Principles of Medical Biology Vol 4 Part II: Cell Chemistry and Physiology

Cell Chemistry and Physiology: Part II PRINCIPLES OF MEDICAL BIOLOGY A Multi-Volume Work, Volume 4 Editors: E. EDWARD ...

Author: Edward Bittar

9 downloads 496 Views 21MB 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

Cell Chemistry and Physiology: Part II

PRINCIPLES OF MEDICAL BIOLOGY A Multi-Volume Work, Volume 4 Editors: E. EDWARD BITTAR, Department of Physiology, University of Wisconsin, Madison NEVILLE BITTAR, Department of Medicine, University of Wisconsin, Madison

This Page Intentionally Left Blank

Principles of l\/ledical Biology

A Multi-Volume Work

Edited by E. Edward Bittar, Department of Physiology, University of Wisconsin, t\/ladison and Neville Bittar, Department of l\/ledicine University of Wisconsin, l\/ladison This work provides: * A holistic treatment of the main medical disciplines. The basic sciences including most of the achievements in cell and molecular biology have been blended with pathology and clinical medicine. Thus, a special feature is that departmental barriers have been overcome. * The subject matter covered in preclinical and clinical courses has been reduced by almost one-third without sacrificing any of the essentials of a sound medical education. This information base thus represents an integrated core curriculum. * The movement towards reform in medical teaching calls for the adoption of an integrated core curriculum involving small-group teaching and the recognition of the student as an active learner. * There are increasing indications that the traditional education system in which the teacher plays the role of expert and the student that of a passive learner is undergoing reform in many medical schools. The trend can only grow. * Medical biology as the new profession has the power to simplify the problem of reductionism. * Over 700 internationally acclaimed medical scientists, pathologists, clinical investigators, clinicians and bioethicists are participants in this undertaking.


Cell Chemistry and Physiology: Part II Edited by E. EDWARD BITTAR Department of Physiology University of Wisconsin Madison, Wisconsin NEVILLE BITTAR Department of Medicine University of Wisconsin Madison, Wisconsin

JAI PRESS INC. Greenwich, Connecticut

London, England

Library of Congress Cataloging-in-Publication Data Cell chemistry and physiology / edited by E. Edward Bittar, Neville Bittar. p. cm.—(Principles of medical biology ; v. 4) Includes index. ISBN 1-55938-805-6 1. Cytochemistry. 2. Cell physiology. I. Bittar, E. Edward. II. Bittar, Neville. III. Series. [DNLM: 1. Cells-Chemistry. 2. CeUs—physiology. QH 581.2 C392 1995] QH611.C4214 1995 574.87-^c20 for Library of Congress 94-37215 CIP

Copyright © 1996 by JAI PRESS INC. 55 Old Post Road, No. 2 Greenwich, Connecticut 06836 JAI PRESS LTD. The Courtyard 29 High Street Hampton Hill, Middlesex TW12 IPD England All rights reserved. No part of this publication may be reproduced, stored on a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, filming, recording or otherwise without prior permission in writing from the publisher. ISBN: 1-55938-806-4 Library of Congress Catalog Na: 95-33561 Manufactured in the United States ofAmerica

CONTENTS

List of Contributors

ix

Preface £ Edward Bittar and Neville Bittar

xl

Chapter 1 Cellular ATP David A. Harris

1

Chapter 2 Purines Charles H. V. Hoyle and Geoffrey Burnstock

49

Chapters The Role of Multiple Isozymes in the Regulation of Cyclic Nucleotide Synthesis and Degradation J. Kelley Bentley and Joseph A. Beavo

77

Chapter 4 The Biological Functions of Protein Phosphorylation-Dephosphorylation Terry A. Woodford, Stephen J. Taylor, and Jackie D. Corbin

123

Chapter 5 The Family of Protein Tyrosine Phosphatases and the Control of Cellular Signaling Responses Nicholas K. Tonks

179

Chapter 6 Cyclic Cascades in Cellular Regulation P. Boon Chock and Earl R, Stadtman

201

vii

viii

CONTENTS

Chapter 7 Mechanisms of Intracellular pH Regulation Greg Coss and Sergio Grinstein

221

Chapter 8 The Membrane Na'*'-K"^-ATPase in Cells Thomas A. Pressley

243

Chapter 9 Intracellular Calcium-Binding Proteins Kevin K.W. Wang

255

Chapter 10 ATP-Ubiquitin-Mediated Protein Degradation A.L Haas

275

Chapter 11 Regulation of Cellular Functions by Extracellular Calcium Edward F. Nemetii

285

Chapter 12 The Basis of Intracellular Calcium Homeostasis in Eukaryotic Cells Francesco Di Virgilio, Daniela Pietrobon, and Tullio Pozzan

305

Chapter 13 Roles of Polyamines in Cell Biology Nikolaus Seiler Chapter 14 Free Radicals in Cell Biology Peter A. Southorn and Garth Powis INDEX

329

349 379

LIST OF CONTRIBUTORS Joseph A. Beavo

Department of Pharmacology University of Washington Seattle, Washington

j. Kelley Bentley

Department of Pharmacology University of Washington Seattle, Washington

Geoffrey Burnstock

Department of Anatomy and Developmental Biology University College London London, England

P. Boon Chock

Laboratory of Biochemistry National Institutes of Health Bethesda, Maryland

Jackie D. Corbin

Howard Hughes Medical Institute Vanderbilt University Nashville, Tennessee

Francesco Di

Institute of General Pathology University of Ferrara Ferrara, Italy

Virgilio

Greg Goss

Division of Cell Biology Hospital for Sick Children Toronto, Ontario, Canada

Sergio Grinstein

Division of Cell Biology Hospital for Sick Children Toronto, Ontario, Canada XI

LIST OF CONTRIBUTORS A.L Haas

Department of Biochemistry Medical College of Wisconsin Milwaukee, Wisconsin

David A. Harris

Department of Biochemistry University of Oxford Oxford, England

Charles H.V. Hoyle

Department of Anatomy and Developmental Biology University College London London, England

Edward F. Nemeth

NPS Pharmaceuticals, Inc.

Daniela

Department of Biomedical Sciences University of Padova Padova, Italy

Pietrobon

Carth Powis

Department of Pharmacology Mayo Clinic and Foundation Rochester, Minnesota

Tullio Pozzan

Department of Biomedical Sciences University of Padova Padova, Italy

Thomas A. Pressley

Department of Physiology and Cell Biology The University of Texas Health Science Center Houston, Texas

Nikolaus Seiler

Institut de Recherche Centre de Cancer Universite des Rennes Rennes Cedex, France

Peter A, Southorn

Department of Pharmacology Mayo Clinic and Foundation Rochester, Minnesota

List of Contributors

XI

Earl R. Stadtman

Laboratory of Biochemistry National Institutes of Health Bethesda, Maryland

Stephen J. Taylor


Nicholas K, Tonks

Cold Spring Harbor Laboratory Cold Spring Harbor, New York

Kevin K,W. Wang

Department oi Pharmacology Warner-Lambert Company Ann Arbor, Michigan

Terry A.


Woodford


PREFACE

This volume illustrates the extent to which the traditional distinction between biochemical and physiological processes is being obliterated by molecular biology. It can hardly be doubted that the revolution in cell and molecular biology is leading to core knowledge that provides an outline of the integrative and reductionist approach. We view this as the beginning of a new era, that of the integration of learning. As in the preceding volumes, the choice of topics has been deliberate not only because of the need to keep the volume within reasonable bounds but also because of the need to avoid information over-load. Several relevant topics are dealt with in other modules; for example, the role of G proteins in transmembrane signaling is covered in the Membranes and Cell Signaling module (i.e.. Volume 7). Omissions are of course inevitable but they are minor. A case in point is the subject of phosphatases, the treatment of which does not take into account calcineurin. One of the key functions of this Ca^'^-activated protein phosphatase that is also regulated by calmodulin is to dephosphorylate voltage-dependent Ca^"^ channels. The mere recognition of such omissions before or after consulting textbooks and journals should be a spur to a more complete discussion by the student of the subject in a small group teaching setting. We should like to thank the many authors for their scholarly contributions and enthusiasm. We also take this opportunity of thanking Ms. L. Manjoney and the staff members of JAI Press for their skill and courtesy. E. EDWARD BITTAR NEVILLE BITTAR xiii


Chapter 1

Cellular ATP DAVID A. HARRIS

Introduction Structure of ATP Chemical Bonds and Conformation What Makes ATP a Good Energy Source? Other Features ofthe ATP Molecule Measurement of Cellular ATP The Freeze-Clamp Technique The Magnetic Resonance Technique Adenine Nucleotide Concentrations Within Cells Spatial Distribution Uses of ATP Contraction of Actomyosin Ion Pumping ATP in Biosyntheses ATP as Phosphate Donor ATP as Charge Neutralizer ATP and Messenger Molecules Structural Role of ATP Reactions Involving Exchange of High Energy Phosphates Creatine Kinase Adenylate Kinase Synthesis of ATP Substrate Level Phosphorylation Oxidative Phosphorylation

Principles of Medical Biology, Volume 4 Cell Chemistry and Physiology: Part II, pages 1-47 Copyright © 1996 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-806-4

2 3 3 3 6 7 7 8 10 10 13 13 16 20 22 23 24 25 25 26 27 28 28 31

2

DAVID A. HARRIS

Control of ATP Levels ATP Levels Are Closely Maintained In Vivo Is ATP a Regulator? Control of Anaerobic ATP Production Controlof Aerobic ATP Production Pathological Disturbances of ATP Levels Malignant Hyperthermia (Malignant Hyperpyrexia) Ischemia Summary

35 35 35 36 40 43 43 43 44

INTRODUCTION In living cells, a variety of processes yield energy. In man, these are typically oxidations (of glucose, amino acids, and fatty acids), although some energy is also produced by the anaerobic breakdown of glucose to lactate (anaerobic glycolysis). The amount of energy yielded in the these processes is very variable—complete oxidation of palmitic acid yields 9,500 kJ/mol, while the conversion of glucose to lactate yields about 170 kJ/mol. Conversely, a variety of processes in the living cell require energy. These include the biosynthesis of small molecules (e.g., glucose from pyruvate) and of large biopolymers (e.g., proteins from amino acids), the transport of molecules and ions, and the performance of mechanical work (e.g., muscle contraction). Because of variety both in the chemical nature of catabolic processes, and in their energy yield, these reactions cannot be used directly to drive the variety of energy-requiring processes. In essence, energy released in catabolic processes is trapped in units of 50-60 kJ/mol, by synthesizing ATP from ADP+Pi, and used in these units in biosyntheses, ion pumping, mechanical processes, etc. Enzymes involved in the latter processes are thus adapted to accept ATP as a convenient and common unit of exchange between themselves and the variety of energy-yielding processes; ATP is sometimes known as the energy currency of the cell. ATP is a short-term store of energy within the cell; the cell content of ATP turns over about once every second. The other short-term energy store in cells is the transmembrane ion gradient, in particular the Na"*" gradient across the plasma membrane and the H"^ gradient across the mitochondrial membrane. The amount of energy stored in these gradients (comprising contributionsfromboth concentration and voltage gradients) is about 15-20 kJ/mol, i.e., 15-20 kJ is released when one mol of ions moves downhill. Thus, energy in ion gradients is stored in smaller units than it is in ATP. However, this seems convenient for most transport processes in animal cells; the plasma membrane Na^ gradient can be used as an energy source for accumulation of glucose and amino acids from the blood. Thus, ion gradients can serve as an energy source in some biological processes. Compared to ATP, however, they are far less versatile in their application. As noted above, the unit of energy stored per ion in a gradient is 15-20 kJ/mol, about one third that per mol ATP. More importantly, this energy is not portable. A gradient can drive processes only at the membrane across which it is located; it cannot power

Cellular ATP

3

the bulk of chemical reactions in the cell, which occur in free solution. Thus, ATP is pre-eminent as a diffusible energy source for biochemical processes.

STRUCTURE OF ATP Chemical Bonds and Conformation The chemical structure of ATP is shown in Figure la. The molecule consists of three notional parts, the purine base adenine, the pentose sugar ribose, and a chain of three linked phosphate groups. The same adenine-ribose-phosphate structure is found in nucleic acids (RNA), and thus the molecule belongs to the class of nucleotides; other aspects of nomenclature are indicated on the figure. In solution, ATP can adopt a variety of conformations, in particular due to rotation about the base-sugar bond (a in Figure la) and to variations in orientation of the phosphate groups. A preferred conformation is the extended form (Figure lb), with the base and to the sugar ring and the phosphate groups extended; in this form it binds to many proteins. Introduction of a bulky group into the five membered ring of the purine (e.g., in 8-bromo ATP) tends to favor the syn conformer, which binds less well to proteins. Normal cytoplasmic conditions are around pH 7.1, with p[Mg^"^]totai * 2.3. This means that >90% of cytoplasmic ATP exists as the fully ionized MgATP^~ complex. All enzymes that use ATP utilize the MgATP^~ complex rather than free ATP, with the exception of the mitochondrial ATPADP exchanger, which utilizes the small amount of ATP"^ in equilibrium with the MgATP^" complex. This ensures that Mg^"*" levels inside mitochondria can be maintained independently of ATP synthesis rates. Under some physiological conditions (e.g., heavy exercise), intracellular pH may fall and MgATP^~+ H"^ MgATP(H)~ equilibrium may shift in favor of the protonated form. This may affect the availability of cellular ATP, although the magnitude of such effects are as yet unknown. What Makes ATP a Good Energy Source? In regard to energy transfer, the critical part of the ATP molecule is its phosphate tail, and, in particular, its two terminal phosphate groups (P and y phosphates). Each is linked to the neighboring phosphate by an acid anhydride link. (Note that phosphate is simply the ionized form of phosphoric acid.) Since acid anhydrides are (thermodynamically) unstable in water, they can serve as a source of energy. Quantitatively, we consider the hydrolysis ATP + H2O-^ ADP + Pj.* Conventionally, Mg^"^ ions and H"^ ions, which are buffered in the cell, are omitted from this equation. Thus, Pj indicates the prevailing ionization state of inorganic phosphate (HPO4") and ATP indicates the complex MgATP^".

PHOSPHATES

OH

OH

Figure 1, The structure of ATP. (a) Chemical structure of the MgATP complex, showing nomenclatures used. Rotation around bond (a) converts antl and syn conformers, (b) Conformation of ATP bound to an enzyme (aspartate transcarbomoylase). Note that (i) the planes of the adenine and ribose rings are at right angles; (ii) the adenine and ribose rings lie syn to each other; and (lii) the phosphate chain is extended.

Cellular ATP

5

The change in free energy (energy available for work) is given by AG = AG°' + RT ln[ADP][Pi]/[ATP]. AG°' is a term reflecting the chemical nature of the compound under consideration; for ATP hydrolysis it is around —30 kJ/mol, larger in magnitude than the value for phosphate esters (e.g., sugar phosphates) of-15 kJ/mol. This is due to the anhydride nature of the bond hydrolyzed, which allows increased resonance stabilization (electron derealization) and increased hydration in the products ADP + Pj, stabilizing them relative to the reactants ATP and H2O. In the cell, ADP levels are kept low such that the ATP/ADP ratio « 200 in the cytoplasm. This means that the actual free energy yielded on ATP hydrolysis in vivo (AG in the above equation) is larger in magnitude than AG°', due to the contribution from the second term in the equation. The free energy released on hydrolysis of intracellular ATP, often written as the phosphorylation potential AGp, is typically -55-60 kJ/mol ATP (Veech et al, 1979): AGp = -30 + RT In [Pi]/100 = -^0 kJ/mol at 37 °C and typical cellular free [Pj] = 1 mM.* ATP is one of a number of cellular phosphates with a highly negative free energy of hydrolysis. It is convenient to designate ATP and compounds with similar AG°' values for hydrolysis (e.g., GTP, UTP, etc.) or greater in magnitude (creatine phosphate, 1,3 diphosphoglycerate, phosphoenol pyruvate) as high energy phosphates; they can all, without fiirther energy input, generate ATP, which can then be used to drive cellular processes. The sum of concentrations of all these high energy phosphates is an indication of the energy status of a cell. The second essential feature of ATP as a temporary biological energy store is its kinetic stability. It is obvious that kinetic stability must be a feature of an energy store: there is no point in producing a high energy compound which hydrolyzes rapidly before it can be used. However, the concept of a thermodynamically unstable and kinetically stable compound might appear patadoxical. It can be understood by consideration of an analogous system, a mixture of hydrogen and oxygen. Such a mixture is kinetically stable: it could stand for thousands of years at room temperature with no noticeable change. However, given a catalyst, or a spark, it changes chemically with the release of large amounts of energy. Similarly, left to itself, ATP is stable in solution for several days, but, given a suitable (enzyme) catalyst, it hydrolyzes to yield large amounts of energy. The kinetic stability of ATP (as compared to, for example, acetic anhydride in water) is chemically due to the high (negative) charge density around the phosphate groups, which discourages the approach of nucleophiles.

6

DAVID A. HARRIS

Other Features of the ATP Molecule

As noted above, ATP is only one of a series of high energy phosphates found within cells (Table 1). It is, however, by far the most versatile, being formed in the bulk of energy-yielding reactions (mitochondrial oxidations) and used in most energy-requiring processes. Although occasionally other compounds may serve directly as an energy source (e.g., phosphoenol pyruvate drives some bacterial transport systems; GTP drives several steps in protein synthesis), such compounds are usually used to generate ATP. The favored role of ATP may be rationalized in several ways: 1. Since it contains rwo acid anhydride links, ATP may hydrolyze either to ADP + Pj or AMP + 2Pi. Thus, the occasional reaction requiring a driving energy of more than 50-60 kJ/mol can be driven by splitting both anhydride bonds. (This is not possible for phosphocreatine, 1,3 diphosphoglycerate, etc.) 2. The adenine ring plays no part in the chemistry or energetics of ATP function; adenosine triphosphate, however, is used much more widely than GTP, CTP, etc. This may be an accident of history, with adenine appearing, by chance, early in prebiotic evolution. However, it is interesting that adenine appears in the structure of a variety of other coenzymes (NAD,

Table 1. Standard Free Energy of Hydrolysis for Biochemical Compounds Compound

AC^' (kj/mol)

phosphoeno/pyruvate ATP (-^ AMP + 2Pi) 1,3 diphosphoglycerate phosphocreatine fatty acyl Coenzyme A (-> fatty acid + CoA) amino acyl tRNA (-• amino acid + tRNA) ATP GTP, UTP, CTP PPi

-61 -58 -49 -43 -35 -35 -31 -31 -28

glucose-6-phosphate AMP glycerol-1 -phosphate

-14 -9.6 -9.2

Note:

Except where indicated, the reaction considered is X - P + H2O - » X + Pi. Compounds above the line are designated high energy compounds in biochemistry (see text). Note that the actual free energy change for hydrolysis of these compounds in vivo (e.g., AGp for ATP hydrolysis) is normally greater than the change under standard conditions, AC°', given here.

Cellular ATP

7

FAD, coenzyme A) where again it plays no part in the reaction. Perhaps it provides a particularly favorable recognition site for enzymes. 3. ATP is an acid anhydride. A major requirement for energy in macromolecule biosynthesis is in driving condensation reactions (removal of water) in an aqueous environment. Formation of a peptide bond, for example, is a dehydration: R-COOH + NH2-R' ^ RCO-NH-R' + H2O. The anhydride nature of ATP allows it to be a good dehydrating agent even, given a suitable reaction mechanism (see below), in an aqueous environment. 4. ATP serves as a source of phosphate groups in biochemical reactions. For example, glucose, on entering the cell, is phosphorylated to glucose-6-phosphate, giving it a negative charge which helps to retain it within a compartment bounded by the (lipophilic) cell membrane. Many metabolic pathways (e.g., glycolysis, histidine biosynthesis) utilize phosphorylated intermediates in this way to limit diffusion out of the cell. A contrasting example is the phosphorylation, by ATP, of enzymes such as glycogen phosphorylase which are switched on (or off) by this process. In both these cases, the important feature of ATP is not its tendency to transfer phosphate to water (high negative free energy of hydrolysis) but its tendency to phosphorylate other hydroxyl groups (high phosphate transfer potential). The energetic role of ATP in these phosphorylation reactions is to ensure the reaction is driven to completion; the loss in free energy in generating a phosphate ester in place of an anhydride is dissipated as heat.

MEASUREMENT OF CELLULAR ATP The Freeze-Clamp Technique Classically, measurement of ATP levels within cells and tissues has involved (a) the rapid arrest of metabolism and of enzyme activity in the tissue; (b) extraction of ATP from the tissue (without destroying it); and (c) assay of its concentration by enzymatic procedures or by high performance liquid chromatography (HPLC). Since the energy status of a tissue is also dependent on ADP, AMP, and Pj concentrations, these are generally measured with ATP in a single extract. This technique is highly sensitive; using firefly luciferase (bioluminescent assay) the ATP content of only a few hundred cells can be measured. This is useful when biopsy material is being studied. In a typical procedure, the tissue is perfused with an oxygenated buffer/salt solution and manipulated (e.g., electrically stimulated, treated with a drug) as desired. The tissue is then rapidly frozen, by crushing it between two flat aluminum

8

DAVID A. HARRIS

plates at liquid nitrogen temperatures, to arrest metabolism. The frozen, powdered tissue is deproteinized with perchloric acid (to remove enzymes) and the soluble extract (containing the tissue metabolites) neutralized. ATP, ADP, etc. are separated by HPLC and detected by ultraviolet absorption. This approach has some disadvantages. Since ATP turns over in the cell within one second, the tissue must be maintained under physiological conditions (oxygenated, neutral pH) until metabolism can be instantaneously stopped. Organ preparations such as heart and muscle from small animals can be perfused, both outside and inside the animal. Muscle biopsies from humans (and in particular post-mortem tissue), in contrast, will not accurately reflect ATP levels in vivo. Secondly, this approach is invasive. It requires removal and destruction of the tissue under investigation which, aside from the obvious clinical problems, means that ATP levels cannot be followed over time in a single tissue. In experiments where time dependent changes are to be followed, multiple tissue samples (and statistical methods of analysis) are required. The Magnetic Resonance Technique The magnetic resonance (NMR) technique utilizes the ability of phosphorus (^^P) nuclei, when placed in a high magnetic field, to absorb radio waves. The wavelength (frequency) absorbed depends on the chemical environment of the nucleus; Pj, phosphocreatine (PCr), and the three phosphorus atoms in ATP will each absorb radiation (shown by peaks on an NMR spectrum) at slightly different wavelengths. The intensity of absorption (peak area) is proportional to the amount of material present which absorbs at that wavelength; thus, from the corresponding peak areas, the amounts of P,, PCr, and ATP in a sample can be quantitatively assessed (Radda, 1986). Since tissues are transparent to magnetic fields and to radio waves, this technique can be used to measure phosphates within the body, i.e., this technique is noninvasive. An arrangement for measuring metabolites within a human arm is shown in Figure 2a. Measurement is clearly made under physiological conditions, without having to freeze metabolism. Furthermore, since spectra can be taken within a few seconds, and the tissue is not altered in the process, the levels of ATP, etc. can be followed in time. Figure 2a shows, in fact, an arrangement for measuring levels of phosphate metabolites within arm muscle, and Figure 2b shows variations in these metabolites, during and after exercise. The main problem with the NMR method is its relatively low sensitivity. It requires gram quantities of tissue, and metabolite concentrations within the tissue of 1 mM or above. Thus, although it will detect Pj, ATP, and PCr, the technique is not sensitive enough to measure ADP or AMP levels, which typically lie below 100 iLiM.

Cellular ATP ^,

^^^^ ^f magnet

V V ^\ V

—)

1

a

blood pressure cuff (ZOOrnm Hg)

PCr,

RECOVERY

pH704 REST Figure 2. NMR measurement of ATP in human organs, (a) Device for exercising human arm in bore of NMR magnet, (b) NMR spectra of phosphate metabolites in human arm during and after anaerobic exercise. Note that nearly all the signal derives from muscle metabolites. During exercise, PCr levels are seen to fall, and ?, levels to rise, while [ATP] is hardly affected. Numbers denote intracellular muscle pH, which also falls due to lactic acid production.

ADP levels in muscle or brain may be calculated from NMR data, assuming creatine kinase to be at equilibrium in the cell, from the equilibrium relationship: Keq = [ATP][Creatine] / [ADP][PCr]. In these calculations, the value of Kgq is a known constant (66 at pH 7.1, 37 °C, etc.) and [ATP], [PCr] are measured by the NMR experiment. The concentration of free creatine must be measured enzymatically after extraction of the tissue (as above); normally it is measured as total (Cr and PCr) creatine at the end of the experiment. Note, howêver, that this calculation is possible only for those tissues (muscle, brain) which contain creatine kinase. In other tissues, [ADP] must be measured by the freeze-clamp procedure.

10

DAVID A. HARRIS

Adenine Nucleotide Concentrations Within Cells Typical values for ATP, ADP, AMP, Pj, and PCr concentrations in heart muscle, measured by each of the above techniques (Veech et al, 1979; Balaban et al., 1986) are given in Table 2. ATP levels at 8 mM are quite high relative to other metabolites (glucose-6-phosphate at 0.5 mM, citrate at 0.1 mM), reflecting the importance of this metabolite in a variety of metabolic processes. Similar ATP levels are observed in many tissues of the body. There is a clear discrepancy in Table 2 between the levels of ADP measured enzymatically (1.4 mM) and the levels calculated from the PCr/ATP equilibrium (0.04 mM). This reflects the fact that the freeze-clamp method extracts total ADP from the tissue, while the equilibrium calculation considers only that part of ADP in equilibrium with ADP and PCr, i.e., ADP that is free in solution. These figures differ, therefore, because most cellular ADP is bound to protein (largely actin) within cells. Since it is free ADP which is a substrate for ATP synthesis—which participates in the equation for the phosphorylation potential, AGp, and which regulates enzymes—ît is the calculated value which is taken as an indicator of cellular energy status. The ratio [ATP]/[ADP]free within these cells is 200 and AGp = -60 kJ/mol ATP hydrolyzed. Again both values are typical not only in heart but in a variety of tissues. Finally, despite the relatively high concentrations of ATP and PCr (together making up about 5% of the dry mass of the heart), their role as an energy store can only be short term. A rat heart uses about 2% of its high energy phosphate per beat; at this rate ATP would last about 3 seconds and PCr about 9 seconds more. Thus ATP generation, from metabolic fuels, must be rapid and continuous in heart as in all other tissues. Spatial Distribution One drawback to both techniques as described above is that they provide only an average value of nucleotide concentration across the tissue. This will obscure

Table 2. Adenine Nucleotide Levels in Rat Heart NMR measurements freeze-clamp methods

ATP

ADP

8 mM^ 8 mM

0.04 mM^ 1.4 mM

ASAP

n.d."^ 0.1 mM

Pi

0.5 mM 2-8 mM

PCr 23 mM 25 mM

Notes: Âbsolute estimation by freeze clamp methods. P,, PCr determined by peak areas relative to ATP peaks. ^Calculated from creatine kinase equilibrium (see text). ''Undetectably low by NMR. Measured values overestimate free phosphate due to some destruction of high-energy phosphates and contamination with extracellular phosphate.

Cellular ATP

11

any differences between cell types within the tissue, between different compartments (e.g., mitochondria and cytoplasm) within the cell, or (particularly in the case of Pj) between the intracellular and extracellular fluids. In many cases, this is unimportant. Adenine nucleotides are present in significant concentrations only within cells, so amounts in the perfusing medium can be ignored. Similarly, mitochondria occupy only a small fraction of cell volume, and thus contain only a small fraction of its metabolites; the figures in Table 2 represent, to a close approximation, cytoplasmic concentrations. Where more precise data are required, the above techniques must be modified as outlined below. Fractionation in Non-Aqueous Solvents In a technique pioneered by Hassinen and coworkers (Kauppinen et al., 1980), a tissue after freeze clamping has the cell water replaced by organic solvents (e.g., heptane/CCU). This prevents both enzyme function and exchange of metabolites between cell compartments. Cells are then fractionated into mitochondria, nuclei, etc., by homogenization and centrifugation—still in organic solvents—ând only then are aqueous extracts made for nucleotide assay. Rapid Cell Lysis/Centrifugal Fractionation This technique (Siess and Wieland, 1976) is suitable for cultured cells in suspension (e.g., hepatocytes). At the time of measurement, cells are mixed with a small amount of digitonin, which ruptures the plasma membrane. A sample is immediately added to the upper aqueous layer of a micro centrifuge tube; this layer is separated from a lower aqueous phase by a layer of (inert) silicone oil (Figure 3). The tubes are then centrifuged, and the unlysed mitochondria spin through the oil into the lower phase (normally perchloric acid), while the cytoplasmic contents remain in the upper aqueous phase. The aqueous phases can then be separately assayed for mitochondrial and cytoplasmic nucleotides. These two methods demonstrate that, in a variety of tissues, the ATP/ADP ratio within mitochondria is around 1:1, much lower than in the cytoplasm, where this ratio is about 200:1 (Table 2). Since ATP is made inside the mitochondrion and exported into the cytoplasm, the relatively low levels of mitochondrial ATP were unexpected. Their explanation lies in the energy dependent system for ATP export/ADP import across the mitochondrial membrane, which continually expels ATP in exchange for ADP (see below). Magnetic Resonance Tomography By using, in effect, a point source of radiofrequency radiation (a surface coil) in the NMR experiment, and a rotating receiver for data collection, ^^P-NMR spectra

12

DAVID A. HARRIS

cytoplasm silicone oil acid mitochondrial pellet

Figure 3. Rapid lysis/centrifugation technique for investigation of metabolite compartmentation. For explanation, see text. can be compiled for various depths within the body (typically 1—5 mm slices). By this method, (analogous to the use of X-ray tomography to map tissue density within the body), differences in the levels of PCr and ATP between cells can be mapped (Radda, 1992). As an example. Figure 4 shows a series of stacked plots, showing the levels of phosphorus metabolites at various depths within the human thorax. The changing

sChest wall muscle Surface coil phantom 2.3DPG

PDE

N

i

l

Blood Heart -• Skeletal muscle

Surface coil

Figure 4. Spatial resolution of phosphate metabolites in the human thorax. Phosphate metabolites were measured in 1 mm slices through the thorax, using ^^P-NMR tomography. The body surface is marked by the surface coil phantom. Muscle ATP is indicated by the three ridges on the right (see Figure 2). The two peaks in the ridge due to PCr mark the chest muscle and the heart muscle; the PCr/ATP ratio of skeletal muscle is seen to be higher than that of heart muscle. Note the left rearmost peak of 2,3 diphosphoglycerate (2,3 DPC); this is an allosteric regulator of hemoglobin, occurring in the heart ventricular blood. The ridge labeled PDE is due to phosphodiesters.

Cellular ATP

13

intensity of the PCr peaks identifies the chest wall (skeletal muscle) and the heart, comparing these with the ATP peaks (to the right on the diagram), we see that skeletal muscle has a higher PCr/ATP ratio than heart muscle. At the very back of the diagram, on the left, we see the heart cavity, marked by a high concentration of 2,3-diphosphoglycerate (bound to hemoglobin in the blood). In a similar, but more precise, study in canine heart (Robitaille et al., 1990), it has been shown that ATP levels in heart muscle remain constant from the outside to the inside of the heart; PCr levels, in contrast, fall significantly towards the endocardial side. This indicates a gradient in PCr and ADP levels from the subepicardial cardiomyocytes (high PCr, low ADP) to the subendothelial cells (low PCr, high ADP). This reflects subtle differences in energy metabolism across the myocardium.

USES OF ATP Contraction of Actomyosin The actomyosin system is designed to convert chemical energy (from ATP hydrolysis) directly into mechanical work. In skeletal and heart muscle, the cells (fibers) are packed with a dense, semicrystalline array of actomyosin, and this protein is responsible for up to 70% of ATP consumption in contracting muscle. (A remaining 20% is consumed in ion pumping; see below.) In other tissues, actomyosin filaments may also contribute to cell motion, e.g., in phagocytes and fibroblasts, but the filaments are organized only locally within the cell, and contribute much less to overall cellular energy consumption. The operation of actomyosin is described in the cross bridge model. In this model, ATP drives the release of the myosin head from one subunit in the actin filament. This is followed by a conformational change in the myosin head such that it rebinds to a different subunit of actin, and a tension generating step in which the myosin returns to its original conformation attached to this new site along the actin filament. ATP and the Energetics of Muscle Contraction It is instructive to consider how ATP powers this overall process, because this also serves as a paradigm for harnessing ATP hydrolysis to drive processes such as ion pumping (see below). It is tempting to imagine that cleavage of the bond between p and y phosphate groups directly energizes proteins in some way, since hydrolysis of this bond in free solution yields energy. This is, however, to reckon without the ability of enzymes to juggle the energy of intermediates in reaction pathways.

14

DAVID A. HARRIS

T

r\ ATP

AGp

ADP^Pi

*P.^

Figure 5. Energy release during ATP hydrolysis, (a) Energy release during hydrolysis of ATP in the absence of an enzyme, (b) Energy release during hydrolysis by myosin (E). Energy is released as ATP binds tightly to the enzyme. The enzyme not only decreases activation energy AG*, but shifts energy release from the expected, bond splitting step (shown dotted) to ATP binding (solid line). Note that the total energy release (AGp) is identical in each case, 'indicates an unstable intermediate species.

Thus, in myosin, hydrolysis of ATP involves four steps, not only (a) cleavage of the ADP-Pj bond, but also (b) formation of an ATP-myosin complex, (c) release of Pj, and (d) release of ADP from the myosin-ADP-Pj complex. The first law of thermodynamics tells us that AG for ATP hydrolysis is fixed (-60 kJ/mol) under cellular conditions; that is to say, when ATP is hydrolyzed to ADP + P,—with or without enzyme—60 kJ/mol must be liberated. It does not, however, tell us at which of these four steps energy is liberated. In the case of myosin, the affinity of the enzyme for ATP is so high (10^ x stronger than for ADP), that nearly all its energy is liberated when ATP binds to myosin before bond cleavage occurs. This is shown schematically in Figure 5. We have thus established the pattern of energy release from ATP during catalysis by myosin. How does this help us understand the contraction mechanism? The answer to this lies in the strength of the actin/myosin interaction. If muscle ATP is depleted (e.g., by stimulation in the presence of metabolic inhibitors), the muscle becomes rigid (tetanus). This is because all the myosin heads are fixed in their complex with actin. Thus, the actin-myosin complex, in the absence of ATP, is a strong one—it is energetically favored. When ATP binds to myosin, sufficient energy is released to dissociate the actin-myosin complex and initiate the cross bridge cycle (Figure 6a). Thus myosin exchanges a favorable interaction with actin for another with ATP; this is brought about mechanistically by the binding of ATP inducing a conformational change in myosin which distorts the actin binding site. Subsequent chemical changes in the myosin-ATP complex (ATP cleavage. Pi release, and ADP release) simply act as triggers for conformational changes in

Cellular ATP

15

myosin conformation (tilting the head, rebinding to actin, and realignment of the myosin molecule) as shown in Figure 6a; tension development occurs at the final stage in the scheme (Hibberd and Trentham, 1986). Energy changes in the system are shown in Figure 6b. ATP binding to myosin liberates energy which is used to dissociate the actin/myosin complex. Dissociated actin and myosin thus constitutes an energized system. Recombination of actin and

hydrolysis

M

^sj^•^'''P

Au-Mro^ A+MC2)

...*/-.N AM*C3)

tension I• C I •I • I

30 k>

'

AM (1)

1

AM (4)

Figure 6. Mechanism of ATP use in muscle contraction, (a) The cross-bridge model for muscle contraction, incorporating ATP binding and hydrolysis steps. Note that each chemical change (ATP binding, bond splitting, ADP, and Pj release) produces a (kinetically) unstable conformation (*), which relaxes in the next stage of the cycle. The numbers (1), (2), etc. represent different conformational states (see text, and Figure 6b). T represents bound ATP, and D represents bound (ADP + Pj). (b) Energetics of the cross-bridge model. Energy is transferred from ATP (solid line) to actin (A) + myosin (M) by dissociation of actomyosin (AM) (dashed line), and on to tension development (dotted line) after reformation of the AM complex. For simplicity, the dissociations of Pj and ADP are shown as a single step.

16

D A V I D A . HARRIS

myosin, permitted after ATP hydrolysis and conformational changes in the myosin head, thus releases this energy which is used in tension development. Ion Pumping

Most small molecules and ions are moved across cell membranes by a variety of porters (symports, antiports, etc.) using energy stored in transmembrane ion gradients (see above). Examples include the gut glucose/Na"*" symport and the Na'^/Ca^'^ antiport at the plasma membrane, and the Pj/H"^ symport and Ca^"*" uniport at the mitochondrial membrane. Some transport systems, however, are powered directly by ATP hydrolysis. These are known as primary transport systems because, typically, they build up the ion gradients (e.g., Na^ across the plasma membrane) which drive the secondary transport systems described above. Examples of such ATP-driven pumps are shown in Figure 7. NrfA

drugs

>NQ

&

P-type

85 ABC-type

fi

F-typg

tt

V-type

Figure 7. ATP driven pumps in eukaryotic cells. The pump orientation is indicated by the position of the ATP binding domain(s) (shown circular); e.g., cytoplasmic-facing for P,V and ABC type pumps. The direction of pumping is indicated by the arrows. H"^ ions are pumped where no specification is given; the plasma membrane Na"^ pump also pumps K^ ions inwards. Mito = mitochondrion; SR = sarcoplasmic reticulum; PM = plasma membrane; SV = storage vesicle (e.g., chromaffin granule, synaptic vesicle); EV = endocytotic vesicle.

Cellular ATP

17

In gross structure, all of these pumps appear similar; they comprise a transmembrane region surmounted by a large aqueous domain facing the cytoplasm. At the molecular level, however, they fall into four distinct families. P'Type Pumps In terms of net energy consumption, P-type pumps are the major ATP-driven transport systems in mammals. They include the ubiquitous NaVK"*" pump which maintains Na"^ and K"^ gradients (and thus the steady state membrane potential) across the plasma membrane. This pump is responsible for up to 40% of all ATP utilization in the brain, rather less in other tissues. Other examples include the Ca^^ pump of sarcoplasmic reticulum (responsible for up to 20% of ATP utilization in active muscle), and the gastric H^ pump, which acidifies the stomach lumen. P-type pumps contain a long polypeptide chain (about 1,000 amino acids) which traverses the membrane up to 10 times. The polypeptide contains its ATP binding site on the large extramembrane domain (between transmembrane helices IV and V) and ion binding sites (ionophores) within the membrane domain (MacLennan, 1990). In some cases, notably the Na'*'/K'^ ATPase, a second smaller polypeptide (P) is present but its function is unknown. P-type pumps are unique in carrying out ATP hydrolysis in two stages, the first involving phosphorylation of the enzyme on an aspartic acid residue (asp 351). The mechanism is thus (i) E + ATP -> E-P + ADP

(ii) E-P + H2O -^ E + Pj

This may be useful in partitioning the energy of ATP hydrolysis between movements of different ions (see below). V-Type Pumps V-type pumps pump only protons (H"^); they are responsible for the acidification of intracellular compartments. An example is the chromaffin granule ATPase, which occurs in the epinephrine-storage vesicles (chromaffin granules) of the adrenal medulla. H^ pumped into these vesicles allows them to trap catecholamines as the charged protonated form which cannot cross the vesicle membrane. Acidification of endocytotic vesicles, synaptic (acetylcholine storage) vesicles, and lysosomes involves related pumps. V-type ATPases contain upwards of 10 separate polypeptide chains, and are separable into a soluble headpiece (containing the ATP binding site) and a transmembrane sector (containing the H"^ channel) (Nelson, 1992). In these, as in the remaining pumps in this section, ATP hydrolysis occurs by direct attack of water on ATP; no phosphorylated enzyme occurs.

18

DAVID A. HARRIS

ATP Binding Cassette (ABC) Pumps ABC pumps typically do not pump cations and, in fact, may have rather wide specificities (Hyde et al, 1990). The best known example in humans is the multidrug resistance (MDR) protein (also known as the P-glycoprotein) which . pumps large organic molecules out of cells. By pumping compounds such as doxyrubicin, vincristine, etc., out of tumor cells, it can be responsible for the low sensitivity of these cells to cytotoxic agents. A similar protein is responsible for chloroquine resistance of the malarial parasite. Other examples include the chloride channel protein defective in cystic fibrosis (CFTR)—^where the compound pumped is still unknown—and the peptide transporter (TAP1-TAP2) involved in antigen presentation in lymphocytes. ABC pumps comprise a dimer of ATP binding domains (outside the membrane) bound to a dimer of transmembrane domains, each comprising six transmembrane helices. Interestingly, the polypeptide organization of these proteins may vary; TAP1-TAP2 contains two polypeptides, each containing one transmembrane and one ATP binding domain, while in the multidrug resistance protein the two are fused in a single polypeptide containing all four domains (1,280 amino acids). F-Type ATPases The F-type ATPases are ATP-driven H"*" pumps. They show similarities in structure to the V-type ATPases, in that they are separable into a soluble headpiece (5-6 different polypeptides) and a transmembrane H^ channel (3-5 polypeptides) (Senior, 1988). However, there are characteristic structural differences, the most obvious occurring in the proton channel; F-type pumps employ a peptide about 80 amino acids long (in 10-12 copies) as a H^ carrier, while in V-type pumps the equivalent peptide is twice as long. More important, however, is the difference in function. In animals, the only F-type pump, which is found in mitochondria, does not act as an ATP-driven H"*" pump but in reverse, as the H"^-driven ATP synthase responsible for all oxidative ATP synthesis. This is dealt with further below. ATP and the Energetics of Ion Pumping In ATP-driven pumps, ATP, which binds to the cytoplasmic domain, cannot interact directly with the ion being pumped, which passes through the membrane sector. Energetic coupling, therefore, is indirect. Ion movement occurs via an alternating access model; the ion binding site is exposed to one side of the membrane in one conformation (EO and to the other side in the other (E2) (Figure 8a). These ideas can be combined with those used to derive the above model for actomyosin function. Critical features are: (a) energy release from ATP is associ-

K.Ei.ATP

JBaSI>B0 VSIDZNE TRIPHOSPHATE

Role of ATP in activating hydroxyl groups to nucleophilic attack.

ATP as Phosphate Donor As ATP has a high phosphate transfer potential, it can transfer its terminal phosphate to an alcohol (OH) group, forming a phosphate ester, in a downhill (thermodynamically favorable) reaction. In contrast to the reactions in the previous section, the resultant compound is not especially reactive; the primary reason for such phosphate transfers is to confer negative charge onto the recipient molecule. Phosphorylation of Sugars The archetype of this class of reaction is the phosphorylation of glucose by hexokinase. This is a downhill reaction—the equilibrium is well over towards glucose-6-phosphate—^which ensures that, within (non-liver) cells, free glucose levels are kept low. The glucose-6-phosphate formed, which is negatively charged, does not readily cross the cell membrane and is thus retained within the cell.* This class of reaction rarely makes significant demands on the cellular ATP content. However, it can do so in the pathological condition of fructose intolerance. The normal pathway of fructose metabolism, which occurs in the liver, requires In liver cells, glucose entry is so fast relative to phosphorylation that free glucose does build up; this allows the liver cell to "sense" blood glucose levels, and is associated with its role in maintaining blood glucose by taking up or releasing glucose.

Cellular ATP

23

two novel enzymes, fructokinase (producing fructose 1 phosphate) and fructose 1 phosphate aldolase (which cleaves fructose 1 phosphate into 3 carbon sugars). In hereditary fructose intolerance, the aldolase is absent, and continued fructose intake will cause a build up in the liver of fructose 1 phosphate with accompanying depletion of cell phosphate and ATP. Phosphorylation of Proteins Protein kinases will transfer phosphate from ATP onto specific hydroxyl residues (serine, threonine, or tyrosine) within proteins. This is commonly associated with enzyme activation (e.g., glycogen phosphorylase, plasma membrane L-type Ca^"^ channel) or inactivation (glycogen synthase, pyruvate dehydrogenase). Thus these phosphorylations play a regulatory role. The high phosphate transfer potential of ATP again ensures that reaction can be virtually complete, i.e., nearly all enzyme molecules are in one form or the other. The sensitivity of the system is, therefore, high compared to allosteric regulation (which is based on reversible, non-covalent, binding equilibria). The role of phosphate in regulation is based largely on its charge. In glycogen phosphorylase, phosphorylation of serine 14 allows this N-terminal region to bind electrostatically to a cationic hole in the protein, triggering a conformational change at the (distant) active site (Browner and Fletterick, 1992). Phosphorylation of membrane proteins may be less precise in its effects; simply changing their surface negative charge may allow aggregation of membrane proteins (e.g., insulin receptors) or cause disaggregation (e.g., in the chloroplast membrane in green plants). ATP as Charge Neutralizer ATP is commonly found in intracellular storage granules. For example, the chromaffin granules of the adrenal medulla contain ATP levels of about 100 mM, 15 times higher than cytoplasmic levels. This ATP is metabolically inert, and seems to exist in a complex with epinephrine such that the positive charge on the catecholamine is neutralized by the negative charge on ATP"^". (Epinephrine, with one positive charge, can reach concentrations of up to 400 mM inside the granules.) Serotonin (platelets), insulin (pancreatic (3 cells), and acetylcholine (synaptic) storage granules also contain ATP. This ATP is released into the blood on exocytosis, along with the hormone. Here it is rapidly hydrolyzed to adenosine. This, too, has some endocrine action in causing the relaxation of vascular smooth muscle, increasing local blood flow, and thus aiding hormone delivery to the target tissues.

24

DAVID A. HARRIS

ATP and Messenger Molecules Generation ofcAMP In a reaction catalyzed by adenylyl cyclase, ATP is converted to 3'5'-cyclic AMP (cAMP). This compound is a ubiquitous signal molecule, generally indicating a stress situation: in both Escherichia coli and man, for example, cAMP is produced in response to nutrient limitation (starvation). In mammals, adenylyl cyclase is a membrane-bound enzyme which is activated in response to a variety of hormone receptors in the cell membrane, notably those for epinephrine and glucagon. cAMP is a second messenger for these hormones, and activates protein kinase A within cells. The reaction producing cAMP is shown in Figure 11. Due to strain in the ring formed, cAMP is (like ATP) thermodynamically unstable. The formation of cAMP from ATP is thus favored only by the hydrolysis of PPj by cellular pyrophosphatase (roughly equivalent to the hydrolysis of one high energy bond). The hydrolysis of cAMP to AMP (roughly equivalent to hydrolysis of the second high energy bond) is catalyzed by phosphodiesterase. The presence in a cell of two enzymes capable of the net uncoupled hydrolysis of ATP is a potential hazard. It can be supported, however, because the maximal capacity of adenylyl cyclase is low (and its activity is generally suppressed even further) such that cAMP is maintained at a basal steady state level of around ICT^ M, five orders of magnitude lower than ATP. When adenylyl cyclase is activated, a new steady state is established with c AMP at about 1 Qr^ M (the phosphodiesterase simply responding passively to increased cAMP levels). Thus ATP provides a nearly infinite pool of potential cAMP: cAMP levels can be changed 10-100-fold with a loss of less than 0.1% of total cell ATP. The maintenance of a large pool, and a low steady state value, of messenger molecules is an essential feature of signaling in biological systems, because it allows a rapid, many-fold change in messenger concentration. The same features, with a rather different organization, occur in Ca^"^ second messenger systems; intracellular [Ca^"*^] is normally around 10"^ M, but can be rapidly increased, in

Qdeninei

0

0

0

laden ine|

^^ ppi

K'

\

r "o

^ — 1 I OH 3 ^ 0 - P = r C / 0 molecule are Figure 11, Formation of cyclic AMP. The 3' and 5' positions of the indicated. PPj produced is rapidly hydrolyzed /n wVo(see above).

Cellular ATP

25

response to a hormone signal, to 10"^ M by transiently opening channels to a large pool (10~-^ M) of intra- or extracellular Ca^^. ATP Dependent IC Channels Besides ion channels controlled by ATP dependent phosphorylations (e.g., plasma membrane L-type Ca^^ channels), cells in a number of tissues contain a plasma membrane K"*" channel which is inhibited by the non-covalent binding of ATP. This channel normally mediates K^ efflux (leading to hyperpolarization), and can be demonstrated as an ATP-sensitive channel in patch clamp experiments. However, since I50 for blocking the channel is only 10-50 |LIM, some 100-fold lower than cellular ATP levels even during ATP depletion, the channel would be expected always to be closed in normal cells. Its role in normal cell function is thus unclear. One interesting suggestion is that this ATP-dependent K"^ channel might trigger insulin release in pancreatic P cells (Ashcroft and Rorsman, 1989). These cells respond to a rise in blood glucose in the range 3—10 mM by increasing their metabolism of glucose, (and thus ATP generation). In this model, then, as blood glucose rises, ATP levels should rise, promoting closure of this K"^ channel. Th6 cell thus depolarizes due to a net cation (Na"*") influx, and insulin release is triggered. This model cannot be regarded as proven, due to discrepancies between the measured changes in ATP concentration in p cells and those predicted for this model to operate. However, it remains an attractive model for the coupling of blood glucose concentration to insulin release. Structural Role of ATP The adenine moiety of ATP is used as part of cellular structures (e.g., coenzymes, RNA, DNA). The amount of ATP consumed in this way will depend upon the biosynthetic activity of the tissue. However, only the roles of ATP as an energy/phosphate source are discussed further in this article.

REACTIONS INVOLVING EXCHANGE OF HIGH ENERGY PHOSPHATES Two important reactions of ATP do not result in a net loss of phosphoanhydride bonds. These are the creatine kinase and adenylate kinase reactions. Creatine kinase: Cr + ATP ^mol

L A D ^ E.ATP ^^^

'. 3H- in

Figure 16. Mechanism of the mitochondrial ATP synthase, (a) Energetics. ATP is formed without energy input on the enzyme surface (solid line), due to the high affinity of the enzyme for ATP. For synthesis of free ATP, H"" ions moving downhill (dotted line) change the enzyme conformation, decreasing ATP affinity, (b) Alternating site mechanism. The three active sites can each exist in three different conformations: t = tight ATP binding; o = open, unable to bind nucleotide; and I = loose, in which ADP and Pj can exchange rapidly (dotted arrows) with the solution. The central mass represents the polypeptides which link Fi with the proton channel (not shown); this associates with the o conformation of active site only. Protons passing through the channel displace these polypeptides from one active site to the next (counterclockwise in this diagram), and the active site conformations thus change in sequence.

33

34

DAVID A. HARRIS

nucleotide translocase, which exchanges internal ATP for cytoplasmic ADP. This uniquely utilizes free ATP rather than the MgATP complex; if MgATP were exported, the mitochondria would lose internal Mg^"*" and charge balance would be upset. Hence, ATP is exported as ATP"*^ and ADP imported as ADP-^~. This process is energetically favored because the interior of the mitochondrion is negative relative to the outside, as a result of pumping protons (H"^) outwards. In principle, the two transporters, the ATP ADP translocase and the ?J¥t symport (which imports Pj into mitochondria) can be considered as a coordinated system in which ATP"*" is exported, while ADP-^ + P?~ (substrates for ATP synthesis) are taken up at the cost of moving 1H"*" down its electrochemical gradient. This requirement for energy for ATP export explains how the cytoplasmic ATP/ADP ratio can exceed the mitochondrial ATP/ADP ratio by a factor of about 100 (see Table 2); ATP is actively expelled and ADP pumped inwards. P/O Ratios The P/O ratio is defined as: P/O = mols ATP made/atom O consumed. This will depend upon the substrate oxidized: NADH, which is a strong reducing agent (E^'=-0.3V), will yield more ATP than succinate (E^'= OV). In mechanistic terms, this is reflected in the ability of NADH/UQ oxidoreductase complex to pump H"*", while succinate/UQ oxidoreductase cannot. This ratio is clearly an important parameter in quantitating cellular energy metabolism. This being so, it is surprising that its value is not known with certainty. Consensus values are P/O = 2.5 for NADH oxidation and P/O = 1.5 for succinate oxidation (Ferguson, 1986). Since, in mitochondria, 3H"^ are used by the ATP synthase, and IH"*" by the translocase in synthesizing 1 mol (cytoplasmic) ATP, this suggests that 10 H"^ are pumped out per mol NADH oxidized. Problems in determining this ratio have been both practical and conceptual. Practical problems center on the metabolic cost of transport; different membrane preparations will show different P/O ratios, depending on whether the synthase is internal (right side out), or external (inside out) where the need for ADP and ATP transport is abolished. The conceptual difficulties arise from a historical tendency to expect this value to be a whole number, in various mechanistic models for ATP synthesis. However, since we now know that several H"^ are required to make 1 ATP, non-integral values for the P/O ratio no longer raise any conceptual problems. Various chemicals known as uncouplers (e.g., 2,4 dinitrophenol, picric acid) can decrease the P/O ratio by increasing the permeability of mitochondrial membranes to H^. In this case, protons leak across the membrane, bypassing the ATP synthase and producing heat. These chemicals, used in explosives manufacture, were responsible for weight loss and tissue wasting among explosives workers early in this

Cellular ATP

35

century. A similar syndrome is observed in a rare genetic disease of mitochondria, Luft's disease, where the membrane again is abnormally leaky to H"^ (probably due to malfunctioning Ca^"^ transport).

CONTROL OF ATP LEVELS ATP Levels Are Closely Maintained In Vivo

Since ATP participates in a wide variety of metabolic processes, it is hardly surprising that its levels are tightly controlled. Nonetheless, the degree of control observed is remarkable; variations in work rate of up to 10-fold in heart, and even more in skeletal muscle, produce no detectable ( O

: N -irJ 3 ^ ^ OJ i^ u . (/) Q Q.

0

"S "8 •^ s: 2^ a^ o^ 'S; o g; a _•SO g; uON '^ (1)

' - - D ' - - D ! : ^ "D - C - C 'S - C 52 nj jf) 03 i2 f^ j D _g; _Q j y

Q.

z < z

A Q. CO A t^

c

U i^ U iî: (>5 U

Q.

A

z uA z

CO A Q-

z < z u

c

OJ >

O

T3 0 3

CO 03

Q

2 •g-o

c"

? 1'! i •£

c E "5

OJ

k.

c

OJ 5^ C 03

5 03

Q

03

Q

CO

U O

r—. (N "^ -^

r—

J^

^-

>- i n

^

7ZZ

E 'OJ

u.

"S ^ < 'x o S O

r^

^ ^ o e:

2 f -^ -S ocy = ^

£

^ ^ i .E

U

o — "^ re

-a i o °

:E J5 M> a •o ^ ^ c

DO (U Si - §

u c

2 is c - ^ E ^ C L c: :5

^

DO(J

I n* ^

DO

(T

=

>

= 92

Regulation ofcNMP Synthesis and Deregulation

93 Name

Progressive sequence alignment of PDE family catalytic domains

r cA-spec r cA-spec r c A-spec r cA-spec r cA-spec -b61kCaM .b63kCaM

c^ 50

-JL.

-1 V

D

S

G Q / V

G \

'

N

D'-â2+ Y I / V E jy Ê LRHVMTNLGEKLTDEEVDEMIREA FVQMMTAK A

A

Helix

Helix

Helix

B

COOH

Figure 2. Two EF-hand domains In calmodulin. (A) Residues 78-148 of rat calmodulin covering the third and the fourth EF-hand structures are shown. In each EF-hand, two eight residue-long hydrophobic alpha-helices (underlined) flanking a 12 residue-long loop with position 1,3,5, 7, and 12 providing side chain oxygen for Ca^^ coordination. (B) The two EF-hand structures are illustrated schematically. Functional EF-hand calcium-binding sites are usually found in pairs and in an orientation that allows hydrophobic interaction between helices from both EF-hands. Amino (NFH2) and carboxyl (COOH) ends are as shown. Modified with permission of Elsevier from Evans etal. (1988).

calcium-binding proteins, there are an even number of EF-hands, e.g., tvvô in calbindin-9K, four in calmodulin and troponin C. Twô adjacent EF-hands are generally arranged to be in close proximity so that the binding of the first calcium ion enhances the binding of the second Ca^"*" (positive cooperativity). All of the X-ray crystallography data relating to EF-hand CaBPs have only been obtained vîth the Ca^^-bound form due to stability reasons. On the other hand, circular dichroism and ^H NMR studies have revealed significant conformational

260

KEVIN K.W. WANG

changes due to calcium binding in an exposed hydrophobic region of the protein (presumably, the helical region). Since most of the EF-hand proteins do not have enzymatic activity, it is assumed that the Ca^'^-dependent exposure of hydrophobicity allows the calcium-binding protein to bind and activate its target protein, which is usually an enzyme or a dynamically regulated structural protein. This will be examined in more detail, using calmodulin as an example, in a later section. EF-Hand Protein Family The availability of the crystal structure of calbindin, calmodulin, and troponin C has clearly confirmed the EF-hand structure found in parvalbumin (see Heizmann and Hunziker, 1991). Many novel proteins have already been cloned, their cDNA sequence obtained and their amino acid sequence deduced. By homology to existing EF-hand sequences, more than 200 new calcium-binding proteins of this family have been identified. They can be further divided into subfamilies based on evolutionary origin (Kretsinger et al., 1991). Table 1 lists a few of the EF-hand calcium-binding proteins with more defined functions. Calbindin-9K is the smallest with only two EF-hands. Parvalbumin and a CaBP highly expressed in tumor cells (oncomodulin) have three EF-hands but the first loop is not functional. The ubiquitous four EF-handed calmodulin (Manalan and Klee, 1984) and myocytespecific troponin C and myosin light chain appear to be closely related. The small subunit of the protein phosphatase, calcineurin, and the Ca^"^-dependent protease, calpain, have four EF-handlike structures. The S-100 sub-family includes S-lOOa and S-100b which have 14 residues instead of 12 in one of their two calcium binding loops (Kligman and Hilt, 1988). Myeloid related proteins (MRP-8 and MRP-14), also called calgranulin A and B, are S-100-related proteins with two EF-hands found in monocytes and neutrophils. They are known to translocate to the cell membrane upon cell activation (and thus calcium-binding). They also are thought to be involved in inflammatory reactions (Odink et al., 1987). Another S-lOObrelated CaBP is calcyclin which is involved in cell cycle progression (Calabretta et al., 1986). Sorcin is a four EF-hand protein that is overexpressed in multi-drug resistant cell lines (Meyers et al, 1987). Calbindin-D28K and calretinin have six EF-hands (Table 1). More recently, two types (beta and delta) of crystallins, which are major soluble proteins present in fiber cells that form vertebrate lens, were found to have an EF-hand structure, and to bind Ca^"^ (Balasubramanian and Sharma, 1991). In the following sections, we will look at several examples of such EF-hand proteins and discuss their proposed functions in more detail. Calbindin-9K and Calbindin-D28K Calbindin-9K is a member of the EF-hand family of calcium-binding proteins with two helix-loop-helix regions that bind two calcium ions. Calbindin-9K is

Calcium-Binding Proteins

261

Table 1. Examples of EF-Hand Calcium Binding Proteins and Their Functions Calcium-Binding Protein Calbindin-9K Calbindin-D36K Calcyclin Calcineurin Calmodulin Calpain Calretinin Crystal I ins (p and 6) Myosin light chain Parvalbumin Recoverin S-100a S-lOOb Sorcin Troponin C

Proposed Functions Vitamin D-dependent calcium deposition Intracellular calcium buffering, intestinal calcium transport Cell cycle progression Calcium-dependent protein dephosphorylation Multi-functional, activates calmodulin-binding proteins and enzymes Calcium-dependent proteolysis Intracellular calcium buffering Lens structural proteins Muscle contraction Calcium buffering Phototransduction Unknown Neurite extension Multidrug resistance (?) Muscle contraction-relaxation

present primarily in bone and cartilage (Balmain, 1991). In cartilage, it occurs only as a cytosolic protein in mature chondrocytes while in bone, it is found in osteoblasts and mature osteocytes. It is also secreted into the extracellular matrix. Its synthesis is strictly dependent on vitamin D. It is likely that calbindin-9K plays an important role in calcium-deposition and subsequent mineral nucleation in the extracellular matrix vesicles of calcifying cartilage and bone. Calbindin-9K has attracted much attention because of its small size which is ideal for molecular structure determination using 2D-protein NMR or X-ray crystallography (Heizmann and Hunziker, 1991). Interestingly, there is also a homologous but larger calbindin-D28K which contains six EF-hand structures in nonmineralized tissues. Its synthesis is also under vitamin-D control. Calbindin D28K is abundant in some regions of the brain and may serve as a calcium buffering system (vide infra). Calbindin-D28K is also found in intestinal absorptive cells and has been shown to activate the plasma membrane Ca^'^-Mg^'^-ATPase which is implicated in active calcium uptake from the intestine (Wasserman et al., 1992). Calmodulin and Calmodulin-Binding Proteins

Calmodulin is present in virtually all eukaryotic organisms and in all cell types (Klee and Vanaman, 1982). It is a small protein (16.5 kDa) composed of four EF-hand structures that bind four calcium ions (Babu et al., 1985). The amino acid sequence from many species has been determined and an exceptionally high degree of conservation was found. Calmodulin is well known to regulate various cellular functions via its interactions with various calmodulin-binding proteins.

262

KEVIN K.W. WANG

Cellular functions regulated by calmodulin include cell motility, mitosis, cAMP metabolism, exocytosis, protein phosphorylation/dephosphorylation, and Ca^"^ transport (Manalan and Klee, 1984). As an exception to most of the EF-hand CaBPs, not one but many target proteins for calmodulin have been identified. The effects of calmodulin on these many cellular functions are thought to be mediated by the calmodulin-binding proteins. Calmodulin-binding proteins can be divided into three groups: (a) enzymes, (b) cytoskeleton (structural) proteins, and (c) miscellaneous. The enzyme group includes two metabolic enzymes (phosphofructose kinase and phosphorylase kinase) (Mayr and Heilmeyer, 1983; Chan and Graves, 1984) and two key enzymes in the control of cAMP levels: adenylate cyclase in the formation of cAMP and cyclic nucleotide phosphodiesterase (PDE) in its breakdown (Cheung, 1971; Yeager et al., 1985). The plasma membrane Ca^'*"-Mg^'*^-ATPase is also calmodulin-activated (Wang et al., 1992). Several calmodulin-dependent enzymes are involved in the protein phosphorylation/dephosphorylation mechanism, including: (a) phosphorylase kinase; (b) myosin light chain kinase (Klee, 1977); (c) calmodulin-dependent protein kinase II (which phosphorylates a large number of cellular proteins) (Kennedy et al., 1987); and (d) calmodulin-dependent phosphatase (calcineurin) (Tallant and Cheung, 1986). Inositol 1,4,5 trisphosphate kinase, a key enzyme in inositol-phosphate signal transduction is also calmodulin-stimulated (Johanson et al., 1988), as well as nitric oxide synthase, the enzyme that produces the highly unstable second messenger nitric oxide (Lowenstein and Snyder, 1992). In the cytoskeletal/structural protein group, erythroid spectrin binds calmodulin with low affinity (Sobue et al., 1981b), while the brain spectrin (fodrin) binds with high affinity (Carlin et al., 1983). Tubulins (a and (3) also bind calmodulin with low affinity (Kumagai et al., 1982). The microtubule associated protein 2 (MAP-2) and Tau factor also appear to bind calmodulin (Sobue et al., 1981a; Lee and Wolff, 1984). Adducin is a calmodulin-binding protein present in the plasma membrane that promotes association of spectrin and actin (Bennett et al., 1988). This activity of adducin is inhibited when it binds calmodulin. Other calmodulin-binding proteins include the calcium release channel of the junctional sarcoplasmic reticulum (Seiler et al., 1984) and both the liver and the lens gap junction proteins (Welsh et al., 1982; Zimmer et al., 1987). Neuromodulin is a neuro-specific calmodulin-binding protein (also called GAP-43) that is involved in cell growth. Interestingly, neuromodulin appears to bind the noncalcium-bound calmodulin more tightly rather than the calcium-saturated calmodulin (Andreasen et al., 1983). In addition, calmodulinbinding proteins as a group appear to be selectively susceptible to proteolytic attack by calpain (Wang et al., 1989). Surprisingly, the amino acid sequence of the calmodulin-binding region of these CaMBP are not highly conserved. Instead, they share a similar three-dimensional conformation: amphiphilic alpha-helix, i.e., an alpha helical structure with hydrophobic amino acids (Ala, He, Leu, Trp, Val) on one side and basic amino acid (Arg,


263

Figure 3. Amphiphillc helix as calmodulin-binding motif in calmodulin-binding protein. Residues 4-17 of myosin light chain kinase from smooth muscle is the target region for calmodulin. The sequence is predicted to form an alpha helix. The residues are fitted into the helix wheel (Kyte and Doolittle, 1982) with the backbone of three consecutive residues. The helix can be visualized to spiral away from the viewer. It is observed that four basic residues (bold letters) cluster on one side of the helix while five out of six hydrophobic residues (outlined letters) concentrate on the opposite side of the helix.

Lys, His) on the other side of the helix (Figure 3) (O'Neil and DeGrado, 1990). Hydrophobic compounds such as phenothiazines interact with calmodulin in a calcium-dependent manner. This is consistent with the view that upon binding calcium, calmodulin exposes the hydrophobic helical region (Weiss et al., 1982). This hydrophobic region in turn interacts with the calmodulin-binding helix. Presumably, electrostatic interactions between the acidic groups on the calciumbinding loop of calmodulin and the basic side chains of the amphiphilic helix of the calmodulin-binding protein are also important. Troponin C

The myosin and actin filaments of myofibrils in skeletal muscle lie in parallel arrays. As illustrated in Figure 4, the cross-bridge is the protruding head of the myosin thick filament. This component possesses ATPase activity and when muscle is not in the active state the myosin molecules are said to be in the myosin-ADP-Pi state. The thin actin filament has several proteins including tropomyosin and three troponins, troponin T, I, and C (Leavis and Gergely, 1984). Very briefly, myosin-actin interaction is inhibited in muscle by tropomyosin and troponin I during the relaxed state. Notice that during this state the cross-bridge is not attached to actin (Figure 4a). However, during activation the myosin molecules undergo a change in state that leads to cross-bridge attachment, release of Pi, and the development of force (Figure 4b). During this phase of power stroke, the cross-bridge rotates from a 90° to a 45° angle (Figure 4c), leading to sliding of the

Tropomy.,^

k

ADP

(c)

Sliding movement

Figure 4. The role of troponin C in skeletal muscle contraction, a) The myosin head (a mechanoenzyme) projects at a 90° angle from the filament backbone in the direction of the actin filament. In muscle at rest, the cross-bridge is not attached to actin, and myosin is in the ADP.Pi state, b) Cross-bridge attachment to actin occurs at a 90° angle when myoplasmic free Ca^* increases allowing troponin C to bind Ca^^. This leads to a conformational change involving the movement of tropomyosin from its blocking action in the groove. Pi is ejected from the myosin head when attachment to actin takes place. The current view is that Pi release is the rate-limiting step in force generation, c) As the attachment increases, so does the development of the active force in the cross-bridge (strain phase). The cross-bridge rotates and assumes a 45° angle relative to actin. (This is a phase of ADP ejection). Such a rotation has been demonstrated in electron micrographs of flight muscle of the giant water bug Lethocerus. The filaments slide past each other without a change in their length. ATP then takes the place of ADP in the myosin head and when bound is eventually split by the myosin Mg^^-ATPase to give ADP.Pi. The resulting conformational change leads to detachment of the cross-bridge from the actin filament and its return to the 90° angle state. 264


265

filaments past each other. ADP is then released from the myosin head. ATP takes its place which when bound is split by the myosin ATPase. The cross-bridge detaches and returns to the 90° angle state (Figure 4a). The trigger event leading to the above sequence of changes is the rise in myoplasmic free Ca^"^ concentration from about ICT^ M to ICT^ M. Step one involves de-inhibition of cross-bridge attachment to the actin filament as the result of the binding by troponin C of Ca^"^. Presumably this occurs after the movement of tropomyosin in the groove of actin (Figure 4b). The return to the relaxed state involves a fall in the myoplasmic free Ca^"*" concentration back to 10"^ M, which is brought about primarily by the Ca^"*" pump of the sarcoplasmic reticulum (SR), along with closure of the sarcolemmal Ca^"^ channels and the SR calcium release channels. If, however, the myoplasmic Ca^"*" level is still raised and Ca^"^ remains bound to troponin C, then the above chemo-mechanical cycle is repeated cyclically. The binding of troponin I to troponin C has been investigated using a 12-residue peptide corresponding to the troponin C binding site on troponin I (residues 104—115). Nuclear magnetic resonance spectroscopy studies reveal that the troponin C-bound form of the peptide is an amphiphilic helix with basic residues on one side and hydrophobic residues on the other (Campbell and Sykes, 1991). Such conformation is essentially identical to the pattern found in calmodulin-binding sequence in many proteins (see above). EF-Hand Structure as a Domain of Protein

Thus far, the EF-hand proteins examined basically involve repeats of the helix-loop-helix. However, EF hands are found as a domain of more complex holoproteins. For example, the 17 kDa subunit of the calmodulin-activated protein phosphatase calcineurin is a calcium binding protein which possesses four EFhands (Figure 5) (Klee et al., 1988; Guerini et al., 1989), while a phosphatase domain and a calmodulin-binding domain exist in its large subunit (Guerini and Klee, 1989) (Figure 5). Moreover, calcineurin has been identified to be the target protein for the immunosuppressant drug, cyclosporin A (Liu et al., 1991). In regard to calmodulin itself, it has also been shown to be one of the five subunits of phosphorylase kinase (Cohen, 1988). Both calcineurin and phosphorylase kinase can bind another molecule of calmodulin and thereby increase their activity. The calcium-dependent cysteine protease, calpain, is another intriguing example. It has a large catalytic subunit and a small regulatory subunit (Suzuki, 1987). The 80 kDa subunit of calpain contains a catalytic cysteine protease domain (similar to papain for example) and an EF-hand Ca^"^-binding domain (Figure 5). The regulatory subunit (29 kDa) has a glycine-rich region on the N-terminal side and four more EF-hand calcium-binding sites on the C-terminal side (about 18 kDa). It is believed that these EF-hand Ca^"*"-binding domains impose strict Ca^'^-dependence on the catalytic activity. More recently, actinin was found to contain an actin binding

KEVIN K.W. WANG

266

Calpain 4EF-hand calciumbinding sites IV

Cysteine protease domain SOKsubunit

Cys

His fi^^

^ ^ ^

4 EF-hand calciumbinding sites VI

^

30 K subunit

Calcineurin Phosphatase domain

Caiyi-blndlng domain

60 K subunit

17 K subunit

Figures. Structural models of calcineurin and calpain. The large (80 kDa) and small (29 kDa) subunits of calpain combine as six domains. Domain II is a cysteine protease domain while domain IV and VI are both calcium-binding domains with four EF-hand structures each. Calcineurin has a larger subunit that contains a phosphatase domain and a calmodulin-binding domain (filled area). The small subunit is the calcium-binding domain with four EF-hand structures (shaded area). In both calcineurin and calpain, the EF-hand calcium-binding sites impose calcium-dependency on the enzymes.

domain as well as an EF-hand calcium binding site (Waites et al, 1992). EF-hand structures have also been reported in Drosophila alpha-spectrin (Dubreuil et al., 1991) and the signal-transducing phosphatidylinositol-specific phospholipase C (Bairoch and Cox, 1990). Calcium Binding Proteins in the Nervous System

Calcium is widely used as a second messenger in the nervous system where it regulates axonal transport of substances, release of neurotransmitters, membrane excitability, and long-term potentiation (memory). Therefore, it is not surprising to find many CaBPs in abundant amounts in the nervous system. Parvalbumin,


267

calmodulin, calbindin-D28K, calretinin, and S-lOOa occur in high concentration in subpopulations of neurons (Heizmann and Braun, 1992). S-lOOb is also found in glial cells (astrocytes, microglial cells, and Schwann cells) (Hyden and McEwen, 1966). Epilepsy and ischemia have been linked to overactivation of glutamate receptors which cause excessive amounts of calcium ion to enter the postsynaptic neurons (Heizmann and Braun, 1992). Parvalbumin, calbindin-D28K, and calretinin were suggested to play a role of buffering increased intracellular calcium levels and, therefore, can be considered as an endogenous neuroprotective mechanism. That parvalbumin, calbindin-D28K, and calretinin-immunoreactive neurons are relatively resistant to glutamate-induced neurotoxicity is still a moot point (Heizmann and Braun, 1992). The S-lOOb protein has been shown to regulate phosphorylation of a microtubule associated protein (tau) which controls microtubule assembly and disassembly (Endo and Hidaka, 1983), and to have neuriteforming activity in neuron cultures (Winningham-Major et al., 1989). Recoverin (visionin) is a 23 kDa CaBP (related to S-lOOb) which is part of the phototransduction process (Stryer, 1991). Upon binding calcium ion, recoverin activates guanylyl cyclase to restore the dark state of the system. Calmodulin-dependent kinase II is a major protein component of the postsynaptic membrane. Its activity has been linked to long-term potentiation of neurons, which is believed to be the first event that leads to memory (Malinow et al., 1988). Calcineurin was found in high density in regions of the brain (Steiner et al., 1992) while calpain has been shown to be activated and to degrade spectrin in neurons exposed to excitotoxin glutamate or hypoxia (Siman and Noszek, 1988). Its degradative activity has been suggested to play a role in the eventual neuronal death. In chronic neurodegenerative diseases, such as Alzheimer's disease and Parkinson's disease, abnormal levels of calbindin D28K and S-100b have been reported (see Heizmann and Braun, 1992). In patients with Alzheimer's disease, both calpain and S-lOOb along with tau have been identified as components of senile plaques—âbnormal extracellular protein aggregates (Heizmann and Braun, 1992).

ANNEXINS Annexin represents a family of calcium-binding proteins that are capable of interacting with phospholipids and cell membranes in a Ca^"*"-dependent manner (Crompton et al., 1988). They have been suggested to participate in membrane fusion, exocytosis, and cell signaling pathways. They serve in vivo as substrates for tyrosine kinase and protein kinase C. When it is membrane bound, annexins inhibit membrane lipid degradation by phospholipases, especially phospholipase A2 and the subsequent release of arachidonic acid, which initiates inflammation. Thus, annexins may be antiinflammatory. Through binding to the cell membrane, they inhibit binding of blood coagulation factors to the cell surface. Thus, they are also considered anticoagulants. So far, eight of the annexins have been identified in human and other animals and named annexin I, II, I I I , . . . VIII. Their affinity and

268

KEVIN K.W. WANG

specificity for different phospholipids appears to vary. There are two major forms of annexin: a 36 kDa form which consists of four internal repeats and a 68 kDa form that has eight repeats. An 11 kDa S-100-like protein (pi 1) complex with the 36 kDa annexin II forms a tetrameric complex. Since pi 1 does not have functional calcium-binding loops, the complexing to annexin II is not calcium-dependent. However, the binding is known to involve the pi 1-binding region on annexin II, forming an amphiphilic helix that interacts with pi 1, in a manner similar to the interaction between calmodulin and calmodulin-binding proteins. The crystal structure of annexin V is now known (Huber et al., 1990). Each repeat is composed of five alpha-helices (a-e) connected by short loops. Three amino acids are proposed to coordinate Ca^"*": a conserved glycine and threonine pair in the loop between helix a and b, and an aspartic acid in helix d of the repeat. When phospholipid is bound, an additional coordinate is thought to be provided by the phosphate group. This structure is generally conserved intramolecularly and intermolecularly. Ironically, these calcium binding sites are similar to those of phospholipase A2 which binds and hydrolyzes phospholipids in a calcium-dependent manner. Its calcium-binding site has a pentagonal coordination of a calcium ion: two carboxylate oxygens of ASP49, carbonyl oxygen from Tyr, Gly, and Ala (Tomoo et al., 1992). This five member coordination is similar to, and yet distinct from, the six member coordination seen in EF-hand proteins. Indeed, these findings are an indication of the emergence of a new family of calcium-binding proteins.

CALCIUM TRANSPORTING PROTEINS A number of proteins are involved in transporting Ca^"^ across biological membranes (Figure 6). These proteins have at least one calcium-binding site. The plasma membrane Ca2"'-Mg2"'-ATPase pump (130-140 kDa) (Wang et al., 1992) and the sarcoplasmic/endoplasmic reticulum Ca^"*"-Mg^'*"-ATPase (110 kDa) (MacLennan et al., 1985) are homologous. Both proteins actively transport Ca^"^ over a membrane barrier against the chemical gradient by utilizing energy derived from ATP hydrolysis. The plasma membrane calcium pump is a high affinity (low K^) system for Ca^"^ and maintains low resting levels of cytosolic calcium by transporting Ca^"^ out of the cells. It is especially important in cells that have no Na'*'-Ca^'^ exchanger, such as erythrocytes. On the other hand, the SR calcium pump functions to rapidly remove excess cytosolic Ca^"*" when muscle fibers contract so that they can return to the relaxed state. It has been suggested that there are two Ca^"^ affinity states for the calcium pump: a high affinity site for binding Ca^"*", and one with low affinity for Ca^"*", thus enabling Ca^"*" to be released to the outside. The calcium binding site was located in several transmembrane helices, presumably lining the lumen of the Ca^^ pore. In particular, two glutamic acids (residue 309 and 771 in the SR Ca^'^-Mg^"^-ATPase) are implicated in the binding of calcium. The Na^-Ca^^ exchanger under normal conditions transports two Ca^"^ outward per three Na"^ inward. The exchanger is a high capacity system with a low affinity (high K^) for


269 Extraceitular space

Plasma membrane calcium pump

Voltage-sensitive calcium ctiannel

Cytosol IP3 receptorcalcium channel

Figure 6. Interplay of various calcium transporters in cells Schematic of a cell showing the various membrane-bound proteins involved in transport of Ca^"^ across a cellular membrane. The plasma membrane has both voltage-gated and ligand-gated calcium channels to allow calcium influx and two proteins to extrude calcium: the plasma membrane Ca^'*"-Mg^"^-ATPase and the Na'^-Ca^'^ exchanger. Internally, calcium can be released via the IP3 receptor-calcium channel (in the endoplasmic reticulum membrane) or the calcium release channel (in the sarcoplasmic reticulum In muscle fibers). Excess Ca^"^ is reuptaken Into the Intracellular Ca^"^ store by the inward pumping ER or SR calcium pump. Note that not all of these transporters are expressed in all cell types.

calcium (Blaustein and Nelson, 1982). However, in certain tissues, e.g., heart, it is the major system involved in the maintenance of the resting cytosolic calcium levels. There are many subtypes of voltage-gated calcium channels which allow calcium to enter the cell when the plasma membrane undergoes depolarization. In skeletal and cardiac muscle, it is primarily the L-type channel (Hofmann et al., 1987), while in the CNS, there are also the N-type, P-type, and T-type channels (Spedding and Paoletti, 1992). Presumably, these heteromeric channels have a central hydrophilic core for Ca^"^ to pass through and have certain acidic residues for binding Ca^"^ at or near the entrance of such pores to facilitate the Ca^"^ channeling process.

270

KEVIN K.W. WANG

In skeletal muscle, the SR is interlinked via the T-tubules to the plasma membrane. When T-tubules undergo depolarization, cisternal Ca^"^ is released into the myoplasm via the Ca^"^ release channel (Inui et al., 1987). This 550 kDa protein has been purified and cloned (Zorzato et al., 1990) but its exact calcium binding sites have not yet been clearly defined. Similar calcium release channels have been found in other excitable cells, such as neurons. Another class of proteins involved in calcium influx are the glutamate receptors found in the central nervous system (Watkins et al., 1990). The subtype AMPA-receptor has several isoforms capable of forming functional homo- or heterodimers. One isoform has a glutamine residue in a putative channel-forming region of the protein which appears to be important for Ca^"*" influx. Another isoform which only conducts Na"*" current contains an arginine residue in the same position instead of glutamine. By comparison, the subtype NMDA-receptor, which is involved in both Na"^ and Ca^"*" influx, contains two asparagines in a homologous region (Bumashev et al., 1992).

OTHER CALCIUM-BINDING PROTEINS Protein kinase C exemplifies a group of more obscure calcium-binding proteins with unidentified calcium binding sites. Protein kinase C can be divided into two halves. Its C-terminal half contains an ATP-binding site and a consensus kinase domain, whereas the N-terminal half, which is the regulatory domain, has two cysteine-rich regions which chelate zinc (zinc fingers) (Nishizuka, 1984). Presumably, the phospholipid and calcium binding sites are in the vicinity of the zinc fingers. Limited proteolysis of protein kinase C by trypsin or calpain yields an enzyme that is independent of both calcium and phospholipid as a result of cleavage in the middle of the molecule. This supports the idea that the calcium-binding sites are located in the N-terminal half of the protein. Calsequestrin (42 kDa) is a major SR membrane component which is highly enriched with acidic residues. It binds calcium with relatively low affinity (IQ about 1 mM) but with high capacity (up to 40-50 mol Ca^'*'/mol protein). This CaBP most likely binds Ca^"^ through negative surface charges, and serves as a sink for Ca^"^ in the SR during muscle relaxation (Scott et al., 1988). Transglutaminase, which catalyzes protein cross-linking, is dependent on millimolar calcium concentrations. It apparently contains an unidentified low affinity calcium-binding site (Friedrich and Aszodi, 1993). There are also other calcium-binding sites in proteins, such as in serine proteases and in metalloproteinases (e.g., carboxypeptidase), which do not actively participate in regulating enzymatic activity. This is not unexpected if the binding of Ca^"^ exerts a structure-stabilizing effect (McPhalen et al., 1991).

SUMMARY In this chapter, we have surveyed a number of intracellular calcium-binding proteins. As calcium is such a diverse second messenger, the number of calcium-


271

binding proteins is equally impressive. The most well studied are the EF-hand superfamily and the annexin family. A typical calcium-binding protein uses its affinity for calcium ion as the sensor for a rise in the intracellular calcium level. Upon binding Ca^"^, the protein undergoes certain conformational changes that allow it to interact with its target protein or with another part of the same protein (the effector). Through the modified activity of the effector, the calcium signal is transduced. It is most likely that more Ca^"*"-binding proteins will be found in the fiiture. REFERENCES Andreasen, T.J., Luetje, C.W., Heideman, W., & Storm, D.R. (1983). Purification of a novel calmodulin binding protein from bovine cerebral cortex membranes. Biochemistry 22, 4615-4618. Babu, Y.S., Sack, J.S., Greenhough, T.J., Bugg, C.E., Means, A.R., & Cook, W.J. (1985). Three-dimensional structural of calmodulin. Nature 315, 37-40. Bairoch, A. & Cox, J.A. (1990). EF-hand motifs in inositol phospholipid-specific phospholipase C. FEBS Lett. 269, 454^56. Balasubramanian, D. & Sharma, Y. (1991). Calcium-binding crystallins. In: Novel Calcium-Binding Proteins: Fundamentals and Clinical Implications (Heizmann, C.W., ed.), pp. 361-374, SpringerVerlag, Berlin. Balmain, N. (1991). Calbindin-D9K. A vitamin-D-dependent, calcium-binding protein in mineralized tissues. Clin. Orthop. 265, 265-276. Bennett, V., Gardner, K., & Steiner, J.P. (1988). Brain adducin: A protein kinase C substrate that mediate site-directed assembly at the spectrin-actin junction. J. Biol. Chem. 263, 5860-5869. Berridge, M.J. (1987). Inositol trisphosphate and diacylglycerol: Two interacting second messengers. Ann. Rev. Biochem. 56, 159-193. Blaustein, M.P. & Nelson, M. (1982). Na'*"-Ca^'^ exchange: Its role in the regulation of cell calcium. In: Membrane Transport of Calcium (Carafoli, E., ed.), pp. 217-236, Academic Press, New York. Bumashev, N., Schoepfer, R., Monyer, H., Ruppersberg, J.P., Gunther, W., Seeburg, P.H., & Sakmann, B. (1992). Control by asparagine residues of calcium, permeability and magnesium blockade in the NMDA receptor. Science 257, 1415-1419. Calabretta, B., Battini, R., Kaczmarek, L., de Riel, J.K., & Baserga, R. (1986). Molecular cloning of the cDNA for a growth factor induced gene with strong homology to S-100, a calcium-binding protein. J. Biol. Chem. 261, 12628-12632. Campbell, A.P. & Sykes, B.D. (1991). Interaction of troponin I and troponin C. J. Mol. Biol. 222, 405-421. Carafoli, E. (1987). Intracellular calcium homeostasis. Ann. Rev. Biochem. 56, 395-433. Carlin, R.K., Bartelt, D.C., & Siekevitz, P. (1983). Identification of fodrin as a major calmodulin-binding protein in postsynaptic density preparations. J. Cell Biol. 96, 443-448. Chan, K.-F.J. & Graves, D.J. (1984). Molecular properties of phosphorylase kinase. In: Calcium and Cell Functions (Cheung, W.Y., ed.). Vol. V, pp. 1-31, Academic Press, New York. Cheung, W.Y. (1971). Cyclic 3',5'-nucleotide phosphodiesterase. J. Biol. Chem. 246, 2859-2869. Cohen, P. (1988). The regulation of phosphorylase kinase activity by calmodulin and troponin. In: Calmodulin (Cohen, P. & Klee, C.B., eds.), pp. 123-144, Elsevier, Amsterdam. Crompton, M.R., Moss, S.E., & Crumpton, M.J. (1988). Diversity in the lipocortin/calpactin family. Cell 55, 1-3. Dubreuil, R.R., Brandin, E., Reisberg, J.H., Goldstein, L.S., & Branton, D. (1991). Structure, calmodulin-binding, and calcium-binding properties of recombinant alpha spectrin polypeptides. J. Biol. Chem. 266, 7189^7193.

272

KEVIN K.W. WANG

Endo, T. & Hidaka, H. (1983). Effect of S-100 protein on microtubule assembly-disassembly. FEBS Lett. 161,235-238. Evans, J.S., Levine, B.A., Williams, R.J.P., & Wormald, M.R. (1988). NMR studies of calmodulin in solution: Structure and dynamics in relation to function. In: Calmodulin (Cohen, P. & Klee, C.B., eds.), pp. 57-82, Elsevier, Amsterdam. Friedrich, P. & Aszodi, A. (1993). Calpains and transglutaminases: Common features in structure, mechanism and functions. Indian J. Chem. 32b, 181—185. Guerini, D. & Klee, C.B. (1989). Cloning of human calcineurin A: Evidence for two isozymes and identification of a polyproline strucmral domain. Proc. Natl. Acad. Sci. USA 86, 9183-9187. Guerini, D., FCrinks, M.H., Sikela, J.M., Hahn, W.E., & Klee, C.B. (1989). Isolation and sequence of a cDNA clone for human calcineurin B, the Ca^'*'-binding subunit of the Ca^Vcalmodul in-stimulated protein phosphatase. DNA 8,675-682. Heizmann, C.W. & Hunziker, W. (1991). Intracellular calcium-binding proteins: More sites than insights. Trends Biochem. Sci. 16, 98-103. Heizmann, C.W. & Braun, K. (1992). Changes in Ca^"^-binding proteins in human neurodegenerative disorders. Trends Neuroscience 15,259-264. Hoffman, F., Nastairczyk, W., Rohrkasten, A., Schneider, T., & Sieber, M. (1987). Regulation of the L-type calcium channel. Trends Pharmacol. Sci. 8, 393-398. Huber, R., Schneider, M., Mayr, I. Romisch, J., & Paques, E.-P. (1990). The calcium binding sites in human annexin V by crystal structure analysis at 2.0 A resolution. Implications for membrane binding and calcium channel activity. FEBS Lett. 275, 15-21. Hyden, H. & McEwen, B. (1966). A glial protein specific for the nervous system. Proc. Natl. Acad. Sci. USA 55, 354-358. Inui, M., Saito, A., & Fleischer, S. (1987). Isolation of the ryanodine receptorfromcardiac sarcoplasmic reticulum and identify with the feet structures. J. Biol. Chem. 262, 15637—15642. Johanson, R.A., Hansen, C.A., & Williamson, J.R. (1988). Purification of D-myo-inositol 1,4,5trisphosphate 3-kinasefromrat brain. J. Biol. Chem. 263, 7465-7471. Kennedy, M.B., Bennett, M.K., Erondu, N.E., & Miller, S.G. (1987). Calcium/calmodulin-dependent kinases. In: Calcium and Cell Functions (Cheung, W.Y., ed.). Vol. VII, pp. 61-107, Academic Press, New York. Klee, C.B. (1977). Conformational transition accompanying the binding of Ca^^ to the protein activator of 3',5'-cyclic adenosine monophosphate phosphodiesterase. Biochemistry 16, 1017—1026. Klee, C.B. & Vanaman, T.C. (1982). Calmodulin. Adv. Protein Chem. 357, 213-321. Klee, C.B., Draetta, G.F., & Hubbard, M.J. (1988). Calcineurin. Adv. Enzymol. Relat. Areas Mol. Biol. 61, 149-200. Kligman, D. & Hilt, D.C. (1988). The S-100 protein family. Trends Biochem. Sci. 13, 437-443. Kretsinger, R.H. & Nockolds, C.E. (1973). Carp muscle calcium-binding protein. II. Structure determination and general description. J. Biol. Chem. 248,3313-3326. Kretsinger, R.H., Tolbert, D., Nakayama, S., & Pearson, W. (1991). The EF-hand, homologs and analogs. In: Novel Calcium-Binding Proteins: Fundamentals and Clinical Implications (Heizmann, C.W., ed.), pp. 17-37, Springer-Verlag, Berlin. Kumagai, H., Nishida, E., & Sakai, H. (1982). The interaction between calmodulin and microtubule proteins. IV. Quantitative analysis of the binding between calmodulin and tubulin dimer. J. Biochem. 91,1329-1336. Kyte, J. & Doolittle, R.F. (1982). A simple method for displaying the hydropathic character of a protein. J. Mol. Biol. 157,105-132. Leavis, P.C. & Gergely, J. (1984). Thin filament proteins and thin filament regulation of vertebrate muscle contraction. CRC Crit. Rev. Biochem. 16, 235-305. Lee, Y.C. & Wolff, J. (1984). Calmodulin binds to both microtubule-associated protein 2 and tau proteins. J. Biol. Chem. 259, 1226-1230.


273

Liu, J., Farmer, J.D., Lane, W.S., Friedman, J., Weissman, L, & Schreiber, S.L. (1991). Calcineurin is a common target of cyclophilin-cyclosporin A and FKBP-FK506 complexes. Cell 66, 807-815. Lowenstein, C.J. & Snyder, S.H. (1992). Nitric oxide, a novel biologic messenger. Cell 70, 705-707. MacLennan, D.H., Brandl, C.J., Korczak, B., & Green, M.N. (1985). Amino acid sequence of a Ca^"^+Mg^"^-dependent ATPase from rabbit muscle sarcoplasmic reticulum, deduced from its complementary DNA sequence. Nature 316, 696-700. Malinow, R., Madison, D.V., & Tsien, R.W. (1988). Persistent protein kinase activity underlying long-term potentiation. Nature 335, 820-^24. Manalan, A.S. & Klee, C.B. (1984). Calmodulin. In: Advances in Cyclic Nucleotide and Protein phosphorylation Research (Greengard, P. & Robison, G.A., eds.), Vol. 18, pp. 227-278, Raven Press, New York. Mayr, G.W. & Heilmeyer, L.M.G., Jr. (1983). Phosphofructokinase is a calmodulin binding protein. FEBSLett. 159,51-57. McPhalen, C.A., Strynadka, N.C.J., & James, M.N.G. (1991). Calcium-binding sites in proteins: A structural perspective. Advances Protein Chemistry 42, 77-144. Meyers, M.B., Schneider, K.A., Spengler, B.A., Chang, T.-D., & Biedler, J.L. (1987). Sorcin (V19), a soluble acidic calcium-binding protein overproduced in multidrug resistant cells. Biochem. Pharm. 36, 2373-2380. Nishizuka, Y. (1984). The role of protein kinase C in cell surface signal transduction and tumor promotion. Nature 308, 693-698. Odink, K., Cerletti, N., Bruggen, J., Clerc, R.G., Tarcsay, L., Zwadlo, G., Gerhards, G., Schlegel, R., & Sorg, C. (1987). Two calcium-binding proteins in infiltrate macrophages of rheumatoid arthritis. Nature 330, 80-82. O'Neil, K.T. & DeGrado, W.F. (1990). How calmodulin binds its target proteins: Sequence independent recognition of amphiphilic a-helices. Trends Biochem. Sci. 15, 59-64. Persechini, A., Moncrief, N.D., & Kretsjnger, R.H. (1989). The EF-hand family of calcium-modulated proteins. Trends Neurosciences 12,462-468. Schachtele, C.N. & Marme, D. (1988). Methods of assay of calcium-binding proteins. In: Calcium Binding Proteins (Thompson, M.P., ed.), Vol. I, pp. 83-96, CRC Press, Florida. Scott, B.T., Simmerman, H.K.B., Collins, J.H., Nadal-Ginard, B., & Jones, L.R. (1988). Complete amino acid sequence of canine cardiac calsequestrin deduced from cDNA cloning. J. Biol. Chem. 263, 8958-8964. Seiler, S., Wegener, A.D., Whang, D.D., Hathaway, D.R., & Jones, L.R. (1984). High molecular weight proteins in cardiac and skeletal muscle junctional sarcoplasmic reticulum vesicles bind calmodulin, are phosphorylated, and are degraded by Ca^"*^-activated protease. J. Biol. Chem. 259, 8550-8557. Sharma, Y. & Balasubramanian, D. (1991). Stains-All is a dye that probes the conformational features of calcium binding proteins. In: Novel Calcium-Binding Proteins. Fundamentals and Clinical Implications (Heizmann, C.W., ed.), pp. 52-61, Springer-Verlag, Berlin. Siman, R. & Noszek, J.C. (1988). Excitatory amino acid activate calpain-I and induce structural protein breakdown in vivo. Neurons 1, 279-287. Sobue, K., Fugita, M., Muramoto, Y., & Kakiuchi, S. (1981a). The calmodulin-binding protein in microtubules is tau factor. FEBS Lett. 132, 137-140. Sobue, K., Muramoto, Y., Fugita, M., & Kakiuchi, S. (1981b). Calmodulin-binding protein of erythrocyte cytoskeleton. Biochem. Biophys. Res. Commun. 100, 1063-1070. Spedding, M. & Paoletti, R. (1992). Classification of calcium channels and the sites of action of drugs modifying channel function. Pharmacol. Reviews 44, 363-376. Steiner, J.P., Dawson, T.M., Fotuhi, M., Glatt, C.E., Snowman, A.M., Cohen, N., & Snyder, S.H. (1992). High brain densities of immunophilin FKBP colocalized with calcineurin. Nature 358, 584-586. Stryer, L. (1991). Visual excitation and recovery. J. Biol. Chem. 266, 10711-10714.

274

KEVIN K.W. WANG

Suzuki, K. (1987). Calcium activated neutral protease: Domain structure and activity regulation. Trends Biochem. Sci. 12, 10^105. Tallant, E.A. & Cheung, W.Y. (1986). Calmodulin-dependent protein phosphatase. In: Calcium and Cell Functions (Cheung, W.Y., ed.). Vol. VI, pp. 71-112, Academic Press, New York. Tomoo, K., Ohishi, H., Doi, M., Ishida, T., Inoue, M., Ikeda, K., Hata, Y., & Samejima, Y. (1992). Structure of acidic phospholipase A2 for the venom of Agkiistrodon halys blomhoffii at 2.8 A resolution. Biochem. Biophys. Res. Commun. 184, 137-143. Tuckwell, D.S., Brass, A., & Humphries, M.J. (1992). Homology modelling of integrin EF-hands. Evidence for widespread use of a conserved cation-binding site. Biochem. J. 285, 325-331. Waites, G.T., Graham, I.R., Jackson, P., Millake, D.B., Patel, B., Blanchard, A.D., Weller, P.A., Eperon, I.e., & Critchley, D.R. (1992). Mutually exclusive splicing of calcium-binding domain exons in chick alpha-actinin. J. Biol. Chem. 267, 6263-6271. Wang, K.K.W., Villalobo, A., & Roufogalis, B.D. (1989). Calmodulin-binding proteins as calpain substrates. Biochem. J. 262, 693-706. Wang, K.K.W., Villalobo, A., & Roufogalis, B.D. (1992). The plasma membrane calcium pump: A multi-regulated transporter. Trends Cell Biol. 2,46-52. Wasserman, R.H., Chandler, J.S., Meyer, S.A., Smith, C.A., Brindak, M.E., Fullmer, C.S., Penniston, J.T., & Kumar, R. (1992). Intestinal calcium transport and calcium extrusion processes at the basolateral membrane. J. Nutr. 122, 662-671. Watkins, J., Krogsgaard-Larsen, P., & Honore, T. (1990). Structure-activity relationship in the development of excitatory amino acid receptor agonists and antagonists. Trends Pharmacol. Sci. 11,25-33. Weiss, B., Prozialeck, W.C., & Wallace, T.L. (1982). Interaction of drugs with calmodulin. Biochemical, pharmacological and clinical implications. Biochem. Pharmacol. 31, 2217-2226. Welsh, M.J., Aster, J.C, Ireland, M., Alcala, J., & Maisel, H. (1982). Calmodulin binds to chick lens gap junction protein in a calcium-independent manner. Science 216, 642-644. Winningham-Major, F., Staecker, J.L., Barger, S.W., Coats, S., & van Eldik, L.J. (1989). Neurite extension and neuronal survival activities of recombinant S-1OOB proteins that differ in the content and position of cysteine residues. J. Cell Biol. 109, 3063-3071. Yeager, R.E., Heideman, W., Rosenberg, G.B., «fe Storm, D.R. (1985). Purification of the calmodulinsensitive adenylate cyclase from bovine cerebral cortex. Biochemistry 24, 3776-3783. Young, A.B. & Fagg, G.E. (1991). Excitatory amino acid receptors in the brain: Membrane binding and receptor autoradiographic approaches. Trends Pharmacol. Sci. Special Report 18-24. Zimmer, D.B., Green, C.R., Evans, W.H., & Gilula, N.B. (1987). Topological analysis of the major protein in isolated intact rat liver gap junctions and gap junction-derived single membrane structures. J. Biol. Chem. 262, 7751-7763. Zorzato, F., Fujii, J., Otsu, K., Phillips, M., Green, N.M., Lai, F.A., Meissner, G., & MacLennan, D.H. (1990). Molecular cloning of cDNA encoding human and rabbit forms of the Ca^"^ release channel (ryanodine receptor) of skeletal muscle sarcoplasmic reticulum. J. Biol. Chem. 265, 2244-2256.

RECOMMENDED READINGS Cohen, P. & Klee, C.B. (eds.) (1988). In: Calmodulin. Elsevier, Amsterdam. Crompton, M.R., Moss, S.E., & Crumpton, M.J. (1988). Diversity in the lipocortin/calpactin family. Cell 55, 1-3. Heizmann, C.W. (ed.) (1991). In: Novel Calcium-Binding Proteins: Fundamentals and Clinical Implications. Springer-Verlag, Berlin. Heizmann, C.W. & Hunziker, W. (1991). Intracellular calcium-binding proteins: More sites than insights. Trends Biochem. Sci. 16, 98-103. O'Neil, K.T. & DeGrado, W.F. (1990). How calmodulin binds its target proteins: Sequence independent recognition of amphiphilic a-helices. Trends Biochem. Sci. 15, 59-64. Persechini, A., Moncrief, N.D., & Kretsinger, R.H. (1989). The EF-hand family of calcium-modulated proteins. Trends Neurosciences 12,462-^68.

Chapter 10

ATP-Ubiquitin-Mediated Protein Degradation A.L. HAAS

Introduction The ATP-Ubiquitin-Dependent Proteolytic Pathway Substrates of the Ubiquitin Proteolytic Pathway Potential Medical Relevance Summary

275 277 280 281 282

INTRODUCTION The concentrations of intracellular proteins are defined by the dynamic balance between their respective rates of synthesis and degradation. Regulation of these opposing processes represents an important mechanism for controlling levels of key structural and catalytic proteins. For example, the muscle hypertrophy associated with exercise largely results from a coordinated increase in rate of synthesis and decrease in rate of degradation for myofibrillar proteins; in contrast, the disuse atrophy observed following denervation or whole-limb casting results from accel-


275

276

A.L. HAAS

erated myofibrillar protein degradation. Since protein synthesis requires a considerable expenditure of cell energy, continuous protein degradation appears initially to be an energetically wasteful futile cycle. However, the marked conservation observed across evolution for the elaborate mechanisms responsible for degrading proteins indicates that protein turnover is a fundamental regulatory process essential for cell viability. The turnover of intracellular proteins rigorously follows first order kinetics. This implies that newly synthesized and existing proteins are indistinguishable to the degradative machinery of the cell; therefore, cellular proteins do not "age" but are degraded randomly. The majority of intracellular proteins are relatively stable and exhibit half lives ranging from several days to several weeks. In contrast, a small subpopulation of intracellular proteins possess short half-lives ranging from several minutes to several hours. Correlative studies have shown that protein half-life is a consequence of structural stability with short half-lives favoring high molecular weight, acidic pi (isoelectric point), and marked surface hydrophobicity. Goldberg and St. John (1976) have shown that within metabolic pathways branchpoint and committed step enzymes generally exhibit short half-lives, while the enzymes within pathways are relatively stable. The half-lives of these important regulatory enzymes can be modulated through a number of factors, including covalent modification by phosphorylation/dephosphorylation and binding of substrates or regulatory subunits. Segal and Kim (1965) have provided afiinctionalrationale for the short half-life of regulatory proteins. If a cell is subjected to some hormonal or metabolic signal that alters the level of a given regulatory protein or enzyme, then that protein will undergo an exponential shift from its initial to final steady-state concentration. The time required to affect this alteration in steady state concentration is proportional to protein half-life; therefore, short lived proteins provide a selective advantage in having the capacity to adjust their concentrations rapidly in response to the stimulus. Paradoxically, half-life equally defines the responsiveness of a protein whether it undergoes a net increase or decrease in concentration. While allostery provides short-term modulation, regulation of the opposing steps of protein synthesis and degradation provides for long-term control of metabolic flux through pathways. This provides an attractive mechanism for controlling metabolism since the cell need only regulate the levels of enzymes catalyzing entry into a pathway, while leaving those enzymes within the pathway intact, thus providing considerable conservation in the energy otherwise required to synthesize the entire cohort of proteins. This strategy also operates for noncatalytic regulatory proteins such as transcription factors, which invariably exhibit short half-lives, and for structural proteins, such as myofibrils where the more abundant subunits are long lived while the small subset responsible for stabilizing the quaternary assembly exhibit short half-lives that allow the cell to rapidly regulate assembly/disassembly.

ATP-Ubiquitin-Mediate Protein Degradation

277

The role of protein degradation in modulating metabolic flux by defining the cellular concentrations of key committed step enzymes appears to have arisen by selective pressure throughout evolution. In addition, the protein degradative machinery is required to distinguish between native and structurally abnormal proteins. Since native protein tertiary structure is only stabilized by 5-15 kcal/mole, proteins are continuously subject to spontaneous denaturation within cells. Proteins can also be destabilized in response to environmental stress such as elevated temperature during heat shock or fever and oxidative damage from free radicals. Errors of transcription/translation, genetic mutation, or incorrect cellular trafficking provide additional mechanisms for generating aberrant proteins. Unless these abnormal proteins are removed by degradation, they form insoluble precipitates that interfere with normal function and lead to cell death. Examples of this latter phenomenon are found in a large class of inclusion body diseases (Mayer et al., 1989) of which the neurofibrillary tangles of Alzheimer's disease is the most cited example.

THE ATP-UBIQUITIN-DEPENDENT PROTEOLYTIC PATHWAY In 1953, Simpson first demonstrated an ATP dependence for intracellular protein degradation within rat liver slices. Subsequent work showed this non-lysosomal process to be present across evolution and to be specific for degradation of various classes of short-lived proteins. While ATP-dependent protein degradation could be readily observed within intact cells, the effect was rapidly lost during preparation of cell free extracts from eukaryotes. This instability precluded further biochemical characterization until 1980 when Hershko and coworkers first demonstrated energy-dependent protein degradation within cell-free rabbit reticulocyte lysates (Ciechanover et al., 1980; Hershko et al., 1980). Our understanding of this important degradative pathway has advanced rapidly in the ensuing years. The defining characteristic of this cytosolic degradative pathway is its absolute requirement for the heat stable polypeptide ubiquitin, an 8.6 kDa protein composed of 76 amino acids (Schlesinger et al., 1975; Wilkinson et al, 1980). Ubiquitin is absent from prokaryotes but almost completely conserved in sequence across eukaryotes, indicating the importance of the polypeptide to cell viability among higher organisms. The biological effect of ubiquitin is manifested through a unique posttranslational modification in which the polypeptide is covalently attached to various intracellular target proteins in an ATP-coupled pathway. The linkage formed is an isopeptide bond between the carboxyl terminus of ubiquitin and s-amino groups of lysyl residues present on the target protein (Hershko et al., 1980). Ubiquitin conjugation signals degradation of the target protein by a 26 S (where S stands for Swedberg) (1.2 MDa) multi-catalytic protease complex (Rechsteiner, 1991). During this process free functional ubiquitin is regenerated; therefore,

278

A.L. HAAS

ubiquitin serves a catalytic role in the overall pathway and the energy requirement provides the specificity of degradative targeting. The mechanism for ubiquitin ligation to 8-amino lysyl group is formally that of a ligase; however, unlike most ligases the two half-reactions of isopeptide bond formation are catalyzed by different enzymes (discussed in Haas et al., 1988). Activation of the carboxyl terminus of ubiquitin is catalyzed by the ubiquitin activating enzyme (El), a 105 kDa single-copy gene product whose mechanism is analogous to the activation of the a-carboxyl groups of amino acids by their respective aminoacyl tRNA synthetases (Haas and Rose, 1982). ATP hydrolysis is coupled to formation of a tightly E1 -bound ubiquitin adenylate possessing a highly reactive mixed anhydride between the ubiquitin carboxyl terminus and the a-phosphate of AMP. Ubiquitin adenylate is subsequently transferred to a second site on El where AMP is released during formation of a high energy ubiquitin carboxyl terminal thiolester to an active site cysteine. The resulting covalent ubiquitin-El adduct is analogous to other high energy acyl thiolesters such as acetyl Coenzyme A. The second half-reaction of isopeptide bond formation is catalyzed by one of several ubiquitiniprotein isopeptide ligases (E3) having molecular weights of ca. 180 kDa (Reiss and Hershko, 1990). In this step the energy of ubiquitin thiolester hydrolysis is coupled to the formation of the isopeptide bond to the target protein. The activation and ligation half reactions are functionally linked by a family of low molecular weight isozymes historically termed ubiquitin carrier proteins (E2). Isozymes of E2 range in molecular weightfrom14 to 210 kDa and share a common catalytic domain, defined by the entire 14 kDa form, that binds both El and E3 (Jentsch et al., 1990). The catalytic domain also possesses a conserved cysteine to which ubiquitin is covalently bound as a thiolester during shuttling between El and E3. Many of the E2 isozymes also contain a carboxyl terminal extension forming a second domain presumably to confer functional specificity. The E2 isozymes are bifunctional in catalyzing both E3-dependent and E3-independent conjugation reactions that differ in substrate and product specificity (Haas et al., 1988, 1991). As anticipated, target protein substrate specificity largely resides in the E3 step where various permutations of subsets of E2 and E3 isozymes may provide additional flexibility to substrate and product specificity. The observed substrate specificity for degradative targeting by ubiquitination embodies stringent requirements. Overall, the system must exhibit rather broad specificity in order to account for the wide range of proteins thought to be degraded by this pathway. It is therefore unlikely that all susceptible proteins contain a single common targeting sequence of amino acids. In contrast, the system must be highly specific in order to selectively target certain proteins under conditions in which the cell does not undergo a general upregulation of total protein degradation. Finally, the system is required to distinguish between native and denatured proteins in order to account for the selective degradation of the latter.


279

All three criteria for substrate specificity can be satisfied by invoking steric accessibility of lysines as targets for ubiquitination. This hypothesis is supported by the few instances in which sites of ubiquitination have been mapped to protein targets. Recombinant calmodulin is rapidly degraded in cell-free extracts by ubiquitination at lysine-115, which physical studies reveal to be highly exposed on the protein surface (Gregori et al., 1987). In contrast, mature vertebrate calmodulin isolated from various tissues is extremely stable. This discrepancy is reconciled by reports that lysine-115 is normally methylated on calmodulin from natural sources, a posttranslational modification that blocks possible ubiquitination. A second example illustrates a mechanism for distinguishing native from denatured proteins by the exposure of susceptible lysine residues following structural changes. The targeting of lysozyme for ubiquitination and subsequent degradation requires reduction of a single disulfide bond which allows the carboxyl terminal helix to swing free into solution (Dunten et al., 1991). This conformational change commits lysozyme to degradation by exposing a lysine residue for ubiquitination. Similar folding transitions could potentially result from other covalent modifications, such as phosphorylation, or by structural alteration following binding of regulatory molecules or subunits. An additional mechanism of targeting specificity has emerged from yeast genetic studies by Bachmair and Varshavsky (1989). This work indicates that the identity of the amino terminal residue can target proteins for degradation. Considerable evidence supports this N-end rule hypothesis in which the 20 amino acids are divided into a large set of stabilizing residues and a smaller set of specific destabilizing residues that can be recognized and bound by E3 for ubiquitination. Bradshaw and others have independently shown that the amino termini of nascent polypeptide chains are subject to a variety of competing posttranslational modifications including sequence-specific acetylation, aminopeptidase cleavage, and addition of new amino terminal residues (reviewed in Arfin and Bradshaw, 1988). The net effect of such amino terminal alterations is to limit exposure of destabilizing residues on cytosolic proteins, analogous to the methylation of lysine-115 on calmodulin, as a means of protecting the former from degradative targeting. In contrast, export proteins generally contain destabilizing residues at their amino termini and are protected from ubiquitination by their rapid translocation into the lumen of the endoplasmic reticulum. This discrimination in amino termini between cytosolic and secreted proteins provides an efficient means of degrading those nascent proteins that elude the normally efficient mechanisms of protein trafficking. The N-end rule also provides a means of targeting intracellular proteins for ubiquitination in response to external signals, providing there has been selective pressure to retain specific sequences that become exposed and cleaved by cytoplasmic endopeptidases to generate new destabilizing amino terminal residues or sites for addition of destabilizing residues by arginyl-tRNA transferases (Elias and Ciechanover, 1990).

280

A.L. HAAS

Ubiquitin conjugates are specifically degraded by an ATP-dependent 26 S multicatalytic protease complex (Rechsteiner et al., 1993). The 26 S complex has also been shown to generate peptides for antigen presentation (Goldberg and Rock, 1992). The core of this complex is the 20 S proteasome, a cylindrical assembly of four stacked rings each of which is composed of six subunits (Eytan et al., 1989). The 20 S complex exhibits multiple protease activities that degrade proteins to free amino acids and small peptides, but shows no specificity for ubiquitin conjugates nor a requirement for ATP. The role of ATP hydrolysis in the action of the 26 S complex is uncertain at present but may be required for release of peptide products or assembly of the complete complex. Additional auxiliary subunits form a cap structure on the 20 S cylinder and confer both specificity for recognition of ubiquitin conjugates and the ATP requirement. In addition, other subunits possess isopeptidase activity required to cleave ubiquitin from the conjugates (Hadari et al., 1992) and ubiquitin carboxyl terminal hydrolase to remove the final lysyl residue from ubiquitin (Eytan et al., 1993).

SUBSTRATES OF THE UBIQUITIN PROTEOLYTIC PATHWAY Under normal conditions of nitrogen balance, the total ubiquitin concentration within cells ranges from 10-20 [xM while the fraction of total ubiquitin present as conjugates is a characteristic of cell type and ranges from 25-80% (Haas, 1988). Approximately 0.01% of total cytosolic protein is present as ubiquitin conjugates that partition between two opposing fates (Haas, 1988). Conjugates can be degraded by the 26 S complex or the ubiquitin can be removed without degradation by soluble isopeptidases in a process termed disassembly. Other work demonstrates that a subset of conjugates within the cytosol is stable to degradation; therefore, ubiquitination alone is' not sufficient to target proteins for destruction. Partitioning of conjugates to degradation requires formation on the target protein of ubiquitin chains ("trees") linked together by isopeptide bonds to lysine-48 of ubiquitin (Chau et al., 1989). Ubiquitin chain formation is catalyzed by several E2 isozymes, either alone or with the participation of E3 (Haas et al., 1988). Target proteins containing ubiquitin chains are degraded many fold faster by the 26 S complex than are conjugates containing single ubiquitin residues, indicating that a capping subunit on the protease complex specifically binds these extended structures. Therefore, target protein specificity with respect to degradation resides in both the initial conjugation event and in the subsequent ability to form ubiquitin chains. Considered in their totality, these results suggest that the proteolytic pathway constantly samples the entire pool of cytosolic proteins by conjugation of single ubiquitin moieties and then commits aberrant target proteins to destruction by chain formation. The factor(s) required for chain formation are presently unknown, but may involve the same features of steric accessibility necessary for the initial conjugation of ubiquitin.


281

Two mechanisms exist to increase the concentration of conjugates and the rate of degradation through the pathway (reviewed in Haas, 1988). Under general enhanced conjugation the total pool of ubiquitin adducts increases, examples of which can be found in the response of cells to radiation, thermal, or chemical stress. General enhanced conjugation is also a feature of the increased degradation that accompanies the terminal differentiation of erythroid cells and in programmed cell death during development. General enhanced conjugation does not result from changes in either total ubiquitin or ATP concentrations since both are always saturating with respect to El (Haas and Rose, 1982); rather, increased conjugation appears to result from an upregulation in the enzymes responsible for ubiquitin ligation. In contrast, under specific enhanced conjugation, the total pool of ubiquitin conjugates remains constant, while specific proteins undergo increased rates of conjugation and subsequent degradation. A number of examples of specific enhanced conjugation have been reported (reviewed in Rechsteiner, 1991). Phytochrome is a regulatory protein utilized by plants to sense light. Absorption of light at 670 nm induces a conformational change in phytochrome that leads to conjugation and subsequent degradation. Loss of phytochrome initiates a regulatory cascade producing the physiological response of plants to daylight. Similarly, cyclins are regulatory proteins that drive mitosis. They are a class of regulatory proteins that accumulate during Gl and S phases but are rapidly degraded by the ubiquitin system during metaphase to allow mitotic progression. Cells become mitotically arrested if the enhanced ubiquitination of cyclins is blocked. Genetic analysis has shown that cyclins possess class-specific targeting sequences required for ubiquitination. The tumor suppressor gene p53 is also subject to ubiquitin-dependent degradation. Several oncogenic viruses appear to transform cells by encoding viral proteins, such as the E6 protein of papillomavirus, that promotes the ubiquitination and destruction of p53. Several other cellular regulatory proteins are also degraded by the ubiquitin system including the MAT a2 repressor of yeast, myc, fos, and El A. It may be a general feature of eukaryotic cells that short-lived regulatory proteins are targeted for degradation by the ubiquitin pathway.

POTENTIAL MEDICAL RELEVANCE To date, no disease state has been demonstrated to result from an aberration within the ubiquitin-dependent degradative pathway. However, the central importance of protein turnover to cellular regulation and the types of proteins shown subject to degradation by ubiquitin conjugation suggests such evidence will eventually emerge. The ubiquitin system participates in the stress response and therefore contributes to the ability of cells to recover from fever, chemical damage from certain drugs, therapeutic radiation, and UV exposure from sunlight. Ubiquitination may also contribute to cell transformation and tumorogenesis through alterations in the turnover of p53 or other regulatory proteins. Since the ubiquitin pathway

282

A.L. HAAS

appears to function in the cellular remodeling accompanying development, inheritable disorders leading to birth defects are reasonable subjects of further study. The pathway may be involved in muscle atrophy since general enhanced conjugation parallels myofibrillar protein degradation in several experimental models. Finally, because mitotic progression requires ubiquitin-mediated degradation, cell senescence during aging may result from loss in ability to carry out specific enhanced conjugation. The only disease indirectly associated with the ubiquitin system is that of the general class of inclusion body diseases for which ubiquitin is conjugated to the intracellular protein precipitates (Mayer et al., 1989). In these cases, ubiquitin conjugation is not the agent for onset of the disease but rather is a normal cellular response to formation of the inclusions. However, the association of ubiquitin with these structures provides a common, specific immunohistochemical marker for diagnosis.

SUMMARY Protein degradation is a fundamental regulatory process within cells for which short half-life favors rapid alteration of protein levels in response to metabolic or hormonal signals. Protein degradation also serves as a structural proofreading mechanism for degrading abnormal proteins arising from spontaneous denaturation, physical damage, mutation, or errors of transcription/translation. Short-lived proteins are targeted for destruction by a novel posttranslational modification in which ubiquitin is covalently ligated to surface lysines on the proteins. Ubiquitin conjugates are subsequently degraded by a 26 S multicatalytic protease complex to regenerate free functional ubiquitin and small peptides from the target protein. The ubiquitin system is required in the response of cells to various forms of stress. Ubiquitin conjugation also participates in the specific degradation of regulatory molecules, including those responsible for mitotic progression, repression of cell transformation, and gene expression. Defects in the enzymes catalyzing ubiquitin ligation, the proteases required for conjugal degradation, or in the specific targeting sequences present on substrate proteins may have the potential of leading to certain disease states.

REFERENCES Arfin, S.M. & Bradshaw, R.A. (1988). Cotranslational processing and protein turnover in eukaryotic cells. Biochemistry 27, 7979-7984. Bachmair, A. & Varshavsky, A. (1989). The degradation signal in a short-lived protein. Cell 56, 1019-1032. Chau, v., Tobias, J.W., Bachmair, A., Marriott, D., Ecker, D.J., Gonda, D.K., & Varshavsky, A. (1989). A multiubiquitin chain is confined to specific lysine in a targeted short-lived protein. Science 243, 1576-1583. Ciechanover, A., Heller, H., Elias, S., Haas, A.L., & Hershko, H. (1980). ATP-dependent conjugation of reticulocyte proteins with the polypeptide required for protein degradation. Proc. Natl. Acad. Sci.USA77, 1365-1368.

A TP-Ubiquitin-Mediate Protein Degradation

283

Dunten, R.L., Cohen, R.E., Gregori, L., & Chau, V. (1991). Specific disulfide cleavage is required for ubiquitin conjugation and degradation of lysozyme. J. Biol. Chem. 266, 3260-3267. Elias, S. & Ciechanover, A. (1990). Post-translational addition of an arginine moiety to acidic NH2 termini of proteins is required for their recognition by ubiquitin-protein ligase. J. Biol. Chem. 265, 15511-15517. Eytan, E., Ganoth, D., Armon, T., & Hershko, A. (1989). ATP-dependent incorporation of 20 S protease into the 26 S complex that degrades proteins conjugated to ubiquitin. Proc. Natl. Acad. Sci. USA 86,7751-7755. Eytan, E., Armon, T., Heller, H., Beck, S., & Hershko, A. (1993). Ubiquitin C-terminal hydrolase activity associated with the 26 S protease complex. J. Biol. Chem. 268, 4668-4674. Goldberg, A.L. & St. John, A.C. (1976). Intracellular protein degradation in mammalian and bacterial cells. Ann. Rev. Biochem. 45, 747-803. Goldberg, A.L. & Rock, K.L. (1992). Proteolysis, proteasomes and antigen presentation. Nature 357, 375-379. Gregori, L., Marriott, D., Putkey, J.A., Means. A.R., & Chau, V. (1987). Bacterially synthesized vertebrate calmodulin is a specific substrate for ubiquitination. J. Biol. Chem. 262, 2562-2567. Haas, A.L. (1988). Immunochemical probes of ubiquitin pool dynamics. In: Ubiquitin (Rechsteiner, M., ed.), pp. 173-206, Plenum Press, New York. Haas, A.L., Bright, P.M., & Jackson, V.E. (1988). Functional diversity among putative E2 isozymes in the mechanism of ubiquitin-histone ligation. J. Biol. Chem. 263, 13268-13275. Haas, A.L., Reback, P.B., & Chau, V.J. (1991). Ubiquitin conjugation by the RAD6 and CDC34 gene products. Comparison to their putative rabbit homologs, E22OK ^^^ ^hiK- J- î^^- Chem. 266, 5104-5112. Haas, A.L. & Rose, LA. (1982). The mechanism of ubiquitin activating enzyme: A kinetic and equilibrium analysis. J. Biol. Chem. 257, 10329-10337. Hadari, T., Warms, J.V., Rose, I.A., & Hershko, A. (1992). A ubiquitin C-terminal isopeptidase that acts on polyubiquitin chains. J. Biol. Chem. 267, 719-727. Hershko, A., Ciechanover, A., Heller, H., Haas, A.L., & Rose, LA. (1980). Proposed role of ATP in protein breakdown: Conjugation of proteins with multiple chains of the polypeptide of ATP-dependent proteolysis. Proc. Natl. Acad. Sci. USA 77, 1783-1786. Jentsch, S., Seufert, W., Sommer, T., & Reins, H.A. (1990). Ubiquitin-conjugating enzymes: Novel regulators of eukaryotic cells. Trends Biochem. Sci. 15, 195-198. Mayer, R.J., Lowe, J., Lennox, G., Doherty, F., & Landon, M. (1989). Intermediate filaments and ubiquitin: A new thread in the understanding of chronic neurodegenerative/diseases. Prog. Clin. Biol. Res. 317, 809-818. Rechsteiner, M. (1991). Natural substrates of the ubiquitin proteolytic pathway. Cell 66, 615-618. Rechsteiner, M., Hoffman, L., & Dubiel, W. (1993). The multicatalytlc and 26 S proteases. J. Biol. Chem. 268, 6065-6068. Reiss, Y. & Hershko, A. (1990). Affinity purification of ubiquitin-protein ligase on immobilized protein substrates. Evidence for the existence of separate NH2-cerminal binding sites. J. Biol. Chem. 265, 3685-3690. Schlesinger, D.H., Goldstein, G., & Niall, H.D. (1975). The complete amino acid sequence of ubiquitin, an adenylate cyclase stimulating polypeptide probably universal in living cells. Biochemistry 14, 2214-2218. Segal, H.L. & Kim, Y.S. (1965). Environmental control of enzyme synthesis and breakdown. J. Cell. Comp. Physiol. Suppl. 66, 11-22. Simpson, M.V. (1953). The release of labeled amino acidsfromthe proteins of rat liver slices. J. Biol. Chem. 201, 143-154. Wilkinson, K.D., Urban, M.K., & Haas, A.L. (1980). Ubiquitin is the ATP-dependent proteolysis factor I of rabbit reticulocytes. J. Biol. Chem. 255, 7529-7532.

284

A.L. HAAS

RECOMMENDED READINGS Band, V., De Caprio, J.A., Delmolino, L., Kulesa, V., & Sager, R. (1991). Loss of p53 protein in human papillomavirus type 16 E6-immortalized human mammary epithlial cells. J. Virol. 65,6671-6676. Chen, P., Johnson, P., Sommer, T., Jentsch, S., & Hochstrasser, M. (1993). Multiple ubiquitin-conjugating enzymes participate in the in vivo degradation of the yeast MAT a2 repressor. Cell 74, 357-369. Hershko, A. & Ciechanover, A. (1992). The ubiquitin system for protein degradation. Ann. Rev. Biochem. 61,761-807. Loeb, K.R. & Haas, A.L. (1992). The interferon-induced 15 kDa ubiquitin homolog conjugates to intracellular proteins. J. Biol. Chem. 267, 7806-7813.

Chapter 11

Regulation of Cellular Functions by Extracellular Calcium EDWARD F. NEMETH

Introduction Regulation of Systemic Ca Metabolism The Parathyroid Cell The C-Cell The Osteoclast Other Extracellular Ca^'*'-Sensing Cells Therapeutic Significance of Extracellular Ca "^ Receptors Summary

285 286 287 294 295 297 300 300

INTRODUCTION A biological phenomenon of increasing recognition is the peculiar ability of extracellular Ca^"^ to regulate the activity of certain specialized cells in the body. While most cells are insensitive to physiological changes in the level of Ca^"^ in the plasma or extracellular fluids, there are a variety of different cell types that can alter their behavior in response to changes in the extracellular Ca^^ concentration. Not


285

286

EDWARD F. NEMETH

surprisingly, many of these cells are involved in the regulation of systemic Ca^"*" homeostasis. Notable among these is the parathyroid cell, whose secretory product, parathyroid hormone, plays a major role in regulating the level of plasma Ca^"^. Other cells sensitive to changes in the concentration of extracellular Ca^"^ are the parafollicular cells of the thyroid that secrete calcitonin and osteoclasts in the skeleton that resorb bone. Certain cells in the kidney, the gastrointestinal tract, the skin, and placental tissue also seem to be responsive to changes in the concentration of extracellular Ca^"^. In fact, for the parathyroid cell, extracellular Ca^"*" is the primary physiological stimulus regulating cellular function. The growing appreciation of the array of different cell types capable of sensing changes in the level of extracellular Ca^"^ has led to the concept that Ca^"^ can function as an extracellular signal, not unlike a hormone or neurotransmitter. This view complements the well-known messenger role of intracellular Ca^"^. Thus, just as intracellular Ca^"^ functions to control a variety of cellular functions as diverse as muscle contraction and cellular secretion, so too does extracellular Ca^^ function to regulate the activity of certain cells in the body. The action of extracellular Ca^"*" on some of these cells involves interaction with a cell surface Ca^"^ receptor protein which is coupled to effector mechanisms that regulate intracellular signals such as Ca^"^, diacylglycerol, and cyclic AMP. Extracellular Ca^"*" receptors are therefore functionally and mechanistically akin to more conventional membrane receptors that initially transduce changes in the concentration of an extracellular ligand into intracellular signals that regulate functional cellular responses. The difference is that the ligand for Ca^"*" receptors is an inorganic ion rather than an organic molecule or protein. This chapter will summarize the data suggesting a messenger role for extracellular Ca^"*" in regulating the activity of diverse cell types. Although all the cells discussed herein have been shown to respond to changes in the concentration of extracellular Ca^"^, the physiological significance of this response is not always obvious. By far the clearest understanding of the molecular events that enable a cell to detect and respond to extracellular Ca^"^, and its physiological significance, derives from studies of cells involved in the regulation of systemic Ca^"^ metabolism, especially parathyroid cells.

REGULATION OF SYSTEMIC Ca^^ METABOLISM Just as intracellular Ca^"^ functions to regulate a variety of cellular responses, so too does extracellular Ca^"*" function to control a variety of life-sustaining functions. Extracellular levels of Ca^"^ are important in maintaining the excitabili^ of nerve and muscle, in permitting thrombosis and cellular adhesion in general, and in proper bone formation. Because of this, the concentration of Ca^"*" in the plasma and extracellular fluids is under tight homeostatic control. In mammals, the level of Ca^"*" in the plasma and exfracellular fluids accounts for only a small percentage (about 0.1%) of the total body systemic calcium content, with

Regulation of Cellular Functions

287

the bulk (99%) stored in the teeth and bones. In humans, the concentration of total calcium in the plasma is 2.4 mM, but only about half of this (1.3 mM) is free, ionized calcium (Ca^"*"). Calcium binds to serum proteins (mostly albumin and globulins) and to various inorganic anions (mostly phosphate and citrate) and in this bound form, calcium is generally considered to be biologically inert. It is the concentration of ionized calcium in the plasma that regulates physiological responses and is the relevant variable sensed by the homeostatic control mechanism. The predominant control mechanism is endocrine and the principal factors regulating the level of plasma Ca^"^ are parathyroid hormone (PTH) and 1,25-dihydroxyvitamin D3. PTH, secreted by cells in the parathyroid gland, guards against hypocalcemia. PTH acts to increase the movement of Ca^"^ from bone to the circulation, and it additionally acts on the kidney to increase distal tubular calcium resorption and proximal tubular synthesis of 1,25-dihydroxyvitamin D3; the latter increases intestinal absorption of Ca^"^. All these actions tend to increase the level of Ca^"^ in the plasma. Increased circulating levels of Ca^"^, in turn, act in a negative feedback capacity to depress secretion of PTH. There is, therefore, a reciprocal relationship between the levels of plasma PTH and Ca^"^, and this simple yet elegant feedback loop is the principal mechanism regulating the level of plasma Ca^"^ (Mundy, 1989). In some species, an additional endocrine factor seems to play an important role in regulating plasma Ca^"^ homeostasis. This is the hormone calcitonin, secreted by parafollicular cells present throughout the thyroid gland. Like PTH, the secretion of calcitonin is regulated by changes in the level of plasma Ca^"^. The difference is that increasing the concentration of extracellular Ca^^ stimulates calcitonin secretion, whereas it inhibits PTH secretion. One site of action by calcitonin is in the kidneys where it stimulates excretion of Ca^"^. The predominant site of action of calcitonin, however, is in bone where it acts to inhibit ongoing osteoclastic bone resorption. This latter action causes a rapid inhibition of Ca^^ flux from bone into the circulation and this results in hypocalcemia. The physiological significance of this effect of calcitonin in adult humans is generally believed to be minor. Nonetheless, calcitonin can be used in pharmacological doses to inhibit bone resorption and is one treatment for bone diseases involving increased bone turnover, such as osteoporosis.

THE PARATHYROID CELL This is the classic cell type long known to be responsive to physiological changes in the concentration of plasma Ca^"^. Perhaps because PTH plays such a crucial role in regulating the level of plasma Ca^"^, its secretion is most responsive to the ambient Ca^^ concentration. In humans and some other species, PTH secretion can be increased by P-adrenergic receptor agonists, but the physiological significance of this is probably minor. The parathyroid glands do not receive significant neural input and, under physiological conditions, PTH secretion is not affected by a wide variety of neurotransmitters, hormones, or other extracellular signaling molecules

288

EDWARD F. NEMETH

(in contrast to calcitonin secretion). It seems safe to say that extracellular Ca^"*" is the primary physiological stimulus regulating PTH secretion. The sensitivity of the parathyroid cell to the ambient Ca^^ concentration is remarkable: minimal and maximal rates of PTH secretion are obtained over a concentration range spanning only 1.5 mM. Significantly, the concentration of extracellular Ca^"^ causing halfmaximal inhibition of PTH secretion or the "set-point" for extracellular Ca^"^, is set precisely near normocalcemic levels (1.3 mM). Moreover, small changes in the level of extracellular Ca^"*" cause rapid (< 1 minute) changes in the rate of secretion of PTH (Brown et al., 1987). Thus, the parathyroid cell is exquisitely constructed to sense and rapidly respond to small, physiological changes in the concentration of extracellular Ca^"*". There have, therefore, been two distinct but related problems in understanding the cellular physiology of parathyroid cells: how do these cells detect such small changes in the concentration of extracellular Ca^"^ and how is this initial recognition event transduced into intracellular signals that regulate PTH secretion? Since the depressive effects of extracellular Ca^"*" on PTH secretion are observed in vitro using dissociated parathyroid cells, it is clear that extracellular Ca^"^ acts directly on parathyroid cells to regulate hormone secretion. While this has been known for many years, it is only quite recently that we have gained some insight into the molecular mechanisms used by parathyroid cells to sense extracellular Ca^"^ levels and thereby regulate PTH secretion. Studies undertaken during the 1970s and early 1980s using dissociated bovine and porcine parathyroid cells demonstrated that agents that cause increases in the levels of cyclic AMP stimulate PTH secretion (Brown et al., 1987). These agents included P-adrenergic agonists, dopamine, prostaglandin E2, and cholera toxin. In contrast, agents that decrease cellular cyclic AMP levels, such as a-adrenergic agonists and prostaglandin F2a, inhibit PTH secretion. Additional studies have suggested that cyclic AMP and extracellular Ca^"*" may regulate secretion of PTH from different intracellular pools: cyclic AMP regulates secretion from a storage pool, whereas extracellular Ca^"^ regulates secretion of PTH from a newly synthesized pool (Watson and Hanley, 1993). It is significant, however, that the magnitude of these responses (cyclic AMP levels and PTH secretion) are dependent on the concentration of extracellular Ca^"*" and increased levels of extracellular Ca^"*" block agonist-induced increases in cyclic AMP and PTH secretion. Moreover, extracellular Ca^"^ alone, while causing large changes in the secretion of PTH, causes relatively small changes in basal levels of cyclic AMP and does not alter the pattern of protein phosphorylation induced by cyclic AMP. Thus, there is an additional mechanism(s) used by extracellular Ca^"^ that can regulate PTH secretion independently of changes in cyclic AMP levels. There is considerable interest in the role cytosolic Ca^"^ may play in the regulation of PTH secretion. Increasing the concentration of extracellular Ca^^ evokes corresponding increases in the concentration of cytoplasmic Ca^"^ ([Ca^"*"]i)


289

and these are associated with an inhibition of PTH secretion (Shoback et al., 1984). The inverse relationship between [Ca^"^]! and secretion is yet another peculiar aspect of parathyroid cell physiology. In most cells, increasing [Ca^"^]i evokes a stimulation of secretion. This general finding has led to the Ca^"^ hypothesis of stimulussecretion coupling which holds that cytosolic Ca^"^ activates or permits exocytotic secretion in diverse cell types (Douglas, 1974). In parathyroid cells, and in some other cells that appear to sense the ambient level of extracellular Ca^"^ (discussed below), cytoplasmic Ca^"^ appears to inhibit secretion. However, the exact role of cytoplasmic Ca^^ in controlling PTH secretion is far from clear. Studies in permeabilized parathyroid cells, in which Ca^"^ has direct access to the exocytotic machinery, have reported either no effect or a stimulation of PTH secretion when exposed to low levels of Ca^"^ that occur within the cell. Additionally, there is data from intact cells suggesting that cytosolic Ca^"^ can have both stimulatory (at low [Ca^"^]!) and inhibitory (at higher [Ca^î) effects on PTH secretion (for review see Brown, 1991). Thus, there are numerous pieces of data suggesting some important signaling role for C3ôplasmic Ca^"^ in parathyroid cells, but the data are often discrepant and no explanatory model has yet emerged. Relatively more progress has been made in understanding the initial steps in stimulus-secretion coupling, namely, how parathyroid cells sense a change in the ambient Ca^"^ concentration and how this detection event is coupled to the regulation of intracellular signals. That extracellular Ca^^ might act through some receptor-like mechanism was initially suggested in 1983 based on electrophysiological measurements (LopezBameo and Armstrong, 1983). Measurements of [Ca^"^]i, however, provided more substantial evidence for an extracellular Ca^"^ receptor and led to a series of biochemical studies consistent with this notion. It was demonstrated that increases in [Ca^"^]i elicited by extracellular Ca^"^ arise from two mechanistically distinct events: the mobilization of intracellular Ca^"^ from a nonmitochondrial pool and the influx of extracellular Ca^"^ through voltage-insensitive channels (Nemeth and Scarpa, 1986,1987a). Moreover, a variety of extracellular di- and trivalent cations were all capable of causing the mobilization of intracellular Ca^"^ in parathyroid cells. Because trivalent cations are impermeant in parathyroid cells, and in cells generally, they must be acting at the cell surface to evoke the mobilization of intracellular Ca^^. Studies using monoclonal antibodies generated against parathyroid cells likewise suggested an action of extracellular Ca^"^ at the cell surface (Gylfe et al., 1990). Together, these results suggested the presence of a Ca^"^ receptor on the surface of parathyroid cells that is coupled to the mobilization of intracellular Ca^"^. Subsequent biochemical studies showed that increased levels of extracellular Ca^"^ evoked rapid increases in the formation of inositol 1,4,5-trisphosphate and diacylglycerol (Brown et al., 1990), two biochemical hallmarks of receptordependent mobilization of intracellular Ca^"^ in various other cells (Berridge, 1987). A significant piece of information was the finding that the inhibitory effects

290

EDWARD F. NEMETH

of extracellular Ca^"^ on cyclic AMP levels is blocked by pertussis toxin (Chen et al, 1989). This observation demonstrated that a Gj-like protein coupled the action of extracellular Ca^"^ at the cell surface to the regulation of adenylate cyclase activity. Such heterotrimeric G-proteins are used to link certain kinds of cell surface receptors to effector mechanisms in diverse cell types (Oilman, 1987). The data obtained with pertussis toxin suggest that the Ca^^ receptor is mechanistically akin to more conventional membrane receptors and is linked to adenylate cyclase by a Gj-like protein. In the aggregate, the results derived from physiological, biochemical, and immunological experiments are complementary and together provide strong evidence for the presence of a cell surface Ca^"^ receptor on parathyroid cells. These studies anticipated the recent functional expression, cloning, and sequencing of the parathyroid cell Ca^"*" receptor (Brown et al., 1993; Racke et al., 1993). Based on the accumulated evidence derived from these various studies, a model of how extracellular Ca^"*" acts on the parathyroid cell to regulate PTH secretion can be formulated (Figure 1). The model reflects to some degree the bias of the author but does incorporate and assemble in a testable manner nearly all the reproducible results obtained in parathyroid cells. On the surface of parathyroid cells is a Ca^"*" receptor protein that enables these cells to detect and respond to small, physiological changes in the concentration of extracellular Ca^"^. The cloning and sequencing of this receptor (Brown et al., 1993) shows that it is a member of the O-protein

Inhibition of PTH secretion Figure 1. Schematic representation of the receptor-dependent regulation of parathyroid cell function by extracellular Ca^"^. Increases in the concentration of extracellular Ca^"^ activate a cell surface Ca^"^ receptor which is linked, by G-proteins, to the inhibition of adenylate cyclase and stimulation of phospholipase C. The net result of Ca^"^ receptor activation is an increase in [Ca'^'*']^ which results from the mobilization of intracellular Ca^"^ and influx of extracellular Ca^"*^ through voltage-insensitive channels. Receptor activation is coupled to the inhibition of PTH secretion. AC, adenylate cyclase; PLC, phospholipase C; IP3, inositol 1,4,5-trisphosphate.


291

receptor superfamily, since the encoded polypeptide exhibits the classic seven transmembrane domain motif common to all such receptors (Strosberg, 1991). The bovine parathyroid cell Ca^"^ receptor is thus structurally homologous to other cell surface receptor proteins that initially transduce extracellular signals into cellular responses. The bovine and human (Garrett et al., 1995a) parathyroid cell Ca^"^ receptors are structurally homologous. Their molecular weight is about 120 kDa and they possess nine (eleven in human) potential glycosylation sites located on the putative extracellular domain. Although the parathyroid cell Ca^"^ receptor is rather large compared to many G-protein-coupled receptors, it is as large as one other subfamily. It turns out that the parathyroid cell Ca^"^ receptor exhibits a 25 to 35% sequence homology with metabotropic glutamate receptors. Glutamate is the principal excitatory neurotransmitter in the central nervous system and these metabotropic subtypes of glutamate receptors are coupled to the mobilization of intracellular Ca^"^ or inhibition of adenylyl cyclase (Schoepp et al., 1990), similarly to the parathyroid cell Ca^"^ receptor. Since the parathyroid cell Ca^"*" receptor responds to physiological changes in the levels of circulating Ca^^ (1-2 mM), it is not surprising that this receptor contains no EF hand domains characteristic of high-affmity Ca^'^-binding proteins like calmodulin (Heizmann and Hunziker, 1991). However, there are proteins that are known to bind Ca^"^ with low affinity. These proteins, such as calsequestrin and calreticulin, are present in the sarcoplasmic reticulum and endoplasmic reticulum, subcellular structures known to serve as intracellular reservoirs for Ca^"^ (Milner et al., 1992). These proteins contain highly acidic regions, especially runs of three or more acidic amino acid residues, which are thought to be responsible for low affinity Ca^"^ binding. The parathyroid cell Ca^"^ receptor contains three regions that are rich in acidic amino acids and these regions are on the putative extracellular portion of the receptor. Studies using chimeric receptor constructs have shown that the extracellular domain is necessary for activation of the receptor by extracellular Ca^-' (Hammerland et al., 1995). The Ca^"^ receptor is coupled to phospholipase C which breaks down inositol phospholipids to form inositol 1,4,5-trisphosphate and diacylglycerol. The former mobilizes intracellular Ca^"^ and the latter activates protein kinase C. It is believed that the Ca^"^ receptor is coupled to phospholipase C by a G-protein. Pertussis toxin does not affect the ability of extracellular Ca^"*" to increase inositol 1,4,5-trisphosphate levels, mobilize intracellular Ca^"*", or inhibit PTH secretion, so this putative G-protein is not Gj. The coupling G-protein might be related to Gq, as this G-protein couples a variety of receptors to the mobilization of intracellular Ca^'^'in other cells. Despite these uncertainties, there are already indications suggesting that the parathyroid cell Ca^"^ receptor uses conventional transmembrane signaling mechanisms to regulate intracellular messengers.

292

EDWARD F. NEMETH

Very little is known about how the Ca^"^ receptor couples to the influx of extracellular Ca^"^. Based on structural comparisons and functional expression studies, the parathyroid cell Ca^"*" receptor does not appear to function as a channel. The influx channel is apparently a distinct protein that is coupled to the Ca^^ receptor either through a G-protein, an intracellular signal(s), or some combination. Parathyroid cells are not electrically excitable and depolarization of the cells does not evoke an increase in [Ca^"^];, suggesting that parathyroid cells do not possess voltage-sensitive Ca^"*" channels. Electrophysiological studies likewise fail to reveal voltage-sensitive Ca^"*^ influx in parathyroid cells. The influx pathway in parathyroid cells is therefore akin to those voltage-insensitive, receptor-operated cation channels observed in various other cells (Nemeth, 1990). The two mechanisms for increasing [Ca^"^]i, mobilization of intracellular Ca^"^ and influx of extracellular Ca^"^, appear to have different functional roles in parathyroid cell physiology. Thus, various extracellular cations that do not promote Ca^"^ influx still inhibit PTH secretion, and secretion is not greatly affected when influx is blocked. It is primarily the mobilization of intracellular Ca^"^, rather than the influx of extracellular Ca^"*", that is associated with the regulation of PTH secretion. This does not necessarily negate a role for Ca^"^ influx in the more global process of stimulus-secretion coupling in parathyroid cells. Maintained hypercalcemic states lasting more than 30 minutes are associated with increased intracellular degradation of PTH, and the relative secretion of PTH fragments of intact hormone is increased at elevated levels of extracellular Ca^"^ (Cohn and MacGregor, 1981). It should be noted that [Ca^"^]i remains high as long as extracellular Ca^"^ remains elevated and it can be promptly decreased by blocking Ca^"^ influx, implying a constant rate of Ca^"*" cycling between cellular and extracellular compartments. The maintained elevation of [Ca^'^Jj under hypercalcemic conditions may be casually involved in regulating intracellular proteolysis of PTH. As noted above, the relationship between cytosolic Ca^"^ and PTH secretion is not clear-cut. There are experimental situations in which PTH secretion can be regulated by extracellular cations independently of changes in [Ca^"^]i. Such findings have led to the proposal that it is Ca^"*" receptor activation, rather than the associated increases in [Ca^"*"]}, that is the critical event regulating PTH secretion (Nemeth and Scarpa, 1987b). Activation of the Ca^"^ receptor presumably regulates the levels of additional or alternative intracellular signals that can influence PTH secretion. An attractive candidate in this regard is diacylglycerol and its target enzyme, protein kinase C. Activation of protein kinase C by diacylglycerol is believed to play a role in the regulation of exocytotic secretion in various secretory cells (Knight, 1986). Although the available data are not entirely consistent, much of it suggests that protein kinase C can modulate PTH secretion regulated by extracellular Ca^"^. It has been shown that activators of protein kinase C, like phorbol myristate acetate, decrease the ability of extracellular Ca^"^ to increase inositol


293

1,4,5-trisphosphate and [Ca^"^], and decrease PTH secretion. This is reflected as a shift to the right in the concentration-response curve for extracellular Ca^"^ for each one of these parameters. Activation of protein kinase C thus decreases the sensitivity of parathyroid cells to regulation by extracellular Ca^"^ (Racke and Nemeth, 1993). In many other cell types, those receptors that are coupled to the mobilization of intracellular Ca^"^ are also sensitive to depressive effects of protein kinase C. In general, protein kinase C often acts in a negative feedback capacity to dampen signaling through the receptor-phospholipase C pathway. This seems to be one of its ftinctions in parathyroid cells, and it has been suggested that protein kinase C could directly phosphorylate the Ca^"^ receptor, thus decreasing its sensitivity to activation by extracellular Ca^"^ (Racke and Nemeth, 1993). In this regard, it is significant that the human parathyroid cell Ca^"^ receptor contains five potential protein kinase C phosphorylation sites on the putative cytoplesmic domain of the receptor (Garrett et al., 1995a). While the general mechanisms depicted in Figure 1 are supported by ample evidence, it should not be considered the penultimate model, and there are still many uncertainties. It is not clear, for instance, whether the same receptor protein couples to both adenylyl cyclase and phospholipase C. In fact, there is some data suggesting that the Ca^^ receptor on parathyroid cells is a much larger protein than the one described here (Juhlin et al., 1990). The mechanism(s) coupling the Ca^^ receptor to the influx of extracellular Ca^"^ is a topic that has been only tangentially studied. And despite much study, the role of cytosolic Ca^^ in the rapid regulation of PTH secretion is still uncertain. Nonetheless, some of the essential mechanisms comprising the initial events of stimulus-secretion coupling in parathyroid cells have been identified. These events, enabling the detection and membrane transduction of the extracellular Ca^"^ signal, are certainly involved in the acute secretory response of parathyroid cells. It seems reasonable to suppose that these same mechanisms are involved in longer term regulation of parathyroid cell functions such as synthesis of PTH and cellular proliferation. The synthesis of PTH (84 amino acids) follows the conventional pattern for proteins entering the regulated secretory pathway and is first transcribed as preproPTH (115 amino acids; Habener et al., 1984). Extracellular Ca^"^ regulates the synthesis of PTH by inhibiting transcription of preproPTH. There is evidence for a negative response element on the PTH gene sensitive to activation by Ca^"^ (Okazaki et al., 1991). Lowering the concentration of plasma Ca^"^ causes a threeto fourfold increase in message for PTH within two to three hours (Naveh-Many and Silver, 1990). Very small decreases from normocalcemia cause profound increases in the rate of synthesis of PTH. The parathyroid cell thus responds to a maintained hypocalcemic challenge by increasing both the secretion and synthesis of PTH. Hypocalcemic states lasting longer than several days are associated with hyperplasia and proliferation of parathyroid cells. It is uncertain if these latter events

294

EDWARD F. NEMETH

are regulated by the Ca^"*" receptor, but it seems possible they are. The new physiology to be learnedfromthe parathyroid cell is the significant role played by extracellular Ca^"*", which functions as an extracellular signaling ligand to control numerous mechanisms in parathyroid cells. All these mechanisms act in concert to protect the animal from hypocalcemia.

THE C-CELL Scattered throughout the thyroid gland are parafollicular or C-cells which secrete the hormone calcitonin. The C-cell, like the parathyroid cell, has long been known to respond to changes in the level of plasma Ca^"^ but the secretory products of these two cells affect plasma levels of Ca^"*" in an opposite manner: PTH causes hypercalcemia, whereas calcitonin causes hypocalcemia. The secretory responses of parathyroid cells and .C-cells are likewise regulated in opposite directions by extracellular Ca^"*". Increasing the concentration of extracellular Ca^"^ stimulates calcitonin secretion. Calcitonin then acts on target tissues to reduce the level of plasma Ca^"^ (Austin and Heath, 1981). C-cells have a different embryological origin than do parathyroid cells and derive from cells of the neural crest. Because of this, they seem to possess many properties of neuroendocrine cells. They synthesize various peptides and biogenic amines and they are electrically excitable. Most of our understanding of the cellular physiology of C-cells derives from studies using rat medullary thyroid carcinoma cells which have the advantage of being reasonably stable cell lines that express many of the characteristics believed to be representative of genuine C-cells. In these cells, increasing the concentration of extracellular Ca^"*" evokes corresponding increases in [Ca^"*']i as does depolarization of the cells by elevated levels of extracellular K^. The increases in [Ca^"^]i elicited by either of these stimuli is associated with a stimulation of calcitonin secretion (Fried and Tashjian, 1986; Muff et al., 1988). Thus, secretion in the C-cell seems to conform to the conventional Ca^"^ hypothesis of stimulus-secretion coupling, wherein cytosolic Ca^^ activates exocytotic secretion. The C-cell uses quite different mechanisms to respond to extracellular Ca^'^than does the parathyroid cell (Nemeth, 1990; Brown, 1991). In the C-cell, nearly all of the increase in [Ca^"^]} elicited by extracellular Ca^"^ resultsfrominflux; there is only a very minor contribution arising from the mobilization of intracellular Ca^"^. Moreover, in C-cells, the influx of extracellular Ca^^ is through voltage-sensitive Ca^"*^ channels. These channels have been characterized biophysically and pharmacologically and are very similar to the high-threshold, L-type Ca^"^ channels present throughout the body (Yamashita and Hagiwara, 1990). Currents through these channels can be affected by dihydropyridines. Dihydropyridines that block influx through these channels inhibit increases in [Ca^"*"]} evoked by extracellular Ca^"^, whereas those that potentiate iiiflux augment cytosolic Ca^"^ responses to extracel-


295

lular Ca^"^. This contrasts with parathyroid cells, where dihydropyridines fail to influence cytosolic Ca^"^ responses evoked by extracellular Ca^^. How extracellular Ca^"^ regulates influx through the voltage-sensitive Ca^"^ channel(s) present on C-cells is far from clear. It apparently involves some novel mechanism because there are no known voltage-sensitive Ca^"^ channels on other cells that display this sensitivity to extracellular Ca^"^. Electrophysiological studies do not reveal any peculiar properties of the L-type channel in rat C-cell lines, so there is at present no reason to suppose that extracellular Ca^"^ affects the channel directly. It seems that there is some alternative mechanism that couples to the Ca^"^ channel and regulates its sensitivity to extracellular Ca^^. Medullary thyroid carcinoma cells and parafollicular cells express a Ca^"^ receptor which is probably identical to that expressed by parathyroid cells (Garrett et al., 1995b). Presumably it is linked to the voltage-sensitive Ca^"^ channel. At present, there is very little known about the longer term regulation of C-cell functions by extracellular Ca^"^. The available data is fragmentary and does not suggest profound regulatory influences of extracellular Ca^"^ on synthesis of calcitonin or cellular proliferation of C-cells.

THE OSTEOCLAST The osteoclast is a relatively new addition to the list of extracellular Ca^'^-sensing cells. The osteoclast is primarily responsible for resorbing bone as part of the bone remodeling process and it accomplishes this task by secreting enzymes and protons. The former digests the organic components of bone (largely collagen), whereas the latter dissolves the inorganic matrix (hydroxyapatite: Caio(P04)6(OH)2). When activated, the osteoclast spreads and attaches tightly to the bone surface, effectively forming a sealed compartment beneath the cell. Actively resorbing osteoclasts are characterized morphologically by the presence of a ruffled border. This specialized part of the membrane is the site of secretion of enzymes, and additionally contains transport ATPases, some of which pump protons into the sealed compartment. The osteoclast, therefore, is a highly polarized cell, and the enzymes function together with the extremely low pH to dissolve the bone (Baron, 1989). There are many humoral and paracrine factors that turn osteoclasts on and off (Heersche, 1992; Mundy, 1992), but how these factors integrate the activity of osteoclasts into the more general scheme of bone remodeling is still far from understood. Certainly calcitonin is one of the more potent and effective hormonal factors that inhibit bone resorption. Osteoclasts possess calcitonin receptors that, when activated, inhibit secretion and cause the cells to round up. As discussed above, the rapid suppression of ongoing osteoclastic bone resorption by calcitonin can be readily monitored in vivo as a hypocalcemic response. PTH, on the other hand, activates osteoclasts. The conventional wisdom is that PTH acts indirectly, perhaps by affecting other cells in bone which then secrete some factor(s) that activates osteoclasts. Some more recent studies suggest that PTH may also have

296

EDWARD F. NEMETH

direct effects on osteoclasts. In addition, there are a wide array of cytokines and growth factors that can alter osteoclastic activity. During the process of osteoclastic bone resorption, the mineralized matrix is dissolved as the pH in the sealed compartment beneath the osteoclast falls to values of four to three. The dissolution of hydroxy appetite releases large amounts of Ca^"*" and its concentration is likely to build up in the forming lacunae in bone. Direct measurements of the concentration of extracellular Ca^"*" beneath osteoclasts in vitro reveal levels as high as 20 to 30 mM (Silver et al., 1988). Under physiological conditions, it is quite possible that the osteoclast is exposed to such high concentrations of extracellular Ca^"*". It was thus suggested that extracellular Ca^"^ controls osteoclastic activity (Teti and Zambonin-Zallone, 1987). It was subsequently shown that extracellular Ca^"^ caused increases in [Ca^'*"]i in isolated rat and avian osteoclasts in vitro (Malgaroli et al., 1989; Zaidi et al., 1989). It was additionally shown that extracellular Ca^"*" inhibited osteoclastic bone resorption in vitro. The concentrations of extracellular Ca^"^ producing these effects fall in the range of 5—20 mM. These concentrations are far in excess of the levels of Ca^"^ in plasma and most extracellular fluids but likely to be physiological for an actively resorbing osteoclast. There is thus mounting evidence suggesting that extracellular Ca^"^ released from the mineralized component of bone might function in a negative feedback capacity to depress osteoclastic activity. The physiological significance of this regulation by extracellular Ca^"^, particularly in the normal bone remodeling process, is yet to be determined. But the analogy to the parathyroid cell is obvious. In both cell types, extracellular Ca^"^ acts to increase [Ca^^]i and inhibit cellular function. We know comparatively little, however, about the mechanisms used by osteoclasts to detect and respond to changes in the concentration of extracellular Ca2^

Studies using isolated avian and rat osteoclasts suggest that increases in [Ca^"^]i evoked by extracellular Ca^"^ arise partly from the mobilization of intracellular Ca^"^ and also from influx of extracellular Ca^"^ (Zaidi et al., 1993). Like the parathyroid cell, the influx of extracellular Ca^"^ is through voltage-insensitive channels; osteoclasts appear to lack voltage-sensitive Ca^"^ channels, at least under the in vitro conditions necessary for their study. There is some evidence suggesting that voltage-sensitive Ca^"*" channels can be differentially expressed, depending on the composition of the substrate to which they are attached. The effects of extracellular Ca^"^ on [Ca^^^li and bone resorption can be mimicked by La^"*", suggesting that extracellular Ca^"*" acts at the osteoclast cell surface, perhaps by binding to an extracellular Ca^"*" receptor. However, because large populations of purified and viable mammalian osteoclasts are so difficult to obtain, there is scant biochemical data characterizing the transmembrane signaling mechanisms linked to the actions of extracellular Ca^"^ that affect [Ca^"*"]i and osteoclast function. The unavailability of tissue also limits efforts aimed at cloning the putative osteoclast Ca^"*" receptor.

Regalation of Cellular Functions

297

So the evidence supporting the presence of an extracellular Ca^"^ receptor on the osteoclast is fragmentary. There are two pieces of evidence suggesting that the extracellular Ca^"^ sensing mechanism(s) on the osteoclast is different from that on parathyroid cells and C-cells. In the first place, the concentration of extracellular Ca^"^ effective in altering osteoclastic activity is significantly higher than that which regulates the activity of parathyroid cells and C-cells (5 to 20 mM vs. 1 to 3 mM). Secondly, the extracellular -sensing mechanisms on osteoclasts and parathyroid cells are pharmacologically distinct. Organic compounds such as neomycin, which activate the parathyroid cell Ca^"^ receptor and increase [Ca^"^]i, are without effect on [Ca^"^]! in mammalian osteoclasts. The differential sensitivity to extracellular Ca^"^ and organic compounds suggests that the putative osteoclast Ca^"^ receptor is structurally distinct from that present on parathyroid cells and C-cells.

OTHER EXTRACELLULAR Ca^^-SENSING CELLS The cells discussed so far, particularly parathyroid cells and osteoclasts, play key roles in the regulation of body Ca^"*" homeostasis. The other main sites in the body that participate in body Ca^"^ metabolism are the kidney, the gastrointestinal tract, and, during pregnancy, the placenta. There is evidence that in each of these tissues there are cells that can sense and respond to changes in the concentration of extracellular Ca^"^. Proximal tubule cells of the kidney are the major site for the 1-hydroxylation of 25-hydroxyvitamin D3 to form 1,25-dihydroxyvitamin D3, the most biologically active form of vitamin D which affects a variety of cellular functions throughout the body, including regulation of PTH synthesis and intestinal uptake of dietary Ca^"*" (Kumar, 1986). 1,25-Dihydroxyvitamin D3 synthesis is increased by elevated plasma levels of PTH and decreased by hypercalcemia or hyperphosphatemia. In a series of elegant in vivo experiments, it was shown that the inhibitory effects of hypercalcemia occur independently of changes in plasma levels of PTH or phosphate (Matsumoto et al., 1987; Weisinger et al., 1989), suggesting that extracellular Ca^"^ might act directly on proximal tubule cells to regulate 1,25-dihydroxyvitamin D3 synthesis. Studies in vitro tend to support this. Thus, increased levels of extracellular Ca^"^ can block the stimulatory effects of PTH on cyclic AMP formation in isolated proximal tubule cells (Mathias and Brown, 1991) and this would be expected to lead to a decrease in the synthesis of 1,25-dihydroxyvitamin D3. Extracellular Ca^"^ might increase [Ca^"^]i in proximal tubule cells and this in itself would depress synthesis of 1,25-dihydroxyvitamin D3 since Ca^"*" can directly inhibit 1-hydroxylase activity. In the kidney, then, many of the ingredients necessary for creating a negative feedback loop akin to that seen in the parathyroid gland are present. A hypocalcemic state would stimulate 1,25-dihydroxyvitamin D3 synthesis, perhaps directly and also by increasing plasma levels of PTH. Elevated plasma levels of 1,25-dihydroxyvitamin D3 would increase the intestinal absorption

298

EDWARD F. NEMETH

of dietary calcium, resulting in increased circulating levels of Ca^"*". The rise of plasma Ca^"^ then acts directly on the proximal tubule cells of the kidney to depress synthesis of 1,25-dihydroxyvitamin D3. Extracellular Ca^"*" also blocks increases in the levels of cyclic AMP evoked by vasopressin in the medullary thick ascending limb of Henle's loop (Takaichi and Kurokawa, 1988). Significantly, this inhibitory effect is blocked by pretreatment with pertussis toxin, suggesting that extracellular Ca^"^ acts through a mechanism coupled to a Gj-like protein to depress adenylyl cyclase activity. The Ca^"^ receptor expressed in parathyroid cells and C-cells is also expressed in the kidney (Riccardi et al., 1995) and likely mediates the effects of extracellular Ca^"^ in the medullary thick ascending limb of Henle's loop. The effects of extracellular Ca^"^ observed in the proximal tubule could be mediated by this Ca^"*^ receptor or by an alternative receptor-like protein (Juhlin et al., 1987). In the gastrointestinal tract, there are only vague indications suggesting a physiologically important role for signaling by extracellular Ca^"^. Extracellular Ca^"*^ might participate in the regulation of gastrin secretion and may play a role in the proliferation of Goblet cells during embryonic development. Additional studies, with an eye towards the role of extracellular Ca^^ in regulating intestinal functions, are certainly warranted. During pregnancy, there are increased demands placed upon the maternal Ca^^ homeostatic system as the mother must now supply the Ca^"^ needed for skeletal development of the fetus (Chesney et al., 1992). One of the cells involved in the transport of Ca^^fromthe maternal to the fetal circulation is the cytotrophoblast of the placenta. There is convincing evidence showing that this cell type responds to increases in the concentration of extracellular Ca^"^ with corresponding increases in [Ca^^]i (Hellman et al., 1992). These evoked increases in [Ca^^]i are blocked by a monoclonal antibody which has been used to isolate a 500 kDa protein from placental cytotrophoblasts and it has been suggested that this protein fimctionsvas an extracellular Ca^"^ receptor (Juhlin et al., 1990). This protein clearly differs from the extracellular Ca^"^ receptor cloned from parathyroid cells, although this same monoclonal antibody blocks increases in [Ca^"*"]! evoked by extracellular Ca^"^ in parathyroid cells (Gylfe et al., 1990). Further studies are required to assess the possible role of this larger protein in regulating parathyroid cell function. The physiological significance of the extracellular Ca^"^ sensitivity of cytotrophoblasts is equally uncertain. Increasing the concentration of extracellular Ca^"^ has been shown to depress secretion of parathyroid hormone-related protein from cytotrophoblasts, and this protein has been implicated in the regulation of Ca^"^ transport in the placenta. Thus, there are various pieces of evidence suggesting that extracellular Ca^^, by actions on the cytotrophoblast, can regulate exchange of Ca^"*^ between the maternal and fetal circulation. Although not directly involved in the regulation of body Ca^^ metabolism, the juxtaglomerular cell of the kidney deserves mention because of the quite solid


299

Table 1. Extracellular Ca2+ -Sensing Cells in the Body Ceil Type Parathyroid cell Parafollicular cell Osteoclast Cytotrophoblast Kidney cells proximal tubule medullary thick ascending limb juxtaglomerular Gastrointestinal cells C-cell goblet Keratinocytes Mammary cells

Function Reguiated by Extraceiiular Ca PTH synthesis and secretion Calcitonin secretion Bone resorption Hormone secretion? Ca "^ transport? 1,25-diOH-vitamin D3 synthesis Urinary concentration Renin secretion Gastrin secretion Proliferation Proliferation Proliferation

evidence demonstrating the sensitivity of this cell to extracellular Ca^^. The juxtaglomerular cell secretes the enzyme renin which converts angiotensinogen to angiotensin I. Angiotensin I, in turn, is converted to angiotensin II by angiotensin converting enzyme. Angiotensin II acts directly and potently on vascular smooth muscle to constrict blood vessels, thus causing an increase in blood pressure. The juxtaglomerular cell therefore plays a key role in the regulation of blood pressure. Elevated levels of extracellular Ca^"^ cause increases in [Ca^"^]i and depress secretion of renin (Fray et al., 1987; Kurtz and Penner, 1989). The physiological importance of these effects of extracellular Ca^"^ on renin secretion are uncertain but an association between plasma Ca^"*" levels and hypertension has long been recognized (Bukoski and McCarron, 1988). From the above discussion, a general pattern emerges: extracellular Ca^"^ generally acts to depress cellular functions. The notable exception is the C-cell, where extracellular Ca^"^ acts to stimulate calcitonin secretion. The cell types known at present to respond to changes in the concentration of extracellular calcium are summarized in Table 1.

THERAPEUTIC SIGNIFICANCE OF EXTRACELLULAR CA^^ RECEPTORS Because extracellular Ca^"*" plays a key role in the regulation of certain cellular responses, it is possible that some disease states are intimately associated with cell surface Ca^"^ receptors. For example, in familial benign hypercalcemia, the set-point for extracellular Ca^"*" regulation of PTH secretion is increased (Khosla et al, 1993). Similar increases in the set-point for extracellular Ca^"^ are also observed in patients

300

EDWARD F. NEMETH

with primary hyperparathyroidism (Brown and Leboff, 1986). Curiously, parathyroid tissue from patients with hyperparathyroidism exhibit reduced staining with a monoclonal antibody that might interact with the parathyroid cell Ca^"^ receptor or some protein closely associated with it (Juhlin et al, 1988). There are thus some reasons for supposing that the expression of Ca^"^ receptors, or mechanisms regulating their activity may be altered in certain pathologic conditions. While it is not certain that all the cells reviewed here possess cell surface Ca^"*" receptors, certainly the parathyroid cell and the C-cell, and certain cells in the kidney do. These extracellular Ca^"^ receptors are structurally akin to many other G-protein-coupled receptors and function similarly to control the response of cells to changes in the concentration of extracellular Ca^"*". Such receptors have long been classic sites for pharmacological intervention in diverse disease states, so there is reason to suppose that extracellular Ca^"*" receptors will likewise be therapeutically relevant targets for new pharmaceuticals effective in the treatment of various disorders, especially those involving bone and mineral-related diseases. Drugs that mimic or potentiate the effects of extracellular Ca^"*" at Ca^"*" receptors are termed "calcimimetics," and act as receptor agonists. Conversely, drugs that block or depress the effects of extracellular Ca^"^ at Ca^"*" receptors are termed "calcilytics," and act as receptor antagonists. For example, calcimimetic drugs acting at the parathyroid cell Ca^"*" receptor would inhibit PTH secretion and be effective in the treatment of hyperparathyroidism. There are already compounds under development that act precisely in this manner. Cell surface Ca^"^ receptors thus provide novel and discrete molecular targets for new classes of drugs that mimic or antagonize the actions of extracellular Ca^"^ throughout the body.

SUMMARY It is now recognized that extracellular Ca^"^ can regulate the functional activity of particular types of cells in the body. Many of these cells are involved in maintaining body Ca^^ homeostasis and are present in certain endocrine glands and in bone, kidney, and the intestine. Notable among these cells are parathyroid cells which secrete parathyroid hormone (PTH). PTH acts on bone and kidney to increase the level of Ca^"*" in blood and extracellular fluids and plays a major role in maintaining body homeostasis. Parafollicular cells in the thyroid, or C-cells, secrete the hormone calcitonin which acts to decrease plasma levels of Ca^"^. The secretion of both PTH and calcitonin is regulated by changes in the concentration of extracellular Ca^"^: increased levels of extracellular Ca^"*^ inhibit PTH secretion and stimulate calcitonin secretion. The effects of extracellular Ca^"^ are mediated by a cell surface Ca^"^ receptor protein. The parathyroid cell Ca^"*" receptor has been cloned and is a member of the G protein-coupled receptor superfamily. In parathyroid cells, the Ca^"^ receptor is coupled to phospholipase C and its activation by extracellular Ca^"*^ results in the inositol 1,4,5-trisphosphate-induced release of intracellular Ca^"^, which is associated with an inhibition of PTH secretion. Other cells, such as


301

osteoclasts in bone are also responsive to changes in the concentration of extracellular Ca^"^, although the structure of the putative Ca^^ receptor on these cells is still unknown. The recognition of a wide array of cells scattered throughout the body that can detect and respond to changes in the concentration of extracellular Ca^"*" provides evidence for a signaling role of extracellular Ca^"*" that is functionally akin to molecular ligands such as hormones and neurotransmitters. The cell surface Ca^"^ receptors expressed on these cells provide novel molecular targets for new drugs to treat a variety of disease states. REFERENCES Austin, L.A. & Heath, H.I. (1981). Calcitonin. Physiology and pathophysiology. N. Engl. J. Med. 304, 269-278. Baron, R. (1989). Molecular mechanisms of bone resorption by the osteoclast. Anat. Record 224, 317-324. Berridge, M.J. (1987). Inositol trisphosphate and diacylglycerol: Two interacting second messengers. Ann. Rev. Biochem. 56, 159-193. Brown, E.M. (1991). Extracellular Ca^"^ sensing, regulation of parathyroid cell function, and role of Ca^"*" and other ions as extracellular (first) messengers. Physiol. Rev. 71, 371-411. Brown, E.M. & Leboff, M.S. (1986). Pathophysiology of hyperparathyroidism. Prog. Surg. 18, 13-22. Brown, E.M., LeBoff, M.S., Getting, M., Posillico, J.T., & Chen, C. (1987). Secretory control in normal and abnormal parathyroid tissue. Rec. Prog. Horm. Res. 43, 337-382. Brown, E.M., Chen, C.J., Kifor, O., LeBoff, M.S., El-Hajj, G., Fajtova, V., & Rubin, L.T. (1990). Ca2- -sensing, second messengers, and the control of parathyroid hormone secretion. Cell Calcium 11,333-337. Brown, E.M., Gamba, G., Riccardi, D., Lombardi, M., Butters, R., Kifor, O., Sun, A., liediger, M., & Lytton, J. (1993). Cloning and characterization of an extracellular Ca^"^-sensing receptor from bovine parathyroid. Nature 366, 575-580. Bukoski, R.D. & McCarron, D.A. (1988). Calcium and hypertension. In: Calcium in Drug Actions (Baker, P.P., ed.), pp. 467-487, Springer-Verlag, New York. Chen, C.J., Bamett, J.V., Congo, D.A., & Brown, E.M. (1989). Divalent cations suppress 3',5'-adenosine monophosphate accumulation by stimulating a pertussis toxin-sensitive guanine nucleotide-binding protein in cultured bovine parathyroid cells. Endocrinol. 124, 233-239. Chesney, R.W., Specker, B.L., Mimouni, P., & McKay, C.P. (1992). Mineral metabolism during pregnancy and lactation. In: Disorders of Bone and Mineral Metabolism (Coe, F.L. & Favus, M.J., eds.), pp. 383-393, Raven Press, New York. Cohn, D.V. & MacGregor, R.R. (1981). The biosynthesis, intracellular processing, and secretion of parathormone. Endocrine Rev. 2, 1-26. Douglas, W. W. (1974). Involvement of calcium in exocytosis and the exocytosis-vesiculation sequence. Biochem. Soc. Symp. 39, 1-28. Fried, R.M. & Tashjian, A.H., Jr. (1986). Unusual sensitivity of cytosolic free Ca^"*" to changes in extracellular Ca^"*" in rat C-cells. J. Biol. Chem. 261, 7669-7674. Fray, J.C.S., Park, C.S., & Valentine, A.N.D. (1987). Calcium and the control of renin secretion. Endocrine Rev. 8, 53-93. Garrett, J.E., Capuano, I.V., Hammerland, L.G., Hung, B.C.P., Brown, E.M., Hebert, S.C, Nemeth, E.F., & Fuller, F. (1995a). Molecular cloning and functional expression of human parathyroid calcium receptor cDNAs. J. Biol. Chem. 270, 12919-12925.

302

EDWARD F. NEMETH

Garrett, J.E., Tamir, H., Kifor, O., Simin, R.T., Rogers, K.V., Mithal, A., Gagel, R.F., & Brown, E.M. (1995b). Calcitonin-secreting cells of the thyroid express an extracellular calcium receptor gene. Endocrinology 136, 5202-5211. Oilman, A.G. (1987). G proteins: Transducers of receptor-generated signals. Ann. Rev. Biochem. 56, 615-649. Gylfe, E., Johlin, C, Akerstrom, G., Klareskog, L., Rask, L., & Rastad, J. (1990). Monoclonal antiparathyroid antibodies—tools for studies of the regulation ofcytoplasmic calcium and function of parathyroid and other antibody-reactive cells. Cell Calcium 11, 329-332. Habener, J.F., Rosenblatt, M., & Potts, J.T. Jr. (1984). Parathyroid hormone: Biochemical aspects of biosynthesis, secretion, action, and metabolism. Physiol. Rev. 64,985-1053. Hammerland, L.G., Krapcho, K.J., Alasti, N., Simin, R., Garrett, J.E., Capuano, I.V., Hung, B.C.P., & Fuller, F.H. (1995). Cation binding determinants of the calcium receptor revealed by functional analysis of chimeric receptors and a deletion mutant. J. Bone Min. Res. 10, SI 56. Heersche, J.N.M. (1992). Systemic factors regulating osteoclast function. In: Biology and Physiology of the Osteoclast (Rifkin, B.R. & Gay, C.V., eds.), pp. 151-169, CRC Press, Boca Raton, FL. Hellman, P., Ridefelt, P., Juhlin, C, Akerstrom, 0., Rastad, J., & Gylfe, E. (1992). Parathyroid-like regulation of parathyroid-hormone-related protein release and cytoplasmic calcium in cytotrophoblast cells of human placenta. Arch. Biochem. Biophys. 293,174-180. Heizmann, C.W. & Hunziker, W. (1991). Intracellular calcium-binding proteins: More sites than insights. TIBS 16,98-103. Juhlin, C, Holmdahl, R., Johansson, H., Rastad, J., Akerstrom, 0., & Klareskog, L. (1987). Monoclonal antibodies with exclusive reactivity against parathyroid cells and tubule cells of the kidney. Proc. Natl. Acad. Sci. USA 84, 2990-2994. Juhlin, C, Klareskog, L., Nygren, P., Ljunghall, S., Gylfe, E., Rastad, J., & Akerstrom, 0 . (1988). Hyperparathyroidism is associated with reduced expression of a parathyroid calcium receptor mechanism defined by monoclonal antiparathyroid antibodies. Endo 122, 2999-3001. Juhlin, C, Lundgren, S., Johnsson, H., Lorentzen, J., Rask, L., Larsson, E., Rastad, J., Akerstrom, 0., & Klareskog, L. (1990). 500-kilodalton calcium sensor regulating cytoplasmic Ca^"^ in cytotrophoblast cells of human placenta. J. Biol. Chem. 265, 8275-8279. Khosla, S., Ebeling, P.R., Firek, A.F., Burritt, M.M., Kao, P.C., & Heath, H. (1993). Calcium infusion suggests a "set-point" abnormality of parathyroid gland function in familial benign hypercalcemia and more complex disturbances in primary hyperparathyroidism. J. Clin. Endo. Metab. 76, 715-720. Knight, D.E. (1986). Calcium and exocytosis. In: Calcium and the Cell Ciba Foundation Symposium, Vol. 122, pp. 250-265, John Wiley & Sons, New York. Kumar, R. (1986). The metabolism and mechanism of action of 1,25-dihydroxyvitamin D3. Kidney Intl. 30, 79S-803. Kurtz, A. & Penner, R. (1989). Angiotensin II induces oscillations of intracellular calcium and blocks anomalous inward rectifying potassium current in mouse renal juxtaglomerular cells. Proc. Natl. Acad. Sci. USA 86, 3423-3427. Lopez-Bameo, J. & Armstrong, CM. (1983). Depolarizing response of rat parathyroid cells to divalent cations. J. Gen. Physiol. 82, Malgaroli, A., Meldolesi, J., Zambonin-Zallone, A., & Teti, A. (1989). Control of cytosolic free calcium in rat and chicken osteoclasts. The role of extracellular calcium and calcitonin. J. Biol. Chem. 264, 14342-14347. Mathias, R.S. & Brown, E.M. (1991). Divalent cations modulate PTH-dependent 3',5'-cyclic adenosine monophosphate production in renal proximal tubular cells. Endocrinol. 128, 3005-3012. Matsumoto, T., Ideda, K., Morita, K., Fukumoto, S., Takahashi, H., & Ogata, E. (1987). Blood Ca^"^ modulates responsiveness of renal 25(OH)D3-la-hydroxylase to PTH in rats. Am. J. Physiol. 253, E503-E507.


303

Milner, R.E., Famulski, K.S., & Michalak, M. (1992). Calcium binding proteins in the sarcoplasmic/endoplasmic reticulum of muscle and nonmuscle cells. Molec. Cell Biochem. 112, 1-13. Muff, R., Nemeth, E.F., Haller-Brem, S., & Fischer, J.A. (1988). Regulation of hormone secretion and cytosolic Ca^"^ by extracellular Ca^^ in parathyroid cells and C-cells: Role of voltage-sensitive Ca^"^ channels. Arch. Biochem. Biophys. 265, 128-135. Mundy, G.R. (1989). Calcium Homeostasis: Hypercalcemia and Hypocalcemia. Martin Dunitz Ltd., London. Mundy, G.R. (1992). Local factors regulating osteoclast function. In: Biology and Physiology of the Osteoclast (Rifkin, B.R. & Gay, C.V., eds.), pp. 171-185, CRC Press, Boca Raton, FL. Naveh-Many, T. & Silver, J. (1990). Regulation of parathyroid hormone gene expression by hypocalcemia, hypercalcemia, and vitamin D in the rat. J. Clin. Invest. 86, 1313-1319. Nemeth, E.F. (1990). Regulation of cytosolic calcium by extracellular divalent cations in C-cells and parathyroid cells. Cell Calcium 11, 323-327. Nemeth, E.F. & Scarpa, A. (1986). Cytosolic Ca^"*" and the regulation of secretion in parathyroid cells. FEBS Lett. 203, 15-19. Nemeth, E.F. & Scarpa, A. (1987a). Rapid mobilization of cellular Ca^"*" in bovine parathyroid cells evoked by extracellular divalent cations. Evidence for a cell surface calcium receptor. J. Biol. Chem. 262, 5188-5196. Nemeth, E.F. & Scarpa, A. (1987b). Are changes in intracellular free calcium necessary for regulating secretion in parathyroid cells? Ann. New York Acad. Sci. 493, Okazaki, T., Zajac, J.D., Igarashi, T., Ogata, E., & Kronenberg, H.M. (1991). Negative regulatory elements in the human parathyroid hormone gene. J. Biol. Chem. 266, 21903—21910. Racke, F.K., Hammerland, L.G., Dubyak, G.R., & Nemeth, E.F. (1993). Functional expression of the parathyroid cell calcium receptor in Xenopus oocytes. FEBS Lett. 333, 132-136. Racke, F.K. & Nemeth, E.F. (1993). Cytosolic calcium homeostasis in bovine parathyroid cells and its modulation by protein kinase C. J. Physiol. 468, 141-162. Riccardi, D., Park, J., Lee, W.-S., Gamba, G., Brown, E.M., & Hebert, S.C. (1995). Cloning and functional expression of a rat kidney extracellular calcium/polyvalent cation-sensing receptor. Proc. Natl. Acad. Sci. USA 92, 131-135. Schoepp, D., Bockaert, J., & Sladeczek, F. (1990). Pharmacological and functional characteristics of metabotropic excitatory amino acid receptors. TIPS 11, 508-515. Shoback, D.M., Thatcher, J., Leombruno, R., & Brown, E.M. (1984). Relationship between parathyroid hormone secretion and cytosolic calcium concentration in dispersed bovine parathyroid cells. Proc. Natl. Acad. Sci. USA 81,3113-3117. Silver, I.A., Murrills, R.J., & Etherington, D.J. (1988). Microelectrode studies on the acid microenvironment beneath adherent macrophages and osteoclasts. Exp. Cell. Res. 175,266-276. Strosberg, A.D. (1991). Structure/function relationship of proteins belonging to the family of receptors coupled to GTP-binding proteins. Eur. J. Biochem. 196, 1-10. Takaichi, K. & Kurokawa, K. (1988). Inhibitory guanosine triphosphate-binding protein-mediated regulation of vasopressin action in isolated single medullary tubules of mouse kidney. J. Clin. Invest. 82, 1437-1444. Teti, A. & Zambonin-Zallone, A. (1987). A working hypothesis: Calcium concentration controls directly osteoclast activity. In: Calcium Regulation and Bone Metabolism. Basic and Clinical Aspects (Cohn, D.V., Martin, T.J., & Meunier, P.J., eds.), Vol. 9, pp. 358-362, Excerpta Medica, New York. Watson, P.H. & Hanley, D.A. (1993). Parathyroid hormone: Regulation of synthesis and secretion. Clin. Invest. Med. 16,58-77. Weisinger, J.R., Favus, M.J., Langman, C.B., & Bushinsky, D.A. (1989). Regulation of 1,25-dihydroxyvitamin D3 by calcium in the parathyroid^ctomized, parathyroid hormone-replete rat. J. Bone Min. Res. 4, 929-935.

304

EDWARD F. NEMETH

Yamashita, N. & Hagiwara, S. (1990). Membrane depolarization and intracellular Ca^^ increase caused by high external Ca^"^ in a rat calcitonin-secreting cell line. J. Physiol. 431, 243-267. Zaidi, M., Datta, H.K., Patchell, A., Moonga, B., & Maclntyre, I. (1989). Calcium-activated intracellular calcium elevation: A novel mechanism of osteoclast regulation. Biochem. Biophys. Res. Comm. 163, 1461-1465. Zaidi, M., Alam, A.S.M.T., Shankar, V.S., Bax, B.E., Bax, CM., Moonga, B.S., Bevis, P.J.R., Stevens, C, Blake, D.R., Pazianas, M., & Huang, C.L.H. (1993). Cellular biology of bone resorption. Biol. Rev. 68, 197-264.

Chapter 12

The Basis of Intracellular Calcium Homeostasis in Eukaryotic Cells FRANCESCO Dl VIRGILIO, DANIELA PIETROBON, and TULLIO POZZAN

Introduction Mechanisms of Intracellular Ca ^ Homeostasis How Cells Handle Ca^^ Ca Transport Systems of the Plasma Membrane Ca -Storing Intracellular Organelles HowIsCa Released From Intracellular Stores?

305 307 308 308 316 321

Intracellular Ca Oscillations: A New Signaling Code Cytoplasmic Ca -Binding Proteins Conclusions

322 323 325

2+

INTRODUCTION The cell, the smallest unit capable of independent life, is surrounded by a lipid barrier, the plasma membrane, at the level of which communication between the external environment and the cell itself takes place. Most extracellular stimuli (hormones, growth factors, neurotransmitters) are unable to cross the plasma membrane; thus, their messages have to be transduced by specialized structures Principles of Medical Biology, Volume 4 Cell Chemistry and Physiology: Part II, pages 305-327 Copyright © 1996 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-806-4

305

306

F. Dl VIRGILIO, D. PIETROBON, and T. POZZAN

G PROTEIN LINKED RECEPTORS

RECEPTORS WITH INTRINSIC TYROSINE KINASE* ACTIVITY

RECEPTORS WITH COUPLED TYROSINE KINASE ACTIVITY

Figure 1. Pathways for intracellular Ca^"*^ mobilization by plasma membrane receptors (modified from Berridge, 1993). All eukaryotic cells, with the exception of erythrocytes, possess Intracellular Ca^* stores which can be mobilized by a diffusible factor, inositol 1,4,5-trisphosphate (IP3). IP3 can be generated by two major receptormediated pathways: receptors coupled to G proteins and receptors with intrinsic or coupled tyrosine kinase activity. G protein-coupled receptors typically have seven transmembrane domains (seven membrane-spanning receptors); a family of trimeric G proteins couple these receptors to the enzyme phospholipase C p1 (PLC p1), one of three main PLC family members. PLC pi hydrolyzes the membrane phospholipid phosphatidyl inositol 4,5-bisphosphate (PIP2) and generates diacylglycerol (DAG) and IP3. PIP2 can also be hydrolyzed by another PLC Isoform (PLC yl) which is activated by tyrosine kinase receptors. Receptors may trigger tyrosine kinase activity either directly (receptors with intrinsic tyrosine kinase activity, e.g., growth factor receptors) or by stimulating an associated tyrosine kinase (receptors with coupled tyrosine kinase activity, e.g., the T-cell receptor). DAG is also a very important intracellular messenger as it specifically activates protein kinase C (PKC), a serin^threonine kinase which has a key role in cell responses. (receptors). In turn, membrane receptors generate additional messengers (second messengers) which convey the information to the cellular effector systems. In contrast to the many molecules with a recognized extracellular messenger function, known intracellular messengers are few: cyclic nucleotides, inositol phosphates, diacylglycerol, and Ca^"*" (Berridge, 1993). In addition to these diffusible messen-

Cs^'*' Homeostasis

307

gers, some receptors are linked to their effector systems via an intrinsic (e.g., growth factor receptors) or coupled (e.g., the T cell receptor) tyrosine specific kinase or phosphatase activity (Figure 1).

MECHANISMS OF INTRACELLULAR Ca^^ HOMEOSTASIS Ca is an abundant element in the body (the total Ca content of an average 70 kg adult man is about 1,160 grams), but most of it is present as Ca phosphate in bone, bound to extracellular or intracellular proteins, or sequestered within cellular organelles. As first recognized by McLean and Hastings, the key parameter which controls cellular functions is not total Ca, but rather the ionized form, which is only a fraction thereof The difference between total Ca and Ca present as the free ion (Ca^"^) is even more striking within the cell, where the total Ca content can be as high as 2-3 mM, while the free cytoplasmic Ca^"*" concentration ([Ca^"^],) under resting conditions never exceeds 100-200 nM, i.e., less than 1/20,000 of total (Pietrobon et al., 1990). A similar difference exists between the extracellular and intracellular compartments, as the Ca^"^ concentration in the blood or in the extracellular fluid is about 1-2 mM. Why living cells established and conserved such a low free [Ca^"^], throughout evolution is not clear; however, it has been suggested that in the primeval aqueous environment the Ca^"^ concentration was low. Hence, cells were able to develop an energy generating system based on the hydrolysis of the phosphate bond. It is believed that in primeval times little or no difference in concentration existed between the intracellular and extracellular Ca^"^. When the Ca^"^ concentration in the primordial sea began to increase, cells were faced with the necessity to keep [Ca^'^]i low in order to prevent precipitation of Ca^"*'-phosphate salts and phosphate esters, a situation which would severely hamper phosphate-based energy metabolism (Rasmussen, 1981). Thus a complex system of Ca^"^-transport and buffering mechanisms had to be developed. The crucial role of Ca^"^ homeostatic mechanisms in preserving physiological cell functions is epitomized by the dramatic effects of uncontrolled [Ca^"^]} elevations due to failure of plasma membrane Ca^'^-Mg^"*'-ATPase (e.g., as a consequence of inhibition of ATP generating systems) or to sustained activation of plasma membrane Ca^"^ channels (e.g., as a consequence of massive release of excitatory amipo acids in the brain). Abnormal increases in [Ca^î are known to irreversibly turn on a number of enzyme pathways (i.e., phospholipases and/or endonucleases) which are implicated in the triggering of cell death, whether occurring via the acute mechanism of necrotic lysis or the slower and more complex process of programmed cell death (Trump and Berezesky, 1992). Furthermore, elevated [Ca^"^]! levels would also result in massive mitochondrial Ca^"^ accumulation, leading to deposition of Ca^"*" salts in the matrix and uncoupling of oxidative phosphorylation. However, an abnormal increase in [Ca^"*"]i can have dire consequences even without causing cell death, but merely by triggering a hyperstimulation of physiological Ca^'^-dependent cellular responses. This is well illustrated by malignant

308


hyperthermia, a syndrome characterized by a disorder of [Ca^^]i regulation which seems to be the resuh of genetic mutation of the Ca^"^ release channel of the sarcoplasmic reticulum (SR) of skeletal muscle. This mutation lowers the threshold for activation of Ca^"^ release in skeletal muscle fibers (Nelson, 1983), thus conferring an increased excitability to the hyperthermic muscle that, under stress conditions (in humans typically during halothane-induced anesthesia), is susceptible to undergoing a sustained contracture followed by a life-threatening rise in body temperature.

HOW CELLS HANDLE Ca^^ Three systems involved in intracellular Ca^"^ handling are commonly recognized: (a) Ca^"^ transport systems of the plasma membrane; (b) Ca^'^-storing intracellular organelles; and (c) Ca^^-binding proteins (Figure 2). These systems are generally present in all eukaryotic cells, although each cell type has developed them to a different degree of sophistication according to its functional specialization. As is often the case during evolution, cells "learned" to exploit to their advantage the new environmental conditions characterized by the huge disequilibrium between the extracellular and intracellular Ca^"*" concentrations, and used the gradient existing across the plasma membrane as a signaling device. This evolutionary adaptation was probably fueled by the growing complexity of the intracellular milieu (formation of many different organelles with specific functions), and the emergence of primordial multicellular organisms which required busy cell-to-cell communication. An easy, fast, and energetically cheap means of communication between the cell and the outside world was urgently needed. Though it is impossible to know how it happened that Ca^"^ was assigned the role of an intracellular second messenger, we think it likely that, given the unusually large difference in Ca^"^ concentration across the plasma membrane, it was easy for the cell to use this ion as a trigger and/or a regulator of metabolic reactions once it learned how to handle its influx and tune its intracellular concentration. An elevation in free [Ca^^j can be triggered by increasing the permeability of the plasma membrane to Ca^"^, by releasing Ca^"*" from cellular stores or by both mechanisms. Whether one pathway—Ca^"^ influx across the plasma membrane or Ca^"*" release from stores—predominates would mainly depend on the trigger stimulus and the cell type involved. Ca^"^ Transport Systems of the Plasma Membrane Since the plasma membrane is basically impermeant to charged species, Ca^"^ transport is mediated by protein structures belonging to three different classes: channels, pumps, and exchangers.

Figure 2. Ca^^ transport and storing systems. All eukaryotic cells possess sophisticated mechanismsforcontrolling intracellular [Ca^"^]! homeostasis. These mechanisms can be basically divided into plasma membrane Ca^"^ transport systems (channels, pumps, and exchangers), intracellular Ca^"'-storing organelles, and cytoplasmic Ca^"*"binding proteins. In the plasma membrane we find Ca^"^ channels which can be activated by changes in the membrane potential {VOC, voltage-operated channels), by IP3 (SMOC, second messenger operated channels), by a putative factor (CIF, calcium influx factor) released by depletion of Ca^"*" stores (CRAC, Ca^"^ release-activated channels), and by direct interaction with an extracellular ligand (LOC, ligandoperated channels). Other plasma membrane receptors (GLR, G protein-linked receptors) are coupled via G proteins to the hydrolysis of PIP2, thus generating diacylglycerol (DAG) and IP3. IP3 triggers Ca^"^ release from the intracellular stores by activating the IP3 receptor (IP3R). Intracellular Ca^^ stores are thought to coincide with specialized portions of the ER, also called calciosomes. The stores are characterized by the presence of Ca^"^-release channels (IP3R or ryanodine receptor, RYR), Ca^^-ATPase (SERCA, sarcoplasmic/endoplasmic reticulum Ca^'^-Mg^^-ATPase), and lumenal Ca^^-binding proteins (^). Ca^"" release from the RYR can also be triggered by Ca^"^ itself via the mechanism of Ca^'^-induced Ca^"^ release. Ca^"^ can also be accumulated by the mitochondria via a uniport transporter driven by the electrochemical potential across the mitochondrial membrane. Ca^"^ is transported across the plasma membrane against its concentration gradient by the plasma membrane Ca^'^-Mg^'"-ATPase (PMCA) or the NaVCa^"" exchanger. In the cytoplasm, Ca^"" is also bound to Ca^'^-binding proteins such as parvalbumin and calmodulin (^^). 309

310


Channels Ca^"^ influx is a passive process driven by the Ca^"*" gradient across the plasma membrane. This ion is in fact four orders of magnitude more concentrated in the extracellular space than in the cytoplasm (1—2 mM versus 100-200 nM, respectively); furthermore, the cytoplasmic side of the membrane is normally negative, thus generating an electrical gradient, which in combination with the chemical gradient and in the absence of constraints to diffusion or gating mechanisms, would bring [Ca^^'l to a level of 0.1-0.2 M. Three main types of Ca^'^-permeable ion channels are known: voltage-operated channels, ligand-operated channels, and second messenger operated channels. A fourth family of channels, activated by depletion of Ca^"*" from intracellular stores, has also been described and named Ca^"^ release-activated channels. Voltage-operated channels. The distinctive feature of voltage-operated channels (VOCs) is that the transition from closed to open conformations is controlled by the membrane potential. Multiple types of voltage-operated channels are present in excitable and also in some nonexcitable cells. Besides the steep voltage-dependence of activation, the different types of C??^ VOCs have in common a very high selectivity for Ca^"^ over Na"*". Ca^"^ VOCs have been classified according to different criteria, which correspond to the 3 different levels, pharmacological, functional or molecular, at which diversity of Ca^^ VOCs can be analyzed. Functionally, on the basis of membrane potential at which the channels start to activate, Ca^^ VOCs have been subdivided in two classes: low voltage-activated (LVA) channels, characterized by a threshold for activation more negative than—50 mV, and high voltage-activated channels (HVA), characterized by a threshold for activation more positive than -50/-40 mV (Bean, 1989; Hess, 1990). LVA Ca^"" channels have a low conductance [8 pS (picoSiemens)] and are also termed "T" (for transient), because they rapidly inactivate (in a few tens of milliseconds) during sustained depolarization. Their structure has not yet been determined, mostly because of the lack of selective ligands. Since LVA Ca^"^ channels are activated by small depolarizations close to the resting potential, their most prominent function is probably to support pacemaker activity or Ca^"*" entry at negative membrane potentials. They are thought to be responsible for neuronal oscillatory activity that is likely implicated in various brain functions such as wakefulness regulation and motor coordination. HVA Ca^"^ channels have been far better characterized than LVA Ca^"^ channels, both functionally and structurally. According to pharmacological criteria, i.e., on the basis of the selective inhibition by specific drugs, four subclasses can be identified: (a) dihydropyridine (DHP)-sensitive or L-type channels; (b) co-conotoxin-GVIA (co-CgTx)-sensitive or N-type channels; (c) co-Agatoxin-IVA (AgaIVA)-sensitive or P-type channels; and (d) Ca^"^ channels resistant to DHP,

Cs?'*' Homeostasis

311

co-CgTx, and Aga-IVA, and for which a specific inhibitor has not yet been found (Tsien et al, 1991; Bertolino and Llinas, 1992; Mintz et al., 1992). L-type channels, the best known class, are found in most excitable and nonexcitable cells and are thought to be implicated in numerous important cell functions. They are the major pathway for voltage-gated Ca^"*" entry in cardiac and smooth muscle and, as such, play a crucial role in excitation-contraction coupling. In skeletal muscle, L-type channels act both to allow influx of extracellular Ca^"^ and also as voltage sensors mediating the release of Ca^"*" from intracellular stores. In endocrine and some neuronal cells, L-type channels are involved in the control of hormone and neurotransmitter release. The wide distribution of L-type channels in the central nervous system and their clustering in cell bodies and at the base of proximal portions of major dendrites suggests that they may be critically involved in initiating Ca^'^-dependent intracellular regulatory events in response to dendritic electrical activity. In fact, 1;here is evidence that L-type channels play a key role in coupling strong synaptic excitation to regulation of gene expression in cortical neurons and to induction of long lasting modifications of neuronal excitability known as "long term potentiation" and "long term depression" in hippocampal neurons. Common functional properties of L-type channels are: (a) threshold for activation around -30/-20 mV; (b) steady-state inactivation starting only at relatively positive voltages (-60 mV); (c) unitary single channel conductance of 22-27 pS; and (d) slow time course of inactivation. However, multiple functionally and structurally different subtypes of DHP-sensitive channels are known to exist in different tissues and to be co-expressed in a given cell type. L-type Ca^"*" channels of skeletal muscle are the best characterized at the structural level. They are composed of four tightly coupled subunits, alpha 1, alpha2-delta, beta, and gamma (Catterall, 1991). The alpha 1 subunit is the voltage-sensitive, pore-forming component of the channel complex responsible for pharmacological sensitivity; in addition, it also contains the binding sites for dihydropyridines and is phosphorylated by several protein kinases. Its structure shares a basic design with other voltage-gated ion channels, consisting of a set of six potential membrane spanning segments (SI-S6) flanked by cytoplasmic hydrophylic sequences, plus an H5 (between S5 and S6) sequence. As in the case of Na"*" channels, this basic design is repeated in four homologous domains. S4, which contains positively charged residues at every third or fourth position, is thought to form part of the voltagesensing machinery, while H5 is involved in pore formation and control of ion selectivity. Following the initial cloning of the skeletal muscle alpha 1 subunit, cDNAs for homologous alpha 1 subunits have been isolated from a variety of tissues. A gene family including at least six different genes encodes different alpha 1 subunits, three DHP-sensitive (alpha IC, alpha ID, and alpha ISM) and three DHP-insensitive (alphalA, alphalB, and alphalE) (Snutch and Reiner, 1992). Further molecular

312


diversity is created by alternative splicing of each gene. Another gene family, including at least four different genes (each giving rise to different isoforms), encodes different beta subunits which have been shown to modulate channel function. Heterologous expression studies have shown that alpha IB encodes N-typeCa^"^ channels that are irreversibly inhibited by a toxic peptide, co-CgTx, isolated from the marine snail Conus geographus (Williams et al., 1992). The N-type channel expression is largely restricted to neurons but, while such channels certainly play a crucial role in controlling neurotransmitter release in lower vertebrates, their role in mammals seems to be minor. This is indicated by the fact that only a small fraction of the Ca^"^ influx and glutamate release from rat brain synaptosomes is inhibited by co-CgTx (Turner et al., 1992). Immunocytochemistry and electrophysiological data indicate that there are multiple structurally and functionally different subtypes of co-CgTx-sensitive channels. It is therefore impossible to generalize about the functional properties of N-type channels. Their most commonly found features are: (a) a threshold for activation similar to that of L-type channels; (b) a biphasic time course of inactivation; (c) a steady-state inactivation at relatively negative voltages (—90 mV); and (d) a single channel conductance of about 14 pS. P-type Ca^"^ channels, first described in cerebellar Purkinje neurons, where they account for 90% of Ca^"*" currents, have a wide distribution in the nervous system. P-type channels are specifically inhibited by Aga-IVA, a peptide isolated from the venom of the funnel web spider Agelenopsis aperta. In Purkinje cells they activate at around -40 mV, inactivate very little during 1 second long depolarizations and show steady-state inactivation at relatively positive voltages (-60 mV). The identity of the gene coding for the P-type channel is not yet clear. Finally, there is an additional class of Ca^"^ channels which are resistant to DHP, co-CgTx, and Aga-IVA. These channels are poorly characterized but seem nonetheless to have an important role in controlling neurotransmitter release in mammals. Ca^"^ VOCs are modulated by a high number of physiological agonists. Particularly well-studied is the stimulatory modulation of cardiac and skeletal muscle L-type channels by beta adrenergic agonists and the inhibitory modulation of neuronal N-type channels by many neurotransmitters. Most of the modulation of L-type channels by beta adrenergic agonists is mediated by cAMP-dependent phosphorylation. In addition, there is evidence for a contribution from a fast parallel direct interaction between the GTP-binding protein Gs and this type of channel. In central neurons a cAMP-dependent stimulatory modulation of N-, L-, and P-type channels has been reported. A wide variety of neurotransmitters inhibit neuronal N-type channels by activating a different class of G proteins (G proteins sensitive to inactivation by the toxin of Bordetella pertussis) that act on N-type channels via fast direct interaction. Neurotransmitter-induced inhibition of Ca^"*" VOCs is thought to contribute to presynaptic inhibition at nerve terminals.

Ca'^'^ Homeostasis

313

Ligand-operated channels. This Ca^"^ influx pathway is activated by binding of an extracellular agonist. The receptor for the extracellular ligand coincides with the channel, which is therefore opened directly by ligand-binding, without the mediation of G proteins or second messengers generated in response to receptor stimulation. The paradigm is the cholinergic nicotinic receptor/channel, which however under physiological conditions mainly transports Na"^ and K"^. More selective for Ca^^ are two other channels, one activated by glutamate (Moriyoshi et al., 1991) and the other by extracellular ATP (Bean and Friel, 1990). Glutamate is the main excitatory transmitter in the central nervous system. Three glutamate receptor subtypes are known, two of which include a channel directly gated by glutamate (ionotropic receptors). Molecular and expression cloning have delineated a mammalian gene family of glutamate ionotropic receptor subunits containing more than 16 members, with distinct expression profiles in the central nervous system (Wisden and Seeburg, 1993). Glutamate ionotropic receptors can be pharmacologically differentiated on the basis of preferential activation by the glutamate analogues, NMDA (N-methyl-D-aspartate), AMPA ((S)-alpha-amino3-hydroxy-5-methyl-4-isoxazolepropionic acid), and the toxin kainate. Kainate and AMPA-activated receptors are mainly permeable to monovalent cations. However, depending on the subunit composition, they may also exhibit a significant permeability to Ca^"". The NMDA receptor is also permeable to Ca^"^ and is characterized by a peculiar behavior in that it is blocked by Mg^"*" in a voltage-dependent manner and requires the co-activator glycine for opening. Strong depolarizations are necessary to remove the Mg^"^ block; such a feature distinguishes the NMDA receptor from other glutamate-activated channels, and is at the basis of its key role in very important neuronal functions such as long-term potentiation and in neuronal cell death following ischemia (Meldrum and Garthwaite, 1990). Another very interesting ligand-operated Ca^^ channel is that activated by extracellular ATP. In fact, it is becoming clear that ATP may serve as an extracellular mediator released from several sources (nerve terminals, platelets, endothelial cells, chromaffin cells) and by ligating specific surface receptors known as "P2 purinergic receptors." Based on pharmacological and functional studies, five different P2 purinergic receptors have been identified, one of which is an ATPactivated channel (Dubyak, 1991). Channels gated by extracellular ATP have been discovered in both excitable and nonexcitable cells (neurons, skeletal muscle cells, mouse thymocytes, and fibroblasts), where they cause Na"^, K^, and Ca^"^ fluxes, thus triggering [Ca^^]i increases and membrane depolarization. These channels have a low selectivity for Ca^"^ over Na"^ (3:1). The physiological role of these channels is unknown, although there is good evidence for ATP being released as a neurotransmitter at sympathetic terminals. An additional and attractive hypothesis is that release of this nucleotide from injured cells may signal tissue damage.

314


Second messenger-operated channels. A common characteristic of second messenger-activated Ca^"^ channels (SMOCs) is the much lower selectivity for Ca^"^ over monovalent cations compared to Ca^"*" VOCs. They may constitute a parallel or (depending on the cell type) alternative mechanism with respect to Ca^"*" release-activated channels (see below) for generating a sustained elevation of [Ca^'^li following receptor stimulation. Ca^"^ fluxes through these channels are activated by second messengers generated in response to receptor stimulation. They are therefore gated from the cytosolic side of the plasma membrane. Diverse SMOCs have been described that differ in the nature of the activating second messenger. Different cell types appear to express different SMOCs. Possibly the best characterized second messenger-operated channel is the cation channel gated by cGMP present in the rod outer segment of the retina (Kaupp and Koch, 1992). In fact, in darkness, the high cGMP concentration within the rods keeps open a membrane channel permeable to Na"^, Ca^"^, and Mg^"^. Light triggers the hydrolysis of cGMP, thus causing closure of the cation channel and rod cell activation. Other cyclic nucleotide-gated channels have been identified in sensory neurons of olfactory epithelium, retinal bipolar cells, renal inner medullary collecting ducts, and sinoatrial myocytes. Plasma membrane Ca^"*" channels can also be opened from the cytoplasmic side by IP3, independently of its known action on the intracellular stores. IP3 receptors, molecularly different from those present in the intracellular Ca^"^ stores, have been identified in lymphocytes and olfactory cells (Kahnetal., 1992). A role in the activation of plasma membrane Ca^"*" channels has also been suggested for inositol 1,3,4,5-tetrakisphosphate (IP4), but a convincing demonstration is still lacking. Both IP3 and IP4 are generated by agonists acting through receptors coupled to the phosphoinositide signaling system via trimeric G proteins (the so-called seven-membrane spanning receptors); thus inositolphosphate-gated channels may represent one of the major pathways for Ca^"^ influx activated upon stimulation of phospholipase C-coupled receptors. Ca^"^ channels activated by a rise in [Ca^"*"]}, or by direct interaction with G proteins have also been described in a variety of cell types (e.g., neutrophils and endothelial cells). Cal^^ release-activated channels. A notable observation is that Ca^"^ influx across the plasma membrane depends on the filling state of the intracellular Ca^^ stores (Penner et al., 1993). The main evidence in support of this is that (a) IP3-generating agonists trigger biphasic Ca^"^ changes (Ca^"*" release from stores, followed by Ca^"^ influx), and (b) inhibitors of the microsomal Ca^^ pump which block Ca^"*" reuptake cause both store depletion and Ca^^ influx from the medium. The most widely accepted interpretation of these results is that specific plasma membrane channels are able to sense the Ca^"^ content of the intracellular stores. How the stores communicate their filling state to the plasma membrane is unknown, but two hypotheses have been put forward. The first states that Ca^^ stores are at least in part in physical contact with the plasma membrane, and communication

Ca^"^ Homeostasis

315

occurs between the IP3 receptor and a putative plasmalemmal protein (a coupling protein or the channel itself). The second states that a diffusible messenger which interacts with a plasma membrane receptor is released from empty intracellular stores. This putative messenger has been tentatively identified as cGMP or with a still-to-be-identified low MW phosphorylated hydrophobic molecule preliminarily named GIF (Ca^"^ Influx Factor). It is thus not unlikely that these channels will be classified in the future as SMOCs (Randriamampita and Tsien, 1993). Although the precise mechanism of communication between the intracellular Ca^"^ stores and the plasma membrane is still speculative, the general consensus favors a retrograde diffusible messenger specifically generated by the empty stores. Although these channels were first described in mast cells, they are now known to exist in almost all nonexcitable cells. The Plasma Membrane Ca^^ Pump

An efficient Ca^'^-extruding system is crucial for keeping [Ca^"^]i within physiological resting levels. The principal pumping device involved in transmembrane Ca^"^ transport is the plasma membrane Ca^'^-Mg^'^ -ATPase (PMCA). Four different genes coding for different isoforms of about 130 kDa have been described (PMCA 1, PMCA 2, PMCA 3, and PMCA 4); in addition to these four gene isoforms, further variants (spliced isoforms) are generated by alternative splicing of gene transcripts (Carafoli, 1992). The pump has a large cytoplasmic domain which comprises about 80% of the molecule, and 10 transmembrane helices linked by short loops. The cytoplasmic domain can be divided into three main units, each with a defined function. Unit one, which protrudes from transmembrane helices 2 and 3, corresponds to the transducing unit which couples (with a stoichiometry of 1:1) ATP hydrolysis to Ca^^ translocation; unit two, which protrudes from the transmembrane helices 4 and 5, contains regulatory phosphorylation sites; and unit three, which protrudes from transmembrane helix 10, contains the Ca^"^-calmodulin-binding sequence and two sites that can be phosphorylated by protein kinase A and C. Direct stimulation of PMCA by Ca^"^-calmodulin decreases the K^ for Ca^"*" from about 20 to 0.5 juM and increases pump activity. Stimulation by Ca^'^-calmodulin is the result of relieving an inhibitory constraint. In fact, in the absence of Ca^"^-calmodulin, the C-terminal domain comprising the calmodulin-binding site of the pump appears to fold over and shield the cytoplasmic protruding units one and two, thus hampering substrate access to the catalytic site located on subunit one. Upon Ca^'^-calmodulin binding, interaction between the C-terminal region and units one and two is relieved, thus allowing free access of ATP to the catalytic site. In the absence of calmodulin, PMCA can be activated by exposure to acidic phospholipids, polyunsaturated fatty acids, phosphorylation mediated by protein kinase A or C, and oligomerization. Among these alternative activating stimuli, phosphatydilinositol might have an important physi-

316


ological role since, on a molar basis, it is more potent than other phospholipids and its concentration in the membrane is modulated by plasma membrane receptors. Digestion by the intracellular protease calpain might also be important in the physiological process of activation of the pump, as this protease is known to be preferentially targeted to proteins containing calmodulin-binding sites. The NaVCa^-" Exchanger This exchanger was initially identified in the late 1960s in cardiac muscle and invertebrate nerve. Its presence is now well documented in numerous excitable and nonexcitable cells, but its role is well defined only in excitable cells. The Na"*'/Ca^'^ exchanger is a crucial point of intersection between Na"^ and Ca^"^ homeostasis as it is driven by the concentration gradients of these two ions across the plasma membrane and the transmembrane potential (Blaustein, 1988). Sensitivity to voltage is conferred by a NaVCa^"*" coupling ratio of 3:1 which implies the net countermovement of a positive charge for each Ca ion translocated across the plasma membrane. At transmembrane potentials positive to the reverse potential of the exchanger, Na"*" is driven out and Ca^"^ in; the reverse happens for transmembrane potentials negative to the reverse potential. This means that the exchanger drives Ca^"*" efflux under resting conditions and Ca^^ influx when the plasma membrane is depolarized. Therefore, in excitable cells the exchanger will synergize with plasma membrane Ca^"*" channels in increasing [Ca^^]i during an action potential, and with the plasma membrane Ca^'^Mg^'^-ATPase in extruding Ca^"^ from the cytosol at rest or during the repolarization phase. The exchanger has a rather high K^ for Ca^"*" (~2 jiM), which may suggests that its major role is in the gross resetting of [Ca^"^]i; however, it cannot be excluded that, in cells expressing a high level of the exchanger, it may also contribute to the fine regulation of [Ca^'^Jj. As a consequence of the presence of the NaVCa^"^ exchanger, any alteration of intracellular Na"^ concentration will also have profound effects on [Ca^"^];. For example, it is well known that cardiac glycosides, which inhibit the Na'^-K'*^-ATPase, have a positive inotropic effect that can be explained in terms of an increase in [Ca^^Jj. In fact, inhibition of the Na"*"-K'^-ATPase induces cytoplasmic Na"^ accumulation; elevation of cytoplasmic Na^ in turn drives Ca^"^ into the cytosol via the exchanger and, as a result, [Cd?'^][ is increased. An elevation in [Ca^"^]i can either directly increase cardiac muscle tension, or cause a larger accumulation of Ca^"*" into the SR, thus increasing the Ca^"*" releasable pool available for muscle contraction. Ca^*-Storing Intracellular Organelles All cells, except erythrocytes, possess an intracellular vesicular compartment (Ca^"^ stores), wherefi-om Ca^^ can be mobilized via activation of specific mem-

C3?^ Homeostasis

317

brane receptors. Such compartments include the SR of muscle cells and the endoplasmic reticulum (ER), or modified portions thereof (calciosome), of nonmuscle cells (Pozzan et al., 1994). Calcium is also contained within mo^t cellular organelles (e.g., mitochondria and secretory granules) wherefrom it can be slowly released, but it is doubtful that this has any relevance under physiological conditions. Although intracellular Ca^^ reservoirs of muscle and nonmuscle cells are structurally very different, they share important functional analogies. Muscle and nonmuscle intracellular Ca^"*" stores are in fact endowed with three major components: (a) Ca^"^ release channels; (b) Ca^"^ binding proteins and (c) Ca^"^ pumping mechanisms. Two Ca^"^ release channels have been purified from intracellular organelles: the IP3 receptor, named after its physiological activator, IP3, and the ryanodine (RY) receptor, first identified in muscle cells and named after the ligand which was instrumental for its purification, the plant alkaloid ryanodine. The SR is present in both smooth and striated muscle cells. Its structural and functional organization is best exemplified in striated muscle where it takes the form of a complex network of tubules and cistemae wrapped around the myofibrils. The longitudinal SR is mainly composed of tubules and small cisternae, and terminal cistemae (TC). TC lie juxtaposed to the invaginations of the plasmalemma which are known as transverse tubules (TT). The TT are responsible for propagating the wave of depolarization. The anatomic unit represented by the TT and the two adjacent TC is highly specialized for fast Ca^"^ release in response to plasmalemmal depolarization. This unit is known as the triad (see Figure 3). No cytoarchitectural organization similar to the SR is present in smooth muscle or nonmuscle cells. Here, the Ca^'^-storing compartment can be identified in terms of the specialized portion(s) of the ER enriched in IP3 receptor, RY receptor, and Ca^'^-binding proteins (Volpe et al., 1988). Much of the information on the structure of the Ca^"^ stores in nonmuscle cells derives from studies on the cerebellar Purkinje neurons of the chicken (Volpe et al., 1990). These cells express a high content of IP3 receptor which is concentrated (500/um^ of membrane) in smooth-surfaced, flat cistemae which seem to be in direct continuity with typical rough ER (RER) cistemae. However, it is still unclear whether the entire population of IP3-rich cistemae is a functionally active intracellular Ca^"^ store. Purkinje neurons contain another subset of IP3 receptor-depleted, RY receptor-enriched ER vesicles which is supposed to also play a role in Ca^"^ storage and release. Mitochondria can also accumulate large amounts of Ca^Vbut only when extramitochondrial Ca^"*" rises above 5—10 JLIM, a condition tnat can be reached locally in the cytoplasm in the proximity of Ca^"^ channels of the plasma membrane or intracellular stores. Ca^"*" uptake by mitochondria, However, rather than having an ion homeostatic function, more likely modulates the activity of a number of dehydrogenases present in the mitochondrial matrix (Denton and McCormack, 1990). Mitochondrial Ca^"^ uptake is driven, via an electrogenic uniport, by the H"*"

8arcol«mma

I

Ca^* L'N^

Ca pump

y CSQ

f

LSR

DIHYDROPYRIDINE RECEPTOR Extracellular TT NH.

V^^^^^V^^V)^ COOH

RYANODINE RECEPTOR

Figure 3. Schematic picture of the SR of skeletal muscle cells. The SR of skeletal muscle cells is a complex intracellular system which controls the myoplasmic Ca^"" concentration. It consists of two portions: the longitudinal SR (LSR) and the terminal cisternae (TC). TC form typical structures, called triads, with plasmalemmal invaginations called transverse tubules (TT). In the triad, two TC and the Intervening TT are connected by structures called "feet." Feet have been Identified with the bulbous head of the ryanodine receptor (lower picture), the Ca^"^ release channel of the SR. The stalk (made of 4 to 10 alpha helices) of the ryanodine receptor Is present in the junctional face of TC, while the head abuts in the junctional space and interacts with the dihydropyrldine receptor of the TT. The dihydropyrldlne receptor senses TT depolarization and opens (probably by direct Interaction) the Ca^"" release channel which is the ryanodine receptor. The TC membrane facing TT is also called junctional face membrane OFM). The TC lumen contains the Ca^'^-binding protein calsequestrin (CSQ). The SR membrane is very rich in Ca^"^-pumping ATPase. 318

Ca'^'^ Homeostasis

319

electrochemical gradient established by the proton-pumping ATPase across the inner mitochondrial membrane. In this respect, mitochondrial Ca^"*" uptake closely resembles Ca^"^ influx through plasma membrane channels, i.e., a passive flux down the electrochemical gradient which, if electrochemical equilibrium were attained, and in the absence of Ca^"*"-extruding systems, would raise the matrix-free Ca^"*" concentration to 0.1 M (about 10^-fold the cytoplasmic concentration). In the intact cell and under resting conditions, matrix Ca^"^ concentration falls in the range of 50-200 nM (Rizzuto et al., 1992). It is noteworthy that massive accumulation of Ca^"*" into the mitochondrial matrix can occur in severely damaged cells, with severe consequences for mitochondrial function and structure, such as uncoupling of oxidative phosphorylation and precipitation of Ca^"^ phosphate. Ca^"^ is also contained in many other organelles, but unlike the SR, ER, and the mitochondria, exchange with the cytosol is extremely slow. Cs^'^ Release Channels: The IP3 and RY Receptors

The IP3 receptor was initially purified from the cerebellum, and later shown to be present, albeit to a lesser extent, in most other tissues tested. The receptor is a single polypeptide of molecular mass of 313 kDa which has numerous glycosylation sites and at least two potential phosphorylation sites for cAMP-dependent kinase (Meldolesi, 1992). Furthermore, the receptor itself has been shown to undergo autophosphorylation. Five isoforms of the IP3 receptor (IPaRla, IPsRlb, IP3R2, IP3R3, and IP3R4) arising from alternative splicing or the existence of separate genes and showing a different tissue distribution, have so far been identified. The IP3 receptor is assembled in a tetrameric structure, each subunit consisting of a large N-terminal region which abuts in the cytoplasm and comprises the IP3-binding site, and a C-terminal domain which contains the membrane-spanning domains (probably six) that form the Ca^"^ channel and anchor the protein to the ER membrane. The Ca^'^-release channel of the SR, the RY receptor, is a large protein composed of four equal subunits (homotetramer) which share many structural and functional homologies with the IP3 receptor (Sorrentino and Volpe, 1993). The two isoforms (RYRl and RYR2) of the RY receptor isolated from skeletal and cardiac muscle share about 66% identity. A third RY receptor (RYR3), different from, but about 70% homologous to, RYRl and RYR2, has been identified in several nonmuscle tissues. Thus, RY receptors, in the past considered to be muscle specific, are widely expressed in nonmuscle cells. The RY receptor is a large molecule (MW about 560 kDa) which, in a tridimensional reconstruction from negative-stain electron microscopy, reveals a four leaf clover (quatrefoil) structure. The N-terminal domain forms a large bulbous head (also known as "foot") that projects into the junctional space between the TC and the TT, while the C-terminal region comprises the membranespanning domains that contribute to the formation of the Ca^"^ channel.

320


Organelle Ct^^-Binding Proteins The ER and SR are able to accumulate large amounts of Ca^"*" (up to several mM in the TC of skeletal muscle) primarily because of their high content in Ca^'^-binding proteins, among which the most abundant and best characterized are calsequestrin and calreticulin, which are typical of muscle SR and nonmuscle cell ER, respectively (Lytton and Nigam, 1992). Calsequestrin exists as at least two isoforms, one specific for fast-twitch skeletal muscle and the other for cardiac muscle, and is defined as a "high capacity, medium/low affinity Ca^"^-binding protein." Its Kd for Ca^"^ lies in the millimolar range and the number of binding sites is 30-50 mol/mol protein. In striated muscle, calsequestrin is exclusively localized in the lumen of the TC, in close proximity to the Ca^"*" release channels. This strategic intraluminal distribution of calsequestrin helps to concentrate Ca^"^ near the release sites, thus increasing the speed of Ca^"^ release. Calreticulin, a glycoprotein of molecular mass (from cDNA) of 47 kDa, was identified in the early 1970s, and then rediscovered several times and given a variety of names. The name "calreticulin" was chosen to stress its role as a calcium-binding protein localized to the endoplasmic/sarcoplasmic reticulum membranes. Calreticulin possesses one high (IQ in the jiM range) and about 20-50 low-affinity (IQ = 1-4 mM) Ca^'^-binding site/mol of protein. Organelle Ca"^"^-Pumping Mechanisms Ca^"^ uptake into the intracellular stores is a vectorial process which translocates this ion fi-om a low (the cytosol) to a high (lumen of the organelles) [Ca^"^] compartment at the expense of energy. Energy is afforded by a family of specialized ATPases (Ca^"^ pumps) which shareftinctionaland structural homology (Carafoli and Chiesi, 1992). These proteins are now referred to as sarco/endoplasmic reticulum Ca^"^-ATPases (SERCAs). Three different genes encoding SERCAs have been described so far: SERCA 1 is expressed in fast-twitch skeletal muscle and has one adult and one neonatal isoform (SERC A la and SERC A lb, respectively), generated by alternative splicing; SERCA 2, which is also expressed in two variants (SERCA 2a and SERCA 2b) predominates in cardiac and slow-twitch muscle, while SERCA 3 is expressed in several nonmuscle cells. The prototype of SERCAs is SERCA la, i.e., the SR Ca^'^Mg^'^-ATPase of skeletal muscle. In the SR, the Ca^"^-Mg^"*'-ATPase is the most abundant protein, since in the longitudinal SR it may represent up to 90% of total protein content. It is a single protein of molecular mass of 110 kDa, with a high affinity for Ca^"*" (IQ 0.5 jLiM) and a Ca^"*':ATP stoichiometry of 2:1. SERCA la is an asymmetrical transmembrane protein with a short lumenal domain, 10 membrane-spanning helices, and a bulky cytosolic head. This latter portion of the molecule is supposed to contain the ATP-binding domains and an asparagine residue which undergoes

Ca^^ Homeostasis

321

phosphorylation upon Ca^^ binding. ER-SERCAs also have a MW of approximately 110 kDa, but are incompletely characterized due to the difficult purification procedure. How Is Ca^^ Released From Intracellular Stores?

Since intracellular Ca^^ stores are physically separated from the plasma membrane, the question arises as to how membrane acting stimuli (depolarization or receptor-directed agonists) can induce Ca^"^ release. This problem, which has attracted the interest of numerous investigators for several years, is now beginning to be solved and the underlying mechanism revealed in its molecular details. Two mechanisms couple plasma membrane stimulation to Ca^"^ release from intracellular stores: mechanical coupling, and chemical coupling (Ca^'^-induced Ca^"^ release and IP3-mediated Ca^^ release). Historically, the most thoroughly investigated model has been Ca^"*" release from the TC of the SR. Ca^^ release at the triadic junction occurs via two main mechanisms: one involves mechanochemical coupling and the other chemical coupling. Mechanochemical coupling occurs in skeletal muscle, where the RY receptor present in the junctional membrane of the TC couples structurally and functionally with the L-type VOCs (DHP receptors) existing in the TT of the sarcolemma. This gives rise to an anatomical structure known as "feet" (Rios and Pizarro, 1991). VOCs are thought to function as voltage sensors which physically transfer TT depolarization to the RY receptor. In contrast, the cardiac muscle RY receptor is activated by Ca^"^ itself via a mechanism known as "Ca^'^-induced Ca^"^ release" (CICR) (Fabiato, 1983). According to this mechanism, Ca^"^ influx across the sarcolemma increases myoplasmic Ca^"*" to a concentration (several juM) sufficient to trigger an explosive burst of Ca^"*" release through the cardiac RY receptor. In contrast, larger rises (mM range) are inhibitory and therefore shut down the channel. Sensitivity to a diffusible messenger such as Ca^"^ explains the mechanism of activation of cardiac RY receptors which are located away from the sarcolemma, in the so-called tubular SR. Both in skeletal and cardiac muscles, the Ca^"^ necessary for muscle contraction comes primarily from the SR. Quantitation of the total SR Ca^^ gives a value of about 8 mM, of which about 3 mM is reckoned to be free in the lumen, i.e., in the range of the K^ of the major segregated Ca^'^-binding protein, calsequestrin. Under resting conditions, calsequestrin is thought to be about 76% saturated with Ca^"*". During a single twitch lumenal Ca^"^ concentration does not change significantly, thus calsequestrin saturation drops only slightly to 72%, and only after a tetanus does lumenal Ca^"*" decrease to below 1 mM. Calsequestrin saturation thus decreases significantly to about 50% of the initial value. After release, Ca^"^ is rapidly pumped back into the SR SERCA present in the membrane of the longitudinal SR and in the TC, whence another cycle of excitation-Ca^"^ release-contraction can start.

322


In sharp contrast, none such rigid and precise organization is found in cardiac and smooth muscle, and in nonmuscle cells, where the Ca^"^ stores are identified with specialized regions of the ER (calciosomes). However, these regions do not seem to establish physical continuity with the plasma membrane (Meldolesi et al., 1990). In nonmuscle cells the only release mechanism known to occur is chemical coupling via IP3 and probably Ca^"*". The IP3 receptor responds to IP3 binding by increasing the opening probability of the channel with reported conductance values between 8 and 26 pS. Though the IP3 receptor is a tetramer, it is not clear whether binding of a single IP3 molecule to one of the subunits is sufficient to open the channel or all four subunits have to be ligated. The IP3 receptor is also sensitive to Ca^^, which is believed to function as a co-agonist with IP3. The Ca^^-releasing activity of IP3 is drastically affected by the [Ca^^^Ji concentration, showing a maximum at about 300 nM [Ca^"*"]i, and then declining. Ca^"*" sensitivity has been invoked to explain the all-or-none property of Ca^"*^ release through the IP3 receptor, as it is thought that above a given threshold Ca^"^ stimulates its own release. Very recently a role in Ca^"^ release from intracellular stores has been suggested for the nucleotide derivative cyclic ADP ribose (cADPR) (Galione, 1993). This agonist has been shown to activate the RY receptor at very low concentrations (jiM or even lower). cADPR is generated in many cell types by two enzymes, one of which is associated with the plasma membrane. However, proof of rapid generation of cADPR after receptor activation is still lacking. Intracellular Ca^"^ Oscillations: A New Signaling Code Ca^^ sensitivity of intracellular Ca^'^-releasing channels is probably the basic property which explains the intriguing observation that many cells stimulated with Ca^'^-mobilizing agonists undergo repetitive oscillations in [Ca^"^]], whose frequency is modulated by both the agonist and extracellular Ca^"^ concentration. These pulsatile [Ca^"*"]} increases have both a temporal and spatial aspect: not only [Ca^"*"]; undergoes repetitive spikes but very often Ca^"*" spiking originates in specific initiation loci wherefrom a Ca^"^ wave propagates throughout the cell (Cobbold and Cuthbertson, 1990). A dramatic example of such a spatiotemporal pattern of oscillations is observed in oocytes of different origin undergoing fertilization. In the eggs of the african toad Xenopus laevis, fertilization induces Ca^"^ release in several intracellular foci (hot spots), thus generating circular waves that annihilate as they merge. It is not completely clear how the Ca^"*" waves are generated, but growing consensus favors models based either on Ca^"*"-stimulated formation of IP3 (Meyer and Stryer, 1991), or CICR (Berridge, 1990), or a combination of the two. The physiological meaning of oscillatory [Ca^"^], increases is not understood. However, it is speculated that an intracellular message can be coded by either a continuous variation in the cytoplasmic concentration of a mediator, or all-or-none repetitive discrete changes in its concentration. In the first case the information

Ca^^ Homeostasis

323

content will reside only in the amplitude of the Ca^"*" change, while in the latter also in the frequency of the spikes. Furthermore, occurrence of spatially-restricted [Ca^"*']i elevations, which may or may not propagate, depending on the intensity of the stimulus and the sensitivity and type of Ca^"^ channels in the intracellular stores, constitutes the structural and anatomical basis for localized cellular responses. Cytoplasmic Ca^^-Binding Proteins

The Ca^"^ homeostatic mechanism ultimately serves two related functions. First, the maintenance of a low [Ca^"^]i essential for cell survival. And second, allowance of controlled changes of [Ca^^^Jj in response to plasma membrane receptor or channel activation. The targets of [Ca^"^]i changes reside in the cytoplasm and are represented by Ca^"^-binding proteins. Since all eukaryotic cells maintain very low (100-200 nM) [Ca^"^]i, it can be predicted that cytosolic Ca^'*'-binding proteins should have affinities in the nM-low |LIM range. However, many exceptions to this rule exist (e.g., protein kinase C), and it should be stressed that efficient buffering of Ca^"^ can be provided not only by high affinity sites, but also by low affinity sites if present in high concentration. Cytosolic Ca^"^-binding proteins can be divided into two major groups: the Ca^"^ buffers, i.e., proteins whose only known function is that of buffering Ca^"^ and the Ca^"*" sensors, i.e., proteins with modulatory activity on cell functions. The Ca^^ Buffers

The most representative members of this group are parvalbumin and related proteins. Parvalbumins are a group of homologous proteins with molecular masses between 9-13 kDa. They were defined as "low-molecular mass albumins" (hence the name parvalbumins,/7arv(3 being Latin for small). Parvalbumins were originally isolated from frog muscle and later found in many other tissues, not just striated muscle, of fish and mammals. Parvalbumins were the first Ca^^-binding proteins to be crystallized and were instrumental in revealing the basic structure of high affinity Ca^"^-binding sites. Carp parvalbumin contains six helical domains (A—F) linked by loops. The Ca^'^-binding sites are located in the loops between helices C—D and E—F (Figure 4). The sequence constraints of the E—F domains (the so-called E-F hand) are so rigid that a protein can be classified in this family of Ca^'^-binding proteins simply by knowing the amino acid sequence (Weeds and MacLachlan, 1974). The two Ca^'^-binding sites have affinities for Ca^"^ in the range 0.1-4 |LiM, and are also reported to bind Mg^"^. The only known role of parvalbumins is to function as soluble cytoplasmic Ca^"*" buffers, especially in fast-twitch fibers, where there seems to be a good correlation between parvalbumin content and speed of muscle relaxation. Several other proteins, whose only known function is to buffer Ca^^ with high affinity, have been

324


helix F Figure 4. The E-F hand (modified from Branden, C. & Tooze, J. (1991). In: Introduction to Protein Structure, Garland Publishing Inc., New York). Most Ca^'^-binding proteins possess in their structure a helix-loop-helix motif which contains the Ca^"^binding site. This helix-loop-helix motif is called E-F hand because its tridimensional arrangement is mimicked by a hand in which the forefinger represents helix E (heavy shadowing), the flexed middle finger represent the loop containing the Ca^^-binding site (light shadowing) and the thumb the F helix (medium shadowing). described, among which is the S-100 intestinal Ca^"^-binding protein family, oncomodulin, a parvalbumin-like protein specifically found in tumor tissues and the calbindin family. All these proteins contain two or more E-F hand motifs. The Cd?* Sensors The most famous protein in this group is calmodulin. This is a small (16 kDa) acidic protein present in large amounts in the cytoplasm of all eukaryotic cells

Cai^^ Homeostasis

325

(Cheung, 1980). The three-dimensional structure was determined in 1985 and the protein looks like a dumb-bell, with two (N- and C-terminal) globular portions linked by an alpha helix. Calmodulin contains four E-F hand motifs, two in each half of the molecule, with Ca^^ affinities which are still being debated. According to one model, Ca^"^ binding exhibits no cooperativity, thus all the four binding sites have similar affinities (K^ = 10 |LIM). According to another model, strong cooperativity exists between the two sites of each half molecule; furthermore, the sites located within the C-terminal half have been suggested to have higher affinity than those located in the N-terminal portion. At present, it is impossible to decide between these two alternative models. Calmodulin acts as a Ca^"^ sensor, thus conferring Ca^"^ sensitivity to many enzymes which would be intrinsically Ca^'^-insensitive. Kinetic models for activation of Ca^"^-calmodulin-dependent enzymes are very complex. In this context, it suffices to recall the following features: (a) the Ca^'^-bound form of calmodulin has an affinity for the target enzyme four orders of magnitude higher than the Ca^^-free form; (b) the complex between Ca^'^-free calmodulin and the target enzyme is inactive; and (c) activation by Ca^'^-calmodulin causes de-repression rather than direct activation of the target enzyme. According to this last concept, the calmodulin-binding domain, in the absence of calmodulin, prevents access of the natural substrate to the enzyme's active site; binding of Ca^'^-calmodulin removes the block. This peculiar mechanism of activation explains why limited proteolysis of Ca^'^-calmodulin-dependent enzymes removes calmodulin sensitivity, and irreversibly activates the enzyme. Ca^'^-calmodulin-dependent reactions can be activated in at least two ways: one is by increasing the [Ca^^^Jj and the other by increasing the calmodulin concentration. Although this latter mechanism may seem speculative, it should not be neglected that intracellular calmodulin concentration is known to change not in different cell types, but also in different steps of the cell cycle. While activation by [Ca^"^]; increases provides a quick and easily reversible means of activation of the target enzyme, activation by increasing the calmodulin concentration might have long lasting effects on cell functions. A closely related homologous protein of calmodulin is troponin C (molecular mass 18 kDa), which mainly differs from calmodulin in being part of a heterotrimeric complex (together with troponin T and I) which is an integral part of the thin-muscle filament, while calmodulin is a soluble monomeric protein which is ubiquitous. The basic structure of troponin C closely resembles that of calmodulin in having two globular domains each containing two E—F hand motifs.

CONCLUSIONS Because of the recent dramatic developments in cell and molecular biology we are becoming increasingly aware that the eukaryotic cell is an incredibly complex living system whose level of structural and metabolic organization is as complex

326

F. or VIRGILIO, D. PIETROBON, and T. POZZAN

as that of much larger muhicellular organisms. While multicellular organisms have functionally and anatomically separated compartments which communicate by means of hormones, growth factors or neurotransmitters, the cell has many different organelles in compartments which communicate between themselves and the plasma membrane by means of diffusible messengers, among which calcium plays a pivotal role.

ACKNOWLEDGMENTS The authors wish to acknowledge the invaluable help of Drs. Davide Ferrari and Paola Chiozzi and Miss Mariella Bergami. The support of the Italian Ministry for Scientific Research, the National Research Council (Target Projects BTBS and ACRO), the Italian Association for Cancer Research (AIRC) and Telethon of Italy is gratefully acknowledged.

REFERENCES Bean, B.P. (1989). Classes of Ca^"" channels in vertebrate cells. Annu. Rev. Physiol. 51, 367-384. Bean, B.P. & Friel, D.D. (1990). ATP-activated channels in excitable cells. In: Ion Channels (Narahishi, T., ed). Vol. 2, pp. 169-203, Plenum, New York. Berridge, M.J. (1993). Inositol trisphosphate and Ca^"*^ signalling. Nature (London) 361, 315-325. Berridge, M.J. (1990). Calcium oscillations. J. Biol. Chem. 265,9583-9586. Bertolino, M. & Llinas, R.R. (1992). The central role of voltage-activated and receptor operated channels in neuronal cells. Ann. Rev. Pharmacol. Toxicol. 32, 399-421. Blaustein, M.P. (1988). Calcium transport and buffering in neurons. Trends Neurosci. 11,438-448. Carafoli, E. (1992). The Ca^"^ pump of the plasma membrane. J. Biol. Chem. 267,2115-2118. Carafoli, E. & Chiesi, M. (1992). Calcium pumps in the plasma and intracellular membranes. Curr. Top. Cell Reg. 32, 209-241. Catterall, W.A. (1991). Functional subunit structure of voltage-gated calcium channels. Science 253, 1499-1500. Cheung, W.Y. (1980). Calmodulin plays a pivotalVole in cellular regulation. Science, 207, 19-27. Cobbold, P.H. & Cuthbertson, K.S.R. (1990). Calcium oscillations: Phenomena, mechanism and significance. Sem. Cell Biol. 1, 311-321. Denton, R.M. & McCormack, J.C. (1990). Ca^"^ as a second messenger within mitochondria of the heart and other tissues. Annu. Rev. Physiol. 52,451-466. Dubyak, G.R. (1991). Signal transduction by P2 purinergic receptors for extracellular ATP. Am. J. Respir. Cell Mol. Biol. 4,295-300. Fabiato, A. (1983). Calcium-induced release of calcium from the cardiac sarcoplasmic reticulum. Am. J. Physiol. 245, C1-C4. Galione, A. (1993). Cyclic ADP ribose: A new way to control calcium. Science 259, 325-326. Hess, P. (1990). Ca^"*" channels in vertebrate cells. Ann. Rev. Neurosci. 13, 337-356. Kahn, A.A., Steiner, J.P., & Snyder, S.H. (1992). Plasma membrane inositol 1,4,5-trisphosphate receptor of lymphocytes: Selective enrichment in sialic acid and unique binding specificity. Proc. Natl. Acad. Sci. USA 89, 2849-2853. Kaupp, U.B. & Koch, K.W. (1992). Role of cGMP and Ca^"^ in vertebrate photoreceptor excitation and adaptation. Annu. Rev. Physiol. 54, 153-175. Lytton, J. & Nigam, S.K. (1992). Intracellular calcium: Molecules and pools. Curr. Biol. 4,220-226. Meldolesi, J. (1992). Multifarious IP3 receptors. Curr. Biol. 2, 393-394. Meldolesi, J., Madeddu, C, & Pozzan, T. (1990). Intracellular Ca^"^ storage organelles in non muscle cells: Heterogeneity and functional assignment. Biochim. Biophys. Acta 1055, 130-140.

Cs?^ Homeostasis

327

Meldrum, B. & Garthwaite, J. (1990). Excitatory aminoacid neurotoxicity and neurodegenerative diseases. Trends Pharmacol. Sci. 11, 290-297. Meyer, T. & Stryer, L. (1991). Calcium spiking. Ann. Rev. Biophys. Biophys. Chem. 20, 153-174. Mintz, I.M., Adams, M.E., & Bean, B.P. (1992). P-type calcium channels in rat central and peripheral neurons. Neuron 9, 85-95. Moriyoshi, K., Masu, M., Ishi, T., Shigemoto, R., Mizuno, N., & Nakanishi, S. (1991). Molecular cloning and characterization of the rat NMDA receptor. Nature (London) 354, 31-37. Nelson, T.E. (1983). Abnormality in calcium release from skeletal sarcoplasmic reticulum of pigs susceptible to malignant hyperthermia. J. Clin. Invest. 72, 862-870. Penner, R., Fasolato, C, & Hoth, M. (1993). Calcium influx and its control by calcium release. Curr. Opin. Neurobiol. 3, 368-374. Pietrobon, D., Di Virgilio, F., & Pozzan, T. (1990). Structural and functional aspects of calcium homeostasis in eukaryotic cells. Eur. J. Biochem. 193, 599-622. Pozzan, T., Rizzuto, R., Volpe, P., & Meldolesi, J. (1994). Molecular and cellular physiology of intracellular Ca^"^ stores. Physiol. Rev. 74, 595-636. Randriamampita, C. & Tsien, R.Y. (1993). A diffusible messenger released by emptying of intracellular Ca^"*" stores. Nature (London) 364, 809-814. Rasmussen, H. (1981). In: Calcium and cAMP as Synarchic Messengers. John Wiley & Sons, New York. Rios, E. & Pizarro, G. (1991). Voltage sensors of excitation-contraction coupling in skeletal muscle. Physiol. Rev. 71,849-908. Rizzuto, R., Simpson, A.W.M., Brini, M., & Pozzan, T. (1992). Rapid changes of mitochondrial Ca^"^ revealed by specifically targeted recombinant aequorin. Nature (London) 358, 325-327. Snutch, T.P. & Reiner, P.B. (1992). Ca^*" channels: Diversity of form and function. Curr. Opin. Neurobiol. 2, 247-253. Sorrentino, V. & Volpe, P. (1993). Ryanodine receptors. How many, where and why. Trends Pharmacol. Sci. 14,98-103. Trump, B.F. & Berezesky, I.K. (1992). The role of cytosolic Ca^"*" in cell injury, necrosis and apoptosis. Curr. Opin. Cell Biol. 4, 227-232. Tsien, R.W., Ellinor, P.T., & Home, W.A. (1991). Molecular diversity of voltage-dependent Ca^"^ channels. Trends Pharmacol. Sci. 12, 349-354. Turner, T.J., Adams, M.E., & Dunlap, K. (1992). Calcium channels coupled to glutamate release identified by w-Aga-IV. Science 258, 310-313. Volpe, P., Alderson-Lang, B.H., Madeddu, L., Damiani, E., Collins, J.H., & Margreth, A. (1990). Calsequestrin, a component of the inositol 1,4,5-trisphosphate-sensitive Ca^"^ store of chicken cerebellum. Neuron 5, 713-721. Volpe, P., Krause, K.-H., Hashimoto, S., Zorzato, F., Pozzan, T., Meldolesi, J., & Lew, D.P. (1988). "Calciosome," a cytoplasmic organelle: The inositol 1,4,5-trisphosphate-sensitive Ca^"*" store of non-muscle cells. Proc. Natl. Acad. Sci. USA 85, 1091-1095. Weeds, A.G. & MacLachlan, A.D. (1974). Structural homology of myosin alkali light chains, troponin C and carp calcium binding protein. Nature (London) 252, 646-649. Williams, M.E., Brust, P.F., Feldman, D.H., Patthi, S., Simerson, S., Maroufi, A., McCue, A.F., Velicelebi, G., Ellis, S.B., & Harpold, M.M. (1992). Structure and functional expression of an co-corotoxin-sensitive human N-type calcium channel. Science 257, 389-395. Wisden, W. & Seeburg, P.H. (1993). Mammalian ionotropic glutamate receptors. Curr. Opin. Neurobiol. 3,291-298.


Chapter 13

Roles of Polyamines in Cell Biology NIKOLAUS SEILER

Introduction Definitions History General Characteristics Metabolism The Polyamine Metabolic Cycle in the Vertebrate Organism Regulatory and Nonregulatory Enzymes Polyamine Transport Interaction with Macromolecuies Free and Bound Polyamines Interactions With DNA Interactions With rRNA and tRNA Interactions With Proteins Polyamines and Growth Polyamines and the Cell Cycle Polyamine Deficient Mutants Inhibitors of Polyamine Biosynthesis Consequences of Polyamine Depletion Development and Differentiation Macromolecular Synthesis Covalent Binding of Polyamines to Proteins Polyamines in Diagnosis and Therapy Summary

Principles of Medical Biology, Volume 4 Cell Chemistry and Physiology: Part II, pages 329-348 Copyright © 1996 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-806-4 329

330 330 330 331 331 331 332 336 336 337 337 337 338 339 339 340 340 340 343 344 345 346 346

NIKOLAUS SEILER

330

INTRODUCTION Definitions

The aliphatic di-, tri-, and tetramines, which are the topic of this chapter, are commonly designated "polyamines." From a chemical point of view, this designation is incorrect because the natural polyamines are, in fact, small molecules. Figure 1 shows their structure and includes the natural acetyl derivatives and amino acids that are formed from the polyamines by oxidative deamination. The term polyamine is used to designate putrescine (1,4-butanediamine), spermidine (N-(3-aminopropyl)-l,4-butanediamine), and spermine (N,N'-bis-(3-aminopropyl)-1,4-butanediamine). Cadaverine (1,5-pentanediamine) and 1,3-propanediamine also occur in nature and form homologues of spermidine and spermine. They are not discussed, however, because of their lack of importance in vertebrates. History The formation of spermine phosphate crystals in aged human semen was observed as early as 1678 by A. van Leeuwenhoek. It was rediscovered by the POLYAMINES

ACETYLPOLYAMINES

TERMINAL

CATABOLITES

1,3-PROPANIOIAMINE 0 0

ACETYLPUTRESCINE CADAVERINE

HN J

N'-ACETYLSPERMIOINE

4-AMINOBUTYRlC ACID

PYRR0LIDIN-2-0NE

PUTREANINE H2N>.,,^S^N'^

^ ^^^^

N'-ACETYLSPERMIOINE

Hg

v J

ISOPUTREANINE

LACTAM

-V^-v)"'

ACETYLISOPUTREANINE LACTAM

HjC^ N -ACETYLSPERMINE

N'-(2-CARB0XYETHYL) SPERMIDINE 0

H,C-^N'""">'^N-' ^ H H N'.N'^ OtACETYLSPERMINE

SPERMIC ACID

Figure 1, Structural formulae of the polyamines and their derivatives.

Polyamines in Cell Biology

331

French chemist Vauquelin in 1791, and again by Charcot, the famous teacher of Freud, in 1853. The chemical structures of the polyamines were definitively established by Rosenheim in 1924. It took another 25 years before the study of the biochemistry of the polyamines began, and a similarly long period before the polyamines attracted more general attention. General Characteristics The polyamines were neglected until recent times by most biochemists and molecular biologists in spite of a remarkable history, and an awareness of several known facts suggestive of functions of basic importance for these biogenic amines: 1. The polyamines appeared at an early stage of evolution and were conserved since then. They occur in all cells. Prokaryotes contain putrescine and spermidine, while eukaryotic cells contain spermine in addition to putrescine and spermidine. Archebacteria, algae, and some higher plants may contain, in addition, homologues and analogues of putrescine, spermidine, and spermine. 2. The polyamines are formed by rather demanding synthetic reactions. Their cellular concentrations are intricately regulated and adapted to physiological needs. 3. Depletion of cellular spermidine prevents proliferation of eukaryotic cells and decreases the growth rate of prokaryotes.

METABOLISM The Polyamine Metabolic Cycle in the Vertebrate Organism Most vertebrate cells produce and require polyamines. Red blood cells do not synthesize polyamines but are able to accumulate and bind polyamines. Ornithine is the exclusive precursor of putrescine, from which it is formed by decarboxylation, as indicated in Figure 2. In bacteria-and plants, and presumably in some parasites, additional reactions exist. One starts with arginine, which is first decarboxylated to agmatine. Agmatine, in turn, is hydrolyzed to putrescine and urea. In order to produce the aminopropyl residues that are required for the formation of spermidine and spermine, methionine first reacts with ATP to form S-adenosylmethionine (AdoMet). This is decarboxylated to decarboxy-S-adenosyl-methionine (dAdoMet) (Figure 2), a compound which donates the aminopropyl residues to putrescine to form spermidine, or to spermidine to form spermine (Figure 3). Spermidine synthase and spermine synthase are the two enzymes which catalyze the transfer of the aminopropyl residues. The second product of dAdoMet, 5'-

332

NIKOLAUS SEILER

Ornithine Decarboxylase

Ornithine

Putrescine

CO2

S-Adenosyl-L-methionine ( AdoMet)

decarboxylated AdoMet ( dAdoMet)

Figure 2. The biosynthetic decarboxylases of polyamlne metabolism, ornithine decarboxylase, and S-adenosyl-L-methionine decarboxylase.

methylthioadenosine (MTA), is formed in equimolar amounts during the production of spermidine and spermine. It can be reused for the formation of ATP. As depicted in Figure 3, spermidine may also be formed from spermine, and putrescine from spermidine. In this case, the monacetyl derivatives of spermidine and spermine are first generated, whereby acetylCoA is the acetyl group donor. The acetyl derivatives are substrates of polyamine oxidase. This flavin-adenine-dinucleotide (FAD)-dependent enzyme splits the N^-acetylpolyamines into an aldehyde (3-acetamidopropanal) (Figure 4), which represents the part of spermidine and spermine that originates from methionine. The spermidine and putrescine formed by this degradative process can be reutilized for de novo polyamine synthesis. Thus, polyamine metabolism is a cyclic process which allows the transformation of putrescine into spermidine and spermine, and vice versa, in accordance with physiological requirements. The metabolic cycle is essential for the regulation of polyamine turnover. Another well-known reaction of the polyamines is their oxidative deamination by serum amine oxidase and by diamine oxidase (Figure 4). The small intestines and the serum during pregnancy are especially rich sources of diamine oxidaselike activities. The amino acids shown in Figure 1 are thefinalproducts of the oxidative degradation of putrescine, spermine, and spermidine by these diamine oxidase-related enzymes. Regulatory and Nonregulatory Enzymes

Three enzymes, ornithine decarboxylase (ODC), S-adenosyl-methionine decarboxylase (AdoMetDC) (Figure 2), and an acetyltransferase (acetylCoA:polyamine

I

UJ

< X 2 -I

is

^'

£ OuJ

ii

HO LUX

?-l® J "

Ou

> 0. X s

ro

in

O in

00

1 1 +1 +1 +1 +1 rs. o o o m i n o CJ^ in in 00 1

0 C

DO C 3 -J

in CN

U^

c

^ ^

a.

03

u c

0)

{A

CO rsi

2 Q-

Q;

in rsj CN

1 +1 +1 +1 +1

U

c/)

E

trt

E t/5

"re

D

> :^ re in

c

U

•o 0)

E c 2 a;

^ ^1^

a; •E

.*i

£ 6

(13

^ re cu

ift

c re

o 00 cy^

^ ^ ^ ¥ E iS

Si

1

336

NIKOLAUS SEILER

In addition to trophic stimuli, the polyamines regulate their intracellular concentrations. A decrease in cellular polyamines leads to induction of the biosynthetic decarboxylases, which, in turn, are repressed as polyamines accumulate. Polyamines regulate their own formation and degradation not only by affecting the formation of mRNA (at the transcriptional level), but also by affecting the rate at which enzyme proteins are formed at the level of translation of the mRNA sequence into a peptide chain. Moreover, the polyamines have posttranslational effects. For example, in vertebrates, putrescine not only enhances the formation of AdoMetDC from a larger pro-enzyme, but also acts as an allosteric regulator of AdoMetDC. Sustained high concentrations of putrescine may induce the formation of a protein, called ODC antizyme, which binds to ODC and inactivates it. In order to decrease cellular polyamine concentration, the induction of the acetyltransferase (cS AT) is required. For the down-regulation of the biosynthetic enzymes, their rate of degradation is enhanced. Antizyme most probably is involved in the enhanced degradation of ODC. In sharp contrast to the regulatory decarboxylases and the acetyltransferase, which occur at low activity, spermidine and spermine synthase are present in cells and tissues in very high concentrations (Table 1). They are stable proteins with biological half-lives of several days. Their activity determines the maximum possible rate of polyamine formation. Polyamine oxidase is also a stable enzyme. Polyamine Transport In addition to regulation by synthesis and degradation, the uptake and release of the polyamines are important features of cellular polyamine regulation. In numerous cell types putrescine, spermidine, and spermine share the same active, Na'^-activated transport system. The intracellular concentration of the free polyamines controls the uptake rate, a situation analogous to the control of synthetic rates. The uptake of polyamines from the circulation can replace intracellular biosynthesis (see Figure 3). The release by cells of polyamines in the form of acetylated products, or of putrescine, is the major elimination pathway in most tissues. In the intact animal, absorption of exogenous polyamines from the gastrointestinal tract and their urinary excretion is a state of affairs resembling cellular uptake and release.

INTERACTION WITH MACROMOLECULES Structurally, polyamines are flexible molecules with positive charges distributed along an aliphatic carbon chain. (Inorganic polycations, such as Mg^^ and Ca^"^represent pointlike charges.) At physiological pH, the amino groups of the polyamines are protonated. The positive charges enable the polyamines to form ionpairs with negatively charged molecules. Binding energy increases with the number of charges (putrescine < spermidine < spermine). Electrostatic interactions with DNA,


337

RNA, proteins, and negatively charged membrane constituents constitute the basis of most of the known functions of polyamines. Free and Bound Polyamines

Bound and free polyamines are in a state of dynamic equilibrium. Although the intracellular concentration of free polyamines is not yet measurable, there is no doubt that it represents only a small fraction of the total spermidine and spermine pool. The concentration of the free, not of the total, polyamines is the factor involved in the regulation of biosynthesis, degradation, uptake, and release. Acetylation moves a positive charge, thus leading to a decrease in electrostatic interaction with negatively charged binding sites. Polyamine acetylation is, therefore, one means of displacing polyamines from binding sites. Interactions With DNA

Spermine, with a binding constant of about 10^ M"^ interacts strongly with DNA. This stabilizes conformation and protects DNA from thermal denaturation and enzymatic hydrolysis. If 80-90% of the negative phosphate charge is neutralized by spermine, DNA collapses into compact structural forms. This "toroidal condensation" by polyamines probably has important implications for the organization of DNA in viral capsids, in nucleosome formation, and in chromosome condensation. A number of investigators have suggested that the conformational transition of B-DNA to Z-DNA is important for the control of DNA function. Through the use of a synthetic polynucleotide as a model (the heteropolymer poly(dG-m^dC)), it was shown that a spermine/nucleotide ratio of 1:40-50 is adequate to bring about the B-Z transition. While toroidal condensation may be explained by nonspecific electrostatic interactions, the facilitation of B-Z transition by spermine is presumably due to the binding of spermine at specific sites on the double helix. X-ray diffraction data for a double-stranded B-DNA dodecamer indicates the binding of one spermine per dodecamer. Interactions With rRNA and tRNA

Crystallographic studies of yeast tRNA^^® reveal the binding of two molecules of spermine per tRNA. In solution, however, several spermine molecules are found to bind to tRNA with high affinity. This interaction causes conformational changes in the anticodon loop. The structural transition caused by spermine may influence the codon-anticodon interaction on the ribosome and thus affect protein synthesis. Furthermore, it has been shown that binding of spermine to tRNA affects its binding

338

NIKOLAUS SEILER

to the appropriate amino-acyl tRNA synthetase. This is another site where polyamines are able to stimulate protein synthesis. Another notable feature of polyamine interaction is shown by studies of bacterial ribosomes. They indicate that the ribosomes contain a small amount of tightly bound spermidine, and that spermidine is able to partially replace Mg^"*" in promoting association of the 20 and SOS ribosomal subunits, a key step in peptide synthesis. Interactions With Proteins Polyamines play a role in actin polymerization and microtubule formation from tubulin. Briefly, microtubules are ubiquitous components of the cytoskeleton in eukaryotic cells. As will be recalled, the cytoskeleton is involved in cell motility and cell division. Purified tubulin, i.e., the globular polypeptide that is the building block of microtubules, polymerizes in the presence of GTP and 0.25 mM spermine (or 1 mM spermidine) to form microtubules. This polyamine-induced assembly of tubulin is a reversible process. Actin, a globular protein consisting of a polypeptide chain and one molecule of ADP or ATP, plays a key role in muscular contraction, as well as in the motility of nonmuscle cells. Actin is known to form double-stranded filaments which, on further association, leads to the formation of bundles. During mitosis, these actin bundles are disassembled to single filaments, but during telophase they reappear and form (together with other proteins) a contractile ring. Studies with spermidine and spermine show that they play the role of inducers of cytokinesis (cytoplasmic division). They do so at physiological concentrations. Additional lines of evidence strongly suggest a physiological role for polyamines in the formation of microtubules and actin filaments: (a) before cells enter mitosis, their polyamine concentration increases (see below); (b) in cells which are unable to synthesize polyamines, cytokinesis is found to be reduced; these cells show a defective cytoskeleton. Actin filaments and microtubules disappear; and (c) microinjection of spermine induces cytokinesis in amoeba 30-60 seconds later. Polyamine-protein interactions are not restricted to cytoskeleton formation. One has to assume the existence of very numerous interactions ranging from effects on chromatin structure to effects on intermediary metabolism. Many enzymes and receptors are changing their conformation, and with it their functional state by micromolar concentrations of spermine and spermidine (allosteric regulation). The most extensively studied examples of allosteric effects of polyamines are the regulation of the activity of membrane-bound acetylcholinesterase, and the glutamate receptor of the N-methyl-D-aspartate (NMD A) type. The latter has a specific binding site for polyamines, which is at present a target for the development of new drugs against the consequences of stroke and brain trauma. More and more evidence is accumulating for a role of the polyamines in Ca^^-signaling and signal transduction, i.e., the processes that are mediating exoge-


339

nous stimuli of hormones, growth factors, and others from the receptor on the cell surface to the interior of the cells.

POLYAMINES AND GROWTH Polyamines and the Cell Cycle

During each cycle, a cell doubles its structural and functional capacities. A major difference between cells that divide rapidly and those that divide slowly is the length of time they spend in the G1 phase of the cell cycle. This phase is assumed to consist of a succession of events that lead to the initiation of DNA replication during the S phase. Polyamine synthesis is one of the events that is stimulated during the Gl phase, as evidenced by the increase in ODC activity. This is shown in Figure 5. A second peak of ODC activity is seen prior to cell division. Similar changes in AdoMetDC activity have also been observed. The changes in ODC and AdoMetDC activity are followed by the accumulation of putrescine, spermidine, and spermine. Taken together, these observations indicate phase-specific changes in the requirement of polyamines. The picture, then, which emerges is that the first surge of

Gl

z o

H O

Principles of Medical Biology Vol 4 Part II: Cell Chemistry and Physiology

Principles of Medical Biology Vol 4 Part II: Cell Chemistry and Physiology

Principles of Medical Biology Vol 4 Part I: Cell Chemistry and Physiology

Principles of Medical Biology Vol 4 Part IV: Cell Chemistry and Physiology

Principles of Medical Biology Vol 10A: Molecular and Cell Endocrinology

Cell Chemistry and Physiology

Immunobiology (Principles of Medical Biology) (Vol 6)

Principles of Medical Biology Vol 9A: Microbiology

Medical Cell Biology, Third Edition (MEDICAL CELL BIOLOGY (GOODMAN))

Cell physiology

The Alkaloids: Chemistry and Physiology Volume II

Biology 4 vol

Principles of Medical Biology Vol 2: Cellular Organelles

Comprehensive Natural Products II: Chemistry and Biology: Volume 4

Cell Physiology (LANGE Physiology Series)

Learning science : Part 4 : Biology and life

Textbook of Medical Physiology

Methods in Enzymology Vol 327: Applications of Chimeric Genes and Hybrid Proteins, Part B: Cell Biology and Physiology

Carbohydrate Chemistry, Biology and Medical Applications

Carbohydrate Chemistry, Biology and Medical Applications

Medical Physiology

Medical Physiology

Review of Medical Physiology

Principles of Medical Biology Vol 15: The Liver Biology in Disease, Hepato - Biology in Disease

Medical Physiology

Medical Physiology: Principles for Clinical Medicine

Principles of Insect Physiology

Biology, Vol. 4 (Pr-Z)

Principles of Bone Biology 2 Vol. Set

Principles of Cell Proliferation

Principles of Medical Biology Vol 4 Part II: Cell Chemistry and Physiology