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 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.

This Page Intentionally Left Blank

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

Howard Hughes Medical Institute Vanderbilt University Nashville, Tennessee

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.

Howard Hughes Medical Institute Vanderbilt University Nashville, Tennessee

Woodford

This Page Intentionally Left Blank

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

This Page Intentionally Left Blank

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 ATP<->ADP 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, how^ever, 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—^it 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: ^Absolute 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—^and 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

J
i2

,., ,

Na.Ei ATP

NaE^-^P

NaJ2± E2.ATP*

/

\EI,ATP

,

„x

/\E2-P

Na in

•

\

E2

Figure 8. Mechanism of ATP use in the NaVK"^ pump, (a) Alternating access model for ATP driven transport of ions. The inward facing (Na"^ binding) form of the enzyme is designated conventionally as E^, and the outv^ard facing (K"^ binding) form as E2. Note that each chemical step (ATP binding, protein phosphorylation) produces a kinetically unstable species (*) which relaxes in the next stage of the cycle (c.f. Figure 6). T represents (non-covalently) bound ATP, P represents covalently bound phosphate, (b) Energetics of .the alternating access model. Energy is transferred from ATP (solid line) to uphill K"^ movement (decrease in K"^ affinity) (dashed line), and from an acyl phosphate to uphill Na"^ movement (decrease in Na"^ affinity) by enzyme conformational changes. Note that energy is used to change ion binding affinity, rather than to change protein conformation per se. For simplicity, the number of ions (2K% 3Na"^) moved in each conformational change is not shown.

19

DAVID A. HARRIS

20

ated with changes in ATP binding to the protein, not with chemical changes; (b) ATP binding provides energy for a decrease in binding affinity of a second hgand (actin, above); and (c) this energy is transmitted via a protein conformational change. Applying these principles to operation of the NaVK"^ pump gives the model in Figure 8b. The analogy is strongest on the top row. K"^ binds very tightly to E2, so it can be picked up at the low concentration outside the cell (low [K]out)- To release it at high [K]in, binding must be weakened, and this is accomplished by binding ATP tightly. Since ATP binds tightly only to Ei, the enzyme changes into the Ei conformation, consequently changing the orientation of its binding site. Coupling between energy release and ion movement thus occurs via conformational changes in the protein. In the second part of the cycle, a similar process occurs. Na"^ binds tightly to El while the aspartyl phosphate residue is destabilized (Ej-P); as the conformation of the protein changes, the aspartyl phosphate is stabilized (E2-P), and the binding energy released loosens Na"^ binding (Na"^ released at high [Na]out)As noted above, the chemical processes (EiATP ~> E p P ; E2P -> E2) are not accompanied by large energy changes and serve simply as triggers for the next stage in the cycle. Similar models can be derived for all types of ATP driven pumps discussed above. ATP in Biosyntheses

In tissues carrying out large amounts of biosynthesis (e.g., liver, exocrine tissues, rapidly dividing cells), a considerable fraction of cell ATP can be devoted to driving NH2 I

NH2 ATP

R - C - H ""v

I

C= 0 0

I

"^^

AMINO ACID

. R-C-H

I

PPi

'

V

"^^^

vR-C-H

C= 0 AMP O-p-O-Qdenosine 0* AMINO ACYL ADENYLATE

CH3

C=0

c=0 O-tRNA

CH3

(c H2)n CoA ppj

I

AMINO ACYL tRNA

CH3

(CH2)nATP C=0

NH2 tRNA

(c Hzln AMP

^=0

i

' Q

'

0"

O-P-O-Qdenosine

0-S-CoA

0" FATTY ACID

FATTY ACYL ADENYLATE

FATTY ACYL Coenzyme A

Figure 9. Role of ATP in generating coenzyme-bound intermediates for macromolecule biosynthesis.

Cellular ATP

21

biosynthetic processes. In the islets of Langerhans, for example, about 50% of ATP is used to drive biosyntheses. The role of ATP in biosynthesis is to activate a molecule for further reaction. Two classes of reaction can be distinguished, ATP as a Dehydrating Agent

As noted above, macromolecules are assembled by reacting component monomers with the elimination of water (condensation). ATP is an acid anhydride; it can therefore, in principle, drive condensation reactions. It must, however, specifically remove that water formed in the reaction, in the presence of a large excess of cell water. This requires the direct reaction of ATP with one of the reacting species. Typically, ATP will activate organic acids (amino acids, fatty acids), forming an adenylated intermediate (mixed acid anhydride). This reaction would be barely favorable if Pj were released; however, PPj is released, and this is rapidly hydrolyzed by cellular pyrophosphatase, ensuring complete reaction. O

II

(i) R-COO" + ATP -> R-CO-P-adenosine + PPj (ii) PPi+H20 -> 2?^ I O"

The adenylated intermediate is too unstable to release into free solution, and so the activated acid is transferred to a carrier molecule, tRNA in the case of amino acids and coenzyme A for fatty acids (Figure 9). However, the carboxylate group is still in an anhydrous state, and these species will react with the acceptor molecule (R-NH2 for amino acids, glycerol(-OH) for fatty acids) without further energy input. In the synthesis of polysaccharides, the dehydrating agent is a different nucleoside triphosphate, UTP. This replaces, in an analogous reaction, a sugar hydroxyl (alcohol) group forming the UDP-sugar compound, which can transfer the activated sugar on to the -OH of a growing polysaccharide chain. Activation of Leaving Groups

A number of biochemical reactions involve the displacement of a hydroxyl group by another group (e.g., an amino group). However, the -OH group is chemically a very poor leaving group, i.e., it shows a strong tendency to remain attached to its neighboring carbon atom. Phosphorylation of the hydroxyl group, using ATP, creates a much better leaving group -P04~, and thus makes this group susceptible to attack. Examples occur in the synthesis of glutamine from glutamate, and of CTP from UTP (Figure 10).

22

DAVID A. HARRIS

(CHzlzATP H-C-NH?

I

COO"

(CH2)2

0\t^,

AOP H-C-NHo

I

coo"

(CH2)2 Pj H-C-NH2

I

COO"

y -OUITAMTL FBOSPHATE

nKIDIME TBIFIiOSPiUTE

Figure 10.

0-P-O' 1 \.s-

NH2

I R

I R

6-I>BaSI>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 <^ PCr + ADP Adenylate kinase: AMP + ATP <-> ADP + ADP In contrast to the reactions described above, these enzymes are primarily involved in maintenance of ATP homeostasis rather than in providing energy for some

26

DAVID A. HARRIS lfH2

^H2 C=NH

C=NH

N-CH3

NH

I

CH2

CH2

COO"

CH2 COO*

creatine

Figure

12.

p-guanidino propionate

Structure o f c r e a t i n e a n d its a n a l o g u e , p - g u a n i d i n o p r o p i o n a t e . T h e

p o s i t i o n o f p h o s p h o r y l a t i o n , t o f o r m PCr is s h o w n (*).

particular cell function. Both enzymes are highly active and freely reversible in vivo. Creatine Kinase

Phosphocreatine represents a metabolic dead end; its only reaction is with ADP, and, in the creatine kinase reaction, it acts as an ATP buffer. This reaction occurs in all muscle and neural tissue, with each tissue having its own isoenzyme of creatine kinase. Other tissues (such as liver and kidney) lack creatine kinase and PCr. The structure of creatine is shown in Figure 12. PCr concentration in heart cells is typically about 25 mM, which is 3—4 times the concentration of ATP. Thus, as ATP is used, if ADP tends to rise, it can be rephosphorylated using PCr as phosphate donor. This is a rapid mechanism of mitochondrion

sarcoplasm

PCr

ADP.

C^: Cr

ATP

Figure 13, Transfer of high-energy phosphates by the phosphocreatine shuttle. CK^^ and CKs represent the mitochondrial and sarcoplasmic isoenzymes of creatine kinase, respectively. Free Pj is omitted for clarity.

Cellular ATP

27

regenerating ATP (buffering ATP concentration) in short-term energy deficiency (see Figure 2b). However, PCr stores cannot last more than 15-30 seconds in the absence of other energy input; PCr seems more likely to function, for example, in smoothing out ATP variation between heart beats than as a significant energy store in the tissue. A second possible role of PCr is in shuttling ADP to and from mitochondria, especially in heart (Jacobus, 1985). This proposal is based on the idea that ADP, present at very low concentration (10-30 juM), may not be able to diffuse from its site of formation on the actomyosin filament to the mitochondria fast enough to allow rapid oxidative phosphorylation. With the PCr shuttle, ATP made inside the mitochondrion emerges into the intramembrane space where it transfers its y phosphate directly to creatine. The resulting PCr moves across the cytoplasm to phosphorylate ADP near actomyosin, for use in muscle contraction, and the resulting creatine diffuses back to the mitochondrion (Figure 13). ADP is formed at the mitochondrion in this process, but can be rephosphorylated rapidly since it does not have to diffuse within the fiber, but simply to reenter the mitochondrion. The only diffusing species are Cr and PCr, present at 100-1,000 x [ADP] concentration. While such a mechanism may well operate, experiments in which the heart's creatine is depleted by treatment with the analogue P-guanidino propionate (Figure 12) (Zweier et al., 1991) with little or no effect on cell function, suggest that it is unlikely to contribute significantly to energy transfer in the heart cell. Adenylate Kinase This enzyme is present at high levels in all tissues. It could clearly act as an ATP buffer if ADP levels were to build up, but the levels of ADP in vivo (about 1% of ATP levels) make this unlikely. It is more likely to be useful in the rephosphorylation of AMP, generated in reactions like amino acid activation (above). AMP itself cannot be used in oxidative phosphorylation, but adenylate kinase will phosphorylate it to ADP, which can. Perhaps a more important role of this enzyme is in signaling the energy status of the cell to regulatory enzymes. In cells, the concentrations of ATP, ADP, and AMP are widely different, with [ATP] at about 5 mM, and free [ADP] and [AMP] at 0.01-0.05 mM (see Table 2). Hydrolysis of a small amount of ATP (say, 0.1 mM) by myosin, for example, will decrease ATP levels by a very small fraction (2%). However, this will tend to increase ADP concentration by the same amount, which is a much larger fraction (300-1,000%) of total ADP. Since adenylate kinase maintains ATP, ADP, and AMP at equilibrium, this change will be transmitted to AMP, whose concentration will also increase by a large fraction. Thus, a small percentage fall in ATP concentration is accompanied by a large percentage rise in [AMP]. AMP, therefore, is a useful regulatory molecule in the cell; changes in

28

DAVID A. HARRIS

[AMP] are amplified versions of changes in [ATP], and enzymes responsive to [AMP] as a regulator are very sensitive to changes in cellular [ATP]. Such enzymes are discussed further below.

SYNTHESIS OF ATP ATP synthesis in animals occurs in one of two modes, one in which oxygen is used (oxidative phosphorylation), which occurs in mitochondria, and one in which oxygen is not involved (substrate level phosphorylation), which occurs, largely, in the cytoplasm. In elementary textbooks, substrate level phosphorylation is often referred to as being less efficient, since it yields only 2 mols ATP per mol glucose used (as compared to 36 mols/mol glucose in oxidative phosphorylation); in mammals, however, this term is misleading since lactate, produced in substrate level phosphorylation in one tissue, can be fiilly oxidized in another with no loss of ATP to the organism. On the contrary, the ability of some tissues to carry out substrate level phosphorylation allows them to be better adapted to their fiinction, and thus increases the efficiency of the organism as a whole. Substrate Level Phosphorylation Substrate level phosphorylation in animals takes place during glycolysis. The pathway for glycolysis is shown in Figure 14. ATP synthesis typically involves two steps: (a) the generation of a high energy phosphate bond; and (b) the transfer of this phosphate to ADP. It occurs twice in the breakdown of each molecule of triose phosphate. The regulation of this pathway is discussed later. In glycolysis, energy is yielded via an internal oxidation-reduction reaction; the aldehyde group of a sugar is oxidized to an acid, and an alcohol group reduced to a methyl group. This is possible only with a carbohydrate energy source. Thus, the fuels for substrate level phosphorylation in vivo must be (blood) glucose or (stored) glycogen; fatty acids and amino acids cannot be used to generate ATP without the involvement of oxygen. Tissues that generate ATP mainly by this route are not, typically, anaerobic. Rather, they lack (sufficient) functional mitochondria to meet their energy demand. They include: 1. Erythrocytes: their (low) energy demand is met by uptake of glucose from the blood, and diffusion of lactate into it. They contain no mitochondria. 2. Lens of the eye: energy requirements are met as for erythrocytes. Here, the presence of mitochondria would impair function, as they are small particles (light scattering) and colored (due to heme). 3. Kidney medulla: in contrast to the kidney cortex, the-medulla, which contains the loops of Henle, is poorly vascularized. This may be due to the osmotic problems of capillaries in such a hypertonic space. Thus oxygen

EXTRACELLULAR FLUID adrenaline

glycogen glycogen

^

GiP

<

phosphorylase

ATP + AMP <

>2ADP

- > ADP + ?\

1,3-diphosphoglycerate

t

(b)

ATP
2-phosphoglycerate

(a)

A

phosphoe«o/pyruvate ATP

(b)

pyruvate . 4 — ^ lactatejn

<—•

lactatCot

figure 14. Glycolysis in skeletal muscle. Stages at which a high energy bond is generated are indicated by V / X / ^ , and designated (a) (see text). Phosphate transfer to ATP occurs in the following step (b). Only key regulatory enzymes are named. Molecules signaling an actual or potential increased ATP utilization are shown in bold. The role of kinases (protein kinase A, phosphorylase kinase) in the actions of cAMP and Ca^"^ on phosphorylase is omitted for clarity. 29

DAVID A. HARRIS

30

supply is low, despite a significant demand for ATP (for pumping ions), and energy derives from glycolysis. Fast twitch (anaerobic) muscle (Type lib fibers): these are responsible for short intense bursts of activity (e.g., sprinting), and are poorly vascularized, but contain large glycogen stores (increased by training). Some mitochondria are present, providing ATP for the cell at rest. On activity, internal glycogen stores are broken down to lactate; entry of glucose and oxygen from the blood would be too slow to fiiel the bursts of high activity, especially since contraction may squeeze blood out of the muscle. Note that internal glycogen stores are limited as increasing them increases muscle mass. Typically, they can maintain ATP levels for 20-30 seconds. Tumor cells: many tumors (e.g., Morris hepatoma, Ehrlich ascites tumor) contain apparently normal mitochondria, but convert a large fraction of glucose taken up to lactate. This may be partly due to limitations in oxygen diffusion to the center of solid tumors. However, it may also reflect the demand by stimulated biosynthetic pathways for glycolytic intermediates in these cells. Lymphocytes and macrophages, similarly, obtain significant amounts of ATP by converting glucose to lactate, despite the presence of functional mitochondria (Ardawi and Newsholme, 1985). Again, this is

fatty acids pyruvate fatty acids

pyruvate L j r dehydrogenase I

^'^^^ Ca^*

acetyl CoA

FADH2 MITOCHONDRIAL MATRIX

FADHj

Figure 15. Oxidative ATP generation in heart muscle. Only key regulatory enzymes are named. Regulatory molecules are shown in bold.

Cellular ATP

31

linked to a requirement for rapid biosynthesis (in response to immune stimulation) in these cells. Oxidative Phosphorylation Oxidative phosphorylation occurs inside mitochondria. It can be divided notionally into two stages, the generation of reduced cofactors (NADH and reduced flavin), and the oxidation of these cofactors by oxygen, coupled to ATP formation. Reduced cofactors are generated by the soluble enzymes of the mitochondrial matrix, notably those of the Krebs cycle and those catalyzing (3 oxidation of fatty acids. These processes are summarized in Figure 15. Typically, oxidation of a-oxo acids or hydroxy compounds yields NADH, which is a soluble cofactor in the mitochondrial matrix; oxidation of-CH2-CH2- groups (e.g., by succinate dehydrogenase, fatty acyl dehydrogenase) yields FADH2 which is bound within mitochondrial inner membrane complexes. Electrons from these cofactors are transferred between the electron transfer complexes of the mitochondrial membrane, ultimately reducing oxygen to water. In this process, protons (H"^) are pumped out of the mitochondria, building up an electrochemical H^ gradient as an intermediate energy store. This is used by another transmembrane complex, the FiFoATPase, for ATP synthesis (see below). Fuels for Oxidative Phosphorylation Compared to substrate level phosphorylation—^where the only available fiiel is glucose (free or in glycogen)—oxidative phosphorylation can be fueled by sugars, fatty acids, or amino acids. It is thus a particularly versatile (and efficient) mode of ATP production. Different oxidative tissues prefer different fuels for ATP production, and the preferred fuel may vary with nutritional status of the organism, the whole producing an integrated, adaptable system for energy production. Examples are: 1. Brain. The brain is one of the few tissues whose ATP production (in the fed state) is fueled largely by glucose oxidation. Fats and fatty acids are inaccessible to the brain because of the blood-brain barrier. In starvation, ketone bodies (partially oxidized fatty acids, produced by the liver) can fuel the brain. 2. Heart. Heart muscle, in contrast, prefers to oxidize triglyceride (from circulating lipoproteins) and lactate in the fed state. Circulating fatty acids and ketone bodies are oxidized in starvation. 3. Skeletal muscle (Types I, Ila). Like heart muscle, skeletal muscle shows a preference for lipid substrates, in particular triglycerides (fed state) and fatty acids/ketone bodies in prolonged exercise or starvation. Skeletal muscle

32

DAVID A. HARRIS does not oxidize lactate due to the low affinity of the muscle isoenzyme of lactate dehydrogenase for this substrate, relative to that of the heart enzyme. 4. Liver. Since the liver is an important center of nitrogen metabolism, its major fuel in the fed state is a-oxo acids, derived from the deamination of dietary amino acids. In starvation, it will oxidize fatty acids released from adipose tissue. 5. Gut. Perhaps surprisingly, small intestine obtains about 60% of its ATP in the fed state from the oxidation of the single amino acid, glutamine (Watford et al., 1979). (This is twice transaminated/deaminated to yield a-oxoglutarate, which is oxidized in the tricarboxylic acid cycle). Oxidation of glucose and ketone bodies make up the remaining 40%.

Mechanism of the Mitochondrial ATP Synthase The mitochondrial ATP synthase utilizes energy stored in a transmembrane H"^ gradient, transducing this into the phosphoanhydride bond of ATP. Current views hold that 3¥t are transferred across the mitochondrial inner membrane for each ATP made by the synthase. Coupling between H"^ movement and ATP synthesis is indirect, i.e., the protons moved do not interact with the ATP made, the two processes communicating solely via changes in protein conformation. The resulting mechanism is essentially the reverse of that described above for ATP-driven pumps, as indicated in Figure 16a. In essence, the synthase has a very high affinity for ATP, such that ADP + P, spontaneously form ATP on the enzyme surface; energy is required to displace ATP from the enzyme. This is supplied by binding H"^ from a solution of high electrochemical potential, inducing a conformational change that displaces ATP. The ATP synthase has a bipartite structure, with the H^ channel lying on one set of polypeptides (FQ, transmembrane fragment) and the ADP binding site on another (Fj, extrinsic fragment) (see Figure 7). Its mechanism is also complicated by the presence of not one but three ADP binding sites which operate alternately. Thus, binding of ADP + Pj, ATP formation and ATP release involve three different conformations of the catalytic site of Fj (designated Loose, Tight, and Open). In the holo-enzyme, each of the three catalytic sites pass through each of these conformations, but at 120° they are out of phase (see Figure 16b) (Boyer, 1987). The reason for this complicated arrangement may be to ensure that ATP does not simply rebind to the enzyme after its expulsion by H"^; any potential ATP binding sites have been previously filled by ADP + Pj (see Figure 16a). The asymmetric structure of Fi at any instant is thought to be due to the preferential interaction of one particular conformation of the catalytic subunits with the (stalk) polypeptides which link the membrane (FQ) and catalytic (Fi) fragments. The ATP synthase generates ATP within the mitochondrion, while most of it is needed in the cytoplasm. ATP is exported from the mitochondrion via the adenine

3H'

E.ATP

30 k> ^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 (<5%) change in ATP concentration. In other tissues too, cytoplasmic ATP levels hardly change over the physiological range of energy utilization. In very severe exercise, ATP levels in skeletal muscle may fall by 50-60%, but this is accompanied by a loss of total adenine nucleotide from the tissue. The level of cellular [ATP] depends on the balance between synthesis and utilization. However, utilization depends on the needs of the cell, and ultimately the organism, the energy-requiring processes (movement, biosynthesis, etc.). Thus ATP levels must be controlled, under physiological conditions, by varying the rate of ATP synthesis. Since cellular ATP turns over about once every second, such control must be both precise and rapid. In addition, processes that require ATP are buffered against changes in its concentration. This is because enzymes that use ATP (e.g., actomyosin, hexokinase, NaVK"^ ATPase, etc.) have K^ values around 0.1-0.2 mM, which means that they are saturated at physiological [ATP] of 4-8 mM; even a 50% change in ATP concentration would hardly affect their ATP utilization rate. Muscle fatigue, for example, is accompanied by a fall in ATP levels to 2-3 mM, but the loss of muscle function in fatigue is due to a rise in intracellular [H"^] (pH falls from 7 to 6.5) rather than ATP depletion (see Figure 2). ATP levels, therefore, seem to be regulated even more precisely than one might expect. It is possible that K^ values measured in vitro may be misleading, and that apparent K^ values in vivo may be closer to physiological [ATP] (if, for example, cellular ADP competes strongly with ATP). Even so, it seems that metabolic processes are doubly insulated against a tendency of [ATP] to fall: ATP synthesis rates will be increased by regulatory mechanisms, and, in addition, enzymes are relatively insensitive to decreases in [ATP] around physiological concentrations. Is ATP a Regulator?

It is tempting to imagine ATP itself as a metabolic regulator. It is conceivable, for example, that ATP might inhibit key enzymes in glycolysis and oxidative

36

DAVID A. HARRIS

phosphorylation; use of ATP would cause its concentration to fall, relieving this inhibition and speeding up ATP synthesis. Indeed, the regulatory enzyme phosphofructokinase (PFK) is inhibited by ATP levels of around 1-2 mM in vitro. The drawback to this simple model is the very constancy of cellular [ATP]. It is difficult to see how a drop in [ATP] of around 5% could allosterically increase the activity of an enzyme by up to 100-fold, and yet this change in glycolytic flux can occur in skeletal muscle during a transition from rest to exercise. In other words, no known system would be sufficiently sensitive for changes in ATP levels directly to cause the observed changes in flux. ATP itself is not a physiological regulator molecule. Mammals have adopted three mechanisms to circumvent this problem. 1. ADP and, in particular, AMP are used to signal changes in ATP concentration. Small percentage changes in [ATP] lead to large percentage changes in both [ADP] and [AMP] levels, due to the large difference in absolute concentration of these nucleotides and the operation of adenylate kinase (above). Thus small changes in [ATP] are amplified into large changes in concentration of these effectors. 2. The sensitivity of an ATP yielding pathway to such regulators is further amplified by the operation of substrate cycles. In a substrate cycle, two enzymes catalyze opposing reactions simultaneously so the net flux through the cycle is small; a small increase in the activity of one enzyme and a corresponding decrease in the activity of the other will, however, produce a large change in net flux (Figure 17). Substrate cycles consume ATP (1 mol per turn of the cycle), as both forward and reverse reactions must be downhill. The fact that such cycles operate in ATP generating pathways attest to the physiological importance of tightly controlling ATP generation; it is even worth paying for with existing ATP. 3. Variations in ATP demand may be signaled by indicators other than adenine nucleotides. In heart and skeletal muscle, for example, mean cytoplasmic [Ca^"*"] rises with increasing work rate due to its role in the contractile process. This can be used as a signal of ATP demand, speeding up ATP production. Extracellular regulators, such as epinephrine, signal an imminent rise in ATP demand rather than an actual one. Such regulators, via cAMP as second messenger, can also stimulate ATP generating pathways, in this case prior to an actual increase in utilization. Control of Anaerobic ATP Production Control in skeletal muscle type lib (white muscle) is taken as the archetype of such regulation, which occurs in the glycolytic pathway (oxidative phosphorylation cannot operate under anaerobiosis). Such muscle (e.g., human quadriceps) is

Cellular ATP

37 ATP , " CPF

fructose

100---^110

6-phosphQte

fructose 1,6-bisphosphQte

1--^20 <J)

>

J'j^--»90 (fBPas§ Figure 17. A substrate cycle in the amplification of a metabolic response. The enzymes depicted are phosphofructokinase (PFK) and fructose 1,6, bisphosphatase (FBPase) (see Figure 14). The figures shown indicate rates (arbitrary units). Initially, the forward rate is 100 and the reverse rate 99; the flux emerging through the pathway (J) is 1 unit. A 10% rise in PFK activity (100 -> 110), and a 10% fall in FBPase (100 -^ 90) leads to a 2,000% change in onward flux.

capable of very intense but short lived bursts of activity, during which glycolytic rates can increase up to 200 x over the rate observed at rest. Control by Nucleotide Levels Crossover analysis has demonstrated that phosphofructokinase (PFK) is the key regulatory enzyme in glycolysis. This enzyme is activated by AMP in vivo (Figure 14). It also participates in a substrate cycle with fructose 1,6 bisphosphatase (FBPase), which hydrolyzes fructose 1,6 bisphosphate to fructose 6 phosphate (F6P), opposing the action of PFK (Figure 17). In muscle, where no gluconeogenesis occurs, the only role of FBPase is to participate in this substrate cycle. Thus, as muscle is stimulated to contract, ATP is hydrolyzed to ADP. ATP levels are buffered by PCr and barely fall; by virtue of adenylate kinase, AMP levels rise significantly. PFK is thus stimulated and glycolysis speeds up. The sensitivity of the system is maximized by using AMP rather than ATP as a regulator (see above). In addition, PFK shows allosteric cooperativity (i.e., a sigmoidal response to AMP) such that a doubling of AMP concentration, in the physiological range, can give a 5—10-fold increase in active enzyme. Furthermore, due to the reverse action of FBPase in the substrate cycle, doubling the activity of PFK can give a 50-fold change in net flux through the pathway. Each of these mechanisms, therefore, amplifies the sensitivity of the pathway to the tiny fall in ATP levels, the rate of ATP synthesis rises sharply, and [ATP] is maintained almost constant. Control, however, cannot be exerted solely at PFK. Unless the supply of hexose phosphate to PFK is maintained, activating it would simply lead to a drastic fall in

38

DAVID A. HARRIS

F6P (and G6P) level, PFK would be deprived of its substrate, and its activity would fall. In rapidly contracting muscle, blood supply is restricted, and with it the source of glucose (and oxygen). In the short term, therefore (sprinting for 10-20 seconds), the source of hexose phosphate is glycogen stored within the muscle. Glycogen phosphorylase is, therefore, the other key enzyme regulated in glycolysis in muscle (Figure 14). Control at the nucleotide level is exerted, again, by AMP which cooperatively activates this enzyme. In addition, this enzyme is inhibited by glucose-6- phosphate. The stimulatory effect of AMP is thus amplified by the reduction in G6P inhibition as PFK is stimulated and sugar phosphates used up. Unlike PFK, glycogen phosphorylase is saturated with its substrate glycogen, and is thus unaffected by even quite large falls in glycogen concentration. When activated (via an increase in Vmax) phosphorylase can maintain a high level of hexose phosphate production until glycogen is nearly exhausted; it catalyzes the flux generating step of the pathway. In severe exercise, muscle AMP levels mayrisehigh enough for this nucleotide to undergo deamination to inosine monophosphate (IMP), producing NH3. The reason for this is uncertain; it may remove AMP so that adenylate kinase activity is not impaired and ADP build up (which would inhibit actomyosin) is prevented. NH3 also stimulates PFK and thus promotes ATP synthesis. However, prolonged activity at this level will deplete the cell of adenine nucleotides and thus cannot be sustained. Up to 20% of adenine nucleotides can be lost in this way; they are regenerated (the lost NH3 being replaced from aspartate) when the muscle returns to rest. Control by Other Agents In sustained contraction, cytoplasmic [Ca^"^]risesfromaround 0.3 |iM to 1.5 \xM. After stimulation by epinephrine (e.g., in anticipation of exercise) cAMP levels in the cell are raised between roughly the same concentrations. The effect of either is to stimulate glycogen phosphorylase via phosphorylation, which is mediated by phosphorylase kinase. Phosphorylation converts phosphorylase from an AMP-dependent form (phosphorylase b) to a form active independently of AMP (phosphorylase a), thus increasing the supply of sugar phosphates to PFK even without significant changes in nucleotide levels. An increased F6P level will stimulate PFK activity, both because F6P is a substrate and the enzyme is normally unsaturated and because F6P is an allosteric activator of PFK. Ca^"^ ions act directly on phosphorylase kinase, whose smallest subunit is calmodulin. This binds Ca^"^ as levels rise and the enzyme is activated. cAMP acts indirectly via protein kinase A; it is the phosphorylation of phosphorylase kinase by this enzyme which activates it. In both these cases magnitude amplification occurs because activation of glycogen phosphorylase involves phosphorylation by

Cellular ATP

39

a controlling enzyme, a few molecules of regulator (e.g., of epinephrine at the cell membrane) can activate a large number of enzyme molecules. Both these effects will stimulate glycolytic ATP synthesis in conditions of increased ATP utilization (or potential utilization) without requiring changes in nucleotide levels, and are thus useful ATP homeostatic mechanisms. However, they are not essential for normal function of muscle; I-strain mice, which lack phosphorylase kinase, can exercise normally. Presumably, in this condition, AMP serves adequately as a sole regulator molecule. A more subtle effect of epinephrine is that it will stimulate muscle FBPase in addition to PFK, possibly via the regulator fructose 2,6 bisphosphate. In other words, this hormone will increase the rate of substrate cycling in muscle. While this in itself will not lead directly to increased ATP production, it will increase the sensitivity of the pathway to regulators such as AMP. In other words, in epinephrine-stimulated muscle, a given rise in AMP will increase glycolytic flux more than in unstimulated muscle. Restoration of Blood Supply

Severe exercise can continue only for 1-2 minutes, since after this time muscle glycogen supplies are exhausted, and the pH of the muscle has fallen sufficiently to cause fatigue. For more prolonged exercise, the blood supply must be restored to remove H"*" ions and provide new substrate (glucose). Moderate exercise, in type lib muscle, can be driven for much longer under these circumstances. Indeed, patients lacking glycogen phosphorylase (McArdle's syndrome) cannot use muscle glycogen; however, they can tolerate moderate exercise entirely using blood glucose as a substrate. As might be expected, glucose entry into muscle is stimulated in exercise. However, it is not yet known how the glucose transporter senses a change in cellular energy demand. It is known that the transporter involved—^the exercise dependent glucose transporter—^is distinct from the insulin-dependent glucose transporter (glut-4) in the cell membrane (Barnard and Youngren, 1992), and is independently stimulated during exercise. It is possible that [Ca^"^] is the internal signal for this stimulation. Defects in Glycolysis

Because some tissues use glycolysis to provide all their ATP, a total absence of a glycolytic enzyme in man would be fatal. In some conditions, however, a partial deficiency is observed, leading to lowered ATP levels in such tissues. An example occurs in a common form of hemolytic anemia, in which pyruvate kinase levels in erythrocytes are reduced to 20% of normal. The erythrocytes lose their shape and lyse readily. In these cells, ATP levels fall and the erythrocyte is unable to maintain

40

DAVID A. HARRIS

its cytoskeleton (hence loss of shape) or Na"^ extrusion via the NaVK"^ pump (hence osmotic lysis). Control of Aerobic ATP Production Control in heart muscle is taken as the archetype of such regulation. Since the fuels used are typically triglyceride and lactate (in the fed state), major control points lie in the mitochondrial oxidative phosphorylation system. The work rate of heart can vary by up to 10-fold from minimal to maximal activity, by changes in heart rate and force of contraction (inotropy). Control by Nucleotide Levels Any discussion of the control of mitochondrial ATP synthesis must consider the fact that ATP utilization occurs in the cytoplasm while synthesis occurs inside the mitochondrion. Thus, some mechanism must exist for signaling changes in energy demand between these different cellular compartments. At the nucleotide level, the metabolic signal appears to be ADP (AMP, the cytoplasmic signal, cannot enter mitochondria). Two mitochondrial enzymes respond directly to ADP levels: the adenine nucleotide translocase and the ATP synthase (above). Both of these use ADP as substrate, and a rise in ADP levels will cause an increase in their activity, and an increase in ATP synthesis rate. (Note that these enzymes are not formally regulated by ADP; they are simply responding to its availability as a substrate.) Which of these enzymes might be limiting in vivo has been a matter of some debate. The answer probably depends on the tissue under consideration. In heart, the capacity of the translocase is high (Doussiere et al., 1984), and mitochondrial ADP levels may simply reflect cytoplasmic levels. Changes in ADP concentration thus affect overall flux by increasing ATP synthase turnover. In liver, the translocase activity is lower and ADP entry may limit access to the synthase. On these considerations, and on the basis of studies with mitochondria in vitro, it was suggested that mitochondrial ATP synthesis rates were simply controlled by varying the levels of substrate, cytoplasmic ADP. Such a model requires that, in vivo, cytoplasmic ADP levels vary around K^^^ for ATP synthesis (ca 30 |LIM) over a sufficient range (5-10-fold) to account for the observed rate changes. In skeletal muscle (human arm) this may well approximate to the true situation; NMR measurements of ATP synthesis and cytoplasmic [ADP] levels show a suitable relationship between ATP synthetic flux and increases in [ADP] with increasing levels of exercise. In heart, however, this is not the case. Over a wide range of heart work rates, both free [ATP] and [ADP] levels (the latter calculated from ATP/PCr ratios, above) remain effectively constant. Thus, again, we find that nucleotide levels are

Cellular ATP

41

very precisely controlled, and that changes in their concentrations are too small to be responsible for observed changes in flux. Control by Ca^^ Levels A second problem with the above model is that substrate ADP controls only the ATP synthase enzyme. This could not sustain an increased rate of ATP synthesis alone, since the proton gradient would soon fall due to a lack of oxidizable substrate. Thus, control of the ATP synthase must be integrated with control of NADH generating reactions: the Krebs cycle dehydrogenases. The messenger which communicates the need for ATP in the cytoplasm to these enzymes is Ca^"^ (Figure 15); other putative messengers such as AMP and cAMP, which control glycolysis, cannot enter the mitochondrion. Ca^"^ enters the mitochondrion through an electrogenic uniport, i.e., it uses the transmembrane potential (inside negative) as a source of energy. It can leave via an electroneutral 2Na"*"/Ca^'^ antiporter. The relative affinities of these systems lead to intramitochondrial [Ca^"^] reaching a steady state at levels (1-10 JLIM) 3-10 x higher than cytoplasmic [Ca^"*"] (0.3-1.5 |LIM). Importantly, however, changes in intramitochondrial [Ca^^] will reflect changes in cytoplasmic [Ca^"*"], and this ion is thus a signal of cytoplasmic contractile activity (energy demand) (Denton and McCormack, 1990). Four mitochondrial enzymes are known to be stimulated by physiological levels of [Ca^"*"]. Ca^"*" stimulates the phosphatase which dephosphorylates (activates) the pyruvate dehydrogenase complex; oxidation of cytoplasmic pyruvate (from lactate and, less commonly, glucose) is stimulated. Ca^"*" allosterically activates isocitrate dehydrogenase and a-oxoglutarate dehydrogenase, two key regulatory enzymes in the Krebs cycle; this stimulates NADH production from all substrates, notably fatty acids. And Ca^"^ activates the ATP synthase, probably by displacing a regulator protein which normally inhibits its activity (Figure 15) (Harris and Das, 1991). To sum up, increased contractile activity in heart leads to an increased mean cytoplasmic [Ca^"*"], leading to a rise in intramitochondrial [Ca^"^]. (The relatively low rate of the mitochondrial transporters smooth out the cyclic variations in cytoplasmic [Ca^^ occurring at each beat cycle.) Oxidation of pyruvate is stimulated, as is oxidation of acetyl CoA derived from fatty acid oxidation, and NADH levels rise. NADH dehydrogenase, which is normally not saturated with NADH, thus increases in activity and oxidation rates rise. Simultaneously, the ATP synthase is switched on and ADP is converted more rapidly to ATP. Thus, a new steady state arises, with increased flux through the respiratory chain and ATP synthase, with a limited change (commonly a small decrease) in the intermediate proton gradient. Cytoplasmic ATP levels are maintained, without any significant change in cytoplasmic ADP or AMP levels.

42

DAVID A. HARRIS

Evidence for the operation of this system comes from studies with ruthenium red, an inhibitor of mitochondrial Ca^"*" uptake (Unitt et al., 1989). Treatment of perfused heart, or cultured heart cells, with this compound inhibits activation of the dehydrogenases and of the ATP synthase which normally occurs with increased work rate, showing the importance of Ca^^ as an intracellular messenger. Furthermore, while untreated cells can maintain their ATP levels constant irrespective of work load, cells in which mitochondrial Ca^"^ uptake is blocked with ruthenium red suffer a drop in [ATP] at high work rates, showing the importance of this mechanism in maintaining normal energy balance in the heart. The effects of Ca^"*" above are discussed largely in terms of Ca^"*" released from the sarcoplasmic reticulum during contraction. However, oxidative phosphorylation can also be stimulated by positive inotropic agents (e.g., epinephrine). In this case, (cAMP-stimulated) protein kinase A phosphorylates an L-type Ca^"*" channel in the cell membrane, promoting (transient) Ca^"^ entry and raising cytoplasmic [Ca^"^]. This then acts on intramitochondrial processes as described above. Defects of Aerobic A TP Production Enzyme deficiencies in Krebs cycle enzymes generally lead to widespread tissue damage (and death), because of the central role of this pathway in a variety of metabolic pathways (including ATP generation). Deficiencies of pyruvate dehydrogenase are known, and generally lead to death in infancy; here, however, damage is localized in the brain because of the dependence of this tissue on glucose oxidation, which requires this enzyme. Tissues such as muscle and liver can utilize alternative fuels. Deficiencies in respiratory chain components also occur in mitochondrial membranes. Indeed, they appear relatively common compared to other genetic disorders. Such defects lead to functional problems in actively oxidizing tissues (i.e., to neuropathies and myopathies). However, location of the damage and severity of the symptoms vary greatly between sufferers; some show chronic tissue damage, while others are able to generate ATP adequately except under conditions of severe stress. This is because many such defects (NADH dehydrogenase, cytochrome oxidase deficiency, etc.) result from lesions on mitochondrial DNA (mDNA) (Harding, 1991). Since one cell contains several autonomously replicating mitochondria (each containing several copies of mDNA), normal and mutant mDNA may be present in the same cell. The severity and location of the functional lesion will then depend on the proportion of mutant mDNA in cells of a given tissue; this in turn will reflect how early in the lineage of the cell type the mutation arose. Less severe changes are observed in conditions (such as hypertension or alcoholic cardiomyopathy) where Ca^"*'-dependent regulation of the ATP synthase is defective. In such cases, ATP levels in the heart cannot be precisely maintained as work rate increases. Thus, ATP levels fall (ca. 20%) with a consequent rise in

Cellular ATP

43

[ADP]. Nonetheless, a new steady state is achieved. It seems that, in the absence of functioning regulation by [Ca^"^], ATP synthesis rates in heart can be increased through substrate control by ADP, as observed in skeletal muscle (above).

PATHOLOGICAL DISTURBANCES OF ATP LEVELS We have seen above how defects in the enzymes of the glycolytic pathway, or the pathway of oxidative phosphorylation, cause severe pathological changes in ATP levels, since ATP utilization continues in the absence of normal ATP synthesis rates. Other conditions exist where an imbalance exists between ATP synthesis and utilization rates, with consequent abnormalities in ATP levels. Malignant Hyperthermia (Malignant Hyperpyrexia)

This condition is characterized by a massive increase in aerobic and anaerobic ATP generation induced by the administration of volatile anesthetics (e.g., halothane). Metabolic activity is so high that body temperature rises (up to 44 °C in extreme cases) and severe lactic acidosis results. The primary defect in this syndrome lies in the Ca^"^ channel of sarcoplasmic reticulum in skeletal muscle. Halothane prolongs opening of this channel, and cytoplasmic [Ca^"^] rises sharply (Heffron, 1988). This activates the actomyosin ATPase and ATP is rapidly hydrolyzed, with consequent heat production. In addition, the Ca^^ pumps of the sarcoplasmic reticulum and plasma membrane are stimulated in an attempt to remove the excess Ca^"^, leading to further ATP hydrolysis. These rises in Ca^"^, ADP, and AMP will stimulate glycolysis and oxidative phosphorylation as described above. Oxidation, in fact, does not reach its maximal capacity, either because pyruvate transport is too slow or because of the effect of high temperature on the mitochondrial membranes; thus oxidation rates rise, typically only 3-4 times. Glycolytic rates are stimulated much more greatly, and considerable lactate is produced even by aerobic muscle. Nonetheless, ATP levels cannot be maintained, and over a 15-30 minute period, [PCr] and [ATP] levels fall below the levels needed to maintain cell viability. Malignant hyperthermia is not a particularly rare genetic disease (incidence ca. 1/12,000) and, while it has little effect on physiological function in everyday life, is a major problem in anesthesiology, particularly as no simple diagnostic test for the condition exists. Ischemia

Most cells can maintain their ATP levels during anoxia; cultured heart cells, for example, retain normal ATP levels for more than 30 minutes in the absence of oxygen. This shows that cells can switch smoothly between glycolysis and oxida-

44

DAVID A. HARRIS

tive phosphorylation as sources of ATP. If the blood supply to heart muscle is interrupted in vivo, however, consequences are much more severe. ATP levels rapidly fall by some 20%, and are maintained at this level. More importantly, however, are the changes that occur when blood supply is restored; if it has been interrupted for more than 20 minutes, reperfusion leads, surprisingly, to further ATP depletion and commonly cell death. This deleterious effect of reintroducing oxygen is known as the oxygen paradox. Since the heart is unable to regenerate new cells, this process leads to permanent damage of some area of the heart—a. myocardial infarct. Similar effects are observed in the brain tissue (stroke). The difference between anoxia and interruption of blood supply lies in the build up, in the latter, of the products of anaerobic metabolism, particularly H"^. Reoxygenation from pH 7.4 has no deleterious effects; reoxygenation (and restoring the tissue to pH 7.4) from pH 6-6.5 causes injury. This is associated with a massive influx of Ca^"*"fromthe blood, probably via the Na"^/Ca^'*" exchanger in the cell membrane (Tani, 1990). It is this abnormal Ca^"*^ load that increases energy utilization (via Ca^"*" pumps and actomyosin stimulation); in addition, it is sufficiently high to overload the mitochondria (possibly precipitating Pj internally) and to inhibit oxidative phosphorylation. Thus, ATP utilization increases, ATP production falls, and ATP levels drop precipitately. It is still unclear why reoxygenation/pH change of heart cells leads to this pathological inflow of Ca^"^. Agents such as free radicals (generated in partial oxidations) and activated phospholipases may damage the cell membrane under these conditions, but it is still uncertain whether the magnitude of such effects is sufficient to explain the dramatic increase in Ca^"^ influx on reperfusion.

SUMMARY ATP is a kinetically stable molecule with a high free energy of hydrolysis/high phosphate transfer potential. This means it can act as a common unit of exchange of energy between a variety of highly exergonic catabolic processes and energy requiring reactions within the aqueous medium of a cell. The chemical nature of the ATP molecule means that it can drive a wide variety of such reactions, including movement of ions and proteins, dehydrations (in macromolecule biosyntheses), activation of small molecules, and imparting a negative charge to sugars and proteins. The relative contributions of these processes to the energy demand of a cell depends on the tissue; muscle expends some 70% of its ATP turnover on movements of actomyosin, brain about 40% of its ATP on Na^ transport, and exocrine cells about 50% on biosyntheses. The molecular basis for the coupling of ATP hydrolysis to non-chemical processes (e.g., muscle contraction, ion pumping) is not precisely known. A model is presented in which the transducing enzyme can manipulate the stages in the release of energy from ATP to link binding energy changes to conformational changes. This rationale unifies existing schemes for the function of these enzymes,

Cellular ATP

45

and also for the functioning of the mitochondrial ATP synthase, which couples transmembrane proton flow to ATP synthesis. In humans, the bulk of ATP synthesis occurs via oxidative phosphorylation in mitochondria, although the fuel oxidized is dependent on tissue. In some tissues (and in rapidly growing tumors), however, considerable amounts of ATP are made by glycolytic conversion of glucose to lactate. This reflects an adaptation of these tissues to specific functions; for example, sustained contraction (and hence restricted O2 supply) in white muscle or a high requirement for biosynthetic precursors (e.g., in lymphocytes). ATP levels inside cells are maintained very precisely (around 8 mM) under all physiological conditions by increasing the rate of ATP synthesis to match demand. In contrast, rates of ATP synthesis can vary greatly (5-100-fold) with the energy demands of the tissue. Together, these facts indicate that ATP itself cannot regulate its own synthesis. Cytoplasmic (glycolytic) ATP synthesis is regulated internally by AMP, changes in whose concentration amplify the changes in [ATP]; other regulators are Ca^"*" and cAMP, which signal an actual or potential increased work rate by the tissue. Mitochondrial (oxidative) ATP synthesis is regulated by cytoplasmic ADP (entering via the adenine nucleotide translocase) and/or by cytoplasmic Ca^"^ (entering via the Ca^"^ uniport), depending on the tissue. Prolonged ATP depletion leads to cell death, largely due to the development of ionic and osmotic imbalance. Such depletion occurs in a variety of clinical conditions. There may be deficiencies in the enzymes of ATP production (e.g., pyruvate kinase, pyruvate dehydrogenase, mitochondrial dehydrogenases) or conditions which lead to abnormal ATP utilization (e.g., fructose intolerance, malignant hyperpyrexia, ischemia/reperfusion). The resulting symptoms vary considerably, depending on the tissue most susceptible to the defect.

ACKNOWLEDGMENTS I thank the Wellcome Trust, and the British Heart Foundation for financial support. I am grateful to Professor G.K. Radda for supplying Figures 2 and 4, and to Dr. J. Clarke and Dr. A.M. Das for their help and encouragement.

REFERENCES Ardawi, M.S.M. & Newsholme, E.A. (1985). Metabolism in lymphocytes and its importance in the immune response. Essays Biochem. 21, 1-44. Ashcroft, F.M. & Rorsman, P. (1989). Electrophysiology of the pancreatic p cell. Prog. Biophys. Mol. Biol. 54, 87-143. Balaban, R.S., Kantor, H.L., Katz, L.A., & Briggs, R.W. (1986). Relation between work and phosphate metabolites in the in vivo paced mammalian heart. Science 232, 1121-1124. Boyer, P.D. (1987). The unusual enzymology of the ATP synthase. Biochemistry 26, 8503-8507. Browner, M.F. & Fletterick, M.F. (1992). Phosphorylase; a biological transducer. Trends. Biochem. Sci. 17,66-72.

46

DAVID A. HARRIS

Barnard, R.J. & Youngren, J.F. (1992). Regulation of glucose transport in skeletal muscle. FASEB J. 6,3238-3244. Denton, R.M. & McCormack, J.G. (1990). Ca^"*" as a second messenger within the mitochondria of the heart and other tissues. Ann. Rev. Physiol. 52,451-^66. Doussiere, J., Ligett, E., Brandolin, G., & Vignais, P.V. (1984). Control of oxidative phosphorylation in rat heart mitochondria: Role of the adenine nucleotide carrier. Biochim. Biophys. Acta 766, 492-500. Ferguson, S.J. (1986). The ups and down of P/0 ratios. Trends Biochem. Sci. 11, 351-352. Harding, A.E. (1991). Neurological disease and mitochondria. Trends Neurosci. 14, 132-138. Harris, D.A. & Das, A.M. (1991). Control of mitochondrial ATP synthesis in the heart. Biochem. J. 280,561-573. Heffron, J.J.A. (1988). Malignant hyperthermia—biochemical aspects of the acute episode. Brit. J. Anaesthesia 60, 274-278. Hibberd, M.G. & Trentham, D.R. (1986). Relationships between chemical and mechanical events during muscular contraction. Ann. Rev. Biophys. Biophys. Chem. 15, 119-161. Hyde, S.C, Emsley, P., Hartshorn, M.J., Mimmack, M.M., Gileadi, U., Pearce, S.R., Gallagher, M.P., Gill, D.R., Hubbard, R.E., & Higgins, C.F. (1990). Structural model of ATP binding proteins associated with cystic fibrosis, multidrug resistance and bacterial transport. Nature 346, 362—365. Jacobus, W.E. (1985). Respiratory control and the integration of heart high energy phosphate metabolism by mitochondrial creatine kinase. Ann. Rev. Physiol. 47, 707—725. Kauppinen, R.A., Hiltunen, J.K., & Hassinen, I.E. (1980). Subcellular distribution of phosphagens in the perfused rat heart. FEBS Lett. 112, 273-276. MacLennan, D.H. (1990). Molecular tools to elucidate problems in excitation/contraction coupling. Biophys. J. 58, 1355-1365. Nelson, N. (1992). Organellar proton ATPases. Curr. Opinion Cell Biol. 4, 654-660. Radda, G.K. (1986). The use of nmr spectroscopy for the understanding of disease. Science 233, 640-645. Radda, G.K. (1992). Control, bioenergetics and adaptation in health and disease: Non-invasive biochemistry from nmr. FASEB J. 6, 3032-3038. Robitaille, P.M., Merkle, H., Lew, B., Path, G., Hendrich, K., Lindstrom, P., From, A.H.L., Garwood, M., Bache, R.J., & Ugurbil, K. (1990). Transmural high energy phosphate distribution in the normal canine myocardium as studied with spatially localised ^'P nmr spectroscopy. Mag. Reson. Med. 16,91-116. Seiss, E.A. & Wieland, O.H. (1976). Phosphorylation state of cytosolic and mitochondrial adenine nucleotides and of pyruvate dehydrogenase in isolated rat liver cells. Biochem. J. 156,91—102. Senior, A.E. (1988). ATP synthesis by oxidative phosphorylation. Physiol. Rev. 68, 177-231. Tani, M. (1990). Mechanisms of Ca^"*" overload in reperfused ischemic myocardium. Ann. Rev. Physiol. 52, 54S-559. Unitt, J.F., McCormack, J.G., Reid, D., Maclachlan, L.K., & England, P.J. (1989). Direct evidence for a role of intramitochondrial Ca^"^ in the regulation of oxidative phosphorylation in stimulated rat heart. Biochem. J. 262,293-301. Veech, R.L., Lawson, J.W.R., Cornell, N.W., & Krebs, H.A. (1979). Cytosolic phosphorylation potential. J. Biol. Chem. 254, 6538-6547. Watford, M., Lund, P., & Krebs, H.A. (1979). Isolation and metabolic characteristics of rat and chicken enterocytes. Biochem. J. 178, 589-596. Zweier, J.L., Jacobus, W.E., Korecky, B., & Brandejs-Barry (1991). Bioenergetic consequences of cardiac phosphocreatine depletion induced by creatine analogue feeding. J. Biol. Chem. 266, 20296-20304.

Cellular ATP

47

RECOMMENDED READINGS Books Newsholme. E.A. & Leech, A.R. (1983). Biochemistry for Medical Students. John Wiley, New York. Devlin, T.M. (1992). Textbook of Biochemistry with Clinical Correlations, 3rd ed. John Wiley, New York. Harris, D.A. (1995). Bioenergetics at a Glance. Blackwell Science, Oxford.

Articles Capaldi, R.A. (1988). Mitochondrial myopathies and respiratory chain proteins. Trends Biochem. Sci. 13, 144-147. Clausen, T., van Hardeveld, C.V., & Everts, M.E. (1991). Significance of cation transport in control of energy metabolism and thermogenesis. Physiol. Rev. 71, 733-774. Erecinska, M., Bryla, J., Michalik, M., Meglasson, M.D., & Nelson, D. (1992). Energy metabolism in the islets of Langerhans. Biochim. Biophys. Acta 1101, 273-295. Hanson, R.W. (1989). The role of ATP in metabolism. Biochem. Educ. 17, 86-92. Pedersen, P.L. & Carafoli, E. (1987). Ion motive ATPases. I. Ubiquity, properties and significance to cell function. Trends Biochem. Sci. 12, 146-150. Pendersen, P.L. & Carafoli, E. (1987). Ion motive ATPases. II. Energy coupling and work output. Trends Biochem. Sci. 12, 186-189. Tanford, C. (1983). Mechanism of free energy coupling in active transport. Ann. Rev. Biochem. 52, 379-409.

This Page Intentionally Left Blank

Chapter 2

Purines CHARLES H. V. HOYLE and GEOFFREY BURNSTOCK

Introduction PpPurinoceptor Pharmacology P2-Purinoceptor Pharmacology Molecular Biology of Purinoceptors Physiological and Pathophysiological Roles of Purines Purines in the Cardiovascular System Purines in the Central Nervous System Therapeutic Potential Summary and Future Directions

49 52 53 55 60 60 63 64 66

INTRODUCTION The aim of this chapter is to provide an overview of the diverse roles that purine compounds play in physiological and pathophysiological situations, including an account of their receptor pharmacology and molecular biology, as well as their therapeutic potential. To this end, it is not intended to present comprehensive experimental evidence of all the various biological roles of purine compounds, but

Principles of Medical Biology, Volume 4 Cell Chemistry and Physiology: Part II, pages 49-75 Copyright © 1996 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-806-4

49

50

CHARLES H. V. HOYLE and GEOFFREY BURNSTOCK

rather to illustrate the main principles and some of the variations of purine physiology. In pharmacological and physiological sciences, the term "purines" has come to mean a class of chemicals with a purine nucleus (in particular adenine or guanine) linked to a ribose sugar, thus forming a nucleoside which may be coupled to a phosphate chain of variable length to form a nucleotide (Figure 1). The most important compounds appear to be adenosine and adenosine-5'-triphosphate (ATP), but many other purine compounds, such as adenosine-5'-monophosphate (AMP), adenosine-5'-diphosphate (ADP), 3',5'-cyclic AMP (cAMP), and the equivalent guanine nucleosides and nucleotides, also have biological significance. Indeed, adenine and guanine themselves are components of nucleic acids. The pharmacological activity of purine compounds has been studied since the 1920s when Drury and Szent-Gyorgyi (1929) observed negative inotropic effects in the heart of extracts from cardiac muscle, kidney and brain, due to their adenylic acid (AMP) content that was derived from ATP present in the original extract (Gillespie, 1933). In the 1930s there were many investigations into the actions of purines in a variety of tissues, including the cardiovascular system (Lidner and Rigler, 1930; Bennett and Drury, 1931; Wedd, 1931; Ostem and Pamas, 1932; Gaddum and Holtz, 1933) and the uterus (Deuticke, 1932). In the 1950's the first hint that a purine compound, ATP, could possibly be a neurotransmitter came from the studies on vasodilatation in the rabbit ear evoked by antidromic stimulation of its sensory innervation (Horton and Holton, 1953,1954;Holton, 1959). Some years later, after obtaining further evidence from autonomically innervated organs (Bumstock et al., 1970,1972), Bumstock (1972) put forward the Purinergic Hypothesis, stating that ATP (or a closely related substance) was released as a neurotransmitter from non-adrenergic, non-cholinergic motor nerves, for which he coined the name "purinergic nerves." Since this time a substantial body of research has accumulated, showing that purine nucleosides and nucleotides are extensively involved in homeostatic mechanisms, with ATP being utilized as a neurotransmitter or cotransmitter in central, sympathetic, parasympathetic, enteric, and sensory nerves. Both adenosine and ATP are utilized extensively as signal molecules released from many different cell types to influence the activity of many different cell types, for example smooth and cardiac muscle, adipocytes, hepatocytes, blood cells, fibroblasts, and nerve cells throughout the body (for reviews see Stone, 1981, 1991; White 1988; Bumstock, 1990,1993; Cooper and Londos, 1988; Hoyle and Bumstock, 1991a;Hoyle, 1992). As the pharmacological activity of purine compounds was investigated, so a receptor classification developed. This has evolved in a complicated manner (see Abbracchio and Bumstock, 1994). The major classes of purinoceptors (also known as purinergic receptors) are Pi and P2 (Bumstock, 1978), both of which are further subclassified, and are detailed below. The Pi-purinoceptors are preferentially activated by adenosine rather than ADP or ATP, while P2-purinoceptors are

NHx

X)

0C>

I KJ

^^kl-^

H

\ H

NHg

ijc:> N^

N

CH 3 v*"~~~ ^

OH OH

OH OH

N'

CHr- 0 - P - O . • * li O

OH

"N

OH

0-

o-

o-

oII

oli

oII

il 0

» 0

OH

i I I CHr-O-P-O-P-O-P-0-

N

I

OH

1 I

O^

OH

IX >

0' OI I CHr-O-P-O-P-0

^N

*

OH

Figure 1. Structures of purine compounds: a) adenine; b) guanine; c) adenosine; d) guanosine; e) adenosine 5'-monophosphate; f) adenosine 5'-diphosphate; g) adenosine 5'-triphosphate.

51

52

CHARLES H. V. HOYLE and GEOFFREY BURNSTOCK

preferentially activated by ATP or ADP rather than adenosine. The most recent developments in this field are the cloning of Pp and P2-purinoceptors of various subclasses. Guanine derivatives tend to be much less pharmacologically active than the adenine derivatives, and are not thought to have extracellular receptors separately identifiable from those for adenosine or adenine nucleotides. Along with the increasing awareness of the biological activity of purine compounds, a therapeutic potential has been realized. This aspect of purines is very much in its infancy. Nevertheless, a number of clinical areas are being examined, including supraventricular tachycardia, cardiac arrhythmia, cerebral ischemia, stroke, hypertension, reperfiision injury, and certain disorders of the central nervous system, male reproductive system, bladder, and immune system (Marangos, 1991; Williams, 1991,1993;Priebeetal., 1992;Rudolphietal., 1992; Bumstock, 1993).

Pi-PURINOCEPTOR PHARMACOLOGY As defined by Bumstock (1978), Pj-purinoceptors are activated by adenosine, and are blocked by methylxanthine derivatives such as caffeine and theophylline (Sattin and Rail, 1970). In membrane fraction studies (Londos and Wolff, 1977) an extracellular R-site and an intracellular P-site for adenosine binding have been identified, occupation of either affects adenylyl cyclase activity. The R-site, which corresponds with the Pppurinoceptor, has been subclassified on the basis that in different tissues its occupation could lead to either activation (Rg) or inactivation (Rj) of adenylyl cyclase (Londos et al., 1980). The same receptor was also subtyped into A] and A2 (Van Calker et al., 1979), again based on the effect on adenylyl cyclase activity. The Ai and A2 subclasses correspond directly with the Ri and Ra subclasses, respectively; the A1/A2 nomenclature has gained general acceptance. These receptors are widely distributed throughout the body (Table 1). Table 1. Distribution of Ai and A2 Pi-Purinoceptors Nervous system striatum hippocampus cortex glial cells peripheral nerve terminals spinal cord Cardiac muscle Adipocytes

Nervous system striatum cortex glial cells Smooth muscle blood vessels gastrointestinal tract trachea Fibroblasts Hepatocytes Leydig cells Platelets Thymocytes

Purines

53

In addition to their coupling with adenylyl cyclase activity, the subtypes of Pi-purinoceptors have other distinguishing features. First, the Appurinoceptor displays a greater affinity for adenosine in ligand binding studies than does the A2-purinoceptor (Van Calker et al, 1979), and in pharmacological studies adenosine is more potent in Aj systems than in A2 systems (Bumstock and Buckley, 1985). Second, and more important as far as receptor classification is concerned, the Aj and A2 receptors display differential rank orders of agonist potency for a series of adenosine analogues (Burnstock and Buckley, 1985), and selective antagonists have also been developed (Hoyle and Burnstock, 1991b). Ai receptors are preferentially activated by N^-substituted analogues of adenosine, such as R-phenylisopropyladenosine (R-PIA) or cyclohexyladenosine, with 5'-substituted analogues of adenosine, particularly 5'-A^-ethylcarboxamidoadenosine (NECA), being orders of magnitude less active. At A2 receptors the converse is true. Further, the Al receptor displays marked stereo selectivity for the enantiomers R-PIA and S-PIA, with often a 100-fold difference between the potencies of these two compounds, whereas at the A2 receptor the preference for R-PIA over S-PIA is much less marked, or even absent. Xanthine substitution has generated some potent and selective antagonists of adenosine receptors: l,3-dipropyl-8-cyclopentylxanthine (DPCPX) is 500-1,000 times more selective for the Aj subclass (see Hoyle and Bumstock, 1991b). Pharmacologically there are some subtle differences between central and peripheral Al receptors, which have led to a tentative subclassification into Aia and Alb (Gustafsson et al., 1990). The A2 receptor has well defined high and low affinity subtypes (A2a and A2b, respectively) (Daly et al., 1983; Bruns et al., 1986). There are two definitions of an A3 receptor. The first was made on the basis of pharmacological evidence and the fact that this adenosine receptor did not appear to be coupled to adenylyl cyclase, but rather to a calcium mobilization controlling mechanism (Ribeiro and Sebastiao, 1986; Sebastiao and Ribeiro, 1988, 1989). However, the second definition was made on the basis of identification by cloning from rat tissues (Zhou et al., 1992). This latter definition seems to be taking precedence over the former (Abbracchio et al., 1993). Another receptor, A4, has been identified recently (Cornfield et al., 1992) from pharmacological rather than molecular studies. It appears to be similar to the cloned A3 receptor, in that although it is activated by adenosine (and is therefore implicitly a Pppurinoceptor of some sort), it is insensitive to methylxanthine-derived antagonists.

P2-PURINOCEPTOR PHARMACOLOGY The field of P2-purinoceptor classification is somewhat less established than that of the Pppurinoceptor, but nevertheless, there are some clear principles. At present there are five well established divisions of the P2-purinoceptor: P2X, P2Y» Piz? P2T? and P2U (Bumstock and Kennedy, 1985; Gordon, 1986: Hoyle and Bumstock,

54

CHARLES H. V. HOYLE and GEOFFREY BURNSTOCK

1991 a; Hoyle, 1992; Abbracchio et al., 1993) with a few other not-so-well accepted subtypes: P2S (Wiklund and Gustafsson, 1988), P2R (Von Kugelgen and Starke, 1990), and a so-called P3-purinoceptor (Shinozuka et al., 1988) that could be regarded as a P2 variant. The P2N receptor is synonymous with P2U (Abbracchio et al., 1993). The locations of these P2-purinoceptor subtypes are summarized in Table 2. As mentioned above, this area of purine receptor taxonomy is currently under review, and it is likely that the P2X and P2Y subclasses will remain as families, each with several members, with the P2T and P2U receptors being reclassified as members of the P2Y family (Abbracchio and Bumstock, 1994). P2Yi, P2Y2, and P2Y3 receptors have recently been identified following cloning and pharmacological definition utilizing receptors induced inXenopus oocytes (see Barnard et al., 1994). The rationale behind the P2X/P2Y superfamily idea is that all known P2x-purinoceptors appear to be ligand-gated ion channels, and that all the non-P2x-purinoceptors (bar P2Z, whose coupling is unidentified) appear to be coupled to GTP-binding proteins (G-proteins). The definitions of P2-purinoceptors given below are based on the report of the lUPHAR Purine Receptor Nomenclature Subcommittee open meeting (Abbracchio et al., 1993). The major divisions of P2-purinoceptors are represented by the P2X and P2Y subclasses (Bumstock and Kennedy, 1985). This division is based on a differential rank order of agonist potency of synthetic analogues of ATP, and on a differential susceptibility to desensitization to a,P-methylene ATP. At the P2x-purinoceptor a,p-methylene ATP is more potent than 2-methylthio ATP, and this receptor is more readily desensitized by a,p-methylene ATP. In contrast, at the P2Y-purinoceptor 2-methylthio ATP is much more potent than a,P-methylene ATP, and there is a resistance to tachyphylaxis by a,P-methylene ATP. Both these subtypes have a wide distribution throughout the body. In general terms, P2x-purinoceptors Table 2. fix Smooth muscle seminal vesicles vas deferens urinary bladder arteries veins nictitating membrane large intestine Cardiac muscle Fibroblasts Brain Celiac ganglion

Distribution of P2-Purinoceptors P2Y

Smooth muscle arteries veins stomach small intestine large intestine gut sphincters Cardiac muscle Endothelial cells Sensory neurons Pancreatocytes Erythrocytes Hepatocytes Pul monocytes

Piz Mast cells Lymphocytes Macrophages Neutrophils

P21

^2U

Platelets Arteries Megakaryocytes Hepatocytes

Purines

55

mediate contractile responses of smooth muscles while P2Y-purinoceptors mediate relaxant responses. P2T-purinoceptors (Gordon, 1986) have so far only been found on platelets and megakaryocytes, and they mediate platelet aggregation. ADP is the principal agonist for this receptor, with ATP being a competitive antagonist (Macfarlane and Mills, 1975). Adenosine inhibits platelet aggregation, but does not act on this receptor (Haslam and Rosson, 1975). The P2z-purinoceptor is activated by ATP'^, i.e., ATP dissociated from divalent cations. These receptors are found on a variety of blood cells, mediating secretion from mast cells (Cockroft and Gomperts, 1979a,b, 1980; Bennett et al., 1981), inhibition of activity of natural killer lymphocytes and monocyte-derived macrophages (Cameron, 1984; Schmidt et al., 1984), and enhanced generation of superoxide anions in polymorphonuclear neutrophils (Kuhns et al., 1988). Activation of P2z-purinoceptor results in the formation of large nonspecific pores in the cell membrane, allowing ions with large radii to pass through. P2U- or P2N-purinoceptors are activated by pyrimidine compounds, particularly uridine triphosphate (UTP), as well as purine compounds (Von Kiigelgen and Starke, 1990; O'Connor et al., 1991; Keppens et al., 1992; Gerwins and Fredholm, 1992; De Souza et al., 1992; Abbracchio et al., 1993). At present there are few antagonists of P2-purinoceptors that are of much pharmacological value, let alone therapeutic use. Reactive blue 2 is an anthraquinone sulphonic acid derivative that is a P2Y-purinoceptor antagonist, but it has limited use as it is only effective over a narrow range of concentrations, and it is time-dependently Cytotoxic (Bumstock and Warland, 1987b). Arylazidoaminopropionyl ATP (ANAPP3) is a photoaffmity analog of ATP that irreversibly blocks P2x-purinoceptors, and because it binds covalently when irradiated by ultraviolet light it is not competitive. Further, it would also be difficult to use in vivo. Desensitization by a,p-methylene ATP is a useful device for blocking P2x-purinoceptors in vitro, but again is not a competitive antagonist, and therefore has limited value. More recently, the trypanocidal agent, suramin, has been found to be a P2-purinoceptor antagonist (Dunn and Blakeley, 1988). However, its drawbacks are low potency, lack of selectivity among P2-purinoceptor subtypes (Hoyle et al., 1990; Leff et al., 1990), lack of general availability, and also because it is toxic. Therapeutically, it is used in a highly controlled manner against trypanosomiasis and onchocerciasis (Voogd et al., 1993), but its actions are unrelated to P2-purinoceptor blockade.

MOLECULAR BIOLOGY OF PURINOCEPTORS Many groups have succeeded in cloning the Ai, A2a, and A2b subtypes of the Pppurinoceptor (Table 3). The findings have been remarkably consistent, and in all cases the recombinant receptors are typical of G-protein coupled receptors with the characteristic seven membrane-spanning segments (TM1-TM7) being highly

CHARLES H. V. HOYLE and GEOFFREY BURNSTOCK

56

Table 3. Cloned Adenosine (Pi) and ATP (P2) Receptors Subtype

Distribution

of mRNA

Number of Mass (kDa) Amino AcidsJ

References

A,

brain, spinal cord, testis, adipocytes, heart

37

326

A2a

striatum

45 45

412 410

A2b

bladder, large intestine, hippocampus, hypothalamus

36 36

332 328

A3

testis

37

319

Libert e t a l . , 1 9 9 1 , 1992; Reppert et al., 1 9 9 1 ; Mahanetal., 1991; Tucker e t a l . , 1992; O l a h e t a l . , 1992; Salvatoreetal., 1992; Townsend-N icholson and Shine, 1992 Maenhautetal., 1990; Fink e t a l . , 1992; Chern e t a l . , 1992; Furlong et al., 1992; Salvatoreet al., 1992 Stehle et al., 1992; Rivkees and Reppert, 1992; Salvatoreetal., 1992; Pierce e t a l . , 1992 Zhou e t a l . , 1992

brain, skeletal muscle, intestine heart, liver, kidney, lung, testis, brain

41

362

Webb e t a l . , 1993

42

373

Lustigetal., 1993

% P2U

conserved (Figure 2). The first adenosine receptors were cloned almost accidentally. Unidentified sequences, RDC7 and RDC8, had been cloned from a canine thyroid gland complementary DNA (cDNA) library through homology to G-protein coupled receptors (Libert et al., 1989); it was only later that these "orphan receptors" were found to encode the Ai (Libert et al., 1991) and A2a (Maenhaut et al., 1990) receptors, respectively. The cloned A3 receptor, which is not the same as the pharmacological A3 receptor, was originally referred to as TGPCR (Zhou et al., 1992). More recently, two types of P2-purinoceptor have been cloned, P2Y (Webb et al., 1993) and P2U (Lustig et al, 1993). The P2Y receptor was cloned from chick brain; when its cRNA was inserted into oocytes of Xenopus laevis the receptors expressed had a pharmacological profile with ATP and 2-methylthio ATP being equipotent, while a,pmethylene ATP was inert: this is indicative of a novel subtype, which was called P2Yi (Webb et al., 1993). The cloned P2u-purinoceptor has been reclassified as P2Y2 (Barnard et al., 1994), and a third recombinant P2-purinoceptor, which responds particularly to ADP, has been discovered and has been designated P2Y3 (Barnard et al, 1994).

57

Purines

HOOC Figure 2. A representation of the structure of an A2 adenosine receptor determined from its gene sequence. Note the seven membrane-spanning domains (l-VIl), the three extracellular loops (E-l, E-ll, E-lll), two histidinyl groups that may be involved in ligand binding (H), a sodium binding site (Na), four hypothetical disulphide bridges (S-S), and thiol groups (HS). The numbers refer to the residue number in the primary sequence. From Jacobson et al., 1993.

The Al, A2b, and A3 subtypes are very similar to one another; the canine Ai and A2a receptors have a 51% sequence homology in their transmembrane domains (Jacobson et al., 1993). They contain a similar number of amino acid residues (326, 332, and 319 for A^, A2b, and A3, respectively), yielding a predicted molecular weight of approximately 37 kDa. The A2a subtype has a longer intracellular carboxy

58

CHARLES H. V. HOYLE and GEOFFREY BURNSTOCK

terminus, and is correspondingly heavier with a molecular weight of approximately 45 kDa from 410-412 amino acid residues. These calculated molecular weights are in close agreement with those determined from receptor solubilization experiments (see Hoyle and Bumstock, 1991b). The three cloned P2-purinoceptors are similar to the P i -purinoceptors in that they are also typical of G-protein coupled receptors with seven highly conserved transmembrane spanning domains. In size they fall between the Ai or A2b and the A2a receptors with the P2Yi receptor having 362 amino acid residues (41 kDa), and the P2U receptor having 373 amino acid residues (Table 3). However, the structure of the cloned P2-purinoceptors is quite different from that of the Pi-purinoceptors, having at most a 21% sequence identity (Lustig et al., 1993; Webb et al, 1993) (Figure 3, Table 4). From analysis of the amino acid sequences of ten recombinant adenosine receptors a close similarity between the basal portions of TM2, TM3, and TM7 has been noted (Linden et al., 1993). This has been suggested to represent the regions

RDC1 lnterleukin-8 Angiotensin B2 Bradykinin 6-Opioid Pa 2Y2 P2YI

' I ^

I .

• I '

P Thrombin C5a FMLP-like PAF GRP Endothelin B CCK Neuropeptide Y Neuromedin B NK3 NK1 NK2 Adenosine A^ Adenosine Aj CAMP

Figure 3. The G-protein coupled receptor superfamily. In the diagram the closer a pair of members are to one another the greater their sequence coidentity. The P2Y family is most closely related to thrombin, and the adenosine receptors are most closely related to cyclic AMP receptors. From Barnard et al., 1993, q.v. for details of database and sequence analysis.

Purines

59 Table 4, Percentage Identity of Cloned P2YI- and P2U-Purinoceptors With Known G-Protein Coupled Receptors % Identity

Receptor

Species

RDCl Angiotensin 11, type 1 Thrombin Platelet activating factor C5a anaphylatoxin Neuromedin K Interleukin 8 Bradykinin B2 Neurotensin Endothelin B Gastrin-releasing peptide Adenosine A^ Substance P Neurokinin 2 Adenosine Aj cAMP

dog human human guinea pig human rat human rat rat human mouse

dog human human

dog slime mold

'^2Y1

— 27 25 25 23 23 22 22 21 21 21 21 20 20 18 17

Piv 22 25 25

23

<12 <12

Note: Data taken from Webb et a!., 1993 and Lustig et a!., 1993.

of the ligand binding sites in the whole receptor. Further, histidyl residues on TM6 and TM7 may be important: site-specific mutation of Ai receptors involving the replacement of histidine with leucine in TM6 or TM7 causes a significant loss of affinity, especially for agonist molecules (Olah et al., 1992; Jacobson et al., 1993). The mRNA for Ai receptors is predominantly located in the brain, spinal cord, vas deferens, and adipose tissue, with lower levels being expressed in the heart, liver, spleen, and testis (Mahan et al., 1991; Reppert et al., 1991; Tucker et al, 1992; Linden et al., 1993). The mRNA for A2a receptors has been found primarily in striatal neurons, but also in the heart, lung, thymus, and adipose tissue (Stehle et al., 1992), while A2b mRNA is poorly represented in the brain, but is found substantially in smooth muscle of the large intestine and urinary bladder (Stehle et al., 1992). The mRNAs for the recombinant P2Yi and P2Y2-purinoceptors have remarkably disparate, and almost complementary, distributions. The P2Y mRNA is found in brain, intestine, and skeletal muscle, with a low level in the spinal cord (being undetectable in the heart, kidney, liver, lung, and stomach) (Webb et al., 1993). The P2U mRNA is more widely distributed in heart, kidney, liver, lung, testis, and brain (Lustig etal., 1993).

60

CHARLES H. V. HOYLE and GEOFFREY BURNSTOCK

PHYSIOLOGICAL AND PATHOPHYSIOLOGICAL ROLES OF PURINES As might be expected from the broad distribution of purinoceptors throughout the body, the physiological actions of purine compounds are diverse. Within the central nervous system ATP has an excitatory action on many neuronal populations, and may be used as a neurotransmitter by certain neurons. In contrast, adenosine tends to have inhibitory activity, reducing the excitability of central neurons, and decreasing the amounts of transmitter release by nerve terminals. The same is also true in the periphery, with ATP being employed as a neurotransmitter, or cotransmitter with norepinephrine and acetylcholine, by sympathetic and parasympathetic nerves, while adenosine is principally an inhibitory modulator of neurotransmission. There are few places where adenosine has been shown to be a neurotransmitter rather than a neuromodulator, but one example is in cat pelvic (vesicle) ganglia (Akasu et al., 1984). In addition, both adenosine and ATP can be released from non-neuronal cells, and act locally in the manner of a paracrine secretion. Further, both these substances may produce effects via intracellular or extracellular mechanisms (Olsson and Pearson, 1990). Adenosine has even been called a retaliatory metabolite (Newby, 1984) because under stressful conditions leading to reduced levels of intracellular energy, e.g., hypoxia, the levels of adenosine increase sharply, and it is then released from the cell to act on nearby receptors. Its receptor-mediated inhibitory actions prevent cells of the tissue from overactivation. In studying the physiological or pharmacological actions of ATP, care has to be taken in interpreting the results because it can be rapidly degraded by ectoenzymes to adenosine, which itself may exert an effect. The degradation of ATP involves sequential dephosphorylation of its terminal phosphates by ecto-ATPase, ecto-ADPase, and 5'-nucleotidase to form ADP, AMP, and adenosine, respectively (Meghji, 1993). This also often makes it difficult to divorce the physiological actions of ATP and adenosine from one another. Because of the limited space in this chapter only two examples of the physiological actions of purines will be illustrated in detail (the cardiovascular system, and central nervous system). Purines in the Cardiovascular System

In many vascular beds, particularly the coronary circulation, episodes of hypoxia cause vasodilatation that permits an increase in blood flow (i.e., hyperemia). ATP is stored in high concentrations in cardiac endothelial cells, much higher than that in the muscle cells, or that of the purines ADP, AMP, and adenosine (Nees and Gerlach, 1983). Although ATP appears in the coronary outflow following hypoxia, ATP is broken down or removed from the circulation so rapidly that the local concentrations of ATP are much higher than would be suggested by the measure-

Purines

61

ments of ATP in perfusates (Paddle and Burnstock, 1974). Adenosine has been proposed as the mediator of hyperemia (Berne, 1963) because it, and its metabolites, are found in high concentrations in coronary effluent following hyperemia. However, taking into account the greater potency of ATP compared with adenosine at causing vasodilatation (Toda et al, 1982), the fact that ATP is released from endothelial cells during hypoxia, and that ATP can act on endothelial cells to cause release of endothelium-derived relaxing factor (EDRF, or nitric oxide), and that ATP is metabolized rapidly, it is highly probable that ATP is an important mediator of hypoxic or ischemic hyperemia (Burnstock, 1989). This is also supported by the observations that antagonists of Pppurinoceptors affect neither vasodilatation due to ATP nor ischemia (Eikens and Wilcken, 1973; Giles and Wilcken, 1977; Olsson et al., 1978), but they do block adenosine-induced vasodilatation. Cerebral arteries are also more sensitive to ATP than adenosine (Hardebo and Edvinsson, 1979; Forrester et al., 1979; Toda et al., 1982) and it has been suggested that the reactive hyperemia associated with migraine headaches could involve an initial vasodilatation due to ATP (Burnstock, 1981). ATP is involved in many of the hemostatic mechanisms that are continuously in play, and that become prominent at sites of vascular injury. When collagen becomes exposed due to endothelial damage, platelets begin to adhere to it, aggregate and release ADP, which promotes further aggregation via activation of P2T-purinoceptors. ADP can also act on P2Y-purinoceptors, either on endothelial cells or on exposed vascular smooth muscle cells, and thereby evoke dilatation. Thrombin, which is formed during clot-formation, is a powerful stimulant of irreversible platelet aggregation, during which platelets degranulate and undergo morphological changes, releasing ATP, ADP, other vasoactive purines such as adenine dinucleotides (Flodgaard and Klenow, 1982; Liithje and Ogilvie, 1983), and other vasoactive substances such as 5-hydroxytryptamine (Mills et al., 1968). ATP is used as an energy source for mechanisms of clot-retraction, but it can also act on the local P2Y"Purinoceptors. Platelet aggregation is inhibited by ATP and adenine dinucleotides, acting on the platelet P2T-purinoceptor (Harrison et al., 1975; Hoyle, 1990), so high local concentrations of these compounds arising as a result of degranulation could serve to control the extent of aggregation. Thrombin also induces release of ATP from endothelial cells, and causes endothelium-dependent vasodilatation (Pearson and Gordon, 1979; De Mey et al., 1982), and 5-hydroxytryptamine can cause endothelium-dependent vasodilatation or a direct vasodilatation. Thus, when hemostatic mechanisms are invoked during vessel injury, the state of the endothelium will be important for determining the effect on vascular tone. It would appear that if an injured vessel maintains its endothelial cells, local vasodilatation will occur, but if endothelial cells are damaged badly, as might occur in a severe injury, then vasoconstriction would be favored.

62

CHARLES H. V. HOYLE and GEOFFREY BURNSTOCK

In many peripheral arteries stimulation of sympathetic nerves causes vasoconstriction mediated by norepinephrine and ATP released as cotransmitters (Bumstock, 1976, 1990; Ktipferman, 1991; Hoyle, 1992). The purinergic component of the pressor response is blocked by P2-purinoceptor antagonists, such as suramin or ANAPP3, or by desensitization with a,P-methylene ATP, both in vitro and in vivo (Ishikawa, 1985; Kennedy and Bumstock, 1985; Muramatsu and Kigoshi, 1987; BumstockandWarland, 1987a; Bulloch and McGrath, 1988; Schlickeretal., 1989; Hoyle and Bumstock, 1991a; Evans and Surprenant, 1992). Thus, ATP released from sympathetic nerves contributes to an increased peripheral resistance and, therefore, a raised blood pressure. Purinergic mechanisms that are blocked by desensitization to a,P-methylene ATP are involved in the reflexly commanded vasodilatation of the muscular vasculature that occurs as a part of evoked alert-responses in rabbits (Shimada and Stitt, 1984). Similarly, in cardiovascular reflexes evoked by carotid artery occlusion (i.e., the pressor sinocarotid reflex) a,P-methylene ATP can attenuate the pressor response, also indicating a physiological role of ATP in this reflex (Tarasova and Rodionov, 1992). Several blood vessels dilate in response to applied ATP. In some cases this action is mediated via P2Y-purinoceptors on the vascular smooth muscle: for example, as in the rabbit mesenteric artery and portal vein (Kennedy and Bumstock, 1985; Mathieson and Bumstock, 1985; Bumstock and Warland, 1987b; Reilly et al., 1987) or dog basilar and middle cerebellar arteries (Fleetwood and Gordon, 1987; Muramatsu and Kigoshi, 1987). In the dog basilar artery, the relaxant responses are evoked by sympathetic purinergic nerve stimulation (Muramatsu and Kigoshi, 1987), while in the rabbit portal vein purinergic relaxation is mediated via parasympathetic nerves (Bumstock et al., 1979; Reilly et al., 1987). In most blood vessels, however, the site of the P2Y-purinoceptor-mediated vasodilatation is the endothelial cell rather than the smooth muscle. This action was first shown for ATP in the dog femoral artery (De Mey and Vanhoutte, 1981), and has since been shown in many other vessels (Gordon and Martin, 1983; Kennedy et al., 1985; Martin et al., 1985; Houston et al., 1987). Activation of this receptor stimulates the endothelial cell to release EDRF or nitric oxide (Palmer et al., 1987). Relaxant actions of ATP may be important components in the complex physiological regulation of vascular tone. For example, during exercise, when there is an increased demand by skeletal muscle for oxygen, there is a vasodilatation in the muscular vascular bed and an increase of ATP in the venous outflow (Forrester and Lind, 1969). In vessels that contract in response to neurally released ATP acting on P2x-purinoceptors, but relax in response to ATP acting on P2Y-purinoceptors, it is suggested that there is an anatomical separation of these populations of purinoceptors that allows for their independent activation by different physiological stimuli. In most vessels the sympathetic innervation is predominantly adventitial, and these nerve

Purines

63

fibers do not penetrate deeply into the muscle, and it is these more peripheral muscle cells that possess P2x-purinoceptors. This is supported by autoradiographic studies in which P2x-purinoceptors were shown to have greater density in the more peripheral region of the vascular wall, while the more intimal regions had much lower densities (Bo and Bumstock, 1993). The more intimal muscle cells possess P2Y-purinoceptors that are activated by ATP delivered from a luminal source. Similarly, P2Y-purinoceptors on endothelial cells are activated by a luminal source of ATP. As ATP has a very short half-life in whole blood, it is thought that these P2Y-purinoceptors are activated by ATP released locally from endothelial cells or other blood cells, such as platelets, due to physiological or pathophysiological stimuli like hypoxia or damage to the vascular wall. Purines in the Central Nervous System It is only recently that strong evidence has come to light that shows that ATP is utilized as a neurotransmitter by central neurons. In the medial habenula of the rat brain, fast synaptic transmission is attenuated by suramin or following desensitization of P2-purinoceptors by a,p-methylene ATP (Edwards et al., 1992; Benham, 1992; Edwards and Gibb, 1993). Similar results were obtained for neural synapses in the celiac ganglion (Evans et al., 1992; Silinsky et al., 1992). Prior to this there was evidence that ATP could be released from nerves, and that ATP could act postsynaptically, but only in the above studies was synaptic transmission actually studied. ATP is present in high concentrations within the brain, up to 4 mM in some regions (Kogure and Olonso, 1978), and release of ATP from synaptosomal preparations of discrete brain areas has been shown (Barberis and Mcllwain, 1976; White, 1978; Potter and White, 1980; White et al., 1980). When ATP is released it can be rapidly degraded to adenosine (Schubert et al., 1979; James and Richardson, 1993) which may contribute to the observed responses. In the locus coeruleus, stable analogues of ATP excite the nerve cell bodies causing an increase in their firing rate, while ATP itself does not. However, in the presence of an adenosine receptor antagonist, excitatory responses to ATP appear (Tschopl et al., 1992; Harms et al., 1992). Further, in this region adenosine has only inhibitory actions (Regenold and Illes, 1990). Similarly, ATP causes biphasic responses when applied to nerve cells in the trigeminal nucleus, with an initial excitation being followed by a period of depression (Krishtal et al., 1983; Salt and Hill, 1983). The excitatory component is likely to be mediated via P2-purinoceptors, and the depression by Pppurinoceptors, because, as in the locus coeruleus, adenosine has only inhibitory activity. In several regions of the brain, for example in the olfactory cortex, cerebral cortex, and CAl hippocampal neurons, ATP can cause an inhibition of synaptic

64

CHARLES H. V. HOYLE and GEOFFREY BURNSTOCK

transmission that is probably due to its conversion to adenosine and subsequent activation of Pppurinoceptors (Okada and Kuroda, 1975, 1980; Kuroda and Kobayashi, 1975; Phillis et al., 1979; Lee et al, 1981). It is thought that this is so because the responses to ATP are blocked by adenosine deaminase or Pppurinoceptor antagonists (Okada and Kuroda, 1980; Lee et al, 1981; Stone and Cusack, 1989) and also because stable analogues of ATP have no activity (Phillis et al., 1979; Stone and Cusack, 1989). Thus, in these regions of the brain there appear to be presynaptic Pppurinoceptors, but not presynaptic P2-purinoceptors. There is also indirect evidence that ATP is a transmitter from both the peripheral and central terminals of primary sensory neurons. Stimulation of the sensory elements traveling in the rabbit great auricular nerve causes dilatation of the capillary bed in the ear (Holton and Perry, 1951). This is mimicked by ATP, which is found in the venous effluent after nerve stimulation (Holton and Holton, 1953; Holton, 1959). Extracts of the dorsal root ganglia that contain the sensory nerve cell bodies contain high levels of ATP, and a microassay puts the concentration of ATP in these cells at close to 2 mM (Holton and Holton, 1954; Andrews and Holton, 1958; Fukuda et al., 1983). At the central terminals in laminae I and II of the substantia gelatinosa in the dorsal horn of the spinal cord, it has been found that those cell bodies which are postsynaptically activated by mechanical stimulation of the skin (C-fiber afferent input) are also excited by applied ATP (Fyffe and Perl, 1984). Further, synaptosomal preparations of the dorsal horn release ATP in response to a challenge by the sensory neurotoxin capsaicin (White et al., 1985; Sweeney et al., 1989; Sawynok et al., 1993). Physiological roles for adenosine and ATP in synaptic transmission in the dorsal horn of the spinal cord have been reviewed recently (Salter et al., 1993).

THERAPEUTIC POTENTIAL The therapeutic potential for intervention at purinoceptors, whether by agonists or antagonists, is summarized in Table 5. Although there are extensive entries in this Table, it is necessary to point out that the only purinergic drug currently approved for human usage in the United States is adenosine itself (Adenocard) for the management of supraventricular tachycardia and for controlled hypotension in aneurysm surgery (Jacobson et al., 1992; Bumstock, 1993; Williams, 1993). In addition to Pj- or P2-purinoceptor agonists or antagonists having possible therapeutic potential, inhibitors of adenosine uptake or degradation may also be useful. The uptake of adenosine is the major route for its clearance from the extracellular biophase. Inhibition of its uptake (for example by dipyridamole) leads to prolongation of its presence outside the cell, and therefore increased availability for its receptor-mediated activity. The actions of adenosine may be terminated by adenosine deaminase, and this enzyme is also a possible therapeutic target. Both inhibition of adenosine uptake and inhibition of adenosine deaminase have beneficial effects against ischemic brain damage (see Rudolphi et al., 1992).

65

Purines

Table 5, Identified Potential Therapeutic Roles for Purinoceptor Agonists and Antagonists Condition

Agent

Alcohol abuse Anxiety

Adenosine Adenosine

Arthritis Asthma Cancer

Adenosine Pi antagonists ATP

Cerebral ischemia Cognition enhancement

Pi antagonists Pi antagonists

Cystic fibrosis

Pi antagonists

Depression Epilepsy

Pi antagonists Adenosine

Hypertension Male fertility Male impotence Pain

Adenosine ATP Adenosine Adenosine

Parkinson's disease

Pi antagonists

Paroxysmal tachycardia

Pi and P2 antagonists Pi agonists Pi and P2 agonists and antagonists

Psychosis Reperfusion injury

Shock Sleep disorders

Stroke Supraventricular tachycardia

ATP Adenosine

Pi antagonists Adenosine

References Diamond and Gordon, 1991 Marangos and Boulenger, 1985 Marangos, 1991 Green e t a l . , 1991 Williams, 1993 Rapaportand Fontaine, 1989 Fangetal., 1992 H o s o i e t a l . , 1992 Rapaport, 1993 Williams, 1993 Evans e t a l . , 1987 Schingnitz et al., 1991 Dudley e t a l . , 1992 N e h l i g e t a l . , 1992 Eidelman et al., 1992 Stuttsetal., 1992 Sargesetal., 1990 Dunwiddie and Worth, 1982 Dragunow et al., 1985 Dragunow, 1990 Honey e t a l . , 1930 Forestaetal., 1992 Takahashietal., 1992 Vapaataloetal., 1975 Yarborough and McGuffin-Clineschmidt, 1981 Holmgren et al., 1983 Delanderand Hopkins, 1987 D o i e t a l . , 1987 Suzuki e t a l . , 1992 Fuxeetal., 1993 Belhassen and Pelleg, 1984 Heffner e t a l . , 1989 Rudolphi et al., 1992 Boissardetal., 1992 Chaudry, 1990 Radulovacki et al., 1982 W a u q u i e r e t a l . , 1987 Marangos, 1991 Jacobsen and McCall, 1990 Sollevietal., 1984 Pantelyand Bristow, 1990

Note: For reviews see Marangos, 1991: Jacobson et al., 1992; Burnstock, 1993; Williams, 1993.

66

CHARLES H. V. HOYLE and GEOFFREY BURNSTOCK

SUMMARY AND FUTURE DIRECTIONS It is now well established that purine compounds act as potent extracellular messengers in the nervous system, neuromuscular transmission, secretomotor transmission, and in the immune system. A subdivision of receptors into those activated by adenosine (Pi) and those activated by ATP or ADP (P2) is clearly established, and subdivisions of Pp and P2-purinoceptors have been developed on the basis of pharmacological criteria and molecular biological approaches. The regulatory roles of both ATP and adenosine in a number of systems is now recognized. Looking to the future, it is evident that more molecular cloning work is needed for identification of purinoceptor subtypes, and selective agonists and antagonists need to be developed for accurate pharmacological assessment. As far as therapeutic applications are concerned, more research into pathological conditions is required. For example, in the abnormal bladder in disorders such as interstitial cystitis, the purinergic or noncholinergic element of parasympathetic motor transmission becomes much more prominent than in the normal bladder (Hoyle and Bumstock, 1993; Palea et al., 1993), and in hypertension the purinergic component of sympathetic cotransmission may be a site of therapeutic intervention. The role of purines in long-term trophic interactions also needs to be considered. For example, ATP acts as a synergistic agent with growth factors (Rathbone et al., 1992), and adenosine may be involved in vascularization and endothelial cell proliferation (Hudlicka, 1984). ATP may also be involved in apoptotic programmed cell death (Di Virgillo et al., 1990; Zheng et al., 1991; Avery et al., 1992).

REFERENCES Abbracchio, M.P., Cattabeni, F., Fredholm, B.B., & Williams, M. (1993). Purinoceptor nomenclature: A status report. Drug Dev. Res. 28, 207-213. Abbracchio, M.P. & Bumstock, G. (1994). Purinoceptors: Are there families of P2X and P2Y purinoceptors. Pharmacol. Ther. 64,445-475. Akasu, T., Shinnick-Gallagher, P., & Gallagher, J.P. (1984). Adenosine mediates a slow hyperpolarizing synaptic potential in autonomic neurones. Nature 311, 62-65. Andrews, T.M. & Holton, P. (1958). The substance P and adenosinetriphosphate (ATP) contents of sensory nerve on degeneration. J. Physiology 143,45-46P. Avery, R.K., Bleier, K.J., & Pastemack, M.S. (1992). Differences between ATP-mediated cytotoxicity and cell-mediated cytotoxicity. J. Immunology 149, 1265-1270. Barberis, C. & Mcllwain, H. (1976). 5'-Adenine nucleotides in synaptosomal preparations from guinea-pig neocortex: Their change on incubation, superfusion and stimulation. J. Neurochem. 26,1015-1021. Barnard, E.A., Bumstock, G., & Webb, T.E. (1994). G-protein-coupled receptors for ATP and other nucleotides: A new receptor family. Trends Pharmacol. Sci. 15, 67-70. Belhassen, B. & Pelleg, A. (1984). Acute management of paroxysmal supraventricular tachycardia: Verapamil, adenosine triphosphate or adenosine? Am. J. Cardiology 54, 225—227. Benham, CD. (1992). ATP joins the fast lane. Nature 359, 103-104.

Purines

67

Bennett, D.W. & Drury, A.N. (1931). Further observations relating to the physiological activity of adenine compounds. J. Physiology 72, 288-320. Bennett, J.P., Cockroft, S., & Gomperts, B.D. (1981). Rat mast cells permeabilized with ATP secrete histamine in response to calcium ions buffered in the micromolar range. J. Physiology 317, 335-345. Berne, R.M. (1963). Cardiac nucleotides in hypoxia: Possible role in regulation of coronary flow. Am. J. Physiology 204, 317-322. Bo, X. & Bumstock, G. (1993). Heterogeneous distribution of [•^H]a,P-methylene ATP binding sites in blood vessels. J. Vascular Res. 30, 87-101. Boissard, C.G., Lindner, M.D., & Gribkoff, V.K. (1992). Hypoxia produces cell death in the rat hippocampus in the presence of an A} adenosine receptor antagonist: An anatomical and behavioural study. Neuroscience 48, 807-812. Bruns, R.F., Lu, G.H., & Pugsley, T.A. (1986). Characterization of the A2 adenosine receptor labeled by [^H]NECA in rat striatal membranes. Molec. Pharmacology 29, 331-346. Bulloch, J.M. & McGrath, J.C. (1988). Blockade of vasopressor and vas deferens responses by a,p-methylene ATP in the pithed rat. Brit. J. Pharmacology 94, 103-108. Bumstock, G. (1972). Purinergic nerves. Pharmacol. Rev. 24, 509-560. Bumstock, G. (1976). Do some cells release more than one transmitter? Neuroscience 1, 239-248. Bumstock, G. (1978). A basis for distinguishing two types of purinergic receptor. In: Cell Membrane Receptors for Drugs and Hormones: A Multidisciplinary Approach, (Straub, R.W. & Bolis, L., eds.) pp. 107-119, Raven Press, New York. Bumstock, G. (1981). Pathophysiology of migraine: A new hypothesis. Lancet i, 1397-1399. Bumstock, G. (1989). Vascular control by purines with emphasis on the coronary circulation. Eur. Heart J. lO(suppl.F), 15-21. Bumstock, G. (1990). Cotransmission. The Fifth Heymans Lecture—Ghent, Febmary 17, 1990. Arch. Intemat. Pharmacodyn. Therap. 304, 7-33. Bumstock, G. (1993). Physiological and pathophysiological roles of purines: An update. Dmg Dev. Res. 28, 195-206. Bumstock, G. & Buckley, N.J. (1985). The classification of receptors for adenosine and adenine nucleotides. In: Methods Used in Adenosine Research (Methods in Pharmacology Series), (Paton, D.M., ed.) pp. 193-212, Plenum Publishing Corp., New York. Bumstock, G. & Kennedy, C. (1985). Is there a basis for distinguishing more than one type of P2-purinoceptor. Gen. Pharmacology 16, 433-440. Bumstock, G. & Warland, J.J.I. (1987a). A pharmacological study of the rabbit saphenous artery in vitro: A vessel with a large purinergic contractile response to sympathetic nerve stimulation. Br. J. Pharmac. 90, 110-120. Bumstock, G. & Warland, J.J.I. (1987b). P2-purinoceptors of two subtypes in the rabbit mesenteric artery: Reactive blue 2 selectively inhibits the responses mediated via the P2Y- but not the P2x-purinoceptor. Brit. J. Pharmac. 90, 383-391. Bumstock, G., Campbell, G., Satchell, D.G., & Smythe, A. (1970). Evidence that ATP or a related nucleotide is the transmitter substance released by non-adrenergic inhibitory nerve in the gut. Br. J. Pharmac. 40, 668-688. Bumstock, G., Dumsday, B., & Smythe, A. (1972). Atropine-resistant excitation of the urinary bladder: The possibility of transmission via nerve releasing a purine nucleotide. Br. J. Pharmac. 44, 451-461. Bumstock, G., Crowe, R., & Wong, H. (1979). Comparative and histochemical evidence for purinergic inhibitory innervation of the portal vein of the rabbit, but not guinea-pig. Br. J. Pharmac. 65, 377-388. Cameron, D.J. (1984). Inhibition of macrophage mediated cytotoxicity by exogenous adenosine 5'-triphosphate. J. Clin. Lab. Immun. 15,215-218. Chaudry, I.H. (1990). Use of ATP following shock and ischemia. Ann. N.Y. Acad. Sci. 603, 130-141.

68

CHARLES H. V. HOYLE and GEOFFREY BURNSTOCK

Chem, Y., Kingk, K., Lai. H-L., & Lai, H-T. (1992). Molecular cloning of a novel adenosine receptor gene from rat brain. Biochem. Biophys. Res. Commun. 185, 304-309. Cockroft, S. & Gomperts, B.D. (1979a). Activation and inhibition of calcium-dependent histamine secretion by ATP ions applied to rat mast cells. J. Physiol. 296, 243-296. Cockroft, S. & Gomperts, B.D. (1979b). ATP induces nucleotide permeability in rat mast cells. Nature 279, 541-542. Cockroft, S. & Gomperts, B.D. (1980). The ATP"^" receptor of rat mast cells. Biochem. J. 188,789-798. Cooper, D.M.F. 8c Londos, C. (1988). In: Receptor Biochemistry and Methodology Vol. II, Adenosine Receptors Alan R. Liss, New York. Cornfield, L.J., Hu, S., Hurt, S.D., & Sills, M.A. (1992). [^H]2-Phenylaminoadenosine ([^H]CV1808) labels a novel adenosine receptor in rat brain. J. Pharmacol. Exptl. Therap. 263, 552-561. Daly, J.W., Butts-Lamb, P., & Padgett, W. (1983). Subclasses of adenosine receptors in the central nervous system: Interaction with caffeine and related methylxanthines. Cell. Mol. Neurobiol. 3, 69-80. Delander, G.E. & Hopkins, C.J. (1987). Involvement of A2 adenosine receptors in spinal mechanisms of antinociception. Eur. J. Pharmac. 139, 215-223. De Mey, J.G. & Vanhoutte, P.M. (1981). Role of the intima in cholinergic and purinergic relaxation of isolated canine femoral arteries. J. Physiol. 316, 347-355. De Mey, J.G., Claeys, M., & Vanhoutte, P.M. (1982). Endothelium-dependent inhibitory effects of acetylcholine, adenosine triphosphate, thrombin and arachidonic acid in the canine femoral artery. J. Pharmac. Exptl. Therap. 222, 166-173. De Souza, L., Raha, S., Lange, A., & Reed, J.K. (1992). Nucleotide receptors in pheochromocytoma (PC 12) cells. Soc. Neurosci. Abstr. 18, 420.1. Deuticke, H.J. (1932). Uber der Einfluss von Adenosin und Adenosinphosphorsauren auf den isolierten Meerschweinchenuterus. Pflug. Arch. 230, 537-555. Di Virgillo, F., Pizzo, P., Zanovello, P., Bronte, V., & CoUavo, D. (1990). Extracellular ATP as a possible mediator of cell-mediated cytotoxicity. Immun. Today 11,274—277. Diamond, I. & Gordon, A. (1991). Use of adenosine agonists and antagonists in the treatment of alcohol abuse. Eur. Patent Appl. O 431,758 A2. Doi, T., Kuzuma, S., & Maki, Y. (1987). Spinal antinociception effects of adenosine in mice. Eur. J. Pharmac. 137,227-231. Dragunow, M. (1990). Adenosine receptor antagonism accounts for the seizure-prolonging effects of aminophylline. Pharmac. Biochem. Behav. 36, 751-755. Dragunow, M., Goddard, G.V., & Laverty, R. (1985). Is adenosine an endogenous anticonvulsant? Epilepsia 26,480-487. Drury, A.N. & Szent-Gyorgyi, A. (1929). The physiological activity of adenine compounds with special reference to their action upon mammalian heart. J. Physiology 68, 213-237. Dudley, M., Racke, M., Ogden, A.M., Peet, N., Secreset, R., & McDermott, R. (1992). MDL 102,234: A selective adenosine A, receptor antagonist reflecting a new binding mode to the receptor. Abst. Soc. Neurosci. 18,998. Dunn, P.M. & Blakeley, A.G. (1988). Suramin: A reversible P2-purinoceptor antagonist in the mouse vas deferens. Br. J. Pharmac. 93,243-245. Dunwiddie, T.V. & Worth, T. (1982). Sedative and anticonvulsant effects of adenosine in the mouse and in the rat. J. Pharmacol. Exptl. Therap. 220, 70-76. Edwards, F.A., Gibb, A.J., & Colquhoun, D. (1992). ATP receptor-mediated synaptic currents in the central nervous system. Nature 359, 144—146. Edwards, F.A. & Gibb, A.J. (1993). ATP—^ fast neurotransmitter. FEBS Lett. 325, 86-89. Eidelman, O., Guay-Broder, C, van Galen, P.J.M., Jacobson, K.A., Fox, C, Turner, R.J., Cabantchile, Z.I., & Pollard, H.B. (1992). Apadenosine antagonist drugs activate chloride efflux from cystic fibrosis cells. Proc. Natl. Acad. Sci. USA 89, 5562-5566.

Purines

69

Eikens, E. & Wilcken, D.E.L. (1973). The effect of dipyridamole and of aminophylline on responses to sixty-second coronary artery occlusion in dogs. Austral. J. Exptl. Biol. Med. Sci. 51, 631-642. Evans, M.C., Swan, J.C., & Meldrum, B.S. (1987). An adenosine analogue, 2-chloroadenosine, protects against long-term development of ischemic loss in the rat hippocampus. Neurosci. Lett. 83, 287-292. Evans, R.J., Derkach, V., & Surprenant, A. (1992). ATP mediates fast synaptic transmission in mammalian neurones. Nature 357, 503-505. Evans, R.J. & Surprenant, A. (1992). Vasoconstriction of guinea-pig submucosal arterioles following sympathetic nerve stimulation is mediated by the release of ATP. Brit. J. Pharmac. 106,242—249. Fang, W-G., Pimia, F., Bang, Y-J., Myers, C.E., & Trepel, J.B. (1992). P2-purinergic receptor agonists inhibit the growth of androgen-independent prostate carcinoma cells. J. Clin. Invest. 89,191-196. Fink, J.S., Weaver, D.R., Rivkees, S.A., Peter Freund, R.A., Pollack, A.E., Adler, E.M., & Reppert, S.M. (1992). Molecular cloning of the rat A2 adenosine receptor: Selective co-expression with D2 dopamine receptors in rat striatum. Mol. Brain Res. 14, 186-195. Fleetwood, G. & Gordon, J.L. (1987). Purinoceptors in the rat heart. Br. J. Pharmac. 90, 219-227. Flodgaard, H. & Klenow, H. (1982). Abundant amounts of diadenosine 5', 5"'-P',P'^-tetraphosphate are present and releasable, but metabolically inactive, in human platelets. Biochem. J. 208, 737-742. Foresta, C , Rossato, M., & Di Virgilio, F. (1992). Extracellular ATP is a trigger for the acrosome reaction in human spermatozoa. J. Biol. Chem. 267, 1-5. Forrester, T. & Lind, A.R. (1969). Adenosine triphosphate in the venous effluent and its relationship to exercise. Fed. Proc. 28, 1280-1283. Forrester, T., Harper, A.M., MacKenzie, E.T., & Thomson, E.M. (1979). Effect of adenosine triphosphate and some derivatives on cerebral blood flow and metabolism. J. Physiol. 296, 343—355. Fukuda, J., Fujita, Y., & Ohsawa, K. (1983). ATP content in isolated mammalian nerve cells assayed by a modified luciferin-luciferase method. J. Neurosci. Meth. 8, 295-302. Furlong, T.J., Pierce, K.D., Selbie, L.A., & Shine, J. (1992). Molecular characterization of a human brain adenosine A2 receptor. Mol. Brain Res. 15, 62-66. Fuxe, K., Ferre, S., Snaprud, P., von Euler, G., Johansson, B., & Fredholm, B. (1993). Antagonistic A2JD2 receptor interactions in the striatum as a basis for adenosine/dopamine interactions in the central nervous system. Drug Dev. Res. 28, 374—380. Fyffe, R.E.W. & Perl, E.R. (1984). Is ATP a central synaptic mediator for certain primary afferent fibres from mammalian skin? Proc. Natl. Acad. Sci. USA 81, 6890-6893. Gaddum, J.H. & Holtz, P. (1933). The localization of the action of drugs on the pulmonary vessels of dogs and cats. J. Physiol. 77, 139-158. Gerwins, P. & Fredholm, B.B. (1992). ATP and its metabolite adenosine act synergistically to mobilize intracellular calcium via the formation of inositol 1,4,5-triphosphate in a smooth muscle cell line. J. Biol. Chem. 267, 16081-16087. Giles, R.W. & Wilcken, D.E.L. (1977). Reactive hyperaemia in the dog heart: Interactions between adenosine, ATP and aminophylline and the effect of indomethacin. Cardiovas. Res. 11,113—121. Gillespie, J.H. (1933). The biological significance of the linkages in adenosine triphosphoric acid. J. Physiol. 80, 345-359. Gordon, J.L. (1986). Extracellular ATP: Effects, sources and fate. Biochem. J. 233, 309-319. Gordon, J.L. & Martin, W. (1983). Endothelium-dependent relaxation of the pig aorta: Relationship to stimulation of ^^Rb efflux from isolated endothelial cells. Br. J. Pharmac. 76, 531-541. Green, P.G., Basbaum, A.I., Helms, C , & Levine, J.D. (1991). Purinergic regulation of bradykinin-induced plasma extravasation and adjuvant-induced arthritis in the rat. Proc. Natl. Acad. Sci. USA 88,4162-4165. Gustafsson, L.E., Wiklund, C.U., Wiklund, N.P., & Stelius, L. (1990). Subclassification of neuronal adenosine receptors. In: Purines in Cellular Signaling. Targets for Future Drugs (Jacobson, K.A., Daly, J.W., & Manganiello, V., eds.), pp. 200-205, Springer Verlag, New York.

70

CHARLES H. V. HOYLE and GEOFFREY BURNSTOCK

Hardebo, J.E. & Edvinsson, L. (1979). Adenine compounds, cerebrovascular effects in vitro with reference to their possible involvement in migraine. Stroke 10, 58-61. Harms, L., Finta, E.P., Tschopl, M., & Illes, P. (1992). Depolarization of rat locus coeruleus neurons by adenosine 5'-triphosphate. Neurosci. 48, 941-952. Harrison, M.J., Brossmer, R., & Goody, R.S. (1975). Inhibition of platelet aggregation and the platelet release reaction by a,co diadenosine polyphosphates. FEBS Lett. 54, 57-60. Haslam, R.J. & Rosson, G.M. (1975). Effects of adenosine on levels of adenosine cyclic 3',5'-monophosphate in human blood platelets in relation to adenosine incorporation and platelet aggregation. Mol. Pharmac. 11, 528-544. Heffner, T.G., Wiley, J.N., Williams, A.E., Bruns, R.F., Coughenour, L.L., & Downs, D.A. (1989). Comparison of the behavioural effects of adenosine agonists and dopamine antagonists in mice. Psychopharmac. 98, 31-37. Holmgren, M., Hedner, J., Nordberg, G., & Mellstrand, T. (1983). Antinociceptive effects in the rat of an adenosine analogue, A/^-phenylisopropyladenosine. J. Pharm. Pharmacol. 35, 679-680. Holton, P. (1959). The liberation of adenosine triphosphate on antidromic stimulation of sensory nerves. J.Physiol. 145,494-504. Holton, F.A. & Holton, P. (1953). The possibility that ATP is a transmitter at sensory nerve endings. J. Physiol. 119, 50-5IP. Holton, F.A. & Holton, P. (1954). The capillary dilator substance in dry powders of spinal roots: A possible role of adenosine triphosphate in chemical transmission from nerve endings. J. Physiol. 126, 124-140. Holton, P. & Perry, W.L.M. (1951). On the transmitter responsible for antidromic vasodilatation in the rabbits ear. J. Physiol. 114, 240-251. Honey, R.M., Ritchie, W.T., & Thompson, W.A.R. (1930). The action of adenosine on the human heart. Quart. J. Med. 23,485-^90. Hosoi, K., Fujishita, M., Sugita, K., Kurihara, K., Atsumi, T., Murai, T., & Ueha, T. (1992). Pj purinergic receptors and cellular calcium metabolism in A431 human epidermoid carcinoma cells. Am. J. Physiol. 262, C635-C643. Houston, D.A., Bumstock, G., & Vanhoutte, P.M. (1987). Different P2-purinergic receptor subtypes of endothelium and smooth muscle in canine blood vessels. J. Pharmac. Exptl. Therap. 241,501—506. Hoyle, C.H.V. (1990). Pharmacological activity of adenine dinucleotides in the periphery: Possible receptor classes and transmitter function? Gen. Pharmac. 21, 827-831. Hoyle, C.H.V. (1992). Transmission: purines. In: The Autonomic Nervous System, Vol. I: Autonomic Neuroeffector Mechanisms. (Bumstock, G. & Hoyle, C.H.V., eds.), pp. 367-407, Harwood Academic Publishers, Switzerland. Hoyle, C.H.V. & Bumstock, G. (1991a). ATP receptors and their physiological roles. In: Adenosine in the Nervous System (Stone, T.W., ed.), pp. 43-76, Academic Press Ltd., London. Hoyle, C.H.V. & Bumstock, G. (1991b). Purinergic receptors. In: Receptor Data for Biological Experiments: A Guide to Dmg Selectivity (van Meel, J.C.A. & Doods, H.N., eds.), pp. 54-61, Ellis Horwood, Chichester. Hoyle, C.H.V. & Bumstock, G. (1993). Postganglionic efferent transmission in the bladder and urethra. In: Nervous Control of the Urogenital System (Maggi, C.A., ed.), pp. 349-381, Harwood Academic Publishers, Switzerland. Hoyle, C.H.V., Knight, G.E., & Bumstock, G. (1990). Suramin antagonizes responses to P2-purinoceptor agonists and purinergic nerve stimulation in the guinea-pig urinary bladder and taenia coli. Br. J. Pharmac. 99, 617-621. Hudlicka, O. (1984). Development of microcirculation: Capillary growth and adaptation. In: Handbook of Physiology: The Cardiovascular System IV, pp. 165-216, American Physiological Society, Bethesda MD.

Purines

71

Ishikawa, S. (1985). Actions of ATP and a,p-methylene ATP on the neuromuscular transmission and smooth muscle membrane of the rabbit and guinea-pig mesenteric arteries. Br. J. Pharmacol. 86, 777-787. Jacobsen, E.J. & McCall, J.M. (1990). Acute ischemic and traumatic injury to the CNS. Ann. Rep. Med. Chem.25,31-40. Jacobson, K.A., van Galen, P.J.M., & Williams, M. (1992). Adenosine receptors: Pharmacology, structure-activity relationships, and therapeutic potential. J. Med. Chem. 35,407-^22. Jacobson, K.A., van Galen, P.J.M., Ji, X-D, Ramkumar, V., Olah, M.E., & Stiles, G. (1993). Molecular characterization of A] and A2 adenosine receptors. Drug Dev. Res. 28, 226-231. James, S. & Richardson, P.J. (1993). Production of adenosine from extracellular ATP at the striatal cholinergic synapse. J. Neurochem. 60, 219-227. Kennedy, C. & Bumstock, G. (1985). ATP produces vasodilatation via P] purinoceptors and vasoconstriction via P2 purinoceptors in the isolated rabbit central ear artery. Blood Vess. 22, 145-155. Kennedy, C, Delbro, D., & Bumstock, G. (1985). P2-purinoceptors mediate both vasodilatation (via the endothelium) and vasoconstriction of the isolated rat femoral artery. Eur. J. Pharmacol. 107, 161-168. Keppens, S., Vandekerckhove, A., & de Wulf, H. (1992). Extracellular ATP and UTP exert similar effects on rat isolated hepatocytes. Br. J. Pharmac. 105,475-479. Kogure, K. & Olonso, O.F. (1978). A pictorial representation of endogenous brain ATP by a bioluminescent method. Brain Res. 154, 273-284. Krishtal, O.A., Marchenko, S.M., & Pidoplichko, V.I. (1983). Receptor for ATP in the sensory neurons of rats. Neurosci. Lett. 35,41-45. Kuhns, D.B., Wright, D.G., Nath, J., Kaplan, S.S., & Baeford, R.E. (1988). ATP induces transient elevations of [Ca^"^] in human neutrophils and primes these cells for enhanced 02-generation. Lab. Invest. 58, 448-453. Kuroda, Y. & Kobayashi, K. (1975). Increase in cyclic AMP levels by adenosine derivatives in mammalian brain synaptosomes. Proc. Japn. Acad. 54 (B), 243-247. Kupferman, I. (1991). Functional studies of cotransmission. Physiol. Rev. 71, 683-711. Lee, K.S., Schubert, P., Emmert, H., & Kreutzberg, G.W. (1981). Effects of adenosine versus adenine nucleotides on evoked potentials in a rat hippocampal slice preparation. Neurosc. Lett. 23, 309-314. Lefif, P., Wood, B.E., & O'Connor, S.E. (1990). Suramin is a slowly equilibrating but competitive antagonist at P2x-purinoceptors in the rabbit isolated ear artery. Br. J. Pharmacol. 101, 645-649. Libert, F., Parmentier, M., Lefort, A., Dinsart, C, Van-Sande, J., Maenhaut, C, Simons, M.J., Dumont, J.E., & Vassart, G. (1989). Selective amplification and cloning of four new members of the G-protein-coupled receptor family. Science 244, 569-572. Libert, F., Schiffman, S., Lefort, A., Parmentier, M., Gerard, C, Dumont, J.E., Vanderhaegen, J.J., & Vassart, G. (1991). The orphan receptor cDNA RDC7 encodes an Aj adenosine receptor. EMBO J. 10, 1677-1682. Libert, F., Van Sande, J., Lefort, A., Czemilofsky, A., Dumont, J.E., Vassort, G., Ensinger, H.A., & Mendle, K.D. (1992). Cloning and functional characterization of a human Aj adenosine receptor. Biochem. Biophys. Res. Commun. 187, 919-926. Lidner, F. & Rigler, R. (1930). Uber die Beeinflussung der Weite der Herzkranzgefdsse durch Produkte des ZelkemstofFwechsels. Pflug. Arch. 226, 697-708. Linden, J., Tucker, A.L., Robeva, A.S., Graber, S.G., & Munshi, R. (1993). Properties of recombinant adenosine receptors. Drug Dev. Res. 28, 232-236. Londos, C, Cooper, D.M.F., & Wolff, J. (1980). Subclasses of external adenosine receptors. Proc. Natl. Acad. Sci. USA 77, 2551-2554. Londos, C. & Wolff, J. (1977). Two distinct adenosine-sensitive sites on adenylate cyclase. Proc. Natl. Acad. Sci. USA 74, 5482-5486.

72

CHARLES H. V. HOYLE and GEOFFREY BURNSTOCK

Lustig, K.D., Shiau, A.K., Brake, A.J., & Julius, D. (1993). Expression cloning of an ATP receptor from mouse neuroblastoma cells. Proc. Natl. Acad. Sci. USA 90, 5113-5117. Luthje, J. «fe Ogilvie, A. (1983). The presence of diadenosine 5',5"'-P',P^-triphosphate (AP3A) in human platelets. Biochem. Biophys. Res. Commun. 115, 253-260. Macfarlane, D.E. & Mills, D.C. (1975). The effects of ATP on platelets: Evidence against the central role of released ADP in primary aggregation. Blood, 46, 309-320. Maenhaut, C, van Sande, S.J., Libert, F., Abramowicz, M., Parmentier, M., Vanderhaegen, J.J., Dumont, J.E., Vassart, G., & Schiffman, S. (1990). RDC8 codes for an adenosine A2 receptor with physiological constitutive activity. Biochem. Biophys. Res. Commun. 173, 1169-1178. Mahan, L.C;, McVittie, L.D., Smyk-Randall, E.M., Nakata, H., Monsma, F.J., Gerfen, C.R., & Sibley, D.R. (1991). Cloning and expression of an A1 adenosine receptor from rat brain. Mol. Pharmacol. 40, 1-7. Marangos, P.J. (1991). Potential therapeutic roles for adenosine in neurologic disease. In: Adenosine in the Nervous System (Stone, T.W., ed.), pp. 217-265, Academic Press, London. Marangos, P.J. & Boulenger, J.P. (1985). Basic and clinical aspects of adenosinergic neuromodulation. Neurosci. Biobehav. Rev. 9,421-430. Martin, W., Cusack, N.J., Carleton, S., & Gordon, J.L. (1985). Specificity of the P2-purinoceptor mediates endothelium-dependent relaxation of the pig aorta. Eur. J. Pharmacol. 108, 295-299. Mathieson, J.J.I. & Bumstock, G. (1985). Purine-mediated relaxation and constriction of isolated rabbit mesenteric artery are not endothelium-dependent. Eur. J. Pharmacol. 118, 221-229. Meghji, P. (1993). Storage, release, uptake, and inactivation of purines. Drug Dev. Res. 28, 214-219. Mills, D.C.B., Robb, I.A., & Roberts, G.C.K. (1968). The release of nucleotides, 5-hydroxytryptamine and enzymes from human blood platelets during aggregation. J. Physiol. 195, 715-729. Muramatsu, I. & Kigoshi, S. (1987). Purinergic and non-purinergic innervation in the cerebral arteries of the dog. Br. J. Pharmacol. 92,901-908. Nees, S. & Gerlach, E. (1983). Adenine nucleotide and adenosine metabolism in cultured coronary endothelial cells: Formation and release of adenine compounds and possible functional implications. In: Regulatory Functions of Adenosine (Berne, R.M., Rail, T.W., & Rubio, R., eds.). pp. 347-360, Martinus Nijhofif, The Hague. Nehlig, A., Daval, J.L., & Debry, G. (1992). Caffeine and the central nervous system: Mechanisms of action, biochemical, metabolic and psychostimulant effects. Brain Res. Rev. 17, 139-170. Newby, A.C. (1984). Adenosine and the concept of "retaliatory metabolites." Trends Bioch. Sci. 9, 42-44. O'Connor, S.E., Dainty, I.A., & Leff, P. (1991). Further subclassification of ATP receptors based on agonist studies. Trends Pharmac. Sci. 12, 137-141. Okada, Y. & Kuroda, Y. (1975). Inhibitory action of adenosine and adenine nucleotides on the postsynaptic potential of olfactory cortex slices of the guinea pig. Proc. Japn. Acad. 51,491-494. Okada, Y. & Kuroda, Y. (1980). Inhibitory actions of adenosine and adenosine analogues on neurotransmission in the olfactory cortex slice of guinea pig—structure-activity relationships. Eur. J. Pharmacol. 61, 137-146. Olah, M.E., Ren, H., Ostrowski, J., Jacobson, K.A., & Stiles, G.L. (1992). Cloning, expression, and characterization of the unique bovine A] adenosine receptor. J. Biol. Chem. 267, 10764-10770. Olsson, R.A., Snow, J.A., & Gentry, M.K. (1978). Adenosine metabolism in canine myocardial reactive hypertension. Circ. Res. 42, 358-362. Olsson, R.A. & Pearson, J.D. (1990). Cardiovascular purinoceptors. Physiol. Rev. 70, 761-845. Ostem, P. & Pamas, J.K. (1932). Uber die Auswertung von Adenosinderivaten am iiberlebenden Froscherz. Biochem Zeitsherift, 248, 389-397. Paddle, B.M. & Bumstock, G. (1974). Release of ATPfromperfused heart during coronary vasodilatation. Blood Vessels 11,110-119. Palea, S., Artibani, W., Ostardo, E., Trist, D.G., & Pietra, C. (1993). Evidence for purinergic neurotransmission in human urinary bladder affected by interstitial cystitis. J. Urol. 150, 2007-2012.

Purines

73

Palmer, R.M.J., Ferrige, A.G., & Moncada, S. (1987). Nitric oxide release accounts for the biological activity of endothelium-derived relaxing factor. Nature 316, 724-725. Pantely, G.A. & Bristow, J.D. (1990). Adenosine. Renewed interest in an old drug. Circulation 82, 1854-1856. Pearson, J.D. & Gordon, J.L. (1979). Vascular endothelial and smooth muscle cells in culture selectively release adenine nucleotides. Nature 281, 384-386. Phillis, J.W., Edstrom, J.P., Kostoploulos, G.K., & Kirkpatrick, J.R. (1979). Effects of adenosine and adenine nucleotides on synaptic transmission in the cerebral cortex. Can. J. Physiol. Pharmacol. 57,1289-1312. Pierce, K.D., Furlong, T.J., Selbie, L.A., & Shine, J. (1992). Molecular cloning and expression of an adenosine A2b receptorfromhuman brain. Biochem. Biophys. Res. Commun. 187, 86-93. Potter, P. & White, T.D. (1980). Release of adenosine 5'-triphosphate from synaptosomes from different regions of rat brain. Neuroscience 5, 1351-1356. Priebe, T., Platzoncas, CD., & Nelson, J.A. (1992). Adenosine receptors and modulation of natural killer cell activity by purine nucleosides. Cancer Res. 50, 4328—4331. Radulovacki, M., Miletich, R.S., & Green, R.D. (1982). A/^-(L-Phenylisopropyl)adenosine (L-PIA) increases slow-wave sleep (S2) and decreases wakefulness in rats. Brain Res. 246, 178-180. Rapaport, E. (1993). Anticancer activities of adenine nucleotides in tumour bearing hosts. Drug Dev. Res. 28,428-431. Rapaport, E. & Fontaine, J. (1989). Anticancer activities of adneine nucleotides in mice are mediated through expansion of erj^hrocyte ATP pools. Proc. Natl. Acad. Sci. USA 86, 1662—1666. Rathbone, M.P., Deforge, S., Deluca, D., Gabel, B., Laurenssen, C, Middlemiss, P., & Parkinson, S. (1992). Purinergic stimulation of cell division and differentiation mechanisms and pharmacological implications. Medical Hypotheses 37, 213-219. Regenold, J.T. & Illes, P. (1990). Inhibitory adenosine A1-receptors on rat locus coeruleus neurons. An intracellular electrophysiological study. Naunyn Schmiedeberg's Arch. Pharmacol. 341,225—231. Reilly, W.M., Saville, V.L., & Bumstock, G. (1987). An assessment of the antagonistic activity of reactive blue 2 at Pp and P2-purinoceptors: Supporting evidence for purinergic innervation of the rabbit portal vein. Eur. J. Pharmacol. 140, 47—53. Reppert, S.M., Weaver, D.R., Stehle, J.H., & Rivkees, S.A. (1991). Molecular cloning of a rat A1-adenosine receptor that is widely expressed in brain and spinal cord. Mol. Endocrinol. 5, 1037-1048. Ribeiro, J.A. & Sebastiao, A.M. (1986). Adenosine receptors and calcium: Basis for proposing a third (A3) adenosine receptor. Progr. Neurobiol. 26, 179-209. Rivkees, S.A. & Reppert, S.M. (1992). RFL9 encodes an A2b-adenosine receptor. Mol. Endocrinol. 5, 1037-1048. Rudolphi, K.A., Schubert, P., Parkinson, F.E., & Fredholm, B.B. (1992). Neuroprotective role of adenosine in cerebral ischaemia. Trends Pharmacol. Sci. 13, 439-445. Salt, T.E. & Hill, R.G. (1983). Excitation of single sensory neurones in the rat caudal trigeminal nucleus by iontophoretically applied adenosine 5'-triphosphate. Neurosci. Lett. 35, 53-57. Salter, M.W., De Koninck, Y., & Henry, J.L. (1993). Physiological roles for adenosine and ATP in synaptic transmission in the spinal dorsal horn. Prog. Neurobiol. 41, 125—156. Salvatore, C.A., Luneau, C.J., Johnson, R.G., & Jacobson, M.A. (1992). Molecular cloning and characterization of human A/ and A2 adenosine receptors. Int. J. Purine Pyrim. Res. 3, 82. Sarges, R., Howard, H.R., Browne, R.G., Lebel, L.A., Seymour, P.A., & Koe, B.K. (1990). 4-Amino[l,2,4]triazolo[4,3a]-quinoxalines. A novel class of potent adenosine antagonists and potential rapid-onset antidepressants. J. Med. Chem. 33, 2240-2254. Sattin, A. & Rail, T.W. (1970). The effect of adenosine and adenine nucleotides on the cyclic adenosine 3',5'-phosphate content of guinea-pig cerebral slices. Mol. Pharmacol. 6, 13-23. Sawynok, J., Downie, J.W., Reid, A.R., Cahill, C, & White, T.D. (1993). ATP release from dorsal spinal cord synaptosomes: Characterization and neuronal origin. Brain Res. 610, 32—38.

74

CHARLES H. V. HOYLE and GEOFFREY BURNSTOCK

Schingnitz, G., Kufner-Muhl, U., Ensinger, H., Lehr, E., & Kuhn, F.J. (1991). Selective A,-antagonists for treatment of cognitive deficits. Nucleos. Nucleot. 10, 1067-1076. Schlicker, E., Urbanek, E., & Goethert, M. (1989). ATP, a,p-methylene ATP and suramin as tools for characterization of vascular P2X receptors in the pithed rat. J. Autonomic Pharmacol. 9, 371-380. Schmidt, A., Ortaldo, J.R., & Herberman, R.B. (1984). Inhibition of human natural killer cell reactivity by exogenous adenosine 5'-triphosphate. J. Immunol. 132, 146-150. Schubert, P., Komp, W., & Kreutzberg, G.W. (1979). Correlation of 5'-nucleotidase activity and selective transneuronal transfer of adenosine in the hippocampus. Brain Res. 168, 419-424. Sebastiao, A.M. & Ribeiro, J.A. (1988). On the adenosine receptor and adenosine inactivation at the rat diaphragm neuromuscular junction. Br. J. Pharmacol. 94, 109-120. Sebastiao, A.M. & Ribeiro, J.A. (1989). 1,3,8- and 1,3,7-substituted xanthines: Relative potency as adenosine receptor antagonists at thefrogneuromuscular junction. Br. J. Pharmacol. 96,211-219. Shimada, S.G. & Stitt, J.T. (1984). An analysis of the purinergic component of active muscle vasodilatation obtained by stimulation of the hypothalamus in rabbits. Br. J. Pharmacol. 83, 577-589. Shinozuka, K., Bjur, R., «fe Westfall, D.P. (1988). Characterization of prejunctional purinoceptors on adrenergic nerves of the rat caudal artery. Naunyn Schmiedeberg's Arch. Pharmacol. 338, 221-227. Silinsky, E.M., Gerzanich, V., & Vanner, S.M. (1992). ATP mediates excitatory synaptic transmission in mammalian neurones. Br. J. Pharmacol. 106, 762-763 Sollevi, A., Lagerkranser, M., Irestedt, L., Gordon, E., & Lindqvist, C. (1984). Controlled hypotension with adenosine in cerebral aneurism surgery. Anesthesiology 61,400-405. Stehle, J.H., Rivkees, S.A., Lee, J.J., Weaver, D.R., Deeds, J.D., & Reppert, S.M. (1992). Molecular cloning and expression of the cDNA for a novel A2-adenosine receptor. Mol. Endocr. 6,384—393. Stone, T.W. (1981). Physiological roles for adenosine and adenosine 5'-triphosphate in the nervous system. Neuroscience 6, 523-555. Stone, T.W. (ed.) (1991). In: Adenosine in the Nervous System, Academic Press, London. Stone, T.W. & Cusack, N.J. (1989). Absence of P2-purinoceptors in hippocampal pathways. Br. J. Pharmacol. 97, 631-^35. Stutts, M.J., Chinet, T.C., Mason, S.J., Fulton, J.M., Clarke, L.L., & Boucher, R.C. (1992). Regulation of C\~ channels in normal and cystic fibrosis airway epithelial cells by extracellular ATP. Proc. Natl. Acad. Sci. USA 89, 1621-1625. Suzuki, F., Shimada, J., Nonaka, H., Ishii, A., & Ichikawa, S. (1992). Potent and selective adenosine A2 antagonist. Intl. J. Purine Pyrim. Res. 3, 107. Sweeney, M.I., White, T.D., & Sawynok, J. (1989). Morphine, capsaicin and K"^ release purine from capsaicin-sensitive primary afferent nerve terminals in the spinal cord. J. Pharmac. Exptl. Therap. 248, 447-454. Takahashi, Y., Ishii, N., Lue, T.F., & Tanagho, E.A. (1992). Effects of adenosine on canine penile erection. J. Urol. 148, 1323-1325. Tarasova, O.S. & Rodionov, l.M. (1992). The role of purinergic neurotransmission in various cardiovascular reflexes. Acta Physiol. Scand. 146, 441—448. Toda, N., Okunishi, H., Taniyama, K., & Miyazaki, M. (1982). Responses to adenine nucleotides and related compounds in isolated dog cerebral, coronary and mesenteric arteries. Blood Vessels 19, 226-236. Townsend-Nicholson, A. & Shine, J. (1992). Molecular cloning and characterization of a human brain A, adenosine receptor cDNA. Mol. Brain Res. 16, 365-370. Tschopl, M., Harms, L., Norenberg, W., & Illes, P. (1992). Excitatory effects of adenosine 5'-triphosphate on rat locus coeruleus neurones. Eur. J. Pharmacol. 213, 71—77. Tucker, A.L., Linden, J., Robeva, A.S., D'Angelo, D., & Lynch, K.R. (1992). Cloning and expression of a bovine adenosine A, receptor cDN A. FEBS Lett. 297, 107-111.

Purines

75

Van Calker, D,, Muller, M., & Hampbrecht, B. (1979). Adenosine regulates via two different types of receptors, the accumulation of cyclic AMP in cultured brain cells. J. Neurochem. 33, 999-1005. Vapaatalo, H., Onken, D., Neuvonen, P.J., & Westermann, E. (1975). Stereospecificity in some central and circulatory effects of phenylisopropyl-adenosine (PIA). Arzneim. Forsch. 25, 407—410. Von Kugelgen, I. & Starke, K. (1990). Evidence for two separate vasoconstriction-mediating nucleotide receptors, both distinct from the P2X receptor, in rabbit basilar artery: A receptor for pyrimidine nucleotides and a receptor for purine nucleotides. Naunyn Schmiedeberg's Arch. Pharmac. 341, 538-546. Voogd, T.E., Vansterkenburg, L.M., Wilting, J., & Janssen, L.H.M. (1993). Recent research on the biological activity of suramin. Pharmac. Rev. 45, 177-203. Wauquier, A., Van Belle, H., Van den Broek, W.A.E., & Janssen, P.A.J. (1987). Sleep improvement in dogs after oral administration of mioflazine, a nucleoside transport inhibitor. Psychopharmacol. 91,434-439. Webb, T.E., Simon, J., Krishek, B.J., Bateson, A.N., Smart, T.G., King, B.F., Bumstock, G., & Barnard, E.A. (1993). Cloning and functional expression of a brain G-protein-coupled ATP receptor. FEBS Lett. 324, 219-225. Wedd, A.M. (1931). The action of adenosine and certain related compounds on the coronary flow of the perfused heart of the rabbit. J. Pharmac. Therap. 41, 355-366. White, T.D. (1978). Release of ATP from a synaptosomal preparation by elevated extracellular K"^ and by veratridine. J. Neurochem. 30, 329-336. White, T.D. (1988). Role of adenine compounds in autonomic neurotransmission. In: Pharmacology and Therapeutics Vol. 38 (Bell, C , ed.), pp. 129-168, Pergamon Press, London. White, T.D., Potter, P., & Wonnacott, S. (1980). Depolarisation-induced release of ATP from cortical synaptosomes is not associated with acetylcholine release. J. Neurochem. 34, 1109-1112. White, T.D., Downie, J.W., & Leslie, R.A. (1985). Characteristics of K"^ and veratridine-induced release of ATP from synaptosomes prepared from the myenteric plexus of the guinea-pig small intestine. J. Neurochem. 40, 1069-1075. Wiklund, N.P. & Gustafsson, L.E. (1988). Indications for P2-purinocpetor subtypes in guinea-pig smooth muscle. Eur. J. Pharmac. 148, 361-370. Williams, M. (1991). Purinergic receptors as potential drug targets. Drug News Persp. 4, 5-12. Williams, M. (1993). Purinergic drugs: Opportunities in the 1990s. Drug Dev. Res. 28, 438-444. Yarborough, G.G. & McGuffm-Clineschmidt, J.C. (1981). In vivo behavioural assessment of central nervous system purinergic receptors. Eur. J. Pharmac. 76, 137-144. Zheng, L.M., Zychlinsky, A., Liu, C-C, Ojcius, D.M., & Ding-E Young, J. (1991). Extracellular ATP as a trigger for apoptosis or programmed cell death. J. Cell Biol. 112, 279-288. Zhou, F.Q-Y., Olah, M.E., Li, C, Johnson, R.A., Stiles, G.L., & Civelli, O. (1992). Molecular cloning and characterization of a novel adenosine receptor: The A3 adenosine receptor. Proc. Natl. Acad. Sci. USA 89, 7432-7436.

This Page Intentionally Left Blank

Chapter 3

The Role of Multiple Isozymes in the Regulation of Cyclic Nucleotide Synthesis and Degradation J. KELLEY BENTLEY and JOSEPH A. BEAVO

Introduction The Adenylate Cyclase Family The Type I Enzyme Is Calmodulin (CaM)-Stimulated The Type II Adenylate Cyclase Is Stimulated by Ga + Gp,Yg Type III Is Stimulated by CaM and Higher Calcium Concentrations Than Type I The Type IV Enzyme Resembles Type 11 but Is Widely Distributed The Type V Isoform Is Enriched in the Heart and Possesses Alternative Splice Variants The Type VI Resembles Type V but Is More Widely Distributed Other Isoforms of Adenylate Cyclase With Different Regulatory Properties Remain To Be Characterized More Fully The Guanylate Cyclase Family The Widely Distributed GC-A Form Binds and Is Activated by ANP in Intact Cells or Membranes GC-B Is the C-ANP Receptor in the Nervous System but not the 80 kDa ANP-C Receptor The GC-C Form Is Enriched in the Intestine The GC-C Form Is Enriched in the Olfactory Epithelia Diverse Forms of Particulate Guanylate Cyclase Are Found in the Retina The Soluble Guanylate Cyclases Are Activated Through Their Heme Functional Group The Cyclic Nucleotide Phosphodiesterase Family Conserved Structural Elements

Principles of Medical Biology, Volume 4 Cell Chemistry and Physiology: Part II, pages 77-121 Copyright © 1996 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-806-4

11

78 79 81 82 82 83 83 84 84 85 88 88 89 89 89 90 91 91

78

J. KELLEY BENTLEY and JOSEPH A. BEAVO

Type I PDEs (Ca^'^/CaM-Stimulated) Type II PDEs (cGMP-Stimulated) Type III PDEs (c-GMP-Inhibited) Type IV PDEs (cAMP-Specific) Type V PDEs (cGMP-Specific) Type VI PDEs (cGMP-Binding) Type VII PDEs (HCPI Family) Conclusion

*

94 99 102 105 106 108 109 109

INTRODUCTION As with all metabolites, the concentrations of cAMP and cGMP in a cell are regulated by their rates of synthesis and degradation. More importantly, each cell type in the body can control its cyclic nucleotide concentration in a unique fashion. This requires distinct mechanisms for regulation of cyclic nucleotide synthesis and degradation for each cell type. It is now recognized that a major mechanism by which differential control is accomplished is through the selective expression of various combinations of distinct cyclic nucleotide cyclases and phosphodiesterases. When coupled with differential, selective expression of receptors and effectors that modulate the activity of these enzymes, each cell type can exhibit a unique and extremely sensitive regulation of cAMP and cGMP. This diversity often also leads to rather complex regulatory pathways for the control of cAMP or cGMP in the cell. For example, some cells respond to hormones that increase intracellular calcium by increasing cGMP levels. This is often accomplished by increasing the activity of nitric oxide synthase which, in turn, regulates a soluble guanylate cyclase in the same or neighboring cell. Other cells might respond to the same calcium signal by lowering cGMP. This can be accomplished because that cell type expresses a calcium and calmodulin dependent phosphodiesterase. Because different isoforms of cyclase and phosphodiesterase have different kinetic properties and respond to different signals, a wide variety of regulatory responses are possible. The repertoire of enzymes expressed can alter not only long term changes in steady state levels of cyclic nucleotide but also the magnitude and duration of a cyclic nucleotide signal caused by a pulse in hormone or neurotransmitter release. Since the turnover of cyclic nucleotides in most cells is very rapid, small changes in activity of either the cyclases or phosphodiesterases can make a large and rapid difference in steady state concentration. It is clear that in order to understand the differential regulation of cyclic nucleotide synthesis and degradation in a specific tissue, we must understand both cellular and molecular details of both synthesis and degradation. Many of the recent studies in this area have focused on how each cyclase and phosphodiesterase is regulated. Since new isozyme members in each family of cyclase and phosphodiesterase are being discovered at a relatively rapid rate, studies on regulation and control are only just being initiated for many of them. Questions of current interest include: What kinds of enzymes, and which isoforms are present

Regulation of cNMP Synthesis and Deregulation in a particular tissue? What is the cellular or subcellular localization of the isoenzyme, and how does its localization contribute to its role in cellular physiology? How are the activities of the distinct proteins regulated? What cellular effectors act on each isoenzyme? How does the enzyme interact with other regulatory systems to form an integrated signal response mechanism? How is the length and magnitude of the cAMP and cGMP signal controlled in a given cell type and how is this influenced by the isozymes of cyclase and phosphodiesterase expressed? What types of feedback regulation (or in some cases perhaps feed forward regulation) are present? Molecular cloning and characterization of the expressed products for the different isoenzymes of cyclase and phosphodiesterase has allowed a variety of different possible regulatory mechanisms to be deduced. It has also allowed structural motifs common to members of a distinct family to be defined. One goal of this review is to demonstrate how functional differences result from unique regulatory characteristics intrinsic to each isoform. Another is to collate and present what is currently known about the relationships of structure to function for these diverse isozyme families. In order to do this, we will first survey the diversity of isoforms known, and briefly examine the structural components that identify them as discretely regulated functional entities. Throughout, references will be made to other current reviews which treat the enzymes discussed briefly here in more depth. Finally, we will provide current representative examples demonstrating how regulation of the distinct cyclase and phosphodiesterase isoforms integrate to produce a precise cellular response.

THE ADENYLATE CYCLASE FAMILY Adenylate cyclases share a common structural motif. Hormones have long been known to stimulate the production of cAMP through specific receptors which, in turn, are coupled through GTP-binding proteins to specific effector enzymes. Only recently has the diversity of GTP binding proteins and effector enzymes been appreciated. Cellular cAMP has recently been shown to be synthesized by multiple isozymes of adenylate cyclase (Krupinski, 1991; Tang and Oilman, 1992; Choi, 1993; Taussig and Oilman, 1995). While each of these enzymes contains the same basic structural motif (see Figure 1), they differ with respect to their interactions with other effector molecules such as calcium, or OTP-binding protein a, p, or y subunits (see Table 1). These distinctions allow a cell-specific response to receptor occupation, dependent upon the kinetics of activation or inhibition that specific effectors have upon the unique isoform. Based on data obtained using the polymerase chain reaction (PCR), the existence of at least ten different adenylate cyclase isoforms have been reported (Krupinski et al., 1992; Taussig and Oilman, 1995). Sequence comparison of the different adenylate cyclase isozymes shows that homology is limited to two large intracellular hydrophilic domains about 30% identical to each other. In all mammalian adenylate cyclases studied thus far, one

79

I E >^ N

O

u >^ u

<

I

^1

If

00 0^

"^ ^91^0> ro-7^

5; _^ ^^ "S

0

r-"

; i ' -D ^

o

s

^

0 (N Q. 0^

03 w

O O^ 0^ O r-

OJ

C2 E_r "DC

r=

"". c 03

5 «

DO

a^ o^ ^ CO

>.

O T3

tn t/> C

p c

in

o o o

•^ 'c 'c • ^

_• "o

0^ T—

^

•Q.

{fi

o

.^ 'c • •<

0 0 00 00 05 05 05

^jj

.E c OS — 2 ^ c ^ ^

O -J

O ^ _i X

-b c/i

>« >- >.

(U 0) (U Q. Q. Q.

^? >«

0)

a

^

•- § "J _: 2

00 OJ

05

-t-

<1»

II c

^ a; (U ^

OS E _ : 05 05 ^

_-:

05

it

11-25

0

~

c

05

E o >- c 05 0

O

I > Q.

Q.

,11 I^

I I

u

03

_;- Ji u

5 +

c

j£)

u Si

<5

"S' 53 "2 ^^•^ o .E Cj 05

CO

^

m

^

>

O

«N

>

Q.

=

^

o -B .

a> ^ .

3 X

O 2

if

II5{

O --

nil c

u 03

'E ^

en:

o u

5

2 "o

03

=

P I

8

2-Q. 0)

Q.

80

^

Regulation of cNMP Synthesis and Deregulation

i

dopamine extracellular

cytosol Figure 1. The type I adenylate cyclase Is activated by both Ca^VCaM and Gas. In receptor-mediated adenylate cyclase activation, a seven-transmembrane domain receptor binds Its ligand and stimulates adenylate cyclase through the action of the a subunit of a GTP-binding protein (Gas). Some interaction between the CI and C2 homologous catalytic domains of the cyclase is required. In some cells, for example neurons containing a Di dopaminergic system, stimulation of cyclase is also accompanied by the mobilization of intracellular Ca^-'CMahan et al., 1990; Monsma et al., 1990). A positive feedback in this system would be accomplished through the action of Ca^VCaM, or through the action of protein kinase C (PKC)-mediated protein phosphorylation at a site not yet specified (*). The py subunits of the GTP-binding protein would inhibit the system either through sequestration of Gas or through a direct interaction with adenylate cyclase. Depending on the cell type, direct stimulation of a CaM-sensitive adenylate cyclase is also possible.

central hydrophilic domain (CI) separates the two sets of six putative membranespanning domains, while the second hydrophilic domain (C2) is found on the carboxy-(C-) terminal end of the molecule. Overall, the protein resembles an ion channel or transporterlike structure, although channellike activity has never been described for it (Krupinski et al., 1989). However, a voltage sensitive adenylate cyclase has recently been reported (Reddy et al., 1995). The Type I Enzyme Is Calmodulin (CaM)-Stimulated

The CaM-activated type I adenylate cyclase transcript encodes a 120 kDa enzyme with intracellular amino-(N-) and C-termini (Krupinski et al., 1989). There is some functional interaction between the two cytoplasmic domains; when each half of the molecule is coexpressed together, catalytic activity is seen. Both hydrophilic domains are needed for enzymatic activity (Tang and Oilman, 1991).

81

82

J. KELLEY BENTLEY and JOSEPH A. BEAVO

In addition to being stimulated by the a subunit of the stimulatory GTP-binding protein, Gsa, the type I enzyme is inhibited by the GTP-binding protein Py complex (Gpy) (Tang et al, 1991). Other effectors also impinge directly upon the type I adenylate cyclase. Phorbol esters, for example, may stimulate the activity of the type I enzyme through Ga^"^-independent C-kinase activation (Choi et al., 1993; Jacobowitz et al., 1993). The type I enzyme is found largely in the brain or in neuronal structures outside the brain (Xia et al., 1993). The distribution of the type I enzyme supports a possible role for this enzyme in associative memory and long-term potentiation of synaptic transmission (Xia et al., 1991). Antisense probes to the type I transcript hybridize strongly with the neocortex, with hippocampus granule cells of the dentate gyrus, and with pyramidal cells in the hippocampal CA1 -CA3 regions (Xia et al., 1991; Matsuoka et al., 1992). Mice in which this gene has been inactivated by targeted disruption show deficient spatial memory and altered long term potentiation (Wu et al., 1995). Furthermore the normal developmental expression of this isoform is disrupted in the mutant mice (Villacres et al., 1995). This enzyme may function as a coincidence detector integrating both Ca"*"*" and cAMP coactivation of gene transcription (Impey et al., 1994). The Type II Adenylate Cyclase Is Stimulated by Ga + Cp.y

The 123 kDa type II adenylate cyclase is also enriched in the nervous system, but it is not stimulated by Ca^VCaM (Feinstein et al., 1991) and may be somewhat inhibited by Ca^"^. The type II enzyme is also reported to be stimulated by protein kinase C (Yoshimura and Cooper, 1993). The type II adenylate cyclase, unlike the type I isoform, is stimulated by GTP-binding protein py subunits (but only in the presence of the a-GTP complex) (Taussig et al., 1993). The site for Py stimulatory action seems to be on the C-terminal end of the molecule. Type III Is Stimulated by CaM and Higher Calcium Concentrations Than Type I

The type III transcript resembles type I most closely in the central hydrophilic domain with lower homology in the C-terminal domain. The 129 kDa type III isoform was originally described as being olfactory epithelia-enriched and is stimulated by the a subunit of Goif, which is also an olfactory epithelium-enriched GTP-binding protein (Bakalyar and Reed, 1990), but type III adenylate cyclase mRNA was also found to be expressed in brain, spinal cord, adrenal medulla, adrenal cortex, heart atrium, aorta, lung, retina, 293 cells, and PC-12 cells (Xia et al., 1992). When compared to the type I enzyme, the type III enzyme was not found to be stimulated by Ca^"*" CaM in the absence of other effectors, but it was stimulated by Ca^"^ through CaM when the enzyme was also activated by either 5'-guanylylimidodiphosphate (GppNHp) or forskolin (Choi et al., 1992). The concentrations of

Regulation of cNMP Synthesis and Deregulation

83

free Ca^"*" for half-maximal stimulation of type I and type III adenylate cyclases were found to be 0.05 and 5.0 |LIM Ca^"^, respectively. Phorbol esters also may stimulate the activity of the type III enzyme through Ca^"^-independent C-kinase activation (Choi et al, 1993; Jacobowitz et al., 1993). The basal activity expressed in 293 cells is Ca^"^-stimulated, and the enzyme normally interacts with Gsa- It has been suggested that the type III adenylate cyclase is stimulated by Ca^"^ when the enzyme is activated by GTP binding protein-coupled receptors, and that increases in free Ca^"^ accompanying receptor activation may amplify the primary cAMP signal in cells with this adenylate cyclase isoform. The situation is complex in vivo, however, since in some cells Ca"^"^ causes only inhibition, presumably by stimulation of CaM-kinase II which inhibits the enzyme (Wayman et al., 1995). The Type IV Enzyme Resembles Type II but Is Widely Distributed

The type II and type IV enzymes are the most highly homologous members of the adenylate cyclase family (Gao and Gilman, 1991). Transcripts for the type IV enzyme encode a 110 kDa protein widely distributed in brain, heart, kidney, liver, and lung. The type IV enzyme was originally cloned from a testis library, but its mRNA is of very low abundance in that tissue as compared to others. Like the type II isoform, the type IV is not stimulated by Ca^VCaM, but Gp^ potentiates the Gsa stimulation of this enzyme like the type II enzyme. Ca^"^ has been reported to inhibit the activity of this isoform in membranes. The Type V Isoform Is Enriched in the Heart and Possesses Alternative Splice Variants

While the type V isoform was originally found in S49 lymphoma cells (Tang and Gilman, 1991), it was also cloned from canine heart (Ishikawa et al., 1992). Northern blotting indicated that the expression of this message was most abundant in heart with a lesser amount in brain. The putative 1184 amino acid product predicted from the cDNA sequence had the standard adenylate cyclase motif The adenylate cyclase activity was stimulated by NaF, GTP analogs, and forskolin, but not by CaM. The activity was inhibited in a concentration-dependent manner with either P-site active agents, such as adenosine, or by Ca^"^. A similar isoform has also been cloned from NCB-20 cells that is inhibited by submicromolar Ca^^ (Yoshimura and Cooper, 1992). Alternative splice products of type V adenylate cyclase exist that seem to be unique among the adenylate cyclase enzymes (Katsushika et al., 1993). A halfmolecule of type V adenylate cyclase diverging at the end of the first cytoplasmic loop (CI) is expressed as half of a normal heterodimer in canine heart (see Figure 1). This 596 amino acid alternative splice product is designated Type Va. Northern blots confirm the presence of an mRNA species approximately 3.5 kilo bases (kb)

84

J. KELLEY BENTLEY and JOSEPH A. BEAVO

in size in the heart. The type Va isoform is generated by use of an early termination and polyadenylation signal found in an alternatively spliced exon of the type V adenylate cyclase gene to produce the type Va exon. When type Va is coexpressed in CMT cells with an artificially generated (C2) half molecule of type V adenylate cyclase, the cyclase activity in transfected cell membranes is significantly higher than that of controls. However, when either is overexpressed alone, no significant increase in catalytic activity takes place. These results indicate that a protein containing six transmembrane spans followed by a single cytoplasmic domain can be generated in vivo, but catalytic activity is lacking unless heterodimerization can occur. Whether the native heterodimer is with the C2 terminal domain of another adenylate cyclase or with a yet undiscovered "P" half-molecule has yet to be determined. The Type VI Resembles Type V but Is More Widely Distributed

A canine cardiac library has also been used to isolate the first reported cDNA sequence for a type VI clone (Katsushika et al., 1992). This isoform was found to be more homologous to type V than to the other adenylate cyclase types. Type VI transcripts were the most abundant in heart and brain. Unlike type V, a low level of expression was also observed in other tissues examined. Type VI adenylate cyclase was stimulated by NaF, guanosine 5'-[y-thio]triphosphate, and forskolin, but not by Ca^"^/CaM, while it was inhibited by adenosine analogs. Type VI has also been obtained fi-om a rat hepatoma cell library (Krupinski et al., 1992). The type VI protein was a 1,166 amino acid protein with a predicted topography similar to that of other adenylate cyclases. The type VI mRNA was expressed in rat heart, kidney, and brain. Human embryonic kidney cells stably expressing the cDNA showed an enhanced response to isoproterenol that could be inhibited by carbachol in intact cells. Increases in intracellular Ca^"^ contributed to the inhibitory effect of carbachol, but the mechanism for this has yet to be determined. The carbachol effect could be mimicked by the Ca^^ ionophore A23187, but only a small fraction of the carbachol effect was blocked by pertussis toxin. Other Isoforms of Adenylate Cyclase With Different Regulatory Properties Remain To Be Characterized More Fully

Polymerase chain reaction (PCR) technology has revealed the presence of at least two more isoforms designated VII and VIII (Krupinski et al., 1992). Furthermore, recent data suggests unique isoforms exist in the striatum of the brain, as well as in erythroleukemia cells (Glatt and Snyder, 1993; Hellevuo et al., 1993). More work will be necessary to fiilly characterize the members of this multigene family. Differences between the adenylate cyclase isoforms allow the expression of cell-specific responses of that cell to hormonal stimulation. Since most of the adenylate cyclase isoforms have been identified only recently, physiological rea-

Regulation of cISIMP Synthesis and Deregulation

85

sons for expression of a specific cyclase are largely speculation based on the known function of the tissues and the kinetic properties of the enzymes. It is expected that many of them will be tested experimentally in the near future. For example, since type I is stimulated by Ca^'^/CaM, as well as Gsa, it might be expected to be expressed in a cell type where it is physiologically advantageous to increase its sensitivity to an "off signal from an inhibitory receptor. This is accomplished mechanistically by increasing its Gp^ sensitivity (see Figure 1) (Tang et al, 1991; Xia et al., 1993). Such a system may occur in Di dopaminergic neurons possessing a type I adenylate cyclase (Mahan et al., 1990; Monsma et al., 1990). On the other hand, type II might be better expressed in a neuronal cell type where Gpy is released in response to the activation of a GTP-binding protein (such as Go) and high activity of adenylate cyclase must be maintained (Taussig et al., 1993). Northem blot analysis suggests that a Ca^'^-inhibited adenylate cyclase exists in cardiac tissue (Yoshimura and Cooper, 1992). Expression of a Ca^"^-inhibitable species of adenylate cyclase makes sense in cardiac tissue, where negative Ca^"^ feedback of adenylate cyclase could allow coupling of oscillations in Ca^"^ and cyclic nucleotide concentrations.

THE GUANYLATE CYCLASE FAMILY Multiple isoforms of guanylate cyclase catalyze cGMP formation in different tissues, and sometimes within the same cell (Koesling et al, 1991a,b; Garbers, 1992; Koller and Goeddel, 1992; Drewett and Garbers, 1994). A list of those currently known is given in Table 2. Guanylate cyclase in mammalian cells can be broadly divided into two distinct groups, the particulate and the soluble guanylate cyclases. All mammalian guanylate cyclases have a conserved C-terminal catalytic domain. Some sequence homology exists between the catalytic domains of the adenylate and the guanylate cyclases. The guanylate cyclases also share the characteristic of only showing full activity upon oligomerization. For the six major membrane-bound or "particulate" isoforms, the exact intersubunit binding sites have not yet been fully mapped, but they are at least in part extracellular (Chinkers and Wilson, 1992). A motif also exists between the protein kinaselike and catalytic domains on the cytoplasmic portion of the molecule that could form a leucine zipper (Garbers, 1992). Between the transmembrane and catalytic domains of the particulate guanylate cyclases is a protein-kinaselike domain absolutely required for activity (Singh et al., 1988; Chinkers and Garbers, 1989; Koller et al, 1992). This kinaselike domain is found in all of the particulate guanylate cyclases cloned to this date and is probably responsible for binding ATP. ATP has been shown to stimulate the action of hormones on every particulate cyclase studied. The guanylate cyclase A (GC-A) isoform expressed in Sf-9 cells is not activated by its ligand, atrial natriuretic peptide (ANP), in the absence of ATP or an ATP analog (Chinkers et al., 1991). This effect does not require ATP hydrolysis or cyclase phosphorylation. For the retinal cyclases,

a;

Q.

2 a. N

E o

J£

<^ — u >^ U

c 03 D

uI -Si

5C

-a

«-

ni

I-

N

o c

fi)

8§

uj ( J

E 3

r—

u u u

-a

,_

J^

o^ in

_T

r o^ -OJs <^ ,_
03 03

U *^ >^ (U

_

r f O^ O^ (M O^ ' ^ O^

^

i2

3l

?8 I2|

£

o o

Q -ti O -C O-

c

O ^O -C 0.

OJ

2 B a. a. U re 7> O


n u u

Q Q Q ^ ^ .^

OJ

OJ 00

D _Q J J 0 .OJ = >: 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 i^i: (>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

<0)(/)

c k.

03

£

-Q

CD

^

86

JD

0^ 0^

^- — 03• 03

§^5

o

a; IP u ^

*^ c ^

—: OJ

'"^

o^

. a^

O 0^ O^ T— 0^

^ -:

U U

cy^

CQL

o d

3

c

DO

O u O z

CN CO.

U U

to

rn

-g

C

O u O z

Z I CO > -

(u c i> -2

O 0^ U ^

3

Q

03

<

3

E

c

0 ^';;, ^ ^ s« a, c 3 ^ -£ fo '5 : c>i :S o iS

b

o u O z

o c OJ

OJ

D

U

s,

<1>

3 ^ 3

l_

CCL

o -1 tn
O c

.

OJ

C «

E -D

— ^

®

Regulation of cNMP Synthesis and Deregulation ATP increases the Vmax of the enzyme without changing the Km for GTP (Gorczyca et al., 1995). A second role of ATP in the GC receptor signal transduction process may be to maintain the phosphorylation state of the enzyme. At least GC-A and -C also undergo similar changes in ANP affinity and coupling efficacy due to a change in phosphorylation state: dephosphorylation desensitizes GC-A to ANP (Potter and Garbers, 1992; Vaandrager et al., 1993). The phosphorylation site is distinct from the ATP-binding site as demonstrated by site directed mutagenesis (Koller et al., 1993). The phosphorylated GC-A more efficiently couples ligand binding at the cell surface to guanylate cyclase activation than the dephospho-form. It has a lower affinity for its ligand than the poorly coupled or desensitized dephospho- form (Garbers, 1989; Jewett et al., 1993). The membrane associated guanylate cyclase family contains at least six different types of isozymes that are coded for by different genes. Four are currently termed GC-A, GC-B, GC-C, and GC-D. Two other retinal specific isoforms, termed RetGC-1 and RetGC-2 have also been identified (Table 2). They are currently termed the GC-A, GC-B, and GC-C isoforms. A recent retina-specific isoform has been cloned which we will designate GC-D here for simplicity's sake. They all share the structural characteristics of being transmembrane proteins with a single hydrophobic domain crossing the plasma membrane (see Figure 2). The amino terminus of the particulate guanylate cyclase, which is extracellular and diverges substantially among the particulate guanylate cyclase isoforms, serves to bind extracellular factors that stimulate the enzyme's activity. The Widely Distributed GC-A Form Binds and Is Activated by ANP in Intact Cells or Membranes In the case of rat GC-A, the most potent ligand is rat atrial natriuretic peptide, but a second natriuretic peptide designated as brain natriuretic peptide (BNP) is almost as potent at this receptor (Chinkers et al., 1989; Suga et al., 1992). A third natriuretic peptide designated C-natriuretic peptide (CNP) stimulates GC-A very poorly. This peptide was named for its interaction with the ANP "clearance" receptor, not because of its interaction with GC-C (which it does not bind at all) (Koller and Goeddel, 1992). The extracellular domain of GC-A shares some homology with the ANP clearance receptor, which is not a guanylate cyclase (Fuller et al., 1988). The extracellular domain also contains sequences important for the oligomerization and hormone-stimulated activation of guanylate cyclase (Chinkers and Wilson, 1992). GC-A oligomerizes in a ligand-independent fashion. This oligomerization is necessary but not sufficient for the stimulation of guanylate cyclase. GC-A has a wide distribution in the brain, adrenal, and cardiovascular tissues. GC-A is desensitized to ANP by dephosphorylation (Potter and Garbers, 1992). A phosphatase that binds to GC-A has recently been reported (Chinkers, 1994).

87

88

J. KELLEY BENTLEY and JOSEPH A. BEAVO extracellular space

ANP

Figure 2. Cellular guanylate cyclases are either soluble or membrane bound. Cellular cGMP synthesis can be activated by the direct interaction of a peptide ligand with a membrane-bound guanylate cyclase molecule, here represented by the interaction of atrial natriuretic peptide (ANP) with its receptor. The ANP receptor requires dimerization for activation, and this interaction may require both intracellular and extracellular sequences. The receptor contains both a protein kinase and a catalytic domain. The catalytic domain of the membrane bound receptor has homology with domains found in both the a and p subunits of the soluble guanylate cyclase. The soluble guanylate cyclase is activated (*) by NO through the heme functional group, which interacts with the catalytic domain of the molecule at an uncharacterized site.

GC-B Is the C-ANP Receptor in the Nervous System but not the 80 kDa ANP-C Receptor

A brain gxianylate cyclase isoforai was first reported to bind and be activated by BNP (Song et al., 1988). When the human GC-B was expressed in COS-7 cells, porcine BNP was shown to be much more effective at stimulating cGMP production than human ANP (Chang et al., 1989); porcine BNP was also more effective than rat ANP on rat GC-B stimulation (Schulz et al., 1989). However, when rat peptides were used to activate the rat GC-B in aortic smooth muscle cells, the order of potency of GC-B activation was CNP > ANP > BNP (Suga et al., 1992). A similar stimulation of human GC-B, but not GC-C, was seen in a eukaryotic expression system with CNP as the ligand (Koller et al., 1991). Therefore, C-ANP is probably the endogenous ligand for the GC-B receptor.

Regulation of cNMP Synthesis and Deregulation The GC-C Form Is Enriched in the Intestine The GC-C form is activated by Escherichia coli enterotoxin (Schulz et al, 1990). This protein exists in the intestinal epithelia as at least two glycosylation variants of 140 and 160 kDa, but both are converted to a 120 kDa form by N-glycosidase F (Vaandrager et al., 1993). Recently, a peptide named guanylin has been discovered which could be the physiological agonist for GC-C (Currie et al, 1992). The molecular cloning of this peptide has predicted a much larger molecular weight precursor (Schulz etal., 1992; Wiegandetal., 1992), andin fact, a 10.3 kDa peptide has recently been isolated from circulating plasma which also stimulates guanylate cyclase (Kuhn et al., 1993). However, as with early reports of BNP stimulation of GC-B, a high concentration of peptide is required for this effect. Therefore, the physiological ligand may be something different. The GC-D Form Is Enriched in the Olfactory Epithelia Recently olfactory neurons have been shown to contain another unique form of membrane associated guanylate cyclase termed GC-D (Fulle, Vassar et al., 1995). This isozyme is highly expressed only in a small subset of olfactory neurons. Since most olfactory cilia contain large amounts of type III adenylate cyclase, the restricted distribution of this enzyme suggests that these neurons may have either a different or additional mechanism of signal transduction or perhaps even a function. It will be interesting to see what other components of the traditional signaling pathways are present in these neurons. No ligand for this guanylate cyclase has yet been identified. Diverse Forms of Particulate Guanylate Cyclase Are Found in the Retina Also contained within this family is a different isoform, the photoreceptorspecific guanylate cyclase. This enzyme is membrane associated and has been isolated in pure form (Koch, 1991). The GC-B isoform has been cloned from the retina (Duda et al., 1993). The molecular cloning of two unique retinal specific isoforms have been performed (Shyjan et al., 1992). In contrast to other membrane guanylate cyclases, the retinal guanylate cyclases have a relatively high basal level of activity when expressed in human 293 cells. cGMP production by RetGC-1 is unaffected by any of the known natriuretic peptides. Until recently, the retinal cyclase was thought to be activated by a 26 kDa protein known as recoverin, but data now suggest that the activator is a different, soluble polypeptide (Koch and Stryer, 1988; Dizhoor et al., 1991; Hurley et al., 1993) that has been termed GCAP for guanylate cyclase activating protein.

89

90

J. KELLEY BENTLEY and JOSEPH A. BEAVO

The Soluble Guanylate Cyclases Are Activated Through Their Heme Functional Group

A second family of guanylate cyclases is found in the cytosol (see Figure 2). This family contains at least four different isoforms (see Table 2). The soluble guanylate cyclase (GC-S) isoenzymes contain a tightly-bound heme cofactor and are activated by nitric oxide (NO). The domains of the cyclase which bind heme have not been characterized. This isozyme family does not have a kinaselike domain, nor does ATP seem to influence the catalytic activity of the soluble isoforms. The native enzyme consists of two subunits (GC-Sa and P) both of which contain a catalytic domain homologous to those found in the membrane-bound guanylate cyclases. In a phenomenon similar to GC-A oligomerization or CI and C2 interaction in adenylate cyclases, expression of both of the GC-Sa and -Sp subunits is required for catalytic activity (Harteneck et al., 1990; Nakane et al., 1990; Buechler et al., 1991). Mutations that prevent catalysis but still allow dimerization will act as dominant negative mutations (Yuen et al., 1994). Another subunit, GC-Sp-2, has been identified from kidney, leading to the suggestion that various heterodimeric enzyme forms may exist (Yuen et al., 1990). An additional a subunit (designated here GC-Sa2) that only has 34% homology to the rat GC-Sa 1 has been cloned from adult human brains. Human genomic Southern analysis demonstrates the presence of multiple genes or pseudogenes for GC-Sa (Giuili et al., 1992). The GC-S may have a general function in neurotransmission. Rat lung GC-Spl subunit mRNA has been found to be correlated with nuclei associated with muscarinic cholinergic neurotransmission (Matsuoka et al., 1992). Hybridization was observed in the striatum, pyramidal and granule cells of the hippocampus, olfactory bulb, inferior and superior colliculus, Purkinje cells of the cerebellum, locus coeruleus, in many pyramidal cells in the layers II-III and V of the cerebral cortex, and in the occipital cortex. Similar results have been found by immunohistochemical staining (Nakane et al., 1983; Nakane et al., 1990). Nitric oxide (NO) activates GC-S. The ligand which activates GC-S is of great physiological interest. Since the discovery that endothelial derived relaxation factor (EDRF) was in fact NO, much work has been done on the enzymatic machinery involved in NO production (Martin et al., 1985; Ignarro et al., 1987a,b; Palmer et al., 1987; Marietta, 1993; 1994). NO synthetase activity has been correlated with a variety of signal transduction phenomena from endothelium-mediated blood vessel relaxation, to immune system function, to action as a retrograde neurotransmitter. There are at least two different isoenzymes in this family. Among other cofactors, the constitutive NO synthetase requires Ca^VCaM for activity (Bredt and Snyder, 1990; Foerstermann et al., 1990; Bredt et al., 1991). A second isozyme of nitric oxide synthase is induced by various endotoxins and cytokines and regulated by glucocorticoids (Radomski et al., 1990b). This enzyme appears to bind

Regulation of cNMP Synthesis and Deregulation

91

Ca^VCaM so tightly and require such a vanishingly small concentration offreeCa^"^ that preliminary studies suggested it was CaM-independent, but it has a unique consensus CaM binding site and CaM copurifies with the enzyme (Cho et al., 1992; Lowenstein et al., 1992). Is carbon monoxide (CO) a physiological stimulator of GC-S? It has recently been suggested that carbon monoxide, like nitric oxide, can function as a neurotransmitter or second messenger in the brain by activating soluble guanylate cyclase (Verma et al., 1993). Since carbon monoxide is formed by the action of the enzyme heme oxygenase, discrete neuronal localization of messenger RNA for the constitutive form of heme oxygenase throughout the brain has been demonstrated by in situ hybridization in brain slices, and since this localization is essentially the same as that for soluble guanylate cyclase messenger RNA, it has been suggested CO regulates cGMP in certain cells. The only biochemical evidence for this at present is that in primary cultures of olfactory neurons, zinc protoporphyrin-9, a potent inhibitor of heme oxygenase, depletes cGMP. The synthesis of cellular cGMP is regulated by either circulating peptide hormones such as AN? acting on the membrane-bound particulate receptors or in a more paracrine fashion by the release of an agent like NO which tends to act more locally (Figure 2). However, both kinds of mechanisms may be at work in the central nervous system. Here, particulate cyclase acts to bind peptide neurotransmitter ligands. The postsynaptically synthesized NO acts as a retrograde neurotransmitter feeding back to the presynaptic terminal; Ca^"^ metabolism therefore influences both the breakdown and synthesis of neuronal cGMP (Figure 4).

THE CYCLIC NUCLEOTIDE PHOSPHODIESTERASE FAMILY The cyclic nucleotide phosphodiesterases (PDEs) are a superfamily of enzymes made up of at least seven different families (Beavo, 1995; Beavo and Houslay, 1990, see Figure 3). Each family is encoded by a different gene or series of very closely related genes. Alternative splicing of many of the mRNAs and posttranslational modification of the protein products create an additional element of diversity. A list of these gene families and their known alternatively processed forms is shown in Table 3. It is likely that additional isozymes will be identified in the coming years. When taken in conjunction with the large number of cyclases discussed in the previous sections, the differential expression of varying combinations of cyclase and PDE provides a cell with an enormous number of possible variations for regulation of cAMP and cGMP steady state levels. Conserved Structural Elements

In mammals, all PDE isoenzymes contain at least one domain involved with their regulation and a more conserved domain located towards the C-terminal end of the molecule that functions catalytically. Alignments within these conserved

O O CL

5r

CD

cN m

2 2;

fo <\3 22 DO 0 0 I ^

(/)

a>

c

"D

tn

Q LO

5

U

6

UJ

00 O 00 ^ r- i n

^2

D f« en c c c a> JO O rz

o o

c c c t o IT)

(U

"D C 05

^^ D E

-C

o o E

(/i

-^^

( J 00 00

c c ^m 'm 'ro

:^ OB

<0Q

0^ ro

UU L U

Q- a . 0 .

QQ Q

a

c ol sf) 0^ c 03

(J <

U

^ 9 ^ ^ c.

to

>» "?

Jc —I

Vrt

15 -a

o

O N

c 03 Q3 n3

o o

- D TD

ro 0^ (J^ r—

Q)

15^ QJ

"K

< 5

00 0^

0 0^ 2 ^

"3 <j^ 0^ T-

*" 00 ^ n3 0^ - = , - 0 0^ > *" '"^ oj

1- c

(U

O

u 2f

rt

03

08

c 2

•-

0)

1

^

^

C

a> t75 a» u o -c

•is S c c a;

O <-M

sD vD O

IX

< oou < LU LU LU LU

QQQQ

Q. 0 . 0 . 0 -

O

UJ

Q.

o

« ~

> . OQ

C

2.

OJ

c

E
o

^

^^ <- o

I

- ^ 0) '^ "D re ^
S E re .c

to _ tft VE: ~ O t 00

^

^^

^ ^ o 3 cy j2 "TO g"

&E 2 ^

o >^ 3 d tr5 ro Si a;

^?:^ §

•• • £

S a; > 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 .2 ^

?5 ^ o; u

nj u


C -Q

on >^

Q.

if) to

^

nj

O ^ _ : .^ T^ 7 5 "3 ^ •=

C

O

_C _^ .— _C

CU _Q

a;

o o

Q. Q.

«/i

(/)

-D to in

O

O 0-

Q-

E

o o

Q) -O

^1 o o

Q Q. O O.5 . E . E Q. a ^•^•^

Q Q Q n

u -^ ^

Q

ro ro K Q 0^ 00 CO m a^ a^ C7^ I 0^

< 00 ( J Q < - ^ Tj- Tf r r uo

to

Qu 0 - 0 - O- CL

u o;

Q. _•

T— r s r - CM

s£)

fO

C C u foc fO 00 ^^ .2 o E O 00 ^ .9- c c f5 ^ "m 'n 03 •C c

h- c>5 5 H

o

2

OJ

<

" ^

a^ I r-

Q

T -

-^ in ro

Q.

to

QD

LU LU L U L U LU

r o fN »— (N
"3 ro O

< 00 ro ro QL

Q Q Q QQ

LU LU

r- ^

^

Q

Gu

I

DO <

o> Q - o ; c ;^ c

_'

a> 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<^ Seciuence Homology

Type IVAI

IV A2 IVD IVB

IVc lA IB

-hHCP-1

VII

-h cGM -r cGI-2

IIlA IIIB

b photo a b photo P b photo a'

VIA VIB Vic

b cG-spec

V

b cGS

II

100

I

Figure 3. A cyclic nucleotide isoenzyme family tree is generated by progressive sequence alignment of the open reading frame in the conserved catalytic domain. Note even within the highly conserved catalytic domain, no more than 50% identity exists between family groupings. For each isoenzyme illustrated, the first initial stands for the organism the sequence from which was derived, (r) indicates rat, (b) indicates bovine, and (h) indicates human. References describing the isoforms shown in this figure are listed in Table 3. The plot was generated by the PILEUP program in the Genetic Computers Group sequence analysis program for the VAX computer. This program is using the progressive alignment method of Feng and Doolittle (1987, J. Mol. Evolution 35, 351-360). The clustering strategy represented by the dendrogram is an unweighted pair-group using arithmetic averages (Sneath, P.H.A. & Sokal, R.R. (1973). Numerical Taxonomy, pp. 230-234, W.H. Freeman and Company, San Francisco, California). Each pairwise alignment in PILEUP uses the method of Needleman and Wunsch (1970, J. Mol. Biology 48, 443-453).

regions show greater than 60% amino acid identity v^ithin individual isozyme families, but less than 30% identity betw^een families. Several of the PDE families also have a second domain that is highly conserved and may contain a second cyclic nucleotide binding site selective for cGMP (Stroop et al., 1989; Charbonneau et al., 1990; Stroop and Beavo, 1991,1992). Most of the PDEs have unique C-termini of unknov^n function (Conti et al., 1991). Sequence alignments identify several amino acids that are very highly conserved among ail PDEs (Charbonneau, 1990; Bentley and Beavo, 1992). The sequence motif HD(L,I,V,M,F,Y)HxHx(A,G)xxNx(L,I,V,M,F,Y) v^ill uniquely define a cyclic nucleotide PDE. An x indicates that any amino acid may occupy this position and parentheses indicate that any of the listed amino acids may occupy the position. Other motifs v^ithin the conserved domain such as EF(F,W)xxQGD(R,K,L)E also v^ill uniquely identify PDEs. From the 21,000 sequences present in the SwissProt Data Base (Release 19), only PDEs

94

J. KELLEY BENTLEY and JOSEPH A. BEAVO

are recognized by these motifs. The importance of the conserved and catalytic domains has been demonstrated by deletion mutagenesis, and the essential role of several conserved residues in catalytic function has been demonstrated by site-directed mutagenesis (Jin et al., 1992). The degree of homology among isoforms is represented in the PDE family tree shown in Figure 3. The catalytic domains of all the isoforms bind cyclic nucleotides in a unique manner as compared to other cyclic nucleotide binding sites. In addition, there are several binding interactions that serve to distinguish the cAMP binding motifs from the cGMP binding motifs (Butt et al., 1993). The types IV and V PDEs appear to interact with the anti conformer, while the types I and II PDEs interact with the syn conformer (Beltman et al., 1993; Butt et al., 1993). Only an interaction at Nl and a lack of binding about the ribose moiety appears to be conserved among all phosphodiesterases and both substrates. All phosphodiesterases that hydrolyze cGMP, with the exception of the type II PDE, utilize the 6-oxo group of cGMP (Beltman et al, 1993). In contrast to cGMP-dependent protein kinase, the 2-amino group on cGMP is not directly required for binding to any PDE except possibly the type III PDE. Large differences exist in the binding requirements for cAMP interaction with PDEs as compared to cAMP-dependent kinase and the catabolite activator protein, or for the binding of cGMP to cGMP-dependent protein kinase or the cGMP-gated cation channel. Type I PDEs (Ca^VCaM-Stimulated)

One well-characterized PDE isozyme family is the Ca^VCaM-stimulated PDE (CaM-PDE) family. All CaM-PDEs are activated by CaM in the presence of calcium. At least eight different members of this family have been described (see Table 3). Additionally, recent reviews are available concerning their structure and function (Wang et al., 1990; Kincaid et al., 1992; Wu et al., 1992). The CaM-PDEs are thought to modulate cyclic nucleotide-dependent processes by signal transduction pathways that increase cytoplasmic calcium concentrations. In many cell types, agonists that increase intracellular Ca^"*^ stimulate cyclic AMP degradation (Miot et al., 1983; Dumont et al., 1984; Buxton and Brunton, 1985; Tanner et al, 1986). All the CaM-PDEs that have been sequenced conform to the general motif shown in Figure 4. The brain is a rich source of CaM-PDEs. Three CaM-PDEs have been purified and found to be brain-enriched (Hansen and Beavo, 1982; Sharma et al., 1984; Shenolikar et al., 1985). Bovine brain CaM-PDEs of 63 and 61 kDa have both distinct and common regions of sequence (Novack et al., 1991; Bentley et al, 1992). High levels of CaM-PDE are found in neurons of the adult rat neocortex, in hippocampus CAl and CA3 regions, and cerebellar cortex Purkinje cells in cell layers V—VI, especially in postsynaptic densities (Kincaid et al., 1987; Ludvig et al., 1991). In the cerebellum, presynaptic regulation of post-synaptic CaM-PDE

Regulation

of cNMP Synthesis and Deregulation

95

* indicates consensus Cal\/I-I
C ^ ^ ^ ^ CaM-binding domain V/////////A Intermediate domain • • • Catalytic domain (conserved) ti:'k':f:i!;i'i:i!;:i4'!&..i C-terminal domain Figure 4. The structure of the CaM-PDEs. CaM-PDEs occur as dinners. These enzymes contain an N-terminal CaM-binding domain, an intermediate domain that participates in the regulation of catalytic activity, and a C-terminal catalytic domain. The conserved catalytic domain residues make up a large part of this region. The 59, 6 1 , and 63 kDa CaM-PDEs are all desensitized to CaM by phosphorylation. Shov^n are consensus sequence sites for protein kinase A (PkA) or CaM-kinase il phosphorylation (*) (Kennelly and Krebs, 1991).

expression occurs (Balaban et al., 1989). On the other hand, further data from the same laboratory suggests that the increased expression of CaM-binding enzymes during rat striatal development is not directly dependent on synaptic inputfromthe nigrostriatal dopamine system, as assessed by CaM-overlay techniques on blots from lesioned animals (Polli et al, 1991). The 61 kDa CaM-PDE is widely distributed in the brain. The mRNA encoding the 61 kDa CaM-dependent PDE is most concentrated in the brain. It appears to be predominantly expressed in neurons (Sonnenburg et al, 1993). Nearly equal concentrations of mRNA are present in basal ganglia, cerebellum, and spinal cord. The 61 kDa CaM-PDE mRNA has the highest level of expression in the cerebral cortex and in the pyramidal cells of the hippocampus (Yan et al., 1993). A moderate hybridization signal was detected in the medial habenula and amygdaloid nuclear complex. In addition, small subsets of neurons in several other regions showed specific hybridization signals. The PDE mRNAs appear to be localized exclusively in neuronal cell bodies. The 63 kDa CaM-PDE is particularly enriched in the striatum. The cDNA for a brain-specific 63 kDa CaM-PDE has been isolated and sequenced (Bentley et al., 1992; Polli and Kincaid, 1992; Repaske et al., 1992). The mRNA for this protein

96

J. KELLEY BENTLEY and JOSEPH A. BEAVO

is found largely in basal ganglia structures in rat and bovine brain and in mouse striatum (Yan et al., 1993), but it is also found to a lesser extent in other areas of the brain and in the adrenal medulla. The mRNA for this CaM-PDE is heavily concentrated in the caudate-putamen and the olfactory bulb of the mouse (Yan et al., 1993). Both the 61 and 63 kDa PDE mRNAs appear to be localized in neuronal cell bodies. This particular isoform is also induced in lymphocytes by mitogens (Hurwitz et al., 1990). The enzyme has a higher affinity and V^ax for cGMP than cAMP (Sharma and Wang, 1985; Hansen and Beavo, 1986). The 61 and 63 kDa brain CaM-PDEs differ slightly in their kinetics of hydrolysis and in their regulation by phosphorylation. cGMP is a better substrate than cAMP for both the 61 and 63 kDa enzymes. The Michaelis constants (K^s) for cAMP are about 100 |LiM and 10 |LIM, respectively, while the K^s for cGMP are 3 JLIM and 1 |iM, respectively (Hansen and Beavo, 1986; Wang et al., 1990). In addition to their activation by Ca^VCaM, CaM-PDEs are regulated by phosphorylation. In in vitro studies, A-kinase will phosphorylate the brain 61 kDa and heart 59 kDa isozymes. A stoichiometry of about 2 mol phosphate per mol subunit is reached, and the phosphorylation results in a decrease in the affinity for Ca^VCaM (Sharma and Wang, 1985; Sharma, 1991). Therefore, an agonist or agonists increasing cellular cAMP above a threshold level might cause a supra additive (or "feed forward") increase in cAMP at a given constant concentration of cellular calcium. In the brain this would be expected to have major effects on both short and longer term changes in synaptic strength. The 63 kDa CaM-PDE is phosphorylated and desensitized to CaM by CaMstimulated protein kinase II (Sharma and Wang, 1986a; Hashimoto et al., 1989). A stoichiometry of about 2 mol phosphate per mol subunit is reached, and the phosphorylation results in a decrease in the affinity for Ca^VCaM. At the CaM concentrations thought to be present in most cells, a higher concentration of Ca^^ is required to activate the phosphoenzyme. The phosphorylation of the CaMstimulated PDEs is partially blocked when the PDE is previously bound to the Ca^"^/CaM complex. Therefore, in cells containing CaM kinase II, agonists that increase cellular Ca^"^ above a threshold level will be expected to attenuate their own effects on the CaM PDE; therefore, the CaM-PDE and CaM-kinase represent pivotal points for the interaction of Ca^"^ and cyclic nucleotides. It is likely that this interaction fine tunes neurotransmitter release and could also modulate some of the effects of CaM kinase II on synaptic plasticity (Sharma and Wang, 1986c). The 59 kDa heart and 61 kDa brain enzymes differ in structure only in their N-terminal and CaM-binding domain. These isozymes are produced by tissue specific alternative splicing of the same gene (Charbonneau et al., 1991; Novack etal., 1991; Sonnenburgetal., 1992,1993). Amino acid and cDNA sequence from a 63 kDa isozyme isolatedfi-ombrain show that it is the product of a different gene than the 59/61 kDa CaM-PDE gene (Novack et al., 1991; Bentley et al., 1992). The 59 kDa isoform is enriched in cardiac tissues and has about a 10-fold lower Ka for

Regulation of cNMP Synthesis and Deregulation

97

CaM than the 61 kDa enzyme (Hansen and Beavo, 1986). The 59 kDa CaM-PDE seems to be the isoform responsible for the a i-adrenergic receptor-mediated activation of cyclic AMP degradation in cardiac myocytes (Buxton and Brunton, 1985). An isoenzyme similar to the 59 kDa isoform has been described in lung; the principal difference between the two is that the lung enzyme purifies with calmodulin as a subunit, even in the presence of the Ca^"*" chelator EGTA (Sharma and Wang, 1986b). No sequence data is currently available on the 58 kDa isoform; it is not clear if it is in fact different from the 59 kDa isoenzyme. Distinct higher molecular weight CaM-PDEs have been described in brain and testis. Another CaM-PDE of about 75 kDa also has been described in bovine brain (Shenolikar et al, 1985). This enzyme shows a much greater specificity for cGMP as compared to cAMP having K^ values of 5 |iM and 15 jiM for cGMP and cAMP, respectively. Addition of calcium and CaM reduces the apparent K^ for cGMP to 2-3 jiM and increases the V^ax by 10-fold, while cAMP hydrolysis shows a similar increase in V^ax with an apparent doubling of K^. No sequence information on the 75 kDa enzyme is available. A 68 kDa CaM-PDE associated with the male reproductive system also has been reported. This isoform seems to share some immunologic determinants with the brain isoforms (Rossi et al., 1988). However, it has several unique characteristics: the purified enzyme hydrolyzes both cAMP and cGMP with high and low affinity components (K^^s of 2 and 20 |iM) observed either in the presence or absence of CaM. Each of the substrates acts as a noncompetitive inhibitor of the other, suggesting either an impure preparation or the presence of two distinct catalytic sites on the enzyme. No sequence information has been reported for this isoenzyme. A high affinity CaM-dependent PDE activity has been identified in olfactory cilia (Borisy et al., 1992). This enzyme has recently been cloned (Yan et al., 1995) and shown to be expressed at high levels in the cilia where it is postulated to play an important role in modulating olfactory signal transduction. CaM-PDEs are a regulatory point for both Ca^"^ and cyclic nucleotide metabolism in the neuron. Muscarinic agonists are known to mobilize CaM and activate CaM-sensitive PDE activity in neuroblastoma cells (Tanner et al., 1986). Agonist occupation of muscarinic cholinergic receptors of 13 21N1 human astrocytoma cells results in an activation of PDE and a resultant 50-75% attenuation of isoproterenol-stimulated cAMP accumulation (Tanner et al, 1986). The similar relative capacities of PDE inhibitors for blocking both the muscarinic receptor-mediated attenuation of cAMP accumulation and CaM-PDE activity in vitro suggest that it is this form of enzyme that is regulated by muscarinic receptor stimulation. In isolated brain synaptosome preparations, CaM-dependent PDE activity strongly modulates cGMP responses elicited by nitric oxide (Mayer et al., 1992). cGMP formation is inhibited at Ca^"^ concentrations required for the activation of NO synthase (200-300 nM), indicating a down-regulation of the signal in NO-producing cells. In cytosolic fractions, the breakdown of cyclic GMP, but

J. KELLEY BENTLEY and JOSEPH A. BEAVO

98 Ca2+

Figure 5. Possible mechanism for CaM regulation of both the synthesis and degradation of cGMP. In response to glutamate release presynaptically, Ca^"" is gated into the postsynaptic neuron. Upon interaction with CaM-NO synthetase, NO is released and diffuses across the synaptic cleft. In the cytosol of the presynaptic neuron, NO activates the ap dimer of soluble guanylate cyclase through the heme functional group. cGMP produced at this point can interact with any of its effector proteins; depicted here is an interaction with a hypothetical cGMP-gated ion channel. The Ca^"" gated into the cell functions to activate the presynaptic CaM-PDE. This in turn acts to lower cGMP, closing the hypothetical cGMP-gated cation channel. A CaM PDE monomer is also depicted here being inactivated by CaM-kinase Il-mediated phosphorylation.

not that of cyclic AMP, was highly stimulated by Ca^"^. The effects of Ca^"" on cyclic GMP hydrolysis and on apparent guanylate cyclase activities were abolished almost completely in the presence of the calmodulin antagonist calmidazolium, whose effect was attenuated by added calmodulin. The PDE arm of this pathway is likely to be limited by the action of CaM-kinase II on a 63 kDa CaM-PDE. Expression of CaM-PDEs in a cell allows one mechanism for the integration of Ca^"^ and cyclic nucleotide-dependent regulatory pathways, and their feedback desensitization by phosphorylation allows a further level of control (see Figure 5). In the model shown, Ca^"^ is gated into the postsynaptic neuron and activates CaM-NO synthetase in response to presynaptic glutamate release. NO diffuses

Regulation of cNMP Synthesis and Deregulation

99

across the synaptic cleft to activate the ap dimer of soluble guanylate cyclase through the heme functional group in the cytosol of the presynaptic neuron. The cGMP produced at this point feeds back to increase cellular Ca^"^. One mechanism illustrated in Figure 5 is an interaction with a hypothetical cGMP-gated ion channel. The Ca^"^ gated into the cell functions to activate the presynaptic CaM-PDE which lowers cGMP and closes the cGMP-gated cation channel. A further interaction is depicted wherein a CaM-PDE is inactivated by CaM-kinase Il-mediated phosphorylation. Type II PDEs (cGMP-Stimulated)

The cGMP-stimulated (cGS) enzymes have a regulatory N-terminal cGMPbinding domain as well as a catalytic domain. They have been extensively reviewed (Manganiello et al., 1990b; Sonnenburg and Beavo, 1992). The cGS-PDEs hydrolyze both cAMP and cGMP, but low concentrations of cGMP enhance cAMP hydrolytic activity (Beavo et al, 1971; Martins et al., 1982). The enzymes all appear to possess a noncatalytic cGMP-binding domain as well as a more C-terminal catalytic domain (Stroop etal., 1989; Stroop and Beavo, 1991,1992). About 1 mole

cGB-PDEs

3', 5'-cNMP

5'-NMP cGMP-binding domain catalytic domain

Figure 6. Cyclic GMP-binding PDEs (types II, V, and VI) share an overall structural motif. All cGMP-binding PDEs have two Internally homologous N-terminal domains involved in noncatalytically binding cGMP. All form active dimers; types II and V form homodimers and type VI forms an active heterodimer.

100

J. KELLEY BENTLEY and JOSEPH A. BEAVO

of cGMP is bound per mole of dimerized enzyme with half-maximal binding at 0.2 laM cGMP (Emeux et al., 1981; Martins et al., 1982). A second class of cGMP binding sites has been reported on the native enzyme (Stroop et al., 1989). Common elements among the cGMP-binding sites. The type II cGMP-binding domain contains sequence elements conserved within all cGMP-binding PDEs (Figure 6) (Charbonneau et al., 1990; Tanaka et al., 1991). The cGMP binding domains of all cGMP binding PDEs (i.e., types II, V, and VI) contain the long sequence motif, L(C,S)(F,L,M)PI(K,V)NXX(E,Q)(E,D)(I,V)(I,V)GVAX(A,F,L)(I,V,Y)N(K,R) (I,K)XG. This motif occurs in a second conserved segment of approximately 340 residues just N-terminal to the catalytic domain that contains an internally homologous repeat (Charbonneau et al., 1990). It is not yet clear whether both parts of the internal repeat are required for cGMP binding. Both soluble and particulate isoforms of cGS-PDE have similar kinetics. The cGS-PDEs exists in two distinct isoforms, a 105 kDa particulate brain-enriched isoform (Whalin et al., 1988; Murashima et al., 1990) and a soluble 102 kDa cardiac, liver, platelet, and adrenal isoenzyme (Martins et al., 1982; Yamamoto et al., 1983; Grant et al., 1990). Both isoenzymes occur as homodimers and have very similar kinetics and peptide maps (Manganiello et al., 1990b; Murashima et al., 1990); the data suggest that they diverge only at their N-termini. Both particulate and soluble isoforms have been identified in liver, but an additional much smaller 67 kDa liver isoform also has been described (Pyne et al., 1986). The particulate and soluble cGS-PDE isoenzymes differ only at their N-terminal residues. Both peptide and cDNA data show that the amino terminal domains differ between the soluble and particulate enzymes (Trong et al., 1990; Sonnenburg et al., 1991). The large regions of ORF identity between the adrenal and brain (unpublished observations) cDNAs suggest that these isozymes are products of an alternately spliced gene. The N-terminal residues present in the particulate brain isoform allow it to bind to the membrane but do not contain an obvious hydrophobic transmembrane spanning sequence. The main regulation of the type II PDEs is the stimulation of its activity by binding of cGMP. In the unstimulated enzyme, the hydrolysis of both cAMP and cGMP is positively cooperative (Sonnenburg and Beavo, 1992). Typical Hill coefficients for cAMP and cGMP are 1.6-1.9 and 1.3-1.6. Apparent K^s are somewhat lower for cGMP than for cAMP (10-16 |iM and 28-40 laM, respectively). On the other hand, apparent Vmax values ranging between 120-200 |imol/min/mg protein for each. The stimulatory effect of cGMP on type II PDE (Kact = 0.35-0.5 inM) results in a less cooperative kinetic behavior lowering the apparent K^ for cAMP with little or no effect on V^ax (Yamamoto et al., 1983; Whalin et al., 1988). Therefore, the degree of activation of cAMP hydrolysis by cGMP varies depending on the substrate concentration used in the assay. At sub-saturating cGMP, high concentrations of cAMP can stimulate [^H]-cGMP PDE activity, but using more physi-

Regulation ofcNMP Synthesis and Deregulation

101

extracellular space ACTH

ANF

cGS-PDE

aldosterone production

cytosol

Figure 7, The cGS-PDE regulates cAMP metabolism and aldosterone production in the adrenal. In adrenal cortical cells, adrenocorticotropic hormone (ACTH) normally binds its receptor to stimulate adenylate cyclase. The cAMP produced activates A-kinase to activate aldosterone production by the phosphorylation of its synthetic enzymes at multiple points in its anabolic pathv^ay. Atrial natriuretic peptide (ANP) antagonizes this process by binding its receptor, GC-A, leading to the activation of cGMP synthesis and an activation of cGS-PDE. The type II PDE has a lower K^ for cAMP than cGMP; therefore, it preferentially hydrolyzes cAMP. As the intracellular cAMP concentration falls, A-kinase is inactivated, and aldosterone synthesis is inhibited as its enzymes are dephosphorylated (MacFarland et al., 1991).

ological levels of cGMP, cAMP inhibits [^H]-cGMP hydrolysis. At the concentrations of cGMP and cAMP in the cell, cAMP hydrolytic activity is enhanced 3-5-fold by cGMP. The cGS-PDE functions in the regulation of aldosterone production and epinephrine release in the adrenal (Figure 7). Using Northern analysis, RNase protection, and specific monoclonal antibodies, both the transcripts and translation products for the cGS-PDE gene can be demonstrated to be particularly rich in the adrenal cortex glomerulosa cells (MacFarland et al., 1991; Sonnenburg et al, 1991). The stimulation of the cGS-PDE occurs normally in the response of the adrenal cortex to ANP. In the adrenal cortical cells, ANP binds its receptor, GC-A, to stimulate guanylate cyclase activity. This stimulates the soluble cGS-PDE to lower cellular cAMP, thus antagonizing the production of aldosterone induced by adrenocorticotropic hormone (ACTH). A similar role has been proposed for cGS-PDE in adrenal chromaffin cells where increasing cGMP leads to a decrease in secretion (Whalin et al., 1990, 1991). Treatment of PC 12 cells with ANP or sodium nitroprusside causes increased intracellular cGMP, attenuation of adenosine-stimulated cAMP accumulation, and increased rates of cAMP decay after removal of the

102

J. KELLEY BENTLEY and JOSEPH A. BEAVO

adenosine stimulus. Treatment of PC 12 cells with inhibitors of cGS-PDE causes an increase in cAMP, potentiation of adenosine-stimulated cAMP, and retardation of the rate of cAMP decay after removal of the adenosine stimulus. These results suggest that in PC 12 cells inhibition of type IIPDE or an increase in intracellular cGMP to activate the cGS-PDE can modulate cAMP metabolism. Type 111 PDEs (c-GMP-lnhibited)

The type III PDEs all share the characteristics of: (a) a relatively low K^ for cAMP; (b) being competitively inhibited by cGMP at a relatively low concentration; (c) inhibition by cardiotonic vasodilator drugs like enoximone (Dage and Okerholm, 1990); and (d) phosphorylation with activation in response to cellular agonists. The cGMP-inhibited (cGI-PDE) isoenzymes have a distinct set of kinetic characteristics: high affinity for both cAMP and cGMP (apparent K^s < 1 jiM), a Vmax for cAMP about lOx greater than for cGMP, and potent (apparently competitive) inhibition of cAMP hydrolysis by cGMP (Manganiello et al., 1990a). As with other PDEs, the catalytic domain is located on the C-terminal end of the molecule, but evidence suggests the cGMP interaction site is at the catalytic site (Reifsnyder et al., 1985; Beltman et al., 1993; Butt et al., 1993). The type III PDEs are produced by at least two distinct genes. The cGI-PDE exists in the heart as both a soluble and particulate enzyme in the sarcoplasmic reticulum where it serves as the site of action of several cardiotonic drugs (Harrison et al., 1986a,b; Weishaar et al., 1989; Smith et al., 1993). Molecular cloning of the human myocardial cGI-PDE has been reported (Meacci et al., 1992). The catalytic domain of the cloned enzyme retains a PDE inhibitor sensitivity profile similar to that of the native enzyme when expressed in bacteria. This isoform has some sequence differences with the translated adipocyte enzyme cDNA, and mRNA hybridization patterns differ between heart and fat cells (Taira et al., 1992). The activity of one of the fat cell isoenzymes is increased by insulin. The insulin-sensitive cGI-PDE enzyme has been purified from fat and liver tissue (Degerman et al., 1987; Boyes and Loten, 1988a,b). Although soluble forms of the cGI-PDE have been described, the insulin-sensitive cGI-PDE is found only in the membrane fraction of adipocytes and hepatocytes. It may be that covalent modification of the adipocyte cGI-PDE is partially responsible for its soluble or membrane-bound distribution (Taira et al., 1992). The cGI-PDE is regulated by phosphorylation. In intact cells, phosphorylation and activation of the fat cardiac or platelet enzyme can be induced by agents as diverse as isoproterenol, prostaglandins, and insulin (see Figure 8) (Macphee et al., 1986,1988a,b; Shibata and Kono, 1990a,b; L'opez et al., 1992). In addition to the metabolic effects in somatic cells, this may also be the basis of the involvement of the type III PDE in the growth-promoting actions of the insulin receptor and p21 ras gene product in Xenopus oocytes (Sadler, 1991).

103

Regulation ofcNMP Synthesis and Deregulation

Recent studies indicate that both the catalytic subunit of A-kinase and an insulin-stimulated unidentified serine kinase act synergistically on this enzyme (Gettys et al., 1988; Macphee et al., 1988b; Smith et al, 1991). The transient increase in cGI-PDE phosphorylation and the enzyme's stimulation also correlate temporally with the insulin-induced reduction in isoproterenol activation of A-kinase and insulin-induced inhibition of lipolysis. Just such a mixed phosphorylation pattern in response to insulin and agents which activate A-kinase is also found in platelets (Macphee et al., 1986, 1988a,b; Hendra and Betteridge, 1989; L'opez et al., 1992). In platelets, phosphorylation occurs relatively quickly in the cell, after the cAMP concentration has reached its peak (Smith and Manganiello, 1989). The phosphorylation of the cGI-PDE by A-kinase is thought to be a negative feedback mechanism that rapidly returns the cell's cAMP concentration back to the nonstimulated levels (Gettys et al., 1987). Additionally, agents which elevate cGMP also inhibit platelet aggregation (Radomski et al., 1990a). In part, this is due to an activation of cGMP-dependent protein kinase (G-kinase), leading to the phosphorylation of the Ca^^-ATPase and cytoskeletal protein phosphorylation (Lincoln and Comwell, 1993a). However, inhibition of the type III PDE by cGMP is a major pathway for cGMP action in platelets (Maurice et al., 1991) (Figure 8). extracellular space

3iy|p

cAJdiP

f

inhibition of aggregation

X.

A-kinase cGMP

G-kinase -^^ cytosol

^

o.Tp^

Figure 8. Phosphorylation activates the cGI-PDE in the platelet. Prostaglandins (here PGEl) bind to their receptors to produce activation of adenylate cyclase. The cAMP produced activates A-klnase, which through phosphorylation-dependent mechanisms leads to an inhibition of platelet aggregation. This process can also produce its own negative feedback due to the phosphorylation and activation of the type III PDE. Another mechanism of cGI-inhibition may occur through the action of NO produced by vasodilators such as sodium nitroprusside (SNP). The cGMP produced would directly inhibit the type III PDE by competition for the catalytic domain.

104

J. KELLEY BENTLEY and JOSEPH A. BEAVO

Type IV PDEs (cAMP-Specific)

The type IV PDEs are a multigene family with a high affinity for cAMP. These PDEs are the products of at least four distinct genes (see Table 3) and have already been reviewed (Conti et al., 1991, 1992, 1995). The type IV PDEs cloned from mammalian cells have been found in brain, heart, kidney, spermatozoa, and Sertoli cells. They are all homologues of the Drosophila learning and fertility mutant dunce gene product, they all hydrolyze cAMP preferentially to cGMP with K^s ranging from 0.1 to 3 ILIM, they all are sensitive to inhibitors such as rolipram, and are all insensitive to cGMP or cardiotonic drugs (Colicelli et al., 1989; Davis et al., 1989; Swinnen et al., 1989a,b; Livi et al., 1990). Due to their ability to rescue yeast mutations in the RAS and PDE locus, yeast expression systems have been used to isolate type IV isoforms (Colicelli et al., 1989, 1991; Michaeli, 1993). The synthesis of the type IV PDEs is induced by agents that increase cellular cAMP. Sertoli cell-derived rat PDE3 type IV transcripts have been characterized as hormone-sensitive due to their 100-fold stimulation by follicle-stimulating hormone (FSH) (Swinnen et al., 1989b). This stimulation is mimicked by agents that act upon cAMP-dependent protein kinase. The hormone-dependent activation of a soluble low K^ PDE has been described for fibroblasts, lymphoma cells, and glioma cells exposed to epinephrine or prostaglandins (Conti et al., 1991). It may also be observed in vivo when FSH is administered to animals. The synthesis of type IV PDE is typically activated at a time when the cellular concentration of cAMP is actually decreasing, and therefore it may serve as a mechanism of desensitization to cellular agonists which would normally be expected to increase cellular c AMP concentrations (Swinnen et al., 1991). The expression of these PDEs may prove to have a crucial role in developmental processes such as spermatogenesis (Welch et al., 1992). An overexpression of type IV PDEs has been linked to disease states in animals. An example of a genetic lesion in c AMP metabolism may be found in mice with nephrogenic diabetes insipidus. In these mice, the entire medullary collecting duct is unresponsive to its normal agonist, arginine vasopressin (AVP) (Valtin et al., 1990). AVP normally activates adenylate cyclase, causing an increase in cAMP and water permeability through the epithelium eventually leading to an antidiuretic effect. Data exist to suggest this is primarily due to the higher activity of the rolipram-sensitive type IV PDE in these animals (Homma et al., 1991; Takeda et al., 1991; Yamaki et al., 1992). The primary mechanism of regulation of the cAMP-specific, rolipram-sensitive PDEs may be at the transcriptional level in cells that express them. However, several reports now have demonstrated regulation by direct phosphorylation (Pyne et al., 1989; Houslay et al., 1992; Conti etal., 1995).

Regulation of cNMP Synthesis and Deregulation

105

Type V PDEs (cGMP-Specific) The type V PDE is a multidomain protein with separate specific cyclic GMP binding and hydro lytic activities. cGMP has a binding K^ of 0.2 |LIM and a K^ of 5 fiM for hydrolysis by this 93 kDa homodimeric enzyme (Thomas et al, 1990a). The enzyme has been found in platelets, spleen, lung, and smooth muscle (Coquil et al., 1980, 1983; Davis and Kuo, 1982; Hamet et al., 1984; Coquil, 1985). The enzyme resembles the type VI PDEs in its sensitivity to specific inhibitors such as zaprinast and dipyridamole (Gillespie, 1990; Thomas et al., 1990a). A cDNA encoding the bovine lung type V PDE has recently been isolated. This cDNA encodes an 875 amino acid, 99 kDa polypeptide with catalytic properties identical to the enzyme isolated from lung (McAllister et al., 1993). As with type II and VI PDEs, it contains a domain from amino acids 142 through 526 with two internal repeats responsible for binding cGMP. The physiological significance of binding cGMP is probably related to its phosphorylation, and the type V PDE is typically found in tissues where guanylate cyclase and cGMP are high (Francis et al., 1990). An interaction between the catalytic and cGMP-binding domains is evident in that an interaction of the PDE inhibitor 3-isobutyl-1-methyl-xanthine with the catalytic site increases the binding of cyclic GMP to the binding domain (Thomas et al., 1992). To date, there has only been one gene product isolated for this isoform, which most closely resembles the type VI PDEs in structure (see Figure 6). Although there is only a single 6.9 kb mRNA species detectable by Northern analysis, PCR data suggests alternative splice products exist (Sonnenburg, personal communication). Homology in the catalytic domain is 48% between residues 578-812 of type V and the catalytic domain of the type Via' isoform (McAllister et al., 1993). Interestingly, the conserved histidines in the catalytic domain form a consensus sequence for Zn"^"^ binding and in fact Zn"*"^ has been shown to bind tightly to this isoform (Francis et al, 1994). The type V PDEs are regulated by phosphorylation. The cGMP-binding PDE contains a single site for phosphorylation by G-kinase or A-kinase; in either case, phosphorylation only occurs when cGMP is bound to the noncatalytic cGMP binding site (Thomas et al., 1990a,b; Robichon, 1991). The serine phosphorylated in vitro is at ORF position 92 (McAllister et al., 1993). Data now exists to suggest the catalytic subunit of protein kinase A triggers activation of the type V PDE from guinea pig lung (Bums et al., 1992). Furthermore, protein kinase A can modulate the inhibitory effects of zaprinast on the type V PDE via perturbations of a noncatalytic binding site, changing the kinetics of inhibition from mixed in the uninhibited (dephospho) form to competitive in the phosphoenzyme (Bums and Pyne, 1992). It has been suggested that cAMP produces this effect through an interaction with the ip form of what is normally considered the cGMP-stimulated protein kinase (Francis et al., 1988; Lincoln and Comwell, 1993b). Selective

106

J. KELLEY BENTLEY and JOSEPH A. BEAVO

inhibitors of this isozyme have major effects on blood pressure and renal function. For example, zaprinast, a selective type V PDE inhibitor, will cause extensive natriuresis even in the face of profound systemic hypotension (McMahon et al., 1989). It is thought that this effect is mediated via inhibition of type V renal PDE activity. Type VI PDEs (cGMP-Binding)

The type VI PDEs are a group of cGMP-specific enzymes found in retinal photoreceptor cells and have an apy heterotrimeric structure. Type VI PDEs play an essential role in photoreceptor function, since their primary role is to hydrolyze cGMP in response to light activation of rhodopsin (Gillespie, 1990; Chabre and Vuong, 1992). This activation occurs by way of the GTP-binding a subunit of transducin (Gta) (Hurley, 1987). The active PDE is a heterodimer of an a and a p subunit, each of which exist in distinct isoforms in the rod and cone cells. The 93.5 kDa cone PDE a' subunit is slightly larger than either the rod PDE a (88 kDa) or rod PDE p (84 kDa) subunit (Gillespie and Beavo, 1988). Molecular cloning of these isoforms has revealed that they are related to each other structurally (Ovchinnikov et al., 1987; Lipkin et al., 1990a). The p-subunit has 853 residues with a calculated molecular mass of 98 kDa and is 72% identical to the rod cGMP phosphodiesterase a-subunit. The P- and a-subunits are more similar to each other than either is to the cone a'-subunit (Lipkin et al., 1990a). A p' isozyme putatively from cones has also been described which seems to be produced by alternative splicing of the p-subunit gene (Baehr et al, 1991). The membrane association of these subunits is related to their isoprenylation (Qin et al., 1992). The type VI PDEs also have an 11 kDa inhibitory y subunit which is trypsinsensitive (Hurley, 1987). The stoichiometry of the purified aPy complex seems to be 1:1:2 (Deterre et al., 1988). The y subunit has been cloned and sequenced in its entirety (Ovchinnikov et al., 1986), and site-directed mutagenesis has revealed that the y fragment 24-33, which is rich in basic amino acids, and, in particular, Arg-24, is essential for PDE y binding both to the PDE catalytic subunits (a and P) and to GtaThe C-terminal fragment ofy participates in PDE inhibition and in binding to Gta:PDE, but not in binding to PDE ap (Lipkin et al., 1990b; Artemyev and Hamm, 1992). Preparations enriched in the cone-specific PDE a' subunit have two other low-molecular weight subunits. It is presumed that many of the physiological differences in the response of rods and cones to light are mediated by differences in the properties of the rod and cone members of the visual transduction cascade. In the case of the PDEs, the cone PDE a' preparation is more sensitive to the rod Gtoi than the rod a-enriched enzyme preparation (Gillespie and Beavo, 1988). Cones have a unique 15 kDa subunit of as yet unknown function. In addition, a 13 kDa PDE inhibitory subunit is unique to a subset of bovine retinal cones (Hamilton and Hurley, 1990). Rod PDE (aP) has similar Kj values (approximately 80 pM) for both

Regulation of cNMP Synthesis and Deregulation

107

the rod and cone y subunits. However, activated cone PDE has Kj values of 200 pM for the cone 13 kDa subunit and 600 pM for the 11 kDa rod y (Hamilton et al., 1993). The interaction of the y subunits with other photoreceptor proteins may prove even more complex than this; it has recently been demonstrated that an immobilized bacterially expressedy fusion protein can also precipitate arrestin (Qin and Baehr, 1993). Retinal degeneration in the rd mouse is caused by a defect in the P subunit of rod cGMP-PDE (Bowes et al., 1990). Mice homozygous for the rJ mutation display hereditary retinal degeneration and the classic rd lines serve as a model for human retinitis pigmentosa. Degeneration is preceded by accumulation of cGMP in the retina and is correlated with deficient activity of the rod photoreceptor type VI. The mouse rd locus encodes the rod photoreceptor type VI p subunit. In mice, this mutation is the result of the integration of a retrovirus into the mouse genome. The Xmv-28 pro virus is integrated into intron I of the re/gene 1511 bp downstream of the exon-intron boundary. The provirus is present in every strain of rd mouse tested so far (Bowes et al., 1993). The phenotype may be rescued by the transgenic expression of p (Lem et al., 1992). It is interesting to note that the human homolog of the rd gene maps very close to the genetic locus for Huntington's disease gene locus (Altherr et al., 1992; Bateman et al., 1992). Additionally, the P subunit cDNA, normally considered to encode a rod-specific enzyme, may be isolated from normal human brain cDNA libraries (Collins et al., 1992). The significance of these facts is unknown at this point. Normally, the expression of photoreceptor genes is tightly regulated in development (Timmers et al., 1993). The y subunit interaction with both transducin and the PDE determines whether or not the PDE is activated. In amphibians, cGMP has been known to suppress Gta GTPase activity (see Figure 9) (Arshavsky et al., 1991). The interaction of Gta with the y subunit accelerates Gt^ GTPase activity by severalfold and can be partially suppressed by cGMP. This has been suggested to be through the N-terminal cGMP-binding sites of PDE a and/or p (Arshavsky and Bownds, 1992). This cGMP regulation may function in light adaptation as a negative feedback that decreases the lifetime of activated cGMP phosphodiesterase as light causes decreases in cytoplasmic cGMP. This nucleotide is known to tightly bind the noncatalytic cGMP binding domains of the ap heterodimer (Gillespie and Beavo, 1989). The N-terminal noncatalytic cyclic GMP-binding sites regulate the binding of aP with y. If the noncatalytic sites are empty, Gta-GTP physically removes y from the heterodimer (Figure 9). If the noncatalytic sites are occupied by cGMP, Gta-GTP releases y inhibitory action but remains bound in a complex with the PDE heterotetramer (Arshavsky et al., 1992). Further evidence for the existence of such a heterotrimer in mammals comes from resonance energy transfer studies (Erickson and Cerione, 1993). It has also been suggested that the tight binding of y with PDE aP when the noncatalytic sites are occupied by cGMP may be responsible for the low level of

108

J. KELLEY BENTLEY and JOSEPH A. BEAVO hv

Vv.

[GtafiyiGD?) - » . (^GtgiGTP

[cGMP;PDEapPytGta;GTP] Long Active Complex (Slow GTPase)

GMP

Figure 9. cGMP binding to the type VI heterodimer facilitates its activation. Light (hv) interaction with rhodopsin (R) activates the receptor (*). This causes the GTP-binding protein, transducin, to release its GTP-bound a subunit (Gta). Recent data (Arshavsky et al., 1992) suggests that this subunit has two potential fates. It either interacts directly with the y subunit of the PDE and removes it when no cGMP is bound to the PDEap heterodimer noncatalytic site. This activates the PDE, but the GTPase has a relatively rapid rate of turnoff. When cGMP is bound to the noncatalytic site of the PDE, the Gta binds to the Py and remains associated with the PDEap as an active heterotetramer. This complex has a relatively slow rate of GTPase activity; therefore, the enzyme is activated longer.

basal PDE activity observed in dark-adapted cells. Therefore, occupancy of the noncatalytic sites ultimately controls the rate of PDE inactivation, for the GTPase activity that terminates PDE activity is slower when these sites are occupied and Gta stays in a complex with PDE holoenzyme (Arshavsky and Bownds, 1992). In contrast, as shown in Figure 3, GTPase acceleration would be maximal when the noncatalytic sites are empty and Gta-y dissociatesfroma|3 (Arshavsky et al., 1992). Type VII PDEs (HCPl Family)

HCPl is a high affinity cAMP-specific PDE that does not share other properties of the type IV family. In early studies, a mutationally activated yeast RAS2 gene, RAS2^^'^^, was used with great success to isolate cAMP-specific PDEsfromboth yeast and animal cDNA expression libraries, since this mutation normally leads to an increase in cellular cAMP and associated phenotypes in yeast (Sass et al., 1986; Nikawa et al., 1987; ColicelH et al., 1989, 1991). More recently, a strain of yeast has been developed in which the two endogenous yeast cAMP-dependent PDEs have been disrupted (Michaeli, 1993). Using this strain, a new human PDE isoform

Regulation of cNMP Synthesis and Deregulation

109

(called HCPl) has been isolated that has a low K^ for cAMP (200 nM). The HCPl PDE differs from the previously described type IV PDEs in that it is not rolipramsensitive and does not have substantial homology to the Drosophila dunce gene product isoforms. This PDE is encoded by a 3.8 kb transcript most abundant in skeletal muscle and also detectable in kidney and heart. An additional 4.0 kb transcript is also detectable in brain, kidney, and pancreas. Within the 300 amino acid C-terminal conserved catalytic domain, HCPl has the highest degree of homology to the type IV PDEs (35% identity, 51% similarity). It should be noted that among the type IV PDEs, the homology within this domain ranges between 85-95%, further establishing that this is a member of a distinct PDE family. The function and the regulation of these proteins will doubtless be of interest.

CONCLUSION It is now very clear that many different isoforms of adenylate and guanylate cyclase and cyclic nucleotide PDE have evolved to control cyclic nucleotide synthesis and degradation. However, it is not yet clear exactly how many exist, how they are regulated, and how they work in concert with other signal transduction systems to regulate cellular homeostasis. Progress is being made and the next few years promise to be exciting ones as these questions begin to be answered. REFERENCES Altherr, M.R., Wasmuth, J. J., Seldin, M.F., Nadeau, J.H., Baehr, W., & Pittler, S.J. (1992). Chromosome mapping of the rod photoreceptor cGMP phosphodiesterase beta-subunit gene in mouse and human: Tight linkage to the Huntington disease region (4pl6.3). Genomics 12(4), 750-754. Arshavsky, V.Y. & Bownds, M.D. (1992). Regulation of deactivation of photoreceptor G protein by its target enzyme and cGMP. Nature 357(6377), 416-417. Arshavsky, V.Y., Dumke, C.L., & Bownds, M.D. (1992). Noncatalytic cGMP-binding sites of amphibian rod cGMP phosphodiesterase control interaction with its inhibitory gamma-subunits. A putative regulatory mechanism of the rod photoresponse. J. Biol. Chem. 267(34), 24501-24507. Arshavsky, V.Y., Gray, K.M.P., & Bownds, M.D. (1991). cGMP suppresses GTPase activity of a portion of transducin equimolar to phosphodiesterase in frog rod outer segments. Light-induced cGMP decreases as a putative feedback mechanism of the photoresponse. J. Biol. Chem. 266(28), 18530-18537. Artemyev, N.O. & Hamm, H.E. (1992). Two-site high-affinity interaction between inhibitory and catalytic subunits of rod cyclic GMP phosphodiesterase. Biochem J. Baehr, W., Champagne, M.S., Lee, A.K., & Pittler, S.J. (1991). Complete cDNA sequences of mouse rod photoreceptor cGMP phosphodiesterase alpha- and beta-subunits, and identification of beta'-, a putative beta-subunit isozyme produced by alternative splicing of the beta-subunit gene. FEBS Lett. 278(1), 107-114. Bakalyar, H.A. & Reed, R.R. (1990). Identification of a specialized adenylyl cyclase that may mediate odorant detection. Science 250(4986), 1403-1406. Balaban, CD., Billingsley, M.L., & Kincaid, R.L. (1989). Evidence for transsynaptic regulation of calmodulin-dependent cyclic nucleotide phosphodiesterase in cerebellar Purkinje cells. J. Neurosci. 9(7), 2374^2381.

110

j . KELLEY BENTLEY and JOSEPH A. BEAVO

Bateman, J.B., Klisak, I., Kojis, T., Mohandas, T., Sparkes, R.S., Li, T.S., Applebury, M.L., Bowes, C, & Farber, D.B. (1992). Assignment of the beta-subunit of rod photoreceptor cGMP phosphodiesterase gene PDEB (homolog of the mouse rd gene) to human chromosome 4pl 6/Genomics 12(3), 601-^03. Beavo, J.A. (1995). Cyclic nucleotide phosphodiestereases: Functional implications of multiple isoforms. Physiol. Rev. 75, 725-748. Beavo, J. & Houslay, M.D. (eds.) (1990). Cyclic Nucleotide Phosphodiesterases; Structure, Function, Regulation, and Drug Action. John Wiley & Sons, Chichester. Beavo, J. A., Hardman, J.G., & Sutherland, E.W. (1971). Stimulation of adenosine 3',5'-monophosphate hydrolysis by guanosine 3',5'-monophosphate. J. Biol. Chem. 246, 3841—3846. Beltman, J., Becker, D.E., Butt, E., Jastorff, B., & Beavo, J.A. (1993). Characterization of cyclic nucleotide phosphodiesterases with cyclic GMP analogs: The topology of the catalytic domain. Mol. Pharmacol, submitted. Bentley, J.K. & Beavo, J.A. (1992). Regulation and function of cyclic nucleotides. Curr. Opin. Cell Biol. 4, 233-240. Bentley, J.K., Kadlecek, A., Sherbert, C.H., Seger, D., Sonnenburg, W.K., Charbonneau, H., Novack, J.P., & Beavo, J.A. (1992). Molecular cloning of cDNA encoding a "63"-kDa calmodulin-stimulated phosphodiesterase from bovine brain. J. Biol. Chem. 267(26), 18676-18682. Borisy, F.F., Ronnett, G.V., Cunningham, A.M., Juilfs, D., Beavo, J., & Snyder, S.H. (1992). Calcium/calmodulin-activated phosphodiesterase expressed in olfactory receptor neurons. J. Neurosci. 12(3), 915-923. Bowes, C, Li, T., Danciger, M., Baxter, L.C., Applebury, M.L., & Farber, D.B. (1990). Retinal degeneration in the rd mouse is caused by a defect in the beta subunit of rod cGMP-phosphodiesterase [see comments]. Nature 347(6294), 677-680. Bowes, C, Li, T., Frankel, W.N., Danciger, M., Coffin, J.M., Applebury, M.L., & Farber, D.B. (1993). Localization of a retroviral element within the rd gene coding for the beta subunit of cGMP phosphodiesterase. Proc. Natl. Acad. Sci. USA 90(7), 2955-2959. Boyes, S. & Loten, E.G. (1988a). Purification of an insulin-sensitive cyclic AMP phosphodiesterase from rat liver. Eur. J. Biochem. 174(2), 303-309. Boyes, S. & Loten, E.G. (1988b). Purification of insulin-stimulated phosphodiesterase from rat liver. Eur. J. Biochem. 174(2), 303-309. Bredt, D.S., Hwang, P.M., Glatt, C.E., Lowenstein, C, Reed, R.R., & Snyder, S.H. (1991). Cloned and expressed nitric oxide synthase structurally resembles cytochrome P-450 reductase. Nature 351(6329), 714-718. Bredt, D.S. & Snyder, S.H. (1990). Isolation of nitric oxide synthase, a calmodulin-requiring enzyme. Proc. Natl. Acad. Sci. USA 87, 682-685. Buechler, W.A., Nakane, M., & Murad, F. (1991). Expression of soluble guanylate cyclase activity requires both enzyme subunits. Biochem. Biophys. Res. Commun. 174(1), 351—357. Bunney, B.S., Chiodo, L.A., & Grace, A.A. (1991). Midbrain dopamine system electrophysiological function: A review and new hypothesis. Synapse 9, 79-94. Bums, F. & Pyne, N.J. (1992). Interaction of the catalytic subunit of protein kinase A with the lung type V cyclic GMP phosphodiesterase: Modulation of non-catalytic binding sites. Biochem. Biophys. Res. Commun. 189(3), 1389-1396. Bums, F., Rodger, I. W., & Pyne, N.J. (1992). The catalytic subunit of protein kinase A triggers activation of the type V cyclic GMP-specific phosphodiesterase from guinea-pig lung. Biochem. J. 283, 487-491. Butt, E., Beltman, J., Becker, D.E., Jensen, G.S., Jastorff, B., & Beavo, J.A. (1993). Characterization of cyclic nucleotide phosphodiesterases using cAMP analogs: The topology of the catalytic site and comparison to other cAMP-binding sites. Mol. Pharm. (submitted). Buxton, I.L. & Bmnton, L.L. (1985). Action of the cardiac alpha 1-adrenergic receptor. Activation of cyclic AMP degradation. J. Biol. Chem. 260(11), 6733-6737.

Regulation of cNMP Synthesis and Deregulation

111

Chabre, M. & Vuong, T.M. (1992). Kinetics and energetics of the rhodopsin-transducin-cGMP phosphodiesterase cascade of visual transduction. Biochim. Biophys. Acta 1101(2), 260-263. Chang, M.S., Lowe, D.G., Lewis, M., Hellmiss, R., Chen, E., & Goeddel, D.V. (1989). Differential activation by atrial and brain natriuretic peptides of two different receptor guanylate cyclases. Nature 341(6237), 68-72. Charbonneau, H. (1990). Structure-function relationships among cyclic nucleotide phosphodiesterases. In: Cyclic Nucleotide Phosphodiesterases: Structure, Function, Regulation and Drug Action (Beavo, J. & Houslay, M.D., eds.), pp. 267-298, John Wiley & Sons, Chichester. Charbonneau, H., Kumar, S., Novack, J.P., Blumenthal, D.K., Griffin, P.R., Shabanowitz, J., Hunt, D.F., Beavo, J.A., & Walsh, K.A. (1991). Evidence for domain organization within the 61-kDa calmodulin-dependent cyclic nucleotide phosphodiesterase from bovine brain. Biochemistry 30(32), 7931-7940. Charbonneau, H., Prusti, R.K., LeTrong, H., Sonnenburg, W.K., MuUaney, P.J., Walsh, K.A., & Beavo, J. A. (1990). Identification of a noncatalytic cGMP-binding domain conserved in both the cGMPstimulated and photoreceptor cyclic nucleotide phosphodiesterases. Proc. Natl. Acad. Sci. USA 87(1), 288-292. Chinkers, M. (1994). Targeting of a distinctive protein-serine phosphatase to the protein kinase-like domain of the atrial natriuretic peptide receptor. Proc. Natl. Acad. Sci. USA 91 (23), 11075-11079. Chinkers, M. & Garbers, D.L. (1989). The protein kinase domain of the ANP receptor is required for signaling. Science 245(4924), 1392-1394. Chinkers, M., Garbers, D.L., Chang, M.S., Lowe, D.G., Chin, H.M., Goeddel, D.V., & Schulz, S. (1989). A membrane form of guanylate cyclase is an atrial natriuretic peptide receptor. Nature 338(6210), 78-«3. Chinkers, M., Singh, S., & Garbers, D.L. (1991). Adenine nucleotides are required for activation of rat atrial natriuretic peptide receptor/guanylyl cyclase expressed in a baculovirus system. J. Biol. Chem. 266(7), 4088-4093. Chinkers, M. & Wilson, E.M. (1992). Ligand-independent oligomerization of natriuretic peptide receptors. Identification of heteromeric receptors and a dominant negative mutant. J. Biol. Chem. 267(26), 18589^18597. Cho, H.J., Xie, Q.W., Calaycay, J., Mumford, R.A., Swiderek, K.M., Lee, T.D., & Nathan, C. (1992). Calmodulin is a subunit of nitric oxide synthase from macrophages. J. Exp. Med. 176(2), 599-604. Choi, E.J., Xia, Z., Villacres, E.C., and Storm, D.R. (1993). The regulatory diversity of adenylyl cyclases. Curr. Opin. Cell Biol. 5, 269-273. Choi, E.J., Wong, S.T., Dittman, A.H., & Storm, D.R. (1993). Phorbol ester stimulation of the type I and type III adenylyl cyclases in whole cells. Biochemistry 32(8), 1891-1894. Choi, E.J., Xia, Z., & Storm, D.R. (1992). Stimulation of the type III olfactory adenylyl cyclase by calcium and calmodulin. Biochemistry 31(28), 6492-6498. Colicelli, J., Birchmeier, C, Michaeli, T., O'Neill, K., Riggs, M., & Wigler, M. (1989). Isolation and characterization of a mammalian gene encoding a high-affinity cAMP phosphodiesterase. Proc. Natl. Acad. Sci. USA 86(10), 3599^3603. Colicelli, J., Nicolette, C, Birchmeier, C, Rodgers, L., Riggs, M., & Wigler, M. (1991). Expression of three mammalian cDNAs that interfere with RAS function in Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. USA 88(7), 2913-2917. Collins, C, Hutchinson, G., Kowbel, D., Riess, O., Weber, B., & Hayden, M.R. (1992). The human beta-subunit of rod photoreceptor cGMP phosphodiesterase: Complete retinal cDNA sequence and evidence for expression in brain. Genomics 13(3), 698-704. Conti, M., Jin, S.L.C., Monaco, L., Repaske, D.R., & Swinnen, J.V. (1991). Hormonal regulation of cyclic nucleotide phosphodiesterases. Endocrine Rev. 12(3), 218-234. Conti, M., Nemoz, G., et al. (1995). Recent progress in understanding the hormonal regulation of phosphodiesterases. Endocr. Rev. 16(3), 370-389.

112

J. KELLEY BENTLEY and JOSEPH A. BEAVO

Conti, M., Swinnen, J.V., Tsikalas, K.E., & Jin, S. L. (1992). Structure and regulation of the rat high-affinity cyclic AMP phosphodiesterases. A family of closely related enzymes. Adv. Second Messenger Phosphoprotein Res. 25, 87-99. Cooper, D.M., Bier, L.C.M., Halford, M.K., Ahlijanian, M.K., & Zahniser, N.R. (1986). Dopamine, acting through D-2 receptors, inhibits rat striatal adenylate cyclase by a GTP-dependent process. Mol. Pharmacol. 29(2), 113-119. Coquil, J.F. (1985). Properties of the cyclic GMP phosphodiesterase from rat lung. Inhibition by unsaturated fatty acids. Biochem. Biophys. Res. Commun. 127, 226-230. Coquil, J.F., Franks, D.J., Wells, J.N., Dupuis, M., & Hamet, P. (1983). Characteristics of a new binding protein distinct from the kinase for guanosine 3',5'-monophosphate in rat platelets. Biochim. Biphys. Acta 743, 359-369. Coquil, J.F., G.B., & Guedon, J. (1980). Occurrence of the methylisobutylxanthine-stimulated cyclic GMP binding protein in various rat tissues. Biochim. Biophys. Acta 631, 148-165. Currie, M.G., Fok, K.F., Kato, J., Moore, R.J., Hamra, F.K., Duffin, K.L., & Smith, C.E. (1992). Guanylin: An endogenous activator of intestinal guanylate cyclase. Proc. Natl. Acad. Sci. USA 89(3), 947-951. Dage, R.C. & Okerholm, R.A. (1990). Pharmacology and pharmacokinetics of enoximone. Cardiology 3,2-13. Davis, C.W. & Kuo, J.F. (1982). Purification and characterization of guanosine-3',5'-specific phosphodiesteraseft-omguinea pig lung. J. Cardiovasc. Pharmacol. 4, 500-508. Davis, R.L., Takayasu, H., Eberwine, M., & Myres, J. (1989). Cloning and characterization of mammalian homologs of the Drosophila dunce+ gene. Proc. Natl. Acad. Sci. USA 86(10), 3604^3608. Degerman, E., Belfi-age, P., Newman, A.H., Rice, K.C., & Manganiello, V.C. (1987). Purification of the putative hormone-sensitive cyclic AMP phosphodiesterasefi^omrat adipose tissue using a derivative of cilostamide as a novel affinity ligand. J. Biol. Chem. 262(12), 5797—5807. Deterre, P., Bigay, J., Forquet, F., Robert, M., & Chabre, M. (1988). cGMP phosphodiesterase of retinal rods is regulated by two inhibitory subunits. Proc. Natl. Acad. Sci. USA 85(8), 2424-2428. Dizhoor, A.M., Lowe, D.G., et al. (1994). The human photoreceptor membrane guanylyl cyclase, RetGC, is present in outer segments and is regulated by calcium and a soluble activator. Neuron 12(6), 1345-1352. Dizhoor, A.M., Olshevakaya, E.V., et al. (1995). Cloning, sequencing, and expression of a 24-kDa Ca(2+)-binding protein activating photoreceptor guanylyl cyclase. J. Biol. Chem. 270(42), 2520025206. Dizhoor, A.M., Ray, S., Kumar, S., Niemi, G., Spencer, M., Brolley, D., Walsh, K.A., Philipov, P.P., Hurley, J.B., & Stryer, L. (1991). Recoverin: A calcium sensitive activator of retinal rod guanylate cyclase. Science 251(4996), 915-918. Drewett, J.G. & Garbers, D.L. (1994). The family of guanylyl cyclase receptors and their ligands. Endocr. Rev. 15(2), 135-162. Duda, T., Goraczniak, R.M., Sitaramayya, A., & Sharma, R.K. (1993). Cloning and expression of an ATP-regulated human retina C-type natriuretic factor receptor guanylate cyclase. Biochemistry 32(6), 1391-1395. Dumont, J.E., Mitot, F., Emeux, C, Couchie, D., Cochaux, P., Gervy, D.C., Van, S.J., & N., a.W.J. (1984). Negative regulation of cyclic AMP levels by activation of cyclic nucleotide phosphodiesterases: The example of the dog thyroid. Biol. Membr. 1(565), 565—572. Erickson, J.W. & Cerione, R.A. (1993). Regulation of the cGMP phosphodiesterase in bovine rod outer segments. Use of resonance energy transfer to distinguish between associative and dissociative activation mechanisms. J. Biol. Chem. 268(5), 3328-3333. Emeux, C, Couchie, D., Dumont, J.E., Baraniak, J., Stec, W.J., Abbad, E.G., Petridis, G., & Jastorff, B. (1981). Specificity of cyclic GMP activation of a multi-substrate cyclic nucleotide phosphodiesterase from rat liver. Eur. J. Biochem. 115(3), 503-510.

Regulation of cNMP Synthesis and Deregulation

113

Feinstein, P.G., Schrader, K.A., Bakalyar, H.A., Tang, W.J., Kruoinski, J., Gilman, A.G., & Reed (1991). Molecular cloning and characterization of a Ca^"^/calmodulin-insensitive adenylyl cyclase from rat brain. Proc. Natl. Acad. Sci. USA 88, 10173-10177. Foerstermann, U., Gorsky, L.D., Pollock, J.S., Ishii, K., Schmidt, H.H., Heller, M., & Murad, F. (1990). Hormone-induced biosynthesis of endothelium-derived relaxing factor/nitric oxide-like material in NIE-115 neuroblastoma cells requires calcium and calmodulin. Mol. Pharmacol. 38(1), 7-13. Francis, S.H., Colbran, J.L., et al. (1994). Cyclic nucleotide phosphodiesterases are zinc hydrolases as indicated by conserved zinc-binding motifs, specific zinc binding, and zinc supported catalysis. FASEBJ. 8(4),A371. Francis, S.H., Noblett, B.D., Todd, B.W., Wells, J.N., & Corbin, J.D. (1988). Relaxation of vascular and tracheal smooth muscle by cyclic nucleotide analogs that preferentially activate purified cGMP-dependent protein kinase. Mol. Pharmacol. 34(4), 506-517. Francis, S.H., Thomas, M.K., & Corbin, J.D. (1990). Cyclic GMP-binding cyclic GMP-specific phosphodiesterase from lung. In: Cyclic Nucleotide Phosphodiesterases: Structure, Regulation and Drug Action (Beavo, J. & Houslay, M.D., eds.), pp. 117-140, John Wiley & Sons, Chichester. Fulle, H.J., Vassar, R., et al. (1995). A receptor guanylyl cyclase expressed specifically in olfactory sensory neurons. Proc Natl. Acad. Sci. USA 92(8), 3571-3575. Fuller, F., Porter, J.G., Arfsten, A.E., Miller, J., Schilling, J.W., Scarborough, R.M., Lewicki, J.A., & Schenk, D.B. (1988). Atrial natriuretic peptide clearance receptor. Complete sequence and functional expression of cDNA clones. J. Biol. Chem. 263(19), 9395-9401. Gao, B. & Gilman, A.G. (1991). Cloning and expression of a widely distributed (type IV) adenylyl cyclase. Proc. Natl. Acad. Sci. USA 88, 10178-10182. Garbers, D.L. (1989). Molecular basis of fertilization. Ann. Rev. Biochem. 58, 719-742. Garbers, D.L. (1992). Guanylyl cyclase receptors and their endocrine, paracrine, and autocrine ligands. Cell 71(1), 1-4. Gettys, T.W., Blackmore, P.F., Redmon, J.B., Beebe, S.J., & Corbin, J.D. (1987). Short-term feedback regulation of cAMP by accelerated degradation in rat tissues. J. Biol. Chem. 262(1), 333-339. Gettys, T.W., Vine, A.J., Simonds, M.F., & Corbin, J.D. (1988). Activation of the particulate low Km phosphodiesterase of adipocytes by addition of cAMP-dependent protein kinase. J. Biol. Chem. 263(21), 10359-10363. Gillespie, P.G. (1990). Phosphodiesterases in visual transduction by rods and cones. In: Cyclic Nucleotide Phosphodiesterases: Structure, Regulation and Drug Action (Beavo, J. & Houslay, M.D., eds.), pp. 163-184, John Wiley & Sons, Chichester. Gillespie, P.G. & Beavo, J.A. (1988). Characterization of a bovine cone photoreceptor phosphodiesterase purified by cyclic GMP-sepharose chromatography. J. Biol. Chem. 263(17), 8133-8141. Gillespie, P.G. & Beavo, J.A. (1989). cGMP is tightly bound to bovine retinal rod phosphodiesterase. Proc. Natl. Acad. Sci. USA 86(11), 4311^315. Giuili, G., Scholl, U., Bulle, F., & Guellaen, G. (1992). Molecular cloning of the cDNAs coding for the two subunits of soluble guanylyl cyclase from human brain. FEBS Lett. 304(1), 83-88. Glatt, C.E. & Snyder, S.H. (1993). Cloning and expression of an adenylyl cyclase localized to the corpus striatum. Nature 361(6412), 536-538. Gorczyca, W.A., Gray, K.M.P., et al. (1994). Purification and physiological evaluation of a guanylate cyclase activating protein from retinal rods. Proc. Natl. Acad. Sci. USA 91(9), 4014-4018. Gorczyca, W.A., Polans, A.S., et al. (1995). Guanylyl cyclase activating protein—a calcium-sensitive regulator of phototransduction. J. Biol. Chem. 270(37), 22029-22036. Grant, P.G., Mannarino, A.F., & Colman, R.W. (1990). Purification and characterization of a cyclic GMP-stimulated cyclic nucleotide phosphodiesterase from the cytosol of human platelets. Thromb. Res. 59(1), 105-119. Hamet, P., Coquil, J.F., Bousseau, L.S., Franks, D.J., & Tremblay, J. (1984). Cyclic GMP binding and phosphodiesterase: Implication for platelet function. Adv. Cyclic Nucleotide Protein Phosphorylation Res. 16, 119-136.

114

J. KELLEY BENTLEY and JOSEPH A. BEAVO

Hamilton, S.E. & Hurley, J.B. (1990). A phosphodiesterase inhibitor specific to a subset of bovine retinal cones. J. Biol. Chem. 265(19), 11259-11264. Hamilton, S.E., Prusti, R.K., Bentley, J.K., Beavo, J.A., & Hurley, J.B. (1993). Affinities of bovine photoreceptor cGMP phosphodiesterases for rod and cone inhibitory subunits. FEBS Lett. 318(2), 157-161. Hansen, R.S. & Beavo, J.A. (1982). Purification of two calcium/calmodulin dependent forms of cyclic nucleotide phosphodiesterase by using conformation-specific monoclonal antibody chromatography. Proc. Natl. Acad. Sci. USA 79, 2788-2792. Hansen, R.S. & Beavo, J.A. (1986). Differential recognition of calmodulin-enzyme complexes by a conformation-specific anti-calmodulin monoclonal antibody. J. Biol. Chem. 261, 14636-14645. Harrison, S.A., Chang, M.L., & Beavo, J.A. (1986a). Differential inhibition of cardiac cyclic nucleotide phosphodiesterase isozymes by cardiotonic drugs. Circulation 73(suppl III), 109-116. Harrison, S.A., Reifsnyder, D.H., Gallis, B., Cadd, G.G., & Beavo, J.A. (1986b). Isolation and characterization of bovine cardiac muscle cGMP-inhibited phosphodiesterase: A receptor for new cardiotonic drugs. Mol. Pharmacol. 29, 506-514. Harteneck, C, Koesling, D., Soeling, A., Schultz, G., & Boehme, E. (1990). Expression of soluble guanylyl cyclase. Catalytic activity requires two enzyme subunits. FEBS Lett. 272( 1-2), 221-223. Hashimoto, Y., Sharma, R.K., & Soderling, T.R. (1989). Regulation of Ca^'^/calmodulin-dependent cyclic nucleotide phosphodiesterase by the autophosphorylated form of Ca^"*"/calmodulin-dependent protein kinase II. J. Biol. Chem. 264(18), 10884^10887. Hellevuo, K., Yoshimura, M., Kao, M., Hoffman, P.L., Cooper, D.M., & Tabakoff, B. (1993). A novel adenylyl cyclase sequence cloned from the human erythroleukemia cell line. Biochem. Biophys. Res. Commun. 192(1), 311-318. Hendra, T. & Betteridge, D.J. (1989). Platelet function, platelet prostanoids and vascular prostacyclin in diabetes mellitus. Prostaglandins Leukot. Essent. Fatty Acids 35(4), 197-212. Homma, S., Gapstur, S.M., Coffey, A., Valtin, H., & Dousa, T.P. (1991). Role of cAMP-phosphodiesterase isozymes in pathogenesis of murine nephrogenic diabetes insipidus. Am. J. Physiol. Houslay, M.D., Griffiths, S.L., Horton, Y.M., Livingstone, C, Lobban, M., Macdonald, F., Morris, N., Pryde, J., Scotland, G., Shakur, Y., & et. al. (1992). Regulation of intracellular cyclic AMP concentrations in hepatocytes involves the integrated activation and desensitization of adenylyl cyclase coupled with the action and activation of specific isoforms of cyclic AMP phosphodiesterase. Biochem. Soc. Trans. 20(1), 140-146. Hurley, J.B. (1987). Molecular properties of the cGMP cascade of vertebrate photoreceptors. Ann. Rev. Physiol. 49,79^-812. Hurley, J.B., Dizhoor, A.M., Ray, R., & Stryer, L. (1993). Recoverin's role: Conclusion withdrawn. Science 260, 740. Hurwitz, R.L., Hansen, R.S., Harrison, S.A., Martins, T.J., Mumby, M.C., & A., B.J. (1984). Immunologic approaches to the study of cyclic nucleotide phosphodiesterases. Adv. Cyc. Nuc. Prot. Phos.Res. 16(89), 89-106. Hurwitz, R.L., Hirsch, K.M., Clark, D.J., Holcombe, V.N., & Hurwitz, M.Y. (1990). Induction of a calcium/calmodulin-dependent phosphodiesterase during phytohemagglutinin-stimulated lymphocyte mitogenesis. J. Biol. Chem. 265(15), 8901-«907. Ignarro, L.J., Buga, G.M., Wood, K.S., Byms, R.E., & Chaudhuri, G. (1987a). Endothelium-derived relaxing factor produced and released from artery and vein is nitric oxide. Proc. Natl. Acad. Sci. USA 84(24), 9265-9269. Ignarro, L.J., Byms, R.E., & Wood, K.S. (1987b). Endothelium-dependent modulation of cGMP levels and intrinsic smooth muscle tone in isolated bovine intrapulmonary artery and vein. Circ. Res. 60(1), 82-92. Impey, S., Wayman, G., et al. (1994). Type I adenylyl cyclase functions as a coincidence detector for control of cyclic AMP response element-mediated transcription: synergistic regulation of transcription by Ca^"" and isoproterenol. Mol. Cell Biol. 14(12), 8272-8281.

Regulation of cNMP Synthesis and Deregulation

115

Ishikawa, Y., Katsushika, S., Chen, L., Halnon, N.J., Kawabe, J., & Homey, C.J. (1992). Isolation and charaeterization of a novel cardiac adenylyl cyclase cDNA. J. Biol. Chem. 267( 19), 13553—13557. Jacobowitz, O., Chen, J., Fremont, R.T., & Iyengar, R. (1993). Stimulation of specific types of Gs-stimulated adenylyl cyclases by phorbol ester treatment. J. Biol. Chem. 268(6), 3829-3832. Jewett, J.R., KoUer, K.J., Goeddel, D.V., & Lowe, D.G. (1993). Hormonal induction of low affinity receptor guanylyl cyclase. EMBO J. 12(2), 769-777. Jin, S.L., Swinnen, J.V., & Conti, M. (1992). Characterization of the structure of a low Km, rolipramsensitive cAMP phosphodiesterase. Mapping of the catalytic domain. J. Biol. Chem. 267(26), 18929-18939. Johansen, P.A., Hu, X.T., & White, F.J. (1991). Relationship between D1 dopamine receptors, adenylate cyclase, and the electrophysiological responses of rat nucleus accumbens neurons. J. Neural Transm. Gen. Sect. 86(2), 97-113. Katsushika, S., Chen, L., Kawabe, J., Nilakantan, R., Halnon, N.J., Homey, C.J., & Ishikawa, Y. (1992). Cloning and characterization of a sixth adenylyl cyclase isoform: Types V and VI constitute a subgroup within the mammalian adenylyl cyclase family. Proc. Natl. Acad. Sci. USA 89(18), 8774-8778. Katsushika, S., Kawabe, J., Homey, C.J., & Ishikawa, Y. (1993). In vivo generation of an adenylylcyclase isoform with a half-molecule motif J. Biol. Chem. 268(4), 2273-2276. Kawamura, S. (1993). Rhodopsin phosphorylation as a mechanism of cyclic GMP phosphodiesterase regulation by S-modulin. Nature 362(6423), 855-857. Kennelly, P.J. & Krebs, E.G. (1991). Consensus sequences as substrate specificity determinants for protein kinases and protein phosphatases. J. Biol. Chem. 266(24), 15555-15558. Kincaid, R.L., Balaban, CD., & Billingsley, M.L. (1987). Differential localization of calmodulindependent enzymes in rat brain: Evidence for selective expression of cyclic nucleotide phosphodiesterase in specific neurons. Proc. Natl. Acad. Sci. USA 84(4), 1118-1122. Kincaid, R.L., Balaban, CD., & Billingsley, M.L. (1992). Regional and developmental expression of calmodulin-dependent cyclic nucleotide phosphodiesterase in rat brain. Adv. Second Messenger Phosphoprotein Res. 25, 111-122. Koch, K.W. (1991). Purification and identification of photoreceptor guanylate cyclase. J. Biol. Chem. 266(13), 8634-8637. Koch, K.W. & Stryer, L. (1988). Highly cooperative feedback control of retinal rod guanylate cyclase by calcium ions. Nature 334(6177), 64—66. Koesling, D., Boehme, E., & Schultz, G. (1991a). Guanylyl cyclases, a growing family of signaltransducing enzymes. FASEB J. 5( 13), 2785-2791. Koesling, D., Schultz, G., & Boehme, E. (1991b). Sequence homologies between guanylyl cyclases and structural analogies to other signal-transducing proteins. FEBS Lett. 280(2), 301-306. Koller, K.J., de, S.F.J., Lowe, D.G., & Goeddel, D.V. (1992). Conservation of the kinase like regulatory domain is essential for activation of the natriuretic peptide receptor guanylyl cyclases. Mol. Cell. Biol. 12(6), 2581-2590. Koller, K.J. & Goeddel, D.V. (1992). Molecular biology of the natriuretic peptides and their receptors. Circulation 86(4), 1081-1088. Koller, K.J., Lipari, M.T., & Goeddel, D.V. (1993). Proper glycosylation and phosphorylation of the type A natriuretic peptide receptor are required for hormone-stimulated guanylyl cyclase activity. J. Biol. Chem. 268(8), 5997-^003. Koller, K.J., Lowe, D.G., Bennett, G.L., Minamino, N., Kangawa, K., Matsuo, H., & Goeddel, D.V. (1991). Selective activation of the B natriuretic peptide receptor by C-type natriuretic peptide (CNP). Science 252(5002), 120-123. Kramer, R.H., Levitan, E.S., Wilson, M.P., & Levitan, LB. (1988). Mechanism of calcium-dependent inactivation of a potassium current in Aplysia neuron Rl 5: Interaction between calcium and cyclic AMP. J. Neurosci. 8(5), 1804-1813.

116

J. KELLEY BENTLEY and JOSEPH A. BEAVO

Kramer, R.H. & Levitan, LB. (1988). Calcium-dependent inactivation of a potassium current in the Aplysia neuron R15. J. Neurosci. 8(5), 1796-1803. Krupinski, J. (1991). The adenylyl cyclase family. Mol. Cell Biochem. 104(1-2), 73-79. Krupinski, J., Coussen, F., Bakalyar, H.A., Tang, W.J., Feinstein, P.O., Orth, K., Slaughter, C, Reed, R.R., & Oilman, A.G. (1989). Adenylyl cyclase amino acid sequence: Possible channel- or transporter-like structure. Science 244(4912), 1558-1564. Krupinski, J., Lehman, T.C., Frankenfield, CD., Zwaagstra, J.C, & Watson, P.A. (1992). Molecular diversity in the adenylylcyclase family. Evidence for eight forms of the enzyme and cloning of type VL J. Biol. Chem. 267(34), 24858-24862. Kuhn, M., Raida, M., Adermann, K., Schulz, K.P., Gerzer, R., Heim, J.M., «fe Forssmann, W.G. (1993). The circulating bioactive form of human guanylin is a high molecular weight peptide (10.3 kDa). FEBS Lett. 318(2), 205-209. L'opez, A.P., Rasc'on, A., Manganiello, V.C, Andersson, K.E., Belfrage, P., & Degerman, E. (1992). Insulin induced phosphorylation and activation of the cGMP-inhibited cAMP phosphodiesterase in human platelets. Biochem. Biophys. Res. Commun. 186(1), 517-523. Lacey, M.G., Mercuri, N.B., & North, R.A. (1987). Dopamine acts on D2 receptors to increase potassium conductance in neurons of the rat substantia nigra zona compacta. J. Physiol. 392, 397-^16. Lem, J., Flannery, J.G., Li, T., Applebury, M.L., Farber, D.B., & Simon, M.I. (1992). Retinal degeneration is rescued in transgenic rd mice by expression of the cGMP phosphodiesterase beta subunit. Proc. Natl. Acad. Sci. USA 89(10), 4422-4426. Lincoln, T.M. & Comwell, T.L. (1993a). Intracellular cyclic GMP receptor proteins. FASEB J. 7(2), 328-338. Lincoln, T.M. & Comwell, T.L. (1993b). Intracellular cyclic GMP receptor proteins. FASEB J. 7(2), 328-338. Lipkin, V.M., Khramtsov, N.V., Vasilevskaya, I.A., Atabekova, N.V., Muradov, K.G., Gubanov, V.V., Li, T., Johnston, J.P., VoIpp, K.J., & Applebury, M.L. (1990a). Beta-subunit of bovine rod photoreceptor cGMP phosphodiesterase. Comparison with the phosphodiesterase family. J. Biol. Chem. 265(22), 1295S-12959. Lipkin, V.M., Udovichenko, LP., Bondarenko, V.A., Yurovskaya, A.A., Telnykh, E.V., & Skiba, N.P. (1990b). Site-directed mutagenesis of the inhibitory subunit of retinal rod cyclic GMP phosphodiesterase. Biomed. Sci. 1(3), 305-308. Livi, G.P., Kmetz, P., McHale, M.M., Cieslinski, L.B., Sathe, G.M., Taylor, D.P., Davis, R.L., Torphy, T.J., & Balcarek, J.M. (1990). Cloning and expression of cDNA for a human low-Km, rolipramsensitive cyclic AMP phosphodiesterase. Mol. Cell Biol. 10(6), 2678-2686. Lowe, D.G., Dizhoor, A.M., et al. (1995). Cloning and expression of a second photoreceptor-specific membrane retina guanylyl cyclase (retgc), retgc-2. Proc. Natl. Acad. Sci. USA 92( 12), 5535—5539. Lowenstein, C.J., Glatt, C.S., Bredt, D.S., & Snyder, S.H. (1992). Cloned and expressed macrophage nitric oxide synthase contrasts with the brain enzyme. Proc. Natl. Acad. Sci. USA 89(15), 6711-6715. Ludvig, N., Burmeister, V., Jobe, P.C, & Kincaid, R.L. (1991). Electron microscopic immunocytochemical evidence that the calmodulin-dependent cyclic nucleotide phosphodiesterase is localized predominantly at postsynaptic sites in the rat brain. Neuroscience 44(2), 491-500. MacFarland, R.T., Zelus, B.D., & Beavo, J.A. (1991). High concentrations of a cGMP-stimulated phosphodiesterase mediate ANP-induced decreases in cAMP and steroidogenesis in adrenal glomerulosa cells. J. Biol. Chem. 266(1), 136-142. Macphee, C.H., Harrison, S.H., & Beavo, J. A. (1986). Immunological identification ofthe major platelet low-Km cAMP phosphodiesterase: Probable target for anti-thrombotic agents. Proc. Natl. Acad. Sci. USA 83, 6660-6663.

Regulation of cNMP Synthesis and Deregulation

117

Macphee, C.H., Reifsnyder, D.H., Moore, T.A., & Beavo, J.A. (1988a). Intact cell and cell free phosphorylation and concomitant activation of a low Km; cAMP phosphodiesterase found in human platelets. J. Cyc. Nuc. Prot. Phos. Res. 11, 487-496. Macphee, C.H., Reifsnyder, D.H., Moore, T.A., Lerea, K.M., & Beavo, J.A. (1988b). Phosphorylation results in activation of a cAMP phosphodiesterase in human platelets. J. Biol. Chem. 263(21), 10353-10358. Mahan, L.C., Burch, R.M., Monsma, F.J. J., & Sibley, D.R. (1990). Expression of striatal Dl dopamine receptors coupled to inositol phosphate production and Ca^"^ mobilization in Xenopus oocytes. Proc. Natl. Acad. Sci. USA 87(6), 2196-2200. Manganiello, V.C, Smith, C.J., Degerman, E., & Belfrage, P. (1990a). Cyclic GMP-inhibited cyclic nucleotide phosphodiesterases. In: Cyclic Nucleotide Phosphodiesterases: Structure, Regulation and Drug Action (Beavo, J. & Houslay, M.D., eds.), pp. 87-116, John Wiley & Sons, Chichester. Manganiello, V.C, Tanaka, T., & Murashima, S. (1990b). Cyclic GMP-stimulated cyclic nucleotide phosphodiesterases. In: Cyclic Nucleotide Phosphodiesterases: Structure, Regulation and Drug Action (Beavo, J. & Houslay, M.D., eds.), pp. 61-86, John Wiley & Sons, Chichester. Marietta, M.A. (1993). Nitric oxide synthase structure and mechanism. J. Biol. Chem. 268, 1223112234. Marietta, M.A. (1994). Nitric oxide synthase: aspects concerning structure and catalysis. Cell 78(6), 927-930. Martin, W., Villani, G.M., Jothianandan, D., & Furchgott, R.F. (1985). Selective blockade of endothelium-dependent and glyceryl trinitrate-induced relaxation by hemoglobin and by methylene blue in the rabbit aorta. J. Pharmacol. Exp. Ther. 232(3), 70^716. Martins, T.J., Mumby, M.C., & Beavo, J.A. (1982). Purification and characterization of a cyclic GMP-stimulated cyclic nucleotide phosphodiesterase from bovine tissues. J. Biol. Chem. 257, 1973-1979. Matsuoka, I., Giuili, G., Poyard, M., Stengel, D., Parma, J., Guellaen, G., & Hanoune, J. (1992). Localization of adenylyl and guanylyl cyclase in rat brain by in situ hybridization: Comparison with calmodulin mRNA distribution. J. Neurosci. 12, 3350-3360. Maurice, D.H., Crankshaw, D., & Haslam, R.J. (1991). Synergistic actions of nitrovasodilators and isoprenaline on rat aortic smooth muscle. Eur. J. Pharmacol. 192(2), 235-242. Mayer, B., Klatt, P., Bohme, E., & Schmidt, K. (1992). Regulation of neuronal nitric oxide and cyclic GMP formation by Ca^"". J. Neurochem. 59(6), 2024-2029. McAllister, L.M., Kadlecek, A.T., Sonnenburg, W.T., Seger, D., Francis, S.H., Thomas, M.K., Corbin, J.D., & Beavo, J.A. (1993). Isolation of a cDNA clone for the bovine lung cGMP-binding cGMP-specific phosphodiesterase. J. Biol. Chem. submitted. McMahon, E.G., Palomo, M.A., Mehta, P., & Olins, G.M. (1989). Depressor and natriuretic effects of M&B 22,948, a guanosine cyclic 3',5'-monophosphate-selective phosphodiesterase inhibitor. J. Pharmacol. Exp. Ther. 251(3), 1000-1005. Meacci, E., Taira, M., Moos, M.J., Smith, C.J., Movsesian, M.A., Degerman, E., Belfrage, P., & Manganiello, V. (1992). Molecular cloning and expression of human myocardial cGMP-inhibited cAMP phosphodiesterase. Proc. Natl. Acad. Sci. USA 89(9), 3721-3725. Memo, M. (1990). Pharmacology and molecular basis for dopamine D-2 receptor diversity. Mol. Neurobiol. 4, 181-196. Michaeli, T., Bloom, T.J., Martins, T., Loughney, K., Ferguson, K., Riggs, M., Rodgers, L., Beavo, J.A., and Wigler, M. (1993). Isolation and characterization of a previously undetected human cAMP phosphodiesterase by complementation of cAMP phosphodiesterase-deficient Saccharomyces cerevisiae. J. Biol. Chem. 268, 12925-12932. Miot, F., Dumont, J.E., & Emeux, C. (1983). The involvement of a calmodulin-dependent phosphodiesterase in the negative control of carbamylcholine on cyclic AMP levels in dog thyroid slices. FEBSLett. 151(2), 273-276.

118

J. KELLEY BENTLEY and JOSEPH A. BEAVO

Monsma, F.J.J., Mahan, L.C., McVittie, L.D., Gerfen, C.R., & Sibley, D.R. (1990). Molecular cloning and expression of a Dl dopamine receptor linked to adenylyl cyclase activation. Proc. Natl. Acad. Sci. USA 87(17), 6723-6727. Murashima, S., Tanaka, T., Hockman, S., & Manganiello, V. (1990). Characterization of particulate cyclic nucleotide phosphodiesterases from bovine brain: Purification of a distinct cGMP-stimulated isoenzyme. Biochemistry 29(22), 5285-5292. Nakane, M., Arai, K., Saheki, S., Kuno, T., Buechler, W., & Murad, F. (1990). Molecular cloning and expression of cDNAs coding for soluble guanylate cyclasefromrat lung. J. Biol. Chem. 265(28), 16841-16845. Nakane, M., Ichikawa, M., & Deguchi, T. (1983). Light and electron microscopic demonstration of guanylate cyclase in rat brain. Brain Res. 273(1), 9-15. Nikawa, J., Sass, P., & Wigler, M. (1987). Cloning and characterization of the low-affinity cyclic AMP phosphodiesterase gene of Saccharomyces cerevisiae. Mol. Cell Biol. 7(10), 3629-3636. Novack, J.P., Charbonneau, H., Bentley, J.K., Walsh, K.A., & Beavo, J.A. (1991). Sequence comparison of the 63-, 61-, and 59-kDa calmodulin-dependent cyclic nucleotide phosphodiesterases. Biochemistry 30(32), 7940-7947. Ovchinnikov, Y.A., Gubanov, V.V., Khramtsov, N.V., Ischenko, K.A., Zagranichny, V.E., Muradov, K.G., Shuvaeva, T.M., & Lipkin, V.M. (1987). Cyclic GMP phosphodiesterasefrombovine retina. Amino acid sequence of the alpha-subunit and nucleotide sequence of the corresponding cDNA. FEBS Lett. 223(1), 169-173. Ovchinnikov, Y.A., Lipkin, V.M., Kumarev, V.P., Gubanov, V.V., Khramtsov, N.V., Akhmedov, N.B., Zagranichny, V.E., & Muradov, K.G. (1986). Cyclic GMP phosphodiesterase from cattle retina. Amino acid sequence of the gamma-subunit and nucleotide sequence of the corresponding cDNA. FEBS Lett. 204(2), 288-292. Palczewski, K., Subbaraya, L, et al. (1994). Molecular cloning and characterization of retinal photoreceptor guanylyl cyclase-activating protein. Neuron 13(2), 395-404. Palmer, R.J.M., Ferrige, A.G., & Moncada, S. (1987). Nitric oxide accounts for the biological activity of endothelium relaxing factor. Nature 327, 524-526. Polli, J.W. & Kincaid, R.L. (1992). Molecular cloning of DNA encoding a calmodulin-dependent phosphodiesterase enriched in striatum. Proc. Natl. Acad. Sci. USA 89(22), 11079-11083. Polli, J.W., Kincaid, R.L., Torris, J., & Billingsley, M.L. (1991). Expression of calmodulin-dependent enzymes in developing rat striatum is not affected by perturbation of dopaminergic systems. Synapse 9(2), 136-143. Potter, L.R. & Garbers, D.L. (1992). Dephosphorylation of the guanylyl cyclase-A receptor causes desensitization. J. Biol. Chem. 267(21), 14531-14534. Pyne, N.J., Cooper, M.E., & Houslay, M.D. (1986). Identification and characterization of both the cytosolic and particulate forms of cyclic GMP-stimulated cyclic AMP phosphodiesterase from rat liver. Biochem. J. 234, 325-334. Pyne, N.J., Cushley, W., Nimmo, H.G., & Houslay, M.D. (1989). Insulin stimulates the tyrosyl phosphorylation and activation of the 52 kDa peripheral plasma-membrane cyclic AMP phosphodiesterase in intact hepatocytes. Biochem. J. 261(3), 897-904. Qin, N. & Baehr, W. (1993). Expression of mouse rod photoreceptor cGMP phosphodiesterase gamma subunit in bacteria. FEBS Lett. 321(1), 6-10. • Qin, N., Pittler, S.J., & Baehr, W. (1992). In vitro isoprenylation and membrane association of mouse rod photoreceptor cGMP phosphodiesterase alpha and beta subunits expressed in bacteria. J. Biol. Chem. 267(12), 8458-8463. Radomski, M.W., Palmer, R.M., & Moncada, S. (1990a). Characterization of the L-arginine:nitric oxide pathway in human platelets. Br. J. Pharmacol. 101(2), 325-328. Radomski, M.W., Palmer, R.M., «fe Moncada, S. (1990b). Glucocorticoids inhibit the expression of an inducible, but not the constitutive, nitric oxide synthase in vascular endothelial cells. Proc. Natl. Acad. Sci. USA 87(24), 10043-10047.

Regulation of cNMP Synthesis and Deregulation

119

Reddy, R., Smith, D., et al. (1995). Voltage-sensitive adenylyl cyclase activity in cultured neurons. A calcium-independent phenomenon. J. Biol. Chem. 270(24), 14340-14346. Reifsnyder, D.H., Harrison, S.A., Macphee, C.H., McCormack, M.B., & Beavo, J.A. (1985). Antibody and inhibitor analysis of cGMP-inhibited phosphodiesterase in mutant S-49 cells and rat adipocytes. Fed. Proc. Fed. Am. Soc. Exp. Biol. 44, 1816. Repaske, D.R., Swinnen, J.V., Jin, S.L., Van, W.J.J., & Conti, M. (1992). A polymerase chain reaction strategy to identify and clone cyclic nucleotide phosphodiesterase cDNAs. Molecular cloning of the cDNA encoding the 63-kDa calmodulin-dependent phosphodiesterase. J. Biol. Chem. 267(26), 18683-18688. Robichon, A. (1991). A new cGMP phosphodiesterase isolated from bovine platelets is substrate for cAMP- and cGMP-dependent protein kinases: Evidence for a key role in the process of platelet activation. J. Cell. Biochem. 47(2), 147-157. Rossi, P., Giorgi, M., Geremia, R., & Kincaid, R.L. (1988). Testis-specific calmodulin-dependent phosphodiesterase. A distinct high affinity cAMP isoenzyme immunologically related to brain calmodulin-dependent cGMP phosphodiesterase. J. Biol. Chem. 263(30), 15521-15527. Sadler, S.E. (1991). Type III phosphodiesterase plays a necessary role in the growth-promoting actions of insulin, insulin-like growth factor-I, and Ha p2 Iras in Xenopus laevis oocytes. Mol. Endocrinol. 5(12), 1939-1946. Sass, P., Field, J., Nikawa, J., Toda, T., & Wigler, M. (1986). Cloning and characterization of the high-affinity cAMP phosphodiesterase of Saccharomyces cerevisiae. Proc. Nad. Acad. Sci. USA 83(24), 9303-9307. Schulz, S., Chrisman, T.D., & Garbers, D.L. (1992). Cloning and expression of guanylin. Its existence in various mammalian tissues. J. Biol. Chem. 267(23), 16019-16021. Schulz, S., Green, C.K., Yuen, P.S., & Garbers, D.L. (1990). Guanylyl cyclase is a heat-stable enterotoxin receptor. Cell 63(5), 941-948. Schulz, S., Singh, S., Bellet, R.A., Singh, G., Tubb, D.J., Chin, H., & Garbers, D.L. (1989). The primary structure of a plasma membrane guanylate cyclase demonstrates diversity within this new receptor family. Cell 58(6), 1155-1162. Sharma, R.K, (1991). Phosphorylation and characterization of bovine heart calmodulin-dependent phosphodiesterase. Biochemistry 30(24), 5963-5968. Sharma, R.K., Adachi, A.M., Adachi, K., & Wang, J.H. (1984). Demonstration of bovine brain calmodulin-dependent cyclic nucleotide phosphodiesterase isozymes by monoclonal antibodies. J. Biol. Chem. 259(14), 9248-9254. Sharma, R.K. & Wang, J.H. (1985). Differential regulation of bovine brain calmodulin-dependent cyclic nucleotide phosphodiesterase isoenzymes by cyclic AMP-dependent protein kinase and calmodulin-dependent phosphatase. Proc. Natl. Acad. Sci. USA 82(9), 2603-2607. Sharma, R.K. & Wang, J.H. (1986a). Calmodulin and Ca^'^-dependent phosphorylation and dephosphorylation of the 63-kDa subunit-containing bovine brain calmodulin-stimulated cyclic nucleotide phosphodiesterase isozyme. J. Biol. Chem. 261, 1322-1328. Sharma, R.K. & Wang, J.H. (1986b). Purification and characterization of bovine lung calmodulin-dependent cyclic nucleotide phosphodiesterase. An enzyme containing calmodulin as a subunit. J. Biol. Chem. 261, 14160-14166. Sharma, R.K. & Wang, J.H. (1986c). Regulation of cAMP concentration by calmodulin-dependent cyclic nucleotide phosphodiesterase. Biochem. Cell Biol. 64(11), 1072-1080. Shenolikar, S., Thompson, W.J., & Strada, S.J. (1985). Characterization of a Ca^^-calmodulin-stimulated cyclic GMP phosphodiesterase from bovine brain. Biochemistry 24(3), 672-678. Shibata, H. & Kono, T. (1990a). Cell-free stimulation of the insulin-sensitive cAMP phosphodiesterase by the joint actions of ATP and the soluble fraction from insulin-treated rat liver. Biochem. Biophys. Res. Commun. 170(2), 533-539.

120

J. KELLEY BENTLEY and JOSEPH A. BEAVO

Shibata, H. & Kono, T. (1990b). Stimulation of the insulin-sensitive cAMP phosphodiesterase by an ATP-dependent soluble factor from insulin-treated rat adipocytes. Biochem. Biophys. Res. Commun. 167(2), 614-620. Shyjan, A.W., de, S.F.J., Gillett, N.A., Goeddel, D.V., & Lowe, D.G. (1992). Molecular cloning of a retina-specific membrane guanylyl cyclase. Neuron 9(4), 727-737. Singh, S., Lowe, D.G., Thorpe, D.S., Rodriguez, H., Kuang, W.J., Dangott, L.J., Chinkers, M., Goeddel, D.V., & Garbers, D.L. (1988). Membrane guanylate cyclase is a cell-surface receptor with homology to protein kinases. Nature 334(6184), 708-712. Smith, C.J., Krall, J., Manganiello, V.C., & Movsesian, M.A. (1993). Cytosolic and sarcoplasmic reticulum-associated low Km, cGMP-inhibited cAMP phosphodiesterase in mammalian myocardium. Biochem. Biophys. Res. Commun. 190(2), 516-521. Smith, C.J. & Manganiello, V.C. (1989). Role of hormone-sensitive low Km cAMP phosphodiesterase in regulation of cAMP-dependent protein kinase and lipolysis in rat adipocytes. Mol. Pharmacol. 35(3), 381-386. Smith, C.J., Vasta, V., Degerman, E., Belfrage, P., & Manganiello, V.C. (1991). Hormone-sensitive cyclic GMP-inhibited cyclic AMP phosphodiesterase in rat adipocytes. Regulation of insulin- and cAMP-dependent activation by phosphorylation. J. Biol. Chem. 266(20), 13385-13390. Song, D.L., Kohse, K.P., & Murad, F. (1988). Brain natriuretic factor. Augmentation ofcellular cyclic GMP, activation of particulate guanylate cyclase and receptor binding. FEBS Lett. 232(1), 125-129. Sonnenburg, W.K. & Beavo, J.A. (1992). Cyclic GMP and regulation of cyclic nucleotide hydrolysis. In: Cyclic GMP: Synthesis, Metabolism, and Function (Murad, F., ed.). Academic Press, San Diego. Sonnenburg, W.K., Mullaney, P.J., & Beavo, J.A. (1991). Molecular cloning of a cyclic GMP-stimulated cyclic nucleotide phosphodiesterase cDNA. Identification and distribution of isozyme variants. J. Biol. Chem. 266(26), 17655-17661. Sonnenburg, W.K., Seger, D., & Beavo, J.A. (1993). Molecular cloning of a cDNA encoding the "61-kDa" calmodulin-stimulated cyclic nucleotide phosphodiesterase. Tissue-specific expression of structurally related isoforms. J. Biol. Chem. 268(1), 645-652. Sonnenburg, W.K., Seger, D., Koopman, W., & Beavo, J.A. (1992). Molecular cloning of a cDNA for the 59 kDa CaM-dependent phosphodiesterase from bovine lung. J. Biol. Chem. submitted. Stroop, S.D. & Beavo, J.A. (1991). Structure and function studies of the cGMP-stimulated phosphodiesterase. J. Biol. Chem. 266(35), 23802-23809. Stroop, S.D. & Beavo, J.A. (1992). Sequence homology and structure—^function studies of the bovine cyclic-GMP-stimulated and retinal phosphodiesterases. Adv. Second Messenger Phosphoprotein Res. 25, 55-71. Stroop, S.D., Charbonneau, H., & Beavo, J.A. (1989). Direct photolabeling of the cGMP-stimulated cyclic nucleotide phosphodiesterase. J. Biol. Chem. 264(23), 13718-13725. Suga, S., Nakao, K., Hosoda, K., Mukoyama, M., Ogawa, Y., Shirakami, G., Arai, H., Saito, Y., Kambayashi, Y., Inouye, K., & et. al. (1992). Receptor selectivity of natriuretic peptide family, atrial natriuretic peptide, brain natriuretic peptide, and C-type natriuretic peptide. Endocrinology 130(1), 229-239. Swinnen, J.V., Joseph, D.R., & Conti, M. (1989a). Molecular cloning of rat homologues of the Drosophila melanogaster dunce cAMP phosphodiesterase: Evidence for a family of genes. Proc. Natl. Acad. Sci. USA 86(14), 5325-5329. Swinnen, J.V., Joseph, D.R., & Conti, M. (1989b). The mRNA encoding a high-affinity cAMP phosphodiesterase is regulated by hormones and cAMP. Proc. Natl. Acad. Sci. USA 86(21), 8197-8201. Swinnen, J.V., Tsikalas, K.E., & Conti, M. (1991). Properties and hormonal regulation of two structurally related cAMP phosphodiesterases from the rat Sertoli cell. J. Biol. Chem. 266(27), 18370-18377.

Regulation of cNMP Synthesis and Deregulation

121

Taira, M., Meacci, E., Belfrage, P., & Manganiello, V.E. (1992). The cGMP-inhibited phosphodiesterase gene family: distinct cardiac and adipose tissue isoforms. In: 8th International Conference on Second Messengers and Phosphoproteins, Abstract C91T, Glasgow, Scotland. Takeda, S., Lin, C.T., Morgano, P.O., Mclntyre, S.J., & Dousa, T.P. (1991). High activity of lowMichaelis-Menten constant 3',5'-cyclic adenosine monophosphate-phosphodiesterase isozymes in renal inner medulla of mice with hereditary nephrogenic diabetes insipidus. Endocrinology 129(1), 287-294. Tanaka, T., Hockman, S., Moos, M.J., Taira, M., Meacci, E., Murashima, S., & Manganiello, V.C. (1991). Comparison of putative cGMP-binding regions in bovine brain and cardiac cGMP-stimulated phosphodiesterases. Second Messengers Phosphoproteins 13(2-3), 87—98. Tang, W.J. & Gilman, A.G. (1992). Adenylyl cyclases. Cell 70(6), 869-872. Tang, W.J. & Gilman, A.G. (1991). Type specific regulation of adenylyl cyclase by G protein py subunits. Science 254, 1500-1503. Tang, W.J., Krupinski, J., & Gilman, A.G. (1991). Expression and characterization of calmodulin-activated (type I) adenyly cyclase. J. Biol. Chem. 266(13), 8595-8603. Tanner, L.I., Harden, T.K., Wells, J.N., & Martin, M.W. (1986). Identification of the phosphodiesterase regulated by muscarinic cholinergic receptors of I32INI human astrocytoma cells. Mol. Pharmacol. 29, 455-460. Taussig, R. & Gilman, A.G. (1995). Mammalian membrane-bound adenylyl cyclases. J. Biol. Chem. 270(1), 1-4. Taussig, R., Quarmby, L.M., & Gilman, A.G. (1993). Regulation of purified type I and type II adenylylcyclases by G protein beta gamma subunits. J. Biol. Chem. 268(1), 9-12. Thomas, M.K., Francis, S.H., Beebe, S.J., Gettys, T.W., & Corbin, J.D. (1992). Partial mapping of cyclic nucleotide sites and studies of regulatory mechanisms of phosphodiesterases using cyclic nucleotide analogues. Adv. Second Messenger Phosphoprotein Res. 25, 45—53. Thomas, M.K., Francis, S.H., & Corbin, J.D. (1990a). Characterization of a purified bovine lung cGMP-binding cGMP phosphodiesterase. J. Biol. Chem. 265(25), 14964-14970. Thomas, M.K., Francis, S.H., & Corbin, J.D. (1990b). Substrate- and kinase-directed regulation of phosphorylation of a cGMP-binding phosphodiesterase by cGMP. J. Biol. Chem. 265(25), 14971-14978. Timmers, A.M., Newton, B.R., & Hauswirth, W.W. (1993). Synthesis and stability of retinal photoreceptor mRNAs are coordinately regulated during bovine fetal development. Exp. Eye Res. 56(3), 257-265. Trong, H.L., Beier, N., Sonnenburg, W.K., Stroop, S.D., Walsh, K.A., Beavo, J.A., & Charbonneau, H. (1990). Amino acid sequence of the cyclic GMP stimulated cyclic nucleotide phosphodiesterase from bovine heart. Biochemistry 29(44), 10280-10288. Vaandrager, A.B., Schulz, S., De, J.H.R., & Garbers, D.L. (1993). Guanylyl cyclase C is an N-linked glycoprotein receptor that accounts for multiple heat-stable enterotoxin-binding proteins in the intestine. J. Biol. Chem. 268(3), 2174-2179. Valtin, H., Coffey, A.K., O'Sullivan, D.J., Homma, S., & Dousa, T.P. (1990). Causes of the urinary concentrating defect in mice with nephrogenic diabetes insipidus. Physiol. Bohemoslov 39(1), 10^111. Verma, A., Hirsch, D.J., Glatt, C.E., Ronnett, G.V., & Snyder, S.H. (1993). Carbon monoxide: A putative neural messenger [see comments]. Science 259(5093), 381-384. Villacres, E.C., Wu, Z., et al. (1995). Developmentally expressed Ca(2+)-sensitive adenylyl cyclase activity is disrupted in the brains of type I adenylyl cyclase mutant mice. J. Biol. Chem. 270(24), 14352-14357. Wang, J.H., Sharma, R.K., & Mooibroek, M.J. (1990). Calmodulin-stimulated cyclic nucleotide phosphodiesterases. In: Cyclic Nucleotide Phosphodiesterases: Structure, Regulation and Drug Action (Beavo, J. & Houslay, M.D., eds.), pp. 19-60, John Wiley & Sons, Chichester. Wayman, G.A., Impey, S., et al. (1995). Ca^"*" inhibition of type III adenylyl cyclase in vivo. J. Biol. Chem. 270(37), 21480-21486.

122

J. KELLEY BENTLEY and JOSEPH A. BEAVO

Weishaar, R.E., Kobylarz, S.D., & Klinkefus, B.A. (1989). New mechanisms for positive inotropic agents: Focus on the discovery and development of imazodan. Cardiovasc. Drugs Ther. 3(1), 29-42. Welch, J.E., Swinnen, J.V., O'Brien, D.A., Eddy, E.M., & Conti, M. (1992). Unique adenosine 3',5' cyclic monophosphate phosphodiesterase messenger ribonucleic acids in rat spermatogenic cells: Evidence for differential gene expression during spermatogenesis. Biol. Reprod. 46(6), 10271033. Whalin, M.E., Scammell, J.G., Strada, S.J., & Thompson, W.J. (1991). Phosphodiesterase II, the cGMP-activatable cyclic nucleotide phosphodiesterase, regulates cyclic AMP metabolism in PC 12 cells. Mol. Pharmacol. 39(6), 711-717. Whalin, M.E., Strada, S.J., & Thompson, W.J. (1988). Purification and partial characterization of membrane-associated type II (cGMP-activatable) cyclic nucleotide phosphodiesterasefromrabbit brain. Biochim. Biophys. Acta 972( 1), 79-94. Whalin, M.W., Strada, S.J., Scammell, J.G., & Thompson, J.G. (1990). Regulation of cAMP metabolism in PC 12 cells by type II (cGMP-activatable) cyclic nucleotide phosphodiesterase. In: Purines in Cellular Signaling Targets for New Drugs (Jacobson, K.A., Daly, J.W., & Manganiello, V., eds.), pp. 323-328, Springer-Verlag, New York. Wiegand, R.C., Kato, J., & Currie, M.G. (1992). Rat guanylin cDNA: Characterization of the precursor of an endogenous activator of intestinal guanylate cyclase. Biochem. Biophys. Res. Commun. 185(3),812-«17. Wu, Z., Sharma, R.K., & Wang, J.H. (1992). Catalytic and regulatory properties of calmodulin-stimulated phosphodiesterase isozymes. Adv. Second Messenger Phosphoprotein Res. 25, 29-43. Wu, Z.L., Thomas, S.A., et al. (1995). Altered behavior and long-term potentiation in type I adenylyl cyclase mutant mice. Proc. Natl. Acad. Sci. USA 92(1), 220-224. Xia, Z., Choi, E.J., Wang, F., Blazynski, C, & Storm, D.R. (1993). Type I calmodulin-sensitive adenylyl cyclase is neural specific. J. Neurochem. 60(1), 305-311. Xia, Z., Choi, E. J., Wang, F., & Storm, D.R. (1992). The type III calcium/calmodulin-sensitive adenylyl cyclase is not specific to olfactory sensory neurons. Neurosci. Lett. 144(1-2), 169-173. Xia, Z.G., Refsdal, CD., Merchant, K.M., Dorsa, D.M., & Storm, D.R. (1991). Distribution of mRNA for the calmodulin-sensitive adenylate cyclase in rat brain: Expression in areas associated with learning and memory. Neuron 6(3), 431-443'. Yamaki, M., Mclntyre, S., Rassier, M.E., Schwartz, J.H., & Dousa, T.P. (1992). Cyclic 3',5'-nucleotide diesterases in dynamics of cAMP and cGMP in rat collecting duct cells. Am. J. Physiol. Yamamoto, T., Manganiello, V.C, & Vaughan, M. (1983). Purification and characterization of cyclic GMP-stimulated cyclic nucleotide phosphodiesterase from calf liver. J. Biol. Chem. 258, 1252612533. Yan, C, Bentley, J.K., Sonnenburg, W.K., & Beavo, J. A. (1993). Differential expression of the 61 kDa and 63 kDa calmodul in-dependent phosphodiesterases in the mouse brain. J. Neuroscience submitted. Yan, C, Zhao, A.Z., et al. (1995). Molecular cloning and characterization of a calmodul in-dependent phosphodiesterase enriched in olfactory sensory neurons. Proc. Natl. Acad. Sci. USA 92(21), 9677-9681. Yoshimura, M. & Cooper, D.M. (1992). Cloning and expression of a Ca(2+)-inhibitable adenylyl cyclase from NCB-20 cells. Proc. Natl. Acad. Sci. USA 89(15), 6716-^720. Yoshimura, M. & Cooper, D.M. (1993). Type-specific stimulation of adenylylcyclase by protein kinase C. J. Biol. Chem. 268(7), 4604^607. Yuen, P.S., Doolittle, L.K., et al. (1994). Dominant negative mutants of nitric oxide-sensitive guanylyl cyclase. J. Biol. Chem. 269(2), 791-793. Yuen, P.S., Potter, L.R., & Garbers, D.L. (1990). A new form of guanylyl cyclase is preferentially expressed in rat kidney. Biochemistry 29(49), 10872-10878.

Chapter 4

The Biological Functions of Protein Phosphorylation-Dephosphorylation TERRY A. WOODFORD, STEPHEN J. TAYLOR, and JACKIE D. CORBIN

Introduction Role of Phosphorylation/Dephosphorylation Historical Overview Physiological Relevance Second Messenger Systems and Protein Phosphorylation Mediation of Hormone Action cAMP-Dependent Protein Kinase cGMP-Dependent Protein Kinase Protein Kinase C Enzyme Regulation by Reversible Phosphorylation Glycogen Synthase and Phosphorylase b Pyruvate Dehydrogenase Acetyl-CoA Carboxylase Hydroxymethylglutaryl-CoA Reductase Tyrosine-3-Hydroxylase Regulation of Muscle Contraction Smooth Muscle Cardiac and Skeletal Muscle Protein Synthesis and Gene Transcription Regulation of Protein Synthesis Regulation of Gene Transcription

Principles of Medical Biology, Volume 4 Cell Chemistry and Physiology: Part If, pages 123-177 Copyright © 1996 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-806-4

123

124 124 125 128 128 128 131 133 134 136 137 139 140 141 142 143 143 145 146 146 147

124

TERRY A. WOODFORD ET AL

Cell Growth and Differentiation Cellular Transformation Retroviral Tyrosine Kinases and Their Cellular Homologs Growth Factor Receptor Tyrosine Kinases Neurological Function Mechanisms of Neural Transmission Neural Phosphoproteins: Ca /Calmodulin Kinase II Modulation of Ion Channel Activity Protein Phosphorylation in Vision Phosphoprotein Dephosphorylation Serine/Threonine Phosphatases Tyrosine Phosphatases Concluding Remarks

151 154 155 158 163 163 164 167 169 170 170 172 174

INTRODUCTION Role of Phosphorylation/Dephosphorylation

Protein phosphorylation is a major mechanism by which intracellular events in eukaryotic and prokaryotic cells are controlled by external stimuli. In eukaryotic cells, phosphoserine and phosphothreonine account for over 99% of all phosphorylated amino acids, while phosphotyrosine represents less than 0.03%. Minute amounts of phosphate may be covalently linked to lysine, arginine, histidine, aspartic acid, glutamic acid, and cysteine. Regulation of cellular functions by reversible phosphorylation involves signal transduction and amplification. In many cases, signal transduction is mediated by protein kinases, which catalyze the phosphotransferase reactions, and phosphoprotein phosphatases, which catalyze the cleavage of phosphate residues from the proteins. The steady-state level of phosphoprotein is determined by the kinase-phosphatase equilibrium: ATP ^

protein

^^ADP

phosphoprotein

Signal transduction also involves activators and inhibitors of kinases and phosphatases as well as other cellular regulators which modulate the signal pathway. Amplification of the cellular signal results from small molecule and protein interaction at various levels to form an integrated relay network. That distinct enzyme pathways exist for phosphorylation and dephosphorylation is an important feature of this dynamic scheme as it allows for greater range of control by separate modifiers and for greater amplification.

Function of Protein Phosphorylation-Depbosptiorylation

125

Most kinases phosphorylate a limited number of proteins in vitro at a physiological rate. Some of the most common protein substrates are the kinases themselves, and the reaction of the kinase catalyzing its own phosphorylation, is referred to as autophosphorylation. This process may reflect the fact that the autophosphorylation sites, or sites with similar structure, are located in the regulatory domains of kinases and act as competitive inhibitors of the protein kinase catalytic domains. In instances in which the residue which would be phosphorylated is missing or altered, but other essential elements of the autophosphorylation site are present, the inhibitory domain is called the "pseudosubstrate" site. It is possible that some protein kinases for which natural substrates have not yet been found catalyze autophosphorylation exclusively, and the phosphate introduced into the kinase causes some as yet unknown functional alteration in the enzyme other than increased catalysis of protein phosphorylation. Among specific substrate proteins other than the kinases themselves, only a few out of a large number of potentially available serine, threonine, or tyrosine residues become phosphorylated, due to the inherent structure and chemistry of the substrate protein itself and to the ability of the kinase to recognize specific phosphorylation site(s). Phosphorylation of a protein may induce subtle changes in its structure by introduction of a negative charge(s). Protein may be phosphorylated at single or multiple sites. Multi-site phosphorylation is a mechanism for increasing the regulatory potential of enzymes or structural and regulatory proteins. Phosphorylation at one site may amplify or antagonize the rate and effects of phosphorylation at other sites. Phosphorylation at different sites on a protein by different protein kinases also allows the cell to respond to several physiological stimuli. It is also likely that in some cases phosphorylation at more than a single site is required for a particular conformational change to occur in the protein substrate. Phosphorylation of a protein may increase or inhibit its activity. Alternatively, a protein may be modified to an intermediate functional state between active and inactive. Therefore, modification of proteins by phosphorylation can be either mediatory or modulatory. Historical Overview The presence of phosphate in proteins has been known for almost 100 years, but its importance has only been realized since the discovery of enzyme regulation by reversible phosphorylation. This type of modulation of protein function was first recognized about 35 years ago with the finding that neural and hormonal control of glycogen metabolism is mediated by changes in the phosphorylation state of three enzymes; glycogen phosphorylase, phosphorylase kinase, and glycogen synthase. The idea that a similar type of bioregulation occurs in other processes besides glycogen metabolism was supported by the identification of cyclic AMP (cAMP)-dependent protein kinase. Most effects of cyclic AMP in eukaryotic systems are mediated by protein phosphorylation reactions catalyzed by this

126

TERRY A. WOODFORD ET AL

enzyme family. The regulatory role of Ca^"^, phospholipid and other molecules in modulating kinases and phosphatases involved in signal transduction is also now apparent. Some of the highlights and discoveries in the development of the field of protein phosphorylation/dephosphorylation are noted below: 1955 1957 1968 1969 1970 1978 1979 1980 1984

Interconversion of phosphorylase by phosphorylation. Cyclic AMP as a second messenger in hormone action. Cyclic AMP-dependent protein kinase as a cellular mediator of cAMP effects. Phosphorylation of the pyruvate dehydrogenase complex as the first example of regulation of a mitochondrial enzyme not directly involved in carbohydrate metabolism. Calmodulin as a mediator of Ca^^ effects and the involvement of Ca^Vcalmodulin in the activation of Ca^'^-dependent protein kinases. The product of the src gene of the oncogenic Rous sarcoma virus as a tyrosine-specific protein kinase. Identification of protein kinase C as a receptor for the tumor-promoting phorbol esters and its activation by diacylglycerol, Ca^"^ and phospholipid. Epidermal growth factor receptor has tyrosine kinase activity triggered by binding of ligand. Cyclic AMP response elements and DNA-binding proteins are involved in regulation of gene transcription.

During the past 20 years, the role of protein phosphorylation-dephosphorylation as a general regulatory mechanism has become apparent. It has been estimated that approximately one out of six cellular proteins is a phosphoprotein and that there are more than 1,000 protein kinases that catalyze phosphorylation reactions in mammals. Although it is not absolutely clear why nature has selected protein phosphorylation as a major regulatory mechanism, one might speculate that the availability of phosphate and ATP, and the presence of proteins which could be readily converted into protein kinases by evolutionary processes, are contributory reasons. The potential for protein conformational changes induced by the strong negative charge introduced by phosphate groups may have been an additional impetus. The simplicity of protein kinase substrate-recognition sites could have facilitated the proliferation of protein phosphorylation. Figure 1 illustrates a few of the systems modified or regulated by phosphorylation in a typical cell. It is clear that phospho-dephospho regulation is involved in carbohydrate, amino acid, lipid, and steroid metabolism, as well as in processes such as gene expression, protein synthesis, cell-cycle control and chromosome condensation. Contractile and membrane proteins, ribosomal, nuclear and cytoskeletal proteins, and proteins of other functional classes such as ion channels and hormone neurotransmitter receptors

127

Function of Protein Phosptiorylation-Depliosplioryiation Cytosolic Enzymes

PFK-FDPase Pyruvate Kinase Phenylalanine Hydroxylase Acetyl CoA Carboxylase

Na/K ATPase Other Transport Systems

Cytoskeleton

Microtubule-Assoc. Proteins Desmin Vimentin Vinculin

Nucleus

HIstones Non-hlstone Proteins (HMG's)

Glycogen Particles Phosphorylase Phosphorylase Kinase Protein Kinase Glycogen Synthase

Rough Endoplasmic Reticulum and Ribosomes

36 Ribosomal Proteins Initiation Factors

Figure 1.

Mitochondrion

Pyruvate Dehydrogenase

Systems modified or regulated by phosphorylation in a typical cell.

undergo phosphorylation-dephosphorylation. Specialized cells possess unique phosphoproteins. In neurons, tyrosine and tryptophan hydroxylases as well as myelin basic protein and synapsin are phosphorylated. In muscle, troponin I and troponin T are phosphorylated, together with the light chain of myosin and components of the sarcoplasmic reticulum implicated in Cd?^ transport. Protein phosphorylation is involved in the binding of divalent cations, e.g., the clustering of phosphoserine residues in phosphovitin, an egg yolk protein, and casein, a milk protein, are likely to be important in the binding of iron and calcium, respectively. Phosphoproteins play a role in the process of bone formation as well as in the molecular pathway of learning. In lower organisms such as yeast and bacteria, phospho-dephospho regulation of chemotaxis, cell division, and gene transcription has been documented. The discovery that some tumor virus-transforming factors and cell-surface growth factor receptors possess protein kinase activity implicates protein phosphorylation in viral infection, cell growth control and cancer. The state of phosphorylation of various cellular proteins indeed influences cell structure and function, as well as the rate of cell growth and state of differentiation and the ability of cells to respond to extracellular stimuli such as hormones, neurotransmitters, and growth factors. It is the purpose of this chapter to describe specific functions in eukaryotic cells which are regulated by phosphorylation-dephosphorylation. Obviously, these areas of study encompass much detail that cannot be covered in this chapter. Therefore, the intent is to provide the reader with a general overview and appreciation of the diverse array of cellular functions that are regulated by phosphorylation-dephos-

128

TERRY A. WOODFORD ET AL

phorylation and to illustrate the way in which bioregulation of these functions is precise and well-coordinated. Physiological Relevance Four criteria have been described by Krebs and Beavo (1979) for establishing that a protein phosphorylation event is physiologically relevant. These include demonstration of the following: 1. The purified enzyme or protein can be phosphorylated stoichiometrically at a significant rate in a reaction(s) catalyzed by an appropriate kinase(s) and dephosphorylated by a phosphatase(s). 2. The functional properties of the enzyme or protein undergo meaningful changes that correlate with the degree of phosphorylation. 3. The enzyme or protein can be modified in vivo or in an intact cell system with accompanying functional changes. 4. Cellular concentrations of kinase and/or phosphatase effectors correlate with the extent of phosphorylation-dephosphorylation of the enzyme or protein. Very few phosphorylation events actually meet the requirements described above. This is hampered by the occurrence of multi-site phosphorylations by multiple enzymes, some with overlapping substrate specificities. In addition, the role of "silenf phosphorylation sites which do not appear to influence protein function, is uncertain. These modifications may have important, albeit as yet, unappreciated roles, such as regulating protein turnover, controlling protein-protein interaction, or specifying subcellular localization. Despite these shortcomings, extensive progress in unraveling subcellular mechanisms that regulate cellular functions by phosphorylation/dephosphorylation has been made in several areas including cell physiology, biochemistry, and biophysics.

SECOND MESSENGER SYSTEMS AND PROTEIN PHOSPHORYLATION Mediation of Hormone Action Adenosine 3',5'-monophosphate (cyclic AMP, cAMP) was discovered in biological systems during the course of investigations on the regulation of glycogen breakdown in liver by epinephrine and glucagon. The role of cAMP as an intracellular mediator of the actions of a wide variety of amine and polypeptide hormones is well-established. The "second messenger" concept has thus evolved from studies in which the hormone in question alters the intracellular concentration of cAMP,

129

Function of Protein Ptiosphorylation-Dephosphorylation

which in turn modifies the rate of a variety of ongoing cellular processes. In mammalian cells, cAMP is generated in response to ligand-receptor interaction at the cell surface. This signal is transduced by binding of cAMP to the regulatory (R) subunit of cAMP-dependent protein kinase (cAMP kinase). Most of the biological effects of cAMP are thought to be caused by the phosphorylation of specific substrates by the catalytic (C) subunit of cAMP kinase. A partial list of hormones that act through the cAMP system are listed in Table 1. Adenylate cyclase (AC), which catalyzes the formation of cAMP from ATP, is an intrinsic membrane protein that is coupled to a receptor protein which spans the membrane via a stimulatory guanine nucleotide-binding regulatory protein (Gs) at the inner surface of the plasma membrane (Figure 2). Binding of a hormone agonist to its specific receptor leads to a conformational change in the receptor molecule and subsequent interaction of the agonist-activated receptor with the heterotrimeric Gs holoprotein. The activated Gsa subunit then interacts with AC to stimulate cAMP formation. Among agonists, catecholamines such as epinephrine and norepinephrine are of special interest since they produce a variety of effects in mammalian tissues. Catecholamines interact with at least two different types of receptors, a- and p-adrenergic receptors. The adrenergic P-receptor to be discussed herein is an integral component of the AC system in liver, spleen, kidney, brain, muscle, adipose tissue, and pancreatic islets. One signaling pathway that involves both hormonal and neuronal stimulation is the regulation of glycogenolysis. Glycogenolysis is triggered in muscle by two extracellular signals, the hormone epinephrine and depolarization of the plasma Table 1, Polypeptide and Catecholamine Hormones that Function through cAMP Tissue Skin Bone Muscle Fat

Brain Thyroid Heart Liver Kidney Adrenal Ovary

Hormone Melanocyte-stimulating hormone Parathyroid hormone Epinephrine Epinephrine Adrenocorticotrophic hormone Glucagon Norepinephrine Thyroid-stimulating hormone Epinephrine Epinephrine Parathyroid hormone Vasopressin Adrenocorticotrophic hormone Luteinizing hormone Follicle stimulating hormone

Principal

Response

Darkening Calcium resorption Glycogenolysis Lipolysis Lipolysis Lipolysis Discharge of Purkinje cells Thyroxin secretion Increased heart rate and force of contraction Glycogenolysis Phosphate excretion Water reabsorption Hydrocortisone secretion Progesterone secretion

TERRY A. WOODFORD ET AL.

130 PLASMA. I

I MIBfTOfCY NOMIONC

-0

STMUtATONV

ATP

*9SSf^V^ W-Sffl^

I PMOSPMATASCH

>llfSPONSC

U

^MiP-

H^ccV"

*-

'^«Fa| PflOTCIN • ATP

Figure 2, Relationships among enzymes involved in cAMP action.

membrane upon acetylcholine release at the neuromuscular junction. Epinephrine binds to the P-adrenergic receptor and initiates a complex reaction sequence in the plasma membrane that leads to the activation of AC. This causes an increase in the concentration of the intracellular-signaling molecule, cAMP. Cyclic AMP binding to cAMP kinase causes the release of free C subunit to phosphorylate a variety of proteins. The first well-characterized substrate for this protein kinase was phosphorylase kinase. After phosphorylation and activation of phosphorylase kinase takes place, this enzyme then catalyzes phosphorylation of phosphorylase b. This converts inactive glycogen phosphorylase b to the active a form, thereby triggering glycogen degradation. The product of glycogen degradation, glucose-1-P, enters glycolysis to fulfill energy requirements of the cell. Alternatively, glycogenolysis is stimulated during muscle contraction, which occurs upon depolarization of the sarcolemma. Ca^"*" release from the sarcoplasmic reticulum through depolarization is probably triggered by phosphatidylinositol breakdown to generate the signal inositol 1,4,5-trisphosphate, IP3. Ca^"^ binds to a receptor, calmodulin, which is a subunit of phosphorylase kinase. This process activates phosphorylase kinase which then triggers the phosphorylase b io a conversion and thereby glycogen degradation. Thus, two entirely different signaling mechanisms, electrical stimulation and a P-agonist, can trigger the same pathway.

Function of Protein Phosphorylation-Depliosphorylation

131

cAMP-Dependent Protein Kinase Since the discovery of cAMP and cAMP-dependent protein kinase, the field of second messengers and protein phosphorylation has grown rapidly. The widely distributed serine/threonine-specific cAMP kinase plays an important role in many metabolic processes and has been a prototype for studies of other kinases. A partial list of substrates for cAMP kinase is shown in Table 2. All mammalian holoenzymes are inactive tetramers composed of two regulatory (R) and two catalytic (C) subunits. The differences in the two major classes of holoenzymes, designated types I and II, lie in their regulatory components, Ri and Rn, respectively. R subunits range in size from 42-45,000 Da and the C subunit is -40,000 Da daltons. Recently, two isoforms for Ri and Rn, a and P, respectively, have been cloned. Three genes for the C subunit have been cloned and designated Ca, Cp, and Cy. This multigene enzyme family shows significant heterogeneity. The P isoforms are localized primarily in the brain, central nervous system, and reproductive cells, suggesting a specialized function in these tissues. The different R subunits apparently do not form heterodimers, but each R readily associates with any of the C subunits Assuming that two different C subunits could be contained in a single tetrameric holoenzyme (e.g., (Ria)2CaCP), it is possible that at least 24 different forms of mammalian holoenzymes exist. The R subunits consist of an amino-terminal dimerization domain and a carboxyl-terminal cAMP-binding domain separated by a region known as the inhibiTable 2,

Substrates o f c A M P - D e p e n d e n t Protein Kinase

Physiological Phosphorylase kinase Hormone-sensitive lipase Pyruvate kinase Glycogen synthase Cholesterol ester hydrolase Acetyl-CoA carboxylase Phosphatase inhibitor 1 Tyrosine hydroxylase ATP-citrate lyase Troponin I Phospholamban Phenylalanine hydroxylase P-receptor (erythrocyte) cAK type 11 R subunit Ca^"^ channel protein 6-phosphofructo-2-kinase

Potentially

Physiological

Chromatin Fructose-2,6-bisphosphatase Myelin basic protein S-6 ribosomal protein Reverse transcriptase Phosphofructokinase Protamine Diglyceride lipase Fructose diphosphatase Tryptophan hydroxylase pp60^'^^ Atrial natriuretic factor Calmodu I in-dependent phosphodiesterase Low K^ phosphodiesterase CREB Methyltransferase

132

TERRY A. WOODFORD ET AL.

tory domain. Each R subunit contains two different types of cAMP-binding sites which differ in cAMP analog specificity and rate of cAMP dissociation. Upon cooperative binding of two molecules of cAMP to each R subunit, dissociation of the holoenzyme occurs and the released C subunits catalyze phosphorylation of substrate proteins: R2C2 + 4cAMP^R2CAMP4 + 2 C The concentration of cAMP in the cell determines the level of activation of cAMP kinase, and is therefore tightly regulated. Hormones acting at their cell surface receptors affect the synthesis of cAMP by either stimulating or inhibiting adenylate cyclase through stimulatory or inhibitory GTP-binding proteins. There are several types of phosphodiesterase (PDE) responsible for degradation of c AMP to 5'-AMP, which are themselves regulated by various effectors such as Ca^"^, calmodulin, and cGMP. One feedback mechanism for regulating cAMP kinase involves activation of a high-affinity PDE, presumably by phosphorylation. Thus, cAMP kinase may regulate intracellular cAMP levels and, consequently, its own level of activity. In the presence of MgATP, the C subunit phosphorylates substrates containing two basic residues (usually arginine) on the amino-terminal side of the phosphorylated residue (usually serine). Thus, the preferred phosphorylation site is Arg-Arg-XSer-X, with X usually being a hydrophobic residue. Other kinases demonstrate different phosphorylation site preferences as shown in Table 3. Rn is autophosphorylated by the C subunit on a serine residue (Ser 95 in bovine Rna) which reduces the interaction between R and C subunits. Ri has a pseudosubstrate site in the same region as the autophosphorylation site of Rn- These regions are thought to inhibit the C subunit function by competing with physiological substrates thereby rendering the enzyme inactive (Corbin et al., 1978). Binding of cAMP to the R subunits relieves this inhibition by causing a conformational change. A specific-heat stable endogenous inhibitor of cAMP kinase called protein kinase inhibitor (PKI) is present in many tissues and also functions as a pseudosubstrate. Interestingly, phosphorylation of a fragment of PKI containing residues 5—22 on tyrosine by the EGF receptor tyrosine kinase can decrease its inhibitory potency. An evolutionary relationship exists among all of the cAMP- and cGMP-dependent protein kinases.

Table 3.

Phosphorylation Site Recognition Sequences of Four Different Serine/Threonine Protein Kinases

Glycogen synthase kinase 3 cAMP-dependent protein kinase Ca^Vcalmodulin-dependent protein kinase casein kinase II Note: *indicates phosphorylated residue.

Ser(Thr)-X-X-Ser /Thr Arg(Lys)-Arg-X-Ser/Thr Arg-X-X-Ser/Thr Ser(Thr)-Glu-Ser /Thr -Glu(Asp)

Function of Protein Phosphorylation-Dephosphorylation

133

The cyclic nucleotide-binding sites are believed to be derived from a common evolutionary precursor, which is also thought to be the ancestor of the bacterial cAMP-binding protein, catabolite gene activator protein (CAP). cGMP-Dependent Protein Kinase

cGMP-dependent protein kinase (cGMP kinase) is a serine/threonine-specific enzyme found in a variety of mammalian tissues, although at lower levels than cAMP kinase. cGMP kinase is relatively abundant in lung, cerebellum, and smooth-muscle tissue. The bovine lung kinase is a dimer of two identical 76,000 Da subunits. Unlike cAMP kinase, which has separate R and C subunits, the R and C components of cGMP kinase reside on the same polypeptide chain. Again, in the absence of cGMP, the regulatory domain inhibits catalytic activity. When cGMP binds to the enzyme, which also has two distinct cGMP-binding sites per subunit differing in kinetic properties and cyclic nucleotide analog specificities, a conformational change occurs and inhibition is released: GK + 4 cGMP ^ GK • CGMP4 In a manner analogous to the cAMP system, the levels of intracellular cGMP are regulated at the level of synthesis by soluble or particulate guanylate cyclases and at the level of degradation by phosphodiesterases. cGMP kinase undergoes autophosphorylation, and the region containing the autophosphorylation sites is also proposed to inhibit the catalytic activity through a competitive mechanism. In the presence of MgATP, the catalytic domain possesses phosphotransferase activity. Multiple isozymic forms of cGMP kinase exist. A unique form of the enzyme (type II) which occurs in intestinal brushborder appears to be an 86,000 Da monomer anchored to the membranes of microvilli by a 15,000 Da protein. cGMP-dependent protein phosphorylation catalyzed by this enzyme may be important in the control of membrane ion permeability in intestinal microvilli. Two forms of the enzyme have been isolated from bovine aorta and designated types l a and ip (Wolfe et al, 1989). Type ip possesses an overall structure and peptide substrate specificity which is similar to that of la, but it has different cGMPbinding behavior and is autophosphorylated almost entirely on serine instead of on both serine and threonine residues. The cDNAs for types l a and ip have recently been cloned. A combination of the cDNA and amino-acid sequence analysis suggests that types l a and ip are produced by alternative splicing of an mRNA from a single gene to generate different amino termini and a common catalytic domain. Recent studies support a role for cGMP and cGMP kinase in the relaxation of smooth muscle in response to agents such as atrial natriuretic factor (ANF), endothelial derived relaxing factor (EDRF), and synthetic nitrovasodilators such as nitroglycerin. Although the substrates for cGMP kinase involved in the process

134

TERRY A. WOODFORD ET AL.

of smooth-muscle relaxation have not been identified, proteins directly or indirectly involved in regulating Ca^"^ levels are good candidates. Cyclic AMP and cGMp kinases have similar substrate specificities, and most substrates are phosphorylated in vitro up to 20 times better by cAMP kinase, which is also more abundant in animal tissues. As compared to cAMP kinase, cyclic GMP kinase has increased selectivity towards substrates that have an additional basic residue on the carboxylterminal side of the phosphorylated residue. One specific substrate for cGMP kinase is a cerebellar protein called G-substrate, which has a structural similarity to phosphatase inhibitor 1. Also a cGMP-binding phosphodiesterasefromlung tissue is preferentially phosphorylated by cGMP kinase. Protein Kinase C

Protein kinase C (PKC) is a Ca^'^/phospholipid-dependent, multi-functional serine/threonine protein kinase. 1,2-Diacylglycerol (DAG), a product of agoniststimulated phosphatidylinositol 4,5-bisphosphate (PIP2) hydrolysis, activates PKC by increasing its affinity for Ca^^. The other product of PIP2 breakdown, inositol 1,4,5-trisphosphate (IP3), mobilizes Ca^"*"fromintracellular stores. PKC activation represents one branch of a complex signal-transduction pathway utilized by a wide variety of hormones, neurotransmitters, and growth factors (Figure 3). Tumorpromoting phorbol esters act as diacylglycerol analogs to activate PKC and concomitantly cause a translocation of PKC activity from the cytosol to membranes. Prolonged exposure of cells to phorbol esters, such as tetradecanoyl-phorbol-13acetate (TPA), leads to a gradual loss or down-regulation of intracellular PKC activity. The fact that PKC mediates most, if not all, of the actions of the phorbol esters and that these agents are potent tumor promoters implicate a role for PKC in cell growth and proliferation. PKC represents a family of related isozymes. Seven different PKC isozymes have been identified and isolated thus far by screening of a variety of mammalian cDNA libraries (Nishizuka, 1988). Four of these, a, (JI and pil (generated by alternate splicing), and y, have been characterized in terms of tissue distribution. Northern blot analysis, in situ mRNA hybridization, and immunochemical techniques have demonstrated that the y isozyme is specific to certain areas of the brain and spinal cord, implying potential specific neuralftmctionsof this isozyme. The a, pi, and pil isozymes are expressed more ubiquitously, though their ratios vary between cell types and subcellular compartments. The a, pi, pil, and y isozymes share a common structure consisting of a regulatory domain, presumably containing the Ca^^, phospholipid and diacylglycerol binding site(s), and a catalytic domain, which shares regions of homology with other kinases. The regulatory domain, which acts as a pseudosubstrate inhibitor for catalytic activity similar to that occurring in the cyclic nucleotide-regulated kinases, contains a repeated cysteine-rich sequence, characteristic of the "zinc-fin-

Function of Protein Phosphorylation-Dephosphorylation

135

hormones neurotransmitters growth factors

i

Figure 3. The role of protein kinase C in the signal transduction mechanisms of hormones and growth factors. The abbreviations represent the following: PLC, phosphoinositide or phosphatidylcholine-specific phospholipase C; PC, phosphatidylcholine; PIP2, phosphatidyllnosltol- 4,5-bisphosphate; DAG, 1,2-sn-diacylglycerol; PKC, protein kinase C; EGF, epidermal growth factor; RTK, receptor tyrosine kinase; IP3, inositol-1,4,5-trisphosphate; Ca^VCaM-kinases, specific and multi-functional calcium/calmodulin-dependent protein kinases.

ger" domain of a number of DNA-binding proteins, although PKC has not been shown to associate with DNA. All four isozymes require phospholipid (particularly phosphatidylserine) for activity and are activated by Ca^"*" and diacylgycerol. However, unlike a and y, the pi and pil species are substantially activated by DAG in the absence of Ca^"^, suggesting the possibility of cellular activation of this isozyme in the absence of Ca^"^ mobilization. The y and, to a lesser extent, a species are activated by low concentrations of arachidonic acid, a product of agonist-stimulated phospholipase A2 activity. The potential role of lipids and lipid metabolites other than phosphatidylserine (PS) and DAG in PKC function has become recognized in recent years. In addition to unsaturated fatty acids such as arachidonate and oleate, lysophospholipids, the other products of phospholipase A2 action, have been shown to regulate PKC activity. Products of sphingolipid metabolism, sphingosine and lysosphingolipids, inhibit PKC and may play a role in the negative control of PKC activity and PKC-mediated processes by certain extracellular signals. In many cell types, agonist-stimulated DAG formation is derived primarily

TERRY A. WOODFORD ET AL.

136

from hydrolysis of phosphatidylcholine by activation of phospholipase C or, indirectly, by phospholipase D. Since phosphatidylcholine breakdown does not release IP3, this mechanism may permit PKC activation in the absence of Ca^"^ mobilization (Exton, 1990). PKC plays a crucial role in the feedback-inhibition of signal transduction by hormones, neurotransmitters, and growth factors that act via G protein-mediated PlP2-phospholipase C activation, and by growth factors that act via receptor tyrosine kinases (see Figure 3). PKC phosphorylates a number of protein substrates and may regulate a wide variety of cellular functions including secretion, ion transport, metabolism, gene expression, smooth-muscle contraction, and cell proliferation. The role of PKC in activation of these cellular processes via phosphorylation of certain key proteins represents an example of a classical second-messenger cascade system.

ENZYME REGULATION BY REVERSIBLE PHOSPHORYLATION The first thoroughly characterized enzyme regulated by covalent modification was glycogen phosphorylase, which is phosphorylated by phosphorylase kinase. Numerous enzymes are now known to be regulated by phosphorylation-dephosphorylation (Table 4). Generally, enzymes involved in biodegradative pathways are activated by phosphorylation. Although some enzymes in their native conformation contain only a single phosphorylation site, e.g., glycogen phosphorylase and pyruvate kinase, many enzymes contain multiple phosphorylation sites. Included in the latter category are glycogen synthase, phosphorylase kinase, pyruvate dehydrogenase, and acetyl-CoA carboxylase. Enzyme activity is generally control-

Table 4. Enzymes which are Regulated by Phosphorylation Enzyme

Metabolic Pathway

Glycogen phosphorylase Phosphorylase kinase Triglyceride lipase Fructose 1,6-biphosphatase Cholesterol ester hydrolase Tyrosine hydroxylase Phenylalanine hydroxylase

Glycogenolysis Glycogenolysis Lipolysis Glycolysis/gluconeogenesis Cholesterol ester hydrolysis Catecholamine synthesis Aromatic amino acid

Glycogen synthase Pyruvic dehydrogenase Hydroxymethylglutaryl CoA Acetyl-CoA carboxylase Pyruvate kinase Phosphofructokinase elF2-kinase

Glycogen synthesis Glycolysis Cholesterol synthesis Fatty acid synthesis Glycolysis/gluconeogenesis Glycolysis/gluconeogenesis Protein synthesis

Activated (A) or Inactivated (1)

A A A A A A A

Function of Protein Phosphorylation-Dephosphorylation

137

led by modification of a few primary sites. Secondary sites appear to alter additional kinetic properties and to modify the rates of phosphorylation-dephosphorylation of the primary sites. For enzyme phosphorylation-dephosphorylation to serve a regulatory function, the protein kinases and/or phosphoprotein phosphatases involved must also be regulated. Kinases are often regulated by specific effectors or messengers such as cyclic nucleotides, Ca^"^, phospholipid, or double-stranded RNA. Phosphatases are often controlled by "substrate-directed effects," in which a given metabolite or metal ion combines with the phosphoenzyme substrate, causing it to be either a more or less effective substrate for the phosphatase. Cellular functions can be coordinately controlled, since different metabolic pathways can be regulated by the same kinases and phosphatases. The activation and inactivation of glycogen phosphorylase and synthase, respectively, allows coordinate control of glycogen breakdown and synthesis in response to epinephrine. The activation and inactivation of triglyceride lipase and glycerol phosphate acyltransferase allow coordinate control oftriglyceride breakdown and synthesis by epinephrine. The inactivation of both glycogen synthase and acetyl-CoA carboxylase allows coordinate control of glycogen and fatty-acid synthesis in response to glucagon. Enzymes regulated by phosphorylation-dephosphorylation are also regulated by a variety of allosteric effectors. This emphasizes their response to alterations in the metabolic state of the cell in addition to physiological stimuli. A close relationship exists between covalent and noncovalent control of enzyme activity. Allosteric effectors alter the activity of enzymes and often affect the rate at which they are modified by phosphorylation or dephosphorylation. Similarly, phosphorylation often changes the affinity of an enzyme for a substrate, an activator or inhibitor, or combinations thereof Thus, amplification or suppression of the effects of phosphorylation are "fine-tuned" by the concentration of substrates or effector molecules in the cell. As previously mentioned, phosphorylation-dephosphorylation is a mechanism for interconverting an enzyme between two (or more than two in the case of multi-site phosphorylations) different states which respond differently to substrates and effectors. In this section, the relationships between single- or multiple-site phosphorylation and enzyme acfivation or inactivation will be discussed, using various metabolic enzymes as examples. The different regulatory systems described in this section will show some striking similarities and will illustrate the role of reversible phosphorylation in the enzymafic control of cellular fimction, including intermediary metabolism. Glycogen Synthase and Phosphorylase b

Glycogen synthase (GS) has been purified and characterized from several issues including skeletal muscle, liver, heart, and adipose tissue. Skeletal-muscle GS exists as a tetramer of identical 90,000 Da subunits. The enzyme is typically bound

138

TERRY A. WOODFORD ET AL.

to glycogen particles in association with phosphorylase and other glycogen-metabolizing enzymes in the cell. Important physiological inhibitors of GS include ATP, ADP, and AMP. Each subunit contains several phosphorylation sites so that multiple forms of the synthase may exist. The nonphosphorylated form is referred to as the a or I form, whereas the phosphorylated forms are designated bovD forms. Increased phosphorylation causes inactivation of the synthase by decreasing its affinity for the substrate UDP-glucose and increasing its dependency on glucose6-P, an allosteric activator. Phosphorylation of GS inhibits its activity when assayed in the absence of glucose-6-P. Phosphorylation of GS is a complex reaction involving multiple (up to seven) sites per subunit and multiple protein kinases. cAMP kinase readily catalyzes partial phosphorylation and inactivation of GS, as does skeletal muscle phosphorylase kinase. Glycogen synthase kinases (GSK) 3, 4, and 5 also phosphorylate GS, and phosphorylation of sites 3a,b,c by GSK3 influences kinetic parameters more than does phosphorylation of other sites, although effects of the various kinases are additive. Glycogen phosphorylase b, which mediates glycogen breakdown, is activated via phosphorylation that is catalyzed by phosphorylase kinase, which in turn is activated the cAMP kinase or by Ca^"^ ions. The cascade of enzymes involved in glycogenolysis due to hormonal stimulation is summarized in Figure 4. Skeletalmuscle phosphorylase kinase has a subunit structure (a,P,y,5)4. The a and P subunits are substrates for cAMP kinase-mediated phosphorylation, the y subunit is the catalytic moiety, and the 8 subunit is calmodulin (CaM). Phosphorylation of a particular serine residue in the P-subunit by c AMP kinase activates phosphorylase kinase by increasing its affinity for Ca^^. Interconversion of GS and phosphorylase b between various forms is under hormonal control in many tissues. It is well-established that hormones such as epinephrine in muscle, heart, and adipose tissue and glucagon in the liver act through cAMP kinase to promote inactivation of GS and coordinate activation of phosphorylase. In these same tissues, insulin treatment results in GS dephosphorylation and activation to promote glycogen deposition in response to hyperglycemia. Phosphatase 1 dephosphorylates GS, glycogen phosphorylase, and phosphorylase kinase. Interestingly, although the phosphatase has a similar affinity for GS containing 1, 2, or 3 phosphates, the apparent reaction velocity differs for these three isoforms. In diabetic rats, analysis of GS from adipose tissue suggests that the enzyme is in a highly phosphorylated form. The decrease in the reaction velocity of phosphatase 1 with highly phosphorylated forms of GS may be important relative to control of synthase bio a activation and could account for a decrease in the rate of synthase bio a conversion in diabetic tissues. Insulin may be required to maintain a high level of phosphatase 1. In diabetic tissues, partial loss of phosphatase would shift the synthase interconversion equilibrium towards the highly phosphorylated synthase, which is a poor substrate for the phosphatase. Two phosphatase inhibitors, inhibitor 1 and 2, have been identified. Inhibitor 1 is active only after phosphory-

Function of Protein Phosphorylation-Dephosphorylation

139

STIMULATORY HORMONE (ACTIVATION) AOENYUTE CYCLASE Mg-f-2 ATP

CAMP

INACTIVE cAK

INACTIVE PHOSPHORYLASE b KINASE

ACTIVE PHOSPHORYLASE b KINASE

PHOSPHORYLASE b

GLYCOGEN + P I

GLUCOSE-1-PHOSPHATE

Figure 4. Cascade for hormone stimulation of glycogenoiysis.

lation by cAMP kinase. Thus, in addition to catalyzing phosphorylation and inactivation of GS, cAMP kinase can function to inhibit GS dephosphorylation. The interdependent enzymatic reactions which regulate glycogenoiysis can provide considerable amplification of the physiological signal, perhaps in the order of one million fold. Such cascades provide a means of simultaneously coordinating the regulation of several processes. In this case, cAMP kinase causes inactivation of glycogen synthase w^hile stimulating glycogen breakdow^n, thus preventing a futile cycle. Also, since both phosphorylase kinase and troponin, which regulates the Ca^^-dependent interaction of actin with myosin in skeletal muscle, have Ca^"^ activation constants of approximately 1.0 |uM, the concentration of Ca^"^ required for activation of glycogenoiysis is comparable to that needed for initiation of muscle contraction. The two processes are thereby coupled. Pyruvate Dehydrogenase

Pyruvate dehydrogenase (PDH) from mammalian tissues is a complex of mitochondrial enzymes which catalyzes the conversion of pyruvate, CoA, and

140

TERRY A. WOODFORD ET AL

NAD"^, into acetyl-CoA, NADH, and CO2. The three major enzymes of the complex, pyruvate decarboxylase, dihydrolipoyl acetyltransferase, and dihydrolipoyl dehydrogenase catalyze sequential reactions through mobile lipoyl residues attached to the acetyltransferase. The native complex of molecular weight 6.5-7.5 X 10^ consists of a core of 60 identical subunits (52,000 Da each) of dihydrolipoyl acetyltransferase to which 20-30 pyruvate decarboxylase molecules and 5-6 dihydrolipoyl dehydrogenase molecules are attached. The latter enzyme is a dimer of 55,000 Da subunits, whereas pyruvate carboxylase has an a2P2 composition of subunits 41,000 (a) and 36,000 (p) Da. Also present in the complex are small molar amounts of PDH kinase and phosphatase. Intramitochondrial ratios of acetyl-CoA andNADH/NAD modulate the activity of the complex through the mechanism of competitive product inhibition. The pyruvate decarboxylase component of the complex also undergoes phosphodephospho regulation. Each a subunit of pyruvate decarboxylase has three different phosphorylation sites. Phosphorylation occurs rapidly at one site, in association with enzyme inactivation, and partially on the other sites, the functions of which are not certain. The PDH complex contains a small amount of tightlybound 50,000 Da protein kinase. Phosphotransfer is stimulated by products of the dehydrogenase reaction, acetyl-CoA and NADH, and inhibited by the reactants, pyruvate, CoA, and NAD. Thus, when the reactants used by the PDH complex are in high concentration relative to the products, PDH is maintained in the nonphosphorylated active form. Low levels of reactants result in complex inactivation. Dephosphorylation and activation of the PDH complex is catalyzed by a specific phosphatase which is loosely associated with the complex in a Ca^'*"-dependent manner. PDH activity is subject to hormonal regulation in a variety of tissues. Generally, treatment of tissues with insulin increases activity. The diabetic condition is associated with a decrease in PDH activity, and these changes in activity can be accounted for by its phosphorylation state, perhaps due to altered activity of the kinase and/or phosphatase. Acetyl-CoA Carboxylase Fatty acids are essential components of membrane phospholipid and important for long-term energy storage as triglycerides. Fatty acid synthesis from cytoplasmic acetyl-CoA is catalyzed by two enzymes, acetyl-CoA carboxylase and fatty acid (FA) synthetase. The reaction catalyzed by acetyl-CoA carboxylase to provide a precursor for the subsequent reaction is as follows: Acetyl-CoA + CO2 + ATP -^ malonyl CoA + ADP + ?,. Fatty acid synthesis is subject to regulation by reversible phosphorylation of acetyl-CoA carboxylase. Acetyl-CoA carboxylase has been purified from a number of tissues including liver, adipose tissue, and mammary gland. The biotin-contain-

Function of Protein Phosphorylation-Dephospliorylation

141

ing enzyme in mammalian systems (molecular weight - 10^) exists as an active polymer of 250,000 Da subunits. Mammalian acetyl-CoA carboxylase has an absolute requirement for citrate or isocitrate and is inhibited by fatty acyl CoA through a feedback mechanism. Enzyme dissociation into an inactive protomeric form occurs in the absence of citrate. Acetyl-CoA carboxylase can be converted from a high-activity form (carboxylase a) to a low-activity form (carboxylase b) by phosphorylation of up to six sites. Acetyl-CoA carboxylase is rapidly phosphorylated by acetyl-CoA carboxylase kinase and by cAMP kinase. Acetyl-CoA carboxylase can be converted back to the high-activity form following dephosphorylation with protein phosphatase 1. Fatty acid synthesis utilizes relatively large amounts of metabolic energy and represents a pathway that is stringently controlled and subject to short-term regulation through covalent modification. Several hormones influence the rate of fattyacid synthesis by acetyl-CoA carboxylase. Glucagon, vasopressin, or angiotensin II treatments cause a decrease in enzyme activity in liver tissue. Epinephrine also causes a decrease in activity in adipose tissue. Insulin treatment results in elevated levels of acetyl-CoA carboxylase activity in both liver and adipose tissue. Hormones which raise cAMP levels, such as glucagon in the liver and epinephrine in adipose tissue, cause parallel phosphorylation and inactivation of acetyl-CoA carboxylase. The analogies between the regulation of glycogen and fatty-acid synthesis is striking. Both processes are affected rapidly and in parallel by several hormones including glucagon, insulin, vasopressin, and P-adrenergic agents such as epinephrine. The two pathways show short-term regulation by alterations in the activity of the rate-limiting enzymes, glycogen synthase, and acetyl-CoA carboxylase. These two enzymes are subject to allosteric regulation as both are activated by precursors of the pathways which they regulate, i.e., glucose-6-P and citrate. Both enzymes are phosphorylated at multiple sites by both cAMP-dependent and independent protein kinases. Phosphorylation by these kinases can be reversed by phosphatase 1. The dephosphorylation reactions can also be regulated by cAMP through phosphatase inhibitor 1 which is active only after phosphorylation by cAMP kinase. It is worth emphasizing that the regulation of two different metabolic pathways by common protein kinases and phosphatases does not imply that the overall rates of the pathways are linked. Superimposed on covalent regulation is allosteric regulation. The concentration of allosteric regulators can affect enzyme activity as well as the rate of phosphorylation-dephosphorylation. Hydroxymethylglutaryl-CoA Reductase Hydroxymethylglutaryl-CoA (HMG-CoA) reductase, the NADPH form, is the principal regulator of cholesterol synthesis in mammalian cells. Mevalonic acid, the product of the HMG-CoA reductase-catalyzed reaction, is converted to cholesterol by way of squalene and lanosterol. In most cells, the enzyme is tightly bound

142

TERRY A. WOODFORD ET AL

to the endoplasmic reticulum. Solubilized liver reductase has a molecular weight of 200,000 Da and appears to be composed of four identical 50,000 Da subunits. Nonliver cells synthesize little cholesterol even though they have the biosynthetic potential. Cholesterol synthesis in these cells is controlled by the release of free cholesterol from plasma low-density lipoproteins (LDL) that are brought into the cells by endocytosis. Hepatic cholesterol biosynthesis is also subject to feedback regulation at the HMG-CoA reductase step by exogenous cholesterol delivered to the liver cells by chylomicron remnants. In short-term regulation of liver metabolism, insulin and glucagon are known to signal dephosphorylation and phosphorylation, respectively, of HMG-CoA reductase. Reductase is converted from an active dephosphorylated form (reductase a) to an inactive phosphorylated form (reductase b) through the action of a specific cytosolic reductase kinase. Purified reductase kinase, a multimer composed of 5 8,000 Da subunits, has a molecular weight of 3 80,000 Da. The activity of reductase kinase is also controlled by reversible phosphorylation. The kinase is converted from an inactive dephospho form to an active phospho form by a second protein kinase. Dephosphorylation of HMG reductase and reductase kinase is catalyzed by phosphatase 1. Inhibition of the phosphatase produces a coordinate response in which the activity of reductase kinase is increased and the activity of reductase is decreased. Conversely, stimulation of phosphatase inhibits reductase kinase and increases reductase activity. This bicyclic system amplifies the signal mediated through phosphatase effectors. Although the kinases in the bicyclic system are c AMP-independent, an alternate route for regulation of the system by glucagon and cAMP is feasible through cAMP kinase-catalyzed phosphorylation and activation of phosphatase inhibitor 1. Tyrosine-3-Hydroxylase Tyrosine-3-hydroxylase (TH) is a mixed-function oxidase which uses molecular oxygen and the cofactor tetrahydrobiopterin (BH4) to catalyze the hydroxylation of tyrosine and form dopa. This initial step in the synthesis of catecholamines such as dopamine, epinephrine, and norepinephrine in neural cells is regulated by hormones as well as by nerve activity and neurotransmitters. TH activity is inhibited by high concentrations of catechols. In response to stimulation of adrenal medullary cells or catecholaminergic neurons, TH is activated by phosphorylation. Two forms of TH exist; a high Km form (nonphosphorylated enzyme) and a low Km form (phosphorylated enzyme). Rat pheochromocytoma (PCI2) cells have been used as a model system for the study of biosynthesis and release of catecholamines, as well as to investigate the role of cAMP-dependent and Ca^'^-dependent phosphorylation in signal transduction. Following stimulation with nerve growth factor, PC 12 cells differentiate from cells exhibiting a chromaffin-like phenotype to cells with sympathetic neuron-like

Function of Protein Ptiosptiorylation-Depliosptiorylation

143

characteristics. TH shows multi-site phosphorylation and different kinases preferentially phosphorylate different sites which may lead to quite subtle regulation of enzyme activity, enzyme half-life, and other aspects of its behavior. TH phosphorylation and activation is observed in PC 12 cells as well as in chromaffin cells and superior cervical ganglia (SCG) in response to cAMP analogs or agents such as cholera toxin or forskolin which increase cAMP levels. TH activation can be blocked by addition of the cAMP kinase inhibitor. In some cases, cAMP analogs mimic the effects of nerve growth factor in cells, which may involve cAMP-mediated pathways. The multifunctional Ca^Vcalmodulin (CaM) kinase II is the probable mediator of Ca^"^ action in response to nicotinic stimulation, K'^-induced depolarization, and calcium ionophore in neural cells. CaM kinase II catalyzes TH phosphorylation which lowers its affinity for cofactor in brainstem tissues. CaM kinase II phosphorylates sites on TH that differ from those phosphorylated by cAMP kinase. Protein kinase C also phosphorylates and activates TH, as does cGMP kinase. Protein kinase C and cAMP-dependent and cGMP-dependent protein kinases can apparently phosphorylate the same peptide fragment of TH. CaM kinase II phosphorylates the same peptide and a second unique peptide. The functional significance of phosphorylation at each residue on TH is not presently known; however, it appears that different sites may be phosphorylated in response to different stimuli. Little is known about the phosphatase(s) that dephosphorylate TH.

REGULATION OF MUSCLE CONTRACTION Smooth Muscle

The particular phosphorylating system to be described herein is dominant in vertebrate nonmuscle cells such as fibroblasts, platelets, or macrophages, as well as in smooth-muscle cells. In these cells, contractile activity is responsible for two kinds of cellular functions: (1) generalized activities, e.g., motility, cytokinesis, or changes in cell surface, and (2) specialized activities, such as phagocytosis in macrophages or secretion in platelets. With respect to smooth-muscle tissue, contractile activity plays a role in the regulation of arterial blood pressure, airway constriction, and peristalsis in the gut. Two major contractile proteins present in both vertebrate muscle and nonmuscle cells are actin and myosin. Each of these proteins forms distinct filaments that interact with each other. As will be described, two different reversible phosphorylation reactions play a role in regulating actin-myosin interaction. The energy for contractile activity is provided by ATP hydrolysis, and the transduction of chemical energy into mechanical energy of the actin-myosin interaction is mediated by myosin itself Myosin (460,000 Da) is a hexamer composed of a pair of heavy chains (200,000 Da) and two pairs of light chains (15,000 and 20,000 Da). The amino-terminal or globular "head" region of the molecule possesses ATPase

144

TERRY A. WOODFORD ET AL

activity, the actin binding site, and binding sites for the two pairs of light chains. The heavy chains intertwine to form a coil in the carboxylterminal "tail" region. Contractile activity can be quantitated by biochemical measurement of the actin-activated ATPase activity of myosin. Under physiological conditions, myosin hydrolyzes ATP at an extremely low rate (~ 4 nmoles/mg per min). The addition of actin to smooth-muscle or nonmuscle myosin increases the rate of ATP hydrolysis by about 10-fold. This occurs only when myosin has been phosphorylated at a particular serine residue on the 20,000-Da myosin light chain (LC2o)- Phosphorylation of LC20 is positively correlated with tension development and contractility in smooth muscle and appears to exert its effect by removing the inhibition imposed on the actin-activated ATPase by the unphosphorylated form of LC20. In addition to its effect on ATPase activity, LC20 phosphorylation plays an important role in the assembly and stability of the myosin filament. The key enzyme that catalyzes the phosphorylation of LC20 is myosin light-chain kinase (MLCK). This enzyme shows absolute dependency on a small globular protein called calmodulin (CaM), which contains four Ca^"*" binding domains and mediates the actions of Ca^"^ in a vast array of cellular reactions. Through its interaction with CaM and MLCK, the actin-myosin complex generates a contractile force in response to Ca^"^. MLCK may itself be phosphorylated, thus enhancing its ability to phosphorylate LC2o- However, phosphorylation of two serine residues by cAMP kinase decreases the aflfmity of MLCK for Ca^VCaM compared to monophosphorylated and nonphosphorylated MLCK. Although regulation of MLCK by cAMP kinase has not been proven to be physiological, these results suggest a mechanism whereby a rise in c AMP due to hormonal or neuronal stimulation might affect contractile activity in smooth-muscle and nonmuscle cells. A fine balance exists between MLCK activity and myosin phosphatase activity in determining the state of myosin phosphorylation. Dephosphorylation of myosin that is catalyzed by smooth-muscle phosphatase I decreases ATPase activity to the basal level found for the unphosphorylated myosin. Relaxation of contracted smooth muscle by vasodilators such as nitroglycerin or endothelial-derived relaxing factor (EDRF), or by peptides released from mammalian cardiac atria called atrial natriuretic factors (ANFs), is correlated with increased levels of cGMP. One type of ANF receptor has been shown to be a membrane-associated guanylate cyclase, which catalyzes the synthesis of cGMP. The mechanism by which cGMP promotes relaxation is not known, but it is likely that cGMP kinase mediates the actions of cGMP in vascular smooth muscle. Treatment of aortic vascular smooth cells with atriopeptin or cGMP analogs causes an increase in the activity of cGMP kinase, which has an overlapping substrate specificity with that of cAMP kinase. The potency of cyclic nucleotide analogs in inducing smooth-muscle relaxation shows good correlation with their ability to activate cGMP kinase but not cAMP kinase (Francis et al., 1988). Elevation in cGMP levels also results in the lowering of intracellular free Ca^^. The decline in

Function of Protein Phosphorylation-Dephosptiorylation

145

free Ca^"*" could lead to dephosphorylation of LC20 and to reversal of other Ca^"^dependent processes that are necessary for maintenance of contraction. Cardiac and Skeletal Muscle A detailed description of the mechanisms of contraction in cardiac and skeletal muscles, which differ in many ways from those of smooth muscle, has been recently reviewed (Robinson-Steiner and Corbin. 1986). Briefly, phosphorylation-dephosphorylation reactions play an important role in the regulation of intracellular Ca^"^ stasis and, hence, of contractile status and in the control of a supply of energy for contraction via regulation of glycogen metabolism. Increased cAMP kinase activity in the heart, following P adrenergic receptor stimulation, results in increased Ca^"^ influx through the slow channels in the sarcolemmal (SL) membrane. The idea that Ca^"^ channel modulation occurs via a phosphorylation-dephosphorylation reaction has been proposed. One 42,000 Da protein, in canine ventricular sarcolemma, thought to be the slow channel modulatory subunit, is a substrate for cAMP kinase. Following p-adrenergic stimulation, an increase in Ca^'^-induced Ca^"^ release from the sarcoplasmic reticulum (SR) augments contraction. Phospholamban, a component of the SR, is a substrate for cAMP kinase, Ca^VCaM kinase and protein kinase C. Phosphorylation catalyzed by these kinases is additive, suggesting the presence of three distinct phosphorylation sites. Phosphorylation of phospholamban is associated with increased Ca^"^ transport and ATP hydrolysis by the SR and may be a hallmark of P-adrenergic stimulation. Troponin I is also an excellent substrate for both cAMP kinase and phosphorylase kinase, and its phosphorylation decreases myofibrillar Ca^"^-binding and increases the level of Ca^"*" required for activation of myosin ATPase. In summary, cAMP and Ca^"^ are key mediators of cardiac muscle contractility and metabolism, and a homeostatic mechanism is operative between the two second messengers. It appears that the primary second messenger in cardiomyocytes is Ca^"^ and its actions during contraction are buffered by cAMP-dependent processes. Thus, cAMP can stimulate Ca^"*" influx through the slow channel, which augments Ca^"^ release by the SR and stimulates the rate and quantity of Ca^"^ sequestered by the SR via phospholamban phosphorylation, resulting in more releasable Ca^"^ for the next contraction, c AMP also leads to an increase in phosphorylation of troponin I which in turn decreases the sensitivity of troponin C for Ca^"*" in the troponin complex. In concert with these changes are the mobilization of carbohydrate and lipid fuels and the increase in heart rate which are brought about by cAMP kinase-catalyzed phosphorylation reactions in cardiomyocytes and pacemaker cells. The regulation of glycogen metabolism and skeletal-muscle contraction is finely integrated. As discussed earlier, phosphorylase kinase, which activates glycogen phosphorylase via phosphorylation, may be controlled by both Ca^"*"- and cAMP-

146

TERRY A. WOODFORD ET AL.

dependent processes. Release of Ca^"^ from the SR not only initiates muscle contraction but also activates phosphorylase kinase, leading to an increased supply of ATP for contraction. Coordinately, glycogen synthase is phosphorylated by cAMP kinase, decreasing its activity. Troponin C, a key Ca^'^-binding protein in the initiation of muscle contraction, is also the dominant regulator of the dephosphorylated form of phosphorylase kinase (while CaM determines Ca^"^ sensitivity of the phosphorylated form). Thus, synchronization between activation of glycogenolysis and contraction may be afforded by the same Ca^"^-binding protein, troponin C.

PROTEIN SYNTHESIS AND GENE TRANSCRIPTION Regulation of Protein Synthesis

Protein phosphorylation acts at several levels during protein synthesis. An interesting cell system for studying protein synthesis is the rabbit reticulocyte, in which regulation by hemin occurs. Reticulocyte lysates deprived of hemin or treated with low levels of double-stranded RNA (dsRNA) are inhibited at the level of protein chain initiation. Inhibition is accompanied by increased phosphorylation of initiation factor 2 (eIF-2) which participates in the formation of a ternary complex with formyl-methionine-charged tRNA (Met-tRNAf) and GTP. This phosphorylation occurs concomitant with the disappearance of the initiation complex in association with the 40S ribosomal subunit. Phosphorylation of the a subunit of eIF-2 (eIF-2a) is catalyzed by two distinct protein kinases; heme-regulated inhibitor (HRI), which is activated in hemedeficient lysates and a protein kinase called dsl, which is activated by dsRNA. Both kinases phosphorylate the same sites on the 38,000 Da eIF-2a. This phosphorylation event is important in the inhibition of protein-chain initiation. Phosphorylation of eIF-2a reduces its ability to interact with other initiation components that are involved in the formation or stabilization of the ternary complex. HRI and dsl are themselves activated by phosphorylation; this illustrates another level of regulation. The activity of dsl is also regulated by protein phosphatase 2A, emphasizing the role of protein dephosphorylation in the inhibition process. An interesting case is the synthesis of interferon (INF) which often occurs when a cell becomes infected with a virus. INFs are best known for their antiviral effects; however, they also exhibit potent antiproliferative effects and can modulate cell differentiation. INF induces synthesis of 2',5' oligoadenylate, a positive cellular effector which activates an endogenous RNAse L to catalyze the degradation of host and viral mRNA. Simultaneously, a dsRNA-dependent protein kinase is induced by INF, which is identical to dsl in its ability to phosphorylate eIF-2. Again, phosphorylation reduces translation of viral mRNA into protein.

Function of Protein Phosphorylation-Dephosphorylation

147

Additional protein kinases differentially phosphorylate components of the protein-synthesizing system, including proteins associated with the 40S ribosomal subunit and initiation factors 2, 3, 4B, and possibly 5. These modulators include cAMP-dependent protein kinase I and II, casein kinase I and II, and two proteaseactivated kinases. Each of the protein kinases phosphorylates two or more components, and substrate proteins are often modified by at least two different protein kinases. Thus, the potential exists for a highly-regulated system although more proof is required for the physiological relevance of such phosphorylation reactions. The phosphorylation in cells of a ribosomal protein called S6 is particularly promoted by hormones which enhance cAMP levels, as well as by insulin, growth factors, transforming viruses, and other agents which stimulate cell proliferation. Multiple phosphorylation of S6 occurs, with some sites being targeted by more than one protein kinase and other sites being distinct. Phosphorylation of S6 by mitogen-stimulated S6 kinase has been shown to stimulate translation of globin mRNA four-fold in a reconstituted protein-synthesizing system (Palen and Traugh, 1987). It is possible that ribosomes containing highly-phosphorylated S6 are preferentially incorporated into translationally-active polysomes. Of interest in the context of protein phosphorylation and its effect on the elongation step of protein synthesis is the recent finding that elongation factor-1 (EF-1) is phosphorylated in the polyribosome fraction of reticulocyte lysates by an as yet undefined kinase, and this phosphorylation apparently causes EF-1 to dissociate from the ribosome. In addifion, phosphorylation of EF-2 by Ca^'*"/calmodulin protein kinase III suppresses its ability to support protein synthesis in a reconstituted system (Nairn and Palfi-ey, 1987). This likely represents a specific mechanism whereby extracellular signals acting via Ca^"^ rapidly regulate the protein synthetic apparatus. Regulation of Gene Transcription Gene expression is subject to multilevel regulation. Expression may be controlled at the level of DNA methylation or chromatin structure; by the cellular level of positive or negative DNA-binding regulatory factors; or by agents which accelerate the rate of basal gene transcription by modulating the activity of transcription factors already bound to target sequences in the promoter regulatory region. Covalent protein modification by phosphorylation is involved in the regulation of transcription of DNA into RNA. There is evidence that regulation of RNA polymerase I occurs via phosphorylation. In addition, phosphorylation of a long repeated motif in the carboxyl-terminal domain of RPBl, a subunit of RNA polymerase II, is thought to control RNA polymerase II activity. A CDC28-containing complex, which will be discussed in a later section, appears to be the kinase catalyzing this phosphorylation.

148

TERRY A. WOODFORD ET AL

Activation of receptor tyrosine kinases following growth stimulation of quiescent cells by mitogens or growth factors results in transcriptional activation of certain mRNAs, including those for actins, ornithine decarboxylase, the protooncogenes c-fos and c-myc, and tumor antigen p53. The chronology of activation and deactivation differs among these genes, with induction of c-fos being quite transient and rapid in onset (within minutes after stimulation). Both the c-fos and c-myc genes encode nuclear proteins which are highly conserved among vertebrates and differentially expressed during development and cell diflferention. PDGF treatment of cells often stimulates c-fos and c-myc gene expression by a mechanism which may at least partially involve activation of protein kinase C. Regulation of c-fos and c-myc expression by an EGF-directed pathway in cells may utilize both Ca^"^ and cAMP kinases as cooperating mediators. Various transcriptional regulators such as SV40 T antigen, glucocorticoid receptor protein, yeast heat shock factor, and a yeast transcriptional activator called ADRl are phosphoproteins. Yeast ADRl controls alcohol dehydrogenase II (ADH2) gene expression. Studies have shown that this activator is phosphorylated by cAMP kinase, resulting in its inactivation and consequently, inhibition of ADH2 expression. The expression of certain genes in higher eukaryotes is stimulated by raising intracellular cAMP, and protein phosphorylation appears to serve an important role. Examples of cAMP-inducible genes, include those which encode tyrosine aminotransferase, prolactin, phosphoenolpyruvate carboxykinase, and plasminogen activator. Several cAMP regulated genes, which have been isolated and characterized to date share common features: (1) they are expressed in tissues that are responsive to hormones or regulatory factors; (2) their rate of transcription is rapidly altered by cAMP; and (3) the mRNA produced by transcription of these genes is relatively short-lived. The transcriptional response of genes to cAMP has been localized to a specific DNA sequence termed the "cAMP response element" or CRE. Most CREs lie within the first 150 base pairs of the 5' flanking region of their respective genes. These elements are regarded as basal enhancers, in addition to ftinctioning as inducible enhancers. The CRE contains a highly conserved 8-bp palindromic sequence T(G/T)ACGTCA. Initial evidence that this sequence was involved in cAMP regulation of gene transcription camefi'omdeletion analysis studies of the promoter-regulatory region of the rat phosphoenolpyruvate carboxykinase (PEPCK) gene (Short et al., 1986). A synthetic oligonucleotide encompassing this region of the PEPCK gene confers cAMP responsiveness when introduced upstreamfroma heterologous promoter and gene. Similar experiments have identified functional CREs in the somatostatin, proenkephalin, and vasoactive intestinal peptide (VIP) genes. When c AMP response elements were first described, there was speculation that cAMP regulation of gene expression in eukaryotes occurred in a manner analogous to that in Escherichia coli. In bacteria, c AMP binds to a receptor protein (catabolite

Function of Protein Phospliorylation-Deptiosptioryiation

149

repressor protein, CRP; cAMP-activator protein, CAP) and the resulting complex interacts with certain DNA sequences that influence gene transcription at nearby promoters. However, there is currently no definitive evidence to suggest that the R subunit of cAMP kinase (structurally similar to CAP in the cAMP-binding domain) is a specific DNA-binding protein in the presence or absence of cAMP. When levels of cAMP are elevated in eukaryotic cells, cAMP binds to the R subunit and the holoenzyme dissociates, yielding an R subunit dimer and two active catalytic (C) subunits. It is the C subunit that mediates the effects of cAMP on gene transcription in eukaryotic cells. Several experimental observations support this role for C subunit. First, the C subunit migrates from the cytoplasm into the nucleus after treatment of cells with agents that increase intracellular cAMP. Second, co-transfection of cultured cells, with a plasmid containing the gene coding for an inhibitor of C together with a cAMP-inducible gene reduces cAMP induction of the co-transfected gene by up to 90%. Third, the cAMP-regulated expression of endogenous genes is blocked by treating cells with inhibitors of cAMP kinase and is stimulated by specific analogs of cAMP. Fourth, microinjection of purified C subunit into tissue culture cells can stimulate expression from a stably integrated fusion gene containing a cAMP-responsive promoter fused to a reporter gene. This latter result indicates that the C subunit alone is sufficient to induce expression of certain cAMP-responsive genes. Perhaps the most direct experimental evidence is the recent demonstration that the C subunit stimulates transcription of the urokinasetype plasminogen activator (uPA) gene in vitro (Nakagawa et al., 1988). More recently, a different cAMP-responsive element has been identified. This sequence, termed the activator protein 2 (AP-2)-binding site, was originally identified in certain promoters such as those for simian virus 40 (SV40) and human metallothionein IIA DNA. While the AP-2 element has not been subjected to the same detailed analysis as the CRE, the two share salient features, including: (1) the presence of more than one copy in the 5'-flanking regions of several genes, and (2) their function as basal transcription enhancers which permits increased rates of gene transcription in the absence of any hormonal stimulation. The AP-2 element, however, is differentfi-omthe CRE in its ability to mediate both c AMP- and phorbol ester-induction of gene transcription (Imagawa et al., 1987). For example, transcription of the human metallothionein IIA gene is rapidly altered by a number of factors including heavy metals, glucocorticoids, cAMP, and tumor-promoting phorbol esters. It contains three AP-2 sequences which are responsive to cAMP and phorbol esters; however, the responsiveness of the enhancer to these agents is cell-specific. When the AP-2 sequence is linked to a heterologous promoter and transfected into mammalian cells, transcription is induced by treatment of cells with either phorbol ester, which activates protein kinase C, or by forskolin, which raises the intracellular level of cAMP. Interestingly, the human proenkephalin gene contains both CRE and AP-2 elements. Deletion analyses of the promoter regulatory region of this gene suggests that both sequences confer cAMP responsiveness.

150

TERRY A. WOODFORD ET AL

The identification of DNA-binding proteins involved in the regulation of gene transcription by cAMP is ongoing. DNase I footprinting and gel shift assays using target DNA sequences are two methods that have been used to identify proteins which function as transcription factors. The best characterized CRE-binding protein, known as CREB (43,000 Da), is a phosphoprotein that has been purified by DNA affinity chromatography from rat pituitary gland (Yamamoto et al., 1988). CREB is phosphorylated in vitro by the C subunit of cAMP kinase and in vivo by treatment of cells with forskolin. Cyclic AMP stimulates somatostatin gene transcription by stimulating phosphorylation of CREB at Ser-133. Protein kinase C may also phosphorylate CREB. Phosphorylation of CREB does not appear to alter the binding affinity of the protein for the CRE, but it may stimulate formation of CREB dimers. Mutation of the phosphorylation sites of CREB causes a loss in transcriptional stimulation. The 50,000 Da protein which binds to the AP-2 regulatory element has also been purified by DNA-affinity chromatography. Direct binding of purified AP-2 to AP-2 element(s) in the promoter-regulatory region of the metallothionein IIA, growth hormone, and proenkephalin genes has been demonstrated. This protein stimulates transcription from the metallothionein IIA promoter in vitro, but forskolin treatment of cells to elevate cAMP levels does not affect its DNA-binding properties. As in the case of CREB, the correlation between the activation of protein kinase by cAMP and the increased level of gene transcription is not fully understood. It is possible that cAMP/protein kinase effects occur indirectly through a regulator of CREB or AP-2. This mechanism of transcriptional activation can be illustrated by that proposed for the regulation of transcription factor NK-kB (Baeuerle et al., 1988). In cell lines derivedfrommature B lymphocytes. NK-kB (~ 60,000 Da) is a constitutively active DNA-binding protein. Binding of NF-kB to the kappa light chain enhancer and the human immunodeficiency virus (HIV) enhancer confers transcriptional activity, inducibility, and cell-stage specificity to genes controlled by these enhancers. In cells which do not have constitutively active NF-kB, this transcription factor is activated by phorbol ester treatment. In pre-B cells, NF-kB forms a complex with a specific inhibitor protein. Stimulation of cells with a phorbol ester such as TPA causes this complex to dissociate. The possibility that TPA-activated protein kinase C can directly phosphorylate the inhibitor in this complex, decreasing its affinity for NF-kB and causing dissociation to liberate NF-kB for transcriptional activation, is being investigated. Regulation of transcription most likely involves a situation whereby inducible enhancer elements can be controlled by positive or negative regulatory elements either alone or in combination. The proximity of two regulatory domains within the promoter may facilitate protein-protein interaction, leading to coordinated control of gene expression. In the case of cAMP-regulated genes, transcription factors appear to bind to the CRE and the AP-2 in a cAMP-independent manner and thus resemble basal transcription factors. Cyclic AMP-inducible enhancer

Function of Protein Pliosptiorylation-Depliospliorylation

151

elements may represent a class of enhancers which are regulated by rapid modification of proteins already bound to their specific DNA elements. These proteins may maintain a basal rate of gene transcription, depending on their tissue concentration. Rapid changes in the intracellular cAMP level would lead to modification of the regulatory proteins already bound to their respective elements in the promoter-regulatory region of the gene. This modification could increase the binding of other basic transcription factors such as RNA polymerase II or increase proteinprotein interactions between CREB or AP-2 and other transcription factors. Cyclic AMP-induced phosphorylation of CREB or AP-2 might also lead to interaction with another non-DNA-binding protein, forming a higher-order complex with basic transcription factors, which would then lead to the increase in gene transcription.

CELL GROWTH AND DIFFERENTIATION The cell cycle in proliferating cells typically consists of a Gi phase, a DNA-synthetic (S) phase, and a mitotic (M) phase. Quiescent cells exit the cell cycle and enter an additional stage termed Go. A burst of phosphorylation occurs as cells undergo mitosis, during which time proteins including histone HI, nucleoplasmin, the nuclear lamins, microtubule-associated protein 2 (MAP2), and centrosomal proteins become hyperphosphorylated. These proteins are involved in events such as chromosome condensation, nuclear membrane breakdown, and assembly of the mitotic spindle. The majority of these phosphorylations occur on serine and threonine residues. During the past 20 years, three general approaches have been used to study eukaryotic cell-cycle regulation. One approach has used yeast as a model system for genetic investigation. Among the 50 genes known to be specifically required for cell-cycle progression in the budding and fission yeast, one has attracted particular attention. The CDC28 (cell division cycle) gene of Saccharomyces cervisiae and its homolog cdc2'^ in Schizosaccharomycespombe play a central role in regulating the cell cycle in both yeasts, being required for the Gi/S phase transition and independently, at least for cdc2'*", for initiation of mitosis. Both CDC28 and cdc2"*" encode protein kinases. The protein encoded by the fission yeast cdc2"*" gene is known as pp34^^^^ and its human homolog has recently been cloned. Lack of pp34 kinase activity during G2 leads to cell-cycle arrest just prior to mitosis, while deregulated pp34 activity leads to mitotic advancement. Another approach has employed cell fusion or microinjection to identify mitotic regulators. The best-characterized regulator is maturation-promoting factor or MPF. This factor causes meiotic maturation when injected into immature oocytes of the African clawed toad, Xenopus laevis, which are arrested at the first meiotic prophase. MPF also induces the G2-M phase transition in mitotic cells. MPF activity changes during meiotic maturation and during mitosis, indicating that its cycling may be an integral part of the oscillator which controls the cell cycle. The fact that mitosis can be induced by MPF in the absence of protein synthesis suggests

152

TERRY A. WOODFORD ET AL.

that MPF post-translationally activates a series of reactions that lead to events such as nuclear membrane breakdown, chromosome condensation, and assembly of the mitotic spindle apparatus. MPF exists as a protein complex that possesses protein kinase activity. An additional approach to studying cell-cycle regulation has been to examine enzyme activities that display genuine periodicity during the cell cycle. Histone HI kinase (HIK), often referred to as "growth-associated" or M-phase-specific-HlK, is transiently activated at the G2/M transition in dividing-cells such as sea urchin eggs, meiotic starfish oocytes, meiotic Xenopus oocytes, and various rodent and human somatic cell lines. In addition to being a serine/threonine phospho-transferase with a high affinity for histone HI, HIK can utilize MgGTP, as well as MgATP, as the phosphate donor and is unresponsive to cyclic nucleotides, Ca^VCaM, Ca^'^-diacylglycerol, and polyamines. Maximum recovery of activity requires stringent measures to inhibit phosphatases, suggesting that HIK is controlled by phosphorylation or by a phosphorylated modulator. Hyperphosphorylation of histone HI at mitosis-specific sites occurs during early metaphase at the time of chromosome condensation. This modification may be associated with the process of condensation and even with the initiation of mitosis. Highly purified MPF f^^omXenopus and M phase phase-specific HI kinase from starfish represent multiprotein complexes in which the kinase component is closely related to pp34^^^^ protein kinase of yeast. Cloning studies, antibody cross-reactivity, and cross-inhibition of kinase activities with species specific inhibitors has provided evidence for structural and fiinctional homology. Activation of pp34 is a multi-step process (Figure 5). In yeast, pp34^^^^ kinase is regulated by reversible phosphorylation during the cell cycle. This is also the case for pp34 kinase in mammalian cells. A complex regulatory network controls entry of yeast cells into mitosis with pp34^^^^ acting downstream of other genefimctions.Two of these gene products, weel+ and nimH-, are likely to encode two additional protein kinases, indicating a key role for protein phosphorylation in mitotic entry. The product of the weel gene has an inhibitory effect on pp34 activation and, hence, entry into mitosis, while niml negatively regulates weel fimction. Another gene, cdc25, positively regulates pp34 activation and mitotic entry, possibly by indirectly decreasing the phosphorylation state of pp34 (the cdc25 product does not resemble known phosphatases or protein kinase inhibitors). Increased kinase activity at M phase is associated with pp34 dephosphorylation. A positive-feedback loop may exist in which active (dephosphorylated) MPF activates a putative-specific phosphatase, which dephosphorylates and activates MPF. A decrease in protein kinase activity occurs upon the exit of cellsfi*omM phase. Rephosphorylation of pp34 may be required for entrance into the subsequent Gi phase. The activity of pp34, but not its protein level, changes during the cell cycle, suggesting that it is not the kinase itself, but rather a modulator of its activity that oscillates and undergoes proteolysis at the end of each cycle. A good candidate for this modulator is one of

Function of Protein Piiospitorylation-Dephosphorylation

G2

inactive

Hh

mitosis

153

G1

active

pp34

rephosphorylation

Figure 5. Regulation of pp34'^^^^ protein kinase activity during entry into and exit from mitosis in Schizosaccharomyces pombe.

the cyclins. These are proteins originally characterized from mitotic sea urchin eggs, whose cellular levels increase at G2/M, then decrease at the end of mitosis. In HeLa cells, MPF is knov^n to be a complex of pp34 kinase together with a phosphorylated 62 kD cyclin. In Xenopus oocytes, pp34 is complexed to a 45 kDa-phosphorylated component. The loss of cyclin and, hence, of active MPF at the end of mitosis, via proteolysis by a protease, may be regulated by another serine/threonine kinase, p39"^°^ This kinase is the product of the c-mos proto-oncogene, the cellular homolog of the transforming gene of the Moloney murine sarcoma virus. It may act to prevent proteolysis of cyclin, by phosphorylation of either the cyclin protease or cyclin, conferring proteolysis-resistance. p39"^^^ which becomes hyperphosphorylated in unfertilized frog oocytes, may itself be regulated, not only by phosphorylation, but by proteolysis by the Ca^'*'-dependent protease, calpain. This kinase may represent a component of cytostatic factor (CSF), which serves to stabilize high levels of MPF during M-phase (Sagata et al., 1989). Other proteins interact with pp34^^^^ in mammalian cells, as well as in yeast, to form a heteromultimeric mitotically active protein kinase complex. In yeast, pp34 is associated with a 13 kDa protein. The role of pi 3, which is encoded by the yeast sucl"^ gene, in regulating pp34 is not clear. The phosphorylation state of pp34, its association with cyclin, and the kinase activity of the complex are each subject to cell-cycle regulation (Draetta et al., 1989). By early G2, pp34 becomes highly phosphorylated in conjunction with its associated cyclin, whereas maximal protein kinase activity is not fully expressed until later. Phosphorylation occurs on serine and threonine residues. The level of phosphorylation decreases as cells enter mitosis. It appears that neither phosphorylation nor association between pp34 and

154

TERRY A. WOODFORD ET AL.

cyclin provides full stimulation for the observed mitotic activation of the pp34/cyclin complex. The final step in activation is as yet unidentified in molecular terms. It could be additional phosphorylation or dephosphorylation of pp34, phosphorylation of cyclin by pp34, addition of another unidentified polypeptide to the pp34 complex, or even expulsion of other protein components from the complex. Interestingly, it appears that pp34 is phosphorylated on tyrosine residue(s), and it may be that dephosphorylation at this site(s) is required for maximum activation. In yeast, pp34^^^^ is phosphorylated on Tyr-15, a residue within the presumptive ATP-binding domain (Gould and Nurse, 1989). Phosphorylation sites in pp34^^^^ might be blocked when it is in the heteromeric complex or phosphorylation on tyrosine might be a transient event preceding complex formation. Substitution of Tyr-15 by phenylalanine advances cells prematurely into mitosis, thus establishing that tyrosine phosphorylation/dephosphorylation directly regulates pp34 function. pp60^'^'^^ is the cellular homolog of the Rous sarcoma virus-transforming protein. Microinjection experiments v/ith Xenopus oocytes and analysis of tyrosine-phosphorylated proteins that are unique to M phase and in mitogenic or virally-transformed cell lines suggest a role for pp60^'^'^^ in cell growth regulation. pp60^"^'*^ is a substrate for multi-site phosphorylation catalyzed by cAMP kinase and protein kinase C. pp60^'^'^^ is also phosphorylated by purified Xenopus MPF or by pp34^^^^ kinase from HeLa cells (Shenoy et al., 1989). The proto-oncogene is modified on novel serine and threonine residues within its amino terminus during mitosis. These modifications are associated with a 4-7-fold increase in pp60^"^'^^ tyrosine kinase activity. In animal cells, pp60^"^^^ is normally phosphorylated at Tyr 527, which is known to strongly inhibit its tyrosine kinase activity. This tyrosine kinase is also dephosphorylated and inactivated at the end of cell division. It is possible that pp34 and pp60^"^'^^ phosphorylate each other. This might constitute an interesting feedback mechanism that is important for mitotic control. Phosphorylation of the simian virus large tumor antigen (T) by cdc2 protein kinase has also been shown. This modification stimulates viral DNA replication, indicating that this replication origin-binding protein is regulated by phosphorylation. The cdc2 protein kinase is required for initiation of DNA replication in yeast and probably also in mammalian cells.

CELLULAR TRANSFORMATION Phosphorylation of tyrosine is a rare protein modification in normal cells. Phosphate esterified to tyrosine was first detected in viral-transforming proteins (Hunter and Sefton, 1980). During their replication cycle, it is thought that some RN A tumor viruses (retroviruses) captured certain cellular genes called proto-oncogenes into their genome and these genes became viral-transforming genes (oncogenes). Thus, the first protein tyrosine kinase (PTK) activities were also detected in association with oncogenes. Later, proto-oncogenes were shown to encode PTK activities. Whether sequence alterations or mutations are required for proto-oncogenes to

Function of Protein Pl^osphorylation-Dephosphorylation

155

assume the transforming activity of oncogenes remains unresolved in many cases. Another type of PTK activity was detected in association with growth-factor receptors. Receptor PTKs are stimulated by their respective ligands. Thus, certain viral-transforming proteins, cellular proto-oncogenes, and growth factor receptors possess intrinsic tyrosine-specific kinase activity and are thought to function in cellular events including growth'control. Transformation of cells and loss of growth control results in changes in cell physiology, morphology, and behavior. When cultured, transformed cells are often rounded rather than spread, usually adhere less well to their substratum, and are not subject to contact inhibition and grow unrestricted. Changes occur in the internal cytoskeletal architecture of the cell, as well as in the exterior architecture. In normal fibroblasts, actin microfilaments aggregate into actomyosin bundles which may have one terminus anchored to the ventral cell surface at distinct sites of contact with the substratum. These sites are called adhesion plaques. Both stress fibers and adhesion plaques are associated with cell surface fibronectin. Transformed cells show marked disruption of actin-containing microfilament bundles and organization of intermediate filaments as well as loss of adhesion plaques. In addition to cytoskeletal changes, transformed cells demonstrate decreased serum-dependency as well as altered enzymology and patterns of gene expression. Retroviral Tyrosine Kinases and Their Cellular Homologs

Rous sarcoma virus (RSV) is a retrovirus that causes sarcomas in chickens and can transform both avian and mammalian cells in culture (Collett et al., 1979). The transforming potential of RSV resides in a single genetic element called src. The product of the src gene is a 60,000-dalton phosphoprotein, pp60^"^'"^. The src gene of RSV is closely related to a gene called c-src, found in all normal vertebrate cells, whose product is also a 60,000-dalton phosphoprotein. pp60^'^''^ is present in normal cells at 50-100-fold lower levels than pp60^'^''^ in virus-transformed cells. However, overexpression of pp60^'^'"^ at a level that is 10-fold higher than that normally found in rat fibroblasts does not morphologically transform cells. pp60^"^^^ and pp60^"^'"^, although structurally similar, can be distinguished by peptide mapping and show significant differences in amino-acid sequence, particularly at the carboxyl-terminal region. The extensive conservation of c-src in animal evolution suggests that is plays an important role in the cell and that transformation by v-src may result from perturbation of a normal growth or developmental pathway mediated by c-src. The complete sequences of the v-src gene and the chicken c-src gene are known and the proteins have been thoroughly studied. pp60^'^'"^ is a protein of 526 amino acids, a substantial subpopulation of which rapidly associates with peripheral membranes upon synthesis. An amino-terminal glycine is modified by covalent attachment of a myristyl group which is thought to be important for membrane association.

156

TERRY A. WOODFORD ET AL.

pp60^'^'"^ is a tyrosine-specific protein kinase whose expression results in the transformation of cultured fibroblasts. In RSV-transformed cells, the level of phosphotyrosine in protein increases 10-fold, which reflects pp60^''^-induced phosphorylation of cellular proteins. The normal cellular homolog, pp60^"^^^, is also a tyrosine-specific protein kinase with similar enzymatic properties. The kinase catalytic domain is contained within the carboxyl-terminal portion of pp60^"^'^^, which also contains determinants for its transforming activity. Recognizable sequence homology between PTKs and other protein kinases occurs in this region. Among the conserved residues in pp60^"^'"^ is Lys 295, which corresponds to Lys 72 in the C subunit of cAMP kinase. This lysine residue participates in ATP binding. In both cases, the lysine is located about 15 residues downstream from a sequence referred to as the glycine loop (Gly-X-Gly-X-X-Gly). A similar sequence is observed in a number of nucleotide-binding proteins and, in cases where the three-dimensional structure is available, this sequence forms a bond which makes contact with the ribose ring of the nucleotide. Alteration of Lys 295 in pp60^"^'^^ abolishes both its kinase and transforming activities. This residue in the C subunit of cAMP kinase is also essential for kinase activity. pp60^"^'"^ is phosphorylated at two major sites, Ser 17 and Tyr 416, the latter of which lies in the catalytic domain of pp60^"^'^^ and is up to 30% phosphorylated at steady state. An equivalent tyrosine is found in the catalytic domain of other viral PTKs and is also a major phosphorylated residue. Phosphorylation of tyrosine 527 in the proto-oncoprotein pp60^"^'"^ near the carboxyl terminus of the protein where its structure diverges from pp60^"^'^^ negatively regulates its kinase activity. Autophosphorylation of pp60^'^'*^ at Tyr 416 causes a three-fold increase in tyrosine phosphotransferase activity. Other minor sites of tyrosine phosphorylation near the amino terminus may also regulate PTK activity. pp60^'^^^ can associate with middle T (mT), the transforming protein of polyoma virus, and a major substrate for tyrosine phosphorylation in this complex. pp60^'^'"^ complexed with mT has greater PTK activity than unassociated pp60^'^'"^. Transformation of cells by polyoma virus may alter the activity or specificity of pp60^'^'"^ and thus disrupt processes involved in the normal regulation of cellular proliferation. About 60% of pp60^'^^^ molecules are phosphorylated at Ser 17 in RSV-transformed cells. This site most likely represents a recognition sequence for cAMP kinase. Phosphorylation at this site is not crucial for biological activity of pp60^'^^^ but may have some other function, such as modulation of enzyme activity, disposition on the membrane, or interaction with other proteins. Among the cytoskeletal substrates thought to be phosphorylated by pp60^'^^^ in RSV-transformed cells is a subpopulation of vinculin. pp60^"^''^, as well as pp60^'^''^, has been found to be present in adhesion plaques, where vinculin is localized. In addition, the intermediate filament protein, vimentin, the microfilament-associated protein, filamin, and 36 kDa and 80 kDa proteins of undefined function, although containing no phosphotyrosine in normal cells, are found to contain low levels of

Function of Protein Phosptiorylation-Deptiosphorylation

157

phosphotyrosine in transformed cells. The steady-state levels of phosphorylation of most of these rather abundant proteins is low (1—10%) and the effects of tyrosine phosphorylation, if any, have not been ascertained. The turnover rate of phosphotyrosine on proteins in RSV-transformed cells is extremely rapid, which may also account for the small fraction of phosphorylated protein observed at any given instant. Crucial targets, unidentified rare proteins, may exist in cells. It is also possible that tyrosine phosphorylation in proteins is largely limited to the PTKs themselves and that autophosphorylation alters other properties besides kinase activity. Although much attention has been focused on tyrosine phosphorylation, analysis of specific proteins, such as ribosomal protein S6, reveals that protein phosphorylation on serine residues in cellular proteins may also be quantitatively altered by pp60^'^^ expression. Some effects of pp60^'"^ may therefore be mediated by regulation of a protein kinase and/or phosphatase specific for serine. The C subunit of cAMP kinase purified from bovine heart can be phosphorylated on tyrosine residues by pp60^''^. Such modification alters enzyme catalytic activity and its regulation by the R subunit and may contribute to the changes in serine phosphorylation observed following cell transformation by pp60^^^. The activities of pp60^'^'"^ and cAMP kinases often have opposing effects on cell growth. In fact, characteristics associated with transformation can be reversed by elevation of intracellular cAMP in many cases. This "reverse transformation" induced by cAMP results in the rapid acquisition of the normal fibroblastic phenotype. An attractive hypothesis to explain these cAMP-induced changes could be mediated through the cAMP kinase system and the phosphorylation of a variety of cellular proteins important in cytoskeletal organization, cell adhesion, contractility, and ceil growth. For example, following treatment of Chinese hamster ovary cells or mouse fibroblasts with dibutyryl-cAMP, a membrane-permeable cAMP analog, and in vivo labeling with [^^PJorthophosphate, enhanced phosphorylation of vinculin, actin, and a- and p-tubulins is observed. Tables. Proteln-Tyrosine Kinases Gene Viral-related c-src c-yes c-fgr c-fps c-fes c-abi c-ros

Protein pp60^-^^^

— —

p98^-^pr p92^-^"^ pi 50^-^^'

—

Intracellular Location Membrane

— —

Soluble

—

Cytoplasm/ nucleus

—

Gene Viral v-src v-yes v-fgr v-fps v-fes v-abi v-ros

Protein

Intracellular Location

pp60'-'''

Membrane/cytoskeleton Membrane Cyloplasm Membrane/soluble Cytoplasm PI 20^^^'^'^' Membrane

pgQgag-yes P708ag-f8r pT40ga8-fps P85gag-fes

p(^QB^g-ros

Membrarie

158

TERRY A. WOODFORD ET AL

RSV is not unique in possessing a transforming protein with PTK activity. The products of the v-yes, v-fgr, v-fps, v-fes, v-abl, and v-ros retroviral oncogenes are other examples of proteins possessing PTK activity (Table 5). The v-src, v-yes, v-fps, and v-ros genes are oncogenes of different chicken sarcoma viruses. The v-fgr and v-fes genes are oncogenes of feline sarcoma viruses and the v-abl gene is the oncogene of a mouse lymphoma virus. Cells transformed by each of these viruses show elevated levels of phosphotyrosine in cellular protein. Like RSV, these viruses all arose by acquisition of cellular sequences that were expressed as part of their respective transforming proteins. Work is ongoing to characterize the functions of their tyrosine kinase activities. Growth Factor Receptor Tyrosine Kinases Serum growth factors stimulate cells by interacting with their cognate cell surface receptors. The receptors for epidermal growth factor (EGF), plateletderived growth factor (PDGF),fibroblastgrowth factor (FGF), insulin-like growth factor I (IGF-I), and insulin show intrinsic ligand-sensitive tyrosine kinase activity (Yardin and Ullrich, 1988). Little is known about the biochemical cascades that receptor tyrosine kinases (RTKs) trigger, including the substrates for tyrosine phosphorylation. Ligand binding to receptors induces pleiotropic cellular responses. Stimulation may result in cell-cycle progression and cell replication. Numerous cell-surface events are initiated, including ion and glucose transport, pinocytosis. and membrane ruffling. These events parallel activation of pathways including glycolysis, polyamine synthesis, and ribosomal protein S6 phosphorylation. Increased protein, RNA, and DNA synthesis are observed 3-20 hours after ligand binding. The expression of specific genes such as c-myc and c-fos is detectable within minutes. The primary structure of a number of RTKs has been determined from cloned cDNA sequences. This consists of a large glycosylated extracellular ligand-binding domain connected by a single hydrophobic membrane-spanning segment to a catalytic domain. Signaling activity is controlled by ligand binding and modulated by intracellular receptor subdomains which transduce the cellular signal. This structural design represents a unique system for allosteric regulation. Three distinct structural subclasses of this family of receptors have been proposed which may reflect their unique molecular pathways of cell activation (Figure 6). The receptor for EGF (EGF-R) and its close relative, the HER2/neu receptor-like gene product (synonymous with c-erb) represent the first subclass of RTKs. Two cysteine-rich sequence repeat regions occur within the extracellular (I-R) domain of these monomeric receptors. Subclass II includes receptors for insulin and IGF-1 which exist as heterotetrameric structures composed of two a and P subunits connected by disulfide bonds. The a subunits contribute to the formation of the ligand-binding domain and are linked to two p subunits which traverse the membrane and contain

Function of Protein Phospborylation-Dephosphorylation

I

EGF-R neu/HER2

n

hR IGF-1 -R

159

in

PDGF-R c-f ms/CSF-1 -R c-kit

Figure 6. Structural subclasses of growth factor receptor tyrosine kinases. Shaded cylinders in subclasses I and II represent extracellular cysteine-rich repeat domains, and open circles in subclass III represent conserved cysteine residues. Open cylinders represent the intracellular tyrosine kinase domains of the receptors.

the kinase domains. One cysteine-rich repeat domain is found in each of the a subunits. The third receptor subclass is the RTK family exemplified by receptors for PDGF and a macrophage growth factor (colony-stimulating factor 1; CSF-1), and is defined by two distinct features: (1) the lack of cysteine-rich repeat clusters within the extracellular domain and its replacement by another type of conserved repeat structure that includes cysteine residues plus specific flanking sequences, and (2) interruption of the catalytic domain by a unique hydrophobic sequence. The amino-terminal half of the catalytic domain contains the ATP-binding site and the carboxyl-terminal half includes the conserved acceptor tyrosine for autophosphorylation. The recently cloned FGF receptor may represent a fourth subclass, but this awaits further study. The mechanisms of activation of RTKs vary. The ability of ligands to induce receptor aggregation followed by endocytosis via coated pits is a feature likely shared by most RTKs and may be a triggering event for kinase activation. Subclass I and II RTKs require continuous presence of ligand (up to 8 hours) to elicit a mitogenic response of target cells, but only a short (0.5 hour) exposure to ligand is sufficient to induce DNA synthesis via the PDGF receptor (subclass III). This

160

TERRY A. WOODFORD ET AL

observation has led to the classification of RTK ligands as either inducers of competence (PDGF) or progression (EGF, IGFs). Unlike other mitogens, PDGF can stimulate DNA synthesis and cell division in fibroblasts in the absence of other factors and has the unique capability of stimulating both protein kinase C through increased phosphatidylinositol turnover and cAMP kinase through production of prostaglandin E and its activation of adenylate cyclase. Elevation of RTK catalytic activity is probably the earliest cytoplasmic effect induced upon ligand binding. A linear relationship between ligand binding and "autophosphorylation" is consistent with direct coupling of these two receptor functions. Autophosphorylation therefore appears to be an intramolecular reaction. Ligand binding resuhs in increased tyrosine kinase activity, utilizing ATP and, in some cases, GTP as the phosphate donor. The majority of autophosphorylation sites in EGF-R are located at the carboxyl-terminus or "C tail" region of the molecule, distal to the catalytic region (Tyr 1068,1198, and 1173). In contrast, such sites are scattered along the entire cytoplasmic domain of the insulin receptor (I-R), including the juxtamembrane segment, the catalytic segment, and the C-tail. Autophosphorylation of EGF-R at Tyr 1173 is likely responsible for kinase activation. The carboxyl-terminal tail of RTKs appears to play a regulatory role and may exert negative control of enzyme functions in a basal, folded conformation. Allosteric activation by ligand binding and autophosphorylation of carboxyl-terminal tyrosine residues may permit cellular proteins with SH2 domains to dock to the tyrosine phosphates of the RTKs. Interestingly, carboxyl-terminal truncations and modifications have been identified in several oncogenic versions of RTKs. The catalytic portion of RTKs displays a high level of conservation, not only when compared within the RTK family, but also when compared to other tyrosine kinases and serine/threonine-specific kinases. Receptors mutated at key positions in the tyrosine kinase domain are biologically inactive, although they retain unaltered ligand binding. For example, the nucleotide binding function can be abolished by the replacement of conserved lysines (721 and 1018 in EGF-R and I-R. respectively). Kinase-negative EGF-R mutants are defective in transduction of the mitogenic effect of EGF, the induction of c-fos and c-myc proto-oncogene expression by ligand, stimulation of Na"*"/H"^ exchange, Ca^"^ influx, and inositol trisphosphate formation. Similarly, I-R mutants at the ATP-binding site are incapable of transducing insulin activation of glucose uptake, thymidine uptake, glycogen synthesis, and S6 kinase activity. Receptor turnover, ligand internalization and degradation, and down-regulation of surface receptors are impaired in various I-R and EGF-R mutants. RTKs isolated from cells show multiple phosphorylations, implying that other protein kinases may regulate RTK activity. Treatment of cells with the phorbol ester TPA, membrane-permeable diacylglycerols, or PDGF induces phosphorylation of EGF-R on serine and threonine residues. Phosphorylation of the EGF-R in response to.these agents, presumably through activation of protein kinase C (PKC), results

Function of Protein Phosptioryiation-Deptiosphiorylation

161

in loss of high-affinity EGF binding and inhibition of EGF-stimulated RTK activity. The major site of TPA-induced EGF-R phosphorylation, Thr-654, is located ten residues from the cytoplasmic face of the plasma membrane. Mutation of EGF-R at this position results in loss of the ability of TPA to reduce EGF-stimulated tyrosine kinase activity. Other phosphorylation sites or modes of action may, however, account for modulation of receptor affinity. PDGF stimulates PIP2 breakdown and consequently causes activation of PKC, which catalyzes Thr-654 phosphorylation in the EGF receptor, leading to decreased EGF-stimulated RTK activity and attenuation of the mitogenic action of EGF (see Figure 3). This mechanism may explain the effect of other biologically active peptides such as vasopressin and bombesin that modulate EGF binding in cells. Treatment of cells with TPA also stimulates phosphorylation of I-R and IGF-I receptors on multiple serine and threonine residues. Decreased receptor kinase activity is observed to be accompanied by inhibition of high-affmity insulin binding. Since PKC activity is influenced by, and at the same time influences, signal transduction pathways involving both G protein-linked receptor-mediated hydrolysis of membrane phospholipids and growth factor receptor tyrosine kinases, it is easy to envisage pertubation of PKC expression resulting in disordered cell proliferation. The observed tumor-promoting activity of phorbol esters, which may act via persistent activation or down-regulation of PKC or both, supports such a possibility. Studies involving over-expression of PKC in cultured cells have demonstrated disordered patterns of cell growth. In addition to PKC-induced phosphorylation, both I-R and EGF-R are phosphorylated in vitro by cAMP kinase, and elevation of intracellular cAMP levels in vivo results in inhibition of ligand binding. Furthermore, phosphorylation of I-R in cells, in response to agonists of cAMP kinase, results in decreased receptor kinase activity and correlates well with the antagonistic effect of cAMP on acute metabolic responses to insulin. Because they can generate mitogenic signals, RTKs possess a latent oncogenic potential. At least two criteria must be met for the expression of RTK-mediated transforming activity: (1) activation of the tyrosine kinase, and (2) abnormal regulation of the signal-generating activity. Additional alterations in components of the normal signaling pathway may further modulate the transforming effect. A number of oncogene products actually represent altered cellular receptor tyrosine kinases (RTKs) and these oncogenic homologs provide clues to the mechanism of growth control by normal receptors and the events involved in cell transformation and tumor formation. Structural alterations caused by chromosome rearrangements of RTK genes and overexpression of certain RTKs in conjunction with autocrine ligand production may play a major role in initiation and progression of neoplasia. Increased levels of RTKs due to gene amplification and overexpression are frequently associated with certain human malignancies; i.e., amplification of the EGF-R gene and HER2/neu genes is observed in human glioblastomas and mammary carcinomas, respectively. The v-erbB oncogene of avian erythroblastosis

162

TERRY A. WOODFORD ET AL.

viruses was generated by recombination of chicken EGF-R sequences into a retroviral genome. The process by which cellular receptor sequences were transduced to create v-erbB led to deletion of most of the EGF-R extracellular domain and of 34 amino acids from the carboxyl-terminus, including the major site of autophosphorylation. In contrast, the product of v-fms oncogene of the McDonough strain of feline sarcoma virus shows homology to the macrophage colony-stimulating factor (CSF-1) receptor and has an intact extracellular domain when compared with its proto-oncogene sequence. The only obvious major structural alteration is a 40-amino acid deletion, at the C-terminus. which includes an autophosphorylation site and its replacement by 11 amino acids of the viral env (envelope) gene. Highly selective modifications in the transmembrane domain alone activate the oncogenic potential of the normal human EGF-R (HER2/neu) proto-oncogene. Replacement of Val-664 by glutamic acid or glutamine confers oncogenicity by possibly activating the tyrosine kinase in a constitutive manner. Thus, conversion of an RTK gene into an oncogene requires preservation of the complete tyrosine kinase domain and involves structural alteration(s) in noncatalytic regions. Modification on both sides of the catalytic region may be required for expression of transforming potential, which may be further enhanced by point mutations within the catalytic domain. Various growth factor receptors support autonomous cell growth by an autocrine mechanism that may result in cell transformation. For example, the v-sis oncogene of the simian sarcoma virus encodes a PDGF-like molecule which, when secreted, induces mitogenesis by interaction with the PDGF receptor. Transforming growth factor-a (TGF-a) is synthesized in a variety of human tumors, in cells transformed by RNA and DNA viruses, and in embryonic tissues. Interaction of secreted TGF-a with the EGF-R leads to tyrosine kinase activation and induces mitogenesis. The phosphorylation cascades which lead to growth stimulation mediated by growth factor receptor tyrosine kinase activity remain a mystery. Overall, the autophosphorylating activity of RTKs is much stronger than their ability to phosphorylate other cellular substrates. One potential cellular substrate which is a target for EGF and PDGF-associated RTK is p36, an abundant membrane protein that binds phospholipids and Ca^"^. p36 is structurally related to lipocortin, which inhibits phospholipase A2. Other examples are p42 (EGF, PDGF, and possibly IGF-1-stimulated phosphorylation), and p240 (identified in human epidermoid carcinoma cell lines; modified in response to insulin, IGF-1, and EGF). With respect to unique substrates, EGF-R mediates the phosphorylation of the HER2/neu proto-oncogene product and p81 (ezerin), a structural protein found in microvilli. Stimulation of I-R leads to phosphorylation of a glycoprotein called pi 20 and pi 5, which may mediate insulin-stimulated glucose uptake in adipocytes. Only recently have protein substrates of RTK-mediated phosphorylation been described with known functions. Among these are two enzymes involved in phosphoinositide metabolism. A phosphatidylinositol kinase activity, leading to phosphorylation of

Function of Protein Phosphorylation-Dephosphorylation

163

PI, PIP, and PIP2 on inositol at the 3'hydroxyl position, is tyrosine phosphorylated and activated by PDGF-R. The role of this kinase in growth factor signal transduction mechanisms is not yet clear. An isozyme of phospholipase G, PLCy, is phosphorylated on tyrosine in vitro and in vivo by EGF-R and PDGF-R, but not by I-R. This event may be involved in the regulation of PIP2 breakdown by these growth factors. Interestingly, PLCy, but not PLCp (which is not a substrate for RTK-stimulated phosphorylation), possesses a region of homology with the src family of nonreceptor tyrosine kinases. This region, the SH2 domain (src homology domain 2), is also present in the v-crk oncogene product and in GAP, a protein which stimulates the GTPase activity of p2^^^ the product of the ras family of proto-oncogenes. GAP is also a substrate for phosphorylation by PDGF-R tyrosine kinase activity. The recent discoveries that the proto-oncogene product Raf-1 is tyrosinephosphorylated by PDGF-R, both in vivo and in vitro, and that PDGF-dependent phosphorylation results in activation of the serine/threonine protein kinase activity of Raf-1, provide examples of cross-talk between the tyrosine and serine kinase systems. Activated Raf-1 then initiates the MAP kinase cascade. One of the protein kinases in this cascade is a "dual specificity" kinase that phosphorylates both tyrosine and serine threonine.

NEUROLOGICAL FUNCTION Mechanisms of Neural Transmission

When released from the terminals of a presynaptic neuron, neurotransmitters induce a change in the membrane permeability of the postsynaptic neuron by interacting with specific receptors located on the cell surface. Synaptic effects induced by neurotransmitters can be classified as mediatory or modulatory. Mediatory action involves direct linkage between transmitter binding and ion channel gating. This is illustrated by the vertebrate neuromuscular junction, where acetylcholine binds to a receptor-ionophore complex and directly mediates the flow of current through a transmitter-gated ion channel. Membrane permeability changes are brief and localized to subsynaptic regions, where they move the membrane potential towards or away from threshold. These properties make this mode of transmission suitable for signals that change rapidly. Modulatory modes of action are more suited for production of long-lasting changes in electrical behavior of postsynaptic cells. One modulatory mechanism involves second messenger-mediated transmission. The receptor for the transmitter and the ion channel it affects are separate molecules that communicate through receptor-mediated alteration in intracellular second-messenger levels. Second messengers play a key role in determining the threshold, duration, and size of the action potential of voltage-gated ion channels.

164

TERRY A. WOODFORD ET AL

The nicotinic and muscarinic acetylcholine receptors and the adrenergic receptors are examples of membrane-associated neurotransmitter receptors that undergo phosphorylation. In fact, phosphorylation of the nicotinic acetylcholine receptor has been shown to increase its rate of desensitization. Neurotransmitter effects on postsynaptic cells may be mediated by cAMP or, in some cases, by cGMP, through their corresponding kinases. cAMP-dependent protein phosphorylation has been shown to affect voltage-gated conductances in neurons and heart muscle. Neurotransmitter signaling also occurs through Ca^"*". Depolarization of excitable tissues by neurotransmitters, such as that induced by acetylcholine binding to nicotinic receptors, leads to activation of voltage-dependent Ca^"*" channels and an elevation in free Ca^^. Ca^"^ can also be elevated by action of cAMP kinase directly on the Cd?^ channel or, indirectly, by modulating K"^ channel activity which also leads to enhanced Ca^^ flux. Another mechanism for signal transduction is the hydrolysis of phosphatidylinositol bisphosphate (PIP2) to form diacylglycerol, which activates protein kinase C, and inositol trisphosphate (IP3), which stimulates the release of Ca^"^ from intracellular pools. Both Ca^'*^/CaM kinase and protein kinase C activities depend upon intracellular Ca^"*" levels. Multiple roles for cAMP kinase, protein kinase C, and Ca^"*"/calmodulin (CaM) kinase in neuronal function include: (1) mediation of neurotransmitter effects at certain types of synapses; (2) regulation of neurotransmitter biosynthesis and release; (3) mobilization of carbohydrate and lipid reserves; (4) transcriptional activation of proteins that have specific roles in neuronal physiology; and (5) regulation of cytoskeletal fiinction. Neural Phosphoproteins: Ca^VCalmodulin Kinase II Over 70 phosphoproteins have been identified in neural tissue. Protein phosphorylation appears to regulate many responses of neurons to neurotransmitters. Many neurotransmitters have been shown to elevate cAMP levels in brain by stimulation of specific neurotransmitter-sensitive adenylate cyclase. Neurotransmitter-sensitive adenylate cyclases in nervous tissue respond to substances such as dopamine, serotonin, norepinephrine, histamine, and octapine. A substantial portion of the cAMP kinase in brain exists in the membrane-bound form, and its association with synaptosomes and synaptic vesicles suggests a role in synaptic transmission. Endogenous substrates in brain tissue include the microtubule-associated protein (MAP-2), synapsins la and lb, and the Rua and Rnp subunits of cAMP kinase. The specific role of the two kinases in neuronal function is unknown, but differential expression of Rn isoforms is observed. For example, astrocytes and neurons exhibit high levels of Rup, whereas myelin-producing oligodendrocytes contain high levels of Riia- Protein III, composed of two polypeptides of 75 kDa and 55 kDa, is also a substrate for cAMP kinase and exhibits substantial immunological cross-reactivity with synapsin I. DARPP-32 (dopamine and cAMP-regulated phosphoprotein; 32 kDa) is only present in certain types of nerve cells. It is composed of a single peptide

Function of Protein Phospliorylation-Dephosplioryiation

165

which is phosphorylated by cAMP kinase. DARPP is Hkely the neural homolog of phosphatase inhibitor 1. DARPP-32 is enriched in neurons that contain dopamine receptors coupled to adenylate cyclase. CaM-kinase II, a member of the Ca^Vcalmodulin-regulated protein kinase family, is involved in a diverse array of processes in neural tissue. These processes include neurotransmitter release, catecholamine biosynthesis, cytoskeletal interactions, and glycogen metabolism (Colbran et al, 1989). This enzyme is highly concentrated in neural tissues and has been purified from mammalian brain, where it exists in both soluble and particulate forms. CaM-kinase II is a heteropolymer (isozymes vary from 300-700 kDa) comprised of structurally related subunits in the range of 50-62 kDa (a, p, P'), depending on the tissue and species. All subunits contain regulatory and catalytic functions and bind calmodulin. Substrates for CaM kinase II include the cytoskeletal proteins MAP-2 and tau, whose states of phosphorylation modify their ability to regulate microtubule assembly. Phosphorylation by CaM kinase (or cAMP kinase) reduces the ability of MAP-2 to stimulate microtubule assembly and causes dissociation of preassembled microtubules. CaM kinase II becomes fully active upon binding Ca^VCaM, and autophosphorylates just before phosphorylating substrate proteins. Autophosphorylation releases the enzyme from the stringent requirement for Ca^VCaM and converts it to an autonomous kinase. Autonomy is correlated with phosphorylation of Thr 286 in the a subunit. As autophosphorylation continues, both Ca^VCaM-dependent and Ca^VCaM-independent phosphorylating activity is eventually lost. Thus, the enzyme has a built-in shut-off mechanism that correlates with phosphorylation of a specific serine residue. Stimulation of cells by neurotransmitters, hormones, or growth factors that elevate intracellular Ca^"^ causes a transient activation of CaM kinase II. Autophosphorylation potentiates this activation by modifying "autonomy" sites while Ca^"^ is elevated. The kinase retains its activity and phosphorylates substrates for some time after Ca^"^ returns to basal levels. The kinase continues to self-phosphorylate at "inhibitory sites," reducing autonomous phosphorylation of substrates, and avoids persistent substrate phosphorylation by eventually losing activity. This process may function as a mechanism for prolonging the effectiveness of transient elevations of Ca^"^ in the brain or in other tissues. The CaM kinase associated with synaptic function may be affected somewhat differently by autophosphorylation, as the sites of autophosphorylation overlap with the CaM-binding regions of the polypeptide. CaM kinase II occurs at high levels in forebrain regions such as the cerebral cortex and hippocampus. These regions show a form of Ca^"^-dependent modulation of synaptic strength called long-term potentiation. This class of synaptic changes is thought to underlie memory formation. CaM kinase II is concentrated in the postsynaptic density, a fibrous cytoskeletal structure that adheres to the neuronal membrane at postsynaptic sites. The kinase is thus strategically positioned to respond to calcium signals generated by the binding of neurotransmitters to their

166

TERRY A. WOODFORD ET AL

receptors. A major brain substrate for CaM kinase II is synapsin I, a neuron-specific phosphoprotein localized on synaptic vesicles in nerve terminals. Synapsin I is present throughout the central and peripheral nervous system. It is composed of two polypeptides of 86 kDa (la) and 80 kDa (lb) which exhibit similar properties, including imriiunological cross-reactivity. The state of phosphorylation of synapsin I is altered by conditions that affect neuronal activity. Synapsin I is capable of complexing with and binding F-actin in the dephospho form. This property suggests a role for synapsin I in concentrating synaptic vesicles at presynaptic terminals through interaction with actin. The enrichment of synapsin I in almost all nerve terminals throughout the brain, its association with vesicles, and its phosphorylation by physiological stimuli that trigger release indicate that it may be involved in neurotransmitter release. cAMP kinase selectively phosphorylates synapsin I at a recognition site shared by CaM kinase I (site 1) with little or no apparent effect. Phosphorylation of synapsin I by membrane-bound CaM kinase II occurs at the vesicle-associated domain (sites 2 and 3). This phosphorylation abolishes its ability to bundle actin filaments and reduces the interaction between synapsin I and synaptic vesicles. Electrical stimulation of preganglionic nerves of rabbit superior cervical ganglia reduces the amount of dephospho-synapsin I. This effect occurs simultaneously with a Ca^"^ influx into the presynaptic terminal, which activates CaM kinase and leads to synapsin I phosphorylation. Depolarization of ganglia with high K"*" has a similar effect on synapsin I. Injection of synapsin I alone and in combination with CaM kinase II into the presynaptic terminal of squid giant synapse supports an inhibitory role of synapsin I on neurotransmitter release. Dephospho-synapsin I is proposed to associate with neurotransmitter-containing synaptic vesicles and cluster the vesicles in the presynaptic terminal. Synapsin I provides a constraint on neurotransmitter release by inhibiting either the movement of vesicles toward the membrane or the actual membrane fusion process. This constraint is relieved by Ca^"^ entry during membrane depolarization, a process that also functions to activate CaM kinase II, which phosphorylates synapsin I. Phosphosynapsin I dissociates from the vesicle surface, and, as Ca^"*" levels decline and synapsin I is dephosphorylated, synapsin I reassociates with vesicles to recycle the process. Evidence indicates that protein kinase C is involved in signal transduction in nervous tissue and may also modulate neuronal functions such as neurotransmitter release, membrane conductance, and potentiation or inhibition of other signaling pathways. PKC is localized in nerve endings as well as in other regions of neuronal cells, such as axons and dendrites. The y-subspecies of PKC appears to be expressed solely in the brain and spinal cord and is particularly concentrated in the hippocampus, cerebral cortex, and amygdaloid complex. Following electrical stimulation of peripheral nerves, the release of acetylcholine from cholinergic nerve endings is potentiated by the simultaneous application of the phorbol ester TPA and the Ca^"^ ionophore A23187. TPA intercalates in the membrane and substitutes for diacylglycerol, thereby activating PKC. Additional experiments designed to show the

Function of Protein Phosphorylation-Dephospliorylation

167

synergistic role off PKC and Ca^"^ mobilization extend this finding to the release of acetylcholine in central nervous system tissues. Many neural proteins have been shown to serve as substrates for PKC, but little is known about the consequence of such protein phosphorylation. Substrates include Fl protein, GABA-modulin, tyrosine hydroxylase, guanylate cyclase, and the nicotinic acetylcholine receptor. The phosphorylation of F1 protein is proposed to regulate synaptic plasticity and increase terminal axon growth after hippocampal long-term potentiation. PKC may play a role in Ca^"^ extrusion after it becomes activated, and Ca^'^-transport ATPase is another possible phosphorylation target. The role of PKC may be extended to modulation of Ca^"^ channel ftinction since microinjection of TPA or PKC itself enhances the voltage-sensitive Ca^"^ current in Aplysia bag-cell neurons. PKC may modulate membrane conductance by phosphorylating proteins that are related to channels, pumps, or transporters. Other protein kinases are involved in neuronal ftinction. Phosphorylation by cGMP-dependent protein kinase is thought to be important in certain neuronal cells. Neurotransmitters which increase cellular cGMP levels include acetylcholine (muscarinic), histamine, norepinephrine, and glutamate. G substrate, a 23 kDa polypeptide, is specifically phosphorylated by cGMP-dependent protein kinase. In its phosphorylated form, G substrate can function as a specific inhibitor of cerebellar phosphatase III. It is found exclusively in the cytoplasm of Purkinje cells in the cerebellum. Gangliosides, sialic acid-containing glycosphingolipids which exist in high concentrations in the vertebrate nervous system, are an essential component of neuronal plasma membranes and are thought to play a role in neuronal functions such as modulation of synaptic transmission, ion permeability, and flux. Gangliosides have a profound modulatory effect on protein phosphorylation, and a ganglioside-stimulated protein kinase from brain tissue has been characterized. Gangliosides may also regulate the responsiveness of PKC to diacylglycerol. Modulation of Ion Channel Activity cAMP-mediated phosphorylation events are involved in modulating the behavior of several types of ion channels, including Ca^"^ and K"*" channels (Rossie and Catterall, 1987). Ca^"^ channels are modulated by neurotransmitters and certain drugs. Modulation of Ca^"^ channels in cardiac cells by isoproterenol, a p-adrenergic agonist that increases intracellular cAMP, has been demonstrated. Electrophysiological recordings made by using the patch clamp technique to monitor singlechannel currents show that cAMP analogs increase the open-state probability of the cardiac Ca^"^ channels and thus cause channel activation. This effect on Ca^"^ channel gating probably occurs through the action of cAMP kinase, and phosphorylation of the channel itself may be the basis of cAMP-dependent regulation. PKC may also modulate cardiac Ca^^ channel functions, and K"^ channel activity appears to be regulated by cAMP kinase in cardiomyocytes. In the case of airway epithe-

168

TERRY A. WOODFORD ET AL

Hum, the C subunit of c AMP kinase mimics the effects of exogenous cAMP and p adrenergic agonists in causing the opening of chloride channels in normal, but not cystic fibrotic, cells. The nicotinic acetylcholine receptor is a neurotransmitter-regulated ion channel that mediates the depolarization of the postsynaptic membrane at the neuromuscular junction. The receptor consists of four subunits in a stoichiometry of 52ay6. The receptor from the electric eel, Torpedo californica, is phosphorylated by cAMP kinase, PKC, and a tyrosine kinase. Endogenous cAMP kinase phosphorylates the Y and 5 subunits of the receptor. cAMP kinase-catalyzed phosphorylation of the acetylcholine receptor increases its rate of desensitization (Yee and Huganir, 1987). Sodium conductance in the initial phase of the action potential in many excitable cells is mediated by voltage-sensitive sodium channels. The sodium channel purified from rat brain consists of three glycoprotein subunits, a (260 kDa), pi (36 kDa)) and p2 (33 kDa). The a subunit is also phosphorylated by cAMP kinase, possibly influencing channel gating or conductance. cAMP and cAMP kinase modulate both presynaptic and postsynaptic transmission. The monosynaptic gill-withdrawal reflex in the marine mollusc, Aplysia californica, is an example of presynaptic modulation by cAMP and provides an interesting biological system for understanding the biochemical steps which underlie a simple process of learning called sensitization (Kandel and Schwartz, 1982). In the short-term form of learning, a single sensitizing stimulus "teaches" the animal to respond more vigorously to subsequent mild tactile stimuli for a period of time ranging from minutes to hours. In Aplysia, neurological behavior is altered when the release of neurotransmitters is enhanced or diminished. Neurotransmitter release is controlled by the flow of Ca^"^ ions into synaptic terminals through channels in the membranes of nerve endings. Individual neurons can be identified and microinjected with a variety of small molecules, such as cAMP, and macromolecules, such as protein kinase and kinase inhibitors. In Aplysia sensory cells, serotonin (5-hydroxytryptamine [5-HT]) produces a slow excitatory post-synaptic potential, mediated by cAMP, that contributes to presynaptic facilitation and sensitization of the animal's gill-withdrawal reflex by increasing the effectiveness of synaptic transmission. When a noxious shock delivered to the head or tail fires a facilitory input to a sensory neuron, a serotoninsensitive adenylate cyclase is activated in the sensory cell. As cAMP levels increase, activation of cAMP kinase occurs and protein phosphorylation increases. Protein phosphorylation leads to prolonged closure of a special K"^ channel. Using the patch clamp technique to record single-channel currents from cell-attached membrane patches, both 5-HT and cAMP decrease the number of active specific K^ channels in the membrane. Depression of the K"*" channel delays repolarization of the action potential, and, consequently, Ca^"^ continues to flow into the terminal for a longer time period. The increase in duration of the Ca^"^ flux leads to nerve transmitter release and a greater response in the gill withdrawal reflex.

Function of Protein Phosphorylation-Dephospborylation

169

Experimental evidence indicates that protein phosphorylation catalyzed by cAMP kinase underlies behavioral sensitization of the Aplysia reflex. Injection of C subunit into sensory neurons simulates the effects caused by cAMP. Presynaptic facilitation can be blocked by introduction of the protein kinase inhibitor into the sensory neuron. Memory or short-term sensitization appears to be dependent upon sustained enhancement of kinase activity. Although the regulation of the overall process is not understood, adenylate cyclase is a key enzyme step in controlling the duration of transmitter release in short-term sensitization. While these results provide a molecular description of the action of cAMP kinase at the level of channel gating, the identity and localization of the phosphoprotein(s) that control channel activity are as yet unknown. Possible substrates through which cAMP kinase exerts its effects include cytoplasmic regulatory proteins, ion pumps, membrane-associated regulatory proteins and perhaps even through direct phosphorylation of the channel itself Protein Phosphorylation in Vision

Protein phosphorylation plays a role in vision. The vertebrate rod outer segment contains a number of enzymes that are activated in a light-dependent manner. Each of these enzymes is temporally linked to the photobleaching of the visual pigment, rhodopsin, located in the photoreceptor membranes of the retina. An active form of bleached rhodopsin, Rho*, catalyzes the exchange of GTP for GDP on a specific GTP-binding protein called transducin, which then activates a cGMP-specific phosphodiesterase. Dynamic changes in cGMP levels alone or in conjunction with Ca^"^ are thought to modulate ion permeability of the rod outer segment (ROS) membrane and regulate hyperpolarization of the photoreceptor cell (Hurley, 1987). The visual response of rod photoreceptor cells is initiated when light absorbed by rhodopsin isomerizes its 11-cis-retinal chromophore to the 11-trans configuration. During this bleaching process, the conformation of the retinal-opsin complex is apparently modified so that the protein moiety becomes an acceptable substrate for phosphorylation by a unique photoreceptor kinase called rhodopsin kinase. This 68 kDa soluble enzyme phosphorylates as many as nine serine and threonine residues on the hydrophobic carboxyl-terminal tail of Rho*. Rhodopsin phosphorylation is capable of producing a graded inhibition of light-stimulated PDE activation over a near-physiological range of bleach levels. A protein called arrestin binds phosphorylated Rho* with high affinity and reduces its effectiveness for transducin and PDE activation. Therefore, as a consequence of photon capture, the ability of rhodopsin to be phosphorylated increases and, thereafter, slowly decreases in the dark. Although the significance of rhodopsin kinase-catalyzed phosphorylation is uncertain, the phosphorylation of bleached rhodopsin and its slow dephosphorylation in the dark by an endogenous phosphoprotein phosphatase implicate this reversible phosphorylation process in the visual mechanism of

170

TERRY A. WOODFORD ET At.

light/dark adaptation. Rhodopsin kinase is part of a family of structurally and functionally related protein kinases which includes P-adrenergic receptor kinase (p-ARK). The latter kinase phosphorylates agonist-occupied P-adrenergic receptors, but not agonist-free receptors, at multiple carboxyl-terminal sites (Sibley et al., 1987). This kinase appears to be essential for a desensitization/adaption process that affects only stimulated receptors. The phosphorylated receptor is a less efficient activator of Gs and adenylate cyclase than is the unphosphorylated receptor. Thus, similarities exist between the function of P-ARK kinase and rhodopsin kinase. A cyclic nucleotide-activated protein kinase is also found in the retina. However, this kinase, when associated with ROS membranes, phosphorylates bleached rhodopsin in a cyclic nucleotide-independent manner.

PHOSPHOPROTEIN DEPHOSPHORYLATION Protein phosphorylation reactions are counterposed by protein phosphatase activities so that the former may serve a regulatory role in biological processes. In this way, phosphoprotein phosphatases play an integral role in modulating phosphorylation-dependent events. Our knowledge of protein phosphatases has not kept up with our understanding of protein kinases. The once popular view that the phosphorylation state of proteins is primarily under control of protein kinases has been overridden by the fact that multiple protein phosphatases have been identified which play important and dynamic roles in regulation. The cDNAs for several of these enzymes have now been cloned so that it will be possible to study their regulation at the level of gene expression. It is clear that the extent of protein phosphorylation at any moment depends on the sum and locations of the activities of relevant kinases and phosphatases. The central questions on the study of protein phosphatases concerns the number of phosphatases, their substrate specificities and the manner in which they are regulated. Serine/Threonine Phosphatases

Four enzymes which have been characterized from higher eukaryotes account for most, if not all, of the cytosolic serine/threonine-specific phosphatases described so far. These activities are divided into two categories. Type I phosphatases dephosphorylate the p subunit of phosphorylase kinase more rapidly than they do the a subunit and are potently inhibited by low concentrations of inhibitor 1 and 2. Type 2 phosphatases, of which there are three distinct catalytic subunits (2 A, 2B, and 2C), dephosphorylate the a subunit of phosphorylase kinase more rapidly than they do the P subunit and are essentially unaffected by the inhibitor proteins. The four phosphatases have been identified in all tissues studied thus far, although their relative concentration and intracellular distributions vary considerably. The activities of serine and threonine phosphatases are known to be regulated by the

Function of Protein Pliosptioryiation-Depliosphorylation

171

phosphoprotein substrate level, by specific inhibitor proteins, by Ca^VCaM, and by phosphorylation. Type I protein phosphatase (PP-1) from higher eukaryotes is involved in regulating major metabolic pathways such as glycogen metabolism, muscle contraction, and protein synthesis. Several forms of this enzyme are known to exist. The enzyme consists of a 37 kDa catalytic subunit whose activity is regulated through interaction with various subunits. Regulation of mammalian PP-1 is complex. The catalytic subunit can be bound to either of two inhibitors, 1 or 2. The potency of inhibitor 1 can be modulated by phosphorylation that is catalyzed by cAMP kinase. In addition to the phosphorylation of inhibitor 1 by the C subunit, the R subunit of cAMP kinase has also been shown to inhibit phosphoprotein phosphatase in vitro (Khatra et al., 1985). In skeletal muscle, the phosphatase catalytic subunit can interact with a glycogen-binding subunit or a myosin-binding subunit. These associations, also controlled by phosphorylation, target PP-1 activity to glycogen particles or to myofibrils, respectively, thereby enhancing the dephosphorylation of glycogen-metabolizing enzymes or myosin. In muscle, the 70 kDa form of the enzyme exists as an inactive complex consisting of a 38 kDa catalytic subunit and a 31 kDa regulatory subunit identified as inhibitor-2. This particular enzyme is one of the major phosphatases involved in the inactivation of phosphorylase and activation of glycogen synthase. As previously mentioned, it dephosphorylates the P subunit of phosphorylase kinase as well as the R subunit of the type II cAMP kinase. One physiological means of activation of this phosphatase requires Mg-ATP and an activating factor subsequently identified as glycogen synthase kinase 3 (GSK3). The exact stoichiometry of phosphorylation is not certain since the phosphatase undergoes self-dephosphorylation. The 31 kDa regulatory subunit is required for both activation and inactivation, so its function extends beyond that of a simple inhibitor; it is also considered to be a modulator of enzyme activity. Phosphatase 1 has been shown to play a role in mitosis and cell-cycle regulation. The bimG ("blocked in mitosis") gene from Aspergillis encodes a protein which is 86% identical to the catalytic subunit of phosphatase 1. The dis2"*" gene, from the fission yeast, Schizosaccharomyces pombe, which regulates chromosome segregation also encodes a protein highly homologous to mammalian phosphatase 1. A role for this enzyme in cell-cycle regulation, specifically in the initiation of M phase, has also been inferred from studies with Xenopus oocytes. Injection of PP-1 inhibitors slows the induction of meiosis in oocytes by progesterone and, in cell-free systems, slows the posttranslational activation of MPF activity that accompanies the first meiotic division (Cyert and Kirschner, 1988). This is further supported by the observation that a yeast homolog of mammalian PP-1, encoded by the bwsl gene, functions in mitotic control. It has been proposed that the cyclin component of the pp34^'^^^-cyclin complex might be a substrate of the bwsl"^ phosphatase.

172

TERRY A. WOODFORD ET AL.

Protein phosphatase 2A (PP-2A) is a cytosolic enzyme found in a number of tissues, including skeletal muscle. It can be resolved into three forms, all of which contain the same 36 kDa catalytic (C) subunit bound to other components. The A subunit (60 kDa) is also identical in each enzyme form, whereas the B subunit (54-55 kDa) differs or is absent. An important characteristic of these enzymes is their ability to be activated by polycations, such as polylysine, and polyamines, such as spermine. The function of PP-2 A is unclear but may play a role in the control of key enzymes regulated by phosphorylation in pathways such as glycolysis, glycogen metabolism, gluconeogenesis, and aromatic acid breakdown. PP-2 A also plays a role in regulating processes that are coupled to the cell cycle. For example, the SIT4 gene in Saccharomyces cerevisiae is 54% identical to mammalian PP-2 A, and mutations in SIT4 are temperature-sensitive for growth. Also, PP-2A from animal cells may be required for T antigen-dependent initiation of SV40 viral DNA replication. Protein phosphatase 2B (PP-2B) is a Ca^Vcalmodulin-dependent phosphatase of narrow substrate specificity and is identical to the major soluble calmodulinbinding protein of brain extracts, calcineurin. PP-2B consists of a 61 kDa catalytic (A) subunit and a 19.2 kDa Ca^"^-binding (B) subunit that is structurally related to calmodulin. The binding of Ca^"*" to the B subunit is essential for enzyme-substrate interaction and, alone, stimulates phosphatase activity slightly. Further binding of Ca^VCaM to the A subunit stimulates activity 10-fold. PP-2B activity is therefore regulated by two Ca^'^-binding proteins. The function of this enzyme also remains to be elucidated, but it is prevalent in brain. PP-2B synthesis peaks during the period of synapse formation. Its localization at postsynaptic densities and dendritic microtubules suggests a role in synaptic function. It may also function as a means by which signals transduced by Ca^"^ may attenuate the effects of hormones which act through cAMP via the dephosphorylation ot inhibitor 1 in muscle or of its brain homolog, DARPP. Protein phosphatase 2C was thought to be a monomeric enzyme of 45 kDa but may actually exist as a dimer. It is completely dependent upon Mg^"^ for activity, and mechanism(s) for its regulation are as yet unknown. Tyrosine Phosphatases

The dephosphorylation of tyrosine in proteins was first detected in the cell membranes of A431 cells, a human epidermoid carcinoma cell line that over-produces the EGF receptor. Treatment of these cells with EGF stimulates tyrosine phosphorylation of the receptor and is followed by slow dephosphorylation. Cultured cells transformed with temperature-sensitive mutants of Rous sarcoma virus have provided another system for study. Phosphotyrosine phosphatases (PTPases) are distinct enzymes which dephosphorylate proteins at specific phosphotyrosyl residues. Some of these enzymes

Function of Protein Phosphorylation-Dephosphorylation

173

exhibit dual specificity, i.e., they will dephosphorylate either protein tyrosine phosphate or protein serine/threonine phosphate. PTPases do not appear to be at all structurally related to the serine/threonine phosphatases. Some members of this enzyme family are now being characterized in order to understand their role in biological functions. For example, injection of the soluble PTPase IB into Xenopus oocytes delays maturation induced by insulin and blocks insulin-stimulated S6 phosphorylation. Likewise, a receptor-linked PTPase known as CD45 has been shown to affect the activation of T and B lymphocytes when conjugated to cell-surface antigens such as CD2, CD3, CD4, and CD9 and to alter interleukin-2 receptor expression. Expression of CD45 in lymphocytes catalyzes the dephosphorylation of p56*^^, a lymphocyte-derived membrane tyrosine kinase, at a specific tyrosine residue. This dephosphorylation subsequently leads to its activation. Recently, a PTPase has been isolated from rat spleen which dephosphorylates and inactivates a tyrosine-specific protein kinase and may be involved in its regulation (Swarup and Subrahmanyan, 1989). Considerable progress has been made in the purification and characterization of PTPases from human placenta, rat spleen, rabbit kidney, bovine brain, and human hematopoietic cells. In addition, endogenous heat-stable proteins which inhibit certain PTPases derived from brain tissue have been reported. A major limitation in the studies of PTPases has been the identification of natural substrates; therefore, artificial substrates have been used. These include the synthetic acidic copolymer poly Glu-Tyr (4:1) bovine serum albumin, casein, histone, tubulin, angiotensin, and IgG. Another complication is that weak non-specific tyrosine dephosphorylation activities have been detected in serine/threonine specific types 2A, 2B, and particularly 2C phosphatases as well as in alkaline and acid phosphatases under certain conditions. Two forms of the enzyme are being intensively studied; a 37 kDa enzyme from human placenta called PTPase IB and a high molecular weight (180-200 kDa) transmembrane-linked glycoprotein, designated the leukocyte common antigen, CD45, found in certain lymphocytes (Charbonneau et al., 1989). The complete amino-acid sequence of placental PTPase IB is known, and the sequences of human, rat, and mouse CD45 have been deduced from cDNAs isolated from lymphocyte libraries. CD45 represents a broad family of proteins which differ in their extracellular amino-terminal structure. Interestingly, CD45 contains in its cytoplasmic segment two catalytic domains, I and II, which have sequence identity with PTPase IB. It appears that either the first catalytic domain only is active, or that the second catalytic domain demonstrates a different substrate specificity. A related protein called LAR (leukocyte antigen-related protein) resembles CD45 and PTPase IB in its cytoplasmic domains but has an external segment containing immunoglobulin-like features and structural repeats similar to those occurring in the extracellular domains of fibronectin and neural cell adhesion molecule (NCaM). The transmembrane proteins CD45 and LAR have extracellular domains characterized by ligand-binding structures, suggesting that these represent recep-

174

TERRY A. WOODFORD ET AL.

tor-linked PTPases, although, at present the ligands are unknown. It may be the case that such receptor-linked PTPases undergo homeotypic interactions and are important in cell-cell or cell—matrix adhesion. These particular PTPases may have a function opposite to that of receptor-linked tyrosine kinases. Clearly, the identification of ligands for receptor-linked PTPases and the cellular substrates for PTPases in general will be most interesting.

CONCLUDING REMARKS This chapter has focused on the regulation of cellular function by phosphorylation/dephosphorylation at various levels of biological activity. Several protein kiases have been delineated, some of which specifically modify serine and/or threonine residues on substrate proteins and some of which phosphorylate tyrosine residues. Considering that they are in the same evolutionary family, it is not surprising that the different protein kinases share certain kinetic properties and even protein substrate specificities. Some protein kinases show coordinate regulation while still maintaining uniqueness by their novel effectors and by their recognition of specific phosphorylation sites in proteins. Phosphoprotein phosphatases counteract the effects of protein kinases and are themselves essential to cellular function. The examples of phospho-Zdephospho regulation summarized herein were mainly based upon those occurring in eukaryotes. It should be mentioned that in bacterial systems, reversible protein phosphorylation, although mediated by biochemically different means, plays a role in processes such as bacteriophage infection, sugar transport, regulation of metabolite flow, chemotaxis, and expression of genes involved in nitrogen assimilation, porin synthesis, and (i-glucoside utilization. Regulation in bacteria often involves autophosphorylation of a key component on a specific histidine residue and the subsequent transfer of this phosphoryl group to other components. The bacterial phosphotransferase systems utilize ATP similarly to those systems in eukaryotic cells, but may also use metabolic intermediates such as phosphoenolpyruvate. In some cases, phosphorylation and dephosphorylation reactions are catalyzed by a single bifunctional enzyme, e.g., isocitrate dehydrogenase kinase/phosphatase from E. coli. An interesting phenomenon that is evident in eukaryotic cells is "cross talk" or interaction between second messenger systems and signaling pathways. Some protein kinases will catalyze phosphorylation of other kinases; e.g., certain serine/threonine kinases will phosphorylate particular tyrosine protein kinases. Cyclic AMP and Ca^"*" signaling systems are often closely interlinked, allowing them to act synergistically or antagonistically, depending upon the target cell. The existence of calmodulin-dependent adenylate cyclase and cAMP phosphodiesterase implies that the level of cAMP can be regulated by Ca^"^. Conversely, the level of Ca^"*" can be regulated by cAMP. For example, phosphorylation of proteins in cardiac sarcoplasmic reticulum by cAMP kinase stimulates Ca^'^-sequestration by these vesicles and may contribute to the increased rate of heart relaxation caused by

Function of Protein Piiosptioryiation-Depliosplioryiation

175

epinephrine. In neural systems, cAMP stimulates the opening of Ca^"^ channels by activating cAMP kinase. At least two calmodulin-dependent enzymes are targets for cAMP kinase. Cyclic AMP-dependent and calmodulin-dependent kinases often phosphorylate common proteins although usually at distinct sites. Such proteins include glycogen synthase, tyrosine hydroxylase, phospholamban, and synapsin I. Some hormones are capable of utilizing both second-messenger systems. For example, epinephrine can bind to at least three types of receptors, a j , a2, and p, causing stimulation of PIP2 hydrolysis, inhibition of adenylate cyclase, and activation of adenylate cyclase, respectively. Sharing of substrates between PKC and cAMP kinase also indicates an interaction between the two signal-transducing systems. The nature of interaction between the cAMP-generating and phosphoinositide degradation pathways also depends upon the cellular system in which the two pathways operate. In some systems, such as endocrine, lymphoma, or fibroblast cells, PKC potentiates cAMP formation, suggesting synergism between the two pathways. In ovarian granulosa or glioma cells, PKC inhibits the adenylate cyclase system, while in platelets, neutrophils, myoblasts, or lymphocytes, cAMP produces an inhibitory effect on phosphatidylinositol turnover and decreases PKC activity. The inactivation, desensitization, and/or down-regulation of various hormone, neurotransmitter, and growth factor receptors by kinase-catalyzed phosphorylation reactions, illustrates a means by which communication between kinases can take place. The complexity of phospho-dephospho regulation of biological ftinction is apparent, and an understanding of the intricate mechanisms whereby this regulation occurs presents quite a scientific challenge.

ACKNOWLEDGMENTS We are very grateful to Mrs. Penny Stelling for typing the manuscript and to Dr. John Shabb for preparing some of the figures.

REFERENCES Bauerle, P.A., Lenardo. M., Pierce, J.W., & Baltimore, D. (1988). Phorbol ester-induced activation of the NF-KB transcription factor involves dissociation of an apparently cytoplasmic NF-kB/Inhibitor complex. Cold Spring Harbor Symposia on Quant. Biol. 53, 789-799. Charbonneau, H., Tonks, N.K., Kumar, S., Diltz, CD., Harrylock, M., Cool, D.E., Krebs, E.G., Fischer, E. H., & Walsh, K. A. (1989). Human placenta protein tyrosine phosphatase: Amino acid sequence and relationship to a family of receptor-like proteins. Proc. Natl. Acad. Sci. USA 86, 5252—5256. Colbran, R.J., Schworer, CM., Hasimoto, Y., Fong, Y.-L., Rich, D.P., Smith, M.K., & Soderling, T.R. (1989). Calcium/calmodulin-dependent protein kinase II. Biochem. J. 258, 313-325. Collett, M.S., Erickson, E., & Erickson, R.L. (1979). Structural analysis of the avian sarcoma virus transforming protein-sites of phosphorylation. J. Virol. 29, 770-775. Corbin, J.D., Sugden, P.H., West, L., Flockhart, DA., Lincoln, T.M., & McCanhy, D. (1978). Studies on the properties and mode of action of the purified regulatory subunit of bovine heart adenosine 3':5'-monophosphatedependent protein kinase. J. Biol. Chem. 253, 3997-4003.

176

TERRY A. WOODFORD ET AL.

Cyert, M. & Kirschner, M. (1988). Regulation of MPF activity in vitro. Cell 53, 185-195. Draetta, B., Lura, F., Westerderf, J., Brizuela, L., Ruderman, J., & Beach. D. (1989). cdc2 protein kinase is complexed to both cyclin A and B: Evidence for proteolytic inactivation of MPF. Cell 56, 829-838. Exton, J.H. (1990). Signalling through phosphatidylcholine breakdown. J. Biol. Chem. 265, 1^. Francis, S.H., Noblett, B.D., Todd, B.W., Wells, J.N., & Corbin. J.D. (1988). Relaxation of vascular and tracheal smooth muscle by cyclic nucleotide analogs that preferentially activate purified cGMP-dependent protein kinase. Molecular Pharmacology 34, 506-517. Gould, K.L. & Nurse, P. (1989). Tyrosine phosphorylation of the fission yeast cdc2 + protein kinase regulates entry into mitosis. Nature 342, 39-45. Hunter, T. & Sefton, B.M. (1980). Transforming gene product of Rous sarcoma virus phosphorylates tyrosine. Proc. Natl. Acad. Sci. USA 77, 1311-1313. Hurley, J.B. (1987). Molecular properties of the cGMP cascade of vertebrate photoreceptors. Ann. Rev. Physiol. 49, 793-812. Imagawa, M., Chiu, R., & Karin, M. (1987). Transcription factor AP-2 mediates induction by two different signal-transduction pathways: Protein kinase C and cAMP. Cell 51, 251-260. Kandel, E.R. & Schwartz, J.H. (1982). Molecular biology of an elementary form of learning: Modulation of transmitter release by cAMP. Science 218, 433-436. Khatra, B.S., Printz, R., Cobb, C.E., & Corbin, J.D. (1985). Regulatory subunit of cAMP-dependent protein kinase inhibits phosphoprotein phosphatase. Biochem. Biophys. Res. Comm. 130, 567— 573. Krebs, E.G. & Beavo, J.A. (1979). Phosphorylation-dephosphorylation of enzymes. Ann. Rev. Biochem. 48, 923-955. Nairn, A.C. & Palfi-ey, H.C. (1987). Identification of the major Mr 100,000 substrate for calmodulindependent protein kinase III in mammalian cells as elongation factor-2. J. Biol. Chem. 262, 17299-17303. Nakagawa, J., Van Der Ahe, D., Pearsen, D., Hemmings, B.A., Shibahara, S., & Nagamine, Y. (1988). Transcriptional regulation of a plasminogen activator gene by cyclic AMP in a homologous cell-free system. J. Biol. Chem. 263, 2460-2468. Nishizuka, Y. (1988). The molecular heterogeneity of protein kinase C and its implications for cellular regulation. Nature 334, 661-665. Palen, E. & Traugh, J.A. (1987). Phosphorylation ofribosomalprotein S6 by cAMP-dependent protein kinase and mitogen-stimulated S6 kinase differentially alters translation of globin MRNA. J. Biol. Chem. 262, 351^-^3523. Robinson-Steiner, A.M. & Corbin, J.D. (1986). Protein phosphorylation in the heart. In: The Heart and Cardiovascular System (Fozzard, H.A., Haber, E., Jennings, R.B., Katz, A.M., & Morgan, H.A., eds.), pp. 887-910. Raven Press, New York. Rossie, S. & Catterall, W.A. (1987). Cyclic AMP-dependent phosphorylation of voltage-sensitive sodium channels in primary cultures of rat brain neurons. J. Biol. Chem. 262, 12735-12744. Sagata, N., Watanabe, N., Vande Woude, G.F., & Ikawa, Y. (1989). The c-mos proto-oncogene product is a cytostatic factor responsible for meiotic arrest in vertebrate eggs. Nature 342, 512-518. Shenoy, S., Choi, J.-K., Bagrodia, S., Copeland, T.D., Mailer, J.L., & Shalloway, D. (1989). Purified maturation promoting factor phosphorylates pp60*^'*"^ at the sites phosphorylated during fibroblast mitosis. Cell 57, 775-786. Short, I.M., Wynshaw-Boris, A., Short, H.P., & Hanson, R.W. (1986). Characterization of phosphoenolypyravate carboxykinase (GTP) promoter regulatory region. II. Identification of cAMP and glucocorticoid regulatory domains. J. Biol. Chem. 261, 9721-9726. Sibley, D.R., Benovic, J.L., Caron, M.G., & Lefkowtiz, R.J. (1987). Regulation of transmembrane signalling by receptor phosphorylation. Cell 48,913-922.

Function of Protein Phosphorylation-Dephosphorylation

177

Swamp, G. & Subrahmanyam, G. (1989). Purification and characterization of a protein-phosphotyrosine phosphatase from rat spleen which dephosphorylates and inactivates a tyrosine-specific protein kinase. J. Biol. Chem. 264, 7801-7808. Wolfe, L., Corbin, J.D., & Francis, S.H. (1989). Characterization of a novel isozyme of cGMP-dependent protein kinase from bovine aorta. J. Biol. Chem. 264, 7734—7741. Yamamoto, K.K., Gonzalez, G.A., Biggs, W.H., & Montminy, M.R. (1988). Phosphorylation-induced binding and transcriptional efficacy of nuclear factor CREB. Nature 334, 494-499. Yarden, Y. & Ullrich, A. (1988). Growth factor receptor tyrosine kinases. Ann. Rev. Biochem. 57, 443-478. Yee, G.H. & Huganir, R.L. (1987). Determination of the sites of cAMP-dependent phosphorylation on the nicotinic acetylcholine receptor. J. Biol. Chem. 262, 16748-16753.

RECOMMENDED READINGS Beebe, S.J. & Corbin, J.D. (1986). Cyclic nucleotide-dependent protein kinases. In: The Enzymes, 3rd ed., (Krebs, E.G. & Boyer, P.D., eds.), p. 43 Vol. 17, Academic Press, Orlando, Florida, Cohen. P. & Cohen, P.T.W. (1989). Protein phosphatases come of age. J. Biol. Chem. 264, 2143521438. Gettys, T.W. & Corbin, J.D. (1990). The protein kinase family of enzymes. CRC Critical Reviews (In press). Greengard, P. (1978). Phosphorylated proteins as physiological effectors. Science 199, 146-152. Hunter, T. & Cooper, J.A. (1985). Protein tyrosine kinases. Annu. Rev. Biochem. 54, 897-930. Murray, A.W. (1989). The cell cycle as a cdc2 cycle. Nature 342, 14-15. Roesler, W.J., Vandenbark, G.R., & Hanson, R.W. (1988). Cyclic AMP and induction of eukaryotic gene transcription. J. Biol. Chem. 263, 9063-9066. Tonks, N.K. & Charbonneau, H. (1989). Protein tyrosine dephosphorylation and signal transduction. Trends in Biochemical Sciences 14,497-500. Ullrich, A., Riedel, H., Yarden, Y., Coussens, L., Cray, A., Dull, T., Schlessinger, J., Waterfield, M.D., & Parker, P.J. (1986). Protein receptors and protein kinase C. Cold Spring Harbor Symposia on Quantitative Biology. Vol. LI, 713-724.

This Page Intentionally Left Blank

Chapter 5

The Family of Protein Tyrosine Phosphatases and the Control of Cellular Signaling Responses NICHOLAS K.TONKS

Introduction Structural Diversity in the PTPase Family Structural Features of the PTPase Family Receptor-like Transmembrane PTPases Cytoplasmic Nontransmembrane PTPases Further Aspects of the Regulation of PTPase Activity Physiological Roles for PTPases: A Role for These Enzymes as Negative Effectors of Signal Transduction In Vivo PTPases as Positive Mediators of Cellular Signaling Responses The Dual Specificity Phosphatases Summary and Perspectives

179 181 182 182 185 187 188 191 193 196

INTRODUCTION Every eukaryotic cell is surrounded by a plasma membrane which forms a barrier to the outside world. However, the cell cannot ignore its environment and must be Principles of Medical Biology, Volume 4 Cell Chemistry and Physiology: Part II, pages 179-199 Copyright © 1996 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-806-4

179

180

NICHOLAS K. TONKS

able to respond to external stimuli. This is largely achieved through specific proteins that span the membrane. Binding of an effector molecule to the extracellular segment of one of these transmembrane receptor proteins elicits a response in the intracellular milieu. One of the principal features that underlies such signal transduction responses is protein phosphorylation—3, reaction catalyzed by protein kinases in which the y phosphate of ATP is transferred primarily to the hydroxyl groups of seryl, threonyl, or tyrosyl amino acid residues in a target protein. The addition of phosphate alters the conformation of the target protein and thus its function. Therefore, through the activation of a protein kinase, triggering phosphorylation with concomitant modulation of the activity of intracellular target substrates, a cell can be induced to respond to extracellular signals. For example, the transmembrane receptors for many growth factors possess intrinsic protein tyrosine kinase (PTK) activity that is stimulated by ligand binding. In response to ligand the activated receptor PTK then phosphorylates a spectrum of substrates that includes the initiation of a cascade of serine/threonine phosphorylation events, culminating in gene transcription in the nucleus and production of the proteins essential for mediating the growth response. Therefore, growth factor binding initiates a complex network of protein phosphorylation events that lead a quiescent cell to enter the cell cycle, undergo DNA replication and ultimately divide. Similarly, such receptor PTKs are involved in differentiation processes in which a precursor cell is driven to assume a specialized phenotype and to cease dividing, for example, in the generation of neurons. The potential for disruption of such systems to cause disease is obvious. For example, aberrant activation of a PTK, so that the cell is constantly being stimulated to divide in an uncontrolled fashion, has been associated with a number of cancers. The receptor for insulin is also a PTK and defects in its function may contribute to type II diabetes. Protein phosphorylation, and of particular interest, protein tyrosine phosphorylation, plays crucial roles in the control of fundamental cellular processes including growth, division, and differentiation. However, before we can begin to pinpoint dysfunctional situations in disease states, we must first understand how these various signaling responses are controlled in normal cells. Much research effort has focussed on this problem from the perspective of the protein kinases. However, in my laboratory we are taking a different approach. It is important to realize that within the cell phosphorylation is not static; it is a reversible dynamic process. Thus, the net level of phosphate in a target substrate reflects the balance between the competing action of the kinase that phosphorylates it and enzymes that catalyze the reverse reaction, the protein phosphatases (Figure 1). In my lab we focus on defining physiological roles for tyrosine phosphorylation but from the perspective of the enzymes that catalyze dephosphorylation, that is, the protein tyrosine phosphatases (PTPases).

181

Protein Tyrosine Phosphatases

PTK

PTPase Figure T. Protein phosphorylation is a reversible process in vivo. It is inriportant to realize that the net level of phosphate in a target substrate reflects the balance betv^een the competing action of the kinase that phosphorylates it and the phosphatase that catalyzes the dephosphorylation reaction. This chapter addresses the physiological functions of members of the PTPase family.

STRUCTURAL DIVERSITY IN THE PTPASE FAMILY Early in the study of the control of glycogen phosphorylase by reversible serine phosphorylation, it was proposed that the major regulation was exerted at the level of the kinase. From there the idea that protein dephosphorylation was catalyzed by a small number of protein phosphatases, serving as constitutively active housekeeping enzymes, pervaded the whole field. However, nothing could be further from the truth. Since protein tyrosine phosphorylation was first observed over 10 years ago, considerable progress has been made in characterizing the PTKs (reviewed in Fantl et al., 1993). The study of PTPases, until recently, lagged significantly behind largely for technical reasons. A major problem in the study of protein phosphatases in general is the requirement for a suitably purified and phosphorylated substrate to measure enzyme activity. In the early days of tyrosine phosphorylation this issue was compounded by the fact that physiologically relevant phosphotyrosyl-containing proteins were largely unidentified. Hence artificial substrates had to be generated. One such protein, reduced carboxyamidomethylated and maleylated (RCM) lysozyme was developed for the assay of PTPase activity in vitro and then used to complete the first purification to homogeneity of a novel, phosphotyrosyl-specific protein phosphatase (Tonks et al., 1988b,c). An enzyme termed PTP IB was isolated in homogeneous form. The amino acid sequence of PTP IB was determined and homology with CD45, a transmembrane glycoprotein that is a surface marker of hematopoietic cells was demonstrated (Charbonneau et al., 1988). This result was important because of the structure of CD45. It possesses a tandem array of two PTPase domains in its intracellular segment, a single transmembrane domain and a highly glycosylated, cysteine-rich extracellular segment that displays the hall-

182

NICHOLAS K. TONKS

marks of a ligand binding motif. Thus, the exciting possibility was raised that CD45 may represent a prototype for receptor-linked PTPases with the capacity to play a direct role in modulating cellular signaling responses. This was further strengthened by the demonstration of intrinsic PTPase activity in CD45 (Tonks et al., 1988a). What followed has been an explosion of interest in the PTPases and now a wide variety of these enzymes have been identified in sources as diverse as mammals, Drosophila, C. elegans, Dictyostelium, yeast, and viruses (reviewed in Charbonneau and Tonks, 1992). PTPases have even been detected in prokaryotes. In fact, it has recently been shown that the essential virulence determinant of the bacterium Yersinia, the causative agent of the plague or Black Death, is a PTPase termed Yop, the activity of which is essential for virulence. Apparently, following infection the Yop PTPase dephosphorylates tyrosyl residues in proteins of the eukaryotic host cell, thus disrupting normal cell function (reviewed in Guan and Dixon, 1993)

STRUCTURAL FEATURES OF THE PTPASE FAMILY We know now that the PTPases rival the PTKs in their structural diversity and complexity. Each PTPase contains at least one conserved segment of-240 residues which is assumed to delineate an independently folding catalytic domain. Within this domain is the sequence motif [IA^]HCXAGXXR[S/T]G (where X can be any amino acid) that uniquely defines the PTPase family of enzymes. The cysteinyl residue, which is absolutely conserved, is involved in the formation of a thiophosphate covalent intermediate as part of the catalytic mechanism. Although the PTPases display a high degree of similarity in their catalytic domains, they can be readily differentiated on the basis of the structure of their noncatalytic segments. Most strikingly, like the PTKs, the PTPases can be subdivided into receptor-like and nontransmembrane isoforms. The observation of receptor-like PTPases is important because it predicts a potential to convert extracellular signals directly into intracellular responses through extracellular ligand-mediated dephosphorylation of intracellular target proteins.

RECEPTOR-LIKE TRANSMEMBRANE PTPASES The distinguishing features of the receptor-like PTPases include the presence of a single transmembrane segment and a tandem array of two PTPase domains within the cytoplasmic tail. Thus far, six receptor-like PTPases have been found to contain a single catalytic domain the prototype being HPTPP, however these remain the minority. The significance of the two-domain arrangement that characterizes the receptor-PTPases is unclear. Although there is the obvious potential for cooperative interactions between domains, it remains a point of controversy as to whether in fact both domains, in particular domain II, are active. In some cases it has been proposed that the function of domain II is to control the specificity of domain I. Nevertheless, it seems clear that, at least for CD45 and PTPa, domains I and II both

Protein Tyrosine

183

Phosphatases

possess intrinsic activity, but differ in their substrate specificity and possibly in their mode of regulation. In contrast to the similarity of their intracellular domains, the receptor-like PTPases can be readily distinguished by their extracellular segments. They may be arbitrarily divided into five subtypes on this basis (Figure 2). One might anticipate that the variation in the extracellular segments of these enzymes may reflect a corresponding diversity in the ligands to which they respond. Until recently, a function had not been ascribed to the extracellular segment of any one of these enzymes, thus they remained putative receptors. It was originally suggested that a B-cell surface protein, CD22, interacted with the smallest, 180 kDa form of CD45. However, subsequently it has become clear that CD22 is a lectin that binds generally to sialoglycoproteins including all forms of CD45. The significance of this association remains unclear, although recent observations suggest that binding of CD22 to CD45 amplifies (i.e. positively regulates) early signals through the T cell receptor (Sgroi et al., 1995). In addition, it was demonstrated recently that the HPTPC

I

LAR

•

HPTPn

DPTP

CD45 H

Ml!

Type!

Type II

DPTP99A HPTPa

I Type III

Type IV

1

Extracellular Intracellular

TypeV

Figure 2. Representatives of the receptor-like PTPases. The conserved catalytic domains are shown in black. At the present time the receptor-like species can be subdivided into five types based on the structure of their extracellular segments. Type I represents the CD45 family, multiple isoforms of which arise from differential splicing of a primary mRNA transcript of a single gene; three exons encoding sequences at the extreme N-terminus (horizontal lines) are differentially expressed. Type II contain immunoglobulin-like (diagonal lines) and tandem fibronectin type Ill-like repeat domains (stippled); this category includes among others LAR (leukocyte common antigen-related), DLAR, DPTP, and HPTP^i. Type ill bear multiple fibronectin type Ill-like repeats. Some type III isoform, such as HPTPp, have only one internal PTP domain. Type IV isoforms such as a and 8 have small glycosylated extracellular segments. Type V possesses an N-terminal motif with homology to carbonic anhydrase.

184

NICHOLAS K. TONKS

carbonic anhydrase-like domain in the extracellular segment of PTP(!;/p on glial cells binds specifically to contactin, a glyosylphosphatidylinositol-anchored cell recognition protein on the surface of neurons. Interaction between this domain in PTP(^/p and contactin promotes neuronal adhesion and the induction of neurite outgrowth and differentiation, suggesting that such interactions may underlie the triggering of unidirectional or bidirectional signaling responses between neurons and glial cells during development (Peles et al., 1995). We have approached this issue of ligands for one of the type II receptor PTPases, PTPjLi, and have now demonstrated that this enzyme has the potential to function in cell-cell adhesion reactions (Brady-Kalnay et al., 1993). The extracellular segment of PTP|LI contains one immunoglobulin (Ig)-like and four fibronectin (FN) type Ill-like domains, and thus displays structural similarity to members of the Ig superfamily of cell adhesion molecules that includes NCAM, the neural cell adhesion molecule. NCAM mediates homophilic adhesion, i.e., NCAM on one cell binds to NCAM on an adjacent cell (see Edelman & Crossin, 1991 for review). To investigate whether PTPji could also serve in cell-cell adhesion, we have followed the strategy of expressing the protein in normally nonadhesive cells to test whether this induces their aggregation. We have generated recombinant baculoviruses that express various forms of PTP^ in Sf9 insect cells and have demonstrated that expression of full length PTPfa, or mutants bearing an intact extracellular segment, whether in the presence or absence of PTPase domains, induced aggregation. However, expression of the catalytic segment of PTP|LI as a soluble protein or in a chimeric molecule fused to the extracellular and transmembrane domains of the EOF receptor was ineffective in inducing aggregation. The degree of aggregation of Sf9 cells induced by PTPja expression is very similar to that observed following expression of other well characterized adhesion molecules in various model systems. Furthermore, we have shown that PTPjn mediates aggregation via a homophilic mechanism. But perhaps most importantly, we have also shown that PTPji, as it is normally expressed on the surface of a lung cell line (MvLu cells), retains the capacity for homophilic binding interactions. Thus, these results suggest that the ligand for this transmembrane PTPase is another PTPji molecule on the surface of an adjacent cell and provided the first indication of the function of the extracellular segment of one of these enzymes. The precise physiological role of such binding interactions remains to be defined but it is tempting to speculate that PTPases such as PTPji may be involved in the control of cell proliferation. As cells grow and divide in culture, they spread out over the surface of the petri dish until eventually forming a uniform, or confluent, layer in which adjacent cells are touching each other. In normal cells this promotes a response termed "contact inhibition of cell growth." When one considers that activation of receptor PTKs, triggering a tyrosine phosphorylation response, can promote cell growth, perhaps activation of a PTPase, triggering tyrosine dephosphorylation, contributes to the mechanism for such growth inhibitory phenomena.

Protein Tyrosine Ptiospitatases

185

Homophilic association between extracellular segments of receptor PTPases, such as PTP|Li, brought into contact as adjacent cells touch in a confluent monolayer, may promote such tyrosine dephosphorylation events. Whether PTPji can itself drive cell-cell association at its normal levels of expression, i.e., function as an adhesion molecule under physiological conditions, remains unclear. Perhaps cell-cell association facilitated by other adhesion molecules serves to juxtapose the extracellular segments of PTP|LI molecules for binding. Furthermore, to date, no direct effect of ligand binding to the extracellular segment of PTPfi on the activity of the intracellular PTPase domains has been detected. Nevertheless, such homophilic binding interactions could serve a tethering role, controlling the activity of the PTPase indirectly by restricting its spatial distribution on the membrane, and thus restricting the spectrum of substrates with which it may interact, and perhaps bringing it into close proximity with an appropriate target. A structural feature that characterizes PTPjLi may be important in this regard; its juxtamembrane domain (that between the transmembrane and first PTPase domains) is longer than the equivalent segment in other receptor PTPases and displays homology to the intracellular segments of members of the cadherin family of cell adhesion molecules. The intracellular segment of the cadherins, which is essential for their adhesive function, interacts indirectly with cortical actin through the binding of proteins termed catenins. Thus, the association of the cadherin cytoplasmic domain, the domain that displays homology to PTPJLI, with catenins is essential for both the cytoskeletal binding and adhesive function of cadherins. We are currently testing whether protein:protein interactions mediated by this segment in PTPju serve to target the phosphatase to particular cytoskeletal structures. Most recently we have obtained evidence to indicate that in a variety of tissues and cell lines PTPju is present in adherens junctions in a complex with cadherins and catenins (Brady-Kalnay et al., 1995). Our data suggest that within these structures PTPja may be one of the enzymes that regulates the dynamic tyrosine phosphorylation and thus the function of the cadherin/catenin complex in vivo. Therefore rather than functioning as an adhesion molecule itself, the physiological role of PTPju may be to regulate other adhesion systems in response to homophilic binding. In transformed, i.e., cancerous, cells the contact inhibitory response breaks down and cells grow and divide essentially without restraint. It will be of interest to determine whether alterations in PTPases such as PTP|Li may contribute to the transformed phenotype.

CYTOPLASMIC NONTRANSMEMBRANE PTPASES Unlike their receptor-like counterparts, nonreceptor PTPases have a single catalytic domain and unrelated, noncatalytic sequences of variable length at either their Nor C-termini (Figure 3). Sequence similarities to other well characterized proteins and, in some cases experimental evidence, suggest that one of the primary functions of these noncatalytic segments may be to control enzyme activity. The concept that a major factor in controlling PTPase activity may be to restrict enzyme location to

186

NICHOLAS K. TONKS P T P I C ^ MEG 2 TC-PTP -PTP liii

B

^

il "-^ PTPH1 PTP1B

PEP n 97 g^

\ \ STEP

PTP1 n Y0P2b

n Cdc25

PTP^PEST

Figure 3. Representatives of the nontransmembrane/cytoplasmic PTPases. The conserved catalytic domains are shown in black. Many of the nonreceptor PTPases bear noncatalytic segments that are structurally related to other well characterized proteins. The position and relative size of these noncatalytic domains are shown as boxes containing distinct symbols; noncatalytic regions that have similar sequences are designated with identical patterns. The noncatalytic segments that have been identified include two SH2 domains in PTPl C, and band 4.1 homology domains in PTPH1, an apparent lipid-binding domain in MEG2, and segments containing PEST sequences in PEP and PTP-PEST. In TC-PTP and PTPIB, the C-terminal noncatalytic segments appear to play a role in modulating activity and controlling subcellular localization. PTP1, STEP, and Yop2b have noncatalytic sequences that are apparently unrelated to sequences in the databases. The protein from Vaccinia virus, VH1, is much smaller than the other PTPs and presumably encodes only essential sequences within the catalytic domain. VHl differs from the other members of the family in that it displays dual specificity, dephosphorylating PSer as well as PTyr.

particular regions of the cell is currently gaining ground. An example of this targeting concept is provided by PTPHl (Yang and Tonks, 1991). The cDNA for PTPHl predicts a protein with a large noncatalytic N-terminal segment containing a domain of-320 residues that displays homology to the superfamily of proteins that includes band 4.1, talin, ezrin, moesin, and the product of the tumor suppressor gene that causes type 2 neurofibromatosis. The presence of this domain apparently defines a family of proteins that are targeted to interfaces between the plasma membrane and the actin cytoskeleton, suggesting that the subcellular location of the PTPase will be similarly restricted. In addition, two closely related but distinct PTPases, one restricted to hematopoietic cells, and the other expressed ubiquitously, have been identified that bear an N-terminal segment containing two SH2 domains (see Neel, 1993). These domains comprise -100 amino acid residues and are found in a number of proteins that are involved in mediating early cellular effects of growth factor stimulation (Pawson and Schlessinger, 1993). SH2 domains bind to phosphotyrosyl residues in sequence specific contexts. In the case of these PTPases, the presence of the SH2 domains leads to their recruitment into receptor PTK signaling complexes in a manner that is dependent upon ligand binding and receptor autophosphorylation. There is evidence to indicate that, once bound to the receptor PTK, the PTPase becomes activated and presumably it then dephosphorylates other members of the signaling complex whose identity remains to be

Protein Tyrosine Phosphatases

187

established. Also these PTPases are themselves phosphorylated on tyrosyl residues in response to ligand-dependent activation of receptor PTKs. Whether this phosphorylation is part of the activation mechanism or even creates new docking sites for proteinrprotein interactions among signaling molecules awaits elucidation. There are also instances of cytoplasmic PTPases in which the noncatalytic segments do not display homology to known motifs but their structure nonetheless implies a targeting function. For example, in PTPIB there is a C-terminal extension of--120 residues following the catalytic domain that serves a regulatory function. Analysis of the distribution of charged and hydrophobic residues in this sequence illustrates a predominantly hydrophilic segment of 80-90 residues followed by a highly hydrophobic segment, with a hydrophobic index close to that observed for transmembrane domains in receptor proteins, comprising the extreme carboxyterminal ~35 residues. This hydrophobic stretch of residues has been shown to be both necessary and sufficient for targeting PTPIB to the cytoplasmic face of membranes of the'endoplasmic reticulum (Frangioni et al., 1992). A closely related enzyme termed TCPTP, which displays a similar distribution of charged and hydrophobic residues in its regulatory C-terminal segment (although the absolute sequence identity with this segment in PTPIB is only ~20%), is also targeted to the membranes of the endoplasmic reticulum. There are even cases where alternative splicing of a PTPase transcript may generate forms of the enzyme that contain the same catalytic domain but which bear different targeting sequences and thus display different subcellular locations. For example, DPTP61F from Drosophila undergoes alternative splicing to generate two nontransmembrane PTPases, each of-61 kDa, that differ in their extreme C-termini. One displays a highly hydrophobic C-terminus and is found associated with cytoplasmic membranes, while the other bears a segment rich in basic residues, conforming to a classic nuclear localization signal, and is found exclusively in the nucleus (McLaughlin and Dixon, 1993). Similar data, in which alternatively spliced forms display different subcellular locations, have now also been obtained for TCPTP (Lorenzen et al, 1995).

FURTHER ASPECTS OF THE REGULATION OF PTPASE ACTIVITY Initial measurements of the activity of PTPIB and CD45 indicated a very high Vmax, 10-1,000-fold in excess of the PTKs in vitro (Tonks et al, 1988c, 1990). Thus, the PTPases have the potential to represent a formidable barrier to the action of PTKs in vivo. This suggests that mechanisms must exist to control PTPase activity so as to permit the normal function of PTKs. The importance of harnessing PTPase activity in vivo has been vividly demonstrated by experiments in which the cytoplasmic enzyme TCPTP was overexpressed in BHK cells (Cool et al., 1992). Overexpression of the full length 48 kDa form of the enzyme, with an intact targeting, regulatory C-terminal segment, did not produce an obvious phenotype. In contrast, overexpression of a truncated 37 kDa form of the enzyme, from which

188

NICHOLAS K. TONKS

the regulatory segment had been deleted and which was now constitutively active and no longer targeted to intracellular membranes, produced a catastrophic effect on the cells. Only a small number of cell lines overexpressing the enzyme were obtained and these displayed a multinucleated phenotype, in which nuclear division was asynchronous. This results from a defect in cytokinesis. Clearly, an uncontrolled PTPase can play havoc in the cell, stressing the importance of characterizing and understanding these control mechanisms. As discussed above many PTPases possess structural motifs suggestive of regulation through intracellular targeting. Also for the receptor-like enzymes there is the potential for modulation of activity by ligand binding. A further tier to the regulation of PTPase activity is the potential for control by reversible phosphorylation of the enzymes themselves. Alteration in the concentration of intracellular free Ca^"*" by treatment of T-cells with ionomycin leads to a decrease in the phosphorylation of CD45 coincident with a decrease in its PTPase activity (Ostergaard and Trowbridge, 1991). More recently, it has been demonstrated that PTP1B is subjected to multisite phosphorylation in vivo, being a point of convergence for the action of at least three distinct Ser/Thr kinases, including PKC and p34^^^^ (Flint et al., 1993). The sites of phosphorylation are found in the hydrophilic portion of the regulatory C-terminal segment of PTP IB. A complex reorganization of phosphate within the molecule accompanies the transition into mitosis in HeLa cells. These and other examples point to the importance of phosphorylation as a regulatory device for the control of PTPase activity. For both the receptor-like and nontransmembrane enzymes, phosphorylation and control of intracellular location, regulating activity indirectly by restricting subcellular distribution, are therefore important concepts that apply to the control of PTPase activity in vivo. Such proteiniprotein interactions may not only illustrate new tiers of control of cellular phosphotyrosine levels and thus signal transduction pathways, but may also eventually point to new families of potential oncogenes and anti-oncogenes.

PHYSIOLOGICAL ROLES FOR PTPASES: A ROLE FOR THESE ENZYMES AS NEGATIVE EFFECTORS OF SIGNAL TRANSDUCTION IN VIVO The initial characterization of several PTPases highlighted the fact that their specific activity in vitro far exceeded that of the PTKs. This implied that the PTPases may exert a dominant influence on PTKs in vivo (Figure 4). This idea was reinforced by the observation that treatment of NRK-1 cells with vanadate, an inhibitor of PTPases, produced enhanced levels of phosphotyrosine and the generation of a transformed morphology (Klarlund, 1985). These observations led to the hypothesis that some PTPases may be the products of tumor suppressor genes, i.e., their deletion or mutation may contribute to the elevation in phosphotyrosine levels that is associated with certain neoplasias. As expected, overexpression of certain PTPases has been shown to suppress or even revert the transformed

189

Protein Tyrosine Phosphatases

Autophosphorylation

<J= PTPase

( inactive )

PTPase Figure 4, PTPases may act negatively to antagonize PTK-induced signaling events in vivo. Since PTKs catalyze the phosphorylation of tyrosyl residues in proteins and PTPases the dephosphorylation of such residues, one can easily see that these enzymes may play antagonistic roles. Many PTKs require autophosphorylation of tyrosyl residues In the kinase Itself for optimal activity. Therefore, one could envisage that an inhibitory effect of PTPase may be exerted through the dephosphorylation and Inactivation of either the PTK or its target substrate. Reproduced from Sun and Tonks (1994).

phenotype associated with oncogenic PTKs such as Fms, Neu, or Src (see Cool and Fischer, 1993). The notion of PTPases as tumor suppressor genes has triggered a considerable research effort on the part of many groups to examine the possibility of coincidence between the chromosomal localization of various PTPase genes and sites of abnormality associated with cancer. At the present time some leads look promising, in particular, the case of RPTPy (LaForgia et al, 1991). RPTPy is a receptor-like PTPase characterized by the presence of an N-terminal segment of ~270 residues that displays homology to carbonic anhydrase. However, only one of the three His residues that are involved in ligating the catalytically essential Zn^"^ ion in carbonic anhydrase is conserved. Therefore, it has been suggested that rather than catalyzing the hydration of CO2, this carbonic anhydrase-like domain in RPTPy may function as a hydrophobic binding pocket for a low Mr ligand. The gene for RPTPy is located on human chromosome 3p21, a region that is frequently deleted in renal and lung carcinomas. A more detailed characterization of the

190

NICHOLAS K. TONKS

RPTPy gene in 31 human lung cancer cell lines did not detect any abnormalities in the segment encoding the intracellular catalytic domains. However, a homozygous deletion has been detected within the RPTPy gene in all derivatives of murine L cells, a connective tissue line which produces sarcomas in syngeneic mice. This deletion removes residues 12-145 from the full length protein and encompasses a segment of the potential ligand-binding carbonic anhydrase domain (Wary et al., 1993). This shorter transcript has been detected in L cell mRNA. Thus, the intriguing possibility exists that in certain tumors a truncated form of RPTPy may be expressed which, ahhough it still possesses functional PTPase domains, may not be able to respond normally to its cognate ligand and thus may not trigger a normal signaling response. Recent studies have highlighted a growth suppressive function for a cytoplasmic PTPase, the SH2 domain-containing PTPase that is expressed exclusively in hematopoietic cells. This enzyme has been given several names, depending upon the group working on it, including HCP, PTPIC, SHP, and SHPTP-1. Mutations in the gene for HCP which result in aberrant splicing of the transcript have been shown to be the cause of the motheaten {me) phenotype in mice (Shultz et al., 1993). Homozygous me mice display severe immunodeficiency and systemic autoimmune disease, and generally only live for 2—3 weeks. In these mice there is a single base deletion in the HCP gene that results in the production of a severely truncated polypeptide of 102 amino acids that is completely devoid of the PTPase domain and most of the SH2 domain sequences. There is a less severe form of the disease, termed motheaten viable {me^\ in which the animals survive for a few months. This results from a distinct mutation that generates protein with sequences either inserted into or deleted from the catalytic domain. This severely impairs catalytic function so that the PTPase activity of we" mutant HCP is reduced by -80% relative to wild type. The broad spectrum of hematopoietic abnormalities associated with the motheaten phenotype suggests that HCP may be a negative regulator of several PTK signaling pathways. For example, there is hyperproliferation of macrophages in these mice that is independent of CSF-1. Therefore, one might anticipate that under normal conditions HCP can exert a negative influence on signaling pathways initiated by the CSF-1 receptor PTK. As might be expected, lysates of macrophages from motheaten mice display an increase in tyrosine phosphorylation of several proteins relative to the wild type. In addition it has been demonstrated that HCP/SH-PTPl plays a major role in downregulation of signaling through the erythropoietin receptor (EPOR) (Klingmuller et al., 1995). HCP/SH-PTPl binds to a site of tyrosine phosphorylation in the intracellular segment of the EPOR (Y429) through its SH2 domains and once recruited into the complex it dephosphorylates and inactivates the JAK2 PTK that associates with the EPOR and is responsible for transmitting the EPO signal. Following expression of Y429F mutant EPORs, in which the docking site for HCP/SH-PTPl has been converted to phenylalanine and which can no longer bind the phosphatase, cells become hyper-

Protein Tyrosine Phosphatases

191

sensitive to EPO and display prolonged EPO-induced autophosphorylation and activation of JAK2. Thus in general the loss of HCP most likely leads to sustained tyrosine phosphorylation with consequent enhanced proliferation, in a manner analogous to that observed with constitutively activated oncogenic PTKs. The possibility that HCP may contribute to human disease, perhaps also being a tumor suppressor gene, is currently under investigation. It is interesting to note that deletions and translocations of human chromosome 12pl2-13, the map position of the HCP gene, are found in patients with acute lymphocytic leukemia.

PTPASES AS POSITIVE MEDIATORS OF CELLULAR SIGNALING RESPONSES Although these and other examples point to a role for PTPases in suppressing tyrosine phosphorylation-dependent signaling pathways, it seems clear that the PTPases cannot simply be regarded as antagonists of the PTKs. Several PTPases have now been shown to act positively in mediating signaling responses. In Drosophila, a mutation termed corkscrew causes abnormal development of the terminal head and tail structures during embryogenesis. This developmental pathway is believed to involve a cascade of phosphorylation events triggered by the activation of a transmembrane PTK termed torso. Cloning of the corkscrew gene revealed that it encodes an SH2 domain-containing PTPase more closely related to the ubiquitously expressed mammalian enzyme (which is termed Syp, SHPTP2, or PTPID) than the hematopoietic cell-specific HCP molecule that underlies the murine motheaten phenotype discussed above (Perkins et al, 1992). Through a variety of genetic approaches it has been demonstrated that corkscrew functions positively in conjunction with a Ser/Thr kinase, D-ra/, in transducing the torso signal. Similarly Syp, the mammalian homolog of corkscrew, is thought to play a role in potentiating the signal from growth factor receptor PTKs. Thus, at present, it appears that two closely related SH2 domain-containing PTPases possess a similar organization of structural motifs, yet serve very different physiological functions, one promoting the other antagonizing PTK-induced signaling pathways. One of the best characterized examples of a positive role for a PTPase in signal transduction involves CD45, the prototype receptor PTPase. CD45 represents a family of proteins expressed exclusively on nucleated hematopoietic cells and which can occupy up to 10% of the surface of a lymphocyte. Multiple isoforms of CD45 have been identified that differ due to the variable expression of three exons encoding sequences at the extreme N-terminus of the protein. This leads to differences in protein sequence and patterns of glycosylation. Different CD45 isoforms are expressed in a highly regulated, cell type-specific fashion. In addition, individual T-cells can express more than one isoform in a pattern that varies with cell activation. These observations imply an important role for this PTPase in lymphocyte fiinction, but only recently has the mechanistic basis of that function become apparent. A breakthrough came with the generation of T-cells that failed

192

NICHOLAS K. TONKS

to express CD45 (reviewed in Woodford-Thomas and Thomas, 1993). It was shown that CD45~T-cell clones failed to proliferate in response to antigen or to cross-linking of CD3, but did respond normally to IL-2. Subsequently, a number of CD45deficient T-cell lines were developed that were shown to be defective in T-cell receptor-induced signaling responses. Thus, unlike the normal parental lines, these CD45" cells did not respond to T-cell receptor stimulation with the release of Ca^"^ from intracellular stores, or with the turnover of phosphatidylinositol that leads to the production of lipid-derived second messengers. Perhaps most striking was the observation that in CD45~ Jurkat cells there was no T-cell receptor-induced increase in cellular phosphotyrosine. These signaling defects could be rescued by restoring expression of the PTPase by transfecting back CD45 cDNA. Thus CD45 is required for coupling stimulation of the T-cell receptor to activation of these signaling pathways, indicating that CD45 plays a positive role in mediating signal transduction in response to T-cell activation. A similar situation is now known to be true of B-cells. These data present something of a paradox: ablation of expression of a PTPase results in the failure of T-cell receptor stimulation to elicit an enhancement in the level of phosphotyrosine. The resolution of this paradox has come from the observation that the Src- family kinases, p56^^^ (found in association with the T-cell surface accessory proteins CD4 or CDS) and p 5 9 ^ (found in association with the CD3 complex that interacts with the T-cell receptor) are abnormally regulated in CD45~ cells. Members of the Src family of PTKs have two major sites of tyrosine phosphorylation. There is a site of autophosphorylation in the catalytic domain and a site at the C-terminus, equivalent to Tyr 527 in Src itself, that when phosphorylated exerts an inhibitory effect on kinase activity. In CD45" cells, these kinases accumulate in a form in which the C-terminal inhibitory site is hyperphosphorylated, and thus activity is repressed, although the relative importance of p56^^^ and p59*^ to a particular pathway varies with cell type. Therefore the signaling defect in CD45~ cells appears to be a failure to activate Src family kinases efficiently. Thus, under normal conditions, CD45, a PTPase, can actually promote tyrosine phosphorylation through the dephosphorylation and activation of a Src-family PTK (Figure 5). Such a concept reinforces the importance of intracellular compartmentalization in the control of PTPase activity to prevent inappropriate dephosphorylation of the downstream targets of the activated Src-family kinase. A similar interaction has been observed between Src itself and the receptor-like PTPase PTPa, and implies a role for this PTK/PTPase couple in neuronal differentiation (see Pallen, 1993). PTPa is characterized by a small, highly glycosylated extracellular segment of 123 amino acid residues. It is expressed ubiquitously and its levels are enhanced during neuronal differentiation of two embryonal carcinoma (EC) cell lines and in neuroblastoma cells, with maximal expression preceding the morphological change to the neuronal phenotype. Overexpression of PTPa in PI 9 EC cells led to a change in the normal pattern of retinoic acid-induced differentiation in vitro. The wild type cells normally differentiate into endodermal or mesoder-

Protein Tyrosine Phosphatases

193 PTPase

Signal inhibited

^ Signal

PTK Figure 5. PTPases may act positively to promote signaling events in vivo. One mechanism by v^hich PTPases may promote signaling responses is through the dephosphorylation and activation of members of the Src family of PTKs. Src family kinases have an inhibitory site of phosphorylation in their C-termini dephosphorylation of which by a PTPase promotes kinase activity and triggers the signaling function. Therefore by this mechanism a PTPase may actually promote tyrosine phosphorylation. Such a system requires strict subcellular compartmentalization to prevent the PTPase from also acting on the downstream targets of the Src family PTK. Reproduced from Sun and Tonks (1994).

mal cells; however, PTPa-overexpressors display a neuronal phenotype following retinoic acid treatment. These effects of PTPa coincide with the dephosphorylation of Tyr 527 in Src and activation of its kinase function (den Hertog et al., 1993). Similarly, overexpression of PTPa in rat embryo fibroblasts results in the dephosphorylation and activation of Src, and is accompanied by the generation of a transformed morphology. The possibility that PTPases may antagonize the effects of oncogenic PTKs and may represent the products of tumor suppressor genes seems intuitively obvious. However, current evidence that increasingly points towards positive roles for certain PTPases in promoting cellular signaling responses raises the possibility that such PTPases may actually be the products of oncogenes. In the future their activation as a result of gene amplification or translocation events may be recognized as underlying certain proliferative disease states.

THE DUAL SPECIFICITY PHOSPHATASES The PTPases that I have described thus far share a common homology domain of -240 residues that contains the unique signature motif that bears the active site cysteinyl residue and defines this family of enzymes. These PTPases are absolutely

194

NICHOLAS K. TONKS

specific for phosphotyrosyl residues in proteins and, at least in vitro, are somewhat promiscuous in their recognition of PTyr-containing substrates. Recently, a large sub-family of dual specificity phosphatases has been uncovered of which VH1 from Vaccinia virus was the first to be identified (Guan et al., 1991). Further examples have been detected in several pox viruses, yeast, mammalian cells, and even in a prokaryote, the cyanobacterium Nostoc commune (reviewed in Guan and Dixon, 1993). These enzymes differ from the classical PTPases in that although they contain the [I/V]HCXAGXXR[S/T]G signature motif they display little other structural similarity with this enzyme family. Furthermore, in many cases, they have been demonstrated to dephosphorylate both Tyr and Ser/Thr residues in proteins. Perhaps most strikingly these enzymes tend to display a very restricted substrate specificity in vivo. At least two of these enzymes have been shown to catalyze dephosphorylation events that are of fundamental importance to the control of cell function. Tyrosine phosphorylation plays a crucial role in controlling the onset of mitosis in eukaryotic cells (see Atherton-Fessler et al., 1993 for review). The key cell cycle regulator p34^^^^, which is a protein Ser/Thr kinase, is phosphorylated on adjacent residues, Thr 14 and Tyr 15, in the GXGXXG motif of the nucleotide binding loop. This phosphorylated form of p34^^^^, which accumulates during interphase, is inactive. Dephosphorylation of Thr 14 and Tyr 15 in p34^^^^ is a key step in activating this protein kinase and driving the transition into mitosis. Both of these dephosphorylation events are catalyzed by a single dual specificity phosphatase, pgQcdc25^ and this mechanism of initiation of mitosis is highly conserved across all eukaryotic species. As observed for the classical PTPases, phosphorylation of Ser/Thr residues in its noncatalytic N-terminal segment and targeting to defined intracellular locations, as well as changes in its mRNA and protein levels, have all been implicated in the control of pSO^'^^^^ function in various cell systems. Recent attention has focused on a widely-expressed dual specificity phosphatase that is the product of a murine immediate early gene, 3CH134. Both the mRNA transcript and the protein derived from this immediate early gene are induced rapidly and transiently following growth factor stimulation of quiescent cells. In addition, in human fibroblasts, this phosphatase has been recognized as the product

Figure 6. Signaling pathways involving MAP kinase. A common feature of the signaling pathways induced by diverse mitogenic stimuli is the activation of MAP kinases. The molecular details of the steps leading to MAP kinase activation are currently being elucidated. This figure summarizes the current appreciation of the cellular responses to activation of a growth factor receptor PTK by binding of its cognate ligand. Autophosphorylation of the receptor creates docking sites for the assembly of multiprotein complexes. These complexes include the receptor itself, adaptor proteins, such as GRB2, and nucleotide exchange factors, such as SOS, which trigger the accumulation of the active, GTP-bound form of Ras. (continued)

Growth Fbctor

V

Growth Fbctor Receptor iosma membrane

C

ircs GDP 1 — ^

n

MAPKKK (Rcf)

MAPKK

Figure 6, (continued) The SeryThr kinase Raf is activated in response to accumulation of Ras-GTP. Although Ras and Raf physically interact this association is not sufficient for activation of Raf and the molecular intricacies of this activation event remain to be elucidated. Raf phosphorylates and activates MAP kinase kinase, which in turn phosphorylates both Tyr and Thr regulatory sites in MAP kinase with concomitant activation. Active MAP kinase is thought to promote transcription of genes required for the growth response either directly, by phosphorylation of transcription factors, or indirectly, through the phosphorylation and activation of another Ser/Thr kinase, p90'^'^, which may in turn phosphorylate transcription factors.

195

196

NICHOLAS K. TONKS

of a stress response gene, the expression of which is induced by stimuli such as heat shock. Complementary DNA has been cloned and the enzyme expressed and characterized. It possesses intrinsic phosphatase activity that is highly specific for MAP kinase in vitro. The MAP kinases have been implicated as common components of signaling pathways induced by diverse mitogenic stimuli. For example, growth factor-induced autophosphorylation of receptor PTKs creates specific binding sites for proteins possessing SH2 domains (reviewed in Pawson and Schlessinger, 1993) (Figure 6). One such class of proteins contains both SH2 and SH3 domains which serve as adaptors to link stimulation of a PTK to activation of Ras. Thus the adaptor protein, GRB2, binds to a particular site of autophosphorylation in the receptor PTK through its SH2 domain, and to a nucleotide exchange protein, SOS, through its SH3 domains. The SH3 domains in GRB2 recognize specific proline-rich motifs in SOS. Assembly of this multiprotein complex in the membrane triggers an SOS-induced conversion of Ras from an inactive GDPbound form to an active GTP-bound state. Activated Ras then initiates a cascade of sequential phosphorylation events in which the Ser/Thr kinase Raf phosphorylates and activates MAP kinase kinase (also known as MEK) which is a dual specificity kinase that in turn phosphorylates both Thr 183 and Tyr 185 in MAP kinase. Phosphorylation of both Tyr and Thr regulatory sites is essential for MAP kinase activation. Once activated, MAP kinase can phosphorylate a number of substrates, including transcription factors, that are essential for triggering the expression of genes that are required for the mitogenic response. One such gene is 3CH134 that encodes the dual specificity phosphatase. Data have now been presented to show that this enzyme not only dephosphorylates MAP kinase in vitro but also dephosphorylates both the Tyr and Thr regulatory sites and inactivates MAP kinase in vivo (Sun et al., 1993). Furthermore, this enzyme appears to be absolutely specific for MAP kinase and is now called MKP-1 (MAP kinase phosphatase) to reflect this restricted specificity. Therefore, this phosphatase may feed back on the mitogenic signaling pathway by dephosphorylating and inactivating MAP kinase, and thus attenuate the signaling response and prevent uncontrolled growth and proliferation. In addition, the high degree of specificity shown by MKP-1 suggests that it may prove to be a powerful tool for defining physiological roles for MAP kinase.

SUMMARY AND PERSPECTIVES This is an exciting time to be working on the protein tyrosine phosphatase family of enzymes. Some 50 members of this family have been identified to date and although important progress remains to be made in identifying novel PTPases and determining their primary structure, we are now also able to move on towards confronting the challenge of elucidating the biological function of these enzymes. The picture has changed from the initial view that PTPases would merely serve as passive antagonists of PTK function. We now know that the PTPases will rival the

Protein Tyrosine Phosphatases

197

kinases both in their structural diversity and complexity, and in the sophistication of their regulation. Perhaps most importantly, the PTPases also have the potential to play an active role in promoting signaling events. Clearly continued study of this family should provide important new insights into the physiological importance of tyrosine phosphorylation.

ACKNOWLEDGMENTS Although I have cited some original papers in this article, due to space restraints I have largely referred to reviews to provide the reader with a source of additional information. I apologize to those authors whose primary papers are not included. Work in my lab is supported by grants from the National Institutes of Health/National Cancer Institute (CA53840 and CA64593), from The Council for Tobacco Research and from the Mellon Family, Hansen Memorial and Lauri Strauss Leukemia Foundations. I am also supported by a Pew Scholarship in the Biomedical Sciences. I thank Carol Marcincuk for typing the manuscript.

REFERENCES Atherton-Fessler, S., Hannig, G., & Piwnica-Worms, H. (1993). Reversible tyrosine phosphorylation and cell cycle control. Sem. Cell Biol. 4, 433-442. Brady-Kalnay, S.M., Flint, A.J., & Tonks, N.K. (1993). Hemophilic binding of PTP^, a receptor-type protein tyrosine phosphatase, can mediate cell-cell aggregation. J. Cell Biol. 122, 961-972. Brady-Kalnay, S.M. & Tonks, N.K. (1995). Receptor protein tyrosine phosphatase PTP|LI associates with cadherins and catenins in vivo. J. Cell Biol. 130, 977-986. Charbonneau, H. (1992). 1002 protein phosphatases? Ann. Rev. Cell Biol. 8, 463-493. Charbonneau, H., Tonks, N.K., Walsh, K.A., & Fischer, E.H. (1988). The leukocyte common antigen (CD45): A putative receptor-linked protein tyrosine phosphatase. Proc. Natl. Acad. Sci. USA 85, 7182-7186. Cool, D.E., Andreassen, P.R, Tonks, N.K., Krebs, E.G., Fischer, E.H., & Margolis, R.L. (1992). Cytokinetic failure and asynchronous nuclear division in BHK cells overexpressing a truncated protein-tyrosine-phosphatase. Proc. Natl. Acad. Sci. USA 89, 5422-5426. Cool, D.E. & Fischer, E.H. (1993). Protein tyrosine phosphatases in cell transformation. Sem. Cell Biol. 4, 443-453 den Hertog, J., Pals, C.E.G.M., Peppelenbosch, M.P., Tertoolen, L.G.J., de Laat, S.W., & Kruijer, W. (1993). Receptor protein tyrosine phosphatase a activates pp60'^'^'^ and is involved in neuronal differentiation. EMBO J. 12, 3789-3798. Edelman, G.M. & Crossin, K.L. (1991). Cell adhesion molecules: Implications for a molecular histology. Ann. Rev. Biochem. 60, 155-190. Fantl, W.J., Johnson, D.E., & Williams, L.T. (1993). Signaling by receptor tyrosine kinases. Ann. Rev. Biochem. 62, 453-481. Flint, A., Gebbink, M.F.B.G., Franza, B.R.J., Hill, D.E., & Tonks, N.K. (1993). Multi-site phosphorylation of the protein tyrosine phosphatase, PTPIB: Identification of cell cycle regulated and phorbol ester stimulated sites of phosphorylation. EMBO J. 12, 1937-1946. Frangioni, J.V., Beahm, P.H., Shifrin, V., Jost, C.A., & Neel, B.G. (1992). The nontransmembrane tyrosine phosphatase PTP-IB localizes to the endoplasmic reticulum via its 35 amino acid C-terminal sequence. Cell 68, 545-560. , Guan, K., Broyles, S.S., & Dixon, J.E. (1991). A Tyr/Ser protein phosphatase encoded by vaccinia virus. Nature 350, 359-362.

198

NICHOLAS K. TONKS

Guan, K-L. & Dixon, J.E. (1993). Bacterial and viral protein tyrosine phosphatases. Sem. Cell Biol. 4, 389-395. Klarlund, J. (1985). Transformation of cells by an inhibitor of phosphatases acting on phosphotyrosine in proteins. Cell 41, 707-717. Klingmuller, U., Lorenz, U., Cantley, L.C., Neel, B.G., & Lodish, H.F. (1995). Specific recruitment of SH-PTPl to the Erythropoietin receptor causes inactivation of JAK2 and termination of the proliferative signals. Cell 80, 729-738. LaForgia, S., Morse, B., Levy, J., Bamea, G., Cannizzano, L.A., Li, P., Nowell, P.C., Boghosian-Sell, L., Glick, J., Weston, A., Harris, C.C., Drabkin, H., Patterson, D., Croce, CM., Schlessinger, J., & Huebner, K. (1991). Receptor protein tyrosine phosphatase y is a candidate tumor suppressor gene at human chromosome region 3p21. Proc. Natl. Acad. Sci. USA 88, 5036-5040. Lorenzen, J.A., Dadabay, C.Y., & Fischer, E.H. (1995). COOH-terminal sequence motifs target the T cell protein tyrosine phosphatase to the ER and the nucleus. J. Cell Biol. 131, 631-643. McLaughlin, S. & Dixon, J.E. (1993). Alternative splicing gives rise to a nuclear protein tyrosine phosphatase in Drosophila. J. Biol. Chem. 268, 6839-6842. Neel, B.G. (1993). Structure and function of SH2-domain containing tyrosine phosphatases. Sem. Cell Biol.4,41SM30. Ostergaard, H.L. & Trowbridge, LS. (1991). Negative regulation of CD45 protein tyrosine phosphatase activity by ionomycin in T cells. Science 253, 1423-1425. Pallen, C.J. (1993). The receptor-like protein tyrosine phosphatase a: A role in cell proliferation and oncogenesis. Sem. Cell Biol. 4,403-408. Pawson, T. & Schlessinger, J. (1993). SH2 and SH3 domains. Curr. Biol. 3,434-441. Peles, E., Nativ, M., Campbell, P. L., Sakurai, T., Martinez, R., Lev, S., Clary, D.O., Schilling, J., Bamea, G., Plowman, G.D., Grumet, M., & Schlessinger, J. (1995). The carbonic anhydrase domain of receptor tyrosine phosphatase p is a functional ligand for the axonal cell recognition molecule contactin. Cell 82, 251-260. Perkins, L.A., Larsen, L, & Perrimon, N. (1992). Corkscrew encodes a putative protein tyrosine phosphatase that functions to transduce the terminal signal from the receptor tyrosine kinase torso. Cell 70, 225-236. Shultz, L.D., Schweitzer, P.A., Rajan, T.V., Yi, T., Ihle, J.N., Matthews, R.J., Thomas, M.L., & Beier, D.R. (1993). Mutations at the murine motheaten locus are within the hematopoietic cell proteintyrosine phosphatase (Hcph) gene. Cell 73, 1445-1454. Sgroi, D., Koretzky, G.A., & Stamenkovic, L (1995). Regulation of CD45 engagement by the B-cell receptor CD22. Proc. Natl. Acad. Sci. USA 92,4026-4030. Sun, H., Charies, C.H., Uu, L.F., & Tonks, N.K. (1993). MKP-1 (3CH134), an immediate eariy gene product, is a dual specificity phosphatase that dephosphorylates MAP kinase in vivo. Cell 75, 487-493. Tonks, N.K., Charbonneau, H., Diltz, CD., Fischer, E.H., & Walsh, K.A. (1988a). Demonstration that the leukocyte common antigen CD45 is a protein tyrosine phosphatase. Biochemistry 27, 86958701. Tonks, N.K., Diltz, CD., & Fischer, E.H. (1988b). Purification of the major protein-tyrosine phosphatases of human placenta. J. Biol. Chem. 263,6722-6730. Tonks, N.K., Diltz, CD., & Fischer, E.H. (1988c). Characterization of the major protein tyrosine phosphatases of human placenta. J. Biol. Chem. 265, 6731-6737. Tonks, N.K., Diltz, CD., & Fischer, E.H. (1990). CD45, an integral membrane protein tyrosine phosphatase. J. Biol. Chem. 265, 10674-10680. Wary, K.K., Lou, Z., Buchberg, A.M., Siracusa, L.D., Druck, T., LaForgia, S., & Huebner, K. (1993). A homozygous deletion within the carbonic anhydrase-like domain of the Ptprg gene in murine L-cells. Cancer Res. 53, 1498-1502. Woodford-Thomas, T. & Thomas, M.L. (1993). The leukocyte common antigen, CD45, and other protein tyrosine phosphatases in hematopoietic cells. Sem. Cell Biol. 4,409-418.

Protein Tyrosine Phosphatases

199

Yang, Q. & Tonks, N.K. (1991). Isolation of a cDNA clone encoding a human protein-tyrosine phosphatase with homology to the cytoskeletal-associated proteins band 4.1, ezrin, and talin. Proc. Natl. Acad. Sci. USA 88, 5949-5953.

RECOMMENDED READING The reader is referred to the following reviews of the most recent advances in the study of the PTPase family: Barford, D., Jia, Z., & Tonks, N.K. (1995). Protein phosphatases take off. Nature Structure Biology 2, 1095-1101. Brady-Kalnay, S.M. & Tonks, N.K. (1995). Protein tyrosine phosphatases as adhesion receptors. Curr. Opin. Cell Biol. 7, 650-657. Hunter, T. (1995). Protein kineses and phosphatases: The yin and yang of protein phosphorylation and signalling. Cell 80, 225-236. Ninfa, E.G. & Dixon, J.E. (1994). Protein tyrosine phosphatases in disease processes. Trends Cell Biol. 4,427-430. Sun, H. & Tonks, N.K. (1994). The coordinated action of protein tyrosine phosphatases and kinases in cell signalling. TIBS 19, 480-485. Trowbridge, I.S. & Thomas, M.L. (1994). CD45: An emerging role as a protein tyrosine phosphatase required for lymphocyte activation and development. Ann. Rev. Immunol. 12, 85-116.

This Page Intentionally Left Blank

Chapter 6

Cyclic Cascades in Cellular Regulation p. BOON CHOCK and EARL R. STADTMAN

Introduction Development of the Cyclic Cascade Model Regulation of Enzymic Activity Transcriptional Regulation The Cyclic Cascade Model Unidirectional Versus Cyclic Cascade Signal Amplification Amplitude Sensitivity Flexibility and Biological Integrators for Metabolic Signals Rate Amplification Energy Consumption Cyclic Cascade Versus Simple Allosteric Control Summary

201 204 204 207 208 208 211 214 214 215 215 217 217 218

INTRODUCTION Reversible covalent modification of a protein was first found to occur in the interconversion between the active a form and the relatively inactive b form of liver

Principles of Medical Biology, Volume 4 Cell Chemistry and Physiology: Part II, pages 201-220 Copyright © 1996 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-806-4

201

202

P. BOON CHOCK and EARL R. STADTMAN

and rabbit skeletal muscle glycogen phosphorylase (Sutherland and Wosilait, 1955; Fischer and Krebs, 1958). The conversion between a and b form involves a covalent attachment and detachment of a phosphate group to and from a specific amino acid residue of the protein, respectively. This phosphate attachment, phosphorylation, is catalyzed by a specific kinase, phosphorylase kinase, with ATP as the phosphate donor. Dephosphorylation is catalyzed by a specific protein phosphatase. To date, there are more than 400 proteins that have been found to undergo similar types of reversible covalent modification. Based on the nature of the modifying group, one can classify this type of covalent modification into several classes as shown in Table 1. Among those shown in Table 1, protein phosphorylation is the most widely observed. Phosphorylation of serine, threonine, and tyrosine occur mainly in eukaryotes while histidine and aspartate phosphorylation occur mainly in prokaryotes. Phosphorylation of aspartate is the result of phosphotransfer of the phosphate from a kinase which contains the phosphorylated histidine. Many of the interconvertible enzymes and proteins can be phosphorylated at multiple sites that Table 1, Reversible Covalent Modifications of Proteins^ Modification

Donor Molecule

Phosphorylation

ATP, GTP

ADP-ribosylation

ATP, acetyl phosphate Phosphorylated histidine NAD"

Modified Amino Acid Residue Serine Threonine Tyrosine Hydroxylysine Histidine Aspartate Arginine Glutamate Lysine (terminal C O O H )

Diphthamide Nucleotidylylation (adenylyla-• ATP, UTP tion and uridylylation) Methylation S-Adenosyl-meth ion i ne

Acetylation Tyrosylation Sulfation Palmitoylation

Acetyl-Co A Tyrosine 3-Phosphoadenosine 5-Phosphosulfate Palmitoyl-CoA

Ubiquitination

Ubiquiton, ATP

Note: ^From Chock and Stadtman (1992); Feng et aL, (1992).

Tyrosine Aspartate Glutamate Lysine Histidine Glutamine Lysine Carboxyl terminus Tyrosine Cysteine Serine Threonine Lysine

Cyclic Cascades

203

are catalyzed by single- or multi-protein kinases. For example, glycogen synthase from skeletal muscle can be phosphorylated at 10 sites in vitro by nine protein kinases, each of which is modulated by different allosteric effectors, and phosphorylation at one site can enhance phosphorylation at other sites. In addition, certain interconvertible proteins, such as phosphorylase kinase, histones, and interleukin-2 receptor are known to undergo more than one form of modification. Multi-site modification can provide a mechanism for progressive degrees of regulation and it can create a cross-talk effect for different regulatory pathways. Although tyrosine sulfation has not been demonstrated to be reversible, nor has it been shown to affect enzymic activity of the sulfated protein, it is included in Table 1 because it is likely that it could have a regulatory role. This is based on reports which show that interleukin-2 receptor contains tyrosine sulfate, and that lower level of tyrosine sulfate are found in cells that are transformed by retroviruses. Interestingly, the enzyme which catalyzes the sulfation reaction, tyrosylprotein sulfotransferase, is activated by a tyrosine-sulfur-binding product. In general, covalent modification of enzymes or proteins often leads to changes in their catalytic activities or their ability to regulate biological events. Covalent modification of an interconvertible enzyme can produce a change in its activity either by decreasing or increasing its K^ (in most cases it is a value for which a substrate concentration is needed to obtain 50% of the maximal activity of an enzyme), or kcat (turnover rate constant of an enzyme), or by altering its response to allosteric effectors. For example, phosphorylation of glycogen phosphorylase results in activation of the enzyme by increasing its kcat, whereas phosphorylation of myosin light-chain kinase by cAMP-dependent protein kinase inhibits the enzyme by decreasing its affinity for the essential activator, Ca(II)-calmodulin complex. Covalent modification of proteins may also lead to translocation of enzymes or proteins. For example, autophosphorylation of growth factor receptors has been shown to induce association of cytoplasmic signaling proteins to form heteromeric complexes through interaction of SH2 and possibly SH3 (Src homology region 2 and 3) domains. In addition, acylation and myristoylation may lead to translocation of the enzyme to form a stable complex with cellular membranes. Therefore, interconvertible enzymes or proteins have been considered to function as "metabolic switches" for turning on or off metabolic pathways in response to biological needs. However, as will be discussed later, covalent interconversion of protein is a dynamic process which leads to the establishment of a steady-state or a transient burst, in which the fraction of modified enzyme can be varied progressively over a wide range by changing the concentration of effectors that modulate the activities of the modifying (converter) enzymes. The importance of covalent interconversion of enzymes and proteins in cellular regulation is evidenced by the findings that many key enzymes in a broad array of cellular pathways undergo covalent modification. These include the metabolism of proteins, carbohydrates, lipids and nucleic acids, muscle contraction, signal

204

P. BOON CHOCK and EARL R. STADTMAN

transduction, viral oncogenesis, interferon action, and DNA repair. In addition, covalent modification of proteins has also been implicated in regulating specific protein degradation in the cell. For example, phosphorylation of 3-hydroxy-3methylglutaryl CoA reductase, pyruvate kinase, yeast glutamate dehydrogenase, fructose-1, 6-bisphosphatase, and cardiac troponin has been shown to increase susceptibility to proteolysis, and ubiquitination of proteins has been shown to be involved in marking proteins for ATP-dependent proteolysis (see Chapter 10).

DEVELOPMENT OF THE CYCLIC CASCADE MODEL Among the interconvertible enzyme systems, the covalent interconversion cascade that regulates glutamine synthetase from Escherichia coli is the best understood and relatively well-defined system. The cyclic cascade model was developed through detailed studies on the regulation of this enzyme. Glutamine synthetase catalyzes the biosynthesis of glutamine. Because the amide nitrogen atom of glutamine is a preferred nitrogen source for the biosynthesis of virtually all amino acids, purine and pyrimidine nucleotides, and glucosamine-6-phosphate, this enzyme is rigorously regulated in E. coli and other enteric bacteria where they synthesize their own amino acids (for review see Stadtman and Ginsburg, 1974; Stadtman and Chock, 1978; Rhee et al., 1989). Figure 1 shows the cyclic cascade of glutamine synthetase regulation. It controls both the enzymic activity and the biosynthesis of the enzyme. Regulation of Enzymic Activity E. coli glutamine synthetase is a dodecameric enzyme of identical subunits. Its activity is mainly regulated by reversible covalent modification. Under most physiological conditions, adenylylation converts the active enzyme to its inactive form. As shown in Figure 1, the reversible covalent modification of the enzyme involves a bicycle cascade which comprises two protein nucleotidylylation cycles. In the first cycle, glutamine synthetase is reversibly adenylylated at Tyr-397 on each subunit. In the second cycle, a regulatory protein Pn is uridylylated at a specific tyrosine residue of each subunit of the Pn tetramer. The adenylylation reaction involves the covalent attachment of the adenylyl moiety from ATP to the hydroxyl group of Tyr-397 via a phosphodiester linkage. The adenylylation and deadenylylation of glutamine synthetase are catalyzed at separate sites (designated in Figure 1 as ATa and AT^j, respectively) on a single adenylyltransferase with a molecular mass of 110 kPa. Up to 12 adenylyl groups can be incorporated into each glutamine synthetase molecule, and the specific activity of the enzyme is almost inversely proportional to the average number of adenylylated subunits per dodecamer. The reactivation of the enzyme involves the phosphorolysis of the adenylyl-0-tyrosyl bond catalyzed by AT^j to form ADP and the unadenylylated enzyme. Since both adenylylation and decadenylylation activi-

Cyclic Cascades

205

GS activity

Pii-(UMP)4 Gin

a-KG

' o. ""I".®'

GS synthesis

Figure 1. The cyclic cascade of glutamine synthetase regulation. Interrelationship between uridylylation, adenylylation, and phosphorylation cycles, the reciprocal controls of these covalent interconversions by L-glutamine (Gin) and a-ketoglutarate (a-KG) are shown. + and - indicates stimulation and inhibition, respectively. Abbreviations: GS, glutamine synthetase; P,, regulatory protein; ATg and AT^, adenylyltransferase activity that catalyzes the adenylylation and deadenylylation reactions, respectively; UT^ and UTd uridylyltransferase that catalyzes the uridylylation and deuridylylation reaction, respectively (UT^ is also known as uridylyl-removing enzyme); NR, (also known as NtrC) and NR,-P, unmodified and phosphorylated glnC product, respectively; NR,, (also known as NtrB) and NR,|-P unmodified and phosphorylated gInL product, respectively. (Modified from Rhee et al., 1989.)

ties are localized on a single polypeptide, without proper control, these two reactions would be intimately coupled, and glutamine synthetase would undergo senseless cycling between its adenylylated and deadenylylated forms. The net result of this cycling is simply phosphorolysis of ATP to form ADP and inorganic

206

P. BOON CHOCK and EARL R. STADTMAN

pyrophosphate. Such futile cycling is prevented by linking the adenylylation cycle with another nucleotidylation cycle in which a regulatory protein Pn undergoes a uridylylation/deuridylylation cycle. The uridylylation and deuridylylation of Pn are catalyzed at separate catalytic sites (designated in Figure 1 as UTu and UT^, respectively) on another bifiinctional enzyme, uridylytransferase, with a molecular mass of 95 kDa. The uridylylation of Pn involves the formation of a phosphodiesterase linkage between UMP from UTP and the hydroxyl group of a specific tyrosyl residue in each of the four identical subunits. The deuridylylation is the hydrolytic cleavage of the uridylyl-O-tyrosyl bond to form UMP and the unmodified Pn. Since a single polypeptide catalyzes both uridylylation and deuridylylation of Pn, strict regulation is needed to prevent senseless cycling of Pn between its uridylylation and unmodified forms, with concomitant hydrolysis of UTP to UMP and inorganic pyrophosphate. This is accomplished through the control of allosteric effectors. The coupling of the uridylylation cycle and the adenylylation cycle is derived from the fact that the unmodified form of Pn stimulates the activity of ATg for adenylylating glutamine synthetase, whereas the uridylylated form of Pn is required to activate AT^ activity for the deadenylylation reaction. Under normal physiological conditions, this cascade functions as a dynamic processing unit in which the interconvertible proteins, the glutamine synthetase, and Pn undergo continual modification and demodification. When the concentration of the metabolites—the allosteric effectors for the modification and demodification reactions—^are maintained at constant levels, a steady-state of fractional modifications of the interconvertible enzyme will be established. Thus, the fractional modifications of Pn and glutamine synthetase are determined by the concentrations of various metabolites that can affect the activities of the four converter enzymes. In fact, about 40 metabolites have been shown to affect one or more of the enzymes involved in the glutamine synthetase cascade. Among them, a-ketoglutarate and glutamine play dominant roles as shown in Figure 1. Glutamine stimulates the adenylylation of glutamine synthetase and the deuridylylation of modified Pn, whereas it inhibits the deadenylylation of adenylylated glutamine synthetase and the uridylylation of Pn. Conversely, a-ketoglutarate inhibits the adenylylation of glutamine synthetase, but it stimulates the deadenylylation of adenylylated glutamine synthetase and the uridylylation of Pn- In other words, with the exception that a-ketoglutarate has no direct effect on the deuridylylation of Pn, all activities which lead to the inactivation of glutamine synthetase are stimulated by glutamine and inhibited by a-ketoglutarate, whereas activities which lead to reactivating glutamine synthetase are stimulated by a-ketoglutarate and inhibited by glutamine. The beauty of these reciprocal effects of glutamine and a-ketoglutarate can be fully appreciated if one considers that a-ketoglutarate is a precursor of glutamine and, therefore, the ratio of glutamine to a-ketoglutarate will vary in response to changes in the levels of ammonia. Consequently, the specific activity of glutamine synthetase will vary

Cyclic Cascades

207

rapidly in response to the availability of ammonia, as has been demonstrated in both in vitro and in vivo experiments. As will be shown later, the efficiency of a multicyclic cascade system can be further enhanced by a pyramidal relationship of the concentration of the enzymes in the cascade. In other words, the concentrations of the enzyme in the first cycle are lower than those in the next cycle, and the concentration of the converter enzyme is lower than that of the interconvertible enzyme it catalyzes. This pyramidal relationship exists in the glutamine synthetase bicyclic cascade. For E. coli K12 grown under derepressed growth conditions, the relative concentrations (in term of subunits) of glutamine synthetase, Pu, adenylyltranferase, and uridylyltransferase were found to exist in a ratio of 411:42:2.6:1. Transcriptional Regulation The synthesis of glutamine synthetase is regulated in response to the availability of nitrogen and carbon source. E. coli culture grown on limited nitrogen and excess glucose medium exhibits a high level of unadenylylated glutamine synthetase. In the presence of excess nitrogen and glucose-limited culture, glutamine synthetase levels are repressed and the enzyme is almost fully adenylylated. The mechanistic studies of the transcriptional regulation of glutamine synthetase revealed that the bicyclic cascade described above, which was originally thought to regulate only the activity of glutamine synthetase, is, in fact, tightly coupled to the transcriptional regulation of the gene which encodes the enzyme (for review see Magasanik, 1982; Keener et al., 1987; Rhee et al., 1989). The structural gene of glutamine synthetase, glnA, is transcribed from a major promoter glnAp2 which is located closest to glnA, and a minor promoter glnApl, which lies upstream of glnAp2. Transcription from the glnAp2 promoter requires RNA polymers containing the alternate sigma factor, a^"^. This polymerase binds tightly to gluAp2, but fails to melt the DNA at the transcription start site. The formation of an open complex by a^"^ RNA polymerase requires the phosphorylated form of a nitrogen regulator I, NRj (also known as NtrC). NRj, the product of glnG gene, exists as a dimer of 55 kDa subunits. As shown in Figure 1, the kinase that phosphorylates NRi is the phosphorylated form of NRn (also known as NtrB), NRn-P. NRji is the glnL product which also exists as a dimer. It catalyzes the autophosphorylation of its histidine residue 139. The phosphorylated NRn would then transfer its phosphoryl group to Asp-54 of NRj. This aspartyl phosphate is relatively labile such that NRi-P can undergo autodephosphorylation. However, in the presence of NRn, Pn, and ATP, the rate of dephosphorylation of the phosphorylated NRj is enhanced by a factor of 4. Thus, the evidence that the unmodified Pji facilitates the dephosphorylation of NRi catalyzed by NRn provides the linkage between transcriptional regulation of glutamine synthetase to the bicyclic cascade which modulates the enzymic activity. The findings that the same cyclic cascade system controls both

208

P. BOON CHOCK and EARL R. STADTMAN

the biosynthesis and the activity of glutamine synthetase lead to the conclusion that there is coordinated control of the specific activity of the enzyme. As a consequence, fluctuation in intracellular concentrations of metabolites, such as glutamine and a-ketoglutarate, can be sensed by their effects on the activities of UTy and UT^, through which the signal is transmitted, via the cyclic cascade, to adjust both the concentration and the activity of glutamine synthetase in cells.

THE CYCLIC CASCADE MODEL The study of the fairly well-defined cyclic cascade that regulates E. coli glutamine synthetase reveals that the analysis of this biological cascade is extremely complex. However, knowledge from this in-depth investigation permits one to formulate a simplified model that contains only the essential features of interconvertible enzyme systems. Unidirectional Versus Cyclic Cascade Covalent modifications of enzymes and proteins are catalyzed by converter enzymes such as kinases or phosphatases. They involve the action of one enzyme upon another and are therefore referred to as cascade systems. They can be classified as unidirectional and cyclic cascades. Unidirectional cascades are irreversible and are usually involved in proteolytic cleavage of a specific peptide bond, such as those that occur in the activation of zymogens. Well-recognized unidirectional cascades include the blood clotting cascades and the cascade involved in complement fixation. They serve as biological amplifiers designed to generate an avalanche of products to meet specific biological challenges. When the need subsides, the cascade is terminated. Therefore, unidirectional cascades are contingency systems functioning as biological switches which can be turned on to meet occasional emergency situations. In contrast, cyclic cascades involve the derivatization of one or more specific amino acid residues within the protein, such as occurs in the phosphorylation of the hydroxyl group of serine, threonine or tyrosine, in the adenylylation of the hydroxyl moiety of tyrosine, and in other modifications shown in Table I. These enzyme or protein derivatizations are cyclic processes due to the coupling of two opposing cascades; one carries out the derivatization reaction(s), while others catalyze the regeneration reaction(s), as shown in Figure 2. In the derivatization or forward cascade, an inactive converter enzyme, Ej, is activated by an allosteric effector, Cj, to its active state, Ea, as in the case of cAMP activation of cAMP-dependent protein kinase. The active Ea then catalyzes the conversion of an interconvertible enzyme, I, from its unmodified form, o-I, to its modified form, m-I, as occurs, for example, when cAMP-dependent protein kinase catalyzes the phosphorylation of phosphorylase kinase. In the regeneration cascade, the inactive converter enzyme, Rj, is activated by an allosteric effector, e2. This active Ra then catalyzes the conversion

Cyclic Cascades

209

ADP

H2O Rj + e2 Figure 2. Schematic representation of a monocyclic cascade: Ki, K2, K^f, and K^^ are dissociation constants for the allosteric effector ei from converter enzyme Eg, e2 from converter enzyme Rg, and for interconvertible enzyme o-l from Eg • o -I complex, and m-l from Rg • m -I complex, respectively; kif and ki^ are specific rate constants for the reaction designate. (From Chock and Stadtman, 1979).

of the modified interconvertible enzyme to its original form, o-I, as when protein phosphatase catalyzes the dephosphorylation of a phosphorylated protein. Dynamic coupling of the forward and regeneration cascades leads to a steady-state in which the rate of m-I formation is equal to the rate of o-I regeneration, and ATP is concomitantly hydrolyzed to ADP and inorganic phosphate. The monocyclic cascade shown in Figure 2 can easily be extended to multicyclic cascades when one form of the interconvertible enzyme functions as a converter enzyme for another interconvertible enzyme, and so on. Figure 3 depicts a multicyclic cascade consisting of n cycles. As will be seen later, the regulatory properties of cyclic cascades can be further amplified as the number of cycles in the cascade increases. In the theoretical analysis, it has been ascertained that the reversible covalent modification of enzymes does not function simply as an on-off switch for various metabolic pathways, but rather that it is part of a dynamic process in which the fractional activities of the interconvertible enzymes can be varied progressively over a wide range. This concept arose from the observation that the adenylylation of glutamine synthetase is not an all-or-none process; instead, a steady-state is established and its level is modulated by the concentration of effectors present (Segal et al., 1974). Similar results have also been reported for the mammalian pyruvate dehydrogenase complex (Pettit et al., 1975). Furthermore, the ATP concentration is maintained in excess relative to the enzymes in the cascade and at a fairly constant level it is treated as a constant factor. To simplify the analysis, it is also assumed that: (a) the formation of the enzyme-enzyme and enzyme-effector complexes proceed via a rapid equilibrium mechanism; (b) the concentration of the

210

P. BOON CHOCK and EARL R. STADTMAN

-12-.

R2i + e3 = r K3

R2a. a2r

m-In

Figure 3. Schematic representation of a multicyclic cascade where the forward converter enzyme for the interconvertible enzyme, 1^, is the modified form of the interconvertible enzyme, ln_i. The parameters a^i and ap^ are defined as: anf = k^/Kpf. ttpr = kn/Knr where n represents the number of cycles. K represents the dissociation constant for the designated reaction. (From Stadtman and Chock, 1978.)

enzyme-enzyme complex is negligible compared to the concentrations of the active and inactive enzymes; and (c) the concentrations of the allosteric effectors are maintained at constant levels for any metabolic state. With these assumptions, the highly complex equations obtained to describe the cyclic cascade system can be reduced to relatively simple equations to quantify the behavior of cyclic cascades. Quantitative analysis of cyclic cascades revealed the following features: 1. They are endowed with an enormous capacity for signal amplification. As a consequence, they can respond effectively to concentrations of primary effector (such as ei in Figures 2 and 3) well below the dissociation constant of the effector-enzyme complex. 2. They can modulate the amplitude of the maximal response that an interconvertible enzyme can accomplish even at saturating concentrations of allosteric effectors. 3. They can enhance the sensitivity of modification of the interconvertible enzyme to changes in the concentrations of allosteric effectors (i.e., they are capable of eliciting apparent positive and negative cooperativity in response to increasing concentrations of allosteric effectors).

Cyclic Cascades

211

4. They serve as biological integrators that can sense simultaneous fluctuations in the intracellular concentrations of numerous metabolites and adjust the specific activity of the interconvertible enzymes accordingly. 5. They are highly flexible with respect to allosteric regulation and are capable of exhibiting a variety of responses to primary allosteric stimuli. 6. They serve as rate amplifiers and, therefore, are capable of responding extremely rapidly to changes in metabolite levels. Signal Amplification This signal amplification is a time-independent parameter defined as the ratio of the concentration of the primary allosteric effector required to attain a 50% activation of the converter enzyme to the concentration required to produce a 50% modification of the interconvertible enzyme. This property derives from the fact that signals (allosteric effectors) exert their effects on the target interconvertible enzyme through the catalytic action of converter enzymes. Equation 1, derived with the assumptions described above, illustrates that fractional modification of the interconvertible enzyme for the monocyclic system shown in Figure 2 is a multiplicative function of the kinetic parameters required to describe the cascade. [m-I]

]£

^1

LV

^1/

Ku

(m m \e{\ \ei\

{\^\e{\/K, 1 + \e^yK^

+1

(1)

where Ki, K2, Kif, and K^r are the dissociation constants for the activated converter enzyme complexes Eg and Ra and enzyme complexes Ea«o-I and Ra •m-I, respectively; kif and kir are rate constants for the modification and demodification reactions, and [E], [R], and [I] are total concentrations of E, R, and I, respectively. As a result, small changes in several parameters can lead to enormous gains in signal amplification. With the exception of \t{\ and [e2], all of the eight parameters in equation (1) can be altered by allosteric interactions of one or more of the cascade enzymes with single or multiple allosteric effectors. Due to signal amplifications, the interconvertible enzymes can yield meaningful responses to effector concentrations that are well below the dissociation constant of the effector-converter enzyme complex. In other words, one only needs to activate a relatively small fraction of the converter enzyme to achieve a significant fractional modification of the interconvertible enzyme. This property is illustrated in Figure 4. Here the computer-simulated fractional modifications of the interconvertible enzyme in a monocyclic cascade were plotted as a function of fractional activation of the converter enzyme E. Curve 1 was obtained with equation (1) by varying [ci], while other parameters were assigned values of 1. Curve 2 was obtained when one varied the eight parameters (excluding [ej and [e2]) each by a factor of only 2 in favor of m-I formation. It can be seen that when 1% of converter enzyme is activated, curve 1 and 2 show that under this condition, one can achieve a 2% and a 75%

212

P. BOON CHOCK and EARL R. STADTMAN T

I.U

2/^

[m-H

ni

°5 ~

/

f

1 /

~i

H n

10-3

1

10-2

10-1

[Ea] [E] Figure 4. Computer-simulated curve showing the relationship between fractional activation of the converter enzyme E and fractional modification of the interconvertible enzyme I. The curves were calculated using equation 1, setting Ki = 1 and varying [eiJ. Curve 1 was obtained when all the parameters were given values of 1 and curve 2 was obtained when the eight parameters (excluding [e^] and [e2]) were each varied by a factor of 2 in favor of m-l formation.

modification of the interconvertible enzyme, respectively. In other words, only a relatively small fraction of converter enzyme is needed to be activated in order to obtain a significant fractional modification of the interconvertible enzyme. Theoretical analysis of a multicyclic cascade, as shown in Figure 3, indicates that signal amplification can be increased enormously as a function of the number of cycles in the cascade. For each additional cycle in the cascade, one introduces eight additional parameters (see Figure 3). Note that Onf andttnreach consists of two constants, because a represents the ratio of rate constant over dissociation constant. In the fractional modification expression of the last interconvertible enzyme, In, these eight parameters occur together with all of the parameters involved in the preceding cycles as a multiplicative function. Therefore, each additional cycle in a cascade results in a substantial increase in signal amplification and allosteric control potential of the system. This amplification effect is illustrated in Figure 5. As a point of reference, curve O shows how the fractional activation of thefirstconverter enzyme E varies with increasing log [Q\] when the equilibrium constant Ki is set at 1.0. With Ki being held constant at 1.0, and all parameters leading to modification are assigned values of 2.0 while those leading to demodification are assigned values of 0.5, one obtains curve 1 to 4 for n = 1 to 4, respectively. Analysis of these curves shows that one can obtain a signal amplification factor of 3.2 x 10^, 1.02 x 10^ 3.28 x 10^ and 1.05 x 10^^ for cycle 1,2, 3, and 4 in the cascade, respectively. Thus, an amplification factor of 10^^ can be

Cyclic Cascades

213

10

lis

:

E

0

/

0

1

2 CYCLE

3

4

J;

1.0?

Kin] Pnl 0-5

n=/4

-12

/3

p

M

/o

-5 log[ei]

Figure 5, Computer-simulated curves for the fractional modification of the interconvertible enzyme, !„. Curve O represents the fractional activation of the first converter enzyme E assuming K^ = 1.0. Curves 1 to 4 were obtained with each of the parameters, except Ki (= 1.0), being varied by a factor of 2 in favor of the modification. The inset depicts the linear relationship between log (signal amplification) and n. (From Stadtman and Chock, 1978.)

obtained in a four-cycle cascade by only a two-fold change in each parameter. The inset in Figure 5 shows that under these conditions, the log of the amplification factor is proportional to the number of cycles in the cascade. It should be pointed out that in unidirectional cascades, signal amplification is infinite because, with sufficient time and ATP, all of the unmodified interconvertible enzymes will be converted to their modified form in response to any level of allosteric effector. In addition, the signal amplification described here is different from catalytic amplification, which is solely a function of the relative concentrations and catalytic efficiencies of the converter and interconvertible enzymes in the cascade. In many cascades, there exists a pyramidal increase in the concentrations of the cascade enzymes; that is, the concentrations of converter enzymes are

214

P. BOON CHOCK and EARL R. STADTMAN

significantly lower than those of the interconvertible enzyme substrates. This high catalytic amplification potential can also enhance the signal amplification. Amplitude Amplitude is defined as the maximal level of fractional modification of the interconvertible enzyme at saturating concentration of an effector. By changing the magnitude of the parameters, such as concentration of the converter enzymes, dissociation constants or rate constants, which can be varied due to allosteric interactions, the amplitude can change smoothly from nearly 100% to almost 0%. Thus, even at saturating levels of an allosteric effector, interconvertible enzyme cascades do not function as on-off switches. Sensitivity Cyclic cascades can generate either apparent positive or negative cooperative responses in fractional modification (i.e., enzymic activity) of the interconvertible enzyme to increasing concentration of an allosteric effector. These apparent cooperativities can be attained from the synergistic or antagonistic effects that a single allosteric effector exerts on two or more steps in the cascade. Under these conditions, a sigmoidal response need not reflect positive cooperativity in the binding of an effector to multiple binding sites on the converter enzyme. Instead, it can be accomplished when an effector activates the forward converter enzyme and inactivates the reverse converter enzyme or vice versa. Therefore, an effective way to obtain high sensitivity is to have both forward and reverse converter enzymic activities located on a single polypeptide such that binding of one effector can lead to activation of one activity and inactivation of the other. Alternatively, high sensitivity can be obtained by activating a forward converter enzyme that leads to the generation of an inhibitor for the reverse converter enzyme, or vice versa. The latter case has been observed when cAMP activates the type II cAMP-dependent protein kinase which leads to the dissociation of its regulatory subunit. The dissociated regulatory subunit inhibits both the activation and the catalytic activity of type I protein phosphatase (Jurgensen et al., 1985). To date, five bifunctional enzymes that are involved in cyclic cascade regulation have been purified and characterized. They are the uridyltransferase that catalyzes the uridylylation/deuridylylation of Pn regulatory protein (Garcia and Rhee, 1983), the adenylyltransferase that catalyzes the adenylylation/deadenylylation of glutamine synthetase (Rhee et al., 1978), the protein kinase/phosphatase that catalyzes the phosphorylation/dephosphorylation of isocitrate dehydrogenase inE. coli (La Porte and Koshland, 1982), the 6-phosphofriicto-2 kinase/fraictose-2, 6-biphosphatase that catalyzes the synthesis and breakdown of fructose-2, 6-biphosphate (ElMaghrabi et al., 1982; van Schaftingen et al., 1982), and the NRn protein ki-

Cyclic Cascades

215

nase/phosphatase that catalyses the phosphorylation/dephosphorylation of NRj (Keener and Kustu, 1988; Weiss and Magasanik, 1988). In addition, apparent cooperativity can also be obtained when the active converter enzyme forms a tight complex with the interconvertible enzyme as shown by Shacter et al., (1984) and Goldbeter and Koshland (1981) who call this effect "zero-order ultrasensitivity." They showed that when the values of one or both of the dissociation constants for the two converter enzyme • interconvertible enzyme complexes is smaller than the total concentration of the interconvertible enzyme, one can obtain a thresholdlike response to changes in effector concentrations. Flexibility and Biological Integrators for Metabolic Signals Cyclic cascades can provide flexibility for generating different allosteric control patterns and for regulation by multiple metabolites. The scheme shown in Figure 2 illustrates just one of many variations that can be obtained by changing the nature of the interactions between the allosteric effectors, Ci and e2, and the converter enzymes, E and R. Analyses have been carried out for four different mechanistic schemes of monocyclic cascades. The results show that the regulatory patterns, expressed in terms of fractional modification of the interconvertible enzyme as a function of increasing effector concentration, are different with respect to their amplitude, signal amplification, and sensitivity (Stadtman and Chock, 1977). This flexibility for obtaining different mechanistic-dependent control patterns increases with the number of cycles in the cascade. Because a minimum of two converter enzymes and one interconvertible enzyme are needed to form a single interconvertible cycle, and each enzyme can be a separate target for one or more allosteric effectors, cyclic cascades have the capability to provide a high degree of flexibility for the input of metabolic signals. By means of allosteric interactions with the cascade enzymes, fluctuations in the concentration of numerous metabolites will lead to automatic adjustments in the activities of the converter enzymes which determine the steady-state levels of fractional modification, i.e., specific activity of the interconvertible enzymes. Multiple entries for metabolic signals via cyclic cascade systems provide both a highly flexible sensor and safeguard mechanism for the cells. In essence, cyclic cascades can serve as biological integrators that sense changes in the concentration of innumerable metabolites and thus modulate the activity of pertinent enzymes accordingly. Rate Amplification The properties of these cyclic cascades are time-independent. For the cells to be able to utilize these properties, the steady-state must be achieved within a reasonable time frame. Kinetic analysis of multicycle cascades reveals that the rate of

216

P. BOON CHOCK and EARL R. STADTMAN 1.0

n= [3 /2 [m-ln] [•n]

0.5 h

1,..—-H

1

0.25

Time, sec.

0.5

Figure 6. Computer-simulated time course showing the rate amplification as a function of cycle (n) in the cyclic cascades. All of the concentrations of interconvertible enzymes were set at 10"^ M while those of the converter enzymes and effectors were set at 5 X 10"^ M. All dissociation constants were equal to 5 x 10"^ M and forward and reverse rate constants were set at 1,000 sec"^ and 10 sec~\ respectively. (From Stadtman and Chock, 1979.)

covalent modification of the last interconvertible enzyme in the cascade is a multiplicative ftmction of the rate constants of all reactions that lead to the formation of the modified enzyme. Therefore, cyclic cascade systems can function as rate amplifiers. Figure 6 illustrates that the rate amplification potential increases with the number of cycles involved in the cascade. Curve 1 shows that for a monocyclic cascade, the fractional modification proceeds with a normal exponential function. However, for a bicyclic cascade (curve 2) and a tricyclic cascade (curve 3), the fractional modification of their last interconvertible enzyme increases with an initial lag phase, followed by a burst. It should be pointed out that the calculated curves in Figure 6 were obtained with the assumption that the concentrations of all of the interconvertible enzymes are the same, and all of the forward and reverse rate constants are set at 1,000 sec~^ and 10 sec~^ respectively. Under these conditions, to obtain a 50% modification of the last interconvertible enzyme, one needs only 85 and 44 msec for the two- and three-cycle cascades, respectively. The extent of the rate amplification can be further enhanced if the multicyclic cascade involved bears a pyramidal relationship with respect to the concentration of its interconvertible enzymes, i.e., [IJ < [I2] < [I3] and so on (see Figure 3) (Stadtman and Chock, 1979). Furthermore, if the converter and interconverter enzymes are topographically positioned close to each other, an even greater rate of response can be achieved. With the cyclic cascade model and reasonable rate constants, one can demonstrate that multicyclic cascades can generate large bio-

Cyclic Cascades

217

chemical responses to primary stimuli within the millisecond time range. Experimentally, Danforth et al. (1962) demonstrated that phosphorylation of phosphorylase b of frog sartorius muscle in response to electrical stimulation at 30 °C can be accomplished with a half-time of 700 msec. The fact that cyclic cascades can respond to stimuli in the millisecond time range, together with their capacity for signal amplification, suggests that they are excellent mechanisms for transmitting neurochemical responses. Energy Consumption Shown in Figure 2 is that for each complete phosphorylation/dephosphorylation cycle, one equivalent of ATP is consumed to generate one equivalent each of ADP and Pj. The capacity of a cyclic cascade system to maintain a steady-state requires a constant supply of metabolic energy (ATP in most cases) to drive the modification reactions. In the absence of adequate energy-rich donor molecules, the interconvertible enzymes or proteins would be converted completely to the unmodified forms. Thus, the constant ATP flux through the cyclic cascade is the energy required to maintain the operation of cyclic cascade in vivo. The rate of ATP hydrolysis is regulated by the parameters needed to describe the cyclic cascade. These include both the constants for rate and equilibrium, and the concentration of enzymes and allosteric effectors. Therefore, the rate of ATP consumption can also be regulated by the parameters shown in Figures 2 and 3. The ATP turnover rate in mammalian cells was estimated to be 1—10 mM/min. Relative to this ATP flux, the amount of ATP needed to maintain a given cyclic cascade is not negligible. Therefore, it is not surprising to find synchronous control of the forward and reverse converter enzymes in cyclic cascade systems. In addition, energy conservation can also be achieved by lowering the steady-state rate for the modification/demodification reactions, or by operating the cyclic cascades in a transient manner in response to stimuli, a phenomenon often observed in biological systems. Cyclic Cascade Versus Simple Allosteric Control Metabolic regulation by cyclic cascades involves both covalent modification of enzymes and proteins, and allosteric interaction between effectors and enzymes. In view of its complexity, and the fact that a similar type of reaction scheme (as shown in Figures 2 and 3) can be obtained without invoking covalent modification of enzymes, it is reasonable to wonder whether allosteric interactions between metabolites and enzymes alone could produce the regulatory properties revealed for cyclic cascades. For signal amplification to be physiologically significant, it has to meet the following conditions: (a) a minimal concentration of metabolites should be required to initiate the biological response(s) within a reasonable time-frame,

218

P. BOON CHOCK and EARL R. STADTMAN

and (b) the reaction(s) induced by the metaboHtes must be catalytic such that one effector can activate more than one target enzyme molecule. Thefirstcondition can be fulfilled either by the cyclic cascade mechanism described in this chapter or by very tight binding between the allosteric effector and the ta;rget enzyme. Since the rate of effector binding is limited by the diffusion rate of the reactants, tight binding can only be achieved by slowing down the off-rate for the enzyme-bound effector. Therefore, high affinity between effector and enzyme would reduce the temporal efficiency of the control process. With the cyclic cascade mechanism, tight binding is not required for high signal amplification. Figure 4 shows it is possible that 1% binding or activation of the converter enzyme to the effector can lead to 75% fractional modification of the interconvertible enzyme in a monocyclic cascade system. In order to obtain a catalytic effect with a simple allosteric model, the effector must first bind to the target enzyme, induce an active conformation, and then dissociate fi-om the active enzyme, which must remain active in order to continuously carry out its catalytic function. When the activity of the target enzyme is no longer needed, the active enzyme would then relax to its inactive state. This type of regulatory mechanism is thermodynamically unfavorable. To make it thermodynamically feasible, one can use the substrate of the enzyme to stabilize its active conformation. However, in this situation, the inactivation of this enzyme will be determined by the time required to consume its available substrate(s). This will not constitute an efficient regulatory mechanism because the control process relies on the catalytic efficiency of the target enzyme and the pool size of its substrate(s), which could be present in large excess. In addition, in the absence of catalytic intermediates, there will be no rate amplification. Furthermore, the capacity for allosteric interactions in cyclic cascade systems is significantly higher relative to a simple allosteric model because there are more proteins—hence, more potential binding sites for various effectors—^involved in a cyclic cascade. On the other hand, the apparent cooperativity which provides the sensitivity observed in the cyclic cascades can be obtained by allosteric interaction alone, particularly if the enzyme contains multiple subunits. However, cyclic cascade systems have the capacity of obtaining much higher cooperative responses through synchronous control by a given allosteric effector. In essence, then, signal amplification and rate amplification cannot be easily achieved without invoking a cyclic cascade mechanism, while the other regulatory properties of the cyclic cascade can be accomplished with a simple allosteric model, but with less regulatory efficiency.

SUMMARY The cyclic cascade model, derived mainlyfi-omdetailed studies ofE. coli glutamine synthetase, is applicable to all covalent interconvertible enzyme systems. The regulatory mechanism makes use of both covalent modification and allosteric interactions. Analysis of this model reveals its regulatory advantages, which

Cyclic Cascades

219

include signal amplification, rate amplification, sensitivity, and flexibility. Most of the properties predicted by theoretical analysis have been verified experimentally by in vitro studies of the phosphorylation/dephosphorylation cycle and the glutamine synthetase cascade, and by studies on permeabilized cells. By means of allosteric interactions with one or more enzymes, cyclic cascades can continuously monitor fluctuations in the concentrations of a multitude of metabolites and adjust the specific activities of the target enzymes in response to biological requirement. Therefore, they serve as biological integrators with the capacity to provide a safeguard mechanism for biological systems. With the enormous capacity for signal and rate amplification, cyclic cascades provide an ideal mechanism for biological signal transduction. Although a cyclic cascade modulates the specific activity of the interconvertible enzyme smoothly and continuously over a wide range of conditions, it can provide transient responses to biological stimuli, and in extreme physiological situations serve as an on-off switch for the activity of the target enzyme. The energy for maintaining such an efficient regulatory mechanism is the consumption of ATP and other energy-rich donor molecules. In view of the unique properties of cyclic cascades, it should not be surprising that they are widely used for transmitting biological signals and for regulating both the activities and biosynthesis of key enzymes. REFERENCES Chock, P.B. & Stadtman E.R. (1979). Covalently interconvertible enzyme cascade and metabolic regulation. In: Modulation of Protein Function (Atkinson, D.E. & Fox, C.F. eds.), pp. 185-202, Academic Press, New York. Chock, P.B. & Stadtman, E.R. (1992). Cyclic Cascades in Cellular Regulation. In: Fundamentals of Medical Cell Biology, Vol 3B (Bittar, E.E. ed.), pp. 391-411, JAI Press, Inc., Greenwich. Danforth, W.H., Helmreich, E., & Cori, C.F. (1962). The effect of contraction and epinephrine on the phosphorylase activity of frog sartorius muscle. Proc. Natl. Acad. Sci. USA 48, 1191—1199. El-Maghrabi, M.R., Claus, T.H., Pilkis, J., Fox, E., & Pilkis, S.J. (1982). Regulation of rat liver fructose 2,6-bisphosphatase. J. Biol. Chem. 257, 7603-7607. Feng, J., Atkinson, M.R., McCleary, W., Stock., J.B., Wanner, B.L., & Ninfa, A.J. (1992). Role of, phosphorylated metabolic intermediates in the regulation of glutamine synthetase synthesis in Escherichia coli. J. Bact. 174, 6061-6070. Fischer, E.H. & Krebs, E.G. (1958). The isolation and crystallization of rabbit skeletal muscle phosphorylase b. J. Biol. Chem. 231, 65-71. Garcia, E. & Rhee, S.G. (1983). Cascade control oiEscherichia coli glutamine synthetase: Purification and properties of Pji uridylytransferase and undylyl-removmg enzyme. J. Biol. Chem. 258, 2246-2253. Goldbeter, A. & Koshland, D.E., Jr. (1981). An amplified sensitivity arising from covalent modification in biological systems. Proc. Natl. Acad. Sci. USA 78, 6840-6844. Jurgensen, S.R., Chock, P.B., Taylor, S., Vandenheede, J.R, & Merlevede, W. (1985). Inhibition of the Mg(II)-ATP-dependent phosphoprotein phosphatase by the regulatory subunit of cAMP-dependent protein kinase. Proc. Natl. Acad. Sci. USA 82, 7565-7569. Keener, J., Wong, P., Dopham, D., Wallis, J., & Kustu, S. (1987). RNA Polymerase and the Regulation of Transcription (Reznikoff, W.S., Burgess, R.R., Dahlberg, J.Z., Gross, C.A., Record, M.T., Jr., & Wickens, M.P., eds.), pp. 159-175. Elsevier Sciences Publishers, New York. ,

220

P. BOON CHOCK and EARL R. STADTMAN

Keener, J. & Kustu, S. (1988). Protein kinase and phosphoprotein phosphatase activities of nitrogen regulatory proteins NRTB and NTRC of enteric bacteria: Roles of the conserved amino-terminal domain of NTRC. Proc. Natl. Acad. Sci. USA 85, 4976-4980. La Porte, D.C. & Koshland, D.E., Jr. (1982). A protein with kinase and phosphatase activities involved in regulation of tricarboxylic acid cycle. Nature 300, 458-460. Magasanik, B. (1982). Genetic control of nitrogen assimilation in bacteria. Ann. Rev. Genetics 16, 135-168. Pettit, F.H., Pelley, J.W., & Reed, L.J. (1975). Regulation of pyruvate dehydrogenase kinase and phosphatases by acetyl-CoA/CoA and NADH/NAD ratios. Biochem. Biophys. Res. Commun. 65, 575-582. . Rhee, S.G., Chock, P.B., & Stadtman, E.R. (1989). Regulation of Escherichia coli glutamine synthetase. Adv. Enzymol. 62, 37-92. Rhee, S.G., Park, R., Chock, P.B., & Stadtman, E.R. (1978). AUosteric regulation of monocyclic interconvertible enzyme cascade systems: Use of Escherichia coli glutamine synthetase as an experimental model. Proc. Natl. Acad. Sci. USA 75, 3138-3142. Segal, A., Brown, M.A., & Stadtman, E.R. (1974). Metabolite regulation of the state of adenylylation of glutamine synthetase. Arch. Biochem. Biophys. 161, 319-327. Shacter, E., Chock, P.B., & Stadtman, E.R. (1984). Energy consumption in a cyclic phosphorylation/dephosphorylation cascade. J. Biol. Chem. 259, 12260-12264. Stadtman, & Chock, P.B. (1977). Superiority of interconvertible enzyme cascades in metabolic regulation: Analysis of moncyclic system. Proc. Natl. Acad. Sci. USA 74, 2761—2766. Stadtman, E.R. & Chock, P.B. (1978). Interconvertible enzyme cascades in metabolic regulation. Curr. Topics Cell. Reg. 13,53-95. Stadtman, E.R. & Chock, P.B. (1979). Advantages of enzyme cascades in the regulation of key metabolic processes. In: The Neuroscience Fourth Study Program (Schmitt, P.O., ed.), pp. 801-817, MIT Press, Cambridge. Stadtman, E.R. & Ginsburg, R.A. (1974). The glutamine synthetase of Escherichia coli: Structure and control. In: The Enzyme Vol. X (Boyer, P.D., ed.), pp. 755-^07, Academic Press, New York. Sutherland, E.W., Jr., & Wosilait, W.D. (1935). Inactivation and activation of liver phosphorylase. Nature (London) 175, 169-1750. Van Schaftingen, E., Davies, D.R., & Hers, H.G. (1982). Fructose-2,6-bisphosphatase from rat liver. European J. Biochem. 124, 143-149. Weiss, V. 8i Magasanik, B. (1988). Phosphorylation of nitrogen regulator I of Escherichia coli. Proc. Natl. Acad. Sci. USA 85, 8918-8923.

RECOMMENDED READINGS Boyer, P.D. & Krebs, E.G. (eds.) (1986,1987). Control by Phosphorylation. The Enzymes, Parts A, B, Vols. 17, 18. Academic Press, Orlando. Rhee, S.G., Chock, P.B., & Stadtman, E.R. (1989). Regulation of Escherichia coli glutamine synthetase. In: Advances in Enzymology (Meister, A., ed). Vol. 62, pp. 37—92, John Wiley and Sons, New York. Shaltiel, S. & Chock, P.B. (eds.) (1985). Modulation by covalent modification. Curr. Topics Cell. Reg., Volume 27. Academic Press, Orlando. Stadtman, E. R. & Chock, P. B. ( 1978). Interconvertible enzyme cascades in metabolic regulation. Curr. Topics Cell. Reg. 13,53-95.

Chapter 7

Mechanisms of Intracellular pH Regulation GREG GOSS and SERGIO GRINSTEIN

Introduction Measurement of Intracellular pH pH-Sensitive Microelectrodes Nuclear Magnetic Resonance (NMR) Weak Acid or Base Partition pH-Sensitive Fluorescent Probes Mechanisms of Cytosolic pH Regulation Na /H Exchange C r / H C 0 3 Exchange Na'^-CoupledCr/HCOj Exchange H"'-ATPases Permeation of H"^, OH", and HCOJ Through Channels Mechanisms of OrganellarpH Regulation Mitochondrial pH Vacuolar System of the Cell Physiological Role of an Acidic pH in Subcellular Organelles Mechanisms of Acidification of Intracellular Compartments Control and Regulation of Vacuolar H -ATPases Concluding Summary

Principles of Medical Biology, Volume 4 Cell Chemistry and Physiology: Part II, pages 221-241 Copyright © 1996 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-806-4 221

222 223 223 223 224 224 225 225 227 228 228 229 230 230 231 231 237 239 240

222

GREG GOSS and SERGIO GRINSTEIN

INTRODUCTION The regulation of intracellular pH is of paramount importance in the maintenance of many cellular functions. Intracellular (cytosolic) pH is regulated at a level which is different from the extracellular pH. In addition, different compartments within the cell display pH values that differ by up to 3 units from the cytosolic pH. The prevalence of specific pH levels within the particular cellular compartments is not fortuitous. Instead, as discussed below in more detail, pH plays a central role in the functioning of whole cells and individual organelles. For this reason, even slight deviations from the physiological pH can have catastrophic consequences on cell function and viability. The activity of many enzymes present in the cytosol is highly pH-dependent. One example is phosphofructokinase, a cytosolic enzyme which constitutes the rate-limiting step in glycolysis. A large acceleration of glycolytic activity is observed when cytosolic pH increases slightly. This highlights the requirement for accurate and effective regulation of intracellular pH. Short term homeostasis of intracellular pH can be favored by cellular buffering, which will tend to mitigate rapid departures from the physiological pH level. The simplest form of intracellular buffering is passive physicochemical buffering, i.e., the ability of weak acids and bases to associate with or dissociatefromH"^, thereby minimizing shifts in the pH of the surrounding medium. In animal tissues, the intracellular buffering power at physiological pH levels has been estimated to range from about 25 to 100 mM/pH unit (i.e., 25 to 100 mmoles of acid equivalents need to be added to one liter of intracellular medium to alter its pH by 1 unit) (Roos and Boron, 1981). Several of the available estimates of buffering power do not include the contribution of CO2/HCOJ to the buffering capacity. In an open system in equilibrium with CO2, the buffering power will increase by 2.3 x [HCOj] (Roos and Boron, 1981). Even though acute intracellular pH changes can be partially counteracted by the cellular buffering capacity, long-term regulation requires continuously operating active (i.e., directly or indirectly energy-dependent) transport processes. Such a need arisesfromthe continuous and spontaneous tendency of H"*^ (or OH") to attain electrochemical equilibrium across biological membranes. Thus, H"*" ions tend to exit the very acidic intralysosomal compartment down their concentration gradient. Conversely, H"*^ ions tend to enter the cytosol across the plasma membrane, drawn by the internally negative transmembrane potential. An additional challenge to pH homeostasis is provided by several metabolic pathways, which yield H"*^ or other acid equivalents as theirfinalproduct. Transport of H"^ equivalents for the purpose of regulating pH need not be "primary active," i.e., directly requiring ATP hydrolysis for the catalytic (transport) event. In fact, in animal cells, most pH regulating systems are "secondary active." In such cases, the energy for the translocation of H"*" equivalents is provided by coupling to the movement of another solute down its electrochemical gradient. Frequently, the gradient of the "driver" solute is itself established by a primary active process. The following sections describe these

Mechanisms of Intracellular pH Regulation

223

primary and secondary active transport systems, the role they play in the regulation of the cytosolic/intraorganellar pH, and their importance to normal cell function. First, however, a brief section is included to describe the methods used for the measurement of intracellular pH.

MEASUREMENT OF INTRACELLULAR pH Prior to discussion of the transporters, it is necessary to understand some of the methods which have been employed to determine the pH of the cytosol and intracellular compartments. It is important to recognize that each method has distinct advantages but is also subject to certain limitations and inherent errors. This chapter will briefly outline some of the most commonly used methods. pH-Sensitive Microelectrodes H"^ selective microprobes are generally made from glass microelectrodes, similar to those used by electrophysiologists. Their principle of operation is identical to that of conventional pH (macro)electrodes. The main body of the electrode is filled with an electrolyte solution, heavily buffered to maintain the desired pH. This compartment is separated from the medium to be measured by an interphase that is selectively permeable to H"^ (usually glass or a liquid ion exchanger). When the electrode is immersed into the test medium, a Nemst-type diffusion potential develops that is proportional to the logarithm of the ratio of H"*" concentrations across the interphase. Since the H"^ activity inside the electrode is constant, its response will vary linearly with the logarithm of the H"*" activity (pH) in the experimental medium (Thomas, 1978). The tip of the electrode can be as small as 0.5 |im, thus permitting penetration of many cell types. However, with the exception of some giant organelles, microelectrodes are not suitable for the measurement of the pH of intracellular compartments other than the cytoplasm. In fact, even small cells are not amenable to impalement with ion-selective microelectrodes. Nuclear Magnetic Resonance (NMR) This techniques makes use of the fact that the electronic environment of a given atom is influenced by the surrounding molecule and its ionic state. The changes in the NMR resonance frequency resulting from protonation or deprotonation of specific molecules have been used as an index of the pH of the surrounding medium (Gadian, 1982). Measurements of intracellular pH by NMR have made use primarily of the resonances of phosphate, an endogenous probe that undergoes protonation/deprotonation reactions in the physiological pH range. However, NMR technology has had only limited use in this field, due mainly to its cost and to the large amounts of biological material and long acquisition times required to generate the data.

224

GREG GOSS and SERGIO GRINSTEIN

Weak Acid or Base Partition In general, the uncharged form of small organic acids or bases is much more permeable across biological membranes than the charged deprotonated or protonated species, respectively. For this reason, when weak electrolytes are added to the medium bathing cells or organelles, they rapidly traverse the membrane and can become protonated or deprotonated in the internal compartment. The distribution of the probe in this compartment is therefore pH dependent. Isotopic or fluorescent weak acids or bases are equilibrated with the cells or organelles of interest and their concentration is then determined by measuring radioactivity or spectroscopically. The pH can next be calculated using the Henderson-Hasselbach equation, provided the pK of the probe is known. This approach has been employed extensively in the determination of the pH of intracellular compartments (Roos and Boron, 1981). However, when whole cells are the experimental system, the resulting measurements provide only a weighted average of all the intracellular compartments, since the probes traverse all cellular membranes. pH-Sensitive Fluorescent Probes One of the most widely used methods in recent years has been the determination of the pH of small cells using pH-sensitive fluorescent compounds. In addition, adaptation of this procedure to the measurement of pH in organelles in suspension or in situ has recently been performed. The principle of the technique is that protonation or deprotonation of a probe leads to changes in its fluorescence intensity, which can be monitored by macro- or micro-fluorimetry. Maximal sensitivity is obtained by devising fluorophores which are half-maximally protonated in the pH range to be studied. There are a number of commercially available probes for use at varied pHs. Sequestration of the probe into the appropriate compartment can be achieved by endocytosis or phagocytosis, by microinjection, or by providing the probe as a membrane permeant precursor, which is chemically transformed inside the cell, rendering it impermeant (see Foskett and Grinstein, 1990 for review). There are a number of advantages to using these probes over other pH measurement systems. That cells can remain intact and viable, with presumably little interference from the probe itself, is an important advantage. In addition, the rapid response time ( « I second) and spatial resolution are far superior to the methods listed above. However, the use of this technique is prone to methodological error due to leakage of the probe from the compartment, photobleaching, or differential loading of dye. However, careful technique and the use of dual wavelength ratio spectroscopy can eliminate these errors and give valid, consistent estimations of intracellular and intraorganellar pH.

Mechanisms of Intracellular pH Regulation

225

MECHANISMS OF CYTOSOLIC pH REGULATION The cytosolic pH of a variety of animal cell types has been extensively investigated. At physiological extracellular pH (7.3—7.4), the cytosolic pH has been found in the vast majority of cases to range betw^een 7.0 and 7.3 (Roos and Boron, 1981). Theoretically, regulation of cytosolic pH can be accomplished either by translocation of acidic or basic equivalents out of the cell across the plasma membrane or by translocation into cytosolic organelles. However, the importance of translocation into intracellular organelles for the purposes of cytosolic pH regulation is dubious, given their finite capacity and the necessity to independently maintain intraorganellar pH (see below). Therefore, this section will highlight the mechanisms present on the plasma membrane which are involved in the regulation of cytosolic pH (see Figure 1). The transfer of acidic and basic equivalents is accomplished by a variety of membrane-spanning proteins (or groups of proteins) which have specific functions, structures, and biochemical/pharmacological characteristics. The principal systems involved in cytosolic pH regulation in animal cells are described below. NaVH^ Exchange The NaVH"^ exchanger (NHE), also termed Na"^/H^ antiport, is an integral membrane glycoprotein with 10-12 membrane spanning domains and a cytoplasmic carboxy-terminus tail. The Na'^/H"^ antiport catalyzes the exchange of extracellular Na"*" for intracellular H"^, thereby counteracting the tendency of the cytosol to become acidic. The stoichiometry of the exchange process is one Na^ for one H"*". For this reason, the transport cycle is electroneutral and insensitive to changes in the membrane potential. This enables excitable cells to undergo electrical potential changes without jeopardizing pH regulation and vice versa. Transport through the Na'^/H'*" exchanger is driven by the combined concentration gradients of the substrate ions, i.e., Na"^ and H"^. Thus, in theory, net flux through the exchanger ceases when the Na"*" concentration gradient is balanced by an identical H"^ gradient. Under physiological conditions, however, the net direction of exchange (Na"^ influx and H"^ efflux) is determined by the prevailing inward Na"^ gradient (extracellular [Na"^] being over 10-fold higher than the cytosolic [Na"^]). This gradient is generated and maintained in most animal cells by the NaVK"*" pump. Therefore, the continued extrusion of intracellular H"^ is directly fueled by the Na"^ gradient and indirectly by the hydrolysis of ATP. In this sense, the Na^^/H^ antiport can be considered a "secondary" active transport mechanism. Recent evidence suggests the existence of at least four isoforms of the antiport, each with specific pharmacological and biochemical properties, specific cellular/tissue localizations, and specific functions. NHE-1 is the best characterized of all the isoforms. The NHE-1 isoform is a near ubiquitous system, present in the plasmalemma of virtually all the mammalian cells studied to date, including the

226

GREG GOSS and SERGIO GRINSTEIN

Figure 7. A schematic representation of the processes responsible for the movement of proton equivalents across the plasma, vacuolar, and mitochondrial membranes. Refer to the text for a discussion of these transport pathways. Abbreviations: OA", organic anion; HOA, organic acid, protonated form. Open circles represent transport processes such as carriers and pumps, whereas gaps in the membrane represent channels. basolateral membrane of epithelial cells. It is characterized by sensitivity to inhibition by amiloride and its analogs and by its ability to respond to growth factors and tumor promoters (e.g., epidermal growth factor, okadaic acid, phorbol esters). The activity of NHE-1 is greatly stimulated when the cytosol is acidified. This

Mechanisms of Intracellular pH Regulation

227

peculiar behavior is partly attributable to the increased availability of internal substrate (i.e., H"^), but is mostly due to the protonation of an allosteric H"^ binding site on the internal surface of the exchanger. When protonated, this putative allosteric site is thought to activate NaVH"^ exchange, thereby counteracting the acidification of the cytosol. Thus, this transport protein is central to the regulation of cytosolic pH and is sometimes referred to as the "housekeeping" isoform of the antiporter. Although other isoforms of the antiport have now been described, they are far less well characterized and their specific role in cytosolic pH regulation is not yet clear. NHE-2, NHE-3, and NHE-4 have a much more discrete tissue distribution, and can differ from NHE-1 in their susceptibility to amiloride and affinity for the transported cations. Ongoing work will shortly clarify the role of these isoforms in pH regulation and ion transport. CI/HCO3 Exchange An electroneutral anion exchange process, analogous to the cation antiport, is also involved in cytosolic pH regulation. This anion exchanger is independent of the cationic composition of the medium and is capable of transporting a variety of halides as well as sulfate and phosphate. However, under physiological conditions, the anion antiport is believed to transport mainly CI" in exchange for HCO3 (Madshus, 1988). The Cr/HC03 exchanger translocates anions with a one-to-one stoichiometry and, like the NaVH"^ antiport, is therefore electroneutral and insensitive to alterations in transmembrane potential. In intact cells, Cr/HCOj exchange is generally inhibited by disulfonic stilbene derivatives. For this reason, covalently binding disulfonic stilbenes have been used to label and identify the molecules responsible for anion exchange (Cabantchik et al, 1978). Anion exchange is driven by the concentration gradients of the substrate ions. In most cell types, the intracellular concentrations of CI" and HCO3 are lower than those in the external medium, due in part to the electronegativity of the cell interior. However, the inward concentration gradient for CI" generally exceeds that for HCO3. This imbalance is expected to drive the net influx of CI" simultaneously with HCO3 efflux, resulting in cytosolic acidification. Consistent with this prediction, recent studies have indicated that Cr/HC03 exchange plays a role in the recovery from experimental alkalosis (Madshus, 1988). The activity of the Cl'/HCOj exchanger is steeply dependent on the intracellular pH. However, its activation is opposite to that of the NaVH"^ exchanger. The anion exchanger is inactivated at acidic pH and greatly stimulated at more alkaline levels. This behavior is consistent with a role in the recovery from alkalosis, since the activated exchange of external CI" for cytosolic HCG3 would tend to restore the physiological pH. At or below the normal pH, anion exchange is greatly reduced, precluding Cr/HCOj exchanger mediated acidification, which would further compromise intracellular pH homeostasis.

228

GREG GOSS and SERGIO GRINSTEIN

Na^-Coupled C I / H C O i Exchange A second type of Cr/HCOj exchange mechanism exists in the plasma membranes of both invertebrates and mammalian cells (Boron, 1983). It differs from the cation-insensitive anion exchanger described above in that the presence of Na"^ is an essential requirement for anion exchange to proceed. Under physiological conditions, this system exchanges extracellular Na"^ and HCOj for intracellular Cl~ The process has been found to be electroneutral, implying that some other cation exits the cell or that an additional anion is transported in. Because two acid equivalents have been found to exit the cell per transport cycle, two combinations have been suggested: extracellular Na"^ could enter the cell accompanied by one COf^ or two HCOj ions, rather than a single HCOj, or alternatively H"^ could be ejectedfromthe cell along with CI" (Boron, 1983). Because of the large inward Na^ gradient and the translocation of HCO3 from outside the cell to the inside, the Na^-dependent C1~/HC03 exchanger tends to prevent acidification of the cytosol. For this reason, Na'^-dependent Cr/HC03 exchange is considered to be a functional pH regulatory system, capable of antagonizing the spontaneous tendency of the cytosol to become acidic. Consistent with this view, the dependence of this system on the intracellular pH resembles that of the Na^/H"^ antiport, i.e., the rate of transport is greater at more acidic levels and the system is quiescent at normal cytosolic pH. Like the cation antiport, the regulatory activity of the Na'*'-dependent Cr/HCOj exchanger relies on the existence of an inward Na"^ gradient and is therefore a secondary active transport system, indirectly fueled by the Na'*'/K"^ pump. Despite their similarities, Na'^-dependent Cr/HCOj exchange and Na'^/H'*' exchange can be readily distinguished pharmacologically. Like the cation-independent Cr/HCOj exchanger, the Na'*'-dependent anion exchanger is sensitive to disulfonic stilbene derivatives, but not to amiloride and its analogues. Conversely, the Na'^/H'*" exchanger is sensitive to amiloride and its analogues, but not to disulfonic stilbene derivatives. H+ATPases One of the most studied transport systems has been the "proton pump." This system is a primary active transport mechanism which hydrolyzes ATP to power the translocation of H"*" across the membrane. The pump itself does not translocate an associated counterion. As such, this system is electrogenic and sensitive to changes in the transmembrane potential. The vacuolar or V-type ATPases from eukaryotic cells have a multiple subunit structure (8-9 subunits ranging in size from 17 kDa to 100 kDa) and can be easily distinguished from other classes of ATPases (e.g., E-type and F-type) on the basis of their pharmacological characteristics (see below). The presence of V-type H"*" ATPases has been convincingly demonstrated in a variety of intracellular membranes (Forgac, 1992) where they play a crucial

Mechanisms of Intracellular pH Regulation

229

role in the maintenance of intraorganellar pH (see below). However, H"^ ATPases have also been demonstrated in the plasma membranes of some tissues, such as some urinary epithelia (Al-Awqati, 1986). Further studies indicate that the pumps in these systems are mobilized rapidly to the plasma membrane in response to cellular acidification. This is accomplished by recruitment of submembranous cytoplasmic vesicles. Similar pumps in the plasma membranes of nonepithelial cells such as macrophages, activated neutrophils, and osteoclasts have also been demonstrated. V-type H"^ ATPases in neutrophils and macrophages likely play a role in the maintenance of cytosolic pH, specifically when the cells are in an acidic environment such as exists within certain tumors or at localized sites of infection. In osteoclasts, V-type ATPases play a role in pericellular acidification and hence in bone resorption. Permeation of H^ OH", and HCO3 Through Channels Experiments using liposomes and planar bilayers have unequivocally demonstrated that H"^ (or OH") can permeate through lipid bilayer membranes, but at extremely low rates. In addition, acid equivalents are thought to traverse biological membranes passively through other components, likely proteins or glycoproteins. Evidence for high conductance, highly H"^ selective, voltage dependent channels has recently been obtained for various cell types (e.g., neurons, alveolar cells, oocytes, and phagocytes; see review by Lukacs et al., 1993). The physiological significance of these channels is still being debated. Under normal conditions, the existence of such pathways would be counterintuitive. At the normal resting potential of most mammalian cells (more negative than -50 mV), these channels would favor net uptake of acid equivalents, thereby exacerbating the normal tendency for the cytosol to become acidic. However, properties of the channels allow only for activation under very specific conditions. Acidification of the exterior of the cell reduces the conductance while acidification of the cytosol activates the conductance. Furthermore, the channel is activated only when cells are depolarized. These properties ensure unidirectional outward flow of H"^. These channels seem only to be localized in specific cell types in which a depolarizing voltage is also accompanied by a metabolic burst (e.g., nerves, phagocytes). Because the channels are activated by depolarization, it is possible that efflux of acid occurs during depolarization, facilitating the maintenance of pH. Conversely, the predicted deleterious entry of acid into repolarized (inside negative) cells would be prevented by closure of the channels (Lukacs et al., 1993). Bicarbonate, a base equivalent, is present at concentrations several orders of magnitude higher than those of H^ or 0H~. It has been demonstrated that HCOj traverses the membrane not only by electroneutral exchange (see above), but also through conductive channels which likely transport also other anions, primarily CI" (Bretag, 1987). Because the membrane potential is generally negative inside.

230

GREG GOSS and SERGIO GRINSTEIN

opening of HCOj channels is expected to induce HCOj efflux with a resulting cytosolic acidification. The physiological significance of this acidification remains obscure.

MECHANISMS OF ORGANELLAR pH REGULATION The importance of controlling the intracellular pH and the different mechanisms employed by the cell to maintain a constant cytosolic pH were described in the preceding section. In addition to the cytosol, many other compartments (or organelles) are found inside the cell. Each of these membrane enclosed compartments is involved in a specific cellular activity. It is often advantageous (and sometimes essential) for the cell to perform some activities in a compartment that is physically separated from the cytosol by an impermeable biological membrane. For example, the production of cellular energy (ATP) depends on a H"^ gradient generated across the inner mitochondrial membrane. Similarly, the enzymatic degradation of molecules ingested by the cell is carried out in a compartment that is not in direct communication with the cytosol, in order to prevent the uncontrolled degradation of useful cellular proteins. The biological membranes enclosing the many intracellular organelles are poorly permeable to most ions (including H"*"), so the intracompartmental ionic composition and pH often differ from those of the cytosol. Using some of the experimental approaches described at the beginning of this chapter, the interior of several intracellular organelles was demonstrated to be acidic, while mitochondria were found to be alkaline. In the following sections, we will describe the physiological importance of the pH of the interior of specific organelles and the mechanisms involved in controlling it. Mitochondrial pH Regulation of mitochondrial pH is a special case for intracellular pH regulation in that the double membrane surrounding the matrix results in two separate compartments in addition to the cytosol. The pH of the mitochondrial matrix is significantly more alkaline (+0.5 pH unit) than that of the cytosol (see Figure 2), while the intermembranous space is significantly more acidic than either of the two aforementioned compartments. The more alkaline pH of the matrix has been attributed to the translocation of protons across the inner mitochondrial membrane by components of the respiratory chain. For each pair of electrons translocated by the chain, two H"^ are extruded from the matrix into the intermembranous space. In accordance with the chemiosmotic theory, redox energy is thereby conserved as a transmembrane protonmotive force. For each oxygen atom that is reduced to water, six H"^ equivalents are transported out of the matrix. Additionally, because the process is electrogenic, an electrical potential (matrix negative) is also generated

Mechanisms of Intracellular pH Regulation

231

across the inner mitochondrial membrane, contributing to the overall electrochemical H"^ gradient (Emster and Schatz, 1981). The primary role of this electrochemical H"^ gradient in mitochondria is to provide free energy for the synthesis of ATP. This process is mediated by an ATPase present in the inner mitochondrial membrane, which given the appropriate H"^ gradient, can function as an ATP synthase. As external H"^ are driven back into the mitochondrial matrix through the enzyme, the energy is utilized for the phosphorylation of ADP to ATP. Two H"*" are transported inward per molecule of ATP synthesized. In addition to energizing the synthesis of ATP, the H"^ electrochemical gradient existing across the inner mitochondrial membrane is also used to propel the translocation of several solutes. Some ions, such as Ca^"^, are driven inward through a conductive pathway. Proteins can also be propelled into mitochondria by the electrical gradient. Other solutes, including organic and inorganic anions and cations, are either cotransported or countertransported with H"*" or possibly with OH". The functional significance of these H"^ (OH") coupled fluxes is not always understood. They have been suggested to play a role in mitochondrial volume regulation, in thermogenesis and also in preventing excessive accumulation of calcium in the mitochondrial matrix. Vacuolar System of the Cell In addition to the cytosol and mitochondria, cells have a vast network of membrane-bound structures involved in synthesis, secretion, and degradation of proteins and glycoproteins. It is important to realize that the luminal side of these organelles is topologically equivalent to the exterior of the cell. For example, soluble proteins secreted outside the cell are transported on the luminal side of a number of successive vesicular organelles until they are delivered to their final destination by fusion of the membrane of secretory granules (in the case of regulated secretion) or transport vesicles (in the case of constitutive secretion) with the plasma membrane. Similarly, the extracellular domain of intrinsic plasma membrane proteins faces the lumen of subcellular organelles when they are transported to, and removed from, the cell surface. Physiological Role of an Acidic pH in Subcellular Organelles Intercompartmental Transport The cell uses small vesicles to transport components between different compartments. Different subcellular organelles are involved in the transport of proteins and lipids along the exocytic pathway from their site of synthesis to their final destination, and also along the endocytic pathway to their site of degradation inside the cell. Several intracellular transport pathways are schematically depicted in Figure

Figure 2. Vacuolar system and intracellular transport pathways.of a typical cell. Schematic representation of Intracellular compartments known to have an intraluminal pH different from that of the cytosol. The intraluminal pH value for each compartment is shown next to the organelle. Large numbers represent transport pathways between different cellular compartments and are briefly described here. Pathway 1 : Regulated secretion. Secretory proteins are transported from the Golgi to secretory granules where they are concentrated. A vacuolar proton pump maintains the intraluminal pH at around 6.0-6.5. In some cells, the acidic interior is required for the final proteolytic processing of the secretory proteins (e.g., maturation of insulin in pancreatic p-cells), and in others, the proton gradient is used to transport small biogenic amines (e.g., catecholamines in adrenal cells) into secretory granules. Following stimulation (nervous or hormonal) of the cell, secretory granules deliver their contents by exocytosis. Pathway 2: Constitutive secretion. Secretory products and resident proteins of the plasma membrane are continuously carried by small transport vesicles from the Golgi to the plasma membrane where they deliver their contents by exocytosis. This pathway has been shown to be regulated by extracellular factors (nervous or hormonal) as well. Pathway 3: Receptor-mediated endocytosis. Ligands bind their receptor on the cell surface and are internalized in cIathrin-coated vesicles. The vacuolar H''-ATPase from bovine brain clathrin-coated vesicles has been extensively studied. After internalization, the clathrin coat is removed and the uncoated vesicle delivers its contents to the endosome. The low pH (6.0) of the endosome promotes dissociation of the ligands from its receptor. Pathway 4: Recycling of receptors back to the plasma membrane so that they can be used in further cycles of internalization. Pathway 5: Delivery of ingested material to the lysosome. The lysosome is the most acidic cellular compartment (pH 5.0) and its acidic interior is maintained by a vacuolar H'*'-ATPase. Ligands ingested by receptor-mediated endocytosis are delivered to the lysosome where they are degraded by hydrolytic enzymes. The acidic interior of lysosomes is required for optimal activity of all lysosomal enzymes, and the proton gradient has been proposed to drive the transport of degradation products out of the lysosome into the cytosol. Pathway 6: Transport of lysosomal enzymes from the Golgi to the endosome. Oligosaccharide chains of lysosomal enzymes are modified by two enzymes in early Golgi cisternae. The signal (mannose-6-phosphate) thus generated is recognized by a specific receptor in the trans-Golgi cisternae, which carries lysosomal enzymes to the endosome. Pathway 7: Recycling of mannose-6-phosphate receptor in the Golgi. Lysosomal enzymes are dissociated from their receptor in the endosome. The unoccupied mannose-6-phosphate is recycled back to the Golgi, and the lysosomal enzymes are transported to the lysosome together with the endocytosed material. The mitochondrial matrix is more basic than the cytosol due to the extrusion of protons into the intermembranous space by the components of the electron transport chain. The electrochemical gradient thus generated across the inner mitochondrial membrane Is used by the FQFI ATPase to generate ATP. The cytosolic pH is maintained at 7.2 through various mechanisms depicted in Figure 1. The abbreviations used are: SG, secretory granule; CP, coated pit; CCV, clathrin-coated vesicle; solid squares represent a ligand taken up by receptor-mediated endocytosis; solid triangles represent lysosomal enzymes; and small solid circles represent products secreted constituitively. 232

EXTRACELLULAR MILIEU pH - 7.3 - 7.4

I^Q^VY V

(QgjZ3l"

ENOOSOME pH-6.0

MITCX^HONDRIA

/

/

CYTOSOL

pH-7.2

MITCX^HONDRIAL MATRIX pH - 8.0 •^ATP

I. OUTER MITCX5H0NDRIAL MEMBRANE

?i%ure 2.

(Continued) 233

234

GREG GOSS and SERGIO GRINSTEIN

2 (pathways 1-7). All of these pathways use similar mechanisms to transport their cargo from one compartment to the next. The intercompartmental transport can be broken down into the following distinct steps: (a) formation and budding of a small transport vesicle from the compartment of origin with concomitant selection of proteins to be transported; (b) diffusion of the transport vesicle in the cytosol; (c) recognition of a cytoplasmic signal molecule on the surface of the transport vesicle by its corresponding receptor on the cytoplasmic surface of the destination compartment; (d) fusion of the membranes of the transport vesicle and the destination compartment, respectively, with ensuing delivery of the content of the transport vesicle to the lumen of the destination compartment; and (e) return of the transport vesicle (either empty or with a different cargo) to the compartment of origin. Although some features are specific for each step in a given transport pathway (especially those steps involved in the recognition between transport vesicles and the destination compartment), it is generally thought that many of the basic transport mechanisms are common to most transport pathways (endoc5^ic and exocytic) inside the cell. Remarkably, despite the rapid and extensive flow of membranes and proteins between intracellular organelles, each individual compartment manages to preserve its own identity. For example, all permanent residents of the endoplasmic reticulum are excluded during the formation of transport vesicles targeted to other organelles. Similarly, in order to accomplish unidirectional transport between compartments, a transport vesicle returning to its compartment of origin after delivery of its content in the destination compartment must exclude the components just delivered. In many transport pathways, the different ionic composition in the lumen of individual cellular compartments plays a central role in the unidirectional transport between two compartments. Two examples of transport pathways utilizing an asymmetrical proton composition to establish the directionality of transport are described. Receptor-mediated endocytosis. The cellular uptake of the cholesterolcarrying low density lipoproteins (LDL) by liver cells provides an excellent and well characterized example of receptor-mediated endocytosis. A receptor present at the cell surface of hepatocytes binds LDL specifically, and can exist in two different conformations. At neutral pH the conformation of the receptor is such that it binds its ligand (LDL) with high affinity (tight binding). However, when exposed to an acidic environment, the LDL receptor assumes a different conformation such that its affinity for LDL is very low (no binding of LDL takes place). When the LDL receptor is exposed outside the cell (facing a medium of neutral or slightly alkaline pH), it binds LDL with high affinity. Following binding of its ligand, the receptor is internalized and the receptor-ligand complex is delivered to a compartment with a slightly acidic luminal pH (6.0-6.5): the compartment of uncoupling of receptor and ligand (CURL). In this acidic compartment, the receptor

Mechanisms of Intracellular pH Regulation

235

undergoes a conformational change resulting in uncoupling of its ligand. The receptor is then recycled back to the cell surface where it will regain its original conformation at neutral pH and will therefore be able to bind LDL for further cycles of internalization and delivery of LDL inside the cell. Thus, the unidirectional delivery of LDL inside the cell requires three essential elements: (a) the affinity of the receptor for its ligand must be pH-sensitive; (b) transport vesicles containing the LDL receptor have to shuttle constantly back and forth between two compartments; and (c) these compartments must have the ability to maintain a different intraluminal pH. Delivery of lysosomal enzymes. The vast majority of integral membrane proteins and soluble proteins destined for secretion or for permanent residence in the different subcellular organelles (with the notable exception of mitochondria and chloroplasts) are translocated to the luminal side of the rough endoplasmic reticulum membrane during their synthesis. These proteins are transported together (by vesicular transport) through several distinct cellular compartments until they reach the outermost cisternae of the Golgi apparatus (called the trans-Golgi network, or TGN). Once in the TON, proteins are sorted and directed to their final destination: the plasma membrane, secretory granules, or lysosomes. Although the sorting mechanisms are not fully understood for most destinations, the delivery of lysosomal enzymes to the lysosome has been well characterized. The oligosaccharide chain of lysosomal enzymes is modified in the early cisternae of the Golgi apparatus (the c/^-cisternae) by two enzymes working sequentially. The first enzyme recognizes a three-dimensional motif unique to lysosomal enzymes and transfers the phosphorylated sugar N-acetylglucosamine-phosphate on carbon 6 of mannose residues on the oligosaccharide chains of lysosomal enzymes. The second enzyme removes the N-acetylglucosamine sugar, thus exposing the phosphorylated mannose residue. In the TGN, a receptor specific for the mannose-6-phosphate group (M6P) binds lysosomal enzymes carrying the M6P signal. Small transport vesicles bring the M6P receptor-ligand complex to a pre-lysosomal compartment with an acidic luminal pH: the endosome. As was the case for LDL, binding of M6P to its receptor occurs at near neutral pH, but the affinity of the M6P receptor for its ligand is significantly lower in the acidic environment of the endosome. The lysosomal enzymes are thus delivered into the endocytic pathway and the M6P receptor is recycled back to the TGN for further cycles of transport. In both receptor-mediated endocytosis and delivery of lysosomal enzymes to the lysosome, the different intraluminal pH of two compartments provides the asymmetry necessary to drive the unidirectional transport between two cellular compartments. It is important to note that although the acid-induced uncoupling of the receptor-ligand complex is a recurring theme in intracellular transport, a similar asymmetry could also be provided (for the transport of other ligands) by differences, other than pH, in the intraluminal compositions of the two compartments involved.

236

GREG GOSS and SERGIO GRINSTEIN

Activation of Lysosomal Hydrolases The lysosome is the compartment where cellular components that are no longer needed and molecules endocytosed from the extracellular milieu are degraded. The lumen of the lysosome is the most acidic of all cellular compartments (pH <5.0), and lysosomal hydrolases (e.g., proteases, lipases, ribonuclease, and deoxyribonuclease) function optimally at such acidic pH values. The acidic optimal pH for lysosomal enzymes provides a double safety feature for the cell. First, it ensures that these hydrolytic enzymes are inactive while in transit in the different cellular compartments (rough endoplasmic reticulum, Golgi, and endosome) before they reach their final destination. If this was not the case, the molecular constituents of several cellular compartments would be susceptible to their hydrolytic activity. Second, it guarantees that if lysosomal enzymes escape the lysosome or are misrouted to a compartment other than the lysosome, they will be inactive and will not hydrolyze useful cellular constituents. Transport of Breakdown Products Out ofLysosomes It has been postulated that the H"*" gradient across the lysosomal membrane could be used by a co-transporter to expel breakdown products (amino acids, nucleotides, etc.) out of the lysosome and into the cytosol, where they can be used by the biosynthetic machinery of the cell. The lysosomal co-transporter is thought to be similar to that found in many bacteria which co-transports H"^ and small molecules into the cell. Uptake of Components into Secretory Granules In many secretory vesicles, such as chromaffin granules of the adrenal medulla, platelet dense granules, and synaptic vesicles, the acidic interior drives the vectorial uptake of biogenic amines (epinephrine, serotonin, and catecholamines, respectively). This uptake is accomplished by a countertransporter which exchanges cytoplasmic biogenic amines for intragranular protons. Maturation of Secretory Products In some endocrine glands, the acidic interior of secretory granules plays an important role in the proteolytic conversion of pro-hormones to their biologically active form. This was demonstrated for the conversion of pro-insulin to mature insulin in the secretory granules of pancreatic islets. The pro-hormone is cleaved by a specific protease yielding the disulfide-linked A and B chains, and the inactive intervening C peptide. This proteolytic conversion depends entirely on the acidification of the lumen of the secretory granules.

Mechanisms of Intracellular pH Regulation

237

Entry of Pathogens

While the acidic interior of several compartments is essential for many cellular activities, various pathogens also use this biochemical asymmetry to their advantage. Several membrane-enclosed viruses, such as influenza, Semliki forest, and vesicular stomatitis viruses, bind to receptors on the cell surface and are taken up by receptor-mediated endocytosis. When the virus enters the endosome, which has a luminal pH of approximately 6.0, one of the viral intrinsic membrane proteins undergoes a conformational change leading to the fusion of the viral membrane with that of the endosome. The fusion of the two membranes results in the delivery of the viral nucleic acid into the cytosol, thus initiating cellular infection. Some bacterial toxins, such as diphtheria toxin, also use the acidic interior of some compartments of the endoc3^ic pathway in order to gain access to the cytosol. Mechanisms of Acidification of Intracellular Compartments

The interior of the intracellular compartments discussed in the previous section is maintained at a pH different from that of the cytosol by the action of ATP-driven proton pumps, or H"^-ATPases. These transmembrane proteins couple the hydrolysis of ATP to the vectorial transport of protons across the lipid bilayer. Two different categories of H'^'-ATPases have been characterized as playing a central role in regulation of cytosolic and compartmental pH. F^FoATPase

This type of H'^'-ATPase is found in mitochondria, chloroplasts, and bacteria. The reaction catalyzed by FIFQ ATPases is fully reversible. Because of the existing proton gradient across the membrane where they are localized, all FIFQ ATPases can function in a direction opposite to that of other proton pumps, namely in the direction of ATP synthesis, and are therefore often referred to as ATP synthetases. In mitochondria, for example, the F i FQ ATPase is located in the inner mitochondrial membrane and couples the synthesis of ATP to the flow of protons back into the mitochondrial matrix. This reaction has already been described in detail above. It is important to remember that the ATPase is not primarily responsible for generation of the mitochondrial pH, but instead makes use of it for ATP generation. Vacuolar ht'-ATPases

All vacuolar H'^-ATPases actively transport protons from the cytosol to the luminal side of the lipid bilayer and are therefore involved in the acidification of the lumen of many subcellular organelles. They are very similar, if not identical to those present and active in the plasma membrane of specific cell types, which were

238

GREG GOSS and SERGIO GRINSTEIN

discussed earlier. However, the bulk of information concerning the biochemical, pharmacological, and structural characteristics of V-type ATPases has been obtained through analysis of organelles. The V-type ATPases are the predominant system involved in the regulation of compartmental pH. Due to the vital function of this transporter in maintaining such cell function, considerable research has been conducted to elucidate its properties. Some of the common characteristics of organellar V-type ATPases are listed below. Distribution. V-type H'^-ATPases have been identified and/or partially purified from several subcellular organelles of diverse tissues and species. They range from manmialian brain clathrin coated vesicles, adrenal chromaffin secretory granules, lysosomes, endosomes, and Golgi apparatus, to vacuoles in plants, Neurospora and yeast. As mentioned, V-type ATPases have also been shown to be present in the plasma membrane of specific cell types where they may function in cytosolic pH regulation (see above). Despite this wide range of distribution, their basic structural integrity is remarkably well preserved, although their function in each locale is highly specific. For example, the yeast vacuolar pH is maintained around pH 6.1 while the mammalian lysosome maintains a pH <5.0. Inhibitor susceptibility. V-type ATPases can be distinguished by their extreme sensitivity to bafilomycins and concanamycins (Ki = 0.02-1.3 nM), a family of antibiotics isolated from Streptomyces sp. Other less selective inhibitors of this type of ATPase include N-ethylmaleimide (NEM) and dicyclohexylcarbodiimide (DCCD). Inhibitors of the FIFQ ATPases, such as oligomycin, do not inhibit the V-type ATPases. Mechanism of action. Vacuolar H"^-ATPases do not transport any other ion (either in the same or in the opposite direction) together with H"^, the way many ion pumps do. The unidirectional transport of positive ions, therefore, generates a membrane potential across the lipid bilayer (luminal side positive), and vacuolar H'^-ATPases are thus said to be electrogenic. Interestingly, vacuolar H"^-ATPases tend to be inhibited by the membrane potential they generate. In the cell, the membrane potential is dissipated by the entry of monovalent anions, most likely C\r, into the lumen of the subcellular organelle. This C\~ conductance therefore prevents the electrical self-inhibition of the proton pump. This can be elegantly demonstrated experimentally. When a subcellular organelle containing a vacuolar H"^-ATPase is incubated with ATP in a buffer containing CI", continuous hydrolysis of ATP and a concomitant acidification of the interior of the organelle are observed. In contrast, when C\~ anions are omitted from the incubation buffer, the vacuolar H"*"-ATPase is quickly inhibited and no acidification occurs. Subunit composition and function. Vacuolar proton pumps have been purified from a variety of sources, but the vacuolar H"*"-ATPases from bovine brain clathrin coated vesicles and from adrenal chromaffin granules are the most extensively

Mechanisms of Intracellular pH Regulation

239

characterized. The vacuolar H"^-ATPase from each source has been purified to homogeneity and consists of a multisubunit transmembrane protein complex. The native clathrin coated vesicle proton pump is composed of three copies each of 73 kDa and 58 kDa nucleotide binding subunits, six copies of a 17 kDa subunit and one copy of each of six remaining subunits. The 73 kDa and 58 kDa subunits are responsible for ATP binding. As mentioned above, proton translocation by vacuolar ATPases is inhibited by DCCD. Radioactive DCCD was shown to react specifically with the 17 kDa subunit of the vacuolar H'^-ATPase, suggesting that proton translocating activity can be assigned to this subunit. The high degree of homology of these proteins with those in other classes of ATPases (e.g., FIFQ), particularly in the nucleotide binding sequence, suggests a common ancestry. The clathrin coated vesicle pump is separated into two domains, the membrane bound VQ (integral) domain consisting of the 17 kDa subunits listed above, plus one copy of each of a 100 kDa, 38 kDa, and 19 kDa subunit. The Vi (peripheral) domain consists of proteins of approximately 73 kDa, 58 kDa, 40 kDa, 34 kDa, and 33 kDa (Forgac, 1992). The specific function of each of the subunits is currently under investigation as they are thought to perhaps hold the key to the differential regulation of vacuolar H"^-ATPases in different compartments (see below). Control and Regulation of Vacuolar H^-ATPases While the different vacuolar H"^-ATPases share a number of common functional and structural characteristics, it is startling to observe that compartments containing these proton pumps have strikingly different intraluminal pH values (see Figure 2). Furthermore, there is increasing evidence that certain vesicles containing the H'*'-ATPases are not active until they are specifically localized in the cell (e.g., clathrin-coated vesicles during receptor-mediated endocytosis are not acidified until they reach the CURL region, see above). How do these seemingly similar H"^ pumps adjust the pH of each subcellular organelle to different values and what are possible mechanisms for turning the pump on and off at specific locations in the cell? Several possible explanations have been examined in recent years. Modulation of Activity The activity of individual H"^-ATPases may be directly controlled by the action of modulating proteins present in some compartments but not in others. Modulation of plasma membrane H"^-ATPases in renal collecting ducts by a 35 kDa cytosolic factor has been shown to occur. This protein which is thought to bind to the H"*"-ATPase complex at low cytosolic pH activates the brush border (apical plasma membrane) pumps. Plasma membrane pumps are activated by the factor to a greater extent than lysosomal pumps, but the reason for this differential regulation has not yet been established.

240

GREG GOSS and SERGIO GRINSTEIN

Alternatively, covalent modifications of various subunits of the proton pump could take place exclusively in some compartments, and consequently modulate their activity. Recent reports suggest that formation of disulfide bonds between cysteine residues in the 73 kDa subunit inactivate the H'^-ATPase. It has been suggested that interconversion between active and inactive states can regulate vacuolar acidification in vivo, although the mechanism(s) controlling this intercoversion have not yet been established. Conductance to Monovalent Anions As mentioned above, H^-ATPases are electrogenic and inhibited by the formation of a membrane potential across the bilayer. The influx of monovalent anions (generally CI") across the membrane prevents this inhibition by dissipation of the electrochemical gradient. One way to achieve different levels of acidification is to control the conductance to monovalent counterions. The different intraluminal pH of lysosomes and endosomes could be controlled by differential C\~ (or other counterion) conductance of their respective membranes (higher conductance in the lysosome), despite the presence of the same number of identical H"*"-ATPases in both compartments. The larger membrane potential that would be generated in the endosomal compartment would be more inhibitory to its H'^^-ATPases, resulting in a higher intraluminal pH as compared to that of the lysosome. Recent evidence has shown that in vitro, the coated vesicle chloride channel is phosphorylated by protein kinase A, resulting in modulation of both conductance and acidification. Subunit Composition As mentioned previously, a number of subunits are common to all vacuolar H'^-ATPases described so far. However, some H'*"-ATPases have additional and/or different subunits. It is possible that the subunits common to all H^-ATPases provide the basic function of proton pumping coupled to ATP hydrolysis and that unique additional subunits control the level of acidification attained in each individual organelle. Reassembly of the monomeric peripheral subunits with the integral subunit (VQ) into a functional V-type ATPase has been demonstrated but activity of the pump is not dependent on the presence of the full complement of subunits. Thus, absence of the 40 kDa subunit during functional reassembly results in a 50% decrease in pump activity but does not completely inhibit the pump.

CONCLUDING SUMMARY Precise control over intracellular pH, both cytosolic and intraorganellar, is critical to the proper functioning of the cell. Intracellular membrane traffic, endocytosis and receptor recycling, and cytosolic enzyme function are but a few of the myriad processes which are pH dependent. This chapter summarizes the recent advances

Mechanisms of Intracellular pH Regulation

241

in our understanding of some of the major mechanisms responsible for this precise regulation. In addition, a brief methodological section is provided. Finally, the functional significance of cj^osolic and organellar pH is discussed in some detail.

REFERENCES Al-Awqati, Q. (1986). Proton translocating ATPAses. Ann. Rev. Cell. Biol. 2, 179-193. Boron, W.F. (1983). Transport of H"^ and of ionic weak acids and bases. J. Membrane Biol. 72, 1-14. Bretag, A.H. (1987). Muscle chloride channels. Physiol. Rev. 67, 618-645. Cabantchik, Z.I., Knauf, P.A., & Rothstein, A. (1978). The anion transport system of the red blood cell. Biochim. Biophys. Acta 515, 239-302. Emster, L. & Schatz, G. (1981). Mitochondria: A historical review. J. Cell. Biol. 91, 227-239. Forgac, M. (1992). Structure, functions and regulations of the coated vesicle V-ATPase. J. Exp. Biol. 172, 155-169. Foskett, J.K. & Grinstein, S. (1990). Noninvasive techniques in cell biology. In: Modem Cell Biology, Vol. 9 (Satir, B., ed.), Wiley-Liss, New York. Gadian, D.G. (1982). Nuclear Magnetic Resonance and its Applications to Living Systems. Oxford University Press, New York. Lukacs, G., Kapus, A., Nanda, A., Romanek, R., & Grinstein, S. (1993). Proton conductance of the plasma membrane: Properties, regulation and functional role. Am. J. Physiol. 265, C3-C14. Madshus, I.H. (1988). Regulation of intracellular in eukaryotic cells. Biochem. J. 250,1-8. Roos, A. & Boron, W.F. (1981). Intracellular pH. Physiol. Rev. 61, 296-434. Thomas, R.C. (1978). In: Ion Sensitive Intracellular Microelectrodes: How to Make Them and Use Them. Academic Press, New York. Thomas, R.C. & Meech, R. W. (1982). Hydrogen ion currents and intracellular pH in depolarized voltage clamped snail neurons. Nature 229, 826-828.

RECOMMENDED READINGS Foskett, J.K. & Grinstein, S. (1990). Noninvasive techniques in cell biology. In: Modem Cell Biology, Vol. 9 (Satir, B., ed.), Wiley-Liss, New York. Grinstein, S. (1988). NaVH"^ Exchange, CRC Press Inc., Boca Raton, FL. Nuccitelli, R. & Deamer, D. W. (1982). In: Intracellular pH: Its Measurement, Regulation and Utilization in Cellular Functions. Alan R. Liss. New York. Harvey, W.R. & Nelson, N. (1992). V-ATPases, J. Exp. Biol. Volume 172 (Supplement).

This Page Intentionally Left Blank

Chapter 8

The Membrane Na"*"-K^-ATPase in Cells THOMAS A. PRESSLEY

Introduction Na K -ATPase and the Origin of Plasma Membrane Ionic Gradients The Pump-Leak Model Discovery of the Na'^-K'^-ATPase Cell Physiology of Na'^-K"^-ATPase Structure and Catalysis Subunit Composition Catalytic Cycle Subunit Heterogeneity Regulation of Na-K"^-ATPase Hormonal Influence on Na -K -ATPase Abundance Substrate Regulation of Catalytic Turnover Pump Inhibition and Increased Cardiac Contractility Summary

243 244 244 245 246 246 246 247 249 250 250 251 251 253

INTRODUCTION Over 200 years ago the British physician, William Withering, published a landmark monograph on the treatment of dropsy, a pulmonary and systemic edema which we would recognize today as a symptom of congestive heart failure (Withering, 1785).

Principles of Medical Biology, Volume 4 Cell Chemistry and Physiology: Part II, pages 243-253 Copyright © 1996 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-806-4

243

244

THOMAS A. PRESSLEY

Presented with a recipe of various herbs used as a folk medicine, he recognized the purple foxglove, Digitalis purpurea, as the important ingredient, and he spent the next ten years evaluating its therapeutic value. Digitalis, the purified agent from Dr. Withering's foxglove extracts, continues to be a potent drug in treating this life-threatening affliction. Indeed, digitalis represents one of a series of steroidlike compounds, the cardiac glycosides, that help relieve the symptoms of heart failure by increasing cardiac contractility. The pharmacological target of the glycosides is the Na"*'K'^-ATPase, a membrane transport system crucial to ionic homeostasis.

NA^-K^-ATPASE AND THE ORIGIN OF PLASMA MEMBRANE IONIC GRADIENTS The Pump-Leak Model

The cytosol of most animal cells is enriched in K"*^ and deficient in Na"^, relative to the extracellular fluid. Because the plasma membrane is permeable to both cations, one might expect that over time, Na"*" would diffuse into the cell and K"^ would leave, yet the electrochemical gradients for these ions persist. To explain this paradox, physiologists in the 1940s postulated the existence of a "Na'*"-K'^pump" that would extrude Na"^ from the interior of the cell and absorb K"*^ at the expense of metabolic energy (Dean, 1941). This transport of Na"*" and K"*^ against their electrochemical gradients would exactly balance the passive "leak" down their gradients, resulting in the distribution of cations observed at steady-state (Figure 1). Numerous observations provide support for the pump-leak model. In isolated cells, radiotracer studies show that rates of Na"^ entry and exit are equal, as are those for K"^. That the efflux of Na"^ and the influx of K"^ are coupled is demonstrated by the inhibition of Na"^ extrusion when K"^ is removed from the extracellular fluid.

Figure 1, The pump-leak model. The high intracellular concentration of K"^ and the low concentration of Na"^ relative to the exterior is explained by the actions of a pump that extrudes Na"^ and absorbs K"^ at the expense of metabolic energy. These movements are balanced by passive diffusion of the cations down the resulting electrochemical gradients, producing a steady state.

Na^-lC-ATPase

245

Poisons such as cyanide or dinitrophenol inhibit the efflux of Na^ and the influx of K"*", as expected if metabolic energy is required to move these cations against their electrochemical gradients. Moreover, the pump-mediated fluxes are inhibited by exposure of the cell to ouabain, a relatively water soluble cardiac glycoside. The importance of this observation will soon become apparent. Discovery of the Na^-K^-ATPase

Although transport studies such as those described above suggested the existence of the Na'^-K'^-pump, its biochemical nature remained unknown until the late 1950s. Working with homogenates of crab nerve, a Danish physiologist, Jens Skou, characterized an enzymatic activity that hydrolyzed Mg^"^-ATP, but only in the presence of Na"^ and K"*" salts (Skou, 1957). Neither cation alone would sustain the reaction. Others quickly confirmed the existence of this Na'^,K'^-dependent ATPase activity in other tissues. Because the hydrolysis of ATP could link the expenditure of metabolic energy to the transport of Na"^ and K"^ against their electrochemical gradients, the Na'^-K'^-ATPase seemed a promising candidate for the proposed Na"^-K"*"-pump. Additional studies have confirmed that the Na'^K'^-ATPase is indeed the enzymatic equivalent of the Na'^-K'^-pump observed in ion transport studies. Like the pump, the Na'^-K'^-ATPase is localized to the plasma membrane of nonpolar cells and the basolateral membrane of most epithelial cells. Its enzymatic activity correlates with the level of pump-mediated Na'^-K"^ transport, and changes in ion concentration have similar effects on both. The orientation of the Na'^-K"^-ATPase within the membrane is also consistent with that expected for the pump (Figure 2). To reach this conclusion, physiologists made extensive use of resealed erythroc3^e membranes. Red cells are lysed in hypotonic medium, and after the cytosol has diffused out, the membranes are resealed in an appropriate isotonic solution. Using

Figure 2. Orientation of the Na'^-K'^-ATPase. Hydrolysis of ATP by a resealed erythrocyte membrane is only observed if the ATP and Na"^ are on the interior and the K"^ is on the exterior. Cardiac glycosides such as ouabain only inhibit the enzyme when added to the extracellular fluid.

246

THOMAS A. PRESSLEY

such red cell "ghosts", ATP was shown to be hydrolyzed only when on the intracellular side of the membrane. Activation of the ATPase only occurred when Na"^ was available on the intracellular side and K"^ on the exterior. As expected from their inhibitory effects on Na'^-K'^ transport, ouabain and other cardiac glycosides are potent blockers of the Na"^-K'*'-ATPase, but only when applied on the extracellular side. Any remaining doubt about the equivalence of the pump and ATPase was dispelled by reconstitution experiments. Highly-purified Na"*"K'^-ATPase was shown to mediate transport of Na"^ and K"^ against their electrochemical gradients when incorporated into lipid vesicles. By these observations and others, the Na"^K"^-pump postulated from the pump-leak model has been given a firm biochemical basis. Cell Physiology of Na'^-K^-ATPase The actions of the Na'^-K'^-ATPase have consequences far beyond the immediate effect on Na"^ and K"^ transport, and the pump plays a central role in cellular function. Together with the selective permeability of the plasma membrane, the catalytic turnover of the pump and the resulting concentration gradients are responsible for the transmembrane potential difference. In excitable cells, this potential difference creates the initial conditions on which action potentials are generated and propagated. Cellular volume is also controlled indirectly by the ion distribution created by the pump. The energy stored within the transmembrane Na"^ gradient is utilized by a number of cotransport systems to regulate the intracellular concentrations of other ions and solutes. For example, Na"*"-dependent glucose uptake is responsible for the absorption of sugar by the intestine against a concentration gradient. NaVamino acid transporters are present as well. Similarly, a Na"^/!" cotransporter in the thyroid enables the gland to accumulate iodine. Transport epithelia such as those in the kidney and intestine depend on the electrochemical gradients and cotransport systems for transepithelial movement of solutes. All this transport of Na"^ and K"^ represents a substantial energy sink for the cell, and Na'*'-K'*^-ATPase may be a major pacemaker for metabolic turnover.

STRUCTURE AND CATALYSIS Subunit Composition The Na'^-K'^-ATPase has been purified to near-homogeneity, usually from epithelia or excitable tissue capable of high transport rates such as the outer medulla of the kidney or electroplax fi-om electric fish. It exists as a membrane-spanning complex with two dissimilar protein subunits: a 110,000 Da a subunit and a 55,000 Da p subunit (Figure 3). Although the subunits are present in equimolar amounts within the purified Na'^'-K'^-ATPase, the actual number of subunits within the

247

Na'-lC'ATPase

cytosol Figure 3. Structure of the Na^-K'^-ATPase. The catalytic a subunit crosses the plasma membrane at least seven times; the glycosylated p subunit crosses once. It is unclear whether the functional complex exists as a single pair of subunits, as depicted here, or as two pairs.

functional complex remains controversial, and it is unclear whether the enzyme consists of one or two copies of each subunit. The a subunit traverses the plasma membrane at least seven times. Its folds contain extracellular domains believed to be involved with the binding of K"^ and cardiac glycosides, as well as extensive cytoplasmic domains that contain the binding sites for Na"*" and MgATP. During the catalytic cycle, one of these cytosolic domains is phosphorylated transiently. For these reasons, a is usually described as the catalytic subunit. Despite the association of most catalytic structures with the a subunit, normal enzymatic activity requires the presence of the P subunit. This subunit is thought to cross the plasma membrane once. Most of its structure consists of an extracellular domain that contains N-linked oligosaccharide chains on as many as three sites. About a third of the subunit's molecular weight is attributable to this extensive glycosylation. Several lines of evidence suggest that the p subunit may play a role in catalysis, perhaps by contributing to the substrate binding sites of the active enzyme, but the details of P's function are unclear. Alternatively, the p subunit may not be involved with catalysis, yet it may be required for proper maturation and processing of the pump during biosynthesis or targeting of the completed complex to the plasma membrane. Catalytic Cycle

Given that the enzyme complex matures correctly and is inserted into the plasma membrane, just how does it convert the free energy released from ATP hydrolysis

248

THOMAS A. PRESSLEY ATP

Na+ El

ADP

-^^:i^—^—•

-^^=-

Ei-P(Na+)

A Na+ E2(K*)

^

^ P|

v^

E2-P K-^

\

Ouabain

figure 4. Simplified model of the catalytic cycle. The Na"^-K"'-ATPase oscillates between two conformations, Ei and E2, and is phosphorylated transiently. Binding of cardiac glycosides increases the stability of the E2-P form of the enzyme.

to the movement of Na"*" and K"^ against their electrochemical gradients? Although a detailed explanation of the molecular events will probably require the elucidation of the pump's three-dimensional structure, studies using highly purified Na'^-K'*"ATPase have provided a crude picture of the catalytic cycle. A major clue is the transient phosphorylation of the a subunit mentioned above. It was found that in the presence of Na"^ and MgATP, but not K"^, the a subunit was phosphorylated at a specific aspartyl residue within the major cytoplasmic domain. The addition of K"*" would promote dephosphorylation. The p-aspartyl phosphate formed during these reactions presumably acts as a common chemical intermediate in the transfer of free energy from the hydrolysis of ATP to the movement of cations. A second clue came to light when it was discovered that digestion of the enzyme with proteases such as trypsin yielded different products in the presence of Na"^ and MgATP, as opposed to K"^. It was proposed that the Na'^-K^-ATPase existed in at least two conformations, Ej and E2, and their existence has been confirmed by changes in fluorescence of the enzyme under the two conditions. Phosphorylation and interconversion of conformations have been incorporated into a very simplified kinetic model for the catalytic cycle (Figure 4). In the reaction that occurs under normal conditions, the Ei conformation of the enzyme binds Na^ and MgATP to their sites on the cytosolic side. Transfer of the y phosphate from ATP to the a subunit produces a relatively unstable EpP form that shifts into the E2-P conformation. During this change in conformation, the bound Na"^ gains access to the extracellular fluid and is released. The E2-P form then binds K"^ at an extracellular site, promoting the dephosphorylation of the subunit and a shift back to the El form. The bound K"^ is then released into the cytosol, and the cycle begins again. This whole process occurs about 100 times per second when the pump is cycling at its maximal rate. Cardiac glycosides inhibit the pump by binding to the E2-P form of the enzyme from the extracellular side, increasing the stability of the conformation. In contrast, vanadate (VO^^) inhibits turnover by binding to the phosphorylation site on the interior, probably because it resembles a transition state in the phosphate transfer reaction.

Na^'lC-ATPase

249

A limitation of the model, as presented, is that it does not account for the observed stoichiometry of the Na'^-K'*"-ATPase. Careful measurements have indicated that for each ATP hydrolyzed, three Na"^ are transported out of the cell and two K"^ are moved in. Because a net charge of+1 is moved out of the cell with each cycle, the pump produces an electric current. Such current-generating transporters are said to be electrogenic. This pump-mediated current contributes to the potential difference across the plasma membrane, although its influence on the magnitude of that difference varies with cell type. One way to assess the electrogenic component of the potential difference is to poison the Na'^-K'^-ATPase with a cardiac glycoside such as ouabain, and then monitor the change in voltage across the membrane with a microelectrode. One must consider only the change produced immediately, however, because in the absence of functioning pumps, the remaining potential difference will gradually decrease as the gradients dissipate. Subunit Heterogeneity For many years the discussion presented thus far would have summarized our knowledge of the Na"*"-K"*"-pump and its structure, but the picture has become considerably more complicated. It is now clear that what we have been calling the Na'^-K'^-ATPase is actually a family of pumps. Molecular cloning has identified at least three isoforms of the a subunit in mammals (Shull et al., 1986). The three a isoforms have nearly identical primary structures (about 85% identity). They are encoded by different mRNAs and are the products of different genes. The a l isoform appears to be expressed in all tissues and is the form originally isolated from kidney. The many biochemical experiments on which the kinetic model is based were probably performed on a l . The a2 isoform is expressed principally in nervous tissue, heart, skeletal muscle, and adipocytes. The a3 isoform is expressed principally in nervous tissue. The expression of the various a isoforms also varies with development and hormonal status. Because of its widespread expression, the a l isoform probably functions as a housekeeping form of the enzyme, fulfilling its role in the pump-leak model described above. By similar reasoning, the a2 and a3 isoforms may play tissue-specific roles. Molecular cloning has also identified isoforms of the p subunit. Investigators working on neuron-astrocyte recognition isolated a nucleotide sequence encoding an adhesion molecule that mediates cell interactions (Gloor et al., 1990). Closer inspection revealed structural similarities with the known form of the P subunit, and it was soon obvious that the adhesion molecule was a second p isoform. This finding has raised exciting possibilities about the role of the Na"^-K^-ATPase in functions unrelated to cation transport. Additional heterogeneity among the P isoforms is introduced by cell-specific glycosylation patterns. Determination of the primary structures for the subunits of the Na^K'^-ATPase also confirmed its membership in a gene superfamily that includes the Ca^"^-pumps

250

THOMAS A. PRESSLEY

and H'^K'^^-pumps. They are all P-type ATPases, so called because they undergo transient phosphorylation during turnover. These transport enzymes share many structural motifs and undergo E1-E2 transitions. All are inhibited by vanadate; however, only the Na'^-K'*'-ATPase is inhibited by cardiac glycosides. Like the Na"^-K'*'-ATPase, the a subunit of the gastric H"^-K^-ATPase requires a P subunit for function, and gene transfer experiments have shown that the H^-K"^-ATPase p subunit can function with the Na"*"K"^-ATPase a subunit. As of yet, there does not appear to be any evidence of a P subunit associated with any of the Ca^"^-ATPases.

REGULATION OF Na^K^-ATPase Hormonal Influence on Na^-K^-ATPase Abundance

As might be imagined for an enzyme that plays such a pivotal role in cellular homeostasis, the Na'*'-K'^-ATPase is subject to tight regulatory control. Mechanisms are in place within the cell to control both the number of functional Na'^'K^-pumps within the membrane and the catalytic rate of those pumps. At least three classes of hormones have been implicated in the long-term regulation of pump abundance: thyroid hormones, mineralocorticoids, and glucocorticoids. An increase in the circulating concentrations of these hormones produces an up-regulation of the Na'^-K"*'-ATPase in target tissues that is attributable, at least in part, to enhanced biosynthesis. Binding of the hormones to high-affinity nuclear receptors produces increases in transcription rate and, ultimately, an increased abundance of the mRNAs encoding the pump. Accordingly, these hormones require hours to elicit their fiill effect. Regulation by the adrenal steroids is linked to salt and nutrient balance, and their target tissues are principally transport epithelia. The role played by the thyroid hormones is less obvious. The regulation of Na'*"-K"^-ATPase by the thyroid hormones appears to be one of several mechanisms elicited to increase metabolic rate during thyroid thermogenesis in target tissues (Ismail-Beigi and Edelman, 1970). Thyroid hormones produce an increase in pump abundance, as well as an increase in membrane passive permeability (Figure 5). These combine to enhance dramatically the catalytic turnover of the enzyme, creating an energy sink that consumes ATP. The ADP generated by the pump, in turn, stimulates oxidative phosphorylation, which yields more ATP. The increase in overall metabolic rate results in the characteristic elevations in oxygen consumption and heat production elicited by the thyroid hormones. In the heart, recent evidence suggests that the o2 isoform responds more dramatically to thyroid hormone than a l . A number of additional hormones also influence Na^-K"*"-ATPase activity, although the mechanisms of action are not as well understood. Rather than altering gene expression, these hormones appear to work at the level of the pump itself, and their actions occur within minutes. In muscle and adipose tissue, insulin appears to

Na^'lC'ATPase

251 Mitochondrion

Plasma Membrane

Heat

Fuels

Na+

Figure 5. Na'^-K'^-ATPase in thyroid thermogenesis. Increased pump abundance and cation permeability in response to thyroid hormones elevates the rate of catalytic turnover in target tissues. The resulting increase in ATP utilization stimulates oxidative phosphorylation, and ultimately oxygen uptake, fuel consumption, and heat production.

alter the affinity of the a2 isoform to intracellular Na"*", leading to an increase in Na'*"K"^-ATPase activity. Recruitment from intracellular compartments may also play a role. Vasopressin and the catecholamines also increase pump activity. As will be discussed below, at least part of the explanation may be a hormone-induced increase in the entry rate for Na"^. In addition, the a isoforms contain a number of potential targets for hormone-stimulated protein kinases. Phosphorylation of these sites may also modulate the pump's activity. Substrate Regulation of Catalytic Turnover Moment-to-moment control of the Na"^-K'^-ATPase is achieved by substrate-dependent regulation. In most cells, the rate of pumping is limited by the intracellular Na"^ concentration, which is well below that required to saturate the enzyme. In contrast, the intracellular concentration of MgATP and the extracellular level of K"^ are near saturation with respect to the pump. If the rate of Na"^ entry increased, as might happen in a neuron that has fired repeated action potentials, the concentration of intracellular Na"^ would begin to increase. This, in turn, would stimulate the Na'^K'^-ATPase to cycle faster. The resulting slight increase in the rate of pumping will compensate for the faster Na"^ leak, minimizing the perturbation. The increase in epithelial Na'^-K'*' transport elicited by some hormones, such as vasopressin, is explained in part by their effects on Na"^ entry and the subsequent stimulation of Na"'-K"'-ATPase activity. Pump Inhibition and Increased Cardiac Contractility We are finally in a position to consider the mechanism by which the cardiac glycosides produce their therapeutic action on the heart, because similar substrate-

252

THOMAS A. PRESSLEY

Cardiac l^+ Glycoside

\

Na-^K-*--transport

Y Ca++ extrusion

Figure 6. Pharmacological action of the cardiac glycosides. Partial inhibition of the Na''-K''-ATPase in cardiac myocytes elevates the intracellular concentration of Na"*^. This in turn decreases the rate of Ca^"" extrusion via its Na'^-dependent cotransporter. The resulting increase in intracellular Ca^"*^ increases cardiac contractility.

dependent regulatory changes occur when the Na'^-K'^-ATPase is inhibited. Administration of the drug causes partial inhibition of the Na"*'K"^-ATPase within cardiac myocytes (Figure 6). The dosage must be carefully controlled, and presumably, inhibition is incomplete because the drug is provided at submaximal concentrations. On the other hand, the heart expresses more than one isoform of the pump, and it may well be that inhibition favors a particular isoform. Either way, the resuh is a decrease in the rate of pumping and the subsequent accumulation of intracellular Na"*". The remaining, uninhibited pumps respond to the elevated Na"^ levels by increasing the pumping rate, and a new steady-state is soon achieved. One important consequence of the elevated intracellular Na"*", however, is a decrease in the rate of Ca^"^ extrusion via the Na'^/Ca^'*" exchanger, also located in the plasma membrane. This exchanger is yet another example of the Na'*^-dependent cotransporters discussed above. The resulting accumulation of intracellular Ca^"*" directly increases the force of contraction within the failing heart by its interaction with the contractile proteins. A number of other mechanisms may also contribute to this increase in Ca^"*" within the myocyte, but all seem to be initiated by inhibition of the Na'^-K^ATPase.

Na'-lC'ATPase

253

SUMMARY The Na'^-K'^-dependent ATPase (Na'^-K^-ATPase) is the major pharmacological target for cardiac glycosides such as digitalis and ouabain. It is a membrane-spanning protein complex that mediates the extrusion of Na"^ from the cell and the absorption of K"^ at the expense of metabolic energy. Critical functions served by this Na'^-K^-pump include the generation and maintenance of the electrochemical gradients of Na"^ and K"*" across the plasma membrane, the absorption of sugars and amino acids via cotransport systems, and the transport of salt and water across epithelia. The Na"^-K"^-ATPase consists of two subunits: a catalytic a subunit and a glycosylated P subunit. The former is phosphorylated transiently during the catalytic turnover of the enzyme. The Na'^-K'^-ATPase is maintained under tight control by its hormonal environment, which produces changes in enzyme abundance as well as enzymatic activity.

REFERENCES Dean, R.B. (1941). Theories of electrolyte equilibrium in muscle. Biol. Symp. 3, 331-348. Gloor, S., Antonicek, H., Sweadner, K.J., Pagliusi, S., Frank, R., Moos, M., & Schachner, M. (1990). The adhesion molecule on glia (AMOG) is a homologue of the p subunit of the Na,K-ATPase. J. Cell Biol. 110, 165-174. Ismail-Beigi, F. & Edelman, I.S. (1970). Mechanism of thyroid calorigenesis: Role of active sodium transport. Proc. Natl. Acad. Sci. USA 67, 1071-1078. ShuU, G.E., Greeb, J., & Lingrel, J.B. (1986). Molecular cloning of three distinct forms of the Na'*^,K'^-ATPase a-subunit from rat brain. Biochemistry 25, 8125-8132. Skou, J.C. (1957). The influence of some cations on adenosine triphosphatase from peripheral nerves. Biochim. Biophys. Acta 23, 394-401. Withering, W. (1785). In: An Account of the Foxglove and Some of its Medical Uses: With Practical Remarks on Dropsy and Other Diseases. Swinney, Birmingham.

RECOMMENDED READINGS Edelman, I.S. (1981). Receptors and effectors in hormone action on the kidney. Am. J. Physiol. 241, F333-F339. Jorgensen, P.L. & Andersen, J.P. (1988). Structural basis for E1-E2 conformational transitions in Na,K-pump and Ca-pump proteins. J. Membr. Biol. 103, 95-120. Pressley, T.A. (1988). Ion concentration-dependent regulation of Na,K-pump abundance. J. Membr. Biol. 105, 187-195. Smith, T.W. (1988) Digitalis: Mechanisms of action and clinical use. N. Engl. J. Med. 318, 35&-365. Sweadner, K.J. (1989). Isozymes of the Na"'/K"'-ATPase. Biochim. Biophys. Acta 988, 185-220.

This Page Intentionally Left Blank

Chapter 9

Intracellular Calcium-Binding Proteins KEVIN K.W. WANG

Introduction Properties and Classes of Calcium-Binding Proteins EF-Hand Calcium-Binding Proteins EF-Hand Calcium-Binding Motif EF-Hand Protein Family Calbindin-9K and Calbindin-D28K Calmodulin and Calmodulin-Binding Proteins Troponin C EF-Hand Structure as a Domain of Protein Calcium Binding Proteins in the Nervous System Annexins Calcium Transporting Proteins Other Calcium-Binding Proteins Summary

255 256 257 257 260 260 261 263 265 266 267 268 270 270

INTRODUCTION Calcium is one of the essential elements of eukaryotic organisms. In vertebrates, including man, over 99% of body calcium is immobilized in the bones and teeth by complexing with phosphate to form hydroxy-apatite. The remaining calcium is Principles of Medical Biology, Volume 4 Cell Chemistry and Physiology: Part II, pages 255-274 Copyright © 1996 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-806-4

255

256

KEVIN K.W. WANG

distributed between the extracellular fluid and intracellular space. Extracellular calcium concentration, including that of the blood plasma, is maintained at about 3 mM. This calcium level is controlled mainly by the mobilization of calcium in and out of bone deposits and the intake of dietary calcium. In addition, about 50% of the extracellular calcium exists in an ionized form (Ca^"^) (see Carafoli, 1987). On the other hand, the total intracellular calcium content varies. Erythrocytes contain only 20 JLIM, brain cells 1.5 mM, and heart cells 4 mM. In contrast to the extracellular pool, a very small fraction of the total intracellular calcium is ionized. Typically, the cytosolic-free Ca^"*" concentration lies between 0.1 JLIM and 1 fiM, which is at least three orders of magnitude lower than the extracellular level. This results in a steep electrochemical gradient of Ca^"^ across the plasma membrane of the cell. A cell can be considered an entity in itself, using the plasma membrane as its barrier to the outside environment. Communication between the outside and the inside of the cell is achieved by a so-called "receptor-effector coupling" mechanism. Generally, stimuli, such as hormones and growth factors, can be regarded as ligands which interact selectively with the extracellular binding site of their respective receptor proteins. Upon ligand binding, the receptor molecule is activated and exerts certain effect(s) on its effector(s). For example, phosphatidylinositol-specific phospholipase C mediates the breakdown of plasma membrane phosphatidylinositol 4,5-bisphosphate (PIP2) to generate inositol-triphosphate (IP3) and diacylglycerol (DAG). IP3 can mobilize Ca^"*" from an intracellular calcium store viz. the endoplasmic reticulum (ER) via the IP3 receptor-calcium channel (Berridge, 1987). In the plasma membrane of excitable cells (such as muscle cells or neurons), there are voltage-dependent Ca^"*" channels and ligandgated Ca^"^ channels such as the glutamate-gated NMDA-receptor/Ca^"^ channel (Young and Fagg, 1991). Ligand binding or depolarization will cause brief opening of these Ca^"*" channels, and thus allow Ca^"^ to enter the cell passively, thereby elevating the cytosolic Ca^"*" concentration. An elevated cytosolic Ca^"*", in turn, activates various Ca^'*"-dependent systems. The first step is usually the binding of Ca^"*" to intracellular target proteins (calcium-binding proteins). In the following sections, we will examine different types of calcium-binding proteins and their proposed functions in Ca^"^ signaling.

PROPERTIES AND CLASSES OF CALCIUM-BINDING PROTEINS An intracellular calcium-binding protein (CaBP) can be defined as a cellular protein capable of binding Ca^"^ (usually reversibly) under physiological conditions. Since the resting levels of cytosolic calcium are very low, most often binding of calcium occurs when there is an elevation in the intracellular Ca^"*" level. Generally, a specific Ca^^-binding site on the CaBP can be found to be made up of several Ca^^-chelating atoms (usually oxygens) contributed by either the amino acid side chains or the

Calcium-Binding Proteins

257

peptide backbone. The architecture of such a binding site is amazingly selective for Ca^"*", usually with little affinity for Mg^"**. This is an important criterion since the intracellular free Mg^"^ is about 0.1-1 mM, while the free Ca^"*" levels, even in excited cells, usually do not exceed 1 \xM. Thus, if the calcium-binding sites do not have 1,000-fold selectivity for Ca^"^ over Mg^"*", these sites would be occupied with Mg^"^ most of the time. The EF-hand CaBPs and annexins are the two largest sub-family of CaBP (Crompton et al., 1988; Kretsinger et al., 1991). Proteins involved in the transportation of Cd?^ across cell or organelle membranes also appear to have site(s) with an afifmity for Ca^"^. These include the Ca^^-Mg^'*"-ATPases (calcium pumps), various calcium channels and the Na'^-Ca^'^ exchanger (Carafoli, 1987). There are also many CaBPs that do not readily fit into any categories, such as the Ca^"^- and phospholipiddependent protein kinase (protein kinase C) (Nishizuka, 1984), the sarcoplasmic reticulum luminal calcium-binding protein calsequestrin, and a list of proteins with calcium-binding sites involved in structural stability (McPhalen et al., 1991). The calcium-binding property of a protein can sometimes be confirmed by the "^^Ca overlaying technique (Schachtele and Marme, 1988). Putative calcium-binding protein samples are resolved on a denaturing electrophoresis gel; "^^Ca is overlayed on the gel or the blotting membrane onto which the protein bands from the gel are first electrotransferred. An autoradiogram can then be obtained. Another technique which applies to many EF-hand CaBP is the use of the carbocyanine dye Stain-All which stains a normal protein band pink on an electrophoresis gel. However, if the protein is Ca^'^-saturated, it stains blue or purple (Sharma and Balasubramanian, 1991). Other techniques used in identifying CaBP are terbium phosphorescence, equilibrium- and flow-dialysis, and equilibrium gel filtration (Schachtele and Marme, 1988).

EF-HAND CALCIUM-BINDING PROTEINS EF-Hand Calcium-Binding Motif A helix-loop-helix calcium-binding structure was first recognized in carp CaBP parvalbumin (Kretsinger and Nockolds, 1973). It is known as the EF-hand structure after the E- and F-helices identified in the crystal structure of parvalbumin. It consists of a 12-residue loop that binds calcium ion, flanked by two perpendicular alpha-helices (Figure 1). Typically, the coordination of the ion is provided by precisely spaced residues 1, 3, 5, 7, and 12, and by a water molecule hydrogenbonded to residue 9, thus forming a 6-member octahedral structure (Figure 1). The EF-hand protein family is now the largest superfamily of CaBP. It was believed that these were evolutionarily derived from an ancestor EF-hand gene (Kretsinger, et al., 1991). However, the overall homology between two EF-hand proteins are relatively low (as little as 25%), confirming that the primary structure

258

KEVIN K.W. WANG

Figure 1. EF-hand Ca^"^ binding site. The helix-loop-helix calcium binding site of parvalbumin is indicated by the ball-and-stick model showing the backbone carbon (left). The side chain oxygens that coordinate the Ca^"^ are shown as black spheres. The first helix (E), the calcium-binding loop and the second helix (F) are visualized, respectively, as the index finger, the curved middle finger, and the thumb of a right hand (right). The two helices are usually perpendicular to each other. The octahedral calcium ion coordination is also illustrated with the coordinating oxygen occupying the six vertices. Reproduced with permission of Elsevier Science Publishers from Persechinietal. (1989).

is not critical (Heizmann and Hunziker, 1991). Being conserved is the general structure—two hydrophobic alpha-helices flanking a calcium-binding hydrophilic loop of about 12 residues (Figure 2). The helices are usually about 9—12 residues long. Within the calcium-binding loop, residue 1, 3, 5, 7, and 12 are almost invariably aspartate (D), asparagine (N), glutamate (E), glutamine (Q), serine (S), threonine (T), or tyrosine (Y) which can contribute a side-chain oxygen atom for Ca^"^ ion coordination (Figure 2). Usually, the K^ (i.e., dissociation constant) for Ca^"^ is about 1—10 ^iM. Also, variation in the amino acids in the calcium-binding loop is thought to alter Ca^"*" affinity. A variant from this theme is a family of membrane proteins, the integrins, which mediate cell adhesion to extracellular matrix proteins in a Ca^'*"-Mg^**'-dependent manner. Interestingly, they have an EF-hand-like structure, with the exception that at position 12, the coordinating residue is replaced with an uncharged valine or isoleucine (Tuckwell et al., 1992). It was proposed that the last coordinate for Ca^"^ is provided by an aspartate residue from a natural protein ligand of intregin, such as fibrinogen. In most EF-hand

Calcium-Binding Proteins

259

Loop

uoop N

G Y / I

G \

'

D—jCa2+ K

SEEEIREAPR'V F ^

> V

D

S

G Q / V

G \

'

N

D'-^a2+ Y I / V E jy ^E 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., tvv^o in calbindin-9K, four in calmodulin and troponin C. Tw^o 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 v^ith 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,

Calcium-Binding Proteins

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<»in

Actin filament

ATP>.,^

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

Calcium-Binding Proteins

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,

Calcium-Binding Proteins

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—^abnormal 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

Calcium-Binding Proteins

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-

Calcium-Binding Proteins

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.

Calcium-Binding Proteins

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-

Principles of Medical Biology, Volume 4 Cell Chemistry and Physiology: Part II, pages 275-284 Copyright © 1996 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-806-4

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.

ATP-Ubiquitin-Mediate Protein Degradation

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.

ATP-Ubiquitin-Mediate Protein Degradation

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- ^i^^- 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

Principles of Medical Biology, Volume 4 Cell Chemistry and Physiology: Part II, pages 285-304 Copyright © 1996 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-806-4

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)

Regulation of Cellular Functions

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^^i) effects on PTH secretion (for review see Brown, 1991). Thus, there are numerous pieces of data suggesting some important signaling role for C3^oplasmic 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.

Regulation of Cellular Functions

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

Regulation of Cellular Functions

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-

Regulation of Cellular Functions

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

Regulation of Cellular Functions

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

Regulation of Cellular Functions

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.

Regulation of Cellular Functions

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^^i 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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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

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.

This Page Intentionally Left Blank

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
el

.•-. s

h^ KO UJ X

20.

in

.-_-_-_-_ir-_-^4--4---t--^~i-L^b^

111

333

OS " ^

C

S < -D ?o ^ Q u .- .:D o

C

TO CD

E .

C

o 0) (D

12 — .CD

c 03

E

. . Q.

Q;

2

^ ^

OS QJ

u

OJ CO _ c

CD ± ; Q. < LD

"" U Q

g C

CD o .E .E

o _>>03 ^CD "-^
CD ; ^

Q . 03

3 ^ S "S

CD <

^'

C

2 ! ^ LD

t/) IS.

CD .--. - 4

•E -°°z I c

.ti

-C:

(D ^

C

05

03CD

.^

.E i-- fo

f^ ^

C CD K^ S

re ^


CL

.E "3 • -

O E ±i CD


E.

S^ E Z O ^ g_,c o o ^ 32 C-

a;

CD 03

s S i '^

:s A^ Q- c

C < < 5

334

NIKOLAUS SEILER

AcetyiCoA +

HM^

AcetylCoA: polyamine N^-acetyltransferase (cytosolic) ( c SAT)

Spermine

H2

+

O2 +

H2O Polyamine oxidase

N^-Acetylspermine

0 N CH3C

H2O2

Acetaminopropanal

Serum amine oxidase or Diamine oxidase

Spermine n

H2O,

SpermineKlerived monoaldehyde

Figure 4. Acetylation and oxidative reactions of the polyamines; polyamine oxidase, serum amine oxidase, and diamine oxidase.

N ^ -acetyltransferase; cytosolic; cS AT) (Figure 4) act as key enzymes for regulating polyamine metabolism. The rate-limiting enzymes have the lowest activities among the enzymes involved in polyamine metabolism (Table 1). To make rate-limiting enzymes useful as regulatory proteins, mechanisms are present to rapidly change their activity. This is usually achieved by induction. Physiological (hormones, growth factors) or nonphysiological (tumor promotors, toxic agents) stimuli may lead to the expression of the appropriate gene, which then produces mRNA. This in turn may increase the rate of formation of the enzyme protein (transcriptional regulation). For example, treatment of female rats with testosterone stimulates ODC formation in the kidney 25-fold, with a proportional increase in ODC mRNA. Under these conditions, ODC protein formation amounts to 1% of total protein synthesis (Persson et al., 1984). A second requirement of regulatory proteins is a short half-life (rapid turnover). The three regulatory enzymes of polyamine metabolism have biological half-lives of 20-40 minutes and, hence, belong to the group of proteins with the highest turnover rate. Rapid induction of the rate-limiting biosynthetic decarboxylases accompanied by a high rate of inactivation permits polyamines to be produced in spurts.

CD D u^

hc

E (J2

O JD

(D C

*E c

S. > to

E N C

T3 c 03


>.

m

"E O Q_

*Q

, >«« •5i ^

0) -C^


QJ

i-s

P 0

Ci.

CU

H:

^ 5

QJ fQ

CD

P


X

E | c c

1 f^

CD

^3

cS-^

CD C

«D

§1

i!

< a: ^

2 c

£.2 cu X

c vT »-. .0 0 ^

^QJ ^? *^ fC

E 0 CD

1

c^ C-0 CD C

CD

1

^

CD

s

3 CL

CD

"h^

CD

to

K: 0)

a

!c

•1

CD

00

.^ p:

1 c

"6

o o

00 ro

N.

^

O

^ v£>

O

CN vO

a^ LO

o

o o fN r—

CD

(N

O LO

o

^ o

^

o

*—

o^ o rv. vO

ro

00

o o 0^ CO

hs

<6 o o

o

0^
in

o

'^ 00

in

o o

^

1

1

in in

o in

00

^

00

CN

O^

o o

CN r—

CD

in

O

in

o

^ CN

O

o

CN

m

K

ro CN

T— o o O o i n vD i n +1 +1 +1 +1 +1 rs. o o o r o 00 o ^ -^ rrs.o CO CN

m

0^

o

t\

in

o

o

LO ro

"^

r—

'^

a» c

T—

r—

o o

in vO

1 1

in

ro

in 0^ CN

O ro

O rv

^

o

1

O

o vD r>4

1

1

O 00

o o

CN

^

o CN rg

*""

^ CN

O O CT^

•^ CN

O

00

o

CM 00

rs.

ro

o

C7^

o o

0^ CN CN

fN CD

O r—

o

o in r>s

CO in

o in ro ro

o o

""^

1 1

o

in in

o

CN v£) ro

^

ro CN

O

in 00

00 CO

0^ 00

o

"^ o

r—

fN

O in

ro

O

00

^

rx

o o

00

0^ r—

o

r—

ro in

in

rx o

+1 +1 +1 +1 o o o

O N. in

O

C3^

hs

T

h>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


\— f N

ro

+1 +1 +1 +1 +1 r\ o O o Ns rr • " ^ CN r\ o

CN

r—

03 > LJ

ro a» X

+1 +1 +1 +1 +1 o o CO i n CO r—

c

^

335

u 3

o

00 0^ r—

00

03

Q;

'6

'oj

c

tS -C

>• 0)

c

Q.

1 T5 C

m c

a

1

03 U (U

-o

C

Q.

c: 00

JJ ^ c

fC

CO

•£ c;

c

"E "1

Si -S c re

o c
JO

re *c re -C

1-

"S E O

t^

S ^ 00

c

1 E o 13

'E 3

"S S 2 m cu

^

c

£

Isz

E .2 in

..

"S

cu

re

>«

11
"^ ^

S ^ 3

^

re >> :^ 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,

Polyamines in Cell Biology

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-

Polyamines in Cell Biology

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

<

0.8

cr

<

0.4

A

\ f

\

S

A

^

, \ A.p2M

I

Q.

J-

O

< \ o o o S

\

LJL_L i / i . - r n

\-\

0.6 0.4

h

0.2

4

8 12 TIME (hr)

16

4

8 12 TIME (hr)

Figure 5. Changes of ODC activity and polyamine concentrations during the cell cycle in Chinese hamster ovary cells. The cells were synchronized by selective detachment of mitotic cells. The upper panels show the fraction of cells in each phase of the cell cycle as a function of time after plating of the mitotic cells. (Redrawn from Hebyetal., 1976.)

340

NIKOLAUS SEILER

polyamines is required for establishing conditions for DNA synthesis, while the second surge is presumably needed to induce cytokinesis. From the requirement of increased polyamine concentrations during certain phases of the cell cycle, it appears that rapidly proliferating cells, such as those of the intestinal mucosa, bone marrow, and tumors, would have a high demand for polyamines. Polyamine Deficient Mutants Mutants of mammalian cells lacking ODC or arginase (the enzyme that makes ornithine from arginine) have an absolute requirement for putrescine. This supports the view that putrescine formation during the G1 phase and early S phase is required for DNA synthesis. In contrast to mammalian cells, bacteria require polyamines only for optimum growth; for instance, Escherichia coli strains containing mutations which render them incapable of producing polyamines show no measurable levels of putrescine and spermidine when they are grown in minimal media. But they continue to grow indefinitely at about one-third the growth rate. Bacteria with a growth deficit of this order are unable to compete with the wild type strain and therefore cannot survive in a natural environment. Inhibitors of Polyamine Biosynthesis It is possible to use enzyme inhibitors instead of deficient mutants in studies of the consequences of polyamine deficiency. These include selective inhibitors of ODC, AdoMetDC, polyamine oxidase, and diamine oxidase (Figure 6), among others. With the exception of aminoguanidine, which forms a hydrazone with the cofactor of the oxidase, the compounds shown in Figure 6 are so-called enzymeactivated irreversible inhibitors. They mimick the natural substrates (or products) of the respective enzyme, and are transformed by the target enzyme from chemically inert molecules into reactive intermediates. These intermediates react with a nucleophilic group (e.g., a thiol group of cysteine) within the active site to form covalent bonds, and thus inactivate the enzyme irreversibly. (Other designations for this type of enzyme inhibitor are suicide substrates and mechanism-based inhibitors.) Selective enzyme inactivators are receiving increasing attention in contemporary drug development. Consequences of Polyamine Depletion Proliferating cells, when exposed to the ODC inactivator DFMO, show, initially, a decline in putrescine levels, followed by the disappearance of spermidine (Figure 7). The decline in spermidine is partly due to its conversion to spermine. Figure 8 indicates that inhibition of ODC causes, after one normal cell cycle, a gradual decrease in proliferation rate of cultured cells. The decrease of proliferation rate is

Polyamines in Cell Biology

341

OH FjCH NHj 2-(Difluoromethyl) ornithine

(DFMO)

selective inactivator of OOC

N ' . N ^ .(2,3-Butadienyl)-1,4-butanediamine selective inactivator of polyamine oxidase

H H2N

.NH, NH2

OHOH

5'{[(2)-4-Amino-butenyl]methylamino}-5'-deoxyadenosine selective inactivator of AdoMet DC

Aminoguanidine irreversible inhibitor of diamine oxidase and serum amine oxidase ( pyridoxal phosphate scavenger)

Figure 6. Inhibitors of polyamine metabolism.

the result of the prolongation of the Gl phase of the cell cycle. If spermidine decreases below a critical level, cells stop replicating in spite of the presence of normal spermine concentrations. In almost all cases, the effect of DFMO and other specific inhibitors of polyamine biosynthetic enzymes is cytostatic, rather than cytotoxic. This is demonstrated by the reversibility of growth inhibition. After the addition of putrescine to the culture medium, the cells resume normal growth, even in the presence of the ODC inactivator (Figure 8). Reversible growth inhibition resulting from polyamine deprivation might also take place in vivo. A near-complete blockade of tumor growth in a mouse is shown in Figure 9. In order to achieve growth inhibition in vivo, it is not sufficient to prevent de novo synthesis of the polyamines in tumor cells; also necessary is the prevention of the utilization of gastrointestinal polyamines by the body. (Polyamines in the gastrointestinal tract are formed by microorganisms, and they are also of alimentary origin.) The formation of putrescine by the polyamine oxidase-catalyzed degradation of N^-acetylspermidine (see Figure 4) should also be prevented. Degradation of N^-acetylspermidine allows the mobilization of putrescine in one tissue, its distribution via the circulation, and its utilization by the tumor and other tissues whose demand for polyamines is high (Figure 3).

0

2 TIME

4

6

8

AFTER

10

12

DILUTION

24

26

( HOURS )

Figure 7, Depletion of putrescine and spermidine in rat hepatoma tissue-culture (HTC) cells as a function of time after exposure to 5 mM DFMO. Note: The depletion of putrescine is mainly due to its conversion into spermidine. Spermidine concentration is decreasing in part because of cell proliferation and, in part, because of its transformation into spermine. Total polyamine content is not affected by DFMO. (Redrawn from Rudkin et al., 1984.)

Y

/CONTROL

10" h o

Q:

o

/

A

DFMO + PUTRESCINE

•

DFMO

> -J 3 O

MEDIUM CHANGES

4 6 8 DAYS

i

10

Figures. Effect of 5 mM DFMO on the proliferation rate pf rat hepatoma tissue culture (HTC) cells. Note: Putrecine (10 |LIM) in the medium restores normal growth rate, even in the presence of 5 mM DFMO. (Courtesy of Mamont, unpublished observations.) 342

Polyamines in Cell Biology

343

o—o so

I

D—D PDC+DFMO + MDL 72527 E E 300 z

o

a 200 (n

(0


T/

^

100

Q:

o

'

B^ChHy^

1/

-L. 0 4 7 9 II DAYS AFTER LLC CELL INOCULATION Figure 9. Tumor growth inhibition by polyamine deprivation. C57B1 mice bearing Lewis lung carcinoma in the hind leg were treated with a polyamine deficient chow (PDC) (containing neomycin in order to (partially) decontaminate the gastrointestinal tract). In addition, they received as drinking fluid a solution containing 2 % DFMO and 0 . 0 5 % N\N'^-(2,3-butadienyl)-1,4-butanediamlne, a potent inactivator of polyamine oxidase. The tumors stopped growing completely in the polyamine deprived animals, but tumor growth was immediately resumed when the treatment was stopped and standard chow (SC) was given to the animals. (From Sarhan et al., 1989.)

Development and Differentiation

In general, rapidly proliferating tissues exhibit high ODC activities and high putrescine concentrations. AdoMetDC, being low, limits the rate of polyamine formation. This pattern is especially typical of embryonal tissues. In rat brain, for example, ODC activity is highest at a fetal age of 14-16 days. It declines thereafter gradually, in parallel with putrescine concentration, reaching adult levels at around day 20 of postnatal life, i.e., a time at which rats show adultlike behavior. Comparisons of various regions of the brain reveal that periods of greatest ODC activity correlate with those of maximal neuronal proliferation. Interestingly, AdoMetDC activity shows the opposite pattern. It remains close to detection limits throughout embryonal development and starts increasing only about 10 days post-partum, reaching adult values at around day 20. Since developmental periods are accompanied by typical changes in polyamine metabolism, a role of polyamines in differentiation seems self-evident. However, it often remains difficult to distinguish proliferative processes from developmental

344

NIKOLAUS SEILER

processes. Cell division is considered a prerequisite for development and differentiation. Nevertheless, there is evidence from several experimental systems suggesting that polyamines may play a role in differentiation processes. For example, treatment of cancer patients with DFMO reveals inhibition of differentiation of hematopoietic stem cells to end-stage erythropoietic cells. Long-term inhibition of ODC by DFMO results in a decreased platelet count. DFMO also inhibits the differentiation of Friend erythroleukemia cells. This inhibition is prevented by the addition of spermine to the culture medium. By contrast, mouse neuroblastoma cells and embryonal carcinoma cells isolated from tumors can be induced to differentiate by being exposed to DFMO. This suggests that alterations in the cellular polyamine pattern and growth rate are prerequisites for the occurrence of differentiation. The most convincing evidence of a role of polyamines in differentiation comes from studies with the mammary gland. Prolactin is known to stimulate in a nonlactating female the differentiation of mammary gland cells and the synthesis of milk proteins (a-lactalbumin and casein). Using explants of mouse mammary glands, Oka et al. (1981) showed that putrescine and spermidine concentrations increased during the proliferation of the mammary epithelial cells, and that, in addition to prolactin, spermidine is required to induce milk-protein production. Macromolecular Synthesis Not only rapidly proliferating cells produce polyamines. Mature, nongrowing organs also generate polyamines continuously. In contrast with proliferating tissues, where de novo biosynthesis (and consequently high ODC activity) is required for the newly formed cells, nonproliferating tissues require the maintenance of a steady-state level (polyamine homeostasis). Since putrescine generated in the metabolic cycle by the degradation of N'-acetylspQrmidine (Figure 3) is reutilized, ODC activity can be low. In nongrowing tissues decarboxylation of ornithine is only needed to provide as much putrescine as is irreversibly lost because of elimination from the metabolic cycle. Brain, liver, pancreas, prostate, and other organs have a high capacity for polyamine synthesis, as is reflected by their enzyme patterns indicated in Table 1. In the case of the prostate, the secretion of polyamines with seminal fluid requires a high rate of polyamine synthesis. As for the other tissues, one may assume that the high capacity for producing polyamines is related to their capacity to produce proteins. One approach to examining the fiinction of polyamines in the synthesis of proteins and other macromolecules is to use cell-free systems and compare the in vitro effect with the predicted effect in intact cells. Since it seemed conceptually simple, this approach has generated a great wealth of data on the effects of polyamines on many reactions involved in the synthesis (and degradation) of DNA,

Polyamines in Cell Biology

345

RNA, and proteins. However, it is difficult to assess the physiological significance of the manifold actions of polyamines observed in cell-free systems. Another approach is to compare the effects of gradual polyamine depletion on cell function. The first observable event during gradual depletion of putrescine and spermidine in exponentially growing tumor cells is, in addition to a reduced growth rate, a decrease in the rate of protein synthesis. This is manifested in a decreased rate of incorporation of labeled amino acids into proteins and in changes in the distribution pattern of polysomes, indicating an impairment of the protein-synthesizing machinery. But even complete depletion of spermidine does not decrease protein synthesis much below 50% of its normal rate. Stimulation of translational processes by the polyamines is well supported by experiments. Spermidine, for example, stimulates the binding of fMet-tRNA^^^ the initiator tRNA, to ribosomes. A protein kinase which leads to inhibition of the eukaryotic initiation factor IF2-a is inhibited by spermidine. Another initiation factor, eIF-5A, contains hypusine, an amino acid whose precursor is spermidine. Prolonged depletion of polyamines in cells prevents hypusine formation, and thus prevents the initiation of new peptide chains on the ribosomes. In addition to initiation, there are several other steps in protein synthesis, such as elongation of the peptide chain, or the fidelity of translation of the mRNA-structure into a peptide sequence, that may be influenced by polyamines. Despite the fact that depletion of spermidine in both eukaryotic and prokaryotic cells causes inhibition of DNA synthesis, there is no compelling evidence as yet for or against direct functional interaction of polyamines at the site of replication. Similarly, it is not clear to what extent polyamines are involved in the regulation of RNA synthesis, in spite of detailed studies on the stimulation of RNA polymerases by polyamines. In view of the difficulties of interpreting in vitro data and the multitude of in vitro effects of polyamines on cell-free systems that contain molecules, which polyamines can bind to by electrostatic forces, it is not surprising that the specific mechanisms involved in the functions of polyamines have not yet been pinpointed and elucidated.

COVALENT BINDING OF POLYAMINES TO PROTEINS Transglutaminases are a group of Ca^^-dependent enzymes which are known to catalyze the reaction of a protein-bound glutamine with a series of primary amines, including the free e-amino group of peptide-bound lysine. The formation of s-(y-glutamyl)-lysine cross-links plays a role in normal blood coagulation, postejaculatory coagulation of seminal fluid (in rats), and epidermis keratinization, among others. Polyamines act as physiological substrates of tranglutaminases. In fact, numerous peptides have been identified that contain bound polyamines. Furthermore, there is increasing evidence that transglutaminase-catalyzed reactions with polyamines are among the most important posttranslational modifications involv-

346

NIKOLAUS SEILER

ing a large variety of proteins. Since polyamines have two primary amino groups, they can react once or twice. In the latter case, cross-links are formed. Such interaction in the regulation of growth and differentiation, not only in normal tissues but also in pathological conditions, e.g., in the formation of psoriatic plaques, is likely to involve posttranslational modifications of certain proteins by polyamines.

POLYAMINES IN DIAGNOSIS AND THERAPY When cells die, their contents are released. Thus, polyamines in the circulation, cerebrospinal fluid, and urine may be increased following cell death. Cancer therapy with cytotoxic agents or irradiation is accompanied by an increase in cell death. Polyamine determinations are, therefore, useful as an indicator of the degree of success of chemotherapy. In the case of rapidly growing brain tumors which exhibit excessive cell death, the recurrence of tumor regrowth may be reflected by a raised cerebrospinal or red blood cell polyamine level. The prominent role played by polyamines in cell growth suggests the possibility that inhibitors of polyamine biosynthetic enzymes could be used as potential therapeutic agents against those diseases that involve increased cell proliferation. Hitherto only DFMO (eflomithine) has been found to be of practical importance, but its effects on tumor growth seem relatively moderate. More striking are the effects seen on certain parasites, such as Trypanosoma b. gambiense, the protozoan that causes West-African sleeping sickness, and Pneumocystis carinii, frequently found infecting AIDS patients and causing pneumonia heralded by fever, tachypnea, and ultimately respiratory failure. Why certain protozoans are especially sensitive to eflomithine is not yet clear.

SUMMARY The natural polyamines, putrescine, spermidine, and spermine are formed by energy-dependent reactions. Their cellular levels are intricately regulated by several mechanisms, including transcriptional, translational and posttranslational processes, as well as transport. Because of their positive charge, polyamines interact with negatively charged macromolecules, such as DNA, RNA, and proteins. Most of the known functions of polyamines are attributable to electrostatic binding and the resulting conformational changes of macromolecules. The manifold functions, primarily their effects on growth-related processes, have yet to be fully clarified. Because polyamines are involved in cell growth and differentiation, knowledge of their points of action may lead to the discovery of effective chemotherapeutic agents. New therapeutic agents against bacterial, viral, and parasitic diseases can also be expected.

Polyamines in Cell Biology

347 REFERENCES

Grillo, M.A., Bedino, S., Testore, G., & Pezzali, D.C. (1984). Regulation of spermidine acetyltransferase activity in animal tissues. In: Advances in Polyamines in Biomedical Science (Caldarera, G.M. & Bachrach, U., eds.), pp. 81-88, CLUEB, Bologna. Heby, O., Marton, L.J., Gray, J.W., Lindl, P.A., & Wilson, C.B. (1976). Polyamine metabolism in synchronously growing mammalian cells. Proc. 9. Congress of the Nordic Society for Cell Biology, pp. 155-164, Odense University Press. Marchant, P.E. & Blankenship, J. (1983). N'-Acetylspermidine and N^-acetylspermidine in normal rat tissues and body fluids. Proc. West. Pharmacol. Soc. 26, 65-68. Oka, T., Perry, J.W., Takemoto, T., Sakai, T., Terada N., & Inoue, H. (1981). The multiple regulatory roles of polyamines in the hormonal induction of mammary gland development. Adv. Polyamine Res. 3, 309-320. Persson, L., Seely, J.E., & Pegg, A.E. (1984). Investigation of structure and rate of synthesis of ornithine decarboxylase protein in mouse kidney. Biochemistry 23, 3777-3783. Raina, A., Pajula, R.L., Eloranta, T., & Tuomi, K. (1978). Synthesis of polyamines and S-adenosylmethionine in rat tissues and tumor cells: Effect of D,L-a-hydrazino-5-aminovaleric acid on cell proliferation. Adv. Polyamine Res. 1, 75-82. Rudkin, B.B., Mamont, P.S., & Seller, N. (1984). Decreased protein-synthetic activity is an early consequence of spermidine depletion in rat hepatoma tissue-culture cells. Biochem. J. 217, 731-741. Sarhan, S., Knodgen, B., & Seller, N. (1989). The gastrointestinal tract as polyamine source for tumor growth. Anticancer Res. 9, 215-224. Seller, N., Bolkenius, F.N., Knodgen, B., & Mamont, P. (1980). Polyamine oxidase in rat tissues. Biochem. Biophys. Acta 615, 480-488. Seller, N., Bolkenius, F.N., & Knodgen, B. (1985). The influence of catabolic reactions on polyamine excretion. Biochem. J. 225, 219-226. Wagner, J., Claverie, N., & Danzin, C. (1984). A rapid high-performance liquid chromatographic procedure for the simultaneous determination of methionine, ethionine, S-adenosylmethione, S-adenosylethionine and the natural polyamines in rat tissues. Anal. Biochem. 140, 108-116.

RECOMMENDED READINGS General Cohen, S.S. (1971). Introduction to the Polyamines. Prentice-Hall, Englewood Cliffs. Bachrach, U. & Heimer, Y.M. (1989). The Physiology of Polyamines. CRC Press, Boca Raton.

Metabolism in Vertebrates Seller, N. & Heby, O. (1988). Regulation of cellular polyamines in mammals. Acta Biochem. Biophys. Hung. 23, 1-36. Schuber, F. (1989). Influence of polyamines on membrane functions. Biochem. J. 260, 1-10.

Metabolism in Microorganisms Tabor, C.W. & Tabor, H. (1985). Polyamines in microorganisms. Microbiol. Rev. 49, 81-99.

348

NIKOLAUS SEILER

Interactions with Macromolecules and Macromolecular Synthesis Marton, L.J. & Morris, D.R. (1987). Molecular and cellular functions of the polyamines. In: Inhibition of Polyamine Metabolism (McCann, P.P., Pegg, A.E., &, Sjoerdsma, A., eds.), pp. 79-105, Academic Press, Orlando.

Polyamines and Cytokinesis Oriol-Audit, C. (1985). Actin and polyamines: A further approach to the cytokinesis mechanism. In: Recent Progress in Polyamine Research (Selmeci, L., Brosnan, M.E., & Seiler, N., eds.), pp. 151-160, Akademiai Kiado, Budapest.

Inhibitors and Therapy Pegg, A.E. (1988). Polyamine metabolism and its importance in neoplastic growth and as a target for chemotherapy. Cancer Res. 48, 759-774. McCann, P.P., Pegg, A.E., & Sjoerdsma, A. (eds.) (1987). Inhibition of Polyamine Metabolism. Academic Press, Orlando. Seiler, N. (1991). Pharmacological properties of the natural polyamines and their depletion by biosynthesis inhibitors as a therapeutic approach. In: Progress in Drug Research (Jucker, E., ed.), Vol. 37, pp. 107-159, Birkhauser Verlag, Basel. Janne, J., Alhonen, L., & Leinonen, P. (1991). Polyamines: From molecular biology to clinical applications. Ann. Med. 23, 241-259. Gilad, G.M. & Gilad, V.H. (1992). Polyamines in neurotrauma. Ubiquitous molecules in search of a function. Biochem. Pharmacol. 44,401-407.

Development and Differentiation Heby, O. (1989). Polyamines and cell differentiation. In: The Physiology of Polyamines (Bachrach, U. & Heimer, Y.M., eds.). Vol. 1, pp. 83-94, CRC Press, Boca Raton. Seiler, N. (1982). Polyamines. In: Handbook of Neurochemistry (Lajtha, A., ed.), Vol. 1, pp. 223-255, Plenum Press, New York.

Chapter 14

Free Radicals in Cell Biology PETER A. SOUTHORN and GARTH POWIS

Introduction Chemical Background of Free Radicals Oxygen Free Radicals Free Radicals in Biology Reactivity of Free Radicals in Biologic Systems Protection Mechanisms Against Free Radicals Free Radicals and Disease Defense Against Invading Microorganisms Inflammatory Diseases Hyperoxic Lung Injury Adult Respiratory Distress Syndrome Emphysemia Retrolental Fibroplasia (Retinopathy of Prematurity) Ischemia-Reperfusion Injury After Myocardial Ischemia Ischemia-FLeperfusion Injury in Other Tissues Cerebrovascular Damage Aging Toxic Effects of Foreign Compounds Carcinogenesis Cancer Therapy Summary

Principles of Medical Biology, Volume 4 Cell Chemistry and Physiology: Part II, pages 349-377 Copyright © 1996 by JAI Press Inc. All rights of reproduction in any form reserved. ISBN: 1-55938-806-4

349

350 350 350 352 354 356 357 358 359 360 361 363 363 364 366 367 367 368 369 370 371

350

PETER A. SOUTHORN and GARTH POWIS

INTRODUCTION Life in an oxygen-containing environment results in constant exposure to reactive oxygen-free radicals that can cause irreversible cellular injury (Dormandy, 1969). To prevent such injury, aerobic life forms have evolved various antioxidant systems. Cellular injury and disease result when there is an imbalance between free radical production and the antioxidant systems (Gerschman et al., 1954). This chapter reviews the nature offreeradicals and their effects on biologic systems and examines the possible mechanisms whereby free radicals might be involved in several types of human disease.

CHEMICAL BACKGROUND OF FREE RADICALS Electrons of atoms and molecules occupy spatial volume elements, called orbitals, which are determined by energy and probabilityftinctionsof quantum mechanics. Each orbital can contain a maximum of two electrons. Unlike most chemical species, which have two electrons in each orbital, with these electrons spinning in opposite directions, free radicals are molecules or molecular fragments with an unpaired electron in their outer orbital (Halliwell, 1984; Slater, 1984). This definition of free radicals includes transition metal ions, the hydrogen atom, and a few common inorganic compounds, for example, nitrogen oxide (NO) and nitrogen dioxide (NO2). The presence of an unpaired electron in the outer orbital of a free radical is conventionally represented by the superscript dot, R*. A compound becomes a free radical either by gaining an additional electron (e"), as in equation (1), which shows the reduction of molecular oxygen to the superoxide anion radical (62^), or when it loses an electron, as in equation (2), which shows the oxidation of ascorbic acid (AH2) to the dehydroascorbic acid (A) through afreeradical intermediate (AH*). 02-e"->02 AH2 - e" -> AH' - e" -^ A

(1) (2)

OXYGEN FREE RADICALS Molecular oxygen is a biradical with one unpaired electron in each of its two outer orbitals. The reactivity that molecular oxygen might be expected to have as a biradical is, however, decreased because the unpaired electrons spin in the same direction (Taube, 1965) (Figure 1). This decreases the ability of molecular oxygen to simultaneously remove two electrons from a nonradical molecule that has the normal configuration of pairs of electrons spinning in opposite directions. An electron spin inversion has to occur before such an oxidative reaction can occur, and the laws of quantum mechanics stipulate that such a spin-inversion process is

Free Radicals in Cell Biology

351

Molecular orbitals

. ^ K ^ ^

a,2poa,2p7r

b.2p,r

-f(-

-(-(-

++

a,2sa-

-H•ff -H-

b,2scr a,1s
Figure /. On the left electron configuration of dioxygen in its ground state. Notice the two unpaired electrons in the outer orbitals (a,2p7r) spin in the same direction. The right side shows dioxygen reduction products. Divalent reduction by a pair of electrons from another molecule (A) cannot occur because two electrons with parallel spins would have to occupy same orbit, which is not permitted. Reduction by series of single electron transfers (B) is permitted and produces free radical intermediates.

restricted. This restriction in the oxidizing capabilities of molecular oxygen is removed and oxygen-free radicals are produced when molecular oxygen acquires one electron at a time, because in this situation electron spin inversion need not occur (Figure 2). During oxidative phosphorylation, the mitochondrial cytochrome oxidase enizyme system links ATP production to the controlled tetravalent reduction of the molecular oxygen to water (equation 3). In this process, the partially reduced oxygen-free radical intermediates are tightly bound to the active sites of the enzyme and present no threat to the cell. 02 + 4H"' + 4 e - - > 2 H 2 0

O2

•02-

•H2O2-:

\

-^OH*

(3)

•H2O

H2O

Figure 2. Sequential reduction of molecular oxygen. From Southorn, P.A. (1988). Free radicals in medicine. I. Chemical nature and biologic reactions. Mayo Clin. Proc, 63, 381-289. By permission of Mayo Foundation.

352

PETER A. SOUTHORN and GARTH POWIS Table 1. Cytotoxic Oxygen Species 02

HOJ H2O2

OH*

ROO*

'0,

Superoxide anion radical Hydroperoxyl radical Hydrogen peroxide Hydroxy 1 radical Peroxide radical (R = lipid ) Singlet oxygen

In other situations, sequential univalent reduction of molecular oxygen produces nonbound, partially reduced oxygen-free radical intermediates which, because of their reactivity, can disrupt normal functioning of the cell (Table 1). The superoxide anion radical is produced by the addition of one electron to molecular oxygen. The superoxide anion radical itself undergoes a dismutation reaction in which one superoxide anion radical acts on another to produce hydrogen peroxide (H2O2). Hydrogen peroxide has no unimpaired electron in its outer orbital and is not a free radical, but, because of its oxidizing properties, it can be toxic to the cell. The addition of an electron to hydrogen peroxide forms the highly reactive and cytotoxic hydroxyl radical (OH*). The hydroxyl radical, on acquiring another electron, is converted to water.

FREE RADICALS IN BIOLOGY Changes in the environment affect thefreeradical exposure of living matter (Pry or, 1982). Ionizing radiation damages tissue by causing the heterolytic bond fission of water to hydrogen atoms, hydrated electrons, and hydroxyl radicals. In the presence of oxygen both superoxide anion radical and hydrogen peroxide are also produced. Light of the appropriate wavelengths can cause photolysis of chemical bonds and, thus, generate free radicals (Varma et al., 1984). Atmospheric pollutants such as ozone or the nitrogen oxides (NO and NO2) can react with biologic molecules to form reactive free radicals (Pryor, 1982). Cigarette smoke contains various free radicals, as does other burning organic matter (Pryor et al., 1983). Normally, free radicals produced during oxidation-reduction reactions are the most important cause of a cell'sfreeradical exposure (Del Maestro, 1980; Halliwell and Gutteridge, 1984). Various enzyme systems catalyze the univalent reduction of molecular oxygen to superoxide anion radical. Examples include xanthine oxidase, aldehyde oxidase, dihydroorotic dehydrogenase, flavin dehydrogenases, and peroxidases (McCord and Fridovich, 1968). Univalent reduction of molecular oxygen also occurs in vivo in nonenzymatic electron transfer redox reactions (e.g., hydroquinone + O2 -> semiquinone '02'^ + H"*") and during auto-oxidation reactions, including those involving catecholamines, flavins, and reduced ferridoxins. Another important source of superoxide anion radical is the "univalent leak" of superoxide anion radical from the mitochondrial electron transport system

Free Radicals in Cell Biology

353

(Boveris, 1977). Superoxide anion radical is also produced by the reduced nicotinamide-adenine dinucleotide (NADPH) oxidase system present on the cell surface of inflammatory cells and is important in these cells' ability to kill invading microorganisms (Babior, 1978). A major source of hydrogen peroxide in the body is the dismutation of superoxide anion radical, a reaction catalyzed by superoxide dismutase (SOD) (Fridovich, 1983) (equations 4 + 5). Divalent reduction of molecular oxygen by certain oxidases, found in specialized cell organelles called peroxisomes, produces hydrogen peroxide. Examples of these enzymes are D-amino-acid oxidase and glycolate oxidase. 02 + H^-^H02 HO2 + O2 + H"^ ^ H2O2 + O2

(4) (5)

Net O2 + O2 + 2H^ -> H2O2 + O2 Hydroxyl radicals can also be formed when hydrogen peroxide comes into contact with ferrous or cupric ions (the so-called "Fenton" reaction (equation 6)) (Halliwell and Gutteridge, 1984). The hydroxyl radical may also be formed by an iron-catalyzed Haber-Weiss type of reaction, the net effect of which is an interaction between hydrogen peroxide arid superoxide anion radical in the presence of traces of transition-metal ions to form the hydroxyl radical, hydroxyl ion, and oxygen (equation 7).

COo •->,

Glucose

^NADPH

.

NADP^

Figure 3. Formation of superoxide anion radical by oxidation-reduction cycling of paraquat pyridinyl cation. The following reactions are involved in paraquat toxicity. A. Pentose phosphate pathway provides reducing equivalents in form of reduced nicotinamide-adenine dinucleotide phosphate (NADPH). B. Paraquat pyridinyl cation (PQ"*""^) is reduced to cation radical (PQ"*"). This reaction proceeds continuously, being catalyzed by NADPH cytochrome P-450 reductase. C. Oxidation-reduction cycling of pyridine cation radical reacting with molecular oxygen forms superoxide anion radical (O2 and pyridinyl cation. From Southern, P.A. (1988). Free radicals In medicine. I. Chemical nature and biologic reactions. Mayo Clin. Proc, 63, 381-389. By permission of Mayo Foundation.

354

PETER A. SOUTHORN and GARTH POWIS Fe^"" + H2O2 ^ Fe^-" + OH" + OH" Fe salt

O2 + H2O2-> O2 + OH'+ OH"

(6) /j\

catalyst

Under pathologic conditions or when certain drugs are present, much larger amounts of oxygen free radicals are formed than normal, and they can overwhelm the cell's defenses against free radicals, leading to cell damage and even death of the cell. A typical sequence for the formation of excess amounts of oxygen free radicals is shown by the injurious effects of the herbicide, paraquat (methyl viologen) (Osheroff et al., 1985) (Figure 3).

REACTIVITY OF FREE RADICALS IN BIOLOGIC SYSTEMS Mostfreeradicals of biologic interest tend to be reactive and unstable (Pry or, 1986). As a result, they have an extremely short life-span, exist only at low concentrations (from 10^ to 10"^ M), and do not travel far from their site of formation. For example, the mean effective radius for action of the hydroxyl radical in a cell is about 30 A and it has a half-life of only a few microseconds (Hutchinson, 1957). Free radicals act as both oxidants and reducing agents. When afreeradical reacts with a nonradical compound, another free radical is formed. This enables free radicals to start chain reactions which may be thousands of events long; for example, lipid peroxidation involving polyunsaturated fatty acids (PUFAs). Although the initial free radical produces localized effects, the secondary radicals formed from it and the degradation products produced by reactions involving free radicals can have biologic effects far removed from the site where the first free radical was formed. When two free radicals react with one another, a stable molecule may be formed. This, and the simultaneous consumption of nonradical species, helps explain the eventual termination of chain reactions induced by free radicals. The superoxide anion radical, is not a particularly reactive or even a toxic species by itself Its importance lies in the fact that it can be transformed into the highly dangerous (in a biologic context) hydroxyl radical. Hydrogen peroxide is likewise not especially toxic to cells but it can cross cellular membranes and this is potentially important because the extracellular environment possesses few antioxidant defense mechanisms. In contrast, the reactivity of the hydroxyl radical is such that it will react with whatever biologic molecule is in its vicinity. The hydroxyl radical will damage proteins, cause DNA strand breakage, initiate lipid peroxidation, and, may be thought of as the ultimate damaging species whenever superoxide is formed (Table 2). Peroxidation of the PUFAs in lipid membranes can severely damage the cell membrane, producing loss of fluidity with breakdown of the membrane secretory

Free Radicals in Cell Biology

355

Table 2, Cellular Components Damaged by Hydroxyl Radical Lipids:

Peroxidation of polyunsaturated fatty acids in organelles and plasma membranes. Proteins: Oxidation of sulfhydryl-containing enzymes -> inactivation of enzymes. Carbohydrates: Polysaccharide depolymerization. Nucleic acids: Base hydroxylation, "nicking," cross-linkage, scission of DNA strands (causing mutation and inhibition of protein, nucleotide, and fatty acid synthesis).

functions and transmembrane ionic gradients (Barber and Bernheim, 1967). This auto-oxidation chain reaction (Figure 4) is initiated by the hydroxyl radical or the hydroperoxyl radical, but not by the less reactive superoxide anion radical nor by hydrogen peroxide. The initiating free radical removes a hydrogen atom from one of the methylene carbons of the PUFA carbon chain. This leaves behind an unpaired electron creating a lipid carbon radical. This rapidly undergoes molecular rearrangement to produce a conjugated diene which reacts with molecular oxygen to form a hydroperoxyl radical. This may in turn abstract a hydrogen atom from a methane carbon of an adjacent CH2 group of the PUFA to form another lipid radical and a lipid hydroperoxide. The lipid radical then combines with another molecular oxygen and continues the chain reaction. The lipid hydroperoxide is a stable compound until it comes into contact with transition metal ions when it produces

Initiation

A^V^A/ ^ ^ " PUFA

—

> A/^^/^^A/ ^ ° ^ "

IvyV^sAcooH

(conjugated diene) (R*)

Propagation

V V V ^ A C O O H

+ O,-

Upid hydroperoxyl radical (ROg*)

±

Upid hydroperoxide Termination

2R*—^RR 2 R 0 2 - - * 0 2 + ROOR RGj- + R-—..ROOR

Figure 4. Lipid peroxidation initiated by hydroxyl radical. PUFA = polyunsaturated fatty acids. From Southorn, P.A. (1988). Free radicals in medicine. I. Chemical nature and biologic reactions. Mayo Clin. Proc, 63, 381-389. By permission of Mayo Foundation.

356

PETER A. SOUTHORN and GARTH POWIS

more radicals which in turn further initiate and propagate other chain reactions. The end products of such a lipid peroxidation process include aldehydes, hydrocarbon gases, and various chemical residues, including malondialdehyde. These degradation products can diffuse awayfromthe site of the chain reaction and can give rise to cell edema and influence vascular permeability, inflammation, and chemotaxis. It also has been suggested that these products may alter phospholipase activity, inducing arachidonic acid release with subsequent formation of stable prostaglandins and various endoperoxides (Del Maestro, 1980).

PROTECTION MECHANISMS AGAINST FREE RADICALS By consuming 90% of the available oxygen, the mitochondrial cytochrome oxidase system acts as a sink to remove oxygen that might otherwise be converted into oxygenfreeradicals. The remaining 10% of oxygen could potentially be converted into free radicals. Several enzyme systems fiinction to remove these free oxygen radicals and their products (Figure 5). Cells also contain chemical free-radical scavengers. Superoxide dismutase enzymes catalyze the dismutation of superoxide anion radical (equations 4 and 5) to proceed at an approximately 10"^ faster rate than spontaneous dismutation would occur at physiologic pH (McCord and Fridovich, 1969,1977). It does this by overcoming the electrostatic repulsion of the negatively charged superoxide anions. The enzyme's metal site (SOD-ME^"^) is reduced by one superoxide anion radical and then reoxidized by a second superoxide anion radical (equation 8). SOD-Me^^ + O2 -> SOD Me^ + O2

(8)

SOD Me"^ + O2 + 2 i r -^ SOD-Me^+ H2O2 Cytochrome oxidase system O2

•Oa^

•H2O2

•OH*

1

•H2O

Superoxide dismutase Peroxidases Catalase Figure 5. Enzyme systems to bypass reactive intermediates being released during sequential univalent reduction of molecular oxygen. From Southorn, P.A. (1988). Free radicals in medicine. I. Chemical nature and biologic reactions. Mayo Clin. Proc, 63, 381 - 3 8 9 . By permission of Mayo Foundation.

Free Radicals in Cell Biology

357

The acceleration of this reaction ensures that no superoxide anion is available to react with hydrogen peroxide to form the hydroxyl radical through the metal catalyzed Haber-Weiss type reaction (equation 7). Superoxide dismutases exist in several forms. One form containing manganese is found in the mitochondrial matrix and another containing copper and zinc occurs in the cytoplasm. A copper-containing superoxide dismutase is present in the plasma. Cells are capable of increasing superoxide dismutase synthesis in response to hyperoxidant stress. Two enzyme systems exist to catalyze breakdown of hydrogen peroxide. At low concentrations most hydrogen peroxide is removed by reaction with reduced glutathione (GSH) to form oxidized glutathione (GSSG) and water, a reaction catalyzed by glutathione peroxidase (equation 9). The enzyme glutathione reductase catalyzes the regeneration of reduced glutathione from oxidized glutathione by using NADPH formed by the pentose phosphate pathway. Glutathione peroxidase also catalyzes the reduction of lipid peroxides by glutathione, thus preventing propagation of lipid peroxidation. At high concentrations of hydrogen peroxide, the enzyme catalase becomes important in its removal (equation 10). glutathione peroxidase

2GSH + H2O2 -^ GSSG+2H2O catalase

(9) H 0^

2H202-^02 + 2H20 Tissues also have various nonenzymatic antioxidants for preventing damage by free radicals (Heffner and Repine, 1989). Vitamin E (a series of isomers of tocopherol) is a lipid phase antioxidant which partitions into all membranes and converts superoxide anion, hydroxyl, and lipid peroxyl radicals to less reactive forms (Tappel, 1962). It acts by donating a hydrogen ion to the radical, thereby confining the latter's effect, and is itself converted into a stable vitamin E radical. Beta-carotene, the metabolic precursor of vitamin A, and bilirubin are other fat-soluble antioxidants. Vitamin C is an aqueous phase antioxidant which is widely distributed in intracellular and extracellular fluids (Varma et al., 1984). Other water-soluble compounds may act similarly. For example, cysteine and cysteamine also may be able to quench free radical reactions by donating electronsfix)mtheir sulphydryl groups. Reduced glutathione, uric acid, and glucose also have free radical scavenging properties.

FREE RADICALS AND DISEASE Abnormal free radical production may be involved in several types of human disease. Studies examining this subject are often necessarily indirect and have to rely heavily on analogy with animal models of particular human disease processes. It is usually not possible to detect free radicals in vivo with techniques such as

358

PETER A. SOUTHORN and GARTH POVVIS

electron spin resonance (ESR) spectroscopy used by chemists for this purpose. In order to demonstrate a role forfreeradicals in a particular type of tissue injury, the following criteria should ideally be fulfilled. 1. Abnormal production of free radicals should be demonstrated by ESR spectroscopy, or, failing that, by chemical means, or by finding their reaction products at the site of the lesion. 2. Dissimilar chemicals or mechanisms for producingfreeradicals at the same cellular locus should produce similar lesions. 3. Agents that remove free radicals should protect against the lesion. It will be seen that in the cases discussed these criteria have not always been fulfilled. Even when it can be demonstrated thatfreeradicals are responsible for a particular type of tissue injury in an animal, the appropriateness of extrapolation of the animal model to human disease often remains a problem. Defense Against Invading Microorganisms Free oxygen radicals are used during phagocytosis to destroy invading microorganisms (Babior, 1978; Klebanoff, 1980; Fantone and Ward, 1982). When activated, polymorphonuclear (PMN) leukocytes and macrophages consume large quantities of oxygen, which is transformed into superoxide anion radicals. This process is called the "respiratory burst" and is due to the enzyme reduced NADPH oxidase located on the exterior surface membrane of the cell, including that which lines the phagocytic vacuoles (Figure 6). The superoxide is subsequently converted into hydrogen peroxide and hydroxyl radical. Hydrogen peroxide forms a complex with myeloperoxidase released from the azurophil granules. This complex can oxidize halides to produce, for example, hypochlorous acid, which has a potent bactericidal action. Various stimuli can initiate the respiratory burst, including opsonized bacteria and viruses. Although not solely responsible for the cytotoxic effects of phagocytes, the importance of this oxidase system in human host defense is exemplified by the genetic disorder, chronic granulomatous disease (CGD) (Tauber et al., 1983). The PMN leukocyte from patients with CGD can ingest microorganisms normally but cannot generate sufficient reactive oxygen species to kill catalase-containing microorganisms. In patients with CGD, there are recurrent infections, and abscesses and tissue granulomas develop. Attempts have been made to provide the PMN leukocytes of patients with CGD with an alternative source of hydrogen peroxide by using the hydrogen peroxide-producing enzyme glucose oxidase bound to latex particles or contained in IgG-coated artificial liposomes. These attempts have met with only partial success (Ismail et al. 1979).

Free Radicals in Cell Biology

359

J-*—Bacteria

f^ f ^ ^ NADPH NADPH oxidase \ on plasma

membrane

\

i#iii

Neutrophil

NADP /

VH-**-*^

Hexose monophosphate shunt

Figure 6. Myeloperoxidase (MPO)-hydrogen peroxide-hallde bactericidal activity of neutrophils. NADP"^ and NADPH = oxidized and reduced forms of nicotinamide-adenine dinucleotide phosphate. From Southorn, P.A. and Powis, G. (1988). Free radicals in medicine. II. Involvement in human disease. Mayo Clin. Proc, 63, 390-408. By permission of Mayo Foundation. Inflammatory Diseases

Reactive oxygen species may contribute to the complex pathophysiology of inflammatory diseases such as rheumatoid arthritis (Fantone and Ward, 1982). Free oxygen radicals are released into the extracellular milieu during phagocytosis and also may be generated by other mechanisms active in inflammation, for example, prostaglandin metabolism by prostaglandin hydroperoxidase (Kalyanaraman et al., 1982). Once formed, free radicals could alter the biochemical and biophysical properties of structural proteins in tissues. Experimentally, oxygen free radicals have been show^n to degrade cartilage both directly and by activating PMN leukocyte collagenase (Burkhardt et al., 1986). The regulators of vascular permeability, the actions of interferon and immunoglobulins, and the immunoreactivity of the lymphocytes may also be modified by free radicals (Lunec et al., 1985). This may be responsible for the self-perpetuation of rheumatoid arthritis, with free radicals released from neutrophils altering immunoglobulins which stimulate the respiratory burst of neutrophils. In rheumatoid arthritis the joint cartilage is eroded and the synovium is swollen with an inflammatory infiltrate containing large numbers of neutrophils. Production of synovial fluid, which lubricates the joint, is increased but its viscosity is decreased due to breakdown of the polymer hyaluronic acid which acts as a

360

PETER A. SOUTHORN and GARTH POWIS

lubricant. Human synovial fluid contains small amounts of iron salts (Senator and Muirden, 1968) but no enzyme free radical scavenger systems (Blake et al, 1981), and any superoxide anion radical formed in it would be converted to the hydroxyl radical. Increased levels of products of lipid peroxidation reactions are found both in the synovial fluid and in the plasma of patients with active rheumatoid arthritis (Lunec et al., 1981). The breath of patients with rheumatoid arthritis also contains pentane which originatesfromthe decomposition of lipid peroxides (Humad et al., 1988). There are reports that free radical scavenging agents may ameliorate symptoms in some inflammatory diseases. For example, bovine superoxide dismutase is used as an anti-inflammatory agent in veterinary medicine and there have been preliminary reports of its effectiveness in treating various inflammatory lesions in man, including rheumatoid arthritis (Goebel and Storck, 1983). Hyperoxic Lung Injury Humans can tolerate breathing 100% oxygen at one atmosphere for perhaps 24 to 48 hours without sustaining lung injury, but a more extended period damages the lungs and is lethal if sustained (Deneke and Fanburg, 1980; Jackson, 1985). The dose-injury response curve is steep, with both the severity of lung injury being less marked and the time course of its development more protracted on exposure to lower concentrations of oxygen. Good evidence exists to implicate oxygen free radicals in hyperoxidant lung injury. Hyperoxia increases superoxide anion formation and leakage at two sites in the mitochondrial respiratory chain, the NADH dehydrogenase complex and the ubiquinone-cytochrome b region (Turrens et al., 1982). Increased oxygen concentrations are toxic to cultured lung parenchymal cells (Martin et al., 1981). The lungs of animals breathing pure oxygen contain lipid peroxidation products (Halliwell, 1978) and human volunteers breathing pure oxygen expire pentane, a breakdown product produced by lipid peroxidation (Morita et al., 1986). Inspiring supplemental oxygen increases the lung damage associated with paraquat, bleomycin, and nitrofurantoin—^agents believed to produce lung damage through production of oxygenfreeradicals (Collie, 1980; Martin et al., 1981; Martin et al., 1985). Young, immature animals of some species demonstrate greater tolerance to oxygen than adult counterparts, with this tolerance being associated with an ability to induce production of superoxide dismutase and glutathione peroxidase within 24 hours of exposure to 100% inspired oxygen (Frank et al., 1978). Rats that ordinarily would die from lung damage caused by breathing pure oxygen can survive on preexposure to sublethal (85%) concentrations of oxygen, and such animals had increased levels of superoxide dismutase in their lungs (Crapo and Tiemey, 1974). Oxygenfreeradicals and other toxic products produced by activated neutrophils that accumulate in the lung on breathing pure oxygen probably contribute to the lethal effects of the associated lung injury (Frank and Massaro, 1980). This

Free Radicals in Cell Biology A. Hyperoxia — - • Macrophage Injury

361 • Chemotaxins — • Neutrophil recruited to lung

Neutrophils Lung endothelial ^— Free radicals <— adhere/ cell Injury released activated

B. Hyperoxia — • Lung endothelial cell injury t

• Neutrophils recognize cell injury

Release free ^— Adhere/activated radicals

Figure 7. Two hypotheses to explain recruitment of neutrophils into lungs exposed to hyperoxia and their involvement in resultant lung injury. A. Neutrophils recruited to lungs because of release of chemotaxins by hyperoxia-damaged macrophages. B. Neutrophils recognize and adhere to hyperoxia-damaged lung endothelial cells. From Southorn, P.A. (1988). Free radicals in medicine. II. Involvement in human disease. Mayo Clin. Proc, 63, 390-408. By permission of Mayo Foundation.

neutrophil lung sequestration and subsequent activation may occur secondary to release of chemotaxins, including leukotrienes produced by phospholipid degradation and mediators released by hyperoxia-darriaged macrophages. An alternative hypothesis for neutrophil sequestration in the lung is that these cells adhere to the hyperoxidant-damaged endothelial cells (Figure 7). In part, the longer survival of rats exposed to 85% oxygen as opposed to 100% may be attributable to fev^er neutrophils accumulating in the lung w^ith exposure to the lower concentration of oxygen (Barry and Crapo, 1985). Neutrophil depletion decreases but does not prevent hyperoxia-induced lung injury (Shasby et al., 1982; Raj et al., 1985). If increased free radical formation is the mechanism for oxygen lung toxicity, it might be possible to mitigate this damage v^ith agents that scavenge free radicals. It has been reported that superoxide dismutase and catalase, when encapsulated in liposomes to allow the enzymes to cross cell membranes, afford protection against oxygen toxicity when administered intravenously or down the trachea (Turrens et al., 1984; Padmanabhan et al., 1985). Intratracheal instillation of red blood cells is also reported to confer similar benefits. This action is mediated through the red blood cells' content of reduced glutathione (Van Asbeck et al., 1985). Agents such as endotoxin, which increase intracellular superoxide dismutase and glutathione peroxidase enzyme levels, mitigate hyperoxidant-produced lung injury in animals (Frank and Roberts, 1979). Adult Respiratory Distress Syndrome

The adult respiratory distress syndrome (ARDS) describes a potentially lethal acute diffuse lung injury characterized, in part, by pulmonary edema that results

PETER A. SOUTHORN and GARTH POWIS

362

from damage to the alveolar capillary membrane. The clinical settings associated with ARDS are diverse and include, for example, aspiration, sepsis, bums, microembolism, and hyperbxia. Based on experimental studies, an attractive hypothesis has emerged to explain the mechanism by which these seemingly unrelated conditions cause ARDS. This suggests that ARDS results from the neutrophil recruitment to the lung which commonly occurs in ARDS, followed by lung injury caused by release of toxic neutrophil products, including oxygen free radicals (Tate and Repine, 1983). Some of the most compelling evidence for this hypothesis comes from studies of isolated rabbit lungs perfused with a balanced salt solution containing albumin (Shasby et al., 1982) (Figure 8). Increased vascular permeability occurred when rabbit or human neutrophils plus phorbol myristate acetate (PMA) were added to the perfusate. PMA is a stimulator of oxygen radical formation by neutrophils. Neutrophils by themselves and PMA alone did not cause this pulmonary edema. Neutrophils from patients with CGD, which cannot produce free oxygen radicals, when added to the perfusate with PMA, did not produce pulmonary edema. Adding enzyme-substrate combinations that generate oxygen free radicals to the perfusion medium also produced pulmonary edema, an effect inhibited by catalase and a potent hydroxyl radical scavenger, dimethyl thiourea (DMTU) (Tate etal., 1982). Despite these and other findings from animal studies, a conclusive association between neutrophils and ARDS in patients has not been established, particularly in ARDS from some causes (Glauser and Fairman, 1985). Indeed, ARDS may occur in septic patients with severe neutropenia (Ognibene et al., 1986). It is possible that other mechanisms, including, for example, in sepsis, the direct injurious effects of

TO

U^^^

E S

1 §

1.000 r ' ^^^^^^^H

000

n^^^^^H

C

E E 5

200

Substance added to perfusate Figure 8.

u^^^l

0 ( 1 [

Normal rteutrophils + PMA

CQD

neutrophils + PMA

l^ormal neutrophils

PMA

A l b u m i n in lung lavages of isolated rabbit lungs perfused with an a l b u m i n -

containing solution w h e n exposed to normal neutrophils and phorbol myristate acetate ( P M A ) , neutrophils from patients with chronic granulomatous disease ( C G D ) and PN^A, normal neutrophils, or P M A by itself. Modified from Shasby et a l . , ( 1 9 8 2 ) .

Free Radicals in Cell Biology

363

bacterial endotoxin, initiate pulmonary endothelial damage. Activated neutrophils, if present, compound this injury (Meyrick et al., 1986). Emphysema

Emphysema, a condition in which there is widespread parenchymal lung damage, can occur in young people with inherited homozygous alpha i-protease inhibitor (ai-PI) deficiency, particularly if they smoke cigarettes. More commonly it occurs later in life in those who do not have this enzyme deficiency but who are heavy smokers (Auerbach et al., 1972; Kueppers and Black, 1974). apPI is the major serum antiprotease in humans. When it is deficient, destruction of lung tissue occurs as a result of uninhibited PMN leukocyte lysosomal protease activity (Snider, 1981). Cigarette smoking may unfavorably alter the balance between protease inhibitors and proteases acting on the lung by at least two mechanisms. First, cigarette smoke contains various oxygen and carbon-based free radicals (Church and Pryor, 1985), and these free radicals inhibit apPI activity (Pryor and Dooley, 1985). Secondly, cigarette smoking causes neutrophil sequestration in the lung (Hunninghake et al., 1980) and activates these neutrophils to release both elastase and more oxidants, which further inhibit apPI (Zaslow et al., 1983). It has been shown in vitro that the antioxidants glutathione and ascorbic acid prevent the loss of elastase inhibitory capacity of a i-PI caused by cigarette smoke (Pryor et al., 1986). In view of the finding that many cigarette smokers have an ascorbic acid deficiency, vitamin C supplements might, in part, protect cigarette smokers from emphysema. A more important health benefit would result from stopping smoking. Retroiental Fibroplasia (Retinopathy of Prematurity)

Advances in neonatal medicine have markedly improved the chances of survival of low birth weight infants. However, they have been associated with a resurgence of retroiental fibroplasia in which, after birth, retinal vessels proliferate abnormally causing retinal detachment, scarring, and blindness (Lucey and Dangman, 1984; Robin and Burke, 1986). The causes of this condition are not fully understood but include hyperoxia. Before the role of hyperoxia was recognized, the practice in the era 1940 to 1950 of giving virtually all newborn infants supplemental oxygen produced an estimated 10,000 cases of blindness worldwide. With oxygen therapy, both the absolute tension of gas used and the duration of its administration are important. The part played by free oxygen radicals remains unclear at this time. Increased production of the arachidonic acid metabolite thromboxane A2 induced by free oxygen radicals may be responsible for the sustained vascular contraction found early in this condition, and lipid peroxidation could contribute to cellular injury (Tripathi and Tripathi, 1984). Experiments in tissue culture (Tripathi and

364

PETER A. SOUTHORN and GARTH POWIS

Tripathi, 1984) and animal studies (Phelps and Rosenbaum, 1979) suggested that vitamin E can decrease but not eliminate oxygen-induced retinal changes. Studies in infants given vitamin E produced conflicting results, but some have suggested that its administration decreases both the incidence and the severity of retinopathy of prematurity (Hittner et al., 1981). Ischemia-Reperfusion Injury After Myocardial Ischemia

Myocardial ischemia and infarction associated with coronary artery occlusion is a major cause of morbidity in our society. Major recent advances in managing acute coronary occlusion have included developing balloon angioplasty and thrombolytic therapy which can reestablish blood flow to the hypoxic, but still potentially viable, heart tissue. In this clinical setting, evidence suggests that the extent of injury that the heart sustains after a partially reversible ischemic insult is, in part, related to damage caused by free radicals, particularly those formed at the time of reperfiision (McCord and Roy, 1982). Thefirstevidence for the existence of reperfiision injury, or at least a variant of it, was obtained by investigators who described a phenomenon they referred to as the "oxygen paradox" (Hearse et al., 1973). They measured intracellular enzyme release, an index of cell injury,fromisolated potassium-arrested rat heart perfiised with a glucose-free solution during a period of hypoxia and then reoxygenation. On reoxygenation after 100 minutes of hypoxia, there was a marked increase in enzyme release (Figure 9). Shorter periods of hypoxia were not associated with this

1

2.050 ®

O Q.

1

1

2.000 1.950

Extended hypoxia

\ 1 1 1 1

^

300

^^

Reoxygenation after 20 min hypoxia - ^ Reoxygenation after 100 min hypoxia

250 200

3

150

Q. O

100 50 0

()20

'

50

12!0

200

I

I

300

Perfusion time, min

I

400

500

Figure 9. Creatinine phosphokinase (CPK) enzyme release Into perfusate from Isolated perfused rat heart with extended hypoxia and reoxygenation after 20 minutes hypoxia and after 100 minutes hypoxia. From Hearst et a!., (1973). By permission of Academic Press.

Free Radicals in Cell Biology

365

injury on reoxygenation, suggesting that it involved tissue-dependent biochemical and cellular changes during the hypoxic period. That free radicals might be involved in this phenomenon was later suggested by the finding that adding vitamin E to the perfusate decreased the extent of the enzyme release (Guarnieri et al, 1982). This has been confirmed more recently by direct identification of free radical production at reperflision by ESR spectroscopy (Zweier et al., 1985). There are several potential mechanisms whereby free radicals could be produced during ischemia and at reperfusion, and it is possible that these mechanisms could act simultaneously. Disruption of the mitochondrial electron transport chain, autooxidation of catecholamines and other intracellular metabolites or components, and arachidonic metabolism through cyclooxygenase and lipooxygenase pathways are potential sources (Rao and Mueller, 1983). Free radicals could also be released by activated neutrophils in the ischemic area. With reperfusion, neutrophils infiltrate previously ischemic tissues and it has been demonstrated that prior neutrophil depletion in myocardial ischemia reperfusion models in the dog decreases the size of the ultimate myocardial infarct (Romson et al, 1983). A burst of free radical production at reperfusion may also result from a sequence of biochemical events initiated during the ischemic period (McCord and Roy, 1982) (Figure 10). The endothelial cells of many tissues, including the heart of some species, contain varying amounts of the enzyme xanthine dehydrogenase. It is postulated that the increased cytoplasmic free calcium concentration associated with ischemia activates an intracellular protease that converts this xanthine dehydrogenase into xanthine oxidase. Whereas the dehydrogenase reduces nicotinamide adenine dinucleotide (NAD"^) and cannot transfer electrons to molecular oxygen, the oxidase enzyme does possess this capability, transferring electrons from hypoxanthine formed during ischemia by ATP catabolism to molecular oxygen supplied at reperfusion to form superoxide anion radical or hydrogen peroxide (McCord and Fridovich, 1968). In dogs, after 30 minutes of coronary artery litigation the concentration of xanthine oxidase in the resulting ischemic myocardium increases 300% (Chambers et al., 1985). Prevention of ischemic reperfusion injury is currently attracting much study. If clinically applicable, this work could not only benefit patients with acute coronary artery occlusion but also possibly mitigate the harmful effects of global myocardial ischemia, such as those associated with coronary artery bypass grafting and cardiac transplantation and after successful cardiopulmonary resuscitation. Data from experimental preparations of regional and global myocardial ischemia and reperfusion have documented that superoxide dismutase alone (Ambrosio et al., 1986) or in combination with catalase (Shlafer et al., 1982; Jolly et al., 1984), has a beneficial effect on myocardial cell viability. AUopurinol, a xanthine oxidase inhibitor, may also decrease the severity of ischemic reperfusion injury (Chambers, 1985). AUopurinol may also be acting as a free radical scavenger to produce this beneficial effect. It has been documented that otherfi^eeradical scavengers, such

366

PETER A. SOUTHORN and GARTH POWIS AMP^^—^^^ iophpmifl iscnemia

Adenosine •, Inosine Hypoxanthine

Xanthine

dehydrogenase | Protease

Reperfusion

H202<< I Iron salts HO* • H 0 + 0 2 Figure 10. Possible mechanism for formation of oxygen free radicals during ischemic reperfusion. During ischemia, adenosine triphosphate (ATP) is catabolized to the substrate hypoxanthine, and the enzyme xanthine dehydrogenase is converted to xanthine oxidase by a protease activated by increased free calcium. O n reperfusion, hypoxanthine reacts with molecular oxygen in presence of xanthine oxidase to form superoxide anion radicals (O2. In the presence of iron salts, superoxide anion radicals can form hydroxyl radicals. From Southorn, P.A. and Powis, G. (1988). Free^radicals in medicine. II. Involvement in human disease. Mayo Clin. Proc, 63, 390-408. By permission of Mayo Foundation.

as dimethylsulfoxide and mannitol, also decrease ischemic reperfusion injury (Ganote et al., 1982; Stewart, 1983). These encouraging results in animal studies suggest that clinical trials are warranted. Ischemia-Reperfusion Injury in Other Tissues

Ischemia-reperfusion tissue injury can occur in several tissues in addition to the heart (Parks et al., 1983). These include the small intestine, gastric mucosa, kidney, liver, and skin. Research in this field has found clinical application in preserving organs used in transplantation. These organs are often subject to extended periods of hypoxia and when implanted in the recipient and reperfiised invariably show at least a temporary loss of function. That this is due in part to oxygen free radicalmediated damage is supported by animal studies. For example, in dogs, allopurinol improves post-transplantation renal function and graft survival (Toledo-Pereyra et al., 1974). These results and those of other animal studies have been applied to developing preservation fluids used to store human transplant organs between their removal and placement in the recipient. The University of Wisconsin preservation fluid introduced recently permits extended hypothermic in vitro preservation of the

Free Radicals in Cell Biology

367

kidney, pancreas, and liver (Jamieson et al., 1988). Included in its ingredients are allopurinol and large quantities of reduced glutathione, the substrate in the chemical reaction catalyzed by glutathione peroxidase in which hydrogen peroxide is reduced to water. Part of the damage associated with hemorrhagic shock has been suggested to be mediated by oxygen free radicals (Parks et al., 1983). This is based on the view that the period of shock represents whole body ischemia and the restoration of adequate tissue perfusion with resuscitation could produce conditions for ischemic reperfusion injury. Evidence for this is that allopurinol improves the survival rate of dogs subjected to hemorrhagic shock (Crowell et al., 1969). Superoxide dismutase also eliminates some of the adverse consequences of hypovolemic shock in dogs, although it does not affect ultimate survival (Schoenberg et al., 1983). Cerebrovascular Damage Vascular injury with subsequent edema and abnormalities in regulation of cerebral blood flow may contribute to secondary neuronal dysfunction in several pathologic conditions of the brain. Recent studies, mostly in animals, suggest that cerebrovascular injury found in acute hypertension, fluid-percussion brain injury, and ischemia may in part be due to formation of oxygen free radicals (Hammond et al., 1985). Acute severe hypertension induced by pharmacologic agents or by fluid-percussion brain injury in animals produces morphologic abnormalities in the endothelium and the vascular smooth muscle (Kontos et al., 1981). These changes are associated with sustained cerebral arteriolar dilation, increased permeability of the blood-brain barrier, and cerebral edema. Pretreatment with superoxide dismutase and catalase or radical scavengers such as nitroblue tetrazolium or mannitol can prevent these morphologic and associated functional changes. Xanthine oxidase and xanthine, a system generating superoxide anion radicals and hydrogen peroxide, when topically applied to the brain surface, produce pathophysiologic abnormalities similar to those seen after hypertension or brain injury (Kontos, 1985). The results have been taken to indicate that superoxide anion radical is the primary free radical species involved in vascular brain injury. Superoxide anion radical formation possibly results from accelerated arachidonic acid metabolism in endothelial cells in the vessel wall. This may present an opportunity to prevent some deleterious effects of cerebral vascular injury by using inhibitors of arachidonate metabolism, such as cyclooxygenase inhibitors like indomethacin, perhaps in combination with scavengers of free radicals. Aging The universality of aging implies that its cause is basically the same in all species. A free radical hypothesis of aging has been proposed that suggests that the

368

PETER A. SOUTHORN and GARTH POWIS

free radicals produced during normal metabolism of the cell over time damage DNA and other macromolecules leading to degenerative diseases, malignancies, and eventual death of vital cells which in turn causes aging and death of the animal (Harman, 1956). This hypothesis is supported by several observations. Radiation produces its biologic effect through free radical formation and causes premature aging (Hempelmann and Hoffman, 1953). Second, the age pigment lipofuscin, which accumulates in all mammalian species with age, is associated with lipid peroxidation (Harman, 1981). The hypothesis that oxygen radicals play a role in aging is also supported by the observation that, in general, long-lived species produce less endogenous free oxygen radicals because of their lower metabolic rate (Sohal, 1981). Long-lived animals also have more superoxide dismutase than do their short-lived counterparts, and animal species with the longest lifespans have the highest levels of superoxide dismutase (Tolmasoff et al., 1980). Oxidative DNA damage is rapidly and effectively repaired. It has been estimated that there are several thousand oxidative DNA damage sites in a human cell every day, and the majority of these are repaired (Ames, 1986). However, a small fraction of unrepaired lesions could cause permanent changes in DNA and might be a major contributor to aging and cancer. A consequence of the free radical hypothesis of aging is the idea that free radical scavenging agents might be used to prevent aging. A number of antioxidants have been tested in animals, including vitamin E (Blackest and Hall, 1981) and butylated hydroxytoluene (Clapp et al., 1979), with equivocal results. Feeding antioxidants to animals is associated in many cases with weight loss which itself can lead to an increased life expectancy in animals (Schneider and Reed, 1985). Thus, interpretation of the results of these studies is difficult. Interestingly, a study of a selfselected group of high dose, human vitamin E users age 65 years or more found an increased mortality associated with consumption of more than 1,000 lU vitamin E a day (Enstrom and Pauling, 1982). This may reflect an attraction felt by less healthy persons to take high doses of vitamin E. A clear association between life expectancy, lifespan, and dietary antioxidant intake in humans remains to be demonstrated. Toxic Effects of Foreign Compounds The toxicity of some foreign compounds and drugs emanates from their capacity to induce free radical-mediated reactions. Two examples will be given. Carbon tetrachloride (CCI4) is used in industry as a solvent for degreasing and, until fairly recently, for dry cleaning clothes. CCI4 can be absorbed through the skin and produces hepatotoxicity in virtually all species, including man. Hepatotoxicity can vary from triglyceride accumulation through necrosis to cirrhosis and cancer, depending on the dose and species studied. Evidence suggests the hepatotoxicity of CCI4 involves its reductive dehalogenation by cytochrome P-450 of the hepatic

Free Radicals in Cell Biology

369

microsomal mixed function oxygenase system to form the trichloromethyl radical (CCI3O (Rosen and Rauckman, 1982; Brattin et al., 1985). This is probably the species that is involved in covalent binding to cellular macromolecules. The trichloromethyl radical can react with molecular oxygen to form the trichloromethylperoxy radical (CC1302-). This radical will, among other things, react with PUFAs to initiate lipid peroxidation. Paraquat (methyl viologen) is a herbicide which, when ingested accidently or taken in cases of suicide by poisoning, causes severe pulmonary toxicity. Paraquat is accumulated by the lung, which accounts for its selective toxicity for this organ, but it can also cause liver and kidney damage. There is congestion and edema of the lung with hyaline membrane formation and inflammatory infiltrates. Later changes in the lung include fibrosis and the end result is almost always death from respiratory failure. Concern has been expressed over the use of paraquat spraying to control the illicit production of marijuana. Residues of paraquat are found in marijuana cigarettes and it has been suggested that persons inhaling smoke fi*om these cigarettes might develop lung injury (Smith, 1978). The mechanisms of paraquat poisoning in man and other animals is generally thought to be mediated through superoxide formed by redox cycling of the paraquat pyridinyl cation (Osheroff et al., 1985) (Figure 3). The heavy fluxes of superoxide anion radical and hydrogen peroxide, in the presence of chelated iron, form hydroxyl radicals and the normal physiologic defenses are overwhelmed. Although the paraquat pyridinyl cation could react with cellular targets, current evidence indicates that the hydroxyl radical is the species responsible for lung toxicity, probably by initiating lipid peroxidation. Experimental intervention with radical scavenging agents to inhibit lipid peroxidation and deferoxamine to chelate iron has produced equivocal or disappointing results. Carcinogenesis Carcinogenesis is thought to involve a two-stage process (Pitot, 1981). In the initiation stage of carcinogenesis a physical, chemical, or biological agent directly causes an irreversible alteration in the molecular structure of DNA of the cell. This is followed by a promotion stage in which there is an alteration in the expression of the genes that regulate cell differentiation and growth. Unlike initiating agents, promoting agents do not usually react directly with the genetic material but alter gene expression by indirect mechanisms. Some agents are complete carcinogens, that is, they possess both initiating and promoting activity, and by themselves can cause neoplasms. Complete carcinogens include polycyclic hydrocarbons, oncogenic RNA and DNA viruses, and ionizing and ultraviolet irradiation. Oxygen free radicals play a role mostly in the promotion phase of carcinogenesis (Cerutti, 1985). Hyperbaric oxygen, superoxide radical, and certain organic peroxides are tumor promoters, but they may also be weak complete carcinogens (Heston

370

PETER A. SOUTHORN and GARTH POWIS

and Pratt, 1959; Sanford et al., 1979; Zimmerman and Cerutti, 1984). In contrast, many antioxidants are antipromoters and anticarcinogens (Weitzman and Stossel, 1982). The chromosome-breaking (clastogenic) effect of the tumor promoter PM A is mediated by the production of oxygen free radicals that either directly damage DNA or activate a component in the cell or the surrounding medium to form a clastogenic factor (Emerit and Cerutti, 1981). A clastogenic factor, whose activity can be suppressed by superoxide dismutase, has been identified in the serum of patients with autoimmune disease (Michelson, 1982). A scenario for an epigenetic mechanism that may result in altered gene expression by oxygen free radicals has recently been proposed (Michelson, 1982). The enzyme poly(ADP ribose) synthetase is known to be stimulated by DNA breaks, such as may be caused by oxygen free radicals, while oxidative stress produced by oxygen free radicals produces an increased amount of NAD"^ which acts as a substrate for the enzyme. The combination of increased enzyme activity and increased supply of NAD"^ results in increased poly ADP-ribosylation of chromosomal material and, consequently, the modulation of gene expression. The carcinogenic activity of polycyclic hydrocarbons has been correlated with their ability to form free radicals (Szent-Gyorgyi et al., 1960). At least four different radicals can be formed from benzo(a)pyrene (BaP), including the benzo(a)pyrene anion radical (BaP*), the benzo(a)pyrene cation radical (BAP*), an unidentified radical species formed by heating BaP (which may be related to radicals formed in cigarette smoke), and oxygenated BaP radicals. The most prominent of the last are the 6-oxyBaP radical and radicals formed from BaP quinones (Sullivan, 1985). BaP* and oxygenated BaP radicals can be formed by the action of cytochrome P-450 or peroxidase-hydrogen peroxide and have been implicated in the binding of BaP to DNA as the first step leading to tumor formation (Cavalieri and Rogan, 1985; O'Brien, 1985). The discovery that phagocyte-derived oxygen radicals product mutation in bacteria (Michelson, 1982) and cytogenetic changes in cultured mammalian cells (Weitberg et al., 1983) provides evidence to support a role for phagocytes and inflammation in the pathogenesis of cancer. Chronic inflammation, it is well known, can be associated with cancer (Templeton, 1975). It has been proposed that activation of carcinogens by leukocytes during the respiratory burst may contribute to the increased incidence of cancer in these conditions. Cancer Therapy The main emphasis in treating human disease caused by free radicals is to decrease the formation of free radicals or to limit their reaction with critical sites within the body. However, when the desired therapeutic effects is to damage or kill certain cells, increasing free radical formation might be considered a usefiil thera-

Free Radicals in Cell Biology

371

peutic strategy, if selectivity for the desired site can be attained. This has been the approach with certain treatments for cancer. The cell-killing effect of ionizing radiation is due primarily to the various ions and short-lived and reactive free radicals formed from the interaction of ionizing radiation and water (Hall, 1973). The two major free radicals resulting from the radiolysis of water are the hydroxyl radical and the aquated electron, with lesser amounts of hydrogen atom. There is considerable evidence that the biologic target for radiation is DNA (Hall, 1973). The hydroxyl radical reacts four times faster than the hydrogen atom or the aquated electron with DNA components, and both the base and sugar moieties are affected. The three most important types of DNA lesion are single-strand breaks, double-strand breaks, and base damage. Well-oxygenated cells are two to three times more sensitive to radiation than hypoxic cells. Unfortunately, tumor cells are often hypoxic. Two mechanisms, both involving free radical processes, have been proposed to explain the oxygen effect on radiation: 1) the formation of damaging species such as hydroperoxy radicals, and 2) the reaction of oxygen with free radicals formed in target biomolecules to produce organic peroxy radicals. Approaches that are used to overcome the fact that tumor cells may have low oxygen concentrations and consequently be relatively resistant to radiation are to fractionate the total dose of radiation, causing the tumor to shrink and allowing hypoxic cells to reoxygenate; to have patients breathe pure oxygen at a pressure of 2 to 3 atmospheres; or to use hypoxic cell radiosensitizers, such as metronidazole or misonidazole, which appear to act in a comparable way to oxygen to "fix" the radical. A chemotherapeutic counterpart to the use of free radicals from ionizing radiation to kill cancer cells may be found in the use of drugs such as doxorubicin and bleomycin. A surprisingly large number of anticancer drugs in use today can form free radicals. This, considered with the observation that tumor cells may be deficient in enzymes such as superoxide dismutase and catalase that normally protect cells from free radical damage (Oberley and Buettner, 1979), has led to suggestions that free radicals might be involved in the antitumor activity of some of these drugs (Powis, 1987). The best evidence for involvement of a free radical in the cytotoxic effect of an anticancer drug is from the glycopeptide antibiotic bleomycin (Lazo et al., 1987). Bleomycin binds to DNA and, in the presence of ferrous iron and oxygen, cleaves DNA. Afreeradical species, possibly the hydroxyl radical or a related species, is formed in close proximity to the DNA and results in its degradation. Radical scavenging agents are generally unimportant for inhibiting DNA cleavage by bleomycin, suggesting that the activated oxygen intermediate is unlikely to diffuse far from its site of formation (Cunningham et al., 1983).

SUMMARY Free radicals are reactive chemical species that differ from other compounds in that they have unpaired electrons in their outer orbitals. They are capable of damaging

372

PETER A. SOUTHORN and GARTH POWIS

cellular components, and accumulating evidence suggests that they may contribute to various disease entities. Biologic systems are exposed to free radicals that have been formed endogenously or that resultfromexternal influences, such as ionizing radiation. Oxygen free radicals are continuously being produced intracellularly by oxidation-reduction reactions. The sequential univalent reduction of molecular oxygen initially forms the superoxide anion radical, which in turn is converted, in the presence of transition metal ions, into the highly reactive hydroxyl radical. All aerobic organisms possess elaborate antioxidant systems that prevent free radicalmediated injury. These include vitamin E and the enzymes superoxide dismutase and glutathione peroxidase. Tissue damage may occur when these host defense mechanisms are abnormally deficient or when there is excess free oxygen radical production. The last two decades have witnessed an increasing body of knowledge implicating free radical-mediated processes in a wide spectrum of different types of human disease. Even when this evidence suggests thatfreeradicals are not totally responsible for a particular disease entity, they often appear to play a part in the disease process. Confirmation of free radical involvement in a particular disease may have clinical relevance. Clinically applicable techniques are currently being developed to removefreeradicalsfromcellular sites where they are injurious and, in other situations such as chemotherapy, techniques or drugs that produce free radicals are available to destroy harmful cells. REFERENCES Ambrosio, G., Becker, L.C., Hutchins, G.M., Weisman, H.F., & Weisfeldt, M.L. (1986). Reduction in experimental infarct size by recombinant human superoxide dismutase: Insights into the pathophysiology of reperftision injury. Circulation 74,1424-1433. Ames, B.N. (1986). Food constituents as a source of mutagen, carcinogens, and anticarcinogens. Prog. Clin. Biol. Res. 206,3-32. Auerbach, O., Hammond, E.C., Garfmkel, L., & Benante, C. (1972). Relation of smoking and age to emphysema: Whole-lung section study. N. Engl. J. Med. 286, 853-857. Babior, B.M. (1978). Oxygen-dependent microbial killings by phagocytes [in two parts]. N. Engl. J. Med. 298, 659-668; 721-725. Barber, A. A. & Bemheim, F. (1967). Lipid peroxidation: Its measurement, occurrence, and significance in animal tissues. Advances Geront. Res. 2, 355-403. Barry, B.E. & Crapo, J.D. (1985). Patterns of accumulation of platelets and neutrophils in rat lungs during exposure to 100% and 85% oxygen. Am. Rev. Respir. Dis. 132, 548-555. Blackett, A.D. & Hall, D.A. (1981). Vitamin E—its significance in mouse ageing. Age Ageing 10, 191-195. Blake, D.R., Hall, N.D., Treby, D.A., Halliwell, B., & Gutteridge, J.M.C. (1981). Protection against superoxide and hydrogen peroxide in synovial fluid from rheumatoid patients. Clin. Sci. 61, 483-486. Boveris, A. (1977). Mitochondrial production of superoxide radical and hydrogen peroxide. In: Tissue Hypoxia and Ischemia (Advances in Experimental Medicine and Biology) (Reivich, M., Corbum, R., Lahiri, S., & Chance, B., eds.). Vol. 78, pp. 67-82. Plenum Press, New York. Brattin, W.J., Glende, E.A., Jr., & Recknagel, R.O. (1985). Pathological mechanisms in carbon tetrachloride hepatoxicity. J. Free Radic. Biol. Med. 1, 27-38.

Free Radicals in Cell Biology

373

Burkhardt, H., Schwingel, M., Menninger, H., Macartney, H.W., & Tschesche, H. (1986). Oxygen radicals as effectors of cartilage destruction. Arthritis Rheum. 29, 379-387. Cavalieri, E. & Rogan, E. (1985). Role of radical cations in aromatic hydrocarbon carcinogenesis. Environ. Health Perspect. 64, 69-84. Cerutti, P. A. (1985). Prooxidant states and tumor promotion. Science 227, 375-381. Chambers, D.E., Parks, D.A., Patterson, G., Roy, R., McCord, J.M., Yoshida, S., Parmley, L.F., & Downey, J.M. (1985). Xanthine oxidase as a source of free radical damage in myocardial ischemia. J. Mol. Cell. Cardiol. 17, 145-152. Church, D.F. & Pryor, W.A. (1985). Free-radical chemistry of cigarette smoke and its toxicological implications. Environ. Health Perspect. 64, 111-126. Clapp, N.K., Satterfield, L.C., & Bowles, N.D. (1979). Effects of the antioxidant butylated hydroxytoluene (BHT) on mortality in BALB/c mice. J. Gerontol. 34, 497-501. Collis, C.H. (1980). Lung damage from cytotoxic drugs. Cancer Chemother. Pharmacol. 4, 17-27. Crapo, J.D. & Tiemey, D.F. (1974). Superoxide dismutase and pulmonary oxygen toxicity. Am. J. Physiol. 226, 1401-1407. Crowell, J.W., Jones, C.E., & Smith, E.E. (1969). Effect of allopurinol on hemorrhagic shock. Am. J. Physiol. 216, 744^748. Cunningham, M.L., Ringrose, P.S., & Lokesh, B.R. (1983). Bleomycin cytotoxicity is prevented by superoxide dismutase in vitro. Cancer Lett. 21,149-153. Del Maestro, R.F. (1980). An approach to free radicals in medicine and biology. Acta Physiol. Scan Suppl. 492, 153-168. Deneke, S.M. & Fanburg, B.L. (1980). Normobaric oxygen toxicity of the lung. N. Engl. J. Med. 303, 76-^6. Dormandy, T.L. (1969). Biological rancidification. Lancet 2, 684-688. Emerit, L & Cerutti, P.A. (1981). Tumor promoter phorobol-12-myristale-13-acetate induces chromosomal damage via indirect action. Nature 293, 144-146. Enstrom, J.E. & Pauling, L. (1982). Morality among health-conscious elderly Califomians. Proc. Natl. Acad. Sci. USA 79, 6023-6027. Fantone, J.C. & Ward, P. A. (1982). Role of oxygen-derived free radicals and metabolites in leukocytedependent inflammatory reactions. Am. J. Pathol. 107, 397-418. Frank, L., Bucher, J.R., & Roberts, R.J. (1978). Oxygen toxicity in neonatal and adult animals of various species. J. Appl. Physiol. 45, 699-704. Frank, L. & Massaro, D. (1980). Oxygen toxicity. Am. J. Med. 69, 117-126. Frank, L. & Roberts, R.J. (1979). Oxygen toxicity: Protection of the lung by bacterial lipopolysaccharide (endotoxin). Toxicol. Appl. Pharmacol. 50, 371-380. Fridovich, L (1983). Superoxide radical: An endogenous toxicant. Annu. Rev. Pharmacol. Toxicol. 23, 239-257. Ganote, C.E., Sims, M., & Safavi, S. (1982). Effects of dimethylsulfoxide (DMSO) on the oxygen paradox in perfused rat hearts. Am. J. Pathol. 109, 270-276. Gerschman, R., Gilbert, D.L., Nye, S.W., Dwyer, P., & Fenn, W.O. (1954). Oxygen poisoning and x-irradiation: A mechanism in common. Science 119, 623-626. Glauser, F.L. & Fairman, R.P. (1985). The uncertain role of the neutrophil in increased permeability pulmonary edema. Chest 88, 601-607. Goebel, K.M. & Storck, U. (1983). Effect of intra-articular orgotein versus a corticosteroid on rheumatoid arthritis of the knees. Am. J. Med. 74, 124-128. Guamieri, C, Ceconi, C, Muscari, C, & Flamigni, F. (1982). Influence of oxygen radicals on heart metabolism. In: Advances in Studies on Heart Metabolism (Caldarera, CM. & Harris, P., eds.), pp. 423-432. CLUEB, Bologna. Hall, E.J. (1973). Radiobiology for the Radiologist. Harper & Row Publishers, Philadelphia. Halliwell, B. (1978). Biochemical mechanisms accounting for the toxic action of oxygen on living organisms: The key role of superoxide dismutase. Cell. Biol. Int. Rep. 2, 113-128.

374

PETER A. SOUTHORN and GARTH POWIS

Halliwell, B. (1984). Oxygen radicals: A common sense look at their nature and medical importance. Med. Biol. 62, 71-77. Halliwell, B. & Gutteridge, J.M.C. (1984). Oxygen toxicity, oxygen radicals, transition metals and disease. Biochem. J. 219, 1-14. Hammond, B., Kontos, H.A., & Hess, M.L. (1985). Oxygen radicals in the adult respiratory distress syndrome in myocardial ischemia and reperfusion injury, and in cerebral vascular damage. Can. J. Physiol. Pharmacol. 63, 173-187. Harman, D. (1956). Aging: A theory based on free radical and radiation chemistry. J. Gerontol. 11, 298-300. Harman, D. (1981). The aging process. Proc. Natl. Acad. Sci. USA 78, 7124^7128. Hearse, D.J., Humphrey, S.M., & Chain, E.B. (1973). Abrupt reoxygenation of the anoxic potassiumarrested perfused rat heart: A study of myocardial enzyme release. J. Mol. Cell. Cardiol. 5, 395-407. Heffher, J.E. & Repine, J.E. (1989). Pulmonary strategies of antioxidant defense. Am. Rev. Respir. Dis. 140,531-554. Hempelmann, L.H. & Hoffman, J.G. (1953). Practical aspects of radiation injury. Ann. Rev. Nucl. Sci. 3, 369-392. Heston, W.E. & Pratt, A.W. (1959). Effect of concentration of oxygen on occurrence of pulmonary tumors in strain A mice. J. Natl. Cancer Inst. 22, 707-717. Hittner, H.M., Godio, L.B., Rudolph, A.J., Adams, J.M., Garcia-Prats, J.A., Friedman, Z., Kautz, J.A., 8L Monaco, W.A . (1981). Retrolental fibroplasia: Efficacy of vitamin E in a double-blind clinical study of preterm infants. N. Engl. J. Med. 305, 1365-1371. Humad, S., Zarling, E., Clapper, M., & Skosey, J.L. (1988). Breathe pentane excretion as a marker of disease activity in rheumatoid arthritis. Free Radic. Res. Commun. 5, 101-106. Hunninghake, G., Gadek, J., & Crystal, R. (1980). Mechanism by which cigarette smoke attracts polymorphonuclear leukocytes to lung. Chest 77, 273. Hutchinson, F. (1957). The distance that a radical formed by ionizing radiation can diffuse in a yeast cell. Radiat. Res. 7,473-^83. Ismail, G., Boxer, L.A., & Baehner, R.L. (1979). Utilization of liposomes for correction of the metabolic and bactericidal deficiencies in chronic granulomatous disease. Pediatr. Res. 13, 769-773. Jackson, R.M. (1985). Pulmonary oxygen toxicity. Chest 88, 900-905. Jamieson, N.V., Lindell, S., Sundberg, R., Southard, J.H., & Beizer, F.O. (1988). An analysis of the components in UW solution using the isolated perfused rabbit liver. Transplantation 46,512-516. Jolly, S.R., Kane, W.J., Bailie, M.B., Abrams, G.D., & Lucchesi, B.R. (1984). Canine myocardial reperfusion injury: Its reduction by the combined administration of superoxide dismutase and catalase. Circ. Res. 54,277-285. Kalyanaraman, B., Mason, R.P., Tainer, B., & Eling, T.E. (1982). The free radical formed during the hydroperoxide-mediated deactivation of ram seminal vesicles in hemoprotein-derived. J. Biol. Chem. 257,4764-4768. Klebanoff, S.J. (1980). Oxygen metabolism and the toxic properties of phagocytes. Ann. Intern. Med. 93,480^89, Kontos, H.A. (1985). Oxygen radicals in cerebral vascular injury. Circ. Res. 57, 508-516. Kontos, H.A., Wei, E.P., Dietrich, W.D., Navari, R.M., Povlishock, J.T., Ghatak, N.R., Ellis, E.F., & Patterson, J.L., Jr. (1981). Mechanism ofcerebral arteriolar abnormalities after acute hypertension. Am. J. Physiol. 240, H511-H527. Kueppers, F. & Black, L.F. (1974). ai-Antitrypsin and its deficiency. Am. Rev. Respir. Dis. 110, 176-194. Lazo, J.S., Sebti, S.M., & Filderman, A.E. (1987). Metabolism of bleomycin and bleomycin-like compounds. In: Metabolism and Action of Anti-Cancer Drugs (Powis, G. & Prough, R., eds.), pp. 194-210. Taylor and Francis, London.

Free Radicals in Cell Biology

375

Lucey, J.F. & Dangman, B. (1984). A reexamination of the role of oxygen in retrolental fibroplasia. Pediatrics 73, 92-96. Lunec, J., Blake, D.R., McCleary, S.J., Brailsford, S., & Bacon, P.A. (1985). Self-perpetuating mechanisms of immunoglobulin G aggregation in rheumatoid inflammation. J. Clin. Invest. 76, 2084-2090. Lunec, J., Halloran, S.P., White, A.G., & Dormandy, T.L. (1981). Free-radical oxidation (peroxidation) products in serum and synovial fluid in rheumatoid arthritis. J. Rheumatol. 8, 233—245. Martin, W.J., II, Gadek, J.E., Hunninghake, G.W., & Crystal, R.G. (1981). Oxidant injury of lung parenchymal cells. J. Clin. Invest. 68, 1277-1288. Martin, W.J., II, Powis, G.W., & Kachel, D.L. (1985). Nitrofurantoin-stimulated oxidant productions in pulmonary endothelial cells. J. Lab. Clin. Med. 105, 23-29. McCord, J.M. & Fridovich, I. (1968). The reduction of cytochrome c by milk xanthine oxidase. J. Biol. Chem. 243, 5753-5760. McCord J.M. & Fridovich, I. (1969). Superoxide dismutase: An enzymatic function for erythrocuprein (hemocuprein). J. Biol. Chem. 244, 6049-6055. McCord, J.M. & Fridovich, I. (1977). Superoxide dismutases: A history. In: Superoxide and Superoxide Dismutases (Michelson, A.M., McCord, J.M., & Fridovich, I., eds.), pp. 1-10. Academic Press, New York. McCord, J.M. & Roy, R.S. (1982). The pathophysiology of superoxide: Roles in inflammation and ischemia. Can. J. Physiol. Pharmacol. 60, 1346-1352. Meyrick, B.O., Ryan, U.S., & Brigham, K.L. (1986). Direct effects of E coli endotoxin on structure and permeability of pulmonary endothelial monolayers and the endothelial layer of intimal explants. Am. J. Pathol. 122,140-151. Michelson, A.M. (1982). Oxygen radicals. Agents Action, Suppl. 11, 179-201. Morita, S., Snider, M.T., & Inada, Y. (1986). Increased N-pentane excretion in humans: A consequence of pulmonary oxygen exposure. Anesthesiology 64, 730-733. Oberley, L.W. & Buenner, G.R. (1979). Role of superoxide dismutase in cancer: A review. Cancer Res. 39,1141-1149. O'Brien, P.J. (1985). Free-radical-mediated DNA binding. Environ. Health Perspec. 64, 219-232. Ognibene, F.P., Martin, S.E., Parker, M.M., Schlesinger, T., Roach, P., Burch, C , Shelhamer, J.H., & Parrillo, J.E. (1986). Adult respiratory distress syndrome in patients with severe neutropenia. N. Engl. J. Med. 315, 547-551. OsherofF, M.R., Schaich, K.M., Drew, R.T., & Borg, D.C. (1985). Failure of desferrioxamine to modify the toxicity of paraquat in rats. J. Free Radic. Biol. Med. 1, 71-82. Padmanabhan, R.V., Gudapaty, R., Liener, I.E., Schwartz, B.A., & Hoidal, J.R. (1985). Protection against pulmonary oxygen toxicity in rats by the intratracheal administration of liposome-encapsulated superoxide dismutase or catalase. Am. Rev. Respir. Dis. 132, 164-167. Parks, D.A., Bulkley, G.B., & Granger, D.N. (1983). Role of oxygen free radicals in shock, ischemia, and organ preservation. Surgery 94,428-432. Phelps, D.L. & Rosenbaum, A.L. (1979). Vitamin E in kitten oxygen-induced retinopathy. II. Blockage of vitreal neovascularization. Arch. Ophthalmol. 97, 1522-1526. Pitot, H.C. (1981). Fundamentals ofOncology. Second edition, pp. 103-113. Marcel Dekker, New York. Powis, G. (1987). Anthracycline metabolism and free radical formation. In: Metabolism and Action of Anti-Cancer Drugs (Powis, G. & Prough, R., eds.), pp. 211-260. Taylor and Francis, London. Pryor, W.A. (1982). Free radical biology: Xenobiotics, cancer, and aging. Ann. N Y Acad. Sci. 393, 1-22. Pryor, W.A. (1986). Oxy-radicals and related species: Their formation, lifetimes, and reactions. Annu. Rev. Physiol. 48, 657-667. Pryor, W.A. & Dooley, M.M. (1985). Inactivation of human alphal-proteinase inhibitor by cigarette smoke: Effect of smoke phase and buffer. Am. Rev. Respir. Dis. 131, 941-943.

376

PETER A. SOUTHORN and GARTH POVVIS

Pryor, W.A., Dooley, M.M., & Church, D.F. (1986). The inactivation of a-1-proteinase inhibitory by gas-phase cigarette smoke: Protection by antioxidants and reducing species. Chem. Biol. Interact. 57,271-283. Pryor, W.A., Prier, D.G., & Church, D.F. (1983). Electron-spin resonance study of mainstream and sidestream cigarette smoke: Nature of the free radicals in gas-phase smoke and in cigarette tar. Environ. Health Perspect. 47, 345-355. Raj, J.U., Hazinski, T.A., & Bland, R.D. (1985). Oxygen-induced lung microvascular injury in neutropenic rabbits and lambs. J. Appl. Physiol. 58, 921-927. Rao, P.S. & Mueller, H.S. (1983). Lipid peroxidation and acute myocardial ischemia. In: Myocardial Injury (Advances in Experimental Medicine and Biology) (Spitzer, J.J., ed.), Vol. 161, pp. 347-363. Plenum Press, New York. Robin, E.D. & Burke, CM. (1986). The recrudescent epidemic of retrolental fibroplasia. Chest 89, 576-578. Romson, J.L., Hook, B.G., Kunkel, S.L., Abrams, G.D., Schorr, M.A., &. Lucchesi, B.R. (1983). Reduction of the extent of ischemic myocardial injury by neutrophil depletion in the dog. Circulation 67, 1016-1023. Rosen, G.M. & Rauckman, E.J. (1982). Carbon tetrachloride-induced lipid peroxidation: A spin trapping study. Toxicol. Lett. 10, 337-344. Sanford, K.K., Parshad, R., Jones, G., Handleman, S., Garrison, C, & Price, F. (1979). Role of photosensitization and oxygen in chromosome stability and "spontaneous" malignant transformation in culture. J. Natl. Cancer Inst. 63, 1245-1255. Schal, R.S. (1981). Age Pigments, pp. 303-316. Elsevier Biomedical Press, Amsterdam. Schneider, E.L. & Reed, J.D., Jr. (1985). Life extension. N. Engl. J. Med. 312, 1159-1168. Schoenberg, M., Younes, M., Muhl, E., Sellin, D., Fredholm, B., & Schildberg, F.W. (1983). Free radical involvement in ischemic damage of the small intestine. In: Oxy Radicals and Their Scavenger Systems (Greenwald, R. & Cohen, G., eds.). Vol. 2, pp. 154-157. Elsevier Science Publishers, New York. Senator, G.B. & Muirden, K.D. (1968). Concentration of iron in synovial membrane, synovial fluid, and serum in rheumatoid arthritis and other joint diseases. Ann. Rheum. Dis. 27,49-53. Shasby, D.M., Fox, R.B., Harada, R.N., & Repine, J.E. (1982). Reduction of the edema of acute hyperoxic lung injury by granulocyte depletion. J. Appl. Physiol. 52, 1237-1244. Shasby, D.M., Vanbenthuysen, K.M., Tate, R.M., Shasby, S.S., McMurtry, I., & Repine, J.E. (1982). Granulocytes mediate acute edematous lung injury in rabbits and in isolated rabbit lungs perfused with phorbol myristate acetate: Role of oxygen radicals. Am. Rev. Respir. Dis. 125, 443-447. Shlafer, M., Kane, P.F., & Kirsh, M.M. (1982). Superoxide dismutase plus catalase enhances the efficacy of hypothermic cardioplegia to protect the globally ischemic, reperfused heart. J. Thorac. Cardiovasc. Surg. 83, 830-«39. Slater, T.F. (1984). Free-radical mechanisms in tissue injury. Biochem. J. 222, 1-15. Smith, R.J. (1978). Poisoned pot becomes burning issue in high places. Science 200,417-418. Snider, G.L. (1981). The pathogenesis of emphysema—twenty years of progress. Am. Rev. Respir. Dis. 124,321-324. Stewart, J.R., Blackwell, W.H., Crute, S.L., Loughlin, V., Greenfield, L.J., & Hess, M.L. (1983). Inhibition of surgically induced ischemia/reperfusion injury by oxygen free radical scavengers. J. Thorac. Cardiovasc. Surg. 86, 262-272. Sullivan, P.D. (1985). Free radicals of benzo(a)pyrene and derivatives. Environ. Health Perspect. 64, 283-295. Szent-Gyorgyi, A., Isenberg, I., & Baird, S.L., Jr. (1960). On the electron donating properties of carcinogens. Proc. Natl. Acad. Sci. USA 46, 1444-1449. Tappel, A.L. (1962). Vitamin E as the biological lipid antioxidant. Vitam. Horm. 20, 493-510. Tate, R.M. & Repine, J.E. (1983). Neutrophils and the adult respiratory distress syndrome. Am. Rev. Respir. Dis. 128, 552-559.

Free Radicals in Cell Biology

377

Tate, R.M., Vanbenthuysen, K.M., Shasby, D.M., McMurtry, I.F., & Repine, J.E. (1982). Oxygen-radical-mediated permeability edema and vasoconstriction in isolated perfused rabbit lungs. Am. Rev. Respir. Dis. 126, 802-«06. Taube, H. (1965). Mechanisms of oxidation with oxygen. J. Gen. Physiol. 49 Suppl, 29-50. Tauber, A.I., Borregaard, N., Simons, E., & Wright, J. (1983). Chronic granulomatous disease:. A syndrome of phagocyte oxidase deficiencies. Medicine (Baltimore) 62, 286-309. Templeton, A.C. (1975). Acquired Diseases. In: Persons at High Risk of Cancer: An Approach to Cancer Etiology and Control (Fraumeni, J.F., Jr., ed.), pp. 69-83. Academic Press, New York. Toledo-Pereyra, L.H., Simmons, R.L., & Najarian, J.S. (1974). Effect of allopurinol on the preservation of ischemic kidneys perfused with plasma or plasma substitutes. Ann. Surg. 180, 780-782. Tolmasoflf, J.M., Ono, T., & Cutler, R.G. (1980). Superoxide dismutase: Correlation with life-span and specific metabolic rate in primate species. Proc. Natl. Acad. Sci. USA 77,2777-2781. Tripathi B.J. & Tripathi, R.C. (1984). Cellular and subcellular events in retinopathy of oxygen toxicity with a preliminary report on the preventive role of vitamin E and gamma-aminobutyric acid: A study in vitro. Curr. Eye Res. 3, 193-208. Turrens, J.F., Crapo, J.D., & Freeman, B.A. (1984). Protection against oxygen toxicity by intravenous injection of liposome-entrapped catalase and superoxide dismutase. J. Clin. Invest. 73, 87-95. Turrens, J.F., Freeman, B.A., Levitt, J.G., & Crapo, J.D. (1982). The effect of hyperoxia on superoxide production by lung submitochondrial particles. Arch. Biochem. Biophys. 217,401-410. Van Asbeck, B.S., Hoidal, J., Vercellotti, G.M., Schwartz, B.A., Moldow, C.F., & Jacob, H.S. (1985). Protection against lethal hyperoxia by trachael insufflation of erythrocytes: Role of red cell glutathione. Science 227, 756-759. Varma, S.D., Chand, D., Sharma, Y.R., Kuck, J.F., Jr., & Richards, R.D. (1984). Oxidative stress on lens and cataract formation: Role of light and oxygen. Cufr. Eye Res. 3, 35-57. Weitberg, A.B., Weitzman, S.A., Destrempes, M., Latt, S.A., & Stossel, T.P. (1983). Stimulated human phagocytes produce cytogenetic changes in cultured mammalian cells. N. Engl. J. Med. 308, 26-30. Weitzman, S.A. & Stossel, T.P. (1982). Effects of oxygen radical scavengers and antioxidants on phagocyte-induced mutagenesis. J. Immunol. 128, 2770-2772. Zaslow, M.C., Clark, R.A., Stone, P.J., Calore, J.D., Snider, G.L., & Franzblau, C. (1983). Human neutrophil elastase does not bind to alpha]-protease inhibitor that has been exposed to activated human neutrophils. Am. Rev. Respir. Dis. 128,434-439. Zimmerman, R. & Cerutti, P. (1984). Active oxygen acts as a promoter of transformation in mouse embryo C3H/10T/C18 fibroblasts. Proc. Natl. Acad. Sci. USA 81,2085-2087. Zweier, J.L., Flaherty, J.T., & Weisfeldt, M.L. (1985). Observation of free radical generation in the post-ischemic heart (abstract). Circulation 72 Suppl 3, 111-350.

RECOMMENDED READINGS Balentine, J.D. (ed.) (1982). Pathology of Oxygen Toxicity. Academic Press, New York. Halliwell, B. & Gutteridge, J.M.C. (1985). Free Radicals in Biology and Medicine. Clarendon Press, Oxford.

This Page Intentionally Left Blank

INDEX

Acetyl-CoA carboxylase, 140-141 Acetylcholine, 23 Actin, 13-16, 18-20, 143-145 {see also "ATP...") Actinin, 265-266 Activator protein 2 (AP-2), 149 Actomyosin, 38, 43 Adducin, 262 Adenine, 3-4, 50, 51 Adenosine, 50, 51 hyperemia, as mediator of, 61 and Pi-purinoceptors, 52-53 as retaliatory metabolite, 60 Adenosine-5'-diphosphate (ADP), 50, 51 and P2-purinoceptors, 55 Adenosine-5'-monophosphate (AMP), 50, 51 Adenosine-5'-triphosphate (ATP), 50,51 and P2-purinoceptors, 55 Adenylate cyclase (AC) family, 79S5, 129 {see also "Isozymes...") Adenylate kinase, ATP and, 27-28, 36 {seealso''ATP.,."^ Adhesion plaques, 155 Adult respiratory distress syndrome (ARDS), 361-363

Aerobic ATP production, control of, 40-43 by Ca'^ levels, 41-42 Krebs cycle dehydrogenases, 41, 42 ruthenium red, 42 defects, 42-43 cardiomyopathy, alcoholic, 42 cytochrome oxidase deficiency, 42 hypertension, 42 mitochondrial DNA, lesions on, 42 myopathies, 42 NADH dehydrogenase, 42 neuropathies, 42 pyruvate dehydrogenase, deficiencies of, 42 in mitochondrial oxidative phosphorylation system, 40 by nucleotide levels, 40-41 Aging: free radicals and, 367-368 ubiquitin system and, 282 Agmatine, 331 AIDS, polyamines and, 346 Allopurinol, 365-367 Alzheimer's disease, 267 and protein degradation, 277

379

380

AMP A, 313 Anaerobic glycolysis, 2 ANAPPs, 62 ANFs, 144 Angiotensin II, 299 Annexins, 257, 267-268 (see also "Calcium-binding.. ,^ annexin II, 268 annexin V, 268 anticoagulant, 267 antiinflammatory, 267 eight, 267-268 protein kinase C, as substrates for, 267 tyrosine kinase, as substrates for, 267 Anticancer drugs, free radicals and, 371 Anti-oncogenes, 188 ARDS, 361-363 Arginine, 331 Arrestin, 169 ATP, cellular, 1-47 introduction, 2-3 as energy currency of cell, 2 glycolysis, anaerobic, 2, 7 ion gradients, 2-3 oxidation, 2 transmembrane ion gradient, 23,16 levels, control of, 35-43 actomyosin, 38 adenine nucleotide translocase, 40 adenylate kinase, 36, 37 aerobic ATP production, control of, 40 {see also "Aerobic ATP...'0 agents of control, other, 38-39 amplification, 38-39 anaerobic ATP production, control of, 36-40 ATP synthase, 40, 41

INDEX

blood supply, restoration of, 39 Ca^ 36, 38, 4 M 2 cAMP,38,42 cooperativity, 37 epinephrine, 36, 39, 42 fructose 1,6 bisphosphatase (FBPase), 37, 39 glucose transporter, 39 glycogen phosphorylase, 38, 39 glycolysis, defects in, 39-40 in hemolytic anemia, 39 isocitrate dehydrogenase, 41 inosine monophosphate (IMP), 38 inotropy, 40, 42 Krebs cycle, 41, 42 L-type Ca^^ channel, 42 lactate, 40,41 mammals, three mechanisms of, 36 McArdle's syndrome, 39 muscle fatigue, 35 NaVK^pump,40 by nucleotide levels, 37-38,4041 oxidative phosphorylation, 36, 40-42,45 a-oxoglutarate dehydrogenase, 41 phosphofructokinase (PFK), 36-38 phosphorylase kinase, 38, 39 problem, circumventing, 36 protein kinase A, 42 proton gradient, 41 pyruvate dehydrogenase, 41,42 pyruvate kinase, 39-40 as regulator, 35-36 ruthenium red, use of, 42 sensitivity of pathway, 37 substrate cycles, 36 type lib muscle, 36-37 in vivo maintenance, 35 white muscle, 36-37

Index

measurement, 7-13 adenine nucleotide concentrations within cells, 10 in biopsies, 7-8 cardiomyocytes, 13 centrifugal fractionation, 11-12 creatine kinase, 9 digitonin, 11-12 fractionation in non-aqueous solvents, 11 freeze-clamp technique, 7-8 in heart muscle, 13 high energy phosphate, 10 high performance liquid chromatography (HPLC), 7 luciferase, 7 magnetic resonance technique, 8-10 phosphocreatine (PCr), 6, 8, 10, 26 rapid cell lysis/centrifugal fractionation, 11-12 spatial distribution, 10-11, 12 tomography, 11-13 pathological disturbances of levels, 43-44 actomyosin, 43-44 free radicals, 44 halothane, use of, 43 hyperpyrexia, malignant, 43 hyperthermia, malignant, 43 ischemia, 43-44 lactic acidosis, 43 myocardial infarct, 44 oxidative phosphorylation, 44,45 oxygen paradox, 44 reperfusion, 44 sarcoplasmic reticulum, 43 stroke, 44 reactions involving exchange of high energy phosphates, 25-28 adenylate kinase, 25, 27-28, 36 creatine kinase reactions, 25-27

381

structure, 3-7 acid anhydride link, 3, 6-7, 21 adenine, 3-4 chemical bonds and conformation, 3 condensation reactions, 7 creatine phosphate, 5 as energy source, reason for, 3-6 extended form, 3 features, other, 6-7 6G% 5 glycogen phosphorylase, 7 hydrolysis, highly negative free energy of, 5, 6, 14 kinetic stability, 5-6 phosphate groups, as source of in biochemical reactions, 7 phosphocreatine, 6, 8, 10 phosphorylation potential, 5 ribose, 3-4 versatility of, 6 summary, 44-45 synthesis of, 28-35 Ehrlich ascites tumor, 30 in erythrocytes, 28 in eye lens, 28 glycogen stores, 30 glycolysis, 28 high energy phosphate bond, generation of, 28 in kidney medulla, 28 loops of Henle, 28 lymphocytes, 30-31 macrophages, 30-31 Morris hepatoma, 30 oxidative phosphorylation, 28, 31-35 (see also "Oxidative phosphorylation") phosphorylase kinase, 29 protein kinase A, 29 substrate level phosphorylation, 28-31

382

uses, 13-25 acetylcholine, 23 actomyosin system, contraction of, 13-16, 18-20, 27, 35 adenyl cyclase, 24 adenylate kinase, 27-28 AMP, cyclic, generation of, 2425 in biosyntheses, 20-22 Ca'', 24-25, 36 cAMP, generation of, 24-25 catecholamine, 23 as charge neutralizer, 23 chromaffin granules, 23 coenzyme A, 21 condensation, 21 CTP, synthesis of, 21 as dehydrating agent, 21 epinephrine, 23, 24, 36 fructose intolerance, 22-23 glucagon, 24 glutamine, synthesis of, 21 glycogen phosphorylase, 23 glycogen synthase, 23 j3-guanidino propionate, 27 hexokinase, 22, 35 hydrolysis of ATP, 14 insulin receptors, 23 insulin release in pancreatic P cells, 25 ion pumping, 16-20 (see also "Ion pumping'O in islets of Langerhans, 20 K^ channels, ATP dependent, 25 leaving groups, activation of, 21-22 and messenger molecules, 24-25 muscle contraction, energetics of, 13-16 oxidative phosphorylation, 27 PCr shuttle, 27 as phosphate donor, 22-23

INDEX

phosphocreatine, 26-27 PPi, 21, 24 proteins, phosphorylation of, 23 pyrophosphatase, 21, 24 pyruvate dehydrogenase, 23 serotonin, 23 signaling, 24 structural role of, 25 sugars, phosphorylation of, 2223 tetanus, 14 transmembrane ion gradients, 16 tRNA, 21 UTP, 21 ATP-ubiquitin-mediated protein degradation, 275-284 introduction, 275-277 Alzheimer's disease, 277 half-lives of regulatory proteins, 276 medical relevance, potential, 281282 aging, cell senescence during, 282 birth defects, 282 inclusion body diseases, 282 muscle atrophy, 282 substrates of proteolytic pathway, 280-281 cycUns, 281 papillomavirus, 281 phytochrome, 281 tumor suppressor gene p53, 281 ubiquitin conjugates, 281 sununary, 282 ubiquitin-dependent proteolytic pathway, 277-280 cahnodulin, 279 N-end rule hypothesis, 279 20S proteasome, 280 26S multicatalytic protease complex, 280

Index

ubiquitin adenylate, 278 ubiquitin carrier proteins(E2), 278 ubiquitin-El adduct, 278 Atrial natriuretic factor (ANF), 133 Autophosphorylation, 125 Beta-carotene, 357 Birth defects, ubiquitin system and, 282 Bleomycin, 360, 371 BNP, 87-88 {see also "Isozymes...") Brain, oxidative phosphorylation in, 31 Brain natriuretic peptide (BNP), 8788 (see also "Isozymes...") C-natriuretic peptide (CNF), 87-88 (see also "Isozymes...") Ca^^calmodulin, 315 Cadherin, 185 cADPR, 322 Calbindin-9K and- D28K, 260-261, 267 in neurodegenerative diseases, 267 Calcilytics, 300 Calcimimetics, 300 Calcineurin, 260, 265 Calciosomes, 309, 322 Calcitonin, 286 hypocalcemia, 294 and osteoclasts, 295-297 osteoporosis, importance in, 287 in plasma Ca^^ homeostasis regulation, 287 Calcium, extracellular, regulation of cellular functions by, 285-304 C-cell, 294-295 calcitonin, 294 dihydropyridines, 294-295 L-type Ca^"^ channels, 294 medullary thyroid carcinoma cells, 294-295

383

introduction, 285-286 Ca^^ receptor protein, 286 calcitonin, 286, 294 cycUc AMP, 286 diacylglycerol, 286 messenger role, 286 osteoclasts, 286, 295-297 parafollicular cells, 286 parathyroid ceU, 286, 287-294 parathyroid hormone, 286 osteoclast, 295-297 bone resorption, 295-296 calcitonin receptors, 295-296 function, 295 neomycin, 297 PTH, 295-296 parathyroid cell, 286, 287-294 Ca^^ receptor, 290-291 calreticulin, 291 calsequestrin, 291 diacylglycerol, 292 glutamate, 291 pertussis toxin, role of, 289-290, 291 protein kinase C, 292-293 schematic, 290 "set point," 288 stimulus-secretion coupling, Ca^^ hypothesis of, 289 synthesis of PTH, 293-294 sensing cells, other, 297-299 angiotensin II, 299 in cytotrophoblast of placenta, 298 gastrin secretion, 298 Goblet ceUs, 298 in gastrointestinal tract, 297,298 in Henle's loop, 298 juxtaglomerular cell of kidney, 298-299 in kidney, 297 1,25-dihydroxyvitamin D3 synthesis, 297-298

384

in placenta, 297, 298 proximal tubule cells of kidney, 297 renin, 299 sununary, 300-301 systemic Ca^^ metabolism, regulation of, 286-287 calcitonin, 287 1,25-dihydroxyvitamin D3,287 osteoporosis, importance in, 287 and plasma PTH, 287 therapeutic significance of Ca^^ receptors, 299-300 calcilytics, 300 calcimimetics, 300 familial benign h3^rcalcemia, 299 hyperparathyroidism, 300 Calcium-binding proteins, intraceUular, 255-274 annexins, 267-268 annexin II, 268 annexin V, 268 anticoagulant, 267 antiinflanunatory, 267 eight, 267-268 protein kinase C, as substrates for, 267 tyrosine kinase, as substrate for, 267 calcium-transporting proteins, 268-270 Ca^^-Mg^^-ATPase, 268 Ca^^-Mg^^-ATPase pump, 268 glutamate receptors, 270 EF-hand CaBPs, 257-267 actinin, 265-266 adducm, 262 Ca^^ binding site, 258 calbindin-9K and -D28K, 260261, 267 calcineurin, 260, 265

INDEX

calcyclin, 260 calgranulin A and B, 260 calmodulin, 260, 261-263 calpain, 260, 265 cyclosporin A, 263 as domain of protein, 265-266 examples of, 259 functions of, 259 GAP-43, 262 integrins, 258 motif, 257-260 in nervous system, 266-267 neuromodulin, 262 oncomodulin, 260 parvalbumin, 260 protein family, 260 protein phosphatase, 260 sorcin, 260 troponm C, 263-265 introduction, 255-256 receptor-effector coupling mechanism, 256 other, 270 calsequestrin, 270 protein kinase C, 268 transglutaminase, 270 properties and classes of, 256-257 annexins, 257 ^^Ca overla5dng technique, 257 definition, 256 EF-hand CaBPs, 257-267 Stain-All, 257 summary, 270-271 Calcium homeostasis, intracellular, in eukaryotic cells, basis of, 305-327 conclusions, 325-326 how cells handle Ca^^, 308-325 Aga-IVA, 312 (S)-alfa-amino-3-hydroxy-5methyl-4-isoxazolepropionic acid (AMPA), 313

385

Index ^2+ 1

Ca -binding proteins, 308, 309, 323-325 Ca'^ buffers, 323-324 Ca^^-calmodulin, 315 Ca^^-induced Ca^^ release, 321 Ca^^ influx, 310, 314 Ca^^ Influx Factor (CIF), 315 Ca^^ oscillations, intracellular, 322-323 Ca^^ release activated channels, 309, 314-315, 319 Ca^^ sensors, 324-325 Ca^^-storing intracellular organelles, 308, 309, 316-319 Ca^^ transport systems of plasma membrane, 308-316 calciosomes, 309, 317, 322 calcium influx factor (CIF), 309 calmodulin, 309, 315-316, 324325 calpain, 316 calreticulin, 320 calsequestrin,318, 320, 321 cardiac glycosides, 316 ceU death, neuronal, 313 cGMP, 314, 315 chemical coupling, 321, 322 cholinergic nicotinic receptor/ channel, 313 co-CgTx, 312 channels, 308, 309, 310-315 cyclic ADP ribose (cADPR), 320 DHP receptors, 321 diacylglycerol (DAG), 309 E-F hand, 323-325 endoplasmic reticulum (ER), 317 exchangers, 308 excitation-contraction coupling, 311 extracellular ATP, 313 G-proteins, 312, 314 GTP-binding protein, 312

gene expression, 309 glutamate, 313 glycine, 313 inositol 1,3,4,5-tetrakisphosphate(IP4), 314 intracellular stores, 314, 315, 320-322, 323 ionotropic receptors, 313 IPs, 309, 314, 315, 317, 319, 322 ischemia, 313 junctional face membrane (JFM), 318 kainate, 313 L-type channels, 311-312 ligand-operated channels, 309, 313 long-term depression, 311 long-term potentiation, 311 mechanical coupling, 321 mitochondria, 317-318 Na'^/Ca^'^ exchanger, 316 Na^-K^-ATPase, 316 N-methyl-D-aspartate (NMDA),313 N-type channel, 312 neuronal cell death, 313 oncomodulin, 324 organelle Ca^^inding proteins, 320 organelle Ca^^-pumping mechanism, 320-321 P-type channels, 312 P2 purinergic receptors, 313 parvalbumin, 309, 323 phosphatydilinositol, 315-316 plasma membrane Ca^^ pump, 315-316 plasma membrane Ca2+-Mg^^ATPase(PMCA),309,315 pumps, 308, 315-316 Purkinje neurons, 312, 317 ryanodine (RY) receptor, 317, 318, 319, 321, 322

386

S-100 Ca^^ binding protein, 324 sarcoplasmic/endoplasmic reticulum Ca^^Mg^^-ATPase (SERCA), 309, 320-321 as second messenger, 308 second messenger operated channels (SMOCs), 309, 314 seven-membrane spanning receptors, 314 SR,315 triads, 316 troponin C, 325 voltage-operated channels (VOC), 309, 310-312 introduction, 305-307 Ca^ 306 cyclic nucleotides, 306 DAG, 306 diacylglycerol, 306 inositol phosphates, 306 membrane receptors, 306 receptors, 306 second messengers, 306 tyrosine specific kinase, 307 mechanisms of, 307-308 ceU death, 307 free cytoplasmic Ca^^, 307 malignant hyperthermia, 308 necrotic lysis, 307 plasma membrane Ca^^Mg^"^ATPase, 305 progranuned cell death, 307 sarcoplasmic reticulum (SR), 308 Calcyclin, 260 Calgranulin A and B, 260 Calmodulin, 144, 259, 260, 261-263, 309, 315 asCa^^ sensor, 324-325 and ubiquitination, 279 Calpain, 260, 265 Calreticulin, 291, 320

INDEX

Calsequestrin (CSQ), 270, 291, 318, 320, 321 CaM-kinases, 164-167 cAMP, 50 generation of by ATP, 24-25 cAMP-dependent protein kinase,131-133 catabolite gene activator protein (CAP), 133 inhibitory domain, 131-132 protein kinase inhibitor (PKI), 132 regulatory components Ri and Rii, 131 substrates of, 131 cAMP response element (CRE), 148-149 Cancer therapy: free radicals and, 370-371 role of polyamines in, 346 {see also "Polyamines...") Carbon tetrachloride (CCU), 368-369 Carcinogenesis,free radicals and, 369-370 Cardiomyocytes, 13 Catabolite gene activator protein (CAP), 132 Catalase, 357, 361, 365 Catecholamine, 23, 251 CeUular ATP, 1-47 {see also "ATP, cellular '0 CGD, 358 Chloroquine resistance of malaria parasite, 18 Chronic granulomatous disease (CGD), 358 CIF, 315 Cigarette smoking, emphysema and, 363 CNP, 87-88 {see also "Isozymes...") Creatine phosphate, 5 CREB, 150-151 CSQ, 318 Cyanide, 245

Index

Cyclic cascades in cellular regulation, 199-218 development of model, 204-208 enzynic activity, regulation of, 204-207 glnA, 207 glutamine synthetase regulation, 204-207 transcriptional regulation, 207208 introduction, 201-204 converter enzymes, 203 covalent interconversion of enzymes, 203-204 interconvertible enzymes, 203 protein phosphorylation, 202 reversible covalent modification of protein, 201-202 model, 208-218 adenylylation/ deadenylylation of glutamine synthetase, 212 amplitude, 214 biological integrators for metabolic signals, 215 converter enzyme, 208, 211-213 energy consumption, 217 flexibility and biological integrators for metabolic signals, 215 four-cycle cascade, 213 interconvertible enzyme, 208, 211-214 monocyclic cascade, 209, 215 multicyclic cascade, 208, 212 rate amplification, 215-217 sensitivity, 214-215 signal amplification, 211-214 versus simple allosteric control, 217-218 unidirectional versus cyclic cascade, 208-211 "zero-order ultrasensitivity," 215 summary, 218-219

387

Cyclic nucleotide phosphodiesterase family, 91-109 {see also "Isozymes...") Cyclic nucleotide synthesis and degradation, role of multiple isozymes in regulation of, 77121 (see also "Isozymes...") Cyclins, 281 Cyclosporin A, 263 Cystic fibrosis, 18 DAG, 309 Dehydrating agent, ATP as, 21 (see afao"ATP...") Dephosphorylation, 123-177 (see also "Protein phosphorylation...") DFMO, 340-344, 346 Diacylglycerol (DAG), 309 Digitalis, 244 (see also "Na^K"" ATPase...") Digitonin, 11-12 Dinitrophenol, 245 Doxorubicin, 371 dsl, 146 EDRF, 133, 144 EF-hand CaBPs, 257-267 (see also "Calcium-binding...") motif, 257-260 Eflomithine, 346 Ehrlich ascites tumor, 30 Emphysema, 363 and cigarette smoking, 363 Endothelial-derived relaxing factor (EDRF), 133, 144 Endotoxin, 361, 363 Epilepsy, 267 Epinephrine, 23, 24, 36, 39, 128-130, 141 Extracellular calcium, regulation of cellular functions by, 285304 (see also "Calcium, extracellular...")

388

Fenton reaction, 353 Free radicals, 44, 349-377 {see also "ATP..." and "Oxygen free radicals...") Friend erythroleukemia cells, 344 Fructose 1,6 bisphosphatase (FBPase), 37, 39 Fructose intolerance, 22-23 {see also "ATP...") G-substrate, 134 Gastrointestinal tract, extracellular Ca^^-sensing cells in, 297, 298 Glucagon, 24 Glucocorticoids, 250 Glucose transporter, 339 {see also "ATP...") Glutamine, 32 Glutathione peroxidase, 357 Glycogen phosphorylase, 38, 39 Glycogenolysis, regulation of, 129130 Glycolysis, anaerobic, 2 Goblet cells, 298 Guanine, 50, 51 Guanine nucleosides, 50 Guanosine, 51 Guanylate cyclase family, 85-92 {see a&o "Isozymes...") Gut, oxidative phosphorylation in, 32 Haber-Weiss reaction, 353, 357 Heart, oxidative phosphorylation in, 31 Heme-regulated inhibitor (HRI), 146 Hemolytic anemia, glycolytic enzyme deficiency in, 39-40 Hemorrhagic shock, 367 Henderson-Hasselbach equation, 224 Henle's loop, 298

INDEX

Hexokinase, 22, 35 High performance liquid chromatography (HPLC), 7 Huntington's disease, 107 Hydrogen peroxide, 352-357 {see also "Oxygen free radicals...") Hydroxymethylglutaryl-CoA reductase(HMG-CoA) as principal regulator of cholesterol synthesis in mammalian cells, 141-142 {see also "Protein phosphorylation...") Hyperoxia, 363-364 Hyperoxic lung injury, role of free radicals in, 360-361 Hyperpyrexia, malignant, 43 {see also "^ ATP.,,"^ Hypertension, severe, free radicals and,367 Hyperthermia, malignant, 43 {see also''ATP..:') Hypusine, 345 IMP, 38 Inclusion body diseases, ubiquitin and, 282 Inflammatory diseases, role of free radicals in, 359-360 Inosine monophosphate (IMP), 38 Insulin: Na^K^-ATPase activity, 250-251 Integrins, 258 Interferon, 146 Intracellular calcium homeostasis in eukaryotic cells, basis of, 305-327 {see also "Calcmm homeostasis...") Intracellular pH regulation, mechanisms of, 221-241 {see also "pH regulation...")

Index

Ion pumping, 16-20 (see also "ATP...") ABC-type pumps, 16, 18 and actomyosin function, combination with, 18-20 alternating access model, 18 antiports, 16 binding cassette pumps, 16, 18 Ca^^ pump, 17 catecholamines, 17 chromaffin granule, 17 cystic fibrosis, 18 endocytotic vesicles, 17 energetics of, 18-20 in eukaryotic cells, 16 F-type pumps, 16, 18 gastric H^ pump, 17 ionophores, 17 lysosomes, 17 multidrug resistance (MDR) protein, 18 NaVK^pump, 17, 19 P-glycoprotein, 18 P-type pumps, 16, 17 primary transport systems, 16 symports, 16 transmembrane ion gradients, 16 V-type pumps, 16, 17 IPs, 309 Ischemia, 43-44,267 (see also "ATP...") Ischemia-reperfusion injury, 364-367 after myocardial ischemia, 364-366 in other tissues, 366-367 oxygen paradox, 364-365 prevention of, 365-366 transplantation, 366-367 xanthine dehydrogenase, 365 Isozymes, multiple, role of in regulation of cyclic nucleotide synthesis and degradation, 77-121 adenylate cyclase family, 79-85 isoforms of, other, 84-85 properties, 80

389 type I CaM-stimulated, 81-82 type II Ga, +0)8, 7, 82 type III CaM-stimulated and higher calcium concentrations, 82-83 type IV, 83 type V, 83-84 type VI, 84 conclusion, 109 cyclic nucleotide phosphodiesterase family, 91-109 CaM-PDE family, 94-99 cGI-PDE, insulin-sensitive, 102103 cGMP-binding, 108 cGS-PDE, 100-102 conserved structural elements, 91-94 families, seven, 91 GMP-binding PDEs, 99 HCPl family, 107-109 Huntington's disease, 107 PDE isoenzymes, 91-93 properties, 92 transducin, 108 type I PDEs, 94-99 type II PDEs, 99-102 type III PDEs, 102-103 type IV PDEs, 104 type V PDEs, 105-106 type VI PDEs, 106-108 type VII PDEs, 108-109 zaprinast, 106 guanylate cyclase family, 85-92 ANP receptor, 87-88 ATP, binding, 85 atrial natriuretic peptide, 87-88 brain natriuretic peptide (BNP), 87 C-natriuretic peptide (CNP), 87-88 carbon monomide (CO), 91 cellular cOMP, 91

390 GC-A form, 87-88 GCAP, 89 GC-B form, 87 GC-C forms, 87-89 GC-D form, 87, 89 GC-S, 90-91 guanylin, 88 nitric oxide, 90 NO synthetase, 90-91, 98 particulate, 85 photoreceptor-specific form, 89 properties, 86 recoverin, 89 RetGC-1, 86-87 RetGC-2, 86-87 retina, photoreceptor-specific form in, 89 soluble, 85, 87, 90-91 introduction, 78-79 Junctional face membrane (JFM), 318 Kainate, 313 Kidney, extracellular Ca^^-sensing ceUs in, 297-298, 299 Krebs cycle dehydrogenases, 41 Lactic acidosis, 43 Lipofuscin, 368 Liver, oxidative phosphorylation in, 31-32 Loops of Henle, 28 Luciferase, 7 Luffs disease, 35 Lymphocytes, ATP and, 30-31(5ee a&o"ATP...'0 Macrophages, ATP and, 30-31 (see Magnetic resonance technique (NMR), 8-10 Marijuana, paraquat and, 369

INDEX

McArdle's syndrome, 39 {see also "ATP...") Mechanisms of intracellular pH regulation, 221-241 {see also "pH regulation...") Methyl viologen, 353-354, 360, 369 MLCK, 144 Mevalonic acid, 141-142 Migraine headaches, 61 Mineralocorticoids, 250 Morris hepatoma, 30 Muscle atrophy, ubiquitin system and, 282 Myocardial infarct, 44 Myosin, 13-16 {see also "ATP...") and actin interaction, 143-145 light-chain kinase (MLCK), 144 N-end rule hypothesis, 279 NaVH' exchange (NHE), 225-227 Na^K^-ATPase in cells, 242-253 introduction, 243-244 digitalis, 244 and plasma membrane ionic gradients, origin of, 244-246 cell physiology of, 246 cyanide, 245 dinitrophenol, 245 discovery of, 245-246 erythrocyte membranes, 245 ouabain, 245, 249 pump-leak model, 244-245 regulation of, 250-252 cardiac contractility, increased, pump inhibition and, 251252 catecholamines, 251 glucocorticoids, 250 hormonal influence on, 250-251 insulin, 250-251 mineralocorticoids, 250 pump inhibition and increased cardiac contractility, 251-252

Index

substrate regulation of catalytic turnover, 251 thyroid hormones, 250 thyroid thermogenesis, 250 vasopressin, 251 structure and catalysis, 246-250 catalytic cycle, 247-249 isoforms, three a, 249 ouabain, 249 subunit composition, 246-247 subunit heterogeneity, 249-250 vanadate, 248, 250 summary, 253 NADPH oxidase, 358-359 NCAM, 184 Neomycin, 297 Neuromodulin, 262 NHE, 225-227 Nitrofurantoin, 360 NMDA, 313 NMR, 8-10, 223 Norepinephrine, 129-130 Northern blot analysis, 83, 85, 101 Nuclear magnetic resonance (NMR), 8-10, 223 Oncogenes, 188-193 Oncomodulin, 260, 324 Orbitals, 350 (see also "Oxygen free radicals...") Osteoclasts, 286 Osteoporosis, importance of calcitonin in, 287 Ouabain, 245, 249 Oxidation, 2 (see also "ATP...") Oxidative DNA damage, 368 Oxidative phosphorylation, 31-35 electrochemical H^ gradient, 31 FiFoATPase, 31 fuels for, 31-32 in brain, 31 in gut, 32 in heart, 31

391 in liver, 32 in skeletal muscle, 31 Krebs cycle, 31 in mitochondria, 31 adenine nucleotide translocase, 34 alternating site mechanism, 33 electrochemical potential, 32 mechanism of ATP synthase, 32-34 MgATP complex, 34 P / 0 ratios, 34-35 Luffs disease, 35 uncouplers, 34 Oxygen free radicals in cell biology, 349-377 in biology, 352-354 Fenton reaction, 353 Haber-Weiss reaction, 353, 357 hydrogen peroxide, 352-354 hydroxyl radicals, 353 paraquat pyridinyl cation, 353354 peroxisomes, 353 superoxide anion radical, 352353 superoxide dismutase (SOD), 353, 361 chemical background of, 350 definition, 350 orbitals, 350 dioxygen, configuration of, 351 and disease, 357-371 adult repsiratory distress syndrome (ARDS), 361-363 aging, 367 allopurinol, 365-367 alphai-protease inhibitor (ai-PI) deficiency, 363 bleomycin, 360, 371 cancer therapy, 370-371 carbon tetrachloride (CCU), 368-369

INDEX

392 carcinogenesis, 369-370 catalase, 361, 365 cerebrovascular damage, 367 chronic granulomatous disease (CGD), 358 criteria for determining role, 358 doxorubicin, 371 emphysema, 363 endotoxin, 361, 363 foreign compounds, toxicity of, 368-369 hemorrhagic shock, 367 hyperoxia, 363-364 hyperoxic lung injury, 360-361 hypertension, severe, 367 inflammatory diseases, 359-360 ionizing radiation, 371 ischemia-reperfusion injury, 364-367 {see also "Ischemiareperfusion. . .'^ lipofuscin, 368 methyl viologen, 369 microorganisms, invading, defense against, 358-359 myocardial ischemia, 364-366 NADPH oxidase, 358-359 neutrophil lung sequestration, 361 oxidative DNA damage, 368 oxygen paradox, 364-365 paraquat (methyl viologen), 369 pentane, 360 phagocytosis, 358 phorbol myristate acetate (PMA), 362 radiation, 368 reperfusion injury, 364-367 "respiratory burst," 358 retinopathy of prematurity, 363364 retrolental fibroplasia, 363-364 rheumatoid arthritis, 359-360

superoxide dismutase, 353, 361, 367, 368 transplantation, 366-367 vitamin C supplements, 363 vitamin E, effect of, 364, 365, 368 xanthine dehydrogenase, 365 xanthine oxidase, 366, 367 introduction, 350 molecular oxygen, sequential reduction of, 351 protection machanisms against, 356-357 beta-carotene, 357 catalase, 357 glutathione peroxidase, 357 hydrogen peroxide, breakdown of, 357 superoxide dismutase, 356-357 vitamins C and E, 357 reactivity of in biologic systems, 354-356 hydroxyl radical, 354-355 lipid peroxidation, 354-355 PUFAs, 354-355 summary, 371-372 Oxygen paradox, 44, 364-365 P2 purinergic receptors, 313 p39"^^ 153 p53,281 Papillomavirus, 281 Parafollicular cells, 286 Paraquat, 353-354, 360, 369 Parathyroid cell, 284, 287-294 (see also "Calcium, extracellular...") in plasma Ca^^ cencentration, 287 "set-point" for extracellular Ca^^, 288 Parkinson's disease, 267 Parvalbumin, 260, 309, 323 PCr, 6, 8, 10

Index

PCR, 84 Pentane, 360 Peroxisomes, 353 (see also "Oxygen free radicals...") pH regulation, intracellular, mechanisms of, 221-241 concluding summary, 240-241 cytosolic pH regulation, mechanisms of, 225-230 CI-/HCO3- exchange, 227 H^ OH- and HCO3-, permeation of through channels, 229-230 H^-ATPases, 228-229 NaVH^ exchange (NHE), 225227 Na^-coupledCl-/HC03exchange, 228 NHE-1, 225-227 proton pump, 228-229 introduction, 222-223 importance of, 222 "secondary active," 222 measurement, 223-224 Henderson-Hasselbach equation, 224 nuclear magnetic resonance (NMR), 223 pH-sensitive fluorescent probes, 224 pH-sensitive microelectrodes, 223 weak acid or base partition, 224 organellar pH regulation, mechanisms of, 230-240 acidification of intracellular compartments, mechanisms of, 237-239 ATP synthetases, 237 breakdown products, transport of out of lysosomes, 236 intercompartmental transport, 231-237

393 lysosomal enzymes, delivery of, 235 lysosomal hydrolases, activation of, 236 M6P signal, 235 mitochondrial pH, 230-231 pathogens, entry of, 237 physiological role of acidic pH in subcellular organelles, 231-237 receptor-mediated endocytosis, 234-235 secretory granules, uptake of components into, 236 secretory products, maturation of, 236 trans-Golgi network (TON), 235 vacuolar H^-ATPases, control and regulation of, 239-240 vacuolar system of cell, 231 Phosphate donor, ATP as, 22-23 {see also''ATP..:") Phosphocreatine (PCr), 6, 8, 10, 26 Phosphofructokinase (PFK), 36-38 Phospholamban, 145 Phosphorylase kinase, 38, 39, 145146 Phosphorylation, protein, 123-177 {see also "Protein phosphorylation...") Phosphotyrosine phosphatases (PTPases), 172-174, 179-198 {see also "Protein tyrosine phosphatases...") Phytochrome, 281 PKI,132 {see also "Protein phosphorylation...") Placenta, extracellular Ca^^-sensing cells in, 297, 298 PMCA,315 Polyamines, roles of in cell biology, 329-348

394 in diagnosis and therapy, 346 AIDS, 346 cancer therapy, 346 DFMO, 346 eflornithine, 346 Pneumocystis carinii, 346 Trypanosoma b,gambiense, 346 West African sleeping sickness, 346 and growth, 339-345 AdoMetDC activity, 339-340, 343 and cell cycle, 339-340 depletion, consequences of, 340343 development and differentiation, 343-344 DFMO, 340-343 eIF-5A, 343 enzyme-activated irreversible inhibitors, 340 and Friend erythroleukemia cells, 344 hypusine, 343 IF2-a, 343 initiators, 343 macromolecular synthesis, 344345 ODC activity, 339, 343-344 polyamine biosynthesis, inhibitors of, 340 polyamine deficient mutants, 340 tumor growth, 343 introduction, 330-331 cadaverine, 330 characteristics, general, 331 history, 330-331 1,3-propanediamine, 330 putrescine, 330 spermidine, 330 spermine, 330 structure, 330

INDEX

macromolecules, interaction with, 336-339 acetylation, 337 acetylcholinesterase, 338 actin polymerization, 338 amino-acyl tRNA synthetase, 338 Ca^^ signaling, 338-339 cytoskeleton, 338 DNA, 337 electrostatic interactions, 336337 free and bound, 337 glutamate receptor of NMD A type, 338 microtubule formation, 338 in mitosis, 338 proteins, 338-339 ribosomes, 338 rRNA and tRNA, 337-338 signal transduction, 338-339 toroidal condensation, 337 tubulin, 338 metabolism, 331-336 acetylCoA, 332-334 AdoMetDC, 332-334, 336 agmatine, 331 arginine, 331 cSAT,334,336 decarboxy-S-adenosylmethionine (dAdoMet), 331 diamine oxidase, 332 enzymes, regulatory and nonregulatory, 332-336 5'-methylthioadenosine (MTA), 332 half-life, short, 334 hormones, growth factors, 334 methionine, 332 ODC antizyme, 332-334, 336 ornithine decarboxylase (ODC), 332-334, 336 oxidative deamination, 332

395

Index

polyamine oxidase, 332, 336 regulatory and nonregulatory enzymes, 332-336 S-adenosyl-methionine decarboxylase (AdoMetDC), 332-334,336 serum amine oxidase, 332 spermidine synthase, 331 spermine synthase, 331, 336 testosterone, 332 transport, 336 tumor promoters, 334 urinary excretion, 336 vertebrate organism, cycle in, 331-332 proteins, covalent binding to, 345346 transglutaminases, 345 summary, 346 Polymerase chain reaction (PCR) technology, 84 pp34 kinase, 152-153 cyclin, phosphorylation of by, 154 pp60^^^ 154, 156 pp60™, 155 ai-protease inhibitor ai-PI) deficiency, 363 Protein kinase C (PKC), 134-136 {see also "Protein phosphorylation...") annexin as substrate for, 267 as calcium-binding protein, 270 Protein kinase inhibitor (PKI), 132 (see also "Protein phosphorylation...") Protein phosphorylationdephosphorylation, biological functions of, 123-177 cell growth and differentiation, 151-154 approaches, three, 151-152 cdc2 protein kinase, 154

cyclin, phosphorylation of by pp34, 154 histone HI kinase (HIK), 152 MPF, 151-153 pj^mos^ 153

pp34 kinase, 152-153 pp60^"^ 154, 156 serine/threonine kinase, 153 cellular transformation, 154-163 adhesion plaques, 155 c-src, 155 glycine loop, 156 growth factor receptor tyrosine kinases (RTKs), 158-163 Lys 295, 156 pp60'"''', 156-157 pp60™', 155-157 protein-tyrosine kinases, 157 proto-oncogenes, 155-163 retroviral tyrosine kinases, 155158 srcgene, 155 tyrosine, phosphorylation of, 154-157 concluding remarks, 174-175 "cross talk," 174 enzyme regulation by reversible phosphorylation, 136-143 acetyl-CoA carboxylase, 140141 allosteric effectors, 137, 141 cholesterol synthesis in mammalian cells, 141-142 in diabetic tissues, 138 glycogen phosphorylase, 136 glycogen synthase (GS), 137-139 glycogen synthase kinases (GSK), 138 hormonal control, 138 hydroxymethylglutaryl-CoA reductase, 141-142 inhibitors, 138-139 mevalonic acid, 141-142

396 phosphatase inhibitor 1, 142 phosphorylase b, 137-139 pyruvate dehydrogenase (PDH), 139-140 reductase kinase, 142 tyrosine-3-hydroxylase, 142-143 gene transcription, regulation of, 147 activator protein 2 (AP-2), 149 cAMP response element (CRE), 148-149 covalent protein modification by phosphorylation, 147 CREB, 150-151 RNA polymerase I, regulation of, 147 tyrosine kinases, 148 introduction, 124-128 autophosphorylation, 125 historical overview, 125-128 importance of, 126-127 multi-site, 125 physiological relevance, 128 protein kinases, 124-125 "pseudosubstrate" site, 125 reversible phosphorylation, 125 role of, 124-125 signal transduction, 124 muscle contraction, regulation of, 143-146 calmodulin, 144 cardiac and skeletal muscles, 145-146 cGMP kinase, 144 smooth, 143-145 actin, 143-145 myosin, 143-145 myosin light-chain kinase (MLCK), 144 phospholamban, 145 skeletal muscles, 145-146 troponin C, 146

INDEX

neurological function, 163-170 arrestin, 169 Ca^^/calmodulin kinase II, 164167 gangliosides, 167 ion channel activity, modulation of by phosphorylation, 167-169 long-term potentiation, 165 neural transmission, mechanisms of, 163-164 rhodopsin kinase, 169-170 sensitization, 168-169 serotonin, 168 synapsin 1,166 transducin, 169 in vision, 169-170 phosphoprotein dephosphorylation, 170-174 glycogen synthase kinase 3,171 phosphatase 1,171 PP-2A, 172 PP-2B, 172 serine/threonine phosphatases, 170-172 tyrosine phosphatases, 172-174 protein synthesis,regulation of, 146-147 dsl, 146 elongation factor-1 (EF-1), 147 heme-regulated inhibitor (HRI), 146 interferon, 146 S6, 147 reversible, enzyme regulation by, 136-143 {see also ".. .enzyme regulation by...'0 second messenger systems and, 128-136 adenylate cyclase (AC), 129 atrial natriuretic factor (ANF), 133,144 cAMP, 128-129

Index

cAMP-dependent protein kinase, 131-133 {see also "cAMP-dependent...") catabolite gene activator protein (CAP), 133 catecholamines, 129-130 cGMP-dependent protein kinase, 133-134 endothelial-derived relaxing factor (EDRF), 133, 144 epinephrine, 128-130 G-substrate, 134 glycogenolysis, regulation of, 129-130 hormone action, mediation of, 128-130 norepinephrine, 129-130 phosphorylase kinase, 130 protein kinase C, 134-136 tetradecanoyl-phorbol-13acetate (TPA) tyrosine kinases, 136 Protein tyrosine phosphatases, family of, and control of cellular signaling responses, 179198 cytoplasmic nontransmembrane PTPases, 185-187 alternative splicing, 187 PTPHl, 186 SH2 domains, 186 superfamily of proteins including band 4.1, 185 dual specificity phosphatases, 193196 cell cycle, 194 growth factor stimulation, 194195 immediate early gene, 194-195 MAP kinase, 195-196 MKP-1, 196 mitogenic stimuli, 196 Ser/Thr kinase, 194-195

397 stress response gene, 196 3CH134, 196 VHl from Vaccinia wiruSy 194 introduction, 179-181 and diabetes, 180 kinases, 180 PTK activity, 180 reversible process, phosphorylation as, 181 physiological roles for, 188-191 HCP, 190-191 motheaten {me) phenotype, 190 RPTP7, 189-190 tumor suppressor genes, 188189 vanadate, 188 receptor-like transmembrane PTPases, 182-185 cadherin family, 185 cell-cell adhesion reactions, 184 "contact inhibition of cell growth," 184 contactin, 184 homophilic adhesion, 184-185 Ig superfamily of cell adhesion molecules, 184 NCAM, 184 subtypes, five, 183 regulation of PTPase activity, further aspects of, 187-188 anti-oncogenes, 188 intracellular targeting, 188 oncogenes, 188 phosphorylation, importance of, 188 signaling responses, cellular, as positive mediators of, 191193 CD45, 191-192 corkscrew mutation, 191 intracellular compartmentalization, 192 lymphocyte function, 191-192

398 neuronal diferentiation, 192 PTPa, 192 Src family kinases, 192-193 Syp, 191 structural diversity in, 181-182 cellular signalling responses, modulating, 182 PTPIB, 181 RCM, 181 Yersinia bacterium, 182 Yop PTPase, 182 structural features of, 182 summary and perspectives, 196-197 Proto-oncogenes, 155 PTPases, 172-174, 179-198 {see also "Protein tyrosine phosphatases...'^ Purines, 49-75 introduction, 49-52 adenine, 50 adenosine, 50 adenosine-5'-diphosphate (ADP), 50, 51 adenosine-5'-monophosphate (AMP), 50, 51 adenosine -5'-triphosphate (ATP), 50, 51 guanine 50 guanine nucleosides, 50 history, 50 Purinergic Hypothesis, 50 "purinergic nerves," 50 purinergic receptors, 50-52 purinoceptors, 50-52 structures of, 51 therapeutic potential, 52 3',5'-cycUc AMP, 50 what they are, 50 molecular biology of, 55-59 cloned adenosine (Pi) and ATP (P2) receptors, 56 G-protein coupled receptor family, 58

INDEX

membrane-spanning segments, seven, 55-56, 57 TGPCR, 56 Pi-purinoceptor pharmacology, 52-53 adenosine, activated by, 52 distribution of, 52 R-site, 52 P2-purinoceptor pharmacology, 53-55 and ADP, 55 distribution, 54 G-proteins, 54 ligand-gated, 54 suramin, 55 physiological and pathophysiological roles of, 60-64 in cardiovascular system, 60-63 in central nervous system, 63-64 retaliatory metabolite, 60 sununary and future directions, 66 in interstitial cystitis, 66 in trophic interactions, longterm, 66 therapeutic potential, 64-65 Adenocard, 64 dipyridamole, 64 hypotension, controlled, in aneurysm surgery, 64 supraventricular tachycardia, management of, 64 Purkinje neurons, 312, 317 Pyruvate dehydrogenase, deficiencies m,42 Pyruvate kinase, 39-40 Radiation, free radicals and, 368 Recoverin, 89, 267 Retaliatory metabolite, 60 Retinopathy of prematurity, 363-364 Retrolental fibroplasia, 363-364 Rheumatoid arthritis, role of free radicals in, 359-360

399

Index

Rhodopsin, 169-170 Ribose, 3-4 RNA,3 Rous sarcoma virus (RSV), 126, 154 (see also "Protein phosphorylation...") src gene, 155 (see also "Protein phosphorylation...") RPTP7, 189-190 Ruthenium red, studies using, 42 Ryanodine (RY) receptor, 317, 318, 319,321,322 Sarcoplasmic/endoplasmic reticulum Ca^^Mg^^-ATPase (SERCA), 309 Sensitization, 168-169 SERCA, 309 Serine/threonine kinase, 153 Serotonin, 23, 168 Skeletal muscle, oxidative phosphorylation in, 31-32 SMOCs,309,314 Smoking, cigarette, and emphysema, 363 Sorcin, 260 Stain-All, 257 Stroke, 44 Superoxide dismutase, 353, 361, 367, 368 Suramin, 55, 62 Tetanus, 14 (see also "ATP...") Tetradecanoyl-phorbol-13-acetate (TPA), 134 (see also "Protein phosphorylation...") TGN, 235 3',5'-cyclic AMP (cAMP), 50 Thrombin, 61 Thyroid hormones, 250 ^ TPA, 134 (see also "Protein phosphorylation...") Trans-Golgi network (TGN), 235

Transducin, 169-170 Transglutaminases, 270, 345 Transplantation, ischemia-reperfusion injury and, 366-367 Troponin C, 146, 263-265, 325 Tumor suppressor genes, 188-189 p53, 281 26S multicatalytic protease complex, 280 Tyrosine, phosphorylation of, 154-155 Tyrosine kinases, 136, 148 annexin as substrate for, 267 growth factor receptor, 158-163 protein-tyrosine kinases, 157 retroviral, 155-158 Tyrosine -3-hydroxylase, 142-143 Tyrosine phosphatases, 172-174 Ubiquitin, 275-284 (see also "ATPubiquitin-mediated...") adenylate, 278 -El adduct, 278 calmodulin, 279 jcarrier proteins (E2), 278 conjugates, 281 N-end rule hypothesis, 279 20S proteasome, 280 26S multicatalytic protease complex, 280 Uncouplers, 34 (see also "Oxidative...") Vanadate, 188, 248, 250 Vasopressin, 251 Vision, protein phosphorylation in, 169-170 rhodopsin, 169-170 Visionin, 267 West African sleeping sickness, polyamines and, 346 Xanthine dehydrogenase, 365 Xanthine oxidase, 366, 367

This Page Intentionally Left Blank