bovine aortic endothelial cell NKCC was also cloned and sequenced, and the predicted protein exhibits ~95% amino acid identity with the T84 cell cotransporter sequence (Yerby et al., 1997). NKCC 1 comprises ~ 1200 amino acids and has a molecular weight of 130-132 kDa. The predicted secondary structure for NKCC1 from shark rectal gland shows 12 ce-helical transmembrane (TM) domains and hydrophilic amino- and carboxy-terminal regions flanking a central hydrophobic domain (Fig. 8). Whereas the amino terminus displays considerable variability, the large carboxyl terminus is well con-
served among the different NKCC proteins. All identified Na+-K+-C1 - cotransporters are glycoproteins and possess consensus sites for N-linked glycosylation within a large hydrophilic (extracellular) loop between TM segments 7 and 8, as shown in Fig. 8 (Haas and Forbush, 1998). In fact, the protein in its physiological plasma membrane environment is glycosylated. It is believed that glycosylation enhances cell-surface expression of these proteins. In addition, NKCC1 has several phosphorylation sites within the predicted amino- and carboxy-terminal domains (labeled "P" in Fig. 8). Northern blot analysis has revealed the presence of NKCC1 in a wide variety of tissues, including salivary gland, stomach, lung, trachea, and pancreas, all possessing secretory epithelia. This analysis as well as immunohistochemistry and in situ hybridization also demonstrated the presence of NKCC1 m R N A in nonepithelial tissues such
310
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
TM1 NH 2
COOH
111
FIGURE 8. Model of the shark rectal gland Na+-K+-2CI- cotransport protein (NKCC1), based on its cDNA sequence and hydropathy and secondary structural analysis. Each link symbolizes a single amino acid residue (1191 total). "NH2" and "COOH" indicate amino- and carboxytermini, respectively. Branched lines indicate potential glycosylation sites in the extraceUular loop formed between putative transmembrane (TM) segments 7 and 8. Biochemically identified phosphorylated threonine (Thr) residues (positions 189 and 1114) are indicated by open diamonds with an adjacent "P." The proposed secondary structure of this protein with 12 TM helices and large intracellular amino- and carboxyterminal domains is shared by other members of the cation-chloride cotransporter (CCC) superfamily that have been cloned and sequenced (see Fig. 13). (Modified from Haas and Forbush HI, 1998.)
as heart, skeletal muscle, and vertebrate sensory neurons and axons, as well as in Schwann cells (Delpire et al., 1994; Xu et al., 1994; Payne et al., 1995; Plotkin et al., 1997; Sung et al., 2000; Alvarez-Leefmans et al., 1998, 2001). Immunofluorescence studies using anticotransporter antibodies (Lytle et al., 1995) have shown that in intact secretory epithelia, NKCC1 protein is localized in the basolateral membrane. The exception is the rat choroid plexus (CP) where NKCC1 is located in the apical membrane (Plotkin et al., 1997). The cotransporters that are localized in the basolateral membrane mediate a net salt influx into the cells, and in doing so act in concert with apical CI-channels and basolateral K + channels to produce net salt and fluid secretion in these epithelia. The location of NKCC1 in the apical membrane of CP argues against its involvement in ion secretion into cerebrospinal fluid (see Section IV.D). It has been proposed that the CP cotransporter is constitutively active and it functions in series with Ba2+-sensitive K § channels to reabsorb K § from cerebrospinal fluid to blood (Wu et al., 1998). The location of NKCC in the paranodal region of the Schwann cells suggest a similar role in K + uptake from
the peri-axonal space, thereby preventing K + accumulation which could produce abnormal changes in axonal excitability (Alvarez-Leefmans, et al., 2001). The NKCC2 isoform has thus far been identified exclusively in the kidney (Gamba et al., 1994; Mount et al., 1998). NKCC2 shares only---60% amino acid sequence identity with NKCC1. In fact, these two proteins are the products of independent genes. In the mouse, Delpire and his coworkers (1994) have localized the mNKCC1 gene to chromosome 18, whereas the mNKCC2 gene has been localized to chromosome 2. Studies in rabbit kidney revealed that there are at least three splice variants of N K C C 2 u A , B, and F (Payne and Forbush, 1994)--all of which are differentially distributed within the rabbit kidney: one form restricted to the cortex (variant B), one restricted to the medulla (variant F), and a third equally distributed between both the cortex and the medulla (variant A). It is now known that the three splice variants are likely to be a general feature of the mammalian kidney (Gamba et al., 1994). Additional splice variants of mouse NKCC2 have been recently reported (Mount et al., 1999a). The functional role of all these splice variants is not well understood but it is conceivable that they underlie heterogeneity of ion transport in the thick ascending limb of the loop of Henle (TALH). NKCC2 is a protein smaller than NKCC 1, with a deduced amino acid sequence of--1100 residues and a molecular mass of 120-121 kDa, which is consistent with the size (150 kDa) of the glycosylated transporter. The difference between the two isoforms can be accounted for almost entirely by the additional 80 amino acids at the amino terminus of NKCC1. In fact, the amino terminus is the most divergent region of the entire protein both between the two isoforms and comparing the same isoform between species (Haas and Forbush, 1998). NKCC2 is located at the apical membrane of the epithelial cells of the thick ascending limb of the loop of Henle (TALH), the main pharmacological target of the loop diuretics. NKCC2 plays a central role in transcellular absorption of Na § and C1- by the medullary and cortical TALH and plays a secondary role in the paracellular transport of Na § Ca 2+, and Mg 2§ by this nephron segment (Mount et al., 1998). NKCC2 is considered one of the major proteins involved in salt and fluid reabsorption in the kidney; 15-25% of filtered Na § is reabsorbed in the TALH, most of it via the NKCC2. Salt absorption by the TALH is crucial to countercurrent multiplication in the renal medulla and the excretion of concentrated urine. The TALH also functions in renal acid secretion, since NH~ can compete with the K+-binding site on the NKCC2 (Good, 1994). The NH~ generated by the renal proximal tubule is thus reabsorbed from the tubule lumen in the TALH and excreted from the renal interstitium by more distal nephron segments. NKCC2 is also located in the apical membrane of the macula densa cells (Obermiiller et al., 1996; Mount et al., 1998). The macula densa is a specialized group of tubular cells situated at the point of apposition of renal tubules with their parent glomeruli. These cells control two important physiological processes, tubuloglomerular feedback and tubular regulation of renin release. An increase
20. lntracellular Chloride Regulation
3 11
in luminal Na+-C1 - activity at the macula densa decreases glomerular filtration by inducing constriction of the afferent renal arteriole. Increases in luminal Na+-C1 - activity at the macula densa also inhibit renin release from juxtaglomerular cells in the afferent arteriole. Thus N K C C 2 plays a crucial role in the regulation of fluid volume and osmolarity.
C. A Kinetic Model of Na+-K§
CI- Cotransport
Normal operation of the vertebrate N K C C requires the binding to the cotransporter of all the ions involved (i.e., 1 Na +, 1 K +, and 2 C1-) on the same side of the membrane before ion translocation occurs. The order of binding of the cotransported ions does not seem to occur at random. Evidence shows a preferred order of ion binding resulting from an allosteric effect such that the binding of one ion increases the apparent affinity of a binding site for the next ion. The besttested model to date is one based on a systematic study of ion dependencies of self-exchange fluxes mediated by the cotransporter in duck and human red blood cells (Duhm, 1987; Haas, 1994; Lytle et al., 1998). Two partial reactions of the cotransported cycle have been revealed: K+-K + exchange in normal high-K + cells and Na+-Na + exchange in high-Na + cells. The internal and external ion requirements of each mode can be explained by a reaction cycle based on ordered binding and glide symmetry. The latter means that the first ion to bind on one side of the "carrier" will be the first ion to be released on the other.
Figure 9 shows the cotransport model based on ordered ion binding and glide symmetry. Net cotransport influx would require that the ions bind to the cotransporter in the order shown. In the initial loading of the empty carrier at the outer face of the membrane, binding of a Na + in the pocket (reaction 2) induces formation of a C1- site, and binding of that C1- (reaction 3) creates a site that can bind K +, and so forth. All four ions are occluded momentarily in a transitional state (E4) without access to either side of the membrane. Reaction 5 represents a conformational change of the N K C C such that the binding sites are now accessible from the intracellular compartment, and reactions 6-9 represent the ordered release of ions. Reaction 10 leads to a conformational change that involves the reorientation of the totally unloaded cotransporter to an outwardly facing conformation. The model has been quantitatively tested using simulations based on flux data from duck red blood cells, HeLa cells, and epithelial cells (Benjamin and Johnson, 1997). A critique of the model showing its weaknesses and strengths has appeared recently (Russell, 2000).
D. Functions of the Na+-K+-2 CI- Cotransport The net transport of Na +, K +, and C1- can serve a number of physiological functions depending on the cell type in question. Some of these functions have already been highlighted in Section IV. The central role that N K C C plays in net salt and fluid secretion across absorptive and secretory
FIGURE 9. Hypothetical model of ordered ion binding with glide symmetry of Na§ - cotransport. Ten intermediate states of the transporter are depicted. The model assumes that cotransport is completely reversible and that only fully loaded (E4) and completely unloaded (E0) forms of cotransporter protein are capable of changing orientation from inward- to outward-facing (and vice versa). The idea is that when all sites are occupied (E4), a major conformational change occurs, allowing bound ions to unload on the opposite side of the membrane in the order they were bound (first on, first off). When the carrier is completely unloaded (E0), the empty form can also undergo a major conformational change, allowing it to begin reloading ions at the opposite side of the membrane, thus facilitating net cotransport. Numbers refer to order of binding, reorientation, and release for one complete influx cycle. (Modified from Lytle et al., 1998.)
312
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers A
Chloride Secretion
._____J ..
~
Na +
.,
membrane ] (Lumen) [
-. Na +"
/
[
'=~ " K + ~,) .. +
-""NBasolateral ~ membrane
I
"(~*
B
I
Chloride Absorption
Apical membrane (Lumen) Na +
~
.
~
~ 1~
K§ --
"a +
~"J
c,-
|
| I I
I
~ t
Basolateral membrane .
I
FIGURE l 0. Mechanism for active CI- absorption and secretion in C1- transporting epithelia. (A) Transport model for exocrine gland CI-secreting epithelium (e.g., salivary glands or trachea epithelium). Intracellular [CI-] is maintained above electrochemical equilibrium by NKCC across the basolateral membrane. Na+ influx is balanced by efflux via Na+,K+-ATPase; K+ influxes are balanced by effiux via K+ channels. Transepithelial secretion is induced by activation of apicalmembrane C1- channels. (B) Model of C1- transport in thick ascending limb of Henle's loop, a C1--absorptive epithelium. C1- entry across the apical membrane is mediated by NKCC. Most basolateral C1-effiux proceeds through C1-channels. (Modified from Reuss, 1997.)
epithelia has already been mentioned. In cells of C1-absorptive (e.g., distal renal tubules) and Cl--secretory (e.g., salivary glands) epithelia, NKCC serves as the major CIentry pathway and functions in concert with three other transport proteins, namely Cl- and K + channels and the Na +K+pump, to carry out net transepithelial movements of salt (Fig. 10). The possible involvement of NKCC in cell volume regulation has also been mentioned. In some cells, NKCC appears to be involved in regulatory volume increase in response to osmotic cell shrinkage, by promoting net gain of K +, CI-, and osmotically obligated water (see Chapter 21 and Russel (2000) for a more detailed account). Another role of NKCC is to maintain [C1-]i at higher than equilibrium values in vertebrate sensory spinal neurons, making possible the phenomenon of presynaptic inhibition. In vertebrate spinal cord, presynaptic inhibition is primarily due to depolarization of primary afferent nerve terminals by the neurotransmitter GABA (y-aminobutyric acid), which leads
to a reduction in neurotransmitter output (Rudomin and Schmidt, 1999). The prevailing idea is that GABA released from interneurons depolarizes primary afferent fibers via axoaxonic synapses. These depolarizations inactivate Na + channels sufficiently to block action potential invasion into the primary afferent terminals, thereby inhibiting transmitter release. The depolarizing inward current results from opening of GABAA-gated C1- channels, with the consequent effiux of C1-which transiently drives the transmembrane potential toward the equilibrium potential for C1-. This peculiar depolarizing action of an inhibitory transmitter is possible because the [C1-] inside the nerve terminals of primary sensory neurons is higher than predicted from a passive distribution, thereby providing the necessary driving force for electrodiffusional C1- effiux. The nonpassive distribution of C1- is generated and maintained by NKCC (Alvarez-Leefmans et al., 1988, 1998; Sung et al., 2000). Moreover, due to their small volume/surface ratio, presynaptic nerve terminals are expected to be partially depleted of solutes (C1- and K*) and water upon GABA action. It has been postulated that NKCC is capable of restoring ionic gradients and osmotic balance altered as a consequence of neurotransmitter action (Fig. 11). The Na+-K+-C1-cotransporter present in sensory neurons is similar to that characterized in the plasma membrane of a variety of non-neuronal cells, and has been recently identified with immunofluorescence techniques with anticotransporter antibodies (Plotkin et al., 1997; Alvarez-Leefmans et al., 1999, 2001; Sung et al., 2000). It has been found that the expression of NKCC is developmentally regulated in postnatal rat brain (Plotkin et al., 1997; Clayton et al., 1998). Consistent with this finding, in embryonic neurons, GABA has a widespread depolarizing action (Ben-Ari, 1994; Cherubini et al., 1991). The latter seems crucial in promoting Ca 2+ influx (Yuste et al., 1991; Obrietan and van den Pol, 1996), influencing important developmental events such as neuronal proliferation, differentiation, and migration and neurite extension and targeting (LoTurco et al., 1995; Obfietan and van den Pol, 1996). Again, the depolarizing effect of GABA results mainly from an effiux of C1through GABAA-gated anion channels. This outward C1-current is possible because the equilibrium potential for chloride (Ecl) is kept at a more positive value than the transmembrane potential (Era) by means of an inwardly directed NKCC that actively accumulates CI-. NKCC exhibits some of the features required for an efficient extracellular K + buffer; that is, it has a strong net inwardly directed driving force and a high affinity for external K + (Payne et al., 1995). In fact, there is some evidence for its role as a K+-uptake mechanism that contributes to buffeting and regulation of extracellular K + in many tissues. For instance, NKCC1 is present in the apical membrane of mammalian choroid plexus and in concert with other transport systems coexisting in these polarized cells, plays a central role in the reabsorption of K + from the cerebrospinal fluid to the blood, thereby contributing to the buffeting and regulation of brain interstitial [K +] (see Xu et al., 1998, and Fig. 12). NKCC is also prominently located in the paranodal region of Schwann cells of myelinated axons, a region prone to extracellular K + accumulation resulting from axonal impulse activity (Chiu, 1991). This suggests the possible involvement
20. lntracellularChlorideRegulation
3 13 Furosemide 3 Na+ Bumetanide
Ouabain. . ~ i ~ L Piretanide.~ .,,,~'~J"',~ ~ 2 CI" //~T,P~ 1 ~ ~ / / / N a I /
~/~ur21 ~ : : 0 ~ Primary iJp I (~-~ ~ j//."~.__.-.---~ __. afferent M~176176 "~(~~~~\@ ? [ ~+Cl" ~--~/~'~~ fiber
//
,~- "~"~
1
,
i::
o Glutamate A GABA
,'> ~" l
~ %
Aresynaptic
I
axon
F I G U R E 1 1. Schematic representation of a synapse made by a primary afferent on a motoneuron and the axo-axonic contact made by a GABAergic interneuron (presynaptic inhibitory axon). Release of GABA from the presynaptic inhibitory axon causes an effiux of C1- from the primary afferent ending, resulting in a depolarization that is "recorded" by a microelectrode inserted into the terminal (lower left). The depolarization activates K + channels with consequent K§ effiux. The outwardly directed C1gradient is generated and maintained by Na+-K§ - cotransport. Putative osmotic water fluxes accompanying GABA-induced ion movements are indicated with dashed arrows. (Reproduced from Alvarez-Leefmans et al., 1998.)
of N K C C in K + uptake from the periaxonal space, thereby regulating the external [K +] and maintaining the axonal excitability in this critical location (Alvarez-Leefmans et al., 2001).
CSF
BLOOD
V. Electroneutral K+-CI - Cotransporters The K+-C1- cotransporter (KCC) is an integral membrane protein that mediates the obligatorily coupled, electroneutral movement of K + and C1- across the plasma membrane of many animal cells. As an electroneutral secondary active transport mechanism, the direction of the net movement of K § and C1- by the cotransporter is determined solely by the sum of the chemical potential gradients of the two ions (Lauf and Adragna, 1996). Under normal physiological conditions, the outwardly directed K § chemical potential gradient maintained by the Na§ + p u m p drives C1- uphill against its chemical potential gradient. Hence, under normal conditions, the K C C is an effiux pathway for K + and C1-. However, the cotransporter is bidirectional and can mediate net ion effiux or influx, depending upon the prevailing K + and C1-chemical potential gradients. Although the stoichiometry has actually never been determined, it is believed that the transport process involves one-for-one m o v e m e n t of K § with C1-. Whether it is (n + 1) K § : (n + 1) C1- remains to be determined. This means that the overall transport process of K §
3Na+~C~)I ~I~2K+
Cl'
10 kcal-mol -l, a value typical of artificial lipid bilayers (see Table 2). Mercurials inhibit water transport primarily by targeting protein SH groups because their effect is fully and rapidly reversed by cysteine. The second inhibitor is radiation. High doses of radiation inhibit water transport in both erythrocytes (van Hoek et al., 1992) and renal brushborder membrane vesicles (van Hoek et al., 1991). The characteristics of radiation inactivation suggest that the target is the size of a 30-kDa protein and are inconsistent with the entire membrane or transient defects serving as the major water pathway. Although these studies and others, especially in epithelia (e.g., Verkman, 1993), made it clear that water pores or channels must exist, until recently their identity and characteristics remained mysterious. Are what we call "water channels" simply water moving through open ion channels or through an ion-exchanger or ion-cotransporter? Are water channels specific, admitting water but excluding ions? These questions have been answered with the cloning and expression of a family of water channels called aquaporins
(Agre et al., 1993; King and Agre, 1996; Verkman et al., 1996; Heymann et al., 1998; Nielsen et al., 1999).
1. Aquaporins Water channels are related to the membrane integral protein (MIP) family (20-40% homology). The first water channel identified was cloned from a human bone marrow library by Preston and Agre (1991) based on the sequence of a purified protein of uncertain function. Initially called CHIP28 (channel-forming integral protein, 28 kDa), this protein was later redesignated aquaporin-1 (AQP1), and MIP-1 is referred to as AQP0. Ten mammalian aquaporins are known, and more than 100 related proteins have been found in amphibians, Drosophila, plants, E. coli, and yeast, but not all of these conduct water (Heyman et al., 1998; Nielsen et al., 1999; Heymann and Engel, 1999). The cloned AQP protein exhibits all of the characteristics of a water channel. AQP1 expressed in Xenopus oocytes induces up to 30-fold increases in PT that is blocked by HgC12 (Preston et al., 1992). Reconstitution of purified AQP1 protein in liposomes verified that AQP1 itself, rather than modulation of an endogenous oocyte membrane protein, was responsible for mercurial-sensitive water permeation (Zeidel et al., 1992). Furthermore, incorporation of AQP1 into liposomes reduced the activation energy of Pf from 16.1 kcal-mo1-1, characteristic of permeation through the bilayer, to 3.1 kcal.mol -~, characteristic of water passing through water-filled pores. Moreover, AQP1 appears to be active and highly selective for water without requiting regulatory subunits or cofactors (Preston et al., 1992; Zhang et al., 1993). There are about 2 • 105 AQP1 molecules per erythrocyte (Zeidel et al., 1992). Taking the erythrocyte's membrane area as 1.35 • 10 -6 cm 2 (Solomon, 1989), this corresponds to a density of about 1.5 • 1011AQP1 cm -2 or 1500 AQP1 pm -2. Despite this remarkable density of pores, the water permeability of a human erythrocyte is only -10 to 50 times greater than that of a phosphatidylcholine/cholesterol bilayer (Fettiplace and Haydon, 1980). Thus, 30-150 aquaporin channels are needed to equal the permeability of 1 pm 2 of bilayer. AQP1 contains 269 amino acid residues and was postulated to have 6 membrane-spanning domains based on an analysis of hydrophilicity (Preston and Agre, 1991) and selective proteolysis of protein loops that face the intra- or extracellular side (Preston et al., 1994). The topology for the AQP 1 and the other members of the AQP family is shown in Fig. 12. Mercurial-inhibition, N-glycosylation, and PKA phosphorylation sites have been identified. There is an internal homology between the halves of AQP designated repeat-1 and -2. The greatest homology among AQPs is in the segments surrounding asparagine-proline-alanine (NPA) motifs in loops B and E. These loops are thought to be arranged anti-parallel and are postulated to dip back into the membrane to form an hourglass-shaped pore represented in the cartoon in Fig. 12 (Jung et al., 1994). It is hypothesized that four highly conserved polar residues, El7, N76, N192, E142, comprise a critical portion of the water permeation pathway (Heymann et al., 1998). The biochemical behavior of solubilized erythrocyte AQP1 suggested that it is a noncovalently-linked tetrameric
21. Osmosisand Regulationof CellVolume
335
structure, and electron microscopy of negatively stained AQP1 confirmed this (Walz et al., 1994). Coexpression of Hg-sensitive and -insensitive recombinant proteins indicates that each AQP monomer forms an independent functioning pore, however (Preston et al., 1993). An exception to the tetrameric design of AQPs is AQP4. AQP4 forms large multimeric square arrays in the end-feet of astrocytes that surround capillaries (Rash et al., 1998). X-ray diffraction patterns at 3.5- and 6-,~ resolution have been obtained from two-dimensional crystalline AQP1 arrays (Jap and Li, 1995; Walz et al., 1997). Electron density contour maps confirmed a tetrameric arrangement of monomers, and each
!
monomer appeared to have a central low density core, presumably the permeation pathway, surrounded by six tilted high density regions thought to represent membrane-spanning a-helicies. Immunohistochemistry, western blot (protein determination), and northern and in situ hybridization and RNase protection assays (mRNA determinations) have identified broad but only partially overlapping distributions of AQP homologs in regions where water permeability is high (Hasegawa et al., 1993; Nielsen et al., 1993; Zhang et al., 1993; Umenishi et al., 1996; Nielsen et al., 1999; Ma and Verkman, 1999). For example, AQP1 is located in erythrocytes, renal proximal
Glycosylation sites (~) loop A-AQPI loop C-AQP2, 3, 4,.....5
]Lo'op E ] /lhemicha, nel !
j ( ~ Colton polymorphism
i v loop A-AQP1. . . . . .
out .
.
.
~
.
/I(~)Hg inhibition site | IoopE-AQP1,2,5
.
# NH2
,
PKA motifs
iPC
,
....
(~) loop B-AQP4 loop D-AQP5 COOH-AQP2
.. repeat-
i
.....
1
A
COOH
repeat-2
_
E w Out
~ln
H2N
HOOC F I G U R E 12. AQP family portrait. Cartoon illustrating the postulated topology of AQP as a structure with six membrane spanning a-helices (top). Cysteine (C) responsible for inhibition by mercurials and consensus PKA phosphorylation (S) and glycosylation (N) sites are also shown. Loops B and E are envisioned to dip into the membrane with the two NPA motifs forming the narrow neck of an hourglass-shaped, water-filled pore (bottom). Single letter codes are standard amino acid abbreviations. The position of the Colton polymorphic blood group antigen (Co) is noted. [Reproduced from King and Agre (1996) with permission, from the Annual Review of Physiology, volume 58, 9 1996, by Annual Reviews, Inc., and from Jung et al., (1994) with permission.]
336
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
tubules and the descending thin limb of the Loop of Henle (but not in the collecting duct, where water permeability is controlled by vasopressin), the choroid plexus, the iris, ciliary and lens epithelia and corneal endothelium of the eye, lung alveolar capillaries and epithelium, red splenic pulp (with erythrocyte precursors), colonic crypt epithelium, and nonfenestrated capillary and lymphatic endothelium in a number of organs including cardiac, skeletal, and smooth muscle. The exception with regard to broad distribution is AQP2. AQP2 underlies the vasopressin-regulated water permeation pathway of the renal collecting duct apical membrane that maintains water balance (Nielsen et al., 1995; Nielsen et al., 1999). Mutation of AQP2 is responsible for inherited nephrogenic diabetes insipidus, and targeting and expression defects may cause acquired forms of this disease. On the other hand, AQP2 expression is upregulated in pregnancy and congestive heart failure, states associated with enhanced water retention (Nielsen et al., 1999). Several cautions are warranted when drawing physiological interpretations from the localization of AQP. For example, AQP1 mRNA is higher in cardiac homogenate than in homogenate from any other organ (Umenishi et al., 1996), but functionally, the story is different. In isolated myocytes, Lp is very low, and the activation energy is high, about 10 kcal-mo1-1 (Suleymanian and Baumgarten, 1996) (see Table 2). Thus, AQPs do not significantly contribute to water transport across the myocyte membrane. Organs comprise cells with diverse functions and often different requirements for water transport. Because of its localization in vascular tissue, AQP detected in homogenates should not be attributed to the principal cells of the organ without confirmation. Even careful immunohistochemistry and in situ hybridization may lead the physiologist astray. The problem is the high density of AQP necessary to significantly affect Pry" If, for example, the density of AQP1 in the erythrocyte membrane was 10 AQP1 pm -2 rather than 1500 AQP1 pm -2, AQP1 would still be detected by modem techniques, but it would make a physiologically insignificant contribution to Pf.
2. Other Channels and Transporters The identification of specific water channels does not exclude the possibility that water flux through ion channels or transporters significantly contributes to the water permeability of the membrane. One interesting example is the cystic fibrosis conductance regulator (CFTR). CFTR functions as a cAMP-regulated C1-channel. Hasegawa et al. (1992) recently found that CFTR expressed in oocytes also acts as a water channel with an estimated single channel Pf comparable to that of AQP1. Both Pf and C1-conductance were increased by cAMP. The effect of the CFTR channel on Pf is large but not unique. Pore-forming antibiotics such as gramicidin, nystatin, and amphotericin, are reported to induce a more modest Pf. Transporters also may contribute to Pf. Fischbarg et al. (1990) found that expression of Na+-independent glucose transporters in oocytes increases water permeability, and Solomon et al. (1983) suggested that some of the water permeability of erythrocytes was contributed by the C1--HCO~ exchanger.
However, water flux mediated by these transporters does not account for the macroscopic properties of water transport.
V. Regulation of Cell Volume under lsosmotic Conditions A. Gibbs~Donnan Equilibrium Because ions exert an effective osmotic pressure, the distribution of ions affects water flow across the cell membrane and thus cell volume. The starting point for understanding these effects is the theoretical ideas of Gibbs that were first demonstrated experimentally by Donnan. The phenomenon is now called Gibbs-Donnan equilibrium or simply Donnan equilibrium. Macknight and Leaf (1977) elegantly describe the history of how these ideas were applied to cell volume regulation, and Overbeek (1956) provides a detailed derivation and considers nonideal solution behavior. In a Donnan system, the membrane permits the movement of small charged solutes (e.g., K § and C1-) between two compartments but restricts the movement of large charged species such as proteins, which are usually polyvalent and negatively charged at intracellular pH. It is the inability of one (or more) charged species to distribute freely between the compartments that profoundly influences the distribution of the mobile ions and, consequently, water. For a cell, it is the cell membrane that restricts the movement of large ions. A membrane is not required to establish a Donnan equilibrium, however. Allthat is needed is a means of restricting one charged species to a single compartment. Donnan equilibria can arise in gels consisting of charged structural components (e.g., an ion-exchange column) because the charges on the gels are fixed in place. Donnan equilibria can also arise in cells after their membranes have been removed, leaving a cytoplasmic "gel" consisting of interconnected cellular proteins. A Donnan system represents a true equilibrium, although as we will see, an equilibrium is not attained under the usual biological conditions. There are three important characteristics in a Donnan equilibrium system: (1) an unequal distribution of ions; (2) a potential difference between the compartments; and (3) an osmotic pressure. These characteristics can be understood from the application of a basic thermodynamic principle to the system. Species that can cross the membrane, including both mobile ions and water, will distribute themselves so that their electrochemical potential (/zi) is the same in both compartments. To understand how a Donnan equilibrium develops, imagine starting an experiment by "filling" the cell and the extracellular space with solutions of concentrations indicated in Fig. 13A. Since ions move down their electrochemical gradients, the first question is: what are the electrochemical gradients? Initially, there is no potential difference across the membrane because both internal and external solutions start with an equal number of positive and negative charges. There is no concentration gradient for K § because the amount of K § is the same on both sides of the membrane. Due to the presence of impermeant anions inside the cell, however, [C1-]o must be greater than [C1-]i. Consequently, C1- will enter the cell, moving down its electrochemical gradient, which is initially just the concentration gradient.
21. Osmosis and Regulation of Cell Volume
337
Inward m o v e m e n t of C1- makes the inside of the cell negative with respect to the outside. The developing inside negative potential has two consequences. First, it slows the rate of further influx of C1-. Second, the electrical gradient causes the accumulation of K § inside the cell. As K § moves down its electrochemical gradient in the direction set by the potential gradient, an opposing concentration gradient is created. Finally, an equilibrium is reached when the electrical and concentration gradients for both K + and C1- are equal in magnitude and opposite in direction. This satisfies the requirement that/x K and/Xcl, the electrochemical potential for K § and CI-, respectively, are the same in both compartments. During the evolution of the equilibrium, the influx of K § and C1- are equal on a m a c r o s c o p i c scale and are said to be c o u p l e d by the r e q u i r e m e n t for m a c r o s c o p i c electroneutrality. This is simply a s h o r t h a n d for the idea that any difference in rates of anion and cation t r a n s m e m brane flux (a separation of charge) gives rise to a potential that equalizes the fluxes (e.g., if the influx of K § was greater than that of CI-, an inside positive potential w o u l d develop, slowing K § influx and accelerating C1influx until the rates m a t c h e d exactly). In the e x a m p l e in Fig. 13A, both [K+]i and [C1-]i increased by - 7 8 . 5 m M . T h e total entry of K § and C 1 - c a n n o t be precisely equal,
A
however, b e c a u s e establishing a potential difference implies that a separation of charge, albeit quite small, m u s t have taken place. If this were a spherical cell with a radius of 20 lam, an inequality of the K § and C1- fluxes of less than 1 ion per 100 000 w o u l d cause a potential difference of 100 mV. 1 The basis for Donnan equilibrium can be expressed in terms of the electrochemical potentials 2 of ions that cross the membrane, here K + and C1-. For ideal solutions,/x is given by -/x 0 iX~o + R T In [K § + zI~ ~q/o (72a) -/x 0 ~. + R T I n [ K + ] / + ZK ~ i and tzcl ~ - / x 0 + R T In[ C1- ] o + Zcl ~Oo
(72b)
tzcli --].L 0 + R T In [C1-]i + ZC1 ~ ~ti
where the subscripts o and i represent the outside and inside of a cell,/z ~ is the chemical potential of the standard state, is the potential of the compartment, z i is the valence of species i, and R, T, and ~ have their usual meanings. At equilibrium, each permeant species distributes so t h a t / x is
1The amount of charge, q, necessary to establish a potential difference, Em, is related to the specific capacitance of the membrane, C m, and its area, A, by
150 K* 150 Cl"
q = Cm AEm
Assuming a specific membrane capacitance of 1 ixF/cm2, as is typical of most biological membranes, the amount of charge necessary to develop a potential of 100 mV in a spherical cell with a radius of 20 lxm is q = (lxl0-6 F/cm2)[ 47r (20x10-4 cm)2]( 0.1 V) ( - ~ )
Em= 0 A~ = 0
E~ = - 1 1 . 4 mV A~ = 157 mosmol/L
q = 5.0 x 10 -12 C This can be converted to a change in the concentration, AC, of ions in the cell
5 K* 150 CI"
AC = (q/~)/'O
where ~ is Faraday's constant, and O is used for the volume of the cell here to distinguish it from voltage. For the same 20-~m-radius cell AC =(5.0 • 10-12 C/96 500 C/mol)/[~Tr (20• Em = 0
LX~=0
H20
Em =-91.1 mV
1L 1000 cm3 ) ]
AC = 1.5 x 10 -6 mol/L
A~=0
F I G U R E 13. Development of a Donnan equilibrium. (A) Cell bathed in 150 mM KC1 is "filled" with 150 mM K§ 20 mM CI-, and 130 mM A-. Initially, there is no potential or osmotic gradient (left). K+ and C1- enter the cell until [K+]o.[C1-]o = [K+]i.[C1-]i, and the ionic fluxes establish a potential and osmotic gradient (fight). An equilibrium is achieved only if the membrane is rigid and cell volume is constant. With a biological membrane, however, water enters the cell, cell volume increases, and the [K+]i 9[C1-]i decreases. This causes further entry of solute and solvent, and the cycle repeats until the cell bursts. (B) DoubleDonnan system. Part of the extracellular K§ is replaced by Na§ which is assumed to be impermeant. With impermeant ions on both sides of a membrane (left), equilibrium is attained (right) without an osmotic pressure gradient. In this case, K+, CI-, and water leave the cell, until at equilibrium [K+]o9[C1-]o = [K+]i. [C1-]i. The amount of A- in the cell is fixed, and thus, the decrease in cell volume increases [A-]i.
cm)a(
Thus, a potential gradient of 100 mV would develop if ion influx were to increase [K+]i by only 1.5 IxM more than it increased [C1-]i. 2Formally, as presented previously in Eq. 3 for the chemical potential of a neutral species, Pi should be defined in terms of the mole fraction rather than the concentration of a component and an additional term, PV, representing the pressure times the partial molar volume, should be added. For dilute solutions, the mole fraction of a solute closely approximates its concentration (see Eqs. 5 and 9). The PV terms are ignored in the derivation of the ion distribution for a Donnan equilibrium because to simplify we initially assume the membrane is rigid (see Fig. 13A), and then, we consider a situation without a pressure gradient where the PV terms cancel (see Fig. 13B). Derivations of Donnan equilibrium retaining the mole fraction and the pressure-volume terms can be found in Overbeek (1956) and Lakshminarayanaiah (1984). m
w
m
338
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
identical inside and outside the cell. Equating the expressions for/x and simplifying gives
R T In [K+]o + z K ~@o - R T In [K+]i + ZK ~@i RTln[C1-]o + z c l ~ b o _ R T l n [ C l _ ] i
+ Zcl~l]ti
(73)
Membrane potential, Em, is measured as i ~ i Rearranging and substituting in for E m and z gives
~o-
RT [K+]i E m - E K - - - - ~ In [K+---~o
[Cl-]o
RT E m -
E cI
(74)
These expressions are the Nernst equilibrium potentials for K § (EK) and C1- (Ecl). Thus, when the mobile ions attain their equilibrium distribution in a Donnan system, the potential between the compartments, Era, simultaneously equals both E K and Ecl. Equating the Nernst potentials and simplifying gives the ratio of intra and extracellular ions predicted by Donnan equilibrium
[K+]i
[Cl-]o
[K+]o
[C1-]i
(75)
or expressed another way
[K+li'lc1-]i-[K+]o
" [Cl-]o
(76)
In a Donnan equilibrium, the KC1 product inside a cell equals the KC1 product outside. This simple rule implies that increasing extracellular K + (e.g., by replacing Na +) will cause a cell obeying Donnan equilibrium to take up K + and C1- and swell. Ion movements in a number of tissues appear to follow Donnan equilibrium, at least for some conditions. Boyle and Conway (1941) made careful measurements of [K+]i, [C1-]i, and cell water in frog sartorius muscles while varying extracellular KCI. For [K+]o greater than 6 mM, the experimental ratio of the KCI products, ([K+]o- [C1-]o)/([K+] i. [CI-]i), was very nearly 1.0, as predicted by Eq. 76. The deviation at low [K+]~ occurs because Na + influx causes E to deviate from E K. Accompanying changes in cell water in these experiments and in separate experiments in which K + replaced Na + also agreed with theory. The expectations for a Donnan equilibrium are calculated for the system shown in Fig. 13A. When the intracellular and extracellular KC1 products are equal, E = E K = E o -- -11.4 mV. In addition, because the number of ions inside the cell is much greater than that outside, the Donnan system has established a significant osmotic gradient. Assuming ideal behavior, the osmolarity inside the cell,-~457 mosmol-L -l, is about 1.5 times that outside, 300 mosmol.L -~. The resulting osmotic pressure can be calculated as follows:
A,rr - R T AC
(77)
Because most cells are readily permeable to water, the osmotic pressure generated here would cause water to enter the cell. This influx of water dilutes the intracellular ion content, and the product [K+]i-[C1-]i must fall below [K§ o. As a result, more KC1 would enter the cell, followed again by more water, in an endless cycle that would lead to destruction of the cell. That is to say, the simple Donnan system described in Fig. 13A fails to reach equilibrium when typical cell membrane properties are assumed.
B. Double-Donnan or Pump-Leak Hypothesis How can we reconcile the failure of the simple Donnan equilibrium model (Fig. 13A) and the experimental studies demonstrating behavior consistent with Donnan equilibrium? The Donnan system can be stabilized in two ways. A hydrostatic pressure could be applied to balance the osmotic pressure and arrest transmembrane water movement. In view of the enormous pressures required, this is not a realistic solution for animal cells. Alternatively, equilibrium can be attained by restricting an ionic species to the extracellular compartment just as one is restricted to the intracellular compartment. This arrangement is referred to as a double-Donnan (Leaf, 1959) or pump-leak system (Tosteson and Hoffman, 1960) and is illustrated diagrammatically in Fig. 13B. Assume that both Na § and macromolecular anions, A-, are restricted to the extracellular and intracellular compartments, respectively. Filling the cell with the same concentrations of K § C1- and A- as before, we find now that [ K + ] i 9 [C1-]i > [K§ 9[C1-] o, and thus KC1 must leave the cell to establish Donnan equilibrium. In the process, water follows the KC1, adjusting cell volume until the KC1, products are equal. A consequence of water flow is that the concentration of the impermeant intracellular ion exactly equals the concentration of the impermeant extracellular ion. At equilibrium, E m = E K = Ecl = -91.1 mV. In contrast to the simple Donnan system, the osmotic pressure developed by intracellular macromolecules and their counterions (sometimes referred to as colloid osmotic pressure) is exactly balanced in the double-Donnan system by the osmotic pressure developed by ions restricted to the extracellular fluid and their counterions. As a consequence, there is no net osmotic pressure across the membrane, and the system is stable. Another approach for considering the volume change expected in the system in Fig. 13B is to calculate the effective osmotic pressure of each compartment. We can suppose that the reflection coefficients, o-, are 1.0 for both A- and Na + because they are impermeant, and that the reflection coefficients are 0 for both K § and C1- because they freely pass the membrane. Extracellular and intracellular osmotic pressures are given by "it - E ~ Cn n
= (0.082 L-atm-K-l.mol-1) (310 K)(0.457- 0.300 mol-L-1)
% - 1.[Na+]o + 0.[K+]o + 0.[C1-]o
= 3.99 atm
%.- 1-[A-]i + 0-[K+li + 0.[C1-]i
(78)
21. Osmosis and Regulation of Cell Volume
Because only the impermeant species contribute to osmotic pressure, the only way for the cell to reach osmotic equilibrium, 7r~ 7"gi, is to alter cell volume until [Na§ = [A-]i. As a result, the ratio of volumet==/volumet= o must equal [A-]i/[Na+]o . It is important to realize that an ion does not need to be impermeant to be effectively restricted to the extracellular space and thus counterbalance the osmotic pressure developed by intracellular macromolecules. Na § is permeant, but it adequately plays this role anyway. As long as the leak of Na + down its electrochemical gradient into the cell is matched by its transport back out, cell volume will remain stable. Consequently, the existence of a pump that actively extruded Na + against its concentration gradient was postulated to explain cell volume (Leaf, 1956), and subsequently the Na +, K+-ATPase, which extrudes 3 Na + while taking up 2 K + at the cost of ATP hydrolysis, was identified. Because energy is consumed to extrude Na § the cell is in a steady state rather than a true equilibrium. =
C. Modulation of the Na+-K § Pump One implication of the double-Donnan or pump-leak model is that the Na+-K + pump is ultimately responsible for cell volume regulation. Perturbations that alter passive Na § entry must lead to offsetting changes in the rate of Na § extrusion by the Na+-K + pump, or cell volume will change. Because the K m of the Na§ + pump, for intracellular Na + is close to the physiological [Na§ alterations in Na § influx automatically give rise to a compensatory modulation of Na § effiux. Nevertheless, metabolic or pharmacologic inhibition of the Na§ § pump should lead to a net gain of Na § and anions coupled by macroscopic electroneutrality and result in a swelling of the cell. The effect of pump inhibition has been examined extensively (Macknight and Leaf, 1977; Macknight, 1988). The predicted cell swelling has been reported in many tissues, including brain slices, kidney slices, renal tubules, hepatocytes, and sheep erythrocytes, when the Na+-K § pump is inhibited by cardiac glycosides (e.g., ouabain) or by depleting ATP. It is equally clear, however, that swelling after pump inhibition in renal cortex, liver slices, various muscle preparations, lymphocytes, and human erythrocytes is very slow or even absent, perhaps reflecting a low Na § permeability. Several processes may affect the response of cell volume to Na+-K § pump inhibition, and the outcome in some tissues depends on the experimental conditions. Rather than accumulating C1- with Na § cells might instead lose K § to satisfy macroscopic electroneutrality. An equivalent gain of Na § and loss of K § replaces one osmotically active particle with another, causing no change in cell volume. Closer consideration of this mechanism indicates that it can only be a holding action, however. The loss of intracellular K § eventually must lead to a reduction of the K § gradient and a less negative Em. This will lead to accumulation of C1- with Na § and cell swelling. Nevertheless, a loss of K § must slow the swelling that otherwise would have occurred, and this would explain the absence of swelling in cells with appropriate Na § and K § permeabilities over the time course of experiments.
339
The possibility that a mechanism other than the Na§ § pump can extrude Na § has been considered. Although controversial, a cardiac glycoside-insensitive but metabolically dependent volume regulation mechanism in kidney that does not incorporate the Na§ + pump has been described (for review, see Macknight and Leaf, 1977). In addition, circulating erythrocytes from a number of carnivores, including dog, cat, bear, and ferret, lack a functioning Na§ § pump and must regulate their volume by a different mechanism (Sarkadi and Parker, 1991). In these cells, Na § effiux must depend on the gradient of other ions. D. isosmotic Volume Regulation Although principles of the pump-leak or double-Donnan model are correct and still relevant to the regulation of cell volume, it has become apparent that neither the leak nor the pump is constant. Not only is the control of these fluxes more complex than originally envisioned, but a myriad of other transport processes also contribute to cell volume regulation. Stated simply, constant cell volume under isosmotic conditions implies an equality of intra- and extracellular osmolarity that is perpetuated by a continuous balance of the effiux and influx of osmolytes. The transport processes involved include, but are not limited to: (1) ion and organic osmolyte channels; (2) the Na § K § pump; (3) the Na§ -, K+-C1-, and Na§ - cotransporters, which transport the specified ions in one direction; (4) Na§ sugar and amino acid cotransport; (5) Na § Ca2§ exchange, which exchanges 3 Na + for 1 Ca2§ and (6) osmotically neutral exchangers that indirectly provide a net solute flux. For example, Na+-H§ exchange allows the cell to accumulate Na § but the H § removed is replaced by dissociation of H § from intracellular buffers. C1--HCO 3 and Na§ § exchange can operate in parallel to mediate a net influx of Na § and C1- in exchange for H § and HCO 3, which are converted to CO 2 and 1-120 by the action of carbonic anhydrase. It should be noted that CO 2 freely crosses the cell membrane (o-= 0) and does not directly contribute to solution tonicity. Also, the direct extrusion of water by these two parallel exchangers is negligible compared to the osmotic water gain caused by the accumulation of Na § and C1-. Of the number of transporters that participate in volume regulation in any given type of cell, which are most important in regulating cell volume under isosmotic conditions? No simple answer can be offered. The importance of each process to cell volume regulation depends critically upon the tissue and species under consideration as well as the conditions. Moreover, transporters and channels are modulated by multiple signaling pathways, and they extensively interact by altering membrane potential or the concentration of the transported species, ff the rates of ion transport are not correctly matched, cells will inappropriately shrink or swell. The precise maintenance of cell volume exemplifies the need for sensitive and complex regulatory mechanisms. Attempts to mathematically integrate the fluxes and study their interaction have been made based on the cell's requirement for macroscopic electroneutrality and osmotic equilibrium and the equations governing ion fluxes (Jakobsson, 1980). For erythrocytes, the nonideal behavior of hemoglobin has been added (Bookchin et al., 1989). Although these simplified models correctly predict a number of
340
SECTIONII MembranePotential,TransportPhysiology,Pumps,and Exchangers
observations, they fail to explain others. In short, we remain a long way from a complete quantitative description of the processes underlying cell volume regulation. In the following sections, we will discuss three examples of isosmotic regulation of cell volume that illustrate some of the underlying principles.
Control i
8-Br-cG
j
Wash
(D E r-I
i i
o1.0 03
.r-I
1. Na+-K+-2C1 - Cotransport in Heart Recent fmdings indicate that the Na+-K+-2CI - cotransporter plays a critical role in regulating cardiac myocyte cell volume under isosmotic conditions. As in other tissues, Na+-K+-2C1cotransport conveys osmolytes into cardiac cells under physiological conditions. Because net transmembrane fluxes control cell volume, a decreased osmolyte influx is equivalent to increased effiux. Therefore, inhibition of the Na§247 - cotransport by bumetanide, for example, favors a reduction of cell volume. Consistent with this idea, bumetanide decreases the volume of atrial and ventricular myocytes by about 10% in less than 5 min, and myocyte volume is stable at this new level (Drewnowska and Baumgarten, 1991). The Na§247 - cotransporter cannot operate without Na § and C1- in the extracellular fluid, and removing either ion renders bumetanide ineffective. These data imply that ion uptake by Na§247 cotransport in the heart must be responsible for a significant osmolyte flux under isosmotic conditions and that other transport processes are incapable of fully compensating when this flux is removed. In contrast, myocyte volume was unchanged after inhibiting the Na§ § pump with 10 pM ouabain (Drewnowska and Baumgarten, 1991) or by cooling to 9 ~ (Drewnowska et al., 1991) for 20 min. At least in the short term, cardiac cell volume in isosmotic solution is influenced more by Na§247 - cotransport than by the Na§ § pump. Modulation of Na§ - cotransport by intracellular messengers such as cGMP may provide a physiological means of modulating cell volume in heart tissue. Figure 14 shows the effects of elevating intracellular cGMP in three ways: (1) by adding 8-Br-cGMP, a membrane-permeant analog of cGMP; (2) by adding atrial natriuretic factor (ANF), a natriuretic, diuretic, and vasodilatory hormone released by the heart that elevates cGMP by activating guanylate cyclase; and (3) by adding sodium nitroprusside (SNP), a vasodilator, that also activates guanylate cyclase. In each case, cell volume decreased. Furthermore, blocking cGMP-specific phosphodiesterase with zaprinast (M&B22948) augmented the effect of ANF. Based on its sensitivity to bumetanide and the requirement for ions transported by Na+-K+-2C1- cotransport, cGMP-dependent volume decreases were shown to be due to an inhibition of Na§247 - cotransport by cGMP (Clemo et al., 1992; Clemo and Baumgarten, 1995). Interestingly, lowering cGMP levels by inhibiting guanylate cyclase with LY83583 resulted in a small amount of cell swelling. Thus, changing cGMP from its physiological level in either direction altered cell volume. The mechanism and evidence for isosmotic regulation of cell volume in the heart are summarized in Fig. 15. Does the same mechanism regulate cell volume under isosmotic conditions in other tissues? Perhaps it does in some cells, such as vascular endothelium, in which cGMP inhibits Na+-K§ - cotransport. In other cells, Na+-K+-2C1- cotransport is stimulated by cAMP, cGMP, or a PKC-dependent pathway, and perhaps by a number of other signaling pathways
-foo 0
9
rr
l
6
I'o
do
do
Control I
ANF
]
Wash
6
iS
do
3'o
Control[
SNP
j
Wash
-10
4'0
~1.0 >
> .r-I
-~0 9 I--i II
f
-10
4'0
oJ E
o
1.0
-----
~
/
....
0J > .r-i
*JO 9
ro r--4 Off
I
I
o
do
Time (min)
3'o
4'o
F I G U R E 14. Changes in the volume of isolated ventricular myocytes during a 20-min exposure to: (A) 10 pM 8-Br-cGMP (8-Br-cG), a permeable cGMP analog; (B) 1 pM atrial natriuretic factor (ANF); and (C) 100
pM sodium nitroprusside (SNP). These agents reversibly decreased cell volume 11%, 8%, and 11%, respectively (n = 5 for each time point). The decreases in cell volume were caused by inhibition of Na§ - cotransport by cGMP. Relative cell volume was measured and calculated as described in the legend of Fig. 1. (From Clemo et al. (1992). Reproduced from The Journal of General Physiology, 1992,100, 89-114, by copyright permission of The Rockefeller UniversityPress.) (Palfrey and O'Donnell, 1992; Palfrey, 1994). This diversity in the control of Na+-K+-2C1- cotransport may be related to the variety of transporter isoforms that have been identified and cloned (Haas, 1994; Kaplan et al., 1996). Although the physiological significance of this diversity in the control of ion transport is not well understood, it must lead to diversity in the regulation of cell volume. 2. Hormones and Substrate Transport in Liver
Another interesting example of isosmotic volume regulation is found in hepatocytes. An impressive number of hormones induce either cell swelling or cell shrinkage at physiological concentrations, and these actions are related to their control of liver metabolism (H~iussinger and Lang, 1991; H~iussinger et al., 1994; Agius et al., 1994; H~iussinger, 1998). Na§ + exchange, Na§ - cotransport, and the Na§ § pump are
21. Osmosis and Regulation of Cell Volume
FIGURE 15. Schematic diagram of the action of atrial natriuretic factor (ANF) and cGMP on cardiac cell volume. Binding of ANF activates guanylate cyclase and increases intracellular cGMP levels. By one or more steps, cGMP inhibits Na+-K+-2C1- cotransport. Reducing ion influx by this means is equivalent to increasing n e t ion effiux, and cell shrinkage ensues. LY83583 inhibits guanylate cyclase, thereby blocking the effect of ANF, and zaprinast (M&B22948) potentiates the effect of ANF by inhibiting cGMP-specific phosphodiesterase (PDE). Sodium nitroprusside (SNP) also increases cGMP levels, and bumetanide (BUM) directly inhibits the cotransporter; both cause cell shrinkage. (From Clemo et al. (1992) Reproduced from The Journal of General Physiology, 1992, 100, 89-114, by copyright permission of The Rockefeller University Press.)
stimulated in rat liver cells by insulin. The net effect is that insulin increases [K+]i, [Na+]i and [C1-]i and causes cells to swell by about 12%. This swelling is prevented by bumetanide, a blocker of Na+-K+-2C1- cotransport, or by amiloride, a blocker of Na+-H + exchange. In contrast, glucagon shrinks hepatocytes by about 14%. Instead of directly opposing the action of insulin, glucagon reduces cell volume by increasing K + and C1effiux through ion channels. Other agents that swell hepatocytes include bradykinin and phenylephrine. Shrinking is initiated by adenosine, 5-HT, vasopressin, and cAMP. Hepatocytes also swell as a result of Na+-dependent amino acid cotransport in isosmotic media (H~iussinger and Lang, 1991; Boyer et al., 1992; H~iussinger et al., 1994; H~iussinger, 1998). Exposure to amino acids that are accumulated with Na + (e.g., alanine, glutamine, glycine, hydroxyproline, phenylalanine, proline, and serine) causes cell swelling. These effects occur at amino acid levels found physiologically in the portal vein. For example, glutamine provokes up to a 10% swelling with a half-maximal effect at approximately 0.7 mM. Amino acid cotransport gives rise to an inward current, and C1-enters to maintain macroscopic electroneutrality. Instead of loading the cell, Na + is pumped out, mainly by the Na+-K + pump, leaving an accumulation of K + and C1-. In contrast to these amino acids, substances not accumulated by the liver, such as glucose and leucine, do not affect hepatocyte volume. Cell swelling caused by Na+-dependent amino acid cotransport has also been observed in intestine and renal proximal tubule. 3. Na+-Ca2 + Exchange in Carnivore Erythrocytes
The Na+-K + pump in erythrocytes from dogs, cats, ferrets, and bears ceases to function as cells mature. As a result, [K+]i
341
and [Na+]i are similar to [K+]~ and [Na+]o. This poses a special problem for cell volume regulation. How can these cells offset the osmotic pressure generated by impermeant intracellular molecules and avoid swelling without a Na+-K + pump to make Na + effectively impermeant? Carnivore erythrocytes solve this dilemma by extruding 3 Na + in exchange for 1 Ca 2+ via the Na+-Ca 2+ exchanger (Parker, 1973; Sarkadi and Parker, 1991). In most cells, the electrochemical gradients for Na + and Ca 2+ favor the effiux of Ca 2+. In these erythrocytes, however, the gradients favor Ca 2+ entry because the Na + gradient is reduced, and Na+-Ca 2+ exchange operates in what is called the reverse mode. To stabilize their volume, carnivore erythrocytes also must have a means of maintaining the Ca 2+ gradient (i.e., low [Ca2+]i). This is accomplished by an ATPdependent Ca 2+ pump in the plasma membrane. Thus, as in other cells, maintaining cell volume in the face of impermeant intracellular colloids requires the expenditure of energy in a pump-leak mechanism. In this case, ATP is expended by a plasma membrane Ca 2+ pump rather than by the Na+-K + pump.
VI. Regulation of Cell Volume under Anisosmotic Conditions A. Osmometric Behavior of Cells Because the permeability of most cell membranes to water is much greater than that to solutes, ceils swell or shrink when placed in an environment that is hyposmotic or hyperosmotic, respectively. Water rapidly flows to equalize its chemical potential, ~w, inside and outside the cell. The initial volume response is often close to that predicted for an ideal osmometer from van't Hoff's law. An example is shown in Fig. 16, which illustrates the response of rabbit ventricular myocytes to solutions with osmolarities ranging from 195 to 825 mosmol/L (0.60 to 2.55 times isotonic). Relative cell volume (V), calculated as Vtest/Visosmotic, is plotted against the inverse of relative osmolarity, "/'/'isosmotic/"/rtest, and the data are fit to V - (1
Vb)( gr isosmotic/"/7"test)+ Vb
(79)
By definition, relative volume is 1.0 at a relative osmolarity of 1.0. Two conclusions can be reached from the data in Fig. 16. First, as expected from van't Hoff's law, the relationship between relative cell volume and the inverse of relative osmolarity is linear. Second, the intercept of the relationship on the volume axis, V b, is 0.34, which is significantly different from 0. This is interpreted as meaning that a fraction of cell volume is osmotically inactive, that is, it apparently does not participate in the response to anisosmotic solutions. Several arguments can be made to justify the observation of an osmotically inactive volume. The expectation from the simplest model is that cell water should vary in proportion to osmolarity. However, not all of cell volume is water. The volume of nonaqueous components such as small solutes and proteins, which represent 25-30% of the cell on a weight/weight basis, is unaffected by water movements. Even measurements of cell water show nonideal behavior, however (Macknight and Leaf, 1977; Solomon, 1989), so
342
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
additional explanations are necessary. One suggestion is that a fraction of cell water is intimately associated with cell proteins or membranes and thereby is bound or structured and unavailable as solvent (e.g., LeNeveu et al., 1976; Hinke, 1980). Although the state of water molecules adjacent to proteins and membranes must be different from that in the bulk phase of the cytoplasm, in light of NMR, intracellular ion activity, and other data, most investigators believe that virtually all water (-95%) is available as solvent (Shporer and Civan, 1977; Hladky and Rink, 1978). Another possibility is that the behavior of intracellular macromolecules is concentration-dependent. For example, the osmotic coefficient and charge on hemoglobin increase with its concentration as red cells shrink, and anions are drawn in to maintain electroneutrality. These phenomena are important in explaining water movement in red cells (Freedman and Hoffman, 1979), but their importance in other tissues remains uncertain. A third possibility is that intracellular compartments such as mitochondria, nuclei, endoplasmic reticulum, and sarcoplasmic reticulum of muscle cells may undergo volume changes that are not proportional to those of the whole cell. Differential responses to an osmotic challenge are expected because the plasmalemma and intracellular membranes possess distinct arrays of transporters and ion channels, and each sees a unique environment. Most methods for determining cell water or cell volume fail to distinguish between cytoplasmic and total cell water or volume (for a method that does distinguish these, see Reuss, 1985). A crucial assumption made in determining osmotically inactive volume also may affect the value obtained for Vb in
2.0
-
IJJ E~
J
1.5-
_.1
O > tu 1 . 0 >
J
I-< _.1
uJ 0 . 5 n,"
Eq. 79. The analysis assumes that only water has moved at the time volume is measured. That is to say, transmembrane ionic fluxes can be ignored. Although the permeability of the cell membrane to water is many times greater than the permeability to ions, the net fluxes of both water and ions start at the instant extracellular osmolarity is changed. If ion fluxes significantly affect intracellular osmolarity at the time of measurement, the extrapolated osmotically inactive volume will be imprecise. If in addition the ion fluxes depend on the direction or magnitude of the osmotic gradient, the plot of relative volume v e r s u s ~isosmotic/7/'test can become nonlinear (e.g., Grinstein et al., 1984).
B. Compensatory Regulation of Cell Volume Although an osmotic gradient initiates cell swelling or shrinkage, the initial volume response is not maintained in a wide variety of cells. Cell swelling activates compensatory processes that lead to an effiux of osmolytes and a reduction of cell volume. This is called a regulatory volume decrease (RVD). Similarly, cell shrinking activates an influx of osmolytes in some cells, leading to a compensatory swelling referred to as a regulatory volume increase (RVI). RVD and RVI nearly restore the original cell volume in some cells, are far less complete in others, and are absent in a few types of cells. Regulatory volume effects are thought to be adaptive and were first identified in nucleated duck erythrocytes (for a review, see Kregenow, 1981). Examples of an RVD and RVI taken from work by Grinstein et al. (1983) are shown in Fig. 17. Exposure of human peripheral blood lymphocytes to a solution made hypotonic by 50% dilution with water leads to a rapid, 1.6-fold increase in cell volume (Fig. 17B). Then, over about 10 min, an RVD returned cell volume almost completely to its initial value. On switching back to isotonic solution, cell volume shrank to less than the control value and then was restored by an RVI. RVDs and RVIs may differ in magnitude, however. A much less complete RVI was observed when lymphocytes in isotonic solution were shrunk in media made hyperosmotic by adding of 300 mosmol.L -1 of NaC1, and the RVI was absent when the lymphocytes were challenged with 300 mosmol-L -1 of sucrose instead of NaC1 (Fig. 17A). The RVD also could be eliminated, for example, by cooling lymphocytes to 4 ~ (Grinstein et al., 1984). Thus, regulation of cell volume following an osmotic challenge depends on the particulars of the perturbation as well as the cell under study (compare Figs. 1 and 17).
_ I
0
I
0.5
I
I
1.0
I
I
1.5
I
I
2.0
1/(RELATIVE OSMOLARITY) FIGURE 16. Relationship between relative cell volume and the inverse of relative osmolarity in cardiac ventricular myocytes. Data from 38 measurements of relative cell volume in anisosmotic solutions (0.6T, 0.8T, 1.8T, and 2.6T) were fit to a least squares regression line constrained to pass through (1, 1) (square) because relative cell volume is 1 in isotonic solution by definition. The extrapolated intercept on the relative volume axis, 0.34, represents the fraction of cell volume that is osmotically inactive (see Eq. 79). Volume measurements, calculation of relative cell volume, and means of adjusting osmolarity are as described for Fig. 1. (Reproduced from Drewnowska and Baumgarten (1991).)
C. Transport Processes Responsible for RVD and RV! How do cells gain or lose osmotic equivalents in anisosmotic media? The mechanisms underlying RVDs and RVIs have been extensively characterized in a variety of cell types and exhaustively reviewed (Hoffmann and Simonsen, 1989; Chamberlin and Strange, 1989; Grinstein and Foskett, 1990; Sarkadi and Parker, 1991; H~iussinger and Lang, 1991; McCarty and O'Neil, 1992; Strange, 1994; Hoffmann and Dunham, 1995; Lang et al., 1998; O'Neil, 1999). In general, cells undergo RVD or RVI by translocating Na § K § and CI-,
343
21. Osmosis and Regulation of Cell Volume
and a variety of channels, exchangers, cotransporters, and pumps can participate. In some cases, organic osmolytes (e.g., taurine, betaine, sorbitol, and urea) are transported instead of, or in addition to, inorganic ions. Table 3 lists processes that are activated by altering the volume of various cells, and Fig. 18 illustrates some of the mechanisms diagrammatically. The compensatory mechanisms invoked vary with the tissue, species, and conditions under which osmotic stress is applied. The primary mechanism of RVD in a number of cell types is activation of conductive pathways, and this process will be discussed in more detail. Cell swelling, acting directly or via a messenger, opens ion channels that allow increased effiux of K + and CI-, and H20 follows. Although the openings of cation and anion channels are independent events, K + and C1- effiux are tightly coupled by the need to maintain macroscopic elec-
A
HypertonK:
l
1.0-
0.8e-.-o NoCI
0.6-
o-.---o Sucrose
9
0
|
B
0 >
.~
1.6
,, I
I
10 0
10
I
2O
I
9
30
40
Hyl:)otonK:
-
i
n,-
1.4
~
IsotonK:
1.2
1
1.0 i~
A
o
A
I ~o 9
0.8
o-a- o
0.6 !
10 0
!
10
I
I
I
I
I
10
20
30
20
0 30
I
4O
Time (mm)
FIGURE 17. (A) Effect of hypertonic solution (2T) on relative cell volume of human peripheral blood mononuclear lymphocytes. Solution osmolarity was increased by adding 300 mosmol. L -1 of NaC1 or sucrose. Cells shrank 40% in hypertonic solution, and a small regulatory volume increase (RVI) was observed in the NaC1 solution. Points are representative of four experiments. (B) Response to hypotonic solution (0.5T). After swelling by nearly 60%, a regulatory volume decrease (RVD) virtually restored cell volume to its initial value in 10 min. On returning to isotonic solution, cell volume decreased below its control value, and an RVI lead to full recovery. Different symbols represent separate experiments selected from over 40. Cell volume was measured with a Coulter counter that determines volume from the change in electrical resistance of a column of solution as cells pass through an aperture. (From Grinstein et al. (1983). Reproduced from The Journal of General Physiology, 1983, 82, 619-638, by copyright permission of The Rockefeller University Press.)
troneutrality. If only the anion or cation channel were to open, the resulting change in E would rapidly make ion effiux selflimiting and arrest volume regulation. The effect of this coupling of K + and C1- fluxes is illustrated in Fig. 19A, which shows RVDs in Ehrlich ascites cells (Hoffmann et al., 1986; Hoffmann and Dunham, 1995). At time zero, cells were switched from 300 to 150 mosmol.L -] media. Cells rapidly swelled to about 1.9 times their initial volume and then underwent an RVD that returned relative cell volume to about 1.3 within 5 min. When K + conductance was increased by pretreating cells with 0.5 pM gramicidin, a K + ionophore, the compensation by the RVD was more rapid and larger in magnitude, returning cell volume to nearly its control level in 2 min. When K + channels were blocked with 1 pM quinine, however, only a feeble RVD took place. These data argue that conductive K + effiux cannot keep up with C1- effiux in control cells and limits the rate of RVD (Hoffmann et al., 1986). Thus cell swelling must increase C1- conductance more than K + conductance, leading to greater effiux of C1- than of K +. Consistent with this idea, a depolarization is observed during the RVD. The activation of C1- conductance by swelling is only transient, however. The ability of gramicidin to induce RVD decays with time, as shown in Fig. 19B. In contrast to the high selectivity of C1--dependent cotransporters, RVD is also supported by Br-, NO 3 and SCN, suggesting the C1- channel responsible for RVD poorly discriminates among these anions. Furthermore, RVD can be suppressed by C1- channel blockers such as inacrinone (MK-196) and diphenylamine-2-carboxylate (DPC) but not by the cotransport inhibitors. Taken together, the data in Ehrlich ascites cells provide strong evidence that two independent channels are responsible for RVD instead of, for example, activation of a single K+-C1- cotransporter. Organic osmolytes also permeate a class of swelling-activated C1-channels referred to as volume-sensitive organic osmolyteanion channels (VSOAC) and are released by a number of cells to regulate volume (Jackson et al., 1994). VSOAC have a high permeability to taurine (Ptaurine/PCl = 0.75) and glutamate (PglutamatJPc1 = 0.20) and other amino acids and small organic molecules are permeant (Bandarali and Roy, 1992). Whereas many types of cells exhibit extensive RVDs, fewer cells exhibit robust RVIs. RVIs are usually due to an accumulation of Na +, CI-, and in some cases, K +, which occurs over minutes. The primary mechanisms underlying RVIs are acceleration of Na+-K+-2C1- or Na+-C1- cotransport and coupled Na+-H + and C1--HCO 3 exchange. Volume recovery can be blocked by application of appropriate inhibitors such as bumetanide for Na+-K+-2C1- cotransport, amiloride for Na+-H + exchange, and SITS (4-acetamido-4'-isothiocyano-stilbene2,2'-disulfonic acid) for C1--HCO 3 exchange. RVI in most cells is readily observed when simple salt solutions bathe the cells, but in renal cortical collecting duct and proximal tubules, butyrate, acetate, or other metabolizable fatty acids must be in the perfusate to support RVI. Substrate metabolism may provide H + and HCO 3 to support Na+-H + and C1--HCO 3 exchange in these cells (for a review, see McCarty and O'Neil, 1992).
D. Organic Osrnolytes Adaptation to hyperosmolarity over a longer term also occurs in some cells and is mediated by an accumulation of
344
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
TABLE 3 Ionic M e c h a n i s m s of R e g u l a t o r y R e s p o n s e s to A n i s o s m o t i c S o l u t i o n s Transport mechanism activated
Cell types
Cell swelling-induced regulatory volume decrease K+ and C1-conductances
Frog urinary bladder Chinese hamster ovary cells Ehrlich ascites tumor cells Frog skin HeLa carcinoma cells Human platelets Human granulocytes Human lymphocytes Intestinal 407 cells Madin-Darby canine kidney (MDCK) cells Necturus enterocytes Necturus gallbladder Rabbit renal proximal convoluted tubule Rat hepatocytes
K+-C1- cotransport
Avian, dog, fish, human, rabbit, and low-K + sheep erythrocytes Ehrlich ascites tumor cells (Ca2+ depleted) Necturus gallbladder
Coupled K+-H+ and C1--HCO3 exchange
Amphiuma erythrocytes
Na+-Ca2+ exchange
Dog and ferret erythrocytes
Organic osmolyte effiux
Crustacean muscle and myocardium Ehrlich ascites tumor cells Elasmobranch and molluscan erythrocytes
Cell shrinkage-induced regulatory volume increase Na+-K+-2CI- cotransport
Astrocytes C6 glioma cells Duck, fish, rat, and human erythrocytes Ehrlich ascites tumor cells Frog skin HeLa cells Rat kidney medullary thick ascending limb 3T3 cells
Coupled Na+-H+ and C1--HCO3 exchange
Amphibian gallbladder Dog and amphibian erythrocytes Human lymphocytes Mouse medullary thick ascending limb Rabbit renal proximal straight tubule Ehrlich ascites tumor cells
Na+-C1- cotransport
Ehrlich ascites tumor cells Necturus gallbladder
K+ and C1-conductances, inhibited
Madin-Darby canine kidney (MDCK) cells
Organic osmolyte influx
Many animal and plant cells, bacteria and fungi
For references, see Yancey et al. (1982), Hoffmann and Simonsen (1989), Chamberlin and Strange (1989), Grinstein and Foskett (1990), Wolff and Balaban (1990), Sarkadi and Parker (1991), H~iussinger and Lang (1991 ), McCarty and O'Neil (1992), Boyer et al. (1992), Strange (1994), Hoffmann and Dunham (1995), Lang et al. (1998), and O'Neil (1999).
organic osmolytes, including amino acids, polyols, and urea (Yancey et al., 1982; Chamberlin and Strange, 1989; Wolff and Balaban, 1990; Garcia and Berg, 1991" Yancey, 1994; Berg, 1995). Three of these organic osmolytes, taurine, betaine, and inositol, are taken up by Na+-dependent or Na +- and
Cl--dependent cotransporters. The taurine and betaine transporters have been cloned and belong to the same family as the Na § and C1--dependent norepinephrine, GABA, dopamine, serotonin, and proline transporters (Uchida et al., 1992). The taurine transporter is found in a variety of cell types. The
21. Osmosis and Regulation of Cell Volume
FIGURE 18. Schematic diagrams of the transport processes involved in (A) regulatory volume decrease (RVD) and (B) regulatory volume increase (RVI). Following cell swelling, a compensatory reduction in cell volume (RVD) may result from increased K + and C1- conductance (channels), activation of functionally coupled C1--HCO 3 and K§ § exchange (exchangers), or activation of K+-C1- cotransport (cotransporters). Following cell shrinking, a compensatory increase in cell volume (RVI) may result from activation of the Na+-K+-2C1- or Na§ - cotransporters (cotransporters) or activation of functionally coupled C1-HCO 3) and Na+-H + exchange (exchangers). The Na§ § pump extrudes the Na § that enters during RVI, so that cells gain K + and C1-.
345
346
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
A 201
Gra~nicidin
1.5-
",u
Gramicidin
1.0
Time (rnin)
B
Gramicidin
1500-
o!
1000.
6
~
~b
l"s
Time ( rain )
FIGURE 19. (A) Effects of quinidine (1 mM) and gramicidin (5 pM) on regulatoryvolumedecrease (RVD) in Ehrlich ascites cells suspended in a Na+-free choline medium with 1 mM Ca2+. Swelling was induced by diluting the medium to 0.5T. Quinidine blocks Ca2+-activatedK§ channels, and gramicidin increases cation conductance (K+ conductance in the absence of Na+). The data indicate that K+ permeability is rate-limitingduring the RVD. (B) The swelling-induced increase in C1- conductance is transient. Same protocol as in as (A), except that quinidine (1 mM) was added to the hypotonic media to reduce K+ conductance (control). At selected times, 0.5 pM gramicidin was added to increase cation permeability. The ability of gramicidin to induce an RVD follows the time-course of the swelling-induced C1- conductance. Relative cell volumes measured with a Coulter counter, as described for Fig. 17. (Reproduced with permission from Hoffmann et al. (1986).)
distribution of mRNA levels is kidney > ileum > brain > liver > heart (Berg, 1995). This sequence for taurine transporter mRNA is surprising because heart has the highest intracellular taurine concentration among these tissues. The mRNA level for the betaine transporter is also highest in the kidney (Yamauchi et al., 1992). A different family of transporters is responsible for Na§ inositol uptake. The inositol
transporter is related to those for glucose and nucleosides and is highly expressed in renal medulla (Kwon et al., 1992). A common feature of all three osmolyte transporters is that gene transcription is up-regulated by hypertonic stress. Cells cultured under hypertonic conditions slowly increase their maximum rates of transport, and intracellular concentrations rise. For example, betaine transporter mRNA levels in MDCK cells peak about 16 h after the osmotic challenge, and the transport rate peaks at about 24 h (Uchida et al., 1993). Another mechanism for accumulating organic osmolytes while adapting to hypertonicity and cell shrinkage is to modulate their synthesis. For example, renal medullary cells, which experience osmolarities of more than 1200 mosmol. L -1 during antidiuresis, accumulate sorbitol and inositol to regulate their volume (Wolff and Balaban, 1990; Garcia and Berg, 1991; Sands, 1994). Sorbitol levels rise 12-24 h after osmotically challenging cultured renal medullary cells, and accumulation depends on the induction of aldose reductase by enhanced transcription. Aldose reductase converts exogenous glucose to sorbitol. Sorbitol is synthesized from stored glycogen in other tissues. This requires coordinated up-regulation of glycogen phosphorylase and hexosekinase to produce glucose 6-phosphate, an intermediate for sorbitol production, and the down-regulation of phosphofructokinase to prevent consumption of substrates for sorbitol production. Inhibition of degradation is another mechanism for accumulating organic osmolytes when cells are challenged by hypertonicity. This is the main mechanism by which MDCK and medullary collecting duct accumulate glycerophosphocholine (GPC) (Berg, 1995). In response to hypertonicity or elevated urea, GPC-choline phosphodiesterase is inhibited. Under some conditions, synthesis of GPC by phospholipase also may be enhanced. The accumulation of organic osmolytes is metabolically expensive, especially considering that many cells dump organic osmolytes to respond acutely to swelling (Chamberlin and Strange, 1989; Rasmusson et al., 1993; Jackson et al., 1994, Hoffmann and Dunham, 1995). Why does the cell expend extra energy to use organic compounds rather than inorganic ions? Apparently the reason is that accumulation of inorganic ions may perturb protein structure within cells. The Hofmeister seties, first described more than 100 years ago, lists inorganic ions according to their ability to alter the solubility and conformation of proteins. Nonspecific effects of inorganic salts include changes in Vmax, K m, tertiary structure, and subunit assembly of enzymes (Yancey et al., 1982; Somero, 1986; Yancey, 1994). Why are some organic compounds accumulated rather than others? Yancey et al. (1982) categorized osmolytes as nonperturbing (stabilizing) or perturbing (destabilizing). Elevated concentrations of nonperturbing osmolytes are compatible with normal enzyme function, whereas high concentrations of perturbing osmolytes are not. This distinction could be due either to direct interaction of osmolytes with enzymes or substrates or to effects on hydration, solubility, or charge interactions of proteins. Organic solutes that are nonperturbing generally are uncharged [e.g., trimethylamine N-oxide (TMAO), glycerol] or zwitterionic (e.g., betaine, taurine), although negatively charged octopine is used by some cells. In contrast, most perturbing organic osmolytes are positively charged (e.g., arginine, guanidinium). Neutral urea, however, perturbs proteins.
21. Osmosis and Regulation of Cell Volume
Finally, nonperturbing osmolytes can counteract the destabilizing effects of perturbing osmolytes. Several organisms accumulate perturbing osmolytes, such as urea, in fixed ratios with nonperturbing osmolytes.
E. Signaling Pathways Underlying RVD and RVI How do cells detect alterations in their volume and activate the transport processes underlying compensatory volume regulation? Do the transporters themselves sense cell volume and respond by changing their activity? Alternatively, does water movement simply change the concentration of a critical regulatory substance? Answers to these questions are just beginning to emerge. It appears that the signaling pathways modulated by cell volume are as diverse as the transporters that respond. 1. Anisosmotic M e d i a
An obvious candidate for signaling a change in cell volume is the composition of the anisosmotic media itself; that is, the ionic strength or concentrations of ions in the bathing media might initiate a regulatory volume response. Changing concentration or ionic strength must affect various ion transport processes and transmembrane ion fluxes to some degree. Nevertheless, these factors seem relatively unimportant in volume regulation because comparable responses are observed when nonelectrolytes such as mannitol or sucrose replace a large fraction of the electrolytes in the bathing media. RVDs also are initiated in isosmotic solutions after swelling caused by sugar or amino acid uptake. Thus it appears that regulation is initiated by the volume change itself rather than the composition of the bathing media. 2. M e m b r a n e Potential
Another possibility is that regulatory volume responses are initiated by dilution or concentration of intracellular K § via an effect on E m. A two-fold change in volume without compensatory K + fluxes would alter [K+]i and cause an 18 mV change in E in a cell that conforms to the Nemst equation. Fluxes through K + and C1- channels are voltage-dependent, reflecting both the electrochemical driving force and the voltage-dependent conductance. For many K § channels, the current-voltage relationship is highly non-linear. Furthermore, the Na+-K + pump and Na+-Ca 2§ exchange are both voltage-dependent because they mediate a net movement of charge across the membrane. Despite this, it is unlikely that changes in E m are the p r i m a r y cause of RVDs or RVIs in most cells. Regulatory responses have been observed with changes in cell volume of phosphatidylethanolamine > phosphatidylcholine. The flipase exhibits stereospecificity in acting only on Lisomers and is inhibited by vanadate and by oxidation of protein sulfhydryl groups. Elevated intracellular calcium, such as occurs during normal red-cell senescence as well as in red cells from patients with sickle cell anemia, causes loss of the normal phospholipid asymmetry, leading especially to excess phosphatidylserine in the outer hemileaflet. Under these conditions, the external membrane surface acquires procoagulant activity and promotes thrombosis (for review, see Diaz and Schroit, 1996). The external surface of the red cell membrane is coated with adsorbed albumin and some plasma globulins. Because of the sialic acids on glycophorin, an integral membrane glycoprotein, the cell surface is negatively charged (for review, see chapter 27 by Seaman in Surgenor, 1974). The cytoskeleton is an organized polygonal fibrous network, about 60 nm thick containing spectrin and actin (for review, see chapter 1 by Gardner and Bennett in Agre and Parker, 1989; Bennett and Gilligan, 1993). Sides of the regular five- or sixsided polygons are formed by spectrin, a flexible filamentous protein 200 nm in length. Vertices of the polygons are formed by/$-actin protofilaments 30--40 nm in length consisting of 12-14 actin monomers; both grooves of the actin filaments contain tropomyosin, which binds to tropomodulin at junctional complexes. At the mid-region of the spectrin filaments a junctional ternary complex formed by ankyrin and band 4.2 protein links spectrin to the membrane-spanning integral band 3 protein. At the ends of the spectrin filaments, an additional linkage site to band 3 and to glycophorin C may be provided by protein 4.1. Spectrin-
actin junctional complexes also contain the actin-bundling protein dematin (band 4.9) and the calmodulin-binding protein adducin, which functions to cap and to regulate the length of the actin filaments and which may be involved in Ca2§ alterations in cytoskeletal structure (Kuhlman et al., 1996). Measurements of lateral translational diffusion showed that a fraction of band 3 protein is immobile because of its linkage with spectrin, whereas another fraction is capable of slow diffusion in the plane of the membrane (Golan and Veatch, 1980; for review, see chapter 13 by Golan in Agre and Parker, 1989). A small amount of myosin is also found in the red-cell cytoskeleton, but whether active tension is generated or regulated in mature red cells is not known. Presumably, the cytoskeleton is actively involved during enucleation in developing red cells and during diapedesis, or egress of red cells from the bone marrow to the peripheral circulation.
Iil. Intracellular E n v i r o n m e n t The red-cell membrane and associated cytoskeleton enclose a viscous cytoplasmic solution of the oxygen-binding pigment hemoglobin at a concentration of 34 g/100 ml cells, corresponding to 5.2 mM, 7.3 millimolal, or 44 g Hb/100 g cell water. This concentration is near the threshold for gelation, but hemoglobin is one of the most soluble of all proteins and constitutes more than 98% of red-cell protein by mass. The hemoglobin ce2/32tetrameters have a molecular weight of 64 373 Da and are approximately spheroidal with dimensions of 65/~ • 55 ~ • 50/~. In the interior of red blood cells, hemoglobin is nearly close-packed, with the distance between the surfaces of neighboring proteins averaging only about 20 ~ (Ponder, 1948). With the 86 carboxylates of aspartic and glutamic acids, and with the 98 basic amino and amine groups of arginine, lysine, and histidine per hemoglobin tetramer, the total cellular concentration of the titratable amino acids of human hemoglobin is around 0.9 M. Thus, hemoglobin contributes significantly, depending on intracellular pH, to a high intracellular ionic strength. Since the charged amino acids are located on the surface of hemoglobin (see Antonini and Brunori, 1971), and taking the radius of hemoglobin to be 28 A, the average distance between charged sites is only about 7 A. About 7600 water molecules per hemoglobin tetramer occupy the narrow interstices between the protein molecules. The hydration of hemoglobin in dilute solution is 0.2--0.3 g water/g Hb (see Antonini and Brunori, 1971), representing about 15% of the intracellular water. When the tortuosity of the surface of soluble proteins is taken into account, as much as 30% of redcell water could reside in the first monolayer around the protein surface. To determine the mean ionic activities of the intracellular KC1 and NaC1 in this concentrated charged environment, studies were conducted in which the red-cell membrane was rendered permeable to cations by exposure of the cells to the channel-forming antibiotic nystatin, thus allowing K § Na § and C1- to reach Gibbs-Donnan equilibrium. In these experiments sufficient extracellular sucrose was added to prevent cell swelling by balancing the colloid osmotic pressure
23. Membrane Transport in Red Blood Cells
of hemoglobin and other impermeant cell solutes (see Chapter 15). The mean ionic activity coefficient of KC1 and NaC1 in the concentrated intracellular hemoglobin solution was found to be within 2% of that in the extracellular solution (Freedman and Hoffman, 1979a). Moreover, permeant nonelectrolytes also have equilibrium ratios of intracellular to extracellular concentrations within 10% of unity (Gary-Bobo, 1967). In view of the high intracellular ionic strength and the high volume fraction of cell water in direct contact with hemoglobin, it is both curious and remarkable that intracellular salts and nonelectrolytes appear to behave as if in dilute solution. Either the intracellular solution is indeed like a dilute solution, or alternatively, the expected effect of protein-solvent interactions in altering the activity of intracellular solutes is offset by the effect of interactions between proteins and the solutes themselves.
IV. M e t a b o l i s m and Life S p a n The red cell also has relatively simple metabolic pathways (Grimes, 1980; Beutler, 1986), at least in comparison with most other cells. Catalogs of red-cell enzymes list about 140 enzymes (see the chapter by Friedemann and Rapoport in Yoshikawa and Rapoport, 1974; and chapter 3 by Pennell in Surgenor, 1974). Glycolysis produces ATP and lactate from glucose, inorganic phosphate, and exogenous purine in the form of adenine, adenosine, or inosine. A mathematical model of red-cell glycolysis was proposed by Rapoport et al. (1974). The pentose shunt provides reducing equivalents in the form of glutathione, NADH, and NADPH, which, together with catalase, superoxide dismutase, glutathione peroxidase, glutathione reductase, and methemoglobin reductase, act to prevent the oxidation of protein sulfhydryl groups and of Fe z+ in hemoglobin. No tricarboxylic acid cycle, cytochrome system, or lipid catabolism or utilization are known to occur in red cells, although cholesterol and phospholipids do exchange with plasma lipids, and some fatty acids may be incorporated into membrane phospholipids (for review, see Shohet, 1976). Protein synthesis does not occur in mature red cells, and there is no DNA replication or transcription, and no RNA metabolism or gene action. Red cells in humans make up 40-45% of the blood volume, a fraction known as the hematocrit; thus, they are readily accessible, and are easily separable from leukocytes and platelets by centrifugation or filtration. In short-term experiments, red cells may be studied in simple buffered isotonic salt solutions (e.g., 145 mM NaC1, 5 mM KC1, 5 mM Hepes buffer, pH 7.4). For longer experiments, glucose is added to prevent the decline of ATP; for even longer term experiments, red cells survive in vitro at room temperature or at 37 ~ for many days, albeit with morphological heterogeneity, in a chemically defined culture medium that includes vitamin cofactors, an exogenous purine for synthesis of ATE and amino acids to support the synthesis of glutathione (Freedman, 1983). In vitro, red blood cells can be studied free from the uncontrollable variables of the intact organism. Each of the 25 trillion red blood cells in normal adult humans lives for about 120 days. At a turnover rate of about
379
1%/day, some 250 billion new red cells are released from the bone marrow each day, a rate that corresponds to 3 million cells per second! This may seem like a lot, but 3 million red blood cells occupy less than a microliter of volume, since each biconcave discoidal cell occupies only 87 ixm3 and has a surface area of 133 txm2, a diameter of 8 txm, and a thickness of 2.4 ixm at the rim and 1.0 Ixm at the center. After 120 days, senescent human red cells bind IgG autoantibody and are then recognized by macrophages in the initial stage of erythrophagocytosis, a process that leads to the recycling of iron, amino acids, and other essential redcell constituents. Considerable evidence indicates that the antigenic recognition sites are composed of clusters of an oxidatively denatured form of band 3 protein (for review, see chapter 9 by Low in Agre and Parker, 1989; Kay, 1991). The normal function of band 3 protein, also called capnophorin or AE1 (anion exchange protein 1), is to mediate the obligate electroneutral exchange of C1- for HCO 3 across the red-cell membrane during gas exchange in the pulmonary and systemic capillaries. The comparative biochemistry and physiology of red blood cells is instructive, with many differences in metabolic and membrane transport properties, as well as oxygen transport properties, known to occur among different animal species (for examples, see Willis, 1992). Whereas human red cells live for 120 days, the life span of dog red cells is 60 days, and that of mouse red cells is only 40 days. The life spans L of mammalian red blood cells correlate with body weight W according to the relation L = 69 W 0"12, as illustrated in the loglog plot in Fig. 1, but the physiological basis for this striking correlation is not understood (for discussion, see V~icha and Znojil, 1981). The life span data suggest that red cells contain a biological clock that is obviously not directly determined or controlled in the mature cell by gene transcription or translation.
200
I
"I
I
I
0.12 Life span (L) = 69W
Horse
SheeP~w
U~ >, m "0
~")Man
r
m
m
100
go
Guinea A / D o g pig V~Cat
8o o
m 70 er. r 60
nm
~
.
50 40
0.01
l~Mouse I 0.1
I 1
...... I 10
I 100
1000
Body weight (kg)
FIGURE 1. Mean life span (days) of mammalian red blood cells versus body weight (kg). (Data from V~isha and Znojil (1981) and various other sources.)
380
SECTION I1 Membrane Potential, Transport Physiology, Pumps, and Exchangers
i
V. Membrane Transporters in Red Blood Cells Among eight mammalian species (Table 1), the intracellular concentration of K + ranges from 135 mM in human red blood cells to only 8 mM in cat and dog red cells, whereas intracellular Na + varies from only 17 mM in humans to 142-162 mM in dog and cat; in contrast, the extracellular low K § and high Na § are relatively constant among species, as are intracellular and extracellular C1-. Cation transport in red blood cells was understood classically in terms of the pump-leak theory (Tosteson and Hoffman, 1960), which quantitatively accounted for the differing steady-state Na § and K + concentrations in red blood cells found in two genetic phenotypes of sheep designated as high K + (HK) and low K + (LK). High-K + sheep red cells were found to have a relatively high number of Na+-K § pumps per cell and high Na+-K + pump fluxes with relatively low ouabain-insensitive leakage fluxes. Anions such as CI-, HCO 3, and OH- are passively distributed and appear to follow the double-Donnan equilibrium, as explained later (for reviews on red-cell transport, see chapter 3 by Passow in Bishop and Surgenor, 1964; chapter 15 by Sachs, Knauf, and Dunham in Surgenor, 1974; Ellory and Lew, 1977; Agre and Parker, 1989; Raess and Tunnicliff, 1990). The ionic composition of the intracellular and extracellular solutions may be depicted on bar graphs, as shown for human red blood cells in Fig. 2. In human red cells, the Na+,K+-ATPase (see Chapter 17) specifically selects the K § from the Na+-rich medium, and pumps it against the K § concentration gradient (and electrochemical gradient) into the cytoplasmic solution. The same
TABLE i Concentrations (mM) of Potassium, Sodium, and Chloride in Mammalian Red Blood Cells and Plasma Intracellular Species
[K +]
[Na +]
Extracellular [C1-]
[K +]
[Na +]
[CI-]
Man a
135
17
77
3.7
138
116
Baboon b
145
24
78
4.7
157
115
Rabbit b
142
22
80
5.5
150
110
Rat b
135
28
82
5.9
152
118
Horse b
140
16
85
5.2
152
108
Sheep c HK LK
124 17
13 119
Dog d
8
162
80
4.6
165
123
Cat b
8
142
84
4.6
158
.
.
.
.
.
Mean b
SD
.
.
112 .
.
.
81
4.9
153
115
3
0.7
8
5
aFunder, J., and Wieth, J. O. (1966a,b). Scand. J. Clin. Lab. Invest. 18, 151-166, and Acta Physiol. Scand. 68, 234-245. bBemstein, R. E. (1954). Science 120, 459-460. CDunham, P. B. (1992). Comp. Biochem. Physiol. 102A, 625--630. dparker, J. C. (1973). J. Gen. Physiol. 61, 146-157.
o _=i
§
1
K 143
+
K
CI 105
Na + 145 cI 15o
Na
§
Hb33 =.=
7
P12 pH o 7 . 4 2 8 9 rnOSM Vw/V~ - 1.o
FIGURE 2. Ionic composition of the intracellular solution in human red blood cells and of the extracellular solution, designated by i and o, respectively. Organic phosphates are represented by P-.
ionic pump specifically selects Na § from the K§ cytoplasmic solution and extrudes it from the cell against the Na § concentration gradient. This coupled active transport of three internal Na § for two external K § uses metabolic energy obtained from the hydrolysis of ATP (for review, see Hoffman, 1986; chapters 6 and 16 by Mercer et al. and by Kaplan in Agre and Parker, 1989). In the steady state, at the same time that K § is actively accumulated in the cell, it is also continually leaking out of the cell at the same rate through parallel pathways down its concentration gradient. The same is true for Na § leaking into the cell. Steady-state distributions of K § and Na § are thus achieved by a balance between active pumping and passive leakage. If the Na+-K § pump of human red blood cells is completely inhibited, the loss of K § and gain of Na § is so slow that the leakage fluxes would continue unabated for at least 30 days before GibbsDonnan equilibrium is approached (Fig. 3); however, the cells would hemolyze before reaching equilibrium because of the inability of the plasma membrane to withstand any significant osmotic pressure associated with the GibbsDonnan equilibrium, as discussed in Section VI. A striking correlation between the properties of the Na+,K+ATPase and cation fluxes in red blood cells constitutes impressive evidence that the Na+,K+-ATPase is the ion pump that mediates the fluxes of Na + and K § across the red-cell membrane: (1) Both the Na+,K+-ATPase and the control of cation transport are located in the membrane; (2) the Na+,K+-ATPase and cation transport are both stimulated specifically by intracellular ATP, by intracellular Na § and by extracellular K+; (3) the cation concentrations required for half-maximal activation of the Na+,K+-ATPase and cation transport are the same; (4) cardiac glycosides, which are impermeant, specifically inhibit the Na+,K§ and cation transport without directly affecting other transporters or fluxes; (5) cardiac glycosides in-
23. Membrane Transport in Red Blood Cells
381
I
I
I
I
I
I
5
10
15
20
25
30
120
v
E
80
o z
40
0
Time (days) FIGURE 3. Loss of K+ and gain of Na+ when the Na+-K+ pump of human red blood cells is completely inhibited. (Theoretical plot by P. R. Pratap using the integrated model of Lew and Bookchin, 1986.)
hibit the Na+,K+-ATPase and cation transport half-maximally at the same concentration and with the same molecular specificity in that inhibition of both processes requires an unsaturated lactone group; (6) K + antagonizes the effect of cardiac glycosides on both the Na+,K+-ATPase and cation transport; and (7) the Na+,K+-ATPase is reversible, and net synthesis of ATP occurs when the Na + and K + concentration gradients are reversed. Moreover, the activity of the Na+,K+-ATPase parallels the magnitude of cation fluxes over a twenty-five thousandfold range in a variety of tissues, and active transport of Na + and K + against their respective concentration gradients has been observed directly by measurement of net fluxes in resealed ghosts. L/~uger (1995) considered that the fluctuating energy barriers in ion pumps and exchangers may be similar to those in ion channels (see also DeFelice and Blakely, 1996). With this idea in mind, it is of considerable interest that palytoxin, isolated from a marine soft coral, reversibly increases the cation conductance of red cell membranes over the native C1- conductance, and that the effects of the toxin are totally blocked by external ouabain and prevented by internal vanadate. In the presence of the toxin, the conductive cation selectivity becomes K + > Rb + > Cs + > Na + > Li +, corresponding to Eisenman sequence IV (see Chapter 1). It appears that palytoxin "opens a 10 pS c h a n n e l . . , at or near each pump site" (Tosteson et al., 1991). Although this channel probably represents a perturbed form of the native cation permeation pathway in the Na+,K+-ATPase, further characterization might provide new insights as to how channels might function within pumps. A related problem in red cells concerns the relationship of DIDS-sensitive C1- conductance to the DIDSsensitive C1--HCO~ exchanger, as described in Section VII. Whereas the Na+-K + pump is by far the most widely distributed transporter that regulates the monovalent cation composition of cells, the red blood cells of carnivores (e.g., dogs, cats, ferrets, bears) contain high Na + and low K + (see Table 1). The immature nucleated red cells of dogs do contain the usual high K + and low Na + contents and Na+-K + pumps, but during maturation primary active transport of
Na + and K + ceases. In mature dog red cells, intracellular [Na +] is slightly less than extracellular [Na+], whereas intracellular [K +] is slightly greater than extracellular [K+]. These cells, like human red cells, also contain a potent Ca 2+ATPase in their membranes that maintains submicromolar concentrations of intracellular [Ca 2+] in the face of millimolar extracellular [Ca 2+] (for review, see chapter 17 by Vincenzi in Agre and Parker, 1989; see also Chapter 18). However, dog, ferret, and bear red cells, unlike human red cells, also contain a Na+-Ca 2+ countertransporter that couples the passive influx of Ca 2+ to the effiux of Na +, thus resulting in maintenance of a slight reduction in intracellular [Na +] in the steady state (see Chapter 19). The slight excess of intracellular K + is consistent with an inside negative membrane potential. In addition to the many structural and mechanistic studies on active pump fluxes mediated by the Na+,K+-ATPase and the Ca2+-ATPase, much progress has also occurred in understanding the mechanisms of passive cation transport (for review, see chapter 18 by Parker and Dunham in Agre and Parker, 1989). Unlike muscle and nerve, red cells have a low resting permeability to both Na + and K +, but a high conductive permeability to C1- that is about 100 times greater than that to cations. Consequently, C1- is in equilibrium with the resting membrane potential E m,
gr [C1-]i gm- T In Cl_-----~o [ where R is the gas constant, T is the absolute temperature, is the Faraday constant, and RT/~ is 25.5 mV at 23 ~ For the ratio of [C1-]i/[C1-]o = 77/116 = 0.66, the resting potential E m is -11 mY, inside negative. In the steady state, ionic currents of K + and Na + both flow passively across the membrane down their respective concentration gradients. Assuming that the currents of monovalent ions are independent, that the membrane is symmetrical, that the transmembrane electric field is constant, and that ions first partition into the membrane and then diffuse across driven by concentration and electrical gradients, the theory of electrodiffusion gives the membrane potential as
RT PK[K+]i+PNa[Na+]i+Pc'[C1-]o E m= - ~ In PK[K+]o + PNa[Na+]o + Pcl[C1-]i where PK, PNa' and Pc1 are the constant field permeabilities (s-l). Since C1- is at equilibrium, we may substitute Cli = C1~ eem~/gT into the preceding Godman-Hodgkin-Katz equation. Rearranging and simplifying gives Em
=-~
RT
In
PK[K+]i+ PNa[Na+]i PK[K+]o+ PNa[Na+]o
In other words, all permeant ions--cations and anions--have some relationship to the electrical potential E m across the membrane. Solving this expression for the ratio PK/PNa gives PK _ [Na+]oe-4'-[Na+]i PNa
[K+]i - [K+] o e-~b
where ~ is the reduced potential Em~/RT. Using standard values for human red cells from Table 1 for [K+]i, [K+]o,
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
382
Ouabain
K
due to the electrogenic Na § current is normally less than 1 mV, but is demonstrable (Fig. 4) using fluorescent dyes after optimizing the signal by increasing internal Na +, by increasing the membrane resistance through replacing cell C1with SO42-, and by inhibiting the anion conductance with DIDS (Dissing and Hoffman, 1990). As illustrated in Fig. 5, red cell passive fluxes, either conductive or electroneutral, are composed of at least nine phenomenologically separable components, a much more complex situation than assumed in classical theories (Van Slyke et al., 1923; Jacobs and Stewart, 1947; Tosteson and Hoffman, 1960). The pump-leak theory for red cells has been extended to model some of the effects of the additional cotransport pathways (Milanick and Hoffman, 1986; Lew and Bookchin, 1986). Pathways for passive electrolyte transport in human red blood cells include cotransporters for Na+-K+-2 C1- and K+-C1 - (for review, see Lauf et al., 1992; Hoffman and Dunham, 1995), countertransporters for Na+(Li+)-Na + (e.g., Duhm and Becker, 1977) and Na§ § (Escobales and Canessa, 1986), Ca2§ K + channels (for review, see Schwarz and Passow, 1983), C1--HCO 3 exchange mediated by capnophorin (AE1 or band 3 protein) (for review, see Knauf, 1979; Passow, 1986; chapter 19 by Gunn et al., in Agre and Parker, 1989; Jennings, 1992; Reithmeier, 1993), C1-conductance (Freedman et al., 1988, 1994; Freedman and Novak, 1997), and HC1 cotransport (Bisognano et al., 1993). Specific red-cell transporters also mediate the transmembrane movement of glucose, nucleosides, lactate and other organic anions, oxidized glutathione, choline, and amino acids (for review, see Lefevre, 1961; the chapters by Srivastava and by Martin in Ellory and Lew, 1977; Agre and Parker, 1989; Raess and Tunnicliff, 1990). Water flows through pores made of aquaporin (for review, see Agre, 1996). Some of the most frequently used inhibitors in studies of red-cell membrane transport are listed in Table 2. In isotope flux studies, active transport of Na § and K § has often been equated with the "ouabain-sensitive fraction" of Na + effiux or K + influx. Similarly, Na+-K+-2C1 - cotransport is operationally measured as the "ouabain-insensitive, bumetanidesensitive" fraction of a Na + flux, whereas Na§247247 countertransport is the "ouabain- and bumetanide-insensitive,
1
1
LL.
~,
o
-1 ---
-1
m,
|
0
||
1
i
I
2
,3
T i m e (rain)
FIGURE 4. Electrogenic potential of the Na§ § pump in human red blood cells, as monitored with the oxonol dye WW781. Intracellular C1- has been replaced with SO42-, followed by treatment with DIDS. In Na§ medium, the pump is activated by addition of 10 mM K§ and then inhibited with 33 millimolal ouabain, as indicated. An upward deflection indicates hyperpolarization. (Data of P. R. Pratap and J. C. Freedman.)
[Na+]i, [Na§ , and ~b yields PK/PNa = 1.5, much less than the corresponding ratio in nerve and muscle. Thus, the resting cation diffusion potential of human red blood cells is consistent with the membrane having a slightly greater constantfield permeability for K + than for Na +. However, the measured ground permeabilities to Na + and K + appear to be nonselective and more consistent with a single-barrier model than with electrodiffusion (Zade-Oppen et al., 1988). Away from the steady state, such as following cell shrinkage by exogenously added extracellular sucrose, the gradient of the highly permeant C1-will dominate and determine the membrane potential of human red blood cells until a new steady state is reached (see Bisognano et al., 1993). The electrogenic current of the Na§247 is one-third of the active Na § effiux; this current flows across the membrane electrical resistance, which for normal human red blood cells has been estimated at about 10 6 ~~ c m 2 (for review, see Hoffman et al., 1980). The electrogenic potential
Ca2 §
Na* Ca 2*
3Na +
K§
- ATP
CI-
ADP, Pi CI-
~
~
f
K§ \ ~
. C I - ADP, K§ 2K §
/ ~ HCO~-
X
2CI-
Na* r~-
-
~.._...~
FIGURE 5. The principal membrane transport systems of human red blood cells. Starting with the Na+-K+ pump and the passive leakage pathways for Na+ and K+ on the right, and proceeding clockwise around the membrane, are K+-C1- cotransport, Na§ K+-2C1- cotransport, C1--HCO3 exchange, the Ca2§ pump with a Ca2§ leakage pathway, the Ca2§ K§ channel, and C1conductance. The steady-state concentrations of K § Na§ and Ca2§ represent a balance between active pumping and passive leakage. C1- and HCO 3 are passively distributed at equilibrium with a membrane potential of-11 mV.
383
23. Membrane Transport in Red Blood Cells
TABLE 2
Some Inhibitors of M e m b r a n e Transport in Red Blood Cells Transporter
Pumps Na+,K+-ATPase Ca2+-ATPase GSSGa Cotransport Na+_K+_2C1K+_C1H+_C1Countertransport C1--HCO3Na+(Li+)-Na+ Na+_H+ Facilitated diffusion Glucose Nucleosides Lactate Choline Channels Ca2+(K+)
Inhibitor Ouabain, vanadate La3+, vanadate, eosin Fluoride Bumetanide, furosemide DIOAb Calcyculin A, okadaic acid DIDSc DIDSc, DNDSa Phlorizin, phloretin Phloretin Amiloride Phlorizin, phloretin Cytochalasin B Nitrobenzylthioinosine PCMBSe, DTNBf NEMg Charybdotoxin
aOxidized glutathione b[(Dihydroindenyl)oxy]alkanoic acid c4,4'-Diisothiocyano-2,2'-disulfonic acid stilbene d4,4'-Dinitro-2,2'-dinitrosulfonic acid stilbene
epara-Chloromercuribenzenesulf onate f5,5'-Dithio-bis(2-nitrobenzoate) gN-Ethylmaleimide
phloretin-sensitive" fraction. Fluoride and phloretin are relatively nonspecific and act on more than one transporter. Eosin is a relatively new inhibitor for the Cae+-ATPase. Some inhibitors bind reversibly to the transporter (e.g., DNDS to band 3), whereas others inhibit irreversibly by formation of covalent bonds (e.g., DIDS to band 3). Inhibitors may also act indirectly on associated enzymes that modulate transport. For example, the inhibition of a Type 1 protein phosphatase by calyculin A or okadaic acid prevents the activation of K+-C1- cotransport during cell volume regulation.
VI. Ionic and Osmotic Equilibrium and Cell Volume Regulation Human red blood cells contain 66% water (100 • g water/g packed cells), as is easily determined by drying a weighed pellet of sedimented cells to constant weight in a vacuum oven, taking into account suitable corrections for trapped volume. According to the Boyle-van't Hoff law, which is derived from the Gibbs equation (see Dick, 1959), the osmotic pressure ('w) of a solution containing N moles
of solute dissolved in V w liters of water is given by
chNRT "w- ~ = qScRT Vw
where ~b is the osmotic coefficient of the solute, R is the gas constant, T is the absolute temperature, and c (= N/V w) is the total concentration of dissolved solutes. The osmolality 4~c is the total osmolal concentration of all dissolved solutes (see Chapter 21). The osmotic coefficient ~b is unity for ideal dilute solutions; deviations of r from unity are due to solute-solvent interactions and other nonidealities. The osmotic pressure "wi of the intracellular solution is the sum of the osmotic pressures of each component of the mixture,
'We ~ dpj Nj RT =RT~qbjC. vw
where cj (= N j / V w) is the concentration of the jth solute. Alternatively,
N T RT V - ~ cr RT
'Wi--
w h e r e (l) i is the osmotic coefficient of the intracellular mixture, a parameter that equals the sum of the osmotic coefficients of the components of the mixture, weighted according to their respective mole fractions. Water will cross the membrane until the intracellular osmotic pressure equals the extracellular osmotic pressure, or "wi = "wo,as expressed by
E c~j,i f j, i - E c~j,o f j, o An isosmotic solution is defined as having an osmolality equal to that of normal plasma, or 289 milliosmolal; an isotonic solution will maintain the normal volume of cells incubated during physiological experiments. In isotonic solution, designated by the superscript 0,
0
"Wi--
vo
Dividing "we by "w0, setting "we "- 7"/'0and "w0_ 3T0, and rearranging yields the following expression for the equilibrium cell water content relative to that of cells in isotonic solution: 0
V
1to m
0 ~
This expression quantitates how much cell shrinkage will occur in hypertonic solution (VJV~~ < 1 when "wo> "wo o), and how much cell swelling will occur in hypotonic solution (Vw/V~ > 1 when "wo> "wo ~ The osmotic coefficient of hemoglobin 4~Hbrises approximately with the square of hemoglobin concentration (Fig. 6B; see also Ross and Minton, 1977), resulting in significant deviations of red cells from ideal osmotic behavior (Dick, 1959; Freedman and Hoffman, 1979a). For an ideal osmometer, 4~Ub and the osmotic coefficients of other solutes would be unity. At a normal [Hb] of 7.3 millimolal, 4~Hbis 2.85; in contrast, the osmotic coefficients of KC1 and NaC1
384
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
*r ' 1 2
0.70
4) 0
==
A
-
..,.....
W
Human .....
4)
4) 0
o
1.0
0 .,.=
J~.__
0
__.__.
-_ .
_ . . . .. .
_~ ~._ ~ ~.-~ - --- _ - . ._ ._ ~
Na __CI _._
E
o
O.8
0.0
I 0.2
,.r
.
KCI
] 0.4
Hypotonic
O O}
I 0.6
Isotonic
c m e~ o o
.--- -.
0.65
L_
9
o
Concentration (mMolal)
Hypertonic
0.60
-B
--/~ 0
r-
4
. 4) m o 4)
o 0
I
, 2
8
Time 4
8
0
B
(hr)
Amphiuma
"o 2 . 5
0
E W o 0 0
I .... 5
I 10
,
I 15
Concentration (mMolal)
FIGURE 6. (A) Osmotic coefficients of NaC1, KC1, and sucrose. (B) Osmotic coefficient of hemoglobin. (Data for NaC1, KC1, and sucrose are from the "CRC Handbook of Chemistry and Physics," 58th Ed. (1978-79) (R. C. Weast and M. J. Astle, Eds.), p. D-261, and from R. A. Robinson and R. H. Stokes, "Electrolyte Solutions," 2nd Ed. (1959). Butterworths, London. Data for hemoglobin are from Adair, as adapted from Fig. 7 of Freedman and Hoffman, (1979), J. Gen. Physiol. 74, p. 177, by copyright permission of The Rockefeller University Press, and is drawn according to ~bHb= 1 + 0.0645[Hb] + 0.0258[Hb]2.)
decrease with increasing concentration and are 0.93 at physiological concentrations, whereas that of sucrose rises slightly with increasing concentration (Fig. 6A). Mature human red blood cells maintain stable volumes when placed in isotonic solution, and rapidly adjust their volumes to new stable levels that are maintained over time in hypotonic or hypertonic salt solutions (Fig. 7A). In contrast, human reticulocytes and red cells from many other species, including duck, dog, the salamander Amphiuma, trout, skate, rabbit, and a low-K + genetic variant of sheep, all possess the capability of altering their membrane permeability in response to an altered extracellular osmolarity in such a way as to alter their total intracellular solute concentration N 7. and return their cell volumes back to normal (for review, see Hoffmann and Dunham, 1995; see also Chapter 21). Examples of the volume r e g u l a t o r y decrease (VRD) and volume regulatory increase (VRI) exhibited by Amphiuma red blood cells are shown in Fig. 7B. The swelling-activated response usually involves an increase in K+-C1- cotransport, as in duck, sheep, and rabbit red cells, but may also involve formation of a channel related to band 3 that permits effiux of osmolytes such as taurine, as in red cells from skate and trout. An exception seems to be net KC1 effiux resulting from activation of K§ + exchange coupled with C1--HCO 3 exchange, as proposed for Amphiuma red cells. The shrinkage-activated response involves activation of Na§ - cotransport in
r
VRD
4) r
0 20 O
"
~
t_
V
R
I
4)
1.5 0
I 50
I ...... 100 Time
J 150
(min)
FIGURE 7. (A) Stable osmotic response of human red blood cells. Washed human red cells were incubated in hypotonic, isotonic, or hypertonic media, and the water contents were determined at selected times by drying packed cells to constant weight in a vacuum oven. (Data of J. C. Freedman and S. Fazio.) (B) Volume regulatory decrease, VRD (upper trace), or volume regulatory increase, VRI (lower trace), when Amphiuma red cells are incubated in hypotonic or hypertonic media, respectively. For red cells from Amphiuma, but not from humans, the red-cell water content spontaneously returns back toward its normal value. (Bottom panel adapted from Figs. 1 and 2 of P. M. Cala, The Journal of General Physiology, 1980, Vol. 76, pp. 688-689.)
duck red cells, Na+H + countertransport in parallel with C1-HCO 3 exchange in Amphiuma red cells, and Na+-H + exchange in dog red cells. The mechanisms by which cells "sense" their altered volume prior to switching on a transport pathway, and "sense" their normal volume when turning off a transport pathway, are the subjects of current active investigation. Thus, the amount of water in red blood cells is determined by the total concentration (osmolality) of salts (cations and anions), and by proteins and organic metabolites, as well as by the nonideal effects expressed by the osmotic coefficients of hemoglobin and other intracellular solutes. An understanding of the G i b b s - D o n n a n e q u i l i b r i u m is critically important for properly incubating red cells and for understanding their response to altered pH and osmolarity. Human red blood cells, for example, swell in acid and shrink in alkaline media (Fig. 8). Red cells swell and hemolyze on exposure to compounds such as nystatin that elevate
23. Membrane Transport in Red Blood Cells
385
0
,_ "~
!
0.7
i
I
'
!
.......
'
C1; _ Na + + K~- + zHb rC1- - -
~
06 ~
_
= eE,,,~/RT
Ko++ Nao+
O
05 -
:
'
i
'
I
1.4 "to
"
1.2
"~ 1 0 d ,,,,,
o
Cl o
This expression illustrates most clearly that the doubleDonnan equilibrium consists of one Donnan equilibrium being established by the slowly permeant ("functionally impermeant") cations, and the other by hemoglobin (Hb) and organic phosphates. The net charge on hemoglobin depends on intracellular pH as its constituent amino acids are titrated. The titration curve of hemoglobin is approximately given by
z--m(PHi-PI ) 0,4
oo 0
0.2 0.0
n 6
I 7
I 8
I 9 9
_
10
PHo
F I G U R E 8. Comparison of measured (points) and predicted (lines) Donnan ratios in human red blood cells. The equilibrium Donnan ratios for C1- (bottom panel) and the cell water contents (top panel) were determined at varied extracellular pH. (Adapted from Fig. 3 of R. B. Gunn, M. Dalmark, D. C. Tosteson, and J. O. Wieth, (1973). J. Gen. Physiol. 61, p. 193, by copyright permission of The Rockefeller University Press.) The solid lines are predictions of the program IONIC. (Adapted from Fig. 3 of J. C. Freedman and J. E Hoffman, (1979) J. Gen. Physiol. 74, 174, with copyright permission of The Rockefeller University Press.)
the normal low permeability of the membrane to cations, a phenomenon termed colloid osmotic hemolysis by Wilbrandt. As a human red blood cell swells, its membrane surface area remains constant as the biconcave disc gains water and converts to a spherical shape at the critical hemolytic volume, which is 1.65 times the original volume (for review, see Hoffman, 1992). The membrane, which has a surface tension of about 1 dyn/cm, but an area expansion modulus of-+450 dyn/cm, is unable to withstand more than about 4% stretching (see chapter 15 by Berk, Hochmuth, and Waugh in Agre and Parker, 1989), and consequently ruptures with further swelling. A Gibbs-Donnan equilibrium occurs whenever a membrane separates an aqueous salt solution from a salt solution that also contains impermeant charged electrolytes, such intracellular charged proteins and organic phosphates (see Chapter 15). For human red cells, the permeability of the membrane to Na + and K + is 100 times less than that to C1- and HCO3-, so that certain cellular responses can be understood by considering the cells to approximate a double-Donnan system. From the principle of bulk electroneutrality in the extracellular and intracellular solutions, C l o - K+o+ Na+ and C 1 ; - Na + + K~- + z Hb where z is the average net charge (Ixeq/mol Hb) on impermeant cell solutes including hemoglobin and organic phosphates. The Donnan ratio rcl, for C1- is
where m is the slope of the titration curve or buffer capacity (m = -12 eq/mol Hb4/PHo), and pI is the isoelectric point of the cell contents, or the pH i where the net charge is zero (pI = 6.8 at 25 ~ in human red cells). In ectothermic vertebrates, maintaining the constancy of the charge z on intracellular proteins may be more important for homeostasis than the constancy of extracellular pH (for review, see Reeves, 1977). If HC1 is added to a suspension of red cells, the intracellular pH will decrease, hemoglobin will become more positively charged, and since Na + and K + are so much less permeant than CI-, the highly permeant anion will enter the cells to balance the increased positive charge on hemoglobin. Since the total intracellular solute concentration has now increased because of entry of CI-, the red cells swell (see Fig. 8). In contrast, if NaOH is added to a suspension of red cells, the intracellular pH increases, hemoglobin becomes more negatively charged and C1- leaves the cell, resulting in cell shrinkage (see Fig. 8). Model calculations show that at the isoelectric pH, the sum of the cation concentrations is greater in plasma than in the intracellular solution, in accordance with the double-Donnan concept; at the physiological pH of 7.4, however, cell shrinkage raises the intracellular cation concentration so that it nearly equals the plasma cation concentration, thus obscuring the doubleDonnan effect. In human red cells, NO 3, I-, SCN-, and methane sulfonate (CH3SO3), but not methyl sulfate (CH3SO4), all bind to hemoglobin and alter its net charge, with a consequent small but significant change in cell volume in accordance with the Gibbs-Donnan equilibrium (Payne et al., 1990). The Donnan equilibrium shifts in the expected manner on experimental alteration of intracellular phosphates (Duhm, 1971) and in abnormal human red cells containing HbC, which has an increased positive charge (Brugnara et al., 1985). Colloid osmotic hemolysis occurs when the red-cell membrane is exposed to pore-forming ionophores such as nystatin or gramicidin, or when the normal low permeability to cations increases on exposure to detergents or other chemicals. At equilibrium at the isoelectric pH, where the Donnan ratio equals unity and the membrane potential is zero, the intracellular concentrations of Na +, K +, and C1would equal those of the extracellular solution. The osmotic pressure of hemoglobin would be unbalanced by any extracellular solute and would draw water into the cell, leading to swelling and hemolysis. Colloid osmotic hemolysis occurs at any pH when the membrane is permeant both to cations and to anions. The Na + pump, working against the
386
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
leakage pathways and any significant cotransport pathways, establishes the steady-state intracellular cation concentrations, while C1- and HCO 3 move to maintain electroneutrality. In red blood cells, for given steady-state concentrations of K + and Na +, as set by the balance of pumps and leaks, the equilibrium cellular electrolyte concentrations and water contents are determined by the following three principles: 1. Bulk electroneutrality 2. Osmotic equality of the intracellular and extracellular solutions 3. Equality of the Donnan ratios for all permeant ions of the same charge The Donnan ratio for monovalent anions is equal to the inverse of the Donnan ratio for protons or any other highly permeant monovalent cations. For human red blood cells, the nonideal equations expressing these constraints, including the activity and osmotic coefficients of the salts and of hemoglobin, have been formulated into a computer program designated IONIC (Freedman and Hoffman, 1979a; B isognano et al., 1993; see also Raftos et al., 1990). For a given composition of the extracellular medium, and certain cellular parameters, the program predicts the cell volume, the intracellular pH and ion concentrations, the membrane potential, and the charge on hemoglobin and organic phosphates. The predictions of the model generally agree with experimental results within a few percent; for example, the predicted and experimentally determined dependence of the Donnan ratios for C1- on pH are shown in Fig. 8. For a
FIGURE 9.
Jacobs-Stewartcycle for red blood cells.
model describing blood acid-base status at constant or variable temperature, see Rodeau and Malan (1979). Beginning with a model of the Gibbs-Donnan equilibrium, Lew and Bookchin (1986) developed an integrated model of red-cell transport that includes kinetic parameters describing most of the known transport systems, and begins to describe the time course of the changes in cell volume, intracellular salt concentrations, and membrane potential that occur when the extracellular conditions are changed.
VIi. Anion E x c h a n g e and C o n d u c t a n c e The transport of carbon dioxide from the peripheral tissues to the lungs is facilitated by the production of bicarbonate by red blood cells in a cyclic reaction scheme known as the Jacobs-Stewart cycle. In systemic tissue capillaries, carbon dioxide diffuses easily across the red cell membrane into the cells, whereupon carbonic anhydrase (CA) catalyzes its hydration to carbonic acid (Fig. 9). The acid then dissociates into a proton and bicarbonate. The proton is buffered by hemoglobin, which when protonated has a reduced affinity for oxygen, a phenomenon known as the Bohr effect. The bicarbonate passes across the red cell membrane into the venous blood by undergoing a rapid electroneutral exchange with CI-, a reaction known as the chloride shift. A small proportion of carbon dioxide forms a covalent carbamino compound with amino groups on hemoglobin. Production of HCO 3 by red blood cells minimizes acidification of venous blood and increases its carbon dioxide car-
23. MembraneTransport in Red Blood Cells
rying capacity. Red cell C1--HCO 3 exchange is mediated by the membrane domain of capnophorin, a protein also denoted as anion exchanger AE 1. This integral glycoprotein, having a molecular weight of 101 700 Da, makes up 25% of the mass of membrane protein in red blood cells (Steck, 1974) and is the predominant protein located in band 3 of SDS polyacrylamide gels. DIDS is the most potent of a series of stilbene derivatives originally found by Cabantchik and Rothstein to inhibit red cell anion transport. DIDS inhibits 99.999% of C1--HCO 3 exchange, but only 67% of net C1- effiux. Molecular biological experiments have identified the location of the DIDS binding site. The cDNA from mice and humans that codes for band 3 protein (AE1) was cloned and sequenced (for review, see Alper, 1991). Band 3 protein has also been functionally expressed in X e n o p u s laevis toad oocytes microinjected with mRNA prepared from human and mouse cDNA clones. Site-directed mutagenesis of the mouse protein indicates that one of the isothiocyanate (NCS) groups on DIDS binds covalently to Lys 558, which is located on the extracellular side of the 65-kDa chymotryptic N-terminal fragment. The stoichiometry of binding is one DIDS per capnophorin monomer (for review, see Passow et al., 1992 The membrane domain of capnophorin has been reconstituted with lipids and crystallized in two-dimensional arrays (Reithmeier, 1993; Wang et al., 1994). At 20 ~ resolution, reconstructed images indicate a dimeric structure with a cavity between the two monomers of band 3.35C1 nuclear magnetic resonance studies of capnophorin suggest an hourglass type of structure in which internal and external hemichannels lead to the transport site, which can exist in inward- and outwardfacing conformations (Falke and Chan, 1986b). C1--HCO 3 exchange follows Ping-Pong kinetics in which a conformational change occurs only after binding an anion alternately from the inward- and outward-facing conformations (for review, see Frrhlich and Gunn, 1986). DIDS lies in the outer hemichannel between the transport site and the extracellular medium, partially blocking the outward-facing transport site (Falke and Chan, 1986a). Structure-activity studies with a series of analogs of DIDS suggested a model of the DIDS binding site that included positively charged groups providing electrostatic stabilization of the sulfonates on DIDS, with adjacent hydrophobic and electron-donor centers. Experiments using the technique of fluorescence resonance energy transfer indicate that DIDS is located only 34--42/~ from sulfhydryl reagents bound to cysteine residues on the 40 000-Da aminoterminal cytoplasmic domain of capnophorin, a fmding that is also consistent with DIDS residing in a cleft in the outer hemichannel. Moreover, substrate anions such as C1- traverse only 10-15% of the full transmembrane electric field when they move from the extracellular medium and bind to the outward-facing transport site (Jennings et al., 1990). This observation is consistent with the hourglass model and indicates a low electrical resistance for the outer hemichannel, because if the resistance of the outer hemichannel were high, the transmembrane voltage would have dropped more than the 10-15% that was observed. The classical view that the distribution of red-cell C1- is simply determined by the double-Donnan equilibrium was consistent with the observations that 38C1 equilibrates within
387
1 s, which is 106 times faster than K § and that the Donnan ratios for CI-, HCO 3, and OH- are all equal. The concept of rapid net movement of C1- across the red-cell membrane changed dramatically when experiments with gramicidin and valinomycin revealed that C1- conductance is about 104 times slower than electroneutral C1- exchange, yet still about 102 times greater than the cation conductance. Thus net transport of cations normally occurs over a time course of hours, net movement of C1- occurs over minutes, and exchange of C1- for HCO 3 occurs in seconds. In addition to electroneutral anion exchange, capnophorin is also believed to mediate the net flux of CI-, which could contribute to loss of salt and water during pathophysiological dehydration of the red cell, such as occurs during the formation of irreversibly sickled cells. Studies have revealed DIDS-sensitive and DIDS-insensitive components of the net C1- fluxes, and the DIDS-insensitive component increases with increasing extents of membrane hyperpolarization (Freedman et al., 1988, 1994; Freedman and Novak, 1997). Because it has not proven possible to use microelectrodes to drive currents or to clamp the voltage across the membrane of human red cells, ionophores have been used instead to increase the permeability to cations, thus permitting ionic currents to flow under the influence of diffusion potentials, which can then be measured indirectly. Consider the equivalent circuit shown in Fig. 10. The concentration gradients of K § Na § and C1- across the membrane are represented by batteries (E K, ENa, and Ecl) in parallel with the membrane capacitance C m. Each battery is in series with its respective ionic conductance (gK, gNa' and gcl)" The electrical potential inside the cell is E m, while that outside is at ground. Treatment of red cells with valinomycin or with gramicidin in Na+-free medium, increases the normally low value of g~: above that of gcl and allows the flow of the ionic currents, i~:
n
~
"T c--lL. gK -q ~
gcI
ENa ~
Cm
gNa
Out
m
FIGURE 10. Equivalent electrical circuit for human red blood cells treated with valinomycin,or with gramicidin in Na§ medium. Arrows indicate the direction of positive current flow when the potassium conductance gr~ is increased by the addition of valinomycin or of gramicidin in Na+-free medium.
388
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
and icy around the loop (see Fig. 10, arrows). At low [K+]~ the net effiux of K + corresponds to the outward current iK, which discharges the potassium battery. The magnitude of iK depends on [K+]o, which is easily varied. The effiux of K + hyperpolarizes the membrane potential from its resting value near-11 mV to a value of some -60 mV, thus driving a net effiux of CI-, corresponding to an inward current ic] that charges the chloride battery. The K + and C1-currents are opposite in sign and approximately equal in magnitude; a small disparity of less than 10% is accounted for by a proton current. Whereas the total net current across the membrane must be zero in the absence of an external circuit, the individual ionic currents induced by the addition of ionophores can be either measured directly or inferred from the rate of cell shrinkage. Fluorescent potentiometric indicators that monitor the voltages continuously show that the voltage remains at a steady value after the addition of valinomycin or gramicidin. Steady levels of dye fluorescence are seen with the oxacarbocyanine dye, diO-C6(3 ) (Hoffman and Laris, 1974), or with the thiadicarbocyanine, diS-C3(5 ) (Fig. 11), or indodicarbocyanine, diI-C3(5), dyes (Freedman and Hoffman, 1979b; Freedman and Novak, 1989). Human red blood cells treated with valinomycin or with gramicidin thus behave as if voltage-clamped by their intrinsic ion concentration batteries, E K and Ecl, instead of by an external circuit. This alternative voltage-clamp experiment yields currentvoltage curves characterizing the conductance to C1-. The change in voltage upon addition of ionophores has been estimated from the change in the extracellular pH of unbuffered DIDS-treated red-cell suspensions (Macey et al., 1978) using the proton ionophore FCCP, a method that is also useful for calibrating fluorescent potentiometric in-
OiS-C~(5)
Ko
(rnM)
20 --'--
+Val
--
-
dicators (Freedman and Novak, 1989; Bifano et al., 1984). The inward-rectifying current-voltage curve found for the DIDS-insensitive fraction of C1- net transport (Fig. 12) deviates markedly from the nearly linear outward-rectifying curve predicted by the constant field theory for the inward concentration gradient of C1-. The inward-rectifying current-voltage curve is also inconsistent with certain single-occupancy multiple-barrier models, but instead is consistent either with a single barrier located near the center of the transmembrane electric field, or with a voltagegated mechanism according to which half the channels are open a t - 2 7 mV, and with an equivalent gating charge of -1.2 + 0.3 (Freedman and Novak, 1997). Further experiments are needed to clarify the relationship of DIDS-insensitive C1- conductance inferred for red cells voltage-clamped with ionophores to the anion-selective channels recorded with the patch-clamp technique (Schwarz et al., 1989), and to those seen after fusing vesicles from red-cell suspensions into planar lipid bilayers (Freedman and Miller, 1984). It is also important to know if the DIDS-sensitive fraction of C1conductance is indeed mediated by the band 3 exchanger, in which case further studies could attempt to discover a unified kinetic scheme for exchange and conductance mediated by the same transporter.
VIIi.
Cytotoxic C a l c i u m
Cascade
Despite the importance of Ca 2+ as a trigger and modulator in a variety of cell activities (see Chapter 1), too much intracellular Ca 2+ is harmful to cells in general and to red blood cells in particular. Elevated intracellular Ca :§ stimulates a cytotoxic cascade of pathophysiological and biochemical changes. Some of these events are depicted in Fig. 13. Subjecting human red blood cells to shear stress increases the passive influx of Ca 2+ as seen in suspension (Larsen et al., 1981), or during flow through synthetic microlattices (Brody
150
~
A
c
0
m .c r
~
. . . . . .
-
100
.
Icm ( m A / g
Hb)
v
u
C 4) (,1 W 4) k. 0
-20
3 -
m
,--;- - 4 0
-100
m
n -J~-- ,
-
,
,
50
100
Em- Ecl -80 0
' 50
, 100
, 150
, 200
, 250
, 300
T i m e (s)
FIGURE 1 1. Steadylevels of fluorescence of diS-C3(5 ) following the addition of valinomycin to suspensions of human red blood cells in x mM KCI (as indicated), 150 - x mM NaC1, and 5 mM Hepes buffer (pH 7.4 at 23 ~ Downward deflections indicate hyperpolarization, whereas upward deflections indicate depolarization. (Adapted from Fig. 2 of Freedman, J. C., and Novak, T. S. (1989). Methods Enzymol. 172, 108.)
_.[
(mv)
FIGURE 12. Current-voltage curve for DIDS-insensitive C1- conductance of human red blood cells. The solid line represents the best fit either of a single-barrier model, or of a mechanism involving voltage gating of ionic channels. (Adapted from Fig. 5 of Freedman, J. C., and Novak, T. S. (1997). J. Gen. Physiol. 107, 207, with copyright permission of The Rockefeller University Press.)
389
23. Membrane Transport in Red Blood Cells
I Shear stress I !
[ Tca9 1
\
Hyperpolarization
Te, [Echin~
I
I!Tpl' , '
11N---a:K~se [ Cell shrinkage l Kc, I CIc, i H20 / I Spherocytosis
Prolytic state TNa c
]
PL Asymmetry T Phospholipase C, DG, PA
IHemolysis[ T Transglutaminase T Protease FIGURE 13. The cytotoxic calcium cascade of human red blood cells. The scheme illustrates some of the effects of elevated intracellular Ca 2+. (Adapted from Fig. 3 of Freedman, J. C. et al., (1988). "Cell Physiology of Blood." p.222, with copyright permission of The Rockefeller University Press.)
et al., 1995). When intracellular Ca 2+ rises to about 3 IxM, the K + permeability dramatically increases, a phenomenon first described by Gardos and known as the Gardos effect. The increased K + permeability is due to the opening of a CaZ+-activated K + channel, resulting in loss of intracellular K + and CI-, accompanied by cell shrinkage (for review, see Schwarz and Passow, 1983). Elevated [Ca2+]c also stimulates the CaZ+-ATPase, causing depletion of ATP within minutes (Crespo et al., 1987), which in turn inhibits the Na+,K +ATPase. In addition to causing substrate depletion, Ca 2+ also inhibits the Na+,K+-ATPase with a potency that depends on a regulatory factor (for review, see Yingst, 1988). At higher levels of intracellular Ca 2+, the smooth biconcave discoidal form of human red cells converts first to an echinocytic form with spicules protruding from the membrane, and then to a spherocytic form with release of microvesicles. ATP depletion is prevented by inhibiting the ATPases with vanadate, but echinocytosis still develops (Freedman et al., 1988). The echinocytogenic effect of Ca 2+ is a striking example of the profound influence that Ca 2+ can exert on cell morphology. At still higher levels, Ca 2+ activates a transglutaminase that cross-links cytoskeletal proteins and reduces cell deformability. Ca 2+ also activates degradative proteases and a phospholipase C that converts membrane phosphatidylcholine (lecithin) to diacylglycerol. The various effects of elevated intracellular Ca 2+ are induced at widely differing concentrations. Addition of A23187 stimulates loss of KC1 at 0.1 I~M free external Ca 2+, and the membrane hyperpolarizes within 2 s (Freedman and Novak, 1983). The threshold for CaZ+-induced echinocytosis (assessed by the morphologic index) is 2 orders of magnitude greater, at 10 IxM [CaZ+]f. It is possible to set external Ca 2+ at a level sufficient to induce hyperpolarization, but below the threshold for echinocytosis, thus showing that the membrane potential does not directly influence cell shape
(Bifano et al., 1984). Under these conditions, the cells are able to crenate because simply increasing extracellular Ca 2+ produces echinocytes. Under similar conditions, the Ca 2+induced decrease in cell deformability, as assessed by a decreased elongation index during flow at set shear stress (measured in an Ektacytorneter), occurs at 1 mM Ca 2+. The separate apparent potencies of Ca 2+, spanning 5 orders of magnitude, for inducing hyperpolarization, echinocytosis, and reduced deformability are consistent with the idea that Ca 2+ acts at independent sites to initiate these elements of the cytotoxic calcium cascade (Freedman et al., 1988). The relationships of the activation by Ca 2+ of a transglutaminase, a phospholipase C, and a protease to the loss of phospholipid asymmetry (see Lin et al., 1994), and to the changes in permeability, cell volume, and morphology are complex. Lipid alterations, including phosphoinositide hydrolysis and phosphatidic acid production, occur at 1-10 txM Ca 2+. Whereas IxM Ca 2+ dramatically affects the viscosity of complexes of spectrin, actin, and band 4.1 (Fowler and Taylor, 1980), the relative contributions of lipid and cytoskeletal alterations to the CaZ+-induced shape changes are still unclear. The final consequence of increased [Ca2+]C in vitro, after prolonged cell shrinkage due to loss of KC1, is the development of prolytic cells leaky to Na that then undergo colloid osmotic hemolysis. Experiments suggest that cell shrinkage, rather than Ca 2+ itself, results in the gain of Na + (Crespo et al., 1987). The observation that comparable gains of Na + occur in cells simply shrunken by sucrose, as upon exposure to valinomycin or to Ca 2+ plus calcium ionophore A23187 at low external K +, suggests that Ca 2+ is not essential in causing elevated [Na+]c. Other flux studies and assessment of relative cell volume distributions indicate that a subpopulation of cells becomes highly permeable to Na + after shrinkage, induced either osmotically or by CaZ+-induced loss of KC1. An apparent steady-state level of elevated [Na]c in the population of cells is attained because a fraction of the cells become prolytic while another fraction hemolyzes. An interesting question is what determines which cells enter the prolytic state. A role of the cytotoxic calcium cascade in the formation of dehydrated irreversibly sickle cells (ISCs) is indicated by the high levels of intracellular Ca 2+ sequestered in sickle cells, and by the ability of charybdotoxin and clotrimazole, two blockers of CaZ+-K+ channels, to inhibit the formation of ISCs (Ohnishi et al., 1989; Brugnara et al., 1996). It should be noted that K+-C1- cotransport has very low activity in normal mature red cells, but has elevated activity in immature reticulocytes and in sickle cells, and is also activated by acidosis (Brugnara et al., 1989). The relative contributions of K+-C1- cotransport and net loss of K + and C1after activation of CaZ+-K+ channels is an important question pertinent to the development of pharmacological interventions designed to ameliorate the painful vaso-occlusive crises suffered by patients with sickle cell disease.
IX. Summary The availability, structural and metabolic simplicity, and ease of manipulating the intracellular as well as the extra-
390
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
cellular solutions of red blood cells have made them a favorite subject for the study of membrane transport. The plasma membrane and underlying cytoskeleton have been extensively characterized. A flipase that transports phospholipids from the outer to the inner hemileaflet contributes to phospholipid asymmetry in the membrane. The high concentration of hemoglobin, nearly close-packed and hydrated with a large volume fraction of intracellular water, contributes to a high intracellular ionic strength. Curiously, the intracellular activities of KC1, NaC1, and nonelectrolytes appear to be nearly ideal. Red-cell metabolism supplies ATP as substrate for the Na+,K+-ATPase and Ca 2+ATPase ionic pumps and prevents the oxidation of Fe 2+ in hemoglobin. The life spans of red blood cells from various mammalian species correlate with body mass, suggesting the existence of a biological clock that is obviously independent of gene transcription or translation in the mature nonnucleated cells. The red blood cells of all species normally maintain a submicromolar intracellular concentration of Ca 2+, as set by the CaZ+-ATPase pumping against a small passive inward leakage. The red blood cells of most species have high steady-state concentrations of K + but low Na +, as set by a balance between pumping by the Na+,K+-ATPase and parallel leaks. An exception is the red cells of carnivores, which, unlike other red cells, have high intracellular Na + and low intracellular K +. These cells lack a Na+,K+-ATPase, but do have a Na+-Ca 2+ exchanger that couples the effiux of Na + to the influx of Ca 2+. Whereas human red blood cells are unable to regulate their volume in anisotonic media, the red cells of many other species use the gradients of K + and Na + and activate their K+-C1- and Na+-K+-2C1- cotransporters to return their volume to normal in hypotonic and hypertonic media, respectively. Alternatively, in red cells of trout and skate, an effiux of the osmolyte taurine is used in cell volume regulation. In human red blood cells, the conductive permeability to C1- is 100 times greater than that to K + or Na § Consequently, the membrane potential is consistent with the C1-equilibrium potential o f - 1 1 mV, and with a constant field permeability ratio PK/PNa,of 1.5. The electrogenic potential of the Na+,K§ is normally less than 1 mV. Away from the steady state, C1- will dominate the membrane potential. Treatment of human red blood cells with valinomycin, or with gramicidin in Na+-free media, results in clamping of the voltage by the intrinsic ion concentration batteries. Analysis of individual ionic currents flowing under the influence of diffusion potentials yields currentvoltage curves that characterize the conductance to C1-. An important problem for future study is to understand the "channel-like" properties of exchangers and pumps. According to the Boyle-van't Hoff law, the total intracellular and extracellular osmolalities are equal at osmotic equilibrium. Significant deviations from osmotic ideality are due to the osmotic coefficient of Hb rising with the square of Hb concentration. For given cation contents of human red cells, the double-Donnan equilibrium describes the passive distribution of C1- and water across the red-cell membrane and accounts quantitatively for the effects of pH on cell swelling and shrinkage, as well as for colloid osmotic
hemolysis. Computer models are available for (1) red cell glycolysis, (2) the nonideal double-Donnan equilibrium, (3) acid-base balance, and (4) the kinetics of membrane transport. These models begin to enable prediction of the time course of changes in the cell volume, the intracellular pH, the membrane potential, and the intracellular salt concentrations when the extracellular conditions are altered. When intracellular Ca 2+ rises too much for the Ca 2+ATPase to maintain the steady state, a cytotoxic calcium cascade leads to depletion of ATE inhibition of the Na+,K +ATPase, activation of CaZ+-K+ channels, cellular dehydration, loss of phospholipid asymmetry, echinocytosis, membrane vesiculation, activation of phospholipase C, transglutaminase, and protease, and hemolysis. Blockers of CaZ+-K+ channels reduce the formation of dehydrated, irreversibly sickled cells both in vitro and in vivo and are currently being tested as a new treatment for sickle cell disease.
Acknowledgment The author thanks Dr. Philip B. Dunham of the Biology Department, Syracuse University, for reading this chapter.
Bibliography Agre, R (1996). Pathophysiology of the aquaporin water channels. Annu. Rev. Physiol. 58, 619-648. Agre, P., and Parker, J. C. (Eds.). (1989). "Red Blood Cell Membranes: Structure, Function, Clinical Implications." Marcel Dekker, New York. Alper, S. L. (1991). The band 3-related anion exchanger (AE) gene family. Annu. Rev. Physiol. 53, 549-564. Antonini, E., and Brunori, M. (1971). "Hemoglobin and Myoglobin in Their Reaction with Ligands." North-Holland, Amsterdam. Bennett, V., and Gilligan, D. M. (1993). The spectrin-based membrane skeleton and micron-scale organization of the plasma membrane. Annu. Rev. Cell Biol. 9, 27-66. Bessis, M. (1974). "Corpuscles. Atlas of Red Blood Cell Shapes." Springer-Verlag, New York. Bessis, M., Weed, R. I., and Lebiond, P. E (Eds.). (1973). "Red Cell Shape. Physiology, Pathology, Ultrastructure." Springer-Vedag, New York. Beutler, E. (1986). "Red Cell Metabolism, Vol. 16. Methods in Hematology." Churchill Livingstone, New York. Bifano, E. M., Novak, T. S., and Freedman. J. C. (1984). The relationship between the shape and the membrane potential of human red blood cells. J. Membr. Biol. 82, 1-13. Bishop, C., and Surgenor, D. M. (1964). "The Red Blood Cell. A Comprehensive Treatise." Academic Press, New York. Bisognano, J. D., Dix, J. A., Pratap, R R., Novak, T. S., and Freedman, J. C. (1993). Proton (or hydroxide) fluxes and the biphasic osmotic response of human red blood cells. J. Gen. Physiol. 102, 99-123. Brody, J. P., Han, Y., Austin, R. H., and Bitensky, M. (1995). Deformation and flow of red blood cells in a synthetic lattice: evidence for an active cytoskeleton. Biophys. J. 68, 2224--2232. Brugnara, C., Kopin, A. S., Bunn, H. E, and Tosteson, D. C. (1985). Regulation of cation content and cell volume in hemoglobin erythrocytes from patients with homozygous hemoglobin C disease. J. Clin. Invest. 75, 1608-1617. Brugnara, C., Ha, T. V., and Tosteson, D. C. (1989). Acid pH induces formation of dense cells in sickle erythrocytes. Blood 232, 487-495. Brugnara, C., Gee, B., Armsby, C. C., Kurth. S., Sakamoto, M., Rifai, N., Alper, S. L., and Platt, O. S. (1996). Therapy with oral clotrimazole induces inhibition of the Gardos channel and reduction of cry-
23. Membrane Transport in Red Blood Cells
throcyte dehydration in patients with sickle cell disease. J. Clin. Invest. 97, 1227-1234. Crespo, L. M., Novak, T. S., and Freedman. J. C. (1987). Calcium, cell shrinkage, and prolytic state of human red blood cells. Am. J. Physiol. 252 (Cell Physiol. 21), C138-C152. Davson, H., and Danielli, J. E (1943). "The Permeability of Natural Membranes." Cambridge University Press, Cambridge, UK. DeFelice, L. J., and Blakely, R. D. (1996). Pore models for transporters? Biophys. J. 70, 579-580. Diaz, C., and Schroit, A. J. (1996). Role of translocases in the generation of phosphatidylserine asymmetry. J. Membr. Biol. 151, 1-9. Dick, D. A. T. (1959). Osmotic properties of living cells. Int. Rev. Cytol. 8, 387-448. Dissing, S., and Hoffman, J. E (1990). Anion-coupled Na effiux mediated by the human red blood cell N a ~ pump. J. Gen. Physiol. 96, 167-193. Dodge, J. T., Mitchell, C., and Hanahan, D. J. (1963). The preparation and chemical characteristics of hemoglobin-free ghosts of human erythrocytes. Arch. Biochem. Biophys. 100, 119-130. Duhm, J. (1971). Effects of 2,3-diphosphoglycerate and other organic phosphate compounds on oxygen affinity and intracellular pH of human erythrocytes. Pfliigers Arch. 326, 341-356. Duhm, J., and Becker, B. E (1977). Studies on the lithium transport across the red cell membrane. IV. lnterindividual variations in the Na+-dependent Li § countertransport system of human erythrocytes. Pfliigers Arch. 370, 211-219. Ellory, J. C., and Lew, V. L. (1977). "Membrane Transport in Red Cells." Academic Press, New York Ellory, J. C., and Young, J. D. (1982). "Red Cell Membranes--A Methodological Approach." Biological Techniques Series. Academic Press, New York. Escobales, N., and Canessa, M. (1986). Amiloride-sensitive Na § transport in human red cells: evidence for a Na/H exchange system. J. Membr. Biol. 90, 21-28. Falke, J. J., and Chan, S. I. (1986a). Molecular mechanism of band 3 inhibitors. I. Transport site inhibitors. Biochemistry 25, 7888-7894. Falke, J. J., and Chan, S. I. (1986b). Molecular mechanism of band 3 inhibitors. 2. Channel blockers. Biochemistry 25, 7895-7898. Fowler, V., and Taylor, D. L. (1980). Spectrin plus band 4.1 cross-link actin. Regulation by micromolar calcium. J. Cell Biol. 85, 361-376. Freedman, J. C. (1983). Partial requirements for in vitro survival of human red blood cells. J. Membr. Biol. 75, 225-231. Freedman, J. C., and Hoffman, J. E (1979a). Ionic and osmotic equilibria of human red blood cells treated with nystatin. J. Gen. Physiol. 74, 157-185. Freedman, J. C., and Hoffman, J. E (1979b). The relation between dicarbocyanine dye fluorescence and the membrane potential of human red blood cells set at varying Donnan equilibria. J. Gen. Physiol. 174, 187-212. Freedman, J. C., and Miller, C. (1984). Membrane vesicles from human red blood cells in planar lipid bilayers. Ann. NY Acad. Sci. 435, 541-544. Freedman, J. C., and Novak, T. S. (1983). Membrane potentials associated with Ca-induced K conductance in human red blood cells. Studies with a fluorescent oxonol dye, WW781. J. Membr. Biol. 72, 59-74. Freedman, J. C., and Novak, T. S. (1989). Optical measurements of membrane potential in cells, organelles, and vesicles. Methods Enzymol. 172, 102-122. Freedman, J. C.., and Novak, T. S. (1997). Electrodiffusion, barrier, and gating analysis of DIDS-insensitive chloride conductance in human red blood cells treated with valinomycin or gramicidin. J. Gen. Physiol. 109, 201-216. Freedman, J. C., Bifino, E. M., Crespo, L. M., Pratap, P. R., Wallenga, R., Bailey, R. E., Zuk, S., and Novak, T. S. (1988). Membrane potential and the cytotoxic Ca cascade of human red blood cells. In "Cell Physiology of Blood" (R. B. Gunn and J. C. Parker, Eds.), pp.
391
218-231. Society of General Physiologists Series, The Rockefeller University Press, New York. Freedman, J. C., Novak, T. S., Bisognano, J. D., and Pratap, P. R. (1994). Voltage dependence of DIDS-insensitive chloride conductance in human red blood cells treated with valinomycin or gramicidin. J. Gen. Physiol. 104, 961-983. Fr/Shlich, O., and Gunn, R. B. (1986). Erythrocyte anion transport: the kinetics of a single-site obligatory exchange system. Biochim. Biophys. Acta 864, 169-194. Gary-Bobo, C. M. (1967). Nonsolvent water in human erythrocytes and hemoglobin solutions. J. Gen. Physiol. 50, 2547-2564. Golan, D. E., and Veatch, W. (1980). Lateral mobility of band 3 in the human erythrocyte membrane studied by fluorescence photobleaching recovery: evidence for control by cytoskeletal interactions. Proc. Natl. Acad. Sci. USA 77, 2537-2541. Grimes, A. C. (1980). "Human Red Cell Metabolism." Blackwell, Oxford. Harris, J. W., and Kellermeyer, R. W. (1970). "The Red Cell. Production, Metabolism, Destruction: Normal and Abnormal," rev. ed. Harvard University Press, Cambridge, MA. Henderson, L. J. (1928). "Blood. A Study in General Physiology." Yale University Press, New Haven, CT. Hoffman, J. E (1986). Active transport of Na § and K § by red blood cells. In "Physiology of Membrane Disorders" 2nd ed. (T. E. , Andreoli, J. F. Hoffman. D. E Fanestil, and S. G. Schultz, Eds.), pp. 221-231. Plenum Press, New York. Hoffman, J. E (1992). On red blood cells, hemolysis and resealed ghosts. In "The Use of Resealed Erythrocytes as Carriers and Bioreactors" (M. Magnani and J. R. DeLoach, Eds.). Adv. Exper. Biol. Med., 326, 1-15. Plenum, New York. Hoffman, J. E, and Laris, P. C. (1974). Determination of membrane potential in human and Amphiuma red blood cells by means of a fluorescent probe. J. Physiol. (London) 239, 519-552. Hoffman, J. E, Kaplan, J. H., Callahan, T. J., and Freedman, J. C. (1980). Electrical resistance of the red cell membrane and the relation between net anion transport and the anion exchange mechanism. Ann. NY Acad. Sci. 341, 357-360. Hoffmann, E. K., and Dunham, P. B. (1995). Membrane mechanisms and intracellular signalling in cell volume regulation. Int. Rev. Cytol. 161, 173-262. Jacobs, M. H. (1962). Early osmotic history of the plasma membrane. In "Symposium on the Plasma Membrane, New York Heart Association, Inc." Circulation 26, 1013-1021. Jacobs, M. H., and Stewart, D. R. (1947). Osmotic properties of the erythrocyte. XII. Ionic and osmotic equilibria with a complex external solution. J. Cell Comp. Physiol. 30, 79-103. Jennings, M. L. (1992). Inorganic anion transport. In "The Structure of Biological Membranes" (P. Yeagle, Ed.), pp. 781-832. CRC Press, Boca Raton, FL. Jennings, M. L., Schulz, R. K., and Allen, M. (1990). Effects of membrane potential on electrically silent transport. Potential-independent translocation and asymmetric potential-dependent substrate binding to the red blood cell anion exchange protein. J. Gen. Physiol. 96, 991-1012. Kay, M. M. B. (1991). Drosophila to bacteriophage to erythrocyte: the erythrocyte as a model for molecular and membrane aging of terminally differentiated cells. Gerontology 37, 5-32. Knauf, P. A. (1979). Erythrocyte anion exchange and the band 3 protein: transport kinetics and molecular structure. Curr. Top. Membr. Transp. 12, 249-363. Kuhlman, P. A., Hughes, C. A., Bennett, V., and Fowler, V. M. (1996). A new function for adducin. Calcium/calmodulin-regulated capping of the barbed ends of actin filaments. J. Biol. Chem. 271, 7986-7991. Larsen, R. L., Katz, S., Roufogalis, B. D., and Brooks, D. E. (1981). Physiological shear stresses enhance the Ca 2§ permeability of human erythrocytes. Nature 294, 667-668. Lauf, P. K., Bauer, J., Adragna, N. C., Fujise, H., Zade-Oppen, A. M. M., Ryu, K. H., and Delpire, E. (1992). Erythrocyte K-C1 cotrans-
392
SECTION II Membrane Potential, Transport Physiology, Pumps, and Exchangers
port: properties and regulation. Am. J. Physiol. 263 (Cell Physiol. 32), C917-C932. Lauger, E (1995). Conformational transitions of ionic channels. In "Single Channel Recording," 2nd ed. (B. Sakmann and E. Neher, Eds.), ch. 22, pp. 651-662. Plenum, New York. LeFevre, E G. (1961). Sugar transport in the red blood cell: structureactivity relationships in substrates and antagonists. Pharm. Rev. 13, 39-70. Lew, V. L., and Bookchin, R. M. (1986). Volume, pH and ion content regulation in human red cells: analysis of transient behavior with an integrated model. J. Membr. Biol. 92, 57-74. Lew, V. L., Hockaday, A., Sipulveda, M.-I., Somlyo, A. P., Somlyo, A. V., Ortiz, O. E., and Bookchin, R. M. (1985). Compartmentalization of sickle-cell calcium in endocytic inside-out vesicles. Nature 315, 586-589. Lin, S., Yang, E., and Huestis, W. H. (1994). Relationship of phospholipid distribution to shape change in Ca2+-crenated and recovered human erythrocytes. Biochemistry 33, 7337-7344. Macey, R. I., Adorante, J. S., and Orme, F. W. (1978). Erythrocyte membrane potentials determined by hydrogen ion distribution. Biochim. Biophys. Acta 512, 284-295. Mercer, R. W., and Dunham. P. B. (1981). Membrane-bound ATP fuels the Na/K pump. J. Gen. Physiol. 78, 547-568. Milanick, M., and Hoffman, J. F. (1986). Ion transport and volume regulation in red blood cells. Ann. NY Acad. Sci. 488, 174-186. Murphy, E., Berkowitz, L. R., Orringer, E., Levy, L., Gabel, S. A., and London, R. E. (1987). Cytosolic free calcium levels in sickle red blood cells. Blood 69, 1469-1474. Ohnishi, S. T., and Ohnishi, T. (1994). "Membrane Abnormalities in Sickle Cell Disease and Other Red Blood Cell Disorders." CRC Press, Boca Raton, FL. Ohnishi, S. T., Katagi, H., and Katagi, C. (1989). Inhibition of the in vitro formation of dense cells and of irreversibly sickled cells by charybdotoxin, a specific inhibitor of calcium-activated potassium efflux. Biochem. Biophys. Acta 1010, 199-203. Passow, H. (1986). Molecular aspects of band 3 protein-mediated anion transport across the red blood cell membrane. Rev. Physiol. Biochem. Pharmacol. 103, 61-203. Passow, H., Wood, P. G., Lepke, S., MUller, H., and Sovak, M. (1992). Exploration of the functional significance of the stilbene disulfonate binding site in mouse band 3 by site-directed mutagenesis. Biophys. J. 62, 98-100. Payne, J. A., Lytle, C., and McManus, T. J. (1990). Foreign anion substitution for chloride in human red blood cells: effect on ionic and osmotic equilibria. Am. J. Physiol. 259 (Cell Physiol. 28), C819C827. Penniston, J. T., Vaughan, L., and Nakamura, M. (1979). Endocytosis in erythrocytes and ghosts: occurrence at 0 ~ after ATP preincubation. Arch. Biochem. Biophys. 198, 339-348. Ponder, E. (1948). "Hemolysis and Related Phenomena." Grune and Stratton, New York. Proverbio, E, Shoemaker, D. G., and Hoffman, J. F. (1988). Functional consequences of the membrane pool of ATP associated with the human red blood cell Na/K pump. In "The Na+,K§ Part A: Molecular Aspects" (J. C. Skou, J. G. Norby, A. B. Maunsbach, and M. Esmann, Eds.). Prog. Clin. Biol. Res. 268A, 561-567, Alan R. Liss, New York. Raess, B. U., and Tunnicliff, G. (Eds.). (1990). "The Red Cell Membrane: A Model for Solute Transport." Humana Press, Clifton, NJ. Raftos, J. E., Bulliman, B. T., and Kuchel, P. W. (1990). Evaluation of an electrochemical model of erythrocyte pH buffering using
31p nuclear magnetic resonance data. J. Gen. Physiol. 95, 1185-1204. Rapoport, T. A., Heinrich, R., Jacobasch, G., and Rapoport, S. (1974). A linear steady-state treatment of enzymatic chains. A mathematical model of glycolysis of human erythrocytes. Eur. J. Biochem. 42, 107-120. Reeves, R. B. (1977). The interaction of body temperature and acidbase balance in ectothermic vertebrates. Annu. Rev. Physiol. 39, 559-586. Reithmeier, R. A. E (1993). The erythrocyte anion transporter (band 3). Curr. Opin. Struct. Biol. 3, 515-523. Rodeau, J., and Malan, A. (1979). A two-compartment model of blood acid-base state at constant or variable temperature. Resp. Physiol. 36, 5-30. Ross, P. D., and Minton, A. P. (1977). Analysis of non-ideal behavior in concentrated hemoglobin solutions. J. Mol. Biol. 112, 437-452. Salhany, J. (Ed.). (1989). "Erythrocyte Band 3 Protein." CRC Press, Boca Raton, FL. Schwarz, W., and Passow, H. (1983). Ca2+-activated K + channels in erythrocytes and excitable cells. Annu. Rev. Physiol. 45, 359-374. Schwarz, W., Grygorczyk, R., and Hof, D. (1999). Recording singlechannel currents from human red cells. Methods Enzymol. 173, 112-121. Shohet, S. B. (1976). Mechanisms of red cell membrane lipid renewal. In "Membranes and Disease" (L. Bolis, J. E Hoffman, and A. Leaf, Eds.), pp. 61-74. Raven Press, New York. Shohet, S. B., and Mohandas, N. (Eds.). (1988). "Red Cell Membranes, Vol. 19, Methods in Hematology." Churchill Livingstone, New York. Steck, T. L. (1974). The organization of proteins in the human red blood cell membrane. A review. J. Cell Biol. 62, 1-19. Surgenor, D. MacN. (Ed.). (1974). "The Red Blood Cell," 2nd ed., Vols. I and II. Academic Press, New York. Tosteson, D. C., and Hoffman, J. E (1960). Regulation of cell volume by active cation transport in high and low potassium sheep red cells. J. Gen. Physiol. 44, 169-194. Tosteson, M. T., Halperin, J. A., Kishi, Y., and Tosteson, D. C. (1991). Palytoxin induces an increase in the cation conductance of red cells. J. Gen. Physiol. 98, 969-985. V~icha, J., and Znojil, V. (1981). The allometric dependence of the life span of erythrocytes on body weight in mammals. Comp. Biochem. Physiol. 69A, 357-362. Van Slyke, D. D., Wu, H., and McLean, E C. (1923). Studies of gas and electrolyte equilibria in the blood. V. Factors controlling the electrolyte and water distribution in the blood. J. Biol. Chem. 56, 765-849. Wang, D. N., Sarabia, V. E., Reithmeier, R. A. E, and Ktihlbrandt, W. (1994). Three-dimensional map of the dimeric membrane domain of the human erythrocyte anion exchanger, band 3. EMBO J. 13, 3230-3235. Whittam, R. (1964). "Transport and diffusion in red blood cells." Williams and Wilkins, Baltimore. Willis, J. S. (1992). Symposium on diversity of membrane cation transport in vertebrate red blood cells. An overview. Comp. Biochem. Physiol. 102A, 595-596. Yingst, D. R. (1988). Modulation of the Na,K-ATPase by Ca and intracellular proteins. Annu. Rev. Physiol. 50, 291-303. Yoshikawa, H., and Rapoport, S. M. (1974). "Cellular and Molecular Biology of Erythrocytes." University Park Press, Baltimore. Zade-Oppen, A. M., Adragna, N. C, and Tosteson, D. C. (1988). Effects of pH, potential, chloride and furosemide on passive Na § and K § effluxes from human red blood cells. J. Membr. Biol. 102, 217-225.
Nicholas Sperelakis
Cable Properties and Propagation of Action Potentials
I. I n t r o d u c t i o n The resting potential (RP) of cells enables the electrogenesis of action potentials (APs) and excitability. In this chapter, we examine the mechanism for propagation of the APs and excitability from one part of a neuron or muscle to a distal part. It is imperative that the body be able to transmit a signal from one point to another very rapidly. The only way that this can be accomplished is by an electrical mechanism. Blood flow and diffusion and signaling molecules are much too slow to allow rapid signaling. In contrast, electricity flows very quickly, at the speed of light (3 x 108 m/s) in a copper wire or about one-ninth the speed of light in a water solution (like the composition of the body). Therefore, the body makes use of electricity for rapid signaling in the nervous system, skeletal muscle, heart, and smooth muscles. Propagation velocity is about 120 m/s in our fastest nerve fibers, about 6 m/s skeletal muscle, about 0.5 m/s in heart, and about 0.05 m/s in smooth muscle (Table 1). One example of the need for very fast communication or signaling is the process of walking. Very rapid signals must travel from the motor cortex of the brain, down to the lower spinal cord region, and out the motor axons to the skeletal muscles of the lower extremities (Fig. 1). In this process, the signal crosses one or more synapses, which are regions in which one neuron ends and the next one begins, and in which a special chemical neurotransmitter signal is involved (Fig. 2). At the termination of each branch of a motor nerve axon on the skeletal muscle fiber, there is another synapse, known as the neuromuscular junction or motor end plate (see Fig. 1). The signal crosses the neuromuscular junction and gives rise to an AP in the muscle fiber that propagates in both directions from the motor end plate. The muscle AP elicits contraction. Receptors in the muscles (e.g., stretch receptors) transmit information (in the form of propagating APs) back into the central nervous system (CNS). Thus, in walking, there is a continuous rapid flow of information and instructions to the muscles in both directions: out of the CNS and into the CNS. Therefore, even a relatively simple Cell Physiology Sourcebook: A Molecular Approach, Third Edition
395
skeletal activity such as walking would not be possible without a very rapid signaling system. To illustrate, in various demyelinating diseases (e.g., caused by some viruses, heavy metals, autoimmune reactions), loss of the myelin sheath around the myelinated nerve fibers causes propagation to become slowed and impaired in the affected nerve fibers, with associated uncoordination and partial paralysis.
I!. F r e q u e n c y - M o d u l a t e d
Signals
Because propagating all-or-none APs are all very similar to each other (in shape, duration, amplitude, rate of rise, and propagation velocity), to make the signal stronger or weaker, the body increases or decreases, respectively, the frequency of the APs. That is, the body uses a frequency-modulated (FM) system, rather than an amplitude-modulated (AM) system (see
TABLE 1 Conduction Velocity as a Function of Fiber Diameter in Nerve Axons and Muscle Fibers ,,,
Fiber type Myelinated axons
Nonmyelinated axons
Fiber Propagation Velocity/ diameter v e l o c i t y diameter (lam) (m/s) (m/s/lam) 20
120
6.0
12
70
5.8
5
30
6.0
1.5
1.0 Squid giant axons (20 ~
500
2.0
1.3 25
1.3
1.3 0.05
Skeletal muscle fibers
50
6
0.12
Cardiac muscle fibers
15
0.5
0.03
Smooth muscle fibers
5
0.05
0.01
Copyright 9 2001 by Academic Press. All rights of reproduction in any form reserved.
396
SECTION III Membrane Excitability and Ion Channels
FIGURE 1. Schematic diagram of a motor axon, with the cell body (soma) in the anterior horn of the spinal cord and its terminal branches ending on skeletal muscle fibers (only one muscle fiber depicted) to form the neuromuscular junctions (motor end plates). Each motor axon with its attached skeletal muscle fibers is known as the motor unit. The motor axons are of large diameter (e.g., 20 pm) and are myelinated, and therefore propagate at fast velocities (e.g., 120 m/s). As a consequence of the chemical synaptic transmission process at the neuromuscular junction, an AP is initiated in the muscle fiber and propagates in both directions, bringing about contraction. Fig. 2). It is a digital system composed of on-off identical signals. At each synapse, the signal becomes graded in amplitude rather than all-or-none: the greater the amplitude and duration of the local postsynaptic potential, the higher the frequency of APs triggered. The same is true of the local graded receptor potential generated at some sensory organs/receptors. Stronger signals translate into a higher frequency of impulses, and weaker signals correspond to a lower frequency of impulses.
separated by rubber. Usually one of the conductors is arranged as a tubular sleeve surrounding a central solid rod (wire), as depicted in Fig. 3A. The equivalent electrical circuit for a cable is shown in Fig. 3B: two parallel conductors (wires) separated by a transverse resistance (R) shown distributed along the length of the cable. The resistance of the conductors is so small compared with the transverse insulation resistance that it is assumed to be zero. In the case of the biological cable (a long narrow nerve fiber or skeletal muscle fiber), one parallel conductor is the inside fluid (cytoplasm) and the second parallel conductor is the outside fluid surrounding (bathing) the cell (the interstitial fluid). Because the conductivity of biological fluid is much less (i.e., much higher resistance) than that of copper wire, and because the cross-sectional area of the cell is so small, the inside longitudinal resistance is high and cannot be ignored (Fig. 3C). The outside longitudinal resistance is relatively small, as compared with the inside, because of the larger volume (cross-sectional area) of fluid available to carry the outside current, and therefore is assumed to be negligible. In addition, there is a stray capacitance distributed along the length of the cable (Fig. 3D), because a capacitance occurs when two parallel conductors ("plates") are separated by a high-resistance dielectric material. The dielectric constant of materials is related to vacuum, which is assigned a value of 1.0000; air has a value very close to vacuum, oils have a value of 3-6, and that of pure water is 81. The biological membrane, which has a matrix of phospholipid molecules, has a dielectric constant of about 5, typical of oils. The higher the dielectric constant, the higher the capacitance; the closer the parallel plates, the higher the capacitance. Because the biological membrane is so thin (approx. 70 A or 7 nm), its capacitance is relatively high: all cell membranes have a membrane capacitance (Cm) of about 1.0 pF/cm 2 (where F stands for farads). Capacitors and the dielectric constant are discussed timber in Chapter 70.
I!I. C a b l e P r o p e r t i e s A. Biological Fiber as a Cable
B. Length Constant
An electrical cable consists of two parallel conductors separated by insulation material, for example, two copper wires
In the electric cable depicted in Fig. 3A and B, a voltage (or signal) applied at one end would be transmitted to a dis-
Synapse APs
~ ~ ' ~
Synaptic vesicles
All-or-none Digitalsignal Frequency-modulated APs
Excitatory postsynapticpotential (local, graded, nonrefractory)
FIGURE 2. Diagram of a chemical excitatory synapse between two nerve fibers. An AP in the presynaptic fiber (left) brings about the release of the neurotransmitter at its nerve terminal. The transmitter molecules rapidly diffuse the short distance (e.g., 0.1 pm) across the synaptic cleft, bind to receptor sites on the postsynaptic membrane, and open the associated nonselective ion channels (Na+,K+ mixed conductance) complexed to the receptor (i.e., these are ligand-gated ion channels). The associated synaptic current depolarizes the postsynaptic membrane, producing the excitatory postsynaptic potential (EPSP); this depolarization spreads passively into the adjacent conductile membrane (excitable) because of cable properties, thereby triggering one or more APs in the postsynaptic axon. The EPSPs are local, graded in amplitude, and nonrefractory whereas the APs are all-or-none (maximal), refractory, and propagated actively. Thus, the AM synaptic process gives rise to an FM or digital signal. The strength of a biological response (e.g., contraction, secretion, sensation) is a function of the frequency of the signal. That is, higher AP frequency corresponds to stronger contraction or stronger sensation (in the sensory nervous system).
397
24. Cable Properties and Propagation of Action Potentials
L-Vo Z
(2a)
e
=V
1 ~
=0.37V o The mathematical solution to Eq. 1 is
( x)
V-
antiln In Vo - -~
V-
antilog
(2b)
2.303 log Vo - (x/A) 2.303
(2c)
Hence, the distance at which the voltage decays to 37% of the initial value gives the length constant, A. In nerve fibers and skeletal muscle fibers, A has a value of only about 1-3 mm. Therefore, the relatively short A, c o m p a r e d with the length of the neuron (e.g., over 1.0 m for a lumbar anterior horn cell motor neuron) or skeletal muscle fiber, means that a signal applied at one end (or midpoint) would fall off very quickly with distance along the fiber (Fig. 5). If the length constant were 1.0 mm, then at 4 m m , the signal would bec o m e negligible. Hence, the electrical signal cannot be conducted passively in the biological cable, because it would decrement and disappear over relatively short distances. The AP (signal) is amplified to a constant value at each point (or each node) in the m e m b r a n e , as discussed in Chapter 25. That is, conduction is active, not passive, with
F I G U R E 3. A coaxial cable and associated electrical equivalent circuit (A, B), and the electrical equivalent circuit for a biological cable (C, D). (A) The coaxial cable consists of an inner conductor (e.g., copper wire) and an outer concentric conductor separated by a layer of insulation (e.g., rubber). (B) The equivalent circuit for a coaxial cable. The inner and outer conductors are depicted as having nearly zero resistances, and the transmembrane insulation resistance is shown distributed along the length of the cable. (C) In the biological cable, the inner conductor is the cytoplasm (axoplasm or myoplasm), which is not of negligible resistivity, and therefore is depicted as r i distributed along the length of the fiber. The transverse insulation resistance is the cell membrane resistance (rm). (D) Addition of the capacitance elements to the biological cable (Cm), which arise due to the lipid bilayer matrix of the cell membrane.
A
0 B
() 1oo 5O 3O 2O
rei V 10
tant end with little or no d e c r e m e n t (diminution or attenuation), and the so-called l e n g t h c o n s t a n t would be very long or nearly infinite. In the biological cable (Fig. 3C and D), however, a signal applied at one end rapidly falls off (decays) in amplitude as a function of distance, with a relatively short length constant (A). This decay in voltage is exponential (Fig. 4A). An exponential process gives a straight line on a semilogarithmic plot (log V versus distance) (Fig. 4B). In a cable, the relationship between the voltage at any distance (x) from the applied voltage (V o) is
VThus, when x = A
V e -x/A
(1)
(log scale)
5 3 2
10-
i
I
I
I
1
2
3
4
Distance (rnm) F I G U R E 4. Length constant (A) of the biological cable (fiber). (A) The exponential decay of voltage on both sides of an applied current (voltage) as a function of distance is diagrammed. The distance at which the voltage falls to I/e, or 36.7% of initial voltage (at x = 0), gives the h value. (B) When the voltage is plotted on a log scale against distance, a straight line is obtained.
398
SECTION III M e m b r a n e Excitability and Ion Channels
Io
,',j / ,
,~V 1
I I I I
va I
F I G U R E 5. Diagram of a motor axon and how the voltage signal would decay with distance if the neuron were only a passive cable (nonexcitable). The voltage traces illustrate the voltage signals that would be simultaneously recorded at three different points along the axon (V l, V2, and V3) from the site of injection of a rectangular current pulse (I o). As depicted, the amplitude of the steady-state voltage pulse rapidly falls off with distance, because the length constant (A) of the biological cable is short (e.g., 1 mm) compared with the length of the axon (e.g., 1000 mm). Therefore, an active response of the cell membrane at each point is required to faithfully propagate the signal. The voltage recorded at point V1 (AV1) also illustrates that the membrane potential changes in an exponential manner, both at the beginning of an applied rectangular current pulse and at the end. This exponential change results from the capacitance of the cell membrane, the time constant (7) being a product of the resistance and capacitance (Rmfm). The time it takes for the voltage to decay to 1/e (36.7%) of the initial (maximal) value at the end of the pulse, or to build up to 63.3% (1 - I/e) of the maximal value at the beginning of the pulse, gives the ~- value.
energy being put into the signal at each point to prevent any decay of the signal. The parameters that determine the length constant of a cable are the square root of the ratio of the transverse resistance (r m) to the sum of the inside (ri) and outside (r o) longitudinal resistances, A--~/
rm
(3)
r/+r ~"~"c m
cm-~
f~ +__~ cm
cm
where rm, ri, and r ~ are the resistances normalized for a unit length (1 cm) of fiber. For surface fibers of a nerve or a muscle bundle bathed in a large volume conductor, r o is negligibly small, and Eq. 3 reduces to
._/ft.
(4a)
A-
cm -
J
Rm a Ri 2
(4b)
~
O . c m 2.cm f~.cm
where a is the fiber radius, and R m and R i are the m e m b r a n e resistance and longitudinal cytoplasmic resistance, respectively, normalized for both length (1 cm) and cell diameter. Thus, the greater the m e m b r a n e resistance and the smaller the internal longitudinal resistance (larger cell diameter), the greater the A value. We will see later that the propagation velocity is a function of A; that is, larger diameter fibers propagate faster. We will also see that myelination increases the effective m e m b r a n e resistance (R m) and lowers the effective capacitance, thereby increasing propagation velocity. C. Time Constant Because of the large capacitance of the cell m e m b r a n e (Cm) the m e m b r a n e potential ( E m ) c a n n o t change instantaneously upon application of a step current pulse. Instead, E m
24. Cable Properties and Propagation of Action Potentials
399
changes in an exponential (negative) manner (see Fig. 5) on both the charge and the discharge. The membrane time constant (rm) is a product of the resistance (Rm) and capacitance of the membrane and can be expressed as Tm
(5a)
Frn C m =emC
If the microelectrode is near the middle of a very long fiber (or "infinite" cable), then Rin is related to the other cable parameters (e.g., to the internal longitudinal resistance r i and the DC length constant ADC)" Rin- 0.5 r. ADC
(5b)
m
~'~
-
~
where 1~ is the resistance in ohms, and F is the capacitance in farads. The discharge of the membrane (parallel R C network) is given by
Vt -
Wmaxe-t~7
(6)
where V t is the voltage at any time t (at the site of current injection), and Vmax is the final maximum voltage attained during the pulse. When t = ~-, Vt -
1 Wmax e
=V
(7a)
1 max 2.718
=0.37Vma x Hence, the time at which the voltage decays to 37% of the initial (maximal) value gives the time constant, ~" (see Fig. 5). In nerve fibers and skeletal muscle fibers, Tm has a value of about 1.0 ms. When the membrane is charging, there is a similar exponential (negative) process, with the identical time constant (see Fig. 5). The corresponding relationship is given by
c
cm
s-l~.F
(9)
m
The factor of 0.5 is present because current flows in both directions from the microelectrode. If the microelectrode is at one end of the long fiber, then the factor of 0.5 is removed. Equation 9 indicates that the Rin is greater when r i and/or ADC are greater. Combining Eqs. 8 and 9 gives
Vo
I-- = 0.5 r ADC o
(10a)
V - I ~ 0.5 r/)tDC
(10b)
or
This equation indicates that the change in membrane potential at x - 0 is equal to the applied current times the input resistance. To obtain the change in membrane potential at any point x along the fiber cable (Vx), then an exponential term (e -x/A) must be added to account for the exponential decay of voltage over distance" V - I o 0.5 r. ADC e-X/a
( 11 )
This equation is identical to Eq. 1, because the loO.5riADc term is equal to V o (Eq. 10b)" V - V ~ e -x/A
Because ADC = V/ r / r .
(1)
(Eq. 4a), Eq. 9 can be given as /
V t - Vmax(1 - e -'/~)
(7b)
The time it takes for the voltage to build up to 63% of its final value gives the time constant. When t = ~-, gt- ( 1 -1)
max
(7c)
=(1 - 0.37)Vma x = 0.63Vma x Thus, the time constant can be measured on the build-up of the pulse (time to reach 63% of the final voltage) or on the decay (time to reach 37% of the initial voltage).
D. Input Resistance The input resistance (Rin) of a muscle or nerve fiber is essentially the resistance that an intracellular microelectrode "looks" into when a small current (DC pulse) is injected. Thus, Rin is determined by the change in membrane potential ( V or Vx=0) at steady state produced at the site of injection (x = 0) by the injection of a known (measured) amount of current ( / o r / x = 0 ) , based on Ohm's law.
Vo Bin
io
(8)
R i n - 0 . 5 F i 1/~Im. --0.5 /FmF /
(12)
and Eq. 11 can also be given as V-
I ~ 0.5 rf~mr/ e -x/~
(13)
Thus, input resistance is proportional to v / r 5" This is logical because the higher the longitudinal resistance of the fiber's axoplasm (or myoplasm), which is a function of the resistivity of the axoplasm and fiber diameter (see Section XI.A-C of Chapter 70), and the higher the membrane resistance, the higher the input resistance should be. The input resistance can be calculated from the electrical equivalent circuit for a cable. The input impedance can also be given in terms of R m and Ri, the specific resistance of the cell membrane and resistivity of the axoplasm, respectively (see Section XI.A, B of Chapter 70). Since R m = 27tar m and R i = 7ra2ri , then Eq. 12 can be converted to Rin- 0.5
--0.5
IR
R m
i
27ra 7ra 2
(14a)
IRR
m i 2,n-2a 3
(14b)
Thus, the input resistance is directly proportional to the square root of the membrane-specific resistance and the re-
SECTIONIII MembraneExcitabilityand Ion Channels
400
sistivity of the axoplasm, and inversely proportional to the square root of the fiber radius raised to the third power. Calculation of the input impedance (Zin) is more complicated because it also depends on the membrane capacitance (C m) and on the frequency ( f ) of the alternating current (AC) used. According to Katz (1966), the input impedance of a long fiber (current injected at one end) is given by Z m
R Ri m _ ~ 2"n'2a 3 /1 + 47r2f2R2 C 2 V
m
A ~
()' Monophasic 9
m
cm 3 / ( s _ 2 ) ( s 2)
,c
Diphas--~. . . . ~
:3 External
-4Nw-
recording
'~
i
External
/'[~
.. I d2V V_ . _ / _ A_~_ _ 't d t 2
triphasic ~
recording
~,.) and
I dV ~ L
(15)
l (I~-cm 2 )(I~.cm) ~-
I~
Intracellular
recording
l rn
t i-
()
/Y~=-- ....
A discussion of impedance is given in Section VIII of Chapter 70 and a discussion of the AC length constant (AAC) is given in Appendix 2 to this chapter. E. Local Potentials In contrast to the active propagation of APs, synaptic potentials and sensory receptor potentials are not actively propagated. Such potentials decay exponentially (from their source of initiation) along the fiber cable, as described previously. Therefore, postsynaptic potentials and receptor potentials are local potentials. When local potentials are in the depolarizing direction, they can give rise to APs, which are propagated; when hyperpolarizing, they act to inhibit production of APs. These local potentials are similar to the local excitatory response (see Chapter 25), in that both are confined to a local region; however, the electrogenesis of the two is different. As stated before, the neuromuscular junction is an excitatory type of chemical synapse and produces excitatory postsynaptic potentials (EPSPs), known here as end-plate potentials (EPPs). Most synaptic potentials are graded, that is, they can add on to one another, both in time and in space (temporal summation and spatial summation), to produce larger responses. Larger synaptic potentials exert a greater stimulatory or inhibitory effect on the production of APs.
0-
_
-()
-
,,
Monophasic AV / m Triphasic Diphasic /eand I c I e Diphasic Triphasic Im FIGURE 6. (A) Schematic representation of the first (I)') and second (V) time derivatives of the AP spike and the longitudinal and radial currents associated with the propagating spike. The first derivative (dV/dt or V) is proportional to the capacitive current (I c = C m dV/dt) and to the longitudinal (axial) current and is diphasic, having an intense forward phase and a less intense backward phase, as depicted in the diagram at the lower right. The second derivative (d2V/dt 2 or V) is proportional to the radial transmembrane current (/), and is triphasic, having a moderately intense initial outward phase, then a very intense inward phase (the "current sink"), followed by a least intense second outward phase, as depicted in the lower diagrams. (B) The arrows at the lower left also depict the three phases of the membrane current and the two phases of the axial current, dV/dt can be recorded externally by a pair of closely spaced (relative to the spike wavelength) electrodes arranged parallel to the fiber axis, as illustrated, d2V/dt 2 can be recorded by a pair of electrodes arranged perpendicular to the fiber axis, as depicted. The vertical dashed line indicates that when the slope of the spike goes to zero at the peak of the spike, dV/dt is zero; dV/dt is maximum at about the middle of the rising phase of the spike.
IV. C o n d u c t i o n o f A c t i o n P o t e n t i a l s A. Local-Circuit Currents The generation of APs is described in Chapter 25. This section examines the mechanism for their rapid propagation (conduction). Propagation occurs by means of the local-circuit currents that accompany the propagating APs, as depicted in Fig. 6. Such currents exist because, when two points are at a different potential (voltage) in a conducting medium, current (I) will flow between the two points, as governed by O h m ' s law (I = V/R). At the peak of the AP in one region of the fiber, the inside of the membrane at that region becomes positive with respect to the outside. The inside is also positive with respect to the inside cytoplasm at a region downstream from the active region. Therefore, current flows through the cytoplasm from the active region (current source) to the adjacent inac-
tive region, then out of the fiber across the cell membrane, then through the interstitial fluid back to the active region (current "sink"), and finally through the membrane of the active region. This completes the closed loop for the current. The outward current through the membrane of the inactive region produces an I R voltage drop (Ohm's law), positive inside to negative outside. This acts to depolarize this region, because the polarity of the voltage drop is opposite to that of the resting potential (negative inside, positive outside). When the depolarization exceeds the threshold potential, an AP is triggered. Thus, the inactive region becomes converted to an active region. This process is repeated in each segment of fiber, thus resulting in movement (propagation) of the impulse sequentially down the fiber. If we examine a propagating AP in the middle region of a fiber (Fig. 6), we see that there is also a small backflow of cur-
24. Cable Properties and Propagation of Action Potentials
401
rent internally, coupled with a corresponding small forward flow externally, associated with the repolarizing phase of the AP. Thus, as the AP propagates down the fiber, from right to left, there is a simultaneous double flow of local-circuit current: clockwise flow associated with the rising phase of the AP, and counterclockwise flow associated with the repolarizing phase of the AP. The internal longitudinal current, sweeping past a transverse plane of the fiber, has two phases, first forward (right to left), and then reverse (left to fight). The external longitudinal current also has two phases: first left to right and then fight to left. The transverse membrane current, flowing outward in a plane perpendicular to the membrane, has three consecutive phases: first outward (still passive membrane), then inward (active membrane), and finally outward again (still active membrane). It is the local-circuit current flow that enables the electrocardiogram (ECG), electromyogram (EMG), electroencephalogram (EEG), and electroretinogram (ERG) to be recorded from the body surface over the tissue of interest (heart, skeletal muscle, brain, and eye). The internal longitudinal current is confined to the cytoplasm of the fiber, but the external current can use whatever conducting fluid is available (e.g., the entire torso volume conductor) because of the principle that parallel resistors give a lower total resistance (or current takes the path of least resistance). Thus, this external local-circuit current causes the skin to be at different potentials, and these differences can be recorded (as the ECG, etc.).
B. Propagation Velocity Determinants The factors that determine active velocity of propagation (0) include (1) fiber diameter, (2) length constant (A), (3) time constant (Tm), (4) local-circuit current intensity, (5) threshold potential, and (6) temperature. Some of these factors are interrelated, such as fiber diameter and length constant (since A is proportional to the square root of the radius), and length constant and time constant (since both have a dependence o n Rm). Propagation velocity is directly proportional to length constant and inversely proportional to time constant as
0a: A
(16a)
7" m
By substituting Eq. 4a for A and Eq. 5a for ~'m' we have
1 /
O cx: -~m
a
Rm Ri
(16b)
Thus, propagation velocity is directly proportional to the square root of fiber diameter or radius (a) and inversely proportional to membrane capacitance (Cm). The larger the fiber diameter, the lower the absolute longitudinal resistance of the intracellular cytoplasm (principle of resistors in parallel), and therefore the greater the amount of local-circuit current flowing longitudinally and the greater the length constant. For example, it is well known that the larger the diameter of nerve fibers, the faster they propagate (Table 1).
Equation 16b shows that if C m c a n be reduced (by myelination), then 0 should increase in proportion. This is discussed in Section IV.C. In addition, 0 depends on the intensity of the local-circuit current, and hence on the rate of rise of the AP. The greater the AP rate of rise (max d V / d t ) , the greater the longitudinal current and the transmembrane capacitive current (Ic). Therefore, all other factors being constant, faster rising APs propagate faster. The AP rate of rise depends on the density of the fast Na § channels that carry inward current, on C m, and on temperature. Max d V / d t decreases with increased C m, with cooling, and with partial depolarization (due to the h~ versus E m relationship discussed in Chapter 25). Cooling slows the rate of all chemical reactions, particularly those with a high Q10 (activation energy), such as the ion conductance changes in activated membrane. Finally, the threshold potential (Vth) affects propagation velocity. If the threshold were to be shifted to a more positive voltage (more depolarized), then it would take longer for a given point in the membrane to reach threshold (and explode) during propagation of an AP from upstream. A greater critical depolarization (difference between RP and Vth) would be required to bring the membrane to threshold. Therefore, propagation velocity would be slowed. As stated earlier, some of these factors are interrelated, and some actually exert opposing effects. The foregoing discussion applies to nonmyelinated nerve axons and skeletal muscle fibers. An electron micrograph of small bundles of nonmyelinated nerve fibers enveloped by Schwann cells is shown in Fig. 7. In myelinated nerve fibers, propagation velocity is greatly increased by the myelin sheath, as discussed in Section IV.C.
C. Saltatory Conduction The nerve cable has been vastly improved by the evolutionary development of myelination in vertebrates. An electron micrograph of a myelinated nerve fiber is shown in Fig. 8. The myelin sheath improves the cable by increasing the effective R m by about one-hundred-fold and decreasing the effective C m by about one-hundred-fold. This increases the length constant, h, and tends to decrease the time constant r m. Although z m tends to increase with the increase in effective R m (due to the myelin sheath), the decrease in C m counteracts this effect, thus acting to hold ~'malmost constant. Thus, Eqs. 16a and 16b predict that propagation velocity should increase with myelination. One consequence of myelination, therefore, is that propagation velocity is greatly increased (Table 1). Another consequence is that the energy cost of signaling is greatly decreased, because passive ion leaks are limited and active current losses are restricted to the small nodes of Ranvier, which are spaced relatively far apart. An electron micrograph of a node of Ranvier is shown in Fig. 9. At each node, the length of exposed (naked) cell membrane is only a few micrometers. The intemodal distance is about 0.5-2.0 mm (depending on fiber diameter), and the width (length) of each node is only about 0.5-3 ~am. The node forms an annulus around the entire perimeter of the fiber. Therefore the degree of energy-requiring active ion transport (Na§ § and Ca2§ required to maintain the steady-state ion
402
SECTION III Membrane Excitability and Ion Channels
F I G U R E 7. Electron micrograph of an autonomic nerve of mouse heart cut in transverse section. The unmyelinated axons (Ax) are arranged in small groups (bundles) that are engulfed by the cytoplasm of a Schwann cell. That is, the nonmyelinated axons are embedded in, and surrounded by, a Schwann cell. Two Schwann cell nuclei (Sch) can be seen. The axon bundles are separated and surrounded by a collagenous matrix. The entire nerve is surrounded by a perineurial sheath (PN). The scale bar at lower right equals 2 Ixm. (Micrograph courtesy of Dr. Mike Forbes, University of Virginia.)
distributions and to hold the system in a state of high potential energy is greatly reduced. For example, the amount of Na § gained and K + lost per impulse is reduced as a result of myelination. The rate of oxidative metabolism in myelinated fibers reflects this lowered energy requirement. The myelin sheath is produced by the wrapping of the Schwann cell repeatedly around the nerve fiber in a spiral, forming 20-200 wrappings, depending on axon diameter. That is, the larger axons have a thicker myelin sheath. For the purpose of our discussion, we will assume an average of 100 wrappings. The myelin sheath covers the nerve axon as a coat sleeve and is interrupted at each node. The cytoplasm of the Schwann cell in the region of the myelin sheath is nearly completely extruded during its formation, so that the sheath consists essentially of 100 cell membranes stacked in series. Because resistors in series are added to calculate the total resistance, the effective transmembrane resistance is increased one-hundred-fold, but because the total capacitance of capacitors in series is calculated like resistors in parallel, the effective capacitance is reduced one-hundred-fold. Since A is directly proportional toy/Rm, A is increased accordingly. As described in the section on the factors influencing propagation velocity, increasing A and lowering C m increase O. Myelinated nerves usually have an optimal amount of myelin, which is an amount such that the ratio of diameter
of axon cylinder (naked axon) to total fiber (including myelin sheath) is about 0.64).8. Assuming a maximal total diameter feasible (the body must pack many circuits within a limited space, e.g., sciatic nerve bundle), a greater fraction of myelin, by infringing on the diameter of the axis cylinder, would raise r i too high, causing 0 to decrease. Thus, there are two opposing factors in determining the degree of optimal thickness of the myelin sheath: the more the myelin, the greater the decrease in C m and the increase in effective R m, but the smaller the diameter of the axis cylinder and therefore the higher the r i. In saltatory conduction (Latin s a l t a r e , to jump), the impulse jumps from one node to the next. The internodal membrane does not fire an AP. This is due to two reasons: (1) the internodal membrane is much less excitable (e.g., much fewer fast Na § channels) and (2) the depolarization of the neuron cell membrane at the internodal region is only about 1/100th of that at the node. The latter occurs because the I R voltage drop across the internodal cell membrane is only 1/100th of that across the entire series resistance network (neuron cell membrane plus 100 layers of Schwann cell membrane) (Kirchhoff's laws dealing with voltage drops across resistors in series). Even though the internal potential in the internodal region swings positive (e.g., to +30 mV) when the adjacent nodes fire, the potential at the
403
24. Cable Properties and Propagation of Action Potentials
outer surface of the internodal membrane also swings nearly as positive (e.g., to +29 mV). Therefore, the depolarization of the neuronal m e m b r a n e at the internode is only about 1 mV, which is well below threshold, and so the internodal m e m b r a n e does not fire. The potential that controls the m e m b r a n e conductances (activates the voltage-dependent ion channels) is the pd (potential difference) directly across the m e m b r a n e and not the absolute potential on either side. As stated previously, the internodal m e m b r a n e has only a few fast Na § channels, whereas the nodal m e m b r a n e has a very high density. Since the cell m e m b r a n e is fluid and proteins can diffuse (float) laterally in the lipid bilayer matrix, what keeps the fast Na § channel proteins confined (at high density) in the nodal region? It appears that there are special anchoring proteins (e.g., ankyrin) that anchor the ion channel proteins to the cytoskeletal framework, thus preventing their lateral m o v e m e n t into the internodal m e m brane. The effect of myelin and saltatory propagation is to make propagation m u c h faster. For example, a 20-pm-diameter myelinated nerve fiber conducts even faster than a 1000-pm (1-mm)-diameter nonmyelinated nerve fiber (e.g., the giant axon in squid, lobster, earthworm): 120 m/s versus 2 5 - 5 0 m/s. Thus, invertebrates, to achieve fast conduction in some essential circuits, must resort to giant neurons, resulting in a lower r i and hence fast conduction. Because of space/size limitations, only a few critical neurons can be made giant in diameter. In vertebrates, on the other hand, a large fraction of the nerve fibers in the peripheral nerves is myelinated for fast propagation. We saw earlier that, in nonmyelinated axons and skeletal muscle fibers, 0 should vary with the square root of the cell diameter or radius (a~ In myelinated axons, 0 varies with the first power of the cell radius (a 1) because 0 varies with A2 as indicated by F I G U R E 8. Myelin sheath of myelinated nerve fibers. (A) Myelinated nerve fiber in spinal cord of rat cut in cross-section. There are about seven wrappings of the Schwann cell around the axon, thus giving about 14 membranes in series with the cell membrane of the axon (axolemma). The internal mesaxon is visible at the left side (lower) of the axon, and represents the beginning of the spiraling of the Schwann cell membranes (cytoplasm squeezed out) to form the myelin sheath surrounding the axon. On the upper left side, a process of an oligodendrocyte abuts on the myelin sheath. The axon shown is closely plastered up against the myelin sheath of a neighboring axon (bottom of figure). Microtubules and neurofilaments are visible in the cytoplasm of the axon. Magnification of 168000x. (From J. Rhodin, "An Atlas of Histology," Oxford University Press, New York, 1975.) (B) Higher magnification of two adjacent nerves fibers (in phrenic nerve of rat) cut in longitudinal section to illustrate an axon with a thicker myelin sheath (approx. 23 wrappings = 46 membranes). Note that between the thicker (dense) lines are two thinner lines, which can be seen more clearly in C. The extracellular space (ECS) between the two axons is labeled. The calibration bar at the lower fight equals 0.1 I~m. (C) Still higher magnification of a portion of one of the myelin sheaths shown in the middle panel to better illustrate the periodicity of the myelin membranes. The dense lines alternate with a pair of thin lines. The space in between the pair of thin lines is extracellular space. The dense lines represent the intracellular (cytoplasmic) region of the Schwann cell process, with the cytoplasm extruded, thus allowing the inner (cytoplasmic face) leaflets of the cell membrane of the Schwann cell to come into close contact. The outer
~2 a O a : ~ - ~ Tm 2 R ' C m
(17)
The dependence of conduction velocity on diameter of myelinated and nonmyelinated fibers is summarized in Table 1.
D. Wavelength of the Impulse We can calculate the wavelength of the AP, which is the length of the axon simultaneously undergoing some portion of the AP. The wavelength is equal to propagation velocity
(extracellular face) leaflets of the Schwann cell membrane are seen as the thin lines. Thus, each cell membrane has the appearance of a double line, representing the hydrophilic surfaces of the cell membrane; the hydrophobic region of the membrane is the clear region between the dense line and the contiguous thin line on either side. UM = unit cell membrane. The scale bar on the lower fight represents 0.1 Ixm. (Micrographs B and C courtesy of Dr. Mike Forbes, University of Virginia.)
404
SECTION III Membrane Excitability and Ion Channels
FIGURE 9. Electron micrograph of one node of Ranvier in a single myelinated nerve axon of rat sciatic nerve cut in longitudinal section. As can be seen, the dark myelin bordering the axon on each side is interrupted near the middle of the micrograph, leaving the neuronal cell membrane nude of myelin at the node region. However, cytoplasmic processes of the Schwann cells cover the nodal membrane. Collagenous fibrils of the endoneurium are visible at the region of the node, peripheral to the Schwann cell cytoplasm. The myelin tapers and thins as it approaches the node, and there is a frayed appearance caused by the successive laminae of the myelin sheath terminating as cytoplasmic swellings. The axoplasm contains neurofilaments and microtubules. Magnification of 9000x. (From J. Rhodin, "An Atlas of Histology," Oxford University Press, New York, 1975.)
(0) times the duration of the
AP (mPD100); thus,
wavelength = 0 x APD100
(18)
cm = (cm/s) x s distance = velocity• time Note the similarity of this relationship to that for the wavelength of an electromagnetic radiation: wavelength = velocity of light/frequency of the radiation. The reciprocal of frequency is the period (duration of one cycle). The wavelength in a large myelinated nerve axon is about 12 cm: 120 m/s x 1.0 ms. In a skeletal muscle fiber, it is about 1.8 cm (6 m/s • 3.0 ms). In a smooth muscle bundle, the wavelength is only about 1.5 mm (5 cm/s x 30 ms).
V. External Recording of Action Potentials A. Monophasic, Diphasic, and Triphasic Recording As discussed previously, local-circuit currents accompany the propagating AP in each fiber. The intracellular and extracellular longitudinal currents are diphasic; that is, initially they travel in the forward direction intracellularly and then in the reverse direction. The forward direction current is intense (high current density) and the reverse direction current is weak (low current density). The transmembrane radial currents are triphasic; that is, the first phase is outward (moderate intensity), the second phase is inward (high intensity), and the third phase is outward (low intensity). The first phase (outward) gives rise to the passive exponential foot of the AP, and is due to the passive cable spread of voltage and current. The second phase (inward) corresponds to the large inward fast Na + current, which occurs during the later portion of the rising phase and peak of the AP. The third phase (outward) corresponds to the net outward current (K+), which occurs during the repolarizing phase.
These longitudinal and radial currents can be recorded by suitably placed external electrodes. The extracellular longitudinal currents can be recorded by two electrodes (bipolar) placed close together along the length of the fiber. If the interelectrode distance is short (relative to the wavelength), an approximate first (time) derivative of the AP is obtained (see Figs. 6 and 10C). The extracellular radial currents can be recorded by two electrodes placed close together in a plane perpendicular to the fiber axis. This gives an approximation of the second (time) derivative of the AP (see Fig. 6). The internal axial currents are confined to the cytoplasm, whereas the external longitudinal currents can use the entire interstitial fluid space of the nerve bundle or muscle or even the entire torso (so-called volume conductor), since current takes the path of least resistance (resistors in parallel). As mentioned in Section IV.A, this allows the recording of the ECG from the body surface and the EMG from the skin overlaying an activated skeletal muscle. The ECG and EMG consist essentially of diphasic potentials, reflecting the external longitudinal currents during propagation of APs. When the two external electrodes are placed far apart (with respect to the wavelength) along a nerve or muscle fiber, the diphasic recording has two phases that are about equal (Fig. 10A). The proximal electrode records the wave of negativity (associated with the propagating AP) first, and then returns to isopotential. When the wave reaches the second electrode, the wave is recorded by it in reversed polarity (because current flow through the voltmeter is reversed). If the AP were now prevented from reaching the second (distal) electrode by crushing this region of the fiber or elevating [K+]o to depolarize it, then a monophasic recording would be obtained (Fig. 10B). This monophasic recording would most resemble the true AP recorded by a microelectrode impaled into a fiber to record the transmembrane potential, but would be much smaller in amplitude.
405
24. Cable Properties and Propagation of Action Potentials
trode is placed on the skin of the ventral forearm and the other (reference) electrode on the wrist of the same arm. Then, as the subject voluntarily produces stronger and stronger contraction to flex the hand, the electrical signals picked up become greater and greater in amplitude and frequency. The amplitude becomes larger because more muscle fibers are simultaneously activated. This is known as fiber recruitment. The frequency also increases, because the motor nerves fire at a higher frequency, causing the muscle fibers to fire at a higher frequency and thus producing a more powerful tetanic contraction.
VI. Summary
FIGURE 10. Diagram of the waveforms that would be recorded externally during propagation of an AP in a single fiber. (A) When the two electrodes are far apart (relative to the AP wavelength), a diphasic recording is obtained, with the two phases being symmetrical and separated by an isopotential segment. The two phases are due to the current flow through the voltmeter recorder being first in one direction and then in the opposite direction. (B) If the fiber between these two electrodes is damaged (e.g., by crushing) or depolarized (by elevated [K+]o), so the AP cannot sweep past the second electrode (2), then this second phase is prevented and the recording is rnonophasic. (C) If the two electrodes depicted in part A are brought progressively closer, then the isopotential segment would shorten and disappear. If electrode 2 is brought very close to electrode 1, so that the interelectrode distance is short relative to the wavelength, then the second phase is smaller than the first phase, and the record resembles the first derivative of the true AP.
B. Compound Action Potential When one records the APs externally, the records are graded and not all-or-none, as in the case of the true APs recorded intracellularly from single fibers (see Chapter 25). That is, the signal recorded becomes larger and larger, up to a maximum amplitude as the intensity of stimulation is increased. This is the so-called compound action potential. It is graded because, as a greater and greater fraction of the fibers is activated, the external longitudinal currents associated with the all-or-none AP in each fiber cut across the recording electrodes, thereby producing a larger signal. The AP in each individual fiber is always all-or-none. The amplitude of the signal is determined by the resistance between the electrodes multiplied by the amount of current flowing through this resistance (V --- IR). The recording of compound action potentials is diagrammed in Fig. 11. The compound action potentials can be demonstrated by recording the E M G from a human subject when one elec-
Although the biological cable (i.e., nerve fiber or skeletal muscle fiber) is the best possible, it is a relatively poor cable with a short length constant and relatively long time constant. Therefore, for faithful and rapid signal transmission over long distances, energy must be put into the system at each point along the way. The system evolved is that of AP generation, which are all-or-none signals of constant amplitude and constant propagation velocity, in addition to having refractory periods and sharp thresholds. This is a frequencymodulated system, in which increasing strength of sensation or motor response follows from an increase in frequency of the AP signals. AP propagation occurs by means of local-circuit currents. The transmembrane current has three phases: outward, inward, and outward. The internal and external longitudinal currents have two phases: forward and backward (for internal) or backward and forward (for external). The external currents use the path of least resistance, enabling electrograms (e.g., ECG, EMG) to be recorded from the body surface. The compound AP is graded in amplitude, reflecting the summation of the external currents generated from each fiber that is activated; that is, the more fibers simultaneously activated, the greater the amplitude of the electrogram signal. Propagation velocity is faster the larger the diameter of the fiber, the longer the length constant, and the lower its time constant and capacitance. The myelin sheath evolved by vertebrates enables much faster propagation velocity and at a lower energy cost. Myelination raises the effective membrane resistance and lowers the effective capacitance, and excitability occurs only at the short nodes of Ranvier that periodically interrupt the myelin sheath. Therefore, the AP signal jumps from node to node in a saltatory pattern of conduction.
Bibliography Cole, K. S. (1968). "Membranes, Ions and Impulses: A Chapter of Classical Biophysics." University of California Press, Berkeley. Davis, L., Jr., and Lorente de No, R. (1947). Contribution to the mathematical theory of the electrotonus. Stud. Rockefeller Inst. Med. Res. 131, 442-496. Hodgkin, A. L., and Rushton, W. A. H. (1946). The electrical constants of a crustacean nerve. Proc. R. Soc. Lond. (Biol.) 133, 444--479. Jack, J. J. B., Noble, D., and Tsien, R. W. (1975). "Electric Current Flow in Excitable Cells." Clarendon Press, Oxford.
406
SECTION III Membrane Excitability and Ion Channels
F I G U R E 1 1. Diagram of a compound action potential in an isolated nerve trunk, such as the frog sciatic nerve, recorded externally by a pair of longitudinal electrodes. The voltmeter (oscilloscope) records the voltage (IR) drop across the resistance (fluid) between the two electrodes. If only fiber 1 is activated, the current passing between the electrodes is small, and the voltage recorded is small. If fiber 2 is simultaneously activated with fiber 1, then the amount of current is doubled, and the voltage is doubled. When all three fibers are simultaneously activated, the current is tripled, and the voltage is tripled. Therefore, the externally recorded compound action potential is graded because it reflects the electrical activity of numerous fibers, each of which produces an all-or-none (nongraded) AP.
Katz, B. (1966). "Nerve, Muscle, and Synapse." McGraw-Hill, New York. Lakshminarayanaiah, N. (1984). "Equations of Membrane Biophysics." Academic Press, Orlando, FL. Rail, W. (1977). In "Handbook of Physiology" (J. M. Brookhart and V. B. Mounteastle, Eds.), Vol. 1, pp. 39-97. American Physiological Society, Bethesda, MD. Sperelakis, N. (1979). Origin of the cardiac resting potential. In "Handbook of Physiology" (R. M. Berne and N. Sperelakis, Eds.), Vol. 1, pp. 187-267. American Physiological Society, Bethesda, MD. Sperelakis, N. (1992). Cable properties and propagation mechanisms. In "Physiology" (N. Sperelakis and R. O. Banks, Eds.), pp. 83-97. Little, Brown, Boston.
Sperelakis, N., and Fabiato, A. (1985). Electrophysiology and excitation-contraction coupling in skeletal muscle. In "The Thorax: Vital Pump" (C. Roussous and P. Macklem, Eds.), pp. 45-113. Marcel Dekker, New York. Sperelakis, N., and Mann, J. E., Jr. (1977). Evaluation of electric field changes in the cleft between excitable cells. J. Theor. Biol. 64, 71-96. Taylor, R. E. (1963). Cable theory. In "Physical Techniques in Biological Research" (W. L. Nastuk, Ed.), Vol. 6, pp. 219-262. Academic Press, New York.
Appendix 1: Propagation in Cardiac Muscle and Smooth Muscles
1. Background Chapter 24 discusses propagation in cells that are long cables, such as nerve fibers and skeletal muscle fibers. Propagation is more complex in tissues composed of assemblies of short cells, such as cardiac muscle and visceral smooth muscles. These short cells may be considered to be short or truncated cables. In such a truncated cable cell, its true length constant h is much longer than the cell's length. Therefore, there is relatively little voltage fall-off (decay) over the length of each cell. It follows that the entire cell undergoes an action potential (AP) nearly simultaneously. Yet propagation velocity 0 is much slower in cardiac muscle (ca. 0.5 m/s) and smooth muscles (ca. 0.05 m/s) than in skeletal muscle (ca. 5 m/s) or nonmyelinated nerve fibers (ca. 2 m/s). Part of the reason for the slower 0 concerns fiber diameter (i.e., diameter of cardiac muscle fiber is about 15 Ixm compared to about 60 txm for skeletal muscle fibers). Equation 16b in Chapter 24 indicates that in a simple cable, 0 is a function of v/-d (where a is the fiber radius). Thus, the theoretical 0 for cardiac muscle (Oc) should be 0
-
c
f 15txm 60 txm
•
0sk
= 0.5 X 5 m/s = 2.5 m/s where 0sk is the propagation velocity for skeletal muscle, and it is assumed that the other parameters given in Eq. 16b in Chapter 24 are the same in the two types of muscle. This theoretical velocity of 2.5 m/s is substantially greater than the actual 0.5 m/s. Therefore, at least one other factor must determine velocity in cardiac muscle, and that is the high resistance of the junctional membranes (the intercalated disc membranes in the case of cardiac muscle). It is still controversial as to exactly how high the junctional resistance is. When gap junctions are present, the gap junction channels span the intercellular junction, and so serve to lower the cell-to-cell resistance (see Chapter 31). However, the hearts of the lower vertebrates (e.g., amphibians, reptiles) either have no gap junctions or they are very sparse and tiny. In addition, even parts of mammalian hearts and some visceral smooth muscles (e.g., longitudinal muscle layer of intestine) do not appear to contain gap junctions. If so, another mechanism may be involved for cell-to-cell
407
propagation in those muscles in which there is a virtual absence of gap junctions. One of the mechanisms proposed is known as the electric field model (Sperelakis et al., 1989). This model is discussed in this appendix. First, however, let us return to the problem that propagation in cardiac muscle is about five times slower than it should be if the gap junction channels served to interconnect thoroughly (electrically) the cytoplasm of two contiguous cells lying end to end. We also stated previously that the short cardiac muscle cell should show no voltage decay, and so the entire length of the cell should undergo the AP nearly simultaneously (i.e., there is nearly infinite propagation velocity within each cell). Yet the overall propagation velocity in the tissue is relatively slow. Clearly, a large time delay must occur at each cell junction. In fact, most of the propagation time is consumed at the cell junctions. This has been demonstrated experimentally. Propagation in cardiac muscle has been shown to be actually discontinuous or saltatory in nature (Spach et al., 1981; Rudy and Quan, 1987a, b; Sperelakis et al., 1989). Therefore, the presence of a large number of gap junction channels is insufficient to reduce the junctional resistance enough to allow a chain of cells to behave as a simple cable. This relatively high junctional resistance has relevance to heart block and fibrillation, but this is outside the scope of this chapter. A similar analysis can be done for visceral smooth muscle. The theoretical 0 for visceral smooth muscle (0 s) should be 0
_ / 5~zm ~ 60 Ixm • 0~k = 0.289 • 5 m/s = 1.44 m/s
This theoretical velocity of 1.44 m/s is much greater than the actual velocity of 0.05 m/s. Thus, propagation velocity in smooth muscle is only about 3.5% of what it should be based on fiber diameter. The lower maximum rate of rise of the AP (+ max dV/dt) in cardiac muscle (ca. 200 V/s) and smooth muscle (ca. 5 V/s), as compared with skeletal muscle (ca. 600 V/s), is another factor that contributes to the lower than expected 0 in cardiac muscle and smooth muscle. In addition, the higher the extracellular resistance, which depends on the tightness of packing of the fibers in the muscle, the slower the velocity (e.g., see Jack et al., 1975).
408
SECTION III Membrane Excitability and Ion Channels
!i. Experimental Facts Some of the key experimental facts relevant to the transmission of excitation from one cell to the next in cardiac muscle and visceral smooth muscle can be summarized as follows. These tissues can be enzymatically separated into their individual cells, and the individual single cells are viable and functional. Gap junctions are absent in the hearts of lower vertebrates and in some regions of mammalian hearts, as well as in some visceral smooth muscles, as stated previously. The length constant h of cardiac muscle and smooth muscle tissues, when measured properly, is relatively short, that is, less than 0.5 mm (i.e., not much more than about one cell length). The input resistance (gin) of cardiac muscle and smooth muscle, when measured properly, is relatively high, about 5--40 MI~. The short h and high Rin suggest that the cells are not profusely connected by low-resistance pathways (e.g., by gap junction channels). Thus, even when gap junctions are present, the junctional membranes constitute a substantial barrier to current flow from one cell to thenext. As stated in Section I, the true h of individual cells is much greater than the cell length, so that there is almost no voltage decay in a single cell and that the entire cell fires an AP nearly simultaneously. Therefore, most of the propagation time is consumed at the cell junctions, and propagation in these tissues is a discontinuous process, in contrast to a continuous process for skeletal muscle fibers and nonmyelinated nerve fibers. Propagation velocity in cardiac muscle and smooth muscle is slower than what can be accounted for by the smaller fiber diameter and lower + max dV/dt. The cell-to-cell transmission process is quite labile, somewhat like that in synaptic transmission. The reader is referred to several review-type articles by Sperelakis and colleagues for additional details and evidence for some of the statements made here. Considerable data have been published concerning the degree of spread of electrotonic current between neighboring cells in cardiac muscle and visceral and vascular smooth muscles. Only one example is presented here for frog cardiac (venlricular) muscle. In these experiments, electrode 1 was used to record voltage and to inject current (using a bridge circuit), whereas electrode 2 recorded voltage only. As illustrated in Fig. A-1, using a pair of microelectrodes whose tips were spaced 11 txm apart, in some double impalements there was no electrotonic current spread between the two electrodes (e.g., Fig. A-1A), whereas in other impalements there was substantial spread of current (e.g., Fig. A-1B). In one unusual case (Fig. A-1C-D), the double impalement first showed good spread (i.e., interaction) between the electrodes (C), but then due to muscle contraction, one electrode left that cell and impaled a neighboring cell having a normal resting potential (fight portion of Fig. A-1C); now there was no significant interaction between the two neighboring cells (D). In the impalements in which there was substantial interaction between the electrodes (e.g., Fig. A-1B, C), it was proposed that the two electrodes had impaled the same cell (e.g., both electrodes recorded low resting potentials, perhaps due to damage caused by the two electrodes impaling one cell). In the impalements in which there was little or no interaction (e.g., Fig. A-1A, D), it was proposed that the electrodes had impaled neighboring cells. This interpretation is most clear in Fig. A-1D, in which the two electrodes
A
_
_ A a
C ..........
_ _
O
Imv
_
-60
F I G U R E A-1. Three typical experiments (A, B, and C-D) measuring the spread of electrotonic current between two closely spaced intracellular microelectrodes in intact frog ventricular trabeculae at rest and during the plateau of the action potential. The interelectrode distance was 11 p.m. Rectangular hyperpolarizing current pulses (approx. 150 ms in duration) were applied. Two successive sweeps of the oscilloscope are superimposed in each panel. (A) Interaction at rest and during the plateau was nearly 0%. Capacitive transients are only seen on the trace from electrode 2, whereas a large maintained hyperpolarization occurred at electrode 1. (B) In another impalement in which the cell was injured by the two electrodes, the resting potential was low and the degree of interaction was high. (C-D) In another impalement in which the cell was damaged, the resting potential was low, and the degree of interaction was high (electrode 1 deflection not shown because of bridge imbalance). The contraction accompanying the action potential caused one of the electrodes to become dislodged from that cell and penetrate into a neighboring cell that had a normal resting potential; the degree of interaction then became nearly zero. (Reproduced with permission from Tarr, M., and Sperelakis, N. (1964). Am. J. Physiol. 207, 691-700.)
recorded markedly different resting potentials. Thus, there appears to be little or no spread of current between neighboring cells in frog cardiac muscle, ff so, then propagation of excitation must occur by some other means.
!11. Electric Field M o d e l A. Electric Field Effect Sperelakis and colleagues developed an electric field hypothesis for propagation of APs in cardiac muscle for situations in which there were no functioning gap junction channels. A computer simulation model for cell-to-cell propagation in cardiac muscle was developed and progressively improved since the mid-1970s (Sperelakis and Mann, 1977; Sperelakis et al., 1985; Sperelakis, 1987; Picone et al., 1991). The model allows electrical transmission to occur between adjacent excitable cells by means of the electric field effect in the very narrow junctional cleft between the contiguous cells (Sperelakis and Mann, 1977). This electric field model does not require low-resistance channels (gap junctions) between cells. The major requirements of the model are that the pre- and postjunctional membranes (pre-JM and post-JM) be ordinary excitable membranes, and that these membranes be very closely apposed to one another (i.e., the junctional cleft be very narrow, about 10 nm). When the pre-JM fires, the cleft between the cells becomes negative with respect to ground (the interstitial fluid sur-
24. C a b l e P r o p e r t i e s and P r o p a g a t i o n of A c t i o n P o t e n t i a l s
409
rounding the cells), and this negative cleft potential (about - 4 0 mV) acts to depolarize the post-JM by an equal amount (namely, 40 mV) and brings it to threshold. (The inner surface of the post-JM remains at nearly constant potential with respect to ground.) This, in turn, brings the surface membrane of the postjunctional cell to threshold. Figure A-2A illustrates propagation of an AP along a chain of 10 cells by the electric field effect. In this computer simulation of cardiac muscle, propagation of overshooting APs occurred at a constant velocity of 32 cm/s, and the maximum rate of rise of the AP averaged 209 V/s. As can be seen, the upstroke of each AP exhibited a b r e a k or step, reflecting the junctional transmission process. In the model, a plot of propagation time as a function of distance along a chain of cells has a staircase shape, indicating that almost all propagation time is consumed at the cell junctions and that excitation of each cell is virtually instantaneous (Picone et al., 1991; Sperelakis et al., 1991) (see Fig.
Action potential propagation under standard conditions 40 30 20 10 0 >-E -10 'E -20 -30 -40 -50 -60 -70 -80 -90
)
"
/ ~'rnax = 209 V/s
;~ 0.0
......
1.0
2.0
3.0
, ..........
4.0
5.0
Time (ms)
, ..........
6.0
,
7.0
5.0
..... ./--'./ ....
4.5 4.0 , 9 o.j.ei
3.5 IE
3.0
r E "~
2.5
.m
J, ee 9 /ee
2.0
.=_
1.5
u_
1.0
/
2 3 4 5
0.5
........
o
eoe,o/t'
~~
o=.o
'~
9 1
l, e l e
0.0
/
9
,
I
~,
~
;~
s
o
Cell number
7
8
9
~o
FIGURE A-2. Electric field model for propagation in cardiac muscle using a computer simulation. (A) Successful propagation of an AP at constant velocity along a chain of 10 cells under standardconditions. Propagation velocity was 32 cm/s; max dV/dt was 209 V/s. Note the small step, or prepotential, on the AP upstroke. (B) Plot illustrating that junctional delays occupy most of the propagation time. Firing time for an AP to spread along the chain of 10 cells under standard conditions (corresponding to part A) is plotted against distance. Within each cell, the units are numbered 1 through 6 from left to right, 1 corresponding to the post-JM (input), 2-5 the surface membrane, and 6 the pre-JM (output). The AP travels along the surface membrane at a high velocity. At the junctions between cells, there is a significant conduction delay as the AP "jumps" across the cleft. This clearly demonstrates the discontinuous nature of AP propagation. (Adapted with permission from Figs. 2 and 3 of Sperelakis, N., Ortiz-Zuazaga, H., and Picone, J. B. (1991). Innov. Tech. Biol. Med. 12, 404-414.)
A-2B). In this figure, it can also be seen that the prejunctional membrane fires a fraction of a millisecond before the surface membrane of the same cell, as required by the electric field model. Propagation was found to be strongly dependent on radial cleft resistance (R,o) and the junctional membrane properties. There was an op[lmal Rjc for maximum propagation velocity under any given conditions. This model is consistent with many experimental facts about propagation in cardiac muscle, and provides an alternative mechanism for AP propagation that does not require low-resistance pathways to transfer excitation directly between adjacent cells. The electric field model can also account for the fact that propagation in cardiac muscle is actually discontinuous or saltatory in nature (Spach et al., 1981; Rudy and Quan, 1987a, b).
B. High Density of Fast Na + Channels at Intercalated Discs For this mechanism to work efficiently, there is a requirement for the prejunctional membrane to fire an AP a fraction of a millisecond before the contiguous surface membrane (Sperelakis and Mann, 1977). That is, the prejunctional membrane should be more excitable than the surface membrane, which would cause it to reach threshold first. It was suggested that this situation would be achieved if there were a greater density of fast Na + channels in the junctional membranes (intercalated discs) than in the contiguous surface membranes (Sperelakis and Mann, 1977). This would make the intercalated discs more excitable and give them a lower threshold than the surface membrane. To examine this possibility, a polyclonal antibody raised against fast Na + channels (from rat brain) was used to immunolocalize the fast Na + channels in rat atrial and ventricular tissues. In immunofluorescence examination, intense labeling was observed associated with the intercalated discs of both atrial and ventricular cells (Ferguson, D., Sperelakis, N., and Angelides, K. J., unpublished observations) (Fig. A-3). This fluorescence was more intense than that of the cell surface membrane. These findings are in agreement with results reported by Cohen and Levitt (1993). Therefore, it is likely that, not only are fast Na + channels present at the intercalated disc membranes, they are present in a higher concentration (density) than in the surface cell membrane. A higher density of fast Na + channels in the junctional membrane would cause it to have a lower threshold than the surface membrane, and so it would reach threshold and discharge first. Thus, not only are the intercalated discs composed of excitable membranes, but they also may have greater excitability than the contiguous surface membranes. This would be analogous to the initial segment of the axon of the anterior horn neuron having a lower threshold, and therefore discharging before the soma or proximal dendrites (where the excitatory synapses are actually located). Consistent with these findings, it was reported that K + channels are also localized at the intercalated discs (Mays et al., 1995). Therefore, the electric field model is a plausible mechanism for cell-to-cell transmission of excitability in cardiac muscle and in visceral smooth muscle. Electric field effects may also occur between closely spaced contiguous neurons in the central nervous system.
410
SECTION III Membrane Excitability and Ion Channels
FIGURE A-3. Immunofluorescence localization of Na + channels in adult rat cardiac muscle (ventricle) using polyclonal antibody to the fast Na + channels of rat brain. The antibody is most densely localized at the cell junctions (intercalated discs), but also stained the surface cell membrane. These results suggest that the intercalated disks are highly excitable membranes, even more excitable than the surface cell membrane, consistent with the requirement of the electric field model for propagation between cells not connected by low-resistance tunnels (gap junction connexons). (From Ferguson, D., Sperelakis, N., and Angelides, K. J., unpublished observations.)
IV.
Electronic
Model
for Simulation
of Propagation
Sperelakis and colleagues (1990) constructed an electronic model to simulate an excitable membrane, and this model was used as a new circuit model of propagation. Four such circuits were successfully made to interact with one another through capacitive coupling to simulate propagation over four cells. Adjustment of one parameter either slightly depolarized and caused repetitive spontaneous APs, or hyperpolarized slightly and depressed excitability, causing partial block (e.g., 2:1, 3:1, 4:1) or complete block of propagation from cell to cell at the cell junctions. When four such cells were connected head to tail in a closed loop, reentry of excitation occurred and could keep going for many seconds before dying out (due to failure at one of the labile junctions). Several configurations of external networks were used in this model to study possible electrical field coupling in cell chains. In a subsequent study (Ge et al., 1993), 12 such units were arranged to model two adjacent cells: 4 units for each surface membrane and 1 for each junctional membrane. The first unit was stimulated to threshold, and AP propagation spread over the two cells was recorded. The time delay at each junction between two adjacent cells was measured as a function of R. the radial shunt resistance at the cell junction. The experimental results showed that the time delay was about 1-2 ms w h e n Rjc values were changed JC'
from infinite down to about 100 k12; the time delay increased at lower R.jc values, and propagation was blocked when Rjc was. below 10 kl~ " Excitation of cell 1 caused . . hy. perpolanzatlon of the post-JM (unit 7) and depolarlzaUon of the other units in cell 2, due to current flow prior to triggering of the AP in cell 2. Raising the effective coupling resistance (Rc) between the two cells increased the conduction delay. As expected, the junctional delay increased when CJ was lowered. Stimulating at the pre-JM (unit 6) resuited in bidirectional propagation. Making the pre-JM inexcitable did not prevent excitation of cell 2, reflecting the presence of longitudinal current flow. However, the postJM (unit 7) only fired an AP when unit 6 was active, reflecting the electrical field effect across the junction. Thus, in this model, both local-circuit current and the electrical field effect play roles in the transfer of excitation. Records from an electronic model of two myocardial cells with a cell junction between them are illustrated in Fig. A-4. In this example, the first cell was set to be spontaneously active and fire a spontaneous cardiac-like AP. All four surface-membrane units in cell 1 (U2, 3, 4, 5) fired an AP nearly simultaneously. However, there was a short delay (e.g., ca. 1 ms) in the firing of the unit (U6) representing the prejunctional membrane (at the intercalated disc). Firing of U6 led to the firing of all units (U7, 8, 9, 10,11, 12) of cell 2 after a junctional delay of about 2 ms. As can be seen, the firing of the prejunctional membrane (U6) drove inward hyperpolarizing current through the postjunctional membrane (U7) and outward depolarizing current through the other units (U8, 9, 10, 11, 12) of cell 2.
-~~--------~ v : " --------~ j ;
I0mV
' : . . . . . . . . .
j
. . . . . . . .
'
U12
: " " .
~
::~ -
..
e.
.
'":
-
,
. . . . . . . . . . . . . . . .
-
.
~,
i
0
10
20 30 Time (ms)
40
50
F I G U R E A-4. Action potentials produced in an electronic model of two heart cells with a cell junction between them. There were six excitable-circuit units in each cell: four representing the surface cell membrane (U2, 3, 4, 5 and U8, 9, 10, 11) and one for each junctional membrane (U1, 6 and U7, 12). The parameters in the first cell (cell 1) were adjusted so that it fired spontaneously in response to pacemaker potential depolarization. Note the nearly simultaneous firing of all units in cell 1, and after a slight junctional delay time, the nearly simultaneous firing of all units of cell 2. (Reproduced with permission from Fig. 5A of Ge, J., Sperelakis, N., and Ortiz-Zuazaga, H. (1993). Innov. Tech. Biol. Med. 14, 404-420.)
24. Cable Properties and Propagation of Action Potentials
Bibliography Cohen, S. A., and Levitt, L. K. (1993). Partial characterization of the rill sodium channel protein from rat heart using subtype-specific antibodies. Circ. Res. 73, 735-742. Cole, W. C., Picone, J. B., and Sperelakis, N. (1988). Gap junction uncoupling and discontinuous propagation in the heart: a comparison of experimental data with computer simulations. Biophys. J. 53, 809-818. Ge, J., Sperelakis, N., and Ortiz-Zuazaga, H. (1993). Simulation of action potential propagation with electronic circuits. Innov. Tech. Biol. Med. 14, 404--420. Jack, J. J. B., Noble, D., and Tsien, R. W. (1975). "Electric Current Flow in Excitable Cells." Oxford University Press, Oxford. Mann, J. E., Jr., Sperelakis, N., and Ruffner, J. A. (1981). Alterations in sodium channel gate kinetics of the Hodgkin-Huxley equations on an electric field model for interaction between excitable cells. IEEE Trans. Biomed. Eng. 28, 655-661. Mays, D. J., Foose, J. M., Philipson, L. H., and Tamkun, M. M. (1995). Localization of the Kvl.5K + channel protein in explanted cardiac tissue. J. Clin. Invest. 96, 282-292. Picone, J. B., Cole, W. C., and Sperelakis, N. (1989). Discontinuous conduction in cardiac muscle. In "Cell Interactions and Gap Junctions" (N. Sperelakis and W. C. Cole, Eds.), Vol. II, pp. 143-154. CRC Press, Boca Raton, FL. Picone, J. B., Sperelakis, N., and Mann, J. E., Jr. (1991). Expanded model of the electric field hypothesis for propagation in cardiac muscle. Math. Comp. Mod. 15, 17-35. Rudy, Y., and Quan, W. L. (1987a). A model study of the effects of the discrete cellular structure on electrical propagation in cardiac tissue. Circ. Res. 61, 815-823. Rudy, Y., and Quan, W.-L. (1987b). Effects of the discrete cellular structure on electrical propagation in cardiac tissue. In "Activation, Metabolism, and Perfusion of the Heart--Simulation and Experimental Models" (S. Sideman and R. Beyar, Eds.), pp. 61-76. Martinus Nijhoff, Dordrecht.
41 1
Spach, M. S., Miller, W. T., III, Geselowitz, D. B., Barr, R. C., Kootsey, J. M., and Johnson, E. A. (1981). The discontinuous nature of propagation in normal canine cardiac muscle. Evidence for recurrent discontinuity of intracellular resistance that affects the membrane currents. Circ. Res. 48, 39-54. Sperelakis, N. (1987). Electrical field model for electric interactions between myocardial cells. In "Activation, Metabolism, and Perfusion of the Heart--Simulation and Experimental Models" (S. Sideman and R. Beyar, Eds.), pp. 77-113. Martinus Nijhoff, Dordrecht. Sperelakis, N., and Mann, J. E., Jr. (1977). Evaluation of electric field changes in the cleft between excitable cells. J. Theor. Biol. 64, 71-96. Sperelakis, N., and Picone, J. (1986). Cable analysis in cardiac muscle and smooth muscle bundles. Innov. Tech. Biol. Med. 7, 433--457. Sperelakis, N., Marschall, R., and Mann, J. E. (1983). Propagation down a chain of excitable cells by electric field interactions in the junctional clefts: effect of variation in extracellular resistances, including a "sucrose gap" simulation. IEEE Trans. Biomed. Eng. 30, 658--664. Sperelakis, N., LoBrocco, B., Mann, J. E., Jr., and Marschall, R. (1985). Potassium accumulation in intercellular junctions combined with electric field interactions for propagation in cardiac muscle. Innov. Tech. Biol. Med. 6(1), 24-43. Sperelakis, N., Picone, J. B., and Mann, J. E., Jr.(1989). Electric field model for electric interactions between cells: an alternative mechanism for cell-to-cell propagation. In "Cell Interactions and Gap Junctions" (N. Sperelakis and W. C. Cole, Eds.), Vol. II, pp. 191-208. CRC Press, Boca Raton, FL. Sperelakis, N., Rollins, C., and Bryant, S. H. (1990). An electronic analog simulation for cardiac arrhythmias and reentry. J. Cardiovasc. Electrophysiol. 1, 294-302. Sperelakis, N., Ortiz-Zuazaga, H., and Picone, J. B. (1991). Fast conduction in the electric field model for propagation in cardiac muscle. Innov. Tech. Biol. Med. 12(4), 404--414. Tarr, M., and Sperelakis, N. (1964). Weak electronic interaction between contiguous cardiac cells. Am. J. Physiol. 207, 691-700.
Richard D. Veenstra
Appendix 2: Derivation of the Cable Equation and the AC Length Constant
I. M e m b r a n e P r o p e r t i e s The biological membrane has both resistive and capacitive elements associated with it due to the ion transporters and channels that permit the translocation of electrical charge across the membrane lipid bilayer that separates the cytoplasm from the extracellular fluid. Fig. A-5 illustrates an axon and the equivalent electrical circuit representation in a manner analogous to Fig. 3 in Chapter 24. For a uniform linear cable of infinite length (>5A), the voltage ( V ) at distance x from the point source of the applied voltage (Vo) decays exponentially according to the expression V = V e --x/A where h is the length constant for the axon (see Eq. 1 in Chapter
A
I.dx.I
_.."
-
"
.."
"
.."
-
-
-
.."
,.."
"
B A
A
~
E.
A
24). Since the equivalent membrane circuit at any point also has a capacitive element (Cm) in parallel with the membrane resistance (rm), the voltage at point x will decay following the termination of a long-duration applied voltage (>5z) in proportion to the value of rmC m o r time constant for the membrane. Another important concept that is not intuitively obvious from the linear cable diagram is that r m and the internal resistance (ri) a r e not constants, but depend inversely on the radius of the axon. This is because, in simplest terms, two equivalent resistors in parallel yield half the resistance of their initial value (like two traffic lanes can carry twice as many cars as one, all else being equal). The membrane resistance varies according to the surface area of the membrane while the extracellular (ro) and cytoplasmic resistances vary according to the volume of the extracellular fluid or cytoplasm. For a simple fight cylinder as illustrated in Fig. A-5, the circumference of the axonal membrane is equivalent to 27ra and the cross-sectional area is equivalent to 7ra 2, where a is the radius of the axon. It follows that for a cylindrical axon, the surface to volume ratio (Sv) is proportional to 2/a. Resistivity (R) is classically defined as the resistance of a 1 cm 2 area of membrane. For these reasons the extracellular and internal resistivities, R ~ and Ri, have units of fbcm. Since r m varies inversely with membrane surface area, r m = R/area and the membrane resistivity has units of l).cm 2. Every centimeter of membrane also has a specific membrane capacitance, C m, equivalent to 1 ~aF/cm 2 for biological lipid bilayers. Hence, C m is expressed in units of F/cm 2 since the capacity increases in direct proportion to the area. For a right cylindrical cable, the values of the resistance and capacitive elements at any point x are related to their specific resistivities and capacitance by the following expressions:
v
FIGURE A-5.
Diagram of a linear cable. (A) Cylindrical model of a linear cable of length x and radius a. In order to derive equations for axial and transverse current flow, the cable is evaluated in small increments of distance (dx) along its entire length. (B) Equivalent electrical circuit for the cable illustrated in A with an axial internal core resistance, r i, a n external resistance, ro, and a parallel resistor-capacitor, (rmCm) circuit which represents a biological membrane. (C) The equivalent circuit for the cable assuming there is no external voltage gradient (i.e., r~ = O, no resistance to external current flow). Cell Physiology Sourcebook: A Molecular Approach, Third Edition
4 12
rm (Jn ~ ~ m ) = Rm / 27ra
(A-I)
r i (in l~/cm) = R i / 7ra 2
(A-2)
r (in l ) / c m ) = R ~
(A-3)
and r
(in IxF/cm) = 27raC m
(A-4)
The actual value for any length d x of the cable can be determined by r m / d x , r i d x , r ~ d x , and c m dx, assuming that the radius a remains constant. The volume of the extracellular space is usually considered to be infinite, so the value of R ~ Copyright 9 2001 by Academic Press. All rights of reproduction in any form reserved.
413
24. Cable Properties and Propagation of Action Potentials
is determined only by the ionic composition of the extracellular fluid. This latter assumption is not true when restricted extracellular spaces are encountered (e.g., tight junctions or tortuous extracellular pathways in bundles of packed fibers). In this case, significant barriers to ionic diffusion are encountered in the extracellular space and the geometry of the intercellular space must be considered. Commonly, the external resistance is assumed to be negligible and r ~ is omitted from further consideration. For now, we will consider the first case when r ~ = 0 (Fig. A-5C).
!!. Linear Cable Theory
From Ohm's law, we know that V = IR, so in order for a voltage to develop across the membrane (Vm) there must be current flowing across the membrane. For a point source of constant current (DC = direct current) injection at the center (x = 0) of an infinitely long cable, Vm will decay uniformly along the length of the cable in both directions. At each point along the cable, the membrane current, im, will generate a membrane voltage, Vm, equivalent to imr m, o r i m = Wm/r m. Since each segment of membrane has a capacitive element, there must also be a capacitive current (ic) component equivalent to c m dVm/dt. That is, the total membrane current ( / ) is the sum of the ionic c u r r e n t (Ii) through the resistance and the capacitive current (Ic) through the capacitance"/m = li + Ic" Hence, + Cm d V m / d t
(A-5)
However, Vm cannot be constant along the entire length of the cable, because the internal axial current flow, i i, would have to remain constant along the entire cable in order for this to occur; that is, there must no leakage of internal axial current across the membrane along the length of the cable. Simply put, Im--ii
/dx
(A-6)
where dx is an infinitesimally small distance along the length of the cable. It follows that the change in V m over the distance dx is d V m / d x = - i i r i. Combining this expression with E,qs. A-5 and A-6 gives
I m-
(l/F/)
d2V/dx 2
(V/Fm) q- Cm N V / N t
(A-7a) (A-7b)
This is referred to as the linear cable equation. Multiplying both sides of the equation by r m yields the following expression:
( r / t ) ) d2Vm/dx2-Vm + (c m rm)dVm/dt
(1-8)
From this version of the cable equation, it is apparent that the value of Vm at the point x + dx (dVm/dx) is directly proportional to the square root of r m/r i. T h e value of v/ r /r/ is referred to as the length constant of the cable: A - / r / r/
A-
(A-9a)
These expressions refer to the passive spread of voltage along the length of a linear cable. It should now become ap-
~r
m
vr+r
(A-9b)
Also, cardiac muscle fibers, unlike axons or skeletal muscle fibers, do not possess a continuous cytoplasm. Instead, each cardiac cell is coupled to its neighbor by gap junctions that add to the internal resistance of the muscle fiber. Under these conditions, A_I/
A. Passive Cable Properties
I m - (Vm/r)
parent that an infinite length of cable is considered to be > 51 since Vx/V o < 1% when x > 51. Other conditions will alter the value of A. For example, when r o > O,
r
r/+r+r
(A-9c)
It is also apparent that the change in voltage with respect to time at the point x + dx is directly proportional to (Cmrm), which is defined as the membrane time constant:
Tm m Cm r m
(A-10)
For the same reasons given for considering an infinite length of cable to be > 51, an infinitely long duration current pulse is considered to be > 5r.
B. Active Cable Properties: Action Potential Propagation The previous conditions apply to the resting muscle fiber or nerve axon, provided that the fiber or axon is several A long and its geometry approximates a linear cable (e.g., no branch points for axial current flow). However, conduction of the action potential involves the spread of current flow along a passive membrane cable segment from an advancing active membrane segment. The transition from a passive to an active membrane response (i.e., the initiation of an action potential) requires that a certain threshold voltage level be achieved in order for the segment of membrane undergoing a voltage change in response to passive current spread (as defined by the cable equation in Section II.A) to generate an active response. We will see later that the excitation threshold is not actually a voltage value; rather, it is equivalent to the amount of charge required to depolarize the membrane and produce an active current response. During the action potential, the active portion of the membrane becomes a source of depolarizing current which contributes to the distal spread of the voltage deflection, r m, or 1/gm, is actually the sum of several ionic components resulting from the presence of specific proteins (ion channels and pumps) that permit the selective movement of ions across the membrane. These discrete resistive membrane elements are associated with the activity of ion channels (e.g., Na § K § Ca 2§ and C1+ + channels) and pumps (e.g., Na ,K -ATPase, Na+-Ca 2§ exchanger) associated with the cell membrane. Since each ion possesses a net charge (valence), the translocation of an ion across the membrane produces a current (li). The activities of the protein channels and transporters are not constant, but are regulated by a variety of mechanisms including membrane voltage (Vm). Hence, at different points in time (dt), r m (and R m ) is not constant, but rather is a function of voltage and time, r m ( V , t). So a voltage deflection dVm alters r m,
414
SECTION III Membrane Excitability and Ion Channels
which further affects V . This constitutes a feedback loop where r m and Vm are dependent variables (see Chapter 24). In a passive cable, any voltage deflection caused by the decrement in axial current flow in that particular segment of membrane is initially opposed by existing membrane currents, which counteract the voltage deflection and maintain the resting (passive) state. This is a negative feedback mechanism and is predominant at all times except during an action potential. When the voltage deflection exceeds the threshold for excitation, certain ion channels are activated, which leads to the generation of more ionic current and a greater voltage deflection in the same direction as the initiating response. Hence, the voltage response of the membrane becomes additive to the initial passive dVm and a positive feedback loop is transiently produced. If allowed to persist, a positive feedback loop leads to prolonged excitation, which prevents the return to the resting state and the generation of any further action potentials. ~ Therefore, the positive feedback loop, in order to be effective in generating an action potential, must also be terminated once that action potential is initiated. The threshold for excitation is the critical value of Vm (Vth) that must be achieved in order for the active membrane response to commence. However, as observed from the equivalent circuit and cable equation, there are both resistive and capacitive elements to the membrane current. Current is also equivalent to the derivative of charge flowing per unit time (e.g., coulomb per second). Hence, the resistive and capacitive current components of i m c a n be expressed in terms of dQ/dt, where Q is the charge in coulombs. From Faraday's law applied to our segment of membrane,
Om-fmWm
(m-11)
So the threshold for excitation is actually the amount of charge (Qth) that must flow across the membrane to achieve the required membrane voltage (Vth) to initiate the action potential. Once initiated, the action potential propagates down the length of the cable at a rapid rate. The conduction velocity (0) is the distance traveled per unit time, or dx/dt. Returning to our cable equation (Eq. A-8) and solving for dx/dt, we obtain /~2 (1/02) d 2 V / d t 2 _ V +
T m
dV /dt
(A-B)
(1/r.) d 2 V / d x 2
(A-7a)
I m --
(Fm/Fi) d 2 W / d x 2 - W -k- (c m Frn) d V / d t h2
or
(A-13)
Now we have successfully related the cellular membrane properties (rm, ri, Cm, etc.) to the passive cable properties (a, r), and subsequently to the active cable properties (0) involved in the conduction of an action potential. To summarize, the cable equation can be expressed in several equivalent forms-
(A-8)
(1/02) d2Vm/dt 2 _ Vm + Tm d V / d t
(A-12)
Another useful derivation of the cable equation provides insight into how cell size, or more specifically cell surface/volume ratio (Sv), influences propagation (Joyner et al., 1983):
rm 1 d2Vm/dt2 - V + c r d V m / d t Ri + Sv 0 2 m m
(A-14)
Several additional derivations are possible which depend on the boundary conditions applied to the cable such as length, extracellular isopotentiality (r ~ = 0), voltage uniformity (i.e., voltage-clamp), and uniformity of geometry (e.g., fiber radius a or branch points).
C. Finite Spatial and Temporal Responses 1. Uniform Cable Let us consider further our linear cable of uniform diameter 2a, length x, and external resistance r o = 0. Throughout the passive cable analysis described above we have only considered the case where the voltage displacement along the length of this cable is produced by a constant current (DC) source, I 0, located at the exact center (x = 0) of the cable. We have shown that A2 = rm/r i or A2
=
(Rm/27ra)
(A-15a)
( Ri / 7ra2 )
aR m 2R i or
(A-15b)
/Rma
(m-12)
It is now evident that the rate at which the action potential propagates along the length of passive cable elements depends upon the length and time constants of the membrane
O2 oc h / r
I m - ( V m / r m ) + Cm d V m / d t
Ri 2
(A-15c)
and that r m = Cmrm (Eq. A-10) or
"rm-(Rm/27ra)(2"traCm)
(A-16a)
=RmC m
(A-16b)
This implies that/~2 (and consequently 02) oc (a/2), whereas zm is independent of fiber radius and membrane area (provided R m is constant). Under steady-state conditions dVm/dt = 0, so the cable equation becomes
0 -- i~2 dZVm/dX 2 - T m d V / d t -
V
(A-lVa)
which further reduces to 1 One example of overamplification and positive feedback is the annoying hum of the microphone heard over the audio speaker system of any soundstage production.
0 - A2 d Z V / d x 2 - V
(A-lVb)
To solve this differential equation, let us def'me X = x/A (Taylor, 1963; Jack et al., 1983). It follows that dX/dx = l/A, so/~2 _
24. Cable Properties and Propagation of Action Potentials
41 5
(dx/dX) 2 and the equation becomes 0 = d 2 V m / d X 2 - V m. T h e solution to d X / d x is eXand 1/(dX/dx) = e-x. The solution to our differential equation therefore becomes V
(A-18a)
- V e -x
or
(A-18b)
V - V ~ e -x/x
where V ~ - the value of Vm at x = 0. As x ~ 0, V m ~ V ~ and spatial uniformity (V m = Vo) of membrane voltage is approached. This is identical to Eq. 1 in Chapter 24. Another consideration is the time required for the membrane to decay from its steady-state voltage once a current stimulus is terminated. Again, 0 - h 2 d2Vm/dx 2-
7" m
dV/dt-
V
(A-17a)
except in this case d Z V m / d x 2 = 0, since we are observing the same point x at different times during the current injection and dVm/dt r O. Now our equation reduces to 0
-- T m
dV/dt
(A-19)
+ V
which is again a differential equation. Let us define T = lh" m, and it follows that d T / d t = 1/'r m (Taylor, 1963; Jack et al., 1983). Our differential equation takes the form (A-Z0)
0 - dV / dT + Vm
which has the solution (A-Z1)
W m - V o e -t/T
where V ~ = the value of V when t = 0. This is identical to Eq. 6 in Chapter 24. In conclusion, for a uniform fiber, V decays over distance and time by a first-order exponential decay process.
2. Nonuniform Cable
Vm - Vo e f f v/ t / r
which differs from the previous derivation of the timedependent voltage response for a spatially uniform cable. Since err (1) = 0.84, when t = T, Vm = 0.84V o for the nonuniform fiber or axon instead of 0.63V o in the case of the uniform cable (Eq. 24, Jack et al., 1983). This means that V m will rise and fall faster in a long fiber with a point source of current injection than in a short cable where spatial voltage uniformity is approximated (x/A--l) or a long fiber made short by injecting current uniformly over distance x. Hence, under non-voltage-clamp conditions, the 84% time provides the measure of rmC m, not the 63 % time as indicated in Eq. A21 (Jack et al., 1983).
D. Input Resistance The magnitude of the steady-state membrane voltage deflection produced by a DC current source will be determined by Ohm's law, Vo - / o R i n where the input resistance (R in) of the cable represents the effective (net) resistance of the preparation. The input resistance is dependent on the geometry of the preparation and is not a direct measure of r m unless the preparation is under the conditions of spatial voltage uniformity and leakage around the current and/or voltage electrodes is negligible. Rin is best defined as the resistance of the preparation resulting from the internal resistance and membrane resistance over the length of the fiber. For intracellularly injected current, all current must flow across the internal resistance and will proportion itself between axial and membrane current flow according to the length constant as previously derived. Since half of the injected current will flow in either direction down the long axis of the cable,
These solutions to the differential equations apply only to the specified conditions described above. For a nonuniform cable (but r ~ still = 0), the generalized cable equation 0 = d Z V m / d X 2 - d V m / T - V will have different complex solutions. The solutions to several different boundary conditions are derived in Rall (1977). In situ, the condition of voltage uniformity does not apply to a muscle fiber or nerve axon due to a large r i. Hence, for a constant current injection (Io) at X = 0 and time T , the cable equation has the general solution of (Hodgkin and Rushton, 1946)
-eXIl-eff[X+2~/T
/11
~-~
For x = x and t = o% this equation reduces to V m - V o e -x/~t
(A- 18b)
which is the same as the solution derived for the decay in V m as a function of distance. However, the special solution for V m as a function of time at x = 0 has the solution
(A-23)
V ~ - ri I ~ A / 2
(A-24)
gin - / Frn rt" / 2
(A-25a)
where Rin -- ri/~, or
IIR R. = 5 2~'~a 3
(A-25b)
Hence, the input resistance of a fiber is proportional to v/1/a3(inversely proportional to the square root of the volume). With these derivations it should be apparent how the length constant, time constant, voltage, conduction velocity, and input resistance of a one-dimensional fiber or axon are related to the parameters of radius, membrane resistance, axial resistance, and current. Other derivations are based on different boundary conditions (e.g., Rall, 1977).
E. Alternating Current When an alternating current (AC) is applied to an RC circuit, the solution is more complex. Capacitive current,/c, is C m dV/dt, and for a sine wave, d V / d t = 0 when the current is at its maximum (In). Simply stated, the voltage across the capacitor (Vc) lags b~hind the current across the resistor by 90 ~ The current across the resistor is now a sine wave and takes the form of Vp sin (tot)/r m where Vp = peak voltage of the sine
416
SECTION III Membrane Excitability and Ion Channels
wave. Since dV sin (tot)/dt = V to cos (tot) = Vp.to sin (coto P 90 ), the current equation (Eq. AP-5) for our circuit becomes
AAC-/2/27rfCmr i -/1/TrfCm
Ip - V sin(wt) / r + c m V w sin(ogt - 90 ~
(A-26)
where w = 2rrf and f = frequency of the sine wave (in Hz). Ohm's law states that V = IR, so it follows that 1/ooC has units of resistance (12). Let us define the capacitive reactance, (A-27)
Xc--1/wC
Let us further define impedance, (A-28)
Z- R + jX
where R is the purely resistive component, X is the reactance of the circuit, and j = v/-1. Actually, the membrane should always be thought of as having an impedance. In the DC case, X = 0 because ~o = 0 (since there is no frequency component) and the membrane exhibits only a resistance. Since the axial and external resistances in our linear cable do not contain a capacitive component, R i and R ~ are unaffected. Recall that the steady-state solution for voltage as a function of distance x is (A-18b)
V - V e -x/a
where Vo = ( r i l a ) / 2 , and a = v / r / r . . We can rewrite Eq. A18b by substituting for V~ and A in the above expression, thus obtaining V - (Io/2)
~
e-X~,/
r,,,Ir,
(A-29a)
Substituting z m for r m in the above expression results in the following expression: ri
e-X~/Zm/ri
1 / R m Ri e i n - / F m r. / 2 - - ~ ~ 2~)a--3
(A-25a, b)
The resistance and capacitor in series form a low-pass filter as we have already observed in Eqs. A-21 and A-23. When f < 1/(2rrRC), the voltage output (Vout) relative to the voltage input (Vi~) (in response to an AC source) is unity. The ratio of Vout/Vi~ is defined as the gain of the circuit and is equivalent to Vout/Win - 1//~
C2"+-m1 - 1 / / 4 r r 2 f 2 e
2mC2m + 1 (A-36)
since the magnitude of the output voltage is determined by the expression V u t - (Vin X c / Z m )
(A-37)
The equivalent expression for the input impedance of the cable becomes
1J (RmRi )
Zin- 2-
2rr2a 3 4rr2f2R2 C 2 + 1
(A-38)
m
(A-29b)
The membrane impedance, z m, is a complex number consisting of real (resistance) and imaginary (reactance) components. So the above expression should be further subdivided into real and complex numbers. Let us define a propagation constant, y, which is a complex number composed of an attenuation factor, a, and a phase constant,/3, such that y - a + j/3
(A-35)
When f < f/~, r m > X m, and Z m = r m , SO/~ -" / r m /It. , which is equivalent to the DC case. The net effect of the frequency dependence of Aac is that the length constant becomes shorter at higher frequencies (e.g., instantaneous steps in membrane voltage). There is also an analogous expression for the input impedance of the cable that is derived from the previous expression for the input resistance:
m
V - ( I o / 2 ) ,/Z m
r.
This equation was originally derived by Falk and Fatt (1964) in their analysis of frog sartorius muscle fiber linear cable properties. The expression is again divided by two since the source of current injection was the center of the cable and current will flow in both directions along the long axis of the fiber.
(A-30)
The expression for the peak voltage of our circuit is V-(ip/2)
v/IZmlr, e " x
(1-31)
where I :m is the magnitude of the impedance, v / R 2 + X 2 (Eisenberg and Johnson, 1970; see also Chapter 70). It follows from Eq. A-31 that Aac = 1/3'. It is easier to solve for the above expression by assuming the frequency of the injected current is high and z m = j X =j/ogc m. This occurs when f > 1/(2rrrmCm). X = r m when f = 1/(2rrrmCm) =fb where f,, is the cut-off or corner frequency for the circuit. Equation A-31 now becomes V-(ip/2)
v/jr./ogc m e ~ x
(A-32)
and 3' = 1/A = 1/v/ z m / r. - 1/v / j / wc m r. = a + j ~ . Solving for a and/3, we obtain ce - 1 / / 2 / o 9 c m r. - / o 9 c m r//2
(A-33)
j,8 = j v/ ~OCmri / 2
(A-34)
and
Since AAC = 1/% it follows that AAC = 1/c~ = v/2/COCm t) or
Bibliography Eisenberg, R. S., and Johnson, E. A. (1970). Three-dimensional electrical field problems in physiology. Prog. Biophys. Molec. Biol. 20, 5-65. Falk, G., and Fatt, P. (1964). Linear electrical properties of striated muscle fibres observed with intracellular electrodes. Proc. R. Soc. Lond. B 160, 69-123. Hodgkin, A. L., and Rushton, W. A. H. (1946). The electrical constants of a crustacean nerve fibre. Proc. R. Soc. Lond. B. 133, 444-479 Jack, J. J. B., Noble, D., and Tsien, R. W. (1983). "Electric Current Flow in Excitable Cells." Clarendon Press, Oxford. Joyner, R. W., Picone, J., Veenstra, R., and Rawling, D. (1983). Propagation through electrically coupled cells. Effects of regional changes in membrane properties. Circ. Res. 53, 526-534. Rail, W. (1977). Core conductor theory and cable properties of neurons. In "Handbook of Physiology" (J. M. Brookhart and V. B. Mountcastle, Eds.), Vol. 1, pp. 39-97. American Physiological Society, Bethesda, MD. Taylor, R. E. (1963). Cable theory. In "Physical Techniques in Biological Research" (W. L. Nastuk, Ed.), Vol. 6B, pp. 219-262, Academic Press, New York.
Nicholas Sperelakis
Electrogenesis of Membrane Excitability
By the 1940s, improvements had been achieved in electronic instrumentation, especially in the high-input impedance amplifiers necessary to record bioelectric phenomena using tiny intracellular microelectrodes. In addition, biophysicists began to study the squid giant axon (500-1000 lam in diameter), which permitted insertion of relatively large intracellular electrodes, yielding the first measurements of the true transmembrane potential. The transmembrane potential is recorded as the difference between an intracellular and an extracellular electrode (Fig. 1). The findings from the squid giant axon were successfully applied to the smaller diameter (1-20 pm) neurons found in the vertebrate nervous system (Fig. 2).
I. Introduction Excitability is an intrinsic membrane property that allows a cell to generate an electrical signal or action potential (AP) in response to stimuli of sufficient magnitude. The elongated nerve axon serves to transmit information in the form of APs over long distances. The AP mechanism is required to propagate a uniform signal in a nondecremental manner. In muscle cells, the AP serves to spread excitation over the entire cell surface and is involved in triggering cell contraction. The energy source for the generation of the AP is stored in the excitable cell itself. An initial depolarization is produced by a stimulus and triggers the intrinsic AP mechanism. The immediate source of energy for the AP comes from the transmembrane ionic gradients for K + and Na +, which act like a battery. The K + ion concentration gradient is mainly responsible for the generation of the resting potential, which causes an excess of negative charge to build up on the inner surface of the membrane. Upon depolarization to threshold, the Na + ion electrochemical driving force, which is directed inward, causes a large and rapid inward Na + current that generates the AP upstroke. Over a longer time frame, the Na+-K + pump is responsible for the generation of the Na + and K + ionic gradients and for their maintenance and restoration after repetitive AP activity. The Na+-K + pump derives chemical energy from the hydrolysis of ATE The bases of the resting potential and active ion transport have been discussed in earlier chapters. Important technological improvements have led to significant advances in the understanding of the basis of membrane excitability. In the early 1900s, several theories were proposed to explain the mechanism that produces the AE Julius Bernstein (1902, 1912) proposed that the excitable cell membrane, at rest, was selectively permeable to K + ions (producing the resting potential), and thatduring excitation the membrane became permeable to all ions (producing the AP). About the same time, Overton (1902) had demonstrated that Na + ions were essential for excitability.
Cell Physiology Sourcebook: A Molecular Approach, Third Edition
4l
11. Action P o t e n t i a l C h a r a c t e r i s t i c s Action potentials in a given fiber (e.g., myelinated nerve fiber) have the properties of being all-or-none, having a sharp threshold, and having a refractory period. All impulses look alike, being very similar in shape, amplitude, and duration. Thus, they constitute a digital system. Strength of the signal is conveyed by changing the frequency of the impulses, as discussed in the preceding chapter on propagation. The all-or-none property means that the single signal is not graded in amplitude (i.e., not an amplitude-modulated system) and that the signal is either zero ("off" or "no") or maximum ("on" or "yes"). These characteristics of the AP are discussed in the next sections.
A. Local-Circuit Currents During propagation of an AP down a nerve fiber, current flow accompanies the propagating change in membrane voltage. This current is called the local-circuit current, and has both longitudinal and radial (transverse) components that make a complete circuit (Fig. 3A, B). Propagation and localcircuit currents have been thoroughly discussed in the preced-
7
Copyright 9 2001 by Academic Press. All rights of reproduction in any form reserved.
418
SECTION III Membrane Excitability and Ion Channels
A
,)
() Extracellular
0 Intracellular
T
!)
C 0
a
(mY) -70 . . . . . . . . . . .
b
FIGURE I. Recording the transmembrane potential of a squid giant nerve axon. The transmembrane potential is measured as the potential difference between the intracellular and extracellular electrodes. (A) The microelectrode is outside the axon, and measures 0 mV (segment a in C). (B) As the electrode is advanced and crosses the membrane, the resting potential of-70 mV is measured (segment b in C). The cell membrane makes a tight seal around the glass electrode. (C) The membrane potential (Em) is depicted. Stimulation (at arrow) elicits an action potential.
ing chapter. For our purpose here, we focus on the longitudinal component. The intracellular and the extracellular (external) longitudinal current are exactly equal in amplitude, but flow in opposite directions (Fig. 3A, B). The intracellular current is, of course, confined to the cross-sectional area of the nerve fiber (neuroplasm), whereas the extracellular current can use the entire extracellular volume (so-called volume conductor). If a single nerve fiber or nerve bundle is mounted in air, then the extracellular action current is confined to the surface film of fluid adhering to the single fiber or the interstitial fluid space between fibers in the bundle. To illustrate the principles involved, let us record externally with a pair of electrodes from a single nerve fiber (e.g.,
Cell body (soma) ~..~/~j,~
.\
/ ~ ~ / / '~'~
Initial segment
Nucleus Dendrite
Node of Ranvier |
S chwann cell
//~ Terminal buttons
FIGURE 2. Diagram of a motor neuron with a myelinated axon. The major structural elements are diagrammed, including the cell body, initial segment, Schwann cell, and node of Ranvier.
squid giant axon) bathed in air (Fig. 3C). As the AP moves from left to right past the pair of recording electrodes, the first electrode becomes negative with respect to the second electrode. Therefore, the voltmeter swings in one direction (an upward deflection is defined as negative in extracellular recording) due to the potential difference (pd). When the wavefront moves to a position between the two electrodes, the voltmeter will record almost zero voltage. When the AP reaches the second electrode, the voltmeter will swing in the opposite direction, because the second electrode is now negative with respect to the first. Thus, the electrodes will record a biphasic change in voltage as the AP sweeps past. Thus, when recording externally, the AP is often described as a wave of negativity sweeping down the fiber; when recording internally, the AP is a wave of positivity. The reason for the wave of negativity in external recording is that the fluid between the two electrodes constitutes a resistance (R), and longitudinal current (/) flowing through this fluid resistance produces an IR voltage drop (Ohm's law: V = IR). As the wavefront approaches the first electrode, the external local-circuit current is from fight to left and produces an IR drop, positive (second electrode) to negative (first electrode). Conversely, when the AP passes beyond the first electrode, the action current is now reversed in direction and produces an IR drop, positive (first electrode) to negative (second electrode). This produces the biphasic voltage recording. Propagation velocity is a function of axon diameter and myelination, such that the larger the diameter, the greater the velocity. Myelinated axons conduct much faster than nonmyelinated axons. The factors that determine propagation velocity and the mechanism of saltatory conduction are discussed in the preceding chapter.
B. Threshold and All-or-None Property Nerve membrane responses near the site of application of brief current pulses vary depending on the magnitude and direction of the pulses (Fig. 4). Anodal (inward) currents produce hyperpolarization, and cathodal (outward) currents produce depolarization. Hyperpolarizing and subthreshold depolarizing responses are graded in magnitude according to the stimulus current. However, as can be seen, a somewhat higher intensity outward current produces a depolarizing response with a different waveform and a longer lasting duration. This condition is referred to as the local excitatory state (Fig. 4A, B). It occurs when a small area of the membrane near the stimulus electrode comes close to threshold, but it does not generate an AP. This local membrane activity is not propagated and decays with distance along the axon. A slightly stronger stimulus is, however, sufficient to bring the membrane potential of a large enough membrane area to the threshold potential, thereby initiating an all-or-none AP. At the threshold potential, there is a greater amount of inward (depolarizing) than outward (repolarizing) current, and the membrane will continue to depolarize. In regard to the applied stimulus, outward current depolarizes the membrane (IR drop across the resting membrane is positive inside, negative outside), whereas inward current hyperpolarizes (IR drop is negative inside).
25. Electrogenesis of Membrane Excitability
4 19 C
A
J
Cathode-ray oscilloscope
C) - - + +
,
Nerve TTTT ,T
!i
:3
* §§
§
§247 § ...... i)
§
B
I i i 11111' I .....
4-
~-i
1
:3
~
mv
+§-
0
-':)
- ' - § *'i) mV
Distance
Time F I G U R E 3. Diagram of the local-circuit current in a nerve fiber during an AE (A) The current during the rising phase. (B) The current during the entire AP. The AP is depicted as propagating from right to left. (C) The actual registration of the signal is depicted for an AP propagating from left to right past a pair of electrodes spaced relatively far apart.
A
B m
40
(mV)
20 o -20
~h-40 -60 -80
f
tito~ t~
_ / / _ . . . I SubtraCt
0.0
~
"
0.2
-100 -120
~
0.4
0.5
1.0
ms
. response
-ENa
C +50
j(~ -- -- ~-Overshoot
0 -- -- T -'~- - -
(mY) T
h
r
e
s
h
Hyperpolarizing
~
~
_
7
0
--Ek I
0 Stimulating 0
current
I
1
I
2
I
3
.... I
4
9
5ms
J~e,~ Incr si g stimulus strength
F I G U R E 4. Diagrammatic sketches depicting initiation of the nerve impulse by membrane depolarization. (A) Membrane responses to depolarizing and hyperpolarizing current pulses are shown. The nonlinear local excitatory response (LER) occurs just below the threshold for the all-or-none AP. (B) Illustration of the LER obtained by subtraction of the membrane responses to hyperpolarizing current pulses (passive only) from the responses to depolarizing pulses (passive plus active). (C) A nerve AP is depicted, illustrating the sharp threshold, overshoot, and hyperpolarizing afterpotential.
420
SECTION III Membrane Excitability and Ion Channels
In contrast, for the active membrane, inward currents are depolarizing (bringing positive charge inside the cell), and outward currents are hyperpolarizing. That is, when the membrane is behaving passively, applied inward current is hyperpolarizing and outward current is depolarizing; this is equivalent to a circuit external to a battery. In contrast, when the membrane is behaving actively, inward current generated is depolarizing, and outward current is hyperpolarizing; this is equivalent to an internal circuit within a battery. This difference is related to the fact that current flows from positive to negative in the external circuit, and from negative to positive within the battery itself. A subthreshold depolarization is defined as one that does not reach threshold to elicit an AP. It is proportional to the applied stimulus, and it decrements with distance along the nerve axon cable. If a stimulus current is of sufficient magnitude (i.e., enough positive charges are transferred into the cell), then the resulting membrane depolarization reaches a critical value, called the threshold potential, at which an AP is initiated (see Fig. 4). The AP parameters, including overshoot, duration, and rate of rise, are characteristic for each type of excitable cell. For example, the duration of the AP of the squid giant axon is about 1 ms, whereas the cardiac AP lasts for several hundred milliseconds. These differences in the APs subserve the functions performed by the difference excitable tissues. The overshoot and a hyperpolarizing afterpotential (following the spike) are illustrated in Fig. 4C.
C. Refractoriness Once an AP is initiated, a finite and characteristic time must elapse before a second AP can be generated. This time interval is called the refractory period, and its value depends on the type of excitable cell. Cells with long-duration APs (e.g., myocardial cells) have long refractory periods; cells with brief APs (e.g., neurons) have short refractory periods. That is, the refractory periods are proportional to the AP duration. Two types of refractory periods are usually defined: an absolute refractory period and a relative refractory period (Fig. 5). The absolute refractory period denotes the in-
A
B
ARP
+50-O
terval during which a second AP cannot be elicited, regardless of the intensity of the applied stimulus. During the relative refractory period, a second AP may be elicited, provided that a greater than normal stimulus is applied. The second AP often is subnormal in amplitude and rate of rise. Therefore, the physiologically important refractory period is the functional (effective) refractory period, which is defined by the highest frequency of APs that the excitable cell (e.g., neuron) can propagate. For example, if a myelinated nerve axon can propagate impulses up to 1000/s, then the functional refractory period is 1.0 ms. The triggering of a second impulse at a given point (or node) is limited by the amount of action current available from an active point (or node) upstream (unlike an electronic stimulator). Therefore, the functional refractory period encompasses all of the absolute refractory period and part of the relative refractory period. The absolute refractory period extends from when threshold (Vth) is reached at the initial portion of the rising phase of the AP to when repolarization has reached a level of about - 5 0 mV. During further repolarization beyond -50 mV (e.g., to -70 mV), a larger and larger fraction of the fast Na + channels recovers from inactivation, and so more fast channels are again available to be reactivated to produce another AE This is the period that inscribes the relative refractory period. The greater the degree of repolarization (toward the resting potential), the larger the subsequent AP. In addition to voltage, time is a factor in the recovery of the ion channels. Therefore, the absolute and relative refractory periods persist briefly beyond the theoretical voltages, and the relative refractory period actually slightly exceeds the AP duration. Membrane excitability is greatly altered during the refractory periods. Excitability is zero during the absolute refractory period, and is depressed during the relative refractory period, becoming less and less depressed as the membrane repolarizes back to the resting potential (Fig. 6). The afterpotentials that many cells exhibit also affect membrane excitability: hyperpolarizing afterpotentials depress excitability (greater critical depolarization required to reach Vth) and depolarizing afterpotentials enhance excitability. The latter produces a supernormal period of ex-
~~|!I
ARP
+500_-
I
I
-50 -
j S.~.,- i---
RRP
I I
I I I I
I~" I I I
i
~
-50 -70
I I
4
-70
i_
RRP
i ~Hyperpol'arizing afterpotential
FRP I
I
I
0
1
2
....
I
I
I
3
4
5
Time (ms)
F I G U R E 5. Refractory periods of nerve action potentials, both without (A) and with (B) a hyperpolarizing afterpotential. The absolute (ARP), relative (RRP), and functional (FRP) refractory periods are labeled.
421
25. Electrogenesis of Membrane Excitability 100
Recovery 5 o - , of 0 excitability ~ i
Absolute refractory
'l ai/ ~
/
refractory . =
I
o
J
??.
Time (ms) FIGURE 6. The time course of recovery of nerve excitability during the relative refractory period. As shown, during the absolute refractory period, an AP cannot be elicited, and there is a progressive increase in excitability during the relative refractory period.
citability, and the former blends into and extends the relative refractory period. Thus, for example, since neurons usually exhibit hyperpolarizing afterpotentials, the functional refractory period described previously actually includes part of the afterpotential. Propagation velocity is slowed during the relative refractory period, achieving the normal value at the end of the relative refractory period. Propagation velocity is slightly faster than normal during the supernormal period of excitability.
D. Strength-Duration Curve Whether the threshold potential is reached depends on the amount of charge transferred across the membrane. Figure 7 shows that the total charge transfer across the membrane required to produce excitation is approximately constant (since Q = IT). It is an approximate rectangular hyperbola (xy = k) over the sharply bending region of the curve. The strength-duration (S-D) curve can be derived from the equation for the exponential charge of the membrane capacitance.
Minimal = '----stimulation time
J L_ __~
Intensity of stimulus (relative)
xy=k
1.44 o-
The S-D curve deals only with the stimulus parameters (i.e., strength and duration of the applied current pulses) necessary to bring the membrane to threshold. It shows that the greater the duration of the applied pulse, the smaller the current intensity required to just excite the fiber. The asymptote parallel to the x-axis is the rheobase, which is the lowest intensity of current capable of producing excitation, even when the current is applied for infinite time (practically, > 10 ms for myelinated nerve fibers). The asymptote parallel to the y-axis is the minimal stimulation time, which is the shortest duration of stimulation capable of producing excitation, even when huge currents are applied. The rheobase is useless when comparing the excitability of one nerve with another because only the relative current intensity is meaningful. Furthermore, it is difficult to measure the stimulation time of a current with the intensity of the rheobase because it is an asymptote. Thus, a graphic measurement is made of the time during which a stimulus of double the rheobasic strength must act in order to reach threshold. This time is the chronaxie. Chronaxie values tend to remain constant regardless of geometry of the stimulating electrodes. The shorter the chronaxie, the more excitable the fiber. The chronaxie value for normal myelinated nerve fibers is about 0.7 ms. Some nerve pathologies in humans can be detected early by changes in their chronaxies. Measurement of chronaxie in the laboratory is also valuable because it provides an easy method for measuring the value of the membrane time constant % (see Chapter 24 on cable properties and Chapter 70 for a review of electricity). In brief, the relationship between chronaxie (0.) and time constant (~'m) is tr_ tr = 1 4 4 o ~'m - ln2 - 0.69 "
(1)
Thus, ~m is 1.44 times the value of 0-. Therefore, 0. is analogous to a half-time for a first-order reaction, whose rate constant is the reciprocal of the Tm (k = 1/%). The S-D curve indicates that current pulses of very short duration (e.g., < 0.1 ms) are less effective for stimulation. Thus, sinusoidal alternating current (AC) at frequencies above 10 000 Hz is less capable of stimulation. Another way to view this is that, because the membrane impedance decreases greatly at high frequencies (since the cell membrane is a parallel RC network), the potential difference that can be produced across the membrane by current flow across it (IR or IX drops) is very small. Hence, AC of very high frequency has less tendency to electrocute, and the energy of such currents can be dissipated as heat in body tissues, and thus may be used in diathermy for therapeutic warming of injured tissues.
E. Accommodation 2
. . . . .
|
1
0~ ;, Chronaxio
~Rheobase
Duration of stimulus (ms)
FIGURE 7. Strength-duration curve for AP initiation in excitable membranes. The intensity of rectangular stimulating pulses is plotted against their duration for stimuli that are just sufficient to elicit an AP. The rheobase current and chronaxie (o-) are indicated.
Accommodation, in electrophysiology, refers to the loss of sensitivity of a cell to an applied stimulus. Sensory organs exhibit the property of accommodation, as do many neurons and other excitable membranes, such as skeletal muscle fibers. When a rectangular (square-wave) current pulse is used to depolarize a quiescent motor neuron (step depolarization) from the resting potential ( e . g . , - 7 0 mV) to the
422
SECTION III Membrane Excitability and Ion Channels
threshold potential (Vth) or beyond, the neuron quickly responds with an all-or-none AP. However, if the applied pulse is ramp-shaped (triangular), the neuron may or may not respond, even if the normal Vth is exceeded, depending on the slope of the ramp. If the slope of the ramp is steep, the neuron will respond, but at a higher Vth level (more critical depolarization is required). If the slope is shallow, the neuron will fail to fire an AP, regardless of the level to which it is depolarized. This is accommodation. That is, when the membrane is depolarized gradually, the stimulus is ineffective in producing an AP response (Fig. 8A). This is not true of pacemaker cells (Fig. 8B). Automatic (spontaneously discharging) cells do not exhibit accommodation of their membrane to stimuli capable of evoking AP responses (e.g., nodal pacemaker cells of the heart and some sensory neurons). Such cells will discharge an AP no matter how gradually the membrane is brought to the Vth point. In fact, the pacemaker potential in cardiac nodal cells is of the ramp type, producing depolarization to Vth over a period of about 200-800 ms. The explanation for this phenomenon of accommodation is as follows. As the membrane is slowly depolarized toward Vth, the positive feedback cycle between E m, Na + conductance (gNa) ' and INa begins to operate (beginning at about 80% of the critical depolarization). Therefore, some of the fast Na § channels are turned on (activated), only to inactivate spontaneously (1-gates close) within 1-2 ms. If a critical number (critical mass) of fast Na + channels are not activated simultaneously, then the positive feedback cycle does not become explosive, and a regenerative AP is not produced. That is, the slow depolarization does not allow a critical number of fast Na § channels to be simultaneously in the
A
open conducting state. The channels that opened and spontaneously inactivated cannot be reactivated until they return to the resting state, which requires repolarization to near the resting potential. Hence, they are lost from the pool of available channels. Added to this is the fact that K § channels (delayed rectifier type) will open during the slow depolarization, thus increasing K § conductance (gK)" Whenever gK is increased, excitability becomes depressed, because it tends to repolarize and keep the membrane from depolarizing (to produce an AP); lowering R m also lowers the effectiveness of the depolarizing current. Therefore, accommodation to low-slope stimuli occurs for two reasons: (1) spontaneous inactivation of fast Na § channels that have been activated, and therefore lack of a simultaneously open critical mass and (2) increase in gK, which depresses excitability. The lack of accommodation in automatic cells (e.g., SA nodal cells of the heart) then may be due to (1) less spontaneous ion channel inactivation and (2) less gK increase during the applied ramp stimulus or natural ramp pacemaker potential. Both of these conditions apparently apply to cardiac nodal cells. The inward current responsible for the rising phase of the AP is not a fast Na § current, but rather a slow Ca 2+ current, which inactivates very slowly, and the kinetics of the turn-on of the delayed rectifier K § current is also very slow. As stated previously, accommodation also occurs in sensory organs. For example, some stretch receptors accommodate to a sustained stretch. When the stretch is first applied, there is a burst of APs. But the bursting frequency of discharge gradually slows down and then stops, even though the stretch is maintained. F. Anodal-Break Excitation
e~ (mV) 0
--50 -70
B (mY) 0
--40
Vth
-60
FIGURE 8. (A) The process of accommodation in response to a ramp stimulus in a motor neuron and (B) contrasted with the lack of accommodation in a nodal pacemaker cell from the heart. In A, as the slope of the ramp stimulus is decreased (from a to c), the action potential is delayed (b), and then fails completely (c).Vth, threshold potential.
Excitation occurs on the make (the beginning) of a squarewave depolarizing stimulus. If the applied stimulus duration is very long (relative to the AP duration), repetitive firing of APs will occur if the membrane is of the nonaccommodating type. If the membrane is of the accommodating type, then only the initial AP is produced, because accommodation occurs. If the cathode (negative) and anode (positive) electrodes are placed directly on an isolated single nerve axon (e.g., squid giant axon), then an AP will be triggered at the cathode region on the make of the square-wave stimulus. This happens because, as depicted in Fig. 9A, depolarization occurs under the cathode, whereas hyperpolarization occurs under the anode. However, something unexpected occurs under the anode on the break of the stimulating pulse; namely, an AP is triggered from this hyperpolarized region of the axon (Fig. 9B). The explanation for this is that ion channel changes occur during the hyperpolarization, such that the excitability of that membrane is transiently increased (lower Vth point) immediately following cessation of the applied pulse. The increase in excitability is due to two factors: (1) there is an increase in hoo during the hyperpolarization, reflecting that almost 100% of the fast Na § channels have their I-gates open, and hence are capable of conducting (open state) when the
25. Electrogenesis of Membrane Excitability
423
v
20 fl. c m 2
FIGURE 9. Current flow under the anode and cathode during extracellular stimulation (A) and anodal-break stimulation of an action potential (B). (A) Excitation occurs under the cathode on the make of a current pulse and under the anode at the break of the rectangular pulse. (B) An intracellularly applied hyperpolarizing current pulse produces an anodal-break response. RP, resting potential.
membrane is depolarized on removal of the hyperpolarizing pulse (gNa = max gNa m3h); and (2) there is a decrease in n during the hyperpolarization, hence decreasing gK (gK = max gK n4). These equations are discussed in the following section. The changes in the h and n parameters persist for a short period after termination of the applied pulse, and hence increase membrane excitability during this brief period and trigger an AP. This is called anodal-break excitation and postanodal enhancement of excitability (due to the postanodal depolarization). In contrast, under the cathode, after termination of the applied pulse, an opposite change occurs in E m and excitability. The membrane is hyperpolarized transiently and excitability is depressed. This is known as postcathodal hyperpolarization and postcathodal depression of excitability. In Hodgldn-Huxley terms, this phenomenon also is due to two factors: (1) decrease in h during the depolarization, reflecting a smaller fraction of fast Na + channels having their 1-gates open and hence incapable of conducting and thereby decreasing gNa; and (2) increase in n , hence increasing gK- The increased gK and decreased gNa produce hyperpolarization (refer to Chapter 4), and thereby depress excitability.
!11. E l e c t r o g e n e s i s
FIGURE 10.
Time course for the decrease in membrane resistance
(Rm) to the passage of current measured during the action potential (V) in a nerve a x o n . R m falls from about 1000 fI.cm~ at rest to about 20 12.cm2 near the peak of the AP. (Adapted from Cole, K. S., and Curtis, H. J. (1939). J. Gen. Physiol. 22, 649-687.)
toward varying the concentrations of the different ions bathing the axon. Figure 11 shows a classic experiment in which the concentration of Na § ions bathing the squid axon was altered. It was found that the overshoot and the rate of rise of the AP were proportional to [Na+]o . This result was
Stimulus ~(~
Squid axon
O-
-
'="=="=~~ ~ > ~ E rn
Em2il
2
(mY).2
-40~ -60 -80 C
E 20~-
oh
<my, b :!~
1/~3
i1\\
9
33%
Na +
of A c t i o n P o t e n t i a l t
Early experimentation in electrophysiology focused on determining the mechanism for the generation of the AP. One important finding by Cole and Curtis (1939) was that during the AP, the membrane resistance (but not the capacitance) changed dramatically (Fig. 10). The large reduction in membrane resistance during the AP supported the hypothesis that the AP resulted from a large increase in the ionic permeability of the membrane. To determine which ionic species might be involved in generating the AP, subsequent experimentation was directed
0
J
I
1 2 Time after stimulus (ms)
I
3
FIGURE 1 1. Na§ dependence of the nerve action potential. (A) Diagram depicts the method used to record the transmembrane AP (Era) from a squid axon. (B and C) The voltage traces show the reduction in AP amplitude as the Na+ concentration of the solution bathing the axon was reduced to 50% (B) or to 33% (C) of normal. APs were recorded with 100% of the normal Na+ concentration (traces labeled 1), after reduction of the Na§ concentration (traces 2), and after return to normal Na§ concentration (traces 3). (Adapted from Hodgkin, A. L., and Katz, B. (1949). J. Physiol. (London) 108, 37-77.)
424
SECTION III Membrane Excitability and Ion Channels
the first indirect demonstration that the AP resulted from an increase in the membrane permeability to Na + ions. Several years later, this hypothesis was confirmed directly by the voltage-clamp method, as discussed in the next section.
be electronically subtracted and thus removed by a special procedure. Since Vm is constant during the clamp pulse, dV/dt - 0
and
A. Voltage-Clamp Method
Ic- 0
The membrane current (/m) that generates the AP is composed of ionic current (li) and capacitive current ( / ) according to the relationship (2)
Im=li+Ic
The flow of ionic currents across their respective resistive membrane pathways (or channels) causes a change in the membrane potential (from Ohm's law: V = I R ) . T h e change in membrane voltage causes a capacitive current to flow as (3)
I - Cm dV/dt
where d V / d t is the rate of change of the AP. Substituting Eq. 3 into Eq. 2 yields Im
- I
+ Cm
(4)
dV/clt
Since the membrane potential during an AP is constantly changing, it would be difficult to separate the contributions of these interacting ionic and capacitive components. In addition, the total ionic current is composed of multiple individual currents carried by specific ions. To analyze and separate the membrane currents into their capacitive and ionic components, a revolutionary method, called voltage clamping, was introduced in the early 1950s by Cole and by Hodgkin and Huxley. A diagram of the voltageclamp method is shown in Fig. 12. During a voltage-clamp experiment, the membrane potential is held constant (clamped) by a negative-feedback amplifier, and the amount of current that is necessary to perform this task is recorded. Since the membrane potential (Vm) is held constant, the capacitive current is equal to zero. However, at the very beginning of the depolarizing clamp pulse, there is a large transient capacitive current that occurs, and again at the "off" of the pulse; these can
Therefore,
tm--/i The voltage-clamp experiment gives the magnitude and time course of the ionic currents at a given clamp potential. By clamping the membrane to many different potentials, information about the flow of ionic currents and the underlying conductance changes during the AP is obtained. B. V o l t a g e - C l a m p Analysis
In the voltage-clamp experiments, the various ion currents (such as Na +, Ca 2§ or K § currents) can be isolated from the total ionic current and analyzed individually. For example, in the squid axon experiments, the total ionic current consists of an early inward current followed by a delayed outward current (Fig. 13). By varying [Na+]o , the early inward current can be shown to be carried by Na + ions. Similarly, by changing [K+]o, the delayed outward current can be shown to be carried by K + ions. The Na + and K § currents can also be separated by blocking their pathways through the membrane. Na + channels can be blocked with tetrodotoxin (TTX), derived from the ovaries of Japanese puffer fish, and K + channels can be blocked by several inorganic ions and organic compounds, including tetraethylammonium ions (TEA +) and 4-aminopyridine (4-AP). The current remaining can then be
+20
_70---J
IK I ~ [Na]o or T T X ~
,..
"
L
~
7
/L
J
0 (9
c:
2_
__F J
.-.7_..---
"
u
()
A (Control) . . . . . . . . . . . . . . . . .
() Comm.n~ pulse
E i 0
ial
amplifier (neg. feedback)
.
FIGURE 12. The voltage-clamp method used in a squid giant axon. The two wires inserted into the axon are used to measure membrane potential (V) and to pass current (/). The high-gain negative-feedback amplifier compares the command pulse with the membrane potential, and outputs the amount of current necessary to hold the membrane potential constant (clamped). The magnitude of the feedback current can be measured as the IR voltage drop across a resistor and displayed on a cathode-ray oscilloscope (CRO).
Subtracted
a I 2
I
I 4
I
Time(s) FIGURE 13. The ionic currents that flow when a squid giant axon in seawater is clamped from its resting potential (-70 mV) to a transmembrane potential of +20 mV. Trace A shows the net inward Na+ current (INa) and outward K+ current (IK) in normal medium. Trace B shows the net ionic current when the axon is placed in artificial seawater with most of the Na+ replaced by choline+ (an impermeant cation), so that the intracellular and extracellular Na+ concentrations are equal. This current is due to K+ only. TTX also can be used to block Iia" Trace C shows the difference between curves A and B, which represents INa" (Redrawn from Hodgkin, A. L., and Huxley, A. F. (1952a). J. Physiol. (London) 116, 449.)
25. Electrogenesis of Membrane Excitability
425
subtracted from the total ionic current to reveal the time course for the current that was blocked. The current-voltage relationship is obtained from measurements of the peak inward Na + current and peak outward K + current during a series of voltage clamp steps (Fig. 14). Depolarizing voltage steps to just above the resting potential first produce a small outward current. In this voltage region, the membrane behaves in an ohmic fashion. With greater depolarization, the inward INa is activated and the I-V relationship displays a negative slope or negative resistance region. At potentials above the peak current of the I-V curve, a positive slope is seen, and the current magnitude decreases as ENa is approached, and actually becomes outward at voltages above ENa. The voltage at which the current reverses in direction is the reversal potential. The reason that the current diminishes and reverses at potentials approaching ENa and beyond is that the net electrochemical driving force for Na + ions first becomes smaller and smaller and then outwardly directed, whereas the conductance for Na + ions remains constant and high over this entire voltage range, as indicated by I N a - gNa
(Era- ENa)
(5)
A 100mV
gN a - ' - ~ ~ 25mV 0
B
1
2
3
4
gKl 100mv 50mY 25mV
i
0
1 2 3 4 Time (ms)
FIGURE 15. Voltage dependence and time dependence of the changes in Na + conductance (gNa) and K + conductance (gK) during voltage-clamp of the squid giant axon. The numbers refer to the magnitude of depolarization (in mV) from the resting potential. (A) gNa turns on rapidly and then spontaneously declines over a brief time period. (B) gK turns on more slowly and is sustained in amplitude during the entire clamp pulse. (Redrawn from Hodgkin, 1964).
The outward K + current activates a b o v e - 2 0 mV and increases with depolarization as
IK- grr (Em-EK)
(6)
The voltage-clamp experiments have revealed the most fundamental property of the ionic conductances of excitable membrane: namely, that the conductances are both voltage dependent and time dependent (Fig. 15). Both gNa and g~: activate with depolarization, but with different time courses (Figs. 15 and 16). With time, gNa spontaneously turns off, or inactivates. That is, the INa currents shuts off within 1-2 ms.
Outward
/~
A number of biological toxins that act on specific ion channels have been discovered. For example, TTX mentioned previously has a very high affinity for the fast Na + channel of nerve and some types of muscle cells. It binds to the fast Na + channel and blocks the passage of Na + ions through the channel. Several different types of toxin, including batrachotoxin (BTX), inhibit the inactivation process of the Na + channel, so that the Na + currents are greatly prolonged once activated. Such toxins have proven to be valuable tools in analyzing voltage-clamp currents and understanding ion channel function. The ion channel toxins are discussed in great detail in a Chapter 37.
& Off
IN Vth
RP
'my'
0
Em -10 (mV)
Off
! ,'i
'' t ,
',
70
gNa
_t.
20 F / ~ ~a-" Inact ivation 10t i~tdeactivati~
(mmho/cm a)
N gat've \ resistance~.._~/
/
P~t;r~i;a'
Inward
FIGURE 14.
Current-voltage relationship for the peak early inward current and delayed outward current obtained from a squid axon under voltage-clamp. The inward current is carried by Na § (INa)' the outward current by K + (IK). The reversal potential for INa is the voltage at which the current changes from inward to outward. The region of the INa curve that has a negative slope is known as negative resistance. RP, resting potential; Vth, threshold; ENa, Na + equilibrium potential;/m ' total membrane current; Vm, membrane potential. (Redrawn from Hodgkin, 1964.)
gK (mmho/cma)
10 ~ ~ ~ D e l a y 20
I 0
I 1
I 2
e " d (slow) I 3
1 4
Time
I 5
__l 6
"-..~-slow "-.. deactivation ........ I 7
i 8
i 9
I 10
(ms)
F I G U R E 16. Na + conductance (gNa) and K + conductance (gK) changes are shown in response to voltage-clamp of the squid giant axon from the resting potential to a membrane potential (Em) o f - 1 0 mV. gNa inactivates after a short time, even when the voltage-clamp is maintained; in contrast, gK remains elevated until the clamp is released. When the clamp pulse is terminated early, the gNa increase is quickly turned off (deactivated); the turn-off of grr is considerably slower. (Redrawn from Hodgkin, A. L. (1958). Proc. R. Soc. London (Biol.) B148, 1.)
426
SECTION III M e m b r a n e Excitability and Ion Channels
C. Whole-Cell Voltage-Clamp A new electrophysiological method, the whole-cell voltageclamp technique, has enabled researchers to examine the basis of excitability at the single cell level, for example, for myocardial cells or smooth muscle cells. The method allows the recording of the macroscopic currents that flow through the assembly of ion channels in the cell membrane. The details of this method are given in the following chapter. In brief, to record the whole-cell current, a small-tipped glass pipette, known as a patch pipette, is pressed against the cell membrane, and negative pressure is applied to the interior of the pipette to draw a small patch of membrane into its tip. A high-resistance seal (e.g., 10 l~ f~) spontaneously forms (Fig. 17). Then, strong suction is applied to the patch pipette to blow out the membrane patch and allow the lumen of the pipette to be continuous with the lumen of the cell. This technique allows recording of the whole-cell current, and a complete voltage-clamp analysis can be done for small cells, as described previously for giant axons. This technique also allows some control over the intracellular content of the cell; for example, a substance can be introduced into the cell by diffusion from the patch pipette solution. Whole-cell voltage-clamp records obtained from isolated single uterine smooth muscle cells (18-day pregnant rat) are illustrated in Fig. 18. The condition in Fig. 18 is arranged so that outward K § currents are blocked (by high Cs § concentration in the patch pipette), and any inward currents carried by Ca 2§ or Na § ions can be recorded. As shown in Fig. 18, there are two inward currents: an initial fast current and a later slow current. The initial fast current is carried by Na § ion and this Na § current is blocked by TTX; the later slow current is carried by Ca 2§ ions. The presence of functional fast Na § channels is unusual for most smooth muscles, and
F I G U R E 1 7. The whole-cell voltage-clamp technique in isolated single cells to record the macroscopic current (e.g., lNa' /Ca' or IK)from the entire cell membrane. (Upper left) Cell-attached patch mode, produced by applying light suction to the patch pipette to produce a gigaseal (109 ~'~). This mode is used to record the microscopic currents from only one channel or from a few channels. (Upper right) Whole-cell clamp mode produced by applying strong suction to the patch pipette to blow out the membrane patch and allowing the lumen of the pipette to be continuous with the lumen of the cell. This mode is used to record the macroscopic currents from the entire complement of ion channels in the cell membrane. (Bottom) Preparation of isolated membrane patches. (Reprinted with permission from Hamill, O. P., and Sakmann, B. (1981). Nature 294, 462-464.)
H P = - 9 0 mV
C P = - 5 0 mV - 4 0 mV
-
..............................................
omv
,;
-20 mV t,,,~,........................................................................................................................
- 1 0 mV
0 mY
i
Iiiiiiiiiiiiiiiiiiiiiii111111111111111111111111111111111 i i i 111111111111111111111111161111111111111111111111111111111if
100 pA
20 ms
FIGURE 18.
Two types of inward current recorded from isolated single myometrial cells from pregnant rat using whole-cell voltage-clamp with a patch pipette technique. Depolarizing potential steps (-50 to 0 mV) were applied from a holding potential (HP) o f - 9 0 mV. The arrow indicates the beginning of a voltage step that continues to the end of the trace. The bath contained K+-free solution with 150 mM Na § and 2 mM Ca 2§ The pipette solution contained high Cs § Cell capacitance was 80 pF. CP, command potential. (Reproduced with permission from Ohya, Y., and Sperelakis, N. (1989). Am. J. Physiol. 257, C408-C412.)
25. Electrogenesis of Membrane Excitability
427
it was discovered that such channels developed during pregnancy and reached their maximum close to term. The information gained from such whole-cell voltageclamp experiments, performed on a variety of excitable cell types, has greatly increased our knowledge of the properties and regulation of the many types of ionic currents. The macroscopic or whole-ceU current (/) is the product of the number of functional channels in the cell membrane (N) times the probability that the "average" channel is open during the clamp step (Po) times the single-channel current (i), giving (7)
I = iNPo
The single-channel current (in pA), open probability (Po), and single-channel conductance (3/) are obtained from patchclamp experiments, in which the activity of a single channel (or a few channels) can be monitored in a small patch of membrane (e.g., 1-10 ixm 2) that is electrically isolated from the rest of the cell membrane (cell-attached patch) or excised
A 0mV
,
D. Overview of Action Potential Generation The increase in gNa and resulting increase in inward INa cause the regenerative depolarization of the AP. The depolarization is limited by the approach of the membrane potential toward ENaand by the Na inactivation process. As the membrane is depolarized, both gK and the driving force for I K increase, and the outward I~ repolarizes the membrane.
B
300ms
-8omvJ
(isolated patch). Opening and closing of the gates of the single channel reflect conformational changes in the channel protein. This sophisticated technique, which earned the Nobel Prize in physiology and medicine in 1991 for E. Neher and B. Sakmann, allows the biophysical study of the behavior of a single protein molecule and thus represents electrophysiology at the molecular level. Although this patch-clamp technique and analysis is discussed in detail in the following chapter, a brief illustration of actual records is given here (Fig. 19).
I
300 ms
-80mY j';
~i
iiii
' .....
, '~'
i
V
u
0
"~ . . . . . . . . . . . . . .
o. . . . . .
- ~ 0 -, . . . . . . . . . . . . . . . . .
--30
9 ....
~"
"""
..............
~ , ~ - ~ ~ .
v~"~'~
~-
, ~
"--~-_,,~_-L~ j_m _;--~_-r_ -_-__~ --_-c - - - ' - _ - - : - ~ ---~. . . . . . .
IlpA 30ms
12pA 30ms
/(pA)
f
n=29
|
1
I
I
I
l
I
l
I
(mV v
FIGURE 19. Single-channel currents recorded from a slow (L-type) Ca2+ channel, using the cell-attached patch-clamp technique, in a single cardiomyocyte isolated from a young (3-day-old) embryonic chick heart (ventricular cell). These Ca2§ channels in young embryonic/fetal chick and rat heart cells exhibit a high incidence of long openings. (A) Inward currents were evoked by seven depolarizing voltage step pulses (300-ms duration; 0.5 Hz) to 0 mV from a holding potential (HP) of-80 mV. Pulse protocol is given at the top. The ensemble-averaged current from 29 such pulses is given at the bottom. (B) Similar recordings in another cell illustrating the inward currents recorded at five different command step potentials (-30,-20,-10, 0, and + 10 mV) from a HP of-80 mV. Again note the presence of numerous long openings. Also note that the amplitude of the unitary current became smaller at the higher (more positive) command potentials. The current amplitudes are plotted below in the current-voltage curve (each point being the mean + SE of 5-10 experiments). The data points were fitted by a straight line giving a slope conductance of 26 pS for the single-channel conductance. (Modified from Tohse, N., and Sperelakis, N. (1990). Am. J. Physiol. 259, H639-H642; and Tohse, N., and Sperelakis, N. (1991). Circ. Res. 69, 325-331.)
428
SECTION III M e m b r a n e E x c i t a b i l i t y a n d Ion C h a n n e l s
- +70
ENa
Em
+50
l
Ern
-ao (mV)
(rel)
-50 -70
EK
-90 I
I
0
1
I
I
2
3
Time (s)
. 4
FIGURE 20. The relative conductance for Na+ (gNa) and K + (gK) during an action potential in a nerve fiber. The rising phase of the AP is caused by an increase in gNa"The falling phase of the AP is due to the rise of gK (delayed rectification) and to the decrease in gNa (Na+ inactivation). The hyperpolarizing afterpotential is explained by the fact that gK remains elevated for a short time following repolarization, tending to hold the membrane potential (Em) near the K+ equilibrium potential (E~:). ENa, Na§ equilibrium potential; Ag, change in conductance. (Redrawn from Hodgkin, A. L., and Huxley, A. E (1952). J. Physiol. 117, 500-544.)
The increase in gK is self-limiting; that is, the increase in gK produces repolarization, which, in turn, shuts off the increase in gK" The slow kinetics of the turn-off of gK result in a transient hyperpolarization, the hyperpolarizing afterpotential. During the hyperpolarizing afterpotential, the m e m b r a n e potential is brought closer to E K than at rest. The membrane conductance changes that occur during the AP are shown in Fig. 20. The time course for the ionic currents during the nerve AP is shown in Fig. 21. The total ionic current (li) is sepa-
V 6 =
Passive
exponential
/i..".. \
I: '
/! :.
L - '~ ( E
",. \
9
K - :~K
9 9
, 9
9
. -
.
"
.
.
-
o .
, o
., "i K ,
Na+
-
-
m EK)
oo 9
9
IN. +/K
x.. =
" \
.
.
.
.
.
.
.
.
ENa
.
0
+10 -10
.
(mV)
+30
~g
.
o~
l
R P
'
4
1
: .
Ec I - - ~ :
"T"
~----~ EK
F I G U R E 22. Driving forces for Na + and K + currents during the action potential. The total electrochemical driving force is equal to membrane potential (Em) minus the equilibrium potential for the ion (Ei): (E m - El). As depicted, the driving force for the Na § current decreases during the AP, whereas that for K § increases. Even when C1- is passively distributed, a net driving force for C1- influx (outward C1- current) occurs during the AP.
rated into its two major components, INa and I K. Since I K is slower to activate than INa, the inward INa predominates initially, giving rise to the upstroke of the AR Later, I K dominates, causing a net outward current that repolarizes the membrane, that is, is partly responsible for the downstroke of the AR As stated in Eqs. 5 and 6, the specific ionic currents are a product of the membrane conductance for the ionic species and the electrochemical driving force exerted on the ion. Thus, the driving forces on Na + and K + ions continually change during the time course of the AP, as diagrammed in Fig. 22. At the resting potential, there is a large driving force for Na + to flow into the cell, since (E m - ENa) is large. Conversely, at the peak of the AP, (E m - ENa) is at its lowest value, and the driving force for Na + entry is small. In contrast, the driving force for K + effiux is largest at the peak of the AP, when (E m - E K) is maximal. It is important to remember that ionic current flow depends on both conductance and driving force. There is no net current if only one, but not the other, is present. The biological elements of the excitable m e m b r a n e may be represented in terms of an electrical equivalent circuit model, as shown in Fig. 23. In the circuit model, currents flow through the individual conductance pathways for each type of ion. The conductances for Na + and K + ions are variable and depend on the transmembrance potential and time. Batteries (positive pole directed inwardly for Na + ions and outwardly for K + ions) provide the driving forces for current flow. A passive leak conductance for C1- ions is also included in the model. If the correct values for the elements are incorporated and varied over time, the model circuit will generate an AR
E. Fast Na § Channel Activation F I G U R E 21. Ionic currents that flow during the nerve action potential. The total current (li) is separated into an inward Na + current (INa) and an outward K + current (I~:). I i is the algebraic sum of INa and I K. The appropriate equations for I i,/Na' and I K are given. Also depicted is the fact that a net inward Na + flux occurs during the rising phase of the AP and a net K + effiux occurs during the repolarizing phase. (Redrawn from Hodgkin and Huxley, 1952a.)
During an AP, the increase in gNa is related to the membrane potential E,, in a positive-feedback fashion, or "vicious cycle" (Fig. 24A). That is, a small depolarization leads to an increase in gNa' which allows a larger inward INa, which causes further depolarization. This greater depolarization produces a greater increase in gNa" This positive feedback process is "explosive,"
429
25. Electrogenesis of Membrane Excitability
trast, the I-gate is open at the resting E m and closes slowly on depolarization. The time course of the changes in the m and h variables during a depolarizing clamp step is schematized in Fig. 26. In the Hodgkin-Huxley (1952) analysis, the opening of the A-gate requires simultaneous occupation of three negatively charged sites by three positively charged (m +) particles. The activation variable, m, is the probability of one site being occupied, and m 3 is the probability that all three sites are occupied; therefore,
Outside
OK
9Na
EK
e --L ENa
gCl
(leak)
c~
~(~
let
~
ECI
eT
gNa- gNa m 3 h
Inside FIGURE 23. Hodgkin-Huxley electrical equivalent circuit for the squid giant nerve axon. The K+ conductance (gr~)is in series with K+equilibrium potential (EK) and gNa is in series with ENa. The arrows indicate that gNaand gK vary with voltage and time. The low conductance for C1(ga) was termed the leak conductance. C m, membrane capacitance. (Redrawn with permission from Hodgkin, A. L. (1964). "The Conduction of the Nervous Impulse." Charles C Thomas, Springfield, IL.)
with a sharp trigger point (threshold) resulting from the exponential (positive) relationship between gNa and E m (Fig. 24B). It is this positive-feedback relationship that accounts for the negative resistance (slope) in the current-voltage curve (Fig. 14). As Fig. 24B shows, gNa reaches a maximum (saturates) at positive potentials that gives maximal activation of the population of fast Na § channels. The fast Na + channels (and the slow Ca 2+ channels) have a double gating mechanism: an inactivation gate (I-gate) and an activation gate (A-gate) (Fig. 25). For a channel to be conducting, both the A-gate and the I-gate must be open; if either one is closed, the channel is nonconducting. The A-gate is located somewhere near the middle of the channel; it is not at the outer surface because even TTX does not prevent the movement of this gate, and it is not at the inner surface because proteases perfused internally do not affect it. The A-gate is closed at the resting E m and opens rapidly on depolarization; in con-
AE m
where h is the inactivation variable and gNa is the maximum conductance. Hodgkin and Huxley found that, in squid giant axon, there was an e-fold increase in gNa per 4-6 mV of depolarization. To convert to In from log10, the 2.303 R T / z ~ factor of 58 mV (at 20 ~ becomes 58 mV/2.303, or 25.2 mV. Therefore, they concluded that the voltage-sensitive activation gate must have about 4 to 6 unitary charges (25/6 to 25/4).
17. G a t i n g Current
Since the gating of the ionic channels is voltage dependent, a part of the channel protein contains a charged group or dipole that can sense the electric field across the membrane and move in response to a change in transmembrane voltage. When the gating region moves, it causes a shift in the overall conformation of the channel, which allows it to conduct ions. A gating current (I~ that corresponds to the movement of the charged m + panicles (or rotation of an equivalent dipole) has been measured. The gating current is very small in intensity and is measured by subtracting the linear capacitive current (from a hyperpolarizing clamp step) from the total capacitive current (linear plus nonlinear) that occurs with a depolarizing clamp step beyond threshold. The outward I g precedes the inward INa" Tetrodotoxin does not block/g, although it does block INa" Thus, the gating cur-
B
A
viciouscycle
& 9Na
Peak gNa
efe INa E m
/Na = g N a ( E m
-
(8)
Depolarization
ENa)
FIGURE 24. The positive-feedback relationship between Na+ conductance (gNa) and membrane potential E m, leading to the all-or-none action potential. (A) The increase in gNa allows an increase in the inward Na§ current, INa-" gNa (Em- ENa)' which is depolarizing and so triggers a further increase in gNa"This explosive feedback cycle is caused by the voltage dependency of the gated fast Na+ channels. (B) Plot of gNa versus depolarization, showing the initial exponential (positive) increase in gNa as a function of voltage, followed by saturation with greater depolarization, thus giving a sigmoidal relationship. (Adapted from Hodgkin, A. L. (1964). "The Conduction of the Nerve Impulse." Charles C. Thomas, Springfield, IL.)
SECTION III Membrane Excitability and Ion Channels
430
F I G U R E 2 5 . The three hypothetical states of the fast Na § channel, based on the Hodgkin-Huxley model. In the resting state, the activation gate (A or m) is closed and the inactivation gate (I or h) is open: m = 0, h = 1. Depolarization to the threshold or beyond activates the channel to the active state, the A-gate opening rapidly and the I-gate still being open: m = 1, h = 1. The activated channel spontaneously inactivates to the inactive state because of delayed closure of the I-gate: m = 1, h = 0. The recovery process on repolarization returns the channel from the inactive state to the resting state, thus making the channel again available for reactivation. Na + is depicted as being bound to the outer mouth of the channel and poised for entry down its electrochemical gradient when both gates are open (active state of channel). The reaction between resting state and the active states is reversible, whereas the other reactions may be less reversible. The Ca 2§ slow channels pass through similar states.
rent is a nonlinear outward capacitive current (not an ionic current) obtained during depolarizing clamps that reflects movement of the A-gates from the closed to the open configuration. The linear capacitive current results mainly from the lipid bilayer matrix (see Chapter 70), whereas the nonlinear capacitive current arises from the charge movement associated with the A-gates of the protein channels. Specific charged residues are located along the primary sequence of amino acids that comprise the channel polypeptide,
and they can "sense" the transmembrane electric field and move in response to changes in the electric field. The movement of these voltage-sensing or gating residues initiates a conformational change in the channel structure, which then permits ions to flow through the central pore region of the alpha subunit. The movement of the gating charge produces the small, but measurable, gating current. Gating currents have been recorded from several types of ion channels and provide information concerning the steps leading to channel opening.
G. N a § I n a c t i v a t i o n
h t
t ~mn ~ 1//\\
~ "h
-.,..
1
i I
A gNa
gK
i
:Jr
F I G U R E 2 6 . Sketch of the time course of the changes in the Hodgkin-Huxley m, n, and h factors during a depolarizing voltage-clamp step. The lower traces give the resulting time course for the changes in Na + and K + conductances during the step. Because gNa is proportional to the product of the m and h variables, it returns to about the original level within approximately 2 ms after the clamp step is applied.
The fast INa lasts only for 1-2 ms because of the spontaneous inactivation of the fast Na § channels. That is, the fast Na + channels inactivate quickly, even if the membrane were to remain depolarized (Fig. 26). (In contrast, the slow Ca 2+ channels inactivate slowly.) Inactivation is produced in the fast Na + channels (and in the slow Ca 2+ channels) by the voltage-sensitive slow closing of the inactivation gate (Igate) (see Fig. 25). The I-gate is located near the inner surface of the membrane, as evidenced by the fact that addition of proteolytic enzyme to the inside of a perfused giant axon chops off the I-gate and eliminates inactivation (protease added outside does not have this effect). The I-gate is presumably charged positively to allow it to move with changes in the membrane potential. During depolarization, the inside of the membrane becomes less negative or more positive,
25. Electrogenesis of Membrane Excitability
431
A
C 1.0
1.0
boo
---m ,
hoo 0.5
or 0.5 moo -70
-50 Em
-90 -70 -50 -30 -10
(mY)
B
Em (mY)
D| 1.0
1.0
=
~ +
0
-30
= max
E
0.5
o
0.5
Q
0
-70
-50
, -30
Em (mY)
m
0
=
_= ._~='"
~
-90 -70 -50 -30
'
-10
I
,
10
Membrane potential
FIGURE 27. Voltage inactivation of the fast Na+ channels as a function of the membrane potential (Em). (A) hoo plotted against Em, where hoois the inactivation factor of Hodgkin-Huxley. The hoorepresents h at infinite time or steady state. This graph illustrates that fast Na§ channels begin to inactivate at about -75 mV, and nearly complete inactivation occurs at about -30 mV. (B) Maximal rate of rise of the AP (max dV/dt) as a function of resting E m. Max dV/dt is a measure of the inward current intensity, which is dependent on the number of channels available for activation. Therefore, max dV/dt decreases as hoodecreases. (C) Plot of both steady-state inactivation (hoo)and activation (moo)against E m to illustrate the overlap of the two curves, depicting the window current region. (D) Steady-state activation and inactivation curves for the fast Na§ current [INa(f~] recorded using the whole-cell voltage-clamp technique from uterine smooth muscle cells isolated from late pregnant (18-day) rats. For the inactivation curve, conditioning pulses of various amplitudes were applied for 2 s before a test pulse to 0 mV. The activation curve was obtained by measuring peak amplitude of INa elicited by various command potentials from a HP of-90 mV. The two curves were obtained by fitting the data to Boltzmann distributions. Each point represents the mean + SE. Note the overlap between the activation and inactivation curves, resulting in a steady-state window current between about -50 and -30 mV. (Data from Y. Inoue and N. Sperelakis, unpublished observations.)
and this causes the I-gate to close. At the normal resting potential, the I-gate is open, but the A-gate is closed. The voltage dependency of inactivation is given by the h a versus E m curve (Fig. 27). The inactivation variable h (probability function) varies between 0 and 1.0, perhaps reflecting occupation of a negatively charged site by a positively charged h particle; ho~ is the value of h at infinite time (> 10 ms) or at steady state. When h = 1.0, the I-gates of all of the fast Na § channels are in the open configuration; conversely, when h = 0, all the 1-gates are closed. Since gNa at any time is equal to the maximal value (gNa) times m3h (see Eq. 8), when h = 0, gNa = 0. When h = 1.0, gNa = gNa (if m = 1.0). At the normal resting potential, h a is nearly 1.0 and it diminishes with depolarization, becoming zero at a b o u t - 3 0 mV. The maximal rate of rise of the AP (max dV/dt) is directly proportional to the net inward INa' which is directly proportional to gNa' the decrease in h a is the cause of decrease in gNa" At about - 3 0 mV, h a = 0 and there is complete inactivation of the fast Na § channels. Therefore, depolarization by any means (e.g., elevated [K+]o or applied depolarizing current pulses) decreases gNa and excitability disappears at a b o u t - 5 0 mV (Fig. 27B). The AP disappears at - 5 0 mV (rather than - 3 0 mV) because of a minimum current density requirement for a regenerative and propagating response.
A plot of the inactivation curve along with the activation curve is given in Fig. 27C to illustrate the overlap in these curves, giving rise to a so-called w i n d o w c u r r e n t (or steady-state inward current) over a certain voltage region. The presence of the window current means that, over the voltage range of a b o u t - 5 0 to - 3 0 mV, there is a steady inward Na + current passing through about 10-20% of the fast Na + channels that are always open (on a rotating population basis). The slow Ca 2+ channels (Fig. 28B) behave much the same way as the fast Na + channels (Fig. 28A) with respect to activation and inactivation, with one main difference being the voltage range over which the slow channels operate. For example, inactivation occurs between - 5 0 and 0 mV for the slow channels, compared with b e t w e e n - 1 0 0 a n d - 3 0 mV for the fast Na + channels. Another major difference is that slow channels inactivate much more slowly than the fast channels; that is, they have a long inactivation time constant (rinact). In myocardial cells, the h variable for the slow channel is referred to as the f variable, and the m variable as the d variable and the exponent is 2.0: g c a - gca d 2 f
(9)
The amino acid sequences of Na + and Ca 2+ channels are known, as well as their putative tertiary structure (Fig. 29).
432
SECTION III Membrane Excitability and Ion Channels
F I G U R E 28. (A) Illustration of a fast Na + channel and (B) a slow C a 2+ channel. The ionic channels are proteins that float in the lipid bilayer matrix of the cell membrane. These voltage-dependent channels have two gates, as depicted: the activation gate (A or m) and the inactivation gate (I or h). Conformational changes in the protein could serve as the gates. The 1-gate is located near the inner surface of the membrane because proteolytic enzymes added internally destroy this gate. The A-gate is located somewhere in the middle of the channel; TTX does not prevent movement of this gate. TTX binds to the outer mouth of the channel and plugs it physically, as depicted. The gates are presumably charged positively, so that on depolarization--inside going less negative (i.e., in a positive direction)--the gates are electrostatically repelled outward. The A-gate opens relatively quickly (e.g., 0.25 ms), whereas the I-gate closes relatively slowly (e.g., inactivation time constant of 1-2 ms). Therefore, for a short period, both gates are open and the channel is in the conducting mode with Na + entering the cell down its electrochemical gradient. However, when the I-gate closes (after about 2 ms), the channel again becomes nonconducting. During recovery (reactivation) upon repolarization, the gates return back to their original resting positions. The slow Ca 2+ channels (B) behave similarly to the fast channels (A), except that their gates apparently move more slowly on a population basis. The slow-channel conductance activates, inactivates, and recovers more slowly. The slow-channel gates operate over a different voltage range than the fast channels (i.e., less negative, more depolarized). TTX does not block the slow channels. Drugs such as nifedipine block the slow channels (but not the fast Na+ channels) by binding to the channel and somehow inhibiting it. In addition, the voltage inactivation curve of the Ca 2§ slow channels is shifted to the right, so that inactivation begins at about -45 mV and is not complete until about-5 mV. The slow channels also have a higher activation (threshold) potential of about-35 mV (compared with about-55 mV for the fast Na+ channel). The activation gate variable is known as d, and the inactivation gate variable as f.
Na + and Ca 2+ channels consist of several subunits, one of which contains the water-filled pore through which the ions pass. Ion channels have one or more sites that can be phosphorylated, and phosphorylation alters their behavior. For example, in cardiac muscle, the slow Ca 2+ channel activity is increased by adrenaline, a hormone that increases the cyclic A M P level and leads to Ca 2+ channel phosphorylation (see Chapter 33). The molecular structure of ion channels is discussed in more detail in a later chapter. H. R e c o v e r y Any Na + or Ca 2+ channel that has been activated and then spontaneously inactivated must go through a recovery process before it can return to the resting state from which it can be reactivated (see Fig. 25). The recovery process is dependent on voltage and time. The membrane must be repolarized beyond a b o u t - 5 0 m V before the recovery process can begin (i.e., traveling up the h a v e r s u s E m curve). At any given Em, time is necessary for the recovery process to occur, namely, the time required for the charged A-gates and 1-gates to move back to their resting configuration (A-gate closed, I-gate open) with the electric field. The recovery process is rapid for fast Na + channels (e.g., 1-10 ms) and less rapid for the slow Ca 2§ channels. The recovery process of the slow channels is slowed by organic calcium antagonist drugs.
!. K § A c t i v a t i o n The K + channel (outward-going delayed rectifier) is generally believed to have only an A-gate, because it does not inactivate. This gate is thought to be located near the inner surface of the membrane, because TEA + blocks the K + channel more readily from the inner surface. The block is use dependent or frequency dependent: as the A-gate opens, the TEA + molecule can bind in the channel behind the gate. The A-gate is believed to be positively charged, and depolarization (inside going positive) opens the gate. In the Hodgkin-Huxley analysis of squid giant axon, the A-gate opens when four positively charged n + particles simultaneously occupy four negatively charged sites. If n is the probability that one site is occupied, then n 4 is the probability that all four sites are occupied; therefore, g K - ~ K n4
(10)
The power to which n is raised varies in different tissues.
J. M o d e l for A c t i v a t i o n a n d I n a c t i v a t i o n of N a § a n d K § C h a n n e l s As stated previously, in the H o d g k i n - H u x l e y analysis, the Na + conductance gNa is controlled by three charged activating m particles and one blocking h particle, which move with changes in the electric field across the m e m b r a n e to ei-
25. Electrogenesisof Membrane Excitability
433
A
B
I
-
! i I I
t'l/~ Outside U~ j " u
'i
Inside
_
CO 2
I I
II
+H3N.~
| D
C Ca 2+
I
I
I
I
tZ1 t~l
II
+
F I G U R E 29. Models of the fast Na + channel and the slow Ca 2+ channel proteins based on structural data. (A) The Na + channel has multiple protein subunits, labeled a,/31, and f12"The a subunit is the one that contains the water-filled pore through which Na + passes. One site that can be phosphorylated by cyclic AMP--dependent protein kinase (PK) is present on the ce subunit. The /32 subunit has a disulfide bond (SS). (B) The structure of the central pore-forming a subunit for the Na + channel. There are four homologous intramembrane polypeptide repeat domains, connected by intracellular polypeptide segments or loops. The four domains are arranged into a circular structure within the plane of the membrane to form a channel. Each domain consists of six units that span the membrane, as depicted. Both the carboxy (COO-) and amino (NH~) termini are intracellular. (C) The Ca 2+ channel also is composed of several subunits: eel, a2, fl, y, and t~. The a 1 subunit contains the pathway by which Ca 2+ ions traverse the channel. Two sites can be phosphorylated by the cAMP-PK, one on the a 1 subunit and the other on the/3 subunit. As depicted, the/3 and t~ subunits do not span the lipid bilayer, and the t~ subunit possesses a disulfide bond. Ribosylation sites are present on the outer surfaces of the subunits. (D) The polypeptide structure of the a 1 subunit of the slow Ca2+ channel is similar to that for the fast Na + channel, with four repeat membrane-spanning domains connected by intracellular and extracellular loops. (Reprinted with permission from Catterall, W. (1988). Science 242, 50-61.)
ther occupy or unoccupy certain sites on the Na + channel protein. If m is the probability of one favorable site being occupied, then m • m • m or m 3 is the probability that all three sites are occupied simultaneously. The activation Agate cannot be fully opened unless all three sites are occupied. A simplistic variant is to consider that the A-gate of the Na + channel actually consists of three separate subgates or subdoors, as illustrated in Fig. 30. For a Na + ion to pass through the channel, all three subdoors must be open. The value of h is the probability of one favorable site being occupied that opens the I-gate, and this site is already occupied at the normal resting potential (I-gate open), and becomes unoccupied with depolarization. Therefore, g N a - gNa m3h
(8)
where gNa is the m a x i m u m gNa possible (primarily a function of density of channels), m is the activation variable, and h is the inactivation variable.
The K + conductance gK is controlled by four charged activating n particles, which move with depolarization to occupy four sites on the K + channel protein. If n is the probability of one favorable site being occupied, then n • n • n X n or n 4 is the probability that all four sites will be occupied simultaneously. There is no inactivation variable for the K + channel, because gK does not exhibit a major decrease during a short depolarizing voltage-clamp pulse (e.g., 20 ms). Therefore, g K - ~ K n4
(10)
where gK is the m a x i m u m gK and n is the activation variable. Based on the preceding analysis, Hodgkin and Huxley (1952a, b, c) then gave the differential equations that govern the values of n, m, and h using a pair of rate constants, c~ and /3, for the forward reaction and reverse reaction, respectively. The subscripts for the ce and/3 rate constants identify which variable they pertain to (i.e., n, m, or h). The ce a n d / 3 rate constants depend only on membrane potential (at constant
434
SECTION III Membrane Excitability and Ion Channels
F I G U R E 30. Diagram depicting the modified Hodgkin-Huxley view of the three states of the fast Na + channel. As shown, there is evidence for the existence of three closed states of the channel (C 3, C 2, and C1). The activation gate (A-gate) is depicted as consisting of three parts, each one being controlled by one m+ particle. For the A-gate to be open to allow Na § ions to pass, all three subgates must be open, and therefore all m sites must be occupied. The C 3 closed state is depicted as having all three subgates closed, the C 2 closed state with two subgates closed (one open), and the C 1 closed state with one subgate closed (two open). For the open state (O), of course, all three subgates must be open. P, plugged channel; I, inactive.
dn
temperature and [Ca2+]o ). A schematic model is given in Fig. 31, and the corresponding equations are dn dt = a ( 1
- n ) - flnn
dm dt - O~m(1 - m ) - tim m dh dt = a h ( 1 - h ) - f l h h
(ll)
(12)
(13)
where n is the fraction of sites occupied, and therefore (1 n) is the fraction of sites unoccupied. Therefore, rate of o c c u p a t i o n - a (1 - n)
(14)
rate of unoccupation = / 3 n
(~5)
rate of change of n - rate of occupation - rate of unoccupation
or
(16)
d---7 = cen(1 - n ) -
~n n
(11)
The same analysis can be applied to the m and h factors. The effect of making the inside of the fiber more positive (i.e., depolarizing) is to increase a , a m, and/3 h and to decrease/3 n,/3 m, a n d a h. One must remember that the h particles occupy the favorable site at the resting potential (1-gate open), and this site becomes unoccupied with depolarization (I-gate closes). This is opposite to the situation for the m and n particles (and A-gate). In voltage-clamp experiments, at a given clamp step, the membrane potential is fixed and the differential equations (Eqs. 11-13) lead to exponential expressions for n, m, and h, and the K + and Na + conductances can then be calculated. When this was done, there was a good fit of the calculated curves with the experimental data points for the changes in gNa and gK with time at various clamp steps (see Fig. 15). K. M e c h a n i s m s
of R e p o l a r i z a t i o n
The AP is terminated primarily by the turn-on of gK, the d e l a y e d rectification. It is called the delayed rectifier be-
25. Electrogenesis
of Membrane
435
Excitability
p,~te o4 occupation
Unoccupi / ~ ~ @ s i t eOccupi s i t eesd@ s ed (depol. inc.)
(depol. dec.)
Rate of unoccupat~o~ (depol. dec.)
higher than PK, the Pcl/PK ratio being about 3-7. However, as discussed in Chapter 14, the C1- ion is passively distributed, or nearly so, and thus cannot determine the resting potential under steady-state conditions. However, net C1movements inward (hyperpolafizing) or outward (depolarizing) can and do affect E m transiently until reequilibration occurs. There is no net electrochemical driving force for C1current (Icl) at the resting potential, since
Ec1
(17a)
Eel ) = 0
(17b)
E m -
and (E m -
(depol. inc.)
However, during AP depolarization, there is a larger and larger driving force for outward Icl (i.e., C1- influx), since
(depol. inc.)
Ic1 = gc1 ( E m
(depol. dec.) F I G U R E 31. Schematic diagram based on the Hodgkin-Huxley model for activation and inactivation of Na + channels and K + channels. a an(]/3 are the rate constants for occupation and unoccupation, respectively, of sites on the channel protein by the m +, h +, and n + particles. See text for further explanation.
cause its turn-on is slower and delayed with respect to the turn-on of the fast Na § channels. The increase in gK acts to bring E m toward E~: (about-90 mV), since the membrane potential at any time is determined mainly by the ratio of gNa/gK (see Chapter 14). This type of gK channel is activated by depolarization and turned off by repolarization. Therefore, this gK channel is self-limiting, in that it turns itself off as the membrane is repolarized by its action. In addition to the gK turn-on, there is also some turn-off of gNa' which contributes to repolarization. Two reasons for gNa turn-Off are (1) spontaneous inactivation of fast Na § channels that had been activated, that is, closing of their 1-gate (inactivation z of 1-3 ms); and (2) a reversible shifting of activated channels directly back to the resting state because of the rapid repolarization occurring due to the g~ mechanism. Theoretically, it would be possible to have an AP that would repolarize (but slowly) even if there were no gK mechanism, because the gNa channels would spontaneously inactivate, and so the gNa/gI~ ratio and E m would slowly be restored to their original resting values. Turn-on of gK acts to sharpen repolarization of the AP and allows higher frequency of impulses.
L. Skeletal Muscle Repolarization The two mechanisms for sharp repolarization of the AP discussed previously for neurons also apply to skeletal muscle. But there is an important third factor involved in repolarization of the skeletal muscle AP, the C1- current. The C1permeability (Pcl) and conductance (gel) are very high in skeletal muscle. In fact, Pc1 of the surface membrane is much
-
Ec1)
(18)
where Eel is the C1- equilibrium potential calculated from the Nernst equation (see Fig. 22). In other words, the large electric field that was keeping C1- out (i.e., [C1-]i C1- > I- > F- is consistent with a weak field strength binding site. I- blocks the multi-ion pore which is also permeable for HCO~. The channel is activated by a cAMPmediated stimulation of PKA and in the intestine also by a cGMP-mediated stimulation of PKG. The R-domain has 10 PKA phosphorylation sites, between $660 and $813. In human tissue, nine sites lie in a monobasic RxS sequence. In all other species these sites present a dibasic motif RRxS. Two serines belong to PKC consensus sites ($686 and $790, Fig. 11B). Gating only proceeds after several phosphorylation steps in the R-domain, which then allows ATP binding to the nucleotide-binding fold N 2. This step precedes channel opening. Gating is then further regulated by a cycle of ATP hydrolysis at N1 and N2 when R is highly phosphorylated. Dephosphorylation by protein phosphatases returns the channel into the closed form (for gating see Fig. 12). CPTR, which is much more than a C1- channel, regulates other ion channels and transporters, indicating that the name CFT-regulator was chosen with much foresight. These regulatory properties might even be more important than its function as a C1- channel. The epithelial Na + channel ENaC is inhibited by CFFR. Its activity is up-regulated when CPTR is not expressed in the plasmalemma. ORCC is normally activated by CFTR and is defective in cystic fibrosis. CaCC and the volumeregulated anion channel VRAC (discussed later) are downregulated in cells that coexpress CFTR. Basolateral ROMK
499 channels are activated by CFTR. The KvLQT-1 channel that has recently been detected in epithelial cells probably is a complex of IsK (minK) and eag channels (Chapter 27), and seems to be activated by CFTR and is therefore down-regulated in the absence of CFTR. CFFR also regulates water channels (aquaporins) and several transporters such as the Na§ § exchanger and probably the members of the AE (anion exchanger) family. Importantly, CFTR is involved in the cAMP-stimulated release of intracellular ATP, but the exact mechanism is unknown. This CFFR-mediated ATP release is probably responsible for the functional coupling between CPTR and ORCC, since the latter is modulated by ATP agonists. Some proteins which are important for exocytosis either inhibit (e.g., syntaxin 1A) or stimulate CFTR (MUNC 18). CFTR also affects intracellular processes, such as the traffic of exocytotic vesicles and the acidification of the trans-Golgi network, which is related to sialylation of secretory proteins. The accumulation of asialo moieties can create a proinflammatory environment attracting neutrophils and enhancing the binding of Pseudomonas aeruginosa and other bacteria. Potential mechanisms of interaction with other proteins might involve a direct protein-protein interaction via the PDZ-binding domain located at the C-terminus of CFTR. The last for AA residues 1477-1480 (DTRL) provide the PDZ-binding motif, which seems essential for at least the functional interaction with the Na+-H § exchanger (NHE) regulator protein.
FIGURE 12. Gating scheme of the cAMP-activated CFTR C1- channel. Shown are only the N- and C-terminus nucleotidebinding folds, N 1 and N2. If R is highly phosphorylated, ATP binding and hydrolysis at both sites gate the channel. If N1 binds ATP and N2 binds ADP, the channel is still closed. Hydrolysis of ATP at N 1 induces short channel openings. Openings are stabilized when ADP at N2 is exchanged by ATP. Hydrolysis and ADP unbinding at both sites close the channel. The channel is also closed when phosphorylation of R is terminated by phosphatases.
500
SECTION III M e m b r a n e E x c i t a b i l i t y and Ion Channels
Various CFTR mutants lead to loss-of-functions which are phenotypically coupled to mucus abnormalities, defects in electrolyte transport, salty sweat, ileus, cilia dyskinesia in the airway epithelium, dilated bronchi, bronchiectasis, biliary cirrhosis, gallstones, pancreas insufficiency, severe malnutrition, and impairment of HCO 3 transport. The most important mutant types are arranged in classes. Class 1 comprises defective translation mutants that all constitute truncated forms, and stop mutations in N1. Class 2 contains proteins which are incompletely glycosylated and cannot maturate. They are retained in the endoplasmic reticulum. The most important mutant, a deletion of phenylalanine at position 508 in N1, AF508, accounts for approximately 80% of the CF patients. Class 3 mutations interfere with ATP binding and hydrolysis and thus influence gating of the CFTR channel: Class 4 mutants represent mutations in the pore, causing functional channel defects due to changes in permeation. Typical mutations of this kind are R117H, R334W, and R234E Class 5 summarizes mutations that do not cause CF, but CBAVD, the congenital bilateral absence of vas deferens. The exceptional not yet understood defects in regulator functions of CFTR are comprised in class 6.
5. Volume-Regulated Anion Channels (VRAC) Anion channels that open in response to an increase in cell volume are abundantly expressed in most excitable and nonexcitable cells. The current through this volume-regulated anion channel (VRAC) is mainly carried by CI-, but
~ ~ [ j _
0
1 0
d
/
0
\
the channel is also permeable for amino acids and organic osmolytes. The effiux of C1- and organic osmolytes through VRAC is functionally important for the regulation of cell volume. In addition, VRAC also contributes to the control of membrane potential and the driving force for Ca 2+ influx, vectorial transport of solutes, and intracellular pH homeostasis. More speculatively, VRAC could assist in regulating cell growth and the transition from proliferation to differentiation. VRAC can also be activated by mechanical forces (shear stress, biaxial tensile stress) which induce changes in cell shape and possibly fold and unfold the plasma membrane. This function might be especially important for cells that sense mechanical forces, such as endothelial cells. Currents through VRAC are outwardly rectifying, slowly inactivate at potentials positive to +60 mV, and show a voltagedependent recovery from inactivation (Fig. 13). The permeability sequence of VRAC for anions is SCN-> I - > NO 3 > B r - > CI-> HCO 3 > F - > gluconate > glycine > taurine > lactate > aspartate, glutamate. The sequence for amino acids is consistent with that for the rates at which these amino acids are lost from cells during hypo-osmotic swelling. The pore diameter estimated from the fit of the relative permeabilities of these anions to their Stoke's diameter was approximately 11 /k. The single-channel conductance is approximately 40-50 pS at positive potentials and 10-20 pS at negative potentials. Rectification is therefore a property of the open channel. It is not known how mechanostimulation or cell swelling affect gating of VRAC. The annexin II-pl 1 complex, which is part of the cortical cytoskeleton and is involved in the formation of caveolae, seems to be important for activation of VRAC in
"2
II
--,~-.~_-~ cont (1) III I ---1 "
O
0
\ o\ o
/ oO oo
< 500 ms
2
2 rain
- I [nA] ~ hypo (2)
[
2 ~
_
.
,
. . . . . . . . .
,
-
.
-
I
I
!
. ]
2-___1
2
-100 L
I
0 . . . . . . ~'~
hyper (3)
V [m V] -2 FIGURE 13. Activation of VRAC by hypotonic solutions. (A) Time course of the activation and deactivation of currents through VRAC (measured at -80 mV) in an endothelial cell from cultured pulmonary artery. The solid bar indicates the application of a solution which, with a 25% lower osmolality than the control solution, induces cell swelling. (B) I-V curves obtained at the time-points 1,2 indicated in (A). (C) Current traces under isotonic, hypotonic, or hypertonic conditions in response to a stepvoltage protocol. Holding potential is 0 mV, 2-s steps range from -80 to + 100 mV and are 20 mV spaced.
29. Ion Channels in N o n e x c i t a b l e Cells
endothelial cells. The major protein in caveolae is the 178 AA caveolin-1 which is linked to many other signaling proteins such as tyrosine kinase, small GTPase (e.g., RhoA), Ca 2+ pump proteins of the exocytotic machinery, and in endothelial cells NO synthase. The current density of VRAC correlates with the caveolin-1 content in various cell types. In cells with little caveolin-1, such as FRT and CaCo2 cells, VRAC has a very low current density. VRAC can also be activated under isovolumetric conditions. A decrease in intracellular ionic strength directly activates VRAC under conditions of constant osmolality and C1- concentration, a possible interpretation being an influence of ionic strength on protein tyrosine kinases (PTK). Intracellular dialysis with GTPTS also activates VRAC without any increase in cell volume, an effect probably mediated via small G-proteins. A possible candidate is the GTP-hydrolyzing RhoA-GTPase, since VRAC is down-regulated in cells pretreated with the Clostridium limosum exo-enzyme C3, which inactivates p21RhoA, and is enhanced by CNF1 (cytotoxic necrosis factor), a bacterial toxin that constitutively activates Rho GTPases. Inhibition of the RhoA-associated protein kinase (ROCK), a possible downstream target of p21 RhoA, with the specific blocker Y-27632, also down- regulates VRAC. A tyrosine phosphorylation step downstream from the Rhosignaling cascade may be a critical event in the swellinginduced activation of the channel, since inhibitors of protein tyrosine kinases inhibit VRAC, whereas the protein tyrosine phosphatase (PTP) inhibitors Na3VO 4 and dephosphatin potentiate it. It has recently been shown that the tyrosine kinase p561ck mediates activation of the swelling-induced C1- currents in lymphocytes. A scheme summarizing the possible steps involved in VRAC-gating is shown in Fig. 14. The molecular nature of this elusive VRAC channel is still a matter of debate. P-glycoprotein, the product of the multidrug resistance 1 gene which functions as a drug transporter and possibly as an anion channel, has been proposed as a putative candidate for VRAC, but now even its role as a regulator of VRAC is questioned. A second candidate for VRAC was PIcl n, where n refers to a putative extracellular nucleotide blocking site. Since the primary amino acid sequence of PIcl n lacks transmembrane helices, it has been proposed that the PIc~n channel is a homodimer consisting of four/3 strands which form an eight-stranded, antiparallel/3barrel transmembrane pore similar to that of porins. The most compelling evidence for PIcl n being a plasma membrane channel was the presence of a GXGXG motif, being an essential component of the extracellular nucleotide binding site, which was believed to be essential for the nucleotide sensitivity of the VRAC current. It was reported that mutations of this site (G54A;G56A;G58A) prevent current inhibition by extracellular nucleotides, but these data could not be reproduced, which undermine the idea that Plc1n is VRAC. Recent evidence has also shown that Plcl n is a cytosolic rather than membrane protein that probably binds spliceosomes and regulates protein import into the nucleus. C1C-3, a membrane protein of the C1C family, is currently considered as a likely candidate for VRAC. Within the same family, C1C-2 is obviously volume-sensitive but its biophysical and pharmacological properties are completely different from those of VRAC. Also another candidate, the 72-aminoacid intrinsic membrane protein phospholemman with a single membrane-spanning domain is obviously not VRAC.
501 GT P,,/S shear?
C3
Ionic strength (ri) Cell swelling
tyrphostins, genistein ....
l
::
~ p21RhoA
~,
o CNF1
,I, Rho-kinase
Stimulation ::
~ Inhibition
F I G U R E 14. Gating Model of VRAC channels. Changes in intracellular ionic strength, F i (likely due to water influx into the cell) or cell swelling activate VRAC probably via tyrosine protein kinase(s), PTK. VRAC can be blocked by diverse PTK inhibitors and can be potentiated-when activated---by inhibitors of protein tyrosine phosphatases (PTP). A second pathway might be involved during isovolumetric activation of VRAC. GTPyS and probably also shear stress (in endothelial cells) can activate a small G-protein, RhoA, which responds to downstream targets such as Rho-kinases. This pathway has been elucidated by use of specific activators or inhibitors of RhoA (CNF1, C3, respectively) and Rho-kinase (Y-27632). Downstream targets of Rho-kinase might involve myosin light chains. Rho-kinase inhibits MLC phosphatases, resulting in an increased level of phosphorylated MLC. This might be a step in VRAC activation. Inhibitors of myosin light chain kinase (MLCK), the specific peptide AV25, and ML-9 also inhibit VRAC.
!Ii. Functional Role of Ion Channels in N o n e x c i t a b l e Cells A. Ion Channels and Cell Volume Regulation Most if not all nonexcitable cells regulate their volume by activating either a regulatory volume increase (RVI) in shrunken cells or a regulatory volume decrease (RVD) in swollen cells. RVI is associated with an accumulation of [Na+]i and [C1-]i and a concomitant entry of water to restore the original cell volume. RVD is induced by a cellular loss of K § CI-, and water. Cell swelling opens stretch-activated channels (SACs), either K+-selective or non-selective cation channels. K+-selective SACs with a single-channel conductance between 30 to 50 pS are activated by stretch between 0 a n d - 5 0 cm H20 in most cells (e.g., in the basolateral renal proximal tubule), and induce a loss of intracellular K + which is necessary for RVD. Nonselective cation SACs are permeable for Ca 2+ and provide a pathway for Ca 2+ influx, which increases [Ca2+]i and opens CaZ+-activated K + channels. K + effiux associated with C1- efflux induces RVD. Quinine and quinidine block these K + channels and inhibit RVD. Both mechanisms for RVD mediated by SACs are schematically depicted in Fig. 15. It is important to emphasize that RVD depends on coactivation of channels carrying ions of opposite charge and with a different equilibrium potential. This is illustrated in Fig. 16 for an epithelial cell from toad distal nephron in which cell swelling coactivates K + and C1- currents. Under isotonic conditions, cells are very fight and only a small current, Iiso' is
502
A
SECTION III Membrane Excitability and Ion Channels
K§
SAC
B
NSC - SAC
K§
Cl, osmolytes H20
H20 VRAC
VRAC
KCa
F I G U R E 15. Involvement of ion channels in volume regulation. (A) Swelling of an epithelial cell activates K§ stretch-activated ion channels (SACs) and possibly also CI- and osmolyte permeable volumeregulated anion channels (VRAC). Both events decrease the osmolality of the cell followed by outflow of water and decrease of the cell volume. (B) Cell swelling activates a Ca2§ nonselective SAC and VRAC. Ca 2+ influx activates K + channels, likely BKca, IKca, SKca. Both events induce loss of osmolytes and decrease the cell volume.
A
1.0 ,r i [nA] Ihypo
present (Fig. 16A). Swelling in hypotonic solution activates a large current (I,nypo) which represents the sum of a C1- and a K § current. The C1- component can be inhibited by the C1- channel blocker NPPB (a nitro-phenylpropylamino-benzoic acid derivative), the K § component by quinine (Fig.16B). The current in the presence of either blocker represents the current through the other channel, which reverses at its corresponding equilibrium potential. The cell depolarizes during swelling to the zero-current potential between E K and Ecr There is no net current flow at this potential, but equal inward C1-and outward K § currents, and thus a C1- and K § osmolyte efflux. Block of either current component shifts the membrane potential towards the equilibrium potential of the other ion and thus prevents net efflux of both ions, which impedes RVD (Figs. 16C and D). It is only the coactivation of both currents that can induce RVD. Since VRAC is permeable for non-ionic osmolytes it also provides a voltage-independent pathway for efflux of osmolytes and RVD. RVD and RVI do not only depend on activation of ion channels. The functional cooperation between volume-sen-
C Ihypo
Icl (QUIN)
0.5
block Ic~
~
Ic=
IK (NPPB) V [mV]
liso
-100
J
IK
100
B
block IK
D
VM [mV] I
V
1I
-0.5 t
100
hypo I
NPPB~ QUIN~
Ecj
-50
-100 I 0
,
I 2
,
I 4
,
I 6
EK
block of gK or gcJ
,
NO RVD
co-activation gK and gc=
RVD
t [min]
F I G U R E 16. Coactivation of C1- and K § channels is necessary for regulatory-volume decrease (RVD). (A) A distal nephron epithelial cell (A6 cell from toad) is challenged with a 40% hypotonic solution. Swelling induces a large current (see change from /iso to lhypo)" This current can be partially blocked by quinine (500 IxM). The remaining current is a C1- current (Icl, QUIN, reversal potential close to Ecl= -15 mV). The other component of the current can be blocked by NPPB (100 IxM) and is a K + current (In, NPPB, reversal potential close to E n = -85 mV). (B) Changes in membrane potential during cell swelling and after application of NPPB and quinine (for further explanation see text). (C) Schematic presentation of the swellings-activated currents. Block of one component will shift the membrane potential toward the reversal potential of the remaining component. Zero-current represents (thick arrow) inward C1-and outward K + currents, thus efflux of both osmolytes. (D) Block of one conductance will block RVD because of shifting the membrane potential to the reversal potential of the remaining component and thus inhibition of osmolyte efflux. Only co-activation of two currents can induce RVD.
29. Ion Channels in Nonexcitable Cells sitive ion channels, cotransporters, and antiporters (e.g., Na+-H + and C1--HCO 3 antiporters, Na+-K+-2C1 - and K+-C1 cotransporters) in the regulation of cell volume will be discussed in Chapter 21.
B. Role of Ion Channels in Vectorial Transport Many functions of nonexcitable cells are associated with vectorial secretion and reabsorption. These functions are widespread in the body and only a few examples will be discussed here. Na + r e a b s o r p t i o n is a well-understood mechanism for vectorial transport that occurs in renal cells, intestine, colon, frog skin, and various other cell types. Ion channels and transporters are anchored by the cytoskeleton to strategically important sites in these cells. The E N a C channel is confined to the outer, apical, or luminal face, whereas Na+-K + pumps and K + channels are localized at the inner, serosal, basolateral, or blood side (Fig. 17). The apical Na + channels are always open and provide an influx pathway for Na + into the cell. The Na + ions are extruded at the basolateral side by the Na+,K+-ATPase, which also transports K + into the cell that
F I G U R E 17. Na+ reabsorption. In Na+-reabsorbing epithelial cells, amiloride-sensitive Na + channels (ENaC) are located at the outer, mucosal, apical face. Na+-K+ pumps are mainly localized at the inner, serosal, basolateral face. This topology creates a Na + sink. K § recycles via pump and basolateral K § channels, e.g., ROMK.
503 flows along its electrochemical gradient through a variety of K + channels in the basolateral membrane. This concerted action of ion channels and the Na+-K + pump generates a pathway for Na + across the epithelial cell. CI- secretion represents an example of the functional importance of C1- channels in controlling vectorial transport (Fig. 18). Acinar cells in secretory glands transport C1- against its electrochemical gradient from the interstitial side (blood side, basolateral side) into the cell mainly via the Na+-K+-2C1 - cotransporter (NKCC) fueled by the Na+-gradient generated by the Na+-K + pump. Na + and K + recycle via the Na+-K + pump and basolateral K + channels. C1- secretion can be activated by
F I G U R E 18. Transcellular transport of C1- in secretory epithelial cells-Transcellular C1- transport results from the asymmetric distribution of Na+-K+-2 C1- cotransporters, mainly located at the blood or interstitial site (abluminal), and cAMP-gated (CFTR) or Ca2+-gated (CaCC, CLCA) C1- channels at the luminal side. C1~-entry proceeds via the abluminal Na+-K+-2 C1- cotransporter, luminal secretion either via cAMP-activated CFTR channels stimulated by PKA (top) or via Ca2+activated C1- channels (bottom). The elevation in [Ca2+]i occurs via release from intracellular stores due to activation of G-protein-coupled receptors (GPCR). C1-movement is passively followed by water flux through a paracellular pathway (for explanation see text).
504
SECTION III Membrane Excitability and Ion Channels
two mechanisms: one is cAMP-dependent, the other Ca 2+dependent. C F I ~ channels at the apical side of epithelial cells are stimulated by cAMP-mediated phosphorylation. Opening of these cAMP-activated C1- channels induces a transcellular transport of C1-. Defective CFTR channels in patients with cystic fibrosis, a secretory malfunction (see Section II.E4), cannot be activated even if the secretory stimulus still generates the same amount of cAMP. Therefore the secretion and the coupled transport of water and Na + no longer occur in these patients. The opposite effect occurs during overproduction of cAMP in the presence of the Gs-protein activating cholera toxin. Secretory C1- channels are maximally activated, causing a massive secretion of fluid toward the luminal side. The other mechanism that activates C1- secretion is initiated by the binding of agonists, e.g., acetylcholine or cholecystokinin, to their respective G-protein coupled receptors. Activation of these receptors releases Ca 2+ from intracellular
Ins(1,4,5)P 3 sensitive stores and increases [Ca2+]i, which opens C1- channels at the luminal side (CaCC, CLCA, see Section II.E3). These channels drain C1- into the lumen of the secretory duct, increasing the abluminal C1- gradient and activating NKCC. The inwardly transported Na + and K + recycle via the Na+-K + pump and presumably via CaZ+-activated K + channels in both luminal and abluminal membranes. The vectorial transport of C1- would be accompanied by a passive movement of Na + and H20 through a paracellular pathway (see Fig. 18). Approximately 50% of the filtered C1- is reabsorbed in the proximal tubules. An important component of the CI- reabsorption is passive and paracellular, and driven by the electrochemical gradient (Fig. 19A). In addition, C1--formate (and also C1--oxalate) exchangers have been identified that transport C1-uphill through the apical membrane. The recycling of formate (oxalate) occurs by an H+-coupled transport in parallel with the Na+-H + exchanger. C1- exits at the basolateral
FIGURE 19. C1-reabsorption in the kidney. (A) Proximal tubule. C1- is reabsorbed from the luminal side facing the tubule fluid via exchange with a formate transporter (HCOO-) and oxalate transporter. Formate (or oxalate, not shown) is transported into the cell via a Ht-HCOO- cotransporter. Ht recycles in exchange for Nat transported by the Na+-Ht exchanger, NHE. The transcellular transport of C1- at the basolateral (abluminal) side proceeds via passive transport through C1- channels, most likely C1C-Ka and b. In addition, C1- leaves the cells via a K+-CI- cotransporter. Nat is also secreted through the basolateral side via cooperative action of NHE and the Nat,Kt-ATPase. Kt recycles via pump and Kt channel. (B) Distal tubule. (TAL: thick ascending limb; DCT: distal convolute tubules). At the luminal side CI- is cotransported with Nat and Kt via the NKCC transporter. Kt leaves the cells via ROMK1, a member of the Kir family. At the basolateral side, C1- exits the cells via the C1C-Kb channel. Nat passages through the cells via NKCC and the Na+, Kt-ATPase. ROMK1, NKCC, and C1C-Kb are targets in Bartter's syndrome.
29. Ion Channels in Nonexcitable Cells
membrane via channels (possibly C1C-Ka,b) and the K+-C1cotransporter. Na § which enters the cell at the apical side via the Na§ + transporter exits through the basolateral membrane via the Na+-K § pump, and K + recycles via basolateral K § channels (see Fig. 19). Therefore, a net C1- uptake is driven by the combined action of various exchangers, cotransporters, and pumps. As already mentioned, C1- secretion in the distal tubule (TAL, thick ascending limb, and DCT distal convolute tubules) is different: C1- enters the cell via NKCC, K + recycles through ROMK. The abluminal transport of C1- and Na § occurs via C1C-Kb channels and via NKCC and the basolateral Na+-K § pump respectively. Defects in NKCC, ROMK, and C1C-Kb induce hypokalemic acidosis, low blood pressure, and renal salt-wasting, but also hypercalciuria by a mechanism not yet understood (see Fig. 19B). The total blood Ca 2§ includes a fraction bound to plasma proteins (albumin) and an unbound fraction which is ultrafiltrated. The ultrafiltrated fraction contains Ca 2§ complexed by anions and free Ca 2+. Importantly, part of the free Ca 2+ is reabsorbed in the kidney and the intestine. Vitamin D and PTH (parathyroid hormone) activate Ca 2+ reabsorption. Transepithelial uptake of Ca 2+ seems to be a three-step process consisting of Ca 2+ entry into the cell down its electrochemical gradient, binding to calbindin, a 28-kDa Ca 2+binding protein, and active extrusion out of the cell via a basolateral plasma membrane CaZ+-pump (PMCa). Part of the active Ca 2+ reuptake may also occur by a Na+-Ca 2+ ex-
FIGURE 20. Reabsorption of C a 2+ in kidney and intestine. C a 2+ enters the epithelial cell passively via the recently cloned ECaC channel, binds to calbindin, and is extruded at the abluminal side via the PMCa Ca2+ pump and probably also via the Na+-Ca2§ exchanger.
505
changer (Fig. 20). Recently a channel called EcaC has been cloned which is closely related structurally to the trp channel family and may be responsible for Ca 2+ entry at the apical membrane. C h a n n e l co-operation, as a nice example of the cooperation between ion channels in the organization of vectorial transport, has been proposed by E. Neher (Fig. 21). A rise in [Ca2+]i in a stimulated pancreatic acinar cell opens Ca 2+activated C1- channels at the luminal side, leading to a C1effiux down its electrochemical gradient. At the abluminal side this rise in [Ca 2+ ], opens Ca 2+ -activated nonselective cation channels and Ca~+-activated C1-channels, which depolarizes the abluminal membrane and reverses the electrochemical gradient for CI-, leading to a C1- influx. Because
FIGURE 21. The push-pull hypothesis of vectorial transport. In a resting secretory pancreas acinus cell C1- channels and Ca2§ dependent NSCs are closed. Stimulationby a secretagogue induces first a luminal increase in [Ca2+]i that activates C1- channels and induces a C1- effiux via CaCC, because the luminal membrane potential is more negative than the C1- equilibrium potential. The abliminal channels are still closed. A Ca2+ wave reaches the abluminal site with a delay. [Ca2+]iis now higher at the abluminal side. Opening of Ca2+-dependent NSC and C1- channels at the abluminal side depolarizes the membrane to a value more positive than the C1- equilibrium potential, resulting in C1- influx. The luminal channels are closed again. Sequestration of Ca2§ leads again to the resting state. [Modified after Kasai and Augustine (1990), with permission from Nature 348:735-738 (1990), copyright 1990 Macmillan Magazines Ltd.]
506
SECTIONIII Membrane Excitability and Ion Channels
Ca2+-waves originate at the luminal side and reach the abluminal side with some delay, these events occur sequentially, that is, C1- is first "pushed" into the lumen and then "pulled" into the cell from the abluminal side (push-pull model).
C. Role of Ion Channels in Ca2+-Signaling Many cellular functions modulated by agonists are triggered by changes in intracellular Ca 2+, consisting of an initial transient Ca 2§ rise generated by Ca 2§ release from Ins(1,4,5)P3-sensitive Ca 2+ stores followed by a more sustained Ca2+-plateau coupled to Ca 2+ influx. It is mainly the latter increase that is essential to maintain, for example, the secretion or synthesis of various messengers. An important example is the Ca2+-dependent synthesis of endotheliumderived relaxing factor EDRF. This factor, identical with the simple chemical molecule nitric oxide (NO), acts as an activator of the soluble guanylate cyclase that increases the intracellular cGMP concentration in neighboring cells. EDRFNO synthesis and release are triggered during a long lasting elevation of [Ca2+]i via a Ca2+-calmodulin-NADPH-depen dent NO synthase that catalyses oxidation of the N-guanidine terminus of L-arginine, and releases the easily diffusable NO. This sustained Ca 2+ influx occurs via Ca2+-selective channels (DACs, SOCs, CRACs) or via Ca2+-permeable nonselective ion channels. Activation of Ca2+-activated NSCs creates positive feedback, since the increase of [Ca2+]i via these channels would further activate them. Another feedback is the opening of Ca2+-activated K + channels, which induces hyperpolarization and increases the driving force for Ca 2+ influx. The renin secretion from juxtaglomerular epitheloid cells is another exciting example of cooperation between changes in [Ca2+]i and ion channels. Exocytosis of renin in these cells is enhanced by swelling, but inhibited by an increase of [Ca2+]i. Increased [Ca2+]i activates Ca2+-dependent C1channels and depolarizes these cells by shifting the membrane potential to the C1-equilibrium potential, which in turn promotes K + effiux through K § channels. This coactivation of K + and C1-channels decreases the osmolarity of the cell, which causes it to shrink and reduces renin secretion. Stimulation of these cells, for example, with angiotensin II (AT II), increases [Ca2+]i, which exerts negative feedback on the renin secretion.
IV. S u m m a r y Ion channels in nonexcitable cells have been less extensively studied at the functional level than the channels involved in fast signaling processes via voltage-operated ion channels. Also, the molecular biological information about these channels is rather restricted. Nevertheless, these channels appear to be involved in the control of biologically important functions. We have discussed several of these channels in this chapter: 9 Na+-selective, amiloride-blocked channels (ENaC) which regulate Na + absorption, and Na + secretion in epithelial cells
9 Nonselective cation channels (NSC) that provide a pathway for Ca 2+ entry and cause depolarization. They can be activated by an increase in [Ca2+]i or by the binding of cyclic nucleotides. Recently, a voltage- and cyclic-nucleotide-gated family of hyperpolarizationactivated channels (HAC) has been discovered. Other nonselective cation channels can be activated by extracellular agonists indirectly via G-protein dependent mechanisms or directly by extracellular messengers, such as ATP (Pzx family), or by intracellular cyclic nucleotides (e.g., cGMP, cAMP) 9 K + channels, especially those activated by intracellular Ca 2+ (voltage-dependent small-, intermediate-, and big-conductance channels), which modulate cellular signals caused by various agonists, and inwardly-rectifying K + channels 9 C1-channels, which are mainly involved in secretory functions. They are activated by an increase in [Ca2+]i (CaCCs or the recently identified CACL family), by cAMP via PKA-dependent protein kinase, or by changes in cell volume. Some of them are voltage-dependent (CIC) 9 Most nonexcitable cells have developed a Ca 2+ entry mechanism dependent on intracellular Ca 2+ stores (SOCs, DACs, CRAC) and are responsible for longlasting Ca 2+ signals. Putative molecular candidates may belong to the recently described trp-channel family 9 Various mechanically activated ion channels (K + channels, nonselective cation channels, C1- channels) involved in, for example, volume-regulation of many cells, and in shear-stress mediated release of EDRF in endothelial cells We have also addressed the functional role of the ion channels in volume regulation and fine-tuning of the driving force for Ca 2§ influx, and their involvement in vectorial transport, which is intimately linked to their localization at strategic sites of the cells and usually requires the cooperation of several channel types.
Bibliography Abriel, H., Loffing, J., Rebhun, J. F., Pratt, J. H., Schild, L., Horisberger, J. D., Rotin, D., and Staub, O. (1999). Defective regulation of the epithelial Na§ channel by Nedd4 in Liddle's syndrome. J. Clin. Invest. 103, 667-673. Arreola, J., Melvin, J. E., and Begenisich, T. (1996). Activation of calcium-dependent chloride channels in rat parotid acinar cells. J. Gen. Physiol. 108, 35-47. Begenisich, T., and Melvin, J. E. (1998). Regulation of chloride channels in secretory epithelia. J. Membr. Biol. 163, 77-85. Birnbaumer, L., Zhu, X., Jiang, M. S., Boulay, G., Peyton, M., Vannier, B., Brown, D., Platano, D., Sadeghi, H., Stefani, E., and Birnbaumer, M. (1996). On the molecular basis and regulation of cellular capacitative calcium entry: Roles for Trp proteins. Proc. Natl. Acad. Sci. USA 93, 15195-15202. Canessa, C. M., Horisberger, J. D., Schild, L., and Rossier, B. C. (1995). Expression cloning of the epithelial sodium channel. Kidney Int. 48, 950-955. Davies, P. F. (1995). Flow-mediated endothelial mechanotransduction. Physiol. Rev. 75, 519-560.
29. Ion Channels in Nonexcitable Cells Davies, P. E, and Tripathi, S. C. (1993). Mechanical stress mechanisms and the cell. An endothelial paradigm. Circ. Res. 72, 239-245. Foskett, J. K. (1998). C1C and CFTR chloride channel gating. Annu. Rev. Physiol. 60, 689-717. Gadsby, D. C., and Nairn, A. C. (1999). Control of CFTR channel gating by phosphorylation and nucleotide hydrolysis. Physiol. Rev. 79, $77-S107. Gandhi, R., Elble, R. C., Gruber, A. D., Schreur, K. D., Ji, H. L., Fuller, C. M., and Pauli, B. U. (1998). Molecular and functional characterization of a calcium-sensitive chloride channel from mouse lung. J. Biol. Chem. 273, 32096-32101. George, A. L., Jr. (1998). Chloride channels and endocytosis: C1C-5 makes a dent. Proc. Natl. Acad. Sci. USA 95, 7843-7845. Gruber, A. D., Elble, R. C., Ji, H. L., Schreur, K. D., Fuller, C. M., and Pauli, B. U. (1998). Genomic cloning, molecular characterization, and functional analysis of human CLCA1, the first human member of the family of CaZ+-activated C1-channel proteins. Genomics 54, 200-214. Gruber, A. D., and Pauli, B. U. (1999). Molecular cloning and biochemical characterization of a truncated, secreted member of the human family of CaZ+-activated C1- channels. Biochim. Biophys. Acta. 1444, 418-423. Gruber, A. D., Schreur, K. D., Ji, H. L., Fuller, C. M., and Pauli, B. U. (1999). Molecular cloning and transmembrane structure of hCLCA2 from human lung, trachea, and mammary gland. Am. J. Physiol. 45, C 1261-C 1270. Hamill, O. E, and McBride, D. W., Jr. (1994). The cloning of a mechano-gated membrane ion channel. Trends Neurosci. 17, 439-443. Hille, B. (1992). "Ionic Channels of Excitable Membranes." Sinauer Associates, Sunderland, MA. Hoenderop, J. G. J., van der Kemp, A., Hartog, A., van de Graaf, S. E J., van Os, C. H., Willems, E, and Bindels, R. J. M. (1999). Molecular identification of the apical Ca 2+ channel in 1,25-dihydroxyvitamin D-3-responsive epithelia. J. Biol. Chem. 274, 8375-8378. Hoffmann, E. K., and Dunham, E B. (1995). Membrane mechanisms and intracellular signalling in cell volume regulation. Int. Rev. Cytol. 161, 173-262. Hofmann, T., Obukhov, A. G., Schaefer, M., Harteneck, C., Gudermann, T., and Schultz, G. (1999). Direct activation of human TRPC6 and TRPC3 channels by diacylglycerol. Nature 397, 259-263. Jan, L. Y., and Jan, Y. N. (1997). Cloned potassium channels from eukaryotes and prokaryotes. Annu. Rev. Neurosci. 20, 91-123. Jensen, B. S., Strobaek, D., Christophersen, E, Jorgensen, T. D., Hansen, C., Silahtaroglu, A., Olesen, S. P., and Ahring, E K. (1998). Characterization of the cloned human intermediate-conductance CaZ+-activated K + channel. Am. J. Physiol. 275, C848-C856. Jentsch, T. J., Friedrich, T., Schriever, A., and Yamada, H. (1999). The CLC chloride channel family. Pflugers Arch. Eur. J. Physiol. 437, 783-795. Kasai, H., and Augustine, G. J. (1990). Cytosolic Ca 2+ gradients triggering unidirectional fluid secretion from exocrine pancreas. Nature 348, 735-738.
507 Kellenberger, S., Gautschi, I., Rossier, B. C., and Schild, L. (1998). Mutations causing Liddle syndrome reduce sodium-dependent downregulation of the epithelial sodium channel in the Xenopus oocyte expression system. J. Clin. Invest. 101, 2741-2750. Kirk, K., and Strange, K. (1998). Functional properties and physiological roles of organic solute channels. Annu. Rev. Physiol. 60, 719-739. Kunzelmann, K., and Schreiber, R. (1999). CFTR, a regulator of channels. J. Membr. Biol. 168, 1-8. Kurtz, A. (1990). Do calcium-activated chloride channels control renin secretion? News Physiol. Sci. 5, 43-46. Lansman, J. B., Hallam, T. J., and Rink, T.J. (1987). Single stretchactivated ion channels in vascular endothelial cells as mechanotransducers? Nature 235, 811-813. Ludwig, A., Zong, X., Jeglitsch, M., Hofmann, E, and Biel, M. (1998). A family of hyperpolarization-activated mammalian cation channels. Nature 393, 587-591. Montell, C. (1998). TRP trapped in fly signaling web. Curr. Opin. Neurobiol. 8, 389-397. Nilius, B., Eggermont, J., Voets, T., Buyse, G., Manolopoulos, V., and Droogmans, G. (1997). Properties of volume-regulated anion channels in mammalian cells. Progress Biophysics and Molec. Biol. 68, 69-119. Nilius, B., Sehrer, J., De Smet, E, Van Driessche, W., and Droogmans, G. (1995). Volume regulation in a toad epithelial cell line: Role of coactivation of K + and C1- channels. J. Physiol. (London) 487, 367-378. Nilius, B., Viana, F., and Droogmans, G. (1997). Ion channels in vascular endothelium. Annu. Rev. Physiol. 59, 145-170. North, R. A. (1996). Families of ion channels with two hydrophobic segments. Curr. Opin. Biol. 8, 474--483. Okada, Y. (1997). Volume expansion-sensing outward-rectifier C1channel: Fresh start to the molecular identity and volume sensor. Am. J. Physiol. 273, C755-89. Oleson, S.-O., Clapham, D. E., and Davies, P. F. (1988). Haemodynamic shear stress activates a K + current in vascular endothelial cells. Nature 331, 168-170. Parekh, A. B., and Penner, R. (1997). Store depletion and calcium influx. Physiol. Rev. 77, 901-930. Rossier, B. C., Canessa, C. M., Schild, L., and Horisberger, J. D. (1994). Epithelial sodium channels. Curr. Opin. Nephrol. Hypertens. 3, 487--496. Schwiebert, E. M., Benos, D. J., Egan, M. E., Stutts, M. J., and Guggino, W. B. (1999). CFTR is a conductance regulator as well as a chloride channel. Physiol. Rev. 79, S 145-S 166. Sheppard, D. N., and Welsh, M. J. (1999). Structure and function of the CFTR chloride channel. Physiol. Rev. 79, $23-$45. Strange, K., Emma, E, and Jackson, E S. (1996). Cellular and molecular physiology of volume-sensitive anion channels. Am. J. Physiol. 270, C711-C730. Tsien, R. W., and Tsien, R. Y. (1990). Calcium channels, stores, and oscillations. Annu. Rev. Cell Biol. 6, 715-760. Zhu, X., and Birnbaumer, L. (1998). Calcium Channels Formed by Mammalian Trp Homologues. News Physiol. Sci. 13, 211-217.
A. Li(vano and A. Darszon
Ion Channels in Sperm
I. I n t r o d u c t i o n
Fertilization is one of the most important biological events. It allows not only the generation of a new individual, but an opportunity for genetic recombination as well as an increased probability of species preservation over time. Complex signaling events between gametes are required for their final encounter and fusion, which starts a developmental program. The spermatozoa, a motile cell playing one of the key roles in this process, is very specialized. It lacks the machinery for protein or nucleic acid synthesis and has only a nucleus, mitochondria, a flagellum, an acrosomal vesicle, and a centriole pair. Figure 1A depicts sperm from the sea urchin (upper drawing) and from the mouse (lower drawing). Sperm, like other types of cells, have ion channels that participate in fundamental responses to the outer layer of the egg that are required for fertilization. It has been shown that the flow of ions through the plasma membrane of sperm, particularly Ca 2+, participates crucially in the events leading to fertilization (Schackmann, 1989; Florman et al., 1998; Darszon et al., 1999). Indeed, sperm quickly respond to components of the outer layer of the egg with changes in their plasma membrane ion permeability. During its life span, a sperm must respond to different stimuli, thereby changing behavior until fusion is achieved with the egg to create a zygote.
Sea urchin sperm are metabolically arrested in semen. When spawned into seawater, their respiration and motility quickly activate, and chemotaxis to egg components may help them to find the egg (rev. in Garbers, 1989; Darszon et al., 1999). Contact with the egg jelly induces dramatic morphophysiological changes in sperm through a complex process called the acrosome reaction (AR), which is necessary for fertilization. The most conspicuous morphological change in this reaction is the exocytosis of the acrosomal vesicle and the extension of the acrosomal tubule (Fig. 1B, upper panel). The AR leads to the release of hydrolytic enzymes present in
13
A
|
Acrosomal vesicle Nucleus Plasma membrane Centrioles Mitochondria Flagellum
A. Sea Urchin Sperm Sea urchins are among the most useful and better known experimental models in fertilization because they undergo external fertilization in a simple medium (seawater) and each male can spawn about 10 l~ cells, making it possible to have large amounts of biological material for biochemical and biophysical manipulations. Furthermore, the cells respond to environmental stimuli rapidly (in seconds), synchronously, and in a compulsory order. 5 09
Mitochondria j
Flagellum
~bulo
Bindin
0 o o o ooo} Hybrid oo vesicles
Plasma membrane Acrosomal vesicle Nucleus Centrioles
11. S p e r m R e s p o n s e s to Egg C o m p o n e n t s
Cell Physiology Sourcebook: A Molecular Approach, Third Edition
I,---~ Acrosomal
)
FIGURE 1. Schematic diagram of a sea urchin (upper half) and mouse (lower half) sperm (A) before and (B) after acrosome reaction. The sperm head contains all the cell organelles except the flagellum. Copyright 9 2001 by Academic Press. All rights of reproduction in any form reserved.
510
the acrosome that are required for sperm penetration of the egg's external layers, and to the exposure of new membrane surfaces specialized for fusion with the egg plasma membrane. When sperm undergo the AR, a protein called bindin (see Fig. 1B) is exposed and interacts species specifically with components in the egg. The AR is triggered by plasma membrane ion permeability changes that lead to modifications in membrane potential (Em), intracellular pH (PHi), and intracellular Ca 2+ concentration([Ca2+]i). 1. Responses to Egg Peptides
Small peptides contained in the egg jelly surrounding the sea urchin egg profoundly influence sperm physiology. For instance, pM concentrations of speract, a decapeptide isolated from S. purpuratus egg jelly, and resact, a similar peptide isolated from Arbacia punctulata, stimulate sperm phospholipid metabolism and respiration in sperm suspended in acidified seawater. At nM concentrations these peptides stimulate 22Na+ and 45Ca2+ uptake, H + and K + effiux, and increases in cGMP, cAME [Ca2+]i, and pH i (Garbers, 1989; Schackmann, 1989; Ward and Kopf, 1993). They also regulate sperm motility (Ward et al., 1985; Cook et al., 1994) and appear to enhance fertilization (Suzuki and Yoshino, 1992). Chemotaxis has been demonstrated only in A. punctulata, where nM concentrations of resact attract sperm in a Ca 2+dependent manner (Ward et al., 1985). Cross-linking experiments indicate that speract binds to a 77-kDa sperm plasma membrane protein in the flagella of S. purpuratus sperm; this protein was purified, sequenced, and cloned. It has been proposed that this receptor protein modulates a membrane guanylyl cyclase (rev. in Garbers, 1989). In A. punctulata, resact binds to a membrane-bound guanylyl cyclase, transiently activating it and changing its phosphorylation state (Trimmer and Vacquier, 1986). The phosphorylation state of guanylyl cyclase is pH i- and [Na+]o-dependent (Trimmer and Vacquier, 1986; Garbers, 1989). The resact receptor is the first cloned and sequenced member of a family of guanylyl cyclases that are surface receptors participating in a new signal transduction pathway (Drewett and Garbers, 1994). Speract triggers a transient hyperpolarization in S. purpuratus sperm flagella, in flagellar membranes, and in sperm, probably activating cGMP-modulated K + channels (rev. in Darszon et al., 1999). This hyperpolarization stimulates a voltagedependent Na+-H + exchange in swollen (Babcock et al., 1992; Reynaud et al., 1993; Cook and Babcock, 1993a) and nonswollen sperm (Lee, 1984a, b; Schackmann and Chock, 1986), even though its Na+:H + stoichiometry was estimated to be 1:1 (Lee, 1984a; Lee and Garbers, 1986). In flagellar membranes, GTPyS stimulates the speract-induced hyperpolarization, suggesting the possible participation of a G protein (Lee, 1988). Though Gi, Gs, and several low-molecular weight G proteins have been detected in sea urchin sperm (Ward and Kopf, 1993; Curllar-Mata et al., 1995), their role in sperm physiology has not been established. The relationship between the speract-induced increase in cGMP and cAMP levels and the resulting changes in ionic permeability of the sperm holds the key to understanding how sea urchin gametes manage to meet. Patch-clamping sea urchin sperm is difficult due to their tiny size (see Fig.
SECTION III Membrane Excitability and Ion Channels
5C; Guerrero et al., 1987); thus, to facilitate clamping, they must be swollen in diluted seawater (see Fig. 5D; Babcock et al., 1992). Swollen sperm are spherical. (ca. 4 ]am diameter) and retain their Em, PHi, and [Ca2+]i regulation (Fig. 2A). Picomolar concentrations of speract activate a K +selective channel in swollen sperm, as indicated by patchclamp experiments (Babcock et al., 1992). It is not known how the increase in [cGMP] opens the TEA+-insensitive K +selective channels that hyperpolarize sperm. Higher speract concentrations (>25 pM) transiently hyperpolarize the cells near the K + equilibrium potential. (EK) and immediately repolarize them towards the resting potential. (Er; Fig. 2A, upper left panel). As mentioned earlier, the hyperpolarization activates Na+-H + exchange (Lee and Garbers, 1986; Gonzalez-Martinez et al., 1992; Reynaud et al., 1993). The increase in pH i inhibits guanylyl cyclase (Trimmer and Vacquier, 1986) and stimulates adenylyl cyclase (Cook and Babcock, 1993a, b), which is also activated by a membrane potential hyperpolarization (Beltr~in et al., 1996) and increases in [Ca2+]i (Garbers, 1989). The decrease in [cGMP] would diminish K + permeability (Cook and Babcock, 1993a) and contribute to repolarizing sperm. In addition, nM concentrations of speract increase [Ca2+]i and induce a CaZ+-dependent depolarization beyond E r in swollen sperm (see Fig. 2A; Babcock et al., 1992; Reynaud et al., 1993; Cook and Babcock, 1993a). These changes are inhibited by CaZ+-channel-blockers like Co 2+, Ni 2+, and Zn 2+ (Reynaud et al., 1993; Cook and Babcock, 1993b). These CaZ+-permeable channels allow Mn 2+ to pass through and are regulated by cAMP (Cook and Babcock, 1993b). In normal sperm the hyperpolarization is small and fast and the depolarizing phase (1) is only partly diminished in the absence of external Ca 2+, (2) depends on external Na +, (3) is poorly sensitive to Ca 2+ channel-blockers, and (4) is partially blocked by high concentrations of TEA + and Ba 2+ (Labarca et al., 1996, 1997). Two (or more) ion channels with distinct selectivity and pharmacology might contribute to the depolarization triggered by speract in normal sea urchin sperm: a cAMP- and/or PHi-regulated Ca 2+ channel (Babcock et al., 1992; Cook and Babcock, 1993b) and a cAMP-regulated K + channel that allows Na + flux into sperm (Labarca et al., 1996). A cAMP-modulated K + channel whose open probability increases at hyperpolarizing potentials has been detected in flagellar membranes incorporated into planar lipid bilayers. Because this channel has a P~+/PNa+ of 5, its opening would depolarize sperm. This channel would be activated by the hyperpolarization and the increase in [cAMP] induced by speract, and could explain part of the Na+-dependence of the speract-induced repolarization (Labarca et al., 1996). This channel could be involved in the modulation of motility (Darszon et al., 1999). A cAMP-regulated mildly selective K + channel whose properties resemble those of the channel just described in planar bilayers has been cloned from sea urchin testis and shown to be present in sperm flagella (Gauss et al., 1998; see Section III.A). 2. Acrosome Reaction
Contact of sperm with a glycoprotein-fucose sulfate polymer complex contained in the egg jelly, called factor (FSG), triggers the AR (Trimmer and Vacquier, 1986; Keller and
30. Ion Channels in Sperm
A
5 11
B
S
(,c"2)2
10 FAU
(,c"2h
CH3
20 mV
CH 3
DisC3-(5)
30 s IL |
.....
PHi
a 9
o
o~~,o~ox
I 0 . 2 pH H~
0.2 pH BCECF
m
i
[Ca2+]i o
o
It ( OCCH212N NICH2CO12 0~/~ O-CH2CH2-O~ _ II
riM
N'
CH3
~
I~,o
_ _-2"-,,, _ - - -
1
I 0 . 5 riM
FURA-2
C-O-
FIGURE 2.
Membrane potential (Era), intracellular pH (PHi), and intracellular Ca2+ concentration ([Ca2+]i) changes. (A) S.
p u r p u r a t u s swollen sperm exposed to speract. (B) Normal sperm exposed to FSG. In all records, the arrow indicates 100 nM sper-
act (S) or FSG (F) addition. An upward deflection indicates an increase in the measured parameter. The fluorescent probe used is indicated in the center of the figure: a cyanine dye DisC3-(5) for E m, BCECF for pHi, and FURA-2 for [Ca2+]i. Sperm were swollen in tenfold diluted artificial seawater containing 20 mM MgC12 (DASW) (A). For pHi and [Ca2+]i measurements, cells were loaded overnight with the permeant FURA-2-AM or BCECF-AM dyes at 4 ~ in 0 Ca artificial seawater, pH 7.0. For E m determinations, the cells were pre-equilibrated with 500 nM of the Em-sensitive dye DisC3-(5) during 2-3 min. (Records kindly provided by Marco Gonzfilez-Martfnez, Enrique Reynaud, and Lucia de De La Torre.)
Vacquier, 1994). This reaction involves acrosomal vesicle exocytosis, which exposes material required for sperm-egg binding. These events lead to the extension of the acrosomal tubule, the latter being surrounded by the membrane destined to fuse with the egg (Ward and Kopf, 1993; Darszon et al., 1999). The AR requires external Ca 2+ and Na + in seawater at pH 8.0. Exposure of sperm to FSG induces, within seconds, Na + and Ca 2+ entry and H + and K + effiux (Schackmann, 1989; Ward and Kopf, 1993). These ion fluxes result in interrelated changes in E m (Gonzfilez-Martfnez and Darszon, 1987; Darszon et al., 1999), [Ca2+]i (Schackmann, 1989; Guerrero and Darszon, 1989a), and pH i (Lee et al., 1983; Guerrero and Darszon, 1989b) (Fig. 2B). The FSG also raises cAMP levels, protein kinase A activity, turnover of inositol trisphosphate (InsP3) and phospholipase D activity. It is not clear how these changes relate to the FSG-induced permeability changes (Garbers, 1989; Ward and Kopf, 1993).
When exposed to FSG, L. p i c t u s sperm first transiently hyperpolarize and then depolarize (Fig. 3A). This hyperpolarization is K+-dependent, probably being mediated by K + channels (Gonzfilez-Martfnez and Darszon, 1987), and activates a Na+-H + exchange that leads to an increase in pH i (Schackmann, 1989). It is not known if speract and FSG modulate the same Na+-H + exchange. Antagonists to Ca 2+ channels (verapamil and dihydropyridines) and K + channels (TEA +) inhibit Ca 2+ uptake and the AR in S. p u r p u r a t u s sperm, indicating their mandatory participation in this process (Schackmann, 1989; Darszon et al., 1994). Figure 3B shows the changes in E m (upper trace), [Ca2+]i (middle trace), and pH i (lower trace) in L. p i c t u s sperm suspended in 0 K + seawater associated with a valinomycin-induced hyperpolar+ ization; the latter is followed by a K -induced depolarization about 60 s after the hyperpolarization. The first arrow indicates the addition of valinomycin to increase the K + permeability of the membrane and to bring the E m close to E K. At
512
SECTION III Membrane Excitability and Ion Channels
!
A 9 ~ mM
s
!
700,4 10 FAU I
1
O%
40
Em
60
Val ~
KCI ~
54%
IIOFAU /
76%
--
08
- o6
[Ca2+]i ~.._~
~
0.4
0.2 30 s 2% ~ PHi
~
'
I0.3pH
F I G U R E 3. FSG-induced changes in E m in normal sea urchin sperm. Percent numbers indicate AR; FAU indicates fluorescence arbitrary units. (A) L. p i c t u s sperm's FSG-induced E m changes at the indicated KC1 concentrations (numbers on the left of each record; modified from Gonz~ilez-Martfnez and Darszon, 1987). The upper trace shows a control in normal artificial seawater (ASW). (B) Effect of valinomycin (Val; 2 !uM) and subsequent KCI addition on E m (upper trace), [Ca2+]i (middle trace), and pH i (lower trace) in L. p i c t u s sperm incubated in 0 K ASW. Intracellular pH was measured with the fluorescent probe DMCF, and [Ca2+]i with QUIN-2. Cells were loaded overnight with 10 pM of the permeant form of the dyes at 4 ~ in 0 Ca ASW, pH 7.0, and E m was measured as in Fig. 2. (Modified from Gonz~ilez-Martfiaez et al., 1992.)
this point, only a small percentage of sperm have reacted (at most 16%). However, the hyperpolarization has clearly increased pH i which at about 60 s reaches the value attained during a normal AR. Addition of KC 1 to the medium at this time (second arrow) immediately depolarizes and increases [Ca2+]i and the percentage of reacted sperm (>40%). A valinomycin-induced depolarization in high-KC 1 seawater is unable by itself to trigger the AR. Such experiments indicate that it is possible to artificially manipulate E m and pH i to induce an increase in [Ca2+]i, that the induction of the AR requires coordination between the increase in [Ca2+]. and oH, and that sea urchin sperm have voltage-dependent Ca 2+ ch"an'nels (Gonz~ilez-Martfnez et al., 1992). Measuring [Ca2+]i in sea urchin sperm has provided evidence of the participation of two different Ca 2+ channels in the AR (Guerrero and Darszon, 1989a, b). Figure 4A depicts the changes in [Ca2+]i associated with the FSG-induced AR. After the addition of FSG, [Ca2+]i levels increase ten- to twenty-fold and remain high. In fact, once the AR has occurred, even
though [Ca2+]i remains constant, Ca 2+ influx continues, and Ca 2+ accumulates in the mitochondria until cell death. Figure 4B shows [Ca2+]i changes associated with the addition of FSG to sperm in seawater with Mn 2+ added. 1 The record illustrates a biphasic behavior: immediately after FSG addition, a Ca 2+ channel opens transiently; this channel is very selective for Ca 2+ and does not allow Mn 2+ influx. Subsequently, a second type of channel opens that is less selective than the first one, and allows the influx of Mn 2+ that quenches the FURA-2 fluorescence. Thus, the first type of Ca 2+ channel is a channel that opens by a still unknown mechanism based upon receptor occupancy. The first type of channel is blocked by verapamil and dihydropyridines (DHPs), and it shows inactivation. The other type is not blocked by these compounds, does not inactivate, and allows Mn 2+ to permeate. This second type of channel is blocked by conditions that inhibit the increase in pH i and the AR, but still support a transient increase in intracellular Ca 2+. Thus, the second channel is modulated by pH i. The opening of the first type of channel is required for the opening of the second; blocking of the first inhibits Ca 2+ uptake through the second and blocks the AR. How the two types of channels are coupled is still a mystery; however, both are important to fully achieve AR (Darszon and Gonz~ilez-Martfnez, unpublished). The first Ca 2+ channel could be inactivated at the normal sperm resting potential (around -45 mV; Gonz~ilez-Martfnez and Darszon, 1987), like the low-threshold T-type channels
1This divalent cation has a forty-fold higher affinity for FURA-2 than Ca 2§ and quenches its fluorescence.
~ w
--
96%
,,
_%:
~
08
06 0
A
~ 0.4 ~ 0.2
%
FSG
20 s I,
!
s .....
C j
1o0%
3s~ t
FSG
Mn 2+
98%
F | G i J R E 4. FURA-2 detection of Mn 2+ influx through Ca 2+ channels during the S. purpuratus sperm FSG-induced AR. (A) Control record showing the profile of [Ca2+]i change induced by the addition of FSG. Mn 2+ (3 mM) was added (B) before or (C) after FSG. Mn 2+ influx is indicated by a decrease in fluorescence. Numbers on the right of (A) indicate the fraction of calcium-bound FURA-2. The percentage of FSG-induced AR is shown at the end of each record. (Modified from Guerrero and Darszon, 1989b.)
30. Ion Channels in Sperm
(Hille, 1992). The FSG-induced transient hyperpolarization could remove this inactivation and subsequently trigger the AR. It is known that rat and mouse spermatogenic cells express T-type Ca 2§ channels in their membrane (Hagiwara and Kawa, 1984; Li6vano et al. 1996; Arnoulta et al., 1996; see Section II.B.2), and that the mouse AR is blocked by laM concentrations of dihydropyridines and Ni 2+ (Florman et al., 1992), matching the T-type Ca 2§ channel's sensitivity to those compounds (Hille, 1992; Li6vano et al., 1994, 1996; Arnoulta et al., 1996, 1999). An approximately 210-kDa plasma membrane protein is the best candidate for the FSG sperm receptor. It is not known how the activated receptor triggers the AR, but it could directly regulate ion channels. This protein has species-specific affinity for egg jelly, and some monoclonal antibodies to it induce the AR (Moy et al., 1996). These antibodies bind to a narrow plasma membrane collar over the acrosome and along the entire flagellum (Trimmer and Vacquier, 1986). Recently this glycoprotein receptor, composed of 1450 residues and approximately 50% carbohydrate, has been cloned (Moy et al., 1996). It has 17 potential sites for N-linked, and 12 for O-linked, glycosylation, and only the extreme C-terminal appears to have a putative transmembrane region, strongly suggesting that this protein is not an ion channel. The amino-terminal portion has an EGF domain and two contiguous carbohydrate recognition domains with significant relatedness to those of the human macrophage mannose receptor (important in the activation of the classical complement pathway). The receptor contains a novel module (700 residues) that shares extensive homology with the human polycystic kidney disease protein (PKD1) (Moy et al., 1996). Autosomal dominant polycystic kidney disease is one of the most frequent human genetic diseases and is caused by mutations in PKD 1 and PKD2. The mechanism by which these two proteins function is unknown, though it has been proposed that they modulate ion fluxes (Sanford et al., 1997).
B. Mammalian Sperm Mammalian spermatozoa must undergo changes after leaving the testis to become competent for fertilization. These changes occur in the male reproductive tract (epididymal maturation) and in the female reproductive tract (capacitation and the AR). During maturation the sperm plasma membrane is modified with respect to surface charge, reactivity and distribution of its components, addition of epididymal-secreted proteins, and ionic permeability (rev. in Florman and Babcock, 1991). 1. C a p a c i t a t i o n
Capacitation is essential for successful gamete interaction in mammals (Chang, 1951 Austin, 1951). It regulates the efficiency of acrosome exocytosis in sperm and coordinates it with egg contact to ensure fertilization (Baldi et al., 1996; Visconti and Kopf, 1998). It is worth stressing that unlike somatic secretory cells, which have many secretory vesicles, sperm only have a single large vesicle, the acrosome. Gamete interaction action and fusion requires exocytosis of this vesicle (Yanagimachi, 1994). Preventing premature AR
5 13
is vital for fertilization since sperm that have undergone this reaction too early have difficulties reaching the egg and fusing with it (Florman et al., 1998). Mammalian sperm capacitation can also be accomplished in vitro by incubating ejaculated sperm in defined medium (Yanagimachi, 1994). Three key components are required for this process in mouse sperm: Ca 2§ NaHCO 3, and serum albumin (Visconti and Kopf, 1998). In addition to the changes in composition and distribution of plasma membrane lipids that accompany in vitro capacitation in mouse sperm, a timedependent and cAMP-regulated increase in the protein tyrosine phosphorylation occurs. The serum albumin requirement has been related to its ability to remove cholesterol from the membrane. The decrease in cholesterol content is thought to alter membrane architecture, which results in changes, amongst them an elevation of cAMP levels in sperm (Visconti and Kopf, 1998). Furthermore, during capacitation, [Ca2+]i and pH i increase and membrane potential becomes more negative. These changes are important to endow sperm with the capacity to undergo the AR (Florman et al., 1998). The pH i increase that occurs during capacitation mainly involves a Na +, CI-, and HCO3--dependent mechanism (Zeng et al., 1996) and has been related to sperm cholesterol content (Cross and Razy-Faulkner, 1997). As discussed earlier, pH i may influence sperm Ca 2+ permeability, therefore, an acidic pH. may contribute to maintain membrane potential and low [Ca2+]i, thus avoiding premature AR (Babcock and Pfeiffer, 1987; Darszon et al., 1999). A K + permeability increase accompanies capacitation and significantly contributes to the hyperpolarization (----30 mV) that occurs during this process (Zeng et al., 1995; Arnoult et al., 1999). If mammalian sperm AC is sensitive to membrane potential, like the one from sea urchin sperm (Belmin et al., 1996), this hyperpolarization could stimulate it. Increases in [cAMP] would activate PKA and result in protein phosphorylation. Voltage-dependent channels would also be affected by this membrane potential change, particularly the T-type Ca 2+ channels likely to be present in sperm (see Section II. B.2) (Lirvano et al., 1996; Arnoult et al., 1996a). It is interesting that 30 mM external K + can inhibit capacitation in mouse sperm (Arnoult et al., 1999). This indicates that K + channels play a role in this event. K + -selective channels blocked by TEA + have been recorded in bilayers containing rat sperm plasma membranes (Chan et al., 1997). On the other hand, voltage-dependent K+-selective and TEA+-sensitive currents were observed in rat spermatogenic cells. These currents were insensitive to external Ca 2+ and decreased during spermatogenesis (Hagiwara and Kawa, 1984). Recently, transcripts for Kir5.1, a subunit of inwardly rectifying K + channels, were detected in rat testis. Antibodies against an external epitope of this subunit allowed its detection in spermatogenic cells and in mature sperm. Kir5.1 is unable by itself to form functional channels but can do so when coexpressed with Kir4.1 (Salvatore et al., 1999). Little is known about the regulation of K + channels in spermatogenic cells or in sperm. 2. A c r o s o m e Reaction
The z o n a p e l l u c i d a (ZP), a thick extracellular glycoprotein coat surrounding the egg, is the main mediator of the
514
sperm AR in mammals. ZP3 is the murine ZP s u l f a t e d glycoprotein that displays the sperm-binding and AR-inducing activity of unfertilized eggs. Specific receptors for ZP3 on the plasma membrane of an acrosome-intact sperm overlying the acrosome must mediate its binding and induction of the AR. This Ca2+-dependent reaction involves the fusion of the plasma and outer acrosomal membranes of the sperm head (see Fig. 1B, lower drawing). It exposes the sperm's inner acrosomal membrane, whose surface contains proteases and/or glycosidases thought to allow its penetration through the zona pellucida to reach the egg plasma membrane (Bleil, 1991; Ward and Kopf, 1993). Identification of the sperm surface receptor for ZP3 has been attempted by several laboratories. Among the candidate proteins are the following: in mouse sperm, a/3-1,4 galactosyl transferase (Gong et al., 1995), a hexokinase (Leyton and Saling, 1989), and the lectin sp56 (Bookbinder et al., 1995); in guinea pig sperm, a hyaluronidase (Gmachl and Kreil, 1993). Trypsin-like proteins (Boettger-Tong et al., 1993) and spermadhesins (Gao and Garbers, 1998) have been also proposed as receptors. The physiological relevance of many of these candidates is under active debate (McLeskey et al., 1998; Darszon et al., 1999). Multiple concerted and cooperative interactions between ZP3 and the sperm surface, possibly involving receptor aggregation, may be needed to achieve the signal transduction events that result in the AR (Leyton and Saling, 1989; Ward and Kopf, 1993). How these receptors convey information to initiate signal transduction in mammalian sperm is not known. The ZP-induced AR is inhibited by pertussis toxin (PTX), a specific inactivator of the G i class of heterotrimeric G proteins, in mouse, bovine and human sperm. Among the multiple species of G proteins found in mouse sperm, apparently Gil and Gi2 are preferentially activated by ZP (Ward et al., 1994). It was reported that mouse sperm fl-l,4 galactosyl transferase interacts with a G protein (Gong et al., 1995). There are still many questions that need to be solved: How does the ZP3-activated receptor turn these G i proteins on, which activities do they regulate, and are there ion channels among them? A rise in [Ca2+]i is an essential step in the ZP3 signaling path leading to the AR. External Ca 2+ is required for the physiological AR, and ZP induces [Ca2+]i and pH i increases that precede exocytosis in single sperm loaded with fluorescent dyes as ion indicators. These changes are somehow mediated by G i proteins since both are inhibited by PTX (Florman et al., 1992). It turns out that the PTX-sensitive step in the ZP-induced AR is the pH i increase (Arnoult et al., 1996b). Evidence has been provided for the presence of voltagedependent Ca 2+ channels in the plasma membrane of mammalian sperm (Babcock and Pffeifer, 1987; Florman et al., 1992). Micromolar concentrations of DHPs are needed to block the AR and the increase in [Ca2+]i. Because typical Ltype Ca 2+ channels are blocked by submicromolar concentrations of these blockers, doubts emerged as to the nature of this voltage-dependent Ca 2+ channel (Florman et al., 1998; Darszon et al., 1999). Depolarizing conditions, at elevated pH i, that open Ca 2§ channels sensitive to DHPs bypass the inhibition of the ZP3-induced exocytosis produced by PTX.
SECTION III Membrane Excitability and Ion Channels
It is not known how pH i and [Ca2+]i are intimately related and finely tuned in mammalian sperm. The activation of voltage-dependent Ca 2+ channels is a required step in the ZP3 signal transduction pathway (Florman et al., 1992). It has been suggested that, like in the sea urchin sperm (Guerrero and Darszon, 1989b), at least two different Ca 2+ channels are present in mammalian sperm (Florman, 1994). Two phases of the ZP3-induced increase in [Ca2+]i have been resolved using ion-selective fluorescent probes. First, transient elevation of [Ca2+]i occurs within 40-50 ms to values of approximately 10 laM and subsequently relaxes to resting values within the next 200 ms (Arnoult et al., 1999). The characteristics of this transient in terms of its kinetics of activation and inactivation and pharmacology are consistent with the properties of low voltage-activated (LVA), T-type voltage-sensitive Ca 2+ channels (Arnoult et al., 1999). T-type Ca 2+ currents are the only voltage-dependent Ca 2+ currents expressed during the later stages of rodent spermatogenesis (Hagiwara and Kawa, 1984; Li6vano et al., 1996; Santi et al., 1996; Arnoult et al., 1996a). Like the LVA T-type channels from spermatogenic cells, both the AR and the transient increase in [Ca2+]i are inhibited by micromolar concentrations of DHPs, pimozide, and Ni 2+ (Amoult et al., 1996a, b). Therefore, it is likely that the fast transient increase in sperm [Ca2+]i is mediated by a ZP3-dependent activation of T-type Ca 2+ channels (Florman et al., 1998; Darszon et al., 1999). How does ZP depolarize sperm to open Ca 2+ channels? ZP or ZP3 have been reported to induce a 30-mV depolarization in bovine or mouse sperm (Arnoult et al., 1996b); however, this membrane potential change seems too slow to activate T-type Ca 2+ channels. Two newly cloned channels from testis mRNA, if present in mature sperm, could be candidates to accomplish a ZP3-induced depolarization: mSlo3, a channel with extensive sequence similarity to large-conductance K + channels activated by Ca 2+ and voltage (Schreiber et al., 1998), and a homologue of sea urchin sperm SPIH called hHCN4, which is a member of the pacemaker channel family and thus activated by hyperpolarization and regulated directly by CN (Seifert et al., 1999) (see Section III for more details). Second, the fast transitory response is followed by a much slower elevation in [Ca2+]i which remains high while ZP is present. This slow response requires several seconds to several minutes to develop and AR occurs only after the high, sustained [Ca:Z+]/is reached (Arnoult et al., 1996a, b). The kinetic characteristics of the slow sustained elevation in [Ca2+]i are incompatible with the properties of T-type Ca 2+ channels (Bean and McDonough, 1998); therefore, at least another pathway for Ca 2+ is necessarily involved in triggering the AR. Antagonists of LVA Ca 2+ channels added before ZP3 also inhibit the sustained elevation in [Ca2+]i (Arnoult et al., 1996a). These results indicate that the transient increase in [Ca2+]i due to the ZP-induced activation of T-type channels is necessary to open a second Ca 2+ pathway that keeps [Ca2+]i elevated to allow the AR (Florman et al., 1998; Darszon et al., 1999). ZP3 signaling may achieve the sustained elevation of [Ca2+]i, releasing Ca 2+ from an InsP3-sensitive intracellular store (Walensky and Snyder, 1995) by a mechanism requir-
30. Ion Channels in Sperm
ing prior Ca 2+ influx through a T-type Ca 2+ channel. Consistent with this possibility are the following findings: there are several isoforms of phospholipase C in sperm (Srivastava et al., 1982; Vanha-Perttula and Kasurinen, 1989) and ZP3 stimulates InsP 3 production in these cells (Tomes et al., 1996); InsP 3 receptors are present in acrosomal membranes (Walensky and Snyder, 1995; Trevino et al., 1998); in digitonin-permeabilized mouse sperm exogenous InsP 3 causes 45Ca2+ efflux from a nonmitochondrial pool (Walensky and Snyder, 1995); and agents that are expected to promote Ca 2+ release from smooth endoplasmic reticulum pools in nonpermeabilized sperm promote elevations in [Ca2+]i and the AR (Meizel and Turner, 1993; Blackmore, 1993). Nevertheless, how T-type Ca 2+ channels, Ca 2+ release from intracellular stores, and the AR are coordinated is yet not known. Recently, it was reported that a store-operated Ca 2+ channel is present in the plasma membrane of mouse spermatogenic cells, as well as in immotile testicular sperm. This channel could be responsible for the sustained [Ca2+]i elevation necessary for the AR (Santi et al., 1998). CaZ+-permeable channels that are activated by the depletion of intracellular Ca 2+ stores (SOCs) have been detected in many cell types. SOCs replenish the [Ca 2+] in intracellular stores when it is are depleted during repetitive cycles of Ca 2+ release and may also participate in signaling. Several types of SOCs have been described, such as the highly CaZ+-selective ICRAC pathway, a number of distinct cation-selective channels, and certain mammalian homologs of the Drosophila trp gene (rev. in Parekh and Penner, 1997; Barritt, 1999). At least one member of the trp gene family is expressed in mammalian spermatogenic cells (Wissenbach et al., 1998). Further work is needed to identify the protein components of this influx mechanism in sperm. Progesterone and other progestins can induce PTX- and DHP-insensitive large extracellular CaZ+-dependent increases in [Ca2+]i which result in the AR in mammalian sperm (Thomas and Meizel, 1989; Meizel, 1997; Blackmore, 1999). Considering that this steroid activates phospholipase C (Thomas and Meizel, 1989), it could also lead to depletion of sperm Ca 2+ stores and activation of a SOC. This would be in agreement with observations that progesterone and ZP3 initially activate different transduction systems (Tesarik et al., 1993; Murase and Roldan, 1996) and yet act cooperatively to cause the AR (Roldan et al., 1994). Progesterone metabolites have been shown to enhance the interaction of ~/-aminobutyric acid (GABA) with the GABA receptor in central nervous system neurons. The GABA receptor is a multisubunit protein containing a C1-channel that has been detected in boar and ram sperm. It has been proposed that the fast progestin-induced human sperm responses may involve steroid interaction with a sperm steroid receptor-Cl-channel complex similar to the GABAA-C1-channel complex (Wistrom and Meizel, 1993). The first single-channel recordings made on mouse sperm have indicated the presence of cation channels and a C1channel sensitive to niflumic acid, a blocker of anion channels, which in addition inhibits the mouse AR (Espinosa et al., 1998).
51 5
III. S p e r m Ion C h a n n e l s In spite of the explosion of knowledge on ion channels generated by the patch-clamp technique in the last few years, the properties of sperm ion channels are still relatively unknown. The reason for this is their small size and complex geometry (Fig. 1), which has hindered the use of conventional electrophysiological strategies in the characterization of the channels involved in sperm physiology. As mentioned before, one of the main advantages of working with sea urchin sperm is the large quantity of biological material available. This allows the isolation and characterization of different plasma membrane fractions, which can be reassembled to study sperm ion channels in model systems by different reconstitution strategies (rev. in Darszon et al., 1994, 1996).
A. Sea Urchin Sperm Ion Channels Single K + channels were first recorded in bilayers made at the tip of patch-clamp pipettes from monolayers generated from a mixture of lipid vesicles and isolated sperm flagellar membranes. Three types of K § channels were identified with conductances of 22, 46, and 82 pS (Fig. 5A). Two of them are blocked by TEA § which inhibits the AR (Li6vano et al., 1985). Although with great difficulty, single channels were recorded directly from sea urchin sperm heads using the patch-clamp technique. Single-channel events of 40, 60, and 180 pS were detected, and one of the channels observed was a K + channel (Fig. 5C; Guerrero et al., 1987). As mentioned earlier, swelling S. purpuratus sperm significantly improved the success rate of patch formation and allowed the detection of a 2-5 pS K + channel that is activated by speract. Swollen sperm have opened new possibilities of directly studying the ion channels modulated by egg components and their regulation (Fig. 5D; Babcock et al., 1992). Two types of Ca 2+ channels have been detected in S. purpuratus sea urchin sperm by fusing isolated plasma membranes into planar lipid bilayers (Fig. 5B): (1) a voltagedependent channel of 50 pS conductance observed in 10 mM Ca 2+, and (2) a high-conductance channel having a main conducting state of 172 pS in 50 mM CaC12 and several subconductance states (Fig. 5B; Li6vano et al., 1990). This channel is strongly voltage dependent, showing a single main-conducting state, with rare closing events at voltages more positive t h a n - 2 5 mV, and displaying several subconductance states of lesser conductance at more negative potentials. The main-state conductance size sequence is Ba2+> Sr2+> Ca 2+, as in many other Ca 2+ channels (Bean, 1989). The channel discriminates poorly between divalent and monovalent cations (Pca2§ = 5.9), and is also permeable to Mg 2+ when it is added to the cis side (Pc.a2+/PM,~+= 2.8). In contrast, addition of Mg 2+ to the trans sideb]ock-s~the channel in a voltage-independent manner (Li6vano et al., 1990). Two findings suggest the possible participation of the highconductance Ca 2+ channel in the sperm AR. Both Cd 2+ and Co 2+ block the channel at concentrations similar to those required to inhibit the AR and the Ca 2+ uptake induced by egg jelly (Li6vano et al., 1990). In addition, Mg 2+ blocks the highconductance Ca 2+ channel only when present in the trans
516
SECTION III Membrane Excitability and Ion Channels
A
9
9
~/ (pS) K + 82
Ca 2+ 172
K § 46
K+
130
K + 22
C1-
150
E h (my)
z
K+ 5
(~
(~
0 -20
K§
-40
65
~"~'~'~'*~'~"'~*-
-60
(
-~
!I
!
II
|
...... /f-
K+
130
BLM F I G U R E 5. Strategies used for ion channel characterization in sperm cells. The channels detected with each technique are shown, indicating the main ion transported and their single-channel conductance value. (A) Bilayers at the tip of a patch-clamp pipette. (B) Black lipid membranes (BLM) with fused sperm plasma membrane vesicles. (C) On-cell patch-clamp recordings in sperm (left) and on deflagellated heads (right). (D) On-cell patch-clamp recordings in osmotically swollen sperm. (E) Direct ion channel transfer from cells to the BLM; see Section III.B (Modified from Darszon et al., 1994.)
side. This could indicate that Mg 2+ in seawater may modulate the influx of Ca 2+ into sperm, either from the outside or after entering the cell. Verapamil and nisoldipine do not significantly modify the kinetics nor the conductance of the high-conductance channel seen in planar bilayers. Therefore, the high-conductance Ca 2+ channel could be the second type of channel that participates in the AR, allows Mn 2+ influx, and is PHi-modulated. Little is known about anion channels in these cells. The fusion of sperm plasma membranes into lipid bilayers allowed identification of a 150-pS anion channel (see Fig. 5B). This anion channel was enriched from detergentsolubilized sperm plasma membranes using a wheat germ agglutinin-Sepharose column. Vesicles formed from this
preparation were fused into black lipid membranes (BLM), yielding single channel anion-selective activity with similar properties to those found in the sperm membranes. The anion selectivity sequence found was NO~ > CNS-> B r - > C1-. This anion channel has a high open probability at the holding potentials tested, is partially blocked by DIDS, and often displays substates. DIDS blocks the AR in S. p u r p u ratus sea urchin sperm by a still unknown mechanism. These results suggest that this C1- channel could be involved in the events that lead to the AR, or in determining the resting potential of sperm, which modulates this reaction (Morales et al., 1994). Ion channels are subject to multiple forms of regulation (Hille, 1992), and cyclic nucleotides can directly modulate
30. Ion Channels in Sperm
them (Yau, 1994). Spermatozoa respond to the outer layer of eggs with changes in the second messenger levels ([Ca2+]i, cyclic nucleotides, IP3) (Schackmann, 1989; Garbers, 1989; Ward and Kopf 1993). Recently a K+-selective channel derived from sea urchin plasma membranes, upwardly regulated by cAME has been detected in planar bilayers. Its single-channel conductance is 103 pS in 100 mM KC1. The channel has a low open probability and is weakly voltage dependent. Addition of cAMP on the cis side (the side of membrane addition) up-regulates channel activity in a dosedependent (KD = 200 ~JM) and reversible fashion, increasing the open probability. Millimolar concentrations of Ba 2+ or TEA on the trans side blocked the channel in a voltagedependent fashion. The channel exhibited a low PK+/PNa§ = 5, indicating a sizable permeability to Na+; therefore, its opening would depolarize sperm. Hence, this channel could contribute to the depolarizing phase of the response to speract, and perhaps even during the AR as cAMP levels increase (Labarca et al., 1996). A cAMP-regulated K + channel with similar characteristics as the one described above in planar bilayers has been cloned from sea urchin testis and functionally expressed in HEK 293 cells (Gauss et al., 1998). The cDNA encodes a 767amino-acid polypeptide ( M r .~ 88 kDa) named SPIH with significant sequence similarity to cyclic nucleotide-gated (CNG) and ether-a-gogo (EAG/HERG) channels. The selectivity filter contains the GYG residues typical of K + channels; however, a conserved threonine on the N-terminal side of the GYG sequence is replaced by cysteine, and the aspartate immediately following GYG is replaced by lysine. The channel is only about four times more permeant to K + than to Na + and is much more sensitive to cAMP than to cGME Antibodies to this channel revealed its presence in sperm flagella (Gauss et al., 1998). Because this channel is activated both by cyclic nucleotides and hyperpolarizing potentials, it is a member of a growing family of channels named HCN channels (Clapham, 1998). The channels of this new family are important in shaping the autonomous rhythmic activity of single neurons and the periodicity of network oscillations. B. lon Channels of Mammalian Sperm Fewer reports have been published on single-channel activity from mammalian sperm. For example, it was shown that addition of a partially purified 110-kDa human sperm plasma membrane protein to a preformed lipid bilayer resulted in incorporation of cationic channels having a singlechannel conductance of 130 pS (0.1 M NaC1). Apparently the channel was formed from functionally aggregated triplets (Young et al., 1988). Another study described the detection of two types of Ca 2+ channels in tip-dip bilayers 2 formed from liposomes containing boar sperm plasma membrane. The single-channel conductances detected were between 10-20 pS and 50-60 pS, and the smaller channels were partially blocked by unknown concentrations of ni-
2Dip-tip bilayers are formed at the tip of patch-clamp electrodes by apposition of two monolayers; see Fig. 5A.
5 17
trendipine and verapamil, and completely blocked by 0.5 mM La 3+ (Cox and Peterson, 1989). A nonselective cation channel was also reported, both from cauda epididymal or ejaculated boar sperm plasma membranes incorporated into planar lipid bilayers. Both monovalent and divalent cations permeate through the channel. The channel displays voltage-independent kinetics and is blocked by high concentrations of verapamil or nitrendipine and by ruthenium red (Cox et al., 1991). CNG channels are heterooligomeric complexes formed from at least two subunits (c~ and/3). The oz subunit displays the channel activity, while/3 alone is not functionally active. Coexpression of oz and /3 subunits yields channel species with different properties, when compared to homooligomeric channels (rev. in Kaupp, 1995). The first subunit cloned from bovine testis was ce (Weyand et al., 1994). It has 78% amino acid sequence homology to CNG channels in chicken photoreceptors and contains the cyclic nucleotide-binding site, pore sequence, transmembrane segments, and $4 voltage sensor motif characteristic of the CNG channel family. The single-channel conductance of the channel expressed in X e n o p u s oocytesis is 20 pS. It selects poorly between Na + and K +, is blocked by Mg 2+, is permeable to Ca 2+, and has a much higher affinity for cGMP (>100 fold) than for cAME Small cGMP-induced currents associated with single-channel transitions of < [From Larsen, W. J. (1977). Structural diversity of gap junctions: A review. Tissue Cell 9, 373, with permission.] (h) The large gap junction caps may invaginate into the cell (13) through an endocytotic mechanism. Magnification, 14,400• [From Larsen, W. J., and Tung, H. (1978). Origin and fate of gap junction vesicles in rabbit granulosa cells. Tissue Cell 10, 585, with permission.] (i) Bimembranous gap junction vesicles (14) may pinch off from the invaginating junction to fuse with lysosomes before undergoing degradation. Magnification, 52,200• [From Larsen, W. J., and Tung, H. (1978). Origin and fate of gap junction vesicles in rabbit granulosa cells. Tissue Cell 10, 585, with permission.] (j) Freeze-fracture planes within the cytoplasm may reveal the P-face particles (P) and E-face pits (E) within cytoplasmic gap junction vesicles. Magnification, 38,700• [From Larsen, W. J., and Tung, H. (1978). Origin and fate of gap junction vesicles in rabbit granulosa cells. Tissue Cell 10, 585, with permission.]
525
526
SECTION III Membrane Excitability and Ion Channels
than originally envisioned, but tenacious investigators in a number of laboratories were able to overcome initial obstacles. It is now known that proteins that make up gap junctions in vertebrate tissues are basic and contain significant stretches of hydrophobic sequence. This is not surprising given the fact that these are integral membrane proteins. For these reasons, they proved difficult to isolate and purify. Extraction of tissues with boiling sodium dodecyl sulfate (SDS) proved to be the only reliable method for obtaining morphologically pure fractions of gap junction membranes for many years, but the proteins extracted from these fractions ran at variable molecular weights on polyacrylamide gels. It was controversial as to whether some of these bands were breakdown products, or other proteins associated with gap junctions in the cell membrane. It was not until gentler, nondetergent extraction procedures were developed in the early 1980s that a consistent band on polyacrylamide gels with a predicted molecular weight of 26 kDa could be routinely isolated from rodent livers. Once this native protein was obtained and purified, it was possible to produce a polyclonal antibody. This antibody was used to screen an expression cDNA library to isolate and sequence its gene. This first sequence was published in 1986. Hydropathy analysis of the deduced amino acid sequence of this cDNA revealed a 32-kDa protein which could be interpreted as possessing four potential transmembrane regions, two extracellular loops, and an internal loop and amino and carboxyl termini that extended into the cytoplasm (Fig. 4). The mapping of these proteins with antibodies directed against specific sequences, and their cutting with specific proteases have largely confirmed this initial interpretation of the relationships of gap junction protein segments to the membrane. It was also postulated that six gap junction protein molecules called connexins constitute each gap junction intramembrane particle or connexon, thus providing a molecular basis for models of the permeable membrane junction deduced from earlier physiological and structural studies (Figs. 4 and 5). Indeed, more recent electron cryomicroscopy studies of frozen, rehydrated 2D crystals of a recombinant C-terminal truncated form of Cx43 has revealed even more detail regarding the relationship of connexins to the connexon, and the relationship of their specific transmembrane domains to the plasma membrane. These studies show that each connexin subunit contributes a transmembrane alpha helix that lines the aqueous pore, and an additional alpha helix adjacent to the surrounding membrane lipids. In addition, the images reveal that the apposing connexons that form each channel are staggered by about 30 degrees.
Vll. Two Large Families of Gap Junction Proteins Early structural studies, particularly those utilizing the freeze-fracture technique, demonstrated structural diversity among gap junctions distributed throughout the animal kingdom and within a variety of different tissues. Variability in some of these structural qualities was shown to depend upon
F I G U R E 4. Structure and topology of a connexin. Both amino and carboxyl termini (B) are located in the cytoplasm along with an intermediate loop (A). Every connexin contains four membrane-spanning regions and two extracellular loops containing highly conserved cyteine residues (C). [From Beyer, E. C., Paul, D. L., and Goodenough, D. L. (1990). Connexin family of gap junction proteins. J. Membr. Biol. 116, 187, with permission.]
the tissue in which the gap junction was located, or the phylum or species in which the gap junction was identified. While relationships between detailed gap junction structure and composition are not completely understood, it is now known that gap junctions in different tissues and in different phyla and species may be formed by different protein isoforms within two large families of gap junction proteins. The gap junctions in vertebrate tissues are formed by members of the connexin family of gap junction proteins. The gap junctions of invertebrate tissues appear to be formed by members of an analogous innexin family of gap junction proteins. Members of the connexin family of gap junction proteins range in molecular weight from 26 kDa to 57 kDa (Fig. 6). Each connexin is named for its deduced molecular weight in kilodaltons so that the 26-kDa connexin in humans, for example, is called human connexin 26 (human Cx26), while the 30.3-kDa connexin in mice is called mouse connexin 30.3 (mouse Cx30.3). Sequence analysis supports the possibility that two major phylogenetic subfamilies of connexins diverged from each other between 1.3 and 1.9 billion years ago; one group ranging in size from about 26 kDa to 32 kDA and the other from about 33 kDa to 57 kDa. About a dozen different connexins have been identified in rodents; another dozen or so related isoforms have been described in other species, and the list is growing. Typically, it appears that only one or a few connexin species may comprise the gap junctions within any given tissue. For example, Cx26 and Cx32 constitute gap junctions
527
31. Biology of Gap Junctions
FIGURE 5. This diagram of a region of a gap junction isolated from mouse liver is based on electron microscope images and X-ray diffraction data. Each gap junction particle or connexon is composed of six gap junction protein molecules called connexins. Connexons in apposed membranes meet within the intercellular space. [Reproduced from The Journal of Cell Biology, 1977, 74, p. 449, by copyright permission of the The Rockefeller University Press.]
of mammalian liver and have been shown to coexist within the same gap junction aggregates. On the other hand, the Cx32 and Cx43 documented in thyroid epithelium, appear to be segregated into separate gap junctions formed in different regions of the lateral cell membrane. Likewise, the distribution of three different connexins (Cx40, Cx43, and Cx45) within the human heart appears to be nonrandom. For example, Cx45 is the first connexin to appear in the embryonic mouse heart. Ultimately, however, its expression is restricted to the region of the SA node, while Cx40 is localized in the AV node and in myocytes of the atrium and conductive myocardium. In contrast, Cx43 expression occurs within the crista terminalis and other cardiomyocytes of the heart. As many as six different connexin genes (Cx30.1, Cx31, Cx31.1, Cx40, Cx43, and Cx45) are transcribed and translated in mouse embryos as early as the eight-cell stage. These studies imply differential functions for gap junctions composed of different connexins and consequently stimulate the following question: what part of the connexin molecule encodes functional specificity and how do these specificities differ from connexin to connexin? It is known that the amino terminus, portions of the third transmembrane helix, cystine-containing regions of the extracellular loops, and serine residues of the carboxyl terminus are well conserved (Figs. 4 and 6). However, the cytoplasmic loops and cytoplasmic carboxyl termini vary significantly in length and amino acid sequence from one connexin to another. Moreover, heterotypic gap junctions can be formed only with connexons made of particular connexins, suggesting that sequence specificity within the extracellular domains may be relevant to gap junction assembly. Recent studies, therefore, have focused on the potential functional relevance of specific differences in amino acid sequences within the various domains of the connexin proteins (see Sections IX-XIV). In invertebrates, proteins of the innexin family are able to form gap junctions based on ultrastructural criteria. In addi-
tion, studies with a Xenopus oocyte expression system have shown that some of the innexin proteins are able to form permeable cell-to-cell channels that may be voltage- or pHsensitive. As many as 24 innexin genes have been identified in the nematode, Caenorhabditis elegans, while 6 innexin genes have been described in Drosophila. While the innexins bear no sequence homology to the connexin family of proteins, the structural similarities of innexins and connexins are dramatically similar. For example, like the connexins, the innexin proteins have four transmembrane domains, two extracellular loops, and cytoplasmic N- and C-termini. Moreover, the functions of the innexin proteins in invertebrates parallel functions of the connexins in vertebrate tissues. For example, the innexin encoded by the shak-B(neural) locus forms electrical synapses of the Drosophila nervous system, while the innexin encoded by the eat-5 locus forms permeable junctions responsible for synchronization of pharyngeal musculature in Drosophila.
VIII. Channels within Gap Junctions Following the identification of the gap junction as the site of cell-to-cell transfer of ions and small hydrophilic molecules, a hypothesis developed that the pathway for this exchange consisted of an array of parallel aqueous channels. The central aqueous pore resides in each IMP particle in the freeze-fracture images or hexagonal bridge in the negatively stained micrographs of a gap junction plaque. An aqueous pore 16 nm in length with a channel diameter of approximately 14 A should have a unitary conductance of 10-l~ f~ (= 100 pS). Such a relatively high conductance should produce a detectable electrical signal in support of the channel
Cytoplasmic
Extracellular Intramembrane
cx38, 4..,3, 46 ~' ~ cx 26, 30, 32
I
ooo.oo.o .oo.ooo.oo.x , \ o
m
_J oooomomIooomoIII
l
[,, [] . . . . . ,] ooooooooooooooooooooooo~,o 'I,o "o$ o t : . . , i , ~ m ~ 8 o o o %o%oO, ~,,43,468cx 8 ~ cx26 ., "8 ~8"8 ~.,,- cx 30, 32 I ......1 0
O
C12]
mOlD
r~
OOOOOOOOOO00000000IO000IOI ~ O 0 9 ~"IDOOOOOEICK]O000000
%o%o'o
I cx 26
I 13
i
c, 26, 3o. 3 2 - . ~ cx 38, 43, 4 6 - - ~
OOO0000000[]OOOO0[.l[]O00000000000C]I2)90 O O ~I~IIR~Im~=n.or-.-~. o cx 3 0 1 n O 0 ~-~]0 m,-~ O 9 9 9 0
9
u~--
oooooooooooooooooooooooooooooo
DOO 0 0 O O 0 0 0 [ ] O 0 0 0 0 0 n o o o o o o 0 0 0 0 0
o
cx 38~
oooooooooooooooooooooooooooooo
;
~x43
8
DOOOOO[]OOOOOOOOOOO[]OO[]Oono00oO
cx 46;
ooo ooooooooooooooooooooooooo
FIGURE 6.
This diagram depicts the predicted structure of rat and
Xenopus connexins. Each square represents an amino acid. The diagonal arrows mark points within the cytoplasmic and extracellular loops where the connexins differ in length. The vertical arrows indicate the amino and carboxyl termini for each connexin. Closed squares represent amino acids that are the same in all connexins. [From Bennett, M. V. L., Barrio, L. C., Bargiello, T. A., Spray, D. C., Hertzberg, E., and Saez, J. C. (1991). Gap junctions: New tools, new answers, new questions. Neuron 6, 305, with permission. Copyright 1991 Cell Press.]
528
SECTION III Membrane Excitability and Ion Channels
hypothesis for gap junction permeability. Small stepwise changes in the electrical coupling between paired Xenopus embryonic cells provided evidence for the existence of discrete coupling elements, but the low input resistance owing to the large cell size limited the resolution of the electrical signals. Adaptation of the two-cell voltage clamp to cell pairs of high input resistance achieved precise determination of unitary gap junction channel conductance. Concomitant current records from both voltage-clamped cells always appear as mirror images since each cell resides on opposite sides of the junction (Fig. 7). This equal amplitude and opposite polarity is the criterion used to define junctional current signals in the dual whole cell recording configuration. Channel conductance values of 120 and 160 pS were close to the original predicted value and provided supportive evidence for the existence of relatively large nonselective aqueous channels. Conductance values range from 30 to 300 pS for different gap junction channels under essentially physiological salt conditions. For an aqueous diffusion limited pore, the channel conductance will increase in direct proportion to 27rr 2. For a long pore of 16 nm in length, a 300pS gap junction channel requires a 26-,~ diameter fight cylindrical pore. It follows that higher conductance gap junction channels should exhibit less ionic selectivity and a higher molecular permeability than channels with smaller conductance values.
IX. Evidence for Charge Selectivity The apparent permeability to a variety of water-soluble (hydrophilic) molecules (ions, cAMP, sugars, ATP, nucleosides, etc.) suggests the presence of a large-diameter aqueous pore with minimal selectivity based on electrical charge (equivalent valence at physiological pH). Several permeability studies on mammalian cell and invertebrate gap junctions were instrumental in assigning the commonly accepted molecular permeability limit of 1 kDa or diameter < 14 ~. However, two of these same studies used tagged fluorescent tracers with different valences and demonstrated that the molecular permeability limit was lower for molecules with higher negativity. Selective poermeability of molecules > 600 Da and approximately 10 A in diameter is not surprising since the size of the permeant molecule is approaching the estimated diameter of the pore. Placing any charged surfaces of the pore and the permeable molecule in close proximity to each other would enhance any electrostatic attractive or repulsive forces that might be present. This suggests that the pore of the gap junction channel contains fixed electronegative sites within the pore which reduce the permeability of large negatively charged molecules relative to their neutral or less negative counterparts. Evidence for selectivity at the ionic level did not exist until the patch clamp methodology was adapted to the recording of gap junction channel currents. In two cellular preparations, the junctional membranes of the earthworm septate axon and rat lacrimal gland cells were found to be less permeable to C1- relative to K § by ratios of 0.52 and 0.69, respectively. This modest selectivity among ions with nearl~r identical aqueous mobilities and diameters of less than 4 A
FIGURE 7. (A) Dual whole-cell recording configuration. Equivalent resistive circuit for the whole-cell patch-clamp recording configuration from two coupled cells. The voltage command potentials (V1 and V2) are applied via each negative feedback patch-clamp amplifier (FBA) through whole-cell patch electrodes. The difference between the command potentials and the cell membrane potentials (Vml and Vm2) is minimized by reducing the series resistance (Rsl and Rs2) to less than 5% of the cell input resistance (RinI and Rin2) and junctional resistance (Rj). Accurate junctional current signals are achieved when a majority of the applied current flows across the junction (white arrowheads) instead of the cell membrane (black arrowheads). This condition is readily obtained by using high input resistance cells (Rin > 1 GI)). Conductances (g) are the reciprocal of the corresponding resistance element. (B) The whole-cell currents for each amplifier (11 and 12) appear as simultaneous signals of opposite sign. These are unaltered whole-cell currents obtained from a Cx43-transfected N2A cell pair digitized at 1 kHz after low-pass filtering at 100 Hz. The transjunctional voltage was -30 mV.
suggests that there are weak interactions between the permeant ion and the wall of the pore. In a simple diffusion-limited pore, ions interact with water molecules but not with other ions. Recent observations from four different homotypic connexin hemichannels or gap junction channels report K § to C1-permeability ratios of approximately 10:1. Selectivity among monovalent cations followed their relative aqueous mobility sequence. These observations suggest a common theme to the structure of the selectivity filter within the connexin channel pore. This proposed selectivity filter readily permits the passage of hydrated cations while restricting the simultaneous passage of similar aqueous anions by ten-fold. These observations provide direct evidence for cation-anion interactions within the gap junction pore that contradict aqueous diffusion theory. The observed permeability to
31. Biology of Gap Junctions
atomic and large organic cations is consistent with a pore diameter of 11-13 A for mammalian connexin channels. The ability to transfer these electrical and chemical signals from cell to cell is important to tissue function. The most direct comparison of connexin-specific molecular permeability differences comes from the use of structurally similar molecules of varying size and valence. Two fluorescein derivatives with varying valences but essentially constant physical dimensions (~10/k), 2',7'-dichlorofluorescein (diC1F), and 6-carboxyfluorescein (6-CF), were readily permeable through rat Cx43 channels. In contrast, both dyes were less permeable through chicken Cx43, human Cx37, and rat Cx40 gap junctions. Chicken Cx45 gap junctions were permeable to diC1-F, but not 6-CE This indicates that the more anionic 6CF is restricted in its junctional permeability relative to diC1F in a connexin-specific manner. Since Cx43 has a lower channel conductance than Cx40 and Cx37, no correlation between the maximum conductance and molecular permeability of the connexin channels exists. This lack of correlation between charge selectivity and channel conductance suggests that pore size and conductance are not directly related. Alternatively, recent channel theory proposes that the pore with the smaller diameter and same electrostatic (fixed structural) charge will exhibit the higher channel conductance due to the increased charge density. This updated charged membrane theory can explain the conductance and selectivity properties of known ion channels, including gap junctions.
529
tional (cell-to-cell) voltage are also connexin-specific. The most significant differences in the amino acid sequences of the connexins occur within the cytoplasmic loop and carboxyl-terminal domains. These two cytoplasmic domains provide the basis for distinct conductance and regulatory properties upon the assembled multimeric connexin channels. Since a specific gene encodes for each connexin, tissue- and developmental-specific patterns of expression exist. Limited comparisons of the channel conductance and regulatory properties of the connexin channels to gap junctions in native cell types demonstrate the presence of at least two distinct connexin channel types. The presence of multiple connexin channels can explain the properties of the gap junctions found in many cell types. In one study, the expression of three embryonic chicken heart connexins, Cx42, Cx43, and Cx45, demonstrated that each connexin exhibited unique conductance properties, which coincided with the observance of multiple conductance states in cultured cardiac myocytes. Furthermore, mixing these three connexins in different proportions modeled the response to transjunctional voltage during different stages of cardiac development. Thus, the observed decline in Cx45 expression could significantly explain the decrease in the transjunctional voltage sensitivity of junctional conductance in the developing chicken heart. Developmental and hormonal changes in connexin .~pression occur in other tissues, suggesting a possible role for this mechanism in the modulation of gap junction communication in a variety of tissues.
X. Channel Properties of Different Connexins XI. Gating by lons and Second Messengers The functional expression of different connexins, either by mRNA injection into Xenopus oocytes or stable expression in communication-deficient cell lines derived from mammalian tumors, has yielded tantalizing results. Single-channel unitary conductance values (the ease with which current flows) are connexin-specific. Channel conductance values vary by ten-fold, from 30 to 300 picosiemens (pS = 1/1012 1~), with a majority of the channel conductance values in the 90 to 180 pS range. The reporting of channel conductance became more complicated due to the presence of multiple conductance states for several of the connexin channels. The ability of some connexins to form heterotypic (two connexons of different connexin composition) or heteromeric (two different connexins within the same connexon) gap junction channels adds another level of complexity to the function of channel conductance. The physiological relevance of these conductance differences is poorly understood, but one recent demonstration indicates that connexin composition modulates gap junction channel molecular permeability limits. Rat hepatocyte gap junctions contain almost exclusively Cx32 while mouse hepatocyte gap junctions are composed of a mixture of Cx26 and Cx32. Isolated mouse liver gap junctions exhibited a lower size limit to uncharged sugar molecules and cAMP than rat liver gap junctions, consistent with a smaller pore size, and they exhibited possible charge selectivity, contributed by Cx26 to otherwise Cx32-containing gap junctions. The regulation of gap junction conductance by intracellular pH, intracellular calcium, protein kinases, and transjunc-
Perhaps of greater physiological relevance than transjunctional voltage is the modulation of junctional communication by highly buffered intracellular cations (protons and calcium) and organic second messengers such as cAME diacylglycerol, and inositoltrisphosphate. Whether cations directly bind to a regulatory site (or sites) on the connexin or act through intermediate accessory proteins is unknown. This issue remains an important topic for investigation since it has been observed that interventions that lower intracellular pH also increase intracellular free calcium. Conversely, there is direct evidence that several of the connexins are substrates for phosphorylation by protein kinases such as PKA and PKC. Protein kinase-dependent phosphorylation often increases or decreases junctional communication, depending on the connexins expressed in each tissue. Physiological correlates at the channel level are rare, owing to the difficulty of observing channel activity under distinct ionic or phosphorylation conditions. Generally, cAMP increases junctional conductance and PKC has the opposite effect, at least in Cx43-containing cells. PKA and PKC may both phosphorylate the same serine residue ($233) of Cx32. In contrast, tyrosine phosphorylation by pp60 v-src produces a rapid and reversible uncoupling of Cx43 gap junctions but has no effect on Cx32 gap junctions. Cx43 contains specific v-src SH2- and SH3-binding domains near a tyrosine phosphorylation site (Y265) by pp60 v-src. Cx32 lacks these binding and tyrosine phosphorylation sites. There is increasing evidence that the diversity in the cytoplasmic loop and carboxyl-tail domains of different connexins
530
SECTION
III M e m b r a n e
confers different conductance and regulatory properties on the connexin-specific gap junctions in response to protein kinases and intracellular ions.
9
Excitability
.."-.~.~.i
and Ion Channels
~(,~:;~...
9 :. .:. :. f.:...*-'.~,~i'~.~----~'_&..~-7 ~.I
9 .-...:.:,.:..r .... 9 .:.~:.~:.~:~ ~:~ ,.':,,-.:,,--.:~.::~-;',..o-,.,.'..,. -.......~,,.
XII. R e g u l a t i o n of Functions of C o n n e x i n - B a s e d Gap Junctions at Multiple Levels
P,~
As suggested above, a likely reason for differences in the sequence of the carboxyl tail of connexins may be in the regulation of gap junction-mediated tissue-specific or cell-specific biological activities. Indeed, gap junction regulatory mechanisms described so far appear to fall into one of two general categories. On the one hand, rapid changes in cell coupling that occur within seconds or minutes may reflect regulation of gating of preexisting pores. On the other hand, long term changes which occur within minutes or a few hours to days are likely to involve the modulation of synthesis, assembly, or degradation of connexins (Fig. 8). Existence of this latter scheme of modulation is supported by studies that show that the half-lives of several different connexins are relatively short, ranging from 1 to 3 hours, and by more recent direct videomicrography of living cells expressing green fluorescent protein tagged connexins. In these latter studies, connexins have been observed as they concentrate within the developing Golgi; as they move to the surface to form plaques in regions of cell-cell contact; and as large fragments of the connexin-positive plaques are internalized as apparent vesicles (Fig. 8). Moreover, a variety of studies support the idea that gap junctions may be regulated at virtually any stage of their formation, maturation, and degradation. For example, the possible effect of cAMP on transcription of connexin mRNA and its ultimate translation is supported by studies which show that cAMP-mediated increases in coupling or in junction formation in some cells can be blocked by inhibitors of mRNA or protein synthesis. The potential for control at the level of transcription of connexins is also implied by the characterization of their 5' flanking sequences. For example, the 5' flanking sequence of the myometrial Cx43 gene has been shown to possess several consensus activator protein-1 (AP-1) binding sites as well as half-palindromic estrogen response elements. Since the AP1 proteins, Fos and Jun, are expressed in response to increased estrogen, transcription of Cx43 may be indirectly up-regulated through its AP-1, cis-acting elements. In vascular smooth muscle, the transcription of Cx43 may be upregulated by mechanical load induced by stretch, presumably via an up-regulation of c-fos, which may act in turn on AP-1 binding sites or upon other sites in the 5' flanking region of the Cx43 gene. Promoter analysis of other connexin genes have shown that their transcription may also be regulated by transcription factor binding sites. For example, two GC boxes, two GT boxes, a TTAAAA box, a YY l-like binding site, and a mammary gland factor binding site have been identified in the proximal promoter region of the human Cx26 gene. Likewise AP-1, AP-2, SP-1, TRE, and p53 transcription factor binding sites have been identified in the 5' flanking sequence of the mouse Cx40 gene while Cx-B2, and P1 and P2 promoters are implicated in the transcrip-
9.-'~:~.'.:'-'.'~.~','~-~:.:.".-....- :-a~r I ~,~ .,-.:',-,~:':.,~.~..-"~,~::..,:2...-
\ ......::..:,~',~;i. ~,~.,..-......."
.
~
~
._,~.-
~-,' ~,
x
c$
FIGURE 8. Schemedepicting (1) the synthesis of connexins within rough endoplasmic reticulum (rer), packaging in the Golgi (g), and transport of connexons from the cytoplasm (CS) to the plasma membrane; (2) their assembly in particle-poorregions of the membranecalled formation plaques (fp) into small primary plaques (p), functional gap junction aggregates (a), and then into very large terminal caps (tc); (3) their endocytosis, which is facilitated by clathrin and an actin-mediated contractile mechanism; and (4) their degradation, following fusion with lysosomes resulting in the formation of residual bodies (rb).
tional regulation of Cx32. Yet another level of control of gap junction function is revealed by studies that show the up-regulation of coupling or gap junction formation in the absence of transcription or translation, perhaps at a post-translational level of control. The direct application of dibutyryl cAMP (dbcAMP) to mammary tumor cells in culture, for example, increases the number of gap junctions as well as cell-cell coupling in the absence of an increase in Cx43 mRNA or protein. Likewise, connexin 43 trafficking within several cultured cell lines appears to be regulated by post-translational mechanisms. Specifically, the trafficking of Cx43 from the cytoplasm to the plasma membrane, and its assembly into functional gap junction aggregates may require phosphorylation of two serine phosphorylation consensus sites in the carboxyl tail of this connexin. For example, two cell lines were studied that could not form gap junctions or become coupled, but were able to synthesize significant amounts of Cx43, which was stored within vesicles in the cytoplasm. The Cx43 in these cells was phosphorylated at only one of its serine residues. When one of these lines was transfected with a gene for a cell adhesion molecule, L-CAM, the Cx43 in the transfected line was phosphorylated again, translocated to the plasma membrane and assembled into gap junctions. Further evidence for the phosphorylation-dependent trafficking and assembly of Cx43 in cardiomyocytes has been obtained in experiments with monensin, a reagent that inhibits translocation of proteins between Golgi cisternae. These studies suggest that phosphorylation of Cx43 to the mature form occurs in a compartment of the cell between the Golgi apparatus and the
31. Biology of Gap Junctions
plasma membrane. Similar results are obtained in studies of effects of brefeldin A, which blocks both Cx43 phosphorylation and trafficking to the plasma membrane in a mammary tumor cell line. Finally, it has also been suggested that the process of connexon clustering that gives rise to functional gap junction plaques may be facilitated by the action of microfilaments. Thus the increase in Cx43-positive gap junctions and coupling in the absence of Cx43 mRNA synthesis following treatment of cultured prostatic cells with forskolin may be explained by effects of cAMP on post-translational processing of Cx43, namely their assembly into functional plaques and their translocation from the cytoplasm to the membrane. Numbers of functional gap junctions within the plasma membrane may also be affected by regulation of their removal and degradation. For example, the treatment of some cells with phorbol esters such as 12-o-tetradecanoylphorbol13-acetate (TPA) may stimulate the removal and degradation of gap junctions through phosphorylation of connexins by protein kinase C. Similarly, it is thought that the phosphorylation of mitogen-activated protein (MAP) kinase serine phosphorylation sequences in cultured rat liver cells treated with epidermal growth factor, may mediate the loss of functional gap junctions from the cell membrane. In another example, it has been shown that massive endocytosis of gap junctions in granulosa cells of the ovary is stimulated by high titers of follicle stimulating hormone; that endocytosis is facilitated by clathrin and microfilaments; and that the internalized gap junction vesicles fuse with lysosomes prior to degradation. (Fig. 8). Alternatively, it has been suggested that gap junction plaques and their connexins may be disposed of via a ubiquitin-mediated proteasomal proteolytic pathway. This mechanism is supported by studies showing that connexins accumulate in cells treated with proteasomal inhibitors and by immunogold studies that demonstrate the association of ubiquitin with gap junction plaques (Fig. 9). In summary, it appears that the amount of functional gap junction membrane at the cell surface may be controlled at virtually any point in the gap junction's life history; namely through regulation of (1) transcription of connexin mRNA, (2) translation of connexin protein, and then through posttranslational modifications, which affect the (3) assembly of functional plaques, (4) gating of junctional channels, and (5) degradation of functional gap junctions.
XIII. Specific Biological Functions of Gap Junctions Early speculations regarding the general functional significance of gap junctions, such as the electrical coupling of some neurons and the buffeting of metabolites within a tissue mass, have not been seriously challenged for the past three decades (see Sections I, II, and IX). Several recent studies, however, have begun to establish even more specific functional relationships between gap junctions and tissue activities and behaviors, most notably with respect to the regulation of smooth muscle contraction and ligand-mediated secretory activity. Perhaps one of the best studied cases is that of the myometrial connexin, Cx43. It is known, for example, that the
531
F I G U R E 9. Image of gap junction in cultured HeLa cell transfected with Cx40 and Cx43. The replica was immunostained with antibodies to Cx40, Cx43, and ubiquitin. The 6-nm gold particles indicate Cx40, the 10-nm gold particles stain for Cx43, and the 15-nm gold particles indicate epitopes for ubiquitin.
myometrium of the nonpregnant uterus in a variety of species is virtually devoid of gap junctions and expresses very little, if any myometrial Cx43 mRNA or protein. In contrast, however, Cx43 mRNA is rapidly synthesized and translated and then gap junctions appear coincident with rising estrogen and declining progesterone titres just prior to parturition. Experimental manipulations of these hormones also predictably induce or inhibit the appearance of gap junctions in experimental models of preterm labor or tocolysis respectively (effective labor coincides with their induction while myometrial quiescence parallels their inhibition). For example, removal of the ovaries or the injection of RU486 during the latter part of pregnancy effectively increases the estrogen : progesterone ratio, the expression of Cx43 mRNA and protein, the assembly of myometrial gap junctions, and concomitant preterm labor. Conversely, the injection of progesterone several days prior to term inhibits the formation of gap junctions and onset of labor. Thus, it is likely that transcription and translation of Cx43 are required for gap junction formation and coordinated myometrial contractions in humans. Moreover, the evidence for the presence of AP-1 binding sites in 5' flanking sequences of the myometrial Cx43 gene and the estrogen-mediated expression of Jun and Fos described above (see Section XII) is consistent with this notion. The function of myometrial gap junctions and contractility may also be regulated at the level of post-translational processing. For example, high levels of myometrial Cx43 appear a full day prior to delivery in rats (Fig. 10) but most of the Cx43 in these cells is located within perinuclear vesicles which also stain with an antibody
532
SECTION III Membrane Excitability and Ion Channels
against a Golgi-associated protein (Fig. 10B). Six to twelve hours prior to onset of delivery, however, Cx43 immunopositive plaques are observed in the plasma membrane (Fig. 10C). Conversely, cessation of delivery is accompanied by the loss of gap junctions from the cell surface by a process of endocytosis (Fig. 10E). Gap junctions have also been implicated in the process of secretion and other ligand-mediated functions in many dif-
ferent kinds of cells. For example, gap junctions are well-developed in differentiated endocrine and exocrine cells and are typically sparse or absent in proliferating stem cells. Moreover, the induction of secretory activity itself appears to be correlated with increased numbers of gap junctions in luteal cells, adrenal cortical cells, pancreatic exocrine cells, mammary alveolar cells, and thyroid cells. Studies of osteoblasts have demonstrated a close correlation between the
FIGURE 10. Immunofluorescent staining of myometrium sections incubated with site-specific antibodies to the carboxyl terminus of Cx43 from pregnant rat sacrificed on day 19 (a) or day 21 (b), at midnight on day 21 (c), during delivery on day 22 (d), 6 h after delivery (e), or 24 h after delivery (f). Magnification, 500x. [From Hendrix, E. M., Mao, S. J. T., Everson, W., and Larsen, W. J. (1992). Myometrical connexin 43 trafficking and gap junction assembly at term and in preterm labor. Mol. Reprod. Dev. 33, 27, with permission. Copyright 1992 Wiley-Liss, a division of John Wiley & Sons, Inc.]
31. Biology of Gap Junctions
effects of hormones on the generation of a second messenger, formation of gap junctions, and cell function. In addition, gap junctions appear to be more abundant in cultured confluent osteoblasts than proliferating cells and osteoblast cell coupling in culture and cAMP production are enhanced by the application of parathyroid hormone (PTH). Conversely, an analog of PTH which binds to PTH receptors but attenuates cAMP accumulation, results in a decrease in cell coupling. Finally, transfection of osteoblasts with antisense Cx43 mRNA results in significant concomitant reductions in coupling and in cAMP synthesis in response to PTH. It has been suggested that the development of gap junctions may enhance the sensitivity of hormonally responsive cells to theft specific ligand as a consequence of the transfer of cAMP through gap junctions. Thus cells not receiving direct stimulation by binding of the secretogogue to its receptor may also respond to the ligand-mediated production of cAMP in neighboring cells which possess the appropriate receptor. Such an hypothesis has been invoked to explain the maintenance of meiotic arrest of mammalian primary oocytes. These germ cells initiate the first meiotic division during embryonic life but are then almost immediately arrested at the first meiotic prophase stage. Typically, a primary oocyte does not resume meiotic maturation until it responds to an ovulatory surge of gonadotropic hormones following puberty. However, since meiosis spontaneously resumes when the oocyte is removed from the follicle and cultured in medium-lacking hormones, it has been postulated that the follicular environment is inhibitory to resumption of meiotic maturation. Meiotic arrest may be maintained by follicle cell-generated cAMP which enters the oocyte through a well-developed network of gap junctions. Conversely, it has been suggested that meiotic resumption may be signalled by the LH-mediated disruption of the gap junction pathway within the cumulus mass (following an ovulatory surge of gonadotropins), thus preventing cAMP manufactured within the follicle cells from entering the oocyte. Other hypotheses of both negative and positive regulation of meiotic resumption have also been proposed. It has recently been proposed that gap junction connexins (specifically Cx43) may function in a role somewhat different than those described above; namely in protein trafficking. It has been shown, for example, that the second PDZ domain (of three) of the tight junction-associated protein ZO-1 binds to the extreme C-terminal end of the Cx43 molecule typically resulting in cell membrane colocalization of Cx43 and ZO-1 in cells that coexpress these proteins. In support of the possibility that Cx43 is required for shuttling of ZO-1 to the cell membrane, it was found that both ZO-1 and Cx43 fail to accumulate at the cell membrane in Cx43 knock-out astrocytes.
XIV. Gap Junctions in Human D i s e a s e and in Murine Models of Human D i s e a s e A. Carcinogenesis In the early 1970s, a hypothesis was formulated stating that the regulated and coordinated growth of cells in normal tissues requires the presence of permeable (communicating) cell junctions, while cells incapable of forming communicat-
533
ing junctions exhibit uncontrolled growth like the uncontrolled growth observed in cancer cells. Studies that compared the presence or absence of coupling or gap junctions in normal cells and in several tumors and cancer cell lines initially supported this hypothesis. For example, coupling or gap junctions were found to be deficient in certain hepatoma cells and in L-cell derivatives such as the clone-lD cell line in contrast to their normal counterparts. Indeed, more recent studies support these early observations. For example, an extensive study has examined the role of gap junctions in tumor progression in rat liver following initiation by diethylnitrosamine (DEN) and promotion by either phenobarbitol (PB) or 2,3-dichloro-7,8-dibenzo-p-dioxin (TCDD). Transcripts and protein levels of Cx26, Cx32, and Cx43 were measured and reductions in Cx26- and Cx32-positive gap junctions were observed in all resulting neoplasms. However, it was also demonstrated that these decreases were not always associated with reductions in specific m-RNA transcripts for these connexins, suggesting that the loss of junctions could result from the modulation of transcription or translation or from post-translational modification affecting assembly or degradation (see Section XII). Consistent with these observations, the transfection of HeLa cells with cDNA encoding Cx26, but not Cx40 or Cx43, results in inhibition of tumor formation in nude mice, suggesting that Cx26 gap junctions play a pivotal role in growth control. Similarly, the transfection of communication-deficient hepatoma cells (SKHepl) with cDNA encoding Cx32 results in slowing of the growth rate of tumors in nude mice, compared with tumors arising from communication-deficient parental SKHepl cells. In conflict with these findings, however, other highly malignant tumor cells have been shown to be well coupled or to possess large numbers of gap junctions. These include Novikoff hepatoma cells, human SW-13 adrenal cortical carcinoma cells, and murine B16 melanoma cells. Moreover, other studies have analyzed gap junctions during tumor progression in skin tumors and in hepatocarcinoma and have found that while gap junctions or coupling may disappear during the transformation of normal cells by cancer-causing agents (such as tumor promoters), their loss seems to follow, rather than lead the changes which transform these normal cells to tumor cells. In response to results such as these, a modified defective communication hypothesis was proposed to include cancer cells which might not necessarily exhibit an absence of, or obvious structural defect in the gap junction itself, but which might be defective with respect to some other part of the gap junction communication mechanism, for example, lack of ability of the cell to synthesize the postulated regulatory message, or defects in the receptor which translates the cell's activity. Other modifications of the hypothesis have also been advanced, proposing, for example, that some cancer cells (for example, cultured human SW13 cells and diethylstilbestrol-induced tumors of the proximal tubule in the Syrian hamster) are not able to maintain a steady-state gap junction level through a regulated balance of synthesis or degradation of gap junctions. Under some circumstances, these cells possess normal or even excessive gap junction complements and are mitotically quiescent, while under other conditions the cells lose most or all of their junctions, resulting in cancerous growth. A temporal relationship
534
between gap junction formation and degradation, and the rate of proliferation is also exhibited by normal liver cells in partially hepatectomized rats. Normal rat liver cells remaining after partial hepatectomy initially lose their gap junctions and proliferate rapidly. By 40 hours after partial hepatectomy, the liver cells regain their gap junctions coincident with a reduction in the rate of DNA synthesis and cell division. Another variation of the hypothesis is based upon a series of experiments that demonstrate that normal cells may have the ability to suppress tumor cell behavior in some normaltumor cell cocultures as long as the tumor cells can also form gap junctions with the cocultured normal cells. Other tumor cells may not be able to form gap junctions with normal cells but may be able to form gap junctions between themselves. In these tumor cells, contact with normal cells does not suppress the transformed phenotype of the tumor cell. However, the ability of certain tumor cells to make gap junctions with one another may result in more efficient killing by antitumor agents by allowing the cell-to-cell movement of lethal metabolites formed in some, but not all of the tumor cells within the interconnected network. This phenomenon is called the bystander effect. In conclusion, it is unlikely that the progression of all cancers is dependent upon the activity of gap junctions. It is also unlikely that mechanisms of gap junction-mediated growth control will be identical in all cells or tumors where the activity of gap junctions can be demonstrated to play a pivotal role. B. Charcot-Marie-Tooth D i s e a s e and Mutations of Cx32
Spontaneous mutations of Cx32 result in a common peripheral neuropathy in humans, Charcot-Marie-Tooth disease, which affects 1 of every 2500 individuals. While this demyelinating disease may arise as a consequence of mutations of a peripheral myelin protein PMP22 (on chromosome 17) or of the myelin constituent Po (on chromosome 1), numerous families have been identified with an X-linked dominant form of the disease characterized by as many as 150 different mutations of Cx32 (CMTX). Mutations may involve single-base substitutions, formation of a premature stop codon, frame-shift, or elimination of an amino acid residue. They may occur in amino or carboxyl termini, cytoplasmic or extracellular loops, or in transmembrane regions of the connexin. The causal relationship between these diverse mutations of Cx32 and demyelination of peripheral nerves is obviously varied since the mutations are quite diverse and may or may not affect channel properties. In some cases, it seems possible that mutations may directly affect the function of Cx32 gap junctions in peripheral myelin by interfering with their possible function as ATP-sensitive hemichannels, their function in the exchange of nutrients between the perinuclear region of the Schwann cell and the Schmidt-Lantermann incisures and paranodal processes at the Node of Ranvier, or their possible function in signaling between internodes. In other cases, it is thought that the mutations may indirectly affect Cx32 function through effects on Cx32 synthesis and trafficking within the cell. Targeted knockout of connexin 32 in mice also results in changes in liver enzyme activities and in glucose mobilization, but most
SECTION III Membrane Excitability and Ion Channels
interestingly, these Cx32-deficient mice have an increased susceptibility to hepatocarcinogenesis. Otherwise, they are vital and fertile. C. Heart Function and Mutations of Cx43 and Cx40
While no specific heart defects related to mutations of Cx43 have been observed in humans, a series of studies with animal models of ischemia support a role for Cx43 gap junctions in recovery from injuries to the myocardium. For example, the number of gap junction plaques within the intercalated disc has been shown to decrease in the region of the border of the healing infarct. In addition, the spatial distribution of gap junctions is also abnormal in these areas. These observations along with studies that show the critical role of Cx43 in conduction (particularly in the ventricle) have led to suggestions that alterations in Cx43 levels and distributions may play an important role in development of infarct-mediated arrhythmias. Targeted knockouts of Cx43 and mice overexpressing Cx43 are characterized by abnormal development of the pulmonary (right ventricular) outflow tract resulting in their death immediately following birth. Enlargement of the fight ventricle with thinning of the wall, attenuation of the ductus arteriosus, and formation of one or two abnormal pouches at the base of the pulmonary outflow tract were also observed. It has been suggested that these defects of conotruncal development may result from disruption of the cardiac neural crest. In addition, Cx43 knockout mice of both sexes have small gonads, resulting in part from deficiencies in germ cell development. Since these mice die shortly after birth, the gonads were cultured to determine their capacity for folliculogenesis. It was found that follicles in these ovaries develop only to the primary follicle stage. Targeted knockout of Cx40 results in defects consistent with its localization in the AV node and in conductive myocardium. Electrocardiography of Cx40 knockout animals was characterized by reduced conduction velocities and abnormalities characteristic of first-degree atrioventricular block with associated bundle branch block. D. Cataract Formation and Mutations of Cx50
Mutations of connexin 50 have recently been linked to inherited congenital cataract in human lens. Since lens fibers are linked by extensive gap junctions composed of Cx50 (and Cx46), it is thought that the maintenance of junctional communication between lens fibers is necessary for the maintenance of normal transparency of the lens. Targeted knockout of Cx50 in mice leads to development of zonular pulverulent cataracts of the lens but also results in micropthalmia. These results thus implicate Cx50 not only in the maintenance of lens transparency but also in the growth of the eye. E. Deafness and Mutations of Cx26
A variety of mutations of Cx26 including missense mutations and mutations resulting in frameshifts and formation of premature stop codons have been linked to both autosomal recessive and autosomal dominant hearing loss in humans. It has been suggested that these mutations affect the
31. Biology of Gap Junctions
formation of gap junctions in the fibrocytes in the spiral ligament that underlies the stria vascularis of the cochlea leading to disruptions of ion and e n d o l y m p h circulation. Targeted knockout of Cx26 results in embryonic lethality presumably as a consequence of placental dysfunction.
E Fertility and Targeted Knockout of Cx37 While no h u m a n syndrome resulting from mutations of Cx37 is known, the targeted knockout of Cx37 in mice has been shown to result in the disruption of folliculogenesis. Follicles develop only to early antral stages, the oocytes fail to acquire meiotic competence, and ovulation does not occur.
XV. Summary A structure serving as an electrical synapse in excitable tissues and as permeable cell-to-cell ion channels in inexcitable cells is the gap junction. Although the function of these structures as electrical synapses in neurons and cardiac muscle is generally understood, new studies relating the activity of specific connexins (and innexins) in particular conduction pathways (for example in the heart) m a y begin to specify the functions of gap junctions c o m p o s e d of different connexins. Developmental studies now seem to hold particular promise, while studies in the field of cancer cell biology continue to make progress, including analysis of the relationship between connexin expression, processing, assembly, gating, and degradation and transformation to malignancy. In addition, current studies of the properties of gap junction channels and of the cellular regulation of gap junction gating, formation, and turnover in relation to normal cell physiology m a y yield more specific functional information. Connexin-specific gap junction channels exhibit differences in electrical conductance and molecular permeability that are not directly correlated and dispute the contention that all gap junctions are large aqueous pores. Fortunately, DNAs encoding several connexins have been cloned and significant advances have been made in the correlation of sequence and function. Moreover, the application of this knowledge to the development of transgenic animals exhibiting gain or loss of function of specific connexins continue to lead to more specific insights into the functional significance of particular connexins and innexins.
Bibliography Alcolea, S., Theveniau-Ruissy, M., Jarry-Gnichard, T., Maries, I., Tzouanacou, E., Chavin, J. E, Briand, J. E, Moorman, A. E, Lamers, W. H., and Gros, D. B. (1999). Downregulation of connexin45 gene products during mouse heart development. Circ. Res. 84,1365-1379. Bai, S., Schoenfeld, A., Pietrangelo, A., and Burk, R. D. (1995). Basal promoter of the rat connexin 32 gene: Identification and characterization of an essential element and its DNA-binding protein. Mol. Cell. Biol. 15, 1439-1445. Bennett, M. V. L., and Verselis, V. K. (1992). Biophysics of gap junctions. Sem. Cell Biol. 3, 29-47.
535 Bennett, M. V. L., Barrio, L. C., Bargiello, T. A., Spray, D. C., Hertzberg, E., and Saez, J. C. (1991). Gap junctions: New tools, new answers, new questions. Neuron 6, 305-320. Bennett, M. V. L., Zheng, X., and Sogin, M. L. (1994). The connexins and their family tree. In "Molecular Evolution of Physiological Processes," pps. 223-233. Rockefeller University Press, New York. Bevans, C. G., Kordel, M., Rhee, S. K., and Harris, A. L. (1998). Isoform composition of connexin channels determines selectivity among second messengers and uncharged molecules. J. Biol. Chem. 272, 2808-2816. Beyer, E. C., Paul, D. L., and Goodenough, D. L. (1990). Connexin family of gap junction proteins. J. Membrane Biol. 116, 187-194. Brink, P. R., and Dewey, M. M. (1980). Evidence for fixed charge in the nexus. Nature 285, 101-102. Brink, P. R., and Fan, S.-E (1989). Patch clamp recordings from membranes which contain gap junction channels. Biophys. J. 56, 579-593. Budunova, I. V., Carbajal, S., Viaje, A., and Slaga, T. J. (1996). Connexin expression in epidermal cell lines from SENCAR mouse skin tumors. Mol. Carcinog. 15, 190-201. Chiba, H., Sawada, N., Oyamada, M., Kojima, T., Nomura, S., Ishii, S., and Mori, M. (1993). Relationship between the expression of the gap junction protein and osteoblast phenotype in a human osteoblastic cell line during cell proliferation. Cell Struct. Func. 19, 419-426. Coppen, S. R., Kodama, I., Boyett, M. R., Dobrzynski, H., Takagishi, Y., Honjo, H., Yeh, H. I., and Severs, N. J. (1999). Connexin45, a major connexin of the rabbit sinoatrial node, is co-expressed with connexin43 in a restricted zone at the nodal-crista terminalis border. J. Histochem. Cytochem. 47, 907-918. Cowan, D. B., Lye, S. J., and Langille, B. L. (1998). Regulation of vascular connexin43 gene expression by mechanical loads. Circ. Res. 82, 786-793. Curtin, K. D., Zhang, Z., and Wyman, R. J. (1999). Drosophila has several genes for gap junction proteins. Gene 232, 191-201. Darrow, B. J., Laing, J. G., Lampe, P. D., Saffitz, J. E., and Beyer, E. C. (1995). Expression of multiple connexins in cultured neonatal rat ventricular myocytes. Circ. Res. 76, 381-387. Donahue, H. J., McLeod, K. J., Rubin, C. T., Andersen, J., Grine, E. A., Hertzberg, E. L., and Brink, P. R. (1995). Cell-to-cell communication in osteoblastic networks: Cell line-dependent hormonal regulation of gap junction function. J. Bone Min. Res. 10, 881-889. Ewart, J. L., Cohen, M. E, Meyer, R. A., Huang, G. Y., Wessels, A., Gourdie, R. G., Chin, A. J., Park, S. M., Lazatin, B. O., Villabon, S., and Lo, C. W. (1997). Heart and neural tube defects in transgensis mice overexpressing the Cx43 gap junction gene. Development 124, 1281-1292. Filson, A. J., Azarnia, R., Beyer, E. C., Loewenstein, W. R., and Brugge, J. A. (1990). Tyrosine phosphorylation of a gap junction protein correlated with inhibition of cell-to-cell communication. Cell Growth Diff. 1, 661-668. Fishman, G. I., Moreno, A. P., Spray, D. C., and Leinwand, L. A. (1991). Functional analysis of human cardiac gap junction channel mutants. Proc. Natl. Acad. Sci., USA 88, 3525-3529. Flagg-Newton, J., Simpson, I., and Loewenstein, W. R. (1979). Permeability of the cell-to-cell membrane channels in mammalian cell junction. Science 205, 404-407. Forge, A., Becker, D., Casalotti, S., Edwards, J., Evans, W. H., Lench, N., and Souter, M. (1999). Gap junctions and connexin expression in the inner ear. Novartis Found. Symp. 219, 134-150. Furshpan, E. J., and Potter, D. D. (1959). Transmission at the giant motor synapses of the crayfish. J. Physiol. 145, 289-325. Furshpan, E. J., and Potter, D. D. (1968). Low-resistance junctions between cells in embryos and tissue culture. Curr. Top. Dev. Biol. 3, 95-127.
536
Giepmans, B. N., and Moolenaar, W. H. (1998). The gap junction protein connexin43 interacts with the second PDZ domain of the zona occludens-1 protein. Curr. Biol. 8, 931-934. Goldberg, G. S., Martyn, K. D., and Lau, A. E (1994). A connexin 43 antisense vector reduces the ability of normal cells to inhibit the foci formation of transformed cells. Mol. Cytogen. 11, 106-114. Goodenough, D. A., Goliger, J. A., and Paul, D. (1996). Connexins, connexons, and intercellular communication. Ann. Rev. Biochem. 65, 475-502. Goodenough, D. A., Simon A. M., and Paul, D. L. (1999). Gap junctional intercellular communication in the mouse ovarian follicle. Novartis Found. Symp. 219, 226-235. Haeflinger, J.-A., Bruzzone, R., Jenkins, N. A., Gilbert, D. J., Copeland, N. G., and Paul, D. L. (1992). Four novel members of the connexin family of gap junction proteins. J. Biol. Chem. 267, 2057-2064. Hendrix, E. M., Mao, S. J. T., Everson, W., and Larsen, W. J., (1992). Myometrial connexin 43 trafficking and gap junction assembly at term and in preterm labor. Mol. Reprod. Develop. 33, 27-38. Hendrix, E. M., Myatt, L., Sellers, S., Russell, E T., and Larsen, W. J. (1995). Steroid hormone regulation of rat myometrial gap junction formation: Effects on Cx43 levels and trafficking. Biol. Reprod. 53, 547-560. Hille, B. (1992). "Ionic Channels of Excitable Membranes." 2nd Ed. Sinauer Assoc., Sunderland, MA. Hoh, J. H., John, S. A., and Revel, J.-P. (1991) Molecular cloning and characterization of a new member of the gap junction gene family, connexin-31. J. Biol. Chem. 266, 6524-6531. Huelser, D. E, Rehkopf, B., and Traub., O. (1997). Dispersed and aggregated gap junction channels I identified by immunogold labeling of freeze-fractured membranes. Exp. Cell Res. 233, 240-251. Jordan, K., Solan, J. L., Dominguez, M., Sia, M., Hnad, A., Lampe, E D., and Laird, D. W. (1999). Trafficking, assembly, and function of a connexin 43-green fluorescent protein chimera in live mammalian cells. Mol. Biol. Cell 10, 2033-2050. Juneja, S. C., Barr, K. J., Enders, G. C., and Kidder, G. M. (1999). Defects in the germ line and gonads of mice lacking connexin43. Biol. Reprod. 60, 1263-1270. Kamibayashi, Y., Oyamada, Y., Moil, M., and Oyamada, M. (1995). Aberrant expression of gap junction proteins (connexins) is associated with tumor progression during multistage mouse skin carcinogenesis. Carcinogen. 16, 1287-1297. Kanemitsu, M. Y., Loo, L. W. M., Simon, S., Lau, A. E, and Eckhart, W. (1997). Tyrosine phosphorylation of connexin 43 by v-src is mediated by SH2 and SH3 domain interactions. J. Biol. Chem. 272, 22 824-22 831. Kanter, H. L., Saffitz, J. E., and Beyer, E. C. (1992). Cardiac myocytes express multiple gap junction proteins. Circ. Res. 70, 438-444. Kenne, K., Fransson-Steen, R., Honkasalo, S., and Wamgard, L. (1994). Two inhibitors of gap junction intercellular communication, TPA and endosulfan: Different effects on phosphorylation of connexin 43 in the rat liver epithelial cell line, IAR 20. Carcinogen. 15, 1161-1165. Khan-Dawood, E S., Yang, J., and Dawood, M. Y. (1996). Expression of gap junction protein connexin-43 in the human and baboon corpus luteum. J. Clin. Endocrinol. Metab. 81,835-842. Kiang, D. T., Jin, N., Tu, Z. J., and Lin, H. H. (1997). Upstream genomic sequence of the human connexin26 gene. Gene. 199, 165-171. Kirchhoff, S., Nelles, E., Hagendorff, A., Kruger O., Traub, O., and Willecke, K. (1998). Reduced cardiac conduction velocity and predisposition to arrythmias in connexin40-deficient mice. Curr. Biol. 8, 299-302. Laird, D. W., Puranam, K. L, and Revel, J.-P. (1991). Turnover and phosphorylation dynamics of connexin43 gap junction protein in cultured cardiac myocytes. Biochem J. 273, 67-72. Laird, D. W., Castillo, M., and Kasprzak, L. (1995). Gap junction turnover, intracellular trafficking and phosphorylation of con-
SECTION III Membrane Excitability and Ion Channels
nexin43 in brefeldin A-treated rat mammary tumor cells. J. Cell. Biol. 131, 1193-1203. Landesman, Y., White, T. W., Starich T. A., Shaw, J. E., Goodenough, D. A., and Paul, D. L. (1999). Innexin-3 forms connexin-like intercellular channels. J. Cell Science 112, 2391-2396. Larsen, W. J., Wert, S. E., and Brunner, G. D. (1987) Differential modulation of follicle cell gap junction populations at ovulation. Dev. Biol. 122, 61-71. Lo, C. W., Cohen, M. E, Huang, G. Y., Lazatin, B. O., Patel, N., Sullivan, R., Pauken C., and Park, S. M. (1997). Cx43 gap junction gene expression and gap junctional communication in mouse neural crest cells. Dev. Genet. 20, 119-132. Loewenstein, W. R., Kanno, Y., and Socolar, S. J. (1978). Quantum leaps of conductance during formation of membrane channels at cell-cell junction. Nature 274, 133-136. Loewenstein, W. R. (1985). Regulation of cell-to-cell communication by phosphorylation. Biochem Soc. Symp. 50, 43-58. Loewenstein, W. R. (1987). The cell-to-cell channel of gap junctions. Cell 48, 725-726. Loo, L. W., Berestecky, J. M., Kanemitsu, M. Y. and Lau, A. E (1995). pp60src-mediated phosphorylation of connexin 43, a gap junction protein. J. Biol. Chem. 270, 12751-12761. Maestrini, E., Korge, B. P., Ocana-Sierra, J., Calzolari, E., Cambiaghi, S., Scudder, P. M., Hovnanian, A., Monaco, A. P., and Munro, C. S. (1999). A missense mutation in connexin26, D66H, causes mutiliating keratoderma with sensorineural deafness (Vohwinkel's syndrome) in three unrelated families. Hum. Mol. Genet. 8, 1237-1243. Makowski, L., Caspar, D. L. D., Phillips, W. C., and Goodenough, D. A. (1977). Gap junction structures. II. Analysis of the x-ray diffraction data. J. Cell Biol. 74, 629-645. Mesnil, M., Krutovskikh, V., Piccoli, C., Elfgang, C., Traub, O., Willicke, K., and Yamasaki, H. (1995). Negative growth control of HeLa cells by connexin genes: connexin specificity. Cancer Res. 55, 629-639. Moennikes, O., Buchmann, A., Ott, T., Willecke, K., and Schwarz, M. (1999). The effect of connexin32 null mutation on hepatocarcinogenesis in different mouse strains. Carcinogenesis 20, 1379-1382. Musil, L. S., and Goodenough, D. A. (1990). Gap junctional intercellular communication and the regulation of connexin expression and function. Curr Opin. Cell Biol. 2, 875-880. Neuhaus, I. M., Bone, L., Wang, S., Ionascescu, V., and Werner, R.. (1996). The human connexin32 gene is transcribed from two tissuespecific promoters. Biosci. Rep. 16, 239-248. Neveu, M. J., Hully, J. R., Babcock, K. L., Hertzberg, E. L., Nicholson, B. J., Paul, D. L., and Pitot, H. C. (1994). Multiple mechanisms are responsible for altered expression of gap junction genes during oncogenesis in rat liver. J. Cell Sci. 107, 83-95. Neyton, J., and Trautmann, A. (1985). Single-channel currents of an intercellular junction. Nature 317, 331-335. Nonner, W., and Eisenberg, B. (1998). Ion permeation and glutamate residues linked by Poisson-Nernst-Planck Theory in L-type calcium channels. Biophys. J. 75, 1287-1305. Orsino, A., Taylor, C. V., and Lye, S. J. (1996). Connexin-26 and connexin-43 are differentially expressed and regulated in the rat myometrium throughout late pregnancy and the onset of labor. Endocrinol. 137, 1545-1553. Pal, J. D., Berthoud, V. M., Beyer, E. C., Mackay, D., Shiels, A., and Ebihara, L. (1999). Molecular mechanism underlying a Cx50-1inked congenital cataract. Am. J. Physiol. 276, c1443-c1446. Peters, N. S. 1996. New insights into myocardial arrythmogenesis: Distribution of gap-junctional coupling in normal ischaemic and hypertrophied human hearts. Clin. Sci. 90, 447-452. Piersanti, M., and Lye, S. J. (1995). Increase in messenger ribonucleic acid encoding the myometrial gap junction protein, con-
31. Biology of Gap Junctions nexin-43, requires protein synthesis and is associated with increased expression of the activator protein-l, c-fos. Endocrinol. 136, 3571-3578. Reaume, A. G., de Sousa, E A., Kulkarni, S., Langille, B. L., Zhu, D., Davies, T. C., Juneja, S. C., Kidder, G. M., and Rossant, J. (1995). Cardiac malformation in neonatal mice lacking connexin43. Science 267, 1831-1834. Revel, J.-E, and Karnovsky, M. J. (1967). Hexagonal array of subunits in intercellular junctions in the mouse heart and liver. J Cell Biol. 38, C7-C12. Revel, J.-E, Hoh, J. H., John, S. A., Laird, D. W., Puranam, K., and Yancey, S. B. (1992). Aspects of gap junction structure and assembly. Sem. Cell Biol. 3, 21-28. Saez, J., Nairn, A. C., Czemik, A. J., Spray, D. C., Hertzberg, E. L., Greengard, E, and Bennett, M. V. L. (1990). Phosphorylation of connexin32, a hepatocyte gap-junction protein, by cAMP-dependent protein kinase, protein kinase C and CaZ+/calmodulin-dependent protein kinase II. Eur. J. Biochem. 192, 263-273. Seul, K. H., Tadros, E N., and Beyer, E. C. (1997). Mouse connexin40: Gene structure and promoter analysis. Genomics 46, 120-126. Simon, A. M., Goodenough, D. A., and Paul, D. L. (1998). Mice lacking connexin40 have cardiac conduction abnormalities characteristic of atriventricular block and bundle branch block. Curr. Biol. 8, 295-298. Simpson, I., Rose, B., and Loewenstein, W. R. (1977). Size limit of molecules permeating the junctional membrane channels. Science 195, 294-296. Spray, D. C., and Dermiietzel, R. (1995). X-linked dominant CharcotMarie-Tooth disease and other potential gap junction diseases of the nervous system. TINS 18, 256-262. Swenson, K. I., Piwnica-Worms, H., McNamee, H., and Paul, D. L. (1990). Tyrosine phosphorylation of the gap junction protein connexin43 is required for the pp60v-src-induced inhibition of communication. Cell Reg. 1, 989-1002. Tu, Z. J., and Kiang, D. T. (1998). Mapping and characterization of the basal promoter of the human connexin26 gene. Biochim. Biophys. Acta. 1441, 169-181.
537 Unger, V. M., Kumar N. M., Gilula, N. B., and Yeager, M. (1999). Three-dimensional structure of a recombinant gap junction membrane channel. Science. 283, 1176-1180. Veenstra, R. D. (2000). Ion permeation through connexin gap junction channels: Effects on conductance and selectivity. In "Current Topics in Membranes," (C. Perrachia, Ed.), Vol. 49, pp. 95-129. Veenstra, R. D., and DeHaan, R. L. (1986). Measurement of single channel currents from cardiac gap junctions. Science 233, 972-974. Veenstra, R. D., Wang, H.-Z., Beblo, D. A., Chilton, M. G., Harris, A. L., Beyer, E. C., and Brink, E R. (1995). Selectivity of connexinspecific gap junctions does not correlate with channel conductance. Circ. Res. 77, 1156-1165. Veenstra, R. D., Wang, H.-Z., Westphale, E. M., and Beyer, E. C. (1992). Multiple connexins confer distinct regulatory and conductance properties of gap junctions in developing heart. Circ. Res. 71, 1277-1283. Waldo, K. L., Lo, C. W., and Kirby, M. L. (1999). Connexin43 expression reflects neural crest patterns during cardiovascular development. Dev. Biol. 208, 307-323. Warn-Cramer, B. J., Lampe, P. D., Kurata, W. E., Kanemitsu, M. Y., Loo, L. W., Eckert, W., and Lau, A. F. (1996). Characterization of the mitogen-activated protein kinase phosphorylation sites on the connexin-43 gap junction protein. J. Biol. Chem. 271, 3779-3786. White, T. W., Goodenough, D. A., and Paul, E L. (1998). Targeted ablation of connexin 50 in mice results in micropthalmia and zonular pulverulent cataracts. J. Cell Biol. 143, 815-825. Wohlburg, H., and Rohlmann, A. (1995). Structure-function relationships in gap junctions [review]. Int. Rev. Cytol. 157, 315-373. Yamaguchi, D. T., Huang, J. T., and Ma, D. (1995). Regulation of gap junction intercellular communication by pH in MC3T3-E1 osteoblastic cells. J. Bone Min. Res. 10, 1891-1899. Yamasaki, H. (1991). Aberrant expression and function of gap junctions during carcinogenesis. Env. Health Prosp. 93, 191-197. Yu, W., Dahl, G., and Werner, R. (1994). The connexin43 gene is responsive to oestrogen. Proc. Roy. Soc. (London) Series B: Biol. Sci. 255, 125-132.
Louis I. De Felice and Michele Mazzanti
Biophysics of the Nuclear Envelope
!. Introduction Eukaryotic cells contain a nucleus, a spherical body that encloses characteristic organelles. These organelles include a thin nuclear envelope, one or more nucleoli, chromatin (the chromosomes, DNA attached to proteins, primarily histones), a proteinaceous nuclear lamina (which contacts the inner nuclear membrane, chromosomes, and the nuclear RNA), irregular granules of chromatin material, and linin (fine threads that associate with the chromatin granules). Amorphous nucleoplasm occupies the rest of the nuclear volume. The nuclear envelope transports macromolecules, and the transport selectivity changes during the cell cycle and throughout development. Most textbooks describe the envelope as forming no barrier to small ions such as Ca 2+, Na +, K +, or C1-. In this view, the nucleus is a sievelike structure containing large aqueous pores. However, certain data suggest that the nuclear envelope can restrict the movement of intermediate-sized molecules, including peptides, amino acids, and sugars. Moreover, although the data are controversial, particular experiments support the idea that inorganic ions like Ca 2+ can accumulate in, or be excluded from, the nucleus. Furthermore, the nucleus has an electrical potential, and it can swell or shrink in response to osmotic forces. One theory is that such phenomena derive from selective absorption by the nucleoplasm. This chapter explores another view, that the nuclear envelope selectively transports small ions and has electrical properties similar to the semipermeable plasma membrane. The nuclear pore governs the movement of nucleic acids and large-molecular-weight proteins into and out of the nucleus: enzymes enter the nucleus to polymerize nucleic acids, RNA leaves for translation in the cytosol, and DNA-binding proteins penetrate for transcription and gene regulation. Selection is precise, and the nucleus eliminates from its domain the majority of proteins, many of which are smaller than the proteins it admits. During singular events, such as fertilization or viral infection, or under laboratory manipulations, such as gene transfection, the nuclear pore exhibits transitory
Cell Physiology Sourcebook: A Molecular Approach, Third Edition
5 3 9
selection. This chapter explores the possibility that ion selection accounts for the electric and osmotic properties of the nucleus. The movements of ions across the nuclear envelope may act as regulatory counter currents for the transport of charged macromolecules. This assertion derives from the biophysical characteristics of nuclei and the presence of ionselective channels in the nuclear envelope. Ion channels in the nuclear envelope have properties similar to ion channels in the plasma membrane. We further suggest that some channels observed in the envelope are part of the nuclear pore complex. This proposition introduces a paradox: how could a pore that transports large molecules through a large opening also select small ions? This chapter addresses the implications of this question and offers an explanation of its consequences.
11. Permeability of the Nuclear Envelope Two mechanisms may explain the separation of molecules by the nuclear envelope. One mechanism involves specific solubility in the nucleoplasm (or exclusion from the cytoplasm). Separation of solutes between adjacent phases falls under the heading of the Donnan equilibrium. The Donnan equilibrium relies on the differential absorption of solutes into immiscible phases. If charged molecules partition into adjacent compartments, an electric potential difference results. If the nucleoplasm and the cytoplasm form such domains, and if they connect through open pores, the Donnan equilibrium could explain the nuclear resting potential. Another mechanism involves a semipermeable membrane that separates the nucleoplasm from the cytoplasm. This mechanism, which characterizes cell plasma membranes, comes under the heading of the Nernst-Planck regime. Whereas one lipid bilayer forms the plasma membrane that surrounds the cell, two such membranes form the envelope that surrounds the nucleus. However, the double-membrane structure simplifies because nuclear pores span the envelope and act as channels between nucleoplasm and cytoplasm. An early model of the double-membrane-spanning nuclear pore depicts the pore as two parallel rings, one on the
Copyright 9 2001 by Academic Press. All rights of reproduction in any form reserved.
540
cytoplasmic side and the other on the nucleoplasmic side. Each ring consists of eight subunits that surround a central pore. We refer to this structure as the nuclear pore complex. The phrase nuclear pore complex applies to the entire structure, not merely to the opening in the center. The opening in the center (the nuclear pore) contains an electrondense central granule that connects the two rings. The size and shape of the central pore help determine its selectivity to macromolecules. For example, RNA molecules elongate into cylindrical rods (diameter N100 /~) as they enter the pore, and certain proteins larger than 100/k in diameter cannot pass through the pore. However, the size of the pore does not entirely explain its selectivity, because 170 ,~, diameter gold particles coated with nucleoplasmin can enter the nucleus. Without nucleoplasmin, or after treatment with trypsin to remove nucleoplasmin, the large gold particles cannot enter. Therefore it is thought that certain proteins contain a signal that permits nuclear transport. This signal is specific, because a single amino acid substitution can abolish, for example, the transport of specific tumor antigens into the nucleus. Furthermore, a single base substitution can dramatically reduce the rate of tRNA transport. Some substances apparently block transport merely by getting in the way, for example, wheat germ agglutinin (WGA) accumulates on the cytoplasmic face of the nucleus and inhibits the uptake of nucleoplasmin. Such blocking cannot explain all the data, because WGA does not influence the rate of dextran entry. Thus nuclear transport is complex and highly varied and is unlikely to be explained by one mechanism. The high rates of tRNA translocation in mammalian cell nuclei (109 molecules/min) exclude simple diffusion as the transport mechanism. Some experiments suggest a carrier mechanism for macromolecules, augmented by the observation that RNA transport requires ATP. Although diffusion accounts for only a small fraction of macromolecular transport, smaller molecules are translocated at rates below diffusion. For example, inulin (5.5 kDa) enters the nucleus at one-fifth its diffusion rate in water. Another example is sucrose, which passes through the nuclear envelope more readily than it does through the plasma membrane, but at one-third the rate predicted by diffusion. Furthermore, nuclei may even restrict small inorganic anions and cations. Evidence for this comes from Drosophila salivary gland cells, whose nuclei maintain an electrical potential o f - 1 5 mV with respect to the cytoplasm. Electrical potentials exist across other nuclear envelopes, although not always of the same sign or amplitude. Finally, Na + and K + ions have different concentrations in the nucleoplasm than in the cytoplasm, and nuclear Ca 2§ levels change during cell differentiation and fertilization. These data are controversial, and not everyone agrees that ion gradients and electrical potentials exist across the nuclear envelope. Even if they do, it is unclear whether the nucleus restricts ions via the Donnan equilibrium or whether the nuclear envelope acts as a semipermeable membrane. These mechanisms are not exclusive; however, if some selection of small ions does occur via a semipermeable membrane, how can we reconcile this phenomenon with the passage of large macromolecules through the same openings? The next section offers a possible explanation.
SECTION III Membrane Excitability and Ion Channels
II1. Structure of the N u c l e a r E n v e l o p e Two lipid bilayers surround the nucleus, and in this regard, the nucleus resembles a mitochondrion. In contrast, a single lipid bilayer surrounds other internal organelles, such as the Golgi apparatus, storage vesicles, and the endoplasmic reticulum. The double membrane structure of isolated nuclei remains intact in vitro and, in many ways, isolated nuclei may be treated as cells maintained in culture.
A. Isolation of Nuclei Nuclei from invertebrate cells are easily dissected from starfish oocytes in the germinal vesicle stage, when the nuclei are large (30 pm in diameter). After the dissection, the nuclei can be kept for hours at cold temperatures (5 ~ in an intracellular-like solution (200 mM K2SO 4, 20 mM NaC1, 10 mM Hepes, 10 mM EGTA, and sucrose to 1000 milliosmol). Nuclei are tolerant to other physiological solutions, such as extracellular media and seawater. M a m m a l i a n cell nuclei (10 txm in diameter) are readily obtained from mouse oocytes 12 h after superovulation. Upon fertilization, pronuclei or early embryo nuclei are collected 12 h after mating. Prior to removing the nucleus, the oocytes, zygotes, or embryos should be placed in an intracellular-like solution (120 mM KC1, 2 mM MgCI 2, 1.1 mM EGTA, 0.1 mM CaC12, 5 mM glucose, 10 mM Hepes, pH 7.4) so that when the nucleus comes out of the cell it experiences a normal ionic environment. Mammalian nuclei can also be isolated from differentiated tissue, such as liver cells, by shearing the tissue and centrifuging the homogenate. This results in a pellet of pure nuclei, which can then be re-suspended into a physiological solution. These techniques result in four categories of nuclei that we have used for comparative studies: germinal vesicles, pronuclei, two-cell embryo nuclei, and adult liver cell nuclei. It is unclear whether nuclei preserve their characteristics in culture conditions; however, zygote enucleation results in pronuclei that have the same resting potential as in vivo. Mouse liver nuclei removed by centrifugation have zero resting potential, but it is uncertain whether this condition comes from the isolation procedure. B. Electron Microscopy of Nuclei The procedures for obtaining the photographs and electron micrographs of nuclei become important when evaluating nuclear pore density. The dissected starfish nuclei were fixed in 1% glutaraldehyde, postfixed in 1% osmium tetroxide, dehydrated in alcohol, and dried in CO 2. The nuclei were then coated with gold for scanning electron microscopy. The mouse nuclei were fixed in 1% glutaraldehyde and 1% tannic acid, post-fixed in OsO 4, and dehydrated in alcohol, and 400-500 A sections were stained with uranyl acetate and lead citrate. These dehydration and fixation procedures may introduce shrinkage and distortion of the nuclear envelope, causing an apparent density of nuclear pores that exceeds the in vitro density.
32. Biophysics of the Nuclear Envelope
Figure 1 is an electron micrograph of three mouse oocytes packed together. Only a segment of each oocyte appears. The single darkly stained line, designated as plasma membrane (pm), represents the bilayer that surrounds the oocyte. The plasma membrane encompasses the cytoplasm (cy), which includes internal organelles such as mitochondria (mt) and the endoplasmic reticulum (er). The term cytosol (cy) refers to the cytoplasmic volume, excluding the organelles. The designation (cy) will stand either for the cytoplasm or the cytosol, depending on the context. In Fig. 2, parallel lines represent the lipid bilayer membrane. Figure 2 depicts the extracellular volume (ex), the cytosol (cy), the single-membrane-enclosed endoplasmic reticulum (er), and the double-membrane-enclosed nucleus (nu). Openings that penetrate the double membrane connect the nucleoplasm with the cytoplasm. The membranes that encircle the nucleus enclose a volume topologically equivalent to the lumen of the endoplasmic reticulum. This restricted volume has a particular name, the cistern (cs). In effect, cisterns are specialized organelles bound by one membrane. The membrane that defines the cistern is further categorized into the outer membrane and the inner membrane. Because the inner and outer membranes are close together, they are given the cumulative name, nuclear envelope (ne). Figure 3 indicates that cistern membranes and endoplasmic reticulum membranes may join one another. However, cisterns may exist without any connection to the endoplasmic reticulum. The anatomy of the nucleus and its associations to cell structures varies with the developmental stage or the specialization of the cell. At some stages, the endoplasmic reticulum is virtually absent. Even in cells that have a well-developed endoplasmic reticulum, large regions of the envelope do not connect with the endoplasmic reticulum. Where they do connect, cisterns adjoin the endoplasmic reticulum via the outer membrane. The endoplasmic reticulum is either smooth or rough, depending on the absence or presence of ribosomes (dots in Fig. 3). The outer membrane of the nuclear envelope is also smooth or rough. The ribosomes on the endoplasmic reticulum are evident in the photograph in Fig. 1; if there are ribosomes on the outer membrane, they are not obvious. The openings between the cisterns depicted in Figs. 2 and 3 are more than mere breaches in the nuclear envelope. They contain the nuclear pore complex. The nuclear pore complex is sometimes referred to simply as the nuclear pore (np). The reader must be careful, because "nuclear pore" also refers to the large central opening in the nuclear pore complex.
IV. Structure of the Nuclear Pore The nuclear pore complex has a molecular weight of about 100 MDa, and it is comprised of nearly 100 gene products. The complex consists of eight symmetrical units that form a large central opening. We note the following difference between pore-forming proteins in the nuclear pore and poreforming proteins in the plasma membrane: ion channels, gap junction proteins, receptor subunits, and neurotransmitter transporters span a single bilayer. These we call transmembrane (TM) proteins. Some nuclear-pore proteins also span a
541
single bilayer, but with a much different relationship to the flow of material. To illustrate, consider the structure of gap junctions. This example has special relevance to our topic because nuclear pore proteins and gap junction proteins both form openings through abutting membranes. Figure 4 illustrates membrane-spanning proteins of the nuclear pore complex, which insert at the division of the inner and outer membrane. Eight units (for simplicity, only two appear in Fig. 4), each composed of many proteins, form the central opening (center arrow, top drawing in Fig. 4). In addition, eight lateral openings exist (side arrows in Fig. 4 indicate two lateral openings). In gap junctions, six pairs of proteins (only one pair appears in Fig. 4, bottom drawing) define the opening. The distinction may seem artificial, because gap-junction proteins (connexins), as well as the subunits of ion channels and receptors, form parallel structures. However, the nuclear envelope has a distinct structure-function relationship between the TM proteins and transport. As far as we know, the functional connections made within the nuclear pore do not include cisternto-cistern communication, but only cytoplasm-to-nucleoplasm communication. Figure 5 illustrates the three-dimensional structure in more detail. The dark regions in the Fig. 5 reconstruction are the TM domains of the eight repeated units. These units also define eight additional pathways (as indicated in Fig. 4 by the lateral arrows). These pathways surround the central opening. We refer to this structure as eight-plus-one configuration, in analogy with the axoneme nine-plus-two nomenclature. Surprisingly, in a minority of cases more than eight units can form a nuclear pore complex. The significance of these higher order structures is unknown. Figure 6 defines the proposed relationship between the nuclear pore (np), the outer membrane (om), and the inner membrane (im) of the nuclear envelope (ne). The inner and outer membranes form the connecting loops that define openings between cisterns. The crosshatched regions in the drawing indicate the lateral openings that exist within the complex. Figure 6 also shows a feature mentioned earlier~ the nuclear granule (or nuclear plug), illustrated as spoolshaped, that resides in the central opening. Recent work indicates that the nuclear plug moves within the pore to regulate transport in response to cistern Ca 2+. It has been suggested that the nuclear plug may include newly synthesized ribosomes. The nuclear pore complex therefore, consisting of a large number of proteins, some of which are embedded in lipid bilayers, forms an eight-plus-one assembly that spans the double membrane. Some nuclei are virtually covered with nuclear pores. The high density of pores raises several interesting questions. What is the physical chemistry of the lipid/protein ratio in dense arrays? How is it possible to patch-clamp such arrays? In nuclei with a high density of pores, the relative amount of outer lipid membrane diminishes, and connections to the endoplasmic reticulum become sparse. Pore density varies between 1 and 120 per lum2, depending on the specialization of the cell, and local pore densities may differ within the same nucleus. Figure 7 illustrates the patch-clamp technique applied to isolated nuclei. In this technique, a glass electrode isolates a
542
SECTION III Membrane Excitability and Ion Channels
32. Biophysics of the Nuclear Envelope
FIGURE 2. Cell spaces. The diagram defines the volumes within and around a cell: the extracellular space (ex); the cytosol (cy); the endoplasmic reticulum (er); and the nucleus (nu). The double lines represent lipid bilayers.
nO
/
543 patch of membrane on the outer surface of the nucleus. The applied voltage V (shown as a battery in the diagram) appears at the electrode tip, and the current meter (I) records the flow of ions through the patch. An isolated patch is 1-5 ~am2 in area. The voltage across the envelope is the difference between the nuclear resting potential and the pipette potential. This voltage also appears across the nuclear pores in the patch if the bath and the pipette solution are at ground potential. The possibility exists that channels in the outer and inner membranes connect the pipette solution to the nucleoplasm. This hypothetical pathway through the cistern would be in parallel with the nuclear pore. Let us now consider an actual patch-clamp experiment on a nucleus with a high density of pores. Figure 8 shows a scanning electron micrograph of a starfish oocyte at the germinal vesicle stage. The nucleus is ~30 lam in diameter, larger than many specialized cells. The inset shows an isolated area on the outer m e m b r a n e about 1 Ixm in diameter. The diameter of a nuclear pore complex is -~1000 A (0.1 pm), which implies 100 nuclear pore complexes side-by-side in 1.0 lam 2 of membrane. Each doughnut-shaped object in the expanded field in Fig. 8 represents one nuclear pore complex; thus the pores cover the surface. Thus even if shrinkage has occurred during the fixation procedure, it seems likely that a nucleus-attached patch on a nucleus in this preparation contains dozens of pores.
V. Electrophysiology of the Nucleus
FIGURE 3. The nuclear envelope and the endoplasmic reticulum. This illustration shows the relationship between the outer membranes (om), the inner membranes (im), and the endoplasmic reticulum (er). The inner and outer membranes and the nuclear pore complexes (np) define the nuclear envelope. The central egg-shaped structure inside the nucleus represents the nucleolus. The dots along the outer membrane and on the endoplasmic reticulum represent ribosomes. The cistern (cs) of the nuclear envelope and the lumen of the endoplasmic reticulum form a continuous compartment within the cell.
If nuclei like the one shown in Fig. 8 are dispersed in a dish, they can be manipulated and prepared for electrophysiology as if they were cells. As mentioned above, some nuclei have resting potentials. For example, in a mouse zygote, the in situ pronuclei have a resting potential o f - 1 0 m V with respect to the cytoplasm. If we eliminate the resting potential of the zygote plasma membrane (see Section II, Chapter 14), no immediate change occurs in the nuclear potential. We expect this, because the nuclear potential appears across the envelope; altering a voltage in series should not affect it. For excised pronuclei in an intracellular-like solution (120 m M K§ the nuclear potential has the same value it had in the cell. Lowering K § concentration in vitro to 60 m M K § changes the nuclear potential to - 2 0 mV. Table 1 summarizes experiments on in vivo and
FIGURE 1. Structure of the nucleus. Transmission electron micrograph of three mouse oocytes at the germinal vesicle stage. The plasma membrane (pm) surrounds the cytoplasmic compartment (cy), and the nuclear envelope (ne) surrounds the nuclear compartment (nu) within the cytoplasm. The micrograph demonstrates two additional membrane-bound compartments in the cytoplasm, mitochondria (mt) and the endoplasmic reticulum (er). The endoplasmic reticulum has most likely swollen during the fixation procedure. The nuclear envelope consists of a double-membrane structure interrupted by nuclear pores (np). The double membranes, designated as the inner membranes and the outer membranes (see Fig. 3), define a separate cytoplasmic compartment called the cistern (cs). The cistern may be continuous with the endoplasmic reticulum (see Fig. 3). The nuclei contain darkly stained material adjacent to the inner membrane, called nuclear lamina, and they contain the diffuse darkly stained material called chromatin. Samples were dehydrated with alcohol to propylene oxide, infiltrated and embedded in epon, thin sectioned, and stained in uranyl acetate and lead citrate. The bar length indicates one micron, and the bar width indicates 1000 angstroms. (From Mazzanti et al., J. Memb. Biol. 121:189-198 (1991).)
544
F I G U R E 4. Nuclear pores versus gap junctions. A comparison between the structure and function of the nuclear pore (top) and the gap junction (bottom). The nuclear pore complex contains eight membrane-spanning subunits (two are shown here). Each subunit is a multiprotein structure that repeats eight times. Material flows through a central pore (the center arrow in the top figure) defined by the surrounding subunits. In addition, lateral nucleocytoplasmic pathways are located within each subunit (see Fig. 5). The shorter arrows in the top drawing illustrate the lateral pathways. The complete nuclear pore complex contains eight lateral pathways surrounding the central pore. Gap junctions have a total of six pairs of abutting membrane-spanning proteins. One such pair is illustrated in Fig. 4 (bottom). Apart from a different number of repeating subunits, gap junctions also have a different structural relationship to function when compared to nuclear pore complexes.
SECTION III Membrane Excitability and Ion Channels
F I G U R E 5. Structure of the nuclear pore. A three-dimensional reconstruction based on an electron density map of the 100-MDa structure that forms the nuclear pore complex. The entire structure has an external diameter of about 1300/~. The model proposes eight lateral channels that surround a large central pore. The central pore normally contains a proteinaceous plug that occupies most of the opening (see Fig. 6). The central plug is absent in Fig. 5. The reconstruction is based on electron microscopy and image analysis of detergent-released nuclear pore complexes from macronuclei of mature Xenopus laevis oocytes. [From Hinshaw et al., Cell 69:1133-1141 (1992). Copyright 1992 by Cell Press.]
in vitro nuclei u n d e r different ionic conditions. T h e data s u g g e s t that the n u c l e a r e n v e l o p e selects K § ions. Different conditions m a y lead to other potentials. Frog o o c y t e and adult liver cell nuclei have no apparent resting potential, and starfish nuclei can m a i n t a i n a positive potential with respect to the bath.
A. Measurement of Resting Potential To detect the nuclear resting m e m b r a n e potentials in intact cells requires a conventional glass microelectrode of high resistance (70 to 90 MI2 w h e n filled with 3 M KC1). In such experiments, we penetrate both the outer and inner m e m b r a n e s . Thus the nuclear potential m e a n s the voltage across the entire d o u b l e - m e m b r a n e nuclear envelope. Ion channels exist in both the outer and inner m e m b r a n e s . For example, C1- channels have been reported on the surface of nuclei, as well as ligand-gated channels. However, no measurements exist from isolated cisterns, and we do not know the electrical potential of the cistern with respect to the cytoplasm or nucleoplasm. For in situ experiments on the nucleus, cells that are m a i n t a i n e d in physiological solutions
F I G U R E 6. The nuclear pore and the nuclear envelope. Illustration of the position of the nuclear pore (np) within the doublemembrane structure that envelopes the nucleus. The two planar membranes illustrated in the diagram represent the outer membrane (om) and the inner membrane (im). The interconnecting subunits that make up the supramolecular nuclear pore complex insert into the lipid bilayer at the border between the inner and outer membranes. These subunits form eight channels that surround the main opening of the pore (see Fig. 5). The main opening contains a proteinaceous granule or "plug" (not shown in Fig. 5), which we have pictured as a nearly space filling, spool shaped object. Thus two kinds of pathways connect the cytoplasm and the nucleoplasm: the middle pore that contains the plug, and the eight encircling channels that riddle the proteins that form the central pore.
32. Biophysics of the Nuclear Envelope
545
Y F I G U R E 7. Patch-Clamp Technique. An illustration of a nucleusattached patch-clamp experiment. The glass pipette contains a physiological solution connected by a Ag/AgC1 electrode to a current meter (I) and a battery. The battery supplies a voltage, Vp, to the pipette that ideally appears at the outer membrane of the nucleus. The nucleus bathes in a physiological cytoplasm-like solution connected to ground. The potential across the nuclear envelope is the resting potential of the nucleus minus Vp.
(133 mM NaC1, 2 mM KC1, 1.5 mM CaC12, 0.5 mM MgC12, 10 mM Hepes, 5 mM dextrose, pH 7.4) have normal resting potentials. To eliminate the cell resting potential, we bathe in a solution that mimics the cytoplasm (120 mM KC1, 2 mM MgCI2, 1.1 mM EGTA, 0.1 mM CaCI2, 10 mM Hepes, 5 mM dextrose, pH 7.4). Among the technical difficulties associated with measuring the nuclear resting potential, electrode tip potential gives the most problems. The structures that compose the nuclear envelope, the inner and outer membranes and the lamina may stick to the microelectrode and increase resistance and tip potential. To guard against such errors, check that the tip potential (the electrode voltage with the tip in the bath solution) returns to the baseline upon removal. This does not prove that the observed potential comes from the nuclear envelope, but it is an essential first step. Similar technical difficulties related to the value of nuclear potentials arise in measuring plasma membrane potentials.
A patch-clamp experiment on the nucleus results in singlechannel currents similar to those observed in the plasma membrane of most cells. Nuclear envelope channels were first observed in mouse oocytes. Figure 9 shows the results of a patch-clamp experiment on a starfish oocyte nucleus (as in Fig. 8). In Fig 9 9, Vp is the voltage applied to the pipette (the battery in the external circuit, Fig. 7). The channel conductance in this experiment was about 100 pS. Not only longlasting openings, but also numerous fast openings occur. These are too brief to resolve in this record. The downward deflections in Fig. 9 indicate positive electric current flowing into the nucleus. The downward deflection could also result from negative ions moving out of the nucleus. Mouse oocyte nuclei have a lower pore density than starfish germinal vesicles; nevertheless, they display similar ion channel activity. The nuclear pore density depends on the tissue and on the stage of development of the cell, as we have stated. The freeze-fracture image shown in Fig. 10A is from an oocyte germinal vesicle, and that in Fig. 10B is from an adult liver cell nucleus. The pore density is 3-10 per ~m 2 in Fig. 10A and 14-18 per lum2 in Fig. 10B. Recordings from such surfaces show channels similar to those shown in Fig. 9. Figure 11 summarizes results from four different kinds of nuclei at various stages of cellular development: the mouse germinal vesicle, the pronucleus, a two-cell embryo nucleus, and a differentiated adult cell. In Fig. 11, the current is upward because the applied potential is negative. We may conclude from experiments on starfish oocyte nuclei and mammalian cell nuclei that channels in the nuclear envelope are a ubiquitous phenomenon 9 In addition, Fig. 11 shows for embryonic preparations (A, B, C) only one or two conductance levels, while in the adult preparation there are consistently more levels, as illustrated in the histograms to the right of the figure. Combining morphological and functional data, we have hypothesized a relationship between nuclear pores and ion channels. Specifically, the channels could represent an ionic pathway through the nuclear pore complex itself.
B. Measurement of Single Channels from Nuclei In mouse liver cell nuclei, about 50% of the attempts to patch succeed. Seals up to 50 GI) are obtained using pipettes having an external tip diameter of about 1 txm and a resistance between 5 and 10 MI). In the single-channel experiments, 120 mM K + (cytoplasmic-like) solutions may be used to fill the electrode and to bathe the nucleus. To investigate the ionic selectivity of the nuclear envelope, NaC1 may replace KC1. Single-channel currents can be recorded with a standard current-to-voltage converter (List EPC-7). Positive pressure in the pipette (-~2 cm H20 ) keeps the electrode solution from mixing with the bath solution, flattens and cleans the nuclear surface, and helps form the seal. For unknown reasons, a negative potential (---Vp= - 4 0 mV) applied to the electrode tip before touching the nucleus facilitates seal formation. Such negative potentials help form seals on most cells. After touching, a slight increase in electrode resistance can be noticed: the pressure can be reversed until a tight seal occurs, and then the suction can be released.
VI. O s m o t i c Effects in the N u c l e u s Figure 12 shows that the nucleus undergoes changes in response to osmotic forces. However, these changes are not the simple osmotic effects expected from solution chemistry. In this experiment, the ionophore nystatin was used to make the plasma membrane permeable to monovalent cations. The extracellular bath and the cytoplasm are in electrical communication through nystatin channels. Adjusting bath osmolarity changes nuclear volume. Such permutations of the volume are difficult to see and even more difficult to quantify. Some responses occur in the way we would expect from osmotic changes: in hypertonic solutions the nucleus shrinks, and in hypotonic solutions the nucleus swells. However, too great an osmotic shock produces different effects, some of which are irreversible: if the nucleus shrinks in external 480 mM KC1, it subsequently swells irreversibly to the configuration shown in
546
SECTION III Membrane Excitability and Ion Channels
F I G U R E 8. Scanning electron micrograph of a nucleus from a starfish oocyte. The figure shows a scanning electron micrograph of the surface of a germinal-vesicle-stage nucleus removed from a starfish oocyte (Mathasterias glacialis) by manual dissection. The diameter of the nucleus is approximately 30 txm. The diameter of the inset, which gives a magnified view of the outer membrane, is approximately 1 txm. Large areas of the external face of the nucleus appear relatively free of extraneous material. The whitish, globular matter is cellular debris stuck to the surface after dissection. The exterior surface of the dissected nucleus appears studded with numerous doughnut-shaped particles that we have interpreted as nuclear pore complexes (average density 60-110 per ~xm2). This anatomy makes it impossible to patch the starfish nuclear envelope without encountering numerous NPC. (Courtesy of Luigia Santella, Stazione Zoologica, Naples, Italy.)
32. Biophysics of the Nuclear Envelope
547
TABLE 1 Results of Experiments on Nuclei Subjected to Different Ionic Conditions (1) 2 K/133 Na (2) 120 K/0 Na (3) 120 K/0 Na (4) 60 K/60 Na cy
nu
cy
nu
nu
nu
-23
-11
0
-10
-10
-28
-18
-12
2.5
-9.5
-7
-18
-28
-9
3
-9
-9
-23
-25
-11
1
-11
-12
-25
Note: The ratios of K to Na (in mM) refer to extracellular (bath) concentrations. In columns (1) and (2) (zygote pronuclei in situ), the voltages (in mV) are as follows, cy = the potential of the cytoplasm with respect to the extracellular solution nu = the potential of the nucleoplasm with respect to the cytoplasm. In columns (3) and (4) (pronuclei in vitro), the voltages are the nuclear potential with respect to the bath.
Fig. 12D. The swelling that occurs in 480 mM KC1 is due in part to internal structural changes in the chromatin. Distended nuclei remain stable for up to 30 min. Returning the nucleus to 120 mM KC1, after swelling with 480 mM KC1, causes the nucleus to disrupt. With V = 0 and 120 mM KC1 in the patch and the bath, channels in the patch carry small inward currents. After perfusion with 480 mM KC1 and waiting 5 min for the nucleus to swell, the direction of the current changes from inward to outward, and the amplitude increases. In 120 mM KC1, the conductance of the channels is 230 pS, as in the pronucleus. In 480 mM KC1, the conductance is 510 pS, and their reversal potential shifts negative by 33 mV. The size and direction of the potential shift indicate that the channel is K+-selective. The existence of K+-selective channels would explain the nuclear resting potential discussed above. More importantly, these results imply that ion channels in the patch span the double membrane, because the current pathway must join the pipette with the bath. It is conceivable that channels in the patch reside in the outer membrane, and that other channels in the inner membrane complete the circuit. Because the only structure known to connect cytoplasm to nucleoplasm is the nuclear pore, and because the patch contains many pores, it is likely that some structure within the nuclear pore complex itself forms the observed channels.
could play a major role in the transport of macromolecules. We will now apply to the nucleus a well-known approach from cellular physiology. The flux of any substrate through a boundary that sustains diffusional and electrical forces is -do-
dn dV u k T -~x + z e n u --~x
( 1)
The minus sign comes about because flux occurs down a negative gradient (i.e., downhill). The symbol 9 represents the flux of the substrate, that is, the number of molecules that move per unit area per unit time. We use the symbol o
Vp (mV) 120
100
80
60
40
20
20 c--------n
100 rn
VII. Electrical and Diffusional Forces across the Nuclear Envelope Solution chemistry is insufficient to explain the complex phases that make up nucleoplasm and cytoplasm. Nevertheless, the data suggest that the nuclear envelope can sustain electric and diffusion gradients similar to those that exist across the plasma membrane. In particular, channels recorded from the nuclear surface respond to transenvelope voltages, and the nucleus changes shape and volume with osmolarity. Thus we consider the hypothesis that the nuclear envelope acts as a barrier to the diffusion of small ions. Such gradients
FIGURE 9. Nucleus-attached patch recordings from starfish germinal vesicle. The figure shows sample currents recorded from the surface of a germinal-vesicle-stage nucleus isolated from a starfish oocyte similar to Fig. 8. Channel openings appear as downward deflections, which signify positive charges flowing into the nucleus. Two categories of openings appear: brief inward currents barely resolved on this time scale, and longer openings that last several hundred milliseconds. In some traces (e.g., at 60 mV), simultaneous openings occur giving two current levels of the same amplitude. The voltages (Vp) at the side of each trace indicate the pipette potential with respect to the bath. The nucleus-attached electrode contains 200 mM K S O 4, 20 NaC1, 200 sucrose, 10 EGTA, 10 Hepes (pH 7.0) and the bath solution is comprised of natural seawater at room temperature. (From Santella et al., 5th Intl. Conf. Cell Biol., Madrid (1992).)
5411
SECTION III Membrane Excitability and Ion Channels
F I G U R E 10. Nuclear envelope freeze fracture from mouse oocyte and adult mouse liver cells. The figures show scanning electron micrographs of a fracture through the outer membrane of the nuclear envelope of (A) mouse oocyte nucleus and (B) adult mouse liver cell nucleus. The doughnut shaped objects represent the nuclear pore complexes (NPC). Note the different pore density in the two preparations. The bar length indicates 1 micron. (Modified from Innocenti and Mazzanti, J. Memb. Biol. 131:137-142 (1993).)
to stand for "has the units of." Thus 9 o #]cm2-s, n(x) is the substrate concentration, and n o #/cm 3 at position x, V(x) is the voltage (mV) at position x, u is the mobility of the substrate through the barrier, and u e, (velocity/force), k is the Boltzmann constant, T is the absolute temperature in kelvins, z is the valence, and e is the electronic charge. These latter quantities will appear as the ratio
tractable and is usually a free parameter. The flux equation (in its differential form, Eq. 1, or in its integral form, Eq. 2) is called the Nernst-Planek equation. If we assume V(x) to be a linear function of x, the Nernst-Planck equation leads to the G o l d m a n equation. Rather than assume a particular V(x), we leave the integral undetermined and replace it by the symbol L
L - I e a V(x) dx
a = ze/kT= z ~ / R T = z/25 mV -1 at 23 ~ where ~ is the Faraday constant and R is the gas constant. By considering the convention in Fig. 2 and integrating the flux equation between the nucleoplasm (nu) and the cytosol (cy), the expression for the flux becomes r/nu e-a V~. _ n e-a Wcy - ~ = ukT cy (2)
I eaV(x)dx This assumes that u is constant and flux is in a steady state (no buildup or rundown of substrate; thus 9 is constant). Note that D = ukT, where D is the diffusion constant of the substrate. In the derivation of Eq. 2, no assumption was made about the spatial distribution of the voltage or concentration gradients, V(x) or n(x) except that they are constant. Only the values of n and V far from the envelope come into play. An unevaluated integral appears in the denominator of Eq. 2. To integrate this term, V(x) must be known, the shape of the voltage gradient across the barrier; however, V(x) is never known exactly and, in any case, it varies during transport. The term f e aV(x) dx is therefore in-
(3)
where L o length (cm) and ukT/L = D/L o velocity (cm/s); we call this ratio the permeability p, where p o (cm/s). Thus we can write for the flux of any substrate --~-P
( nnu e-aVnu-n cy e-aVcy)
(4)
where n nu is the concentration of the substrate in the nucleus and n cy is its concentration in the cytoplasm. For a further discussion see ch. 2 in DeFelice (1981) and DeFelice (1997). How would a theory used for cell physiology apply to the transport of macromolecules across the nuclear envelope? To answer this, we overlook the double membrane and consider the nucleus as a cell within a cell (Fig. 13). This ignores the anatomy of the cisterns and replaces the double membrane with a single boundary penetrated by nuclear pores. Let us first consider a carrier model. For specificity, suppose that K + ions cannot get through the envelope unless they combine with the nuclear pore complex
32. Biophysics of the Nuclear Envelope
549
A
50-
O~ 5
,'5
'
e's
5
l's
'
e's
B 300 -
0 ~
C 200 -
o. 5
D
500
5pA 0 ,
5
15
,
25
' pA
1 sec
F I G U R E 1 1. Nucleus-attached patch recordings from mouse nuclei during development. The figure shows large-conductance ion channels that appear on the surface of (A) germinal-vesicle-stage nucleus from a mouse oocyte, (B) pronucleus from a mouse zygote, (C) nucleus from two-cell mouse embryo, and (D) nucleus from an adult mouse liver cell. Amplitude probability histograms appear to the right of the representative trace. The traces and histograms result from stepping the pipette voltage from 0 mV to -25 mV. In this experiment, the bath and the nucleus-attached patch pipette contained an intracellular-like solution (120 mM KC1, 2 mM MgC12, 1.1 mM EGTA, 0.1 mM CaC12, 5 mM glucose, 10 mM Hepes, pH 7.4) at room temperature. (From Mazzanti e t al., J. M e m b . B i o l . 121:189-198 (1991).)
(np), and eight ions must interact before transport can occur. We select eight because of the geometry of the complex. Thus we expect eight (or a multiple of eight) interactions between the complex and the ions. Under this assumption, we write 8K + + P = K8P
[K]8 [P] Kkp-[K8P ]
where P stands for the nuclear pore complex and K,Kp. is the equilibrium constant for the reaction. Next consider a macromolecule, M, that undergoes a similar interaction with P, but one molecule at a time M + P - ~ MP K
(5)
= [MI[P] mp
(6)
[MP]
K m is the equilibrium constant for the reaction. Assume that K8~ and MP can move across the envelope, but that K +, M,
5.50
SECTION III Membrane Excitability and lon Channels
32. Biophysics of the Nuclear Envelope
551
number of pores constrains the equations. Thus, we may write the flux of the macromolecules, ~mp, in terms of [K]cy, [K]nu, [M]cy, and [M]nu. Now consider a channel model. Assume the nuclear pore complex consists of eight channels for ions and one channel for charged macromolecules, and no that other pathways exist between cytoplasm and nucleoplasm. Then the fluxes of ions and macromolecules are coupled through their electric charge without a separate assumption about the carrier nature of the pore. We may write the flux equations FIGURE 13. The nucleus as a cell within a cell. This diagram suppresses the double-membrane structure of the nuclear envelope, and it regards the nucleus (nu) as a membrane-bound organelle. In this model, only the nuclear pores (np) connect the nucleoplasm (nu) to the cytoplasm (cy). The nuclear envelope is, therefore, a semipermeable membrane analogous to the cell plasma membrane that separates the cytoplasm from the extracellular space (ex). and P by themselves cannot. Now consider two fluxes: the flux of KsP bound to the pore
-(I~kp=Pkp([K8P nue-aVnu-[K8P]cye-aV~y)
(7)
and the flux of MP bound to the pore
-dPmp=Pmp([MP]nue-a~'nu-[MP]cye-aVcy )
(8)
where Pkp is the permeability of K8P, and p mpis the permeability o t MP through the envelope. By hypothesis, let the transport of macromolecules and the transport couple to one another. We assume this is a property of the nuclear pore. Then the steady-state condition reduces to
(I)kp+ %p- 0
(9)
This development treats the nuclear pore complex as a carrier and indicates that the concentration of the carrier does not change: if a certain number of these carriers in the kp state move to the left, then an equivalent number in the mp state move to the right. We may solve these equations for (I)mp, the flux of macromolecules, in terms of the concentrations of the substrates, [K] and [M], and the concentration of the pores, [P], using
[KsP] - [K]S [P]/Kkp (~0) [MP] - [ M l [ P l / K m p The values for Kkp, Kmp, as well as the concentrations of the substrates, may differ on each side of the pore. The fixed
-dPk= Pk ([K ]nu e-aV,u- [K ]cye-aVcy)
(11)
-~ m- pm([M]nue-aVnu-[M]cye-a Vcy)
(12)
We have dropped the p in the subscripts kp, etc., to distinguish the channel model from the carrier model. In the channel model, the steady-state condition reduces to this
8(I)k+2m%= 0
(13)
where zm is the valence of the macromolecule. Equation 13 leads to an expression for the flux of the macromolecules, m' in terms of the substrate concentrations on either side of the envelope. The flux of any charged molecule, including a macromolecule, is an electrical current. In general, the relationship between flux, ~ o #]cmZ-s, and current, I o pA/cm 2, is
I - zen~
(14)
The parameter z in this expression is the valence on the molecule whose concentration is n, and e is the electronic charge (e = 1.6 x 10 -I9 coulombs). For example, if ze were the charge of the K + ion, then ze = +le, and would represent the charge on the macromolecule under consideration. Thus we may use Eq. 14 to write an expression for the K + current and for the macromolecular current, and these two currents would be constrained by Eq. 13. The carrier model or the channel model amounts to a balance of fluxes and currents. For example, if the charged macromolecule being transported is RNA, then we could write from the above channel model
Zme
I K + IRNA = 0
(15)
A similar equation would hold for charged proteins. It may seem unusual to consider the m o v e m e n t of R N A or proteins as a current. However, we are already familiar with this concept in gel electrophoresis, in which nucleotides and proteins move under an electric field. Equation 15 hypothesizes that a comparable p h e n o m e n o n to gel electrophoresis occurs at the nuclear envelope. A s s u m e a rate of 10 9 R N A molecules/min, one electronic
FIGURE 12. The nucleus as an osmometer. Nuclei (nu) from 3T3 fibroblasts under four osmotic conditions. In all cases, the bath solution contains nystatin, a monovalent cation ionophore, at a concentration of 50-100 ug/ml. The nystatin pores form holes in the plasma membrane that connect the cytoplasm to the bath. (A) 120 mM KC1 bath solution (isosmotic), (B) 240 mM KC1 (hyperosmotic), (C) 60 mM KC1 (hypo-osmotic), and (D) 480 mM KC1. In (A, B, and C) the effects are reversible, but in (D) the effects are irreversible. (Unpublished data.)
SECTION III Membrane Excitability and Ion Channels
552
unit charge/base, and 104 bases/molecule. Then IRNA would be the order of 10 nA for the entire nucleus and 1 pA for an individual nuclear pore. The point we wish to emphasize is that transport of a charged macromolecule must have a counter current. If electrical coupling between macromolecules and ions does occur, the ionic current and the voltage across the nuclear envelope could regulate transport of nucleic acids and proteins into and out of the nucleus.
Viii. Modulation of ionic Nuclear Permeability by ATP Patch-clamp experiments on isolated nuclei suggest that the channel activity recorded on the nuclear envelope surface represent a direct nucleocytoplasmic communication pathway. As previously stated, several paradoxes arise from this hypothesis. Results obtained from photobleaching techniques indicate that 40 kDa molecules are the upper limit for free diffusion through the pore. In terms of diameter this would mean 9 nm particles. A channel that is 9 nm in diameter, 80 nm in length, and contains a solution of 100 l~-cm, has a conductance of about 1 nS. This value is significantly greater than the lowest channel conductance levels obtained using the patchclamp recording technique, especially if we consider that the surface isolated by the patch pipette contains many nuclear pores. Pores of 1 nS conductance would shunt the patch and make the observation of smaller conductances impossible. Thus if the observed channels are in the outer membrane, the nuclear pores must be closed. On the other hand, if the channels are within the nuclear pore, their conductance is too low. Is it possible that, under the experimental conditions of the patch clamp recordings, the pores are closed? As mentioned previously, the nuclear pore contains a plug (Fig. 6) that may move under altered Ca conditions. Nuclear isolation procedures are traumatic: cytoplasmic components are lost and nucleo cytoplasmic cytoskeletal structures are disrupted. Patchclamp recordings from nuclei inside the cell help overcome these shortcomings. Figure 14 shows single-channel events recorded in situ from Xenopus oocyte nuclei (developmental stage one). Adding 1 mM ATP to the pipette solution (a physiological concentration of ATP typical for cytoplasm) transforms sporadic channel openings into a macroscopic current with characteristic kinetics. Figure 15 shows the results of such experiments under three different conditions. The presence of ATP in the recording pipette is essential to record the macroscopic current illustrated in (Fig. 15B). In the absence of the nucleotide, either in situ or after nuclear isolation, individual channel openings are observed, as illustrated in Fig. 15C and E. However, the ensemble averages of (C) or (E) have the same kinetics as (B). Furthermore, whole-cell nucleus recordings have resulted in current traces similar to Fig. 15B, although the amplitude of the current is much larger. We can estimate the upper limit of the current that should flow if all the pores in the patch were open. Freezefracture measurements on first-stage Xenopus oocytes indicate 5-8 pores per pm 2. A patch area of 2 pm 2 would contain 10-16 pores. Assuming the maximum expected value of 1 nS per p o r e , - 2 5 mV would generate a patch
current between 250 and 375 pA. In experiments like that of Fig. 15B, the average peak current a t - 2 5 mV is 318 + 34 pA (mean + SD; n = 22). These results show that the channels observed in isolated nuclei are similar to the channels observed in situ, that the channel pathways responsible for the currents are the same with or without ATE and that ATP increases the open probability. Thus, the experimental conditions of in vitro patch-clamp recordings apparently do not alter the properties of individual channels, although the open probability may vary. These data, and data from whole-nucleus preparations, support the hypothesis that channels recorded from the nuclear envelope represent a direct nucleocytoplasmic communication pathway.
IX. Cytoskeletal Interaction with the Nuclear ionic Flux Recent experiments have tested the hypothesis that most of the current flows through the nuclear pore. The first approach was to combine single-channel recording techniques on isolated rat liver nuclei with Atomic Force Microscopy (AFM). The sequence was as follows: after forming a seal on an isolated nucleus, the patch was stimulated with voltage steps to observe channel activity. The pipette was then withdrawn, detaching a patch from the nuclear envelope. It was then possible to transfer the patch on a rigid support and visualize its surface using AFM. In this way we showed that membrane patches with ion channel activity contained several nuclear pores (Fig. 16). Unfortunately this procedure was unable to demonstrate how the number of channels and nuclear pores are related. Using a different approach, Tonini and colleagues have made a comparative study using neonatal and adult mouse liver nuclei. Using freeze-fracture techniques, Tonini showed that pore density in two preparations is substantially the same, however, electrophysiology experiments show that the number of channels having the same conductance (300 pS) are three- to fourfold less in adult preparation. Biochemical analysis of the cytoskeletal nuclear components, actin and myosin, also show marked differences, with more structural proteins present in adult nuclei. The possibility exists that the acto-myosin complex located in the nuclear pore represents an engine for gating a molecular iris formed by the pore subunits. The nuclear pore as a diaphragm is supported by other experiments: ATP and Ca 2+ ions are nuclear envelope modulators and they are implicated in the actin-myosin interactions. AFM demonstrates that addition of ATP to the nuclear membrane modifies the pore structure. Finally, in the comparison between neonatal and adult nuclei, one way to increase the number of channels in adult preparations is to add ATP to the external solution. Thus the nuclear pore may act like a diaphragm with cycles of contraction and relaxation depending on the ATP or Ca 2+ concentration. In this view, the cytoskeleton would then be an active part of the regulatory mechanism for nuclear transport. Adult preparations have no more than a few active channels per patch. In nucleusattached experiments, with patches containing two current
32. Biophysics of the Nuclear Envelope
FIGURE 14. In-situ patch-clamp experiment on a Xenopus oocyte. Enzymatically isolated first-stage oocytes were transferred into experimental chamber. The oocytes were held in position by a holding pipette (A) via light suction. After perforating the plasma membrane (B) the patch-pipette (right) was pushed gently against the nuclear surface. (D) Before touching the nuclear envelope, positive pressure was applied to the patch-pipette solution. (C) The clear shadow in the oocyte cytoplasm is due to solution outflow, which was used to keep the electrode tip clean and to remove cytoplasmic material from the nuclear surface. (From Mazzanti et al., FASEB J. 8:231-236 (1994).)
553
554
SECTION III Membrane Excitability and Ion Channels
in the cell
in the cell
cell-free
1 m M ATP
no ATP
no ATP
( )
( ) C
Vp= - 2 5
Vp= - 2 5
mV
Vp = 25 m V
._~"'-''
mV
'"~r~'L."
---.-IL
!
Vp= - 2 5
B
mV
.
D
F
9 1
,
i
ls
,
t
ls
,
,
ls
F I G U R E 15. ATP modulation of the ionic permeability of the nuclear envelope. The upper panels in the figure show schematic drawings of the patch-clamp configurations corresponding to the experimental traces plotted below. (A) Voltage protocol used in the experiments. (B) Current traces at eight different test voltages from a nucleus-attached experiment in an intact oocyte. The pipette contained 1 mM ATP. (C) Nucleus-attached single-channel traces with -25 mV in the pipette. The patch electrode did not contain ATP. (D) Average current of 20 single-channel recordings obtained at each voltage step. In (E) and (F) experiments were performed after manual isolation of nuclei. The patch pipette contains no ATP. (E) depicts single-channel activity at -25 mV. (F) shows average currents at different test potentials. (From Mazzanti et al., FASEB J. 8:231-236 (1994).)
F I G U R E 16. Atomic Force Microscopy view of nuclear envelope surface. Comparison between a nuclear membrane patch that was transferred from a patch pipette to the supporting substrate (a) and a nuclear membrane that was manually prepared (b). Nuclear pore complexes (indicated by the solid arrows) are clearly preserved by the manual preparation and resolved by AFM. In the transferred patch, the structures are less well preserved, probably due to the transfer process. The arrows indicate NPCs. Both the manually prepared nuclear envelopes and the transferred membrane patch shown in this figure are from Xenopus laevis oocyte nuclei. (From Danker et al., Cell Biol. Int., 21,11:747-757 (1998).)
32. Biophysics of the Nuclear Envelope
555
levels corresponding to 300 pS conductance, the actin filament disrupter, cytochalasin, was slowly perfused around the nucleus. After 2-5 seconds, the number of single-channel current levels increased to 10 before loosing the seal (Fig. 17). Taking into account that the number of nuclear pore structures per patch is between 8 and 15 in this preparation, this experiment suggests that the acto-myosin complex gates the nuclear pore.
X. S u m m a r y The intricacy of the nuclear pore complex, which contains perhaps as many as 100 proteins and possibly multiple nucleocytoplasmic pathways, indicates that the central pore is not simply a watery hole. The complex is rather a coupled transporter for ions and macromolecules. This conclusion
seems plausible because the nucleus can partition ions, the nuclear envelope can maintain a resting potential, osmotic forces are manifest, and ion-selective channels exist in patches that contain dozens of nuclear pores. How can the same structure that transports macromolecules be selective to small ions? One viewpoint would be that the observed channels lie not in the pore but in the inner and outer membranes. Nuclear pores would then be the sites for macromolecular transport, and channels in cisterns would be the sites of ion transport. Channels do exist in the cistern membranes, but we hypothesize that they also exist within the nuclear pore complex itself. The lateral openings shown in Fig. 5 may be the site of these channels. This hypothesis is attractive for several reasons, for example, macromolecular transport requires ATE and in situ patch-clamp experiments indicate that ATP opens channels. If proteins and nucleic acids move as ions, ATP would regulate macromolecular transport
400
A
300
200
1 oo O
,
I.g
ls
0
A ......
-120
I
.... I
-90
,I
-60 pA
pA 150-
G
i
-30
9 .
9
~ .
,
0
o
~
.
..'..~,,~..-
100
+
I
A
.." ~,....~ ... .~" ..'~... .... 9" "'.':O" . . J .... ,4"':::"..:':"
= 1.5 nS 50 ..:;i'.-?;;'" = 1.8nS ,e-" = 2 . 1 nS " 2'0 " 4b 2.4 nS -60 ........-zl0 " - 2 0 . " "~'~ . ...:~;!.'.'::" = 2.7 nS ~::~'5.-"' 0::Y -_.= 3.0 nS :~:.-. -'..v:. .-" o .," . 9-
6'0
mV
oo
9 1 4 9
9 .."
"
. ,' 9
.." .." .
-150
.-
FIGURE 17. Cytochalasin effect on nuclear permeability. Cytocalasin modulates the nuclear pore9The panel on the top presents 3 different examples of single-channel current recordings obtained in steady-state conditions with -50 mV holding potential in the patch pipette9 Before the interruption of the experimental traces (two parallel bars), the current has 1 or 2 levels9After a variable time of exposure to 10 ~M cytocalasin, the number of levels increases to eight in this example before loosing the seal. The bottom panel shows amplitude histograms before (left) and after (right) cytocalasin treatment. (From Tonini et al., FASEB J. 13:1395-1403).
556
SECTION III Membrane Excitability and Ion Channels
through channel openings. It is thus crucial to test whether the nuclear pore complex and the ion channels are linked to the same structure. The majority of experiments indicate that the ion channels in the envelope select K + ions; however, C1-channels, ligandgated ion channels, and Ca 2+ pumps are also present. Channels in the outer and inner membranes do not preclude channels in the nuclear pore complex. Furthermore, the Donnan equilibrium and the semipermeable membrane may coexist at the nuclear envelope, as they do in the plasma membrane. Attributing the electrical potential of the nucleus to either a Donnan equilibrium or a selective membrane reflects early controversies concerning the plasma membrane. Curiously, a similar debate exists in another area of cell biology, the release of hormones and neurotransmitters from secretory vesicles. Secretory vesicles contain a gel matrix that restricts the movement of transmitters even after membrane fusion has occurred. In such cases, mechanisms that rely on the matrix, and mechanisms that rely on the membrane that surrounds the matrix, are not mutually exclusive. It is interesting to speculate whether the structure of the nuclear pore complex resembles other macromolecular assemblies. One similarity regards the central granule (or plug), a protein within a protein motif that is c o m m o n to other macromolecules. Criss Hartzell (Emory University) has noticed that the eight-plus-one structure of the nuclear pore complex is reminiscent of chaperones, which have a seven-plus-one structure, and we have already pointed to the axoneme with its nine-plus-one structure. Mu-Ming Poo (UC Berkeley) has asked whether parts of the assembled nuclear pore complex might not exist as independent units in the endoplasmic reticulum, to be reassembled into the complete nuclear pore complex. K+-selective channels and proteinconducting channels reside in the reticulum and bear some similarity to the ionic and macromolecular pathways that we have discussed. The possibility that channel proteins and transporters in endoplasmic reticulum are components of the nuclear pore complex would have broad implications for pore assembly during development. These speculations may stimulate new experiments to help resolve the structural basis of nuclear pore function.
Acknowledgments We wish to thank B.J. Duke for helping with experiments and preparing the figures. Unpublished work presented here was done at Stazione Zoologica in Naples, Italy with Luigia Santella and Brian Dale. NATO-CGR 190025 to BD and LJD supported this work.
Bibliography Agutter, E S. (1991). Role of the cytoskeleton in nucleocytoplasmic RNA and protein distributions. Biochem. Soc. Trans. 19, 1094-1098. Aidley, D. J. (1989). "The Physiology of the Excitable Cell," Third edition. Cambridge University Press, Cambridge, UK. Akey, C. W. (1989) Interaction and structure of the nuclear pore complex revealed by cryoelectron microscopy. J. Cell Biol. 109, 955-970. Alberts, et al. (1989). "Molecular Biology of the Cell," 2nd ed., Ch. 8 Garland Publishing, New York.
Blobel, G., and Potter, V. R. (1966) Nuclei from rat liver: Isolation method that combines purity with high yield. Science 154, 1662-1665. Bonner, W. M. (1975). Protein migration into nuclei: I. Frog oocyte nuclei in vivo accumulate microinjected histones, allow entry to small proteins, and exclude large proteins. J. Cell Biol. 64, 421-430 Bonner, W. M. (1975). Protein migration into nuclei: II. Frog oocyte nuclei accumulate a class of microinjected oocyte nuclear proteins and exclude a class of microinjected oocyte cytoplasmic proteins. J. Cell Biol. 64, 431-437. Bustamente, J. O. (1992). Nuclear ion channels in cardiac myocytes. Pflugers Arch. 421, 473-485. Bustamente, J. O. (1992). Nuclear electrophysiology. J. Memb. Biol. 138, 105-112. Carmo-Fonesca, M., and Hurt, E. C. (1991). Across the nuclear pores with the help of nucleoporins. Chromosoma 101, 199-205. Carter, K. C., Bowman, D., Carrington, W., Fargarty, K., McNeil, J. A., Fay, E S., and Lawrence, J. B. (1993). A 3-D view of precursor messenger RNA metabolism within the mammalian nucleus. Science 259, 1330-1335. Century, T. J., Fenichel, I. R., and Horowitz, S. B. (1970). The concentration of water, Na and K ions in the nucleus and cytoplasm of amphibian oocytes. J. Cell Sci. 7, 5-13. Dale, B., DeFelice, L. J., Kyozuka, K., Santella, L., and Tosti, E. (1994). Voltage clamp of the nuclear envelope. Proc. Royal Soc. B 225, 119-124. Danker, T., Mazzanti, M., Tonini, R., Rakowska, A., and Oberleithner, H. (1998). Using atomic force microscopy to investigate patchclamped nuclear membrane. Cell Biol. Int. 21, 747-757. Dargemont, C., and Kuhn, L. C. (1992). Export of mRNA from microinjected nuclei of Xenopus laevis oocytes. J. Cell Biol. 118, 1-9. DeFelice, L. J. (1981). "Introduction to Membrane Noise." Plenum Press, N.Y. DeFelice, L. J. (1997). "Electrical Properties of Cells: Patch Clamp for Biologists." Plenum Press, N.Y. Dingwall, C. (1990). Plugging the nuclear pore. Nature 346, 512-514. Dworetzky, S. I. and Feldherr, C. M. (1988). Translocation of RNAcoated gold particles through the nuclear pores of oocytes. J. Cell. Biol. 106, 575-584. Feldherr, C. M., and Akin, D. (1990). EM visualization of nucleocytoplasmic transport process. Elec. Micro. Rev. 3, 73-86. Finkelstein, A. (1987). "Water Movement through Lipid Bilayers, Pores, and Plasma Membranes." John Wiley and Sons, New York. Finlay, D. R., Newmeyer, D. D., Price, T. M., and Forbes, D. J. (1987). Inhibition of in vitro nuclear transport by a lectin that binds to nuclear pores. J. Cell Biol. 104, 189-200. Forbes, D. J. (1992). Structure and function of the nuclear pore complex. Annu. Rev. Cell Biol. 8, 495-527. Goldberg, M. W., Blow, J. J., and Allen, T. D. (1992). The use of field emission in-lens scanning electron microscopy to study the steps of assembly of the nuclear envelope. J. Struct. Biol. 108, 257-268. Hernandez-Cruz, A., Sala, E, and Conner, J. A. (1991). Stimulusinduced nuclear Ca signals in Fura-2 loaded amphibian neurons. Ann. N.Y. Acad. Sci. 635, 416-420. Hille, B. (1992). "Ionic Channels of Excitable Membranes." Second edition. Sinauer Associates, Sunderland, MA. Hinshaw, J. E., Carracher, B. O., and Milligan, R. A. (1992). Architecture and design of the nuclear pore complex. Cell 69, 1133-1141. Horowitz, S. B., and Moore, L. C. (1974). The nuclear permeability intracellular distribution, and diffusion in Inulin in the amphibian oocyte. J. Cell Biol. 60, 405-4 15. Innocenti, B., and Mazzanti, M. (1993). Identification of a nucleocytoplasmic ionic pathway by osmotic shock in isolated mouse liver nuclei. J. Memb. Biol. 131, 137-142. Jarnik, M., and Aebi, U. (1991). Toward a more complete 3-D structure of the nuclear pore complex. J. Struct. Biol. 110, 883-894.
32. Biophysics of the Nuclear Envelope
Lauger, P. (1991). "Electrogenic Ion Pumps." Second edition. Sinauer Associates, Sunderland, MA. Loewenstein, W. R., and Kanno, Y. (1963). Some electrical properties of a nuclear membrane examined with a microelectrode. J. Gen. Physiol. 46, 1123-1140. Matzke, A. J. M., and Matzke, M. A., (1991). The electrical properties of the nuclear envelope, and their possible role in the regulation of eukaryotic gene expression. Bioelectrochem. Bioenerg. 25, 357-370. Maul, G. G. (1977). The nuclear and cytoplasmic pore complex: Structure, dynamics, distribution, and evolution. Int. Rev. Cytol. 6 (Suppl.), 75-186. Mazzanti, M., DeFelice, L. J., Cohen, J., and Malter, H. (1990). Ion channels in the nuclear envelope. Nature 343, 764-767. Mazzanti, M., DeFelice, L. J., and Smith, E. E (1991). Ion channels in murine nuclei during early development and in fully differentiated adult cells. J. Memb. Biol. 121, 189-198. Mazzanti, M., Innocenti, B., and Rigatelli, M. (1994). ATP dependent ionic permeability of nuclear envelope in in-situ xenopus oocyte nuclei. FASEB J. 8, 231-236. Moore, M. S., and Blobel, G. (1992). The two steps of nuclear import, targeting to the nuclear envelope and translocation through the nuclear pore, require different cytosolic factors. Cell 69, 939-950. Newport, J. W. and Forbes, D. J. (1987). The nucleus: Structure, function and dynamics. Ann. Rev. Biochem. 56, 535-65. Nicotera, E, McConkey, D. J., Jones, D. E, and Orrenius, S. (1989). ATP stimulates Ca uptake and increases the free Ca concentration in isolated rat liver nuclei. PNAS 86, 453-457. Nigg, E. A., Baeuerle, E A., and Luhrmann, R. (1991). Nuclear importexport: In search of signals and mechanisms. Cell 66, 15-22. Overbeek, J. T. G. (1956). The Donnan Equilibrium. Prog. Biophys. Mol. Biol. 6, 57-84. Paine, E L., Johnson, M. E., Lao, Y. T., Tluczek, L. J., and Miller, D. S. (1992). The oocyte nucleus isolated in oil retains in vivo structure and functions. BioTechniques 13, 238-246. Paine, E L., Horowitz, S. B. (1980). The movement of material between nucleus and cytoplasm. Cell Biol. 4, 299-338. Perez-Terzic, C., Pyle, J., Jaconi, M., Stehno-Bittle, L., and Clapham, D. E. (1996). Conformational states of the nuclear pore complex induced by depletion of nuclear Ca stores. Science 273, 1875-1877.
557 Richardson, W. D., Mills, A. D., Dilworth, S. M., Laskey, R. A., and Forbes D. J. (1992). Structure and function of the nuclear pore complex. Annu. Rev. Cell Biol. 8, 495-572. Santella, L. (1996). The cell nucleas: an Eldorado to future Ca research? J. Membr. Biol. 153, 83-92. Simon, S. M., and Blobel, G. (1991). A protein-conducting channel in the endoplasmic reticulum. Cell 65, 371-380. Stewart, M., Whytock, S., and Moir, R. D. (1991). Nuclear envelope dynamics and nucleocytoplasmic transport. J. Cell Sci. Supp. 14, 79-82. Stochaj, U., and Silver, P. (1992). Nucleoplasmic traffic of proteins. Eur. J. Cell Biol. 59, 1-11. Stricker, S. A., Centonze, V. E., Paddock, S. W., and Schatten, G. (1992). Confocal microscopy of fertilization-induced Ca dynamics in sea urchin eggs. Dev. Biol. 149, 370--380. Stehno-Bittel, L., Perez-Terzic, C., and Clapham, D. E. (1995). Diffusion across the nuclear envelope inhibited by depletion of the nuclear Ca store. Science 270, 1835-1838. Tabares, L. M., Mazzanti, M., and Clapham, D. E. (1991). C1 channels in the nuclear envelope. J. Memb. Biol. 123, 49-54. Tonini, R., Grohovaz, E, LaPorta, C. A. M., and Mazzanti, M. (1999). Gating mechanism of the nuclear pore complex channel in isolated neonatal and adult mouse. FASEB J. 13, 1395-1403. Unwin, P. N. T., and Milligan, R. A. (1982). A large particle associated with the perimeter of the nuclear pore complex. J. Cell Biol. 93, 63-75. Wagner, P., Kunz, J., Koller, A., and Hall, M. N. (1990). Active transport of proteins into the nucleus. FEBS Lett. 275, 1-5. Waybill, M. M., Yelamarty, R. V., Zhang, Y. L., Scaduto, R. C., Jr., LaNoue, K. F., Hsu, C. J., Smith, B. C., Tillotson, D. L., Yu, F. T., and Cheung, J. Y. (1991). Nuclear Ca transients in cultured rat hepatocytes. Am. J. Physiol. 261, E49-E57. Wente, S. R., Rout, M. P., and Blobel, G. (1992). A new family of yeast nuclear pore complex proteins. J. Cell Biol. 119, 705-723. Williams, D. A., Becker, P., and Fay, E S. (1987). Regional changes in Ca underlying contraction of single smooth muscle cells. Science 235, 1644-1648. Yarmola, E. G., Zarudnaya, M. I., and Lazurkin, Y. S. (1985). Osmotic pressure of DNA solutions and effective diameter of the double helix. J. Bio. Structure and Dynamics 2, 981-993. Zasloff, M. (1983). tRNA transport from the nucleus in a eukaryotic cell: Cartier-mediated translocation process. Proc. Natl. Acad. Sci. USA 80, 6436-6440
Nicholas Sperelakis and Gordon M. Wahler
!3!3 Regulation of Ion Channels by phosphorylation
I. Introduction Considerable attention has been given during the past 25 years to phosphorylation of ion channels as a means whereby the activity of the channels can be regulated or modulated. There is evidence for such regulation or modulation of function of Ca 2§ K +, Na § and C1- channels by phosphorylation, and biochemical evidence shows that one or a few sites on the channel proteins can be phosphorylated by various protein kinases. Most physiological evidence for such changes in ion channel function is based on the L-type Ca 2§ channels of nerve, skeletal muscle, cardiac muscle, and vascular smooth muscle (VSM) and on the K § channels (delayed-rectifier type) of cardiac muscle and nerve. This chapter focuses primarily on the slow L-type Ca 2§ channel of cardiac muscle in order to illustrate the important principles that are involved. The voltage-dependent L-type Ca 2§ channels in the myocardial cell membrane are the major pathway by which Ca 2§ ions enter the cell during excitation for initiation and regulation of the force of contraction of cardiac muscle. The Ltype Ca 2§ channels have some special properties, including functional dependence on metabolic energy, selective blockade by acidosis, and regulation by the intracellular cyclic nucleotide levels. Because of the special properties of these slow channels, Ca 2+ influx into the myocardial cell can be controlled by both extrinsic factors (such as autonomic nerve stimulation or circulating hormones) and intrinsic factors (such as intracellular pH or adenosine triphosphate (ATP) levels). In myocardial cells, the Ca 2§ influx that occurs during each cardiac cycle is regulated by cyclic nucleotides. This regulation is presumably mediated by phosphorylation(s) of the L-type Ca 2§ channel protein. Phosphorylation of the Ca 2+ channels (or of an associated regulatory protein) by cAMP-dependent protein kinase (PK-A) (Fig. 1) (1) increases the number of L-type Ca 2§ channels available for voltage activation during the action potential (AP), (2) increases the probability of channel opening, and (3) increases channel mean open time. The greater density of open Ca 2§ channels increases the inward Ca 2+ current (/Ca) during the
Cell Physiology Sourcebook: A Molecular Approach, Third Edition
5 5 9
AE The resulting increase in Ca 2+ influx increases the force of contraction of the heart. In contrast to the response of the channel to phosphorylation by PK-A, phosphorylation of the channels or an associated regulatory protein by cGMPdependent PK (PK-G) depresses the activity of the L-type Ca 2+ channels (Wahler and Sperelakis, 1985; Wahler et al., 1990).
II. Types of Ca+ Channels Four or five different subtypes of voltage-dependent Ca 2+ channels have been described for nerve and muscle cells. The first three, found in sensory ganglion neurons (rat dorsal root ganglion) of the spinal cord, were called L-type (i.e., long-lasting or kinetically slow), T-type (i.e., transient or kinetically fast), and N-type (i.e., neither L-type nor Ttype) Ca 2+ channels (Nowycky et al., 1985). More recently, another subtype was found initially in Purkinje neurons of the cerebellum, and hence called P-type (Llinas et al., 1992). The P-type channel has since been identified at the nerve terminals of neuromuscular junctions of both vertebrates and invertebrates. Muscle fibers, in general, apparently possess only the L-type and T-type channels, with the T-type being very sparse or absent in some types of muscles. That is, in muscle cells, the major inward Ca 2+ current (involved in excitation-contraction coupling) is the current through the L-type Ca 2+ channels. This is true of skeletal muscle, cardiac muscle, and smooth muscles. The N-type Ca 2+ channel has a single-channel conductance (y) of about 14-18 pS, which is between that of the T-type (8-12 pS) and L-type (18-26 pS). The N-type channel has been localized only to neurons so far. The P-type Ca 2+ channel is high threshold (like the L-type), having an activation voltage o f - 4 5 t o - 3 5 mV, but is not blocked by the L-type channel blockers. The values reported for singlechannel conductance (y) in various cells range from 9-20 pS. Table 1 summarizes the major differences between the Ltype and T-type Ca 2+ channels. As indicated, the kinetics of
Copyright 9 2001 by Academic Press. All rights of reproduction in any form reserved.
560
SECTION III Membrane Excitability and Ion Channels
FIGURE 1. Schematic model for an L-type Ca2+ channel in myocardial cell membrane in two hypothetical forms: dephosphorylated (or electrically silent) form (left diagrams) and phosphorylated form (right diagrams). The two gates associated with the channel are an activation gate and an inactivation gate. The phosphorylation hypothesis states that a protein constituent of the slow channel itself (A) or a regulatory protein associated with the slow channel (B) must be phosphorylated for the channel to be in a state available for voltage activation. Phosphorylation of a serine or threonine residue occurs by PK-A in the presence of ATP. Phosphorylation may produce a conformational change that effectively allows the channel gates to operate. The L-type channel (or an associated regulatory protein) may also be phosphorylated by PK-G (C), thus mediating the inhibitory effects of cGMP on the Ca2§ channel. (Adapted from Sperelakis and Schneider, 1976.)
activation and inactivation are slower for the L-type. That is, the L-type current I o,,, turns on (activates) more slowly and turns off (inactivates) more slowly. In addition, the voltage range over which these channels operate is different, the threshold potential and inactivation potential being higher (more positive or less negative) for the L-type Ca 2+ channels. Therefore, the L-type channels are high threshold and the T-type channels are low threshold. The single-channel conductance is greater for the L-type Ca 2§ channel: 18-26 pS versus 8-12 pS. The L-type Ca 2+ channels are regulated by cyclic nucleotides and phosphorylation, whereas the fast T-type Ca 2§ channels are not. Finally, the L-type Ca 2+ channels are blocked by Ca 2§ channel antagonist drugs (such as verapamil, diltiazem, and nifedipine) and opened by Ca 2+
agonist drugs (such as Bay K-8644, a dihydropyridine that is chemically very close to nifedipine), whereas the T-type Ca 2§ channels are not. In some respects, the T-type Ca 2§ channels behave like fast Na § channels, except that the Ttype channels are Ca 2+ selective (rather than Na § selective) and are not blocked by tetrodotoxin (TTX). Table 2 summarizes the blocking or opening action of several drugs and toxins on the various subtypes of Ca 2+ channels. As indicated, the L-type Ca 2§ channel is blocked by the prototype Ca 2§ antagonist drugs (verapamil, diltiazem, and nifedipine) and opened by the dihydropyridine Bay K-8644. The T-type channels are relatively selectively blocked by tetramethrine and by low concentrations of Ni 2+ (e.g., 30 pM). Higher concentrations of Ni 2+ block the L-
33. Regulation of Ion Channels by Phosphorylation
TABLE 1
561
Summary of Major Differences between the L-Type and
T-Type Ca 2+ Channels
Properties
L-type (Slow)
T-type (Fast)
Duration of current
Long-lasting (sustained)
Transient
Inactivation kinetics
Slower
Faster
Activation kinetics
Slower
Faster
Threshold
High (ca. -35 mV)
Low (ca.-60 mV)
Half-inactivation potential
ca. -20 mV
ca. -50 mV
Single-channel conductance
High (18-26 pS)
Low (8-12 pS)
Regulated by cAMP and cGMP
Yes
No
Regulated by phosphorylation
Yes
No
Blocked by Ca2§ antagonist drugs
Yes
No (slight)
Opened by Ca2§ agonist drugs
Yes
No
Permeation by Me2+
Ba > Ca
Ba _=_Ca
Inactivation by [Ca]/
Yes
Slight (?)
Recordings in isolated patches
Runs down
Relatively stable
type channels as well. Possible blockers of the F-type channel (see below) are not known. In addition to the Ca 2§ channel subtypes described previously, a new subtype was discovered in 18-day-old fetal rat ventricular (heart) cells (Tohse et al., 1992). A substantial fraction (e.g., 30%) of the total/Ca remained in the presence of a high concentration (3 laM) of nifedipine (nifedipineresistant/Ca)' and it was not blocked by diltiazem (another L-type channel blocker) or og-conotoxin (N-type channel blocker). This novel Ca 2+ current had a half-inactivation potential about 20 mV more negative than that of the L-type Ca 2+ current. It was called F-type (fetal-type) Ca 2+ current (1Ca(F)) " The single-channel conductance (3') is not known Another difference discovered during the embryonic or fetal period, first observed in chick and later in rat, is that the L-type Ca 2§ channels exhibit an unusually high incidence of
very long openings, as observed in single-channel recordings (cell-attached patch) (Tohse and Sperelakis, 1990; Masuda et al., 1995). The incidence of long openings diminishes during development and approaches the adult channel behavior. The adult behavior primarily consists of bursting patterns (rapid openings and closings of short durations, or flickering).
III. Cyclic AMP Stimulation of L-type Ca 2§ Channels A. Cyclic AMP It has long been known that cyclic AMP (cAMP) modulates the functioning of the L-type Ca 2§ channels (Shigenobu and Sperelakis, 1972; Tsien et al., 1972; Reuter and Scholz,
TABLE 2
S u m m a r y of Drugs or Toxins t h a t Block or O p e n t h e Various S u b t y p e s of Ca 2+ C h a n n e l s Channel type
Blockers
L-type
Verapamil, diltiazem, nifedipine
T-type
Tetramethrine, Ni 2+ (30 ~M)
N-type
oJ-Conotoxin
P-type
og-Agatoxin-IVA a Polyamine (FTX) b
Openers Bay K-8644
F-type aw-Agatoxin-IVA, a polypeptide (5202 Da) from funnel-web spider venom (Agenopsis aperta), blocks P-type Ca2§ channels of rat Purkinje neurons with a Ko of about 2 nM. bA smaller (ca. 254 Da ) polyamine (FTX) extracted from venom of this spider also blocks the P-type channels (Mintz et al., 1992). FTX block is antagonized by Ba2+ ion. P-type Ca2+ channels are involved in presynaptic Ca2+ influx and associated neurotransmitter release.
562
1977)./3-adrenergic agonists, after binding to theft specific receptors, lead to rapid stimulation of adenylate cyclase with resultant elevation of cAMP levels. Other neurotransmitters, hormones, and so on that stimulate cAMP production (e.g., histamine) similarly enhance/Ca" Drugs that inhibit phosphodiesterase (the enzyme that hydrolyzes cyclic nucleotides) also cause an elevation of cAMP and increase of /Ca (Fischmeister and Hartzell, 1991; Mubagwa et al., 1993; Kawamura and Wahler, 1994). Additional evidence for the regulatory role of cAMP in heart cells includes the following: (1) Forskolin and the GTP analog GPP(NH)P, which directly activate adenylate cyclase, both induce Ca2+-dependent slow APs (Josephson and Sperelakis, 1978; Wahler and Sperelakis, 1986) and enhance /Ca (Hescheler et al., 1986). (2) cAMP injection into ventricular muscle cells induces Ca2+-dependent slow APs in the injected cells within seconds (Vogel and Sperelakis, 1981; Li and Sperelakis, 1983; Bkaily and Sperelakis, 1985) or enhances/Ca in isolated single cardiac cells (Irisawa and Kokubun, 1983). (3)A photochemical activation method for suddenly increasing the intracellular cAMP level enhances /Ca in bullfrog atrial cells (Nargeot et al., 1983). (4) Singlechannel analysis suggests that cAMP increases the number of functional L-type Ca 2+ channels available and/or the probability of opening of a given channel (Cachelin et al., 1983; Trautwein and Hofmann, 1983; Bean et al., 1984). The/3-adrenergic agonist isoproterenol increases the mean open time of single Ca 2§ channels and decreases the intervals between bursts; the conductance of the single channel is not increased (Reuter et al., 1982). Therefore, the increase in the Ca 2+ current produced by isoproterenol could be produced by the observed increase in mean open time of each channel and the probability of opening, as well as by an increase in the number of available channels. There are gender differences in the cardiovascular responses to/3-adrenergic stimulation. Thus, the/3-adrenergicmediated increases in the Ca 2+ current (Vizgirda et al., 1997) and in cell shortening (Vizgirda et al., unpublished observations) of rat ventricular myocytes have been found to be approximately two-fold greater in males than females.
SECTION III Membrane Excitability and Ion Channels
ylation of the Ca 2+ channel may also act to increase/Ca by reducing the normal blockade of the channel by Mg 2+ (i.e., by altering the affinity of binding sites in the channel pore for Mg 2+) (Yamaoka and Seyama, 1998). Cyclic AMP-dependent phosphorylation also modulates the activity of other channel types. For example, in the heart alone, other channels regulated by cAMP-dependent phosphorylation include the CFTR (cystic fibrosis transmembraneconductance regulator) chloride channel (rev. in Gadsby and Naim 1999) and the delayed-rectifier channel (IK(del)) (e.g., Walsh and Kass, 1988). Whatever the precise subcellular mechanism for the effect phosphorylation has on channel activity, in the phosphorylation model, the phosphorylated form of the L-type Ca 2+ channel is the active (operational) form, and the dephosphorylated form is the inactive (inoperative) form. The dephosphorylated channels are virtually silent electrically (i.e., their opening probability approaches zero). Phosphorylation increases the probability of channel opening with depolarization. An equilibrium would exist between the phosphorylated and dephosphorylated forms of the channel under a given set of conditions. For example, fluoride ions (< 1 mM) increase the force of contraction of the heart and potentiate the slow CaZ+-dependent APs and Ca 2+ influx (/Ca) without increasing the level of cAME Fluoride may act by inhibiting the phosphatase, which dephosphorylates the channel protein, thus prolonging the life span of the phosphorylated channel. This suggests that the rate of dephosphorylation, in addition to the rate of phosphorylation (see below), may also be an important determinant of the amplitude of/Ca" Based on the rapid decay of the response to microinjected cAMP (Fig. 2, top), the mean life span of a phosphorylated channel is probably only a few seconds at most, and it is possible that the channels are phosphorylated and dephosphorylated with every cardiac cycle. Agents that affect or regulate the phosphatase that dephosphorylates the channel would affect the life span of the phosphorylated channel. Thus, channel stimulation can be produced either by increasing the rate of phosphorylation (by PK-A activation) or by decreasing the rate of dephosphorylation (inhibition of the phosphatase).
B. Phosphorylation Hypothesis Because of the relationship between cAMP and the number of available L-type Ca 2+ channels and because of the dependence of the functioning of these channels on metabolic energy, it was postulated that the L-type Ca 2+ channel must be phosphorylated for it to become available for voltage activation (Shigenobu and Sperelakis, 1972; Tsien et al., 1972). The elevation of cAMP by a positive inotropic agent activates PK-A, which phosphorylates a variety of proteins in the presence of ATE One protein that is phosphorylated is the L-type Ca 2+ channel protein itself or a contiguous regulatory type of protein (see Fig. 1). Agents that elevate cAMP increase the fraction of the channels that are in the phosphorylated form and hence readily available for voltage activation. Phosphorylation could make the Ca 2+ channel available for activation by a conformational change that allowed the activation gate to be opened upon depolarization (or increased the pore diameter), cAMP-mediated phosphor-
C. Protein Kinase A Intracellular injection of the catalytic subunit of PK-A induces and increases the slow Ca2+-dependent APs and potentiates/Ca (Osterreider et al., 1982; Bkaily and Sperelakis, 1984). Injection of an inhibitor (protein) of the PK-A into heart cells inhibits the spontaneous slow CaZ+-dependent APs and/Ca (Bkaily and Sperelakis, 1984; Kameyama et al., 1986). These results verify that the regulatory effect of cAMP is exerted by means of PK-A and phosphorylation. Consistent with the phosphorylation hypothesis, L-type Ca 2+ channel activity disappears within 90 s in isolated membrane inside-out patches (Reuter, 1983), but can be restored (in neurons) by applying the catalytic subunit of PKA together with MgATP (Armstrong and Eckert, 1987). This is consistent with the washing away of regulatory components of the Ca 2+ channels or of the enzymes necessary to phosphorylate the channel. Even in whole-cell voltage-
33. Regulation of Ion Channels b y Phosphorylation
563
cAMP INJ.
20mY
k_--
60ms 0-25S
1 rain
30s
0-25s
cGMP
Control[
Slow AP
1-2min
2.7min
4-5min
6.0rain
o
-40mV
~
I 20ws
_ . . . _ . . . _ _..~_.. _.__ ~ t
FIGURE
2.
O.ls
;
Effects of intracellu]ar injections of cyclic nucleotides. Upper row: Induction of CaZ+-dependent slow APs
in guinea pig papillary muscle by intracellular pressure injection of cAME The muscle was depolarized in 22 mM Ko to voltage-inactivate fast Na+ channels. (A) Small graded response (stimulation rate 30/min). (B) Superimposed records showing the gradual appearance of slow APs upon cAMP injection over a 25-s period. (C) Presence of stable slow APs after injection for 1 min. (D) Gradual spontaneous depression of slow APs over a period of 25 s after the injection is stopped. (E) Complete decay of slow APs 30 s after cessation of cAMP injection. All records are from one impaled cell. (From Li and Sperelakis, 1983.) Lower row: Transient abolition of Ca2+-dependent slow APs by pressure injection of cGMP. (A) Control slow AP. (B and C) 1-2 min after the onset of cGMP injection (10-s duration), the slow APs were depressed and then abolished. (D and E) At 4-6 min, the slow APs recovered spontaneously to control levels. All records from the same cell. (From Wahler and Sperelakis, 1985.)
clamp, there is a progressive run-down of the Ca 2+ current, which is slowed or partially reversed by conditions that enhance PK-A phosphorylation.
D. Phosphatases In ventricular cells, phosphatases have been shown to inhibit/Ca (e.g., Kameyama et al., 1986; Hescheler et al., 1987; duBell et al., 1996), consistent with the phosphorylation hypothesis. Additionally, phosphatase inhibitors, such as okadaic acid and microcystin, cause large increases in/Ca (Hescheler et al., 1988; Frace and Hartzell, 1993). There is some disagreement about the relative effectiveness of specific phosphatases (particularly type 1 versus type 2A) in reducing/ca' and whether basal/ca is significantly inhibited by phosphatases, or if only the cAMP-stimulated/Ca (e.g., by a fl-adrenergic agonist) is affected. There is overall agreement that dephosphorylation of the Ca 2§ channel by phosphatase is an additional important regulatory step in determining the amplitude of/Ca" In addition to cardiac muscles, phosphatases have been shown to decrease the Ca 2§ current in other cell types, such as neurons (Chad and Eckert, 1986) and vascular smooth muscle cells (e.g., Schumann et al., 1997). Other types of channels have also been shown to be modulated by phosphatases. For example, the CFTR C1-channel is inactivated by phosphatase 2B (calcineurin) (Fischer et al., 1998). Type 1 protein phosphatase activity is inhibited by (at least) two low-molecular weight proteins, known as protein phosphatase inhibitor-1 (PPI-1) and protein phosphatase
inhibitor-2 (PPI-2). PPI-1 is present in ventricular myocytes, and it may be an important additional component of the modulation of the phosphorylation/dephosphorylation cycle of the Ca 2+ channel. The activity of PPI-1 is enhanced with phosphorylation by PK-A (Ahmad et al., 1989; Gupta et al., 1993, 1996). Thus, PK-A not only phosphorylates the Ca 2+ channel, it also decreases the rate of channel dephosphorylation. Both actions stimulate channel activity.
IV. Cyclic GMP Inhibition of the Ca 2§ Current A. Cyclic GMP The physiological role played by cyclic GMP (cGMP) on cardiac function is still controversial. A number of years ago, it was proposed that cGMP plays a role antagonistic to that of cAMP, namely, that there was a Yin-Yang relationship between cAMP and cGMP (Goldberg et al., 1975). Since that time, considerable evidence has accumulated supporting this hypothesis. For example, 8-Br-cGMP (10 -4 M) shortens the AP duration in rat atria (accompanied by a negative inotropic effect), and it was suggested that cGMP might decrease the Ca e§ conductance (Nawrath, 1977). Both acetylcholine (ACh), which increases cGMP levels, and 8-Br-cGMP, a more permeable analog of cGMP, reduce the upstroke velocity and duration of the Ca2+-dependent slow AP in guinea pig atria (Kohlhardt and Haap, 1978). The abbreviation of AP duration also occurs following injection of cGMP into isolated guinea pig cardiomyocytes (Trautwein et al., 1982).
564
SECTION III Membrane Excitability and Ion Channels
Superfusion of isolated ventricular muscle with 8-Br-cGMP abolishes the CaZ+-dependent slow APs and accompanying contractions (Wahler and Sperelakis, 1985). A similar inhibition by cGMP was shown for the slow APs of atrial muscle and Purkinje fibers (Mehegan et al., 1985). Intracellular pressure injection of cGMP into ventricular cells was found to transiently depress or abolish slow APs more quickly (e.g., 1-2 min) (Wahler and Sperelakis, 1985) (Fig. 2, bottom). It was also demonstrated that 8-Br-cGMP inhibits the basal/Ca (i.e., unstimulated by cAMP) in voltage-clamped ventricular myocytes (Wahler et al., 1990; Haddad et al., 1995) (Figs. 3 and 4). Cyclic GMP inhibition of Ca 2+ channel activity of embryonic chick heart was also demonstrated at the single-channel level (Tohse and Sperelakis, 1991) (Fig. 5). Cyclic GMP did not change the unit amplitude and slope conductance of the Ca 2+ channel, but prolonged the closed times and shortened the open times. The Ca 2+ slow channels of young (3-day-old) embryonic chick heart cells often exhibit long-lasting openings (e.g., 300 ms) under normal conditions, especially at more positive command potentials (Tohse and Sperelakis, 1990, 1991). Long-lasting openings were much less frequently observed in 17-day-old embryonic cells. That is, the Ca 2+ channels in early development naturally possess some mode 2 behavior, which is normally produced by Ca 2+ channel agonists such as the dihydropyridine BayK-8644. Addition of 8-Br-cGMP to the bath of cells (3-day) exhibiting long openings completely inhibited Ca 2+ channel activity (see Fig. 5). Long openings were also observed in fetal (12-day) rat ventricular cardiomyocytes (Masuda et al., 1995). In whole-cell voltage-clamp experiments on single ventricular cardiomyocytes from 17-day embryonic chicks, the
8-Br-cGMP (1 mM) o
J ,
.
120-
-40
E
v
t--
-40
0
-50
r-
4 day 0.34
E
e
6 da 0.17
a) - 2 0
-20 9 Of) o
"
e
O
o 9
O
9149
9
o
% oqb*9
x
"5 day
-,v
X~
0.67
o
1
t
500
1000
_
_
t,,
1500
Length (lam)
1...
2000
o
1
5
I
1 I11
J9
50
JO0
200
[K*]o (rnM)
FIGURE 8. Developmental changes of resting potential in the skeletal my9149 from (A) embryonic chick and (B) fetal rat. (A) The relation between the RP and length of the chick skeletal my9149 in culture. The filled circle enclosed within the box is the mean (+ 2SE) resting potential of 20 my9 (B) The relationship between the RP and external K+ ion concentration ([K+]0) for rat skeletal my9 of different periods in culture. The Nernst equilibrium potential for K+ is indicated by the straight line labeled EK. The solid line for each my9 is the theoretical curve predicted by the Goldman equation for the PNa/PK ratios given at the left and the extrapolated [K+]ivalue. [Reproduced with permission from (panel A) Fischbach et al., 1971. J. Cell. Physiol. 78, 289-300. Copyright 1971 John Wiley & Sons, Inc. Reprinted by permission of Wiley-Liss, Inc., a subsidiary of John Wiley & Sons, Inc., and (panel B) Ritchie and Fambrough, 1975. Reproduced from The Journal of General Physiology by copyright permission of The Rockefeller University Press.]
TTX in almost all my 9 tested (Fig. 9A, right side, and Fig. 9B), differentiation of the spikelike APs is due to progressively increased intensity of inward current through fast Na § channels (Kano and Yamamoto, 1977). That is, an increase in the number (density) of fast Na + channels allows a regenerative AP to be produced, and the maximum rate of rise to become faster and faster. In myocytes of a marine tunicate (ascidian), AP duration abbreviates during the embryonic period (Greaves et al., 1996). There is a progressive decrease in AP duration from the middle embryonic to the late embryonic period, and a corresponding increase in the rate of rise and fall of the AE Spontaneous firing of APs is observed in most ascidian myocytes in the middle embryonic period. The automaticity progressively disappears during development. B.
Inward-Rectifier
K + Channels
The RP of maturing my 9 gets closer to E K because the PNa/PK ratio is gradually decreased during development (see Fig. 8B). These characteristics are also observed in
chick embryonic cardiomyocytes during development (Sperelakis and Shigenobu, 1972) and are produced by marked expression of the inward-rectifier K + channels in the surface membrane of the myocytes. Therefore, it seems likely that the hyperpolarization of the RP of skeletal myocytes in culture is produced by a similar developmental change of inward-rectifier K* channels. The IK~IR) is present in skeletal muscle cells from early embryonic amphibian (Linsdell and Moody, 1995). Although there is a brief period (approximately 4 h) during which its density decreases, the overall trend is an increase during development. In ascidian myocytes, IK(IR) exhibited dramatic changes during development (Fig. 10C) (Greaves et al., 1996). In the ascidian, the inward-rectifier K + current is gained after fertilization of the egg. (The same change also occurs in other species.) When gastrulation ends (at 16 h after fertilization), the current density suddenly decreases from 4 to 0.5 pA/pE After the tailbud stage (22 h after fertilization), the current density progressively increases again and reaches a value of 5 pA/pF before hatching. Because IK(IR) is one of the most
593
35. Development Changes in Ion Channels A
Control
Tetrodotoxin
0
V/s 30"
5 days
_/
k._.
m t(9 m
O r
20 -
.u Q.
7 days
O (9 (/} L
{
O (9 lmM 50 p~M
Persistent activation Persistent activation Persistent activation
Polypeptides a-Scorpion Toxins /3-Scorpion Toxins Anemone Toxins
1-2 nM 0.5 nM 10-1000 nM
Delayed inactivation Shifts voltage dependence of activation Delayed inactivation
IxM range 0.6 ng/ml 30-50 nM
Shifts voltage dependence of activation Induces depolarization Delayed inactivation
Dinoflagellate toxins Brevetoxins Ciguatoxin Goniopora
37. Ion Channels as Targets for Toxins on ion selectivity. Subsequent structure-function analysis of TTX and STX, using both natural and synthetic analogs (Mosher, 1986; Shimizu, 1986), clearly delineated the importance of the guanidinium group of TTX (and the corresponding 7, 8, 9 guanidinium of the bifunctional STX) for activity. In addition, the C4, C9, and C10 hydroxyl groups of TTX and the C12 hydroxyl of STX are required. It is assumed that these hydroxyl groups form hydrogen bonds to as yet unknown acceptor sites on the channel. Recent data suggest that the TTX binding site is located near the external face of the channel and may lie outside of the transmembrane field. In the absence of direct information on the three-dimensional structure of the channel, precise localization of the TTX binding site is impossible. While current models for the binding of both toxins still emphasize the importance of fixed negative charges at or near the channel mouth, it is clear that both molecules make multiple contacts to channel residues lying within the linker connecting the putative $5 and $6 helices. Site-directed mutagenesis of both Na + channels and K + channels indicates that this region is important both for ion selectivity and binding of channel-specific toxins. Thus, replacement of Arg 377 of the TTX-resistant cardiac (RH1) channel by asparagine results in a small increase in TTX-sensitivity (Stuhmer et al., 1989a), whereas the mutation of Cys 374 to tyrosine renders the channel approximately 700-fold more sensitive to TTX (Satin et al., 1992). This latter result raises the possibility that the positively charged guanidinium groups of TTX and STX are actually bound not by ionic interactions but rather by interactions with the 7r-clouds contributed by a cluster of electron-rich aromatic residues located near the channel mouth. Analogous models have been proposed to account for acetylcholine binding to the nicotinic acetylcholine receptor, and tetraethylammonium binding to a variety of potassium channels. However, it is clear that interaction with aromatic residues is not the sole binding determinant, since mutagenesis has also identified a carboxyl group, Glu 387, essential for high-affinity TTX binding in the brain RII channel (Noda et al, 1989). Like Cys 374 and Arg 377, this residue is also located in the SS2 region. Thus, whatever the nature of the specific interactions involved in TTX binding, important determinants are clearly found in the SS 1-SS2 region. A very recent development has utilized the ability of TTX to protect mutagenically created channel cysteine residues from reaction with sulfhydryl-specific reagents to develop topographic maps of the channel pore. The corresponding $5-$6 linker regions from domains III and IV are also important determinants of ionic selectivity, as shown by the recent finding that the mutations Lys 1422 to glutamate and Ala 1714 to glutamate give the R n Na t channel an ionic selectivity similar to that of a Ca 2§ channel. Interestingly, mutagenesis has also been used to show that the $5-$6 region of shaker K § channels is directly involved in both ion transit and interaction of inhibitors such as tetraethylammonium (TEA) and charybdotoxin. Tyr 379 of the K § channel RBK-1 has been identified as an important determinant of tetraethylammonium (TEA) binding (MacKinnon and Yellen, 1990), while the introduction of a tyrosine at position 449 of the shaker H4 channel greatly in-
627 creases its sensitivity to TEA. Thus, a leitmotif stressing the importance of aromatic residues in the binding of cationic ligands to ion channels is emerging (Dougherty, 1996). A structurally unrelated toxin, designated/z-conotoxin or geographutoxin, is a competitive inhibitor of STX binding to muscle, but not nerve, Na t channels (K d < 100 nM in muscle versus > 1 IxM in brain; Cruz et al., 1985; Ohizumi et al., 1986)./x-Conotoxin is a 22-residue, hydroxyprolinecontaining peptide, tightly cross-linked by three disulfide bonds. A model for its three-dimensional structure, based on two-dimensional NMR measurements and simulated annealing protocols, has been proposed. The overall structure (Fig. 3) includes a series of tight turns in the N-terminal region and a short stretch of right-handed helix, while the core of the peptide encompasses what the authors refer to as a disulfide cage. A key feature of the structure is that the cationic side chains of arginines 1, 13, and 19, and lysines 8 and 16 all project into solution, away from the core of the molecule as defined by the disulfide bonds. Analysis of synthetic variants of ~-conotoxin has shown that Arg 13 is essential for activity: even conservative replacement by lysine reduces potency by about tenfold. In contrast, Arg 19 can be replaced by lysine with essentially no change in activity, while substitution at Arg 1 yields an intermediate result (Cruz et al., 1989). It is interesting that binding of both TTX/STX and the ~-conotoxins is at least partially dependent on guanidinium functions, despite the disparate chemistries of these molecules. Unlike TTX and STX, the binding of/z-conotoxin is dependent on membrane potential, with K d decreasing e-fold for every 34 mV depolarization (Cruz et al., 1989). TTX/STX-resistant (i.e., cardiac) channels are likewise resistant to ~-conotoxins. Very recently, a mutated form of ~-conotoxin has been used as a monitor of distance between the voltage sensor and channel pore (French et al., 1996).
B. Site 2: Batrachotoxin, Grayanotoxin, Veratridine, and Related Compounds The existence of low molecular weight substances having dramatic effects on the gating properties of voltage-dependent Na t channels has been known for over 20 years. The most commonly used of these alkaloids are batrachotoxin, grayanotoxin, and veratridine. Despite their different biological origins (respectively, skin secretions of the frog P. aurotaenia, Lilaceae, and E r i c a c e a e species) and doseresponse curves (ICs0s are approximately 0.5, 10, and 25 txM for batrachotoxin, grayanotoxin, and veratridine), these molecules have very similar effects on channel function. They all appear to interact at a common binding site, which is allosterically coupled to both the TTX binding site described in the previous section and the a-scorpion toxin site to be discussed later (Catterall, 1977). Binding of any of the alkaloids to this site causes persistent activation of the channel, and leads to membrane depolarization. In addition, alkaloid-treated channels display a more relaxed ionic selectivity than do their naive counterparts. Thus, while the permeability order of untreated Na t channels is Na t : guanidinium : K § : Rb § = 1 : 0.13 : 0.09 : 0.01, veratridine increases the relative permeability for guanidinium, K § and Rb § by factors of 2.7,
628
SECTION IV Ion Channels as Targets for Toxins, Drugs, and Genetic Diseases
energy transfer experiments. Because the effects of BTX on membrane potential are dependent on the presence of extracellular Na § and are blocked by treatment with TTX, they are attributed to the toxin's ability to cause voltage-dependent Na § channels to activate at the resting potential, and to persist in the activated state. BTX also causes increased permeability to larger cations, consistent with allosteric coupling between the BTX binding site and the selectivity filter of the Na § channel. Other alkaloids that appear to interact with the BTX site with similar functional effects include veratridine, grayanotoxin (GTX), and aconitine. While structurally distinct from one another, these highly hydrophobic compounds do have certain common features, such as the bridged ring systems and esterified aromatic moieties found in BTX, veratridine, and aconitine (Fig. 4). In the case of BTX, both the bridge and the esterified group are essential for activity. For all of these toxins, the available data strongly suggest that interaction of a single toxin molecule per channel is sufficient for activation. The known biological effects of these toxins are ascribed to their ability to interact with the voltage-dependent Na + channel, although only BTX binding has been measured directly. Catterall (1977) proposed that the alkaloids bind preferentially to the activated conformation
4.5, and 10, respectively. Qualitatively similar results are obtained with the other alkaloids of this class and with combinations of alkaloid and polypeptide toxins from scorpion or sea anemone venoms. The steroidal alkaloid batrachotoxin (BTX) was purified and characterized by Witkop, Daly, and their coworkers. In collaboration with Albuquerque, the effects of BTX have been measured in a variety of tissues including axons, brain slices and isolated synaptosomes, neuromuscular junctions, heart papillary muscle, and cardiac Purkinje fibers. Batrachotoxin binding has been measured directly in rat brain synaptosomes using a tritiated derivative, and the measured density of binding sites correlates well with that for TTX and STX. Binding of BTX to this site is competitively inhibited by veratridine with a K i very similar to the K0.5 for its activation of the channel. Because BTX acts from either side of the membrane, it had been speculated that this binding site might lie within the lipid bilayer, and a very recent study has demonstrated that a photoactivatable derivative of batrachotoxin specifically labels a site within the $6 transmembrane segment of domain I of the rat brain channel (Trainer et al., 1996). As discussed later, fluorescent derivatives of BTX have also been important for measuring distances between the various toxin binding sites by resonance
A
K16 ( ~ ~
R13
B R19
K18 R
,
R13
( N
1
K8
Variant
Sequence
GIIIA GIIIB GIIIC
RDCCTPPKKC RDCCTPPRKC RDCCT.PPKKC
KDRQCKPQRC KDRRCKPMKC KDRRCKPLKC
CA CA CA
Consensus
RDCCTPP+KC
KDROCKPX+C
CA
9
FIGURE 3. /~-Conotoxin structures. Top: A space-filling model of the three-dimensional structure of/x-GIIIA, as deduced from two-dimensional nuclear magnetic resonance data. Black spheres indicate sidechain nitrogens of Arg, Lys, and the N-terminal group. A and B depict different faces of the molecule, while the structure of tetrodotoxin is shown in scale in C for comparison. (From Lancelin et al. (1991) Biochemistry 30, 6908-6916.) Bottom" Amino acid sequences of three ~-conotoxin variants are shown in the single letter code with P designating hydroxyproline. Disulfide bonds link Cys 3 to Cys 15, Cys 4 to Cys 20, and Cys 10 to Cys 21. In the consensus sequence + = cationic residue; X = ambiguous character.
37. Ion Channels as Targets for Toxins
629
A
"T ~ OH~ v
v
NH
v
/ ~ _ _O_H..... _ (/OCH3
H3C--OL'~ . , ' ~ ...........L,"--OCOCBH5 CH3CH2~~ ...... -ococ.. "
"OH
HO('I'~-~~C H30," OCH3
CHa
C ?H
CHaONj---, 0 GH30---~O~O
H
~ H H
F I G U R E 4. Structures of some major alkaloid agonists of the voltage-sensitive sodium channel. (A) Batrachotoxin, (B) Aconitine, (C) Veratridine.
of the Na + channel, shifting a pre-existing equilibrium between active and inactive states toward the former. Thus, toxin action would best be approximated by the allosteric model of Monod, Wyman, and Changeaux, in which the binding energy for a given toxin is expressed as a shift in the voltage dependence for channel activation. As will be discussed later, conformational changes in the channel induced by binding of toxins have been directly demonstrated, thus supporting this model. C. Sites 3 and 4: Scorpion and S e a A n e m o n e Toxins
An effect of scorpion venom on frog Na + channels was first demonstrated almost 30 years ago, and individual toxins from Androctonus venom were purified and sequenced beginning in the mid-1970s. Subsequently, toxins active on the sodium channel have been purified from the venoms of Androctonus, Leiurus, Centruroides, Tityus, and Buthus species. While these toxins exhibit certain common structural features, the pharmacology of the Androctonus and Leiurus toxins is quite distinct from that of the corresponding polypeptides from Centruroides and Tityus. The former (a-toxin) group delays or abolishes channel inactivation and thus delays repolarization of the membrane through binding to neurotoxin site 3, while the latter (fl-toxin) group shifts the voltage dependence
of activation to more negative potentials as a result of binding to site 4 (in the Catterall nomenclature), resulting in spontaneous activity. Binding of the a-toxins is dependent on membrane potential, with K d values in the nanomolar range at the resting potential. In contrast, /3-toxin binding is potentialindependent and generally of slightly lesser affinity. The pharmacologic distinction between the a- and/3-toxins should be emphasized, in view of the close relationship among these toxins at the level of covalent and tertiary structure. The primary structures of representative scorpion toxins are depicted in Fig. 5. All scorpion toxins characterized to date are cationic polypeptides of molecular mass approximately 7000 Da that contain 4 disulfide bonds whose integrity is essential for activity; in addition, they display a significant degree of sequence homology, with approximately 30% of all positions being conserved among the known a-toxins. The/3-toxins, exemplified by Centruroides toxin 3, are generally similar to those of the a group, but display less extensive homology. It is, however, possible that much of this sequence conservation within a group is essential simply for allowing proper folding of these toxins, similar to results obtained recently with the conotoxin family, but until residues important for interaction with the channel binding site have been definitively identified, this question will remain unresolved. Nonetheless, three-dimensional structures of representatives of each group show a good deal of similarity in the peptide backbones. With respect to structure-activity correlations, it has been shown that alkylation of Lys 56 of AaH-I is correlated with its inactivation, suggesting the importance of a cationic group at this position in the a-scorpion toxins, and that modification of Arg 2 and Arg 60 likewise results in inactivation. While other residues have been implicated in activity by treatment of one or another of the scorpion toxins with group-selective reagents, the results are somewhat fragmentary, and no overall patterns have as yet emerged. While at first glance this paucity of information might appear surprising, given the small size of the toxins and their ready availability, it must be remembered that prior to the advent of molecular biological approaches, the most accessible techniques (e.g., chemical modification using group-specific reagents) for identifying essential residues in proteins were developed to probe substrate binding to enzymes and their catalytic mechanisms. In general, these methods do not adapt well to proteins lacking superreactive residues, such as the toxins now under consideration. The ability to mutagenize peptide neurotoxins has in the past few years led to more extensive knowledge of the relationship between theft structure and function, as exemplified by analysis of anemone toxins described in a later section. The three-dimensional structures of a number of scorpion toxins are now known, including members of both the a- and /3-classes (Fig. 6). As might be anticipated from the highly cross-linked nature of these proteins, both possess a dense core of secondary structure connected by two of the four disulfides; this core contains three strands of antiparallel/3sheet (residues 1-4, 37-41, and 46-50) and a short strand of a-helix (residues 23-32). In addition to several loops that protrude from this core region, a prominent feature of both structures is a relatively nonpolar surface region, which contains a number of conserved aromatic residues. Because chemical modification of sites external to this surface region
630
SECTION IV Ion Channels as Targets for Toxins, Drugs, and Genetic Diseases
Representative a-toxins
F I G U R E 5. Primary structures of representative a- and B-scorpion toxins. Sequences are shown in the single letter code, and disulfide bonds link Cys 12 to Cys 63, Cys 16 to Cys 36, Cys 22 to Cys 46, and Cys 26 to Cys 48 in both groups. Numbering is based on the sequence of AaH-II. Highlighted positions are either invariant or occupied by functionally equivalent residues within the a or fl subgroup. In the consensus sequence, the coding is: + = cationic residue; O = polar residue; X - character undefined; U = nonpolar residue; ~ - aromatic residue; [3 = branched side chain. AaH, Androctonus australis Hector; Lqq, Leiurus quinquestriatus quinquestriatus; BoT, Buthus occitanus Tunetanus; CsE, Centruroides sculpturatus Ewing; TsS, Tityus serrulatus.
FIGURE 6. Three-dimensional structure of Centruroides toxin variant 3. Structural coordinates were obtained from the Brookhaven Protein Database, with permission. Although variant 3 is a/3-toxin, the overall folding patterns of the a- and fl-toxins are highly similar (Fontecilla-Camps et al. (1988) Proc. Natl. Acad. Sci. USA 85, 74437447). The backbone of the molecule is depicted as a ribbon, and the hydrophobic surface proposed to be important in toxin-receptor interactions is shown as a dot surface. All side chains have been omitted for the sake of clarity.
37. Ion Channels as Targets for Toxins
Consensus
XUXCXCOODG
631
POXROOOUOG
f~UXUXXXOCX
OGWOOCOOXX
XXXOXCC+O
FIGURE 7. Primary structures of representative sea anemone toxins. Sequences are shown in the single letter code, with the numbering based on the sequence of anthopleurin A. Disulfide bonds link Cys 4 to Cys 46, Cys 6 to Cys 36, and Cys 29 to Cys 47. Highlighted positions are either invariant or contain functionally conservative replacements. In the consensus sequence, the coding is + = cationic residue; O = polar residue; X = character undefined; U = nonpolar residue; 9 = aromatic residue; 13= branched side chain. As, Anemonia sulcata; Ap, Anthopleura xanthogrammica; Sh, Stichodactyla helianthus; Rp, Radianthus paumotensis.
is, in general, not deleterious to toxin function, and because Lys 56, referred to previously, lies within it, FontecillaCamps and coworkers first suggested that this region was essential for receptor binding. Interestingly, this surface is fairly well conserved in both the a- and/3-toxins, despite their distinct pharmacologies and nonoverlapping binding sites. Indeed, all of the structural features discussed here are common amongst scorpion toxins regardless of binding site or species specificity. In addition to the a-scorpion toxins, site 3 also binds a variety of polypeptide toxins derived from marine invertebrates: the best characterized of these are a diverse family of sea anemone toxins derived from A n e m o n i a , Anthopleura, Stichodactyla, and Radianthus species (Catterall and Beress, 1978). Under appropriate conditions, certain of these polypeptides can function as cardiac stimulants. While also basic and rich in disulfides, the anemone toxins differ from those discussed previously in being somewhat smaller (46-50 residues versus 60-65 residues) and having three, rather than four, disulfides. Moreover, despite the wealth of structural information now available, no sequence homology has been found between neurotoxins from scorpions and anemones. Representative anemone toxin sequences are shown in Fig. 7. Within the past few years, the solution structures of four anemone toxins have been solved (Fig. 8). These structures, all having a core of twisted, four-stranded, antiparallel /3pleated sheet (residues 2-4, 18-23, 31-34, and 42-47 in anthopleurin A numbering), demonstrate unequivocally that anemone and scorpion toxins are structurally unrelated. Nonetheless, anemone and a-scorpion toxins have very similar pharmacologies and display mutually competitive binding to the Na + channel, suggesting that at least a limited degree of relatedness may yet remain to be found. It may be significant that all anemone toxins contain a large loop region (approximately residues 9-20) whose solution structure is very poorly defined. At least two of the amino acid residues known to be important for anemone toxin binding are found within this loop. In addition, biophysical analyses have provided evidence for the exposure of a number of hydrophobic residues on the surface of homologous anemone toxins. In anthopleurin B, the functional roles played by a subset of these residues have been analyzed using site-directed mutagenesis
FIGURE 8. Three-dimensional structure of anthopleurin-B, a representative anemone toxin. The structure shown represents a model developed in the author's laboratory (Khera et al., 1995) based on the coordinates of the homologous toxin from Stichodactyla helianthus (Fogh et al. (1990) J. Biol. Chem. 265, 13016). The overall folding patterns for anemone toxins of known structure are similar. The model is also consistent with the recently determined solution structure of ApB (Monks et al., (1995) Structure 3, 791). The backbone is shown in ribbon format, and the side chains of residues which have been shown to be important determinants of binding affinity are annotated. The van der Waals surface depicts a region of partially exposed hydrophobic residues, including the required Leu 18 and Trp 33, which may be the functional equivalent of the hydrophobic surface involved in scorpion toxin binding to the sodium channel.
of a synthetic gene capable of encoding the toxin polypeptide. These studies have clearly demonstrated that both Leu 18 and Trp 33 play important roles in regulating high-affinity binding of this toxin to both neuronal and cardiac sodium channels, and further show that two other exposed hydrophobic sites, Ile
632
SECTION IV Ion Channels as Targets for Toxins, Drugs, and Genetic Diseases
43 and Trp 45, are external to this binding epitope. Thus, the former pair of residues may act as the functional equivalent of the hydrophobic surface implicated in scorpion toxin binding (Dias-Kadambi et al., 1996a,b). Indeed, charge-neutralizing mutations of all the ionizable groups in this toxin reveal that the contribution to affinity made by hydrophobic contacts is far more significant than any involving electrostatic interactions. Surprisingly, even Arg 14, which is conserved in all known anemone toxins, is revealed by mutagenesis to play at most a minor role in defining binding affinity (Khera and Blumenthal, 1994). While sea anemone and c~-scorpion toxins bind to the channel in a membrane-potential-dependent manner and their binding sites overlap at least in part,/3-scorpion toxins bind to a demonstrably distinct site on the same protein and their binding is potential-independent. All of the known polypeptide toxin binding sites are structurally distinct from those for the alkaloid activators and guanidinium inhibitors. Affinity labeling experiments, using derivatives of either TTX or Leiurus toxin, have demonstrated that the binding sites for both toxins are found on the c~-subunit of the channel. Like the TTX/STX binding site discussed previously, a photoactivatable derivative of Leiurus toxin labels residues between the $5 and $6 helices (i.e., SS 1-SS2) of domains I and IV of the rat brain channel. In some cells, the/31-subunit of the channel is also photolabeled, leading to the suggestion that the scorpion toxin binding site may be located near a subunit interface. Labeling of both subunits is specific, being blocked in the presence of excess unlabeled ligand, or by membrane depolarization. While these experiments demonstrate convincingly that scorpion toxin is capable of labeling the channel in the SS 1-SS2 region of domains I and IV, they do not prove that interaction at this site is responsible for the physiologic consequences of scorpion toxin binding. Interestingly, recent genetic studies have highlighted the role of other extracellular sequences in modulating channel inactivation kinetics. The dominantly heritable disorder paramyotonia congenita, which displays an electrophysiologic phenotype identical to that of an anemone or scorpion toxin-treated, wild-type channel, has been mapped to a set of missense mutations in the $3-$4 linker region of domain IV of the muscle sodium channel (Cannon and Corey, 1993; Chahine et al., 1994; Yang et al., 1994). Biochemical analyses have demonstrated that Glu 1613, located in this region of the neuronal channel (Rogers et al., 1996), is an important determinant of Leiurus toxin binding, while Benzinger et al. (1998) have shown that the analogous Asp 1612 of the cardiac channel interacts electrostatically with Lys 37 of the anemone toxin anthopleurin B.
D. Other Toxins Specific for the Na § Channel A variety of nonproteinaceous substances of diverse (and in some cases, unknown) structure have been shown to interact with the Na + channel at sites distinct from those discussed previously. These include lipid-soluble dinoflagellate toxins (brevetoxins, ciguatoxin, and maitotoxin), the Goniopora toxins, and the pyrethroid insecticides. The brevetoxins induce the Na § channel to open at membrane potentials o f - 8 0 t o - 1 6 0 mV, and also abolish rapid inactivation; these effects lead to both generation of repetitive ac-
tion potentials and ultimately membrane depolarization. Both effects are blocked by TTX. Analyses of brevetoxin binding suggest the existence of a fifth ligand site on the channel, since brevetoxin binding fails to displace toxins bound to sites 1-4. The physiologic effects of ciguatoxin are very similar to those of the brevetoxins, while the mechanism of maitotoxin toxicity is unresolved at this time. Goniopora toxins, while not completely characterized, appear to be protein in nature, with molecular weights between 12 000 and 19 000 being reported from different preparations. The mode of action of at least one of these proteins appears to be similar to that of sea anemone toxins, but there is evidence that multiple Goniopora toxins with distinct molecular targets exist. Finally, the pyrethroids delay inactivation via interaction with a site distinct from that for polypeptide toxins having the same effect. Physiologic effects of pyrethroids are antagonized by TTX.
E. Relationships among Sites 1-4 The best data available from direct binding studies support the notion that each channel molecule contains single sites for each class of toxins. The next obvious question relates to identification of these sites and characterization of interactions among them. Angelides and coworkers have estimated the distances between the various toxin sites using resonance energy transfer between fluorescent derivatives of the bound ligands. In addition to confirming that the a and /3 sites are unique and separated by about 22 ,~, these studies indicate that the TTX site is some 33/k removed from that for/3-toxins and the BTX site lies about 37/k distant from that of o~-toxins. Analogous experiments have also been used to verify that the sites involved are conformationally linked, with distances between a given pair of ligands changing upon binding of a third. Because multiple subtypes of the Na + channel have been characterized, it is natural to wonder whether different toxins are directed against specific channel isoforms, either in a tissue-specific manner or not. As discussed below, this may be the case for K + channels. However, to date no convincing evidence has been presented that a given (e.g.) scorpion toxin distinguishes among Na + channel subtypes within an individual tissue, although it is clear that toxins do differ substantially in their affinities for channels that are expressed in a tissue-specific manner (Khera et al., 1995). Obvious examples of this phenomenon would include the specificity of/x-conotoxins for muscle Na + channels, and the tenfold affinity preference displayed by some sea anemone toxins for cardiac, as compared to neuronal, Na + channels.
I!!. V o l t a g e - A c t i v a t e d a n d Ca2§ Potassium Channels Our knowledge of the structure, function, and diversity of potassium channels has increased dramatically within the past five years, with the application of combined molecular genetic, electrophysiologic, and crystallographic approaches (Miller, 1991). As will be seen, despite the enormous power
633
37. Ion Channels as Targets for Toxins
of the genetic approach, the ability of polypeptide toxins specific for this family of channels to distinguish among subtypes is proving an invaluable tool. In addition, both polypeptides and nonproteinaceous blockers, such as tetraethylammonium ions and 4-aminopyridines, have proven to be valuable tools for development of structural models of the K + channel pore. These models are likely to apply to certain aspects of the homologous Na + and Ca 2+ channels as well. In the case of K + channels, toxins from both arthropods and snakes have provided a wealth of pharmacologically significant data (summarized in Table 2), and very recently a new class of polypeptide K + channel blockers has been purified from sea anemone cnidocytes. While voltage-dependent Na + channels exist in a relatively small number of molecular forms, the diversity seen in K + channels is vastly greater: some are ligand-coupled, while others are activated by either Ca 2+ or depolarization. A large number of channels seem to exist even within this last category. In Drosophila, K + channels are encoded by genetic loci designated as shaker, shal, shab, shaw, and eag, and mammalian counterparts for each of these channel types have been identified. Shaker apparently encodes channels of the rapidly inactivating (or A) type, and transcripts from this locus display alternative splicing. It seems likely that this gives rise to channel forms that differ with respect to virtually all important properties: pharmacology, voltage-dependence, and unit conductance (Stuhmer et al., 1989b). Multiple transcripts from the shab locus have also been observed, and the gene products have many of the properties associated with the delayed rectifier channel. The shal and shaw genes seem to encode A-type and delayed rectifier channels, respectively. In addition to this multiplicity of voltage-gated K + channels, there exists an array of channels gated by Ca 2+ as well. Because the unit conductances within this group vary quite widely, it is generally divided into high-conductance (BK or maxi-K, 130-300 pS) and small conductance (SK, 10-50 pS) subtypes. The amount of structural information presently
TABLE 2
Polypeptide Blockers of K +
Channels Toxin
Channel type
Kd
Charybdotoxin
Ca2+-activated K + Voltage-sensitive K +
3.5 nM 140 pM
Iberiotoxin
CaZ+-activated K + (BK)
250 pM
Kaliotoxin
Ca2+-activated K + (BK)
20 nM
Noxiustoxin
Ca2+-activated K+ (BK)
>300 nM
Scyllatoxin
Ca2+-activated K+ (SK)
80 pM
Apamin
SK
10--400 pM
Dendrotoxins
Voltage-sensitive K +
100 pM
available on these channels is significantly less than for the A-type channels, and because low-stringency hybridization screening using known K + channel probes has failed to identify members of this group, it is inferred that significant differences exist at the primary structure level despite some toxin cross-reactivity. Toxin binding may thus provide a means for purification of (e.g.) BK channels, and partial sequence analysis would then allow design of oligonucleotide probes, accessing additional structural information using classic molecular biological approaches. The channels encoded by shaker and related genes have many structural features in common with the Na + channels discussed above, and these proteins display a limited degree of homology at the amino acid level. In both cases, the channel "subunit" contains six putative transmembrane helices with the N-terminus of the first being cytoplasmic. Helix $4, which contains a series of cationic groups separated by two hydrophobic residues, is postulated to serve as the voltage sensor, perhaps with some participation by charged residues in $2 and $3. As with the Na + channel, sequences joining the helices designated $5 and $6 are thought to provide the inner lining of the pore, and a great deal of evidence that these residues also provide the binding sites for channel-specific toxins has been accumulated in the past few years. The most obvious structural difference between these channels is that whereas the Na + channel c~-subunit has a molecular mass of about 260 kDa and contains some 1800 amino acid residues grouped into 4 homologous domains, the K + channel monomer is approximately onequarter this size. While the monomeric unit of the K + channel is thus small relative to its Na + and Ca 2+ counterparts, coexpression of mRNAs encoding distinct channel isoforms in Xenopus oocytes results in the appearance of functional channels having intermediate properties, strongly suggesting that the physiologically relevant form of this channel is a noncovalent tetramer. The possibility of functional diversity arising from both alternative splicing and formation of hetero-oligomeric channels thus arises. In 1998, MacKinnon's lab solved the crystal structure of the Streptomyces lividans K + channel (Doyle et al., 1998). While this protein is only about 30% the size of the eukaryotic voltage-gated K + channels, it displays significant homology to (e.g.) the shaker and other eukaryotic voltage-gated channels, particularly within the putative pore-forming regions. As a result, the S. lividans channel represents an attractive starting point for modeling the three-dimensional structures of the eukaryotic pore. In the bacterial channel, this region contains a 10-residue a-helix (the pore helix) whose C-terminus is oriented toward the center of the pore joined to the C-terminal transmembrane helix by a 20residue linker that lacks regular secondary structure. The helices flanking the pore region are proposed to be analogous to the $5 and $6 helices of eukaryotic K + channels. Most of the known K+-channel-specific toxins have been purified from scorpion, honeybee, and snake venoms. The first such agent to be characterized was purified by Miller from Leiurus venom, and given the colorful, if nondescriptive, name eharybdotoxin (Miller et al., 1985). Homologs of this polypeptide have been found in venoms from Leiurus (scyllatoxin and agitoxin), Androctonus (kaliotoxin), Buthus
634
SECTION IV Ion Channels as Targets for Toxins, Drugs, and Genetic Diseases
FIGURE 9. Primary structures of polypeptide toxins specific for K+ channels. Sequences are depicted in the single letter code, with disulfide bonds linking Cys 7 to Cys 28, Cys 13 to Cys 33, and Cys 17 to Cys 35 in charybdotoxin. Highlighted residues are invariant within a class. In the consensus sequence, the coding is + = cationic residue; O = polar residue; X = character undefined; U = nonpolar residue; ~ = aromatic residue; [3 = branched side chain. ChTX, charybdotoxin; ScTX, scyllatoxin; KaTX, kaliotoxin; IbTX, iberiotoxin; NxTX, noxiustoxin; aDTx, a-dendrotoxin; Dtx 1, dendrotoxin I; kalicludine, K§ channel blocker from Anemonia sulcata.
(iberiotoxin), and Centruroides (noxiustoxin) species. In addition, the honeybee peptides apamin and mast cell degranulating peptide, and a homologous family of snake (Dendroaspis) dendrotoxins have been shown to have K +channel-blocking activity. Analogous to the case for sodium channels, the potassium channel toxins from scorpion venom comprise an homologous group in which essentially the only absolutely conserved residues are involved in disulfide bond formation (Fig. 9). Thus, one would reasonably expect these toxins to fold into a generally similar overall structure, but to differ in detail. It is therefore not surprising to realize that at least some of the K + channel toxins are able to distinguish among channel subtypes. The honeybee toxins and dendrotoxins display no homology to any of the scorpion polypeptides. Charybdotoxin (ChTX) was first identified in 1985 as a polypeptide from Leiurus venom capable of inhibiting Ca 2+activated K + channels from skeletal muscle. Amino acid sequence analysis of the purified toxin revealed a primary structure completely distinct from that of other known scorpion toxins. The three-dimensional structure of ChTX was solved by NMR spectroscopy in 1991 (Fig. 10; Bontems et al., 1991 a,b; 1992). Its major features include a small, triplestranded /3-sheet encompassing residues 1-2, 25-29, and 32-36 linked by disulfide bonds (Cys 13-Cys 33 and Cys 17-Cys 35) to an a-helix including residues 10-19. Surprisingly, however, the overall fold of this molecule displays significant similarities to both the Centruroides and
Androctonus Na+-channel-specific toxins for which crystal structures had been obtained during the 1980s. It is important to recall that comparison of the sequences of the three toxins shows that only the positions of the disulfide bonds are conserved. However, in three dimensions, additional common features of these toxins are revealed, including
FIGURE 10. Three-dimensional structure of charybdotoxin as deduced from NMR experiments. Positions of the backbone atoms of each of the 12 contributing structures in the PDB file pdb2crn.ent (available at www.rcsb.org) are depicted. Original data contributing to this structure were reported in Bontems et al. (1991a).
37. Ion Channels as Targets for Toxins
three disulfide bonds occurring in similar positions and orientations, the triple-stranded/3-pleated sheet (residues 1-4, 37-41, and 46-50 of Centruroides toxin versus residues 1-2, 25-29, and 32-36 of charybdotoxin), the single strand of c~-helix (residues 23-32 versus 10-19) and a significant proportion of extended structure (Fig. 11). The differences in comparing the structures of these long and short neurotoxins are most pronounced in loop regions connecting the elements of defined secondary structure. Thus, it would appear that the scorpion neurotoxins as a group share a common structural motif, apparent only at the level of three-dimensional structure, regardless of the nature of their target channel. The apparent similarities in channel folding thus seem to be mirrored in the structures of their "ligands." Interestingly, this motif appears to be similar to one observed in insect defensins, small peptides produced in response to injury that are thought to play a role in protecting insects against bacterial infections (Bontems et al., 1992). It seems counterintuitive that the scorpion toxins, despite their vastly different pharmacologies, display conservation of tertiary structure, while anemone and scorpion ce-toxins, which show very similar pharmacologies, do not. Perhaps this is simply indicative of the fact that our knowledge of the molecular interactions important for productive binding of these toxins to their target sites remains incomplete (however, see the following). However, it should be noted that the very recently characterized K+-channel-blockers from anemones (kalieludines and kaliseptines, e.g.), which are functional homologs of charybdotoxin and dendrotoxin, appear to share neither this structural motif, sequence homology, nor even disulfide connectivities. Because ChTX possesses a net charge of +5, it was predicted that ionic interactions with the channel would be important in its binding. This prediction is supported by the observations that (1) ChTX binding affinity is decreased 100-fold in a medium of increased ionic strength (300 mM salt); this reduction in binding is due to a decrease in toxin association and (2) chemical modification of the channel with the carboxylate-specific reagent trimethyloxonium (TMO) ion greatly reduces ChTX binding. The structure of the S. lividans channel reveals that the binding site for these cationic toxins is rich in acidic amino acid residues. Affinity also depends on membrane potential, with ChTX preferring the open conformation by approximately seven-fold. More recent studies with charybdotoxin mutants are consistent with hydrophobic interactions between channel sites and Met 29 and Tyr 36 of ChTX as also being important for binding (MacKinnon and Miller, 1989). Although charybdotoxin was discovered because of its ability to block BK channels, evidence has since accumulated that its specificity is less well-defined. It now seems clear that voltage-dependent K + channels in frog ganglia, and SK channels from Aplysia among others, are also blocked by ChTX. However, because it affords us the ability to precisely map ChTX-channel interactions, the most interesting new target for ChTX is the shaker channel. To test the importance of ionic interactions in ChTX binding to this target, acidic residues located in extracellular sequences of the channel were altered by site-directed mutagenesis, and the kinetics of ChTX blocking measured. This strategy iden-
635
FIGURE 1 1. Comparison of the three-dimensional structures of (A) Centruroides variant 3 and (B) charybdotoxin. Backbone structures are shown as ribbons, representing the regions of highest similarity between the two polypeptides: a-helices 10-20 (charybdotoxin) and 22-32 (variant 3), and/3-strands 1-2, 25-29, and 32-36 (charybdotoxin) and 1-4, 37-42, and 44-48 (variant 3). Images were created from the PDB files pdb2sn3 and pdb2crd, respectively, using InsightlI for graphics display. Structural coordinates are reported in Bontems et al. (1991), and may be found in the PDB files cited at www.rcsb.org. tiffed Glu 422, located in the SS1 region, as important for binding: substitution by aspartic acid had no effect while conversion to glutamine or lysine decreased blocking activity by 3- and 12-fold, respectively. Very recent data have identified another point mutation (Phe 425 to Gly) whose major consequence is to render the shaker channel approximately 2000 times more sensitive to ChTX. From both the nature of the substitution made and the magnitude of the change observed one might reasonably ascribe this increase to a major conformational change in the toxin binding site. More recently, application of charybdotoxin and other homologous peptide blockers of potassium channels, notably agitoxin-2, as molecular calipers capable of providing topographical information on the vestibule of the potassium channel, has provided exciting results. For the sake of clarity, description of these results will be restricted to the charybdotoxin/agitoxin:shaker system, although analyses of
636
SECTION IV Ion Channels as Targets for Toxins, Drugs, and Genetic Diseases
other toxin:channel pairs have also proven quite useful. Miller's laboratory first systematically mutated all solventexposed residues of charybdotoxin, allowing identification of a subset of residues whose alteration had major effects upon its affinity for the shaker channel (Goldstein et al., 1994). Subsequently, MacKinnon and coworkers (Hidalgo and MacKinnon, 1995; Ranganathan et al., 1996) demonstrated by co-mutagenesis of agitoxin and the $5-$6 region of the shaker channel the existence of interactions between Arg 24 (toxin) and Asp 431 (channel), Ser 10 and Gly 425, Lys 27 and Thr 449, and Arg 25 and Met 448. Miller's group found that Met 29 and Thr 449 and Lys 31 and Asp 427 also form interacting pairs (Naranjo and Miller, 1996; Naini and Miller, 1996). Interestingly, this last interaction was first predicted based on previous models of the charybdotoxin:shaker complex, knowledge of the toxin's solution structure, and considerations of symmetry in the tetrameric channel. Importantly, docking and mutagenic analyses based on the crystal structure of the bacterial channel validate all of the predictions of interacting pairs made in the absence of structural data (MacKinnon et al., 1998). From such studies as these, a low-resolution structural map of the outer vestibule of the eukaryotic channel is beginning to emerge (Fig. 12). Since nonpeptide pore blockers that function at the internal mouth of the channel are available, and channel chimeras and mutagenesis have identified residues at or near the cytoplasmic ends of $5 and $6 as essential to their binding, an analogous approach to mapping its structure may also prove fruitful. That the high-affinity polypeptide K + channel blockers from sea anemone venom are structurally distinct
FIGURE 12. Proposed topography of the external pore and vestibule of the shaker K+channel. A hypothetical single shaker subunit, a 90~ sector of the outer pore region and vestibule, is displayed. Residue numbers given are referenced to the P-region (P0-P20) or to shaker numbering. Push pins indicate residues whose positions relative to the pore axis were established by mapping with pore-blocking polypeptides. Shaded residues are predicted to project into the aqueous pore or vestibule, while unshaded residues are proposed to project away from the aqueous phase. No predictions have been made regarding side-chain orientations for those residues having diagonal striping, Reproduced from Lu and Miller (1995) Science 268, 304, with permission.
from the charybdotoxin family makes them yet another potentially fruitful probe of pore three-dimensional structure (Tudor et al., 1996). The effects of ChTX on shaker homologs cloned from rat brain (RCK channels) have also been examined. These experiments indicate a correlation between toxin affinity and the presence of negative charges in the $3-$4 loop. While this correlation remains to be verified experimentally, involvement of the $3-$4 loop in charybdotoxin binding would be of special interest in view of the demonstrated involvement of this region of the Na + channel in inactivation (see page 332). Because ChTX displays high-affinity interactions with so many different K + channels, the search for more specific probes has continued and a number of groups have purified and characterized new toxins whose targeting may be more stringent. Iberiotoxin (IbTX), isolated from Buthus venom on the basis of its ability to inhibit ChTX binding, is 70% homologous to ChTX but significantly less basic. IbTX inhibits ChTX binding to smooth muscle noncompetitively, consistent with its having a distinct binding site, and data available to date are consistent with IbTX recognizing only Ca2+-activated channels in a variety of tissues (Galvez et al., 1990). Studies with chemically-synthesized chimeras of the two toxins suggest that C-terminal sequences of each toxin dictate channel specificity to a large degree: a construct containing the N-terminal 19 residues of ChTX and the C-terminal 18 from IbTX has IbTX-like properties, and vice versa. Two additional scorpion toxins have been shown to target the B K-type channel, at least preferentially. Kaliotoxin, recently purified and characterized from Androctonus venom, is approximately 50% homologous to ChTX, IbTX, and Noxiustoxin and has been shown to suppress the whole-cell Ca2+-activated K + current in mollusk and rabbit sympathetic nerve. The affected channel is insensitive to both apamin and dendrotoxins, and is blocked by ChTX and TEA (Crest et al., 1992). Noxiustoxin (NxTX), purified from Centruroides venom, is 50% identical to ChTX, but due to its relatively low affinity in many systems, and its apparent interaction with a multiplicity of channel types, it seems unlikely that NxTX will be widely useful for channel purification and/or discrimination studies. NxTX appears to bind most tightly to Jurkat cells and to the ChTX-sensitive channel from brain. In contrast, its affinity for Ca2+-activated channels lies in the 0.1-1.0 pM range. Two distinct, nonhomologous toxins directed at the SK channel are now known: apamin and scyllatoxin; the latter is 25% identical to ChTX, although most of this relatedness defives from alignment of the cysteine residues. Apamin is an 18-residue polypeptide containing two disulfide bonds; it has an a-helical core comprising residues 9-15 flanked by regions rich in r-turns. Analysis of synthetic analogs of apamin highlights the importance of Arginines 13 and 14 for activity. Electrophysiologic analyses and ion flux experiments indicate that apamin is directed against the SK channel in guinea pig hepatocytes, murine neuroblastoma cells, and cultured rat muscle cells. Iodinated apamin has been used to demonstrate the existence of high affinity (K d = 10-400 pM) receptors in rat brain, with affinity increasing with [K+]. The chemical na-
637
37. Ion Channels as Targets for Toxins
ture of the apamin-binding protein has been studied by radiation inactivation and chemical cross-linking, but the results have been inconclusive, with multiple polypeptides being labeled. Scyllatoxin seems to be directed against the same target as apamin, despite the fact that the two polypeptides display no sequence homology. In all systems analyzed to date, scyllatoxin competes with apamin for binding and/or exhibits the same functional effects as apamin. Although the latter toxin appears to bind more tightly (Kd is 15 pM for apamin and 80 pM for scyllatoxin), both proteins display very high affinities (Auguste et al., 1992). While the positions of defined helix and sheet structures are conserved in charybdotoxin and scyllatoxin (with the exception of the first/3-strand, which is absent in the latter polypeptide), recent results indicate that these toxins, which act on different K + channel subtypes, are structurally distinct. In addition to lacking strand/3-1, scyllatoxin also lacks many of the residues known to form an important hydrophobic cluster in charybdotoxin. Sequence homology suggests that this hydrophobic cluster is also maintained in noxius- and iberiotoxins. In addition, scyllatoxin can be distinguished from the others based upon positioning of cationic side chains. Three positively charged residues (Arg 25, Arg 34, and Lys 31) of charybdotoxin are replaced by leucine, aspartic acid, and glutamic acid, respectively, in scyllatoxin. Two of these residues, Arg 25 and Arg 34, have been shown to be important for interaction of charybdotoxin with the rat muscle CaZ+-activated K + channel: mutagenesis to glutamine causes a large increase in the rate of toxin dissociation. Thus, scyllatoxin, which targets the SK channel, lacks these cationic sites but contains alternative ones that appear to be important for channel interaction. Modification of Arg 18 and His 38 of scyllatoxin appears to abolish its binding to rat brain channels. One possible interpretation of these data is that the charybdotoxin subgroup utilizes mainly positive residues in the/3-sheet in its primary binding interaction, while scyllatoxin uses cationic sites i n both the/3-sheet and the c~-helix. Dendrotoxins comprise yet another family of toxins which target the K + channel and at present are the only snake (Dendroaspis species) toxins known to do so. Dendrotoxins (Dtx; four homologs known) are grossly similar to the toxins discussed previously, being highly cationic molecules containing 57-59 amino acid residues crosslinked by three disulfide bonds. While there is no apparent sequence homology between the Dtx's and the other toxins described previously, the Dtx's are clearly related to the small subunit of/3-bungarotoxin and to bovine pancreatic trypsin inhibitor (BPTI). VerYorecently, the crystal structure of ce-Dtx has been solved at 2 A resolution, showing the similarity between Dtx and B PTI. Although the authors failed to discuss this structure in the context of those of other polypeptide neurotoxins, examination of the relevant structures does not reveal any obvious relationships. The similarity between Dtx and BPTI is even more curious in view of the fact that the latter protein binds to an internal site of the BK channel from skeletal muscle and that this binding is associated with the generation of subconducting states. Dendrotoxins were first identified by their ability to stimulate neurotransmitter release at both central and peripheral
neurons. Available electrophysiologic data now indicate that ~-Dtx recognizes a variety of voltage-dependent K + channels. The Dtx sensitivity of RCK channels expressed in Xenopus oocytes has been analyzed, showing that rapidly inactivating channels are less sensitive than slowly inactivating ones, although all four types were responsive. To date, no data demonstrating an effect of the Dtx's on CaZ+-activated currents have been published. Based on immunologic cross-reactivity, the ce-Dtx receptor in mammalian brain is likely related to proteins encoded by the shaker locus, and mutations in the $5-$6 linker of RBK1 have been shown to strongly influence Dtx binding to this shaker homolog.
IV. Voltage-Dependent Calcium Channels Voltage-dependent Ca 2+ channels are a good deal more complex than those discussed in Section III, containing 5 subunits ranging in size from approximately 30 kDa to 170 kDa (Catterall, 1991). Nonetheless, certain structural features of the Ca 2+ channel are strikingly similar to those of the Na + and K + channels. Most importantly, the ~1 subunit of the calcium channel is 29% identical to the c~ subunit of the rat brain R H Na + channel, and the two proteins have functionally related but chemically distinct residues at about the same frequency. Application of the usual predictive structural algorithms to the c~1 subunit suggest the protein to be organized into four homologous structural domains, each containing the now-familiar six transmembrane helices. In addition, the sequence of the $4 helix suggests that, like its counterpart in Na + and K + channels, this region functions as the voltage sensor. Heterologous expression experiments have shown that the c~1 subunit is essential to the properties of dihydropyridine-sensitive Ca 2+ channels. Knowledge of the toxicology of Ca 2+ channels is not extensive, at least as regards polypeptide effectors. There are no known scorpion, anemone, or snake toxins directed at this channel molecule. As of this writing, the only known peptide blockers of the Ca 2+ channel are the omega conotoxins and a family of peptides from funnel web spiders that are structurally unrelated to the conotoxins; both groups prevent transmitter release at presynaptic terminals (Venema et al., 1992). At least seven to-conotoxins have been described; their common features include their cationic nature, the presence of three disulfide bonds, amidated C-termini, and a large number of hydroxylated amino acids, including in some cases hydroxyproline (Fig. 13). Two basic classes of these conotoxins are known, and homology across these classes amounts to about 40%. However, it should again be emphasized that this value may be misleading in view of the fact that approximately half the identical residues are involved in disulfide bonds. At least two classes of spider toxins, distinguishable by size and amino acid sequences, are now known and have been designated either as agatoxins (not to be confused with agitoxin, discussed previously) or curtatoxins. The shorter group (/z-toxins) contains 36-38 residues and is cross-linked by four disulfides; in addition, these/z-toxins are distinguished from the polypeptides discussed previously in being generally less cationic. A second group of spider toxins (omega toxins) contains 65-75 residues and as many as six disulfides, and is more cationic
638
SECTION IV Ion Channels as Targets for Toxins, Drugs, and Genetic Diseases
eAga- IVA
LA
FIGURE 13. (A) Primary structures of omega-conotoxin variants. The sequences are shown in the single letter code, with invariant residues or functionally conservative residues highlighted. Disulfide bonds link Cys 1 to Cys 16, Cys 15 to Cys 27, and Cys8 to Cys20. (B) Primary structures of representative spider toxins specific for Ca2+ channels. Highlighted positions are invariant within a group or occupied by functionally conservative replacements. Where complete sequences have not been determined, N-terminal sequences are shown. Aga, agatoxins; CuTX, curtatoxins.
than the shorter group. Sequences of representative spider toxins are also shown in Fig. 13, emphasizing their lack of relatedness to the conotoxins at the level of primary structure. Omega conotoxins irreversibly block stimulus-evoked transmitter release in frog neuromuscular preparations; because the muscle retains the ability to respond to a direct stimulus, a presynaptic target for the toxin is indicated. That this target is the Ca 2+ channel has since been confirmed in chick dorsal root ganglia (DRG). The effect of this toxin seems to be irreversible: in experiments in which DRG was treated with 100 nM toxin, then washed for 2 hours, no recovery of Ca 2+ activity was observed. Toxins isolated from funnel web spider venom are in general less well-characterized than the conotoxins as regards
their target channel. The unfractionated venom has been shown to block a dihydropyridine and ~oCgTX-insensitive Ca 2§ current detectable in Xenopus oocytes following microinjection of rat brain mRNA. However, because these spider venoms are known to contain both peptide and nonpeptide toxins, the chemical nature of the active substance in these experiments is unknown. More recent analyses using purified oJ-agatoxins support the conclusion that these peptides target the Ca 2§ channel. ~o-Aga-IIA has been shown to inhibit binding of CgTX GVIA to chick brain membranes, while ~o-Aga-IA and -IB do not, suggesting that the spider toxins may be able to discriminate among channel subtypes. Additional evidence consistent with this hypothesis arises from analysis of the effects of Aga-IA, -IIA, and -IliA on
639
37. Ion Channels as Targets for Toxins Ca 2+ entry into depolarized chicken synaptosomes. In this system, the latter two polypeptides block entry, with 50% effective doses in the nanomolar range, while w-Aga-IA is ineffective at 1 IxM. Very recently, a new w-agatoxin, a)-AgaIVA, has been characterized and found to target P-type Ca 2+ channels in synaptosomes; these channels are distinct from the N-type, and are resistant to both dihydropyridine and wCgTX blockade. Based on its sequence, w-Aga-IVA may represent yet another class of spider toxins. Because many distinct Ca 2+ channel subtypes have been characterized electrophysiologically, it is natural to ask whether all are equally susceptible to the w-conotoxins. Clearly, from the experiments described above, muscle channels are unaffected. However, even amongst neuronal channels, at least three subtypes are distinguishable by a variety of criteria, including gating kinetics, unit conductance, and antagonist pharmacology: Nowycky and coworkers, who studied them in chick DRG, designated these channels as T, N, and L-types. The L-type channel is distinguished by its sensitivity to dihydropyridines (DHPs), whereas T-types are not affected by these agents, oJ-Conotoxins block T-type channels only weakly, whereas inhibition of neuronal L-type channels is essentially complete and irreversible in almost all cases; the irreversibility precludes an accurate determination of K d. Binding of w-conotoxins to a variety of tissues has been analyzed, with K a values in the subnanomolar range commonly reported. However, the validity of these values may fairly be questioned, given the irreversible nature of the toxin's effects. What is most clear is the lack of competition between the conotoxins and any class of organic channel blockers: dihydropyridine, benzothiazepine, and phenylalkylamine binding is unaffected in the presence of high concentrations of any of these drugs. Reported densities of o)-conotoxin binding sites are in the range of picomoles per milligram membrane protein in a variety of vertebrate neuronal preparations. Binding analyses also suggest that channels have either a DHP or a conotoxin site, although since L n channels can be blocked by both agents, exceptions to this rule must exist. In interpreting the literature on og-conotoxin binding, it must be remembered that distinct forms of this toxin are made by different Conus species, and these toxins do not invariably have the same specificity. co-Conotoxins have been cross-linked to putative receptors in rat and chick brain. The variation in the molecular weight of this complex upon reduction suggests that the coconotoxin receptor is associated with the % - 6 subunit of the Ca 2+ channel. However, given the ability of the a~-conotoxins to block this channel, and the fact that the al subunit is integral to ion permeation, it is reasonable to postulate interactions of the toxins with a l as well.
V. Other Toxins a n d C h a n n e l s
A. a-Conotoxins and the Nicotinic Acetylcholine Receptor-Associated Channel Although not targeted to ion channels, the ot-conotoxins will be briefly discussed in the interest of covering all of the conotoxin classes. The a-conotoxins are the smallest of the
conotoxin peptides known, comprising 13-15 amino acid residues, of which four are half-cystines involved in disulfide bonds. All of the known c~-conotoxins are cationic, with formal charges of +1.5 to +3.5 at neutral pH. Sequences of the a-conotoxins are shown in Fig. 14. Conservative changes are tolerated at most positions without significant loss of activity, so that relatively little information on structure-function relationships can be derived from synthetic analogs. It is clear that the presence of two disulfides in such a short peptide imposes relatively severe constraints upon its three-dimensional structure. The proposed structure for aconotoxin, based upon NMR measurements, includes /3turns at residues 5-8 and 9-12. a-Conotoxins are directed at the nicotinic acetylcholine receptor channel, and mechanistically act like the snake toxins exemplified by a-bungarotoxin. In fact, the most widely accepted model for a-conotoxin structure places several of its side-chain functional groups into positions very similar to those seen in the active site of bungarotoxin. The postsynaptic site of action of a-conotoxin is consistent with its observed ability to compete with other nAChR antagonists like d-tubocurarine and a-BgTX. On this basis, a-conotoxin is, strictly speaking, not truly directed against the ion channel, and will not be discussed further.
B. Chlorotoxin and Chloride Channels The toxicology of anion-selective channels is not nearly as well understood as that of the cation channels described previously, and therefore represents a fertile ground for future enquiry. The recent literature is replete with reports of chloride channels in a variety of tissues that are activated by diverse stimuli including membrane potential, cyclic AMP, volume, pH, etc. (Cuppoletti et al., 1993; Pusch et al., 1995a, b; Malinowska et al., 1995). These channels, which are all members of the family designated as C1C, are formed by proteins of M r 90-100 kDa, which most likely function as dimers (Fahlke et al., 1997a). Hydropathy-based models predict that both their N- and C-termini are located in the cytoplasm, and are consistent with the presence of twelve transmembrane domains, which are modeled as a-helices. It should be noted, however, that these models are clearly still in an evolutionary state, with domains moving into or out of the membrane as recently as 1998. Structure-function analysis of C1C channels is still at a rather early stage, although N-terminal cytoplasmic sequences essential for channel inactivation have been identified by deletion scanning mutagenesis (Grunder et al., 1992). This region is hypothesized to physically occlude the internal opening of the channel in a ball-and-chain-type inactivation mechanism, perhaps
G1 G1A GII HI FIGURE 14. Primary structures of a-conotoxins. The sequences are shown in the single letter code, with highlighted residues either invariant or allowing only functionally conservative substitution.
640
SECTION IV Ion Channels as Targets for Toxins, Drugs, and Genetic Diseases
docking to sequences within the D7-D8 cytoplasmic linker (Jordt and Jentsch, 1997). More recently, mutagenesis coupled with chemical modification has implicated residues 240-260 and Cys 546 of human C1C-1 as lying in the outer vestibule of the channel pore (Fahlke et al., 1997b; Kurz et al., 1999). Clearly, high-affinity toxins analogous to TTX or charybdotoxin would be an invaluable tool in helping to limit the target size of such analyses. However, to date only a single peptide blocker of chloride channels has been identified. This polypeptide, isolated from Leiurus venom and designated as chlorotoxin, contains 36 amino acid residues cross-linked by four disulfide bonds, and displays no sequence homology to toxins from the same venom that target Na § or K § channels (DeBin et al., 1993). It is, however, highly homologous to the scorpion insectotoxins derived from Buthus eupeus venom. Interestingly, while chlorotoxin blocks enterocyte chloride conductances with an affinity in the low micromolar range, it does so only when applied to the interior aspect of the channel; thus, its biological role is unlikely to involve chloride channel blockade. The homologous insectotoxins have been suggested to target the insect post-synaptic glutamate receptor, although supporting data for this proposal have not been presented, and the effect of chlorotoxin on this receptor is unknown. The solution structure of chlorotoxin was recently solved by Lippens et al. (1995) using multidimensional NMR. Not surprisingly, the averaged structure, which contains a small three-stranded fl-sheet packed against a single a-helix, is very similar to that of the insectotoxin. More unexpected is the observation that the overall fold of the molecule is rather similar to that of charybdotoxin, despite the lack of significant sequence homology between the two proteins. While structure-function relationships have not been addressed experimentally for either the insectotoxins or chlorotoxin, our recent ability to express active toxin in bacteria opens up this area for further exploitation.
VI. Summary Neurotoxins have proven to be invaluable tools in the purification of many of the best-characterized cation channels, and in addition have contributed significantly to our understanding of how these macromolecules function in excitable tissues. As should be clear from the foregoing outline, this collection of molecules displays a great deal of diversity in both a chemical and a functional sense. Although many of the compounds discussed in the preceding sections are cationic, additional unifying structural themes are conspicuous by their absence. Even within the subgroup of polypeptide toxins, the most well-studied groups (i.e., anemone, scorpion, snail, and snake toxins) fail to display any intergroup relatedness at the level of primary structure. Indeed, this is rather surprising in view of the ability of (e.g.) anemone and a-scorpion toxins to compete for a common receptor. While all of the polypeptides do share a cationic character, and are rich in disulfide bonds, it remains to be proven that the former property is intimately related to their activity; the latter property is almost certainly a consequence of the necessity for stabilizing small peptides in an extracel-
lular environment. Perhaps the most striking observation is the relatedness between charybdotoxin and the Na § channeldirected scorpion toxins at the level of their three-dimensional structures. This clearly raises the possibility of common binding mechanisms having evolved. Given the vast diversity of the venoms from which each of these toxins is derived, and the growing array of ion channels purified and/or channel activities characterized, it is likely that new and interesting toxins will continue to emerge in the future, and that their characterization will add significantly to our knowledge of channels and receptors in excitable tissues. Two areas of research that might prove especially fruitful involve using existing (or newly discovered) toxins to probe for specific channel isoforms in different tissues, and using behavioral assays to identify new venom components likely to have neuronal activities. This latter approach has already been applied to components of Conus geographus venom with very intriguing results, among them being the discovery of the so-called "sleeper" and "King Kong" peptides. It is likely that this surface has as yet only been scratched, and that future studies will unveil additional toxins that interact with new, and perhaps presently unknown, target sites.
Bibliography Armstrong, C. M., and Hille, B. (1998). Voltage-gated ion channels and electrical excitability. Neuron 20, 371-380. Auguste, P., Hugues, M., Mourre, C., Moinier, D., Tartar, A., and Lazdunski, M. (1992). Scyllatoxin, a blocker of Ca-activated Kchannels: structure-function relationships and brain localization of the binding sites. Biochemistry 31, 648--654. Benzinger, G. R., Kyle, J. W., Blumenthal, K. M., and Hanck, D. A. (1998). A specific interaction between the cardiac sodium channel and site-3 toxin anthopleurin B. J. Biol. Chem. 273, 80-84. Bontems, E, Roumestand, C., Boyot, P., Gilquin, B., Doljansky, Y., Menez, A., and Toma, E, (1991a). Three-dimensional structure of natural charybdotoxin in aqueous solution bylH-NMR. Charybdotoxin possesses a structural motif found in other scorpion toxins. Eur. J. Biochem. 196, 19-28. Bontems, E, Roumestand, C., Gilquin, B., Menez, A., and Toma, E (1991). Refined structure of charybdotoxin: common motifs in scorpion toxins and defensins. Science 254, 1521-1523. Bontems, E, Gilquin, B., Roumestand, C., Menez, A. and Toma, E (1992). Analysis of side-chain organization on a refined model of charybdotoxin: structural and functional implications. Biochemistry 31, 7757-7764. Cannon, S. C., and Corey, D. P. (1993). Loss of Na-channel inactivation by anemone toxin mimics the myotonic state in HPP. J. Physiol. 466, 501. Catterall, W. A. (1977). Activation of the action potential Na-ionophore by neurotoxins: an allosteric model. J. Biol. Chem. 252, 8669-8676. Catterall, W. A. (1991). Functional subunit structure of voltage-gated calcium channels. Science 253, 1499-1500. Catterall, W. A. (1995). Structure and function of voltage-gated ion channels. Annu. Rev. Biochem. 64, 493-531. Catterall, W. A., and Beress, L. (1978). Sea anemone toxin and scorpion toxin share a common receptor site associated with the action potential sodium ionophore. J. Biol. Chem. 253, 7393-7396. Chahine, M., George, A. L. Jr., Zhou, M., Ji, S., Sun, W., Barchi, R. L., and Horn, R. (1994). Na channel mutation in paramyotonia congenita uncouple inactivation from activation. Neuron 12, 281.
37. Ion Channels as Targets for Toxins Crest, M., Jacquet, G., Gola, M., Zerrouk, H., Benslimane, A., Rochat, H., Mansuelle, E, and Martin-Eauclaire, M.-E (1992). Kaliotoxin, a novel peptidyl inhibitor of neuronal BK-type Ca-activated K-channels characterized from Androctonus mauretanicus mauretanicus venom. J. Biol. Chem. 267, 1640-1647. Cruz, L. J., Gray, W. R., Olivera, B. M., Zeikus, R. D., Kerr, L., Yoshikami, D. and Moczydlowski, E. (1985). Conus geographus toxins that discriminate between neuronal and muscle sodium channels. J. Biol. Chem. 260, 9280-9288. Cruz, L. J., Kupryszewski, G., LeCheminant, G. W., Gray, W. R., Olivera, B. M., and Rivier, J. (1989)./z-Conotoxin GIIIA, a peptide ligand for muscle sodium channels: chemical synthesis, radiolabeling and receptor characterization. Biochemistry 28, 3437-3442. Cuppoletti, J., Baker, A. M., and Malinowska, D. H. (1993). Chloride channels of the gastric parietal cell that are active at low pH. Am. J. Physiol. 264, C 1609. DeBin, J. A., Maggio, J. E., and Strichartz, G. R. (1993). Purification and characterization of chlorotoxin, a chloride channel ligand from the venom of the scorpion. Am. J. Physiol. 264, C361-C369. Dias-Kadambi, B. L., Combs, K. A., Drum, C. L., Hanck, D. A., and Blumenthal, K. M. (1996a). The role of exposed tryptophan residues in the activity of the cardiotonic polypeptide anthopleurin B. J. Biol. Chem. 271, 23828-23835. Dias-Kadambi, B. L., Drum, C. L., Hanck, D. A., and Blumenthal, K. M. (1996b). Leucine-18, a hydrophobic residue essential for high affinity binding of anthopleurin-B to the voltage sensitive sodium channel. J. Biol. Chem. 271, 9422-9429. Dougherty, D. A. (1996). Cation-Tr interactions in chemistry and biology: a new view of benzene, Phe, Tyr and Trp. Science 271, 163-168. Doyle, D. A., Cabral, J. M., Pfuetzner, R. A., Kuo, A., Gulbis, J. M., Cohen, S. L., Chair, B. T., and MacKinnon, R. (1998). The structure of the potassium channel: molecular basis of K-conduction and selectivity Science 280, 69-77. Fahlke C., Knittle, T., Gurnett, C. A., Campbell, K. E, and George, A. L. (1997a). Subunit stoichiometry of human muscle chloride channels. J. Gen. Physiol. 109, 93-104. Fahlke C., Yu, H. T., Beck, C. L., Rhodes, T. H., and George A. L. Jr. (1997b). Pore-forming segments in voltage-gated chloride channels. Nature 390, 529-532. Fogh, R. H., Kem, W. R., and Norton, R. S. (1990). Solution structure of neurotoxin 1 from the sea anemone Stichodactyla helianthus. J. Biol. Chem. 265, 13 016-13 028. Fozzard, H. A. and Hanck, D. A. (1996). Structure and function of voltage-dependent sodium channels: comparison of brain II and cardiac isoforms. Physiol. Rev. 76, in press. French, R. J., Sochaczewski, E. E, Zamponi, G. W., Becker, S., Kularatna, A. S., and Horn, R. (1996). Interactions between a poreblocking peptide and the voltage sensor of the sodium channel: an electrostatic approach to channel geometry. Neuron 16, 407-413. Galvez, A., Gimenez-Gallego, G., Reuben, J. E, Roy-Contancin, L., Feigenbaum, P., Kaczorowski, G. J., and Garcia, M. L. (1990). Purification and characterization of a unique, potent, peptidyl probe for the high conductance calcium-activated potassium channel from venom of the scorpion Buthus tamulus. J. Biol. Chem. 265, 11083-11090. Goldstein, S. A. N. (1996). A structural vignette common to voltage sensors and conduction pores: canaliculi. Neuron 16, 717-722. Goldstein, S. A. N., Pheasant, D. J., and Miller, C. (1994). The charybdotoxin receptor of a shaker K + channel: peptide and channel residues mediating molecular recognition. Neuron 12, 1377-1388 Grunder, S., Thiemann, A., Pusch, M., and Jentsch, T. J. (1992). Regions involved in the opening of C1C-2 chloride channel by voltage and cell volume. Nature 360, 759-762. Hartshorne, R. E and Catterall, W. A. (1984). The sodium channel from rat brain: purification and subunit composition. J. Biol. Chem. 259, 1667-1675.
641 Hidalgo, E, and MacKinnon, R. (1995). Revealing the architecture of a K + channel pore through mutant cycles with a peptide inhibitor. Science 268, 307-310. Hoshi, T., Zagotta, W. N., and Aldrich, R. W. (1990). "Biophysical and molecular mechanisms of shaker channel inactivation. Science 250, 533-538. Jordt, S.-E., and Jentsch, T. J. (1997). Molecular dissection of gating in the C1C-2 chloride channel. EMBO J. 16, 1582-1592. Kerr, L. M., and Yoshikami, D. (1984). A venom peptide with a novel presynaptic blocking action. Nature 308, 282-284. Khera, P. K., and Blumenthal, K. M. (1994). Role of the cationic residues Argenine-14 and Lysine-48 in the function of the cardiotonic polypeptide anthopleurin B. J. Biol. Chem. 269, 921-926. Khera, P. K., Benzinger, G. R., Lipkind, G., Drum, C. L., Hanck, D. A., and Blumenthal, K. M. (1995). Multiple cationic residues of anthopleuruin B that determine high affinity and channel isoform discrimination," Biochemistry 34, 8533-8541. Kurz, L. L., Klink, H., Jakob, I., Kuchenbecker, M., Benz, S., Lehmann-Horn, E, and Rudel, R. (1999). Identification of three cysteines as targets for the Zn2+-blockade of the human skeletal muscle chloride channel. J. Biol. Chem. 273, 11687-11692. Lippens, G., Najib, J., Wodak, S. J., and Tartar, A. (1995). NMR sequential assignments and solution structure of chlorotoxin, a small scorpion toxin that blocks chloride channels. Biochemistry 34, 13-21. MacKinnon, R., and Miller, C. (1989). Mutant potassium channels with altered binding of charybdotoxin, a pore-blocking inhibitor. Science 245, 1382-1385. MacKinnon, R., and Yellen, G. (1990). Mutations affecting TEA blockade and ion permeation in voltage-activated K channels. Science 250, 276-279. MacKinnon, R., Cohen, S. L., Kuo, A., Lee, A., and Chait, B. T. (1998). Structural conservation in prokaryotic and eukaryotic potassium channels. Science 280, 106-109. Malinowska, D. H., Kupert, E. Y., Bahinski, A., Sherry, A. M., and Cuppoletti, J. (1995). Cloning, functional expression and characterization of a PKA-activated gastric C1- channel. Am. J. Physiol. 268, C 191-200. Miller, C. (1991). 1990: Annus mirabilis of potassium channels. Science 252, 1092-1096. Miller, C., Moczydlowski, E., Latorre, R., and Phillips, M. (1985). Charybdotoxin, a protein inhibitor of single Ca-activated K channels from mammalian skeletal muscle. Nature 313, 316-318. Monks, S. A., Pallaghy, E K., Scanlon, M. J., and Norton, R. S. (1995). Solution structure of the cardiostimulant polypeptide AnthopleurinB. Structure 3, 791-803. Mosher, H. S. (1986). The chemistry of tetrodotoxin. Ann. N.Y. Acad. Sci. 479, 32-43. Naini, A. A., and Miller, C. (1996). A symmetry-driven search for electrostatic interaction partners in charybdotoxin and a voltage-gated K + channel. Biochemistry 35, 6181-6187. Naranjo, D., and Miller, C. (1996). A strongly interacting pair of residues on the contact surface of charybdotoxin and a shaker K + channel. Neuron 16, 123-130. Noda, M., Suzuki, H., Numa, S., and Stuhmer, W. (1989). A single point mutation confers tetrodotoxin and saxitoxin insensitivity on the sodium channel II. FEBS Lett. 259, 213-216. Ohizumi, Y., Nakamura, H., Kobayashi, J., and Catterall, W. A. (1986). Specific inhibition of saxitoxin binding to skeletal muscle sodium channels by geographutoxin II, a polypeptide channel blocker. J. Biol. Chem. 261, 6149-6152. Pusch, M., Steinmeyer, K., Koch, M. C., and Jentsch, T. J. (1995a). Mutations in dominant human myotonia congenita drastically alter the voltage dependence of the C1C-1 chloride channel. Neuron 15, 1455-1463.
642
SECTION IV Ion Channels as Targets for Toxins, Drugs, and Genetic Diseases
Pusch, M., Ludewig, U., Rehfeldt, A., and Jentsch, T. J. (1995b). Gating of the voltage-dependent chloride channel C1C-0 by the permeant anion. Nature 373, 527-531. Ranganathan, R., Lewis, J. H., and MacKinnon, R. (1996). Spatial localization of the K + channel selectivity filter by mutant cycle-based structure analysis. Neuron 16, 131-139. Rogers, J. C., Qu, Y., Tanada, T. N., Scheuer, T., and Catterall, W. A. (1996). Molecular determinants of high affinity binding of a-scorpion toxin and sea anemone toxin in the $3-$4 extracellular loop in domain IV of the Na-channel a subunit. J. Biol. Chem. 271, 15950-15962. Satin, J., Kyle, J. W., Chen, M., Bell, P., Cribbs, L. L., Fozzard, H. A., and Rogart, R. B. (1992). A mutant of TTX-resistant cardiac sodium channels with TTX-sensitive properties. Science 256, 1202-1205. Shimizu, Y. (1986). Chemistry and biochemistry of saxitoxin analogues and tetrodotoxin. Ann. N.Y. Acad. Sci. 479, 24-31. Stuhmer, W., Conti, E, Suzuki, H., Wang, X., Noda, M., Yahagi, N., Kubo, H., and Numa, S. (1989a). Structural parts involved in activation and inactivation of the sodium channel. Nature 339, 597-603. Stuhmer, W., Ruppersberg, J. P., Schroter, K. H., Sakmann, B., Stocker, M., Giese, K. P., Perschke, A., Baumann, A., and Pongs, O. (1989b).
Molecular basis of functional diversity of voltage-gated potassium channels in mammalian brain. EMBO J. 8, 3235-3244. Tamkun, M. M., Talvenheimo, J. A., and Catterall, W. A. (1984). The sodium channel from rat brain: reconstitution of neurotoxin-activated ion flux and scorpion toxin binding from purified components. J. Biol. Chem. 259, 1676-1687. Trainer, V. L., Brown, G. B. and Catterall, W. A. (1996). Site of covalent labeling by a photoreactive batrachotoxin derivative near transmembrane segment IS6 of the sodium channel a subunit. J. Biol. Chem. 271, 11261-11267. Tudor, J. E., Pallaghy, P. K., Pennington, M. W., and Norton, R. S. (1996). Solution structure of ShK toxin, a novel potassium channel inhibitor from a sea anemone. Nature Stuct. Biol. 3, 317-320. Venema, V. J., Swiderek, K. M., Lee, T. D., Hathaway, G. M., and Adams, M. E. (1992). Antagonism of synaptosomal calcium channels by subtypes of omega agatoxins. J. Biol. Chem. 267, 2610-2615. Yang, N., Ji, S., Zhou, M., Ptficek, L. J., Barchi, R. L., Horn, R., and George, A. L. Jr. (1994). Na channel mutations in paramyotonia congenita exhibit similar biophysical phenotypes in vitro. Proc. Natl. Acad. Sci. USA 91, 12785.
I. Rivolta, H. Abriel, and Robert S. Kass
Ion Channels as Targets for Drugs
Many drugs exert their therapeutic effects by acting on the ion channels of various tissues. As illustrations, we will concentrate on only a few of the ion channels of the cardiac muscle. Electrical activity in the heart is generated by the summation of currents through multiple ion channels (Kass, 1995). The ventricular action potential is characterized by a long-lasting plateau period in which a balance is maintained between small inwardly and outwardly directed exchange and ion channel currents, and small changes in this balance can have severe functional consequences. During the past 10 years, a wealth of information has evolved characterizing the molecular structures that form the protein pathways for ion conduction in heart and other excitable cells. With the emerging molecular and structural information, true molecular pharmacological approaches to disease management are evolving with the concept of rational drug design being brought to fruition in the novel approaches applied to pharmacological intervention based on genetic defects in ion channel expression. This chapter will focus on recent progress in the development of a molecular pharmacological approach to targeting two types of cardiac ion channels: L-type calcium (Ca 2+) and voltage-gated sodium (Na +) channels.
A
Na +
[
I. Calcium Channels At least six types of voltage-gated calcium channels (N-, L-, P-, Q-, R-, and T-type) have been identified based on their pharmacological and/or biophysical properties (Tsien et al., 1991; Bimbaumer et al., 1994). In the heart, both T- and L-type channels contribute to cardiac electrophysiology, but L-type channels, the targets of organic calcium channel modulators and substrates for cAMP-dependent protein kinase A and protein kinase C, are unique in their importance to the maintenance of calcium homeostasis because they can be under pharmacological and/or neurohormonal control (Gutierrez et al., 1994; Zhao et al., 1994; Ma et al., 1992; Tsien et al., 1991; Sculptoreanu et al., 1993a, b; Catterall et al., 1991). Skeletal muscle L-type calcium channels are heteromultimeric proteins consisting of OZl-, ~2-' and a2/6-subunits (Hille, 1992; Catterall et al., 1988; Gutierrez et al., 1991) and it is likely that L-type calcium channels in other tissues are also multi-subunit proteins (Takahashi et al., 1987; Schneider and Hofmann, 1988; Takahashi and Catterall, 1987; Ahlijanian et al., 1990). Figure 1 shows a comparison of putative subunit structure of skeletal and cardiac calcium channels as well as potential phosphorylation sites.
B
Ca2+
o
~i ~i ~i ~i ~il/311
F I G U R E 1. Block diagram illustrating subunit structures of voltage-gated sodium channels (a) and L-type calcium channels (b). Sodium channels consist of a central pore-forming subunit (o~) and two auxiliary fl-subunits. L-type calcium channels also contain a pore-forming a-subunit and fl-subunits, but also a23- and y-subunits (Catterall, 1988). Cell Physiology Sourcebook: A Molecular Approach, Third Edition
643
Copyright 9 2001 by Academic Press. All rights of reproduction in any form reserved.
644
SECTION IV Ion Channels as Targets for Toxins, Drugs, and Genetic Diseases
Functional roles of the auxiliary subunits have been studied by several groups with particular focus on the role of the /32-subunit (Ruth et al., 1989; Singer et al., 1991; Gao et al., 1997). When both a- and /3-subunits of the L-type Ca 2+ channel are coexpressed, the kinetics of/Ca a r e more nearly normal than when the a-subunit is expressed alone (Lacerda et al., 1991; Lory et al., 1993; Nishimura et al., 1993; Varadi et al., 1991) and expression levels are enhanced Neely et al., 1993). One role of the/3-subunit is thus thought to involve the targeting of the a l-subunit to the surface membrane. Similarly, Cez/6-subunits increase gating charge and ionic current in recombinant channel activity (Bangalore et al., 1996). The roles of auxiliary subunits in modulating native channel function continues to be an important area of investigation that will be important in unraveling drug and neurohormonal modulation of L-type calcium channels Hosey et al., 1996). L-type calcium channels inactivate in a voltage- and calcium-dependent manner (Yue et al., 1990; Kass and Sanguinetti, 1984) and are the targets of the most extensively developed calcium channel pharmacology (Kass, 1994a). Perhaps best studied is the marked modulation of L-type calcium channels by the/3-adrenergic (/3AR) signaling cascade. Action potentials in the heart are sensitive to catecholamines. Exposure of isolated tissue to norepinephrine increases pacemaker activity (Tsien, 1977), increases the height and affects the duration of the ventricular action potential plateau (Reuter and Scholz, 1977), and increases the strength of contraction, all of which have been shown to be due, at least in part, to modulation of L-type channels by/3-AR stimulation (Cachelin et al., 1983; McDonald et al., 1994; Morad and Trautwein, 1968; Morad and Rolett, 1972; Reuter and Scholz, 1977; Reuter, 1983). Calcium entry via L-type channels contributes the major source of calcium ions to load intracellular stores for subsequent release and activation of contractile proteins. The process linking calcium entry via sarcolemmal channels and release of calcium from the sarcoplasmic reticulum (SR) is thought to rely on calcium-induced calcium release from the SR due to rapid accumulation of calcium ions in a restricted space surrounding L-type calcium channels and ryanodine receptors (RYRs), the intracellular calcium channels of the SR (Marks, 1997; Oh et al., 1997; Yamazawa et al., 1997; Cannell et al., 1994). This work does not rule out the interesting possibility that L-type calcium channels and RYRs are coupled with molecular bridges in heart as they are in skeletal muscle (Chu et al., 1991), a possibility presently under experimental investigation. Modulation of L-type Ca 2+ channel activity by calcium channel antagonists has become a key clinical therapeutic approach to the management of hypertension and certain types of cardiac rhythm disturbances (Kass, 1994a). Several more recent reports, however, have raised questions about the effectiveness and safety of some calcium channel antagonists in the treatment of hypertension (Furberg and Psaty, 1995), making it clear that an understanding of the precise molecular targets of this important drug family is crucial to improved therapeutic efficacy. The drugs that have received the most attention belong to three distinct chemical classes: (1) phenylalkylamines (ver-
apamil, D-600); (2) benzothiazepines ((+) cis diltiazem); and the 1,4-dihydropyridines (PN 200-110, nitrendipine, nifedipine, nisoldipine) (Kass, 1994b) (Fig. 2). These drugs bind to distinct but allosterically coupled sites on the channel protein (Glossmann and Striessnig, 1990). The eel-SUbunit of the L-type calcium channel contains the binding sites for the three major classes of calcium channel modulators described above (Catterall and Striessnig, 1992), and, when expressed in heterologous expression systems, is sufficient to encode channels with most of the biophysical and pharmacological properties of intact native channels (Hofmann et al., 1993; Welling et al., 1993a; Mori et al., 1991; Mikami et al., 1989) (see Fig. 1). Studies of the molecular site(s) and mechanisms of action of calcium channel blockers have focused on the biochemical and biophysical role of the ce1subunit and its relationship to modulation of native L-type channels. Site-directed mutagenesis studies have revealed specific residues that form the binding domains for all three classes of drugs. Emerging from these studies is the consensus view that domains III and IV of the al-SUbunit are crucial to modulation of L-type Ca 2+ channels, and that it is very likely that multiple residues on the a l-subunit interact in an allosteric manner to cause voltage-dependent modulation of channel gating (Ito et al., 1997; Grabner et al., 1996; Peterson et al., 1996, 1997; Schuster et al., 1996; Tang et al., 1993). The most provocative model for the actions of these drugs has been developed by Catterall and his colleagues, who have proposed a domain interface model (Fig. 3) for the mode of action of dihydropyridine derivatives (Hockerman et al., 1997). Most importantly, in this model, allosteric interactions between bound drug and two domains of the ce1subunit cause conformational changes in the channel, which modify channel gating and control the entry of divalent ions. Most DHP derivatives previously studied in detail are neutral compounds at physiological pH (Rodenkirchen et al., 1982) and it is well established that neutral DHPs such as nitrendipine and nisoldipine inhibit L-type Ca 2+ channels in a voltage-dependent manner, being approximately 100-fold
CH30
CN
CH30
CH3
OCH3
--CH2~CH2--CH2 CH
/\
CH3 CH3
H
Verapamil OCI-I3
I CH3~l Ni ' ~ cH3 II u'l
II
0 '~L o
~
NO2
..
~%""~-{o (~_~CH3 CH2CH2NH "~CH 3
Nifedipine
Diltiazem
2. The major classes of calcium channel blockers: phenylalkylamines (verapamil); dihydropyridines (nifedipine); and benzothiazepines (diltiazem) (Hille, 1992). FIGURE
38. Ion Channels as Targets for Drugs
FIGURE 3. Domain interface model for dihydropyridine (DHP) binding in the a-subunit of the L-type calcium channel. Indicated are key amino acids in domain IIIS6 and domain IVS6 that when mutated change dihydropyridineby factors greater than five-fold. Black letters inside shaded circles represent amino acids that when mutated have significant but less than five-fold reduction in DHP binding. A schematized DHP ligand is illustrated by the triangle contacting the putative key binding residues. It is proposed that channel regulation is a consequence of allosteric interactions between the drug molecule and the two helices IIIS6 and IVS6. (From Catterall et al., 1997.)
more potent at inhibiting channel activity at depolarized voltages. Subsequent repolarization to negative holding potentials readily reverses voltage-enhanced channel inhibition (Sanguinetti and Kass, 1984). Addition of a charged headgroup appears to change this relationship between channel modulation and membrane potential: channel inhibition by charged DHPs is not relieved upon repolarization (Kass, 1994a,b). The voltage dependence of neutral dihydropyridine channel modulation has been interpreted within the general framework of the modulated receptor hypothesis which has been very successful in predicting differences in the activity of charged and neutral Class Ib local anesthetics (see Section II). For dihydropyridines, binding affinity to inactivated channels is high; binding affinity to channels in the resting state is low (Sanguinetti and Kass, 1984), consistent with this hypothesis. The interrelationship between the activity of DHPs and divalent ions is well established (Glossmann et al., 1987). Highaffinity dihydropyridine labeling of L-type channels in brain, cardiac, or smooth muscle membranes depends on the presence of divalent ions (Glossmann and Striessnig, 1990), as does the
645
binding of phenylalkylamines and DHPs to the purified DHP receptor (Flockerzi et al.,). High-affinity Ca 2+binding to glutamate residues in the putative pore-lining regions (SS1-SS2) of repeats I-IV in the calcium channel alC_a-SUbunit is also a critical determinant of ion selectivity and permeation in alc_a Ca 2+ channels (Yang et al., 1993). Subsequent studies showed that high-affinity Ca2+ binding to the glutamate residues in the SS2 segments of domains I11 and IV also stabilizes the DHP receptor site in a high-affinity state (Hockerman et al., 1995). Thus, for the first time, two independent studies demonstrated that high-affinity DHP-binding depends on Ca2+ coordination by glutamate residues within the L-channel pore. Conversely, these results imply that DHP binding modulates the binding of Ca 2+ to these glutamates, which, in turn, will control Ca2+ movement through the channel. In combination with recent studies identifying the IVS6 residues critical to PAA and DHPbinding (Hockerman et al., 1995), these data raise the possibility that the glutamate residues and the IVS6-DHP-binding site may be within close proximity to each other. Therefore, a positively charged headgroup added to DHP derivatives may promote ionic or allosteric interactions with the negative charges of the pore-lining glutamate residues in addition to the binding of the DHP moiety to its high-affinity site. L-type calcium channels encoded by the smooth muscle splice variant of the a (alc_b)-subunit respond to DHP derivatives differently than channel encoded by the cardiac splice variant (alC_a) (Welling et al., 1993b). In particular, it was found that channels encoded by CelC_bcDNA were more sensitive to the DHP antagonist nisoldipine than were channels encoded by the alC_a splice variant (Fig. 4). Subsequent work has shown that this distinct pharmacological profile is due to alternative splicing of the IS6 segments of the 1C gene in cardiac and smooth muscle (Welling et al., 1997). In addition, Welling and colleagues confirmed that CelC_bis expressed in smooth muscle, but not cardiac, cells, linking the molecular pharmacology to the physiology of the tissues. These experimental results are important because they show that regions of the 1C subunit distinct from the DHP-binding domain influence drug activity, most likely through allosteric interactions. These results for DHP antagonists have been confirmed by others (Hu and Marban, 1998; Zuhlke et al., 1998); however, differences in DHP agonist modulation of smooth muscle versus cardiac muscle L-type calcium channels has not yet been systematically addressed. T-type calcium channels have voltage-dependent kinetics, ion permeability, and pharmacological properties that distinguish them from L-type calcium channels. They are resistant to block by dihydropyridines, inactivate in a voltage-dependent manner, and, most importantly, activate at voltages much more negative than L-type calcium channels (Bean, 1989). Because of the voltage dependence of T-type channels and the relatively small size ofT-channel current in the ventricle (Bean, 1985), direct activation of contractile proteins and/or calcium-induced release of calcium from the SR is not likely (Balke et al., 1993; CanneU et al., 1995). T-type channel activity has been suggested to contribute to pacemaker activity in nodal cells (Irisawa and Hagiwara, 1991), which may be particularly important in the Purkinje fibers of the ventricle. The cloning of the neuronal T-type channel has opened the possibility of determining the molecular basis for these differences (Perez-Reyes et al., 1998).
646
SECTION IV Ion Channels as Targets for Toxins, Drugs, and Genetic Diseases
(a)
Native channels
Channels ~
10 m M C a 2+
9 to o
10.0
10 m M C a 2+
; l~o.o m,
ms
L L
i
2 0 m M Ba 2+
to d
,r. -
10.0
9"
ms
30 mM
t
e ~1,
..
~.~
.
.
.
.
Ba 2+
="
~.
o
10.0
_
ms
-
.
9~
; ~r-" . . . . . . .
,
9
.
*~
-. 2.
".,L_ . . . . .
. ,
..~
---.-'. _ .i
(b) 80 > E o60 r I
0
O40 JD C fJ
L_
m20 t,"
m~a
na~e
hrt
Channel type F I G U R E 4 . L-type calcium channels expressed in Chinese hamster ovary (CHO) cells, showing the greater sensitivity of smooth muscle (sm) than heart (hrt) channel variants to dihydropyridine block. (a) Native (left) and recombinant (right) L-type calcium channel activity when calcium (upper row) and barium (lower row) carries the charge. (b) Tonic block of recombinant channel activity by the DHP nisoldipine for alC_a (hrt) and alC_b (sm) a subunits. (From Welling, et al.)
38. Ion Channels as Targets for Drugs
Ii. S o d i u m (Na § Channels Voltage-gated Na + channels are also integral membrane proteins (Catterall, 1993) that not only control the movement of Na + and underlie the spread of excitation in ventricular and atrial muscle cells and in the Purkinje fiber network throughout the heart (Kass, 1994c), but also can contribute so-called window inward current, which prolongs action potential duration (Attwell et al., 1979). In most tissues the voltage-gated Na + channel is a heterotrimeric protein consisting of c~ (33 kDa)-,/31 (36 kDa)-, and/or/32 (33-kDa)-subunits (Catterall, 1993; Catterall and Epstein, 1992; Catterall et al., 1992), but only the ce-subunit is needed for expression of recombinant channels, particularly for heart channels (Stuhmer et al., 1989a; Noda et al., 1989; Suzuki et al., 1988). In heart, Na + channel activity underlies the rapid spread of electrical impulses in specialized conducting tissue of the atria and ventricle (Purkinje fibers) as well as the working muscle of both atrial and ventricular tissue (Kass, 1994d). Consequently, cell activity can be indirectly read from the electrocardiogram: the QRS interval reflects conduction time through the ventricle and hence the number of available Na § channels. Prolongation of the QRS interval occurs when Na + channel activity is reduced by drugs or cellular conditions. Although the role of Na + channel currents in impulse propagation is well known, Na + channel activity can also contribute to the duration of the ventricular action potential. The action potential is shortened in the presence of TTX (a Na § channel-blocking toxin) or after sodium removal (Hiraoka et al., 1986; Colatsky, 1982) because a very small fraction of channels fail to enter the absorbing inactivated state of the channel and thus create what has been referred to as a window current through TTX-sensitive Na + channels. Most importantly, inherited mutations of the human Na § channel can affect this property of the channel and promote reopening of a larger fraction of channels after inactivation. This produces enhanced inward current that prolongs the action potential in forms of the long-QT syndrome linked to the gene that encodes the Na + channel c~-subunit (Clancy and Rudy, 1999). The selective pore of Na + channels is regulated by voltagesensitive channel gates. According to Hodgkin and Huxley (Hodgkin and Huxley, 1952), at resting membrane potentials, Na § channels are in a resting state where the pore is closed by activation (m) gates. Membrane depolarization induces conformational changes (gating) that open m-gates, resulting in conduction of Na + ions through the pore (permeation) and Na § current. Continued depolarization triggers closure of an inactivation (h) gate, occlusion of the channel pore, and termination of the Na § current. Membrane repolarization returns the channel to the resting state by shutting m-gates and opening the hgate. Na § channel gating is more complex than the fundamental workings detailed in this model (rev. in Patlak, 1991). However, the addition of a single-state, slow inactivation provides a satisfactory backdrop for considering binding and gating interactions relevant to local anesthetic (LA) action. Local anesthetic drugs, such as lidocaine and mexilitine, block Na § channels in a voltage-dependent manner. Investigation of the molecular basis for this action has provided valuable information about the structure and function of the Na + channel with a particular emphasis on channel
647
gating. LA modulation of Na + currents and underlying channel gating is commonly investigated using electrophysiological studies; both single-channel and macroscopic current. Our present understanding of LA action is founded on a number of key experimental observations that lead to hypotheses that account for tonic and phasic inhibition. Permanently charged, hydrophilic analogs of LAs (e.g., QX-314) are poorly membrane permeable and only block Na + currents when delivered intracellularly (Strichartz, 1977). In addition, QX-314 binding requires open channels, traverses 60% of the membrane electric field from the cytoplasmic channel mouth to reach its receptor, and does not prevent channel gating. Na + currents recover rapidly from phasic inhibition induced by hydrophobic LAs compared with permanently charged analogs, suggesting that uncharged molecules leak out from closed channels. Reducing external pH slows recovery from phasic inhibition by tertiary amine LAs, suggesting that hydrogen ions can pass the Na + channel selectivity filter and ionize the LA molecule (Hille, 1977). LAs enhance Na + channel inactivation such that voltage-dependent recovery from inactivation is shifted leftward, hyperpolarization relieves inhibition, and recovery from inactivation is slow (Schwarz et al., 1977). These results lead to the proposal of an intrapore LA receptor just inside the Na + channel selectivity filter involving hydrophobic and hydrophilic access routes (Hille, 1977). Hille (1977) proposed the modulated receptor hypothesis in which a single LA receptor lies within the pore, between the selectivity filter and channel gates. Receptor occupation leads to cessation of ion flow and promotion of inactivation. Uncharged molecules gain access through a hydrophobic, transmembrane pathway. Charged molecules approach the receptor by an aqueous, hydrophilic pathway through the open cytoplasmic mouth (channel gates open) and the pore (see Fig. 1). Also, receptor affinity is modulated by channel state, where open and inactivated states bind drug avidly and resting states do not. Alternatively, Starmer and Courtney (1986) proposed that state-dependent affinity of an intrapore receptor arises from a changeless receptor that is guarded by channel gates, a guarded receptor. Here, receptor access is regulated by channel gates, where open channels provide greater receptor access. Such state-dependent accessibility manifests as affinity modulation but with greater mechanistic parsimony.
A. Mutation-Specific Drug Actions Voltage-dependent block of Na + channel currents by antiarrhythmic drugs is a consequence of distinct interactions with different states of the voltage-gated Na + channel. According to the modulated receptor theory of drug/channel interactions, state-dependent activity of these drugs is a consequence of low-affinity interactions with channels in the rested, or closed, state and higher affinity interactions with channels in open and/or inactivated states that are evoked by depolarization (Hille, 1977; Hondeghem and Katzung, 1977). Two drugs, lidocaine and flecainide, differ in their modes of action in that lidocaine interacts preferentially with inactivated channels, and drug block is not necessarily dependent on channel openings (Bean et al., 1983; Ragsdale et al., 1996). In contrast, flecainide
648
SECTION IV Ion Channels as Targets for Toxins, Drugs, and Genetic Diseases
requires channels to open, and does not require channels to enter the inactivated state to promote block (Anno and Hondeghem, 1990; Ragsdale et al., 1996). The molecular determinants of inactivation of the voltage-gated Na + channel have been well studied and have been shown to be interrelated to the molecular determinants of local anesthetic modulation of sodium channels (Ragsdale and Avoli, 1998; Kambouris et al., 1998). Multiple groups have provided molecular and biophysical evidence that fast inactivation of this channel occurs by the binding of an intracellular inactivation gate, the loop connecting homologous domains III and IV of the Na + channel, to regions around the inner mouth of the Na + channel pore through hydrophobic interactions (Stuhmer et al., 1989b; Vassilev et al., 1988; Rohl et al., 1999; Chahine et al., 1997). The docking site of the inactivation gate has been less clearly identified, but the short intracellular loops connecting the $4 and $5 segments in each domain of the channel (McPhee et al., 1998) have been revealed as interaction sites, and two residues in the $6 segment of domain IV (F1764 and Y1774 in rat brain IIA channel) transmembrane segment IVS6 have been found to be key in modulating inactivation (McPhee et al., 1995). Importantly, overlapping residues on IVS6 (F1764 and Y 1771), when mutated, decrease the affinity for lidocaine and, to a lesser extent, flecainide (Ragsdale et al., 1996) (Fig. 5). That lidocaine and other local anesthetics stabilize the inactivation state of the channel is evidenced by the hyperpolarizing shift in the voltage dependence of inactivation and by the voltage range that potentiates its activity (Ragsdale et al., 1996). Furthermore, mutation of the hydrophobic IFM motif of the III-IV intracellular loop into QQQ removes fast inactivation of brain IIA (West et al., 1992) and cardiac Na + channels (Bennett et al., 1995), and strongly decreases lidocaine TB and UDB (Bennett et al., 1995). However, recent data suggest that transitions to (or conformations of) the active state may play a more important role in the actions of lidocaine block than previously considered (Vedantham and Cannon, 1999; Scheuer, 1999). Multiple mutations of SCN5A, the gene that encodes the human heart Na + channel c~-subunit, have been discovered and linked to two inherited cardiac arrhythmias: the long QT syndrome and the Brugada's syndrome. The fact that these mutations cause functional changes in expressed channel activity has created the unique opportunity to develop specific molecular therapeutic approaches to disease management based on specific functional changes in the channel proteins encoded by mutant genes (Keating and Sanguinetti, 1996). Pharmacological analysis of mutant channels expressed heterologously has provided evidence that Na + channel blockers that preferentially interact with the inactivated state of the channel (Hille, 1977; Hondeghem and Katzung, 1977) block in a targeted manner the maintained current conducted by some mutant LQT (AKPQ mutant) channels (An et al., 1996; Compton et al., 1996; Dumaine et al., 1996; Priori et al., 1996; Wang et al., 1997; Dumaine and Kirsch, 1998), shorten action potential duration in cellular studies (Shimizu and Antzelevitch, 1997; Schwartz et al., 1995), and in preliminary studies correct QT prolongation in patients
SSl-SS2 !
\
iVS6
H5C2C3H7 \N"
HC-C2~
o-c H0
F I G U R E 5. Schematic diagram of the putative interaction between lidocaine and key residues on domain IV transmembrane segment $6 of the rat brain Na channel. (Reprinted with permission from Science. Copyright 1994 American Association for the Advancement of Science.)
(Schwartz et al., 1995; Rosero et al., 1997). Figure 6B illustrates the maintained current carried by LQT-3 mutant (AKPQ) channels and preferential block of this current by lidocaine. More recently, flecainide, which preferentially blocks open but not inactivated channels (Ragsdale et al., 1996), has also been shown to be effective in inhibiting persistent Na + channel current in AKPQ mutant channels (Nagatomo et al., 1997) and in shortening and normalizing QT intervals in patients carrying this gene mutation (Windle et al., 1999). However other LQT mutations, such as the D1790G mutation of the Na + channel ce-subunit (An et al., 1998; Benhorin et al., 1998) do not promote sustained inward currents and are not likely to be targeted by lidocaine. Interestingly, clinical data suggest that the open channel blocker flecainide might be effective in controlling arrhythmias caused by the D 1790G mutation (Benhorin et al., 1999). Preliminary experimental data suggest that flecainide might be a reasonable candidate drug to selectively target D 1790G mutant channels in carriers of this LQT-3 mutation due to unique effects of this mutation on flecainide interactions with the Na + channel (Abriel et al., 2000). In order to extrapolate to possible modes of action in patients, estimates of plasma drug concentrations are useful.
38. Ion Channels as Targets for Drugs A
hill
,l,.,,~,a_ t.
. . . .
649
....
KPQ , i . . J l ~ i ...... , Id
,l-t,~,J
!F---o- ....
a,..Aa,
._.
?e--.r.-
.....
IillV
-so ,5 ,x, ~,_4~ 6
6
6
4i
6
W' -5
B 0.8
"13
,~ 0.6 O mO ..(3
t = 5.20054 p = 0.0004 At the 0.02 level, the two means are significantly different
cQ~ z-
=
0.4
O O tO .i--, O
0.2
It
0.0
1
Peak
-
I
Maintained
F I G U R E 6. (a) The AKPQ deletion mutation of the cytoplasmic linker between domains III and IV of the human heart Na § channel oLsubunit promotes maintained sodium channel current that is not present in wild-type a (hill) channels. (b) Summary of experiments in which the fraction of peak inward Na + channel current (left bar) was compared with the fraction of late or maintained Na § channel current blocked (fight bar) by the same concentration of lidocaine (100 lxM). The late current is more than twice as sensitive to lidocaine as peak current. (From An et al., 1996.)
T h e s e studies represent the first a p p r o a c h to selectively target ion channels that have b e e n e n c o d e d by diseasecausing m u t a n t genes. D e s p i t e the fact that both B r u g a d a ' s s y n d r o m e and l o n g - Q T s y n d r o m e are relatively rare diseases, the targeted therapy a p p r o a c h can serve as a m o d e l for m o r e general cases that are likely to e m e r g e , such as the broad category of acquired forms of cardiac arrhythmias.
Bibliography Abriel, H., Wehrens, X. H. T., Benhorin, J., and Kass, R. S. (2000). Molecular pharmacology of the D1790G LQT-3 mutation: biophysical determinants of drug- and mutation-specific therapeutics. Biophys. J., in press.
Ahlijanian, M. K., Westenbroek, R. E., and Catterall, W. A. (1990). Subunit structure and localization of dihydropyridine-sensitive calcium channels in mammalian brain, spinal cord, and retina. Neuron 4, 819-832. An, R. H., Bangalore, R., Rosero, S. Z., and Kass, R. S. (1996). Lidocaine block of LQT-3 mutant human Na + channels. Circ. Res. 79, 103-108. An, R. H., Wang, X. L., Kerem, B., Benhorin, J., Medina, A., Goldmit, M., and Kass, R. S. (1998). Novel LQT-3 mutation affects Na § channel activity through interactions between alpha- and beta l-subunits. Circ. Res. 83, 141-146. Anno, T., and Hondeghem, L. M. (1990). Interactions of flecainide with guinea pig cardiac sodium channels. Importance of activation unblocking to the voltage dependence of recovery. Circ. Res. 66, 789-803. Attwell, D., Cohen, I., Eisner, D., Ohba, M., and Ojeda, C. (1979). The steady state TTX-sensitive ("window") sodium current in cardiac Purkinje fibres. Pfliigers Arch. 379, 137-142. Balke, C. W., Rose, W. C., O'Rourke, B., Mejia-Alvarez, R., Backx, P., and Marban, E. (1993). Biophysics and physiology of cardiac calcium channels. Circ. 87 Suppl. 7, VII49-VII53. Bangalore, R., Mehrke, G., Gingrich, K., Hofmann, E, and Kass, R. S. (1996). Influence of the L-type Ca-channel %/8 subunit on ionic and gating current in transiently-transfected HEK 293 cells. Am. J. Physiol. 270, H1521-H1528. Bean, B. P. (1985). Two kinds of calcium channels in canine atrial cells. Differences in kinetics, selectivity, and pharmacology. J. Gen. Physiol. 86, 1-30. Bean, B. P. (1989). Classes of calcium channels in vertebrate cells. Ann. Rev. Physiol. 51, 367-384. Bean, B. P., Cohen, C. J., and Tsien, R. W. (1983). Lidocaine block of cardiac sodium channels. J. Gen. Physiol. 81, 613-642. Benhorin, J., Goldmit, M., MacCluer, J., Blangero, J., Goffen, R., Leibovitch, A., Rahat, A., Wang, Q., Medina, A., Towbin, J., and Kerem, B. (1998). Identification of a new SCN5A mutation associated with the long QT syndrome. Human Mutations 12, 72. Benhorin, J., Medina, A., Taub, Y., Windman, M., Goldmit, M., Kerem, B., and Kass, R. S. (1999). Effects of flecainide in patients with a new SCN5A mutation: mutation-specific therapy for Long QT Syndrome? Circulation, in press. Bennett, P. B., Valenzuela, C., Chen, L. Q., and Kallen, R. G. (1995). On the molecular nature of the lidocaine receptor of cardiac Na § channels. Modification of block by alterations in the alpha-subunit III-IV interdomain. Circ. Res. 77, 584-592. Birnbaumer, L., Campbell, K. P., Catterall, W. A., Harpold, M. M., Hofmann, E, Home, W. A., Mori, Y., Schwartz, A., Snutch, T. P., and Tanabe, T. (1994). The naming of voltage-gated calcium channels. Neuron 13, 505-506. Cachelin, A. B., de Peyer, J. E., Kokubun, S., and Reuter, H. (1983). Ca channel modulation by 8-bromocyclic AMP in cultured heart cells. Nature 304, 462-464. Cannell, M. B., Cheng, H., and Lederer, W. J. (1994). Spatial nonuniformities in [Ca2+]i during excitation-contraction coupling in cardiac myocytes. Biophys. J. 67, 1942-1956. Cannell, M. B., Cheng, H., and Lederer, W. J. (1995). The control of calcium release in heart muscle. Science 268, 1045-1049. Catterall, W., and Epstein, P. N. (1992). Ion channels. [Review]. Diabetologia 35 Suppl. 2, $23-$33. Catterall, W. A. (1988). Structure and function of voltage-sensitive ion channels. Science 242, 50--61. Catterall, W. A. (1993). Structure and function of voltage-gated ion channels. [Review]. Trds Neurosci. 16, 500-506. Catterall, W. A., Scheuer, T., Thomsen, W., and Rossie, S. (1991). Structure and modulation of voltage-gated ion channels. [Review]. Ann. N. Y. Acad. Sci. 625, 174-180. Catterall, W. A., Seagar, M. J., and Takahashi, M. (1988). Molecular properties of dihydropyridine-sensitive calcium channels in skeletal muscle. J. Biol. Chem. 263, 3535-3538.
650
SECTION IV Ion Channels as Targets for Toxins, Drugs, and Genetic Diseases
Catterall, W. A., and Striessnig, J. (1992). Receptor sites for Ca channel antagonists. TIPS 13, 256-262. Catterall, W. A., Trainer, V., and Baden, D. G. (1992). Molecular properties of the sodium channel: a receptor for multiple neurotoxins. [Review]. Bull. Soc. Pathol. Exot. 85, 481-485. Chahine, M., Deschenes, I., Trottier, E., Chen, L. Q., and Kallen, R. G. (1997). Restoration of fast inactivation in an inactivation-defective human heart sodium channel by the cysteine modifying reagent benzyl-MTS: analysis of IFM-ICM mutation. Biochem. Biophys. Res. Commun. 233, 606-610. Chu, A., Fill, M., Entman, M. L., and Stefani, E. (1991). Different Ca 2+ sensitivity of the ryanodine-sensitive Ca2+ release channels of cardiac and skeletal muscle sarcoplasmic reticulum. Biophys. J. 59, 102a. Clancy, C. E., and Rudy, Y. (1999). Linking a genetic defect to its cellular phenotype in a cardiac arrhythmia. Nature 400, 566-569. Colatsky, T. J. (1982). Mechanisms of action of lidocaine and quinidine on action potential duration in rabbit Purkinje fibers. An effect on steady state sodium currents? Circ. Res. 50, 17-27. Compton, S. J., Lux, R. L., Ramsey, M. R., Strelich, K. R., Sanguinetti, M. C., Green, L. S., Keating, M. T., and Mason, J. W. (1996). Genetically defined therapy of inherited long-QT syndrome--correction of abnormal repolarization by potassium. Circulation 94, 1018-1022. Dumaine, R., and Kirsch, G. E. (1998). Mechanism of lidocaine block of late current in long Q-T mutant Na + channels. Am. J. Physiol. 274, H477-H487. Dumaine, R., Wang, Q., Keating, M. T., Hartmann, H. A., Schwartz, E J., Brown, A. M., and Kirsch, G. E. (1996). Multiple mechanisms of Na + channel linked long-QT syndrome. Circ. Res. 78, 91 6-924. Flockerzi, V., Oeken, H. J., Hofman, E, Pelzer, D., Cavalie, A., and Trautwein, W. (1986). Purified dihydropyridine-binding site from skeletal muscle T-tubules is a functional calcium channel. Nature 323, 66--68. Furberg, C. D., and Psaty, B. M. (1995). Should dihydropyridines be used as first-line drugs in the treatments of hypertensionmthe con side. Arch. Intern. Med. 155, 2157-2161. Gao, T., Puff, T. S., Gerhardstein, B. L., Chien, A. J., Green, R. D., and Hosey, M. M. (1997). Identification and subcellular localization of the subunits of L-type calcium channels and adenylyl cyclase in cardiac myocytes. J. Biolog. Chem. 272, 19 401-19 407. Glossmann, H., and Striessnig, J. (1990). Molecular properties of calcium channels. Rev. Physiol. Biochem. Pharmacol. 114, 1-105 Glossmann, H., Striessnig, J., Ferry, D. R., Goll, A., Moosburger, K., and Schirmer, M. (1987). Interaction between calcium channel ligands and calcium channels. Circ. Res. 61, 130-36 Grabner, M., Wang, Z., Hering, S., Striessnig, J., and Glossmann, H. (1996). Transfer of 1,4-dihydropyridine sensitivity from L-type to class A (BI) calcium channels. Neuron 16, 207-218. Gutierrez, L. M., Brawley, R. M., and Hosey, M. M. (1991). Dihydropyridine-sensitive calcium channels from skeletal muscle. I. Roles of subunits in channel activity. J. Biolog. Chem. 266, 16 387-16 394. Gutierrez, L. M., Zhao, X. L., and Hosey, M. M. (1994). Protein kinase C-mediated regulation of L-type Ca channels from skeletal muscle requires phosphorylation of the alpha 1 subunit. Biochem. Biophys. Res. Communic. 202, 857-865. Hille, B. (1977). Local anesthetics: hydrophilic and hydrophobic pathways for the drug-receptor reaction. J. Gen. Physiol. 69, 497-515. Hille, B. (1992). "Ionic Channels of Excitable Membranes," 2 ed. Sinauer, Sunderland, MA. Hiraoka, M., Sawada, K., and Kawano, S. (1986). Effect of quinidine on plateau currents of guinea-pig ventricular myocytes. J. Molec. Cell. Cardio. 18, 1097-1106. Hockerman, G. H., Peterson, B. Z., Johnson, B. D., and Catterall, W. A. (1997). Molecular determinants of drug binding and action on Ltype calcium channels. [Review]. Annu. Rev. Pharmacol. Toxicol. 37, 361-396.
Hodgkin, A. L., and Huxley, A. E (1952). A quantitative description of membrane current and its application to conduction and excitation in nerve. J. Physiol.(London) 117, 500-544. Hofmann, E, Biel, M., Bosse, E., Flockerzi, V., Ruth, E, and Welling, A. (1993). Functional expression of cardiac and smooth muscle calcium channels. In "Ion Channels in the Cardiovascular System: Function and Dysfunction." (A. M. Brown, W. A. Catterrall, G. J. Kaczorowski, E S. Spooner, and H. C. Strauss, Eds.) AAAS Press, Washington. Hondeghem, L. M., and Katzung, B. G. (1977). Time- and voltagedependent interactions of antiarrhythmic drugs with cardiac sodium channels. [Review]. Biochim. Biophys. Acta 472, 373-398. Hosey, M. M., Chien, A. J., and Purl, T. S. (1996). Structure and regulation of L-type calcium channels - - a current assessment of the properties and roles of channel subunits. Trds. Cardiovasc. Med. 6, 265-273. Hu, H., and Marban, E. (1998). Isoform-specific inhibition of L-type calcium channels by dihydropyridines is independent of isoformspecific gating properties. Mol. Pharmacol. 53, 902-907. Irisawa, H., and Hagiwara, N. (1991). Ionic current in sinoatrial node cells. J. Cardiovasc. Electrophys. 2, 531-540. Ito, H., Klugbauer, N., and Hofmann, E (1997). Transfer of the highaffinity dihydropyridine sensitivity from L-type to non-L-type calcium channels. Mol. Pharmacol. 52, 735-740. Kambouris, N. G., Hastings, L. A., Stepanovic, S., Marban, E., Tomaselli, G. E, and Balser, J. R. (1998). Mechanistic link between lidocaine block and inactivation probed by outer pore mutations in the rat micro 1 skeletal muscle sodium channel. J. Physiol. (London) 512 (Pt 3), 693-705. Kass, R. S. (1994a). Dihydropyridine modulation of cardiovascular Ltype calcium channels: molecular and cellular pharmacology. In "Ion Channels in the Cardiovascular System: Function and Dysfunction." (E M. Spooner, A. M. Brown, W. A. Catterall, G. J. Kaczorowski, and H. C. Strauss, Eds.), pp. 425-440. Futura Publishing Co., Armonk, NY. Kass, R. S. (1994b). Molecular pharmacology of cardiac L-type calcium channels. In "Handbook of Membrane Channels: Molecular and Cellular Physiology" C. Peracchia, pp. 187-198. Academic Press, Orlando, FL. Kass, R. S. (1994c). Ionic basis of electrical activity in the heart. In "Physiology and Pathophysiology of the Heart," (N. Sperelakis, Ed.) Kluwer Academic, Norwell, MA. Kass, R. S. (1994d). Ionic basis of electrical activity in the heart. In "Physiology and Pathophysiology of the Heart," (N. Sperelakis, Ed.) Kluwer Academic, Norwell, MA. Kass, R. S. and Sanguinetti, M. C. (1984). Calcium channel inactivation in the cardiac Purkinje fiber. Evidence for voltage- and calciummediated mechanisms. J. Gen. Physiol. 84, 705-726. Kass, R. S. (1995). Ionic basis of electrical activity in the heart. In "Physiology and Pathophysiology of the Heart," (N. Sperelakis, Ed.) pp. 77-90. Kluwer Academic, Norwell, MA. Keating, M. T., and Sanguinetti, M. C. (1996). Pathophysiology of ion channel mutations. Curr. Opin. Gen. Devel. 6, 326-333. Lacerda, A. E., Kim, H. S., Ruth, E, Perez-Reyes, E., Flockerzi, V., Hofmann, F., Birnbaumer, L., and Brown, A. M. (1991). Normalization of current kinetics by interaction between the c~1 and /3 subunits of the skeletal muscle dihydropyridine-sensitive Ca 2+ channel. Nature 352, 527. Lory, E, Varadi, G., Slish, D. E, Varadi, M., and Schwartz, A. (1993). Characterization of beta subunit modulation of a rabbit cardiac Ltype Ca 2+ channel alpha l subunit as expressed in mouse L cells. FEBS Lett. 315(2), 167-172. Ma, J., Gutierrez, L. M., Hosey, M. M., and Rios, E. (1992). Dihydropyridine-sensitive skeletal muscle Ca channels in polarized planar bilayers. 3. Effects of phosphorylation by protein kinase C. Biophys. J. 63, 639-647. Marks, A. R. (1997). Intracellular calcium-release channels: regulators of cell life and death. [Rev.] Am. J. Physiol. 272, H597-H605.
38. Ion Channels as Targets for Drugs
McDonald, T. E, Pelzer, S., Trautwein, W., and Pelzer, D. J. (1994). Regulation and modulation of calcium channels in cardiac, skeletal, and smooth muscle cells. [Rev.] Physiol. Rev. 74, 365-507. McPhee, J. C., Ragsdale, D. S., Scheuer, T., and Catterall, W. A. (1995). A critical role for transmembrane segment IVS6 of the sodium channel alpha subunit in fast inactivation. J. Biolog. Chem. 270, 12 025-12 034. McPhee, J. C., Ragsdale, D. S., Scheuer, T., and Catterall, W. A. (1998). A critical role for the $4-$5 intracellular loop in domain IV of the sodium channel alpha-subunit in fast inactivation. J. Biolog. Chem. 273, 1121-1129. Mikami, A., Imoto, K., Tanabe, T., Niidome, T., Mori, Y., Takeshima, H., Narumiya, S., and Numa, S. (1989). Primary structure and functional expression of the cardiac dihydropyridine-sensitive calcium channel. Nature 340, 230-233. Morad, M. and Rolett, E. L. (1972). Relaxing effects of catecholamines on mammalian heart. J. Physiol. 224, 537-558. Morad, M., and Trautwein, W. (1968). The effect of the duration of the action potential on contractions in mammalian heart tissue. Pfliigers Arch. 299, 66-82. Mori, Y., Friedrich, T., Kim, M.-S., Mikami, A., Nakai, J., Ruth, P., Bosse, E., Hofmann, E, Flockerzi, V., Furuichi, T., Nikoshiba, K., Imoto, K., Tanabe, T., and Numa, S. (1991). Primary structure and functional expression from complementary DNA of a brain calcium channel. Nature 350, 398--402. Nagatomo, T., Fan, Z., Ye, B., January, C. T., and Makielski, J. C. (1997). Effects of flecainide on the long QT sodium channel syndrome. Circulation 96, 677. Neely, A., Wei, X., Olcese, R., Birnbaumer, L., and Stefani, E. (1993). Potentiation by the beta subunit of the ratio of the ionic current to the charge movement in the cardiac calcium channel. Science 262, 575-578. Nishimura, S., Takeshima, H., Hofmann, E, Flockerzi, V., and Imoto, K. (1993). Requirement of the calcium channel beta subunit for functional conformation. FEBS Lett. 324, 283-286. Noda, M., Suzuki, H., Numa, S., and Stuhmer, W. (1989). A single point mutation confers tetrodotoxin and saxitoxin insensitivity on the sodium channel II. FEBS Lett. 259, 213-216. Oh, S. T., Yedidag, E., Conklin, J. L., Martin, M., and Bielefeldt, K. (1997). Calcium release from intracellular stores and excitation-contraction coupling in intestinal smooth muscle. J. Surg. Res. 71, 79-86. Patlak, J. (1991). Molecular kinetics of voltage-dependent Na + channels. [Rev.] Physiol. Rev. 71, 1047-1080. Perez-Reyes, E., Cfibbs, L. L., Daud, A., Lacerda, A. E., Barclay, J., Williamson, M. P., Fox, M., Rees, M., and Lee, J. H. (1998). Molecular characterization of a neuronal low-voltage-activated T-type calcium channel. Nature 391, 896-900. Peterson, B. Z., Johnson, B. D., Hockerman, G. H., Acheson, M., Scheuer, T., and Catterall, W. A. (1997). Analysis of the dihydropyridine receptor site of L-type calcium channels by alanine-scanning mutagenesis. J. Biolog. Chem. 272, 18 752-18 758. Peterson, B. Z., Tanada, T. N., and Catterall, W. A. (1996). Molecular determinants of high affinity dihydropyridine binding in L-type calcium channels. J. Biolog. Chem. 271, 5293-5296. Priori, S. G., Napolitano, C., Cantu, F., Brown, A. M., and Schwartz, P. J. (1996). Differential response to Na + channel blockade, betaadrenergic stimulation and rapid pacing in a cellular model mimicking the SCN5A and HERG defects present in the long-QT syndrome. Circ. Res. 78, 1009-1015. Ragsdale, D. S., and Avoli, M. (1998). Sodium channels as molecular targets for antiepileptic drugs. Brain Res. Rev. 26, 16-28. Ragsdale, D. S., McPhee, J. C., Scheuer, T., and Catterall, W. A. (1996). Common molecular determinants of local anesthetic, antiarrhythmic, and anticonvulsant block of voltage-gated Na + channels. Proc. Natl. Acad. Sci. USA 93, 9270-9275. Reuter, H. (1983). Calcium channel modulation by neurotransmitters, enzymes and drugs. Nature 301, 569-574.
651
Reuter, H., and Scholz, H. (1977). The regulation of Ca conductance of cardiac muscle by adrenaline. J. Physiol. 264, 49-62. Rodenkirchen, R., Bayer, R., and Mannhold, R. (1982). Specific and non-specific Ca antagonists: a structure-activity analysis of cardiodepressive drugs. Progr. Pharmacot. 5, 9-23. Rohl, C. A., Boeckman, E A., Baker, C., Scheuer, T., Catterall, W. A., and Klevit, R. E. (1999). Solution structure of the sodium channel inactivation gate. Biochem. 38, 855-861. Rosero, S. Z., Zareba, W., Robinson, J. L., and Moss, A. (1997). Genespecific therapy for long QT syndrome: QT shortening with lidocaine and tocainide in patients with mutation of the sodium channel gene. Ann. Noninvasive Electrocardiol. 2, 274-278. Ruth, P., Rohrkasten, A., Biel, M., Bosse, E., Regulla, S., Meyer, H. E., Flockerzi, V., and Hofmann, F. (1989). Primary structure of the 13 subunit of the DHP-sensitive calcium channel from skeletal muscle. Science 245, 1115-1118. Sanguinetti, M. C., and Kass, R. S. (1984). Voltage-dependent block of calcium channel current in the calf cardiac Purkinje fiber by dihydropyridine calcium channel antagonists. Circ. Res. 55, 336-348. Scheuer, T. (1999). Commentary: a revised view of local anesthetic action: what channel state is really stabilized? J. Gen. Physiol. 113, 3-6. Schneider, T., and Hofmann, E (1988). The bovine cardiac receptor for calcium channel blockers is a 195-kDa protein. Eur. J. Biochem. 174, 369-375. Schuster, A., Lacinova, L., Klugbauer, N., Ito, H., Birnbaumer, L., and Hormann, E (1996). The IVS6 segment of the L-type calcium channel is critical for the action of dihydropyridines and phenylalkylamines. EMBO J. 15, 2365-2370. Schwartz, P. J., Priori, S. G., Locati, E. H., Napolitano, C., Cantu, F., Towbin, J. A., Keating, M. T., Hammoude, H., Brown, A. M., Chen, L. S. K., and Colatsky, T. J. (1995). Long QT syndrome patients with mutations of the SCN5A and HERG genes have differential responses to Na § channel blockade and to increases in heart rate: implications for gene-specific therapy. Circulation 92, 3381-3386. Schwarz, W., Palade, P. T., and Hille, B. (1977). Local anesthetics. Effect of pH on use-dependent block of sodium channels in frog muscle. Biophys. J. 20, 343-368. Sculptoreanu, A., Rotman, E., Takahashi, M., Scheuer, T., and Catterall, W. A. (1993a). Voltage-dependent potentiation of the activity of cardiac L-type calcium channel alpha 1 subunits due to phosphorylation by cAMP-dependent protein kinase. Proc. Natl. Acad. Sci. USA 90, 10 135-10 139. Sculptoreanu, A., Scheuer, T., and Catterall, W. A. (1993b). Voltagedependent potentiation of L-type Ca 2§ channels due to phosphorylation by cAMP-dependent protein kinase. Nature 364, 240-243. Shimizu, W., and Antzelevitch, C. (1997). Sodium channel block with mexiletine is effective in reducing dispersion of repolarization and preventing torsade des p o i n t e s in LQT2 and LQT3 models of the long-QT syndrome. Circulation 96, 2038-2047. Singer, D., Biel, M., Lotan, I., Flockerzi, V., Hofmann, E, and Dascal, N. (1991). The roles of the subunits in the function of the calcium channel. Science 253, 1553-1557. Starmer, C. E, and Courtney, K. R. (1986). Modeling ion channel blockade at guarded binding sites: application to tertiary drugs. Am. J. Physiol. 251, H848-H856. Strichartz, G. (1977). Effects of tertiary local anesthetics and their quaternary derivatives on sodium channels of nerve membranes. [Proceedings]. Biophys. J. 18, 353-354. Stuhmer, W., Conti, E, Suzuki, H., Wang, X., Noda, M., Yahagi, N., Kubo, H., and Numa, S. (1989b). Structural parts involved in activation and inactivation of the sodium channel. Nature 339, 597-603. Stuhmer, W., Conti, F., Suzuki, H., Wang, X., Noda, M., Yahagi, N., Kubo, H., and Numa, S. (1989a). Structural parts involved in
652
SECTION IV Ion Channels as Targets for Toxins, Drugs, and Genetic Diseases
activation and inactivation of the sodium channel. Nature 339, 597-603. Suzuki, H., Beckh, S., Kubo, H., Yahagi, N., Ishida, H., Kayano, T., Noda, M., and Numa, S. (1988). Functional expression of cloned cDNA encoding sodium channel III. FEBS Lett. 228, 195-200. Takahashi, M., and Catterall, W. A. (1987). Dihydropyridine-sensitive calcium channels in cardiac and skeletal muscle membranes: studies with antibodies against the alpha subunits. Biochemistry 26, 5518-5526. Takahashi, M., Seagar, M. J., Jones, J. F., Reber, B. F. X., and Catterall, W. (1987). Subunit structure of dihydropyridine-sensitive calcium channels from skeletal muscle. Proc. Natl. Acad. Sci. USA 84, 5478-5482. Tang, S., Yatani, A., Bahinski, A., Mori, Y., and Schwartz, A. (1993). Molecular localization of regions in the L-type calcium channel critical for dihydropyridine action. Neuron 11, 1013-1021. Triggle, D. J. (1994). Molecular pharmacology of voltage-gated calcium channels. [Rev.]. Ann. NY Acad. Sci. 747, 267-281. Tsien, R. W. (1977). Effects of epinephrine on the pacemaker potassium current of cardiac Purkinje fibers. J. Gen. Physiol. 64, 293-319. Tsien, R. W., Ellinor, P. T., and Home, W. A. (1991). Molecular diversity of voltage-dependent Ca 2§ channels. [Rev.]. TIPS 12, 349-354. Varadi, G., Lory, P., Schultz, D., Varadi, M., and Schwartz, A. (1991). Acceleration of activation and inactivation by the/3 subunit of the skeletal muscle calcium channel. Nature 352, 159-162. Vassilev, P. M., Scheuer, T., and Catterall, W. A. (1988). Identification of an intracellular peptide segment involved in sodium channel inactivation. Science 241, 1658-1661. Vedantham, V., and Cannon, S. C. (1999). The position of the fastinactivation gate during lidocaine block of voltage-gated Na § channels. J. Gen. Physiol 113, 7-16. Wang, D. W., Yazawa, K., Makita, N., George, A. L., and Bennett, P. B. (1997). Pharmacological targeting of long QT mutant sodium channels. J. Clin. Invest. 99, 171 4-1720. Welling, A., Kwan, Y. W., Bosse, E., Flockerzi, V., Hofmann, F., and Kass, R. S. (1993a). Subunit-dependent modulation of recombinant
L-type calcium channels: molecular basis for dihydropyridine tissue selectivity. Circ. Res. 73, 974-980. Welling, A., Kwan, Y. W., Bosse, E., Flockerzi, V., Hofmann, F., and Kass, R. S. (1993b). Subunit-dependent modulation of recombinant L-type calcium channels: molecular basis for dihydropyridine tissue selectivity. Circ. Res. 73, 974-980. Welling, A., Ludwig, A., Zimmer, S., Klugbauer, N., Flockerzi, V., and Hofmann, F. (1997). Alternatively spliced IS6 segments of the alpha 1C gene determine the tissue-specific dihydropyridine sensitivity of cardiac and vascular smooth muscle L-type Ca 2§ channels. Circ. Res. 81, 526-532. West, J. W., Patton, D. E., Scheuer, T., Wang, Y., Goldin, A. L., and Catterall, W. A. (1992). A cluster of hydrophobic amino acid residues required for fast Na(+)-channel inactivation. Proc. Natl. Acad. Sci. USA 89, 10 910-10 914. Windle, J. R., Geletka, R. C., Moss, A. J., and Atkins, D. L. (1999). Normalization of ventricular repolarization with flecainide in patients with the LQT3 form (SCN5A) of Long QT Syndrome. Circulation, in press. Yamazawa, T., Takeshima, H., Shimuta, M., and Iino, M. (1997). A region of the ryanodine receptor critical for excitation-contraction coupling in skeletal muscle. J. Biolog. Chem. 272, 8161-8164. Yang, J., Ellinor, P. T., Sather, W. A., Zhang, J. E, and Tsien, R. W. (1993). Molecular determinates of Ca2§ selectivity and ion permeation in L-type Ca2§ channels. Nature 366, 158-161. Yue, D. T., Backx, P. H., and Imredy, J. P. (1990). Calcium-sensitive inactivation in the gating of single calcium channels. Science 21, 1735-1738. Zhao, X. L., Gutierrez, L. M., Chang, C. F., and Hosey, M. M. (1994). The alpha 1-subunit of skeletal muscle L-type Ca channels is the key target for regulation by A-kinase and protein phosphatase-lC. Biochem. Biophys. Res. Commun. 198, 166-173. Zuhlke, R. D., Bouron, A., Soldatov, N. M., and Reuter, H. (1998). Ca 2§ channel sensitivity towards the blocker isradipine is affected by alternative splicing of the human alpha l C subunit gene. FEBS Lett. 427, 220-224.
Shirley H. Bryant and JamesMaylie
Ion Channels as Targets for Disease
I. Introduction Ion channels are involved in many critical cellular processes as varied as the transmission of impulses in excitable tissues including skeletal muscle to transport across epithelial tissue. Altered function of membrane ion channels then would be expected to have serious consequences for the survival of the organism. Ion channels can be affected in a variety of ways by disease, both directly and indirectly, thereby serving as targets for diseases. Direct action on the channel protein structure occurs as a result of genetic mutations within the gene coding a channel subunit. Many of the genetic mutations that markedly alter the channel function result from a single base pair change affecting only a single amino acid. We recognize these as the hereditary diseases whose targets are ion channels. Indirect action includes any of the following: (1) abnormalities in regulatory mechanisms such as phosphorylation required for normal channel function, (2) presence of any number of kinds of toxic materials that block channels or prevent their synthesis, and (3) development of autoimmune disease directed at a channel protein or its regulators, and (4) altered gene expression of channel subunits. Naturally occurring channel diseases have been reported within the general families of CI-, Na +, Ca 2+, K +, and transmitter-activated channels. With the recent introduction of rapid cloning methods and improvements in functional analysis of channels, especially the patch-clamp methodology, the rate of identification of dysfunctional channels causing disease has mushroomed. The diseases themselves, now being referred to as channelopathies, are becoming too numerous to discuss with any depth. It is our task in this chapter to make the reader aware of pertinent examples of disease for each of these general families of channels. The chapter is organized by channel type (i.e., CI-, Na § Ca 2§ K § and neurotransmitter gated) rather than by disease. Some disease entities (e.g., trinucleotide repeats in myotonia) may be caused by mutations that affect transcription or translation. Such diseases are discussed
Cell Physiology Sourcebook: A Molecular Approach, Third Edition
6 5 3
under each of the channel type headings. For a more detailed discussion of the molecular basis of ion channel diseases, consult Ion Channels and Disease (Ashcroft, 1999).
II. Ion C h a n n e l D i s e a s e s A. Mutations of Ion Channel Genes Genetic mutations can affect channel function in several ways: (1) a mutation in a gene coding a channel subunit giving rise to altered channel function, (2) altered gene expression of channel subunits affecting the number of channels incorporated into the membrane, and (3) mutation within a regulatory subunit affecting regulation of channel function. Identification of genes involved in ion channel diseases involves two principal search strategies: (1) functional or expression cloning and (2) positional cloning. In expression cloning the gene is isolated on the basis of information about the expressed protein product, for example, its amino acid sequence or antibody reactivity, or its function (e.g., receptor/ligand reactivity or ion channel characteristics), cDNA libraries are screened using different types of probes including antibodies and oligonucleotides. The polymerase chain reaction (PCR) technique can be employed in several ways including the amplification of cDNA using oligonucleotides derived from the protein sequence. Functional cloning was used to characterize the genes for the CLC-1 C1-channel and the SkM-1 Na + channel discussed later. In contrast, positional cloning is the process of isolating the gene starting from information about its genetic or physical location in the genome. Often little is known about the function of the product. The work involved is enormous since it relies on the method of chromosome walking and identification of expressed sequences. In spite of the difficulties, 19 disease genes were identified in this way between 1989 and 1993 (Ballabio, 1993), including myotonic dystrophy and cystic fibrosis, which are discussed in this chapter.
Copyright 9 2001 by AcademicPress.All rights of reproduction in any form reserved.
654
SECTION IV Ion Channels as Targets for Toxins, Drugs, and Genetic Diseases
B. Autoimmune Disease and Pleiotropic Effects Ion channel diseases can be due to the production of antibodies directed at the ion channel protein. Three examples in which autoimmune disease affects channels are discussed. Two of these are myasthenia gravis and the Lambert-Eaton syndrome. These conditions lead to bouts of skeletal muscle weakness because neuromuscular transmission is compromised. The third example is a voltage-gated K § channel, which is the target of an autoimmune disease called neuromyotonia. Mutations on one gene may influence the expression or functions of channels coded by different genes; these are referred to as pleiotropic effects. In myotonic dystrophy or Steinert's disease (see Section II.D), the aberrant expression of an apamin-sensitive small-conductance calciumactivated K + (SK) channel leads to skeletal muscle hyperexcitability and myotonia. The mechanisms for pleiotropic effects are not well understood.
C. Diseases, Linkage, Candidate Genes, and Candidate Channels Common methods are evolving for the study of familial diseases. Two basic strategies are used: positional cloning and the candidate gene. The first approach, positional cloning, is used to identify the genetic locus even when there is no knowledge of the biophysical or biochemical abnormalities underlying the disease. This is done by examining affected individuals for genetic linkage to polymorphic markers located throughout the genome. In this way a region of the gene is found that is linked to the disease. Linkage is established when there is a low rate of recombination between the marker locus and the disease locus. The gene can be used to express the protein it encodes and then determine its function and how the disease mutations disturb this function. The cystic fibrosis transmembrane-regulator channel was discovered in this way. The second approach, or candidate gene strategy, requires knowledge of functional abnormalities brought about by the disease. Genes responsible for maintaining the normal function of the process affected by the disease are then considered to be candidate genes for the disease. Genetic linkage to the disease is then sought, and if one is found the candidate gene is screened for mutations. These mutations can be incorporated into the wild-type DNA and expressed in heterologous systems. Then one can examine the biophysical and biochemical behavior and determine if the abnormality could account for the disease. The new rapid cloning techniques are largely responsible for the increased pool of known channel genes having known functions. This makes it easier to identify candidate channel genes; thus the candidate gene approach has been very successful for the study of the channelopathies. Most of the new channelopathies including the CLC-5 mutations causing kidney stones, the episodic ataxias, and the long-QT-interval diseases were discovered using the candidate gene approach. From the small number of channelopathies now reported, we can begin to make certain generalizations with respect to excitable cells. In excitable cells channels often assume both excitatory and stabilizing roles. Therefore blocking the sta-
bilizing C1- channel (e.g., the low-C1- channel myotonias) is tantamount to blocking the inactivation of the excitatory Na § channels (e.g., the Na § channel myotonias). Similarly, in the human heart the long-QT syndrome is linked either to K § channels that have a decreased Popen or to Na § channels that have an increased Po-en" In the central nervous system the phenomenon of episodic ataxia is linked both to K § channels that have a decreased P open (episodic ataxia type 1) or to Ca 2+ channels that have an increased P open (episodic ataxia type 2). Sometimes a linkage to a mutant gene is present but the expressed channels containing the mutation do not behave abnormally. This is the case for hypokalemic periodic paralysis (HoPP) discussed in Section V.B.2. The linkage to a mutation in the gene coding for a skeletal muscle Ca 2§ channel is clear, but heterologously expressed channels do not display behavior that would explain the disease. In another example, some mutant C1- channels linked to myotonia in patients when expressed in heterologous systems have normal behavior. One explanation for discrepancies of this type is that important factors necessary for phenotypic expression may be absent in the expressions system, be it a cultured mammalian cell or a Xenopus oocyte. The missing factors could include regulatory paths or chemical modifications (e.g., disulfide linkages) necessary for proper folding of the channel protein (George, 1995).
D. Phenomenon of Myotonia Myotonia is a clinical sign of certain diseases that target skeletal muscle membrane ion channels. It is a relatively rare phenomenon, usually not life threatening, yet it has held the interest of scientists and physicians for more than a century after it was first described by Dr. J. Thomsen in 1875 (see Rudel and Lehmann-Horn, 1985, for early references). Myotonia is defined as an abnormal contraction of skeletal muscles following an evoked contraction. In a typical example, a patient with myotonia is asked to grip an object or hand for a few seconds and then asked to release his grip. In spite of the fact that the patient is no longer willing a contraction, i.e., causing nerve action potentials to excite the muscle, the contraction continues for several more seconds. The myotonic muscle fibers are hyperexcitable (i.e., have low threshold potential) and tend to fire repetitive action potentials on their own. The afterdischarge of action potentials leading to the aftercontraction has been shown to result from a slowly decaying depolarization of the transverse tubules by accumulated K + in their lumen (Adrian and Bryant, 1974). Myotonic fibers fire repetitively for many seconds to steady depolarization, whereas normal fibers fire a short burst for about one-tenth of a second. This characteristic produces a diagnostically distinct electromyographic recording, which sounds like a dive-bomber when played through a sound system. Myotonic disease occurs as a result of ion channel dysfunction in the skeletal muscle excitable membranes, surface and tubular. Hereditary myotonias occur indirectly as a result of mutations in enzymes regulating the ion channels, or directly through mutations of genes coding for ion
655
39. Ion Channels as Targets for Disease
channels. Although normal muscle fiber excitability requires Na +, K +, and C1- channels, to date, only errors in Na § and C1- channels have been associated with myotonia. K § channels play an important role in action potential generation, so it is surprising that there are no natural K § channel myotonias. The most frequently encountered myotonic disease is human myotonia dystrophy (DM). The mutation responsible for this autosomal dominant disease (actually, the most common inherited neuromuscular disorder) is an expanded CTG repeat within the 31 untranslated region of a gene encoding a protein with homology to serine-threonine protein kinases, DM protein kinase (DM kinase). The CTG repeat does not produce a defective ion channel and the original proposition was that the disease results from a reduction or elevation of steady-state levels of DM kinase. However, transgenic DM kinase-null mice do not replicate the symptoms of DM, suggesting that changes in DM kinase alone do not induce DM (Jansen et al., 1996; Reddy et al., 1996). Regardless, the hallmark feature of muscle biopsies obtained from DM patients is that they contain higher levels of SK channels than normal muscle (Renaud et al., 1986). Additionally, application of the bee venom peptide toxin apamin, a blocker of SK channels, dramatically reduces muscle myotonia (Behrens et al., 1994). Thus the aberrant expression of a K + channel gives rise to hyperexcitability and myotonia in DM muscle. Therefore the underlying genetic and molecular mechanisms in DM ultimately exert effects on the expression of an ion channel in muscle. However, the disease is generalized to many organ systems and the underlying molecular mechanisms may be different in other tissue. Less frequently occurring, but better understood, are several nondystrophic muscle diseases due to mutations in the skeletal muscle membrane ion channels. These can be classified into (1) CI- channel type and (2) Na + channel type. The membrane properties characteristic of these two types are illustrated in Fig. 1.
1. C1- Channel Type
The main C1-channel-type myotonias include human
dominant myotonia congenita (DMC, Thomsen's disease), human recessive myotonia congenita (RMC; also called recessive generalized myotonia), and, among animals, the dominant myotonia of goats (which we also consider a DMC) and the recessive myotonia adr mouse. The basic physiological problem in these myotonias is a low membrane C1-conductance (gel) due to dysfunctional CLC-1 C1-channels. Normally, the mammalian CLC-1 C1channels provide the major stabilizing steady conductance (i.e., gcl) necessary for normal excitability. When gcl falls below 25% of normal, the skeletal muscle membrane becomes hyperexcitable and myotonia results. C1- channel myotonia can be induced by agents that specifically block the CLC-1 channel such as anthracene-9-carboxylic acid (9-AC), and simulated by computer models of action potentials in which gcl is lowered (Bryant and MoralesAguilera, 1971; Furman and Barchi, 1978; Adrian and Marshall, 1976). See Fig. 1 A-C and G.
2. Na § Channel Type
The main Na § channel myotonias, which are all dominant, include human and equine hyperkalemic periodic paralysis, human paramyotonia congenita, and human Na + channel myotonia. The common physiological factor in this group is a small fraction of Na § channels that inactivate more slowly and/or show abnormal reopenings compared to the wild type. The abnormally increased Na § currents result in hyperexcitability because a smaller stimulus can initiate an action potential and can then cause reexcitation with myotonic repetitive firing. Na § channel myotonia can be induced by agents that block Na § channel inactivation like the sea anemone toxin, ATX-II. Like the low-gcl myotonia, Na § channel myotonia has been simulated using computer models (Cannon et al., 1993). See Fig. 1D-F and H.
III. CI- Channels A. General Comments C1-channels and nonspecific anion channels have only recently gained attention, yet the number of distinctly different types of C1- channels may in time exceed that of other channels (Vaughan and French, 1989; Pusch and Jentsch, 1994). Unfortunately, many of the newly described channels have not yet been assigned physiological functions. The functions already identified include osmotic pressure and stretch sensitivity, C1- concentration regulation, and membrane potential stabilization. With respect to the latter function, the channels can furnish a constant stabilizing membrane conductance in skeletal muscle, a transient agonist-activated stabilizing conductance in neurons (known as inhibition), and a hormonally activated stabilizing conductance in heart fibers. Based on structural similarity and function we can identify many classes of C1- channels. Examples of genetic diseases are given for three: (1) the CFTR channel important for normal airway function whose dysfunction produces cystic fibrosis, (2) the CLC-1 channel present mainly in skeletal muscle membranes that controls the excitability of muscle fibers and whose dysfunction leads to forms of myotonia, and (3) the CLC-5 channel of the kidney whose dysfunction leads to formation of kidney stones.
B. Cystic Fibrosis Transmembrane Regulator Cystic fibrosis (CF) is an autosomal recessive disease that affects about 1 in 2000 Caucasians. The recessive gene is present in about 5% of that population, and it is the most common inherited lethal disease. It is caused by mutations (more than 100 have been identified) in the gene for the cystic fibrosis transmembrane regulator, which is a C1- channel known as CFYR. Figure 2A shows the primary structure of CFrR and the site of the major (70%) mutation, A508f, a deletion of a single phenylalanine at position 508. The primary sequence suggests 12 transmembrane helices, two hydrophilic nucleotidebinding sites, and a large cytosolic regulatory (or R) domain, which has many potential phosphorylation sites. This channel is regulated by phosphorylation and by nucleoside triphosphates. These agents act at the proposed regulatory areas.
656
SECTION IV Ion Channels as Targets for Toxins, Drugs, and Genetic Diseases
+50
k -100 ..
i
j
Ill
I , 100 m s
J=
.
.
L -100
-IOOL
t
Io = 48 n A
l
E
100 m s
F
+400
+40 0
1111111 I1!!!11 -80 100 ms
+50
Io = 48 n A
Io = 8 7 n A
100 m s
-100
-100
10 m s
../--
t~
Io = 7 5 n A 40 ms
+40 I
0
0
-100
-100
100 m s
Id = 45 laA/cm2
10 m s
F I G U R E 1. Myotonia due to C1- and Na § channels. Tracings A through F are actual microelectrode recordings of membrane potentials from single mammalian skeletal muscle fibers. Except for trace D, in which the fiber was spontaneously active, the fibers were stimulated by long square-current pulses passed through a second microelectrode inserted into the same fiber close to the voltagerecording microelectrode. A lower trace representing this depolarizing current is shown in several of the traces and the value I 0 is the intensity of this pulse in nanoamperes. (A-C): Low gcl myotonia. (A) From a myotonic mutant goat fiber. Note the repetitive action potential and an afterdischarge when the current is turned off. (B) From a normal goat fiber for control. Note that a larger current than in trace A produced no myotonia. (C) Impermeant sulfate ion replaced C1- ion bathing a normal fiber, thus blocking ga" Myotonic responses are induced identical to those of the mutant fiber in trace A. Other gcl blockers such as 9 AC at 50 IxM would give the same resuit. These goat fibers were studied at 38 ~ (Adrian and Bryant, 1974). D-F Na § channel myotonia. (D) From a human patient with the PC mutation (Lehmann-Hom et al., 1981). PC patients have normal gel yet the fibers are myotonic. Cooling this fiber from 37 to 30 ~ caused the depolarization, the small action potentials, and myotonia. (E and F) From normal rat fibers (Cannon et al; 1993). In trace E are five superimposed control measurements with 50-ms current pulses ranging from 90 to 175 nA (threshold at 115 nA); there is no myotonia. In trace F the fiber was exposed to 10 I~M ATX-II, the sea anemone toxin that blocks Na § channel inactivation, and stimulated with a long pulse. Note the induction of myotonia, the baseline depolarization, and an aflerdischarge. G-H Computer simulations of myotonia. (G) Simulates low-gel myotonia. Shown are the first five surface membrane potentials calculated for propagating the action potential model, which has a simulated T-tubular system that accumulates K § These action potentials are from a myotonic train during a long depolarizing current pulse. After the pulse was turned off there was a short afterdischarge. The model was made myotonic by lowering the gcl term in a Hodgkin-Huxley mathematical model of the action potential adapted to skeletal muscle. The control calculation (not shown) with normal gcl gave only one action potential for a large stimulating current. (From Adrian and Marshall, 1976.) (H) Computer simulation of Na+-channel myotonia (Cannon et al., 1993). The mathematical model was similar to that used for the lowgc1 myotonia except in this case the Na + channel currents were modified by adding 2% of the slowly inactivating HYPP mutant-type Na + currents to the normal Na + currents. Trace H shares many details with trace F from the toxin model of Na+-channel myotonia, for example, the tendency for the average potential to move toward a depolarizing plateau during the pulse, followed by an afterdischarge after cessation of the current pulse. Both low-ga and Na + myotonias require an intact T-tubular system for an afterdischarge. Detubulated real or simulated myotonic fibers remain myotonic but lack the afterdischarge.
N o r m a l o p e r a t i o n is a s s u m e d to be that the c h a n n e l o p e n s w h e n the sites are o c c u p i e d . L e v e l s o f c A M P d e t e r m i n e d b y p h y s i o l o g i c a l n e e d s o f the cell w o u l d t h e n r e g u l a t e n o r m a l f u n c t i o n . W h e n m u t a n t C F T R is p r e s e n t in the p u l m o n a r y e p i t h e l i a it d o e s n o t r e s p o n d to r e g u l a t i o n
b y c A M P a n d r e m a i n s c l o s e d , n o t a l l o w i n g C1- i o n s to e n t e r the cell, w h i c h p r e v e n t s n o r m a l s e c r e t i o n in a i r w a y p a s s a g e s . T h i s l e a d s to a c c u m u l a t i o n o f a t h i c k m u c u s , w h i c h b l o c k s the p a s s a g e s a n d l e a d s to d e a t h in 9 0 % o f the p a t i e n t s b e f o r e r e a c h i n g a d u l t h o o d . T h e r e are a d d i t i o n a l
657
39. Ion Channels as Targets for Disease
F I G U R E 2. C1- channels. The primary structures as deduced from hydropathy plots and other information are given for the following C1-channels: (A) The CFTR channel located in pulmonary and other epithelia. The lower arrow points to NBD1 (nucleotide-binding domain 1), the location of the principal mutation PheA508, a deletion of phenylalanine at position 508, which prevents the channel from being regulated by cAMP and leads to 70% of the cases of severe CE M6 is indicated as the site of two mutations at which arginine (334 and 347) is substituted by tryptophan and proline, respectively. M2 is the site of a similar mutation at position 117. These latter three mutations decrease the C 1- conductance of the channel and lead to mild CE (B) The CLC-1 channel of mammalian skeletal muscle produces the large stabilizing gcl necessary for normal excitability. Shown are some of the known mutations with the theoretical structural alterations of the CLC-1 channel leading to the following phenotypes exhibiting low-C1- conductance myotonia: DMC, RMC, DML (dominant myotonia levior), adr, and adrrot~(recessive myotonia in the myotonic mouse). All of the diseases indicated occur in human patients except for those conditions labeled as adr mouse or myotonic goat. Following each disease designation is the mutation given in standard single-letter amino acid codes and the residue position. For example, DMC: G230E is interpreted as dominant myotonia congenita in which at residue 230 a glutamate (E) is substituted for the wild-type glycine (G). The adr-mouse syndrome is due to the insertion of a transposon into the gene, which prevents expression of the channel. Shown here is the position (between 467 and 468) on the protein corresponding to the region on the gene affected by the transposon. This is for illustration only since the protein is never completely expressed. For more complete information on mutations of the CLCN 1 gene that encodes for the human CLC- 1 channel, see Lehmann-Horn and Rtidel (1996). (Redrawn and modified from Lehmann-Horn and Rtidel, 1996, using data from their Table 2.)
problems with secretion by the epithelia of the pancreas, gut, salivary glands, and other organs. One diagnostic feature of CF is a high NaC1 content of sweat. A mutant C F T R can account for most of the pathological signs of the disease but some additional factors may also be involved (rev. in Warner, 1992). It has been recently argued that in spite of the fact that C F T R is a channel it m a y not be the actual C1- channel that furnishes the ion transport. C F T R m a y indeed act as a regulatory protein, which controls this channel as originally thought. The discovery that an abnormal C1- channel was ultimately at fault came from resting potential measurements in epithelial cells from airways or sweat glands that always
showed the CF cells to be hyperpolarized. Perfusing the luminal side of the cells with a C1--free (sulfate) m e d i u m increased the potential of normal cells as predicted, but had no effect on the diseased cells; in other words, C1- channels were not conducting properly. The major breakthrough occurred when molecular biologists sequenced the gene responsible for C E The initial analysis of the proposed product suggested that it was a regulatory enzyme; hence, it received the name cystic fibrosis transmembrane regulatory factor or CFTR. Patch-clamp studies later showed that C F T R codes for a nonrectifying cAMP-activated channel of about 8 pS. The A508f mutation prevents the channel from responding to regulation by c A M E although the defective
658
SECTION IV Ion Channels as Targets for Toxins, Drugs, and Genetic Diseases
CFTR protein is expressed. Impressive proof that CFTR was a channel came from constructing a CFTR mutant that exhibited an altered permeability sequence to various anions, suggesting that the CFTR product was indeed a channel. As mentioned earlier, it is not completely clear that CFTR and the responsible channel are one and the same. Gene therapy by transfecting genes for the normal channel regulator into diseased pulmonary epithelia is under investigation. It is too early to evaluate these results.
C. Skeletal Muscle CLC-I Channels Skeletal muscle fiber membranes in their resting or unstimulated state have a relatively high gcr This amounts to about 80-90% of the resting conductance in mammals, the remainder being K + and less than 1% Na + and other ions. The resting gcl has an equilibrium potential (Ecl) about equal to the resting potential ( c a . - 8 2 mV), and thus it acts to stabilize the membrane toward this potential, which in effect prevents myotonia. The gc1 of skeletal muscle fibers has been studied for a long time, but identification of the responsible channel, known as CLC-1, has only occurred recently. The CLC family of CI- channels cloned by Jentsch and his colleagues (1995) presently includes about a dozen diverse members, and the number appears to be still increasing. The first member cloned was CLC-0 from electrocytes of the marine electric fish Torpedo, where it stabilizes the membrane on one side of the electric cells, thus allowing the cells to act as electric batteries (see Chapter 60). Because these electrocytes were derived embryologically from skeletal muscle, polypeptides from short sequences of the CLC-0 channel were used to probe a rat skeletal muscle cDNA library, which resulted in cloning of the CLC-1 channel (Steinmeyer et al., 1991b). Development and innervation control the CLC-1 channel, whose primary structure is shown in Fig. 2B. The channel is not seen in embryonic muscle or in cultured muscle cells, or following denervation of the adult fiber. The corresponding mRNA was seen only in trace amounts in other tissues. Monoclonal antibodies to CLC-1 react only with sarcolemmal membrane fragments, suggesting a purely sarcolemmal location of this channel (Gurnett et al., 1995). This is in contrast to earlier functional studies, which placed mammalian gcl in both the T-tubular and surface membranes (e.g., Dulhunty, 1979). Possibly, there is a T-tubular isoform that does not bind the antibody, or the channels are not accessible in the T-tubular fragment. The location problem awaits further resolution. The CLC C1- channel family has continued to expand with the reports of new members. At present this family, in addition to CLC-0 and CLC-1, consists of CLC-2, CLC-3, and CLC-4, which are ubiquitous mammalian channels of little understood function, possibly involved in cell volume regulation; CLC-K1 and CLC-K2, which are involved in reabsorption in the mammalian kidney; and CLC-S, an additional kidney channel. So far, only two members of the CLC family are associated with genetic disease. Mutations of the CLC-1 channel account for several classical hereditary myotonias in humans, mice, and goats; mutations in
the CLC-S channel produce human hereditary hypercalciuric nephrolithiasis (kidney stones). These conditions are discussed next. 1. Myotonia Congenita
Thomsen's disease or dominant myotonia congenita (DMC) is an autosomal dominant condition in which there is myotonia due to loss of the stabilizing gcr As shown in Fig. 2B (Koch et al. 1992; George et al. 1993), the defect producing the Thomsen family form of DMC is a single point mutation where the conserved glycine ((3) (residue 230) is replaced by glutamic acid (E), shown on the figure as DMC'(3230E. Some of the other known dominant mutations are also shown in Fig. 2B. The fact that a dominant mutation in a channel gene can cause a large reduction in the total number of functional normal channels requires explanation. In the heterozygous individual where only half the subunits expressed are abnormal, the gc1 is much less than 50%. One of the mechanisms proposed is that the CLC-1 channel functions as a multimer, probably a dimer based on homology with CLC-0 (Ludewig et al. 1996; Middleton et al. 1996), and that the presence of even a single mutant subunit out of two makes the channel dysfunctional. From the binomial theorem about one-fourth of the channels will have wild-type subunits and the remaining three-fourths of the channels will have at least one mutant subunit and therefore be inactive. Experimental evidence for this possibility is that coexpression of a DMC mutant CLC-1 with the wild-type CLC-1 produces a reduction in gCl that is greater than one-fourth of that predicted for homomeric wild-type channels expressed by themselves (Steinmeyer et al. 1994). It is also of interest that the gating process of CLC-O and CLC-1 is fairly unique in that these channels utilize the permeating chloride ion itself as the gating particle. 2. Recessive Myotonia Congenita
A large population of nondystrophic myotonic patients was studied, and two separate diseases corresponding to their dominant or recessive inheritance were distinguished (Becker, 1973). The smaller group (30%) were the autosomal dominant DMC patients who showed less myotonia on the average and whose physiology we discussed previously. The larger group (70%) were the autosomal recessive RMC patients who were generally more severely myotonic. RMC is also called recessive generalized myotonia. In RMC there is a reduced gcl due to a variety of mutations, some of which are shown in Fig. 2B. Additionally, some dysfunctional Na + channels may contribute to the myotonia in certain patients. This Na + channel abnormality may be a pleiotropic effect consequent to the CLC- 1 mutation. 3. Myotonic Mouse A mouse that had difficulties in righting itself when placed on its back was discovered at the Jackson laboratories (Mehrke et al. 1988). This condition was due to a recessive mutation, and at first was thought to be a neurological condi-
659
39. Ion Channels as Targets for Disease
tion affecting the righting mechanism and thus received the name of arrested development of righting response or adr mouse. Later it became clear that the major problem was the myotonia of the skeletal muscles, and thus the term myotonic mouse is also used. The repetitive firing and other defects in excitability were shown to be due to a specific loss of gcl" The heterozygous animals that had nearly normal gc] behaved normally, which indicates that the normal allele in the heterozygous mouse is capable of coding for and expressing a normal density of C1- channels. Jentsch and colleagues (Steinmeyer et al. 1991a, b) have shown that the CLC-1 channel (see Fig. 2B) is also the major skeletal muscle C1channel in the mouse. They further showed that the mutation was caused by a transposon (i.e., a nonsense code) inserted into this gene at a region that would otherwise code for an essential part of the channel protein. Actually, the defect is so severe that no mRNA for this channel was detected in the myotonic mouse. In addition to the transposon defect of the adr mouse, other missense and nonsense mutations have been reported in the mouse CLC-1 that lead to myotonic conditions. One of these is the adr mt~mouse.
4. Myotonic Goat A dominant mutation in goats in which myotonia is the only clinical sign has been known for more than a century (see Bryant, 1979). These animals (so-called stiff, falling, or nervous goats) are still being raised for their interesting behavior on sudden movement, which often causes bizarre stiffness and sometimes falling. Based on human experience it is believed that the muscle effects are not painful. Electrophysiological studies on biopsied muscle fibers from these animals advanced our knowledge of the biophysics of myotonia, and in particular led to the discovery of myotonia due to low gc] discussed earlier. Also shown was the necessity of K + accumulation in the T-tubular system for maintaining an afterdischarge (Adrian and Bryant, 1974). The myotonia can be simulated by blocking gc] with 9-AC or other C1- channel blockers (Bryant and MoralesAguilera, 1971), and by computer modeling of action potentials when the C1- conductance term, modeled as a leak conductance, is made low (Adrian and Marshall, 1976). The mutation responsible for CLC-1 dysfunction in myotonic goat is the replacement of a highly conserved alanine by proline in the carboxy-terminus region, A885P. The functional effect of this mutation, determined in a human CLC-1 construct, is to shift the curve relating open probability (Pop,n) to membrane potential to the fight by 45 mV. This is illustrated in Fig. 3. These results imply that at the physiological resting potential of-82 mV, if only mutant channels were present, the C1conductance would be roughly 25% of a membrane having wild-type channels; this is similar to the reduction of C1- conductance recorded in native myotonic goat membrane. An exact comparison with the older studies would be difficult, since these data were usually averaged from mixed genetic populations having both mutant and wild-type alleles. The pattern of shifting the conductance curves toward more depolarized voltages also occurs in four dominant human myotonias (I290M, R317C, P480L, and Q552R), and may represent a biophysical phenotype for dominant C1- channel myotonic disease.
CLC-I Chloride Channel 1o0 ""
0.5--
~r
~/'~
~
0.0-
"A885P
~
J
I
-1 O0
(myotonic goat)
I
0 Voltage (mV)
+ 1wOo
F I G U R E 3. Myotonic goat CLC-1 C1-channel function. The relative open probabilities at the different membrane voltages were determined with a two-electrode voltage-clamp from Xenopus oocyte expressing wild-type (WT) or mutant (A885P) CLC-1 channel currents. The curves were fitted to single Boltzmann functions. The principal change observed is the right shift of the center potential, V1/2, on the mutant channel by 45 mV from the wild-type control. This kind of shift is typical of shifts observed in several dominant human myotonias involving the CLC- 1 channel. At the resting potential o f - 8 0 mV the mutant channels would have a lower probability of being open, and this would account for the low gcl and myotonia in the affected goat (Redrawn from Beck et al, 1996.)
D. Kidney CLC-5 Channel: Kidney Stone Diseases The second member of the CLC family of Cl-channels to be associated with genetic disease is CLC-5. This channel is strongly outwardly rectifying, unlike CLC-1, which is an inward rectifier. Kidney stones affect about 12% of males and 5% of females in the Western world, and in 45% of these patients, the disease is inherited. Dent's disease and two other types of kidney stone disease have been linked to chromosome Xpll.22. A microdeletion in a Dent's disease patient allowed the identification of a candidate gene (CLCN5) that codes for the renal C1-channel, CLC-5. Many types of mutations of this gene were then detected; these errors include nonsense, missense, donor splice mutations, and various deletions (Lloyd et al., 1996). A speculative explanation for the pathological effects of the mutant CLC-5 is that the protein leakage through the glomerulus is taken up by endocytosis in the proximal tubules, and subsequently degraded in a lysosomal compartment. The essential endocytotic uptake is postulated to require acidification of the endosomes, and the CLC-5 channel provides the C1- transport necessary to maintain the low pH in this compartment. Because the mutant CLC-5 does not transport CI-, the endosomal pH rises, endocytosis is impaired, protein accumulates in the tubule, and kidney stone formation is promoted.
IV. Na § C h a n n e l s A. General Comments Voltage-gated Na + channels play an essential role in excitable cells where they are responsible for the rapid rise in the action potential upstroke and for the rapid conduction. Voltage-gated Na + channels belong to the general family of
660
SECTION IV Ion Channels as Targets for Toxins, Drugs, and Genetic Diseases
F I G U R E 4. Skeletal muscle Na + channel. (A) Primary structure of the a-subunit of the SkM-1 skeletal muscle Na + channel coded by the gene SCN4A. Mutations are indicated that lead to PC (paramyotonia congenita), HYPP (hyperkalemic periodic paralysis), and SCM (Na + channel myotonia). *The amino acid position numbers for PC and HYPP follow those of Cannon and Strittmatter (1993); however, for the mutations at the conserved glycine causing SCM, the numbering system of Lerche et al., (1993) was used. Unfortunately, the position numbers overlap here although the real amino acid locations are different. (B) Single Na + channel currents recorded in HEK293 cells after expressing the wild type and two genetically engineered constructs, Thr6698Met and Met1585Val, which correspond to natural mutations indicated in part A that cause HYPP. Ten single-channel current traces are shown above, and the averaged ensemble currents are shown below. Note that the wild-type channel shows brief single openings and a transient ensemble current. In contrast, the mutant Na § channels display long openings and reopenings giving averaged Na + currents with long noninactivating tails. These prolonged depolarizing inward currents lead to the myotonic behavior and, if sufficiently large, to the block of excitation and a flaccid paralysis. The HYPP mutant channels, unlike thePC channels, show a sensitivity to K + ions that further decreases their ability to inactivate. The PC channels, on the other hand, are sensitive to lowering the temperature, which has the same effect.
v o l t a g e - g a t e d cation c h a n n e l s , w h i c h includes Ca 2+ and K + channels. T h e p r i m a r y a - s u b u n i t f o r m i n g the N a + c h a n n e l (Fig. 4A) consists of about 2000 a m i n o acids. T h e r e are four repeats, e a c h c o n t a i n i n g six putative t r a n s m e m b r a n e helices. A m i n o acid analysis and site-directed m u t a g e n e s i s
studies have d e t e r m i n e d that the fourth t r a n s m e m b r a n e segm e n t in each repeat, called $4, acts as the v o l t a g e sensor. T h e c y t o p l a s m i c loop c o n n e c t i n g repeats III and IV is necessary for inactivation, and the r e g i o n b e t w e e n $5 and $6 in e a c h repeat, the P region, contributes to f o r m a t i o n of the
39. Ion Channels as Targets for Disease
hydrated, ion-conducting pore. Ca 2+ channels share a similar molecular design. The K + channel, however, departs from the scheme in that genes encode for only a single repeat and the channel is assembled from four identical or differing c~-subunit. Well-described disease entities involving voltage-gated Na + channel mutations include those of the human SCN4A gene on chromosome 17q, which codes for skeletal muscle Na + channels. These are discussed as a group next.
B. Skeletal Muscle SkM-I and SkM-2 Channels Multiple subtypes of Na + channels in skeletal muscle have been identified based on sensitivity to toxins and to antibodies, and these subtypes may be located preferentially in the T-tubules or surface membrane. Two Na + channels specific to skeletal muscle, known as SkM-1 (George et al., 1992) and SkM-2 (Kallen et al., 1990), have been cloned and sequenced. These channels are sometimes designated ~1 and /~2, respectively. Skeletal muscle Na + channels are heterodimers of a 260-kDa a-subunit and a 38-kDa/3-subunit (Kraner et al., 1985). Type SkM-1 is expressed in both innervated and denervated adult muscle and is blocked by nanomolar concentrations of tetrodotoxin (TTX) and p-conotoxin. The TTX-insensitive SkM-2 is not found in innervated adult mammalian muscle, but appears within hours following denervation, reaching a maximum at 48 h and then decreasing. Type SkM-2 is expressed in early development and disappears as type SkM-1 increases; thus SkM-1 is the only adult form. Type SkM-2 is also seen in vitro in skeletal muscle culture (absence of innervation) where it accounts for about 30% of the functioning channels; its sequence is essentially the same as the normal TTX-insensitive Na + channel of heart (Rogart et al., 1989).
1. Periodic Paralysis and Paramyotonia In a number of inherited diseases, periodic bouts of paralysis occur. Some of these conditions are associated with either hyperkalemia or hypokalemia, and myotonia may also be present. Recent studies have brought considerable light to understanding many of these conditions. Several genetic forms of myotonia and periodic paralysis may be due to failure of a small fraction of the skeletal muscle voltage-gated Na + channels to inactivate normally (Rudel and Lehmann-Horn, 1985). Hyperkalemie periodic paralysis (HYPP) and paramyotonia eongenita (PC) are related because they are due to autosomal dominant mutations of the ce-subunit of the SkM-1 Na + channel located in the surface membrane and T-tubular membrane of mammalian skeletal muscle fibers. These channels are normally responsible for conducting the skeletal muscle action potential (AP) throughout the tubular membranes, which is necessary for excitation-contraction coupling. Figure 4A shows the locations of several PC and HYPP mutations of the c~-subunit of the Na + channel (Fontaine et al., 1990). In Fig. 4B, the lack of effect on activation and the profound effect on inactivation are clearly seen in mutant constructs having point mutations in these regions (residue numbers
661
differ slightly from Fig. 4A due to species). Note in the single-channel records the longer openings and repeated reopenings of the mutant constructs. This accounts for the longer persistence of the averaged mutant sodium currents. Another way of stating this is to say that the mutant Na + channels exhibit a decreased tendency to inactivate. HYPP is an autosomal dominant mutation in which recurrent episodes of weakness occur with small elevations in the serum K + concentration. In this particular form of the disease, there is marked depolarization of skeletal muscle fibers during an attack, which makes the fiber inexcitable and leads to a flaccid paralysis of the muscle. Diseased biopsied fibers studied under voltage-clamp conditions yield TTX-sensitive Na + currents that do not completely inactivate; this has been shown to be the mechanism for depolarization (Lehmann-Horn et al., 1987). Single-channel recordings confirming this finding (Cannon et al. 1991) show that single Na + channels recorded from HYPP biopsied fibers have prolonged open times and tend to open repetitively (see Fig. 4A, B). In high extracellular K + concentrations (around 10 mM compared with a normal of 4-5 mM), a small fraction (5-10%) of the mutant channels do not inactivate and this produces a constant open probability of between 0.02 and 0.05, yielding a steady depolarizing current. This current is large enough to account for the depolarization that inactivates the unaffected Na + channels, resulting in action potential failure (Cannon et al., 1993). To demonstrate that lack of Na + channel inactivation is the possible cause of some forms of myotonia and periodic paralysis, two approaches have been taken. In the first approach, an in vitro model was created in rat muscle exposed to a polypeptide toxin (ATX-II) obtained from a sea anemone. The toxin at 10 pm produced a noninactivating open probability a t - 1 0 mV of approximately 0.02, similar to that reported for myotubes from HYPP patients, and myotonia was apparent (Cannon and Corey, 1993). In the second approach, a computer simulation was developed based on modifications of the Hodgkin-Huxley equations adapted to skeletal muscle fibers. This approach is similar to that discussed for computer simulation of low-gc~ myotonia. The simulated fiber, like the low-C1- model, required a Ttubular compartment to act as a diffusion-limited space in which activity-induced K + can accumulate. The computed APs for increasing degrees of incomplete inactivation of Na + channels effectively simulated normal, myotonic, and paralytic muscle. The simulated APs with altered Na + inactivation compared favorably with the abnormal APs produced pharmacologically with ATX-II (Cannon et al., 1993). PC, an autosomal dominant condition, has been also referred to as paradoxical myotonia because the signs of myotonia increase with use of the skeletal muscles, rather than diminish, and there is an increase in myotonia and a dramatic increase in the response of a muscle to percussion with local cooling. At least in vitro, low-C1-conductance myotonias are quite different since the signs are diminished with cooling below 27 ~ (Furman and Barchi, 1978). If one replaces the sensitivity to cooling with the worsening of symptoms with elevated serum K + levels, there is a physiological resemblance
662
SECTION IV Ion Channels as Targets for Toxins, Drugs, and Genetic Diseases
A Domain
I
II
IV
III
HoPP:R1239H
EA-2:AC4073 Extracellular
Intracellular FHM:T666M FHM:R192Q
I "EA-2:splice site /
\
FHM:I1811L
FHM:V714A
BB
NH2 " MH:R614C
COOH Cytoplasm SR Lumen F I G U R E 5. Ca2§ channels. (A) The primary structure of the al-subunit of the human skeletal muscle type-1 voltagesensor/channel (i.e., the DHP receptor) of the T-tubular membrane, and the neuronal P/Q-type Ca2§ channel are shown here superimposed on the same illustration. In skeletal muscle, two mutations linked to human hypokalemic periodic paralysis (HoPP) are indicated. Both mutations occur in segment 4 of domains II and IV, which is a region believed to be involved in gating of the channel. Using the same schematic, the primary structure is superimposed on the a 1-subunit of the neuronal P/Q-type Ca 2§ channel. The mutations causing familial hemiplegic migraine (FHM) and EA-2 are indicated. Note that the mutations causing FHM are of the missense type, resulting in single amino acid substitutions, whereas the two mutations for EA-2 both cause premature stops. (Redrawn and modified from Ophoff et al., 1996.) (B) The primary structure of the human ryanodine receptor indicating the main mutation responsible for malignant hyperthermia (MH). The mutation at the conserved Arg 614 is a substitution for Cys in the cytoplasmic region of the molecule. In porcine MH, Arg 615 is substituted in the same way. One should note that except for the four membrane-spanning segments MI through M4 (indicated here as simple 1 through 4), 90% of this molecule (5037 amino acids) is cytoplasmic. The functioning Ca 2§ release channel consists of four of these units (i.e., a homotetramer) in the SR membrane. Activation to the T-tubular DHP receptor causes activation of the Ca2§ release channel in the process of excitationcontraction coupling. One of many speculations as to how this is accomplished is that a third molecule, triadin, interacts with the connector between domains II and III of the DHP receptor (shown in A), giving a mechanical coupling between the two systems.
between PC and HYPP. Both of these conditions have myotonia with episodic weakness and the similarities appear to be explained by the findings that both diseases are due to different mutations but on the same skeletal muscle Na § channel gene (Ptacek et al., 1993; McClatchey et al., 1992). 2. Na § C h a n n e l M y o t o n i a
S o m e mutations in the S C N 4 A gene for the SkM-1 channel can p r o d u c e in patients a relatively pure form of m y o t o n i a k n o w n as h u m a n N a + c h a n n e l m y o t o n i a ( S C M ) (Lerche et al., 1993). This s y n d r o m e is devoid of dystrop h y and weakness, and does not r e s e m b l e PC or HYPP, al-
though the latter pair is also due to mutations in the same gene (see Fig. 4A). G l y c i n e at position 1306 is c o m p l e t e l y c o n s e r v e d in all Na § channel c~-subunits in all species. This is an important region for control of fast N a § channel inactivation and it has even been p r o p o s e d that the two glycines at 1306 and 1307 m a y act as a h i n g e to b l o c k the channel pore ( K e l l e n b e r g e r et al., 1997). In families with SCM, the glycine at 1306 is substituted by glutamine, valine, or alanine. The g l u t a m i n e substitution p r o d u c e s the m o s t severe m y o t o n i a . N o n e of the mutations in the patients affects inactivation as m u c h as the substitution for phenylalanine at 1311, but the degree is sufficient to account for the m y o t o n i a . P a t c h - c l a m p studies of the S C M channel
663
39. Ion Channels as Ta~ke~s for Disease
show a slowing of the rate of fast inactivation and an increased frequency of late channel openings. These parameters lead to an increase in the ratio of late current to peak current from control of 0.4% to 6%. The simulation experiments using ATX-II anemone toxin (Cannon and Corey, 1993) with computer modeling showed that myotonia occurred when only 2% of Na + channels fail to inactivate, which corresponds to a late over-peak ratio of 5%. It should be clear that a larger late depolarizing current could result in restimulation and repetitive firing, that is, myotonia. Thus, SCM and its severity can be explained by the length and charge of the group substituted for Gly 1306.
C. Altered Neuronal Na + Channel Function in Febrile Epilepsy Febrile seizures are the most common seizure disorder, affecting approximately 3% of all children under six years of age. Some of these children develop ongoing epilepsy with afebrile seizures later in life. Linkage studies in one large family mapped this disorder to chromosome region 19q13.1 and a mutation in the voltage-gated Na + channel/31 subunit gene (SCN1B) was identified (Wallace et al., 1998). The mutation changes a conserved cysteine residue and coexpression of mutant fll subunits with a brain Na + channel a subunit in Xenopus oocytes shows that the mutation reduces the rate of Na + channel inactivation. The complex inheritance pattern of this disorder raises the possibility of involvement of other Na + channel subunit genes in febrile seizures and generalized epilepsies.
the distal nephron, distal colon, and the airways. ENaC is a heteromere of three different but homologous subunits (a,/3, and 3/). Mutations within any of these subunits result in either loss or gain of function (rev. in Hummler and Horisberger, 1999). Loss of ENaC function results in pseudohypoaidosteronism type 1, which is characterized by urinary loss of Na + and reduced K + excretion. Severe forms of this syndrome are inherited as an autosomal recessive trait that may result in lethal episodes of hyponatremia, hypotension, and hyperkalemia. Certain mutations in the/3- and y-subunits result in a gain of ENaC function, Liddle's syndrome, giving rise to hypertension due to Na + retention. Metabolic alkalosis and hypokalemia are often associated with this disorder.
V. Ca 2+ C h a n n e l s A. General Comments There are several types of Ca 2+ channels, both agonist activated and voltage controlled. The primary structure of the a subunit of the L-type Ca 2+ channel of skeletal muscle is shown in Fig. 5A. This channel is blocked by dihydropyridines (DHPs), so it is also known as the skeletal muscle DHP receptor; in addition it is the T-tubule voltage sensor. (For a review of the important voltage-gated Ca 2+ channels, see Tsien et al., 1991.) The Ca 2+ release channel of the sarcoplasmic reticulum (Fig. 5B) can be reviewed from the references given in MacLennan and Phillips (1992).
B. Dihydropyridine Receptor D. Cardiac Na + Channel: Long-QT-Interval Disease
1. Muscular Dysgenesis
Like neuronal and skeletal muscle action potentials, the Na + current is responsible for the initial upstroke of the action potential in cardiac ventricular and Purkinje cells. The Na + channels quickly inactivate but the action potential remains depolarized for its duration due to the rectification of the inwardly rectifying K + current and the activation of Ca 2+ currents. The action potential is repolarized by the Ca 2+dependent inactivation of the Ca 2+ current and activation of the delayed-rectifier K + currents, lie and/Ks" The cardiac Na + channel is coded by the SCN5A gene, and mutations in this gene produce aberrant Na + channels in which fast inactivation is delayed (Roden et al., 1995). The end result is a cardiac action potential that has a prolonged repolarization, which is seen as a distinctive long QT interval on the electrocardiogram. Associated with the repolarization defect is a serious arrhythmia leading to syncope (fainting), seizures, and sudden death from a distinctive ventricular tachycardia known as torsades de pointes. The long-QT syndrome is also genetically linked to sites other than the Na + channel gene (Keating and Sanguinetti, 1996). These sites include mutations in the genes that encode lie and/Ks" (See Section VI.)
Mice that are homozygous for the autosomal recessive dysgenic gene mutation (mdg/mdg) lack excitation-contraction (EC) coupling and the slow Ca 2+ current. These animals are therefore incapable of muscular movement and die at birth, presumably from respiratory paralysis. The physiological problem is related to lack of function of the DHP receptor. The mutant dysgenic animals have a five-fold decrease in DHP binding in skeletal muscle and a low level of mRNA encoding the DHP receptor; monoclonal antibodies detect no DPH receptors. On the other hand, cardiac DHP binding is normal, illustrating the fact that the L-type Ca 2+ channels (or DHP receptors) are distinctly different in these two tissues. Muscular dysgenesis appears to have been the first documented example of a genetic disease in vertebrates that is produced by a defect in a structural gene coding for an ion channel (Tanabe et al., 1988). Although a human counterpart has not been reported for this mutation, basic studies with the dysgenic mouse model have helped confima the role of the DHP receptor as both a voltage sensor and channel in the T-tubular membrane. A critical experiment was performed in which both EC coupling and slow Ca 2+ current were restored in the cultured dysgenic muscle cells by microinjection of an expression plasmid (pCAC6) carrying cDNA for the missing DHP receptor (Tanabe et al., 1988). The restoration of both voltage sensing and ion channel functions with the single cDNA confirmed the dual role of this protein in the T-tubule. These mice have been used extensively
E. Non-Voltage-Gated Na + Channels in Disorders Involving Epithelial Transport Non-voltage-gated Na + channels such as the epithelial Na + channel (ENaC) control transepithelial Na + reabsorption in
664
SECTION IV Ion Channels as Targets for Toxins, Drugs, and Genetic Diseases
to examine the coupling between the DHP receptor and the sarcoplasmic reticulum (SR) Ca 2+ release channel and to show conclusively the difference between cardiac- and skeletal-like EC coupling (Beam and Franzini-Armstrong, 1997).
2. Hypokalemic Periodic Paralysis Hypokalemic periodic paralysis (HoPP) is a disease of skeletal muscle in which episodes of weakness occur in association with lowered serum K + levels. It is an autosomal dominant trait, but in one-third of the cases it appears spontaneously. A typical attack of weakness induced by low serum K + can begin a few hours after a high carbohydrate meal, often when asleep, and can last for 4-24 h. Other precipitating factors include rest after exercise and situations where K + moves into muscle cells. For years this disease was thought to be involved somehow with abnormal pump (Na+,K +ATPase) or cation channel function and the syndrome can be reasonably modeled in rats that have their skeletal muscle K + contents altered through insulin and/or low K + diet. It came as a surprise then when linkages were discovered to the gene that encodes the skeletal muscle L-type Ca 2+ channel. Two mutations were described and are shown in Fig. 5A. In each mutation there is a replacement of a highly conserved arginine by histidine in the $4 segment, a region believed to be involved in channel gating. One replacement occurs in the D4 domain at Arg 1239 (Ptacek et al., 1994), and the other occurs in the D2 domain at Arg 528 (Jurkat-Rott et al., 1994). Recent studies of these mutations expressed in heterologous systems have thus far been disappointing in that they have not revealed abnormal functions of this channel that can explain HoPE (For further discussion see LehmannHorn and Rudel, 1996.) C. Ca 2§ R e l e a s e Channel
1. Malignant Hyperthermia Malignant hyperthermia (MH) is a rare autosomal dominant condition in humans that predisposes these individuals to react to anesthesia with muscle rigidity, hypermetabolism, and high fever; the muscles become highly stimulated metabolically, with a consequent rise in body temperature (MacLennan and Phillips, 1992). A popular form of anesthesia involving use of halothane in conjunction with a depolarizing neuromuscular blocker, succinylcholine, can trigger MH. If not treated immediately these patients may die within a few minutes from ventricular fibrillation or within hours to days from neurological or renal complications. Although MH is rare, it is serious enough to be fatal in apparently healthy individuals undergoing anesthesia. MH is also seen in animals where it is often triggered by heat stress. MH-susceptible pigs have been recognized for many years (Lucke et al., 1979) and genetic and physiological studies of the MH pigs have hastened our understanding of the comparable condition in human patients. The mechanism of MH is a mutation in the skeletal muscle SR Ca 2+ release channel (Fig. 5B). The specific mutations that account for the dysfunction of this channel are different in human and pigs. The Ca 2+ release channel is a key
element of the EC coupling process. The AP invades the Ttubular membrane causing movement of gating charge in the DHP-sensitive Ca 2+ channel/voltage sensor, which in turn activates the SR Ca 2+ release channel to release Ca 2+ into the myoplasma to initiate contraction. The exact means by which the voltage sensor of the T-tubular membrane couples with the release channel of the SR is not yet fully understood. In addition, Ca 2+ pumps located in the SR reaccumulate the released Ca 2+. These events and a scheme to explain the pathophysiology of MH are depicted in Fig. 6A and B. The Ca 2+ release channel is opened specifically by nanomolar concentrations and blocked by micromolar concentrations of the plant alkaloid ryanodine. Thus, this channel is also referred to as the ryanodine receptor. Using this high-affinity ligand, the ryanodine receptor was purified and several fulllength cDNAs from different species were subsequently cloned. Two different genes encode the channel: RYR1 on chromosome 19 codes for the skeletal muscle channel, and RYR2 on chromosome 1 codes for the heart and brain channel. In pigs the cDNA sequences for RYR1 between normal and MH pigs predict a single amino acid difference, a cysteine for an arginine in the N-terminus (Fig. 5B). In human studies two mutations of RYR 1 are associated with MH, but some MH patients have no linkage with this gene; further, mutations of this gene may cause unrelated forms of muscle disease. Human MH may therefore involve other gene mutations and interactions (see Gillard et al., 1992). These major isoforms of the ryanodine receptor are the only known cellular binding sites for the ligand. The protein is a large homotetrameric complex constructed from 565-kDa subunits. The channel region is probably located in the membrane-spanning segments, which are 20% of the subunit from the C-terminal. The remainder of each subunit is cytoplasmic and the four subunits together form a cytoplasmic structure that make up the feet proteins of the triadic junction (Block et al., 1988). This cytoplasmic region also contains the sites for coupling to the DHP receptor. All of the signs of MH as seen in Fig. 6A and B can be explained by a defect in the regulation of intracellular Ca 2+. Sustained Ca 2+ levels in the fiber would cause contracture, increased glycolytic and anaerobic metabolism with depletion of ATE glucose, and oxygen, and overproduction of CO 2, lactic acid, and heat. A specific antagonist for the MH reaction is dantrolene Na +, which is capable of blocking release of intracellular Ca 2+ from the SR.
2. Central Core Disease Another autosomal dominant disease of the RYR1 gene is central core disease (CCD). It is characterized by hypotonia and proximal weakness first seen in infants, but appears to be nonprogressive. The severity of the disease is variable within families. To date, three mutations causing the disease have been identified among affected families. The substitution of histidine for a conserved arginine (R2434H), which is absent in the general population, was identified in a large Canadian family. Two other missense mutations, I404M and R163C, have also been identified. One might presume that these channel defects should cause EC coupling malfunction and explain the observed pathology; however, this is still an open question (George, 1995).
39. IonChannelsas TargetsforDisease
665
AT
~AI] ATP1 ~ C a~2+
Ca2+~ - ~
k.c,,/
C a 2+
2+
II
] ]' Kinase Phosphorylase 'I _1
Lact
/
Ca2+
,b---
~tycogen ATe
~P
i,l,l,i,l,i1
p
~~
ATP
ATP
Phosphorylase Kinase
Gluc ,se-l-P
~ ~ ADP C ATP
Lactic acid
k
~d
f
ATP
Heat CO2
FIGURE 6. Malignant hyperthermia. (A) Normal physiological mechanism of EC coupling. The C a 2+ release channel (shown here as square elements) is a key element of the process known as excitation-contraction coupling. (B) MH muscle contraction cycle. Mutant C a 2+ release channels can explain the basic mechanism of MH. The mutations produce a release channel that is more sensitive to opening stimuli and fails to close rapidly, thus leading to abnormally high C a 2+ levels. Sustained C a 2+ levels in the fiber then would cause contracture, increased glycolytic and anaerobic metabolism with depletion of ATE glucose, and oxygen, resulting in overproduction of CO2, lactic acid, and heat. M indicates a mitochondrion. (Modified from MacLennan and Phillips, 1992; copyright 9 1992 AAAS.)
D. Neuronal Ca2§ Channels The neuronal Ca 2+ channels are like the L-type skeletal muscle channels discussed earlier. They are multimeric entities consisting of one each of or1-, c~2-, ~-, and/3-subunits. The central Crl-SUbunit has the voltage sensor and the conducting pore. The neuronal channels discussed here differ from the skeletal muscle L-type channel in lacking the DHP receptor on the a 1-subunit. Two different classes of neuronal Ca 2+ channel diseases are considered. The first, LambertEaton syndrome, is due to autoimmune block of Ca 2+ channels in motor nerve terminals leading to skeletal muscle weakness. The second class is a pair of neurological diseases related by the fact that they are caused by different types of mutations in the same gene that codes for a central nervous system Ca 2+ channel. The types of mutations in the same channel determine the resultant form of the disease.
1. Lambert-Eaton Syndrome
In Lambert-Eaton syndrome autoimmune-generated antibodies reduce the number of voltage-activated Ca 2+ channels in the presynaptic nerve terminal. This decreases the amount of Ca 2+ entering the nerve terminal with each
nerve impulse. Because Ca 2+ entry is necessary for the release of transmitter release at the presynaptic terminal, the amount of acetylcholine delivered to the postsynaptic AChR channels will be reduced causing neuromuscular transmission failure and muscle weakness. There are many subtypes of neuronal Ca 2+ channels; however, the type involved in Lambert-Eaton syndrome is specifically labeled by the marine snail toxin, eo-conotoxin (see Penn et al., 1993). Lambert-Eaton syndrome has a presynaptic origin that is to be contrasted with myasthenia gravis (another form of neuromuscular weakness, discussed in Section VII.B), which has a postsynaptic origin. Patients with Lambert-Eaton syndrome often have an accompanying carcinoma and it has been asserted that the antibodies to the cancer cells also recognize the Ca 2+ channel.
2. Hemiplegic Migraine and Episodic Ataxia Type 2 Migraine is a common neurological disorder characterized by recurrent severe headaches accompanied by phonophobia, photophobia, nausea, and vomiting. It occurs in about 24% of females and 12% of males, and varies greatly with regard to frequency, duration, and severity. Familial hemiplegie migraine (FHM3) is a rare autosomal dominant
666
SECTION IV Ion Channels as Targets for Toxins, Drugs, and Genetic Diseases
form of migraine, which in addition to the usual migraine symptoms has an accompanying hemiparesis (paralysis of one side of the body) and sometimes progressive cerebellar atrophy. Episodic ataxia type 2 (EP-2) is an autosomal dominant neurological disorder characterized by ataxia of cerebellar origin, and resembles FHM in also involving migraine attacks and cerebellar atrophy. The episodic ataxia component responds to treatment with acetazolamide (a carbonic anhydrase inhibitor); otherwise it is similar to the K + channel defect episodic ataxia type 1 (see Section VI.B). FHM and EP-2 are described together because these diseases have both been mapped to chromosome 19p13 and are due to mutations in the gene (CACNL1A4) that codes for a brain Ca 2+ channel known as the P/Q-type Ca 2+ channel. Four missense mutations causing FHM and two premature stops causing EP-2 have been identified. The primary structure of the brain P/Q-type Ca 2+ channel and the sites of the mutations causing the aforementioned diseases are shown in Fig. 5A. The electrophysiological dysfunctions of the mutant channels in FHM and EP-2 have not been yet been identified. The four missense mutations associated with FHM could produce a situation analogous to the missense mutations of the skeletal muscle Na + channel that cause hyperkalemic periodic paralysis, paramyotonia, and Na + channel myotonia (see Section IV). In these latter conditions, the Na + channels have impaired inactivation, and by analogy in FHM the Ca 2+ channels would exhibit increased activity. The cellular mechanism proposed is that increased Ca 2+ activity would promote the process known as spreading cortical depression (Ophoff et al., 1986), which would in turn facilitate initiation of a migraine attack. On the other hand, the mutations occurring in EP-2 are classed as premature stops (i.e., frame shifts and splice-site mutations), and these would lead to truncated a-subunits consisting only of repeats I and II and part of III. Thus, the EP-2 channels would not be expected to function and probably would be present in lower density, the end result being decreased Ca 2+ channel function in EP-2 with consequent dysfunction of the cerebellum, which would account for the ataxia.
VI. K § C h a n n e l s A. General Comments K + channels serve many important functions in excitable tissues such as action potential repolarization and stabilization of the resting membrane potential. Therefore, it is not surprising that dysfunction of K + channels might be expected in neurological or cardiac disorders related to electrical activity. In fact, the very first mutation described in a mammalian K + channel causes a human neurological condition, episodic ataxia type l, EA-1. Subsequently, the congenital long QT syndrome and inherited idiopathic epilepsy were linked to several different K + channels. From a historical perspective, the defective K + channel in EA-1, K v 1.1, is a member of the Shaker K + channel family. The finding that Kvl.1 missense mutations underlie EA-1 presents analogies to the Shaker disorder in Drosophila.
The Shaker locus from Drosophila was the first K + channel cloned and characterized at the molecular level. Using Shaker sequences as probes, numerous other voltagedependent K + channels of similar structural motif have been identified, with Kvl.1 being the first mammalian homolog cloned. Mutations induced in voltage-gated K + channels in Drosophila have been very useful in determining the molecular basis for gating and inactivation. Many descriptive terms have been introduced to describe the families of K + channels that came out of the Drosophila studies. For example, Shaker and ether a-go-go refer to those K + channels having mutations in the channel gene that give rise to phenotypic behaviors in response to the challenge of ether vapor. In essence, some mutants would shake their legs, while others would wriggle their tails provocatively. The Drosophila studies have become increasingly useful as more structural similarities are being found to Drosophilarelated channels in vertebrates. Voltage-dependent Na + and Ca 2+ ion channels are sometimes thought of as pseudotetramers since their main functioning subunit consists of four interconnected domains, each having six membrane-spanning segments (Fig. 7A, B). K + channels, in contrast, are true tetramers with each subunit coded by a single gene. There are six membrane-spanning segments for the voltage-gated K + channels and only two in the case of non-voltage-gated inward-rectifying K + channels. The three main functional types are voltage gated, Ca 2+ activated, and inward rectifying. Many have been cloned and they consist of homotetramers or heterotetramers within the same class. Thus, there is the possibility for mixing subunits that would allow for the evolution of designer channels to match particular different functional circumstances. There is a systematic classification of mammalian K + channels, and for greater detail one should consult Gutman and Chandy (1993).
B. Episodic Ataxia Type 1 Hereditary episodic ataxia type 1 (also called myokymia) is an autosomal dominant disease characterized by periodic attacks of motor imbalance, uncoordination, and involuntary tremor. These attacks are precipitated by emotional or physical stress. The condition is caused by point mutations in the Kvl.1 (or KCNA1) gene (located on chromosome 12p) of the Shaker K § subfamily, which codes for a delayed outward-rectifying K + channel present in neurons (Browne et al., 1994). The predicted primary structure of the Kvl.1 channel and the known mutations causing EA-1 are shown in Fig. 7A. The Kvl.1 K § channel displays the C-type inactivation (conformational change at the mouth of the pore) and N-type inactivation is conferred by an associated/3-subunit, Kv/31. A common defect shared by many of the mutant channels thus far examined is a reduction in current amplitude for homomeric EA-1 channels when expressed in Xenopus oocytes. Some of the mutations shifted the V1/2 for activation in the depolarizing direction accelerated the rate of channel closure (Adelman et al., 1995; Zerr et al., 1998). Because heterozygous individuals show neurological signs and no homozygous individuals have yet been found, the many other types of K § channels that might be present are
667
39. Ion Channels as Targets for Disease
FIGURE 7. K + channels. (A) Structure of the Kvl.1 K + channel (KCNA1 gene), which is implicated in the human neurological condition EA-1. Point mutations from 14 families having this condition are indicated. (B) Structure of the HERG (human ether a-go-go related gene) inward-rectifying K + channel of the human heart, which is implicated in the K + channel type of the long QT syndrome. Several types of mutations found in this condition are indicated. There are three point missense mutations, N470D, A561 V, and G628S. There is also a deletion between I500 and F508, a protein truncation due to a single base-pair deletion in S 1, and a splicing error in the carboxy-terminus. The errors in S 1 and $3 are unlikely to form functional channels, whereas the point mutations might do so. (Redrawn and used with permission of Roden et al., 1995.)
unable to take over the function of the Kvl. 1 channel. This speaks against a current notion that a large number of different types of K + channels have evolved to prevent just such a dependence. Coexpression studies show that the mutant subunits, in addition to forming homotetramers characteristic of the Kvl. 1 channel, mix with those from the wild-type allele and assemble to form a variety of functionally defective heterotetramers in vivo (Zerr et al., 1998). The cellular mechanism suggested by these biophysical findings is that nerve cells having mutant channels would not effectively repolarize after each AP due to the impaired delayed rectifier function of the mutant channels. Though an animal model of EA-1 has not yet been developed, mice deficient of Kvl.1 may provide useful clues for the cellular role of K vl. 1. Homozygous K vl. 1 knockout mice, while not replicating the symptoms of EA-1, do have frequent
generalized epileptic seizures (Smart et al., 1998). Physiological studies show that deletion of Kvl. 1 gives rise to tonic inhibitory tone on cerebellum Purkinje neurons and enhances excitability of the basket cells in the cerebellum by selectively increasing the likelihood of action potential propagation past axonal branch points (Zhang et al., 1999). It is therefore likely that missense mutations of K v 1.1 linked to EA-1 will similarly affect inhibitory tone in the cerebellum, resulting in an imbalance between inhibitory and excitatory input and thereby destabilizing motor control under stress or exercise. C. Long QT Interval Disease
Inherited long QT syndrome is a cardiac disorder characterized by fainting (syncope), seizures, and sudden death
668
SECTION IV Ion Channels as Targets for Toxins, Drugs, and Genetic Diseases
due to a particular ventricular tachycardia called torsades de pointes. Patients having the disease have a characteristically prolonged QT interval, which indicates dysfunction of ventricular repolarization. Mutations in both a Na + (Section IV.D) and K + channels have been linked to long QT. The Na + channel remains open too long and the K + channel opens less efficiently. Both conditions produce a similar physiological effect, namely, a prolonged cardiac AP as reflected in the long QT interval of the electrocardiogram. The delayed-rectifier K + current, which is a key regulator of cardiac repolarization, is composed of two components: a rapid (IKr) and a slow (/Ks) activating K + current (Sanguinetti and Jurkiewicz, 1990). Mutations in the genes underlying these channels have been linked to long QT. HERG encodes the primary a-subunit for IKr and minK (KCNE1) and KVLQT 1 (KCNQ1) encode subunits that coassemble to form IKs. Mutations in HERG and KVLQT1 are the most common cause of long QT. The primary structure of the HERG channel and some of the identified mutations causing long QT syndrome are shown in Fig. 7B. Though not all mutations in HERG or KVLQT1 have the same effect on K + channel function, the functional consequence of each is delayed cardiac repolarization with increased risk for torsades de pointes and sudden death.
D. Inherited Epilepsy of Newborns Benign neonatal familial convulsions (BNFC) is a subtype of idiopathic generalized epilepsy that accounts for --40% of epilepsy up to the age of 40. A search for the gene underlying this form of epilepsy converged with channel specialists seeking to identify the molecular mechanism underlying cholinergic modulation of excitability in the central nervous system. The acetylcholine sensitive M-current, which is a key regulator of neuronal excitability, has gained prominence as a potential contributor to normal cognition, dementia, and epilepsy. Thus when linkage studies identified mutations in genes encoding two new K + channel subunits, KCNQ2 and KCNQ3, as responsible for B NFC (Biervert et al., 1998; Charlier et al., 1998; Singh et al., 1998), speculation developed that these subunits might underlie the Mcurrent. KCNQ2 and KCNQ3 belong to the KCNQ family that includes KCNQ1, one of the genes affected in long QT syndrome. KCNQ2 and KCNQ3 are expressed in overlapping patterns in the brain, and their coexpression in Xenopus oocytes yielded currents with gating and pharmacological properties similar to the neuronal M-current, which supported the idea that KCNQ2 and KCNQ3 do indeed underlie the M-current. To date the known B NFC mutations in KCNQ2 and KNCQ3 cause a reduction in current amplitude. BNFC thus results from a loss of function of the M-current in agreement with experimental data showing that muscarinic agonists, which suppress the M-current, cause acute status epilepticus and induce a chronic seizure.
long QT that was linked to mutations in either KCNQ1 or KCNE1 as described for long QT disorders. KCNQ1 and KCNE1 coassemble to form/Ks and are expressed in the stria vascularis endothelium in the inner ear. The stria is responsible for secretion of the K+-rich endolymph that fills the middle chamber and bathes the sound-receptive hair cells. In Jervell and Lange-Nielsen syndrome, secretion of endolymph does not occur normally, presumably due to reduced/Ks function. Consequently, the middle chamber of the cochlea collapses and the hair cells degenerate. A form of nonsyndromic dominant deafness has been linked to mutations in a fourth member of the KCNQ K + channel family, KCNQ4. This mutation exerts a dominant negative effect on wild-type KCNQ4. KCNQ4 is expressed in the outer hair cells, which serve to mechanically amplify sound vibrations within the cochlea.
F. Acquired Neuromyotonia In addition to the two human diseases due to mutations of K + channels, there is an autoimmune disease involving a gated K + channel known as acquired neuromyotonia (Sinha et al., 1991). In this disease the peripheral motor nerves are hyperexcitable, resulting in repetitive firing and abnormal contractions of skeletal muscle. This must be clearly distinguished from the true myotonia discussed previously, in which the muscle fibers but not the nerve fibers are hyperexcitable. The two syndromes can be distinguished by administration of a neuromuscular blocker such as tubocurarine and, with direct muscle stimulation, the true myotonic fibers remain hyperexcitable, whereas in the neural type of myotonia the muscle fibers have normal excitability.
Vii. N e u r o t r a n s m i t t e r - G a t e d Channels A. General Comments Several types of channels are gated open by neurotransmitters. The major ones can be classed by the neurotransmitter effective at their receptors. These include acetylcholine (ACh), glutamate, 5-HT, adenosine triphosphate (ATP), gammaaminobutyric acid (GABA), and glycine. The selectivity of the channels can be for nonspecific cations, such as the ACh nicotinic channel, the K+-selective ATP channels, or the Cl-selective glycine or GABA channels. This class of channels usually has five subunits, with one of the units occurring more than once. Only a few mutations among this class of channels have been described for human disease, although the channels have been targets of autoimmune disease. Our disease examples include a clinically relevant one involving the skeletal muscle nicotinic receptor channel, and a rare but instructive one involving the strychnine receptor glycine channel.
E. Deafness
B. Nicotinic Acetylcholine Receptor: Myasthenia Gravis
Two forms of deafness are related to mutations in different K + channels. Jervell and Lange-Nielsen syndrome is a recessive disorder characterized by bilateral deafness and
In the human disease myasthenia gravis, the number of nicotinic acetylcholine receptors (AChRs) at the postsynaptic side of the neuromuscular junction is decreased, resulting
669
39. Ion Channels as Targets for Disease
in smaller postsynaptic currents and the tendency to block neuromuscular transmission. As a result patients experience weakness of skeletal muscles. The AChR is a membranebound glycoprotein that exists as a pentamer. Mammals have two main isoforms: the mature or innervated form having two c~- and one each of/3-, ~-, and E-subunits, and the immature or denervated form in which a y-subunit replaces the E-subunit. The a-subunit has the ACh receptor and all five subunits are necessary for ligand-gated channel function during neuromuscular transmission. When ACh is released from the nerve terminal it diffuses to the AChR, and when two molecules occupy each c~-receptor, the AChR channel can open, causing both Na + and K + currents to flow and resulting in depolarization of the postsynaptic membrane and excitation of the muscle fiber. Most of the myasthenia gravis syndromes are apparently due to a T-lymphocyte-dependent serum auto-antibody against AChR (see Penn et al., 1993). Experi-mentally produced monoclonal antibodies have been shown to cause the channel to make kinetic transitions, leading to desensitization rather than activation. Myasthenia gravis may also occur very rarely due to heredity, reportedly by a mutation in the e-subunit.
C. Glycine Receptor: Hyperekplexia Glycine-activated C1- channels are found in many central neurons, where they function to produce inhibition. Activation of these channels causes an increase in gcl, which tends to clamp the membrane potential to the C1equilibrium potential, which is often in the hyperpolarizing direction. The effects of the excitatory depolarizing channels are thus reduced causing inhibition. The mammalian glycine-activated channel is a heterooligomeric C1- channel that is inhibited by the plant toxin strychnine. The adult receptor is a pentamer consisting of three 48-kD ce-subunits and two 58-kDa/3-subunits. The a-subunit of the channel protein has the strychnine binding site and behaves as a channel when expressed in Xenopus oocytes. There are several subtypes of this channel and a large degree of sequence homology to the GABA, glutamate, and ACh ligand-gated ion channels. In mammals small doses of strychnine cause highly exaggerated startle reactions to unexpected acoustic or mechanical stimulation. A response pattern resembling strychnine poisoning is seen in patients afflicted with the autosomal dominant disorder known as hereditary hyperekplexia (HHE) or familial startle disease (STHE) (Shiang et al., 1993). Mutations in the a-subunit of the glycine-activated C1-channels have been shown to be the cause. The proposed structure of the a-subunit seen in Fig. 8 shows position 271 between domains M2 and M3, where a positively charged arginine is substituted by a neutral leucine or glutamine in the major mutations. These mutations lie at the presumed mouth of the channel, where altering the charge has a significant effect on channel function. The fact that three a-subunits are required in the functional adult receptor helps explain the dominance of the inheritance. If the channel is rendered dysfunctional if only one mutant subunit is present, then even if there is only one mutant allele (the other producing normal subunits), the het-
FIGURE 8. Glycine receptor C1- channel. The c~-subunit of the glycine-activated, strychnine-blocked C1- channel (or GLRA1 channel) present in the mammalian nervous system. The functional pentameric channel is assembled from three of these ce-subunits and two/3-subunits (not shown). The arrow points to position 271 near the presumed mouth of the assembled channel where autosomal dominant mutations cause a charged lysine to be replaced by either an uncharged leucine or glutamine. The result is a blocked inhibitory channel leading to the nervous hyperexcitability called hereditary hyperekplexia (HHE). erozygous individual would have only one-eighth or 12.5% of the normal functioning channels. This theory was discussed earlier for dominant MC.
VIII. Summary Ion channels are often targets of disease, either directly through a mutation in the gene coding for some critical part of the channel protein structure, or indirectly through autoimmune disease or by defective coding of a critical ion channel regulator. We have seen examples of each case. When ion channel function is abnormal, we may get malfunction of excitable cells as occurs in the myotonias, periodic paralyses and episodic ataxias, or problems with ion transport as occurs in cystic fibrosis and malignant hyperthermia. No region of the channel protein appears to be immune to the effects of mutation. Effects are seen on gating, selectivity, and unit conductance. Often a change in a single amino acid accounts for the altered function of a channel or of its regulator. Ion channel diseases are often called channelopathies. Modern electrophysiological methods, including the patch-clamp method, have made the monitoring of channel function very direct and convenient. Combining the new electrophysiology with the new molecular biology has been successful in offering us a staggering amount of information on normal and diseased channels in a remarkably short period of time. Common methods are evolving for the study of familial diseases. Two basic strategies are used, positional cloning and the candidate gene. The first approach, positional cloning, is used to identify the genetic locus even when there is no knowledge of the biophysical or biochemical abnormalities underlying the disease by examining affected individuals for genetic linkage to polymorphic markers located
670
SECTION IV Ion Channels as Targets for Toxins, Drugs, and Genetic Diseases
throughout the genome. In this way a region of the gene is found that is linked to the disease. Linkage is established when there is a low rate of recombination between the marker locus and the disease locus. From the gene one can express the protein it encodes and then determine its function and how the disease mutations disturb this function. The CFI'R channel of cystic fibrosis was discovered in this way. The second approach, or candidate gene strategy, requires knowledge of functional abnormalities brought about by the disease. Genes responsible for maintaining the normal function of the process affected by the disease are then considered to be candidate genes for the disease. Genetic linkage to the disease is then sought, and if one is found the candidate gene is screened for mutations. These mutations can be incorporated into the wild-type D N A and expressed in heterologous systems. Then one can examine the biophysical and biochemical behavior and determine if the abnormality could account for the disease. The new rapid cloning techniques are largely responsible for the increased pool of known channel genes having known functions. This makes it easier to identify candidate channel genes; thus the candidate gene approach has been very successful for the study of the channelopathies. Most of the new channelopathies including the CLC5 mutations causing kidney stones, the episodic ataxias, and the long QT interval diseases were discovered using the candidate gene approach. With the increase in the number of channelopathies now described we can begin to make certain generalizations. We have seen that if more than one channel type is involved in control of cellular excitability, then either decreasing the P oven of the stabilizing channel or increasing the P o-n of the depolarizing channel will lead to hyperexcitabihtWy of the excitable cell. Examples to date are reduced C1-channel function or decreased Na + channel inactivation in the myotonias, aberrant K + channel block expression or decreased Na + channel inactivation in the long QT syndromes, and reduced K + channel function or increased activation of Ca 2+ channels in the two types of episodic ataxias.
Bibliography Adelman, J. P., Bond, C. T., Pessia, M., and Maylie, J. (1995). Episodic ataxia results from voltage-dependent K § channels with altered functions. Neuron 15, 1445-1454. Adrian, R. H., and Bryant, S. H. (1974). On the repetitive discharge in myotonic muscle fibres. J. Physiol. 240, 505-515. Adrian, R. H., and Marshall, M. W. (1976). Action potentials reconstructed in normal and myotonic muscle fibres. J. Physiol. (London) 258, 125-143. Ashcroft, E M. (1999). "Ion Channels and Disease." Academic Press, San Diego, CA. Ballabio, A. (1993). The rise and fall of positional cloning? Nature Genet. 3, 277-279. Beam, K. G., and Franzini-Armstrong, C. (1991). Functional and structural approaches to the study of excitation-contraction coupling. Methods Cell Biol. 52, 283-306. Beck, C. L., Fahlke, C., and George, A. L., Jr. (1996). Molecular basis for decreased muscle chloride conductance in the myotonic goat. Proc. Natl. Acad. Sci. USA 93, 11248-11252 Becker, P. E. (1973). Generalized non-dystrophic myotonia: the dominant (Thomsen) type and the recently identified recessive type. New Devel. Electromyography Clin. Neurophysiol. 1, 407-412.
Behrens, M. I., Jalil, R, Serani, A., Vergara, E, and Alvarez, O. (1994). Possible role of apamin-sensitive K + channels in myotonic dystrophy. Muscle Nerve 17, 1264-1270. Biervert, C., Schroeder, B. C., Kubisch, C., Berkovic, S. E, Propping, E, Jentsch, T. J., and Steinlein, O. K. (1998). A potassium channel mutation in neonatal human epilepsy. Science. 279, 403-406. Block, B. A., Imagawa, T., Campbell, K. E and Franzini-Armstrong, C. (1988). Structural evidence for direct interaction between the molecular components of the transverse tubule/sarcoplasmic reticulum junction in skeletal muscle. J. Cell Biol. 107, 2587-2600. Browne, D. L., Gancher, S. T., Nutt, J. G., Brunt, E. R. P., Smith, E. A., Kramer, E, and Litt, M. (1994). Episodic ataxia/myokymia syndrome is associated with point mutations in the human potassium channel gene, KCNA 1. Nature Genet. 8, 136-140. Bryant, S. H. (1979). Myotonia in the goat. Ann. NY Acad. Sci. 317, 314-325. Bryant, S. H., and Morales-Aguilera, A. (1971). Chloride conductance of normal and myotonic goat fibres and the action of monocarboxylic aromatic acids. J. Physiol. (London) 219, 367-383. Bryant, S. H., Mambrini, M., and Entrikin, R. K. (1987). Chloride and potassium conductances are decreased in skeletal muscle fibers from the (mto) myotonic mouse. Neuroscience 465, 17. Cannon, S. C., and Corey, D. P. (1993). Loss of Na + channel inactivation by anemone toxin (ATX II) mimics the myotonic state in hyperkalaemic periodic paralysis. J. Physiol. (London) 466, 501-520. Cannon, S. C., and Strittmatter, S. M. (1993). Functional expression of sodium channel mutations identified in families with periodic paralysis. Neuron 10, 317-326. Cannon, S. C., Brown, R. H., Jr., and Corey, D. P. (1991). A sodium channel defect in hyperkalemic periodic paralysis: potassium-induced failure of inactivation. Neuron 6, 619-626. Cannon, S. C., Brown, R. H., Jr., and Corey, D. P. (1993). Theoretical reconstruction of myotonia and paralysis caused by incomplete inactivation of sodium channels. Biophys. J. 65, 270-288. Charlier, C., Singh, N. A., Ryan, S. G., Lewis, T. B., Reus, B. E., Leach, R. J., and Leppert, M. (1998). A pore mutation in a novel KQT-like potassium channel gene in an idiopathic epilepsy family Nature Genet. 18, 53-55. Dulhunty, A. E (1979). Distribution of potassium and chloride permeability over the surface and T-tubule membranes of mammalian skeletal muscle. J. Membr. Biol. 45, 293-310. Fontaine, B., Khurana, T. S., Hoffman, E. P., Bruns, G. A. P., Haines, J. L., Trofatter, J. A., Hanson, M. P., Rich, J., McFarlane, H., Yasek, D. M., Romano, D., Gusella, J. E, and Brown, R. H. (1990). Hyperkalemic periodic paralysis and the adult muscle sodium channel alpha-subunit gene. Science 250, 1000-1002. Furman, R. E., and Barchi, R. L. (1978). The pathophysiology of myotonia produced by aromatic carboxylic acids. Ann. Neurol. 4, 357-365. George, A. L., Jr. (1995). Molecular genetics of ion channel diseases. Kidney Int. 48, 1180-1190. George, A. L., Kamisarof, J., Kallen, R. G., and Barchi, R. L. (1992). Primary structure of adult human skeletal muscle voltage-dependent Na§ channel. Ann. Neurol. 31, 131-137. George, A. L., Crackower, M. A., Abdalla, J. A., Hudson, A. J., and Ebers, G. C. (1993). Molecular basis of Thomsen's disease (autosomal dominant myotonia congenita). Nature Genet. 3, 305-310. Gillard, E. E, Otsu, K., Fujii, J., Duff, C., DeLeon, S., Khanna, V. K., Britt, B. A., Warton, R. G., and MacLennan, D. H. (1992). Polymorphisms and deduced amino acid substitutions in the coding sequence of the ryanodine receptor (RYR1) gene in individuals with malignant hyperthermia. Genomics 13, 1247-1254. Gurnett, C. A., Kahl, S. D., Anderson, R. D., and Campbell, K. P. (1995). Absence of the skeletal muscle sarcolemma chloride channel CLC-1 in myotonic mice. J. Biol. Chem. 270, 9035-9038.
39. Ion Channels as Targets for Disease Gutman, G. A., and Chandy, K. G. (1993). Nomenclature of mammalian voltage-dependent potassium channel genes. Neurosciences 5, 101-106. Hummler, E., and Horisberger, J. D. (1999). Genetic disorders of membrane transport. V. The epithelial sodium channel and its implication in human diseases. Am. J. Physiol. 276, G567-G571. Jansen, G. P., Groenen, P. J., Bachner, D., Jap, P. H., Coerwinkel, M., Oerlman, F., van den Broek, W., Gohlsch, B., Pette, D., Plomp, J. J., Molenaar, P. C., Nederhoff, M. G., van Echteld, C. J., Dekker, M., Berns, A., Hameister, H., and Wieringa, B. (1996). Abnormal myotonic dystrophy protein kinase levels produce only mild myopathy in mice. Nature Genet. 13, 316-324. Jentsch, T. J., Gunther, W., Pusch, M., and Schwappach, B. (1995). Properties of voltage-gated chloride channels of the CLC gene family. J. Physiol. (London) 482, Suppl. P, 19S-25S. Jurkat-Rott, K., Lehmann-Horn, F., Elbaz, A., Heine, R., Gregg, R. G., Hogan, K., Powers, P. A., Lapie, P., Vale-Santos, J. E., Weissenbach, J., and Fontaine, B. (1994). A calcium channel mutation causing hypokalemic periodic paralysis. Human Mol. Genet. 3, 1415-1419. Kallen, R. G., Sheng, Z.-H., Yang, J., Chen, L., Rogart, R. B., and Barchi, R. L. (1990). Primary structure and expression of a sodium channel characteristic of denervated and immature rat skeletal muscle. Neuron 4, 233-242. Keating, M. T., and Sanguinetti, M. C. (1996). Molecular insights into cardiovascular disease. Science 272, 681-685. Kellenberger, S., West, J. W., Catterall, W. A., and Scheuer, T. (1997). Molecular analysis of potential hinge residues in the inactivation gate of brain type IIANa + channels. J. Gen. Physiol. 109, 607-617. Koch, M. C., Steinmeyer, K., Lorenz, C., Ricker, K., Wolf, F., Otto, M., Zoll, B., Lehmann-Horn, E, Grzeschik, K.-H., and Jentsch, T. J. (1992). The skeletal muscle chloride channel in dominant and recessive human myotonia. Science 257, 797-800. Kraner, S. D., Tanaka, J. C., and Barchi, R. L. (1985). Purification and functional reconstitution of the voltage-sensitive sodium channel from rabbit T-tubular membranes. J. Biol. Chem. 25, 6341-6347. Lehmann-Horn, E, and Rudel, R. (1996). Molecular pathophysiology of voltage-gated ion channels. Rev. Physiol. Biochem. Pharmacol. 128, 195-268. Lehmann-Horn, E, Rudel, R., Dengler, R., Lorkovic, H., Haass, A., and Ricker, K. (1981). Membrane defects in paramyotonia congenita with and without myotonia in a warm environment. Muscle Nerve 4, 496-506. Lehmann-Horn, E, Kuther, G., Ricker, K., Grafe, P., Ballanyi, K., and Rudel, R. (1987). Adynamia episodica hereditaria with myotonia: a non-inactivating sodium current and the effect of extracellular pH. Muscle Nerve 10, 363-374. Lerche, H., Heine, R., Pika, U., George, A. L., Jr., Mitrovic, N., Browatzki, M., Weiss, T., Rivet-Bastide, M., Franek, C., Lomonaco, M., Ricker, K., and Lehmann-Horn, E (1993). Human sodium channel myotonia: slowed channel inactivation due to substitutions for a glycine within the Ill-IV linker. J. Physiol. (London) 470, 13-22. Lloyd, S. E., Pearce, S. H. S., Fisher, S. E., Steinmeyer, K., Schwappach, B., Scheinman, S. J., Harding, B., Bolino, A., Devoto, M., Goodyer, P., Rigden, S. P. A., Wrong, O., Jentsch, T. J., Craig, I. W., and Thakker, R. V. (1996). A common molecular basis for three inherited kidney stone diseases. Nature 379, 445-449. Lucke, J. N., Hall, G. M., and Lister, D. (1979). Malignant hyperthermia in the pig and the role of stress. Ann. NY Acad. Sci. 317, 326--337. Ludewig, U., Pusch, M., and Jentsch, T. J. (1996). Two physically distinct pores in the dimeric CLC-0 chloride channel. Nature 383, 340-343. MacLennan, D. H., and Phillips, M. S. (1992). Malignant hyperthermia. Science 256, 789-794. McClatchey, A. I., Van den Bergh, P., Pericak-Vance, M. A., Raskind, W., Verellen, C., McKenna-Yasek, D., Rao, K., Haines, J. L., Bird, T., Brown, R. H., and Gusella, J. E (1992). Temperature-sensitive muta-
671 tions in the III-IV cytoplasmic loop region of the skeletal muscle sodium channel gene in paramyotonia congenita. Cell 68, 769-774. Mehrke, G., Brinkmeier, H., and Jockusch, H. (1988). The myotonic mouse mutant ADR: electrophysiology of the muscle fiber. Muscle Nerve 11, 440--446. Middleton, R. E., Pheasant, D. J., and Miller, C. (1996). Homodimeric architecture of a CLC-type chloride ion channel. Nature 383, 337-340. Ophoff, R. A., Terwindt, G. M., Vergouwe, M. N., van Eijk, R., Oefner, P. J., Hoffman, S. M., Lamerdin, J. E., Mohrenweiser, H. W., Bulman, D. E., Ferrari, M., Haan, J., Lindhout, D., van Ommen, G. J., Hofker, M. H., Ferrari, M. D., and Frants, R. R. (1996). Familial hemiplegic migraine and episodic ataxia type-2 are caused by mutations in the Ca 2+ channel gene CACNL1A4. Cell 87, 543-552 Penn, A. S., Richman, D. P., Ruff, R. L., and Lennon, V. A. (Eds.). (1993). Myasthenia gravis and related disorders. Ann. NYAcad. Sci. 681, 1-622 Ptacek, L. J., Gouw, L., Kwiecinski, H., McManis, P., Mendell, J. R., Barohn, R. J., George, A. L., Barchi, R. L., Robertson, M., and Leppert, M. F. (1993). Sodium channel mutations in paramyotonia congenita and hyperkalemic periodic paralysis. Ann. Neurol. 33, 300-307. Ptacek, L. J., Tawil, R., Griggs, R. C., Engel, A. G., Layzer, R. B., Kwiecinski, H., McManis, P. G., Santiago, L., Moore, M., Fouad, G., Bradley, P., and Leppert, M. E (1994). Dihydropyridine receptor mutations cause hypokalemic periodic paralysis. Cell 77, 863-868. Pusch, M., and Jentsch, T. J. (1994). Molecular physiology of voltagegated chloride channels. Physiol. Rev. 74, 813-828. Reddy, S., Smith, D. B., Rich, M. M., Leferovich, J. M., Reilly, P., Davis, B. M., Tran, K., Rayburn, H., Bronson, R., Cros, D., BaliceGordon, R. J., and Housman, D. (1996). Mice lacking the myotonic dystrophy protein kinase develop a late onset progressive myopathy. Nat. Genet. 13, 325-335. Renaud, J. F., Desnuelle, C., Schmid-Antomarchi, H., Hugues, M., Serratrice, G., and Lazdunski, M. (1986). Expression of apamin receptor in muscles of patients with myotonic muscular dystrophy. Nature. 319, 678-680. Roden, D. M., George, A. L., Jr., and Bennett, P. B. (1995). Recent advances in understanding the molecular mechanisms of the long QT syndrome. J. Cardiovasc. Electrophysiol. 6, 1023-1031. Rogart, R. B., Cribbs, L. L., Muglia, L. K., Kephart, D. D., and Kaiser, M. W. (1989). Molecular cloning of a putative tetrodotoxin-resistant rat heart Na + channel isoform. Proc. Natl. Acad. Sci. USA 86, 8170-8174. Rudel, R., and Lehmann-Horn, E (1985). Membrane changes in cells from myotonia patients. Physiol. Rev. 65, 310-356. Sanguinetti, M. C., and Jurkiewicz, N. K. (1990). Two components of cardiac delayed rectifier K + current. Differential sensitivity to block by class III antiarrhythmic agents. J. Gen. Physiol. 96, 195-215. Shiang, R., Ryan, S. G., Zhu, Y. Z., Hahn, A. E, O'Connell, R, and Wasmuth, J. J. (1993). Mutations in the c~-subunit of the inhibitory glycine receptor cause the dominant neurological disorder, hyperekplexia. Nature Genet. 5, 351-358. Singh, N. A., Charlier, C., Stauffer, D., DuPont, B. R., Leach, R. J., Melis, R., Ronen, G. M., Bjerre, I., Quattlebaum, T., Murphy, J. V., McHarg, M. L., Gagnon, D., Rosales, T. O., Peiffer, A., Anderson, V. E., and Leppert, M. (1998). A novel potassium channel gene, KCNQ2, is mutated in an inherited epilepsy of newborns. Nature Genet. 18, 25-29. Sinha, S., Newson-Davies, J., Mills, K., Byrne, N., Lang, B., and Vincent, A. (1991). Autoimmune aetiology for acquired neuromyotonia (Isaac's Syndrome). Lancet 338, 75-77. Smart, S. L., Lopantsev, V., Zhang, C. L., Robbins, C. A., Wang, H., Chiu, S. Y., Schwartzkroin, R A., Messing, A., and Tempel, B. L., (1998). Deletion of the K(V)1.1 potassium channel causes epilepsy in mice. Neuron. 20, 809-19. Smith, R L., Baukrowitz, T., and Yellin, G. (1996). The inward rectification mechanism of the HERG cardiac channel. Nature 379, 833-836. Steinmeyer, K., Lorenz, C., Pusch, M., Koch, M. C., and Jentsch, T J. (1994). Multimeric structure of CLC-1 chloride channel revealed by mutations in dominant myotonia congenita (Thomsen). EMBO J. 13, 737-743.
672
SECTION IV Ion Channels as Targets for Toxins, Drugs, and Genetic Diseases
Steinmeyer, K., Klocke, R., Ortland, C., Gronemeier, M., Jockusch, H., Grunder, S., and Jentsch, T. (1991a). Inactivation of muscle chloride channel by transposon insertion in myotonic mice. Nature 354, 304-308. Steinmeyer, K., Ortland, C., and Jentsch, T. (1991 b). Primary structure and functional expression of a developmentally regulated skeletal muscle chloride channel. Nature 354, 301-304. Tanabe, T., Beam, K. G., Powell, J. A., and Numa, S. (1988). Restoration of excitation-contraction coupling and slow calcium current in dysgenic muscle by dihydropyridine receptor complementary DNA. Nature 336, 134-139. Tsien, R. W., Ellinor, E T., and Home, W. A. (1991). Molecular diversity of voltage-dependent calcium channels. Trends Pharmacol. Sci. 12, 349-354. Vaughan, E C., and French, A. S. (1989). Non-ligand activated chloride channels of skeletal muscle and epithelia. Prog. Biophys. Mol. Biol. 54, 59-79.
Wallace, R. H., Wang, D. W., Singh, R., Scheffer, I. E., George, A. L., Jr., Phillips, H. A., Saar, K., Reis, A., Johnson, E. W., Sutherland, G. R., Berkovic, S. F., and Mulley, J. C. (1998). Febrile seizures and generalized epilepsy associated with a mutation in the Na+-channel/31 subunit gene SCN1B. Nature Genet. 19, 366-70. Warner, J. O. (1992). Immunology of cystic fibrosis. Br. Med. Bull. 48, 893-911. Zerr, P., Adelman, J. P., and Maylie, J. (1998). Episodic ataxia mutations in Kvl.1 alter potassium channel function by dominant negative effects or haploinsufficiency. J. Neurosci. 18, 2842-2848. Zhang, C. L., Messing, A., and Chiu, S. Y. (1999). Specific alteration of spontaneous GABAergic inhibition in cerebellar Purkinje cells in mice lacking the potassium channel Kvl.1. J. Neurosci. 19, 2852-2864.
Gary L. Westbrook
Ligand-Gated Ion Channels
I. I n t r o d u c t i o n Ion channels are fundamental signaling molecules in virtually all cells. Although channels are present on intracellular membranes of organelles, our current understanding of ion channel function originated with studies of ion channels in the plasmalemma of excitable cells. In the nervous system, voltage-gated ion channels mediate action potentials (APs) and trigger transmitter release, whereas ligand-gated channels are responsible for chemical signaling mediated by classical fast-acting neurotransmitters. Neurotransmitters also trigger slower synaptic responses that are mediated by G-protein-coupled receptors, as discussed in Chapter 41. Until the 1980s, knowledge about the molecular properties of ion channel proteins was largely inferred from physiological and biophysical studies (reviewed in Hille, 1992). In retrospect, many of the predictions from these biophysical studies have been confirmed in dramatic fashion with the elucidation of the amino acid sequence of many voltage- and ligand-gated ion channels by molecular cloning. However, the increasing knowledge of the molecular structure has also revealed a number of unexpected findings and has begun to permit sophisticated correlations of protein structure with function (see, e.g., Miller, 1989; Montal, 1990; Unwin, 1993). These advances have led to an explosive increase in our understanding of the molecular operation of this important class of membrane proteins. These interesting molecules have captured the interest of many outstanding scientists whose studies of ion channels have led to several Nobel Prizes (see, e.g., Neher, 1992; Sakmann, 1992). Studies of ion channels were pioneered by studies of the voltage-gated ion channels in squid axon. These studies established that a conductance change was associated with the AP, and that this conductance change could be attributed to selective increases in the membrane permeability to Na § and K § ions, with the energy for the process derived from the transmembrane ion gradients created by the Na+,K § ATPase. Both the Na § and K § conductances increased with membrane depolarization. These findings suggested that
Cell Physiology Sourcebook: A Molecular Approach, Third Edition
6 7 5
there were discrete pores in the membrane that accounted for the Na + and K § conductance. However it was not until the 1970s that the existence of ion channels was confirmed, first by measurements of the statistical properties of currents through populations of ion channels, a technique called fluctuation or noise analysis. Later, this was refined by the introduction of patch-clamp recording, which allowed measurement of current through single ion channels (for review, see Neher, 1992; Sakmann, 1992). Patch-clamp recording revolutionized the study of ion channels, because it allowed detailed biophysical studies of the electrical activity of single molecules. Finally, protein purification and subsequent molecular cloning have confirmed that ion channels are a large and heterogeneous family of membrane proteins. Although the ligand-gated ion channels vary in their structural features, they share a common basic structure in that they are composed of multiple subunits arranged around a central water-filled pore. Each subunit is a polypeptide encoded by a separate gene. Generally, there are a large number of possible subunit combinations; thus the particular subunits expressed in a certain class of neurons can result in channels with distinct functional characteristics. This chapter is divided into three topic areas. In the first, the categories of ligand-gated ion channels are briefly reviewed. The basic physiological properties and general molecular structure of the ligand-gated channels are then discussed using examples from the muscle nicotinic acetylcholine receptor (AChR) and the N-methyl-D-aspartate (NMDA)-type glutamate receptor. In the third section, the features of neuronal ligand-gated channels that are responsible for fast synaptic transmission are reviewed with an emphasis on the distinctive characteristics of each class of channel. Channels gated by acetylcholine (ACh), 3'aminobutyric acid (GABA), glycine, and glutamate are included as representative examples. Related topics on ion channels are covered extensively in other chapters in Sections III through V of this book. The reader is also referred to a large number of excellent monographs and reviews listed in the references for further details and original citations in this rapidly advancing field.
Copyright 9 2001 by Academic Press. All rights of reproduction in any form reserved.
676
SECTION V Synaptic Transmission and Sensory Transduction
I!. Classes of Ligand-Gated Ion Channels Ligands that activate ion channels in nerve cell membranes can be divided into two major categories, neurotransmitters and intracellular ligands, as listed in Table 1. The best studied are neurotransmitters that are involved in fast chemical synaptic transmission. These include ACh, which is the transmitter at the vertebrate neuromuscular junction and in autonomic ganglia, and the amino acids, Lglutamate and GABA, which mediate the majority of fast excitatory and inhibitory synaptic transmission, respectively, in the vertebrate central nervous system (Betz, 1990). Each of these ligands also activates receptors that are coupled to second messengers via GTP-binding (G) proteins. Gprotein-coupled receptors generally mediate slower and neuromodulatory transmembrane signaling (reviewed in Nicoll et al., 1990). Other neurotransmitters that activate ligand-gated ion channels in various cell types include serotonin, glycine, histamine, and adenosine triphosphate (ATP). For example, the 5HT 3 receptor activated by serotonin is a ligand-gated ion channel (reviewed in Julius, 1991). ATP is released from nerve terminals in several pathways and activates an ATP-gated channel in some spinal neurons that process incoming sensory information (reviewed in Bean, 1992; Burnstock, 1999). In addition to neurotransmitter-gated ion channels, increasing attention has focused on ligand-gated ion channels that are activated by intracellular ligands. These ligands include Ca 2+, cyclic nucleotides, ATE and inositol trisphosphate (IP3). Intracellular ligands can increase (Ca 2§ cAMP, IP3) or decrease (cGMP, ATP) activity of the associated channel. These channels play important roles in cell function. For example, Ca2+-dependent potassium, chloride, and nonspecific cation channels modify the excitability of the nerve cell membrane and thus can alter the AP and release of transmitter from nerve terminals. The cyclic nucleotide (cAMP,
TABLE 1
Ligand-Gated Ion Channels Ion
Ligand Neurotransmitters Acetylcholine Glutamate GABA
Glycine Serotonin ATP Intracellular ligands Calcium Cyclic nucleotides ATP IP 3
Receptor
selectivity a
Muscle, neuronal AChRs AMPA, kainate, NMDA GABA A, GABA c GlyR 5HT 3 receptor ATP receptor
NS NS CI C1 NS NS
Calcium-dependent channels cGMP and cAMP receptors ATP-dependent channel Calcium release channel
K, C1, NS NS K Ca
aAbbreviations for ion selectivity are NS (nonselective cation permeability, some with calcium permeability); C1 (permeable to chloride); K (permeable to potassium); and Ca (permeable to calcium).
cGMP)-gated channels mediate sensory transduction in the visual and olfactory system. Light results in the closing of cGMP channels in vertebrate photoreceptors; odors trigger the opening of cAMP-gated channels in olfactory receptor neurons. See Chapters 47 and 49 for more discussion of channels gated by cyclic nucleotides.
!!!. Basic Physiological Features Modern patch-clamp techniques allow investigators to define the three fundamental properties of any ion channel-conductance, selective permeability, and gatingmin molecular terms. In addition to these fundamental properties, modulation, channel block, and desensitization are processes that affect the activity of ligand-gated ion channels, and are important in shaping the impact of channel activity on membrane excitability. The biophysical basis of these properties will be discussed first, and then the structural features of ligand-gated channels that control these properties will be considered. For a more detailed discussion, see Hille (1992). Conductance reflects the flux of charged ions through the channel, and is measured in picoSiemens (pS). Most ligand-gated ion channels have a conductance of 5-50 pS, which corresponds to the movement of more than 1 x 106 ions per second through the channel pore. These high flux rates initially suggested that ligand-gated receptors contained a water-filled pore, because this rate is much higher than that predicted for other transport mechanisms, such as pumps or exchangers. The size of the single-channel conductance (y) is characteristic for a given channel, although many ligand-gated channels also have subconductance levels (states). An example of a glutamate-activated channel in a hippocampal neuron with several subconductance states is shown in Fig. 1. Note that the conductance of the open channel is usually near 50 pS, but drops occasionally to 8, 35, 40, and 45 pS. The conductance of a channel increases as the permeant ion concentration is raised, but eventually saturates at concentrations well above physiological levels. This behavior can be described by Michaelis-Menten kinetics, suggesting that ions bind to sites within the pore rather than simply obeying the laws of free diffusion. Most permeant ions have a low K m of approximately 100 mM (Km is the concentration at which the conductance is half-maximal, as defined by the Michaelis-Menten equation); thus the ions bind for only a microsecond or so before continuing through the pore. The current carried by the opening of a single-channel is measured in picoamperes (pA) with the single-channel current given by i = y ( V - Veq)
(1)
where V is the membrane potential and Veq is the equilibrium potential, at which no net current is measured. This equation is simply Ohm's law, where ( V - Veq) is the electrochemical driving force for ions that can pass through the channel. For example, if the extracellular and cytoplasmic concentrations of permeant ions are equal, then Veq is 0 mV. Thus, as the membrane potential changes, the size of the sin-
40. Ligand-Gated Ion Channels
677
/ ii_ l// ~^_
=.A.
A =
,~..*
.. A / ~ , ,
I~ . . . .
A.,_
9
2 pA 5 ms
_1 L r _-~'~ ,/~M ' 9
~-
9 -
40pS 50
pS
F I G U R E | . Ligand-gated channels can open to multiple conductance levels. Examples of NMDA-type glutamate channel recording in an outside-out membrane patch obtained from a cultured hippocampal neuron. Downward deflections indicate opening of the channel in the presence of ligand (in this case, 20 laM NMDA). The channel usually opens to the 50 pS level, but may switch to lower levels of 8, 35, 40 or 45 pS before closing to the baseline level. [Modified with permission from Jahr and Stevens (1987). Nature 325, 522-525. Copyright 1987 by Macmillan Magazines Limited.[
gle-channel current also changes; a plot of this relationship is called a current-voltage or I - V plot. For channels such as the muscle AChR, that are permeable to Na + and K +, the reversal potential is 0 mV and the I - V curve is nearly linear. However some channels deviate from linear behavior and show inward or outward rectification (that is, the channel passes current in one direction better than in the other), analogous to the behavior of some voltage-gated K + channels. Channels are not equally permeable to all ions, that is, they exhibit selective permeability. As listed in Table 1, most ligand-gated ion channels are permeable to either monovalent cations or anions. The selective permeability of channels is partly due to the physical size of the pore. For example, the muscle AChR is permeable to cations with diameters up to about 6.5 A. In general, cationic ligand-gated ion channels, are less selective than voltagegated ion channels, which are highly selective for Na +, K +, or Ca 2+, but more selective than gap junction channels, which are permeable to small molecular weight molecules such as cyclic nucleotides as well as to ions. Because anions and cations are approximately the same size, it is obvious that the pore dimensions cannot totally account for selective permeability. As discussed later, positively and negatively charged amino acid residues in the entrance to the channel and within the pore provide a means for this discrimination. Selective permeability is extremely impor-
tant in considering the function of an ion channel. For example, channels that are permeable to Na + and K + do not significantly alter the ion concentrations on either side of the membrane, but instead provide an electrical signal that depolarizes the neuron and brings the membrane potential closer to the threshold for AP generation. However, channels with significant Ca 2+ permeability can transiently increase the cytoplasmic Ca 2+ concentration, and thus act as a biochemical signal. Thus the relative Ca 2+ permeability of ligand-gated channels such as the NMDA receptor and some nicotinic receptors is an important aspect of their role in neuronal function. The reader is referred to the monograph by Hille (1992) for a full discussion of channel permeability and its measurement. Channel gating refers to the conformational change in the ion channel that is triggered by ligand binding. The conformational change results in a rapid switch between conducting (open) and nonconducting (closed) states of the channel. The gating behavior of channels has many parallels with the allosteric behavior of enzymes, with the binding of ligand providing the free energy necessary to maintain the channel in the open conformation. Gating between the open and closed configuration is extremely rapid (approximately 10 las), and thus is beyond the resolution of standard recording methods. Each channel opening usually lasts a few milliseconds, but can, on some occasions, last up to several hundred milliseconds. Binding of more than one agonist molecule is usually necessary to open a channel. The binding of multiple agonist molecules creates a sigmoidal doseresponse relationship, characteristic of the cooperative binding of substrates to enzymes. The steep activation created by sigmoidal activation kinetics prevents the channel from opening in the presence of a low concentration of agonist. This may be quite important in preventing desensitization of receptors at synapses. In general, the higher the binding affinity of the agonist, the longer the channel will remain open. This is because the channel can remain open (or reopen) until the agonist dissociates. The gating of ligand-gated ion channels can be described by multistate kinetic diagrams as shown in Fig. 2A. Such state diagrams have been developed for many ligand-gated ion channels, based on the analysis of open and closed time distributions obtained in single channel recording. For example, the presence of two exponentials in an open time histogram implies the existence of two open states. The minimal kinetic model for a ligand-gated channel usually has at least four states, including closed and unbound, closed and bound, open and bound, and nonconducting (desensitized). For the case shown in Fig. 2A and B, note that the channel opens only when two agonist molecules are bound. For fast-acting neurotransmitters, the concentration of transmitter in the synaptic cleft is typically very high for a brief period. At high concentrations of agonist, the receptors quickly reach the fully liganded but closed state, this transition is determined by the product of the binding rate, r b and the concentration of agonist. Thus the rising and falling phase of the synaptic response largely reflect the unbinding rate r u, the channel opening and closing rates (/3 and a) and, in some cases, the rates in and out of the desensitized state, R d. Classical pharmacological terms can be interpreted in terms of these kinetic
678
SECTION V Synaptic Transmission and Sensory Transduction
schemes (see, e.g., Colquhoun, 1998). Agonists open the channel, while antagonists bind, but apparently do not cause a sufficient conformational change in the channel protein to open the channel. The fraction of time the bound channel spends in the open state is called the open probability, and may differ between agonists--an agonist that activates the channel with higher probability therefore has a higher efficacy in the nomenclature of classical receptor pharmacology. Earlier studies of the activity of ligand-gated channels were generally limited to equilibrium measurements due to the limits in the speed of application of agonists to cells or cell-free patches. Such equilibrium measurements are difficult to interpret in terms of state diagrams such as that shown in Fig. 2. In addition, the duration of fast-acting transmitters such as glutamate is generally brief, suggesting that the synapse itself operates under nonequilibrium conditions. However, rapid solution exchange techniques now allow agonists to be applied within several hundred microseconds to membrane patches, usually in the "outside-out" configuration to allow agonist in the bath to access the extracellular face of the receptor. Many investigators have used these methods to analyze the gating of ligand-gated channels. Such approaches also appear to more closely resemble conditions at synapses and are particularly useful in evaluating the role of channel modulation, channel block, and desensitization on synaptic responses. An example of the response of a membrane patch
A
2rb rb 13 2 A + R ~-~ r. A + A R - -2~--~. ~ A i ~ " a A2R* A2Rd
fj
A2R* A2R AR - R
Receptor state
Channel activity l
J
-closed open
Time
FIGURE 2. Kinetics of ligand-gated channels. (A) Example of a four-state kinetic scheme used to describe the behavior of ligand-gated ion channels. In this example, two agonist molecules (A) must bind before the channel can open (A2R*)or enter a nonconducting desensitized state (A2Ra). The rate constants for agonist binding and unbinding are designated rb and r, and the rate constants for channel opening and closing are designated as/3 and a. (B) Relationship of the state of the receptor to the open and closing of the channel. The channel opens only when the receptor is in the A2R* state as indicated by the arrows.
from a hippocampal neuron to a brief application of glutamate is shown in Fig. 3. In general, the conductance and selective permeability of ligand-gated channels are not subject to regulation. However a number of allosteric control mechanisms can profoundly affect the gating (for review, see Changeux and Edelstein, 1998). These mechanisms can be categorized as desensitiza-
tion, channel block, and allosteric modulation. Desensitization refers to the loss of the response during the continued presence of agonist, and was first noted in studies of the muscle AChR. At the single-channel level, desensitization reflects a nonconducting state of the channel. This phenomenon has now been seen for most ligand-gated channels, and can be described by including an extra bound, but nonconducting, state in the kinetic scheme (see Fig. 2A). Depending on the channel, the nonconducting state may be accessible from either the closed state (AzR) or the open state (AzR*). Desensitization is thought to represent a distinct conformation of the receptor, but this has yet to be demonstrated at the molecular level. Very rapid desensitization may play a role in terminating the neuronal response to neurotransmitters at some synapses, whereas slower desensitization may actually prolong synaptic responses as channels reenter open states after transiting through desensitized states (see, e.g., Jones and Westbrook, 1996). Ligand-gated channels can also be plugged by ions or drugs, a phenomenon referred to as channel block. If the blocking particle is charged, such as a large ion, the block will be influenced by the voltage across the membrane. Thus the reduction in channel activity will be more pronounced at some membrane potentials. The voltage-dependent block of NMDA-type glutamate channels by extracellular Mg 2§ is an important example of this phenomenon (see Ascher and Nowak, 1987). At the single channel level, channel block is usually seen as rapid interruptions ("flickers") of the open state as illustrated in Fig. 4A and B. A characteristic feature of most forms of channel block is that the blocker is only effective when the channel is open (see, e.g., Pascual and Karlin, 1998). Thus the binding site for the blocker is thought to be within the pore of the channel, or at least at a site that is only available in the open conformation of the channel. Channel block is the mechanism of action for a number of drugs, including psychoactive compounds such as phencyclidine (PCP). Channel gating is also subject to modulation by either covalent modifications such as phosphorylation, or via the noncovalent binding of modulators to the channel (e.g., Swope et al., 1999; Changeux et al., 1990). Most ligandgated channels undergo phosphorylation of intracellular portions of the channel protein; however, the resulting effect on channel function is dependent on the specific channel and the type of kinase involved. Phosphate groups are highly charged, and thus can modify interactions in local regions. For example, phosphorylation of the AChR enhances desensitization, and may also be important in clustering of AChR molecules at sites of innervation. The list of noncovalent modulators is long and includes Ca 2§ toxins, drugs, and some endogenous substances such as protons and steroid hormones. These modulators can usually act in a noncompetitive manner, that is, they do not directly interfere with
40. Ligand-Gated Ion Channels
679
Hippocampal Neuron
Patch pipette with outside-out membrane patch containing glutamate channels
Fast application of glutamate to patch ]
|
piezoelectric
: man ulato
Electrical activity ofsingle glutamate channeis
,
F I G U R E 3. Use of rapid application methods to study kinetics of ligand-gated channels. Use of the patch-clamp technique allows small membrane patches to be removed from a cell such as a hippocampal neuron with the extracellular side of the membrane facing the bath (outside-out configuration). The tip of the pipette is then placed in the interface of two rapidly moving streams of solution. A piezoelectric manipulator moves the solutions a few microns to change the solution flow over the patch. Such methods allow solution exchange in less than 1 ms. Thus the response to short applications of ligand, similar as is thought to occur at synapses, can be recorded as openings of single channels. Note, in the response at the right, the opening of single glutamate channels of the NMDA type greatly outlasts the brief application of agonist. The duration of the glutamate application is indicated by the bar above the channel activity.
ligand binding. Most of these reagents affect gating by altering the rate of channel opening, or the time spent in the open state. For example, glycine binds to the NMDA channel, and results in an increased probability of opening (Fig. 4C and D). This is a dramatic example of allosteric regulation because the channel will not open unless both glycine and the transmitter (glutamate) are bound (for review, see McBain and Mayer, 1994). A less profound, but equally important, example is the upregulation of GABA A receptor activity by the benzodiazepines (Olsen et al., 1991).
IV. Molecular Structure The AChR from skeletal muscle is the prototypic ligandgated ion channel, and its physiological properties have been extensively studied during the past 35 years (Changeux et al., 1990; Karlin, 1991). The AChR also led the way in determining the molecular structure of ligand-gated channels. This analysis took advantage of the high density of AChRs in the electric organ of the ray (Torpedo californica), and the high-affinity snake toxin, cebungarotoxin, that binds to muscle-type AChRs. Protein pu-
rification studies revealed an approximately 250-kDa complex with two c~-bungarotoxin binding sites per complex, and separation on denaturing gels revealed polypeptides of 40, 50, 60, and 65 kDa designated as the ~x,/3, % and 6 subunits. Because of the molecular weight of the complex, and the binding of toxin molecules to the c~ subunit, a pentameric structure with two c~ subunits and single copies of /3, % and ~ was proposed. Using the partial amino acid sequence of the purified subunits, the cDNAs encoding AChRs in the Torpedo electric organ were cloned and sequenced in the early 1980s. Incorporation of the purified protein complex in lipid bilayers or expression of AChR mRNA in Xenopus oocytes demonstrated channels activated by acetylcholine, confirming the identity of the molecule. Consistent with the proposed combination of subunits in intact receptors, the expression of AChRs in oocytes was much greater when all four subunit mRNAs were included. The c~ subunit was essential in order to obtain responses to acetylcholine. The c~2/33/6 pentamer occurs in embryonic muscle, but later in development the 3/ subunit is replaced by an E subunit. This substitution results in an increase in the single-channel conductance and a decrease in the time the channel is open. A much greater
680
SECTION V Synaptic Transmission and Sensory Transduction
heterogeneity of receptor subunits appears to be involved in the expression of ligand-gated channels at central synapses (Sections V-VII of this chapter). During the past 20 years, great strides have been made in understanding the structural characteristics of the AChR channel. Based on the hydrophobicity of amino acid residues in the primary sequence, a transmembrane topology was proposed that included four putative transmembrane domains. A large N-terminal domain and a short C-terminal domain extend from the extracellular side of the membrane. This configuration predicts two cytoplasmic l o o p s w a short one between M 1 and M2 and a longer loop containing phosphorylation sites between M3 and M4. For the AChR, several residues in the N-terminal domain have been shown to be involved in forming the ACh binding site. The actual binding pocket appears to lie at the interface between the Nterminal domain of an c~ subunit and that of an adjacent subunit. The N-terminal domain may also determine subunit interactions during assembly of the pentamer, and contains carbohydrate residues whose function is largely unknown. The structures of the neuronal nicotinic and GABA/glycine receptor subunits have the same general characteristics (Fig.
A
5A), and have therefore been grouped together as a gene superfamily, perhaps suggesting a common ancestral gene (reviewed in Betz, 1990).The length of the transmembrane domains (approximately 20 residues) is consistent with an c~-helical configuration, although structural studies suggest that the actual configuration of the transmembrane domains is not strictly c~-helical. The M2 domains of the five subunits make a major contribution to the lining of the channel pore as shown schematically in Fig. 5B. The glutamate receptor subunits differ substantially in their primary amino acid sequence (Nakanishi, 1992, Sommer and Seeburg, 1992; Hollmann and Heinemann, 1994). Notable differences include the much longer N-terminal domain (approximately 500 residues) of the glutamate subunits (Fig. 5A). The glutamate subunit transmembrane topology was initially modeled by homology with the AChR receptor family with both the N- and C-terminal domains as extracellular. However, subsequent analysis has revealed marked differences between the transmembrane topology of the AChR family and the glutamate receptor family. For example, the second hydrophobic domain of the glutamate subunits (M2 in the AChR subunits) appears to be a reentrant loop that enters and exits the mem-
C
[lom s
MAGNESIUM
9
GLYCINE
~
FIGURE 4. NMDA-typeglutamate channels illustrate two common mechanisms of channel regulation: channel block and allosteric modulation. (A and B) Channels activated by NMDA in a membrane patch from a hippocampal neuron in the absence of extracellular Mg2+remain open for several milliseconds before closing (A). However in the presence of Mg2+ (B), the openings are interrupted by brief closures due to plugging of the pore by Mg2+ ions. Downward deflections are due to channel opening. (C and D) The frequency of NMDA channel opening is increased in the presence of extracellular glycine (D) compared to glycine-free solutions (C). The stepwise increase in current in part D indicates that at least two channels were present in the membrane patch. Because glycine is now known to be required for NMDA channel opening, the occasional openings in part C probably reflect contamination of the solution with low levels of glycine. [Modified with permission from Ascher, R. and Nowak, L. (1987). Trends Neurosci. 10, 284-288.]
40. Ligand-Gated Ion Channels
FIGURE 5. Schematictopology of ligand-gated channels. (A) The proposed membrane arrangement of polypeptide subunits for the AChR and glutamate receptors, diagrammed from the amino-terminal (NH2) to carboxyl-terminal (COOH) ends. The hydrophobic segments M1-M4 span the membrane with a long cytoplasmic loop between M3 and M4. The region marked CC in the AChR subunits contains cysteines that form disulfide bridges and is involved in ligand binding. Note the much larger extracellular domain of the glutamate receptors. (B) Cross-section through the membrane showing the proposed distribution of 5 subunits around a central pore of the AChR. The M2 domain lines the pore.
brane from the cytoplasmic side without fully traversing the membrane. This "P" loop shows striking homology to the P loop of voltage-gated channels that line the channel pore. A recently discovered prokaryotic glutamate receptor may provide the missing link between the voltage-gated potassium channels and glutamate receptors (Chen et al., 1999). Secondly, the region between the third and fourth hydrophobic domains is extracellular in glutamate subunits whereas it is cytoplasmic in the AChR subunits. Finally the C-terminus of the glutamate subunits is cytoplasmic. These unique features provide important clues to the location of the transmitter binding site(s), the behavior of the channel pore and sites of interaction with other proteins that anchor and/or localize glutamate receptor subunits at synapses (see Section VII of this chapter). The analysis of the primary amino acid sequence gives limited information about the three dimensional structure of the channel. Recently, crystals of AChRs from the Torpedo electric organ have been analyzed with the electron microscope (Unwin, 1993; Miyazawa et al., 1999). The resolution of these studies provides a good representation of the overall shape and features of the channel (Fig. 6). Several features are striking,
681 including the large component of the receptor that protrudes out of the membrane into the extracellular space. In addition, the recent higher-resolution structure (4.6 ]~, Miyazawa et al., 1999) shows that the entrance to the channel (vestibule) is rather long and connected to the putative ligand binding sites by tunnels lined by twisted/3-sheet strands of the contributing subunits. The vestibule is 25 A in diameter, and ion flux may be influenced by electrostatic interactions with residues in the wall of the pore. The gate of the channel is formed by a narrow bridge across the transmembrane segment with narrow openings between the a-helical structure of the channel lining segments. It is also interesting that a cytoplasmic protein remains attached to the base of the channel protein. The attached protein is likely to be the 43-kDa protein that copufifies with AChRs, and is involved in clustering of receptors. Similar proteins linking ligand-gated ion channels to the cytoskeleton may be involved in localizing, anchoring, or influencing the gating of other channels, including the glycine receptor and the NMDA receptor (see, e.g., Sheng, 1996, 1997). The evidence for M2 as the pore lining comes from several approaches. For example, a 23-amino-acid peptide fragment from the M2 region of the Torpedo 6 AChR subunit forms a cation channel when expressed in lipid bilayers. More direct evidence comes from site-directed mutagenesis of conserved residues in this putative membrane-spanning segment. It has been suggested that the M2 region forms an a-helix with 3.6 residues per turn. However, studies using a technique known as scanning cysteine mutagenesis (see Karlin and Akabas, 1995, 1998) suggest that some portions of M2 are in a/3-sheet configuration with the gate located near the intracellular extent of M2. This method has great
FIGURE 6. Cross-section of AChR obtained from crystallized postsynaptic membrane of Torpedo electric organ at 9-A. resolution. Parts of several receptors are shown, with a complete profile of one receptor along the axis of the pore. The dark lines indicate the channel protein, with arrows identifying the pore and a presumptive area ca. 30 ~ above the membrane that may be the site of ACh binding. The pore narrows in the membrane-spanning region. The square contour lines at the cytoplasmic face of the receptor may be the 43-kDa protein that is involved in receptor clustering. [Modified with permission from Unwin (1993). Copyright 1993 Cell Press.]
682
SECTION V Synaptic Transmission and Sensory Transduction
application in structure-function studies of ion channels because substitution of other amino acid residues by cysteines usually does not affect channel properties. The cysteine residue can then form disulfide bonds with applied reagents, allowing investigators to systematically map residues that are exposed in particular conformations of the molecule, such as during channel opening. Rings of negatively charged residues (glutamate or aspartate) at the ends of M2 for the cationic AChR and glutamate receptor subunits are thought to attract permeant cations, because mutation of these residues to neutral or positively charged amino acids results in a decrease in the single-channel conductance. Consistent with this hypothesis, glycine and GABA subunits have positively charged residues (lysine or arginine) near the ends of the M2 region that may attract permeant anions (Fig. 7B). The M2 region was assumed to be membrane-spanning because the amino acid residues are hydrophobic, that is, they have no charged sidechains (except of course for the charged residues at the ends of M2). Several small uncharged residues near the middle of the bilayer appear to line the channel lumen, as they are involved in the binding of some open channel blockers to the AChR. Mutations of the conserved leucine residue (position 12 in Fig. 7B) also cause a profound change in the response to ACh in the neuronal AChR subunit c~7, and is involved in the binding of the noncompetitive antagonist chlorpromazine (Section V of this chapter).
V. Neuronal Acetylcholine Receptor Channels Similar to muscle AChRs, neuronal AChRs (nAChRs) are ligand-gated ion channels that are permeable to Na § and K +, and in some cases Ca 2§ (for review, see Luetje et al., 1990; McGehee and Role, 1995). ACh has long been known to activate channels mediating excitatory synaptic transmission in autonomic ganglia, and at the recurrent collateral synapse of spinal motoneurons onto Renshaw cells (inhibitory interneurons). However, the functional role of nAChRs in the central nervous system has been more difficult to establish despite the well-known effect of nicotine (via cigarette smoke) on human behavior (Dani and Heinemann, 1996). nAChRs differ from prototypic muscle AChRs both pharmacologically and structurally. First, ce-bungarotoxin blocks muscle AChRs, but not most types of nAChRs. However, binding studies revealed ce-bungarotoxin binding sites on neurons whose function was unclear for many years. It is now clear that the neuronal c~-bungarotoxin receptor is a subtype of the nAChRs that has faster kinetics and a higher calcium permeability than other nAChR subtypes. The Ca 2§ permeability of some nAChRs and changes in channel properties during synaptogenesis suggest that various nAChR subtypes play a role in synaptic plasticity and development. The activity of nAChRs may also be modulated by phosphorylation. The kinase activation may be due to the effects
FIGURE 7. The M2 segment lines the channel pore. (A) Schematic of a multiple subunit AChR channel, as suggested by electron micrographs (See Fig. 6) and physiological data. The membrane-spanning region of the pore is lined by the M2 segment of each subunit, with negatively-charged residues at either end of M2. (B) Amino acid sequence of the a 1 subunit of the AChR that forms a cation channel is compared with the c~1 subunit of the GABAA receptor that forms an anion channel. The charged residues are in bold, and conserved small uncharged residues are highlighted by gray bars. Note that the negatively-charged residues (glutamate, Glu) in the AChR subunit are replaced by positively charged residues (arginine, Arg) in the GABAA subunit. These differences may account respectively for the cation and anion selectivity.
683
40. Ligand-Gated Ion Channels
of neuropeptides such as vasoactive intestinal peptide and substance P that are coreleased with acetylcholine at some synapses. Perhaps even more than for other ligand-gated channels, the cloning and characterization of subunit genes has greatly accelerated studies of both the structure and function of nAChRs (Steinbach, 1989; Role, 1992). The structure of nAChRs is comprised of a pentamer of subunits assembled around a central pore. But unlike the muscle AChRs, the nAChR pentamer can consist of only one or two classes of subunits, aptly named a and/3. The subunits with cysteines at residues 192 and 193 (involved in binding of ACh to the a subunit of muscle AChRs) are designated as ce; subunits lacking these cysteines are called/3. To date, the genes for 8 a (a2_9) and 3/3 (/32_4) subunits have been cloned from rat and chick brain cDNA libraries. The stoichiometry of each channel molecule appears to be 2 a and 3/3 subunits, at least for channels comprised of Ce4 and ~32 subunits. Expression of single subunits and combination of subunits in Xenopus oocytes (by injection of subunit mRNA) has begun to narrow the possible combinations of subunits in functional channels. For example, O~2-4 do not express as homo-oligomeric (e.g., five Og2 subunits) channels, whereas ce7, c~8, and OL9 do. On the other hand, as_ 6 do not express in oocytes even in combination with any of the current/3 subunits, although it is of course possible (even likely) that these subunits contribute to functional receptors in vivo. Likewise, combinations of subunits differ in their single-channel properties, ligand affinity, and sensitivity to antagonists such as neuronal bungarotoxin. Homo-oligomeric O/7 channels produce a rapidly desensitizing response to ACh. This subunit is being used to probe the structural substrates of channel properties such as desensitization, ion flux, and selectivity. Homo-oligomeric ce7, a8, and OL9 channels are also sensitive to a-bungarotoxin, suggesting that a 7 (which is the predominant subunit of these three subunits in brain) is the a-bungarotoxin receptor. Recent studies suggest that presynaptic a-bungarotoxin receptors, particularly of the a 7 type with their high calcium permeability, can enhance the release of glutamate and other transmitters. The recently cloned OL9 subunit is unique due to its limited expression in the outer hair cells of the cochlea as well as its mixed muscarinic/nicotinic pharmacology. Although many central neurons have responses to exogenous application of nicotinic agonists, the distribution of nAChR subunit genes revealed by in situ hybridization is wider than might have been predicted based on physiological studies. However, there is a reasonable correlation between pathways with functional nAChRs (e.g., the cholinergic input to the hippocampus and medial habenula from the medial septum and diagonal band), the presence of nicotinic binding sites, and the expression of nAChR subunits. Recent studies of nAChRs in different neuronal cell types has revealed a spectrum of channel properties and pharmacological characteristics, suggesting that individual neurons may express several combinations of nAChR subunits. Thus the simplest scenariomthat each cell type expresses only one combination of nAChR subunitsmis certainly incorrect. For example, in situ hybridization studies of neurons in the medial habenular nucleus have revealed two ce and three /3 subunit genes.
However a 3 f l 4 and a 4 f l 2 appear to be common heteromers in the brain and autonomic ganglia, respectively. As for many classes of ligand-gated channels, the role of specific nAChR subunits is currently being investigated using subunit knockout mice. Interest in the nAChRs has been further fueled by recent studies indicating that mutations in the ce 4 subunit cause a genetic form of human epilepsy.
VI. ~/-Aminobutyric Acid and Glycine Receptor Channels Synaptic inhibition in the central nervous system is mediated by channels gated by GABA and glycine. Virtually all neurons have GABA channels, while the distribution of glycine channels is restricted primarily to the brainstem and spinal cord. These two channels are selectively permeable to anions, principally C1-under physiological conditions. GABA-gated C1- channels are designated as GABA A receptors to distinguish them from the G-protein-coupled GABA B receptor (reviewed by Bormann, 1988). GABA A channels are often localized on proximal dendrites of central neuron, but also are expressed on axon initial segments and distal dendrites. Because the C1- equilibrium potential in many neurons is more negative than the resting potential, the opening of GABA A channels hyperpolarizes the cell, and thus reduces excitability. In addition to the hyperpolarization, the opening of large numbers of GABA A channels lowers the resistance of the membrane, and effectively "shunts" excitation traveling down the dendrite from excitatory synapses on more distal dendritic branches. However, in some neurons, particularly during early development, the C1-equilibrium is more positive than the resting potential resulting in excitatory GABA A responses. Such excitatory responses may act as a development signal in some cases. The behavior of single GABA A and glycine channels can be described by a kinetic scheme similar to that of the AChR, with the binding of two agonist molecules required for channel opening (see Macdonald and Twyman, 1992). Three conductance levels are apparent. These are approximately 19, 30, and 45 pS when the C1- concentration is equal on both sides of the membrane. The 30-pS level is the most frequently observed level for GABA A channels, and 45 pS is generally more common for glycine channels. Analysis of the openings and closings of single GABA A channels suggests that the channel may open briefly following the binding of a single GABA A molecule, and into two longer-lived open states from the doubly liganded configuration. Receptors may close and reenter the longer-lived open states before the agonist dissociates, so-called bursts where short closings interrupt a series of openings. These bursts last tens of milliseconds. Desensitization of GABA A channels results in long closed intervals that are grouped with bursts into dusters lasting up to several hundred milliseconds. These clusters are important in determining the duration of inhibitory postsynaptic potentials at some synapses (see Jones and Westbrook, 1996). By measuring the permeability of anions of different diameters, the pore size has been estimated to be 5.6 A. GABA A channels also become less active during prolonged recording. This process has been called rundown,
684 and appears, at least in part, to be due to channel dephosphorylation, and may also be sensitive to intracellular Ca 2§ These are themes that are common to several ligand-gated ion channels. The drugs that act on GABA A and glycine channels comprise a fascinating assortment of clinically important compounds (Olsen et al., 1991). Because these channels are so important in synaptic inhibition, either enhancement or reduction in their activity can lead to profound changes in brain function, including amnesia (increased GABA A activity) or seizures (decreased GABA A activity). Antagonists for these receptor channels include strychnine, which blocks glycine receptors; bicuculline, which inhibits GABA A channels; and picrotoxin, which inhibits both channel types. The GABA A channel is also the target of sedative-hypnotic drugs, such as the benzodiazepines and barbiturates. Benzodiazepines (BDZ) increase the probability of channel opening, whereas barbiturates appear to act by prolonging long channel openings (bursts). The pharmacology of benzodiazepine modulation of the GABA A receptor is particularly interesting, because compounds can either enhance channel opening (BDZ agonists), reduce channel opening (BDZ inverse agonists), or block the effects of BDZ agonists (BDZ antagonists). GABA A receptor activity is also modulated by alcohol, volatile anesthetics such as isoflurane, and some steroid anesthetics (or their endogenous equivalents, the neurosteroids). Using benzodiazepines and strychnine as selective ligands, GABA A and glycine receptors were purified as several polypeptides, each with molecular weights of approximately 50-60 kDa (reviewed in Betz et al., 1999; DeLorey and Olsen, 1992). The solubilized receptor complex had a molecular weight of approximately 250 kDa, suggesting that, as for the AChR, multiple (four or five) subunits constitute a receptor. Molecular cloning has now identified a series of receptor subunits for both the GABA A and glycine receptor. The glycine subunits include the strychnine-binding subunit (a) and a/3 subunit, with a proposed stoichiometry of a3/32. A larger number of GABA A subunits (seventeen) have been identified, including six a, three/3, three % 6, e, 7r, and two p (see Schofield et al., 1990; Wisden and Seeburg, 1992; Macdonald and Olsen, 1994). The GABA and benzodiazepine binding sites appear to reside at the interface between an a subunit and a/3 or 3' subunit (usually Y2), respectively. Interestingly, the C~6 subunit has a low affinity for BDZ agonists, but still can bind BDZ inverse agonists or antagonists. This observation may explain the occurrence of benzodiazepine-insensitive GABA A receptors in some neurons. As for the nAChRs and glutamate receptors, the large number of GABA A subunits provides a formidable challenge in determining which combinations are the predominant ones in neurons. The combinations of subunits that form functional receptors is just beginning to be explored, but it is clear that different combinations can alter receptor properties. Subunit expression also varies during development and with neuronal cell type. Molecular studies are also beginning to reveal anchoring and regulatory proteins that interact with glycine and GABA A receptors such as gephyrin (reviewed in Fallon, 2000) and GABARAP (Wang et al., 1999).
SECTION V Synaptic Transmission and Sensory Transduction
There is now increasing evidence for a distinct GABAmediated chloride channel, termed the GABA c receptor, in some retinal neurons. This receptor can be activated by several conformationally restricted GABA analogs, including cis-4aminocrotonic acid. The GABA c receptor is bicucullineinsensitive, weakly antagonized by picrotoxin, and not modulated by B DZs, barbiturates, or neurosteroids. These channels show distinct gating properties and conductance compared to GABA A receptors. Because the GABA c receptor shares some features with the Pl subunit, that is highly expressed in retina, Pl is likely to be one of the GABA c receptor subunits.
VII. Glutamate Receptor Channels Although it was known from the early 1950s that L-glutamate was a neuroexcitant, it was not until the 1980s that the role of glutamate-gated ion channels in central synaptic transmission was widely accepted (see Collingridge and Lester, 1989; Westbrook and Jahr, 1989). It is now clear that glutamate receptors mediate a substantial portion of fast excitatory transmission in the brain and spinal cord through the simultaneous activation of two types of ion channels colocalized at excitatory synapses. Characterization of this family of ligand-gated ion channels was initially based on selective activation by the exogenous amino acid ligands: NMDA, kainate, and quisqualate. Because quisqualate also activates a G-protein-coupled receptor, AMPA (a-amino-3hydroxy-5-methyl-4-isoxazole proprionic acid) has replaced quisqualate as one of the prototypic ligands. It soon became apparent that NMDA receptors were distinct from kainate and AMPA receptors, but for some time it was debated whether kainate and AMPA (i.e., non-NMDA) responses were due to the same or different receptors. It is now clear that AMPA and kainate receptors are distinct molecular entities (see below), although the agonist kainate can activate both classes of receptors. Initial studies of native AMPA/kainate receptor channels in neurons demonstrated many of the same features as AChRs. In hippocampal pyramidal neurons, AMPA channel activation leads to brief openings (1-5 ms) of monovalent cation channels that show little or no voltage dependence. However, in some neurons, particularly interneurons, AMPA channels have a higher calcium permeability and are inwardly rectifying. AMPA receptors rapidly desensitize when gated by glutamate, and both agonist dissociation (unbinding) and desensitization (binding, but nonconducting) can contribute to the decay of the synaptic current. The behavior of AMPA channels seems well suited to the fast relay of information that characterizes the fast component of excitatory postsynaptic potentials in the brain. The NMDA channel has been intensively investigated by neuroscientists in recent years, and provides perhaps the best example of the linkage between fundamental properties of ion channels, and the electrical activity of single cells and complex behavioral phenomena (for review see Bekkers and Stevens, 1990; McBain and Mayer, 1994). This linkage is primarily based on three features of NMDA channels: (1) voltage-dependent block of the open channel by extracellular Mg2+; (2) a relatively high permeability to Ca2+; and (3) slow
40. Ligand-Gated Ion Channels
channel kinetics resulting in long-lasting channel activity (see, e.g., Jahr and Lester, 1992). Because Mg 2+ is positively charged, these ions sense the membrane electric field and are drawn into the open channel at negative membrane potentials. Mg 2+ binds in the channel pore and impedes the flow of permeant ions, even though the ligand (glutamate) remains bound. However, as the cell membrane depolarizes during synaptic activity, Mg 2+ falls out of the channel and permeant cations flow through the channel. Thus ion flux through the channel is voltage-dependent. This mechanism differs from other voltage-dependent ion channels such as the sodium or calcium channel, where voltage-dependence results from an intrinsic conformational change in the channel protein. However, the end result is the same; the NMDA channel becomes more effective with depolarization, and thus acts as a positive feedback on synaptic activation. Also important to the cellular function of NMDA receptors is ion permeability. Unlike many AMPA channels, Ca 2+ contributes a substantial fraction of the flux through NMDA channels. Based on measurements of the equilibrium (reversal) potential in different ionic solutions, the permeability ratio of Ca 2+ to Na + has been estimated to be approximately 10. However, the rate of flux of Ca 2+ through the NMDA channel is slower than Na + because Ca 2+ transiently binds with higher affinity to sites within the channel (energy wells). Thus most of the current through open NMDA channels is due to Na +. Nonetheless, the approximately 5-10% of the current that is carried by Ca 2+ concentration is sufficient to act as a biochemical signal for such processes as the induction of long-term potentiation in the hippocampus, an experimental model of associative learning. Excessive activation of glutamate receptors may also cause neuronal damage, presumably through elevations of intracellular Ca 2+ concentration with subsequent activation of proteases and free radical formation. This mechanism, called exeitotoxieity, has been implicated in brain damage due to prolonged seizures, following strokes, and may also contribute to loss of neurons in several degenerative neurological diseases. Because no high-affinity ligands were available for purification of glutamate channel proteins, the structure of this class of ligand-gated channels defied analysis for a number of years. Thus molecular biologists were forced to use the somewhat laborious task of expression cloning to isolate cDNA clones for glutamate receptor subunits. The first AMPA receptor subunit (GluR1) was isolated by this technique in 1989, and the first NMDA receptor clone (NMDAR1) was also isolated by expression cloning in 1991. As would be predicted from analysis of other ligand-gated ion channels, the proposed structure of glutamate channels involves multisubunit complexes surrounding a central pore. Using the initial AMPA and NMDA clones to screen for homologous sequences, approximately 17 related glutamate receptor subunits have now been isolated from rat and mouse cDNA libraries. Surprisingly, the sequence of glutamate channel subunits is not highly homologous to AChR, GABA, or glycine channels. In particular, glutamate channels have a very large extracellular domain constituting approximately 50% of the protein. Likewise, as discussed in Section IV of this chapter, the topology of glutamate subunits is very different from that of the AChR family of receptor channels.
685
The two major extracellular domains of the glutamate subunits, the N-terminus and the loop between the third and fourth hydrophobic domains, are homologous to a group of bacterial proteins that bind amino acids. This homology is also shared by the N-terminus of the single polypeptide metabotropic glutamate receptors. This led to a proposed model for the glutamate binding site based on the crystal structure of the periplasmic binding proteins (O'Hara et al., 1993). Using chimeric receptors composed of domains from two glutamate subunits with distinct pharmacology, the role of these two domains in binding of glutamate was confirmed (Stem-Bach et al., 1994). The analysis of crystals of the extracellular ligand binding domains of a glutamate receptor subunit has provided a more detailed understanding of this homology (Armstrong et al., 1998). On the basis of sequence homology and the characteristics of the expressed receptors, the subunits can be grouped into three categories: four AMPA subunits (GluR1-4 or A-D), five high-affinity kainate subunits (GluR5-7, KA-1, KA-2) and six NMDA subunits (NMDA1, NMDA2A-D, NR3). Other related glutamate subunits (delta 1, delta 2) have been isolated, but their function remains unclear. The delta subunit is particularly intriguing because a mutation in this orphan subunit is responsible for the neurological phenotype of the Lurcher mouse. AMPA receptors have a low affinity for glutamate and kainate, compared to the so-called high-affinity kainate subunits. Some of the subunits can combine to form functional homo-oligomeric receptors, and this has revealed some interesting and curious phenomena. For example, in the M2 region of GluR1-4 (which forms at least part of the channel pore), the genes all code for a glutamine residue at one location. The RNA for GluR2, 5, and 6 is edited, however, in such a way that the expressed protein has an arginine residue at this site. This switch has a profound effect on the behavior of the channel. The current evoked by homo-oligomeric expression of unedited subunits shows marked inward rectification and an increased permeability to Ca 2+. The edited versions have a linear current-voltage relationship and are permeable only to monovalent cations (Na + and K+). Because the synaptic response mediated by AMPA receptors in hippocampal pyramidal cells has a linear current-voltage relationship, it seems likely that most AMPA receptors in these neurons contain one or more edited copies of GluR2. However, in interneurons in cortex and hippocampus as well as in dorsal horn neurons of the spinal cord, AMPA receptors lacking edited GluR2 subunits show inward rectification and increased permeability to Ca 2+. Physiological studies have demonstrated a more restricted distribution of high-affinity kainate responses, including dorsal root ganglion neurons and both presynaptic and postsynaptic localization on some central neurons (for review, Chittajallu et al., 1999). Expression studies and in situ immunohistochemistry suggest that combinations of GluR5 or 6 with KA-2 may account for at least some of the high-affinity kainate responses. Glutamate channels, particularly NMDA channels, are regulated both by phosphorylation and by a variety of allosteric mechanisms. The most dramatic is the action of glycine as a eoagonist that is required for the opening of NMDA channels, and extracellular Mg 2+ that blocks NMDA channels (Fig. 4). It appears that two molecules of glutamate and two
686
SECTION V Synaptic Transmission and Sensory Transduction
molecules of glycine are required to activate an NMDA channel. Because it now appears that glutamate channels are tetramers, it is thought that each subunit probably contributes to agonist binding. Although AMPA channels do not require a coagonist, given the homology between AMPA and NMDA channels, it may be that four molecules of glutamate are also necessary to activate AMPA channels (e.g. see Rosenmund et al., 1998). Other regulator factors that have been shown to affect NMDA channel activity include extracellular protons, Zn 2+, polyamines, redox potential, and intracellular Ca 2+. NMDA receptors can be regulated by protein kinase A and C and tyrosine kinases, whereas AMPA channel activity is upregulated by cAMP-dependent protein kinase. The impact of these regulatory mechanisms on synaptic function is an important area for future investigation. As is the case for all synaptic receptor channels, AChR and glutamate channels are clustered at synaptic sites on muscle and neurons. For AchRs on muscle, the clustering-anchoring involves a complex cascade involving factors released from the neuron as well as a number of protein-protein interactions in the postsynaptic membrane, including the 43-kDa protein rapsyn (for review, see Froehner, 1993). Recent studies have revealed a number of candidate proteins in central neurons that are involved in the clustering and localization of glycine and glutamate receptor channels. Glutamate channels are imbedded in the postsynaptic density that contains receptors, regulatory proteins, and cytoskeletal proteins. Biochemical studies suggest that calmodulin and cytoskeletal proteins can interact with C-terminal domains of several NMDA receptor subunits at domains that are involved in the regulation of NMDA receptor desensitization, suggesting an intriguing link between dynamic regulation of channel activity (e.g. by compartmentalized regulatory proteins) and structural features such as channel anchoring and clustering. A family of cytoskeletal proteins with so-called PDZ domains appears to be involved in interactions with AMPA and NMDA subunits (for review, see Sheng, 1996; Ehlers et al., 1996). Studies of these interactions may be important in synapse development as well as in the dynamic regulation of activity of ligand-gated channels at synapses.
Channel activity is terminated both when the channel closes or when it enters a nonconducting (desensitized) state. The ACh and glutamate receptors are cation channels, whereas GABA and glycine receptors are anion channels. Many of the genes coding for these channels have been cloned. The primary structure of the ACh, GABA and glycine channel subunits, deduced from the primary amino acid sequence, is highly homologous. Thus, this group of cDNAs constitutes a gene superfamily that probably evolved from a common ancestral gene. Glutamate channels comprise a separate but related family of channel subunits. The ligand-binding domain of glutamate channels shares homology with bacterial amino acid binding proteins, whereas the pore shares homology with the pore of voltage-dependent ion channels. Each of the neurotransmitter-gated channels appears to be comprised of a hetero-oligomeric complex, which is comprised of four or five closely related subunits surrounding a central ion pore. For each receptor type there are a number of possible subunit combinations that can form functional channels. AChR subunits have four transmembrane domains with the second transmembrane domain lining the pore. Charged amino acid residues near the mouth of the pore are important in determining the selectivity and conductance of the individual channel types. The glutamate subunits have three transmembrane domains and a reentrant loop domain that forms a portion of the pore. A number of important drugs act by binding to the walls of the channel pore and obstructing the flow of ions through the open channel. Molecular studies of these subunits have revealed more and more detailed structural features of the channel proteins that determine their gating behavior, conductance, and selective ion permeability. Ligand-gated ion channels are also highly regulated by allosteric and covalent modifications. Allosteric regulation constitutes an important mechanism for drug action in the brain. Recent progress in the study of ligand-gated ion channel function and structure has answered many long-standing questions, but has also raised many new and interesting issues concerning the function of this important class of membrane proteins.
Bibliography VIii. Summary Ligand-gated ion channels are membrane proteins that are fundamental signaling molecules in neurons. These molecules are localized in the plasmalemma and on intracellular organelles and can be gated by both intracellular and extracellular ligands. The neurotransmitter-gated ion channels discussed in this chapter mediate fast excitation and inhibition in the nervous system, and they have now been well characterized by physiological and molecular studies. Studies using the technique of voltage and patch-clamp recording have examined the three basic features of an ion channel: gating, conductance, and selective permeability. In general, ligand-gated ion channels can be described by kinetic models involving binding of two or more ligand molecules that induce a conformational change in the protein. As a result, a central, water-filled pore opens and conducts ions at very high rates of up to 107 ions per second.
Armstrong, N., Sun, Y., Chen G. Q., and Gouaux, E. (1998). Structure of a glutamate-receptor ligand-binding core in complex with kainate. Nature 395, 913-917. Ascher, P., and Nowak, L. (1987). Electrophysiological studies of NMDA receptors. Trends Neurosci. 10, 284-288. Bean, B. P. (1992). Pharmacology and electrophysiology of ATP-activated ion channels. Trends Pharmacol. Sci. 13, 87-90. Bekkers, J. M., and Stevens, C. E (1990). Computational implications of NMDA receptor channels. Cold Spring Harbor Symp. Quant. Biol. 55, 131-135. Betz, H. (1990). Ligand-gated ion channels in the brain: the amino acid receptor superfamily. Neuron 5, 383-392. Betz, H., Kuhse, J., Schmieden, V., Laube, B., Kirsch, J., and Harvey, R. J. (1999). Structure and functions of inhibitory and excitatory glycine receptors. Ann. NY Acad. Sci. 868, 667-676. Bormann, J. (1988). Electrophysiology of GABAAand GABABreceptor subtypes. Trends Neurosci. 11, 112-116. Burnstock, G. (1999). Current status of purinergic signalling in the nervous system. Prog. Brain Res. 120, 3-10.
40. Ligand-Gated Ion Channels Changeux, J.-P., and Edelstein, S. J. (1998). Allosteric receptors after 30 years. Neuron 21, 959-980. Changeux, J. P., Benoit, P., Bessis, A., Cartaud, J., Devillers-Thiery, A., Fontaine, B., Galzi, J. L., Klarsfeld, A., Laufer, R., Mulle, C., Nghiem, H. O., Osterlund, M., Piette, J., and Rehav, F. (1990). The acetylcholine receptor: functional architecture and regulation. Adv. Second Messenger Phosphoprotein Res. 24, 15-19. Chen, G. Q., Cui, C., Mayer, M. L., and Gouaux, E. (1999). Functional characterization of a potassium-selective prokaryotic glutamate receptor. Nature 402, 817-821. Chittajallu, R., Braithwaite, S. P., Clarke, V. R., and Henley, J. M. (1999). Kainate receptors: subunits, synaptic localization and function. Trends Pharmacol. Sci. 20, 26-35. Collingridge, G. L., and Lester, R. A. (1989). Excitatory amino acid receptors in the vertebrate central nervous system. Pharmacol. Rev. 41, 143-210. Colquhoun, D. (1998). Binding, gating, affinity and efficacy: the interpretation of structure-activity relationships for agonists and of the effects of mutating receptors. Brit. J. Pharmacol. 125, 923-947. Dani, J. A., and Heinemann, S. (1996). Molecular and cellular aspects of nicotine abuse. Neuron 16, 905-908. DeLorey, T. M., and Olsen, R. W. (1992). T-Aminobutyric acid A receptor structure and function. J. Biol. Chem. 267, 16747-16750. Ehlers, M. D., Mammen, A. L., Lau, L. E, and Huganir, R. L. (1996). Synaptic targeting of glutamate receptors. Curr. Opin. Cell. Biol. 8, 484-489. Fallon, J. R. (2000). Building inhibitory synapses: exchange factors getting into the act? Nat. Neurosci. 3, 5-6. Froehner, S. (1993). Regulation of ion channel distribution at synapses. Annu. Rev. Neurosci. 16, 347-368. Hille, B. (1992). "Ionic Channels of Excitable Membranes." Second edition, Sinauer Associates Inc., Sunderland, MA. Hollmann, M., and Heinemann, S. (1994). Cloned glutamate receptors. Annu. Rev. Neurosci. 17, 31-108. Jahr, C. E., and Lester, R. A. J. (1992). Synaptic excitation mediated by glutamate-gated ion channels. Curr. Opin. Neurobiol. 2, 270-274. Jones, M. V. and Westbrook, G. L. (1996). The impact of receptor desensitization on fast synaptic transmission. Trends Neurosci. 19, 96-101. Julius, D. (1991). Molecular biology of serotonin receptors. Annu. Rev. Neurosci. 14, 335-360. Karlin, A. (1991). Explorations of the nicotinic acetylcholine receptor. Harvey Lectures 85, 71-107. Karlin, A., and Akabas, M. H. (1995). Toward a structural basis for the function of nicotinic acetylcholine receptors and their cousins. Neuron. 15, 1231-1244. Karlin, A. and Akabas, M. H. (1998). Substituted-cysteine accessibility method. Methods Enzymol. 293, 123-145. Luetje, C. W., Patrick, J., and Seguela, P. (1990). Nicotine receptors in the mammalian brain. FASEB J. 4, 2753-2760. Macdonald, R. L., and Olsen, R. W. (1994). GABA A receptor channels. Annu. Rev. Neurosci. 17, 569-602. Macdonald, R. L., and Twyman, R. E. (1992). Kinetic properties and regulation of GABA A receptor channels. Ion Channels 3, 315-343. McBain, C., and Mayer, M. (1994). N-methyl-D-aspartic acid receptor structure and function. Physiol. Rev. 74, 723-759. McGehee, D. S., and Role, L. W. (1995). Physiological diversity of nicotinic acetylcholine receptors expressed by vertebrate neurons. Annu. Rev. Physiol. 57, 521-546.
687 Miller, C. (1989). Genetic manipulation of ion channels: a new approach to structure and mechanism. Neuron 2, 1195-1205. Miyazawa, A., Fujiyoshi, Y., Stowell, M., and Unwin, N. (1999). Nicotinic acetylcholine receptor at 4.6 A resolution: transverse tunnels in the channel wall. J. Mol. Biol. 288, 765-786. Montal, M. (1990). Molecular anatomy and molecular design of channel proteins. FASEB J. 4, 2623-2635. Nakanishi, S. (1992). Molecular diversity of glutamate receptors and implications for brain function. Science 258, 597-603. Neher, E. (1992). Ion channels for communication between and within cells. Science 256, 498-502. Nicoll, R. A., Malenka, R. C., and Kauer, J. A. (1990). Functional comparison of neurotransmitter receptor subtypes in mammalian central nervous system. Physiol. Rev. 70, 513-565. O'Hara, P. J., Sheppard, P. O., Thogersen, H., Venezia, D., Haldeman, B. A., McGrane, V., Houamed, K. M., Thomsen, C., Gilbert, T. L., and Mulvihill, E. R. (1993). The ligand-binding domain in metabotropic glutamate receptors is related to bacterial periplasmic binding proteins. Neuron 11, 41-52. Olsen, R. W., Sapp, D. M., Bureau, M. H., Turner, D. M., and Kokka, N. (1991). Allosteric actions of central nervous system depressants including anesthetics on subtypes of the inhibitory gammaaminobutyric acid A receptor-chloride channel complex. Ann. NY Acad. Sci. 625, 145-154. Pascual, J. M., and Karlin, A. (1998). Delimiting the binding site for quaternary ammonium lidocaine derivatives in the acetylcholine receptor channel. J. Gen. Physiol. 112, 611-621. Role, L. W. (1992). Diversity in primary structure and function of neuronal nicotinic acetylcholine receptor channels. Curr. Opin. Neurobiol. 2, 254-262. Rosenmund, C., Stem-Bach, Y., and Stevens, C. E (1998). The tetrameric structure of a glutamate receptor channel. Science 280, 1596-1599. Sakmann, B. (1992). Nobel Lecture. Elementary steps in synaptic transmission revealed by currents through single ion channels. Neuron 8, 613-629. Schofield, P. R., Shivers, B. D., and Seeburg, P. H. (1990). The role of receptor subtype diversity in the CNS. Trends Neurosci. 13, 8-11. Sheng, M. (1996). PDZs and receptor/channel clustering: rounding up the latest suspects. Neuron 17, 575-578. Sheng, M. (1997). Excitatory synapses. Glutamate receptors put in their place. Nature 386, 221-223. Sommer, B., and Seeburg, P. H. (1992). Glutamate receptor channels: novel properties and new clones. Trends Pharmacol. Sci. 13, 291-296. Steinbach, J. H. (1989). Structural and functional diversity in vertebrate skeletal muscle nicotinic acetylcholine receptors. Annu. Rev. Physiol. 51, 353-365. Stem-Back, Y., Bettler, B., Hartley, M., Sheppard, P. O., O'Hara, P. J., and Heineman, S. E (1994). Agonist selectivity of glutamate receptors is specified by two domains structurally related to bacterial amino acid-binding proteins. Neuron 13, 1345-1357. Swope, S. L., Moss, S. J., Raymond, L. A., and Huganir, R. L. (1999). Regulation of ligand-gated ion channels by protein phosphorylation. Adv. Second Messenger Phosphoprotein Res. 33, 49-78. Unwin, N. (1993). Neurotransmitter action: opening of ligand-gated ion channels. Neuron 10, Suppl., 31-41. Wang, H., Bedford, E K., Brandon, N. J., Moss, S. J., and Olsen, R. W. (1999). GABA(A)-receptor-associated protein links GABA(A) receptors and the cytoskeleton. Nature 397, 69-72. Westbrook, G. L., and Jahr, C. E. (1989). Glutamate receptors in excitatory neurotransmission. Sem. Neurosci. 1, 103-114. Wisden, W., and Seeburg, P. H. (1992). GABAa receptor channels: from subunits to functional entities. Curr. Opin. Neurobiol. 2, 263-269.
lanusz B. Suszkiw
Synaptic Transmission
i. Introduction
Gap junctions and electrotonic transmission are discussed in greater detail elsewhere in this volume.
The function of nerve cells is to receive, process, and transmit information. Intercellular transfer of signals among neurons or from neurons to effector cells is accomplished at specialized intercellular junctions called synapses and is by and large chemically mediated. A characteristic feature of a chemical synapses is the presence of a 20-100 nm wide extracellular space, the synaptic cleft, which separates the presynaptic (transmitting) and postsynaptic (receiving) elements of the synapse. This physical discontinuity prevents efficient transfer of current between the cells and necessitates intervention of a diffusible chemical transmitter substance. The fundamental aspect of chemically mediated transmission is the transduction of voltage signal into the release of neurotransmitter from the presynaptic neuron and transduction of transmitter binding to a specific receptor into voltage change in the postsynaptic cell. The basic principles of chemical neurotransmission have been elaborated in a series of seminal experiments conducted on the vertebrate skeletal neuromuscular junction by B. Katz and his collaborators in the 1950s and 1960s (Katz, 1966). At about the same time, the pioneering work by J. C. Eccles and his coworkers on spinal motor neurons showed that the general principles of chemical transmission established at the vertebrate neuromuscular junction also apply to central synapses (Eccles, 1964). Since then, new insights into the mechanism of synaptic transmission have been gained on the molecular level through the application of contemporary techniques of molecular biology and electrophysiology. This chapter focuses on chemical synaptic transmission; however, the reader should be cognizant of another form of signal transfer, referred to as electrical transmission. Electrical transmission is mediated by the gap junctions (Bennett, 1997). Gap junctions consist of channel aggregates that bridge the closely apposed cell membranes and thus provide electrical continuity between the communicating cells. Unlike the chemical synapses, gap junctions and the electrical coupling they mediate are not unique to the nervous system but are also found in many other tissues. Cell Physiology Sourcebook: A Molecular Approach, Third Edition
689
II. Structure and Function of Chemical Synapses: An Overview Based on their morphofunctional characteristics, synaptic junctions may be classified as directed, fast-acting or nondirected, slow-acting synapses. Although the configurations of synaptic junctions can vary considerably, all share common features (Fig. 1). The presynaptic element, usually an axon terminal bouton or axon varicosity, is characterized by the presence of membrane-bound spherical organelles, the synaptic vesicles, which store the transmitter prior to its release. A subset of synaptic vesicles clusters near the presynaptic active zone, a structural specialization of the presynaptic membrane for synaptic vesicle docking and exocytosis. The juxtaposed postsynaptic membrane contains specific receptors for the transmitter released from the presynaptic terminal. A thickening of postsynaptic membrane, the postsynaptic density, is frequently observed and thought to play a role in localizing the neurotransmitter receptors at the synaptic membrane. The width of the synaptic cleft at highly directed (point-to-point), fast-acting synapses typically varies from about 20-50 nm but can be considerably wider at nondirected, slow-acting synapses. The synaptic cleft contains proteinaceous material, including various cell adhesion molecules (Hall and Sanes, 1993), thought to serve as an adherent for stabilizing the synapse. The junction is enveloped by glial elements, which insulate the synapse as well as perform other functions such as, for example, transmitter removal from the synaptic cleft. For a detailed discussion of synaptic ultrastructure, the student may wish to consult the articles in Pappas and Purpura (1972) and a more recent review by Burns and Augustine (1995). Transmission at chemical synapses is unidirectional from the presynaptic to postsynaptic element. The presynaptic terminals are specialized for transmitter biosynthesis, Copyright 9 2001 by Academic Press. All rights of reproduction in any form reserved.
690
SECTION V Synaptic Transmission and Sensory Transduction
F I G U R E I. Schematic representations of neuro-neuronal and neuro-effector synaptic junctions. (A) A directed neuro-neuronal synapse subserving fast synaptic transmission in the CNS. The presynaptic terminal bouton contains mitochondria (M), clear-cored, about 40 nm in diameter, synaptic vesicles (SV), a subset of which is seen docked at the presynaptic vesicular grid. The receptors are localized in the juxtaposed thickening of postsynaptic membrane, the postsynaptic density (PSD). The presynaptic vesicular grid together with the juxtaposed PSD define the active zone of the synapse. The synaptic cleft is typically about 20 nm wide and contains dense material, with filamentous structures seen sometimes to span the cleft from pre-to-postsynaptic membrane. Glial cells (GC) cap and insulate the synapse. (B) The vertebrate skeletal neuromuscular junction (NMJ), a directed synapse subserving fast transmission from motor neuron to striated muscle. These large terminals are filled with many clearcored, acetylcholine-containing synaptic vesicles. A few dense-cored vesicles can be also seen. A subset of clear-cored synaptic vesicles aggregate at the presynaptic active zone, which is defined by the presence of the presynaptic density. The presynaptic density appears to have the configuration of a bar defined by parallel rows of intramembrane particles, thought to represent the calcium channels, with two rows of vesicles attached on either side of the bar (bottom). The receptors for acetylcholine (AchR) are localized in the crests of the postjunctional folds (PJF) directly opposite the presynaptic active zones. The synaptic cleft at the vertebrate NMJ is typically 50 nm wide. The basal lamina within the synaptic cleft contains ACHE, an enzyme that degrades ACh released into the synaptic cleft. The junction is insulated by Schwann cell (SC) processes. (C) A nondirected, slow-acting synapse between a presynaptic axon varicosity (AV) of a sympathetic neuron and smooth muscle. The varicosity contains large (>60 rim) dense-core, norepinephrine-containing vesicles. This synapse is characterized by the absence of distinct active zone structures and a variable but usually wide (> 100 nm) separation between the presynaptic and postsynaptic elements of the junction.
packaging of transmitter into the synaptic vesicles, and vesicular transmitter exocytosis. The transmitter receptors in the postsynaptic membrane transduce transmitter binding into ionic current which generates the postsynaptic potential (PSP). Postsynaptic potentials at fast synapses are generated by activation of i o n o p h o r i c , or channel-forming, receptors. The synaptic potentials are typically fast in onset and last for only few milliseconds. The generation of the postsynaptic potentials is, however, not instantaneous but rather registers 0.3-0.5 ms after the arrival of the action potential at the presynaptic terminal. This s y n a p t i c delay is a characteristic feature of chemical synapses and reflects, for the most part, the time required for the molecular events associated with
the transmitter release. At nondirected, slow synapses, channels are not directly transmitter-gated but rather are coupled to the receptor via G protein and second messenger systems. These receptors are referred to as m e t a b o t r o p i c . The synaptic potential generated by activation of the metabotropic receptors is slower in onset and longer lasting. The process of chemical neurotransmission involves (1) synthesis and vesicular uptake of neurotransmitter in presynaptic terminal, (2) exocytotic release of vesicular transmitter into the synaptic cleft, (3) diffusion and binding of transmitter to postsynaptic receptors and generation of synaptic potential, and (4) termination of synaptic activity. As indicated above, transmitters are synthesized in the nerve ending
41. SynapticTransmission of the presynaptic neuron and are concentrated and stored in synaptic vesicles. The intravesicular packet or quantum of transmitter that is stored in the vesicles is released by Ca 2§ dependent exocytosis. Exocytosis is triggered when an action potential (AP) in the presynaptic neurons invades and depolarizes the nerve terminal, thereby opening the presynaptic voltage-gated Ca 2+ channels and allowing influx of Ca 2+ into the terminal. The resulting transient rise in [Ca2+]i triggers fusion of synaptic vesicles with the plasmalemma and transmitter exocytosis. The requirement for Ca 2§ influx is absolute. In absence of Ca 2+ the presynaptic AP will fail to trigger release. Following its release into the synaptic cleft, transmitter combines with specific receptors in the postsynaptic membrane, causing a change in its permeability to specific ions. Change in membrane permeability to ions gives rise to synaptic current which, depending on ions involved, depolarizes or hyperpolarizes the postsynaptic membrane (Fig. 2). The depolarizing potential is called an excitatory postsynaptic potential (EPSP) because it tends to bring the cell membrane potential toward the threshold for an AP. EPSPs are associated with a change in the postsynaptic membrane permeability to Na + and K § ions and a net influx of positive charges (inward current carried by Na+). The hyperpolarizing potential is called an inhibitory postsynaptic potential (IPSP), because it tends to move or hold the membrane potential away from the threshold, thus decreasing the likelihood of an AP being fired. The IPSPs elicited by activation of ionophoric receptors at fast synapses are associated with an increase in the postsynaptic membrane permeability to C1-. PSPs associated with activation of metabotropic receptors usually involve change in membrane permeability to K § ions. Transmitter release normally terminates within less than 1 ms. As the presynaptic AP decays and the terminal repolarizes back toward the resting potential, the depolarization-activated calcium channels reclose, Ca 2§ influx ceases, and [Ca2+]i is rapidly lowered to the prestimulus level. Transmitter action on the receptors in the postsynaptic membrane is terminated by diffusion and enzymatic degradation into an ineffective substance or clearance from the synaptic cleft by reuptake into the nerve terminals and/or glial cells. Reuptake and subsequent catabolism is the primary route of transmitter inactivation for most transmitters. The exception is acetylcholine, in which case inactivation is primarily by means of extracellular degradation of ACh to inactive acetate and choline.
III. N e u r o t r a n s m i s s i o n A. Neurotransmitters and Neurotransmitter Receptors Several criteria define a chemical substance as a neurotransmitter. (1) The biosynthetic enzymes for the synthesis of the substance must be present in the identified presynaptic neuron to catalyze the synthesis of transmitter in the nerve terminals. (2) The substance must be released by stimulation of the presynaptic neuron in a Ca2+-dependent
691 1. AP
Vm = 0- C mM VP _ _R~M_P_
~
2. EPSP
~
3. AP
. . . . (depol _ a~rizati_on)__ ~ _
Na §
CE
PR
Vrn
LL
-~-~~'-
ES
=0mY CMP RMP
A_
...... ...... - - ---~/-
- ~- . . . . .
1. AP
~
2. IPSP
=- 3 . I n h i b i t i o n
(hyperpolarization)
FIGURE 2. Fast excitatory versus inhibitory synaptic transmission. At an excitatory (E) synapse, the presynaptic action potential releases transmitters that activate cationic channels, resulting in an inward synaptic current carried by Na§ ions and depolarizing potential, the excitatory postsynaptic potential (EPSP). EPSPs increase the likelihood of an action potential (AP) being fired by the postsynaptic cell. At an inhibitory (I) synapse, the transmitter activates receptor-gated chloride channels and influx of C1- ions, resulting in hyperpolarizing potential, the inhibitory postsynaptic potential (IPSP). The likelihood of an AP is diminished by activation of I-synapses, because IPSPs hold or move the membrane potential of the postsynaptic cell away from the threshold. RMP, resting membrane potential; CMP, critical membrane potential (threshold) at which an AP is triggered. manner. (3) Application of the substance to the postsynaptic cell must mimic the actions of the neurally released substance. (4) A specific mechanism for inactivation of the transmitter substance, such as a selective uptake system in the presynaptic terminals or the presence of degradative enzyme(s), must be demonstrated at the synapse investigated. Amino acid glutamate is the major excitatory neurotransmitter and y-aminobutyric acid (GABA) and glycine serve as inhibitory transmitters in the CNS. Acetylcholine (ACh), dopamine (DA), norepinephrine (NE), epinephrine, serotonin (5-HT), histamine, and ATP are all recognized as neurotransmitters or neurotransmitter candidates. Acetylcholine and norepinephrine are the established transmitters in the peripheral nervous system. In addition there is good evidence that in the vascular smooth muscle, ATP is co-released and acts as a cotransmitter with NE. Neurons may also co-store and co-release various peptide hormones. These may act as cotransmitters or
SECTIONV SynapticTransmissionand SensoryTransduction
692
Glutamate
0 HO-C-C,H-(CH2}2-COOH II
NH 2
GABA
H2N-CH2-CH2-CH2-COOH
Gly
NH2 CH3-CH-COOH
Acetylcholine
0 ~/CH3 CH3-C-O-CH2-CH2- N ~--CH3 CH3 II
HO~cH2-CH2-NH2
Dopamine
HO
Norepinephrine
HO
OH CH2.CH2_NH 2
HO
Epinephrine
OH HO~~-(~H2-CH2-NH-CH 3 HO- ",,~"
E~N/~
Serotonin
H i starnine
F I G U R E 3.
CH2"CH2"NH2
~----~ CH2"CH2"NH2 HN.~ N
Structural formulas of common neurotransmitters.
TABLE 1
Common
Neurotransmitters
Transmission type
and Major Receptor Types
Neurotransmitter
Receptor a
Glutamic acid
Kainate, AMPA, NMDA, mGluR
Gamma-aminobutyric acid (GABA)
GABAA, GABA B, GABA c
Glycine
Gly-R
Cholinergic
Acetylcholine
nAChR, mAChR
Aminergic
Dopamine
D1, D 2
Norepinephrine
al' a2' J~l' J~2
Serotonin
5HT1, 5HT 2, 5HT 3
Amino acidergic
Purinergic
Histamine
H1, H2, H 3
ATP, adenosine
P l, P2x, P2y
alonophoric or channel-forming receptors are indicated in bold. Regular lettering indicates metabotropic or G-protein-coupled receptors.
693
41. Synaptic Transmission
modulate the synaptic actions of conventional transmitters. Common low-molecular-weight transmitter substances and their chemical structures are shown in Fig. 3. Neurotransmitter receptors (Table 1) are integral membrane proteins containing the transmitter recognition site and the transducer site. Ionophoric or channel-forming receptors include the muscle and neural nicotinic acetylcholine receptors; three pharmacologically distinguishable glutamate receptors that are selectively activated by analogs of glutamate--kainate, a-amino-3-hydroxy-5methylisoxazole-4-propionic acid (AMPA), or N-methylD-aspartic acid (NMDA); the 5HT 3 serotonin receptor; and the Pzx purinoceptor for ATE These relatively nonspecific cationic channels permit passage of Na +, K +, and in some cases Ca 2+ ions, but exclude anions. The anionic (C1-) channel-forming receptors include the GABA A and GABA c receptors and the glycine receptors (Gly-R). Metabotropic receptors transduce the transmitter binding into a physiological response via G-protein/second messenger-coupled mechanisms. The synaptic actions of monoamines, dopamine, norepinephrine, epinephrine, serotonin, and histamine are exerted via the G-protein/second messengercoupled mechanisms systems. The G-protein-coupled receptors also include the muscarinic ACh receptors, metabotropic glutamate receptor (mGluR), GABA B, and several receptors for adenosine and ATE Ligand-gated and G-protein-coupled channels are discussed elsewhere in this volume.
B. Biosynthesis, Storage, and Inactivation of Neurotransmitters Conventional, low-molecular-weight neurotransmitters are synthesized locally in the nerve terminals and packaged into synaptic vesicles (SV) prior to release. The local synthesis of transmitter in the nerve terminals assures that transmitter is available for refilling the vesicles after they have released their contents into the synaptic cleft. The vesicular uptake of transmitters is mediated by specific transporters and is driven by an electrochemical gradient generated by electrogenic proton pump in the vesicular membrane (McMahon and Nicholls, 1991). Following release, the synaptic neurotransmitter action is terminated by diffusion, reuptake into presynaptic terminals or glial cells, and degradation. Reuptake of transmitters from the extracellular space is energetically coupled to a transmembrane Na + gradient and is mediated by specific high affinity transporters that are distinct from the vesicular transporters. Two
FIGURE 4.
families of plasmalemmal transporters have been identified. The carriers for excitatory amino acids require the presence of extracellular Na +, whereas the carriers for biogenic amines, GABA, and glycine require both Na + and C1- for uptake (Fig. 4). (Amara and Azzaro, 1993; Sonders and Amara, 1996). The key enzyme(s) involved in transmitter synthesis are selectively expressed in the neurons and define their neurotransmitter phenotype. Choline acetyltransferase (CHAT) is the marker enzyme for cholinergic neurons where it catalyzes the synthesis of acetylcholine (ACh) from acetyl coenzyme A (AcCoA) and choline (Ch): ChAT
AcCoA+ Ch
; ACh
Acetyl coenzyme A derives from the tricarboxylic acid (TCA) cycle whereas choline is transported into nerve terminals from the extracellular medium by the sodium-dependent, highaffinity carrier system that is localized in the presynaptic terminals. Transport of choline and synthesis of ACh are tightly coupled and both the rate of choline uptake and ACh synthesis increase during activity, ensuring adequate supply of the transmitter. The vesicular ACh transporter (VAChT) mediates uptake of ACh from the cytosol into synaptic vesicles. The plasmalemmal Ch uptake is selectively inhibited by a chemical hemicholinium-3 (HC-3) whereas vesamicol blocks vesicular uptake of ACh. Either of these drugs will cause depletion of ACh and failure of cholinergic transmission. Following its release, the synaptic action of ACh is terminated by diffusion and acetylcholinesterase (AChE)-catalyzed hydrolysis to inactive acetate and choline. Much of the choline is recaptured by the nerve terminals and reutilized in new ACh synthesis. Catecholaminergic neurons and their synaptic connections comprise dopaminergic, noradrenergic (norepinephrine), and adrenergic (epinephrine) systems. The rate-limiting step in catecholamine biosynthesis is the tyrosine hydroxylase (TH)-catalyzed hydroxylation of tyrosine to L-dihydroxyphenylalanine (L-DOPA), which is then decarboxylated by a nonspecific aromatic L-amino acid decarboxylase (L-AADC) to dopamine. Dopamine is taken up into synaptic vesicles and serves as a transmitter at dopaminergic synapses. In noradrenergic nerve endings, dopamine is further converted to norepinephrine by intravesicularly localized dopamine-/3-hydroxylase (D/3H). Methylation of norepinephrine to epinephrine at adrenergic synapses is catalyzed in cytosol by phenylethanolamineN-methyltransferase (PNMT) in a reaction requiring Sadenosylmethionine as methyl donor:
Neurotransmittertransport systems in vesicular and plasmalemmal membranes. (Based on Amara and Arriza, 1993.)
694
SECTION V Synaptic Transmission and Sensory Transduction
tyrosine ~
TH
in turn exported to the nerve terminals and reutilized in the formation of glutamate:
L-DOPA
L-AADC
L-DOPA
transaminase
; dopamine
ce-KA(TcA) ~
D/3H dopamine
glutamine
~norepinephrine glutamate
PNMT
norepinephrine
~epinephrine
Vesicular monoamine transporters (VMATs) mediate the uptake of catecholamines into synaptic vesicles. Reserpine inhibits the VMATs and thereby depletes vesicular stores of catecholamines. Synaptic actions of catecholamines are terminated by high-affinity neuronal reuptake followed by intracellular degradation or reuptake into synaptic vesicles. The plasmalemmal catecholamine transporters are distinct from VMATs. Plasmalemmal reuptake is insensitive to reserpine but is blocked by psychoactive drugs such as cocaine and amphetamine, which thereby increase extracellular levels of catecholamines. The major catabolic pathway for catecholamines involves the mitochondrial monoamine oxidase (MAO)-catalyzed oxidative deamination to aldehydes followed by their rapid conversion by hydrogenases and reductases to corresponding acids and alcohols. Serotonergic neurons utilize serotonin as the transmitter. Serotonin, an indoleamine, is formed by tryptophan hydroxylase-mediated hydroxylation of tryptophan to 5hydroxytryptophan (5-HTP), followed by L-AADC-mediated decarboxylation of 5-HTP to 5-hydroxytryptamine (5-HT, serotonin): Trp- hydroxylase
tryptophan
glutaminase
~glutamate "-
~ 5-HTP
C. T r a n s m i t t e r
Release
Transmitter release is evoked by arrival of an action potential at the presynaptic terminal. Depolarization of the nerve terminal membrane activates (opens) the plasmalemmal voltage-gated Ca 2§ channels, allowing influx of extracellular Ca 2§ into the terminal. The resultant transient increase of Ca 2§ concentration near the plasmalemmal release sites triggers a cascade of biochemical events that culminate in synaptic vesicle-plasmalemma fusion and exocytosis of transmitter into the synaptic cleft. The process terminates upon nerve terminal repolarization when the Ca 2+ channels deactivate (close), Ca 2+ entry ceases, and intracellular Ca 2+ returns to prestimulus levels.
A _
b~L-- _
L -AADC
%,.,.,~..,_~
~ 5-HT
Uptake of serotonin into synaptic vesicles appears to be mediated by the same or a closely related VMAT that mediates vesicular transport of catecholamines. The synaptic action of serotonin is terminated primarily by neuronal sodiumdependent reuptake, mediated by a transporter that is homologous to the plasmalemmal transporters for catecholamines. Following reuptake, serotonin is degraded by MAO-catalyzed oxidative deamination to 5-hydroxyindole acid aldehyde and 5-hydroxyindole acetic acid. Glutamic acid, y-aminobutyric acid (GABA), and glycine are the major amino acid neurotransmitters in the CNS. Because glutamic acid and glycine are common constituents of amino acid pools found in all cells, their use as neurotransmitters at glutamatergic and glycinergic synapses implies subcompartmentation within the respective nerve terminals. This subcompartment in all likelihood corresponds to synaptic vesicles (SV), which presumably selectively accumulate glutamic acid and glycine from the cytosol and release them during synaptic transmission. The amino acid neurotransmitters are synthesized in nerve terminals from intermediates of the tricarboxylic acid (TCA) cycle. Glutamate is formed by transamination of a-ketoglutaric acid (a-KA) and GABA is formed from glutamate in a reaction catalyzed by glutamic acid dehydrogenase (GAD), a marker enzyme for GABAergic neurons. Inactivation of synaptic actions of glutamate and GABA is through reuptake into nerve terminals as well as the glial cells, where both amino acids are converted to glutamine, which is
GAD ~ GABA
1 mV
50ms
B
50mY 2ms
C 10mY 20ms
FIGURE 5. Intracellular recordings from the frog end plate. (A) Spontaneous, miniature end-plate potentials (MEPPs) recorded from resting (not stimulated) junction. Note that these small depolarizing potentials are less than 1 mV in amplitude and occur randomly. (B) Postsynaptic response to a presynaptic action potential. The initial hump on the recorded waveform is the end-plate potential (EPP) elicited by ACh released by the presynaptic AP. Note that the EPP is a large depolarization (> 40 mV), sufficient to bring the end plate to threshold and trigger a muscle AP. (C) Fluctuations in EPPs when transmitteroutput has been reduced by adding 10 mM Mg2§to the bathing medium. (Adaptedfrom Fatt and Katz, 1952, and del Castillo and Katz, 1954.)
41. SynapticTransmission 0.~~
695 many quantal units released nearly synchronously by a presynaptic A E For detailed discussion of the statistics of transmitter release, the interested reader is referred to Martin (1977). For the present purpose, it suffices to state that transmitter release can be described by a simple statistical expression:
failures
2
40
18
o 30 >
16
"~ 2O
--
m=nP
O
: "~
~ 10
0
1,
i i i i i i i 0 0.5 1.0 Amplitude of spontaneous end-plate potentials (mV)
i 4
2
o
l
-a
o 0'., o'., o'.~ 0'., ,'.o ,'.~ ~'., ,'.~ ,'., ,'.o ~'., ,'., ~'.~ ~'., 3.',
A m p l i t u d e of evoked e n d - p l a t e p o t e n t i a l s (mV)
FIGURE 6. Distribution of EPP amplitudes recorded from mammalian end plate under conditions of reduced transmitter release in high (12.5 mM) Mg 2§ The inset shows a histogram of spontaneous potentials (MEPPs) recorded from resting junction. Note that EPP amplitudes group around multiples of mean MEPPs amplitude, and the number of experimentally observed failures (0 quanta released) and single, double, triple, or more quantal responses, fit the theoretical distribution (solid curve) calculated from the Poisson equation. (From Boyd & Martin, 1956.)
1. Ouantal-Vesicular Hypothesis of Transmitter Release The evolution of the current understanding of the mechanism of transmitter release began with the formulation of the q u a n t a l - v e s i c u l a r hypothesis of transmitter release by Bernard Katz and his collaborators (Fatt and Katz, 1952; delCastillo and Katz, 1954). Katz and colleagues observed that in the resting neuromuscular junction, that is, in the absence of stimulation, nerve terminals spontaneously release ACh, giving rise to small depolarizations that occur at random intervals and average about 0.5 m V in amplitude (Fig. 5A). These small potentials behaved in all respects as miniature replicas of the end-plate potential (EPP) evoked by presynaptic APs (Fig. 5B) and therefore were called miniature end-plate potentials (MEPPs). The crucial insight into the relationship between MEPPs and EPPs was provided by the observation that when evoked release was reduced in low-CaZ+/high-Mg 2+ solutions, the size of EPPs fluctuated in a random manner (Fig. 5C) such that the EPP amplitudes appeared to be made up of integral multiples of the average MEPP amplitude and could be described by Poisson distribution (Fig. 6). Katz concluded that transmitter release is a stochastic process, consisting of random release of multimolecular packets or quanta of ACh each producing a unit response (MEPPs), and that the EPP is a summation of
The parameter m is the average number of quanta released per presynaptic impulse when a large number of trials are performed and is called the quantal content of the E P E The parameter n represents the number of quanta immediately available for release and most likely corresponds to either the population of synaptic vesicles associated with the presynaptic active zones or the number of release sites. P is the probability of any single quantum being released and primarily reflects the probability of a productive CaZ+-dependent vesicle fusion with the plasm a l e m m a as a function of Ca 2+ concentration at the release sites. Reducing the availability of Ca 2+ to enter terminals would reduce P, thus accounting for a reduction of transmitter release in low-CaZ+/high-Mg 2+ media. Conversely, increasing the concentration of Ca 2+ at or near the release sites would tend to enhance the probability of exocytosis, that is, facilitate transmitter release. 2. Essential Role of Ca 2+ in Depolarization-Release Coupling The voltage-gated calcium channels in the presynaptic plasma membrane couple membrane depolarization to transmitter exocytosis. The essential role of Ca 2+ influx in transmitter release was demonstrated by Katz and Miledi
A
B
P 4 m V (, 40 ms
_
j
==-=.--
Ca2+
.,.
~
Mg 2+
FIGURE 7. Calcium is required for transmitter release evoked by depolarization of nerve terminals. In the absence of C a 2+, a depolarizing pulse (p) applied to a presynaptic nerve fails to evoke transmitter release and the postsynaptic response is nearly absent (A, top). Application of a pulse of Ca 2+ to the neuromuscular junction shortly before applying stimulus (p) to the presynaptic nerve enables transmitter release as shown by the appearance of a postsynaptic response (A, bottom). The normal postsynaptic response in CaZ+-containing media (B, top) is blocked by applying a pulse of high Mg2+ prior to the depolarizing pulse (p) (B, bottom). The failure to elicit transmitter release in the presence of high M g 2+ is due to block of C a 2+ influx into the terminals. (Adapted from Katz and Miledi, 1967.)
696
SECTION V Synaptic Transmission and Sensory Transduction
A I" .r
m
B
!
-
20mV
2ms
Presynaptic voltage
ynaptic 150On, PCar e scurrent
C
I
2mV
Postsynaptic potential
F I G U R E 8. Experiments in the squid giant synapse illustrating the relationship between magnitude of presynaptic Ca 2+ current and transmitter release monitored by recording postsynaptic potentials. Graded depolarizations of presynaptic terminals (A) evoke graded inward Ca2+currents (B) that correlate with graded postsynaptic potentials (C), which reflect the amount of transmitter released. (Adapted from Llinas, 1977.)
(1967), who showed that depolarization of presynaptic terminals failed to evoke transmitter release when Ca 2+ was absent or its entry into nerve terminals was prevented by high Mg 2+ concentration in the extracellular medium (Fig. 7). In subsequent experiments carried out at the squid giant synapse, where it is possible to make intracellular recordings from both pre- and postsynaptic cells simultaneously, Miledi (1973) showed that injection of Ca 2+ into presynaptic terminals elicited transmitter release, thus providing direct evidence that a rise in intracellular Ca 2+ alone is sufficient for activation of the release process. Furthermore, using voltage-clamping to control the membrane potential of the presynaptic terminals at the squid giant synapse, Llinas (1977) showed that the quantity of synaptic transmitter released, monitored as the size of the postsynaptic potential, is related to the size of the presynaptic calcium current, which in turn depends on the extent of nerve terminal depolarization (Fig. 8). The presynaptic calcium channels that couple membrane depolarization to transmitter release at fast synapses appear to be strategically localized near the release sites (Robitaille et al., 1990). Opening of these channels results in domains of calcium entry, giving rise to localized subplasmalemmal calcium concentrations that may reach 0.1 mM or higher very rapidly and for a very brief duration (Smith and Augustine, 1988). These subplasmalemmal calcium transients represent more than a thousand-fold increase of [Ca 2§ relative to the resting cytosolic calcium level of about 0.1 ~M and provide a powerful trigger for focal synaptic vesicle exocytosis. The relationship between intracellular calcium concentration and the rate of exocytosis investigated in the terminals of goldfish retinal bipolar neurons indicates half-saturation at about 200 luM and cooperative interactions involving at least four calcium ions in activation of synaptic vesicle exocytosis (Heidelberger et al., 1994). Trans-
mitter release occurs within about 60 txs following calcium entry into presynaptic terminals, indicating that Ca2+-triggered exocytosis involves activation of a preassembled vesicle-plasmalemma docking/fusion complex (Bruns and Jahn, 1995; Sabatini and Regehr, 1996).
3. Exocytosis and Recycling of Synaptic Vesicles The active zones in the presynaptic nerve terminals provide plasmalemmal specializations for synaptic vesicle docking and exocytosis. In ingenious experiments, Heuser and coworkers (1979) employed a specially constructed quick-freeze apparatus that enabled them to capture images of vesicle exocytosis at the active zone of the frog neuromuscular junction (Fig. 9). The comparison of the number of exocytotic pores captured in the quick-freeze experiment correlated with the estimated number of quanta released under similar conditions of stimulation, providing the most direct, structural evidence that exocytosis of synaptic vesicles at the active zone region is the likely mechanism of quantal transmitter release. Synaptic vesicles within the terminal are distributed among the so-called readily available and reserve pools. The readily available pool is thought to correspond to the vesicles docked at the active zone and immediately available for release, whereas the reserve pool comprises the vesicles distributed within the terminal at some distance from the active zone. It is evident that in order to maintain the availability of quanta for release, exocytosis must be accompanied by mobilization of new vesicles from the reserve pool within the terminal to the plasmalemmal release sites. Mobilization of vesicles is thought to involve alteration in vesicle cytoskeleton interactions that are regulated by phospho-dephosphorylation of synaptic vesicle-associated proteins, synapsins. Dephosphorylated synapsin seems to stabilize vesicle-cytoskeletal interactions and inhibit vesicle mobilization, whereas phosphorylation of synapsin by the Ca2+-calmodulin-dependent protein kinase II is thought to promote vesicle mobilization (Llinas et al., 1991; Greengard et al., 1993). The mechanism of vesicle docking and initiation of exocytosis by Ca 2§ has not been yet completely worked out; however, remarkable progress has been made in this direction during the past few years. Studies initially conducted in model systems such as mast cells or chromaffin cells and recently extended to synaptic preparations indicate that secretion is accompanied by a stepwise increase in cell capacitance, consistent with fusion of secretory granules with the cell membrane. Electrical measurements further suggest that the first event in exocytosis may be the formation of a pore that connects vesicle lumen with the extracellular space and may provide a channel for the release of soluble contents into the synaptic cleft and/or promote collapse and complete fusion of vesicle with the plasma membrane (Lindau and Almers, 1995; Matthews, 1996). Beginning with the identification and cloning of synaptic proteins in the early 1990s (rev. in Stidhof, 1995), extraordinarily rapid progress has been made during the last few years in defining the molecular machinery of exocytosis. Current evidence indicates that the synaptic exocytotic apparatus makes use of constitutive membrane fusion ma-
41. Synaptic Transmission
697
FIGURE 9. Synaptic vesicle exocytosis captured by a rapid-freezing technique. (A) P-face of freeze-fractured replica of a stimulated and rapidly frozen nerve terminal showing images of synaptic vesicles caught in the process of exocytosis at the active zones. The junction was activated in presence of 1 mM 4-aminopyridine to enhance the level of transmitter release per single presynaptic impulse. (Adapted from Heuser et al. J. Cell Biol. 81: 275-300, 1979.) (B) Three-dimensional representation of pre- and postsynaptic membranes in relation to the freeze-fracture image above. The freeze-fracture splits membrane bilayers into protoplasmic (P-face) and extracellular (E-face) leaflets of the bilayer. The rows of intramembrane particles seen in P- and Efaces are thought to be the presynaptic Ca2§ channels. The openings in the P-face and their craterlike continuations in the E-face are the exocytotic pores formed when synaptic vesicles fuse with the plasmalemma along the parallel bars of the active zone. (Adapted from Heuser et al. J. Cell Biol. 81: 275-300, 1979.)
chinery that has been placed under control of a calcium sensor. The elements of the constitutive docking-fusion apparatus at synapses are the synaptic vesicle membraneassociated protein (VAMP) synaptobrevin and the plasma m e m b r a n e proteins syntaxin and SNAP-25 (25-kDa, synaptosome-associated protein). The VAMP, syntaxin, and SNAP-25 belong to a class of highly conserved membrane-targeting proteins that serve as SNAP (soluble NSF attachment protein) receptors and hence are referred to as SNAREs (S611ner et al., 1993a, b). The SNAREs have been shown to associate into a stable, 7-S core complex that links the apposed vesicle and plasma membranes through formation of parallel bundles of four interacting a-helices,
with syntaxin and synaptobrevin each contributing one helix and SNAP-25 contributing two helices (Sutton et al., 1998). The evidence for the crucial role of SNARE complexes in exocytosis has been provided by the observation that selective proteolysis of either VAMP, syntaxin, or SNAP-25 by clostridial neurotoxins prevents formation of the complex and inhibits transmitter release (Niemann et al., 1994; Hayashi et al., 1994). Biochemical studies indicate that the core complex is primed by ATP-dependent, SNAP-assisted binding of NSF (N-maleimide-sensitive factor). NSF is an ATPase whose activation is thought to play an important role in fusion by inducing a conformational change in syntaxin and disassembly of the SNARE
698
SECTION V Synaptic Transmission and Sensory Transduction
F I G U R E 10. A simplified model of a vesicle docking-fusion complex. The model incorporates interactions among the proteins of synaptic vesicles (synaptobrevins and synaptotagmins), plasma membrane proteins (syntaxins, SNAP-25, neurexin), and the soluble elements of constitutive fusion machinery (NSF, SNAP). Synaptobrevin, syntaxin, and SNAP-25 form a stable, 7-S core complex to which SNAP and NSF (an ATPase) bind to form a 20-S complex. Activation of NSF ATPase leads to disassembly of the complex, perhaps assisting in the final step of membrane fusion. The complex is thought to form in close proximity to the plasmalemmal calcium channels, with synaptotagmin serving as the key Ca 2§ sensor. The plasma membrane protein neurexin interacts with synaptotagmin and may modulate its interactions with the core complex.
complex (Whiteheart et al., 1994; Hanson et al., 1997). The synaptic vesicle membrane protein synaptotagmin 1, whose cytosolic domains contain two Ca2§ C2 motifs homologous to the C2 regulatory domain of protein kinase C, is believed to serve as the Ca 2+ -sensor of exocytosis (Brose et al., 1992; Shao et al., 1977). Biochemical experiments indicate that Ca 2§ activation of synaptotagmin is associated with simultaneous binding of its C2 domains to the membrane phospholipids and the SNARE complex. The kinetics of synaptotagmin-Ca2§ complex interactions occur on a microsecond time scale consistent with its postulated function as Ca2§ receptor of exocytosis (Davis et al., 1999). Another protein that has been implicated in exocytosis is neurexin (Petrenko et al., 1991), a plasmalemmal protein that provides a target for c~latrotoxin, a component of black widow spider venom that induces massive, Ca2§ transmitter exocytosis. Through its capacity to associate with synaptotagmin, neurexin could somehow modulate interactions between
synaptotagmin and SNAREs. A simplified model of a vesicle docking-fusion complex is illustrated in Fig. 10. Several other synaptic vesicle- and plasma membrane-associated proteins have been identified that are likely to have either accessory or regulatory functions in vesicle exocytosis/endocytosis (Stidhof, 1995); however, their precise function and mechanism of action still remain by and large unresolved. Following the release of transmitter, exocytosis may be terminated in at least two ways. One is simply through reclosure of the pore and fission of vesicles at the active zone. These postexocytotic vesicles may then refill with locally synthesized transmitter and, by virtue of being already positioned close to or at the active zone, release the newly formed transmitter in preference to the reserve pool. This could explain the well-documented phenomenon of the preferential release of newly synthesized transmitter. A more generally accepted idea is that, following fusion and exocytosis, vesicle membrane collapses
699
41. Synaptic Transmission into the presynaptic plasmalemma to be retrieved in a series of transformations starting with clathrin-mediated endocytosis followed by endosomal fusion-budding and reformation of functional vesicles (Fig. 11). It is likely that both types, the rapid exo-endocytosis as well as the slower, clathrin-mediated endocytosis occur, but the extent to which one predominates over the other depends on the rate of presynaptic stimulation and possibly other factors. The evidence for local recycling of synaptic vesicles in the nerve terminals has been provided by the observation that vesicles become labeled with high-molecularweight markers such as horseradish peroxidase (HRP) or dextrans when motor nerve terminals are stimulated with these markers in the extracellular medium (Ceccarelli et al., 1973). More recent evidence for vesicle recycling has been provided by tracking the movement of fluorescentlabeled synaptic vesicles during stimulation of motor nerve terminals at the frog neuromuscular junction (Betz and Bewick, 1992) and in other synaptic preparations (Ryan et al., 1993; Lagnado et al., 1996). Vesicles undergoing local recycling within the nerve terminals are progressively degraded and replaced by vesicles that are formed de n o v o in the cell body and transported to the nerve terminals by fast axoplasmic transport. The halflife of vesicles has been estimated at 7-14 days. Evidently, synaptic vesicles can undergo numerous cycles of transmitter release-reloading before being replaced with new ones.
D. Generation of Postsynaptic Potentials at Fast Synapses The nature of postsynaptic responses is determined by the type of receptor-gated ionic conductances that are activated in the postsynaptic membrane. At excitatory synapses, the transmitters activate receptor-gated channels that conduct cations, principally Na § and K § The net current through the synaptic channels is inward and carried by Na § ions, causing a depolarization of the postsynaptic membrane, or EPSP. The EPSPs generated at the skeletal neuromuscular junction are called end-plate potentials (EPPs) and those at other peripheral synapses, for example at nerve-smooth muscle junctions, are frequently referred to as excitatory junctional potentials, or EJPs. At inhibitory synapses, transmitters activate receptor-gated channels that conduct C1- ions. Influx of chloride ions through the synaptic channels tends to increase the negativity of the cell interior and hyperpolarizes the postsynaptic membrane. The hyperpolarizing postsynaptic potentials are called inhibitory postsynaptic potentials (IPSPs) because they tend to move the membrane potential away from the threshold.
1. Synaptic Current and Synaptic Equilibrium Potential The synaptic current (i s) flowing through a single transmitter-activated channel is determined by the channel
FIGURE 11. Synapticvesicle cycling in nerve endings. Following vesicle docking and fusion, synaptic vesicles can recycle through a short or long pathway. The short pathway (left) involves transient fusion between vesicle and plasma membrane, followed by fission and recharging of the emptied vesicle with new transmitter (T). In the long pathway (fight), vesicle membrane collapses into the plasmalemma to be eventually retrieved and reformed via a sequence of steps, then reloaded with new transmitter.
700
SECTION V Synaptic Transmission and Sensory Transduction
conductance (Ys) and the electrochemical driving force ( V - E s) acting on the ions moving through the channel: is-Ys(V-Es)
(1)
where V and E s are membrane potential and the synaptic equilibrium potential, respectively. In resting synapse, most of the receptor-gated channels are closed and the conductance of the postsynaptic membrane to ions is very low. When transmitter is released by presynaptic impulse, it binds to its receptors in the postsynaptic membrane and opens the associated channels for a short, 1-2 ms duration. The resultant total synaptic current, I s, is a sum of currents through all opened channels: Is-nYs(V-Es)-gs(V-Es)
(2)
where n is the number of active channels and gs is the total synaptic conductance (nYs). At excitatory synapses, the transmitter-activated channels permit the passage of both Na + and K + ions with about equal ease and the excitatory postsynaptic current I s CE) is the sum of Na + and K + currents" IS(E)-INa+IK-gNa(V-ENa)+gK(V-EK)
INa - --IK
(4)
gNa(Vm- ENa )---gK (Vm- EK )
(5)
or The membrane potential Vm at which this occurs is the synaptic equilibrium potential (Es). By solving for Vm, it is seen that E s is a weighted average of sodium and potassium equilibrium potentials:
(6)
E s is the limiting potential to which the postsynaptic membrane can be depolarized during the transmitter action. Any further depolarization beyond this point would result in IK > INa and reversal of net synaptic current from inward to outward direction. Therefore, the E s is also called a reversal
potential (Er). It is evident that activation of receptor-gated C1- channels at inhibitory synapses results in the synaptic current given by Is(1)- g s ( V m - E s )
The relationship between the synaptic current and postsynaptic potential can be analyzed in terms of an equivalent electrical circuit consisting of parallel synaptic and nonsynaptic branches, as is illustrated for an excitatory synapse in Fig. 12. The synaptic branch represents the synaptic receptor-gated conductance (gs) in series with the synaptic battery of the synaptic equilibrium potential E s. The nonsynaptic branch consists of membrane capacitance C m and leakage channels (gm) in series with the battery of the resting membrane potential, E m. During the synaptic action of transmitter at the E-synapses, the synaptic current (I s) flows inward through the synaptic branch and outward through the parallel capacitive and resistive elements of the nonsynaptic branch as I c and/m" The direction of current flow at the I-synapses is a mirror image of that at excitatory synapses. Is(E)--(I c + I m)
(8a)
-Is(t)--(I c + I m)
(8b)
I c - CmdVm/dt
(9)
Im-gm(Wm-Em)
(10)
and
(3)
where gNa and gK are Na + and K + ion conductances, and ENa and E K are the Na + and K + equilibrium potentials, respectively. The direction and magnitude of ion flux is determined by the electrochemical driving forces (Vm-Ei). Because gNa (Vm-ENa) > gK(V-EK), that is, INa > IK' the net synaptic current is always inward and carried by Na + ions, resulting in membrane depolarization. As the membrane potential (Vm) becomes depolarized, the term (Vm- ENa) decreases and (Vm-E K) increases until equilibrium is reached, where the inward Na § current is exactly equal to the outward K § current,
Es-(gNa/(gNa+gK))ENaW(gK/(gK+gNa))EK
2. Relationship Between Synaptic Currents and Postsynaptic Potentials
where
and
At the onset of synaptic action most of the synaptic current flows through the capacitive branch because the outward driving force (V m - E m) on current flow through the nonsynaptic channels (gm) is small. Once the membrane capacitance is discharged (depolarization) or charged (hyperpolarization) to its final value, all synaptic current exits through the leakage channels (gm)" Thus at the peak of synaptic activation, I c = 0 and IS(E)---I m
(11a)
-Is(i)- I m (1 lb) at E- and I-synapses, respectively. Substituting Eqs. 2 and 10 into Eq. 11 and solving for V m yields the expression for the postsynaptic membrane potential at the peak of synaptic activation" W-(gs/(gs"l-gm))Es-l-(gm/(gs+gm))Em
(12)
Equation 12 shows that the value of membrane potential (Vm) at the peak of synaptic activation is a weighted average of E s and E m, where the weighting factors are the relative magnitudes of synaptic (gs) and nonsynaptic (gm) conductances. During peak activation of synaptic channels, gs > gm and the membrane potential will tend toward the E s but the amplitude of PSP (i.e., Vm - Em) will be influenced by gm of the nonsynaptic membrane.
(7)
where the equilibrium potential E s is the chloride equilibrium potential, Ecl. This is the limiting potential to which a synaptic membrane can be hyperpolarized during the action of transmitter at inhibitory synapse.
3. Time Course of PSPs The synaptic potentials are electrotonic potentials" they decay passively as a function of time and distance. The time course of the rising phase of the synaptic potential is deter-
41. Synaptic Transmission
701
FIGURE 12. Equivalent electrical circuit description of excitatory synaptic potential. Four phases of the postsynaptic potential (A) and the corresponding equivalent circuit representations (B) are illustrated. (1). At rest, the transmitter-gated synaptic channels are closed (circuit open; no current flow; dV/dt = 0). (2) Onset of synaptic action (active phase). Synaptic channels are opened by transmitter binding to the receptor (the synaptic switch is closed) and/s flows inward through the synaptic branch (gs) and outward as I c and/m through the nonsynaptic branch. (3) At the peak of synaptic action, capacitance C m has discharged to its final value, I c = 0 and/s = -/m (dV/dt = 0). (4) Passive decay phase. Synaptic channels have reclosed (synaptic switch open) and /s = 0. The current through the nonsynaptic membrane consists of outward/m and inward/c' which recharges the membrane capacitance and repolarizes the membrane back to the prestimulus membrane potential. (Based on Kandel et al. "Principles of Neural Science" (2000). McGraw-Hill, New York. Reproduced with permission of The McGraw-Hill Companies.) mined by both active and passive properties of the membrane. As the synaptic channels spontaneously reclose, the PSP decays in an exponential fashion. The decay of the PSP is purely a passive process whose time course is a function of the membrane time constant, z. The membrane time constant is the time required for an electrotonic potential to decay to 1/e or 37% of its peak value. The time constant of neurons is in the range of 1-20 ms. The amplitude of the synaptic potentials decreases exponentially from the maxim u m recorded focally at the synapse as a function of the membrane length constant, 7. The length constant is the distance at which electrotonic potentials decay to 1/e or 37% of their amplitude at the point of origin. The length constant of dendrites is typically in the range 0.1 to 1 mm.
E. Slow Synaptic Transmission Mediated by G-Protein-Coupled Receptors In contrast to fast synaptic potentials mediated by directly transmitter-gated channels, slow synaptic responses are mediated by a distinct family of proteins that transduce the transmitter binding into cellular responses through activation of GTP-binding regulatory proteins or G proteins. Binding of a transmitter to a specific receptor transforms the associated G protein into an active form that modulates activity of ionic channels at some distance from the receptor either through direct interaction with the channel and/or through second messenger systems. The slow synaptic transmission may involve either increase or
702 decrease in postsynaptic membrane conductance due to a channel opening or channel closure, respectively. For example, activation of muscarinic receptors in heart muscle causes a G-protein-mediated opening of a certain class of voltage-gated K § channels and relatively long-lasting (seconds) membrane hyperpolarization. Similar, G-proteinmodulated potassium channels K(G) are present in central neurons, where they mediate slow IPSPs. In contrast, activation of a muscarinic receptor in sympathetic ganglion neurons and certain neurons in the central nervous system is associated with a second-messenger-mediated closure of K § channels and membrane depolarization. The responses to transmitter activation of G-protein-coupled receptors need not be limited to modulation of ionic channels but also involve modifications of other regulatory proteins resulting in long-term modifications of the postsynaptic cell's physiology. For a more detailed discussion of G proteins and second messengers in slow synaptic actions of neurotransmitters, the reader is referred to Schulman and Hyman (1999).
E Synaptic Integration versus Amplification Central neurons can receive from several dozens to several thousand excitatory (E) and inhibitory (I) synaptic connections converging on the target neuron from a variety of other neurons in the brain. Whether or not a neuron discharges a propagated AP is determined by the number of E-synapses and I-synapses active at any one time. It should be noted that even in absence of IPSPs, activation of a single excitatory synapse would be insufficient to discharge an AP, because individual PSPs generated at central synapses are very small, usually in the range of 0.5-2 mV in amplitude. Therefore, summation of several EPSPs is usually necessary to bring the membrane of the postsynaptic neuron to the threshold potential. The postsynaptic neurons integrate the synaptic potentials by adding the EPSPs and subtracting the IPSPs from the membrane potential at any instant of time. Algebraic summation of two or more topographically separated synapses that are activated nearly simultaneously is called spatial summation. The effectiveness of spatial summation depends on the membrane space constant which, it will be recalled, is the distance at which electrotonic potentials decay to 37% of their amplitude at the point of origin. Clearly, synaptic inputs separated by a distance smaller than the space constant can sum together more effectively than those that are separated by distances larger than the space constant. When a presynaptic neuron fires at a rate such that the interval between successive presynaptic APs is less than the duration of the PSP, each succeeding PSP adds to its predecessor. This process is referred to as temporal summation. Effectiveness of temporal summation depends on the membrane time constant (i.e., the time required for an electrotonic potential to decay to 37% of its peak value). The larger the time constant, the longer is the duration of the PSP and thus the greater the opportunity for summation of successive PSPs to occur. A propagated action potential will be triggered only if the net current is of sufficient magnitude to depolarize the neuronal membrane to the threshold. An action
SECTION V Synaptic Transmission and Sensory Transduction potential is normally triggered at the initial segment of an axon, the region of the neuron with the lowest threshold. Since summation of synaptic currents at the initial segment is the principal determinant of whether or not an AP will be fired, the initial segment is referred to as the integrative zone of the neuron. In contrast to the integrative activity at central synapses, the function of the neuromuscular junction is to transfer without failure the AP from the presynaptic motor neuron to the postsynaptic muscle fiber. In this case, the excitatory postsynaptic potential (EPP) is normally always suprathreshold and sufficient to trigger the muscle AP. It may be noted that because the nerve terminal is very small in diameter compared with the muscle fiber it innervates, even if these two membranes were contiguous, the current generated during the invasion of presynaptic terminal by an AP would be insufficient to depolarize the postsynaptic membrane to threshold due to impedance mismatch between the two membranes; that is, the small nerve terminal cannot provide enough action current to depolarize the large-diameter skeletal muscle fiber much more than about 1 mV. Thus, the function of transmitter at the neuromuscular junction is to amplify the presynaptic signal.
G. Modulation of Synaptic Transmission The efficacy of signal transmission at chemical synapses can be modulated by extrinsic and intrinsic factors, including the pattern of ongoing activity as well as history of previous activity. The efficacy of transmission may be altered by mechanisms that affect the dynamics of presynaptic transmitter release and/or modify the postsynaptic receptormediated events. This modification may be short lasting or may persist for some time. Thus, synaptic modulation provides for fine-tuning of ongoing synaptic activity as well as for longer lasting changes that are likely to play an important role in learning processes.
1. Depression Depression or fatigue of synaptic transmission refers to progressive reduction in the amplitudes of postsynaptic potentials in the course of prolonged, relatively high-frequency activation of presynaptic neurons, reflecting progressive depletion of releasable transmitter stores. Recovery from fatigue may take from minutes to hours.
2. Facilitation Facilitation is a frequency-dependent increase in the amplitude of postsynaptic potentials evoked by closely spaced presynaptic action potentials. The increase in the amplitudes of succeeding PSPs reflects the progressively larger amount of transmitter released with each nerve impulse in the course of stimulation. The mechanism is thought to involve build-up of ionized Ca 2§ within the terminals. That is, when presynaptic neuron is stimulated at certain frequency, the diffusion and clearance of Ca 2+ from the release sites begins to lag, and the residual Ca 2+ adds to the Ca 2+ transient evoked by the next arriving impulse. Build-up of Ca 2+ increases the probability (P) of transmit-
41. Synaptic Transmission
ter quanta being released. Facilitation is a relatively shortlived process that lasts for a few seconds. 3. Post-Tetanic Potentiation When the nerve is activated with relatively prolonged and/or high-frequency (tetanic) bursts of impulses, the quantity of transmitter is increased upon subsequent stimulation, even after a relatively long intervening rest period. This phenomenon is known as post-tetanic potentiation. In contrast to facilitation, post-tetanic potentiation may last for minutes and sometimes hours, suggesting a long-term modification of presynaptic function secondary to increase in cytosolic Ca 2+ levels. 4. Long-Term P o t e n t i a t i o n
Long-term potentiation (LTP) refers to a long-lasting increase in EPSP following tetanic stimulation in the presynaptic neurons. It is distinguished from post-tetanic potentiation (PTP) in that the latter is a strictly presynaptic phenomenon, whereas induction and expression of LTP involves both postsynaptic and presynaptic elements. LTP was first described (Bliss and Lomo, 1973) and analyzed most extensively in the hippocampus. In the CA1 region, the LTP involves a special glutamate receptor subtype, the NMDA receptor. The NMDA receptor-gated channel is normally blocked by Mg 2§ but can be activated by glutamate when the postsynaptic neuron is sufficiently depolarized so that the Mg 2§ blockade of the channel is relieved. The induction of LTP appears to be associated with influx of Ca 2§ and activation of Ca2+-dependent protein kinases CaM kinase II and protein kinase C, and possibly other protein kinases in the postsynaptic dendritic spine, which leads to increased efficacy of synaptic transmission that may last for days to weeks. It is thought that while induction of LTP involves the postsynaptic events, the maintenance of LTP may be associated with long-term increase in the probability of transmitter release, a presynaptic event.
H. Presynaptic Receptors and Transmitter Release Receptors for neurotransmitters are not confined to postsynaptic sites, but are also found on presynaptic nerve terminals. The modulation of transmitter release by presynaptic receptors that respond to transmitter released by another neuron is referred to as heterosynaptie modulation. Heterosynaptic modulation may involve either inhibition or facilitation of transmitter release. In addition, certain presynaptic autoreceptors recognize the cell's own neurotransmitter. In this case, the neuron's transmitter may modulate its own release by interacting with these receptors. This is called automodulation. For example, many cholinergic neuron terminals possess muscarinic autoreceptors, and ACh released from these terminals acts on the autoreceptors to inhibit its own release. Although the physiological role of presynaptic receptors has been a subject of debate, it is evident that they provide a potential mechanism for fine-tuning of transmitter release.
703
IV. Summary The transmission of synaptic signals is mediated by chemical neurotransmitter substances. Neurotransmitters are synthesized in presynaptic terminals and stored in synaptic vesicles. Transmitter release is evoked by presynaptic APs, which activate influx of Ca 2+ into terminals and trigger a CaZ+-dependent exocytosis of transmitter from synaptic vesicles into the synaptic cleft. Once released, neurotransmitters activate specific receptor-gated channels in the postsynaptic cell and elicit a transient change in the membrane permeability to cations or anions. Fast synaptic transmission is mediated by ionophoric receptors. Slow synaptic transmission is mediated by G-protein-coupled receptors. Excitatory postsynaptic potentials (EPSPs) are associated with transmitter-induced increase in Na + and K + conductance of the synaptic membrane, resulting in net entry of positive charge carried by Na + and membrane depolarization. Inhibitory postsynaptic potentials (IPSPs) are associated with transmitter-activated influx of C1- and membrane hyperpolarization. The EPSPs at the skeletal neuromuscular junction are called end-plate potentials (EPPs). In a healthy neuromuscular junction, the EPPs are always large enough to depolarize the muscle membrane to threshold and trigger muscle APs. The EPSPs generated at any single neuro-neuronal synapse are usually too small to depolarize the postsynaptic neuron to threshold. Synaptic signals converging onto a neuron are normally integrated through summation of EPSPs and-IPSPs, and an AP is triggered only when the resultant membrane potential reaches or exceeds the threshold. Chemical synaptic transmission is subject to modulation by intrinsic and extrinsic factors, including frequency and pattern of AP firing, which can either facilitate or depress the transmission across any given synapse.
Bibliography Amara, S. G., and Arriza J. L. (1993). Neurotransmitter transporters: three distinct gene families. Curr. Opin. Neurobiol. 3, 337-344. Bennett, M. V. L. (1997). Gap junctions as electrical synapses. J. Neurocyt. 26, 349-366. Bennett, M. K., and Scheller, R.H. (1993). The molecular machinery for secretion is conserved from yeast to neurons. Proc. Natl. Acad. Sci. USA 90, 2559-2563. Betz, W. J., and Bewick, G. S. (1992). Optical analysis of synaptic vesicle recycling at the frog neuromuscular junction. Science 255, 200-203. Bliss, T. V. P., and Lomo, T. (1973). Long-lasting potentiation of synaptic transmission in the dentate area of the anaesthetized rabbit following stimulation of the perforant path. J. Physiol. 232, 331-356. Boyd, I. A., and Martin, A. R. (1956). The end-plate potential in mammalian muscle. J. Physiol. 132, 30-38. Brose, N., Petrenko, A. G., Stidhof, T. C., and Jahn, R. (1992). Synaptotagmin: a calcium sensor on the synaptic vesicle surface. Science 256, 1021-1025. Bruns, D., and Jahn, R. (1995). Real-time measurement of transmitter release from single synaptic vesicles. Nature 377, 62-65. Bums, M. E., and Augustine, G. J. (1995). Synaptic structure and function: dynamic organization yields architectural precision. Cell 83, 187-194.
704
Ceccarelli, B., Hurlbut, W. R, and Mauro, A. (1973). Turnover of transmitter and synaptic vesicles at the frog neuromuscular junction. J. Cell Biol. 54, 30-38. Davis, A. E, Bai, J., Fasshauer, D., Wolowick, M. J., Lewis, J. L., and Chapman, E. R. (1999). Kinetics of synaptotagmin responses to CaZ+and assembly with the core SNARE complex onto membranes. Neuron 24, 363-376. delCastillo, J., and Katz, B. (1954). Quantal components of the endplate potential. J. Physiol. 124, 560-573. Dreyer, E, Peper, K., Akert, K., Sandri, C., and Moor, H. (1973). Ultrastructure of the "active zone" in the frog neuromuscular junction. Brain Res. 62, 373-380. Eccles, J. C. (1964). "The Physiology of Synapses." Academic Press, Inc, New York. Fatt, E, and Katz, B. (1952). Spontaneous subthreshold activity at motor nerve endings. J. Physiol. 117, 109-128. Greengard, E, Valtorta, E, Czernik, A. J., and Benfenati, E (1993). Synaptic vesicle phosphoproteins and regulation of synaptic function. Science 259, 780-784. Hall, Z. W., and Sanes, J. R. (1993). Synaptic structure and development: the neuromuscular junction. Neuron 72(suppl.), 99-121. Hanson, P. I., Roth, R., Morisaki, H., Jahn, R., and Heuser, J. E. (1997). Structure and conformational changes in NSF and its membrane receptor complexes visualized by quick-freeze/deep etch electron microscopy. Cell 90, 523-535 Hayashi, T., McMahon, H., Yamasaki, S., Binz, T., Hata, Y., Siidhof, T. C., and Niemann, H. (1994). Synaptic vesicle fusion complex: action of clostridial neurotoxins on assembly. EMBO J. 13, 5051-5061. Heidelberger, R., Heinemann, C., Neher, E., and Matthews, G. (1994). Calcium dependence of the rate of exocytosis in a synaptic terminal. Nature 371, 513-515. Heuser, J. E., Reese, T. S., Dennis, M. J., Jan, Y., Jan, L., and Evans, L. (1979). Synaptic vesicle exocytosis captured by quick freezing and correlated with quantal transmitter release. J. Cell Biol. 81, 275-300. Kandel, E. R., and Siegelbaum, S. A. (2000). Signalling at the nervemuscle synapse: directly gated transmission, ch. 11, pp. 187-206. In "Principles of Neural Science." E. R. Kandel, J. H. Schware, and T. M. Jessell (Eds.). McGraw-Hill, New York. Katz, B. (1966). "Nerve, Muscle, and Synapse." McGraw-Hill Book Co., Inc. Katz, B., and Miledi, R. (1967). The timing of calcium action during neuromuscular transmission. J. Physiol. 189, 535-544. Kuttler, S. W., and Nicholls, J. G. (1977). "From Neuron to Brain: A Cellular Approach to the Function of the Nervous System." Sinauer Associates, Inc., Sunderland, MA. Lagnado, L., Gomis, A., and Job, C. (1996). Continuous vesicle cycling in the synaptic terminal of retinal bipolar cells. Neuron 17, 957-967. Lindau, M., and Almers, W. (1995). Structure and function of fusion pores in exocytosis and ectoplasmic membrane fusion. Curr. Opin. Cell Biol. 7, 509-517. Llinas, R. R. (1977). Calcium and transmitter release in squid synapse. In "Approaches to the Cell Biology of Neurons" (W. M. Cowan and J. A. Ferrendelli, Eds.) Society for Neuroscience Symposia, II, pp. 139-169. Society for Neuroscience, Bethesda. Llinas, R. R., Gruner, J. A., Sugimori, M., McGuinness, T. L., and Greengard, P. (1991). Regulation by synapsin I and Ca2+-calmod ulin-dependent protein kinase II of transmitter release in squid giant synapse. J. Physiol. 436, 257-282.
SECTION V Synaptic Transmission and Sensory Transduction
Martin, R. A. (1977). Junctional transmission II. Presynaptic mechanisms. In "Handbook of Physiology. The Nervous System," Vol. 1., ch. 10, pp. 329-355. American Physiological Society, Bethesda, MD. McMahon, H. T., and Nicholls, D. G. (1991). The bioenergetics of neurotransmitter release. Biochim. Biophys. Acta 1059, 243-264. Matthews, G. (1996). Synaptic vesicle exocytosis and endocytosis: capacitance measurements. Curr. Opin. Neurobiol. 6, 358-364. Maycox, E R., Hell, J. W,. and Jahn, R. (1990). Amino acid neurotransmission: spotlight on synaptic vesicles, Trends Neurosci. 13, 83-87. Miledi, R. (1973). Transmitter release induced by injection of calcium ions into nerve terminals. Proc. Royal Soc. 183, 421-425. Niemann, H., Blasi, J., and Jahn, R. (1994). Clostridial neurotoxins: new tools for dissecting exocytosis. Trends Cell Biol. 4, 179-185. Pappas, G. D., and Purpura, D. P. (Eds.). (1972). "Structure and Function of Synapses." Raven Press Publishers, New York. Petrenko, A. G., Perin, M. S., Davletov, B. A., Ushkaryov, Y. A., Geppert, M., and Stidhof, T. C. (1991). Binding of synaptotagmin to the a-latrotoxin receptor. Nature 353, 65-68. Robitaille, R., Adler, E. M., and Charlton, M. P. (1990). Strategic location of calcium channels at transmitter release sites of frog neuromuscular junction. Neuron 5, 773-779. Ryan, T. A., Reuters, H., Wendland, B., Schweizer, E E., and Smith, S. J. (1993). The kinetics of synaptic vesicle recycling measured at single presynaptic boutons. Neuron 11, 713-724. Sabatini, B., and Regehr, W. G. (1996). Timing of neurotransmission at fast synapses in the mammalian brain. Nature 384, 170-172. Schulman, H., and Hyman, S. E. (1999). Intracellular signaling. In "Fundamental Neuroscience" (M. J. Zigmond, E E. Bloom, S. C. Landis, J. L. Roberst, and L. R. Squire, Eds.), pp. 269-316 Academic Press, San Diego. Shao, X., Davletov, B. A., Sutton, R. B., Stidhof, T. C., and Rizo, J. (1997). Bipartite Ca binding motif in C2 domains of synaptotagmin and protein kinase C. Science 273, 248-251. Smith, S. J., and Augustine, G. J. (1988). Calcium ions, active zones and synaptic transmitter release. Trends in Neurosci. 11,458-464. Srllner, T., Bennett, M., Whiteheart, S., Scheller, R., and Rothman, J. (1993a). A protein assembly-disassembly pathway in vitro that may correspond to sequential steps of synaptic vesicle docking, activation and fusion. Cell 75, 4 0 9 4 18. Srllner, T., Whiteheart, S. W., Brunner, M., Erdjument-Bromage, H., Geromanos, S., Tempst, P., and Rothman, J. E. (1993b). SNAP receptors implicated in vesicle targeting and fusion. Nature 362, 318-324. Sonders, M. S., and Amara, S. G. (1996) Channels in transporters. Curr. Opin. Neurobiol. 6, 294-302. Spray, D. C., Scemes, E., and Rozental, R. (1999). Cell-cell communication via gap junctions. In "Fundamental Neuroscience" (M. J. Zigmond, E E. Bloom, S. C. Landis, J. L. Roberts, and L. R. Squire, Eds.), pp. 317-343, Academic Press, San Diego. Stidhof, T. C.(1995). The synaptic vesicle cycle: a cascade of proteinprotein interactions. Nature, 375, 645-653. Sutton, R. B., Fasshauer, D., Jahn, R., and Brunger, A. T. (1998). Crystal structure of a SNARE complex involved in synaptic exocytosis at 2.4 ]k resolution. Nature 395, 347-353. Whiteheart, S. W., Rossnagel, K., Buhrow, S. A., Brunner, M., Jaenicke, R., and Rothman, J. E. (1994). N-ethylmaleimide-sensitive fusion protein: a trimeric ATPase whose hydrolysis of ATP is required for membrane fusion. J. Cell. Biol. 126, 945-954.
Nicole Gallo-Payet and Marcel Daniel Payet
Excitation-Secretion Coupling
I. Introduction The process of excitation-secretion coupling is completely different depending on the peptide/amine or lipid nature of the secretory products. The secretory process for peptide and amine molecules begins with the synthesis, modification, and sorting of the molecules to be secreted. Synthesis occurs in the rough endoplasmic reticulum (RER) and sorting occurs in the Golgi complex. The secretory molecules are packaged in secretory granules or vesicles, which are then transported to the cell periphery before they are released in the extracellular space by fusion with the plasma membrane. This complex process is named exocytosis or reverse pinocytosis and can be operated either by a constitutive or a regulated mechanism. Constitutive secretion is unregulated and closely follows the rate of synthesis of the secretory products. This form of secretion occurs in many cell types, including lymphocytes, hepatocytes, and pancreatic/3 cells. In regulated secretion, fusion of the secretory granules with the plasma membrane is initiated by a specific signal (ligand-receptor coupling), triggered by an increase in cytosolic calcium concentration ([Ca2+]i). In steroid-secreting cells (adrenal cortex, ovary, testis), the process of synthesis begins with cholesterol stored in lipid droplets followed by subsequent steps occurring in mitochondria and smooth endoplasmic reticulum. It is generally assumed that steroids are free to diffuse throughout the aqueous cytoplasm and lipid phase of the plasma membrane. Secretory vesicles are not present and secretion and/or release of steroids is tightly coupled to steroid synthesis. Although the process of synthesis differs, stimulation of secretion of peptide hormones, neurotransmitters, and steroids involves similar cascades of molecular events. After binding to their specific receptors, the stimuli activate second messenger production, several cascades of phosphorylation/dephosphorylation of intracellular proteins, cytoskeleton reorganization, synthesis of new products, and
Cell Physiology Sourcebook:A Molecular Approach, Third Edition
705
release of secretory products. Cytoskeleton and Ca 2+ ion are certainly the most important players involved in this excitation-secretion coupling. However, while disruption of the actin network is necessary to trigger fusion of secretory vesicles with the cell membrane, a well-preserved organization seems important for steroid release.
II. Cellular C o m p o n e n t s Involved in Excitation-Secretion Coupling A. Interaction of Cytoskeletal Structures with Transmembrane Signaling Molecules The cytoskeletal elements are described in Chapter 6. Therefore the brief descriptions given here on microfilaments and microtubules are aimed at understanding the mechanisms involved in excitation-secretion coupling (Schmidt and Hall, 1998). In the living cell, actin filaments, F-actin (consisting of two staggered, parallel rows of monomers, G-actin, noncovalently bound and twisted into a helix), interact with several proteins, such as vinculin, ~-actinin, villin, and fodrin, which cross-link and bundle microfilaments into a well-organized three-dimensional network. They can also cross-link myosin, forming a contractile network, or form a very dense network at the cell periphery, the cell cortex (Fig. 1A and Fig. 2A). This dynamic organization of microfilaments depends on a large and diverse group of actin-binding proteins, including profilin and caldesmon (which bind G-actin), and gelsolin and scinderin (which cap and sever F-actin). On the other hand, polymerization, stabilization, and modulation of microtubules depend on several microtubule-associated proteins (MAPs) that adorn the tubulin-containing core of the tubules. The complete cell cytoskeletal network includes interaction between microfilaments, microtubules, and intermediate filaments (see Figs. 1 and 2 and Chapter 6).
Copyright 9 2001 by Academic Press. All rights of reproduction in any form reserved.
706
SECTION V Synaptic Transmission and Sensory Transduction
F I G U R E 1. Interaction of cytoskeletal structures with transmembrane signaling. In resting cells (A), receptors, G proteins, phospholipase C (or adenylyl cyclase), and phosphoinositides are not linked together, but are associated with microfilaments or actin-associated proteins. For example, profilin interacts with high affinity with eight molecules of PtdlnsP2, protecting them from hydrolysis by phospholipase C; gelsolin is also associated with PtlnsP2; the protein caldesmon is associated with both actin monomers and calmodulin; MARCKS (myristoylated, alanine-rich C kinase substrate) is a specific protein kinase C substrate associated with the cytoplasmic face of the membrane under resting conditions. In its nonphosphorylated form, MARCKS crosslinks actin, favoring a rigid actin meshwork at the membrane level. In activated cells (B), ligand binding to its receptor activates phospholipase C, which hydrolyzes PtdlnsP 2 and leads to InsP 3 and diacylglycerol (DAG). The rise in [Ca2+]i (due to the InsP 3 binding to intracellular pools of Ca 2§ induces F-actin depolymerization. Moreover, calmodulin-caldesmon complex binds to Ca2+, gelsolin binds, severs, and caps actin filaments into 2:1 complexes to the barbed ends. Activated protein kinase C (now at the membrane) phosphorylates MARCKS, which is released from the membrane. All these modifications decrease cell rigidity, making the actin network more plastic.
42. Excitation-Secretion Coupling
F I G U R E 2. Dynamic changes in cytoskeleton during exocytosis. In resting cells ([Ca2+]i = 100 nM) (A), microfilaments interact with several proteins (including fodrin and a-actinin) that cross-link and bundle microfilaments into a well-organized three-dimensional network. In vesicle-secretory cells, microfilaments can also cross-link myosin, forming a contractile network, or form a very dense network at the cell periphery, the cell cortex. Activation of cells by appropriate stimulation (B) is associated with a reduction in the cell cortex rigidity, due to characteristic changes in cytoskeletal properties from "gel" (solid) to "sol" (liquid) states. These changes are primarily under the control of Ca 2§ which cooperates with phospholipid messengers to induce microfilament disruption. An increase in intracellular calcium ([Ca2+]i (= 0.5-1 I~M) induces several cellular modifications. In addition to capping and severing of the actin microfilaments (see Fig. 1), there is dissociation of actin from actin-associated proteins, such as fodrin; patching of fodrin along the plane of the plasma membrane (docking site); and dissociation of caldesmon from secretory vesicles. These events result in a decrease in viscosity, which favors movement of granules toward the plasma membrane releasing sites. Actin-myosin interactions could facilitate granule displacement in cytosol.
707
708 Association of cell surface molecules with the cytoskeleton is widely believed to be one of the earliest consequences of cellular activation for many systems. These cell surface molecules include not only receptors, but also integrins, the receptors of the components of extracellular matrix. Appropriate stimulation of receptors and integrins has immediate and profound effects on the organization and activity of the microfilaments (Aplin et al., 1998). Several studies have shown that microfilament disruption (with cytochalasins) or microtubule disruption (with colchicine or vinblastine) increased exocytosis (i.e., peptide or amine secretion), although steroid hormone secretion is decreased or abolished. In other words, these observations indicate that cytoskeleton is a barrier in the former, but a requirement for the latter. In both cases, microfilaments are active players in the process of excitation-secretion coupling. Several studies indicate that receptors (either G-protein-coupled or tyrosine kinase types); a subunits of heterotrimeric G-proteins (a i, as, C~q); mono-meric G proteins (Rho, Rac, Cdc42); second-messenger-activating enzymes (adenylyl cyclase, phospholipase C (PLC)), Ca 2+, K +, or C1- channels (see Chapter 36); and proteins up or downstream from second messenger production (phosphatases, protein kinase C (PKC), mitogenic associated protein kinase (MAPK)) are associated with microfilaments. For example, anchoring of PKC-e to F-actin is required for glutamate release in nerve endings of neuronal cells (Janmey, 1998). In general, these associations occur mainly via acting-binding proteins, which bind effectors or phosphoinositides through specific domains, called plecktrin homology domains, (PH domains). PH domains are essential for the membrane recruitment of several proteins that contain them and are frequently accompanied by other motifs (Src homology domains SH2 and SH3, and prolinerich, Dbl homology (DH), or GTP exchange domains), indicating that the recruitment of the PH domain protein would nucleate sites on the membrane for protein complex assembly (Inglese et al., 1995; Janmey, 1998; Martin, 1998). In addition, close functional association between G proteins and microtubules has also been extensively described during the past five years (Popova et al., 1997). Such observations indicate that the cytoskeleton operates as a matrix improving the efficiency of the signal transduction cascade and that actin-binding proteins are in large part responsible for these cytoskeleton-membrane receptor interactions.
B. Actin-Binding Proteins Important in Signaling The physical properties of actin networks depend on the length of microfilaments and the architecture of the threedimensional network formed by interaction between actin and several actin-binding proteins. The aim of this section is to point attention to the F-actin cross-linking proteins which may be regulated by Ca 2+ and thus involved in the process of excitation-secretion coupling (Schmidt and Hall, 1998). Profilin was the first actin-binding protein described in signaling. Profilin also contains a phosphatidylinositol bisphosphate (PtdlnsP2) binding site, and binding of PtdlnsP 2 to this site triggers the dissociation of profilin from actin, promoting
SECTIONV Synaptic Transmission and Sensory Transduction actin polymerization (Fig. 1A). PtdlnsP 2 molecules bind to profilin with a stoichiometry of about 8" 1. This 8:1 association of PtdlnsP2:profilin complexes protects PtdlnsP 2 from cleavage by phospholipase C y. In a similar fashion, also in resting conditions, PLCy may also bind profilin. However, when PLCy becomes tyrosine phosphorylated (by appropriate ligand activation), its affinity for PtdinsP 2 increases to a level where profilin protection can be overcome, resulting in PtdlnsP 2 hydrolysis. Hydrolysis of one or two of the eight PtdlnsP 2 molecules bound to each profilin leads to a rapid decrease in PtdinsP 2 affmity and release of the remaining profilin-bound PtdinsP 2 molecules (Fig. 1B). Thus, the binding of profilin to PtdlnsP 2 not only liberates polymerizationcompetent G-actin, but also affects hydrolysis of PtdinsP 2 by PLCy (Forscher, 1989; Janmey, 1998). Gelsolin is a Ca2+-dependent F-actin severing molecule which, along with villin, fragmin, adseverin, and scinderin, is able to sever actin filaments. These proteins bind to actin filaments and bend and cleave them in a Ca2+-dependent manner and afterwards cap the barbed filament ends. Severing actin filaments promotes cell cortex transition from a "gel" state to a "sol" state upon addition of Ca 2§ Gelsolin contains two spatially separate binding sites: a G-actin Ca2§ site in the C-terminal domain and a Ptdlns-sensitive site closer to the N-terminal portion. In resting cells ([Ca2+]i = 100 nM), the bulk of gelsolin is cytoplasmic and a small proportion is PtdlnsP2-associated. Gelsolin has little affinity for actin under these conditions and is in an actin-free state (Fig. 1A). When the cell is activated, there is a transient rise in [Ca2+]i due to InsP 3 action or Ca 2§ influx that then activates gelsolin, causing a 200fold increase in its affinity for F-actin. This results in rapid filament side-binding, followed by severing and capping of any free barbed filament ends, inducing a dramatic disruption of existing actin network structure in the vicinity of Ca 2+ elevation (Fig. 1B). The activities of these severing proteins are inhibited by binding to PtdlnsP 2. In the absence of Ptdlns turnover, much of the PtdlnsP 2 is likely to be tightly associated with profilin and thus unavailable to gelsolin. If gelsolin is activated under these conditions (for example by an increase in [Ca2+]i, independent of the Ptdlns turnover), severing of actin networks is observed, but without subsequent actin reassembly. This leads to two possible modes of gelsolin activation: (1) Calcium influx produced by activation of PtdlnsP-independent pathways (i.e., voltage- or agonist-gated Ca 2+ channels) leads to actin severing and capping only. (2) In contrast, activation of these same Ca 2§ channels concomitant with PtdlnsP turnover results in severing and capping followed by actin polymerization, that is, actin remodelling (Forscher, 1989; Janmey, 1998). In summary, association of actin-binding proteins with phosphoinositides causes opposite effects to those observed upon association with Ca 2+. In the former, such associations (profilin:PtInsP 2) promotes actin polymerization, favoring the cortical actin "gel" near the plasma membrane, while increased Ca 2+ concentration favors association of actin-binding proteins (profilin, gelsolin) with Ca 2+, inducing depolymerization and thus "sol" state of actin filaments.
709
42. Excitation-Secretion Coupling
Caldesmon is a calmodulin-dependent actin-binding protein that, at low Ca 2+ concentrations (100 nM), binds and cross-links actin monomers, inhibiting actin polymerization. Under these conditions, caldesmon interacts reversibly with secretory granules. At greater concentrations of Ca 2+ (~M ranges), CaZ+-calmodulin complex binds to caldesmon and reverses this inhibition. The flip-flop regulation of caldesmon may be important for secretory vesicle function during the changes in intracellular Ca 2+ levels observed upon stimulation (Figs. 1 and 2). MARCKS (myristoylated, alanine-rich C kinase substrate) is a specific protein kinase C substrate that is targeted to the membrane by its amino-terminal binding domain. In resting cells, MARCKS associates with the cytoplasmic face of the membrane. In its nonphosphorylated form, MARCKS cross-links actin, favoring a rigid actin meshwork at the membrane level (Fig. 1A). Activated protein kinase C phosphorylates MARCKS, which remains associated with actin filaments, but can no longer crosslink actin fibers, making the actin network more plastic (Fig. 1B). In addition, an increase in intracellular Ca 2+ concentration promotes binding of calmodulin to MARCKS, inhibiting its actin cross-linking activity, again resulting in a less rigid actin meshwork. Thus, PKC induces a local destabilization of the actin skeleton through the phosphorylation of MARCKS. MARCKS is phosphorylated when synaptosomes are depolarized, suggesting a role in secretion (Arbuzova et al., 1998). Several newly identified proteins, such as the focal adhesion molecule (p125FAK), paxillin, and the small GTPbinding protein Rho are also closely implicated in actin polymerization after hormonal stimulation. Tyrosine phosphorylation of p125 FAK and paxillin and their association with cytoskeleton and/3y subunits of G proteins have been recently identified as early events in the action of several growth factors and G-protein-coupled receptors (such as angiotensin II and vasopressin). However, to date, their role has been ascribed to regulating cell adhesion, motility, or proliferation rather than secretion (Inglese et al., 1995; Tapon and Hall, 1997; Hall, 1998).
C. Interaction between Microtubules and Microtubule-Associated Proteins in Signaling As mentioned earlier, polymerization, stabilization, and plasticity of microtubules depend on several microtubuleassociated proteins (MAPs), which differentially cross-link microtubules (Matus, 1988). Recent studies have shown that some MAPs could be implicated in the secretory process (Gundersen and Cook, 1999). MAPs are substrates for several protein kinases, including a CaZ+-calmodulindependent kinase, a cAMP-dependent kinase, tyrosine kinases, and protein kinase C. Both cAMP (via protein kinase A) and Ca 2+ (via CaZ+-calmodulin kinase) lead to phosphorylation of MAP-2. Moreover, MAP-1 and MAP-2 are responsible for binding secretory granules to microtubules, either in cells from the anterior pituitary gland, the/3 cells of the endocrine pancreas or in synaptic vesicles of neurons. These results strengthen the probable role of MAPs in secretory processes.
D. Actin-Binding, Docking and Fusion Proteins of the Secretory Granules Granule vesicles for peptides or amine products not only allow secretory tissues to store large amounts of secretory products in a relatively small volume, but also protect this material from intracellular degradation while providing a very efficient means for transporting and releasing fixed quantities of secretory material. 1. Actin-binding Proteins Associated with Secretory Granules Induction of exocytosis is associated with a loss in the cell cortex rigidity, due to characteristic changes in cytoskeletal properties from "gel" to "sol" states. These changes are primarily under the control of Ca 2+, which cooperates with phospholipid messengers to induce microfilament disruption (Fig. 1B and 2B) (Tchakarov et al., 1998). Moreover, several pieces of evidence indicate that microfilaments, rather than microtubules or intermediate filaments, are responsible for the mechanical changes initiated by Ca 2+. Chromaffin cells of the adrenal medulla and mast cells from the immune system synthesize and, along with neuronal synaptic vesicles, store and secrete large amounts of neurotransmitters or neuropeptides. These cells contain numerous electron-dense secretory granules, which discharge their contents into the extracellular space by exocytosis. The subplasmalemmal area is characterized by the presence of a highly organized cytoskeletal network. F-actin seems to be exclusively localized in this area and, together with specific actin-binding proteins, forms a dense viscoelastic gel. Fodrin, vinculin, a-actinin, and caldesmon, four actinbinding proteins, as well as gelsolin and scinderin, two actin-severing proteins, are found in the plasmalemmal region. Moreover, fodrin, caldesmon, and ~-actinin binding sites also exist on secretory granule membranes, indicating that actin filaments can also link to secretory granules (Fig. 2A). Chromaffin granules can be entrapped in this subplasmalemmal lattice, and thus the cytoskeleton acts as a barrier preventing exocytosis (Burgoyne, 1995; Burgoyne and Morgan, 1998a). Synapsin I, a phosphoprotein and a substrate for protein kinase A and calmodulin-dependent protein kinase II (CaM kinase II), is associated with synaptic vesicles. Synapsin I also binds to spectrin and actin microfilaments, and may serve as an anchor between synaptic vesicles and the cytoskeleton. The affinity of synapsin I for synaptic vesicles is decreased by phosphorylation, and neurotransmitter release is preceded by a reversible phosphorylation of synapsin I. Therefore, synapsin I phosphorylation results in the release of synaptic vesicles from their anchorage sites on the cytoskeleton, thus allowing the vesicles to move to the active exocytosis zones. 2. Docking and Fusion Proteins In addition to the actin-binding proteins, several proteins from the secretory vesicle membrane, the plasma membrane, and the bulk cytosol interact together to form the core complex (or a docking complex) with docking, priming, and fusioning properties (or budding properties). The most
710
SECTION V Synaptic Transmission and Sensory Transduction
important of these properties are described next (Stidhof, 1995; Fernandez-Chacon and Stidhof, 1999). The vesicular proteins synaptobrevins (or vesicle-associated membrane proteins, VAMPS) and the plasma membrane proteins syntaxin and synaptosome-associated protein (SNAP-25, M r 25 kDa) form the three complex membrane proteins (called the core complex) essential to the regulated exocytosis machinery in neurons and neuroendocrine and endocrine cell types. Moreover, the potential involvement of the cytosolic proteins, N-ethylmalimide-sensitive fusion protein (NSF) and the soluble NSF attachment proteins (SNAPs) (c~,/3, 7 isoforms), was demonstrated by the discovery that all these proteins interact to form a complex named SNARE (for SNAP receptor) (Fig. 3). Based on in vitro studies, a model for neurotransmitter release was proposed, in which synaptic vesicles become docked at the plasma membrane by a specific pairing of VAMP (the vesicle membrane, v-SNARE) with syntaxin 1 and SNAP-25 (the target membrane, t-SNARE). The proposed sequence for docking and priming is described hereafter (Stidhof, 1995; Aroeti et al., 1998; Burgoyne and Morgan, 1998a). Before and/or during docking, syntaxin is bound to Munc 18 and synaptophysin to synaptobrevin. These two complexes, syntaxin-Muncl8 and synaptophysin-synaptobrevin, must dissociate in order for the core complex to be formed. Formation of this complex is considered to be the first step of vesicle priming. Syntaxin and SNAP-25 (tSNARE) can then bind tightly together to form a highaffinity site for synaptobrevin (v-SNARE) located on the vesicle membrane; the core complex has a stoichiometry of
1:1:1. The trimeric core complex serves as a receptor (SNARE) for the soluble SNAPs (not related to SNAP-25). NSF will only interact with SNAPs (a,/3) bounded on the trimeric complex. The NSF-SNAP receptor forms a multisubunit particle that sediments at 20S; it may form the core of a generalized apparatus catalyzing bilayer fusion (Srllner et al., 1993). NSF is a trimeric protein that crosslinks multiple core complexes into a network. The core complex is then disrupted by enzymatic activity of NSF under ATP hydrolysis. Botulinum A and tetanus toxins are toxin proteases able to digest synaptobrevin, SNAP25, and syntaxin. When entering in the nerve terminal, they irreversibly inhibit exocytosis. However, the number of docked granules is not decreased by the toxins, indicating that the primary function of the core complex is fusion and not docking. Hydrolysis of ATP by NSF is followed by ATP-independent steps (see later) sensitive to temperature, H +, and Ca 2+. The last step is Ca2+-sensitive and likely involves a Ca 2§ sensor at the site of exocytosis. Synaptotagmins (Syt) are membrane glycoproteins found in brain secretory vesicles of which eight forms have been cloned. One of these, synaptotagmin I (Syt I), plays a pivotal role in the Ca2+-triggered neurotransmitter release as a Ca 2§ sensor (Stidhof and Rizo, 1996; Goda and Stidhof, 1997; Burgoyne and Morgan, 1998b). The functional implication of multiple synaptotagmins is unknown. Syt I binds Ca 2§ cooperatively and undergoes a Ca2+-dependent conformational change; the coefficient of cooperativity (4) is similar to that observed for Ca2+-triggered release. Syt I also binds phospholipids as a function of Ca 2+ with high affinity (half max-
FIGURE 3. SNARE, the receptor involved in docking and fusion. The 20S particle that forms the core complex contains several interacting proteins, v-SNARE, related to synaptobrevin (VAMP) and located on the vesicle, binds to t-SNARE, related to syntaxin and SNAP-25 and located on the plasma membrane, to form the SNARE or SNAP receptor. SNAPs and NSF can then bind to the receptor to complete the core complex. Numerous SNARE-related proteins, each specific for a single kind of vesicle or target membrane, ensure vesicle-to-target specificity. Binding of synaptobrevin relies on the inhibition of the Ca2§ channel.
42. Excitation-Secretion Coupling
imal binding, 5-6 txM). Syntaxin, one of the proteins of the core complex, in addition to its role in the docking and fusion process of the vesicle, controls the entry of Ca 2§ by the voltage-dependent Ca 2+ channels (L- and N-type). Indeed, binding of syntaxin to the Ca 2§ channel prevents the opening of the channel by membrane depolarization. When the vesicle docks to the membrane by binding to syntaxin and SNAP-25, the inhibition of syntaxin on the channel is relieved, allowing voltage-dependent opening of the channel (Fig. 3) (Geppert and Stidhof, 1998). In addition to phospholipids and syntaxin, Syt I binds to neuroxins, a family of neuronal cell surface proteins, and to AP-2, a protein complex involved in synaptic vesicle endocytosis. In addition to a Ca 2+ sensitivity similar to Ca2+-triggered release, the role of Syt I as a Ca 2§ sensor has been illustrated in several ways. Knockout mice for Syt I have impaired Ca 2+triggered transmitter release, but release can still be obtained by Ca2+-independent agents such as hypertonic sucrose or the excitatory neurotoxin a-latrotoxin (the receptor for a-latrotoxin belongs to the neuroxin family). Synaptotagmin Drosophila mutants show a severe but incomplete block of neurotransmission with an altered Ca2+-dependence in some mutants. Injection of synaptotagmin peptide in squid nerve terminal inhibits release and vesicles accumulate possibly by competing for a common effector. Synaptotagmins are also able to interact with non-neuronal syntaxin, indicating that they can play a role in a variety of cell types from endocrine and immune systems. Synaptophysin and the related protein synaptogyrin are major integral membrane proteins of small presynaptic vesicles. The presence of phosphorylation sites for tyrosine kinase of the Src family could indicate that phosphorylation modulates protein activity. The primary structure was deduced from the cDNA sequence, leading to the proposition that synaptophysin and synaptogyrin span the membrane four times with N- and C-terminals located in the cytoplasm. Ubiquitous isoforms of these two proteins are co-expressed in all cells and could represent invariant components of trafficking organelles. The annexin family includes several Ca 2+- and phospholipidbinding proteins with conserved structure. Two of these are thought to be involved in the mechanism of exocytosis. Synexin (annexin VII) is a calcium-binding protein (47 kDa) with four transmembrane domains. This protein demonstrates a voltage-dependent Ca 2+ channel activity. Synexin is able to induce the aggregation of chromaffm vesicles in the presence of Ca 2+. A model for the synexin-driven Ca2+-dependent membrane fusion has been proposed in which synexin monomers polymerize as the concentration of Ca 2§ increases. The polymerized synexin forms a hydrophobic bridge between the two membranes. Annexin II (calpactin) is located on the cytoplasmic side of the plasma membrane in chromaffin cells. Sites of phosphorylation for protein kinase C and cAMP- and calmodulin-dependent protein kinases have been localized within the N-terminal domain; phosphorylation of these sites could inactivate the protein. A possible role for annexin II could be to link the granule with the plasma membrane (docking) and/or to induce fusion (Burgoyne, 1991). Exocytotic fusion mechanisms also involve Ptdlns transfer protein, Ptdlns 4-kinase (PI4K), and PI5K to pro-
71 1
mote formation of Ptlns4,5P 2 on the vesicle membrane (Martin, 1998).
III. Cellular and Molecular Events in Chromaffin, Mast Cells, and Neuronal Synaptic Vesicles Exocytosis is an all-or-none phenomenon in which Ca 2§ plays a pivotal role. Ca 2§ is required for second messenger activity, for the control of cytoskeletal dynamics, and for the vesicle-plasma membrane fusion process. In neurons, neuroendocrine cells, and some endocrine cells, the electrical activity of the cell, the action potential, leads to the opening of voltage-dependent Ca 2§ channels with a subsequent increase in cytosolic Ca 2§ Neurotransmitter release at synaptic and neuromuscular junctions, peptide hormone secretion, and catecholamine release by the adrenal medulla all belong to this class. In nonexcitable cells, the triggering Ca 2§ signal is provided by release of Ca 2§ from intracellular stores after appropriate stimulation. Depletion in Ca 2§ from these intracellular pools activates an influx of Ca 2+, which is responsible for the sustained increase in [Ca2+]i observed in many cell types. Moreover, this Ca 2+ influx provides Ca 2+ ions for the replenishment of internal stores.
A. Dynamic Changes in the Cytoskeletal Networks Are Required for Exocytosis Secretion is a process that requires (1) the movement of secretory vesicles toward the plasma membrane, (2) the fusion of vesicles with the plasma membrane, and (3) subsequent release of secretory contents in the cell exterior. As explained before, in resting conditions, actin filaments (F-actin) are preferentially localized in the cell surface and act as a barrier to the secretory granules, impeding their contact with the plasma membrane (Trifar6 et al., 1992). Stimulation of chromaffin or mast cells as well as neuronal synaptic vesicles produces disassembly of the actin network and removal of the barrier (Tchakarov et al., 1998) (Fig. 2). Several observations support this concept: (1) Direct evidence for an actin barrier has come from the use of drugs that affect actin assembly and disassembly. Cytochalasin B and DNase I prevent actin assembly and drive the system toward net disassembly and increased secretion in permeabilized chromaffin cells. (2) Studies using fluorescent rhodamine-labeled phalloidin (a drug that stabilizes actin filaments in vitro and stops actin disassembly on stimulation) and actin antibodies have shown, in resting cells, a strong cortical fluorescent ring of filamentous actin. Cholinergic receptor stimulation produces a fragmentation of the fluorescent ring, leaving cell cortical areas devoid of fluorescence. These changes are accompanied by a decrease in F-actin associated with a concomitant increase in G-actin (Tchakarov et al., 1998) (Fig. 4). (3) Results from the use of toxins also support the concept of a cortical actin barrier. Botulinum C2 toxin, which ADP-ribosylates actin and inhibits actin polymerization, enhances secretion in PC12 cells. In contrast, tetanus and botulinum A toxins, which
712
SECTION V Synaptic Transmission and Sensory Transduction
FIGURE 4. Rhodamine-phalloidinfluorescence of chromaffin cell cytoskeleton. Cultured chromaffin cells were incubated for 40 s in Locke's solution in the absence (A) or in the presence (B) of 10 I~M nicotine. Cells were fixed and processed for fluorescence studies. Resting chromaffin cells showed a bright and continuous cortical fluorescent ring (A), while nicotinic-receptor-stimulated cells showed disruption in the cortical fluorescent ring. Some fluorescent patches are shown by arrowheads (B). Three-dimensional image analysis of the same patterns is shown below each cell. In control cells, there is a uniform cortical fluorescence intensity pattern, whereas in stimulated cells, the cortical fluorescent intensity pattern shows irregularities such as valleys and peaks. These peaks correspond to the patches observed in cells (Adapted with permission from Tchakarov et al. (1998)).
block actin disassembly, inhibit exocytosis upon cholinergic stimulation. Several actin-binding proteins present in the cell cortex undergo changes upon stimulation. In a resting cell, fodrin is localized in the cell cortex. On stimulation, it rearranges into patches beneath the plasma membrane. This redistribution could be related to the clearing of exocytotic sites at the plasma membrane. Scinderin is a cytosolic protein that shortens actin filament length when Ca 2+ is present in the medium. Stimulation induces both redistribution of scinderin from cytosol to cell cortex and F-actin disassembly, which precedes exocytosis. Thus, stimulation-induced redistribution of scinderin and F-actin disassembly produce subplasmalemmal areas of decreased cytoplasmic viscosity and high secretory vesicle mobility. All these processes require the presence of Ca 2+ in the extracellular medium. Therefore, only secretagogues that induce Ca 2§ entry are able to produce these effects (Fig. 2B). In isolated chromaffin cells, stimulation with nicotinic agonists can result in secretion of about 30% of the total catecholamine. Electron microscopic observations show that a small number of granules lie within the exclusion zone of the cell cortex. This demonstrates the importance of changes in cortical actin to allow movement of the bulk of the granules involved in a full secretory response. Nevertheless, low levels of exocytosis, due to these granules in the cortical exclusion zone, could occur without generalized changes in cortical actin. Thus, in a physiological situation, where relatively few exocytotic events occur per stimulus, changes in cortical actin may not be necessary for the initial wave of exocytosis, but are required for the movement, into the cortical exclusion zone, of granules ready for the next stimulus. A similar picture has emerged from studies on the nerve terminal cytoskeletal phosphoprotein, synapsin I, which is believed to cross-link synaptic vesicles and release them following depolarization and phosphorylation of the synapsin I.
B. Physical Events Associated with the Fusion of Vesicles to Plasma Membrane In regulated exocytosis, fusion of the secretory granules with the plasma membrane is triggered by an appropriate signal. The fusion process, following the triggering signal, can be very fast, such as in mammalian nerve terminals, with delays of less than 0.2 ms between the action potential and exocytosis. In some cells, however, delays of 0.2 s (chromaffin cells) to 50 s (mast cells) are observed. This delay is thought to be caused by the time required for production of second messengers possibly involved in exocytosis and removal of the cytoskeletal barrier that immobilizes the vesicles. However, the physical interactions between the granule membrane and the plasma membrane remain similar whether the exocytotic delay is fast or slow. Accordingly, the following fusion events will be described along general lines based on a sequence proposed by Almers (1990). 1. Capacitance Jump
Each time a secretory granule fuses with the plasma membrane, the total capacitance of the cell increases by a value proportional to the surface area of the new membrane added to the existing cell membrane. Assuming that biological membranes have a constant specific capacitance of about 1 ~F/cm 2 allows a simple calculation of the granule size. Upon fusion of a single vesicle, the capacitance value increases abruptly to a new stable value as the membrane of the vesicle and the plasma membrane become continuous and the vesicle lumen opens into the extracellular space. Figure 5 illustrates the equivalent circuitry of a resting cell (Fig. 5A) and that of a cell undergoing exocytosis (Fig. 5B). During degranulation, several granules fuse with the plasma membrane, which produces a typical staircase recording (Fig. 5C). Degranulation in three different cell types are presented in Fig. 6: human neutrophil (Fig. 6A), guinea pig eosinophil
42.Excitation-SecretionCoupling A
71 3
B
C
Oo ../------
T
T
-Oo
]
c.
To- t + t,,
-u-u-
FIGURE 5. Capacitancemeasurement and equivalent electrical circuit. (A) The cell is patch-clamped in the whole-cell configuration. The cell is at rest and an unfused granule is shown near the plasma membrane. The equivalent circuit is represented by membrane resistance R m in parallel with membrane capacitance C m. The series resistance due mainly to the micropipette has been omitted for clarity. (B) Once the granule has fused with the plasma membrane, its capacitance Cv, proportional to the surface of membrane added, is added to the total capacitance of the cell. Rp is the fusion pore resistance, which will rapidly decrease as the pore dilates. (C) The fusion of a granule with the plasma membrane increases the capacitance by step. Granules of identical sizes induce the same increases in capacitance.
(Fig. 6B), and horse eosinophil (Fig. 6C). Note that the amplitude of the individual step capacitance is lower in human neutrophil than in horse eosinophil, reflecting the different sizes of the vesicles. A capacitance step amplitude histogram is built by measuring the step height of a large number of individual events. An example of this is illustrated in Fig. 6D for neutrophils. The capacitance step amplitudes range from 1 to 6 fF (1 to 6 x 10-15 F), with a greater number of events having an amplitude of 2 fF. Assuming a spherical shape, the diameter of the granule can be calculated from the step change in capacitance, 6 C m in fF, by the relation r
m -
7rd 2
(1)
The frequency distribution of the sizes of the granules obtained from the capacitance step amplitude histogram is represented in Fig. 6E for guinea pig eosinophils. The distribution is fitted (smooth line) by the sum of two Gaussian curves with means of 520 and 590 nm (Lindau and Gomperts, 1991). This size distribution fits in well with the morphometric data obtained by direct microscopic observation of the secretory vesicles. Capacitance step measurements have been achieved in a variety of cell types, including adrenal chromaffin cells, mast cells, pancreatic acinar cells, neutrophils, eosinophils, basophils, pituitary lactotrophs, and nerve terminals derived from the posterior pituitary. Capacitance step values generally range between 1 and 30 fF, corresponding to granule diameters of 0.2 to 1 Ixm. Giant vesicles with diameters ranging from 1 to 5 txm are found in mast cells from a strain of genetically defective beige mice (strain C57BL/6J-bgJ/bgJ). In these mice, mast cells and other granulocytes are unable to limit the size of their secretory vesicles; mast cells contain 10 to 40
giant vesicles that can easily be observed under photonic microscopy. These cells thus provide an ideal material for exocytosis studies and, for this reason, have been extensively used. The capacitance method offers the possibility to study degranulation in real time online; time analysis of the secretory process reveals that the granules fuse sequentially, one by one, with the plasma membrane. However, in mast cells, capacitance step analysis demonstrates the presence of step values greater than 60 fF, which could not be produced by the fusion of a single vesicle. A detailed analysis of the capacitance step histogram reveals a multimodal distribution of granule size, which indicates that the larger granules could be formed by the fusion of two to five single granules with each other.
2. Capacitance Flickering The pattern of the staircase increase in capacitance during degranulation demOnstrates that each step builds upon the previous one, indicating that the fusion event is irreversible. However, closer observation of capacitance jumps in mast cells reveals the existence of "on" and "off" steps. The "on" step is produced by the opening of a small-diameter pore, called a fusion pore, which adds the surface of the vesicle to that of the cell. Once opened, the fusion pore can close quickly and reopen (Fig. 7). Rapid oscillation between open and closed states gives the appearance of a flickering of the capacitance (Almers, 1990). Size distribution of capacitance steps during the flickering period shows that large steps are absent and that all steps remain in the range of values expected for single vesicle fusion events. The size of the fusion pore is proportional to its conductance. An unexpected result is that the reclosing of the pore occurs, not only in smalldiameter pores (low conductance), but also in larger-sized
B
A
C
: . - : -
,.:~y_::.
J - I
30 fF
5s
2~
_~ ' ~ i 1~176176 fF i
lOs
20 s
,i
E
D 150 f
5O =3
02
O'0i 4 6 s step size/fF
io
0
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
granule diameter/~tm
FIGURE 6. Analysisof step capacitance. Capacitance recordings obtained from different cells: (A) human neutrophils, (B) guinea pig eosinophils, (C) horse eosinophils. Note the staircase appearance of the recordings, which reflects the sequential fusion of individual granules and the size of the step capacitances, which are proportional to the size of the granules. (D) Frequency distribution of step sizes for human neutrophils. (E) Distribution of granule size for guinea pig eosinophils derived from capacitance measurement. (Adapted with permission from Lindau and Gomperts (1991)).
714
FIGURE 7. Flickeringof the membrane capacitance. Fusion of the granule is reversible and can oscillate between an unfused state and a fused state. Each time the granule fuses with the plasma membrane, the capacitance of the cell increases by a step. If the fusion pore closes, the capacitance recovers its initial value; incomplete recovery indicates very brief closures exceeding the speed of the recording system. Eventually, a stable fused state can be reached.
pores having a conductance of several nS. Capacitance flickering is rather frequent in mast cells, but very few have been observed in eosinophils. This raises the question of the undetectable presence of flickering during the fusion of all vesicles before the irreversible fused state is attained. Indeed, the initial phase of fusion is a very fast process, which cannot be faithfully recorded by the speed of the recording techniques used. The earliest steps as well as short-lived flickering are certainly missed. 3. F u s i o n Pore
Freeze-fracture images of exocytosis in mast cells and neutrophils reveal the presence of narrow pores formed between the granules and the plasma membrane. The size of the pore increases as fusion progresses, allowing the release of vesicle contents. The presence of pores joining two secretory vesicles is also observed (Lindau and Gompertz, 1991). Electrically, the first opening of the fusion pore generates a brief transient current from which several parameters can be deduced. a. Size of Fusion Pore. The conductance of the fusion pore can be calculated from the current transient produced by movement of the charges between two differently charged membranes, that is, the plasma membrane and vesicle membrane. The initial value of the current, I 0 , is related to the initial pore conductance, go, by the relationship
where E c is the potential across the plasma membrane and E v the potential across the vesicle membrane. The initial conductance of the fusion pore in mast cells was found to be 230 pS. The corresponding pore diameter was calculated assuming a resistivity of 100 f~.cm and a pore length of 15 nm. The abrupt increase in conductance corresponds to the all-or-none opening of a pore having an inner diameter of less than 2 nm. For comparison purposes, gap junction channels have conductances that vary from 80 to 240 pS and a diameter of approximately 2 nm. It can thus be proposed that the fusion pore is a large protein spanning across two membrane thicknesses, and having a structure and function resembling that of a channel. Data obtained from electron microscopic observations reveal that the smallest pores that can be observed
SECTION V Synaptic Transmission and Sensory Transduction
have diameters between 30 and 50 nm, values far higher than the values reported from conductance measurements. A rapid dilatation of the pore soon after its formation might explain this discrepancy. Once the pore has been formed, the conductance increases abruptly to a value near 250 pS. Thereafter, the pore begins to dilate, possibly by infiltration of lipid molecules between the subunits of the protein structure, followed by an increase in conductance. The pore conductance increases from 500 pS to 3000 pS in about 25 ms; a plateau is reached after 150 ms, corresponding to a pore diameter of more than 16 nm. b. Does the Fusion Pore Leak? Once established, the diameter of the fusion pore is similar to that of the gap junction, between 1.5 and 2 nm. Since gap junctions allow the intercellular passage of molecules weighing up to 1.9 kDa, the question can be asked as to whether granule contents can leak through the fusion pore. In guinea pig eosinophils, the irreversible fusion of the granule with the plasma membrane is occasionally preceded by a long-lived fusion pore having a conductance of 70-250 pS. When granules were loaded with the fluorescent dye quinacrine, no release of the dye in the extracellular medium could be observed during the life of the fusion pore. Only after the fusion pore had completely dilated, and the vesicle had reached its irreversible state of fusion, was the dye released. These results indicate that the fusion pore is too narrow for the release of granule contents. However, computations based on the diffusion of a small molecule such as histamine predict that a granule with a diameter of 0.8 mm should release its contents rapidly through the opening of the fusion pore. Recently, experimental proof was provided by Neher and collaborators using bovine chromaffin cells (Chow et al., 1992). Secretion was measured by voltametry, while cells were studied by voltage-clamp. The cells were stimulated by depolarizing the membrane from a holding potential o f - 6 0 mV to a step potential of + 10 mV for 25 ms to activate the Ca 2§ channels. The amperometric signals, which represent the detection of the released catecholamine molecules by the carbon electrode (potential of 800 mV), were transient with a fast or slow rising phase and variable amplitudes. A histogram of integrals of current transient amplitude obtained on several cells showed that the mean charge transfer had a value of 0.76 pC, which is equivalent to the release of 2.36 • 106 molecules of catecholamine. Sometimes, larger events were detected, presumably corresponding to multigranular exocytosis. One interesting and surprising feature was that the majority of the fast-rising events were preceded by a small "foot" or "pedestal," as illustrated in Fig. 8A. The mean duration of the foot was 8.26 ms and the mean charge 34 fC, equivalent to 1.05 • 105 molecules (Fig. 8B). The foot was interpreted as reflecting a slow leakage of catecholamine molecules through the fusion pore formed during the early step of the fusion process. A second important result of this study was the discovery and quantification of a long latency period between the end of the stimulus and catecholamine release. The majority of the secretory events occurred 5 to 100 ms after the end of the electrical stimulus, which is rather long when compared to nerve endings, where the delay is about 1 ms. A complex cascade of intracellular events triggered by
42. Excitation-Secretion Coupling
71 5
A
thus establishing a relationship between the dilatation of the pore and decreased plasma membrane capacitance. Once the flickering stops, the conductance recovers its initial value but the whole-cell capacitance remains lower. These results are interpreted as reflecting a decrease of the plasma membrane surface due to a net transfer of material to the granule membrane. The difference between the "on" and "off" steps (found in one-half of the transient fusions with values b e t w e e n - 2 t o - 4 fF) is proportional to the duration of contact between the plasma membrane and the secretory vesicle (Fig. 9C). The rate of cell surface area reduction is 0.16 txm2-s-1. The transfer of membrane is facilitated by the fact that the membrane of the secretory vesicle is under tension. Upon fusion, a movement of phospholipid molecules occurs from the plasma membrane to the granule membrane. A possible mechanism for generating tension in the granule membrane is osmotic swelling. However, fusion can proceed in isotonic, hypotonic, or hypertonic solutions with no change in the kinetics of capacitance increase.
A
to intermediate metabolism, ATP production/ADP consumption and K+(ATP) channel closure (2), ~ membrane depolarization (3), ~ opening of voltage-dependent Ca2§ channels (4), and finally Ca2§ binding to granule docking complex and granule fusion (5). Non selective cation channels that form the counterweight to K+(ATP) channels at rest are shown to the left of the Glut-2 transporter. Further to the left are the Gprotein-linked receptor cascades for muscarinic, cholinergic agonists, and the incretin GLP-1. The latter two "priming" compounds produce second messengers capable of releasing Ca2§ from intracellular stores and/or activating protein kinases to modulate channel activity and granule mobilization. Lastly, the contents of the secretory granule, namely insulin and ATP, may "feed back" on the/3 cell via cell surface receptors and complex signaling cascades.
(Step 3). When V reaches the threshold for activating volt% + age-dependent Ca and Na channels, electrical activity and voltage-dependent Ca 2§ entry begins (Step 4). The resultant rise in cytosolic Ca 2§ triggers Ca2+-dependent fusion of insulin granules with the plasma membrane (Step 5). Steps 4 and 5 are reminiscent of the process of depolarization-secretion coupling, well studied at nerve terminals and other electrically excitable endocrine cells (adrenal chromaffin cells, pituitary lactotropes and melanotropes, and terminals of the
supraoptic nucleus cells), with subtle differences that shall be discussed later. However, this scheme does not answer some rather old questions. (1) How do enhancers of cAMP serve to restore stimulus-secretion coupling in metabolically and electrically active but nonsecreting fl cells? (2) How, aside from possibly depolarizing the/3 cell, do anticipatory or modulatory factors such as gut hormones and acetylcholine work? The anticipatory-modulatory pathways of receptor-mediated
43. Stimulus-Response Coupling in Metabolic Sensor Cells
stimulus-secretion coupling, as well as possible autocrine effects of insulin and ATP released by exocytosis, are outlined on the left of Fig. 3 and also will be discussed later. In fact, it may be possible that under very extreme conditions, modulatory pathways may result in some secretion independent of an inciting Ca 2+ current or a detectable rise in global cytosolic Ca 2+. This might occur through the combination of a very localized rise in cytosolic Ca 2+ with a vastly increased readiness of the secretory apparatus. However, it should never be forgotten that the "sui generis" of the/3 cell remains metabolic regulation of cell electrical activity via K+(ATP) channels. A. Role of ATP-Sensitive K§ Channels in Coupling Metabolism to/3-Cell Depolarization How do fuel metabolites produce/3-cell depolarization? Early patch-clamp experiments, using the cell-attached mode of recording, revealed the presence of a very interesting K + channel that was open at fasting levels of extracellular glucose (2-3 mM), but rapidly closed on raising ambient glucose concentrations to those encountered after a meal (5-7 mM). Usually, channel closure is followed closely by the onset of electrical activity. This channel is open at the cell's resting potential, and its activity is not dependent on membrane voltage. However, the channel is inward rectifying (i.e., single channels pass inward flow of potassium better than outward). These channels have many features predictable from the fuel metabolism hypothesis: addition to the bath of any metabolized fuel closes the channel, yet the channel is reopened subsequently by the addition to the bath of inhibitors of metabolite transport or oxidation. The time course and relative change in channel activity seen in a cell-attached patch during a variety of metabolic maneuvers closely parallels the changes in whole-cell resting conductance (Gm) , monitored as the change in membrane potential in response to a given test pulse of current applied (see Fig. 3A, B). The latter strongly suggested that the activity of this channel, which tends to move Vm to a level near the potassium equilibrium potential, E~, critically controls the cell's resting potential. Since this channel population is the predominant type open at rest, its closure en masse would be needed to make the K + conductance of the cell comparable to that of other conductance pathways operative at rest. The latter are nonselective cation conductance channels and anion conductance channels with estimated equilibrium potentials of 0 or N -30 mV respectively. This would permit Vmto depolarize to the threshold for activating cell electrical activity (~ -45 mV). Another hint of the importance of this type of channel was that it is closed by oral hypoglycemic agents. How is this channel gated? In the cell-attached patchrecording configuration, in which the exterior of the channel was isolated from bath glucose by a very tight pipette-tomembrane seal, it is safe to assume that the channel is gated by one or more metabolic intermediates. A priori, cell metabolism might affect a host of metabolic intermediates, ranging from high-energy phosphate compounds (e.g., ATP and ADP), to redox equivalents, to cytosolic H + or Ca 2+. Early experiments (see Fig. 3C) showed that after excision of the patch, in an inside-out excised patch configuration,
729
the metabolite-regulated K + channel was avidly gated by ATP at its inner surface, hence the name K+(ATP) channel. Free concentrations of ATE or its nonhydrolyzable analogs (e.g. AMP-PNP) as low as 10 laM reduce channel activity by half even in the absence of Mg. This suggests that ATP closes this channel by directly "gating" it rather than "phosphorylating" it. In parallel with this, in whole-cell patch experiments where ATP dialyzes out of the cell (due to the absence of ATP in the pipette), resting G m increases, and the cell hyperpolarizes. Interestingly, ATP inhibition of channel activity in the excised (inside-out) patch can be partially reversed by addition of MgADP to the bath. Hence, changes in the relative concentrations of ATP and ADP, which usually change in a reciprocal fashion as metabolism is altered, should be an important factor in physiological channel gating. In most cells, ATP concentrations change little with metabolism, but percentage-wise, ADP levels change more dramatically. Hence the channel may have been misnamed. To be sure, the activity of the K+(ATP) channel is also altered by other potential intermediates whose cytosolic concentrations change with metabolic stimulation (e.g., NADH/NAD ratio, H + and Ca2+), but these effects are much smaller and require changes beyond the physiological range. A puzzling question regarding the function of this channel is how its low Kd value (luM) for ATP-induced channel closure in the excised membrane patch can be reconciled with the millimolar-range concentration of ATP in the cytosol. Several factors mitigate this discrepancy. (1) Complex effects of ADP on channel activity. ADP competition for an ATP binding site, combined with allosteric action of ADP at an independent binding site, adjusts the half-maximal effective concentration (Kd) of ATE measured in the presence of ADR closer to the mM range. (2) Compartmentation of the ATP and steep ATP gradients from source to sink. It is likely that the overwhelming majority of ATP is of mitochondrial origin: mitochondrial inhibition totally prevents glucose-induced channel closure, while ketone bodies, amino acids, and their deamination products (e.g., leucine and ketoisovalerate), which directly enter mitochondria, stimulate channel closure even in the face of inhibition of glycolysis. In addition, given the random localization of these channels over the surface of cultured cells, it is likely that the channels are interspersed among ATPases. Hence, steep cytoplasmic gradients of ATP and ADP, due to en passant consumption of ATP and concomitant local production of ADP by neighboring plasma membrane ATPases, should make the ATP/ADP available to the channel different from the average cytosolic concentrations actually measured. (3) Modulation of gating by other metabolic intermediates such as long chain fatty acids. (4) Artifactual enhancement of channel sensitivity to ATP in the excised patch by progressive washout of intrinsic channel inhibitors such as anionic lipids (phosphoinositol phosphates), from the exposed membrane. Polyanionic molecules might specifically compete with ATP for binding to the channel or else contribute to a negative surface charge at the inner surface of the plasma membrane, thereby electrostatically repelling soluble anions. Two features of the metabolite-regulated K+(ATP) channel proved critical to its cloning and structure-function analysis. (1) The channel functions as an inward rectifier K +
[ [
2 mmol/L glucose +3 mmol/L NAN3. ]
[
+3 mmol/L NaN 3
]
Vm
(mY)
i
9
9
l
.......
l
0 ,D
- 7 2 ,o J
C 0 mmol/L ATP 0 ] 2
3
I00 lamol/L ATP + I00 lamol/L ADP
0.
100 Iamol/L ATP 0-l 2
0pA 10pA
,
0 mmol/L ATP
.0,21.
2 3
F I G U R E 3. ATP-sensitive K + channels and electrical activity in canine pancreatic islet 13 cells. (A) A cell-attached patch was formed with the pore-forming antibiotic nystatin in the pipette. Shortly thereafter, ion-channel activity was recorded before, during, and after application of sodium azide (NAN3, 3 mM) to the low-glucose bath. (B) Electrical activity recorded 20 minutes later. By that time, the patch of membrane encompassed by the pipette had been permeabilized by insertion of nystatin channels, thereby permitting recording of membrane current or voltage from the whole cells. Reapplication of NaN 3 results in membrane hyperpolarization, abolition of electrical activity, and increased membrane conductance (signified by the reductions in transient hyperpolarizations accompanying-5 pA currents applied at arrows). V m denotes cell membrane potential. (C) Nucleotide gating of K+(ATP) channel activity in the excised patch. Addition of 100 micromolar ATP (100 ~M) reduces channel activity by threefold (C, upper right panel); further addition of ADP restores channel activity (C, lower left panel), whereas washout of ATP and ADP results in overshoot of channel activity (C, lower right panel) (D.M. Pressel and S. Misler, unpublished data).
43. Stimulus-Response Coupling in Metabolic Sensor Cells channel and hence could be cloned as a member of that family. (2) The channel is the major locus of action of hypoglycemic (glucose-lowering) sulfonylureas and could be cloned as a sulfonylurea receptor (SUR). The sulfonylureas (e.g., tolbutamide and glyburide (glibenclamide)) enhance glucose-induced insulin release by/3 cells and are useful in the treatment of non-insulin-dependent diabetes mellitus, in which/3 cells produce significant quantities of insulin, but release it out of synchrony with the appearance of the glucose load. At concentrations equal to free plasma levels that enhance glucose-induced electrical activity and insulin secretion in intact animals, the sulfonylureas close K+(ATP) channels in cell-attached or excised patches alike, while leaving voltage and calcium gated K + channels unaffected. Curiously, though K+(ATP) channels in myocytes show similar conductance, kinetics, and ATP sensitivity, they have greatly reduced sulfonylurea sensitivity. To reconstitute channel activity in a naive cell (e.g., Xenopus oocyte or clonal cell) it was necessary to transfect or inject the cell with cDNA or mRNA coding for (1) a standard inward rectifier type subunit (Kir6.2), comprised of two membrane-spanning domains separated by an H loop dipping into the bilayer, and (2) the SUR from /3 cells (SUR1) (see Fig. 4A, B). SUR 1 is a complex protein with multiple transmembrane domains (hydropathy plot analysis suggests 17!). Structurally it contains two motifs resembling those of the ATP-binding cassettes of members of the ABC transporter superfamily, the most famous of which are the cystic fibrosis gene product, CFTR, and the multidrug resistance p-glycoprotein. SUR1 has two distinct cytoplasmic loops (or facettes) that bind nucleotide di- and tri-phosphates in the presence of Mg (NBFs). The channel itself is likely to be an octamer consisting of an inner ring of four Kir6.2 subunits (the confluence of the four H loops forming the ion selective pore) and an outer ring of four SUR1 subunits. The two types of subunits are needed to chaperone each other into the membrane as well as to provide full channel gating; without SUR1, Kir6.2 is not expressed. Clinically, mutations in SUR1 have been identified in persistent hyperinsulinemic, hypoglycemia of infancy, a disease where K+(ATP) channel activity is not detected even after metabolic inhibition and/3 cells are persistently depolarized and secreting. Point mutations to NBF result in channels that are closed by ATP in excised patches, but are not reopened by MgADE while more extensive mutations result in the absence of channel activity under any test circumstance. In contrast, a C-terminal truncated mutant of Kir6.2 (Kir6.2AC26) forms K + channels gated by ATP in the absence of Mg. These channels have lower affinity for ATP (Kd -~10 ~M in the absence of Mg) than the native or Kir6.2/SUR1 cloned channel (Kd .--100 laM) but show similar need for ATP cooperativity in channel gating (n = 4). However, these channels lack both inhibition of activity by sulfonylureas and enhancement of activity by MgADE Hence, while Kir6.2 might be considered the ATP gating subunit, SUR1 with its NFBs is the ATP-hypersensitizing subunit as well as the subunit endowing the channel with special sensitivity to drugs and MgADE A schematic model of K+(ATP) channel gating is presented in Fig. 4C. SUR might have a natural ligand, a cytoplasmic protein called en-
731 dosulfine, recently demonstrated to displace sulfonylureas from SUR binding sites. Analysis of mutant channels has also elucidated the origin of a number of previously puzzling effects. One example is the "refreshment effect" of MgATE That is, rapid wash-in and wash-out of MgATR even in concentrations of 10-100 ~M, can restore channel activity that otherwise would run down in time in its absence (see Fig. 3C). This effect can now be partially explained by the finding that small concentrations of MgATP as well as MgADP activate the channel by binding to a "Walker motif" of the NFB in native forms of SUR1 but not in those SURs whose NBFs have been mutated. Likewise, channel modulation by Mg-guanidine nucleotides also occurs at these sites, suggesting that there is no need to invoke G proteins in GTP or GDP modulation of channel gating (though G-protein-coupled receptors may independently modulate channel activity). Both effects require hydrolyzable nucleotides and Mg, suggesting that some phosphorylation reaction may be involved. A second example is the ability of certain vasodilator drugs, such as diazoxide, to open K+(ATP) channels, hyperpolarize the/3 cell and inhibit glucose-induced insulin secretion. These drugs also bind to the "Walker motif." A third example is the differential sulfonylurea sensitivity of K+(ATP) channels in 13 cells as compared with those in cardiac myocytes. While both channels have similar Kir6.2 subunits, their SURs differ. Replacing transmembrane domains 13-16 of cardiac SUR (SUR2) with the corresponding region of SUR by engineering of a chimeric SUR restores sulfonylurea sensitivity. Is this finely tuned coupling of cell metabolism to K+(ATP) channel activity used as a sensory mechanism by the two other major glucose receptor cells of the body, namely the satiety center of the hypothalamus and the peripheral sweet taste receptors? Recently, K+(ATP) channels have been sought in glucose-receptive (GR) neurons of the ventromedial nucleus (VMN) whose activity is associated with suppression of food intake. In these neurons, reduction of bath glucose or addition of the glycolytic inhibitor mannoheptulose results in hyperpolarization and cessation of electrical activity accompanied by an increase in resting G m; oppositely, tolbutamide produces excitation. In cell-attached patch recordings, cessation of electrical activity is associated with increased activity of a K+-selective channel, which is voltage-insensitive. After excision of the membrane patch, the latter channel is inhibited by cytoplasmic ATE However, this K + channel is quite distinct from the K+(ATP) channel in/3 cells: under identical excised patch recording conditions, the K/for ATP inhibition is 2.3 mM (rather than 20 luM in/3 cells) and the singlechannel conductance is 135 pS (rather than 55-65 pS in/3 cells). One very interesting feature of this channel is that it is opened, and the cell is hyperpolarized by leptin, a protein released by adipocytes, or fat-storing cells, in proportion to their triglyceride storage. Leptin also opens K+(ATP) channels in/3 cells and appears to do this in a manner dependent on activation of PI-3 kinase. Does leptin serve as a negative feedback hormone, that is signaling adequate energy stores by downregulating both appetite-induced caloric intake and insulin-dependent depot storage? One indication of this
732
SECTION V Synaptic Transmission and Sensory Transduction
A
SUR 1
Kir 6.2
Outside _
_
Inside
NBD1
C
NBD2
Mg ADP; D ATP
9 ATP Closed
Mg ADP
K*
.
ADP ,ATP
ATP Open
F I G U R E 4. Illustrations depicting K+(ATP) channel structure and function. (A) Transmembrane orientations of SUR1 and Kir6.2 subunits based on structural homology, hydropathy, and site-directed mutagenesis. (B) Proposed octamer array of SUR1 and Kir6.2 subunits. (C) Simplified reaction scheme in channel gating.
comes from obese, hyperphagic, hyperinsulinemic, and hyperglycemic db/db rats, where leptin has little influence on the firing rate of VMN neurons. In peripheral taste receptors, glucose chemoreception may occur through G proteins and a second messenger cascade. In these cells, application of sucrose or saccharin reduces the amplitude of a voltage-dependent outward K + current. This effect can be mimicked by application of membrane-permeable analogs of cAMP. In excised patches, K + channels with similar features are closed by exposure of the patch to ATP and cAMPdependent protein kinase, but not by exposure to ATP alone, suggesting that this K + channel is "gated" by phosphorylation.
B. Metabolic Regulation of DepolarizationSecretion Coupling? Recall that closure of K+(ATP) channels only depolarizes the cell. Calcium must enter the cells and insulin granules must be ready to be released. Does metabolism alter the "calcium trigger" or maintain the "readily releasable pool" of insulin granules?
1. The "Ca trigger" When/3 cells, exposed to sufficient glucose, a cAMP enhancer, and near physiological temperature, are depolarized for many tens to hundreds of milliseconds, they exocytose in a CaZ+-entry-dependent manner. Under these conditions, several studies have shown a ---3rd power dependence of exocytosis, as measured by amperometry or capacitance, on the total Ca 2+ entering the cell (i.e., the integral of the Ca 2+ current). Though most of the Ca 2+ current is carried by L-type Ca 2+ channels, in a subset of human/3 cells, where they are present in sufficient density, lower threshold T-type Ca 2+ channels also support exocytosis. We might expect changes in cell metabolism to affect the gating of Ca 2+ channels via phosphorylation or lipidation. To date, data supporting the hypothesis that enhanced glucose enhances Ca 2+ channel activation or slows inactivation have been inconsistent. Another paradigm, however, which illustrates other aspects of depolarization-secretion coupling may provide some other clues. Beta cells in intact islets secrete pulses of insulin that are time-linked to "bursts" of electrical activity. Longer "bursts"
733
43. Stimulus-Response Coupling in Metabolic Sensor Cells
evoke larger pulses of insulin. Contrary to the situation at most nerve terminals and despite quite sizeable Ca 2§ currents, it is very rare to see an isolated action potential, or a brief depolarization (10-50 ms in duration), provoke quantal release (here measured by amperometry). However, when the depolarization is extended out to many tens to several hundreds of milliseconds, release routinely commences in the course of the depolarization and often continues for several seconds after Ca 2+ entry ceases, while cytosolic Ca 2§ is falling (see Fig. 5). These data suggest that Ca 2§ might have to "jump through a few hoops" before triggering secretion and that it might set up a "chain reaction" that outlasts the peak Ca 2§ concentration (Fig. 6). Several possibilities for this "slow-to-start, slow-to-stop" pattern of secretion are currently under investigation. First, with large-sized granules that are not clustered in any regular array, Ca 2§ channels and granule release sites may not be all that closely co-localized. To be sure, the interspersed cytosol has many Ca2+-buffering molecules. Hence, if the site where Ca 2§ triggers exocytosis were, on average, a distance of 1-2 granule radii (say 200-300 nm) from the inner mouth of the Ca 2§ channel site, local cytoplasmic buffers would need to be saturated before Ca 2§ could appear at the release site. This would cause a slight delay in the onset of exocytosis. At this distance, during continuous Ca 2§ entry there might be good spatial equilibration of Ca 2§ so that the Ca 2+
A
1412-
.._.. Q) 0 t"
"~
B
~L_ ls
1 Hz
lamp
................
~,.a.-,
10-
2s 8--
6-
omv -1--t
g
0 121
u_
concentration at the release site might closely follow the average Ca 2+ concentration of the cell. This is in no way counterintuitive because it is well appreciated that Ca 2§ and cytosolic Ca 2§ concentrations as low as 1.5 pM evoke respectable rates of release (see Fig. 7B). Conversely, once Ca 2+ no longer enters the cell, the buffer might serve as a continuing source for Ca 2§ and support a "secretagogue" level of free cytosolic Ca 2+ for some time. This feature has been examined and modeled extensively in adrenal chromaffin cells but has only been explored in a cursory manner in/3 cells. Second, Ca 2§ entry might be causing regenerative release of Ca 2§ from some intracellular stores, thus providing a sort of propagating Ca 2§ wave, while the increase in cytosolic Ca 2§ might be stimulating Ca 2§ uptake by other stores for later use. To date, while it is well appreciated that Ca 2§ can be released from ER stores by the generation of IP 3, blockade of Ca 2+ release from ER stores has yielded only small changes in depolarization-induced rises in cytosolic Ca 2§ or exocytosis. However, the picture may be more complex in other situations. For example, when GLP-1 is applied to voltage-clamped/3 cells in the presence of 20 m M Ca 2+, "spontaneous" spikes of cytosolic Ca 2+ occur between depolarizations. To be sure, there are multiple stores of intracellular Ca 2§ aside from the IP 3 -releasable ones in ER. For example, mitochondria- and hormone-containing granules
,..
O!
I
I
2
4
6
~
_•
Vm ,J
-+---I
I
8 10 Charge (pC)
I
I
I
12
14
16
[ 2 oA 5s
4Hz
~.
120pA
~
0.5s
~ +5mY -70 mV
F I G U R E 5. Aspects of Ca2+ dependence of exocytosis in/3 cells (A) Relationship of capacitance increase to total Ca2+ entry (integral of Ca2+ current Qca) during/3-cell depolarization. Ca2+ entry was altered by changing the membrane potential to which the cell was voltage-clamped for the 200-ms test stimulus (unpublished data by D. Barnett and S. Misler). (B) Time course of amperometric events during trains of action potentials stimulated at 1 Hz (upper panel) and 4 Hz (lower panel) Note that none of the 11 action potentials occurring at 1 Hz stimulation results in quantal release, while at 4 Hz stimulation quantal release begins after several action potentials. (C) Time course of amperometric events recorded during continuous depolarization to +10 mV, where Ca2§ entry is maximal. Note that the first release event is seen --120 ms after start of voltage-clamp pulse, while release events continue, albeit at a low frequency, for nearly 3 s after depolarization is completed (unpublished data by Z. Zhou and S. Misler).
71t4
SECTIONV SynapticTransmissionand SensoryTransduction
,tl14,
Mo'Tp Ca2+ Ca2+ ~Ca2+ Ca2+ Ca2+
Ca2+ FIGURE 6. Calciummetabolism in fl cells: potential sites for modulation of depolarization-secretion coupling. After calcium enters the cytoplasm, it is likely that much of it rapidly binds to either fixed or mobile buffers, the latter perhaps resembling the tetra-clawed EGTA. Some of the bound Ca2§ unbinds to diffuse further. Diffusing Ca2§ may then bind to one of the following: (1) a Ca2+-binding protein (e.g. synaptotagmin) that is part of the docking complex that anchors a granule to the membrane and thereby directly triggers exocytosis, (2) a protein kinase, whose activity enhances the movement of granules along "motorized" tracks, (3) a binding site on Ca2+-gated channels in Ca2§ storage organelles, including the secretory granule itself, which thereby triggers Ca2+-induced Ca2§ release. Calcium uptake into organelles (granules, mitochondria, and ER) is probably slower (on the time scale of hundreds of milliseconds to seconds). After fusion with the plasma membrane, granule membrane may be retrieved piecemeal by a local "pinch off" mechanism, by using dynamin, or as part of a large endocytic "slurp."
might be expected to release as well as "mop-up" Ca 2+, but little is known about their release and uptake rates during various patterns of electrical stimulation or exposure to hormones and transmitters. Some recent evidence suggests that depletion of the Ca 2+ content of the acidic compartment of cytoplasmic organelles, which includes secretory granules, reduces depolarization-induced Ca 2+ release. In a metabolically stimulated cell where the Krebs cycle intermediates, NAD-derived redox equivalents and local ATP concentrations may be changing, it would not be surprising that changes in cell Ca 2+ metabolism in response to Ca 2+ influx might be a very dynamic feature. Third, entering Ca 2+ might rapidly recruit granules into a small, "readily releasable pool" waiting to be triggered.
2. The "readily releasable pool" (RRP) In cells designed for rapid secretion, it is convenient to assume that vesicles are lined up against the plasma membrane and ready to fuse as soon as the appropriate level of second messenger is achieved in the cytoplasm. Morphologically, in /3 cells, this does not appear to be the case (see Fig. 6). When viewed under the electron microscope, very few of the -~10,000 granules in the/3 cell are within one granule diameter of the membrane. Physiologically, experiments with single cells reveal that secretion is actually quite "wimpy"
unless "revved up" by other maneuvers; with repeated depolarizations, exocytosis "poops out" and may take many tens of seconds to recover. This suggests that the RRP of available granules is actually quite small and that maintaining it, under high output demand, is a non trivial matter. However, in/3 cells, two ideal sources of "physiological revving" of secretion are (1) activation of protein kinases (A and C) and (2) the provision of ATP by enhanced substrate metabolism. Protein kinases are well known to facilitate the attachment of a variety of granules to tracks (microtubules, actin filaments) that bring them in close proximity to the membrane. In/3 cells, during feeding, insulin secretion is "primed" even before serum glucose rises, by release of two substance that activate protein kinases, acetylcholine from vagal fibers and incretins (e.g., glucagon-like peptide (GLP-1) from enterochromaffin cells. Activation of muscarinic ACh receptors results in activation of phospholipase C and production of IP 3 and diacylglycerol. Activation of incretin receptors results in activation of protein kinase A pathway and enhanced cytosolic cAMP levels. These ACh and GLP-1 pathways might enhance distal processes in excitation-secretion coupling, along with increasing the rate of/3-cell depolarization in response to glucose, by affecting ATP-sensitive K + channels and/or the nonselective cation channels. Hence both background electrical activity and insulin release- and metabolite-induced electrical activity and
43. Stimulus-Response Coupling in Metabolic Sensor Cells
735
insulin release could be enhanced. Changes in cytosolic MgATP would be expected to alter the fueling of vesicle transport processes as well as the molecular motors (e.g., NSF) that untwine paired attachment proteins (SNAREs) on granules and the target membrane. The latter permits SNAREs to subsequently intertwine with their cohorts on the complementary membrane, thereby accomplishing vesicle docking. In/3 cells, glucose is known to enhance insulin secretion evoked by elevated [K+]o even when K§ channels are already closed, while metabolic inhibition can block insulin secretion, even as cytosolic [Ca 2§ rises into the micromolar range. Experiments monitoring exocytosis from single/3 cells support the motion that the RRP is small, easily depletable, and requires constant replenishment by processes that are cAMP, MgATP and temperature dependent. Three approaches have been used. In the first approach, the cell is rapidly depolarized, in the presence versus absence of a modulator, until there is no further increase in capacitance. An estimate of the total number of granules exocytosed provides a measure of RRP, while the time to recovery of full response provides a measure of the refilling rate. Then, a rough estimate of size and dynamics of vesicle pools, in the presence versus absence of a modulatory factor, is obtained by examining the time course of release after introduction of a given concentration of Ca 2+ into cell. A final approach is to measure the capacitance increase, in the presence versus absence of the modulator, in response to instantaneous "uncaging," by flash photolysis, of Ca2+ bound to a chelator introduced into the cytosol, or "uncaging" a fraction of the modulator, now bound to a chelator, in the presence of a fixed level of cytosolic Ca2§ Sample experiments demonstrating that stimulators of cytosolic cAMP enhance initial Ca2+-entry-depen dent, depolarization-evoked release, as well as maintain release
A
with repeated depolarization, all with little attendant change in Ca 2+ entry are shown in Fig. 7A. Sample experiments demonstrating that cAMP enhances both the initial maximum rate of exocytosis and the total exocytosis in cells dialyzed against a pipette with fixed concentrations of Ca2+ are shown in Fig. 7B. More recently, similar experiments using all three approaches have been performed at varying cytosolic levels of MgATP. ATP appears to modulate RRP as well. Lastly, to be sure, anticipation of future high rates of release requires activity-stimulated, energy-dependent insulin synthesis, followed by processing in specialized membrane compartments. With prolonged continuous stimulation,/3 cells undergo several phases of insulin release; after an initial spurt of release lasting several minutes, insulin secretion wanes and then, beginning 10-15 min later, rises to a plateau that is sustained for hours. Also, islets rechallenged with glucose several hours after an initial brief bout of secretion display greater "peak" insulin release on the second round of stimulation and do this with newly synthesized insulin. Thus islets display a crude but effective form of "memory" for previous stimulation. As activity-stimulated protein synthesis can be triggered by changes in cell Ca 2+, cAMP, and ATP levels, their relative contributions to synthesis of competent new insulin granules is receiving intense scrutiny.
C. The Paradox of Stimulus-Secretion Coupling In the Glucagon-Secreting a Cell Alpha cells secrete glucagon at preprandial levels of glucose (2-3 mM); secretion is further enhanced by amino acids such as arginine and depressed by higher levels of glucose. Curiously, under some conditions, the glucagon secretion is
1.5 txM Ca 2+ 150 IxM cAMP
B
___.J
1.s ~
tree Ca
0
150
15
+ 1mMIBMX
ACm ~ 1.0, 0~
ff
I
,
,
a
I
I
I__
I
l
I
60 i20 180 240 30C)360 420 480 540 600
Time (s)
~
0.$
~"~
~1 O
FIGURE 7. Enhancement of cytosolic cAMP appears to increase the pool of readily releasable granules. Test by repetitive stimulation. (A) Note that in the "control," repetitive depolarization produces rapid rundown of capacitance response so that by the fourth depolarization there is no evidence of further net exocytosis. In contrast, after addition of membrane permeable cAMP or a phosphodiesterase inhibitor, isobutylmethylxanthine (IBMX), not only is the response to initial depolarization enhanced, but response to 4th depolarization is nearly as large as the response to the 1st depolarization in the "control". In this cell, total Ca 2+ entry over the entire series of depolarizations differed by less than 15% across the two runs. Perforated patch recording from rat /3 cell. Test by cell dialysis. (B) Cells with similar baseline capacitance (4-6 pF at break-in to the cell at time 0) were dialyzed against pipette solutions containing 1.5 pM Ca 2§ versus 1.5 pM Ca 2+ + 150 luM cAMP. Note that, in the presence of cAMP, not only was the maximum rate of rise of capacitance, abbreviated (dCm/dt) max, on average nearly 2.5-fold greater, but the total capacitance increase over the entire 6 min of recording was on average nearly 2-fold greater.
736 stimulated by tolbutamide and inhibited by diazoxide, both at concentrations that alter insulin secretion from/3 cells. So, do a cells have K+(ATP) channels and, if so, do these channels contribute to c~-cell function? In the presence of 0--3 mM glucose, a cells generate spontaneous Nao+- and Ca2o%dependent action potentials from their resting potential of---60 mV. They hyperpolarize on their exposure to increased glucose (5-20 mM); they depolarize and show enhanced spike frequency on exposure to 10 mM arginine. Under voltage clamp control, depolarization of the a-cell results in Ca2+-entry-dependent increases in cytosolic Ca 2+ and increases in membrane capacitance, the latter enhanced by agents that increase cytosolic cAMP (e.g./32 adrenergic agonists). Single-channel and whole-cell patch-clamp recording from ce cells reveals inward rectifier K + currents that have similar single-channel conductance and kinetics to those in fl cells as well as similar K d for inhibition by ATP and activation by MgADP and PIP 2. In situ hybridization identifies Kir6.2 and SUR1 subunits of characteristic of/3-cell K+(ATP) channels on glucagon-bearing cells; in fact, the densities of these subunits are even higher than those in fl cells. With apparently functional K+(ATP) channels in place, the a cell's inability to respond to glucose or metabolic inhibition is probably related to their extraordinarily high cytosolic [ATP] and [ATP]/[ADP] ratio that remains virtually unchanged after increases in extracellular glucose. (a cells are much less efficient transporters and metabolizers of glucose than are/3 cells.) In contrast, the a cell's response to arginine may be related to the depolarizing effect of the electrogenic transport of this cationic amino acid coupled to the low threshold for excitability of this cell type (i.e., low resting membrane conductance and abundance of low-voltage-activated (T-type) Ca 2+ channels). (Arginine also depolarizes /3 cells without closing K+(ATP) channels, although it usually does not enhance electrical activity unless suprathreshold concentrations of glucose are also present.) However, the crucial element of physiological regulation of the ce cell, the inhibition of its electrical activity and secretion by glucose, remain unexplained on the basis of K+(ATP) channel activity. It is possible that the inhibitory effect of glucose results largely from a paracrine interaction within the islet, a cells display a robust GABA activated C1- current, which, when stimulated, can abolish arginine-enhanced electrical activity./3 cells are known to secrete GABA along with insulin. Hence secretagogue-induced t-cell exocytosis might be the "suppressor" of a-cell activity.
D. Implications for Pathophysiology and Therapeutics Diabetes mellitus, a disease characterized by hyperglycemia and dysregulation of metabolite use and culminating in multiorgan system damage, comes in two general varieties. The insulin-dependent, ketosis-prone variety (IDDM) is usually an autoimmune disease of rapid onset, characterized by destruction of/3 cells as a result of massive assault by antibodies and cytokines. In contrast, the noninsulin-dependent, nonketosis-prone variety (NIDDM) is a spectrum of diseases characterized by increasing insulin resistance coupled with poorly timed, inadequate insulin secretion. One of its earliest indicators of this condition is the loss of the transient "first
SECTION V Synaptic Transmission and Sensory Transduction
phase" of glucose-induced insulin secretion, thought to be "primed" by acetylcholine released by the vagus nerve and GLP-1 released by enterochrommaffin cells. Accompanying this is a "compensatory" increase in insulin release during the prolonged "second phase." Unfortunately, this contributes to downregulation or "tonic tune out" of insulin receptors. Possible defects in NIDDM have been proposed for every link along the/3-cell's stimulus-secretion coupling cascade (i.e., from reduced affinity of glucose transporters to reduced efficiency of intracellular handling of Ca2+). A recently explored model for NIDDM features insulin resistance of the/3 cell with chronically decreased release of Ca 2+ from intracellular (ER) stores. No doubt, the single cell approaches outlined above will be increasingly brought to bear to dissect out cellular mechanisms underlying NIDDM. Those approaches may be expected to spur the development of more "physiologically targeted" pharmacotherapy for NIDDM, including carefully timed, pre-prandial administration of (1) GLP-1 and other gut-secreted incretins and (2) highly selective sulfonylurea-type agents of shorter onset and duration of action than those currently marketed. To be sure, two realizations, (1) that a cells contain sulfonylurea-inhibited K+(ATP) channels possibly of identical structure, and (2) that constant activation of ce cells coupled with waning of/3cell function, may contribute to the progressive sulfonylurea insensitivity, are now promoting synthesis of hypoglycemic drugs more selectively targeted at t-cell function. It is worth noting that sulfonylureas also modify the gating of some C1 conductance channels in/3 cells as well as the activity of Na+-K + ATPase and the efficacy of depolarization-secretion coupling in a manner similar to the effects of protein kinase C enhancement. As mentioned above, the NBFs of SUR resemble those of CFTR. Spare CFTR subunits appear to interact with a variety of C1 channels and even with the epithelial Na + channel (EnaC) in epithelial cells. Do spare SUR subunits likewise interact with integral membrane or membrane associated proteins in 13 cells to modulate a variety of events in stimulus-secretion coupling?
i!I. Metabolic Sensing as Protection from Hypometabolic Injury The hyperpolarization and increased K+(ATP) channel activity displayed by/3 cells exposed to metabolic inhibitors that reduce cytosolic [ATP]/[ADP] suggests that a safeguard against hypoxic damage of an excitable cell might be hypoxia-induced opening of K+(ATP) channels. This mechanism appears to apply to skeletal and cardiac myocytes and to some neurons. As in/3 cells, in these cells metabolically sensitive channels are very often K+(ATP) channels consisting of Kir and SUR subunits, though often the SUR subunit (SUR2) has much lower affinity for some sulphonylureas than does SUR1. In addition, in skeletal myocytes, which produce lactate under hypoxic conditions, the K+(ATP) channel displays pH i sensitivity that is the reverse of K+(ATP) channels in/3 cells; in these myocytes, intracellular acidosis vigorously opens K+(ATP) channels even at fixed [ATP]/[ADP]. Recent evidence suggests that in highly active regions of the brain, those cell types with higher den-
43. Stimulus-Response Coupling in Metabolic Sensor Cells sity of K+(ATP) channels show most rapid excitability block in response to hypoxia and survive it best. However, what may be an adaptive feature to individual cells may have disastrous consequences for the entire organism. In skeletal muscle, a drop in tension development by some motor units results, via spinal reflex, in recruitment of other less active motor units. More extended fatigue due to inexcitability may preserve enough ATP to maintain sarcolemmal integrity and avert leak of myoplasmic contents such as myoglobin. (The latter condition, known as rhabdomyolysis, is the chief cause of extended incapacity and systemic illnesses, such as with acute renal failure, seen after intense bouts of exercise.) However, in contrast, in cardiac ventricle, block of excitability or excitation contraction coupling in one part of the electromechanical syncitium may predispose to an arrhythmia or severely dyskinetic contraction, while in neurons, excitability block can result in a depressed level of central consciousness and even coma. As an example of hypoxia-induced excitation block, let us examine block of excitation-contraction coupling in the cardiac ventricular myocyte. Historically, this phenomenon was key to the original discovery of the K+(ATP) current. In these cells, deprivation of substrate or oxygen, or inhibition of substrate metabolism, was found to result in shortening of Ca2o+dependent plateau phase of the action potential with attendant abbreviation or block of generation of contractile force. This was found to have occurred prior to detectable changes in membrane potential, or voltage-dependent Na + and Ca 2+ currents, and it was accompanied by major augmentation of resting K + conductance. Critically, plateau phase shortening and reduced contraction were reversed by intracellular injection of MgATE As shown in Fig. 8, the key to this phenomenon is very basic. The plateau phase of the ventricular action poten-
tial represents a delicate moment-to-moment balance between two opposing tendencies. These are (i) the "depolarizing tendency" of the voltage-dependent, but slowly inactivating Ca 2§ conductance and (ii) the "repolarizing tendency" of the sum of a variety of K § currents that are either slowly activating or slowly emerging from "inward rectification" (i.e., polyamine or Mg block). Increases in background K § current provided by opening of K+(ATP) channel tip the balance in favor of repolarization, as soon as the huge inward Na + current, underlying the rapid upstroke of the action potential, wanes. However, shortening of the broadly propagating action potential may reduce the refractory period and promote local impulse re-entry or rebound excitation. In addition, the period of reoxygenation may enhance arrhythmogenicity by promoting large transient inward currents (lti) which result in spontaneous depolarizations. Hence, it would appear to be better for the ventricle to anticipate rather than react to metabolic deprivation. Physiologically, this is exactly what happens: ventricular hypoxia results in dilation of coronary arterioles leading to greater supply of oxygen and metabolites (see Section V.A). However, as often happens in critical physiological processes, there is yet another factor at work. Short bouts of hypoxia, with opening of K+(ATP)channels, may actually "pre-condition" myocytes and ameliorate some of the effects of subsequent, more prolonged bouts. In that case, does prevention of hypoxia-induced opening of K§ channels, as might occur in the presence of a sulfonylurea, predispose to hypoxia-induced cell injury? Controversy still surrounds this issue. It is worth mentioning other mechanisms that might contribute to ischemia-induced contractile failure. These include (i) altered [Ca2+]i "homeostasis," (ii) changes in the binding affinity of the contractile apparatus for Ca2+ induced by accumulation of acid equivalents and (iii) activation of other K § channels, such as muscarinic and G-protein gated K + channels, by metabolic intermediates generated during hypoxia.
IV. Stimulus-Secretion Coupling in Carotid Chemoreceptor Cells
o V~ (mY)
-85
Gca~+
737
-~-:
"--..__~2~..-
~
GK§ GK+(ATP)
Tension I
250 ms
I
FIGURE 8. K+(ATP)channels and cardiac excitability. A steadystate enhancement of K+(ATP) conductance (GK+(ATp)),resulting from metabolic inhibition, shortens the time courses of the ventricular action potential and twitch tension and alters the time courses and magnitudes of the underlying ionic conductances. Solid lines indicate control; dashed lines indicate effects of metabolic inhibition.
Neural control of respiratory drive is critical in maintaining physiologic levels of plasma 02, CO 2 and H +. In the rhythmic, neuromuscular process of breathing, trains of action potentials, generated by pacemaker cells in the CNS, trigger motor neurons and activate periodic contractions of the inspiratory muscles (diaphragm and intercostal muscles). This action expands the chest, thereby sucking air of high-O 2, low-CO 2 content into the lungs. Gas exchange consists of net CO 2 diffusion from capillary to adjacent lung air spaces (alveoli) and net O 2 diffusion from alveoli to the capillary. Relaxation of inspiratory muscles, sometimes combined with the contraction of expiratory muscles, expels the air of high CO 2, low O 2 content from the lungs. A major stimulus for altering the pattern of respiratory drive is a drop in the partial pressure of O 2 dissolved in plasma (i.e., plasma PO2); a secondary stimulus is a rise in plasma pCO 2 or a fall in pH. The carotid body is the organ that senses changes in plasma pO 2, pCO 2, and pH, and mediates changes in CNS respiratory drive. Located at the bifurcation of the carotid artery, a branch of the aorta, the carotid body consists of a central core of chemoreceptor (or glomus) cells;
738
SECTION V Synaptic Transmission and Sensory Transduction
these originate from the neural crest. Glomus cells synapse on dendritic endings of sensory nerve fibers that comprise the carotid sinus nerve travelling into the CNS. The glomus cell core is surrounded by more superficial glial, or sustentacular, cells, as well as by a dense network of highly porous capillaries. The glomus cell is the site of transduction of changes in plasma pO 2, pCO 2, and pH into changes in electrical activity of the afferent carotid sinus nerve. Decreases in pO 2, as well increases in pCO 2, or decreases in pH, lead to an increase in release of dopamine and probably to an increase in as-yet-unidentiffed transmitters from glomus cells, as well as to an increased action potential frequency in the dopamine-sensitive carotid sinus nerve. This contributes to increased respiratory drive. A. Role of O2-Sensitive K + C h a n n e l s in T r a n s d u c t i o n of the Hypoxic S t i m u l u s
Two divergent views have emerged conceming chemotransduction of hypoxia by glomus cells (see Fig. 9A). The first theory proposed that hypoxia induces cell depolarization, followed by opening of voltage-gated Ca 2§ channels, Ca 2§ influx, and synchronized exocytotic release of dopamine. This might give rise to sufficiently large excitatory post-synaptic potentials in the dendritic regions of single sinus nerve fibers to trigger propagating action potential. An alternative view proposed that hypoxia induces quantal release of transmitter in a Ca2+-depen dent manner that is independent of voltage-gated Ca 2+ entry but dependent on the discharge of Ca 2+ from intracellular stores. The latter should increase asynchronous quantal release of transmitter, generating a rise in the frequency of miniature excitatory post-synaptic potentials, and should enhance ongoing
A
/1~K+
electrical activity in the low-threshold carotid sinus nerve fiber. These contrasting viewpoints have arisen from data generated with different preparations from different species. There are several lines of evidence supporting the classical scheme for "depolarization-secretion coupling" in glomus cells. First, glomi are electrically excitable sensory cells and contain HVA Ca 2+ currents, delayed rectifier K + currents, and in some cases voltage-dependent Na + currents. Second, in some isolated cell preparations, glomus cells respond to hypoxia by depolarizing and generating action potentials. (However, in other preparations, glomus cells fire action potentials only in response to current injection or release of the cell from sustained hyperpolarization. These cells show no change in passive electrical activity in response to hypoxia.) Third, in glomus cells that fire in response to hypoxia, depolarization increases both cytosolic Ca 2+, measured with intracellular dyes, and dopamine release, measured by amperometry. In these cells, depolarization-secretion coupling is reduced by exposure to blockers of the HVA-type Ca 2+ channels or by reduction in extracellular Ca 2+. A very exciting development in chemoreceptor physiology consistent with the depolarization-secretion coupling scheme outlined above is the discovery that, in whole-cell recordings, lowering ambient pO 2 selectively and reversibly reduces the outward K + current flowing through delayed rectifier K+(DR) channels of the glomus cells. Reducing pO 2 from 160 mm Hg to 90 mm Hg reduces peak K+(DR) current by -30%. This effect is not dependent on the concentrations of ATP (0-3 mM) or Ca 2+ ( 100
750
SECTION V Synaptic Transmission and Sensory Transduction
per ixm 2) have never been reported for MS channels. A cell may have only one type of MS channel--for example, the SA cation channels in X e n o p u s oocytes (Hamill and McBride, 1992)--or they may have multiple types, as indicated in Table 2. Neither excision of a patch nor removal of all cytoplasmic-side calcium abolishes the mechanosensitivity of MS channels (e.g., Vandorpe et al., 1994). In fact, the more stripped-down and variously traumatized a membrane patch becomes, the more readily some MS channels are activated by a standard stimulus (Wan et al., 1999), suggesting that tension applied to the naked bilayer is the most effective stimulus. T h e r e a r e nonselective and cation-selective (i.e., monoand divalents) as well as K + and C1- or anion-selective MS channels (see Table 2). Most MS channels are stretch activated (SA) and a few are stretch inactivated (SI). The MS activity of some channels depends on the direction of membrane curvature (Bowman et al., 1992; Marchenko and Sage, 1997; Maingret et al., 1999). From their diversity, it is evident that MS channels belong to many different channel families. Many are activated by a variety of stimuli other than mechanical stress (Table 3), though some are known only from their MS gating (e.g., the SA cation channel in X e n o p u s oocytes). The champion so far, with four types of activation, is the smooth muscle K + channel reported by Kirber and colleagues (1992), although this channel may really be a ligand-gated channel in disguise. Fatty acids, liberated from the membrane by stretch, appear to be the mediator of stretch activation for this channel (Ordway
et al., 1995). This possibility needs to be considered more widely. A variety of molecularly identified (cloned) channels are now known to exhibit MS gating at the single level. Even a cursory glance at this collection reveals that the requirements for MS gating must be rather catholic. The cloned channels include MscL, the nonselective channel of nanoseimen conductance, from E. coli (Sukharev et al., 1994); the NMDA channel, a glutamate-activated cation channel (Paoletti and Ascher, 1994); smooth muscle Ca2+-activated K + channels (Dopico et al., 1994); a G-protein-regulated K+-permeant inward rectifier, GIRK (Pleusamran and Kim, 1995; Ji et al., 1998); and, most recently, several members of the TREKlike family (Maingret et al., 1999). Of these, only MscL was cloned because of its mechanosensitivity. McsL forms homomultimers, probably homopentamers (Chang et al., 1998), but bears no sequence resemblance to either the heteropentameric NMDA channel, the heterotetrameric GIRK, or the (presumably) dimeric TREK-like channels. MS channels are defined by their responses to membrane deformation but not all are equally responsive. Dynamic ranges of several orders of magnitude are not uncommon for SA channels in, say, molluscan neurons, chick muscle, or Xenopus oocytes (Morris and Sigurdson, 1989; Guharay and Sachs, 1984; Hamill and McBride, 1992); this compares to the dynamic ranges of voltage- and ligand-gated channels. For some MS channels, however, mechanical stimulation only slightly changes open probability. For example, the NMDA-type glutamate channel increases its open probability
TABLE 2 Cells with Multiple Types of MS Channel
Cell
Channel(s)
Reference
Amphibian kidney cells
All demonstrable K+ channels in the basolateral membrane of amphibian proximal tubule cells are mechanosensitive
Cemerikic and Sackin, 1993
Molluscan heart
(Lymnaea stagnalis) SA K+ and (zero extracellular Ca2+only) SA Na+ channels
Morris, 1990; Gardiner and Brezden, 1990
Molluscan neurons
Lymnaea stagnalis SA K§ and SI K§ channels; the latter are less common and have a smaller conductance Cepaea nemoralis SA K+ channels and (excised patches only) SA C1- channels
Morris and Sigurdson, 1989 Bedard et al., 1992
Amphibian smooth muscle
SA cation and SA K§ channels
Kirber et al., 1992
Chick heart cells
5 distinct SA channels (3 K+-selective, 2 cationselective)
Ruknudin et al., 1993
Mammalian outer hair cells
The lateral walls of guinea pig outer hair cells have a SA cation channel and one that appears K+-selective
Ding et al., 1992
Mammalian osteoblast-like cells
SA K+ channel, 60-pS SA cation channel, 20-pS SA channel
Davidson et al., 1990
Crustacean stretch receptor neurons
SA cation channels, rectifying SA cation channels
Erxleben, 1989
E. coli
Several molecularly distinct nonselective conductance channels
Sukharev et al., 1994
44. Mechanosensitive Ion Channels in Eukaryotic Cells
751
TABLE 3 MS Channels Whose Gating Is Also Controlled by Other Factors Channel
Factor(s)
Reference(s)
Avian SA cation channel
Depolarization
Sachs, 1992
Amphibian smooth muscle SA K+ channel
Ca2+, fatty acids, and membrane depolarization
Kirber et al., 1992
Amphibian smooth muscle SA cation channel
Hyperpolarization
Hisada et al., 1991
Crayfish rectifying SA cation channel
Hyperpolarization
Erxleben, 1989
E. coli MS channel
Depolarization
Martinac, 1993
Mammalian atrial SA K+ channels
Arachidonic acid and depolarization ATP-inhibited
Kim, 1992 Van Wagoner and Russo, 1992; but see Kim, 1992
Liver SA cation channel
ATP-activated (via purinergic receptors)
Bear and Li, 1991
Osteosarcoma SA cation channel
Parathyroid hormone via cytoplasmic messengers increases open probability and single-channel conductance
Duncan et al., 1992
Aplysia S(erotonin)-channel
Stretch and (via FMRFamide receptor pathway) arachidonic acid metabolites
Vandorpe et al., 1994; see Patel et al., 1998
Mammalian neuron NMDA channela
Potentiated by mechanical stimuli
Paoletti and Ascher, 1994
Cardiac GIRKa
Enhanced by membrane stretch
Pleusamran and Kim, 1995
aThese channels are only known to be mechanosensitive when they are ligand bound.
approximately three-fold in response to pipette suction (Paoletti and Ascher, 1994), and the bovine epithelial Na § channel ENaC only about two-fold in bilayers and not at all when expressed in Xenopus oocytes (Awayda et al., 1998). Ideally one would reserve the term "mechanosensitive" for channels that recognize mechanical deformation as a physiological signal. Then, whether the dynamic range was 2-3 or 2-3-orders-of-magnitude, the channel would qualify as "MS" only if mechanosensitivity mattered in the life of the cell. Various MS channels can be activated via the ligands of second messenger systems as well as by membrane stretch. They include the K+-selective Aplysia S-channel (Vandorpe and Morris, 1992), their mammalian relatives TREK-1 and TRAAK (Maingret et al., 1999), and cation channels in osteoblasts (Duncan et al., 1992), hepatocytes (Bear and Li, 1991), and kidney cells (Verrey et al., 1995). Physiologically, the Aplysia S-channel (Vandorpe et al., 1994) is activated by arachidonic acid metabolites and inhibited by an A-kinase, ligands controlled via the neurotransmitters FMRFamide and serotonin, respectively. This voltage-independent K § channel's physiological job is to modulate electrical excitability according to neurotransmitter signalling. Under patch-clamp conditions (cell-attached or excised), however, the S-channel can be activated by stretch alone (see Fig. 1B); in fact, the maximal effect of stretch on a given patch considerably exceeds activation via second messenger paths (Vandorpe et
al., 1994). MS K + channels are ubiquitous in neurons of Aplysia and other molluscs (Bedard et al., 1992; Morris, 1992), but only in the identified Aplysia neurons are the signaling molecules, which include arachidonic acid, known. In situ the channels do not readily activate with mechanical stimuli (Morris and Horn, 1991; Wan et al., 1995), yet "SA K + channel" was until 1998 (see Patel et al., 1998) the best designation available for these "S-like" channels. Many other channels may have been designated as "MS channels" simply because the primary physiological stimulus, perhaps a second messenger ligand, has not been identified. The Xenopus oocyte MS channel (Steffensen et al., 1991) may be a case in point. For all MS channels, tension must be conveyed to the channel via the surrounding lipids, membrane skeleton, extracellular matrix elements, or some combination. Although these structures are universal, suction does not affect gating in all channels. The classical example is the nicotinic acetylcholine channel of chick skeletal muscle. This is a cation channel with very similar permeation properties to a cationselective SA channel in the same membrane (and a heteropentameric structure not unlike the NMDA channel), but the acetylcholine channel's kinetics are not affected by stretch (Guharay and Sachs, 1984). Mechanotransduction at membranes need not be mediated by MS ion channels. Recently, it was shown that ex-
752
SECTION V Synaptic Transmission and Sensory Transduction
pression of a heterologous metabotropic ATP receptor confers mechanosensitivity on Xenopus oocytes. The explanation is that oocyte deformation releases ATP, which then acts on the receptor (Nakamura and Strittmatter, 1996). Here, the mechanical event that needs explanation is the stretch-induced release of ATP.
V. MS Channels in Patches: Stretch Versus D a m a g e Plus Stretch Because MS channels are almost ubiquitous, the question of whether their mechanosensitivity is meaningful or merely epiphenomenal is not trivial. If, in situ, many MS channels are not used for cellular mechanotransduction, why is MS gating in membrane patches so common? Perhaps for many channels anchored to the membrane skeleton, spurious (essentially pathological) MS gating can occur when the integrity of the cortical cytoskeleton in the channel's immediate environs is inadequate (Table 4). The assumption here is that an intact cortex absorbs mechanical loads that would otherwise be felt by the membrane skeleton and by attached channels. MS channels designed for use as physiological mechanotransducers would, by contrast, require an environs specifically constructed to enhance mechanical loading of the channel via the membrane skeleton or bilayer. Patch recording is inevitably accompanied by mechanical damage to the patched membrane region. Small and Morris (1994) showed that in molluscan neurons, starting from a near-intact patch, successive rounds of mechanical stimulation augment the responsiveness of MS channels to the mechanical stimuli. "Stimulation" means "aspiration of the membrane and the underlying cytoskeleton," and it would be specious to argue that this is entirely benign. The ability of patch-induced damage to facilitate sustained MS channel responses may explain much of the widespread discordance between single-channel and whole-cell recordings (Morris and Horn, 1991; Wan et al., 1999). In Xenopus oocytes, too, MS channel behavior is damage dependent: the rapid adaptation behavior is irreversibly destroyed by patch trauma (i.e., by successive stimuli). Interestingly,
TABLE 4
pore (a) subunits of voltage-gated Na +~channels expressed (in the absence of their auxiliary /3-subunit) in Xenopus oocytes exhibit a physiologically nonstandard slowly inactivating gating mode that, by some unknown mechanism, irreversibly converts during membrane stretch into kinetically normal fast inactivation (Tabarean et al., 1999). This irreversible effect probably says something important about the susceptibility of protein complexes to bilayer tension, but should not be taken to mean that Na § channels are mechanotransducers. What then, is the meaning of channel mechanosensitivity? There is no unique answer, but an encapsulation of my viewpoint is: many eukaryotic channels are inherently mechanosusceptible to changing their transition rates among open and closed states (either reversibly or irreversibly) when mechanical stimuli increase tension in the plane of the bilayer. The channels' complex cellular environment can either act as a restraint, absorbing mechanical energy so the channel can attend to other stimuli, or it can "tune-up" the inherent mechanosusceptibility to yield a mechanotransducer of membrane tension. For channels that are not mechanotransducer specialists, mechanosusceptibility may vary as physiological and pathological conditions alter the membrane skeleton (e.g., during cell swelling, cell division, cell growth, ischemia, mechanical trauma, and nociceptive stimulation). By contrast, mechanotransducer specialist channels such as those at the tips of auditory hair cells may prove to be exceedingly insensitive to bilayer tension at all times, thus ensuring that they "listen" to and are gated exclusively by the relevant tension, found in their specialized accessory structures (e.g., tip links). Whether or not it is used by the cell, mechanosusceptibility is an inherent trait in diverse integral membrane proteins. Should we marvel that this trait evolved independently as a "selected-for" feature in each protein or should we surmise that a tendency for mechanical perturbations to trigger gating transitions is, instead, a difficult to suppress trait for membrane-spanning enzymes? I favor the latter, but both postulates are valid, and deeper understanding will require molecular studies of eukaryotic channels that exhibit physiologically relevant mechanosensitivity as well
Trauma, Pathology, Nociception, and MS Channel A c t i v i t y
Dystrophic muscle
Lacks membrane skeleton dystrophin so spectrin may experience abnormal stress which may be transferred to
channels linked to spectrin. Dystrophic cells are fragile and leaky to calcium. MS cation channels in dystrophic muscle exhibit abnormal MS gating (Franco and Lansman, 1990). There may be a causal relationship. Brain ischemia
In the presence of agonist, NMDA-gated channels are mechanosensitive. In swelling neurons, NMDA channels generate whole-cell currents that grow larger as swelling progresses (Paoletti and Ascher, 1994). In stroke, this could exacerbate glutamate excitotoxicity.
Nociception and mechanical trauma
Neuronal SA channels are easier to activate following trauma (Small and Morris, 1994). Osteoclast MS channel mechanosensitivity is enhanced in chronically and intermittently strained cells (Duncan and Hruska, 1994). Sustained rather than transient responses of MS cation channels predominate after membrane trauma (Hamill and McBride, 1992), a phenomenon that could potentially be exploited for nociceptive signaling.
Ectopic foci in demyelinated neurons
Hypermechanosensitive (see references in Waxman, et al., 1994). Abnormalities in membrane and cortex may increase the mechanosensitivity of normally refractory channels.
753
44. Mechanosensitive Ion Channels in Eukaryotic Cells
as those that appear to be mechanosusceptible in spite of having no evident role as mechanotransducers.
VI. The Role of the Membrane Skeleton When MS channels were first described (Guharay and Sachs, 1984) it was clearly demonstrated that neither tubulin nor F-actin was required for the MS gating. It was postulated that MS channels are linked into the membrane skeletal network and that when the in-parallel actin cytoskeleton is depolymerized, stress is readily transferred to the channel-linked component. What that parallel component may be has not been resolved, although it is not dystrophin since MS channels are evident in dystrophic muscle cells (Franco and Lansman, 1990). Spectrin is a candidate that has not been ruled out. For eukaryotic systems, the bilayer may seldom experience significant stress except near lysis. Supporting this notion is the fact that the elastic constant of patches (Sokabe et al., 1991) is much less than for lipid membranes. The bilayer is not, therefore, the patch's load-bearing element. That leaves the cortical cytoskeleton and the membrane skeleton. If it is generally true that the elastic constant of patches is insensitive to the actin disrupter cytochalasin (Sokabe et al., 1991), that leaves, by further elimination, a cytochalasin-insensitive membrane skeletal network as the only candidate. In an intact cell, one would expect cortical actin to contribute to the elasticity of the membrane-cortex region. In patches, the fact that it does not suggests that the act of patch formation disrupts the cortical actin network, consistent with the finding of Small and Morris (1994) who showed that cytochalasin, like mechanical damage, augments rather than diminishes the responsiveness of SA channels to membrane stretch. Other membrane skeletal filaments like spectrin may constitute the load-bearing elastic elements that transfer mechanical energy to the channels. A variety of well-characterized membrane channels (and other transporters) are tethered to the membrane skeleton, either to /3-spectrin via ankyrin (Lambert and Bennett, 1993) or directly to c~-spectrin (Rotin et al., 1994). Additionally, some voltage- and ligand-gated channels are anchored to postsynaptic density-family proteins (Kim et al., 1995). Although we surmise that eukaryotic MS channels receive mechanical energy via the membrane skeleton, there are no grounds for turning the argument around and invoking channel mechanosensitivity as a reason for linkages. Transporters may anchor to the membrane skeleton in order to stay fixed at some cellular locale. In other cases, as in erythrocytes, the transporters may serve as anchor points for a membrane skeleton whose major job is to strengthen the membrane. It is probably counterproductive for many anchored membrane enzymes to allow their activity to be influenced by mechanical stress. Thus, in asking how and if specialized channel-membrane skeleton linkages might be used by MS channels to convey force to the channel, we should also remember to ask the reverse question: what features can make such linkages stressproof?. (See Glogauer et al., 1998.)
Vil. Delay and Adaptation: Mechanically Fragile Aspects of MS Channel Behavior The mechanical properties of patches are labile and history dependent (Hamill and McBride, 1997; Small and Morris, 1994). In fibroblasts stretch causes F-actin to depolymerize within 10 s, although when the cells are left to rest briefly, F-actin reorganizes, reappearing at greater than control levels, stiffening the cortex as it does so (see Glogauer et al., 1998). Patch formation and stretch stimulation of cells may have qualitatively similar influences on the state of actin. Thixotropic properties of actin, in which mechanical stimulation of actin gels promotes the sol-state of actin (see Heidemann and Buxbaum, 1994) may affect MS channel mechanosensitivity. In molluscan neurons and in Xenopus oocytes (Small and Morris, 1994; Hamill and McBride, 1992; Wan et al., 1999), the behavior of SA channels following a step of applied suction has been studied in patches formed with a minimum of mechanical disruption--gentle patches. The responses of MS channels in the gentle patches differ dramatically from those in standard patches. The designations "gentle" and "standard" patch are operational; gentle patches inevitably become progressively more standard with successive mechanical stimuli. In gentle oocyte patches, SA channels exhibit rapid adaptation; adaptation is abolished in standard patches. In gentle snail neuron patches, S A channels show a prolonged delay before activation; delay is abolished in standard patches. Changes in patch size and changes in cortical cytoskeleton integrity are probably not mutually exclusive, so the interpretation of input-output relations for MS channels can be fraught with difficulty. However, the mechanical fragility of the two dynamic phenomena, adaptation and delay, cannot be explained by the increases in patch size that may accompany a patch's progressive transformation from gentle to standard. As indicated, molluscan neuron SA channel activation in response to a first hit (a first large step of suction applied to a gentle patch) proceeds only after a delay, as if the patch were viscoelastic. The delay is substantial (> 2 s a t - 130 mm Hg) and changes produced by treatments are easily detected. Treatments that dramatically decrease delay include repetition of mechanical stimulation, use of larger mechanical stimuli, pretreatment with cytochalasin (a drug that promotes actin depolymerization), use of recently isolated (i.e., recently disrupted) as opposed to well-established cultured neurons, hyposmotic swelling of neurons, exposure to the sulfhydryl reagent N-ethylmaleimide, and elevation of intracellular Ca 2+ (Small and Morris, 1994; Wan et al., 1999). In addition to shortening delay, these treatments increase the activation elicited for the same stimulus. In summary, the response time is long and the extent of activation is small when the cortical cytoskeleton is intact, but not when it is compromised. It is reasonable to conclude (Wan et al., 1999) that the channels are not functionally mechanosensitive in situ when the cortex is perfectly intact. Mechanical inflation of neurons under whole-cell clamp does not activate MS current except just before the cell rapidly expands and ruptures; a catastrophic disruption of the cortex
754
SECTION V Synaptic Transmission and Sensory Transduction
could explain both the MS currents and the ensuing membrane rupture (Wan et al., 1995) (Fig. 5). Interestingly, this echoes the situation described earlier for E. coli MS channels. The fragility of MS channel dynamic responses can cause consternation and confusion but it also opens the possibility that substantial modification of MS responses may be possible through hormones that modify the cytoskeleton and by cell motility-related and volume regulation-related changes in the cytoskeleton (Schweibert et al., 1994). For example, vasopressin causes both actin depolymerization and SA channel activation in kidney cells (Verrey et al., 1995). Another example is during the cell cycle (Bregestovski et al., 1992), when major cytoskeletal rearrangements may partly account for the observation that SA channel sensitivity varies during the cycle. Since the formation of focal adhesions is accompanied by major cytoskeletal rearrangements, it seems likely that MS channels should show altered mechanosensitivity in connection with changed states of cell adhesivity. Hints that this might be so are beginning to accumulate. Macrophages have a MS K + channel whose activity, monitored by single-channel recording at the upper surface of the cell, increases when the lower surface of the cell becomes adherent to a substrate (Martin et al., 1995). MS channel currents in osteoblasts subjected to chronic intermittent strain are more readily activated than those in quiescent cells (Duncan and Hruska, 1994). It is, thus, possible that the mechanical fragility of delay is telling us about variable in situ states of MS channels. But what about adaptation? Here there is, as yet, little insight. In phasic mechanosensory neurons, the purpose of rapid adaptation is to provide a mechanical high-pass filter, but what it might mean for amphibian oocytes, tunicate eggs, and yeast MS channels is a mystery. By contrast, the other dynamic property, delay, could in principle act like a low-pass filter; it could be useful for, say, muscle or bone cells MS channels, enabling them to ignore momentary stresses, yet inform the cell of sustained mechanical inputs and hence modify some aspect of its metabolism (volume regulation, Ca 2+ loading, protein synthesis).
VIII. P h y s i o l o g y of M S C h a n n e l s What establishes that a putative mechanotransducer channel performs a MS physiological (or developmental) task? Suggestions about functions of the channels are legion, but experimental evidence conclusively linking MS channel activity to physiological processes is rare. A dearth of specific pharmacological tools continues to hinder progress, as does the difficulty of mechanically stimulating membrane in a calibrated and reproducible manner. Finding MS channels in patches does not mean that MS currents can readily be recorded from the channels in the intact cell.
A. Criteria for Establishing that a MS Channel Performs a Mechanical Cellular Task The criteria that need to be met are comparable to those that have been met by ligand-gated and voltage-gated channels with established physiological roles. The following criteria summarize a list given by Morris (1992). 1. Establish the MS channel's biophysical characteristics by recording at the single-channel level. 2. Establish a way of altering channel function (e.g., a channel blocker or a mutant channel). 3. Demonstrate that some physiological or developmental aspect of the cell is MS. 4. Demonstrate that the MS cellular function is specifically impaired by the channel blocker (or by the mutation). 5. Make macroscopic recordings under conditions of minimal damage to the membrane-cortex, and demonstrate MS currents that correspond to the single-channel recordings according to (a) selectivity; (b) pharmacology; (c) noise characteristics; (d) saturability at large mechanical stimuli; (e) the expected magnitude of saturating macroscopic current given the singlechannel membrane density; (f) expected MS current density variations--spatial and/or temporal--as predicted from any spatial or temporal variations noted by
FIGURE 5. Inflationof molluscan neurons under whole-cell clamp rarely elicits MS K+ currents in spite of the abundant MS channels. Macroscopic MS currents are observed only at pressures marginally less than those that rupture the plasma membrane. The left and middle graphs are ramp-clamp currents for the stated pressures (applied via the recording pipette as shown). As expected by Laplace's law, large cells can withstand only small pressures. In the same neurons, K+ currents are elicited by osmotic swelling (fight, a difference current obtained by subtracting ramp currents before swelling from those during swelling). Are these volume-activated currents the direct result of swelling (mechanical) stress? One cannot say, since swelling causes an increase in membrane capacitance (possibly recruiting new channels) and alters cytoplasmic chemistry (possibly stimulating channel activity via ligand changes) as well as increasing tension. (Modified from Wan et al., 1995.)
755
44. Mechanosensitive Ion Channels in Eukaryotic Cells
single-channel recording; (g) degree of mechanosensitivity; and (h) sensitivity to other known physiological influences (e.g., hormones, intracellular ions).
B. A S! Cation Channel and Hypothalamic Osmotransduction For one MS channel, many of the criteria listed above have been addressed. A subpopulation of neurons in the hypothalamus detects and responds to blood osmolarity. These magnoceUular osmosensory neurons (OSNs) possess MS cation channels that may underlie the osmotransduction currents that hyperpolarize OSNs during hyposmotic stimuli and depolarize them during hyperosmotic stimuli (see Fig. 4A, B). This in turn leads to respective decreases and increases in release of the osmoregulatory hormone vasopressin from the OSNs. Under cell-attached recording conditions with physiological solutions, the --30 pS OSN channels reverse n e a r - 4 0 mV, sufficiently depolarized with respect to resting potential for active channels to be excitatory. The idea that MS channels mediate osmotransduction currents is supported by the demonstration that gadolinium ions block both the macroscopic osmosensor currents and the single-channel MS currents. Activation curves for the OSN single-channel currents in cell-attached patches are bell-shaped rather than the sigmoid curve characteristic of voltage-gated channels and of other MS channels (see Fig. 4C). This is not inconsequential; wellestablished physical explanations exist for sigmoid activation curves ("Boltzmanns"), but there is no precedent for "Gaussians" when it comes to activation curves. Does the bell shape reflect an inherent peculiarity of the channels? It may not. Rather, the unusual input-output relation may stem from an acknowledged systematic error in assessing the true pipette pressure. Although seeming to be data obtained at applied pressures symmetric about zero (see Fig. 4C, middle curve), the activation curves were obtained entirely with suction (negative pressures) (see Fig. 4C, fight); the published bell curve peaks at "zero" only because an arithmetic offset was applied. Macroscopic currents from the same cells show a monotonic dependence on hypotonicity that fits well to a sigmoid (see Fig. 4B). If both single-channel and intact-cell currents reflect activity of the same channels, what explains the dramatic qualitative difference in input-output relations? To answer, we need to visualize how applied tensions in patch recordings would correlate with osmotic stimuli for whole-cell recording. With whole-cell data plotted on an x-axis going from lowto-high osmotically induced membrane stretch (as in Fig. 4B), the Oliet and Bourque data fit a sigmoid Boltzmann relationship (akin to Hodgkin-Huxley h ) as expected for the intrinsic tension-activity relations of SI channels (Morris and Sigurdson, 1989; Lecar and Morris, 1993). The expected sigmoid can be found in the single-channel bell curve (see Fig. 4C, right) if the following two assumptions hold: (a) the MS channel is indeed a SI channel and (b) MS channel activity was recorded with net positive pressure in the pipette, even though suction was applied. These assumptions may be incorrect but are not strictly posthoc rationalizations--they have a precedent. For snail neuron SI channels (Morris and Sigurdson, 1989), sigmoids are more
common but bell-shaped activation curves are obtained if the appropriate experimental pressure range is used. The bell corresponds to the expected hoo type sigmoid plus its mirror image reflected on the zero-tension point of the pressure axis. The xaxis is the estimated experimental pressure (see Fig. 4C), not membrane tension; neither the true zero pressure nor tension is known. The authors contend that the bell-shape "illusion'" arises because the membrane patch is sealed firmly in the electrode rim, and the patch is concave, flat, or convex depending on the effective pipette pressure. Tension sufficient to gate the SI channels can be achieved in both the concavely and convexly stretched patches. If residual pressure indeed persists in the pipette after gigaseal formation, the bell-shaped curve will center not at the origin but elsewhere along the axis of experimentally applied pressure, according to the sign and magnitude of the residual pressure (see Fig. 4C). To summarize, the basis for the bell shape is critical. If the bell arises from the stimulus mechanics of patches (a bell is the SI channel congener to the U-shape of SA channels over the full stimulus range (pressures positive and negative to zero; see Sachs and Morris, 1998)) and if all the single-channel OSN channel recordings were indeed obtained starting with positive residual (not zero) pressure in the pipette, then the channel is indeed a SI channel and the microscopic and macroscopic input-output relations are not at odds.
C. Other Roles Proposed for MS Channels This section quotes from a potpourri of papers in which mammalian or avian MS channels were tentatively linked to a wide range of cellular phenomena. The microscopic to macroscopic correspondences have yet to be established. There is a pressing need to determine if various MS channels perform in situ the mechanical tasks suggested for them based on their single-channel behavior. 9 Bear, 1990: "It is proposed that stretch-activated channels in liver cells permit the transient influx of Ca 2§ which in turn acts to trigger changes in ion conductance or cytoskeletal components involved in cell volume regulation." (See Bear and Li, 1991.) 9 Suleymanian et al., 1995: "Stretch-activated channel blockers modulate cell volume in cardiac ventricular myocytes . . . support[ing] the idea that [the channels] are involved in cardiac cell volume regulation." 9 Puro, 1991: "Stretch-activated channels may help mediate a compensatory response of gila to s w e l l i n g . . . [because] activation of calcium-permeable stretchsensitive channels is associated with an increase in the activity of calcium-activated potassium channels. Activation of potassium channels to produce an efflux of potassium with a subsequent l o s s . . , of cell water c o u l d . . , decrease glial cell volume." 9 Hansen et al., 1991: "Our results indirectly implicate stretch-activated channels in the genesis of stretchinduced [cardiac] arrhythmias and provide preliminary evidence for a potential new mode of antiarrhythmic drug actionmblockade of stretch-activated channels." 9 Kent et al., 1989: "sodium entry [--an initial cellular response requisite to the growth-inducing activity of many substances--] through stretch-activated ion
756
9
9
9
9
9
SECTION V Synaptic Transmission and Sensory Transduction
channels is stimulated by deformation of the sarcolemma . . . . Streptomycin, a cationic blocker of the mechanotransducer ion channels . . . inhibited, in a dose-dependent manner, [load-dependent] protein synthesis otherwise observed in contracting [myocardial] muscles developing tension." Ding et al., 1992: "SA channels . . . could affect the motile response [of outer hair cells, via] membrane potential or by allowing the entry of free Ca 2+ which could lead to [cellular length changes via] actin and myosin. SA channels could a l s o . . . [regulate the cell's osmotic pressure] thereby influencing its electroosmotic response." Iwasa et al., 1991: In auditory outer hair cells, "stretchactivated channels may play an important role in producing a mechanical feedback, an indispensible element in cochlear tuning." Coleman and Parkington 1992: "Propagation of electrical and mechanical activity in uterine smooth muscle: a functional role for stretch-sensitive channels." Davidson et al., 1990: "We propose that one or more of these [three] mechanosensitive ion channels [in osteoblast-like cells] is involved in the response of bone to mechanical loading." Sigurdson et al., 1992: "Gently prodding [tissue-cultured chick heart] cells with a pipette produced a Ca 2+ influx that often led to waves of calcium-induced calcium release spreading from the site of stimulation . . . . The mechanical sensitivity probably arose from stretch-activated ion channels."
IX. Other Explorations of MS C h a n n e l s A few widely ranging examples are used to provide some flavor of potentially exciting findings. In each case, the question testifies to the need for more information about the biological impact of the readily observed mechanosensitivity. (1) Fungal germling of bean rust: topographical sensing? Rust is a parasite whose germling needs to find the stomatal openings of host-plant leaves, a feat it achieves by sensing minute topographical features. Kung and colleagues (Zhou et al., 1991 ) showed that germ tube protoplasts of rust have SA channels evident in both macroscopic and single-channel recordings. Saturating MS currents were readily obtained in both types of recording. Moreover, gadolinium blocks both single-channel and whole-cell currents in protoplasts and inhibits germ tube growth and differentiation in vivo. The rust channels are nonselective and have a very high conductance (> 0.5 nS), so that channel openings could profoundly alter the ionic composition of the minute volume of cytoplasm in the growing tips. Presumably fine control could be achieved in vivo by having exceedingly brief openings. (2) Hyphal tips in fungi: MS channels and development? On the basis of numbers of SA channels per patch, Levina and colleagues (1994) report variations in the density of the channels toward hyphal tips in fungi. Based on the tip-high staining pattern of filamentous actin, and on data with cytochalasin, the authors suggest that actin causes
the channels to cluster, either by linking channels to the cytoskeleton or by increasing transport of channels to the tip. Alternately, regional variations in membrane mechanics may explain the data; channels may be present at uniform density but they may be less mechanosensitive in "old" membrane far from the tip and easy to stimulate in "new" membrane near the tip. This would echo the case of neuronal cytochalasin-sensitive SA channels which become increasingly difficult to activate as the cell culture matures (Small and Morris, 1994). (3) Fish embryo: a role in the cell cycle? SA K § channels in loach embryos (2-256 cells) show dramatic changes in activity during the cell cleavage cycle, perhaps explaining the membrane voltage oscillations. At the beginning of interphase, channels tend to be inactive and less responsive to stretch, whereas at prometaphase they are highly active. Evidence is presented showing that the activity of these channels is regulated through cAMP-dependent phosphorylation (Medina and Bregestovski, 1991). (4) Mammalian neurons: a role in neuronal development? The SA K § in fish embryos may well be related to both molluscan growth cone SA K + channels (Morris and Sigurdson, 1989) and to the TREK family. The TRAAK version of this family, found exclusively in nervous tissue, has been suggested to play a role in such processes as growth cone motility (Maingret et al., 1999); while this idea was not supported for the molluscan case (Morris and Horn, 1991), it will be very worthwhile to revisit it with the more powerful tools available once a channel's molecular identity is known. (5) Fish epithelial keratocytes: a role in modulating locomotion? During the locomotion of keratocytes, transient changes in intracellular calcium, evidently mediated by calcium-permeant SA channels, occur more frequently in cells temporarily stuck to the substratum or subjected to stretch; the increased calcium is implicated in detachment of the rear cell margin (Lee et al., 1999). (6) X e n o p u s oocytes MS channels: why the rapid adaptation? On gentle patches, a pressure step elicits a transient increase in MS channel activity--the response shows adaptation, comparable perhaps to the phasic responses of some mechanosensory cells. Is adaptation a type of slow inactivation (comparable to inactivation in voltagegated channels) or is it associated with a change in mechanosensitivity? Double-step protocols indicate that adapted channels will reactivate with stronger stimulation, suggesting that inactivation of the MS gates has not occurred. By elimination, the transient change seems to be associated with gating sensitivity. As described earlier, adaptation is a fragile phenomenon, critically dependent on the mechanical history of the patch. A surprising aspect of adaptation is that it is not seen at depolarizing potentials. This is also true of adaptation in yeast MS channels (see Martinac, 1993; Gustin, 1992). From the point of view of potential physiological significance, these are interesting but puzzling findings because there would seem to be a Catch-22 operating. Adaptation occurs only at hyperpolarized potentials. In vivo, therefore, activation of the MS cation channels would depolarize the membrane, thereby precluding adaptation. Hyperpolarizing influences working in parallel with the MS
44. Mechanosensitive Ion Channels in Eukaryotic Cells
channel to override their depolarizing effect (e.g., Ca2+-acti vated K + channels) might, however, undo the Catch. (7) Epithelial SA CI- channels: a role in volume regulation? In shark rectal gland cells and in renal cells, C1channel activity depends on the state of actin (Schwiebert et al., 1994; Mills et al., 1994). In renal glands, filamentous actin is associated with inactive channels and depolymerized actin with active channels. Membrane stretch also stimulates channel activity. It is suggested that during swelling-induced actin depolymerization, cell swelling could be limited through activation of MS chloride channels. (8) NMDA (glutamate) channels in mouse neurons: pathology? physiology? Membrane tension potentiates the activity of ligand-bound NMDA channels in a variety of recording configurations (Paoletti and Ascher, 1994). By contrast, activated kainate-type glutamate channels in the same patches are not mechanosensitive. The potentiation by tension is relatively small (at best, three-fold) and stretch alone will not activate the channels. Successive stimuli yield progressively larger effects, suggesting that mechanosensitivity is increased by disrupting the cytoskeleton. NMDA responses to stretch in the whole-cell mode support this view; osmosensitivity becomes evident after prolonged dialysis. The simplest explanation of this "run-up" of mechanosensitivity is that as mechanoprotection from the cytoskeleton is progressively lost, force transfer to the NMDA channels directly via the bilayer improves (consistent with newer findings that amphipaths mimic stretch in NMDA channels (Casado and Ascher, 1998)). Potentially, mechanosensitivity in NMDA-activated channels could have far-ranging implications for neuronal plasticity. If it reflects pathology, it is a phenomenon worth understanding in the context of the glutamate excitotoxicity and neuronal swelling associated with stroke.
X. Models for Gating of MS Channels Formal models for mechanosensitive gating provide a framework for relating experimentally applied force to kinetic data, and now that channels can be crystallized, to protein structure. If a channel has two states and is in some sense a spring, then the mechanical features that should matter are the spring's position and its elasticity. There are various candidates for MS channel springs. Is the entire channel or some part of it elastic? Is an extracellular or intracellular tether (tip links in hair cells, strands of membrane skeleton or extracellular matrix) mechanically in series with the channel? Is it the collection of noncovalent bonds between lipid and protein at the channel-bilayer interface (or "the bilayer")? Is it some combination? Whatever the identity of the spring, a two-state model assumes that force acting through the spring biases a channel's open probability; the channel would have a free energy difference of several k T between open and closed states and the force helps surmount this barrier. When the larger state has a pore in the conducting configuration, then the channel is stretch activated, whereas when the larger state has a nonconducting pore, the channel is stretch inactivated. A simple and satisfying two-state model is that developed by Hudspeth and colleagues (see Corey and Howard, 1994)
757
for hair cell channels. It posits, on excellent empirical grounds, that an extracellular tip link acts like an elastic spring in series with a stochastic gate. An external force changes the spring's tension and thereby its mean position. Two extremes of position are helpful to visualize: a position in which the spring is fully relaxed, with its attached gate occluding the channel (the small state), and another position in which the spring is under maximal tension with the gate away from the now-open pore (the large state). If external force alters the mean position of the gate and thereby the channel's open probability, then reciprocally, thermally induced opening and closing of the channel should randomly stress and relax the gating spring. Measurements of channel noise (which indicates the rates of gating transitions) and hair cell compliance (assumed to be an indicator of gate position) in stimulated hair cells support this prediction. Note that in this model, the spring's elasticity is critical to its action, but it is regarded as a fixed quantity. But for many eukaryotic MS channels, as for bacterial MS channels (see Batiza et al., 1999) the important positional effect may be even simpler. Consider stretch-activated K + channels like TREK-1. It is simplest to assume they lack specialized spring structures and to postulate that, over a limited range, the bilayer is the spring around the stochastic channel. Stochastic (thermally driven) conformational transitions of the channel occur among closed and open states. If, say, a particular open conformation occupies more membrane area than all other states, it will be favored by tension because, to open, the opening channel protein has less work to do against the bilayer spring when the bilayer is under tension. Again, it is helpful to visualize two extremes of position: a position in which the spring (bilayer) is fully relaxed and its "attached" channel is closed (this is the smaller state; note that the channel protein per se is the gate) and another position with the spring (bilayer) under maximal tension and the channel (gate) open. In the low-tension condition, if the channel is to open, thermal energy in the channel will need to be sufficiently large to compress the already quite compressed bilayer. When the bilayer is expanded (i.e., under tension), however, the opening event will happen more frequently because a more frequently experienced low dose of thermal energy will suffice. Again, note the reciprocal compression and relaxation of the spring (whose elasticity constant is fixed) as the channel opens and closes. This describes stretch activation. Inversely, a scenario in which one or more of the closed states occupied more membrane area than other states would yield stretch inactivation. The possibility that a MS channel's two major states have different elasticities has also been formalized. In this case (Sachs and Lecar, 1991; Lecar and Morris, 1993) the MS gate is depicted as a parabolic harmonic oscillator. At a given applied force (membrane tension) the free energies of the open and closed states depend on both mean gate position and the state-specific elasticities of the two states. Using reasonable quantities for channel transition rates and positional changes, Corey and Howard (1994) estimate, however, that differences in the second factor--elasticity--would be too small to have appreciable impact. Intriguingly, though, they point out that were the closed state significantly "softer" than the open state, the prediction would be that a channel's open
758
SECTION V Synaptic Transmission and Sensory Transduction
probability will go through a maximum as force increases, rather than generating a sigmoid activation curve. It may be worth beating this latter prediction in mind as further studies of SI channel mechanisms are undertaken. As discussed above, Oliet and Bourque (1993) had to assume in explaining their data that their experimental conditions generated a persistent tension offset in the osmosensory neuron membrane. Should it be found that there is no such offset and that the bell-shaped activation curves represent the true inputoutput relations of these MS channels, then the possibility of state-specific elasticities could be explored through detailed single-channel kinetics. Finally, some MS channels may undergo mechanosensitive open-closed transitions along a reaction coordinate whose units are not well described as a linear position (Lecar and Morris, 1993).
XI. Summary and Conclusions MS channels are ubiquitous, abundant, and diverse. They have few unifying features other than a propensity to change open probability when stretched under patch-clamp conditions. For single-channel recordings, the name "MS channel" carries an implicit assumption that some susceptible gating structure experiences tension in the plane of the membrane and that this biases the channel's open probability. The available information on possible mechanisms of ion channel mechanosensitivity suggest that two extremes are both important. At one extreme, the specialized receptors, cytoarchitecture is central to function. Specialized animal mechanoreceptors (e.g., hair cells in vertebrates, mechanosensory neurons in the nematode C. elegans.) require that mechanotransducer channels be precisely aligned with respect to special extracellular and intracellular structural proteins as well as more distant cellular structures and the whole array of hair cells. At the other extreme are bacterial MS channels, which can be reconstituted into artificial bilayers and retain their mechanosensitivity. As mechanistic details become available, definitions will need refining. A mechanically responsive channel that might not qualify as a MS channel would be one whose open probability changed (perhaps irreversibly) because a mechanical deformation of the membrane induced a critical lipid phase transition or altered the constituents of the bilayer or disrupted interactions among channel auxiliary subunits. It is easy to see that this could degenerate into semantics. Eventually, we may wish to reserve the term MS channel for physiological mechanotransducer channels and to speak of "mechanosusceptible" channels to describe those that show MS gating in patch recordings but not in situ. In the meantime, ingenious ideas about how MS channels might contribute to cellular control processes are not in short supply, and work is underway on many fronts to test these ideas. Although MS channels detectable at the single-channel level have been on the scene for approaching two decades, they should only be afforded a status coequal with voltagegated and ligand-gated channels once it is abundantly clear that their MS gating has a biological meaning. It should not be forgotten, however, that the inherent mechanosusceptibility of many other channels also acquires biological meaning (rather than solely biophysical meaning) if it is shown that
evolution has had to contrive ways to protect cells from unwanted mechanically driven membrane currents.
Bibliography Awayda, M. S., and Subramanyam, M. (1998). Regulation of the epithelial Na+ channel by membrane tension. J. Gen. Physiol. 112, 97-111. Bargmann, C. I. (1994). Molecular mechanisms of mechanosensation? Cell 78, 729-731. Batiza, A. E, Rayment, I., and Kung, C. (1999). Channel gate! Tension, leak and disclosure. Structure Fold Des. 7, R99-R103. Bear, C. E., and Li, C. (1991). Calcium-permeable channels in rat hepatoma cells are activated by extracellular nucleotides. Am. J. Physiol. 261, C1018-C1024. Bedard, E., Sigurdson, W. J., and Morris, C. E. (1992). Channels activated by stretch in neurons of a helix snail. Can. J. Physiol. 70, 207-213. Bourque, C. W., and Oliet, S. H. R. (1995). Mechanosensitive ion channels and osmoreception in magnocellular neurosecretory neurons. In Neurohypophysis: recent progress in vasopressin and oxytocin research. (T. Saito, K. Kurokawa, and S. Yoshida, Eds.) Elsevier Science, Amsterdam. pp. 205-213. Bowman, C. L., Ding, J. P., Sachs, E, and Sokabe, M. (1992). Mechanotransducing ion channels in astrocytes. Brain Res. 584, 272-286. Bregestovski, P., Medina, I., and Goyda, E. (1992). Regulation of potassium conductance in the cellular membrane at early embryogenesis. J. Physiol. Paris 86, 109-115. Casado, M., and Ascher, P. (1998). Opposite modulation of NMDA receptors by lysophospholipids and arachidonic acid: common features with mechanosensitivity. J. Physiol. (Lond.) 513, 317-330. Cemerikic, D., and Sackin, H. (1993). Substrate activation of mechanosensitive, whole-cell currents in renal proximal tubule. Am. J. Physiol. 264, F697-F714. Chang, G., Spencer, R. H., Lee, A. T., Barclay, M. T., and Rees, D. C. (1998). Structure of the MscL homolog from Mycobacterium tuberculosis: a gated mechanosensitive ion channel. Science 282, 2220--2226. Corey, D. P, and Garcia-Anoveros, J. (1996). Mechanosensation and the DEG/ENac ion channels. Science 273, 323-324. Corey, D. P., and Howard, J. (1994). Models for ion channel gating with compliant states. Biophys. J. 66, 1254-1257. Dai, J., and Sheetz, M. P. (1995). Mechanical properties of neuronal growth cone membranes studied by tether formation with laser optical tweezers. Biophys. J. 68, 988-996. Dai, J., Sheetz, M. P., Wan, X., and Morris, C. E. (1998). Membrane tension in swelling and shrinking molluscan neurons. J. Neurosci. 18~ 6681-6692. Davidson, R. M., Tatakis, D. W., and Auerbach, A. L. (1990). Multiple forms of mechanosensitive ion channels in osteoblast-like cells. Pflugers Arch. 416, 646-651. Ding, J. P., Salvi, R. J., and Sachs, F. (1992). Stretch-activated ion channels in guinea pig outer hair cells. Hearing Res. 56, 19-28. Dopico, A. M., Kirber, M. T., Singer, J. J., and Walsh, J. V., Jr. (1994). Membrane stretch directly activates large conductance Ca2§ vated K§ channels in mesenteric artery smooth muscle cells. Am. J. Hyperten. 7, 82-89. Duncan, R. L. and Hruska, K. A. (1994). Chronic, intermittent loading alters mechanosensitive channel characteristics in osteoblast-like cells. Am. J. Physiol. 267, F909-F916. Duncan, R. L., Hruska, K. A., and Misler, S. (1992). Parathyroid hormone activation of stretch-activated cation channels in osteosarcoma cells (UMR- 106.1). FEBS Lett. 307, 219-223.
44. Mechanosensitive Ion Channels in Eukaryotic Cells Erxleben, C. (1989). Stretch-activated current through single ion channels in the abdominal stretch receptor organ of the crayfish. J. Gen. Physiol. 94, 1071-1083. Franco, A., Jr., and Lansman, J. B. (1990). Calcium entry through stretchinactivated ion channels in mdx myotubes. Nature 344, 670-673. French, A. S. (1992). Mechanotransduction. Annu. Rev. Physiol. 54, 135-152. Glogauer, M., Arora, P., Chou, D., Janmey, P. A., Downey, G. E, and McCulloch, C. A. (1998). The role of actin-binding protein 280 in integrin-dependent mechanoprotection. J. Biol. Chem. 273, 1689-1698. Guharay, E, and Sachs, E (1984). Stretch-activated single ion channel currents in tissue-cultured embryonic chick skeletal muscle. J. Physiol. 352, 685-701. Gustin, M. C. (1992). Mechanosensitive ion channels in yeast. Mechanisms of activation and adaptation. Adv. Compar. Environ. Physiol. 10, 19-38. Gustin, M. C., Sachs, E, Sigurdson, W., Ruknudin, A., Bowman, C., Morris, C. E., and Horn, R. (1991). Single-channel mechanosensitive currents (Refereed Commentary and Reply). Science 253, 801-802. Hamill, O. E, and McBride, D. W., Jr. (1992). Rapid adaptation of single mechanosensitive channels in Xenopus oocytes. Proc. Natl. Acad. Sci. USA 89, 7462-7466. Hamill, O. E, and McBride, D. W., Jr. (1995). Mechanoreceptive membrane channels. Am. Scientist 83, 30-37. Hamill, O. E, and McBride, D. W., Jr. (1996). The pharmacology of mechanogated membrane ion channels. Pharmacol. Rev. 48, 231-252. Hamill, O. E, and McBride, D. W., Jr. (1997). Induced membrane hypo/hyper-mechanosensitivity: a limitation of patch-clamp recording. Annu. Rev. Physiol. 59, 621-631. Hase, C. C., Le Dain, A. C., and Martinac, B. (1995). Purification and functional reconstitution of the recombinant large mechanosensitive ion channel (MscL) of Escherichia coli. J. Bio. Chem. 270, 18 329-18 334. Heidemann, S. R., and Buxbaum, R. E. (1994). Mechanical tension as a regulator of axonal development. Neurotoxicol. 15, 65-108. Hisada, T. R. W., Ordway, R. W., Kirber, M. T., Singer, J. J., and Walsh, J. V., Jr. (1991). Hyperpolarization-activated cationic channels in smooth muscle cells are stretch-sensitive. Pfliigers Arch. 417, 493-499. Hong, K., and Driscoll, M. (1994). A transmembrane domain of the putative channel subunit MEC-4 influences mechanotransduetion and neurodegeneration in C. elegans. Nature 367, 470-473. Huang, M., Gu, G., Ferguson, E. L., and Chalfie, M. (1995). A stomatin-like protein necessary for mechanosensation in C. elegans. Nature 378, 292-295. Hudspeth, A. J. (1997). How hearing happens. Neuron 19, 947-950. Ji, S., John, S. A., Lu, Y., and Weiss, J. N. (1998). Mechanosensitivity of the cardiac muscarinic potassium channel. A novel property conferred by Kir3.4 subunit. J. Biol. Chem. 273, 1324--1328. Kell, A., and Glaser, R. W. (1993). On the mechanical and dynamic properties of plant cell membranes: their role in growth, direct gene transfer and protoplast fusion. J. Theor. Biol. 160, 41-62. Kim, D. (1992) A mechanosensitive K + channel in heart cells. Activation by arachidonic acid. J. Gen. Physiol. 100, 1021-1040. Kim, E., Niethammer, M., Rothschild, A., Jan, Y. N., and Sheng, M. (1995). Clustering of Shaker-type K + channels by interaction with a family of membrane-associated guanylate kinases. Nature 378, 85-88. Kirber, M. T., Ordway, R. W., Clapp, L. H., Walsh, J. V., Jr., and Singer, J. J. (1992). Both membrane stretch and fatty acids directly activate large conductance Ca2+-activated K + channels in vascular smooth muscle cells. FEBS Letts. 297, 24-28. Lambert, S., and Bennett, V. (1993). From anemia to cerebellar dysfunction. A review of the ankyrin gene family. Eur. J. Biochem. 211, 1-6.
759 Lecar, H., and Morris, C. E. (1993). Biophysics of mechanotransduction. In "Mechanoreception by the Vascular Wall" (G. Rubanyi, Ed.), Futura Pub. Co. Inc., Mount Kisco, NY. pp. 1-11. Lee, J., Ishihara, A., Oxford, G., Johnson, B., and Jacobson, K. (1999). Regulation of cell movement is mediated by stretch-activated calcium channels. Nature 400, 382-386. Levina, N. N., Lew, R. R., and Heath, I. B. (1994). Cytoskeletal regulation of ion channel distribution in the tip-growing organism of Saprolegnia ferax. J. Cell Sci. 107, 127-134. Lewis, R. S., Ross, E E., and Cahalan, M. D. (1993). Chloride channels activated by osmotic stress in T lymphocytes. J. Gen. Physiol. 101, 801-826. Liu, J., Schrank, B., and Waterston, R. H. (1996). Interaction between a putative mechanosensory membrane channel and a collagen. Science 273, 361-364. Maingret, E, Fosset, M., Lesage, E, Lazdunski, M., and Honore, E. (1999). TRAAK is a mammalian neuronal mechano-gated K + channel. J. Biol. Chem. 274, 1381-1387. Marchenko, S. M., and Sage, S. O. (1997). A novel mechanosensitive cationic channel from the endothelium of rat aorta. J. Physiol. (Lond.) 498, 419425. Martinac, B. (1993). Mechanosensitive ion channels: biophysics and physiology. In "Thermodynamics of Membrane Receptors and Channels" (M. B. Jackson, Ed.), pp. 327-352. CRC Press. McBride, D. W., Jr., and Hamill, O. E (1992). Pressure clamp: a method for rapid step perturbation of mechanosensitive channels. Pfliigers Arch. 421, 606-612. Medina, I. R., and Bregestovski, E D. (1991). Sensitivity of stretch-activated K + channels changes during cell-cleavage cycle and may be regulated by cAMP-dependent protein kinase. Proc. Roy. Soc. Lond. B 245, 159-164. Mills, J. W., Schwiebert, E. M., and Stanton, B. A. (1994). Evidence for the role of actin filaments in regulating cell swelling. J. Exp. Zool. 268, 111-120. Morris, C. E. (1990) Mechanosensitive ion channels. J. Memb. Biol. 113, 93-107. Morris, C. E., (1992). Are stretch-sensitive channels in molluscan cells and elsewhere physiological mechanotransducers? Experimenta 48, 852-858. Morris, C. E., and Horn, R. (1991). Failure to elicit neuronal macroscopic mechanosensitive currents anticipated by single-channel studies. Science 251, 1246-1249. Morris, C. E., and Sigurdson, W. J. (1989). Stretch-inactivated ion channels coexist with stretch-activated ion channels. Science 243, 807-809. Nakamura, E, and Strittmatter, S. M. (1996). P2Y1 purinergic receptors in sensory neurons: contribution to touch-induced impulse generation. Proc. Natl. Acad. Sci. USA 93, 10465-10470. Oliet, S. H. R., and Bourque, C. W. (1993). Mechanosensitive channels transduce osmosensitivity in supraoptic neurons. Nature 364, 341-343. Oliet, S. H. R., and Bourque, C. W. (1996). Gadolinium uncouples mechanical detection and osmoreceptor potential in supraoptic neurons. Neuron 16, 175-181. Ordway, R. W., Petrou, S., Kirber, M. T., Walsh, J. V. Jr., and Singer, J. J. (1995). Stretch activation of a toad smooth muscle K + channel may be mediated be fatty acids. J. Physiol. 484, 331-337. Paoletti, E, and Ascher, E (1994). Mechanosensitivity of NMDA receptors in cultured mouse central neurons. Neuron 13, 645-655. Patel, A. J., Honore, E., Lesage, E, Fink, M., Romey, G., and Lazdunski, M. (1999). Inhalational anesthetics activate two-poredomain background K + channels. Nat. Neurosci. 2, 422-2426. Patel, A. J., Honore, E., Maingret, E, Lesage, E, Fink, M., Duprat, E, and Lazdunski, M. (1998). A mammalian two pore domain mechano-gated S-like K + channel. EMBO. J. 17, 4283-4290. Pleusamran, A., and Kim, D. (1995). Membrane stretch augments the cardiac muscarinic K + channel activity. J. Membr. Biol. 148, 287-297.
760
Rotin, D., Bar-Sagi, D., O'Brodovich, H., Merilainen, J., Lehto, V. E, Canessa, C. M., Rossier, B. C., and Downey, G. E (1994). An SH3 binding region in the epithelial Na + channel 0irENaC) mediates its localization at the apical membrane. EMBO. J. 13, 4440--4450. Sachs, E (1992). Stretch-sensitive ion channels: an update. In "Sensory Transduction" (D. Corey, Ed.), pp. 242-260, Rockefeller University Press. Sachs, E, and Lecar, H. (1991). Stochastic models for mechanical transduction (letter). Biophys. J. 59, 1142-1145. Sachs, E, and Morris, C. E. (1998). Mechanosensitive ion channels in non-specialized cells. Rev. Physiol. Biochem. Pharmacol. 132, 1-78. Sachs, E, Morris, C. E., Hamill, O., Chakfe, Y., and Bouraque, C. W. (2000). Does a stretch-inactivated cation channel integrate asmotic and peptidergic signals? (Refereed Commentary and Reply) Nat. Neurosci. 3, 847-848. Sackin, H. (1995). Mechanosensitive channels. Annu. Rev. Physiol. 57, 333-353 Schwiebert, E. M., Mills, J. W., and Stanton, B. A. (1994). Actin-based cytoskeleton regulates a chloride channel and cell volume in a renal cortical collecting duct cell line. J. Biol. Chem. 269, 7081-7089. Sheetz, M. E, and Dai, J. (1996). Modulation of membrane dynamics and cell motility by membrane tension. Trends Cell Biol. 6, 85-89. Small, D. L., and Morris, C. E. (1994). Delayed activation of single mechanosensitive channels in Lymnaea neurons. Am. J. Physiol. 267, C598-C606. Sokabe, M., Sachs, E, and Jing, Z. (1991). Quantitative video microscopy of patch clamped membranes stress, strain, capacitance, and stretch channel activation. Biophys. J. 59, 722-728. Steffensen, I., Bates, W. R., and Morris, C. E. (1991). Embryogenesis in the presence of blockers of mechanosensitive ion channels. Dev. Growth Differ. 5, 437--442. Sukharev, S. I., Blount, E, Schroeder, M., and Kung, C. (1996). Multimeric structure of bacterial mechanosensitive channel MscL. Biophys. J. 70, A366.
SECTION V Synaptic Transmission and Sensory Transduction
Sukharev, S. I., Blount, E, Martinac, B., Blattner, E R., and Kung, C. (1994). A large conductance mechanosensitive channel in E. coli encoded by MscL alone. Nature 368, 265-268. Tabarean, I. V., Juranka, E, and Morris, C. E. (1999). Membrane stretch affects gating modes of a skeletal muscle sodium channel. Biophys. J. 77, 758-774. Tavernarakis, N., and Driscoll, M. (1997). Molecular modeling of mechanotransduction in the nematode Caenorhabditis elegans. Annu. Rev. Physiol. 59, 659-689. Vandorpe, D. H., and Morris, C. E. (1992). Stretch activation of the Aplysia S channel. J. Memb. Biol. 127, 205-214. Vandorpe, D. H., Small, D. L., Dabrowski, A. R., and Morris, C. E. (1994). FMRFamide and membrane stretch as activators of the Aplysia S-channel. Biophys. J. 66, 46-58. Vanoye, C. G., and Reuss, L. (1999). Stretch-activated single K + channels account for whole-cell currents elicited by swelling. Proc. Natl. Acad. Sci. USA 96, 6511-6516. Van Wagoner, D. R., and Russo, M. (1992). Whole-cell mechanosensitive K + currents in rat atrial myocytes. Biophys. J. 61, A251. Verrey, E, Groscurth, E, and Bolliger, U. (1995). Cytoskeletal disruption in A6 kidney cells: impact on endo/exocytosis and NaC1 transport regulation by antidiuretic hormone. J. Memb. Biol. 145, 193-204. Wan, X., Harris, J. A., and Morris, C. E. (1995). Response of neurons to extreme osmomechanical stress. J. Membr. Biol. 145, 21-31. Wan, X., Juranka, E, and Morris, C. E. (1999). Activation of mechanosensitive currents in traumatized membrane. Am. J. Physiol. 276, C318-C327. Waxman, S. G., Kocsis, J. D., and Black, J. A. (1994). Pathophysiology of demyelinated axons. In "The Axon: Structure, Function and Pathophysiology" (S. G. Waxman, J. D. Kocsis, and E K. Stys, Eds.) Oxford University Press. Zhou, X. L., Stumpf, M. A., Hoch, H. C., and Kung, C. (1991). A mechanosensitive channel in whole cells and in membrane patches of the fungus Uromyces. Science 253, 1415-1417.
Andrew S. French and Piiivi H. Torkkeli
Sensory Receptors and Mechanotransduction
!. Introduction While most living cells can detect a variety of changes in their external physical and chemical environments and respond appropriately, we usually reserve the term sensory receptor for those cells that transmit information about such changes to the animal's nervous system. Even this classification must be qualified. For example, the heart and other hollow organs have localized, mainly autonomous nervous systems that allow sensory receptors to produce very local functional changes. Moreover, some of the mechanisms that cells use to detect external events seem to be very similar as we move from unicellular animals to the complex sensory organs of humans. With these caveats in mind, this chapter is restricted to those cells (usually neurons) that transmit sensory information into one or more divisions of the central nervous system. Several general principles can be applied to all sensory receptors. The concept of modality means that each sensory cell transmits information about only one type of environmental stimulus. Photoreceptors detect light but are insensitive to touch, while stretch receptors do not respond to odorant molecules. Within the modalities there are further specializations, so that cold temperature receptors do not respond to hot stimuli, and sweet taste receptors do not respond to bitter substances. Each type of receptor seems to produce a single specific molecular machinery for detecting just one type of stimulus. An exception to this rule is provided by the cutaneous polymodal pain receptors (Perl, 1996) which can be stimulated by mechanical, thermal, or chemical stimuli. However, it is not yet known how this multimodal behavior is produced, or if each sensory ending can detect all the different modalities. Neurons carrying information into the nervous system form labeled lines that signal specific modalities and specific locations. Artificially stimulating a sensory nerve by electrical or mechanical stimulation produces a sensation that reflects the modality and location of the sensory ending, not the actual stimulus. Another general principle is that the intensity of the stimulus is encoded as the amplitude of the Cell Physiology Sourcebook: A Molecular Approach, Third Edition
761
electrical signal passing along the sensory neuron. In most cases, this means the action potentials propagating along the sensory axon, with more action potentials per second representing a stronger stimulus. However, there are several situations where the primary sensory cell does not produce action potentials at all, but a graded change in membrane potential that is conducted decrementally along the cell. Graded potentials can usually be propagated for only very short distances (a few millimeters at most), so such cells are generally small. Well-known examples are the rods and cones of the retina and the hair cells of the inner ear. Some invertebrate sensory axons are sufficiently short that both graded and action potential sensory signals can reach the central nervous system (Pasztor and Bush, 1982).
11. Sensory Transduction Sensory cells respond to an external stimulus by changing their membrane potential, although several processes may occur before and after this transduction step. This change in membrane potential is called a receptor potential, and it is a graded potential, increasing and decreasing with the intensity of the stimulus. Nonsensory neurons are insensitive to external stimuli such as mechanical stress, temperature, light, and chemicals (other than synaptic transmitters), unless these reach such intensity as to threaten the cell's normal function or integrity. Therefore, sensory receptors must have specializations to detect external stimuli (Fig. 1). Conceptually, the simplest receptors are the electroreceptors, which use an external electrical current to change their own membrane potential by direct current flow. Although the concept is simple, these cells are highly specialized to maximize their electrical sensitivity. Current flows through resting membranes by the passage of ions through ion channels that are open at rest. The salt taste receptors of the tongue and mouth also rely on direct ion flow through ion channels. In this case the channels are selective for sodium and can be blocked by amiloride, so that placing a small amount of amiloride in
Copyright 9 2001 by Academic Press. All rights of reproduction in any form reserved.
762
SECTION V Synaptic Transmission and Sensory Transduction
many chemoreceptors, specialized receptor molecules in the outer cell membrane activate membrane ion channels through a second messenger cascade when they receive an appropriate stimulus molecule. In the case of rod photoreceptors, light activates specialized receptor molecules in intemal cell membranes, again producing a second messenger cascade that eventually closes ion channels in the extemal membrane. It is not known which types of sensory receptor mechanisms evolved first, but it is clear that several, if not all, of these processes have close parallels in other nonsensory tissues, such as the sodium channels of epithelia, the mechanically activated channels of muscle and many other tissues, and the second messenger cascades of synapses and secretory cells. It seems likely that the evolution of different sense modalities has taken advantage of preexisting cellular mechanisms, and vice-versa.
I!!. Sensory Adaptation
F I G U R E 1. Fundamental mechanisms of sensory transduction organized by the level of complexity. In each case the cell has specialized ion channels that produce a receptor current. Modulation of the current is achieved by changing the electrochemical force for the permeant ion in the cases of electroreceptors and salt taste receptors. In other receptors the external stimulus changes the channel open probability, either by acting directly on the channel protein or via a second messenger
cascade. Chemoreceptors have receptor proteins in the cell membrane. In vertebrate rods the rhodopsin receptors are in the internal disk membranes, but cones and invertebrate photoreceptors have rhodopsins in the cell membrane.
the mouth removes the ability to taste salt. Sodium diffuses through these channels to depolarize the taste cell membrane. An additional level of complexity is illustrated by sour taste receptors, where potassium-selective ion channels are closed by external hydrogen ions, reducing an outward current and depolarizing the cell. The fundamental transduction step in mechanoreceptors is not well understood, but is assumed to involve mechanosensitive ion channels. It seems likely that these channels are opened by links to intracellular cytoskeletal components (Sachs, 1997), although this is not yet proven, ff true, this is another case where the external stimulus acts directly to open or close an ion channel. Finally, there are cases where the external stimulus acts indirectly through intermediate membrane receptors. In
Sensory receptors often have to deal with very wide ranges of stimulus amplitudes. Therefore, they need high sensitivity to detect weak stimuli, plus the ability to reduce their sensitivity if the stimulus is strong. In some cases the maximum sensitivity approaches the limits imposed by physics or chemistry. Human rod photoreceptors and some arthropod photoreceptors can detect single photons arriving in the eye (Fein and Szuts, 1982), which is clearly the physical limit. The human ear can detect air movements of about 0.01 nm, close to the diameter of a hydrogen atom (Hudspeth, 1989). In such cases the sensory cells must amplify the initial signal considerably, which at least partly explains the complex morphology of rod photoreceptors and the human cochlea. The reduction in sensitivity following an increase in stimulus is called adaptation. It may be seen as a decrease in receptor potential with time during a constant stimulus, or as an increase in the strength of stimulus required to produce a constant response. All sensory receptors adapt to some extent but there is a very wide range of adaptation speeds and amounts. At one extreme are receptors such as Pacinian corpuscles (Loewenstein and Mendelson, 1965) and spider slit sensilla (Seyfarth and French, 1994), which fire only one or two action potentials with even a strong, continuous stimulus. At the other extreme are Ruffini endings (Malinovsky, 1996), which continue to fire steadily for long periods. The type of adaptation is always appropriate to the function of the receptor. Muscle and joint receptors that signal limb position would not be useful if they adapted to silence in a few seconds, because the sense of limb position, or kinesthesia, would vanish if one did not keep moving. On the other hand, photoreceptors must function under a wide range of light intensities and it is essential that they can adjust their sensitivity to the ambient light level. The rapid adaptation of Pacinian corpuscles makes them ideally suited for detecting vibration, since a rapidly changing stimulus will repeatedly stimulate the receptor, while a steady stimulus will produce little response. The mechanisms of adaptation vary widely and may occur at different stages of the process. Some involve components outside the sensory cell, such as the mechanical creep of the muscle in muscle spindles or the movements of screening pigments in insect eyes. Others may involve chemical signals
45. Sensory Receptors and Mechanotransduction
within and between cells, as found in photoreceptors and olfactory receptors. Mechanical adaptation by the capsule of the Pacinian corpuscle is very well known because of the pioneering work of Loewenstein and Mendelson (1965), who described the dramatic reduction in receptor potential adaptation that can be achieved by removing most of the capsule (see Fig. 2). Unfortunately, many descriptions stop at this point, leaving the impression that mechanical adaptation dominates Pacinian corpuscle behavior, but Loewenstein and Mendelson showed that even after decapsulation the receptor will only fire one or two action potentials in response to a prolonged step stimulus (Fig. 2). They described this as electrical adaptation, and it now seems probable that many receptors use voltage- or calciumactivated ion channels in their membranes to produce electrical adaptation by raising the threshold for action potential production. This has been clearly established in several arthropod and vertebrate mechanoreceptors and is likely to occur in many other types of receptors that use action potentials to encode the sensory signal (French and Torkkeli, 1994).
IV. Information Transmission by Sensory Receptors Nervous systems and individual nerve cells deal with information. Cells with long axons transmit information over considerable distances, and we assume that all nerve cells process information to some extent. Unfortunately, it is usually difficult to decide what the nature of the information is, and how it is being processed, except in vague and general terms. Sensory cells have been particularly important in trying to understand quantitatively how information is encoded and transmitted by nerve cells because it is usually much easier to define the input signal of a receptor neuron than other nerve cells. For example, the intensity and wavelength of light entering a photoreceptor or the length of a muscle spindle receptor can be accurately controlled while the out-
c
FIGURE 2. Adaptationin the mammalian Pacinian corpuscle occurs in two stages. (A) Stimulating with a mechanical step (bottom) causes rapidly adapting receptor potential responses at the start and end of the step in a normal receptor. (B) Removing most of the lamellae surrounding the sensory ending (decapsulation) eliminates most of the adaptation to reveal a slowly adapting receptor potential. (C) Suprathreshold stimulation of a decapsulated receptor still produces only two action potentials, showing that there is rapid adaptation in the conversion of receptor potential to action potentials. (Redrawn from Loewenstein and Mendelson, 1965.)
763 put of the receptor in terms of receptor potential or action potentials is recorded. Quantitative investigations of information flowing through sensory neurons have been based on ideas developed by engineers for dealing with artificial communication systems, such as telephone or television signals. The recent widespread use of " digital communications devices, like telephone modems, has made the concept of an information transmission rate (usually in bits per second, or bps) familiar to most of us. If we know the nature of the information being transmitted we can measure the actual rate of transmission, but even if nothing is known about the kind of information being transmitted, it is possible to calculate the theoretical maximum rate that could be achieved by an optimum encoding scheme. This is called the information capacity of system, and it is closely related to the inherent noise in the system. The basic idea is very simple. Suppose that one person (the sender) wants to send a message to another (the receiver) by changing the voltage at one end of a piece of copper wire, so that the receiver can measure the voltage change to read the message. If the copper wire were completely free of noise, the sender and receiver could agree to use voltages of any value. For example, the voltages 1 V, 2 V, and 3 V could represent the letters A, B, and C. Since any size of voltage can be used, there is no limit to the number of different signals (voltages) sent and received in any given time, and the information capacity is infinite. However, if the copper wire actually has random voltage noise of about 1 V amplitude, a signal of 1 V is difficult or impossible to read, while one of 10 V may still be detected. The higher the noise level, the more difficult it is to send and receive information. If the receiver can take time to average the output from the wire, it may still be possible to observe the 1 V signal, but this reduces how much information can be received per unit time. The temporal properties of the noise are also important, because slowly changing noise may not interfere with a quickly changing signal, and vice-versa. All these arguments can be applied, with suitable dimension changes, to other information transmission systems, such as optical fibers, satellite broadcasts, or sensory neurons. Note that when a noisy neuron fires action potentials the noise appears primarily as variability or randomness in the timing of the action potentials. These ideas were formulated into a standard expression for calculating the information capacity from the signal-to-noise ratio by Shannon and Weaver (1949) and this approach has been used in most estimates of receptor cell information capacity. Although relatively few measurements have been made, it is clear that cells using action potentials to transmit information have lower information capacities than those that do not. Spiking mechanoreceptors in cricket cercal mechanoreceptors and spider slit sensilla had capacities of 300 bps and 200 bps respectively, compared to values of 1650 bps for fly photoreceptors and 2240 bps for the receptor potential in spider slit sensilla before encoding (Juusola and French, 1997). However, it was possible to transmit information at rates up to 500 bps for short periods using a defined encoding scheme in a cockroach mechanoreceptor (French and Torkkeli, 1998), so we must be cautious about these estimates until we understand the actual encoding schemes that neurons use. These findings support the concept that nerve cells use action potentials to transmit information faithfully over long
SECTIONV SynapticTransmissionand SensoryTransduction
764 distances, but there is a significant cost in information capacity. This probably explains why some sensory systems have elaborate neural structures located peripherally, close to the primary sensory neurons. A familiar example is the vertebrate retina, with two layers of nonspiking cells and complex synaptic interactions before the spiking ganglion cells. This arrangement presumably allows important information processing to take place on the high-capacity input from the photoreceptors before it is encoded into the lower capacity ganglion cell axons.
V. Mechanoreceptors A. Vertebrate Receptors Vertebrate mechanoreceptors have such a wide variety of shapes and functions that classification is difficult. Many receptors bear the names of their original discoverers or rediscoverers, such as Meissner corpuscles and Ruffini endings. A variety of classification schemes have been used, based on features such as the morphology of the sensory ending and its associated tissues, rates of adaptation, and location in the body. The hair cells of the vertebrate auditory and vestibular systems have been studied extensively, and are described in Chapter (46). Vertebrate somatic mechanoreceptors have recently been classified according to the embryonic origin of the tissues associated with the sensory ending, and the complexity of the sensory ending itself (Malinovsky, 1996), as shown in Fig. 3. Type I receptors are associated with tissues of mesodermal origin (connective tissues and muscle) and include the Ruffini endings of the joints and skin, the Golgi tendon organs, and the muscle spindles. Type II receptors are transitional between Type I and Type III, and are associated with tissues of endodermal
Ib
Ruffini, Golgi
~ Ilia
IIIb
~~)Hair
~
Merkel
~Meissner Pacinian
FIGURE 3. Classificationof vertebrate mechanoreceptors by embryonic origin of associated tissues and complexity of ending. Type I is associated with tissues of mesodermal origin, Type II with endodermal or epidermal, and Type III with epidermal. Subtypes (Ia, Ib, etc.) indicate morphological complexity. (Redrawn from Malinovsky, 1996.)
or ectodermal origin. They include the Merkel endings, where each sensory ending is closely apposed to a specialized Merkel cell. Type III receptors are clearly associated with tissues of ectodermal origin, and comprise a wide range of receptors including the bulbous endings found in hair follicles, the Meissner corpuscles, and the Pacinian corpuscles. This classification broadly accompanies a functional shift from slowly adapting receptors (such as Ruffmi endings) in the Type I group to very rapidly adapting receptors in the Type III groups (such as Pacinian corpuscles). Phylogenetically, Type I and Type II receptors are found in all classes of vertebrates from fish to primates, while Type HI receptors are unknown in fish, start to occur in amphibia, and become increasingly common in higher vertebrates. A more detailed description of this classification and a history of vertebrate mechanoreceptor classification is provided by Malinovsky (1996).
B. Arthropod Receptors Arthropods (insects, spiders, and crustaceans) have provided some important preparations for the investigation of mechanoreceptor function. Their receptors can be divided into two major groups, cuticular and multipolar, sometimes called Type I and Type II mechanoreceptors (McIver, 1985). Cutieular receptors, as the name implies, are generally associated with the arthropod cuticle and have their cell bodies in the periphery, close to the sensory ending. This arrangement is very different than most vertebrate and other invertebrate mechanoreceptors, which have their cell bodies in the central nervous system, distant from the sensory ending, and it allows recordings to be made close to the site of mechanotransduction. In addition, many arthropod receptor cells are relatively large, allowing penetration by microelectrodes, which is impossible in most vertebrate receptors. There are three major groups of cuticular receptors (Fig. 4). Hairlike receptors are found all over the outer surfaces of most arthropods in a variety of shapes and sizes from long, thin hairs to short pegs. The hair is supported by flexible cuticle within a socket and moves relative to the skin. In insects, a single mechanoreceptor neuron is closely apposed to the base and its sensory ending contains microtubules that end in a dense tubular body. It is assumed that movement of the hair compresses the ending, with the tubular body perhaps ~/dding a rigid structure that the compression can work against. Crustacean and spider hairs are similar, but with two or three mechanosensory neurons in each hair. All arthropod hair types can contain other sensory neurons in addition to the mechanoreceptors, such as chemoreceptors in taste hairs. Campaniform (bell-shaped) sensilla are found in insects, where they detect stress in the cuticle. The stress again leads to compression of a dendritic tip containing a tubular body by squeezing of the bell as it is pushed downwards. Spiders have analogous stress-detecting receptors called slit sensilla, where the neurons are located in cuticular slits. Chordotonal receptors are generally found further beneath the integument, although they can be connected to the integument by attachment structures. They serve a variety of functions, including hearing and joint movement detection. They generally lack tubular bodies but have dense scolopale structures surrounding the dendrite and can have multiple mechanosensory neurons.
45. Sensory Receptors and Mechanotransduction
765
FIGURE 4. Threecommon types of arthropod cuticular mechanoreceptors.TB, tubular body in dendrite, formed from microtubules embedded in electron-dense material. Ep, epithelial layer, connected by tight junctions to create a lymph space (L). At, attachment from cap of chordotonal sensillum to another structure, such as a joint. Sc, scolopale surrounding a bundle of sensory dendrites.
C. Muscle Mechanoreceptors One of the best known arthropod mechanoreceptors is the crayfish stretch receptor, which has some interesting similarities to the vertebrate muscle spindle (Fig. 5). These receptors are found on many muscles in crustacea and are multipolar or Type II arthropod mechanoreceptors. The large cell bodies are again located in the periphery, but the sensory endings consist of finely divided dendrites that cover part of the muscle surface. Unlike the muscle spindle, crayfish stretch receptors are located on the main, force-producing muscle, and so are directly excited by passive stretch of the muscle or activation of motoneurons. However, crayfish stretch receptors also have inhibitory innervation that can inhibit the sensory neuron directly, as well as the muscle. Vertebrate muscle spindles are prominent and numerous mechanoreceptors, believed to be crucially involved in the control of muscle activation. However, they also make a major contribution to the sense of body position and movement, or kinesthesia, because appropriate artificial stimulation of muscle spindles gives sensations that the limbs are moving or in incorrect positions (Matthews, 1981). Mammalian muscle spindles are highly specialized receptors, located on their own exclusively activated intrafusal muscle fibers within the muscle, and are innervated by two types of sensory endings with different adaptation properties.
VI. Experimental Mechanoreceptor Preparations Although many different preparations have been developed to investigate mechanoreception, only three basic methods have been used to measure the receptor current or receptor potential (Fig. 6). The decremental conduction method relies on insulating the sensory axon from the extracellular fluid as close as possible to the sensory ending, usually with a nonconducting substance such as paraffin wax or petroleum jelly,
FIGURE 5. Vertebrateand invertebrate muscle mechanoreceptors. Mammalian muscle spindles contain two types of specialized muscle fibers with their own motor innervation via two types of y-motoneurons that are activated separately from the main muscle fibers. There are also two types of sensory endings with different adaptation characteristics. Crustacean stretch receptors are multipolar mechanoreceptors with many fine dendrites embedded in muscle fibers. There are two types of sensory neurons with different adaptation properties and inhibitory neurons that synapse on the sensory neurons as well as the muscle fibers. (Redrawn from Matthews, 1981; Swerup and Rydqvist, 1992).
766
FIGURE 6. Three methods that have been used widely to observe the receptor current and potential in mechanoreceptors. (A) Decremental conduction measures the current flowing along the sensory axon from a Pacinian corpuscle by insulating the axon as close as possible to the receptor to prevent the current from reaching ground, except via the measuring bath. (B) Transepithelial experiments in an arthropod hair measure the voltage as close as possible to the lymph space surrounding the sensory ending, and rely on the high resistance of the surrounding epithelium. (C) lntracellular recording from a spider slit sensillum neuron measures the actual receptor potential during stimulation. Current measurement requires reliable voltage-clamp of the sensory ending.
before leading it to a separate conducting solution. As the receptor is stimulated, some of the receptor current flows decrementally along the axon but cannot follow its normal path to the extracellular solution because of the insulation. It can only return through the second bath, where it can easily be measured. This method has been used to observe the receptor potential in Pacinian corpuscles, muscle spindles, and insect cuticular receptors. It allows a relatively stable preparation, because the receptor is not very disturbed and fine positioning is not needed after the initial setup. However, only a relatively small fraction of the total receptor current can usually be observed, and the decremental conduction includes a substantial selective loss of high-frequency components through the membrane capacitance to ground, so that the observed receptor potential is strongly attenuated and filtered. It is generally difficult to estimate the amplitude and waveform of the original receptor potential with any accuracy. Arthropod cuticular receptors have the sensory neurons so close to the surface that it is possible to observe part of the current flowing through the neuron from the exterior. For a hair
SECTION V Synaptic Transmission and Sensory Transduction
receptor, a small pool of conducting solution is placed in the hair socket and an electrode of some sort (glass, wire, wick) measures the potential in the pool. These transepithelial measurements rely on the high-resistance epithelium that surrounds cuticular receptors (see below) and the relatively low resistance of the thin, flexible socket. A similar technique is used for cuticular chemoreceptors, where it is common to cut the hair and place the tip of an electrode over the cut end to obtain even better electrical contact. However, this is more difficult for mechanoreceptors, where the hair must be moved to stimulate transduction. Transepithelial measurements allow a relatively stable recording situation, but the high-resistance access pathway attenuates the signal and emphasizes high frequencies, which can pass more easily through the cuticle. Here again, it is difficult to estimate the original receptor potential amplitude and time course with accuracy. IntraceUular recording close to the site of sensory transduction is clearly the technique of choice if it can be accomplished. However, it requires stable penetration of a small neuron while the sensory ending is moved nearby, and has only been successfully achieved in a few preparations. The real goal is voltageclamp of the sensory receptor membrane, since this allows the voltage across the ion channels to be controlled while the current flowing through them is recorded. Crustacean stretch receptors are relatively easy to penetrate, even with two microelectrodes, and have been used for many pioneering studies of the receptor currents and other currents that contribute to the electrical properties of the neurons during sensory transduction (Swerup and Rydqvist, 1992). However, the receptor currents probably originate in the very fine endings that are embedded in the muscle (see Fig. 5) so that accurate voltage-clamp of the receptor current is difficult. Voltage-clamp of arthropod cuticular receptors was first performed in the cockroach tactile spine (Torkkeli and French, 1994). A more successful cuticular preparation is that of spider slit sensilla, where it is possible to clamp the receptor ending close to the sensory ending while stimulating the slits (Juusola and French, 1995). The search for improved preparations continues, and recent development of the first successful tissue-cultured mechanoreceptor system offers the promise that we may eventually be able to study single receptor cells in vitro as they perform mechanotransduction and sensory encoding (Torkkeli and French, 1999).
Vii. Steps in Mechanoreception Mechanoreception can be viewed as a three-stage process comprising coupling, t r a n s d u c t i o n , and encoding (Eyzaguirre and Kuffier, 1955; Loewenstein, 1959). The stimulus is first mechanically coupled from its origin to the membrane of the sensory cell where it is transduced into a receptor potential, and if the cell is excitable this is encoded into action potentials, usually for transmission along a sensory axon.
A. Coupling Living cells are not normally exposed to the outside surface of an animal, so most, if not all, external mechanoreceptors have some kind of tissue between them and the source of the me-
45. Sensory Receptors and Mechanotransduction chanical input. In some cases these surrounding tissues are very elaborate and obviously designed to modify the input signal by affecting its amplitude and possibly its dynamic properties. Internal mechanoreceptors also have a variety of surrounding tissues with presumably similar functions. The morphology of these structures varies so strongly between different mechanoreceptors that it is widely used for classification, as described above. The Pacinian corpuscle is probably the most illustrated vertebrate mechanoreceptor, with its complex lamellar structure, but it represents only one extreme of a wide range of vertebrate mechanoreceptor structures (Malinovsky, 1996). Invertebrate mechanoreceptors also display many morphological forms, including the well-known hair receptors that cover the outside surfaces of insects and spiders, and the stretch receptors of crayfish and lobsters. Although we generally assume that these elaborate extracellular structures modify the spatial and temporal sensitivities of the receptors, there is relatively little quantitative information about these functions. An additional problem is to separate the effects of external structures from the mechanically activated ion channels themselves, since we do not yet understand the temporal properties of the channels, or how they are linked to other mechanical structures. In most cases the external structures attenuate the stimulus, so that the displacement of the receptor cell membrane is much smaller than the original movement. For example, hairs on the human skin can be moved by several millimeters, but the resuiting movements of the follicle receptors are only a few micrometers. Estimates of this attenuation have been made in some cases. For the Pacinian corpuscle, it has been suggested that the pressure reaching the inner capsule is attenuated by about two orders of magnitude (100) compared to the pressure on the outer lamellae (Bell et al., 1994), but this force may already be greatly reduced compared to the initial stimulus at the skin. An attenuation of about 100 was also reported for insect hairs, based on the morphology of the sensory structure and the location of the moving elements (French and Sanders, 1979). Estimates of attenuation during coupling also allow us to estimate the threshold, or minimum amplitude of movement at the cell membrane that leads to sensation. For Pacinian corpuscles, movements of 1-10 nm at the capsule are probably adequate to produce action potentials at the optimum vibration frequency of about 250 Hz, and sinusoidal movements of about 0.3 nm can stimulate auditory hair cells (Hudspeth, 1989). For insect cuticular hair receptors, threshold estimates of 4 nm and 3 nm were obtained for bee (Thurm, 1965) and cockroach (French and Sanders, 1979). The few attempts that have been made to quantify the mechanical effects of coupling over a range of input movements indicate that it can have significant temporal effects and be substantially nonlinear. The viscoelastic properties of Pacinian corpuscles are assumed to be responsible for most of the time dependence of the receptor current, but have not yet been quantified. The mechanical properties of crayfish stretch receptors have been studied very thoroughly (Swerup and Rydqvist, 1992) and require nonlinear springs plus at least one dashpot (or Voigt element) to account for the passive properties of the receptor without any muscular activity. No mechanical description is available that includes active contraction. Muscle spindles also have strongly timedependent mechanical properties, with probably more com-
767
plex dynamic behavior during 7-motoneuron activation (Matthews, 1981). B. Transduction
We assume that transduction in mechanoreceptors involves mechanically activated ion channels in the receptor cell membrane (see Chapter 44), but no single-channel recordings have yet been made that can reliably be linked to transduction in a mechanosensory neuron. Instead, most information about these channels comes from indirect measurements. Single, mechanically activated ion channels have been seen in crayfish stretch receptor neurons (Erxleben, 1989) and in tissue-cultured moth antennal mechanoreceptor neurons (Torkkeli and French, 1999), but we cannot be sure in either case that the channels were located on the fine sensory endings where transduction is thought to occur (Fig. 7). The crayfish channels had maximum conductance to potassium ions (-70 pS) but were also permeable to other monovalent and divalent cations. The moth channels were also permeable to potassium with a conductance of -40 pS, but their selectivity has not yet been established. Although we lack definitive single-channel measurements, we have information about some transduction channels, including the ions that flow through them, their ionic conductance, blocking chemicals, and the numbers of such channels in each receptor cell (Fig. 8). In Pacinian corpuscles, muscle spindles, and crayfish stretch receptors, development of a mechanical response depends upon the presence of sodium ions in the external solution, indicating that the channels are permeable to sodium (Loewenstein, 1971; Swerup and Rydqvist, 1992; Ito et al., 1983). However, a variety of other cations can
A 0 mm Hg
Crayfish stretch receptor
14 mm Hg 'i r r r- - lfT'" 10 pA 100 ms Manduca mechanoreceptor
20 mm Hg
I
10 pA 1000 ms
FIGURE 7. Single-channel recordings of mechanically activated ion channels in mechanoreceptor cells. (A) Channel in the membrane of a crayfish stretch receptor opens rarely at zero pressure but becomes very active with a suction of 14 mm Hg (channel opening downwards). (B) A channel in a tissue-cultured moth mechanoreceptor is activated by suction of 20 mm Hg. (Redrawn from Erxleben, 1989; Torkkeli and French, 1999).
768
SECTION V Synaptic Transmission and Sensory Transduction
Choline
Sodium
50
Amiloride
1.0
ms
--
#
"O (D
.__N 0.75 t~
o e-
L--
0.5-
L.
a) o n,'
0.25
0,0
-
--
I
10 -6
I 10 -5
I 10 -4
I 10 -3
Amilodde concentration (M)
FIGURE 8. Ionic selectivity and block of receptor current in two mechanoreceptors. (A) Step movements of a spider slit sensillum activate an inward current that is reduced to zero by replacing the sodium ions with impermeable choline, or by adding 1 mm amiloride to the bath. (B) Amiloride blocks the receptor current in mouse cochlear hair cells in a dose-dependent manner. (Redrawn from H6ger et al., 1997; Rtisch et al., 1994).
pass through the channels of Pacinian corpuscles (Bell et al., 1994) and crayfish stretch receptors (Rydqvist and Purali, 1993), and there is evidence that muscle spindle channels are also unselective (Ito et al.,1983). Spider slit sensilla neurons also pass sodium ions, but in this case the channels are very selective, rejecting most other cations except lithium (H6ger et al., 1997). In each of these cases, sodium entry during transduction causes a depolarization that leads to action potentials if the stimulus is strong enough, and each type of neuron has a receptor ending that is surrounded by an elevated concentration of sodium ions (H6ger et al., 1997). A different situation exists in vertebrate auditory and vestibular hair cells, where the channels are permeant to a wide range of cations, and the fluid bathing the sensory endings contains a high concentration of potassium ions (Hudspeth, 1989). Similarly, in insect cuticular mechanoreceptors, the sensory ending is surrounded by lymph that is rich in potassium ions and has a positive electrical potential relative to the normal hemolymph (French, 1988). The cases where potassium ions are used to carry the receptor current always seem to rely on separating
two regions of the sensory neurons by embedding them in an epithelial layer with tight junctions so that a potassium concentration gradient, usually combined with an electrical gradient, can drive potassium ions through the cell, depolarizing the membrane; similar arrangements exist for some of the sodium-dependent systems. Cells with nonselective receptor channels may also receive a significant calcium flux during transduction, because the inward calcium electrochemical gradient is usually high. Calcium influxes have been demonstrated in vertebrate hair cells (Ohmori, 1985) but their contributions to transduction or sensory adaptation are not clear in any mechanosensory cell. Like many other ion channels, mechanically activated channels can be blocked or closed by several chemical agents, although it is unfortunate for mechanoreceptor researchers that no evidence for highly specific blocking molecules has been found as for voltage-activated sodium channels and a variety of ligand-gated channels. The best known blocking chemical is the trivalent positive cation of gadolinium, and sensitivity to Gd 3§ is sometimes used as an indicator that mechanically activated channels are involved in a physiological process. Gadolinium blocked the mechanoreceptor current in the crayfish stretch receptor (Swerup et al., 1991) and the spider slit sensilla (H6ger et al., 1997). However, Gd 3§ is not specific for mechanically activated channels. It also blocks many calcium channels, as well as some ligand-gated and voltageactivated cation channels (H6ger et al., 1997). Amiloride is a well-known blocker of epithelial sodium channels, and it also blocks several types of mechanically activated ion channels (see Fig. 8). Amiloride blocks the receptor current in vertebrate hair cells (Riisch et al., 1994) and in spider slit sensilla neurons (H6ger et al., 1997), which has led to the suggestion that mechanoreceptor channels, and mechanically activated channels in general, may be members of the same gene family as the epithelial sodium channels. However, a direct molecular test of this theory in mammalian auditory hair cells was negative (Rtisch and Hummler, 1999). Some local anesthetics have also been used to block the receptor current in the crayfish stretch receptor (Lin and Rydqvist, 1999b) and certain antibiotics are known to block transduction in auditory hair cells (Hudspeth, 1989). In addition to the single-channel conductance estimates described above, total mechanoreceptor currents have been estimated in some cells, and noise, or fluctuation, analysis has been used to estimate single-channel conductance and the numbers of channels per cell in several cases. Noise analysis relies upon the idea that the variance of the receptor current arises from receptor channels opening and closing, because all known channels regulate their average conductance by switching rapidly between open and closed states. Therefore, the variance is minimal when all of the receptor channels are either fully open or fully closed, and maximal when the mean open probability is 0.5. Assuming that the noise variance is due to the summation of currents flowing through many independent, identical channels that are randomly opening and closing, it is possible to quantify the relationships between the membrane potential, total cell membrane current variance, total current, number of channels, single-channel conductance, and mean channel open probability. In a few cases, noise analysis measurements in nonreceptor systems have been confimaed by later single-channel
45. Sensory Receptors and Mechanotransduction
experiments. Noise analysis measurements are challenging because of the inherent difficulty of making good recordings of mechanoreceptor membrane currents. However, the three cases where they have been made have produced rather consistent results. In statocyst cells of the snail Hermissenda, noise analysis gave values of 5 pS for the single-channel conductance and 40 channels per cell (DeFelice and Alkon, 1977). In vertebrate hair cells, the corresponding values were 12 pS and 280 channels per cell (Hudspeth, 1989), and in spider slit sensilla neurons 7.5 pS and 250 channels per cell (HOger and French, 1999a). These single-channel conductances are all at the low end of the range of known mechanically activated ion channels and significantly lower than the only two single-channel measurements described above. The total membrane conductances due to sensory channels observed during these measurements agree with several estimates made in other mechanosensory neurons, and would provide enough current to depolarize the neurons to produce action potentials. Therefore, it seems likely that the number of receptor channels in each cell is limited to a few hundred at most, which will make them difficult to locate by patch-clamp methods. Another interesting property of mechanotransduction currents is their temperature sensitivity. This has been measured in at least six invertebrate and vertebrate preparations with activation energy values in the range of 12-22 kcal/mol (-50-100 kJ/mol) (H6ger and French, 1999b). These energy values are similar to those required to break chemical bonds, and significantly more than the energy barriers associated with ionic diffusion or conductance through ionic channels. Little is known about the activation energy of mechanically activated ion channels, but it seems likely that some crucial stage in the link between membrane tension and ion channel opening leads to this relatively high energetic barrier.
C. Encoding Most mechanoreceptors use action potentials to transmit information to the central nervous system, and the action potentials are produced by the conventional combination of voltagedependent sodium and potassium channels. Accurate characterization of these currents and other currents involved in the control of membrane excitability relies on voltageclamp recordings, which have only been possible in a few cases. Tetrodotoxin-sensitive sodium channels are clearly responsible for the action potential upswing in Pacinian corpuscles (Loewenstein, 1959), lobster stretch receptors (Edman et al., 1987a), crayfish stretch receptors (Swerup and Rydqvist, 1992), the cockroach tactile spine (Torkkeli and French, 1994), and spider slit sensilla (Seyfarth and French, 1994). The distribution of these channels in the cell membrane probably determines where the receptor potential is converted to action potentials. In the slowly adapting crayfish stretch receptor the channels are mainly located on the axon and the immediately adjacent cell body (Swerup and Rydqvist, 1992), but in the rapidly adapting receptor the channels are restricted to the axon itself (Lin and Rydqvist, 1999a). In spider slit sensilla, sodium channels are more evenly distributed, with a significant concentration on the sensory dendrite and strong excitability in the cell soma (Seyfarth et al., 1995). An addi-
769
tional slow component of sodium inactivation was needed to model the firing behavior of lobster stretch receptors (Edman et al., 1987a) and is probably present in the slowly adapting crayfish stretch receptor (Purali and Rydqvist, 1998) and the cockroach tactile spine (French and Torkkeli, 1994). Delayed-rectifier potassium currents seem to be responsible for action potential repolarization in all of the mechanoreceptors studied, but other types of potassium currents are unevenly distributed among different receptors. Rapidly inactivating potassium currents resembling the A-type current have been described in the cockroach tactile spine (Torkkeli and French, 1994) but not in spider slit sensilla (Sekizawa et al., 1999) or crayfish stretch receptors (Swerup and Rydqvist, 1992). Calcium-activated potassium currents were found in the cockroach tactile spine (Torkkeli and French, 1995) and vertebrate auditory hair cells (Hudspeth, 1989) but not spider slit sensilla (Sekizawa et al., 1999). Single calcium-activated channels were seen in crayfish stretch receptors (Erxleben, 1993), although there have been no reports of whole-cell currents. An inwardly rectifying potassium current is present in the slowly adapting lobster stretch receptor (Edman et al., 1987b) but has not been described elsewhere. Calcium currents are present in auditory neurons, and negative feedback from depolarization-induced calcium entry to hyperpolarization via calcium-activated potassium current has been suggested to cause frequency tuning (Hudspeth, 1989). Calcium currents have also been measured in spider slit sensilla (Sekizawa et al., 2000), although these cells do not seem to have calcium-activated potassium currents. Electrogenie sodium pumping is well established in crayfish stretch receptors (Sokolove and Cooke, 1971) and contributes to adaptation of action potential discharge by repolarizing the cell after sodium entry. There is also evidence for its contribution to adaptation in the cockroach tactile spine neuron (French, 1989). The terms rapid adaptation and slow adaptation are qualitative and have been used to describe a wide range of adaptation behaviors. However, some mechanoreceptors adapt so rapidly that only one or a few action potentials are produced in response to a step stimulus (Fig. 9) and so can truly be called "rapidly adapting." It has been known for many years that the encoding stage of the Pacinian corpuscle limits its response to one or two action potentials (Loewenstein and Mendelson, 1965), although the ionic basis for the effect is not known. In the cockroach tactile spine, removal of the rapidly inactivating potassium A-current increases the overall rate of firing, but the receptor continues to adapt, while removal of the calciumactivated potassium current removes most of the adaptation (Torkkeli and French, 1994, 1995). However, blockade of these currents does not affect the shape of individual action potentials (Fig. 10) because their rapid repolarization is due to the delayed-rectifier current. Following the encoding stage, action potentials propagate into the central nervous system along nerve axons. In vertebrates, the cell bodies of the neurons are located in the dorsal root ganglia of the spinal cord, and the axons enter via the dorsal roots. Mechanoreceptor axons are generally in the Ace, and A/3 groups of myelinated fibers with conduction velocities in the range of 40-120 m/s. Bare nerve endings (Group Ia
770
SECTION V Synaptic Transmission and Sensory Transduction
VIII. Efferent Control of Mechanoreceptors Pacinian corpuscle
Tactile spine
Slit sensillum
[
_J
100
ms
FIGURE 9. Rapidadaptation in three mechanoreceptors. Following a step mechanical stimulus, a Pacinian corpuscle, cockroach tactile spine, and spider slit sensillum each fire a few action potentials and are then silent. While tactile spine receptors can fire longer bursts with large movements, Pacinian corpuscles and spider slit sensilla cannot normally produce more than two action potentials, however strong the stimulation. (Redrawn from French and Torkkeli, 1994).
of Fig. 3) are innervated by myelinated A8 or unmyelinated C fibers, with conduction velocities in the range 0.5-30 m/s. The major transmitter at the first central synapse is probably the excitatory amino acid glutamate (Willis and Coggeshall, 1991). Arthropod mechanoreceptors have their cell bodies in the periphery and send axons into the central nervous system via nerve roots of the segmental ganglia, which are sometimes fused into larger structures. Conduction velocities are typically 1-5 m/s. The dominant transmitter for arthropod mechanoreceptors is acetylcholine (Burrows, 1996).
trol
iI}lllJl] 50mV
2s
~IrIIr r r r
.
.
.
.
.
I~
1111111!1111
.....................~.mlIm,m~ mn
FIGURE 10. Removal of adaptation during encoding by potassium channel block. (A) A rapidly adapting burst of action potentials in the cockroach tactile spine, following a step depolarization of the neuron. (B) 4-aminopyridine removes the rapidly inactivating potassium A-current, making the neuron fire much more, but the underlying adaptation can still be seen. (C) Cadmium blocks calcium entry and the calcium-activated potassium current, removing adaptation. Note that the shape of the individual action potentials (left) is not strongly affected by either treatment because its repolarization depends on a delayed-rectifier potassium current. (Redrawn from French and Torkkeli, 1994).
All mechanoreceptors studied so far receive inhibitory efferent innervation close to the output synapses of their centrally located axon terminals. Well-known examples are the efferent innervation of muscle spindle and tendon organ (Frank and Fuortes, 1957). Peripheral mechanoreceptors (Semba et al., 1985) and pain receptors (Schmidt and Schaible, 1998) receive similar synaptic inputs. This type of presynaptic innervation of mechanosensory neurons is remarkably similar in all vertebrate and invertebrate species studied so far (Frank and Fuortes, 1957; Semba et al., 1985; Cattaert et al., 1992; Pearson, 1993; Burrows, 1996). Most thoroughly studied of these synapses are GABAergic ones in insects (Burrows, 1996), crustaceans (Cattaert et al., 1992), and vertebrates (Nicoll and Alger, 1979). Evidence is now accumulating that modulation of mechanosensory neurons is more complex than the inhibitory control of axon terminals, with synaptic regulation of the sensory endings themselves and the sensory cell bodies. Transmitter receptors are found in the vertebrate dorsal root ganglion neuron somata (Robertson, 1989), and unmyelinated nerve fibers are seen in slowly adapting cutaneous mechanoreceptors and in the capsules of Pacinian corpuscles. Some of these endings are probably efferent and have been suggested to modulate sensory reception (Bell et al., 1994). There is also evidence that the Merkel cell functions as a secretory cell rather than a sensory neuron and modulates the associated nerve ending (Zaccone et al., 1994). Modulation of vertebrate cutaneous mechanoireceptors and pain receptors by circulating factors has also been recognized for a long time. For example, sympathetic efferents have modulatory effects on muscle spindles, Pacinian corpuscles (Akoev, 1980), several types of lowthreshold skin mechanoreceptors, and pain receptors (Raja, 1995), and these neurons express adrenergic receptors (Birder and Per, 1999). Rohon-Beard neurons, which are developmentally early amphibian and fish touch receptors, have serotonergic efferent innervation and 5-HT is believed to alter their sensitivity to mechanical stimuli by inhibiting both low- and high-voltage-activated calcium currents (Sun and Dale, 1997). Insect mechanosensory neurons are modulated by biogenic amines, especially octopamine (Burrows, 1996). The dorsal unpaired median neurons of locust, which secrete octopamine, were recently shown to have terminals in the periphery, closely associated with the dendrites of mechanoreceptors (Br/~unig and Eder, 1998). Previous findings of octopamine effects in situations where axon terminals were removed during recordings (Zhang et al., 1992) also suggest that octopamine receptors are located in the neurons' peripheral parts. Although octopamine modulation of insect mechanoreceptors is probably the most thoroughly studied neuromodulatory system, there are no clear conclusions about the mechanisms involved. Octopamine may increase or decrease spiking frequency, even in the same neuron (Br~iunig and Eder, 1998). The most conclusive evidence that peripheral regions of mechanoreceptors can receive efferent innervation comes from immunocytochemical findings in arachnid and crustacean mechanoreceptors (see Fig. 5; Fig. 11A). Both types of receptors have dense efferent innervation, forming several types of synapses on all parts of the sensory neurons, includ-
45. Sensory Receptors and Mechanotransduction
771
FIGURE 1 1. Efferent innervation of mechanosensory neurons has been mapped by immunohistochemistry in spider slit sensilla. (A) Schematized distribution of synapses surrounding a pair of neurons innervating one slit. Several fine, efferent fibers run parallel to the neurons and form numerous synapses onto their initial axon segments, somata, and dendrites, as well as onto surrounding glial cells and onto each other. The density of synapses on the dendrites is lower than other regions. At least three types of efferent fibers show immunoreactivity to GABA, and at least one efferent fiber type shows immunoreactivity to glutamate. (B) Four types of efferent synapses onto sensory neurons. Numerous other synaptic combinations between efferents, sensory neurons, and glia have been demonstrated. (Redrawn from Fabian-Fine, 1998).
ing their somata and dendrites (Foelix, 1975; Elekes and Florey, 1987; Fabian-Fine, 1998; Fabian-Fine et al., 1999). The spider slit sensilla neurons and other cuticular mechanoreceptors have at least four different types of synaptic structures on the efferent terminals, based on the shapes and sizes of synaptic vesicles and immunogold labeling (Fig. 11B). These different types of synapses could have several functions, including direct inhibition or excitation of mechanosensory neurons and sensitization or desensitization. They could also control the dynamic responses of the neurons and affect their transmission of repetitive signals. It seems likely that central nervous systems are able to adjust the responsiveness of many types of mechanosensory neurons, although the functional significance of this control is not yet evident.
IX. Summary The large variety of morphological forms of mechanoreceptors indicates that the extemal components surrounding the sensitive endings play important roles in controlling the receptor behavior. Most of this control is assumed to involve coupling of initial movement into deformation of the membrane containing mechanically activated channels. However, this has not been thoroughly investigated or modeled in any receptor, and there may be other functions involved, such as chemical modulation of excitability. Characterization of mechanically activated channels is proceeding, but it is important to realize that no single-channel recordings have yet been unequivocally linked to mechanotransduction in any mechanosensory neuron. Therefore, it may yet emerge that an unknown family of channel proteins is responsible for this function in true mechanoreceptors. The encoding of action potentials from receptor current and the general control of receptor excitability involves voltage- and calcium-activated ion
channels that are very similar to those in other tissues, and complete models of encoding in some receptors are now available. Adaptation, and general dynamic properties of mechanoreception, occur in both the coupling and encoding stages of the process. There may also be significant dynamic behavior of the mechanically activated channels, but this will not become clear until the channels themselves are better known. Efferent control of transduction and adaptation in mechanoreceptors is probably common but generally not well understood. Mechanoreception is a widespread and crucially important process for many physiological functions. Recent developments in electrophysiological techniques and new experimental preparations have provided important information about each stage of mechanoreception from initial deformation to action potential production, but much remains to be learned. In particular, the lack of vertebrate preparations that would allow voltage-clamp of receptor currents or detailed examination of action potential encoding is a major problem. It is also true that the molecular biological revolution has so far failed to play an important role in unraveling the machinery of mechanoreception, except insofar as providing general information about some of the families of voltage-activated ion channels that are involved in encoding. When the molecules responsible for mechanotransduction are known, it should be possible to discover how they are linked to other cellular components to confer mechanical sensitivity, and from there to move on to complete models of mechanoreception in intact cells.
Bibliography Akoev, G. N. (1980). Catecholamines, acetylcholine and excitability of mechanoreceptors. Prog. Neurobiol. 15, 269-294.
772
Bell, J., Bolanowski, S., and Holmes, M. H. (1994). The structure and function of Pacinian corpuscles: a review. Prog. Neurobiol. 42, 79-128. Birder, L. A., and Perl, E. R. (1999). Expression of c~2-adrenergic receptors in rat primary afferent neurones after peripheral nerve injury of inflammation. J. Physiol. (Lond). 512, 533-542. Br~iunig, P., and Eder, M. (1998). Locust dorsal unpaired median (DUM) neurones directly innervate and modulate hindleg proprioceptors. J. Exp. Biol. 201, 3333-3338. Burrows, M. (1996). "The Neurobiology of an Insect Brain." Oxford University Press, Oxford. Cattaert, D., Manira, A. E., and Clarac, E (1992). Direct evidence for presynaptic inhibitory mechanisms in crayfish sensory afferents. J. Neurophysiol. 67, 610-623. DeFelice, L. J., and Alkon, D. L. (1977). Voltage noise from hair cells during mechanical stimulation. Nature 269, 613--615. Edman, A., Gestrelius, S., and Grampp, W. (1987a). Analysis of gated membrane currents and mechanisms of firing control in the rapidly adapting lobster stretch receptor neurone. J. Physiol. 384, 649-669. Edman, A., Gestrelius, S., and Grampp, W. (1987b). Current activation by membrane hyperpolarization in the slowly adapting lobster stretch receptor neurone. J. Physiol. 384, 671-690. Elekes, K., and Florey, E. (1987). Immunocytochemical evidence for the GABAergic innervation of the stretch receptor neurons in crayfish. Neuroscience 22, 1111-1122. Erxleben, C. E (1989). Stretch-activated current through single ion channels in the abdominal stretch receptor organ of the crayfish. J. Gen. Physiol. 94, 1071-1083. Erxleben, C. E (1993). Calcium influx through stretch-activated cation channels mediates adaptation by potassium current activation. Neuroreport 4, 616-618. Eyzaguirre, C., and Kuffier, S. W. (1955). Processes of excitation in the dendrites and in the soma of single isolated sensory nerve cells of the lobster and crayfish. J. Gen. Physiol. 39, 87-119. Fabian-Fine, R. (1998). Thesis: Synaptische Kontakte an sensorischen Neuronen von Spinnen: Immuncytochemische und ultrastrukturelle Untersuchungen. Erlangung des Doktorgrades, Fachbereich Biologic, Johann Wolfgang Goethe-Universit~it, Frankfurt am Main, 1-133. Fabian-Fine, R., H6ger, U., Seyfarth, E. A., and Meinertzhagen, I. A. (1999). Peripheral synapses at identified mechanosensory neurons in spiders: three-dimensional reconstruction and GABA immunocytochemistry. J. Neurosci. 19, 298-310. Fein, A., and Szuts, E. Z. (1982). "Photoreceptors: Their Role in Vision." Cambridge University Press, Cambridge. Foelix, R. E (1975). Occurrence of synapses in peripheral sensory nerves of arachnids. Nature 254, 146-148. Frank, K., and Fuortes, M. G. E (1957). Presynaptic and postsynaptic inhibition of monosynaptic reflexes. Fed. Proc. 16, 39-40. French, A. S. (1988). Transduction mechanisms of mechanosensilla. Annu. Rev. Entomol. 33, 39-58. French, A. S. (1989). Ouabain selectively affects the slow component of sensory adaptation in an insect mechanoreceptor. Brain Res. 504, 112-114. French, A. S., and Sanders, E. J. (1979). The mechanism of sensory transduction in the sensilla of the trochanteral hair plate of the cockroach, Periplaneta americana. Cell Tiss. Res. 198, 159-174. French, A. S., and Torkkeli, P. H. (1994). The basis of rapid adaptation in mechanoreceptors. News Physiol. Sci. 9, 158-161. French, A. S., and Torkkeli, E H. (1998). Information transmission at 500 bits/s by action potentials in a mechanosensory neuron of the cockroach. Neurosci. Lett. 243, 113-116. Hrger, U., and French, A. S. (1999a). Estimated single-channel conductance of mechanically activated channels in a spider mechanoreceptor. Brain Res. 826, 230-235. Hrger, U., and French, A. S. (1999b). Temperature sensitivity of transduction and action potential conduction in a spider mechanoreceptor. Pfliigers Arch. 438, 837-842.
SECTION V Synaptic Transmission and Sensory Transduction
Hrger, U., Torkkeli, E H., Seyfarth, E.-A., and French, A. S. (1997). Ionic selectivity of mechanically activated channels in spider mechanoreceptor neurons. J. Neurophysiol. 78, 2079-2085. Hudspeth, A. J. (1989). How the ear's works work. Nature 341, 397-404. Ito, E, Fujitsuka, N., and Kim, N. (1983). The spindle potential in the frog muscle spindle does not require external Na +. Brain Res. 277, 352-354. Juusola, M., and French, A. S. (1995). Recording from cuticular mechanoreceptors during mechanical stimulation. Pfliigers Arch. 431, 125-128. Juusola, M., and French, A. S. (1997). The efficiency of sensory information coding by mechanoreceptor neurons. Neuron 18, 959-968. Lin, J. H., and Rydqvist, B. (1999a). Different spatial distributors of sodium channels in the slowly and rapidly adapting stretch receptor neuron of the crayfish. Brain Res. 830, 353-357. Lin, J. H., and Rydqvist, B. (1999b). The mechanotransduction of the crayfish stretch receptor neurone can be differentially activated or inactivated by local anaesthetics. Acta Physiol. Scand. 166, 65-74. Loewenstein, W. R., (1959). The generation of electrical activity in a nerve ending. Ann. NY Acad. Sci. 81, 367-387. Loewenstein, W. R. (1971). Mechano-electric transduction in the Pacinian corpuscle: initiation of sensory impulses in mechanoreceptors. Handbook Sensory Physiol. 1, 269-290. Loewenstein, W. R., and Mendelson, M. (1965). Components of receptor adaptation in a Pacinian corpuscle. J. Physiol. 177, 377-397. Malinovsky, L. (1996). Sensory nerve formations in the skin and their classification. Micro. Res. Tech. 34, 283-301. Matthews, R B. (1981). Evolving views on the internal operation and functional role of the muscle spindle. J. Physiol. 320, 1-30. Mclver, S. B. (1985). Mechanoreception. In "Comprehensive Insect Physiology Biochemistry and Pharmacology" (G. A. Kerkut and L. I. Gilbert, Eds.), pp. 71-132. Pergamon Press, Oxford. Nicoll, R. A., and Alger, B. E. (1979). Presynaptic inhibition: transmitter and ionic mechanisms. Int. Rev. Neurobiol. 21, 217-258. Ohmori, H. (1985). Mechano-electrical transduction currents in isolated vestibular hair cells of the chick. J. Physiol. 359, 189-217. Pasztor, V. M., and Bush, B. M. (1982). Impulse-coded and analog signalling in single mechanoreceptor neurons. Science 215, 1635-1637. Pearson, K. G. (1993). Common principles of motor control in vertebrates and invertebrates. Annu. Rev. Neurosci. 16, 265-297. Perl, E. R. (1996). Cutaneous polymodal receptors: characteristics and plasticity. Prog. Brain Res. 113, 21-37. Purali, N., and Rydqvist, B. (1998). Action potential and sodium current in the slowly and rapidly adapting stretch receptor neurons of the crayfish (Astacus astacus). J. Neurophysiol. 80, 2121-2132. Rtisch, A., and Hummler, E. (1999). Mechano-electrical transduction in mice lacking the a-subunit of the epithelial sodium channel. Hear. Res. 131, 170-176. Rtisch, A., Kros, C. J., and Richardson, G. P. (1994). Block by amiloride and its derivatives of mechano-electrical transduction in outer hair cells of mouse cochlear cultures. J. Physiol. 474, 75-86. Raja, S. N. (1995). Role of the sympathetic nervous system in acute pain and inflammation. Ann. Med. 27, 241-246. Robertson, B. (1989). Characteristics of GABA-activated chloride channels in mammalian dorsal root ganglion neurones. J. Physiol. (Lond.) 411, 285-300. Rydqvist, B., and Purali, N. (1993). Transducer properties of the rapidly adapting stretch receptor neurone in the crayfish. J. Physiol. 469, 193-211. Sachs, E (1997). Mechanical transduction by ion channels: how forces reach the channel. In "Cytoskeletal Regulation of Membrane Function" (S. C. Froehner, Ed.), pp. 209-218. Rockefeller University Press, New York. Schmidt, R. F., and Schaible, H.-G. (1998). Modulation of nociceptive information at the presynaptic terminals of primary afferent fibers. In "Presynaptic Inhibition and Neural Control" (E Rudomin, R.
45. Sensory Receptors and Mechanotransduction
Romo, and L. M. Mendell, Eds.), pp. 424-439. Oxford University Press, New York. Sekizawa, S. I., French, A. S., H6ger, U., and Torkkeli, E H. (1999). Voltage-activated potassium outward currents in two types of spider mechanoreceptor neurons. J. Neurophysiol. 81, 2937-2944. Sekizawa, S. I., French, A. S., and Torkkeli, E H. (2000). Low-voltage activated calcium current does not regulate the firing behavior in paired mechanosensory neurons with different adaptation properties. J. Neurophysiol. 83, 746-753. Semba, K., Masarachia, E, Malamed, S., Jacquin, M., Harris, S., Yang, G., and Egger, M. D. (1985). An electron microscopy study of terminals of rapidly adapting mechanoreceptive afferent fibers in the cat spinal cord. J. Comp. Neurol. 232, 229-240. Seyfarth, E.-A., and French, A. S. (1994). Intracellular characterization of identified sensory cells in a new spider mechanoreceptor preparation. J. Neurophysiol. 71, 1422-1427. Seyfarth, E.-A., Sanders, E. J., and French, A. S. (1995). Sodium channel distribution in a spider mechanosensory organ. Brain Res. 683, 93-101. Shannon, C. E., and Weaver, W. (1949). "The Mathematical Theory of Communication." University of Illinois Press, Urbana, IL. Sokolove, P. G., and Cooke, I. M. (1971). Inhibition of impulse activity in a sensory neuron by an electrogenic pump. J. Gen. Physiol. 57, 125-163. Sun, Q. Q., and Dale, N. (1997). Serotonergic inhibition of the T-type and high voltage-activated Ca 2+ currents in the primary sensory neurons of Xenopus larvae. J. Neurosci. 17, 6639-6649.
773
Swerup, C., and Rydqvist, B. (1992). The abdominal stretch receptor organ of the crayfish. Comp. Biochem. Physiol. A 103, 423-431. Swerup, C., Purali, N., and Rydqvist, B. (1991). Block of receptor response in the stretch receptor neuron of the crayfish by gadolinium. Acta Physiol. Scand. 143, 21-26. Thurm, U. (1965). An insect mechanoreceptor I. Fine structure and adequate stimulus. Cold Spring Harbour Syrup. Quant Biol. 30, 75-94. Torkkeli, P. H., and French, A. S. (1994). Characterization of a transient outward current in a rapidly adapting insect mechanoreceptor neuron. Pfliigers Arch. 429, 72-78. Torkkeli, P. H., and French, A. S. (1995). Slowly inactivating outward currents in a cuticular mechanoreceptor neuron of the cockroach (Periplaneta americana). J. Neurophysiol. 74, 1200-1211. Torkkeli, P. H., and French, A. S. (1999). Primary culture of antennal mechanoreceptor neurons of Manduca sexta. Cell Tiss. Res. 297, 301-309. Willis, W. D., and Coggeshall, R. E. (1991). "Sensory Mechanisms of the Spinal Cord." Plenum Press, New York. Zaccone, G., Fasulo, S., and Ainis, L. (1994). Distribution patterns of the paraneuronal endocrine cells in the skin, gills and the airways of fishes as determined by immunohistochemical and histological methods. Histochem. J. 26, 609-629. Zhang, B. G., Torkkeli, P. H., and French, A. S. (1992). Octopamine selectively modifies the slow component of sensory adaptation in an insect mechanoreceptor. Brain Res. 591, 351-355.
Daniel C. Marcus
Acoustic Transduction
The detection of sound by mammals depends on a series of biological systems beginning with the collection of sound pressure waves by the external ear, followed by the mechanical transmission through the middle ear ossicles to the cochlea of the inner ear. The sound pressure waves in the cochlea induce motion of the basilar membrane to which the sensory organ, the organ of Corti, is attached. Motion of the organ of Corti with respect to another structure (tectorial membrane) causes movements of the sensory cilia (hairs) on the hair cells and modulation of the flow of current through these hair cells, leading to modulation of the rate of firing of the afferent auditory nerve fibers, which synapse to the base of the hair cells. The nerve fibers carry the auditory information to the brain where the signal undergoes central processing, leading to the perception of sound. Peripheral auditory processing is modulated by efferent signals originating from the brain. This chapter focuses on the cellular aspects of acoustic transduction in the auditory periphery. Much of what we know about the function of the mammalian cochlea is derived from experiments performed on preparations from the vestibular labyrinth of mammals, birds, and amphibians, and from the cochlea of birds.
on the basilar membrane, a fibrous sheet that transmits the acoustic stimulus. In the medial direction from the organ of Corti, the duct is comprised of inner sulcus cells and interdental cells of the spiral limbus. The tectorial membrane is a gelatinous, acellular structure in the cochlear lumen apparently secreted by the interdental cells. Lateral from the organ of Corti, the cochlear duct consists of outer sulcus cells, spiral prominence cells, and marginal cells of the stria vascularis. Reissner's membrane forms the remaining wall of the triangular-shaped cochlear duct and is comprised of a thin, avascular sheet of epithelial cells. The apical and basolateral membranes of all of these cells are separated by tight junction complexes near the endolymphatic surface, which serve to join each cell to its neighbor and to complete the barrier between endolymph and the fluid bathing the basolateral membranes. The basolateral fluid is perilymph for all cell types except strial marginal cells, as described later. The cochlear duct is closed at the apex of the cochlea and is joined at the base of the cochlea via a constriction in the epithelial lumen (ductus reuniens) to the vestibular system. Cellular physiologists have focused most of their attention on the strial marginal cells, which provide the energy source for the transduction process, and on the sensory hair cells themselves.
I!. M a m m a l i a n Inner Ear Structure
!II. Cell Physiology of Endolymph H o m e o s t a s i s
!. Introduction
The transduction apparatus is part of an epithelium forming the cochlear duct and separating two distinct cochlear fluids, endolymph and perilymph. The composition and importance of these fluids is related later. A diagram of a cross-section of the mammalian cochlear duct is shown in Fig. 1. The organ of Corti is comprised of the sensory inner and outer hair cells (Fig. 11, shown later), which are surrounded by Deiters' cells, pillar cells, and Hensen's cells. A mechanically stiff apical surface of the organ of Corti, the reticular lamina, is formed by cuticular plates just under the apical membrane of these cells. The organ of Corti sits
Cell Physiology Sourcebook:A Molecular Approach, Third Edition
7 7 5
A. Composition Even in the absence of an acoustic stimulus, the cochlea is highly active, maintaining the electrolyte composition of endolymph (Table 1) and a standing current analogous to the "dark current" of photoreceptors (Chapter 48) (Wangemann and Schacht, 1996). The transduction current from endolymph through the hair cells is carried by K + and so depends on the high concentration of that ion in endolymph. Gross changes in endolymph composition either by a tear in Reissner's membrane or by genetic interference in K + secretion lead to degeneration of the hair cells with subsequent
Copyright 9 2001 by Academic Press.All rights of reproduction in any form reserved.
SECTION V Synaptic Transmission and Sensory Transduction
776
TABLE 1 Approximate Ion Composition and Electrical Potential of Cochlear Endolymph and Perilymph a Ion
Endolymph
Potassium (mM)
Perilymph
157
5
Sodium (mM)
1
145
Calcium (mM)
0.02
Chloride (mM) Bicarbonate (mM) pH
1
132
125
31
25
7.4
Potential (mV)
7.3
+80
0
aFrom Wangemann and Schacht, 1996. Copyright 1996 by SpringerVerlag.
FIGURE I. Diagram of the epithelial cell boundary of the cochlear duct. The triangular cross-section is about 0.3--0.5 mm on each side. The lateral wall is the site of the K+-secretory stria vascularis, a tissue consisting of two barrier layers of strial marginal cells (SMCs) and basal cells (BCs), which enclose the intrastrial space (ISS). The organ of Corti sits on the fibrous basilar membrane (BM), forms most of the wall separating scala media and scala tympani, and consists of the inner hair cells (IHCs), outer hair cells (OHCs), and several supporting cell types including Deiters' cells (DCs). The sensory hairs at the apical membranes of the OHC project into the gelatinous tectorial membrane (TM), whereas for the most part those of the IHC do not. The spiral prominence (SP) and outer sulcus (OS) epithelial ceils lie on the lateral wall between the SV and organ of Corti. The OS forms a para-sensory pathway for the absorption of cations from endolymph. The spiral ligament is the connective tissue between the stria vascularis and the bone (not shown). (Adapted with permission from Wangemann, 1997, Kalium-Ionensekretion und Entstehung des endokochlearen Potentials in der Stria vascularis. HNO 45, 205-209. Copyright 1997 by Springer-Verlag.)
transduction current. The driving force for that current consists almost exclusively of the voltage across the transduction channels in the stereocilia (Section IV) and that voltage is the sum of the intracellular potential of the hair cells and the EP. The large EP therefore heightens the sensitivity of the cochlear transduction process compared to vestibular organs, which do not have this high transepithelial electrical polarization. The cellular transport model by which the stria vascularis secretes potassium and generates the EP is shown in Fig. 2 and described next. Note that all epithelial cells that produce a vectorial transport of substances do so by virtue of different membrane properties of their apical and basolateral membranes.
Stria vascularis !
Marginal cell
Basal cell
+80 m~
+80 mV
0mV
II
[]
K
K Na K~--- ----
Na
CI K
degeneration of the synapsing afferent auditory nerve (Vetter et al., 1996). The dimensions of the gelatinous tectorial membrane are sensitive to the levels of K +, Na +, H +, and Ca 2+ as well as to unknown chemical factors (Shah et al., 1995). Swelling has been associated with substitution of Na + for K +, and with decreases of Ca 2+. Movements of water into and out of the membrane are believed to be related to the level of screening of fixed charges.
B. Stria Vascularis The stria vascularis has two primary known functions: secretion of K + and generation of the lumen-positive endocochlear potential (EP; ca. +80 mV). The high endolymphatic K § concentration provides the carrier of the
Endoiymph BE ap
la Intrastria! ! bl Compartme
ymph
F I G U R E 2. Model of ion transport by the stria vascul~s. The marginal cell epithelium secretes K + into endolymph, and the basal cell layer produces the voltage (endocochlear potential) between endolymph and perilymph. The ion transport processes ascribed to the marginal cells have a strong experimental basis. The transport processes proposed in the model of the basal cells have not yet been experimentally demonstrated and may be distributed among basal cells, intermediate cells of the stria vascularis, and fibrocytes of the spiral ligament, which are all connected by gap junctions. Arrows indicate ion channels; open circle, ion carrier; filled circles, primary-active ion "pump." (Adapted from Hear Res. 90, Wangemann, P., Comparison of ion transport mechanisms between vestibular dark cells and strial marginal cells, pp. 149-157, Copyright 1995 with kind permission of Elsevier Science-NL, Sara Burgerhartstraat 25, 1055 KV Amsterdam, The Netherlands.)
46. Acoustic Transduction
200"
777 Bum iO pM 2 min
> v
== lOO 2o .
FIGURE 3. K+ secretory flux from stria vascularis measured with the K+-selective self-referencing probe. The ordinate reflects the difference in microvolts measured at the two positions of the probe. Flux was decreased by basolateral perfusion of the inhibitor of Na§ cotransport, bumetanide (Bum). (Adapted from Hear. Res. 84, Wangemann, P., Liu, J., and Marcus, D. C., Ion transport mechanisms responsible for K§ secretion and the transepithelial voltage across marginal cells of stria va,scularis in vitro, pp. 19-29, Copyright 1995 with kind permission of Elsevier Science-NL, Sara Burgerhartstraat 25, 1055 KV Amsterdam, The Netherlands.)
1. Division of Function between Marginal and Basal Cells The stria vascularis consists of two barriers formed by the marginal cells and the basal cells. Each barrier consists of a continuous sheet of cells joined by tight junction complexes (Fig. 2). Between these barriers is the intrastrial space with the capillary bed for which the tissue is named and a discontinuous layer of intermediate cells. The basal cells are joined via gap junctions (Chapter 31) to the intermediate cells and to fibrocytes in the adjacent connective tissue, suggesting a level of cooperation among these three cell types (Kikuchi et al., 1995). Unlike most sheets of epithelial cells, strial marginal cells are not coupled to each other nor to other cells by gap junctions (Kikuchi et al., 1995; Sunose et al., 1997). The physiological significance of this functional independence of marginal cells is not known. There is strong evidence that there is a division of function between strial marginal cells and basal cells. The basal cell barrier produces and supports the endocochlear potential and the strial marginal cells secrete K +. In spite of the separation of cellular function, the two processes are closely tied together through the composition of the intrastrial space. For example, inhibition of K + secretion by the marginal cells leads to a decline of the endocochlear potential, presumably due to a rise in intrastrial K + concentration and the consequent depolarization of the intrastrial membrane of the basal cells. Several lines of indirect evidence, including histochemical, biochemical, electrophysiological, and flux studies, have strongly suggested that the stria vascularis is responsible for secretion of K + into the cochlear lumen and for production of the EP (reviewed in Wangemann and Schacht, 1996). Recent evidence of a more direct nature is presented here.
duced in either the perilymphatic or vascular space and its appearance in endolymph (Konishi et al., 1978; Sterkers et al., 1982). These fluxes were inhibited by transport blockers and anoxia. The stria was assumedto be the site of secretion since the flux was nearly the same when the radiotracer was added to either scala vestibuli or scala tympani; possible contributions by the spiral limbus were disregarded. More recently, the stria was isolated from the cochlear duct, a K+-selective self-referencing probe (see Appendix at the end of this chapter) was placed near the tissue in vitro and a K § gradient was found directed away from the luminal surface (Wangemann et al., 1995a) (Fig. 3). The marginal cell layer of the stria vascularis produces this K § flux using the constellation of transport processes shown in Fig. 2. The cell model is identical to that for the vestibular dark cells, the vestibular homolog of strial marginal cells (Wangemann, 1995). K + is taken up across the basolateral m e m b r a n e from the i n t r a s t r i a l space fluid by two transporters: the Na+,K+-ATPase and the Na+-K+-C1-cotransporter. The first is a primary-active process, which uses the energy from cytosolic adenosine triphosphate (ATP) and the second is secondary-active and uses the large Na § concentration gradient between extracellular and intracellular compartments created by the Na+,K+-ATPase. Na § and C1-taken up by the cotransporter are removed from the cytosol by the Na+,K+-ATPase and basolateral C1channels, respectively (Wangemann and Schacht, 1996). K + secretion occurs passively via channels in the apical membrane (Fig. 4). These channels are products of the genes KCNQ1 and KCNE1 for the KvLQT1 K + channel and its beta subunit, IsK (or mink). The resulting channel complex* and current carried by the channel are often referred to as IKs. The essential contribution of the IsK protein to hearing and balance is most dramatically illustrated by the development of a mouse in which the KCNE1 (isk) gene was knocked out (Vetter et al., 1996). In these mice the entire endolymphatic space collapsed (Fig. 5) after the point in embryonic development at which endolymph secretion normally commences,
*Historically, the channel and gene were first named after the beta subunit before the channel itself was discovered.
100-
Q,. v ..i-, t--
9
50-
x,..
+20 mV
0
.
2. K+ S e c r e t i o n by Strial Marginal Cell Epithelium
Active K + secretion in the cochlear duct has been demonstrated by flux measurements of radiolabeled K + intro-
Vhold +40 mV
1$
200 pS
Smaller
Block by TEA
Yes
No
Block by CTX
Yes
Not tested
Block by 4-AP
No
Yes
Apamin
Not tested
No
TEA, tetraethylammonium; CTX, charybdotoxin; 4-AP, 4-amino pyridine.
/m(pA) 0
t
A 1-
A__m. w
g
~-50
0
-s'o
-2's Voltage
o
2;
s;
(mV)
FIGURE 16. Basolateralcalcium currents (whole-cell recordings): (A) the avian equivalent of inner hair cells, representative traces (a) and current-voltage relationship (b) and (B) outer hair cells. [Reprinted from (A) Prog. Neurobiol. 39, Fuchs, E A. Ionic currents in cochlear hair cells, pp. 493-505. Copyright 1992 with kind permission from Elsevier Science Ltd, The Boulevard, Langford Lane, Kidlington OX51GB, UK. and (B) from Neurosci. Lett. 125, Nakagawa, T., Kakehata, S., Akaike, N., Komune, S., Takasaka, T., and Uemura, T., Calcium channel in isolated outer hair cells of guinea pig cochlea, pp. 81-84. Copyright 1991 with kind permission of Elsevier Science-NL, Sara Burgerhartstraat 25, 1055 KV Amsterdam, The Netherlands.]
discharge rates to the acoustic stimulus (Kros, 1996). By contrast, the Ca 2+ currents in outer hair cells are activated only above -30 mV (see Fig. 16B), a range that is far removed from the resting membrane potential n e a r - 8 0 mV found in vivo and never reached by receptor potentials, which apparently saturate at 15 mV (Kros, 1996). (The curve would be expected to shift slightly to the left at physiological perilymphatic Ca 2+ levels, but the channel would still not likely show significant activation at voltages observed in OHC.) There is therefore no evidence at this time for a function of these voltage-gated Ca 2+ channels in OHCs; there are also no measurements of neural activity in the Type II afferent fibers in the auditory nerve that synapse exclusively to the OHCs. The K + channels in hair cells (Table 3) vary considerably among type and species. The cochlear hair cells of many nonmammalian vertebrates and the vestibular hair
cells of mammals have Ca 2+- activated K + channels that act in synergy with the voltage-gated Ca 2+ channels to electrically resonate at the characteristic frequency of the cell. This forms part of the basis of frequency discrimination in these species. Electrical tuning is not, however, significantly present in mammalian cochlear hair cells. Sharp tuning of mammalian IHCs in vivo is thought to be the result of active mechanical feedback from other cochlear elements such as OHCs (see later section on reverse transduction). In inner hair cells, the K + currents are activated by depolarization in the range o f - 6 0 t o - 2 0 mV, a range that correlates well with that of the resting and receptor potentials found in recordings made in vivo and with that of the voltage-gated Ca 2+ channels described earlier. The resting potential of IHCs in vivo is about --40 mV, which reflects a compromise of the highly negative EMF from the basolateral K + channels by the depolarized EMF of the nonselective cation conductance in the apical membrane (Dallos, 1986). The total cell K + current was found to be composed of both a fast and slow component (Fig. 17) The fast component is carried by channels resembling maxi K (or BK) channels that are characterized by rapid activation by membrane depolarization, block by TEA and charybdotoxin, insensitivity to 4-amino pyridine (4-AP), and single-channel conductance with symmetrical high K + concentration of more than 200 pS (Gitter et al., 1992; Fuchs, 1992; Kros, 1996). The slow component is a delayed rectifier K + channel exhibiting slow activation by depolarization, no inactivation, a smaller single-channel conductance than the fast component, block by 4-AP but not TEA. Surprisingly, the fast component of the current is apparently independent of extracellular Ca 2+ (Kros and Crawford, 1990), although the equivalent current in other types of hair cell is highly sensitive to Ca 2+ (Fuchs, 1992). The profile of K + currents in outer hair cells is different from that of IHCs (Kros, 1996; Fuchs, 1992). The primary K + conductance is due to Ii~ n channels, which are deactivated below -90 mV and fully activated a t - 5 0 mV, have a single-channel conductance of about 45 pS, are activated by Ca 2+ (Po = 0.5 at
786
SECTION V Synaptic Transmission and Sensory Transduction
r
_
. . . . . . . .
21
A
1.2
--8
0.8
PDE~I3+ PDE~.T~.GTP '- GMP
ATP ADP
\
Ta'G
PgEi
TI3~
FIGURE 1 1. The enzyme cascade controlling visual transduction in rods. Abbreviations: A, arrestin; channel, cGMP-gated outer-segment channel; GC, guanylate cyclase; PDE, cGMP-specific phosphodiesterase (*, active form): Rh, rhodopsin (*, photoexcited, active form); T, the G protein, transducin. There is increasing evidence that in spite of its complexity, this diagram is quite incomplete. (Modified with permission from Stryer, 1991, and drawing on information from Hargrave and McDowell. 1992, and Hofmann et al., 1992.)
The recovery of the dark current after photon absorption is accomplished by many processes, including these: 1. Shutoff of rhodopsin, as discussed earlier. 2. Inherent GTPase activity of Tee, converting Ta back to the GDP-bound, inactive form that reassociates with T/3% 3. Reassociation of PDE~/with PDEa/3, returning PDE to its more inhibited, dark state. 4. Synthesis of new cGMP from GTP by guanylate cyclase (GC). 5. Reopening of cGMP-gated ion channels, as more cGMP becomes available. Although these processes are the most established ones, there is recent evidence that other processes and factors may be involved in regulating the cGMP cascade, mostly through feedback mechanisms. Some of the factors that may play a role are regulatory cGMP-binding sites on PDE; calciumbinding proteins that may affect guanylate cyclase, PDE, rhodopsin, transducin, and/or the cGMP-gated channel; members of the inositol phosphate system whose function has not yet been established; and various kinases and phosphatases whose functions are also not yet clear. The field of regulation of the photoreceptor cGMP cascade is currently very unsettled and very intriguing. Some interesting findings on this subject are reviewed in Lagnado and Baylor (1992), Detwiler and Gray-Keller (1992), and Yarfitz and Hurley (1994). One aspect of visual transduction that has been proposed to be controlled by feedback regulation is light adaptation. The following argument is suggested: coupled with calciumbinding proteins, Ca 2+ has been found to inhibit guanylate cyclase (GC), to stimulate PDE (via inhibition of Rh phosphorylation) and to inhibit the cGMP-gated channels (see Chapter 47). In the dark, when the cGMP, concentration is relatively high, Ca 2§ enters the cell through the cGMP-gated channels, raising intracellular [Ca 2+] ([Ca2+]i). When the
channels close following photon absorption, [Ca2+]i decreases because its influx through the channels is decreased while its extrusion by Na+:Ca2§ § exchangers continues. The decreased [Ca2+]i, would be expected to stimulate guanylate cyclase and to inhibit PDE, leading to a subsequent partial recovery of the cGMP level and a reopening of some of the channels. The feedback mechanisms involving Ca 2+ are mediated by several calcium binding proteins (reviewed in Roof and Makino, 2000). While this scheme sounds very logical and appealing, some results suggest that it is very oversimplified. Detwiler and Gray-Keller (1992) have found that high [Ca2+]i speeds the recovery from a light response, as if facilitating the return to dark levels of [cGMP] and the consequent reopening of cGMP-gated channels. This behavior is inconsistent with the argument just presented, which predicts that high [Ca2§ would d e c r e a s e [cGMP] by stimulating PDE and inhibiting guanylate cyclase. These and other recent findings reveal that many pieces of the transduction puzzle are still missing (also see Chapter 47).
V. Summary The process of vertebrate vision begins with photon capture and visual transduction in the rods and cones. The rods are outstanding single photon detectors that are used in dim light, whereas the cones provide excellent visual acuity and color vision in bright light. There are slight differences in the structures of rods and cones, but these photoreceptors have the same basic design for optimal photon capture and visual transduction. The first step in the response to light is the absorption of a photon by a visual pigment molecule. The visual pigment (or photopigment) consists of a particular protein (an opsin) containing a ubiquitous chromophore, retinal. The differences in
814
SECTION V Synaptic Transmissions and Sensory Transduction
color sensitivity of the rods and cones result from spectral tuning of retinal by their different opsins. The photopigment absorption spectra determine the probability that a photon of a particular wavelength will be absorbed. Visual transduction consists of a kind of signal conversion: the message of photon absorption is converted into a decrease in membrane conductance, which results in a hyperpolarization that decreases the release of neurotransmitter onto other retinal neurons. The light-induced decrease in membrane conductance is accomplished by the closure of nonselective cation channels in a specialized region of the photoreceptor called the outer segment. These channels conduct the so-called "dark current" that flows into the outer segment in the dark. The channels are kept open in the dark by cGMP, and they close in the light when the intracellular cGMP concentration falls. The light-induced fall in [cGMP] occurs when a photoexcited visual pigment molecule activates a G protein, which in turn activates a phosphodiesterase to hydrolyze cGMP. There are elaborate mechanisms for returning the molecular machinery of the photoreceptor to its dark state, and for light and dark adaptation. Although many of the fundamental molecular steps in visual transduction have been determined, numerous mysteries still remain.
Bibliography Bacigalupo, J., Johnson, E., Robinson, P., and Lisman, J. E. (1990). Second messengers in invertebrate phototransduction. In "Transduction in Biological Systems" (C. Hidalgo, J. Bacigalupo, E. Jaimovich, and J. Vergara, Eds.), pp. 27-45. Plenum, New York. Bader, C. R., Bertrand, D., and Schwartz, E. A. (1982). Voltageactivated and calcium-activated currents studied in solitary rod inner segments from the salamander retina. J. Physiol. 331, 253-284. Bames, S., and Hille, B. (1989). Ionic channels of the inner segment of tiger salamander cone photoreceptors. J. Gen. Physiol. 94, 718-743. Baylor, D. A. (1987). Photoreceptor signals and vision. Invest. Ophthalmol. Visual Sci. 28, 34--49. Baylor, D. A. (1996). How photons start vision. Proc. Natl. Acad. Sci. USA 93,560-565. Baylor, D. A., and Nunn, B. J. (1986). Electrical properties of the lightsensitive conductance of rods of the salamander Ambystoma tigrinum. J. Physiol. 371, 115-145. Baylor, D. A., Lamb, T. D., and Yau, K.-W. (1979). The membrane current of single rod outer segments. J. Physiol. 288, 589-611. Baylor, D. A., Nunn, B. J., and Schnapf, J. L. (1987). Spectral sensitivity of cones of the monkey Macaca fascicularis. J. Physiol. 390, 145-160. Birge, R. R. (1990). Nature of the primary photochemical events in rhodopsin and bacteriorhodopsin. Biochim. Biophys. Acta 1016, 293-327. Bok, D. (1985). Retinal photoreceptor-pigment epithelium interactions. Invest. Ophthalmol. Visual Sci. 26, 1659-1694. Detwiler, E B., and Gray-Keller, M. E (1992). Some unresolved issues in the physiology and biochemistry of phototransduction. Curr. Opin. Neurobiol. 2, 433-438.
Dowling, J. E. (1987). "The Retina: An Approachable Part of the Brain." pp. 195-197. The Belknap Press of Harvard Univ. Press, Cambridge, MA. Dratz, E. A., and Hargrave, E A. (1983). The structure of rhodopsin and the rod outer segment disk membrane. Trends Biochem. Sci. 8, 128-131. Fain, G. L., Matthews, H. R., and Cornwall, M. C. (1996). Dark adaptation in vertebrate photoreceptors. Trends Neurosci. 19, 502-507. Hagins, W. A. (1972). The visual process: Excitatory mechanisms in the primary receptor cells. Annu. Rev. Biophys. Bioeng. 1, 131-158. Hagins, W. A., Penn, R. D., and Yoshikami, S. (1970). Dark current and photocurrent in retinal rods. Biophys. J. 10, 380-412. Hargrave, E A., and McDowell, J. H. (1992). Rhodopsin and phototransduction. Int. Rev. Cytol. 137B, 49-97. Hofmann, K. E, Pulvermuller, A., Buczylko, J., Van Hooser, E, and Palczewski, K. (1992). The role of arrestin and retinoids in the regeneration pathway of rhodopsin. J. Biol. Chem. 267, 15701-15706. Koutalos, Y., and Yau, K.-W. (1996). Regulation of sensitivity in vertebrate rod photoreceptors by calcium. Trends Neurosci. 19, 73-81. Lagnado, L., and Baylor, D. A. (1992). Signal flow in visual transduction. Neuron 8, 995-1002. Lamb, T. D. (1980). Spontaneous quantal events induced in toad rods by pigment bleaching. Nature 287, 349-351. Lamb, T. D., and Pugh, E. N., Jr. (1990). Physiology of transduction and adaptation in rod and cone photoreceptors. Neurosciences 2, 3-13. Leibrock, C. S., Reuter, T., and Lamb, T. D. (1998). Molecular basis of dark adaptation in rod photoreceptors. Eye 12, 511-520. McNaughton, P. A. (1990). Light response of vertebrate photoreceptors. Physiol. Rev. 70, 847-883. Nathans, J. (1992). Rhodopsin: Structure, function, and genetics. Biochemistry 31, 4923-4931. Pugh, E. N., Nikonov, S., and Lamb, T. D. (1999). Molecular mechanisms of vertebrate photoreceptor light adaptation. Curr. Opin. Neurobiol. 9, 410--418. Roof, D. J., and Heuser, J. E. (1982). Surfaces of rod photoreceptor disk membranes: Integral membrane components. J. Cell Biol. 95, 487-500. Roof, D. J. and Makino, C. L. (2000). The structure and function of retinal photoreceptors. In: The Principles and Practice of Ophthalmology, 2rid edition, volume 3, pp. 1624-1673. (D. M. Albert and E A. Jakobiec, Eds.) W. B. Saunders Company. Philadelphia. Schnapf, J. L., Nunn, B. J., Meister, M., and Baylor, D. A. (1990). Visual transduction in cones of the monkey Macaca fascicularis. J. Physiol. London 427, 681-713. Stryer, L. (1991). Visual excitation and recovery. J. Biol. Chem. 266, 10711-10714. Wald, G. (1968). The molecular basis of visual excitation. Nature 219, 800-807. Yau, K.-W., and Baylor, D. A. (1989). Cyclic GMP-activated conductance of retinal photoreceptor cells, Annu. Rev. Neurosci. 12, 289-327. Yarfitz, S., and Hurley, J. B. (1994). Transduction mechanisms of vertebrate and invertebrate photoreceptors. J. Biol. Chem. 269, 14329-14332. Zuker, C. S. (1996). The biology of vision in Drosophila. Proc. Natl. Acad. Sci. USA 93, 571-576.
Stephen D. Roper
Z[I:) Gustatory and Olfactory Sensory Transduction
!. I n t r o d u c t i o n Many sensory receptor cells are modified epithelial cells and consequently share properties common to epithelial tissues. These properties include (a) an ongoing renewal of the cell population and (b) the existence of two separate regions of the cell surfacewapical membrane and basolateral membrane. These regions are kept separate by intercellular junctional complexes (including tight junctions) that also serve to bind adjacent cells together. Apical and basolateral cell surfaces in epithelial cells and sensory receptor cells are exposed to two vastly different milieux. This is especially the case in gustatory and olfactory receptor cells. Their apical membrane tips confront external chemical environments in the oral and nasal cavities, respectively, that can vary profoundly in composition. In contrast, their basolateral membranes are bathed in a protected, relatively constant medium--the interstitial fluid. The exposed apical membrane of gustatory and olfactory sensory cells is the site of chemosensory transduction, the conversion of external chemical stimuli into intracellular electrical signals. Tight junctions confine most chemical stimuli to the specialized apical membrane and prevent the stimuli from penetrating deeper into the sensory tissue (although there may be exceptions in gustatory sensory cells, see Section II). This has two implications for gustation and olfaction. First, the molecular machinery for the initial events of signal transduction must reside in the apical membrane of the sensory cells, where chemical stimuli contact the cells. Second, because chemosensory stimulation is usually associated with a receptor (or generator) current (that is, a flux of ions) across the apical chemosensitive membrane, the ionic environment at the exposed apical region has a significant impact on transduction. For example, the apical chemosensitive membrane of olfactory sensory cells possesses a Ca2+-dependent C1- conductance, as will be discussed later. This conductance exploits the low (relative to
Cell Physiology Sourcebook: A Molecular Approach, Third Edition
8 | 5
interstitial fluid) C1- concentration in the external environment. When the Ca2+-dependent C1-channels are opened during odor stimulation, C1- ions leave the cell, generating an inward (depolarizing) current. This C1- current augments the sensory signal. Receptor currents generated across the chemosensitive tips of olfactory and gustatory sensory cells flow into the apical membrane and out of the sensory cell across its basolateral membrane. This will be discussed in greater detail later. Olfactory receptor cells transmit their signals directly to the brain via their axons (Fig. 1). Olfactory receptor cells are examples of primary sensory cells. In contrast, gustatory sensory cells transmit their signals synaptically to sensory neurons whose peripheral processes (sensory afferent fibers) innervate taste buds. The central terminals of the sensory neurons communicate with the CNS. The cells within a taste bud are also believed to communicate with each other via chemical synapses and electrical synapses (Roper, 1992). This suggests that some degree of information processing via lateral synaptic interactions, that is, cross talk, occurs in gustatory sensory organs, but the details of this are sketchy.
II. T a s t e R e c e p t o r Cells A. General Comments Peripheral sensory organs for taste consist of clusters of receptor cells, the taste buds, found throughout the oral cavity. There are 2000-10000 taste buds in humans. This number varies considerably from individual to individual. Each taste bud consists of 50-100 cells. Taste buds are embedded in the lingual epithelium in specialized mounds of tissue termed papillae (fungiform, foliate, and circumvailate papillae). Taste buds are also localized on the soft palate, uvula, epiglottis, pharynx, larynx, and have even been reported in the upper esophagus.
Copyright 9 2001 by Academic Press. All rights of reproduction in any form reserved.
[316
SECTION V Synaptic Transmission and Sensory Transduction
F I G U R E 1. Schematic drawing comparing a gustatory sensory cell (A) and an olfactory receptor neuron (B). The apical or chemosensitive portions of the receptor cells extend above the epithelial junctional complex. The initial events in chemosensory transduction occur here. The basolateral membrane of the cells lies below the apical junctional complex and is surrounded by interstitial fluid. Gustatory sensory cells synapse onto primary sensory afferent fibers from sensory ganglion cells. Olfactory receptor neurons are primary receptor cells and send axons directly to the olfactory bulb, as shown. (C) Micrograph of a single olfactory receptor neuron isolated from a frog. (D) Micrograph of an isolated rat taste bud, containing several gustatory sensory cells. The taste bud has been immunostained for gustducin, a taste-specific G protein. This reveals three to four intensely immunostained gustatory sensory cells within this taste bud (I. Wanner and S. Roper, unpublished). Calibrations = 10 lam in C and D. (Micrograph in C modified from Kleene, S. J., and Gesteland, R.C. (1981) Brain Res. 229, 536-540.)
Taste buds respond to a number of basic taste qualities, including sweet, sour, salty, bitter, u m a m i (discussed later), and perhaps others. Early investigators tested whether particular regions of the tongue were selective for these different qualities. For example, there is a somewhat greater sensitivity to bitter in the posterior tongue and sweet at the tip (Fig. 2). However, these original findings have often been misinterpreted or overinterpreted in the intervening years, leading to the popular misconception that there is a discrete regional specialization of the tongue for certain tastes. The truth is that all parts of the tongue sense the basic qualities, as shown in Fig. 2. Most of the cells in a taste bud are narrow elongate cells, extending nearly the full thickness of the lingual epithelium. The basal processes reach to the basement membrane and the apical processes stretch up to a tiny cavity in the lingual surface, the taste pore. The elongate, columnar morphology of gustatory sensory cells sets taste buds apart from the surrounding flat laminar tissue. That is, taste buds form tiny pockets of simple columnar epithelium within the larger sheet of stratified squamous lingual epithelium. Such an arrangement positions
i!
~o..,-.".~~ .... I "
sweet
~!:.:::
bitter
:-2-'
sour
salty
F I G U R E 2. Regional distribution of sweet, bitter, sour, and salty taste on the human tongue, as originally determined by D. P. H~inig (1901). The sensitivity for a particular taste quality is represented by the density of stippling. There is a slight tendency for the anterior tip of the tongue to be more sensitive to sweet, the lateral regions to be more sensitive to sour, and the posterior to bitter. However, there is complete overlap for all these taste qualities. Also, see discussion in Lindemann (1999).
49. Gustatory and Olfactory Sensory Transduction
817
taste cells directly across the electric field generated by transepitheliai ion t r a n s p o r t currents. This may have implications for taste transduction mechanisms (discussed later). The apical tips of the taste cells extend into the taste pore. Microviili on the exposed tips of gustatory sensory cells are the primary sites of chemosensory transduction. The cytoplasm of taste cells varies in appearance from cell to cell. Based on the cytological and ultrastructural features of taste bud cells, some morphologists have categorized taste cells as lOmV d a r k cells and light cells. Other morphologists have used categories such as Type I, II, and III cells. Some ultrastructural studies have reported that all taste bud cells possess synaptic lnA connections and thus are believed to be chemoreceptor cells. 20 ms Other studies report that only one of the cell types (Type III) possesses synaptic connections and is the actual receptor cell. Some taste bud cells are believed to be n e u r o m o d u l a t o r y cells; they release serotonin to modify the chemosensitivity of the taste bud (Ewald and Roper, 1994; Delay et al., 1997; Herness and Chen, 1997). Other cells may be sustentacular, or glial-like, cells and function to regulate potassium in the intercellular spaces of the taste bud (Bigani, 2001). Ongoing investigations are attempting to correlate these proposed functions of the different taste bud cells with the different cellular morphotypes (dark, light, Type I, II, III). Taste buds also contain a population of small ovoid stem cells near the base of the gustatory organ. Taste cells, like ' fl _3 ...... 20mV CaC12 BaC12 the adjacent non-taste epithelium, represent a r e n e w i n g population. Taste bud cells have an estimated life span of ls approximately 10 days in the mammal, slightly longer than BaCI2 I(C1 the surrounding non-sensory epithelium. As stem cells divide and their daughter cells differentiate into mature chemoreceptive cells, they extend a long narrow process up to the taste pore. One interpretation of the morphological classes in the taste bud is that the different cell types represent different stages of maturation of gustatory sensory ~ ~ rnrp~n~"n-m cells, but the data supporting this are not conclusive. I I0mV Chemosensory transduction, the focus of this chapter, is 0.1 nA primarily concerned with events occurring at the apical 2s chemosensitive m e m b r a n e of gustatory sensory cells. FIGURE 3. Intracellular recordings from gustatory sensory cells. However, it is important to be aware that events at the baso(A) Gustatory sensory cells are excitable and generate action potentials lateral membrane can also influence gustatory signals. For when depolarized, as shown for a taste cell from the amphibian example, the basolateral membrane is the site where seroNecturus maculosus. Upper, two superimposed traces showing retonin, an important neuromodulator in taste buds, (a) acts to sponses to brief depolarizing current pulses passed through the intracelmodify the passive (electrotonic) spread of receptor currents lular microelectrode. Lower, current pulse. Reprinted from Roper throughout the taste cell and (b) modulates Ca 2+ currents (1983). Copyright 1991 American Association for the Advancement of that are important in synaptic transmitter release (Ewald and Science. (B) Subthreshold receptor potentials elicited by brief focal apRoper, 1994; Delay et al., 1997; Herness and Chen, 1997). plications of CaC12 and BaC12, two potent taste stimuli, to the taste pore For many years, gustatory sensory cells were considered of a taste bud in Necturus. (C) Same as B but showing action potentials passive transducers of the chemical environment. However, elicited when receptor potentials exceed threshold. Lower traces in B and C show the application of the chemical stimuli. (D) Intracellular remicroelectrode studies then revealed that these cells, like ceptor potentials and membrane conductance changes elicited by focal neurons, are excitable and possess a wide variety of voltageapplication of BaC12 and KC1 to the taste pore in Necturus. Both stimd e p e n d e n t ionic conductances (Fig. 3A; Roper, 1983; uli elicited depolarizing receptor potentials. However, divalent cations Kashiwayanagi et al., 1983). These conductances include such as Ba2§ decrease membrane conductance. This is shown by the involtage-dependent, tetrodotoxin (TTX)-sensitive Na + chancreased amplitude of hyperpolarizing responses produced by a train of nels; delayed-rectifier K + channels; type A K + channels; brief constant current pulses applied through the recording electrode. In inward-rectifier K + channels; CaZ+-activated K + channels, contrast, KC1 depolarized the cell, accompanied by an increased memCaZ+-activated C1-channels; and voltage-dependent Ca 2+ brane conductance. Upper traces, intracellular recordings. Lower traces, channels (reviewed in Lindemann, 1996). Certain of these current pulses applied through the recording electrode. B, C, and D ion channels are intimately involved in the taste transduction modified from Bigiani and Roper, 1991. Copyright 1991 American Association for the Advancement of Science. process, as described later. m
m
i
ill
Jl.,
i i
818
B. Nature of Taste Stimuli Most water-soluble chemicals elicit a taste. Chemical stimuli have traditionally been grouped into the primary qualities sweet, sour, salty, and bitter. The existence of fundamental taste qualities argues there are distinct transduction mechanisms and signaling pathways for each of the taste qualities, and this notion has guided taste research. There is growing evidence that there are more than just these four categories, such as the unique taste of monosodium glutamate, termed umami (described later). Chemicals that elicit taste usually do so only at high concentrations (a few millimolar to hundreds of millimolar). Certain bitter-tasting substances (such as quinine, caffeine, and strychnine) and artificial sweeteners are relatively potent stimuli and elicit responses at concentrations in the pM to low-mM range. Nonetheless, it is clear that taste cells have low stimulus sensitivity relative to the exquisite responsiveness of photoreceptors (i.e., single photons) or olfactory and pheromone chemoreceptors (i.e., picomolar stimulus concentrations, possibly single molecules). Primary functions of taste organs include regulating the intake of important nutrients and protecting against the ingestion of harmful substances. Therefore, the generally low sensitivity of gustatory sensory cells to many chemical stimuli (with the exception of bitter taste) may represent optimization of responses to physiologically relevant concentrations of nutrients. Sensitivity to bitterness may be related to the fact that many bitter-tasting chemicals are toxic and, conversely, many toxic chemicals elicit bitter taste--an aversive taste quality. Thus, it is understandable that bitter taste transduction has evolved to detect low concentrations of stimuli and to prevent the consumption of even small quantities of potentially harmful substances. Taste stimuli are dissolved in saliva and reach only the chemosensitive microvillar (apical) surface of gustatory sensory cells. Most chemical stimuli are prevented from gaining access to the basolateral regions of the taste bud by the tight junctions of the junctional complex that seals off the apical membrane in the taste pore (Fig. 1). Tight junctions, however, are leaky to certain stimuli, such as Na +, K +, and C1- (Ye et al., 1993). These ions can pass from the lingual surface into the intercellular spaces within the taste bud to raise the concentration there above levels normally found in interstitial fluid. It is possible that chemotransduction mechanisms for some taste stimuli, especially Na +, K +, CI-, and possibly other stimuli, exist on the basolateral membranes of taste receptor cells as well as on their exposed apical tips. Basolateral transduction mechanisms may explain intravenous taste, where certain chemicals that are dissolved in the blood elicit gustatory sensations. The bitter taste of some drugs injected intravenously may be an example of this phenomenon.
C. Initiation of Taste Receptor Potentials The arrival of a taste stimulus at the apical chemosensitive membrane is signaled by the generation of a depolarizing receptor potential (Fig. 3B). The conversion of a gustatory chemical signal into an intracellular receptor potential is
SECTION V SynapticTransmission and SensoryTransduction
termed taste transduction. Unlike signal transduction in most other sensory organs, there is no single unifying hypothesis that explains how gustatory sensory cells respond to taste stimuli. The necessity of taste organs to sense a wide range of chemical substances, from simple ions to complex molecules, has led to the evolution of multiple taste transduction mechanisms (see Corey and Roper, 1992; DeSimone, 1991; Finger et al., 2000; Kinnamon and Margolskee, 1996; Lindemann, 1996; Simon and Roper, 1993). Presumably, different gustatory sensory cells possess different transduction mechanisms and respond to different taste stimuli. Recent findings from Ca 2+ imaging studies support this view. That is, an individual taste receptor cell may respond to a somewhat limited repertoire of chemical substances (Bernhardt et al., 1996; Caicedo and Roper, 2001). However, the f'mal answer has yet to be established. Even if an individual gustatory sensory cell is "tuned" to a subset of chemical stimuli, a taste bud is a population of gustatory sensory cells. Thus, a single taste bud with its 50-100 cells can respond to a diverse array of chemical stimuli. The multiple signaling pathways utilized in taste transduction can be divided into three basic categories:
1. Passive ion permeation through ion channels. Some taste stimuli enter taste cells through ion channels in the apical membrane, thereby generating transmembrane receptor currents and receptor potentials. This is often the case when the taste stimulus is itself an ion (for example, Na + or K + salts). 2. Block of ion channels. As explained later, certain ions (e.g., Ca 2+ and H +) block apical membrane channels, particularly those for K +. This also generates a receptor potential because when K + channels are blocked, leak conductances for other ions such as Na + (which tend to depolarize the cell) can then exert a greater influence on the resting potential. 3. Receptor-ligand binding. Interactions occur between chemical stimuli (ligands) and specific membranebound receptors (i.e., ligand-receptor binding) and ultimately lead to receptor currents or the release of intracellular Ca 2+. Parenthetically, amphipathic stimuli such as saccharin and quinine may insert into the plasma membrane and activate taste cells independently of these three basic mechanisms. Taste transduction mechanisms involving ion permeation, channel blockage, and receptor-ligand interactions are described next. 1. Ion Permeation, Channel Blockage, and Current Flow through Apical lon Channels
a. Na + Channels. Sodium salts ("salty taste") are believed to be transduced in a subset of taste cells by the direct permeation of Na + through passive Na+-selective channels (as opposed to the TI'X-sensitive, voltage dependent Na § channels that underlie action potentials) (Heck et al., 1984). These ion channels resemble amiloride-sensitive epithelial Na + channels (ENaC) found in renal tubules and elsewhere. Indeed, mRNA isolated from lingual epithelium shows that all three ENaC subunits (a, /3, and 7) are expressed in taste cells
49. Gustatory and Olfactory Sensory Transduction (Kretz et al., 1999; Lin et al., 1999). Additionally, immunohistochemistry has confirmed the presence of ENaC subunits in taste cells. Yet, these channels are also expressed in lingual tissues other than taste buds, leading to speculation that nonsensory epithelial cells in the tongue somehow participate in salt taste (Li et al., 1994). The proportion of a,/3, and y ENaC subunits expressed in gustatory sensory cells appears to vary between the anterior and posterior tongue, which may explain known regional differences of the tongue in salt taste. Where it has been measured (in frog taste cells), these taste Na + channels have small unitary conductance (1-2 pS) and are blocked by low concentrations of amiloride (Kin= 0.2-1 laM). As stated previously, the apical amiloride-sensitive Na + channels in gustatory sensory cells are not voltage-dependent. ENaC channels are open at rest and allow an ongoing flux of Na +, driven by this cation's electrochemical gradient. The concentration of Na + in human saliva (the solution bathing the apical tips of gustatory sensory cells) varies from about 3 to 63 mM, depending on the rate of salivary secretion. This resuits in an equilibrium potential for Na + (ENa) of about-30 mV to +50 mV (assuming [Na+]i is 10 raM) across the apical chemosensitive membrane. The resting potential of gustatory sensory cells is estimated to be -60 t o - 8 0 mV. Thus, there should be a small steady-state influx of Na + through the apical membrane at rest in those taste receptor cells that possess apical ENaC channels. During stimulation with NaC1 when [Na § in the oral cavity exceeds salivary [Na§ the electrochemical gradient driving Na + into the cells increases. This generates an inward receptor current through the apical membrane, over and above the resting Na + influx, and further depolarizes the cell. This current exits across the entire basolateral cell membrane, thus forming a return current loop by way of an extracellular (paracellular) pathway that includes the intercellular fight junctions near the taste pore. Consequently, any factors that affect the basolateral membrane conductance or the paracellular resistance, especially at the fight junctions near the taste pore, will also affect the magnitude of the receptor currents. The density of amiloride-sensitive Na + channels, which is under hormonal control, determines the magnitude of Na + influx at rest and during salt stimulation. Specifically, arginine-
vasopressin (antidiuretic hormone, ADH) and aldosterone increase the number of ENaC channels, albeit via different mechanisms. Na + conductance in taste cells may be tonically depressed by salivary Na + ("self-inhibition," Gilbertson and Zhang, 1998), but this inhibition is overridden during chemostimulation with Na + salts. Accumulation of Na + within the taste cell is prevented by the action of the Na+,K+-ATPase distributed on the basolateral membrane. Na+,K+-ATPase pumps excess Na + from the cell into the surrounding interstitial spaces. Vascular elements below taste buds presumably drain off any surplus Na +, preventing a buildup in the intercellular spaces within the taste bud. The depolarizing current produced by Na + influx at the apical tip may be offset in part by the electrogenic basolateral Na+-K + pumps (which hyperpolarize the membrane) and partly by the compensatory effiux of K + through basolateral K + channels (i.e., an outward, repolarizing K + current). Although one might anticipate that ENaC channels would be localized to the apical chemosensitive tips of the gustatory sensory cells, in fact, immunostaining for these channels reveals
819 their presence on apical and basolateral membranes. The function of basolateral ENaC channels is not yet understood. A substantial resting Na § conductance through basolateral ENaC channels would chronically depolarize gustatory sensory cells due to the steady-state influx of Na § from the interstitial spaces within the taste bud. This conundrum remains to be solved. In addition to their role in salt taste, the amiloride-sensitive Na § channels on gustatory sensory cells may participate in other taste qualities, particularly "sour" (discussed later). Amiloride reduces acid ("sour") taste in behavioral studies on animals and psychophysical studies on humans. Indeed, the role of ENaC channels for "saltiness" per se in human taste has been questioned. Amiloride appears to affect only the sourness, not saltiness, of NaC1 solutions in human studies (Ossebaard and Smith, 1996). Na § salts have an important secondary effect on taste buds apart from permeating through Na§ ion channels in the apical tips of taste cells. The presence of Na § in saliva resuits in a steady transepithelial transport of Na + across the non-sensory lingual epithelium, from mucosal (apical, saliva) to serosal (basolateral, interstitial fluid) spaces (Fig. 4). This Na § transport generates a standing voltage across the entire lingual epithelium, estimated to be about 10 mV or higher (serosal side positive). Consequently, transepithelial transport imposes an electric field across taste buds which, as described previously, represent islands of simple columnar cells embedded in a stratified squamous epithelium. In the presence of Na § salts, transepithelia! Na § transport is increased and the transepithelial voltage gradient increases, thereby increasing the electric field across the taste buds. The transepithelial voltage drives a return current through the paracellular and transcellular pathways back to the mucosal surface (Fig. 4). The net effect of the transcellular current is to hyperpolarize the basolateral membrane of taste cells, where synapses are located, and reduce the potential difference across their apical membranes, where chemosensory transduction takes place. Hyperpolarizing the membrane near synapses might be expected to reduce transmitter release. These secondary effects oppose the direct stimulation of taste receptor cells by Na § salts. Furthermore, when the electrical resistance of the paracellular current pathways is increased, for example by the presence of organic anions (e.g., stimulation with sodium acetate), proportionately more of the transepithelial transport return current is shunted through the taste cells (transcellular pathway, Fig. 4). This even further hyperpolarizes the basolateral membrane and decreases the apical membrane potential. Thus, current loops generated by widespread transepithelial Na + transport in the tongue may modulate taste cell activity, independent of any direct stimulation by Na § of taste receptor cells p e r se. This modulation may have implications for salty taste and for taste mixtures when combinations of NaC1 and other taste stimuli are present, as normally occurs during food intake. These considerations emphasize how the overall function of taste buds may be highly dependent on their histological context, that is, being situated in an intact ion-transporting epithelium. b. K § and Other Cation Channels. Duringchemostimulation
with K + salts ("bitter-salty taste"), K + presumably permeates through cation channels located in the gustatory sensory cell
820
FIGURE 4. Na+ transport across the non-sensory lingual epithelium produces current loops through taste buds. (A) Magnification of apical region of taste bud. (B) In lingual epithelial cells, Na+ influx through passive apical channels is actively pumped out of the basolateral membranes by electrogenic Na+,K+-ATPase, generating a flux of Na+ and transepithelial currents (open arrows). This results in a transepithelial potential of about 10 mV (serosal positive). This transepithelial voltage drives current out around taste cells (paracellular pathway) and through taste cells (transcellular pathway) (A). The current through taste cells (transcellular pathway) hyperpolarizes the basolateral membrane where synapses are located and depolarizes the apical membrane where transduction occurs. The relative contribution of current through these two pathways is determined in large part by the resistance of the junctional complex in the region of the taste pore (striped region in A) and the resistance of the taste cell membranes. Certain taste stimuli, such as organic anions, alter the resistance of the junctional complex (paracellular resistance) and thereby change the balance of current through taste buds, possibly altering synaptic transmission and/or apical transduction.
membrane. For example, when mucosal [K +] is elevated, the passive influx of K + through apical K + channels generates a depolarizing receptor potential (Fig. 3C). Apical K + channels have been demonstrated in taste cells from amphibia. However, in taste cells from mammalian taste buds, K + channels are present mainly on the basolateral membrane (Fume and Yoshii, 1997), and it is not known whether there is an api-
SECTION V Synaptic Transmission and Sensory Transduction cal K + conductance. Most likely, in mammalian taste buds, when stimulated with potassium salts, K + ions follow a paracellular pathway, penetrating the apical tight junction barrier and depolarizing gustatory sensory cells by permeating K + channels in the basolateral membrane. I n w a r d l y - r e c t i f y i n g K + channels that open when the cell is hyperpolarized much below El( are also present in taste cells, and together with the delayed rectifier and other K + channels, may serve to stabilize the resting potential near El(. Taste cells have a high input resistance (20-50 Gf~) near their resting potential. Thus, the r e c e p t o r cell at rest is in a state of m a x i m a l sensitivity. The inward rectifier will buffer a taste cell against influences that might tend to hyperpolarize the membrane excessively and depress excitability. Such influences could include electrogenic Na + pumps in taste cells or transcellular currents generated by transepithelial transport (discussed earlier). How acids ("sour taste") are transduced may differ from species to species. In the amphibian, Necturus, it has been shown that protons block apical K + channels. This blockage results in a depolarization. That is, in the absence of stimulation with K + salts (i.e., without elevated [K+]o), the equilibrium potential for K + (El() is near the resting potential and there is likely to be a small but finite ongoing effiux of K +. Indeed, as detailed in Chapter 14, the resting K + conductance in large part determines a cell's resting potential. When this resting K + conductance is blocked by protons, K + effiux is curtailed and conductances for other ions (Na +, Ca 2+, C1-) with equilibrium potentials positive with respect to resting potential determine the membrane potential. Thus, the net effect of blocking K + channels is a depolarization. Additionally, protons may activate acid-sensitive cation channels and generate an inward depolarizing receptor current. An acid-gated cation channel, MDEG1, has been reported in rat gustatory sensory cells (Ugawa et al., 1998). Lastly, in hamsters it has been shown that protons themselves permeate the apical membrane, and this influx constitutes a depolarizing receptor current (Gilbertson et al., 1993). Interestingly, protons permeate hamster taste cells via amiloride-sensitive Na + channels (ENaC) (Fig. 5). In mixtures of Na + salts and acids, protons bind to the ion selectivity site in the ENaC channel more strongly than do Na + ions and as a result, H + conductance is less than that of Na +. This also means that protons interfere with the influx of Na +. Psychophysical studies in humans also implicate a role for amiloride-sensitive Na + channels in sour taste. (The interference between Na + and H + permeation of amiloride sensitive Na + channels may help explain the culinary wisdom that acidifying food, for example by adding vinegar or lemon juice, reduces its salty taste). The fact that there is a common ion channel pathway for H + and Na + ions (ENaC), at least in some species, raises the question of how sour is distinguished from salty. The answer is probably that protons affect taste cells multiple ways. For example, as already mentioned, H + can block K + channels and H + can stimulate acid-gated channels. Protons may have additional less-well-characterized actions on the apical chemosensory membrane that contribute to sour taste. It remains to be determined whether different subsets of taste cells are activated by acids versus by Na + salts and whether
49. Gustatory and Olfactory Sensory Transduction
amiloride:'30 uM ....
1.... .
.
.
.
.
i1"
1 1
'
821
'
-1
I
.....
!
utt
!_f
tl
t i
--
7.4 4.5 50 pA
(c) (a) (a) ~ - = -
"---',,a
(e)
s..
50 pA
F I G U R E 5. Acid taste stimuli are transduced, in part, by proton permeation of amiloride-sensitive Na + channels. This figure shows a patch-clamp record from a taste cell in an isolated fungiform taste bud from the hamster. The cell was held a t - 8 0 mV throughout the record. Acid stimulation (2.5 mM citric acid, pH 4.5) elicits an inward current (a) due to influx of H +. This response is blocked by perfusing the bath with amiloride (30 pM) (b). Washing out amiloride from the bath restores the acid response (c). Top two traces indicate the periods when amiloride and citric acid were applied. Bottom traces are expanded records to show the increases in membrane conductance during citric acid stimulation more clearly. Membrane conductance was monitored by applying a train of small, constant, hyperpolarizing pulses throughout the record. From Gilbertson et al. (1993). Copyright 1993 by Cell Press.
any of the sour transduction mechanisms are common across species. Other salts, such as those of the divalent cations Ca 2+ and Ba 2+, are also transduced by reducing ion permeation through the apical membrane. To humans, these salts elicit a bitter taste. For instance, Ca 2+ and Ba 2+ block apical K + channels in amphibia (Fig. 3C), and this may constitute one mechanism for bitter taste transduction. Gustatory sensory cells appear to detect dietary fats in the oral cavity in part by the ability of fatty acids such as linoleic acid and arachidonic acid to block K + channels (Gilbertson et al., 1997). This action inhibits outward current, especially through delayed rectifier K + channels, and thereby enhances depolarizations elicited by other taste stimuli and prolongs the duration of action potentials. Apparently, transduction mechanisms for salty, sour, certain bitter substances, and fatty acids overlap and affect many of the same ion channels. If all taste cells were to possess identical ion channels, it would be a mystery as to how taste perceptions of these substances can be differentiated. One likely answer is that different taste receptor cells may express different combinations of ion channels. Thus, different subsets of receptor cells are activated during stimulation with salts, acids, and bitter stimuli. This is a topic of ongoing investigation. 2. M e m b r a n e - B o u n d Receptor Events
a. Receptor-Ligand Binding. Certain taste stimuli are believed to interact with specific membrane-bound receptors rather than permeate or block ion channels. Sweet-tasting substances, some bitter-tasting compounds, and amino acids are examples of such stimuli. Particular amino acids, espe-
cially L-glutamate, evoke a unique taste termed u m a m i that is fundamentally different from the other taste primaries sweet, sour, salty, and bitter. Consequently, the taste of Lglutamate (umami) is believed to have its own receptor mechanism (discussed later). Umami is becoming recognized as another primary taste quality, signaling the concentration of free glutamate found in many protein-rich foods such as meats, cheeses, and certain vegetables. The most straightforward mechanism for receptor-mediated signal transduction in gustatory sensory cells is ligand-gating of ion channels (ionotropic channels). This is similar to the action of acetylcholine at the neuromuscular junction. In catfish taste cells, arginine and proline are believed to act this way, stimulating nonselective cation channels. More complex receptor mechanisms involve G-proteincoupled receptors (GPCRs). GPCRs are members of a large family of proteins with seven transmembrane domains. The binding of a taste stimulus to a GPCR triggers intracellular second messengers and a cascade of enzymatic reactions inside the cell. These cascades amplify and/or modulate the initial signal and ultimately converge upon an effector mechanism, usually specific ion channels in the plasma membrane that generate a receptor potential. For instance, the taste of sucrose is believed to be transduced by GPCRs, as follows (reviewed in Lindemann, 1996). Investigators postulate that there is a GPCR for sucrose that activates Gs proteins that, in turn, stimulate adenylyl cyclase (AC) and in-
crease intracellular cyclic adenosine monophosphate (cAMP) concentration. The most plausible interpretation of studies to date is that the increase in intracellular cAMP concentration stimulates a protein kinase that phosphorylates and blocks basolateral K § channels, thereby indirectly depolarizing the taste cell. Stimulating taste tissue with
822 sucrose has been shown to enhance AC activity and increase cAMP concentration in taste buds. Injecting cAMP intracellularly into a subset of taste cells produces a depolarization that is associated with a decreased membrane conductance. This depolarization depends on the presence of ATP and is blocked by a protein kinase A (PKA) inhibitor delivered through the patch pipette. In isolated patches derived from the basolateral membrane of taste cells, PKA has been shown to block certain K + channels. Collectively, these findings indicate a primary role for AC and cAMP in sucrose taste transduction with K § channels being the final effectors. Compounds other than sugars also elicit sweetness. These include such diverse compounds as lead salts, certain amino acids, proteins, and artificial sweeteners (e.g., saccharin). Cross-adaptation studies and cellular analyses of taste cell responses reveal that there are multiple transduction mechanisms, and possibly multiple receptors, for sweet taste. For example, in rat taste cells, saccharin has been shown to stimulate the synthesis of intraceilular inositol t r i s p h o s p h a t e (IP3) and thus increase [Ca2+]i by release from intracellular stores (Bernhardt et al.,1996). Yet, in the same cells, sucrose increases intracellular [Ca 2+] by influx from the extracellular medium. This influx of Ca 2§ is consistent with activation of depolarization-gated Ca 2+ channels following an increase in cAMP, as described previously. That is, in some gustatory sensory cells there appears to be a dual transduction pathway for sweet, involving the second messengers cAMP (for sucrose) and IP 3 (for saccharin). Based on the involvement of the second messengers cAMP and IP 3, it is reasonable to infer that sucrose and saccharin are transduced via GPCRs. However, to date, no receptor has been cloned or identified that has been convincingly linked with the taste of sugars or nonsugar sweeteners. 1 Other taste stimuli, for example certain bitter compounds, are also thought to activate membrane receptors that are coupled to G proteins. For instance, bitter compounds such as denatonium, strychnine, and caffeine transiently generate IP 3 in taste cells and stimulate Ca 2+ release from internal stores, presumably via specialized GPCRs. These findings are consistent with the existence of bitter receptors that activate phospholipase C, converting membrane phos-
phatidylinositol 4,5-bisphosphate, releasing diacylglycerol (DAG) and IP 3 into the cytosol. An additional receptor mechanism for bitter taste has been proposed that involves a reduction of intracellular cAMP, possibly via a transducin-coupled receptor mechanism similar to that in pho-
1Two putative GPCRs of unknown function, TR1 and TR2, have been cloned from rat taste tissues (Hoon et al., 1999). Based on the distribution of TR 1 and TR2 expression in the tongue, it was suggested that TR1 underlies sweet taste and TR2 bitter. However, ligands for these putative GPCRs have not been identifiedto date and their role in taste remains to be established. Other receptors for bitter taste (T2Rs) have recently been discovered (Alder et al., 2000; Matsunami et al., 2000). Additionally, there is suggestive evidence that saccharin (as well as quinine and other amphipathic taste stimuli) may penetrate the taste cell membrane and directly activate intracellular G proteins, bypassing any specific receptors (Naim et al.,1994) but nevertheless triggering intracellular signaling cascades.
SECTION V Synaptic Transmission and Sensory Transduction
toreception (Kolesnikov and Margolskee, 1995; Ruiz-Avila et a1.,1995). Finally, applying the bitter tasting compounds caffeine and theophylline to taste membrane fragments has been shown to elevate cGMP transiently, suggesting yet another signaling pathway for bitterness. Recently, a large family of bitter receptors had been identified in humans and rodents (Alder et al., 2000; Matsunami et al., 2000). These receptors (T2Rs) have properties that are consistent with several of the bitter taste transduction mechanisms (Chandrashekar et al., 2000). The taste quality umami is now believed to be transduced by a metabotropic glutamate receptor (mGluR). A novel, tastespecific variant of mGluR4 (taste-mGlvR4) has been localized selectively to taste buds in lingual epithelium (Chaudhari et al., 2000). In the brain, mGluR4 is a receptor for glutamatergic synaptic transmission between neurons. However, in taste cells, a novel variant of this GPCR appears to have been put to another use--to detect dietary glutamate, a naturally-occurring component of many foods. Umami taste transduction pathways activated by taste-mGluR4 remain to be elucidated. b. G Proteins. As implied in the preceding section, certain receptors in taste are presumed to be coupled to G proteins. A G a protein that is unique for taste, a-gustducin, has been identified in taste buds (McLaughlin et al., 1992) and is a candidate for the Gce protein that couples to bitter and sweet receptors. Gene-knockout mice lacking a-gustducin show deficits in their ability to sense bitter and sweet tastes. Additionally, a G a protein found in photoreceptors, transducin, has also been shown to be expressed in taste cells (McLaughlin et al., 1993; Yang et a1.,1999). This suggests that there may be molecular mechanisms in common between photoreception and chemoreception. Lastly, certain G protein 13 and y subunits have been found in taste cells, specifically ill,/33, and ),13 (Huang et al., 1999). These/33/ subunits appear to activate a particular phospholipase C (PLC/32) also found in taste cells (Rossler et al., 1998). This leads to the generation of IP 3 and DAG, as described for bitter taste. Thus, specific a,/3, and y G protein subunits are present in taste cells. Activating taste GPCRs can initiate transduction paths associated with G proteins (i.e., changing the activity of AC and phosphodiesterase) as well as pathways associated with Gt~r (i.e., stimulating phospholipase C). The several transduction mechanisms postulated for gustatory sensory cells are summarized in Fig. 6.
D.
Termination of Taste Signals
Termination of chemostimulation in taste cells is probably due to two factors: diffusion of the stimulus away from the apical chemosensitive tips and receptor cell adaptation. Very little is known about adaptation in gustatory sensory cells. However, it is plausible that the Ca2+-dependent C|conductance that exists in taste bud cells contributes to sensory adaptation in taste (Taylor and Roper, 1994). An influx of C1- across the basolateral membrane would repolarize the gustatory sensory cell even in the presence of a maintained chemical stimulus due to the increased [Ca2+]i consequent to taste stimulation, as follows. Excitatory receptor currents result in an increase in [Ca2+]i either (a) by way of
49. Gustatory and Olfactory Sensory Transduction
FIGURE 6. Summary of chemosensory transduction in taste. Events at the apical membrane of taste receptor cells, top, include (shown here from left to right) (1) block of K + channels, for instance by H + (sour) and Ca 2+ (bitter); (2) cations (Na + and H +) passing through apical amiloride-sensitive Na + channels (salty and sour, respectively); (3) ligand-gating of cation channels, such as protongated MDEG1 channels (sour) or L-arginine-gated channels (found in catfish gustatory sensory cells); (4) activation of G-proteincoupled receptors that stimulate adenylyl cyclase, for example, by sucrose (sweet); (5) activation of G-protein-coupled receptors that are negatively coupled to cAMP, for example, by monosodium glutamate (umami); (6) activation of G-protein-coupled receptors that stimulate phospholipase C and increase inositol trisphosphate, for example, by saccharin (sweet) or denatonium (bitter). Different cells are believed to possess different transduction mechanisms. Events illustrated on the basolateral membranes of taste receptor cells, bottom, include (1) depolarization of the membrane from a reduction of K + conductance due to cAMP-stimulated phosphorylation; (2) depolarization of the membrane consequent to an influx of cations at the apical membrane; (3) generation of an action potential when depolarizations reach threshold; (4) influx of Ca 2+ ions through voltage-gated Ca 2+ channels, or (5) intracellular release of Ca 2+ from IP3-gated stores; (6) release of neurotransmitter consequent to increased [Ca2+]i. Abbreviations: AC, adenylyl cyclase; DAG, diacyl glycerol; G, G protein; GPCR, G-protein-coupled receptor; IP 3, inositol 1,4,5-trisphosphate; PDE, phosphodiesterase; PIP 2, inositol 4,5-bisphosphate; PICA, phosphokinase A; PLC, phospholipase C.
823
824
SECTION V Synaptic Transmission and Sensory Transduction
Ca 2+ permeation during stimulation, (b) by influx through voltage-dependent Ca 2+ channels opened by depolarizing receptor and action potentials, or (c) by release from intracellular Ca 2+ stores. Increased [Ca2+]i activates the Ca 2+dependent C1- conductance and thereby triggers an influx of Cl- from the interstitial fluid surrounding taste cells, that is, an outward, repolarizing current. This would serve to reduce the effectiveness of a maintained excitatory stimulus.
!!!. Olfactory R e c e p t o r Cells A. General Comments Receptor cells for odorants are distributed in specialized patches of sensory epithelium embedded within the nasal respiratory epithelium. These patches comprise 1-2 cm 2 of surface in each nostril in the human, or approximately 5% of the total nasal epithelium. Olfactory receptor cells are specialized neurons and are often termed olfactory receptor neurons. Olfactory receptor neurons extend the full thickness of the epithelium. Receptor neurons are interspersed among elongate supporting cells in the olfactory epithelium. Olfactory receptor neurons possess apical membrane specializations for receiving chemical signals. However, unlike gustatory sensory cells, the apical tip of an olfactory receptor neuron consists of a tiny knob from which extend 6 to 12 long cilia (see Fig. 1). The opposite (basal) pole of olfactory receptor neurons transforms into an axon that extends to the olfactory bulb in the CNS. Consequently, olfactory receptor neurons bypass any synaptic intervention or modulation in the periphery and transmit action potentials directly to the brain. Parenthetically, this raises the intriguing possibility that certain agents, such as drugs or toxins, can be taken up by olfactory receptor neurons and transported directly into the brain along the axons of the olfactory neurons. That is, the olfactory epithelium represents a "window" into the brain that might be exploited for pharmaceutical therapies or may be at the root of certain central nervous system disorders that are linked to environmental contaminants. A population of stem cells resides at the base of the olfactory epithelium, similar to the stratum germinativum of epithelial tissues. Stem cells are capable of dividing and differentiating into olfactory receptor neurons. The lifespan of olfactory receptor neurons was originally thought to be approximately 30 days (Graziadei and Monti Graziadei, 1978), but more recent findings indicate that receptor neurons may live for many months (Mackay-Sim and Kittel, 1991a,b). Mitotic activity of stem cells provides a reserve population of immature olfactory receptor neurons that can rapidly differentiate and replace mature olfactory receptor neurons if necessary, but that otherwise die. The vomeronasai organ (VNO), a sensory structure situated near the olfactory epithelium in many animals, detects a special category of volatile and nonvolatile chemical stimuli. These stimuli include pheromones. Pheromones are chemical signals discharged by one member of a species to communicate with other members of that same species.
Pheromones play a key role in endocrine responses and in shaping social and sexual interactions (e.g., territoriality, mating) among members of a species, perhaps even including humans. Pheromone stimuli for VNO sensory neurons include airborne molecules (such as dihydro-exo-brevicomin, a volatile constituent of male mouse urine) and nonvolatile molecules (such as aphrodisin, a component of vaginal secretions in hamsters). 2 In humans, the VNO consists of shallow bilateral pits located in the nasal cavity, anterior to the olfactory epithelium. There is considerable controversy whether the VNO is a functional structure or a vestigial organ in humans. In rodents, the VNO consists of bilateral cigar-shaped structures encased in a bony capsule and lying beneath the rostral end of the nasal cavity. Receptor neurons in the VNO, as in the olfactory epithelium, possess axons. However, unlike most olfactory receptor neurons, sensory neurons of the VNO possess apical microvilli, not cilia. The initial events of transduction are believed to occur on the microvilli. VNO sensory neurons send their axons to the accessory olfactory bulb, a separate portion of the olfactory bulb that is reserved for processing pheromone stimuli. Sensory mechanisms for transducing pheromones and odorants are believed to be quite dissimilar, as will be discussed next. This chapter will focus mainly on olfactory transduction, which has been studied much more extensively to date (see reviews by Ache and Restrepo, 1999; Breer et al., 1994; Corey and Roper, 1992; Dionne and Dubin, 1994; Doty, 1995; Firestein et al., 1996; Gold, 1999).
B. Nature of Olfactory Stimuli Odorants for land-dwelling animals are volatile compounds, typically diluted in large volumes of air. Odorants are only poorly soluble in aqueous solutions such as those that bathe the chemosensitive olfactory epithelium (i.e., mucus layer). Olfactory receptor neurons and VNO sensory neurons are much more sensitive to chemical stimulation than are taste receptor cells. Odorants and pheromones in the range of subnanomolar to micromolar concentrations can be detected. Many animals have a keen sense of smell; for example, dogs can detect as few as 1000 molecules/cm 3. However, reports of extreme chemosensitivity such as the ability of certain species (e.g., moths, dogs, and sharks) to detect single molecules of pheromones or odorants may be overestimates based on theoretical calculations assuming uniform diffusion of the stimulus in air or water. Yet, experiments show that the actual dispersion of odorant and pheromone molecules is more often in the form of irregular plumes rather than diffusion, somewhat like smoke rising from burning incense. Within the plume, the concentrations 2The VNO is likely to be a primary site for pheromone detection. However, some pheromones (such as 5-a-androstenone, a volatile steroid found in boar saliva) instead stimulate the olfactory system. Androstenone is also found in human male axillary sweat. Some researchers have speculated that androstenone functions as a pheromone in humans although the data are not compelling. Lastly, the VNO may serve functions in addition to pheromone detection. For instance, in snakes, the VNO is used to detect prey in addition to pheromones.
825
49. Gustatoryand OlfactorySensoryTransduction of odorants and pheromones are much higher than predicted by simple diffusion. Nonetheless, olfactory and pheromone receptors are notoriously responsive to chemical stimuli. The exquisite sensitivity of these receptor neurons might be considered an optimization of the sensory neuron to the low concentrations of biologically important chemicals that are present in the environment. The low concentration of odorants imposes certain constraints on the possible receptor mechanisms for olfaction (chemosensory transduction in VNO sensory neurons will be discussed later). Namely, amplifying intraeellular second messenger cascades are likely to be involved in transduction, as will be shown to be the case. Additionally, olfactory receptor neurons possess a high membrane resistance. The input resistance of olfactory receptor neurons is substantial (up to 30 GI~). Consequently, tiny receptor currents--some investigators believe even those generated by the opening of a single ion channel---can produce significant voltage changes in an olfactory receptor neuron. These currents are sufficient to depolarize an olfactory receptor neuron to threshold for generating an action potential. Intracellular second messenger cascades and a high input resistance combine to increase the sensitivity of olfactory receptor neurons. Volatile odoriferous compounds partition into the mucus layer covering the olfactory epithelial surface. A specialized carrier or transport protein has been identified in the mucus layer, odorant binding protein (OBP). It has been hypothesized that OBP facilitates the partitioning of odorants into the mucus. According to this hypothesis, odorants bound to OBP are carried and presented to the chemoreceptive surface on the cilia that protrude up into the mucus layer. OBP may also help rid the olfactory epithelium of odoriferous molecules.
C. Initiation of Olfactory Receptor Potentials As far as is known, olfactory transduction involves odorant binding to specific receptors localized in the ciliary membrane of olfactory receptor neurons. 1. Membrane-Bound Receptors The transduction pathways that have been characterized for olfaction to date involve GPCRs that are integral proteins in the ciliary membrane. When occupied by odorant, the receptor activates an intracellular G protein. Unlike gustatory sensory cells, olfactory stimuli do not permeate the plasma membrane through ion channels and carry charge (current). Volatile compounds, being lipophilic, probably do partition directly into the membrane, but this is not believed to be a principal mechanism in olfactory transduction. Receptors for odorants have been cloned and sequenced (Buck and Axel, 1991; Raming et al., 1993; Reed, 1992). Odorant receptors are molecules with seven membranespanning regions and are members of the large G-proteincoupled receptor supeffamily. It has been estimated that in rodents there are 1000 odorant receptor genes. This represents a sizeable fraction of the estimated total genome (~100 000 genes). Data suggest that any given olfactory receptor neuron
expresses a single odorant receptor. Researchers have now been able to express and demonstrate function of cloned odorant receptors in several different systems, including an insect cell line (Sf9, Breer et al., 1998), rat olfactory epithelium (Zhao et al., 1998; Touhara et al., 1999), and HEK-293 cells (Krautwurst et al., 1998). These studies have reported responses to specific odors, thereby identifying a particular receptor with a group of odoriferous ligands. 2. G Proteins in Olfactory Receptor Neurons Odorant binding to receptors triggers second messenger pathways, beginning with activation of G proteins. Researchers postulate that there are two separate G protein pathways for olfaction. The better characterized pathway involves Goif, a G protein that is enriched in olfactory epithelium. Golf is a member of the G s family and activates type III adenylyl eyclase expressed in olfactory epithelium (Pace et al., 1985). It has long been known that exposing isolated olfactory cilia to low concentrations (txM) of certain fruity, floral, or herbaceous odors results in an increase in intracellular cAME These observations were key in establishing adenylyl cyclase in the intracellular pathway for signal transduction in olfaction, especially for what are considered pleasant odors. A less well-established G protein pathway, believed by some to act as an intermediary for signaling in putrid and unpleasant odors, will be discussed later. 3. Cyclic AMP and Cyclic Nucleotide-Gated Channels in Olfactory Receptor Neurons The effector channels for the intracellular second messengers in olfactory receptor neurons are specialized ion channels. Cyclic AMP binds to and rapidly opens cation-selective (Na +, K +, Ca 2+) cyclic nueleotide-gated, or CNG, channels from the inside of the ciliary membrane. This was demonstrated by applying cAMP and cGMP to isolated patches of membrane pulled from olfactory cilia (Nakamura and Gold, 1987) or to isolated olfactory receptor neurons (Firestein et al., 1991) and observing cyclic nucleotide-gated, cationselective conductance increases. CNG channels have little or no voltage-dependence and under physiological conditions (i.e., in the presence of extracellular Ca 2+ and Mg 2+) have a small conductance, ~100 fS. The consequence of odorant stimulation and thus cAMP-gated channel activity is a depolarizing receptor potential (Fig. 7A). CNG channels from olfactory receptor neurons have been cloned and sequenced (Dhallan et al., 1990). Olfactory CNG channels are quite similar to CNG channels found in photoreceptors. CNG channels from photoreceptors and olfactory receptor neurons differ in their sensitivity to cyclic nucleotides. The CNG channels in olfactory receptor neurons respond to cAMP and cGMP alike, with cGMP being somewhat more potent. In contrast, cAMP is far less effective than cGMP in gating photoreceptor CNG channels. It is an enigma that olfactory CNG channels are more sensitive to cGMP than cAMP because odorant stimulation is believed principally to elevate cAME A possible role for an indirect elevation of cGMP in olfactory adaptation (discussed later) may resolve this puzzle. Details of CNG channels are given in Chapter 47.
826
SECTION V Synaptic Transmission and Sensory Transduction
A
[
I 30 mV
300 pA "-
ls
B 0
=•-10 r
-20
1 control i
C
1
i
i
0
10M
4. IP 3 in Olfactory R e c e p t o r N e u r o n s
i -20
k/\ -40
, 1
control , 2
brane potential to threshold and initiates impulses that are conducted by the axon to the olfactory bulb. A key component of the cation current through olfactory CNG channels is Ca 2+ influx. The influx of Ca 2+ has several important consequences in addition to supplying inward (depolarizing) current. First, Ca 2§ influx activates a Ca2+-dependent CIc o n d u c t a n c e in the olfactory cilia. The olfactory cilia are suspended in a mucus that has a C1- concentration of about 55 mM. [C1-]i at the apical tips of olfactory receptor neurons appears to be maintained at a remarkably high level (ca. 70 mM), yielding a C1-equilibrium potential a few mV positive to zero and considerably more positive than resting potential. Consequently, activating the Ca2+-activated anion conductance in these cells causes an efflux of Ci-, that is, a d e p o l a r i z i n g c u r r e n t that complements the influx of cations through CNG channels. This C1- current contributes substantially to the depolarizing receptor current. Indeed, in rat olfactory receptor neurons, the Ca2§ C1- current i s even greater than the initial inward Na § and Ca 2§ current through CNG channels. The combined action of the cAMPgated cation channels and Ca2+-dependent C1-channels may exist to ensure a depolarizing receptor current even in the face of fluctuating extracellular (mucosal) cation concentrations. 3 Second, Ca 2§ influx through CNG channels during odorant stimulation exerts negative feedback on the C N G channel itself. Third, Ca 2§ regulates the activity of a number of the enzymes involved in the second messenger cascades in olfactory signal transduction. These actions of Ca 2§ are particularly important for signal adaptation (discussed later).
, 3
, 4
time (s)
F I G U R E 7. Patch-clamp recordings from olfactory receptor neurons. (A) Depolarizing receptor potential (top) and receptor current (bottom) elicited by applying an odorant (10 mM n-amyl acetate) to the olfactory cilia of an isolated newt olfactory receptor neuron. Modified from Kurahashi (1989). Inhibitory effects of (B) carbon monoxide, CO, and (C) cGMP on receptor currents in salamander olfactory receptor neurons after brief applications of an odorant (100 pM cineole). Traces 1 and 3 in B were recorded before and after recovery from the presence of 800 nM CO in the bathing solution. Traces in C were recorded before and during the presence of 500 nM and 1 pM 8-BrcGMP. CO stimulates the formation of cGMP during odorant stimulation, which ultimately decreases odor responses and may underlie long-lasting olfactory adaptation. Top traces in B and C mark the application of odorant. Modified from Leinders-Zufall et al. (1996).
During odorant stimulation, Ca 2+ and Na + enter the olfactory cilium through CNG channels. This cationic current depolarizes the sensory cell. The cation influx, in concert with a depolarizing C1-current (discussed next), raises the mem-
There is evidence, albeit quite controversial for vertebrate species, that another second messenger cascade may be activated in olfactory receptor neurons. This cascade appears to be triggered by a different set of odorants than those that are transduced by Gol f, adenylyl cyclase, and cAMP. The main impetus to search for alternative signal transduction pathways stems from observations that certain odorants stimulate the rapid generation of inositol t r i s p h o s p h a t e (IP 3) in olfactory receptor cilia. G proteins other than those of the G s category, described previously, are believed to activate the membrane-bound enzyme phospholipase C (PLC). PLC hydrolyzes a phospholipid, phosphatidyl inositol 4,5-bisphosphate, in the plasma membrane of olfactory receptor neurons. The resultant by-products, IP 3 a n d diacyglyceroi (DAG), are powerful bioactive compounds. In other tissues such as
3As stated previously, activating the Ca2+-dependent anion conductance in gustatory sensory cells results in the opposite effect than at the cilia of olfactory receptor neurons. Namely, in gustatory sensory cells this conductance produces an influx of C1- (i.e., a repolarizing, outward current). This is because the basolateral membrane of taste cells-where the C1-conductance channels are situated--is exposed to interstitial fluid and [C1-]i in taste cells is believed to be low. Consequently, in taste cells, the equilibrium potential for C1- is likely to be near the resting potential. Opening C1- channels, especially during depolarizations, will allow C 1- to enter the taste cell and repolarize the membrane.
827
49. Gustatory and Olfactory Sensory Transduction
taste buds (discussed earlier), IP 3 receptors on intracellular Ca 2+ storage compartments (e.g., endoplasmic reticulum) control the release of Ca 2§ into the cytosol. However, Ca 2§ stores have not been demonstrated in olfactory cilia. Instead, IP 3, if it is involved in olfactory transduction, is believed to activate specialized channels in the ciliary plasma membrane. IP3-gated, non-selective cation channels have been demonstrated in olfactory cilia immunocytochemically and in certain species, especially invertebrates, electrophysiologically. Both mechanisms--adenylyl cyclase---~cAMP, and PLC---~ IP3----may coexist in a single olfactory receptor neuron and may be activated by different odorants. This has been clearly shown to occur in invertebrate olfactory receptor neurons, where the two pathways mediate opposing effects---excitation (IP3-gated cation channels) versus inhibition (cAMP-gated K + channels) (Ache, 1994). However, if this is also the case in mammals, it might require the expression of two or more different odorant receptors in the same cell. This would seem to contradict the prevailing belief. Lastly, although certain findings in vertebrate species support the involvement of IP 3 in olfactory signal transduction, the data are controversial. In fact, recent evidence from gene knockout mice lacking olfactory CNG channels indicates that odorant responses to a wide selection of odors--even odors that are known to elevate intracellular IP 3 and complex odor mixtures--are undetectable (Brunet et al., 1996). The ability to eliminate odor responses by selectively eliminating CNG channels in olfactory receptor neurons seriously challenges the hypothesis that IP 3 is directly involved in mammalian olfactory transduction, at least in any straightforward scheme (reviewed by Gold, 1999). The transduction mechanisms for olfactory receptor neurons are summarized in Fig. 8. 5. Transduction in VNO Sensory Neurons The transduction pathways in VNO sensory neurons for pheromones and other chemical stimuli are presently under investigation. It appears that VNO sensory neurons utilize distinctive GPCRs, a different set of G proteins, and different downstream effectors than olfactory receptor neurons. For example, VNO sensory neurons express two families of GPCRs that are unrelated to the large family of odorant receptors (Dulac and Axel, 1995). In rodents, it is estimated that there are 50-100 genes in each of these two families of VNO receptors. The two families of VNO receptors are expressed in two anatomically separate regions of the VNO sensory epithelium. The G proteins Gao and G ai2 are highly expressed in VNO sensory neurons but the olfactory-specific Go1f is not (Berghard and Buck, 1996). Furthermore, stimulating VNO receptors does not appear to prompt the synthesis of cAMP and cation influx through CNG channels, as occurs in olfactory receptor neurons. Instead, VNO sensory neurons express TRP2, an ion channel of the transient receptor potential (TRP) family (Liman et al., 1999). Researchers currently speculate that activation of VNO receptors elevates intracellular second messengers (IP 3 and DAG) that directly gate TRP2 and depolarize the VNO sensory cell.
D. Termination of Olfactory Signals Olfactory receptor neuron excitation is terminated by a number of mechanisms. An obvious one is the unbinding and disappearance of the odorant from the chemoreceptive surface of the cilia, perhaps aided by odorant binding protein (OBP, see Fig. 8). The removal of odorants is complicated by the lipophilic nature of most odoriferous compounds; they will tend to partition into the plasma membrane and thus may linger in the vicinity of the receptors. Although the details are only now unfolding, one proposed mechanism for ridding the chemosensitive surfaces of odorants is the enzymatic modification of odorants by broad-spectrum biotransformation enzymes found in adjacent supporting cells. These enzymes are similar to the detoxification enzymes found in the liver. Olfactory epithelial supporting cells adjacent to receptor neurons contain high concentrations of the detoxifying enzymes glutathione transferase, cytochrome P-450, and UDP gluconosyl transferase (Lazard et al., 1991). Indeed, the lipophilic nature of many odorants and their absorption into the sensory epithelium may necessitate the existence of such degradative mechanisms. Besides disappearance of the odorant, sensory adaptation is also key in terminating olfactory signals. Adaptation in olfactory receptor neurons is explained by multiple mechanisms, many of which involve Ca 2+, as follows. First, shortterm adaptive mechanisms are believed to be set in motion by the Ca 2+ that enters through CNG channels during odorant stimulation. Ca 2§ inhibits CNG channels themselves (Kurahashi and Menini, 1997), a direct negative feedback on the olfactory signal. That is, an increase in Ca 2§ in the ciliary cytosol will depress odorant responses by shutting down effector (CNG) channels. Second, heightened intracellular Ca 2+ also depresses the cascade of enzymes that are triggered during olfactory transduction. At high (mM) concentrations, Ca 2§ inhibits adenylyl cyclase, thereby reducing the continued generation of cAMP during maintained stimulation. In olfactory receptor neurons, this inhibition appears to be mediated by Ca2+-calmodulin-induced phosphorylation via the enzyme Ca2+-calmodulin-dependent protein kinase II (CaMKII). However, at nM to pM concentrations, Ca 2§ has the opposite effect, namely, it stimulates adenylyl cyclase. Consequently, the net effect of Ca 2+ influx on adenylyl cyclase activity during odor stimulation is concentrationdependent. Additionally, Ca 2§ influx stimulates phosphodiesterase (PDE), thereby accelerating the degradation of cAMP and reducing the odorant response. Third, Ca 2+ influx activates Ca2+-dependent K + channels that initiate an outward, repolarizing current. Counterbalancing this, Ca 2§ also activates a C1-conductance in the olfactory cilia that produces an inward depolarizing current, as described previously. Thus, the actions of Ca 2+ are complex. The net effect of Ca 2+ entry through CNG channels during olfactory stimulation appears to be to reduce the responsiveness of olfactory receptor neurons perhaps as much as 20-fold, moving the stimulusresponse relationship to a higher range of odorant concentrations. Longer-lasting adaption to a maintained odor stimulus involves another mechanism. A prolonged adaptation in
828
SECTION V Synaptic Transmission and Sensory Transduction
F I G U R E 8. Transduction mechanisms in olfactory receptor neurons. A portion of the shaft of an olfactory cilium is illustrated. In the mucus layer, odorants may be bound to cartier molecules, odorant binding protein (OBP). (1) Some odorants (odorant l) activate receptors (OR1) that (2) stimulate a G protein (Golf) and (3) activate adenylyl cyclase (AC). This converts ATP to cAMP. (4) cAMP is hydrolyzed by phosphodiesterase, PDE. (5) More importantly, cAMP activates CNG channels, allowing an influx of Ca 2+ (also Na+). Ca 2+ influx opens Ca2+-dependent C1-channels, allowing an effiux of C1- (i.e., a depolarizing current). (6) cAMP also stimulates protein kinases that phosphorylate and desensitize the odorant receptor, thereby quenching the response and contributing to adaptation. Intracellular Ca 2+ decreases the long-term responsiveness of the olfactory receptor neuron, also producing adaptation. Ca 2+ accomplishes this by exerting negative feedback on the CNG channel, positive feedback on PDE, and concentration-dependent effects on AC. Some of these actions are indicated by white wavy lines. (7) Odorant activation also generates carbon monoxide, which stimulates soluble guanylyl cyclase (sGC) to produce cGMP. (8) cGMP exerts a tonic, subthreshold excitatory influence on CNG channels, allowing a small but steady influx of Ca 2+ into the olfactory receptor neuron. Other odorants (odorant2) may activate receptors (OR2) that are coupled via a G protein to phospholipase C (PLC). When these receptors are activated, the intracellular second messenger, inositol trisphosphate (IP3), is formed. IP 3 opens channels in the olfactory receptor neuron membrane. The existence and significance of odorant-activated IP3 pathways in olfactory transduction in vertebrates is currently under debate but is well-established in invertebrates.
olfactory r e c e p t o r n e u r o n s , involving an intriguing role for the novel g a s e o u s n e u r o t r a n s m i t t e r c a r b o n m o n o x i d e ( C O ) , has b e e n p r o p o s e d ( L e i n d e r s - Z u f a l l et al., 1996). CO, like the related n e u r o t r a n s m i t t e r nitric oxide (NO), is a h i g h l y - d i f f u s i b l e substance and is g e n e r a t e d during olfactory t r a n s d u c t i o n via a m e c h a n i s m that is not yet wellcharacterized. CO, in turn, activates s o l u b l e g u a n y l y l cyc l a s e (sGC) and leads to the p r o d u c t i o n of c G M P in the
s t i m u l a t e d o l f a c t o r y r e c e p t o r n e u r o n as well as in surr o u n d i n g cells. As d i s c u s s e d previously, c G M P is a very effective ligand for o l f a c t o r y C N G channels. At first glance, g e n e r a t i n g c G M P m i g h t s e e m c o n t r a r y to adaptation and r e p r e s e n t positive f e e d b a c k during o d o r a n t stimulation. However, activation of olfactory C N G c h a n n e l s via the CO p a t h w a y p r o d u c e s only a low-level, s u b t h r e s h o l d i n w a r d c u r r e n t . This m a i n t a i n s a persistent
49. Gustatory and Olfactory Sensory Transduction
829
tion controls the active state of odorant receptors and ultimately, the responsiveness of olfactory receptor neurons. Collectively, short-term adaptation, long-term adaptation, and receptor desensitization act to shape and terminate olfactory responses to prolonged odor stimulation. Sensory adaptation in many sensory systems, including olfaction, serves a dual purpose. First, over time a maintained stimulus ceases to excite responses, that is, the sensory organ adapts to the stimulus. In this sense, adaptation tends to accentuate signals from stimuli that fluctuate and emphasizes the dynamic aspects of sensory stimuli. Second, adaptation reduces responsivity to increasingly intense stimuli. This prevents the output from the sensory organs from becoming saturated by only moderately intense stimulus intensities while at the same time preserving the ability of the organ to respond well to very weak stimuli. That is, adaptation greatly extends the effective range of stimulus intensities over which a sensory organ can respond.
trickle of Ca 2+ into the olfactory receptor neurons, c h r o n ically r e d u c i n g C N G c h a n n e l activity. The rapid transient influx of Ca 2+ through C N G channels (the immediate effect of odorant stimulation), c o m b i n e d with a sustained low-level influx via c G M P activation of these channels (the CO pathway), overall leads to a long-lasting (minutes) reduction of the sensitivity of olfactory receptor neurons after the initial excitation. 4 In addition to these sensory adaptation mechanisms, desensitization of o d o r a n t receptors following odorant stimulation also plays a key role in t e r m i n a t i n g o d o r a n t responses (Schleicher et al., 1993). cAMP produced by odor stimulation opens CNG cation channels and secondarily activates protein kinase A (PKA). PKA and a specialized G protein-coupled receptor kinase, GRK35, phosphorylate odorantbond receptors (Fig. 9). These protein kinases, together with regulatory proteins such as arrestin, render the odorant receptor inactive and quench the subsequent transduction cascade. Receptors are resensitized by the action of phosphatases. Thus, a cycle of p h o s p h o r y l a t i o n / d e p h o s p h o r y l a -
IV. S u m m a r y
Taste and olfaction share certain common features. For example, transduction mechanisms in both these chemical senses involve G-protein-coupled receptors and changes in intracellular cyclic nucleotides, particularly cAMP. This is the case for certain sweet, bitter, and u m a m i stimuli as well as odoriferous stimuli. IP 3 mechanisms may also play a role.
4previously, it was believed that NO played this role in olfactory adaptation. However, evolving data utilizing blockers of CO synthesis now implicate CO, not NO, as the diffusible intra- and intercellular controller of CNG channel activity in olfactory receptor neurons. 5GRK3 was formerly named /3 adrenergic receptor kinase 2, or /3ARK2.
1200 -
1200 "
anti-GRK3 10oo "
ex0
1000 "
800
800 "
O
600
Ill
alsh inhibitor
600antl-~ARK-
< 400 "
400 "
9
Control
200 "
a
a
n
,
I
I ....
I
i
\
200 "
t
,I
'
i ~ a
Control
........
I
I
a
a
i
- 100 0 100 200 300 400 500 - 100 0 100 200 300 400 500 FIGURE 9. Desensitization of odorant receptors during prolonged stimulation. Olfactory second messenger signaling is quenched by phosphorylation cascades. Left, inhibiting protein kinase activity with Walsh inhibitor or heparin markedly prolongs increases in cAMP induced by odorant stimulation in cilia preparations from the rat. (O) odorant alone (1 pM citralva); (A) odorant stimulation of cilia pretreated with Walsh inhibitor (3.8 pM); (*) odorant stimulation of cilia pretreated with heparin (1 pM). Right, pretreating the cilia preparation with antibodies to GRK3 specifically prevents desensitization of the odorant receptor and prolongs its activation(measured here as an increase in cAMP). (Q) odorant alone (1 pM citralva); (A) odorant stimulation of cilia pretreated with anti-/3ARK-1 antibodies (diluted 1:5000); (*) odorant stimulation of cilia pretreated with anti-GRK3 antibodies (diluted 1:5000); abscissae, time in ms. Modified from Schleicher et al. (1993).
830
SECTION V Synaptic Transmission and Sensory Transduction
In taste, IP 3 stimulates the release of Ca 2+ from intracellular stores. In olfaction, IP 3 is believed to gate specific channels. Transduction channels and integral membrane receptors appear to be concentrated on the exposed apical tips of both types of chemosensory receptor cells, although the basolateral membranes of taste cells may be additional sites for chemosensory transduction. In some regards, sensory transduction in taste and olfaction is similar to that in other modalities, notably photoreception. Photoreceptors, olfactory receptor neurons, and gustatory sensory cells possess ion channels that are modulated by cyclic nucleotides. Further, transducin, the G protein in photoreceptors, is also found in taste cells. However, there are key differences between olfaction and taste. Olfactory receptor neurons are optimized for low concentrations of chemical stimuli, gustatory sensory cells for higher concentrations. Furthermore, whereas an increase in c A M P opens cation channels in olfactory receptor neurons, c A M P closes K + channels in taste receptor cells (by activating PKA, leading to channel phosphorylation). An additional difference is that taste cells possess ligand-gated (ionotropic) ion channels and ion channels that are either permeated or blocked by certain chemical stimuli (especially ions). These mechanisms do not participate in olfactory transduction. Areas that are current topics of intensive research in the peripheral transduction mechanisms for taste and olfaction include, among others, extending our knowledge of the G-protein-coupled effector cascades in the receptor cells, discovering the molecular structure of olfactory and taste receptor proteins, understanding how taste and olfactory receptor neurons distinguish among the multitude of different chemical stimuli, learning how the output from the peripheral sensory organs is encoded, and determining the synaptic mechanisms that transmit signals between cells within the taste bud and on to sensory afferent fibers that innervate the taste bud.
Bibliography Ache, B. (1994). Towards a common strategy for transducing olfactory information. In "Seminars in Cell Biology." (H. Breer, Ed.), pp. 55-63, Academic Press, San Diego, CA. Ache, B. W., and Restrepo, D. (2000). Olfactory transduction. In "Neurobiology of Taste and Smell, 2nd edition." (T. E. Finger, D. Restrepo, and W. L. Silver, Eds.) Wiley, New York, 159-177. Adler, E., Hoon, M. A., Mueller, K. L., Chandrashekar, J., Ryba, N. J. P., and Zuker, C. S. (2000). A novel family of mammalian taste receptors. Cell 100, 693-702. Berghard, A., and Buck, L. B. (1996). Sensory transduction in vomeronasal neurons: evidence for Gao, Gai2 and adenylyl cyclase II as major components of a pheromone signaling cascade. J. Neurosci. 16, 909-918. Bernhardt, S. J., Naim, M., Zehavi, U., and Lindemann, B. (1996). Changes in IP 3 and cytosolic Ca2+ in response to sugars and nonsugar sweeteners in transduction of sweet taste in the rat. J. Physiol. 490, 325-336. Bigiani, A. (2001). Mouse taste cells with glia-like membrane properties. J. Neurophysiol. In press. B igiani, A., and Roper, S. D. (1991). Mediation of responses to calcium in taste cells by modulation of a potassium conductance. Science 252, 126--128.
Breer, H., Krieger, J., Meinken, C., Kiefer, H. and Strotmann, J. (1998). Expression and functional analysis of olfactory receptors. Ann. NYAcad. Sci. 855, 175-181. Breer, H., Raming, K., and Krieger, J. (1994). Signal recognition and transduction in olfactory neurons. Biochim. Biophys. Acta 1224, 277-287. Brunet, L. J., Gold, G. H., and Ngai, J. (1996). General anosmia caused by a targeted disruption of the mouse olfactory cyclic nucleotidegated cation channel. Neuron 17, 681-693. Buck, L., and Axel, R. (1991). A novel multigene family may encode odorant receptors: a molecular basis for odor recognition. Cell 66, 175-187. Caicedo, A., and Roper, S. D. (2001). Taste receptor cells that discriminate between bitter stimuli. Science, In press. Chandrashekar, J., Mueller, K. L., Hoon, M. A., Adler, E., Feng, L., Guo, W., Zuker, C. S., and Ryba, N. J. E (2000). T2Rs function as bitter taste receptors. Cell 100, 703-711. Chaudhari, N., Landin, A. M., and Roper, S. D. (2000). A metabotropic glutamate receptor variant functions as a taste receptor. Nat. Neurosci. 3, 113-119. Corey, D. E, and Roper, S. D. (1992). "Sensory Transduction." Rockefeller University Press, New York. Delay, R. J., Kinnamon, S. C., and Roper, S. D. (1997). Serotonin modulates voltage-dependent calcium current in Necturus taste cells. J Neurophysiol. 77, 2515-2524. DeSimone, J. A. (1991). Transduction in taste receptors. Nutrition 7, 146-147. Dhallan, R. S., Yau, K.-W., Schrader, K. A., and Reed, R. R. (1990). Primary structural and functional expression of a cyclic nucleotideactivated channel from olfactory neurons. Nature 347, 184-187. Dionne, V., and Dubin, A. E. (1994). Transduction diversity in olfaction. J. Exper. Biol. 194, 1-21. Doty, R. L. (1995). "Handbook of Olfaction and Gustation." Marcel Dekker, New York. Dulac, C., and Axel, R. (1995). A novel family of genes encoding putative pheromone receptors in mammals. Cell 83, 195-206. Ewald, D. E., and Roper, S. (1994). Bidirectional synaptic transmission in Necturus taste buds. J. Neurosci. 14, 3791-3801. Finger, T. E., Restrepo, D., and Silver, W. L. (2000). "Neurobiology of Taste and Smell, 2nd edition." Wiley, New York. Firestein, S., Breer, H., and Greer, C. A. (1996). Olfaction: what's new in the nose. J. Neurobiol. 30, 1-176. Firestein, S., Darrow, B., and Shepherd, G. (1991). Activation of the sensory current in salamander olfactory receptor neurons depends on a Gprotein mediated cAMP second messenger system. Neuron 6, 825-835. Furue, H., and Yoshii, K. (1997). In situ tight-seal recordings of taste substance-elicited action currents and voltage-gated Ba currents from single taste bud cells in the peeled epithelium of mouse tongue. Brain Res. 776, 133-139. Gilbertson, T. A., and Zhang, H. (1998). Self-inhibition in amiloridesensitive sodium channels in taste receptor cells. J. Gen. Physiol. 111, 667-677. Gilbertson, T. A., Fontenot, D. T., Liu, L., Zhang, H., and Monroe, W. T. (1997). Fatty acid modulation of K+ channels in taste receptor cells: gustatory cues for dietary fat. Am. J. Physiol. 272, C1203-1210. Gilbertson, T. A., Roper, S. D., and Kinnamon, S. C. (1993). Proton currents through amiloride-sensitive Na+ channels in isolated hamster taste cells: enhancement by vasopressin and cAMP. Neuron 10, 1-20. Gold, G. H. (1999). Controversial issues in vertebrate olfactory transduction. Annu. Rev. Physiol. 61, 857-871. Graziadei, P. E C., and Monti Graziadei, G. A. (1978). Continuous nerve cell renewal in the olfactory system. In "Handbook of Sensory Physiology." (M. Jacobson, Ed.).Vol. IX, pp. 55-82. SpringerVerlag, New York. H~inig, D. E (1901). Zur Psychophysik des Geschmackssinnes. Philosophische Studien 17, 576-623.
49. Gustatory and Olfactory Sensory Transduction Heck, G. L., Mierson, S., and DeSimone, J. A. (1984). Salt taste transduction occurs through an amiloride-sensitive sodium transport pathway. Science 223, 403-405. Herness, S., and Chen, Y. (1997). Serotonin inhibits calcium-activated K + current in rat taste receptor cells. Neuroreport 8, 3257-3261. Hoon, M. A., Adler, E., Lindemeier, J., Battey, J. E, Ryba, N. J., and Zuker, C. S. (1999). Putative mammalian taste receptors: a class of taste-specific GPCRs with distinct topographic selectivity. Cell 96, 541-551. Huang, L., Shanker, Y. G., Dubauskaite, J., Zheng, J. Z., Yan, W., Rosenzweig, S., Spielman, A. I., Max, M., and Margolskee, R. E (1999). Gyl3 colocalizes with gustducin in taste receptor cells and mediates IP 3 responses to bitter denatonium. Nat. Neurosci. 2, 1055-1062. Kashiwayanagi, M., Miyake, M., and Kurihara, K. (1983). Voltagedependent Ca 2+ channel and Na + channel in frog taste cells. Am. J. Physiol. 244, C82-88. Kinnamon, S. C., and Margolskee, R. F. (1996). Mechanisms of taste transduction. Curr. Opin. Neurobiol. 6, 506-513. Kolesnikov, S. S., and Margolskee, R. E (1995). A cyclic-nucleotidesuppressible conductance activated by transducin in taste cells. Nature 376, 85-88. Krautwurst, D., Yau, K. W., and Reed, R. R. (1998). Identification of ligands for olfactory receptors by functional expression of a receptor library. Cell 95, 917-926. Kretz, O., Barbry, P., Bock, R., and Lindemann, B. (1999). Differential expression of RNA and protein of the three pore-forming subunits of the amiloride-sensitive epithelial sodium channel in taste buds of the rat. J. Histochem. Cytochem. 47, 51-64. Kurahashi, T., and Menini, A. (1997). Mechanism of odorant adaptation in the olfactory receptor cell. Nature 385, 725-729. Kurahashi, T. (1989). Activation by odorants of cation-selective conductance in the olfactory receptor cell isolated from the newt. J. Physiol. 419, 177-192. Lazard, D., Zupko, K., Poria, Y., Nef, E, Lazarovits, J., Horn, S., Khen, M., and Lancet, D. (1991). Odorant signal termination by olfactory UDP glucuronosyl transferase. Nature 349, 790-793. Leinders-Zufall, T., Shepherd, G. M., and Zufall, F. (1996). Modulation by cyclic GMP of the odour sensitivity of vertebrate olfactory receptor cells. Proc. R. Soc. Lond. [Biol.] 263, 803-811. Li, X.-J., Blackshaw, S., and Snyder, S. H. (1994). Expression and localization of amiloride-sensitive sodium channel indicate a role for non-taste cells in taste perception. Proc. Natl. Acad. Sci. USA 91, 1814-1818. Liman, E. R., Corey, D. E, and Dulac, C. (1999). TRP2: a candidate transduction channel for mammalian pheromone sensory signaling. Proc. Natl. Acad. Sci. USA 96, 5791-5796. Lin, W., Finger, T. E., Rossier, B. C., and Kinnamon, S. C., (1999). Epithelial Na + channel subunits in rat taste cells: localization and regulation by aldosterone. J. Comp. Neurol. 405, 406-420. Lindemann, B. (1996). Taste reception. Physiol. Rev. 76, 719-766. Lindemann, B. (1999). Receptor seeks ligand: on the way to cloning the molecular receptors for sweet and bitter taste. Nat. Med. 5, 381-382. Mackay-Sire, A., and Kittel, E (1991a). Cell dynamics in the adult mouse olfactory epithelium: a quantitative autoradiographic study. J. Neurosci. 11, 979-984. Mackay-Sire, A., and Kittel, E (1991b). On the life span of olfactory receptor neurones. Eur. J. Neurosci. 3, 209-215.
831 McLaughlin, S. K., McKinnon, P. J., and Margolskee, R. E, (1992). Gustducin is a taste-cell-specific G protein closely related to the transducins. Nature 357, 563. McLaughlin, S. K., McKinnon, P .J., and Margolskee, R. E (1993). Gustducin and transducin: a tale of two G proteins. In "The Molecular Basis of Smell and Taste Transduction." Ciba Foundation Symposium No. 179, pp. 186-200. Wiley, Chichester. Matsunami, H., Montamayeur, J. E, and Buck, L. B. (2000). A family of candidate taste receptors in human and mouse. Nature 404, 601-604. Naim, M., Seifert, R., Nurnberg, B., Grunbaum, L., and Schultz, G. (1994). Some taste substances are direct activators of G-proteins. Biochem. J. 297, 451-454. Nakamura, T., and Gold, G. H. (1987). A cyclic nucleotide-gated conductance in olfactory receptor cilia. Nature 325, 442-444. Ossebaard, C. A., and Smith, D. V. (1996). Amiloride suppresses the sourness of NaC1 and LiC1. Physiol. Behav. 60, 1317-1322. Pace, U., Hanski, E., Salomon, Y., and Lancet, D. (1985). Odorantsensitive adenylate cyclase may mediate olfactory reception. Nature 316, 255-258. Raming, K., Krieger, J., and Strotmann, J. (1993). Cloning and expression of odorant receptors. Nature 361, 353-356. Reed, R. R. (1992). Signalling pathways in odorant detection. Neuron 8, 205-209. Roper, S. (1983). Regenerative impulses in taste cells. Science 220, 1311-1312. Roper, S. (1992). The microphysiology of peripheral taste organs. J. Neurosci. 12, 1127-1134. Rossler, E, Kroner, C., Freitag, J., Noe, J., and Breer, H. (1998). Identification of a phospholipase C beta subtype on rat taste cells. Eur. J. Cell Biol. 77, 263-261. Ruiz-Avila, L., McLaughlin, S. K., Wildman, D., McKinnon, E J., Robichon, A., Spickofsky, N., and Margolskee, R. F. (1995). Coupling of bitter receptor to phosphodiesterase through transducin in taste receptor cells. Nature 376, 80-85. Schleicher, S., Boekhoff, I., Arriza, J., Lefkowitz, R. J., and Breer, H. (1993). A beta-adrenergic receptor kinase-like enzyme is involved in olfactory signal termination. Proc. Natl. Acad. Sci. USA 90, 1420-1424. Simon, S. A., and Roper, S. (1993). "Mechanisms of Taste Transduction." CRC Press, Boca Raton, FL. Taylor, R., and Roper, S. (1994). CaZ+-dependent C1- conductance in taste receptor cells from Necturus. J. Neurophysiol. 72, 475-478. Touhara, K., Sengoku, S., Inaki, K., Tsuboi, A., Hirono, J., Sato, T., Sakano, H., and Haga, T. (1999). Functional identification and reconstitution of an odorant receptor in single olfactory neurons. Proc. Natl. Acad. Sci. USA 96, 4040-4045. Ugawa, S., Minami, Y., Guo, W., Saishin, Y., Takatsuj, K., Yamamoto, T., Tohyama, M., and Shimada, S. (1998). Receptor that leaves a sour taste in the mouth. Nature 395, 555-556. Yang, H., Wanner, I. B., Roper, S. D., and Chaudhari, N. (1999). An optimized method for in situ hybridization with signal amplification that allows the detection of rare mRNAs. J. Histochem. Cytochem. 47, 431-446. Ye, Q., Heck, G. L., and DeSimone, J. A. (1993). Voltage dependence of the rat chorda response to Na + salts: implications for the functional organization of taste receptor cells. J. Neurophysiol. 70, 167-178. Zhao, H., Ivic, L., Otaki, J. M., Hashimoto, M., Mikoshiba, K., and Firestein, S. (1998). Functional expression of a mammalian odorant receptor. Science 279, 237-242.
Stephen D. Roper and Michael S. Grace
Appendix: Infrared Sensory Organs
I. Introduction Infrared radiation is an important and ubiquitous component of the diverse physical environments in which animals live. All warm objects radiate infrared energy. Naturally occurring infrared sources include terrestrial, atmospheric, and astronomical bodies. Animals, especially ectothermic species, use radiant infrared energy to assist in maintaining their body temperature and must avoid excess infrared radiation to avoid overheating. In a few species, infrared radiation is also used in a novel manner--to form spatial images of the environment and to hunt prey. These animals possess specialized infraredreceptive sensory organs, the topic of this Appendix. Infrared-receptive organs occur in both vertebrate and invertebrate animals. Buprestid beetles, including the European fire beetle Melanophila acuminata, have specialized infraredreceptive organs on the ventral surfaces of their abdomens. Melanophila and similar beetles congregate in large numbers at forest fire sites to deposit eggs in recently bumed wood. They likely use chemoreceptors and mechanoreceptors (to detect wind direction) to guide them to sites of fires, and their infrared detectors to identify potential egg-laying sites of appropriate temperature. The deep-sea shrimp Rimicaris exoculata also possesses an elaborate organ capable of detecting 350 ~ "black smoker" thermal vents in the deep Atlantic ocean. Their infrared sensory detection system may help the shrimp maintain close proximity to life-sustaining thermal vents in the deep sea without directly contacting the steep thermocline (350 ~ to 2 ~ over a 2 to 3 cm lateral distance) adjacent to the thermal vent. The common vampire bat (Desmodus rotundus) may possess infrared radiation detectors, but its infrared sensory system has been little investigated. Behavioral experiments show that vampire bats can detect infrared radiation as low as 50 pW/cm 2, and thus should be able to detect mammalian prey at distances of up to 16 cm. Vampire bats may use their infrared detection system to locate suitable prey. However, by far the best-studied infrared sensory detectors are those of snakes. As stated previously, ectothermic animals such as snakes rely on infrared radiation to regulate their body temperature. In some snakes (Boidae, which include boa constrictors and pythons, and Crotalinae, which include rattlesnakes and water moccasins), infrared radiation is also used for accurate and precise targeting of prey. These snakes preferentially feed on endothermic (warmblooded) prey. Unlike other infrared-sensing animals, boid and crotaline snakes are the only animals known to form spatial images of their environments using extremely sensitive infrared-receptive organs.
Cell Physiology Sourcebook: A Molecular Approach, Third Edition
8 3 2
Boid and crotaline snakes are efficient predators. While vision may be useful for some aspects of their predatory behaviors, infrared imaging alone can allow accurate prey targeting. Snakes experimentally blinded by occluding both eyes and snakes born without eyes are capable of accurately placing strikes at mammalian prey (Kardong and Mackessey, 1991). Snakes lacking one eye target their strikes as well as do normally sighted snakes. Monocularly occluded snakes do not target preferentially on the sighted side. Even in snakes with an intact visual system, vision may be relatively unimportant for prey targeting. Snakes efficiently and accurately target prey regardless of visual contrast (black versus white mice against a black background) (Grace et al., 2001).
I!. Nature of the Stimulus: What is Infrared (IR) Radiation? Radiant infrared energy is a region of the spectrum of electromagnetic radiation that lies between visible light (wavelengths of 400 to 750 nm) and microwaves (wavelengths of --1 mm to 1 cm). The English astronomer Sir William Herschel discovered infrared radiation at the beginning of the nineteenth century. Herschel was exploring the ability of light passing through a prism to heat a thermometer that he positioned at different points in the color spectrum. Herschel noticed that visible light was capable of heating the thermometer to some extent but that this effect was much more pronounced when the thermometer was placed beyond the red end of the spectrum. Electromagnetic radiation in the infrared region is absorbed as the radiation passes through objects in its pathway. Most of this absorbed infrared energy is dissipated by the resulting increase in molecular collisions (i.e., heat). Infrared radiation penetrates body tissues deeper than does visible light, and thus IR generators produce the so-called deep heat used in physical therapy. Any object above absolute zero radiates electromagnetic waves in the infrared, the more so as it is warmed. With increasing heat, an object radiates increasingly shorter wavelengths. Objects at room temperature radiate entirely in the IR and not at all in the visible light spectrum. At 500 ~ objects begin to radiate also in the visible spectrum and they appear dull red. Endothermic (warm-blooded) animals such as rats and mice radiate maximally in the infrared region at a wavelength of approximately 10 ~m. IR radiation can thus be used to detect objects in the absence of visible light (i.e., in the dark) with an object's detection being enhanced by having a temperature that contrasts with its background temperature.
Copyright 9 2001 by Academic Press.All rights of reproduction in any form reserved.
833
49. Gustatory and Olfactory Sensory Transduction
Water vapor and carbon dioxide in the atmosphere absorb and severely attenuate the transmission of certain wavelengths of the IR spectrum and allow other wavelengths to pass. That is, there are windows for IR transmission in the atmosphere. One of these windows is in a region where certain IR-sensitive photographic emulsions absorb well, thereby allowing IR images of landscapes and cityscapes to be collected by satellite cameras. Another atmospheric transmission window includes the region around 10 pm, that is, the region of the IR spectrum in which endothermic animals maximally radiate. IR sensory organs in snakes are believed to have a maximum sensitivity in this region (Grace et al., 1999) and thus are well suited to detect endothermic prey.
Iil. Infrared-Sensitive Pit Organs in S n a k e s Infrared sensory organs in snakes consist of pit organs. Pit organs are invaginations within or between scales in the head. These invaginations are up to 3 to 4 mm wide and 3 to 4 mm deep. In boid snakes (e.g., pythons) the pit organs are arrayed in rows within labial scales of both the upper and lower jaws. 1 In crotaline snakes (e.g., rattlesnakes), there is a single facial pit organ on each side of the face between the nostril and eye (Fig. A-1). There may be additional infraredsensitive receptor terminals in the oral cavity (especially the palate) that may be important for guiding the predatory strike when the snake's mouth is open and fangs extended (Dickman et al., 1987) (Fig. A-2). A dense plexus of infrared-receptive primary sensory afferent nerve terminals is present in pit organs. In boid snakes, these terminals lie in the epidermis just under the surface of the pit organ. Immediately below the layer of infrared-receptive neuronal terminals is a thick capillary network. In crotaline snakes, a thin membrane (15 pm) stretches across the cavity of the pit (Fig. A-1B). Infrared-receptive nerve terminals are embedded within this membrane, also over a thick capillary bed. Infrared receptor terminals and the dense capillary plexus are mainly confined to the pit organs. 2 The anatomical configuration of the pit organ, namely that the receptive tissue is stretched across the bottom of a cavity that has a narrowed opening to the outside environment, provides the snake with a means to direct its pit organs towards prey. Thus, the pit organs can provide information about the location of an infrared-emitting source in front of the snake. A distinct image, p e r se, such as a pin-hole camera might afford, is not formed in the pit organ, but the shadows cast by the infrared radiation passing through the narrowed opening into the pit project a crude approximation of the infrared source onto the sensory terminals. This "image" is conveyed to the snake's brain, much like visual information, to establish a spatiotopic map on the optic tectum (see Section III.A).
1Boid snakes that do not have obvious pit organs possess clusters of infrared-receptive neuronal terminals in these same labial areas. 2The oral cavity of rattlesnakes is also sensitive to infrared and possesses dense ramifications of infrared-receptive sensory nerve terminals, similar to those in pit organs (Dickman et al., 1987).
Infrared receptor terminals form large, highly-branched, entwined clusters surrounding an aggregate of Schwann cells. Collectively, this structure is called a terminal nerve mass (--30-100 pm diameter) (Fig. A-3). Each terminal branch is densely packed with mitochondria. Clusters of electron lucent vesicles occur just beneath the plasma membrane. These specialized endings are the sites where infrared transduction takes place (see Section III.B). A complex and extensive bed of capillary loops embraces the terminal nerve masses. This capillary bed is believed to have a dual function: (1) to provide a robust oxygen supply for the mitochondria in the receptor terminals, and (2) to act as a thermal exchanger to carry away excess heat and prevent the epithelial tissues from retaining temperature increases produced by infrared radiation (Amemiya et al., 1999). The thermal exchange function of the capillary bed, combined with the low heat capacity of the thin epithelial tissues in the pit organs, may ensure that the sensory receptors are more ready to respond to the transient temperature increases that are expected to occur when an infrared source (e.g., prey) passes in front of the pit organ (discussed in Section III.B).
A. Innervation and Central Nervous System Pathways of Pit Organs Pit organs are innervated by branches of the trigeminal nerve (reviewed by Molenaar, 1992). These same nerve branches also subserve mechanoreception, nociception, and thermoreception in the facial region of reptiles, mammals, and other vertebrates. Cell bodies for the infrared-sensitive nerve terminals are located in sensory ganglia of the trigeminal nerve. In infrared-sensitive snakes, sensory neurons in the trigeminal ganglia synapse with neurons in two distinct nuclei located in the hindbrain--the nucleus of the solitary tract (NST) and a unique group of cells termed the nucleus of the lateral descending trigeminal tract (LTTD) (see Molenaar, 1992). Neurons in the NST subserve such functions as mechanoreception, thermoreception, and nociception, as in other vertebrates. However, the LTTD is found only in infrared-sensing snakes and is responsible for processing information specifically from the pit organs (Stanford et al., 1981). This nucleus is present in all infrared-sensitive snakes studied thus far, and is absent from animals that are not infrared-sensitive. Ultimately, infrared information from the LTTD reaches the optic tectum. Spatial relationships among the axons and synapses of the infrared receptor pathway are preserved such that a systematic topographical map of the receptive membrane in the pit organ is projected onto the tectum, much like the orderly retinotopic mapping of visual information that also occurs in the optic tectum. Indeed, the infrared sensory map overlies and is aligned with the retinotopic map. Individual neurons in the optic tectum may be excited by visual and infrared stimuli arising from approximately the same point in space 3 (Hartline et al., 1978). Thus,
3The two maps--visual and infrared---do not appear to overlap well in the peripheral fields of view. The overlap is more precise in the central field, that is, directly in front of the snake.
834
SECTION V Synaptic Transmission and Sensory Transduction
835
49. Gustatory and Olfactory Sensory Transduction
FIGURE A-2. Photograph of a rattlesnake striking at prey, demonstrating how infrared sensing receptors in the oral cavity may play a role in guiding the strike. Photograph by Pete Carmichael from Florida's Fabulous Reptiles and Amphibians (1991) World Publications, Tampa, FL.
snakes possessing infrared-sensitive pit organs view their environments simultaneously using two distinct regions of the electromagnetic spectrum.
B. Infrared Sensory Transduction Infrared-detecting pit organs in snakes are exquisitely sensitive and selective (Bullock and Diecke, 1956; Gamow and Harris, 1973; de Cock Buning et al., 1981; Newman and Hartline, 1982; Hartline, 1999). The adequate stimulus is a sharp thermal gradient, or contrast, in the surrounding environment, such as an endothermic animal (e.g., a mouse) against thermoneutral vegetation, or cool prey (e.g., a frog) against a cooler background (e.g., a pond). The temperature of the snake's body itself is not a signifi-
cant factor, nor is the average temperature of the surroundings. Air temperature is irrelevant; the pit organ responds only to radiant energy in the infrared, not conductive energy transferred by warm air. Thus, a snake is an effective hunter in the heat of midday as well as in the cool of the night. The sensitivity of pit organs is such that a pit viper can use its infrared detector to orient to and accurately strike a rat at a distance of about 1/2 meter. At this distance, a rat with a surface temperature 10 ~ above background emits infrared radiation (~10 ]am wavelength) that would be expected to warm the pit organ tissues by only 0.001 ~ (Bullock and Diecke, 1956). This low threshold for temperature detection was also observed experimentally. Researchers infused water of different temperatures into pit organs and measured threshold nerve responses to
FIGURE A- 1. Infraredpit organ in the rattlesnake. (A) Photograph of the head of a rattlesnake (Crotalus viridis) showing the distinctive pit organ between the nostrils and the eye. Photograph by Bill Love from The WorM's Most Spectacular Reptiles and Amphibians (1997) World Publications, Tampa, FL. (B) Schematic drawing of the internal structure of the infrared sensory organ in rattlesnake (Crotalus viridis). Reproduced from Gamow and Harris (1973). (C) Higher magnification schematic view of the rattlesnake pit organ. Reproduced from Newman and Hartline (1982).
836
SECTION V Synaptic Transmission and Sensory Transduction
FIGURE A-3. Reconstruction of the nerve terminals for infrared sensing trigeminal sensory axons. The branches form nerve terminal masses that include Schwann cells and receptor terminals. The nerve terminals are filled with numerous mitochondria. Reproduced from Terashima et al. (1970).
temperature changes as small as 0.003 ~ in 0.06s (Bullock and Diecke, 1956). Rapid temperature increases as small as 0.1 ~ in the pit organ elicit robust responses (i.e., well above threshold detection) in the receptors (de Cock Buning et al., 1981).4 How infrared radiation is transduced into neuronal excitation remains a mystery. It is unproven whether snake pit organ infrared receptors operate on a photochemical or thermal mechanism. That is, it is not yet known whether pit organ receptors are stimulated by specific wavelengths of infrared radiation, like photoreceptors in the retina, or whether the receptors are stimulated by temperature increases. The spectral sensitivity of snake pit organs has not been systematically investigated in much detail. Thus, the specific wavelengths that maximally excite infrared sensory terminals and therefore that might be correlated with the absorbance spectra of candidate phototransduction molecules are unknown at present. Infrared spectrometry experiments on intact pit organs in pythons show that the pit organs absorb in the ranges of 3-5 Ixm and 8-12 Ixm significantly more than do the surrounding tissues (Grace, et al., 1999a). Further, the epidermal tissue overlying the IR receptor nerve terminals is essentially transparent to windows of the IR spectra centered at 4 pm and 10/am. These results correlate with the peak emission (approximately 10 pm) of endothermic prey normally targeted by these snakes. However, the findings are not sufficiently precise to define any candidate
4Bullock and Diecke (1956) and de Cock Buning et al. (1981) alike point out that when factors such as the relative thickness of the overlying tissues, the low heat capacity of the thin pit organ membrane, and the density of innervation are taken into account, the threshold for snake pit organs is similarto that of other warm receptors, including human cutaneous thermoreceptors. One reason for the apparent higher sensitivity to small temperature increases in snakes is that the density of innervation in pit organs is very high and the epithelial tissue overlying the sensory nerve terminals is extremely thin. Thus, infrared radiation can readily and rapidly increase the temperature around a dense plexus of sensory nerve terminals.
photoreceptive molecule(s) that absorb in a defined infrared action spectrum. The argument against the notion that a photochemical reaction underlies infrared sensory reception is primarily that photons in the effective spectral region (~ 10 ]am wavelength) have insufficient energy to produce molecular photoisomerization of any known transduction molecules. This may be true for retinaldehydes, which together with opsins form visual pigments. However, other molecular structures can be efficiently isomerized by infrared radiation, including wavelengths in the 10-pm region. For example, highly branched hydrocarbon dendrimers absorb infrared photons, leading to photoisomerization of an azobenzene core (Jiang and Aida, 1997). In principle, a similar mechanism could allow efficient infrared phototransduction in snake or other infrared receptors. More likely, however, is the hypothesis that infrared receptors function as thermoreceptors. Indeed, infrared pit organ receptors are commonly termed heat receptors in the literature. This would imply that the sensory terminals are stimulated by all wavelengths of infrared radiation absorbed by the pit organ in proportion to the heat they generate in the tissues. The evidence that pit organs are temperature receptors includes the observation that responses to infrared stimulation can be mimicked by infusing warm water directly into the pit organ. More importantly, it is possible to calculate how much the temperature in the tissues surrounding the sensory terminals increases when infrared radiation at threshold intensity for eliciting responses is focused in the pit organ. This calculated temperature agrees well with the observed threshold temperature change determined experimentally by perfusing water of specified temperature into the pit organ (Bullock and Diecke, 1956). Little is known about heat transduction in thermoreceptive sensory fibers. The only known thermal transduction molecules in animals are VR1, an ion channel that is activated by moderate temperatures (~43 ~ Caterina et al., 1997), and
49. Gustatory and Olfactory Sensory Transduction
VRL-1, a related heat-activated channel that responds to higher temperatures (52 ~ Caterina et al., 1999). Low pH (acid) and capsaicin also stimulate VR1. VR1 is found in small-diameter polymodal nociceptor sensory nerve terminals. There is no indication that VR1 or VRL-1 are expressed in infrared sensory neurons, although this has not been specifically tested to date. However, given the high specificity of effective stimuli that excite infrared receptors and the extreme sensitivity of pit organs to very small thermal changes, it is unlikely that a polymodal, heat-transducing molecule such as VR1 or a high-temperature receptor like VRL-1 is involved.
Bibliography Amemiya, E, Nakano, M., Goris, R. C., Kadota, T., Atobe, Y., Funakoshi, K., Hibiya, K., and Kishida, R. (1999). Microvasculature of crotaline snake pit organs: possible function as a heat exchange mechanism. Anat. Rec. 254, 107-115. Bullock, T. H., and Diecke, E P. J. (1956). Properties of an infra-red receptor. J. Physiol. 134, 47-87. Caterina, M. J., Rosen, T. A., Tominaga, M., Brake, A. J., and Julius, D. (1999). A capsaicin-receptor homologue with a high threshold for noxious heat. Nature 398, 436-441. Caterina, M. J., Schumacher, M. A., Tominaga, M., Rosen, T. A., Levine, J. D., and Julius, D. (1997). The capsaicin receptor: a heatactivated ion channel in the pain pathway. Nature 389, 816-824. de Cock Buning, T., Terashima, S., and Goris, R. C. (1981). Crotaline pit organs analyzed as warm receptors. Cell. Mol. Neurobiol. 1, 69-85. Dickman, J. D., Colton, J. S., Chiszar, D., and Colton, C. A. (1987). Trigeminal responses to thermal stimulation of the oral cavity in rat-
837 tlesnakes (Crotalus viridis) before and after bilateral anesthetization of the facial pit organs. Brain Res. 400, 365-370. Gamow, R. I., and Harris, J. S. (1973). The infrared receptors of snakes. Sci. Am. 228, 94-100. Grace, M. S., Church, D. R., Kelley, C., and Cooper, T. M. (1999a). The python pit organ: immunocytochemical and imaging analysis of a sensitive natural infrared detector. Biosens. Bioelectron. 14, 53-59. Grace, M. S., Woodward, O. M., Church, D. R., and Calisch, G. (2001). Prey targeting by the infrared-imaging snake python: effects of experimental and congenital visual deprivation. Behav. Brain Res. 119, 23-31. Hartline, P. H. (1999). Infrared sense. In "Encyclopedia of Neuroscience" (G. Adelman and B. H. Smith, Eds.). Elsevier, New York. pp. 957-959. Hartline, P. H., Kass, L., and Loop, M. S. (1978). Merging of modalities in the optic tectum: infrared and visual integration in rattlesnakes. Science 199, 1225-1229. Jiang, D.-L., and Aida, T. (1997). Photoisomerization in dendrimers by harvesting of low energy photons. Nature 388, 454-456. Kardong, K., and Mackessey, G. (1991). The strike behavior of a congenitally blind rattlesnake. J. Herpetol. 25, 208-211. Molenaar, G. J. (1992). Anatomy and physiology of infrared sensitivity of snakes. In "Biology of the Reptilia" (C. Gans and E S. Ulinski, Eds.). University of Chicago Press, Chicago. pp. 367--453. Newman, E. A., and Hartline, E H. (1982). The infrared "vision" of snakes. Sci. Am. 246, 116-127. Stanford, L. R., Schroeder, D. M., and Hartline E H. (1981). The ascending projection of the nucleus of the lateral descending trigeminal tract: a nucleus in the infrared system of the rattlesnake (Crotalus viridis). J Comp. Neurol. 201, 161-174. Terashima, S., Goris, R. C., and Katsuki, Y. (1970). Structure of warm fiber terminals in the pit membrane of vipers. J. Ultrastruct. Res. 31, 494-506
John G. New and Timothy C. Tricas
Electroreceptors and Magnetoreceptors
I. I n t r o d u c t i o n That certain fishes could produce potent shocks has been known since antiquity, and shortly after the discovery of electricity it was determined that the shock from these animals was electrical in nature (see Wu, 1984). It was not, however, until the discovery in the 1950s that some fishes produce a much weaker electrical discharge, in the millivolt range, that the use of electrical signals for purposes other than attack or defense was considered (Lissman, 1958; Lissman and Machin, 1958). Because these weak electrical discharges offered a possible channel for communication among conspecifics, the necessity arose to determine the identity of the sensory receptors that detected them, and thus was born the study of electroreception. It has since been discovered that electroreception is more widespread among vertebrates than was initially thought, and an electrosense is possessed by many animals that lack electric discharge organs as well as by those that possess them. Electroreception is a primitive vertebrate character, found in the common ancestor of jawless and jawed vertebrates (Bullock et al., 1983). This primitive electrosense is retained in numerous extant taxa (Fig. 1) including lampreys, chondrichthyan fishes, bichirs, sturgeon and paddlefishes, lungfishes, coelacanths, and nonanuran amphibians (salamanders and caecilians). Phylogenetic analysis of character traits indicates that the electrosenses of all of these animals are homologous, reflecting their common phylogenetic origin (Bullock et al., 1983; Bullock and Heiligenberg, 1986). Electroreception, however, has been largely lost in the teleost fish, the lineage to which most modem bony fishes belong. Most teleosts, together with their sister groups of gars and bowfin, which collectively comprise the Neopterygii, lack an electrosense. Remarkably, only two distantly related lineages of teleost fishes do possess electroreceptors, and it is apparent from character analysis that these have evolved independently. In addition, it has been recently discovered that two species of monotreme mammal, the platypus and the echidna, are also electroreceptive. Because electrosenses are not present in most other tetrapod
Cell Physiology Sourcebook: A Molecular Approach, Third Edition
839
lineages, this clearly represents yet another "reinvention" of electroreception (Scheich et al., 1986; Gregory et al., 1989). Electroreceptors in anamniotic vertebrates display clear affinities with the receptors of the mechanosensory lateral line. Both types of receptors probably develop from the same embryonic ectodermal placodes (Northcutt et al., 1995; Vischer, 1995), are innervated by the lateral line nerves, and possess either a kinocilium, microvilli, or both. In animals with primitive electrosensory systems, it is not known whether the electroreceptors are derived directly from lateral line receptors, or whether both were derived from a primitive mechanosensory cell (Bodznick, 1989; JCrgensen, 1989; Bullock and Heiligenberg, 1986). In teleosts, it is evident that electroreception has evolved as a specialization of the lateral line system. In contrast, the electrosense of mammals has evolved as a specialization of the trigeminal cranial nerve system since amniotes have lost the lateral line concurrent with the evolution of reproductive life histories that are not of necessity linked to aquatic environments (Andres and von During, 1988; Manger and Hughes, 1992). All electroreceptors, whether primitive or derived, can be broadly classified as belonging to one of two categories: ampullary or tuberous (see Zakon, 1986a, 1988). Ampullary receptors are broadly tuned and respond to low-
Craniates - ~
Myxinee Pteraspidomorpha
! I
Vertebrates I
[
Myopterygii [
]
Gnathostomes [
Chondrichthyes * I
Osteichthyes I 9 Eiectroreceptive taxa Extinct taxa - Lacks electroreception e
Petromyzontidae *
Actinopterygii */I
Sarcopterygii I
Actinistia * t ]
f |
Dipnoans * Rhipidistia e
k__... Tetrapods*/FIGURE 1. Cladogram illustrating the distribution of electroreception among craniates. Electroreception in tetrapods is found only in nonanuran amphibians and monotreme mammals.
Copyright 9 2001 by Academic Press. All rights of reproduction in any form reserved.
840
SECTION V Synaptic Transmissions and Sensory Transductions
frequency stimuli, with best frequencies generally between 0.1 and 20 Hz. This range of frequencies encompasses most bioelectric fields produced by aquatic organisms as well as those produced by nonliving physicochemical sources in aquatic environments (Peters and Bretschneider, 1972). Detection of these extrinsic sources by ampullary receptors is often referred to as electroreception in the passive mode (Kalmijn, 1974,1988). Tuberous electroreceptors are tuned to much higher frequencies, with best frequencies in the 0.1-1.0 kHz range. Tuberous organs are found only in fishes possessing weak electric organs of the sort initially described by Lissman (1958), and these receptors have best frequencies that correspond closely to the peak spectral frequency of the discharge of the animal's electric organ. The tuberous organs detect changes in the intensity and temporal pattern of the electric field produced across the body by the electric organ discharge (EOD), whether by the presence of items in the field of objects with different conductances than the water, or by the addition of the electric organ discharges of another individual. Such forms of electrodetection responding to alterations of the standing EOD are referred to as electroreception in the active mode (Kalmijn, 1988). In this chapter we consider the morphology and physiology of both ampullary and tuberous electroreceptors. Space limitations preclude a discussion of the central processing of electrosensory information, which has become one of the most engrossing stories in modem neuroethology. The reader is referred to several excellent recent reviews on this topic (Bullock and Heiligenberg, 1986; Zakon, 1988; Heiligenberg, 1991). The ability to detect geomagnetic fields directly, or magnetoreception, has been described in a number of different taxa (for review, see Tenforde, 1989). These studies have been primarily behavioral; no studies exist concerning the physiology of magnetoreceptors or magnetoreception, and even the behavioral studies have been difficult to replicate or confirm (Moore, 1988). The direct identification of a magnetoreceptor organ in these taxa has itself been problematic; such organs are believed to be associated with localized deposits of magnetite (Fe304). Localized magnetite domains have been described in bees (Gould et al., 1978), salmon (Kirschvink et al., 1985), tuna (Walker et al., 1984), turtles (Perry et al., 1981), pigeons (Walcott et al., 1979; Preston and Pettigrew, 1980), dolphins (Zoeger et al., 1981), and humans (Baker et al., 1983), but the organs, transduction mechanisms, and innervation have yet to be positively identified. A recent description of a putative magnetoreceptor in honeybees (Hsu and Li, 1994) has been disputed (Nesson, 1995; Nichol and Locke, 1995; Kirschvink and Walker, 1995). The exception is the description of magnetotactic behavior in magnetite-containing bacteria (Blakemore, 1975; Kalmijn and Blakemore, 1978). Magnetoreception, however, is believed to be possible in some electroreceptive taxa via the detection of electric fields induced by an animal's movement through a geomagnetic field (see discussion later).
I!. Ampullary Electroreceptors The wide phylogenetic distribution of ampullary electroreceptors among extant vertebrate taxa suggests that this
class of electroreceptor has served important biological functions for hundreds of millions of years and has subsequently "re-evolved" several times (Fig. 1; see also Fig. 8). With the exception of weakly-electric fishes, which also possess tuberous electroreceptors, most species with ampullary electroreceptors lack electric organs. Thus, behaviorally relevant electric field stimuli for most species are thought to originate primarily from extrinsic sources. Ampullary electroreceptors are known to be important for the detection of prey (Kalmijn, 1971; Peters and Meek, 1973; Tricas, 1982), mates (Tricas et al., 1995), and orientation to local inanimate electric fields (Kalmijn, 1982; Pals et al., 1982). In addition, the ampullary electroreceptor system is theoretically capable of mediating navigation by detecting electric fields induced by movement of the animal through the earth's magnetic field (Kalmijn, 1974, 1982; Paulin, 1995), which would represent a form of electroreception in the active mode, as discussed earlier (Kalmijn, 1988).
A. Morphology The ampullary electroreceptor organ in lampreys is known as an end bud (Fig. 2) and differs considerably in morphology from the ampullary electroreceptors found in cartilaginous and teleost fishes (Ronan and Bodznick, 1986). Each end bud consists of numerous support cells and 3 to 25 sensory cells in the epidermis that are in direct contact with the surrounding water. Individual receptor cells have numerous small microvilli on the apical surface but lack a kinocilium. Small groups or lines of end buds are distributed over the head and body surface with multiple buds being innervated by a single sensory lateral line nerve fiber (Ronan, 1986; Bodznick and Preston, 1983). Little is known about transduction in end buds, but the similarity of their responses to primitive gnathostome electroreceptors supports a homologous transduction mechanism (see discussion later). It is not known whether end buds represent the primitive electroreceptor state or whether they are a derived condition unique to the lampreys, which first appeared in their current form in the Carboniferous era considerably after the appearance of jawed fishes (Moy-Thomas and Miles, 1971). The anatomy and physiology of primitive ampullary electroreceptor organs have been studied most extensively in the chondrichthyan fishes. Chondrichthyan fishes include the elasmobranchs (rays, skates, and sharks) and the rat fishes (Fields et al., 1993), all of which possess ampullary electroreceptor organs of a similar morphology. In elasmobranch fishes, the electroreceptive unit is a prominent and highly specialized structure known as an ampulla of Lorenzini (Fig. 3). The ampulla proper in the marine skate is composed of multiple alveolar sacs, which share a common lumen (Waltman, 1966). The apex of each ampulla chamber is connected by a highly insulated marginal zone to a single subdermal canal, which is approximately 1 mm in diameter and terminates as a small epidermal pore. The canal wall is 1-2 lam thick and composed of two layers of flattened epithelial cells, which are separated by a basement membrane to which the lumenal layer is also united by tight junctions. Both the canal lumen and the ampullary chambers are filled with a K§ mucopolysaccharide, jelly-like matrix that is secreted by the superficial layer and has an ex-
50. Electroreceptors and Magnetoreceptors
841
FIGURE 2. Light photomicrograph of the electrosensory end bud of the silver lamprey. (From Ronan 1986. Copyright 9 1986 John Wiley & Sons. Reprinted by permission of John Wiley & Sons, Inc.)
tremely low resistivity (Murray and Potts, 1961; Waltman, 1966). These remarkable morphological and cytological features make each canal a low-resistance core conductor that connects the internal ampullary lumen with the surface pore. The sensory epithelium within the alveolus is composed of two cell types, which form a monolayer that is approximately 15 lam thick (Fig. 4). The vast majority of the alveolar surface is formed by accessory cells that are highly resistive to transmembrane currents and are bound together by tight junctions that prevent ionic leakage across the lumenal and basal surfaces of the epithelium. Interspersed among the
accessory cells are flask-shaped receptor cells (thought to be modified hair cells), which possess a single kinocilium on the apical surface and lack microvilli. This physical arrangement results in only a small fraction of the receptor cell surface being exposed to the ampullary chamber. The basal membrane surface of the receptor cell forms a ridge seated in a postsynaptic invagination that is separated by a distance of 100-200 A (Waltmann, 1966). A synaptic ribbon about 250 A wide and 2 pm in length is located within the presynaptic ridge. A single layer of synaptic vesicles covers the ribbon, and exocytotic release of chemical
842
SECTION V Synaptic Transmissions and Sensory Transductions
CANAL ".._..~...
CE PAT
MZ
AE
PAN
F I G U R E 3. Ampulla of Lorenzini from the marine skate, Raja. The ampulla proper consists of multiple alveoli formed by the alveolar epithelium (AE). A high-resistance marginal zone (MZ) connects the sensory walls of the ampulla to the high-resistance canal epithelium (CE), which projects to the surface of the skin and terminates as a small pore confluent with the surrounding water. The ampulla lumen (LU) and canal are filled with a highly conductive jelly, which makes the lumen isopotential with voltages present at the pore. Myelinated primary afferent neurons (PANs) innervate the base of the ampullae and their unmyelinated primary afferent terminals (PATs) receive chemical excitation from the basal region of the sensory cells in the epithelial layer. (Modified from Waltmann, 1966.)
neurotransmitter contained within these vesicles depolarizes the postsynaptic membrane of the innervating fibers of the anterior lateral line nerves. Unlike the hair cell receptors of the mechanosensory lateral line and octaval systems, all ampullary electroreceptors, both primitive and derived, lack efferent innervation. Chondrichthyan fishes typically possess hundreds of ampullae that are associated in specific regional clusters that are closely bound by a dense matrix of connective tissue (Fig. 5A). From these clusters the subdermal canals radiate omnidirectionally and terminate in surface pores on the head, and on the enlarged pectoral disk of rays and skates. The multiple orientations of the receptor canals, and the copious distribution of the ampullary pores over the cephalic surface provide an extensive array of receptors with a high degree of spatial resolution. The morphology of the ampullary electroreceptors in freshwater elasmobranchs is thought to reflect sensory adaptations to their highly resistive environment (Kalmijn, 1974, 1982; Raschi and Mackanos, 1989). The freshwater rays,
F I G U R E 4. Receptor cell of the skate ampulla of Lorenzini. (A) Photomicrograph of flask-shaped receptor cells (R) and adjacent accessory cells (A) that are united by tight junctions to form the alveolar epithelium. A single kinocilium projects from each receptor cell into the lumen and, together with a small portion of the apical surface, is exposed to electric stimuli. Primary afferent neurons (n) innervate the basal portion of the receptors. The basement membrane (bm) and fibroblast cells (f) lie beneath the sensory epithelium. (From Waltmann, 1966.) (B) Diagrammatic representation of the receptor cell illustrating current flow during excitation. In the case of the elasmobranch, a cathodal (negative) stimulus relative to the basal region of the receptor excites the apical surface of the cell, which causes an increase in outward current flow (arrows). This in turn causes release of chemical transmitter at the cell synapse and excitation of the primary afferent. Anodal potentials in the lumen will decrease outward current flow at the apical surface and subsequently decrease the rate of transmitter release. (Reproduced from The Journal of General Physiology, 1972, vol. 60, pp. 534-557, by copyright permission of The Rockefeller University Press.)
Pomatotrygon and Dasyatis garouaensis, have a hypertrophied, thick epidermis function to increase transcutaneous electrical resistance. The ampullary electroreceptors are greatly reduced in size and are referred to as r n i n i a m p u l l a e or m i c r o a m p u U a e , which are distributed individually across the skin rather than in clusters, and which have very
50. E l e c t r o r e c e p t o r s
A
and Magnetoreceptors
I
DORSAL /~,~X
843
VENTRAL
..'::~ \ 2.,
"
",,
,
r/
\
.......- ...y..'j"
"-.....
:";".:",. :,:'.,e;,-.~ ~ . .. ~ ~ ' i !
,
"i~,, "~.-.
"
...7/. . . .
t ~'(,
: ,.
~. ,.',,
2.L..-~_
~,/
;-/
)?~J i
i
...,t't'-
~
sl"
'~':' -"
"
DORSAL
~":'~'.
0
LATERAL
) VENTRAL
I
FIGURE 5. Distribution of ampullary receptors on nonteleost fishes. (A) In the skate, Raja laevis, ampullary canals originate at specific regional clusters and radiate in multiple directions. These canals function to measure voltage gradients that pass across the body of the animal. (Adapted from Raschi, 1986). (B) In the sturgeon, Scaphirhynchus, ampullary canals are extremely short and are not arranged in clusters. Instead they measure transcutaneous potential relative to the internal potential of the animal. (From Northcutt, 1986. Copyright 9 1986 John Wiley & Sons. Reprinted by permission of John Wiley & Sons, Inc.) short canals (about 0.3-2.1 mm long) that traverse the integument. The anatomy and organization of ampullary electroreceptor organs in other nonteleost fishes and amphibians are generally similar to those of elasmobranch fishes, with which
they are believed to behomologous. Ampullary electroreceptors in chondrostean (sturgeon and paddlefishes), cladistian (bichirs), and dipnoan fishes (lungfishes) share in most respects a similar morphology among alveoli, canals, and ampullary pores. However, the ampullary organs in these fishes are most commonly arranged as single units (as opposed to large clusters) and are located immediately beneath the skin with very short (generally E
O
-10
O O t'~
m -~
113
-20
"~
"0 0 t'm
x:
F-
-10
O O
-20
"0 0
,-I---
-30
-40 - I . . . . . 40
1
100
l
1
300
, I ,~1
1000
~
I
3000
-30
-40
1 t ~tttl
40
100
t
.
lit.J,!
1000
FIGURE 14. (A, B) EOD waveforms and power spectra and (C, D) corresponding frequency response functions of tuberous electroreceptors from two species of the gymnotiform Hypopomus. Tuning curves with open symbols are from pulse markers (M units); those with closed symbols are from burst duration coders (B units). (From Bastian, 1977.)
Knollenorgans possess a small number of receptor cells (1 to 10), although larger numbers are reported in some species. Each of the large (40-50 lam in diameter) spherical receptor cells itself is individually encapsulated by a layer of covering cells within the larger capsule of the organ. The receptor cells are attached to the walls of the capsule only at their basal membrane surface, leaving most of the receptor membrane exposed to the lumen within the covering cell capsule. The apical surface of the cell membrane is densely supplied with microvilli, and numerous mitochondria are apparent just beneath the apical membrane. Unlike most teleost electroreceptors, the membrane of Knollenorgans is believed to be electrically excitable: Stimulation of the afferent nerve fiber results in a spike that can be recorded from the organ. Although it has been suggested that an electrotonic synapse exists between Knollenorgan and afferent fiber (Bennett, 197 la; Steinbach and Bennett, 1971), only chemical synapses have been observed (Derbin and Szabo, 1968; Schessel et al., 1982). Although a synaptic ribbon is present,
it does not extend into an evagination of the presynaptic membrane as is the case in gymnotids. Mormyromast electroreceptors (D units) are more complex organs, with two distinct populations of electroreceptor cells, each innervated by different afferent nerve fibers (Szabo and Wers~ill, 1970). Mormyromast receptors consist of two chambers; a superficial chamber that resembles an ampullary receptor and a deep chamber that is similar in morphology to a Knollenorgan. The deep chamber (or sensory chamber) possesses three to five small (2-3 pm) receptor cells (B cells). The receptor cells are covered with microvilli and, as in other tuberous organs, are attached to the capsule only at the basal membrane surface. A narrow canal connects the deep chamber to the superficial chamber, which is filled with a jelly-like matrix. The five to seven receptor cells of the superficial chamber (A cells) are embedded within the walls of the receptors with only the apical membrane exposed to the chamber lumen in a manner similar to ampullary electroreceptors. These cells lack
852 microvilli, but possess a curious small dense conical structure on the apical membrane. Each of the receptors in the upper chamber is innervated by one of two to four nerve fibers, each of which innervates a series of adjacent cells. Tuberous electroreceptor organs in gymnarchids are quite different from those of their mormyrid sister group. Gymnarehomast receptors are divided into two morphologically distinct types (Fig. 11D, E). In both types of organs, the cell bodies of the receptors are embedded in the capsule walls with the apical face exposed to the jelly-like matrix of the lumen. Gymnarehomast type I organs possess two distinct receptor cell morphologies, usually with one or more pairs of each type per organ. One type of receptor cell possesses a deep, groove-like invagination of the apical membrane surface, whereas the other possesses only a shallow indentation. Both types of receptor cells have apical microvilli, and all of the receptors within an organ are innervated by a single afferent nerve fiber. Gymnarehomast type II organs possess a single type of large receptor cell, with a characteristic deep invagination of the apical membrane and long apical microvilli. All of the sensory cells (7 to 12) in a type II receptor are innervated by a single afferent fiber (Szabo, 1965, 1974). Ultrastructural studies have demonstrated relatively high levels of Ca 2+ in the cytoplasm of the receptor cells (Djebar et al., 1995) presumably as a result of the inward depolarizing current produced at the membrane basal face.
B. Physiology As in gymnotids, the tuberous electroreceptors of mormyrids can be differentiated into rapid-timing units and amplitude-coding units. The broadly tuned Knollenorgans respond to presentation of EOD-like stimuli by generating a single, highly phase-locked spike per pulse (Bennett, 1967; Szabo, 1974; Hopkins and Bass, 1981). The temporal pattern of the EOD pulse is critical to the receptor response, and only stimuli consisting of positive-negative transitions with rise-fall profiles similar to the animal's own EOD will reliably elicit a spike: Phase-shifted stimuli with power spectral peaks similar to the EOD are also poor at eliciting receptor responses (Hopkins and Bass, 1981). The Knollenorgans play a functional role in the communicative behavior of mormyrid fishes by detecting EODs produced by other conspecific individuals, since responses of Knollenorgans to the animal's own EOD are suppressed in the central nervous system. Hopkins and Bass (1981) suggested that different classes of Knollenorgan (KI and KII) may respond to different portions of the EOD cycle; the interval between the firing of the two receptor types will thus be specific for a given EOD waveform forming a basis for conspecific recognition at the level of the receptor. Mormyromast receptors respond to the amplitude of the EOD, producing afferent fiber spikes as a function of EOD intensity. Unlike the B units of gymnotiforms, mormyromasts typically generate fewer spikes (1 to 5) per stimulus, and both response latency and interspike intervals shift with increasing stimulus amplitude (Szabo and Hagiwara, 1967). The threshold of mormyromast receptors is relatively high, indicating that these receptors are stimulated primarily by the animal's own EOD, and very little by the attenuated EODs of con-
SECTION V Synaptic Transmissions and Sensory Transductions
specifics at a distance. Alteration of the electric field produced by the EOD via presentation of items of varying conductances near the fish's skin produces systematic shifts in spike latency and number in afferents innervating mormyromasts. Afferent nerve fibers innervating either of the two populations of receptor cells in the mormyromast organ project to different regions of the first-order medullary electrosensory nucleus (Bell et al., 1989) and exhibit differing responses to presentation of EOD stimuli. Fibers innervating A cells exhibit higher thresholds and smaller numbers of maximum spikes per burst per stimulus, whereas those innervating B cells have lower thresholds and higher numbers of maximum spikes per burst (Bell, 1990a,b). The dynamic range of the mormyromast response intensity function is relatively limited (6 dB) suggesting that these cells are very sensitive to slight modulations of EOD amplitude within that range (Szabo and Hagiwara, 1967). The properties of the mormyromast response to EOD stimulation suggest that EOD amplitude is coded by a latency code, in which the intensity of the EOD is encoded by the latency of the first spike of the response, rather than by a burst duration code, as in the B units of gymnotids, in which the amplitude is encoded by the number of spikes generated in the burst (Szabo and Hagiwara, 1967; Bell, 1990b). Mormyromasts also exhibit extreme sensitivity to EOD phase modulation (vonder Emde and Bleckmann, 1992). The different behavioral functions of Knollenorgans and mormyromasts (namely, communication versus electrolocation) in the ethogram of mormyrids is mainmined throughout the central nervous system, and information from the two receptor types is segregated into two parallel pathways through the neuraxis. The gymnarchomast receptors show striking similarities to those of gymnotids producing wave-type EODs, exhibiting either rapid-timing (S units) or amplitude-coding (O units) responses to presentation of EOD-like stimuli. S units have a low threshold and narrow dynamic range. At "normal" EOD intensities, the saturated S units show little temporal variation in their response to changes in stimulus intensity. O units have higher thresholds and dynamic ranges, and changes in EOD amplitudes are encoded as rapidly adapting bursts of afferent fiber spikes (Bullock et al., 1975).
Bibliography Akoev, G. N., Ilyinsky, O. B., and Zadan, E M. (1976a). Physiological properties of electroreceptors of marine skates. Comp. Biochem. Physiol. 53A, 201-209. Akoev, G. N., Ilyinsky, O. B., and Zadan, E M. (1976b). Responses of electroreceptors (ampullae of Lorenzini) of skates to electric and magnetic fields. J. Comp. Physiol. 106A, 127-136. Andres, K. J., and von During, M. (1988). The platypus bill. A structural and functional model of a pattern-like arrangement of cutaneous receptor receptors. In "Sensory Receptor Mechanisms" (A. Iggo, Ed.), pp. 81-89. World Scientific Publishing Company, Singapore. Andrianov, G. N., Broun, G. R., and Ilyinsky, O. B. (1984). Frequency characteristic of skate electroreceptive central neurons responding to electrical and magnetic stimulation. Neurophysiol. 16, 365-376. Baker, C. (1980). Jamming avoidance behavior in gymnotoid electric fish with pulse-type discharges: Sensory encoding for a temporal pattern discrimination. J. Comp. Physiol. 136, 165-181.
50. Electroreceptors and Magnetoreceptors Baker, R. R., Mather, J. G., and Kennaugh, J. H. (1983). Magnetic bones in human sinuses. Nature 301, 78-80. Bass, A. H. (1986). Electric organs revisited: Evolution of a vertebrate communication and orientation organ. In "Electroreception" (T. H. Bullock and W. Heiligenberg, Eds.), pp. 13-70. John Wiley and Sons, New York. Bass, A. H., and Hopkins, C. D. (1983). Hormonal control of sexual differentiation: Changes in electric organ discharge waveform. Science 220, 971-974. Bass, A. H., and Hopkins, C. D. (1984). Shifts of frequency tuning in electroreceptors in androgen treated mormyrid fish. J. Comp. Physiol. 155, 713-724. Bastian, J. (1976). Frequency response characteristics of electroreceptors ill weakly electric fish (Gymnotoidei) with a pulse discharge. J. Comp. Physiol. 112, 165-180. Bastian, J. (1977). Variations in the frequency response of electroreceptors dependent on receptor location in weakly electric fish (Gymnotoidei) with a pulse discharge. J. Comp. Physiol. 121, 53-64. Bastian, J., and Heiligenberg, W. (1980). Neural correlates of the jamming avoidance response in Eigenmannia. J. Comp. Physiol. 136, 135-152. Bell, C. C. (1990a). Mormyromast receptor organs and their afferent fibers in mormyrid fish. II. Intra-axonal recordings show initial stages of central processing. J. Neurophys. 43, 303-318. Bell, C. C. (1990b). Mormyromast receptor organs and their afferent fibers in mormyrid fish III. Physiological differences between two morphological types of fibers. J. Neurophys. 43, 319-332. Bell, C. C., Zakon, H., and T. E. Finger (1989). Mormyromast receptor organs and their afferent fibers in mormyrid fish. I. Morphology. J. Comp. Neurol. 286, 391-407. Bemis, W. E., and Hetherington, T. E. (1982). The rostral organ of Latimeria chalumnae: Morphological evidence of an electroreceptive function. Copeia 1982, 467-471. Bennett, M. V. L. (1967). Mechanisms of electroreception. In "Lateral Line Detectors" (P. Cahn, Ed.), pp. 313-393. Indiana University Press, Bloomington, IN. Bennett, M. V. L. (197 l a). Electroreception. In "Fish Physiology" (W. S. Hoar and D. S. Randall, Eds.), Vol. 5, pp. 493-574. Academic Press, New York. Bennett, M. V. L. (1971 b). Electric organs. In "Fish Physiology" (W. S. Hoar and D. S. Randall, Eds.), Vol. 5, pp. 347-491. Academic Press, New York. Bennett, M. V. L., and Clusin, W. T. (1978). Physiology of the ampulla of Lorenzini, the electroreceptor of elasmobranchs. In "Sensory Biology of Sharks, Skates, and Rays" (E. S. Hodgson and R. E Mathewson, Eds.), pp. 483-505. Office of Naval Research, Arlington, VA. Bennett, M. V. L., and Obara, S. (1986). Ionic mechanisms and pharmacology of electroreceptors. In "Electroreception" (T. H. Bullock and W. Heiligenberg, Eds.), pp. 157-181. John Wiley and Sons, New York. Blakemore, R. (1975). Magnetotactic bacteria. Science 190, 377-379. Bodznick, D. (1989). Comparisons between electrosensory and mechanosensory lateral line systems. In "The Mechanosensory Lateral Line: Neurobiology and Evolution" (S. Coombs, E G6rner, and H. Mtinz, Eds.), pp. 655-680. Springer-Verlag, New York. Bodznick, D., and Montgomery, J. C. (1992). Reafference in the elasmobranch electrosensory system: Medullary neuron receptive fields support a common-mode rejection mechanism. J. Exp. Biol. 171, 127-137. Bodznick, D., Hjelmstad, G., and Bennett, M. V. L. (1993). Accommodation to maintained stimuli in the ampullae of Lorenzini: How an electroreceptive fish achieves sensitivity in a noisy world. Japan J. Physiol. 43, Suppl. 1, $231-$237.
853 Bodznick, D., Montgomery, J. C., and Bradley, D. J. (1992). Suppression of common-mode signals within the electrosensory system of the little skate, Raja erinacea. J. Exp. Biol. 171, 107-125. Bodznick, D., and Preston, D. G. (1983). Physiological characterization of electroreceptors in the lampreys, Ichthyomyzon unicuspis and Petromyzon marinus. J. Comp. Physiol. 152, 209-217. Braford, M. R. (1986). African knifefishes: The Xenomystinae. In "Electroreception" (T. H. Bullock and W. Heiligenberg, Eds.), pp. 453-464. John Wiley and Sons, New York. Bullock, T. H., Behrend, K., and Heiligenberg, W. (1975). Comparison of the jamming avoidance response in gymnotoid and gymnarchid electric fish: A case of convergent evolution of behavior and its sensory basis. J. Comp. Physiol. 103, 97-121. Bullock, T. H., Bodznick, D. A., and Northcutt, R. G. (1983). The phylogenetic distribution of electroreception: Evidence for convergent evolution of a primitive vertebrate sense modality. Brain Res. Rev. 6, 25-46. Bullock, T. H., and Heiligenberg, W. (1986). "Electroreception." John Wiley and Sons, New York. Derbin, C., and Szabo, T. (1968). Ultrastructure of an electroreceptor (Knollenorgan) in the mormyrid fish, Gnathonemus petersii Int. J. Ultrastruct. Res. 22, 469--484. Dijkgraaf, S., and Kalmijn, A. J. (1966). Versuche zur biologischen Bedeutung der Lorenzinischen Ampullen an Haifischen. Z. Vgl. Physiol. 47, 438-456. Djebar, B., Bensouilah, M., and Denizot, J.-E (1995). Ultrastructural distribution of calcium in cutaneous electroreceptor organs of teleost fish. Biotech. Histochem. 70, 81-89. Echague, A., and Trujillo-Cenoz, O. (1981). Innervation patterns in the tuberous organs of Gymnotus carapo. In "Advances in Physiological Sciences," Vol 31: "Sensory Physiology of Aquatic Lower Vertebrates" (T. Szabo and G. Czeh, Eds.), pp. 29-40. Akademiai Kiado, Budapest. Feng, A. S., and Bullock, T. H. (1977). Neuronal mechanisms for object discrimination in the weakly electri fish, Eigennmania viriscens. J. Exp. Biol. 66, 141-158. Fields, R. D., Bullock, T. H., and Lange, O. D. (1993). Ampullary sense organs, peripheral, central and behavioral electroreception in chimeras (Hydrolagus, Holocephali, Chondrichthyes). Brain Behav. Evol. 41, 269-289. Fritzsch, B., and Mtinz, H. (1986). Electroreception in amphibians. In "Electroreception" (T. H. Bullock and W. Heiligenberg, Eds.), pp. 483-496. John Wiley and Sons, New York. Gould, J. L., Kirschvink, J. L., and Deffeyes, K. E (1978). Bees have magnetic resonance. Science 201, 1026-1028. Gregory, J. E., Iggo, A., McIntyre, A. K., and Proske, U. (1989). Responses of electroreceptors in the snout of the echidna. J. Physiol. (Lond.) 414, 521-538. Hagedorn, M., and Heiligenberg, W. (1985). Court and spark: Electric signals in the courtship and mating of gymnotoid fish. Anim. Behav. 33, 254-265. Hagiwara, S., Kusano, K., and Negishi, K. (1962). Physiological properties of electroreceptors of some gynmotids. J. Neurophys. 25, 430-449. Hagiwara, S., and Morita, H. (1963). Coding mechanisms of electroreceptor fibers in some electric fish. J. Neurophys. 26, 551-567. Hagiwara, S., Szabo, T., and Enger, E S. (1965). Electroreceptor mechanisms in a high-frequency weakly electric fish, Sternarchus albifrons. J. Neurophys. 28, 784-799. Harder, W. (1968). Die Beziehungen zwischen Elektrorezeptoren, elektirischem Organ, Seitenlinienorganen und Nervensystem bei den Mormyridae (Teleostei, Pisces). Z. Vgl. Physiol. 59, 272-318. Heiligenberg, W. (1977). "Principles of Electrolocation and Jamming Avoidance Response in Electric Fish." Springer, New York. Heiligenberg, W. (1986). Jamming avoidance responses. In "Electroreception" (T. H. Bullock and W. Heiligenberg, Eds.), pp. 613-649. John Wiley and Sons, New York.
854 Heiligenberg, W. E (1991). Neural nets in electric fish. MIT Press, Cambridge. Hopkins, C. D. (1972). Sex differences in signaling in an electric fish. Science 176, 1035-1037. Hopkins, C. D. (1974). Electric communication: Functions in the social behavior of Eigenmannia viriscens. Behaviour 50, 270--305. Hopkins, C. D. (1976). Stimulus filtering and electroreception: Tuberous electroreceptors in three species of gymnotoid fish. J. Comp. Physiol. 103, 171-207. Hopkins, C. D. (1983). Functions and mechanisms in electroreception. In "Fish Neurobiology," Vol. 1" "Brain Stem and Sense Organs" (R. G. Northcutt and R. E. Davis, Eds.), pp. 215-259. University of Michigan Press, Ann Arbor, MI. Hopkins, C. D., and Bass, A. H. (1981). Temporal coding of species recognition signals in an electric fish. Science 212, 85-87. Hopkins, C. D., and Heiligenberg, W. E (1978). Evolutionary designs for electric signals and electroreceptors in gymnotoid fishes of Surinam. Behav. Ecol. Sociobiol. 3, 113-134. Hsu, C.-Y., and Li, C.-W. (1994). Magnetoreception in honeybees. Science 265, 95-97. Istenic, L., and Bulog, B. (1984). Some evidence for the ampullary organs in the European cave salamander Proteus anguineus (Urodela Amphibia). Cell Tissue Res. 235, 393-402. JCrgensen, J. M. (1980). The morphology of the Lorenzinian ampullae of the sturgeon Acipenser ruthenus (Pisces: Chondrostei). Acta Zool. 61, 87-92. J0rgensen, J. M. (1982). Fine structure of the ampullary organs of the bichir Polypterus senegalus Cuvier, 1829 (Pisces: Brachiopterygii) with some notes on the phylogenetic development of electroreceptors. Acta Zool. 63, 211-217. J0rgensen, J. M. (1984). On the morphology of the electroreceptors of the Australian and one African lungfish species. Vidensk. Meddr. Dansk Naturh. Foren. 145, 77-85. J0rgensen, J. M. (1989). Evolution of octavolateralis sensory cells. In "The Mechanosensory Lateral Line: Neurobiology and Evolution" (S. Coombs, P. Grmer, and H. Mtinz, Eds.), pp. 115-146. SpringerVerlag, New York. J0rgensen, J. M., Flock, A., and Wersall, J. (1972). The Lorenzinian ampullae of Polyodon spathula. Z. Zellforsch. 130, 362-377. Kalmijn, A. J. (1971). The electric sense of sharks and rays. J. Exp. Biol. 55, 371-383. Kalmijn, A. J. (1974). The detection of electric fields from inanimate and animate sources other than electric organs. In "Handbook of Sensory Physiology" (A. Fessard, Ed.), Vol. 111/3, pp. 147-200. Springer-Verlag, New York. Kalmijn, A. J. (1982). Electric and magnetic field detection in elasmobranch fishes. Science 218, 915-918. Kalmijn, A. J. (1988). Detection of weak electric fields. In "Sensory Biology of Aquatic Animals" (J. Atema, R. R. Fay, A. N. Popper, and W. N. Tavogla, Eds.), pp. 151-186. Springer-Verlag, New York. Kalmijn, A. J., and Blakemore, R. (1978). The magnetic behavior of mud bacteria. In "Animal Migration, Navigation and Homing" (K. Schmidt-Koenig and W. T. Keeton, Eds.), pp. 347-353. SpringerVerlag, New York. Kirschbaum, E (1983). Myogenic electric organ precedes the neurogenic organ in apteronotid fish. Naturwissenschaften 70, 205. Kirschvink, J. L., and Walker, M. M. (1995). Honeybees and magnetoreception. Science 269, 1889. Kirschvink, J. L., Walker, M. M., Chang, S. B., Dizon, A. E., and Peterson, K. A. (1985). Chains of single-domain magnetite particles in chinook salmon. J. Comp. Physiol. A 157, 375-381. Lissman, H. W. (1958). On the function and evolution of electric organs in fish. J. Exp. Biol. 35, 156-191. Lissman, H. W., and Machin, K. E. (1958). The mechanisms of object location in Gymnarchus niloticus and similar fish. J. Exp. Biol. 35, 451-486.
SECTION V Synaptic Transmissions and Sensory Transductions
Manger, R R., and Hughes, R. L. (1992). Ultrastructure and distribution of epidermal sensory receptors in the beak of the echidna, Tachyglossus aculeatus. Brain Behav. Evol. 40, 287-296. McKibben, J. R., Hopkins, C. D., and Yager, D. D. (1993). Directional sensitivity of tuberous electroreceptors: Polarity preferences and frequency tuning. J. Comp. Physiol. 173, 415--424. Meyer, J. H. (1983). Steroid influences upon the discharge frequency of a weakly electric fish. J. Comp. Physiol. 153, 29-38. Meyer, J. H., and Zakon, H. H. (1982). Androgens alter the tuning of electroreceptors. Science 217, 635-637. Millot, J., and Anthony, J. (1956). L'organe rostral de Latimeria (Crossoptrrygien Coelacanthid~). Ann. Sci. Nat. Zool., 11 e Sdrie, 18, 381-387. Montgomery, J. C. (1984). Frequency response characteristics of primary and secondary neurons in the electrosensory system of the thornback ray. Comp. Biochem. Physiol. 79A, 189-195. Montgomery, J. C., and Bodznick, D. (1993). Hindbrain circuitry mediating common-mode suppression of ventilatory reafference in the electrosensory system of the little skate, Raja erinacea. J. Exp. Biol. 183, 203-315. Moore, B. R. (1988). Magnetic fields and orientation in homing pigeons: Experiments of the late W. T. Keeton. Proc. Natl. Acad. Sci. USA 85, 4907-4909. Moy-Thomas, J. A., and Miles, R. S. (1971). "Paleozoic Fishes." W. B. Saunders, Philadelphia. Murray, R. W. (1962). The response of the ampullae of Lorenzini of elasmobranches to electrical stimulation. J. Exp. Biol. 39, 119-128. Murray, R. W., and Potts, T. W. (1961). The composition of the endolymph and other fluids of elasmobranches. Comp. Biochem. Physiol. 2, 65-75. Nachtrieb, H. E (1910). The primitive pores of Polyodon spathula (Walbaum). J. Exp. Zool. 9, 455-468. Nelson, J. S. (1994). "Fishes of the World," 3rd ed. John Wiley and Sons, New York. Nesson, M. H. (1995). Honeybees and magnetoreception. Science 269, 1889-1890. New, J. G. (1990). Medullary electrosensory processing in the little skate. I. Response characteristics of neurons in the dorsal octavolateralis nucleus. J. Comp. Physiol. 167, 285-294. New, J. G., and Bodznick, D. (1990). Medullary electrosensory processing in the little skate. II. Suppression of electrosensory reafference via a common-mode rejection mechanism. J. Comp. Physiol. 167, 295-307. Nichol, H., and Locke, M. (1995). Honeybees and magnetoreception. Science 269, 1888-1889. Northcutt, R. G. (1983). The primary lateral line afferents in lepidosirenid lungfishes. Soc. Neurosci. Abstr. 9, 1167. Northcutt, R. G. (1986). Electroreception in nonteleost bony fishes. In "Electroreception" (T. H. Bullock and W. Heiligenberg, Eds.), pp. 257-285. John Wiley and Sons, New York. Northcutt, R. G., Br~indle, K., and Fritzsch, B. (1995). Electroreceptors and mechanosensory lateral line organs arise from single placodes in axolotls. Dev. Biol. 168, 358-373. Obara, S. (1976). Mechanisms of electroreception in ampullae of Lorenzini of the marine catfish Plotosus. In "Electrobiology of Nerve, Synapse and Muscle" (J. P. Reuben, D. P. Purpura, M. V. L. Bennett, and E. R. Kandel, Eds.), pp. 128-147. Raven Press, New York. Obara, S., and Bennett, M. V. L. (1972). Mode of operation of ampullae of Lorenzini of the skate, Raja. J. Gen. Physiol. 60, 534-557. Pals, N., Valentijn, P., and Verwey, D. (1982). Orientation reactions of the dogfish, Scyliorhinus canicula, to local electric fields. Neth. J. Zool. 32, 495-512. Paulin, M. G. (1995). Electroreception and the compass sense of sharks. J. Theor. Biol. 174, 325-339.
50. Electroreceptors and Magnetoreceptors Perry, A., Bauer, G. B., and Dizon, A. E. (1981). Magnetite in the green turtle. Trans. Am. Geophys. Union 62, 850. Peters, R. C., and Bretschneider, E (1972). Electric phenomena in the habitat of the catfish, lctalurus nebulosus LeS. J. Comp. Physiol. 81, 345-362. Peters, R. C., and Evers, H. P. (1985). Frequency selectivity in the ampullary system of an elasmobranch fish (Scyliorhinus canicula). J. Exp. Biol. 118, 99-109. Peters, R. C., and Meek, E (1973). Catfish and electric fields. Experientia (Basel) 29, 299-300. Pfeiffer, W. (1968). Die Fahrenholzschen Organe der Dipnoi und Brachiopterygii. Z. Zellforsch. 90, 127-147. Preston, D., and Pettigrew, J. D. (1980). Ferromagnetic coupling to muscle receptors as a basis for geomagnetic field sensitivity in animals. Nature 285, 99-101. Raschi, W. (1986). A morphological analysis of the ampullae of Lorenzini in selected skates (Pisces-rajoidei). J. Morphol. 189, 225-247. Raschi, W., and Mackanos, L. A. (1989). The structure of the ampullae of Lorenzini in Dasyatis garouaensis and its implications on the evolution of freshwater electroreceptive systems. J. Exp. Zool. 2, 101-111. Ronan, M. C. (1986). Electroreception in cyclostomes. In "Electroreception" (T. H. Bullock and W. Heiligenberg, Eds.), pp. 209-224. John Wiley and Sons, New York. Ronan, M. C., and Bodznick, D. (1986). End buds: Non-ampullary electroreceptors in adult lampreys. J. Comp. Physiol. 158, 9-16. Scheich, H. (1977). Neural basis of communication in the high frequency electric fish, Eigenmannia viriscens (jamming avoidance response). III. Central integration in the sensory pathway and control of the pacemaker. J. Comp. Physiol. 113, 229-255. Scheich, H., Bullock, T. H., and Hamstra, R. H. (1973). Coding properties of two classes of afferent fiber: High frequency electroreceptors in the electric fish, Eigenmannia. J. Neurophys. 36, 39-60. Scheich, H., Langner, G., Tidemann, C., Coles, R. B., and Guppy, A. (1986). Electroreception and electrolocation in platypus. Nature 319, 401-402. Schessel, D. A., Ginzburg, R. D., and Highstein, S. M. (1982). Ultrastructural and physiological evidence for chemical synaptic transmission between the vestibular type I hair cell and its primary nerve chalice. Anat. Rec. 202, 168. Srivastava, C. B. L. (1973). Peripheral nerve ending types in tuberous electroreceptors of a high frequency gymnotid, Sternarchus albifrons. J. Neurocytol. 2, 77-83. Steinbach, A. B., and Bennett, M. V. L. (1971). Effect of divalent ions and drugs on synaptic transmission in phasic electroreceptors in a mormyrid fish. J. Gen. Physiol. 58, 580-598. Suga, N. (1967). Coding in tuberous and ampullary organs of a gymnotid electric fish. J. Comp. Neurol. 131, 437-453. Sugawara, Y. (1989). Two Ca ++ current components of the receptor current in the electroreceptors of the marine catfish Plotosus. J. Gen. Physiol. 93, 365-380. Sugawara, Y., and Obara, S. (1989). Receptor Ca ++ current and Ca ++gated K § current in tonic electroreceptors of the marine catfish Plotosus. J. Gen. Physiol. 93, 343-364. Szabo, T. (1965). Sense organs of the lateral line system in some electric fish of the Gymnotidae, Mormyridae and Gymnarchidae. J. Morphol. 117, 229-250. Szabo, T. (1970). Morphologische und funktionelle Aspekte bei Elektrorezeptoren. Verh. Dtsch. Zool. Ges. 64, 141-148. Szabo, T. (1974). Anatomy of the specialized lateral line organs of electroreception. In "Handbook of Sensory Physiology" (A. Fessard, Ed.), Vol. III/3, pp. 13-58. Springer-Verlag, New York.
855 Szabo, T., and Hagiwara, S. (1967). A latency-change mechanism involved in sensory coding of electric fish (mormyrids). Physiol. Behav. 2, 331-335. Szabo, T., and Wers~ill, J. (1970). Ultrastructure of an electroreceptor (mormyromast) in a mormyrid fish. Gnathonemus petersii. J. Ultrastruct. Res. 30, 473-490. Szamier, R. B., and Wachtel, A. W. (1970). Special cutaneous receptor organs of fish. VI. Ampullary and tuberous organs of Hypopomus. J. Ultrastruct. Res. 30, 450-471. Teeter, J. G., and Bennett, M. V. L. (1981). Synaptic transmission in the ampullary electroreceptor of the transparent catfish, Kryptopterus. J. Comp. Physiol. 142, 371-377. Teeter, J. G., Szamier, R. B., and Bennett, M. V. L. (1980). Ampullary electroreceptors in the sturgeon Scaphirhynchus platorynchus (Rafinesque). J. Comp. Physiol. 138, 213-223. Tenforde, T. S. (1989). Electroreception and magnetoreception in simple and complex organisms. Biomagnetics 10, 215-221. Tricas, T. C. (1982). Bioelectric-mediated predation by swell sharks, Cephaloscyllium ventriosum. Copeia 1982, 948-952. Tricas, T. C., and New, J. G. (1998). Sensitivity and response dynamics of elasmobranch electrosensory primary afferent neurons to near threshold fields. J. Comp. Physiol. 182, 89-101. Tricas, T. C., Michael, S. W., and Sisneros, J. A. (1995). Electrosensory optimization to conspecific phasic signals for mating. Neurosci. Lett. 202, 129-132. Viancour, T. (1979a). Electroreceptors of a weakly electric fish. I. Characterization of tuberous receptor tuning. J. Comp. Physiol. 133, 317-325. Viancour, T. (1979b). Electroreceptors of a weakly electric fish. II. Individually tuned receptor oscillations. J. Comp. Physiol. 133, 327-338. Vischer, H. A. (1995). Electroreceptor development in the weakly electric fish Eigenmannia: A histological and ultrastructural study. J. Comp. Neurol. 360, 81-100. vonder Emde, G., and Bleckmann, H. (1992). Extreme phase sensitivity of afferents which innervate mormyromast electroreceptors. Naturwissenschaften 79, 131-133. Wachtel, A. W., and Szamier, R. B. (1966). Special cutaneous receptor organs of fish: The tuberous organs of Eigenmannia. J. Morphol. 119, 51-80. Walcott, C., Gould, J. L., and Kirschvink, J. L. (1979). Pigeons have magnets. Science 201, 1027-1029. Walker, M. M., Kirschvink, J. L., Chang, S. B. R., and Dizon, A. E. (1984). A candidate magnetic sense organ in the yellowfin tuna, Thunnus albacares. Science 224, 751-753. Waltmann, B. (1966). Electrical properties and fme structure of the ampullary canals of Lorenzini. Acta Physiol. Scand. 66, Suppl. 264, 1-60. Watson, D., and Bastian, J. (1979). Frequency response characteristics of electroreceptors in the weakly electric fish, Gymnotus carapo. J. Comp. Physiol. 134, 191-202. Wu, C. H. (1984). Electric fish and the discovery of animal electricity. Am. Sci. 72, 598-607. Yager, D. D., and Hopkins, C. D. (1993). Directional characteristics of tuberous electroreceptors in the weakly electric fish, Hypopomus (Gymnotiformes). J. Comp. Physiol. 173, 401-414. Zakon, H. H. (1984a). Postembryonic changes in the peripheral electrosensory system of a weakly electric fish: Addition of receptor organs with age. J. Comp. Neurol. 228, 557-570. Zakon, H. H. (1984b). The ionic basis of the oscillatory receptor potential of tuberous electroreceptors in Sternopygus. Soc. Neurosci. Abstr. 10, 193. Zakon, H. H. (1986a). The electroreceptive periphery. In "Electroreception" (T. H. Bullock and W. Heiligenberg, Eds.), pp. 103-156. John Wiley and Sons, New York. Zakon, H. H. (1986b). The emergence of tuning in newly-generated tuberous electroreceptors. J. Neurosci. 6, 3297-3308.
856 Zakon, H. H. (1987). Variation in the mode of receptor cell addition in the electrosensory system of gymnotiform fishes. J. Comp. Neurol. 262, 195-214. Zakon, H. H. (1988). The electroreceptors: Diversity in structure and function. In "Sensory Biology of Aquatic Animals" (J. Atema, R. R. Fay, A. N. Popper, and W. N. Tavolga, Eds.), pp. 813-850. Springer, New York.
SECTION V Synaptic Transmissions and Sensory Transductions
Zakon, H. H., and Meyer, J. H. (1983). Plasticity of electroreceptor tuning in the weakly electric fish Sternopygus dariensis. J. Comp. Physiol. 153, 477-487. Zipser, B., and Bennett, M. V. L. (1973). Tetrodotoxin resistant electrically excitable responses of receptor cells. Brain Res. 62, 253-259. Zoeger, J., Dunn, J. R., and Fuller, M. (1981). Magnetic material in the head of the common Pacific dolphin. Science 213, 892-894.
Harvey M. Fishman
Appendix: The Biophysics of Electroreception in Ampullary Organs of Elasmobranch Fishes
Specialized organs, the ampullae of Lorenzini, make electroreception possible in marine elasmobranch fishes (sharks, skates, and rays) and other animals that possess this unusual sensory modality (see Chapter 50). These organs transduce variations in electric field intensity into modulations of the basal spike-firing rate in the primary afferent nerve, which innervates the ampullae and conveys this signal to the brain. Most animals, including humans, do not have an electroreceptive sense, but ampullary epithelial cells and mechanosensitive hair cells of the inner ear have some common morphological and functional properties (see Chapter 46). Electroreception in marine elasmobranchs 1 also ranks among the most sensitive sensory systems (for a comparison see Block, 1992). Recently, the mechanism that underlies this acute sensitivity has become of interest with respect to understanding how weak electromagnetic fields could possibly affect humans (Hafemeister, 1998). Based on behavioral responses, sharks and rays respond to electric field intensities as low as, and perhaps lower than, 0.5 laV/m (Kalmijn, 1966). This acute electrical sense apparently enables them to find prey and to orient themselves with respect to the earth's magnetic field (Kalmijn, 1988). In this brief appendix, we consider the biophysical properties of ampullary organs that enable transduction of electric field intensities into neural signals, which are perceived in the brain. Some of these properties were deduced from voltage-clamp and complex admittance measurements (Lu and Fishman, 1994) using the isolated, single-organ preparation 2 of Clusin and Bennett (1977) excised from skates. Such whole-organ preparations allow determination of an organ's output (neural spike activity) to various electrical stimuli applied to the ampullary epithelium (Fig. A-l). Based on these isolated organ data, we describe the evidence that supports the following four assertions. (1) The behavioral responses of animals to low-intensity electric fields are
1Here we consider only the biophysical properties of ampullary organs in marine elasmobranchs, which are the most sensitive with respect to behavioral responses to exogenous electric fields. Some details of ampullary organ function in freshwater elasmobranchs, such as the electric field orientation for excitation versus inhibition of afferent nerve activity, are significantly different (see Chapter 50). 2Each excised organ preparation consists of a 4-cm length of ampullary canal attached to an ampulla, lined with epithelial cells, which are innervated by a 1-cm length of primary afferent nerve. The canal is filled with a highly conductive gel and lined with nonconducting epithelial cells. Cell Physiology Sourcebook: A Molecular Approach, Third Edition
85 7
primarily due to the properties of ampullary organs. (2) The electric-field orientation (stimulus polarity) that produces an excitatory output (i.e., increased afferent nerve activity) is due to the conduction properties of the specific ion channels in apical and basal membranes of ampullary epithelial cells. (3) Signal amplification is achieved by interaction of physically separated and different ion conductances in apical and basal membranes of the ampullary epithelial cells. (4) The acute electric-field sensitivity is achieved by a convergent system that enhances the signal relative to background noise in the signal generation and amplification processes of the sensory epithelium. When marine elasmobranchs are in their seawater environment, electrically conductive paths through seawater and their low skin resistance (Murray, 1967) shunt all ampullary organs. Consequently, the high shunt resistance of "opencircuit" (current-clamp) measurements of transorgan voltages in response to currents applied to an electrically isolated organ do not correspond to the operational conditions of an ampullary organ in situ in an animal (Lu and Fishman, 1994). In contrast, the low shunt resistance or "short-circuit" (voltage-clamp) measurements of ampullary transepithelial current in response to control of transepithelial voltage more closely approximate the electrical "load" on an organ in an animal swimming in seawater. Thus, voltage-clamp responses are more likely to reflect organ operational characteristics and to provide the electrical data that are essential for understanding ampullary electroreceptor function. From such voltage-clamp data (Lu and Fishman, 1994), the relationship between the spike-firing rate of the afferent nerve (organ output) and applied step-voltage changes across an epithelium (input) was determined. This transfer curve (Fig. A2) shows that an organ provides a discernible signal to the brain of an animal, in response to applied transepithelial voltages of a few microvolts or less! Furthermore, when the transepithelial voltage (luminal with respect to basal side) is more negative than -100 pV (Fig. A-2), no additional increase in the rate of afferent spike firing occurs (i.e., the output becomes saturated). Thus, the saturated output for inputs more negative t h a n - 1 0 0 laV is indicative of the upper limit of the excitatory response of an organ. With respect to accounting for marine elasmobranch sensitivity (Kalmijn, 1966; Kalmijn, 1982), a 0.5-pV/m field applied to a 0.2-m length canal should produce an ampullary transepithelial voltage drop of 0.1 pV. Although this transepithelial voltage is still an order of magnitude lower than the lowest level (1 ~V) step clamp response
Copyright 9 2001 by Academic Press. All rights of reproduction in any form reserved,
858
SECTION V Synaptic Transmissions and Sensory Transductions
FIGURE A-1. (A) Diagram showing the location of various ion channels and transporters in the ampullary epithelium as identified by the effect of specified antagonists, agonists, or ion substitutions on admittance data from isolated organs (from Lu and Fishman, 1995a). Conduction by Ca 2§ channels in apical membranes constitutes the negative conductance reflected in electrophysiological data (Fig. A-3) in that they provide the inwardly directed current (arrow) in response to depolarization of apical membranes (i.e., luminal negative-polarity voltages). Conduction by ion channels (C1ca (calcium-activated C1-), Ca 2§ and K § in basal membranes produces an oscillation involved in the release of neurotransmitter at presynaptic membrane (see Lu and Fishman, 1995b). The ion transporters (Na§ § and Na+-Ca2§ maintain a highly regulated membrane voltage and intracellular Ca 2§ concentration to enable detection of fields that produce sub-microvolt changes in transepithelial voltage. (B) Electron micrograph of the ribbon synapse (boxed region in A) formed at the basal membrane of an ampullary epithelial cell with the afferent nerve, which conveys electroreceptor signals to the brain. From Waltman (1996).
A
.4-1
100-
I,., ,,..,,. 'l,il-.
O
r
r :> ,,.,..
O_
inhibitory
Q -100-120
I
I
-80
I
I
-40
1
I
0
"1""
I,
40
I
I
R"
80
I
120
Ampullary transepithelial v o l t a g e (laV) F I G U R E A-2. Transfer characteristic curve of an isolated ampullary organ excised from the skate Raja erinacea. Data (filled circles) were obtained by measuring the afferent spike-firing rate relative to the basal firing rate (35 spikes/s at a holding potential of 0 mV) in response to transepithelial step voltage clamps from the holding potential. Applied step changes in ampullary transepithelial voltage polarized the apical (luminal) side relative to the basal side (see Fig. A-1A), Increasingly negative luminal voltages ( E E I.H
-20
v
-40 -60
,
0
1
,
2
,'
3
4
5
,
6
,
7
8
,
~
,
)
T i m e (rain)
FIGURE 6. Depolarization and corresponding tension allow a steady state to be maintained in a rat caudal artery during stimulation by 300 nM norepinephrine. These drawings are traced from original records and illustrate the positive correlation between membrane potential and tension that is typical of small arteries. Both membrane potential and tension levels can remain relatively constant for many minutes during stimulation by vasoconstrictors, and may explain a major part of how tone is developed and maintained in blood vessels. (Reproduced with permission from Hermsmeyer, 1982.)
906
SECTION Vl M u s c l e a n d O t h e r C o n t r a c t i l e S y s t e m s
A
-
0
Control (mY) -
B
~
,..._
.
.
.
.
.
.
.
.
~
-
-
-
-------
-56
Hypoxia
I
Em (mV) -56
C Hypoxia (Verapamil 10 -7 M)
0
-
-
9
.
.
.
.
(mY)
-56 10 min FIGURE 7. Microelectroderecording of membrane potential in a cannulated, pressurized (80 mmHg) pulmonary resistance artery in: (A) control oxygenated conditions (pO 2 = 300 Torr), (B) the presence of hypoxia (pO 2 - 30 Torr), and (C) in the continued presence of hypoxia after the addition of the L-type C a 2+ channel blocking drug, verapamil (10-7 M). The presence of hypoxia triggered depolarization and generation of action potentials in pulmonary smooth muscle cells, which were eliminated by verapamil. [Adapted from Harder et al. (1985) with permission.]
smooth muscle cells, which may be related to the inhibition of K + channels, are inhibited by drugs that block L-type Ca 2+ channels, indicating a key role for voltage-gated Ca 2+ influx in spike generation. Importantly, the contractile response of the pulmonary arteries to hypoxia reduces blood flow to poorly ventilated areas of the lung, but enhances blood flow to ventilated alveoli, where effective gas exchange can take place to support normal respiration. At least two general families of Ca 2+ permeant channels, the receptor-operated (nonselective) cation channels and the voltage-gated L-type Ca 2+ channels, appear to act in concert to mediate cell excitation and depolarization in response to receptor activation. Receptor-operated cation channels conduct inward Na + and Ca 2+ currents in response to ligand binding to G protein-coupled receptors, which induces a conformational change in the pore-forming protein (Bolton et al., 1999). Under some conditions, this influx of cations is thought to provide the initial depolarizing signal that activates L-type Ca 2+ channels to trigger smooth muscle contraction. The molecular structure of nonselective channels in smooth muscle tissues remains unclear, but the channel complex is thought to consist of both receptor-associated and channel domains, and may have only two transmembrane-spanning segments, similar to the KIR channel. These channels may arise from multiple gene families, because they show variable properties of ion selectivity, voltagesensitivity, and regulatory profiles in various smooth muscle cells. One of the first nonselective cation channels to be detected by patch-clamp techniques in smooth muscle membranes was the ATP-gated channel in vascular smooth muscle, depicted in Fig. 8 (Benham and Tsien,
1987). This channel opens in the presence of the ATP ligand molecule to conduct Na + as well as Ca 2+ ions through the channel pore. In any given smooth muscle tissue, dee-
tromechanical (voltage-gated) and pharmacomechanical (iigand-gated, voltage-insensitive) coupling may interact to determine the final level of contraction. Finally, receptor activation may permit the direct signaling of intracellular Ca 2+ release, without the opening of surface membrane Ca 2+ channels. In this case, the membranedelineated enzyme phospholipase C causes the conversion of a principal substituent of the membrane, phosphatidyl inositol into the metabolite, inositol 1,4,5-trisphosphate (InsP3), which causes localized (but highly effective) release of intracellular Ca 2+ (Berridge and Irvine, 1984; Bolton et al., 1999). Intracellular Ca 2+ release is primarily from membranous structures thought to function analogously to the sareoplasmie retieulum (SR), the prominent Ca 2+ release organelle in skeletal muscle. One portion of the SR that regulates intracellular Ca 2+ by release in response to membrane signals is located immediately inside the surface membrane, and functions as the superficial buffer barrier (Daniel et al., 1995). This sub-sarcolemmal SR is important in Ca 2+ homeostasis in vascular muscle cells because Ca 2+ entering through sarcolemmal Ca 2+ channels must pass through this region of SR, giving rise to the buffer barrier concept. Intracellular stores are modulated by the Ca 2+ content of the superficial buffer barrier, but at least in localized sites, the Ca 2+ concentration may increase more than onehundred fold from its resting level of 100-300 nM when Ca 2+ is released from the larger central component of the SR. Diacylglycerol (DAG) is also formed by the phospholi-
53. Smooth Muscle Action Potentials and Electrical Profiles
A 130 mM Na + ATP
B -100 mV
2s
C
907
130 mM Na §
~ , - , v . - - - - ~
-100
~
-120 , =~T-~-~_,-~-'~.-'a-~,~:- 140
110 mM Ca 2* ATP
-140 mV
m
40 ms
E V (mY)
0.7s
D
-150
-100
i
I
-50 I
B a 2 + ~ 110 mM Ca 2+
-1I~1
gl
(pA)
40 ms -2-
FIGURE 8. Individualcurrents activated by ATP in outside-out membrane patches tom from a cell are shown here. (A) Currents carried by Na§ and in (B), the same single channel events are spread out for individual inspection. Values to the right of each trace indicate the membrane potential at which the trace was obtained. (C, D) The same kind of membrane patch, but with Ca2§ as the current carrier. From these records, the amplitude of current for a given potential was determined and plotted, as shown in (E). From such plots, the conductance (Acurrent/Avoltage) was determined as 25 picosiemens (pS) for the divalent ion current and 200 pS for the Na§ currents. Reprinted by permission from Nature [Benham and Tsien (1987) copyright 1987 Macmillan Magazines Ltd.]
pase C-induced metabolism of phosphatidyl inositol, and DAG causes the activation of protein kinase C (PKC). The PKC signal triggers contraction through multiple mechanisms. Activation of PKC causes intracellular Ca 2§ release, phosphorylation of the contractile proteins leading to an increased contraction for a given level of Ca 2+, and a markedly prolonged contraction. This combination of activation responses triggered by PKC results in strong, long-lasting contractions that correlate poorly, or at times not at all, with the level of depolarization, especially at later times during the contraction. Thus InsP 3 and PKC are other examples of substances that can induce p h a r m a c o m e c h a n i c a l coupling, that is, changes in contraction without requisite participation of membrane potential change (Somlyo and Somlyo, 1990). In summary, once membrane excitation has begun, a complex set of membrane-related processes amplify the original discrete set of signals. The intracellular concentration of Ca 2§ rises as a result of Ca 2§ influx or release from intracellular stores (or both). In many types of smooth muscle cells, the key event appears to be Ca 2§ entry through ion channels, with perhaps even the opening of a single Ca 2+ channel beginning the messenger cascade leading to contraction. In addition, some smooth muscle activators also are believed to produce contraction through a direct action on the intracellular signaling processes, whether or not membrane potential is allowed to change. The latter event is often triggered by the receptor-linked production of InsP 3 from phosphatidyl inositol during the activation of phospholipase C in the cell membrane. This process is a biochemical signaling cascade, capable of causing activation of multiple intracel-
lular processes, through amplification by intracellular messengers, that ultimately result in smooth muscle contraction. Interference with this process can be at several points, but Ca 2+ entry and release are particularly notable and vulnerable. Hence, Ca 2+ channel blocking drugs that interfere with voltage-gated Ca 2+ influx and the Ca 2+ signal process are very effective smooth muscle relaxants.
V. Heterogeneous Electrical Properties of Smooth Muscle Cells The electrical properties of different smooth muscle cells, including resting potentials and electrical spikes, have many fundamental similarities to other types of excitable cells. Among smooth muscles, there are c o m m o n characteristics such as the co-expression of different ion channel types, but it is difficult to generalize about them. In fact, smooth muscle cells, like striated muscles and perhaps even more like neurons, have a diverse array of electrical properties that correspond with functional differences in various organs, and even parts of an organ. For example, the resting membrane potential of smooth muscles varies over a range between a b o u t - 3 5 m V a n d - 6 5 mV. This variation reflects real differences in the ionic conductances of the cells being studied. In fact, within a smooth muscle organ, there is a diversity of resting membrane potentials distributed even within a very finite distance, suggesting different subpopulations of smooth muscle cells within an organ. The range of resting potentials in arteries has been reported by many
908
SECTION VI Muscle and Other Contractile Systems
investigators to span more than 10 mV. Studies of visceral smooth muscle show a similar range, possibly contributing to the composite characteristics of an organ. Figure 9 emphasizes the vast differences in spontaneous contractile activity found in circular smooth muscle isolated from different layers of the human colon, which reflect the localized differences in electrical patterns for smooth muscle excitation within this tissue (Rae et al., 1998). Thus, one way to think of an organ, such as the intestinal tract, is as a composite of smooth muscle cells consisting of subpopulations of cells which have a range of electrical properties that, when regarded together, constitute the electrical characteristics of that organ. At the basis of electrical heterogeneity is the diverse population of K + channels expressed in smooth muscle cells. Because the opening of K + channels sets the membrane potential level, it also influences the availability of voltagegated Ca 2+ and Na + channels that mediate the amplitude of the electrical spike. The K + channels also can modify the form (shape and duration) of the electrical spike, because they mediate the flow of electrical current (K + effiux) which repolarizes the cells after excitation. Thus, the outward current generated by K + channels contributes to the resting membrane potential, the pacemaker mechanism, and spike repolarization. If the K + channels are blocked by such drugs as tetraethylammonium or by the cation, barium, there is depolarization of the cells (usually producing contraction), and slowed repolarization of the spikes (that often leads to enhanced electrical activity, especially in organs showing rhythmic contractions). The ion that carries the inward (depolarizing) current that causes the electrical spikes found in smooth muscle tissues usually is calcium (Ca2+). In cells that are known for their
Circular Smooth Muscle of Human Colon
Whole Myenteric ~
s _._~~~~.~f~~
Interior
Submucosal ~
3gr 2 rain
F I G U R E 9. The highly variable patterns of spontaneous contractile activity observed in different circular smooth muscle layers of the human colon. The contraction patterns of the isolated smooth muscle strips were examined using isometric tension recording. [Adapted from Rae et al. (1998) with permission.]
spontaneous spiking activity, there is a higher expression of T-type Ca 2+ channels. However, the same spontaneously active cells also express L-type Ca 2+ channels in high density, and generate electrical spikes which are inhibited by Ca 2+ antagonists (blockers of L-type Ca 2+ channels). Thus the role of the T-type Ca 2+ channel in spike generation may be permissive rather than mandatory. Examples of cells showing prominent Ca2+-dependent action potentials or spikes include the azygos vein of newbom animals, the hepatic portal vein in all animals, numerous smooth muscle cells in the intestine, and the pacemaker cells of the progesteronedominated uterus. Under some conditions, the mechanism for smooth muscle spikes includes Na + channels, which can be found in certain smooth muscle cells, especially those that show prominent pacemaker activity. The function of Na + channels in smooth muscle cells that have a Ca 2+ channel pacemaker which functions independently of Na + channels is enigmatic. One useful way to think of a function for Na + channels in smooth muscle is the role of an adjunct to the pacemaker mechanism (Sturek and Hermsmeyer, 1986; Bolton et al., 1999; Farrugia, 1999). If the Ca 2+ pacemaker should fail to initiate spontaneous spikes, the Na + pacemaker is present and available for initiating spikes. Thus, it would provide a safety function for maintenance of pacemaker function. In comparison to neuronal and cardiac tissues which rely on the dense expression of Na § channels to mediate the upstroke of the action potential, Ca 2+ channels are expressed in lower densities in the plasma membrane of smooth muscle cells, and have a slower rate of activation (onset of channel opening). Thus, the electrical spikes in smooth muscle cells show a slower upstroke velocity and lower amplitude than the action potentials observed in neuronal cells or cardiac myocytes. Even more importantly, the inactivation (turn-off of current) of Ca 2+ channels occurs over a time course that is at least a thousandfold as long as for Na + channels. These channel kinetics permit intracellular Ca 2+ and contractile force to reach higher levels, and for contractions to be sustained much longer in smooth muscles than in cardiac cells expressing Na + channels. The prolonged process of inactivation also allows for a diversity of function, and for custom tailoring of the excitation event by hormones and regulatory substances that act through variables of time and space. Thus pacemaker activity in smooth muscle cells depends on a balance of ion currents. The dynamic balance between inward and outward currents determines whether a cell will depolarize to reach the threshold voltage that triggers an electrical spike. A model of pacemaker activity based on Ca 2+ and K + channels is shown in Fig. 10 (Hermsmeyer and Akbarali, 1989). In pacemaker cells, the spontaneous depolarization caused by the influx of Ca 2+ and/or Na + is balanced by an outflow of K + current that tends to repolarize the membrane. If there is a slight increase in outward K § channel current, there will be hyperpolarization and suppression of pacemaker activity. If K + conductance slightly decreases, the cell will depolarize sufficiently to activate voltage-gated Na + or Ca 2+ channels and generate spontaneous electrical activity. It is only under conditions where K § conductance is in an optimum range and there is a balance of currents, that tireless, repetitive cycles of sponta-
909
53. Smooth Muscle Action Potentials and Electrical Profiles
tractions and continuously graded regulation. With the multiplicity of receptors and response mechanisms found in the deceptively small smooth muscle cell, the complexity of signaling mechanisms possible with the simple smooth muscle spike should be regarded as a compact masterpiece of multiple responsiveness.
!
Bibliography
/Ca
-
j,,,
_
_
ACa2+ /K(Ca) FIGURE 10. Pacemaker potentials are modeled by this diagram of membrane potential (Em) , C a 2+ current (/Ca)' intracellular C a 2+ c o n c e n t r a t i o n (ACa2+), and Ca2+ -dependent K § current (Ir~(Ca))"In this hypothetical vascular muscle cell, pacemaker potentials are evident as the slow upward drift of membrane potential toward zero to an inflection point, at which the slow upward movement suddenly bends to a nearly vertical rise that is the electrical spike. The spike is brief because of inactivation of Ca2+ channels that is evident in the return to baseline of Ca2+ current. The increase in intracellular Ca2+ (ACa2+) enhances outward current through the CaZ+-sensitive K + channels (Ir~(Ca)),which accelerate the repolarizing phase and allow the downsweep to the most negative E m at the beginning of the pacemaker depolarization. [Adapted from Hermsmeyer and Akbarali (1989) with permission.] neous activity can occur. Because of the multiplicity of regulators that modulate ion channels, spontaneous activity is somewhat irregular in smooth muscle cells, distinguishing it from the more clocklike timing mechanism found in the pacemaker cells of the heart.
VI. Summary Smooth muscle cells are small cells arranged in groups to form the contractile wall of hollow organs and blood vessels. In most cases, the muscle cell function is specialized for slow, maintained contraction (tone) rather than speed. Smooth muscle spikes show a high degree of diversity which translates into varied excitation. This diversity is in contrast to the rapid (but inflexible) signals that result from fast Na + channels in nerve and skeletal muscle. In contrast, the variability in the electrical signal in smooth muscles, generally based on the dynamic balance between outward K + current and inward Ca 2+ current, provides a range of gradation of activity for internal organs that permits normal function. Smooth muscle pacemakers, although relatively less studied than those in the heart, provide vital functions for blood vessels and visceral organs, and are important for understanding regulatory mechanisms and drug actions. The frequency and configuration of action potentials and electrical spikes are regulated by a multiplicity of stimulatory and inhibitory inputs to a diverse array of receptors on a single cell. Coordination through chemicals released by nerve fibers, secretory cells, or other local factors influences an array of ion conductances that allows for long-duration con-
Benham, C. D., and Tsien, R. W. (1987). A novel receptor-operated CaZ+-permeable channel activated by ATP in smooth muscle. Nature 328, 275-278. Benham, C. D., Hess, E, and Tsien, R. W. (1987). Two types of calcium channels in single smooth muscle cells from rabbit ear artery studied with whole-cell and single-channel recordings. Circ. Res. 61, 10-16. Berger, M. G., and Rusch, N. J. (1999). Voltage and calcium-gated potassium channels: functional expression and therapeutic potential in the vasculature. Perspect. Drug Disc. Design 15/16, 313-332. Berridge, M. J., and Irvine, R. F. (1984). Inositol trisphosphate, a novel second messenger in cellular signal transduction. Nature 312, 315-321. Bolton, T. B., Prestwich, S. A., Zholos, V., and Gordienko, D. V. (1999). Excitation-contraction coupling in gastrointestinal and other smooth muscles. Annu. Rev. Physiol. 61, 85-115. Bozler, E. (1938). Electric stimulation and conduction of excitation in smooth muscle. Am. J. Physiol. 122, 614-623. Bozler, E. (1948). Conduction, automaticity, and tonus of visceral muscles. Experientia 4, 213-218. Daniel, E. E., Van Breemen, C., Schilling, W. E, and Kwan, C. Y. (1995). Regulation of vascular tone: Cross-talk between sarcoplasmic reticulum and plasmalemma. Can. J. Physiol. Pharmacol. 73, 551-557. Farrugia, G. (1999). Ionic conductances in gastrointestinal smooth muscles and interstitial cells of Cajal. Annu. Rev. Physiol. 61, 45-84. Harder, D. R., Madden, J. A., and Dawson, C. A. (1985). Hypoxic induction of CaZ+-dependent action potentials in small pulmonary arteries of the cat. J. Appl. Physiol. 59, 113-118. Hermsmeyer, K. (1982). Electrogenic ion pumps and other determinants of membrane potential in vascular muscle. Physiologist 2g(6), 454--465. Hermsmeyer, K., and Akbarali, H. (1989). Cellular pacemaker mechanisms in vascular muscle. Prog. Appl. Microcirc. 15, 32-40. Hernandez-Nicaise, M.-L., Mackie, G. O., and Meech, R. W. (1980). Giant smooth muscle cells of Bero& Ultrastructure, innervation, and electrical properties. J. Gen. Physiol. 75, 79-105. Horowitz, B., Ward, S. M., and Sanders, K. M. (1999). Cellular and molecular basis for electrical rhythmicity in gastrointestinal muscles. Annu. Rev. Physiol. 61, 19-43. Meera, E, Wallner, M., Song, M., and Toro, L. (1997). Large conductance voltage- and calcium-dependent K + channel, a distinct member of voltage-dependent ion channels with seven N-terminal transmembrane segments (S0-$6), an extracellular N terminus, and an intracellular ($9-S10) C terminus. Proc. Natl. Acad. Sci. USA 94(25), 14066-14071. Nelson, M. T., Patlak, J. B., Worley, J. E, and Standen, N. B. (1990). Calcium channels, potassium channels, and voltage-dependence of arterial smooth muscle tone. Am. J. Physiol. 259, C3-C 18. Prosser, C. L., Burnstock, G., and Kahn, J. (1960). Conduction in smooth muscle: Comparative structural properties. Amer. J. Physiol. 199, 545-552. Quayle, J. M., Nelson, M. T., and Standen, N. B. (1997). ATP-sensitive and inwardly rectifying potassium channels in smooth muscle. Physiol. Rev. 77, 1165-1231.
910
Rae, M. G., Fleming, N., McGregor, D. B., Sanders, K. M., and Keef, K. D. (1998). Control of motility patterns in the human colonic circular muscle layer by pacemaker activity. J. Physiol. 510, 309-320. Rich, A., Kenyon, J. L., Hume, J. R., Overturf, K., Horowitz, B., and Sanders, K. M. (1993). Dihydropyridine-sensitive calcium channels expressed in canine colonic smooth muscle cells. Am. J. Physiol. 264, C745-C754. Shmigol, A. V., Eisner, D. A., and Wray, S. (1998). Properties of voltage-activated [Ca2+] transients in single smooth muscle cells isolated from pregnant rat uterus. J. Physiol. 511, 803-811.
SECTION VI Muscle and Other Contractile Systems
Somlyo, A. E, and Somlyo, A. V. (1990). Flash photolysis studies of excitation-contraction coupling, regulation, and contraction in smooth muscle. Ann. Rev. Physiol. 52, 857-874. Sturek, M., and Hermsmeyer, K. (1986). Calcium and sodium channels in spontaneously contracting vascular muscle cells. Science 233, 475-478. Tomita, T. (1970). Electrical properties of mammalian smooth muscle. In "Smooth Muscle" (E. Btilbrin, A.E Brading, A.W. Jones, and T. Tomita, Eds). The Williams and Wilkins Company, Baltimore. pp. 197-243.
Judith A. Hein y
Excitation-Contraction Coupling in Skeletal Muscle
I. Introduction
neuron arrives at the neuromuscular junction, resulting
Skeletal muscle is activated to contract by a sequence of fast, electrically driven events that are collectively termed excitation-contraction coupling (EC coupling). EC coupling is a highly organized process of signal transduction that utilizes specialized membranes, membrane junctions, and ion channels on both the exterior and interior of the cell. This chapter describes the cellular and molecular processes that mediate electrical excitation of the outer membranes and transduce it into a signal for intracellular Ca 2§ release, focusing on vertebrate fast-twitch skeletal muscle which is the best characterized. Chapter 55 discusses molecular mechanisms for control of Ca 2§ release and reuptake by the sarcoplasmic reticulum (SR). Chapter 56 discusses mechanisms of force generation by the contractile proteins actin and myosin. Skeletal muscle is the fastest contracting of the three major muscle types and is related to cardiac and smooth muscle types evolutionarily and embryonically. The proteins that mediate excitation-contraction in cardiac, smooth, and skeletal muscle share a high degree of homology. Nonetheless, each muscle type has evolved highly differentiated mechanisms of using these proteins to serve its specialized functions. While cardiac and smooth muscle primarily mediate enteric contractile processes that are under autonomic and humoral control, skeletal muscle primarily mediates rapid, willed bodily movements that are under the control of the central nervous system. Consequently, the defining features of skeletal muscle activation are speed and voluntary control, as needed for fine control of body movement.
II. Overview of EC Coupling The sequence of EC coupling in a vertebrate fast-twitch skeletal muscle fiber is shown schematically in Fig. 1. Activation begins when an action potential from a motor Cell Physiology Sourcebook: A Molecular Approach, Third Edition
9 1 1
in the release of the neurotransmitter acetylcholine (ACh) into the synaptic clefts (see Fig. 1 of Chapter 41). Binding of ACh to ACh receptors on the adjacent end plate (see Chapter 40) locally depolarizes the postsynaptic membrane to threshold for exciting an action potential on the muscle sarcolemma. The action potential rapidly propagates the depolarization to the entire sarcolemma by a mechanism similar to action potential generation in the nervous system (see Chapters 24 and 25). The action potential also propagates into the fiber interior via a specialized system of tubular membranes (transverse-tubules or T-tubules) which are continuous with and invaginate transversely from the sarcolemma at periodic intervals. The T-tubules provide the conduit for the action potential to reach the fiber interior, and also bring the outer membranes into close proximity with the internal sarcoplasmic reticulum (SR) at specialized intracellular junctions called triads. The triad junction mediates rapid communication between the extracellular sarcolemma/T-tubules and the intracellular SR. At the triad junctions, voltage-dependent Ca 2+ channels in the T-tubular membrane (also called voltage sensors because they respond to the action potential depolarization, or dihydropyridine receptors (DHPRs) because of their sensitivity to the dihydropyridine class of Ca 2+ channel blocking drugs) detect the depolarization and transduce it into a signal for opening Ca 2+ release channels (also called the ryanodine receptors (RyRs) because they bind this plant alkaloid with high affinity) on the closely opposed SR membrane. The molecular interactions between these two distinct types of Ca 2+ channel are not completely known but it is likely that they form a macromolecular complex that interacts allosterically. It is hypothesized that membrane depolarization drives a conformational change on charged, intramembrane domains of the T-tubule Ca 2+ channel which, in turn, drives a conformational change on the SR Ca 2+ release channel, which results in its opening. Whether this Copyright 9 2001 by Academic Press. All rights of reproduction in any form reserved.
912
SECTION VI Muscle and Other Contractile Systems
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.
Resting [Ca 2+]i -0.1 pM Neuromuscular transmission Action potential propagation along sarcolemma and into T-tubules Signal transduction from Ca 2+ channels/DHPRs to Ca 2+ release channels/RyRs at triad junctions Ca 2+ release from SR Cytosol [Ca 2+] reaches 1-10 pM Diffusion and binding of Ca 2+ ions to TnC Removal of troponin inhibition of the contractile proteins Contractile proteins shorten to generate force Reuptake of Ca2+ into SR by Ca2+-ATPase Inactivation of Ca 2+ release channels Binding of Ca 2+ to calsequestrin
FIGURE 1.
Sequenceof muscle activation and relaxation.
interaction requires accessory proteins as part of a larger complex remains to be established. The SR, related evolutionarily to the endoplasmic reticulum, is a highly specialized membrane compartment for controlling cytosolic Ca 2§ In a resting muscle fiber, the SR actively maintains cytosolic [Ca 2+] at submicromolar levels, typically 0.1 laM, by active transport mediated by Ca2+-ATPase molecules present at high density along its length. Upon receiving a stimulus from the T-tubule the SR rapidly releases Ca 2+ into the cytosol at rates approaching 100 pM/ms to raise cytosolic [Ca 2+] to micromolar levels (typically 1-10 pM). Ca 2+ efflux occurs through RyRs localized on the junctional regions of the SR (jSR) where the SR membrane nearest the T-tubules form cisternae (terminal cisternae). Released Ca 2+ ions diffuse and bind to troponin-C (TnC), the regulatory subunit of troponin, thereby removing troponin's inhibitory effect on the contractile proteins, actin and myosin, which shorten to generate force (see Chapter
56). Force output is proportional to cytosolic Ca 2+ concentration (Fig. 2). Upon repolarization, Ca 2§ release terminates and multiple mechanisms act in concert to return cytosolic Ca 2§ to resting levels. High-density Ca2+-ATPase pumps on the SR membrane rapidly pump Ca 2§ back into the SR where it is largely bound to the Ca 2§ binding protein calsequestrin. Cytosolic cae+-binding proteins with rapid kinetics buffer Ca 2§ transiently while the slower SR pumps return it to the SR. Some Ca 2§ is also extruded extracellularly by active Ca 2§ transport mechanisms in the sarcolemma. The RyR is itself inactivated by a Ca2+-de pendent mechanism. Force terminates when cytosolic Ca 2+ returns to resting levels. Thus, cytosolic [Ca 2§ is the central link between membrane excitation and activation of the contractile proteins (Fig. 3). Contraction is inhibited at low resting Ca 2§ levels and proceeds when Ca 2§ is elevated transiently in response to membrane excitation.
54. Excitation-Contraction Coupling in Skeletal Muscle
9 13
100 -
80-
60 0 L_ 0 LL
40
20Relaxed 0 4 10-8
FIGURE 2. concentration.
I I 10-7 10-6 Myoplasmic [Ca2+]
I 10-5 (M)
I 10-4
Relationship between force and myoplasmic Ca2+
III. S p e e d of Skeletal Muscle Activation Activation of skeletal muscle is rapid despite the fact that skeletal muscle cells are among the largest of mammalian cells. Skeletal muscle fibers are multinucleated cells that form during embryogenesis by fusion of precursor myoblasts. They are cylindrically shaped with diameters from 5 0 - 1 5 0 laM and lengths from millimeters to centimeters. To produce effective mechanical force, all parts of a muscle fiber must develop tension simultaneously. It was recognized early that a trigger process other than one involving a diffusion-limited second messenger must be involved (Hill, 1948). For example, a small ion entering through the outer membrane of a 100-pm muscle fiber would require 100-200 ms to diffuse and bind to putative activating sites distributed throughout the cross-section. By comparison, a fast-twitch muscle fiber is activated in 2 0 - 3 0 ms.
Synchronous activation of skeletal muscle is achieved via a rapidly propagating action potential that spreads the depolarization along the sarcolemma and into the cell interior along the specialized conducting membranes of the Ttubules. The final trigger, Ca 2§ diffuses at most a few micrometers from its release site on the SR to the activator sites on TnC. Figure 4A shows the typical delay between membrane excitation and contraction for a frog fast-twitch muscle at 18 ~ Tension starts to develop within 3-5 ms of the action potential and peaks in 2 0 - 3 0 ms. Table 1 summarizes the speed of the key events following neuromuscular transmission that contribute to this delay in a frog muscle fiber at 18 ~ For example, propagation of the action potential laterally along the sarcolemma from the middle of the fiber to both tendons introduces a delay of 5 ms. Signal transduction into the fiber interior, including all events from sarcolemma depolarization to activation of the innermost contractile proteins, requires 4-5 ms. Each of these events is faster in mammalian muscle at 37 ~ Figure 4B shows the time course of the cytosolic Ca 2§ change and tension in a mammalian fast-twitch muscle at 37 ~ stimulated with a single action potential or at tetanic frequency to produce a maintained contraction.
[mV]
[Nlcm 2]
+50 0
20
-50 10 -100 I 0
I 10
1 20
I 30
I 40
,,,
0 I 50 [ms]
B A[Ca2+] 20 pM
.... ~
- - -~-~- - . . . . ' - . . . . . '. . . . . .
o
Tension
lOO
Time (ms)
FIGURE 3. A rise in cytosolic [Ca2+] links membrane excitation to activation of contractile proteins.
FIGURE 4. Speedof EC coupling in skeletal muscle. (A) Speed of EC coupling in a fast-twitch frog skeletal muscle fiber at 18 ~ showing the action potential (top trace) and tension (lower trace). (Modified from Hodgkin and Horowicz, 1957.) (B) Speed of EC coupling in a mouse skeletal muscle fiber at 37 ~ The top traces show the myoplasmic Ca2§ change measured with a Ca2+ indicator dye. The lower trace shows tension. The fiber was stimulated at time zero with a single action potential, and with repetitive action potentials at high frequency, to simulate a muscle tetanus. (Reproduced from The Journal of General Physiology, 1996,108, 455--469 by copyright permission of The Rockefeller University Press.)
914
SECTION Vl Muscle and Other Contractile Systems TABLE 1 D u r a t i o n of Key S t e p s in t h e Activation of a Fast-Twitch Skeletal Muscle Durationa
EC coupling steps Action potential propagation along sarcolemma
5-10 ms
Action potential propagation to center of fiber along T-tubules
--0.7 ms
Signal transduction at triad junction, from T-tubule depolarization to activation of RyR on SR
-- 0.5 ms
Peak rate of Ca2+ release to peak Peak myoplasmic
C a 2+
C a 2+
binding to TnC (start of tension)
change to peak tension
2-3 ms 15-25 ms
aDurations were calculated for a hypothetical frog fiber of 50 pm diameter and 5 cm length, having a central end plate. Literature values for conduction velocity and duration of intermediate steps (GonzalesSerratos, 1971; Vergara and Delay, 1986; Jong et al., 1996) were adjusted to 18 ~
IV. M e m b r a n e Architecture of EC Coupling Synchronous activation depends on effective, rapid communication between the specialized extracellular membranes of the sarcolemma/T-tubules and the intracellular SR. To achieve this, the membranes of skeletal muscle are highly organized to bring into close physical proximity the proteins involved in each of the key events: membrane excitation, Ca 2§ release, and force generation (rev. in Franzini-Armstrong and Jorgensen, 1994; Franzini-Armstrong and Protasi, 1997). Figure 5 illustrates the three-dimensional architecture of the sarcolemma, T-tubules, triad junctions, and SR membranes of a fast-twitch skeletal muscle fiber, and their physical relationship to the contractile proteins. The sarcolemma and T-tubule
membranes are organized to conduct the action potential rapidly to all parts of the fiber. The T-tubules compose the majority of the sarcolemma membrane, representing 50-80% of the total sarcolemma area (Peachey, 1965). They invaginate from the sarcolemma in a transverse plane approximately twice every 1-2 pM (at two planes per sarcomere in mammalian muscle). Within this plane they branch extensively to form a network that covers the entire cross-sectional area of the muscle cell, as shown in Fig. 6. Up to 80% of the tubular membrane is associated with the SR at triad junctions (Peachey, 1965; Dulhunty, 1984). Given this membrane architecture, no part of the sarcolemma is more than a few tenths of a micrometers and at most a few milliseconds signal transduction time from the Ca 2§ release channels of the SR.
FIGURE 5. Three-dimensional reconstruction of a longitudinal section of a skeletal muscle fiber from the frog. The sarcolemma surrounds bundles of contractile proteins called myofibrils. T-tubules invaginate transversely from the sarcolemma at periodic intervals and form junctions with the SR along their entire length. At the triad junctions, the longitudinally oriented SR membranes widen into sacs called terminal cisternae, which closely oppose the T-tubules and contain the Ca2§ release channels/ RyRs. The terminal cisternae also closely overlay the activating Ca 2§ sites on TnC. The functional unit of force generation is the sarcomere, consisting of overlapping actin and myosin filaments anchored at each end. The longitudinally oriented tubular regions of SR contain the Cae+-ATPase which takes up released Ca2§ all along the sarcomere. (Reproduced from The Journal of Cell Biology, 1965, 25, 209-232 by copyright permission of The Rockefeller University Press.)
54. Excitation-Contraction Coupling in Skeletal Muscle
FIGURE 6. Reconstruction of the T-tubule network over the whole cross-section of a frog twitch muscle fiber. The reconstruction was made by tracing electron micrographs of transverse serial slices of muscle perfused with the electron-dense and membrane-impermeant molecule peroxidase, which diffused into the T-tubules. (From Peachey & Eisenberg, 1978.)
Likewise the triad junctions are optimized for rapid signal transduction from the T-tubule to the SR. Figure 7A shows the morphology of the triad junction, at which discrete junctional domains of the SR interact with specialized junctional domains of the T-tubule. Figure 7B shows the spatial organization of the two key proteins of the triad junction, the DHPRs and the RyRs, which associate at the junctions to form a functional signal transduction complex. At the junctions, the T-tubule and SR membranes flatten and face each other across a narrow gap of about 10 nm. The junctional surface of the temainal cistemae of the SR contains two rows of proteins corresponding to the Ca 2+ release channels/RyRs. The RyRs are packed in highly ordered arrays, possibly touching at the comers, in a skewed pattern with a center-to-center spacing of about 30 nm. The RyRs are also termed foot proteins because of their unique appearance as dense, bridging structures in transverse electron micrographs of the triad. Each RyR/Ca 2+ release channel is composed of four identical subunits. Each subunit has a membrane-spanning domain and a large cytosolic domain that extends across the junctional gap and comes to within 1 nm of the T-tubule membrane. In a top view, the large cytosolic domains of the tetramer roughly resemble four spheres assembled in a four-leaf clover or quatrefoil pattern. The large cytoplasmic domain is identified with the foot structure observed in electron micrographs of the triad junction. In a side view, the tetramer assumes a mushroom shape with a smaller central region composed of four equal lobes that presumably insert into the jSR membrane to form the ion-conducting pore of the molecule. The junctional T-tubule membranes also contain a high density of proteins that are aligned with a regular periodicity in parallel rows of four-particle arrays, termed junctional tetrads (Block et al., 1988). The tetrads have been identified as groups of four voltage sensors/DHPRs and are arranged in a pattern approximately corresponding to the outer comers of the cytosolic spheres of the RyR tetramer. Tetrads are asso-
9 15
ciated with RyRs in an alternate disposition such that one tetrad is associated with every other RyR (indicated by dotted lines in Fig. 7B). Figure 7B was taken from a fish skeletal muscle fiber, in which tetrads line up opposite every other RyR in a 1:2 ratio. A similar arrangement of tetrads is suggested in triads and peripheral couplings of other skeletal muscles from the frog, rat, chicken, mouse, and human (FranziniArmstrong and Jorgensen, 1994; Franzini-Armstrong and Kish, 1995). However, the ratio determined from binding studies may be less than that inferred from electron micrographs, so the 1:2 ratio of DHPR tetrads to RyRs may be a lower limit (Anderson et al., 1994). Variations on this architecture theme for the intracellular junctions occur in different species. These include dyads, in which a junctional region of terminal cisterna is apposed to a single short segment of T-tubule, and peripheral couplings, in which a junctional region of terminal cisterna is apposed to an approximately circular junctional domain of the surface membrane. Triads, dyads, and peripheral couplings are structurally and functionally equivalent. Each is the intracellular site at which the interaction between DHPRs and RyRs links electrical excitation of the outer membrane to intracellular Ca 2§ release. Collectively, these intracellular junctions are called calcium release units (CRUs; Franzini-Armstrong et al., 1999). A further functional subdivision of the system, termed a couplon has been applied to a functional grouping of RyRs and DHPRs that act in concert during EC coupling (Stem et al., 1997). Accordingly, the frog triad junction contains two couplons and a peripheral coupling contains a single couplon.
A. Proteins of the Triad Junction Figure 8 shows the molecular structure and membrane topology of DHPRs and RyRs at the triad junction. The DHPR/T-tubule Ca 2§ channel is a voltage-gated ion channel belonging to the L-type family of Ca 2+ channels (rev. in DeWaard et al., 1996; Perez-Reyes and Schneider, 1995; see also Chapters 27 and 33). Multiple subtypes of the L-type calcium channel with tissue-specific distribution have been identified, including skeletal muscle, heart, brain, and endocrine isoforms. The L-type Ca 2§ channels belong to the superfamily of voltage-gated ion channels that share common structural and functional motifs. All L-type calcium channels are multisubunit complexes composed of at least three subunits associated in an equimolar ratio: a, O~26, and/3. The skeletal channel has a fourth y subunit. The alpha isoforms are distributed in a highly tissue-specific manner and are the product of different genes and/or alternative splicing mechanisms. Apparently all L-type channels contain 0/2-6 , which is encoded by a single gene. The/3 isoforms are tissue specific and result from different genes and alternative splicing. The skeletal-specific 3' subunit is a single gene product. The complete skeletal muscle calcium channel consists of als, a2-6,/31a, and 3' subunits (Fig. 9A). These have been purified, sequenced, cloned, and expressed. Expression of als alone or in various combinations with a2-6, flla and y can reconstitute the Ca 2+ current. However, all four subunits are required for the formation of
916
SECTION VI Muscle and Other Contractile Systems
FIGURE 7. (A) Electron micrograph of a longitudinal section from a fish swim bladder muscle triad junction. (From FranziniArmstrong and Peachey, 1981.) (B) Three-dimensional reconstruction of a half-triad showing the organization of the junctional membranes and proteins involved in signal transduction. T-tubule and SR membranes face each other across a gap of about 10 nm. The junctional surface of the terminal cisternae of the SR bears two rows of proteins corresponding to the RyR/Ca2+ release channels. RyRs project into the gap and come to within 1 nm of the T-tubule membrane. The junctional T-tubule membrane contains parallel rows of four-particle arrays, termed junctional tetrads, identified as DHPRs. In fast-twitch fish muscle, tetrads are arranged in a zig-zag spacing that corresponds exactly to the position of every other RyR on the adjacent SR (indicated by dotted lines). Nonjunctional and longitudinal regions of the SR are continuous with the terminal cisternae and contain densely packed Ca2+-ATPase molecules. The lumen of the SR contains calsequestrin, a high-capacity Ca2+binding protein. (Reproduced from The Journal of Cell Biology, 1988, 107, 2587 by copyright permission of The Rockefeller University Press.) a functional DHP receptor with properties of the native channel (Suh-Kim et al., 1996). The primary a ls subunit contains the major functional domains of the channel. It is a large integral membrane protein (212 kDa) that contains the pore-forming and voltagesensing regions as well as the binding site for dihydropyridines and other calcium channel blockers. The a 2 and 8 chains (125 and 27 kDa, respectively) are linked through a disulfide bond to form a single integral membrane glycoprotein (125 kDa). The 8 chain contains a transmembrane segment that may anchor the a2-6 complex to the mem-
brane. The c~2-t~ subunit may play a role in determining the spatial distribution and membrane localization of the Ca 2+ channel. The flla subunit is a smaller cytoplasmic protein (58 kDa) that associates specifically with the skeletal isoform of alpha at an interaction domain on the I-II cytoplasmic linker. This subunit increases the activation kinetics of a is and is required for the proper expression and membrane targeting of als (Neuhuber et al., 1998; Gregg et al., 1996). It also participates in regulation of the voltage-sensing and EC coupling functions of the Ca 2+ channel (Beurg et al., 1999). The ~/ subunit is a small glycosylated membrane
54. Excitation-Contraction Coupling in Skeletal Muscle
Extracellular space
9 17
Dihydropyridine Receptor I
II
IV
III
T-tubule
i
Cytoplasm
,
Ryanodine Receptor
Foot region
4-------- Modulatory region ,
COOH r--
Sarcoplasmic reticulum
| [.~
_ Channel region
? --Nil= Lumen
S-/
i +
i" §
O00H
Calsequestrin
COOH
F I G U R E 8. Key proteins of the triad junction, the DHPR (top) and RyR (bottom), showing primary sequence and presumed secondary membrane topology. For simplicity, only the ce1 subunit of the DHPR and only one subunit of the ryanodine receptor homotetramer are shown. DHPRs are composed of one each of four subunits (a 1, ce2-6, ,8, y). The DHPR ce1 subunit contains the key functional domains including the voltage sensor, pore, and DHP-binding site. It consists of four homologous hydrophobic domains (I-IV) connected by cytosolic linker sequences. Each of the four domains in turn is composed of eight transmembrane helices (S 1-$6 alpha helices are drawn schematically as cylinders plus two shorter intramembrane helices that fold into the bilayer to form the central pore). Alternative models of the RyR suggest as many as 12 transmembrane domains. A question mark indicates the interaction between triadin and calsequestrin. Alternatively, the cytosolic carboxy end of triadin may fold parallel to the lumenal face of the SR membrane and bind calsequestrin. (Adapted from McPherson and Campbell, 1993.)
protein (25 kDa) of u n k n o w n function that associates with the skeletal C~s isoform but not with heart or brain alpha subunits. It contains four transmembrane domains and two N-linked glycosylation sites. The alpha subunit (Fig. 9B) is composed of four hydrophobic internal repeats, termed domains I, II, III, and IV, that share a high degree of homology. Each domain consists of eight transmembrane alpha-helical segments, denoted S 1,
$2, $3, $4, $5, $ 5 - 6 linker (or P-linker), and $6. Domains I - I V are connected by longer segments of more hydrophilic amino acids (denoted I-II, II-III and I I I - I V linkers). The linkers as well as N- and C-terminals are cytosolic. The domains assemble to form a transmembrane protein with a central pore. The major functional domains on C~lshave been identified. The P-linker helices of each domain fold into the membrane
918
F I G t l I ~ E 9. (A) Subunit structure of the skeletal muscle calcium channel showing the four subunits, C~ls, c~2-6 complex, J~la and y. (B) Assembly of the primary a ls subunit, showing the four hydrophobic domains that fold into the bilayer to form an integral membrane protein with a central ion-conducting pore.
to form the pore for ion flux. The $4 transmembrane helix within each domain is highly charged, containing a positively charged residue on every third amino acid. $4 is presumed to form the voltage sensor that responds to the depolarization during the action potential. Upon depolarization, all four $4 helices move in a cooperative fashion to activate both the DHPR and the allosterically coupled RyR. Movement of the $4 segments is presumed to generate the macroscopic intramembrane charge movement currents that are detected electrically during EC coupling. The skeletal muscle Ca 2+ channel is capable of functioning both as a voltage sensor for EC coupling and as a Ca 2+ ion channel. As a Ca 2+ channel, it is characterized by highvoltage activation, brief openings, and low single-channel ion flux. Activation of the skeletal Ca 2+ channel is slow compared with the closely related cardiac channel. The $4 charge movements may be followed by slower, less steeply voltage-dependent conformational changes leading to opening. It is thought that the early rapid movement of $4 domains, and not the Ca 2§ current that follows after a delay of several hundred milliseconds, has been exploited evolutionarily by skeletal muscle to control RyR activation. Due to the slow kinetics of opening, Ca 2§ influx through the channel pore is not significant during a normal action potential of a few milliseconds duration. However, it may build up to measurable levels during a tetanus, when a skeletal muscle
SECTION VI Muscle and Other Contractile Systems
is activated repetitively at rates of 50-100 Hz to elicit a maintained contraction. In this case Ca 2§ entry may add to and perhaps modulate intracellular Ca 2+ release (Oz and Frank, 1991). The RyR/SR Ca 2§ release channel belongs to the superfamily of ligand-gated ion channels (rev. in Franzini-Armstrong and Protasi, 1997; see also Chapter 55). RyRs share significant sequence and structural homology with the inositol 1,4,5trisphosphate (IP 3) receptors that release Ca 2+ from internal stores in other cell types. RyRs are larger molecules with a significantly greater Ca 2§ conductance than IP 3 receptors and are found in cells such as muscle that generate large, fast changes in intracellular Ca 2§ concentration. Three main types of RyR have been identified with tissuespecific distributions. The predominant isoform in mammalian skeletal muscle is RyR1. RyR2 is the major isoform in heart and is also widely distributed in the brain. RyR3 is present in brain, epithelium and selected muscles of some species. The RyR isoforms are products of different genes and share 65-70% homology. Alternatively spliced transcripts have also been detected. The skeletal RyR1 isoform is associated either directly or indirectly with a number of other proteins and therefore must be viewed as a larger complex when considering its functional interactions. RyR1 interacts primarily with the DHPR on the cytoplasmic side and with calsequestrin on the luminal side of the jSR. In addition, it interacts with other components of the jSR including junctin, FKBP12 protein (FK506 binding protein), triadin, and a novel 90-kDa jSR protein that may associate with RyR1 in fast-twitch skeletal muscle (Froemming et al., 1999). Calsequestrin is localized on the luminal jSR membrane, closely associated with the RyR. Its main function is to increase the total capacity of the SR for Ca 2+. Junctin is a small calsequestrin-binding protein that is enriched in the jSR and may anchor calsequestrin to the membrane. FKBP12 protein (FK506 binding protein) is stably bound to the comers of the cytosolic domain of the RyR in a 1:1 ratio and may mediate an endto-end interaction. Triadin is present in high abundance in jSR and also binds the RyR in a 1:1 ratio. Its functional role is not known. The RyRs are extremely large molecules (N5000 amino acids and--560 kDa). The tetramer assembles to form a molecule of approximately 30 x 30 x 12 nm size with a fourleaf clover-shaped extracellular domain and a smaller intramembrane domain with a central pore. Figure 10 shows a three-dimensional reconstruction of RyR1 in closed and open conformations. Sequence analysis indicates a large hydrophilic N-terminal region thought to constitute the cytoplasmic or foot domain and a smaller mostly hydrophobic C-terminal region predicted to form the intramembrane channel. Both N- and C-termini are predicted to be cytoplasmic. The foot region contains four repeat motifs that occur in two tandem pairs. The membranecrossing region is highly conserved and has strong similarity with the same region of the IP 3 receptor. The number of membrane crossings of the C-terminal region has not been established, but is proposed to be between 4 and 10. The skeletal RYR1 is capable of dual modes of activation, both allosteric and ligand gating. It can be allosteri-
54. Excitation-Contraction Coupling in Skeletal Muscle
9 19
FIGURE 10. The open and closed states of the skeletal muscle RyR/Ca2+release channel are shown in three different views: top, side, and bottom. The putative transmembrane and cytoplasmic (TM and CY) parts of the protein resemble the stem and cap of a mushroom-shaped protein. The cytoplasmic side of the tetramer contains clamp-shaped and handle (C and H) domains which alternately touch and overlap in the open and closed conformations. The cytoplasmic side of the tetramer contains the regions of spatial superposition with the T-tubule tetrads/DHPRs. The transmembrane region of the channel appears open towards the SR, whereas in the closed state a central opening is not seen in this region. (Reproduced from The Journal of General Physiology, 1996, 106, 659-704, by copyright permission of The Rockefeller University Press.)
cally opened by an associated skeletal DHPR or opened by direct Ca 2+ binding when not associated with a skeletal DHPR. When isolated and purified in the absence of triad junctions or DHPRs, the skeletal RYR1 behaves as a ligand-gated ion channel with a high single-channel ion flux. Ca 2+ and adenine nucleotides at mM levels open the pore, and mM Mg 2+ inhibits it. In this respect, the skeletal RyR1 is similar to cardiac RyR2, although its Ca 2+ sensitivity is lower. Studies in which DHPRs have been inserted into dysgenic myotubes that congenitally lack DHPRs (see also Section V.C) have demonstrated that RyR1 can be activated by Ca 2+ entering through DHPRs when the inserted DHPR contains key cardiac-specific sequences. On the other hand, RyR1 can be opened allosterically and mediate a skeletal-type EC coupling (independent of extracellular Ca 2+) only when associated with a skeletal-type DHPR. Thus, the complete skeletal DHPR-RyR1 complex is required to initiate Ca 2+ release by the physiologically relevant mode in fast-twitch skeletal muscle.
V. M e c h a n i s m s of Interaction b e t w e e n DHPRs and RyRs At the triad junction a unique intracellular signal transduction occurs whereby a voltage-gated Ca 2+ channel on an extracellular membrane controls the opening of a different, normally ligand-gated Ca 2+ channel on an intracellular membrane. The essential steps in signal transduction at the triad junctions are detection of the action potential depolarization by voltage-sensing molecules/DHPRs in the
T-tubule, communication from the voltage sensor/DHPR to the Ca 2+ release channel/RyR1 of the SR, and Ca 2+ efflux from the SR.
A. Voltage Sensing It was recognized early that EC coupling in fast skeletal muscle of higher vertebrates is directly controlled by membrane potential, without a requirement for extracellular Ca 2+. This contrasts with EC coupling in skeletal muscle of some lower invertebrates and cardiac muscle, in which Ca 2+ influx across the sarcolemma/through sarcolemmal DHPRs is absolutely required for contraction. In cardiac muscle the Ca 2+ that enters through the sarcolemmal DHPR functions as a second messenger to activate the cardiac RyR2, a process termed calcium-induced calcium release (CICR). In a classic experiment, Armstrong and colleagues (1972) demonstrated that a frog skeletal muscle fiber continues to twitch in the absence of extracellular Ca 2+ when stimulated with action potentials. Similarly, voltage-clamp experiments demonstrated that a skeletal muscle fiber can develop tension provided only that the membrane is depolarized to a minimum potential and duration, termed the mechanical threshold. That is, when the Na +, K +, and C1- ion channels that mediate the action potential are blocked, all that is required to activate contraction is a suprathreshold change in membrane potential. The relationship between tension and membrane potential is very precise in skeletal muscle as shown in Fig. 11. Subsequently, Schneider and Chandler (1973) discovered the presence in skeletal muscle membranes of mobile
920
S E C T I O N VI M u s c l e a n d O t h e r C o n t r a c t i l e S y s t e m s
1 O0 ms
3.0
-----'--'~
-
'
~
50 ms
Od
E
k~ 2.O
_o O9 Z UJ v < i.U s
tO
/ 0.0
INTERNAL POTENTIAL CHANGE: A V i (mV)
FIGURE 1 1. Relationshipbetween peak tension and membrane potential in the absence of an action potential. The fibers were depolarized by holding the membrane potential constant for 50 ms with a voltage-clamp. Tension was measured simultaneously with a transducer attached to one tendon. (Reproduced from The Journal of General Physiology, 1984, 84, 133-154, by copyright permission of The Rockefeller University Press.) intramembrane charges whose movement could be detected electrically as a voltage-dependent dielectric current, termed c h a r g e m o v e m e n t . Based on the similarity of the kinetics and voltage dependence of charge movement to mechanical activation (Fig. 12), they proposed that it reflected the movement of charged intramembrane domains of a molecule in the T-tubules that functioned as the essential voltage sensor for EC coupling. They further suggested that the voltage sensor might interact directly with
30-
2O LI_
/
C5 (-
v
0
10
t -100
I ,, ,! -80 -60
t -40
~ ~,,, -20
I 0
I 20
I 40
!. 60
I 80
Test potential (mV)
FIGURE 12. Relationship between charge movement (nC/]aF) and membrane potential (mV). Charge moved at each potential was obtained by integrating the voltage-dependent capacity current, normalized to the fiber linear capacitance near the resting potential. Symbols represent the mean values of charge moved at the "on" and "off" of the pulse. The continuous curve is a fit of the data to a two-state Boltzmann function, with parameters: Vmid =-47.7 mV, k = 8 mV, and Qmax = 21.5 nC/]aE (From Chandler et al., 1976.)
a hypothetical Ca2+-conducting protein on the SR to cause its opening. The mechanical or allosteric hypothesis of EC coupling (electromechanical coupling) proposes that movement of charged voltage-sensing domains of the skeletal DHPR allosterically alters an activation domain of RyR1, either directly or via an intermediate protein. In this model, the triad junction exists to bring together the key proteins in a macromolecular complex. This early allosteric model of EC coupling, formulated before the identity of the proteins of the triad junction had been established, has formed the framework for subsequent investigations into the nature of the molecular interactions at the triad junction complex. Several lines of evidence led to the identification of the voltage sensor as the DHPRff-tubule Ca 2+ channel which was shown to be localized in skeletal muscle T-tubules at one of the highest densities found in nature (rev. in Rios and Pizarro, 1991; Schneider, 1994; Melzer et al., 1995). A definitive identification came following the cloning of the skeletal muscle Ca 2+ channel (Tanabe et al., 1987) and its use in experiments on dysgenic mice (rev. in Adams and Beam, 1990). Muscular dysgenesis arises from a single lethal mutation in the gene for the DHPR asubunit, which results in reduction of DHPR protein, absence of slow Ca 2+ currents and charge movement, and failure of EC coupling. Tanabe and colleagues (1988) demonstrated that Ca 2+ currents and EC coupling could be restored in dysgenic muscle fibers transfected with the cDNA for the a-subunit of the DHPR, confirming the essential role of the DHPR as the voltage sensor in signal transduction to the SR.
B. Molecular Interactions b e t w e e n the DHPR and RyR The molecular mechanism by which the Ca2+-conducting pore of the RyR moves from a closed to open configuration under control of the T-tubule Ca 2+ channel is the subject of intense research (rev. in Meissner and Lu, 1995; Leong and MacLennan, 1998a). The unique structure of the R y R m a Ca 2+ channel domain in the SR bilayer and a huge cytosolic domain that spans most or all of the junctional gap--suggests that opening and closing of the pore could be remotely controlled by a cytosolic region that makes contact with the DHPR across the junctional gap. For example, the DHPR in its resting state could exert an inhibitory influence on the RyR/Ca 2+ release channel that could be removed upon depolarization, or the RyR could be activated by conformational movements of the DHPR during depolarization. Inhibition of the RyR by the DHPR in the resting state and its removal by depolarization is attractive because of the inherent speed of this type of on-but-inhibited switching mechanism and because positively charged transmembrane helices in the $4 domain of the DHPR would move outward in the T-tubule membrane upon depolarization, away from the RyR. The following sections and Fig. 13 summarize the major domains of the skeletal DHPR ce subunit that have been shown to mediate EC coupling interactions with the RyR.
C. Information from Genetic and Transgenic Models Results from studies with dysgenic mice first identified specific domains of the DHPR that can activate RyR1 in in-
54. Excitation-Contraction Coupling in Skeletal Muscle
921
FIGURE 13. Sequences of the DHPR identified with functional roles in EC coupling. $4 voltage-sensing domains are indicated with a positive charge. Pore sequences between transmembrane segments $5 and $6 are indicated in bold. The 138amino-acid cytosolic linker between domains II-III is a key interaction domain with the RyR and is essential for activating skeletal-type EC coupling. An Arg residue in the III-IV cytosolic linker of the human DHPR is critical for controlling Ca2§ release from the skeletal RyR 1 and causes malignant hyperthermia when mutated. The I-II linker is the site of interaction with the/3-subunit and influences EC coupling. Inset: Within the II-III linker a subdomain from approximately residues 666-726 is important for activation of RyR1 by the skeletal DHPR and contains a cluster of five basic residues (striped) that are required for the interaction. Residue Ser 687 inhibits RyR1 activation when phosphorylated. A second overlapping region from approximately residues 711-760 is important for both activation and blocking of RyR1 by the DHPR, and includes a region that receives a retrograde signal from RyR 1 to enhance the function of DHPRs as calcium channels. Numbering corresponds to the rabbit skeletal muscle DHPR sequence.
tact muscle under physiological conditions. The ce subunits of the L-type Ca 2+ channel/DHPR of skeletal and cardiac muscle share a high degree of homology and both play a key role in triggering Ca 2+ release. The main differences in primary structure between the skeletal and cardiac isoforms reside in the large cytosolic regions at the N- and C-termini, and the linker sequences II-III and III-IV. A key functional difference is that Ca 2+ entry through the channel is required to elicit Ca 2+ release in cardiac but not in skeletal muscle. Beam and collaborators (Tanabe et al., 1990) exploited these differences to identify molecular regions of the ce1 subunit of the DHPR that are critical for initiating a skeletal-type EC coupling. When the skeletal muscle a ls subunit was expressed in dysgenic myotubes, a Ca 2+ release independent of Ca 2+ entry could be elicited in response to electrical depolarization, whereas Ca 2+ entry was required to elicit Ca 2+ release when the cardiac c~lc subunit was expressed. By constructing and expressing chimeric DHPRs composed of various cardiac and skeletal domains, Tanabe and collaborators (1990) demonstrated that the 138-amino-acid cytoplasmic linker between domains II and III is a critical determinant of skeletal-type EC coupling. When a chimeric DHPR having the cardiac protein backbone but the skeletal II-III linker was expressed (see Fig. 13), contraction could be
elicited by electrical stimulation even when Ca 2+ entry was blocked. This indicates that the II-III linker is involved in signal transmission from the skeletal DHPR to RyR1 on the SR. Interestingly, expression of both the wild-type skeletal and cardiac DHPRs as well as the chimeric channels all led to the restoration of charge movement. This indicates that the voltage-sensing step in EC coupling does not depend on the nature of the transmission mechanism, and likely involves separate molecular domains of the a subunit. Other cytosolic regions on the a subunit of the skeletal DHPR have also been implicated in direct interactions with the RyR 1. A mutation in the I I I - I V cytosolic linker of the human DHPR (A1086H) is associated with susceptibility to one form of malignant hyperthermia (MH), a hereditary skeletal muscle disorder that is characterized by an uncontrolled, sustained Ca 2+ release (Monnier et al., 1997). This suggests that the I I I - I V linker also participates in control of RyR1 activation by the skeletal DHPR, or, alternatively, it may participate in turning off Ca 2+ release under control of the skeletal DHPR. Earlier studies have shown that Ca 2+ release, which is initiated by membrane depolarization, is also terminated upon membrane repolarization, under tight control of the voltage sensor (rev. in Rios and Pizzaro, 1991; Suda, 1995). The mutation in MH
922 may delay or prevent closure of RyR1 upon repolarization (Leong and MacLennan, 1998a). A minor role of the I-II cytosolic linker in determining the mode of EC coupling has also been shown (De Waard et al., 1996). EC coupling is blocked when the DHPR 13 subunit gene, which binds to the I-II linker of skeletal DHPR, was knocked out. Expression of/~la but not the predominantly cardiac ~2a isoform in/3-null myotubes could fully restore a skeletal-type EC coupling, suggesting that the interaction of/3 la with skeletal DHPR is required for a skeletal, Ca2+-independent type of EC coupling. D. Evidence from in Vitro Studies Information from in vitro studies supports the existence of protein-protein interactions between the skeletal DHPR and RyR (rev. in Meissner and Lu, 1995; Leong and MacLennan, 1998a). Early biochemical evidence for a direct association between proteins in the T-tubule and the SR came from the observation that purified fractions of T-tubule membrane could form junctions with SR membranes and that triad fractions could be isolated intact, despite stringent fractionation procedures (Caswell et al., 1979). Strong protein-protein associations of some kind at the triad junction would be required to preserve the structural integrity of the triad junction as it sustains the strong forces of contraction. Further studies have confirmed the importance of the II-III cytoplasmic linker in conferring skeletal-type EC coupling. Peptides composed of either the skeletal or cardiac ce subunit II-III linker sequence can activate the skeletal RyR1 by increasing open channel probability and the affinity of ryanodine binding (Lu et al., 1994). The peptides did not activate the cardiac RyR2 Ca 2+ release channel. This suggests that the skeletal isoforms of both the RyR and the DHPR are required to achieve a skeletal-type EC coupling. Studies with shorter peptides have identified subdomains of the II-III linker which are essential to activate Ca 2§ release (see Fig. 13, inset). Amino acid segments within the region 666-725 confer skeletal-type EC coupling (666-726: Lu et al., 1995; 671--690: E1-Hayek et al., 1995; 1998). A 25amino-acid segment (Glu 666-Pro 691) within this region of a ls, but not the corresponding cardiac sequence, is able to activate the skeletal ryanodine receptor. A cluster of five basic residues within this segment (681RKRRK685) is absolutely required to activate the skeletal RyR1 Ca 2§ release channel (Zhu et al., 1999). A second overlapping region from amino acids 711-760 appears to be important both for activation and blocking of RyR1 activation by the DHPR (711-742: Nakai et al., 1998b; 725-742: Saiki et al., 1999). Conversely, as expected for direct protein-protein interactions, regions on RyR1 that interact specifically with the skeletal DHPR have also been identified. An N-terminal sequence on RyR1 (Leu 922-Asp 1112) interacts strongly with the II-III linker and weakly with the III-IV linker of the skeletal DHPR, suggesting that the DHPR linkers bind to contiguous and possibly overlapping sites on RyR1. A 37amino-acid segment within this region of RyR1, but not the corresponding cardiac sequence on RyR2, binds specifically to the II-III linker of the skeletal DHPR (Leong and MacLennan, 1998b).
SECTIONVI Muscle and Other Contractile Systems Much useful information on the skeletal RyR1 has come from studies using mice homozygous for a disrupted RyR1 gene (dyspedic mice). Dyspedic mice lack EC coupling, as expected, and in addition have a thirty-fold lower density of calcium channel current (Nakai et al., 1996). Expression of the cDNA for RyR1 in dyspedic myotubes restores EC coupling and enhances the calcium channel current. This suggests that RyR1, in addition to receiving the EC coupling signal from the DHPR, also transmits a retrograde signal that enhances the function of skeletal DHPRs as calcium channels. The interactions between the DHPR and the RyR leading to EC coupling are specific for the skeletal RyR1, indicating that the identity of the RyR isoform may determine the type of EC coupling. Expression of the predominant cardiac isoform, RyR2, did not restore skeletal-type gating of Ca 2+ release by the DHPR or retrograde signaling to enhance calcium channel activity (Nakai et al., 1997). By expressing chimeras of the skeletal RyR1 and cardiac RyR2, Nakai and collaborators (1998a) identified distinct regions of RyR1 that are involved in the reciprocal interactions of RyR1 with the skeletal DHPR. Residues 1635-2636 of RyR1 mediated skeletal-type EC coupling and enhanced calcium channel function, whereas a chimera containing adjacent RyR1 residues (2659-3720) was only able to enhance calcium channel function. Residues 720-765 within the II-III linker of als are required for receiving the feedback signal from RyR 1 that leads to enhancement of Ca 2§ current (Grabner et al., 1999). E. Evidence from in Vivo Studies Early functional studies demonstrated that manipulations that alter T-tubule Ca 2+ channel function influence SR Ca 2§ release, and conversely that manipulations that affect Ca 2+ release alter the function of the DHPR. Such bidirectional cross-talk between the DHPR and RyR, involving both forward and retrograde interactions, is expected for an allosteric interaction along a macromolecular complex. In studies using intact skeletal muscle cells, conformational transitions of the DHPR are detected as charge movement and SR Ca 2+ release is detected by monitoring myoplasmic Ca 2+ transients using Ca 2§ indicator dyes introduced into the myoplasm. Much useful information on the kinetics and voltagedependence of the voltage-sensing and Ca 2+ release steps under physiological conditions has come from these studies (rev. in Rios and Pizarro, 1991; Melzer et al., 1995). Additional information has come from investigating the actions of pharmacologic agents that are known to alter EC coupling in skeletal muscle, including caffeine and perchlorate (C104-). C104- is a chaotropic anion that is a potent and specific potentiator of EC coupling in skeletal muscle. It shifts the voltage dependence of charge movement by 15 to 30 mV to more negative voltages, slows the kinetics of both on and off charge movements, and reduces the minimum threshold charge required to activate Ca 2§ release. It was demonstrated early that several of the actions of C104- could be explained by a simple model in which the two proteins interact allosterically as shown in Fig. 14 (Rios et al., 1993). In this model, the RyR/Ca 2§ release channel is a homotetramer
54. Excitation-Contraction Coupling in Skeletal Muscle
923
9
-
"~ .j'4. ,~.':
e. 9
FIGURE 14. Allostericmodel of EC coupling. Four independent voltage-sensor molecules/DHPRs in the T-tubule membrane compose a tetrad that is aligned with and contacts one RyR/Ca2§ release channel, depicted as a homotetramer. Two states of the release channel (closed and open) are pictured, as well as two of five possible dispositions of the dihydropyridine receptor tetrad. (A) All DHPRs in the inactive conformation. (B) One DHPR in the active conformation and three in the inactive state. (Reproduced from The Journal of General Physiology, 1993,102, 449-481, by copyright permission of The Rockefeller University Press.)
that can assume two conformations, open and closed. The four DHPRs aligned above it (one tetrad) act as heterotropic li-gands to modulate this transition. Each shift of one voltage sensor/DHPR to its fully active state increases the probability that the RyR will open. In this model, CIO 4- acts primarily on the RyR/Ca 2+ release channel to increase its probability of opening and secondarily on the DHPR via allosteric feedback. As part of this feedback to the DHPR, one kinetic component of charge movement, termed Q. charge, could be simulated. A related hypothesis in this model is that released Ca 2§ ions in the triadic gap modify the voltagesensing process in a positive-feedback manner to generate additional charge movement. Thus, both allosteric interactions between the RyR and the DHPR, and Ca 2§ ions released into the triadic gap may contribute to the generation of the Qe charge movements detected electrically from intact muscle. Other studies have likewise shown that Ca 2+ release can influence the DHPR and vice-versa, but have led to a somewhat different model of the interaction (rev. in Jong et al., 1996). An alternate proposal is that the steep voltage dependence of Ca 2§ release is an intrinsic function of the voltage sensor in the membranes of the T-system, without requiring additional positive feedback from Ca 2+ release. In this model, the SR Ca 2§ content or release or a related process directly alters the kinetics of charge movement, but does not generate a separate species of charge. Both of these findings suggest a close association between the DHPR and the SR Ca 2+ release channel. Although Ca 2+ entry is not required for contraction in skeletal muscle, the fact that Ca 2§ release and/or the RyR through retrograde allosteric interactions can modify the slow Ca 2§ current also supports the idea of a close interaction between RyR1 and the skeletal DHPR. Intracellular BAPTA,
a rapid Ca 2+ buffer that is capable of buffering released Ca 2+ ions in the triadic space near the release sites (Jong et al., 1996) eliminates the slow Ca 2+ current (Feldmeyer et al., 1993).
F. Overall Control and Integration of Release Events
C a 2+
Although there is general agreement that the initial event leading to activation of calcium release is a depolarizationdriven allosteric interaction between a skeletal DHPR tetrad and one RyR1 (electromechanical coupling), the mechanism by which this signal is communicated to adjacent RyRs and perhaps amplified or synchronized with other RyRs, is less understood. The position of DHPR tetrads alternating with every other RyR1, and the packing of RyRls in highly ordered arrays possibly touching at the comers suggests a dual control of RyR1 activation in which RyRls linked to a skeletal DHPR are activated allosterically by the DHPR while neighboring RyRs are ligand gated by binding of released Ca 2+ (Fig. 15A). The latter is analogous to Ca2+-induced Ca 2§ release in heart muscle, with the source of activator Ca 2§ being intracellular rather than extracellular. According to the dual control theory, each RyR is opened independently either by allosteric or ligand gating and the global myoplasmic change arises from the summation of thousands of individual single Ca 2§ release events from both populations of RyRs. An alternate model of RyR gating, termed coordinated gating, proposes that clusters of physically linked RyRs open simultaneously in a coordinated fashion (Fig. 15B; Marx et al., 1998). In the later model, clusters of RyRs operate synergistically as all-or-none Ca 2§ release units (all
924
SECTION VI Muscle and Other Contractile Systems
F I G U R E 15. Two models of integration of Ca 2+ release from RyRs. A tetrad of four DHPRs is shown aligned above one RyR1. The adjacent RyR is not associated with a DHPR tetrad and is not necessarily attached to the first RyR. (A) Dual gating model of RyR1 activation. Upon depolarization, one DHPR tetrad allosterically activates its associated RyR; the resulting local Ca 2+ change activates the unassociated RyR. Thus, one population of RyRls is gated allosterically while the second population of RyR 1s is ligand gated by released Ca 2+. The overall myoplasmic Ca 2+ change results from the summation of individual Ca 2§ release events from both populations of RyRls. (B) Coordinated gating model. A tetrad of four DHPRs is aligned above one RyR1. The DHPR-associated RyR and adjacent RyRs are linked to each other in such a way that one DHPR and a cluster of RyRs form a functional macromolecular complex. Upon depolarization, one DHPR tetrad allosterically activates its associated RyR and that allosteric change is transmitted down the complex to activate simultaneously all RyRs in the cluster. In this model, both populations of RyR 1s, those associated and those unassociated with DHPR tetrads, are activated by the DHPR without intermediate Ca 2+ ligand. The overall myoplasmic Ca 2+ change results from the coordinated gating of linked RyRls.
channels in a cluster open or close together). Opening of one RyR1 by its linked DHPR results in simultaneous opening of all contiguous RyRs in the cluster. RyRs operating as functional units would be expected to generate faster local Ca 2+ release events and a more synchronous global change in myoplasmic Ca. Small, local Ca 2+ release events, termed Ca 2+ sparks, have been detected intracellularly in skeletal and cardiac muscle using Ca 2+ indicator dyes (Cheng et al., 1993; Tsugorka et al., 1995). It is generally thought that the Ca 2+ spark is the elementary unit of SR Ca 2+ release. The Ca 2+ spark, which exhibits a stereotypical amplitude and duration, may arise from the stochastic opening of several RyRs or may represent the coordinated opening of a functional RyR cluster.
G. Modulation of EC Coupling There is growing evidence that EC coupling can be modulated under physiological conditions, and a number of potential sites and mechanisms have been identified. The skeletal DHPR ce subunit contains a number of consensus phosphorylation sites including a serine residue, Ser 687 within the critical II-III loop that is rapidly phosphorylated by a cAMP-dependent kinase. Phosphorylation of this residue is required for binding of the II-III loop to RyR1 (but not RyR2) but only the dephosphorylated II-III loop can activate the RyR (Lu et al., 1995). In studies using dysgenic myotubes, substitution of Ser 687 with alanine, the corresponding cardiac residue, did not abolish skeletal-type EC coupling (Nakai et al., 1998b). Together, these data suggest that phosphorylation of the DHPR II-III linker may be a potential regulatory mechanism for EC coupling. The a ls subunit contains sites that can inhibit as well as activate Ca 2+ release from the RyR (residues 671-690 in the II-III
linker: E1-Hayek et al., 1995; residues 1487-1506 in the Cterminus: Slavik et al., 1997). Similarly, subdomains of RyR1 have been identified that participate in the modulation, inhibition, or inactivation of Ca 2§ release. In addition to the regions of RyR1 involved in retrograde enhancement of DHPR function as a Ca 2§ channel, it has been shown that the C-terminal one-fourth of RyR1 has a major role in both the activation and inactivation of RyR1 by Ca 2§ (Nakai et al., 1999; Du and MacLennan, 1999). Other proteins of the triad may also modulate EC coupling as well as participate in the structural architecture of the junction. The triad region of skeletal muscle forms supramolecular membrane complexes that include, in addition to the DHPR and RyR1, FKBP12, calsequestrin, junctin, triadin, and possibly other proteins. Finally, a long-term regulation of EC coupling may occur at the level of gene expression. Studies have shown that the expression of eels can be altered during adaptation to physiological demands, including altered patterns of nerve activity, exercise, changes in the loadbearing state and metabolic or endocrine status (Ray et al., 1995; Kandarian et al., 1996; Pereon et al., 1997).
VI. Summary Excitation-contraction coupling in skeletal muscle is a fast signal transduction process by which an action potential on the sarcolemmal/T-tubule membranes is transduced into a biochemical trigger for the contractile proteins to shorten and generate force. The immediate intracellular trigger is Ca 2§ released from intracellular stores. Contraction is inhibited when cytosolic Ca 2+ is at resting, submicromolar levels, and proceeds when Ca 2§ rises to micromolar levels. The key transduction steps that link membrane depolarization to Ca e§ release from the SR occur at specialized
54. Excitation-Contraction Coupling in Skeletal Muscle
junctions that bring together the sarcolemmal and intracellular SR membranes. At the junctions, a sarcolemmal Ca 2+ c h a n n e l / D H P R senses the depolarization during the action potential and c o m m u n i c a t e s it to the Ca 2+ release channels/RyRs on the SR, resulting in the release of stored Ca 2+. The molecular m e c h a n i s m s of interaction b e t w e e n these two distinct Ca 2+ c h a n n e l s involves an allosteric process. D H P R s and R y R s associate at the junctions to form a functional m a c r o m o l e c u l a r c o m p l e x that is capable of distinct molecular interactions at identified sites. Additional associated junctional proteins m a y participate in or modulate this interaction. The result of this signal transduction is a fast, coordinated activation of muscle contraction.
Bibliography Adams, B. A., and Beam, K. G. (1990). Muscular dysgenesis in mice: a model system for studying excitation-contraction coupling. [Review]. FASEB J. 4, 2809-2816. Anderson, K., Cohn, A. H., and Meissner, G. (1994). High-affinity [3H]PN200-110 and [3H]ryanodine binding to rabbit and frog skeletal muscle. Am. J. of Physiol. 266, C462-C466. Armstrong, C. M., Bezanilla, E M., and Horowicz, P. (1972). Twitches in the presence of ethylene glycol bis(-aminoethyl ether)-N,N'tetracetic acid. Biochim. Biophys. Acta 267, 605-608. Beam, K. G., Tanabe, T., and Numa, S. (1989). Structure, function, and regulation of the skeletal muscle dihydropyridine receptor. [Review]. Ann. NY Acad. Sci. 560, 127-137. Beurg, M., Sukhareva, M., Ahem, C.A., Conklin, M.S., Perez-Reyes, E., Powers, P.A., Gregg, R.G., Coronado, R. (1999). Differential regulation of skeletal muscle L-type Ca2+ current and excitationcontraction coupling by the dihydropyridine receptor beta subunit. Biophys. J. 76, 1744-1756. Block, B. A., Imagawa, T., Campbell, K. E, and Franzini-Armstrong, C. (1988). Structural evidence for direct interaction between the molecular components of the transverse tubule/sarcoplasmic reticulure junction in skeletal muscle. J. Cell Biol. 107, 2587-2600. Caputo, C., Bezanilla, E, and Horowicz, P. (1984). Depolarizationcontraction coupling in short frog muscle fibers. A voltage clamp study. J. Gen. Physiol. 84, 133-154. Caswell, A.H., Lau, Y.H., Garcia, M., and Brunschwig, J.E (1979). Recognition and junction formation by isolated transverse tubules and terminal cisternae of skeletal muscle. J. Biol. Chem. 254, 202-208. Chandler, W. K., Rakowski, R. E, and Schneider, M. E (1976). A nonlinear voltage-dependent charge movement in frog skeletal muscle. J. Physiol. 254, 243-283. Cheng, H., Lederer, W. J., and Cannell, M. B. (1993). Calcium sparks: elementary events underlying excitation-contraction coupling in heart muscle. Science 262, 740-744. De Waard, M., Gurnett, C. A., and Campbell, K. E (1996). Structural and functional diversity of voltage-activated calcium channels. Ion Channels 4, 41-87. Du, G. C. and MacLennan, D. H. (1999). Ca(2+) inactivation sites are located in the COOH-terminal quarter of recombinant rabbit skeletal muscle Ca(2+) release channels (ryanodine receptors). J. Biol. Chem. 274, 26 120-26 126. Dulhunty, A. E (1984). Heterogeneity of T-tubule geometry in vertebrate skeletal muscle fibres. J. Muscle Res. Cell Motility 5, 333-347. E1-Hayek, R., and Ikemoto, N. (1998). Identification of the minimum essential region in the II-III loop of the dihydropyridine receptor alpha 1 subunit required for activation of skeletal muscle-type excitationcontraction coupling. Biochem. 37, 7015-7020.
925 E1-Hayek, R., Antoniu, B., Wang, J., Hamilton, S. L., and Ikemoto, N. (1995). Identification of calcium release-triggering and blocking regions of the II-III loop of the skeletal muscle dihydropyridine receptor. J. Biol. Chem. 270, 22 116-22 118. Feldmeyer, D., Melzer, W., Pohl, B., and Zollner, E (1993). A possible role of sarcoplasmic Ca2+ release in modulating the slow Ca2+ current of skeletal muscle. Pfliigers Arch. 425, 54-61. Franzini-Armstrong, C., and Jorgensen, A. O. (1994). Structure and development of E-C coupling units in skeletal muscle. [Review]. Ann. Rev. Physiol. 56, 509-534. Franzini-Armstrong, C., and Kish, J. W. (1995). Alternate disposition of tetrads in peripheral couplings of skeletal muscle. J. Muscle Res. Cell Motility 16, 319-324. Franzini-Amastrong, C., and Peachly, L. D. (1981). Striated muscle-contractile and control mechanisms. [Review].J. CellBiol. 91,166s-186s. Franzini-Armstrong, C., and Protasi, E (1997). Ryanodine receptors of striated muscles: a complex channel capable of multiple interactions. Physiol. Rev. 77, 699-729. Franzini-Armstrong, C., Protasi, E, and Ramesh, V. (1999). Shape, size and distribution of Ca2+ release units and couplons in skeletal and cardiac muscles. Biophys. J. 77, 1528-1539. Froemming, G. R., Pette, D., and Ohlendieck, K. (1999). The 90-kDa junctional sarcoplasmic reticulum protein forms an integral part of a supramolecular triad complex in skeletal muscle. Biochem. Biophys. Res. Commun. 261, 603-609. Gonzalez, A., and Rios, E. (1993). Perchlorate enhances transmission in skeletal muscle excitation-contraction coupling. J. Gen. Physiol. 102, 373-421. Gonzalez-Serratos, H. (1971). Inward spread of activation in vertebrate muscle fibres. J. Physiol. (Lond.) 212, 777-799. Grabner, M., Dirksen, R. T., Suda, N., and Beam, K. G. (1999). The II-III loop of the skeletal muscle dihydropyridine receptor is responsible for the bi-directional coupling with the ryanodine receptor. J. Biol. Chem. 274, 21 913-21 919. Gregg, R. G., Messing, A., Strube, C., Beurg, M., Moss, R., Behan, M., Sukhareva, M., Haynes, S., Powell, J. A., Coronado, R., Powers, P. A. (1996). Absence of the beta subunit (cchb 1) of the skeletal muscle dihydropyridine receptor alters expression of the alpha 1 subunit and eliminates excitation-contraction coupling. Proc. Natl. Acad. Sci. USA 93, 13 961-13 966. Hill, A. V. (1948). On the time required for diffusion and its relation to processes in muscle. Proc. R. Soc. B 135, 446-453. Hodgkin, A. L., and Horowicz, P. (1957). The differential action of hypertonic solutions on the twitch and action potential of a muscle fibre. J. Physiol. 136, 17-18. Hollingworth, S., Zhao, M., and Baylor, S. M. (1996). The amplitude and time course of the myoplasmic free [Ca2+] transient in fasttwitch fibers of mouse muscle. J. Gen. Physiol. 108, 455-469. Jong, D. S., Pape, P. C., Geibel, J., Chandler, W. K. (1996). Sarcoplasmic reticulum calcium release in frog cut muscle fibers in the presence of a large concentration of EGTA. Soc. Gen. Physiol. Ser. 51, 255-268 Kandarian, S. C., Peters, D. G., Favero, T. G., Ward, C. W., and Williams, J. H. (1996). Adaptation of the skeletal muscle calciumrelease mechanism to weight-bearing condition. Am. J. Physiol. 270, C1588-C1594. Leong, E, and MacLennan, D. H. (1998a). Complex interactions between skeletal muscle ryanodine receptor and dihydropyridine receptor proteins. Biochem. Cell Biol. 76, 681-694. Leong, E, and MacLennan, D. H. (1998b). A 37-amino-acid sequence in the skeletal muscle ryanodine receptor interacts with the cytoplasmic loop between domains II and III in the skeletal muscle dihydropyridine receptor. J. Biol. Chem. 273, 7791-7794. Lu, X., Xu, L., and Meissner, G. (1994). Activation of the skeletal muscle calcium release channel by a cytoplasmic linker of the dihydropyridine receptor. J. Biol. Chem. 269, 6511-6516.
926
Lu, X., Xu, L., and Meissner, G. (1995). Phosphorylation of dihydropyridine receptor II-III linker peptide regulates skeletal muscle calcium release channel function. Evidence for an essential role of the beta-OH group of Ser687. J. Biol. Chem. 270, 18 459-18 464. Marx, S. O., Ondrias, K., and Marks, A. R. (1998). Coupled gating between individual skeletal muscle Ca 2§ release channels. Science 281, 818-821. Mcpherson, P. S., and Campbell, K. P. (1993). The ryanodine receptor/Ca 2§ release channel. J. Biol. Chem. 268, 13 765-13 768. Meissner, G., and Lu, X. (1995). Dihydropyridine receptor-ryanodine receptor interactions in skeletal muscle excitation-contraction coupling. Biosci. Rep. 15, 399407. Melzer, W., Herrmann-Frank, A., and Luttgau, H. C. (1995). The role of Ca 2+ ions in excitation-contraction coupling of skeletal muscle fibres. [Review]. Biochim. Biophys. Acta 1241, 59-116. Monnier, N., Procaccio, V., Stieglitz, P., and Lunardi, J. (1997). Malignant-hyperthermia susceptibility is associated with a mutation of the alpha 1-subunit of the human dihydropyridine-sensitive Ltype voltage-dependent calcium-channel receptor in skeletal muscle. Am. J. Hum. Genet. 60, 1316-1325. Nakai, J., Gao, L., Xu, L., Xin, C., Pasek, D. A., and Meissner, G. Evidence for a role of C-terminus in Ca(2+) inactivation of skeletal muscle Ca(2+) release channel. (1999). FEBS Lett. 459, 154-158. Nakai, L. J., Dirkson, R. T., Nguyen, H. T., Pessah, I. N., Beam, K. G., and Allen, P. D. (1996). Enhanced dihydropyridine receptor channel activity in the presence of ryanodine receptor. Nature 380, 72-75. Nakai, J., Ogura, T., Protase, E, Franzini-Armstrong, C., Allen, P. D., and Beam, K. G. (1997). Functional nonequality of the cardiac and skeletal ryanodine receptors. Proc. Natl. Acad. Sci. USA 94, 1019-1022. Nakai, J., Sekiguchi, N., Rando, T. A., Allen, P. D., and Beam, K. G. (1998a). Two regions of the ryanodine receptor involved in coupling with L-type Ca 2+ channels. J. Biol. Chem. 273, 13 403-13 406. Nakai, J., Tanabe, T., Konno, T., Adams, B., and Beam, K. G. (1998b). Localization in the II-III loop of the dihydropyridine receptor of a sequence critical for excitation-contraction coupling. J. Biol. Chem. 273, 24 983-24 986. Neuhuber, B. Gerster, U., Mitterdorfer, J., Glossmann, H., Flucher, B. E. (1998). Differential effects of Ca 2+ channel betala and beta 2a subunits on complex formation with alpha l S and on current expression in tsA201 cells. J. Biol. Chem. 273, 9110-9118. Orlova, E. V., Serysheva, I. I., Van Heel, M., Hamilton, S. L. & Chiu, W. (1996). Two structural configurations of the skeletal muscle calcium release channel. J. Gen. Physiol. 106, 659-704. Oz, M., and Frank, G. B. (1991). Decrease in the size of tetanic responses produced by nitrendipine or by extracellular calcium ion removal without blocking twitches or action potentials in skeletal muscle. J. Pharmacol. Exper. Ther. 257, 575-581. Paul, R. J., Ferguson, D. G., and Heiny, J. A. (1996). Muscle physiology: molecular mechanisms. In "Essentials of Physiology", (N. Sperelakis and R. O. Banks, Eds.) pp. 203-216. Little, Brown and Co., Boston. Peachey, L. D. (1965). The sarcoplasmic reticulum and transverse tubules of the frog's sartorius. J. Cell Biol. 25, 209-232. Peachey, L. D., and Eisenberg, B. R. (1978). Helicoids in the T system and striations of frog skeletal muscle fibers seen by high voltage electron microscopy. Biophys. J. 22, 145-154. Pereon, Y., Navarro, J., Hamilton, M., Booth, E W., and Palade, P. (1997). Chronic stimulation differentially modulates expression of
SECTION VI Muscle and Other Contractile Systems
mRNA for dihydropyridine receptor isoforms in rat fast twitch skeletal muscle. Biochem. Biophys. Res. Commun. 235, 217-222. Perez-Reyes, E. and Schneider, T. (1995). Molecular biology of calcium channels. Kidney International 48, 1111-1124. Ray, A., Kyselovic, J., Leddy, J. J., Wigle, J. T., Jasmin, B. J., and Tuana, B. S. (1995). Regulation of dihydropyridine and ryanodine receptor gene expression in skeletal muscle. Role of nerve, protein kinase C, and cAMP pathways. J. Biol. Chem. 270, 25 837-25 844. Rios, E., and Pizarro, G. (1991). Voltage sensor of excitation-contraction coupling in skeletal muscle. [Review]. Physiol. Rev. 71, 849-908. Rios, E., Karhanek, M., Ma, J., and Gonzalez, A. (1993). An allosteric model of the molecular interactions of excitation-contraction coupling in skeletal muscle. J. Gen. Physiol. 102, 449-481. Saiki, Y., E1-Hayek, R., and Ikemoto, N. (1999). Involvement of the Glu724-Pro760 region of the dihydropyridine receptor II-III loop in skeletal muscle-type excitation-contraction coupling. (1999). J. Biol. Chem. 274, 7825-7832. Schneider, M. E (1994). Control of calcium release in functioning skeletal muscle fibers. [Review]. Ann. Rev. Physiol. 56, 463--484. Schneider, M. E & Chandler, W. K. (1973). Voltage-dependent charge movement in skeletal muscle: a possible step in excitation-contraction coupling. Nature 242, 244-246. Slavik, K. J., Wang, J. P., Aghdasi, B., Zhang, J. Z., Mandel, F., Malouf, N., and Hamilton, S. L. (1997). A carboxy-terminal peptide of the alpha 1-subunit of the dihydropyridine receptor inhibits Ca(2+)-release channels. Am. J. Physiol. 272, C1475-C1481. Stem, M. D., Pizarro, G., and Rios, E. (1997). Local control model of excitation-contraction coupling in skeletal muscle. J. Gen. Physiol. 110, 415-440. Suda, N. (1995). Involvement of dihydropyridine receptors in terminating Ca 2+ release in rat skeletal myotubes. J. Physiol. (Lond.) 486, 105-112. Suh-Kim, H., Wei, X., Klos, A., Pan, S., Ruth, P., Flockerzi, V., Hofmann, F., Perez-Reyes, E., and Birnbaumer, L. (1996). Functional interaction among alpha 1, beta, gamma and alpha 2 delta subunits. Receptors Channels 4, 217-225. Tanabe, T., Beam, K. G., Adams, B., Niidome, T., and Numa, S. (1990). Regions of the skeletal muscle dihydropyridine receptor critical for excitation-contraction coupling. Nature 346, 567-569. Tanabe, T., Beam, K. G., Powell, J., and Numa, S. (1988). Restoration of excitation-contraction coupling and slow calcium current in dysgenic muscle by dihydropyridine receptor complementary DNA. Nature 336, 134-139. Tanabe, T., Takeshima, H., Mikami, A., Flockerzi, V., Takashiki, H., Kangawa, K., Kojima, M., Matsuo, H., Hirose, T., and Numa, S. (1987). Primary structure of the receptor for calcium channel blockers from skeletal muscle. Nature 328, 313-318. Tsugorka, A., Rios, E., and Blatter, L. A. (1995). Imaging elementary events of calcium release in skeletal muscle cells. Science 269, 1723-1726. Vergara, J. and Delay, M. (1986). A transmission delay and the effect of temperature at the triadic junction of skeletal muscle. Proc. R. Soc. Lond. B Biol. Sci. 229, 97-110. Zhu, X., Gurrola, G., Jiang, M. T., Walker, J. W., and Valdivia, H. H. (1999). Conversion of an inactive cardiac dihydropyridine receptor II-III loop segment into forms that activate skeletal ryanodine receptors. FEBS Lett. 450, 221-226.
Gerhard Meissner
Ca 2+ Release from Sarcoplasmic Reticulum in Muscle
I. Introduction In excitable cells, the release of Ca 2§ ions from intracellular membrane compartments is triggered by an excitatory electrical signal, or it occurs via a chain of voltage-independent steps that involve the agonist-induced formation of inositol 1,4,5trisphosphate (IP3) and subsequent activation of an intracellular membrane receptor/Ca 2+ channel complex, the IP 3 receptor (Berridge, 1993). The voltage-dependent mechanism, commonly referred to as excitation-contraction (EC) coupling, is the focus of this chapter. It has been most thoroughly studied in vertebrate striated muscle. Figure 1 illustrates the major components involved in EC coupling: A transverse (T-) tubule membrane system of invaginations through which muscle contraction is triggered and an intracellular Ca 2+ storing and Ca 2§ releasing membrane system, the sarcoplasmic reticulum (SR), which contains a Ca 2+ pump and a Ca 2+ release channel. Lumenal ionized SR Ca 2+ has been estimated at about 1.0 mM and the ionized cytosolic Ca 2+ at about 0.1 IxM, yielding an approximately tenthousand-fold Ca 2+ gradient across the SR membrane. Sequestration of Ca 2§ into the SR against the Ca 2§ gradient is mediated by a Ca2+-ATPase that transports two Ca 2§ into the SR at the expense of one ATE The release channels span the narrow gap where the SR (the Ca 2§ store) and the T-tubule (the conduit of the action potential) are within --15 nm of each other (for reviews, see Fleischer and Inui, 1989; FranziniArmstrong and Protasi, 1997). The SR Ca 2+ release channels are also known as feet or junctional processes, and as ryanodine receptors (RyRs) because they have the ability to bind the plant alkaloid ryanodine with high affinity and specificity. They are often viewed, at least in skeletal muscle, as directly linked to four particles located in the T-tubule membrane and presumed to represent another Ca 2+ channel (L-type), also known as the dihydropyridine receptor (DHPR). The relative number of RyRs and DHPRs in striated muscle varies greatly, ranging from about 10 RyRs per DHPR in cardiac muscle to a 1:1 stoichiometry in fast-twitch mammalian skeletal muscle.
Cell Physiology Sourcebook: A Molecular Approach, Third Edition
92 7
surface membrane
T-tubule DHPR/Ca2+Channel RyR/ Ca"+ Channel
Ca2§
Myofibrils
calcium ~mmp
ADP +Pi
SR
D
(
RyR/
Ca"+ Channel
(
DHPR/Ca2+ Channel
T-tubule
1
FIGURE 1. Schematic representation of a segment of a mammalian skeletal muscle cell illustrating the major components involved in EC coupling. Copyright 9 2001by Academic Press.All rights of reproduction in any form reserved.
928
SECTION VI Muscle and Other Contractile Systems
Accordingly, not all RyRs may be linked or closely apposed to DHPRs in striated muscle. In addition to the Ca 2+ release channel, the SR of skeletal and cardiac muscle contains monovalent ion-selective channels: (1) a K + channel, (2) a C1- channel, and (3) a H + (OH) permeable pathway. Movement of ions through the monovalent cation- and anion-selective channels has been proposed to compensate charge movements across the SR membrane, and thereby to support rapid Ca 2+ release and reuptake (for review, see Meissner, 1983).
II. M e c h a n i s m s of EC C o u p l i n g Different excitation-contraction coupling mechanisms exist in skeletal and cardiac muscles (for reviews, see Fabiato, 1983; Rios and Pizarro, 1991). A distinguishing feature is that EC coupling in cardiac muscle is dependent on extracellular Ca 2+, whereas skeletal muscle EC coupling is not (in the short run). In cardiac muscle, dihydropyridine-sensitive, L(longlasting)-type Ca 2+ channels located in the surface membrane and T-tubule mediate the influx of Ca 2+ during an action potential by functioning as voltage-dependent Ca 2§ channels (Fig. 2). The resulting rise in intracellular Ca 2+ concentration triggers the massive release of Ca 2+ by opening SR Ca 2+ release channels. This process is known as calcium-induced Ca z+ release (CICR). Intracellular Ca 2+ transients have been suggested to arise as the sum of localized Ca 2§ release events called Ca z+ sparks. The opening of a single T-tubule Ca 2+ channel is apparently sufficient to evoke a Ca 2+ spark by activating a functional Ca 2+ release unit that may consist of one or more Ca 2+ release channels (Santana et al., 1996). Ca 2+ also plays a major role in regulating ryanodine-sensitive SR Ca 2+ release channels in mammalian smooth muscle and neurons.
Mammalian heart cells contain two additional Ca 2+ entry pathways, T(transient)-type Ca 2+ channels and the Na+-Ca 2+ exchanger; however, there is little evidence that they have a significant role in initiating SR Ca 2+ release in normal hearts. In vertebrate skeletal muscle, a different mechanism of EC coupling is present. A unique mechanism, referred to as the mechanical coupling mechanism, has been formulated and suggests that the SR Ca 2+ release channel is opened via a direct physical interaction with a voltage-sensing molecule in the T-tubule (see Fig. 2). Biophysical and pharmacological evidence, as well as molecular expression studies, suggest that the T-tubule DHPR is the voltage sensor for EC coupling in vertebrate skeletal muscle (Rios and Pizarro, 1991). The mammalian skeletal muscle DHPR is composed of five subunits: a l, a 2,/3, y, and 6 (Catterall, 1995). a 1 alone forms a Ca 2+ channel, binds Ca 2+ channel antagonists, and belongs to a voltage-sensitive ion channel superfamily that includes Na + and K + channels. Co-immunoprecipitation and cross-linking of skeletal muscle DHPRs and RyRs as well as morphological evidence suggest a well-defined interaction between the two receptors. Clusters of four particles, tetrads presumed to represent four DHPRs, are located opposite four subunits of every other RyR (FranziniArmstrong and Protasi, 1997). The formation of DHPR tetrads is dependent on the presence of the skeletal muscle RyR (Protasi et al., 1998). Other SR junctional proteins that may assist in the formation of the triad are triadin and junctin (for review, see Meissner and Lu, 1995). The existence of two populations of skeletal muscle RyRs, one coupled to tetrads and one not, raises the question of how unlinked RyRs are activated. One suggestion is that Ca 2+ released by DHPR-linked RyRs activates DHPR-unlinked RyRs in skeletal muscle by a CaZ+-induced mechanism resembling that in cardiac muscle (Rios and Pizarro, 1991). An alternative
Muscle Excitation-Contraction Coupling Skeletal Muscle: Mechanical Coupling
Cardiac Muscle: Calcium-Induced Calcium Release 9
DHPR: Voltage Sensor
Ca
2+ R: C ~ + Channel
T-tubular membrane Ca
2+
SR membrane +
RyR: Ca~
Release Channel
RyR: Ca~ + Release Channel
F I G U R E 2. Models of skeletal and cardiac muscle EC coupling. Depicted are the mechanical coupling model for vertebrate skeletal muscle, and the calcium-induced calcium release (CICR) model for cardiac muscle.
929
55. Ca2+Release from Sarcoplasmic Reticulum in Muscle mechanism is that neighboring skeletal muscle RyRs are physically linked, leading to simultaneous channel opening and closing, termed coupled gating (Marx et al., 1998). We shall discuss below studies carried out to determine the structural and functional features of the SR RyR/Ca 2+ release channel. These studies have shown that mammalian tissues, including muscles, express three RyR/Ca 2§ release channel isoforms that are encoded by three different genes: ryrl, ryr2, and ryr3. The three RyR isoforms are known as the skeletal muscle (RyR1), cardiac muscle (RyR2), and brain (RyR3) RyRs. All three share a high-affinity binding site for [3H]ryanodine and a high-conductance pathway for Ca 2§ and monovalent cations, but display isoform- and species-dependent differences in their in vitro regulation by Ca 2§ and other effector molecules.
0
OH
HO--
H
,,p
o
HO" HOi~
"" OH
111. Isolation of M e m b r a n e Fractions Enriched in RyR/Ca 2+ R e l e a s e Channels The molecular properties of the RyR/Ca 2+ release channels have been extensively studied with rabbit skeletal muscle as the source for RyRs. Fragmentation of the SR during homogenization and subsequent fractionation by differential and density-gradient centrifugation yields a "heavy" SR vesicle fraction enriched in Ca 2+ release channels and [3H]ryanodinebinding activity and that corresponds to the junctional region of the SR, junctional SR (see Fig. 1). Another important advance allowing study of the coupling between T-tubule depolarization and SR Ca 2+ release in skeletal muscle (Ikemoto et al., 1985) has been the isolation of membrane fractions composed of a T-tubule segment sandwiched between two junctional SR vesicles (Fleischer and Inui, 1989). These junctional complexes are known as triads. Microsomal membrane fractions enriched in ryanodinesensitive Ca 2+ release channels have been also isolated from other excitable tissues, including cardiac muscle, smooth muscle, and brain. In these cases, however, the membrane fractions are typically of a lower purity than those from skeletal muscle.
IV. Isolation and Structure of RyRs The isolation and structural determination of the SR Ca 2+ release channel has been greatly facilitated by the identification of ryanodine as a channel-specific ligand. Ryanodine is a neutral plant alkaloid that is obtained from the stems of the South American shrub, Ryania speciosa, and is composed of two major compounds: ryanodine and 9,21didehydroryanodine (Fig. 3). Ryanodine is a highly toxic compound. Its pharmacological effects have been most clearly shown in muscle, where depending on muscle type and activity, it can cause either contracture or a decline in contractile force (Jenden and Fairhurst, 1969). Ca 2+ flux studies with isolated SR vesicles and single-channel recordings have shown that ryanodine specifically affects the SR Ca 2+ release channel. Ryanodine activates the channel at low (nanomolar) concentrations, but inhibits the channel at high (micromolar) concentrations (see Section VI. C, Figs. 9 and 10). Because the drug binds with high specificity and dissociates slowly from the high-affinity site of the receptor (ei-
FIGURE 3. Structure of ryanodine (R1 = CH3, R2 = H) and 9,21didehydroryanodine (Rl, R2 = CH2).
ther membrane bound or detergent solubilized), [3H]ryanodine has been found to be an ideal probe in the isolation of the RyR from a variety of tissues and species. The RyR was first isolated from striated muscle because its role in regulating free cytoplasmic Ca 2+ levels was originally recognized in muscle, and relatively large amounts of membranes enriched in [3H]ryanodine-binding activity can be obtained from skeletal muscle and cardiac muscle. In most studies, the membrane-bound Ca 2§ release channels were solubilized in the presence of [3H]ryanodine, using the zwitterionic detergent Chaps and high ionic strength (1.0 M) NaC1; they were then purified by sequential column chromatography (Inui et al., 1987). Purification of a functional receptor can be also obtained in essentially one step by immunoaffinity chromatography (Smith et al., 1988) or density gradient centrifugation through a linear sucrose gradient (Lai et al., 1988). Figure 4A shows the [3H]ryanodine pattern on the sucrose gradients and sedimentation profile of the proteins associated with junctional SR vesicles isolated from rabbit skeletal muscle. A single peak of bound radioactivity, comigrating with a small protein peak possessing an apparent sedimentation coefficient of 30S, is observed in the lower half of the gradients. Binding to the small protein peak is specific since no radioactivity is present in the lower half of the gradients when the membranes are incubated with an excess of cold ryanodine. A sedimentation coefficient of 30S suggests that the RyR is a very large protein complex with a molecular weight (M) in excess of 1000 000. The sucrose gradient centrifugation procedure is relatively simple and straightforward. The procedure results in efficient separation of the large 30S RyR complex from the other solubilized smaller SR proteins because of its faster sedimentation rate. This method has been used to isolate a functional 30S RyR from several species and tissues including skeletal muscle, cardiac muscle, smooth muscle, and brain (Meissner, 1994).
930
SECTION VI Muscle and Other Contractile Systems
A i
/.
30
0.9
0.6
v
v
0~
0.3 O e-*-
o a:l
r~
O
0
0.0
5
Fraction
I
lO
Number
SDS polyacrylamide gel electrophoresis has shown that the 30S protein complexes of mammalian skeletal and cardiac muscles are composed of a major high-molecular-weight RyR polypeptide and isoform-specific low-molecular-weight immunophilins (FK506 binding protein) which migrate with an apparent M r > 340 000 (Fig. 4B) and M r ~12,000 (Timerman et al., 1996; not visible on the gels in Fig. 4B), respectively. Cloning and sequencing of the complementary DNA of the mammalian skeletal and cardiac muscle RyR isoforms have revealed an open reading frame of about 15 kb encoding RyR polypeptides of M r ~560,000 (rev. in Sutko and Airey, 1996; Ogawa et al., 1999). Purified RyR preparations from mammalian brain, crustacean skeletal muscle, and the nematode C a e n o r h a b d i t i s elegans also displayed a single protein band of high molecular weight on SDS gels. In contrast, the presence of two immunologically distinct high-molecular-weight RyR protein bands (corresponding to the mammalian RyR1 and RyR3) has been described for the main skeletal muscles of chicken, frog, and fish (Sutko and Airey, 1996). The two RyR isoforms are present as discrete homooligomers in amphibian and avian skeletal muscles. The appearance of the RyR1 isoform alone in some very fast-contracting muscle of fish suggests that this isoform is selectively expressed when rapid contraction is required in nonmammalian vertebrate muscles (O'Brien et al., 1993). Electron microscopy has revealed that the purified mammalian RyRs have a morphology nearly identical to that of the protein bridges (junctional feet) that span the Ttubule-SR junctional gap in vertebrate skeletal muscles (Franzini-Armstrong and Protasi, 1997). The skeletal muscle RyR is composed of four polypeptides of M r -560 000 as evidenced by (a) the four-leaf clover-like (quatrefoil) appearance of negatively stained samples (Fig. 5), (b) a high apparent sedimentation coefficient of 30S (see Fig. 4A), and (c) cross-linking studies (Lai et al., 1989). Threedimensional reconstruction of images from electron microscopy indicates that the skeletal and cardiac muscle RyRs consist of a large loosely packed cytosolic foot region and a smaller transmembrane region that extends ~7 nm toward the SR lumen and is thought to contain the Ca 2§ channel pore (for review, see Wagenknecht and Radermacher, 1997).
V. Molecular Cloning and Expression of RyRs FIGURE 4. Sedimentation profile and SDS gel electrophoresis of rabbit skeletal muscle and canine cardiac RyRs. (A) Junctional skeletal muscle SR vesicles were solubilized in Chaps, centrifuged through a linear sucrose gradient, and fractionated. Fractions were analyzedfor protein and 3H-radioactivity.Unbound [3H]ryanodine and the majority of the solubilized proteins sedimented near the top of the gradient (fractions 1-6), whereas the [3H]ryanodine-labeledRyR comigrated with a small protein peak to the bottom of the gradient (fractions 11-13). (In modified form from Lai et al., Nature 331, 315-319, 1988. Copyright 1988 by Macmillan Magazines Limited.) (B) Silver-stained SDS-polyacrylamide gel of whole rabbit skeletal muscle homogenate (H) and rabbit skeletal and canine cardiac muscle heavy SR membranes (M) and purified RyRs (R). Sizes of molecular weight standards are shown (x 10-3). (From Meissner et al., Mol. Cell. Biochem. 82, 59-65, 1989.)
Mammalian tissues express three types of RyRs that are encoded by three different genes. The three RyR isoforms are also known as the skeletal muscle (RyR1), cardiac muscle (RyR2), and brain (RyR3) RyR because they were first identified and isolated from skeletal muscle, cardiac muscle, and brain, respectively (Sutko and Airey, 1996; FranziniArmstrong and Protasi, 1997; Ogawa et al., 1999). Each of the large RyR polypeptides in RyR1, RyR2, and RyR3 is comprised of ~5000 amino acids with a predicted molecular mass of ~560 kDa and amino acid sequence identity of 65-70%. RyR1 and RyR2 are the predominant isoforms in skeletal muscle and cardiac muscle, respectively; however, both isoforms are also expressed in brain and other tissues at low levels. In turn, the brain RyR is expressed as a minor
55. Ca2+ Release from Sarcoplasmic Reticulum in Muscle
931
F I G U R E 5. Negative-stain electron micrograph of the purified rabbit skeletal muscle RyR. Shown is a selected panel of particles displaying the characteristic four-leaf clover (quatrefoil) structure of the 30S RyR complex. Dimensions of the quatrefoils are 34 nm from the tip of one leaf to the tip of the opposite one, with each leaf 14 nm wide. The central electron-dense region has a diameter of 14 nm with the central hole of a diameter of 1-2 nm (from Lai et al., Nature 331, 315-319, 1988. Copyright 1988 by Macmillan Magazines Limited).
component in skeletal and cardiac muscles, cDNAs encoding the amphibian, avian, and fish homologs of mammalian RyR1 and RyR3 have been isolated and sequenced. Isolation and sequencing of a gene for RyR in the fruit fly (Drosophila melanogaster) and Caenorhabditis elegans revealed 40-47% identity between the amino acid sequences of the Drosophila and C. elegans RyRs and the three mammalian RyRs (Sakube et al., 1997). Hydropathy plots of the amino acid sequences of RyRs have suggested that the M r ~560 000 polypeptides have two major structural regions: (1) a C-terminal pore region that exhibits a high extent of similarity in amino acid sequence and traverses the membrane at least four times (hence 16 or more transmembrane segments per tetrameric RyR); and (2) a large, more variable extramembrane region, which is thought to correspond to the cytoplasmic foot structure. The four membranespanning segment model (Takeshima et al., 1989) is supported by studies with SR lumenal site-directed antibodies and singlechannel recordings with tryptic fragments and deletion mutants. Studies with site-directed antibodies suggest that the Nand C-termini of RyR1 are cytoplasmically localized, with the C-terminus being important for the expression of a functional RyR1 complex (Gao et al., 1997). Expression studies in cul-
tured cells and Xenopus oocytes have suggested that the M r 560 000 polypeptides are sufficient to form Ca 2+ channels sensitive to ryanodine, caffeine, and Ca 2+. Primary sequence analysis suggests the presence of several phosphorylation sites and sites of Ca 2+, ATE and calmodulin binding on the large cytoplasmic foot structure. Experimental evidence for one phosphorylation site (Witcher et al., 1991), three calmodulin binding sites (Guerrini et al., 1995), and several Ca2+-sensitive channel domains has been reported (see Section VII, Fig. 12). RyRs constitute a potential target for reactive nitrogen and oxygen species because they contain a large number of sulfhydryls (~100/RyR subunit) whose oxidation modulates function (for review, see Zucchi and Ronca-Testoni, 1997). In favor of a regulatory role of sulfhydryls, the cardiac channel is S-nitrosylated endogenously, and S-nitrosylation and oxidation by exogenous NOgenerating molecules lead to channel activation and inactivation (rev. in Eu et al., 1999). RyR gene knockout mice were constructed to address the question of a functional requirement of the predominant (RyR1) and minor (RyR3) RyR isoforms for skeletal muscle EC coupling. Mutant mice lacking RyR1 died perinatally and the skeletal muscle fibers failed to show a contractile response
SECTION Vl Muscle and Other Contractile Systems
932
to electrical stimulation under physiological conditions. RyR3 knockout mice showed an impairment of contraction in neonatal but not adult skeletal muscle (Takeshima et al., 1996; Sonnleitner et al., 1998). Deletion of RyR2 resulted in embryonic lethality and altered cardiomyocytes (Takeshima et al., 1998).
VI. RyRs Are High-Conductance Ligand-Gated Channels Function of the Ca 2+ release channels has been extensively studied in vitro in SR vesicle Ca 2+ flux and singlechannel measurements. A third, more indirect method relies
on the use of [3H]ryanodine, a plant alkaloid that is widely used as an indicator of channel activity due to its preferential binding to open channels.
A. SR Vesicle Ca 2+ Efflux Measurements In skeletal and cardiac muscles, the SR releases its Ca 2+ stores in milliseconds in response to an action potential. To measure similarly rapid Ca 2§ fluxes in vitro, rapid mixing and filtration devices must be employed. The released Ca 2§ can be measured using Ca 2+ indicator dyes such as fura-2 or a radioisotope of Ca 2+ (45Ca2+). Figure 6 shows a typical 45Ca2+ efflux experiment. A rapid mixing device (Fig. 6A) is used to measure 45Ca2+ re-
I -cm/s r m
I0O
Samplein 5_mM45Ca2+ - - -o--.....
IOmM Mg, IOjuMRR
--._ .c_ o
S_ 50
9-, 5./uM
I I
I I
I i
Activation Ca z +
I I
I00
I I Quench
"5
-
5O
~
5
)
u
M
Ca 2+
~' 25
g
'-
8 7
o o~ 0
.
.
. 9
. ,
.
FIGURE 5.
I 20
-9
,
, I 40
0
.,
I-
0
,
Systems
I 60
......
I 80
m ~00
~t $~
I 500
TIME (ms) Excitation-contraction coupling in Spirostomum. Temporal relationship between excitation, increase in free cal-
cium, and the contraction of a Spirostomum cell preinjected with aequorin, a Ca 2+-dependent luminescent protein. Contraction begins when aequorin-induced photon emission is maximal. Cessation of electrical stimulation is followed by a decrease in photon emission and the cell relaxes. (Reproduced from The Journal of General Physiology, 1970, vol. 56, pp. 168-179, by copyright permission of The Rockefeller University Press.)
58. Centrin-Based Contraction and Bacterial Flagella
989
0.8
0.6 3
o ,rx
contracted o
c 0.4 cD ci.. CD L_
e,.
m 0.2 1
9
0
I
,
I
0.2
I
J
0.4
I
I
I
0.6
......
!
0.8
I
I
1.0
Stress ts (kg.cm2)
FIGURE 6. Birefringence changes of spasmonemes. In the presence of high calcium (o) the spasmoneme birefringence is directly proportional to stress (ts -- tensile stress). Thus in the high calcium state, the spasmoneme behaves optically and mechanically as rubber. However, in the low calcium state at 10-8 M concentration (X), the relationship between birefringe and stress is not the same. In the low calcium medium, the spasmoneme actively elongates and exhibits a pushing force, and the unstressed organelle becomes birefringent. [Reprinted with permission from Amos, W. B. (1975). Contraction and calcium binding in the vorticellid ciliates, In "Molecules and Cell Movement" (S. Inou6 and R. E. Stephens, Eds.), pp. 411-436. Raven Press, New York.]
energetic known muscle (insect flight muscle), which is only 0.05-0.2 kW/kg wet wt. Spasmonemes have a high degree of passive extensibility and can be stretched to four times their resting lengths and then return to their normal lengths. At low Ca 2+ concentrations, a measurable pushing force and increased birefringence develop. If the stalk is prevented from elongating, the spasmoneme is seen elongating and bending within the stalk. This pushing force observed during spasmoneme extension is not thoroughly understood, but it may involve ATP hydrolysis, which is consistent with the presence of abundant mitochondria closely associated with spasmonemes. Elemental analysis (calcium K peak) by electron microscopic microprobe of glycerinated Z o o t h a m n i u m giant spasmonemes indicates that the organelle has 1.5 and 0.36 g bound calcium/kg dry wt during contraction and extension, respectively (Fig. 7). The change in chemical potential of calcium (Atx,~ = R T In [Ca2+],.. / [CaZ+]]ow) was examined to deterLa nlgn mine whether Ca 2+ alone can account for work done by spasmonemes during contraction. The value of 104 J/mol calcium was calculated for a change from 10-8 to 10-6 M; therefore, for each mole of calcium bound to the spasmoneme, 104 J of energy would be available for work. As mentioned previously, the work done in a single twitch of the spasmoneme is 11 J/kg wet wt. Therefore, to produce the estimated energy output, the spasmoneme must bind at least 11 • 10 --4 moles of Ca 2+, or 44 mg calcium&g wet wt. The microprobe elemental analyses indicated values of 1.5 (contracted spasmoneme) and 0.36 (extended spasmoneme) g calcium bound/kg wet wt. Therefore, the chemical potential of calcium can account for the work done by spasmoneme contraction.
o o
0.5
1 .o
1.5
2.0
Total Ca (mM)
FIGURE 7. Electron microprobe analysis of the calcium content of isolated glycerinated Zoothamnium giant spasmonemes. Contracted spasmonemes were prepared by treatment with 10- 6 M Ca 2+. Extended spasmonemes were treated with 10-8 MCa2+, except at the ordinate, at which a 20 mM EDTA solution was used [Reprinted with permission from Amos, W. B. (1975). Contraction and calcium binding in the vorticellid ciliates. In "Molecules and Cell Movement" (S. Inou6 and R. E. Stephens, Eds.), pp. 411-436. Raven Press, New York.]
3. Spasmins and the Mechanism of Contraction
The major proteins of spasmonemes, spasmins, are acidic proteins with pI values of 4.7 and 4.8. They constitute 50-60% of the total stainable proteins of the isolated organelles separated by SDS-PAGE. Their migrations, which are not affected by/3-mercaptoethanol (no - S - S - bonds), indicate that their masses are about 18 kDa (spasmin A) and 20 kDa (spasmin B). In EGTA-CaZ+-buffered gels, their mobilities (anodally) decrease in the presence of high Ca 2+ (10 -6 M), suggesting a decrease in net negative charge, an increase in the Stokes radius of the molecules, or both. However, in alkaline urea gels with Ca 2+, spasmin B exhibits increased mobility, which is not consistent with a decrease in net negative charge due to the binding of calcium. This observation provided strong evidence against an earlier hypothesis, the electrostatic model of spasmoneme contraction. This hypothesis was based on the idea that Ca 2+ binding neutralized repulsive negative charges, resulting in the folding of polymeric molecules. Using the microprobe data for bound calcium in contracted and extended spasmonemes, it was calculated that Ca 2+ binds to spasmin with stoichiometry of one or two Ca 2+ per molecule. Spasmins are high in serine, aspartate + asparagine, and glutamate + glutamine. Unlike troponin C and parvalbumin, spasmins lack cysteine. Coiling of microfibrils is a plausible hypothesis for myoneme contraction. An alternative hypothesis is the entropic rubber model, which generally equates spasmonemes with material such as elastin or rubber that possesses oriented isoprene units. In this model, contraction results by the release of stored elastic energy. 4. Excitation-Contraction C o u p l i n g Less examined and understood is the nature of the coupling of excitation with contraction (EC coupling). Since the
990
SECTION VI Muscle and Other Contractile Systems
activation time is on the order of 2-3 ms in Zoothamnium, and its giant spasmoneme is 30 Ixm in diameter, the signal is transduced 10 times faster than can be explained by simple diffusion of Ca 2+ through the surface. Currently, the linkage complex associated with spasmonemes appears the most likely candidate for EC coupling (see Fig. 4). Allen has suggested that the striated filament that links the midpiece of the linkage complex with cortical structures near the surface membrane (e.g., basal bodies) may transduce the electrical signal to the midpiece. The midpiece is physically coupled to the ER and hence could trigger Ca 2+ release from the cisternae. Furthermore, the rails may be involved in the movement of Ca 2+ in and out of the ER and may be the sites of ATPase activity for concentrating Ca 2+ into the ER. B. Striated Rootlets Striated rootlets (rhizoplasts), well-developed in unicellular organism such as flagellated algal cells (e.g., Tetraselmis), are organelles that are associated with the flagellar basal bodies (Fig. 8). They extend, branch, and link to adjacent basal bodies. In some cases they extend and pass along the nuclear envelope and anchor near the cell membrane opposite the flagella. The anchor site has been termed laminated oval, half-desmosome, or rhizankyra. The rootlet exhibits crossstriations of greater than 160-nm periodicity and is composed of 3- to 8-nm microfibrils aligned parallel to the long axis of the organelle. That these were contractile organelles was clearly demonstrated by fixation of Tetraselmis cells in 0.1-5 mM Ca 2+. Salisbury (1989) found that, compared with organelles in cells fixed without Ca 2+, the organelle was 65% shorter, 30% wider, and had a reduced cross-striation periodicity (see Fig. 8). Other studies of striated rootlets demonstrate that regions proximal to the flagellar basal bodies have smaller periodicities than distal regions, suggesting a polarity in the organelle. However, this observation may reflect the direction in which contraction propagates. In living cells, it was observed that the rootlets undergo cyclic contractions and extensions with millimolar concentrations of Ca 2+ and ATP. Much like spasmonemes, neither muscle HMM nor myosin S-1 fragments decorate extracted rootlets. Also, striated rootlets and other related organelles contract rapidly within less that 20 ms; reextension is slow (a few seconds to about 1 hour.) Contraction is initiated by elevated intracellular Ca 2+ and is independent of ATP hydrolysis. When contracted, the microfibrils in striated rootlets assume a supercoiled configuration. Although the direct requirement for ATP in the reextension of spasmonemes is not clearly established, there is good evidence that ATP is required by striated rootlets during their reextension process. The major proteins of striated rootlets, centrins (called caltractin in Chlamydomonas), make up greater than 60% of the total proteins of the isolated organelle. Centrins are lowmolecular-mass (20-kDa) acidic proteins (a centrin, pI 4.9; and/3 centrin, pI 4.8). Metabolic radiolabeling with 32po 4 demonstrates that the /3 isoform is phosphorylated whereas the a isoform is not. The role of phosphorylation of these contractile proteins is currently unclear. However, it has been suggested that the phosphorylated /3 and dephosphorylated c~ isoforms are
FIGURE 8. Striated rootlet of the marine flagellate, Tetraselmis. Cells were fixed for electron microscopy under various conditions. The insets show the binding patterns of flourescence-labeled anti-centrin antibodies. (A) Fixation with low Ca2+ illustrates the normal width of the organelle and the periodicity of its cross-striations. (B) Fixation with high Ca2+ plus ATP may be equivalent to the reextension phase after contraction. (C) Fixation with high Ca2+ (0.1-5mM) caused a dramatic shortening of the striated rootlets. The organelle in this state is characterized by increased thickness and reduced periodicity of the cross-striations. (Courtesy of Jeffrey Salisbury.)
correlated with the extended and contracted states, respectively. Both isoforms of centrin exhibit Ca2+-sensitive mobilities in electrophoresis gels. However, in alkaline urea gels, the mobilities decrease in the presence of Ca 2§ similarly to the mobility of spasmins in SDS-PAGE under denaturing conditions. These observations on centrin may be
58. Centrin-Based Contraction and Bacterial Flagella
991
FIGURE 9. Anti-centrin antibody binding to cells. The wide occurrence of centrin-like molecules is demonstrated by a HeLa cell double-stained with fluorescence-labeled anti-tubulin and anti-centrin antibodies. (A) Anti-tubulin binding pattern. (B) Anticentrin binds to the centrosomes of cells from broad phylogenetic backgrounds. (Courtesy of Jeffrey Salisbury.)
related to those made with troponin C, which binds Ca 2+ in urea gels, causing an increase in mobility. This has been interpreted as a change to a more compact molecular conformation, analogous to the interaction of troponin I and troponin C when the two proteins interact and form a slower migrating complex.
C. Occurrence and Functions of Centrin and Related Molecules Monospecific antibodies to flagellar rootlet centrin proteins have been reported to bind to centrioles and centrosomes of mitotic spindle poles, as well as to the mitotic spindle matrix (Fig. 9). These observations, made on a wide
range of cell types, revealed the ubiquitous occurrence of centrin-like molecules and inspired Salisbury (1989) to coin the term centrin for the protein. These findings indicate that centrin may have a universal function in cells, since it is found in cell division structures. The anti-centrin antibodies also bind to a nine-pointed stellate structure in the transition zone between the flagellar shaft and the basal body of some flagella. The antibody binding occurs at the plane of scission during the deflagellation process. The stellate structure was shown to contract during CaZ+-induced deflagellation, implicating that structure in the generation of force required for autotomy (Fig. 10). However, autotomy occurs in many cilia and flagella that normally do not have the nine-pointed stellate structure,
FIGURE 10. Transition zone of Chlamydomonas flagella at thenine-pointed stellate structure. (A) The configuration in a normal flagellated cell. (B) The nine-pointed stellate structure appears contracted in a cell treated with high concentrations of Ca2+, a treatment that induces autotomy. Constriction of the stellate structure may be the direct cause of weakening at this plane leading to severing of the flagella. (Reproduced from The Journal of Cell Biology, 1989, vol. 108, pp. 1751-1760, by copyright permission of The Rockefeller University Press.)
992
SECTION VI Muscle and Other Contractile Systems
and recently Jarvik has described autotomy in a Chlamydomonas mutant lacking this structure. The centrin-related proteins are related to the E-F hand superfamily of calcium-binding proteins. Huang isolated and sequenced Chlamydomonas caltractin (centrin) cDNA and demonstrated that the protein contains four E-F hands. Sequence identity with calmodulin is 45-48%, and sequence identity with the yeast spindle pole CDS31 gene is 50%. Anti-caltractin antibodies, however, do not cross-react with calmodulin. An alternative hypothesis for the function of centrin has been proposed for the biflagellated alga Spermatozopsis, whose rhizoplast (rootlet) does not exhibit cross-striations. A connecting fiber joins the two basal bodies. During forward swimming, the two flagella of the cell exhibit asymmetrical breast stroke movements. The organism displays a photophobic response to light (avoidance reaction), resulting in backward swimming. During this response, the flagella undulate and the basal bodies reorient by contraction of the connecting fiber. This switch for basal body reorientation is independent of the motor that drives flagellar beating. Anti-centrin antibodies also bind to a structure called the paraxial rod, which is found within the transverse flagellum of some dinoflagellates. The paraxial rod may cause a rapid and strong contraction of the dinoflagellates transverse flagellum that causes a stop reaction of the swimming cell (equivalent to an avoidance reaction). Thus, centrin may function as a modulator of motile structures.
11. Prokaryote Locomotion Locomotion, characteristic of some bacteria, serves as a useful taxonomic criterion. Motility assays are standardly employed in identifying bacterial isolates. Two types of locomotion are recognized in prokaryotic cells: (1) flagella driven swimming and (2) gliding. Gliding prokaryotes and eukaryotes is discussed later in Section III.
bacterial flagellum is composed of three parts: the basal body, hook, and filament (Fig. 12).
1. Basal Body The basal body is relatively small, but is a complex structure made up of a rod and rings. The rod is analogous to a transmission shaft. Basal body tings differ in gram-negative and gram-positive bacteria, and these substructures correlate with the differences in cell wall structures. In gram-negative species such as Escherichia coli and Salmonella typhimureum, there are usually four rings: (1) the M ring resides within the cell membrane; (2) the S ring is located within the periplasmic space, (3) the P ring is in the peptidoglycan layer; and (4) the L ring is within the lipopolysaccharide-containing outer layer. The outer two tings probably serve as the motor's bushing and allow the rod to pass through the outer layers of the cell envelope. The inner two are involved in rotation of the motor. How rotation is generated is not yet clear. It has been suggested that the M ring may serve as the stator and the P ring as the rotor. Alternatively, some have argued that a stator needs a large mass offering sufficient stability into which it can anchor, since the filament needs to overcome high viscous drag. The most likely structure for greater mass is the peptidoglycan layer. In this case, the rotation would then be generated with the cell membrane and the M ring directly involved in the generation of rotation. In gram-positive species that lack the outer layer, only two tings are found in the basal body: one is the equivalent of the M ring within the cell membrane, and the other, the outer ring, is associated with the teichoic acid component of the cell wall.
2. Hook The hook functions like an elastic universal joint between the basal body and filament. Its 90-nm length is genetically predetermined and involves hook-associated proteins (HAP) at the hook-filament junction. Mutants lacking HAP continuously secrete unassembled filament protein into the medium.
3. Filament
A. Bacterial Flagella Flagellated bacteria swim by the rotation of rigid helical flagella driven by motors at their bases. Bacterial flagella are thin (ca. 20 nm in diameter) extracellular organelles that extend about 15-20 Ixm from the surface cells (Fig. 11). They can occur as a single flagellum (monopolar) or a bundle of several flagella (lophotrichous) at one end of the cell. When bundles occur at both poles, the pattern is called amphitrichous. Cells with randomly distributed flagella are peritrichous. In spirochetes, flagella do not protrude from the cell surface. Instead, an axial filament consisting of two sets of fibrils that extend from pole to pole, enclosed within the outer layer of the cell surface, is responsible for the flexing movements exhibited by these organisms.
The filament, which acts like a propeller, is a rigid, brittle helical polymer of a single 51-kDa protein, flagellin. Some bacterial flagella (such as in Caulobacter) are composed of two types of flagellin molecules. The flagellin subunits determine the final shape of the filament structure by non-equivalent interactions between the subunits. Left-handed helices are more common than right-handed helices. Flagellin self-assembles in vitro, forming a cylindrical wall of about 11 subunits per 2 turns. A mixture of two different flagellins also self- assembles in vitro. Unlike that of actin and tubulin, polymerization of flagellin does not require energy from nucleotide triphosphates. Flagellins are among the most antigenic proteins known, which is an adaptation of the vertebrate immune system for effective elimination of bacterial infections.
B. Structure
C. Assembly of Flagellar Substructures
When stained, bacterial flagella are visible by light microscopy, but details of their structure can be elucidated only by special microscopic and spectroscopic techniques. The
Much like that observed in growing eukaryote flagellar axonemes, the assembly of the growing flagellum occurs by the addition of individual protein subunits from the proximal to
58. Centrin-Based Contraction and Bacterial Flagella
993
F I G U R E 1 1. Electron micrograph of Pseudomonas marginalis. Flagella emerge from one pole of the bacterial cell. (Courtesy of Arthur Kelman.)
distal end; the tip is the last to be added. This base-to-tip mechanism probably evolved in bacteria, since assembly occurs primarily outside the cell membrane and the export and addition of simple subunits to a single structure are easier than the export and addition of larger, more complex modular structures. Also, since most of the torsional load is at the proximal end, and because the motor rotates while the organelle is being elongated, any proximal separation in the nascent structure for insertion of new material would be mechanically difficult. The precursors, therefore, must be exported, single-file, through the growing structure; they do not reach the tip on the external surface. The information for the entire flagellum structure is probably inherent in the structure of each subunit, since self-assembly in vitro reconstitutes the different componant~ v a r y cln~a in ~i7.~ nnd ~t~.ctll~ne~ tN tht~. intact nrcranalla
and the export of the subunits. A central channel of 5-nm diameter has been detected in bacterial flagella by x-ray fiber diffraction techniques. This channel is sufficiently large to serve as a conduit for single subunits to get to the growing tip.
D. Energy Source The energy for rotation of the motor does not come from the hydrolysis of high-energy phosphate bonds of ATP. It is derived from the dissipation of an ionic transmembrane electrochemical gradient (ion-motive force). In most neutrophilic species, such as E. coli, the ionic species that has been actively pumped out of the cell by the electron transport chain in the bacterial membrane during respiration is H § (proton-motive fnrce,~ The, Colmlin~ of energy released in the resoiratorv chain
994
SECTION Vl Muscle and Other Contractile Systems
f-~
Filament cap assembly of flagellin)
t
Filament (helical propeller)
..Hook-filament junction (HAP & adaptor proteins) ook (universaljoint) /L ring " ~ (bushings) ~/ / P ring)
Rod (transmission shaft)
..
/[f//'/////S/////SS///~J { I I I !"V I'//////'////~/.//'/S//]~ Outersu fface layer ~ { ! ! ~'~.gl--Peptidoglycanlayer periplasm
=-11 !1 ~ ~ [ i)_ ~i -'_~ ~33_~ - 1 IC [ "q - ~~
~
& S rang . -q-Cellmembrane
Motor complex (motor r o t a t i o n ) ~ Switch complex (motor rotation switching)
M ring cytoplasm (mounting plate)
Export apparatus F I G U R E 12. Structure of the flagellum of a gram-negative bacterium. The filament is flagellin polymer that is rigid and helical. The hook, L ring, P ring, and rod lie distal to the cell surface membrane. The S ring and M ring are embedded in the membrane bilayer and are associated with the complexes that serve as the motor and the motor switch. The export apparatus allows molecules into a central channel for flagellar growth at the site of assembly at the tip. (Modified from MacNab, 1990, with permission of Cambridge University Press.)
electrical gradient, the proton-motive force becomes available to drive flagellar rotation. Under anaerobic conditions, the translocation of protons is achieved by proton ATPase activities. In marine and alkalophilic bacteria, the ion is Na +, which has been pumped out of the cell by a Na+-H + antiporter or by the electron transport chain. Since the energy can come from the electrochemical gradients of either H + or Na +, the rotation of the motor must involve electrostatic interactions and events involving association and dissociation of ions. Although no morphological structures have been identified, genetic and biochemical evidence indicates that there must be structures involved in energy transduction for converting ionmotive force to motor rotation. Since it is difficult to control specific membrane potentials and pH gradients experimentally in gram-negative bacteria, energy coupling has been best examined in gram-positive bacteria such as Streptococcus and Bacillus subtilis. It has been clearly established that rotation is coupled to the inward current of protons and that the speed of rotation is proportional to the proton-motive force. A flux of 1200 protons per revolution, independent of load or veloc-
ity, has been calculated. Membrane potentials or pH gradients of opposite polarity also drive the rotation of the motor. Gliding motility in some species of bacteria has also been shown to be powered by proton-motive force.
E. Swimming Behavior In general, peritrichous bacteria swim slower and rotate in a relatively straight line; those with flagella restricted to poles swim in many directions and spin rapidly. During swimming, the flagella of peritrichous bacteria rotate as a coordinated bundle (Fig. 13). In amphitrichous cells, the bundle at one end rotates clockwise (CW) and the one at the other end rotates counterclockwise (CCW) to propel the cell in one direction. The rotary nature of the flagellar motor can be visualized in tethered cells with their flagella adhered onto the substratum. Tethered cells rotate; the shorter the cell, the less the hydrodynamic drag, and thus the faster the cell rotates. The force created by the dissipation of the transmembrane potential propels the cell at a speed of 20-80 p~m/s or
995
58. Centrin-Based Contraction and Bacterial Flagella
re7 Cl- q r
als, random walk becomes biased (see Fig. 13). This is unlike chemotaxis of Dictyostelium amoebae, which move directly toward or away from attractants and repellents, respectively. Taxis is possible because the flagellar motor has an all-ornone switch. The morphological identity of the motor switch has yet to be determined, but there is good genetic and physiological evidence for its existence. When E. coli is placed in a spatial chemical gradient of an attractant, suppression of tumbling leads to net progress toward the attractant; repellents increase tumbling. Actually, bacteria do not sense a spatial gradient; that is they cannot detect differences in concentrations between their anterior and posterior parts. Berg found that chemotactic behavior is the result of a response to a temporal gradient that develops as the cell moves through the medium. Thus, bacteria must be able to measure the attractant concentration, store that information, and compare it with a value measured at a later time. To do this, they must have memory. The time between stimulus and response (response latency) is on the order of 200 ms. In signal transduction, the cell can integrate information from one type of receptor or it can integrate information from different receptors; the effect of an attractant (CCW flagellar rotation) is canceled by the effect of a repellent (CW flagellar rotation). Chemotactic responses are transient; thus the cell exhibits adaptation.
~eN~
2. Receptor Binding and Transduction of Signal to Switch
~f-twiddle
No Gradient
"~,run
d 6 d_qZ,~
q
Gradient
of repellent
I ,4q~;
FIGURE 13. Swimtracks illustrating chemotaxis in bacteria. Top: The cell exhibits spontaneous runs and tumbles (twiddles) with random orientations when attractant or repellent molecules are uniformly distributed in the medium. During runs, flagella form a bundle and rotate together. Tumbling results from the flagella splaying out of the bundle. Middle: In the presence of a gradient of an attractant, a random walk changes to a biased walk characterized by longer runs toward the region of higher attractant concentration. Bottom: The cell exhibits longer runs away from higher repellent concentration.
>10 lengths/s. In cells like E. coli, swimming is a threedimensional random walk with gentle curves called runs. The flagellar bundle revolves CCW at a rate of 12 000 rpm. The current consumption of each flagellum has been calculated at 10-5 A. Considering that bacteria commonly live in dilute aqueous solutions, with Reynolds numbers of 104-105, inertia is insignificant. That is, if the cell stopped swimming, it would come to a halt estimated at
-4.0
-4.5 Reaction coordinate
FIGURE 6. (A) Energy diagram for the simplistic enzyme reaction scheme described in the text, Eq. 19. (B) shows the AV component of AG in the same scheme, with, for purposes of illustration, a-AV shown in the first, binding step followed by +AV# for the catalytic stage. [(A) from Northrop, 1996; (B) from Morild (1981)]
The first part of the reaction scheme in Eq. 19 is an enzyme-substrate binding equilibrium and the second part is a catalytic reaction. Each passes through its activation (transition) stage; thus the AV # relating to k2 is the activation volume in forming the transition stage ES*, en route to releasing products from the enzyme-substrate complex. In high substrate concentration, the reaction velocity at high pressure is limited by the AV # of k2; hence the slope of plot of In K against pressure will reveal A V #. An example of the effect of pressure on the ES ~ P + E step is given in Fig. 7 A. The enzyme alkaline phosphatase catalyses the release of phosphate from P-nitrophenyl phosphate. High substrate concentrations ensure the reaction velocity is maximal, Vmax, and the data enable AV# to be calculated a t - 2 1 . 6 m l . m o l -l, that is, pressure accelerates the catalysis. If the reaction is run in the presence of the neutral salt K I , - A V # is much reduced by, it is argued, affecting the hydration changes which attend the activation process in the ES ~ P + E stage. (Low and Somero, 1975). An example of pressure affecting the E + S step, that is, the binding of an enzyme and its substrate, is shown in Fig. 7B. As the enzyme phophofructokinase binds to its substrate, fructose 6-phosphate, it undergoes a decrease in intrinsic fluorescence. Thus the substrate binding equilibrium can be measured optically at high pressure. In Fig. 7B the log of K, the dissociation constant, the inverse of the binding constant, plotted against pressure, shows a linear positive slope, of AV 45 ml-mo1-1 for the binding reaction. The enzyme's regulatory ligands, phosphoenol pyruvate (PEP), which inhibits, and MgADR which activates, also bind to
~o _.J
-5.0
-5.5
0
I
I
I
30
60
90
I
,,i
120 150
Pressure (MPa)
FIGURE 7. Effects of high pressure on enzyme-catalyzed reactions. (A) Alkaline phosphatase catalyzes the release of phosphate from P-nitrophenyl phosphate. Vmax obtained at high substrate concentrations is plotted against pressure, and the slope reflects the -AV# of 21.6 ml.mo1-1 (salt absent) and-8.6 ml mo1-1 (salt present) for the ES* activation stage. See text. Reprinted with permission from Greaney and Somero (1979), copyright (1979)American Chemical Society. (B) The binding of the enzyme phosphofructokinase to its substrate fructose6-phosphate may be monitored by fluorescence. The dissociation constant K=([E][S]/[ES]) for the binding process is increased by pressure, AV 45 ml.mo1-1. (From Johnson and Reinhart, 1996).
the enzyme with +AV, (+40 and +44 ml mo1-1 respectively). In all these cases the +AV of binding (not a V #) is likely to arise from the release of electrostricted water from the binding surfaces. If substrate concentrations are low, the rate determining step may not be a single reaction, the AV # of substrate binding can be involved, and interpreting AV #, the slope of a plot as in Fig. 7A, is more complicated. These examples show that the reaction conditions, in vitro, the concentration of enzyme, substrate, and regulatory ligands, and the nature of the ionic environment and the molecular changes involved in particular reactions, all influence the effect which pressure has on enzyme-catalyzed reaction. If we consider such reactions in vivo, there are other important variables, such as the structured environment in which catalysis takes place (lipid bilayer, ribosome), which add to the complexity of relating the overall
1010
SECTION IV Muscle a n d O t h e r Contractile S y s t e m s "13
effect of pressure on a reaction to the molecular volume changes underlying the rate limiting step.
40
4
--I
[
o
"13
3O o f-.
3
m
E
III. Cellular Effects
r
20
w 2 cn
Pioneer experiments with fertilized sea urchin eggs and other oocytes, carried out by Marsland and colleagues, showed that high hydrostatic pressure (20-60 MPa) reduced the gel strength of the cortical cytoplasm, simultaneously blocking and reversing the contractile cleavage furrow (Marsland, 1970). More recent experiments (Begg et al., 1983) have shown that filament bundles of actin in microvilli are disorganized, but recover on decompression. Generally the cytoplasm of eukaryote cells is reversibly "liquefied" by high pressure, largely caused by the depolymerization of actin stress fibers. As a result, cells round up and amoeboid movement and cytokinesis are reversibly blocked. Microtubules, polymers of tubulin (Chapter 6), are variously susceptible to high pressure (Fig. 8). In the projecting cytoplasmic spines of the protozoan A c t i n o p h r y s , the microtubules are somewhat differently organized but similarly susceptible to pressure (Kitching, 1970), whereas the microtubules in HeLa cells, and osteosarcoma cells are relatively resistant (Crenshaw et al., 1996). Also in HeLa cells, intermediate filaments composed of cytokeratin (Chapter 36) are readily depolymerized by pressure, but myosin II filaments are not. Although pressure generally depolymerizes proteins in vitro (Section II), intracellular structures such as actin filaments, microtubules, and cytokeratin intermediate filaments are now thought to be indirectly affected by pressure. There are at least two reasons for this. As we have seen, the same protein, for example, tubulin, polymerized into microtubules in different cells, shows radically different susceptibilities to high pressure, presumably reflecting the state of the polymerizing process in vivo. Furthermore, the state of cells in culture influences the susceptibility of their intracellular structures. In particular HeLa cells growing in contact with each other fail to round up at high pressure (30 MPa), whereas isolated cells round up and lack actin stress fibers (Crenshaw et al., 1996). A specific attempt to measure cytoplasmic [Ca 2+] in mouse fibroblasts caused to round up by 40 MPa failed to reveal any significant change. Nevertheless, other protein-regulating mechanisms such as phosphorylation could be responsible for the structural changes (Crenshaw and Salmon, 1996). If indeed pressure acts indirectly then the apparent A V cited in the case of oocyte microtubules (Fig. 8) is merely an arbitrary numerical index and does not directly relate to tubulin polymerization. The widespread disruption of cytoskeletal
i"1"i Z
~ T
When cells and tissues are subjected to hydrostatic pressure, numerous processes and reactions are affected, usually reversibly. This section should be read with due reference to Chapters 6, 13-18, 24-27, and 51-56, and with two particular questions in mind: (1) Are the effects caused directly or indirectly by pressure? (2) What useful and illuminating generalizations can be drawn from these effects?
A. Cell Structures
9 rrn
tll _J c3
3=
z
10 g
1
n
__
,
!
~-~2
1
I
4
I
-
I
I
6
I
"1 I
I
I
8R~2
I
4
I
!
I
6
I
8
|
I
1
I
10
T i m e (min)
B
2.0 v
1.0 O~ c" O
0
0.4
E
.,=_ .-9_
= 0.2
uJ 0.1
0
5
10
15
Pressure (MPa)
FIGURE 8. High pressure causes microtubules to depolymerize. (A.) Birefringence retardation, (BR, left hand axis ( i ) a measure of the concentration of orientated tubulin, and length of the metaphase spindle (right hand scale (0) in an oocyte of Chaetopterus, subjected to 13.6 MPa, 22 ~ at P and decompressed to atmospheric pressure 9 minutes later at R. The approach to the equilibrium birefringence value is rapid (11). (B) The polymer-monomer equilibrium constant (K) is [B]/[Ao-B], in which A o is the maximum birefringence (fully polymerized state) and B the value under given conditions. K is plotted on a log scale as a function of pressure, and from Eq. 6 an apparent AV of 400 ml.mo1-1 at 22 ~ is calculated. (From Inou6 et al., 1975).
proteins which high pressure causes, directly or indirectly, may well affect cellular processes which seem somewhat remote from cell structure per se. For example, the distribution and activity of ion channels are intimately associated with the state of cytoskeletal proteins (Chapter 36) and so too are transport and permeability processes in epithelia. Furthermore, the association of mRNA with the cytoskeletal proteins is affected (for example, actin mRNA) (Zimmerman et al., 1994). This is one of several aspects of protein synthesis which high pressure inhibits. Another is the pressure dissociation of ribosomes (Gross and Jaenicke, 1994). The pressure blockade of cell cleavage has enabled the ploidy of cells to be manipulated for research and commercial purposes. In fish farming the production of large quantities of triploid salmon and trout are achieved by pressurizing the fertilized oocyte to block the shedding of a polar body (Lakra and Das, 1998). Triploid female fish do not sexually mature, and their marketable quality does not decline upon reaching maturity.
59. Effects of High Pressure on Cellular Processes
B. Muscle Muscle, a tissue whose cells are conspicuously filled with polymerized proteins, is affected by high pressure in ways determined by the contraction cycle. Some of these were discovered by the first physiologists to apply high pressure techniques to eukaryote cells (Table 3). Generally pressure has no depolymerizing effect on muscle proteins in vivo. Consider first the mechanical responses of skeletal muscle fibers (rabbit psoas) whose plasma membrane has been removed. Such "skinned" fibers generate tension which can be initiated by the experimenter controlling the perfusing solution (Chapter 56). The tension in relaxed fibers is unaffected by 10 MPa, but fibers maximally activated by Ca 2+ show a reduction in tension. In the latter condition, cross-bridge cycling is inhibited by pressure, whereas in the relaxed state cross bridges are absent. A third state, rigor, is interesting because cross bridges are formed but not cycling. In this state, high pressure increases tension by the bulk compression of the protein matrix, in much the same way as it affects collagen (Ranatunga et al., 1990). In intact skeletal and cardiac muscle, twitch tension is accelerated and relaxation slowed by high pressure (Fig. 9). A major cause of this is an enhanced [Ca2+]i transient following the action potential. The [Ca2+]i transient is the net result of Ca 2+ release from, and simultaneous transport back into, the sarcoplasmic reticulum, and pressure's main effect appears to be the inhibition of Ca 2+ transport. The enhanced [Ca2+]i more than offsets the simultaneous and lesser inhibition of cross-bridge cycling. This latter effect is manifest in the reduced tetanic tension which is produced at high pressure (Fig. 9). However, when the tetanic tension is markedly reduced in severely fatigued muscle, pressure again increases the tension because the tension-limiting process switches from cross-bridge cycling to CaZ+-activation (Vawda et al., 1996). Thus positive and negative inotropic effects of pressure can be revealed by manipulating the limiting reaction. This was elegantly demonstrated by Brown (1936) in brilliant pioneering work. He applied 40 ms pulses of hydrostatic pressure at different times in the twitch process. A pressure pulse applied just before or at the same time as the electrical stimulus caused the same increase in tension as is seen during continuously applied pressure. The same pressure pulse applied 40 ms after the stimulus, had no effect. These pressure jump experiments, recently confirmed by Vawda et al. (1995), showed that pressure enhanced twitch tension by affecting an early stage in excitation-contraction coupling. This is now known to be the increase in [Ca2+]i. In vertebrate cardiac muscle, pressure also increases the force generated but two important differences from skeletal muscle are noted. The first is that an isolated heart shows spontaneous, rhythmic contractions. Pressure reduces the rate of spontaneous beating; for example, in the isolated rat atria, 10 MPa reduces the beat frequency by 22%, and increases the force of contraction by approximately 50% (Ornhagen and Sigurdsson, 1981). Part of the latter is caused by a prolonged action potential, which can mobilize more calcium to initiate contraction (see later discussion). If the beat frequency is electrically paced, however, the same pressure
1 01 1
causes a 20-30% increase in force of contraction, which probably arises from the way cardiac muscle regulates [Ca2+]i, the second distinctive feature to consider. The idea here is that pressure acts like ouabain, inhibiting the Na+-K + pump. The result is that [Na+]i increases and causes [Ca2+]i to increase, by biasing the Na+-Ca 2+ exchange mechanism. (Chapter 52; and Hogan and Besch, 1993).
C. Membrane Permeability and Transport Simple processes, such as the diffusion of gases or water in aqueous solution, are little affected by the pressure range of interest here. However, the diffusion of K + across liposome bilayers is a more complicated process, involving adsorption and translocation, and is much more sensitive to pressure. When facilitated by valinomycin, the AV# for K + flux through a liposome bilayer is 40 ml.mo1-1 compared to ---20 ml.mo1-1 in the absence of carrier. The translocation of the K-valinomycin complex across a planar bilayer is limited by the viscosity of the interior which pressure increases and the resultant activation volume AV~ is 12 ml.mo1-1 (Benz and Conti, 1986). The formation of the K-valinomycin complex, which probably occurs at the bilayer water interface, appears to involve a small positive AV, and its dissociation, a negative AV. Overall the positive activation volume of translocation predominates, and K + flux is reduced at high pressure. When the carrier-mediated mechanisms of erythrocytes are blocked, the resultant passive diffusion of K, Na, Cs, and Rb is accelerated by pressure (AV# -50 to -90 ml.mo1-1) (Hall et al., 1993). This clearly indicates an unexpected underlying complexity. Additionally there is a KC1 co-transport pathway which Godart and Ellory (1996), and Gibson and Hall (1995) have shown to be indirectly activated by pressure, which favors dephosphorylation. Reassuringly perhaps, the passive flux of some amino acids (e.g. alanine, glycine) in red cells is reduced by high pressure to an extent similar to liposome fluxes. Primary active transport processes, which use metabolic energy directly, are susceptible to high pressure through a variety of mechanisms. As with enzyme kinetics we may distinguish changes in K m (a measure of the affinity of the transporter for the molecule it transports) and changes in the maximum rate of transport (i.e., Vmax). The ubiquitous Na+-K + pump is inhibited by pressure in erythrocytes (human, fish, frog skin, and fish gill) (Hall et al., 1993). In human erythrocytes, 40 MPa halves the K m and Vmax for K + influx and Na + effiux at 37 ~ hence [Na+]i increases. The three main reaction steps (Chapter 17) ion binding, occlusion (translocation through the bilayer), and release are each possible sites of action. Another is the uncoupling of the ATPase, energy capturing process, from the pump. Insofar as pressure (15 MPa) activates the ATPase yet inhibits the pump in human erythrocytes, uncoupling is an attractive possibility (Goldinger et al., 1980). The uncoupled ATPase, isolated from dog kidney and fish gill, is inhibited by pressure (40 MPa) which also reduces fluidity of the bilayers which support it (Chong et al., 1985; Gibbs, 1997). The Ca 2+ pump in the sarcoplasmic reticulum of mammalian muscle (Chapter 18) has been studied by Hasselbach and colleagues (Stephan and Hasselbach,
1012
SECTION IV Muscle and Other Contractile Systems Atmospheric Pressure
Z[
8 MPa
1.6
E
2v
t~
c5 t~
I
I 0.9
100 ms
12 1-6
-
E ::L O .m
.~_ 1-4
-
c 1-2
-
~-
eq)
..9
0 J~
.9 9
9
I I
/% /
"~.
400
ms
o
e-
9~-
~o 1.2
-
O .0-, 0"8
0
I
I
I
I
I
2
4
6
8
10
Pressure (MPa)
cO c" .0-, .m c-
o
10
O
....
Oo
0.8 -
0.6
I
0
2
I
I
4 6 Pressure (MPa)
I
I
8
10
FIGURE 9. Positiveand negative inotropic effects of high hydrostatic pressure. (A) Twitch contractions (tension, mN) in a bundle of fibers from rat skeletal muscle are accelerated and enhanced at 2, 4, and 6 MPa. Relaxation is slowed. The lowest curve was obtained at normal atmospheric pressure. (B) Pressure dependence of the twitch tension, as in A, normalized to the atmospheric pressure value. (C) The electrically stimulated rise in [Ca]/(in relative units) in an individual cardiac myocyte (guinea pig) is enhanced at 8 MPa. These data were obtained with an optical high pressure apparatus, enabling both cell length and fluorescence ([Ca]i), to be simultaneously monitored. (D) The negative inotropic action of pressure is manifest in the reduced tetanic tension of fiber bundles normalized to the atmospheric pressure value. [(A), (B), and (D) from Ranatunga and Geeves, (1991), and (C) from Besch and Hogan (1996).]
1991), who have shown that its inhibition by pressure probably causes the increased [Ca2+li transient in cardiac muscle, mentioned above. Both this (~a2+pump and the equivalent one in erythrocyte membranes appear to undergo subunit dissociation at very high pressure (VerjovskiAlmeida et al., 1986; Coelho-Sampaio et al., 1991). Transport across the amphibian epithelia provides an interesting example in which pressure affects the passive permeability to sodium (Hong et al., 1984, Brouha et al., 1970) similar, perhaps, to the pressure diuresis in human divers (Takeuchi et al., 1995). It also affects the insertion of aquaporins (Coluccio et al., 1983). D. Excitable
Membranes
The hyperexcitable effects of pressure, first seen in marine animals in the 19th century and then rediscovered in deep simulated dives with laboratory mammals and humans, has stimulated an interest in electrophysiology at high pressure.
The pressure reversal of anesthesia in whole animals has also been another stimulus. Virtually all electrical recording techniques have been adapted for use at high pressure (Table 3) but our understanding of how pressure "excites" animals (Section IV) is far from clear. 1. S q u i d G i a n t A x o n
Spyropoulos (1957a) was the first to record the prolonged time course of the action potential in the squid axon at high pressure, and Table 3 shows the development of voltageclamp experiments in the giant axon and other neurons. The most refined recordings show pressure slows both the activation and inactivation of the sodium current (Fig. 10). The AV # for the opening of the m gate is 58 ml-mo1-1, and for the equivalent potassium channel n gate it is 37 ml.mo1-1 (Chapter 25). The AV # of the gating current for sodium channels is much smaller (17 ml.mol-1), reflecting the movement of the voltage sensitive part of the channel prior
59. Effects of High Pressure on Cellular Processes TABLE 3
10 13
Electrophysiological Techniques Used with Simple Preparations at High Pressures
Method
Experimental description
Reference
1. Single motor nerve fibres, amphibian,108 MPa
Spyropoulos ( 1 9 5 7 b )
2. Motor nerve, amphibian, 68 MPa
Gershfeld and Shanes (1958)
3. Vagus nerve, deep-sea fish, 40 MPa,
(Glycerol)
2._ H20
Promitochondrion 02 F I G U R E 5. Summary of glycosomal metabolism in bloodstream forms of Trypanosoma brucei. Redox shuttle: glycosomally produced glycerol-3-phosphate (G3P) is oxidized in the promitochondrion to dihydroxyacetone phosphate (DHAP), and this oxidized three-carbon unit returns to the glycosomes. The P* denotes where net formation of ATP occurs under aerobic conditions. Under anaerobic conditions, glycerol is produced in equal amounts to pyruvate. Additional abbreviations: glyceraldehyde-3-phosphate, GAP; 3-phosphoglycerate, 3-PGA; phosphoenolpyruvate, PEP. (Adapted with permission from Fairlamb and Opperdoes, 1986.)
the glycosome (Opperdoes, 1987). Under aerobic conditions, pyruvate is the sole product of the cells, but oxygen is consumed by the promitochondria, a process quite distinct from that found in normal mitochondria (see Chapter 8). Without oxygen, no Pasteur effect is seen (i.e., the rate of glucose consumption does not change), but now glycerol is produced with pyruvate (Fig. 5). Glycerol production occurs through a selective permease in the cell membrane. This process ensures removal of glycerol and thus helps overcome an unfavorable thermodynamic gradient for formation of ATP by a glycerol kinase in the cytosol. Most interestingly, under aerobic conditions the amount of ATP consumed in activating the six-carbon sugars is balanced by formation of 2 mol of 3-phosphoglycerate (3PGA). Thus, the glycosome under these conditions does not provide net formation of ATP for the cell, despite the considerable turnover of ATP. The excretion of pyruvate by the cell leaves unanswered the question of how the glycosome oxidizes reduced nicotinamide adenine dinucleotide (NADH), a rather typical problem for a fermentative pathway. Studies of enzyme activities and their subcellular distribution lead to the conclusion that an unusual C 3 redox shuttle connects the glycosome with the mitochondrion of the cell (see Fig. 5). In the shuttle, dihydroxyacetone phosphate (DHAP) reduction to glycerol-3-phosphate (G3P) results in the regeneration of NAD from NADH for the pathway leading to pyruvate production. The G3P moves to the mitochondrion, where an alternative oxidase regenerates DHAP to resupply the glycosome. The mitochondrion in bloodstream forms is regarded as a promitochondrion since it lacks several components of the mitochondrion. Bloodstream forms do not obtain ATP by the alternative oxidase.
Glycosomal metabolism is controlled to a large extent by glucose transport across the cell membrane. Under certain conditions of glucose excess, several glycolytic enzymes appear to also control the metabolism of glucose. Entry of glucose into the cell occurs by facilitated diffusion with different transport proteins used by bloodstream forms and procyclic forms. These proteins belong to a superfamily of glucose transporters such as Glut l of the mammalian erythrocyte, another cell that lacks mitochondria. The significance of the glycosomal membrane is not understood. This membrane presents a significant barrier to metabolites, since treatment of glycosome enriched fractions with triton is required in order to measure the enzyme activities. Glucose and the products of the organelle appear to have free access. An antiport may function to exchange G3P and DHAP, but two antiport transporters cannot be excluded (Borst and Fairlamb, 1998). Pyruvate and glycerol leave the cell by energy-independent permeases, and at least for pyruvate, transport occurs by facilitated diffusion. While glycosomes do function in bloodstream forms to degrade glucose, several other functions also occur in the organelle, depending on the species and the stage of the life cycle. Procyclic forms rely on respiration by a complete mitochondrion. Glycosomes function in C02 fixation. In addition, they show key enzyme activities for both fatty acid oxidation and ether lipid peroxidation. These enzyme activities, along with the presence of catalase (in a few strains) and protein sequences of glycosomal enzymes, suggest a close relationship between glycosomes and peroxysomes (Opperdoes, 1987; Sommer and Wang, 1994). The glycosome is unique for the presence of the glycolytic enzymes. How this compartmentation benefits the cell is not clear, however. More detailed studies are needed to clarify the transport mechanisms and how the transport processes are regulated, especially during transformation between the vertebrate and insect forms.
B. Hydrogenosomes Hydrogenosomes have been found in several unrelated groups, including parasitic flagellates and free-living protists (MUller, 1993, Yarlett, 2000). Besides these protozoa, at least one rumen fungus has hydrogenosomes. They were first discovered in 1973 in the cattle parasite Tritrichomonas foetus and in the human parasite Trichomonas vaginalis. Hydrogenosomes represent about 5% of the cell protein in the trichomonads. No mitochondria are present in these cells, and they lack the ATPase characteristic of oxidative phosphorylation (MUller, 1993). The organelle has a double membrane and was named for its ability to produce hydrogen gas while oxidizing pyruvate, produced by cytosolic glycolysis (Fig. 6). Hydrogenosomes are morphologically distinct from mitochondria since for most hydrogenosomes no inner membranes are present except for the organelle in a free-living ciliate. Electron microscopy reveals that the organelles elongate and divide, or can be consumed by autophagy (Benchimol et al., 1996). Inhibition of polyamine metabolism in Tritrichomonasfoetus by the putrescine analog 1,4-diamino-2-butanone results in inhibition of cell growth and a degraded appearance of the
61. Physiological Adaptations of Protists
1047
Hydrogenosome
glyeolysis Glucose ........ ,--
/ -)PEP ~
i ....................................
i .......................
Glycerol
Pfo, Fd, H2a.~ Pyruvate ~ ~ .........
: H2 +
,,,
......................
Lactate
............................
Acetate
FIGURE 6. Summaryof T. vaginalis hydrogenosomal metabolism. Gly-colysis occurs by the Embden-Meyerhoff pathway in the cytoplasm. Oxidation of pyruvate occurs in the hydrogenosome by the sequential action of pyruvate: ferredoxin oxidoreductase (Pfo), ferredoxin (Fd), and hydrogenase (H2ase). Net formation of ATP is designated by P* and phosphoenolpyruvate by PER The dashed line indicates that more than one step occurs in the formation of acetate and of ATR (Adapted with permission from MUller, 1993.)
hydrogenosomes, yet the cells retain motility (Reis et al., 1999). The ability to produce hydrogen gas is a highly unusual feature for a eukaryote cell. Cells ferment significant amounts of pyruvate through this organelle with ATP formation occurring by a substrate level phosphorylation step (Fig. 6). The iron-sulfur proteins that metabolize the pyruvate are now used in molecular approaches to identify hydrogenosomes when only a few cells can be obtained. Removal of pyruvate by the hydrogenosome eliminates its use as an electron acceptor for reoxidizing glycolytically reduced NAD. It was discovered recently that T. vaginalis oxidizes the NADH under anaerobic conditions by the production of glycerol. In contrast to trypanosomes, trichomonads lose energy by producing glycerol. Trichomonads have a G3Pase rather than a glycerol kinase. Trichomonads prefer growth under anaerobic or microaerophilic conditions, but do have the ability to consume oxygen (Mtiller, 1993). Likewise, the isolated hydrogenosomes consume oxygen, but in contrast to whole cells this activity is rapidly lost. The cytosol of trichomonads contains a soluble NADH oxidase that directly reduces oxygen to water; this would protect the oxygen-labile components of the hydrogenosome, such as the two iron-sulfur enzymes. Oxygen does not change the rate of glucose consumption, in contrast to aerobic eukaryotic cells, and cells gain no ATP by oxygen reduction. (Oxygen stops hydrogen production by intact cells, but acetate and CO 2 production continue.) The presence of the hydrogenosomal pathway is not essential for growth of trichomonads since several strains of T. foetus and T. vaginalis have been obtained by long-term culture in the presence of metronidazole. This drug is reductively activated to a cytotoxic form by the oxidation of pyruvate. The cells grow at a decreased rate and show an increased rate of glucose consumption, however (Kulda et al., 1993). Free-living ciliates with hydrogenosomes show an unusual symbiotic relationship by having established methanogenic bacteria as endosymbionts (see Fig. 12 Section V.B). The hydrogenosomes in these ciliates have an unusual membrane since it forms invaginations, and thus the organelle structurally resembles a mitochondrion. The methane production pulls reoxidation of NADH against a standard midpoint po-
tential difference o f - 1 0 0 mV by (presumably) removing H 2 produced via NADH. The presence of hydrogenosomes in such diverse anaerobic organisms as flagellates, free-living and symbiotic ciliates, and rumen fungi provides for many variations in structure and function. Additional functions for hydrogenosomes, such as accumulation of Ca and formation of a transmembrane potential, have also been reported for some species (Mtiller, 1993). The hydrogenosome can be regarded as a fermentative analog of a mitochondrion, since both organelles degrade pyruvate and form ATP as a major function. The hydrogenosome is believed to have originated in an endosymbiotic event. DNA has not been found in hydrogenosomes, except for a report that hydrogenosomes of the ciliate Nyctotherus were stained with DNA-specific antibodies (Akhmanova et al., 1998). It has been suggested that the Eukaryotes arose through symbiotic association of an anaerobic, strictly hydrogendependent, strictly autotrophic Archaebacterium host with a Eubacterium symbiont that was able to respire, and which generated molecular hydrogen as a waste product (Martin and Mtiller, 1998).
V. Sensory Adaptations, Membrane Potentials, and Ion Channels Like other cells, protists exhibit sensitivity and respond in various ways to environmental stimuli. The basic physiological and biochemical mechanisms, as they are known, are similar to those found in metazoan systems. Paramecium, for example, has been referred to informally as a "swimming neuron." In this section several examples of photoreceptors and gravireceptors are examined, and then recent work on the underlying transduction mechanisms is discussed.
A. Photoreceptors Ambient light is a source of energy to autotrophic profists. In addition, for motile species, it can serve as a directional signal. Many flagellates, and probably all the motile photosynthetic species, exhibit a positive phototaxis. For this purpose a directional receptor is needed. Directional receptors, or "eyespots," fall into two main categories: (1) receptors with opaque screens, which detect direction by the screen's shadow, and (2) receptors with antennae based on interference and diffraction to detect direction, essentially the same optical principle as that seen in a reflection hologram. In addition to these directional detectors, one group of predatory dinoflagellates has developed a system of ocelloids with intracellular lenses. In this section we discuss the antennae and the ocelloids, two unique but understudied systems. 1. Receptors with Light Antennae The interferometric light antenna is found in a phylogenetically diverse array of flagellates, including many genera of green algae, dinoflagellates, and cryptophytes (Foster and
1048
SECTION VII Protozoa and Bacteria
Smyth, 1980; Melkonian and Robenek, 1984; Smyth et al., 1988). In the cases where experiments have been done, these antennae appear to be selective with regard to wavelength and light direction. The presence of essentially similar organelles in a wide array of groups considered on many other grounds, including molecular homology, to be phylogenetically diverse raises a question: are these all cases of convergent evolution? If not, have these organelles spread through endosymbiotic events? The antennae operate through an interference mechanism similar to that which produces structural color in iridescent objects. The essential structure consists of a number of parallel reflective layers that act as partial mirrors. These are
spaced evenly, separated by a distance of a quarter of the wavelength to be selected. Thus, incident light passes through a series of partially reflecting pigment layers spaced at quarter-wavelength intervals. Normally incident light of the appropriate wavelength is reflected back at each partial mirror. These reflected light rays are in phase with each other and with the incoming radiation, leading to positive reinforcement of this light signal in front of the eyespot, where the receptor pigment is located. The basic principle is illustrated in Fig. 7. Light from other angles, and light of other wavelengths, also passes through this filter, but the reflections are out of phase with each other, and the resultant interference minimizes their contribution. The most intensive studies of light antennae were done with the green flagellate Chlamydomonas (Fig. 8) (Foster and Smyth, 1980), but the antennae are also found in dinoflagellates, chrysophytes, and many other light-responsive flagellates.
Er
2. Intracellular Lenses in Dinoflagellates
"
"N~ / /
\~
//""~
/
\
Z/ X
P.
Y
~-.
s i ::i
L
HLH
Perhaps the most complex and surprising photoreceptors are the ocelloids (ocelli), eyespots found in a family of nonphotosynthetic dinoflagellates, the Warnowiaceae. A typical ocelloid consists of a lens, a retinoid, and an opaque pigment cup (Fig. 9). The pigment cup, or melanosome, consists of a wall of pigment granules, and it surrounds the paracrystalline retinoid. In front of this is the lens, or hyalosome, consisting of a peripheral corneal zone and a central crystalloid body. The entire organelle is approximately 24 ~am long and 15 pm wide (Greuet, 1978, 1987). Francis (1967) measured the refractive index of the lens in Nematodinium by three methods, and found it to be approximately 1.52. He determined the focal plane by tracing rays parallel to the optic axis and found that light was focused in the retinoid layer. The field of view in these species was determined to be about 30 ~ but may be wider in other species. From these observations we conclude that the ocelloids could probably act as directional light receptors.
L
FIGURE 7. Reflection. (A) Perfect mirror. Incident w a v e Ei, reflected w a v e E r , and the sum E = Ei + Er a standing wave: the dashed lines show the extrema. (B) Interference reflector. A segment of a wave one wavelength long at successive instants, showing reflections from partial mirrors spaced quarter-wavelengths apart. The reflected waves are in phase. (From Foster and Smythe, 1980.)
FIGURE 8. Sectional view of the eyespot of Chlamydomonas, showing reflecting pigment layers forming a quarter-wavelength stack. (From Foster and Smyth, 1980.)
61. PhysiologicalAdaptations of Protists
.
1049 image on the retinoid is maximal. While certainly plausible, such behavior has not been reported yet. Unfortunately, these species have not been cultured, and there appear to be no studies of ocelloid function, so that all we have at this time are speculations based on the structure.
t
mit '
micr
H ,
;
o
i
OCh
,~
-pig
A
B. Gravity Receptors in Ciliates Many single-celled organisms orient with respect to the earth's gravitational field while swimming. In some, such as the well-studied ciliate Paramecium, this is thought to be a purely hydrodynamic effect due to the shape and the distribution of mass in the cell: as the organism swims, the center of mass pulls the rear of the cell downward, orienting it so that it swims upward (Roberts, 1981). In this case, then, there seems to be no need for a gravity receptor. However, many protists can migrate either up or down, switching between a positive and a negative geotaxis in response to external signals or to a circadian or a tidal rhythm, suggesting a more complex type of response. In particular, a group of ciliates containing the genera Loxodes and Remanella has a gravity response that depends on the dissolved oxygen level. In nature, these ciliates collect at the interface between anaerobic and aerobic zones in the water column. In oxygen-containing water they swim downward, while in an anoxic environment they swim upward. They all contain a characteristic organelle, the MUller body, which acts as a gravity receptor. This is a fluid-filled vesicle, in which a membrane-covered mineral body is suspended (Fig. 10). The mineral body, consisting of a strontium or barium salt, acts as a statolith, and its position on the interior surface of the vesicle appears to inform the cell of its orientation in the gravitational field (Fenchel and Finlay, 1986). The vesicle is anchored to the system of connected cilia of the cell, and it is suggested, by analogy to other mechanoreceptors, that oriented mechanical stress on the cell membrane at this point of connection may lead to a depolarization, which in turn would affect ciliary beating and determine the orientation of the swimming motion (Fenchel and Findlay, 1986) (see also Section VIII).
B FIGURE 9. (A) Diagram of structure of Nematodinium ocellus, showing interior structure: crystalline body (Cr), ocelloid channel (f), constricting ring (cr), periocelloid gallery (g), hyalosome (H), lamellae of retinal body (1), microcrystalline layer (micr), mitochondrion (mit), ocelloid chamber (OCh), pigmentary ring (pig), basal plate (lb), microtubular layer (t), vesicular layer of the retinal body (v). Magnification: 3900X. (B) Position of ocellus in the Nematodinium cell, showing parts of the cell: episome (E), hyposome (HO), intercingular sulcus (sic), cingular sulcus with transverse flagellum (str), posterior flagellum (fl). (From Greuet, 1978.)
Given the presence of nematocysts (see Section VI.C) in these same dinoflagellate species, and the fact that they are predators rather than autotrophs, it has been suggested that the ocelloids might act as "range finders," leading to discharge of nematocysts when the contrast of the focused
C. Sensory Transduction: Membrane Potentials, Ion Channels, and Intracellular Components 1. Membrane Potentials, Calcium, and Behavior As with many aspects of protistan physiology, membrane potentials are best known in the large ciliate Paramecium. A classic feature of ciliate behavior is the avoidance reaction, in which the cell stops, swims backward for a short time, then swims forward again, usually in a somewhat different direction. Early workers observed prolonged backward swimming in cells subjected to various stresses: high temperatures, sudden changes in pH or osmotic pressure, exposure to solvents and other deleterious chemicals. The common denominator in all these proved to be that the cell membrane became "leaky" or permeable to Ca 2+ and other ions. Using the methods developed by Szent-Gyorgy and others for muscle tissues, Paramecium cells were exposed to
1050
SECTION VII Protozoa and Bacteria
9
"~
,-
9
, - . 9
~
Q
02+ 6 H20
AG o - - 6 8 6 kcal/mol Many of the proteins photosynthetic bacteria used for photosynthesis could also be used for oxidative phosphorylation. Several lines of evidence suggest that purple photosynthetic bacteria might have been the ancestors of nonphotosynthetic bacteria that perform oxidative phosphorylation, and that the mitochondria of eukaryotic cells evolved from such bacteria, which had formed symbiotic associations with primitive eukaryotic cells (Woese, 1987). Thus, it is no coincidence that the mechanisms of photosynthesis and respiration are similar in many respects. Much of our understanding of oxidative phosphorylation has come from studies of photosynthesis, which have exploited the experimental advantages of a system that can be driven by light. In algae and higher plants, photosynthesis takes place in organelles known as chloroplasts. There is evidence that chloroplasts evolved from oxygen-evolving photosynthetic
bacteria such as cyanobacteria or Prochloron, which had formed symbiotic associations with eukaryotes. This process appears to have occurred several times. The evolution of photosynthesis is discussed in excellent on-line photosynthesis tutorials (Whitmarsh and Govindjee, 2000; Cramer, 1999). A shorter tutorial is offered by Kramer (2000). Links to a number of very useful photosynthesis resources, many of which include interactive graphics and animations, are presented by Orr and Govindjee (2000) and Roberts (2000).
I!. Chloroplasts In higher plants, most chloroplasts are found in the mesophyll cells of leaves (Fig. 2). The chloroplast consists of a collection of flattened membranous vesicles called thyiakoids, which are enclosed in an envelope formed by a double membrane (Fig. 3). The outer membrane contains channels formed by the protein porin and is freely permeable to substances whose molecular mass is below about 10 kDa. The inner membrane is selectively permeable and contains a number of metabolite transporters, which mediate the interaction between metabolism in the plant cytosol and within the chloroplast (Fltigge and Heldt, 1991). The volume surrounding the thylakoids is known as the stroma. The thylakoid membranes contain the pigments that capture light, and, like the cristae of mitochondria, they contain an electron transport system coupled to an adenosine triphosphate (ATP) synthase. The soluble enzymes responsible for carbon dioxide assimilation are found in the stroma. In many chloroplasts of land plants and green algae, the thylakoid membranes are stacked tightly together in some regions to form structures called grana, which look like green grains under the light microscope. The grana are connected by nonstacked thylakoid extensions, the stroma lamellae. Chloroplasts still retain genes, which are located on a chromosome found in the stroma, for many of the proteins involved in photosynthesis and for transfer and ribosomal RNA (Rochaix et al., 1998). Proteins coded by chloroplast genes are synthesized on ribosomes located in the stroma. Genes for other chloroplast proteins are now located on nuclear chromosomes of the plant cell. Remarkably, many multi-subunit proteins of the chloroplast thylakoid, including the ATP synthase, contain both subunits coded on nuclear genes and synthesized on cytosolic ribosomes and subunits coded on chloroplast genes and synthesized on stromal ribosomes. How the plant coordinates the synthesis of these subunits and the assembly of the proteins presents an intriguing research problem. Chloroplast proteins that are synthesized in the plant cytosol are initially formed with amino terminal extensions, termed transit peptides, which allow them to bind to receptors on the outer envelope membrane and then enter the chloroplast. The process requires GTP, ATP, and molecular chaperones (Schnell, 1995). After the proteins have entered the chloroplast stroma, the transit peptides are removed by a highly specific protease. Proteins destined for the membrane or lumen of the thylakoid are formed with an additional presequence, which is proteolytically cleaved after it has allowed the protein to enter the thylakoid membrane or lumen.
64. Photosynthesis
1099
FIGURE 2. Transmission electron micrograph of corn (Zea mays) chloroplasts. Upper left: chloroplast within mesophyll cell. Lower right: chloroplast within bundle sheath cell. Arrows indicate plasmodesmata through which metabolites are exchanged between the cells. The mesophyll cells lie near the leaf surface. The bundle sheath cells lie near the leaf vascular bundles. Note the grana in the mesophyll chloroplast. (Micrograph courtesy of lain M. Miller.)
I11. Biochemistry of Carbon Assimilation A. The Reductive Pentose Phosphate Cycle To determine the sequence of reactions involved in the conversion of carbon dioxide to sugars, in the early 1950s Melvin Calvin and his associates exposed algae to 14C02 and light for different periods. After treatment with boiling alco-
hol, the carbon compounds that had been formed were separated by paper chromatography. The manner in which the distribution of 14C among the carbon atoms of the various carbon compounds changed with time revealed the operation of a cyclic process, which has become known as the reduc-
tive pentose phosphate cycle, or Calvin cycle (Fig. 4). In the first step, CO 2 combines with ribulose 1,5-bisphosphate to form two molecules of 3-phosphoglycerate. The
1100
SECTION VIII Plant Cells, Photosynthesis, and Bioluminescence
TRIOSE PHOSPHATE Transporter
outer membrane
~ 1 ~ ~ ,GRAINt~TARCH"" - m ' ~ " ' ~ ~ . ........ :
.......
TRANSIT PEPTIDE
'"" " - ~ ~ ~ _
~ LAMELLA THYLAKOIDS ~ ~
.~ ENTERING PROTEIN
FIGURE 3.
Diagram of a chloroplast.
CO2
2 ATP
2 1,3-bisphosphoRibulose 1,5L~ ~ 2 3-phosphoglycerate L glycerate bisphosphate Rubisco NADPH p . ~ i
Glyceraldehyde 3-phosphate
t
Starch
Dihydroxyacetone phosphate
P.
ATP
I
Fructose 69 phosphate
._L
._ Erythrose4"- phosphate Xylulose 5phosphate
Fructose 1,6bisphosphate
~" -~ Sucrose
-I
Sedoheptulose ~ J 1,7-bisphosphate
..Sedoheptulose "7-phosphate
1
Ribulose 5phosphate
i
i
Hi
F I G U R E 4. The reductive pentose phosphate or Calvin cycle. Stoichiometries are shown only for the first two steps. Erythrose, xylulose and sedoheptulose are sugars that contain 4, 5, and 7 carbons, respectively. Stars indicate steps catalyzed by light-regulated enzymes.
64. Photosynthesis
1 10 1
reaction is catalyzed by ribulose 1,5-bisphosphate carboxylase/oxygenase, often referred to as "rubisco." This enzyme has a rather low affinity for CO 2. To compensate, it is present in extremely high concentrations in the chloroplast and is probably the most abundant protein on earth. Next, the 3phosphoglycerate is phosphorylated to form 1,3-bisphosphoglycerate, which in turn is reduced to glyceraldehyde 3-phosphate by nicotinamide adenine dinucleotide phosphate (NADPH). In the remaining steps of the cycle, ribulose 1,5bisphosphate is regenerated by a pathway that includes a complicated series of rearrangements catalyzed by transketolases and aldolases. Intermediates in the cycle, including fructose 6-phosphate, glyceraldehyde 3-phosphate, and dihydroxyacetone phosphate, serve as precursors for the starch, sucrose, and amino acids that are the final products of photosynthesis. ATP is required for the conversion of 3-phosphoglycerate to 1,3-bisphosphoglycerate and for the conversion of ribulose 5-phosphate to ribulose 1,5-bisphosphate. NADPH is required for the reduction of 1,3-bisphosphoglycerate to glyceraldehyde 3-phosphate. For each molecule of CO 2 fixed, three molecules of ATP and two molecules of NADPH are needed. It is the function of the light-driven reactions in the thylakoid membranes to furnish this ATP and NADPH. B. Photorespiration
and C 4 Plants
Not only does ribulose 1,5-bisphosphate carboxylase/oxygenase have a low affinity for CO 2. It also catalyzes a competing reaction, in which 0 2 rather than CO 2, is added to ribulose-1,5-bisphosphate (Fig. 5). The products are 3-phosphoglycerate and phosphoglycolate. This reaction seems to serve no useful purpose. Many plants grow much faster in a CO2-enriched atmosphere in which the carboxylation reaction can compete more effectively with the oxygenation reaction. It appears that in the 2 billion years since plants began to fill the atmosphere with oxygen while removing CO 2 from it, they have been unable to modify the enzyme so that its affinity for C O 2 is increased or its affinity for 0 2 is decreased significantly. Molecular biologists are now trying to accomplish that task.
Part of the carbon appearing in phosphoglycolate is rescued. In a series of reactions occurring in peroxisomes and in mitochondria, two molecules of phosphoglycolate are converted to one molecule of glycerate, which is returned to the chloroplast and rephosphorylated. But one carbon atom is lost as CO 2, and ATP and 0 2 are consumed. This process is known as photorespiration. It is not coupled to ATP formation, and the net result is waste of ATP and fixed carbon. A number of plants, including corn, sugarcane, and crabgrass, partially avoid photorespiration by concentrating CO 2 in the cells that contain the Calvin cycle enzymes, so that carboxylation can compete more effectively with oxygenation. In these plants, the Calvin cycle enzymes are located in the chloroplasts of the bundle sheath cells, which surround the vascular bundles deep within the leaves (Fig. 6). CO 2 is first captured in the mesophyll cells, which lie near the leaf surface, by carboxylation of phosphoenolpyruvate (PEP): PEP carboxylase
PEP + CO 2
~ oxaloacetate + Pi
The carboxylation is catalyzed by PEP carboxylase, an enzyme that has a high affinity for CO 2 and does not catalyze an oxygenation reaction. Oxaloacetate is next reduced to malate (or transaminated to form aspartate in some plants). The malate or aspartate is transferred to the bundle sheath cells through fibers known as plasmodesmata. Some of these can be seen at the arrows in Fig. 2. Malate is oxidatively decarboxylated to pyruvate in the bundle sheath cells in a reaction that also generates NADPH. The CO 2 that is released enters the Calvin cycle in the bundle sheath cells. Pyruvate returns to the mesophyll cells, where it is again transformed to PEP. The net result is the transfer of CO 2 and NADPH to the bundle sheath cells. The CO2-concentrating mechanism consumes energy, since ATP is converted to adenosine monophosphate (AMP) and phosphate. Nevertheless, plants that use it are often referred to as efficient plants because avoidance of photorespiration more than compensates for this expense. Such plants are usually known as C 4 plants because of the involvement of four-carbon acids. Many C 4 plants are native to tropical areas, where the C 4 pathway is especially advantageous since high temperatures and bright sunlight encourage photorespiration.
H2C- O- POzo
I
C=O
I HC-OH
+
I HC-OH
02
Ribulose bisphosphate carboxylase/ oxygenase
c% I
HC-OH
I
+
I H2C-O-PO 2-
H2C- O- PO 2 -
3-Phosphoglycerate
2-Phosphoglycolate
I H2C-O-PO 2Ribulose 1,5bisphosphate
FIGURE 5.
The oxygenation reaction catalyzed by ribulose 1,5-bisphosphate carboxylase.
1102
SECTION VIII Plant Cells, Photosynthesis, and Bioluminescence
IV. Formation of ATP
CO2
AIR
A. Light-Driven Electron Transport and ATP Synthesis Carbon dioxide fixation requires ATP and NADPH. It seemed reasonable to suspect that the role of light is to provide the energy necessary for their formation. Photosynthetic membranes contain electron transport chains much like those of mitochondria, and light can drive electron transport along the chains (Figs. 7 and 8). As in mitochondria, ATP formation is coupled to the electron transport. How can electron flow through the electron carriers shown in Fig. 7 cause ATP formation? Mitchell (1961) suggested that electron transport and ATP formation are both coupled to the movement of protons across membranes. His chemiosmotic hypothesis helps us understand the reason for the arrangement of the electron carriers within photosynthetic membranes (Mitchell, 1979).
MESOPHYLL CELL
Pi CO,~NADPHNADP + ~ ' ~ ~ ~ . ~ ( ~ H2 ~ i 0'~ ,
=O
H2
CO}
/ Oxaloacetate
HC-OH
,c-o-Po:~-
CO}
CH2
,
C,H2
a
H?-OH CO~ Malate NADP+
CO2 Phosphoenolpyruvate
. ~ A M P + PPi I~ Pi + ATP H3 C--O
Pyruvate '
i~ ~NADPH CO2
B. The Chemiosmotic Hypothesis It was known that electron transport in mitochondria and chloroplasts is accompanied by the movement of protons across the cristae or thylakoid membranes. Mitchell suggested that the transmembrane pH difference (often called a proton gradient) that is formed is an intermediate in phosphorylation. Electron transport would drive proton translocation, and the proton gradient in turn would drive ATP synthesis. Two experiments involving photosynthetic material played a critical role in convincing scientists that the chemiosmotic hypothesis must be taken seriously. Chloroplasts will synthesize ATP even if they are illuminated in the absence of ADP and Pi and then left in the dark
1~O,~ Pyruvate BUNDLESHEATHCELL
FIGURE 6. The mechanism used by C 4 plants to concentrate CO2 and NADPH in bundle sheath cells. CO2 is captured in the mesophyll cells by carboxylation of phosphoenolpyruvate. Malate is transferred to the bundle sheath cells, where CO2 is released and NADPH is formed to enter the Calvin cycle. Pyruvate returns to the mesophyll cells.
-1.5
P700,~~50ps -1.0 ~k
P680~3 ps
"A0 30 ps "~.~ 2OO ns
FX FA FB
-0.5
FD
~300 ps
2 Its
/ c y t b H
i
> i_
o
]
?
i
c/) i
I
_.m 0.01 "I,
"~I \
/
_o~"
0
Ceils noncycling
i
1)
and I
i
0.001
'
0
!
'
2
4 6 8 Hours between two doses (7.63 and 7.95 Gy)
I
'
I
'
I
'
I
10
/
'
12
'
0
!
2
'
I
'
4
I
6
Hours
between
'
I
"
8 two
I
10
'
I
12
doses
(7.42 and 8.04 Gy)
FIGURE 12. (A) Survival of Chinese hamster cells exposed to two fractions of X-rays and incubated at room temperature for various time intervals between the two exposures. (B) Survival of Chinese hamster cells exposed to two fractions of X-rays and incubated at 37 ~ for various time intervals between the two doses. The survivors of the first dose are predominantly in a resistant phase of the cell cycle (late S). When the interval between doses is about 6 h, these resistant cells will have moved to the G2/M phase, which is sensitive. (Redrawn from Elkind et al., 1965.)
|
14
1194
separated by 30 min, the surviving fraction is already appreciably higher than for a single dose. As the time interval is extended, the surviving fraction increases until a plateau is reached at about 2 h, corresponding to a surviving fraction of 0.02. This represents about four times as many surviving cells as for the dose given in a single exposure. A further increase in the time interval between the dose fractions is not accompanied by any additional increment in survival. The increase in survival in a split-dose experiment is due to the repair of sublethal radiation damage. This illustrates the phenomenon of repair without the complication of movement of the cell through the cell cycle. Figure 12B shows the results of the parallel experiment in which cells were exposed to split doses while being maintained at their normal growing temperature of 37 ~ The pattern of repair seen in this case differs from that observed for cells kept at room temperature. In the first few hours prompt repair of sublethal damage is again evident, but at longer intervals between the two split doses the surviving fraction of cells decreases again. An understanding of this phenomenon is based on the age-response function described earlier. When an asynchronous population of cells is exposed to a large dose of radiation, more cells are killed in the sensitive than in the resistant phases of the cell cycle. The surviving population of cells, therefore, tends to be partly synchronized. In Chinese hamster cells, most of the survivors from a first dose are located in the S phase of the cell cycle. If about 6 h are allowed to elapse before a second dose of radiation is given, this cohort of cells will progress around the cell cycle and will be in G 2 or M, a sensitive period of the cell cycle, at the time of the second dose. If the increase in radiosensitivity in moving from late S to the G2/qVI period exceeds the effect of repair of sublethal damage, the surviving fraction will fall. The pattern of repair shown in Fig. 12B is, therefore, a combination of three processes occurring simultaneously. First, there is the prompt repair of sublethal radiation damage. Second, there is progression of cells through the cell cycle during the interval between the split doses, which has been termed reassortment. Third, there is an increase in surviving fraction due to cell division, or repopulation, when the interval between the split doses is 10-12 h because this exceeds the length of the cell cycle of these rapidly growing cells. A simple experiment, performed in vitro, illustrates the three "R's" of radiobiology" repair, reassortment, and repopulation. It should be emphasized that the dramatic dip in the split-dose curve at 6 h, caused by reassortment, and the increase in survival by 12 h, because of repopulation, are seen only for rapidly growing cells. Hamster cells in culture have a cycle time of only 9 or 10 h. The time sequence of these events would be longer in more slowly proliferating normal tissues in vivo. Repair of sublethal radiation damage has been demonstrated in just about every biological test system for which a quantitative endpoint is available. Figure 13 shows that when a dose of radiation is split into several fractions, each separated by a time interval sufficiently long for sublethal damage to be repaired, more cells survive than for the same total dose given in a single fraction, because the shoulder of the curve must be repeated
SECTION IX Cell Division and Programmed Cell Death
tim
O1 r~ > >
om, ~
Fractionated doses
s_
01 0 ..i
dose
!
!
I
Dose (Gy)
FIGURE 13.
The effect of fractionation when radiation is deliv-
ered in a series of fractions. With a time interval between fractions sufficiently long for sublethal damage to be repaired, fewer cells are killed than for the same dose delivered in a single exposure.
with each fraction. In general, there is a good correlation between the extent of repair of sublethal damage and the size of the shoulder of the survival curve. This is not surprising, since both are manifestations of the same basic phenomenon: the accumulation and repair of sublethal damage. Some mammalian cells are characterized by a survival curve with a broad shoulder, and split-dose experiments then indicate a substantial amount of sublethal damage repair. Other types of cells show survival curves with a minimal shoulder, and this is reflected in more limited repair of sublethal damage. In the terminology of the linear-quadratic (c~//3) description of the survival curve, it is the quadratic component (13) that causes the curve to bend and results in the sparing effect of a split dose.
X. The Mechanism of Sublethal D a m a g e Repair As previously mentioned, there is a good correlation between cell killing and the production of asymmetrical chromosomal aberrations, such as dicentrics. This in turn is a consequence of an interaction between two (or more) double-strand breaks in the DNA. On this interpretation, the repair of sublethal damage is simply the repair of DSBs. When a dose is split into two parts separated by a time interval, some of the DSBs produced by the first dose are rejoined and repaired before the second dose. The breaks in two chromosomes that must interact to form a dicentric may be formed by (1) a single track breaking both chromosomes or (2) separate tracks breaking the two chromosomes. The component of cell killing that results from single-track damage will be the same whether the dose is given in a single exposure or fractionated. The same is not true of multiple-track damage. If the dose is given in a single exposure (i.e., two fractions
69. Effects of Ionizing Radiation on Cells
with t = 0 between them), all breaks produced by separate electrons can interact to form dicentrics. On the other hand, if the two dose fractions, D/2, are separated by (for example) 3 h, breaks produced by the first dose may be repaired before the second dose is given. Consequently, there will be fewer interactions between broken chromosomes to form dicentrics and more cells will survive. On this simple interpretation, the repair of sublethal damage reflects the repair and rejoining of DSBs before they can interact to form lethal lesions. This interpretation readily accounts for the repair of radiation damage in cells where mitotic death dominates. In cells that die an apoptotic death, survival tends to be an exponential function of dose, that is, the survival curve is straight with no shoulder on the usual semilog plot. In this case, there is no sparing from a split-dose exposure.
A. Repair and Radiation Quality For a given biological test system, the size of the initial shoulder on the acute survival curve and, therefore, the amount of sublethal damage repair indicated by a split-dose experiment varies with the type of radiation used. The shoulder is largest for X-rays, smaller for neutrons, and nonexistent for densely ionizing a particles, where survival is an exponential function of dose. The amount of sublethal damage repair follows the same pattern; largest for X-rays, smaller for neutrons, and nonexistent for ~ particles.
B. The Dose-Rate Effect For X- or gamma-ray doses, rate is one of the principal factors that determines the biological consequences of a given absorbed dose. As the dose rate is lowered and the exposure time extended, the biological effect of a given dose is generally reduced. Continuous low dose-rate irradiation may be considered to be an infinite number of infinitely small fractions; consequently the survival curve under these conditions would also be expected to have no shoulder and to be shallower than for single acute exposures. The magnitude of the dose-rate effect from the repair of sublethal damage varies enormously between different types of cells. Cells characterized by a survival curve for acute exposures that has a small initial shoulder usually exhibit a modest dose-rate effect. This is to be expected, since both are expressions of the cell's capacity to accumulate and repair sublethal radiation damage. This is generally true of cells for which apoptosis is an important mechanism of cell death. Cell lines characterized by a survival curve for acute exposures, which has a broad initial shoulder, exhibit a dramatic dose-rate effect. An example is shown in Fig. 14 for Chinese hamster cells; in this cell line, mitotic death dominates and apoptosis is unimportant.
XI. The Oxygen Effect A. Nature of Oxygen Effect Many chemical and pharmacologic agents that modify the biological effect of ionizing radiations have been discov-
1 195
100
E c
10-1 8.6 mGy/min
-~ "r:: ~
0_2
c6 0
m
"-~
0.16 Gy/min
lO-a
10-"
Gy/min 1.07 Gy/min
1 0 -5
5
10
15
20
25
Dose (Gy)
FIGURE 14. Dose-response curves for Chinese hamster cells (CHL-F line) grown in vitro and exposed to cobalt-60 gamma rays at various dose rates. At high doses a substantial dose late effect is evident even between 1.07, 0.3, and 0.16 Gy/min. The decrease in cell killing becomes even more dramatic as the dose rate is further reduced. (Redrawn from Bedford et al., 1973).
ered. None is simpler than oxygen, and none produces such a dramatic effect. The oxygen effect was observed by Schwarz (1910), who noted that the skin reaction produced on his forearm by a radium applicator was reduced if the applicator was pressed hard onto the skin. In England, Mottram (1936) explored the question of oxygen in detail and was the first to discuss the possible importance of the oxygen effect in the radiotherapy of human cancer. Survival curves for mammalian cells exposed to X-rays in the presence and absence of oxygen are illustrated in Fig. 15. The ratio of hypoxic to aerated doses needed to achieve the same biological effect is called the oxygen enhancement ratio (OER). For sparsely ionizing radiations, such as X- and gamma rays, the OER at high doses has a value of between 2.5 and 3. The OER has been determined for a wide variety of chemical and biological systems with different endpoints, and its value for X-rays always tends to fall in this range.
B. Mechanism of Oxygen Effect In practical terms, oxygen must be present during irradiation for its sensitizing effect to be observed. In fact, sophisticated experiments have shown that oxygen still sensitizes if if is added a few m i c r o s e c o n d s after a brief pulse of radiation. This is the clue to the mechanism of the oxygen effect since microseconds correspond to the lifetime of the free radicals formed when X-rays interact with water.
I 196
SECTION IX Cell Division and Programmed Cell Death 10 0 ~
i
I
i
I
i
I
D. Concentration of Oxygen Required
i
~\,, r
,.
1.o ~
I=
O r L_ r ~ > .-> t=.
10-1
Only a small amount of oxygen is required to produce the dramatic and important oxygen effect characteristic of X-rays. An oxygen concentration of 3 mm Hg (about 1/2%) results in a radiosensitivity halfway between complete hypoxia and 100% oxygen. By a concentration of 30 mm Hg, characteristic of most normal tissues in the body, the radiosensitivity of cells is indistinguishable from that in air or even in 100% oxygen.
03
Hypoxic
OER 3.0
L)
10 -2
E. Oxygen Effect in Radiotherapy of Human Cancer
Aerated
10-3
t 0
~ 10
2O
30
40
Dose (Gy)
FIGURE 15. Survivalcurves for mammalian cells exposed to Xrays under aerated and hypoxic conditions produced by passing a stream of pure nitrogen over the cells. These data are typical of many in the literature. Oxygen is a dose-modifying factor; that is, at all levels of cell survival, the dose required under hypoxic conditions is three times greater than that required under aerated conditions to produce the same biological effect. This ratio of doses is known as the oxygen enhancement ratio (OER).
Oxygen is essential to "fix" the damage to DNA resulting from radiation-produced free radicals. In the absence of oxygen, radical induced damage is readily repaired, but if oxygen is present, the damage is made permanent, that is, fixed. (The word fix here is used in the European sense of making something permanent, as in fixing a film, not in the American sense of repairing, as in fixing a flat tire.) Consequently, oxygen modifies only that component of radiation damage mediated by free radicals--the indirect damage. Oxygen has no effect on the direct damage caused by ionization of DNA.
C. Oxygen Effect for Different Types of Radiation The oxygen enhancement ratio for X- or gamma rays is approximately 3. It is this large because two-thirds of the damage produced by X-rays is mediated by free radicals. Densely ionizing c~ particles kill cells primarily by direct action, that is, by ionization of the DNA, a process not dependent on oxygen. Consequently, the OER for low energy (2-MeV) ce particles is 1.0. The OER for neutrons has an intermediate value of about 1.6, because while the direct action is dominant there is still a contribution from free radical damage. In summary, the oxygen effect is large and important in the case of sparsely ionizing radiations, such as X-rays; is absent for densely ionizing radiations, such as c~ particles; and has an intermediate value for fast neutrons.
Malignant tumors tend to outgrow their blood supply so that areas of necrosis (dead cells) are a common feature of most cancers. Oxygen diffusing from blood vessels has a limited range because it is used up by rapidly growing and respiting cancer cells. Consequently, there are pockets of cells distant from capillaries that are hypoxic--still with sufficient oxygen to be viable, but at a low enough oxygen concentration to be resistant to killing by X-rays. This diffusionlimited hypoxia is known as "chronic" hypoxia. Regions of "acute" hypoxia develop in tumors as a result of the temporary closing of a particular blood vessel. If this blockage were permanent, the cells downstream would, of course, eventually die and would be of no further consequence. There is, however, good evidence that tumor blood vessels open and close in a random fashion so that different regions of the tumor become hypoxic intermittently. At the moment when a dose of radiation is delivered, a proportion of the tumor cells may by hypoxic, and consequently resistant to killing by X-rays. The problem of hypoxia, both chronic and acute, is largely overcome in practice by delivering radiotherapy in a large number of small fractions, typically 30 to 40, over a period of 6-8 weeks. Other attempts to eliminate the problem of hypoxia cells has been to use neutrons rather than X-rays (because of their smaller OER) and to develop chemicals that specifically sensitize or kill hypoxic cells.
XII. Radiation Quality and Biological Effects A. Deposition of Radiant Energy When radiation is absorbed in biological material, ionizations and excitations occur that are not distributed at random but tend to be localized along the tracks of individual charged particles in a pattern that depends on the type of radiation involved. For example, photons of X-rays give rise to fast electrons, particles carrying unit electric charge and having very small mass; neutrons, on the other hand, give rise to recoil protons, particles again carrying unit electric charge but having a mass nearly 2000 times greater than that of the electron. Alpha particles carry two electric charges on a particle four times as heavy as a proton. The charge-to-mass ratio of a particles therefore differs from that for electrons by a factor of about 8000. As a result, the spatial distribution of the ionizing events produced by different particles vary enormously. For X-rays, the primary ionizing events are well separated in space and for this reason X-rays are said to be "sparsely ionizing." The ionizing
69. Effects of Ionizing Radiation on Cells
1 197
events produced along the track of a low-energy c~ particle form a dense column, and for this reason c~ particles are referred to as "densely ionizing."
1 -
10
B. Linear Energy Transfer
C .O m
Linear energy transfer (LET), a term introduced by Zirkle (1940), is the energy transferred per unit length of the track. The special unit usually used for this quantity is kiloelectron volt per micron (keV/Ixm) of unit density material. The International Commission of Radiological Units (1962a, 1962b) defined this quantity as follows: the linear energy transfer (L) of charged particles in medium is the quotient of dE/dl, where dE is the average energy locally imparted to the medium by a charged particle of specified energy in traversing a distance of dl. That is, L - dE/dl
(5)
Since most radiations in practice consist of a wide spectrum of energies, LET can only be an average quantity. Typical LET values for commonly used radiations are listed in Table 1. Note that for a given type of charged particle, the higher the energy, the lower the LET and therefore the lower its biological effectiveness. For example, gamma rays and X-rays both give rise to fast secondary electrons; therefore, 1.1-MV cobalt-60 gamma rays have a lower LET than 250-kV X-rays and are less effective biologically by about 10%. By the same token, 150-MeV protons have a lower LET than 10-MeV protons and are, therefore, slightly less effective biologically.
C. Relative Biological Effectiveness Figure 16 illustrates the survival curves obtained for three different types of radiation, namely, X-rays, 15-MeV neutrons, and ce particles. As the LET increases from about 2 keV/pm for X-rays up to 150 keV/pm for a particles, the survival curve changes in two important respects. First, the survival curve becomes steeper. Second, the extrapolation number tends toward unity; that is, the shoulder of the curve becomes progressively smaller as the LET increases. It is evident from Fig. 16 that equal doses of different types of radiation do not result in equal biological effects. One Gray
TABLE
1
Typical LET Values
Radiation
Track avg.
LET (keV/Ixm)
Cobalt-60 X-rays
0.2
250-kV X-rays
2.0
10-MeV protons
4.7
150-MeV protons
0.5
14-MeV neutrons 2.5-MeV ce particles 2-GeV Fe ions
-1 "
Energy avg.
100
12 166 1000
O
c~ c
10 -2 -
.m > CO
10-a -
10-4
I
0
I
2
I
I
4
I
I
I
6
I
8
I
I
I
10
I
12
D o s e (Gy)
FIGURE 16. Survivalcurves for cells of human origin exposed to 250-kV X-rays, 15-MeV neutrons, and 4-MeV a particles. As the LET of the radiation increases, the slope of the survival curves gets steeper and the size of the initial shoulder gets smaller. (Redrawn from Broerse et al., 1968.)
of c~ particles produces more cell killing than 1 Gy of X-rays. The key to the difference lies in the pattern of energy deposition at the microscopic level. In comparing different radiations, it is customary to use X-rays as the standard. The formal definition of relative biological effectiveness (RBE) is as follows: The RBE of some test radiation (r) compared with Xrays is defined by the ratio D x/D r where D x and D r are, respectively, the doses of X-rays and the test radiation required for equal biological effect. To determine the RBE of some test radiation, say, a particles, one chooses a biological system in which the effect of radiation may be scored quantitatively, and one compares a particles with the standard radiation, that is, X-rays. The data in Fig. 16 are an example. To achieve a surviving fraction of 10-2 requires a dose of about 9.5 Gy or X-rays or 2.5 Gy of a particles. The RBE of ce particles relative to X-rays is the ratio of doses required to produce the same biological effect, that is, 9.5/2.5 = 4.6. Neutrons have a somewhat lower RBE; the dose to result in a surviving fraction of 10 -2 in 4.3 Gy, so that the RBE is about 9.5/4.3, or 2.2. It is important to note that, because the survival curves for X-rays (low LET) and neutrons or a particles (high LET) have a different shape, with the survival curve for X-rays having an appreciable shoulder, RBE does not have a unique value. At lower doses, corresponding to higher levels of survival, the RBE will assume a larger value; the maximum RBE will correspond to the ratio of the initial slopes for Xrays and for the test radiation.
D. Relative Biological Effectiveness for Different Cells and Tissues Even for a given total dose, the RBE varies significantly according to the tissue or endpoint used to measure it. Survival
1198
SECTION IX Cell Division and Programmed Cell Death
curves for mammalian cells exposed to X-rays have a large but variable shoulder region, while for high LET radiations the shoulder region is reduced or eliminated. As a consequence, the RBE will be different for each cell line. In general, cells characterized by an X-ray survival curve with a large shoulder, indicating that they can accumulate and repair a large amount of sublethal radiation damage, will show a large RBE for neutrons or a particles. Conversely, cells for which the X-ray survival curve has little if any shoulder will exhibit small RBE values for c~ particles or neutrons.
E. Relative Biological Effectiveness as a Function of Linear Energy Transfer Figure 16 shows data for just three types of radiation to illustrate the way in which the shape of the survival curve changes as the density of ionization (LET) increases. Studies have been performed with many types of radiation covering a wide range of LET values. Figure 17 is a plot of the RBE as a function of LET. As the LET increases, the RBE increases slowly at first, then more rapidly as the LET increases beyond 10 keV/pm. Between 10 and 100 keV/l~m, the RBE increases rapidly with increasing LET and in fact reaches a maximum of about 100 keV/pm. Beyond this value for the LET, the RBE falls to lower values. The LET at which the RBE reaches a peak is much the same (about 100 keV/pm) for a wide range of mammalian cells, from mouse to human, and is the same for all biological endpoints. It reflects the "target" size and is related to the DNA content, which is similar for all mammalian cells. The
~200
keV/pm
X-
_
~ 0 t0 .m
> 0 0
e::
7 6 5
0
~o
4-
0
~
3-
o>
3"
.n
_
nO
....
0.1
i
i
i
i i i ii I
1
10 LET (keV/IJm)
1O0
1000
FIGURE 1 7. Variation of RBE with LET. RBE rises to a maximum at an LET of about 100 keV/pm and subsequently falls for higher LET values. The inset diagram illustrates that radiation with an LET of 100 keV/lam is most biologically effective because the average separation between ionizing events coincides with the diameter of the DNA double helix (2 nm); consequently a single-track of this quality can produce double-strand breaks with greatest efficiency. (Graph redrawn from Barendsen, 1968.)
reason why radiation with a LET of 100 keV/pm is optimal in terms of producing a biological effect can be understood from the upper panel of Fig. 17. At this density of ionization, the average separation between ionizing events just about coincides with the diameter of the DNA double helix (20/k or 2 nm). Radiation with this density of ionization is most likely to cause a double-strand break by the passage of a single charged particle, and DSBs are the basis of most biological effects, as discussed earlier. This is illustrated in the top panel of Fig. 17. In the case of X-rays, which are more sparsely ionizing, the probability of a single track causing a doublestrand break is low, and in general more than one track will be required to produce a double-strand break. As a consequence, X-rays have a low biological effectiveness. At the other extreme, much more densely ionizing radiations (with a LET of 200 keV/lam, for example) will readily produce DSBs but energy will be "wasted" because the ionizing events are too close together. Since RBE is the ratio of doses to produce equal biological effect, this more densely ionizing radiation will have a lower RBE than the optimal LET radiation. The more densely ionizing radiation will be just as effective per track, but less effective per unit dose. It is possible, therefore, to understand why RBE reaches a maximum value in terms of the production of DSBs, since the interaction of two DSBs to form an exchange type aberration is the basis of most biological effects. Radiation having this optimal LET includes neutrons of a few hundred kiloelectron volts, as well as low-energy protons and a particles.
F. Radiation Weighting Factors It is evident from the preceding discussion that different types of radiation differ in their biological effectiveness per unit of absorbed dose. The complexities of RBE are too difficult to apply in specifying dose limits in radiation protection; it is necessary to have a simpler way to consider differences in biological effectiveness of different radiations. One cGy of neutrons, for example, is more hazardous than 1 cGy of X-rays. The term radiation weighting factor (W r) has been introduced for this purpose. The quantity produced by multiplying the absorbed dose by the weighting factor is called the equivalent dose. When dose is expressed in Gray (Gy), the equivalent dose is in Sievert (Sv). Radiation weighting factors are chosen by the International Commission on Radiological Protection (1991), based on a consideration of experimental RBE values, biased for biological endpoints relevant to radiation protection at low dose and low dose rate. There is a considerable element of judgment involved. The W r is set at unity for all low LET radiations (X-rays, gamma rays, and electrons), with a value of 20 for maximally effective neutrons and a particles. Thus, using this system, an absorbed dose of 0.1 Gy of radiation with a radiation weighting factor of 20 would result in an equivalent dose of 2 Sv.
XIII. R a d i o p r o t e c t o r s In the late 1940s it was discovered that either cysteine or cysteamine could protect mice from the lethal effects of total
69. Effects of Ionizing Radiation on Cells
1 199
body X-irradiation if administered in large amounts prior to exposure. The structures of these compounds are: qH2 SH-CH2-~H
cysteine
(6) Indirect action
COOH SH-CH2-CH2-NH 2
cysteamine (7)
Animals receiving these radioprotectors could tolerate almost double the radiation dose. The ratio of the radiation dose to produce the same biological effect in the presence and absence of the drug is known as the dose reduction factor (DRF). Many similar compounds have been tested and found to be effective as radioprotectors. The most efficient tend to have certain structural features in common: a free SH group (or potential SH group) at one end of the molecule and a strong basic function such as amine or guanidine at the other end, separated by a straight chain of two or three carbon atoms. Sulfhydryl compounds are efficient radioprotectors against sparsely ionizing radiations, such as X- or gamma rays. The mechanism of action involves the scavenging of free radicals. X-ray photons give up their energy to produce fast electrons as described earlier. These electrons may hit the DNA and cause a break directly (direct action) or interact with a water molecule to form a hydroxyl radical (OH.), which diffuses to the DNA and causes a break (indirect action). The protective effect of sulfhydryl compounds stems from their ability to "scavenge" free radicals. This process is illustrated in Fig. 18. The protective effect of sulfhydryl compounds tends to parallel the oxygen effect, being maximal for sparsely ionizing radiations (e.g., X- or gamma rays) and minimal for densely ionizing radiations (e.g., low-energy ce particles). It might be predicted that with effective scavenging of all free radicals the largest possible value of DRF would equal the oxygen enhancement ratio, with a value of 2.5 to 3.0. It is not surprising that the discovery in 1948 of a compound that offered protection against radiation excited the interest of the U.S. Army, since the memory of Nagasaki and Hiroshima was vivid in the years immediately after World War II. Although cysteine is a radioprotector, it is also toxic and induces nausea and vomiting at the dose levels required for radioprotection. Consequently, the Walter Reed Army Hospital in Washington, D.C., synthesized more than 3000 compounds in an attempt to fmd the perfect radioprotector, one that would protect against radiation without debilitating side effects. At an early stage, the important discovery was made that the toxicity of the compound could be greatly reduced if the sulfhydryl group was covered by a phosphate group. This is illustrated in Table 2. The LDs0 of the compound in animals can be doubled and the protective effect in terms of the DRF greatly enhanced if the SH group in cysteamine is covered by a phosphate. This tends to reduce systemic toxicity. Once in the cell, the phosphate group is stripped, and the SH group begins scavenging free radicals. Many compounds have been synthesized, but only two have been put to practical use. The structure and effectiveness of these is summarized in Table 3. The first compound, WR-638, called Cystaphos, was said to be carried routinely in the field pack of Russian infantry in Europe during
-OH ~
,S-A=T-S~
?
P\
S- C--------- G-S, ~P's~,T ' P" A-S P P~S_G=C_S /
H20
Radical scavengers
e-~. H2 0 Direct action
FIGURE 18. Photonsof X-rays either interact directly with DNA to produce a strand break (direct action) or indirectly via the production of a free radical (indirect action). Radioprotectors that are radical scavengers "intercept" the direct action. (Redrawn from Hall, 1994.) the era of the Cold War for use in the event of a nuclear conflict. Its usefulness would have been largely psychological, since the compound was carried as a tablet to be administered orally, when in fact these sulfhydryl compounds break down in stomach acid and are only effective when administered intravenously or intraperitoneally. A further factor, of course, is that such compounds will protect only from sparsely ionizing radiation; consequently, they would offer little protection against the prompt release of neutrons produced by the detonation of a nuclear device. They would be effective only against the gamma rays from the resulting fallout. The second compound, WR-2721, now known as amifostine, is perhaps the most effective of those synthesized in the
Effect of Adding a Phosphate-Covering Function on the Free Sulfhydryl of/3-Mercaptoethylamine
TABLE 2
Drug
Formula
LDs0 in mice
DRF
MEA
NH2-CH-CH2-SH
343 1.6 at 200 (323-364) mg/kg
MEA-PO3
NH2-CH2-CH-SH2PO3
777 2.1 at 500 (700-864) mg/kg
1200
SECTION IX Cell Division and Programmed Cell Death
TABLE 3
Two Protectors in Practical Use
Compound
Drug dose (mg/kg)
DRF 7 days
DRF 30 days
Structure
WR-638
NH2CH2CH2SPO3HNa
500
1.6
2.1
WR-2721
NH2(CH2)3NHCH2CH2SPO3H2
900
1.8
2.7
Walter Reed series (Kligerman et al., 1992). It gives good protection to the blood-forming organs, as can be seen by the DRF for 30-day death in mice, which approaches the theoretical maximum value of 3. It was probably the compound carfled by the U.S. astronauts on their trips to the moon to be used if a solar event occurred. On these missions, when the space vehicle left earth's orbit and began coasting toward the moon, the astronauts were committed to a 14-day mission, since they did not have sufficient fuel to turn around without first orbiting the moon and using its gravitational field. If there had been a major solar event in that period, the astronauts would have been exposed to a shower of high-energy protons, resulting in an estimated dose of several hundred Gray. The availability of a radioprotector with a DRF of between 2 and 3 would have been very important in such a circumstance. As it turned out, no major solar event ever occurred during any manned lunar mission. Amifostine also has a potential in radiotherapy. One application would be for local topical use, to reduce for example the mucosal reaction that occurs during radiotherapy of head and neck cancers. Another possibility is to protect all normal tissues by delivering the drug systemically; there is some evidence that the drug gets into tumors more slowly than into normal tissues, so that if the radiation is delivered within minutes of the drug, a differential effect is obtained with more protection of the normal tissues than of the tumor.
XIV. Summary Life on earth has always been exposed to radiation, but there is now an additional exposure from medical X-rays, nuclear power, and journeys in high-flying aircraft. Of concern are ionizing radiations, so called because they have a sufficient photon energy to knock electrons out of the atoms through which they pass, break chemical bonds, and cause a variety of biological changes. Ionizing radiations come in a variety of types. X-rays are usually generated in an electrical device. Gamma rays are emitted by a radioactive material during decay. X- or gamma rays are electromagnetic waves, similar to radiant heat and visible light, but having a shorter wavelength. Neutrons are uncharged particles, often associated with nuclear fission. Highenergy high-z charged particles are a hazard to astronauts in long-duration space flights. The principal target for all biological effects of ionizing radiations is the DNA. The basic lesion is a double-strand break, caused when a charged particle breaks both strands of
the DNA double helix, causing the chromatin strand to snap. Chromosomal aberrations represent one of the earliest detectable biological changes; if DSBs occur in two separate chromosomes, the broken ends may rejoin in incorrect and bizarre ways. These rearrangements can often be seen at the first metaphase after irradiation. Following irradiation, cells may die from one of two mechanisms: 1. Mitotic cell death, due to the failure of cells to cope with severely damaged chromosomes 2. Apoptosis, or programmed cell death, similar to the way that cells die and are removed during embryogenesis The quantity of radiation, or dose, is measured in terms of the energy absorbed per unit mass. The current mass is the Gray (Gy), defined to be an energy absorption of one joule per kilogram. A number of factors influence the biological effects of a given absorbed dose of radiation. First, the response of cells to radiation depends critically on their position in the division cycle at the time that they are exposed. In general, cells are most radiosensitive when they are close to mitosis. Second, cells vary greatly in their ability to repair radiation damage. In addition, a given dose of radiation is much less effective when spread out over a long period of time, because much of the radiation damage can be repaired during a prolonged exposure. Third, in the case of X- and gamma rays (but less so for neutrons and a particles), the biological consequences of a given dose can be modified by chemical means. For example, cells are more radiosensitive if molecular oxygen is present than if it is absent. On the other hand, compounds that scavenge free radicals are often found to be radioprotectors, and such compounds were used by the astronauts during lunar missions. Fourth, the biological effect of a given absorbed dose depends on the "quality" of the radiation. X- and gamma rays are said to be "sparsely ionizing" because the ionizing events are well separated along the tracks of the electrons set in motion. Consequently, they are less effective than neutrons and a particles that are said to be "densely ionizing" because the ionizing events are clustered together and more likely to produce a DSB in a chromosome.
Bibliography Barendsen, G. W. (1968). Responses of cultured cells, tumors and normal tissues to radiations of different linear energy transfer. Curr. Top. Radiat. Res. 4, 293. Bedford, J. S., and Mitchell, J. B. (1973). Dose-rate effects in synchronous mammalian cells in culture. Radiat. Res. 54, 316-327. Broerse, J. J., Barendsen, G. W., and Van Kersen, G. R. (1968). Survival of cultured human cells after irradiation with fast neutrons of different energies in hypoxic and oxygenated conditions. Int. J. Radiat. Biol. 13, 559-572. Elkind, M. M., Sutton-Gilbert, H., Moses, W. B., Alescio, T., and Swain, R. B. (1965). Radiation response of mammalian cells in culture: V. Temperature dependence of the repair of x-ray damage in surviving cells (aerobic and hypoxic). Radiat. Res. 25, 359-376. Hall, E. J. (1994). "Radiobiology for the radiologist." (4th ed.) J. B. Lippincott Company, Philadelphia.
69. Effects of Ionizing Radiation on Cells
Howard, A., and Pelc, S. R. (1953). Synthesis of deoxyribonucleic acid in normal and irradiated cells and its relation to chromosome breakage. Heredity 6, Suppl., 261. International Commission on Radiological Protection. (1991) "Recommendations," Report No. 60. Pergamon Press, New York. International Commission on Radiological Units and Measurements. (1962a). "Radiation Quantities and Units," Report 10a, Handbook 84. National Bureau of Standards, Washington, DC. International Commission on Radiological Units and Measurements. (1962b). "Physical Aspects of Irradiation," Report 10b, Handbook 85. National Bureau of Standards, Washington, DC. Kligerman, M. M., Liu, T., Liu, Y., Scheffier, B., He, S., and Zhang, S. (1992). Interim analysis of a randomized trial of radiation therapy of
1201
rectal cancer with/without WR-2721. Int. J. Radiat. Oncol. Biol. Phys. 22, 799-802. Mottram, J. C. (1936). Factor of importance in radiosensitivity of tumors. Br. J. Radiol. 9, 606-614. Schwarz, W. (1910). Wein. Klin. Wochenschr. Nr. l l S , 397. Sinclair, W. K., and Morton, R. A. (1966). X-ray sensitivity during the cell generation cycle of cultured Chinese hamster cells. Radiat. Res. 29, 450-474. Terasima, R., and Tolmach, L. J. (1963). X-ray sensitivity and DNA synthesis in synchronous populations of HeLa cells. Science 140, 490-492. Zirkle, R. E. (1940). The radiobiological importance of the energy distribution along ionization tracks. J. Cell. Comp. Physiol. 16, 221.
Nicholas Sperelakis
Review of Electricity and Cable Properties
I. Introduction
~'~o
An appreciation of some elementary principles of electricity is essential to the understanding of many aspects of cell physiology and biophysics and the electrophysiology of excitable cells. The purpose of this section is to give a brief review of the most relevant aspects of electricity and electric circuits that pertain to the cell membrane.
+
E (9 i n
@
Ii. Definition of Circuit Elements and Ohm's Law The simple electric circuit shown schematically in Fig. 1 consists of a battery, a resistor, an ammeter, a switch, and copper wire. Closing the switch allows electrical current to flow from the positive side of the battery, through the resistor and ammeter, and back to the negative side of the battery. Electric current is considered, by convention, to flow from positive to negative through the external circuit, even though the dectrons--the negatively charged particlesmactually flow in the opposite direction, that is, from negative to positive. One can consider that holes, that is, the absence of an electron in a space that previously contained one, flow in the direction opposite that of the electrons, that is, from positive to negative around the external circuit. Within the battery, the flows are just the opposite; namely, electrons flow from positive to negative, and current flows from negative to positive. In the circuit of Fig. 1, the relationship between the applied voltage, the total resistance, and the current that flows is given by O h m ' s law. This relationship, first formulated by George S. Ohm, states that the current (I) that flows is equal to the applied voltage (V) divided by the resistance (R): I -V R volts amperes= ohms
Cell Physiology Sourcebook: A Molecular Approach, Third Edition
(la)
I
Atom F I G U R E 1. A simple electric circuit containing a battery (E) and a resistor with resistance R. When the switch is closed, electrical current (I) flows in the direction indicated by the arrows, from the positive pole of the battery to the negative pole in the external circuit, and can be measured by the ammeter (Amm) placed in series with the resistor. Current through the battery flows from the negative pole to the positive pole. The relationship between, E, R, and I is given by Ohm's law: I = E/R.
An a m p e r e (A) of current is the movement past a point of one coulomb (C) of electrical charge per second (1 A = 1 C/s). The unit of resistance is the ohm (f~), which may be operationally defined as that value of resistance through which there is a current of 1 A when the applied potential is 1 V (R = V/1). Resistance is a measure of the opposition to current flow; namely, for a given applied voltage, the greater the resistance the less the flow of current (inverse proportionality). The volt can be operationally defined from Ohm's law as that potential difference (PD) necessary to produce 1 A of current flow through a resistance of 1 ~2:
V= IR
(lb)
volts = (amperes)(ohms)
| 2 0 5
Copyright 9 2001 by Academic Press. All rights of reproduction in any form reserved.
1206
Appendix
Two points are at a PD of 1 V when it takes 1 joule (J) of energy to transport 1 C of charge from one point to the other, given as
1 V - 1 J/C
R1
R2
(2a)
RT
Note that energy equals voltage times the charge:
R1
(2b)
j o u l e s - volts • coulombs
o
(Thus the electron-volt unit of atomic physics is a unit of energy, the charge on an electron being 1.60 • 10-19 C.) A coulomb of charge is equal to that represented by 6.24 • 1018 electrons or univalent ions (reciprocal of 1.60 • 10-19 C/electron).
(3)
1 C - 6.24 x 1018 univalent ions
'
o
i I| G1
G2
(160
0'9 univa,ention
e.
~- 96 500 C eq This derived value is known as the Faraday constant (@), used in electrochemistry in equations such as the Nernst equation. Multiplying the Faraday constant by the valence of the ion (eq/mol) then gives the number of coulombs per mole:
G3
G1
G2 G3
(4)
qe • N A
I
C
Multiplication of the charge on an electron (qe) by Avogadro's n u m b e r (NA), which is the number of molecules per mole, gives the number of coulombs in an equivalent (eq) of charge or mole of univalent ions -
R3
FIGURE 2. (A) Resistors placed in series-the total resistance (RT.) is the sum of the resistances: Rr = R 1+ R2 + R 3. (B) Resistors placed in parallel--the reciprocal of the total resistance is equal to the sum of the reciprocals of the individual resistances" 1/R 7, = 1/R 1 + 1/R 2 + 1/R 3. (C) Reciprocal of the total conductance (1/G~) of conductances in series is equal to the sum of the reciprocals of the individual conductances: 1/G~ = 1/G 1 + 1/G 2 + I[G 3. (D) Total conductance of conductances in parallel is equal to the sum of the conductances: G T = G 1 + G 2 + G 3.
R T - R 1+ R 2 + R 3 + ... z@
-
eq mol
x
C eq
- -
-
~
C mol
( 5 )
Conductance (G) is the reciprocal of the resistance: G:
1
The greater the resistance, the lower the conductance, and vice versa. Therefore, Ohm's law can also be written as I-
V _ 1 V-
R
GV
If all the resistors are of the same value (R 1 = R 2 = R3), then the total resistance is R~
(6)
-R
(7)
R
This states that the current is a product of conductance and voltage. This form is useful for electrophysiology, since conductances in parallel simply add. The unit of conductance is the m h o (ohm spelled backwards) or E2-1 or siemens (S).
!II. R e s i s t o r s a n d C o n d u c t a n c e s in S e r i e s a n d in P a r a l l e l A. Resistors in Series For resistors in series (Fig. 2A.), the total resistance (Rr) is a simple sum of the values of the resistances in series"
(8)
:
R1N
(9)
where N is the number of resistors in series. Because myelin consists of many wrappings of membrane around the axon membrane (membranes in series), it should be obvious why myelination effectively raises membrane resistance.
B. Resistors in Parallel For resistors in parallel (Fig. 2B), the reciprocal of the total resistance (1/R~) is equal to the sum of the reciprocals of the resistances: __1 : __I + __1 + __1 + ... R~r R 1 R 2 R 3
(10)
If all the resistors are the same value ( R 1 - R 2 = R3), then the total resistance is equal to any one resistance divided by the total number of resistors in parallel: RT
R1 = -N-
(lla)
If there are only two resistances in parallel, then Eq. 10 can be reduced algebraically to
1207
70. Review of Electricity and Cable Properties
1 Rr
1 R1
1 R2
R 1+ R 2
(llb)
A
il
R1R 2 +
Taking the reciprocal gives
_
R2
(llc)
(R1 x R2)
+t;-
R r = (R 1+ R2) Equation 11 c states that the total resistance of two resistors in parallel is equal to the product of the resistances divided by the sum (often called the rule of product over the sum). Note that R r is always less than the lowest R for resistors in parallel.
O
R1
+
l:/2
C. Conductances in Series Conductances in series (Fig. 2C) add like resistances in parallel; namely, the reciprocal of the total conductance (1/Gr) is equal to the sum of the reciprocals of each conductance:
1 :__I +__I +__I Gr
G1
G2
(12)
G3
FIGURE 3.
Diagrammatic representation of Kirchhoff's two laws: (1) Current approaching a point is equal to that leaving the point. Thus in the simple series circuit (A) the current is constant throughout, and in the simple parallel circuit (B), the sum of the currents in each leg is equal to the total current (It): Ir = il + i2 + i3" (2) The sum of the voltage drops (IR) in the external circuit must equal the source emf. Thus in the simple series circuit (C), the sum of the IR drops is equal to the source emf (battery voltage): source emf = IR 1 + IR 2 + IR 3.
D. Conductances In Parallel Conductances in parallel (Fig. 2D) add like resistances in series; namely, the total conductance (Gr) is simply equal to the sum of each conductance: Gr + G 1+ G 2 + G 3 + ...
(13a)
That this is true is demonstrated by substituting G = 1/R into Eq. 13a and obtaining
1 =l+!+!+... Rr R1 R2 R3
(13b)
which is identical to Eq. 10 for resistances in parallel. It is convenient to use conductances in the study of the electrical properties of membranes, since the conductance pathways for the various species of ions are in parallel to each other, and therefore can simply be added to give the total conductance.
IV. Kirchhoff's Laws Kirchhoff's two laws may be stated in the following manner: (1) The current going toward a point is equal to that leaving it. For example, in a simple circuit composed of three resistors in series and a battery in series (Fig. 3A), the current is constant throughout the circuit; that is, an ammeter placed anywhere in the circuit gives the same reading. In a simple parallel circuit composed of a network of three resistors in parallel connected to a series battery (Fig. 3B), the current is not constant throughout the circuit. The sum of the currents in each branch is equal to the current in the main branch; in other words, the sum of the currents converging toward a branch point is equal to that leaving it: I r - i 1+ i2 + i3
(14)
(2) The algebraic sum of the voltage drops (IR) across each of the resistors in any closed network is equal to the source electromotive force (emf), e.g., battery. For the series circuit
shown in Fig. 3C, source emf = I R 1 + I R 2 + I R 3
(15)
The I R or voltage drop across each resistor is always in the polarity depicted; namely, the proximal side with respect to the direction of the current flow is always positive, and the distal side is negative. Current is always considered to flow from positive to negative in the external circuit. By convention, a positive potential is higher (or greater) than a negative one; hence, the potential is progressively dropped or lowered around the external circuit.
V. N a t u r e of C a p a c i t o r s A capacitor is a device that can store electric charge or electricity. It consists of two parallel plates (conductors) separated by a dielectric. All effective dielectrics have a high electrical resistance, but the reverse is not necessarily true; namely, all high resistors are not necessarily very good dielectrics (e.g., air). The dielectric constant (e) of a material is a relative index of its ability to function as an efficient dielectric compared to a vacuum, which has an e value of 1.000. The dielectric constants of some selected materials are given in Table 1. The greater the dipolar nature of the molecules, the greater the dielectric constant. For example, water is an efficient dielectric because of its highly dipolar nature. The dipoles must be free to flip-flop and orient with changes in the electric field, much like the water dipole orients around small charged ions (e.g., Na § K § to form the hydration layers. Polar phospholipids in the cell membrane should have a higher dielectric constant than less polar oils. It is seen from Coulomb's law, for the force (F, in dynes) of interaction between point charges, that a large dielectric constant reduces the force of interaction between electrical
Appendix
1208 TABLE 1
the specific capacitance (C/A), that is, the capacitance normalized to unit area (e.g., 1 cm2) 9
Dielectric Constants of S o m e Materials
Material
Dielectric constant (E)
Vacuum Air Oils Water (distilled) Cell membrane Solid insulators Paraffin wax Quartz Mica Glass Thin layers of oxidized metal
C A
-- =
1.00000 1.0006 3-5 80 3-6 2-10 2-3 4-5 6-7 5-7 >30
charges, that is, the dielectric shields or buffers (see Section IX): (16)
where q l and q2 are the magnitudes of the charges (in statcoulomb), d is the distance between the charges, and 9 is the dielectric constant. The distance squared term in Coulomb's law reflects the geometrical factor: the surface area of a sphere equals 4~a 2 or ~'d2, in which a is the radius and d is the diameter. Therefore, the surface area over which the total force is distributed increases with the square of the radius (or diameter) of the sphere. From Eq. 16, a statcoulomb may be defined as that charge which, when placed 1 cm from an equal and like charge in a vacuum, repels the charge with a force of 1 dyne (dyn). A dyne is the force necessary to give a mass of 1 g an acceleration of 1 cm/s 2. Because the gravitational constant is 980 cm/s 2, 1 dyn is equal to approximately 1 mg of force (1 g/980). The magnitude of a capacitance (C) is equal to 9A 1 C - --~ 4~'----k
(17b)
The shorter the distance between the plates, the greater the capacitance (Eq. 17a). Therefore, the cell membrane has a very high capacitance, because the distance between the plates is only about 3.5-7.0 nm or 35-70 ~ (the nonpolar phospholipid tail length of the membrane bilayer). For example, if the measured specific capacitance of the cell membrane (Cm) is 1.0 pF/cm 2, and assuming a dielectric constant of 5, then from Eq. 17b d-
F - qlq2 9d 2
e 1 d 47rk
9
1
(C/A) 4"rrk
(17c)
5 1 1 x 10-6F/cm 2 (12.56)(9.0• 1011cm/F) = 44.2 • 10-8 cm =44 = 4.4 nm All capacitors have a breakdown voltage; that is, at more than a critical voltage level, the dielectric material cannot withstand the PD impressed across it. Therefore the capacitor breaks down, and electric current can now flow through the dielectric material from one plate to the other. The capacitor can no longer function as a capacitance, because the two plates are almost shorted to one another. A typical breakdown voltage of a commercial radio capacitor may be 200 V. In contrast, the capacitance of the nerve or muscle membrane has a breakdown voltage of only about 200 mV. Because the two plates of the membrane capacitor are much closer (about 50 ~ or 0.5 • 10-6 cm), however, the voltage gradient necessary for breakdown is 200 mV/0.5 • 10-6 cm or 400 000 V/cm. At a resting potential o f - 8 0 mV, the voltage gradient is about 160 000 V/cm. Thus, the cell membrane is indeed a very effective capacitor and dielectric.
(17a)
in which A is the area of each plate (in cm2), d is the distance (in cm) between the plates (i.e., thickness of the dielectric), and k is a constant (9.0 • l0 ll cm/F). The unit of capacitance is the farad (F), named after Michael Faraday. Thus, the greater the dielectric constant, the greater the capacitance (Eq. 17a). The greater the area, the greater the capacitance (Eq. 17a). Therefore, the total capacitance of several capacitors in parallel should be the simple sum (see Section VI.A) because the area of the plates is effectively increased. For example, the capacitance of the tuning capacitor (dial) of a radio is varied by altering the degree of overlap between two sets of interdigitating plates: the more the overlap, the greater the capacitance, because the area (A) is greater; air is the dielectric in this case. (Commercial capacitors may use substances such as air, oil, paper, or mica for the dielectric; the area can be made very large by many wrappings of double sheets of aluminum foil separated by paper or mica). The area term is often incorporated with the capacitance to give
VI. C a p a c i t o r s
in P a r a l l e l a n d S e r i e s
A. C a p a c i t o r s In Parallel
For capacitances in parallel (Fig. 4A), the total capacitance (Cr) is a simple sum of the capacitances C T = C 1 + C 2 + C 3 + ...
(18)
The reason for this should be clear from Eq. 17a, which demonstrates that capacitance is directly proportional to the area of the plates; hence placing capacitors in parallel effectively increases the area of the plates, as can be deduced from Fig. 4A. If the number (N) of capacitors in parallel all have the same value (C 1 = C 2 = C3), then the total capacitance is
CT-C1N
(19)
Thus, capacitors in parallel add like resistors in series. Invaginations of the skeletal muscle fiber or ventricular myocardial cell surface membranes to form transverse (T)
1209
70. Review of Electricity and Cable Properties
C1
higher the frequency, the lower the capacitive reactance (X c, in f~). The relationship is given by
C2
X c = 2"rrfC
1
It
1 o~C
C3
in which C is capacitance and the angular velocity (o~) in radians/s is equal to 27rf. For DC, f = 0 and X c = oo. At f = oo, X c = 0. For a given f, the larger C is, the lower X c is. For a nerve or muscle membrane, if C is 1.0, laF/cm2 and f is 1000 cycles/s, then
tt B Cl
(22a)
Ca
C3
1 FIGURE 4. (A) Capacitors placed in parallel. The total capacitance is the sum of the individual capacitances: Cr - C1 + C2 + C3. (B) Capacitors placed in series. The reciprocal of the total capacitance is equal to the sum of the reciprocals of the individual capacitances: I/C T = I/C 1 + 1/C 2 + 1/C 3.
(22b)
=0.159 x 103 s'cm2 F = 159 O-cm 2
tubules increase the total measured fiber capacitance, because the increased membrane area effectively adds capacitors in parallel.
B. Capacitors In Series For capacitances in series (Fig. 4B), the reciprocal of the total capacitance (1/C7.) equals the sum of the reciprocals of each capacitance:
1 =l+!+ Cr
X c = (6.28)(1000/s)(1 x 10 -6 F/cm 2)
C1
1 C2
+...
(20)
C3
If there are several capacitors in series that are all of equal value (C 1 = C 2 = C 3 = ...), then the total capacitance is equal to any capacitance divided by N: C1
(21a)
If there are only two capacitors in series, regardless of values, the total capacitance equals the product over the sum: C1C2 CT-
(21b)
C 1 -'k C 2
Note that Cv is always less than the lowest C. Because myelin consists of many (20-200) wrappings of membrane around the axon membrane, it should be obvious why myelination reduces the total capacitance of a nerve fiber. Capacitors in series effectively increase the distance between the plates (d), and thereby lower the total capacitance (see Eq. 17a).
VIII. M e m b r a n e I m p e d a n c e A. Parallel RC Circuit Because the cell membrane has a capacitive component (because of the continuous phospholipid bilayer matrix) in parallel with a resistive component (because of protein ion channels embedded in the lipid bilayer matrix), membrane impedance (Z) should diminish as the AC frequency is increased (Fig. 5A). The relationship between frequency and relative impedance for a typical cell membrane is given in Fig. 5B. When f is extremely low, X c is extremely high, and Z approaches R. As f is increased, X c diminishes, and more and more of the current passes through C. Current does not actually pass through the dielectric of C, but it appears so because of the phenomenon of induced charge that occurs (see Section IX). When f is extremely high, X c approaches zero, and therefore Z approaches zero. This is analogous to two resistors in parallel, one of whose impedance decreases as a function of the applied frequency. In diathermy, a very high frequency AC is used to heat damaged tissues without producing excitation of the nerve and muscle membranes; this is possible because of the capacitive component of the membrane, which causes Zm to be extremely low and which does not allow a significant membrane potential change to be produced. At the frequency at which X c = R, however, Z is not equal to 0.5 • R, but is actually 0.707 X R, as determined by the following equation for a resistor and capacitor in parallel (note its similarity to that for two resistors in parallel, except for the squared function):
VII. Capacitive R e a c t a n c e Capacitors have almost infinite resistance to direct current (DC). Their impedance (Z, in f~) or so-called "AC resistance" to sinusoidal alternating current (AC) is less, however, and is a function of the frequency (f) of the AC. The
-
Z-
+
l
(23a)
2R 2
Xc 2
Xc+
R2
(23b)
1210
Appendix
B
A
I.O
"
"" "
R
Xc 0.71
z R
RAPIDLY ISHING
(REL)
t
ao
C
'1'
Ioo Iooo ,o 4 ,b s
D
I
I
:1
|
I,S. ",,|
j
RI
- V
"~,AAN%----
I
R
FIGURE 5.
(A) Parallel resistance-capacitance (RC) network model of the cell membrane. (B) Impedance (Z) of such a parallel resistance-capacitance network decreases as the sinusoidal frequency (f) is increased: a plot of Z versus log frequency is an inverse sigmoidal shape. Impedance decreases because the capacitive reactance (X c) decreases according to X c = 1/27rfC. (C) Parallelogram method of obtaining the impedance of a parallel resistance-capacitance circuit. Parallelogram constructed of 1/X c against 1/R gives the vector that is equal to 1/Z; the phase angle (0) between current and voltage is also obtained. (r approaches 90 ~ as resistance (R) approaches infinity; i.e., the circuit becomes purely capacitive.) Also, as can be reasoned, the larger the capacitance (for a given resistance), the lower the X c for a given f and hence the larger the 0. Thus the circuit behaves as the most purely capacitive (0 ~ 90~ the larger the capacitance and the larger the resistance. (D) Circuit diagram illustrates why current leads the voltage (by a certain phase angle) in a resistance-capacitance circuit. The instant the switch (Sw) is closed the frequency appears to be infinite, and so X c is zero; therefore no voltage appears across the capacitance, although the (induced) current through the capacitance is maximum; that is, current leads the measured voltage.
The impedance (Z) and the phase angle (0) between current and voltage can also be measured graphically by the parallelogram method, as illustrated in Fig. 5C. In this method, the reciprocal of the capacitive reactance is plotted on the ordinate against the reciprocal of the resistance on the abscissa, and the parallelogram is completed by drawing lines parallel to the two axes. The diagonal line is equal to the reciprocal of the impedance, and the angle between the diagonal and the abscissa gives the phase angle (0) between voltage and current. In a purely resistive circuit, the current and voltage are exactly in phase (0 = 0~ In a purely inductive circuit, current lags the voltage by 90 ~ whereas in a purely capacitive circuit, current leads the voltage by 90 ~ In a resistance-capacitance circuit, parallel or series, 0 is between 0 ~ and 90 ~ depending on the relative values of R and X c. The reason that current leads the voltage in a resistancecapacitance circuit may be discerned from Fig. 5D. At the instant the switch is closed, the frequency appears to be infinite (step change), and so X c = 0. Therefore, there can be no potential difference across the capacitor, and the entire voltage is dropped across R 2. Thus, the current through the capacitor is maximum, while the voltage across the capacitor is minimum, that is, the current leads the voltage by 90 ~. The identity of the parallelogram method with Eq. 23a is given by the Pythagorean theorem of trigonome.try,
C2 -- a 2 + b 2
(24)
where c is the length of the hypotenuse of a right triangle, and a and b are the lengths of the other two sides. By substituting for a, b, and c from Fig. 5C, the following expression, which is identical to Eq. 23a is obtained: (23a)
B. Series RC Circuit For a series resistance-capacitance network (Fig. 6A), the corresponding analysis is Z 2 - X c2 + R 2
(25a)
or
Z-/X
2 + R2
(25b)
which is similar to having two resistors in series, except for the squared function. When X c = R , then Z = 1.414R (not 2 X R). At f = oo, X c = 0, and Z = R. At f = 0 (DC), X c = oo, and Z = oo. The parallelogram analysis is shown in Fig. 6B.
70. Review of Electricity and Cable Properties
A
c
PD across the membrane during the rising phase of the AP is relatively large (e.g., 100 mV/0.5 ms or 200 V/s); therefore capacitive current is also relatively large. There is little or no change in the value of the membrane capacitance during an AP. As stated in Section VIII, current does not actually flow through the capacitor, but only appears to do so because of the phenomenon of induced charge. When electrons flow into one plate of the capacitor, the negative charges on this plate repel the electrons from the other adjacent plate, since like charges repel one another. Thus, as diagrammed in Fig. 7, a flow of electrons into the lower plate of the capacitor leads to a flow of electrons out of the upper plate. Hence it appears as though the flow of electrons was directly through the dielectric. The deficiency of electrons in the upper plate makes the upper plate become positive, whereas the lower plate is negatively charged.
R
1]1
r
12 1 1
~xAAAA,
O
1 I I
x~
I I
R
F I G U R E 6. (A) Series resistance-capacitance (RC) network. (B) Parallelogram method of obtaining the impedance (Z) of a series resistance-capacitance circuit. Parallelogram constructed with capacitative reactance (Xc) on the ordinate and resistance (R) on the abscissa gives the vector which is equal to Z. The phase angle (0) between current and voltage approaches 90 ~ as R approaches zero.
IX. Capacitive Charge and Capacitive Current The total charge (Q, in coulombs) stored on the plates of a capacitor is equal to the product of the capacitance and applied voltage:
O = CV
(26)
X. M e m b r a n e Time Constant When current is through a simple resistance or series of resistors, the PD across the network is developed instantaneously. This is not true for a biological membrane or for the parallel resistance-capacitance network shown in Fig. 8A. Instead, the PD builds in a nearly exponential (negative) manner until the final maximal value is attained, as illustrated in Fig. 8B. The slow charging occurs because of the presence of the capacitance. The time it takes for 63% of the final PD to be reached is, by definition, the time constant (~-). The equation describing this PD buildup is Wt = Wmax (l - e -t/~)
_[:F
A
c - (F)(v) Thus, if a certain charge is placed on a capacitor (air dielectric) and then the distance between the plates is decreased so that capacitance increases, it follows that the PD across the capacitor must decrease in proportion. It also follows that because an increase in dielectric constant increases capacitance, an increase in dielectric constant must also decrease the PD for a given charge separation; that is, it reduces the force acting between the separated charges. The electric charge stored in a capacitor can be used to do work. Kinetic energy is needed to charge a capacitor, the charged capacitor representing potential energy (like a battery), and discharge of the capacitor converts the potential energy back into kinetic energy. From Eq. 26, and because electric current flow (I) is a charge flowing past a point per unit of time (I = Q/t), the capacitive current flow (I o in amps) during charge or discharge of a capacitor is equal to the capacitance (in farads) times the rate of change of the PD:
dV I c - C d--t-
ELECTRONS REPELLED (LIKE CHARGES REPEL)
CHARGE INDUCTION
L_
FULLY CHARGED
ELECTRON FLOW
.+ § + +
(27)
C=FV s
(28)
s
Thus for a given capacitance, the greater the rate of change of voltage, the greater the capacitive current is. In a myocardial cell membrane, for example, the maximum rate of change in
F I G U R E 7. Illustration of the apparent flow of current through the dielectric of a capacitor. Current does not actually flow through the capacitor but appears to because of the phenomenon of charge induction. (A) When one plate of the capacitor is made negative, this plate repels the electrons from the other plate nearby, because like charges repel. (B) Thus the second plate has an electron deficiency that gives the plate a positive charge.
1212
Appendix
where V t is the voltage across the network at any time (t) and Vmax is the maximum or final PD when the steady state is reached. When t = ~-, then e -t/~ becomes e_l=l= e
1 = 0.37 2.717
(29a)
Therefore, gt
-
g m a x (1
-
W t - Wmax -
- 0.37)
(29b)
= 0.63Vmax For all other values of t, Eq. 28 is solved by taking the logarithms and obtaining
Vt -
t
(30)
antiln In Vmax- T
Vmax- antilog
2.303 log Wmax 2.303
(t/T)
On discharge of the resistance-capacitance network, the voltage across the capacitor ( V c ) decreases exponentially, as indicated in Fig. 8C. Again, the time it takes for 63% of the final PD to be reached (which is to the level of 37 % of the initial value) is the time constant (~-). The equation describing the decay in PD as a function of time (analogous to that for the decay in voltage as a function of distance along a cable) is
A
(32)
Wt - V m a x e - t / r
where Vo
ve
(31)
Wmax
is the maximum or initial PD at t = 0. When
t - - T,
(Applied V)
(33)
V t - O.37 Vmax
For all other values of t, Eq. 32 is solved by taking the logarithms t
In V t - In Vmax- ~"
vR2 B
vo
ioo o-I00
VMAX
,o~j.... o
Vc %
63
/
~ OFA i CAPACITOR '/-T" ~ FOLLOWS(NEG)
50
f
l
I00~
EXPONENTIAL
I \
50 1
o
0
w
DISCHARGE
\
antiln In Wma x -
t
(_
t(time)
0
V~
,
Vt -
]__
i EXPONENTIAL
0
C
I
v~
oe,
tltime) (A) Exponential (negative) charge and discharge of a ca-
FIGURE 8. pacitor for the parallel resistance-capacitance (R1C) circuit in series with
a second resistance (R2). Applied voltage is from a 100-V battery upon closing of the switch and is measured by voltmeter Va. Voltmeter V c measures the voltage across the capacitor, and voltmeter VR2 measures the voltage across R 2. (B) Upon closing the switch, Va suddenly jumps from 0 to 100 V. Potential recorded by V c, however, slowly increases exponentially; the time constant (~-) is the time it takes for V c to build to 63% (1 - e -1) of its final (or maximal) value. The instant that the switch is closed, the frequency effectively is infinite, and so capacitive reactance ( X c ) is zero; therefore no voltage appears across the parallel resistance-capacitance network, although the induced current through the capacitor is maximum. The entire voltage drop of the battery thus appears across R 2. After a long time, that is, when the circuit reaches steady state,f= 0 and X c is infinite; therefore the voltage across the parallel resistance-capacitance network is equal to (RIlR 1 + Re) • 100 V. This charging of a capacitor can be viewed as a compartment filling with electrons and obeying first-order kinetics. (C) Upon opening of the switch, the capacitor discharges exponentially through R 1. Time constant is the time it takes for discharge to 37% (e -1) of the initial (maximal) value.
Vt-
antilog
(34a) t
T
2.303 log Wmax 2.303
(34b) (t/T)
(34c)
In the circuit illustrated in Fig. 8A, upon closing the switch, the capacitor (C) charges exponentially with a time constant about equal to R 1 C (if R 1 is much greater than Re). At the first instant after the switch is closed, there is a step change in the applied potential that appears as an infinite frequency. Therefore, because X c = 1/2 7rfC, X c ~- O. According to Kirchhoff's laws pertaining to the relative voltage drops across series impedances, the voltage drop is in proportion to the relative impedance (i.e., the higher impedance has the greater voltage drop across it), and the sum of all voltage drops must equal the source emf. Thus, all the emf of the battery should be dropped across R 2 (VR2 = 100 V) and none across R 1 ( V c = 0 V). At the final steady-state value (plateau of applied voltage pulse), the applied voltage pulse is steady, and therefore the frequency is zero; hence X c ~ 0% and from analysis of the impedance of a parallel resistance-capacitance network, Z = R1. At this time, nearly all of the battery emf should be dropped across C (or R l) because R 1 >> R 2. In this condition, the current continues to flow through R 1 and R2. At all intermediate instants, V c > 0 V but X L and current leads voltage; at high frequencies, X L > X c and current lags voltage. At f0, the current is in phase with the applied voltage and X c = X c. Thus, 1 : wL (56a) coC or (.02.-
1 LC
(56b)
1 LC
(56c)
'
RcI>>RK F I G U R E 1 1. Circuit diagram (K. S. Cole circuit) depicting the apparent inductance of biological membranes (Lm). L m is in parallel to the membrane capacitance (Cm) and is due to the peculiar behavior of the K + resistance (RK); that is, there is no physical basis for the inductance as there is for the capacitance (namely, the lipid bilayer matrix). When C1- resistance (Rcl) is high compared to R K, as is true for most myocardial cells and neurons, the parallel inductance-capacitance circuit tends to oscillate at a resonant frequency ~0): f0 = 1/2"n'v/1/LC" The arrow through R K indicates that it is a variable resistance, depending on membrane potential. When the membrane is depolarized, R K increases (because of the K + channels responsible for so-called "anomalous rectification"). Because I K ( E m - E K ) / R K, the increase in driving force ( E m - EK) for outward K + current (IK) is counterbalanced by the increase in R K, the latter acting to hold the K + current constant. This is behavior typical of an inductance and gives rise to the apparent membrane inductance.
w h i c h is the same as Eq. 55, since ~o = 2-rrf. If a significant resistance (in series with the inductive branch, as for the excitable m e m b r a n e ) is present in the parallel (L + R)C circuit, then a d a m p i n g f a c t o r (R2/L 2) m u s t be introduced, and f0 is n o w given by
fo- ~
1 V// 1 LC
R2 L2
(57)
=
O n e characteristic of the parallel i n d u c t a n c e - c a p a c i tance tank circuit is that w h e n suitable e n e r g y is s u p p l i e d to such a circuit, o s c i l l a t i o n s in p o t e n t i a l tend to o c c u r at
TABLE 5
R a n g e of t h e E l e c t r o m a g n e t i c Frequency
Rcl
m
and
w-
,
Spectrum Wavelength (m)
(Hz) (10 kHz)
30 ooo (30 km)
1 X 10 6 1 X 10 7 3 x 108
(1,000 kHz) (10 MHz) (300 MHz)
300 (0.3 km) 30 1.0
TV broadcast
3 x 108
(300 MHz)
1.0
Microwave
3 x 10 l~
(30 GHz)
0.01 (1.0 cm)
Light Infrared Visible Ultraviolet
3 x 1012 3 x 1014 3 x 1016
1 • 10 -4 (0.1 mm) 1 • 10 -6 (1.0 Ixm) 1 • 10-8 (0.01 i~m;10 nm)
X-rays
3 x 1018
1 x 10 -1~ (1.0 ,~; 0.1 nm)
Gamma rays
3 x 102~
1•
Longwave
1X
Radio broadcast AM Shortwave FM
10 4
10 -12
(0.01 ~)
Source: Now You're Talking, edited by L. D. Wolfgang, J. Kearman, and J. P. Kleinman. Published by the American Radio Relay League, Newington, CT, 1996. Frequencies listed are for about the midrange of each spectrum. The corresponding wavelengths were calculated from the following equation, where c is the velocity of light, f is the frequency, and L is the wavelength. The sample calculation is for an AM radio wave of 1000 kHz. c m/s m L- ~ f cycles/s cycle _- 3 0 0 0 0 0 0 0 0 m / s _ _ 3• --3x102m 1000 kHz 1 x 106 cycles/s
70. Review of Electricity and Cable Properties the resonant (or natural) frequency. In effect, at one instant, all the energy is stored in the magnetic field around the inductor; this field collapses when the current through the inductor goes to zero. This effect induces an emf that drives current in the opposite direction and charges the capacitor. W h e n fully charged, all the energy is stored as electrical charge on the plates of the capacitor. W h e n the capacitor is fully charged and the current in the circuit goes to zero, the inductance now appears as a zero resistance (f = 0) across the capacitor, causing the capacitor to discharge into the inductor. This changing current induces a magnetic field in the inductor, and the cycle is repeated over and over. The oscillations produced are sinusoidal in nature. A small amount of energy needs to be put into the system; otherwise the oscillations dampen out because of the small resistance of the components and the resultant power (I2R) loss as heat. Oscillations of membrane potential, namely, the pacemaker potentials, are known to occur in many nerve and muscle cells under natural conditions and can be made to occur in all nerve and muscle membranes under experimental conditions. The nerve membrane (e.g., squid giant axon) has been demonstrated by Cole, Mullins, and others to also contain an inductive component (see Cole, 1968). Estimates of the apparent inductance for biological membranes (Lm) range from 0.5 to 10 H-cm 2. This inductance does not have a simple physical molecular basis, as do the resistance and capacitance components; hence it is described as an apparent inductance only. The L m seems to be intimately associated with the K + resistive component of the membrane (RK), as depicted-in the circuit diagram of Fig. 11. As indicated by the arrow through the K + resistance, R K varies as a function of the membrane potential (Em); that is, there is K + rectification. In K + anomalous rectification, when the membrane is depolarized, thereby increasing the driving force (E m - E K) for outward K + current, gK decreases; this decrease in gK tends to keep the outward
1219 K + current constant, since I K = gK (Em-EK), thereby mimicking the behavior of an inductance.
XIV. E l e c t r o m a g n e t i c S p e c t r u m The electromagnetic spectrum from radio waves and light waves through g a m m a rays is summarized in Table 5. As Table 5 shows, the higher the frequency of the electromagnetic wave, the shorter its wavelength. The equation relating wavelength (L) to frequency ( f ) and to relativistic velocity (c) is given in the footnote to Table 5. Note the similarity of this relationship to that for calculating the wavelength of a nerve impulse discussed in Chapter 24. The information given in this table has relevance to Fig. 1 of Chapter 69.
Bibliography Cole, K. S. (1968). "Membranes, Ions, and Impulses." University of California Press, Berkeley. Josephson, I. R., Sanchez-Chapula, J., and Brown, A. M. (1984). A comparison of calcium currents in rat and guinea pig single ventricular cells. Circ. Res. 54, 144-156. Sperelakis, N. (1979). Origin of the cardiac resting potential. In "Handbook of Physiology, The Cardiovascular System, Vol. 1: The Heart" (R. M. Berne and N. Sperelakis, Eds.), pp. 187-267, American Physiological Society, Bethesda, MD. Sperelakis, N. (1995). "Electrogenesis of Biopotentials." Kluwer Academic Publishers, New York. Sperelakis, N., and Lehmkuhl, D. (1966). Ionic interconversion of pacemaker and nonpacemaker cultured chick heart cells. J. Gen. Physiol. 49, 867-895. Sperelakis, N., Hoshiko, T., and Berne, R. M. (1960). Non-syncytial nature of cardiac muscle: membrane resistance of single cells. Am. J. Physiol. 198, 531-536. Tarr, M., and Sperelakis, N. (1964). Weak electrotonic interaction between contiguous cardiac cells. Am. J. Physiol. 207, 691-700.
Index
A Acetylcholine, inhibition of calcium channel, 568-569 Acetylcholine receptor channel acoustic transduction, 787-788 dysfunction in myasthenia gravis, 668-669 electrocytes, 1029, 1034-1037 neuronal channels, 682-683 nicotinic, s e e Nicotinic acetylcholine receptor/channel structure, 679-683 Acetylcholinesterase, T o r p e d o electrocyte, 1035-1036 Acoustic transduction, s e e a l s o Cochlea; Echolocation; Endolymph; Hair cell; Stria vascularis cell physiology, 782-788 Acquired neuromyotonia, potassium channel mutation, 668 Actin, see a l s o F-actin amoeboid movement, 959-966 binding proteins, 708-709, 959-961 cell motility, 959-960 comparison to other motility proteins, 960 crossbridge interaction with myosin, 945-949, 953, 955-956 depolymerization of stress fibers, 1010 filament, pH dependence, 368 ion channels, 602, 616 microvilli, 965 neuronal cytoskeleton, 480-481 polymerization, 959-960 Action potential (AP), s e e a l s o Action potential propagation; Afterpotential accommodation, 421-422 anodal-break excitation, 422-423 calcium channel activation and inactivation, 428-432 calcium-dependent, 873-875 cardiac, 585-586, 887-897 compound recording, 405-406 electrocytes, 1025, 1027, 1029, 1030 feedback loops, 414 frequency-modulated signaling, 395-396 gating currents of ion channels, 429-430 generation, overview, 427-428 local circuit currents, 400--401, 417-4 19 local excitatory state, 418 neuron developmental changes, 594-595 patch clamp analysis, 426 phases in external recording, 404-405,419
potassium channel activation and inactivation, 432-434 refractoriness, 420-421 repolarization, 434-435 resting potential effects, 436 skeletal muscle, 591-592, 865-883 smooth muscle, 899-909 sodium channels, 409, 428-429, 430--434 strength-duration curve, 421 threshold of excitation, 414, 418, 420 voltage clamp analysis, 424-427 wavelength determination, 403-404 Action potential propagation alternating current, 415-4 16 biological fiber as cable, 396, 412-416 conduction velocity calculation, 414 determinants of velocity, 401 electric field model, 407-410 fiber diameter effects, 395 input resistance, 399-400, 415 length constant, 396-398 linear cable theory, 413-4 16 local potentials, 400 membrane properties, 412-4 13 myelination effects, 401-403 saltatory conduction, 401-403 time constant, 398-399 velocities, 395,412-416 Active transport cotransporters, 250 countertransporters, 250 nutrients, 128 primary versus secondary, 250 Activin, development of electrical activity in skeletal muscle, 594 Activity coefficient, ions, 8,228-229 Adenosine triphosphate, s e e ATP Adenylate cyclase cyclic AMP synthesis, 155-156 regulation by G proteins, 155-156 /3-Adrenergic receptor, stria vascularis, 780 Afterpotential, s e e a l s o Action potential early depolarizing, 436, 872 early hyperpolarizing, 437 electrogenesis, 436--438 late depolarizing, 437, 872 late hyperpolarizing, 437, 872-873 physiological significance, 437-438 Agatoxin, 637-639 Agonist-activated channel, 446-447 Allosteric regulation, metabolic enzymes, 123
1221
Amiloride-sensitive sodium channel gene locus, 486-487 hormonal stimulation, 485-486 ion selectivity, 485 structure, 486 taste transduction, 818-821 tissue distribution, 485 Amino acid acid-base properties, 28-29 degradation for energy production, 126 essential, 126 interactions in a protein chain, 23-24 structures, 20 7-Aminobutyric acid (GABA) receptor conductance, 683 desensitization, 683 drug modulation, 684 structure, 683-685 Ammonium chloride prepulse technique, alteration of intracellular pH, 375-376 Amoeboid movement actin-based systems, 959-966 chemotaxis, 964 contractile proteins, 961-962 cytoplasmic streaming, 962, 963 mechanism, 960-961 myosin distribution, 962-964 organelle movement, 964 sol-gel transformation, 962-963 AMPA (a-amino-3-hydroxy-5-methyl4-isoxazole propionic acid) receptor postsynaptic density, 604, 606-607 properties, 684-686 Ampulla of Lorenzini, s e e Ampullary electroreceptor Ampullary electroreceptor active mode, 840 bias current, 844 biophysics, 857-861 chondrichthyan (cartilaginous) fishes, 840-843, 844, 857-861 clustering, 842-843 elasmobranchs, 840-842, 844, 857-861 frequency response, 839-840, 845, 859 lampreys, 840 morphology, 840-844 passive mode, 840 physiology, 844-845 sensitivity, 844, 845, 860-861 species distribution, 839-840, 843-844
1222 Ampullary electroreceptor (continued) synaptic ribbon, 841-842 teleost fishes, 843-844 Anchoring protein distribution, 604-606 scaffolding functions, 608 Anesthesia, reversal of propofol effects, 1018 Annexins cellular function, 172 sequence organization, 171 structure comparisons, 174 Antiadrenergic effect, cyclic GMP, 564 Ankyrin, interaction with sodium channels, 602-604 AP, see Action potential Apamin, effect on potassium channel, 636--637 Apoptosis bcl-2 regulation, 1176-1178 biochemical pathways, 1177 calcium-dependent, 1174-1176 caspase family, 1180 compared to necrosis, 1171-1173 disruption in cancers, 1166-1168 DNA damage induction, 1174 effectors, 1180-1181 mitochondria role, 149-150, 1178-1179 morphology of cells, 1172-1173 ontogeny role, 1173 p53 regulation, 1168 regulators, 1176-1180 stress induction, 1174 tumor necrosis factor receptor, 1173-1174, 1179 Aquaporin calcium receptor activation, 187-188 erythrocytes, 334-336 types and cellular distribution, 334-336 Arachidonic acid, modulation of exocytosis, 718 Archaea acidophiles, 1073 anaerobic, 1073 cell membranes, 1068 halophiles, 1072-1073 hyperthermophiles, 1073 Ataxia telangiectasia (AT), cell cycle defect, 1151 Atomic absorption spectroscopy, ion measurement, 12 Atomic Force Microscopy (AFM), 552 view of nuclear envelope surface, 554 ATP channel gating role, 729-732 high-energy bonds, 125 ion transport regulation in stria vascularis, 780 modulation of exocytosis, 717 rate of synthesis in cells, 141-142 structure, 125 Atrioventricular (AV) node action potential, 895 automaticity, 896 Autophosphorylation, calcium function, 170-171 Autosomal dominant hypoparathyroidism (ADH), calcium receptor mutation, 183 Autotrophy, bacteria, 1068 Avoidance reaction ciliary motion, 973-975 Protista, 1049-1051, 1057
Index Axon, see Neuron Axostyle, properties, 998-1000
B Bacterial flagellum assembly of substructures, 992-993 basal body, 992 chemotaxis, 995-996, 997 energy source, 993-994 filament structure, 992 hook, 992 swimming behavior, 994-995 Bacterium acidophiles, 1073 anaerobic respiration, 1069 autotrophy, 1068 bioluminescence, 1119-1122 capsules, 1067 cell wall, 1066, 1067 chemiosmotic coupling, 1069-1070 chemotactic behavior, 995-996, 997 classification, 1063, 1064 cytoplasm, 1065 cytoplasmic membrane, 1065-1066 endosymbionts in Protista, 1052-1054 fimbriae, 1067 flagella, 1066-1067 gliding, 996-998 Gram stain, 1063, 1067 halophiles, 1072-1073 hyperthermophiles, 1073 mycolic acids, 1067 nitrogen fixation, 1069 nucleoid, 1064-1065 outer membrane, 1066 photosynthesis, 1068 pili, 1067 porins, 1066 protonmotive force, 1069-1070 signal transduction, 995-996, 1071-1072 solute transport systems, 1070-1071 stress responses, 1071-1072 substrate-level phosphorylation, 1069 surface layers, 1067 Barium dance, ciliary motion, 974 Basal body bacterial flagella, 992 eukaryotic cilia and flagella, 971,990-992 Batrachotoxin, effect on sodium channel, 425, 627-629 Bcl-2, apoptosis regulation, 1176-1178 Binary phospholipid systems, mixing behavior, 59-61 Biogenesis, cell membrane, 91-92 Biological cable, structure and properties, 1215-1217 Bioluminescence annelids, 1127-1128 applications in research, 1129 bacteria, 119-1122 biochemistry, 1121-1122, 1123, 1125, 1126-1127 circadian clock, 1123-1124 coelenterates and ctenophores, 1124-1125 crustaceans, 1128
dinoflagellates, 1122-1124 evolutionary origins, 1115 fireflies, 1126-1127 fish, 1128 functions, 1116-1120, 1122-1123, 1124-1125, 1126 mechanisms, 1115-1116 mollusks, 1127 regulation, 1122, 1125, 1127 scintillon, 1123 species distribution, 1116 Birefringence, study technique, 67 Biscarboxyethyl carboxyfluorescein, measurement of pH, 374-375 Bisphosphate, modulation of potassium channel activity, 577 Boltzmann distribution, formula, 9 Boyle-van't Hoff law, osmotic pressure calculation, 383 Buffer buffering power of solutions, 357-359 closed, 358-359 open, 358-359
C Caffeine, effects on calcium release, 935-936 Calcilytics, calcium receptor antagonists, 184 Calcineurin, apoptosis induction, 1176 Calcium cellular function studies, 167, 174-176 cytotoxic cascade, red blood cells, 388-389 distribution and resting membrane potential, 224-226 excitation-secretion coupling role, 708-712 extracellular calcium-sensing cells, 188-189 hypoxia signaling, 737-740 insulin secretion, 726-729, 732-734 intracellular homeostasis, 175 intracellular signaling, 168 mechanism of release, 169 myonemes and spasmonemes, 988 nerve terminal depolarization-release, 694, 695-696 physiological processes, 167-168 regulation in skeletal muscle, 169 release from sarcoplasmic reticulum, 911, 927-938 signal creation, 168-169 signaling for exocytosis, 716-717 signal mediation, 169-170 signal transduction, 158-159 spikes, 595 systemic homeostasis, 179-180 waves, 595 Calcium-ATPase cardiac diseases, 277-278 in vitro studies on regulation, 275-276 in vivo studies on regulation, 276-277 plant cells, 1090 plasma membrane, 279-280 sarcoplasmic reticulum, 271-278 Calcium channel, see also Dihydropyridine receptor (DHPR); Ryanodine receptor (RyR) acoustic transduction, 784-785
1223
Index activation and inactivation, action potential generation, 428-432 antagonists, activity modulation, 644 apoptosis regulation, 1175 cardiac, 586-588, 616, 643-646 dihydropyridines, activity, 644-645 dysfunction and disease, 663-666 fetal-type, 588 gating-related domains, 470-471 inhibition by cyclic GMP, 563-568 inhibition by G proteins, 578-580 inhibition by muscarinic agonists, 568-569 ligand-gated, 596 mutagenesis studies, 644 neurons, 594-596, 615,665-666 pharmacology, 644-645 plant cells, 1089-1090 pore formation, 471-473 recovery process following activation, 432 regulation by cytoskeleton, 615 schematic model, 560 skeletal muscle, 593, 866-867, 874 smooth muscle, 901,903-904, 908 sperm, 510-519 stimulation by cyclic AMP, 561-563 structure, 433 synaptic transmission, 694-696 targets for toxins, 637-639 trp family, 492-495 types, 559-561 voltage-dependent, 466, 468, 579-580, 595-596, 637-639, 784 Calcium current cardiac, 890-891 L-type, 891 Calcium flux apoptosis, 1174-1176 basal, inhibition, 566-567 ciliary motility, 973-975 cyclic GMP inhibition, 563-568 effect of nitric oxide, 567-568 mediation of signal, 176 sensory receptors, 769 Calcium receptor allosteric modulation, 184 calcilytics, 184 calcimimetics, 184 calcium-binding domains, 181-182 gene regulation, 182 G protein coupling, 179, 181-182, 188 ion specificity, 183-184 ligand sensing, 181-182 medullary thyroid carcinoma, 187 mutations in disease, 182-183 parafollicular cells, 186-187 parathyroid cells, 182-187 renal epithelial cells, 187-188 structure, 180-181 Calcium-sodium exchange, s e e a l s o Sodiumcalcium exchange reaction, 224-226 Calcium transient, developing neurons, 595 Calmodulin electrocyte expression, 1034-1035 hair cell adaptation, 784 modulation of calcium effects, 159, 170-171
modulation of exocytosis, 718 regulation of nonstriated muscle contraction, 171 Calsequestrin, excitation-contraction coupling, 912 Calvin cycle, overview, 1099-1101 cAMP, s e e a l s o Cyclic AMP Campaniform receptor, arthropod, 764 cAMP/phosphorylation cascade, 131-132 Cancer cell adhesion, 1164-1166 apoptosis, 1166-1168 cell cycle, 1161-1163 DNA repair defects, 1163-1164 genome instability, 1163-1164 migration, 1165-1166 tumor suppressor gene mutations, 1162-1164 Candidate gene, ion channel mutations in disease, 654 Capacitance flickering in vesicle fusion, 713-714 jump in vesicle fusion, 712-713 membranes, 220 Capacitor charge, 1211 current, 1211 definition, 1207-1208 parallel and series, 1208-1209 reactance, 1209 specific capacitance, 1214-1215 Cardiac glycoside affinity for sodium,potassium-ATPase, 267 ligand-induced conformational change, 267 primary domains, 268 sodium-potassium pump, 266-268 sodium pump lag hypothesis, 267 Cardiac muscle ATP-sensitive potassium current, 893-894 atria, 894 atrioventricular node, 895,896 automaticity, 895-897 calcium channels, 586-588, 616, 643-646 calcium current, 890-891 cAMP-stimulated chloride current, 893 cell volume versus osmolarity, 342 delayed rectifier, 891-892 developmental changes, 585-591 electric field model of propagation, 407-410 excitation-contraction coupling, 928 hyperpolarization-activated inward current, 590-591 inward rectifier, 892-893 pacemaker current, 590-591 phases of action potential, 888-893 potassium channels, 589-590, 610, 616, 736-737 Purkinje fibers, 895 resting potential, 585,887-888 sinoatrial node, 894, 896 sodium-calcium exchange, 295-296, 893 sodium channels, 586, 616, 647-649, 663 sodium current, 888-889 sodium-potassium pump, 893 transient outward current, 889-890 Cardiac myopathy, sarcoplasmic reticular calcium-ATPase activity, 278
Cardiomyocyte, s e e Cardiac muscle Carotid body, chemoreceptor cells, 737-740 Carrier-mediated transport cotransport, 254-257 countertransport, 257-259 facilitated diffusion, 250-254 mechanisms, 250-259 Cation-chloride cotransporter, targets for disease, 316 C cell, s e e Parafollicular cell's calcium functions ce Cell, stimulus-secretion coupling, 735-736 /3 Cell, s e e a l s o Insulin secretion hypotheses of insulin secretion, 726-729 role of potassium channels, 729-732 Cell cycle, s e e Mitosis Cell volume regulation, s e e a l s o Osmosis; Regulatory volume decrease (RVD); Regulatory volume increase (RVI) anisosmotic conditions, 342-352 betaine transporter, 346-347 compensatory regulation, 342-343 Gibbs-Donnan equilibrium, 336-338 isosmotic conditions, 336-341 nonexcitable cells, 501-503 pump-leak hypothesis, 338-339 red blood cells, 383-386 signaling pathways, 347-352 sodium-calcium exchanger, 341 sodium-potassium pump, 339 taurine transporter, 346-347 Cell wall biosynthesis, 1080-1082 development, 1083-1084 functions in plants, 1082-1083 stress relaxation, 1083 structures, 1080-1082 Central core disease, calcium release channel mutation, 664 Centrin, eukaryotic flagellar rootlets, 991-992 cGMP, s e e Cyclic GMP Channel, s e e a l s o Calcium channel; Chloride channel; Ion channel; Potassium channel; Sodium channel in gap junction, 527-528 Channel gating correlations and two-time distributions, 448 dwell-time distributions, 447 gap junction, 529-530 gated channel families, 445 Channelopathy, s e e s p e c i f i c d i s e a s e Charcot-Marie-Tooth disease, 534 Charybdotoxin, effect on potassium channels, 634-636 Chemical gradient, measurement, 792-793 Chemical potential, ions in solution, 17 Chemiosmotic theory, s e e Mitchell chemiosmotic hypothesis; Photosynthesis Chemoreceptor cell, stimulus-secretion coupling, 737-740 Chloride active transport mechanisms, 302-303 distribution across plasma membrane, 302 distribution and resting membrane potential, 223-224 reabsorption, vectorial transport, 504 role in cellular functions, 301
12 2 4 Chloride-bicarbonate exchanger intracellular pH regulation, 364 Chloride channel, s e e a l s o Cystic fibrosis transmembrane regulator (CFTR) calcium-activated, 497 cell volume regulation, 501-503 CLC-1 skeletal muscle channels, 658-659 cyclic AMP activation, 498-500 double-barreled gating, 1036 dysfunction and disease, 655-659 electrocytes, 1036 epithelial, 614--615 kidney CLC-5 channels, 659 mammilian, 496 myotonia, 655,656, 658-659 regulation by cytoskeleton, 614--615 skeletal muscle, 868-870 smooth muscle, 901 target for toxins, 639-640 vectorial transport role, 503-506 voltage-dependent, 495-497, 869-870 volume-regulated, 500-501 Chloride current cAMP-stimulated, 893 cardiac action potential, 893 Chloride formate exchanger inhibition, 366 intracellular pH regulation, 366 Chlorophyll, structure, 1097 Chloroplast, s e e a l s o Photosynthesis ATP synthase, 1105-1106 bacterial origin, 1063, 1098 electron transport, 1106-1111 genes, 1098 protein transport, 1098 structure, 1098, 1099, 1100 Chlorotoxin, effect on chloride channel, 639-640 Cholesterol, conformation, 50-51 Cholesterol transbilayer domains, structure, 84 Chord conductance equation, 231-232, 239-240 Chordotonal receptor, arthropod, 764 Chord resistance equation, 232, 240 Chromosome walking, ion channel gene mutations, 653 Cilium axoneme, 968, 970-971 basal body, 971 beat cycle, 966, 973,975-976 bend formation, 972-973 ciliary membrane, 970 discovery, 966 dynein arms and ATPases, 968-969, 974 flagellar assembly, 971-972 linkers and accessory structures, 969-970 metachrony, 976-977 microtubules, 968, 972-973 modulation and control, 973-977 reversal, 973-975 sensory receptors, 977 shaft, 968-970 sliding microtubule hypothesis of motility, 972 striated rootlets, 990-991 structure, 966-971,972 transition zone, 970-971
Index Circuit analysis, resting membrane potential, 240-241 Circular dichroism, protein secondary structure analysis, 25, 27 Cistern, description, 541 C1C channel, s e e Chloride channel Cochlea, s e e a l s o Endolymph, Hair cell; Stria vascularis basilar membrane, 775 endocochlear potential (EP), 776, 778-779 endolymph, 775-782 hair cell, 775 perilymph, 775,776 sensitivity, 782 stria vascularis, 775 structure, 775,776 tectorial membrane, 775 vestibular dark cells, 779-781 Colloid osmotic pressure colloid osmotic hemolysis, 385-386 effects on Gibbs-Donnan equilibrium, 246 Competitive/noncompetitive inhibition, mediated transport systems, 213 Concentration cell, development of equilibrium potential, 227-228 Conduction saltatory, 401-403 velocity calculation, 414 Connexin, s e e a l s o Gap junction cataracts role, 534 channel selectivity, 529 characterization, 526-527 electrical properties, 529 functions, 529, 530-531 mutation, in Charcot-Marie-Tooth disease, 534 c~-Conotoxin effect on nicotinic acetylcholine receptor/channel, 639 structure, 639 /x-Conotoxin, binding site, 626--627, 628 w-Conotoxin effect on calcium channel, 637-639 structure, 638 Constant-field equation, calculation of resting membrane potential, 230-231,237-239 Contractile ring, formation, 964-965 Coil cycle, overview, 136-137 Cotransport carrier-mediated transport, 254-257 coupling, 255-256 electrogenic mechanisms, 256 functional properties of solute transport, 256-257 inhibition, 255 kinetic model, 254 sodium-glucose isoforms, 256-257 unidirectional solute transport, 255 Coulomb's Law, 4, 8 Countertransport carrier-mediated transport, 257-259 functional properties of solute transport, 259 homoexchange of solutes, 257 heteroexchange of solutes, 257 inhibition, 257-258 kinetic model, 257
mechanism, 258-259 significance, 259 unidirectional solute transport rate, 257 Creatine kinase, energy storage, 134 Crenator, effect on mechanosensitive ion channels, 748 Crossbridge theory, s e e Muscle contraction Cuticular receptor, arthropods, 764 Cyclic AMP (cAMP), s e e a l s o Cyclic nucleotide-gated (CGN) ion channel gap junction, effects, 529-531,533 hormone response mediation, 156, 157 ion transport regulation in stria vascularis, 780 nonselective cation channel activation, 490-491 olfactory transduction, 796, 798, 825,827, 829 signal transduction, 155-156 stimulation of calcium channels, 561-562, 569 sucrose taste transduction, 822 synthesis, 155-156 Cyclic AMP response element binding protein (CREB), transcription factor, 199-200 Cyclic GMP (cGMP), s e e a l s o Cyclic nucleotide-gated ion (CNG) channel endogenous hydrolysis, 567 inhibition of calcium channels, 563-567, 569 mediation of biological processes, 157 nonselective cation channel activation, 490--491 olfactory transduction, 825, 828 pathophysiological effects, with nitric oxide, 568 signal transduction, 157-158 synthesis, 155-156, 158 visual transduction, 795-796, 798 Cyclic GMP-dependent protein kinase (PKG), signal transduction, 158 Cyclic nucleotide, s e e a l s o Cyclic AMP; Cyclic GMP modulation of exocytosis, 717-718 Cyclic nucleotide-gated (CNG) ion channel, s e e also specific channels
conductance, 800 control by cyclic nucleotide enzyme cascades, 796-797 functional modulation, 802-803 gating mechanism, 797-799 ion selectivity, 799-800 molecular structure, 801-802 olfactory transduction, 796, 798-800, 825-826, 827-829 permeation, 799-800 pharmacological blocking, 800-801 rationale for gating, 795 tissue distribution, 795-796 visual transduction, 795-796, 798, 800-801, 808-809 Cyclin cyclin A, 1139 cyclin D, 1139 types, 1137-1139, 1142-1143 Cyclin-dependent kinase (CDK) cyclin activation, 1136 G1-S phase transition regulation, 1136-1141 G2-M phase transition regulation, 1141 inhibitors, 1136, 1144-1146, 1148 types, 1137, 1143-1144
1225
Index Cystic fibrosis transmembrane regulator (CFTR) cyclic AMP activation, 498-500 epithelial sodium channel, 486--487 mutation in disease, 655-658 water transport, 336 Cytochalasin, effect on nuclear permeability, 555 Cytokine receptor, signal transduction, 202-203 Cytoplasm, electron microscopy, 541 Cytoplasmic streaming amoeboid movement, 962, 963 plant cells, 1085 Cytoskeleton excitation-secretion coupling role, 705-708 exocytosis and changes, 711-712 filaments, 107-109 function, 107-113 hydrostatic pressure effects, 1010 intermediate filaments, 109 ion channel distribution, 601-6 12 ion channel function, 613-616 microfilaments, 107, 109, 480--481,705-708 microtubules, 109-113, 481-482 neurofilaments, 481 neuron, 480--482 plant cells, 1085 postsynaptic density, 604-608 red blood cell, 378 relation to mechanosensitive gating, 617, 753 thick filaments, 109 Cytosol, electron microscopy, 541
disease and calcium channels, 663-664 excitation-contraction coupling, 911 genetic studies, 920-922 in vitro studies, 922 in vivo studies, 922-923 interaction with ryanodine receptor, 917-924, 936-937 properties, 915 structure, 917 voltage sensing, 915-917, 919-920 Dinoflagellate, bioluminescence, 1122-1124 Dipole moment induced, 6 water, 4 DNA damage apoptosis, 1166-1168 dose-rate effect, 1195 ionizing radiation, 1187-1188 oxygen effect, 1195-1196 repair defects in cancer, 1163-1164 repair of ionizing radiation damage, 1193-1195 Donnan equilibrium, see Gibbs-Donnan equilibrium Double membrane, structure, 67 Down-regulating, metabolic enzymes, 129 Dynein, cilia, 968-969, 973,974 Dystroglycan, rapsyn role in clustering, 602 Dystrophin-glycoprotein complex (DGC), 602
E D Davson-Danielli paucimolecular membrane, model of structure, 66-67 Deafness, potassium channel mutation, 668 Debye equation, 7 Debye-Hiickel theory, 8-10 Decremental conduction, sensory receptor measurement, 765-766 Dendrotoxin, effect on potassium channel, 637 DGC, see Dystrophin-glycoprotein complex D-Glucose transport, isoforms function, 254 Diabetes mellitus altered glucose metabolism, 126 insulin-dependent (IDDM), 736 noninsulin-dependent (NIDDM), 736 Diacyl phospholipid hydrophobic tail, 45 interfacial region, 45 plasmalogens, 44 platelet-activating factors, 44 polar headgroup, 45 structure, 43-46 zigzag plane, 45 Dielectric constant, water, 4 Diffusion across membrane with partitioning, 212-213 coefficient, 21 0-212 Einstein's law, 214 exponential time course, 217 membrane transport overview, 209-210, 213 Dihydropyridine receptor (DHPR) calcium release, 918,919, 920, 922-924, 927, 928, 936, 937
E2F, cell cycle regulation, 1149-1150 Echolocation, 788 Effectors of membrane receptors, 573 EF hand, intracellular calcium binding, 169-170 Einstein-Stokes equation, 211-212 Eisenman's selectivity sequences, binding of alkali cations to negatively charged sites, 10 Electrical potential, calculation for proteins, 30-31 Electric field model, action potential propagation electric field effect, 408-409 electronic model for simulation of propagation, 410 experimental data, 408 sodium channel density, 409 Electricity, see also Capacitor; Impedance; Inductor ;Resistance capacitive charge and current, 1211 capacitors, 1207-1209 circuit elements, 1205 Faraday constant, 1206 Kirchhoff's laws, 1207 membrane impedance, 1209-1211 membrane time constant, 1211-1213 Ohm's law, 1205 resistance-capacitance networks in parallel and in series, 1209-1211 resistors and conductances in series and in parallel, 1206-1207 Electric organ discharge (EOD), 840, 846-849 Electrochemical potential equation, 18, 215,321,337-338 solute activity effects, 249-250
Electrocyte acetylcholine receptors, 1029, 1034-1037 acetylcholinesterase in Torpedo electrocyte, 1035-1036 action potential, 1025, 1027, 1029, 1030 anatomy of Electrophorus, 1025-1027 calmodulin expression, 1034-1035 chloride channels, 1036 comparative physiology of Electrophorus and Torpedo, 1032-1036 electric organ discharge, 846-849 electromotor neurons, 1026 end-plate potentials, 1027, 1029, 1034-1035 equivalent circuits, 1029-1032 mechanism of electrical discharge, 1025-1027 membrane and extracellular potentials, 1027, 1029 resting potential, 1027 sodium channels, 1032-1034 sodium,potassium-ATPase, 1034 structure, 1028 transcellular potentials, 1027, 1029, 1030 Electrodiffusion overview, 213-215 theory, 381-382 Electroneutral cotransport magnitude and direction, 307-308 potassium-chloride cotransporters, 313-315 sodium-chloride, 315 sodium-potassium-chloride cotransporters, 303-313 stoichiometry, 306-307 Electron paramagnetic resonance, electron donors in photosynthesis, 1108 Electron probe microanalysis, ion measurement, 12 Electron transport system, mitochondria, 140--141 Electrophoresis, protein charge, 31-32 Electroreceptor, see Ampullary electroreceptor; Tuberous electroreceptor Electrostatic field effect, charged protein on dissociation, 30 Embryogenesis, amoeboid movement, 965 End bud, see Ampullary electroreceptor Endocrine cells docking and fusion of secretory granules, 719 excitation-secretion coupling, 718-720 Endolymph composition, 775-776 control by stria vascularis, 776-781 reabsorption in outer sulcus epithelium, 781-782 Endoplasmic reticulum organization, 101 relationship to nuclear pore, 543 ribosomes, 101 rough, 101 smooth, 101 Endosomes, formation, 105 Endothelial cell calcium-activated nonselective cation channels, 492-495 mechanically activated ion channels, 493 End-plate potential, see Electrocyte Enthalpy of hydration, ions, 6-7
1226 Enzyme inhibitors, 122-123 kinetics, 120-122 maximal velocity, 120 metabolic pathway regulation, 119-120, 123, 129-130 protein catalysts, 119-120 transition state, 119 Epilepsy potassium channel mutation, 668 sodium channel mutation, 663 Epinephrine, tissue-specific regulation, 132-133 Episodic ataxia type 1, potassium channel mutation, 666-667 Episodic ataxia type 2, neuronal calcium channel mutation, 665-666 Epithelial cell amiloride-sensitive sodium channels, 485-487, 613-614 calcium-activated nonselective cation channels, 492-495 Epithelial sodium channel (ENaC), regulation by cytoskeleton, 613-6 14 Epo, s e e Erythropoietin Equilibrium processes in aqueous solution, 1004-1005 Equilibrium potential development in concentration cell, 227-228 energy wells, 229 equivalent electrical circuit, 226 Gibbs-Donnan equilibrium, 336-338 measurement, 226--227 Nernst equation, 226-227 synapses, 699-700 Erythrocyte, s e e a l s o Red blood cell aquaporin density, 334-336 Erythropoietin, secreting cell, 741-742 Escape reaction ciliary beat frequency, 975-976 EVH (enabled/VASP homology) domain, protein-protein interactions, 611 Excitation-contraction (EC) coupling activation speed, 913-914 calcium receptors, 918-922 calcium release, 911-912, 927-938 cardiac muscle, 591 charge movement, 918, 920 mechanical threshold, 915-916 membrane architecture, 915-919 modulation, 924 signal transduction at triad junctions, 914-919 skeletal muscle, 911-925 sodium-calcium exchange, 296-297 spasmonemes, 989-990 Excitation-secretion coupling, s e e a l s o Exocytosis calcium signaling, 708-712 cytoskeleton role, 705-708 Exocytosis actin-binding proteins, 708-711 calcium signaling, 716-717 cytoskeleton changes, 705-708, 711-712, docking and fusion proteins, 709-711 effectors, 71 6-717 fusion of vesicles to plasma membrane, 712-715
Index guanine nucleotides, 717 hormone release, 718-720 modulators, 717-718 proteins in synaptic vesicle docking and fusion, 696-698 secretory granule pools, 718 synaptic vesicles, 696--699 termination in synaptic vesicles, 698-699 Extracellular potential, measurement, 792-793 Extrusome, s e e Protista
F Facilitated diffusion cartier-mediated transport, 250-254 competitive versus noncompetitive inhibition, 252 conceptual and kinetic model, 250 electrogenic facilitated diffusion of charged solutes, 252 kinetic parameters, 252 mechanisms, 252 overview, 250-251 structure-function relationship of transport protein, 252-253 unidirectional solute transport, 251-252 F-actin, neuronal NMDA channel regulation, 615 Familial benign hypocalciuric hypercalcemia (FBHH), calcium receptor mutation, 182 Familial hemiplegic migraine, neuronal calcium channel mutation, 665-666 Faraday constant, calculation, 1206 Fatty acid, roles, 126 Fatty acid transbilayer domains, structure, 84 Ferredoxin, electron transport, 1111 Fetal-type calcium channel, 588 Fiber recruitment, recording by electromyography, 405 Fick's diffusion equations, 210, 212 Flagellin, comparison to other motility proteins, 960 Flagellum eukaryotic, s e e Cilium bacterial, s e e Bacterial flagellum Flame photometry, ion measurement, 12 Fluid mosaic model illustration, 97 membrane structure, 72 Fluorescence, mechanism, 1106 Fluorescent chelator dyes, ion measurement, 12 Fodrin, description, 611-612 Free energy, ion transport, 16-17 Fusion pore conductance, 715 formation, 713-714 leakage, 71 4-715 size, 714
G Gap junction, s e e a l s o Connexin biological functions, 531-533 carcinogenesis, 533-534 channel gating and regulation, 529-530 channel conductance, 527-528
Charcot-Marie-Tooth disease, 534 charge selectivity, 528-529 compared to nuclear pore, 544 deafness, 534-535 discovery, 523-524 disease, 533-535 electrical properties, 529 heart function, 534 intracellular pH effects, 368 ion selectivity, 528-529 protein characterization, 524, 526--527 stria vascularis basal cells, 777 ultrastructure, 524 Gardos effect, cytotoxic calcium cascade, 389 Gating current ion channels in action potential generation, 429-430 sodium channel inactivation, 430-432 Geographutoxin, s e e / x - C o n o t o x i n Gephyrin, glycine receptor clustering, 604 Ghost, isolated membranes, 67 Gibbs-Donnan equilibrium, 539 cell volume regulation, isosmotic conditions, 336-338 colloid osmotic pressure, 246 development, 337 electrochemical potential, equation, 337-338 ion distributions, 245 osmotic considerations, 246 principle of electroneutrality, 245 quantitation, 245-246 red blood cells, 378-379, 384-386 Gibbs-Donnan equilibrium potential, mechanism for development, 243-244 Gibbs equation, 17 Gigaseal technique, s e e Patch clamp GK, s e e Guanylate kinase Gliding bacteria, 996-998 unicellular eukaryotes, 998 Glomus cell, s e e Chemoreceptor cell Glucagon secretion, stimulus-secretion coupling, 735-736 Glucocorticosteroid, ion transport regulation in stria vascularis, 780 Glutamate receptor/channels distribution, 604-608 synaptic transmission, 684--686 Glyburide, effect on insulin secretion, 731 Glycerophospholipid, structure, 43-49 Glycine receptors (GlyR) chloride channel, 669 conductance, 683 drug modulation, 684 inhibitory, 604 mutation in hyperekplexia, 669 structure, 684-685 Glycogen metabolism regulation, 131 storage, 128 synthesis and degradation, 126 Glycolysis, catabolic pathway, 125 Glycosome ATP formation, 1046 Protista, 1045-1046 Goldman equation, 548
Index Goldman-Hodgkin-Katz constant-field equation, s e e Constant-field equation Golgi apparatus cisternae, 101 functions, 101,103 plant cells, 1084 G protein activation, 154-155 calcium channel inhibition, 578-579 calcium receptor, 179, 181-182, 188 cycle, 573,574 ion channel modulation, 155,445 olfactory transduction, 825 potassium channel activation, 573-574 regulation by RGS proteins, 578 signal transduction, 153-160 slow synaptic transmission mediation by coupled receptors, 701-702 target tissue responses, 154 taste transduction, 821-822 toxins and inhibition, 155 transcriptional activation, 199-200 Gradient, chemical, 792-793 Graham's law, 211 Gramicidin ion channels, 75-78 ion selectivity, 13 Gravity receptor, plant cells, 1091 Grayanotoxin, sodium channel binding, 627-629 Grotthus chain mechanism, transport, 8 GTPase activating protein (GAP), in signal transduction, 162 Guanine nucleotide, exocytosis effector, 717 Guanine nucleotide-binding proteins, s e e G protein Guanylate cyclase, regulation, 158 Guanylate kinase (GK) domain, protein-protein interactions, 611 Gustatory receptor, s e e Taste receptor
H Hair cell, s e e a l s o Acoustic transduction; Cochlea; Endolymph adaptation, 783-784 apical membrane channels, 782-783 basolateral membrane channels, 784-786 gating-spring model, 783 kinocilium, 966 receptors, 787-788 reverse transduction, 786-787 stereocilia, 966 synaptic release of vesicles, 786 tip links, 783 transduction channels, 782-783 Hairlike receptor, arthropods, 764 Half-cell potential, artifact in membrane potential measurements, 237 Helical hairpin hypothesis, 33 Hematocrit, definition, 379 Hemisodium, ion selectivity, 13 Henderson-Hasselbalch equation, 357 Hepatocytes, isomotic volume regulation, 341 Heterotrimeric GTP-binding protein, s e e G protein Hexose monophosphate (HMP) shunt, roles, 125 Hodgkin-Huxley analysis, ion channel activation, 430-435
1227 Hofmeister series, ion effects on proteins, 10 Homeostasis, definition, 3 Homer ligand motif, linking mGlu receptors, 608 Homer protein, binding to mGlu receptors, 607-608 Hormone, binding to cells, 130 Hydrogen bond, strength, 3-5 Hydrogenosome, Protista, 1045, 1046-1047 Hydronium ion, s e e a l s o Proton physical properties, 7-8 Hydrophobic effect, bimolecular membrane formation, 70 Hydrostatic pressure adaptation of animals, 1018-1019 aquatic animals, 1014-1015 breathold diving, 1016 buoyancy effects, 1015 cellular effects, 1010-1014 chondrocytes, effects, 1014 cortical cytoplasm gel strength, 1010 cytoskeletal protein disruption, 1010 depolymerization of actin stress fibers, 1010 effects on animals and humans, 1014-1018 excitable membranes, effects, 1013-1014 homeoviscous theory of membrane adaptation, 1018-1019 inert gas interactions, 1017-1018 inotropic effects, 1012 lipid bilayers phase state, 1005-1006 membrane permeability and transport, effects, 1013 molecular effects, 1004-1009 muscle effects, 1011-1013 protein adaptations, 1019 rates of chemical reactions, 1006-1009 reactions catalyzed by enzymes, 1008-1009 scuba diving, 101 6-1017 stability of proteins and supramolecular structures, 1005 thermodynamics of equilibrium processes in aqueous solution, 1004-1005 transition state theory, 1006 units and occurrence, 1003 Hyperekplexia, glycine receptor mutation, 669 Hyperpolarization-activated inward current, developmental changes, 590-591 Hyperthyroidism, sarcoplasmic reticular calcium-ATPase activity, 277-278 Hypokalemic periodic paralysis, dihydropyridine receptor mutation, 664 Hypothyroidism, sarcoplasmic reticular calciumATPase activity, 277-278 Hypoxia histotoxic, 739 signaling in carotid chemoreceptor cells, 737-740 signaling in vascular smooth muscle, 740-741
I Iberiotoxin, effect on potassium channel, 636 Impedance, cell membrane, 1209-1211 Inductor definition, 1217 impedance, 1217 oscillations, 1219
reactance, 1217 resonant frequency, 1217-1219 Information capacity, sensory receptor, 763 Infrared receptor occurrence in animals, 832 pit organs in snakes, 833-837 sensory transduction, 835-837 Inositol trisphosphate (IP3) olfactory transduction, 826-827 taste transduction, 822 Insulin secretion consensus hypothesis, 726-728 defects and pharmacology, 736 depolarization-secretion coupling hypothesis, 726-729 description, 718-719 fuel hypothesis, 726-727 readily releasable pool, 734-735 role of calcium, 726-729, 732-734 Intercellular junction, classification, 113-114 Intracellular membrane organelle lipid domains, 90-91 properties, 90 protein functions, 91 structure and functions, 81-83 Intracellular pH acidic intracellular organelles, 360-361 active membrane transport of acids and bases, 362-366 buffering power, 357-359 calculation, 359 cell proliferation effects, 369 cell volume regulation, 369-370 chloride-bicarbonate exchanger, 364 effects on cell coupling, 368 effects on cytoskeleton, 368 effects on membrane flow, 370 effects on metabolism, 367 effects on muscle contraction, 368 effects on signaling molecules, 368-369 maintenance of steady state, 361 membrane conductance, 368 metabolic production and consumption of acids, 361-362 organelle control, 359-361 passive transmembrane proton flux, 362 potassium-proton exchanger, 364 proton-ATPase, 365 recovery, 376 sodium-bicarbonate-chloride exchanger, 364-365 sodium-bicarbonate cotransporter, 365 sodium-organic anion cotransporter, 365-366 sodium-proton exchanger, 362-364 Intracellular recording, sensory receptor, 766 Intraflagellar transport (IFT), 972 Ion anions in cells, 11 cations in cells, 10-11 chemical potential, 8-10 Debye-Hiickel theory, 8-10 enthalpy of hydration, 6-7 Hofmeister series of protein effects, 10 ionic strength, 8 ionic selectivity, 10 ionization energy, 7
1228 Ion (continued) hydration, 5-7 measurement of electrolytes and membrane potential, 12-13 mobility, 6, 8 physicochemical properties, 211 single-file diffusion, 10 Ion channel, see also specific channels cardiac, 616 cyclic nucleotide-gated, 795-804 developmental changes, 585-597 domain-dependent distribution, 601-6 12 epithelial, 613-6 15 gating, 429-430 hydrostatic pressure, effects, 1014 inactivation mechanisms, 37-38 intracellular pH effects, 368 ligand-gated, 596, 675-686 mechanosensitive, 491-492, 617, 745-758, 768 neuronal, 615 perm-selectivity, 37 phosphorylation, role in clustering, 612-6 13 properties in planar lipid bilayers, 74-75 regulation by cytoskeletal proteins, 613-6 16 skeletal muscle action potentials, 866-876 structure, 35, 37 target for disease, 653-670 target for drugs, 643-649 target for toxins, 625-640 vectorial transport role, 503-506 voltage activation, 35-36 voltage-gated, 595-596 Ion channel disease autoimmune disease, 654 mutations, 653 myotonia, 654-655 Ion channel gene mutations, effects, 653 Ionizing radiation background radiation, 1185 cell cycle phase and sensitivity, 1191-1193 cell survival curves, 1189-1191 direct versus indirect ionization, 1186-1187 DNA damage, 1187-1188 DNA repair, 1193-1195 interactions with matter, 1186-1187 linear energy transfer, 1197 measurement of dose, 1187 radioprotectors and dose reduction factors, 1198-1200 relative biological effectiveness, 1197-1198 scattering of neurons, 1187 types, 1185-1186 weighting factors, 1198 Ion leaks flux expression, 143-144 significance in mitochondria, 142-1 44 Ionophore, ion selectivities, 13 Ion-specific glass microelectrodes, ion measurement, 12 Ischemia sarcoplasmic reticular calcium-ATPase activity, 278 Ischemia-reperfusion injury mitochondria protection, 147-148 Islet cell, see Pancreatic islet cell
Index Isoelectric point, proteins, 29 Isoionic point, proteins, 29
J Jacobs-Stewart cycle, bicarbonate production, 386-387 Jak-Stat pathway, cytokine receptor signal transduction, 202-203 Janus kinase, cytokine receptor signal transduction, 202-203
K Kainate receptor/channel, postsynaptic density, 606 Kalicludine, effect on potassium channel, 635 Kaliotoxin, effect on potassium channel, 636 Kaliseptine, effect on potassium channel, 635 Keratinocyte, calcium sensing, 183 Ketoacidosis, 133-134 Kinesin, structure and properties, 482 Kinetochore, microtubules, 980 Kirchhoff's laws, definition, 1207 Krebs tricarboxylic acid (TCA) cycle, overview, 126 Knockout mouse, transgenic mice studies, 1146, 1155
planar lipid bilayer, 51-52, 73-74, 329-330 protein distribution, 87-90 rafts, 86 sphingolipids and glycolipids, 85 structure, 46 transbilayer lipid distribution, 83-85 vesicle, 329-330 water transport mechanism, 332-333 Lipophilic hormones, signal transduction mechanisms, 191-197 Liposome formation, 52 types, 52 Liquid ion exchange microelectrodes, ion measurement, 12 Loop diuretics, sodium-potassium-chloride cotransporter inhibition, 308 Long-QT syndrome potassium channel mutation, 667-668 sodium channel mutation, 663 Low density lipoprotein receptors, calcium sensing, 183 Lysosomes autophagy, 103 functions, 103, 105 heterophagy, 103 phagocytosis, 105 pinocytosis, 105 receptor-mediated endocytosis, 105
L Lambert-Eaton syndrome, neuronal calcium channel mutation, 665 Large unilamellar vesicles, formation, 52 Lecithin, see Phosphatidylcholine Ligand-activated nuclear receptors invertebrate receptor genes, 192-193 nuclear receptor gene family, 192-193 Ligand binding theory, proteins, 29-30 Ligand-gated channel, see also specific channels channel block, 678, 680 classes, 676 conductance, 676 definition, 441,445 desensitization, 678 developmental changes, 596 gating mechanism, 677-678 kinetics, 678, 679 modulation, 678-679, 680 selective permeability, 677 Lineweaver-Burke plot, 121-122 Lipid bilayer biological processes role, 89-90 caveolae, 86 cholesterol lateral microdomains, 85-86 clathrin-coated pits, 85 desmosomes, 85 formation, 83 gel-to-liquid-crystalline phase transition temperature, 55-59 illustration, 219 lateral lipid microdomains, 85-86 macroscopic domains, 83, 87 microscopic plasma membrane lateral domains, 87 phase transition behavior, 53-61
M Magnetoreceptor species distribution, 840 MAGUK, see Membrane-associated guanylate kinase (MAGUK) protein Malignant hyperthermia, calcium release channel mutation, 664 MARCKS, cellular substrate for protein kinase C, 172-173 Markov process, channel gating, 445 Mechanoreception coupling, 766-767 encoding, 769-770 transduction, 767-769 Mechanoreceptor, see also Sensory receptor arthropod, 764-765 efferent control, 770-771 muscle, 765 vertebrate, 764 Mechanosensitive (MS) ion channel acoustic transduction, 782-783 bacteria, 746 blocking by chemical agents, 768 cloning, 750 delay and adaptation, 753 diversity, 749-752 electrophysiology, 746-747, 755 gating, 617, 752-753, 757-758 ligand activation, 750-751 mechanical quantities, 749 nematode, 746 nonexcitable cells, 491-492 osmotransduction, 755 patch clamp analysis, 748-750, 752, 753 physiological roles, 754-757
Index plant cells, 1091 relation to cytoskeleton, 753 selectivity, 750 stretch-activated, 747, 750 stretch-inactivated, 747, 750, 755 structure, 747, 748 trauma, pathology, and nociception, 752 Medullary thyroid carcinoma, calcium receptors, 187 Melittin mechanism of cell lysis, 34 sequence, 33 structure, 34 Membrane capacitance and resistivity, 220 fluidity, 220 potential profile, 220-221 Membrane-associated guanylate kinase (MAGUK) protein, protein-protein interactions, 610-611 Membrane lipids classification, 43-51 importance, 43 molecular structures, 43-51 structural organization, 51-52 Messenger-activated channel, properties, 445 Metabolic sensor cell, see also specific cell types protection from hypometabolic injury, 736-737 types, 725 Metabolism anabolism versus catabolism, 119 enzyme pathway regulation, 119-120, 123 fat storage, 133-134 starvation-induced changes, 134, 135 synthesis versus degradation, 129-130 tissue-specific regulation, 132-133 mGlu (metabotropic glutamate) receptor postsynaptic density, 607-608 Micelle, definition, 69-70 Michaelis constant, definition, 120-121 Michaelis-Menten equation, 121 Microfilament cross-linking protein, ion channel expression, 611-612 Microsatellite instability, cancer cells, 1164 Microtubule components, 109-113 excitation-secretion coupling, 709 functions, 979 hydrostatic pressure effects, 1010 mitotic spindle formation, 979-981 nonaxonemal, 977 particle and organelle movement, 981-982 polymerization and depolymerization, 977-979 sliding, in ciliary motility, 972-973 structure, 967, 968 Microtubule-associated proteins (MAP) function, 111, 113 postsynaptic membranes, 612 Microvillus, structure, 965 Mineralocorticosteroid, ion transport regulation in stria vascularis, 781 Mitchell chemiosmotic hypothesis generation of energy, 126-128 mitochondrial function, 139-140, 360
1229 Mitochondria anion exchange carriers, 145 apoptosis, 149-150, 1178-1179 ATP synthesis, 141-142 bacterial origin, 1063 calcium cycle, 145-146 cardiac function, 147 chemiosmotic theory, 139-140 cristae, 105 electron transport system, 140-141 function, 105-106 ion leaks, 142-144 membrane structure, 139 oxidative phosphorylation, 105-107, 127 permeability transition, 148-149 pH control, 360 potassium cycle, 146-147 proton cycle and uncoupling proteins, 144 structure, 105 volume homeostasis, 146 Mitogen activated protein (MAP) kinase, signal transduction pathway, 162-163 Mitosis cancer cell, 1161-1163 cell cycle overview, 1136-1141 cell cycle time, 1191 checkpoints, 1193 checkpoints in cell cycle, 1136, 1147 cyclin-dependent kinases, 1137, 1139, 1143-1144 cyclins, 1137-1139, 1142-1143 E2F transcription factors, 1149-1150 exit from mitosis, 1141 G1-S phase transition, 1136-1141 G2-M phase transition, 1141 inhibitors of cyclin-dependent kinase complexes, 1136, 1144-1146, 1148 knockout mice studies, 1146, 1155 phases, 1136 pocket proteins in regulation, 1137, 1140, 1148-1150 proteolysis, 1136, 1141, 1151-1155 radiosensitivity, 1192-1193 START, 1161 Mitotic spindle, formation, 979-981 Monensin, ion selectivity, 13 Multilamellar vesicle, formation, 52 Muscarinic agonist, 568-569 Muscarinic potassium channel, 574 Muscle, see Cardiac muscle; Excitationcontraction (EC) coupling; Muscle contraction; Skeletal muscle; Smooth muscle Muscle contraction afterload, 944 crossbridge theory, 945-946, 947-949, 953, 955-956 energetics, 949-951 force-length relationship, 942, 943,946 interdigitating filament systems, 942-943 metabolic requirements, 951-952 myosin ATPase as energy source, 950-951 phosphocreatine breakdown, 950-952 regulation in smooth muscle, 955-957 skeletal muscle fiber types, 952-953 sliding-filament theory, 943-944, 945
smooth muscle structure-function relationships, 953-955 steady-state modeling, 946-949 thick filament structure, 946, 948 velocity, 944-945 work, 944-945 Muscular dysgenesis, dihydropyridine receptor mutation, 663-664 Myasthenia gravis, acetylcholine receptor dysfunction, 668-669 Myelin sheath, effect on propagation velocity, 401-403 Myoneme calcium effects, 988 contraction speed, 987-988 energetics, 988-989 excitation-contraction coupling, 989-990 linkage complex, 986 optical properties, 988 role of spasmins, 989 ultrastructure, 986 Myosin ATPase in muscle contraction, 946, 950-952 calcium signal transduction, 956.957 content in muscle types, 953 crossbridge interaction with actin, 945-949, 953,955-956 role in amoeboid movement, 962-964 Myotonia chloride current, 868-869 ion channel dysfunction, 654-655,656, 658-663, 868-869 sodium channels, 869
N nAChR, see Nicotinic acetylcholine receptor/channel Narp (neuronal activity-regulated pentraxin), description, 607 Necrosis, compared to apoptosis, 1171-1173 Neonatal severe hyperparathyroidism (NSHPT), calcium receptor mutation, 182-183 Nernst equation derivation, 237 equilibrium potential, 226-227 solute equilibrium, 17 Nernst-Planck equation, 228-229, 548 Nernst-Planck regime, 539 Net diffusion potential, determination, 232 N-ethylmaleimide-sensitive fusion protein, see NSF Neurofilament (NF), NMDA receptors, 612 Neuromuscular junction action potential, 395 formation, 603 nicotinic acetylcholine receptor/channel, 594, 601-602 Neuron, see also Sensory receptor action potential, 594-595 axonal transport regulation, 483 axoplasmic flow, 482-483 classification, 479 components, 479-481 cytoskeleton, 480-482 developmental changes, 594-596
1230 fast retrograde transport, 483 functions, overview, 479 ion channel function, 615 polarity, 479 protein synthesis, 479-480 slow axonal transport, 483 ultrastructural features, 114-116, 479-480 Neuronal apoptosis inhibitory protein (NAIP), apoptosis regulation, 1179 Neurotransmitter defining criteria, 691 inactivation, 693-694 quantal-vesicular hypothesis, 695 receptor types, 692 release, 694-699 storage, 693-694 synthesis, 693-694 types and structures, 691-692 Neurotransmitter-gated channels, s e e a l s o s p e cific c h a n n e l s
mutations and disease, 668-669 Nicotinic acetylcholine receptor/channel (nAChR) developmental changes, 594 domain-specific clustering, 601-602 effect of a-conotoxin, 639 role of tyrosine kinase in clustering, 612-613 Nigericin, ion selectivity, 13 Nitric oxide (NO) effects on calcium flux, 567-568 synthases, 568 Nitrogen fixation bacteria, 1069 NMDA (N-methyl-D-aspartate) receptor/channel neurofilaments, interaction with, 612 neuronal, 615 postsynaptic density, 604-606 properties, 684-686 rundown, 615 Yotaio interaction with neurofilaments, 612 N-methyl-D-aspartate, see NMDA Nonselective cation channel ATP activation, 490 calcium activation, 490 cyclic nucleotide activation, 490-491 Noxiustoxin, effect on potassium channel, 636 NSF (N-ethylmaleimide-sensitive fusion protein), description, 606-607 Nuclear envelope cytochalasin effect on permeability, 555 electrical and diffusional forces, 547-552 freeze fracture, 548 modulation of permeability by ATP, 552 permeability, 539-540 relationship to endoplasmic reticulum, 543 structure, 540-541 Nuclear magnetic resonance spectroscopy (NMR), protein structure determination, 27-28 Nuclear Overhauser effect (NOE), 28 Nuclear pore channel between nucleoplasm and cytoplasm, 539-540 compared to gap junction, 544 cytoskeletal interaction with ionic flux, 552-555
Index structure, 541-543 Nuclear pore complex central granule, 540 channel model, 551 definition, 540 macromolecular assemblies, similarity to other, 556 Nuclear receptor contact with transcription apparatus, 196-197 dimerization, 196 DNA-binding domain, 194 DNA-binding modes, 196 DNA-binding motifs, 195 gene family, 192-193 inhibitory protein dissociation, 196 ligand-activated receptors, 192 ligand-binding, 194-195, 196-197 orphan receptors, 192 structural domains, 193-195 Nucleoplasmin, enabling large gold particles to enter nucleus, 540 Nucleotide, cyclic, see Cyclic AMP, Cyclic GMP Nucleus cell within cell, 551 electron microscopy, 540-541 electrophysiology, 543-545 isolation, 540 measurement of single channels, 545 osmometer, 551 osmotic effects, 545-547 pH, 361 resting potential, 544-545 structure, 98-101,543 Nystatin, ion selectivity, 13
O Ohm's law, equation, 1205 Olfaction, cyclic nucleotide-gated ion channels, 796, 798-800 Olfactory receptor adaptation, 827-829 calcium-dependent chloride channels, 815 calcium-dependent potassium channels, 826-827 chemical stimuli, 824-825 cyclic AMP and cyclic nucleotide-gated channels, 825-826 G proteins in transduction, 825 initiation of receptor potentials, 825-827 inositol trisphosphate in transduction, 826-827 membrane-bound receptors, 825 olfactory signals, termination of, 827-829 pheromone receptors, 824 signal transmission, 815 stem cells, 824 structure, 816 vomeronasal organ, 824, 825, 827 Orphan receptor, identification, 192 Osmosis definition, 319-320 diffusional permeability of porous membranes, 326 diffusional permeability through lipid membranes, 331
discovery, 320 dissolution-diffusion model calculation, 331 filtration permeability, 326 flow mechanism models, 325-329 lipid bilayer membranes, dissolution-diffusion model, 329-331 narrow pores, water flow, 331-332 pressure-driven flow, dissolution-diffusion model, 330-331 solute flux calculation, 328 water transport across membranes, 333-336 Osmotic coefficient nonideal solutions, 322-323 sucrose concentration effects, 323 Osmotic pressure Boyle-van't Hoff law, 383 definition, 320 effective osmotic pressure, 320 hemoglobin dependence, 324 hydrostatic pressure, 324, 325 measurement, 321-325 nonideal solutions, 322-324 physical origin, 326-329 reflection coeffiecient, 324-325 van't Hoff's law, 321-325 Oxidative phosphorylation Mitchell's chemiosmotic hypothesis, 127 mitochondria, 105-107 Oxygen effect, ionizing radiation damage enhancement, 1195-1196
P Pacemaker current, cardiomyocyte, 590--591 Pacemaker potential, cardiomyocyte, 585 Palytoxin, membrane transporter in red blood cells, 381 Pancreatic islet cell, s e e a l s o ce Cell,/3 Cell, stimulus-secretion coupling, 725-736 Parafollicular cells, calcium functions, 186-187 Paramyotonia, sodium channel dysfunction, 661-662 Parathyroid cell, calcium sensing, 182-187 Parathyroid hormone regulation of calcium, 185-186 secretion, 185-186 Passive transport, definition, 12 Patch clamp capacitance measurements, 451-452 cell-attached patch, 426, 443 circuitry, 442 correlation analysis, gating mechanism, 448 dwell-time distributions in single-channel analysis, 447 electrocytes, 1033-1034 gap junctions, 528 gentle, 753 high pressure, 1013-1014 inside-out patch, 426, 444 mechanosensitive ion channels, 748-750, 752, 753 membrane noise analysis, 451 noise, sources and minimization, 442-443 nonstationary single-channel analysis, 448-449 nuclei, 541,545
Index outside-out patch, 426, 444 perforated patch technique, 450 plant cells, 1089 single-channel analysis, 444--449 tight-seal technique, 449-452 transition rates in single-channel analysis, 444-445 two-time distributions in single-channel analysis, 448 whole-cell technique, 449-452 PDZ domains, protein-protein interactions, 610-611 Pentose shunt, red blood cell metabolism, 379 Permeability calculation, 213 coefficient calculation, 213 membrane transport overview, 209-210 Peroxisome, photorespiration, 1101 pH, see also Intracellular pH buffering power of solutions, 357-359 electrodes for measurement, 373-374 Henderson-Hasselbalch equation, 357 measurement using biscarboxyethyl carboxyfluorescein, 374-375 Phosphatase, ventricular cells, 563 Phosphatidylcholine (PC), structure, 46--47 Phosphatidylethanolamine (PE), structure, 47 Phosphatidylglycerol (PG), structure, 48-49 Phosphatidylinositol (PI), structure, 47-48 Phosphatidylserine (PS), structure, 49 Phosphocreatine, muscle energy storage, 950-952 Phospholamban phosphorylation, 273-276 regulation of calcium-ATPase, 273-277 structure, 273-275 Phospholipase A2, modulation of exocytosis, 718 Phospholipase C, stimulation, 169 Phospholipase D, signal transduction role, 161-162 Phospholipid, see Glycerophospholipid Phospholipid transbilayer distribution mechanisms, 84-85 structure of domains, 83-84 Phosphorylation cell volume regulation, 350 role in ion channel clustering, 612-613 Photoperception plant cells, 1090-1091 Protista, 1047-1049 Photoreceptor, see Visual transduction Photosynthesis, see also Chloroplast ATP formation, 1102-1106 bacteria, 1068 Calvin cycle, 1099-1101 chemiosmotic hypothesis of ATP synthesis, 1102-1104 electron donors, 1107-1108 electron transport, 1106--1111 Emerson enhancement, 1106-1107 evolution, 1098 free energy change, 1097, 1104-1105 microorganisms, 1097-1098 photorespiration and C4 plants, 1101 photosynthetic units, 1108 quantum yield, 1106
1231 reaction center structure, 1108-1109 regulation, 1111-1112 species distribution, 1097-1098 structures of molecules and cofactors, 1103 Pituitary hormone, secretion, 719 Plant cell, see also Photosynthesis calcium transport, 1089-1090 carrier-mediated transport, 1089 cell wall structure and functions, 1080-1083 communication with other organisms, 1087-1088 comparison to animal cells, 1079 cytoskeleton, 1085 defensive signals, 1086-1087 Golgi apparatus, 1084 gravity perception, 1091 hormones, 1086, 1087 ion channels, 1089 lipid bodies, 1085 mechanosensory mechanisms, 1091 membrane potentials, 1089 meristems, 1079-1080 microbodies, 1085 photoperception, 1090-1091 plasma membrane, 1084 plasmodesmata, 1085-1086 plastids, 1085 proton-ATPases, 1088 receptor proteins for hormones, 1091-1092 signal transduction, 1092-1093 stomata, 1083 vacuoles, 1084-1085 Plasma membrane calcium-ATPase, 279-280 chloride distribution, 302 electron microscopy, 541 structure, 65-73, 81 Plasmodesmata intercellular transport, 1086 photorespiration role, 1101 structure, 1085-1086 Plastocyanin, electron transport, 1111 Pocket proteins, cell cycle regulation, 1137, 1140, 1148-1150 Polypeptide potassium channel blockers, 633 proteins in membranes, 32-4 1 Positional cloning, ion channel gene mutations, 654 Post-Albers reaction mechanism, sodium pump, 262-264 Postsynaptic density (PSD) glutamate receptor/channels, 604-608 potassium channels, 608-6 10 Postsynaptic ion channel/receptor matrix, role of gephyrin, 604 Potassium, secretion and transport in cochlea, 776-780 Potassium channel acoustic transduction, 784-786 activation and inactivation, action potential generation, 432-434 amino acid sequence, 577 ATP-regulated, 436, 729-732, 870-872 calcium-activated, 436, 632-637, 769, 784-785,902
cardiac muscle, 589-590, 610, 616, 736-737 cell volume regulation, 501-503 coupling of subunits to G protein, 574-577 crystal structure of bacterial pore region, 473-474 delayed rectifier, 435,593-594, 769, 872 dysfunction and disease, 666-668 epithelial, 615 hypoxia, 736-737, 740-741 inactivation mechanism, 468-469 insulin secretion, 729-732, 736-737 inward rectifying, 436, 488-489, 589, 592-593,608-610, 769 modulation by bisphosphate and sodium ions, 577 modulation by G proteins, 573-574, 575 modulation by RGS proteins, 578 muscarinic, 574 neuronal G protein-gated, 610 nonexcitable cells, 487 nucleic acid clones coding, 465 outward rectifying, 435 oxygen-sensitive, 738-741 plant cells, 1089 polypeptide blockers, 633 pore formation, 471-473 PSD-95-related proteins, 608 regulation by cytoskeleton, 615 S h a k e r mutants of D r o s o p h i l a , 465-472, 474-475,666 signature sequence, 471 skeletal muscle, 592-593,593-594, 736-737, 866, 868, 870-872 smooth muscle, 901-903,908 sperm, 510-511,513,515,516-518 subfamily properties, 474 targets for toxins, 632-637 taste transduction, 819-821 transient outward current, 436 types, 435-436 voltage-dependent, 466, 468, 487-488, 589-590, 608, 632-637 Potassium-chloride cotransporters basic features, 314 molecular structure, distribution, and functions, 314-315 overview, 313-314 thermodynamics, 315 Potassium current ATP-sensitive, 893-894 delayed rectifier, 891-892 inward rectifier, 892-893 Potassium proton exchanger, intracellular pH regulation, 364 Profilin, role in actin-based motility, 960 Prokaryote physiology, see Bacteria Protein, see also specific p r o t e i n s circular dichroism analysis, 25, 27 conformation, 24-25 nuclear magnetic resonance to determine structure, 27-28 prediction of structures, 25-28 primary amino acid structure, 19 secondary structure, 19-22 structures, geometry and dihedral angles, 24 tertiary structure, 21-23
1232 Protein (continued) titration, 29-31 X-ray crystallography to determine structure, 27 Protein kinase A effect on calcium channels, 562-563 synaptic channel density, 613 Protein kinase A pathway, cyclic AMP concentration, 156-157 Protein kinase C calcium mediation pathway, 172-174, 569 control of parathyroid hormone secretion, 185-186 modulation of exocytosis, 718 regulation, 175 Protein tyrosine phosphatase, in signal transduction, 164 Protista apicoplastids, 1054 avoidance reaction, 1049-1051, 1057 bacterial endosymbionts and xenosomes, 1052-1054 behavioral mutants, 1051 calcium effects on behavior, 1049-1051 classification, 1041-1042 contractile vacuoles, 1044-1045 crystalline bodies, 1056 defecation, 1044 digestion, 1043-1044 diversity, 1058-1059 endocytosis, 1043-1044 energetic adaptations, 1045-1047 extrusomes, 1042, 1055-1056 filter-feeding, 1042-1043 glycosomes, 1045-1046 gravity receptors in ciliates, 1049 hydrogenosomes, 1045, 1046-1047 intracellular lenses in dinoflagellates, 1048-1049 ion channel types, 1051 kinetoplast DNA, 1058 light antennae, 1047-1048 organelle capture and use, 1041, 1052-1054 oxygen gradient response, 1056-1058 paraflagellar rod, 1054-1055 photoreceptors, 1047-1049 rhoptries, 1054 second messengers and transduction pathways, 1051-1052 sensory adaptations, 1047-1052, 1056-1058 sensory transduction, 1049-1052 Proton channels in pH maintenance, 362 conductance, 359 Proton-ATPase intracellular pH regulation, 365 plant cells, 1088 types, 365 Protonmotive force (PMF) bacteria, 1069-1070 equation, 127, 139-140 photosynthesis, 1105 Proton pump, see Proton-ATPase Protozoan, gliding behavior, 998 PSD, see Postsynaptic density Pump-leak hypothesis, cell volume regulation, 338-339
Index Purkinje fiber action potential, 895 automaticity, 895-896 Push-pull model, vectorial transport, 505-506 Pyruvate kinase, regulation, 130
13 Q cycle, proton translocation, 1111 Quantum energy, 1106 yield in photosynthesis, 1106
R Rapsyn, ion channel clustering, 601-602 ras-p21 protein, amino acid substitutions effects, 38-41 Red blood cell anion exchange and conductance, 386-388 cell volume regulation, 383-386 cytoskeleton, 377-378 cytotoxic calcium cascade, 388-389 double-Donnan equilibrium, 380 echinocyte form, 377 erythrophagocytosis, 379 Gibbs-Donnan equilibrium, 378-379, 384-386 hemoglobin content, 378 inhibitors of membrane transporters, 382-383 intracellular environment, 378-379 intracellular mean ionic activity, 378-379 ionic and osmotic equilibrium, 383-386 ionophore studies, 387-388 life span, 379 membrane composition, 377-378 membrane transporters, 380-383 metabolism, 379 morphology, 377-378 production coupled to oxygen sensing, 741-742 pump-leak theory, 380 resealed ghosts, 378 resting membrane potential, 381-382 stomatocyte form, 377 Reductive pentose phosphate cycle, see Calvin cycle Reflection coefficient definition, 324-325 physical interpretation, 329 Regulator of G protein signaling (RGS protein), 578 Regulatory volume decrease (RVD) anisosmotic media composition in signaling, 347 calcium signaling, 349 cytoskeleton in signaling, 347-349 definition, 342-343 intracellular pH effects, 370 macromolecular crowding role, 351-352 mass action model, 350-351 membrane potential in signaling, 347 nonexcitable cells, 501-503 phosphorylation signaling, 350 red blood cells, 384 transport processes, 343-346
Regulatory volume increase (RVI) anisosmotic media composition in signaling, 347 calcium signaling, 349 cytoskeleton in signaling, 347-349 definition, 342-343 intracellular pH effects, 370 macromolecular crowding role, 351-352 mass action model, 350-351 membrane potential in signaling, 347 nonexcitable cells, 501-503 phosphorylation signaling, 350 red blood cells, 384 transport processes, 343-346 Renal epithelial cells, calcium receptors, 187-188 Resin microelectrode, measurement of pH, 374 Resistance axoplasm and myoplasm, 1213-1214 cell membrane, 1214 Ohm's.law, 1205 resistors in parallel and in series, 1206-1207 specific resistance, 1213 Respiration neural control, 737-740 rate in cells, 140-141 Resting membrane potential, see also Equilibrium potential calcium distribution, 224-226 cardiac, 585, 887-888 chloride distribution, 223-224 chord conductance equation, 231-232 circuit analysis, 240-241 constant-field equation, 230-231,237-239 determining factors, 230 effects on action potential, 436 electrochemical driving forces, 229-230 ion distributions, 220-226 measurement, 221 membrane ionic currents, 230 Nemst-Planck equation, 228-229 skeletal muscle developmental changes, 591-592 sodium-potassium pump effects, 222-223 values for various cell types, 222 Reverse cholesterol transport, process, 86 RGS protein, see Regulator of G protein signaling Rhodopsin, activation and shutoff, 812-813 Ryanodine effects on calcium release from sarcoplasmic reticulum, 935-936 toxicity, 929 Ryanodine receptor (RyR) activation, 911-912 calcium binding sites, 937-938 calcium release, 918, 919, 920, 922-924, 927-938 cloning and expression, 930-932 functional regions, 936-938 genetic studies, 920-922 high-conductance ligand-gated channel, 932-936 interaction with dihydropyridine receptor, 917-924, 936-937 in vitro studies, 922
1233
Index studies, 922-923 isoforms, 918 isolation, 929-930 single-channel planar lipid bilayer measurements, 933-935 structure, 917,929-930 voltage sensing, 919-920 in v i v o
S Sarcoplasmic reticulum (SR) calcium-ATPase, 271-278 calcium release, 927-938 communication with transverse tubules, 880-882 fractionation studies, 929, 930 Sarcotubular system, 878 Saxitoxin, binding site, 626-627 Scatchard plot, 30, 130-131 Scorpion toxin sodium channel effects, 629-632 structure, 630 Scyllatoxin, effect on potassium channel, 636-637 Sea anemone toxin sodium channel effects, 629-632 structure, 631 Self-referencing electrode, description and applications, 792-793 Sensor cell, metabolic, 725-742 Sensory receptor adaptation, 762-763,769 arthropod, 764-765 biogenic amine modulation, 770 cilia, 977 coupling, 766-767 discrimination of stimuli, 761-762 efferent control, 770-771 encoding, 769-770 information capacity, 763 information transmission, 763-764 internal, 725 measurement, 765-766 muscle, 765 noise analysis, 768-769 satiety, 731 single-channel conductance estimates, 767 taste, 731 temperature sensitivity, 769 transduction, 761-762, 767-769 types, 761-762 vertebrate, 764 Shear-stress activated channel, endothelial cells, 491-492 Signal transducers and activators of transcription, cytokine receptor signal transduction, 202-203 Signal transduction bacteria, 995-996, 1071-1072 cyclic AMP, 153, 155-157 cyclic GMP, 157-158 G proteins, 153-160 health and disease effects, 204-205 heptahelical membrane receptors, 198-200 intracellular receptors, 191-192 Jak-Stat pathway, 202-203
lipophilic hormones via intracellular receptors, 191-197 myosin light-chain phosphorylation, 956-957 phosphatidic acid as second messenger, 161-162 phosphatidylinositol cycle, 159-160 plasma membrane, 197-198 protein hormones via plasma membrane receptors, 197-204 protein tyrosine kinases and phosphatases, 204 receptor phosphorylation signaling, 162-163 sensory receptors, 767-769 serine/threonine kinases, 204 signal amplification, 156-157 sphingolipids as second messengers, 160-161 triad junctions, 914-919 trimeric GTP-binding proteins Single-channel analysis gated channel families, 445 nonstationary analysis, 448-449 random nature of channel gating, 445-448 sensory receptor estimates, 767 Sinoatrial (SA) node action potential, 894 automaticity, 896 Skeletal muscle action potential, 435, 591-592, 865-883 activation speed, 913-914 afterpotentials, 872-873 calcium channels, 593, 866-867, 874 calcium-dependent slow action potentials, 873-875 chloride channels, 868-870 conduction down fiber, 876-878 conduction into fiber, 878-882 developmental changes, 591-593,875 electromyography, 877 excitation-contraction coupling, 911-925,928 potassium channels, 592-594, 736-737, 866, 868, 870-872 repolarization, 435, 867-869 resting potential, 591-592 ryanodine receptor functional regions, 936-938 slow fibers, 876 sodium channel, 593,660-663, 866, 868, 870-872 sodium-potassium pump, 875-876 types, 952-953 Sliding-filament theory, s e e Muscle contraction Small unilamellar vesicles, formation, 52 Smooth muscle action potential, 407, 899-909 calcium channels, 901,903-904, 906, 908 chloride channels, 904 classification, 900 depolarization, 901 electrical properties, 899 hyperpolarization, 901 ionic currents, 906-907, 908 membrane potential determinants, 901 morphology in relaxation and contraction, 954 neurotransmitter effects, 905 pacemaker activity, 908-909 potassium channels, 901-903,908
receptors, 904-907 regulation of contractility, 955-957 resting membrane potential, 907-908 state of activation, 900-901 structure and function, 953-957 vascular, 740-741 Sodium-bicarbonate-chloride exchanger, intracellular pH regulation, 364-365 Sodium-bicarbonate cotransporter, intracellular pH regulation, 365 Sodium-calcium exchange, s e e a l s o Calciumsodium exchange cardiac action potential current, 295-296, 893 carnivore erythrocytes, 341 energetics, 284-285 excitation-contraction coupling, 296-297 giant patches for measurement, 287 ionic dependencies of current, 289-290 isoforms, 284 isolation of exchange current, 288-289 mechanism, 294-295 mutations, 289 phosphatidylinositol-4,5-bisphosphate, 292 phosphorylation, 292 Ping-Pong (consecutive) mechanism, 294-295 problems with measurement of current, 285-288 proteins in exchange family, 284 reaction, 224-226 regulation of exchange current, 290-292 reversal potential, 289 sequential (simultaneous) mechanism, 294-295 sodium-dependent inactivation, 290 stoichiometry, 285 structure of exchanger, 283-284 voltage dependence of current, 292-294 whole-cell patch clamp measurement, 285-286, 288-289 Sodium channel amiloride-sensitive, 485-487, 613-614 batrachotoxin effects, 425,627-629 cardiac, 586, 616, 647-649, 663 density in selected tissues, 456 dysfunction and disease, 661-663 electric field model, fast channel density~ 409 electrocytes, 1032-1034 epithelial, 485-487, 613-615 epithelial transport disorders, 663 fast, 428--429 gating-related domains, 470-471 grayanotoxin, 627-629 high pressure and squid giant axon, 1014 hydropathy analysis, 460, 462-463 inactivation of fast channels, action potential generation, 430-434 ligand-gated, 596 mutations and epilepsy, 663 mutation-specific drug actions, 647-649 myotonia, 656, 660-663 node of Ranvier, 602-604 neuron developmental changes, 595-596 pharmacological inactivation, 647-648 pharmacology, 647-649 pore formation, 471-473 purification, 455-458
1234 Sodium channel (continued) recovery process following activation, 432 resting, active, and inactive states, 430, 434 scorpion toxin, 629-632 sea anemone toxin, 629-632 selectivity filter, 626 sequence homology of transmembrane domains, 460--461,463 skeletal muscle, 593,660-663, 866, 868, 870-872 structure, 433,457 target for toxins, 625-632 taste transduction, 818-819 tetrodotoxin-sensitive, 424--426, 432, 769 topographical structure, 463-464 veratridine, 627-629 voltage-dependent activation domains, 466, 468 voltage-gated, 595-596 voltage-sensitive, 625-632 Sodium-chloride cotransporter, electroneutral cotransport, 315 Sodium current cardiac action potential, 888-889 window, 431,890 Sodium-organic anion transporter, intracellular pH regulation, 365-366 Sodium,potassium-ATPase, see also Sodiumpotassium pump cardiac glycoside binding site, 266-268 cation binding sites, 266 electrocytes, 1034 isoforms, 264-265 mechanism, 261 red blood cells, 380-382 smooth muscle, 901 subunits and structure, 261,263-267 Sodium-potassium-chloride cotransporter cell volume regulation, 340-341 electroneutral cotransporters, 303-313 heart, 340-341 kinetic model, 311 molecular structure and distribution, 308-311 physiological functions, 311-313 properties, 304-308 Sodium-potassium pump, see also Sodium,potassium-ATPase activity, 261 cardiac action potential, 893 cardiac glycoside effects, 266-268 coupling ratio, 235 density on cell surface, 234 electrogenic pump potentials, 232-235 inhibition by cardiac glycosides, 262 intracellular ATE 262 ion selectivity, 262 Post-Albers reaction mechanism, 262-264 red blood cells, 380-382 regulation mechanisms, 261-262 regulation of cell volume, 243 skeletal muscle, 875-876 transport mechanism, 262-263 turnover rate, 235 visual transduction, 810 Sodium-proton exchanger intracellular pH regulation, 362-364 structure, 363
Index Solute transport, definition, 12 Solutions, colligative properties, 322 Spasmin comparison to other motility proteins, 960 role in spasmonemes, 989 Spasmoneme calcium effects, 988,989 contraction speed, 987-988 energetics, 988-989 excitation-contraction coupling, 989-990 linkage complex, 986 optical properties, 988 role of spasmins, 989 ultrastructure, 986 Sperm acrosome, reaction, 509-513, 513-515 calcium channels, 51 0-519 capacitation, 513 egg peptide responses, 510 mammalian model, 513-515, 517-519 potassium channels, 510-511,513,515, 516-518 sea urchin model, 509-513, 515-517 zona pellucida, acrosome reaction, 513-515 Sphingolipid, structure, 49-50 Sphingolipid and glycolipid transbilayer domains, structure and function, 84 Starling hypothesis, 246 Starvation, metabolic adaptation, 134, 135 Steroid hormones, secretion mechanisms, 719-721 Stimulus-contraction coupling, vascular smooth muscle cells, 740-741 Stimulus-response coupling, metabolic sensor cells, 725-742 Stimulus-secretion coupling a cells, 735-736 /3 cells, 725-735 Strength-duration curve, 421 Stretch-activated channel cell volume regulation, 501-502 endothelial cells, 491-492 properties, 445 Stria vascularis basal cells, 777, 778-779 hormone receptors, 780 ion transport model, 776 ion transport regulation, 779-781 marginal cells, 777-778 potassium secretion, 776-781 Sulfonolipid, role in bacterial gliding, 997 Sulfonylurea, effect on insulin secretion, 731 Synapse, see also Gap junction automodulation, 703 calcium role in depolarization-release coupling, 695-696 chemical neurotransmission overview, 689-691 classification, 689 depression of transmission, 702 excitatory postsynaptic potential, 691,701 facilitation of transmission, 702-703 heterosynaptic modulation, 703 inhibitory postsynaptic potential, 691 integration versus amplification, 702 long-term potentiation, 703
neurotransmitter storage, 693-694 postsynaptic potential generation, 699-700 postsynaptic potentials, 700-701 post-tetanic potentiation, 703 presynaptic inhibition, 703 receptor types, 690 reversal potential, 700 slow transmission mediated by G protein-coupled receptors, 701-702 structure, 689-691 vesicle docking-fusion complex, 697-698 vesicle exocytosis, 696--699 Synaptic transmission, 689-703
T Taste receptor apical chemosensitive membrane, 817 chemical stimuli categories, 818 initiation of receptor potentials, 818-822 ion channel blocking, 818-821 passive ion permeation through channels, 818-821 receptor-ligand binding, 818-821 signal transmission, 815, 817-824 stem- cells, 817 structure, 816 taste buds, 815-817 taste transduction, 822 termination of signals, 822, 824 voltage-dependent ionic conductances, 817 Tetrodotoxin (TTX) binding site, 626-627 sodium channel blocking, 424-426, 432, 586, 647, 769 Thigmomorphogenesis, plant cells, 1091 Thioredoxin, regulation of photosynthesis, 1112 Thomas recessed-tip microelectrode, measurement of pH, 373 Time constant, calculation for membranes, 1211-1213 Titration, proteins, 29-31 Tolbutamide, effect on insulin secretion, 731 Trace elements, cofactors, 11-12 Transducin, visual transduction, 157-158 Transepithelial measurement, sensory receptors, 766 Transgenic mouse, see Knockout mouse Transmembrane electrical potential, measurement, 17 Transmembrane protein, function, 34-35 Trans to g a u c h e isomerizations lipid bilayers, 53-55 rotamers, 54 Transverse tubule calcium release, 927 communication with sarcoplasmic reticulum, 880-882 length constant, 880 skeletal muscle fibers, 878-882 triad junctions, 91 4-919 Triacylglycerol, storage, 128 Triad junction properties, 915-919 skeletal muscle excitation, 91 4-919 T-Tubule, see Transverse tubule
Index Tuberous electroreceptor anatomy, 846-849, 850-852 electric organ association and stimuli, 840, 846, 847, 849-850, 852 equivalent circuit, 849 frequency response, 840, 849, 851 gymnarchids, 848, 852 gymnotids, 846-849 Knollenorgans (K units), 850-851 mormyrids, 848, 850-852 mormyromasts, 851-852 physiology, 849-850, 852 species distribution, 839-840, 846 Tubulin comparison to other motility proteins, 960 neuronal cytoskeleton, 481-482 pH changes, 368 Tumor, s e e Cancer cell Tumor necrosis factor receptor (TNFR) apoptosis, 1173-1174, 1179 Tyrosine kinase activity and hormone receptors, 201-202 clustering of nicotinic acetylcholine receptors, 612-613 signal transduction, 162 Tyrosine phosphorylation hormone receptors, 200-201 kinases in signal transduction, 200-201 phosphatases in signal transduction, 204
IJ Ubiquitin, proteolysis targeting in cell cycle control, 1136, 1141, 1151-1155 Ultrastructure, defined, 95 Unidirectional influx and effiux, definitions, 213-214 Unit membrane, trilamilar structure, 67 Up-regulating, metabolic enzymes, 129 Ussing flux ratio equation, 215-216
V Vacuole, plant cells, 1084-1085 Valinomycin, ion selectivity, 13
1235 van't Hoff's law derivation, 321-325 nonideal solutions, 322-324 osmotic pressure calculation, 321-325 Vectorial transport, of nutrients, 128 Veratridine, sodium channel binding, 627-629 Vesicle docking-fusion, proteins in docking and fusion, 696-698,709-711 Vesicle fusion capacitance flickering, 713-714 capacitance jump, 712-713 fusion pore properties, 714-715 kinetics, 712 membrane tension as a driving force, 715 steps in fusion, 715 Visual transduction cone properties, 807-808 dark current and photoreceptor light response, 808-810 cyclic nucleotide cascade, 812-813 cyclic nucleotide-gated ion channels, 795-796, 798, 800-801,808-809 ion transport and resting potential, 808-810 photopigment activation and shutoff, 81 0-812 photoreceptor adaptation, 810 photoreceptor structure, 807-808 rod properties, 807-808 suction electrode measurement, 809 Voltage clamp, s e e a l s o Patch clamp action potential generation analysis, 424-427 ion-conducting system analysis, 450-451 sensory receptor, 766 whole-cell voltage clamp, 426-427 Voltage-gated channel definition, 441,445 expression systems, 466 gating-related domains, 470-471 hydropathy analysis, 460, 462-463 inactivation mechanism, 468-469 interaction with G proteins, 579-580 isoforms, 474-475 large polypeptide channels, 460, 462 mutagenesis studies, 466-470 pore formation, 471-473 purification, overview, 455-458 reconstitution, 458
sequence homology of transmembrane domains, 460-461,463 sequencing, 458-465 structure investigation, 458-465 subunit composition, 458 topographical structure, 463-464 voltage-dependent activation domains, 466, 468 Volume-regulated anion channel functions, 500-501 gating model, 501 Vomeronasal organ pheromone receptors, 824, 825, 827
W Water cellular content, 3 channels in biological membranes, 334-336 dielectric constant, 4 dipole moment, 4 hydrogen bond strength, 3-5 ion hydration, 5-7 physical properties, 3-5 transport across cell membranes, 319-325, 332-336 viscosity, 5 Window current, sodium channels, 586
X Xenosome, endosymbiotic bacteria in Protista, 1052-1054 X-ray crystallography, protein structure determination, 27
u Yotiao, interaction with NMDA receptors, 606
Z Zeta potential, cell membrane, 221 Zona pellucida, sperm acrosome reaction, 515-515