Membrane Receptors (Current Topics in Membranes and Transport ; V. 18)

Current Topics in Membranes and Transport Volume 18 MEMBRANE RECEPTORS Advisory Board M . I? Blaustein A . Essig ...

Author: Arnost Kleinzeller

23 downloads 685 Views 25MB 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

DOWNLOAD PDF

Current Topics in Membranes and Transport

Volume 18

MEMBRANE RECEPTORS

Advisory Board

M . I? Blaustein A . Essig

R . K . H . Kinne I? A. Knauf Sir H . L. Kornberg

P. Lauger C. A . Pasternak W. D. Stein W. Stoeckenius K. J . Ullrich

Contributors

Michael C. Lin Suzanne K. Beckner Robert Dale Brown B. Richurd Martin Steen E. Pedersen Brian A. Cooke Dermot M . E Cooper John I? Perkins W. D. Rees Manjitsri Das Elliott M . Ross Vincent A. Florio J . Gliemann Paul Schlesinger Larry M . Gordon Virginia Shepherd Jeffrey M . Stadel Miles D. Hoidslay Serge Jard Philip Stahl Duvid A . Johnson Palmer Taylor Robert J . Leflowitz A . M . Tolkovsky Simon van Heyningen

Current Topics in Membranes and Transport Edited by

Arnost Kleinzeller Department of Physiology University of Pennsylvania Philadelphiu. Pennsylvania

Volume 18

MEMBRANE RECEPTORS Guest Editors

Arnost Kleinzeller

B. Richard Martin

Department of Physiology University of Pennsylvania Philadelphia, Pennsylvania

Department of Biochemistry Uniwrsitv of Cumbridge Cumbridge. England

1983

@

ACADEMIC PRESS

A Siihlidion of Hon owi Bruce J o ~ u n oiii h Pihlidirr,

New York London Paris San Diego San Francisco Slo Paulo Sydney Tokyo Toronto

COPYRIGHT @ 1983, BY ACADEMIC PRESS,INC. ALL RIGHTS RESERVED. NO PART OF THIS PUBLICATION MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM OR BY ANY MEANS, ELECTRONIC OR MECHANICAL, INCLUDING PHOTOCOPY, RECORDING, OR ANY INFORMATION STORAGE AND RETRIEVAL SYSTEM, WITHOUT PERMISSION IN WRITING FROM THE PUBLISHER.

ACADEMIC PRESS,INC.

1 1 1 Fifth Avenue, New York, New York 10003

United Kingdom Editiott pubtisfzed by ACADEMIC PRESS, INC. ( L O N D O N ) LTD. 24/28 Oval Road. London NWl7DX

LIBRARY OF CONGRESS CATALOG CARD NUMBER: 70-1 17091 ISBN 0-12-153318-2 PRINTED IN THE UNITED STATES OF AMERICA 83 84 85 86

9 8 1 6 5 4 3 2 1

Contents Contributors, xi Preface, xiii Yale Membrane Transport Processes Volumes, xv Contents of Previous Volumes, xvii

PART I .

ADENYLATE CYCLASE-RELATED RECEPTORS

Hormone Receptors and the Adenylate Cyclase System: Historical Overview

B. RICHARD MARTIN Text. 3 References. 8

The Elucidation of Some Aspects of Receptor Function by the Use of a Kinetic Approach A. M. TOLKOVSKY

I. Introduction, II 11. Signal-Response Coupling and Receptor Theory, 13 111. Applying Kinetic Theory to Data Generated by Turkey Erythrocyte Adenylate Cyclase. 22 IV. Conclusions, 40 References. 43

The p-Adrenergic Receptor: Ligand Binding Studies Illuminate the Mechanism of Receptor-Adenylate Cyclase Coupling

JEFFREY M . STADEL and ROBERT J. LEFKOWITZ

I. 11. Ill. IV.

Introduction. 45 Development of Radioligands Specific for Adrenergic Receptors, 47 Study of Adrenergic Receptors in Membranes. 49 Characterization of Detergent-Solubilized Adrenergic Receptors, 57 References. 63 V

vi

CONTENTS

Receptor-Mediated Stimulation and Inhibition of Adenylate Cyclase

DERMOT M. F. COOPER I. 11. 111. IV.

V. VI. VII. VIIl. IX. X.

Introduction. 67 Stimulation of Adenylate Cyclase, 68 GTP-Dependent Inhibition of Adenylate Cyclase. 70 Bimodally Regulated Adenylate Cyclase Systems, 71 Receptor Binding of Inhibitory Ligands, 73 The Role of GTP Hydrolysis in Inhibition of Adenylate Cyclase, 75 The Relationship between N, and N,, 76 Structural Studies on Dually Regulated Adenylate Cyclase Systems, 77 Future Directions. 78 Conclusion, 81 References, 8 I

Desensitization of the Response of Adenylate Cyclase to Catecholamines

JOHN I? PERKINS 1. Introduction. 85 11. Scope of the Review, 87 111. Catecholamine-Induced Desensitization of Intact Cells, 88 IV. Catecholamine-Induced Changes in Adenylate Cyclase and in PAR Binding Properties, 93 V. Separation of Native and Desensitized PAR, 95 VI. Receptor Endocytosis as a Mechanism for Agonist-Induced Desensitization, 97 V11. A Kinetic Model for Agonist-Induced Desensitization, 98 VIII. Differential Expression of PAR during Growth of 1321NI Cells, 100 IX. Down-Regulation of PAR and the Recovery of Lost Receptors. 100 X. Isoproterenol-Induced Changes in Agonist Binding Properties of Intact 1321N1 Cells, 103 XI. Conclusions, 104 References. 106

Hormone-Sensitive Adenylate Cyclase: Identity, Function, and Regulation of the Protein Components

ELLIOTT M. ROSS, STEEN E. PEDERSEN, and VINCENT A. FLORIO

I. 11. 111. IV.

Overview, 109 Protein Components of Hormone-Sensitive Adenylate Cyclase, I 10 Protein-Protein Interactions and the Regulation of Adenylate Cyclase, 127 Assessment of Progress, 137 References, 137

CONTENTS

The Regulation of Adenylate Cyclase by Glycoprotein Hormones

BRIAN A. COOKE

I. Introduction. 143 11. Nature of the Hormones. 144

111. Nature of the Receptors, 145 IV. Involvement of Cyclic AMP in Hormone Action. 149 V. Important Features of the Hormone Receptor-Adenylate Cyclase System, 151 VI. Desensitization and Down-Regulation by Homologous Hormone, 152 References. 172

The Activity of Adenylate Cyclase Is Regulated by the Nature of Its Lipid Environment

MILES D. HOUSLAY and LARRY M . GORDON 1. Structure of Biological Membranes. 180 11. Structural Aspects of Hormone Receptor-Adenylate Cyclase Interaction. 183 111. Membrane Fluidity as a Regulator of Adenylate Cyclase Activity, 185 1V. Selective Modulation of Adenylate Cyclase by Asymmetric Perturbations of the Membrane Bilayer. 20X V. Phospholipid Headgroup Composition and Adenylate Cyclase Activity. 223 VI. Disease States. 225 References. 226

The Analysis of Interactions between Hormone Receptors and Adenylate Cyclase by Target Size Determinations Using Irradiation Inactivation

B. RICHARD MARTIN 1. Irradiation Inactivation: General Considerations, 233 11. Practical Considerations in irradiation Inactivation Sttidies

on Membranes, 236 111. Analysis of Data, 237

IV. The Application of Target Size Analysis to Rat Liver Plasma Membrane Adenylate Cyclase, 239 V. Model of Hormone Action. 241 VI. Effects of Fluoride, 246 V11. Evaluation of the Model in Relation to the Results of Other Approaches. 248 References, 253

vi i

...

CONTENTS

Vlll

PART 11.

RECEPTORS NOT INVOLVING ADENYLATE CYCLASE

Vasopressin lsoreceptors in Mammals: Relation to Cyclic AMP-Dependent and Cyclic AMPIndependent Transduction Mechanisms SERGE JARD I. Introduction, 25.5 11. Methodological Basis for the Characterization of Vasopressin Isoreceptors, 256 111. Kinetics of Hormone Binding to Vasopressin Receptors, 259 1V. Transduction Mechanisms Triggered by Vasopressin Receptors, 265 V. VI. VI1. VIII.

Effects of Nucleotides and Other Putative Effectors on Vasopressin Receptors, 272 Physicochemical Characteristics of Solubilized Vasopressin Receptors, 274 Recognition Patterns of Vasopressin Isoreceptors, 275 Summary and Conclusions. 279 References. 280

Induction of Hormone Receptors and Responsiveness during Cellular Differentiation MICHAEL C. LIN AND SUZANNE K. BECKNER 1. Introduction, 287 11. Model Systems, 290 111. Conclusion, 307 References, 310

Receptors for Lysosomal Enzymes and Glycoproteins VIRGINIA SHEPHERD, PAUL SCHLESINCER, and PHILIP STAHL

I. Introduction, 317 11. The Phosphomannosyl Recognition Pathway, 3 19 111. Role of Oligosaccharide Moiety in Recognition of Extracellular Lysosomal

Enzymes and Glycoproteins, 323 IV. Lysosomal Enzymes and the Mannosyl Recognition System, 324 V. Receptor-Mediated Endocytosis of Glycoconjugates, 327 VI. Conclusion, 335 References, 335

The Insulin-Sensitive Hexose Transport System in Adipocytes

J. GLIEMANN and W. D. REES

I. Summary of the Present Status, 339 11. Historical Background, 340

CONTENTS

111. IV. V. V1.

VII. VIII. IX. X. XI. XII. XIII. XIV.

Critical Steps in the Methodology, 342 Kinetic Approaches to the Study of Hexose Transport. 348 Transport of Nonmetabolizable Sugars and Sugar Analogs in the Adipocyte, 355 The Requirements for D-Glucose Binding to t h e Adipocyte Hexose Transport System, 359 Nontransported Competitive Inhibitors of Transport. 360 Sugars Which Are Both Transported and Phosphorylated-Rate-Limiting Steps. 362 Modulation of the Transport System by Glucose Metabolites, 366 Mechanism of Insulin’s Ability to Increase V,,,, 367 Human Adipocytes. 371 The Transport System in Obesity and Diabetes, 371 Reconstitution of the Hexose Transporter, 372 Concluding Remarks, 373 References, 373

Epidermal Growth Factor Receptor and Mechanisms for Animal Cell Division MANJUSRI DAS I. 11. 111. IV. V.

Introduction, 381 Properties of EGF, 382 The EGF Receptor, 383 The Pathway to Nuclear DNA Replication. 393 A Family of EGF-like Polypeptides and Their Role in Animal Development and Growth, 398 References. 400

The Linkage between Ligand Occupation and Response of the Nicotinic Acetylcholine Receptor PALMER TAYLOR. ROBERT DALE BROWN, and DAVID A. JOHNSON

I. Introduction, 407 11. Structure of the Isolated Receptor, 408 111. Biophysical Properties of the Receptor Channel, 412 IV. The Behavior of Partial Agonists, Antagonists, and Anesthetics in Relation to Channel Activation. 414 V. Desensitization of the Receptor, 415 VI. Ligand Occupation and Transitions in Receptor State, 416 V11. Other Ligands Affecting Receptor Function, 423 VIII. Analysis of Receptor Activation. 425 1X. Toward the Understanding of Coupling between Occupation of the Receptor and the Permeability Response. 426 X. Occupation and Activation by Agonists. 427 XI. Association of Antagonists with the Receptor and Functional Antagonism, 430 XII. Quantitation of Antagonist Occupation and Functional Antagonism, 434

ix

CONTENTS

X

X111. Structural Implications and Arrangement of Subunits, 435 XIV. Analysis of the Bound Ligand States. 437 References, 438

The Interaction of Cholera Toxin with Gangliosides and the Cell Membrane

SIMON VAN HEYNINGEN I. Structure and Action of Cholera Toxin, 446 11. The Role of Ganglioside GMl as a Cell-Surface Receptor, 450 111. The Nature of the Reaction between Ganglioside and Toxin, 455

IV. Transport of Cholera Toxin across the Cell Membrane and the Role of Binding to Ganglioside. 460 V. Other Compounds That Bind to Gangliosides, 465 References, 466

Index, 473

Contributors Numbers in parentheses indicate the pages on which the authors' contributions begin. Suzanne K. Beckner, Laboratory of Cellular and Developmental Biology, National Institute of Arthritis. Diabetes, Digestive and Kidney Diseases, National Institutes of Health, Bethesda. Maryland 20205 (287) Robert Dale Brown, Division of Pharmacology. Department of Medicine, University of California, San Diego, La Jolla. California 92093 (407) Brian A. Cooke, Department of Biochemistry, Royal Free Hospital School of Medicine, University of London. London WClN IBE England (143) Dermot M. F. Cooper,' Section on Membrane Regulation. Laboratory of Nutrition and Endocrinology, National Institute of Arthritis. Diabetes. Digestive and Kidney Diseases, National Institutes of Health, Bethesda, Maryland 20205 (67) Manjusri Das, Department of Biochemistry and Biophysics, University of Pennsylvania School of Medicine. Philadelphia. Pennsylvania 19104 (381) Vincent A. Florio, Department of Pharmacology, University of Texas Health Science Center at Dallas. Dallas, Texas 75235, and Department of Pharmacology, University of Virginia, Charlottesville, Virginia (109) J. Gliemann, Institute of Physiology, University of Aarhus, DK-8000 Aarhus C, Denmark (339) Larry M. Gordon, California Metabolic Research Foundation. La Jolla. California 92038, and Rees Stealy Research Foundation. San Diego, California 92101 (179) Miles D. Houslay, Department of Biochemistry. University of Manchester Institute of Science and Technology, Manchester, England. and California Metabolic Research Foundation, La Jolla, California 92038 (179) Serge Jard, Centre CNRS-INSERM de Pharmacologie-Endocrinologie, 34033 Montpellier Cedex. France (255) David A. Johnson, Division of Pharmacology, Department of Medicine, University of California, San Diego, La Jolla. California 92093 (407) Robert j. Lefkowitz, Department of Medicine (Cardiology), Howard Hughes Medical Institute, Duke University Medical Center, Durham. North Carolina 27710 (45) Michael C. Lin, Laboratory of Cellular and Developmental Biology, National Institute of Arthritis. Diabetes. Digestive and Kidney Diseases. National Institutes of Health, Bethesda, Maryland 20205 (287) B. Richard Martin, Department of Biochemistry. University of Cambridge. Cambridge CB2 IQW, England (3. 233) Steen E. Pedersen, Department of Pharmacotogy, University of Texas Health Science Center at Dallas, Dallas. Texas 75235, and Department of Biochemistry, University of Virginia. Charlottesville, Virginia (109)

Present address: Department of Pharmacology, University of Colorado Medical School, Denver, Colorado 80262. xi

xii

CONTENTS

John P. Perkins, Department of Pharmacology, University of North Carolina, Chapel Hill, North Carolina 27514 (85) W. D. Rees, Institute of Physiology. University of Aarhus. DK-8000 Aarhus C, Denmark (339) Elliott M. Ross, Department of Pharmacology, University of Texas Health Science Center at Dallas, Dallas, Texas 75235 (109) Paul Schlesinger, Department of Physiology and Biophysics, Washington University School of Medicine. St. Louis. Missouri 631 10 (317) Virginia Shepherd, Department of Physiology and Biophysics, Washington University School of Medicine, St. Louis, Missouri 631 10 (317) Jeffrey M. Stadel, Department of Medicine (Cardiology), Howard Hughes Medical Institute, Duke University Medical Center, Durham, North Carolina 27710 (45) Philip Stahl, Department of Physiology and Biophysics, Washington University School of Medicine, St. Louis, Missouri 631 10 (317) Palmer Taylor, Division of Pharmacology, Department of Medicine, University of California, San Diego, La Jolla, California 92093 (407) A. M. Tolkovsky, Department of Pharmacology. Hadassah Medical School, The Hebrew University, Jerusalem, Israel ( I I) Simon van Heyningen, Department of Biochemistry, University of Edinburgh, Edinburgh EH8 9XD, Scotland (445)

Preface The articles in this volume are intended to provide a survey of current ideas about the mechanisms by which plasma membrane receptors function. Part I deals with hormone receptors that modulate the activity of adenylate cyclase and, hence, the concentration of the second messenger cyclic 3',5'-AMP inside the cell. The discovery of hormone-sensitive adenylate cyclase in the early 1960s by Sutherland and his colleagues was perhaps the single most important advance in our understanding of the mechanism of hormone action. It led to an explanation of the action of a large class of hormones. It also led to the formulation of the second messenger hypothesis, which was an important concept in the study of hormone action in general. We now know a great deal about the mechanisms by which hormones activate adenylate cyclase. Rodbell and his colleagues showed that the activation requires the presence of GTE and it has since been shown, notably by the work of Gilman and his colleagues, that the effects of GTP and other guanine nucleotides are mediated through a separate protein subunit distinct from the hormone receptor and the catalytic subunit. Thus the system contains at least three components, and the major focus of interest at the present time is to determine precisely how these three components interact to modulate adenylate cyclase activity. More recently it has been shown that certain hormones and other agents, such as prostaglandins, directly inhibit adenylate cyclase in some tissues. It has also been found that prolonged exposure to activator hormones can lead to a loss of response, a process commonly referred to as desensitization. These effects are less well characterized than the activation process. The aim of Part I is to provide an overview of a single system in some depth. The aim of Part I1 is to give an impression of the diversity of mechanisms by which membrane receptors are thought to act. We have attempted to cover as many aspects as possible, ranging from the acetylcholine receptor, which functions by altering the membrane ion permeability in a very specific rapid and direct way, to toxins and growth factors, which ultimately enter the cell after binding and require comparatively long periods to act. We hope that the selection of contributors has resulted in a reasonably xiii

xiv

PREFACE

comprehensive overview of both aspects. However, the authors were asked to concentrate primarily on the work of their own laboratories rather than to attempt comprehensive reviews of the literature. As a result there are inevitably many topics equally worthy of consideration that have been omitted. Such deficiencies are entirely the responsibility of the Editors, ARNOSTKLEINZELLER B. RICHARD MARTIN

Yale Membrane Transport Processes Volumes

Joseph F. Hoffman (ed.). (1978). “Membrane Transport Processes,” Vol. 1. Raven, New York. Daniel C. Tosteson, Yu. A. Ovchinnikov, and Ramon Latorre (eds.). (1978). “Membrane Transport Processes,” Vol. 2. Raven, New York. Charles F. Stevens and Richard W. Tsien (eds.). (1979). “Membrane Transport Processes,” Vol. 3: Ion Permeation through Membrane Channels. Raven, New York. Emile L. Boulpaep (ed.). (1980).“Cellular Mechanisms of Renal Tubular Ion Transport”: Volume 13 of Current Topics in Membranes and Trunsport (F. Bronner and A. Kleinzeller, eds.). Academic Press, New York. William H. Miller (ed.). (1981). “Molecular Mechanisms of Photoreceptor Transduction”: Volume I5 of Ciirrcnt Topics in Mernhrcincs rind Transport (F. Bronner and A. Kleinzeller, eds.). Academic Press, New York. Clifford L. Slayman (ed.). (1982). “Electrogenic Ion Pumps”: Volume 16 of Cirrrent Topics in Mernhrrinrs rind Trrinsport (A. Kleinzeller and F. Bronner, eds.). Academic Press, New York.

xv

This Page Intentionally Left Blank

Contents of Previous Volumes Volume 1

Volume 3

Some Considerations about the Structure of Cellular Membranes MAYNARD M. DEWEY AND LLOYDBARR The Transport of Sugars across Isolated Bacterial Membranes H. R. KABACK Galactoside Permease of Escherichia cnli ADAMKEPES Sulfhydryl Groups in Membrane Structure and Function ASER ROTHSTEIN Molecular Architecture of the Mitochondrion H. MACLENNAN DAVID Author Index-Subject Index

The Na+, K+-ATPase Membrane Transport System: Importance in Cellular Function ARNOLDSCHWARTZ, GEORGE E. LINDENMAYER, AND JuLius C. ALLEN Biochemical and Clinical Aspects of Sarcoplasmic Reticulum Function ANTHONYMARTONOSI The Role of Periaxonal and Perineuronal Spaces in Modifying Ionic Flow across Neural Membranes W . J. ADELMAN, JR. AND Y. PALTI Properties of the Isolated Nerve Endings GEORGINA RODRiGUEZ DE LORES ARNAIZ AND EDUARDO DE ROBERTIS Transport and Discharge of Exportable Proteins in Pancreatic Exocrine Cells: I n Vitro Studies J . D. JAMIESON The Movement of Water across Vasopressin-Sensitive Epithelia M. HAYS RICHARD Active Transport of Potassium and Other Alkali Metals by the Isolated Midgut of the Silkworm AND WILLIAM R. HARVEY KARLZERAHN Author Index-Subject Index

Volume 2 The Molecular Basis of Simple Diffusion within Biological Membranes w. R. LIEBA N D w . D. STEIN The Transport of Water in Erythrocytes ROBERTE. FORSTER Ion-Translocation in Energy-Conserving Membrane Systems AND M. MONTAL B. CHANCE Structure and Biosynthesis of the Membrane Adenosine Triphosphatase of Mitochondria ALEXANDERTZAGOLOFF Mitochondria1Compartments: A Cornpanson of Two Models HENRY TEDESCHI Aurhor Index-Subject Index

Volume 4 The Genetic Control of Membrane Transport W.SLAYMAN CAROLYN xvii

xviii Enzymic Hydrolysis of Various Components in Biomembranes and Related Systems MAHENDRA KUMAR JAIN Regulation of Sugar Transport in Eukaryotic Cells HOWARD E. MORGANA N D CAROL F. WHITFIELD Secretory Events in Gastric Mucosa RICHARD P. DURBIN Author Index-Subjeci Index

Volume 5 Cation Transport in Bacteria: K+, Na+, and H+ FRANKLIN M. HAROLD AND KARLHEINZ ALTENDORF Pro and Contra Camer Proteins: Sugar Transport via the Periplasmic GalactoseBinding Protein WINFRIED Boos Coupling and Energy Transfer in Active Amino Acid Transport ERICHHEINZ The Means of Distinguishing between Hydrogen Secretion and Bicarbonate Reabsorption: Theory and Applications to the Reptilian Bladder and Mammalian Kidney AND WILLIAM A. BRODSKY THEODORE P. SCHILB Sodium and Chloride Transport across Isolated Rabbit Ileum G. SCHULTZ AND STANLEY PETER F. CURRAN A Macromolecular Approach to Nerve Excitation ICHIll TASAKI AND EMILIO CARBONE Subject Index

Volume 6 Role of Cholesterol in Biomembranes and Related Systems MAHENDRA KUMAR JAIN Ionic Activities in Cells A. A. LEVAND W. MCD. ARMSTRONG Active Calcium Transport and Ca*+-Activated ATPase in Human Red Cells H. J. SCHATZMANN The Effect of Insulin on Glucose Transport in Muscle Cells TORBEN CLAUSEN

CONTENTS OF PREVIOUS VOLUMES

Recognition Sites for Material Transport and Information Transfer HALVOR N. CHRISTENSEN Subject Index

Volume 7 Ion Transport in Plant Cells E. A. C. MACROBBIE H+ Ion Transport and Energy Transduction in Chloroplasts AND RICHARD A. DILLEY ROBERT T. G~AQUINTA The Present State of the Camer Hypothesis PAULG . LEFEVRE Ion Transport and Short-circuit Technique S. REHM WARREN Subjeci Index

Volume 8 Chemical and Physical Properties of Myelin Proteins M. A. MOSCARELLO The Distinction between Sequential and Simultaneous Models for Sodium and Potassium Transport P. J. GARRAHAN AND R. P. GAMY Soluble and Membrane ATPase of Mitochondria, Chloroplasts, and Bacteria: Molecular Structure, Enzymatic Properties, and Functions RIVKAPANET AND D. h0 SANADI Competition, Saturation, and InhibitionIonic Interactions Shown by Membrane Ionic Currents in Nerve, Muscle, and Bilayer Systems ROBERT J. FRENCH AND WILLIAM J. ADELMAN, JR. Properties of the Glucose Transport System in the Renal Brush Border Membrane R. KINNE Subject Index

Volume 9 The State of Water and Alkali Cations within the Intracellular Fluids: The Contribution of NMR Spectroscopy AND MORDECHAI SHPORER MORTIMER M. CIVAN

xix


Electrostatic Potentials at Membrane-Solution Interfaces STUART MCLAUGHLIN A Thermodynamic Treatment of Active Sodium Transport S. ROYCAPLAN AND ALVINESSIG Anaerobic Electron Transfer and Active Transport in Bacteria WIL N. KONINGS AND JOHANNES BOONSTRA Protein Kinases and Membrane Rosphorylation M. MARLENEHOSEYAND MARIANO TAO Mechanism and Physiological Significance of Calcium Transport across Mammalian Mitochondria1 Membranes LEENAMELA Thyroidal Regulation of Active Sodium Transport F. ISMAIL-BEIGI Subject Index

Volume 10 Mechanochernical Properties of Membranes E. A. EVANS AND R. M.HOCHMUTH Receptor-Mediated Protein Transport into Cells. Entry Mechanisms for Toxins, Hormones, Antibodies, Viruses, Lysosomal Hydrolases, Asialoglycoproteins, and Carrier Proteins JR. A N D DAVID M. NEVILLE, TA-MINCHANG The Regulation of Intracellular Calcium CARAFOLI AND ERNESTO MARTINCROMITON Calcium Transport and the Properties of a Calcium-Sensitive Potassium Channel in Red Cell Membranes L. LEWA N D VIRGILIO HUGOG. FERREIRA Proton-Dependent Solute Transport in Microorganisms A. A. EDDY Subject Index

Volume 11 Cell Surface Glycoprotelns: Structure, 6losynth~ls,and Blologlcal Functions

The Cell Membrane-A Short Historical Perspective ASER ROTHSTEIN The Structure and Biosynthesis of Membrane Glycoproteins JENNIFER STURGESS, MARIOMOSCARELLO, AND HARRY SCHACHTER Techniques for the Analysis of Membrane Glycoproteins R. L. JULIANO Glycoprotein Membrane Enzymes JOHNR. RIORDAN AND GORDON G. FORSTNER Membrane Glycoproteins of Enveloped Viruses W. COMPANS AND RICHARD MAURICEC. KEMP Erythrocyte Glycoproteins MICHAELJ. A. TANNER Biochemical Determinants of Cell Adhesion LLOYDA. CULP Proteolytic Modification of Cell Surface Macromolecules: Mode of Action in Stirnulating Cell Growth KENNETHD. NOONAN Glycoprotein Antigens of Murine Lymphocytes MICHELLELETARTE Subject h d e x

Volume 12 Carriers and Membrane Transport Protelns

Isolation of Integral Membrane Proteins and Criteria for Identifying Carrier Proteins MICHAELJ . A. TANNER The Carrier Mechanism S. B. HLADKY The Light-Driven Proton Pump of Halobacterium halobium: Mechanism and Function MICHAELEISENBACH AND S. ROY CAPLAN Erythrocyte Anion Exchange and the Band 3 Protein: Transport Kinetics and Molecular Structure PHiLip A. KNAUF

xx The Use of Fusion Methods for the Microinjection of Animal Cells R. G. KULKA AND A, L ~ Y T E R Subject Index

Volume 13 Cellular Mechanlsms of Renal Tubular

Ion Transport PART I: ION ACTIVITY AND ELEMENTAL COMPOSITION OF INTRAEPITHELIAL COMPARTMENTS lntracellular pH Regulation WALTERF. BORON Reversal of the pH,-Regulating System in a Snail Neuron R. C. THOMAS How to Make and Use Double-Barreled Ion-Selective Microelectrodes THOMAS ZUETHEN The Direct Measurement of K, CI, Na, and H Ions in Bullfrog Tubule Cells MAMORUFUJIMOTO, KUNIHIKO KOTERA, AND YUTAKAMATSUMURA Intracellular Potassium Activity Measurements in Single Proximal Tubules of N e c turus Kidney TAKAHIRO KUBOTA, BRUCE BIAGI,AND GERHARD GIEBISCH Intracellular Ion Activity Measurements in Kidney Tubules RAJAN. KHURI Intracellular Chemical Activity of Potassium in Toad Urinary Bladder JOEL DELONG AND MORTIMERM. CIVAN Quantitative Determination of Electrolyte Concentrations in Epithelial Tissues by Electron Microprobe Analysis ROGERRICK,ADOLFDORGE, RICHARD BAUER,FRANZBECK, JUNEMASON,CHRISTIANE ROLOFF, AND KLAUS THURAU PART 11: PROPERTIES OF INTRAEPITHELIAL MEMBRANE BARRIERS IN THE KIDNEY


Hormonal Modulation of Epithelial Structure JAMESB. WADE Changes in Cell Membrane Surfaces Associated with Alterations of Transepithelial Ion Movement MICHAELKASHGARIAN The Dimensions of Membrane Bamers in Transepithelial Flow Pathways LARRY w . WELLING AND DANJ. WELLING Electrical Analysis of Intraepithelial Barriers AND EMI LE L. BOULPAEP HENRY SACKIN Membrane Selectivity and Ion Activities of Mammalian Tight Epithelia SIMON A. LEWIS,NANCYK. WILLS, A N D DOUGLAS C. EATON Ion Conductances and Electrochemical Potential Differences across Membranes of Gallbladder Epithelium Luis REUSS A Kinetic Model for Ion Fluxes in the Isolated Perfused Tubule BRUCE BIAGI, ERNESTO GONZALEZ, A N D GERHARD GIEBISCH The Effects of Voltage Clamping on Ion Transport Pathways in Tight Epithelia ARTHURL. FINN AND PAULA ROGENES Tubular Permeability to Buffer Components as a Determinant of Net H Ion Fluxes G . MALNIC, v. L. COSTA SILVA, s. s. CAMPIGLIA, M. DE MELLO AIRES, AND G. GIEBISCH Ionic Conductance of the Cell Membranes and Shunts of Necturus Proximal Tubule AND GENJIRO KIMURA KENNETH R. SPRING Luminal Sodium Phosphate Cotransport as the Site of Regulation for Tubular Phosphate Reabsorption: Studies with Isolated Membrane Vesicles HEINIMURER,REINHARD STOLL, CARLA EVERS,ROLFKINNE, JEAN-h1LIPPE BONJOUR, AND HERBERT FLEISCH The Mechanism of Coupling between Glucose Transport and Electrical Potential in the Proximal Tubule: A Study of Potential-

xxi


Dependent Phlorizin Binding to Isolated Renal Microvillus Membranes PETERS. ARONSON Electrogenic and Electroneutral Na Gradient-Dependent Transport Systems in the Renal Brush Border Membrane Vesicle BERTRAM SACKTOR PART 111: INTRAMEMBRANE CARRIERS AND ENZYMES IN TRANSEPITHELIAL TRANSPORT Sodium Cotransport Systems in the Proximal Tubule: Current Developments A N D H. MURER R. KINNE,M. BARAC, ATPases and Salt Transport in the Kidney Tubule DE LA MARCARITA PEREZ-GONZALEZ MANNA, FULGENCIO PROVERBIO, AND GUILLERMO WHITEMBURY Further Studies on the Potential Role of an Anion-Stimulated Mg-ATPase in Rat b o x imal Tubule Proton Transport AND R. KINNE E. KINNE-SAFFRAN Renal Na+- K+-ATPase: Localization and Quantitation by Means of Its K+-Dependent Phosphatase Activity 111 AND REINIER BEEUWKES SEYMOUR ROSEN Relationship between Localization of N+K+-ATPase, Cellular Fine Structure, and Reabsorptive and Secretory Electrolyte Transport STEPHEN A. ERNST, CLARA v. RIDDLE, AND JR. KARLJ. KARNAKY, Relevance of the Distribution of Na+ Pump Sites to Models of Fluid Transport across Epithelia JOHNW. M u s AND DONALD R. DIBONA Cyclic AMP in Regulation of Renal Transport: Some Basic Unsolved Questions THOMAS P. DOUSA Distribution of Adenylate Cyclase Activity in the Nephron F. MOREL,D. CHABARDBS, AND M. IMBERT-TEBOUL Subject Index

Volume 14 Carrlers and Membrane Transport Protelns

Interface between Two Immiscible Liquids as a Tool for Studying Membrane Enzyme Systems L. I. BOGUSLAVSKY Criteria for the Reconstitution of Ion Transport Systems ADILE. SHAMOO AND WILLIAM F. TIVOL The Role of Lipids in the Functioning of a Membrane Protein: The Sarcoplasmic Reticulum Calcium Pump J . P.BENNETT, K. A. MCGILL,AND G.B. WARREN The Asymmetry of the Hexose Transfer System in the Human Red Cell Membrane w. F. WIDDAS Permeation of Nucleosides, Nucleic Acid Bases, and Nucleotides in Animal Cells PETER G . w. K A G E M A " AND ROBERT M. WOHLHUETER Transmembrane Transport of Small Peptides D. M. MAITHEWS ANDJ.W. PAYNE Characteristics of Epithelial Transport in Insect Malpighian Tubules S. H. P. MADDRELL Subject Index

Volume 15 Molecular Mechanlsms of Photoreceptor Transductlon

PART I: THE ROD PHYSIOLOGICAL RESPONSE The Photocurrent and Dark Current of Retinal Rods G. M A ~ H E W AND S D. A. BAYLOR Spread of Excitation and Background Adaptation in the Rod Outer Segment K.-W. YAU. T. D. LAMB,A N D P. A. MCNAKJGHTON Ionic Studies of Vertebrate Rods W. GEOFFREY OWENAND VINCENT TORRE

xxii Photoreceptor Coupling: Its Mechanism and Consequences GEOFFREY H. GOLD PART 11: THE CYCLIC NUCLEOTIDE ENZYMATIC CASCADE AND CALCIUM ION First Stage of Amplification in the CyclicNucleotide Cascade of Vision LUBERT STRYER, JAMES B. HURLEY, A N D BERNARD K.-K. FUNG Rod Guanylate Cyclase Located in Axonemes DARRELLFLEISCHMAN Light Control of Cyclic-Nucleotide Concentration in the Retina G. EBREY,PAULKILBRIDE, THOMAS JAMES B. HURLEY, ROGERCALHOON, AND MOTOYUKITSUDA Cyclic-GMP Phosphodiesterase and Calmodulin in Early-Onset Inherited Retinal Degenerations G. J. CHADER, Y. P. LIU, R. T. FLETCHER, G . ACUIRRE, R. SANTOS-ANDERSON, A N D M. T’so Control of Rod Disk Membrane Phosphodiesterase and a Model for Visual Transduction P. A. LIEBMAN AND E. N. WGH, JR. Interactions of Rod Cell Proteins with the Disk Membrane: Influence of Light, Ionic Strength, and Nucleotides HERMANN KUHN Biochemical Pathways Regulating Transduction in Frog Photoreceptor Membranes M. DERICBOWNDS The Use of Incubated Retinas in Investigating the Effects of Calcium and Other Ions on Cyclic-Nucleotide Levels in Photoreceptors ADOLPHI. COHEN Cyclic AMP Enrichment in Retinal Cones DEBORA B. FARBER Cyclic-Nucleotide Metabolism in Vertebrate Photoreceptors: A Remarkable Analogy and an Unraveling Enigma M. W. BITENSKY, G. L. WHEELER, A. YAMAZAKI, M. M. RASENICK, AND P. 3. STEIN


Guanosine Nucleotide Metabolism in the Bovine Rod Outer Segment: Distribution of Enzymes and a Role of GTP HITOSHI SHICHI Calcium Tracer Exchange in the Rods of Excised Retinas ETE z. SZUTS The Regulation of Calcium in the Intact Retinal Rod: A Study of Light-Induced Calcium Release by the Outer Segment GEOFFREY H. GOLDAND JUANI. KORENBROT Modulation of Sodium Conductance in Photoreceptor Membranes by Calcium Ions and cGMP ROBERT T. SORBI PART 111: CALCIUM, CYCLIC NUCLEOTIDES, AND THE MEMBRANE POTENTIAL Calcium and the Mechanism of Light Adaptation in Rods BRUCEL. BASTIAN AND GORDON L. FAIN Effects of Cyclic Nucleotides and Calcium Ions on Bufo Rods AND JOELE. BROWN GERALDINE WALOGA The Relation between Caz+and Cyclic GMP in Rod Photoreceptors STUART A. LIFTONAND JOHNE. DOWLING Limits on the Role of Rhodopsin and cCMP in the Functioning of the Vertebrate Photoreceptor SANFORD E. OSTROY, EDWARD P. MEYERTHOLEN, PETERJ. STEIN, ROBERTAA. SVOBODA, AND MEEGAN J . WILSON [Ca2+],Modulation of Membrane Sodium Conductance in Rod Outer Segments BURKSOAKLEY I1 A N D LAWRENCE H. PINTO Cyclic-GMP-Induced Depolarization and Increased Response Latency of Rods: Antagonism by Light WILLIAM H. MILLERAND GRANT D. N~COL

xxiii


PART IV: AN EDITORIAL OVERVIEW Ca2+and cGMP WILLIAM H. MILLER Index

Volume 16 Electrogenic Ion Pumps PART I. DEMONSTRATION OF PUMP ELECTROGENICITY IN EUKARYOTIC CELLS Electrophysiology of the Sodium Pump in a Snail Neuron R. C. THOMAS Hyperpolarization of Frog Skeletal Muscle Fibers and of Canine Purkinje Fibers during Enhanced Na+-K+ Exchange: Extracellular K+ Depletion or Increased Pump Current? DAVIDC. GADSBY The Electrogenic Pump in the Plasma Membrane of Nitella ROGERM. SPANSWICK Control of Electrogenesis by ATP, Mg2+. H+, and Light in Perfused Cells of Cham MASASHI TAZAWA AND TERUO SHIMMEN PART 11. T H E EVIDENCE IN EPITHELIAL MEMBRANES An Electrogenic Sodium Pump in a Mammalian Tight Epithelium S. A. LEWIS A N D N. K. WILLS A Coupled Electrogenic Na+- K+ Pump for Mediating Transepithelial Sodium Transport in Frog Skin ROBERTNltLSEN Transepithelial Potassium Transport in Insect Midgut by an Electrogenic Alkali Metal Ion Pump MICHAEL G . WOLFERSBERGER, AND WILLIAM R. HARVEY. MOIRAClOFFl The ATP-Dependent Component of Gastric Acid Secretion G. SACHS, B. WALLMARK, E. RABON, G . SACCOMANI, H. B. STEWART. D. R. DIBONA,AND T. BERGLINDH

PART 111. REVERSIBILITY: ATP SYNTHESIS DRIVEN BY ELECTRIC FIELDS Effect of Electrochemical Gradients on Active H+ Transport in an Epithelium QAISAL-AWQATI AND TROYE. DIXON Coupling between H+Entry and ATP Synthesis in Bacteria PETERC. MALONEY Net ATP Synthesis by H+-ATPase Reconstituted into Liposomes YAW0 KAGAWA Phosphorylation in Chloroplasts: ATP Synthesis Driven by A$ and by ApH of Artificial or Light-Generated Origin PETERGRABER PART IV. SOME THEORETICAL QUESTIONS Response of the Proton Motive Force to the Pulse of an Electrogenic Proton Pump ERICHHEINZ Reaction Kinetic Analysis of CurrentVoltage Relationships for Electrogenic Pumps in Neurospora and Acetabularia DIETRICH GRADMANN, AND ULF-PETERHANSEN, CLIFFORD L. SLAYMAN Some Physics of Ion Transport J. MOROWITZ HAROLD PART V. MOLECULAR MECHANISMS OF CHARGE SEPARATION An H+-ATP Synthetase: A Substrate Translocation Concept I. A. KOZLOVA N D V. P. SKULACHEV Proton Translocation by Cytochrome Oxidase MARTENWIKSTROM Electrogenic Reactions of the Photochemical Reaction Center and the UbiquinoneCytochrome b / c z Oxidoreductase P. LESLIEDUTTON,PAULMUELLER, P. O'KEEFE, DANIEL NIGELK. PACKHAM, ROGERc. PRINCE, AND DAVIDM. TIEDE


xxiv Proton-Membrane Interactions in Chloroplast Bioenergetics R. A. DILLEY. L. J. PROCHASKA, G. M. BAKER,N. E. TANDY, AND P. A. MILLNER Photochemical Charge Separation and Active Transport in the Purple Membrane BARRY HONK Mitochondrial Transhydrogenase: General Principles of Functioning I. A. KOZLOV Membrane Vesicles, Electrochemical Ion Gradients, and Active Transport H. R. KABACK PART V1. BlOLOGlCAL SIGNIFICANCE OF ELECTROGENIC ION PUMPS The Role of Electrogenic Proton Translocation in Mitochondrial Oxidative Phosphorylation JANNA P. WEHRLE Electrogenic Reactions and Proton Pumping in Green Plant Photosynthesis WOLFGANG JUNCE The Role of the Electrogenic Sodium Pump in Controlling Excitability in Nerve and Cardiac Fibers MARIO VASSALLE

Pumps and Currents: A Biological Ferspective FRANKLIN M. HAROLD Index

Volume 17 Membrane Lipids of Prokaryotes Lipids of Prokaryotes-Structure and Distribution HOWARD GOLDFINE Lipids of Bacteria Living in Extreme Environments A. LANGWORTHY THOMAS Lipopolysaccharides of Gram-Negative Bacteria OTTO LUDERITZ, M A R I N A A. FREUDENBERG, CHRIS GALANOS, VOLKER LEHMANN. AND ERNSr TH. RIETSCHEL, DEREK H. SHAW Prokaryotic Polyterpenes: Phylogenetic Precursors of Sterols GUY OURISSON AND MICHELROHMER Sterols in Mycoplasma Membranes SHMUEL RAZIN Regulation of Bacterial Membrane Lipid Synthesis CHARLES 0. ROCKAND JOHNE. CRONAN. JR. Transbilayer Distribution of Lipids in Microbial Membranes SHLOMO ROTTEM Lipid Phase Transitions and Regulation of Membrane Fluidity in Prokaryotes D O N A L D L. MELCHIOR Effects of Membrane Lipids on Transport and Enzymic Activities RONALDN. MCELHANEY Index

Current Topics in Membranes and Transport

Volume 18

MEMBRANE RECEPTORS


Part I

Adenylate Cyclase-Related Receptors


CURRENT TOPICS IN MEMBRANES AND TRANSPORT. VOLUME IS

Hormone Receptors and the Adeny late Cyclase System: Histo rical Overview B . RICHARD MARTIN Deparrment of Biochemistry University of Cambridge Cambridge, England

Text . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3 8

When Sutherland and his co-workers first demonstrated the existence of hormone-stimulated adenylate cyclase a very simple model was sufficient to explain the available information on hormone regulation of cell function. It could be suggested that the enzyme consisted of a single protein which spanned the plasma membrane. The catalytic site responsible for the conversion of ATP to cyclic 3',5'-AMP would be located on the inner surface of the plasma membrane and a recognition or receptor site for the hormone on the outer surface of the plasma membrane. The binding of the hormone to the recognition site on the receptor would then lead to a conformational change resulting in an increase in the catalytic activity of the enzyme (Robinson et al., 1967). Thus, it was suggested that adenylate cyclase functioned in a fashion similar to any other regulatory enzyme with the additional feature that the regulatory and catalytic sites were located asymmetrically on opposite sites of the plasma membrane. This would allow the enzyme to be used as a means of transfer of information from the outer surface of the cell to the cytoplasm. This relatively simple formulation was rapidly shown to be inadequate, and we now know that the hormone-sensitive adenylate cyclase system consists of at least three distinct protein components. The nature of the physical interactions between these components is the subject of many of the articles in this volume. 3

Copynght Q 1983 by Academic Press. Inc. All rights of reproduction in any form reserved. ISBN 0-12-153318-2

4

E. RICHARD MARTIN

Among the most significant early observations were those of Rodbell and his colleagues (for review see Rodbell et al., 1975). They showed that in rat fat cell plasma membranes several different hormones were capable of activating adenylate cyclase. In their initial study they examined the effects of P-adrenergic agonists, glucagon, ACTH, and secretin, but the enzyme is also activated by a number of other hormones. They found that the effects of the hormones were not additive. If adenylate cyclase was maximally activated by the most effective agonist, isoproterenol, the addition of any combination of the other hormones had no effect on either the activation or inhibition of the enzyme. In view of the very marked dissimilarities between the structures of the different agonists it seemed unlikely that they were competing for the same receptor, and, indeed, a large body of evidence from many groups has shown that each hormone has its own distinct receptor protein, Thus, it seemed likely that the hormone recognition site was on a protein component distinct from that containing the catalytic site of adenylate cyclase and that several different hormone receptors were able to compete for the activation of a common pool of adenylate cyclase catalytic units. To explain the interaction of a number of different receptors with a common pool of catalytic units, the mobile receptor hypothesis was proposed (Birnbaumer, 1973; Bennett et al., 1975). This suggested that in the absence of hormone, the catalytic unit of adenylate cyclase and the hormone receptor were separate from each other but able to move in relation to each other, laterally within the fluid matrix of the membrane lipid. The binding of the hormone then results in a conformational change in the hormone receptor, which increases its affinity for the adenylate cyclase catalytic unit so that when they make contact a long-lived association is set up. It was proposed that this association results in the activation of the catalytic unit. Direct evidence for the existence of separate components responsible for hormone binding and adenylate cyclase catalytic activity came from a number of studies in which plasma membranes were solubilized using nonionic detergents. When this was done it was found that the capacity of the enzyme to be activated by hormone was invariably lost, and in some cases it was possible to separate physically the hormone binding activity from the adenylate cyclase catalytic activity (Welton et al., 1977, 1978). Direct evidence for the mobile receptor hypothesis came from the studies of Schramm and his colleagues (Orly and Schramm, 1976). They selected two distinct cell lines, a Friend erythroleukemia cell which contained adenylate cyclase catalytic activity but no P-adrenergic receptor and a turkey erythrocyte which contained both adenylate cyclase catalytic activity and a P-adrenergic hormone receptor. The catalytic activity of adenylate cyclase in the turkey erythrocytes was then selectively destroyed by treatment with N-ethylmaleimide, leaving the specific P-adrenergic binding activity intact. Thus, they had one cell line which contained a hormone receptor but no adenylate cyclase catalytic activity and a second cell line which contained cata-

HISTORICAL OVERVIEW

5

lytic activity but no hormone receptor. The two cell lines were then fused using Sendai virus and an isolated plasma membrane fraction. The hybrid plasma membranes were found to contain adenylate cyclase, which was activated by catecholamines. This provided a convincing demonstration that the two separate entities, the receptor and the enzyme, were able to migrate within the plane of the plasma membrane and to produce a productive functional interaction. The same group has been able to extend this approach using a number of different sources to supply the catalytic unit and receptors for a number of different hormones again derived from several different tissues (Schramm et al., 1977; Schulster ef al., 1978). This has conclusively demonstrated that the recognition site for the hormone and the catalytic site of adenylate cyclase are located on distinct protein components. It also incidentally shows that the structure of the sites on the two different components which interact must be very highly conserved since it is possible to generate hybrid membranes in which the adenylate cyclase will demonstrate a hormone response and in which the catalytic unit and the hormone receptor are derived from completely different species. Rodbell and his colleagues were also responsible for a second major advance in our understanding of the mechanism of activation of adenylate cyclase by hormones. They showed that in rat fat cell plasma membranes and rat liver plasma membranes, adenylate cyclase was activated by GTP at very low concentrations of the order of lo-’ M .They also showed that GTP and hormone used in combination gave a markedly synergistic activation (for review see Rodbell et al., 1975). Subsequent studies have shown that if care is taken to exclude GTP from the system and, in particular, to ensure that the ATP used as substrate for adenylate cyclase is not contaminated with GTP then there is very little activation by hormone alone and the hormonal activation becomes largely dependent upon added GTP (Kimura er al., 1976). The next stage was to consider the possible role of transfer of the terminal phosphate of GTP in the activation process. This was approached by making use of an analog of GTP, guanylyl imidodiphosphate [p(NH)ppG], in which the terminal phosphate linkage is resistant to hydrolysis or transfer. Far from being less effective as an activator of adenylate cyclase than GTP, p(NH)ppG gave a much larger and essentially irreversible activation. In the case of adenylate cyclase, in most plasma membranes from mammalian sources p(NH)ppG is capable of activating the enzyme to its maximum extent and the effect of addition of hormone is only to increase the rate at which activation is achieved. In the widely studied turkey erythrocyte plasma membrane and in other avian erythrocyte plasma membranes the activation of the enzyme by either GTP or p(NH)ppG is largely dependent upon the presence of hormone and is much slower than in mammalian membranes (Rodbell et al., 1975; Tolkovsky and Levitzki, 1978). This has been taken advantage of in the kinetic studies of Tolkovsky and Levitzki which are discussed in Part I of this volume. The involvement of GTP as an additional component required for

6

B. RICHARD MARTIN

hormonal activation led Rodbell to postulate the existence of a third component of the hormone-sensitive adenylate cyclase. He suggested that the primary activator of the enzyme was GTP and that the role of the hormone receptor complex was to render the activation of adenylate cyclase by GTP more effective. The existence of a component which he designated the transducer was proposed to act as a link between the hormone receptor and the catalytic unit and to mediate the effects of GTP. The greater effectiveness of p(NH)ppG as an activator and the irreversible nature of the activation were taken to suggest that the binding of GTP was responsible for the activation of adenylate cyclase and that the reversal of activation required the hydrolysis of GTP to GDP and Pi. More recent studies have shown that at least in outline both of these suggestions are correct. Using turkey erythrocyte plasma membranes Cassel and Selinger (1976) demonstrated the existence of a specific GTPase which was activated by catecholamines. They suggested that the increase in the rate of hydrolysis of GTP resulted from an increase in the rate of formation of an enzyme GTP complex which was promoted by the hormone receptor complex. Cholera toxin was known to activate adenylate cyclase irreversibly, the effect being both time dependent and dependent upon the presence of NAD (Gill, 1975). They were able to show that pretreatment of turkey erythrocyte plasma membranes with cholera toxin in the presence of NAD caused an irreversible loss of hormonesensitive GTPase activity and that the activation of adenylate cyclase by cholera toxin was dependant upon the presence of GTP (Cassel and Selinger, 1977). Later it was demonstrated that in both avian erythrocyte plasma membranes and liver plasma membranes cholera toxin catalyzed the ADP-ribosylation of a protein with a molecular weight of 42,000. In the case of the avian erythrocyte plasma membranes the same protein could be shown to be covalently labeled by a photoreactive analog of GTP, whereas in liver plasma membranes the ADPribosylation was shown to be dependent upon the presence of GTP (Cassel and Pfeufer, 1978; Doberska et al., 1980). It seems likely, therefore, that the protein is a component of the hormone-activated GTPase. Thus, the effect of cholera toxin is to modify this GTPase covalently by the incorporation of ADP-ribose using NAD as donor and resulting in the inhibition of the GTPase activity. This has the effect of rendering GTP more effective as an activator since the rate of reversal of activation by hydrolysis of GTP to GDP and Pi is reduced. Subsequent work using a variety of guanine nucleotide analogs and examining the rate of release of GDP from the membranes led to the conclusion that the role of the hormone receptor complex is to open the guanine nucleotide binding site and to allow the exchange of one guanine nucleotide for another (Cassel and Selinger, 1978). The effect of this under physiological conditions will be to allow the exchange of bound GDP remaining at the end of a cycle of activation for GTP to initiate a fresh cycle of activation. We should now address the question of the location of the guanine nucleotide

HISTORICAL OVERVIEW

7

binding site. There are two possibilities: it could be located on the same protein component as the catalytic site of adenylate cyclase or it could be located on a distinct "transducer" protein as suggested by Rodbell. In most cases detergentsolubilized preparations of adenylate cyclase retain the ability to respond to p(NH)ppG, and in early studies of this type of preparation it was not found to be possible to resolve the catalytic activity of adenylate cyclase from the guanine nucleotide binding activity of the preparation (Pfeufer and Helmreich, 1975). On the basis of this type of circumstantial evidence it seemed possible that the guanine nucleotide binding site was located on the same protein component as the catalytic site of adenylate cyclase. Subsequent work by Gilman and his coworkers has demonstrated conclusively that the guanine nucleotide binding site and the adenylate cyclase catalytic site are on separate protein components. They made use of a strain of S49 lymphoma cells known as cyc- . In contrast to the wild type, these cells appear to lack a hormone-sensitive adenylate cyclase. It can be shown, however, that the binding of catecholamines to the cell surface is normal. Furthermore, it can be shown that the plasma membranes contain a normal adenylate cyclase catalytic unit. This can be demonstrated by making use of the activation of the enzyme by Mn2+ ions. In the presence of Mn2+ the activity of adenylate cyclase in cyc- cell plasma membranes is the same as the activity in wild-type membranes under the same conditions. Thus, these membranes contain a functional hormone receptor and a functional catalytic site, but the two components lack the ability to interact. The component which is either missing or impaired in function is the guanine nucleotide binding site. The membranes lack the 42,000-dalton protein which is ADP-ribosylated by cholera toxin. The guanine nucleotide response can be restored by fusion with membranes which contain a functional guanine nucleotide binding protein (Ross er al., 1978). In fact, the transfer of guanine nucleotide binding protein between membranes may be much simpler than the transfer of hormone receptors, which requires membrane fusion. Lad et al. (1980) have reported that the guanine nucleotide binding protein can be transferred to turkey erythrocyte plasma membranes from human erythrocyte plasma membranes which, while lacking the other components of the system, appear to contain the guanine nucleotide binding component. A recent report by Bhat er al. (1980) showed that the guanine nucleotide response can be restored by fractions derived from the cytoplasm of cells as well as by fractions derived from the plasma membrane. These observations imply that the guanine nucleotide binding protein may be a membrane peripheral protein rather than a membrane integral protein. In any case the use of cyc - membranes as a means of detecting the presence of the guanine nucleotide binding protein is well established and has allowed considerable progress toward the purification of this component of the system. At this point, in order to explain the activation of adenylate cyclase by hormones we must take account of the role of at least three separate protein compo-

8

B. RICHARD MARTIN

nents: (1) the hormone receptor usually designated (R), (2) the catalytic unit usually designated (C), and (3) a guanine nucleotide binding component usually designated (N) or, by some workers, ( G ) .The situation may well be even more complex with other components responsible for the action of inhibitory hormones and for the down-regulation of adenylate cyclase. However, at the present time there seems to be no reason to invoke any protein components in the activation process other than those discussed above. Obviously, it is of great interest to determine the sequence of interactions between these three components; that is, which interacts with which and at which stage of the cycle of activation and reversal of activation of adenylate cyclase. A number of approaches to this question have been used. Tolkovsky and Levitzki have made use of a detailed analysis of the kinetics of activation by hormones. Houslay and his colleagues have examined the effects of alterations in the membrane lipid mobility on adenylate cyclase activity in the presence of different activators. Both of these studies are described in Part I of this volume. We have used the method of target size analysis by irradiation inactivation in an attempt to obtain a direct physical measure of size changes of the components of the hormone-sensitive adenylate cyclase and, accordingly, of the association and dissociation of the different components. The approaches discussed above represent significant steps toward the elucidation of the molecular mechanisms involved in the interaction of hormones and membrane receptors and the adenylate cyclase system, REFERENCES Bennett, V., O’Keefe, E., and Cuatrecasas, P. (1975). Mechanism of action ofcholera toxin and the mobile receptor theory of hormone receptor adenylate cyclase interactions. Proc. Natl. Acud. Sci. U.S.A. 72, 33-37. Bhat, M. K., Iyengar, R., Abramowitz, J., Bordelon-Riser, M. E., and Birnbaumer, L. (1980). Naturally soluble components that confer guanine nucleotide and fluoride sensitivity to adenylate cyclase. Proc. Nafl. Acad. Sci. U.S.A. 77, 3836-3840. Birnbaumer, L. (1973). Hormone sensitive adenylate cyclases. Useful models for studying hormone receptor functions in cell free systems. Biochim. Biophys. Acra 300, 129-158. Cassel, D., and Pfeufer, T . (1978).Mechanism of cholera toxin action: Covalent modification of the guanyl nucleotide binding protein of the adenylate cyclase system. Proc. Nurl. Acad. Sci. U.S.A. 75, 2669-2673. Cassel, D., and Selinger, Z. (1976).Catecholamine stimulated GTPase activity in turkey erythrocyte membranes. Biochim. Biophys. Acra 452, 538-551. Cassel, D., and Selinger, Z. (1977). Mechanism of activation of adenylate cyclase by cholera toxin. Inhibition of GTP hydrolysis at the regulatory site. Proc. Nafl. Acud. Sci. U.S.A. 74, 3307-331 I . Cassel, D., and Selinger, Z. (1978). Mechanism of activation of adenylate cyclase through the padrenergic , receptor: Catecholamine-induced idisplacement of bound GDP by GTP. Proc. Natl. Acad. Sci. U.S.A. 75, 4155-4159. Doberska, C. A., Macpherson, A. J. S., and Martin, B. R. (1980). Requjrement for guanosine triphosphate for cholera-toxin-catalysed incorporation of adenosine diphosphate ribose into rat liver plasma membranes and for activation of adenylate cyclase. Biochem. J . 186, 749-754.

HISTORICAL OVERVIEW

9

Gill, D. M. (1975). Involvement of nicotinamide adenine dinucleotide in the action of cholera toxin in vitro. Pror. Natl. Acad. Sci. U.S.A. 72, 2064-2068. Kimura, N., Nakane, K., and Nagata, N. (1976). Activation by GTPof liveradenylate cyclase in the presence of high concentrations of ATP. Bi{Jchm.Biophvs. Res. Commun. 70, 1250- 1256. Lad, P. M., Nielsen, T. B., and Rodbell. M. (1980). A probe for the organisation of the P-adrenergic receptor-regulated adenylate cyclase system in turkey erythrocyte plasma membranes by the use of a complementation assay. FEBS Lett. 122, 179-183. Orly, J., and Schramm, M. (1976). Coupling of catecholamine receptor from one cell with adenylate cyclase from another cell by cell fusion. Proc. Nail. Acad. Sci. U.S.A. 73, 4410-4414. Pfeufer. T., and Helmreich, E. J . M. (1975). Activation of pigeon erythrocyte plasma membrane adenylate cyclase by guanyl nucleotide analogues and separation of a nucleotide binding protein. J . B i d . Chem. 250, 867-876. Robinson, G. A., Butcher, R. W., and Sutherland, E. W. (1967). Adenyl cyclase as an adrenergic receptor. Ann. N.Y. Acad. Sci. 139, 703-723. Rodbell, M., Lin, M. C., Salomon, Y.,Londos, C., Harwood, J . P., Martin, B . R., Rendell, M., and Berman, M. (1975). Role of adenine and guanine nucleotides in the activity and response of adenylate cyclase systems to hormones. Evidence for multisite transition states. Adv. Cyclic Nucleotide Res. 5, 3-29. Ross, E. M., Howlett, A. C., Ferguson, K. M . , and Gilman. A. G. (1978). Reconstitution of a hormone-sensitive adenylate cyclase activity with resolved components of the enzyme. J . B i d . Chem. 253, 6401-6412. Schulster, D., Orly, J., Seidel, G., and Schramm, M. (1978). lntracellular cyclic AMP production enhanced by a hormone receptor transferred from a different cell. J . B i d . Chem. 253, I201 - 1206. Schramm, M., Orly, J., Eimerl, S . , and Komer, M. (1977). Coupling of hormone receptors to adenylate cyclase of different cells by cell fusion. Nature (London) 268, 310-313. Tolkovsky, A , , and Levitzki, A. (1978). Collision coupling of the P-adrenergic receptor with adenylate cyclase. In “Hormones and Cell Regulation” (J. Dumont and J. Nunez, eds.), Vol. 2, pp. 89-105. North-Holland, Amsterdam. Welton, A. F., Lad, P. M., Newby, A. C., Yamaniura, H., Nicosia, S . , and Rodbell, M. (1977). Solubilisation and separation of the glucagon receptor and adenylate cyclase in guanine nucleotide sensitive states. J . Biol. Chem. 252, 5947-5950. Welton, A. F., Lad. P. M., Newby, A. C., Yamamura, H.. Nicosia, S . , and Rodbell. M. (1978). The characteristics of lubrol solubilised adenylate cyclase from rat liver plasma membranes. Biorhim. Biophys. Acra 522, 625-639.


CLlRRENT TOPICS IN MEMBRANES AND TRANSPORT, VOLUME 111

The Elucidation of Some Aspects of Receptor Function by the Use of a Kinetic Approach A.

M . TOLKOVSKY

Department of Pharmacology Hadassah Medical School The Hebrew University Jerusalem, Israel

............................. r Theory. . . . . . . . . ....... A. The Basic Kinetic Formulation: One-Step Model .................. B. The Separation of Sensation and Function . . . . . .................. e Cyclase . . . . . . . . . . C. Guanylnucleotides and the Rate of Activation of 111. Applying Kinetic Theory to Data Generated by Turkey Erythrocycte Adenylate Cyclase ............................................. A. Methodological Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Slow Activation versus Rapid Equilibrium in a Precoupled System.. . . . . . . . . . C. Slow Activation by a Sequential Mechanism: The Collision Coupling Model. . . D. The Role of GppNHp as Modulator and the Position of Subunit in the Activation Pathway.. . . . . . . . . . . . . . . . .

.......................................... List of Symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.

I1 13 13 15 20 22 22 23 29 35 40 42 43

INTRODUCTION

The purpose of this article is to describe what we have learned in the past few years about the mechanism whereby hormone receptors and guanylnucleotides cause the activation of adenylate cyclase, by applying a kinetic approach. What use, if any, are kinetic analyses of complex biological phenomena? Kinetic analyses enable one to come to terms with quantitative aspects of phe11 Copyright 0 1983 by Academic Press. Inc. All rights of reproduction in any form reserved. ISBN 0-12-153318-2

12

A. M . TOLKOVSKY

nomena which evolve with time. They also provide a framework for creating molecular models which yield predictions concerning quantitative aspects of temporal events. It is fundamental to the kinetic approach that such predictions be subjected to experimental tests. A prediction which does not provide the means to test its own validity is of no value or consequence. As such, kinetics define and predict the behavior of a molecular mechanism of action. The power of refutation, a fundamental component of kinetics, lies in the hand of the experimentalist. The particular problem of the activation of adenylate cyclase by hormone receptors belongs in the realm of signal-response relationships. In the past few years, the chemical models which pertain to the adenylate cyclase response system have become much more complex than the first model of receptorresponse coupling as formulated by Clark (1937). This development did not occur in one step. Structural and functional studies which were rapidly forthcoming from many laboratories (Rodbell et af., 197 1; Bourne et al., 1975; Orly and Schramm, 1976; Neer, 1976; Ross and Gilman, 1977; Ross et al., 1978) provided strong evidence for the multicomponent nature of the hormone receptor adenylate cyclase system. These studies served as conceptual catalysts, which forced the kinetic studies to seek out and formulate more complex models. New structural entities had to be integrated into old models. As a result, new modes of coupling were devised, and tested experimentally. The experiments in turn suggested the participation of new structural or isomeric entities in the activation process. This developmental, iterative process is the essence of the methodology used in the kinetic approach. In this artkle I have chosen to trace the interplay of facts and concepts which has led to the stepwise reapplication of more and more complex kinetic formulations to the problem of adenylate cyclase activation by receptors. Perhaps this description will promote the understanding of kinetics on a more intuitive basis and shed some light on the unique solutions which kinetics can provide for the study of signal-response coupling. I also hope that this will create a feel for the valid use of theory in presenting new and valuable experimental approaches which can be used to derive facts concerning signal transmission from the outside environment into the cell. Finally, a word about paradigms. Some years ago, Kuhn (1970) advanced the idea that the development of a particular field of science occurs by the periodic reformulation of a leading paradigm or a shared example. I wish to pay homage to what may be identified as a paradigm in the receptor field-the hypothesis of the mobile receptor (Cuatrecasas, 1974a,b). This paradigm, I believe, has probably been the most influential element in creating a new conceptual framework for relating the kinetic approach to the problem of adenylate cyclase activation by hormones.

13

KINETIC APPROACH TO THE STUDY OF RECEPTOR FUNCTION

II.

SIGNAL-RESPONSE COUPLING AND RECEPTOR THEORY

A. The Basic Kinetic Formulation: One-Step Model The kinetic approach to the study of receptor function and to receptor-response coupling originates in the formulation by Clark (1937) and by Ariens (1954) of the idea that the response of a biological system to a ligand is proportional to the occupancy of a receptor by that ligand. This statement is formally equivalent to the following scheme: L

I

u

+ R S LR + response 7

(1)

in which L is the ligand, R the receptor, LR the ligand-receptor complex, 1 the ligand receptor association rate constant, 2 the ligand receptor dissociation rate constant, and a a proportionality constant between LR and the response. What constitutes a quantitative solution to this formal statement depends on our ability to express the magnitude of the response in terms of the components which cause its creation. In this case the components are L,, the total ligand concentration, R,, the total receptor concentration, and the rate constants, 1 , 2, and a . For example, when equilibrium between L and R is achieved and L, = Lfree the following expression is obtained ( t is time):

response

= dRTtl(211

+ L)

(2)

Three parameters are sufficient to characterize the behavior of this system: the equilibrium constant 2/1, the total receptor concentration R,, and the proportionality constant a. The ability to formulate an idea and to translate it into a chemical statement which defines and predicts the temporal nature of events which ensue from the moment of signal initiation to the final point at which a response is generated establishes the fundamental principles and boundaries of the kinetic discipline as relating to receptor function. 1 . TESTINGTHE FORMULATION AND

FOR

HIDDEN ASSUMPTIONS

LIMITATIONS

One then asks, does this model provide a sufficient foundation upon which to build a fundamental molecular mechanism of action which both defines and predicts how a ligand may cause a response? Is this formulation exclusive and complete? On purely kinetic grounds the answer is yes. On conceptual grounds the answer is no. A choice exists between these two answers. Choosing involves

14

A. M. TOLKOVSKY

(1) understanding and accepting all the assumptions of the model, those upon which the formulation is based, and those which are consequential to the basic hypothesis (often elusive and hidden from cognitive assessment) and (2) accepting the limitations which are set by the assumptions. Thus, the foundation provided by Eq. (1) is sufficient and complete in sofur as one has chosen to formulate a prediction about the nature of response systems which concerns only L and R, on the one hand, and the response, on the other hand. What are the assumptions that were made? The formulation of this model was founded on the principle that the response is proportional to the concentration of LR. But the model also contains an implied assumption, which stems from the nature of the chemical statement: that the receptor molecule is not only the acceptor of L but also the exclusive physical vehicle for the response which it creates. Figuratively speaking, it would mean that the acetylcholine receptor would also be the carrier of ion fluxes, that the insulin receptor would also be the glucose transport system, and that the P-adrenergic receptor would also be an adenylate cyclase. What limitations are therefore imposed? The simple, one-step model limits the extent to which the temporal mechanism of informational transfer from the receptor to an effector or response-carrying molecule can be examined, because it simply does not consider, nor describe in physicochemical terms, the separate existence of any of the molecules which participate in the response. Also, because of the underlying assumption regarding the role of the receptor in generating the response as well as the signal, the nature of a is obscure. Is it really the rate constant which defines the rate at which the response is generated’? If so, is this the rate-limiting step which one is interested in pursuing? 2, THE VALUEOF THE BLACK Box APPROACH TO RECEPTORFUNCTION What then is the strength and the weakness of Clark’s model? The strength of the formulation lies in its power to distill the idea of receptor-response coupling to a point where a pure and focused conceptual framework is obtained. This model thus becomes the most abstract expression of the idea of receptor-response coupling. All the events which occur between the input and output are buried in a black box. As such, this model has had great impact on our current understanding of ligand-receptor interactions and receptor function. From the abstract, probabalistic nature of a,the concept of efficacy was derived to explain partial agonist action (Stephenson, 1956). The major conceptual weakness of this model lies in its failure to account more precisely for an entity which carries the functional properties of the response system. Thus, in order to arrive at chemical statements of consequence concerning the separate, sequential nature of signals and responses, a black box approach


15

is insufficient. Models must tie formulated which will allow functional entities to be separate.

B. The Separation of Sensation and Function The separation of sensation and function to two distinct molecular entities was first suggested by Cuatrecasas (1974a,b) in the form of the mobile receptor hypothesis. This hypothesis, in which it was proposed that a nonstoichiometric relationship exists between receptors which activate adenylate cyclase and the catalytic moiety, arose in response to newer concepts in membrane biology which were advanced by Singer and Nicolson (1 972) that the membrane is a fluid solvent of finite dimensions in which proteins are free to move laterally in the plane of the membrane. The lateral nature of the membrane matrix, on the one hand; the perpendicular nature of the information whose signal originated on the outside of the cell but the response to which was generated inside the cell, on the other hand; and the concept of separate molecules which contain sensory and response roles were the basis of a paradigm concerning receptor function which has had a great impact in the field of adenylate cyclase research. The reason that the hypothesis of the mobile receptor was first quantitatively applied to research on receptor-adenylate cyclase coupling (and not to membrane receptors whose physical and structural characterization was much more advanced, such as the nicotinic acetylcholine receptor) was probably due to three factors. (1) The identity of the first biochemical event following receptor occupancy was thought to be most clearly defined, namely, CAMPelevation. Therefore, the molecular formulation found fertile grounds in the reciprocal, biochemical nature of the adenylate cyclase response system. Allosteric properties and enzyme kinetics could be adapted to this system with ease. (2) The output of this system was slow enough to be easily measured under the same conditions as one could measure ligand binding. Faster response systems, such as ion gating responses to acetylcholine, and the quanta1 nature of this system left the single ion channel firmly attached to the acetylcholine receptor (Werman, 1975). (3) A qualitative experiment was designed and executed by Orly and Schramm (1976) which established that the catalytic units belonging to Friend erythroleukemiacells could be coupled to P-adrenergic receptors belonging exclusively to turkey erythrocyte cells. This experiment of cell fusion provided tangible evidence for the concept of the mobile receptor by crossing experimental techniques which belonged to the discipline of membrane biology with biochemical techniques of enzyme catalysis. 1. FORMULATION OF THE MOBILERECEPTOR HYPOTHESIS

The actual formulation of the mobile receptor hypothesis in molecular terms was presented by Jacobs and Cuatrecasas (1976) in the following form:

16

A. M. TOLKOVSKY

LRE

response

in which E generates the response. As observed by Jacobs and Cuatrecasas (1976), the outstanding implication of this model is the prediction of a nonlinear relationship between ligand binding as a function of ligand concentration and the response generation as a function of ligand concentration. This model also predicts that the binding function of L has a complex pattern from which a highaffinity-like site and a low-affinity-like site can be derived. The two classes of sites which emerge result from L binding to two species of the receptor, R and RE, and from the random, boxlike sequence of interactions which occurs. The response is found to be proportional to the high-affinity site alone, thereby dissociating the binding behavior from the response. This model constitutes the most general formulation of the mobile receptor hypothesis, barring additional subsidiary hypotheses concerning state transitions of the receptor from R to R‘ and of the catalytic unit from E to E’ as proposed by De Haen (1976).

2. TWO-STEPSEQUENTIAL MODELSOF THE MOBILERECEPTOR

A critical analysis of this model entails simplifying it to limiting cases. Boeynaems and Dumont (1975, 1977a,b) actually preceded the most general formulation presented above by suggesting in its stead three separate linear sequences of L to R to E interactions, all of which could be applied to account for adenylate cyclase activation by hormones. Upon inspection, these models turn out to be special cases of the general mobile receptor formulation. The three models are L

+ RE

LRE

response

(4)

KI

L t R

s LR

+ E e LRE Ir, response

KI

L

(5)

K2

+ RE e LRE s LR + E 5 response Kl

K?

In Eq. (4) R and E are always in a complex, and only LRE can generate a response. Kinetically, it is a degenerate, overformulated form of Clark’s model. In Eq. (5) R and E are separate and the union of LR with E to form a complex LRE gives rise to a response. In Eq. (6) RE is preassociated but E is inactive.

BINDING A N D RESFWNSE FUNCTIONS Ok ONE-STEP Function

Binding

TABLE I A N D TWO-STEP S E Q U E N T I A L RAPID EQLJILIERIUM MODELSO Response

L + R E S LRE Kl

LRE 3 response

LR

+E

LRT LRE = _ _ L + Kl

C LRE KI e,

2

,

C LRE K,

+

4 5

+ RE

g

L

2 response

E

LRE

W

E 1;response The solutions to the two-step models are approximations of square root functions, derived by expansion into a power series. Only the first terms were taken. Concentrations are designated by italics.

18

A. M. TOLKOVSKY

The binding of L causes a dissociation, producing an active form of E which generates a response. In contrast to the general case, each of the limiting cases allows only one binding step involving L, similar to the one-step model. Can such simple models, in which in addition to the ligand binding step a second step occurs which involves either the association or the dissociation of R and E, generate patterns of behavior which deviate from the one-step model? If so, there may be room to consider such two-step models in the testing of the mobile receptor hypothesis. Therefore, one would also like to know whether these two-step models are at all distinguishable from each other. In order to answer these questions, it is necessary to solve the equations in terms of binding and response patterns as a function of the parameters L, K , K,, R,, and E,. Approximate solutions to all three models are presented in Table I. The binding and response patterns as a function of L are also illustrated in Fig. 1, for the particular case where (R,) = (E,) = K, = K , = 1. Complete simulations are provided by Boeynaems and Dumont (1977a,b). From inspection of the formulations presented in Table I one can conclude that, indeed, each sequential two-step model predicts a different pattern of bebavior. The dependence of the response on the concentration of the ligand is different. Also, the binding patterns are different from the response patterns for each model. In each case, the terms which include the species E, and R , are instrumental in determining the differential patterns of each model. To what factor can this difference between models be ascribed? The answer does not entail structural arguments. All three models were set up to contain the

binding

response

v/F

B/F

111

I

V

B

FIG. 1 . Theoretical binding and response curves of one-step and two-step sequential rapid equilibrium models. Models which are presented in Table I are shown as Eadie plots ( v . rate of response output; F , free L concentration) or Scatchard plots ( E , bound ligand; F . free L concentration) in the particular case where K , = Kz = ET = RT = 1 . L concentration varies from 0.2 to 50 K units. I, Precoupled; 11, sequential association; 111, sequential dissociation.


19

same number of components. The functional roles of the components are identical. Also, the models share similar combinations of structural elements. Therefore one must conclude that it is not structure alone which determines a specific pattern of signal-to-response coupling. Rather, the different patterns of behavior stem from the fact that different sequences of events, and a different number of steps, ensue from the moment of signal initiation (concomitant in all models with L occupancy by definition) to the point at which the response is generated. Each sequence creates new, transient structural entities, providing the system with unique properties. Since each system is composed of sequential events, a kinetic analysis will distinguish between these models. Kinetics will therefore allow a glimpse into functional relationships between predefined structures. A structural analysis will, at best, establish their existence.

3. ASSUMPTIONS AND LIMITATIONS OF THE SEQUENTIAL MODELS: THE RAPIDEQUILIBRIUM The sequential models depart from the one-step models in two important respects. 1. The ligand has changed its role. In the one-step models the role of the ligand is to initiate a signal and a response. The nature of the signal vis-h-vis the response in molecular terms remains obscure. In the sequential models, the role of the ligand is to alter the affinity between R and E. This in itself gives the concept of efficacy a sound physical basis: each ligand would have its own potential to alter the affinity between R and E. Thus, while cx controls the efficacy parameter in the Stephenson formulation (Stephenson, 1956) in the twostep models the efficacy will be expressed by a term which involves the affinity constant between R and E. 2. The mere separation of sensation and function defines and allows one to focus on an event which follows ligand binding but which occurs prior to response evolution. We shall term this event coupling. Coupling in its most elementary form may occur by more than one pathway. In fact, by assuming one coupling step, three minimal models replace the singular formulation of the precoupled systems. The multiplicity is a necessary consequence of the added degree of functional freedom we have allowed the system. Each of these models displays unique properties. Each could account for events which occur subsequent to the binding of the ligand to the receptor.

Thus, the foundation provided by the sequential models is almost sufficient to begin to examine experimentally the evolution of coupling events. In one important respect, though, the sequential models still impose a barrier to the elucidation of coupling phenomena. The barrier stems from the fact that all these models bury the temporal nature of the coupling event beneath a rapid equilibrium assumption.

20

A. M. TOLKOVSKY

The rapid equilibrium assumption implies that the mechanism of coupling between R and E is not time consuming compared to the slowest step in signal generation. Under rapid equilibrium conditions between L and R and between LR and E, the rate of the slow step is still controlled by a. In essence, the sequential and the precoupled models propose a concerted mechanism of action: the binding event of L is synonymous in time with the formation of LRE and RE. We shall illustrate this with an example: In the mechanism L t R

e LR KI

+ E s LRE K:

u

response

(7)

when K , is small relative to L and R , and K , is large relative to LR and E , most of the receptor is in the LR form. The binding of L is almost exclusively to the LR form, and is therefore controlled by K , . The response, on the other hand, is still generated at a rate defined by a and by the extent of LRE which is formed. Thus, although LRE takes up a negligible portion of R, its formation is crucial for the response to occur. The formation of LRE occurs concomitant with the binding of L, but it cannot be probed by L since it does not accumulate. The coupling step is buried and lost. In summary, molecular models which allowed the separation of sensation and function paved the way for focusing on coupling events. Understanding the coupling process would lead to an appreciation of crucial aspects of the mechanism whereby a receptor causes a response, namely, functional interactions. At the same time equilibrium assumptions somewhat destroyed the potential of sequential models to enable the elucidation of the coupling process because they converted an inherently dynamic process which occurs sequentially in time into a static, concerted process. In some respects, then, sequential models were still not providing a good enough foundation for the study of signal-response coupling. In practice as well as in theory, rapid equilibrium had to be dissected further into rate processes.

C. Guanylnucleotidesand the Rate of Activation of Adenylate Cyclase For many years there was no experimental evidence in the field of adenylate cyclase research which could be applied to contradict the assumption of a concerted mechanism, namely, that a hormone-receptor adenylate cyclase ternary complex accumulated prior to the generation of CAMP. The first experimental evidence to the contrary was provided from observing the effects of guanylnucleotides on adenylate cyclase activation (Rodbell ef al., 1971). GTP was shown to elevate the catalytic activity of adenylate cyclase synergistically with the hormone (Salomon el af., 1975). Rodbell (1975) then proposed that


21

GTP works by binding to a subunit which he termed a transducer molecule. The concept of a guanylnucleotide-sensitivetransducer molecule which was interposed between R and E upset the foundation provided by the sequential models as pertaining to a molecular mechanism of receptor coupling to an effector as applied to adenylate cyclase research in three basic respects. 1. It upset the contention that the RE complex needs to accumulate (or dissociate) in order to make E active. It suggested, on the contrary, that LR and E need not ever accumulate. The transducer molecule could conceivably shuttle between the two entities. Thus, E would become a second effector in the signal-response system of adenylate cyclase. Its activation would be a product of tbe primary interaction between R and N (N will be used throughout to represent the guanylnucleotide subunit), the transducer molecule. As we shall see, the level of complexity in formulating such models increases considerably, just as when the separation of R and E replaced one model with three alternative models. 2 . It upset the basic hypothesis that the rate of formation of the active form of the catalytic unit is rapid. In the presence of guanylyl imino-py-diphosphate (GppNHp), a nonhydrolyzable analog of GTP, cAMP accumulation was no longer a linear function of time. Rather, cAMP now appeared with a considerable lag time. Thus, cx could not be the determinant of the rate-limiting step. The lag time in AMP accumulation in the presence of GppNHp appeared in three very different experimental systems: in the P-adrenergic receptor-activated system of turkey erythrocytes (Sevilla et al., 1976) and the P-adrenergic receptor-activated system of the frog erythrocyte (Schramm and Rodbell, 1975); in the glucagon receptor-activated system of liver membranes (Salomon ef af., 1975); and in the P-adrenergic receptor-activated system of S49 lymphoma cell mernbranes (Ross et a/., 1977). The fact that the slow activation phenomenon was shared by different systems implied that it may have some fundamental mechanistic significance. 3. The concept of a guanylnucleotide-activated transducer molecule also upset the contention that the role of L is to allow a tighter or looser association between R and E and thereby to amplify the extent (or concentration) of the active form of the catalytic unit. In contrast, early studies by Rodbell (1975) suggested that the role of the activating ligand L is to promote a faster rate of appearance of the active form of E. The extent of amplification of catalytic activity seemed to be controlled by the type of guanylnucleotide used.

The use of GppNHp as an experimental handle in probing receptor-adenylate cyclase relationships changed the nature of the simple, sequential two-step formulations even before they were applied rigorously to test the mechanism of adenylate cyclase activation by hormones. Instead, new fonnulations which departed from rapid equilibrium assumptions were introduced (Sevilla et a/.,

22

A. M. TOLKOVSKY

1976; Ross et al., 1977; Tolkovsky and Levitzki, 1978a,b). These formulations now contained an irreversible step whicb defined the slow transition from an inactive to an active form of E following receptor-effector complex formation. It is this step which now assumed the identity of a coupling event. For example, even the precoupled mechanism L

+ RE + LRE 5 response

(8)

could be formulated into an irreversible coupling-containing process such as L

slow + RE + LRE + LRE’ + response K ( I

(9)

The irreversible nature of this slow step was based on experimental evidence which suggested that once the enzyme was activated in the presence of GppNHp, it remained in an active form even after washing or in the presence of an antagonistic ligand which is bound to the receptor (Schramm and Rodbell, 1975). By introducing an irreversible step, an added level of rigor in quantitation of temporal phenomena in relation to adenylate cyclase was established. In practice, defining the rate law by which such activation is achieved and defining the parameters which control the rate of activation enabled the examination and testing of the mechanism of adenylate cyclase coupling to a P-adrenergic receptor and to an adenosine receptor in the simple system of the turkey erythrocyte membrane.

111.

APPLYING KINETIC THEORY TO DATA GENERATED BY TURKEYERYTHROCYTEADENYLATECYCLASE

A. Methodological Considerations The historical survey in the preceding sections is a biased description of some of the developments which occurred in the field of adenylate cyclase research presented in conjunction with receptor theory. It was intended as a developmental exposition of concepts which have led to the application of kinetic theory and terminology as tools for the elucidation of the mechanism of activation of adenylate cyclase by receptors. We have seen that in order to analyze the kinetics of coupling events in relation to the activation of adenylate cyclase by hormones one wants to (1) define a molecular mechanism of action, (2) convert it to a statement which includes all of the species which participate in the response pathway, (3) allow coupling events to be separate from signaling events on the receptor, and (4) eliminate rapid equilibrium assumptions, replacing them with rate constants unless a rapid equilibrium can be verified experimentally.


23

In order to test such modcls experimentally one needs to (5) evaluate critically all the implications of each model, (6) look for vulnerable points which make the model conflict with a simpler version of itself, (7) use conflicts in order to test the models against the experimental system, and (8) reject inconsistent models and reapply processes (1) to (7) to an extended form of the model which is most consistent with the data, by altering one variable. This process is theoretically complete at a point where the most extended model recapitulates precisely all the predictions of the previous model. Such a protocol will be followed in this section. We will first formulate and compare the simplest rapid equilibrium model with its congener version in which a slow activation step occurs. The salient features of each model are listed in Table 11. A simulation is presented side by side in Fig. 2. We shall then set up discriminatory criteria to distinguish between both models. These criteria are also listed in Table 11. The resultant functions are portrayed in Fig. 3. Data derived from an analysis of the activation of turkey erythrocyte adenylate cyclase by an adrenergic receptor in the presence of GppNHp are presented and examined in relation to each model in Fig. 4. It will be shown that all the sequential models in which slow activation occurs after the formation of the complex LRE are not consistent with the data (Table 111). The precoupled slow activation model is the only model to which the data conform. This model is then reformulated by altering one variable-the position of the slow step in the activation process. The new model so derived' (termed collision coupling) is compared with the parent model in Table IV. It will be shown that the new model is indistinguishable from the parent model by the first discriminatory criteria which were established (otherwise it would have to be rejected outright). New rejection criteria to distinguish between the two models are also described in Table 1V as a second level discrimination and are portrayed in Fig. 5. Experiments which were designed to test the new model are shown in Fig. 6 . Four probes are used to test each model: (1) the dependence of the response evolution in time on the concentrations of the interacting species L, R, and E (2) the relation of the binding function to the response function; (3) the dependence of the apparent rate constant which determines the rate-limiting step on the concentrations of the interacting species and on the equilibrium constants; and (4) the rate law which governs the activation process. Detailed comment regarding each figure and table follows. 6. Slow Activation versus Rapid Equilibrium in a

Precoupled System Table 11 compares the mechanisms depicted in Eq. (8) and (9) in terms of the response evolution and the intermediate accumulation. The mathematical solutions to the molecular formulations and the general features are also listed. Four

24

A. M. TOLKOVSKY

features, all inherently kinetic, make these models distinguishable from each other: ( I ) the pattern of LRE or of LRE‘ accumulation, or of the response output as a function of time, (2) the levels of LRE and of LRE’ at infinite time, (3) the rate law of LRE or of LRE’ accumulation, and (4) the patterns of each feature (1)-(3) as determined by the ligand concentration and by K , the dissociation constant of LR. These are illustrated graphically in Fig. 2. In Fig. 2A and B the accumulation of the response as a function of time at different ligand concentrations is described. The rapid equilibrium model (A) predicts a linear accumulation of response at all ligand concentrations. The slow activation model (B) predicts that the response will accumulate with a lag time. The lag time will increase when the ligand concentration is reduced. At long times the response becomes independent of ligand concentration, and all the slopes are equivalent. The response reflects the extent of the active enzyme which has been formed (Fig. 2C and D). In the rapid equilibrium model the active form of the enzyme was established instantaneously. Therefore the extent of active enzyme does not change as a function of time at any ligand concentration (C). It is this behavior which generates (upon integration) the linear response output in Fig. 2A. In the slow activation model, the time-consuming reaction is the formation of active enzyme (D). Since each active enzyme molecule will produce a response output as function of time, the response is generated in a cooperative rather than in a linear manner (B). At long times, say 10 half-lives, all the active enzyme has been formed. Therefore, independent of ligand concentration, the response output is linear and maximal. The behavior of L is determined largely by the magnitude of its dissociation constant. How is the dissociation constant for L derived from such data? This is shown in Fig. 3. If the slopes of the response output (which are equivalent to response units divided by time, or normalized response units) are plotted as a function of L, a saturation function for L (Fig. 3A) is obtained. In the case of the rapid equilibrium model, because response output is linear with time, one such function will be generated regardless of the time at which the data were collected (A, dashed line). This function can then be linearized ( C , dashed line). The slope of the linear plot is equal to the negative value of the affinity constant and so K, the dissociation constant, is obtained. In the case of the slow activation model, each time point at which the value of the normalized response units as a function of ligand concentration is calculated will yield a different curve (A). The slopes of the linear expressions of these curves ( C ) are seen to change as a function of time. But, if the slope is always K , it should remain invariable, since K is a constant and not subject to change in time. Since that is not the case, the slopes are not K, and therefore such plots cannot be used to derive K for a mechanism involving slow activation. On the other hand, precisely for this reason such plots can be used as a very sensitive probe for nonlinear response output.

TABLE 11 A COMPARISON OF THE ONE-STEP RAPIDEQUILIBRIUM MODELWITH Mechanism

ONE-STEP SLOW ACTIVATION MODEL"

Rapid equilibrium

+ RE

Formulation

L

Concentration of intermediate

LRE=

Response evolution

THE

LRE

5 response

RETL K+L (Fig. 2C)

dETL K+L (Fig. 2A)

Slow activation

L

+ RE e LRE J, LRE' 5

LRE' = RET{~- exp[-sLr/(K

response

+ L)])

(Fig. 2D) d E ~ {exp[-sLr/(K + L)I - 1) sLI(K + L ) (Fig. 2B)

dET1 + Discriminating criteria

Accumulation of LRE or LRE'

So rapid that it is fully formed at zero time

Slow, rate determined by s, L, and K (Fig. 2D)

(Fig. 2C) Levels of LRE or LRE' at infinite time

Equivalent to zero time levels (Fig. 2C)

One common maximal level equivalent to RET (Fig. 2D)

Rate law of LRE or LRE' accumulation

Zero order

First-order, slope of semilogarithmic plot gives sL/(K + L) (Fig. 3D)

Measurable factors which are determined by L and by K

LRE (Fig. 2C) Slope of response output (Figs. 2A and 3A)

Extent of LRE' before r is infinite (Fig. 2D) Lag time in response output (Figs. 2 9 and 3B-D)

A simulation of each mode! is presented in Figs. 2 and 3. These are indicated in parentheses

A.

26

60

K=l

-c "7

L

60-

K-1

M.TOCKOVSKY

L

-

a 0 y1 0

-

n

T I M E , min

Fic. 2. Theoretical response output and active enzyme accumulation curves of a one-step rapid equilibrium model compared to a one-step slow activation model. The behavior of each system is shown as a function of ligand concentration. Ligand concentration is given in K units. K is the LRE dissociation constant. (A, C) Rapid equilibrium; ( B , D) slow activation. Details are given in Section

111,B.

In order to derive the dissociation constant in this case, the fractional activation of the enzyme, which is a calculated entity E',/EfT, can be plotted as a function of time in a sernilogaritbrnic plot (Fig. 3D). Such plots are linear because the saturation function for L appears in an exponential expression (Fig. 2D). When the slopes so derived are replotted as a function of ligand concentration (Fig. 3B) a curve which is identical to the dashed curve in Fig. 3A is obtained. The ordinate of this plot (Fig. 3B) no longer describes the normalized response units as in Fig. 3A. It describes kobs, or the values of the slopes derived from the semilogarithmic plots. Thus, at infinite ligand concentrations a measure of total active enzyme concentration, R&, is not obtained. Instead one obtains the value of the rate constant for the conversion of LRE to LRE',, a (Table 11). The half


27

FIG. 3. A critical analysis of the one-step slow activation model. The slow activation model (Fig. 2D) yields nonequivalent patterns of enzyme activation (E‘,/E’T)as a function of L at different times (A). In contrast, thc pattern of enzyme activation as a function of L which is predicted by a rapid equilibrium model is indifferent to time (dashed line). This function yields a linear Scatchard plot (C). A slow activation model predicts linear semilogarithmic plots (D). (D) is actually a replot of the - E‘,)/ErTat different L concentrations lines in Fig. 2D. The slopes of these lines which depict are the values of k,,,,, characteristic of each ligand concentration and are used to construct a kobs versus L plot (B). Although this curve is identical to the dashed line in (A), the ordinate units of each plot are different: at infinite time, E’, + ElT in (A), but kobs + a in (B). Thus one can derive the values of EIT and of a , the rate constant for activation. In addition to the values of E’T or a. the parameter which is derived from (A) and (B) is K . the ligand dissociation constant. K can be derived from the slope of the Scatchard-like plot shown in (C). Note how time causes a change in the apparent slope of the lines which are derived from the curves depicted in (A) and how this is rectified in (B). Further details are given in Section 1II.B. EIT. total active enzyme concentration at infinite time; E’,, active enzyme concentration at time I ; E’,/EIT,fraction of active enzyme; (E’T-E’,)IE’T. fraction of inactive enzyme. (ElT

saturation constant derived from this function is now equivalent to the dissociation constant. Figure 4 shows data generated by following the activation of adenylate cyclase of turkey erythrocytes at various L-epinephrine concentrations, in the presence of 0.1 mM GppNHp. The data describing CAMP accumulation as a function of epinephrine concentration and time were collected at 25°C (Fig. 4A). The lines

28

A. M. TOLKOVSKY B

A

4

I

c

K.37uM

I

min

I

1

60 80

0

0

I

1

120

rnin

.3.

- .2 c E

;.l-

_*

0I

i

5

0 L-epinephrine. pM

FIG.4. Activation of turkey erythrocyte adenylate cyclase by L-epinephrine in the presence of GppNHp. Activation is at 25OC (A) or at 37°C (B-D). In all the experiments 0.1 mM GppNHp is present. In (A), the accumulation of cAMP is shown as a function of time and hormone concentra3 @I; 0 , 15 p M . The lines all fit to one equation (Fig. 2B). The tion. A, 0.5 pM; A,1 @I; 0, parameters derived from fitting the data to the model are shown in the figure. Note curvature. In (B), data concerning the accumulation of the E’ form is shown at various hormone concentrations: 0 , O .I phf; 0.0.4 @I; A, 1 pM; 5 pM; 17,100 pM. Linear replots are shown in (C) from which the values of k&s are derived and used to construct the lines which are drawn into the data in (B): 0, 0.035 min-I; 0 , 0.066 min-1; A, 0.154 min-I; A,0.245 min-I. Each slope is dependent on hormone concentration (D). From the kobs versus L curves one can derive the values of K ,the ligand dissociation constant, and a, the rate constant for the activation of E to E’.

A,

through all the points were derived by fitting the data to the function which was used to construct Fig. 2B. The following parameters are obtained: REi-,., the maximal catalytic activity (Vmax); a, the slow-step rate constant; and K , the dissociation constant. The rest of the data (B-D)were collected by initially incubating the membranes with epinephrine and GppNHp. At each point in time a sample was removed into propranolol (a potent 6-adrenergic receptor antagonist). This stops the activation process instantaneously. The extent of active enzyme is then measured by following cAMP accumulation for a fixed amount of time (Tolkovsky and Levitzki, 1978b). The cAMP will evolve linearly.


29

In Fig. 4B an entire time course of enzyme activation was collected at different hormone concentrations. Each curve was linearized by the semilogarithmic plot (Fig. 4C). The slopes derived from the semilogarithmic plot are apparent rate constants, kobs. The calculated values of kohs (each slope in Fig. 4C) are given in the legend to Fig. 4. Each curve was constructed using this value and the equation given above Fig. 2D. The values of kobs are also replotted as a function of hormone concentration in Fig. 4D. The curve through the points was theoretically derived from fitting the data to the following expression, which relates kohs to L and to K (Fig. 3B): kohs

=

0.295 X L K + L

where 0.295 min- I is the value of a. At this point one would conclude that the activation of turkey erythrocyte adenylate cyclase by epinephrine and by GppNHp is consistent with the precoupled slow activation model. There is no instance in which the precoupled rapid equilibrium mechanism could be shown to conform to the data. On the other hand, there are also no data which are consistent with similar two-step slow activation models (Table 111; Tolkovsky and Levitzki, 1978b, 1980). Were the mechanism of adenylate cyclase activation related to the two-step sequential models as formulated in Table 111, linear semilogarithmic plots would not have been observed when the fractional activation of the enzyme as a function of time was examined. Also, the binding of the hormone was observed to be simple and not apparently negatively cooperative, as these models would predict. In addition, the binding was also completely independent of the state of the enzyme.

C. Slow Activation by a Sequential Mechanism: The Collision Coupling Model The conclusion that the P-receptor and the enzyme are tightly coupled did not fit with results of the fusion experiments which showed that the turkey erythrocyte P-receptor will form an active complex with a foreign catalytic unit. But none of the sequential models could account for the activation data either. This disturbing conflict made us seek a model which would accommodate the idea that the receptor need not be precoupled tightly to the enzyme, and yet remain consistent with the data. In designing this model two criteria of consistency had to be fulfilled (this is the first level of discrimination): (1) Simple binding of the ligand. The binding of the ligand must be governed by the same constant as the half saturation constant which is derived from the activation data. (2) First-order activation. In order to allow the model an added degree of freedom, we added the assumption that the receptor is no longer required to bind to the active catalytic

TABLE 111 GENERALFEATURESOF THE SEQUENTIAL, SLOWACTIVATION MODELS* Formulation

L+R=

LR+E=LRE

Kl

LRE

L + R E = LRE= LR + E Kl

K2

i LRE’ 5 response

E

5

Kl

E‘ 5 response

Binding

Simple (note equivalence in affinity of L to R and RE forms)

Two-component (apparent negative cooperativity )

Two-component (apparent negative cooperativity)

Rate law

Non-first-order (deviates more at low hormone concentrations)

Second-order

Neither first- nor second-order

Effect of reducing ET

Reduces extent of activation

Reduces rate and extent of activation

Reduces rate and extent of activation

Effect of reducing

Reduces rate of activation, not extent



RT

aThe solutions are correct for LT = Lf,,; s is the rate constant governing the slow step


31

unit in order to maintain its activity in the presence of GppNHp. In order to retain first-order behavior we introduced the idea that both LRE and LRE’ maintain a steady-state concentration, similar to the Briggs-Haldane assumption so often used in enzyme kinetics. In effect, we converted the receptor into a catalyst, and obtained a mechanism which contained both an association and a dissociation between R and E: L

+ R+ LR + E .-(LRE-LRE‘) K “ slow

rani

+ LR h

a + E’ + response

(1 1)

Three conditions make this formulation distinct from all the previous two-step slow activation models: (1) LRE and LRE‘ must never accumulate to any significant extent compared to LR. These conditions are met by assuming a to be much smaller than b. (2) LR dissociates from E’. E’ can maintain its activity indefinitely as long as ample GppNHp is present. (3) The slow step is no longer sequential to the formation of the LRE complex; it precedes LRE formation. The catalytic role assigned to the receptor is the most extreme case of nonstoichiometric coupling. R is the catalyst, E‘ is the product of the catalysis, and both LRE and LRE’ are negligibly small compared to LR. Because LR remains attached to E for a very short time, we called this mechanism of activation “collision coupling.” The accumulation of the active enzyme is predicted to occur via a first-order process:

The discriminatory features of this model as compared to the precoupled, slow activation model are presented in Table IV. The rejection criteria focus on the altered role of the receptor, which is reflected in the position of the receptor in Eq. (12) compared to Eq. (9). The way this changes the kinetics of the activation process is illustrated in Fig. 5. The activation of the enzyme as a function of time and as a function of receptor concentration is shown in Fig. 5A and B. By reducing the receptor concentration, the precoupled model predicts a reduced extent of activation (A). The collision ‘coupling model predicts that the rate of activation becomes slower but that the final extent of activation remains unaltered and maximal. Response generation is given in Fig. 5C and D. For both models there is a lag time in response output, because both include an exponential form. For the precoupled slow activation model the final response is proportional to the concentration of the receptor in a linear manner. For the collision coupled model, the kinetics are very similar to those shown for reduced ligand concentration in Fig. 2B: the receptor, like the ligand in that system, increases the lag time of the response, but

TABLE IV A COMPARISON OF THE COLLISION COUPLING MODEL WITH Mechanism Formulation

THE

PRECOUPLED SLOWACTIVATION MODEL"

Recoupled slow activation L

+ RE

LRE

2

Collision coupled

L

LRE'

+ R sK LR + E 2

(LRE ~ L R E ' )

1LR + E' ( b >> a ) First level of discrimination Rate law

First-order

First-order

E' or LRE' at t + =

RET (independent of L)

ET (independent of L)

Reduction in L concentration

Slows rate of activation. depends on LI(K

+ L)

Slows rate of activation. depends on L/(K + L )

Second level of discrimination Apparent rate constant of activation

aL (Fig. 5A) K+L

K + L

(Fig. 5B)

CAMP accumulation (Fig. 5C) Reduction in

RT

Reduction in ET

(Fig. 5D)

Limits extent (Fig. 5A and '.-'

Limits rate of activation, does not limit extent (Fig. 5B and D)

Limits extent (Fig. 5A and C )

Limits extent (Fig. 5B and D)

Simulation of each model is given in Fig. 5, indicated in parentheses.

33


I

I

r

1

1

1

0

C

-

RT

6o

1

r 0

1

I

1

I

60

0

I

1

60

TIME ,min FIG. 5 . Theoretical response output and fractional activation curves of a one-step slow activation model versus a collision coupling model. The behavior of each system is identical with respect to dependence on ligand concentration. The test to distinguish between them i s given by observing the dependence of the activation on the receptor concentration. ( A . 0 Slow activation; (B.D) collision coupling. Note how reducing RT reduces the extent of activation in (A), but reduces the rate of activation in (B). A detailed analysis is given in Section 1II.B.

at longer times all responses will be generated at an equal rate. Reduction of the extent of E will reduce, in both cases, the extent of activation and response. Data to test the collision coupling model were obtained with the aid of an irreversible P-adrenergic antagonist affinity label which was synthesized by Dr. D. Atlas (Atlas et a/., 1976). The treatment of turkey erythrocyte membranes with this ligand caused a reduction in the number of receptors which could bind epinephrine. After treatment with this ligand, adenylate cyclase was activated in the presence of 0.1 mM GppNHp and a saturating concentration of L-epinephrine (0.1 mM). The data were derived as previously described and are given in Fig. 6A. The data were replotted on a semilogarithmic scale and the apparent rate constant was derived (as in Fig. 4C). The value of kobs derived for each curve is given beside Fig. 6A. The lines are theoretical constructions according to the collision coupling equation. The relationship between the maximal number of receptors, remaining after affinity label treatment as determined by a binding

34

A. M. TOLKOVSKY A kobs, r n d

.033 0

,052

A

,094 -217 374

A v

I

1

0

50

I

100

B

t

I

0 'obs

1

I

.4

0

1

i

30

min

FIG.6. Activation of turkey erythrocyte adenylate cyclase by limiting receptor concentration. Data are derived at 37°C. in the presence of 0.1 mM CppNHp and 0. I mM (saturating) L-epinephrine. Each curve in (A) is based on the values of kobs given in the figure. These values are derived from a semilogarithmic plot of the data generated in (A) (like Fig. 4C). Note how similar (A) and Fig. 4B are. This is due to the fact that both RT and the expression LI(L + K ) modify the exponential term in a linear manner (Fig. SB,or Table IV). In (B), the kobsvalues are compared with the RT values. Note the linear dependence. In (C), CAMPaccumulation is shown as a function of ET. Since ET determines the extent of activation, not the rate, one should compare these data with the theoretical behavior depicted in Fig. 5C. The value of a, the rate constant remains unaltered.

experiment, and the measured kobs is given in Fig. 6B. The linear relationship between these two measured entities attests to one further prediction of the collision coupling model: that the rate of activation is dependent on the receptor concentration in a linear rather than in a saturable manner. We also examined the effect of reducing the enzyme concentration by treating membranes with the mercurial PHMB (Fig. 6C). It can be observed that reducing the enzyme concentration altered the extent of activation but had no influence on the rate constant, which remained at a value of 1 min- even after the activity of the catalytic unit was reduced to 32% of its value in native membranes (these results conform to the pattern presented in Fig. 5C). Thus a collision coupling mechanism can account for all the experimental situations encountered so far.

KINETIC APPROACH TO THE STUDY

OF RECEPTOR FUNCTION

35

D. The Role of GppNHp as Modulator and the Position of the Guanylnucleotide Subunit in the Activation Pathway The process of progressive reformulation of simple models in order to account for complex data has gone through one cycle. No doubt, consistency of models with data is the pitfall of the modeling procedure. Analogies are drawn by conjecture. For example, we have seen that data could be generated which were consistent with a simple precoupled slow activation model, in which R is precoupled to E. But, to conclude that R and E are precoupled would have been wrong. The internal consistency between data and the precoupled slow activation model failed upon further scrutiny. Actually, this model would have been rejected independently of structural data had it occurred to the experimentalist to test whether R is precoupled to E by altering the concentration of R independently of the concentration of E. Since slow activation sequential models did not fit the data, it was difficult to arrive at the idea of such an experiment unless one could conceive of a mechanism in which R assumes other than a modifier role. The conclusion is methodological: in testing such models one must formulate, modulate, and probe experimentally all the species involved in the signal response pathway. This then allows one to determine the position and independence of each component. Clearly, then, one serious omission in the collision coupling model remains. This model considered a two-component system in which informational transfer from R to E occurred. At the same time, we know that this information is mediated in an obligatory fashion by the guanylnucleotide subunit. Without GppNHp no activation of E is obtained. Therefore, it is necessary to account for a stable activation of the adenylate cyclase subunit by GppNHp and the hormone receptor, where the receptor, which is the promoter of this process, acts in a catalytic manner. In order to do so it must be assumed that N, the guanylnucleotide subunit to which GppNHp is bound, is the message of the coupling process. When N now is in its active mode, and is coupled to E, the E form of the catalytic subunit is transformed to E’. This is an obligatory process; the act of E to E’ transition is a passive reflection of events which occur in N. What is the molecular nature of the arrangement between R, N , and E which allows this process to occur? How does N assume the role of messenger and message? In order to answer the first of these questions one would like to know (1) whether N is separate from R and from E; (2) whether the hormone receptor causes the activation of N in a ternary complex which involves RNE, which would then dissociate upon activation; (3) whether alternatively R can cause the activation of N in a binary complex, RN; and (4) whether the site for GppNHp exists independently of the state of the receptor, or whether the site is created only upon the interaction between R and N, and therefore GppNHp binding follows the activation of the guanylnucleotide subunit by the receptor. We begin by formulating models in which N assumes an independent role. We

TABLE V ( T H E GUANVLNUCLEOTIDE SLIBUNIT) ASSUMES A N INDEPENDENT ROLP

A COMPARISON OF COLLISION COUPLING MODELSI N WHICHN

Mechanism

NE are always coupled

z

36

-.

N is activated first, it then couples to E (N’ is the message from R to E)

Rate-limiting step

Formulation (A) LR

A

+ NE LR

f

(LRNE

= LRNE’)

Order of reaction second-level test

Rate law

Reducing N and E third-level test

~.LR.NE

In F = -uRTI

First-order

Rate remains constant for anyNorE levels

u*RT.N

In F

First-order

Slows rate because N’ to E becomes rate limiting (Fails secondlevel test) (Fails secondlevel test) (Fails secondlevel test)

N’E’

+ N 5 LRN LRN A LR + N’

(B) LR

N’ + E 1,N‘E’ ( I ) R to N rate iimiting

=

-uR+

Zero-order

(2) N to N’ conversion rate limiting

Complex Zero-order to first-order

(iii) NT > R7

k

I

(iv) NT epinephrine > norepinephrine. In addition, P-adrenergic antagonists should potently compete for the binding of the radioligand, reflecting their ability to inhibit cyclase stimulation. Furthermore, the biologically active (-) stereoisomers of adrenergic agonists and antagonists should be more potent in competing for the binding sites than the less active (+) stereoisomers. The radioligands currently employed to identify adrenergic receptors (see above)

50

JEFFREY M. STADEL AND ROBERT J. LEFKOWITZ

have been shown to fulfill these criteria and therefore label physiologically relevant receptors in a variety of tissues. One of the earliest applications of the radioligand binding technique to the study of P-adrenergic receptors was the investigation of the interaction of a wide variety of unlabeled P-adrenergic agonist and antagonist compounds with the receptors. This is accomplished through competition binding experiments in which the dose-dependent inhibition of binding of a fixed concentration of radioligand is examined. By comparing the affinities of these unlabeled agents for the ligand binding sites with their abilities to modulate adenylate cyclase activity, it could be convincingly shown that agonists and antagonists compete for the same set of P-adrenergic receptor binding sites. The affinity of both agonists and antagonists for the P-receptors is primarily determined by their stereo configuration and the substitutions on the amino nitrogen. These studies also reinforced the notion that agonist “activity” is not simply related to binding affinity, but rather involves additional interactions not triggered by antagonists. Following the validation of the radioligand binding approach for studying adrenergic receptors, investigative emphasis shifted toward utilizing these new tools to explore the mechanism of agonist activation of adenylate cyclase. Agonist agents interact with P-adrenergic receptors in a fundamentally different way than do antagonists in order to initiate the chain of events leading to characteristic biological responses. Radioligand binding assays provided an opportunity to examine these differences at the receptor level. One of the first unique properties of agonist binding to (3-adrenergic receptors which was documented was that guanine nucleotides modulate receptor affinity for agonists but not antagonists (Maguire et al., 1976; Lefkowitz et al., 1976). The rationale for exploring the effects of guanine nucleotides on adrenergic receptor binding properties was based on the studies of Rodbell er al. (1971), who demonstrated that these nucleotides were essential regulators of the glucagon receptor-adenylate cyclase complex in rat liver membranes. For (3-adrenergic receptors, the initial demonstrations of guanine nucleotide regulation of agonist binding were accomplished using partially purified plasma membranes prepared from C6 glioma cells (Maguire et al., 1976) and frog erythrocytes (Lefkowitz et al., 1976). It was found that agonist competition binding curves vs radiolabeled antagonists were “shallow” in the absence of guanine nucleotides but became steeper and shifted toward the right, indicating a lower apparent affinity of the receptor for agonist, in the presence of exogenously added guanine nucleotides such as GTP or Gpp(NH)p (Fig. 1). The addition of guanine nucleotides to the binding assay did not affect the binding of the radiolabeled antagonist nor did it affect the shape or position of the competition binding curves generated by unlabeled antagonists (Fig. 2). These curves are “steep” and uniphasic under all experimental conditions. The guanine nucleotide-dependent ‘‘shift to the right” of the agonist binding curves, i.e., to lower receptor affinity, was observed for a series of p-

51

THE (3-ADRENERGICRECEPTOR: LIGAND BINDING STUDIES

9

8

7

-loglo

6

5

I-)Isoproterenol]

4

3

2

(M)

FIG. I . Computer modeling of competition binding data of isoproterenol for r3H]DHA in frog erythrocyte membranes. Competition of the agonist isoproterenol for ['HIDHA in the absence (0) and presence (0) of CTP. The curve in the absence of nucleotide was significantly ( p < 0.00 I ) better fit by a model for two binding states of the receptor. See text for details. (From Kent et a / ., 1979.)

25 5

5 100

-

-

F

0

u

0

s

80-

(-1 Alprenolot

0

UL= KH * 12nM

m

-

5

60-

C

ea

0

40-

-s5

20-

e

D %

I

.-.. I

0

1

I

1

1

I

,. r

FIG. 2. Computer modeling of competition binding data of alprenolol for ['HIDHA in frog erythrocyte membranes. The competition curve of the antagonist alprenolol for ["IDHA is adequately modeled to a homogeneous class of binding sites. (From Kent er a / . . 1979.)

52


adrenergic agonists (Lefiowitz et a / ., 1976). Interestingly the magnitude of this “nucleotide-dependent shift” correlates with the intrinsic activity of the agonist for activation of the adenylate cyclase. Intrinsic activity is a term used to quantitate the maximum ability of an agonist to stimulate a biological response, such as adenylate cyclase activity. The correlation of intrinsic activity and nucleotide effects on agonist binding suggests a relationship between guanine nucleotide regulation of receptor affinity for an agonist and the drug’s ability to stimulate the cyclase enzyme. Thus agonists, but not antagonists, have the ability to form a high-affinity complex with the P-adrenergic receptor, and this agonist-receptor complex is modulated by guanine nucleotides. Computer-aided analysis of radioligand binding data has added a new quantitative dimension to our understanding of the differences between agonist and antagonist binding to P-adrenergic receptors. As described above, the competition binding curves of adrenergic agonists for a radiolabeled antagonist such as [3H]DHA are shallow (slope factors < 1) indicating complex binding interactions between agonists and the receptors. In contrast, the competition binding curve of unlabeled antagonists for [3H]DHA is always steep (slope factor = I ) indicating a uniform affinity of the receptors for antagonists. Computer modeling of the shallow agonist competition binding curves indicated a statistically significant improvement in the fit of the binding data by a model based on two binding states of the receptor (Fig. I ) (Kent er al., 1979). This two-state model was found to be appropriate for all agonists tested. The two affinity states of the receptor were characterized by specific dissociation constants (K,, KL) and the proportion of the total receptor population in each state was determined (RH, RL). Using a series of full and partial agonists, a significant correlation was shown to exist between the ability of an agonist to stimulate the adenylate cyclase (intrinsic activity) and the ratio of the dissociation constants of the agonist for the high- and low-affinity states of the receptor (KLIKH).A significant correlation was also found to exist between agonist intrinsic activity and the proportion of the receptors binding the agonist with high affinity (% R H ) (Kent et ai., 1979). Quantitative analysis of radioligand binding data thus provides additional evidence for the important role of agonist high-affinity binding in the process of transmembrane signaling by receptor-cyclase complexes. The ability of an agonist to form a high-affinity, nucleotide-sensitive complex with the P-adrenergic receptor is dependent upon the ionic environment of the membranes during the binding assay. Although neither receptor binding properties nor adenylate cyclase activity shows a significant dependence on ionic strength, divalent cations can be shown to be required for both of these activities. Concentrations of Mg2+ or Mn2+ in the millimolar range are necessary for adenylate cyclase activity and these same cations are also required for agonist high-affinity binding to receptors (Williams et al., 1978). Monovalent cations cannot substitute for these divalent metal ions. The regulation of receptor affinity

'THE 5-ADRENERGIC RECEPTOR: LIGAND BINDING STUDIES

53

by divalent cations is observed only with agonist binding; antagonist interactions with the P-adrenergic receptor are unperturbed by the presence or absence of metal ions in the binding medium. Subsequent radioligand binding studies have demonstrated several additional unique binding properties of agonists to P-adrenergic receptors. Pike and Lefkowitz (1978) reported that decreasing the temperature of the binding assay incubation increased the apparent affinity of turkey erythrocyte P-adrenergic receptors for agonists without a significant effect on receptor affinity for antagonists. This observation has recently been extended to show a similar specific temperature effect on agonist binding to (3-adrenergic receptors in a variety of mammalian tissues containing both p,- and @,-receptor subtypes (Weiland e l a / . , 1980). Briggs and Lefkowitz (1980) were able to show that when assayed below physiological temperatures, agonists are not able to induce the high-affinity, nucleotide-sensitive state of the P-adrenergic receptor in turkey erythrocyte membranes and that this observation correlates with an inhibition of the ability of agonists to stimulate the adenylate cyclase in these membranes. Treatment of the turkey erythrocyte membrane with the unsaturated fatty acid cis-vacennic acid, which increases the flujdity of these membranes (Rimon et al., 1978), led to reappearance of the ability of agonists to form a high-affinity complex with the receptor and concomitantly facilitated agonist activation of the adenylate cyclase at low temperature. These effects of temperature on agonist binding characteristics and activation of adenylate cyclase may reflect a specific agonist-induced conformational change in the receptor, specific receptor-lipid interactions, and/ or necessary lateral mobility of the receptor-cyclase components within the lipid matrix. Additional evidence supporting an agonist-induced conformational change in the p-adrenergic receptor comes from studies using N-ethylmaleimide (NEM) in turkey erythrocyte membranes. N-Ethylmaleimide reacts specifically and irreversibly with free sulfhydryl groups of proteins (Means and Feeney, 1971). Pretreatment of turkey erythrocyte membranes with NEM alone does not affect the ability of (3-adrenergic receptors to bind the radiolabeled antagonist [3H]DHA, but pretreatment of these membranes with NEM in the presence of a p-adrenergic agonist resulted in a loss of up to 50% of the (3-receptors as determined by a decrease in binding capacity for ["IDHA (Bottari et al., 1979). Simultaneous treatment of membranes with NEM and antagonist did not result in receptor loss, indicating that agonist binding induces a specific conformational change in the receptor which exposes a cysteine residue to NEM. This agonist-specific effect appears to be related to the mechanism by which agonist binding to receptor results in activation of adenylate cyclase. The rate at which NEM inactivates turkey erythrocyte (3-receptors correlates with the intrinsic activity of the agonist which occupies the receptor (Vauquelin et al., 1979). Guanine nucleotides which

54


regulate receptor affinity for agonists and are required for adenylate cyclase activation also inhibit the inactivation of receptors by NEM in the presence of agonists (Vauquelin et al., 1980). The effect of NEM in the presence of agonist is not restricted to the P ,-adrenergic receptor in turkey erythrocyte membranes. Similar results have been obtained by treating membranes from S49 mouse lymphoma cells (P,-receptor subtype) with NEM and agonist (Vauquelin and Maguire, 1980). It is noteworthy that membranes from S49 mutant cell clones that are functionally devoid of components necessary for guanine nucleotide regulation of receptor affinity failed to show an effect with NEM in the presence of agonists, suggesting that productive coupling between the P-adrenergic receptor and the guanine nucleotide regulatory protein of the adenylate cyclase complex may be necessary for the exposure of the critical sulfhydryl group (Vauquelin and Maguire, 1980). A perplexing question raised by these studies concerns the observation that the maximal effect of NEM and agonist reduces the receptor population only by 50%. Complete inactivation of the receptors was never achieved by this approach. “Heterogeneous” populations of receptors have been offered as a possible explanation, but further investigation will be required to clarify this observation. An important advance in the characterization of agonist binding to P-adrenergic receptors was achieved through the development of an effective radiolabeled agonist ligand. Lefiowitz and Williams (1977) reported the properties of the binding of (+)[ 3H]hydroxybenzylisoproterenolto the P-adrenergic receptor of frog erythrocyte membranes. It became evident that this radioligand would be useful in the direct investigation of P,-adrenergic receptors in purified plasma membrane preparations. Availability of this ligand permitted, for the first time, a direct examination of agonist binding to the P-adrenergic receptor. Previously such studies had been performed by examining agonist competition with radioligand antagonist binding. High-affinity [3H]HBI binding to the P-adrenergic receptor is characterized by a very slow rate of dissociation that is not affected by the addition of competing adrenergic ligands (Williams and Lefiowitz, 1977). In contrast, guanine nucleotides promote the rapid and complete dissociation of [3H]HBI from the membrane receptors. Thus [3H]HBI labels exclusively the high-affinity state of the P-adrenergic receptor in frog erythrocyte membranes. The effect of guanine nucleotides to lower receptor affinity for agonist appears to relate to their ability to destabilize the agonist high-affinity binding state resulting in rapid release of the agonist from the receptor. Chemical treatments of frog erythrocyte membranes which interfere with the ability of agonists to stimulate the adenylate cyclase generally prevent the ability of agonists to form a highaffinity, slowly dissociable complex with the receptor (Williams and Lefkowitz, 1977). Stimulation of adenylate cyclase by guanine nucleotides is therefore associated with a decrease in affinity of the receptor for agonists and rapid dissociation of the high-affinity agonist-receptor complex to free agonist and receptor. Formation of the tight complex between agonist and receptor thus

THE P-ADRENERGIC RECEPTOR:LIGAND BINDING STUDIES

55

appears in some way to facilitate activation of the cyclase enzyme by regulatory nucleotides. It is possible to demonstrate experimentally that the high-affinity, slowly dissociable agonist-receptor complex which is sensitive to guanine nucieotides is an intermediate state on the pathway of coupling between the receptor and adenylate cyclase activation (Stadel et al., 1980). The high-affinity state of the receptor can be isolated by preincubation of purified frog erythrocyte membranes with agonist. These membranes were then washed extensively to remove the free agonist and then assayed for adenylate cyclase activity in the presence of the antagonist propranolol. Basal and NaF-stimulated cyclase activity were unaffected by the preincubation procedures, but the ability of the nonhydrolyzable guanine nucleotide analog Gpp(NH)p to stimulate the enzyme directly was significantly enhanced in membranes preexposed to the agonist compared to membranes preincubated in buffer alone. Experiments of this type suggest that the increased stimulation of the adenylate cyclase by Gpp(NH)p in membranes pretreated with agonist is the result of high-affinity agonist binding to the receptor that persists through the washing procedures. Moreover, they demonstrate that this agonist-receptor complex is an intermediate for agonist activation of cyclase activity. These procedures may be repeated using turkey erythrocyte membranes with qualitatively similar results. Thus the mechanism by which agonists activate adenylate cyclase is similar for both a P I - and a P,-adrenergic receptor (Stadel er af., 1980). Additional insights into the molecular mechanisms of receptor-cyclase coupling can also be gained through the application of computer modeling techniques to radioligand binding data. Quantitative analysis of agonist competition binding curves for [”IDHA in the presence and absence of guanine nucleotides is compatible with the notion that nucleotides mediate a transition between highand low-affinity states of the receptors (Kent er al., 1979). Thus, the agonist competition binding curve is steep in the presence of guanine nucleotides (slope factor = 1) and the uniform dissociation constant for agonist binding to the receptor is identical to the low-affinity dissociation constant ( K J determined for the same agonist in the absence of the nucleotide (Fig. 1). The extent of the transition from high- to low-affinity state is dependent on the concentration of guanine nucleotide in the binding assay. The observations that guanine nucleotide mediates a transition of the agonist high-affinity state of the receptor to the low-affinity state without a similar effect in antagonist binding, and that partial agonists induce differing proportions of the receptor into the high- and low-affinity state at equilibrium (% R H ) ,represent strong evidence that the highand low-affinity states of the agonist-occupied receptors are interconvertible (Kent et al., 1979). A systematic comparison of the ability of several mechanistic models to fit and reproduce agonist competition binding data in the presence and absence of guanine nucleotides led De Lean er af. (1980) to propose a “ternary complex”

56


(precoupled)

FIG.3. Schematic diagram of the ternary complex model. The model involves the interaction of the hormone (H), the receptor (R), and an additional membrane component (X). Nucleotide-dependent coupling between the ternary complex (HRX) and activation of adenylate cyclase (E) is also shown. (From De Lean er al., 1980.)

model as an explanation for the agonist-specific binding properties of P-adrenergic receptors. This model involves the interaction of the receptor (R) with an additional membrane component (X) in the presence of agonist (H) to form a high-affinity ternary complex HRX (Fig. 3). Agonist initially binds to the receptor to form a low-affinity binary complex HR which precedes the ternary complex formation. The modeling indicates that the stoichiometry between the receptor and the component X is close to I:]. The intrinsic activity of an agonist correlates with the affinity constant (L) for the combinations of the agonistreceptor complex (HR) with the additional membrane component X: HR

+X

L

HRX

(Fig. 3). This correlation is entirely consistent with the relationships alluded to above between agonist intrinsic activity and other quantitative parameters of the agonist-promoted high-affinity state (KL/K,l, % RH). The computer-aided modeling of the binding data is independent of the nature of the additional membrane component X, but several lines of evidence suggest that X is a guanine nucleotide regulatory protein (N). The computer analyses of binding data indicate that the presence of guanine nucleotides in the binding assay specifically decreases the ability of an agonist to stabilize the ternary complex between HR and X (De Lean ef al., 1980). Biochemical experiments using high concentrations of Mn2+ ( > l o mM) to uncouple agonist binding to receptors from activation of adenylate cyclase (Limbird et al., 1979) or using the specific sulfhydryl reagent N-ethylmaleimide to inactivate adenylate cyclase catalytic activity (Howlett et al., 1978; Stadel and Lefkowitz, 1979) have demonstrated that a functional cyclase enzyme is not

THE B-ADRENERGIC RECEPTOR: LIGAND BINDING STUDIES

57

required for agonists to promote formation of a high-affinity, nucleotide-sensitive complex with the receptor. However, Stadel and Lefkowitz (1979) have shown that an additional membrane component necessary for agonist high-affinity binding was also sensitive to NEM, but at concentrations 100-fold greater than that necessary to inactivate cyclase catalytic activity. The effect of NEM on agonist high-affinity binding is distal to the ligand binding site of the receptor, since agonist low-affinity binding (in the presence of guanine nucleotide) and the binding of the competitive antagonist [ 3H]DHA were unaffected. Preformation of the agonist high-affinity state of the receptor protected the complex against the effects of NEM, and the complex was still fully sensitive to modulation by guanine nucleotides (Stadel and Lefkowitz, 1979). It is therefore unlikely that the ternary complex (HRX) contains the cyclase enzyme. However, the observation that agonist binding to the P-adrenergic receptor is uniquely modulated by guanine nucleotides is consistent with the notion that component X is a guanine nucleotide regulatory protein. It is of interest to note that a-adrenergic agonist binding to a,-receptors in platelets (Tsai and Lefkowitz, 1979) or neural cell lines (Haga and Haga, 1981) is also characterized by shallow competition binding curves with radiolabeled antagonists that steepen and shift to the right in the presence of guanine nucleotides (Hoffman et al., 1980). The guanine nucleotide sensitivity of the a,receptor is also an agonist-specific property since the affinity of antagonists appears to be unperturbed by the addition of nucleotides. The nucleotide sensitivity of agonist binding to a,-adrenergic receptors again provides clues to understanding the mode of action of a-adrenergic agonists in inhibiting adenylate cyclase, since guanine nucleotides are also stringently required for coupling of these receptors to the catalytic moiety of the enzyme (Jakobs ef al., 1978). Although computer modeling using the ternary complex model (De Lean el al., 1980) has not been applied to binding data for a-adrenergic agonists, it appears likely that such a complex is in fact an intermediate for inhibition as well as stimulation of adenylate cyclase.

IV. CHARACTERIZATION OF DETERGENT-SOLUBILIZED ADRENERGIC RECEPTORS A first step toward the biochemical characterization of membrane-bound hormone receptors is the solubilization of the binding activity from the lipid bilayer through the use of detergents. The P-adrenergic receptor was first solubilized from frog erythrocyte membranes using the plant glycoside digitonin (Caron and Lefkowitz, 1976). Although many different detergents were tested, digitonin was found to be uniquely capable of extracting the receptor in an active form. The binding properties of the soluble receptor sites were in most respects essen-

58


tially the same as those of the membrane-bound receptors (Caron and Lefkowitz, 1976). The soluble receptors demonstrate the appropriate potency order for pzadrenergic receptors (isoproterenol > epinephrine > norepinephrine) and strict stereoselectivity for binding both agonist and antagonist ligands. The single difference in the properties of the soluble receptors is that agonist binding to soluble receptors is of uniformly low affinity. Thus agonists do not promote formation of the high-affinity state of the receptor in soluble preparations. This observation correlates with the inability of agonists to stimulate adenylate cyclase activity in these soluble preparations (Caron and Lefkowitz, 1976). Fractionation of detergent extracts by gel filtration (Limbird and Lefkowitz, 1977) or on sucrose gradients (Haga et al., 1977) has led to a clear resolution of receptor binding activity from adenylate cyclase activity, thus demonstrating that these two activities reside on different polypeptide chains. Studies of soluble receptor preparations have shed light on the unique interactions of P-adrenergic receptors with agonist agents. Soluble extracts from frog erythrocyte or rat reticulocyte membranes which were prelabeled with either the radiolabeled agonist [3H]HBI or the radiolabeled antagonist [3H]DHAwere fractionated by gel filtration over AcA34 resin (Limbird and Lefkowitz, 1978; Limbird et al., 1980b). The ['HIHBI prelabeled receptor was resolved from the antagonist-occupied receptor and appeared to elute with an apparent larger molecular size (Fig. 4A). Several explanations are consistent with this observation including an agonist-promoted asymmetric conformational change in the receptor, agonist-induced receptor aggregation, or the stable association of the agonist-occupied receptor with an additional membrane component. This latter explanation is, of course, consistent with the computer modeling described above. The proposed additional component of the agonist-receptor high-affinity state did not appear to be the cyclase enzyme itself, since the enzyme activity eluted several fractions removed from the receptor binding activity. Additional experiments using rat reticulocyte membranes implicated the nucleotide regulatory protein as a constituent of the agonist-receptor high-affinity complex (Fig. 4B). Prelabeling of rat reticulocyte membranes with [3H]HBI in the presence of guanine nucleotide allows agonist binding to the low-affinity form of the receptor. This low-affinity agonist-receptor complex survives the gel filtration procedures and now coelutes with the smaller antagonist prelabeled receptor. Thus guanine nucleotides which destabilize the high-affinity state of the receptor in the membrane also convert the larger molecular form of the agonist-receptor complex to a smaller species that coelutes with antagonistoccupied receptor. More direct evidence as to the molecular compositions of the agonist-promoted ternary complex was obtained by radioactively labeling of the nucleotide regulatory protein of the adenylate cyclase complex (Limbird et al., 1980a). Cholera toxin catalyzes the covalent transfer of ADP-ribose from NAD to the +

THE p-ADRENERGIC RECEPTOR: LIGAND BINDING STUDIES

59

FIG.4. Gel exclusion chromatography of P-adrenergic receptors solubilized from rat reticulocyte membranes with digitonin after prelabeling with agonist ['HIHBI or antagonist [ 3H]DHA. The column material was AcA34. (A) Prelabeling in the absence of nucleotides. (B) Prelabeling conducted in the presence of 0. I mM Gpp(NH)p. (From Limbird et a / . , 1980b.)

42,000 M , subunit of the guanine nucleotide regulatory protein (for reviews see Ross and Gilman, 1980; Stadel et al., 1982). If 32P-labeledNAD+ is used as the cofactor for the toxin, a radioactive tag is covalently incorporated into the nucleotide regulatory protein. Limbird et al. (1980a) were able to show that agonist pretreatment of rat reticulocyte membranes prior to solubilization resulted in the coelution of the 32P-labeled 42,000 M , protein in the [3H]HBI-receptor region from the gel filtration column. Similar pretreatment of these membranes with antagonist did not cause the labeled subunit of the nucleotide regulatory protein to associate with the receptor. These experiments provide biochemical evidence that agonist occupancy of the P-adrenergic receptor promotes the association of the receptor with the guanine nucleotide regulatory protein of the adenylate cyclase complex.

60


Several recent reports provide additional evidence that the high-affinity, nucleotide-sensitive agonist-receptor complex represents a ternary complex of HRN. Cholate-solubilized extracts of membranes prepared from wild-type S49 lymphoma cells successfully reconstituted hormonally sensitive adenylate cyclase activity in membranes from S49 mutant cell clones that are functionally uncoupled (Sternweis and Gilman, 1979). The critical factor in the detergent extract that allows recoupling of the receptors to the cyclase is the guanine nucleotide regulatory protein (Ross et al., 1978). The P-adrenergic receptors of the mutant cell membranes reconstituted with the cholate extracts of S49 wildtype membranes also demonstrate nucleotide sensitivity of receptor affinity for agonists. It was not possible to separate the component necessary for recoupling of hormonal activation of the cyclase from the component required for restoration of nucleotide-sensitive agonist high-affinity binding to the P-adrenergic receptors. These experiments are consistent with the notion of a single membrane component regulating receptor affinity for agonists and for coupling agonist occupancy of the receptor to activation of the cyclase enzyme. Using reconstitution of lubrol-solubilized components, Stadel et al. (198 1) were able to isolate the nucleotide regulatory protein associated with the 6receptor as a result of agonist binding and subsequently show that this N protein conveyed nucleotide-dependent adenylate cyclase activity to a suitable catalytic unit acceptor. The high-affinity ternary complex HRN was solubilized from frog erythrocyte membranes in the nonionic detergent lubrol and then bound to wheat germ agglutinin immobilized on Sepharose. The ternary complex is bound to the lectin through the carbohydrate moieties of the receptor, which is a glycoprotein (Shorr et al., 1980). After extensive washing of the lectin gel the resin was eluted in the presence of GTPyS. The guanine nucleotide destabilizes the ternary complex HRN resulting in the release of an N-GTPyS complex. The GTPyS eluate from the lectin-Sepharose conveyed nucleotide-sensitive adenylate cyclase to a soluble catalytic unit acceptor. The ability of the GTPyS eluate of the lectin-resin to reconstitute adenylate cyclase activity was strictly dependent on the preformation of the agonist high-affinity state in frog erythrocyte membranes prior to solubilization. The notion that the nucleotide regulatory protein associated with the P-adrenergic receptor by the binding of agonist is the same N that modulates adenylate cyclase activity was supported by additional experimentation (Stadel et at., 1981). Radioactive labeling of the N protein by 32P-labeledNAD+ in the presence of cholera toxin allows the observation of the N protein throughout the solubilization, lectin chromatography, and elution procedures. The amount of 32P-labeled42,000 M, subunit associated with the lectin-resin in soluble extracts from membranes pretreated with agonist or antagonist correlated with the ability of these extracts to stimulate adenylate cyclase activity in the soluble reconstitution assay. These experiments bring together both structural and functional evi-

61

THE B-ADRENERGIC RECEPTOR: LIGAND BINDING STUDIES

p- Adrenergic Receptor Cycle

Nucleotide Regulatory

Adenylate Cyclase

Protein Cycle

Catalytic Moiety Cycle

FIG. 5 . Schematic model of hormonal activation of adenylate cyclase involving agonist (H), receptor (R), nucleotide regulatory protein (N), and enzyme catalytic unit ( C ) . See text for details.

dence that a single nucleotide regulatory protein acts as a “coupler” conveying information from the agonist-occupied receptors to the adenylate cyclase. The observations characterizing the unique binding properties of agonists to the P-adrenergic receptor in both membrane and soluble studies are consistent with a model for receptor-cyclase coupling shown in Fig. 5 . This model is based on the information contained in the experiments reviewed above as well as additional investigations of the properties of the enzyme adenylate cyclase reported by other investigators. In this model agonist occupancy of the receptor (Step 1) promotes or stabilizes the formation of the high-affinity HRN complex (Step 2). As a consequence, GDP is released from N, creating a vacant guanine nucleotide binding site (Step 3). The binding of a guanine nucleotide triphosphate to N (Step 4)results in dissociation of the HRN complex, and N-GTP now associates with C (Step 5 ) to stimulate catalytic activity (Pfeuffer, 1977, 1979). As shown by Cassel and Selinger (1976), hydrolysis of GTP by a GTPase associated with the NC complex (Step 6) is the “turn off” mechanism for adenylate cyclase activity and returns the system to the basal state (Step 7). Binding of agonist to the receptor reinitiates the cycle. The major features of the model are ( 1 ) agonist binding results in the stabilization of the high-affinity ternary complex HRN which facilitates the exchange of nucleotides bound to N; ( 2 ) the guanine nucleotide regulatory protein acts as a coupler between the receptor and the enzyme catalytic unit; (3) the GTPase activity associated with the NC complex deactivates enzyme catalytic activity and dissociates this complex. This model may also provide a starting point for investigating the mechanism of inhibition of adenylate cyclase mediated by qadrenergic receptors. As described above from radioligand binding studies it is apparent that a,-adrenergic receptors are capable of forming a high-affinity, nucleotide-sensitive complex with agonists but not antagonists. Recent studies (Michel er al., 1981; Smith and

62


Limbird, 1981) also show that agonist occupancy of platelet a,-receptors similarly induces an increase in the apparent molecular size of the receptor compared to antagonist-occupied receptors as assessed by centrifugation through sucrose gradients. The soluble agonist-receptor complex is also sensitive to guanine nucleotides, consistent with the notion that the increased molecular size of the agonist-receptor complex is due to the agonist-promoted association of the receptor with a guanine nucleotide regulatory protein (Smith and Limbird, 1981). These data suggest that a-adrenergic inhibition of adenylate cyclase shows many features in common with, and may be analogous to, the mechanism of P-adrenergic stimulation of the cyclase. Further investigation will be necessary to determine how the formulations in the model shown in Fig. 5 for the stimulation of adenylate cyclase might apply to the mechanism of inhibition of the enzyme. A long-range goal of studies of the mechanism of receptor-cyclase coupling is the purification and reconstitution of the individual components of the systems in a functional way. In the past 2 years considerable progress has been made in this regard. The development of sensitive assays for detergent-solubilized components of the complex has allowed the application of biochemical techniques for purification. Purification of these components is a major undertaking since the constituents of the receptor-adenylate cyclase complex exist in very small quantities in the plasma membranes of target cells. Recently, the guanine nucleotide regulatory protein has been purified to apparent homogeneity by classic biochemical techniques (Northup et al., 1980). The purified protein is composed of three heterologous subunits with approximate molecular weights of 52,000, 45,000, and 35,000. The purified guanine nucleotide regulatory protein reconstitutes guanine nucleotide-, hormonal-, and sodium fluoride-dependent stimulation of adenylate cyclase activity in membranes prepared from mutant S49 lymphoma cells that lack a functional regulatory unit. All three subunits appear to be required for successful reconstitution. A key step in the purification of the P-adrenergic receptor was the development of an efficient affinity chromatography gel (Caron et a!., 1979). Alprenolol was immobilized on Sepharose 4B through a hydrophilic spacer arm. The biospecific nature of the interaction of the digitonin-solubilized P-adrenergic receptor with the affinity gel could be demonstrated. Both the adsorption of the soluble frog erythrocyte P-receptor to the resin and its subsequent elution demonstrated typical P-adrenergic specificity. For both processes (blocking adsorption and promoting elution), the agonist potency order was isoproterenol > epinephrine > norepinephrine and stereoselectivity was preserved for both agonist and antagonist agents. The resin adsorbed up to 95% of the receptor in the soluble preparations, and 60% was ultimately specifically eluted. By recycling the soluble receptor preparation through the affinity resin the purification was over 15,000fold from the original membranes (Caron et al., 1979). By coupling the affinity chromatography procedures to ion exchange chro-

THE P-ADRENERGIC RECEPTOR: LIGAND BINDING STUDIES

63

matography the ligand binding site of the P-adrenergic receptor has been purified 55,000-fold (Shorr et al., 1981). The purified receptor demonstrates a 58,000 M, band on SDS-polyacrylamide gel electrophoresis. The purified 58,000Mr protein comigrates with soluble radioligand antagonist prelabeled receptor on sucrose gradients and in isoelectric focusing procedures. Ligand binding experiments using I3H]DHA and the purified protein demonstrate the affinity, specificity, and stereoselectivity expected for the P-adrenergic receptor (Shorr et al., 1981). The purification of the P-adrenergic receptor and the nucleotide regulatory protein leaves the catalytic unit of adenylate cyclase as the only known component remaining to be purified. The rapid progress that has been made in the isolation and characterization of the molecular components of the P-adrenergic receptor-adenylate cyclase complex raises expectations that functional reconstitution of this system will be achieved in the not too distant future. REFERENCES Ahlquist, R. P. (1948). A study of the adrenotropic receptors. Am. J . Phvsiol. 153, 586-600. Atlas, D.. Steer, M. L., and Levitzki, A. (1974). Stereospecific binding of propranolol and catecholamines to the beta-adrenergic receptor. Proc. Narl. Acad. Sci. U.S.A. 71, 4246-4248. Aurbach, G. D., Fedak, S. A , , Woodard, C. J., Palmer, J . S.. Hauser, D., and Troxler, F. (1974). The beta-adrenergic receptor: Stereospecific interaction of an iodinated beta-blocking agent with a high affinity site. Science 186, 1223-1224. Barovsky, K..and Brooker, G . (1980). (-)('2sI]IodopindoloI, a new highly selective radioiodinated P-adrenergic receptor antagonist: Measurement of (3-receptors on intact rat astrocytoma cells. J. Cyclic Nucleoride Res. 6, 297-307. Berthelson, S., and Pettinger, W. A. (1977). A functional basis for the classification of alphaadrenergic receptors. Ljfe Sci. 21, 595-606. Bottari. S . , Vauquelin, 0.. Durien, O., Klutchko, C., and Strosberg, A. D. (1979). The P-adrenergic receptor of turkey erythrocyte membranes: Conformation modification by P-adrenergic agonists. Biochem. Biophys. Res. Commun. 86, 131 1-1318. Briggs, M. M., and Lefkowitz, R. J. (1980). Parallel modulation of catecholamine activation of adenylate cyclase and formation of the high-affinity agonist-receptor complex in turkey erythrocyte membranes by temperature and cis-vaccenic acid. Biochemisrv 19, 4461-4466. Caron, M. G . , and Lefkowitz, R . J. (1976). Solubilization and characterization of the P-adrenergic receptor binding sites of frog erythrocytes. J. B i d . Chem. 251, 2374-2384. Caron, M. G., Srinivasan, Y.,Pitha, J., Kociolek, K., and Letkowitz, R. J. (1979). Affinity chromatography of the P-adrenergic receptor. J. B i d . Chem. 254, 2923-2927. Cassel, D.,and Selinger, 2. (1976). Catecholamine-stimulated GTPase activity in turkey erythrocyte membrane. Biochim. Biophys. Acta 452, 538-55 I . Dale, H. H. (1906). On some physiological actions of ergot. J. Physiol. (London) 34, 165-206. De Lean. A . , Stadel, J. M., and Letkowitz, R. J . (1980). A ternary complex model explains the agonist-specific binding properties of the adenylate cyclase-coupled P-adrenergic receptor. J. B i d . Chem. 255, 7108-71 17. Engle, G . ( 1 980). Identification of different subgroups of beta-receptors by means of binding studies in guinea-pig and human lung. Triangle 19, 69-76. Exton, J. H. (1979). Mechanisms involved in effects of catecholamines on liver carbohydrate metabolism. Biochem. Pharmacol. 28, 2237-2246.

64


Greenberg, D. A , , U'Prichard, D. C., and Snyder, S. H. (1976). Alpha-noradrenergic receptor binding in mammalian brain: Differential labelling of agonist and antagonist states. Life Sci. 191, 69-76. Greengrass, P.. and Brenner, R. (1979). Binding characteristics of [3H]prazosin to rat brain alphaadrenergic receptors. Eur. J. Pharmacol. 55, 323-325. Haga, T., and Haga, K . (1981 ). Characterization by [3H]dihydroergocryptine binding of alphaadrenergic receptors in neuroblastomd X glioma hybrid cells. J. Neurochem. 36, 1152- 1159. Haga. T., Haga, K., and Gilman, A. G. (1977). Hydrodynamic properties of the f3-adrenergic receptor and adenylate cyclase from wild-type and variant S49 lymphoma cells. J . Biol.Chem. 252, 5776-5782. Hoffman, B. B., and Lefkowitz, R. J. (1980). Radioligand binding studies of adrenergic receptors. Annu. Rev. Pharmaml. Toxicol. 26, 581 -608. Hoffmann, B . B . , Mullikin-Kilpatrick, D., and Lefkowitz, R. J. (1980). Heterogeneity of radioligand binding to a-adrenergic receptors. J. B i d . Chem. 255, 4645-4652. Howlett, A. C., Van Arsdale, P. M., and Gilman, A. G . (1978). Efficiency of coupling between the beta-adrenergic receptor and adenylate cyclase. Mol. Pharmarol. 14, 53 1-539. Jakobs, K. H., Saur, W . , and Schultz, G . (1976). Reduction of adenylate cyclase activity in lysates of human platelets by alpha-adrenergic component of epinephrine. J. Cyclic. Nucleotide Res. 2, 381-392. Jakobs, K. H., Saur, W . , and Schultz, G. (1978). Inhibition of platelet adenylate cyclase by epinephrine requires GTP. FEBS Lett. 85, 167-170. Kent, R. S . , De Lean, A., and Lefkowitz, R. J. (1979). A quantitative analysis of beta-adrenergic receptor interactions: Resolution of high and low affinity states of the receptor by computer modeling of ligand binding data. Mol. Pharmacol. 17, 14-23. Lands, A. M., Arnold. A.. McAuliff, J. P., Luduena, F. P., and Braun, T. G . (1964). Differentiation of receptor systems activated by sympathomimetic amines. Nature (London) 214, 597-598. Lefkowitz, R. J., and Williams, L. T. (1977). Catecholamine binding to the beta-adrenergic receptor. Proc. Natl. Acad. Sci. U.S.A. 74, 515-519. Lefkowitz, R. J., Roth, I . , Pricer, W., and Pastan. I. (1970). ACTH receptors: Specific binding of ACTH-['2sI] and its relation to adenyl cyclase. Proc. Natl. Acad. Sci. U.S.A. 65, 745-752. Letkowitz. R . J . , Mukherjee, C., Coverstone, M . , and Caron, M. G. (1974). Stereospecific [3Hl(-)alprenolol binding sites, beta-adrenergic receptors and adenyl cyclase. Biochem. Biophys. Res. Commun. 60, 703-709. Lefkowitz, R. J., Mullikin, D.. and Caron, M. G . (1976). Regulation of beta-adrenergic receptors iw guanyl-5'-yl imidophosphate and other purine nucleotides. J. B i d . Chem. 251, 4680 4692. Limbird, L. E . , and Lefkowitz, R. J. (1977). Resolution of f3-adrenergic receptor lmiding and adenylate cyclase activity by gel exclusion chromatography. J. Biol. Chem. 252, 799-802. Limbird, L. E., and Lefkowitz, R. J. (1978). Agonist-induced increase in apparent P-adrenergic receptor size. Proc. Natl. Acad. Sci. U.S.A. 75, 228-232. Limbird, L. E., Hickey, A. R., and Lefkowitz, R. J. (1979). Unique uncoupling of the frog erythrocyte adenylate cyclase system by manganese. J. Biol. Chem. 254, 2677-2683. Limbird. L. E., Gill. D. M., and Lefkowitz, R. J . (1980a). Agonist-promoted coupling of the padrenergic receptor with the guanine nucleotide regulatory protein of the adenylate cyclase system. Proc. Nail. Acad. Sri. U.S.A. 77, 775-779. Limbird, L. E., Gill, D. M., Stadel, J. M., Hickey, A. R., and Lefkowitz, R. J . (1980b). Loss of padrenergic receptor-guanine nucleotide regulatory protein interactions accompanies decline in catecholamine responsiveness of adenylate cyclase in maturing rat erythrocytes. J . Biol. Chem. 255, 1854-1861. Lyn. S. Y., and Goodfriend, T. L. (1970). Angiotensin receptors. Am. J. Phvsiol. 218, 1319-1328.

THE p-ADRENERGIC RECEPTOR: LIGAND BINDING STUDIES

65

Maguire. M. E., Van Arsdale, P. M., and Gilman, A. G . (1976). An agonist-specific effect of guanine nucleotides on binding to the beta-adrenergic receptor. Mol. Pharmacol. 12, 335-339. Means, ti. E..and Feeney, R. E. (1971). “Chemical Modification of Proteins.” Holden-Day, San Francisco, California. Michel,T., Hoffman. B. B., Lefkowitz, R. J., andcaron, M. G . (1981). Differential sedimentation properties of agonist- and antagonist-labelled platelet alphaz-adrenergic receptors. Biochem. Biophys. Res. Cornmun. 100, 1 I3 I - 1 135. Northup, J. K.,Sternweis, P. C.. Smigel, M. D . , Schleifer, L. S., Ross, E. M., andGilman, A. G. (1980). Purification of the regulatory component of adenylate cyclase. Proc. Natl. Acad. Sci. U.S.A. 77, 6516-6520. Pfeuffer, T. (1977). GTP-binding proteins in membranes and the control of adenylate cyclase activity. J . Biol. Chem. 252, 7224-7234. Pfeuffer, T. ( 1979). Guanine nucleotide-controlled interactions between components of adenylate cyclase. FEBS Lett. 101, 85-89. Pike, L. J . , and Lefkowitz, R. I. (1978). Agonist specific alterations in receptor binding affinity associated with solubilization of turkey erythrocyte membrane beta-adrenergic receptors. Mol. Pharmacol. 14, 370-375. Rimon. G.. Hanski, E., Braun. S., and Levitzki, A. (1978). Mode of coupling between hormone receptors and adenylate cyclase elucidated by modulation of membrane fluidity. Nature (London) 276, 394-396. Robison, G. A,, Butcher, R . W . , and Sutherland, E. W . (1971). “Cyclic AMP.” Academic Press, New York. Rodbell, M., Birnbaumer, L., Pohl. S. L., and Krans. H. M. (1971). The glucagon-sensitive adenyl cyclase system in plasma membranes of rat liver: An obligatory role of guanyl nucleotides in glucagon action. J . Bio/. Chem. 246, 1877-1882. Ross, E. M., and Gilman, A. G. (1980). Biochemical properties of hormone-sensitive adenylate cyclase. Annu. Rev. Biochem. 49, 533-564. Ross. E. M., Howlett, A. C., Ferguson, K. M., and Gilman, A. 0 . (1978). Reconstitution of hormone-sensitive adenylate cyclase activity with resolved components of the enzyme. J . Biol. Chem. 253, 6406-6412. Roth. J. (1973). Peptide hormone binding to receptors: A review of direct studies in vitro. Merab. Clin. Exp. 22, 1059-1073. Shorr, R. G. L., Caron, M . G., and Lefkowitz, R. J . (1980). Isolation and characterization of betaadrenergic receptors from frog erythrocyte membranes. Fed. Proc., Fed. Am. Soc. Exp. Biol. 39, 1616. Shorr, R . G. L., Lefkowitz, R. 1 . . and Caron, M. G . (1981). Purification of the p-adrenergic receptor: Identification of the hormonal binding subunit. J . B i d . Chem. 256, 5820-5826. Smith, S. K.,and Limbird, L. E. (1981). Solubilization of human platelet a-adrenergic receptors: Evidence that agonist occupancy of the receptor stabilizes receptor-effector interactions. Proc. Nail. Acad. Sci. U.S.A. 78, 4026-4030. Stadel, I. M.. and Lefkowitz, R. J. (1979). Multiple reactive sulfhydryl groups modulate the functions of adenylate cyclase-coupled P-adrenergic receptors. Mol. Pharmacol. 16,709-71 8. Stadel, I. M.,De Lean, A . , and Lefkowitz, R. J . (1980). A high affinity agonist P-adrenergic receptor complex is an intermediate for catecholamine stimulation of adenylate cyclase in turkey and frog erythrocyte membranes. J . B i d . Chem. 255, 1436-1441. Stadel, J. M . , Shorr. R . G . L., Limbird. L. E., and Lefkowitz, R. I. (1981). Evidence that padrenergic receptor associated guanine nucleotide regulatory protein conveys guanosine 5’-O-(3-thiotriphosphate)dependent adenylate cyclase activity. J . Biol. Chem. 256, 8718-8723.

66


Stadel, J. M..De Lean, A., and Lefkowitz, R. J. (1982). Molecular mechanisms of coupling in hormone receptor-adenylate cyclase systems. Adv. Enzyrtol. 53, 1-43. Sternweis, P. C., and Gilman, A. G. (1979). Reconstitution of catecholamine-sensitive adenylate cyclase. f . Biol. Chem. 254, 3333-3340. Tharp, D. M., Hoffman, B. B.. and Lefkowitz, R. J . (1981). a-Adrenergic receptors in human adipocyte membranes: Direct determination by ["]yohimbine binding. J. Clin. Endorrinol. Merah. 52, 709-714. Tsai. B. S., and Lefkowitz, R. J. (1979). Agonist-specific effects of guanine nucleotides on alphaadrenergic receptors in human platelets. Mol. Pharmacol. 16, 61-68. U'Prichard, D. C., and Snyder, S. H. (1977). 13H]Epinephrine and [3H]norepinephrine binding to alpha-noradrenergic receptors. L$e Sci. 20, 527-533. U'Prichard, D. C., Greenberg. D. A,, and Snyder, S. H. (1977). Binding characteristics of a radiolabelled agonist and antagonist at central nervous system alpha-noradrenergic receptors. Mol. Pharmacol. 13, 454-473. U'Prichard, D. C., Bylund, D. B., and Snyder. S . H. (1978). (+)-[3H]Epinephrine and ( -)-[7H]dihydroalprenolol binding to P I and Pz-noradrenergic receptors in brain, heart, and lung membranes. J. B i d . Chem. 253, 5090-5102. Vauquelin, G., and Maguire. M. E. (1980). Inactivation of 0-adrenergic receptors by N-ethylmaleimide in S49 lymphoma cells: Agonist induction of functional receptor heterogeneity. Mol. Pharmacol. 18, 362-369. Vauquelin, G., Bottari, S . , and Strosberg. A. D. (1979). Inactivation of P-adrenergic receptors by Nethylmaleimide: Permissive role of P-adrenergic agents in relation to adenylate cyclase activation. Mol. Pharmacol. 17, 163-171. Vauquelin, G., Bottari, S.. Andre, C.. Jacobson. B., and Strosberg, A. D. (1980). Interaction between P-adrenergic receptors and guanine nucleotide sites in turkey erythrocyte membranes. Proc. Nail. Acad. Sci. U.S.A. 77, 3801-3805. Weiland. G. A., Minneman, K. P., and Molinoff, P. B. (1980). Thermodynamics of agonist and antagonist interactions with mammalian P-adrenergic receptors. Mol. Pharmarol. IS, 34 1-347. Williams, L. T., and Lefkowitz, R. J . (1976). Alpha adrenergic receptor identification by 13H]dihydroergocryptine binding. Science 192, 791-793. Williams, L. T., and Letkowitz. R. J. (1977). Slowly reversible binding of catecholamine to a nucleotide-sensitive state of the beta-adrenergic receptor. J. Biol. Chem. 252, 7207-72 13. Williams, L. T., and Lefkowitz, R. J. (1978). "Receptor Binding Studies in Adrenergic Pharmacology." Raven, New York. Williams L. T., Mullikin. D., and Lefkowitz, R . J . (1978). Magnesium dependence of agonist binding to adenylate cyclase-coupled hormone receptors. J. Biol. Ckem. 253, 2984-2989.

CURRENT TOPICS IN MEMBRANES AND TRANSPORT. VOLUME I X

Receptor-Mediated Stimulation and Inhibition of Adenylate Cyclase DERMOT M. F . COOPER' Section on Membrane Regulation Laboratory of Nutrition and Endocrinology National Institute of Arthritis. Diabetes. Digestive and Kidney Diseases National lnstitutes of Health Bethesda. Maryland

I. 11.

Ill.

1v. V. VI. VII. VIII.

IX . X.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Stimulation of Adenylate Cyclase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . GTP-Dependent inhibition of Adenylate Cyclase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bimodally Regulated Adenylate Cyclase Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Receptor Binding of Inhibitory Ligands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Role of GTP Hydrolysis in Inhibition of Adenylate Cyclase . . . . . . . . . . . . . . . . . The Relationship between N, and N , . . . . . .......... Structural Studies on Dually Regulated Ad A. Receptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Cholera Toxin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Calmodulin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D. Ultrastructural Studies. . . . . . . . . . . . . . . . . . . . .. Future Directions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.

67 68 70 71 73 75 76 77 77 77 78 78 78 81 81

INTRODUCTION

The major site of action of many hormones is the adenylate cyclase regulatory complex, which generally responds to these stimuli by increasing cyclic AMP production. In recent years it has become apparent that a variety of hormonal 'Present address: Department of Pharmacology, University of Colorado Medical School, Denver, Colorado 80262.

67 Copyright ,Q 1983 by Academic Prcsr. Inc. All rights of mpmduction in any f0rm rcierved ISBN 0-12-153318-2

68

DERMOT M. F. COOPER

agents are also capable of inhibiting the enzyme, which increases the regulatory flexibility of these systems. Thereby the effects of stimulatory hormones can be attenuated and, in addition, basal cyclic AMP levels can be reduced. GTP plays a central role in determining which regulatory options are available to the enzyme. Evidence accumulated to date would support the notion that distinct GTP regulatory proteins mediate the opposing effects. This article will summarize current studies on dual regulation of adenylate cyclase and contrast these with what is known of the more intensively investigated stimulatory systems, and thus attempt to project future progress in this area.

II. STIMULATION OF AOENYLATE CYCLASE The role of GTP in the stimulation of adenylate cyclase has received intense attention in the last 10 years. A number of excellent reviews of this area have recently appeared (Rodbell, 1980; Ross and Gilman, 1980; Limbird, 1981). As a prelude to a discussion of inhibition of adenylate cyclase it is appropriate to present a brief perspective on stimulation of adenylate cyclase. Many hormones interact with cell surface receptors to transmit a stimulatory signal to the catalytic unit of adenylate cyclase through the intervention of a GTP regulatory protein (termed N,).2 By a mechanism that remains unclear (due largely to a limited knowledge of the components of the system), GTP both decreases the affinity of hormones for their receptors and synergistically amplifies hormonal stimulation of activity (Figs. 1 and 2). In general, the nonhydrolyzable GTP analog Gpp(NH)p promotes the latter action more effectively than the native compound. These and allied findings have led to the development of a general hypothesis, which proposes that hormone binding to receptors leads to the release of previously bound GDP (an ineffective stimulator), which allows occupancy by GTP and, consequently, attainment of a more active R,N,C complex. On hydrolysis of GTP to GDP, the complex reverts to an inactive form, which coincides with the release of hormone (Rodbell, 1980; Ross and Gilman, 1980; Limbird, 1981). Considerable gaps exist in our knowledge of stimulatory adenylate cyclase systems. Quantitative information is lacking on the relationship between hormone occupancy, GTP hydrolysis, and active complex formation. Similarly the relative stoichiometry of R,:N,:C is a subject for speculation. Whether the vari’The following abbreviations are functional assignments which may be represented by one or more distinct proteins: R, and R, are receptors for hornionesineurotransmittersevoking either stimulation o r inhibition, respectively. o f the catalytic activity, C: N, and Ni are the GTP regulatory elements mediating either the stimulation or inhibition of activity. Other abbreviations: CAMP, cyclic 3 ’ 5 adenosine monophosphate; Gpp(NH)p, guanylyl irnidodiphosphate; GP(CH*)P, guanylyl-a,P-methylene phosphonate.

69

STIMULATION AND INHIBITION OF ADENYLATE CYCLASE

-

+ Hormone

Basal

I 10-8

I

I

I

10-7

I 10-6

I 10-5

(GTP)

F w . 1 . Dependence of stirnulatory hormone on GTP for the stimulation of adenylate cyclase. This schematic presentation of the actions of a stimulatory hormone on a "typical" stirnulatory adenylate cyclase shows that, in the absence of GTP, hormone alone (top curve) cannot elicit significant stimulation of activity, but with increasing GTP concentrations, marked amplification is observed. In the absence of hormone (bottom curve) increasing concentrations of GTP elicit little if any increase in activity.

ous elements exist in a preformed complex which is stabilized on interaction with regulatory ligands, or whether there is some degree of independent movement or collision, is unclear. Regulatory components in addition to those already identified may exist. For example, cytoskeletal elements are candidates for supporting roles in these systems. The number of proteins comprising the N, unit appears to

[Hormone]

FIG.2. Effect of GTP on receptor binding of a stirnulatory hormone. This schematic presentation shows displacement of a fixed concentration of a labeled hormone by increasing concentrations of unlabeled hormone in the absence and presence of GTP. A typical effect of GTP on the binding of a stirnulatory hormone to its receptor is shown, i.e., GTP decreases the apparent K d for the hormone. (Compare this effect of GTP with that of the nucleotide on binding to inhibitory receptors: Fig. 5.)

70

DERMOT M . F. COOPER

be either two or three, depending on the source of the purified component. In addition an ADP-ribosylation factor, which permits the N, unit to be ADPribosylated by cholera toxin, may also be an integral component of the N, unit (Sternweis et al., 1981). Notwithstanding the unanswered questions, the central role of the N, unit in mediating the stimulatory effects of hormones is clearly established, and progress in understanding the functioning of this component provides a yardstick against which our understanding of inhibitory regulation can be evaluated.

111.

GTP-DEPENDENT INHIBITION OF ADENYLATE CYCLASE

Early observations on fat cell membranes indicated that GTP concentrations exceeding 1 pM could reduce adenylate cyclase activity (Cryer et n l . , 1969; Harwood et al., 1973; Ebert and Schwabe, 1974). Evidence had also been accumulating that cyclic AMP production could be decreased by a number of agents, such as adenosine and PGE, in adipocytes (Fain er al., 1972), or norepinephrine in platelets (Moskowitz et al., 1971). Several reports demonstrated that adenylate cyclase activity could be inhibited in various broken cell preparations by, for example, muscarinic cholinergic drugs (Murad et al., 1962), norepinephrine (Moskowitz et al., 1971), and opiates (Collier and Roy, 1974). A common factor linking these observations became apparent when it was shown that inclusion of GTP in concentrations exceeding those required for the stimulation of adenylate cyclase by hormones permitted inhibition of the enzyme by

, Stirnulatory Phase

I

Inhibitory Phase

GTP)

FIG. 3. Response of the fat cell adenylate cyclase to GTP in the absence and presence of a stimulatory hormone. This schematic presentation of the effects of GTP on fat cell adenylate cyclase demonstrates characteristic biphasic behavior of a dually regulated adenylate cyclase with GTP. As in Fig. 1, a stirnulatory range of GTP concentrations is required for full amplification of the stirnulatory response. However, after reaching a peak, activity declines in what is referred to as the inhibitory response to GTP.

STIMULATION AND INHIBITIONOF ADENYLATE CYCLASE

71

various putative neurotransmitters, for example, epinephrine in platelets (Jakobs et d.. 1978) and in neuroblastoma x glioma hybrids (Sabol and Nirenberg,

1979) and adenosine in fat cells (Londos et d . , 1978). The GTP requirement for these effects clearly established adenylate cyclase inhibition as a receptor-mediated event, which was somewhat analogous to the stimulation of adenylate cyclase by hormones. Indications that distinct GTP regulatory proteins might mediate stimulation and inhibition of adenylate cyclase came from a series of studies with the fat cell. This enzyme displayed a pronounced biphasic response to GTP; GTP concentrations up to 40 nM increased activity, whereas higher concentrations evoked a steady decline in activity (Fig. 3). The nonhydrolyzable analog Gpp(NH)p did not share the inhibitory response (Cooper et af., 1979). Treatment of fat cell membranes (or cells prior to membrane preparation) with either trypsin (Yamamura et a / . , 1977) or cholera toxin and NAD, or assaying in the presence of Mn2+ resulted in the abolition of the inhibitory response to GTP. In contrast, treatment of membranes with p-hydroxymercuriphenylsulfonic acid eliminated stimulation but retained inhibition by GTP (Cooper e f a / . , 1979). The functional association between the inhibitory response to GTP and GTP-mediated inhibition by adenosine analogs was established by the observation that conditions which eliminated inhibition by GTP also led to the loss of the ability of adenosine analogs to inhibit the enzyme in a GTP-dependent manner (Cooper et al., 1979).

IV. BIMODALLY REGULATED ADENYLATE CYCLASE SYSTEMS A rapid growth in reports of GTP-mediated inhibition of adenylate cyclase by many putative neurotransmitter receptors in a variety of tissues has occurred in recent years. These include opiate (Blume ef a/., 1979), muscarinic cholinergic (Lichahtein et a / . , 1979), and a,-adrenergic (Sabol and Nirenberg, 1979) in neuroblastoma X glioma hybrids; muscarinic cholinergic in myocardium (Jakobs et u l . , 1979; Watanabe et a / ., 1978); dopamine (via a D, receptor) in intermediate pituitary (Cote et a!. , 198 1); adenosine, PGE, , and nicotinic acid in adipocytes (Londos et a/., I98 1 ; Schimrnel et a/., I98 I ; Aktories et a/., 1980); a*adrenergic in platelets (Jakobs et a/., 1978; Cooper and Rodbell, 1979); opiate and adenosine in hippocampus (Girardot et a / . , 1981); opiates in striatum (Law et a / ., 198 I ; Cooper et a/., 1982); angiotensin and a,-adrenergic in liver (Jard et a / ., I98 1); and adenosine in brain cortex (Cooper et a / . , 1980). Common features shared by most of these systems are as follows:' 1 . Biphasic GTP kinetics are almost always encountered (Cooper and Lond-

os, 1982). 'The summary o f the properties of dually regulated adenylate cyciasc systems is drawn from studies in many laboratories. Detailed references are available in the review articles cited in this section.

72

DERMOT M. F. COOPER

FIG. 4. Effect of sodium ion and phenylisopropyladenosine(PIA) on !he response of the fat cell adenylate cyclase to GTP. ( A ) In the absence of NaCI. but in the presence of a stirnulatory hormone. the typical biphasic response of the fat cell enzyme is observed; PIA inhibits activity in the inhibitory GTP phase only (cf. Fig. 3 ) . ( B ) Inclusion of NaCl almost totally reverses the inhibition evoked by GTP alone. so that the response to GTP becomes like that of a stirnulatory system (cf. Fig. 1). However, activity measured in the presence of PIA is virtually unchanged by the presence ofthe salt. The net result is an increase in the absolute inhibition evoked by PIA.

2. GTP concentrations beyond those in the stirnulatory range (ca. l o p 7 M) promote inhibition by putative neurotransmitters and related compounds (Jakobs, 1979; Cooper and Londos, 1982). 3. Where GTP, in the absence of other ligands, causes a decrease in activity, sodium ion (up to 100 mM) reverses this effect (Londos et al., 1981; see Fig. 4). 4. Sodium ion amplifies inhibition by ligands in direct proportion to its reversal of the inhibition promoted by GTP alone (Blume el al., 1979; Londos et a / ., I98 I ) . 5. Gpp(NH)p does not promote inhibition by inhibitory hormones or neurotransmitters, even when it evokes a transient inhibition at early incubation times (Jakobs, 1979; Cooper and Londos, 1982). 6. Inhibition by neurotransmitters is generally less than 60%. except when directed against basal activities, when it may reach 80%. 7. Where multiple inhibitory effectors operate, their effects are nonadditive (Sabol and Nirenberg, 1979; Londos et u l . , 1981). 8. Divalent cations (most effectively MnZ ) selectively abolish GTP-mediated inhibition of activity (Cooper et al., 1979). 9. As with stimulatory ligands, the binding of inhibitory ligands is modulated by GTP (see Section V). 10. Sodium ion and Gpp(NH)p modulate binding of inhibitory ligands, even though these agents may not affect the ability of the inhibitory ligands to attenuate activity (see Section V). +


73

Deviation from these properties is not generally encountered in a wide range of dually regulated systems, unless technical difficulties arise due to, for example, endogenous GTP or inhibitory agent (see Cooper and Londos, 1982).

V.

RECEPTOR BINDING OF INHIBITORY LIGANDS

Extensive studies have been performed on the binding of inhibitory ligands to their receptors. The early studies of Pert and Snyder (1974) on the binding of opiates to brain receptors provided valuable insights for later studies on the inhibition of adenylate cyclase by opiates. In particular, monovalent cations (most effectively sodium) were shown to increase the receptor binding of antagonists and decrease that of agonists. Sub,sequently Blume et al. (1979) found that sodium ion amplified the inhibition of adenylate cyclase in neuroblastoma X glioma cells by opiates. Sodium ions have now been shown to modify the binding of inhibitory ligands in diverse systems, even in situations in which the cation does not modify inhibition of adenylate cyclase by these ligands. For example, in platelets, a decrease in the affinity of the a-adrenergic receptor was encountered (Tsai and Lefkowitz, 1979; Michel et al., 1980), whereas the cation did not modify the inhibition by epinephrine of adenylate cyclase activity (Jakobs, 1979). This latter observation correlates with the lack of inhibition by GTP in the absence of inhibitory ligand in platelet membranes. By contrast, in the fdt cell, a striking effect of sodium on inhibition is seen due to the marked inhibition of enzyme activity by GTP in the absence of inhibitory ligand (Londos et al., 1981). The retention of a sodium ion effect on binding in situations in which it shows no effect on activity may indicate separate loci for these two regulatory events; alternatively, an excess of receptors over catalytic elements would permit discrepant regulation of receptor binding and the function mediated by the receptors. The other alkali metals share these effects of sodium with the following potency: Na+ > K + > Cs+ (Pert and Snyder, 1974; Blume ef a / . , 1979). The sodium effect on binding is not constant with respect to all inhibitory ligands. For example, in brain, the binding of both opiate agonists and antagonists is regulated by sodium (Blume, 1978), whereas only a-adrenergic agonist binding is modified (Greenberg et ul.. 1978). In contrast to the monovalent cations, magnesium ion generally increases agonist affinity (Tsai and Lefkowitz, 1979; U’Prichard and Snyder, 1980). Guanine nucleotides are more consistent in the regulation of the binding of inhibitory ligands. GTP and Gpp(NH)p generally decrease the affinity of inhibitory ligands for their receptors (Tsai and Lefkowitz, 1979; U’Prichard and Snyder, 1980). This is directly analogous to the effects of guanine nucleotides on the binding of stimulatory ligands to their receptors. However, this observation is in conflict with the inability of the nonhydrolyzable GTP analog to promote

74

DERMOT M. F. COOPER

20 mM Mg2+

0 Mg"

PIA(nMI

PIAhM)

PIAbMI

MgC12h M I

FIG. 5 . The effect of GTP on binding to the adenosine receptor of fat cell membranes in the presence of a range of magnesium ion concentrations. Fat cell membranes (30 kg) were incubated with 3HH-labeled Nh-cyclohexyladenosine (CHA) ( 2 nM) in the presence of the indicated concentrations of MgClz in the absence (0)or presence (0) of 20 pM GTP. Note that, in the absence of magnesium ion, GTP decreases binding to the adenosine receptor. However, in the presence of the cation. GTP increases binding to the receptor. A Scatchard analysis of this data reveals that the total number of binding sites is increased by GTP rather than a change in the receptor affinity (Cooper and Gill, in preparation).

inhibition of adenylate cyclase. It might have been anticipated that, since only GTP, and not Gpp(NH)p, can promote adenylate cyclase inhibition by these ligands, the latter compound might not have shared the ability of GTP to modulate binding. The fact that this prediction is not fulfilled again raises the possibility of either separate loci for the regulation of binding compared with function or an excess of inhibitory receptors not in association with catalytic activity, which permits discrepant regulation of total binding compared with a small pool of receptors associated with the enzyme. In tissues in which it has been studied, it appears that magnesium ion can modify the effect of guanine nucleotides on inhibitory ligand binding. In both brain (a,-adrenergic receptor binding) and fat cell membranes (adenosine receptor binding), at magnesium concentrations above 1 mM GTP increases the number of binding sites for the inhibitory ligands, whereas at lower cation concentrations GTP decreases affinity for the ligands (U'Prichard and Snyder, 1980; Cooper and Gill, in preparation, Fig. 5 ) . The apparently wide diversity which exists in the regulation of the binding of inhibitory ligands to their receptors compared with that of stimulatory agents


75

emphasizes our rather primitive understanding of both the interaction of inhibitory receptors with the putative inhibitory GTP regulatory components and the precise transduction mechanisms which translate the binding event into an inhibitory response.

VI.

THE ROLE OF GTP HYDROLYSIS IN INHIBITION OF ADENYLATE CYCLASE

The most persuasive evidence that GTP hydrolysis plays a role in hormonally mediated inhibition of adenylate cyclase is the observation that Gpp(NH)p, the nonhydrolyzable analog, will not promote inhibition under any assay conditions. A characteristic of dually regulated systems is a transient inhibitory response to low concentrations of Gpp(NH)p (Ebert and Schwabe, 1974; Girardot et a / . , 1981; Cooper and Londos, 1982; Sulakhe et al., 1977). Under such conditions, in the absence or presence of sodium ion, inhibitory ligands will not affect activity. An obvious interpretation of this widely encountered finding is that GTP hydrolysis is an absolute requirement for inhibition. The hydrolysis product, GDP, is not required, since GDP (or GP(CH,)P], under conditions in which care is taken to prevent phosphorylation to the triphosphate, will not promote inhibition (Cooper and Schlegel, unpublished; Cooper and Londos, 1982). These observations contrast with the ready interchange of GTP with Gpp(NH)p in stimulation of adenylate cyclase, where GTP hydrolysis does not seem to be a stringent requirement for enzyme stimulation. Supportive evidence for an obligatory role for GTP hydrolysis is available from studies with cholera toxin. As mentioned in Section 111, cholera toxin treatment abolishes both GTP inhibition and inhibition promoted by adenosine in a GTP-dependent manner in fat cell membranes (Cooper ef a / . , 1979). Cholera toxin invariably enhances hormonal stimulation in stimulatory systems by a mechanism believed to involve the inhibition of a specific GTPase activity associated with the stimulatory (N,) unit. Consequently, the observations with the fat cell could be interpreted to imply that toxin treatment also inhibited a GTPase activity associated with the inhibitory (N,) unit. However, this effect of cholera toxin is not universally encountered in dually regulated systems. Inhibitory effects were retained following toxin treatment of Chinese hamster ovarian (CHO), neuroblastoma x glioma hybrid cells, and platelets (Evain and Anderson, 1979; Propst and Hamprecht, 1981; Jakobs and Schultz, 1979). Very recently the first direct evidence pertaining to this issue has been presented. In neuroblastoma X glioma and platelet membranes, a GTPase activity has been detected which can be stimulated by opiates and a-adrenergic agents, respectively (Koski and Klee, 198 1; Aktories and Jakobs, 198 I ) . Substrate spec-

76

DERMOT M. F. COOPER

ificities were not examined in these studies, thus a specific GTPase activity was not unequivocally demonstrated. Nevertheless, a twofold stimulation was achieved, and it is tempting to speculate that the GTPase activity measured is relevant to the inhibition of adenylate cyclase mediated by the receptors in these tissues. The rather complex assay mixtures utilized may allow transphosphorylation of the terminal phosphate of GTP to ATP, which is included in the assay. Neither study has attempted to determine the source of the Pi released into the medium; therefore, the possibility must be considered that inhibitory ligands may be stimulating an ATPase activity in plasma membranes in a manner analogous to that reported for insulin and catecholamines (Resh et al., 1980; Titheradge et al., 1979). However, notwithstanding the reservations raised concerning these recent studies, the accumulated evidence from more indirect studies would perhaps have anticipated a more central role for GTPase in inhibitory regulation than in the case of stimulatory regulation.

VII. THE RELATIONSHIP BETWEEN N, AND N, Definitive evidence is lacking on whether N, and Ni are distinct proteins or merely functional notations. The broadly descriptive options which might be considered include ( I ) that the N unit (or complex, see Section 11) is constant in all adenylate cyclase systems and that the functions associated with N, and Ni merely reflect the association of the N unit with R, or Ri, respectively, (2) N, and N, are distinct regulatory protein complexes which share a common catalytic TABLE 1 S U M M A R Y Ol- THE P R O P E R T l t S ASSOC‘lATtD WITH T H t

TWO GTP RLCUIATOKY FUNCTIONS“

Property

Ni

111 IV 111 111 111, IV

I pM”

GTP requirement (EDs,)) Sodium ions Cholera toxin Me2 Gpp(NH)p Cholera toxin labeled bands Mild trypsin treatment Effect of GTP, Gpp(NH)p on binding GTP on basal cyclase Sodium effect on binding

-. T -. 5 1

+

.

Section

- 3

5

Additional bands -3

.1 Kci.

1 1 Kd

.

7

R

Vlll 111 V 111 V

ampiification of the function; 1 dampening of the function. “Section” refers -, No effect; to sections in the text in which these features are discussed. When stirnulatory systems are considered.


77

unit, or (3) Ni may be a modified form of N, (an oligomeric form; an association with an exogenous factor or protein, such as calmodulin; etc.) which may or may not be stabilized or promoted upon interaction with Ri. Differences in properties between the functional entities referred to as N, and Ni are summarized in Table 1. These differences justify the functional assignments and, with other evidence discussed here, support the view that fundamental differences exist between the two regulatory systems. Nevertheless a full appreciation of their properties will become available only by further structural studies .

VIII.

STRUCTURAL STUDIES ON DUALLY REGULATED ADENYLATE CYCLASE SYSTEMS

A. Receptors

Apart from the preliminary studies on the fat cell indicating selective abolition and retention of one of the two effects mediated by GTP (see Section IV), most attention has focused on the receptors in dually regulated systems. For instance, opiate receptors have been solubilized with full retention of their agonist specificity (Simonds et al., 1980). Another solubilized opiate receptor preparation retained the ability of sodium ion to modify binding (Ruegg et al., 1981), although no effects of GTP were detected. This important observation suggests that the sodium site may be associated with the receptor rather than the GTP regulatory unit. The a-adrenergic receptor from liver has been solubilized and partially purified (Guellaen et al., 1979). The irreversibly binding antagonist, phenoxybenzamine, was utilized to monitor the presence of the receptor through various purification proccclures. Although phenoxybenzaminc cannot readily be removed from the receptor preparation, this material is quite suitable for generating antibodies which may be utilized to identify and purify unoccupied receptors. Prior incubation of platelet plasma membranes with a,-adrenergic agonists stabilizes a higher molecular weight form of the receptor than that observed following incubation with antagonists. This data may suggest that agonists stabilize interaction between receptor and Ni unit (Smith and Limbird, 1981). 8. Cholera Toxin

Since cholera toxin, with NAD, modifies the functions ascribed to both N, and N,, it is conceivable that with [32P]NAD, protein bands additional to those encountered in stimulatory systems might be detected on sodium dodecyl sulfate (SDS) electrophoresis following exposure to the toxin of dually regulated sys-

78

DERMOT M. F. COOPER

tems. When a range of plasma membrane preparations was compared, those subject to dual regulation showed additional labeled bands; in fat cell and CHO cells, a 52,000 and a 54,000 M, band were detected; in platelet, a 58,000 M, band was detected, in addition to the widely observed 42,000 M , band (Cooper et af.,1981). The relationship of these additional proteins to the N, unit, or Ni function. remains to be established.

C. Calmodulin The hippocampal adenylate cyclase is a dually regulated system (Section IV) which can be inhibited by opiates and adenosine analogs in a GTP-dependent manner. Calmodulin appears to play a role in this system, since its removal by EGTA treatment results in the loss of inhibition by the opiates and adenosine. Addition of calmodulin restores the effect (Girardot et af., 1981). This association does not seem to be generally applicable to inhibitory systems, since the platelet and fat cell systems, for example, are not affected by EGTA treatment (Cooper and Londos, 1982). However, in the case of the hippocampal system (and possibly other neural systems), calmodulin may provide a means for identifying inhibitory components.

D. Ultrastructural Studies Recent electron microscopic studies indicate that both opiate and cholinergic receptors occur in clusters on the cell surface (Hazum er a / ., 1979; Peng et a / ., I98 1). Irradiation inactivation studies also indicated that very large structures mediated stimulation and inhibition of fat cell adenylate cyclase (Schlegel et a/., 1980). Such findings are readily accommodated in view of the multiplicity of stimulatory and inhibitory neurotransmitters converging on a common pool of catalytic activity (see Section IV). These structures, which if composed of heterogeneous stimulatory and inhibitory receptors with their associated N units, would resemble multienzyme complexes and would provide a ready means of achieving the nonadditive stimulation and inhibition of adenylate cyclase by different neurotransmitters discussed in Section IV.

IX. FUTURE DIRECTIONS A number of putative neurotransmitters and peptides which have been identified in brain as yet do not have a measurable function in isolated membrane preparations. It seems likely that some of these compounds, including histamine, serotonin, GABA, glutamine, ACTH, substance P, VIP, a-MSH, and others, will turn out to utilize GTP inhibitory pathways.


79

Whether the catalytic unit of adenylate cyclase is identical in dually regulated and in stimulatory systems must be determined. It is conceivable that the complexity of catalytic units varies as a function of the degree of regulation to which they are subject. Further progress in understanding the structural nature of these systems will require a combination of approaches, including ultrastructural studies, solubilization and reconstitution of components, identification of mutants lacking one or another of the regulatory elements, coupled with the fusion/complementation approach pioneered by Orly and Schramm ( 1976). A perplexing question, which must reflect a fundamental property of inhibitory systems, arises from the partial inhibition evoked by inhibitory neurotransmitters. Since inhibition ranges from 20 to 80% maximally (Cooper and Londos, 1982; Jakobs, 1979) and it is generally assumed that a gross overproduction of cAMP is produced in response to stirnulatory hormones, the physiological significance of this regulation may be doubted. However, partial inhibition of adenylate cyclase combined with the presence of phosphodiesterase in intact cells can result in substantial reduction of intracellular cAMP levels. The striking inhibition of cAMP production in the fat cell by adenosine and its analogs (leading to a marked inhibition of lipolysis) and in platelets by epinephrine (resulting in platelet aggregation) must reflect such a situation. A question related to that raised above pertains to the different consequences of inhibiting basal versus hormone-stimulated activity. Clearly, inhibition of hormone-stimulated adenylate cyclase will result in attenuation of the process(es) stimulated by the elevated cAMP levels. However, it is conceivable that in the absence of stimulatory hormone the basal cAMP levels may maintain distinct processes in an activated state due to different sensitivity to phosphorylation. Thus inhibition of basal cAMP production would result in modulation of processes separate from those governed by stimulatory hormones. Potential examples of the types of process maintained by basal cAMP levels might be fundamental cellular mechanisms concerned with maintenance of normal functions, such as ion transport. Thus it is of interest to examine the results of the exposure of intact cells to inhibitory agents, regardless of whether the processes concerned have previously been implicated with cAMP as a second messenger. The recent finding of angiotensin I1 and a-adrenergic inhibition of adenylate cyclase in liver (Jard et al., 1981), which had previously been considered a simple stimulatory system, raises some intriguing possibilities. Inclusion of high concentrations of EDTA was required during all stages of the preparation of the plasma membranes for the observation of this effect. It has also been known for some time that both GTP and sodium ion affect angiotensin binding to adrenal cortex membranes (Glossmann ef al., 1974) (without EDTA treatment), although angiotensin does not inhibit adrenal cortical adenylate cyclase. Thus the possibility is raised that an R N complex existed for angiotensin, which had not been linked to adenylate cyclase prior to the chelator treatments, but to some

80

DERMOT M. F. COOPER

other process. Certainly a number of receptors mediating inhibition of adenylate cyclase have also been implicated in either calcium transport or phosphatidylinositol metabolism, e.g., muscarinic cholinergic, a,-adrenergic, and angiotensin (Jones and Michell, 1978). The findings discussed above raise the possibility that switching of the function served by an RN complex can occur physiologically in addition to the experimental means presented above. An alternative interpretation of these results would be that in normal liver membranes, the expression of inhibitory activity is suppressed by interaction of the Ni unit with divalent cation, as discussed in Section IV, and revealed upon chelator treatment. Sharma el al. (1975) described increased adenylate cyclase activity following prolonged exposure of neuroblastoma X glioma hybrid cells to morphine without alteration in receptor number. The situation was presented as a model system for tolerance and addiction to opiates. Current appreciation of the existence of stimulatory and inhibitory GTP regulatory protein interactions may provide a means of understanding the basis of this observation. Recently, receptor-mediated inhibition of Gpp(NH)p-activated adenylate cyclase activity by progesterone in Xenopus oocytes and by a-mating factor in yeast (Finidori-Lepicard er al., 198 1; Sadler and Maller, 198 1; Liao and Thorner, 1980) has been reported. In both cases fundamental developmental changes correlate with these inhibitions. It is possible that modified forms of Ni with less severe restrictions on the terminal diphosphate bond of the guanine nucleotide mediate these effects.

FIG.6. Schematic representation of dually regulated adenylate cyclase with suggested sites of action of various regulators. In this scheme broken arrows indicate inhibition or suppression of a function: solid arrows indicate promotion or enhancement of a function; the heavy directional lines between R and N units identify the guanine nucleotides which promote communication from R to N or from N to R, respectively.


81

X. CONCLUSIONS The functional components involved in dual regulation of adenylate cyclase and the site of action of some modulators of activity are presented schematically in Fig. 6. The central role of guanine nucleotides is evident. Appreciation of the potential importance of GTP has proved to be the key to our current understanding of these systems. Evidence is accumulating from measurements of both activity and regulation of binding of the two classes of receptor that separate GTP regulatory proteins are associated with inhibition and stimulation. In the next few years there will be an increasing awareness of the physiological importance of these systems as well as insights into their structural composition. ACKNOWLEDGMENTS

I would like to acknowledge the useful comments of my colleagues Drs. D.L. Gill, R. Honnor, and R. T . Simpson on this manuscript. I would also like to thank Dr. M. Rodbell for his continuing support and encouragement and Ms. Bonnie Richards for her expert secretarial assistance. REFERENCES Aktories, K., and Jakobs, K. H. (1981). Epinephrine inhibits adenylate cyclase and stimulates a CTP-ase in human platelet membranes via a-adrenoceptors. FEBS Left. 130, 235-238. Aktories. K.. Jakobs, K . H.. and Schultz, G . (1980). Nicotinic acid inhibits adipocyte adenylate cyclase in a hormone-like manner. FEBS Lert. 115, 11-14. Blume, A. J . (1978). Interaction of ligands with the opiate receptors of brain membranes: Regulation by ions and nucleotides. Proc. Natl. Acad. Sci. U.S.A. 75, 1713-1717. Blume. A. J . . Lichtshtein, D.. and Boone, G . (1979). Coupling of opiate receptors to adenylate cyclase: Requirement for Na+ and GTP. Proc. Nad. Acad. Sci. U.S.A. 76, 5626-5630. Collier. H. 0. J., and Roy, A. C . (1974). Morphine-like drugs inhibit the stimulation by E prostaglandins of cyclic AMP formation by rat brain homogenate. Nature (London) 248, 24-27. Cooper, D. M. F.. and Londos. C. (1982). GTP-dependent stimulation and inhibition of adenylate cyclase. I n “Horizons in Biochemistry and Biophysics” (L. D. Kohn, ed.), Vol. 6, pp. 309- 333. Cooper, D. M. F., and Rodbell, M. (1979). ADP is a potent inhibitor of human platelet plasma membrane adenylate cyclase. Nature (London) 282, 5 17-5 18. Cooper, D. M. F.. Schlegel, W . . Lin, M. C . , and Rodbell. M. (1979). The fat cell adenylate cyclase system. Characterization and manipulation of its bimodal regulation by GTP. J . Biol. Chem. 254, 8927-8930. Cooper. D.M. F.. Londos, C . , and Rodbell, M. (1980). Adenosine-receptor-mediated inhibition of rat cerebral cortical adenylate cyclase by a GTP-dependent process. Mol. Pharmacol. 18, 598-601. Cooper, D. M. F.. Jagus, R., Somers, R . L., and Rodbell, M. (1981). Cholera toxin labels diverse CTP-modulated regulatory proteins. Biochem. Biophvs. Res. Commun. 101, I 179-1 185. Cooper. D. M. F.. Londos, C., Gill, D. L., and Rodbell, M. (1982). Opiate receptor-mediated inhibition of adenylate cyclase in rat striatal plasma membranes. J . Neurocliem. 38, 1164-1167. Cote. T. E . , Grewe, C . W., and Kebabian, J. W. (1981). Stimulation of a D-2 dopamine receptor in the intermediate lobe of the rat pituitary gland decreases the responsiveness of the P-adrenoceptor: Biochemical mechanism. Endocrinology 108, 420-426.

82

DERMOT M. F. COOPER

Cryer, P. E., Jarrett. L., and Kipnis, D. M. (1969). Nucleotide inhibition of adenylate cyclase activity in fat cell membranes. Biochim. Biophy. Acta 177, 586-590. Ebert. R., and Schwabe, U. (1974). Biphasic effect of 5’-guanylylimidodiphosphateon fat cell adenylate cyclase. Naunyn-Srhmiedebergs Arch. Pharmarol. 296, 297-313. Evain, D., and Anderson, W. B. (1979). Inhibitory effect of guanyl nucleotides toward adenylate cyclase activity of Chinese hamster ovary cell membranes activated in vitro by cholera toxin. J . B i d . Chem. 254, 87268729. Fain, J. N., Pointer, R. H., and Ward, W. F. (1972). Effects of adenosine nucleosides on adenylate cyclase, phosphodiesterase, cyclic adenosine monophosphate accumulation, and lipolysis in fat cells. J . Biol. Chem. 247, 6866-6872. Finidori-Lepicard, I., Schorderet-Slatkine, S., Hanoune. J . , and Baulieu, E. E. (1981). Progesterone inhibits membrane-bound adenylate cyclase in Xenopus laevis oocytes. Narure (London) 292, 255-257. Girardot, J. M., Cooper, D. M. F., and Kempf, J. (1981). Regulation of rat hippocampal adenylate cyclase by guanyl nucleotides. Adv. Cyclic Nurleotide Res. 14, 657. Glossmann, H.. Baukal, A., and Catt. K. I. (1974). Properties of angiotensin I1 receptors in the bovine and rat adrenal cortex. J. Biol. Chem. 249, 664-666. Greenberg, D. A., U’Prichard, D. C., Sheehan. P., and Snyder, S. H. (1978). a-Noradrenergic receptors in the brain: Differential effects of sodium on binding of [3H] agonists and pH] antagonists. Brain Res. 140, 378-384. Guellaen, G., Aggerbeck, M., and Hanoune, J. (1979). Characterization and solubilization of the adrenoreceptor of rat liver plasma membranes labeled with (3H]phenoxyhenzamine. J . Biol. Chem. 254, 10761-10768. Hanvood, J. P., Low, H., and Rodbell, M. (1973). Stimulatory and inhibitory effects of guanyl nucleotides on fat cell adenylate cyclase. J. Biol. Chem. 248, 6239-6245. Hazum, E., Chang. K. J., and Cuatrecasas, P. (1979). Opiate (enkephalin) receptors of neuroblastoma cells: Occurrence in clusters on the cell surface. Srienre 206, 1077-1079. Jakobs, K. H. (1979). Inhibition of adenylate cyclase by hormones and neurotransmitters. Mol. Cell. Endocrinol. 16, 147- 156. Jakobs, K. H., and Schultz, G. (1979). Different inhibitory effect of adrenaline on platelet adenylate cyclase in the presence of GTP plus cholera toxin and of stable GTP analogues. NaunynSchmiedeberg’s Arch. Pharmacol. 310, 121-127. Jakobs, K. H., Saur, W., and Schultz, G. (1978). Inhibition of platelet adenylate cyclase by epinephrine requires GTP. FEES Len. 85, 167-170. Jakobs, K. H., Aktories, K., and Schultz, G. (1979). GTP-dependent inhibition of cardiac adenylate cyclase by muscarinic cholinergic agonists. Naunyn-Srhmiedeberg’s Arch. Pharmarol. 310, 113-1 19. Jard, S., Cantau, B . , and Jakobs, K. H. (1981). Angiotensin I1 and a-adrenergic agonists inhibit rat liver adenylate cyclase. J . B i d . Chem. 256, 2603-2606. Jones, L. M., and Michell, R. H. (1978). Stimulus-response coupling at a-adrenergic receptors. Biochem. Sor. Trans. 6, 673-693. Koski, G., and Klee, W. A. (1981). Opiates inhibit adenylate cyclase by stimulating GTP hydrolysis. Proc. Natl. Acad. Sci. U.S.A. 78, 4185-4189. Law, P. Y.,Wu, J., Koehler, I . E., and Loh, H. H. (1981). Demonstration and characterisation of opiate inhibition of the striatal adenylate cyclase. J . Neurochem. 36, 1834-1846. Liao, H., and Thomer, J. (1980). Yeast mating pheromone a factor inhibits adenylate cyclase. Proc. Natl. Acad. Sci. U.S.A. 77, 1898-1902. Lichtshtein, D., Boone, G., and Blume, A. J. (1979). Muscarinic receptor regulation of NG108-IS adenylate cyclase: Requirement for Na+ and GTP. J . Cyclic Nucleoride Res. 5, 367-375. Limbird, L. E. (1981). Activation and attenuation of adenylate cyclase. Biochem. J . 195, 1-13. Londos, C., Cooper, D. M. F., Schlegel, W., and Rodbell, M. (1978). Adenosine analogs inhibit


83

adipocyte adenylate cyclase by a GTP-dependent process: Basis for actions of adenosine and methylxanhines on cyclic AMP production and lipolysis. Proc. Nut/. Aiwd. Sci. U . S . A . 75, 5362-5366. Londos. C.. Cooper. D. M. F.. and Rodbell, M. (1981). Receptor-mediated stimulation and inhibition of adenylate cyclases: The fat cell as a model system. Adv. Cvclic Nucleotide Res. 14, 163- I7 I. Michel, T . . Hoffman. B . B.. and Lefkowitz, R . J. (1980). Differential regulation of a,-adrenegic receptor by Na+ and guanine nucleotides. Nature (London) 288, 709-71 I . Moskowitz, J.. Harwood. J. P.. Reid, W . D.. and Krishna. 0. (1971). The interaction of norepinephrine and prostaglandin E, on the adenyl cyclase system of human and rabbit blood platelets. Biochim. Biophys. Actu 230, 279-285. Murad, F., Chi. Y.-M.. Rall, T. W., and Sutherland. E. W. (1962). Adenyl cyclase 111. The effect of catecholamines and choline esters on the formation of adenosine 3',5'-monophosphate by preparations from cardiac muscle and liver. J . B i d . Chem. 237, 1233-1238. Orly, J . . and Schramm. M . (1976). Coupling of catecholamine receptor from one cell with an adenylate cyclase from another cell by cell fusion. Proc. Nut/. Acud. Sci. U.S.A. 73, 44104414. Peng, H. B.. Cheng. P.-C., and Luther, P. W. (1981). Formation of ACh receptor clusters induced by positively charged latex beads. Nature (London) 292, 831-834. Pert. C. B . , and Snyder. S. H. (1974). Opiate receptor binding of agonists and antagonists affected differentially by sodium. Mol. Phurmucol. 10, R68-879. Propst. F., and Hamprecht. B . (1981). Opioids, noradrenaline and GTP analogs inhibit cholera toxin activated adenylate cyclase in neuroblastoma X glioma hybrid cells. J . Neurochem. 36, 580-588. Resh. M. D., Nemenoff. R . A., and Guidotti, G . (1980). Insulin stimulation of (Na+,K+)-adenosine triphoaphatase-dependent H'Rb+ uptake in rat adipocytes. J . Biol. Chem. 255, 10938-10945. Rodbell, M. (19x0). The role of hormone receptors and GTP-regulatory proteins in membrane transduction. Nuture (London) 284, 17-22, Ross, E. M.. and Gilman. A. G. (1980). Biocheniical properties of hormone-sensitive adenylate cyclase. Annu. Rev. Biochem. 49, 533-564. Ruegg, U. T . , Cuenod, S., Hiller, J. M.. Gioannini, T., Howells, R . D.. and Simon, E. J. (1981). Characterisation and partial purification of solubilized opiate receptors from toad brain. Proc. Nut/. Acud. Sci. U . S . A . 78, 4635-4638. Sabol. S . L.. and Nirenberg, M. (1979). Regulation of adenylate cyclase of neuroblastoma X glioma hybrid cells by a-adrenergic receptors. 1. Inhibition of adenylate cyclase mediated by areceptors. J . B i ~ l Chem. . 254, 1913-1920. Sadler. S . E., and Maller. J. L. (1981). Progesterone inhibits adenylate cyclase in Xenopits oocytes. Action on the guanine nucleotide regulatory protein. J . Biol. Chem. 256, 6368-6373. Schimmel. R . J., McMahon. K. K.. and Serio, R . (1981 j . Interactions between alpha-adrenergic agents. prostaglandin E l , nicotinic acid and adenosine in regulation of lipolysis in hamster epididynial adipocytes. Mol. Phurmucol. 19, 248-255. Schlegel. W.. Cooper, D. M. F.. and Rodbell, M. (1980). Inhibition and activation of fat cell adenylate cyclase by GTP is mediated by structures of different size. Arch. Biochem. Biophys. 201, 678-6x2. Sharma, S. K.. Klee, W. A.. and Nirenberg, M. (1975). Dual regulation of adenylate cyclase accounts for narcotic dependence and tolerance. Proc. Nut/. Acud. Sci. U.S.A. 72, 3092-3096. Simonds. W . F . , Koski. G.. Streaty, R . A.. Hjelmeland, L. M.. and Klee, W. A. ( 1980). Solubilisation of activc opiate receptors. Proc. Nail. Acud. Sci. U . S . A . 77, 4623-4627. Smith. S . K.. and Linibird. L. E. (1981). Solubilization of human platelet a-adrenergic receptors:

84

DERMOT M. F. COOPER

Evidence that agonist occupancy of the receptor stabilizes receptor-effector interactions. Proc. Nut/. Acad. Sci. U.S.A. 78, 4026-4030. Sternweis, P. C.. Northup, I . K., Hanski. E., Schleifer, L. S . . Smigel. M. D., and Gilman, A. G . (1981). Purification and properties of the regulatory component (GIF) of adenylate cyclase. Adv. Cyclic Nucleotide Res. 14, 23-36. Sulakhe, P. V.. Leung, N. L. K., Arbus, A. T., Sulakhe, S . J., Jan. S. H.. and Narayanan. N. ( 1977). Catecholamine-sensitive adenylate cyclase of caudate nucleus and cerebral cortex. Effects of guanine nucleotides. Biochem. J. 164, 67-74. Titheradge, M . A., Stringer. J. L., and Haynes, R. C.. Jr. (1979). The stimulation of the mitochondrial uncoupler-dependent ATP-ase in isolated hepatocytes by catecholamines and glucagon and its relationship to gluconeogenesis. Eur. J . Biochenz. 102, 117-124. Tsai. B. S . , and Lefkowitz. R. J . (1979). Agonist-specific effects of guanine nucleotides on alphaadrenergic receptors in human platelets. Mol. Phurmacol. 16, 61-68. U’Prichard. D. C., and Snyder, S. H . (1980). Interactions of divalent cations and guanine nucleotides at az-noradrenergic receptor binding sites in bovine brain mechanisms. J . Neurochem. 34, 385-394. Watanabe. A. M., McConnaughey, M. M.. Strawbridge. R. A.. Fleming. J . W . , Jones, L. R.. and Besch, H. R.. Jr. ( 1978). Muscarinic cholinergic receptor modulation of P-adrenergic receptor affinity for catecholamines. J . B i d . Chern. 253, 4833-4836. Yamamura, H., Lad. P. M., and Rodbell. M. (1977). GTP stimulates and inhibits adenylate cyclase in fat cell membranes through distinct regulatory processes. J . B i d . Chem. 252, 7964-7966.

CURRENT TOPICS IN MEMBRANES A N D TRANSPORT, VOLUME I 8

Desensitization of the Response of Ade ny Iate Cyclase to CatechoIamines JOHN P . PERKINS Department of’Pharmacology University of North Curolinu Chapel Hill. N m h Carolina

I.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. Scope ofthe Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Catecholatnine-Induced Desensitization of Intact Cells . . . . . . . . A. Origin and Characteristics of the Human Astrocytorna Cell B. Analysis of Rates of Synthesis and Degradation of Cyclic AMP in Whole Cells IV. Catecholamine-Induced Changes in Adenylate Cyclase and in PAR Binding Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Separation of Native and Desensitized PAR. . . . . . . . . . VI . Receptor Endocytosis as a Mechanism for Agonist-lndu VII. A Kinetic Model for Agonist-Induced Desensitization. . . . . . . . . . . . . . . . . . . . . . . . . . VIII. Differential Expression of PAR during Growth of 1321Nl Cells . . . . . . . . . . . . . . . . . 1x. Down-Regulation of PAR and the Recovery of Lost Receptors . . . . . . . . . . . . . . . . . . X . Isoproterenol-Induced Changes in Agonist Binding Properties of Intact 132lN I Cells XI. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . ..................... 11. 111.

1.

85 87 88 88 91 93 95 97 98 100 100 103 104

I06

INTRODUCTION

The initial effect of catecholamines on the j3-adrenergic receptor (PAR) linked adenylate cyclase system is to increase the rate of formation of cyclic AMP. This stimulatory effect occurs essentially instantaneously in most cells. However, secondary effects of exposure to catecholamines can be detected in certain cells within 1 minute. The functional consequence of these secondary reactions is a reduction in the rate and/or extent of cyclic AMP accumulation. Recent studies of such inhibitory events have revealed a complicated series of reactions, in85 Copyrlgh! 0 1983 by Academic Prerr. Inc. All rights 0 1 reproduction in any form rcscrved. ISBN 0-12-IS331R-2

86

JOHN P. PERKINS

duced by receptor agonist, designed to down-regulate responsiveness of the cell to continued or subsequent challenge with a catecholamine. The purpose of such a cellular capability is not yet known, but the process of ligand-induced downregulation of receptor function has been shown to be general in nature; for example, the receptor systems for insulin, glucagon, epidermal growth factor (EGF), thyroid-stimulating hormone (TSH), leutinizing hormone (LH), folliclestimulating hormone (FSH), prostaglandins, histamine, acetylcholine, and adenosine all exhibit agonist-induced modification in function. However, while such down-regulating or desensitizing reactions appear to occur in general, the specific molecular mechanisms involved may be quite different for different receptor systems. About the only property shared in common by the receptor systems listed above is that in each case the receptor is a cell surface protein involved in transduction of the effects of an extracellular information molecule. Our studies of this general phenomenon have focused on those receptors that mediate the stimulatory effects of catecholamines on adenylate cyclase. The

Y

GDP

FIG. 1. This model indicates the hypothetical interactions of a catecholamine (H), the P-adrenergic receptor (R), a guanine nucleotide binding protein (N), the catalytic protein (C), and guanine nucleotides (GDP and GTP). The model indicates that RH is able to bind to N, resulting in the release of free GDP and the formation of HRN. The RN complex has higher affinity for H than does R. The formation of HRN is rate limiting in the activation process, and, in the presence of GTP, HRN is rapidly coverted to N-GTP and HR; thus, the role of the hormone receptor system is to effect the conversion of N-GDP to N-GTP. Once formed, N-GTP interacts with C to form the enzymically active complex C-N-GTP. The lifetime of the active complex is determined by the activity of a GTPase (probably an integral part of N) which hydrolyzes the bound GTP to release Pi, with the subsequent regeneration of C and N-GDP. In the absence of GTP, addition of H leads to formation of HRN in amounts sufficient to change the apparent K, of the system for H. Thus, in the absence of GTP, agonists (H) exhibit binding characteristic of interaction at two sites (R and RN). In the intact cells. or upon addition of GTP to membranes, the amount of HRN would be small because its rate of formation is postulated to be the limiting step in the intact system. Under these conditions, agonists (H) would exhibit binding characteristic of the reaction R + H RH, namely low-affinity binding to a single type of site. (Modified from Su et al., 1980.)For detailed reviews of the hormone-sensitive adenylate cyclase system see Ross and Gilman (1980) and Lirnbird (1981).

*

LIGAND-INDUCED CHANGES IN P-RECEPTOR FUNCTION

87

most compelling reasons for this choice of system include

I . The identities of the major molecular components of the hormone-sensitive adenylate cyclase have been established (see Fig. I ) . 2. A plausible working model for the activation sequence initiated by hormones has been described (see Fig. 1). 3. The amount and functional state of the three major components of the system can be determined by independent assay in subcellular membrane fractions and by reconstitution of the previously resolved components. 4. Assays exist for measurement in intact cells of PAR binding properties, the rate of formation of cyclic AMP, and the rate of cyclic AMP degradation. Thus, as compared to polypeptide hormone receptor systems for insulin and EGF, which have been well studied in terms of down-regulation, the receptorlinked adenylate cyclase systems allow one to investigate the relation between changes in binding events and changes in functional events that occur essentially instantaneously with ligand binding. The PAR-linked adenylate cyclase is the most thoroughly studied receptor-linked adenylate cyclase system, due in large part to the variety of agonist, partial agonist, and antagonist ligands for the PAR that are available for use as experimental probes (e.g., see Minneman et al.. 1981). A variety of high-specific-activity, selectively binding radioligands are available commercially or readily synthesized from available starting materials (Aurbach et al., 1974; Williams and Lefkowitz, 1978; lnnis et a / . , 1979; Barovsky and Brooker, 1980; Engel et d., 1981; Staehelin and Simon, 1982.)

II. SCOPE OF THE REVIEW The observation that chronic exposure to catecholamines of the PAR-linked adenylate cyclase system leads to a reduction in responsiveness has been verified in studies of intact animals and humans (reviewed in Perkins er al., 1979, 1982; also see Aarons et al., 1980; Fraser et a/.. 1981). In addition, it can be shown that chronic underexposure to catecholamines in intact animals leads to an increase in responsiveness. Thus, the concept has evolved that this receptor-enzyme system exists in vivo in various states of responsiveness that depend, inversely, on the recent level of exposure to catecholamines. We will not attempt here to provide evidence of the physiological significance of ligand-induced adaptive changes in the responsiveness of cells to hormones; however, readers are referred to other recent reviews which provide evidence of this nature (Perkins er al., 1979, 1982; Lefkowitz et a/.. 1980). In this article, the biochemical mechanisms involved in such adaptive responses will be reviewed, primarily as they have come to be understood from experiments utilizing either cultured mammalian cells or erythrocytes from mammals, birds, and frogs. The review

88

JOHN P. PERKINS

will further focus on experiments carried out over the past 10 years in my laboratory, the past 6 years in collaboration with T. K . Harden. Work by other investigators will be referred to primarily to demonstrate general points or to highlight areas of controversy.

111.

CATECHOLAMINE-INDUCED DENSENSITIZATION OF INTACT CELLS

A. Origin and Characteristics of the Human Astrocytoma Cell Line 1321N1 Since 1971 we have utilized cultured human astrocytoma cells as a convenient experimental system in which to study the regulation of cyclic AMP metabolism by hormones. The original source of the cell line was a primary culture of a Grade 111 astrocytoma-glioblastoma (multiform) isolated by Ponten and Maclntyre (1968) and designated 118 MG-C. A subclone ( 1 181Nl) of this culture that maintained a more consistent spindle morphology was isolated (MacIntyre er a / ., 1972) and characterized in terms of its adenylate cyclase, cyclic AMP-phosphodiesterase, and cyclic AMP-dependent protein kinase activities (Perkins et al., 1971). For the past 7 years we have primarily studied a subclone of the 1 1 8 1N I line designated 1321 N 1 . These cells respond to catecholamines (Clark and Perkins, 197 I), adenosine (Clark et al., 1974), and prostaglandins (Ortmann and Perkins, 1977), with a rise in cyclic AMP. The cells also express muscarinic cholinergic receptors that mediate inhibitory effects of acetylcholine on cyclic AMP accumulation (Gross and Clark, 1977). The response of 1321NI cells to addition of a catecholamine to the growth medium is shown in Fig. 2. The response is clearly biphasic; the initial rapid rise in cyclic AMP is followed by a decline to near precatecholamine levels. This response is typical of most, if not all, cells that respond to catecholamines. If the cells are washed free of catecholamine, the cellular content of cyclic AMP allowed to fall, and then catecholamine added again, the response observed declines in a manner that is related to the concentration of catecholamine and the time of the initial exposure (Fig. 2). Since 1321N1 cells respond to both catecholamines and prostaglandins, we were able to determine if agonist-induced desensitization was selective' for the inducing hormone or if cross-desensitization occurred (Su er a / . , 1976a). One experimental approach to this question is illustrated in Fig. 3. Cells were first exposed to either isoproterenol or prostaglandin E, for increasing periods of time; the response to the same agonist or the alternate agonist was then tested in a second incubation. Numerous such experiments in our laboratory and others have clearly indicated that the protocol involving a homologous pre- and post-

LIGAND-INDUCEDCHANGES IN f3-RECEPTOR FUNCTION

0

m

20

Y)

40

89

50

60

'ID

Minuter FIG. 2 . The effect of time and agonist concentration on desensitization. Cells were incubated with 100 )uM norepinephrine ( 0 )or 10 phf norepinephrine (m) and 3H-labeled cyclic AMP (CAMP) accumulation was measured. At the times indicated, norepineprhine was removed by washing. The decline of 'H-labeled cyclic AMP content is shown by the dashed line. The cells were subsequently challenged with 100 pA4 norepinephrine for 5 minutes. (From Su ef a/..1976a.)

incubation leads to a more rapid and greater extent of loss of response than does the heterologous protocol. The specificity of the desensitization process of 1321Nl cells is high at short times of incubation, but eventually a 40-60% heterologous desensitization is observed. In this protocol the agonist concentrations used were high enough to fully saturate the receptors and fully activate cyclic AMP production. If instead cells were incubated with low concentrations (1 -5 nM) of catecholamine (Perkins et al., 1979) [or prostaglandin (Leightling et a/.. 1976)j for extended periods of time (12-24 hours), a highly agonist-specific desensitization was observed. The results described above suggested to us that desensitization could occur by more than one mechanism and that the processes might exhibit different time courses and different concentration-effect relationships. Since the concentration-effect relations for catecholamine-induced cyclic AMP formation and for catecholamine-induced desensitization (at 60 minutes) were similar, it seemed reasonable that cyclic AMP might mediate desensitization. Incubation of 1321N1 cells with dibutyryl or 8-methyl thio analogs of cyclic AMP caused a loss of response to both isoproterenol and prostaglandin E, (Su er al., 1976a).

90

JOHN P. PERKINS

2.5

-

5

A 4

2.0

a:

-

4

3

x

QS

0

-

2

-

1

..........................................

ti 0

~~

I

I

I

2

3

0

1

2

3

Hours FIG.3. The time course of loss of homologous and heterologous responsiveness. Cells were first exposed to 10 ph4 isoproterenol (ISO) or 10 ph4 PGE, for varying periods of time as shown on the abscissa. The agonists were then removed by washing and the cells incubated in the absence of agonist for 10 minutes to allow 3H-labeled cyclic AMP content to decline. ( A ) The cells were then challenged with 10 j&f isoproterenol for 5 minutes and the 'H-labeled cyclic AMP content was measured. The homologous response (IS0 then ISO) is represented by the solid circles ( 0 )and the (B) The cells were challenged with heterologous response (PGE, then ISO) by the open circles (0). 10 PGE, for 5 minutes and the 3H-labeled cyclic AMP content was measured. The homologous and the heterologous responsiveness (PGE, then PGE,) is represented by the solid squares Dashed lines represent the basal level of responsiveness (IS0 then PGE,) by the open squares (0). 3H-labeled cyclic AMP in the cells. (From Su et a / . . 1976a.)

(m),

The time courses were similar as were the extents of loss (40-50%), and in this regard the effects of cyclic AMP resembled heterologous desensitization as illustrated in Fig. 3. If cyclic AMP acts as a feedback inhibitor it is reasonable to expect that its action would be nonspecific with regard to the inducing agonist. Our early results and the more extensive studies of Brooker and co-workers (Terasaki et al., 1978; Nickols and Brooker, 1979) are consistent with the conclusion that cyclic AMP mediates a hormone-induced heterologous desensitization by effecting an inhibition of the adenylate cyclase reaction at a site distal to the hormone receptor interaction. A more complete discussion of the evidence in support of this contention can be found in Terasaki et al. (1978) and Perkins et al. (1982).

91


B. Analysis of Rates of Synthesis and Degradation of Cyclic AMP in Whole Cells If a rise in cellular cyclic AMP caused a rise in cyclic AMP-phosphodiesterase activity, a loss in the capacity of hormones to raise cyclic AMP levels also would result. Such an induction of phosphodiesterase activity has been shown to occur in a variety of cells (D’Armiento et al., 1972; Manganiello and Vaughan, 1972). In order to determine the relative roles of changes in synthesis and degradation of cyclic AMP as the basis of desensitization, we developed techniques for such assessments in intact cells. A pulse-labeling technique was used to determine the relative rates of synthesis of cyclic AMP during a series of 3-minute periods throughout a 2-hour exposure of human astrocytoma cells to catecholamines (Su et al., 1976b). The results provided evidence that, within 3 minutes, catecholamines induce a reduction in the rate of cyclic AMP synthesis that further declined over the 2-hour period. In related experiments the initial velocity of accumulation of cyclic AMP (5- to 10-second intervals for 40 seconds) was measured as an indication of changes in the rate of synthesis of cyclic AMP during desensitization. The results of both types of experiments indicated that a rapid loss in the capacity of the cells to synthesize cyclic AMP occurs within several minutes of exposure to catecholamines. Clark and Butcher (1979) have used a related approach to detect, continuously, changes in the rate of cyclic AMP synthesis during exposure of W1-38 fibroblasts to catecholamines. From the results of such studies it has been concluded that a change in the rate of synthesis of cyclic AMP is the primary basis for the loss of response of intact human astrocytoma and WI-38 cells to catecholamines. In order to measure the rate of degradation of cyclic AMP, it was first necessary to raise the intracellular content by exposure of the cells to a catecholamine. The synthesis of cyclic AMP was then reduced rapidly to basal levels by adding the antagonist propranolol. The subsequent decay of cyclic AMP levels appeared to obey first-order kinetics and the rate constant for degradation (K,) was determined as shown in Fig. 4. No significant change in K , was observed during a 60minute incubation with norepinephrine. During this time span the cellular cyclic AMP level underwent a typical biphasic change. No detectable change in the content of cyclic AMP in the medium occurred after addition of propranolol; thus, we could assume that the rate of loss of cyclic AMP was a reflection of the rate of phosphodiesterase activity in the intact cells. We carried out approximately 30 determinations of Kd at various times during the first 60 minutes of exposure of the cells to norepinephrine or isoproterenol. The mean + SD for 30 determinations was 0.32 0.05. Since there was no significant difference in the values of Kd determined after exposure of the cells to

*

92

JOHN P. PERKINS

i . . ’ ” i . ’ r i . *o

Mnute.

rope=-.n

-.2

-

aopc=-.J2

rbp=-.3a

rope= -.u

-.4

2

*P

-.6

-.E -1.0

0

1.0 1.5 Minutes

0.5

2.0

FIG.4. Determination of the rate constant for degradation of‘cyclic AMP. Cells were challenged with norepinephrine (NE) to elevate intracellular 3H-labeled cyclic AMP concentration (upper panel). At the times indicatcd propranolol was added to block completely the siniulatory effect of NE on 3H-labeled cyclic AMP formation. The content of 3H-labeled cyclic AMP in the cells was measured over the next 2 minutes to determine the rate of decline. The results are expressed as the natural logarithm of the fraction of ”-labeled cyclic AMP (A/Ao)remaining versus the time after addition of propranolol. The first-order rate constant of degradation is proportional to the negative slope. See Su rt a / . (1976b) for details of experimental procedure.

norepinephrine or isoproterenol, we concluded that rapidly induced, agonistspecific effects on phosphodiesterase activity did not occur. The idea that desensitization primarily involves a decrease in adenylate cyclase activity rather than an increase in phosphodiesterase activity is supported by these observations. Agonist-induced increases in phosphodiesterase activity have not been observed with less than 90 minutes of exposure to either PGE, or the catecholamines. Thus, the increase in degradative activity (1.5- to 2.0-fold after 2 hours) could play only a minor role late in the overall desensitization process. Based on studies with intact cells we were able to surmise that at least three different processes are induced by receptor agonists that lead to a loss of cellular responsiveness to hormones. The processes exhibit different time courses that allow their distinction. The most rapid process involves agonist-specific modifications and at high hormone concentrations is clearly evident by 3 minutes. At lower agonist concentrations this process is slower but does appear to proceed to

93

LIGAND-INDUCEDCHANGES IN f3-RECEPTOR FUNCTION

near-complete desensitization. A second process results in a nonspecific loss of hormonal responsiveness that can be detected after 30-60 minutes of exposure of 1321N1 cells to either catecholamines or prostaglandins; cyclic AMP analogs appear to induce a similar process over the same time course. Even under optimal conditions nonspecific or heterologous desensitization does not exceed 40-60% inhibition. The third agonist-induced process contributing to loss of response is the induction of phosphodiesterase activity. In 1321N1 cells this occurs after 90- I20 minutes of exposure to an agonist and results in a 50% reduction in the steady-state level of cyclic AMP. It is probable that the three processes are additive in their effects since they appear to occur by distinct mechanisms. Our subsequent studies have focused on an investigation of the mechanism responsible for agonist-specific desensitization.

IV. CATECHOLAMINE-INDUCED CHANGES IN ADENYLATE CYCLASE AND IN PAR BINDING PROPERTIES As an initial hypothesis it seemed reasonable to propose that agonist-specific desensitization would involve changes in the number or the functional state of the PAR. Thus, our first experiments (Su et al., 1980) compared the loss of whole cell response to isoproterenol, the loss of response of adenylate cyclase to isoproterenol, and changes in the number of PAR, during a 24-hour exposure of 1321Nl cells to isoproterenol. PAR were measured with a high-affinity, highly

16

0

24

HOURS

FIG. 5 , Time courses of decrease in PAR density, isoproterenol-stimulated adenylate cyclase activity. and isoproterenol-stimulated cyclic AMP accumulation during exposure to isoproterenol. Data are presented as the percentage of activities expressed in untreated cells assayed at the same time in culture life. (From Su er a/..1980.)

94

JOHN P.PERKINS

specific receptor antagonist, [ 1251]iodohydr~xybenzylpindolol (IHYP). The results of the comparison are shown in Fig. 5. The most striking observation was that loss of response to isoproterenol measured either in whole cells or membranes, was a much more rapid process than was loss of PAR. In fact, detectable loss of PAR did not occur before 1 hour. It also is apparent that the loss of responsiveness of adenylate cyclase to isoproterenol cannot fully account for the loss of response of intact cells. This latter discrepancy is to be expected since by 30-40 minutes of exposure to isoproterenol, heterologous desensitization should begin to contribute to the overall process of desensitization. Also, by 90-120 minutes the effects of a twofold rise in phosphodiesterase activity would contribute to the reduction in whole cell response. Other investigators have observed discrepancies between the magnitude of agonist-induced PAR loss and the extent of desensitization of hormone-stimulated adenylate cyclase activity (Shear et al., 1976; Johnson et al.. 1978; Wessels et al,, 1978, 1979). The distinct lag in PAR loss, while quite evident in most of our experiments with 1321Nl cells, is less evident or not observed in other cell types. Nonetheless, careful kinetic analyses uniformly have exposed a significant discrepancy between the extent of loss of hormone-stimulated enzyme activity and reduction in PAR number. A more detailed kinetic analysis (Su ef al., 1980) led to the conclusion that incubation of 1321NI cells with isoproterenol results in a rapid (t1,2 = 3 minutes) decrease in isoproterenol-stimulated adenylate cyclase activity. This reaction appeared to reach a steady state in which about 50% of control responsiveness remained. Upon removal of isoproterenol full responsiveness was regained with a I,,, for reversal of about 7 minutes. During the first 30-45 minutes of exposure of cells to isoproterenol the only detectable change in the functional status of the adenylate cyclase system was the partial loss of response to catecholamines. Basal enzyme activity and NaF-, Gpp(NH)p (guany 1-5’-yl imidodiphosphate)-, and PGE, -stimulated activities remained unchanged. We have used the term “uncoupled” to describe the state of the desensitized system during this stage of the process, i.e., PAR have not been lost nor have the components of adenylate cyclase been altered; nonetheless, receptor agonists do not stimulate enzyme activity. Our current understanding of the adenylate cyclase system suggests that a change in one of only two of the components of the enyzme system could account for the selective loss of hormone responsiveness; namely, alterations in the receptor per se or in the guanine nucleotide binding component (N) (see Fig. 1). The work of Gilman and co-workers (Ross et al., 1978; Sternweis and Gilman, 1978) suggests that a single class of N-proteins mediates activation of adenylate cyclase by both catecholamines and prostaglandins. Since the loss of responsiveness induced by isoproterenol is agonist specific, it seems less tenable

95

LIGAND-INDUCED CHANGES IN p-RECEPTOR FUNCTION

that changes in the N-protein could account for the observed change. However, the possibility remains that the N-protein could contain independent domains of interaction for each type of receptor expressed in a particular cell. A number of observations led us to propose (Su et al., 1980) that reaction schemes for agonist-induced activation of adenylate cyclase and agonist-induced desensitization share a common intermediate. For example, the uncoupling reaction and activation of cyclic AMP production in whole cells both exhibit a for isoproterenol of 0.03 pM. Also, partial agonists have about equivalent partial effects on the degree of uncoupling and the degree of activation of cyclic AMP synthesis. In addition, agonist-induced loss of PAR does not occur in the cycmutant of the S49 lymphoma cell; such cells exhibit normal PAR and adenylate cyclase but lack functional N-protein; thus, cyccells are “uncoupled. However, recent experiments by Green and Clark (1 98 l ) cast some doubt on our proposal. They exposed cyc - S49 cells to isoproterenol then supplemented membranes from such cells with N-protein extracted with detergent from wildtype S49 cells. Such extracts were sufficient to “recouple” membranes from naive cyccells but did not cause complete recoupling of the response to isoproterenol in cyc membranes from cells previously exposed to isoproterenol. They concluded that isoproterenol-induced desensitization does not require activation coupling of PAR and the N-protein. Experiments by Iyengar et al. (1981) had previously shown that the N-protein extracted from desensitized S49 wild-type cells was capable of “recoupling” cyc- membranes to the same extent as did N-protein from naive wild-type cells. Thus, while there is general agreement that it is the PAR that is modified during desensitization, it is not clear that this reaction requires normal coupling of PAR and N-protein. ”

-

V.

SEPARATION OF NATIVE AND DESENSITIZED PAR

In 1980 we reported that PAR from homogenates of 1321N1 cells exposed to isoproterenol for 15-30 minutes could be separated into two classes based on their respective migration patterns during centrifugation through sucrose density gradients (Harden et al., 1980). The basic observation is illustrated in Fig. 6. Isoproterenol-stimulated adenylate cyclase activity is recovered as a single peak migrating as a dense particle (45-50% sucrose) from both control and desensitized cells. PAR from control cells also are found predominantly in these fractions. However, the PAR from desensitized cells distribute in about equal proportions in the light peak and the heavy peak. The shift in PAR to the light peak appears to be a highly selective process. We have measured the various parameters of adenylate cyclase activity, muscarinic receptors, EGF receptors, Na-KATPase, and total protein in the light fraction before and after desensitization

96

JOHN P. PERKINS

C

.-0 u

e . .L

C

E a \

z

200-

Q, W -

0

a E

FRACTION NUMBER

Fir;. 6 . Sucrose density gradient distribution of P-adrenergic receptors and adenylate cyclase activity after short-term incubation of cells with isoproterenol. Cells were incubated with 1 mM sodium ascorbate ( 0 )or I mM sodium ascorbate plus I isoproterenol (0) for 15 minutes. The cells were treated with concanavalin A, then lysed and centrifuged in a sucrose density gradient. (A) '2s1HYP was used to determine P-receptor density in gradient fractions. ( B ) Isoproterenol-~timulated adenylate cyclase activity was determined. (From Harden ef a / . . 1980.)

with isoproterenol. In no case did these activities shift in their migration pattern. This analysis is not exhaustive, but it is sufficient to indicate a high degree of specificity for the PAR shift reaction. The time course of appearance of receptors in the light vesicle fraction (PAR,,) is similar to the time course of the uncoupling reaction with a t,,* of about 3 minutes; in addition, the rates of reversal of both processes are similar. Also, both processes appear to proceed in a fashion similar to a steady state in which 50% of the receptors remain in the native state (PAR,) and 50% are altered (PAR, or PAR,,). Such correlations suggest a functional role for PAR,, in the reactions leading to loss of responsiveness to isoproterenol.

LIGAND-INDUCED CHANGES IN PRECEPTOR FUNCTION

97

The ligand binding properties of PAR,, have been studied in some detail (Harden et al., 1980). In terms of antagonist binding characteristics no change in Kd values has been detected. Conversely, agonist binding affinity is reduced about 10-fold compared to agonist affinity for PAR, or for receptors in the heavy peak. A similar reduction in the affinity of agonist binding was observed in homogenates of desensitized cells (Harden et al., 1979b). It appears that such homogenates contain receptors with native affinity for agonists as well as modified receptors with lower affinity for agonist. Centrifugation of the homogenate over sucrose gradients separates these two affinity states on the basis of physical differences in the membranes with which they are associated. Electron micrographs of the light fractions of the gradient reveal primarily small vesicles, whereas the heavy peak contains primarily large open sheets of plasma membrane. The plasma membrane fragments behave as open sheets due to a treatment of the cells with concanavalin A prior to lysis. Such treatment appears to stabilize the plasma membrane and reduce fragmentation and vesiculation (see Lutton et al., 1979). Enzymatic markers for the plasma membrane and for the Golgi apparatus migrate exclusively in the heavy and light peaks, respectively (Lutton et al., in preparation). Recent experiments in our laboratory (Hertel and Wakshull, unpublished results) indicate that 1321N1 cell surface EGF receptors migrate exclusively in the heavy peak whereas EGF-induced, internalized receptors migrate primarily in the light peak, corresponding to the migration of PAR,,.

VI.

RECEPTOR ENDOCYTOSIS AS A MECHANISM FOR AGONIST-INDUCED DESENSITIZATION

The results cited above and the analogy to polypeptide hormone-induced receptor endocytosis (Pastan and Willingham, 1981) as a first step in receptor down-regulation support the idea that catecholamine-induced endocytosis of PAR, to yield PAR,, occurs during the first few minutes of agonist exposure to 1321N1 cells. Recent experiments in our laboratory (Toews et af.,in preparation) have examined the binding characteristics of [ '2sl]iodopindolol ( 12sI-Pin)at 4°C to intact cells before and during the desensitization process, and to fractions from sucrose density gradients. Briefly stated, the results indicate that at 4°C '2sl-Pin does nor bind to PAR,, but will bind to the PAR, remaining in the heavy peak. Consistent with this finding is the observation that the number of i2sI-Pin binding sites at 4°C on intact cells is reduced to about 50% of control cells by prior incubation of 1321N1 cells at 37°C with isoproterenol. The loss of binding sites occurs with kinetics of onset and reversal that are consistent with the kinetics of formation and reversal of PAR, and PAR,,. We have tentatively concluded that at 4°C 12sI-Pinbinds only to cell surface receptors, i.e., it does

98

JOHN P. PERKINS

not cross membrane barriers effectively. Such a diffusion limitation would offer a plausible explanation for the results if PAR are lost via agonist-induced endocytosis. Other workers in this field have presented evidence in favor of agonist-induced internalization of PAR. For example Chuang and Costa (1979) reported that incubation of bullfrog erythrocytes with isoproterenol results in the appearance of a soluble fraction of PAR in cell lysates. Staehelin and Simons (1982) have carried out a series of studies of catecholamine-induced desensitization in C6 glioma cells. They utilized a nonhydrophobic radioligand for the PAR, 'Hlabeled CGP-12177, to study cell surface binding much as described above for Iz5I-Pin. Their results from studies carried out in Basel with C6 cells and similar studies carried out in collaboration with us by Dr. C. Hertel working in Chapel Hill using 1321Nl cells all are consistent with an endocytosis model for PAR desensitization. Unfortunately, none of the results reported to date are definitive for such a mechanism.

VII.

A KINETIC MODEL FOR AGONIST-INDUCED DESENSITIZATION

The kinetics of onset and reversal of the early stage of agonist-induced desensitization have been studied in some detail. In summary four different aspects of the process can be measured: (1) the loss of adenylate cyclase activity, (2) the change in agonist binding affinity, (3) the formation of PAR,,, and (4) the loss of whole cell binding to 1251-Pin(or 3H-labeled CGP- 12177) when measured at 4°C. These results all are consistent with the following simple relationship (ISO, isoproterenol): t IS0

BARN C PAR (modified) -IS0

where the forward reaction has a t,,2 at receptor saturation of 2-3 minutes and the recovery reaction has a t,,2 of 6-8 minutes. Recently we have shown that prior exposure of 132I N 1 cells to concanavalin A (Con A) completely prevents the formation of PAR,, but has no effect on the loss of responsiveness of adenylate cyclase to isoproterenol (Waldo, Perkins, and Harden, in preparation). Recovery of PAR, to PAR, occurs upon removal of isoproterenol in the continued presence of Con A. This result suggests the following modification of the kinetics model: 11) t IS0

I21 t

IS0

PARN G PARu S P A R L ~ -IS0

-IS0

99

LIGAND-INDUCED CHANGES IN B-RECEPTOR FUNCTION

0

I

I

I

I

I

I

1

2

3

4

5

6

Time ( h r )

FIG. 7. Recovery of isoproterenol-stimulated adenylate cyclase activity following desensitization. Cells were incubated with 1 p M isoproterenol. At the times indicated the culture dishes were washed with fresh medium. Membrane fractions were then prepared or were prepared after a 20minute further incubation in the absence of isoproterenol. The open circles indicate the loss (00) then recovery (0- -0) of adenylate cyclase activity. The solid triangles indicate the loss then recovery of IHYP binding sites. (From Su el d..1980.)

-A)

(A-

(A--A)

and allows the speculation that a prior modification of PAR leads not only to an inability to couple with the nucleotide binding protein, but initiates conversion of the receptor to PAR,,, possibly by way of endocytosis. The nature of the initial receptor modification is currently under investigation. Reactions ( 1 ) and ( 2 ) have not been distinguished by kinetic analysis and thus occur with similar rates. At early times of desensitization these reactions appear to be completely reversible; however, when loss of PAR binding sites for antagonist radioligands occurs, complete recovery of responsivenss of adenylate cyclase to isoproterenol does not occur (Su et al., 1980). Nonetheless, rapid recovery of enzyme responsiveness does occur (t,,z = 7-10 minutes) to the extent that receptor binding sites remain. This relationship is illustrated in Fig. 7. Such observations are consistent with the following overall kinetic model: (I)

I?)

t IS0

PAR,

t

BARci -IS0

(31

IS0

1

IS0

PAR,, Q PARL

~-Iso

- IS0

where reaction (3) is essentially irreversible within the time span of experimentation. The study of the reversal of reaction (3), to be described below, demonstrates that the reaction does actually reverse with a t,,z of about 12-14 hours. The model does not explicitly explain the lag in the formation of PAR, since the amount of PAR,,, the apparent precursor of PAR,, rises rapidly to a steady-

100

JOHN P. PERKINS

state value long before the rate of PAR, formation is at its maximum. Thus, the formation of PAR, is not a simple first-order reaction with respect to the concentration of PAR,, and must involve at least one intermediate reaction. Alternatively, we must leave open the possibility that PAR, is formed by a set of reactions that does not require prior formation of PAR, or PAR,,, i.e., PAR, may be formed by an independent pathway.

VIII.

DIFFERENTIAL EXPRESSION OF PAR DURING GROWTH OF 1321N1 CELLS

When 1321N1 cells are passaged in culture in the traditional manner, the specific activity of isoproterenol-stimulated adenylate cyclase and the number of PAR per cell vary significantly and in parallel as the dilute cultures ( 5 X 10’ cells/cm2) grow to confluence (3 X lo5 cells/cm2). Basal enzyme activity and NaF- and PGE, -stimulated activities remain relatively constant irrespective of culture density. A series of experiments from our laboratory established that the number of PAR per cell is regulated in a reproducible manner with respect to cell density (Harden et al., 1979a). Cells taken from postconfluent cultures and seeded at low density in fresh culture dishes usually exhibit 2000-3000 PAR/ cell. After a 20- to 24-hour lag the cells begin to accumulate receptors at the rate of 12- 13 X lo3 PAR/cell/cell division until a steady-state receptor density of 10 X 10’ PAR/cell is reached, whereupon the rate declines to 10 X lo3 PAR/celI/ cell division. Upon reaching confluence, PAR synthesis in the culture appears to cease, and, because the cells continue to grow, the number of PAR declines to 2-3 X 103/cell. It is at this point that the culture is usually passaged. Thus, the cells appear to “sense” that the number of PAR is less than lo4 PAR/cell and, if growing at less than confluent density, will begin to make PAR at an accelerated rate. If the receptor depleted cells are plated at confluent or greater density the cells will not activate receptor synthesis. The cells also appear to ‘‘sense” when they attain lo4 PAWcell since the synthesis rate declines to maintain the preconfluence steady-state rate of lo4 PAR/cell/division. A third signal appears to be initiated by cell contact (not by conditioned medium) and terminates receptors synthesis. The physiological significance of the growth-related regulation of PAR is not apparent, but cognizance of the phenomenon allowed us to examine the process of PAR loss and recovery more effectively.

IX.

DOWN-REGULATION OF PAR AND THE RECOVERY OF LOST RECEPTORS

Exposure of 1321N 1 cells to catecholamines for 12-24 hours leads to greater than 90% loss of PAR. If the catecholamine is removed and the cells washed

Ot

a

B

T

1

1

v

A

Control

Control

IS0 W0sho"I +TUN

I

I

I

I

1

I

144

168

192

216

240 264

StlnOH

'

, /t

021

I20

Cbl

E

168

I

144

891

SmOH

HOURS

120

261

96

1

72

912

48

ObZ

24

32

f/

0

1

d /

FIG.8. Recovery of down-regulated PAR in pre- and postconfluent cultures. Cells were exposed to isoproterenol (ISO) for 12 hours. The cultures were then washed and the incubation continued in fresh growth medium in the absence or presence of cycloheximide (CHX) or tunicamycin (TUN). At the time indicated the density of PAR was determined by I25IHYP binding. (A) Preconfluent cultures. (B) Postconfluent cultures.

0

102

JOHN P. PERKINS

with fresh medium, PAR reappear as does catecholamine responsiveness of adenylate cyclase. Depending on the cell density of the cultures two patterns of recovery have been observed (Doss et al., 1981). These patterns are illustrated in Fig. 8 . When cells were down-regulated and then allowed to recover under preconfluent growth conditions, recovery could not be differentiated from new receptor synthesis. As shown in Fig. 8A, the cultures regained PAR at a rate and to an extent identical to control cells. However, if recovery was allowed to occur in the presence of cycloheximide, PAR returned only after a lag and at a reduced rate; nonetheless, PAR eventually recovered to the level present in the cultures at the time of exposure to isoproterenol. It should be noted that cycloheximide added to control cultures stopped cell growth immediately and completely inhibited the synthesis of PAR (Doss ef al., 198I). The effects of cycloheximide on both cell proliferation and PAR synthesis were completely reversible even after 48 hours of exposure to the protein synthesis inhibitor. Our initial conclusion from such studies (Fig. 8A) was that although PAR, were not detectable by standard binding assays, the primary amino acid structure of the receptor was not destroyed during down-regulation and apparently the full complement of PAR, could be recovered in the absence of protein synthesis. We also have utilized tunicamycin, which inhibits dolichol phosphate-mediated protein glycosylation, in a similar set of experimental protocols (Doss et al., 1982). Like cycloheximide, tunicamycin added to control cultures completely inhibits the appearance of PAR; however, it does not completely block cell proliferation. In fact, cells continue to grow at about half the normal rate. Recovery of down-regulated PAR in the presence of tunicamycin occurs more rapidly than in the presence of cycloheximide but to about the same extent, i.e., to the level present in the cultures at the time of first exposure to isoproterenol. The most obvious interpretation of the recovery experiments carried out with preconfluent cultures is that essentially all of the lost receptors are recovered even in the presence of agents that completely block PAR synthesis. Thus, PAR, does not represent a receptor modified in its primary amino acid sequence and probably it is not modified in terms of its core glycosylation. An unexpected result was obtained when down-regulated receptors were allowed to recover in postconfluent cultures; namely, recovery was completely blocked by cycloheximide (Fig. 8B). This result suggested that new receptors were synthesized during recovery in postconfluent cells. However, in contrast to cycloheximide, tunicamycin had little or no effect on either the rate or extent of receptor recovery. We attempted to resolve the dichotomy of the contrasting effects of tunicamycin and cycloheximide by a method capable of a direct identification of newly synthesized proteins. Postconfluent cultures were treated with isoproterenol for 12 hours, then allowed to recover in the presence of heavy isotope 2H,15N,13C-containingamino acids. Any newly synthesized PAR would necessarily be of increased mass and could be identified by their behavior during

LIGAND-INDUCED CHANGES IN O-RECEPTOR FUNCTION

103

centrifugation in sucrose-D,O gradients (see Fambrough, 1979, and Reed and Lane, 1980, for similar application of this technique to other receptor systems). The receptors that recovered under such conditions, however, migrated on the gradients like native, 1H-,14N-,12C-containingPAR. Taken together the results suggest that whereas the recovery of PAR in postconfluent cells does not require synthesis of the receptor per se, other proteins involved in the recovery process must be synthesized during the recovery period for recovery of PAR to occur. If the turnover of such proteins was greater in post- than in preconfluent cultures it would provide an explanation for the differential effects of cycloheximide (compare Fig. 8A and B). It is of some interest that 1321N1 cells seem not to turn over PAR at any appreciable rate. Inhibition of new PAR synthesis with tunicamycin or cycloheximide does not result in the degradation of previously synthesized receptors. Even when the number of PAR per cell is declining as in postconfluent cultures, the total number of PAR per culture remains constant. The cells even preserve PAR upon down-regulation; as noted above, PAR recover to the level present in the cell at the time of initiation of desensitization. It is not clear what purpose is served by the stringent maintenance of cellular PAR once they are synthesized. In contrast, insulin receptors and nicotinic cholinergic receptors turnover with a /, 7 -9 (Reed and Lane, 1980) and 17-21 hours (Fambrough, 1979), 1 cspectively . The results presented above also suggest that PAR are not continuously synthesized during long-term exposure to isoproterenol; at least the number of PAR, that can be recovered is related directly to the number present upon exposure to isoproterenol (see Fig. 8). This relationship was examined (Doss, unpublished observations) quantitatively by exposing cells in preconfluent culture to isoproterenol for increasing periods of times (12, 24, 36, and 48 hours). In each case the number of PAR that reappeared upon removal of isoproterenol was the same. If PAR synthesis had continued at the same rate as in control cells, and if these receptors had been shuttled into the form PAR,, approximately three times as many receptors should have been recovered after 24 hours of exposure than after 12. Thus the desensitization process appears to inhibit new PAR synthesis as well as initiate functional loss of PAR.

X.

ISOPROTERENOL-INDUCEDCHANGES IN AGONIST BINDING PROPERTIES OF INTACT 1321N1 CELLS

The formation during the early stage of desensitization of a PAR variant (PAR,,) with lowered affinity for agonists is a change that theoretically should be demonstrable in whole cells. The problem with such an experiment is that the typical binding reaction requires 30-60 minutes to come to equilibrium at 37°C

104

JOHN P. PERKINS

and the desensitization reaction has a t,,, for induction of about 2-3 minutes. Thus we were required to design binding protocols that would allow determination of changes in the affinity of agonists for the PAR of intact cells during the first few minutes of exposure to an agonist. An accurate equilibrium dissociation constant for a competing agonist can be obtained in nonequilibrium binding assays if initial velocity conditions are met for the radioligand binding reaction, and if the competing agonist is at equilibrium with the receptor throughout the time of radioligand binding. Using '251-Pin we have been able to carry out experiments to determine the initial binding parameters for the interaction of isoproterenol and epinephrine with PAR on intact cells (Toews et a l . , 1982). The binding competition is conducted over short times (10-60 seconds) at low 1251Pin concentrations. Using this protocol it can be demonstrated that isoproterenol binds to naive I32 IN 1 cells with a Kd of 0.1 pM; however, after exposure of the cells to isoproterenol the Kd of binding of isoproterenol shifts from 0.1 to 20-40 IJ.n with a tIlz for the agonist-induced change of 1-2 minutes. Upon removal of isoproterenol the Kd returns to about 0.1 pM with a c , , ~of 6-8 minutes. The shift is not induced by antagonists and the Kd values for antagonists are not changed by exposure of the cells to isoproterenol. Thus the rates of change in PAR function during the onset and reversal of desensitization are similar when assessed using intact cells or membrane fractions. The proportion of PAR in lowand high-affinity states cannot be accurately assessed from the data of such binding experiments; however, it would appear that a 20-minute exposure of cells to receptor-saturating concentrations (>1 pM) of isoproterenol results in the conversion of about 70% of PAR to the low-affinity state.

XI.

CONCLUSIONS

Exposure of intact 1321N 1 cells to isoproterenol not only causes the activation of adenylate cyclase and the accumulation of cyclic AMP but sets in motion a complex series of events that results in the down-regulation of responsiveness if exposure to the catecholamine is extended in time. Three general categories of down-regulating responses have been identified: ( I ) a rapid uncoupling of the PAR-adenylate cyclase system with subsequent loss of PAR; (2) a slower, nonspecific desensitization of adenylate cyclase to the effects of all classes of receptor agonists by a process that is mediated by cyclic AMP; and (3) a slow induction of phosphodiesterase activity mediated by cyclic AMP. The overall process of agonist-induced desensitization to the further effects of agonists is probably the summation over time of these three processes. The receptor-specific process of desensitization is currently best described

105

LIGAND-INDUCEDCHANGES IN p-RECEPTOR FUNCTION

kinetically by the following set of reactions: PARN

(1)

(21

(3)

tlSO

+IS0

+ IS0

-IS0

-IS0

-IS0

* PAR” * PARLv T PARL

Reaction ( 1 ) results in a reversible modification of native receptors that prevents their interaction with the nucleotide binding protein of the adenylate cyclase system; this results in an “uncoupling” of the effects of receptor binding from activation of adenylate cyclase. Reaction (2) involves conversion of the receptor to a form that behaves as if the PAR or its environment had been physically altered. We have speculated that PAR,, are formed by a process of endocytosis (see Fig. 9). The reactions to this point in the sequence are rapidly and completely reversible. Reaction (3) represents a slowly reversible further change in the properties of the receptor. PAR, no longer bind to antagonist ligands; however, it probably is not a degraded protein since full recovery of PAR, to PAR, occurs even in the presence of cycloheximide or tunicamycin, each of which blocks new synthesis of PAR,. It is possible that the receptor is recycled through the Golgi apparatus and, after refurbishing, returned to the plasma membrane along the normal pathway of receptor insertion. This idea and the details of the sequence of events depicted in Fig. 9 are speculative, but clearly our results to date are compatible with such a model, which is similar to the models proposed for the downregulation of polypeptide hormone receptors (Pastan and Willingham, 198 1) and the internalization and recycling of LDL-receptor proteins (Goldstein et al., 1979).

FIG.9. Agonist-induced receptor internalizationas a model for catecholamine-induced desensitization. The model depicts the binding, uncoupling, endocytosis, and receptor loss steps of catecholamine-induced down-regulation in graphic terms commonly used to illustrate down-regulation of polypeptide hormone cell surface receptors.

106

JOHN P. PERKINS

REFERENCES Aarons, R. D.. Nies. A. S.. Gal, J.. Hegstrand, L. R.. and Molinoff, P. B. (1980). Elcvation of P-adrenergic receptor density in human lymphocytes after propranolol administration. J. Clin. Invest. 65, 949-957. Aurbach, G . D., Fedak, S. A., Woodard, C. J . , Palmer, I . S., Hauser, D., and Troxler, T. (1974). The beta-adrenergic receptor: Stereospecific interaction of iodinated beta-blocking agent with a high affinity site. Science 186, 1223-1224. Barovsky. K., and Brooker, G. (1980). 12slodopindolol. a ncw highly selective radioiodinated P-adrenergic receptor antagonist: Measurement of P-receptors on intact rat astrocytoma cells. J . Cyclic Nuclentide Res. 6 , 297-307. Chuang. D. M.. and Costa, E. (1979). Evidence for internalization of the recognition site of P-adrenergic receptors during receptor sub-sensitivity induced by (-)-isoproterenol. Proc. Nail. Acad. Sci. U.S.A. 76. 3024-3028. Clark, R. B., and Butcher, R. W . (1979). Desensitization of adenylate cyclase in cultured fibroblasts with prostaglandin El and epinephrine. J. B i d . Chem. 254, 9373-9378. Clark, R. B., and Perkins, J . P. (1971). Regulation of adenosine 3’:s’-cyclic monophosphate concentrations in cultured human astrocytoma cells by catecholamines and histamine. Proc. Nail. Acad. Sci. U.S.A. 68, 2757-2760. Clark, R . B., Gross, R., Su, Y. F., and Perkins, J. P. (1974). Regulation of adenosine 3‘5’monophosphate content in human astrocytoma cells by adenosine and the adenine nucleotides. J. Bid. Chem. 249, 5296-5303. D‘Armiento, M., Johnson, G. S., and Pastan. I. (1972). Regulation of adenosine 3’:5’-cyclic monophosphate phosphodiesterase activity in fibroblasts by intracellular concentrations of cyclic adenosine monophosphate. P roc. Nail. Arad. Sci. U.S.A. 69, 459-62. Doss. R. C., Perkins. J. P.. and Harden, T. K. ( 1981). Recovery of P-adrenergic receptors following long temi exposure of astrocytoma cells to catecholamine. J. B i d . Chem. 256, 12281- 12286. Doss. R. C., Harden, T. K.. and Perkins, J. P. (1982). Role of protein glycosylation in the synthetic processing of P-adrenergic receptors (PAR). Fed. Proc. Fed. Am. Soc. Exp. Eiol. 41, 7392 abs. a new Engel, G., Hoyer, D., Berthold, R., and Wagner, H. (1981). ( 2 ) I~~lodocyanopindolol, ligand for P-adrenoceptors: Identification and quantitation of subclasses of P-adrenoceptors in guinea pig. Nauyn-Schmied. Arch. Pharmacoi. 317, 277-285. Fambrough, D. M. (1979). Control of acetylcholine receptors in skeletal muscle. Phvsiol. Rev. 59, 165-227. Fraser, J . , Nadeau, J., Robertson, D., and Wood, A. J. J. (1981). Regulation of human leukocyte beta receptors by endogenous catecholamines. J. Clin. Invest. 67, 1777-1 784. Goldstein, J. L., Anderson, G. W., and Brown, M. S. (1979). Coated pits, coated vesicles. and receptor-mediated endocytosis. Nature (London) 279, 679-685. Green, D. A,, and Clark, R. B. (1981). Adenylate cyclase coupling proteins are not essential for agonist-specific desensitization of lymphoma cells. J. B i d , Chem. 256, 2105-2108. Gross, R. A., and Clark, R. B. (1977). Regulation of adenosine 3’5’-monophosphate content in human astrocytoma cells by isoproterenol and carbachol. Mof. Pharmacol. 13, 242-250. Harden, T. K., Foster, S. J., and Perkins, J. P. (1979a). Differential expression of components of the adenylate cyclase system during growth of astrocytoma cells in culture. J. Eiol. Chem. 254, 4416-4422. Harden, T. K., Su Y-F, and Perkins, J. P. (l979b). Catecholamine-induced desensitization involves an uncoupling of baa-adrenergic receptors and adenylate cyclase. J. Cvclic Nucleotide Res. 5 , 99- 106. Harden, T. K., Cotton, C. U.,Waldo, G. L., Lutton, J. K.. and Perkins, J. P. (1980). Cate-


107

cholamine-induced alteration in the sedimentation behavior of membrane-bound P-adrenergic receptors. Science 210, 44 1-443. Innis, R. B.. Cornea, F. M. A.. and Snyder, S. (1979). Carazolol, an extremely potent p-adrenergic blocker: Binding to P-receptors in brain membranes. Life Sci. 24, 2255-2264. Iyengar, R., Bhat, M. K . , Riser, M. E., and Bimbaumer. L. (1981). Receptor-specific desensitization of the S49 lymphoma cell adenylyl cyclase. J . B i d . Chem. 256, 4810-4815. Johnson, G. L. Wolfe, B. B., Harden, T. K.. Molinoff, P. B., and Perkins, J. P. ( 1978). Role of padrenergic receptors in catecholamine-induced desensitization of adenylate cyclase in human astrocytoma cells. J. B i d . Chem. 253, 1472- 1480. Lefkowitz. R . J., Wessels, M. R., and Stadel. J. M. (1980). Hormones, receptors, and cyclic AMP: Their role in target cell refractoriness. Current TO\>. Cell. Regul. 17, 205-230. Leichtling. B. J . , Drotar, A. M., Ortmann, R.. and Perkins, J. P. (1976). Growth of astrocytoma cells in the presence of prostaglandin E l : Effect on the regulation of cyclic AMP metabolism. J . Cvclic, Nideotide Res. 2, 89-98. Linibird. L. E. (1981). Activation and attenuation of adenylate cyclase: The role of GTP-binding proteins as macromolecular messengers in receptor-cyclase coupling. Biochcm. J . 195, I- 13. Lutton, J. K.. Frederich, R. C., and Perkins, J. P. (1979). Isolation of adenylate cyclase-enriched membranes from mammalian cells using concanavalin A. J. B i d . Chem. 254, I 1181-11184. Maclntyre, E. H.. Ponten, J.. and Vatter, A. E. (1972). The ultrastructure of human and murine astrocytes and of human fibroblasts in culture. Actu Pathol. Microbiol. Scand. 80, 267-283. Mangdniek. V . . and Vaughan. M. ( 1972). Prostaglandin El effects on adenosine 3':5'-cyclic monophosphate concentration and phosphodiesterase activity in fibroblasts. Proc. Nail. Acad. Sci. U . S . A . 69, 269-273. Minneman, K . P . . Pittman. R. N., and Molinoff, R. B. (1981). p-Adrenergic receptor subtypes: Properties. distribution and regulation. Annic. Rei.. Neurosci. 4, 419-462. Nickols, 0. A , . and Brooker, G. (1979). Induction of refractoriness to isoproterenol by prior treatment of C6-2B rat astrocytoma cells with cholera toxin. J . Cyclic Nurleotide Rrs. 5 , 435-447. Ortniann. R.. and Perkins. J. P. ( 1977). Stimulation of adenosine 3'3-monophosphate formation by prostaglandins in human astrocytoma cells. J . B i d . Chcm. 252, 6018-6025. Pastan, I . H., and Willingham, M. C. (1981). Receptor-mediated endocytosis of hormones in cultured cells. Annu. Rev. Phvsiol. 43, 239-250. Perkins, J. P.. Maclntyre, E. H., Riley. W. D., and Clark, R. B. (1971). Adenylate cyclase, phosphodiesterase. and cyclic AMP dependent protein kinase of malignant glial cells in culture. Life Sci. 90 (Part l ) , 1069-1080. Perkins, J . P.. Su Y.-F., and Harden, T . K. (1979). Adaptive changes in the responsiveness of hol 4, 279-294. adenylate cyclase to catecholamines. Drug A l ~ ~ ~Depend. Perkins. J . P., Harden, T. K . , and Harper, J . F. (19x2). Acute and chronic modulation of the responsiveness of receptor-associated adenylate cyclases. Hundh. Exp. Pharmacol. 58, 185-224. Ponten. J . . and Maclntyre, E. H. (1968). Long term culture of normal and neoplastic human glia. A m Pathol. Microbiol. 74, 465-486. Reed, B. C.. and Lane, M. D. (1980). Insulin receptor synthesis and turnover in differentiating 3T3LI preadipocytes. Proc. Nail. Acud. Sci. U.S.A. 77, 285-289. Ross. E. M.. and Gilman, A. G. (1980). Biochemical properties of hormone-sensitive adenylate cyclase. Annri. Rev. Biochem. 49, 533-564. Ross. E. M.. Howlett, A. C . , Ferguson. K. M., and Gilman, A. G . (1978). Reconstitution of hormone-benbitive adenylate cyclase activity with resolved components of the enzyme. J . B i d . Chcm. 253, 6401-6412.

108

JOHN P. PERKINS

Shear, M.. Insel, P. A., Melmon, K. L., and Coffino, P. (1976). Agonist-specific refractoriness induced by isoproterenol. J . B i d . Chem. 251, 7572-7576. Staehelin, M.. and Simons, P. (1982). Rapid and reversible disappearance of P-adrenergic cell surface receptors. EMBO J . I, 187-190. Sternweis. P. C.. and Gilman, A. G. ( 1979). Reconstitution of catecholamine-sensitive adenylate cyclase. J . B i d . Chem. 254, 3333-3340. Su. Y .-F., Cubeddu-Ximenez, L., Perkins, J . P. (1976a). Regulation of adenosine 3’:5’-monophosphate content of human asrrocytoma cells: Desensitization to catecholamines and prostaglandins. J . Cvclic. Nucleoride Res. 2, 257-270. Su. Y.-F., Johnson, G. L., Cubeddu-Ximenez. L.. Leichtling, B. H., Ortmann, R.. and Perkins, I. P. ( 1976b). Regulation of adenosine 3’5’-monophosphatecontent of human astrocytoma cells: Mechanism of agonist-specific desensitization. J . C.vclic Nucleotide Res. 2, 271-285. Su, Y.-F., Harden, T. K . , and Perkins, J . P. (1980). Catecholamine-specific desensitization of adenylate cyclase: Evidence for a multistep process. J . Biol. Chem. 255, 7410-4719. Terasaki. W. L.. Brooker, G . , de Vellis, J . , Inglish, D., Hsu. C.-Y.. and Moylan. R. D. (1978). Involvement of cyclic AMP and protein synthesis in catecholamine refractoriness. Adv. Cyclic Nuclcotide Res. 9 , 33-52. Toews, M. L., Harden, T. K . , and Perkins. J . P. (1982). Detection of high-affinity agonist binding to P-adrenergic receptors on intact cells. Fed. Proc. Fed. Am. Soc. Exp. Eiol. 41, 7393 (Abstr.). Wessels. M. R., Mullikin, D., and Lefkowitz, R . J. (1978). Differences between agonist and antagonist binding following beta-adrenergic receptor desensitization. J . Biol. Chrm. 253, 3371-3373. Wessels, M. R., Mullikin, D., and Lefkowitz, R. J . (1979). Selective alteration in high affinity agonist binding: A mechanism of beta-adrenergic receptor desensitization. Mol. Pharmucol. 16, 10-20. Williams, L. T., and Lefkowtiz, R. J . (1978). “Receptor Binding Studies in Adrenegic Pharmacology.” Raven, New York.

CURRENT TOPICS IN MEMBRANES A N D TRANSPORT, VOLUME 18

Hormone-Sensitive Adenylate Cyclase: Identity, Function, and Regulation of the Protein Components ELLIOTT M . ROSS," STEEN E . PEDERSEN,".' AND VINCENT A . FLORIO*,i *Department of Pharmacology University c$ Texas Health Science Center at Dallas Dallas. Texas and Depurtments of' tBiochemistr?, and $Pharmacology University of Virginia Charlottesville. Virginia

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Protein Components of Hormone-Sensitive Adenylate Cyclase A. Identities of the Proteins.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B. Resolution of the Catalytic and Regulatory Proteins of Adenyla C. Catalytic Component of Adenylate Cyclase: C . . . . . . . . . . . . . . D. The Stimulatory GTP-Binding Regulatory Protein: G/F. . . . . . . . . . . . . . . . . . . . . E. Cell Surface Receptors That Stimulate Adenylate Cyclase . . . . . . . . . . . . . . . . . . 111. Protein-Protein Interactions and the Regulat lase.. . . . . . . . . . . . A. General Considerations. . . . . . . . . . . . . ..... ..... B. Activation of C by GI F . . . . . . . . . . . . . . . . . . . . . . . . . . ................ C. Regulation of the Activation of GIF by Receptor and Hormone. . . . . . . . . . . . . . IV. Asscssment of Progress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

I.

109

11.

1.

121 I25 127 127 128 131 I37 137

OVERVIEW

It is now clear that the hormone-sensitive adenylate cyclase system in the plasma membrane of target cells consists of at least three distinct protein species. There may be as many as seven. This complexity is magnified when multiple receptors for different hormones, both inhibitory and stimulatory, exist. Our 109 Copyright 0 1983 by Academic Press. Inc All righis of reproduclion in any [om reserved ISBN 0-12-153318-2

110

ELLIOTT M. ROSS ET AL.

laboratory has decided to approach the study of this complex network by trying first to identify and separate these individual protein components, to study them in isolation using increasingly pure preparations, and then to study their interactions and regulation after suitable reconstitution. This is in marked contrast to the approach of studying adenylate cyclase in intact plasma membranes and inferring mechanistic information from the modulation of that activity by regulatory ligands. In this article we will discuss results both from our own laboratory and from others, but it will stress our basic analytical approach to a multienzyme regulatory complex. We will implicitly assume that all vertebrate, membranebound, hormone-sensitive adenylate cyclases are qualitatively similar in their regulation and composition. It follows that observed differences among different cells will merely reflect differences in the concentration of each protein or quantitative differences in their kinetic or equilibrium constants rather than qualitative differences in mechanism. As adenylate cyclase components prepared from different cells are shown to be capable of interacting with each other, this assumption is increasingly supported.

II. THE PROTEIN COMPONENTS OF HORMONE-SENSITIVE ADENYLATE CYCLASE A. Identities of the Proteins Much of the recent research activity in the field of adenylate cyclase has involved the enumeration and identification of the proteins that compose the system. From the early 1970s, most investigators have at least tacitly assumed that the protein that catalyzes the adenylate cyclase reaction on the inner face of the plasma membrane is distinct from the receptor that binds hormone on the cell surface. This assumption rested primarily on kinetic and developmental studies (reviewed by Perkins, 1973) and gained increasing support from chemical and genetic approaches to the resolution of receptor and enzyme (reviewed by Maguire et al., 1977; Ross and Gilman, 1980). In 1977, both Limbird and Lefkowitz (1977) and Haga et al. (1977a) chromatographically separated (3-adrenergic receptors from adenylate cyclase after detergent solubilization of plasma membranes. The receptors were measured by assaying the binding of appropriately specific radioactive ligands , and adenylate cyclase was assayed according to its enzymatic activity. This experiment has now been duplicated in several laboratories using receptors for various hormones. It is thus clear that the population of cell surface receptors that activate adenylate cyclase constitutes a large family of proteins that, on a simple level, interact with adenylate cyclase in a common fashion.

PROTEIN COMPONENTS OF ADENYLATE CYCLASE

111

Although hormonal stimulation of activity is lost upon the disruption of the plasma membranes, the adenylate cyclase activity that is solubilized by detergents usually retains its characteristic responsiveness to guanine nucleotides and to fluoride. Ross and Gilman (1977b; also Ross et af., 1978) and Pfeuffer (1977) showed that the activity that is stimulated by fluoride or by analogs of GTP reflects the interaction of two separate proteins (see below). The isolated catalytic protein, referred to as C, is insensitive to these compounds. Its activity is stimulated by these ligands only in the presence of a regulatory GTP-binding protein, referred to as GIF (and also as G, N , etc.). These three proteins-receptor, catalyst (C), and regulatory protein (G/F)constitute the essential hormone-sensitive adenylate cyclase. A fourth protein, calmodulin (calcium-dependent regulatory protein), has been shown to mediate the Ca2 -dependent stimulation of adenylate cyclase in membranes of brain and of C6 glioma cells (Cheung et af., 1978; Wolff and Brostrom, 1979). The mechanism of calmodulin’s effects on adenylate cyclase, and even its site of action, is still in dispute and will not be discussed here (see Toscano et al., 1979; Salter et af., 1981; Sano and Drummond, 1981). In addition, Schleifer et al. (1982) have suggested that yet another protein may interact with G/F. A physiological function for this protein, which so far is assayable only by its effects on the interaction of G/F with cholera toxin, is unknown. Hormonal inhibition of adenylate cyclase further suggests the presence of one or two other proteins that are involved with regulation of adenylate cyclase. It has been known for 20 years that a number of hormones inhibit the enzymatic activity of adenylate cyclase (Murad et al., 1962), and each presumably has a unique receptor in the plasma membrane of target cells. Since the receptormediated inhibitory effects also depend upon the presence of GTP, Rodbell and his colleagues suggested that there might be a distinct, inhibitory GTP-binding protein associated with this process (see Cooper, this volume). The groups of Limbird and Ui have tentatively supported this idea (Limbird et af., 1981; Limbird, 1981; Hazeki and Ui, 1981; Katada and Ui, 1981, 1982). Of the seven proteins mentioned above, only one, G/F, has been purified in useful quantity such that functional studies may be carried out (Sternweis et af., 1981). Other components may remain to be identified. Thus, one should consider hormone-sensitive adenylate cyclase to represent almost a rnultienzyme, information-transducing organelle, in the same category as the enzymes of oxidative phosphorylation. It cannot be likened to the simpler nicotinic cholinergic receptor that has been studied in such detail. This article will concentrate on the structures, activities, and interactions of C, GIF, and the stimulatory receptors that compose one segment of this organelle. The other proteins referred to above are discussed more fully in the works cited. +

112


B. Resolution of the Catalytic and Regulatory Proteins of Adenylate Cyclase The complex kinetic behavior of hormone-sensitive adenylate cyclase led Rodbell (1972) and others to propose the existence of a separate protein that acts to couple hormone receptors to adenylate cyclase. Since the site at which fluoride and guanine nucleotides stimulate the activity of adenylate cyclase seemed to be more closely linked to the enzyme than were hormone receptors, it was also suggested that these ligands might act upon such a regulatory protein or coupling factor. The existence of a coupling factor was also supported by the isolation of a variant S49 lymphoma cell in which adenylate cyclase activity and the ligandbinding activity of the receptor were both essentially at normal levels, although hormone binding no longer stimulated adenylate cyclase activity (Haga et ul., 1977b). These variant cells, referred to as UNC (for uncoupled), seemed to be defective in such a coupling factor. Direct evidence for the existence of a distinct regulatory protein in the adenylate cyclase system came from the work of Pfeuffer (1977) and Ross and Gilman (1977b). Pfeuffer found that chromatography of a detergent extract of pigeon erythrocyte plasma membranes on GTP-agarose decreased its responsiveness either to fluoride or to Gpp(NH)p.' If the GTP-agarose was washed with GTP or Gpp(NH)p, a factor was eluted that partially restored these responses. While neither the resolution of the two fractions nor the reconstitution of stimulation was quantitative, these experiments were the first to argue strongly for physically dissociable regulatory and catalytic components of adenylate cyclase. Ross and Gilman (1977b) demonstrated the presence of these two proteins by taking advantage both of complementary cell lines that were deficient in one or the other protein and of the proteins' differential stabilities to denaturation. These authors originally attempted to reconstitute hormonal stimulation of adenylate cyclase activity by reincorporating detergent-solubilized enzyme into membranes that already contained hormone receptors. They showed that a detergent extract of plasma membranes that contained adenylate cyclase activity could, under appropriate conditions, recombine with membranes of a phenotypically adenylate cyclase-deficient S49 lymphoma cell (denoted cyc - ) to yield hormonesensitive activity (Fig. 1) (Ross and Gilman, 1977a). The cyc- variant cells retain P-adrenergic receptors (Insel et al., 1976) but lack adenylate cyclase activity that is assayable in the presence of Mg2+ and ATP alone (Fig. 1). These experiments suggested that the mechanism of the reconstitution might be the interaction of solubilized enzyme with hormone receptors in or on the cyc-

'Abbreviations used: Gpp(NH)p, guanyl-5'-yl irnidodiphosphate; GTPyS, guanosine-5'-(3-thiotriphosphate); GDPpS, guanosine-5'-(2-thiodiphosphate).

113


400-

300-

[AC- Membranes]

(mglrnl)

FIG. I . Reconstitution of hormone-sensitive adenylate cyclase in membranes of cyc- S49 lymphoma cells. Adenylate cyclase activity was solubilized from plasma membranes of 882 fibroblasts using the detergent Lubrol 12A9. These cells have no P-adrenergic receptors. A fixed amount of the detergent extract was diluted into increasingly concentrated suspensions of cyc - plasma membranes. The membranes contain P-adrenergic receptors but essentially no adenylate cyclase activity when assayed in the presence of Mg* and the absence of Mn2 . Adenylate cyclase activity in the reconstituted mixture was assayed in the presence of Mg2+ plus the activators shown. Also shown is the stimulation of activity by isoproterenol relative to the GTP basal (dashed line). In retrospect, the increased activity stimulated by NaF or Gpp(NH)p probably reflects the contribution of extra C by the cyc membranes. The decline in activity with higher concentrations of cyc- membranes is probably caused by the addition of insufficient detergent to promote reconstitution. AC- is the original name of the cyc- variant. (From Ross and Gilman, 1977a.) +

+

~

membranes. However, thermal denaturation of the soluble enzymatic activity at 30°C led to only slightly decreased levels of activity in the reconstituted mixture. Thus, a heat-inactivated detergent extract from originally active plasma membrane could combine with the inactive cyc- S49 membranes to yield relatively high levels of adenylate cyclase activity that could be stimulated by fluoride, Gpp(NH)p, or hormone (Fig. 2) (Ross and Gilman, 1977b; Ross et a l ., 1978). Similarly the heated extract could reconstitute soluble fluoride- or Gpp(NH)pstimulatable activity upon combination with a detergent extract of cyc - membranes, These authors argued that the cyc - membanes (or extracts therefrom) were supplying a heat-labile factor intrinsic to adenylate cyclase that was destroyed during the heating of the complementary extract. The heated extract was hypothesized to provide a second, more stable component in which the cyc-

114

ELLIOTT M. ROSS ET AL. 100

Donor a1

5

IS

10

20

25

30'C

30

Time (min)

FIG. 2. Selective thermal denaturation of the catalytic component of adenylate cyclase. Adenylate cyclase activity was solubilized from plasma membranes of wild-type S49 lymphoma cells using Lubrol 12A9, and the extract was heated at 3OoC for various times. Activity in the extract was assayed after dilution in detergent-free buffer, and the decay of activity with time is shown by the dashed lines [upper line, Gpp(NH)p-stimulated; lower line, NaF-stimulated]. Aliquots of extract were also mixed with cycplasma membranes and the mixtures were assayed for activity (solid lines). Assays were performed in the presence of the stimulating ligands shown in the figure. Note that after all assayable activity in the extract is lost (20-30 minutes of heating), its ability to reconstitute activity in the cyc- membranes is decreased only slight if at all. INE, (-) Isoproterenol (isopropylnorepinephrine). (From Ross er a/., 1978.)

cells were deficient. Sensitivity to proteases, heat, and sulfhydryl reagents suggested that both factors were proteins (Ross and Gilman, 1977b). In addition to establishing the existence of the two proteins, this work also provided novel, reconstitutive assays for their activities, thereby allowing their continued study and fractionation. Studies with the crude preparations described above led these authors to propose that the more thermostable protein, which was missing in cyccells, served a regulatory function and mediated the effects of guanine nucleotides, since its activity was stabilized by either GTP or Gpp(NH)p and it bound to Pfeuffer's GTP affinity matrix. The identification of the more labile protein in cyc- membranes as the catalyst was facilitated by the finding that these membranes contain a Mn2 -stimulated adenylate cyclase activity which had not previously been observed. This Mn2 -stimulated enzyme displayed hydrodynamic properties similar to the protein that was active in reconstitution with the thermostable factor. The reconstitutive and catalytic activities +

+

115


also displayed similar thermostabilities that could be varied in parallel by ATP, Mn’ , or MgZ+ . These studies, taken together, demonstrated that the catalytic protein, C, and the regulatory protein are indeed separate. They also showed that guanine nucleotides and fluoride (hence the name “G/F”), as well as hormone receptors, exerted their stimulatory effects on the regulatory protein rather than directly upon C. G/F was further implicated as a true coupling factor between catalyst and receptor by the observation that it is also responsible for mediating the negatively cooperative binding interaction between hormones and guanine nucleotides. The affinity of agonist, but not antagonist, ligands of adenylate cyclase-linked receptors is usually decreased in the presence of those purine nucleotides that permit hormonal activation of adenylate cyclase (see, for example, Ross et al., 1977; or Ross and Gilman, 1980). Ross et al. (1977) noted that receptors from cell lines that had unaltered G/F showed this interaction, but that loss or chemical modification of G/F destroyed the effect of nucleotides on the affinity of hormone binding. Sternweis and Gilman (1979) confirmed the role of G/F in this effect by showing that reconstitution of cyc- or UNC plasma membranes with G/F restored the negative heterotropic interaction of hormone and nucleotide. +

C. Catalytic Protein of Adenylate Cyclase: C 1. PREPARATIONS A N D PROPERTIES Physical and enzymological studies of C have been limited by the difficulty of preparing C that is free from G/F and that still retains reasonable stability. Most of the initial observations related to C came from studies of plasma membranes from cyc- S49 lymphoma cells or of crude detergent extracts therefrom. Hydrodynamic studies suggested that the molecular weight of C from cyccells is about 1.9 X los and that it binds a significant amount of detergent (Ross et a!., 1978). The latter property is suggestive of a large hydrophobic surface area that is characteristic of intrinsic membrane proteins (Clarke, 1975). These early results are roughly consistent with more recent data on cholate-solubilized preparations (Strittmatter and Neer, 1980; Ross, 1981), but the tendency of C to aggregate, even in the presence of detergent, requires more detailed study using the purified protein. Preliminary characterization of Lubrol-solubilized C from cyccells also showed that C is an extremely heat-labile protein, and that it is stabilized somewhat against denaturation by Mn2 plus ATP (Ross and Gilman, 1977b). It is also stabilized by high ionic strength (Ross, 1981). phosphatidylcholine (Strittmatter and Neer, 1980; Ross, 1981, 1982), and forskolin (E. M. Ross, unpublished). C contains at least two sulfhydryl residues that are sensitive to N-ethylmaleimide, the more reactive of which is required for interaction with G/F but not for catalytic activity (Ross et a / . , 1978). +

116


fE I

10

15

20

Elution Volume (ml)

FIG. 3. Separation of C and G/F by gel filtration of a cholate extract of hepatic plasma membranes. C was assayed either in the presence of Mn2+ ( 0 )or of added G/F, M g z + , and GTPyS (A), GIF was assayed according to its ability to mediate the activation of added C by GTPyS (0). The separation was carried out in buffer containing 0.5 M (NH&S04 and 12.5 mM cholate. In more recent preparations of C, the activity recovered has been increased up to fourfold. (From Ross, 1981.)

More recently, other preparations in which active C has been resolved from G/F have become available. Strittmatter and Neer (1980) and Ross (1981) demonstrated that C could be resolved by gel filtration in cholate solution if manipulations were carried out at high ionic strength (Fig. 3). Hepatic C, prepared in this way, displayed properties essentially identical to those of the enzyme from cyc- S49 cells. It was judged to be free of contamination by G/F according to (1) a high ratio of Mn2+-stimulated activity to basal activity in the presence of Mg2+, (2) lack of stimulation by either GTPyS or fluoride, both of which responses could be restored by the addition of pure G/F, (3) undetectable levels of G/F activity as assayed by the stimulation of a large excess of added C (see below), and (4) the absence of measurable amounts of the 45,000-dalton substrate for cholera toxin (Table I). Little progress has been made toward purification of C, primarily because it appears to be quite unstable in detergent solution. The t,,* for denaturation of soluble, resolved C is 24 hours at best. Storm's group has reported a specific activity of 15 nmoles min- I mg- I for a partially purified fraction from cerebral cortex (Westcott et al., 1979), but this value is at best about 0.01 of that which might be expected if the enzyme were homogeneous. The purification of C is clearly necessary for the study of its regulation, and the development of strategies to deal with its apparent lability deserves a high priority.

117


TABLE 1 ADENYLATE CYCLASEACTIVITIES OF C PREPARATION FROM RABBITLIVEROR CYC- S49 LYMPHOMA CELLSO Specific activity (nmole min- I mg-

1)

C

Additions None

Plus GIF

Plus cyc- membranes

Assay conditions MnCI, MgC12 MgC12. NaF MgC12, GTPyS MgCl2 MgC12, NaF MgC12. GTPyS MgC12, NaF MgC12, CTPyS

Rabbit liver

cyc- S49 cells

61

174 12 9 7 12 I39 256

8 7 7 7 57 151 4 8 ~

Liver (C

58

I1 89 276 10 86 25 1 757 1335

3 N.D.“ ~~

+ CIF)

~~

Adenylate cyclase activities in preparations of C from rabbit liver or from cyc- S49 cells, both added as cholate solutions. are compared with those of cholate-solubilized. unfractionated hepatic plasma membranes. Activities were assayed either in the absence of additions, after addition of a saturating amount of hepatic GIF, or after addition o f a large excess of cycplasma membranes. The latter assay gives an estimate of total GIF activity. Data are from Ross (1981). N.D., Not determined. ‘I

2. REGULATION

C was originally described as being totally inactive in the presence of Mg2+ and active only in the presence of Mn2+ or of G/F. It is now clear that the catalytic activity of C is exquisitely regulated by at least four ligands and that the response to these ligands is modulated by the hydrophobic environment in which C is located. While G/F is presumably the physiologically most important ligand, C is also stimulated by Mn2+ and by forskolin (7P-acetoxy-8,13-epoxyla,6P,9a-trihydroxylabd-14-en-I I-one) (Seamon and Daly, 1981). Neer and Salter (1981) and our laboratory (Florio and Ross, 1982a,b) have also shown that adenosine acts directly upon C at the so-called P site to inhibit activity. The interaction of C with G/F will be dealt with in Section III,B and regulation by small molecules will be considered here. Divalent manganese was first shown to stirnulate C in plasma membranes of cyc- S49 cells. Relative to an extremely low “basal” activity measured in the presence of Mg2+, Mn2+ could stimulate activity up to 20-fold. However, the extent of stimulation by Mn2 was variable, particularly in native membranes. +

118


The source of this variability has not been defined, but data obtained using soluble preparations are much more uniform. In a Lubrol 12A9 extract of cycplasma membranes, free Mn2+ in the 0.5-2 mM range reproducibly stimulates activity 15- to 20-fold (Ross et al., 1978). A 7- to 10-fold stimulation by Mn2 is more typical of resolved hepatic C (Ross, 1981), in agreement with data obtained for cerebral cortical C (Strittmatter and Neer, 1980). Careful study of the stimulation of isolated C by Mn2+ has not been undertaken. However, it appears that free Mn2 is the activating ligand as opposed to MnATP acting as a preferred substrate. Stimulation is maximal at calculated concentrations of free Mn2+ of 0.5-2 mM over a wide range of ATP concentrations (20-2000 and the addition of 10 mM MgCI, or 1 mM EDTA does not alter activity significantly (V. A. Florio and E. M. Ross, unpublished data). Neer (1979) suggested further that Mn2+ is also necessary for the stimulation of C by G/F, citing the ability of Mn2+ to reverse the blockade of G/F-mediated stimulation that is caused by EGTA. It will be difficult to substantiate this claim using currently available preparations. Possibly, a second low-molecular-weight activator of C has recently been identified by Seamon and Daly (1981). Forskolin, a substituted diterpene extracted from Coleus forskohlii, had been identified originally according to its cardiovascular regulatory activities (Lindner et d . , 1978). Seamon et d. (1981) demonstrated that forskolin in the 1-100 pkif range was active in elevating adenylate cyclase activities in numerous tissues, and Seamon and Daly (198 1) found that forskolin also activated adenylate cyclase in plasma membranes of cyc- S49 cells. Since this preparation lacks G/F, these authors suggested that forskolin directly activates C. We have recently shown that forskolin also stimulates resolved rabbit hepatic C (Ross, 1982). Thus, while none of these test systems is pure, it is likely that forskolin is a ligand of C or of a closely associated protein. In our hands, forskolin reversibly stimulates the activity of C up to &fold in the presence of either Mn2 or Mg2 . Quantitation of stimulation is difficult, however. Forskolin is not soluble in water at concentrations greater than about 0.2 mM, and it may be difficult to demonstrate maximal activation due to this limited solubility. The potency of forskolin is also hard to quantitate, since forskolin presumably can partition into the lipid phase of membranes or into detergent micelles. Regardless, forskolin frequently activates adenylate cyclase in membranes more than any other agent. Given these experimental qualifications, forskolin also seems to activate C with greater potency in the presence of G/F, suggesting that forskolin and activated G/F bind with positive cooperativity to independent sites. Forskolin is slightly less potent in activating adenylate cyclase in cyc- S49 plasma membranes as compared to wild type, and the addition of GTPyS-activated G/F to resolved hepatic C increases forskolinstimulated activity and decreases the concentration of forskolin that gives halfmaximal activation (E. M. Ross, unpublished). +

+

w),

+

+

119


Negative regulation by adenosine and by its ribose-modified analogs has also been observed in all preparations of C studied so far. Such inhibition was noted previously in plasma membranes from numerous tissues. Since this inhibition is specific for analogs with an unmodified purine moiety, this site has been called the P site to distinguish this function from the inhibition and stimulation of adenylate cyclase caused by ribose-specific analogs that act on cell surface receptors (R sites) (Londos and Wolff, 1977; Londos et al., 1979). Wolff et al. (1978) proposed that the P site might exist on C because P site-mediated inhibition was retained after solubilization of membranes by detergent and because it did not require guanine nucleotides. This idea was supported by Premont et al. ( 1979), who observed P site-mediated inhibition after functionally uncoupling C and GIF using cholate. We have confirmed that the P site resides on C or on a closely associated protein by demonstrating inhibition of activity by 2' ,5'-dideoxyadenosine (DDA), a P site-specific compound using either plasma membranes of cyc- S49 cells or our preparation of resolved hepatic C (Florio and Ross, 1982a,b). Adenosine also has no direct effect on the rate of activation or deactivation of GIF, although it may promote the binding of activated G/F to C. The inhibition of C by P site agents is complex. Fractional inhibition of activity of resolved C is greatly increased in the presence of activators. Thus, DDA can decrease the activity of C nearly to basal values whether activity is initially stimulated 7- to 10-fold by M n * + , 30- to 50-fold by forskolin plus M g 2 + , or over 100-fold by either activated G/F or by forskolin plus Mn2+

"

10-8

10-6

10-2

[2',5'-Dideoxyadenosine] (M)

FIG. 4. Inhibition of C in cycplasma membranes by 2'5'-dideoxyadenosine (DDA). Adenylate cyclase activity in plasma membranes of cyc - S49 lymphoma cells was assayed in the presence of increasing amounts of DDA. Reaction volumes contained 25 mM MgC12 (a),2.5 mM MnCI2 (0). or MnClz plus 0.3 mM forskolin (A).

I

4 1.0

i1:: 0.8

Q6

0

Y

'3%

0

1u3

[Forskolin] (M) FIG.5 . Forskolin makes DDA a more potent inhibitor of C. The concentration dependence upon DDA as an inhibitor of Mn2+-stimulated adenylate cyclasc activity in plasma membranes of cyccells was determined at different concentrations of forskolin. Two measures of potency are shown: the concentration of DDA at which total activity is decreased by 20% (0)and the concentration of For DDA at which inhibition is 50% of the maximum inhibition observed with 3 mM DDA (0). comparison, the stimulation of adenylate cyclase activity by forskolin under similar conditions is also Note that forskolin by itself causes no increase in activity at concentrations below lo-" shown M .A similar disparity exists for GTPyS-activated GIF between the concentration needed to stimulate the activity of C and the concentration needed to potentiate inhibition by DDA.

(A).

C'A

--

KIO

CoA

FIG. 6. A minimal three state model for the action of P site inhibitors. Activating ligand (L) is assumed to bind preferentially to an active state of the enzyme (C*). Analogs of adenosine (A), in the presence of L. bind preferentially to the third, inhibited state (Co). The activities of the basal state (C) and inhibited state (C") are assumed to be zero. Note that the preferential binding of P site agents to basal C will yield apparent competitive inhibition with respect to L. Cooperative binding of L and A to the Co state is required to simulate experimental data. This model and the derivable constants for DDA and for Mn2+, forskolin, or GTPyS-activated G/F are described in more detail elsewhere (Florio and Ross, 1982b).


121

(Fig. 4). Inhibition of basal activity (i.e., Mg2+ alone) is rarely greater than 30%. This pattern would be consistent with a simple, two-state allosteric model in which activators bind preferentially to an active state of C and DDA binds to an inactive basal state, except that this simple scheme would predict an apparent competition between activator and inhibitor. As shown in Fig. 4, DDA inhibits at lower concentrations in the presence of activators, and the increase in apparent affinity for DDA at the P site is directly related to the maximum efficacy of the activator. By this we mean that DDA is most potent in the presence of activated G/F or forskolin (K& = 0.5-2 pM) and least potent at inhibiting basal activity (I& > 5 mM). The effect of activating ligands of C on its apparent affinity for DDA is reciprocal, as would be predicted thermodynamically. Thus, while forskolin stimulates the activity of C half-maximally at about 50 pk! (Seamon and Daly, 1981; Florio and Ross, 1982a,b), it potentiates the inhibitory activity of DDA half-maximally at less than 1000-fold lower concentrations (Fig. 5). We have noted a similar, quantitatively reciprocal interaction of DDA with activated G/F, and Mn2+ appears to display a similar effect. Unless a previously undetected protein that mediates the response of one ligand or the other exists, these data minimally demand a three-state model for C (Fig. 6), in which a nonactivated (basal) state, an active state, and an inhibited state coexist. P site agents bind with low affinity to the basal or active state and most tightly to the inhibited state. Activators alone bind preferentially to the active state. To fit our data quantitatively, it must be assumed further that activator and P site agent bind positively cooperatively to the inhibited state. If the shifts in apparent affinity shown in Figs. 5 and 6 represent true changes in K,s for the ligands, then the free energy of coupling for this interaction is a strikingly large number (AGA.L = -6 kcal) (Weber, 1975). It should be stressed that this formalism, although adequate, does not demonstrate that a simple three-state, cooperative model is correct. The large, negative AGA,L almost indicates that the mechanism must be more complex. Fortunately, the observed interactions should allow us to develop ligandbinding assays to explore the mechanism in greater detail. The major conclusion is that, regardless of detail, the entity (or entities) that we now refer to as C is a highly regulated and complex enzyme rather than a mere mirror of the activation of G/F.

D. The Stirnulatory GTP-Binding Regulatory Protein: G/F G/F is by far the best studied and understood protein of the adenylate cyclase system, largely because of technical reasons. It was noted soon after its discovery that it was the most stable protein (Ross, et al., 1978) and that it behaved as a monodisperse species in solutions of several detergents (Howlett and Gilman,

122

ELLIOlT M. ROSS ET AL.

1980; Kaslow et al., 1980; Sternweis et al., 1981). It can be [3ZP]ADP-ribosylated on one of its subunits, facilitating its identification in crude mixtures (Gill and Meren, 1978; Cassel and Pfeuffer, 1978). It can also be identified by its binding of a 32P-labeled photoaffinity analog of GTP and by its binding to a GTP-agarose affinity matrix (Pfeuffer, 1977). These properties motivated the initial purification of G/F from rabbit liver by Northup er al. (1980). This preparative procedure has been markedly improved and applied to G/F from turkey and human erythrocytes (Sternweis et al.. 1981; Hanski et al., 1981). A more detailed description of G/F and its properties is available in these references and in a discussion by Smigel et al. (1982). Purified preparations of G/F contain two or three polypeptides of molecular weights 35,000, 42,000-45,000, and 52,000, as determined by electrophoresis of the dodecyl sulfate-denatured protein (Northup et al., 1980). The purified native protein behaves hydrodynamically as a particle with a molecular weight of about 80,000 (Sternweis et al., 1981; Hanski et af.,1981). The discrepancy between the native molecular weight and the sum of the molecular weights of the three subunits, combined with other data, suggests strongly that a molecule of G/F is a dimer composed of one 35,000-dalton subunit and one subunit of either 45,000 or 52,000 daltons (Sternweis et al., 1980). It is likely that the 45,000dalton protein is a proteolytic product of the 52,000-dalton protein, since the two are similar functionally (see below) and yield similar peptide maps (Hudson and Johnson, 1980). The 52,000-dalton form is entirely lacking in avian erythrocytes (Gill and Meren, 1978; Cassel and Pfeuffer, 1978; Pfeuffer, 1977; Hanski et al., 1981), and Larner and Ross (1981) have shown that it is preferentially lost as rat erythrocytes mature from the reticulocyte stage. Purified hepatic G/F has been shown to bind at least 1 mole/mole of [35S]GTPySwith a concentration dependence similar to that observed for activation, and other nucleotides compete with [3sSS]GTPySfor binding to this site with appropriate potency (Smigel et al.. 1982). It is uncertain whether binding sites of lower affinity exist, but it is likely that this high-affinity site resides on the 52,000 (45,000) -dalton polypeptide. This conclusion derives from the labeling of this protein with [32P]GTP-y-azidoanilide(Pfeuffer, 1977). This polypeptide also contains the site at which G/F is specifically ADP-ribosylated by cholera toxin, thus suggesting that it also contains the active site for the hydrolysis of GTP (Cassel and Selinger, 1977b). All data so far available suggest that this larger polypeptide is the proximal activator of C. The key question in the study of G/F and its function is the process whereby it is activated. Unliganded G/F does not appreciably alter the activity of C. It is its activation, either by fluoride or a nonhydrolyzable analog of GTP (or, transiently, by GTP), that converts G/F into a potent stimulator of the adenylate cyclase activity of C. Howlett et al. (1979) first showed that G/F, in the absence of C, could be stably activated by fluoride or Gpp(NH)p in detergent solution.

123


Activated G/F was assayed subsequently according to its ability to stimulate C’s catalytic activity. Activation of G/F, i.e., its conversion to a form which can stimulate C , is not synonymous or necessarily coincident with binding of the activating ligand. Activation and deactivation of G/F are both first-order processes, and their rates are dependent upon both the concentration of ligand and the concentration of the 35,000-dalton subunit (Smigel et af., 1982). Activation by nucleotide is also dependent on rather high concentrations of Mgz+ (Sternweis er al., 1981; Hanski et al., 1981), and activation by fluoride has recently been shown to be dependent on the presence of A13 or Be2 (Sternweis and Gilman, 1982). The accurate measurement of activation and inactivation is complex, and the molecular mechanism whereby it is regulated is still speculative. Interested readers are directed to an excellent discussion of the problem by Smigel and co-workers ( I 982). G/F is best assayed reconstitutively by virtue of its ability, when activated by fluoride or a guanine nucleotide, to stimulate the catalytic activity of C (Ross and Gilman, 1977b; Ross et al., 1978; Sternweis et af., 1981). The sample to be assayed is mixed in an appropriate medium with an excess of C and the mixture is allowed to anneal in the presence of a stirnulatory ligand of G/F (fluoride or GTPyS). Adenylate cyclase activity is then measured in the presence of Mg2+ (Northup et af.. 1980; Sternweis et af., 1981). This assay is sensitive, easy, and reliable. It is strictly quantitative, however, only when certain conditions are met (Sternweis er al., 1981). First, the amount of C used in the assay must truly be saturating with respect to the amount of G/F that is to be assayed (Fig. 7), reflecting the bimolecular nature of the G/F-C interaction. It is clearly not sufficient to use an amount of G/F small enough so that reconstituted activity is linear with added G/F (Sternweis er al., 1981; Lamer and Ross, 1981). The second major problem encountered in assaying G/F is that detergents used to solubilize G/F, primarily cholate, can inhibit or denature the C that is used in the assay, necessitating the dilution of G/F into a standard medium prior to assay. These two limitations have detracted from the quantitative accuracy and reproducibility of several published measurements of G/F, including the early work of Ross et al. (1978) (see Lamer and Ross, 1981, for more discussion). G/F has been recognized recently as one of the prominent loci at which variation in the regulation of adenylate cyclase systems among different cells is expressed. Kaslow et al. (1979) first suggested that differences in the quantitative responses to hormones and guanine nucleotides that exist between adenylate cyclases from different tissues might be due to differences in the G/F protein. These authors compared G/F from wild-type S49 lymphoma cells with G/F from turkey erythrocytes. The former is a typical mammalian system. The latter is relatively unresponsive either to hormone in the presence of GTP [as compared with Gpp(NH)p or GTPyS] or to Gpp(NH)p alone in the absence of hormone. Allowing for inefficient reconstitution and low activities, these properties were +

+

124

ELLlOlT M. ROSS ET AL.

FIG.7. Assay of G/F in rat erythrocyte plasma membranes. G/F activity was solubilized from rat erythrocyte membranes using cholate at high ionic strength. Aliquots of extract, at varying concentrations of protein, were mixed with aliquots of cycplasma membranes (as the source of C). Several different concentrations of cyc- membrane were used, and contributed to each assay the amount of protein shown at the right. The reconstituted mixtures were then activated with GTPyS, and adenylate cyclase was assayed in the presence of MgC12. Note that activity is a relatively linear function of the amount of G/F added to each assay, even at low concentrations of C. Only with greater amounts of C. however, is the measured specific activity of GIF (i.e., the slope of the line) both maximal and constant. (From Larner and Ross, 1981.)

retained when G/F from each of the two different cell types was reconstituted into plasma membranes from cyc- S49 cells. This work has now been confirmed with purified G/F (Hanski et al., 1981). These data argue that G/F itself determines qualitative aspects of the regulation of adenylate cyclase in plasma membranes. This finding is perhaps more interesting when combined with the observation that G/F from avian erythrocytes lacks the 52,000-dalton form of the GTP-binding subunit of G/F-only the 45,000-dalton form is detectable. When Sternweis et al. (1981) compared fractions of rabbit hepatic G/F that were enriched in either one form or the other, they found that the fraction rich in the 45,000-dalton form behaved similarly to turkey erythrocyte GIF with respect to kinetics of activation and Mg2+ dependence. Lamer and Ross (1981) made a similar finding in a study of the maturing rat reticulocyte. They found that the loss of the 52,000-dalton subunit of G/F during maturation correlated well with the increased preference for GTPyS over GTP as cofactor for hormonal stimulation of adenylate cyclase. Changes in Mg2+ dependence for the activation of G/F were also noted by these authors. These observations in several different systems all lead to the speculation that the proteolysis of the 52,000-dalton subunit of G/F to the 45,000-dalton form leads to qualitative changes in the

PROTEIN COMPONENTS OF ADENYLATE CYCIASE

125

regulation of adenylate cyclase. It will be fascinating to determine if this proteolysis is regulated by endocrine or developmental factors. A second modification of G/F which may be of importance in the physiological regulation of adenylate cyclase is suggested by the phenotype of another variant S49 lymphoma cell. Haga et al. (1977b) described a stable variant clone in which hormones (P-adrenergic amines or prostaglandin E,) failed to stimulate adenylate cyclase activity. These cells retain adenylate cyclase activity, which can be stimulated by fluoride or Gpp(NH)p, and also retain P-adrenergic recepThe lesion tors, as measured by the binding of [ ‘251]iodohydroxybenzylpindolol. appears as an uncoupling of receptors from enzyme, leading to the name “UNC,” for uncoupled. The UNC lesion and the cyc- lesion are not complementary with regard to hormonal stimulation of the enzyme, as assayed either by reconstitution protocols (Ross and Gilman, 1977a; Sternweis and Gilman, 1979; Schwarzmeier and Gilman, 1977) or in somatic cell hybrids (Naya-Vigne et al., 1978). It can be inferred, therefore, either that cyccells are deficient both in G/F and a putative “UNC factor” or that G/F in UNC cells is somehow defective. Sternweis and Gilman (1979) argued for the presence of an altered GIF in UNC plasma membranes by demonstrating that a crude preparation of G/F from wild-type S49 cells or rabbit liver can restore hormonal responsiveness to UNC plasma membranes. This ability to reconstitute hormone responses in UNC membranes cofractionates several thousand-fold with G/F, When UNC G/F is [”PJADP-ribosylated using cholera toxin, it is also observed that the 45,000dalton subunit of G/F is shifted to a more acidic isoelectric point (Schleifer e t a l . , 1980). The UNC lesion also abolishes control by guanine nucleotides of the affinity of hormone binding to receptors (Haga et al., 1977b), and this loss is reversed by reconstitution with crude G/F (Sternweis and Gilman, 1979). These results, taken together, provide a strong argument that the UNC lesion represents a modification (or lack of required modification) of the G/F protein such that it can no longer fulfill its role as a coupling factor between receptor and C. Since several physiological states have been described in which G/F appears to be uncoupled (see Ross and Gilman, 1980), it is tempting to speculate that the UNC lesion represents the uncontrolled expression of a normal physiological regulatory function. Understanding the chemical nature of the lesion should allow us to explore the basis of this regulation.

E. Cell Surface Receptors That Stimulate Adenylate Cyclase The ligand-binding properties of a wide variety of hormone receptors that act via adenylate cyclase have been studied in the membranes of target cells by the use of appropriate radioactive ligands; a rather large number of receptors have

126


also been characterized after detergent solubilization. (see Ross and Gilman, 1980, for a brief review). Several receptors have also been affinity labeled, allowing the identification of their subunits by dodecyl sulfate-polyacrylamide gel electrophoresis (Johnson et al., 1981; Ji and Ji, 1980; Rebois et al., 1981; Atlas and Levitzki, 1978; Rashidbaigi and Ruoho, 1981; Lavin et al., 1981), and the P-adrenergic receptor has been substantially purified (Vauquelin et al., 1979; Durieu-Trautmann et al., 1980; Shorr et al., 1981). Ligand-binding characteristics of soluble receptors are generally unchanged from those of the membranebound protein, although some hormone-binding activities are altered by the receptor's environment. In the case of the glucagon receptor, it has not been possible to bind hormone to the solubilized protein, although a receptor-ligand complex has been solubilized from hepatic membranes that were first incubated with [ 1251]iodoglucagon(Welton et al., 1977). This may reflect a stabilizing effect of ligand upon the receptor, implying a relative instability of the unliganded, detergent-solubilized protein. This effect has been noted to a lesser extent with other receptors. Similarly, alkyl polyethylene oxide detergents (Lubrol, Brij) have permitted only solubilization of preformed complexes of (3-adrenergic receptors and their radioactive ligands (Haga et al., I977a), although Caron and Lefkowitz (1976) showed that digitonin could be used to solubilize P-adrenergic adaptors in a form that could bind the antagonist ligand [3H]dihydroalprenoloI (DHA). They further demonstrated the binding of unlabeled agonist and antagonist ligands to soluble receptors by using competitive binding assays that employed [3H]DHA. Similar results have been reported briefly by Witkin and Harden (198 I ) , who used [ 12sI]iodopindolol, also an antagonist, as their probe. Another specific P-adrenergic antagonist ligand, [ '251]iodohydroxybenzylpindolol (IHYP), seems to be most sensitive to small effects of membrane environment on receptor structure. It has not been possible to detect binding to digitonin-solubilized receptors using this ligand-binding activity is lost upon the addition of detergent (Haga et al., 1977a; Fleming and Ross, 1980; Witkin and Harden, 1981). Fleming and Ross (1980) demonstrated that the ability of digitonin-solubilized P-adrenergic receptors to bind [ 1251]IHYP can be restored by their reconstitution into phospholipid vesicles. Starting with a mixture of soluble receptors and dimyristoylphosphatidylcholine,they used gel filtration (and, originally, sucrose density gradient centrifugation), to remove digitonin and to incorporate receptors into unilamellar vesicles of 500- 1000 A diameter. The reconstitution had little effect on the number of receptors that could be assayed using ["IDHA or on the K , for this ligand but was efficient in restoring the binding of [ 1251]IHYPto expected levels. Since this reconstitution of [1251]IHYP-bindingactivity utilizes receptors that were solubilized in digitonin, it should be equally applicable to receptors that are purified in that detergent by the techniques cited above. The procedure has now been applied to receptors derived from rat and turkey erythrocytes and from murine S49 lympho-


127

ma cells. [12sI]IHYPbinding can be similarly restored to P-adrenergic receptor that have been solubilized with deoxycholate by removing detergent and incorporating the receptors into phosphatidylcholine vesicles (Citri and Schramm, 1980; Pedersen and Ross, 1982a,b). In the case of deoxycholate, the solubilized receptors were unable to bind either [3H]DHA or ['251]IHYP, and reconstitution restored the ability to bind both ligands (Pedersen, and Ross, 1982b). The molecular events that cause the reversible inhibition of IHYP binding by digitonin and the inhibition of either IHYP or DHA binding by deoxycholate are unknown. However, these effects may suggest that the hydrophobic environment of the receptor is important in determining its ability to bind P-adrenergic ligands, presumably because of effects on its conformation. The distinct behavior of IHYP and DHA with digitonin-solubilized receptors suggests that detergents do not merely occlude a binding site. This role of lipids on a receptor's ligand-binding activity is apparently distinct from the requirements for a phospholipid membrane for the interaction of receptor with GIF. However, both of these effects may indicate an exquisite sensitivity of the conformation of hormone receptors to their surroundings in the plasma membrane bilayer.

111.

PROTEIN-PROTEIN INTERACTIONS AND THE REGULATION OFADENYLATECYCLASE

A. General Considerations The hormonal regulation of adenylate cyclase activity can be thought of most simply as a sequence of two distinct regulatory interactions between two different pairs of proteins. First, receptor binds hormone, undergoes some conformational change, and acts allosterically on G/F to promote the activation of G/F by guanine nucleotide. Second, activated G/F binds to C and stimulates its catalytic activity. Since this scheme involves two small and two macromolecular allostenc ligands, it is clearly not all that simple. There is also good reason to believe that, in native membranes, several of these reactions may be concerted or that stable assemblies of these proteins may exist (see Tolkovsky et al., 1982; Braun and Levitzki, 1979; Tolkovsky, this volume; and references therein for some of the best arguments in this direction). Such complexes may be further stabilized when inserted correctly in a membrane bilayer or they may be stabilized by cytoskeletal elements (see, for example, Rudolph et af., 1977; Insel and Kennedy, 1978; Rasenick et a!. , 1981 ; Sahyoun et al., 1977, I98 1). Regardless, considering the total system as the sum of two ligand-mediated protein-protein interactions has the virtue of relative conceptual simplicity and seems to be a valid experimental approach. Each interaction has been studied separately. The interaction of receptors with G/F has been studied by measuring

128


the stimulatory effects of hormone binding on the rate of activation of G/F by guanine nucleotides in reconstituted systems (Citri and Schramm, 1980; Pedersen and Ross, 1982b) as well as in membranes of HC-I hepatoma cells, which are essentially deficient in C (E. M. Ross, unpublished; see also Ross et al., 1978). The negatively cooperative binding interactions of guanine nucleotides and hormones, which also reflect the coupling of receptors to G/F, are also observable in HC-1 membranes or in wild-type S49 cell membranes in which C has been chemically or thermally denatured (Ross et al., 1978). Preliminary data from our laboratory indicate that they may also be observed in phospholipid vesicles that contain G/F and P-adrenergic receptors. The stimulation of C by activated G/F has been studied in relatively well resolved systems (see, for example, Sternweis et a!., 1981; Smigel e t a l . , 1982; Ross, 1981, 1982). Reconstituted systems containing only two interacting proteins have the advantage that one can independently manipulate the relative concentrations of each protein in order to estimate their affinities for each other, their rates of interaction and dissociation, and the modulation of these rates by regulatory ligands and by the lipid composition and structure of the membrane. The discussion below will concentrate primarily on work that has utilized this approach.

B. Activation of C by G/F Until 1977, C and G/F together were generally considered to compose a single enzyme, adenylate cyclase, that was variably stimulated in different plasma membranes by Mg2 , Mn2 , F- , and guanine nucleotides. In many cases, this assumption is at least practically useful, in that the binding of C and G/F can be quite stable, such that they apparently do not dissociate over several days. Hydrodynamic studies of the detergent-solubilized C-G/F complex, usually stabilized by an activating ligand such as Gpp(NH)p, have yielded profiles of monodisperse species with properties consistent with a molecular weight of about 2 X lo5 (see Ross and Gilman, 1980, for a review). Tolkovsky et al. (1982) and Tolkovsky and Levitzki (1981) proposed that a stable C-G/F complex, rather than free G/F, is the principal species that is activated by hormone receptors in native membranes. They found that in turkey erythrocytes, the formation of the C-G/F complex was not rate limiting for the hormonal activation of adenylate cyclase and concluded that the C-GIF complex was therefore relatively stable prior to activation. This conclusion may not be general, however. In plasma membranes from S49 lymphoma cells, for example, there appears to be a molar concentration of G/F greater than or equal to the concentration of C plus P-adrenergic receptors. Thus, when adenylate cyclase activity is maximally stimulated by Gpp(NH)p, there is still enough G/F to interact reversibly with receptors (Ross et ai., 1977). This is also true in hepatic plasma membranes and +

+


129

detergent extracts therefrom (Welton et al., 1977; Lad et al., 1977). If a C-GF complex does exist, it must be relatively transient and allow for the fairly rapid exchange of G/F and C molecules, since excess exogenous G/F molecules added to a native membrane have been shown to interact with endogenous C. Thus, G/F from wild-type 549 lymphoma cells can, when added to plasma membranes from UNC cells, compete for C with endogenous UNC G/F; added ADP-ribosyl-G/F can compete with endogenous unmodified G/F (Sternweis et al., 1979); and G/F that has been preactivated with GTPyS competes with native G/F in membranes to which it has been added (Larner and Ross, 1981). These studies argue for the potential of rapid and regulated exchange of C and G/F molecules in membrane, but it is difficult to assess how rapidly exchange occurs in an unperturbed membrane.. It is also frustrating that it has been impossible to count molecules of C in a membrane, so that there is no real way to determine the molar stoichiometric relationship of C and G/F. Such measurements will probably demand reconstitution of purified preparations of each protein in a suitable medium. In crude reconstituted systems and in detergent solution, it has been somewhat easier to deal with the interaction of C and G/F in at least semiquantitative terms. Sternweis et al. (1981) showed that the formation of the activated C-G/F complex could be treated as the simple binding equilibrium, C + G/F e C-G/F. They used the activation of C to measure the amount of complex that was formed and used purified, GTPyS-activated hepatic G/F and cycplasma membranes as their sources of each component. The concentration of G/F was manipulated directly. The concentration of C was altered either by changing the concentration of cyc - membranes or by mixing varying proportions of native membranes and membranes in which C had been inactivated. These findings are consistent with the less extensive data of Larner and Ross (198 l), which were obtained using crude rat erythrocyte G/F. Pfeuffer (1979) used sucrose density gradient ultracentrifugation in Lubrol solution to study the interaction of Lubrol-solubilized C and G/F that had been separated by affinity chromatography. He found that in the presence of active C, G/F could be caused to sediment as a larger particle by the addition of GTPyS. Centrifugation of G/F alone, of G/F plus C but in the absence of nucleotide, or of G/F plus C in the presence of GDPpS gave similarly smaller sedimentation rates for G/F. These data were interpreted to indicate that only nucleotide-activated G/F could form a stable C-G/F complex in Lubrol solution. This conclusion is consistent with numerous observations from other laboratories on the stabilization of detergent-soluble adenylate cyclase activity by fluoride or by analogs of GTP. A major uncertainty about the interaction of membrane-bound C with added soluble G/F is the extent to which G/F can reassociate with the membrane. Ross and Gilman (1977a) initially found that Lubrol-solubilized G/F could stimulate C on cyc - membranes and mediate hormonal stimulation of adenylate cyclase

130


activities under conditions where it was not stably bound to the membrane. Howlett et al. (1979) confirmed these results and showed that such a preparation of G/F became tightly associated with cyc- membranes only when it had been stably activated either by fluoride or by Gpp(NH)p or GTPyS. GTP alone or in the presence of hormone did not promote binding. Sternweis and Gilman (1978) showed that this is not the case when G/F is solubilized in cholate. Cholatesolubilized G/F bound to the membrane after dilution of detergent. Binding required warming and was promoted by the addition of GTP alone, but did not require stable activation of G/F. These data, data on the solubilization of G/F (Ross et al., 1978), and data on the protease susceptibility of G/F (Hudson et al., 1981) have suggested that G/F may interact with membranes through a small hydrophobic domain and may not be a typical, globular, integral membrane protein. As such, it may interact with receptor or C without binding to the lipid bilayer. Recent studies in our laboratory have pointed out a probable role for phospholipids in the stabilization of the C-GIF complex. Rabbit hepatic C, prepared by gel filtration in cholate solution (Ross, 1981), is essentially free of endogenous phospholipids. Using this preparation, we showed that phosphatidylcholine markedly potentiates the ability of activated G/F to stimulate the catalytic activity of C (Ross, 1982). A lipid-free mixture of C and G/F is only slightly stimulated

0

0.4

0.8

1.2

1.4

[DMPC] (mM) FIG. 8. Effect of lecithin on the interaction of C and G/F. Lipid-free preparations cf hepatic C and excess purified hepatic GIF were combined and diluted in the presence of increasing amounts of dimyristoylphosphatidylcholine (DMPC). Adenylate cyclase activity was measured either with Mn2+ (0) or with GTPyS plus Mg2+ (0). Specific activities shown are relative to the amount of protein in the preparation of C. (See Ross, 1982. for experimental details.)


131

by GTPyS (Fig. 8) or by fluoride, and attains a maximal activity less than that elicited by Mn2+ alone. The addition of an optimal concentration of dimyristoylphosphatidy Icholine increases G/F-mediated stimulation by fluoride or GTPyS as much as %fold, to yield the 10- to 20-fold stimulation over basal activity that is typical of native membranes. In contrast, phosphatidylcholine has little effect on activity stimulated by Mn2+ or by forskolin, neither of which acts via GIF. The small increase in Mn2 -stimulated activity that is observed is due primarily to stabilization of C by the lipid during the assay. Phosphatidylcholine also does not potentiate the initial activation of G/F by GTPyS or by fluoride. Thus, by elimination, these results are interpreted as indicating that phosphatidylcholine promotes or is required for the productive interaction of activated G/F with C. Various synthetic and natural phosphatidylcholines display this stimulatory effect, whereas detergents (cholate, digitonin, Lubrol 12A9, lysophosphatidylcholine), cholesterol, and several other phospholipids do not (Ross, 1982). The mechanism whereby phosphatidylcholine acts in this system is unclear, as is its possible physiological significance. These results do, however, point to the sensitivity of the interaction of C and G/F to the nature of their hydrophobic environment. Studies on the role of phospholipids on the interaction of C and G/F have also suggested a novel experimental system in which to pursue these effects. After the addition of phosphatidylcholine to cholate-solubilized C, cholate can be removed by dialysis and centrifugation to yield a preparation of large, unilamellar phospholipid vesicles to which C is bound. In a second stage of reconstitution, G/F can be added to these C vesicles in varying amounts to yield a vesicle-bound adenylate cyclase that can be regulated by fluoride or guanine nucleotides (Ross, 1982). The extent of stimulation by fluoride or by GTPyS is related to the amount of G/F added to the vesicles, and displays saturation with respect to G/F at appropriate concentrations. This system, in which the concentration of C and G/F and their molar ratios can be varied in a membrane of known lipid composition, promises to be useful for further mechanistic studies of the interaction of the two proteins. +

C. Regulation of the Activation of G/F by Receptor and Hormone Several basic properties of hormonal stimulation of adenylate cyclase indicate that the interaction of receptor with G/F is the key to the action of hormone in this system. First, hormonal stimulation of adenylate cyclase activity requires the presence of a guanine nucleotide, as first proposed by Rodbell and co-workers (197 I ) . If the nucleotide is GTP, hormonal activation is rapidly reversible and the steady-state level of adenylate cyclase activity reflects the number of recep-

132


tors and their fractional saturation by hormone. If a p r l y hydrolyzable analog of GTP is present or if G/F has been treated with cholera toxin, activation is slowly reversible or irreversible, and the hormone acts merely to increase the rate at which enzyme is activated (see Maguire er al., 1977; or Ross and Gilman, 1980, for review). More detailed kinetic studies, primarily in the laboratories of Selinger, Levitzki, and Birnbaumer, have led to the notion that the receptor-hormone complex essentially acts catalytically to increase the rate at which adenylate cyclase, or G/F, is activated. The rate of deactivation reflects the cellular source of the G/F and the identity of the guanine nucleotide, but is generally not regulated by hormone (reviewed in Ross and Gilman, 1980; see also Tolkovsky, this volume). Thus, in the presence of GTP, the steady-state level of hormone-stimulated activity reflects the balance of a hormone-catalyzed activation rate and a tonic rate of deactivation. Under special circumstances, the liganded receptor may also catalyze deactivation (Cassel and Selinger, 1977a; Sevilla and Levitzki, 1977; Arad el a / . , 1981). but this is probably not of physiological significance. The nature of the rate-limiting step in the activation of adenylate cyclase by hormone is uncertain. The simplest model for the mechanism of hormone action holds that ( I ) G/F is a GTPase, (2) only the GTPliganded form of G/F is active, and (3) hormone merely facilitates the displacement of GDP, the GTPase product, so that a new molecule of GTP can bind (Cassel et a f . , 1977; review by Ross and Gilman, 1980). More commonly, the conformational change of GTP-liganded G/F from inactive to active states may be rate limiting. Regardless, the physiological stimulation by hormone derives from the ability of H.R to catalyze the activation. In numerous systems, it is the formation of the H - R G / F catalytic intermediate that is the key step in activation by hormone. The questions to be asked now concern the detailed mechanism of the formation of this complex and the molecular nature of the catalytic event. The productive interaction of G/F with liganded receptor and, hence, the hormonal stimulation of adenylate cyclase require that both G/F and receptor be incorporated into a suitably unperturbed lipid bilayer. Furthermore, the chemical composition of the bilayer and its physical structural organization can have profound effects upon the efficiency and extent of hormonal control (for review see Ross and Gilman, 1980; Henis et a f . , 1982; Houslay and Gordon, this volume). Studies of native biological membranes are inherently limited, however, because one cannot independently or specifically alter the concentration of individual proteins or the phospholipid composition of the bilayer. It is also difficult to identify the details of the interaction of receptor and G/F in the presence of C. We therefore have attempted to develop reconstituted experimental systems in which receptor and G/F can interact in a bilayer of known composition and structure. Two such systems have yielded results so far. As discussed above, Fleming and Ross (1980) succeeded in introducing digitonin-solubilized P-adrenergic receptors into phospholipid vesicles by a two-step


133

procedure involving gel filtration and sucrose density gradient centrifugation. Reconstitution into vesicles composed primarily of dimyristoylphosphatidylcholine restored both the stability and the ligand-binding capabilities that are characteristic of the native, membrane-bound receptor, and should provide a population of receptors that are capable of catalyzing the activation of G/F. A second reconstituted system has now been exploited further in our laboratory to study the receptor-G/F interaction. The procedure is an extensive modification of that developed by Citri and Schramm (1980), in which P-adrenergic receptors are first solubilized using deoxycholate, essentially according to Eimerl et al. (1980). The ability to bind either [3H]DHA or [1251]IHYPis lost upon solubilization. Addition of dimyristoylphosphatidylcholine to this extract, followed by gel filtration, yields P-adrenergic binding activity eluted in a turbid fraction at the void volume. Of the binding activity originally in the membranes, greater than 60% is recovered in these fractions, while more than 99% of the deoxycholate is removed. Receptors in this preparation bind both [ *251]IHYPand [3H]DHA with a specific activity greater than I pmole mg - I protein and display appropriate P-adrenergic specificity for unlabeled agonist and antagonist ligands in competitive binding assays. While this preparation has not yet been extensively characterized either hydrodynamically or by electron microscopy, the binding activity is in a particulate fraction containing about 3 mg phospholipid/ mg protein that has displayed, in preliminary experiments, properties expected of vesicles in the 500-2000 A size range. Purified rabbit hepatic G/F added to these P-adrenergic receptor vesicles in a second reconstitution step can be stably activated by GTPyS in a time-dependent reaction. (Activated G/F is subsequently assayed by the addition of Lubrol 12A9 and reconstitution into cyc membranes.) At relatively low concentrations of Mg2+, the rate of activation of G/F in these vesicles is low, as predicted by studies of soluble G/F (Sternweis et al. 1981). However, the addition of (-)isoproterenol stimulates the rate of activation of G/F up to 4-fold (Fig. 9). These results do not compare unfavorably with those of Citri and Schramm (1980), who reported up to 12-fold stimulation in the presence of receptor and hormone. These authors, however, used crude turkey erythrocyte G/F, which has a far lower basal rate of activation than does the rabbit hepatic protein (compare Hanski et al., 1981; and Sternweis ef al., 1981). The increase in rate caused by isoproterenol displays characteristic P-adrenergic chemical and stereochemical selectivity both for stimulation by agonists (Fig. 10) and blockade by antagonists, and is also abolished if P-adrenergic binding sites are destroyed by N,N'dicyclohexylcarbodiimide. This latter procedure does not inactivate G/F. This system thus appears to represent the restoration of the hormone-specific, catalytic interaction of liganded receptors with nucleotide-liganded G/F. Our reconstitution of P-adrenergic stimulation of the activation of G/F is still in its preliminary stages of development. First, only G/F and phospholipid are ~

134


). I

I

0

2

4

6

8

1015

Time (min)

FIG. 9. Activation of GIF by GTPyS is catalyzed by P-adrenergic receptors in reconstituted phospholipid vesicles. P-Adrenergic receptors were solubilized from turkey erythrocyte plasma membranes essentially according to Eimerl et a / . ( 1980), mixed with dimyristoylphosphatidylcholine, and incorporated into phospholipid vesicles by gel filtration in the absence of detergent. Concentrated vesicles were mixed with purified rabbit hepatic G/F (Sternweis et al., 1981)such that the final concentration of Lubrol (which was added with the G/F) war about 20 pg/ml. This mixture was incubated at 3OoC in the presence of 10 @f GTPyS and I @f (-)isoproterenol ( 0 )or of GTPyS, isoproterenol, and 10 p M (-)propranolol The activation reaction was quenched at the indicated times by diluting aliquots of these mixtures with buffer containing 1 0 - 5 M (-)propranolol, 5 mM EDTA, and 0.1 mM GDPPS. The samples were then assayed for activated G/F essentially as described by Sternweis er al. (1981). Each G/F assay tube contained a volume of activation reaction mixture equivalent to 0.46 fmole of P-adrenergic receptors, as assayed using [ '2sI]IHYP. Thus, if the reconstitutive activity of activated G/F is about lo4 U/mg = 160 Ulnmole (Sternweis et ot., 1981), the hormone-catalyzed activation of 1.9 X 1 0 - 3 Uiminute (2.45 x total - 0.55 x 1 0 - 3 basal) represents a turnover rate of more than 5 moles G/F/rnole receptodminute over the first I minute of the activation reaction.

(c).

added as pure components. A purer preparation of receptors must be used in order to accommodate a wider range of receptor:G/F molar ratios in the vesicles and to provide the general security of working with pure proteins. The kinetics of the reaction must also be studied in greater detail. The activation of G/F by receptor-hormone complex is characterized by a burst that is nearly complete in 1 minute. This behavior is not understood, but does not simply reflect denaturation of G/F or receptors during the activation reaction. It is also significant that while the hormone-stimulated rate falls after the burst, it remains higher than the basal rate as both approach linearity. Hypothetically, this nonlinear time course may reflect a rate-limiting exchange of G/F among vesicles. It might also represent the reconstitution of a rapid down-regulation of receptor-G/F coupling (Su et af., 1980). It should be pointed out, however, that initial rates do indicate a catalytic event-we calculate that the initial rate of activation reflects the activation of more than five molecules of G/F per minute per receptor molecule in the experiment shown here, and higher rates have been observed. This calculation does not make allowances for the possible segregation or occlusion of some receptors, so it is probably an underestimate of the catalytic capacity of the active, vesicle-bound receptor molecules. Another estimation of the efficiency of

135


A L

0

10-'0

10-0

10-0

10-4

ISOPROTERENOL (M)

2.5

2 .o 1.5 1.o

0.5

t

04.

I

I

I

I

B

Fic. 10. P-Adrenergic specificity of reconstituted, hormone-stimulated activation of C/F. GIF was activated in 0-adrenergic receptor-containing vesicles as described in the legend to Fig. 9, but in the presence of increasing concentrations of each stereoisomer of isoproterenol (A) or in the presence of I phf (-)isoproterenol and increasing concentrations of each stereoisomer of propranolol (B). In each figure, the open symbol is the ( + ) isomer. In (A), the activation was carried out for I minute. In (B). the activation was carried out for 5 minutes, causing the relatively high basal rate at saturating concentrations of propranolol. Also shown in (A) is the effect of added 5 phf (-)propranolol at two concentrations of (-)isoproterenol

(A).

the reconstitution is the comparison of the hormone-stimulated activation rate with the rate stimulated by high concentrations of M g 2 + . This parameter, evaluated at 1 minute (Table II), suggests a process about 30-40% efficient, but this fraction falls to 15% at longer activation times. Again, this parameter does not account for the possible spatial segregation of some GIF and receptor molecules, since all G/F is presumably accessible to Mg2 , but probably not to receptors. + +

136

ELLIOT M. ROSS ET AL.

TABLE I1 ACTIVATIONOF G/F AI-TITERRECONSTITUTION INTO PHOSPHOLIPID VESICLES CONTAINING P-ADRENERGIC RECEFTORS"

Additions

Rate of activation of G/F (units rnin-1 x 103)

None ( - )Isoproterenol ( 10- M ) ( - )Propranolol (10 - 5 M ) lsoproterenol plus propranolol M) Terbutaline Terbutaline plus propranolol Phentolamine ( 1 0 - 5 M ) Isoproterenol plus phentolamine MgClz (0.05 M )

0.7 2.9 0.7 0.8 2.1 0.7 0.8 2.7 8.7

" Preparation of vesicles and the addition of purified G/F were performed as described in the legend to Fig. 9. Activation was carried out for I minute at 30°C in the presence of the indicated compounds.

Thus, this level of efficiency is not a bad place to begin. A more fundamental test of the efficiency of receptor-G/F coupling in a reconstituted system may be the negatively cooperative binding interactions of guanine nucleotides and agonist ligands (see Maguire et a l . , 1977; or Ross and Gilman, 1980, for a review). The extent to which the addition of GIF increases the affinity with which agonists bind to the receptor and the ability of GTP to block this interaction should provide at least a thermodynamic estimate of the association of the two proteins. In preliminary experiments, we were able to observe a small but definite increase in the affinity with which isoproterenol bound to receptor vesicles when purified G/F was present. We thus foresee multiple ways in which we can study the coupling of the two proteins after reconstitution. The methods described here should ultimately allow independent and welldefined variation of lipid composition, j3-receptor concentration, and G/F concentration during reconstitution. The concentrations of receptors and of G/F can now be varied only within limited ranges. The concentration of receptors within a vesicle is constrained by currently low specific activity in our preparations. Consequently, we are attempting to apply the reconstitutive procedure to purer preparations of receptors. Manipulation of G/F is limited only by the detergent that is introduced into the system with the purified protein, a problem that may be circumvented by the use of more concentrated preparations or a detergent removal step. Those developments, along with physical studies of purified, functional j3-adrenergic receptors, should allow the detailed molecular study of the receptor-G/F interaction.


137

IV. ASSESSMENT OF PROGRESS The last 5 years of research on hormone-sensitive adenylate cyclase have focused on the questions of how many proteins are involved in an observed activity or regulatory events, what the role of each protein is, and the sequence in which individual events occur. Identifying proteins has been a slow process, but is a prerequisite to obtaining the proteins in a pure form, suitable for study. We are now at the point at which the individual protein components must be analyzed in detail by essentially classic physical and enzymological approaches, and we can begin to construct reconstituted model systems in which to test our hypotheses. Many of the outstanding questions are obvious. Direct information on the conformational alteration of a receptor by a hormone requires the study of workable quantities of purified receptors. Questions of how receptors catalyze the activation of G/F demand both kinetic and structural approaches, particularly with regard to the effects of the membrane bilayer on the ability of the two proteins to interact. Whether C is one or several proteins awaits better purification, and the study of the regulation of C is virtually dependent on purer preparations. Ideally, these biochemical problems can be pursued in parallel with the questions of long-term endocrinologic regulation of these proteins and of the control and coordination of their synthesis. ACKNOWLEDGMENTS Studies from this laboratory cited here have been supported by USPHS grant GM30355. E. M. R. is an Established Investigator of the American Heart Association and V. A. F. is the recipient of a graduate fellowship from the Pharmaceutical Manufacturers‘ Association Foundation. REFERENCES Arad, H., Rimon. G . , and Levitzki, A. (1981). The reversal of the Gpp(NH)p-activated state of adenylate cyclase by GTP and hormone is by the ”collision coupling” mechanism. J. B i d . Chem. 256, 1593-1597. Atlas, D., and Levitzki, A. (1978). Tentative identification of P-adrenoceptor subunits. Nature (London) 272, 370-371. Bourne, H. R., Coffino, P., and Tomkins. G. M. (1975). Selection of a variant lymphoma cell deficient in adenylate cyclase. Science 187, 750-752. Braun, S., and Levitzki, A. (1979). Adenosine receptor permanently coupled to turkey erythrocyte adenylate cyclase. Biochemistry 18, 2134-2138. Caron, M. G., and Lefkowitz, R. J . (1976). Solubilization and characterization of the P-adrenergic receptor binding sites of frog erythrocytes. J . Biol. Chem. 251, 2374-2384. Cassel, D., and Pfeuffer, T. (1978). Mechanism of cholera toxin action: Covalent modification of the guanyl nucleotide-binding protein of adenylate cyclase. Proc. Natl. Acad. Sci. U.S.A. 75, 2669- 2673. Cassel, D., and Selinger, 2. (1976). Catecholamine-stimulated GTPase activity in turkey erythrocytes. Biochrm. Biophys. Acta 452, 538-55 I . Cassel, D., and Selinger, Z. ( l977a). Catecholamine-induced release of [‘H]Gpp(NH)p from turkey erythrocyte adcnylate cyclase. J . Cvclic Nucleotide Res. 3, 11-22.

138


Cassel. D., and Selinger, Z. (1977b). Mechanism of adenylate cyclase activation by cholera toxin: Inhibition of GTP hydrolysis at the regulatory site. Proc. Narl. Acad. Sci. U.S.A. 74, 3307-33 1 1 . Cassel. D., Levkovitz. H.. and Selinger. Z . (1977). The regulatory GTPase cycle of turkey erythrocyte adenylate cyclase. J. Cyclic Nuclrotide Res. 3 , 393-406. Cheung. W. Y . . Lynch, T. J., and Wallace, R. W. (1978). An endogenous Caz+-dependent activator protein of brain adenylate cyclase and cyclic nucleotide phosphodiesterase. Adv. Cyclic Nucleotide Res. 9, 233-251. Citri. Y . , and Schramm, M. (1980). Resolution, reconstitution, and kinetics of the primary action of a hormone receptor. Nature (London) 287, 297-300. Clarke, S . (!975). The size and detergent binding of membrane proteins in detergent solution. J . Biol. Chem. 250, 5459-5469. Durieu-Trautmann. O., Delavier-Klutchko, C.. Vauquelin, G . , and Strosberg. A. D. (1980). Visualization of the turkey erythrocyte P-adrenergic receptor. J . Suprumol. Strucf. 13, 41 1-419. Eirnerl. S., Neufeld. G., Korner. M., and Schramm. M. (1980). Functional implantation of a solubilized P-adrenergic receptor in the membrane of a cell. Proc. Natl. Acad. Sci. U . S . A . 77, 760-764. Fleming, J. W., and Ross, E. M. (1980). Reconstitution of beta-adrenergic receptors into phosbinding to digitonin-solpholipid vesicles: Restoration of [ ~~~I]iodohydroxybenzylpindolol ubilized receptors. J. Cyclic Nudeotide Res. 6, 407-4 19. Florio, V. A,. and Ross, E. M. (1982a). Direct inhibition of the catalytic protein of adenylate cyclase at the adenosine P site. Fed. Proc. Fed. Am. Soc. Erp. Eiol. 41, 1408. Florio, V. A , , and Ross, E. M. (l982b). Regulation of the catalytic component of adenylate cyclase: Cooperative interaction of stimulatory ligands and ribose-modified adenosine analogs. Mol. Pharmacol., in press. Gill. D. M., and Meren, R. (1978). ADP-ribosylation of membrane proteins catalysed by cholera toxin: Basis of the activation of adenylate cyclase. Proc. Nut/. Acud. Sci. U . S . A . 75, 30503054. Haga, T., Haga, K.. and Gilman. A. G. (1977a). Hydrodynamic properties of the P-adrenergic receptor and adenylate cyclase from wild type and variant S49 lymphoma cells. J. Biol. Chem. 252, 5776-5782. Haga, T., Ross, E. M., Anderson, H. J . , and Gilman, A. G. (1977b). Adenylate cyclase permanently uncoupled from hormone receptors in a novel variant of S49 mouse lymphoma cells. Proc. Natl. Acad. Sci. U . S . A . 74, 2016-2020. Hanski, E., Sternweis, P. C., Northup, J . K.. Dromerick, A. W., and Gilman, A. G. (1981). The regulatory component of adenylate cyclase. Purification and properties of the turkey erythrocyte protein. J. BkJl. Chem. 256, 1291 1-12919. Hazeki, 0.. and Ui. M. (1981). Modification by islet-activating protein of receptor-mediated regulation of cyclic AMP accumulation in isolated rat heart cells. J. Biol. Chrm. 256, 2856-2862. Henis, Y . I . , Rimon, G., and Felder, S. (1982). Lateral mobility of phospholipids in turkey erythrocytes. Implications for adenylate cyclase activation. J. B i d . Chem. 257, 1407- 141 I. Howlett, A. C., and Gilman, A. 0. (1980). Hydrodynamic properties of the regulatory component of adenylate cyclase. J. Biol. C h r m . 255, 2861-2866. Howlett, A. C., Sternweis, P. C., Macik. B. A,, Van Arsdale, P. M . , and Gilman, A. G . (1979). Reconstitution of catecholamine-sensitive adenylate cyclase. Association of a regulatory component of the enzyme with membranes containing the catalytic protein and P-adrenergic receptors. J. Biol. Chem. 254, 2287-2295. Hudson, T. H., and Johnson, G. L. (1980). Peptide mapping of adenylate cyclase regulatory proteins that are cholera toxin substrates. J. B i d . Chem. 255, 7480-7486. Hudson, T. H., Roeber, J. F.. and Johnson, G. L. (1981). Conformational changes of adenylate cyclase regulatory proteins mediated by guanine nucleotides. J . B i d . Chem. 256, 1459- 1465.


139

Insel. P. A., and Kennedy, M. S. ( 1978). Colchicine potentiates P-adrenoceptor-stimulated cyclic AMP in lymphoma cells by an action distal to the receptor. Nature (London) 273, 471473. Insel, P. A.. Maguire, M. E.. Gilman, A. G., Bourne, H. R.. Coffino. P., and Melmon. K . L. ( 1976). Beta adrenergic receptors and adenylate cyclase: Products of different genes'? Mol. Pharmacol. 12. 1062- 1069. Ji. I . , and Ji. T. H. (1980). Macromolecular photoaffinity labeling of the lutropin receptor on granulosa cells. Proc. Narl. Acad. Sci. U.S.A. 77, 7167-7170. Johnson, G . L., MacAndrew, V. I.. and Pilch, P. F. (1981 1. Identification of the glucagon receptor in rat liver membranes by photoaffinity cross-linking. Pruc. Narl. Acud. Sci. U.S.A. 78, 875-878. Kaslow. H . R.. Farfel, Z . , Johnson, G. L.. and Bourne, H. R. (1979). Adenylate cyclase assenibled in virrot Cholera toxin substrates determine different patterns of regulation by isoproterenol and guanosine S'-triphosphdte. Mol. Pharmacol. 15, 472-483. Kaslow, H. R., Johnson. G . L., Brothers. V. M., and Bourne, H. R. ( 1980). A regulatory component of adenylate cyclase from human erythrocyte membranes. J . B i d . Chem. 255, 3736-3741. Katada. T.. and Ui. M. (1981). Islet activating protein. J . Biol. Chem. 256, 8310-8316. Katada. T.. and Ui, M. (1982). Direct modification of the membrane adenylate cyclase system by islet-activating protein due to ADP-rihosylation of a meinbrane protein. Proc. Nut/. Acad. Sci. U.S.A. 79, 3129-3133. Lad. P. M.. Welton. A. F.. and Rodbell, M. (1977). Evidence fordistinct guanine nucleotide sites in the regulation of the glucagon receptor and of adenylate cyclase activity. 1. Biol. Chem. 252, 5942-5946. Lamer, A. C.. and Ross, E. M. (1981). Alteration in the protein components of catecholaminesensitive adenylate cyclase during maturation of rat reticulocytes. J . Biol. Chcm. 256, 9551-9557. Lavin. T. N., Heald. S. L.. Jeffs. P. W . . Shorr, R . G . L.. Letkowitz, R. J.. and Caron, M. G . (19x1). Photoaffinity labeling of the P-adrenergic receptor. J. B i d . Chrtn. 256, 11944-1 1950. Limbird. L. E. (1981). Activation and attenuation of adenylate cyclase: GTP-binding proteins as molecular messengers in receptor-cyclase coupling. Biorhrm. J. 195, I- 13. Limbird. L. E.. and Letkowitz, R. J. (1977). Resolution of P-adrenergic receptor binding and adenylate cyclase activity by gel exclusion chromatography. J. B i d . Chem. 252, 799-801. Limbird. L. E.. McMillan, S . T.. and Smith. S. K. (1981). Solubilization of human platelet aadrenergic receptors: Evidence that agonist occupancy of the receptors stabilizes receptoreffector interactions. Proc. Narl. Acad. Sei. U.S.A. 78, 4026-4030. Lindner. E.. Dohadwalla, A. N.. and Bhdttdcharya, B. K . (1978). Positive inotropic and blood pressure lowering activity of a diterpene derivative isolated from Colrit.s,~orsX.ohlii:Forskolin. Armeim. Forsch. 28, 284-289. Londos, C.. and Wolff. J. ( 1977). Two distinct adenosine-sensitive sites on adenylate cyclase. Pro(,. Nutl. Acad. Sci. U.S.A. 74, 5482-5486. Londos. C., Wolff. J.. and Cooper, D. M. F. (1979). Action of adenosine on adenylate cyclase. In "Physiological and Regulatory Functions of Adenosine and Adenine Nucleotides" (H. P. Baer and 0.I . Drummond. eds.), pp. 271-281. Raven. New York. Maguire. M. E.. Ross, E. M . . and Cilman. A.G. (1977). 0-Adrenergic receptor: Ligand binding properties and the interaction with adenylyl cyclase. Adv. Cyclic Nircleotide Res. 8, 1-83. Murad. F.. Chi. Y . M., Rall, T. W.. and Sutherland. E. W. (1962). The effect of catecholamines and choline esters on the formation of adenosine 3',5'-phosphate by preparations from cardiac muscle and liver. J . B i d . Chem. 237, 1233-1238. Naya-Vigne. J . , Johnson, G . L., Bourne. H. R.. and Coffino. P. (1978). Complementation analysis of hormone-sensitive adenylate cyclase. Nature (London) 272, 720-722.

140

E L L I O T M. ROSS ET AL.

Neer. E. J. (1979). Interaction of soluble brain adenylate cyclase with manganese. J . Eiol. Chem. 254, 2089-2096. Neer. E. J . , and Salter, R. S. (1981). Reconstituted adenylate cyclase from bovine brain. J . Biol. Chem. 256, 12102-12107. Northup. J . K.. Sternweis, P. C., Smigel, M. D.. Schleifer, L. S., Ross, E. M.. and Gilman, A. G . (1980). Purification of the regulatory component of adenylate cyclase. Proc. Nut/. Acud. Sci. U.S.A. 77, 6516-6520. Pedersen. S. E., and Ross. E. M. ( 1982a). Functional reconstitution of P-adrenergic receptors and the guanine nucleotide-binding regulatory protein of adenylate cyclase. Fed. Proc. Fed. Am. Sor. Exp. B i d . 41, 141I . Pedersen. S. E.. and Ross. E. M. (1982b). Functional reconstitution of 0-adrenergic receptors and the stimulatory GTP-binding protein of adenylate cyclase. Proc. Nut/. Acrid. Sci. U.S.A. In press. Perkins. J . P. (1973). Adenyl cyclase. Adv. Cyclic Nucleotide Res. 3, 1-64. Pfeuffer, T . (1977). GTP-binding proteins in membranes and the control of adenylate cyclase activity. J . Biol. C h m . 252, 7224-7234. Pfeuffer, T . ( 1979). Guanine nucleotide controlled interactions between components of adenylate cyclase. FEES Lett. 101, 85-89. Premont. J . , Cuillon. G.. and Bockaert. J . (1979). Specific Mg’+ and adenosine sites involved in a bireactant mechanism for adenylate cyclase inhibition and their probable localization on this enzyme’s catalytic component. Eiochein. Eiopphys. Res. Coinmun. 90, 5 13-5 19. Rasenick, M. M.. Stein. P. I . , and Bitensky, M. W. (1981). The regulatory subunit of adenylate cyclase interacts with cytoskeletal components. Nature (London)294, 560-562. Rashidbaigi, A.. and Ruoho, A. E. (1981). Iodoazidopindolol. a photoaffinity probe for the P-adrenergic receptor. Proc. Nut/. Acad. Sci. U.S.A. 78, 1609-1613. Rebois, R. V . , Omedeo-Sale, F.. Brady. R. 0..and Fishman. P. F. (1981). Covalent crosslinking of human chorionic gonadotropin to its receptor in rat tcstes. Proe. NmI. Acad. Sri. U.S.A. 78, 2086-2089. Rodbell. M. (1972). Regulation of glucagon action at its receptor. I n “Clucagon: Molecular Physiology, Clinical Therapeutic Implications” (P. S. Lefebvre and R. H. Unger, eds.), pp. 61-75. Pergamon, Oxford. Rodbell. M., Birnbaumer, L., Pohl, S. L., and Krans, H. M. J . (1971). The glucagon-sensitive adenyl cyclase system in plasma membranes of rat liver. V. An obligatory role of guanyl nucleotides in glucagon action. J . Eiol. Chein. 246, 1877-1882. Ross, E. M. (1981). Physical separation of the catalytic and regulatory proteins of hepatic adenylate cyclase. J. Eiol. Chein. 256, 1949-1953. Ross, E. M. ( 1982). Phosphatidylcholine-promoted interaction of the catalytic and guanine nucleotide-binding proteins of adenylate cyclase. J . B i d . Chem., in press. Ross, E. M.. and Gilman. A. G. (l977a). Reconstitution of catecholamine-sensitive adenylate cyclase activity: Interaction of solubilized components with receptor-replete membranes. Proc. Nut/. Arad. Sci. U.S.A. 14, 3715-3719. Ross. E. M.. and Gilman. A. G. (1977b). Resolution of some components of adenylate cyclase necessary for catalytic activity. J . Eiol. Chem 252, 6966-6970. Ross. E. M., and Gilman. A. G. (1980). Biochemical properties of hormone-sensitive adenylate cyclase. Anriu. Rev. Biot,hem. 49, 533-564. Ross, E. M., Maguire, M. E., Sturgill, T. W . . Biltonen. R. L.. and Gilman, A. G. (1977). Relationship between the P-adrenergic receptor and adenylate cyclase. Studies of ligand binding and enzyme activity in purified membranes of S49 lymphoma cells. J . B i d . Chem. 252, 5761-5775. Ross, E. M.. Howlett, A. C . , Ferguson, K. M., and Gilman. A. G . (1978). Reconstitution of


141

hormone-sensitive adenylate cyclase activity with resolved components of the enyzme. J . Biol. Chem. 253, 6401-6412. Rudolph, S . A.. Greengard, P., and Malawista, S . E. (1977). Effects of colchicine on cyclic AMP levels in human leukocytes. Proc. Natl. Acad. Sci. U . S . A . 74, 3404-3408. Sahyoun. N., Hollenberg. M. D.. Bennet. V . . and Cuatrecasas. P. (1977). Topographic separation of adenylate cyclase and hormone receptors in the plasma membrane of toad erythrocyte ghosts. Proc.. Natl. Acad. Sci. U.S.A. 74, 2860-2864. Sahyoun. N. E.. LeVine, H., 111, Hebdon, G. M., Hemadah, R., and Cuatrecasas, P. (1981). Specific binding of solubilized adenylate cyclase to the erythrocyte cytoskeleton. Proc. Nail. A m d . Sci. U . S . A . 78, 2359-2362. Salter. R. S . . Krinks. M . H.. Klee, C. B.. and Neer, E. J. (1981). Calmodulin activates the catalytic unit of brain adenylate cyclase. J. Biol. Chem. 256, 9830-9833. Sano. M . . and Drummond, G. I. (1981). Properties of detergent-dispersed adenylate cyclase from cerebral cortex. Presence of an inhibitor protein. J . Neurochem. 37, 558-566. Schleifer. L. S . , Garrison. J . C.. Sternweis, P. C . . Northup. J. K., and Gilman, A. G.(1980). The regulatory component of adenylate cyclase from uncoupled S49 lymphoma cells differs in charge from the wild type protein. J . Biol. Chem. 255, 2641-2644. Schleifer. L. S . . Kahn. R. A,, Hanski, E., Northup, J . K., Sternweis, P. C.. and Gilman, A. G. ( 1982). Requirements for cholera toxin-dependent ADP-ribosylation of the purified regulatory component of adenylate cyclase. J . Biol. Chem. 257, 20-23. Schwarzmeier, J. D.. and Gilman, A. G. ( 1977). Reconstitution of catecholamine-sensitive adenylate cyclase activity: Interaction of components following cell-cell and membrane-cell fusion. J . Cvclic Nideotide Res. 3, 227-238. Seamon. K., and Daly, J . W. (1981). Activation of adenylate cyclase by the diterpene forskolin does not require the guanine nucleotide regulatory protein. J . Biol. Chem. 256, 9799-9801. Seamon. K. B., Padgett. W., and Daly, J . W. (1981). Forskolin: Unique diterpene activator of adenylate cyclase in membranes and in intact cells. Proc. Natl. Acad. Sci. U.S.A. 78, 3363-3367. Sevilla. N., and Levitzki. A. (1977). The activation of adenylate cyclase by 1-epinephrine and guanylimidodiphosphate and its reversal by I-epinephrine and GTP. FEBS Lett. 76, 129- 134. Shorn. R. G . L.. Lefkowitz, R. J . , and Caron. M. G. (1981). Purification of the P-adrenergic receptor. Identification of the hormone-binding subunit. J . Biol. Chem. 256, 5820-5826. Sniigel. M. D., Northup. J . K., and Gilman, A. G . (1982). Characteristicsoftheguanine nucleotidebinding regulatory component of adenylate cyclase. Recent Prog. Horm. R e x . 38, 601-622. Skmweis, P. C.. and Gilman, A. G. ( 1979). Reconstitution of catecholamine-sensitive adenylate cyclase. Reconstitution of the uncoupled variani of the S49 lymphoma cell. J. Biol. Chem. 254, 3333-3340. Sternweis, P. C . , and Gilman, A. G . (1982). Aluminum: A requirement for activation of the regulatory component of adenylate cyclase by fluoride. Proc. Natl. Acad. Sci. U.S.A. 79, 4888-489 I , Sternweis, P. C . . Northup. J . K . . Smigel, M. D., and Gilman, A. G . (1981). The regulatory component of adenylate cyclase. Purification and properties. J . Biol. Chem. 256, 1151711526. Strittmatter, S . , and Neer. E. J . (1980). Properties of the separated catalytic and regulatory units of brain adenylate cyclase. Proc. Natl. Acad. Sci. U.S.A. 77, 6344-6348. Su. Y . - S . , Harden. T. K . , and Perkins, J . P. (1980). Catecholamine-specific desensitization of adenylate cyclase. Evidence for a multistep process. J . Biol. Chem. 255, 7410-7419. Tolkovsky. A. M . . and Levitzki, A. (1981). Theories and predictions of models describing sequential interactions between the receptor, the GTP regulatory unit. and the catalytic unit of hormone dependent adenylate cyclase. J . Cvclic Nucleotide Res. 7, 139-150.

142


Tolkovsky, A. M.. Braun, S., and Levitzki, A. (1982). Kinetics of interaction between P-receptors, GTP protein. and the catalytic unit of turkey erythrocyte adenylate cyclase. Proc. Narl. Acad. Sci. U.S.A. 79, 213-217. Toscano, W. A,, Jr., Westcott, K. R., LaPorte, D. C.. and Storm, D.R. (1979). Evidence for a dissociable protein subunit required for calmodulin stimulation of brain adenylate cyclase. Proc. Nafl. Acad. Sci. V.S.A. 76, 5582-5586. Vauquelin, G.,Geynet, P., Hanoune, J . , and Strosberg, A. D. (1979). Affinity chromatography of the P-adrenergic receptor from turkey erythrocytes. Eur. J. Biochem. 98, 543-556. Weber, G . (1975). Energetics of ligand binding to proteins. Adv. Protein Chem. 29, 1-83. Welton, A. F., Lad, P. M., Newby, A. C., Yamamura, H., Nicosia, S., and Rodbell, M. (1977). Solubilization and separation of the glucagon receptor and adenylate cyclase in guanine nucleotide-sensitive states. J. Biol. Chem. 252, 5947-5950. Westcott, K. R., LaPorte, D. C., and Storm, D. R. (1979). Resolution of adenylate cyclase sensitive and insensitive to Ca2+ and calcium-dependent regulatory protein (CDR) by CDR-Sepharose affinity chromatography. Proc. Nafl. Acad. Sci. U.S.A. 76, 204-208. Witkin, K. M., and Harden, T. K. (1981). A sensitive equilibrium binding assay for soluble P-adrenergic receptors. J. Cyclic Nucleotide Res. 7 , 235-246. Wolff, D. J., and Brostrom, C. 0. (1979). Properties and functions of the calcium-dependent regulatory protein. Adv. Cyclic Nucleotide Res. 11, 27-88. Wolff, J., Londos, C., and Cook, G. H. (1978). Adenosine interactions with thyroid adenylate cyclase. Arch. Biochem. Eiophys. 191, 161-168.

CURRENT TOPICS IN MEMBRANES AND TRANSPORT. VOLUME I8

The Regulation of Adenylate Cyclase by Glycoprotein Hormones BRIAN A . COOKE Department of Biochemistry Roval Free Hospital School of Medicine University of’ London London. England

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nature of the Hormones., . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nature of the Receptors . ................... IV. Involvement of Cyclic AMP in Hormone Action.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . V . Important Features of the Hormone Receptor-Adenylate Cyclase System. . . . . . . . . . Effect of Guanine Nucleotides on Binding of the Hormone to Its Receptor. . . . . v1. Desensitization and Down-Regulation by Homologous Hormone . . . . . . . . . . . . . . . . . A. Desensitization and Down-Regulation of a LH-Responsive Leydig Cell Tumor . B. The Desensitizing Effect of GTP on Isolated Plasma Membranes. . . . . . . . . . . . . C. Determination of the Site of Lesion in LH-Desensitized Leydig Tumor Cells. . . D. Possible Mechanisms Involved in Desensitization . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I.

11. 111.

1.

143 144 149 151 152

I52 154 161 165 170 172

INTRODUCTION

Most mammalian cells controlled by hydrophilic hormones contain plasma membrane bound adenylate cyclase and specific hormone receptors. The target cells controlled by the glycoprotein hormones, lutropin (LH), follitropin (FSH), and thyrotropin (TSH), are no exception. What distinguishes these hormones from other hormones working through the same mechanism is their structure and binding characteristics of the hormone receptor interaction. They each contain two essential subunits with different but not well-defined functions. These hormones bind to their receptors in a manner which is not readily reversible and are unaffected by guanine nucleotides in terms of their binding characteristics. As 143 Copyright D 19x3 by Academic Press. Inc. All rights of reproduclion In any lbml reserved. ISBN 0-12-153318-2

144

BRIAN A. COOKE

pointed out by Abramowitz et (11. (1981) this appears to make the glycoprotein hormone receptors unique, because all other receptor systems that affect adenylate cyclase activity bind to hormones in a manner that is readily reversible and are affected by guanine nucleotides; in the presence of the latter the affinity of the nonglycoprotein hormone receptor interactions is lowered by as much as 10-fold. The glycoprotein hormone-adenylate cyclase systems have, in common with other systems, a protein (or proteins) which couple their receptors to the adenylate cyclase catalytic protein. This coupling protein contains a GDP/GTP binding site. The interaction of the hormone, receptor, and guanine nucleotide binding protein (referred to as G or N or GIF protein) to the adenylate cyclase results in the formation of cyclic AMP from ATP and requires the presence of Mg2+. In addition to the stimulatory effect on cell function it has been shown that the glycoproteins desensitize their target cells, resulting in a loss of response to further stimulation. Initially this involves a decrease in adenylate cyclase activity, which is followed by loss of hormone receptors (down-regulation). The mechanisms involved have not been fully elucidated, but probably an uncoupling of the hormone receptor complex from the G-protein-adenylate cyclase occurs followed by internalization of the hormone receptor. In some systems these processes have been found to depend on protein synthesis. The purpose of this article is to highlight some of the more important aspects of the recent advances in our understanding of the above-mentioned mechanisms. Because of the current state of knowledge and the interests of this author emphasis will be placed on the mechanism of action of one of the glycoproteins-lutropin-and its action on the Leydig cell. For more detailed reviews on other aspects of the regulation of receptors and adenylate cyclase the reader is referred to Abramowitz et ul. (1979), Jacobs ( 1979), Ross and Gilman (1980), Schulster and Livitzky (l980), Limbird (1981), Spiegel e t a l . (1981), and Cooke (1982).

II. NATURE OF THE HORMONES The glycoprotein hormones follitropin (follicle-stimulating hormone, FSH), lutropin (luteinizing hormone, LH), thyrotropin (thyroid-stimulating hormone, TSH), and choriotropin (human choriogonadotropin, hCG) are structurally related. Each molecule is composed of two nonidentical polypeptide chains (a and p). The carbohydrate moiety represents 7-30% of the total molecular mass. The carbohydrate residues are linked to the polypeptide chains through N-glycosidic linkages to either the amide group of asparagine or the hydroxyl group of serine or threonine residues. Qualititative and quantitative variations occur in the carbohydrate contents of the glycoproteins from different species. Sequence analysis of the amino acid residues has shown that the a-chains of

REGULATION BY GLYCOPROTEIN HORMONES

145

the glycoproteins are remarkably similar and the P-chains are quite different. Dissociation of the hormone into the a and P subunits can be achieved under nondenaturing conditions indicating that they are linked by noncovalent bonds. The dissociated units are biologically inactive. Reassociation at 4°C can be achieved with restoration of most of the original activity. It has been suggested that the p subunit of the glycoprotein hormone determines the biological specificity of the molecule, but that this can be achieved only in the presence of the (Y subunit. Specific gangliosides may be involved in the binding of the glycoproteins to their receptor sites (see review by Kohn, 1978). The role of the carbohydrate moiety in the glycoprotein hormones is not clear, although it has been shown that removal of the sialic acid residues from hCG makes this hormone more susceptible to metabolism in vivo but has little effect on its binding to testicular plasma membranes (Van Hall et a l . , 1971). In the female both FSH and LH are involved in the maturation of the ovarian follicle. FSH, together with estradiol- 17p, prepares the follicle for ovulation, and LH (and prolactin) is mainly concerned with the maintenance of the corpus luteum and steroid production (Richards, 1979). In the male, FSH specifically controls the activity of the Sertoli cells in the seminiferous tubules (see review, Dorrington and Armstrong, 1979), and LH the Leydig cells where the male androgen testosterone is formed (see review by Cooke et af., 1981b). hCG is secreted by the placenta and, unlike the other trophic hormones, is independent of the hypothalmus or pituitary gland. This hormone, together with human placental lactogen, is secreted in large amounts during the first few weeks of pregnancy and during this time may well play an important luteotrophic role. The roles of these hormones after this time, however, are still unclear. The biological properties of hCG are very similar to those of LH, but they differ in chemical structure and in their metabolic half-life (hCG has a much longer halflife than LH). Because of the stability of hCG and its similarity to LH and availability in a highly purified form, it is often used in place of LH in in vivo and in v i m studies on the testes and ovaries. TSH has been shown to be capable of stimulating almost every metabolic process in the thyroid. This includes iodide trapping, formation and release of the thyroid hormones, and general metabolic effects on glucose oxidation, ribonucleic acid synthesis, and phospholipid formation. 111.

NATURE OF THE RECEPTORS

A11 of the hormones listed in Table 1 have been shown to have plasma membrane receptors in their target cells with the characteristics necessary for fine control of their functions, i.e., high specificity, sensitivity, and affinity. The affinity constants obtained are of the order of 0.1 nM. These and other physical

TABLE I PROPERTIES OF THE GLYCOPROTEIN HORMONES

Molecular weight

Number of amino acid residues

Origin

Hormone

Molecular nature

Pituitary

Follitropin (follicle-stimulating hormone, FSH)

GI ycoprotein 2 subunits (a,P)

36,000

204 89 (a)

Lutropin (luteinizing hormone, LH)

GIycoprotein

Thyrotropin (thyroid-stimulating hormone, TSH)

Glycoprotein 2 subunits (a,p)

28,850 15,750 (a) 15.350 (P) 25,000

Choriotropin (human choriogonadotropin, hCG)

Glycoprotein 2 subunits (a,P)

Placenta

36,000 14.500 (a) 22,200 (P)

Site of action

26

Ovary (follicle), testis (Sertoli cells)

218 89 (a) 129 (P) 215

23 16 (a) 7 (P) 20

Ovary (follicle), testis (Leydig cells)

90 (a) 125 ($1

14 (a)

23 I 92 (a) 139 (p)

45

115

2 subunits (a,f3)

Carbohydrate content (res mole-1)

(P)

6

(PI

16 (a) 19 (P)

Thyroid, adipose tissue


147

properties and the data obtained on the purification of hormone receptors are summarized in Table 11. In order to characterize the hormone binding sites it is necessary to prepare radioactively labeled hormones of high specific radioactivity. For the protein hormones, usually the i2sl-labeled derivatives are used. The development of mild labeling techniques (e.g., with lactoperoxidase, Thorell and Johansson, 1971) has enabled the preparation of '2s1-labeled hormone usually without a loss in biological activity. Using these labeled hormones it has been possible to demonstrate specific binding in target issues. High-affinity capacity receptors have been reported in target cells. However, in view of the virtual irreversibility of the binding (see p. 152 for references), especially with lutropin, the theoretical assumptions on which many of the binding data are based (e.g., reversible binding, equilibrium conditions) may not be valid. Nonionic detergents have been extensively used to solubilize hormone receptors, and this has been achieved for LH/hCG (see Dufau et al., 1975, for references) and FSH (Dufau et al., 1977a). In the male, the plasma membrane receptor for LH is located in the testis Leydig cell, and in the female the ovarian granulosa cell. The LH receptor has been solubilized and purified from both the testis and ovary (Dufau and Catt, 19761, and various physicochemical parameters have been measured (see Table 11). The results obtained suggest that a common macromolecular configuration is shared by the gonadotropin receptors extracted from testis and ovary (Dufau et a / . . 1974). The LH receptor was extracted from rat testis interstitial cells by treatment with Triton X- 100. The solubilized receptors retained their hormonal specificity but the association constant at 24°C was lower (0.5-1 X loioM - '1 than that of the original particulate receptor for hCG (2.4 x I O ' O M - I). Exposure of particulate and soluble receptors to trypsin caused loss of gonadotropin binding activity, indicating the protein nature of an essential component of the receptor site. In addition, a significant role of phospholipid in the receptor was suggested by the reduced binding activity observed after treatments of particulate and soluble receptors with phospholipase A. By reference to the behaviors of standard proteins during filtration on Sepharose 6B, the hydrodynamic radius of the receptor was calculated to be 64 A. The sedimentation constant of the free receptor was 6.5 S, and that of the hormone-receptor complex was 7.5 S. Based on these physical analyses, on gel filtration and density gradient centrifugation, it was concluded that both the free and hormone-bound receptor exist in solution as elongated molecules with molecular weights of 194,000 and 224,000, respectively (Dufau et af.. 1973). The solubilized receptors were purified 15,000-fold by affinity chromatography on agarose coupled to hCG columns (Dufau et al., 1975). A water-soluble hCG binding protein has also been isolated from Leydig

TABLE I1 PROPERTIES OF HORMONE RECEFTORS Affinity consiani

Hormone Luvopin

Folh~ropin Thyroid-stimulating hormone

Tissue Ovary

Solvent used for solubilizaiion of receptor Tnton X-I00

Changes in specificity after solubilirarion

Before roiubiiirauon (M- '1

None

After solubilization (M-1) 0.5-1

Testii

Tnton X-100

None

24 x

Testis Thyroid

Tnton X-I00 Lithium diiodooxalicylate

None None

8.5 x 10'0

10'0

Y

10"'

Naiure of receptor Protein,

phospholipid 0.5-1 x 10i" Protein1 phospholipid 1.6 X lo9 Protein

Purification Estimdicd achieved molecular I +fold1 weight

250

wdimenlation constant 6.0. 6 75 Stokec radius 6 4 nm (61A,. red,mentation constant 7.55

Slokei radius 6.0 nrn (60 Al.

15.ooO

15,oOO

Other propenies

194.000

15.-

270.000

References Dufau er 01. (19741;

Haour and Saxena (19741 Duf~ucr a/ (1973) Dufau er a / . ( 1 9 7 7 ~ ) Tate cr 01. (1975a.b)


149

cell membrane fractions (Pahnke and Leidenberger, 1978). The molecular weight, sedimentation coefficients, and K, were found to be 71,500,4.35 S , and 0.75-1 .O X 1O1() M - ’ , respectively. The evidence obtained suggested that the hCG water-soluble binding protein was derived from membrane-bound LH receptors possibly cleaved by lysosomal or other proteolytic enzymes or, alternatively, they might be newly formed receptor material that had not been inserted into the membrane. An antiserum to soluble rat luteal LH receptors has been raised in rabbits (Lubrosky and Behrman, 1979). LH-dependent progesterone secretion from isolated rat ovarian cells was reduced by the antiserum, although there was no change in the LH binding, indicating that the antiserum was bound at a site different from the LH binding site. FSH receptors are present in the Sertoli cell plasma membranes in the rat testes seminiferous tubules (see review by Dorrington and Armstrong, 1979). These receptors have been solubilized by extracting 20-day-old male rat testes preparations with 1% Triton X-100 (Dufau ef al., 1977b). The association constant of the detergent-solubilized FSH receptors was higher than that of the particulate receptors (8.5 x lo”, I .6 X lo9 M - I , respectively), which is in contrast to the fall in binding affinity observed during solubilization of testicular LH receptors (Dufau et al.. 1973, 1975). Water-soluble binding sites with a high affinity for 12sI-labeledFSH ( K , I . 17 X 10yM- I ) were also detected. These water-soluble receptors represented about 20% of the FSH receptors in the testis. Bovine TSH receptors from thyroid tissue have been solubilized with lithium diiodooxalicylate and were found to be heterogeneous in size, in that they had binding components with molecular weights of 268,000, 160,000, 75,000, and 15,000-30,000 (Tate et a/., 1975a,b). Tryptic digestion converted all three higher molecular weight components to the same 15,000- to 30,000-dalton species. The latter had all the binding properties of the higher molecular weight forms including nonlinear Scatchard plots. The tryptic fragment of the solubilized receptor was purified approximately 250-fold by affinity chromatography on TSH-Sepharose columns.

IV. INVOLVEMENT OF CYCLIC AMP IN HORMONE ACTION It has been clearly shown that cyclic AMP can mimic the action of FSH, LH, and TSH, and there is good evidence that it plays an obligatory role in the action of these hormones. The early investigations revealed that cyclic AMP (or more active derivatives) mimicked the effects of the hormone on the cell response (e.g., steroidogenesis in ovarian and testicular cells) and that the effect of cyclic AMP was not additive to the maximum stimulating level of the hormone. The

150

BRIAN A. COOKE

hormone-induced increase in cyclic AMP preceded the effect on cell response and phosphodiesterase inhibitors potentiated the hormonal effect on cyclic AMP production and the cellular response. These data satisfied the criteria proposed by Sutherland and co-workers (Robison et al., 1971) for the mediation of cyclic AMP as the second messenger. However, in some systems cyclic AMP production was observed to be undetectable at levels of the hormone that gave a submaximal response of the cell; for example, in ovarian and testicular cells it was demonstrated that increased steroidogenesis could be detected with hormone concentrations 10 times lower than those required to detect changes in cyclic AMP levels (Marsh 1966; Catt and Dufau, 1973; Moyle and Ramachandran, 1973; Rommerts et al., 1973). This led to the conclusion that the role of cyclic AMP at physiological concentrations of hCG or LH may not be obligatory. However, it was argued that small changes in cyclic AMP could occur within the cells which were not detectable by the methods used; that the latter was true was indicated by the stimulatory effects of phosphodiesterase inhibitors on testosterone release in testis cells, but not on detectable cyclic AMP production, by low amounts of hCG (Catt and Dufau, 1973). LH was known to stimulate cyclic AMP-dependent protein kinases in Leydig cells (Cooke and van der Kemp, 1976) (the only known mechanism of cyclic AMP action), and therefore it was investigated whether at low concentrations of LH cyclic AMP-dependent protein kinase activation in testis Leydig cells was a more sensitive parameter than the cyclic AMP concentration itself (Cooke et al., 1976). It was clearly demonstrated that all concentrations of LH which stimulated testosterone production also stimulated protein kinase activation. Again with the lower stimulating amounts of LH (' L, Linear.

-

61 60

Break point ("CI

Above hreak

Below bredk

28 22 L

25

84

61

I60

71

el ul. ( 1976a.h).

whereas defined phospholipids, with low phase transitions, depress the lipid phase separation temperature (Table 111). Essentially similar results were obtained using cultured Chang liver cells (Bakardjieva et nl., 1979). E. Cholesterol

Although cholesterol can be expected to exert a profound effect on membrane fluidity, and hence the activity of integral enzymes, it can also perturb enzyme function in a number of other ways. For example, it can interact preferentially with certain phospholipid species (Demel er d., 1977) to form cholesterol-rich and -poor domains. As proteins tend to segregate in cholesterol-poor domains (Kleeman and McConnell, 1976; Houslay and Palmer, 1978), alterations in membrane cholesterol will modulate both the chemical composition of the lipid pool available to interact with the protein and the concentration of protein within its available lipid pool. Each of these parameters could influence enzyme activity, as perhaps could the direct interaction of cholesterol with the protein itself. All of the studies attempting to describe the actions of cholesterol on adenylate cyclase activity that have been reported so far have involved either treating intact cells with cholesterol-rich and -poor liposomes or by using cell mutants defective in cholesterol biosynthesis (Sinha et al., 1977; Insel et al., 1978; Sinensky et al., 1979). Both of these methods can have severe drawbacks that markedly affect interpretation of results. If viable whole cells are employed, then attempts to manipulate plasma membrane cholesterol levels by adding it to the serum are indeed successful, but it is likely that the cells will attempt to adapt to this change in fluidity by altering their phospholipids (see Kimelberg, 1977) and have indeed been noted recently in such mutant cell lines (Sinensky, 1980). While mam-

THE LIPID ENVIRONMENT

203

malian cells are not exceptionally successful at achieving this, adaptive changes in membranes have always been observed (see Kimelberg, 1977). An additional difficulty in analyzing such experiments is the finding that serum lipoproteins can activate adenylate cyclase activity (Pairault et al., 1977; Ghiselli et al.. 19811. Conceivably, the introduction of cholesterol to cells grown in serum also modifies the expression of adenylate cyclase activity by perturbing the ability of serum lipoproteins to activate the enzyme. In cells incubated with liposomes, it is probable that considerable fusion can occur which would be expected to introduce exogenous phospholipids capable of affecting both the lipid composition of the membrane and the lipid-protein ratio and hence adenylate cyclase activity. Klein et NI. (1978) demonstrated that incubation of fibroblasts with cholesterol-enriched liposomes led to an increase in cell, and presumably plasma membrane, cholesterol content. Such treatment also mediated a loss of basal, fluoride-, and prostaglandin E,-stimulated adenylate cyclase activity. Although these observations would be consistent with elevated cholesterol levels’ decreasing membrane fluidity and inhibiting the enzyme, further experimentation is needed before making such an assignment. Comparable results were obtained using either cholesterol-egg lecithin or cholesterol-dipalniitoyl phosphatidylcholine, even though fusion of the cells with liposomes occurred (Klein et al., 1978), which implies that the inhibitory effect was achieved by gross overloading with cholesterol. Similar studies were performed on human platelets t o assess the relationship between membrane cholesterol and adenylate cyclase activity (Sinha et a / . , 1977; Insel e t a / ., 1978). However, Insel eral. (1978) were unable to confirm the earlier findings of this group working on platelets (Sinha et d.,1977) that elevated cholesterol levels suppressed fluoride-, hormone-, and guanine nucleotide-stimulated activities. They now maintain that platelets incubated with cholesterol-rich liposomes exhibit an enhanced basal activity but similar hormone-stimulated activities, i.e., there is a net decrease in the fold stimulation achieved by hormone. Since lipid analyses were not carried out on the native and cholesterol-manipulated platelet plasma membranes, it is not possible to evaluate accurately the action of membrane cholesterol on adenylate cyclase activity in these studies. Sinensky el al. ( 1979) have modulated the plasma membrane cholesterol content of a mutant Chinese hamster ovary (CHO) cell line, defective in cholesterol biosynthesis, by supplementing the medium with cholesterol. These workers observed an enhanced basal activity with increasing cholesterol levels, although the relative stimulation achieved by fluoride or prostaglandin E, fell dramatically. The increased cholesterol content of the membrane was associated with a decrease in the bilayer fluidity detected with a fatty acid spin label, and it was suggested that the increased membrane acyl chain ordering was responsible for the activation of basal adenylate cyclase (Sinensky et al., 1979). Their view

MILES D. HOUSLAY AND LARRY M. GORDON

204

that decreases in membrane fluidity play a controlling factor in activating basal adenylate cyclase is somewhat at odds with their finding that decreases in fluidity achieved by temperature reductions are associated with decreases in basal adenylate cyclase activity (Sinensky et a / ., 1979). The interpretation of Sinensky et al. (1979) also appears somewhat anomalous, inasmuch as there is considerable evidence that for adenylate cyclase and other enzymes (see Gordon et af.,1980a) increased membrane fluidity leads to enhanced activity. As the basal activity appears to be relatively insensitive to the lipid fluidity (see Section III,C), and considering the myriad interactions of cholesterol in biological membranes, it would seem more likely that in platelets and CHO cells the increased cholesterol content is complexing an inhibitory phospholipid species. This would enhance the basal activity, but the decreased fluidity would be expected to inhibit the hormone responses. Such factors would explain the reduced net stimulation over basal observed activity in these systems. Thermodependence studies of the adenylate cyclase activity of membranes from native and cholesterol-supplemented CHO cells were also carried out (Sinensky et al., 1979). Unfortunately, the number of data points (4-5) is insufficient to make any meaningful evaluation, save that the activation energies for adenylate cyclase in cholesterol-rich membranes are greater than for the native membranes. This would be consistent with our interpretation stated above that the more rigid environment would render the reaction less favorable. In an attempt to define the effect of cholesterol upon adenylate cyclase A. D. Whetton, L. M. Gordon, and M. D. Houslay (unpublished results) have developed a technique, using liposomes, to both increase and decrease the cholesterol content of rat liver plasma membranes. The manipulations are carried out at 0°-4"C, so that adenylate cyclase is not denatured, and conditions are such that fusion of liposomes with the plasma membrane is negligible. From the results of such experiments it would appear that cholesterol optimizes the functioning of adenylate cyclase in these membranes (Fig. 15). Any increase or decrease in the cholesterol content from that exhibited by native membranes [0.65-0.72, cholesteroUphospholipid ( U P ) molar ratio] leads to a reduced activity of the enzyme. These effects are reversible upon further manipulation of cholesterol levels. We should note that if suboptimal cholesterol concentrations exist in a membrane, and the responses are similar to those described here, then over an appropriate range an increase in cholesterol concentration may actually enhance enzyme activity. Furthermore, basal activity seems remarkably insensitive to increased cholesterol concentrations compared to the hormone-stimulated activity. Thus the degree of activation of adenylate cyclase, over basal activity, by glucagon will drop precipitously upon increased cholesterol levels up until they become very high ( U P 1 .O). A decreased cholesterol content will also diminish the net stimulation by hormone. Such results imply that the cholesterol content of membranes has an important influence on adenylate cyclase function,

-


205

FIG, 15. Cholesterol optimizes the functioning of adenylate cyclase. The cholesterol content of native rat liver plasma membranes (CIP = 0.72) was manipulated using liposomes either loaded with or free from cholesterol under conditions where all the cholesterol was transferred by exchange and no liposomal fusion occurred with the membranes. The steady-state activity of the enzyme at 30°C was followed as before (see Houslay et d . . 1980a) with various activating ligands (A. D. Whetton and M. D. Houslay. unpublished).

206


TABLE IV ALTERATIONS IN THE CHOLESTEROL CONTENT O t RAT LIVER PLASMAMEMBRANESMODULATE THE BILAYEK FLUIDITV WITH THE 5-NlTROXIUE STEARATE FATTY ACID DETECTED SPIN PROBE CholesteroUphospholipid molar ratio

S(Tl1)”

0.36

0.43

0.72“

0.82

0.94

0.715c (0.005)d

0.686 (0.003)

0.665 (0.004)

0.720 (0.004)

0.715 (0.005)

Native membrane preparation. The polarity-uncorrected order parameter S(T11) was determined from native and cholesterol-modulated rat liver plasma membranes at 30°C. Mean determined from duplicate measurements. Values in parentheses indicate I SD calculated from duplicate determinations.

as changes can lead to as much as a 50% decline in the degree of hormone stimulation. How does cholesterol exert its effects on adenylate cyclase? Certainly an increased membrane cholesterol content is usually accompanied by decreased membrane fluidity (see Kimelberg, 1977), and this is just what we observed from spin-label studies (Table IV). Increases in cholesterol levels may inhibit adenylate cyclase activity, at least in part, by making the membrane more rigid. Another possibility is that incorporation of cholesterol into the bilayer decreases the size of those cholesterol-poor domains sampled by the protein, thereby inhibiting enzyme activity either by promoting protein-protein interactions or by restricting the availability to the enzyme of phospholipids that are critical for maintaining maximal activity. This may be somewhat analogous to lowering the temperature below T , (see Fig. 2). Additional experiments show that the enzyme in a high-cholesterol environment is more potently activated by benzyl alcohol than is the native enzyme. Indeed the maximal specific activity achieved by optimal benzyl alcohol concentrations is identical for the enzyme in both native and high-cholesterol membranes. Accordingly, inhibition at high cholesterol concentrations may be due entirely to the increased rigidity of the membrane and to the restricted size of lipid domains sampled by the enzyme, as it can be fully reversed by benzyl alcohol, an agent capable of ‘fluidizing liver plasma membranes (Gordon et al., 1980a) and disrupting cholesterol-rich domains in model membranes (Colley and Metcalfe, 1972). Surprisingly, decreasing the cholesterol content over the range examined also appears to lower the fluidity of the membrane (Table IV). This at first rather


207

unexpected result may well reflect the inherent domain structure of the membrane. Here, cholesterol in the native membrane may be preferentially clustering specific, rather rigid lipids such as sphingomyelin or acidic phospholipids to form domains excluding spin probe and adenylate cyclase. However, the release of such lipids, upon decreasing the cholesterol content, would lead to their mixing in the lipid pool, thereby increasing the rigidity of the environment of the enzyme and the probe. Certainly, adenylate cyclase appears to be somewhat more activated by benzyl alcohol, but the effect is nowhere near as dramatic as with high-cholesterol membranes. On this basis we would like to suggest that the inhibition observed at low cholesterol/phosphoIipid levels is due to the release of inhibitory phospholipid species from complexes in cholesterol-rich domains. Indeed, these observations are entirely in accord with our earlier experiments using the polyene amphoptericin B (Dipple and Houslay, 1979b) and with those of others using filipin (Puchwein et a / . , 1974; Lad et al.. 1979). These drugs form high-affinity, specific complexes with cholesterol in the membranes and so would be expected to mimic the effects of depleting cholesterol. This is in fact what happens as both fluoride- and glucagon-stimulated activities are inhibited by increasing concentrations of these drugs. These effects can be shown to be fully reversible in the case of amphoptericin B (Dipple and Houslay, 1979b). In line with our arguments, identical changes in lipid phase separation temperatures and the production of a new lipid phase separation are seen either by decreasing the cholesterol content (A. D. Whetton, L. M. Gordon, and M. D. Houslay, unpublished experiments) or by using amphoptericin B (Dipple and Houslay, 1979b). Presumably this is due both to the depletion of cholesterol and the release of specific phospholipids into the lipid domains interacting with adenylate cyclase and sampled by the spin probe. We cannot, however, rule out entirely the possibility that cholesterol itself is required to interact with the protein in order to optimize its function.

F. Hormone-Mediated Alterations in Lipid Fluidity There have been numerous reports employing extrinsic fluorescent probes that in vitro additions of hormones (e.g., insulin and growth hormone) could elicit changes in plasma membrane fluidity, thereby altering cellular functions (see Sauerheber et al., 1980, for review). All of the above studies must, however, be viewed as preliminary inasmuch as none have been confirmed by independent investigators using similar or complementary techniques. Indeed, recent experiments employing spin labels tend to disprove such contentions (Amatruda andFinch, 1979; Sauerheber et a/., 1980). The view that changes in the membrane fluidity will mediate the action of hormones appears untenable, since perturbations in the bilayer fluidity will influence the activity of a wide range of enzymes (see Kimelberg, 1977; Gordon et ul., 1980a) and thus lack the specificity that is

208


the hallmark of a classic hormone effector. Moreover, if such a hypothesis were correct, then agents that modulate the membrane fluidity (e.g., local anesthetics and cholesterol) should act as hormones; this clearly is not the case. Recently, it has been suggested (Hirata and Axelrod, 1980) that agonist occupancy of P-adrenergic receptors, benzodiazepine receptors, and mast cell IgE receptors stimulates the plasma membrane to synthesize phosphatidylcholine from phosphatidylethanolamine.This is apparently achieved by the liganded receptor activating two asymmetrically oriented S-adenosylmethionine (SAM)required methyltransferases in the plasma membrane. The first enzyme exposed at the cytosol surface of the membrane monomethylates phosphatidylethanolamine and is presumed to trigger the transfer of this lipid to the external half of the bilayer, where it cad be further methylated. In in vitro experiments with DPH-labeled erythrocytes (Hirata and Axelrod, 1978), it has been claimed that monomethylation induces an increase in bilayer fluidity, leading to the activation of hormone-responsive adenylate cyclase. As the fluoride-stimulated activity was unaffected, it is possible that any bilayer perturbation is restricted to the outer monolayer. The interpretation of these results in terms of a perturbation of the membrane fluidity (Hirata and Axelrod, 1978, 1980) has been severely criticized (Vance and de Kruijff, 1980) on the basis that the minute amounts of lipid transformed in the process would be insufficient to fluidize the bilayer. As yet, the report by Hirata and Axelrod (1 978) that monomethylation triggers a membrane fluidization has not received independent corroboration. Thus the effects may not be due to a change in fluidity detectable by extrinsic reporter groups but instead may be exerted by a direct action of the modified lipid on the protein. However, it is unlikely that this mechanism will be of general importance in hormone action, although it may prove to have a key role in the transfer of lipid molecules across the membrane of the endoplasmic reticulum (Higgins, 1981) during the biosynthesis of lipids and the generation of membrane lipid asymmetry.

IV. SELECTIVE MODULATION OF ADENYLATE CYCLASE BY ASYMMETRIC PERTURBATIONS OF THE MEMBRANE BILAYER The previous section established that alterations in the bilayer fluidity may act as effectors of the activity of adenylate cyclase. Since the adenylate cyclase complex is asymmetrically oriented with respect to the bilayer, it is reasonable to suppose that agents which perturb the fluidity of one or the other half of the lipid bilayer will selectively modify the expression of this enzyme activity. Here, we consider the results of earlier structural and functional studies to determine whether such a relationship indeed exists for such perturbants as charged local anesthetics, Ca2+, and mitogenic agents.


209

A. Positively and Negatively Charged Local Anesthetics Sheetz and Singer (1974) have suggested that, owing to the now well-documented asymmetry of the lipid composition of plasma membranes (see Rothman and Lenard, 1977), in which negatively charged phospholipids predominate at the cytosol-facing surface, drugs of opposite charge may act preferentially at one or the other side of the bilayer. One might well expect such drugs to act selectively on functioning proteins that are themselves asymmetrically orientated in the bilayer. In recent studies (Houslay et al., 1980b, 1981a) we have sought to test the above hypothesis by examining the actions of positively and negatively charged local anesthetics on the activity of hepatic adenylate cyclase. Such an investigation seemed warranted, since Higgins and Evans (1978) have demonstrated that, similar to the surface membranes of other eukaryotic cells, the acidic phospholipids are restricted to the cytosol surface of the bilayer of rat liver plasma membranes. Previous work has also defined the topology of glucagon-stimulated adenylate cyclase relative to the bilayer, where the receptor is exposed at the external surface of the plasma membrane and the catalytic unit and guanine nucleotide regulatory component are at the internal surface (Houslay et af., 1980a). In the absence of glucagon, the receptor and catalytic units are able to undergo free lateral diffusion as independent entities (Houslay et al., 1977; Houslay, 1981a) and the activity of the free catalytic unit (i.e., the uncoupled activity) is sensitive to the lipid environment of the cytosol side of the bilayer only (Houslay and Palmer, 1978; Houslay, 1979). However, in the presence of glucagon (coupled activity), the occupied receptor and catalytic unit interact to form a transmembrane complex under conditions in which a mobile receptor model is obeyed, spanning the lipid bilayer. This complex allows adenylate cyclase to respond to the lipid environment of both halves of the bilayer (see Section 111,B). Thus, the measurement of adenylate cyclase in the coupled and uncoupled states provides us with a useful system to probe the lipid properties of both halves of the bilayer. The experimental protocol that we employed to study the effects of charged drugs on rat liver plasma membranes is as follows. Dose-response curves were performed at 30°C (i.e., above the high-temperature onset of the lipid phase separation) to assess the actions of the agents on the glucagon-stimulated (coupled) activity and the fluoride-stimulated (uncoupled) activity of the membranebound enzyme. Parallel ESR studies were conducted on 5-nitroxide stearatelabeled membranes to determine the role that changes in the bilayer fluidity might play. To discriminate between direct effects on the catalytic unit and those transmitted through the native lipid bilayer, the actions of the agent on the activity of detergent-solubilized adenylate cyclase were also examined. Finally, thermodependence studies on the activity in the membrane-bound and solubilized states and on the membrane fluidity were conducted with selected con-

210


6ol

40 20

1, 1

I

-0.5

0

I

I

1.0 1.5 log [Rilocaine] mM

0.5

2.0

Fic. 16. Effect of prilocaine on the adenylate cyclase activity of rat liver plasma membranes.

(A) Glucagon-stimulated activity (mobile receptor model); ( 0 ) fluoride-stimulated adenylate lubrol-solubilized, fluoride-preactivated enzyme at 30°C. (Data from Houslay cyclase; (0) 1980b.)

rt

a / .,

centrations of the agent. These experiments allow us to evaluate whether a given agent is perturbing the thermotropic lipid phase separation occurring in the outer half of the bilayer, or, alternatively, inducing a lipid phase separation in the inner half of the bilayer. The effects of the positively charged amine local anesthetics prilocaine, nupercaine, and carbocaine on the structural and functional properties of rat liver plasma membranes were investigated (Houslay er a / . , 1980a; Gordon er a / . , 1980b). As shown in Fig. 16, concentrations of prilocaine above 4 mM led to an augmentation of fluoride-stimulated adenylate cyclase activity, up to a maximum of 150% of its original activity at 10 mM prilocaine; further increases in prilocaine concentration led to a progressive inhibition of the activity. When these experiments were repeated on an enzyme preparation that had been solubilized with the nonionic detergent Lubrol 12A9 the fluoride-stimulated activity exhibited no response to these concentrations of prilocaine. In contrast to its effect on the fluoride-stimulated activity, the glucagon-stimulated activity began to be inhibited by prilocaine at concentrations where the fluoride-stimulated activity began to increase. Further addition of this compound led to a progressive inhibition of the glucagon-stimulated activity (Fig. 16). All these functional effects were fully reversible upon washing to remove the anesthetic. Spin-label studies of liver plasma membranes demonstrated that, at concentrations of prilocaine greater than 4 mM, the bilayer fluidity was progressively augmented up to the highest drug concentration tested (33 mM) (Fig. 17). Carbocaine and nupercaine

21 1


*r

-I

-+= - 4 v)

a

-6

d.5 l o ;1. log [Prilocaine] mM

&

FIG. 17. Prilocaine incrcases the fluidity of rat liver plasma membranes labeled with 5-nitroxide stearate. Dependence of AS(T11 ) on prilocaine concentration at 30°C. (Data from Houslay ef ul., I980b. )

(dibucaine) exerted similar effects on the activities of the coupled and uncoupled adenylate cyclase and the fluidity of liver plasma membranes (Gordon et al., 1980b). Although carbocaine achieved these perturbations over a drug concentration range identical to that of prilocaine, nupercaine manifested its actions on both membrane fluidity and adenylate cyclase activity at a 10-fold lower concentration. The activation of the fluoride-stimulated activity mediated by prilocaine, carbocaine, and nupercaine is most likely due to the drug-induced fluidization of the rat liver plasma membrane bilayer. Such an interpretation agrees with our previous observations that low benzyl alcohol concentrations (up to 30-40 nM) progressively activated the fluoride-stimulated activity and also increased the lipid fluidity to the same degree seen with the cationic drugs (Figs. 3,5,and 16). Further support for this hypothesis comes from the finding the nupercaine initiates increases in both the bilayer fluidity and the fluoride-stimulated activity at a concentration that is an order of magnitude less than that noted for either prilocaine or carbocaine. Clearly, these effects are not due simply to the direct action of the drug on those portions of the enzyme exposed to the aqueous buffer, since incubation of the cationic drugs did not perturb the activity of the detergentsolubilized enzyme (Fig. 16). Since anionic lipids predominate at the cytosolfacing surface of the bilayer, and cationic local anesthetics efficaciously act to increase the fluidity of such lipids in model systems, it is reasonable to suppose that the increase in fluidity achieved by the positively charged drugs (Fig. 17) is due to the interaction of the drug with the inner half of the bilayer (see Houslay et

21 2

MILES D. HOUSLAY AND LARRY M. GORDON Anionic drugs act here

TSrn 'Native

+phewdtai

+pttbcairA

28

16

20

-

11

outer

Acidic phospholipids Fluoride., guanine nuclaotldestirnulated

\

Catlonlc drugs act here Giucagon.stlrnulated

Frc. 18. Sensitivity of ligand-stimulated adenylate cyclase in liver plasma membranes to asymmetric perturbations. Ts,Lipid phase separation temperature; R. receptor; C. catalytic unit; g, G/F coupling protein. This demonstrates the states of the enzyme when stimulated by fluoride or guanine nucleotides and when stimulated by glucagon and suboptimal concentrations (0.08 F M ) of GTP where a mobile receptor model is obeyed. ( f ) indicates inserted fatty acid spin label; h, glucagon. (Adapted from Houslay ef (11.. 1981 .)

1980b, 1981). This selective fluidization then leads to the activation of the fluoride-stimulated adenylate cyclase, an uncoupled activity that responds only to the lipid environment of the cytosol-facing half of the bilayer (Fig. 18). Of particular interest is our finding that these cationic drugs did not cause an enhancement of the glucagon-stimulated activity but instead began to inhibit the enzyme at concentrations where activation of the fluoride-stimulated activity commenced (Fig. 16). One might expect that, since the uncoupled catalytic unit was activated due to an increase in the fluidity of the surrounding lipid, the coupled catalytic unit would be similarly activated; this clearly was not the case. That the inhibition occurred at a concentration similar to the activation of the uncoupled catalytic unit suggests that they have a common origin. Presumably, this is the interaction of the cationic drug with the bilayer leading to an increase in fluidity, an event that we have suggested is preferentially localized to the inner half of the bilayer. However, this may well affect the activity of the coupled enzyme differently than that of the uncoupled enzyme. It was earlier proposed that the vertical positioning of the catalytic unit in the bilayer alters upon its coupling to the glucagon receptor (Dipple and Houslay, 1978), and such a change might lead it to react unfavorably with the surrounding lipid or drug. An alternative explanation focuses on the demonstration that negatively charged phospholipids might be involved in the coupling process between the receptor and catalytic unit (see Section 111,E). Because there is good evidence for a strong interaction between cationic anesthetics and negatively charged phospholipids, the involvement of such lipids in the coupling process may well yield an inhibitory response that would mask any activation caused by an increase in bilayer fluidity. Additional evidence that the cationic anesthetic prilocaine perturbs rat liver a/.,


1.0

213

-

0.9-

f

0.8 -

d

-

/

# 0 0.7 0.6 -

0.5 3.1

I

3.2

I

3.3

1

3.4

I

35

1

3.6

1

3.7

103

T O FIG. 19. Effect of 10 mM prilocaine on the temperature dependence of the order parameter for a S(T11) (0).and S ( T , ) fatty acid spin probe incorporated into rat liver plasma membrane. S (A), (W). (Adapted from Houslay ef al., 1980b.)

plasma membranes asymmetrically has been provided by a study of the thermodependence of both the activity of adenylate cyclase and of the membrane fluidity detected with a fatty acid spin label (Houslay er ul., 1980b). Prilocaine (10 mM) had no significant effect on the thermotropic lipid phase separation occurring at 28°C that could be detected with either the spin probe (Fig. 19) or Arrhenius plots of glucagon-stimulated adenylate cyclase (Fig. 20). These data suggest that 10 mM prilocaine does not perturb the lipids of the external half of the bilayer, but instead selectively fluidizes the lipids of the inner half of the bilayer. Consistent with this hypothesis are our observations that the Arrhenius plots of the fluoride-stimulated adenylate cyclase activity, which were normally linear over the temperature range studied, exhibited a well-defined break at around 11°C when 10 mM prilocaine was present in the assays (Fig. 20). The

214


3.c

2.5

- 2.0 T

0,

E

ln

c .-

C

->3 1.5 0 7

0 -

1.0

0.5 3.1

I

3.2

1

3.3

1

I

3.4 3.5

I

I

3.6

3.7

FIG. 20. Effect of 10 mM prilocaine on Arrhenius plots of glucagon- and fluoride-stimulated adenylate cyclase activity in liver plasma membranes. For fluoride- (01and glucagon- (A)stimulated activies. (Adapted from Houslay el a/.. 1980b.)

occurrence of this prilocaine-induced break was also apparent in Arrhenius plots of glucagon-stimulated adenylate cyclase activity (Fig. 20), which, unlike the fluoride-stimulated activity, is sensitive to the lipid environment of both halves of the bilayer. As Arrhenius plots for the solubilized enzyme were unaffected by 10 mM prilocaine, the break in the Arrhenius plots occurring at around I 1"C does not appear to be due to an effect of prilocaine on the protein, but is instead a consequence of a lipid phase separation that is reported by the spin probe (Fig. 19). The simplest explanation of these data is that the prilocaine- (10 mM) induced lipid phase separation is localized to the inner (cytosol-facing) half of the bilayer (Fig. 18). The inhibitory effects of fluoride-stimulated adenylate cyclase activity, observed at the higher cationic anesthetic concentrations (Fig. 16), have been attributed to the drug preventing annular/boundary lipid from interacting with


21 5

sites of the protein (i.e., the annular lipid displacement model; Gordon et al., 1980a,b; Houslay et a l . , 1980b). Clearly, this inhibition is not simply due to a direct action of cationic anesthetics on the protein, as the detergent-solubilized enzyme was unaffected for each agent tested. By contrasting the effects on adenylate cyclase activity induced by increases in fluidity achieved by temperature elevations or anesthetic additions, we have also ruled out the cationic drugmediated inhibition as being a consequence of a “too-fluid” bilayer (Gordon et al., 1980a; Houslay et al., 1980b). Instead, we suggest that occupancy of annular sites by the anesthetic leads to reduced activity, because either the anesthetic itself was inhibitory or the displaced lipid was essential for activity. Since the lipid asymmetry of cell plasma membranes, in which anionic lipids predominate at the cytosol-facing surface, appears to be a widespread phenomonon (see Rothman and Lenard, 1977), the adenylate cyclase activity of a number of other tissues might well be expected to respond to cationic anesthetics in a manner similar to that noted in rat liver plasma membranes. Indeed, the fluoride- and hormone-stimulated activities in such diverse systems as beef thyroid membranes (Wolff and Jones, 1970), rat adipocyte plasma membranes (Hepp et al., 1978), pigeon erythrocyte membranes (Salesse and Gamier, 1979), and rat liver plasma membranes (Houslay et ul., 1980b, 1981a) are all modulated similarly by positively charged drugs. Although the model presented by us to account for the actions of cationic drugs on rat liver plasma membranes (see above) may also be invoked to explain the functional effects of cationic anesthetics on these other membranes, additional experimentation will be needed before making any definitive assignments. The anionic drugs phenobarbital, pentobarbital, and salicylic acid were also examined for their actions on rat liver plasma membranes (Gordon et al., 1980b; Houslay et al., I98 1). Phenobarbital initially inhibited the glucagon-stimulated (coupled) activity at concentrations of about 1 mM, above which concentrations a dramatic activation ensued, that reached a maximum between 4 and 6 mM. Any further increases in phenobarbital concentration led to a progressive loss in the coupled activity. On the other hand, no effect of phenobarbital was observed on the fluoride-stimulated (uncoupled) activity of either native membranes or a Lubrol-solubilized enzyme (Fig. 21). All of the changes in activity were readily reversed upon washing to remove phenobarbital. Qualitatively similar results were obtained with the anionic drugs pentobarbital and salicylic acid. Thermodependence studies were also conducted to assess the effects of phenobarbital on the activity of hepatic adenylate cyclase and the membrane fluidity detected with a fatty acid spin probe. Addition of optimum (4 mM) phenobarbital concentrations to the assays had a dramatic effect on the Arrhenius plot for the glucagon-stimulated activity, such that the break was depressed from 28” to 16°C and the activation energies detected above and below the break were considerably less than those exhibited in the absence of phenobarbital (Fig. 2 2 ) . Howev-


216

" 2oo/

h

Glucagon

[phenobarbital 1 rnM

FIG. 21. Sensitivity of liver plasma membrane adenylate cyclase to phenobarbital. (0) Glucagon + suboptimal GTP; (0) fluoride and (B) lubrol-solubilized. fluoride-preactivated enzyme at 30°C. (From Houslay el d., 1981.)

3.5

3.0

-

i"

\ C .-

E 2.5

-

\ u

-5 0

>

m

2 2.0

1.5

& ;

1.0 3.1

Fluorlds Native

3.2

3.4

3.3 (OIT '

3.5

3.6

3.7

(K)

FIG.22. Arrhenius plots,of liver plasma membrane adenylate cyclase activity in the presence of 4 mM phenobarbital. Fluoride; (0) glucagon + suboptimal CTP concentration. (Data from Houslay et al., 1981.)

(m)

21 7

THE LIPID ENVIRONMENT 1.o

0.90

-

OB0

-

A m -

+Phenobarbital

Sprll) 1ooc

-

0.60 f I 1 3.2

I 3.3

3.4

3.5

3.6

FIG. 23. Phenobarbital's effect on the temperature dependence of the order parameter ST for a fatty acid spin probe inserted into liver plasma membranes. ( 0 )Treated and (A)native membranes. (Data from Houslay er al., 198 1 , )

er, the Arrhenius plot of the fluoride-stimulated (uncoupled) activity was apparently unaffected by the presence of phenobarbital. Arrhenius-type plots of the order parameter of 5-nitroxide stearate-labeled rat liver plasma membranes indicated that phenobarbital not only lowered the high-temperature onset of the lipid phase separation from 28" to 16°C but also increased the membrane fluidity for temperatures between 10" and 30°C (Fig. 23). The activation of the coupled adenylate cyclase activity that we observe appears to be due to a selective fluidization of the external half of the bilayer by phenobarbital (Fig. 18), since the drug-induced depression of the break temperature in the Arrhenius plot of the coupled activity (Fig. 22) closely parallels the reduction of the high-temperature onset of the lipid phase separation identified with our spin probe (Fig. 23). Further evidence of a selective fluidization of the external half of the bilayer by phenobarbital is gained from our observation that the activation energies for the coupled activity were reduced, whereas for the uncoupled activity not only does the Arrhenius plot stay linear but the activation energy for the reaction remained identical (Fig. 22). The fluidization of rat liver plasma membranes for temperatures between 10" and 30°C achieved by phenobarbital is consistent with the disordering that this drug exerts on spin-labeled phospho1ipid:cholesterol (2: 1) bilayers (Pang and Miller, 1978), and these effects may be a consequence of phenobarbital disrupting intrinsic phospho1ipid:cholesterol associations. It is significant that phenobarbital is much more effective in activating glucagon-stimulated adenylate cyclase at temperatures below 28°C than above (Fig. 22). The coupled enzyme is probably restricted to relatively cholesterol-

218


free lipid domains; otherwise, Arrhenius plots of the glucagon-stimulated activity of native membranes would be unable to exhibit a well-defined break at 28°C with a sharply increased activation energy below the break temperature (Fig. 22; see Houslay and Palmer, 1978; Dipple and Houslay, 1979b, for discussion). Indeed, increases in the cholesterol content of rat liver plasma membranes (from cho1esterol:phospholipid ratios of 0.65 to 0.94) using the liposome methodology described in Section III,E. abolished the break at 28°C normally observed in Arrhenius plots of the coupled activity and the order parameter of 5-nitroxide stearate-labeled membranes (A. D. Whetton, L. M. Gordon, and M. D. Houslay, unpublished results). The elevated activation energy noted for the coupled enzyme of native membranes below 28°C may be due directly or indirectly to the formation of cholesterol-rich QCC in the L matrix. For example, the increase in QCC as the temperature is lowered may segregate the coupled enzyme into a relatively small L phase, perhaps inhibiting enzyme activity by promoting protein-protein interactions (see Fig. 3). Support for such a hypothesis has been provided by earlier reports indicating that integral proteins are nonrandomly distributed in liver plasma membranes (Montesano et al., 1979) and that the presence of either QCC or S in other membranes inhomogeneously segregates proteins in the plane of the bilayer (Melchior and Steim, 1979). Alternatively, the formation of QCC may restrict the availability to the coupled enzyme of phospholipids that are critical for maintaining maximal activity. Consequently, phenobarbital may be such a potent activator of glucagon-stimulated adenylate cyclase at low temperatures partly because the bulk lipid fluidity is increased, but perhaps also because this drug facilitates disruption of clusters of QCC domains, or proteins, or both. A similar explanation has been proposed to account for the marked activation of glucagon-stimulated adenylate cyclase induced by 40 mM benzyl alcohol at low temperatures (Gordon et al., 1980a). The inhibition of the glucagon-stimulated activity that is observed at low concentrations of phenobarbital (Fig. 21) and other anionic drugs (Gordon et al., 1980b) may be due to these drugs having a direct effect on the receptor or acting at the coupling interface (see Houslay and Palmer, 1979; Whetton and Houslay, 1980), although it is possible that they could disrupt the interaction with specific lipids that may be required for activity (see Rubalcava and Rodbell, 1973). The inhibition of activity seen at high phenobarbital concentrations (Fig. 21) is similar to that observed with a number of neutral and charged local anesthetics (Figs. 5 and 16), and is probably due to the displacement of annular lipid from around the enzyme (see above). Thus, our studies of the effects of positively and negatively charged anesthetics on rat liver plasma membrane adenylate cyclase activity and fluidity present us with striking confirmation of the "bilayer-couple" theory of Sheetz and Singer (1974). However, the above results suggest only that charged drugs exhibit a greater tendency to interact with one or the other half of the bilayer of


219

liver plasma membranes and do not prove that any of these agents exclusively reside in only one half of the bilayer. It is likely, for example, that the anionic fatty acid spin probe used in this study, 5-nitroxide stearate, distributes between both halves of the bilayer. Clearly, some fraction of the label samples the outer half of the bilayer, since 5-nitroxide stearate detects the lipid phase separation in the outer leaflet (Houslay et al. 1979b,c; Gordon et al., 1980a) and is acutely sensitive to phenobarbital, which depresses the phase separation (Fig. 23). Inasmuch as the incorporated 5-nitroxide stearate probably retains a partial negative charge, it would not be unexpected for this probe to reside in the outer half of the bilayer, since that is where positive and neutral lipids are concentrated. Nevertheless, there is evidence suggesting that 5-nitroxide stearate also partitions into the inner half of the bilayer (Houslay et al., 1980b). Since the cationic local anesthetics nupercaine, carbocaine, and prilocaine each begins to stimulate fluoride-stimulated adenylate cyclase (i .e., an enzyme localized in the cytosol-facing leaflet) at the same concentrations that decrease S(T II ), a certain fraction of the 5-nitroxide stearate probe must be restricted to the inner half of the bilayer to report on this fluidization (Houslay et af., 1980a; Gordon et al., 1980b). That phenobarbital induces such dramatic effects on the lipid phase separation monitored by 5-nitroxide stearate (Fig. 23) indicates that the proportion of probe may be somewhat greater in the outer than in the inner half of the bilayer, perhaps owing to selective charge interactions (Fig. 18). The use of cationic and anionic drugs provides us with powerful tools capable of modulating the fluidity of each half of the bilayer of rat liver plasma membrane independently, Since a wide variety of liver plasma membrane enzymes are influenced by changes in the bilayer fluidity (Gordon et al., 1980a), one might predict that charged drugs may be used to define the vertical positioning of these integral proteins with respect to the bilayer. Indeed, recent studies on the effects of positively and negatively charged anesthetics on hepatic 5'-nucleotidase [i.e., an enzyme activity sensitive to the 28°C thermotropic lipid phase separation (Dippleand Houslay, 1978)and responsive to increases in lipid fluidity achieved by the neutral anesthetic benzyl alcohol (Gordon et af., 1980a)l demonstrate that this ectoenzyme is regulated by only the external half of the bilayer (Dipple et al., 1982). We propose that, once appropriate concentrations of anionic and cationic drugs have been determined for a given plasma membrane system from structural and functional studies, these agents may be of general use in defining the topology of integral enzymes relative to the bilayer.

B. Calcium Ca2+ exerts multiple effects on biological membrane processes. It is well known that Ca2 participates in a number of membrane-associated functions, +

220


including regulation of enzyme activities, transduction of hormonal information, stimulus secretion coupling, functioning of transport systems, neuronal conduction, and muscular contraction (see for reviews Rasmussen and Goodman, 1977; Triggle, 1972; Gordon and Sauerheber, 1982). In view of the facts that millimolar concentrations of Ca2+ reduce the fluidities of plasma membranes from a wide variety of tissues (see Gordon and Sauerheber, 1982) and that changes in bilayer fluidity modulate the activities of a number of membrane enzymes (Gordon et al., 1980a), it would not be untoward to hypothesize that Ca2+ perturbs certain of the aforementioned membrane functions by altering the lipid fluidity. L. M. Gordon, A . D. Whetton, S. Rawal, J. A. Esgate, and M. D. Houslay (unpublished results) have recently explored in some detail the relationships that may exist between the effects of Ca2 on the fluidity and lipid organization of rat liver plasma membranes and adenylate cyclase activity. As demonstrated by us before (Gordon et al., 1978), ESR studies on rat liver plasma membranes probed with 5-nitroxide stearate demonstrated that Ca2 decreased the lipid fluidity, as indicated by marked increases in the order parameter S. These effects are a concentration-dependent binding process reaching saturation at about 2.8 mM CaCI, and could be reversed with the use of a divalent cation-chelating agent. Over a range of cation concentrations similar to that which lowered the bilayer fluidity, the fluoride-stimulated (uncoupled) activity of the liver plasma membrane-bound enzyme was inhibited by Ca2 with an ID,, (concentration yielding 50% inhibition) of I mM. To discriminate between direct effects on the catalytic unit and those transmitted through the native lipid bilayer, the actions of Ca2+ on the activity of adenylate cyclase in the solubilized state were also examined. The fact that the fluoride-stimulated activities of the membrane-bound enzyme of a Lubrol-solubilized preparation were identically inhibited for Ca2 concentrations less than 2 mM suggests that Ca2+ achieves this effect not by reducing the bilayer fluidity but instead either by liganding to the protein and substrate (ATP) components or by binding to associated annular lipid. Only at Ca2 concentrations greater than 2 mM, where the membrane-bound enzyme was somewhat more inhibited than the solubilized preparation, is it probable that Ca2 +-mediated increases in lipid ordering influence the uncoupled activity. The glucagon-stimulated (coupled) activity is, on the other hand, more sensitive to Ca2+ inhibition with an ID,, of 0.2 mM. Such effects are clearly not due to either changes in lipid fluidity or alterations in the binding of ’2s1-labeled glucagon to its receptor, which was unaffected by CaZ+ concentrations up to 10 +

+

+

+

+

mM.

Although the inhibitory effects on the coupled and uncoupled hepatic adenylate cyclase activities are undoubtedly primarily due to the direct interaction of Ca2+ with the enzyme complex, it is, nevertheless, possible that the actions of Ca2+ on the lipid structures of the outer and inner halves of the bilayer might exert “second-order” perturbations on the temperature dependence of these

22 1


activities. Arrhenius-type plots of the order parameters indicated that the thermotropic lipid phase separation, which occurs in the outer half of the bilayer of native membranes from 28" to 19"C, is perturbed by millimolar concentrations of Ca2 due to the binding of cation to the outer leaflet such that the high-temperature onset is elevated to 32"-34"C, but the low-temperature onset is apparently undisturbed. Ca*+ was also found to bind to the cytosol-facing half of the bilayer, inasmuch as pretreatment of liver membranes with 10 mM prilocaine, an agent that selectively fluidizes the inner leaflet by interacting with acidic lipids residing there, inhibited the ability of Ca2+ to decrease the fluidity detected by the spin probe. It is of particular interest then that, with 1 mM CaCI,, Arrhenius plots of the glucagon-stimulated activity indicated breaks at 32" and 16"C, while those of the fluoride-stimulated activity showed a single break at 17°C. We achieves these effects by asymmetrically perturbing the suggest that Ca' bilayer, such that the high-temperature onset in the 'outer leaflet is elevated to 32"C, and a second lipid phase separation at 16"-17"C is induced by Ca2+ binding to acidic phospholipids residing in the inner leaflet. The glucagonstimulated activity responds to Ca2 -induced perturbations occurring in both halves of the bilayer because it is a transmembrane complex, while the fluoridestimulated activity senses only the lipid phase separation at 16-17°C because the catalytic unit, when dissociated from the glucagon receptor, is restricted to the inner half of the bilayer. The relative insensitivity of the activity of adenylate cyclase to Ca2 -dependent changes in the fluidity of 5-nitroxide stearate-labeled liver plasma membranes is somewhat anomalous, since increases in fluidity achieved by temperature elevations or low concentrations of anesthetics markedly activate this enzyme (see Sections III,C and D). The stimulation that ensued in these earlier studies was attributed to an increase in the conformational flexibility of the enzyme achieved by a relief of a physical constraint imposed by the bilayer upon the protein (Dipple and Houslay, 1978). One explanation of these disparate findings is that the decrease in fluidity mediated by Ca2 occurs in lipid domains of the membrane from which adenylate cyclase is excluded. On the other hand, the direct binding of Ca2+ that takes place under our assay conditions may "desensitize" adenylate cyclase to alterations in lipid fluidity. The latter hypothesis was tested by examining whether 1 mM CaCI, modified the ability of either temperature alterations or benzyl alcohol additions to perturb liver plasma membrane fluidity and adenylate cyclase activity. Although pretreatment of liver membranes with CaCI, did not alter the percentage change in S(T 11 ), S(T,) induced by raising the temperature from 30" to 37"C, the increases in the fluoride- and glucagon-stimulated activities achieved by this temperature elevation were each significantly reduced for the Ca2 -treated membranes when compared to that observed for native membranes (Fig. 24). Incubation of liver membranes with CaCI, also did not influence the response of either the bilayer +

+

+

+

+

+

222


I

Initial Condition: Native at 30%

I

I

T

Perturbant: TDC ralsed

t037

1 mM CaCi, at 30%

Native at 30°C

TC raised to 37

50 mM Benzyl Alcohol

mi CaCl,

31 50 mM Benzyi Alcohol

Native at 3oDc

Native at 3(pc

1 rnMCaC1,

50rnM Benzyi Alcohol + 1 mM CaCI,

FIG. 24. Does Ca2+ desensitize adenylate cyclase to changes in lipid fluidity? The presence of Ca2+ apparently reduces the sensitivity of adenylate cyclase to changes in fluidity caused by temperature and benzyl alcohol, and to Ca2+ itself. G SAC, Glucagon-stimulated activity; F SAC, fluoride-stimulated activity.

fluidity or the fluoride-stimulated activity to 50 mM benzyl alcohol at 30°C, but did inhibit the ability of this neutral anesthetic to activate the glucagon-stimulated activity (Fig. 24). Finally, Fig. 24 shows the respective actions of 1 mM CaCl,, or the joint addition of 50 mM benzyl alcohol and I mM CaCI,, on the fluidity and adenylate cyclase activity of liver membranes. Treatment with CaCl, decreased the fluidity of the liver membranes and was a potent inhibitor of the fluoride- and glucagon-stimulated activities. Simultaneous addition of 50 mM benzyl alcohol and 1 mM CaCl, served to override the Ca2+-dependent lipid ordering such that the bilayer was more fluid than untreated membranes, but was unable to return the fluoride- and glucagon-stimulated activities to the baseline values observed with native membranes. The results in Fig. 24 suggest that the binding of Ca2+ “desensitizes” the activity of the coupled enzymes and, to a lesser extent, that of the uncoupled enzyme to increases in bilayer fluidity. Once bound to adenylate cyclase, Ca2+ may alter the protein conformation such that the enzyme exists in a relatively stable state that resists changes mediated by the lipid fluidity. However, Ca2+ does not affect the ability of the enzyme to detect lipid phase separations occurring in the liver plasma membrane.


223

C. Mitogenic Agents The stimulation of lymphoid cells to mitosis achieved by a variety of agents may be viewed as a model of gene activation in higher organisms. Mitogenic agents appear to ligand to specific receptors on the surface membranes of these cells, and trigger a number of alterations in membrane-linked processes, inciuding influx of K + , Ca2+, amino acids, glucose, and uridine and changes in activity of nucleotide cyclases. Mitogenic agents trigger the patching and capping of plasma membrane proteins components and, in doing so, mediate the redistribution of plasma membrane lipid to form relatively rigid, glycosphingolipid-rich domains and more fluid, glycosphingolipid-poor domains in the outer half of the bilayer (Curtain et al., 1980). The stimulated lymphocytes have increased cellular cyclic AMP levels, which, using fluorescent antibody techniques, exhibit an identical localization to the patched mitogen and glycosphingolipid (Curtain, 1979; Curtain ef al., 1980). It may be that the glycosphingolipids are hydrogen bonded to receptor-bearing membrane components and migrate with them into the ligand-induced patches and caps. Although other explanations may be invoked to account for the above observations, the activation of lymphocyte adenylate cyclase may simply be a consequence of changes in the enzyme's lipid environment induced by the clustering of the glycosphingolipids. The formation of glycosphingolipid-rich and -poor domains, initiated by mitogenic agents, might well be expected to modulate the fluid properties of the outer half of the lipid bilayer. However, this lipid redistribution could also influence the fluidity of the cytosol-facing half of the bilayer, since in lymphocytes more than 50% of the glycosphingolipid fatty acids contain 22 or more carbon atoms. If the longer acyl chains of the glycolipid were to interdigitate into the inner leaflet, then alterations in the topographical distribution of this lipid could modulate the activity of adenylate cyclase (Curtain, 1979) either through direct interaction with the catalytic unit or by perturbing the fluid properties of the cytosol-facing lipid leaflet. Support for such a view has been obtained from experiments in which addition of gangliosides to rat cerebral cortex membranes substantially increases the activity of basal adenylate cyclase (Partington and Daly, 1979).

V.

PHOSPHOLIPID HEADGROUP COMPOSITION AND ADENYLATE CYCLASE ACTIVITY

To date there is only one clear example of an enzyme that requires a specific type of phospholipid in order to function. This enzyme is P-hydroxybutyrate dehydrogenase, which requires the choline headgroup of phosphatidylcholine in

224


order to bind its coenzyme, NAD (see Houslay et al.. 1975). Many attempts have been made to try to assess whether a specific type of phospholipid is required for the functioning of adenylate cyclase (see Fain, 1978; Ross and Gilman, 1980). However, the inability, at present, to reconstitute homogeneous preparations of receptors, guanine nucleotide coupling protein, and catalytic unit in their appropriate orientations in defined lipid bilayers has meant that, of necessity, indirect approaches or impure preparations have been used. Indeed studies with nonionic-detergent-solubilized preparations which purported to show that acidic phospholipids could reconstitute hormone-stimulated activity have not been able to be reproduced by others (see Ross and Gilman, 1980). Moreover, the reconstitution of such a system would prove to be extremely difficult as, in order to function, all three componets need to be present in their correct, asymmetric orientations within the same liposome. If this is to be achieved by random insertion during reconstitution, we can expect that the number of fruitful reconstitutions is likely to be very low indeed. The effect of specific lipid headgroups on the functioning of an enzyme can be ascertained by a number of different means. These methodologies include detergent-mediated lipid substitution, the insertion of defined lipids into membranes by lipid fusion, phospholipase treatment, supplementing the cell culture medium either with precursors or inhibitors, and also dietary manipulation (see Kimelberg, 1977). However, all of these methods require careful consideration as, in many cases, a number of parameters other than merely headgroup composition can be affected. For example, although certain phospholipases can preferentially attack specific species of phospholipids, lysophospholipids, which are a product of the action of phospholipase A,, can both enhance and inhibit the activity of adenylate cyclase depending upon their concentration (Houslay and Palmer 1979; Lad et al., 1979). Furthermore, alterations of the phospholipid headgroup composition of membranes by dietarylculture media manipulation can lead to adaptive responses which affect the fluidity of the membranes, their fatty acyl chain, and cholesterol content which would lead to alterations in domain structure. Lipid fusion studies involving defined synthetic lipids can be expected to have effects on the cholestero1:phospholipid ratio, domain structure, and 1ipid:protein ratio, all of which have been shown to affect the activity of adenylate cyclase. Thus it is fair to state that at present there is no unambiguous evidence for the involvement of specific headgroup phospholipids in the action of adenylate cyclase. Nevertheless, there are a number of intriguing reports which do suggest that alterations in the levels of certain phospholipid species may affect the functioning of adenylate cyclase. Treatment of liver plasma membranes with a pure preparation of phospholipase C from Bacillus ureus, which hydrolyzes preferentially acidic phospholipids by cleaving off their headgroup to leave the innocuous diglyceride, causes a marked loss of the glucagon-stimulated response. This is in contrast to the action of the

225


enzyme from Clostridium perfringens which hydrolyzes neutral phospholipids preferentially and, even after a substantial hydrolysis, has little effect on the hormone receptors (Rubaclava and Rodbell, 1973). These effects imply that certain acidic phospholipids are necessary for the hormone response. However, both the structure and properties of the membrane are being perturbed, and so the altered activities could be due to second-order effects. Indeed such a loss of activity would need to be shown to be reversed by the fusion of defined acidic phospholipids with the membrane. In contrast to these results rather large changes in the lipid pool content achieved by fusion with defined synthetic phosphatidylcholines or phosphatidylethanolamines caused no dramatic loss of hormone-stimulated activities in liver membranes, although some reduction was observed (Houslay et al., 1976b; Bakardjieva et al., 1979). However, as acidic phospholipids are only a minor component of these membranes, it could well be argued that the enzyme would have a high affinity for any such specific phospholipid. Engelhard et al. (1978) have investigated the effect of supplementing the medium of mouse LM cells with polar headgroup analogs. These led to rather large changes in the ratio of phosphatidylethanolarnine (PE) to phosphatidylcholine (PC) in the plasma membranes of these cells as well as to the incorporation of the analogs into phospholipids. However, the amounts of acidic phospholipids present remained relatively constant. In this case there were marked changes in both basal and prostaglandin E ,-stimulated adenylate cyclase activity, whose increases seemed to correlate well with the PE/PC ratio. However, it remains to be seen whether these changes in activity are due to specific headgroup effects or changes in the physical properties (fluidity) of the lipid domains occupied by the enzyme. In contrast to these studies the incorporation of a solubilized enzyme from brain into liposomes consisting of mixtures of phosphatidylcholine and another lipid has demonstrated that increasing amounts of phosphatidylethanolamine, phosphatidylserine, and cholesterol progressively inhibit enzyme activity (Hebdon et af., 1979). However, such studies are no doubt complicated by the presence of a nonionic detergent, impure lipids and proteins, and the fact that pure phosphatidylethanolamine and cholesterol do not form bilayers into which the enzyme can insert. It would thus seem that the existing evidence for the involvement of phospholipid headgroup composition on the functioning of adenylate cyclase is at best equivocal.

VI.

DISEASE STATES

There have been reports indicating. that the lipid structures/fluidities of the surface membranes from a number of tissues are altered in such disease states as

226


transformation, atherosclerosis, spur-cell anemia, abetalipoproteinemia, muscular dystrophy, cystic fibrosis, diabetes mellitus, and multiple sclerosis (see Kimelberg, 1977, and Houslay and Stanley, 1982, for reviews). In view of the sensitivity of adenylate cyclase to its lipid environment (see above), it might well be profitable to explore whether any abnormalities in the functioning of the adenylate cyclase complex observed in the above disease states are due to alterations in the plasma membrane lipid structure/fluidity. In this regard, Gidwitz et al. ( 1980) have demonstrated that the kinetic properties, thermodependence, and thermostability of adenylate cyclase in chicken embryo fibroblasts are markedly different from the properties of this enzyme obtained from cells transformed with Rous Sarcoma virus. However, upon detergent solubilization the enzyme activity from the two sources behaved identically. This implies that transformation caused a change in the membrane environment of the enzyme which led to marked effects on the enzyme's properties. As it is possible to alter membrane lipid composition and properties by dietary means (see Kimelberg, 1977) and also to influence membrane properties using drugs, it may be fruitful to explore their uses in alleviating certain problems associated with such cellular malfunctions. Clearly, however, rigorous investigations employing the experimental protocols outlined in this article must be performed in these various disease states before assignments as to the significance of membrane perturbation can be made. ACKNOWLEDGMENTS Work in the authors' laboratories was supported by a Medical Research Council (U.K.) project grant (M.D.H); NATO Research Grant RG/218.80 (M.D.H); a Wellcome Trust Travel Fellowship (M.D.H); grants-in-aid from the American Diabetes Association, Southern California Affilitate, Inc.; and NIH Grants AM-21290. AM-28431. and HLIAM-27120 (L.M.G). We would like to thank Elizabeth M. Wright for typing the manuscript. REFERENCES Ahkong, Q. F., Botham, G. M., Woodward. A. W., and Lucy, J. A. (1980). Calcium-activated thiol-proteinase activity in the fusion of rat erythrocytes induced by benzyl alcohol. Biochem. J . 192, 829-836. Amatruda, J. M., and Finch, E. D. (1979). Modulation of hexose uptake and insulin action by cell membrane fluidity. J. B i d . Chem. 254, 26 19-2625. Amir, S. M., Mubrow, N. I . , and Ingbar, S. H. (1981). Phenol, a potent stimulator of adenylate cyclase in human thyroid membranes. Endocrine Res. Commun. 8, 83-95. Bakardjieva, A., Galla, H. J., and Helmreich, E. I . M. (1979). Modulation of the P-receptor adenylate cyclase interactions in cultured Chang liver cells by phospholipid enrichment. Biochemisrry 18, 3016-3023. Bar, H.-P. (1974). On the kinetics and temperature dependence of adrenaline-adenylate cyclase interactions. Mol. Pharmacol. 10, 597-604. Brasitus, T. A., and Schachter, D. (1980). Lipid dynamics and lipid-protein interactions in rdt enterocyte basolateral and microvillus membranes. Biochemistry 18, 2763-2769. Colley, C. M., and Metcalfe, J. C. (1972). The localisation of small molecules in lipid bilayers. FEES Lett. 24, 241-246.


227

Curtain. C. C. ( 1979). Lymphocyte surface modulation and glycosphingolipids. Immunology 36, 805-8 10. Curtain. C. C.. Looney. F. D.. and Smelstorius. J. A. (1980). Lipid domain formation and ligandinduced lymphocyte membrane changes. Biuchim. Biophys. Aria 596, 43-56. Demel, R . A., Jansen. J. W. C. M., van Dijck, P. W. M.. and van Deenan, L. L. M. (1977). The preferential interaction of cholesterol with different classes of phospholipids. Biochim. Bio p h w . Actu 465, 1-10, Dipple. I . , and Houslay, M. D. (1978). The activity of glucagon-stimulated adenylate cyclase from rat liver plasma membranes is modulated by the fluidity of its lipid environment. Biochem. J . 174, 179-190. Dipple. I . . and Houslay, M. D. (1979a). The temperature dependence of adenylate cyclase activity solubilized using various Lubrol detergents. Biochem. B i o p h y . Res. Commun. 90, 663-666. Dipple. I . , and Houslay. M. D. (l979b). Aniphoptericin B has very different effects on the glucagonand fluoride-stimulated adenylate cyclase activities of rat liver plasma membranes. FEBS Lett. 106, 21-24. Dipple, I.. Elliot, K . R. F.. and Houslay, M. D. (1978). Detergents modify the form of Arrhenius plots of 5’-nucleotidase activity. FEBS Lcrt. 89, 153- 156. Dipple. I . , Gordon. L. M., and Houslay, M. D. (1982). The activity of 5’-nucleotidase in liver plasma membranes is affected by the increase in bilayer fluidity achieved by anionic drugs but not by cationic drugs. J . B i d . Chem. 257, I81 I-1815. Engelhard, V. H.. Esko, J . D., Storm. D. R . , and Glaser, M. (1976). Modification of adenylate cyclase activity in LM cells by manipulation of the membrane phospholipid composition in vivo. Proc. Null. Acud. Sci. U.S.A. 73, 4482-4486. Engelhard. V. H., Glaser, M.. and Storm. D. R . (1978). Effect of membrane compositional changes on adenylate cyclase in LM cells. Biochemistry 17, 3 I9 1-3200. Fain. J . N. (1978). Hormones. membranes and cyclic nucleotides. I n “Receptors and Recognition” ( M . F. Greaves and P. Cuatrecasas. eds.). Ser. A.. Vol. 6. pp. 3-62. Chapman & Hall, London. Ghiselli, G . , Sitori. C. R.. and Nicosia, S . (1981). Effect of serum lipoproteins on the adenylate cyclase activity of rat liver plasma membranes. B m h r m . J . 196, 899-902. Gidwitz, S., Toscano, W. A.. Toscano, D. G . . Weber, M. J., and Storm, D. (1980). A comparison between adenylate cyclase solubilized from normal and Rous Sarcoma virus-transformed chicken embryo fibroblasts. Biochim. Biophys. Acrm 627, I - 16. Cordon. L. M . , and Sauerheber, R. D. (1982). Calcium and membrane stability. In “Calcium in Normal and Pathological Biological Systems” (L. J . Anghileri. ed.). CRC. Boca Raton, Florida, in press. Gordon. L. M . , Sauerheber, R. D., and Esgate, J. A . (1978). Spin label studies on rat liver and heart plasma membranes: Effects of temperature, calcium and lanthanum on membrane fluidity, J . Siiprumoi. Srruct. 9. 299-326. Gordon, L. M., Sauerheber, R. D., Esgate. J. A,. Dipple, I.. Marchmont. R. J.. and Houslay, M. D. (1980a). The increase in bilayer fluidity of rat liver plasma membranes achieved by the local anesthetic benzyl alcohol affects the activity of intrinsic membrane enzymes. J. Biol. Chrm. 255, 4519-4527. Gordon. L. M., Dipple, I . . Sauerheber, R. D., Esgate, J . A,. and Houslay, M. D. (1980b). The selective effects of charged local anaesthetics on the glucagon- and fluoride-stimulated adenylate cyclase activity of rat-liver plasma membranes. J . Suprumol. Struct. 14, 21-32. Hanski, E . , Rimon. G.. and Levitzki. A. (1979). Adenylate cyclase activiation by the P-adrenergic receptors as a diffiision-controlled process. Biochemisrry 18, 846-853. Hebdon, G . M.. LeVine, H., Minard. R . B.. Sahyoun. N. E.. Schmitges, C. J.. and Cuatrecasas, P. ( 1979). Incorporation of rat brain adenylate cyclase into artificial phospholipid vesicles. J . Biol. Chem. 254, 10459-10465.

228


Hepp, K. D., Rinninger, J . , Langley, 1.. and Renner, R. (1978). Inhibition of catecholaminestimulated adenylate cyclase in fat cells by local anaesthetics. FEES Lett. 91, 325-328. Higgins, J . A. ( 1981). Biogenesis of endoplasmic reticulum phosphatidylcholine. Translocation of intermediates across the membrane bilayer during methylation of phosphatidylethanolamine. Eiochim. Biophys. Acta 640, 1-15. Higgins, J . A,, and Evans, W. H. (1978). Transverse organisation of phospholipids across the bilayer of plasma-membrane subfractions of rat hepatocytes. Biochem. J. 174, 563-567. Hirata, F.. and Axelrod. J. ( 1978). Enzymatic methylation of phosphatidylethanolamine increases erythrocyte membrane fluidity. Narure (London) 275, 2 19-220. Hirata, F., and Axelrod, J. (1980). Phospholipid methylation and biological signal transmission. Science 209, 1082-1090. Houslay. M. D. (1979). Coupling of the glucagon receptor to adenylate cyclase. Eiochem. Soc. Trans. 7 , 843-846. Houslay, M. D. (1981a). Mobile receptor and collision coupling mechanisms for the activation of adenylate cyclase by glucagon. Adv. Cyclic Nucleoride Res. 14, I 1 I - 119. Houslay, M. D. (1981b). Membrane phosphorylation: A crucial role in the action of insulin, EGF and pp60src. Eiosci. Rep. 1, 19-34. Houslay, M. D., and Palmer, R. W. (1978). Changes in the form of Arrhenius plots of the activity of glucagon-stimulated adenylate cyclase and other hamster liver plasma-membrane enzymes occurring on hibernation. Biochem. 1. 174, 909-919. Houslay, M. D., and Palmer, R. W. (1979). Lysophosphatidylcholines can modulate the activity of the glucagon-stimulated adenylate cyclase in rat liver plasma membranes. Eiochern. J. 178, 2 17-221, Houslay, M. D., and Stanley, K. K. (1982). “Dynamics of Biological Membranes: Influence on Synthesis, Structure and Function.” Wiley, New York. Houslay, M. D., Warren. G. B., Birdsall, N. J . M., and Metcalfe, J. C. (1975). Lipid phase transitions control P-hydroxybutyrate dehydrogenase activity in defined-lipid protein complexes. FEESLert. 51, 146-151. Houslay, M. D.. Metcalfe, J . C., Warren, G. B., Hesketh, T. R., and Smith, G. A. (1976a). The glucagon receptor of rat liver plasma membranes can couple to adenylate cyclase without activating it. Eiochim. Biophys. Acra 436, 489-494. Houslay. M. D., Hesketh, T. R., Smith, G. A., Warren. G . B., and Metcalfe, J. C. (1976b). The lipid environment of the glucagon receptor regulates adenylate cyclase activity. Eiochim. B i o p h p . Acru 436, 495-504. Houslay. M. D., Johannsson, A,, Smith, G. A,. Hesketh, T. R., Warren, G. B., and Metcalfe. J. C. (1976~).On the coupling of the glucagon receptor to adenylate cyclase. Nobel Found. S.ymp. 34,331-344. Houslay, M. D., Ellory, J . C., Smith, G . A . , Hesketh, T. R . , Stein, I. M . , Warren, G. B., and Metcalfe, I. C. (1977). Exchange of partners in glucagon receptor-adenylate cyclase complexes: Physical evidence for the independent mobile receptor model. Biochim. Eiophys. Acra 467, 208-2 19. Houslay. M. D., Palmer, R. W., and Duncan. R. J . S . (1978). The action of the local anaesthetic, benzyl alcohol and the monoamine oxidase inhibitor, clorgyline on the P-hydroxybutyrate dehydrogenase activity of adult and weanling rat brain mitochondria. J. Pharm. Pharmucol. 30, 711-714. Houslay, M. D., Dipple, 1.. and Elliot, K . R. F. (1980a). Guanosine 5’-triphosphate and guanosine 5’-[pa-imido]triphosphate effect a collision coupling mechanism between the glucagon receptor and a catalytic unit of adenylate cyclase. Eiochem. J. 186, 649-658. Houslay, M. D., Dipple, I . , Rawal. S., Sauerbeber, R. D., Esgate, J. A , , and Gordon, L. M. ( 1980b). Glucagon-stimulated adenylate cyclase detects a selective perturbation of the inner


229

half of the liver plasma membrane bilayer achieved by the local anaesthetic prilocaine. Biochem. J . 190, 131-137.

Houslay, M. D., Dipple, I . , and Gordon, L. M. (1981). Phenobarbital selectively modulates the glucagon-stimulated activity of adenylate cyclase by depressing the lipid phase separation occurring in the outer half of the bilayer of liver plasma membranes. Biochem. J . 197, 675-68 I . Insel, P. A., Nirenberg. P., Turnbull, J . , and Shattil, S. J . (1978). Relationships between membrane cholesterol, a-adrenergic receptors and platelet function. Biochemistry 17, 5269-5274. Iyengar, R., Birnbaumer, L., Schulster, D.. Houslay. M. D.,and Michell, R. W. (1980). Mode of membrane receptor-signal coupling. In “Cellular Receptors for Hormones and Nuerotransmitters” (D. Schulster and A. Levitzki, eds.), pp. 55-81. Wiley, New York. Katz, A . M., and Messineo, F. C. (1981). Lipid-membrane interactions and the pathogenesis of ischemic damage in the myocardium. Circ. Res. 48, 1-16. Keirns, J. J., Kreiner, P. W . , and Bitensky, M. W. (1973). An abrupt temperature-dependent change in the energy of activation of hormone-stimulated hepatic adenylyl cyclase. J . Supramol. Struct. I, 368-378. Kimelberg, H. K . (1977). The influence of membrane fluidity on the activity of membrane-bound enzymes. In “Dynamic Aspects of Cell Surface Organisation” ( G . Poste and G. L. Nicolson, eds.), pp. 205-293. Elsevier, Amsterdam. Kimura, N., and Nagata, N. (1977). The requirement of guanine nucleotides for glucagon stimulation of adenylate cyclase in rat liver plasma membranes. J . Biol. Chem. 252, 3829-3835. Kleeman, W., and McConnell, H. M. (1976). Interactions of proteins and cholesterol with lipids in bilayer memhranes. Biochim. Biophys. Acta 419, 206-222. Klein, I . , Moore, L., and Pastan, I . (1978). Effect of liposomes containing cholesterol on adenylate cyclase activity of cultured mammalian fibroblasts. Biorhim. Biophys. Acta 506, 4253. Krdll, J . F., Leshw, S . C., Frolich, M., and Korenman, S. G. (1981). Activation of uterine smooth muscle adenvlste cyclase by guanyl nucleotide. J . Biol. Chem. 256, 5436-5442. Lad, P. M., Preston, M. S . , Welton, F., Nielsen, T. B . , and Rodbell, M. (1979). Effects of phospholipase A2 and filipin on the activation of adenylate cyclase. Biochim. Biophys. Acta 551, 368-381. Martin, B. R., Stein, J. M., Kennedy, E. L., Doberska, C. A,, and Metcalfe, J. C. (1979). Transient complexes. Biochem. J . 184, 253-260. Marinetti, G . V . , and Crain, R. C. (1978). Topology of amino phospholipids in the red cell membrane. J . Supramol. Strucr. 8, 191-213. Melchior, D. C., and Steim, J. M. (1979). Lipid-associated thermal events in biomembranes. In “Progress in Surface and Membrane Science” (D. A. Cadenhead and J . F. Danielli, eds.), Vol. 13, pp. 21 1-289. Academic Press, New York. Montesano, R., Perrelet, A , , Vassali, P., and Orci, L. (1979). Absence of filipin-sterol complexes from large coated pits on the surface of cultured cells. Proc. Nutl. Acad. Sci. U.S.A. 76, 6391 -6395. Neer, E. J . (1976). The size of adenylate cyclase and guanylate cyclase from the rat renal medulla. J . Supramol. Strurt. 4, 5 1-6 I . Orly, J . , and Schramm, M. (1975). Fatty acids as modulators of membrane functions: Catecholamine-activated adenylate cyclase of the turkey erythrocyte. Proc. Nuti. Acad. Sci. U.S.A. 72, 3433-3437. Pairault, J . , Levilliers, J . , and Chapman, M. J . (1977). Human serum lipoproteins activate adipocyte plasma membrane adenylate cyclase. Nature (London) 269, 607-609. Pang, K.-Y. Y., and Miller, K. W. (1978). Cholesterol modulates the effect of membrane perturbers in phospholipid vesicles and biomembranes. Biochim. Biophys. Acra 511, 1-9.

230


Partington, C. R., and Daly, J. W. (1979). Effect of yangliosides on adenylate cyclase activity in rat cerebral cortical membranes. Mol. Pharmucol. 15, 484-491. Pliego, J . A , , and Rubalcava, B. (1978). The effect of temperature on the activity of the adenylate cyclase of liver plasma membranes. Eiochem. Eiophys. Res. Commun. 80, 609-615. Puchwein, G., Pfeuffer, T.. and Helmreich, E. J . M. (1974). Uncoupling of catecholamine activation of pigeon erythrocyte membrane adenylate cyclase by filipin. J . Biol. Chem. 249, 3232-3240. Rasrnussen, H., and Goodman, D. B. P. (1977). Relationship between Ca2+ and cyclic nucleotides in cell activation. Physiol. Rev. 57, 421-509. Rene, E., Pecker, F., Stengel, D., and Hanoune, I. (1978). Thermodependence of basal and stimulated rat liver adenylate cyclase. J . Eiol. Chem. 253, 838-84 I . Rimon, G., Hanski, E., Braun, S . , and Levitzki, A. (1978). Mode of coupling between hormone receptors and adenylate cyclase elucidated by modulation of membrane fluidity. Narure (London) 276, 394-396. Rimon, G., Hanski, E., and Levitzki, A. (1980). Temperature dependence of P receptors, adenosine receptor, and sodium fluoride-stimulated adenylate cyclase from turkey erythrocytes. Eiochemistry 19, 445 1-4460. Ross, E. M., and Gilman A. G . (1980). Biochemical properties of hormone-sensitive adenylate cyclase. Annu. Rev. Biochem. 49, 533-564. Roihman, J. E., and Lenard, J. (1977). Membrane asymmetry. Science 195, 743-753. Rubaclava, B., and Rodbell, M. (1973). The role of acidic phospholipids in glucagon action on rat liver adenylate cyclase. J . Eiol. Chem. 248, 3831-3837. Salesse, R., and Gamier, J. (1979). Effects of drugs on pigeon erythrocyte membrane and asymmetric control of adenylate cyclase by the lipid bilayer. Eiochim. Eiophjs. Acta 554, 102-1 13. Sauerheber, R. D., Lewis, U. J . , Esgate, I. A,, and Gordon, L. M. (1980). Effect of calcium, insulin, and growth hormone on membrane fluidity. A spin label study of rat adipocyte and human erythrocyte ghosts. Eiochim. Eiophys. Acfu 597, 292-304. Schrdmm, M., Orly, J . , Eimerl, S . , and Komer, M . (1977). Coupling of hormone receptors to adenylate cyclase of different cells by cell fusion. Nature (London) 268, 3 10-3 13. Sheetz, M. P., and Singer, S. J . (1974). Biological membranes as bilayer couples. A molecular mechanism of drug-erythrocyte interactions. Proc. Narl. Acad. Sri. U.S.A. 71, 4457-4461. Sinensky, M. (1980). Adaptive alteration in phospholipid composition of plasma membranes from a somatic cell mutant defective in cholesterol biosynthesis. J . Cell Eiol. 85, 166-169. Sinensky, M., Minneman, K. P., and Molinoff, P. B. (1979). Increased membrane acyl chain ordering activates adenylate cyclase. J . B i d . Chem. 254, 9135-9141. Sinha, A. K., Shattil, S. J . , and Colman, R. W .(1977). Cyclic AMP metabolism in cholesterol-rich platelets. J . Eiol. Chem. 252, 3310-3314. Tolkovsky, A. M., and Levitzki, A. (1978). Mode of coupling between the P-adrenergic receptor and adenylate cyclase in turkey erythrocytes. Biochemistry 17, 3795-38 10. Triggle, D. J . (1972). Effects of calcium on excitable membranes and neurotransmitter action. In “Progress in Surface and Membrane Science” (J. Danielli, P. M. Rosenberg, and D. Cadenhead, eds.), Vol. 5, pp. 267-331. Academic Press, New York. Vance. D. E., and de Kruijff, B. (1980). The possible functional significance of phosphatidylethanolamine methylation. Nature (London) 288, 277-278. Warren, G. B., and Houslay, M. D. (1980). Membrane structure and receptor organisation. In “Cellular Receptors for Hormones and Neurotransmitters” ( G . Schulster and A. Levitzki, eds.), pp. 29-54, Wiley, New York. Warren, G. B., Houslay, M. D., Metcalfe. J . C., and Birdsall, N. J . M. (1975). Cholesterol is excluded from the phospholipid annulus surrounding an active calcium transport protein. Nafure (London) 255, 684-687.


23 1

Whetton. A. D., and Houslay, M. D. (1980). The effect of vinblastine on the glucagon, basal and GTP-stimulated states of the adenylale cyclase from rat liver plasma membranes. FEES Leu. 111, 290-294. Wolff. J . . and Jones. A . B. (1970). Inhibition of hormone-sensitive adenylate cyclase by phenothiazines. Proc. Nail. Acad. Sci. U.S.A. 65, 454-459.


CURRENT TOPICS IN MEMBRANES AND TRANSPORT. VOLUME I 8

The Analysis of Interactions between Hormone Receptors and Adenylate Cyclase by Target Size Determinations Using Irradiation Inactivation B . RICHARD MARTIN Department of Biochemistry Universit.v of Cambridge Cambridge. England

I. If. 111. IV .

Irradiation Inactivation: General Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Practical Considerations in Irradiation Inactivation Studies on Membranes ......... ................. ............

Membrane Adenylate Cyclase . . . . . . . . . . . . . . ....................... Model of Hormone Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VI . Effects of Fluoride . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . VII. Evaluation of the Model in Relation to the Results of Other Approaches . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V.

1.

233 236 231 239 241 246 248 253

IRRADIATION INACTIVATION: GENERAL CONSIDERATIONS

The method of irradiation inactivation as an approach to the determination of the molecular weight of proteins had been used as long ago as the 1930s. The first systematic study was conducted by Kepner and Macey in 1968. They examined a number of proteins whose molecular weights had already been determined by conventional methods. In most cases the activity of the enzyme decayed as a monoexponential with increasing doses of electron irradiation, and they were able to determine empirically the expression Molecular weight = (6.5 233

X

1OS)/D37 Copyright Q 1983 by Academic Press, Inc. All rishi? of re~roduclionin any form rcwrved.

ISBN 0-12-153318-2

234

B. RICHARD MARTIN

where D37 is the radiation dose in Mrad at which 37% of the original activity remains. A theory to account for this behavior was formulated-the single hit hypothesis. This states that a single hit by one electron is sufficient to destroy the activity of the protein completely. Thus for a given dose of electron radiation the likelihood of a protein molecule being hit will depend upon its size and, accordingly, upon the molecular weight. This method of protein size determination was particularly attractive as a means of examining association and dissociation of proteins in membranes. All that is required is a measure of the activity of the protein under study. There is no requirement for the protein to be pure or for the protein to be in solution. Thus, the method offers the opportunity to examine the size of proteins in intact biological membranes. Before applying the method to adenylate cyclase in plasma membranes, however, a number of potential problems need to be considered. At the time we started these experiments there were very few determinations of molecular weights of adenylate cyclase or indeed of other membrane integral enzymes available for comparison with the results of the target size analysis. Many of the values which had been determined for adenylate cyclase largely depended on gel filtration experiments using preparations solubilized in nonionic detergents (Welton et af., 1977, 1978). Under these conditions there is a tendency for large polydisperse aggregates to form, giving overestimates of the molecular weight. The possibility must therefore be considered that the target size determined for a membrane integral protein may reflect the molecular weight of not only the protein but also some attached lipid. In fact, subsequent molecular weight determinations for adenylate cyclase by a variety of methods have yielded values which are in quite good agreement with the values obtained by irradiation inactivation (Neer, 1974; Newby et af., 1978). In any case, we are more interested in alterations in the molecular weight of the components in response to specific effectors than in the absolute values of the molecular weight. A second, more serious consideration relates to the validity of the method in determining changes in states of associations. Where multisubunit enzymes have been examined by irradiation inactivation the target size sometimes corresponds to the whole enzyme, which is obviously the state of affairs which is desirable in this case, and sometimes corresponds to the monomer molecular weight (for review see Kempner and Schlegel, 1979). If this is the case with adenylate cyclase, the method will not yield any useful results. The only way around this problem is to perform the experiments on an empirical basis and to see if the results conform to a rational model which can be supported by other types of data. There are two types of approach to the study of protein-protein interaction by target size analysis. In the first case the preparation, in our case rat liver plasma membranes, is irradiated in the absence of any effectors or in the basal state. The

TARGET SIZE ANALYSIS

235

activity under study, for example, the catalytic activity of adenylate cyclase, is then determined in the usual way, and the target size for the basal state can be calculated. The activity can also be determined in the presence of an activator such as a hormone. If the increased activity does not depend upon the association of a second component, in other words, if in this case the binding site for the hormone is on the same protein unit as the catalytic site of adenylate cyclase then the target size will not change. If, on the other hand, the increased activity in the presence of the effector does require the association of a second distinct component, then the target size will increase. This is because the increased activity depends upon both the catalytic unit and the receptor, and both are affected by the irradiation. It is important to stress, however, that the extent of the increase in target size will depend upon the extent to which the activity is dependent on the second component, that is, the larger the activation the larger the increase. The increase will also depend on the relative size of the two components. If, for example, the hormone receptor is small relative to the catalytic unit the loss of receptor activity will also be less for a given radiation dose and the increase in target size will be small. In theory this approach should also be applicable to twocomponent systems where the regulatory component results in an inhibition, although in this case the addition of the inhibitor in the assay will result in a reduction rather than an increase in the target size. However, I am not aware of any publications in which irradiation inactivation has been used to examine inhibitory processes. It should be stressed that the target size increases determined under these conditions are not a direct measure of a change in molecular weight but simply reflect the fact that more than one separate component is involved in the change in activity. In the second type of general approach the membranes are preincubated with the effector prior to irradiation. In this case the preincubations should, so far as possible, be identical to the conditions which are normally employed in the assay of the activity which is under study, although minor modifications may be necessary. For example, it has been found that the presence of thiol reagents such as dithiothreitol during irradiation leads to extensive nonspecific loss of activity of many enzymes. In experiments in which the system is exposed to the effector before the irradiation, any changes in target size should reflect an actual change in the molecular weight of the component responsible for the activity which is measured. However, in interpreting this type of result care must be taken to ensure that the conditions under which the enzyme is assayed do not lead to any further apparent changes in target size resulting from the involvement of another component. In general the interpretation of the results of the first type of approach is relatively straightforward. The pretreatment approach yields more interesting information but more care is required in the interpretation of the results.

236

8 . RICHARD MARTIN

II. PRACTICAL CONSIDERATIONS IN IRRADIATION INACTIVATION STUDIES ON MEMBRANES A number of approaches have been used in the preparation of plasma membranes for irradiation. Schlegel et al. (1979) irradiated membranes in aqueous suspension. In this case the suspension must be frozen, and they conducted their irradiations at - 170°C using liquid nitrogen as the freezing agent. This approach has the advantage that there is no possibility of a damaging rise in temperature due to the irradiation which may result in a loss in activity due to heating rather than to the impact of the electrons. It also has the advantage that recoveries of activity in nonirradiated samples in frozen suspension are likely to be close to loo%, and since the preparation simply has to be allowed to thaw the preparation of the samples for assay is simple. The main disadvantage with the use of frozen samples is that with a temperature as low as - 170°C the expression derived by Kepner and Macey ( 1968) no longer holds and a conversion factor of 1.7 has to be used. This results in the need for very high radiation doses of the order of 20-30 Mrad to achieve 90% inactivation of quite large proteins of molecular weights of 100,000or more. One then has to be concerned about'the calibration of the radiation dose. This is most commonly done by measuring changes in the optical density of pieces of Perspex of standard thickness and quality. The operative range for this type of calibration is 0-3 Mrad and the advisability of extrapolating the calibration to doses 10 or 20 times higher is doubtful. As discussed above, a general cause for concern in the application of the method is whether or not association of different components is always reflected in an increase in the target size. This seems more likely to be a problem when the irradiation is carried out at very low temperatures. We have preferred to use the alternative approach of freeze-drying the membranes after preincubation when appropriate and before irradiation. At the end of the preincubation the preparation is rapidly frozen by immersing the tube in liquid nitrogen and is then lyophilized. Obviously a critical first step is to ensure that the bulk of the activity of a preparation which has not been irradiated can be recovered after the rehydration of the membranes. In the case of adenylate cyclase we were able to recover more than 90% of the enzyme activity routinely. It is important, however, that the freeze-drying process should be carried out very carefully since in our experience any loss of vacuum during the drying process leads to substantial and variable loss of activity. When freeze-dried membranes are used the irradiation is carried out at about 10"-15°C and the apparatus can be cooled by a flow of chilled air. It should be borne in mind, however, that large irradiation doses of 10 Mrad or more may lead to unacceptable increases in temperature. In either approach oxygen should be excluded from the system. In the case of irradiation in aqueous suspension the medium should be flushed with nitrogen before use. In the case of dry samples the ideal

237


solution is to irradiate under vacuum, or, if this is not possible, under a nitrogen atmosphere. In the majority of cases the irradiations will be carried out using a linear accelerator belonging to a radiotherapeutics facility. One is therefore dependent upon the goodwill of technical staff whose prime concern is the treatment of cancer patients, so to some degree the simpler the irradiation protocol that can be used the better.

111.

ANALYSIS OF DATA

Typical irradiation plots are shown in Fig. 1. In this case all the plots are linear, which indicates that the protein which is responsible for the activity under study is present as a single homogeneous species. If this is not the case a curved plot will result (Fig. 2). Great care is needed in the analysis of this type of data. It may be tempting to use a computer to fit the data to two straight lines to determine the target sizes of the two species present. However, a number of problems should be borne in mind. First, it is necessary to be satisfied that there are only two species and not three or four, in other words, that the curve is not fitted better by assuming three or four species. In order to do this with any degree of confidence the curve must be very well defined indeed, with a very large number of data points. Care must be taken in the choice of the mathematical expression used to fit the curve and it is probably advisable to take expert advice. If other data are available to give an independent measure of the target size, that is, target sizes obtained from linear plots obtained under other conditions, the situation is more straightforward. In this case one can ask the question, Does the nonlinear plot give a good fit to a combination of two linear plots whose parameters are known'? Of course one should have a biochemical rationale for the assumption that the curve represents these two particular species. It should always be borne in mind, however, that given a particular set of instructions a computer will usually batter a set of data into some sort of submission but the results may not be particularly useful. In dealing with nonlinear plots alternative explanations should be considered as well as the existence of more than one species. One should not lose sight of the fact that a preparation such as a plasma membrane contains many enzymes, all of which will be affected by irradiation. The possibility that this may give rise to spurious effects should be considered. In the case of adenylate cyclase, for example, care must be taken to ensure that the ATP regenerating system in the assay is adequate to maintain ATP levels throughout the assay time course. If this is not the case, the fact that the ATPase activity in the plasma membranes will also be reduced by irradiation may cause an apparent curvature in the irradiation plot. This is particularly likely to be the case if very low concentrations of ATP are used so as to increase the specific activity of the ATP used in the assay in an

238

8.RICHARD MARTIN

30' C

30'C

(315,000)

I , 0

.I

1

2

3

4

5

I,...., 0

1

2

o'c

3

4

5

0%

0

1

2

3

4

5

Mrad

Mrad

FIG. 1. Effect of glucagon and p(NH)ppG on irradiation inactivation of adenylate cyclase. Rat liver plasma membranes were preincubated in the presence of 25 mM Tris-HCI buffer, pH 7.4, 0. I mM ATP, 0. I mM cyclic 3'5'-AMP, 10 mM MgC12, 5 mM phosphocreatine, and 5 units of creatine kinase for 10 minutes at either 30 or 0°C. Further additions were (0) no additions, (0)p(NH)ppG (0.1 mM), (0) glucagon ( I pM), and glucagon (1 pht) + p(NH)ppG (0.1 mM). After irradiation adenylate cyclase assay was determined at 30°C for 20 minutes in a medium containing 25 mM Tris-HCI, pH 7.4,0.5 mM ATP, 1 FCi of [a-32P]ATP, 0.1 mM cyclic 3'5'-AMP, 10 mM MgCI,, I mM dithiothreitol, 5 mM phosphocreatine, and 5 units of creatine kinase. Numbers in parentheses indicate target sizes. From Martin et al. (1979).

(m)

239


30°C

30°C

No Additions

(331,000)

,

3

I

1

2 3 M rad

4

5

I

1

0

1

1

2 3 Mrad

L

5

FIG. 2. Effect of GTP on irradiation inactivation of adenylate cyclase. Rat liver plasma membranes were preincubated at 30°C in the presence of (0) no additions, ( 0 )GTP (1 mM), and ((?) GTP ( 1 mM) + glucagon ( 1 w). Other conditions were as described in the legend to Fig. I . From Martin et al. (1979).

attempt to make the detection of very low adenylate cyclase activities easier. In general, it is desirable to conduct irradiations under extreme conditions. If an activation is under study it should be ensured that activation is maximal. This increases the chances of generating a single homogeneous species and hence a linear irradiation activation plot which will be much easier to interpret. Conditions which give a partial effect and hence nonlinear plots are also much easier to interpret in the light of additional information derived from extreme states.

IV. THE APPLICATION OF TARGET SIZE ANALYSIS TO RAT LIVER PLASMA MEMBRANE ADENYLATE CYCLASE The approach of irradiation inactivation was first used on rat liver plasma membrane adenylate cyclase by Houslay and his colleagues (Houslay et al., 1977). The aim was to see if the method would yield any evidence in support of the mobile receptor hypothesis. They examined the effect of glucagon on both the target size of adenylate cyclase catalytic activity and on the target size of the specific glucagon binding activity which was taken to reflect the amount of functional glucagon receptor. When liver plasma membranes were irradiated in the absence of any effectors the target size for adenylate cyclase in the presence of either fluoride or guanylyl imidodiphosphate [p(NH)ppG] was 160,000. In the presence of glucagon, however, added in the assay after the irradiation, the target

240

8. RICHARD MARTIN

size increased to 389,000. Under the same conditions the specific glucagon binding activity had a target size of 217,000. The increase in target size on addition of glucagon after the irradiation indicates the involvement of a second component in the glucagon activation other than the component containing the catalytic site. When the plasma membranes were exposed to glucagon at 0°C prior to irradiation the target size for both the catalytic activity and the glucagon binding activity increased to about 380,000. Thus the data were consistent with the binding of the hormone to the receptor causing the association of the receptor with the catalytic unit of adenylate cyclase. A particularly attractive aspect of this study was that the activity of both of the components could be determined independently, the catalytic unit by measuring adenylate cyclase activity and the hormone receptor by measuring glucagon binding. The results of both types of measurements were internally consistent. This study provided convincing support for the mobile receptor hypothesis in which the binding of hormone promotes the association of the hormone receptor with the catalytic unit and the complex formed represents the activated state. In the initial study of Houslay and his colleagues, all the preincubations with glucagon were carried out at 0°C in the absence of any other effectors and in the absence of the substrates necessary for catalytic activity. In particular, no attempt was made to examine the effects of guanine nucleotides which are essential for the full expression of hormone stimulation of adenylate cyclase. We therefore decided to examine the effects of glucagon and also of guanine nucleotides under conditions in which we could be certain that activation of adenylate cyclase had taken place (Martin et al., 1979). To this end all the pretreatments of the plasma membranes were carried out under conditions identical with those routinely used for the determination of adenylate cyclase activity with the single exception that dithiothreitol was omitted from the medium. Under these conditions we were able to essentially repeat Houslay’s original observations of the effects of pretreatment of membranes with glucagon at 0°C. In this case the target size of adenylate cyclase catalytic activity increased from 300,000 to 470,000, again reflecting the association of the hormone receptor with the catalytic unit of the enzyme. At 0°C the addition of either GTP or p(NH)ppG had no effect on the target size in either the presence or absence of glucagon, and hence apparently no effect on the association of the hormone receptor with the enzyme. The effect of glucagon and guanine nucleotides on pretreatment at 30°C presented a very different picture, however. Pretreatment with p(NH)ppG caused a reduction in the target size from 300,000 to 200,000 and this was not affected by the presence of glucagon (Fig. 1). Thus, in the presence of p(NH)ppG which in liver plasma membranes is capable of activating adenylate cyclase to its maximal extent, there was no evidence of any association of the hormone receptor with the catalytic unit of adenylate cyclase. It also appeared that some component of the system dissociates. We then examined the effect of the natural guanine nucleotide GTP

241


at 30°C. On preincubation in the presence of GTP alone we found a nonlinear inactivation plot indicating that more than one species was responsible for the activity of adenylate cyclase. Preincubation in the presence of GTP together with glucagon gave a linear plot indicating a reduction in target size from 300,000 to 200,000, comparable to that seen with p(NH)ppG either alone or in the presence of glucagon (Fig. 2 ) . Thus it appears that GTP promotes the same dissociation observed with p(NH)ppG but that the effect is incomplete. At this stage we attempted to devise a model which would account for the data and provide a framework for the design of further experiments.

V.

MODEL OF HORMONE ACTION

The model for the effects of glucagon and guanine nucleotides on the target size of adenylate cyclase is outlined in Fig. 3 . We suggest that the effect of p(NH)ppG is to cause the dissociation of a regulatory subunit (G) from the catalytic subunit (C). The catalytic subunit is assumed to be fully activated in the

b

GI ucagon

I

FIG. 3. Model of the mechanism of activation of adenylate cyclase by glucagon and guanine nucleotides. The figure describes the model proposed for the alterations in the aggregation state of adenylate cyclase during activation by (a) guanine nucleotides and (b) guanine nucleotides in the presence of glucagon. The glucagon receptor is represented by R, the catalytic unit by C, and the regulatory unit by C. From Martin el al. (1979).

242

0. RICHARD MARTIN

dissociated state and the regulatory subunit is assumed to be a guanine nucleotide binding protein. The dissociation promoted by p(NH)ppG is unaffected by the independent hormone receptor which, in the absence of glucagon, does not interact with the (GC) dimer (Fig. 3a). The effect of p(NH)ppG is assumed to be essentially irreversible and complete. In contrast, the effects of GTP alone are reversible and incomplete in that an equilibrium is established between the activate dissociated state (C) and the inactive associated state (GC); this is consistent with the nonlinear irradiation inactivation plot and with the lower degree of activation observed in the presence of GTP than that observed with p(NH)ppG (Table 1). All the effects of guanine nucleotides showed a marked dependence on temperature and did not occur at 0°C. The model for the action of glucagon is summarized in Fig. 3b. In the presence of glucagon the receptor (R) associates with the inactive complex (GC) to form a ternary complex (RGC). This effect is not temperature dependant and can be detected at 0°C. In fact, for reasons to be discussed below, it is more readily detected at 0°C when the effects of guanine nucleotides are not present than at 30°C. At 30°C in the presence of either p(NH)ppG or GTP the complex (RGC) dissociates, releasing free (C) and hence activating adenylate cyclase. Glucagon and GTP together are able to promote complete dissociation of the complex (RGC) to release the activated state ( C ) , whereas GTP alone will promote only TABLE I EFFECTSOF GUANINE NuCLEOTlDES A N D GLUCACON ON ADENYLATE CYCLASE I N RAT LIVERPLASMAMEMBRANES".~

Additions to assay No preincubation None GTP ( I mM) p(NH)ppG (0.1 mM) Glucagon ( 1 pkf) Glucagon ( I pkf) + GTP ( I mM) Glucagon ( I pM) + p(NH)ppG (0.1 mM)

A.

Adenylate cyclase (nmole/20 minutes per mg of protein)

0.465 1.40 3.69 1.16 4.64 5.01

k

0.012

f

0.047

5.11 5.15

f 0.136

2 0.126 f 0.014

2 0.008 2 0.057

2 0.104

From Martin et a/. (1979). Adenylate cyclase was determined in rat liver plasma membranes as described in the text. Membranes either without pretreatment (A) or pretreated with p(NH)ppG (0.I M ) (B) were incubated for 10 minutes at 30°C in MgC12 (10 mM). Results are means 5 SEM for three observations. 1'

h


243

partial dissociation of the complex (GC). Thus, in common with Cassel and Selinger and many other workers, we propose that the effect of the hormone receptor complex is to facilitate the activation of adenylate cyclase by GTP. At this stage a number of features of the model were undefined and at least one important assumption had been made. In this study only the target size of adenylate cyclase catalytic activity was determined, so that we were unable to say whether the ternary complex dissociates completely to give free (C), (G), and (R) or whether component (G) and the receptor (R) remain associated as a complex (GR) and a further dissociation event is required to regenerate free (R) and (G). Under physiological conditions the free regulatory component (G) must eventually become available to regenerate the inactive complex (GC), and in the absence of any evidence one way or the other we made the simpler assumption that the dissociation takes place in one step to release free (C), (G), and (R). The second, more important assumption relates to the nature of the regulatory subunit which dissociates. As described in the first article, there is a large body of evidence to suggest that the adenylate cyclase system contains a protein component which mediates the activation of the enzyme by guanine nucleotides and also has a GTPase activity which is responsible for the decay of the activation. We made the assumption that the regulatory subunit (G) which dissociates under the influence of guanine nucleotides can be identified with this component of the system. It should be pointed out that there is no direct evidence for this assumption. To obtain such evidence we would have to conduct irradiation inactivation analysis of the guanine nucleotide binding activity associated with the activation of adenylate cyclase and of the specific hormone-sensitive GTPase activity. Unfortunately, in rat liver plasma membranes this is not technically possible. Together with a number of other groups (Cassel and Selinger, 1977; Lin et al., 1978), we have found that the background of nonspecific GTPase activity and nonspecific guanine nucleotide binding activity in rat liver plasma membranes is far too high to allow the measurement of the specific activities associated with the regulation of adenylate cyclase. Thus it was not possible to conduct an experiment analogous to those performed by Houslay et al., who were able to measure both adenylate cyclase catalytic activity and the specific glucagon binding activity. This means that we cannot exclude the possibility that the binding site for guanine nucleotides is on the same component (C)as the catalytic site of adenylate cyclase or indeed that both the catalytic unit (C) and the regulatory unit ( G ) contain a binding site for guanine nucleotides. The model gave rise to a number of predictions, some of which were very simple to test. It appears that p(NH)ppC alone can cause complete dissociation of the regulatory subunit from the catalytic subunit, and we propose that the free catalytic subunit (C) represents the fully activated state of the enzymes. It follows from this that glucagon should not cause any further increase in the activity of the enzyme in the presence of p(NH)ppG. At first sight this did not appear to

244

B. RICHARD MARTIN

be the case. If glucagon and p(NH)ppG were added simultaneously at the beginning of the assay, then glucagon apparently caused a small but consistent and significant increase in activity (Table IA). However, if the liver plasma membranes were preincubated in the present of p(NH)ppG for 10 minutes, the period used for preincubations prior to irradiation, the addition of glucagon produced no further activation (Table IB). Rodbell et al. (1975) showed that there was a marked lag in the response of rat liver plasma membrane adenylate cyclase to p(NH)ppG and that this lag was abolished in the presence of glucagon. The final extent of the activation was not, however, affected by the hormone. It seems, therefore, that glucagon is able to increase the rate of activation by p(NH)ppG but not the extent of the activation, and that our preincubation was sufficiently long to allow full activation by the guanine nucleotide. The target size analysis implies that GTP in the presence of glucagon is as effective as p(NH)ppG in promoting the dissociation of the regulatory subunit. These conditions should therefore be equally effective in activating adenylate cyclase. Table I A shows that this was the case. A further prediction was that at 0°C there should be little or no activation by any effector or combination of

FIG. 4. Effect of variation of temperature on the activation of adenylate cyclase by glucagon and guanine nucleotides. Adenylate cyclase activity was determined as described in the legend to Fig. 1 at GTP ( I M).([7)p(NH)ppG (0.1 mM). ( 0 )glucagon various temperatures in the presence of (0) ( I pM). glucagon ( I p W ) + GTP (I mM), and (A) glucagon ( I p M ) + p(NH)ppC (0.1 mM). From Martin el a / . (1979).

(m)

245


effectors, since the target size analysis gave no indication of any dissociation to release the free active catalytic unit at 0°C. The maximum activation at 30°C was 800%, observed with glucagon together with p(NH)ppG. At 0°C the corresponding activation was 50%. This would represent a proportional dissociation to yield free (C) which would be well below the limits of sensitivity of the target size analysis (Fig. 4). The model suggests that at 30°C in the absence of guanine nucleotides glucagon should cause an increase in target size reflecting the formation of the ternary complex (RGC), which in the absence of guanine nucleotides should persist. In fact, under these conditions, we observed a nonlinear plot which suggested the presence of a mixture of several different species (Fig. 5). This is to be expected, since glucagon alone activates the enzyme to a limited extent (Table IA). This is probably due to the presence of small amounts of GTP contaminating either the plasma membrane preparation or the ATP used as substrate in the assay. It is well established that the complete purification of the system from contaminating GTP is very difficult (Kimura et al., 1976). More recently, however, a number of ATP preparations have become available with greatly improved purity with respect to GTP compared to the preparations available at the time of this study. If we were to repeat this experiment now, the highmolecular-weight species corresponding to the ternary complex should predominate to a much greater extent.

Glucagon

GT P

.?

\

bGlucagon

\

ol 0

1.0.

Control

I

0

1

2

3

4

5

Mrad

FIG. 5 . Effect of glucagon on irradiation inactivation of adenylate cyclase at 30°C. Rat liver plasma membranes were preincubated at 30°C as described in the legend to Fig. I in the presence of (0)no additions, (0) glucagon (1 @), and ( 0 )glucagon ( 1 @) + GTP ( 1 mM). Adenylate cyclase activity was determined as described in the legend to Fig. 1 .

246

B. RICHARD MARTIN

VI.

EFFECTS OF FLUORIDE

The mechanism of activation of adenylate cyclase by F- ions i s still not entirely clear. One point which has been established, however, is that the effect appears to be mediated by the same protein component of the system that mediates the action of guanine nucleotides. This conclusion is based on the studies of Gilman and his colleagues (Ross er al., 1978), who showed that cyc- S49 lymphoma cells, where the adenylate cyclase lacks the capacity to respond to guanine nucleotides, have also lost the capacity to respond to fluoride. When we pretreated rat liver plasma membranes with F- in the presence of an adenylate cyclase assay medium at 30"C, the target size was reduced from 300,000 to 200,000 (Fig. 6, Martin et al., 1980). This effect was exactly comparable with the effect of p(NH)ppG or of GTP in the presence of glucagon. It seemed likely, therefore, that the activation of the enzyme by F- ions involved the dissociation of the same regulatory subunit which was released on activation by guanine nucleotides. In contrast to the effects of guanine nucleotides, however, the reduction in target size with F- was observed when the preincubation was carried out at 0°C as well as at 30°C (Fig. 6). Since we proposed that the reduction in target size reflects a necessary step in the activation mechanism,

2.0-

--

>1

> .c

"1.5-

a & =

D

,o 1.0-

No Addltions

No Additions (329.000)

0

0

1

2

3 M rad

4

5

6

b

1

2

3

4

5

6

Mrad

FIG.6 . Effect of F- ions on the irradiation inactivation of rat liver plasma membrane adenylate cyclase. Rat liver plasma membranes were preincubated as described in the legend to Fig. 1 at 0 and or presence (0)of 10 mM NaF. Adenylate cyclase activity was deterat 30°C in the absence (0) mined after incubation as described in the legend to Fig. I . From Martin et a/. (1980).

247


TABLE I1 EWM-TOF TEMPEKATURt O N T H t ACTIVATION Ob R x r LIVERPLASMA MEMBRANE ADtNYl.ATh C Y C L A SB~Y F- "." Adenylate cyclase activity (nmolei10 minutes per mg of protein) Temperature ("C)

No addition

10 mM NaF

Increase in activity (9%)

0 30

0.01 19 f 0.0004 0.365 t 0.016

0.09.5 !I0.0014 3.587 2 0.067

no0 980

From Martin rr a / . (1980). Rat liver plasma membranes were incubated in 0.1 ml of assay medium as described in the text with further additions as given in the table. Incubations were for 40 minutes with 0. I mg of membrane protein at 0°C and for 10 minutes with 0.04 mg of membrane protein at 30°C. Results are means 2 SEM for three parallel incubations.

then F- ions in contrast to p(NH)ppG should be capable of activating adenylate cyclase at 0°C. Table I1 shows that this was in fact the case. The extent of activation in the presence of F- ions at 0°C was comparable with the activation observed at 30°C. There was a very marked temperature dependence of the adenylate cyclase activity in e.ither the presence or absence of F- ions. In both cases, the increase in activity from 0" to 30°C was more than 30-fold. Similar results have been reported by Houslay et al. (this volume), in a much more extensive study of the effects of temperature on adenylate cyclase activity, and other membrane integral enzymes have been shown to have a similar marked dependence of temperature (Houslay and Palmer, 1976). In common with other workers (for review see Bimbaumer, 1973), we observed that the activation of adenylate cyclase by F- ions was consistently about 20 or 30% less than the activation observed with p(NH)ppG. Since F- ions were as effective as p(NH)ppG in promoting a reduction in target size and since we proposed that the low-molecular-weight species represents the fully activated state of the enzyme, we would expect the two effectors to produce the same activation. When the effects of varying the concentration of F- were examined, it was found that the response of adenylate cyclase activity was biphasic. The activity first increased with increasing F- concentration and then decreased (Fig. 7). If the adenylate cyclase was first fully activated by preincubation in the presence of glucagon and p(NH)ppG before the addition of F - , then all concentrations of F- were inhibitory (Fig. 8). It seems, therefore, that F- has two effects, one to activate and one to inhibit. The lower activation observed in the presence of F- ions in comparison to p(NH)ppG can therefore be explained by the inhibitory effect's becoming significant at F- concentrations at which the

248

B. RICHARD MARTIN 5r

10

20

30

40

[F-] (mM1

FIG.7. Effect of F- ions on rat liver plasma membrane adenylate cyclase activity. Adenylate cyclase activity was determined as described in the legend to Fig. I for 20 minutes at 30°C in the presence of varying concentrations of NaF. From Martin er nl. (1980).

activation effect is not complete. The data are therefore consistent with the view that the activation of adenylate cyclase by F- involves the release of the same regulatory subunit that is released on activation by guanine nucleotides.

VII. EVALUATION OF THE MODEL IN RELATION TO THE RESULTS OF OTHER APPROACHES In conclusion, we should consider how consistent our model, derived from target size analysis, is with data derived by other workers by more conventional approaches. The model makes two major suggestions. The first is that the association of the hormone receptor with the catalytic unit is transitory and that, as the activation process is completed by the involvement of a guanine nucleotide, the receptor dissociates. An implication of this proposal is that the receptor will then be available to promote the activation of further adenylate cyclase complexes. Its effectiveness in doing this will depend upon the relative rates of interaction with further complexes compared to the rate of dissociation of the bound hormone which will, of course, render the receptor ineffective as an activator of the enzyme. The second major proposal was that the activation of

249


-G E

0

4-

\

c

c

e

a

0'

' 0

10

20

30

40

[F-] ( m M )

FIG.8. Effect of F- ions on rat liver plasma membrane adenylate cyclase after preactivation with glucagon and p(NH)ppG. Rat liver plasma membranes were irradiated for 10 minutes at 30°C in 25 mM Tris, pH 7.4, in the presence of glucagon ( I M),p(NH)ppG (0.1 mM), and MgCI2 (10 mM). Adenylate cyclase activity was determined at the same concentrations of glucagon and p(NH)ppG and varying concentrations of NaF as described in the legend to Fig. I . From Martin efal. (1980).

adenylate cyclase by guanine nucleotide involves the release of a regulatory unit. We also suggested that this regulatory unit corresponds to the guanine nucleotide binding protein, whose existence has been established by a large number of different groups. The suggestion that the association of the hormone receptor is transient helps to explain a number of earlier observations and is also supported by the work of other groups which was conducted at about the same time as our original study. Rodbell and his colleagues examined the relationship between the specific binding of glucagon to rat liver plasma membranes and the activation of adenylate cyclase (Rodbell et al., 1974). They found that in the absence of GTP, the rate of dissociation of glucagon from the plasma membranes was very slow, and that accordingly the apparent binding affinity was very tight. Our model predicts that in the absence of GTP and in the presence of glucagon, the predominant species will be the ternary complex of receptor, regulatory unit, and catalytic unit (RGC). We suggest that the rate of dissociation of glucagon from this complex is slow. In the presence of GTP, the ternary complex will break down to release the free hormone receptor (R). We suggest that the rate of dissociation of glucagon

250

B. RICHARD MARTIN

from the free receptor is relatively fast. This would explain the effects of GTP on the binding of glucagon to the hormone receptor. A more careful consideration of Rodbell’s data suggests that this may be an oversimplification, since the GTP concentration dependence for the effect on glucagon binding and the activation of adenylate cyclase differs. This might suggest that the complex of receptor and regulatory unit (RG) does persist after dissociation of the ternary complex (RGC). The same group also found that at physiological concentrations of glucagon, the addition of GTP resulted in a marked increase of adenylate cyclase activity, while the amount of glucagon bound was reduced to 20% or less (Rodbell et al., 1974). To put this another way, maximal activation by glucagon in the absence of added GTP or more likely in the presence of very low concentrations of GTP, since the problem of GTP contamination had not been recognized when these experiments were done, requires 100% occupancy of the glucagon receptors. In the presence of optimal GTP, maximal activation by glucagon, which was also greatly increased in comparison to the activity in the absence of GTP, required only 20% occupancy of the glucagon receptors. It seems likely, therefore, that in the presence of GTP, one hormone receptor complex is able to promote the activation of several adenylate cyclase catalytic units. When the complex (RGC) is dissociated by the action of GTP at the same time as the activation process is completed, the free hormone receptor complex will be available to promote the activation of further catalytic units. The concentration of free hormone receptor complex which is available to do this will be dependent upon the concentration of free hormone to which the receptors are exposed. This model proposing the transient nature of the association of the hormone receptor with the catalytic unit is very similar to the collision coupling model proposed by Tolkovsky and Levitzki (1978). They came to essentially the same conclusions using a very different experimental approach. Their model was based upon a detailed study of the kinetics of activation of turkey erythrocyte plasma membrane adenylate cyclase by catecholamines. Houslay and his coworkers have made use of the effects of phase transitions in rat liver aqd hamster liver plasma membranes to examine the state of association of the glucagon receptor with the catalytic unit of adenylate cyclase (Houslay et al., 1980). This study gave rise to the same conclusion, that, in the presence of GTP, the association of the receptor with thd’catalytic unit does not persist. Both these studies are described in detail elsewhere in this volume. Lefkowitz and his colleagues have examined the effects of GTP on the binding affinity of P-adrenergic agonists and antagonists (Lefkowitz et al., 1981). In the case of P-adrenergic agonists, they found a situation similar to that observed by Rodbell and his co-workers in their studies of glucagon binding. In the absence of GTP, the agonists bound with an apparently high affinity which was reduced by the addition of GTP. Adrenergic antagonists, however, displayed only the low-affinity binding, and the binding affinity was unaffected by GTP. Antago-


251

nists are thought to compete for binding to the recognition site for a hormone on a hormone receptor but to be incapable of producing the appropriate conformational change in the receptor to produce the appropriate response. Lefiowitz and his colleagues suggested that agonists are capable of promoting an interaction of the hormone receptor with the GTP binding protein and that, in the absence of GTP, this association is long-lived. The rate of release of hormone from this complex is relatively slow, with the result that the apparent binding affinity is high. On the binding of GTP, they suggest that the free receptor is released and that the rate of release of hormone from the free receptor is relatively fast, resulting in an apparent reduction in binding affinity. Since the antagonists are incapable of promoting the interaction of the hormone receptor with the GTP binding protein in the first place, the receptor remains free in both the presence and absence of GTP and only the low-affinity state is observed. There is a considerable body of evidence arising from several different groups of workers using a wide variety of different experimental techniques to support the concept of the transient association of the hormone receptor with the other components of the system. The essential conclusion is that the role of the hormone receptor in the activation is catalytic and that the coupling of the hormone receptor to the adenylate cyclase catalytic unit is not necessary in order to maintain activation. This type of mechanism has been described by Tolkovsky and Levitzki (1978) as collision coupling. The second major aspect of our model was the proposal that the reduction in target size on activation by p(NH)ppG alone or by GTP in the presence of glucagon or by F- reflected the dissociation of the GTP binding protein. As already discussed, the evidence for this aspect of the model was less satisfactory due to the technical impossibility of determining the target size of the guanine nucleotide binding protein independently of its effects on the adenylate cyclase catalytic activity. A consideration of subsequent work by other groups suggests that this aspect of the model, while consistent with the original data, is likely to be an oversimplification of the true situation. There is now a large body of evidence to support the view that the component of the adenylate cyclase system which binds guanine nucleotides and mediates the activation of the enzyme by GTP and p(NH)ppG remains associated with the catalytic unit in the activated state. At the simplest level, we can consider the properties of the adenylate cyclase of cyc- S49 lymphoma cell membranes. These cells appear to lack a functional guanine nucleotide binding protein but retain a fully competent catalytic unit (Ross et al., 1978). If activation requires the dissociation of the guanine nucleotide binding protein, we might expect that the adenylate cyclase in these cells would be fully active, but in fact it displays very low activity. It is possible to argue that the guanine nucleotide binding protein is in fact still present but has lost its ability to interact with GTP while retaining its ability to suppress the activity of adenylate cyclase. However, this is difficult to reconcile with the

252

6. RICHARD MARTIN

ability of active guanine nucleotide binding protein to restore the response of adenylate cyclase to guanine nucleotides. Pfeuffer ( 1977) has shown that the treatment of detergent-solubilized preparations with GTP immobilized on agarose beads and the removal of the agarose beads lead to a loss in adenylate cyclase activity in the presence of p(NH)ppG. The conclusion is again that the removal of the guanine nucleotide binding protein leads to a loss of activity rather than to increased activation. A problem with this particular study is that treatment with GTP agarose does not affect the response of the enzyme to F- ions to the same extent as would be expected if the same protein subunit is involved in activation by fluoride as is involved in activation by guanine nucleotides. It has also been shown by gel filtration studies of detergent-solubilized preparations that treatment with p(NH)ppG tends to increase the molecular weight of adenylate cyclase rather than to decrease the molecular weight (Neer et af., 1980). These studies can be criticized on the grounds that detergent solubilization causes a major disruption of the system and that the size changes determined, while consistent, are very small in relation to size determined for the enzyme in the absence of p(NH)ppG. While none of these studies is conclusive, they do all point to the same conclusion that the component containing the guanine nucleotide binding site needs to be associated with the catalytic unit for the maintenance of guanine nucleotide activation. In the light of this, we should reconsider our conclusion that the regulatory subunit which dissociates under the influence of guanine nucleotides can be identified with the guanine nucleotide binding protein responsible for activation. In this connection, a recent suggestion by Rodbell (1980) is of considerable interest. He points out that in fat cell plasma membranes, GTP demonstrates an inhibitory effect on adenylate cyclase as well as an activation (Harwood et a/., 1973). Furthermore, it is well established that the effects of inhibitory hormones such as a2-agonists on adenylate cyclase also appear to be mediated through a guanine nucleotide binding protein (Jakobs et af., 1981). Thus, the possibility arises that there are two distinct guanine nucleotide binding proteins, one responsible for mediating activation and the other responsible for mediating inhibition. The strength of the approach of target size analysis by irradiation inactivation is exemplified by the results achieved in studies of the effects of hormones. In this case, the method has yielded direct physical evidence of the nature of the interactions of the hormone receptor with other components of the system and the conclusions are well supported by other types of evidence. The major potential pitfall of the approach is illustrated by our original conclusions regarding the guanine nucleotide binding protein. It is not much of an exaggeration to say that, in this case, it was possible to do only half the experiment, since independent determination of the activity of the guanine nucleotide binding protein was not possible. We started the study with the idea derived from a large body of

253


literature that there are three components in the system and, indeed, the data provided evidence for the existence of three separate components. It was then very difficult to avoid identifying the three components for which we had experimental evidence with the three components which made up our mental picture of the system. As it turns out in the case of the guanine nucleotide binding protein, the picture is not so simple, and other possibilities should be considered. ACKNOWLEDGMENTS The work described in this article was supported by grants from the MRC and the SRC. I would also like to thank Dr. J. M. Stein, who was closely involved in the studies, and Drs. J. C. Metcalfe and M. D. Houslay, with whom I have had many helpful and stimulating discussions. REFERENCES Birnbaumer, L. (1973). Hormone sensitive adenylate cyclases. Useful models for studying hormone receptor functions in cell free systems. Biochim. Biophys. Acru 300, 129-158. Cassel, D., and Selinger, Z. (1977). Mechanism of activation of adenylate cyclase by cholera toxin. Inhibition of GTP hydrolysis at the regulatory site. Proc. Nut/. Acad. Sci. U.S.A. 74, 3307-33 1 I . Cassel, D., and Selinger, Z. (1978). Mechanism of adenylate cyclase activation through the p adrenergic receptor: Catecholamine induced displacement of bound GDP by GTP. Proc. Nutl. Acad. Sci. U.S.A. 75, 4155-4159. Harwood, J . P., Low, H., and Rodbell, M. (1973). Stimulatory and inhibitory effecs of guanyl nucleotides on fat cell adenylate cyclase. J . Eiol. Chem. 248, 6239-6245. Houslay, M. D., and Palmer, R. W. (1978). Changes in the form of Arrhenius plots of the activity of glucagon-stimulated adenylate cyclase and other hamster liver plasma membrane enzymes occurring on hibernation. Biochem. J . 174, 909-919. Houslay, M. D., Ellory, J. C., Smith, G . H . , Hesketh, T. R . , Stein, J . M., Warren, G. B . , and Metcalfe, I. R. (1977). Exchange of partners in glucagon receptor adenylate cyclase complexes. Physical evidence for the independent, mobile receptor model. Biochim. Biophys. Acta 467 208-2 19. Houslay, M. D., Dipple, I . , and Elliot, K. R. F. (1980). Guanosine triphosphate and guanosine 5'[py-imido]triphosphate effect a collision coupling mechanism between the glucagon receptor and catalytic unit of adenylate cyclase. Biochem. J . 186, 649-658. Jakobs, K. H., Aktories, K., and Schulz, G. (1981). Inhibition of adenylate cyclase by neurotransmitters and hormones. Adv. Cyclic Nucleotide Res. 14, 173- 189. Kempner, E. S . , and Schlegel, W. (1979). Size determinations of enzymes by radiation inactivation. Anal. Biochem. 92, 2- 10. Kepner, G. R., and Macey. R. I. (1968). Molecular size determinations by radiation inactivation. Eiochim. Biophvs. Acfa 163, 188-203. Kimura, N., Nakane, K . , and Nagata, N. (1976). Activation by GTP of liver adenylate cyclase in the presence of high concentrations of ATP. Biochem. Biophys. Res. Commun. 70, 1250-1256. Lefkowitz, R. J . , DeLean. A , , Hoffman, B., Stadel, J. M.. Kent, R., Michel, T., and Limbird, L. (1981). Molecular pharmacology of adenylate cyclase-coupled a and p adrenergic receptors. Adv. Cvclic Nucleoride Res. 14, 145-163. Lin, M. C., Welton, A. F., and Berman, M. F. (1978). Essential role of GTP in the expression of adenylate cyclase activity after cholera toxin treatment. J . Cyclic Nucleotide Res. 4, 159-168. Martin, B. R. Stein, J. M., Kennedy, E. L., Doberska, C. A , , and Metcalfe, J . C. (1979). Transient I

254

6. RICHARD MARTIN

complexes. A new structural model for the activation of adenylate cyclase by glucagon. Biochem. J . 184, 253-260. Martin, B. R., Stein, J. M., Kennedy, E. L., and Doberska, C. A. (1980). The effect of fluoride on the state of aggregation of adenylate cyclase in rat liver plasma membranes. Biochem. J . 188, 137-140. Neer, E. J. (1974). The size of adenylate cyclase. J. Biol. Chem. 249, 6527-6531. Neer, E. J . , Echeverria, D., and Knox, S. (1980). Increase in the size of soluble brain adenylate cyclase with activation by guanosine 5’-(py-imino)triphosphate.J . B i d . Chem. 255, 9782-9789. Newby, A. C., Rodbell, M., and Chramback, A. (1978). Adenylate cyclase in polyacrylamide gel electrophoresis solubilised but active. Arch. Biochem. Eiophys. 190, 109-1 17. Pfeuffer. T. (1977). GTP-binding proteins in membranes and the control of adenylate cyclase activity. J . Biol. Chem. 252, 7224-7234. Robinson, G. A., Butcher, R. W., and Sutherland, E. W. (1967). Adenyl cyclase as an adrenergic receptor. Ann. N.Y. Acad. Sci. 139, 703-723. Rodbell, M. (1980). The role of hormone receptors and GTP regulatory proteins in membrane transduction. Nature (London) 284, 17-22. Rodbell, M., Lin, M. C., and Salomon, V. (1974). Evidence for interdependent action of glucagon and nucleotides on the hepatic adenylate cyclase system. J . Eiol. Chem. 249, 59-65. Rodbell, M., Lin, M.C.. Salomon, Y.,Londos, C., Harwood, J. P., Martin, B. R., Rendell, M., and Berman, M. (1975). Role of adenine and guanine nucleotides in the activity and response of adenylate cyclase systems to hormones. Evidence for multisite transition states. Adv. Cyclic Nucleotide Res. 5, 3-29. Ross, E. M., Howlett, A. C., Ferguson, K. M., and Gilman, A. G . (1978). Reconstitution of a hormone-sensitive adenylate cyclase activity with resolved components of the enzyme. J . Eiol. Chem. 253, 6401-6412. Schlegel, W., Kempner, E. S., and Rodbell, M. (1979). Activation of adenylate cyclase in hepatic membranes involves interactions of the catalytic unit with multimeric complexes of regulatory proteins. J. Biol. Chem. 254, 5168-5176, Tolkovsky, A., and Levitzki, A. (1978). Collision coupling of the P-adrenergic receptor with adenylate cyclase. In “Hormones and Cell Regulation” (J. Dumont and 1. Nunez, eds.), Vol. 2, pp. 89-105. North-Holland, Amsterdam. Welton, A. F., Lad, P. M., Newby, A. C., Yamamura, H.,Nicosia, S., and Rodbell, M. (1977). Solubilisation and separation of the glucagon receptor and adenylate cyclase in guanine nucleotide sensitive states. J. Eiol. Chem. 252, 5947-5950. Welton, A. F., Lad, P. M., Newby, A. C., Yamamura, H., Nicosia, S., and Rodbell, M. (1978). The characteristics of lubrol solubilised adenylate cyclase from rat liver plasma membranes. Eiochim. Biophys. Arlo 522, 625-639.

Part I I

Receptors Not InvoIving Adeny Iate Cy clase


CURRENT TOPICS IN MEMBRANES AND TRANSPORT. VOLUME I8

Vasopressin Iso receptors in Mammals: Relation to Cyclic AMPDependent and Cyclic AMPIndependent Transduction Mechanisms SERGE JARD Centre CNRS-INSERM de Pharmarologie-Endorrinologie Montpellier. France

. . . . . . . . . . . . . 255 1. Introduction ...................... 11. 111. Kinetics of Hormone Binding to Vasopressin Receptors . . . . . . . . . . .

IV.

D. Other Vasopressin-Responsive Cells ............ V. Effects of Nucleotides and Other Putative ceptors. . . . . . A. Kidney Receptors ............................................ ........

270 272 272 272

vr. v11. Recognition Patterns of Vasopressin lsoreceptors . . . . . . . . . . . . . . . . . . . . . . . . 275 VIII. Summary and Conclusions. . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280

1.

INTRODUCTION

Vasopressin was first recognized as a pressoric principle present in postpituitary extracts (Oliver and Shafer, 1895). Shortly thereafter its antidiuretic action 255

Copyright 0 19x3 by Academic Press, Inc. All rights of reproduction i n any form reserved. ISBN 0- 12-153318-2

256

SERGE JARD

was discovered and its role in the overall regulation of body fluid osmolarity was clearly established (Verney, 1947). The physiological relevance of the vasopressor action of vasopressin in mammals has often been questioned (see, for instance, Saameli, 1968). Indeed, in intact animals the doses of vasopressin needed to elicit a vascular response are much higher than those producing a clearcut antidiuretic response. However, there is increasing experimental evidence that endogenous vasopressin might play a role in controlling blood pressure under several physiopathological conditions such as hemorrhage (Rocha da Silva and Rosenberg, 1969) or hypertension induced by deoxycorticosterone and high salt intake (Mohring et al., 1977). In these situations where plasma argininevasopressin rises, vasopressin antagonists (Cowley et a l . , 1980; Crofton et al., 1979) or antivasopressin antibodies (Mohring et al., 1977) produce the hypotensive responses expected from blocking of the effects of endogenous vasopressin on vascular tone. In the past few years a large variety of biological effects of vasopressin have been described (see Table I), the physiological relevance of which in most instances has yet to be established. Most of the biological effects cited in Table I are elicited by vasopressin doses which are in the range (or at the upper limit) of physiological blood levels. Pharmacological data obtained with vasopressin structural analogs revealed in most instances a high degree of specificity, suggesting that all the observed effects are mediated by specific vasopressin receptors. It is also clearly established (Sawyer et al., 1981) that several structural modifications of the natural vasopressin could affect its biological activities in a differential manner depending on the target tissue considered. In addition, it has been recognized (see for instance Orloff and Handler, 1967; de Wulf et al., 1980) that vasopressin exerts its biological activities through cyclic AMP-dependent (antidiuretic) or cyclic AMP-independent (pressoric and glycogenolytic, among others) effects. It is therefore very likely that several types of vasopressin receptors (vasopressin isoreceptors) exist. The main purpose of the present article is to review available pharmacological and biochemical data on vasopressin receptors in mammals and to discuss the validity of the different criteria which have been or could be used to distinguish vasopressin isoreceptors.

II. METHODOLOGICAL BASIS FOR THE CHARACTERIZATION OF VASOPRESSIN SORECEPTORS The only absolute criterion for the identification of vasopressin isoreptors is the determination of their primary structure in terms of amino acid sequence. This determination is not presently possible, and the use of indirect criteria is therefore necessary. Michell et al. (1979) have proposed two types of vasopressin receptors, V1 and V2, distinguished on the basis of their functional

257

VASOPRESSIN ISORECEPTORS

TABLE I Bioi.oC‘icAi. Ef-wcrs of- VASOPRE\is likewise the V,,, for unidirectional influx from 1 to 2. In a zero trans exit experiment, the cells are loaded with substrate and radiolabeled substrate until equilibrium is achieved. The extracellular substrate is then rapidly diluted with a large quantity of buffer (usually at least IW-fold), and the loss of radiolabel from the cells is followed. This method does not give true zero trans conditions but the small amount of substrate in the trans solution relative to the cis sohion Ieads to a minimal error. The practically infinite volume on the trans face of the membrane in a zero trans exit experiment leads to minimal flC

C mM

4.0

'

2.0

u 0.1

Tlma

imesl

0.2

0.3

0.4

In~l-C/s,J+C/&,

C

FIG. 3 . (a) A time course for net entry (zero trans entry) of I mM 3-O-methyl-~-glucosein cells pretreated with 10 nM insulin at 37°C. The apparent initial rate calculated from the I second measurement was 0.16 Mlsecond. (b) An integrated rate equation replot of I mM 3-0-methylD-glucose net entry. The initial rate V , = 0.210 &/second. (From Taylor and Holman, 1981.)

353

HEXOSE TRANSPORT IN ADIPOCYTES

backflux of the substrate trans to cis, and for the simple carrier model the rate of exit will be given by the Michaelis-Menten equation. However, there is a rapid change in the substrate concentration at the cis face as efflux proceeds so that an integrated rate equation treatment is required. The data from zero trans exit experiments can be conveniently analyzed by an integrated Michaelis-Menton equation (Karlish eruf., 1972) which can be rearranged (Baker and Naftalin, 1979) to give

-In(S,/S,)/(So - S,)

=

V,,,,t/[K,(S,

- S,)] - 1/K,

(9)

where S, is the starting substrate concentration inside the cell and S, is the internal substrate concentration at time t . This equation is equivalent to a form of the Lineweaver-Burk transformation of the Michaelis-Menton equation 1/S = (VmaX/K,,,)(l/v0) - 1/K,

(10)

Thus, by plotting -ln(S&)/(S,

-

S,) vs

r/(S, - S,)

a value of - I/Knl is obtained from the intercept on the ordinate and a value of l/Vmxxis obtained from the intercept on the abscissa. By comparing Eqs. (9) and (10) it can be seen that when -ln(SI/So)/(So - S,)

=

I/S,

then t/(S, - S,) is equal to I/V,, i.e., the reciprocal of the initial rate at a given substrate concentration. Therefore, one can alternatively calculate the initial rate of exit at a number of internal substrate concentrations and plot the data using Eq. (6). This procedure has the advantage of giving a more faithful reflection of the experimental error. The K , measured by the zero trans exit experiment (K$ ) measures the K, at the internal face of the membrane. The zero trans exit V,,, (Vf, ) measures the V,, for unidirectional efflux from side 2 to side 1. It should be noted that this analysis is based on a carrier model such as Scheme b and assumes that there is a single operational affinity (single K,) for substrate at the inner face of the membrane. There is now evidence for two operational affinities at the inside face of the human erythrocyte hexose transporter (Ginsberg and Stein, 1975) and this may lead to deviations from the curve predicted by Eq. (9) (see Holnian, 1980).

0. Infinite cis Experiments The zero trans experiments provide a means of measuring the kinetic constants of one side when no substrate is available on the opposite side. The K,,, of one side can also be measured when the opposite side is saturated with substrate. Thus, for

354

J. GLIEMANN AND W. D. REES

the infinite cis experiment the net flux cis to trans is measured when the substrate concentration in the cis solution is at a saturating concentration, that is for practical purposes at least 10 times the zero trans K,. The cis side of the transport system is saturated with substrate, and therefore the rate of unidirectional flux from cis to trans will be maximal. On the other hand, the backflux process is dependent on the substrate concentration in the trans solution, and K , on the trans side can therefore be determined by following the rate of net flux into solutions containing different substrate concentrations. In other words, the net flux (cis to trans minus trans to cis) is reduced as the substrate concentration on the trans side increases and the trans to cis flux depends on the K , on the trans side. The infinite cis entry experiment is most easily performed by measuring the time course for the net uptake of a single high concentration of substrate. As the uptake proceeds, the internal substrate concentration will increase from zero until it is finally at equilibrium with the external solution. Thus, the substrate concentration in the trans solution is changing with time, and this allows the determination of the infinite cis entry parameters. Eilam and Stein (1974) showed that their integrated rate equation for net entry could be applied to the infinite cis experiment (see also Ginsburg and Stein, 1975) so that a plot of

ttc vs

ln(l

+ CIS,) + CIS, C

[cf. Eq. (8)l

yields a straight line giving I/Vmax for infinite cis entry as the intercept on the ordinate. The intercept on the abscissa will be -K,ISi ( I + S o h ) , where 7~ is the effective osmotic concentration of the buffer, and K i can thus be calculated (Ginsburg and Stein, 1975; Holman, 1979). K k will be the K , for the internal side. In order to perform infinite cis exit experiments [also known as the Sen and Widdas experiment (Sen and Widdas, 1962)] the cells are first loaded with a saturating concentration of substrate (e.g., 40 mM) and the net efflux of this substrate into solutions containing different concentrations of substrate is followed. With a saturating concentration of substrate in the cis solution the net exit rate remains linear until the internal substrate concentration ceases to be saturating. In adipocytes this means 10-20 seconds even when the cells are treated with insulin. Thus, the rate of net efflux can easily be measured without the need for integrated rate equations. A plot of I / V vs S gives minus the infinite cis exit K, ( K g ) as the intercept on the abscissa. Kt;' measures the K,, for the external side. The infinite cis experiments are technically simpler to perform than the zero trans experiments since they offer the advantage of longer time courses.

E. Infinite trans Experiments In these experiments (which are equivalent to counterflow experiments) the substrate concentration in the trans solution is at a saturating concentration.


355

These experiments allow an additional measure of the K , for one side (cis) when the other side is saturated with substrate. The infinite trans entry experiment will therefore measure the K , for the external site while the infinite trans exit experiment will measure the K , for the internal site. It is possible to formulate rejection criteria for different kinetic models by comparing the results of the different experimental protocols with the predictions of the models (Lieb and Stein, 1974a,b). Thus, using these criteria, Hankin et al. (1972) were able to show that the simple carrier model can be rejected as a model for hexose transport in the human erythrocyte.

V.

TRANSPORT OF NONMETABOLIZABLE SUGARS AND SUGAR ANALOGS IN THE ADIPOCYTE

Since D-glucose entering the adipocyte is rapidly metabolized, it is impossible to study its transport directly. When the rate of D-glucose transport is rate limiting for metabolism, a measure of its transport rate can be obtained through the use of indirect measurements such as the rate of D-glucose incorporation into lipids or CO,. The value of this approach is limited and it does not allow a full kinetic characterization of the transport system. Before considering the results obtained with nonmetabolizable sugars, it is important to note that the sugar permeability in the isolated adipocyte due to nonmediated diffusion is negligible under most conditions. This conclusion is in part derived from the finding that L-glucose at a tracer concentration exhibits an equilibration half time of about 60 minutes under conditions where the half time is 2-3 seconds for 3-O-methyl-~-glucose. The half time for equilibration of L-glucose is further increased to several hours in the presence of 40 mM methylglucose (Whitesell and Gliemann, 1979). In addition, Vinten (1978) found that the total methylglucose permeability, which could not be inhibited by a large concentration of cytochalasin B, was only a small fraction of the total permeability. The transport of the nonmetabolizable C-3 epimer of D-glucose, D-allose, was studied by Loten et al. (1976) who showed it to be transported slowly. Foley et al. (1978) reported that L-arabinose (a D-galactose derivative lacking the C-6 hydroxymethyl group) was also transported slowly by the adipocyte and not metabolized. Both analogs are transported by the glucose transport system since their transport is competitively inhibited by glucose and the rate of transport is stimulated by insulin. These sugars are transported slowly due to their high K,’s for the transport system. D-Allose has a K , of about 270 mM (Rees and Holman, 1981), and L-arabinose has a K , > 50 mM (Foley et al., 1978). These high K,’s limit the usefulness of D-allose and L-arabinose in a conventional kinetic characterization of the transport system since saturation of the transporter with these sugars requires very high concentrations, which are outside a practical concentration range. The low affinity, and hence slow transport of these sugars does,

356


however, offer a distinct advantage in inhibition experiments allowing uptakes to be followed over a period of several minutes as opposed to seconds with a high affinity sugar. 3-O-Methyl-~-glucoseis a rapidly transported D-glucose analog which is not metabolized (Czaky and Wilson, 1956), and Gliemann et al. (1972) showed that it has a distribution space not different from that of tritiated water and urea. Equilibrium exchange experiments carried out as demonstrated in Fig. 1B showed only one K , value of about 5 mM, and insulin caused a marked increase in V,,, without changing K , (Vinten et al., 1976). Similar data were obtained by Vinten (1978) using the same method and by Whitesell and Gliemann (1979) and Taylor and Holman (1981) using the influx method shown in Fig. 1C. An experiment of this type is shown in Fig. 4. It should be noted that the 3-0methylglucose concentration range used in these experiments was fairly narrow (up to about 20 mM). However, experiments have been carried out with substrate concentrations up to 60 mM (with the correction for nonmediated diffusion, which is necessary under these conditions) and this gave a K , value of 4.5 ? 0.6 mM at 37°C in cells stimulated maximally with insulin (G. D. Holman and W. D. Rees, unpublished data). It should be mentioned at this point that the reason for the insulin-induced increase in V,,, is probably that more transporters are available in the plasma membrane after treatment of the cells with insulin. This hypothesis is a result of the work of Kono, Cushman, and their co-workers, and will be discussed below.

S Methylglucose (mM) FIG. 4. The concentration dependence of 3-@methyl-~-glucoseequilibrium exchange at 37°C in the absence of insulin and in cells pretreated with 10 nM insulin. K,, was 4.4 mM and not significantly different for “basal” and insulin-treated cells. V,,,,, was 0.08 mM s e c - 1 for “basal” cells and 1.12 mM sec-1 for insulin-pretreated cells.


357

It implies that one is probably looking at the same species of transporters whether or not the cells are treated with hormone. The properties of the transporter will be discussed in the light of this hypothesis. Adrenalin increases the rate of 3-0-methyl-~-glucosetransport in adipocytes by approximately twofold (Ludvigsen et al., 1980), and this stimulation also occurs through an increase in the V,,,, with no change in the K,. As with insulin, the adrenalin effect is preserved in isolated plasma membranes and it is tempting to speculate that insulin and adrenalin stimulate transport through a common mechanism. The effect of adrenalin appears to be mediated through preceptors (Ludvigsen er al., 1980). Glucocorticoids cause a marked decrease in hexose transport rate. However, no kinetic characterization has been carried out of the transport system in glucocorticoid-treated cells (Foley et al., 1978). It is also possible to measure the transport of hexoses in isolated plasma membranes even though the permeability due to nonmediated diffusion is much higher than in intact adipocytes. Since the membrane preparation does not metabolize D-glucose, the transport of D-glucose can be studied directly without the need for nonmetabolizable analogs. Ludvigsen and Jarett (1979, 1980) reported that the K , for D-glucose uptake was 9-26 mM in plasma membranes isolated from insulin-stimulated adipocytes. The value of the measured K , , was dependent on technical details in the preparation of membranes. The question of whether adipocytes show asymmetric transport parameters was first approached by Whitesell and Gliemann (1979), who measured the net entry of 20 mM 3-0-methylglucose (i.e., “almost” an infinite cis experiment). The progress curve did not deviate significantly from that predicted by symmetrical transport parameters. It was concluded that the system was probably symmetrical and that any asymmetry, if present, was certainly not like that described in human red blood cells. Taylor and Holman (1981) carried out a complete kinetic analysis following the principles of Eilam and Stein (1974) as outlined in the preceding section and found no evidence for kinetic asymmetry of 3-0methylglucose transport. Table 1 shows the transport parameters obtained using different protocols. It should be noted that some authors have reported much lower V,,,, values, particularly in insulin-treated cells. The reason may be that the initial velocities were underestimated. These values are not given in Table I [for discussion, see Whitesell and Gliemann ( 197911. Table 11 shows for comparison the kinetic parameters of the most intensively studied hexose transport system, that of the human erythrocyte. It appears that this transporter shows marked asymmetry as recently reviewed by Widdas ( 1980) in Volume 14 of this series. The data are generally obtained using glucose but the results using 3-0-methylglucose also show marked asymmetry ( G . D. Holman, unpublished observations). The kinetic constants vary depending on the experimental protocol, and the most marked asymmetry is observed in the zero trans experiments with zero trans entry showing a low K , and V,,,, while zero trans

TABLE I KINETICPARAMETERS FOR 3-0-METHYL-D-GLUCOSE TRANSPORT IN

THE

RATADIFQCYTE

K, (mM) Experiment Equilibrium exchange

Zero trans entry (measures outside site)

Zero trans exit (measures inside site) Infinite cis entry (measures inside site) Infinite cis exit (measures outside site)

Reference

Basal

Vinten el a/. ( I 976) Whitsell and Gliemann (1979)" Taylor and Holman ( I98 1) I Whitesell and Gliemann ( 1979)" Taylor and Holman (1981)' Taylor and Holman (1981) r Holman and Rees ( 1982) Taylor and Holman (1981F Taylor and Holman (1981)c

About 5 2.5-5 4.22 2 1.24 2.5-5 5.41 2 0.98 4.09 5 1.05

9.03 5 3.28 4.54 2 1.32

Plus insulin About 5

2.5-5 4.45 5 0.26 2.5-5 6.10 5 1.65 2.66 5 0.26 5.65 5 2.05 6.51 t 0.83 3.60 5 1.33

V,,,

(mM sec - 1)"

Basal

Plus insulin

0.07-0.2 0.058 0.058 ? 0.001

1.6- I .9 0.8 0.84 t 0.002

-

0.034 5 0.034 0.153 2 0.023

1.20 -C 0.19 1.19 t 0.07

-

-

0.066 2 0.013 0.106 2 0.026

0.98 2 0.09 1.76 t 0.63

Equivalent to millimolesiliter intracellular waterkcond. Range of values obtained. 2 SE (from regression analysis). Values from Whitesell and Gliemann (1979) were obtained at 22"C, and the other values at 37°C

359


TABLE I1 K I N ~ I I CPARAMETERS . FOR

D-GLUCOSEIN T H E HUMANERYTHROCYTE

Experiment

Reference

K , (mM)

Equilibrium exchange Zero trans entry (measures outside site) Zero trans exit (measures inside site) Infinite cis entry (measures inside site) Infinite cis exit (measures outside site)

Naftalin and Holman (1977) Lacko el a!. (1972)

34 1.6

6.0 0.6

Karlish et a / . (1972)

25.0

2.15

Hankin et (I/. (1972)

2.8

-

Lacko et a / . (1972)

1.8

-

V,,,

(mM sec

~

1)

exit shows a high K,,, and V,,,,,. However, when the Kn,’s of the inner and outer sites are measured by the two infinite cis procedures, both show symmetrical low K,’s. Equilibrium exchange experiments have until recently been reported to have a high K,,, and V,,,,,. Holman et af. (1981a) have presented evidence for negative cooperativity in the equilibrium exchange of D-glucose in the human erythrocyte, showing nonlinearity in reciprocal plots which reveal two apparent K,’s of 2 and 26 mM. Other cell types also show different kinetic constants for equilibrium exchange and zero trans entry. Whitesell et al. (1977) reported that sugar in the trans solution increased the rate of uptake in thymocytes. Plagemann et af. (1981) have reported similar results with a range of cultured cell types. On the other hand, hepatocyte preparations are similar to the adipocyte with symmetrical zero trans transport parameters for 3-0-methylglucose (Craik and Elliot, 1979). On the basis of kinetic studies it is thus possible to identify at least two classes of mammalian sodium-independent facilitated diffusion systems for hexoses, those which show symmetric, kinetic parameters, and those which show asymmetric parameters.

VI. THE REQUIREMENTS FOR D-GLUCOSE BINDING TO THE ADIPOCYTE HEXOSE TRANSPORT SYSTEM From the K,,, values for transported substrates it is apparent that the relative affinities decrease in the order 2-deoxy-~-glucose (deoxyglucose) > 3-0methyl-D-glucose > (n-glucose) >> L-arabinose > D-allose. To characterize further the binding of D-glucose to the transporter, Rees and Holman (1981) studied the inhibition of D-allose transport by a range of D-glucose epimers, deoxy sugars, fluoro sugars, and other D-glucose analogs [cf. Eq. ( 7 ) ] . From the

360

J. GLIEMANN AND W.

D.REES

relative Ki’s of these analogs it was possible to determine which atoms of the glucose molecule are important for its binding to the transporter. These experiments revealed that the D-glucose molecule binds to the adipocyte transporter and is transported by it in a pyranose ring form. Hydrogen bonds are directed toward the ring oxygen and the oxygen atoms of the hydroxyls at C-1, C-3, and to a lesser extent C-6. The role of the C-4 hydroxyl is not as clear as the other positions but appears to be more important in the absence of a hydroxyl at C-6. There is no requirement for a gluco-configuration C-2 hydroxyl as is the case for the sodium-dependent active sugar transport systems of the intestine and kidney (Crane, 1960; Silverman, 1976). However, it should be noted that these hydrogen bonds need not form simultaneously, but that all are formed at some stage of the transport process. in solution the hexoses will be hydrogen bonded to water and changes in this hydration shell with different analogs may also influence the binding of the molecule to the transporter. The hydrogen bonding requirements of the adipocyte hexose transporter are very similar to those reported for the human erythrocyte by Kahlenberg and Dolansky (1972) and Barnett et af. (1973a), with only slight differences in the affinities of C-4K-6-modified sugars between the two systems. Similar hydrogen bonding requirements have also been reported for hexose transport across the blood-brain barrier (Betz ef af., 1975) and the sodium-independent system of the basal lateral membranes of the small intestine (Wright ef at., 1980).

VII. NONTRANSPORTED COMPETITIVE INHIBITORS OF TRANSPORT Not all competitive inhibitors of transport are transported by the transport system, apparently since spatial restrictions prevent them from passing through the membrane. The use of alkylated D-glucose analogs and disaccharides (Holman et d . , 198I b) has revealed that D-glucose binds to the external site of the transporter through the reducing part of the molecule (C-I) with the nonreducing part of the molecule (C-4) facing the external solution. There is a close approach of the molecule at C-l and C-2 with little space being available around these hydroxyls. There is rather more space around the C-3 hydroxyl which accounts for the transport of 3-O-methyl-~-glucose.Glucose molecules with bulky hydrophobic substitutions at C- 1 or C-416, for example, 4,6-0-ethylidene-~-glucopyranose (4,6-O-ethylidine-~-glucose), n-propyl, or n-butyl-P-~-glucosides,are not transported by the insulin-sensitive hexose transporter, but these compounds are able to enter the cell through an alternative route, probably by nonmediated diffusion (Holman and Rees, 1982). These analogs show asymmetric side-specific competitive inhibition of 3-U-methylglucose transport; 4,6-U-ethylideneD-glucose is an inhibitor on the outside of the cell but does not inhibit on the


361

inside, while the alkyl-@-r>-glucosidesare effective inhibitors at the inside of the cell membrane but do not inhibit at the outside. A similar situation exists in the human erythrocyte. Baker and Widdas (1973) reported that 4,6-O-ethylidene-~-glucosewas a much more effective inhibitor of the outside site than the inside site. Barnett et al. (1973b, 1975) showed that 6-0alkyl derivatives of D-glucose were inhibitors only outside the membrane while alkyl-@-~-glucosideswere inhibitors only at the inner face of the membrane. On the basis of the results from both the human erythrocyte and the rat adipocyte, similar models for the mechanism of D-glucose transport have been put forward (Fig. 5). The glucose molecule in the external solution binds to the transporter through the C-1 end of the molecule. The transporter protein is then proposed to undergo a conformational change and the glucose molecule is transferred to the inner site with C-1 facing the internal solution. Inhibitor studies have thus revealed an asymmetry of the inner and outer binding sites of the adipocyte glucose transport system which is not shown by the kinetics of hexose transport. In this context it is interesting to note that the

FIG.5 . The proposed structure of the transporter. (a) In the absence ofthe substrate the system is closed. (b) Binding to the external site destabilizes the interface between subunits. Sufficient spacc is available to accommodate a bulky group at C-4. (c) Binding to the internal site opens the internal subunit interface. Sufficient space is available to accommodate a bulky group at C-I. i, 0, Inner and outer sites, respectively. (From Holman and Rees, 1982.)

362


inhibition constants of both the inside and outside-directed side-specific analogs are very similar in both the erythrocyte and the adipocyte. There is no evidence for the 10-fold asymmetry of affinities between the inner and outer sites of the human erythrocyte, which is evident in the zero trans experiments, when the inhibition constants of these analogs are compared in the two cell types. Models such as the asymmetric carrier (Geck, 1971; Regen and Tarpley, 1974) predict such as asymmetry and this observation may provide a further reason for rejecting such models. The observation that D-glucose may inhibit transport by binding to the transporter on the cytoplasmic facing side through the nonreducing part of the molecule should also be considered in experiments with isolated plasma membranes. The currently favoured membrane preparation (McKeel and Jarett, 1970) uses a sucrose-containing buffer for the isolation procedure. If sucrose can gain access to the inner site of the transporter (the inner face of the membrane) as in a membrane preparation it may cause competitive inhibition of transport through the free C-4/C-6 part of the glucose molecule. If so, this will lead to an increase in the apparent K , for transport and may explain the reported K,, of 26 mM for D-glucose (Ludvigsen and Jarett, 1980). It has been suggested that cytochalasin B inhibits hexose transport through binding to the inside facing site in the human erythrocyte (Basketter and Widdas, 1978), and it might therefore by analogy also bind to the inside site of the adipocyte transporter. Therefore, sucrose may also compete for the cytochalasin B binding site reducing the apparent binding in a competitive manner.

VIII.

SUGARS WHICH ARE BOTH TRANSPORTED AND PHOSPHORYLATED-RATE-LIMITING STEPS

2-Deoxy-~-glucoseis phosphorylated by the hexokinase and is not believed to be metabolized any further to any major extent (Wick et a l . , 1951). 2-Deoxyglucose phosphate is trapped in the cells and the total rate of 2-deoxyglucose uptake might therefore be taken as a measure of the rate of 2-deoxyglucose transport when the transport step is rate determining. Olefsky (1 978) used 2-deoxyglucose (3-minute uptakes) in an attempt to characterize the transport system and found a K,, for deoxyglucose of about 1.2 mM (Fig. 2 of the reference) as well as an inhibition constant (Ki) for glucose of about 2 mM. However, it turns out that the hexokinase becomes partially rate limiting for the uptake of 2-deoxyglucose at deoxyglucose or glucose concentrations as low as about 50 pM (Foley er al., 1980b). This shift in the rate-limiting step from transport at a trace sugar concentration to hexokinase at higher sugar concentrations is particularly evident in insulin-stimulated cells due to the high sugar permeability of the plasma membrane. Therefore, the measured inhibition


363

constants of phosphorylated sugars will depend on the time of incubation: at short times (a few seconds) the measured inhibition constants will reflect mainly K , on the transport system and at infinite time mainly Ki of the hexokinase. At infinite time the apparent Ki for deoxyglucose and glucose is of the order of 100 @ in insulin-stimulated cells; on the other hand, the inhibition constant of deoxyglucose on the initial velocity of methylglucose uptake ( 1 second measure1980b). ments) is about 5 mM (Foley et d.. Recent results have revealed some surprising characteristics of the 2-deoxyglucose transport (Foley and Gliemann, 1981a). In the presence of 2-deoxyglucose at a very low concentration (7 pM) the hexokinase should act as a sink and the uptake should continue at a linear rate in the absence of efflux of 2deoxyglucose phosphate (or 2-deoxyphosphogluconate which is a minor metabolic product of 2-deoxyglucose phosphate). In fact, the uptake curve was linear for only about 10 minutes and this was caused by a slow efflux of free deoxyglucose. Furthermore, a high fraction of the intracellular sugar was present in the free form. Time course studies showed that the intracellular concentration of free deoxyglucose remained essentially zero for about 1 minute. The ratio of intracellular deoxyglucose concentration to extracellular concentration (accumulation ratio) exceeded unity by 3-5 minutes and then continued to increase. By 60 minutes, the intracellular deoxyglucose concentration had exceeded the extracellular concentration by 50-fold. In other words, free deoxyglucose was markedly accumulated in the cell against its concentration gradient. This accumulation was absent in cells depleted of ATP by treatment with dinitrophenol. The mechanism of the accumulation is in part explained by the phosphorylation of newly transported 2-deoxyglucose followed by dephosphorylation. However, it remains to be explained why the 2-deoxyglucose generated by dephosphorylation does not equilibrate with the extracellular medium within seconds. One possihility is that the transport rate of free 2-deoxyglucose out of the cell is much slower than the inward transport. However, this seems highly unlikely, first because the transport system is symmetric with respect to methylglucose and second because the internal deoxyglucose concentration is so low when the accumulation starts that it is difficult to understand how it could exert any inhibition of the transport system. Therefore, it seems necessary to postulate a diffusion barrier between the site of dephosphorylation and the transporter (Fig. 6 ) . In this connection it is worth noting that deoxyglucose phosphate, generated by phosphorylation of deoxyglucose in a tumor cell line, appears to be located in a compartment of a much lower pH (6.4) than that of inorganic phosphate (cytosol, pH 7.1) (Griffith et al., 1981). Other time course experiments (Foley and Gliemann, 1981a) showed that at higher deoxyglucose concentrations, the accumulation of intracellular free deoxyglucose started earlier whereas the steady-state accumulation ratio de-

364

J. GLIEMANN AND W.

D.REES

plasma membrane

5

Hexokinase

/

[14C] DO

-;

l7pMI

(I

rephosphorylation

a

Slow efflux of accumulated [14d DG from postulated compartment

FIG.6 . A model proposed to explain the accumulation of free 2-deoxyglucose against its concentration gradient in adipocytes. DG, Deoxyglucose; DGP, deoxyglucose phosphate. For further explanation, see text.

creased progressively. Thus, a maximum accumulation ratio of 3.5 was reached by 7 minutes using I mM and a ratio of about 1.6 was reached by 3 minutes using 10 mM extracellular 2-deoxyglucose. This phenomenon is probably related to the limited capacity of the hexokinase. It is difficult to predict the effect of intracellular deoxyglucose on the measured transport parameters of other sugars since the accumulation ratios are not necessarily indicative of the internal concentration at the transport site. Recent experiments have shown that high concentrations of phloretin cause a rapid drop in the ATP level of adipocytes and that this is associated with a dephosphorylation of 2-deoxyglucose phosphate (Wieringa et al., 1981). Phloretin was not used in the experiments cited above showing intracellular accumulation of free deoxyglucose. However, the results of Wieringa et al. (1981) demonstrate that the ATP level may be important in regulating the adipocyte phosphatase activity. Marked accumulations of 2-deoxyglucose have previously been reported in mammalian cells, for example, hamster kidney cortex slices (Elsas and McDonell, 1972). However, in this system sugar transport is sodium dependent and active (uphill), and transport clearly precedes phosphorylation. On the other hand, Kleinzeller and McAvoy (1973) have found evidence for a slight accumulation of 0.5 mM deoxyglucose against its concentration gradient following sodium-independent transport across the basolateral membrane of flounder renal cells. Since dephosphorylation of deoxyglucose phosphate also occurs in this system, the mechanism of accumulation might be similar to that postulated for adipocytes. From a physiological point of view, D-glucose is of course the most interesting


365

sugar. Its inhibition constant on the initial velocity of methylglucose transport is about 8 m M , and this is in agreement with experiments designed to measure uptake of glucose itself (Whitesell and Gliemann, 1979). Similar values have been reported for the inhibition constants of glucose on allose uptake ( 1 3 mM, Loten et al., 1976; and 9 mM, Rees and Holman, 1981) and arabinose uptake (8 mM, Foley et al., 1978). Using the “slow” sugars, the measured inhibition constant would be one of equilibrium exchange ifthe rapidly transported glucose was not metabolized. However, glucose is actually metabolized and transport is rate limiting at low concentrations (Foley et al., 1980d) but not at concentrations above 0.5 mM in insulin-stimulated cells (Gliemann, 1967, 1968). In fact, it has been shown that 2 mM glucose equilibrates across the membrane in insulinstimulated cells (Foley er al., 1980a). Therefore it seems unlikely that K , for net entry is different from that of equilbrium exchange. In other words, the system is probably symmetric with respect not only to 3-O-methyl-~-glucosetransport (as described above) but also to D-glucose transport. In view of the unexpected results with accumulation of free intracellular deoxyglucose, similar experiments were carried out with 7 pV glucose. However, we were unable to detect any accumulation of glucose, which is perhaps not surprising in view of the rapid further conversion of glucose 6-phosphate to metabolites (Foley and Gliemann, unpublished observations). However, this does not rule out that a phosphorylation-dephosphorylation cycle might occur in analogy to the findings with 2-deoxyglucose. The glucose concentration giving half-maximal glucose metabolism is about 1 mM in insulin-stimulated cells (Gliemann, 1968). This is the reason why insulin assays based on glucose metabolism are carried out at a low glucose concentration (Gliemann, 1967; Moody er al., 1974). However, from a physiological point of view this seems paradoxical considering that the plasma glucose concentration varies roughly between 4 and 8 mM. The low “metabolism Km” agrees neither with experiments on epididymal fat pads (Gliemann, 1968) nor with insulin action in vivo. It seems likely that interstitial diffusion gradients exist not only in incubated pieces of adipose tissue, as shown by Crofford and Renold (1965a,b) but also in vivo. Several other metabolizable sugars are transported via the insulin-sensitive glucose transport system but rather little information is available. Mannose at tracer concentration is transported at a slightly slower rate than glucose and is rapidly phosphorylated and metabolized (Foley et al., 1980~). Galactose is transported at about half of the rate of glucose but is phosphorylated at a much slower rate (Vega and Kono, 1978). Fructose is transported slowly by the glucose transporter discussed in this article. However, it should be stressed that the adipocytes possess a specific transport system for fructose which is not influenced by insulin (Schoenle et al., 1979). In “basal” cells, the fructose uptake is almost entirely accounted for by transport via its specific system, whereas the

366

J. GLIEMANN AND W.

D.REES

insulin-sensitive system accounts for about half of the total fructose uptake in insulin-stimulated cells.

IX. MODULATION OF THE TRANSPORT SYSTEM BY GLUCOSE METABOLITES Using 2-deoxyglucose as a probe, it has been found that a high rate of glucose metabolism modulates the transport system. The ability of 2-deoxyglucose to inhibit the initial velocity of 3-0-[ 14C]methylglucoseis slightly greater than that of methylglucose (Foley et af., 1980~).Therefore, it would be expected that 2deoxyglucose was transported at least as rapidly as methylglucose. It is, however, transported at only about one-third of this rate indicating some resistance to the transfer of 2-deoxyglucose across the membrane after its initial binding. Incubation of the insulin-stimulated adipocytes with 10 mM glucose for 30 minutes at 37°C increases the permeability of deoxyglucose to the same level as that of methylglucose (Foley et a/., 1980~).Mannose, which is transported rapidly, phosphorylated, and further metabolized to glucose intermediates, has the same effect. Sugars which are either transported or metabolized slowly have no effect. Thus, a high rate of glucose metabolism modulates the transport system and removes the resistance to transfer of 2-deoxyglucose at a tracer concentration. It is possible that the effect is caused by a feedback of an intermediate of glucose metabolism but no direct evidence is available. It also remains to be clarified whether a high rate of glucose metabolism affects v,, or K,, for transport of 2-deoxyglucose. A model for the hexose transport system would have to account for this phenomenon and we have proposed that shown in Fig. 7. The initial sugar binding occurs at step 1 and here deoxyglucose and methylglucose have equal affinities. The putative glucose metabolite is presumed to act at the cytoplasmic side of the membrane, i.e., at step 3; in the absence of glucose the resistance of step 3 is higher for 2-deoxyglucose than for 3-0-methylglucose, whereas a high rate of glucose metabolism causes a modification of step 3 to give the same resistance for the two sugars. The transmembrane distance between the two points of solution contact of an intrinsic protein or assembly of proteins is rather large as compared with the size of a hexose molecule, and for this reason a diffusive step (step 2) is proposed between the two “discriminators” or “microcarriers” (steps 1 and 3 ) . The properties of the “microcarriers” would be as described by Holman and Rees (1982): the C-1 region of glucose binds to the extracellularly facing (at step 3 pore facing) side; this induces a conformational change and the sugar is let loose on the other side (cf. Fig. 5). Conversely, the C-4 region binds to the cytosolic facing (at step 1 the pore facing) side followed by a conformational change and transfer of the sugar. This model of two re-


367

Step i 0

FIG. 7. A model of the glucose transporter in adipocytes. The model was proposed to explain the acceleriltion of glucose metabolism on the initial velocity of 2-deoxyglucose uptake. Each of the microcamers is proposed to function as illustrated in Fig. 5. (From Foley and Gliemann, 1981b.)

sistances in series separated by a pore (Fig. 7) eliminates the necessity of a very large protein carrier moving the sugar molecule across the membrane. The kinetic transport parameters of a model of this type will be described by complex equations. The resistances at step 1 and step 3 and the pore volume will determine the flux through the model and detailed predictions depend on the value of these parameters. The model is not incompatible with the available data for transport of 3-0-methylglucose in the adipocyte but more refined experiments are necessary to assign values for the basic model parameters.

X.

MECHANISM OF INSULIN’S ABILITY TO INCREASE V,,,,,

Two groups have recently reported important observations clarifying the cause of the insulin-induced increase in V,,, for 3-0-methylglucose transport. Cushman and co-workers (Wardzala et al., 1978) characterized a specific class of D-glucose inhibitable cytochalasin B binding sites in adipocyte plasma membranes and found that insulin treatment of the cells prior to preparation of the membranes caused a marked increase in the number of sites. The cytochalasin B binding site was taken to be a marker for the transporter and the insulin-induced increase in specific plasma membrane cytochalasin B binding sites should therefore be analogous to the insulin-induced increase in V,,, as shown in Fig. 4. Cushman and Wardzala ( 1980) also found glucose-displaceable cytochalasin B binding sites in a low-density microsomal fraction, and, moreover, treatment of the cells with insulin before the subcellular fractionation caused a marked shift in the distribution of binding sites so that the insulin-induced increase in plasma

368


membrane sites was accompanied by a comparable decrease in the microsomal sites. This shift does not occur in cells depleted of ATP, and neither does the insulin effect on transport (Kono er al., 1977; Siege1 and Olefsky, 1980) or the reversal of the transport activity to the basal state in insulin-pretreated cells (Vega et al., 1980; Laursen et al., 1981). The increment in the plasma membrane cytochalasin B binding sites after treatment with insulin at a high concentration corresponds well to the insulin-induced increase in transport of 3-0methylglucose. Moreover, there is good quantitative agreement between the steady-state insulin dose-response relationships of the transport increase on the one hand and the appearance of cytochalasin B binding sites on the other, as well as between the time course of the two phenomena (Karnieli et af., 1981a). Suzuki and Kono (1980) used a modification of the method described by Shanahan and Czech ( 1977) to solubilize from isolated membranes components which catalyze stereospecific glucose transport. These authors found that insulin caused an increase in the transport activity derived from the plasma membrane and a decrease in the activity derived from a light microsomal fraction. Also the experiments of Kono and co-workers show adequate quantitative correlations between the insulin-induced increase in transport activity of the whole cell and the transport activity that can be extracted from the plasma membrane fraction (Kono et al., 1981). Taken together, these independent experiments provide convincing evidence that hexose transporters are in two pools, one (functional) in the plasma membrane and another (nonaccessible or nonfunctional) at some other location. Moreover, insulin causes a translocation of the transporters from the nonfunctional to the functional pool. The model proposed by Cushmann and co-workers is shown in Fig. 8. There is little doubt that an increase in the number of functional transporters in the plasma membrane is a major effect of insulin in adipocytes and the same mechanism has been proposed in striated muscle (Wardzala and Jeanrenaud, 1981). However, criticism has been raised (Carter-Su and Czech, 1980) and an alternative mechanism has been proposed in the adipocyte (Pilch er al., 1980). The translocation hypothesis explains an observation made by several authors and first by Martin and Carter (1970), namely, that insulin has no effect when added directly to plasma membrane vesicles retaining the stereospecific glucose transport system. This hypothesis is also in agreement with the identical K , values for transport of 3-0-methylglucose in basal and insulinstimulated cells using different experimental protocols (Taylor and Holman, 1981) and with identical K , values for a range of different sugars (Holman et al., 1981b). The question is whether the hypothesis explains the entire insulin effect. Some observations favor at first glance the proposal that insulin also increases the sugar transport across each individual transport unit. Thus, an increase in temperature from 20 to 37°C increases transport of 3-0-methylglucose in insulinstimulated but not in “basal” cells (Czech, 1976a; Whitesell and Gliemann,

369

HEXOSE TRANSPORT IN ADIPOCYTES Dissociation

(9 Translocation

“V

Glucose

-

0 + -. Transport

Glucose

Glucose

\b

\\

\@Fusion

lntracellular Pool

I

d a n s l o c a tion

\-@Binding

7

Plasma

0 Association FIG.8. Schematic representation of a hypothetical mechanism of insulin’s stirnulatory action on 1981.) glucose transport in adipocytes. (From Karnieli et d,,

1979). However, Kono et al. (1981) have observed that decreasing temperature causes a shift in the steady-state distribution of transporters toward the plasma membrane pool, and this may explain the apparent difference in the behavior of transporters in the “basal” and insulin-stimulated state. Sonne et al. (1981) observed that the “basal” but not the insulin-stimulated transport rate increased with increasing pH. There is n o information as to whether this difference might also be due to a shift in the distribution between the pools of transporters. It also remains to be established whether sections of the plasma membrane are pinched off and transferred to an intracellular pool as depicted in the model of Cushrnan and co-workers (Fig. 8). It should be noted that insulin influences various transport systems to different degrees. Thus, transport of the nonmetabolizable amino acid a-aminoisobutyric acid is not stimulated by insulin in the adipocyte (Minemura ef al., 1970) and neither is transport of adenosine (W. D. Rees, unpublished observations). Also, as noted above, transport of fructose through the fructose transporter is not modulated by insulin. Therefore, a model of the type proposed by Kono, Cushman, and their co-workers seems to

370


demand that some areas of the plasma membrane are specialized to contain the glucose transporters or that the transporters are able to cluster in such areas. In any case, the transfer, at least from the putative intracellular pool and to the plasma membrane, is surprisingly rapid since the maximal effect of insulin at a high concentration on 3-0-methylglucose transport is manifest by about 1 minute (Whitesell and Gliemann, 1979). At lower (physiological) insulin concentrations the rate-limiting step for activation or deactivation of the transport system appears to be the association or dissociation of insulin from the receptors (Karnieli et af., 1981a). This is in agreement with previous studies on the time course of insulin binding and insulin-induced activation of lipogenesis from glucose (Gliemann et al., 1975). The next question is whether insulin after binding to its receptor causes the formation of a chemical signal which in turn mediates the transfer of transport units from the inactive to the active pool. Larner et al. ( 1979) have extracted a heat- and acid-stable factor from muscle which inhibits cyclic AMP-dependent protein kinase and activates glycogen synthase phosphoprotein phosphatase. Jarett and Seals (1979) have shown that insulin activates pyruvate dehydrogenase in mitochondria provided that plasma membranes are present in the mixture. Later studies showed that insulin can induce the release of a material from adipocyte membranes which appears to mediate its action on pyruvate dehydrogenase and this material was indistinguishable from the “Larner material” (Kiechle et al., 1981). The factor seems to be a peptide with a molecular weight of 1000-4000 (Seals and Czech, 1981; Kiechle et al., 1981) and is perhaps produced from an endogenous substrate by a protease which becomes activated after binding of insulin to its receptor (Seals and Czech, 1980). The putative mediator is probably not a part of the insulin molecule since its release from adipocyte plasma membranes can be initiated by proteases such as trypsin in the absence of insulin (Seals and Czech, 1980). It should also be noted in this connection that irreversible binding of photoaffinity-labeled insulin to adipocytes appears to cause an irreversible activation of the transport system (Ushkoreit et al., 1981). The question remains open as to whether the “Larner material” has a function as a signal for the stimulation of hexose transport. Another possibility is that the intracellular concentration of calcium ions plays a critical role, even though several authors have noted that the effect of insulin on hexose transport is not influenced by extracellular calcium. Clausen and co-workers (Sorensen et al., 1980) have shown that insulin increases the efflux of radiolabeled calcium ion from preloaded adipose or muscle tissue with a similar time course as the increase in transmembrane sugar transport. Thus, insulin may increase the release of calcium ions from intracellular stores and thereby cause a shift in the distribution of transporters. Several other mechanisms have been proposed [see Gliemann et al. (1981) for review].


XI.

371

HUMAN ADIPOCYTES

Transport of 3-0-methylglucose has been studied using the technique shown in Fig. IC (Ciaraldi et al., 1979; Pedersen and Gliemann, 1981). The transport system appears very similar to that of the rat adipocyte in that K , for net entry as well as for equilibrium exchange of 3-0-methylglucose is about 4 mM. In other words, the system seems to behave symmetrically with respect to 3-0-methylglucose transport. The inhibition constant of glucose on the initial velocity of 3-0-methylglucose uptake is about 8 mM as in the rat adipocyte. Transfer of glucose across the plasma membrane by nonmediated diffusion is also insignificant in the human adipocyte. The main difference between epididymal adipocytes from 200-g rats and abdominal subcutaneous adipocytes from normal weight adult humans is the relatively small response to insulin (two- to threefold) in the human cells. In the “basal” state, the permeability of the human cells is about half of that of the rat cells. Assuming that the turnover of sugar molecules is the same on each transporter from the two species, this indicates that the density of transporters in the human cell is about half of that in the rat cell. In the presence of insulin the permeability of the human adipocyte is about one-tenth of that of the rat adipocyte. It is likely, therefore, that the adipocyte of adult humans is able to recruit only a limited number of additional transport units when treated with insulin. The questions of the orientation of the glucose molecule in the transporter and the spatial and hydrogen binding requirements have not been studied in the human adipocyte but there is no a priori reason to expect any important differences between the species. As in the rat, the rate of metabolism of glucose appears to be limited by the hexokinase and not by transport when insulin is present and the glucose concentration exceeds a few millimoles per liter (Pedersen and Gliemann, 1981).

XII. THE TRANSPORT SYSTEM IN OBESITY AND DIABETES Results from our laboratory using 3-0-methylglucose have shown that the permeability (cdsecond) in the absence of insulin is about the same in small cells from small lean rats and large cells from large obese rats (Foley el al., 1980d). The number of transporters per unit surface area is thus probably quite independent of the cell size. On the other hand, after treatment with insulin the permeability is much smaller in cells from large rats than in cells from small rats. Therefore, it is likely that the cells from obese rats are able to recruit less transporters from the inactive pool after treatment with insulin. Recently, Cushman et al. (198 I ) showed that the number of cytochalasin B binding sites per unit

372

J. GLIEMANN AND W.

D.REES

area of the plasma membrane was independent of the cell size whereas the insulin-induced increase in the number of binding sites in the plasma membrane was markedly reduced in cells from obese rats. This agrees with the transport studies of Foley et af. (1980d). Earlier studies have shown that the decreased insulin responsiveness with respect to glucose metabolism was related to the degree of obesity of the animals rather than to the adipocyte size per se (Gliemann and Vinten, 1974; Hansen et uf., 1974). These studies were carried out using a low glucose concentration (0.5 mM) and the same conclusion probably applies, therefore, to the glucose transport step. Other authors studying transport in small and large cells (Livingston and Lockwood, 1974; Czech, 1976b; Olefsky, 1976) have obtained different results, probably because 2-deoxyglucose uptake was taken as a measure of transport or because the initial velocity was missed in transport studies using 3-O-methylglucose (for discussion, see Foley et af., 1980d). Streptozotocin-induced diabetes in the rat is associated with a marked reduction in the ability to stimulate glucose transport and metabolism in the adipose cell (Kasuga et af., 1978; Kobayashi and Olefsky, 1979). Recent studies by Cushman and co-workers (Karnieli et af., 1981b) have shown that this, in fact, is associated with a depletion of the pool of cytochalasin B binding sites (and therefore probably hexose transporters) that can be recruited by insulin treatment.

XIII. RECONSTITUTION OF THE HEXOSE TRANSPORTER The reconstitution of intrinsic membrane proteins into an artificial phospholipid bilayer offers a powerful technique for the study of hexose transport. These techniques have been pioneered with the human erythrocyte hexose transporter (Kasahara and Hinkle, 1976) and have now been refined, giving a high efficiency of reconstitution [for recent reviews see Baldwin and Lienhard (1981) and Jones and Nickson (1981)l. Briefly, the results indicate that the purified transporter is a protein of 46,000 molecular weight (Gorga et af., 1979) which binds cytochalasin B in what Baldwin and Lienhard (1981) have suggested to be a 1:l ratio. Studies on the native membrane using radiation inactivation (Jung et al., 1980) suggest, however, that the transporter is larger with a molecular weight of 2 X lo5. Wheeler and Hinkle (1981) have shown that the reconstituted transporter-like the transporter of the intact erythrocyte-shows accelerated sugar transport when sugar is added to the trans side (accelerated exchange). The transporter is incorporated at random in the liposome and the reconstituted system is therefore not asymmetric. However, asymmetry becomes manifest when the liposomes are treated with trypsin which cancels transport in the transporters incorporated “upside down” (Wheeler and Hinkle, 1981).

373


Reconstitution of the adipocyte hexose transporter has also been reported (Shanahan and Czech, 1977), but since there are few copies of the transporter in the adipocyte membrane, this system has presented greater difficulties. Carter-Su et al. (1980, 1981) have partially purified a protein from rat adipocyte membranes and report that an integral membrane protein can be reconstituted into liposomes which then show stereospecific glucose transport. The transport protein is-in contrast to the glycoprotein insulin receptor-not retained by column chromatography using immobilized concanavalin A (Carter-Su et al., 1981). This does not rule out the possibility that the two proteins are noncovalently associated within the native membrane. The molecular weight of the adipocyte transport protein has not been determined, but the molecule has been reported to have a Stokes radius of 60-80 A (Carter-Su er af., 1981). These molecular dimensions would be sufficient to allow the protein to span the membrane consistent with a model in which substrate binding sites are in contact with each solution. The molecule could therefore provide a channel through which the sugar can move (cf. Fig. 7).

XIV.

CONCLUDING REMARKS

Studies over the last decade have elucidated the kinetics of the binding of insulin to its receptor and the relation between insulin binding and biological effects such as the enhancement of glucose transport. Furthermore, the subunit structure of the insulin receptor has been clarified (for reviews see, for example, Czech, 1980; and Gliemann ef al., 1982). In addition, the characteristics of the insulin-sensitive glucose transporter have been elucidated using adipocytes as a model system. This transporter is similar to that of human erythrocytes with respect to substrate specificity but is different with respect to the kinetic properties of glucose transport. The insulin-induced increase in hexose transport seems to be brought about by a transfer of transporters to the plasma membrane from a storage pool. Future experiments may clarify the important questions of the precise nature of the transfer process and the properties of a possible chemical mediator. ACKNOWLEDGMENTS W. D. Rees is a recipient of a NATO-SERC overseas postdoctoral fellowship. The authors wish to thank Drs. S . W. Cushman, G . D. Holman, and J . Vinten for allowing us to reproduce their published figures. REFERENCES Andreasen, P . , Schaumburg, B . , Plsterlind, K., Vinten, .I.Gammeltoft, , S . , and Gliemann, J . ( 1974). A rapid technique for isolation of thymocytes from suspension by centrifugation through silicone oil. Anal. Biochcm. 59, 110-1 16.

374


Baker, G . F., and Naftalin, R. J. (1979). Evidence for multiple operational affinities for D-glucose inside the human erythrocyte membrane. Biochim. Biophys. Acza 550, 474-484. Baker. G. F., and Widdas, W. F. (1973). The permeation of human red cells by 4,6-0-ethylidene-a-D-glucopyranose(ethylidene glucose). J . Physiol. (London) 231, 129- 142. Baldwin, S . A , , and Lienhard, G. E. (1981). Glucose transport across plasma membranes: Facilitated diffusion systems. Trends Biochem. Sci. 6 , 208-2 I 1. Bang, 0.. and (Zlrskov, S. L. (1937). Variations in the permeability of the red blood cells in man. J. Clin. Invest. 16, 279-281. Bamett, J. E. G., Holman, G . D., and Munday, K. A. (l973a). Structural requirements for binding to the sugar transport system of the human erythrocyte. Biochem. J. 131, 21 1-221. Bamett, J. E. G . , Holman, G , D., and Munday, K. A. (1973b). An explanation of the asymmetric binding of sugars to the human erythrocyte sugar transport system. Biochem. J. 135, 539541. Bamett, J: E. G., Holman, G . D., Chalkley, R. A., and Munday, K. A. (1975). Evidence for two asymmetric conformational states in the human erythrocyte sugar transport system. Biochem. J. 145, 417-429. Basketter, D. A., and Widdas, W. F. (1978). Asymmetry of the hexose transfer system in human erythrocytes. J. Physiol. (London) 278, 389-401. Betz, A. L., Drewes, L. R., and Gilboe, D. D. (1975). Inhibition of glucose transport into brain by phlorizin, phloretin and glucose analogues. Biochim. Biophys. Acta 406, 505-5 15. Carter-Su, C.. and Czech, M. P. (1980). Reconstitution of o-glucose transport activity from cytoplasmic membranes. J. B i d . Chem. 255, 10382- 10386. Carter-Su, C., Pillion, D. J., and Czech, M. P. (1980). Reconstituted D-glucose transport from the adipocyte plasma membrane. Chromatographic resolution of transport activity from membrane glucoproteins using immobilized Con A. Biochemistry 19, 2374-2385. Carter&, C., Pillion. D. I., and Czech, M. P. (1981). Chromatographic resolution of the insulin receptor from the insulin insensitive D-glucose transporter of adipocyte plasma membranes. Biochemistry 20, 216-221. Ciaraldi, T. P.,Kolterman, 0.G., Siegel, J. A,, and Olefsky, J. M. (1979). Insulin stimulated glucose transport in human adipocytes. Am. J . Physiol. 236, E621LE625. Craik. J. D., and Elliot, K. R. F. (1979). Kinetics of 3-O-methy~-o-gh1cosetransport in isolated rat hepatocytes. Biochem. J. 182, 503-508. Crane, R. K. (1960). Intestinal absorption of sugars. Physiol. Rev. 40, 789-825. Crofford, 0. B., and Renold, A. E. (1965a). Glucose uptake by incubated rat epididymal adipose tissue. Rate limiting steps and site of insulin action. J. Biol. Chem. 240, 14-21. Crofford, 0. B., and Renold, A. E. (1965b). Glucose uptake by incubated rat adipose tissue. Characteristics of the glucose transport system and action of insulin. J. Biol. Chem. 240, 3237-3243. Crofford, 0. B., Stauffacher, W., Jeanrenaud, B., and Renold, A. E. (1966). Glucose transport in isolated fat cells. Procedures for measurement of the intracellular waterspace. Helv. Physiol. Pharmacol. Acta 24, 45-57. Cushman, S . W., and Wardzala, L. J. (1980). Potential mechanism of insulin action on glucose transport in the isolated rat adipose cell. J. Biol. Chem. 255, 4758-4762. Cushman, S. W.. Hissin, P. I., Wardzdla, L. J., Foley, J. E., Simpson, 1. A., Kamieli, E., and Salans, L. B. (1981). Mechanism of insulin resistance in the adipose cell in the aging rat model of obesity. Biochem. Sac. Trans. 9, 518-522. Czaky, T. Z., and Wilson, J. E. (1956). The fate of 3 - 0 4 14C]methyl-~-g~ucose in the rat. Biachim. Biophys. Acta 22, 185-186. Czech. M. P. (1976a). Regulation of the o-glucose transport system in isolated fat cells. Mol. Cell. Biochem. 11, 51-63.


375

Czech, M. P. (1976b). Cellular basis of insulin insensitivity in large adipocytes. J. Clin. Invest. 57, 1523-1532. Czech, M. P. ( 1980). Insulin action and the regulation of hexose transport. Diabetes 29, 399-409. Eilam, Y., and Stein, W. D. (1974). Kinetic studies of transport across red blood cell membranes. In “Methods in Membrane Biology” (E. D. Korn, ed.), Vol. 2, pp. 283-354. Plenum, New York. Elsas, L. J., and McDonell, R. C. (1972). Hzxose transport and phosphorylation by hamster kidney cortex slices and everted jejunal rings. Biochim. Biophys. Acta 255, 948-959. Foley, J. E., and Gliemann, J. (1981a). Accumulation of 2-deoxyglucose against its concentration gradient in rat adipocytes. Biochim. Biophys. Acru 648, 100-106. Foley, J. E., and Gliemann, J. (I981 b). Glucose transport in isolated adipose cells. Int. J . Obesity 5, 679-684. Foley, J. E., Cushman, S. W., and Salans, L. B. (1978). Glucose transport in isolated rat adipocytes with measurements of L-arabinose uptake. Am. J. Physiol. 234, El 12-El 19. Foley, J. E., Cushman, S. W., and Salans, L. B. (1980a). Intracellular glucose concentration in small and large adipose cells. Am. J . .Physiol. 238, E180-El85. Foley, J. E.. Foley. R.. and Gliemann, 1. (1980b). Rate limiting steps of 2-deoxyglucose uptake in rat adipocytes. Biochim. Biophys. Acfo 599, 689-698. Foley, J. E., Foley, R., and Gliemann, J. ( 1 9 8 0 ~ )Glucose-induced . acceleration of deoxyglucose transport in rat adipocytes: Evidence for a second barrier in sugar entry. J . Biol. Chem. 255, 9614-9677. Foley, J. E., Laursen, A. L., Sonne, O., and Cliemann, J . (1980d). Insulin binding and hexose transport as related to fat cell size. Diahetologia 19, 234-241. Geck, P. (1971). Properties of a carrier model for the transport of sugars by human erythrocytes. Biochim. Biophys. Acra 241, 462-47:!. Ginsburg, H., and Stein, W . D. (1975). Zkro-trans and infinite-cis uptake of galactose in human erythrocytes. Biochim. Biophys. Aria 382, 353-368. Gliemann, J. (1967). Assay of insulin-likc activity by the isolated fat cell method. I. Factors influencing the response to crystalline insulin. Diubetologin 3, 382-388. Gliemann, J . (1968). Glucose metabolism and response to insulin of isolated fat cells and epididymal fat pads. Aria Physiol. Scand. 72, 481-491. Gliemann. J., and Vinten, J. (1974). Lipcigenesis and insulin sensitivity of single fat cells. J. Physiol. (London) 236, 499-516. Gliemann, J., Osterlind, K., Vinten, J., and Gammeltoft, S. (1972). A procedure for measurement of distribution spaces in isolated fat c(:lls. Biochim. Biophys. Acra 286, 1-9. Gliemann, J., Gammeltoft, S., and Vinten, I . (1975). Time course of insulin-receptors binding and insulin-induced lipogenesis in isolated rat fat cells. J . B i d . Chem. 250, 3368-3374. Gliemann, J., Laursen, A. L., Foley, J. E., and Sonne 0. (1981). The insulin receptors: Looking at the present. In “Current Views on Insillin Receptors (D. Andreani, ed.), pp. 1-12. Academic Press, New York. Ghemdnn, J . , Foley, J. E., Sonne, 0..and Laursen, A. L. (1982). Insulin. In “Polypeptide Hormone Receptors (B. I. Posner. ed.). Dekker, New York, in press. Gorga, F. R., Baldwin, S. A,, and Lienhard, C.E. (1979). The monosaccharide transporter from human erythrocytes is heterogenously glycosylated. Biochem. Biophys. Res. Commun. 91, 955-961. Griffith, J. R., Stevens, A. N., Iles, R. P I . , Gordon, R. E., and Shaws, D. (1981). 31P-NMR investigation of solid tumors in the living rat. Biorci. Rep. 1, 319-325. Hankin, B. L., Lieb, W. R., and Stein, W. D. (1972). Rejection criteria for the asymmetric carrier and their application to glucose transport in the human red blood cell. Biochim. Biophys. Acta 288, I 14- 126.

376


Hansen, F. M., Hejriis Nielsen, J., and Gliemann, J. (1974).The influence of body weight and cell size on lipogenesis and lipolysis of isolated rat fat cells. Eur. J. Clin. Invest. 4, 41 1-418. Holman, G. D. (1979). Infinite-cis influx of cyclic AMP into human erythrocyte ghosts. Biochim. Biophvs. Acts 553, 489-494. Holman, G . D. (1980). An allosteric pore model for sugar transport in human erythrocytes. Biuchim. Biophys. Aria 599, 202-213. Holman. 0. D., and Rees, W. D. (1982). Side specific analogues for the rat adipocyte sugar transport system. Biochim. Biophys. Acfa 685, 78-86. Holman, G. D., Busza, A. L., Pierce, E. J., and Rees, W. D. (1981a). Evidence for negative cooperativity in human erythrocyte sugar transport. Biochim. Biophys. Acfa 649, 503-5 14. Holman, G. D., Pierce, E. J., and Rees, W. D. (1981b). Spatial requirements for insulin-sensitive sugar transport in rat adipocytes. Biochim. Biophvs. Arfa 646, 382-388. Jarett, L.. and Seals, J. R. (1979). Pyruvate dehydrogenase activation in adipocyte mitochondria by an insulin-generated mediator from muscle. Science 206, 1407- 1408. Jones, M. N . , and Nickson. J . K. (1981). Monosaccharide transport protein of the human erythrocyte membrane. Biochim. Eiuphys. Acfa 650, 1-20. Jordan, J. E., and Kono, T. (1980). Elimination of insulin-like activity present in certain batches of crude bovine serum albumin by trypsin treatment. Anal. Biochem. 104, 192-195. lung, C. Y., Hsu, T. L., Hah, J. S . , Cha, C., and Haas, M. N. (1980). Glucose transport carrier of human erythrocytes: Radiation-target size of glucose sensitive cytochalasin-B binding protein. J. Biol. Chem. 255, 361-364. Kahlenberg, A , , and Dolansky, D. (1972). Structural requirements of u-glucose for its binding to isolated human erythrocyte membranes. Can. J. Biochem. 50, 638-643. Karlish, S . J. D., Lieb, W . R., Ram, D., and Stein, W. D. (1972). Kinetic parameters for glucose efflux from human red blood cells under zero-trans conditions. Biochim. Biophys. Acfa 255, 126- 132. Karnieli, E., Zarnowski, M. J . , Hissin, P. J., Simpson, I. A,, Salans, L. B., and Cushman, S . W. (198 I a). Insulin-stimulated translocation of glucose transport system in the isolated rat adipose cell. Time course, reversal, insulin concentration dependency, and relationship to glucose transport activity. J. Biol. Chem. 256, 4772-4777. Kamieli, E., Hissin, P. J., Simpson, I. A., Salans, L. B., and Cushman, S . W. (1981b). A possible mechanism of insulin resistance in the rat adipose cell in streptozotocin-induced diabetes mellitus. J. Clin. Invest. 68 81 1-814. Kasahara, M., and Hinkle, P. C. (1976). Reconstitution of u-glucose transport catalysed by a protein-fraction from human erythrocytes in sonicated liposomes. Proc. Nail. Acad. Sci. U.S.A. 13, 396-400. Kasuga. M., Akanuma, Y., Iwamoto, Y.,and Kosaka, K. C. (1978). Insulin binding and glucose metabolism in adipocytes of streptozotocin diabetic rats. Am. J. Physiol. 235, E175-El82. Kiechle, F. L., Jarett, L., Kotagal, N., and Popp, D. A. (1981). Partial purification from rat adipocyte plasma membranes of a chemical mediator which stimulates the action of insulin on pyruvate dehydrogenase. J. B i d . Chem. 256, 2945-295 1. Kleinzeller, A., and McAvoy, M. (1973). Sugar transport across the peritubular renal cells of the flounder. J. Gen. Physiol. 62, 169-184. Kobayashi, M., and Olefsky, 1. M. (1979). Effects of streptozotocin induced diabetes on insulin binding, glucose transport and intracellular glucose metabolism in rat adipocytes. Diabetes 28, 87-95. Kono, T. (1969). Destruction of insulin effector system of adipose tissue cells by proteolytic enzymes. J. Biol. Chem. 244, 1772-1778. Kono, T., Robinson, F. W.. Sarver, J. A., Vega, F. V., and Pointer, R. H. (1977). Actions of insulin in fat cells. Effects of low temperature, uncouplers of oxidative phosphorylation and respiratory inhibitors. J. Biol. Chem. 252, 2226-2233.

HEXOSE TRANSPORT IN ADIPOCMES

377

Kono, T., Suzuki, K . , Dansey, L., Robinson, F. W., and Blevins. T. L. (1981). Energy dependent and protein synthesis independent recyclings of the insulin sensitive glucose transport mechanism in fat cells. J. Biul. Chem. 256, 6400-6407. Lacko, L., Wittke. B . . Komphardt, H. (1972). Zur Kinetik der glucose-Aufnahme in Erythrocyten. Effekt der Transkonzentration. Eur. J . Biochem. 25, 447-454. Lamer, J., Galasko, G., Cheng, K.. DePaoli-Roach, A. A , . Huang, L., Daggy, P., and Kellog, I. ( 1979). Generation by insulin of a chemical mediator that controls protein phosphorylation and dephosphorylation. Science 206, 1408- 1410. Laursen, A. L.. Foley. J. E.. Foley, R., and Gliemann. J. (1981). Termination of insulin-induced hexose transport in adipocytes. Biochim. Biuphvs. Acra 673, 132- 136. LeFevre, P. G. (1948). Evidence of active transfer of certain non-electrolytes across the human red cell membrane. 1. Gen. Ph.vsiul. 31, 505-527. Levine, R., Goldstein, M., Klein. S., and Huddlestun, B. (1949). The action of insulin on the distribution of galactose in eviscerated nephrectoinized dogs. J. Biul. Chem. 179, 985986. Lieb, W. R., and Stein, W. D. (1974a). Testing and characterizing the simple pore. Biurhim. Biuphys. Acra 373, 165- 177. Lieb, W. R., and Stein, W. D. (1974b). Testing and characterizing the simple carrier. Biuchim. Biuphvs. Acta 373, 178-196. Livingston. J. N.. and Lockwood, D.H. (1974). Direct nieasurement of sugar uptake in small and large adipocytes from young and adult rats. Biochem. Biuphvs. Res. Cummun. 61, 989-996. Loten, E. G., Regen, D. M., and Park, C. R. (1976). Transport of D-allose by isolated fat cells. An effect of ATP on insulin stimulated transport. J. Cell Physiul. 89, 651-659. Ludvigsen, C . , and Jarett, L. (1979). Kinetic analysis of D-glucose transport by adipocyte plasma membranes. J. Biul. Chem. 254, 1444-1446. Ludvigsen, C., and Jarett, L. (1980). A comparison of basal and insulin stimulated glucose transport in rat adipocyte plasma membranes. Diabetes 29, 373-378. Ludvigsen, C., Jarett, L., and McDonald. J. M. (1980). The characterization of catecholamine stimulation of glucose transport by rat adipocytes and isolated plasma membranes. Enducrinulo ~ Y106, 786-790. Lundsgaard, E. (1939). On the mode of action of insulin. Uppsala Lak. Fiiren. Fiirh. 45, 1-4. McKeel, D. W.. and Jarett, L. (1970). Preparation and characterisation of a plasma membrane fraction from isolated fat cells. J . Cell. Biul. 44, 417-432. Martin, D. B., and Carter, J. R. (1970). Insulin stimulated glucose uptake by subcellular particles from adipose tissue. Science 167, 873-874. Meldahl, K. P., and 0rskov. S. L. (1940). Photoelektrische Methode zur Bestimmung der Permeierungsgcschwindigkeit von Anelektrolyten durch die Membran von roten Blutkorperchen. Scand. Arch. Physiul. 83, 266-280. Minemura, T., Lacy, L. W., and Crofford, 0. B. (1970). Regulation of the transport and metabolism of amino acids in fat cells. J. Biul. Chem. 245, 3872-3881. Moody, A. J., Stan. M. A,, Stan, M., and Gliemann, J. (1974). A simple free fat cell bioassay for insulin. ffurm. Merab. Res. 6, 12-16. Morgan, H.E., Regen, D. M., and Park, C. R. (1964). Identification of a mobile carrier-mediated sugar transport system in muscle. J . B i d . Chem. 239, 369-374. Naftalin, R. J.. and Holman, G. D. (1977). Transport of sugars in human red cells. In "Membrane transport in Red Cells" (C. J. Ellory and V . L. Lew, eds.), pp. 257-299. Academic Press, New York. Olefsky, J. M. (1976). The effect of spontaneous obesity on insulin binding, glucose transport and glucose oxidation of isolated adipocytes. J. Clin. Invest. 57, 842-851. Olefsky, J. M. (1978). Mechanisms of the ability of insulin to activate the glucose transport system in the rat adipocytes. Biochcm. J. 172, 137-145.

378


0rskov. S . L. ( I 935). Eine Methode zur fortlaufenden photographischen Aufzeichung von Volumanderungen der roten Blutkorperchen. Biochem. Z. 279, 241-249. Pedersen, O., and Gliemann, I. (1981j. Hexose transport in human adipocytes; factors influencing the response to insulin and kinetics of methylglucose and glucose transport. Diahetologia 20, 630-635. Pilch. P. F., Thompson, P. A.. and Czech, M. P. (1980). Coordinate modulation of D-glucose transport activity and bilayer fluidity in plasma membranes derived from control and insulintreated adipocytes. Proc. Natl. Acad. S r i . U.S.A. 77, 915-918. Plageman, P. G . W., Wohlheuter, R. M., Graff. J.. Erbe. J., Wilkie, P. (1981). Broad specificity hexose transport system with differential mobility of loadcd and empty carrier but directional symmetry is a common property of mammalian cell lines. J . B i d . Chem. 256, 2835-2842. Rees. W. D., and Holman, G. D. (1981). Hydrogen bonding requirements for the insulin-sensitive sugar transport system of rat adipocytes. Biochim. Eiophys. Acta 646, 251-260. Regen. D. M., and Tarpley, H. L. (1974). Anomalous transport kinetics and the glucose carrier hypothesis. Biochim. Biophvs. Acta 339, 2 18-233. Rodbell, M. (1964). Mctabolism of isolated fat cells. I. Effects of hormones on glucose metabolism and lipolysis. J. B i d . Chem. 239, 375-380. Schoenle, E., Zapf, J., and Froesch, E. R. (1979). Transport and metabolism of fructose in fat cells of normal and hypophysectomized rats. Am. J. Physiol. 237, E325-330. Seals, J. R., and Czech, M. P. (1980). Evidence that insulin activates an intrinsic plasma membrane protease in generating a secondary chemical mediator. J. Eiol. Chem. 255, 6529-6531, Seals. J. R., and Czech, M. P. (1981). Characterization of a pyruvate dehydrogenase activator released by adipocyte plasma membranes in response to insulin. J . B i d . Chem. 256, 2894-2899. Sen. A. K., and Widdas, W. F. (1962) Determination of the temperature and pH dependence of glucose transfer across the human erythrocyte membrane measured by glucose exit. J . Physiol. (London) 160, 392-403. Shanahan, M. F., and Czech, M. P. (1977). Purification and reconstitution of the adipocyte plasma membrane o-glucose transport system. J . Biol. Chem. 252, 8341-8343. Siegel, J . , and Olefsky, J. M. (1980). Role of intracellular energy in insulin’s ability to activate 3-0methylglucose transport by rat adipocytes. Biochemistr?, 19, 2183-2 190. Silverman, M. (1976). Glucose transport in the kidney. Biochim. Biophvs. Acta 457, 303-351. Sonne. O., Gliemann, J., and Linde, S . (1981). Effect of pH on binding kinetics and biological effect of insulin in rat adipocytes. J. Biol. Chem. 256, 6250-6255. Sbrensen, S. S., Christensen, F., and Clausen, T. (1980). The relationship between the transport of glucose and cations across cell membranes in isolated tissues. X . Effect of glucose transport stimulation on the efflux of isotopically labelled calcium and 3-0-methylglucose from soleus muscles and epididymal fiat pads of the rat. Biochim. Bivphys. Acta 602, 433-445. Suzuki, K . . and Kono, T. (1980). Evidence that insulin causes translocation of glucose transport activity to the plasma membrane from an intracellular storage site. Proc. Nail. Acad. Sci. U.S.A. 77, 2542-2545. Taylor, L. P., and Holman, G. D. (1981). Symmetrical kinetic parameters for 3-O-methyl-~-glucose transport in adipocytes in the presence and in the absence of insulin. Biochim. Biophys. Aria 642, 325-335. Thorsteinsson, B., Gliemann, J., and Vinten, J. (1976). The content of water and potassium in fat cells. Biochim. Biophys. Acta 428, 223-227. Uschkoreit, J . , Brandenburg. D., and Gliemann, J. (1981). Photoaffinity labelling of insulin receptors: Correlation of receptor occupancy and stirnulation of glucose transport in adipocytes. I n “Current Views on Insulin Receptors” (D. Andreani, ed.), pp. 317-322. Academic Press, New York.


379

Vega. F. V., and Kono. T. (1978). Effects of insulin on the uptake of o-galactose by isolated rat epididymal fat cells. Biochim. Biophvs. Actu 512, 221-222. Vega. F. V . , and Kono, T. (1979). Sugar transport in fat cells. Effects of mechanical agitation cellbound insulin. and temperature. Arch. Biochem. Biophvs. 192, 120- 127. Vega. F. V., Key, R. J . , Jordan, J. E., and Kono. T. (1980). Reversal of insulin effects of fat cells may require energy for an activation of glucose transport but not for an activation of phosphodiesterase. Arch. Biochem. Biophys. 203, 167- 173. Vinten, J. (1978). Cytochalasin B inhibition and temperature dependence of 3-0-methylglucose transport in fat cells. Biochim. Biophys. Actu 511, 259-273. Vinten. J.. Gliemann, I., and Dsterlind. K. (1976). Exchange of 3-0-methylglucose in isolated fat cells. Concentration dependence and effect of insulin. J . Biol. Chem. 251, 794-800. Wardzdla, L. J., and Jeanrendud, B. (1981). Potential mechanism of insulin action on glucose transport in the isolated rat diaphragm. J . B i d . Chem. 256, 7090-7093. Wardzala, L. J.. Cushman, S. W.. and Salans, L. B. (1978). Mechanism of insulin action on glucose transport in the isolated rat adipose cell. J . B i d . Chem. 253, 8002-8005. Wheeler. T. 1.. and Hinkle, P. C. (1981). Kinetic properties of the reconstituted glucose transporter from human erythrocytes. J . Biol. Chem. 256, 8907-8914. Whitesell, R. R., and Gliemann, J. (1979). Kinetic parameters of transport of 3-0-methylglucose and glucose in adipocytes. 1. B i d . Chem. 254, 5276-5283. Whitesell, R. R . . Tarpley, H. L.. and Regen, D. M. (1977). Sugar transport kinetics of the rat thymocyte. Arch. Biochem. Biophvs. 181, 596-602. Wick. A. N . , Drury, D. R., Nakada, H . , a'nd Wolfe, J . B. (I951). Location of the primary metabolic block produced by 2-deoxy-o-glucose. J . B i d . Chem. 224, 963-969. Widdas, W. F. (1980). The asymmetry of the hexose transfer system in the human red cell membrane. Curr. Top. Membr. Transp. 14, 165-223. Wieringa, T. J., van Putten, J . P. M . , and Krans, H. M. J. (1981). Rapid phloretin-induced dephosphorylation of 2-deoxy-o-glucose-6-phosphatein rat adipocytes. Biochem. Biophys. Res. Commun. 103, 841-847. Wilbrandt, W. (1954). Secretion and transport of nonelectrolytes. Symp. SOC. Exp. Biol. 8, 136-162. Wright, E. M.. van Os, C. H., and Mircheff, A . K. (1980). Sugar uptake by intestinal basolateral membrane vesicles. Biochim. Biophvs. Acta 597, 112-124.


CURRENT TOPICS IN MEMBRANES AND TRANSPORT. VOLUME 18

Epidermal Growth Factor Receptor and Mechanisms for Animal Cell Division MANJUSRl DAS Depurtmenr of Biochemistry and Biophysics University of Pennsylvania School of Medicine Philadelphia. Pennsylvunia

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11. Properties of E G F . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111. The EGF Receptor . . .......... A. Identification o f t .......................................... B . Clustering, Internalization, and Degradation of EGF-Receptor Complexes . . . . . C. Protein Kinase Domain of the EGF Receptor and Its Relationship to the Oncogene Product, pp6oSrc . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . D. Antibodies Directed against the EGF Receptor.. . . . . . . . . . . . . . . . . . . . . . . . . . . E. Receptor Regulation.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . F. Studies on Location of the EGF-Receptor Gene.. ......................... G. lnsertion of Exogenous EGF Receptors into Receptor-Negative Variant Cells.. . IV. The Pathway to Nuclear DNA Replication.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A. The Mitogenic Capability of EGF ............. B. The Mitogenic Pathway . . . . . . . . ............................. C. Biochemical Signals for Mitogenesis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. A Family of EGF-like Polypeptides and Their Role in Animal Development and Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References ... ........................................

1.

381 382 383 383 386 387 388 390 391 392 393 393 393 396 398 400

INTRODUCTION

One of the central problems in modern biology is the understanding of the mechanism by which extracellular molecules such as hormones, toxins, and neurotransmitters interact with surface receptors to regulate intracellular events. 381 Copyright ((1 l Y X 3 by A w k r n i c Pres. Inc All rights of reprodu~cmIn any forni rc\erved ISBN 0-12-1533 18-2

382

MANJUSRI DAS

Epidermal growth factor (EGF), a single chain polypeptide of -6000 daltons (Carpenter and Cohen, 1979), belongs to a new class of cytomodulatory factors that are hormone-like in their biological potencies. EGF stimulates DNA replication and cell division in a wide variety of cells including those of nonepidermal origin, and the EGF receptor, a cell surface polypeptide of 170,000- 180,000 daltons, has a wide tissue distribution. Among the various growth factors isolated to date, EGF is one of the most potent and best characterized as to its physical, chemical, and biological properties, and the EGF-receptor system has been an important stimulus to the development of new ideas on receptor action and mitogenesis. During the last decade there has been an exponential growth of literature and reviews on EGF (Carpenter and Cohen, 1979; Hollenberg, 1979; Adamson and Rees, 1981). The present article briefly recapitulates some of the earlier findings, and discusses in detail some of the more recent developments in this area.

II. PROPERTIES OF EGF EGF was first isolated from mouse submaxillary glands by Cohen (1962). It was described as an epidermal tissue stimulatory factor that caused precocious eyelid opening and tooth eruption in newborn mice, and was hence named epidermal growth factor. In an independent study, Gregory and his colleagues isolated a polypeptide from human urine that inhibited gastric acid secretion, and named it urogastrone (Gregory, 1975; Gregory and Willshire, 1975; Gregory and Preston, 1977). Comparison of amino acid sequences revealed regions of structural homology between the 53 residues long mouse and human polypeptides (Savage et al., 1972; Gregory, 1975). This led to a comparison of biological properties. It was found that both the mouse and human polypeptides shared tissue growth-stimulatory and gastric acid-inhibitory properties, and were capable of competing equally for the same receptor sites in a variety of animal tissues. This suggests that both peptides belong to a family of mitogenic, acid-inhibitory polypeptides that show some interspecies structural variations, but are probably near-identical in their active site regions that are responsible for receptor binding and biological activity. Normal plasma concentration of EGF in adults is 0.1-0.2 nM, and this is subject to hormonal modulations (Bynny et al., 1974; Barka et al., 1978). a-Adrenergic agents stimulate the release of EGF from submaxillary glands into plasma. Androgens increase the levels of EGF in submaxillary glands (Bynny et al., 1972; Barthe et a / . . 1974) but do not appear to stimulate its release into plasma. An in vivo role for the EGF-receptor system in embryonic and organ development is suggested by various studies on EGF binding and EGF action on embry-

EPIDERMAL GROWTH FACTOR RECEPTOR

383

onic, amniotic, placental, and other developing tissues (Hassel, 1975; Hassel and Pratt, 1977; Ladda et al., 1979; Nexo et al., 1980). In adult animals, EGF in the gastrointestinal tract is likely to mediate regulation of gastric acid secretion (Gregory, 1975) and replenishment of the rapidly turning over intestinal epithelial cells (Forgue-Laffite et al., 1980). In addition, EGF in adults could mediate other vital processes, the nature of which is not yet known. Given the wide tissue distribution of the EGF receptor (O'Keefe et al., 1974), it is tempting to propose that EGF plays an important role as fundamental as other hormones in the well being and survival of animals.

111.

THE EGF RECEPTOR

A. Identification of the Receptor Specific and high-affinity receptors for EGF are present in a wide variety of cells including those of nonepidermal origin. Rapid and saturable binding of '251-labeled EGF has been demonstrated in fibroblasts (Hollenberg and Cuatrecasas, 1973; Carpenter et al., 1979, corneal cells (Frati et al., 1972; Gospodarowicz et al., 1977), lens cells (Hollenberg, 1975), kidney cells (Holley et al., 1977), intestinal epithelial cells (Forgue-Laffite et al., 1980), human glial cells (Westermark, 1977), 3T3 cells (Pruss and Henchman, 1977; Aharonov et al., 1978), granulosa cells (Vlodavsky et al., 1978), human epidermal carcinoma cells (Fabricant et d . , 1977), and human vascular endothelial cells (Gospodarowicz et al., 1978). Apparent dissociation constants for binding are in the range of 0.1-1 nM (Hollenberg and Cuatrecasas, 1973; Carpenter et al., 1975; Aharonov et al., 1978). The number of receptor sites per cell varies from 4- 10 x lo4 in fibroblastic cells (Hollenberg and Cuatrecasas, 1973; Carpenter et al., 1975; Das et al., 1977) to 1-2 X lo6 in human epidermal carcinoma cells (Haigler et al., 1978). The EGF receptor was first identified using a chemical cross-linking technique (Das et al., 1977; Das and Fox, 1978). Photoreactive derivatives of EGF were used to label and identify specifically the membrane receptor for EGF in murine 3T3 cells. Photoreactive arylazide derivatives of radioiodinated EGF were prepared using arylazide heterobifunctional cross-linking reagents, which are useful in identification of ligand-binding components in complex biological systems (Das and Fox, 1979). Photoactivable EGF, labeled with IZ5I, was incubated with 3T3 cells and then photolyzed in situ to generate a nitrene capable of reacting with a wide variety of chemical bonds. Analysis of the system by sodium dodecyl sulfate-polyacrylamide gel electrophoresis revealed besides the band of EGF, only one other major radioactive band at a position indicating an apparent molecular weight of 190,000. This band was absent when a nonresponsive and

384

MANJUSRI DAS

nonbinding variant of 3T3 was used. A direct proportionality between binding activity and cross-linked complex formation was demonstrated using a variety of binding conditions. The photoactivable derivative of EGF thus acted as a typical affinity label for its receptor, and there appeared to be only one protein (M, 184,000) involved in specific recognition and binding of EGF to 3T3 cells (Das et al., 1977) (Table I). This finding on receptor molecular weight was confirmed by Sahyoun et al. (1978) and Hock et al. (1979), who undertook the labeling of the 1251-labeled EGF binding components in placental and liver membranes by glutaraldehyde cross-linking followed by sodium borohydride reduction (Table I). In experiments with human placental membranes (Hock et al., 1979), two labeled components of M, 160,000 and 180,000 were observed. The latter value is in good accord with the value of M, 184,000 for the murine receptor (Das et ul., 1977). Same labeled components (M, 160,000 and 180,000) were observed after labeling human placental membranes with a photoaffinity analog of EGF. It was suggested that the two constituents observed in human placenta could be interrelated by either a biosynthetic or a degradative process. More recently, Cohen et ul. (1980, 1982) purified the human EGF receptor by using a procedure involving solubilization and affinity purification. Plasma membranes from human A-431 carcinoma cells which are exceptionally rich in EGF receptors were used in these studies. The receptor was solubilized with Triton X-100 and was purified by affinity chromatography on columns of agarose containing covalently bound EGF. Plasma membranes prepared using the procedure of Thom et al. (1977) yielded a receptor protein of M, 150,000 (Cohen et al., 1980). This receptor molecular weight is slightly smaller than the earlier reported molecular weights on the A-431 receptor (Wrann and Fox, 1979), the human placental receptor (Hock et al., 1979), and the murine 3T3 receptor (Das el al., 1977) (Table I). However, when plasma membrane vesicles were prepared using a rapid “hypotonic shedding procedure,” a higher receptor molecular weight (170,000) was observed (Cohen et al., 1982). It was suggested that the 170,000 M, protein is proteolytically degraded to a 150,000 form which retains its EGF binding function. A glycoprotein structure for the EGF receptor was proposed based on the ability of various lectins to inhibit reversibly the binding of ‘2sI-labeled EGF to human fibroblasts and to placental membranes (Carpenter and Cohen, 1977). In fact, lectin affinity columns have been useful in effecting considerable purification of the receptor (Hock ef al., 1980). Additional evidence for a glycoprotein structure comes from the finding that treatment of cells with tunicamycin, a potent inhibitor of dolichol-mediated glycosylation, results in a progressive loss of EGF-receptor activity (Bhargava and Makman, 1980). Also a mutant 3T3 cell line defective in protein glycosylating activity was found to be deficient in EGF binding activity (Pratt and Pastan, 1978).

TABLE I MOLECULAR WEIGHTDETERMINATIONS ON THE EGF RECEPTOR Source Murine 3T3 cells Human placental membranes Human placental membranes Human Munne Human Human

foreskin fibroblasts 3T3 cells A-431 carcinoma cells A-43 1 carcinoma cells

Method Specific labeling of the receptor with photoaffinity analogs of 1251-labeledEGF Glutaraldehyde cross-linking of 1251-labeled EGF to the receptor Specific labeling of the receptor with photoaffinity analogs of '251-labeled EGF Direct labeling of the receptor with 1251-labeled EGF Direct labeling of the receptor with 1251-labeled EGF Direct labeling of a receptor with ~251-labeledEGF Solubilization of the receptor from membranes followed by its affinity purification on EGF-agarose columns

Molecular weight 184,000 160,000 and 180,000 160,000 and 110,000 184,000 184,000 175.000 170,000

Reference Das er a/.(1977)

Hock et af. (1979) Hock er af. (1979) Baker el al. (1979) Linsley et a/.(1979) Wrann and Fox (1979) Cohen et a/.(1982)

386

MANJUSRI DAS

An interesting property of the EGF receptor is its ability to become covalently attached to bound EGF at 37°C (Baker et al., 1979; Linsley et al., 1979). This has allowed identification of the receptor in A-43 1 cells (Wrann and Fox, 1979). Only a portion of cell-bound EGF becomes covalently linked in this fashion, and known inhibitors of transglutaminase do not inhibit this reaction. The biological role or nature of this covalent linkage remains unclear.

B. Clustering, Internalization, and Degradation of EGF-Receptor Complexes The EGF-receptor system has been an important stimulus to the development of ideas on the mechanism of receptor-mediated ligand endocytosis. Carpenter and Cohen ( 1976a) showed that at 37°C cell bound 12sI-labeledEGF is degraded very rapidly with the appearance in the medium of [ 12sI]monoiodotyrosine.The degradation at 37°C is blocked by inhibitors of metabolic energy production (azide, cyanide, dinitrophenol) and by a lysosomotropic agent (chloroquine). This suggests that cell surface bound 12sI-labeledEGF is rapidly internalized in an energy-dependent step, and then degraded within lysosomes (Carpenter and Cohen, 1976a). Direct visualization of the process of internalization has been reported by Gordon et af. (1978), Schlessinger et al. (1978), and Haigler et al. (1979). Different techniques were used in these studies, namely, electron microscope autoradiography of 12sI-labeledEGF, tracing of fluorescent derivatives of EGF, and visualizing EGF-ferritin conjugates by electron microscopy. In each case, binding of EGF to dispersed receptors on the cell surface was shown to be followed by surface aggregation and subsequent internalization of the EGF label into pinocytic vesicles or “receptosomes” (Pastan and Willingham, 198I), leading ultimately to its appearance in lysosome-like structures. The studies described above strongly indicate that EGF binding to dispersed receptors on the cell surface leads to receptor clustering and endocytosis. A direct study on the fate of the receptor was performed using the photoaffinity labeling approach outlined earlier (see Section 111,A). Murine 3T3 cells carrying in situ radiolabeled receptor (prepared using a photoreactive derivative of EGF) were incubated at 37°C for increasing time intervals (Das and Fox, 1978). There was a time-dependent reduction of radioactivity from the radiolabeled receptor band of M, 190,000, and the loss was accompanied by the appearance of three distinct low-molecular-weight bands of M, 62,000, 47,000, and 37,000. The radioactivity lost from the receptor band was recovered almost quantitatively from the low-molecular-weight bands, suggesting a precursor-product relationship between these proteins. Subcellular fractionation of cells containing the radiolabeled receptor and its degradation products revealed that the low-molecular-weight proteins banded in sucrose gradients with lysosomes, whereas the


387

receptor cofractionated with the plasmalemmal fraction. These results suggest an endocytic degradative fate for the surface receptor after binding to EGF. It should be noted, however, that the rate of receptor processing/degradation in 3T3 cells is slow (about one-fifth) compared with the rate of EGF degradation. This suggests that most (about 80%) of the endocytosed receptors are perhaps not degraded, but recycled back to the plasma membrane.

C. Protein Kinase Domain of the EGF Receptor and Its Relationship to the Oncogene Product, pp6PrC Addition of EGF to A-43 1 plasma membranes results in a marked stimulation of cyclic nucleotide-independent phosphorylation of endogenous membrane proteins, including the EGF receptor (Carpenter er al., 1978, 1979; King et al., 1980b; Ushiro and Cohen, 1980; Cohen et al., 1980). The reaction was found to involve specific phosphorylation of tyrosine residues in substrate proteins, and that put it outside the common class of protein kinases which phosphorylate serine and threonine. Treatment of A-43 1 plasma membranes with Triton X-100 results in solubilization of both the EGF receptor and the kinase activity. Purification of the receptor on EGF-affinity columns results in a copurification of the kinase activity, suggesting a tight association between the receptor and the kinase (Cohen et al., 1980). More recently, it has been shown that antibodies directed against the EGF receptor can coprecipitate both the 170,000-dalton receptor protein and the kinase activity (Cohen et al., 1982). This strongly suggests that EGF binding activity and kinase activity are covalently linked, and that both activities may reside in the same 170,000-dalton polypeptide, in different domains within the same molecule. Linsley and Fox (1980) showed that EGF receptors on intact A-43 1 cells are autophosphorylated only when the cells are permeabilized with lysolecithin. More recently it has been shown that the purified 170,000-dalton receptor has a good capacity for autophosphorylation, but the degradation product of 150,000 daltons (which can bind EGF, but presumably has lost the tail end of the molecule that extends into the cytoplasm) is a poor substrate for autophosphorylation (Cohen et al., 1982). However, when challenged by exogenous substrates, the 150,000-dalton receptor is a better kinase compared with the 170,000-dalton receptor. This could be due to greater availability of phosphorylation sites in the 170,000-dalton receptor compared with the 150,000 form, which would account for the lower apparent kinase activity of the 170,000 form toward exogenous substrates (Cohen et al., 1982). Thus, the EGF receptor appears to be a multifunctional, multidomain protein, whose inherent kinase function is activated after EGF interacts with its external binding site. Cohen et al. (1982) compared the activation of receptodkinase by

388

MANJUSRI DAS

EGF "to the activation of ribonuclease S by S peptide, where peptide binds to the cleaved ribonuclease and converts an inactive enzyme to an active one." This raises the question of whether EGF and its receptors were parts of one functional protein whose functional domains became separate during evolution. In its specificity for tyrosine phosphorylation, the EGF receptodkinase resembles the tyrosine-specific kinase activities associated with the transforming proteins (oncogene products, pp60v-src) of several RNA tumor viruses (Hunter, 1980; Bishop, 1981; Erikson and Erikson, 1980); i.e., tyrosine-specific protein phosphorylation appears to be intimately associated with both virus-induced cell transformation and the action of a normal stimulant of cell division, EGF [recent studies (Ek et al., 1982) suggest that the action of platelet-derived growth factor, another normal stimulant of cell division, also involves tyrosine-specific phosphorylation]. Although oncogenes were first found in viruses, subsequent studies revealed the gene to be present in normal nontransformed cells. The product of cellular oncogene (named pp60c-src)was found to be indistinguishable (in terms of structure or activity) from the viral product p ~ 6 0 ~ -The " ~ ~version . of c-src found in fishes, birds, and mammals are all closely related to the viral gene v-src and to one another. The small amounts of pp60-src found in normal cells is not sufficient to induce cellular transformation, but it may well be required for the well being of the cells, as suggested by the survival of c-src through long periods of evolution. Perhaps cellular oncogenes (i.e., the genes for tyrosine-specific protein kinases) are part of a delicately balanced network of controls that regulate the growth and development of normal cells. Excessive activity of one of these genes might tip the balance of regulation toward incessant growth. The kinase domain of the EGF receptor (and perhaps of other growth factor receptors) may have evolved from an ancestral oncogene, and there may exist certain structural similarities between the oncogene product, pp6OC-"" or pp6OV-"' , and the mitogenic message transmitting kinase domain of the EGF receptor.

D. Antibodies Directed against the EGF Receptor In 1980, Haigler and Carpenter reported the preparation of an anti-EGFreceptor antiserum which was obtained after immunization of rabbits with human A-43 1 carcinoma cell membranes. The IgG fraction of this immune serum blocked 1251-labeledEGF binding to human and murine EGF receptors and also blocked the induction of DNA synthesis in quiescent fibroblasts by EGF. However, this antiserum was not receptor-specific and was capable of immunological interactions with nonreceptor proteins. Recently, however, three other groups have reported the preparation of specific anti-EGF-receptor antibodies (Schreiber et al., 1981a; Cohen et al., 1982; Carlin and Knowles, 1982). Schreiber et af. (1981a) have described the preparation of monoclonal murine


389

antibodies against the human A-431 EGF receptor. The antibodies, of IgM type, were capable of inhibiting 1251-labeledEGF binding to both human and murine EGF receptors. In addition, these monoclonal IgM antibodies induced EGF-like biological effects. Like EGF, they enhanced protein phosphorylation in A-43 1 membranes, and stimulated DNA synthesis in human fibroblasts. In their ability to activate the EGF receptor, they resemble the anti-insulin-receptor antibodies that have been shown to exert potent insulin-like effects on cells (Kahn et a/., 1977). Although these monoclonal IgM antibodies are EGF-like, polyclonal IgG antibodies directed against the EGF receptor are incapable of producing any EGF-Iike biological effects (Haigler and Carpenter, 1980; Carlin and Knowles, 1982). Schreiber et a/. (1981a) suggest that this difference could be a consequence of the better receptor cross-linking ability of the decavalent IgM compared with the bivalent IgG antibodies. A different type of anti-EGF-receptor antibody was prepared by Cohen et a/. (1982). The affinity-purified A-431 EGF receptor was subjected to SDS-gel electrophoresis and the Coomassie blue-stained 170,000-dalton band in the gel was minced and used for immunizing rabbits. The antiserum obtained against the denatured receptor was capable of immunoprecipitating the solubilized EGF receptor and the associated kinase, but it did not inhibit the binding of 1251labeled EGF to the A-43 1 receptor, and did not inhibit basal or EGF-stimulated phosphorylation. Yet another type of anti-EGF-receptor antibody became available through a serendipitous route. It has been known for some time that specific IgG antibodies are produced against a M , 165,000 human protein when human-mouse somatic cell hybrids (containing chromosome 7 as the only human chromosome) are injected into syngeneic mice (Aden and Knowles, 1976; Ford et af., 1978). After it became known that the human EGF-receptor gene was associated with chromosome 7 (see Section II1,G) (Shimizu et a / . , 1980; Davies et al., 1980), the antibody was tested for anti-EGF-receptor activity, and it was found to be a potent inhibitor of EGF binding in human cells but not in murine cells (Carlin and Knowles, 1982). Immunoaffinity chromatography of human A-43 I cellular proteins on an antibody-agarose column resulted in the purification of a protein ( M , 175,000) which comigrated with the human EGF receptor during electrophoreses under reducing, denaturing conditions (Das et a/., 1982). The interaction of this antibody with the human EGF receptor was studied in further detail using affinity-purified 1251-labeledantibodies (Das et al., 1982). The IgG fraction of antisera was labeled with 1251. The 1251-labeledantireceptor antibody, which initially represented about 0.2% of total I2'I-labeled IgG, was enriched by selective adsorption to and subsequent elution from human WI-38 cells which contain EGF receptors. The purified I 251-labeled antireceptor antibody bound to human tissue culture cells (A-43 1 and human fibroblasts) and to human placental membranes in a time-, temperature-, and concentration-dependent manner. No

390

MANJUSRI DAS

binding was observed with the murine EGF receptor in 3T3 cells or in hepatic membranes. The binding of antibody to human cells was inhibited by unlabeled antibody and EGF, but not by nonimmune mouse IgG or hormones such as insulin and fibroblastic growth factor (FGF). For human EGF receptors of diverse origin (fibroblasts, A-43 1, and placenta) the ratios of 1251-labeledantibody binding activity to 12sI-labeledEGF binding activity were about the same, suggesting a close molecular similarity between these receptors from different sources. The preparation of high specific activity It51-labeled antireceptor antibodies by cytoadsorption and elution thus provides a sensitive method for detection and characterization of receptors.

E. Receptor Regulation 1. EGF-INDUCED REGULATION

A “down-regulation” phenomenon, quite similar to other receptor-macromolecule interactions such as antigenic modulation, has been observed with the EGF-receptor system (Carpenter and Cohen, 1976a; Aharonov et al., 1978; Das and Fox, 1978). It has been suggested that the hormone-induced loss of receptors is due to endocytic remova: of surface EGF-receptor complexes, without the concomitant production of new receptors. It could also be partly due to an EGFinduced drastic increase in receptor affinity leading to the formation of nondissociable EGF-receptor complexes. Down-regulated cells can be stimulated to regain receptors by removal of EGF and addition of serum (Carpenter and Cohen, 1976a). A 100% regain of receptors can be achieved within 9 hours. This recovery process is inhibited by cycloheximide or actinomycin D. This suggests the involvement of de now synthesis of receptors or of nonreceptor labile proteins which play a crucial role in receptor recycling. 2. REGULATIONBY STRUCTURALLY UNRELATEDAGENTS Lee and Weinstein ( 1978) demonstrated that tumor-promoting phorbol esters caused a rapid and marked inhibition of EGF binding to its receptors. It was initially thought that the reduction in binding was due to a decrease in the number of available EGF-receptor sites; but later it was shown that the number of EGFreceptor sites was not reduced, only the binding affinity was markedly decreased (Lee and Weinstein, 1979; Brown et al., 1979). The inhibitory effect was reversed upon the removal of phorbol esters from the medium. The effect of phorbol esters is EGF receptor-specific, because other non-EGF-


391

receptor activities are unaffected. The inhibitory effect of phorbol esters is temperature sensitive. It is inhibitory at 37°C but not at 4°C. More recently, it has been shown that vitamin K, (a quinone) and vasopressin (a neurohypophyseal nonapeptide hormone) also markedly reduce the affinity of EGF receptors for '251-labeled EGF in a time- and temperature-dependent fashion (Shoyab and Todaro, 1980; Rozengurt e t a l . , 1981a). The properties of the inhibition of EGF binding by these agents have many similarities with those of the phorbol ester family. The inhibition of EGF binding by these agents does not require protein synthesis or degradation, but is completely blocked by reducing the temperature to 4°C. Such findings have led investigators to suggest that these inhibitory agents (vitamin K,, vasopressin, phorbol ester) bind to sites which are separate from the EGF receptor. (This is consistent with the absence of any structural analogy between these molecules.) It was proposed that the phorbol esterhasopressidvitamin K,-occupied receptors interact with the EGF-receptor sites in a diffusionally controlled, temperature-sensitive step, and thereby reduce the affinity of the EGF receptors for EGF.

F. Studies on Location of the EGF-Receptor Gene Somatic cell genetic techniques were used for studies on chromosomal location of the EGF-receptor structural gene (Shimizu et al., 1980; Davies et al., 1980). Fusion of mouse and human cells results in the formation of hybrids that usually retain all the mouse complement of chromosomes and a small random subset of human chromosomes. Because both the murine and human genes are usually functional, one can assay each hybrid clone for the presence of a given gene product. In these studies mouse cell mutants deficient in hypoxanthine phosphoribosyltransferase, and devoid of '251-labeled EGF binding activity were fused with human diploid cells, possessing EGF binding ability. The humanmouse cell hybrids were isolated after hypoxanthine/aminopterin/thymidineselection. Analysis of isozyme markers and chromosomes of a number of these human-mouse clones indicated that the expression of EGF binding ability is correlated with the presence of human chromosome 7. These results suggest that a gene on chromosome 7 could code for human EGF receptor or complement a deficiency in the mutant mouse cells. Immunologic analysis confirmed that the receptor in these human-mouse clones is nonmurine, and of human origin (Carlin and Knowles, 1982). A number of EGF-receptor-negative variants of murine 3T3 cells have been isolated by Pruss and Henchman (1977) using the colchicine selection technique. These variants have served as excellent specificity controls for studies on identification and insertion of the EGF receptor (Das et al., 1977; Bishayee et

392

MANJUSRI DAS

a / . , 1982). However, the genetic lesions that are responsible for the

loss of

receptor activity in these cells remain uncharacterized.

G. Insertion of Exogenous EGF Receptors into ReceptorNegative Variant Cells Polyethyleneglycol-mediated membrane fusion techniques have been used for putting foreign receptors into recipient cells, and for combining components from two different types of cells for studies on coupling between hormone receptor complexes and adenyl cyclase (Schramm e r a / ., 1977). It is intriguing to note that an exogenous EGF receptor can be inserted into a recipient cell (in a biologically active orientation) by a novel mechanism requiring no added fusogenic agent (Bishayee er al., 1982). A variant cell line NR-6, derived from mouse 3T3, can neither bind nor biologically respond to EGF (Pruss and Herschman, 1977). When these NR-6 cells were incubated with EGF-receptor-rich mouse hepatic membranes at 26°C for 6 hours in the absence of any added fusogen, there was a transfer of almost 20% of input EGF receptors to NR-6 cells, whereas only 1-2% of bulk hepatic proteins were transferred in a similar fashion. The results suggest a preferential insertion of the EGF receptor over the other hepatic proteins. Experiments with cycloheximide and tunicamycin suggest that the receptor gain by the NR-6 cells was not due to an activation of endogenous protein synthesis or to a glycosylation-induced activation of preexisting aglyco-receptors in NR-6 cells. The inserted receptor bound 12s1-labeledEGF with high affinity, and EGF was found to stimulate D N A synthesis (fourfold maximally) and cell division (twofold maximally) in these membrane-treated NR-6 cells in a concentration-dependent manner (this biological stimulation is dependent upon the quality of the membranes and is not always reproducible). In contrast, NR-6 cells not treated with hepatic membranes were totally unresponsive to EGF. These studies suggest the existence of a natural (affinity-mediated’?) mechanism for specific receptor transfer. Since EGF receptor is an integral membrane protein (detergents are required for its solubilization) it is not easy to visualize a mechanism for its insertion in the absence of added fusogens. It is possible that the preferential insertion of EGF receptors over other hepatic proteins is due to a specific NR-6 membrane protein which possesses a high affinity for the receptor and sequesters the receptor from the hepatic membrane by an unknown mechanism. Such a protein with a tendency to associate with the receptor may also be involved in the biological message transmission mechanism. Therefore, it is of interest to examine the existence of such a protein using purified EGF receptor as a probe.


393

IV. THE PATHWAY TO NUCLEAR DNA REPLICATION A. The Mitogenic Capability of EGF A variety of cell types display a DNA replication response to EGF. This included fibroblasts (Armelin, 1973; Hollenberg and Cuatrecasas, 1973; Carpenter and Cohen, 1976b; Rose etal., 1976; Das and Fox, 1978), glia (Westermark, 1977). lens epithelial cells (Stoker et a l . , 1976). endothelial cells (Gospodarowicz et al., 1978), and kidney cells (Holley et a / . , 1977). It is known that EGF must be continuously present in cell medium for 5 hours for even a small level of DNA synthesis, and for the elicitation of a near-maximal DNA synthetic response a 12- to 15-hour exposure is required (Carpenter and Cohen, I976b). A comparison with other systems (those involving synergistic interactions between different mitogens) reveals the following. It has been shown that in quiescent responsive cells, a transient exposure to platelet-derived growth factor (PDGF) followed by a later exposure to plasma results in G I -+S transi1979; Vogel er a / . , 1978). This led to the suggestion that tion (Stiles et d., PDGF serves to mediate only the earlier events in the mitogenic pathway, and other factors such as those present in plasma are needed for progression through the rest of the pathway (Stiles et a / . , 1979). although Dicker and Rozengurt ( I98 1 ) suggest that a tight and stable association between PDGF and its receptor could make it possible to act at later stages in the mitogenic pathway despite early removal of PDGF from the culture medium. In any event, in the EGF system, both the early and late events in the mitogenic pathway appear to require the presence of EGF in the medium. This suggests that in this system, the mitogen-receptor functional complex is relatively labile, and that one (or more) of the EGF-receptor-generated biochemical signals for transit through the later stages of the G , -+ S pathway is labile. Thus, both early and late events in the mitogenic pathway may be dependent upon the presence of critical concentrations of appropriate signals generated by the EGF-receptor complex.

6. The Mitogenic Pathway Despite the apparent diversity of different mitogenic systems, it appears likely that the cellular protein components other than the receptor that are involved in the expression of mitogenic responses (hormone-induced or otherwise) are very similar in different species of normal and tumor cells. The G I phase of the cell cycle has been shown to be specifically lengthened in the presence of cycloheximide, whereas the remainder of the cell cycle (S, G,, and M ) is only slightly

394

MANJUSRI DAS

1971; Rossow et a/., 1979). These results lengthened (Schneiderman et d., suggest the existence of cycloheximide-sensitive, rapidly turning over G I proteins that can regulate entry into S phase. Considering these, it is of interest to identify the EGF-induced rapidly turning over proteins whose synthesis is markedly enhanced in the presence of EGF, and whose decay occurs abnormally fast in the absence of any mitogenic stimulation. 1. PROTEINS PARTICIPATING I N THE EGF-INDUCED PATHWAY

Some EGF-induced protein factor in cytoplasmic extracts of EGF-treated cell has been shown to stimulate DNA synthesis in sensitive (cell-free) nuclei (Das, 1980). Stationary density-inhibited cultures of 3T3 cells contain only insignificant amounts of the activator of DNA replication as measured by the cell-free assay, However, addition of increasing amounts of EGF to these contact-inhibited 3T3 cells results in an increasing intracellular production of the activator of DNA replication. This EGF-induced increase in activity is inhibited by cycloheximide, suggesting that the increase is mostly due to an enhanced rate of biosynthesis. The concentration of EGF required for half-maximal induction of the activator substance in quiescent 3T3 cells is about 0.1 nM,which is very similar to that required for half-maximal mitogenic response in 3T3 and other animal cells, suggesting a functionally important role for this factor in the initiation of growth and proliferation. The activity is trypsin-sensitive and nondialyzable, and sucrose density gradient centrifugal analysis reveals three peaks of activity corresponding to molecular weights of 46,000, 110,000, and 270,000 (Das, 1980). Activities very similar to the EGF-induced activity described above have been reported to be present in embryonic and tumor cells (Jazwinski et a / . , 1979; Benblow and Ford, 1973, and in concanavalin A-stimulated lymphocytes (J. Gutokowski and S . Cohen, personal communication). As previously suggested the EGF-induced pathway to mitogenesis may be very similar to that induced during tumorigenesis and other mitogenic stimulations. Other proteins that appear to be induced during EGF stimulation include a family of secreted glycoproteins of 34,000 daltons (Nilsen-Hamilton et a / ., 1980). The appearance of these glycoproteins in 3T3 cell medium closely correlates with the DNA synthetic response to EGF and other mitogens such as FGF. The biological function of these secreted proteins remains unclear at present. Recently it has been shown that EGF can stimulate poly(ADP-ribosylation) in 3T3 cells (presumably through induction of the appropriate proteins), and the stimulation appears to be temporally correlated with the cells’ entry into DNA synthesis (Shimizu and Shimizu, 1981). Poly(ADP-ribosylation) has been shown to occur on nuclear proteins including histones and nonhistone proteins (Sugimura, 1973; Hayaishi and Ueda, 1977; Burzio et al., 1979), and it is of interest

395


to examine whether this brings about any alteration in chromatin structure and DNA replicative potential.

2. COMMITMENT TO DNA REPLICATION The mitogenic pathway, despite its enormous complexity and multicomponency, leads ultimately to the formation of a “committed” prereplicative cellular state (Temin, 1971) that appears to be an innate inherent property of the cell type independent of the external mitogen used for achieving commitment to replicate DNA (Das, 1981). It was observed that decay of the induced DNA synthetic ability induced with EGF, serum, or other mitogens is an exponential, first-order process. Internal commitment produced with either EGF or serum had identical half lives and the half-time was the same irrespective of the initial degree of commitment (Das, 198 1). These suggest the production of a preprogrammed common internal state in response to varying levels of diverse stimuli. Left to itself under normal conditions, the committed state would lead to a complete round of DNA replication accompanied perhaps by its own dissolution; but in the presence of inhibitors of DNA synthesis (e.g., hydroxyurea), the state decays in a single step. Thus commitment represents a distinctive state within the cell, a global property of the whole cell. Achieving this end-state must be a multistep

@ t t T r

DNA replication

I

t

zt t t

r

FIG. I . Commiited (C*) state model. The ultimate committed (C*) state appears to be a unit or global property o f the whole cell rather than. for example, a critical concentration of some active induced moleculc. The committed state i s perhaps coded into the configuration of some cellular macrostructure ruch as nuclear membrane or a part of the genome. which might require surpassing a critical concentration of active molecules earlier in the pathway. but the state i t s e l f represents a yes/ no whole-cell deciwn. Normally. attainment of the C* state would lead to DNA replication and suhsequent cell division; but. in the presence o f inhibitors (hydroxyureakxcess thymidine). DNA replication i s blocked and the state decays. ESI. external stimulant (EGF)-induced commitment pathway.

396

MANJUSRI DAS

process, and perhaps requires surpassing of critical concentrations of several protein factors (e.g., the EGF-induced factor described above) earlier in the pathway, but the state itself appears to decay in a single step (Fig. I ) . No presupposition is yet possible on the molecular nature of this global state. It could be coded into the configuration of some cellular macrostructure such as the nuclear membrane or a part of the genome, and it might be of interest to correlate decay of commitment with decay of individual proteins or coordinate decay of a number of different proteins.

C. Biochemical Signals for Mitogenesis Addition of EGF to quiescent cells results in an array of rapid biological changes that include activation of the protein kinase moiety within the receptor (see Section III,D), receptor clustering on the plasma membrane (see Section IlI,C), degradation of EGF within lysosomes (see Section lll,C), activation of Na+-K+-ATPase in the membrane (Rozengurt and Heppel, 1975), and increases in the active transport of nutrients such as amino acids (Hollenberg and Cuatrecasas, 1974) and glucose (Barnes and Colowick, 1976). An examination of these various candidates for signal generation reveals the following. 1 . EGF DEGRADATION

Cellular degradation of EGF is inhibited by various agents including the lysosomotrophic agent chloroquine (Carpenter and Cohen, 1976a), methylamine (Michael et ul., 1980; King er ul., 1980a), the microtubule disrupter colchicine (Brown et al., 1980), and leupeptin and antipain, both cathepsin B inhibitors (Savion ef a/., 1980). Some of these inhibitors (chloroquine, methylamine) can suppress EGF-induced DNA replication, but their suppressive effect could be related to their general cytotoxicity. Other inhibitors (colchicine, leupeptin, and antipain) are relatively nontoxic, and these compounds do not suppress EGFstimulated mitogenic responses. In fact, they somewhat potentiate EGF action. This strongly suggests a noninvolvement of EGF degradation in the signalgenerating mechanism. On the other hand, EGF degradation could serve a regulatory function by effectively reducing the amount of intact receptor-bound EGF.

2. RECEPTOR CLUSTERING AND INTERNALIZATION Earlier studies had shown that both receptor internalization and DNA synthetic stimulation had similar EGF requirements (Das and Fox, 1978; Fox and Das, 1979). Both processes were half-maximally stimulated in 3T3 cells at a EGF concentration (-0.1 nM) that caused 10% receptor occupancy. This suggested


397

that possibility that both events have the same limiting step. A more recent study (Schechter et nl.. 1979) using the cyanogen bromide (CNBr) cleaved analog of EGF (Holladay et d . . 1976) supports those observations. Murine EGF has a single methionine residue, and treatment with CNBr results in the production of two fragments which are disulfide linked. This analog can bind to EGF-receptor (although with less affinity), but it does not induce receptor clustering and is devoid of mitogenic activity. However, the addition of anti-EGF antibody to cells containing bound CNBr-EGF results in the restoration of both activities, namely, surface receptor clustering, and nuclear DNA replication. This strongly suggests that at least one of the biochemical signals necessary for induction for DNA synthesis is generated during the various stages of clusteringiendocytosis. It should be noted however, that certain cells are mitogenically unresponsive to EGF, but they can bind and internalize EGF, and are capable of receptor down regulation (Vlodavsky er a/., 1978). This suggests that clustering or endocytosis (and binding) may be necessary but not sufficient alone for induction of commitment to DNA replication. 3. EGF-ACTIVATED TYROSINE-SPECIFIC PROTEINKINASE The EGF-activated protein kinase within the receptor polypeptide autophosphorylates the EGF receptor and phosphorylates a number of other cellular proteins as well (Cohen e t a / . , 1982). An intriguing aspect of the EGF-receptor/ kinase is that similar tyrosine-specific protein kinase activities are associated with the transforming proteins ( ~ ~ 6) of 0 "several ~ ~ RNA tumor viruses (Hunter, 1980; Erikson and Erikson, 1980; Bishop, 1981). It is therefore tempting to propose that EGF-activated protein phosphorylation could serve as a second messenger in growth stimulation. However, it has been reported by Schreiber et a / . (1981b) that the cyanogen bromide-cleaved analog of EGF is as potent as EGF (at similar receptor occupancy) in enhancing protein phosphorylation, but it fails to induce DNA synthesis (and receptor clustering),. This suggests that even if EGF-induced protein phosphorylation is a necessary initiatory event, it is not a sufficient signal for the induction of DNA synthesis. Also, this shows that receptor clustering is not required for activation of the kinase. 4. N a + ENTRYA N D ACTIVATION OF Na+-K+-ATPAsE

EGF and a number of other growth factors enhance the activity of a plasma membrane-associated Na+ -K pump, as measured by an increased (ouabaininhibitable) XhRb+influx (Rozengurt and Heppel, 1975). The activation of glycolysis that is observed during EGF stimulation (Schneider et d., 1978) could be related to an alteration of this ion pump activity (Racker, 1976). It has been suggested that the stimulation of pump activity is due to an excessive entry of Na+ into the cell. Agents other than conventional growth factors, such as +

398

MANJUSRI DAS

mellitin (an amphipathic polypeptide), can also enhance Na+ influx and increase the activity of the Na+-K+ pump; and at concentrations that promote ion fluxes, mellitin stimulates DNA synthesis in quiescent cells acting synergistically with EGF or insulin (Rozengurt et al.. 1981b). These results suggest indirectly that ion fluxes may provide at least one of the signals necessary for initiation of mitogenesis. 5. MULTIPLESIGNALS I N EGF ACTION

In summary, it appears that a simple single-hit (single-signal) model involving receptor-EGF interaction with a single transducer cannot explain the complexities involved in mitogenic hormone action. It seems likely that multiple signals are necessary, and these are generated by the EGF-receptor complex at various points. The signals are needed at both the early and late stages of the G , + S transition in EGF action (see Section IV,A). Thus, clustering may generate one of these signals and another signal may be generated by protein phosphorylation, but none of these alone is sufficient for entry into the DNA synthetic phase. It is hoped that further exploration of the EGF-receptor system will lead to a better understanding of the biochemical nature and mechanism of induction of these signals and their role in the regulation of mitogenesis.

V.

A FAMILY OF EGF-LIKE POLYPEPTIDES AND THEIR ROLE IN ANIMAL DEVELOPMENT AND GROWTH

An interesting property of the EGF receptor is its ability to bind to certain transforming polypeptides that are produced by tumor cells and sarcomas (Roberts et al., 1980; Todaro et al.. 1980). A group of sarcoma- and tumor cellderived transforming polypeptides (EGF-like or MSA-like) appears to interact with either the EGF receptor or the MSA receptor (Todaro et al.. 1981), while there are other transforming factors which appear nor to compete with either EGF or MSA for the receptor sites, but whose transforming action is clearly potentiated by EGF (Colbum and Gindhurt, 1981; Guinivan and Ladda, 1979; Roberts et al., 1982). The EGF-like factors can compete with EGF for the receptor sites, but they cannot compete with EGF for the anti-EGF antibodies in radioimmunoassays. Thus these factors are antigenically different from EGF. Like EGF, these factors can induce DNA replication and cell division in normal EGF-receptor-containing responsive cells, and, in addition (and unlike EGF), they can induce transformation-specific anchorage-independent growth of cells in soft agar (Todaro et al., 1981). An EGF-like transforming factor (hTGF,), isolated from conditioned medium


399

of human metastatic melanoma cells. was found to be a single chain polypeptide of 7400 daltons (Marquardt and Todaro, 1982). Like EGF, it contains three intrachain disulfide bridges, and no free sulfhydryl groups. However, the amino acid composition of hTGF, is unique, and, unlike human or murine EGF, it lacks tyrosine and methionine, and contains three phenylalanine residues. Despite this lack of structural/antigenic resemblance, hTGF, competed equally with EGF in radioreceptor assays, and completely displaced i2sI-labeled EGF binding to the human A-43 I EGF receptors, suggesting a close similarity in the receptor binding sites of hTGF, and EGF. It has been suggested that the EGF-like polypeptides (hTGF, and sarcoma growth factor) (Marquardt and Todaro, 1982; DeLarco et al.. 1980) may be produced during tumorigenesis by reactivation of genes which are expressed only during development of the embryo (Todaro et a / . , 1981). Search for EGF-like substances in mouse embryos has revealed a discrepancy between the results of radioreceptor assay and radioimmunoassay (Nexo et a/., 1980). Only negligible quantities of EGF were detected using the radioimmunoassay, but the radioreceptor assay revealed the presence of substantial quantities of an EGF-like substance which competed with 12s1-labeled EGF for the receptor sites. This suggests the presence in embryos of a substance that can interact with the EGF receptor, but which is structurallyiantigenically different from adult EGF. Thus a variety of sarcoma- and tumor cell-released polypeptides, and substances present in embryonic tissues, are capable of interacting with the EGF receptor, although they bear no antigenic resemblance to EGF. Also the sarcoma- and tumor cellreleased factors do not appear to be viral gene products (Todaro et a / . , 1981). Thus, clearly. an animal cell has the potential for making a variety of EGFreceptor-reactive polypeptides, and it does so presumably through processes involving gene rearrangements and transpositions. The system may be analogous to the insulin-like growth factor (IGF) system where a family of partially homologous polypeptides shares the same receptor (Blundell and Humbel, 1980). What could be the biological significance of such a diversity in EGF-like substances? The answer lies perhaps in the small and subtle differences in their biological properties and in their capacities for differential receptor activation. It is hoped that further explorations of these “EGF-like polypeptides” will provide vital clues regarding the in vivo role of this most fascinating polypeptide system in animal growth and survival. ACKNOWLEDGMENTS I wish to thank Dr. Subal Bishayee for his useful suggestions. and Mr. Larry Hyland for his help in the preparation of this article. I am also grateful to Dr. Stanley Cohen (Vanderbilt University) and to Dr. George Todaro (NIH) for providing manuscripts before publication. The support of NIH research grants (AM-258 19 and AM-25724) and Research Career Development Award (AM-00693) is acknowledged.

400

MANJUSRI DAS

REFERENCES Adamson. E. D., and Rees, A . R. (1981). Epidermal growth factor receptors. Mol. Cell. Biorhem. 34, 129-152. Aden. D. P.. and Knowles, B. B. (1976). Cell surface antigens coded for by the human chromosome 7. Immunogenetics 3, 209-22 1. Aharonov. A., Pruss, R . M., and Herschman, H. R. (1978). Epidermal growth factor: Relationship between receptor regulation and mitogenesis in 3T3 cells. J . B i d . Chem. 253, 3970-3977. Armelin, H. (1973). Pituitary extracts and steroid hormones in the control of 3T3 cell growth. Proc. Natl. Acud. Sci. U.S.A. 70, 2702-2706. Baker, J. B., Simmer. R. L., Glenn, K . C., and Cunningham, D. D. (1979). Thrombin and epidermal growth factor become linked to cell surface receptors during mitogenic stimulation. Nature (London) 278, 743-745. Barka, T., Gresik, E. W . , and van der Hoen, H. (1978). Stimulation of secretion of epidermal growth factor and amylase by cyclocytidine. Cell T i m e Res. 186, 269-278. Barnes, D., and Colowick, S. P. (1976). Stiniulation of sugar uptake in cultured fibroblasts by epidermal growth factor and ECF-binding arginine esterase. J . Cell. Physiol. 89, 633-640. Barthc. 0. L., Bullock, L. P., Mowszowicz, I . , Bardin, C. W . . and Orth, D. N. (1974). Submaxillary gland epidermal growth factor: A sensitive index of biologic androgen activity. Endo: crinology 95, 1019- 1025. Benblow, R. M., and Ford, C. C. (1975). Cytoplasmic control of nuclear DNA synthesis during early development of Xenopus laevis: A cell-free assay. Proc. N u / / . Acad. S r i . U.S.A. 72, 2437-2441. Bhargava, G., and Maknian, M. H. (1980). Effect of tunicamycin on the turnover of epidermal growth factor receptors in cultured calf aorta smooth muscle cells. Biochim. Biophys. Acta 629, 107-112. Bishayee, S . . Feinman, J., Michael, H . , Pittenger, M.. and Das, M. (1982). Cell surface insertion of exogenous epidermal growth factor receptors into receptor-negative mutant cells: Demonstration of insertion in the absence of added fusogenic agents. Proc. Natl. Acud. Sci. U.S.A. 79, 1893- 1897. Bishop, J. M. (1981). Enemies within: The genesis of retroviral oncogenes. Cell23, 5-6. Blundell, T. L., and Humbel, R. E. (1980). Hormone families: Pancreatic hormones and honiologous growth factors. Nature (London) 287, 78 1-787. Brown. K. D.. Dicker, P., and Rozengurt, E. (1979). Inhibition of epidermal growth factor binding to surface receptors by tumor promoters. Biochem. Biuphys. Res. Commun. 86, 1037-1043. Brown, K. D., Friedkin. M., and Rozengurt, E. (1980). Colchicine inhibits epidermal growth factor degradation in 3T3 cells. Pror. Natl. Acad. Sci. U.S.A. 77, 480-484. Burzio, L. O., Riquelnie, P. T., and Koide, S . S. (1979). ADP ribosylation of rat liver nucleosomal core histomes. J . B i d . Chem. 254, 3029-3037. Bynny, R. L., Orth, D. N., and Cohen. S . (1972). Radioimmunoassay of epidermal growth factor. Endocrinology 90, 1261- 1266. Bynny. R. L., Orth, D. N., Cohen, S . , and Doyne, E. S . (1974). Epidermal growth factor: Effects of androgen and adrenergic agents. Endocrinology 95, 776-782. Carlin, C., and Knowles, B. B. (1982). Identity of the human EGF receptor with SA-7: Evidence for differential phosphorylation of the two components of the EGF receptor from A43 I cells. Prac. Natl. Acud. Sci. U.S.A. 79, 5026-5030. Carpenter, G., and Cohen, S . (l976a). ‘*sI-Labeled human epidermal growth factor: Binding, internalization and degradation in human fibroblasts. J . Cell B i d . 71, 159-171. Carpenter, G., and Cohen, S. (1976b). Human epidermal growth factor and the proliferation of human fibroblasts. J . Cell. Physiol. 88, 227-238.


40 1

Carpenter, G . , and Cohen, S. (1977). Influences of lectins on the binding of 12'I-labeled EGF to human fibroblasts. Eiochem. Eiophys. Res. Commitn. 79, 545-552. Carpenter, G., and Cohen. S. (1979). Epidermal growth factor. Annu. Rev. Eiorhern. 48, 193-216. Carpenter, G . , Lembach, K. J., Morrison, M. M.. and Cohen, S. (1975). Characterization of a binding of 12sI-labeled epidermal growth factor to human fibroblasts. J. Eiol. Chem. 250, 4297-4304. Carpenter, G . , King, L., and Cohen, S. (1978). Epidermal growth factor stimulates phosphorylation in membrane preparations in v i m . Nature (London) 276, 409-4 10. Carpenter, G . , King. L., and Cohen, S. ( 1979). Rapid enhancement of protein phosphorylation in A 43 I cell membrane preparations by epidermal growth factor. J. B i d . Chem. 254, 4884-489 I . Cohen. S. (1962). Isolation of a mouse submaxillary gland protein accelerating incisor eruption and eyelid opening in the newborn animal. J . B i d . Chem. 237, 1555-1562. Cohen, S., and Carpenter. G. (1975). Human epidermal growth factor: Isolation and chemical and biological properties. Proc. Natl. Acad. Sci. U.S.A. 72, 1317- 132 I . Cohen, S., and Stasny, M. (1968). Epidermal growth factor. 111. The stimulation of polysome formation in chick embryo epidermis. Eiochim. Eiophys. Acta 166, 427-437. Cohen, S., Carpenter. G., and King, L. (1980). Epidermal growth factor-receptor-protein kinase interactions. Co-purification of receptor and epidermal growth factor-enhanced phosphorylating activity. J. Eiol. Chem. 255, 4834-4842. Cohen, S., Ushiro, H., Stoscheck, C., and Chenkers, M. (1982). A native 170,000 epidermal growth factor receptor-kinase complex from shad plasma membrane vesicles. J. Eiol. Chem. 257, 1523- 1539. Colburn, N. H., and Gindhurt, T. D. (1981). Specific binding of transforming growth factor correlates with promotion of anchorage independence in EGF-receptor less mouse JB6 cells. Eiochem. Eiophys. Res. Commun. 102, 799-807. Das, M. (1 980). Mitogenic hormone-induced intracellular message: Assay and partial characterizdtion of an activator of DNA replication induced by epidermal growth factor. Proc. Narl. Acad. Sci. U.S.A. 77, 112-1 16. Das. M. (1981). Initiation of nuclear DNA replication: Evidence for formation of a committed prereplicative cellular state. Proc. Narl. Acud. Sci. U.S.A. 78, 5677-5681. Das, M., and Fox, C. F. (1978). Molecular mechanism of mitogen action: Processing of receptor induced by epidermal growth factor. Proc. Natl. Accid. Sci. U.S.A. 75, 2644-2648. Das, M., and Fox, C. F. (1979). Chemical cross-linking in biology. Annu. Rev. Eiophvs. Eioeng. 8, 165- 193. Das, M., Miyakawa. T., Fox, C. F., Pruss, R. M., Aharonov, A., and Herschman. H. R. (1977). Specific radiolabeling of a cell surface receptor for epidermal growth factor. Proc. Natl. Acad. Sci. U.S.A. 74, 2790-2794. Das, M., Hyland, L., and Bishayee, S. (1982). Anti-EGF-receptor antibody: Binding and biological properties. In preparation. Davies. R. L., Grosse, V. A , , Kucherlapati, R., and Bothwell, M. (1980). Genetic analysis of epidermal growth factor action: Assignment of human epidermal growth factor receptor gene to chromosome 7. Proc. Natl. Acad. Sci. U . S . A . 77, 4188-4192. DeLarco, J. E.. Reynolds, R., Carlberg, K., Engle. C., and Todaro, G. (1980). Sarcoma growth factor from mouse sarcoma virus transformed cells: Purification by binding and elution from epidermal growth factor receptor-rich cells. J. Eiol. Chem. 255, 3685-3690. Dicker, P., and Rozengurt, P. (1981). Stimulation of DNA synthesis by transient exposure of cell cultures to TPA or polypeptide mitogens: Induction of competence or incomplete removal. J. Cell. Physiol. 109, 99-109. Ek, B., Westermark, B.. Wasteson, A . , and Heldin, C. H. (1982). Stirnulation of tyrosine-specific phosphorylation by platelet-derived growth factor. Nature (London) 295, 4 19-420.

402

MANJUSRI DAS

Erikson, E., and Erikson, R . L. (1980). Identification of a cellular protein substrate phosphorylated by the avian sarcoma virus transforming gene product. Cell 21, 829-836. Fabricant, R . N.. DeLarco, J . E., and Todaro. (7. J. (1977). Nerve growth factor receptors on human melanoma cells in culture. Proc. Null. Acad. Sci. U.S.A. 74, 565-569. Ford, S. R.. Aden, D. P., Mausner, R.. Trinchieri, G.. and Knowles. B. B. (1978). Partial characterization of cell-surface protein coded for by human chromosome 7. lmmunogenctics 6, 293-300. Forgue-Lafitte, M. E., Laburthe, M., Chemblier. M. C., Moody, A . J., and Rosselin, G. (1980). Demonstration of specific receptors for EGF-urogastrone in isolated rat intestinal epithelial cells. FEBS Lett. 114, 243-246. Fox, C. F., and Das, M. (1979). Internalization and processing of the EGF-receptor in the induction of DNA synthesis in cultured fibroblasts: The endocytic activation hypothesis. J . Supramol. Srrircr. 10, 199-214. Frati. L.. Daniele, S . , Delogu, A., and Covelli, I . (1972). Selective binding of the epidermal growth factor and its specific effects on the epithelial cells of the cornea. ESP. Eve Res. 14, 135- 141. Gordon, P., Carpentier, J . , Cohen, S . , and Orci, L. (1978). Epidermal growth factor: Morphological demonstration of binding, internalization, and lysosomal association in human fibroblasts. Proc. Narl. Acad. Sci. U.S.A 75, 5025-5029. Gospodarowicz, D., Mescher, A . L., Brown, K. D., and Birdwell, C. R. (1977). The role of fibroblast growth factor and epidermal growth factor in the proliferative response of the corneal and lens epithelium. Exp. Eye Res. 25, 631-649. Gospodarowicz, D., Brown, K. D., Birdwell. C. B.. and a l t e r , B. R . (1978). Control of proliferation of human vascular endothelial cells. Characterization of the response of human umbilical fibroblast growth factor, epidermal growth factor and thrombin. J . Cell Biol. 77, 774-788. Gregory, H. (1975). Isolation and structure of urogastrone and its relationship to epidermal growth factor. Nature (London) 257, 325-327. Gregory. H., and Preston. B. M. (1977). The primary structure of human urogastrone. Int. J . Peptide Prorein Res. 9, 107-1 18. Gregory, H., and Willshire, I. R. (1975). The isolation of urogastrones-Inhibitors of gastric acid secretion from human urine. Hoppe-Seylers Z . Physiol. Chem. 356, 1765- 1774. Guinivan. P., and Lddda. R . L. (1979). Decrease in epidermal growth factor receptor levels and production of material enhancing epidermal growth factor binding accompany the temperaturedependent changes from normal to transformed phenotype. Proc. Natl. Acad. Sci. U.S.A. 76, 3377-338 I . Haigler, H., and Carpenter, C. (1980). Production and characterization of antibody blocking epidermal growth factor receptor interactions. Biochirn. Biophys. Acta 598, 3 14-325. Haigler. H.. Ash. J . F., Singer, S. J . , and Cohen. S . (1978). Visualization by fluorescence ofthe binding and internalization of epidermal growth factor in human carcinoma cells A-431. Proc. Nail. Acad. Sci. U.S.A. 75, 3317-3321. Haigler. H. T., McKanna. J . A., Cohen. S . (1979). Direct visualization of the binding and internalization of a ferritin conjugate of epidermal growth factor in human carcinoma cells A-431. J . Cell Biol. 81, 382-395. Hassel, J. R. (1975). The development of rat palatal shelves in virro: An ultrastructural analysis of the inhibition of epithelial cell death and palate fusion by the epidermal growth factor. DPY. B i d . 45, 90-102. Hassel, J . R., and Pratt, R. M . (1977). Elevated levels of CAMPalters the effect of epidermal growth factor in vitro on programmed cell death in the secondary palatal epithelium. Dcv. B i d . 106, 55-62. Hayaishi, 0.. and Ueda, K. (1977). Poly (ADP-ribose) and ADP-ribosylation of protein. Am. Rev. Biochem. 46, 95- I 16.


403

Hock. R . A . . Nexo, E., Hollenberg. M. D. (1979). Isolation of the human placenta receptor for epidermal growth factor-urogastrone. Nurura (London) 277, 403-405. Hock. R. A., Nexo, E.. and Hollenberg, M. D. (1980). Solubilization and isolation of the human placenta receptor for epidermal growth factor-urogastrone. J . B i d . Chem. 255, 10737- 10743, Holladdy, L. A., Savage, C. R.. Cohen, S . , and Puett. D. (1976). Conformation and unfolding thermodynamics of epidermal growth factor and derivatives. Biochemistry 15, 2624-2633. Hollenberg, M. D. (1975). Receptors for insulin and epidermal growth factor: Relation to synthesis of DNA in cultured rabbit lens epithelium. Arch. Biochem. Biophw. 171, 371-377. Hollenberg, M. D. ( 1979). Epidermal growth factor-urogastrone, a polypeptide acquiring hormonal status. Vitcrm. Horm. 37, 69- 110. Hollenberg. M. D., and Cuatrecasas, P. (1973). Epidermal growth factor: Receptors in human fibroblasts and modulation of action by cholera toxin. Proc. Nut/. Acad. Sci. U.S.A. 70, 2964-2968, Hollenberg. M. D., and Cuatrecasas. P. (1975). Insulin and epidermal growth factor: Human fibroblast receptors related to DNA synthesis and amino acid uptake. J . Biol. Chem. 250, 3845-3853. Holley. R. W., Armour, R., Baldwin, J . A,. Brown, K. D.. and Yeh, Y. (1977). Density-dependent regulation of growth of BSC-I cells in cell culture: Control of growth by serum factors. Proc. Natl. Acad. Sci. U.S.A. 74, 5046-5050. Hunter, T. (1980). Proteins phosphorylated by the RSV transforming function. Cell 22, 647-648. Jazwinski, S. M . , Wang. J . L.. and Edelman. G . M. (1976).Initiation of replication inchromosomal DNA induced by extracts from proliferating cells. Proc.. Nut/. Acad. Sci. U.S.A. 73, 22312235. Kahn, C. R., Baird. K.. Flier, J. S., and Jarrett, D. B. (1977). Effect of auto antibodies to the insulin receptor on isolated adipocytes: Studies on insulin binding and insulin action. J . Clin. Invest. 60, 1094-1 106. King. A. C., Hernaez, L. I . , and Cuatrecasas. P. (1980a). Lysosomotropic amines cause intracellular accumulation of receptors for epidermal growth factor. Pror. N d . Acad. Sri. U.S.A. 77, 3238-3287. King. L. E., Carpenter. G., and Cohen, S. (1980b). Characterization by electrophoresis of epidermal growth factor stimulated phosphorylation using A-43 I membranes. Biochemistrji 19, I 524- 1528, Ladda. R. L.. Bullock, L. P., Glanopoulas. T., and McCormick, L. (1979). Radioreceptor assay for epidermal growth factor. Anal. Biochem. 93, 286-294. Lee, L. S . , and Weinstein. I . B. (1978). Tumor-promoting phorbol esters inhibit binding of epidermal growth factor to cellular receptors. Science 202, 313-315. Lee, L. S., and Weinstein, 1. B. (1979). Mechanism of tumor promotor inhibition of cellular binciing of epidermal growth factor. Proc. Nut/. Acad. Sci. U.S.A. 76, 5168-5172. Linsley. P. S . , and Fox, C. F. (1980). Controlled proteolysis of EGF receptors: Evidence for transmembrane distribution of the EGF binding site and the phosphate acceptor site. 1.Suprumol. Struct. 14, 461 -471. Linsley, P. S.. Blifeld, C. B., Wrann, M., and Fox, C. F. (1979). Direct linkage of epidermal growth factor to its receptor. Narure (London) 278, 745-748. Marquardt, H., and Todaro, G. H. (1982). Human transforming growth factors: Production by a melanoma cell line, purification and initial characterization. J . Biol. Chem. 257, 5220-5225. Michael. H., Bishayee, S . , and Das, M. (1980). Effect of methylamine on internalization, processing, and biological activation of epidermal growth factor receptor. FEBS Lett. 117, 125131. Nexo. E.. Hollenberg. M. D.. Figueroa, A,. and Pratt. R . M. (1980). Detection of epidermal growth

404

MANJUSRI DAS

factor-urogastrone and its receptor during fetal mouse development. Proe. Nut/. Arud. Sci. U.S.A. 77, 2782-2785. Nilsen-Hamilton, M., Shapiro, J . M., Massoglia, S. L., and Hamilton, R. T. (1980). Selective stimulation by mitogens of incorporation of Yhnethionine into a family of proteins released into the medium by 3T3 cells. Cell 20, 19-28. O’Keefe, E. J . , Hollenberg, M. D.. and Cuatrecasas. P. (1974). Epidermal growth factor: Characteristics of specific binding in membranes from liver, placenta and other target tissues. Arch. Biochem. Biophys. 164, 5 18-526. Pastan, I . , and Willingham. M. C. (1981). Journey to the center of the cell: Role of the receptoromc. Science 214, 504-509. Pratt, R . M . , and Pastan, I . (1978). Decreased binding of epidermal growth factor to BALBic 3T3 mutant cells defective in glycoprotein synthesis. Nature (London) 272, 68-70. Pruss. R. M., and Herschmann, H. R. (1977). Variants of 3T3 cells lacking mitogenic response to epidermal growth factor. Prac. Natl. Acad. Sci. U.S.A. 74, 3918-3921. Racker, E. (1976). Why do tumor cells have a high aerobic glycolysis. J . Cell. Physictl. 89, 697-700. Roberts, A. B., Lamb, L. C., Newton, D. L.. Sporn. M. B., DeLarco, J . E., and Todaro, G. J . ( 1980). Transforming growth factors: Isolation of polypeptides from virally and chemically transformed cells by acidiethanol extraction. Proc. Natl. Acad. Sci. U.S.A. 77, 349443498, Roberts, A. B . , Anzane, M. A.. Lamb, L. C., Smith, J. M., Frolik. C. A,, Marguardt, H., Todaro. G. J . , and Sporn, M. B. (1982). Isolation from murinc sarcoma cells of novel transforming growth factor potentiated by EGG. Nature (London 295, 417-419. Rose. S. P., Pruss, R. M., and Herschman, H. R. (1976). Initiation of 3T3 fibroblast cell division by epidermal growth factor. J . Cell. Phvsid. 86, 593-598. Rossow. P. W . , Riddle, V. 0. H., and Pardee, A. B. (1979). Synthesis of labile, serum-dependent protein in early G I controls animal cell growth. Proc. Natl. Acad. Sci. U.S.A.77, 3494-3498. Rozengurt, E., and Heppel, L. A. (1975). Serum rapidly stimulates ouabain-sensitive XhRbf influx in quiescent 3T3 cells. Proc. Nail. Acad. Sci. U.S.A. 72, 4492-4495, Rozengurt. E., Brown. K. D., and Pettican, P. (1981a). Vasopressin inhibition of epidermal growth factor binding to cultured mouse cells. J . B i d . Chem. 256, 716-722. Rozengurt, E., Gelehrter, T. D., Legg, A . , and Pettican, P. (1981b). Mellitin stimulates Na+ entry. Na-K pump activity and DNA synthesis in quiescent cultures of mouse cells. Cell 23, 78 1-788. Sahyoun, N., Hock. R. A., and Hollenberg, M. D. (1978). Insulin and epidermal growth factor-urogastrome: Affinity crosslinking to specific binding sites in rat liver membranes. Proc. Nail. Acad. Sci. U.S.A. 75, 1675-1679. Savage, C. R., Inaganii, J.. and Cohen, S . (1972).The primary structure of epidermal growth factor. J . B i d . Chem. 247, 7612-7621. Savion, N., Vlodavsky, I., and Gospodarowicz, D. (1980). Role of the degradation process in the mitogenic effect of epidermal growth factor. Proc. Nutl. Acad. Sri. U.S.A. 77, 1466-1470. Schechter, Y., Hernaez, L., Schlessinger, J . , and Cuatrecasas, P. (1979). Local aggregation of hormone-receptor complexes is required for activation by epidermal growth factor. Narure (London) 278, 835-838. Schlessinger, J., Shechter, Y., Willingham, M. C., and Pastan, I. (1978). Direct visualization of binding, aggregation and internalization of insulin and epidermal growth factor on living fibroblastic cells. Proc. Nail. Acad. Sci. U.S.A. 75, 2659-2663. Schneider, J. A., Diamond, I . , and Rozengurt, E. (1978). Glycolysis in quiescent cultures of 3T3 cells. Addition of serum, epidermal growth factor, and insulin increases the activity of phosphofructokinase in a protein synthesis independent manner. J . B i d . Chem. 253, 872-877. Schneiderman, M. H., Dewey, W. C., and Highfield (1971). Inhibition of DNA synthesis in synchronized Chinese hamster cells treated in G , with cycloheximide. Exp. Cell. Res. 67, 147- 155.


405

Schramm, M. (1979). Transfer of glucagon receptor from liver membranes to a foreign adenylate cyclase by a membrane fusion procedure. Proc. Nail. Acad. Sci. U.S.A. 76, 1174-1178. Schramm, M., Orly, J . , Eimerl. S . . and Korner. M. (1977). Coupling of hormone receptors to adenylate cyclase of different cells by fusion. Nutitre (London) 268, 310-313. Schreiber, A. B., Lax, I . , Yarden, Y . , Eslhar. Z . , and Schlessinger, J . (1981a). Monoclonal antibodies against receptor for epidermal growth factor induce early and delayed effects of epidermal growth factor. Proc. Narl. Acad. Sci. U.S.A. 78, 7535-7539. Schreiber, A. B., Yarden, Y.,and Schlessinger. J . (1981b). A nonmitogenic analog of epidermal growth factor enhances the phosphorylation of endogenous membrane proteins. Biochem. Biophvs. Res. Commun. 101, 517-523. Shimizu, N., Behzadian, M. A., and Shimizu. Y . (1980). Genetics of cell surface receptors for bioactive polypeptides: Binding of epidermal growth-factor is associated with the presence of human chromosome 7 in human-mouse cell hybrids. Proc. Nut/. Acad. Sci. U.S.A. 77, 3600-3604. Shimizu, Y., and Shimizu, N. (1981). Insulin and epidermal growth factor stimulate poly ADPribosylation. Biochem. Biophvs. Res. Commun. 99, 536-542. Shoyab. M., and Todaro, G. J. (1980). Vitamin K 3 (menadione) and related quinones, like tumorpromoting phorbol esters, alter the affinity of epidermal growth factor for its membrane receptors. J . Biol. Chem. 255, 8735-8739. Stiles. C. D., Capone, G . T., Scher, C. D., Antoniades, H. N., Van Wyk. J. J . , and Pledger, W. J. (1979). Dual control of cell growth by somatomedin and platelet-derived growth factor. Proc. Nail. Acad. Sci. U.S.A. 76, 1279- 1283. Stoker, M. G. P., Pigott, D., and Taylor-Papadimitriou. J. (1976). Response to epidermal growth factors of cultured human mammary epithelial cells from benign tumors. Nature (London) 264, 764-767. Sugimura, T. (1973). Poly(adenosine diphosphate ribose). Prog. Nucleic Acid Res. Mol. Biol. 13, 127-15 I . Temin, H. M. (1971). Stimulation by serum of multiplication of stationary chicken cells. J . Cell. Physiol. 78, 161-170. Thorn. D., Powell, A. 1.. Lloyd, C. W . , and Rees, D. A. (1977). Rapid isolation of plasma membranes in high yield from cultured fibroblasts. Biochem. J . 168, 187-194. Todaro, G. J . , Fryling, C., and DeLarco, J . E. (1980). Transforming growth factors produced by certain human tumor cells: Polypeptides that interact with epidermal growth factor receptors. Proc. Nut/. Acad. Sri. U.S.A. 77, 5258-5262. Todaro, G. J., Marquardt, G . , DeLarco, J. E., Fryling, C. M.. Reynolds, F. H., and Stephenson, J. R. (1981). Transforming growth factors produced by human tumor cells: Interaction with EGF membrane receptors. I n “Cellular Responses to Molecular Modulation” (L. W. Mozes, J. Schultz, W . A. Scott, and R. Werner, eds.). Academic Press, New York. Ushiro, H., and Cohen, S. (1980). Identification of phosphotyrosine as a product of epidermal growth factor activated protein kinase in A-431 cell membranes. J . Biol. Chem. 255, 8363-8365. Vlodavsky, I . , Brown, K. D., and Gospodarowciz, D. (1978). A comparison of the binding of epidermal growth factor to cultured granulosa and h e a l cells. J . Biol. Chem. 253, 3744-3750. Vogel, A., Raines. E., Kariya, B., Rivest, M. J., and Ross. R. (1978). Coordinate control of 3T3 cell proliferation by platelet-derived growth factor and plasma components. Proc. Natl. Acad. Sci. U.S.A. 75, 2810-2814. Westermark, B. ( 1977). Local starvation for epidermal growth factor cannot explain density-dependent inhibition of normal human glial cells. Proc. Nut/. A d . Sci. U.S.A. 74, 1619-1621. Wrann, M., and Fox, C. F. (1979). Identification of epidermal growth factor receptors in a hyperproducing human epidermoid carcinoma cell line. J . Biol. Chem. 254, 8083-8086.


CURRENT TOPICS IN MEMBRANES AND TRANSPORT. VOLUME I R

The Linkage between Ligand OccuDation and ResDonse of the NicoGnic Acety Ichol ine Recepto r PALMER TAYLOR, ROBERT DALE BROWN, AND DAVID A . JOHNSON Divisicm of Phurmucologv Department of Medicine Universiry of Culijorniu. Sun Diego La follu. Culijorniu

Introduction. . . .............. ................................ Structure of the Receptor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111. Biophysical Properties of the Receptor Channel, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IV. The Behavior of Partial Agonists, Antagonists. and Anesthetics in Relation to Channel Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V. Desensitization of the Receptor. ............................. VI. Ligand Occupation and Transitions in Receptor State . . . . . . . . . . . . . . . . . . . . . . . . . . VII. Other Ligands Affecting Receptor Function .................................. ... VIII. Analysis of Receptor Activation the IX . Toward the Understanding of Co Permeability Response . . . . . . . . .............................. X . Occupation and Activation by A ..................................... XI. Association of Antagonists with the Receptor and Functional Antagonism . . . . . . . . . ....... XII. Quantitation of Antagonist Occupation and Functional Antagonism XIII. Structural Implications and Arrangement of Subunits . . . . . . . . . . . . . . . . . . . . . . . . . . ............................. XIV . Analysis of the Bound Ligand States References ..................... .......................... 1.

11.

407 408 412 414 415 416 423 425 426 427 430 434 435 437 438

I. INTRODUCTION

Over a decade has passed since Changeux er al. ( 1970) employed the snake a-toxins to identify and monitor isolation of the nicotinic acetylcholine receptor from the electric eel, Electrophorus electricus. This successful initial application of a biochemical approach to characterizing neurotransmitter receptors relied 407

Copyrighl 0 1983 by A d c m i c Pre$a Inc All right5 of rcprodu~iionin m y form rcservcd ISBN 0 I2 153318-2

408

PALMER TAYLOR ET AL.

heavily on earlier observations of Lee and Chang (1963), who found that the elapid a-toxins exhibited an essentially irreversible block of neurotransmission and that antagonism could be localized to a postsynaptic site. Perhaps the most definitive corroborative evidence that the macromolecule purified by monitoring a-toxin binding was the acetylcholine receptor was provided by the finding that the purified receptor contained a 40,000-dalton subunit which could be labeled by treatment of intact cells with the irreversible antagonist p-[3HH]maleimidobenzyltrimethylammonium (MBTA) (Reiter et al., 1972; Karlin and Colburn, 1973). Thus, two labeling techniques which relied on different recognition characteristics of the receptor identified the same macromolecule or component of it. Since then, the biochemical and structural properties of the nicotinic acetylcholine receptor have been examined in considerable detail, its immunochemical properties have been assessed, and in recent studies rec-onstitutionof its functional properties has been demonstrated. Owing to the specific peptide a-toxins and an abundant source of receptor in the electric fish, the acetylcholine receptor is our most thoroughly characterized pharmacologic receptor. Moreover, the a toxins have also facilitated studies of the biosynthesis and turnover of the receptor as a specialized membrane protein, and progress is emerging on the synthesis of the receptor under in vitro conditions. Details on these topics may be found in a number of authoritative reviews (Changeux, 1981; Karlin, 1980; Adams, 1981; Steinbach, 1980). The primary emphasis of this article is devoted to a still-emerging area of study on this receptor, that of the relationship between receptor occupation by ligands and the physiologic or pharmacologic response. It is, of course, the linkage between occupation and response which confers the unique transduction capacities to receptor molecules as well as dictates their pharmacologic specificity for agonists, partial agonists, and antagonists. Since agonists can induce both an open channel and a desensitized state of the receptor and combinations of ligands can alter the character of the open channel state, analysis of the response itself is complex. It is evident that the receptor can exist in various states which differ in their affinity for ligands, conductance behavior, and conformation. Hence our analysis should include a comparison of ligand binding functions with functions of receptor state. With our knowledge of receptor structure and the understanding of its channel properties in biophysical terms, both developing to a relatively advanced stage, it is becoming possible to relate functional behavior to molecular details on structure. Thus, it would be appropriate at the outset to proceed with considerations of receptor structure.

II. STRUCTURE OF THE ISOLATED RECEPTOR Fortunately, in the past 2 years there has been a convergence among investigators regarding the interpretation of the biochemical data on the receptor mole-

LIGAND OCCUPATION AND RESPONSE OF THE CHOLINERGIC RECEPTOR

409

cule. The isolated receptor obtained by solubilization and subsequent purification by affinity chromatography contains four subunits of molecular weights 40,000 (a), 49,000 (p), 60,000 (y), and 67,000 (6) (Weill et al., 1974; Raftery et al., 1975). A similar composition of subunits is found for membrane fragments purified for enrichment in receptor molecules (Reed et a/., 1974). A 43,000-dalton peptide found in the membranes may be removed by alkaline treatment without loss of receptor function (Neubig et al., 1979). Although the 43,000-dalton peptide was thought originally by some to be an integral component of the receptor, this peptide or family of peptides appears to be localized at the cytoplasmic face of the postsynaptic membrane (Cartaud et al., 1981; St. John et al., 1982). Its removal results in a disordering of receptor molecules within the membrane. Thus, this peptide likely plays a role in the alignment and orientation of the receptor molecules in the subsynaptic areas. Based on N-terminal analyses, the subunits in the receptor from Torpedo are present in the stoichiometric ratio of az,p, y, 6 (Raftery et al., 1980). Preparations with gel profiles showing a predominance of the a subunit greatly exceeding the above stoichiometry have now been shown to arise from proteolysis of the p, y, and 6 subunits. The a2Py6 stoichiometry (Reynolds and Karlin, 1978; Lindstrom e t a / . , 1979) would nominally sum to a molecular weight of 250,000, which is close to the value obtained from careful hydrodynamic analysis of the monomeric form of the receptor (Reynolds and Karlin, 1978). Molecular weight estimates by a variety of methods have yielded values between 230,000 and 350,000 (cf. Changeux, 1981; Karlin, 1980). This variance reflects the inherent inaccuracy of the methods for estimation of molecular weight of membrane proteins, neglect of the contribution of dimers within the preparation, and an incomplete or unsatisfactory accounting of the bound detergent. Most, if not all, of the subunits found in the Torpedo preparations appear as corresponding subunits in Electrophorus and mammalian receptors (Lindstrom et al., 1980a; Shorr et al., 1981; Nathanson and Hall, 1979). However, achieving sufficient quantities of receptor devoid of proteolysis from mammalian muscle presents a far more formidable problem. The receptor in Torpedo appears largely as a dimeric (13 S) form in which disulfide association is found between the 6 subunits (Chang and Bock, 1977; Hamilton et a l . , 1977; Hucho et al., 1978; Weiland et al., 1979). Reduction results in cleavage of the disulfide bond with the formation of the 9 S monomer upon solubilization. Cleavage of this interchain disulfide, however, does not affect either the organization in the membrane or the functional properties of the receptor (Kistler and Stroud, 1981; Anholt et al., 1980). The tendency to form dimeric units does not appear to be shared by the mammalian or Electrophorus receptor. A functional role for only the a-chain has been clearly delineated. Initial studies showed that following disulfide bond reduction ['HIMBTA reacted with the a-subunit as did the alkylating agonist, bromoacetylcholine (Damle et al.,

41 0


1978). The predominant sites of a-toxin attachment following bifunctional crosslinking with a-toxin also appear on the a-subunit (Witzemann and Raftery, 1978; Nathanson and Hall, 1980). In the absence of reduction, the irreversible antagonist [ 3H]trimethylbenzenediazonium fluoroborate (TDF) reacts selectively with the a-subunit (Weiland et al., 1979). Thus, it appears that the a-subunit bears the agonist-antagonist recognition site as well as the major surface with which the -7000-dalton a-toxin peptide associates. The 6 subunit is the site of labeling of a covalent analog, 5-azido-trimethisoquin, of the local anesthetic trimethisoquin (Saitoh et al., 1980). The latter compound is a noncompetitive blocker of receptor function whose affinity for the receptor is actually enhanced by agonists. Chlorpromazine exhibits rather similar binding behavior to trimethisoquin, yet, following photoactivation, it reacts with all four subunits (Oswald and Changeux, 1981). Sequence work has been initiated on the individual subunits, and of particular interest is the degree of homology which exists among the peptides. Based on Nterminal sequences 30-50% homology exists in the N-terminal54 residues of the four peptide chains, suggesting their divergences from a common ancestral gene (Raftery et al., 1980). Probable homologous regions have also been detected by monoclonal antibody cross-reactivity (cf. Gullick et ul., 1981). Specific activities of the solubilized and purified receptor preparations in various laboratories have ranged between 5 and I0 pmole/g of protein (cf. Karlin, 1980; Changeux, 198I). The more homogeneous membrane preparations have yielded specific activities of 5040% of this range. Assuming an a*,p, y, 6 stoichiometry, a molecular weight of 250,000, and one a-toxin site per asubunit, we would not expect specific activities to exceed 8.0 pmole/g of protein unless portions of the receptor were lost through proteolysis or additional sites were detected by a-toxin binding. Since a-toxin binding prevents agonist and antagonist association and the latter ligands inhibit the initial rate of a-toxin binding in a competitive fashion, stoichiometry of association of the smaller reversible ligands can be related to a-toxin sites, thus obviating the variance inherent in quantitating protein content and assessing complete homogeneity of the preparation. Weber and Changeux (1974) initially found 1: 1 stoichiometry between acetylcholine and a-toxin binding, and this stoichiometry has been confirmed for a large number of ligands in membrane-associated and solubilized receptor preparations; a spin-labeled analog of decamethonium by electron spin resonance (Weiland et a[., 1976), a dansylacylcholine by fluorescence (Heidmann and Changeux, 1979), and radiolabeled acetylcholine and d-tubocurarine by equilibrium dialysis (Neubig and Cohen, 1979). Other investigators have found stoichiometric ratios of 0.5 for acetylcholine to a-toxin and carbamylcholine to a-toxin; however, in some cases these estimates have recently been revised (cf. Raftery et (11.. 1975; Schimerlik et al., 1979; Dunn et al., 1980). For irreversible antagonists, ratios of 0.5 equivalent of conjugated MBTA


41 1

and bromoacetylcholine per a-toxin site have been found following disulfide bond reduction (Damle and Karlin, 1978). The relationship of a-toxin occupation to the exclusion of [3H]MBTA labeling would suggest an arrangement of nonequivalent sites confined to the same receptor oligomer (Damle and Karlin, 1978). By altering the conditions of bromoacetylcholine labeling, Wolosin ef a f . , ( 1980) were able to achieve labeling of one bromoacetylcholine molecule per atoxin site in both Torpedo and the mammalian receptor. Despite 1: 1 stoichiometry for the alkylating acetylcholine analog the rate of the reaction was not uniform, suggesting that the sites of labeling do not show complete equivalence. TDF conjugation proceeds to 1: 1 stoichiometry (Weiland et a/., 1979). However, this ligand reacts with different residues; disulfide bond reduction is not required prior to TDF labeling. The difference in behavior of a-subunits which is revealed in an apparent chemical inequivalence of certan reactive groups will become important in future considerations. There is no evidence to date that the a-subunits differ in their primary structure. The existence of membrane patches containing high densities of receptor ( I 0 4 / ~ m 2has ) also facilitated ultrastructural characterization of the macromolecule. Electron microscopy following staining or freeze fracture and etching reveals that the receptor traverses the membrane bilayer and extends some 5.0-5.5 nm on the extracellular side. Its overall shape is roughly cylindrical with a long axis of 1 I .O nm and an average diameter of 8.0 nm (Cartaud eta/., 1978; Heuser and Salpeter, 1979; Allen and Potter, 1977; Klymkowski and Stroud, 1979). X-Ray diffraction (Kistler and Stroud, 1981 ) and neutron scattering (Wise et al., 1981a) have yielded further refinements in structure where the molecule is found to be funnel shaped, the widest diameter and largest protrusion (5.5 nm) appearing on the extracellular face. The extension through the cytoplasmic face is only 1 .O- 1.5 nm. In addition, the structure that encircles the central pit does not show three- or sixfold symmetry with respect to an axis perpendicular to the membrane; rather, three nonequivalent maxima can be identified (Kistler and Stroud, 1981). The simplest interpretation is that the three maxima constitute the larger of the subunits, p, y, and 6, with minima residing with the a-subunits, but further data are required before the subunits can be definitively placed. Image reconstruction and electron microscopy of receptor, biotinyl toxin-avidin complexes indicate that the two a-subunits are separated by 100" and hence are not contiguous (Holtzman et al., 1982; Kistler et al. 1982). Moreover, an angle of 45-85' separates the disulfide link in the &-subunitfrom one of the a-subunits. In face view subunit arrangements of 6aPay or 6ayap are most consistent with these data (Wise et a f . , 1981b). Selective labeling at a single membrane surface shows that all of the subunits span the membrane bilayer (Wennogle and Changeux, 1980; Strader and Raftery, 1980) and, hence, as transmembrane peptides, each subunit has both an intracellular and extracellular exposure. Based on the ultrastructure of the receptor and its biochemical composition, a

41 2


provisional model consistent with the large body of data can be developed (Fig. 1) (see also Karlin, 1980; Changeux, 1981). With the individual receptor subunits of a pentameric molecule traversing the membrane, it should not contain an axis of symmetry perpendicular to the membrane bilayer. Yet sequence homology between subunits may be sufficient to dictate the intersubunit associations within a pentameric structure. If we use the structural details of bacteriorhodopsin and cytochrome oxidase as model structures of channel-containing membrane proteins (cf. Stoeckenius and Bogomolni, 1982), the functional channel is not likely to be harbored within a single subunit; rather the perimeter of the channel would be formed by about seven a-helical segments contributed by most or all of the subunits. In this arrangement the two a-subunits will not have equivalent intersubunit contacts and hence we might anticipate that they will not exhibit precise functional equivalence. A number of examples of dimeric enzymes possessing identical sequences but lacking a twofold axis of symmetry are known (cf. Degani and Degani, 1980). In the case of the receptor, the inequivalence in activity of a-subunits is imposed simply by the necessity of including three additional subunits in the pentameric structure.

111.

BIOPHYSICAL PROPERTIES OF THE RECEPTOR CHANNEL

Although biophysical methods have yet to be applied to the rapid detection of ligand occupation on individual endplates or functional cells, techniques for measuring potential or conductance changes associated with transmitter activation have been available for decades. A more recent innovation has been the use of transient methods such as single channel and noise analyses to detect individual steps in the activation process. At low agonist concentrations and under voltage clamp the number of open

I I

I

B

x I6

I

49.000 60,000 67,000

FIG. 1. A plausible structural organization of subunits, a, @, y. 6, would allow for the nonequivalence of both agonist and antagonist binding in each oligomer, the positive cooperativity, and presumed concerted interaction required for receptor activation and desensitization. See text for details. The ligand binding sites are found on the a-subunits and all five subunits form the perimeter of the ion channel which is interior to the subunits. Each of the subunits has an intracellular and extracellular exposure.

LIGAND OCCUPATIONAND RESPONSE OF THE CHOLINERGIC RECEPTOR

413

channels varies as the square of the agonist concentration, and typically halfmaximal responses are obtained at carbamylcholine concentrations from 100 to 400 p l 4 and acetylcholine between 5 and 30 pkl (Adams, 1976; Sheridan and Lester, 1977; Dionne et al., 1978). A simple scheme commonly employed to explain activation data is kl

A f RR

A

k-,

-

vl kl P ARR ARRA-AR*R*A (closed) Z P - , (closed) u (open)

(1)

Thus, the channel exhibits a high probability of opening only when the receptor is in a doubly liganded state. At high agonist concentrations the emergence of desensitization and the loss of linearity of the response upon saturation limit the kinetic analyses that can be accomplished by voltage clamp measurements. The monitoring of individual transients was first carried out by Katz and Miledi (1972) who noted that statistical fluctuations in potential could be interpreted in terms of an average channel conductance and a lifetime of the open channel. Hence information may be obtained on the isomerization step in the above scheme:

P’

closed states

AR*R*A

a

The value of a appeared dependent on the agonist; for example a = 1 msec for acetylcholine whereas for suberyldicholine, a = 10 msec (Colquhoun, 1979). Also, a is much more sensitive to potential than is p’, decreasing approximately e-fold with a negative 70-80 mV change. While a is independent of agonist concentration, an examination of p’ over a range of agonist concentrations shows that it approaches a limiting value (Sakmann and Adams, 1978). Thus it appears possible to select low and high agonist concentrations where the bimolecular step ( k , ) and the isomerization (p) each becomes rate limiting [see Eq. ( l) ]. A comparison of channel opening rates for covalently bound (i.e., tethered) and reversibly associating agonists also indicates that at high agonist concentrations channel opening can become rate limiting (Lester et af., 1980). The development of patch electrodes which cover small areas of the muscle surface and confer large seal resistances has enabled measurements of conductance changes associated with the opening and closing of a single channel to be monitored rather than fluctuations occurring from a family of channels (Neher and Sakmann, 1976). The current pulses have a square wave shape of constant amplitude; their durations are exponentially distributed around defined lifetimes. Both the single channel conductance and the time constants showed good agreement with data derived from fluctuation analysis. Single channel measurements have allowed the determination of both p and a with a higher level of precision. The opening rate constant p is greater than a; thus at saturating agonist con-

414


centrations the open channel state would be a dominant species (Dionne et a l . , 1978; Dionne and Liebowitz, 1982). Single channel opening events may also contain closed intervals of very short duration and are detectable when electrode capacitance is low (Colquhoun and Hawkes, 1981; Colquhoun and Sakmann, 1981). More recently, the stochastic properties of intervals of successive opening and closing events have been analyzed to estimate channel opening p and ligand dissociation rates k . The probability analysis shows a 500 sec - I , p 750 sec-I, and k - I to be -4700 sec-' for acetylcholine (Dionne and Liebowitz, 1982). From these data and an equilibrium constant, k , is estimated to be 1.5 X lo8M - sec- I , a value in close accord with estimates from the reaction kinetics (cf. Section VI). Conductances for agonist activation are approximately 25 pS and would be equivalent to an ion flux rate of lo4 ions/msec. Since this value exceeds the capacity of a rotating carrier system, current appears to be carried by a channel which allows cations to move in accord with their concentration gradients (cf. Karlin, 1980). Reversal potential measurements reveal that cation selectivity depends primarily on ionic radius; cations with diameters up to 0.65 nm permeate freely (Dwyer et al., 1980). ~

,

-

-

IV. THE BEHAVIOR OF PARTIAL AGONISTS, ANTAGONISTS, AND ANESTHETICS IN RELATION TO CHANNEL ACTIVATION Compared to full agonists, partial agonists show saturation of their doseresponse relationships when a considerably smaller fraction of channels are open. Hence their actions might be explained simply on the basis of a small p/a ratio in Eq. (1). However, such compounds may also exhibit less than maximal responses by a secondary channel blocking action (Adams and Sakmann, 1978). Classic antagonists such as d-tubocurarine, gallamine, and pancuronium show competitive antagonism with agonists which is typically reflected as a parallel shift of the dose-response curve (Jenkinson, 1960; Colquhoun et al., 1979). Consistent with this competitive antagonism is the finding that antagonists decrease the frequency of channel opening events elicited b y an agonist without influencing their conductance or duration of opening. However, analyses of functional antagonism over a wide range of concentrations reveal a noncompetitive component of antagonism. The noncompetitive antagonism may reflect direct channel occlusion; it appears to be far more sensitive to membrane potential than competitive antagonism, becoming manifest at larger negative potential (Colquhoun et ul., 1979). The noncompetitive antagonism is also more evident with d-tubocurarine antagonism than with gallamine or pancuronium (Colquhoun and Sheridan, 1981; Katz and Miledi, 1978). Many local anesthetics or related compounds which noncompetitively block


415

steady-state responses directly modify channel conductance properties. This was originally detected by comparing miniature endplate currents following quanta1 release of acetylcholine (Steinbach, 1968). The decay of endplate currents shows a time course virtually identical to the time constant for the average lifetime of single channels opened by acetylcholine. Thus decay rates of miniature endplate currents also give a measure of a.In the presence of certain local anesthetics the normally exponential decay reveals at least two components. The dependences on agonist concentration and membrane potential suggest that the local anesthetic blocks the functional channel primarily while in an open configuration. Although it is an attractive possibility, these observations do not establish that the anesthetic directly enters and physically occludes the channel. Rather, the local anesthetic may simply favor binding to the open channel state of the receptor. Corroborative evidence favoring block of open channels comes from the analysis of single channels where local anesthetics serve to chop the square wave pulses into bursts of pulses of shorter duration (Neher and Steinbach, 1978). Such flickering can be interpreted in terms of repetitive blocking and unblocking of open channels. The anesthetics are a structurally disparate series of compounds, and it would indeed be surprising if a class of compounds with such diverse effects on membrane structure had a unitary mechanism of noncompetitive antagonism. Anesthetics such as QX222, procaine, and tetracaine may act primarily by noncompetitive block of conduction through association with the open pore state of the receptor. The less polar anesthetics such as dibucaine, chlorpromazine, and benzocaine do not exhibit such a marked voltage dependence in their noncompetitive block and hence may affect receptor function by altering annular lipid or associating at a hydrophobic site on the lipid-protein interface (Koblin and Lester, 1979). These agents enhance the rate and extent of desensitization resulting from exposure to agonists (Sine and Taylor, 1982). An inhibitory action which is related but not identical to the local anesthetics is observed for the noncompetitive antagonist histrionicotoxin and its hydrogenated analogs (Albuquerque et al., 1973; Burgermeister el al., 1977).

V.

DESENSITIZATION OF THE RECEPTOR

Upon continuous application of an agonist to the receptor the conductance response, rather than remaining constant, slowly decreases and this loss of responsivity resulting from prior exposure to agonist is termed desensitization. Kinetics of desensitization were initially examined by Katz and Thesleff ( 1957) who found that the onset is dependent on agonist concentration (up to a limiting concentration) and recovery is a slow unimolecular process that was independent of the agonist and its concentration used to promote desensitization. These workers suggested that a cyclic, two-state model best described the kinetics for onset

416


and recovery of desensitization: KR

AR

A+RM

A

1 1

+ R’+

M = R’/R

(3)

AR’ KR

In the absence of agonist the receptor is primarily in the R state (M < 1); the slow conversion to AR’ simply reflects the higher affinity of the R’ state for the agonist A (i.e., the dissociation constant K,.

Membrane Receptors (Current Topics in Membranes and Transport ; V. 18)

CURR TOPICS IN MEMBRANES & TRANSPORT V15, Volume 15 (Current Topics in Membranes and Transport)

Current Topics in Membranes & Transport Vol. 25: Regulation of Calcium Transport Across Muscle Membranes (Current Topics in Membranes and Transport) (v. 25)

Current Topics in Membranes and Transport, Volume 05

Current Topics in Membranes and Transport, Volume 10

Current Topics in Membranes and Transport, Volume 9

Current Topics in Membranes & Transport Vol. 23: Genes & Proteins: Transport Proteins & Receptors

Current Topics in Membranes and Transport, Volume 3

Current Topics in Membranes and Transport, Volume 20

Current Topics in Membranes and Transport, Volume 04

Current Topics in Membranes and Transport, Volume 07

Current Topics in Membranes and Transport, Volume 10

Current Topics in Membranes and Transport, Volume 07

Membrane Protein-Cytoskeleton Interactions, Volume 43 (Current Topics in Membranes)

The Nociceptive Membrane, Volume 57 (Current Topics in Membranes)

Current Topics in Membranes and Transport, Volume 04

Current Topics in Membranes and Transport, Volume 14

Current Topics in Membranes and Transport, Volume 08

Current Topics in Membranes and Transport, Volume 02

Current Topics in Membranes and Transport, Volume 06

Potassium Transport: Physiology and Pathophysiology (Current Topics in Membranes)

Current Topics in Membranes and Transport, Volume 9

Current Topics in Membranes and Transport, Volume 36: Protein-Membrane Interactions

Cell Biology and Membrane Transport Processes, Volume 41 (Current Topics in Membranes)

Regulatory Mechanisms of Intracellular Membrane Transport (Topics in Current Genetics)

Current Topics in Membranes and Transport, Volume 32: Membrane Fusion in Fertilization, Cellular Transport, and Viral Infection

Molecular Neurobiology: Endocrine Approaches (Current Topics in Membranes and Transport, V. 31)

Aquaporins (Current Topics in Membranes, Volume 51)

Claudins (Current Topics in Membranes, Volume 65)

Extracellular Nucleotides and Nucleosides: Release, Receptors, and Physiological & Pathophysiological Effects (Current Topics in Membranes)

Membrane Receptors (Current Topics in Membranes and Transport ; V. 18)

CURR TOPICS IN MEMBRANES & TRANSPORT V15, Volume 15 (Current Topics in Membranes and Transport)

Current Topics in Membranes & Transport Vol. 25: Regulation of Calcium Transport Across Muscle Membranes (Current Topics in Membranes and Transport) (v. 25)

Current Topics in Membranes and Transport, Volume 05

Current Topics in Membranes and Transport, Volume 10

Current Topics in Membranes and Transport, Volume 9

Current Topics in Membranes & Transport Vol. 23: Genes & Proteins: Transport Proteins & Receptors

Current Topics in Membranes and Transport, Volume 3

Current Topics in Membranes and Transport, Volume 20

Current Topics in Membranes and Transport, Volume 04

Current Topics in Membranes and Transport, Volume 07

Current Topics in Membranes and Transport, Volume 10

Current Topics in Membranes and Transport, Volume 07

Membrane Protein-Cytoskeleton Interactions, Volume 43 (Current Topics in Membranes)

The Nociceptive Membrane, Volume 57 (Current Topics in Membranes)

Current Topics in Membranes and Transport, Volume 04

Current Topics in Membranes and Transport, Volume 14

Current Topics in Membranes and Transport, Volume 08

Current Topics in Membranes and Transport, Volume 02

Current Topics in Membranes and Transport, Volume 06

Potassium Transport: Physiology and Pathophysiology (Current Topics in Membranes)

Current Topics in Membranes and Transport, Volume 9

Current Topics in Membranes and Transport, Volume 36: Protein-Membrane Interactions

Cell Biology and Membrane Transport Processes, Volume 41 (Current Topics in Membranes)

Regulatory Mechanisms of Intracellular Membrane Transport (Topics in Current Genetics)

Current Topics in Membranes and Transport, Volume 32: Membrane Fusion in Fertilization, Cellular Transport, and Viral Infection

Molecular Neurobiology: Endocrine Approaches (Current Topics in Membranes and Transport, V. 31)

Aquaporins (Current Topics in Membranes, Volume 51)

Claudins (Current Topics in Membranes, Volume 65)

Extracellular Nucleotides and Nucleosides: Release, Receptors, and Physiological & Pathophysiological Effects (Current Topics in Membranes)

Recommend Documents