Current Topics in Membranes and Transport Volume 16 Electrogenic Ion Pumps
Advisory Board
M . P. Blaustein A . Essig...
9 downloads
685 Views
25MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Current Topics in Membranes and Transport Volume 16 Electrogenic Ion Pumps
Advisory Board
M . P. Blaustein A . Essig R. K . H . Kinne P. A . Knauf Sir H . L . Kornberg
P. Liiuger C. A . Pasternak W . D. Stein W . Stoeckenius K . J . Ullrich
Contributors
Qais Al-Awqati G . M . Baker T. Berglindh Moiru Ci@ D. R . DiBona R . A . Dilley Troy E. Dixon P. Leslie Dutton David C. Gadsby Peter Graher Dietrich Gradmann Ulf-PeterHansen Frunklin M . Harold William R. Harvey Erich Heinz Barry Honig Wovgang Junge H . R. Kuback Yusuo Kugawa I . A . Kozlov S . A . Lewis Peter C. Maloney P. A . Millner Harold J . Morowitz
Paul Mueller Robert Nielsen Daniel P. O'Ketfe Nigel K . Packham Roger C. Prince L . J . Prochaska E . Rubon G . Saccomani G . Sachs Teruo Shimmen V . P. Skulachev Clifford L. Slayrnun Roger M . SpanJwick H . B . Stewart N . E. Tandy Masashi TazawLi R. C . Thomas David M . Tiede Mario Vassa lle B . Wallmark Janncr P. Wehrlt. Mdrten Wikstriim N . K . Wills Michael G . Wolfersherger
Current Topics in Membranes and Transport Edited by Arnost Kleinzeller Department of Physiology University of Pennsylvania School of Medicine Philadelphia, Pennsylvania
Felix Bronner Department of Oral Biology University of Connecticut Health Center Farmington, Connecticut
VOLUME 16 Electrogenic Ion Pumps
Guest Editor Clifford L. Slayman Department of Physiology Yale University School of Medicine New Haven, Connecticut
Volume 16 is part of the series (p. mi) from the Yale Department of Physiology under the editorial supervision o f
Joseph F. Hoffman Department of Physiology Yale University School of Medicine New Haven, Connecticut
Gerhard Giebisch Department of Physiology Yale University School of Medicine New Haven, Connecticut
1982
@
ACADEMIC PRESS
A Subsidiary of Harcourt Brace Jovanovich, Publishers
New York London Paris San Diego San Francisco Sffo Paulo Sydney Tokyo Toronto
COPYRIGHT @ 1982, I BY ACADEMIC PRESS,INC. ALL RIGHTS RESERVED. NO PART O F THIS PUBLICATION MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM OR BY ANY MEANS, ELECTRONIC OR MECHANICAL, INCLUDING PHOTOCOPY, RECORDING, OR ANY INFORMATION STORAGE AND RETRIEVAL SYSTEM, WITHOUT PERMISSION IN WRITING FROM T HE PUBLISHER.
ACADEMIC PRESS, INC. 111 Fifth Avenue, New Yo r k , New
York 10003
United Kinadom Editiori Dublislied bv ACADEM~CPRESS, INC. ( L O N D O N ) 24/28 Oval Road, London N W l 7DX
LTD.
LIBRARY OF CONGRESS CATALOG CARD NUMBER:70-1 17091 ISBN 0-12- 153316-6 PRINTED IN THE UNITED STATES OF AMERICA
82 83 84 85
9 8 7 6 5 4 3 2 1
List of Contributors, xv Preface, xix Yale Membrane Transport Processes Volumes, xxi Contents of Previous Volumes, xxiii Historical Introduction, xxxi
DEMONSTRATION OF PUMP ELECTROGENICITY IN EUKARYOTIC CELLS
PART I .
CHAPTER
1
Electrophysiologyof the Sodium Pump in a Snail Neuron
R. C. THOMAS 1. Introduction, 3 11. Methods, 5
111. Results, 6 IV. Discussion, 14 References, 16
CHAPTER
2.
Hyperpolarizationof Frog Skeletal Muscle Fibers and of Canine Purkinje Fibers during Enhanced Na+-K+ Exchange: Extracellular K+ Depletion or Increased Pump Current?
DAVID C. GADSBY
I . Introduction, 17 11. Experiments Using Frog Skeletal Muscle Fibers, 19 111. Experiments Using Canine Cardiac Purkinje Fibers, 26 IV. Summary and Conclusions, 32 References, 33
vi
CONTENTS
3.
CHAPTER
The Electrogenic Pump in the Plasma Membrane of Nitella
ROGER M. SPANSWICK Introduction, 35 Evidence for the Electrogenic Pump, 36 Identity of the Pumped Ion, 37 Energy Source for the Electrogenic Pump, 38 Dependence of the Membrane Potential on External and Internal pH, 39 VI . The Relationship between the Electrogenic Pump and the Membrane Conductance, 42 VII. Conclusions, 44 References, 45 I. 11. 111. IV. V.
CHAPTER
4.
Control of Electrogenesis by ATP, Mg2+,H+, and Light in Petfused Cells of Chara
MASASHI TAZAWA AND TERUO SHIMMEN
I. Introduction, 49 11. Method for Controlling Intracellular Environment by Internal Perfusion, 51 111. Dependence of Electrogenesis and Net H+ Efflux on Mg'ATP, 53 IV. Dependence of Electrogenesis on pHi, pH,, and W I , , 55 V. Modulation of Electrogenesis by Light, 62 VI. Discussion, 63 VII. Concluding Remarks, 65 References, 66
PART 11.
CHAPTER
5.
THE EVIDENCE IN EPITHELIAL MEMBRANES
An Electrogenic Sodium Pump in a Mammalian Tight Epithelium
S. A. LEWIS AND N. K. WILLS I. Introduction, 71 Electrical Structure of an Epithelium, 72 Basic Transport Properties of Rabbit Urinary Bladder, 72 Electrical Measurements, 74 Epithelial Parameters, 76
11. 111. IV. V.
vii
CONTENTS
VI. VII.
Pump Properties, 79 Summary, 85 References, 86
CHAmER 6.
A coupled Electrogenic Na+-K+ Pump for Mediating Transepithelial Sodium Transport in Frog Skin
ROBERT NIELSEN I. Introduction, 87 11. Coupling between Active Na+ and K+ Transport, 91 111. Conclusions, 106 References. 107
CHAPTER
7.
Transepithelial Potassium Transport in Insect Midgut by an Electrogenic Alkali Metal Ion Pump
MICHAEL G. WOLFERSBERGER, WILLIAM R. HARVEY, AND MOIRA CIOFFl
I. 11. 111. IV.
Introduction, 109 Methods, 111 Behavior of the Midgut K+ Transport System, 114 Membrane Structure and Location of Transport Functions, 123 V. A Potassium Transport ATPase, 129 References, 132
CHAPTER
8.
The ATP-Dependent Component of Gastric Acid Secretion
G. SACHS, B. WALLMARK, G. SACCOMANI, E. RABON, H. B. STEWART, D. R. DIBONA, AND T. BERGLINDH
I. Introduction, 136 Site of Acid Secretion, 136 Energy Source for Acid Secretion, 140 Location of the K+-Dependent ATPase, 142 Nature of the ATPase, 144 Steps in ATP Hydrolysis, 145 H+ Transport by Gastric ATPase, 148 K+ Transport by Gastric ATPase, 150 Electrogenicity of the Pump, 153 X. pH Gradient and Stoichiometry, 154 XI. Structural Aspects of the ATPase, 156
11. 111. IV. V. VI. VII. VIII. IX.
viii
CONTENTS
XII. Summary and Conclusions, 157 References, 158
PART 111.
CHAPTER
9.
REVERSIBILITY: ATP SYNTHESIS DRIVEN BY ELECTRIC FIELDS
Effect of Electrochemical Gradients on Active H+ Transport in an Epithelium QAIS AL-AWQATI AND TROY E. DIXON
I. Introduction, 163 11. Proton Secretion by Turtle Bladder, 164 111. Efficiency of Energy Conversion, 167 IV. Reversibility, 168 V. Stoichiometry, 171 VI. Ion Transport as a Pacemaker of Cellular Metabolism, 172 VII. Conclusions, 173 References, 174
CHAPTER
10.
Coupling between H+ Entry and ATP Synthesis in Bacteria PETER C. MALONEY
1.
Introduction, 175 Voltage-Driven Reversal, 176 Proton Entry Coupled to ATP Synthesis, 178 IV. Stoichiometry of the Coupling between H+ and ATP, 184 V. Rates of ATP Formation and the Nature of the Driving Force, 187 VI. Conclusions, 191 References, 192 11. 111.
CHAPTER
11.
Net ATP Synthesis by H+-ATPase Reconstituted into Liposomes YASUO KAGAWA
I.
Introduction, 195
11. Electrogenic Properties of H+-ATPase, 197 111. Net ATP Synthesis Driven by APH+, 201
IV.
Molecular Properties of H+-ATPase, 207
ix
CONTENTS
V.
Epilogue, 211 References, 212
CHAWER
12.
Phosphorylation in Chloroplasts: ATP Synthesis Driven by A$ and by ApH of Artificial or Light-Generated Origin
PETER G d i B E R
I. 11. 111. IV. V. VI. VII.
Introduction, 215 Background Information, 216 Coupling of Proton Transport to ATP Synthesis, 219 The Functional Unit for ATP Synthesis, 228 The Kinetics of ATP Synthesis, 229 The Problem of Energetic Sufficiency, 239 Epilogue: Conformational Changes Associated with Energization, 241 References, 243
PART IV.
C H A ~ E R13.
SOME THEORETICAL QUESTIONS
Response of the Proton Motive Force to the Pulse of an Electrogenic Proton Pump
ERICH HEINZ I. Introduction, 249 11. Treatment in Terms of Thermodynamics of Irreversible
Processes, 250 References. 256
CHAPTER
14.
Reaction Kinetic Analysis of Current-Voltage Relationships for Electrogenic Pumps in Neurospora and Acetabularia
DIETRICH GRADMANN, ULF-PETER HANSEN, AND CLIFFORD L. SLAYMAN I. 11. 111. IV.
Introduction, 258 Theory: Reduction of Kinetic Models, 258 Results, 266 Extensions of the Model, 273 References, 276
CONTENTS
X
CHAPTER
15.
Some Physics of Ion Transport
HAROLD J. MOROWITZ
I. Free Ion and Ion Carrier Migration, 277 Ion Conductance, 278 References, 28 1
11.
MOLECULAR MECHANISMS OF CHARGE SEPARATION
PART V.
CHAPTER
16.
An H+-ATP Synthetase: A Substrate Translocation Concept
I . A. KOZLOV AND V. P. SKULACHEV I. The Substrate Translocation Hypothesis, 285 11. Determination of the Equilibrium Constant for the
Reaction ATP + H,O = ADP + Pi at the Active Site of H+-ATP Synthetase, 288 111. The Energy-Dependent Release of F,-Bound AMPPNP from the Membrane of Submitochondrial Particles, 290 IV. Comparative Inhibitor Analysis of Solubilized and Membrane-Bound Factor Fl, 292 References. 300
CHAPTER
17.
Proton Translocation by Cytochrome Oxidase
MARTEN
WIKSTROM
I. Introduction, 303 11. The Discovery of True Proton Pumping by Cytochrome
Oxidase, 307 Controversy over Proton Translocation by Cytochrome Oxidase, 310 IV. Molecular Principles and Mechanisms of Proton Translocation, 312 References. 318 111.
CHAPTER
18.
Electrogenic Reactions of the Photochemical Reaction Center and the Ubiquinone-Cytochrome blcz Oxidoreductase
P. LESLIE DUTTON, PAUL MUELLER, DANIEL P. O’KEEFE, NIGEL K. PACKHAM, ROGER C. PRINCE, AND DAVID M. TIEDE
xi
CONTENTS
I. Introduction, 324 11. The Reaction Center Protein, 325 111. The Ubiquinone-Cytochrome b/cz Oxidoreductase, 335 References, 342
CHAPTER
19.
Proton-Membrane Interactions in Chloroplast Bioenergetics
R. A. DILLEY, L. J. PROCHASKA, G. M. BAKER, N . E. TANDY, AND P. A. MILLNER
I. 11. 111. IV.
Introduction, 345 Methods and Rationale, 351 Results and Discussion, 352 Concluding Remarks, 363 References, 367
CHAPTER
20.
Photochemical Charge Separation and Active Transport in the Purple Membrane
BARRY HONIG I. Introduction, 371 11. The Primary Photochemical Event, 372
111. Mechanistic Implications of Steady State Kinetics, 377 IV. Relating Kinetic and Molecular Models, 379 V. Summary, 381 References, 381
CHAPTER
21.
Mitochondria1Transhydrogenase: General Principles of Functioning
I. A. KOZLOV I. Introduction, 383 11. The Hypothesis of the Mechanism of
APH+
Generation by the Transhydrogenase Reaction, 384 111. Known Facts and Forecasts, 387 IV. Conclusion, 391 References, 391
xii
CONTENTS
CHAPTER
22.
Membrane Vesicles, Electrochemical Ion Gradients, and Active Transport
H. R. KABACK
I. Introduction, 393 11. Molecular Architecture of Escherichia coli
Membrane Vesicles, 394 111. Chemiosmotic Phenomena, 395 IV. Carrier Action, 399 References, 402
PART VI.
CHAPTER
23.
BIOLOGICAL SIGNIFICANCE OF ELECTROGENIC ION PUMPS
The Role of Electrogenic Proton Translocation in Mitochondrial Oxidative Phosphorylation
JANNA P. WEHRLE
I. Introduction, 408 Respiration-Dependent Proton Pumping, 409 111. Reversible Electrogenic Proton Translocation by the F,-F, ATPase, 416 IV. The Role of Proton Translocation in Mitochondria1 Oxidative Phosphorylation, 422 V. Electrophoretic Metabolite Transport, 426 VI. Summary, 428 References, 428 11.
CHAETER
24.
Electrogenic Reactions and Proton Pumping in Green Plant Photosynthesis
WOLFGANG JUNGE
I. Introduction, 431 11. The Membrane, 433 111. Electrogenic Reaction Steps, 437 IV. Protolytic Reaction Steps, 449 V. Comments on the Pathway of Protons to the ATP Synthetase, 458 VI. Summary, 459 References, 461
xiii
CONTENTS C H A ~ E R25.
The Role of the Electrogenic Sodium Pump in Controlling Excitability in Nerve and Cardiac Fibers
MARIO VASSALLE I. Introduction, 467 11. The Excitation Process, 468 111. The Sodium Pump and Control of Excitability, 469 1V. Excitability in Nerve and the Electrogenic Sodium Pump, 471 V. Excitability in Cardiac Cells and the Electrogenic Pump, 474 VI. Concluding Remarks, 481 References, 482
C H A ~ E R26.
Pumps and Currents: A Biological Perspective
FRANKLIN M. HAROLD I. Introduction, 485 11. The Role of Ion Currents in the Metabolic Economy of
Bacteria, 487 111. Ion Currents and Energy Coupling in Eukaryotic Cells, 492 IV. Cellular Homeostasis, 495 V. Calcium Currents as Biological Signals, 501 VI. Transcellular Currents and Morphogenesis, 505 VII. A Sense of Direction, 510 References, 513
Index, 517
This Page Intentionally Left Blank
List of Contributors Numbers in parentheses indicate the pages on which the authors' contributions begin.
Qais Al-Awqati, Departments of Medicine and Physiology, Columbia University, College of Physicians and Surgeons, New York, New York 10032 (163) G. M. Baker, Department of Biological Sciences, Purdue University Biochemistry Program, Purdue University, West Lafayette, Indiana 47907 (345) T. Berglindh, Laboratory of Membrane Biology, University of Alabama, Birmingham, Ala-
bama 35233 (135) Moira Cioffi, Department of Biology, Temple University, Philadelphia, Pennsylvania 19122 (109) D. R. DiBona, Nephrology Research and Training Center, University of Alabama, Birmingham, Alabama 35233 (135) I?.A. Dilley, Department of Biological Sciences, Purdue University Biochemistry Program,
Purdue University, West Lafayette, Indiana 47907 (345) Troy E. Dixon,' Departments of Medicine and Physiology, Columbia University, College of Physicians and Surgeons, New York, New York 10032 (163) P. Leslie Dutton, Department of Biochemistry and Biophysics, University of Pennsylvania, Philadelphia, Pennsylvania 19104 (323) David C. Gadsby, Laboratory of Cardiac Physiology, The Rockefeller University, New York, New York 10021 (17) Peter Graber, Max-Volmer-Institut fur Biophysikalische und Physikalische Chemie, Technische Universitat Berlin, D-1000 Berlin 12, Federal Republic of Germany (215) Dietrich Gradmann,' Department of Physiology, Yale School of Medicine, New Haven, Connecticut 06516 (257)
'Present address: Department of Medicine, State University of New York at Stony Brook, Stony Brook, New York 11790. *Present address: Max-Planck Institut fur Biochemie, Abteilung Membranbiochemie, Munchen, Federal Republic of Germany.
xv
xvi
LIST OF CONTRIBUTORS
Ulf-Peter Hansen,JDepartment of Physiology, Yale School of Medicine, New Haven, Connecticut 06510 (257) Franklin M. Harold, Division of Molecular and Cellular Biology, National Jewish Hospital and Research Center, 3800 E. Colfax Avenue, Denver, Colorado, 80206, and Department of Biochemistry, Biophysics, and Genetics, University of Colorado Medical School, Denver, Colorado 80262 (485) William R. Harvey, Department of Biology, Temple University, Philadelphia, Pennsylvania 19122 (109) Erich Heinr, Department of Physiology, Cornell Medical School, New York, New York 10021 (249) Barry H ~ n i g Department ,~ of Physiology and Biophysics, The University of Illinois, Urbana, Illinois 61801 (371) Wolfgang Junge, Schwerpunkt Biophysik, Universitat Osnabriick, Postfach 4469,4500 Osnabriick, Federal Republic of Germany (431) H. R. Kaback, Laboratory of Membrane Biochemistry, Roche Institute of Molecular Biology, Nutley, New Jersey 07110 (393) YaSUO Kagawa, Department of Biochemistry, Jichi Medical School, Minamikawachimachi, Tochigi-ken 329-04, Japan (195) 1. A. Kozlov, Isotope Department and Department of Bioenergetics, A. N . Belozersky Laboratory of Molecular Biology and Bioorganic Chemistry, Moscow State University, Moscow 117234, USSR (285, 383) S. A. Lewis, Department of Physiology, Yale University School of Medicine, New Haven, Connecticut 06510 (71) Peter C. Maloney, Department of Physiology, The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205 (175) P. A. Millner, Department of Biological Sciences, Purdue University Biochemistry Program, Purdue University, West Lafayette, Indiana 47907 (345) Harold J. Morowitz, Department of Molecular Biophysics and Biochemistry, J. W. Gibbs Research Laboratory, Yale University, New Haven, Connecticut 065 10 (277) Paul Mueller, Department of Molecular Biology, Eastern Pennsylvania Psychiatric Institute, Philadelphia, Pennsylvania 19129 (323)
3Present address: Institut fur Angewandte Physik, Universitat Kiel, D-2300 Kiel, Federal Republic of Germany. 4Pre~entaddress: Department of Biochemistry, Columbia University, New York, New York 10032.
LIST OF CONTRIBUTORS
Robert Nielsen, University of Copenhagen, Institute of Biological Chemistry A, Copenhagen, Denmark (87)
xvii
DK 2100
Daniel P. O’Keefe, Department of Biochemistry and Biophysics, University of Pennsylvania, Philadelphia, Pennsylvania 19104 (323) Nigel K. Pa~kharn,~ Department of Molecular Biology, Eastern Pennsylvania Psychiatric Institute, Philadelphia, Pennsylvania 19129 (323) Roger C. Prince, Department of Biochemistry and Biophysics, University of Pennsylvania, Philadelphia, Pennsylvania 19104 (323)
L. J. Prochaska,‘ Department of Biological Sciences, Purdue University Biochemistry Program. Purdue University, West Lafayette, Indiana 47907 (345)
E. Rabon, Laboratory of Membrane Biology, University of Alabama, Birmingham, Alabama 35233 (135) G. Saccomani, Laboratory of Membrane Biology, University of Alabama, Birmingham, Alabama 35233 (135) G. Sachs, Laboratory of Membrane Biology, University of Alabama, Birmingham, Alabama 35233 (135) Teruo Shimmen, Department of Biology, Faculty of Science,Universityof Tokyo, Hongo, Tokyo 113, Japan (49) V. P. Skulachev, Isotope Department and Department of Bioenergetics, A. N. Belozersky Laboratory of Molecular Biology and Bioorganic Chemistry, Moscow State University, Moscow 117234, USSR (285) Clifford L. Slayman, Department of Physiology, Yale University School of Medicine, New Haven, Connecticut 06510 (257) Roger M. Spanswick, Section of Plant Biology, Division of Biological Sciences, Plant Science Building, Cornell University, Ithaca, New York 14853 (35)
H. 6. Stewart, Laboratory of Membrane Biology, University of Alabama, Birmingham, Alabama 35233 (135) N. E. Tandy, Department of Biological Sciences, Purdue University Biochemistry Program, Purdue University, West Lafayette, Indiana 47907 (345) Masashi Tazawa, Department of Biology, Faculty of Science, University of Tokyo, Hongo, Tokyo 113, Japan (49) SPresentaddress: Department of Botany, Imperial College of Science, London SW7 2BB, United Kingdom. ‘Present address: Biological Chemistry Program, School of Medicine, Wright State University, Dayton, Ohio 45435.
xviii
LIST OF CONTRIBUTORS
R. C. Thomas, Department of Physiology, Bristol University, Bristol BS8 ITD, England (3) David M. Tiede, Department of Biochemistry and Biophysics, University of Pennsylvania, Philadelphia, Pennsylvania 19104 (323) MarioVassalle, Department of Physiology, State University of New York, Downstate Medical Center. Brooklyn , New York 11203 (467) B. Wallmark, Laboratory of Membrane Biology, University of Alabama, Birmingham, Ala-
bama 35233 (135) Janna P. Wehrle,' Department of Physiological Chemistry, The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205 (407) Mlrten Wikstrom, Department of Medical Chemistry, University of Helsinki, Siltavuorenpenger 10 A, SF-00170, Helsinki 17, Finland (303) N. K. Wills, Department of Physiology, Yale University School of Medicine, New Haven,
Connecticut 06510 (71) MichaelG. Wolfersberger,Department of Biology, Temple University, Philadelphia, Pennsylvania 19122 (109)
'Present address: Department of Chemistry, University of Maryland-Baltimore County, Catonsville, Maryland 21228.
This volume represents the augmented proceedings of the Sixth Conference an Membrane Transport Processes held under the auspices of the Deparl ment of Physiology, Yale School of Medicine. These meetings were c riginated in 1975 as a memorial to our colleague, Peter Curran. The subject of this volume, which is the phenomenon of charge separation displayed by most active transport systems (pumps) in biological membrane:,, is one which began to interest Curran in the middle 1960s, well before the majority of physiological or biochemical scientists considered it legii imate. Because Curran’s contributions to the understanding of transport processes were broadly based, encompassing physical theory, physiological description, and biochemical mechanism, it is fitting that the volume should cover a similarly broad range of topics for electrogenic ion pumps. After the Historical Introduction, Parts I and I1 present physiological proof5 of pump electrogenicity ; Part I11 demonstrates that electrical gradients can be used to reverse pumps to make net ATP; Part IV considers selected theoretical problems; Part V deals with biochemical analysis and molecular mechanisms for transmembrane charge separation; and Part VI discusses more general biological questions: the physiological role played by electrogenic ion pumps and their integration into organellar and cellular economy. A rrlajor function both of the conference on electrogenic ion pumps and of this volume has been to bring together investigators from different disciplincs who have applied very different technical tools and very different technical languages to the study of active transport. In order to meld the manuscripts into a coherent whole, with a more-or-less common viewpoint and common style, many editorial changes were made, and the Editor is deeply indebted to the authors for their patience and indulgence during this rather lengthy process. And it is the Editor, not the authoi.~,who must take responsibility for the abbreviation of reference lists and restriction of the research reviews mainly to work from single laboratories. We are collectively indebted to the Physiology faculty and staff members who worked to make the conference possible, and particularly to Rita Scott, Marie Santore, and their secretarial staff who have gone xix
xx
PREFACE
through many retypings of manuscripts. In addition to the Yale Department of Physiology, the Squibb Institute for Medical Research, Princeton, New Jersey, and Merck Sharp and Dohme Research Laboratories, Rahway, New Jersey, provided the financial support which made the entire effort possible.
CLIFFORD L. SLAYMAN
Yale Membrane Transport Processes Volumes
Joseph F. Hoffman (ed.). (1978). “Membrane Transport Processes.” Vol. 1. Raven, New York. Daniel C. Tosteson, Yu. A. Ovchinnikov, and Ramon Latorre (eds.). (1978). “Membrane Transport Processes,” Vol. 2. Raven, New York. Charles F. Stevens and Richard W. Tsien (eds.). (1979). “Membrane Transport Processes,” Vol. 3: Ion Permeation through Membrane Channels. Raven, New York. Emile L. Boulpaep (ed). (1980). “Cellular Mechanisms of Renal Tubular Ion Transport”: Volume 13 of Current Topics in Membranes and Transport (F. Bronner and A. Kleinzeller, eds.). Academic Press, New York. William H. Miller (ed.). (1981). “Molecular Mechamisms of Photoreceptor Transduction”: Volume 15 of Current Topics in Membranes and Transport (F. Bronner and A. Kleinzeller, eds.). Academic Press, New York. Clifford L. Slayman (ed.). “Electrogenic Ion Pumps”: Volume 16 of Currenr Topics in Membranes and Transport (A. Kleinzeller and F. Bronner, eds.). Academic Press, New York.
xxi
This Page Intentionally Left Blank
Contents of Previous Volumes Volume 1
Volume 3
Some Considerations about the Structure of Cellular Membranes AND MAYNARD M. DEWEY' LLOYDBARR The Transport of Sugars across Isolated Bacterial Membranes H. R. KABACK Galactoside Permease of Escherichia coli ADAMKEPES Sulfhydryl Groups in Membrane Structure and Function ASERROTHSTEIN Molecular Architecture of the Mitochondrion H. MACLENNAN DAVID Author Index-Subject Index
The Na+, K+-ATPase Membrane Transport System: Importance in Cellular Function ARNOLD SCHWARTZ, AND GEORGE E. LANDENMAYER, JULIUS C. ALLEN Biochemical and Clinical Aspects of Sarcoplasmic Reticulum Function ANTHONY MARTONOSI The Role of Periaxonal and Perineuronal Spaces in Modifying Ionic Flow across Neural Membranes W. J . ADELMAN, JR. A N D Y. PALTI Properties of the Isolated Nerve Endings GEORGINA RODRiGUEZ DE LORES ARNAIZ A N D EDUARDO DE ROBERTIS Transport and Discharge of Exportable Proteins in Pancreatic Exocrine Cells: In Vitro Studies J. D. JAMIESON The Movement of Water across Vasopressin-Sensitive Epithelia RICHARD M. HAYS Active Transport of Potassium and Other Alkali Metals by the Isolated Midgut of the Silkworm WILLIAM R. HARVEY AND KARLZERAHN Author Index-Subject Index
Volume 2 The Molecular Basis of Simple Diffusion within Biological Membranes W. R. Lien A N D W. D. STEIN The Transport of Water in Erythrocytes ROBERTE. FORSTER Ion-Translocation in Energy-Conserving Membrane Systems A N D M. MONTAL B. CHANCE Structure and Biosynthesis of the Membrane Adenosine Triphosphatase of Mitochondria ALEXANDER TZAGOLOFF Mitochondria1 Compartments: A Comparison of Two Models HENRYTEDESCHI Author Index-Subject Index
Volume 4 The Genetic Control of Membrane Transport CAROLYN W. SLAYMAN
xxiii
xxiv Enzymic Hydrolysis of Various Components in Biomembranes and Related Systems MAHENDRA KUMAR JAIN Regulation of Sugar Transport in Eukaryotic Cells HOWARD E. MORGAN AND CAROLF . WHITFIELD Secretory Events in Gastric Mucosa RICHARD P. DURBIN Author Index-Subject Index
Volume 5 Cation Transport in Bacteria: K+, Na+, and H+ FRANKLIN M. HAROLD AND KARLHEIN z ALTENDORF Pro and Contra Carrier Proteins: Sugar Transport via the Periplasmic GalactoseBinding Protein WINFRIED Boos Coupling and Energy Transfer in Active Amino Acid Transport ERICHHEINZ The Means of Distinguishing between Hydrogen Secretion and Bicarbonate Reabsorption: Theory and Applications to the Reptilian Bladder and Mammalian Kidney AND WILLIAM A. BRODSKY THEODORE P. SCHILB Sodium and Chloride Transport across Isolated Rabbit Ileum STANLEY G. SCHULTZ AND PETER F. CURRAN A Macromolecular Approach to Nerve Excitation ICHIJI TASAKIAND EMILIO CARBONE Subject Index
Volume 6 Role of Cholesterol in Biomembranes and Related Systems MAHENDRA KUMAR JAIN Ionic Activities in Cells A. A. LEVAND W. McD. ARMSTRONG Active Calcium Transport and Ca2+-Activated ATPase in Human Red Cells H. J. SCHATZMANN The Effect of Insulin on Glucose Transport in Muscle Cells TORBEN CLAUSEN
CONTENTS OF PREVIOUS VOLUMES
Recognition Sites for Material Transport and Information Transfer HALVOR N. CHRISTENSEN Subject Index
Volume 7 Ion Transport in Plant Cells E. A. C. MACROBBIE H+ Ion Transport and Energy Transduction in Chloroplasts RICHARD A. DILLEYAND ROBERTT. GIAQUINTA The Present State of the Carrier Hypothesis PAULG. LEFEVRE Ion Transport and Short-circuit Technique WARREN S. REHM Subject Index
Volume 8 Chemical and Physical Properties of Myelin Proteins M. A. MOSCARELLO The Distinction between Sequential and Simultaneous Models for Sodium and Potassium Transport P. J. GARRAHAN AND R. P. GARAY Soluble and Membrane ATPase of Mitochondria, Chloroplasts, and Bacteria: Molecular Structure, Enzymatic Properties, and Functions RIVKAPANETAND D. RAOSANADI Competition, Saturation, and InhibitionIonic Interactions Shown by Membrane Ionic Currents in Nerve, Muscle, and Bilayer Systems ROBERT J. FRENCH AND WILLIAM J. ADELMAN, JR. Properties of the Glucose Transport System in the Renal Brush Border Membrane R. KINNE Subject Index
Volume 9 The State of Water and Alkali Cations within the Intracellular Fluids: The Contribution of NMR Spectroscopy SHF'ORER AND MORDECHAI MORTIMER M. CIVAN
xxv
CONTENlS OF PREVIOUS VOLUMES
Electrostatic Potentials at Membrane-Solution Interfaces STUART MCLAUGHLIN A Thermodynamic Treatment of Active Sodium Transport S. ROYCAPLAN A N D ALVINEssrc Anaerobic Electron Transfer and Active Transport in Bacteria AND WIL N. KONINGS JOHANNES BOONSTRA Protein Kinases and Membrane Phosphorylation M. MARLENE HOSEYA N D MARIANO TAO Mechanism and Physiological Significance of Calcium Transport across Mammalian Mitochondria1 Membranes LEENAMELA Thyroidal Regulation of Active Sodium Transport F. ISMAIL-BEIGI Subject Index
Volume 10 Mechanochemical Properties of Membranes A N D R. M. HOCHMUTH E. A. EVANS Receptor-Mediated Protein Transport into Cells. Entry Mechanisms for Toxins, Hormones, Antibodies, Viruses, Lysosomal Hydrolases, Asialoglycoproteins, and Carrier Proteins DAVID M. NEVILLE, JR. A N D TA-MINCHANG The Regulation of Intracellular Calcium ERNESTO CARAFOLI AND MARTIN CROMFTON Calcium Transport and the Properties of a Calcium-Sensitive Potassium Channel in Red Cell Membranes VIRGILIO L. LEWA N D HUGOG. FERREIRA Proton-Dependent Solute Transport in Microorganisms A. A. EDDY Subject Index
Volume 11 Cell Surface Glycoproteins: Structure, Biosynthesis,and Biological Functions
The Cell Membrane-A Short Historical Perspective ASERROTHSTEIN The Structure and Biosynthesis of Membrane Glycoproteins JENNIFER STURGESS, MARIOMOSCARELLO, AND HARRYSCHACHTER Techniques for the Analysis of Membrane Glycoproteins R. L. JULIANO Glycoprotein Membrane Enzymes JOHNR. RIORDAN AND GORDON G. FORSTNER Membrane Glycoproteins of Enveloped Viruses RICHARD W. COMPANS AND MAURICE C. KEMP Erythrocyte Glycoproteins MICHAEL J. A. TANNER Biochemical Determinants of Cell Adhesion LLOYDA. CULP Proteolytic Modification of Cell Surface Macromolecules: Mode of Action in Stimulating Cell Growth KENNETH D. NOONAN Glycoprotein Antigens of Murine Lymphocytes MICHELLE LETARTE Subject Index
Volume 12 Carriers and Membrane Transport Proteins
Isolation of Integral Membrane Proteins and Criteria for Identifying Carrier Proteins MICHAEL J. A. TA N N ER The Carrier Mechanism S. B. HLADKY The Light-Driven Proton Pump of Halubacterium halubium: Mechanism and Function MICHAEL EISENBACH AND S. ROYCAPLAN Erythrocyte Anion Exchange and the Band 3 Protein: Transport Kinetics and Molecular Structure PHILIPA. KNAUF
XXVl
The Use of Fusion Methods for the Microinjection of Animal Cells R. G. KULKA A N D A. LOYTER Subject Index
Volume 13 Cellular Mechanisms of Renal Tubular Ion Transport
PART I: ION ACTIVITY AND ELEMENTAL COMPOSITION O F INTRAEPITHELIAL COMPARTMENTS Intracellular pH Regulation WALTERF. BORON Reversal of the pH,-Regulating System in a Snail Neuron R. C. THOMAS How to Make and Use Double-Barreled Ion-Selective Microelectrodes THOMAS ZUETHEN The Direct Measurement of K, CI, Na, and H Ions in Bullfrog Tubule Cells MAMORU FUJIMOTO, AND KUNlHlKO KOTERA, YUTAKA MATSUMURA Intracellular Potassium Activity Measurements in Single Proximal Tubules of Necturus Kidney TAKAHIRO KUBOTA, BRUCEBIAGI,AND GERHARD GIEBISCH Intracellular Ion Activity Measurements in Kidney Tubules RAJAN. KHURI Intracellular Chemical Activity of Potassium in Toad Urinary Bladder JOELDELONGA N D MORTIMER M. CIVAN Quantitative Determination of Electrolyte Concentrations in Epithelial Tissues by Electron Microprobe Analysis ROGERRICK,ADOLFDORGE, RICHARD BAUER,FRANZ BECK, JUNEMASON,CHRISTIANE ROLOFF, AND KLAUSTHURAU PART 11: PROPERTIES O F INTRAEPITHELIAL MEMBRANE BARRIERS IN THE KIDNEY
CONTENTS OF PREVIOUS VOLUMES
Hormonal Modulation of Epithelial Structure JAMESB. WADE Changes in Cell Membrane Surfaces Associated with Alterations of Transepithelial Ion Movement MICHAEL KASHGARIAN The Dimensions of Membrane Barriers in Transepithelial Flow Pathways AND LARRY W. WELLING DANJ. WELLING Electrical Analysis of Intraepithelial Barriers AND EMILEL. BOULPAEP HENRYSACKIN Membrane Selectivity and Ion Activities of Mammalian Tight Epithelia SIMONA. LEWIS,NANCYK. WILLS, A N D DOUGLAS C. EATON Ion Conductances and Electrochemical Potential Differences across Membranes of Gallbladder Epithelium Luis REUSS A Kinetic Model for Ion Fluxes in the Isolated Perfused Tubule BRUCEBIAGI,ERNESTO GONZALEZ, A N D GERHARD GIEBISCH The Effects of Voltage Clamping on Ion Transport Pathways in Tight Epithelia L. FINN AND PAULA ROCENES ARTHUR Tubular Permeability to Buffer Components as a Determinant of Net H Ion Fluxes G. MALNIC,v. L. COSTA SILVA, s. s. CAMPIGLIA, M. DE MELLOAIRES,A N D G . GlEBlSCH Ionic Conductance of the Cell Membranes and Shunts of Necturus Proximal Tubule GENJIRO KlMURA A N D KENNETH R. SPRING Luminal Sodium Phosphate Cotransport as the Site of Regulation for Tubular Phosphate Reabsorption: Studies with Isolated Membrane Vesicles HEINIMURER,REINHARD STOLL, CARLAEVERS,ROLFKINNE, JEAN-PHILIPPE BONJOUR, AND HERBERT FLEISCH The Mechanism of Coupling between Glucose Transport and Electrical Potential in the Proximal Tubule: A Study of Potential-
xxvii
CONTENTS OF PREVIOUS VOLUMES
Dependent Phlorizin Binding to Isolated Renal Microvillus Membranes PETERS. ARONSON Electrogenic and Electroneutral N a Gradient-Dependent Transport Systems in the Renal Brush Border Membrane Vesicle SACKTOR BERTRAM
Volume 14 Carriers and Membrane Transpo~ Proteins
Interface between Two Immiscible Liquids as a Tool for Studying Membrane Enzyme Systems L . I. BOGUSLAVSKY Criteria for the Reconstitution of Ion PART 111: INTRAMEMBRANE Transport Systems CARRIERS AND ENZYMES IN ADILE. SHAMOO AND TRANSEPITHELIAL TRANSPORT WILLIAM F. T ~ V O L The Role of Lipids in the Functioning of a Membrane Protein: The Sarcoplasmic ReSodium Cotransport Systems in the Proxiticulum Calcium Pump mal Tubule: Current Developments J. P. BENNETT,K. A. MCGILL,A N D R. KINNE,M. BARAC,A N D H. MURER G. B. WARREN ATPases and Salt Transport in the Kidney The Asymmetry of the Hexose Transfer Tubule System in the Human Red Cell Membrane MARGARITA PEREZ-GONZALEZ DE LA W. F. WIDDAS MANNA, FULGENCIO PROVERBIO, AND Permeation of Nucleosides, Nucleic Acid GUILLERMO WHITEMBURY Bases, and Nucleotides in Animal Cells Further Studies on the Potential Role of an PETERG. w. PLAGEMANN AND Anion-Stimulated Mg-ATPase in Rat ProxROBERTM. WOHLHUETER imal Tubule Proton Transport Transmembrane Transport of Small E . KINNE-SAFFRAN A N D R. KINNE Peptides Renal Na+- K+-ATPase: Localization and D. M. MATTHEWS A N D J. W. PAYNE Quantitation by Means of Its K+-DepenCharacteristics of Epithelial Transport in dent Phosphatase Activity Insect Malpighian Tubules REINIER BEEUWKES 111 A N D S. H. P. MADDRELL SEYMOUR ROSEN Subject Index Relationship between Localization of N+K+-ATPase, Cellular Fine Structure, and Reabsorptive and Secretory Electrolyte Volume 15 Transport STEPHEN A. ERNST, Molecular Mechanisms of Photoreceptor CLARA v. RIDDLE, AND Transduction JR. KARLJ. KARNAKY, Relevance of the Distribution of Na+ Pump PART I: T H E ROD PHYSIOLOGICAL RESPONSE Sites to Models of Fluid Transport across Epithelia The Photocurrent and Dark Current of JOHNW. MILLSA N D Retinal Rods DONALD R. DIBONA G. MATTHEWS A N D D. A. BAYLOR Cyclic AMP in Regulation of Renal TransSpread of Excitation and Background Adport: Some Basic Unsolved Questions aptation in the Rod Outer Segment THOMAS P. DOUSA K.-W. Y A U , T. D. LAMB,AND Distribution of Adenylate Cyclase Activity P. A. MCNAUGHTON in the Nephron Ionic Studies of Vertebrate Rods F. MOREL,D. CHABARDES, AND W. GEOFFREY OWENAND M. IMBERT-TEBOUI, VINCENT TORRE Subject Index
xxviii Photoreceptor Coupling: Its Mechanism and Consequences GEOFFREY H . GOLD PART 11: T H E CYCLIC NUCLFOTIDE ENZYMATIC CASCADE AND CALCIUM ION First Stage of Amplification in the CyclicNucleotide Cascade of Vision LUBERTSTRYER, JAMESB. HURLEY, A N D BERNARD K.-K. FUNG Rod Guanylate Cyclase Located in Axonemes FLEISCHMAN DARRELL Light Control of Cyclic-Nucleotide Concentration in the Retina THOMAS G. EBREY,PAULKILBRIDE, JAMESB. HURLEY, ROGERCALHOON, A N D MOTOYUKI TSUDA Cyclic-GMP Phosphodiesterase and Calmodulin in Early-Onset Inherited Retinal Degenerations Y . P. LIU, G. J. CHADER, G. AGUIRRE, R. T. FLETCHER, R. SANTOS-ANDERSON, AND M. T'SO Control of Rod Disk Membrane Phosphodiesterase and a Model for Visual Transduction A N D E. N. PUGH,JR. P. A. LIEBMAN Interactions of Rod Cell Proteins with the Disk Membrane: Influence of Light, Ionic Strength, and Nucleotides HERMANN KUHN Biochemical Pathways Regulating Transduction in Frog Photoreceptor Membranes M. DERICBOWNDS The Use of Incubated Retinas in Investigating the Effects of Calcium and Other Ions on Cyclic-Nucleotide Levels in Photoreceptors ADOLPHI. COHEN Cyclic AMP: Enrichment in Retinal Cones DEBORA B. FARBER Cyclic-Nucleotide Metabolism in Vertebrate Photoreceptors: A Remarkable Analogy and an Unraveling Enigma
CONTENTS OF PREVIOUS VOLUMES
M. W. BITENSKY, G. L. WHEELER, AND A. YAMAZAKI, M. M. RASENICK, P. J. STEIN Guanosine Nucleotide Metabolism in the Bovine Rod Outer Segment: Distribution of Enzymes and a Role of GTP SHICHI HITOSHI Calcium Tracer Exchange in the Rods of Excised Retinas ETE 2. SZUTS The Regulation of Calcium in the Intact Retinal Rod: A Study of Light-Induced Calcium Release by the Outer Segment GEOFFREY H . GOLDA N D JUANI. KORENBROT Modulation of Sodium Conductance in Photoreceptor Membranes by Calcium Ions and cGMP ROBERTT. SORBI PART 111: CALCIUM, CYCLIC NUCLEOTIDES, AND T H E MEMBRANE POTENTIAL Calcium and the Mechanism of Light Adaptation in Rods BRUCEL. BASTIAN AND GORDON L . FAIN Effects of Cyclic Nucleotides and Calcium Ions on Bufo Rods JOELE. BROWN AND GERALDINE WALOGA The Relation between CaZ+and Cyclic GMP in Rod Photoreceptors STUART A. LIFTONA N D JOHNE. DOWLING Limits on the Role of Rhodopsin and cGMP in the Functioning of the Vertebrate Photoreceptor SANFORD E . OSTROY, EDWARD P. MEYERTHOLEN, PETERJ. STEIN, ROBERTA A. SVOBODA, A N D MEEGAN J. WILSON [Ca2+IiModulation of Membrane Sodium Conductance in Rod Outer Segments I1 A N D BURKSOAKLEY LAWRENCE H. PINTO
xxix
CONTENTS OF PREVIOUS VOLUMES
Cyclic-GMP-Induced Depolarization and Increased Response Latency of Rods: Antagonism by Light WILLIAM H. MILLER AND GRANTD. NICOI,
PART IV: A N EDITORIAL OVERVIEW CaZ+and cGMP WILLIAM H. MILLER Index
This Page Intentionally Left Blank
Historical Introduction
It is one of the beautiful ironies of science (and of inductive logic itself) that much of the early evidence educed to support the notion that exergonic metabolic reactions might directly drive electric current through biological membranes can now be seen to have been inspired misinterpretation. The notion orginated with E. J. Lund, who observed that the longitudinal electric polarization of plant roots and stems (1928a) and the transverse polarization of frog skin (1928b) were closely dependent upon tissue respiration, and, in the case of roots (Lund and Kenyon, 1927), quantitatively paralleled the localized ability of cells to reduce methylene blue. Lund hypothesized that the respiratory apparatus can produce redox potential differences at cell surfaces; in ionic solutions, that also implies production of ionic currents. Though he could not have realized it, Lund’s hypothesis was doomed as an explanation of his own data, since he was comparing events which occur in separate and discrete membranes: tissue electrogenesis in the cellular plasma membranes and respiration in the mitochondria. That it was correct in any sense (as applied to mitochondria themselves; Robertson, 1960) is very remarkable, since little of the modern understanding of either bioenergetics or transport was then known. The cytochrome respiratory pigments had only recently been discovered (Keilin, 1925); ATP was just in the process of being discovered (Lohmann, 1929); the concept of energy-rich phosphate bonds was still in the future (Lohmann and Meyerhof, 1934; Lipmann, 1941); and confinement of the respiratory apparatus to the mitochondria was to be demonstrated only 20 years later (Hogeboom et al., 1948). Active transport (i.e., pumping) of ions had been hinted at earlier (Reid, 1892), but had not yet become a substantive scientific idea (Keys, 1931; Krogh, 1937). Furthermore, the preoccupation of thought on bioelectric phenomena, following Bernstein (1912) and Osterhout (1929), was with passive processes, as is now attested, e.g., by the contents of the first Cold Spring Harbor symposium (1933). But even for passive processes, major conceptual problems were 10 -15 years away from solution (see, e.g., Hill and Osterhout, 1938a,b; Cole and Curtis, 1939; Boyle and Conway, 1941; Goldman, 1943). Over the period of two decades during which these major developments were occurring, the notion of redox-driven charge transport was also xxxi
xxxii
HISTORICAL INTRODUCTION
elaborated, but still in connection with the behavior of plasma membranes in eukaryotic cells. LundegArdh (1939) and later Robertson and Wilkins ( 1948) proposed it as a mechanism to account for salt-stimulated respiration in plant roots. Conway and his collaborators extended the notion to acid production by yeast (Conway and Brady, 1950) and to HCl secretion by gastric mucosa (Conway el al., 1950); and the latter suggestion was adopted either explicitly or implicitly by other laboratories studying gastric secretion (Rehm, 1949; Davies and Ogston, 1950). Simultaneously, a separate line of evidence concerning metabolically driven charge transport began to emerge. For historical reasons, this line did not link charge flow to redox phenomena; it was, therefore, more appropriate in connection with the plasma membranes of eukaryotic cells. Lorente de N6 (1947), studying postanoxic overshoot in myelinated nerves, concluded that “metabolism directly establishes and maintains the membrane potential, to a certain extent independently of the external as well as of the internal concentration of ions.” Ussing and his collaborators-who had combined the earlier technique (Francis, 1933) of measuring short-circuit current through frog skin with isotopic tracer techniques for obtaining ion unidirectional fluxes (Hahn et a l . , 1939; Heppel, 1940)found that the short-circuit current was essentially identical with the sodium inward flux, and suggested that active transport of sodium was indeed the source of electric current (Ussing and Zerahn, 1951). A by-product of these studies was an estimation of the apparent EMF of the sodium-transport mechanism, equal to about 100 mV. A complementary juncture between chemical measurements and electrical measurements was also made by Rehm (1949, who demonstrated that imposed electric currents could either accelerate or retard acid secretion by gastric mucosa, depending on the direction of net current flow. In retrospect, it is possible to see that difficulties in formulating the concept of electrogenic ion pumps during the 1940s lodged heavily in the fact that no generally accepted, physically convincing picture had been presented even of passive ionic processes in biological systems. Despite the fundamental contributions of Bernstein, Osterhout, and Conway (already alluded to above), it was not until the early 1950s and the key publications on the nerve action potential (Hodgkin and Katz, 1949; Hodgkin and Huxley , 1952a,b) that the unification of ion-diffusion processes, membrane potentials, and electric circuits was generally perceived. But even this crucial development temporarily impeded the idea that an active transport process could actually displace net charges across cell membranes. It did so by initiating nearly a decade of attempts to account for all bioelectric phenomena on the basis of ion diffusion. Hodgkin and Keynes (1955) found that metabolic inhibitors could halt active sodium
HISTORICAL INTRODUCTION
xxxiii
extrusion from squid nerve with essentially no effect on the resting membrane potential. And while these authors were wary of insensitivity in their measurements, the observation was widely interpreted to mean that the Na+ pump in animal cell membranes functions electroneutrally (see, e.g., Hodgkin, 1957). Posttetanic hyperpolarization in myelinated nerve was supposed to result from K+ depletion in the extracellular space (Ritchie and Straub, 1957); the transepithelial potential difference in frog skin was reinterpreted as the series addition of diffusion potentials in a composite membrane (Koefed-Johnsen and Ussing, 1958); and pump activity in the charophyte algae was missed because of abnornal membrane leakiness, which resulted from deliberate calcium leaching (Hope and Walker, 1961; Dainty, 1962). It was not, therefore, until the early 1960s that the existence of chargetransporting pumps, per se, began to be accepted as legitimate. By that time Rehm and his collaborators had built up an impressive case for electrogenicity of separate chloride and proton transport systems in gastric mucosa (reviewed in Rehm, 1966) based on the parallel behavior of electrical parameters and chemically measured fluxes during selective inhibition of the two systems.' Furthermore, Post and Jolly (1957) had shown convincingly that the ion-exchange ratio for the sodium pump in human red blood cells was 3 Na+ extruded : 2 K+ taken up; this result implied either that the pump must be electrogenic or that it must transport one unidentified ion. The observation which proved pivotal to the field, however, was Kernan's finding (1962) that when Na+-loaded frog sartorius muscles-which had been depolarized and K+-depleted by prolonged soaking in cold K+-free Ringer's solution-were restored to K+-containing Ringer's, they hyperpolarized beyond the potassium equilibrium potential. This experiment was soon repeated and refined (Mullins and Awad, 1965; Frumento, 1965a; Cross et a l . , 1965) even to a demonstration of hyperpolarizations that were large enough to make localized ion depletion impossible as an explanation (Adrian and Slayman, 1966). During the same period, concurring observations were made on a variety of other systems as well. In particular, Frazier and Leaf (1963) found that the electric potential difference across the serosal (actively transporting) surface of the toad bladder epithelium could not be described by diffusion of any common ions, and concluded that the sodium pump itself 'It seems ironic, even retributive, that proton transport carried out by isolated vesicles of gastric mucosa should now prove tightly coupled to the counterflow of potassium so that it appears completely electroneutral (Sachs et a l . , this volume). This stark contrast between earlier results obtained by electrophysiological methods and recent ones obtained by biochemically oriented methods remains to be resolved.
XXXIV
HISTORICAL INTRODUCTION
must be driving current through the membrane. Harvey and Nedergaard (1964) identified the potent Naf-independent, K+-extruding system in Cecropia midgut. Page and Storm (1965) repeated the Kernan experiments, with similar but more dramatic results, on cat papillary muscle. Connelly (1959), re-examining the Ritchie- Straub experiments on myelinated nerves, found a steady hyperpolarization during tetanic stimulation: that is, even during net potassium efflux, so that K+ depletion from the extracellular space could not account for the hyperpolarization. And Kerkut and Thomas (1965) found 20- 30 mV of hyperpolarization associated with the extrusion of injected sodium from snail neurons. Finally, in nonanimal systems, the fungus Neurosporu was reported to have steady membrane potentials that could be rapidly abolished by respiratory inhibitors and were-at -200 mV-far in excess of any plausible ion diffusion potentials (Slayman, 1965). Initial efforts were also being made to incorporate metabolically driven ionic currents into the three major formalisms for transport: constant-field theory (Briggs, 1962; Frumento, 1965b), carrier theory (Finkelstein, 1964), and linear-coupling theory (Hoshiko and Lindley, 1967). These theoretical efforts gave rise to a thermodynamically rigorous definition of electrogenic pumping (see, e.g., Hoshiko and Lindley, 1967) that placed a variety of restrictions on the accepted operational definition then in use: active transport which directly produces either an increment of membrane current under short-circuit conditions or an increment of membrane potential under open-circuit conditions. (Unhappily, in recent years the term “electrogenic” has been thoroughly corrupted, passing into common usage no longer tied to a measurement, such as of electric current or potential, but rather to an inference, such as the transmembrane movement of a single charge.) In the larger scientific context, however, this essential evolution of data and ideas, plus their obvious sequelae, was gradually eclipsed by the realization-stimulated by Mitchell’s provocative Chemiosmotic Hypothesis (1961, 1963, 1968)-that a very large number of formerly disparate phenomena could be unified if certain key reactions, mainly in the realm of bioenergetics, were looked upon as special kinds of charge-transporting pumps. In most cases, protons rather than sodium ions were supposed to be involved. Though initially this idea met with disinterest on the part of physiologists and with total disbelief on the part of biochemists, a growing cascade of experiments in the decade 1964- 1974 made it irresistable. In retrospect, about half a dozen results from the period can be counted as most significant. First was Mitchell’s own finding that classical energy uncoupling agents, such as 2,4-dinitrophenol, accelerate proton flux through mitochondria1 membranes (Mitchell, 1963; Mitchell and Moyle, 1967). Then came Jagendorf and Uribe’s demonstration (1966) that iso-
HISTORICAL INTRODUCTION
xxxv
lated chloroplast thylakoids could synthesize ATP in the absence of both light and oxygen if they were prepared in acidic media and then were shifted to basic media. Cockrell et al. (1967) showed that in the presence of an outward potassium gradient, mitochondria doped with the ionophore valinomycin (plus the respiratory inhibitor rotenone) could also synthesize ATP, and Glynn (1967) interpreted this observation to mean that a mitochondria1 membrane potential, set up by K+ diffusion, was actually driving ATP formation. At about the same time, Garrahan and Glynn (1967) proved that at least one bonafide active transport system, the sodium pump in resealed erythrocyte ghosts, could indeed be made to run backwards, using reversed gradients of Na+ and K+, to synthesize ATP. In the late 1960s Witt and his collaborators (Junge and Witt, 1968; Shliephake et al., 1968; Emrich et al., 1969) identified formation of an electric field through the thylakoid membrane as the source of a lightinduced absorption shqt in chloroplasts. They also showed, for flashes, that the shift could be rapidly reversed, either by allowing phosphorylation to occur or by introducing ionophores to make the thylakoid membrane leaky. Most residual doubts about the physical reality of metabolically generated electric fields were removed by the crucial studies with synthetic lipid-soluble ions, which were first systematically explored by Skulachev and his co-workers (Grinius et al., 1970; Bakeevaet al., 1970). Anions, such as phenyl dicarbaundecaborane or picrate, proved to be extruded by intact mitochrondria and concentrated by submitochondrial vesicles that have everted membranes; but cations, such as dibenzyl dimethylammonium, were distributed in the opposite fashion. All these movements required energy that could be extracted either from ATP or from substrate oxidation, and because the ions had no previous biological (evolutionary) significance, their transport along a metabolically generated electric gradient provided the simplest interpretation of all observations. Finally, discovery of the remarkable proton-pumping properties of the photopigment in purple membrane from halophilic bacteria (Oesterhelt and Stoeckenius, 1973) made it possible to bring most of the above types of experiments to focus on a single biological preparation. Luckily, also, the purple membrane proved simple to remove from bacteria and to insert in functional form into artificial lipid vesicles. It was, therefore, only a short time until both proton transport and electrogenesis by this protein had been completely reconstituted in artificial systems (Kayushin and Skulachev, 1974; Racker and Hinkle, 1974), where they could also be used to drive ATP synthesis through a completely unrelated ATPase (Racker and Stoeckenius, 1974). The period since 1974 has been one in which the important ideas which
xxxvi
HISTORICAL INTRODUCTION
sprang from the previous 15 years have been refined, elaborated, and extended to the broadest possible range of biological circumstances. Andas is elaborated in Part V-the same period has seen the mounting of serious attacks on the deeper molecular problem of just how the energy of photons, electrons, or covalent bonds can be translated into a transmembrane flow of ionic charge. This, of course, broaches what has long been recognized as the central problem of active transport. While the various facts and hypotheses which can be set down at present on the central problem are very exciting, few scientists who have followed the history of transport physiology and bioenergetics over the past 50 years will be surprised if the ultimate answer to this problem should come from a direction which is now only barely perceived on the horizon.
CLIFFORD L. SLAYMAN REFERENCES Adrian, R. H., and Slayman, C. L. (1966). J. Physiol. 184, 970-1014. Bakeeva, L.E., Grinius, L. L., Jasaitis, A. A., Kuliene, V. V., Levitsky, D. O., Liberman, E. A., Severina, I. I., and Skulachev, V. P. (1970). Biochim. Biophys. Acta 216, 13-21. Bernstein, J. (1912). Elektrobiologie. Braunschweig, F., Vieweg. 215 pp. Boyle, P. J., and Conway, E . J. (1941). J. Physiol. 100, 1-63. Briggs, G. E . (1962). Proc. R. Soc. B 156, 573-577. Cockrell, R. S., Harris, E. J., and Pressman, B. C. (1967). Nature (London) 215, 1487- 1488. Cold Spring Harbor. (1933). Cold Spring Harbor Symp. Quant. Biol. 1 . Cole, K. S . , and Curtis, H. J. (1939). J. Gen. Physiol. 22, 649-670. Connelly, C. M. (1959). Rev. Mod. Phys. 31, 475-484. Conway, E. J. and Brady, T . G. (1950). Biochem. J . 47, 347-369. Conway, E. J., Brady, T. G., and Carton, E. (1950). Biochem. J . 47, 369-374. Cross, S. B., Keynes, R. D., and RybovA R. (1965). J. Physiol. 181,865-880. Dainty, J. (1962). Annu. Rev. Plant Physiol. 13,379-402. Davies, R. E., and Ogston, A. G. (1950). Biochem. J . 46, 324-333. Emrich, H. M., Junge, W., and Witt, H. T. (1969). Z. Naturforsch. 24B, 1144-1146. Finkelstein, A. (1964). Biophys. J. 4, 421-440. Francis, W.L. (1933). Nature (London). 131, 805. Frazier, H. S., and Leaf, A. (1963). J. Gen. Physiol. 46,491-503. Frurnento, A. S. (1965a). Science 147, 1442- 1443. Frumento, A. S. (1965b). J . Theoret. Biol. 9, 253-262. Garrahan, P. J. and Glynn, I. M. (1967). J . Physiol. 192,237-256. Glynn, I. M. (1967). Nature (London). 216, 16- 17. Goldrnan, D. E. (1943). J . Gen. Physiol. 27, 37-60. Grinius, L.L., Jasaitis, A. A., Kadziauskas, Yu. P., Liberrnan, E. A., Skulachev, V. P., Topali, V. P., Tsofina, L. M., and Vladimirova, M. A. (1970). Biochim. Biophys. Acta. 216, 1-12. Hahn, L . A., Hevesy, G. C., and Rebbe, 0. H. (1939). Biochem. J. 33, 1549-1558. Harvey, W.R. and Nedergaard, S. (1964). Proc. Nail. Acad. Sci. U . S . A . 51, 757-765. Heppel, L. A. (1940). A m . J . Physiol. 128,449-454.
HISTORICAL INTRODUCTION
xxxvii
Hill, S. E., and Osterhout, W. J. V. (1938a). J . Gen. Physiol. 21, 541-556. Hill, S. E., and Osterhout, W. J . V. (1938b). Proc. Natl. Acad. Sci. U . S . A . 24, 312-315. Hodgkin, A. L. (1957). Proc. R . SOC. London, Ser. B . 148, 1-37. Hodgkin, A. L., and Huxley, A. F. (1952a). J. Physiol. 116,449-472. Hodgkin, A. L., and Huxley, A. F. (1952b). J . Physiol. 117, 500-544. Hodgkin, A. L., and Katz, B. (1949). J. Physiol. 108, 37-77. Hodgkin, A. L., and Keynes, R. D. (1955). J . Physiol. 128,28-60. Hogeboom, G.H., Schneider, W. C., and Palade, G. E. (1948). J. Biol. Chem. 172,619-635. Hope, A. B., and Walker, N. A. (1961). Aust. J. B i d . Sci. 14, 26-44. Hoshiko, T.,and Lindley, B. D. (1967). J. Gen. Physiol. 50, 729-758. Jagendorf, A. T., and Uribe, E. (1966). Proc. Natl. Acad. Sci. U . S . A . 55, 170-177. Junge, W.,and Witt, H . T. (1968). Z. Naturforsch. 23B, 244-254. Kayushin, L. P., and Skulachev, V. P. (1974). FEBS Lett. 39, 39-42, Keilin, D.(1925). Proc. Roy. Soc. London, Ser. B . 98, 312-339. Kerkut, G.A., and Thomas, R. C. (1965). Comp. Biochem. Physiol. 14, 167- 183. Kernan, R. P. (1962). Nature (London) 193, 986-987. Keys, A. B. (1931). 2. Vgl. Physiol. 15,364-388. Koefoed-Johnsen, V.,and Ussing, H. H. (1958). Acta Physiol. Scand. 42, 298-308. Krogh, A. (1937). Skand. Arch. Physiol. 76, 60-74. Lipmann, F. (1941). Adv. Enzymol. 1, 99-162. Lohmann, K. (1929). Naturwissenschafren 17,624-625. Lohmann, K., and Meyerhof, 0. (1934). Biochem. Z . 273, 60-72. Lorente de N6, R. (1947). Stud. Rockefeller Inst. 131, 1- 113. Lund, E. J. (1928a). J . Exp. Zool. 51, 265-290. Lund, E. J. (1928b). J. Exp. Zool. 51, 327-337. Lund, E. J., and Kenyon, W. A. (1927). J. Exp. Zool. 48, 333-357. Lundegardh, H . (1939). Nature (London) 143,203-204. Mitchell, P. (1961). Nature (London) 191, 144- 148. Mitchell, P. (1963). Biochem. SOC. Symp. 22, 142-168. Mitchell, P. (1968). “Chemiosmotic Coupling and Energy Transduction.” Glynn Research, Bodmin. Mitchell, P., and Moyle, J. (1967). Biochem. J. 104, 588-600. Mullins, L. J., and Awad, M. 2. (1965).J. Gen. Physiol. 48, 761-775. Oesterhelt, D.,and Stoeckenius, W. (1973). Proc. Natl. Acad. Sci. U . S . A . 70, 2853-2857. Osterhout, W. J . V. (1929). Bull. Natl. Res. Counc. ( U S . ) 69, 170-228. Page, E.,and Storm, S. R. (1965). J. Gen. Physiol. 48,957-972. Post, R. L., and Jolly, P. C. (1957). Biochim. Biophys. Acta. 25, 108-128. Racker, E.,and Hinkle, P. C. (1974). J. Membr. Biol. 17, 181-188. Racker, E., and Stoeckenius, W. (1974). J. Biol. Chem. 249, 662-663. Rehm, W. S. (1945). A m . J . Physiol. 144, 115- 125. Rehm, W. S. (1949). A m . J . Physiol. 159,586. Rehm, W.S. (1966). Ann. N . Y. Acad. Sci. 137,591-605 Reid, E. W. (1892). Brit. Med. J . 1 , 1133-1134. Ritchie, J. M., and Straub, R. W. (1957). J. Physiol. 136,80-97. Robertson, R. N. (1960). Biol. Rev. 35,231-264. Robertson, R. N., and Wilkins, M. J. (1948). Nature (London) 161, 101. Schliephake, W., Junge, W., and Witt, H. T. (1968). Z. Naturforsch 23B, 1571-1578. Slayman, C. L. (1965). J. Gen. Physiol. 49, 93- 116. Ussing, H . H., and Zerahn, K. (1951). Acta Physiol. Scand. 23, 100- 127.
This Page Intentionally Left Blank
Part I
Demonstration of Pump Electrogenicity in Eukaryotic Cells
This Page Intentionally Left Blank
CURRENT TOPICS IN MEMBRANES AND TRANSPORT, VOLUME 16
Chapter I Electrophysiology of the Sodium Pump in a Snail Neuron R . C . THOMAS Department of Physiology Bristol University Bristol, England
I. Introduction ........................................................................................ 11. Methods ............................................................................................ 111. Results ............................................................................................... A. Sodium Injection and the Membrane Potential ...................................... B. Measurement of Current and Charge Generated by the Pump ................... C. Iontophoretic Transport Number for Sodium Injections .......................... D. Effect of Increasing Em on the Pump Current ........................................ E. Comparison of the Sodium Pump with the pHi-Regulating System ............. 1V. Discussion ........................................................................................... A. Sources of Error ............................................................................. B. Coupling Ratio of the Sodium Pump ................................................... References ..........................................................................................
1.
3 5 6 6
7 9 12 13 14 14 15 16
INTRODUCTION
The sodium (Na) pump is surely the grandfather of all electrogenic pumps, at least in the sense that it was the first active transport system to be studied in detail. As originally proposed in 1940, it was t o be a mechanism in the cell membrane constantly excreting Na. If only Na ions were extruded, as shown in Fig. lA, the pump would generate across the cell membrane a current proportional to its rate of activity. Thus a Na-only pump would inevitably be electrogenic and would hyperpolarize the cell. (It would hyperpolarize the cell, or make the internal potential more negative, because it would be the current generator in a circuit such as that shown in 3
Copyright 0 1982 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-153316-6
4
R. C. THOMAS
A
+++/q.B
C
,
K+
FIG. 1. Sodium pump mechanisms. (A) Pump transporting only Na+ ions. (B) Na+-K+ exchange pump. (C) Electrical circuit with Na pump as the current generator, the voltage across the membrane depending on the membrane resistance R,.
Fig. 1C. The other element in the circuit would be the membrane resistance. Passage of positive ions from outside to inside through this resistance would increase internal negativity.) When the Na pump was investigated in excitable cells, particularly by Hodgkin and Keynes (1955) with cephalopod giant axons, a coupling between Na efflux and potassium (K) uptake was found. The size of the two fluxes was similar, although not the same. There was no significant effect on the membrane potential of inhibiting the pump. This led to a widely held belief that the pump was neutral, and that the efflux of Na+ ions was tightly coupled to the influx of an equal number of K + ions, as in Fig. 1B. It was believed, in other words, that the Na pump was really a Na-K pump with a coupling ratio of 1 Nal 1 K. My own interest in this field began when I injected sodium acetate (NaAc) into a snail neuron by letting it leak out of the end of a lowresistance microelectrode. I had been injecting a wide variety of K salts in an investigation of the ionic mechanism of synaptic inhibition and wanted to try a Na salt for a change. T o my surprise, in a few minutes the membrane hyperpolarized by over 20 mV, something I had never seen with K injections. A rapid search through the available literature showed that Connelly (1959) and Kernan (1962) had already found that a stimulated Na pump could be electrogenic in nerve axons and in frog muscle cells. I was then able to show that the electrogenic property of the Na pump was the most likely explanation of the Na-induced hyperpolarization in snail neurons (Thomas, 1964). In these experiments, I had no control over the rate or duration of Na injection, except that it could be stopped by withdrawing the electrode from the cell. When I returned to the study of the electrogenic Na pump in 1966, I therefore set out to inject Na iontophoretically and to attempt a quantitative investigation. In this chapter I will describe some of the results I obtained then, and more recently, and discuss the early results with the benefit of hindsight. The original results were first described some time ago (Thomas, 1969).
5
1. ELECTROGENIC SODIUM PUMP
II.
METHODS
Snail circumesophageal ganglia were mounted in an experimental bath after the removal of connective tissue over the visceral and right pallial ganglia. The preparation was covered with saline, and cells were exposed by tearing the inner connective tissue. The normal saline contained 80 mM NaCl, 4 mM KCl, 7 mM CaCl,, 5 mMMgCl,, Tris buffer, pH 8, 5 mM, but more recently I have worked with HEPES- or bicarbonate-buffered salines at pH 7.5. Most experiments were done on the largest cell at the rear of the right pallial ganglion. It was penetrated with three or four conventional borosilicate glass microelectrodes, filled with K or Na acetate or chloride, and with resistances of 20-30 Ma. The electrical arrangement is shown diagrammatically in Fig. 2. Once a cell was successfully penetrated with three microelectrodes, one for recording Em and two for current passing, salts were injected by interbarrel iontophoresis. Current was passed between the two electrodes, shown as 2 and 3 in Fig. 2, from a current source which was isolated from ground. The injection current was measured as the voltage drop across a 1.O-Ma resistor. In some experiments the average membrane potential was controlled or “clamped” by a feedback amplifier whose output was fed back to a fourth intracellular mircoelectrode (4 in Fig. 2). This clamp current was measured by an operational amplifier connected as a current-voltage transducer in the path from the bath to ground. The injection and clamp current measurement systems were carefully checked by passing the same current through both in series. Further details of the methods are given in Thomas (1969).
CURRENT SOURCE
FIG. 2. Experimental arrangement, with four microelectrodes in a spherical snail neuron and a reference agar-saline electrode in the bath.
6
R.
C. THOMAS
Solution changes were made by running 5-10 bath volumes of saline through the bath. Between changes the solution in the bath was stagnant. In later experiments with Na+-sensitive microelectrodes it was necessary to have continuously flowing solutions to avoid solution change artifacts. 111.
RESULTS
A. Sodium Injection and the Membrane Potential
Raising intracellular Na by the injection of a Na salt could increase Emin several ways: (1) by stimulating a Na pump which itself directly generates current, (2) by stimulating a neutral pump which lowers external K and thus increases EK,(3) by decreasing the depolarizing effect of the Na gradient by making ENacloser to Em, (4) by decreasing the membrane permeability to Na, and ( 5 ) by increasing the membrane permeability to K. Experiments such as the one illustrated in Fig. 3 strongly favor mechanisms involving the Na pump, mechanism (1) or (2) above, but do not rule out mechanism ( 5 ) . This experiment was done before I had perfected the isolation from ground of the current source. It shows the effect, on both Emand the membrane response to a hyperpolarizing current pulse, of five Na injections and one K injection. The first two Na injections caused hyperpolarizations of about 15 and 20 mV, respectively, and also reduced the membrane resistance. Both Em and membrane resistance recovered within 10-15 minutes. The single K injection had no effect. Then ouabain was applied, causing a small decrease in Em and membrane resistance. After ouabain, injections of NaAc caused only small increases in E,,,. The first postouabain Na injection had little effect on the membrane resistance, but the other two caused significant decreases. Thus ouabain dramatically changed the cell’s response to NaAc injection, showing that the large hyperpolarizations were normally generated by the Na pump. The experiment shown in Fig. 4 confirms this and also rules out mechanism (2). It was done on a cell which proved unusually sensitive to Na injection. During the first injection Em increased by over 25 mV and then returned to the preinjection level within about 15 minutes. The normal Ringer’s solution was then replaced with one containing no K, and a second injection of NaAc made. This injection caused Emto increase by only about 5 mV. When external K was replaced 4 minutes later, there was a further 25-mV transient increase in Em.After a period in K-free Ringer’s solution without NaAc injection (except for that leaking out of the microelectrode) there was only a brief 15-mV hyperpolarization when K was replaced. The
7
1. ELECTROGENIC SODIUM PUMP
KAc
INJECTION CURRENT
U 40
5 E
E
w
50
n
uu
u
-
6070
-
80
L
15 min
FIG. 3 . Pen recording of experiment showing the effect on the snail neuron membrane potential of injecting NaAc or KAc. Spontaneously occurring action potentials were reduced to only a few millivolts (baseline blur) by the slow response of the pen recorder. Ouabain (10 p M ) was added as indicated. Hyperpolarizing current pulses of 2 nA were passed for 5 seconds every 30 seconds except during the injections.
removal of external K itself had very little effect on Em, eliminating any mechanism involving a pump-induced lowering of external K as a n explanation for the large hyperpolarization seen on Na injection. On the other hand, since it is well known that external K removal inhibits the pump, this experiment provides additional evidence for the pump’s role in generating the hyperpolarizations. The Na pump must therefore be itself electrogenic, directly generating a current across the cell membrane. B. Measurement of Current and Charge Generated by the Pump The measurement of pump current is not easy. The current increases during, and decreases after, the injection. If the membrane resistance remained constant, the voltage could be simply converted to current by Ohm’s law. Unfortunately, as shown in Fig. 3, the membrane resistance is K-free
m INaAc
K- f r e e
m
10 min
FIG.4. The effect of two injections of NaAc (40 nA for 1 minute each) and two exposures to K-free Ringer’s solution on the Em of a spontaneously active snail neuron.
a
R. C. THOMAS
lowered by the Na injection, the increase in Em,or both. Indeed, simply hyperpolarizing the cell (without injecting Na) causes a decrease in resistance; and large Na injections can cause an increase in Em and a fall in membrane resistance even when the Na pump is blocked. For these reasons the best way of measuring the pump current is to use a feedback circuit and a fourth intracellular microelectrode to hold Em constant and to measure the current required to do this. In early experiments I found that if I rigidly clamped Em at its resting value (inevitably a little arbitrary since the cell is spontaneously active), the current was unstable, as if the cell’s properties were constant only when action potentials occurred. I therefore gave the clamp circuit a long time constant and kept its output constant during each action potential. In this way action potentials were not eliminated, and only the average Em kept constant. Figure 5 illustrates the operation of this slow clamp circuit. On the left is shown the effect of a NaAc injection on the unclamped, free Em.On the right is the effect of a similar injection made a few minutes after the feedback circuit was switched on and set to maintain the previous average Em. During the first injection Em rose by 18 mV and then declined with the usual complex time course. During the second injection Em was kept almost constant, apart from the action potentials. During the second injection the clamp current rose linearly to a peak of 1.7 nA and then declined exponentially to slightly beyond its preinjection level. The clamp current would be equal to that generated by the pump only if there were no change in membrane resistance. Any increase in K permeability, for example, would tend to generate a clamp current in the same direction as that generated by the pump. In preliminary experiments I therefore carefully checked the effects of Na injections on the conductance
,,:.,, 60
Em CONTROLLED
I
1v
FIG.5 . The effect of NaAc injection (by a current o f 38 nA for 1 minute) on the membrane potential before, and on the membrane potential and clamp current after, switching on a clamp circuit to control Em.
1. ELECTROGENIC SODIUM PUMP
9
of a clamped cell. I measured the conductance at intervals by changing the set potential of the clamp circuit from the resting Em to one 10 mV more negative. The results showed that injections made with charges of less than about 3 pC had no effect, but that larger injections appeared to increase the membrane conductance. Later experiments by Partridge and Thomas (1976) have shown that large injections of either Na or Li increase the membrane K permeability, possibly via an increase in intracellular Ca activity. Such permeability changes following large Na injections can cause serious difficulties in pump current measurements (see Kononenko and Kostyuk, 1976). The clamp current shown in Fig. 5 rose and fell as expected for a current proportional to the rate of a pump whose activity above normal is directly proportional to the increase in internal Na above normal. If so, the recovery of the internal Na and clamp current to their preinjection levels should have the same time course. In the paper published in 1969 I was able to show this for a small number of injections using an inverted-tip N a + sensitive microelectrode, the properties of which I only half understood. I have since confirmed the parallelism between clamp current and internal Na using better electrodes (Thomas 1972a,b). The area enclosed by the clamp current resulting from an injection and the baseline current gives the charge generated by the clamp in holding the average Em constant during an injection. For 31 injections I measured the clamp charge for each and compared it with the charge used to inject Na. The average ratio of clamp charge to injection charge was 0.21 f 0.006 (SE of the mean). This ratio was obtained by assuming a steady current baseline in the absence of Na injection. The Na pump and clamp current, in effect, convert the injected NaAc into KAc, so it might be better to use the clamp current during a KAc injection as the baseline for measuring the charge generated by the pump. Injections of KAc into a cell caused small depolarizations, so that in a clamped cell the current changed in a direction opposite that seen with NaAc. This is shown in Fig. 6 for one injection each of KAc and NaAc. If the current record for KAc was used as the baseline when measuring the area under the NaAc current, the ratio of clamp charge to injection charge became 0.27 rather than 0.21.
C. lontophoretic Transport Number for Sodium Injections I originally assumed (Thomas, 1969) that all the current leaving the NaAc-injected electrode was carried by Na+ ions entering the cell, and that the charge on the Na+ ions injected was equal to the charge used to inject them. That is, I assumed that the iontophoretic transport number was
10
R. C. THOMAS
c
INaAc
FIG.6. Clamp current during the injection of KAc (25 nA for 80 seconds) and NaAc (35 nA for 60 seconds) into the same slow-voltage-clamped cell.
unity. But some of the injection current may have been carried by anions entering the NaAc electrode, reducing the amount of Na injected. Alternatively, the current through the injected electrode might have caused bulk flow of NaAc out of the electrode tip. To reduce these uncertainties I have now measured the amount of Na leaving an iontophoretic electrode using “model” neurons consisting of saline droplets in oil, as shown in Fig. 7. The droplets are placed on the
NaCl
----
FIG. 7 . Arrangement used to measure iontophoretic ejection of Na salts into saline droplets.
11
1. ELECTROGENIC SODIUM PUMP
ends of silver wires connected to ground. The droplets are made of 100 mM KCl or potassium benzenesulfonate with 5 , 10, or 20 mM NaCl, and their diameters carefully measured at 80 x magnification using an eyepiece micrometer. Four microelectrodes are then inserted into a suitable droplet, and the voltage between the Na+-sensitive and KCl reference microelectrodes recorded while iontophoretic injections are made (see also Thomas, 1976). As soon as the NaC1-filled injection electrode enters, the droplet's Na content begins to increase. While the first iontophoretic injection has little additional effect, subsequent injections markedly accelerate the rate of increase in droplet Na. When the current is passed in the reverse injection, to inject KCl, the Na level in the droplet remains constant. From the entire ensemble of droplet experiments, the extra Na injected by a range of injection charges was calculated in two ways: first, by dividing the injection charge by the Faraday constant, and second, by multiplying the measured concentration change (corrected for leakage, which was assumed constant) by the droplet volume. The results are shown in Fig. 8. Injections into chloride droplets produced less Na than injections into benzenesulfonate ones, presumably because chloride entered the Na
40
U
30
n W IX
2 20 a W
I +
z"
10
//+
/
0
10
Na'
20
30
40
50
CALCULATED FROM CHARGE ( p E q )
FIG.8 Graph showing measurements of Na' released by passing current through Na injection rnicroelectrodes. Open circles show injections from NaAc electrodes into droplets of 100 m M potassium benzenesulfonate and 15 mM NaCI; solid circles show injections from NaCl electrodes into droplets of the same composition as for the NaAc electrodes; and crosses show injections from NaCl electrodes into droplets of 100 mM KCI and 5 mM NaCI.
12
R. C. THOMAS
injection microelectrode more readily than benzenesulfonate. Injections from NaAc electrodes gave more Na than injections from NaCl electrodes. The mean transport numbers were as follows: for NaAc electrodes into benzenesulfonate drops, 0.93; for NaCl into benzenesulfonate, 0.8 and for NaCl into KCl, 0.6. For injections smaller than about 25 pEq the transport number for NaAc into benzenesulfonate was close to unity, and this is the model arrangement closest to the experiments on snail neurons. D. Effect of Increasing f, on the Pump Current The Na pump generates a current whose flow is opposed by the resting Em.The larger the Em,the more work the pump must do in extruding Na+ ions. Thus an increase in Emshould slow or even stop current generation by the Na pump. Kostyuk et al. (1972) found that the clamp current in response to Na injection into clamped snail neurons became very small at Em levels above 80 mV, although Kononenko and Kostyuk (1976) concluded that much of this apparent reduction in pump current was due to an increase in the K conductance. I have performed only preliminary experiments in investigating the potential dependence of the Na pump. One is illustrated in Fig. 9, which shows the clamp current change caused by three Na injections at different values of Em. As Emwas increased, the clamp current became much noiser and less stable. The current change during the injection was 1.75 nA at an Em of 40 mV, 1.55 at 60 mV, and about 1.1 at 90 mV. The current was so unstable at the latter two potentials that I could not measure the clamp INJECTION CURRENT 50 nA
FIG. 9. Recordings of clamp currents for three identical Na injections made at different values of
Em.
J
Em 90
10 min
13
1 . ELECTROGENIC SODIUM PUMP
charge. Thus while the pump may operate more slowly, or generate less total charge at an Em of 90 mV, it is apparently still electrogenic.
E. Comparison of the Sodium Pump with the pH,-RegulatingSystem As well as the Na-K pump, it is now clear that excitable cell membranes have many other active transport systems. The best known are those regulating internal Caz+ and pH, but the mechanisms are not well understood. Neither appears to be electrogenic. This is shown for the pHiregulating system in Fig. 10. In this experiment both intracellular pH and free Na+ ion concentration were measured with recessed-tip glass microelectrodes. Two iontophoretic injections were made; first of HCl and then of NaCl. The acid injection caused a decrease in Em,with a partial recovery as the pHi returned toward its preinjection level. As the pHi recovered,
1
-
E
-r
E
40
60
100
.,
c
z
I n
-E.,
7.25
7
14 10
7.50
'0
z
Y
6
4
U FIG. 10. Experiment comparing the effect of injections of HCI and NaCl on Em,pHi, and internal free Na' ion concentration. Arrows above the Em record show where double-barreled injection electrodes were inserted and withdrawn. Leakage from the first pair of injection electrodes proved excessive, so they were withdrawn without attempting an iontophoretic injection. The cell was superfused with a saline equilibrated with 2.2% C 0 2 in oxygen.
14
R. C. THOMAS
there was a rise in internal N a + , consistent with the entry of Na+ ions through the pH,-regulating mechanism (Thomas, 1978). When NaCl was injected, there was a typical hyperpolarization of the membrane, but little effect on pH,. The injection charges for NaCl and HC1 were very similar, so the different effects on Em show that the pH,-regulating system is not electrogenic.
IV.
DISCUSSION
The main conclusion from the results described above is that about a quarter of the charge used to inject Na ions is needed to keep Em constant while the injected Na is pumped out. This suggests that about a quarter of the injected Na ions are extruded unaccompanied by K uptake. The assumptions behind this suggestion and the possible sources of error will now be considered.
A. Sources of Error 1. THE INJECTION
The results shown in Fig. 8 suggest that the quantity of Na+ ions ejected iontophoretically is quite variable, or the technique for measuring the ejection rate is not very accurate, or both. But there does seem to be a consistent effect of the anion in both the electrode and droplet. The closest model situation to that in the neuron is ejection from NaAc electrodes into benzensulfonate droplets. For injection charges of less than about 3 pC the average transport number was close to unity. Perhaps the error originally made in assuming equality between the charge used for Na injection and the charge on the ions actually injected was not very large. In any future experiments it would clearly be useful to measure the transport number not only into droplets with a composition closer to that of the cytoplasm, but using injection electrodes as similar as possible to those used on the cell. 2. CHARGE GENERATED BY THE SODIUM PUMP
Measurement of the total charge generated by the clamp in keeping Em constant during the extrusion of an injected quantity of Na should yield information about the charge generated by the Na pump hence its coupling ratio). The charges generated by the clamp and pump will be equal if: (1) All the injected Na has left the cell by the time the pump current has re-
1. ELECTROGENIC SODIUM PUMP
15
turned t o its preinjection level. In the sense that this is the same as requiring the intracellular Na activity to return to its preinjection level, this has been confirmed using Na+-sensitive microelectrodes (Thomas, 1969, 1972a). (2) All the injected Na leaves the cell by being extruded by pump sites located in or close to the cell body. Given the probable rate of diffusion of Na+ ions inside the cell, it seems possible that some diffuse down the axon and are extruded out of range of the slow voltage clamp system. This is likely to vary from cell t o cell, and may be a serious problem in cells clamped at high values of Em.In future experiments the effects of axotomy or axon ligature should be investigated; for the present my best guess about this error is that it is small. (3) Na ions are only extruded by the Na pump. It is possible, however, that some leave via the Na-Ca exchange system working backward (Baker, 1972), with the entering Ca being sequestered by mitochondria. Some Na+ ions could also leave via reverse operation of the pHiregulating system (Thomas, 1980), but the number is probably very small.
B. Coupling Ratio of the Sodium Pump If the Na pump extrudes only Na+ ions and takes up only K + ions, then the above results suggest a coupling ratio of 4 Na13 K, or 3 Na12 K. The clamp charge as a percentage of the injection charge should be 25 for the former and 33 for the latter. The figure obtained corrected for acetate effects was between these two. Very likely less Na was injected than given by a transport number of 1.O, and probably some Na left the cell by other routes, so the higher figure is perhaps the better one, assuming that each cycle of the pump takes a fixed number of Na+ ions in and a fixed number of K + ions out. As discussed elsewhere, there is considerable biochemical evidence in favor of a coupling ratio of 3 Na12 KI 1 ATP (Thomas, 1972b), with which my data are certainly consistent. Working on marine molluscs, Cooke et al. (1974) and Marmor (1971) have presented data that suggest a ratio of 2 Nal 1 K. Whether this is the result of a real difference due to the much higher ion concentrations in marine animals, or to some experimental error I do not know. As for the reason behind the pump’s electrogenic property, I suspect it simply results from a biochemical requirement of the pump (at physiological ion and ATP levels) to operate by pumping out three Na+ ions for each two K + ions. Its physiological significance has been discussed elsewhere (Thomas, 1972b). Probably the most important role for the pump’s electrogenic property is that it tends to reduce a cell’s excitability after a period of hyperactivity.
16
R.
C. THOMAS
REFERENCES Baker, P. F. (1972). Prog. Biophys. Mol. Biol. 24, 177-223. Connelly, C. M. (1959). Rev. Mod. Phys. 31, 475-484. Cooke, I. M., LeBlanc, G . , and Tauc, L. (1974). Nature (London) 251, 254-256. Hodgkin, A. L., and Keynes, R. D. (1955). J . Physiol. (London) 128, 28-60. Kernan, R. P. (1962). Nature (London) 193, 986-987. Kononenko, N. I., and Kostyuk, P. G. (1976). J. Physiol. (London) 256, 601-615. ICostyuk, P. G., Krishtal, 0. A., and Pidoplichko, V. I . (1972). J . Physiol. (London) 226, 373-392. Marmor, M. F. (1971). J . Physiol. (London) 218, 599-608. Partridge, L. D., and Thomas, R. C. (1976). J . Physiol. (London) 254, 551-563. Thomas, R. C. (1964). “The Effect of Ions on the Inhibitory Post-Synaptic Potential and on the Resting Potential.” (Ph.D. Thesis), Southampton University. Thomas, R. C. (1969). J. Physiol. (London) 201, 495-514. Thomas, R. C. (1972a). J . Physiol. (London) 220, 55-71. Thomas, R. C. (1972b). Physiol. Rev. 52, 563-594. Thomas, R. C. (1976). J. Physiol. (London) 255, 715-735. Thomas, R. C. (1978). Respir. Physiol. 33, 63-73. Thomas, R. C. (1980). Curr. Top. Membr. Transp. 13, 23-29.
CURRENT TOPICS IN MEMBRANES AND TRANSPORT, VOLUME 16
Chapter 2 Hyperpolarization of Frog Skeletal Muscle Fibers and of Canine Cardiac Purkinje Fibers during Enhanced Na+-K+ Exchange: Extracellular K+ Depletion or Increased Pump Current? DAVID C. GADSBY Laboratory of Cardiac Physiology The Rockefeller University New York, New York
Introduction ........................................................................................ Experiments Using Frog Skeletal Muscle Fibers A. The Pump Hyperpolarization ............................................................ B. K + Depletion in the T System ..... ................................. 111. Experiments Using Canine Cardiac Purkinje Fibers ...................................... A . Aftereffects of Brief ............. Membrane Potential B. Aftereffects of Brief ntials ............ C. Direct Measurement of Changes in Sodium Pump Current in Voltage Clamp Experiments IV. Summary and Conclusions ...................................................................... References ..........................................................................................
I.
XI.
I.
17 19 19 21 26 26 27
30 32 33
INTRODUCTION
Although it is well established that enhanced activity of the Na+-K+ exchange pump causes a temperature- and ouabain-sensitive hyperpolarization of Na+-loaded skeletal and cardiac muscle cells (see reviews by 17
Copyright 'CI 1982 by ALademiL Press, Inc All rights of reprodu'tion in any form reserved ISBN 0-12-153316-6
18
DAVID C. GADSBY
Thomas, 1972; Glitsch, 1979), it has proved more difficult to establish conclusively that this hyperpolarization reflects a direct electrogenic effect of pump activity rather than an indirect effect resulting from extracellular K depletion with an associated negative shift of the K + equilibrium potential E, (e.g., Page and Storm, 1965; Adrian and Slayman, 1966; Gadsby ef al., 1977). The pump is electrogenic if it transports unequal quantities of Na+ and K + in opposite directions across the cell membrane, since the difference in flux then appears as membrane current. Whether the pump is electrogenic or not, however, recovery of the transmembrane ionic gradients following Na+ loading requires a net influx of K + into the cells, and whenever this K + uptake occurs across a poorly perfused extracellular space, e.g., at the interior of large multicellular preparations, enhanced pump activity can be expected to lower the K + concentration just outside the cell membrane. Of course, K + depletion is not limited to such large preparations since, even in well-superfused small preparations, important unstirred extracellular spaces still exist, for instance, the narrow lumen of the transverse tubular system in single skeletal muscle fibers and working myocardial cells; if, as now seems likely (see below), high pump rates are associated with appreciable K + influx across the membrane forming the walls of the T system, then the K + concentration in its lumen must be expected to fall. Now, the kind of evidence most usually offered as proof of electrogenic pump activity in skeletal and cardiac muscle is that membrane potentials recorded during recovery from Na+ loading exceed estimates of E, calculated, using the Nernst relation, from experimentally determined values for intracellular K + concentration, [KIi, on the assumption that the K + concentration adjacent to the cell surface is equal to that of the bulk bathing fluid; as already mentioned, this assumption is probably invalid at high pump rates. Some evidence for this assertion will be presented in Section 11. Using membrane K + conductance as an indirect measure of T tubular K + concentration, we will show that, in Na+-loaded skeletal muscle fibers, well superfused in a fast-flow chamber, the K + concentration in the T-tubular lumen falls considerably at high pump rates (see Gadsby ef al., 1977). In Section 111, we will take a different approach to the problem and ask the question: Can we find conditions under which the presence of electrogenic Na+ extrusion can be demonstrated simply and unequivocally? We will show that, in the case of cardiac Purkinje fibers, there are at least three experimental conditions under which a temporary increase in pump activity causes membrane potential changes which are in the direction expected from an increase in sodium pump current but which are opposite in direction to those observed to result from experimental reduction of the extracellular K + concentration, [K], (see Gadsby and Cranefield, 1979a). +
2.
Na+-K+
19
PUMP HYPERPOLARIZATION OF MUSCLE FIBERS
Most of the experiments described here were carried out in chloride-free solutions, in which isethionate was the major anion, in order to avoid possible complications arising from net KCI movements. Under these conditions, cellular Na+ movements can reasonably be assumed to be accompanied by approximately equal but opposite movements of K + . To further facilitate interpretation of the results, we used for these experiments only isolated fibers, or small bundles of fibers, from skeletal muscles, or small bundles of cardiac Purkinje cells, suspended in the narrow channel of a modified Hodgkin-Horowicz (1959) fast-flow system, so that step changes in the Composition or temperature of the fluid at the periphery of the fibers could be made. These step changes were imposed to rapidly alter the level of activation of the sodium pump.
II.
EXPERIMENTS USING FROG SKELETAL MUSCLE FIBERS
A. The Pump Hyperpolarization The hyperpolarization caused by enhanced pump activity in Na -loaded muscle fibers, and the time course of its decline during recovery from Na+ loading, are shown in Fig. 1. These fibers were in a small bundle from a cutaneus pectoris muscle and had been Na+-loaded by exposure to 0.1 m M [K], at 5 2°C for 48 hours (the same procedure was adopted for Na+loading fibers in all the experiments on skeletal muscle), after which [K], was raised to 10 mM while the temperature was kept low. Membrane potentials were then recorded before and after temperature steps between 2 and 22"C, and these potential levels are plotted against the time of measurement in Fig. la. Since pump activity is much reduced at low temperatures, the resting membrane potential in 10 mM K + fluid at 2°C is approximately equal to E,, so that the slow increase in the potentials measured at 2°C reflects the recovery of the intracellular K + concentration, [KIi, resulting from the activity of the Na+-K+ pump, which occurs predominantly at the higher temperature. For this reason, the potentials are shown replotted against incubation time at 22°C in Fig. l b , where the dashed curve indicates the predicted effect of the rise in temperature per se on resting potentials which are determined by only passive ion fluxes (Ling and Woodbury, 1949; Hodgkin and Nakajima, 1972). The difference between the dashed curve and the potentials recorded at 22°C is then the component attributable to enhanced pump activity, the pump hyperpolarization, and this is shown in Fig. lc. +
20
DAVID C. GADSBY
-
2
>
(a 1
0'
-40 -
.
A
0
-
E
v
\
-50
15
.
d
o'a
-
A
0
-
C
-60 -
.+
2 0
0.
I
e
I
I
20
I
I
40
I
I
60
I
80
A
9
n
$
I
0
a C
A
01
-40
2
(b)
.\:
12
-50
-60 I
0
I
1
20
1
I
40
0
20
40
Time /min FIG. 1. Membrane potentials of three Na+-loaded muscle fibers (each represented by a different symbol) during the recovery of normal cellular ion levels in 10 mM K + , CI--free fluid. (a) The top trace indicates the step changes in temperature. The upper points show potentials recorded at 2"C, and the lower points show values at 22°C. (b) Data from (a) replotted against cumulative time spent at 22°C. The solid curves were fitted to the two sets of points by eye. The dashed curve was obtained by correcting the curve at 2°C for the expected effect of warming to 22°C for membrane potentials determined by only passive fluxes (see text). (c) Plot of A V m , the difference between the expected (dashed) and the obtained curves for 2 2 T , i.e., the hyperpolarization attributable to pump activity, against incubation time at 22°C. (From Gadsby el al., 1911.)
The pump hyperpolarization is known to be abolished by ouabain (e.g., Cross et a[., 1965; Adrian and Slayman, 1966; Akaike, 1975), and Fig. 2 confirms that it is also abolished rapidly by another cardiac steroid, acetylstrophanthidin. Membrane potential was continuously recorded in a fiber in a Na+-loaded semitendinosus bundle, exposed to 10 m M K + fluid,
2.
Na+-K+ PUMP
21
HYPERPOLARIZATION OF MUSCLE FIBERS
Acetylstrophanthidin
I -30 J' m
-35
0
-
0 0
mV -40
-45
0
-
0
0
0 0
0
-
0 0
0
-50
0
0
0
-
0 0 0
0
0
0
-551
O O
I 0
O
I
ooooo I
1
min
1
1
I
3
4
5
FIG.2. Effect of 2 pM acetylstrophanthidin (application indicated by bar) on the membrane potential of a Na' -loaded muscle fiber following a step change in temperature from 2 to 20"C, as indiclted by the upper line. A single impalement was maintained throughout these changes, and the points were measured from enlarged photographs of several superimposed sweeps of the oscilloscope trace; 10 mM K + , CI--free fluid throughout.
while the temperature was raised from 2 to 20°C. Subsequent addition of 2 gM acetylstrophanthidin caused a substantial depolarization, 28.5 mV in this case, which occurred with a half-time of 59 seconds. The difference between the initial and final levels of membrane potential in this experiment, -21 - (-27.5) = 6.5 mV, can presumably be attributed partly to the temperature effect mentioned above and partly to the increase in [KIi expected to occur at the higher temperature, both before the addition of acetylstrophanthidin and also after, until pump inhibition was complete. B. K + Depletion in the T System The first clue that K + depletion in the T tubular lumen contributes to the pump hyperpolarization came from analysis of the time course of potential change in Na+-loaded single fibers (or very small bundles) in response to step temperature changes. As shown in Fig. 3, the potential change both on warming and on cooling showed a pronounced slow phase, at least half of
' IFoc
22
DAVID C. GADSBY
-40
-45
1
(a1
1.j
-50 I 0
I
I
2
I
I 4
I
I 6
I
I 0
Time / s
FIG. 3. Membrane potential changes recorded in an isolated Na+-loaded fiber from an extensor digitorum longus muscle in response to step changes in temperature between 2.5 and 19.5"C. The temperature changes, shown by the upper traces in (a) and (b), were recorded using a microthermistor bead positioned just beneath the fiber at the point of microelectrode impalement. Membrane potentials were measured from enlarged photographs and are shown plotted against time after the start of the temperature increase in (a) and temperature decrease in (b). The horizontal dashed lines indicate the levels to which the membrane voltage would be expected to move if the steady potentials were determined by only passive ionic fluxes; 20 mM K', CI--free fluid throughout. (From Gadsby eta!., 1977.)
the pump hyperpolarization occurring after the temperature change was more than 90% complete. The slow phase followed an approximately exponential time course with a half-time of about 1 second, similar to that previously determined for changes of K + concentration in the T system, whether these are evoked either by a sudden change in external K + concentration or by the application of current across the fiber membrane (Hodgkin and Horowicz, 1960; Almers, 1972a; Barry and Adrian, 1973; Kirsch, et al., 1975). This slow phase is unlikely to be due to ion equilibration at the peripheral fiber surface, since the half-time for this process is
2.
Na+-K+ PUMP HYPERPOLARIZATION OF
MUSCLE FIBERS
23
known to be much smaller, i.e. about 0.2 second (e.g., Hodgkin and Horowicz, 1960; Nakajima et al., 1973). Moreover, since there is no a priori reason t o expect changes in pump rate to lag appreciably the changes in temperature, the conclusion from experiments of this kind is that the slow potential changes reflect K + depletion in the T system during high pump rates (on warming) and a reaccumulation of K + when pump activity is low (on cooling). Further evidence of K + depletion in the T system caused by enhanced pump activity came from experiments in which membrane current-voltage relationships were determined in a given Na+-loaded fiber under conditions of both high and low pump activity: The aim of the experiments was to detect a difference in conductance between these two conditions, since a fall in T tubular K + concentration is known to reduce membrane conductance (Adrian and Freygang, 1962; Almers, 1972a). Figure 4 shows current-voltage relationships from one such experiment: The smooth curve on the left was obtained at high pump activity, in 20 m M K + solution at 22"C, and that on the right after inhibiting the pump by adding 20 pM ouabain. The dashed segment of curve was determined also in 20 mMK+solution but at low pump activity, at 2"C, just before adding the ouabain, and has been corrected to apply at 22°C by using experimentally determined Qlovalues for membrane K + conductance. (The Qlovalues were estimated from the observed temperature dependence of membrane chord conductance in the presence of ouabain.) Note that the two curves at low pump activity are closely similar, and that the curve at high pump activity is not simply displaced to the left along the voltage axis but is also shallower, over the entire voltage range shown, than the curves at low pump activity. Since in skeletal muscle fibers most of the K + conductance resides in the T system (see Eisenberg and Gage, 1969; Almers, 1972b; Schneider and Chandler, 1976), these findings strongly suggested that, during enhanced pump activity, membrane K + conductance is reduced as a result of T tubular K + depletion, and so the next step was to attempt to quantify the extent of this depletion. The general approach was to determine, in control experiments, the [K], dependence of membrane K+ conductance so that calibration curves could be constructed for estimating T tubular K + concentration from the change in conductance due to pumping found from current-voltage curves like those in Fig. 4. The K + concentration at the external surface of single fibers, or small bundles, in a fast-flow system can be expected to remain close to [K],, the superfusate concentration (20 m M in these experiments), even during high pump rates, so that K + depletion affects only the conductance of the membrane forming the walls of the T tubules, i.e., about 75% of the total membrane area of the fiber (cf. Peachey, 1965; Mobley and Eisenberg, 1975;
24
DAVID C. GADSBY
Vm/mV
- 80
-100 1
1
I
1
- 20
-40
-60 I
I
I
I
I
0
-
-20
su
5 3.
-40
v
\
k
%
-60’
-80
FIG. 4. Membrane current-voltage relationships obtained from a Na+-loaded fiber from an extensor digitorum longus muscle exposed to 20 mM K +,CI--free fluid under conditions of high and low pump activity. The curves were constructed from current-voltage measurements made with two microelectrodes after correcting the data for (1) the shunt conductance introduced across the cell membrane by the microelectrodes, (2) the finite separation (< 100 pm) of the microelectrodes, (3) the cable properties of the fiber, and (4)the estimated “leakage” conductance of the membrane to Na+ . The curve on the left was obtained at high pump activity at 22°C. The solid curve on the right was also obtained at 22”C, but in the presence of 20 pA4 ouabain. The dashed segment of curve was obtained at low pump activity, but at 2”C, before adding the ouabain; these data were corrected, in addition, for the change in temperature from 2 to 22°C by means of Q l ovalues for membrane K + conductance. (From Gadsby et al., 1977.)
Schneider and Chandler, 1976). A constant conductance component, of 25% of the total K + conductance obtained at 20 mM [K], in the absence of pump activity, was therefore ascribed to the fiber surface, and that constant component was summed with 75% of the total conductance obtained at lower [K], levels (all at low pump activity) to mimic the effects of T tubular K + depletion. The resulting total conductance values, representing different T tubular K + concentrations, were normalized to the total conductance at a uniform K + concentration of 20 mM and then plotted against T tubular K + concentration for comparison with the measured ratio of the membrane conductances obtained at high and at low pump activity. These
2.
Na+-K+
PUMP HYPERPOLARIZATION OF MUSCLE FIBERS
25
conductance ratios were determined in two different voltage ranges, and two corresponding calibration curves (both constructed in the manner described above) were used. The first voltage range used was 40-70 mV negative to the resting potential, since in this region the K + current-voltage relationships are approximately linear and therefore slope conductances are easily measured. The second conductance ratio was determined (by computer differentiation of the current-voltage curves) at a constant potential, chosen as the value of E, for a uniform K + concentration of 20 mM. The conductance ratios so obtained (high divided by low pump activity) ranged from 0.64 to 0.84 with the first method, and from 0.33 to 0.60 with the second, yielding average estimates (fSD) of T tubular K + concentration of 11.1 f 2.2 m M and 13.2 f 1.9 mM, respectively, during high rates of pump activity in Na+-loaded muscle fibers exposed to 20 mM K + fluid. Is it possible that this T tubular depletion results secondarily from a hyperpolarization caused directly by electrogenic pump activity at sites on the peripheral fiber surface; i.e., is it analogous to the K + depletion previously shown to occur during the application of prolonged hyperpolarizing current pulses from an external source (Adrian and Freygang, 1962; Almers, 1972a)? Any component of the K + depletion occurring by this mechanism during enhanced pump activity is expected to be abolished when prolonged depolarizing current pulses are applied to the fiber to make the membrane potential approximately equal to E, for a uniform external K + concentration of 20 mM, since passive net K + flux is then zero. In fact, outward current pulses of 4-seconds duration were used to obtain the positive limbs of the current-voltage curves shown in Fig. 4 and yet, as already described, the marked reduction in slope conductance in that voltage region indicates that considerable tubular K depletion persists at the end of such prolonged depolarizations (second method, above). Since, therefore, pump activity of the peripheral fiber surface is unable to account for the observed K + depletion, it must be the direct result of K + uptake from the T tubular lumen; in other words, there must be Na+-K+ pump sites in the walls of the T system (cf. Barchi et af., 1977; Lau et al., 1979). Although K + depletion clearly plays a major role in the pump hyperpolarization of Na+-loaded muscle fibers, Fig. 3 shows that up to 50% of the potential change following a temperature step occurred quite rapidly. The time course of this component would be compatible with K + depletion and accumulation in the narrow unstirred layer which probably remains at the peripheral fiber surface, but its magnitude is far too great, since the concentration changes expected across such a layer are in the micromolar range. Hence the most likely explanation for this rapid component of potential change is that it reflects the contribution made directly by electrogenic pump activity. +
26
DAVID C. GADSBY
Ill.
EXPERIMENTS USING CANINE CARDIAC PURKINJE FIBERS
A. Aftereffects of Brief Periods of Na+ Loading on Resting Membrane Potentials Fortunately, the presence of electrogenic pump activity can be demonstrated more readily, without the usual complications associated with K depletion, in experiments on small bundles of cardiac Purkinje cells suspended in a fast-flow system. This is partly attributable to the anomalous behavior of the resting potential of these cells at moderately low external K + concentrations. Thus at low [K], levels of, say, 1-4 mM, there are often two possible values for the resting potential of Purkinje fibers (Wiggins and Cranefield, 1976; Gadsby and Cranefield, 1977), as can be seen in Fig. 5 ; more important, when the membrane potential is at the lower (more positive) resting level, a reduction in [K], causes immediate depolarization; i.e., the resting potential moves in a direction opposite t o the change in E, (Figs. 5 and 6). Under these conditions, therefore, enhanced K + uptake by a neutral Na+-K+ exchange pump sufficient to cause extracellular K depletion should result in membrane depolarization, whereas enhanced electrogenic Na + extrusion should increase outward membrane current and so cause hyperpolarization. Figure 6a shows quite clearly that, when a fiber at the lower resting potential in 4 mM [K], is briefly exposed to K+-free fluid, the depolarization at zero [K], is followed, on switching back to 4 mM [K],, by a transient hyperpolarization which reaches a peak within a few seconds and then more slowly decays. The +
+
ot
I
FIG.5 . Membrane potential changes recorded in a small canine Purkinje fiber in response to the step changes in [K], indicated by the upper line. The resting potential at 4 mM [K], was -90 mV both before and after exposure to the other [K], levels. This is evident from the thickening of the first few seconds of the voltage trace, which results from superposition of the beginning of a second sweep. Note that at 2 m M [K], two different levels of resting potential were obtained, one near - 100 mV and the other at about - 45 mV. Temperature, 36°C; Cl-free solutions throughout. (From Gadsby and Cranefield, 1977.)
2. Na+-K+ PUMP HYPERPOLARIZATION OF MUSCLE FIBERS
27
FIG.6. The transient hyperpolarization recorded in small Purkinje fibers on returning to 4 m M K+ fluid following brief periods of exposure to zero [K],. The changes in [K], are indicated by the upper lines, and the traces below them show the resulting changes in Vm, the membrane potential. (a) Effect of a 1-minute exposure to K+-free fluid. The dashed line shows the steady level of resting potential at 4 m M [K],, - 32 mV. (b) Superimposed records, obtained during a single maintained impalement, of responses t o four brief exposures to zero [K], of durations 15,30,60, and 90 seconds, respectively. Temperature, 36.5"C; Cl--free solutions throughout. (From Gadsby and Cranefield, 1979a.)
hyperpolarization reflects the temporary speeding up of the pump in response to the rise in intracellular Na+ concentration, [NaIi, which took place while the pump was slowed during exposure to K+-freefluid; thus the transient hyperpolarization is completely abolished in the presence of L 2 pi" acetylstrophanthidin (see Fig. 4 of Gadsby and Cranefield, 1979a). Furthermore, as shown in Fig. 6b, the peak amplitude of the hyperpolarization becomes larger as the duration of the exposure to K+-free fluid, and hence the degree of Na+ loading, is increased (more accurately, as the magnitude of the increment in [Na], is increased). As already discussed, extracellular K depletion cannot account for the transient hyperpolarization which must therefore reflect a temporary increase in outward membrane current, in other words, an increase in the rate of electrogenic Na+ extrusion. +
B. Aftereffects of Brief Periods of N a + Loading on Action Potentials The same conclusion was reached by studying the aftereffects of similar short periods of Na+ loading, in K+-free fluid, on both driven and spontaneous action potentials. Figure 7 shows that a few seconds after switching back to 4 mM [K],, following a 6-minute exposure to zero [K],, the duration of the action potential in a regularly stimulated Purkinje fiber was reduced by almost 50% with respect to control action potentials recorded either just before the exposure to K+-free fluid or many minutes after. The
28
DAVID C. GADSBY 4K
-4
6minOK
Recovery
I
e-.
I
I
I
a. 4’ 1
I
2oo time
0.8
1
400
-*
-*,
,*-•
- - -
Norm. A.P.D.
(s)
7 r 0 - ’
/
/*
FIG.7. Changes in action potential duration (A.P.D.) recorded in a small Purkinje fiber in 4 mM K + , CI--containing Tyrode’s solution at 36°C during recovery from a 6-minute period of exposure to zero [K],, which is indicated by the upper line. The action potentials were evoked by electrical stimulation of the fiber at a rate of 75 min-’, except during exposure t o K+-free fluid when the fiber was depolarized. The upper set of points shows membrane potential levels (V,) between action potentials, at 4 mM [K],, before and after the period at zero [K],. The lower points show corresponding durations of sample action potentials: These durations have been normalized with respect to the control duration (indicated by the broken horizontal line) either just before or a long time after the exposure to zero [K],. The open circles give normalized durations of the representative action potentials which are illustrated at the bottom: At bottom left, the control action potential (the longer one) recorded just before exposure to K+-free fluid is shown superimposed on that recorded 60 seconds after the return to 4 mM [K],; on the right, the action potential recorded after 60 seconds is shown superimposed on those obtained after 90, 140, and 350 seconds of recovery at 4 mM [K],. The horizontal time calibration for these action potentials also marks the zero-potential level. (From Gadsby and Cranefield, 1979a.)
action potential shortening was accompanied by an increase in resting potential (hyperpolarization) and, as seen in Fig. 7, both action potential duration and membrane potential returned to control values with similar time courses. Electrical stimulation was stopped during the exposure to zero [K],, since the fiber was then depolarized, but this cannot provide an explanation for the subsequent temporary shortening of the action potential observed in 4 mM K + fluid, because it is well known that Purkinje fiber
2.
Na+-K+
PUMP HYPERPOLARIZATION OF MUSCLE FIBERS
29
action potentials are initially lengthened on resuming stimulation after a prolonged pause and only gradually shorten to the steady state duration (see, e.g., Hoffman and Cranefield, 1960; Miller et al., 1971). Although temporary extracellular K depletion might seem a plausible explanation for the transient hyperpolarization, it could not account for the concomitant reduction in action potential duration: Experimental lowering of [K], is known to lengthen the Purkinje fiber action potential, not shorten it (Weidmann, 1956; Vassalle, 1965; Noble, 1965), so that both the transient hyperpolarization and action potential shortening are most readily explained by a temporary increase in the outward current generated by the Na+-K+ pumpFigure 8 shows the effects of a 2-minute exposure to K+-free fluid on a Purkinje fiber initially beating spontaneously at a regular rate of about 30 min-' in a 4 mM K + solution. The depolarization during exposure to zero [K], was associated with spontaneous, slow-response action potentials but, on returning to 4 mM [K],, the membrane hyperpolarized rapidly to a level 6 mV more negative than the original, steady, maximum diastolic potential, and all spontaneous activity was abolished for more than a minute; when spontaneous action potentials reappeared, their rate was initially very low but increased gradually until the control rate was reestablished (cf. Vassalle, 1970). Once again, temporary K + depletion resulting from pump stimulation cannot account for the quiescent period, since experimental reduction in [K], is known to enhance spontaneous activity not abolish it (Vassalle, 1965). Hence both the observed short-term quiescence and the increase in maximum diastolic potential are most readily attributed to a transient increase in the rate of electrogenic Na+ extrusion. +
I
OK
I
FIG.8 Change in the rate of spontaneous activity of a Purkinje fiber, in 4 m M K + , Cl-containing Tyrode's solution at 36"C, in response to a 2-minute exposure to K+-free solution, indicated by the horizontal bar above the voltage record. The dashed line marks the control level of maximum diastolic potential, -87 mV. The fiber depolarized in K+-free fluid, giving rise to spontaneous, slow-response action potentials. The vertical calibration bar represents 100 mV, and its upper end indicates the zero-potential level; the horizontal calibration represents 30 seconds. The voltage trace is a chart recording, so the overshoots of the action potentials at 4 mM [K], are not seen. (From Gadsby and Cranefield, 1979a.)
30
DAVID C. GADSBY
C. Direct Measurement of Changes in Sodium Pump Current in Voltage Clamp Experiments Since, in canine Purkinje fibers, the aftereffects of brief periods of Na+ loading in K+-free fluid are clearly attributable to a transient increase in pump current rather than to temporary K + depletion, it seemed worthwhile to try to measure these current changes under voltage clamp (see, e.g., Thomas, 1969). In these experiments, two microelectrodes are used to clamp small Purkinje fibers, I 2 mm in length and I200 pm in diameter, at a constant holding potential throughout numerous step changes in external K + concentration; the more positive of the two possible levels of resting potential at 4 mM [K], is usually chosen as the holding potential (see Gadsby and Cranefield, 1979b). Figure 9a shows the membrane potential changes, in response to a 1-minute exposure to K+-free fluid, recorded in a fiber impaled with two microelectrodes but with the voltage clamp amplifier switched off. Figure 9b shows that, with the clamp amplifier switched on and the membrane potential held at the resting (zero net current) level at 4 mM [K],, a maintained net inward current is recorded during the exposure to K+-free fluid and a transient net outward current arises on
1
ot u
1rnin
(b) Clamp on
FIG. 9. Changes in membrane potential and in net membrane current recorded in a short, thin Purkinje fiber immersed in CI--free solution at 36.5"C in response to a 1-minute exposure to zero [K],, as indicated at the top of the figure. In both (a) and (b), the upper trace shows the membrane potential, and the lower trace, labeled I, shows the net membrane current. (a) Membrane potential changes (similar t o those in Fig. 6) recorded with the voltage clamp amplifier switched off. The broken line indicates the steady resting potential in 4 m M [K],, -33 mV, i.e., the more positive of the two possible levels of resting potential. (b) Corresponding changes in net membrane current seen after switching on the voltage clamp amplifier to hold the potential at - 33 mV, the zero net current level in 4 m M [K],. The broken line marks zero net current. Note the exponentially decaying, transient net outward current (equivalent to the increment in sodium pump current), following the return to 4 m M [K],, which corresponds to the transient hyperpolarization shown in (a). (From Gadsby and Cranefield, 1979b.)
2.
Na+-K+
PUMP HYPERPOLARIZATION OF MUSCLE FIBERS
31
returning to 4 mM [K],. These net current changes can be seen to have magnitudes and time courses appropriate to account for, at least qualitatively, the membrane potential changes of Fig. 9a. The transient net outward current results from temporary stimulation of the sodium pump and not, for example, from a temporary change in membrane conductance or from a reduction in Na+ influx due to the raised [Na],, since when similar step changes in [K], are repeated in the presence of 2 2 pM acetylstrophanthidin, the return to K+-containing fluid is not accompanied by any transient overshoot of the steady level of holding current (see Fig. 2 of Gadsby and Cranefield, 1979b; cf. Eisner and Lederer, 1979). Moreover, the transient outward current must reflect an increment in the current generated directly by the sodium pump, rather than a change in K + current secondary to K + depletion, since the same current record shows that, at the chosen level of holding potential, lowering [K], causes the net membrane current to become more inward, not outward (Fig. 9b). The increment in pump current recorded on switching from K + -free fluid back to K+-containing fluid, reaches a peak within a few seconds and then declines exponentially with an average time constant, at 4 mM [K],, of 1.4 minutes (Gadsby, 1980; cf. Gadsby and Cranefield, 1979b). This time constant is independent of changes in “a], (Gadsby and Cranefield, 1979b) but depends strongly on the level of [K],, diminishing slightly as [K], is raised above 4 mM and increasing markedly when [K], is lowered below 4 mM (Gadsby and Cranefield, 1978; Gadsby, 1980); its reciprocal provides a measure of the rate constant for Na+ extrusion by the Na+-K+ pump, and the [K], dependence of this rate constant is illustrated in Fig. 10. We can conclude that, in canine cardiac Purkinje cells, as in many other cells, the pump rate constant is half-maximally activated by about 1 m M extracellular K+ (see, e.g., Glynn and Karlish, 1975; Glitsch, 1979). Deitmer and Ellis (1978) found that the maximum rate of recovery of intracellular Na+ activity in sheep Purkinje fibers, following Na+ loading, was half-maximally activated by extracellular K + at a [K], of about 10 mM. Since diffusion equilibration of K + is known to be slow in the extracellular spaces of sheep Purkinje fibers, so that even moderate rates of K + influx lead to measurable K + depletion (Baumgarten and Isenberg, 1977; cf. Cohen et al., 1976), it seems likely that, in the experiments of Deitmer and Ellis (1978), the K + concentration just outside the cells during recovery from Na+ loading might have been considerably lower than the bath concentration, [K],. In other words, the difference between the results in Fig. 10 and those of Deitmer and Ellis (1978) is most likely attributable to extracellular K t depletion at high pump rates in sheep Purkinje fibers, just as was demonstrated for skeletal muscle fibers in Section 11.
32
DAVID C. GADSBY
O A i i
i
8 'h6
4
[Kl0
(mM)
FIG. 10. Dependence on [K], of the exponential rate constant for decay of the pump current transient. Exponential rate constants were determined from semilogarithmic plots of the decay of pump current increments obtained at different (K], levels. These rate constants were normalized with respect to the rate constant at 4 mM [K],, and the mean values are given by the circles: the circle diameters equal 2 x SEM, the vertical bars indicate f SD, and the numbers in parentheses show how many experiments contributed to each mean. The curve is a rectangular hyperbola: The pump rate constant is half-maximally activated by external K + at [K], = 0.94 mM. (From Gadsby, 1980.)
IV.
SUMMARY AND CONCLUSIONS
The overall conclusion from these experiments is that the Na+-K+ exchange pump in cardiac and skeletal muscle cells is electrogenic, more Na+ being pumped out of the cells than K + pumped in. However, it is equally clear that under conditions of a large net K + influx, e.g., during the recovery of muscle cells from extensive periods of Na+ loading, the K + concentration just outside the cells may be considerably lower than the concentration in the bulk bathing fluid. Such K + depletion has previously been shown to occur in both skeletal muscle (Adrian and Freygang, 1962; Almers, 1972a) and cardiac preparations (Maughan, 1973; Baumgarten and Isenberg, 1977) as a result of the increase in K + influx associated with membrane hyperpolarization caused by the injection of current from an external source. It seems safe therefore to draw a further conclusion, namely, when the sodium pump rate is greatly enhanced, separation of the component of pump hyperpolarization attributable to increased pump current from that attributable to K + depletion requires a detailed knowledge of several characteristics of the preparation being studied: It is likely that during recovery from Na+ loading, at least two-thirds of the total K + influx is active (see, e.g., Thomas, 1972), so that the relative sizes of the direct (electrogenic) and indirect (K+ depletion) contributions to the pump hyperpolarization will depend on how the membrane K + conductance varies with changes in membrane potential as well as with changes in exter-
2.
Na+-K+ PUMP HYPERPOLARIZATION OF MUSCLE FIBERS
33
nal K + concentration, in addition to other factors such as the surfaceIvolume ratio of the extracellular space and the time taken for diffusion equilibrium of K + within it. On the other hand, the results obtained by using rapid-flow techniques to subject cells to only brief periods of Na + loading, thereby presumably causing only slight variations in [NaIi and in the pump rate (cf. Gadsby and Cranefield, 1979b; Gadsby, 1980), suggest that measurements of small changes in pump current can be made apparently uncomplicated by extracellular K + depletion (Fig. lo), even in multicellular preparations, if the geometric arrangement of the cells is favorable (cf. Colatsky and Tsien, 1979; Cohen et af., 1979).
ACKNOWLEDGMENTS The experiments described in Section I1 were carried out in the Biophysics Department, University College London, in collaboration with Dr. R. Niedergerke and David C. Ogden. The fast-flow system developed in the laboratory of R. Niedergerke subsequently formed the basis for the experimental approach to the work, presented in Section 111, done in collaboration with Dr. Paul F. Cranefield. I am indebted to Joan Leary and Toni Sachs for technical assistance. The work done in London was supported by grants from the British Heart Foundation, the Medical Research Council, and the Wellcome Trust. The preparation of this article and the work done in New York were supported by U.S. Public Health Service grant HL-14899.
REFERENCES Adrian, R. H., and Freygang, W. H. (1962). J. Physiol. (London) 163, 61-103. Adrian, R. H., and Slayman, C. L. (1966). J. Physiol. (London) 184, 970-1014. Akaike, N. (1975). J. Physiol. (London) 245, 499-520. Almers, W. (1972a). J . Physiol. (London) 225, 33-56. Almers, W. (1972b). J . Physiol. (London) 225, 57-83. Barchi, R. L., Bonilla, E., and Wong, M. (1977). Proc. Nufl. Acud. Sci. U.S.A. 74, 34-38. Barry, P.H., and Adrian, R. H. (1973). J . Membr. Biol. 14, 243-292. Baumgarten, C. M., and Isenberg, G. (1977). Pfluegers Arch. 368, 19-31. Cohen, I., Daut, J., and Noble, D. (1976). J . Physiol. (London) 260, 55-74. Cohen, I., Falk, R., and Kline, R. (1979). J . Physiol. (London) 296, 72P. Colatsky, T. J., and Tsien, R. W. (1979). J . Physiol. (London) 290, 227-252. Cross, S. B., Keynes, R. D., and Rybova, R. (1965). J . Physiol. (London) 81, 865-880. Deitmer, J. W., and Ellis, D. (1978). J . Physiol. (London) 284, 241-259. Eisenberg, R. S., and Gage, P. W (1969). J . Gen. Physiol. 53, 279-297. Eisner, D. A., and Lederer, W. J . (1979). J. Physiol. (London) 296, 75P. Gadsby, D. C. (1980). Proc. Natl. Acad. Sci. U.S.A. 77, 4035-4039. Gadsby, D. C., and Cranefield, P. F. (1977). J. Gen. Physiol. 70, 725-746. Gadsby, D. C., and Cranefield, P. F. (1978). Biophys. J. 21, 166a.
34
DAVID C. GADSBY
Gadsby, D. C., and Cranefield, P. F. (1979a). J. Gen. Physiol. 73, 819-837. Gadsby, D. C., and Cranefield, P. F. (1979b). Proc. Natl. Acad. Sci. U.S.A. 76, 1783-1787. Gadsby, D. C., Niedergerke, R., and Ogden, D. C. (1977). Proc. R . SOC. London Ser. B. Biol. Sci. 198, 463-472. Glitsch, H.G. (1979). A m . J. Physiol. 236(2), H189-HI99. Glynn, I. M., and Karlish, S. J. D. (1975). Annu. Rev. Physiol. 37, 13-55. Hodgkin, A. L., and Horowicz, P. (1959). J. Physiol. (London) 148, 127-160. Hodgkin, A. L., and Horowicz, P. (1960). J . Physiol. (London) 153, 370-385. Hodgkin, A. L., and Nakajima, S. (1972). J. Physiol. (London) 221, 105-120. Hoffman, B. F., and Cranefield, P. F. (1960). “Electrophysiology of the Heart.” McGrawHill, New York. Kirsch, G. E., Nichols, R. A . , and Nakajima, S. (1977). J. Gen. Physiol. 70, 1-21. Lau, Y. H., Caswell, A. H., Garcia, M., and Letellier, L. (1979). J . Gen. Physiol. 74, 335-349. Ling, G . , and Woodbury, J. W. (1949). J. Cell. Comp. Physiol. 34, 407-412. Maughan, D. W. (1973). J. Membr. Biol. 11, 331-352. Miller, J. P., Wallace, A. G., and Feezor, M. D. (1971). J. Mol. Cell. Cardiol. 2, 3-19. Mobley, B. A., and Eisenberg, 9.R. (1975). J. Gen. Physiol. 66, 31-45. Nakajima, S., Nakajima, Y., and Peachey, L. D. (1973). J. Physiol. (London) 234,465-480. Noble, D. (1965). J. Cell. Comp. Physiol. 66 (Suppl. 2), 127-136. Page, E., and Storm, S. (1965). J. Gen. Physiol. 48, 957-972. Peachey, L. D. (1965). J. Cell B i d . 25, 209-231. Schneider, M. F., and Chandler, W. K. (1976). J. Gen. Physiol. 67, 125-163. Thomas, R. (1969). J. Physiol. (London) 201, 495-514. Thomas, R. C. (1972). Physiol. Rev. 52, 563-594. Vassalle, M. (1965). Am. J. Physiol. 208, 770-775. Vassalle, M. (1970). Circ. Res. 27, 361-377. Weidmann, S. (1956). “Elektrophysiologie der Herzmuskelfaser.” Huber, Bern, Switzerland. Wiggins, J. R., and Cranefield, P. F. (1976). Circ. Res. 39, 466-474.
CURRENT TOPICS IN MEMBRANES A N D TRANSPORT, VOLUME 16
Chapter 3
The Electrogenic Pump in the Plasma Membrane of Nitella ROGER M . SPANS WICK Section of Plant Biology Division of Biological Sciences Cornell University Ithaca, New York
I. Introduction ....................................................................................... Evidence for the Electrogenic Pump ......................................................... 111. Identity of the Pumped Ion ....................... ... IV. Energy Source for the Electr ........................................... V. Dependence of the Membrane Potential on External and Internal pH ............... VI. The Relationship between the Electrogenic Pump and the Membrane Conductance ....................................................................... VII. Conclusions ...................... ............. ...... ...................... References ....................... ....................................................... 11.
1.
35 36 37 38 39 42 44 45
INTRODUCTION
Giant algal cells, because of their size, have been attractive to electrophysiologists as experimental material for about 60 years. Several early attempts t o measure membrane potentials in Nitella were made in the laboratories of Umrath and Osterhout in the 1920s and 1930s (see Hope and Walker, 1975, for an historical survey), and Cole and Curtis (1938) made the first measurements of conductance changes during an action potential using Nitelfa. There followed a hiatus, partly due to competition from the giant squid axon, which ended in 1955 when Walker made the first measurements of the membrane potential in Nitella using a modern amplifier and microelectrode techniques. Subsequent attempts to explain the electrical properties of the membranes were heavily influenced by the current successes of animal physiol35 Copyright @ 1982 by Academic Press, Inc. All rights o f reproduction in any form reserved. ISBN 0-12-153316-6
36
ROGER M. SPANSWICK
ogists. Working with nerve and muscle cells, they were able to assume the presence of neutral ion pumps and explain the electrical properties of the membranes in terms of passive ion fluxes (Dainty, 1962). Initial experiments on Chara australis suggested that the electrical properties of plant cell membranes might also be accounted for by the passive diffusion of ions, at least in the absence of external Ca2+ (Hope and Walker, 1959). However, in the presence of Ca2+,it soon became obvious that the membrane potential did not respond to changes in external ion concentrations in the manner expected of a diffusion potential (Kishimoto, 1959; Spanswick el al., 1967). The work of Etherton and Higinbotham (1960) on higher plants and of Slayman (1965a,b) on Neurospora had drawn attention to the presence of electrogenic pumps in these systems. Hope (1965) observed a large hyperpolarization in C. australis on addition of bicarbonate, but this was later shown to be an effect due to the change in external p H rather than to an electrogenic HC0,- influx (Spanswick, 1970a). Meanwhile, Kitasato (1968) published a detailed investigation of the effect of external p H on the membrane potential of Nitella clavata. He observed a large effect of pH on the membrane potential and postulated that it was indicative of a high permeability to H +. To explain how the membrane potential could be maintained at a level considerably hyperpolarized relative to the equilibrium potential for H +, Kitasato postulated that the large passive influx of H + was compensated for by an active electrogenic efflux. Thus arose the first detailed hypothesis for an electrogenic H + pump in Nitella. It seems highly probable that the large unlinked fluxes of HC0,- and OH- are also electrogenic (Lucas, 1976). However, attention here will be confined to the electrogenic phenomena observed in the absence of COz. II.
EVIDENCE FOR THE ELECTROGENIC PUMP
The simplest and least ambiguous demonstration of an electrogenic pump in a cellular system is obtained if conditions can be arranged such that the membrane potential is hyperpolarized beyond the limits of the diffusion potential. In Nitella translucens the negative limit of the diffusion potential is set by EK, the Nernst potential for K + (Spanswick and Williams, 1964). As the external K + concentration is raised from 0.1 to 1.O m M there is very little effect on the membrane potential, and at 0.5 m M K + the potential is about 40 mV more negative than EK(Spanswick, 1972). Further increase in the external K + concentration leads to a depolarization of the membrane to EK. The accompanying increase in membrane conductance (Spanswick, 1972) and in K + fluxes indicates that this is due to an increase in potassium permeability. This phenomenon is used routinely to
3. ELECTROGENICITY IN
Nitella
37
establish EK for individual cells and provides a reference point independent of electrode tip potentials. Hyperpolarization of the membrane potential beyond the limit of the diffusion potential has also been observed in Nitella axilliformis (Saito and Senda, 1973a,b), Charu corallina (Richards and Hope, 1964; Keifer and Spanswick, 1978), and Chara braunii (Oda, 1962). The effects of temperature (Spanswick, 1972; Saito and Senda, 1973b) and inhibitors (Kitasato, 1968; Spanswick, 1973, 1974a; Saito and Senda, 1973b, Richards and Hope, 1974; Keifer and Spanswick, 1978) are consistent with the presence of an electrogenic pump in the membrane, though the interpretation of inhibitor experiments has often been based on unjustified extrapolation of results obtained with isolated organelles (Section IV).
111.
IDENTITY OF THE PUMPED ION
Most of the evidence for H + as the ion pumped electrogenically is circumstantial. Kitasato (1968) postulated that such a pump would be needed to excrete the ions that entered passively. This conclusion was based on the assumption that the response of the membrane potential to external pH was indicative of a high passive permeability to H + . However, Spanswick (1972) offered an alternative explanation of the pH response based on Rapoport’s (1970) theory on a voltage-dependent electrogenic pump in which the effect of external pH was on the “pump” electromotive force (EMF) rather than the diffusion potential (Section V). This hypothesis remained consistent with an electrogenic H + pump. It was also possible to rule out both the C1- pump because the membrane potential was not depolarized immediately in the absence of external C1- (Spanswick, 1974a), and the Na+-K+ pump, because ouabain, an inhibitor of this pump (MacRobbie, 1962), also had no effect on the potential. Hope (1965) suggested that HC0,- may be transported inward electrogenically in C. corallinu. However, in N . trunslucens the maximum hyperpolarization occurs in the absence of C0,-HCO). Indeed 1 mMC02-HCOc at pH 6 has an inhibitory effect similar to darkness in this species (Spanswick, 1974a). Attempts to measure the H + efflux directly have met with limited success. Spear et al. (1969) demonstrated acidification of the regions separating the alkaline bands on the surface of N.clavata and estimated that the H’efflux was at least 50 nmoles m-2 sec-’ 20 minutes after the light was turned on. However, it is difficult to judge the accuracy of this estimate. Barr et ul. (1974) also used the efflux of previously accumulated NH, to estimate the H + efflux and obtained a value of 20-40 nmoles m-2 sec-I.
38
ROGER M. SPANSWICK
Further circumstantial evidence for an H + pump comes from measurements of the cytoplasmic pH. When either the weak acid 5,5-dimethyloxazolidine-2,4-dione(DMO) (Walker and Smith, 1975) or glass microelectrodes (Spanswick and Miller, 1977a) are used, the high value for the cytoplasmic pH (7.5) implies that active transport is required to maintain H + far from equilibrium. It has also been shown that in N. translucens in the light the current required to depolarize the membrane potential to E,, the negative limit of the diffusion potential, was equivalent to an efflux of monovalent cations of 200-300 nmoles m-z sec-I (Spanswick, 1972). This is an order of magnitude larger than the fluxes of the major ions, and it seems likely that H + may carry the current. Measurements of the H + efflux from perfused cells may provide the best evidence to date for a link between the electrogenic pump and the H + efflux (Tazawa and Shimmen, this volume). IV.
ENERGY SOURCE FOR THE ELECTROGENIC PUMP
Kitasato (1968) showed that 2,4-dinitrophenol (DNP) produced a significant depolarization of the membrane potential in N. clavata and implied a dependence of the pump on ATP. This proposition was strengthened by more extensive experiments on N . axilliformis (Saito and Senda, 1973a) and N . translucens (Spanswick, 1973, 1974a) using a variety of inhibitors. The most direct demonstration of the dependence of the membrane hyperpolarization on both ATP and Mg2+has been provided by the vacuolar perfusion experiments of Shimmen and Tazawa (1977) using cells of C. australis in which the tonoplast was removed by treatment with EGTA. While the dependence of the pump on ATP is now well established, its relationship to metabolic processes turns out to be more complex than originally suspected. The wavelength dependence of light stimulation of the membrane potential in N. translucens, the sensitivity of the hyperpolarization to carbonyl cyanide m-chloropheilylhydrazone (CCCP), and its insensitivity to 3-(3 ',4'-dichlorophenyl)-l,l-dimethyIurea (DCMU) were consistent with a dependence of the pump on cyclic photophosphorylation (Spanswick, 1974a). However, it is now evident that light stimulation of the electrogenic pump is not a simple response to a change in the free energy of the driving reaction. The first indication that this might be so came from the ATP measurements of Penth and Weigl(l971) on Chara foetida, which showed little change between light and dark. The absence of an effect of light on the ATP level has since been observed in N. translucens (Spans-
3. ELECTROGENICITY IN
Nitella
39
wick and Miller, 1977b) and C. corallina (Keifer and Spanswick, 1979). In N . translucens there is also no effect of 1 m M CO,-HCO,- at pH 6 on the ATP level, though this treatment has an inhibitory effect similar to darkness on the electrogenic pump (Spanswick and Miller, 1977b). However, inhibitors that do reduce the ATP level [CCCP, dicyclohexylcarbodiimide (DCCD)] also inhibit the pump (Spanswick and Miller, unpublished). This suggests that the effects of light and CO, result from the action of control mechanisms which are independent of the ATP level but are presumably related in some way to photosynthesis. The factor responsible for providing the link between the chloroplasts and the plasmalemma has not been identified. However, Tazawa and Shimmen (1980a) have shown that the light-induced potential change in tonoplast-free perfused cells of C. australis is abolished if the chloroplasts are removed by centrifugation, but may be restored by perfusion of the vacuole with a medium containing spinach chloroplasts or chloroplast fragments. The existence of control mechanisms involved in the action of light on the membrane potential of Nitella has also been inferred from measurements of the response of the potential to light modulated sinusoidally over a wide range of frequencies (Hansen, 1978; Martens et al., 1979). Martens et al. (1979) interpret their results in terms of a model that assumes that light acts on the membrane potential via three parallel pathways. It has not yet been possible to identify the mechanisms involved in these regulatory pathways.
V.
DEPENDENCE OF THE MEMBRANE POTENTIAL ON EXTERNAL AND INTERNAL pH
The most interesting and distinctive feature of the membrane potential in characean cells is its strong dependence on external pH in the neutral to acid range (Kitasato, 1968; Spanswick, 1970a, 1972; Saito and Senda, 1973a,b, 1974). Kitasato interpreted this phenomenon to mean that the membrane was highly permeable to H + . However, Spanswick (1972) pointed out some inconsistencies in this interpretation. For instance, inhibition of the pump by DNP produced a depolarization of the membrane potential to EK and not to the H + equilibrium potential which is more positive. Also, the membrane potential does not respond to changes in the external potassium concentration, [K+],, even when the external pH is high and the term P,H; in the equation for the diffusion potential would be negligible compared to PKK;. As an alternative, it was suggested that the response of the membrane potential to external pH was due to an effect on the electrogenic pump
40
ROGER M. SPANSWICK
rather than on the diffusion potential, Instead of the pump being treated as a current source, it was postulated that it was voltage-dependent and would therefore have the property of conductance. As a first approximation it was then possible to consider the behavior of the membrane in terms of the simple equivalent circuit put forward by Finkelstein (1964) and Slayman (1965b) (Fig. 1). The membrane potential predicted by this circuit will be at a value between the diffusion potential, ED,and the pump EMF, E,, which will be determined by the relative values of the conductances of the pump (g,) and passive (gD) channels according to the equation
An expression for E, was obtained by rewriting the equations of Rapoport (1970) for an electrogenic H + pump and imposing the condition of zero flux through the pump. This yields
E, = (A&/uHF)- R T In (HiflH;)
(2)
where Ap, is the free energy of the driving reaction, v, is a stoichiometric coefficient, and [H '1, AND [H1' , are the internal and external hydrogen ion concentrations, respectively. This equation could obviously account for the observed dependence of the membrane potential on [H'], if g, is much greater than g,. Since the passive fluxes of the major ions only account for
IN
FIG.1. An equivalent circuit for the plasma membrane showing the pump EMF, E,, and conductance, g,, in parallel with the diffusion potential, ED,and the passive conductance, gD.
gD
f gp
ED T
T
OUT
EP
3. ELECTROGENICITY IN
Nitella
41
a small fraction of the observed membrane conductance in N. translucens (MacRobbie, 1962; Williams et al., 1964; Spanswick, 1970b), this possibility would also solve another major problem in the electrophysiology of the Characeae. Although Eq. (2) accounts for the pH dependence of the membrane potential, it can only do so over a limited range. This is because the potential begins to depolarize, rather than continuing to hyperpolarize, as the external pH is increased above 8 (Kitasato, 1969; Spanswick, 1972; Saito and Senda, 1973a; Richards and Hope, 1974; Keifer and Spanswick, 1978). Previously it was suggested that this effect and the transient part of the hyperpolarization on changing from low to high pH might be due to an increase in cytoplasmic pH (Spanswick, 1974b). However, subsequent measurements of the cytoplasmic pH over a range of external pH values (Smith and Walker, 1976; Spanswick, unpublished) show that the changes are not large enough for the [H+Iiterm in Eq. (2) to account for the effects. Nevertheless, perfusion of tonoplast-free cells of C. australis with solutions of varying pH (Fujii et al., 1979) shows that the membrane potential becomes more positive by about 20 mV/pH unit as the internal p H is increased from 6 to 9. In intact cells of N. translucens it is possible that the response to cytoplasmic pH is greater in the neutral range. In cells in which the light-stimulated hyperpolarization has decayed spontaneously, the potential can be repolarized by treatment with the weak acid DMO which, at 5 mM, decreased the cytoplasmic p H to 6 at an external pH of 6 (Spanswick and Miller, 1977b). A similar effect is observed when the membrane has been depolarized by a lengthy treatment in C1--free solutions (Spanswick, 1980). One interpretation of this effect is that DMO reverses an alkalinization of the cytoplasm produced by interruption of the Cl--OHexchange system postulated by Smith (1970). However, if the membrane is already hyperpolarized, acidification produces little further effect (Spanswick and Miller, 1977b). These effects, which appear not to correspond to Eq. (2), may indicate the presence of an internal control site sensitive to H +.This is also true for decreases in cytoplasmic pH to values less than 6.0 which, in perfused cells (Fujii el al., 1979), produce marked depolarization of the membrane potential. This observation may provide an explanation for the depolarization of the membrane potential of N. translucens by 1 mM NaN,, which occurs in the absence of any decrease in the cellular ATP level (Miller and Spanswick, unpublished). However, this treatment reduces the cytoplasmic pH to 5.5 (Spanswick and Miller, 1977b), and this would be sufficient to produce the observed depolarization if the results from tonoplast-free perfused cells of C. australis may be extrapolated to intact cells of N. translucens.
42
ROGER M. SPANSWICK
VI.
THE RELATIONSHIP BETWEEN THE ELECTROGENIC PUMP AND THE MEMBRANE CONDUCTANCE
The hypothesis that the electrogenic pump contributes the major part of the membrane conductance has received support from experiments in which it has been observed that factors which inhibit the pump (darkness, C 0 2 , low temperature, CCCP, DCCD) also decrease the membrane conductance (Spanswick, 1972, 1974a; Table I). Although there are some exceptions to this generalization, it also appears t o apply in most cases to C . corallina (Keifer and Spanswick, 1978). It should be noted that the I-V characteristics of these cells are linear over the range of interest in the light and in the absence of inhibitory conditions; thus the decrease in conductance cannot be attributed to nonlinear I-V characteristics. Attribution of the bulk of the membrane conductance to the electrogenic pump has not, however, achieved universal acceptance in spite of the fact that it appears to resolve the problem created by the discrepancy between the measured membrane conductance and the value calculated from the passive major ion fluxes. For instance, F u j i et al. (1979) and Tazawa and Shimmen (1980b) quote evidence that internal perfusion of tonoplast-free cells of C . australis with ATP-free solutions leads to instantaneous depolarization followed by a slower decline in the membrane conductance over the next 10 minutes. They conclude that the pump conductance is not coupled directly to the electrogenic pump activity and infer that g, is much smaller than g,. It is possible that the passive fluxes are higher in perfused than in normal cells but, if not, the problem of accounting for the high value of g, in terms of the passive ion fluxes would remain. However, there is also a poor correlation between the depolarization of the membrane potential and
EFFECT ON
TABLE I MEMBRANE CONDUCTANCE OF TREATMENTS WHICHINHIBIT THE ELECTROGENIC PUMPOF Nitella translucens
THE
Treatment
Conductance (S m2)
Control conductanceb (S m2)
Darkness 1 mM C02-HCOI1 p M CCCP 50 pM DCCD 9.4"C
0.14 0.17 0.15 0.21 0.29
0.83 0.83
~~
Data from Spanswick (1972, 1974a). Control measurements in the light at 20°C.
0.71 0.46 0.59
3. ELECTROGENICITY IN
43
Nitella
the decrease in the membrane conductance produced by inhibitors both in Neurospora (Slayman, 1965b) and C. coraflina (Keifer and Spanswick,
1978), the depolarization being substantially complete before the decrease in conductance begins. Thus the phenomenon may be a general one. At this stage it may be profitable to ask whether there is necessarily a fixed and obligatory relationship between E, and g,. Although it may seem intuitively probable that there should be such a relationship, there does not appear to be a strong theoretical basis for this view. With Rapoport's (1970) theory it is only possible to show that gP =
F2LrrvH
(3)
where L,, is a conductance relating the free energy change of the driving reaction to the rate of the reaction. Changes in g, are therefore dependent on changes in L,, and thus will have no effect on Ep [Eq. (2)]. However, examination of conductance changes accompanying various cases of pump inhibition in N. translucens suggests that the relationship between the membrane potential and conductance may be related to E, and g, in two different ways, depending on whether or not inhibition is due t o a decrease in the ATP level (Spanswick, 1980). In cases where the ATP level does not change (darkness or CO,) and E, may be assumed to remain approximately constant, the curvilinear relationship between the membrane potential and conductance during the course of inhibition may be fitted quite accurately using Eq. (2) and assuming that the change in membrane conductance is entirely due to a change in g,. The values of E p ( - 180 mV) and ED (- 120 mV) necessary to fit the curve for CO, inhibition (Spanswick, 1980) are entirely reasonable, the former being compatible with a value for v H of 2 H+/ATP and the latter being slightly more positive than the value of - 124 mV for the K + equilibrium potential for the same group of cells. In addition, the value of g, (0.08 S m-*) obtained by this curve-fitting process is on the same order of magnitude as the value calculated from the passive ion fluxes (0.05 S m-2; Spanswick, 1970b). Thus it appears that the inhibition by CO, or darkness can be most easily explained by a control mechanism that switches each pump site from a conducting to a nonconducting state. It is not necessary to postulate a change in E,, and this is consistent with the absence of any change in the ATP level (Spanswick and Miller, 1977b). If this interpretation is correct, it appears that g, is proportional t o the number of activated pump sites per unit area. We can now consider what happens in the case of inhibitors which reduce the ATP level. For the effect of CCCP on N. translucens (Spanswick, 1980) or diethylstilbestrol (DES) or CCCP on C. coraflina (Keifer and Spanswick, 1978) the relationship between the membrane con-
44
ROGER M. SPANSWICK
ductance and potential is linear over most of its range. This means that the change in g, is not sufficient to explain the change in the membrane potential according to Eq. (2). However, the simultaneous decrease in ATP level will translate into a change in E, via the effect on the A& term in Eq. (2). This interpretation is supported by a correlation between the depolarization and the decrease in the ATP level produced by CCCP in both N . trunslucens (Spanswick and Miller, unpublished) and C. corulfinu (Keifer and Spanswick, 1979). In C. corullinu, which does not depolarize in the dark, CCCP inhibits both the membrane potential and the ATP level faster in the dark than in the light. According to Eq. (2) the fall in ATP level produces a change in E,. However, there is no obvious reason why this should result in a change in g,. Nevertheless, as the ATP falls below 60% of the control value in CCCP-inhibited cells of N . trunslucens, there is a marked decrease in the membrane conductance (Spanswick, 1980). If the interpretation of the effects of darkness or C 0 2 given above is correct, the decrease in conductance at low ATP levels may indicate that the electrogenic pump is activated by ATP and that this effect is distinguishable from the role of ATP in providing energy to the system. At the biochemical level this could be interpreted in terms of the pump being an ATPase with two binding sites, one a catalytic site with high affinity and the other a lower-affinity allosteric site, with a K , in the millimolar range, which is responsible for activation of the system. This idea was suggested by studies on the kinetics of ATP hydrolysis by ATPases from animal cells (DuPont, 1977; Glynn and Karlish, 1976; Verjovski-Almeida and Inesi, 1979), which gave sigmoidal kinetics that have been interpreted in this way. There appear to be no comparable studies on plants with measurements at sufficiently low concentrations to detect the high-affinity site.
VII.
CONCLUSIONS
The hypothesis that is emerging for the electrogenic pump in Nitellu is that it transports hydrogen ions and is driven by ATP hydrolysis. Application of Rapoport’s (1970) theory to the pump permits the pH sensitivity of the membrane potential t o be attributed to the pump and requires that the pump channels contribute the major part of the membrane conductance. In this way the two main problems in describing the electrophysiology of Nitellu, the insensitivity of the membrane potential to the external concentrations of the major ions and the high value of the membrane conductance, are resolved. Attempts to account for the response of the membrane potential and
3. ELECTROGENICITY IN
Nitella
45
conductance to conditions which produce inhibition of the pump with or without an accompanying change in ATP level led to the conclusion that there are at least two types of control sites. In one case (CO, or darkness), the change in potential can be accounted for by assuming that some factor is produced by the chloroplasts in the light which changes the pump from a nonconducting t o a conducting state without any effect on the pump EMF. In cases in which inhibition is produced by an effect on the ATP level there is a depolarization which may be attributed, at least in the initial stages, to an effect on the pump EMF. There is also a decrease in the membrane conductance, which may be interpreted to indicate that the pump is activated by ATP in addition to being driven by ATP hydrolysis. The most important consequence of the existence of a proton pump in the Characeae is that it becomes possible to use it as the primary active transport process in a chemiosmotic scheme for transport across the plasma membrane. Indeed, Smith (1970) has suggested that the C1- influx is mediated via a Cl--OH- antiport driven by the pH gradient set up by the H + pump. This hypothesis has yet to be tested definitively, a task which is complicated by control mechanisms such as transinhibition (Sanders, 1980). There is also no evidence for the transport of sugars or amino acids via cotransport systems in the Characeae, although these systems are well established for fungi (Slayman and Slayman, 1974), higher plants (Baker, 1978), and Chforeffa(Komor and Tanner, 1974). Nevertheless, the possibility of measuring cytoplasmic p H in these cells (Walker and Smith, 1975; Spanswick and Miller, 1977a) means they probably will provide the most suitable system for testing this scheme. However, the existence of multiple kinetic controls must be taken into consideration, both because they complicate the experimental evaluation of any chemiosmotic scheme and because they must be understood to provide a complete description of ion transport. It is also desirable at this point to begin to develop a kinetic model for the H + pump, since thermodynamic models are unable to take into account the operation of control systems in a useful manner.
REFERENCES Baker, D. A. (1978). New Phytol. 81, 485-497. Bart, C. E., Koh, M. D., and Ryan, T. E. (1974). In “Membrane Transport in Plants” (U. Zimmermann and J. Dainty, eds.), pp. 180-185. Springer-Verlag, Berlin and New York. Cole, K . S . , and Curtis, H. J. (1938). J. Gen. Physiol. 22, 37-64. Dainty, J . (1962). Annu. Rev. Plant Physiol. 13, 379-402. DuPont, Y. (1977). Eur. J. Biochem. 72, 185-190. Etherton, B., and Higinbotham, N. (1960). Science 131, 409-410.
46
ROGER M. SPANSWICK
Finkelstein, A. (1964). Biophys. J . 4, 421-440. Fujii, S., Shimmen, T., and Tazawa, M. (1979). Plant Cell Physiol. (Tokyo) 20, 1315-1328. Glynn, I. M., and Karlish, S. J. D. (1976). J. Physiol. (London) 256, 465-496. Hansen, U.-P. (1978). J. Membr. Biol. 41, 197-224. Hope, A. B. (1965). Aust. J. Biol. Sci. 18, 789-802. Hope, A. B., and Walker, N. A. (1959). Aust. J . Biol. Sci. 14, 26-44. Hope, A. B., and Walker, N. A. (1975). “The Physiology of Giant Algal Cells.” Cambridge Univ. Press, London and New York. Keifer, D. W., and Spanswick, R. M. (1978). Plant Physiol. 62, 653-661. Keifer, D. W., and Spanswick, R. M. (1979). Plant Physiol. 64, 165-168. Kishimoto, U. (1959). Annu. Rep. Sci. Works, Fac. Sci. Osaka Univ. 7, 115-146. Kitasato, H . (1968). J. Gen. Physiol. 52, 60-87. Komor, E., and Tanner, W. (1974). J. Gen. Physiol. 64, 568-581. Lucas, W. J. (1976). J. Exp. Bot. 27, 19-31. MacRobbie, E. A. C. (1962). J . Gen. Physiol. 45, 861-878. Martens, J., Hansen, U.-P., and Warncke, J. (1979). J . Membr. Biol. 48, 115-139. Oda, K. (1962). Sci. Rep. T6hoku Univ. 4th Ser. 28, 1-16. Penth, B., and Weigl, J. (1971). Planta 96, 212-223. Rapoport, S. I. (1970). Biophys. J . 10, 246-259. Richards, J. L., and Hope, A. B. (1974). J. Membr. Biol. 16, 121-144. Saito, K., and Senda, M. (1973a). Plant Cell Physiol. (Tokyo) 14, 147-156. Saito, K., and Senda, M. (1973b). Plant Cell Physiol. (Tokyo) 14, 1045-1052. Saito, K., and Senda, M. (1974). Plant Cell Physiol. (Tokyo) 15, 1007-1016. Sanders, D. (1980). J . Membr. Biol. 52, 51-60. Shimmen, T., and Tazawa, M. (1977). J. Membr. Biol. 37, 167-192. Slayman, C. L. (1965a). J . Gen. Physiol. 49, 69-92. Slayman, C. L. (1965b). J . Gen. Physiol. 49, 93-116. Slayman, C. L., and Slayman, C. W. (1974). Proc. Natl. Acad. Sci. U.S.A. 71, 1935-1939. Smith, F. A. (1970). New Phytol. 69, 903-917. Smith, F. A., and Walker, N. A. (1976). J . Exp. Bot. 27, 451-459. Spanswick, R. M. (1970a). J. Membr. Biol. 2, 59-70. Spanswick, R. M. (1970b). J . Exp. Bot. 21, 617-627. Spanswick, R. M. (1972). Biochim. Biophys. Acta 288, 73-89. Spanswick, R. M. (1973). In “Ion Transport in Plants” (W. P. Anderson, ed.), pp. 113-128. Academic Press, New York. Spanswick, R. M. (1974a). Biochim. Biophys. Acta 332, 387-398. Spanswick, R. M. (1974b). Can. J. Bot. 52, 1029-1934. Spanswick, R. M. (1980). In “Plant Membrane Transport: Current Conceptual Issues” (R. M. Spanswick, W. J. Lucas, and J. Dainty, eds.), pp. 305-313. Elsevier, Amsterdam. Spanswick, R. M., and Miller, A. G. (1977a). Plant Physiol. 59, 664-666. Spanswick, R. M., and Miller, A. G. (1977b). In “Transmembrane Ionic Exchanges in Plants” (M. Thellier, A. Monnier, M. DeMarty, and J. Dainty, eds.), pp. 239-245. CNRS, Paris. Spanswick, R. M., and Williams, E. J . (1964). J . Exp. Bot. 15, 193-200. Spanswick, R. M., Stolarek, J . , and Williams, E. J. (1967). J. Exp. Bot. 18, 1-16. Spear, D. G., Barr, J. K., and Barr, C. E. (1969). J . Gen. Physiol. 54, 397-414. Tazawa, M., and Shimmen, T. (1980a). In “Plant Membrane Transport: Current Conceptual Issues” (R. M. Spanswick, W. J. Lucas, and J. Dainty, eds.), pp. 589-590. Elsevier, Amsterdam. Tazawa, M., and Shimmen, T. (1980b). In “Plant Membrane Transport: Current Conceptual
3. ELECTROGENICITY IN
Nite//a
47
Issues” (R. M. Spanswick, W. J. Lucab, and J. Dainty, eds.), pp. 349-362. Elsevier, Amsterdam. Verjovski-Almeida, S., and Inesi, G. (1979). J. Biol. Chern. 254, 18-21. Walker, N. A. (1955). Aust. J. Biol. Sci. 8, 476-489. Walker, N. A . , and Smith, F. A. (1975). Plant Sci. Left. 4, 125-132. Williams, E. J . , Johnston, R. J., and Dainty, J. (1964). J . Exp. Bot. 15, 1-14.
This Page Intentionally Left Blank
CURRENT TOPICS IN MEMBRANES AND TRANSPORT, VOLUME 16
Chapter 4
Control of Electrogenesis by ATP, Mg*+,H', and Light in Perfused Cells of Chara MASASHI TAZA WA A N D TERUO SHIMMEN Department of Biology Faculty of Science University of Tokyo Hongo, Tokyo, Japan
I. 11.
Ill.
IV.
V. V1. V11.
Introduction ......... ............... Method for Controll usion .......... Dependence of Electrogenesis and Net H + Efflux on Mg-ATP ....................... A. Involvement of Mg. A T P in Electrogenesis .................... Dependence of Electrogenesis on pHi, pH,, and [K+ 1, .............. A. Dependence of Em and R , on Internal p H in the Presence and Absence of Internal A T P . B. Dependence of Em and R , on
49
51 53 53 54 55 55
......................
56 58
D. Analysis of Results via a Linear Equivalent-Circuit Model ....................... Modulation of Electrogenesis by Light ........................... Discussion .................................................................................... Concluding Remarks ......................................................................
59
62 63 65 66
I. INTRODUCTION' Since Hope (1 965) first suggested the existence of an electrogenic anion pump in the plasmalemma of Cham, to account for the hyperpolarizing effect of bicarbonate ions, substantial progress has been made in the study I Abbreviations: AMP-PNP, Adenylyl irnidodiphosphoric acid; APW, artificial pond water; CCCP, carbonyl cyanide rn-chlorophenylhydrazone;CyDTA, 1,2-cyclohexanediarnine-N,
49
Copyright (c) 1982 by Academic Press, Inc All rights of reproduction in any form reserved. ISBN 0-12-153316-6
50
MASASHI TAZAWA AND TERUO SHIMMEN
of electrogenic pumps in plant cells. Although the hyperpolarizing effect of HC0,- was later proved by Spanswick (1970) to be a pH effect on the membrane potential (Em),Hope's proposal of the existence of a new kind of electromotive force at the plasmalemma was quite stimulating at that time, when the Emof characean cells was exclusively considered to be accounted for by the diffusion potential across the membrane. It was Kitasato (1968) who first postulated the existence of an electrogenic H +-extruding pump in Nitellu plasmalemma, an idea which is now widely accepted not only in algal cells but also in fungi and higher plants. Kitasato tried to explain the large difference between the membrane conductance measured electrically (g,) and the sum of conductances calculated from K + , Na+, and C1- fluxes measured isotopically (Cgj). From the fact that Em was extremely sensitive to the external pH, he assumed that the plasmalemma of Nitella has a high H permeability (P,) and that the proton conductance (gH) may be responsible for the discrepancy between g, and Cg,. If P, is 10,000 times larger than P K ,the response of Em to pH, could be accounted for. However, the membrane potential expected from the Goldman equation is 70-80 mV more positive than the measured Em.The difference between the observed and expected values was explained by electrogenicity of the H + pump, acting to extrude H + which had entered passively from the external medium. This idea implied that the plasmalemma is extremely permeable to H +, whether or not the electrogenic pump is actually working. Kitasato's contention was questioned by Saito and Senda (1973b), who worked with Nitellu flexilis and N. uxilliformis. In these species Emis highly sensitive to pH, in light, but becomes insensitive in the dark, at low temperatures or in the presence of inhibitors such as DCMU and DNP. Thus the strong pH, dependence of Emdoes not necessarily represent a high proton permeability of the membrane but could reflect the dependence of electrogenic pump activity on pH,. Similar results were obtained on Churu corallinu by to Richards and Hope (1974), who calculated the permeability ratio PH/PK be only about 25. In Neurosporu hyphae, [ATP], decreased upon treatment with N2,CN-, or NaN,. At the same time, the net H + efflux decreased and the membrane was depolarized (Slayman, 1970). From the parallel behavior of Em and the intracellular level of ATP, Slayman et ul. (1973) concluded that an H +-extrusion pump fueled by ATP exists in the Neurosporu plasmalemma. Keifer and Spanswick (1978) found the membrane of C. corullinu to be strongly depolarized by treatment with membrane ATPase inhibitors such as DCCD or DES. Since residual membrane conductance (g,) amounted to +
"-tetraacetic acid; DCCD, dicyclohexylcarbodiimide; DCMU, 3-(3 ',4'-dichlorophenyl)-l, 1-dimethylurea; DES, diethylstilbestrol; DNP, 2,4-dinitrophenol; EDTA, ethylenediaminetetraacetic acid; EGTA, ethylene glycol-bis(P-aminoethyl ether) N,N' -tetraacetic acid; HK, hexokinase; LPC, light-induced potential change.
4. CONTROL OF ELECTROGENESIS IN
51
ChaEI
less than 10% of the normal value, they attributed most of the normal g, to the pump conductance (gJ. A major difficulty in all the studies thus far has been that both of the and ATPprincipal substrates for the putative electrogenic pump-H have been implicated only from indirect experiments. To overcome this difficulty, Em must be measured under conditions in which the intracellular ATP concentration and pH are known and controlled. Attempts t o control the chemical composition of the cytoplasm of plant cells were made by Williamson (1975) and Tazawa et al. (1976) using internodal cells of Characeae. The cell models developed by' both groups are similar in that the vacuolar membrane or tonoplast is removed by vacuolar perfusion with a medium containing the Ca2+-chelatingagent EGTA. Subsequently, solutions with a wide range of compositions can be perfused and gain rapid access to the plasmalemma. Details of the perfusion method are presented below. +
II. METHOD FOR CONTROLLING INTRACELLULAR ENVIRONMENT BY INTERNAL PERFUSION
Internal perfusion of characean cells is carried out by cutting both cell ends and forcing the perfusion medium through with a hydrostatic pressure difference between the ends (Tazawa, 1964). As long as the perfusion medium contains Ca2+,the tonoplast remains intact; but when the perfusion medium contains several millimoles of EGTA, to chelate the Ca2+,the tonoplast disintegrates over periods of 3-30 minutes (Tazawa et al., 1976). The compositions of perfusion media used for preparing tonoplast-free cells are listed in Table I. For convenience, cells perfused with the respecTABLE I COMPOSITION OF INTERNALPERFUSION MEDIA Medium composition Component
Mg
EGTA CyDTA M a 2 Hexokinase Glucose Buffer PH KOH Sorbitol
5 mM 5 m M 6 mM 6 mM 6 mM 1 mg/ml 5 mM Tris-maleate, PIPES, or HEPES, 5-30 mM 7.0 mM 7.0 mM 7.0 mM 7.0 mM Between 17 and 71 mM, according to buffer strength Added to adjust total osmolarity to 3 3 0 mM
5 mM
-
Mg. ATP 5 m M
HK
CyDTA
52
MASASHI TAZAWA AND TERUO SHIMMEN
tive perfusion media are labeled with the name of the medium: e.g., Mg-ATP cells or HK cells. HK and CyDTA media are used to remove ATP and Mg2+, respectively. The concentrations of diffusible molecules in the cell after disintegration of the tonoplast (Fig. 1B) can be estimated from the concentrations of substances in the cytoplasm plus the volume ratio between the cytoplasm and the whole cell, which is 1:lO in Chara australis (Tazawa et al., 1974). The concentrations of K, Na, and C1 in the cytoplasm of C. australis are 112, 3, and 21 mM, respectively (Tazawa et al., 1974). The concentration of total Ca is 3.5 mM, while that of free Ca2+ is estimated as less than le7M (Tazawa et al., 1976). The normal level of ATP in cytoplasm is 0.5-1.3 m M (Kikuyama et al., 1979). Since the Mg medium contains no ATP, the concentration of ATP after loss of the tonoplast should be 0.05-0.13 mM. Cells with this level of [ATPIi still maintain a high Emand excitability, like normal cells or cells perfused with Mg-ATP medium (Shimmen et al., 1976; Tazawa et al., 1976; Shimmen and Tazawa, 1977). More exact control of internal chemical composition requires reperfusion with fresh artificial medium. During this second perfusion, most of the endoplasm effuses out (Fig. IC), although the sheet of chloroplasts attached to the cortical gel (Fig. 1C) may still interfere with exact control of the chemical composition near the membrane. The chloroplasts can be removed by centrifuging tonoplast-free cells at 7000- 15000 g for several
A
C
,W
/
-P \C
0
0 3
0
C
gi
0
t‘
v S
-9
B
D -.
- ._
FIG. 1. Schematic representation of the longitudinal section of decorticated characean cells. (A) Normal cell: (B) tonoplast-free celf after the first perfusion with medium containing EGTA; ( C ) tonoplast-free cell after reperfusion: (D) tonoplast-free cells without chloroplasts. w, Cell wall; p, plasmalemma; c, chloroplast; t, tonoplast; v, vacuole: s, sol endoplasm; g, gel ectoplasm; ef, endoplasm fragment. (From Tazawa and Shimmen, 1979.)
4. CONTROL OF ELECTROGENESIS IN
Cham
53 T
FIG.2. Vessel used for the open-vacuole method of measuring Em. The cell (Cha) is open to the external media in pools A and C. Pool B is filled with isotonic APW containing 0.1 mM each of KCI, NaC1, and CaCI,, and sorbitol (Shimmen and Tazawa, 1977).
K
/
n minutes. The resultant chloroplast-free and transparent cells (Fig. lD), when perfused with Mg-ATP medium, maintain the normal level of Emand remain excitable. In simple measurements of Em using glass microelectrodes, the cells are ligated after perfusion, but in most experiments employing the openvacuole method both ends of the cell are kept open to the medium (Fig. 2; Tazawa et al., 1975). The membrane potential for the cell segment in pool B can be measured as the potential difference between electrodes immersed in B and either side pool, A or C. The external medium for the microelectrode method is artificial pond water (APW) containing 0.1 mM each of KCl, NaCl, and CaC12, (PH 5.6) and that for the open-vacuole method is isotonic APW whose osmolarity is adjusted with sorbitol.
111.
DEPENDENCE OF ELECTROGENESIS AND NET H + EFFLUX ON Mg*ATP
A. Involvement of Mg*ATP in Electrogenesis The membrane potential of the Chara plasmalemma is dependent on both [ATP], and [Mg2+],(Shimmen and Tazawa, 1977). Perfusing the cell with the HK medium reduces [ATP], to less than 1 pM (Kikuyama et al., 1979) and simultaneously shifts Em from -200 mV to about - 100 mV. Thus, roughly half the membrane potential of the Chara plasmalemma is ATP-dependent. When Mg2+alone is depleted from the cell interior by perfusion with EDTA medium instead of Mg medium, Emshifts from - 200 to about - 120 mV, so that both ATP and Mg2+are essential for maintenance of the normal Em. The reversible changes in Em with depletion and addition of ATP can readily be observed during continuous perfusion (Shimmen and Tazawa, 1977). As shown in Fig. 3A, at the onset of perfusion with Mg medium (lacking ATP), Em shifts abruptly in the depolarizing direction and soon
54
MASASHI TAZAWA AND TERUO SHIMMEN
attains a steady level. Then, upon reperfusion with Mg-ATP medium, the membrane repolarizes within seconds. Such ATP-dependent transitions between the polarized and depolarized levels can be repeated many times and are essentially identical whether the alternate medium contains Mg or not (CyDTA medium). Repolarization, however, does not occur with a n unhydrolyzable analog such as AMP-PNP (Fig. 3B), suggesting that the energy released by ATP hydrolysis in fact fuels the putative electrogenic ion pump. B. H + Efflux in Relation to [ATP],
No previous direct evidence has existed in characean cells to show that H + is the ionic species actually carried across the plasmalemma by the elec-
Perfusion ATP
I
-
Perfusion ATP
-
+-
-
I
-
-
+
-
-
+ - +
-
-
-
+
--
i+
-
-
+
-
-
-
lmin
I
Imin
MgAMP-PNP
FIG.3. Reversible transition of Em between polarized and depolarized levels by controlling [ATPI,. At the beginning of each recording, the cell contained Mg’ATP medium. T o remove ATP, the cell was perfused with Mg medium. (A) Alternation of Mg medium with Mg’ATP medium. (B) Similar t o (A), but with one test of the nonhydrolyzable A T P analog AMP-PMP. The external medium was isotonic APW to which 5 mM lead acetate had been added to suppress the action potentials (Shimmen and Tazawa, 1977). The E, at the start of record (A) was -202 mV; at the start of record (B), - 184 mV. Brief square pulses of current [shown by the uppermost trace, for (A)] were passed through the membrane, and the corresponding voltage displacements were taken as the measure of membrane resistance R,. Pulse repetition frequency 10 min-’ in (A); 15 min-’ in (B).
4. CONTROL OF ELECTROGENESIS IN
Chara
55
TABLE I1 NET H + EFFLUX IN RELATION TO INTRACELLULAR ATP CONCENTRATION ([ATPIi) [ATPli Perfusion medium
0
Exp. 1
Exp. 2
Exp. 3
HK Mg. ATP
1
22 f 11 (4)
387b
19 f 6 (8) 61 f 16 (9)
18 f 5 (7) 53 f 24 (7)
84
f
46 (4)
Values are in nanomoles per meter squared per second f 1 SE. The number of cells in each experiment is shown in parentheses. Thirty-minute incubation in APW; unpublished ATP data.
trogenic pump, although the evidence of Slayman (1970) on fungal cells has been mentioned above. It has been our objective with perfused cells therefore to check the H + hypothesis by measuring the net efflux of H + in the presence and absence of perfusing ATP. Our experimental procedure is as follows (Shimmen and Tazawa, 1980). Chara cells are first perfused with HK medium or Mg-ATP medium to remove the tonoplast. They are then incubated for 20-30 minutes in the dark, which allows the pump to operate (in the presence of ATP) without CO, assimilation and its resultant alkalinization (Lucas and Smith, 1973). The incubation medium is APW (pH 7.0), to which phenol red has been added; H + efflux is calculated from the volume of 1 mMNaOH required to restore the pH of the medium to 7.0. Net fluxes measured in this fashion are listed in Table I1 for both the ATP-free (HK) and the ATP-containing perfusions. Clearly, the net H + efflux is lower in HK cells than in Mg-ATP cells, the difference between the presence and absence of ATP being 40-60 nmoles m-2 sec-I, or 4-6 mA/m2.
IV.
DEPENDENCE OF ELECTROGENESIS ON W+Io PHi, P H ~ AND ,
A. Dependence of Emand R, on Internal pH in the Presence and Absence of Internal ATP The pH sensitivity of the electrogenic H+-ATPase was studied in isolated plasmalemma vesicles from Neurospora hyphae (Scarborough, 1976, 1977; Bowman and Slayman, 1977). Plasmalemma vesicles have not been isolated from characean cells. However, the pH dependence of the electrogenic ATPase can be studied without using membrane vesicles, since (in the tonoplast-free cell model) the plasmalemma forms large cylindrical
56
MASASHI TAZAWA AND TERUO SHIMMEN
membrane sheets and the pH of the internal solution facing the active component of the membrane ATPase can be freely modified during measurements of Emand R, (Fujii et al., 1979). Figure 4A shows the changes in Emin the presence and absence of ATP at various pHi values with pH, constant at 5.6. During preparation and measurement, the cells were kept under diffuse light of intensity below that required for the light-induced change in Em(which will be described later). In the presence of ATP, Emis maximal at pHi = 7, decreasing with either acidic or alkaline shifts. But in the absence of ATP, Emis low at low pHi and rises linearly with increasing pHi. We assume that the difference between Em(+ ATP) and Em(- ATP) represents the portion of Emsupported by ATP. This fraction, designated E,, is also maximal at neutral pH. The membrane resistance (R,), too, is dependent on pHi (Fig. 4B), but in a manner which does not correlate with E, or the activity of the electrogenic pump. It is high at low pHi and low at high pHi. B. Dependence of E, and R, on External pH in the Presence and Absence of ATP Many studies have been carried out to observe the effect of external pH (pH,) on the membrane potential in characean cells. Em diminishes (becomes less negative) as the pH, falls, and the change sometimes amounts to 50-60 mV per pH unit. Kitasato (1968) interpreted this fact to reflect a of the plasmalemma. Other investigators (Saito high H permeability (PH) and Senda, 1973b; Richards and Hope, 1974) have also found the pH dependence of Em to be strongly suppressed by metabolic inhibitors or +
PHi
5
6
7
8
9
B
-> -
- -a
“E
E
E -100-
W
-200-
.
10I L -
05
6
7
8
9
PHI
FIG. 4. Dependence of membrane properties on pHi in the presence (+ ATP) and absence (-ATP) of ATP. (A) Em; (B) R,. E, is the difference between E,(+ATP) and Em(-ATP) (Fujii et al., 1979).
4. CONTROL OF ELECTROGENESIS IN
57
Cham
lowered temperature, so there seems little doubt that the portion of E, which depends on pH, is the active, metabolically dependent component, not the passive component. Consistent with this interpretation is the fact that light of moderate intensity makes E, more negative (Nagai and Tazawa, 1962; Nishizaki, 1968; Spanswick, 1972; Saito and Senda, 1973a), in a manner suggesting activation of the electrogenic pump. Recently, we have measured the pH, dependence of Em and R , in perfused tonoplast-free Chara cells in the presence and absence of internal ATP (Kawamura et al., 1980). After perfusion with Mg-ATP medium, E, responds to changes in pH, under both light and dark conditions (Fig. 5A), though the effect is stronger in the light than in the dark. In contrast, when [ATP], is reduced below 1 pA4 (perfusion with HK medium, Fig. 5B; Kikuyama et al., 1979), the pH, dependence of Em is greatly reduced and the light-dark difference is abolished. As expected, R , is sensitive to pH, in ATP-perfused cells (Fig. 5C)-where it rises in parallel with pH,-but is almost insensitive to pH, in ATP-free cells (Fig. 5D). Essentially identical results of substrate withdrawal are observed upon perfusion by CyDTA medium to remove Mg2+,(data not shown), instead of by HK medium to PHo
PHO
4 ,
0
2
5
6
7
8
9
-
B
-2 -
E W
-100-
-200-
I
4
5
6
PHO
7
0
9
,
2‘
4
5
6
7
8
9
PHo
FIG.5 . Dependence of membrane properties on pH, under light (L) and dark (D) conditions. (A) Em in Mg-ATP cells; (B) E, in Mg cells; (C) R , in Mg-ATP cells; (D) R , in Mg cells (Kawamura et al., 1980). Note that the voltage scale in (B) is twofold expanded from that in (A).
58
MASASHI TAZAWA AND TERUO SHIMMEN
remove ATP. We interpret this ensemble of results to mean that the conspicuous pH, dependence of both Emand R, is closely related to activity of an electrogenic pump which can be inhibited by depletion of either ATP or Mg2+.
C. Dependence of Emand R, o n External K + Concentration Since E,, the ATP-dependent part of Em,is strongly influenced by [H'],, it would be interesting to know how the ATP-dependent electrogenesis is modified by other external cations, particularly [K+],. An experiment for studying this was conducted by Shimmen and Tazawa (1977). As shown in Fig. 6A, Em in the absence of ATP (HK cells) is slightly more sensitive to [K+], than Em in the presence of ATP (Mg cells), suggesting that the K + permeability (PK)of the plasmalemma not only does not decline on remov-
CK'I,
(mM)
1.0
10 I
I
B
-200
t
A 1
0.I
1.0 CK+l, (mM)
I
10
FIG.6 . Dependence of membrane properties on [K+1, in the presence and absence of internal ATP. (A) Em;(B) R,. Values are 1 SE. Cells containing or lacking ATP were obtained by internal perfusion with Mg or HK medium, respectively. [ATP]i estimated to be 0.05-0.13 m M for Mg medium, and less than 1 pM for HK medium (Kikuyama ef al., 1979). Ea, obtained as the difference Em(+ ATP) minus Em (- ATP), is the component of Em supported by ATP. EK is the equilibrium potential for K + ; [K+Ii = 27 mM. (From Shimmen and Tazawa, 1977.)
*
Char&'
4. CONTROL OF ELECTROGENESIS I N
59
ing ATP but actually increases relative to the permeabilities of other ions. This increased K + sensitivity may be associated with a general elevation of R , seen upon ATP withdrawal (Fig. 6B) and a consequent enhancement of the relative effect of K + conductance with increasing [K'],. E, itself, however, is insensitive to [Kilo, at least over the span 0.1-10 mM. This finding contradicts the argument (Spanswick, 1972; Richards and Hope, 1974) that at higher [K+l,., where E, is equal to EK,the electrogenic pump does not function. D. Analysis of Results via a Linear Equivalent-Circuit Model 1. PUMPCONDUCTANCE (g,)
AND
PUMPELECTROMOTIVE FORCE (E,)
Since the above results all support the notion that the plasmalemma of Cham contains both a passive, ion diffusion regime and an active, H + extrusion pump, the membrane-equivalent circuit should contain at least two parallel limbs: subscripted d for the diffusion regimes, and p for the pump pathway (Fig. 7). (The EMF and conductance of the two channels
in
P---------
out FIG. 7 . An equivalent-circuit model for the Chum plasmalemma. E , EMF; g, conductance; I, current. Subscripts d and p indicate passive diffusion channel and pump channel, respectively. The overall membrane parameters, Em and g, = l / R , , are given by Eqs. (1) and (2).
60
MASASHI TAZAWA AND TERUO SHIMMEN
are designated by E and g, respectively.) From this model, with zero net current flow,
in which I, is the pump current flowing across the pump channel from inside to outside. Because net current flow through the membrane is assumed to be zero, the same amount of current must flow (in the opposite direction) across the diffusion channel. Assuming that g, and Em,in the absence of ATP, are equal to g, and Ed, respectively, it follows that gp and E, can be calculated from Eqs. (1) and (2), while I, can be calculated from Eq. (3). The data on the pH, dependence of Emand R , (Fig. 5 ) have been analyzed via Eqs. (1) and (2), and the three conductances (g", gd, and 8,) are displayed in Fig. 8A. Here g, is smaller than gd for pH, in the range 8.4-5.4 but becomes larger at pH, = 4.4. In Fig. SB, E, is plotted against pH, and is compared with the equilibrium diffusion potential for H + ( E H ) . Ep is highly negative at high pH and less so at low values. When EHis zero (pH, = 7.0), Ep = -490 mV, which should represent the so-called phosphate potential, or the free energy of ATP hydrolysis [cf. Eq. (4)]. A similar analysis has been carried out on the data of Fig. 4 in order to show the dependence of E, and R, on the internal pH. Once again g, is
mV
A
I.Or
-
N
E
3 0.5 0
-
,
-
0
4
5
6
7
8
9
- 800L
pH0
FIG.8. Dependence of the membrane's linear equivalent circuit elements on pH,. (A) Conductances; (B) EMFs. Most symbols are defined in the legend to Fig. 7 . EH,Equilibrium diffusion potential for H + ; g, and gd, measured values of membrane conductance in the presence and absence of internal ATP, respectively. ED was calculated from Eq. (2). Data from Fig. 5 .
4. CONTROL OF ELECTROGENESIS IN
Chara
61
smaller than g,-in this case at all pHi values tested (Fig. 9A). Here g, is nearly zero at both alkaline (pH 8.7) and acidic (pH 5.6) values of pHi, where pump electrogenicity clearly is inhibited (Fig. 4). Ep is most negative at low pHi, as would be expected for a pump carrying H + ions outward, and becomes less negative at high pHi (Fig. 9B). E, equals - 470 mV at the point (pHi = 5.6) where EHis zero. This is essentially the same estimate of the phosphate potential as the one we obtained above, In summary, the pump conductance in Chara accounts for one-fourth to one-third of the total membrane conductance in the physiological range for both internal and external pH. This finding stands in striking contrast to that of Keifer and Spanswick (1978) on N. transfucens, where the operating pump seems to account for about 90% of the total membrane conductance. 2. STOICHIOMETRY BETWEEN H + TRANSPORT AND ATP HYDROLYSIS
The reversal potential or equilibrium potential (E,) for the electrogenic pump can be calculated as
where n is the number of H + ions transported by splitting one molecule of ATP. Assuming the standard free energy of ATP hydrolysis to be -8.54 kcal/mole (pH 7; Slayman et af., 1973) and setting EHto zero @Hi = pH,),
0,
-0.1
L
mV
PHi -6OOL
FIG.9. Dependence of the membrane's equivalent circuit elements on pHi. (A) Conductances; (B) EMFs. Description as for Fig. 8. Data from Fig. 4.
62
MASASHI TAZAWA AND TERUO SHIMMEN
4
5
6 pH0
7
8
9
5
6
7
8
9
PHI
FIG. 10. Dependence of the pump current (Ip)on pH. (A) External pH, in light (L) and dark (D). Ip calculated from data in Fig. 5A and C. (B) Internal pH. Ip calculated from data in Fig. 5B and D.
When Chara cells perfused with MgeATP medium containing 1 mM ATP are incubated in APW, [ATP], decreases to 0.39 f 0.04 mMin 30 minutes (Table 11). Cytoplasmic adenylate kinase leaves some uncertainty in the actual concentrations of ADP and Pi, but the limits appear to be A[ADP], = 0.61 m M = [Pi] = [ADP], at the most, and [Pi] = 1.12 mM, [ADP], = 0.1 mM, at the least. Then Eq. (4) gives 544 I -Ep 5 573 mV. Since Epwas calculated to be - 470 to - 490 mV from the electrical data (Figs. 4, 5, 8B, and 9B), it is reasonable to assume unity for n. 3 . PUMPCURRENT, Zp
As shown in Fig. 10A, Zp [calculated via Eq. (3)] increases with increasing pH, under both light and dark conditions. It is, furthermore, larger in the light than in the dark at all values of pH, tested. As shown in Fig. 10B, on the other hand, variations in internal pH at constant external pH (5.6) give a distinct pH optimum for Zp. This result is consistent with the fact that maximal activity of the plasmalemma ATPase from Neurospora occurs in the range pH 6.5 as determined by Scarborough (1977), to pH 6.7, as determined by Bowman and Slayman (1977, 1979). (The latter authors also found that lower apparent pH optima could arise from the contamination of reagent ATP by traces of vanadate.) V.
MODULATION OF ELECTROGENESIS BY LIGHT
As shown in Fig. 5A, Em is lower (less negative) in the dark than in the light. Since the light-induced potential change (LPC) is completely blocked by an inhibitor of noncyclic electron flow in chloroplasts, DCMU
4. CONTROL OF ELECTROGENESIS IN
Chara
63
(Nishizaki, 1968), it is reasonable t o expect some kind of cytoplasmic factor, produced by photosynthesis, to modulate the pump activity. Further supporting evidence comes from the fact that LPC develops slowly after the light is switched on, requiring more than 10 minutes to reach a steady level (Nishizaki, 1968). Kikuyama et al. (1979) observed that LPC was abolished by ATP-free perfusion of Chara, which raises the question whether ATP itself might be the modulating factor. But the same authors dismissed this hypothesis upon observing that cells saturated with ATP (5 mM) still exhibited LPC. The possibility that a light-induced change in pHi, resulting from chloroplast activation, might modulate the plasmalemmal H pump has also been explored and can be rejected on the basis of several observations. First, although the cytoplasmic pH of Chara cells (pH, = 5-6) increases slightly upon illumination [from 7.4 in the dark to 7.7 in the light (Walker and Smith, 1975)], the pH, dependence of electrogenesis in perfused Chara (Fig. 4A) indicates that any such pH change would be far too small to account for the light-dark difference in Zp(Fig. 10A). Second, LPC occurs in cells perfused with MgSATP medium strongly buffered with 100 mM HEPES (Fujii et al., 1979), as well as in cells perfused with weakly buffered medium (5 m M Tris-maleate; Tazawa et al., 1979). And finally, LPC is observed over a wide range of pHi, from 5.1 to 8.7. It is obvious, therefore, that the search must be continued for a chemical factor mediating the light-induced potential change. +
VI.
DISCUSSION
The experimental results obtained using tonoplast-free Chara cells convince us that the Chara plasmalemma is equipped with an electrogenic H +-extruding pump fueled by Mg*ATP. Quantitatively, however, several problems remain. The first is whether or not the ATP-dependent electrogenesis of the Chara membrane is completely accounted for by activity of the H+-extruding pump alone. The decrease in net efflux of H + upon removal of ATP is 40-60 nmoles m-z sec-1which is equivalent to a current density of 4-6 mA/mZ. On the other hand, the pump current (ZJcalculated under the same conditions as those used to measure the H + flux (i.e., pH, = 6.9, pH, = 7.0)-is 11 mA/mZ, or more than twice as large as the net H + efflux. There are at least two possibilities to explain this discrepancy: Either a H t influx (or a OH- efflux) may occur or a second electrogenic pump, moving ions other than H + , may exist. One candidate for the latter would be a pump actively transporting anions inward. But the facts (1) that Cl--free solutions cause a very slow depolarization in N. translucens (Spanswick, 1973) and (2) that the pH, dependence of the
64
MASASHI TAZAWA AND TERUO SHIMMEN
vacuolar potential in C. cordina is not changed by replacing C1- with Sod2(Richard and Hope, 1974; Spanswick, 1974; Smith and Walker, 1976), indicate that any such C1- electrogenesis is insignificant. The large H + efflux in light, calculated from Ipin Fig. 10A to be about 300 nmoles m-2 sec-I near pH, = 7, has never been measured quantitatively, although Spear et al. (1969) estimated the H + efflux from illuminated cells of N. clavata to be more than 200 nmoles m-* sec-I, based on a color change of external solutions containing the p H indicator, phenol red. As in the case of N. clavata, Lucas and Smith (1973) demonstrated formation of alkaline and acid regions at the surface of C. corallina. From the exact measurement of the pH profile near the cylindrical cell surface, Lucas (1975) could estimate the OH- efflux during illumination to be 250 nmoles m-2 sec-I in the presence of HC0,-. The magnitude of the H + efflux could not be calculated, owing to the mathematical complexity of the acid diffusion profiles, but Lucas (1975) suggested that the value of the H + efflux (occurring from the acid cell surface) might be significantly lower than that reported by Spear et a/. (1969). If this were the case in C. australis, the pump current in light-amounting to 30 mA/m2 (Fig. 10A)-could not be explained solely by H + efflux. We therefore regard the hypothesis that all the ATP-dependent electrogenesis is produced by the H + -extruding pump as not yet proven. In Section IV,D, the experimental data were analyzed on the basis of an equivalent circuit model (Fig. 7) having a pump channel in parallel with a diffusion channel. Since we had no appropriate means to determine the parameters of each channel separately, we assumed that the passive parameters, Ed and gd, are unaffected by the action of the pump and are equal to the membrane potential and membrane conductance measured when the pump is completely stopped by the depletion of ATP. Normally, Em and R, of tonoplast-free Chara cells are measured after confirming loss of the tonoplast, i.e., 10-30 minutes after the end of perfusion. Values of g, (l/R,) for cells lacking ATP (Fig. 5B) are smaller than those for cells containing ATP (Fig. 5A). The difference in g, can be explained by assuming that gp is zero when the electrogenic pump stops. There are, however, several pieces of evidence showing that electrogenic pump activity is not directly coupled with the conductance change. Published data (Slayman, 1965; Keifer and Spanswick, 1978) show that there can be some delay in the decrease of g, when Emis rapidly reduced by metabolic inhibitors. Also, our data (Tazawa and Shimmen, 1979) show that inhibition of electrogenesis occurs without a simultaneous decrease in g,. These facts could be explained by assuming that gpis too small to contribute measurably to g,. Then the measured decrease in g,, after depletion of ATP, would represent a gradually occurring decrease in gd. In the present analysis, therefore, we may have overestimated gpand underestimated
4. CONTROL OF ELECTROGENESIS I N
Chafa
65
gd. The possibility that changes in the passive properties of the plasma-
lemma can occur with time after depletion of ATP or Mg2+is supported by the fact that the action potential in Chara cannot be generated by an electric stimulus after depletion of ATP or Mg2+ (Shimmen and Tazawa, 1977). It is also supported by the observation that Rm(-ATP) is equal to Rm(+ at [K+], = 3 mMbut is less than R,(+ ATP) at [K+], = 10 mM (Fig. 6B). From Eq. (l), this would have to mean that g, is zero with 3 m M potassium and negative with 10 mM potassium. It seems much more reasonable to assume instead that the passive potassium permeability increases in the absence of ATP. Finally, we should consider the disparate proportionality between g, and g, described by Keifer and Spanswick (1978, 1979) for C. corallina, as compared with our results for C. australis. From the fact that gd was decreased greatly by treatment of cells with an uncoupler (CCCP, DNP) or an ATPase inhibitor (DCCD, DES), the former authors concluded that g, must account for most of g,. Quantitatively, g, was calculated to be more than 10-fold as large as gd (Keifer and Spanswick, 1978). In contrast, in perfused cells of C. australis (Figs. 8A, and 9A), g, was calculated as 0.5-fold as large as gd, in the pH, range 6-8. This striking discrepancy may be caused by differences both in the preparation and in the means of stopping the electrogenic pump. To stop the electrogenic pump, we directly removed its fuel, i.e., ATP or Mg2+, while Keifer and Spanswick (1978) used metabolic inhibitors. At least in the dark, the membrane of cells treated with CCCP depolarizes slowly, in parallel with the decline of [ATP],. All inhibitors which cause a strong depolarization reduce the [ATP], to about 10% of the normal level (2-4 mM) (Keifer and Spanswick, 1979). When cytoplasm ATP is diluted to 10% of the normal level (0.5-1.2 mM) (Kikuyama et al., 1979) by perfusion with Mg medium, however, Emremains near normal (Shimmen et al., 1976; Shimmen and Tazawa, 1977). This suggests that the observed inhibition of electrogenesis in C. corallina, by various uncouplers and ATPase blockers, may be caused not via reduction of the ATP level per se but via direct effects of this agent on the pump itself. In support of this, we have found that CCCP or DNP added directly to the cell in the perfusion medium depolarizes the membrane to the same level seen in HK cells, even in the presence of sufficient ATP (1 mM) (Kikuyama et al., 1979).
VII.
CONCLUDING REMARKS
The tonoplast-free cell system as an experimental approach to electrogenesis by the Chara plasmalemma has demonstrated several important facts: (1) the existence of an electrogenic ion pump fueled by Mg-ATP; (2)
66
MASASHI TAZAWA AND TERUO SHIMMEN
the dependence of net H + efflux on ATP; (3) the dependence of the pH, sensitivity of Em upon the electrogenic activity of the plasmalemma; (4) a pHi optimum near neutrality for electrogenesis; ( 5 ) the insensitivity of electrogenesis to external K + concentrations; (6) complete inhibition of the light-induced potential change by removal of intracellular ATP. However, several basic problems remain unsolved. We are not certain whether the electrogenesis supported by Mg-ATP is fully accounted for by the H -extruding pump. Although we can stop active electrogenesis either by removing Mg-ATP or by applying inhibitors, the mechanisms of inhibitor action, whether direct or indirect, have not been fully elucidated. Also, the assumption that Em and g,-when the pump is stopped-are equal to passive components Ed and g,-when the pump is running-should be further examined, and we regard the analysis based on this assumption as tentative at present. The pH dependence of the pump current (Fig. 10) resembles that for the plasmalemma ATPase of Neurospora (Scarborough, 1977; Bowman and Slayman, 1979), but isolation and enzyme characterization of plasmalemma vesicles from Chara cells remain to be carried out. Finally, it will be interesting to investigate whether or not plasmalemma H+-ATPase in Cham can be driven backward to produce ATP [as can be done in prokaryotes (Sone el al., 1977)], when a very large proton-motive force is imposed artificially. +
ACKNOWLEDGMENTS This work was partly supported by a grant in aid for scientific research from the Ministry of Education, Science and Culture, Japan, and also by a grant for promotion of research from the Yamada Science Foundation. REFERENCES Bowman, B. J., and Slayman, C. W. (1977). J . Biol. Chem. 252, 3357-3363. Bowman, B. J., and Slayman, C. W. (1979). J . Biol. Chem. 254, 2928-2934. Fujii, S., Shimmen, T., and Tazawa, M. (1979). Plant Cell Physiol. 20, 1315-1328. Hope, A. B. (1965). Aust. J . B i d . Sci. 18, 789-801. Kawamura, G., Shimmen, T., and Tazawa, M. (1980). Planta 149, 213-218. Keifer, D. W., and Spanswick, R. M. (1978). Plant Physiol. 62, 653-661. Keifer, D. W., and Spanswick, R. M. (1979). Plant Physiol. 64, 165-168. Kikuyama, M . , Hayama, T., Fujii, S., and Tazawa, M. (1979). Plant Cell Physiol. 20, 993- 1002. Kitasato, H. (1968). J. Gen. Physiol. 52, 60-87. Lucas, W. J. (1975). J . Exp. Bot. 26, 271-286. Lucas, W. J., and Smith, F. A. (1973). J . Exp. Bot. 24, 1-14. Nagai, R . , and Tazawa, M. (1962). Plant Cell Physiol. 3, 323-339. Nishizaki, Y. (1968). Plant Cell Physiol. 9, 377-387. Richards, J. L., and Hope, A. B. (1974). J . Membr. Biol. 16, 121-144.
4. CONTROL OF ELECTROGENESIS IN Chara
67
Saito, K., and Senda, M. (1973a). Plant Cell Pbysiol. 14, 147-156. Saito, K., and Senda, M. (1973b). Plant Cell Pbysiol. 14, 1045-1052. Scarborough, G. A. (1976). Proc. Nail. Acad. Sci. U.S.A. 73, 1485-1488. Scarborough, G. A. (1977). Arch. Biocbem. Biopbys. 180, 384-393. Shimrnen, T., and Tazawa, M. (1977). J . Membr. Biol. 31, 167-192. Shimrnen, T., and Tazawa, M. (1980). Plant Cell Pbysiol. 21, 1007-1013. Shimmen, T., Kikuyama, M., and Tazawa, M. (1976). J . Membr. Biol. 30, 249-270. Slayman, C. L. (1965). J . Gen. Pbysiol. 49, 93-116. Slayman, C. L. (1970). A m . Zoo/. 10, 377-392. Slayman, C. L., Long, W. S., and Lu, C. Y.-H. (1973). J . Membr. Biol. 14, 305-338. Smith, F. A., and Walker, N. A. (1976). J . Exp. Bot. 27, 451-459. Sone, N., Yoshida, M., Hirata, H . , and Kagawa, Y. (1977). J . Biol. Cbem. 252, 2956-2960. Spanswick, R . M. (1970). J. Membr. Biol. 2, 59-70. Spanswick, R. M. (1972). Biocbim. Biopbys. Acta 288, 73-89. Spanswick, R . M. (1973). In “Ion Transport in Plants” (W. P. Anderson, ed.), pp. 113-128. Academic Press, New York. Spanswick, R . M. (1974). Biocbim. Biopbys. Acta 332, 399-412. Spear, D. J . , Barr, J . K., and Barr, C. E. (1969). J . Gen. Pbysiol. 54, 397-414. Tazawa, M. (1964). Plant Cell Pbysiol. 5 , 33-43. Tazawa, M., and Shimmen, T. (1980). In “Plant Membrane Transport” (J. Dainty, ed.), pp. 349-362. Elsevier, Amsterdam. Tazawa, M., Kishimoto, U., and Kikuyarna, M. (1974). Plant Cell Pbysiol. 15, 103-110. Tazawa, M . , Kikuyama, M., and Nakagawa, S. (1975). Plant Cell Pbysiol. 16, 611-621. Tazawa, M., Kikuyama, M., and Shimrnen, T. (1976). Cell Struct. Funct. 1, 165-176. Tazawa, M., Fujii, S., and Kikuyarna, M. (1979). Plant Cell Pbysiol. 20, 271-280. Walker, N. A,, and Smith, F. A. (1975). Plant Sci. Lett. 4, 125-132. Williamson, R . E. (1975). J . Cell Sci. 17, 655-668.
This Page Intentionally Left Blank
Part II
The Evidence in Epithelial Membranes
This Page Intentionally Left Blank
CURRENT TOPICS IN MEMBRANES A N D TRANSPORT, VOLUME 16
Chapter 5 An Electrogenic Sodium Pump in a Mammalian Tight Epithelium S. A . LEWIS AND N. K. WILLS Department of Physiology Yale University School of Medicine New Haven, Connecticut
I. 11.
111. IV. V.
Introduction ...................................... Electrical Structure of an Epithelium ..................................
............................
Electrical Measureme A.
......................
Membrane Resis
v1. B. Zero-Gradient Potentials ............... C . Cell Loading ............... ........................ D. Energy ...................... v11. Summary ..........................................................
......................................................
1.
71 72 72 74 76 76 77 78 79 79 80 83 85 85 86
INTRODUCTION
Epithelia are sheets of cells that selectively absorb (lumen to blood) or secrete (blood to lumen) electrolytes and nonelectrolytes against a net electrochemical gradient. To achieve this uphill transport two means of energy are available: (1) direct use of a high-energy compound such as ATP, and (2) indirect use of energy through coupling of the transported species to the electrochemical gradient of another ion. An example of the first type of 71
Copyright $1 1982 b y Academic Press, Inc. All rights of reproduction in a n y form reserved. ISBN 0-12-153316-6
72
S. A. LEWIS AND N. K. WILLS
transport is the translocation of Na+ and K+ across the epithelial membranes by Na+,K+-ATPase. The second type of transport is illustrated by carrier-mediated translocation of NaCl down the Na+ gradient but up the C1- gradient (Reuss, 1979). In this chapter we will focus on active transport of Na+ and in particular investigate some physiological characteristics of the Na+,K+-ATPase located in the basolateral membrane of a “tight” epithelium, rabbit urinary bladder.
II. ELECTRICAL STRUCTURE OF AN EPITHELIUM Epithelia can be divided into two general classes, “tight” epithelia and “leaky” epithelia. One difference between these two classes of epithelia is the resistance of the so-called tight junctions between adjacent cells (Fromter and Diamond, 1972). In tight epithelia the electrical resistance of the junctions is large compared to the resistance of the cellular pathway. Conversely, leaky epithelia have a junctional resistance which is lower in value than the resistance of the cellular pathway. Because of the high relative resistance of the junctions in tight epithelia, the electrical potential developed across the epithelium can be approximated by the sum of the equivalent electromotive forces (EMFs) across the apical and basalar membranes. These EMFs in turn are composed of ionic diffusion potentials and I-R drops caused by any current-generating pumps present. Figure l a is an electrical equivalent circuit for a cell membrane with a current-generating pump. The equation which describes this simple circuit is vm
= [(EK/RK
- &alRNa
+ Ip>1 R K R N a / ( R K
+
RNa)
(1)
where V , is the measured membrane potential, EK and ENa the Nernst potentials for these ions, RK and RNa the resistances, and I,, a pump current (perhaps Na+). Inhibition or activation of this pump current will result in a change in the membrane potential. Unfortunately any real epithelium has a more complex equivalent circuit (Fig. lb) than that shown in Fig. la, and the exact resistors and EMFs are dependent on the ion that is transported.
111.
BASIC TRANSPORT PROPERTIES OF RABBIT URINARY BLADDER
The transepithelial transport properties of the rabbit urinary bladder are very similar to those of more well-known epithelia, frog skin and toad urinary bladder. Electrically the rabbit urinary bladder has a spontaneous
Na+
5. ELECTROGENIC
a
PUMP IN MAMMALIAN TIGHT EPITHELIUM
73
Vm
Pi
b
'bl
-FIG. 1. (a) Simplified equivalent-circuit model of a membrane with a potassium EMF (&) and resistance ( R K ) .a sodium EMF (EN,) and resistance (RN,),and a current source (Ip).V, is the membrane potential ( y- Vo). Subscripts o and i refer to outside and inside respectively. (b) A more realistic electrical equivalent circuit of an epithelium. E, and R, are the resistance and EMF, respectively, of the apical membrane, Ebl and R,I are those for the basolateral membrane. Rj is the junctional resistance, and Ip is the current produced by an electrogenic pump. VT is the electrical potential difference across the epithelium. V,, (basolateral membrane potential) is the potential measured from cell interior to ground. For short-circuited conditions this circuit can be made equivalent to that of (a).
transepithelial potential which ranges from - 40 to - 120 mV, referenced to the serosal solution, a transepithelial resistance that varies from 75,000 to 3000 Q cm2, and a short-circuit current (ZSJ from 0.5 ,uA/cm2 to at least 30 pA/cm2. [The short-circuit current is the current required t o reduce the open-circuit potential across the epithelium to zero and measures the net
74
S. A. LEWIS AND N. K. WILLS
transport of charge.] If only one ion is being actively transported, then Z,,is a direct measure of the net transport of that ion. Using isotope flux measurements Lewis and Diamond (1975, 1976) found that Zsc was indeed equal to the net movement of isotopic Na+ from lumen to blood (the isotope measurements were made under conditions of zero transepithelial electrochemical gradients). Amiloride, a natriuretic drug known to reduce Na+ transport-by decreasing apical Na permeability-in frog skin and toad urinary bladder, also rapidly inhibited Na+ transport in rabbit bladder. Because it decreases apical Na+ permeability, amiloride causes an increase in transepithelial resistance t o a value of 75,000 Q cm'. Among other inhibitors, ouabain (which blocks specifically the Na+, K + ATPase system) rapidly inhibits Na+ extrusion at the basolateral membrane and causes an increase in transepithelial resistance. The major resistance change has been localized at the apical membrane (Lewis et al., 1976). Apparently increased intracellular Na+ activity (> 20 mM, see Wills and Lewis, 1980) causes apical Na+ permeability to fall, thus limiting further Na+ entry. Such negative feedback control may have an important role in the maintenance of cell volume. In contrast to the above agents, aldosterone (an antinatriuretic hormone) stimulates Na' transport across rabbit bladder. It does so by increasing apical Na permeability. This increase in permeability is characterized by an increase in transepithelial spontaneous potential and a decrease in transepithelial resistance. Again the predominant potential and resistance changes occur at the apical membrane. +
+
IV.
ELECTRICAL MEASUREMENTS
Before proceeding to the section on the evidence for an electrogenic pump, a summary of our experimental measuring system is in order. As already mentioned, we measured both spontaneous potential and total resistance across the epithelium. The spontaneous potential across the urinary bladder was measured using a differential amplifier connected to voltage-measuring electrodes placed close to and on opposite sides of the epithelium. The resistance was calculated from the transepithelial voltage change to a current step (R = AV/A& see Lewis et al., 1978). The apical surface area was estimated from the measured membrane capacitance (Lewis and Diamond, 1976; Clausen et al., 1979) assuming IpF = 1 cm2of flat membrane. (In the remainder of this chapter, epithelial resistance, conductance, and currents have been normalized to apical membrane capacitance but will be expressed in cm2 rather than pF.) The net Na+ transport
5. ELECTROGENIC Na+ PUMP IN MAMMALIAN TIGHT EPITHELIUM
75
rate was measured by an automatic voltage clamp which determined the short-circuit current (Zsc = JF; F ) . Such transepithelial measurements do not allow one to calculate changes in the individual membrane resistances, potentials, or ion activities. Since these parameters must be known before an answer can be given concerning the existence or nonexistence of an electrogenic pump, and its localization, conventional microelectrodes were used to determine the separate membrane and junctional resistances and potentials. Glass capillary tubing was pulled to tip diameters less than 0.5 pm o n a Narishige electrode puller and filled with 3 M KCI. The resistance of these microelectrodes was 20-40 Ma. One microelectrode was connected to one channel of a high-impedance differential amplifier (WP Instruments Model 750) mounted on a Stoelting hydraulic-drive micromanipulator. The microelectrode was placed in the mucosal solution and referenced to the second channel of the differential amplifier, which was connected to the voltagemeasuring electrode in the serosal solution. The potential difference between the microelectrode and the reference electrode was then adjusted to equal the transepithelial potential measured as described above. Upon advancing the microelectrode from the mucosal solution through the apical membrane the first potential difference measured was the basolateral membrane potential ( VbJ. The difference between this and the transepithelial potential was the apical membrane potential ( Va). Separate membrane and junctional resistances were determined from the measured transepithelial resistance and the ratio of apical membrane resistance t o basolateral membrane resistance (resistance ratio =a),before and after selectively changing the apical membrane resistance (e.g., with amiloride). The ratio a was determined from the change in the apical and basolateral membrane potentials in response to an applied transepithelial current pulse (see Lewis et al., 1977, 1978; Clausen et al., 1979). When the separate resistances and potentials are known, the equivalent EMFs for each membrane can then be calculated. The intracellular ion activities were estimated using liquid ion-exchanger microelectrodes. K + and C1- liquid ion exchangers were manufactured by Corning, and the Na+ liquid ion was that described by Steiner et al. (1979) and O'Doherty et al. (1979). The micropipets for ion-specific electrodes were pulled identically to the conventional microelectrodes and then made hydrophobic using tri-n-butylchlorosilane (see Lewis and Wills, 1980; Wills and Lewis, 1980). The appropriate ion exchanger was introduced into the tip, and the microelectrode was then backfilled with 0.5 M chloride salt. The ion-specific microelectrodes were calibrated with ion-pure and mixedsalt solutions to determine their slopes and selectivities. Because of the high
76
S. A. LEWIS AND N. K. WILLS
electrical resistance of these ion-specific microelectrodes an amplifier with ultrahigh input impedance was used (WP Instruments F233A, input impedance = 1015Q ) . Actual intracellular ion activities were calculated using the Nicolsky equation:
aix = (a& + Kxyaoy) [exp (nF/RT) ( v x - vbl)l - &yaiy where a is activity, subscripts i and o refer to inside (cell) or outside (bathing) solutions, respectively, x is the ion of interest, y is the competing ion, Kxyis the selectivity term, v b , is the basolateral potential measured with a conventional microelectrode, V, is the total potential measured by the ion-specific microelectrode (the sum of the ion activity potential and vbl), n is the correction term for nonideal electrode behavior, and F, R, and T have their usual meanings. As a test for impalement damage, the resistance ratio was measured using the ion-specific microelectrodes (see Lewis et af., 1978). Transepithelial voltage-measuring and current-passing electrodes as well as conventional and ion-specific microelectrodes were connected t o a strip chart recorder, four-channel storage oscilloscope, and either a digital printer or a small digital computer (North Star).
V.
EPITHELIAL PARAMETERS
The transepithelial potential and conductance of the rabbit urinary bladder increase in value as the rate of Na+ transport increases. Consequently apical and basolateral membrane conductance, EMFs, or both change as a function of increased Na+ transport. In principle the increased EMF could be caused by either an increased passive ion permeability or an electrogenic pump. To dissociate these two possibilities we must first determine the resistance and passive permeabilities of both membranes in the absence of net Na+ transport.
A. Membrane Resistances The condition of zero net Na+ transport can be imposed by exposing the mucosal surface of the epithelium to either amiloride or Na+-free solutions (Lewis et al., 1977, 1978; Clausen et al., 1979). The resultant resistances of apical (R,) and basolateral membranes (Rbl) and tight junctions (Rj) are summarized in Table I . It should be noted that with zero transport there was still a finite apical conductance of 12 pS/cm2 or a resistance of 83,000 Q cm2. Basolateral resistance was also normalized to the apical area. (To
5. ELECTROGENIC
Na+
77
PUMP IN MAMMALIAN TIGHT EPITHELIUM
TABLE 1 EPITHELIAL RESISTANCES AT ZERO AND HIGHRATES OF Na+ TRANSPORT^ Transport rate
Ra
Rbl
4
Zero High
83,000 3000
1500 1500
> 100,000 > 100,Ooo
'All values are in ohm centimeters squared.
convert to the actual basolateral membrane area increase the resistance of the basolateral membrane by a factor of 5; Clausen et al., 1979). As the Na+ transport rate increased, the apical conductance increased from a low of 20 pS/cm2 to a measured high of 300 pS/cm2. The basolateral membrane conductance remained constant, as did the junctional conductance. Thus net Na+ transport is controlled by the permeability of the apical membrane to Na+. One of the actions of aldosterone is then to increase apical Na+ permeability. B. Membrane Potentials Changes in membrane potential as a function of transport are outlined in Table 11. In the absence of Na+ transport the basolateral membrane potential ( Vbl)was z - 54 mV, referenced to the serosal solution, and the apical membrane potential was approximately - 15 mV, referenced to the mucosal solution. The latter increased to - 54 mV when Na+ was replaced with choline (not shown in Table 11). As sodium transport increased, the apical membrane potential changed from cell interior negative to cell interior positive (50 mV), again indicating a specific increase in apical Na+ permeability. In contrast, there was no measurable change in basolateral membrane potential as a function of Na+ transport, which at first sight could mean that there is no electrogenic pump in the basolateral membrane. We shall demonstrate below, however, that this simple interpretation is erroneous. TABLE 11 APICAL AND BASOLATERAL MEMBRANE POTENTIALS AT Low AND HIGHRATESOF Na' TRANSPORT Transport rate
v,(mV)
Low High
- 15
+ 50
vbl (m
v)
- 54 - 54
78
S. A. LEWIS AND N. K. WILLS
C. Membrane Selectivity The apical membrane behaves very much like a Na+ electrode circuited in parallel with a leak permeability that has a discrimination ratio ( P N a / P K ) of 0.8. As already discussed, Na+ permeability of the apical membrane can be increased by aldosterone. The potential and resistance across the basolateral membrane, on the other hand, are both unaffected by Na' transport. What then, is the source of the spontaneous basolateral membrane potential? Is it a consequence of passive ion diffusion as found in most excitable membranes, or could there be, in addition, an electrogenic pump? A preliminary answer t o this question was sought from the effect of ouabain addition (10-4M)to the serosal bathing solution. The basolateral membrane potential did not change for at least 10 minutes after ouabain addition, during which time the sodium permeability of the apical membrane was reduced. This negative result indicates that, if there is an electrogenic pump in the basolateral membrane, either its signal is within the measuring accuracy of the system or it is counterbalanced by a selectivity change in the membrane. The normal selectivity of the basolateral membrane was determined by equimolar replacement of Na+ with K + : at constant KC1 product, to determine P N a / P K ; or at constant C1-, to determine Pc,/PK.With each ion replacement the change in basolateral membrane potential was recorded. The collected data were fitted to a modified form of the constant-field equation. (By measuring the change of the basolateral membrane potential at each partial ion replacement the intracellular cation activity is canceled from the resultant equation when constant.) Table I11 summarizes the basolateral membrane permeabilities. Although C1- permeability appeared somewhat greater than K + permeability, C1- (see below) was in electrochemical equilibrium with V,, and consequently did not influence the resting membrane potential. From the measured membrane potential, intracellular K + activity was estimated at 80 mM, again using the constantfield equation. Actual intracellular ion activities, K + , C1-, and N a + , obtained with
TABLE 111 ION PERMEABILITY OF BASOLATERAL MEMBRN E
Ion
Permeability (cm/second) 0.04
1.2 2 x 10-6
5. ELECTROGENIC
Na+
79
PUMP IN MAMMALIAN TIGHT EPITHELIUM
TABLE IV INTRACELLULARION ACTIVITIES, EQUILIBRIUM, AND BASOLATERAL MEMBRANE POTENTIALS'
Na' K+ C1-
1
+71.5
89 15.8
- 15 - 45
- 55 - 55 - 52
'Activities refer to millimolarity; potentials are in millivolts. liquid ion-exchanger microelectrodes, are summarized in Table IV. The internal C1- activity was found to be 15.8 mM, compared with 92 mMin the serosal medium, yielding a diffusion potential (ECJof - 45 mV, essentially in electrochemical equilibrium with the basolateral membrane potential. Measured and calculated K + activities were in good agreement, thus supporting the use of the constant-field equation for the calculation of permeabilities. K + and Na+ were not in equilibrium with the membrane potential; K + was actively taken into the cell against an electrochemical gradient of approximately 20 mV, and Na+ was extruded from the cell against an electrochemical gradient of nearly 130 mV. From these results, the basolateral membrane potential could be described by the combined diffusion of K + and Na+ with PKG 25 PNa.
VI.
PUMP PROPERTIES
From the ion distributions and cellular membrane potentials reported above it is clear that an energy-dependent pump is located on at least one of the membranes. Its most likely location is in the basolateral membrane, because ouabain is active only from the serosal side. In this section we will describe several groups of experiments which, taken together, demonstrate that the pump is electrogenic-i.e., that it separates charge across the (basolateral) membrane. A. Increasing Na+ Entry Rate
As stated in Section V,C, the sodium pump inhibitor ouabain stops net Na+ transport across the rabbit urinary bladder with no measurable affect on the basolateral membrane potential (Vb,). At least one reason for this lies in the signalhoise ratio of the measuring system. The rate of net Na+ entry into the cell (from both mucosal and serosal solutions) is 0.26 pEq
80
S. A. LEWIS AND N . K. WILLS
cm-2 hr-’ which would give a pump current of 2.3 pA/cmZ assuming an exchange coupling ratio (Na+/K+)of 3:2. Since the apparent resistance of the cell membrane is only slightly greater than 1200 fl cmZ, a nominal I-R drop of 2-3 mV would be expected. This value is within the resolution of the measuring system. In order t o enhance the Na+ transport rate, we decided to increase sodium loading of the epithelium by treating the apical membrane (mucosal solution) with the pore-forming polyene antibiotic nystatin. This antibiotic had previously (Lewis et al., 1977) been found to give elevated apical cation and anion permeability in the ratio (PNa= PK > Pc,) 1: 1:0.3. In the present experiments mucosal nystatin was found t o diminish apical resistance by at least two orders of magnitude and to hyperpolarize the basolateral membrane by 13 f 3 mV. The resistance change is observed under all circumstances, but the hyperpolarization is observed only when Na+ is present in the mucosal solution. As in many systems (Gadsby and Cranefield, 1979), the total hyperpolarization resulting from stimulation of the Na+ pump may have as sources either a direct effect of pump current or an indirect effect due to K + depletion in the pericellular unstirred layers. A calculation based on the constant field equations and the reported ion permeabilities and cell activities indicates that the maximum hyperpolarization expected from a reduction of lateral space K + activity to zero is approximately 8 mV. Since the hyperpolarization was greater than this value, this experiment is evidence that there is an electrogenic pump in the basolateral membrane.
B. Zero-Gradient Potentials
To circumvent the problem of unstirred layers the previous experiment was repeated under zero-gradient conditions, i.e, with the mucosal and serosal K+ activities equal to that of the cell. Figure 2 is a stepwise illustration of the experimental protocol. For the control condition, (a) both sides of the preparation were bathed in NaC1-Ringer’s solution; (b) the mucosal solution was replaced with Na+-free KzS04-Ringer’s solution; and then (c) nystatin was added to the mucosal solution. After the transepithelial resistance and resistance ratio had reached constant values, serosal NaC1Ringer’s solution was replaced with Na+-free K,SO,-Ringer’s solution (d). This ion replacement brought the transepithelial voltage to zero. The K+ activity in the K2S04-Ringer’s solution was equal to the measured cell K+ activity, and consequently the preparation lacked any net electrochemical gradient for K + . At this time NaCl was added sequentially to the serosal and then the mucosal solution, in 13 m M steps. Serosal addition of NaCl
82
S. A. LEWIS AND N. K. WILLS
TIME (minutes)
0
I
2
3
4
5
6
7
8
9
I
I
1
I
I
I
1
1
I
I
10 I
I1
12
1
I
-15' NaCl (mM) I
13
I
I
27
40
1
53
I
ouabain
FIG.3. Response of the transepithelial potential (VT) to stepwise addition of NaCl to first a serosal and then a mucosal bathing solution of a nystatin-treated preparation. Ouabain M ) was added to the serosal solution and reduced the potential from -13.5 to -1.5 mV. The step change in V, after each addition of NaCl represents a diffusion potential across the nystatin-treated apical membrane. (From Lewis el al., 1978.)
From the hyperpolarizing responses of Fig. 3 and the basolateral membrane resistance the pump current (I,) can be calculated. Figure 4 is a plot of I, versus mucosal Na' concentration and reveals the finite order of Na' in the reaction. The curve is the best fit to the equation of highly cooperative binding (using a nonlinear least squares curve-fitting routine):
+
Zp = Imax/[l (KmNa/[Nal)nl
where I, is the current measured at each mucosal Na+ concentration, "a],
n is the number of sites available per ligand, I,,,,,is the maximum current at infinite Na+ concentration, and K,Na is the Na' concentration for halfmaximal current stimulation. The best-fit values were n = 2.3, I,,, = 28.3 =a 28.6 mM. The value for n of 2.3 is a minimum estimate pA/cm2, and Pm of the number of Na+ ions that bind per ligand. In general the shape of the Z,-versus-NaCl concentration curve (Fig. 4) will depend on the intracellular Na' activity rather than the mucosal Na' concentration. By using the rapid hyperpolarizing response of the apical membrane (induced by the addition of mucosal NaCl) intracellular Na'
13 I
5 . ELECTROGENIC
NaCl
Na+
81
PUMP IN MAMMALIAN TIGHT EPITHELIUM
M
NaCl
NaCl
K Z "4
KZ
t
NYSTATIN
,
NYSTATIN
+
NYSTATIN
+
x NaCl
KZ'O,
I
KZS04
+
x NaCl
FIG.2. Schematic of the experimental protocol for producing the experimental situation where no K + gradients exist between the mucosal solution, cell, and serosal solution (see text for details).
resulted in no measurable change in the transepithelial or basolateral membrane potential. In contrast, mucosal addition of NaCl resulted in a rapid hyperpolarization (a diffusion potential across the apical membrane) followed by a slower rise t o some steady state value. Further increases in NaCl caused a hyperpolarizing response up to a NaCl concentration of 50-60 mM. The average hyperpolarization was 15 m V (Fig. 3), which would require a current density for the basolateral membrane of n25 pA/cm2.
5. ELECTROGENIC
Na+
83
PUMP IN MAMMALIAN TIGHT EPITHELIUM
(uA/crn2) 10
5
0 0
10
20
30
40
50
60
70
NaCl (mM) FIG.4. Current response (Ip)of nystatin-treated bladders t o step increases in rnucosal and serosal bathing solution NaCl concentration. The line is best fit to Eq. (2).
activity can be calculated assuming cell Na+ and C1- activities are equal. Such a calculated Na+ activity-versus-pump current is shown in Fig. 5 . The shape of this curve is similar to that in Fig. 4, as it demonstrates saturation behavior. The curve was determined from a model of highly cooperative binding [Eq. (2) modified so that “a] was intracellular Na+ activity a,Na+] with an n of 2.81, a KNma of 14.2 mM, and an I,,,,, of 27.3 pA/cm2 (normalized to apical membrane area; see Fig. 5 ) . Again n is a minimum estimate of the number of Na+ ions binding to the ligand, thus indicating that the value might be three Na+ ions, in agreement with the number of Na+ ions thought to be transported per cycle of the pump by nonepithelial tissues. As expected, ouabain ( M ) reduces the Na+-induced hyperpolarization t o zero within 60-80 seconds. The reaction time is faster than that for ouabain abolition of short-circuit current (> 60 minutes) and is probably dependent on unstirred layers (see Lewis et al., 1978). Despite the unphysiological conditions, these experiments are strong evidence for an electrogenic Na+ pump in the basolateral membrane. C. Cell Loading Under control conditions the basolateral membrane demonstrated no significant depolarization after 3 minutes of ouabain addition. Since increased cell N a + , however, unmasks an electrogenic pump, the question
84
S. A. LEWIS AND N. K. WILLS
30
1
0
10
20
30
40
ai No (mM)
FIG.5 . Same as Fig. 4 except the intracellular Na' activity (qNa+) is calculated from the rapid response of V, after each addition of NaC1. The curve is best fit to Eq. (2).
arises of whether the pump really is electrogenic under more nearly physiological conditions. To investigate this particular question, urinary bladder cells were loaded with Na+ by removal of all K + from both mucosal and serosal solutions. Within 1 hour, cell K + had decreased by 45 m M and cell Na+ had increased by 65 mM. (Similar time-dependent changes in a,K+ and aiNa+ are observed with ouabain treatment.) Upon readdition of K+ to the serosal solution the basolateral membrane hyperpolarized by 20 mV within 1 minute, in the absence of a measurable change in intracellular Na+ or K+ activity. Intracellular Na+ activity decreased to a steady state value of 7 m M within 7 minutes, but intracellular K+ activity did not return to its control value of 90 mMuntil20 minutes after K+ addition. Once again the pump appeared to be clearly electrogenic. Another possible explanation, at least for the initial hyperpolarization, is that the condition of zero serosal K + altered the selectivity of the basolateral membrane. (Since intracellular Na+ and K+ activities, as well as the basolateral membrane potential, were measured in the presence and absence of serosal can be calculated for these two conditions and with and K + , the PNa/PK without ouabain. PNa/PK of the basolateral membrane was not significantly altered by either zero K+ or ouabain. Control PNa/PK was 0.07 f 0.03 and zero K + and ouabain were 0.07 f 0.03 and 0.10 f 0.05, respectively. It is therefore unlikely that the concentration changes per se affected the passive properties of the basolateral membrane.) The lack of correlation between the rate of Na+ extrusion and K + uptake suggests that the stoichiometry of the pump is not constant if one assumes constant cell volume. Initial rates of Na+ extrusion, compared with K + uptake, yield a Na+/K+ stoichiometry of 5 : l . However, previous
-
-
5. ELECTROGENIC
Na+
PUMP IN MAMMALIAN TIGHT EPITHELIUM
85
data (Section V1,B) suggest that only three Na+ ions are transported per pump cycle. A possible explanation is that half of the time three Na+ ions are transported in the complete absence of K', giving an apparent stoichiometry of 6:1 . Variable stoichiometry might mean a "slippage" of the Na+ pump with respect to K + . Alternatively, readdition of K+ might cause cell volume to increase. In the above experiment a volume increase of 22% would account for an apparent variable stoichiometry, with a pump which is actually invariant. D. Energy One of the more interesting observations made, but not yet sufficiently commented on, is the independence of the rate of net Na+ transport and the intracellular Na+ activity. Such independence suggests a very steep kinetic relationship between aiNa+ and the pump. Although it is generally considered that the Na+ pump can act as a current source, this is not necessarily true if the Na+ pump is working near its equilibrium point. In other words, it is commonly thought that small perturbations in ion activities or membrane potential will not alter the translocation properties of the Na+ pump. If the simple case is taken under these experimental conditions the pump is working near or at equilibrium, then the minimal energy requirements for Na+ translocation can be calculated according to the thermodynamic treatment of Chapman and Johnson (1978). The equation is A = ETF(n - m) - R T [ n In (Nao/Nai) + m In (Ki/Ko)]
(3)
where n and m are the number of Na+ ions and K+ ions translocated for each cycle of the pump, A is the free energy of ATP hydrolysis (joules/mole), ET is the membrane potential (referenced to the serosal solution), and R, T, and F have their usual meanings. From this equation, assuming that the pump is near equilibrium and operates at a Na/K+ stoichiometry of 3:2, the minimal free energy is - 10.2 f 0.2 kcal/mole or -42.8 kJ/mole.
VII.
SUMMARY
Under normal physiological conditions the basolateral membrane potential of the rabbit urinary bladder epithelium can be described by the constant field equation with K + as the most permeant ion. The apical membrane has a variable Na+ permeability, subject to increase by aldosterone and to reduction by amiloride or increased intracellular Na+ activity.
86
S. A. LEWIS AND N. K. WILLS
Intracellular K + and Na+ activities are maintained high and low, respectively, compared to plasma by the sodium pump Na+,K+-ATPase in the basolateral membrane. Neither ion is in equilibrium across the membrane. The pump is electrogenic and translocates three Na+ ions per cycle from the cytoplasm to the serosal solution. The number of K + ions translocated per cycle is less than three. Electrogenicity is clearly evident only after the intracellular sodium activity has been artificially elevated. ACKNOWLEDGMENTS This work was supported by NIH grant AM20851 (S.A.L.). N.K.W. was an NIH postdoctoral fellow (AM06033) during this work.
REFERENCES Chapman, J. B., and Johnson, E. A. (1978). J . Gen. Physiol. 72, 403-408. Clausen, C., Lewis, S. A., and Diamond, J. M. (1979). Biophys. J. 26, 291-318. Fromter, E., and Diamond, J. M. (1972). Nature (London), New Biol. 235, 9-13. Gadsby, D. C., and Cranefield, P. F. (1979). J . Gen. Physiol. 73, 819-837. Lewis, S. A., and Diamond, J. M. (1975). Nature (London) 253, 747-748. Lewis, S. A., and Diamond, J. M. (1976). J. Membr. Biol. 28, 1-40. Lewis, S. A., and Wills, N. K. (1979). Fed. Proc. Fed. A m . SOC. Exp. Biol. 38, 1058-1061. Lewis, S. A., and Wills, N. K. (1980). Biophys. J. 31, 127-138. Lewis, S. A., Eaton, D. C., and Diamond, J. M. (1976). J . Membr. Biol. 28, 41-70. Lewis, S. A., Eaton, D. C., Clausen, C., and Diamond, J. M. (1977). J. Gen. Physiol. 70, 427-440. Lewis, S. A., Wills, N. K., and Eaton, D. C. (1978). J. Membr. Biol. 41, 117-148. O’Doherty, J., Garcia-Diaz, J. F., and Armstrong, W. M. P. (1979). Science 203, 1349-1351. Reuss, L. (1979). Fed. Proc. Fed. Am. Soc. Exp. Biol. 38, 2733-2738. Steiner, R. A., Oehme, M., Ammann, D., and Simon, W. (1979). Anal. Chem. 51, 351-353. Wills, N. K., and Lewis, S. A. (1980). Biophys. J . 30, 181-186.
CURRENT TOPICS IN MEMBRANES A N D TRANSPORT, VOLUME 16
Chapter 6 A Coupled Electrogenic Na+-K' Pump for Mediating Transepit he1ial Sodium Transport in Frog Skin R OBER T NIELSEN University of Copenhagen Institute of Biological Chemistry Copenhagen, Denmark
..................................... ....................
87 88 88
90 91 91
92 the Coupling between Na' and K + Transport 111.
................
Conclusions ............................................. ..................................... References ......................... ..................
1.
97 103 106 107
INTRODUCTION
In Na -transporting epithelia such as frog skin, toad bladder, and rabbit bladder and colon, active transport of Na+ takes place across the epithelium from the apical to the basolateral side. A wide range ot experiments have been carried out in order to investigate whether this Na+ pump is a coupled Na+-K+ exchange system or an electrogenic Na+ pump. This article will focus on data recently obtained from isolated frog skins and will +
87
Copyright 0 1982 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-153316-6
88
ROBERT NIELSEN
briefly summarize related findings obtained on other epithelia which have already been reviewed in an earlier volume of this series (Schultz, 1978). A. Anatomy The structure of the frog skin, which may be taken as an example of a multilayered epithelium, is drawn schematically in Fig. 1A and is illustrated by the light micrograph of sectioned material in Fig. 1B. The top layer or stratum corneum consists of dead cells, and-when isolated-is nonselectively permeable to Na+ and other ions (Bracho et al., 1971; Nielsen, 1972; Vieira et al., 1976). Under the stratum corneum is the apical membrane comprising the “outward-facing” surfaces of a second row of cells (in the stratum granulosum). These apical surfaces are firmly interconnected by tight junctions (Farquhar and Palade 1965), such that the apical membrane and tight junctions form a barrier separating the apical bathing solution from the basolateral solution. The rest of the epithelial layer consists of variously packed and interconnected cells whose membranes together form the basolateral membrane of the epithelium. Attached beneath the epidermis, but not shown in Fig. 1, is the dermis, containing blood vessels, glands, chromatophores, and most connective tissue elements and accounting for 80% of the total thickness of the skin. A detailed description of the ultrastructure of isolated frog skin has bee provided by Farquhar and Palade (1965).
B. Models Koefoed-Johnson and Ussing (1958) have found that the apical membrane of the isolated frog skin is selectively permeable to Na+ but almost impermeable to K + , whereas the basolateral membrane is permeable to K + but almost impermeable to free Na+. Qualitatively, the same permeability properties have been found in other Na+-transporting epithelia (e.g. , toad bladder: Macknight and Leaf, 1978; rabbit urinary bladder: Lewis et al., 1977; and rabbit ileum: Schultz, 1978), and on the basis of such observations Koefoed-Johnsen and Ussing proposed a two-membrane hypothesis for the physiological structure of the epithelium. According to this hypothesis (Fig. 2A), active Na+ transport across epithelia-in the inward direction (downward in Fig. 1)-occurs in two steps: passive diffusion of the ion across the apical membrane, followed by active extrusion of Na+ through the basolateral membrane. Koefoed-Johnsen and Ussing (1958) suggested the active mechanism to be a Na+-K+ exchange pump, and a central feature of the model was that this single mechanism regulated both
6.
Na+- K +
PUMP MEDIATING TRANSEPITHELIAL
Na+
TRANSPORT
89
s.c 1RCL
S.G
S.P
S.GER
I
BM
FIG. 1. (A) Schematic drawing of an epithelium. (B) Light micrograph of a frog skin epithelium. S.C, stratum corneum; S.G, stratum granulosum; S.P, stratum spinosum; S.GER, stratum germinativum; IRCL, first reactive cell layer; BM, basement membrane; T J , tight junction; D, desmosome; G, gap junction; AM, apical membrane, BLM, basolateral membrane. The bar in Fig. 1B corresponds to 10 pm. The basolateral membrane consists of the inward-facing membrane of the IRCL, and the plasmalemmas of the underlying cells, to the extent that these cells are coupled to the cells in the IRCL.
90
ROBERT NIELSEN
A
B
C
FIG.2. Models suggested to explain both homocellular regulation of N a + and K + composition and active transcellular transport of N a + . For further details see text.
cellular Na' and K + composition, on the one hand, and active transepithelial Na+ transport, on the other hand. A wide range of observations were found to be consistent with this model (Fig. 2A). Histochemical and autoradiographic studies on intact tissue, as well as enzymic analyses of fragmented epithelial cells, have localized ouabain-binding sites and Na+, K +-activated ATPase activity in the basolateral membranes; little or no ouabain-binding or Na+ ,K+ATPase activity is found in the apical membrane (Farquhar and Palade, 1966; Mills et al., 1977; Bonting, 1970). The involvement of K+ in transepithelial Na+ transport by frog skin was demonstrated by Huf and Wills (1951), who showed that Na+ transport was drastically reduced when K +-free bathing solutions were used. Koefoed-Johnsen (1957) found that addition of ouabain to isolated frog skin resulted in an inhibition of active N a + transport, and a quantitative correlation between the number of ouabain molecules bound and the resultant inhibition of Na+ transport has recently been demonstrated by Cala et al. (1978). While all these observations are consistent with the Koefoed-Johnsen and Ussing model, they are also consistent with the other models, as shown in Figs. 2B and C. Figure 2B is a model suggested by Frazier and Leaf (1963), according to which the Na+ pump is purely electrogenic, K' being pulled into the cells by the electrical potential created by the Na+ pump. Figure 2C (DeLong and Civan, 1978) proposes that transepithelial Na+ transport and the cellular K + concentration are regulated by two different pumps.
C. Electrogenic Na+ Pumps Recent experiments on a variety of tissues strongly suggest that the Na+ pump in Na+-transporting epithelia is indeed electrogenic. Varanda and Lacaz-Vieira (1979) reached this conclusion from studies on toad skin,
6.
Na+-K+
PUMP MEDIATING TRANSEPITHELIAL
Na+
91
TRANSPORT
based on the facts that transepithelial K + fluxes are small and that ouabain transiently increases K + efflux across the apical membrane. In a quite different approach, Miller et al. (1978), using microelectrodes to measure the intracellular potential of bullfrog retinal pigment epithelium, found that the addition of ouabain depolarized the cells in two phases: A fast phase ( t %= 1.5) minutes was identified with the removal of a direct electrogenic component of the pump, and a slow depolarizing phase ( t % hours) due to running down of the transmembrane ionic gradients. Measured changes in the intracellular potential under different conditions have likewise been taken as evidence for electrogenic Na+ pumping in rabbit urinary bladder, rabbit colon, and frog skin (Lewis et al., 1978; Wills et al., 1979; Nagel, 1978). The use of microelectrodes in the study of electrogenic pumps in epithelia has been covered in detail by Lewis and Wills, this volume.
=
II. COUPLING BETWEEN ACTIVE N a + AND K + TRANSPORT In order to distinguish among the three models shown in Fig. 2 it is not enough to establish that the Na+ pump is electrogenic, since in all three models it could be electrogenic or partially so. Rather, one must determine whether active transepithelial Na+ flux is in some way obligatorily coupled to the K + flux across the basolateral membrane and to determine the actual Na+-K+ coupling ratio. For this reason, a great deal of experimental effort has gone into the task of measuring simultaneously the active transepithelial Na+ transport and the basolateral K+ flux in epithelial membranes.
A. Active N a + Transport Active Na+ transport across the isolated frog skin can be measured by the short-circuiting technique (Ussing and Zerahn, 195 l), as diagramed in Fig. 3. The epithelium (E) is placed as a diaphragm between two halfchambers, designated I and 11. The potential difference across the epithelium is detected by calomel electrodes (C) brought into contact with the Ringer’s solutions on either side of the epithelium by a pair of agar-Ringer’s solution bridges (A). Another pair of electrodes (B, usually Ag-AgC1 electrodes) is used to pass current through the membrane and thereby impose a predetermined voltage (zero for a short-circuit measurement) which is set by means of a high-gain differential amplifier (AMP) whose input is the two calomel electrodes. The short-circuit current (SCC)
92
ROBERT NIELSEN B
E
B
FIG.3. Diagram of the apparatus used for short-circuiting an epithelium.
has been demonstrated to be equal to the (chemically measured) active transepithelial Na+ transport when the isolated frog skin (Ussing and Zerahn, 1951) or isolated toad urinary bladder (Leaf et a/., 1958) is bathed on both sides by identical solutions. B. K + Flux
In order to investigate whether the K + flux across the basolateral membrane is coupled to the transepithelial Na+ transport, four different types of measurements have been made: (1) 42K+-K+ exchange across the basolateral membrane, (2) changes in cellular K + content, by means of K +-sensitive microelectrodes, (3) transepithelial K + transport, after the apical membrane has been made permeable to K + , and (4)the effects of inhibitors of passive K + flux. 1. 42K+-K+ EXCHANGE Measurements of K + flux across the basolateral membrane in whole frog skin, rabbit ileum, and toad bladder (K+ flux measured as 42K+ uptake from the inside bathing solution into the tissue) show that no coupling exists between this K + flux and transepithelial Na+ transport (Curran and Cereijido, 1965; Essig and Leaf, 1963; Candia and Zadunaisky, 1972; Schultz, 1978; Nellans and Schultz, 1976; Robinson and Macknight, 1976b). However, in whole-skin experiments K + exchange can be distorted by the thick layer of dermis (corium) which underlies the basolateral membrane. Working with epithelia dissected from frog skin after collagenase treatment, Biber et a/. (1972) found a highly significant correlation between the SCC and 42K+influx from the inside bathing solution. The positive correlation observed was the result of spontaneous parallel variations in SCC and 42K+ influx among epithelia from different animals. The addition of
6.
Na+-K+
PUMP MEDIATING TRANSEPITHELIAL
Na+
93
TRANSPORT
antidiuretic hormone (ADH), to stimulate active sodium transport, resulted in increased 42K+ influx; but when the SCC was abolished by treatment with amiloride, 42K+ influx across the basolateral membrane was not affected. On the other hand, 42K+efflux-from isolated frog skins preloaded with the radiotracer-is drastically reduced by inhibition of the SCC with amiloride (Ferreira, 1979). This finding complements the earlier kinetic analysis of 42K+efflux from toad urinary bladder (Finn and Nellans, 1972), in which ADH stimulation of SCC also increased 42K+ efflux. Thus the isolated epithelia, unlike intact frog skin, yield data which partially support the notion of coupling between transepithelial Na transport and K + flux across the basolateral membrane. The discrepancy between data obtained from whole tissues and those from isolated epithelia is probably due to recycling of the ions. 42K+ diffusing through the narrow interspace system is not likely t o be in equilibrium with 42K+ in the bathing solution (Harris and Burn, 1949). This lack of equilibration results in an unknown degree of recycling (K+, 42K+)which could completely mask the presence of a small K + pool having a high turnover rate. Such an error necessarily increases as the length of the diffusion pathway increases, so that 42K+ flux measurements on whole tissues, and sometimes even on isolated epithelia, may be difficult t o interpret. +
MICROELECTRODES 2. K +-SELECTIVE Removal of K + from the solutions bathing isolated urinary bladder causes the transporting cells to lose 120 nmoles K+/kg dry wt and to gain -45 nmoles Na+/kg dry wt (Robinson and Macknight, 1976a). Restoration of external K + produces rapid and complete recovery of intracellular K + , and DeLong and Civan (1978) have studied the time course of this recovery using K+ -selective microelectrodes. The process is 90-97% complete in 25 minutes, before any appreciable recovery of SCC can be observed. In order to explain these observations, DeLong and Civan (1978) argued: “Although other interpretations are possible, the simplest interpretation of the data is that the processes responsible for K + accumulation and transepithelial Na+ transport are not identical. We propose the existence of a separate transfer mechanism at the basolateral cell membrane, responsible for accumulating intracellular K + , and not directly coupled to active Na+ transport.” This interpretation, however, is based on the assumption that the transport system in the toad bladder does not change during the removal and restoration of K + , say, that the Na+ and K + permeabilities of the basolateral membrane stay reasonably constant and
-
94
ROBERT NIELSEN
that the measured SCC in the two situations is a genuine indicator of pump-mediated active Na+ transport. 3. TRANSEPITHELIAL K + TRANSPORT
The biological effects of polyene antibiotics, e.g., amphotericin B, have been attributed to increased permeability of cell membranes, probably due to an interaction of the polyene compound with the membrane-bound sterols (Kinsky et al., 1966). Lichtenstein and Leaf (1965) and Bentley (1968) have shown that amphotericin B stimulates Na+ transport in toad urinary bladder when added to the apical side. The permeabilities of the bladder and frog skin to small solutes, e.g., chloride and urea, are also greatly increased by amphotericin B (Lichtenstein and Leaf, 1965; Nielsen, 1971). Normal K + fluxes across short-circuited frog skin are small, but the addition of amphotericin B to the outside bathing solution (OBS) results in a considerable increase in transepithelial K + fluxes. In this circumstance the K + efflux is always enhanced more than the K’ influx (Nielsen, 1971, 1972), giving a flux ratio different from unity. [In referring to transepithelial fluxes, the term “efflux” means movement from the anatomical inside of the epithelium to the anatomical outside: from the inner bathing solution (IBS), through the basolateral membrane (BLM), through the apical membrane (AM), and into the outer bathing solution (OBS). The AM BLM term “influx” refers to the opposite movement: OBS IBS. These usages should be kept distinct from the conventional usages applied to single cells, which have only one (plasma) membrane and one (external) bathing solution.] Qualitatively, the same results have been obtained with nystatin and Cu2+on frog skin (Bakhteeva and Natochin, 1975; Ferreira, 1978), and with amphotericin B on rabbit colon and toad bladder (Frizzell and Turnheim, 1978; Gatzy et al., 1979). Thus it is clear that polyene antibiotics increase epithelial K + permeability, as well as permeability to N a + , C1-, and urea. In the experiments mentioned above, the epithelia were short-circuited and the bathing solutions on both sides were identical, so the flux ratio for passive fluxes should have been unity (Ussing, 1949) if K + had been in a steady state. However, during incubation of the epithelia before addition of the antibiotic, the cells exchange some K + with 42K+in the inside bathing solution (IBS, across the basolateral membrane), but not with 42K+ in the OBS (across the apical membrane, which is nearly impermeable to K + ) . Thus there is a difference in the specific activity of cell K + in the skin halves on which the K + influx and efflux are measured. Addition of polyene antibiotics to the OBS increases the K + permeability of the apical membrane.
-
-
-
6.
Na+-K+
PUMP MEDIATING TRANSEPITHELIAL
Na+
TRANSPORT
95
In this situation the K + in the cells is partially in equilibrium with the 42K+ in the IBS but not with the 42K+in the OBS. Therefore the specific activity of the K’ moving from the cells to the OBS (i.e., in the efflux experiment) will normally be higher than the specific activity of the K + moving from the cells to the IBS (i.e., in an influx experiment); and the flux ratio (effluxlinflux) will appear to be greater than 1, even though the fluxes of K + are passive. Clearly, in this circumstance a flux ratio greater than unity cannot be taken as proof of active outward transport of K + across the epithelium. Only if the “net flux of 42K+” (efflux-influx) were very large, compared to total K + in the epithelium, could it be concluded that K+ is transported actively across the epithelium. (The term “net flux of 42K+” has been used in quotation marks because it does not represent steady state flux and the net tracer flux is not equal to the net chemical K + flux across the epithelium.) This indeed has been shown for K + movement in the presence of polyene antibiotics, at least in the cases of frog skin and toad bladder (Nielsen, 1971; Gatzy et a / . , 1979). Another method which can be used to investigate whether K + is transported actively across epithelia is to measure the net changes in the K+ content of the IBS and the OBS (e.g., by flame photometry) under shortcircuit conditions. By this method, it has been shown that-in the presence of the polyene antibiotic filipin-there is a decrease in the K + concentration of the IBS (isolated frog skin) and an increase in the K + concentration of the OBS, as shown in Table I (Nielsen, 1979a). Since the experiments were conducted under short-circuited conditions with identical solutions on each side, and since the changes cannot be ascribed to electrode processes (Ag-AgClelectrodes were used to pass current), it seems safe to conclude that the changes were caused by active transepithelial K + transport. It is also possible to investigate whether the transepithelial K’ transport is passive or active by using a non-steady state flux ratio analysis. However, the experiments must be conducted in a manner which is somewhat different from the usual flux measurement procedure. To investigate, e.g., whether K + is actively transported across the frog skin after addition of filipin, two symmetrical skin halves are incubated under short-circuit conditions with identical solutions on both sides. Filipin is added to the OBS, and when the apical membrane has become permeable to K + , 42K+ is added to the OBS of one skin half and to the IBS of the other skin half. Then samples are taken at suitable time intervals, and the separate 42K+fluxe~ are measured and calculated in the usual way. When the experiment is carried out in this way, the flux ratio equation is valid for non-steady state fluxes, provided the measuring period is small compared with the rate of change in the permeability of the skin. [Even if the permeability of the system changes considerably during the flux measurements, the correct flux ratio,
EFFECTOF 5 x i t 5 M
FILIPINON THE
K+
CONTENT OF THE
TABLE I RINGER’S SOLUTIONS BATHING THE
INSIDE AND THE OUTSIDE OF THE ISOLATED
5 x i t 5 M filipin
Control Inside a 0-2 hours GEq) b 2-4 hours GEq) c 0-4 hours GEq) d Wet Weight e Potassium content GEq) f Potassium content plus loss GEq)
0.24 0.45 0.69
+ +
(8.39
+
0.28 0.15
FROGSKINO
Outside
0.21 f 0.12 0.15 f 0.16 0.36 262 f 28 8.39 f 0.75 0.45 + 0.60) = 9.44
A Loss
0.45 0.60
Inside -0.83 -2.85 - 3.68
(7.43
Outside
* 0.15 zt
+
1.86 f 0.28 3.85 + 0.24 5.71 253 f 26 7.43 f 0.45 1.03 + 1.00) = 9.46 0.14
A Loss
1.03 1 .oo
Filipin was added to the outside bathing solution. Values are the mean f 1 SE of six experiments. Lines a-c show the change in the K + content of the inside and outside bathing solutions. Line d shows the wet weight of the skin halves after incubation. Line e shows the K + content of the skin halves after incubation. Line f gives the starting K + content of the skin halves, computed from line e plus the loss. K + loss from the skin halves (A loss) is calculated from the change in the K + content of the bathing solutions (A outside plus A inside). (From Nielsen, 1979a.)
6.
Na+-K+
PUMP MEDIATING TRANSEPITHELIAL
Na+
TRANSPORT
97
at the time of isotope addition, can be obtained by extrapolation of the measured flux ratios back to this time (Ussing, 1978)l. From such nonsteady state flux experiments, both K + and Rb+ indeed appear to be actively transported from the IBS t o the OBS of the isolated frog skin (Nielsen, 1972, 1979a). 4. EFFECTS OF Na+ AND OUABAIN ON K + TRANSPORT
With Na+ present in the IBS but omitted from the OBS-thus making the active transepithelial Na+ transport zero-the non-steady state 42K+ flux ratio measured after the addition of filipin is 1 (Table 11). Thus movement of K + across the epithelium is clearly passive. But replacement of Na+ in the IBS by Tris or choline, leaving normal Na+ in the OBS, also results in a 30-40% inhibition of Na+ transport. Under these conditions the non-steady state 42K+flux ratio after the addition of filipin is very different from 1 (Table 111). Taken together, these observations mean that active transepithelial K + transport requires Na+ in the OBS but not in the IBS. Furthermore, separate non-steady state flux experiments have shown that ouabain completely inhibits active transepithelial K transport (Nielsen, 1979a) as well as the transepithelial Na+ transport. We now have at last three solid experimental reasons for supposing that the transepithelial movements of Na+ and K + take place via the same mechanism: Both ionic fluxes require Na+ in the OBS and K + in the IBS, both are inhibited by ouabain, and the two fluxes change in a closely correlated manner during inhibition by amiloride (Nielsen, 1979a). +
C. Use of the Polyene Antibiotic Filipin as a Tool for Determining Coupling between Na+ and K + Transport The data in the previous section indicate that filipin-induced transepithelial K + transport is coupled to the transepithelial Na+ transport. In order to estimate the coupling ratio @) between active Na+ and active K + transport one must measure the net Na+ and K + fluxes through the pump. The strategy for doing this, again by the use of filipin-treated frog skins, is discussed below. OF Jfuump 1. DETERMINATION
With filipin present in the OBS, active transport of K + occurs across the isolated frog skin, from the IBS to the OBS (Table I). But the decrease in the K + content of the IBS (AKi) is different from the increase in the K +
€1"
11'1
01 0 5
10'1 58'0
8Z' I P8'0 62' I 8E'I
IE'I ZO' I
56'0 98'0 ZI'I
001
PO' I
10'1
90' I 91'1
PL'O
PO' I
PP' I 61'1 86'0
S 01
6.
Na+-K+
PUMP MEDIATING TRANSEPITHELIAL
Na+
99
TRANSPORT
TABLE 111 FLUXRATIOFOR 42K+ WHENN a + Is ABSENTFROM THE INSIDE BATHINGSOLUTION" Flux ratio Exp. no.
Inside
1
Tris Tris Choline Choline Choline Choline
Minutes 42K+ added after filipin
0-30 minutes
30-60 minutes
60-90 minutes
90-120 minutes
5.42 2.57 5.69 15.29 5.89 7.08
5.35 2.51 3.41 12.41 7.91 22.79
4.95 2.14 1.94 10.71 4.89 -
3.71 1.57 1.68 6.79 4.13
~~
2 3 4 5 6
25 5 0 5 30 10
-
"The 42K+ influx and efflux were measured simultaneously in symmetrical skin halves. The skin halves were incubated with Na-Ringer's solution on the outside and with Tris- or choline-Ringer's solution on the inside. Filipin (50 pl4) was added t o the outside. 42K+ was added 0-30 minutes after the filipin. The first sample (time 0 in the table) was taken 2 minutes after the addition of 42Kt. The flux ratio is calculated as efflux/influx. (From Nielsen, 1979a.)
content of the OBS (AK,,) (Table I), indicating that a net loss of K' occurs from the skin t o the bathing solutions. Thus the skins are not in a steady state, and the non-steady state flux analysis must be applied. Net K + transport across the frog skin (K,,,), then, is equal to the measured uptake from the IBS (AKi) plus the K + loss from the skin to the IBS ( K p ) plus the from the OBS to the IBS (Fig. 4A). At the net K + back transport (Kback) start of the experiment Kback is zero, but during incubation the K + concentration decreases in the IBS and increases in the OBS, because of the active K + transport across the skin; and the resultant K + gradient across the skin drives Kback.
A
FIG.4. (A) Parameters necessary for calculation of K,,,. (B) Diagram showing that net K + flux via the pump is equal to the net flux to the OBS plus the next flux t o the IBS, via passive pathways.
100
ROBERT NIELSEN
The net potassium flux under short-circuit conditions, with identical solutions on the two sides of the epithelium, can be written K,,, = AKi + K F
+ Kback
(la)
AK, - K F
+ Kback
(lb)
or, similarly, K,,,
=
where AK, is the measured increase in the K + content of the OBS and K P is the K + loss from the skin to the OBS. The net K + flux through the pump (J,;;;) is equal to the net K + flux from the transport compartment to the OBS via passive pathways plus the net K + flux from the transport compartment to the IBS via passive pathways (Fig. 4B). Since the total amount of K + which goes to the IBS and the OBS depends on the ratio between the K + permeabilities of the inward- and outward-facing membranes and upon the incubation time (t),
where P i a n d P,",, are the K + permeabilities of the two membranes. The first term on the right side of Eq. (2) is equal to the amount of K + which goes to the IBS, and the second term is equal to the amount of K + which goes to the OBS and also equal to the net amount of K' transported across the skin; thus Knet = t JZib P : u t /(Pi: + Ptut 1
(3)
By substituting Eq. (la) into Eq. (3) and rearranging, we obtain J,":;
t = (AKi + K:OSs
+ Kback)(l+ Pi:/P& )
(44
By substituting Eq. (lb) into Eq. (3) we obtain
J,Ki; t
=
(AKi - K F
+ Kback)(l+ Pi: /P,",, )
(4b)
The use of these equations in determining p will be discussed later 2. DETERMINATION OF J:;:;
Under normal conditions net Na' transport across the isolated frog skin is equal t o the integrated SCC (t x SCC = ISCC, Ussing and Zerahn, 1951). But under steady state conditions, in the presence of an active outward K + transport, the ISCC is equal to ISCC = Na,,, - K,,,
(5)
where Na,,, and K,,, are the net amount of Na+ and K + , respectively, transported across the skin during the incubation. However, the skin is not
6.
Na+-K+ PUMP
MEDIATING TRANSEPITHELIAL
Na+
101
TRANSPORT
in a steady state after the addition of filipin, and under non-steady state conditions Kne,can be expressed by Eq. (la). Similarly, the ISCC in Eq. (5) must be corrected, as is obvious from the fact that the addition of filipin to the OBS results in swelling of the cells (Nielsen, 1977) and uptake of Na+, C1-, and H,O from the bathing solutions. In fact, at least three separate corrections must be made. Since the outward-facing membrane is more permeable to Na+ than the inwardfacing membrane, Na+ uptake should occur mainly from the OBS. The ratio of C1- permeabilities for the apical and basolateral membranes is not known under the circumstances of these experiments, but if it is not very different from unity, then a fraction of the Na+ taken up from the OBS during cellular swelling must be balanced by the uptake of C1- from the IBS. The uptake of C1- from the IBS plus the uptake of Na+ from the OBS will give an electric current without transepithelial Na+ flux; and this contribution to the ISCC is designated (Fig. 5). K + back transport (Pack) also gives a current without concomitant Na+ transport. And, because loss of K + from the cells to the IBS ( K p ) is largely balanced by Na+ uptake from the OBS (due to the predominant Na+ permeability of the apical membrane), this too should give a transepithelial current without a transepithelial sodium flux. In summary, under non-steady state conditions, Eq. ( 5 ) becomes ISCC - Aswell - K b a c k - KIoss = Na net - K n e t (6) CI Substituting Eq. (la) into Eq. (6) and canceling corresponding terms yields ISCC - AFll = Nan,, - AKi
(7)
It has been shown that AZll is small compared with the other terms in Eq. (7) (Nielsen, 1979a), so that under the conditions used in these experiments net transepithelial Na + transport can reasonably be calculated as Nane,=ISCC + AK, OBS
(8) IBS
c1FIG. 5 . Definition of the parameters which give a SCC without a concomitant net Na' flux.
A : '
K back
Na+
Ki loss
102
ROBERT NIELSEN
3. DETERMINATION OF p
The coupling ratio (p) is equal to net Na+ transport through the pump, divided by the net K + transport through the pump: /3 =
J,N,amn~t /J$$
(9)
By substituting Eqs. (4a) and (8) into Eq. (9), we obtain
P=
+ AK, p + K back)(l+ Pi: /P$, ) ISCC
(AKi + K
(1 0 4
By substituting Eqs. (4b) and (8) into Eq. (9), we obtain
ISCC
P=
(AK, - K F
+ AKi
+ K back)(l+ Pi: /Po",,)
(lob)
The only terms which are known in Eqs. (10a and b) are ISCC, AKi and AK,. Therefore, in order to estimate from Eq. (lOa), the experiments must be conducted under conditions where K / O S s and Kbackare small compared with AKi, and where (1 + Pi: /Po",,) approaches unity, which means P; >> Pi:. Ba2+has been shown to decrease the K + permeability of frog heart cell membranes (Hermsmeyer and Sperelakis, 1970), of frog muscle membranes (Henderson, 1974; Sjodin and Ortiz, 1975), and of the basolateral membrane of isolated frog skin (Nagel, 1979; Nielsen, 1979a,b). Therefore, if the experiments are conducted with Ba2+in the IBS (Ba2+decreases P:,) and with filipin in the OBS (filipin increases P$, ), the expression 1 + P i /P&,should approach unity. Furthermore, if the incubation period is short enough that the K + gradient across the skin caused should also by active transport of K + does not become too large, then Kback be small. Then it should be possible to estimate /3 from reduced forms of Eqs. (10a) and (lob):
p
=
(ISCC
+ AKi)/AKi
(1 la)
=
(ISCC
+ AKi)/AK,
(1 1b)
or
[In Eq. (lla), it is assumed that all K + lost from the skin is lost to the OBS, so that K,1"""is equal to zero. Similarly in Eq. (llb), it is assumed that all K+ lost from the skin is lost to the IBS, thus making K F equal to zero.] Now, experiments carried out with filipin in the OBS and Ba2+ in the IBS show a good correlation (r = 0.97) between active Na+ transport (ISCC + Mi)across the isolated frog skin and K + uptake from the IBS (AKJ (Fig. 6 ) . From the slope of the (linear) regression line, /3 was found equal to 2.35. This estimate, however, must be too high, because the use of
6.
Na+-K+
PUMP MEDIATING TRANSEPITHELIAL
Na+ TRANSPORT
103
L
a
4 2.0. n
Regression line: Y = 0.054 + 0 . 4 2 5 ~ r = 0.966
+I
3
+'
Y
4 1.5a
1.01
0.5.
00
1
2
3
4
peq net Nattransport FIG. 6 . Coupling between K + uptake from the IBS (AK,) and Na+ transport. Net Na+ transport is calculated from ISCC + AKi. The skins were incubated in the presence of 10 pM amiloride and 50 pM filipin in the OBS, with 5 mM Ba" in the IBS. (From Nielsen, 1979a.)
Eq. ( l l a ) assumes that all K + lost from the skin is lost to the OBS. In the presence of ouabain, K+ lost from the skin is lost to the IBS (Nielsen, 1979a), and if this occurs in the absence of ouabain, then Eq. ( l l b ) should give a better estimate of /3 than Eq. (1 la). The experiments in Fig. 7 show that there is also a good correlation ( r = 0.97) between active transepithelial Na+ transport and the increase in the K + content of the OBS (AKJ. The slope of this regression line (Fig. 7) gives 0 equal to 1.53. Thus the experiments in Figs. 6 and 7 indicate that there is indeed a coupling between active Na+ and K + transport but shows that the coupling ratio (Na+/K+) must be smaller than 2.34. Other data obtained by the method described above indicate that /3 may be equal to 1.5 (Nielsen, 1979a).
D. Determination of the Coupling Ratio by Blocking the K + Channel The data obtained by the use of filipin indicate that active transepithelial Na+ transport is carried out by a coupled Na+-K+ pump. If the coupling ratio of the pump is 3 Na+/2 K + (= 1.5), then-while essentially all the SCC across the apical membrane is carried by Na+-only one-third of the
104
ROBERT NIELSEN
t '
Regression line: Y=0.098+0 . 6 5 ~ r = 0.972
-f?! 3-
+'Y
ga20
l a
00
1
2
3 4 )req net Na+-transport
FIG.7. Coupling between K+ release, from the skin to the OBS (AK,,), and N a + transport. Net Na+ transport is calculated from ISCC + AKi. The skins were incubated in the presence of 10 pM amiloride and 50 pM filipin in the OBS, with 5 m M Ba2' in the IBS.(From Nielsen, 1979a .)
SCC across the basolateral membrane can be carried by N a + , via the Na+-K+ pump; the rest must be carried by K + , via "K+ channels" (Fig. 8). Thus the addition of a substance which completely blocks the K+ channels (in the absence of filipin) should initially reduce the SCC by two-thirds. Accordingly, Ba2+ added to the IBS of isolated frog skin epithelium
Na+-
FIG. 8. The two-membrane hypothesis, drawn for an Na+-Kf pump having a coupling ratio of 1.5 (3 Na/2 K). P, Na+-K+ pump.
,
.------*
6.
Na+-K+
A
Na+ TRANSPORT
105
E v)
6oa
40
PUMP MEDIATING TRANSEPITHELIAL
4
-
201
0-
-
b
$ 7
b 0
1
2
3
4
5
mM Ba2+ FIG.10. Maximum percentage reduction of the SCC elicited by various concentrations of Ba2+. Values are means + 1 SE. Figures on the curve indicate the number of experiments at each point. (From Nielsen, 1979b.)
106
ROBERT NIELSEN
(without filipin in the OBS) results in a prompt reduction in the SCC (Fig. 9), followed by a much slower recovery (which is presumed to reflect secondary events). The initial Ba2+-induced inhibition of the SCC is plotted against the Ba2+ concentration in the IBS (Fig. lo), revealing a hyperbolic relationship between the BaZf-induced inhibition and increasing Ba2+ concentration. At saturation 65% (f1.9%) of the control SCC is abolished by barium, averaged for 10 experiments. This fraction is not significantly different from the initial two-thirds inhibition of the SCC to be expected for a coupling ratio of 1.5 in the Na+-K+ pump. It is therefore concluded that the coupling ratio for the Na-K pump in frog skin is indeed 1.5, as is generally accepted for symmetric cells such as erythrocytes, muscle, and nerve (Thomas, 1972).
111.
CONCLUSIONS
Measurements of the changes in the intracellular potential under different conditions have revealed that there is an electrogenic Na+ pump in epithelia. When the entire tissue is used (epithelium plus the underlying dermis), no correlation can be found between the transepithelial active Na+ transport and the K + flux across the basolateral membrane. But when isolated epithelia are used, some of the results support the notion of Na+-K+ coupling. The discrepancy between data obtained on whole tissues and those from isolated epithelia probably arises from the recycling of K + ions. When the apical membrane is made K +-permeable with polyene antibiotics, an active outward transport of K + can be observed. This K+ transport is activated by Na+ and inhibited by ouabain and amiloride, indicating strongly that the same mechanism is responsible for both fluxes. The determination of the coupling ratio @) for the active transepithelial Na+ and K + transport (in the presence of filipin and Ba2+)indicates it to be smaller than 2.34. With the K + channel in the basolateral membrane blocked, when the isolated frog skin is incubated in normal Ringer's solution, a value of 1.5 is found for /3. Thus it is concluded that a coupled electrogenic Na+-K+ pump, with /3= 1.5, is responsible for active transepithelial Na+ transport in frog skin. ACKNOWLEDGMENT This work has been supported by a grant from the Danish Natural Science Research Council (511-15846).
6.
Na+-K+
PUMP MEDIATING TRANSEPITHELIAL
Na+ TRANSPORT
107
REFERENCES Bakhteeva, V. T., and Natochin, Y. V. (1975). Lechenow Physiol. J . USSR 61, 1242-1248. Bentley, P. J. (1968). J. Physiol. (London) 196, 703-711. Biber, T. U. L., Aceves, J . , and Mandel, J. L. (1972). A m . J. Physiol. 222, 1366-1373. Bonting, S. L. (1970). In “Membranes and Ion Transport” (E.E. Bittar, ed., Vol. 1, pp. 257-363. Wiley (Interscience), London. Bracho, H., Erlij, D., and Martinez-Palomo, A. (1971). J . Physiol. (London) 213, 50P-51P. Cala, P. M., Cogswell, N., and Mandel, L. J. (1978). J. Gen. Physiol. 71, 347-367. Candia, 0. A,, and Zadunaisky, J. A. (1972). Biochim. Biophys. Acta 255, 517-529. Curran, P. F., and Cereijido, M. (1965). J. Gen. Physiol. 48, 1011-1033. . DeLong, J., and Civan, M. M. (1978). J . Membr. Biol. 42, 19-43. Essig, A., and Leaf, A. (1963). J . Gen. Physiol. 46, 505-515. Farquhar, M. C., and Palade, C. E. (1965). J. Cell Biol. 26, 263-291. Farquhar, M. C., and Palade, C. E. (1966). J. Cell Biol. 30, 359-379. Ferreira, K. T. C. (1978). Biochim. Biophys. Acta 510, 298-304. Ferreira, K . T. (1979). Biochim. Biophys. Acta 555, 13-25. Finn, A. L., and Nellans, H. (1972). J . Membr. Biol. 8, 189-203. Frazier, K., and Leaf, A. (1963). J. Gen. Physiol. 46, 491-593. Frizzell, R. A., and Turnheim, K. (1978). J. Membr. Biol. 40, 193-211. Gatzy, J . T., Reuss, L., and Finn, A. L. (1979). A m . J . Physiol. 237, F145-F156. Harris, E. J . , and Burn, C. P. (1949). Trans. Faraday SOC. 45, 508-528. Henderson, E. C. (1974). Pfluegers Arch. 350, 81-95. Hermsmeyer, K., and Sperelakis, N. (1970). A m . J . Physiol. 219, 1108-1114. Huf, E. C., and Wills, J. (1951). A m . J. Physiol. 167, 255. Kinsky, S. C., Luze, L. A., and Van Deene, L. L. M. (1966). Fed. Proc. Fed. A m . SOC. Exp. Biol. 25, 1503-1510. Koefoed-Johnsen, V. (1957). Acta Physiol. Scand. (Suppl.) 145, 87-88. Koefoed-Johnsen, V., and Ussing, H. H. (1958). Acta*Physiol. Scand. 42, 298-308. Leaf, A., Andersen, J., and Page, L. B. (1958). J. Gen. Physiol. 41, 657-668. Lewis, S. A., Eaton, D. C., Clausen, C., and Diamond, J. M. (1977). J . Gen. Physiol. 70, 427-440. Lewis, S. A., Wills, N. K., and Eaton, D. C. (1978). J . Membr. Biol. 41, 117-148. Lichtenstein, N. S., and Leaf, A. (1965). J. Clin. Invest. 44, 1328-1342. Macknight, A. D. C., and Leaf, A. (1978). J. Membr. Biol. Special Issue, 247-260. Miller, S. C., Steinberg, R. H., and Oakley, B. (1978). J. Membr. Biol. 44, 259-279. Mills, J. W., Ernst, S. A., and DiBona, D. R. (1977). J. Cell Biol. 73, 88-110. Nagel, W. (1978). J. Physiol. (London) 284, 146 P. Nagel, W. (1979). Biochim. Biophys. Acta 552, 346-357. Nellans, H. N., and Schultz, S. C. (1976). J. Gen. Physiol. 68, 441-463. Nielsen, R. (1971). Acta Physiol. Scand. 83, 106-114. Nielsen, R. (1972). J. Steroid Biochem. 3, 121-128. Nielsen, R. (1977). Acta Physiol. Scand. 99, 399-411. Nielsen, R. (1979a). J. Membr. Biol. 51, 161-184. Nielsen, R . (1979b). Acta Physiol. Scand. 107, 189-191. Robinson, B. A., and Macknight, A. D. C. (1976a). J. Membr. Biol. 26, 239-268. Robinson, B. A., and Macknight, A. D. C. (1976b). J. Membr. Biol. 26, 269-286. Schultz, S. C. (1978). In “Membrane Transport Processes” (J. F. Hoffman, Ed.), Vol. 1, pp. 213-227. Raven, New York.
108
ROBERT NIELSEN
Sjodin, R. A., and Ortiz, 0. (1975). J. Gen. Physiol. 66, 269-286. Thomas, R. C. (1972). Physiol. Rev. 52, 563-594. Ussing, H.H . (1949). Acfu Physiol. Scand. 19, 43-56. Ussing, H. H. (1978). I n “Membrane Transport in Biology” (G. Giebisch, D. C. Tosteson, and H. H. Ussing, ed.), Vol. 1, pp. 115-140. Springer-Verlag, Berlin and New York. Ussing, H. H., and Zerahn, K. (1951). Acfa Physiol. Scand. 23, 110-127. Valenzeno, D. P., and Hoshiko, T. (1977). Biochim. Biophys. Acfu 470, 273-289. Varanda, W. A., and Lacaz-Vieira, F. (1979). J. Membr. Biol. 49, 199-233. Vieira, F. L., Nunes, M. A., and Cury, L. (1976). J. Membr. Biol. 27, 251-264. Wills, N. K., Lewis, S. A., and Eaton, D. L. (1979). J. Membr. Biol. 45, 81-108.
CURRENT TOPICS IN MEMBRANES A N D TRANSPORT, VOLUME 16
Chapter 7 Transepithelial Potassiurn Transport in Insect Midgut by an Electrogenic Alkali Metal Ion Pump MICHAEL G . WOLFERSBERGER, WILLIAM R . HAR VEY, AND MOIRA CIOFFI Department of Biology Temple University Philadelphia, Pennsylvania
1. 11.
Introduction ..................................................................... Methods ...............................
...............
.................................
..................... D. A Special Inhibitor
.......
A. Morphology and Fine Structure ..................................... B. Pump Location ........................................ C. Pool Location ......................................................................... V . A Potassium Transport ATPase .....................................
I.
109 111 111 112 113 114 114 116 116 119 122 123 123 125 126 129 132
INTRODUCTION
Voracious consumption of leafy material by lepidopteran larvae places an unusual requirement upon these insects’ digestive apparatus: Because 109
Copyright @ 1982 by Academic Press, lnc. All rights of reproduction in any form reserved. ISBN 0-12-1 533 16-6
110
MICHAEL G. WOLFERSBERGER
et a/.
leaf tissue is rich in potassium and poor in sodium, it is potassium ions which must be either retained in the gut or secreted from the blood in order to maintain ionic balance. As a point of reference, the normal cation concentration in midgut contents, midgut tissue, and the blood of fifth-instar cecropia larvae (Hyalophora cecropia) are summarized in Table I. With respect t o potassium and sodium distribution, two major facts stand out: First, at all three loci potassium is the dominant cation, being about 860-fold more concentrated than sodium in the midgut contents and 12-fold more concentrated than sodium even in the blood. This suggests, by comparison with the situation for other insects, that lepidopteran larvae have partially adapted to the dietary potassium stress by allowing blood potassium to rise and blood sodium t o fall (Florkin and Jeuniaux, 1973). One striking feature of this adaptation is that lepidopteran midgut apparently does not contain the supposedly ubiquitous cellular Na+-K+ pump (Section V). The second outstanding fact, however, is that potassium is nevertheless 12-fold less concentrated in the blood than in the midgut contents. In fact, except for protons, potassium sustains the largest concentration ratio between gut and blood (or vice versa) of any of the common inorganic cations. Although it is not generally recognized, (Stobbart and Shaw, 1974), the midgut potassium pump of lepidopteran larvae plays a major role in determining blood K + levels (Maddrell, 1971; Harvey and Blankemeyer, 1975; Harvey, 1980). It has turned out that the lepidopteran midgut K + transport system is useful to study not only for an understanding of insect physiology (and, from a practical standpoint, insect population control) but also for insight TABLE I CATION CONCENTRATIONS IN NORMAL, MATURE, FEEDING, FIFTH-INSTAR Hyulophoru cecropia LARVAE~
Cation
K+ Na' MgZt Ca2H + (pH)
Midgut contents
Midgut tissue
284.0 f 51.0 0.33 f 0.10 18.1 f 2.2 38.2 f 6.9 9.7c
90.2 f 4.1b 0.60 f 0.13 26.9 f 1.2 6.86 f 1.71
Blood
22.8 f 1.0 1.98 f 0.32 70.5 f 5.5 13.6 f 0.6 6.5
Data from Harvey er a/. (1975). Ion concentrations in millequivalents per liter of water 1 SEM. Zerahn (1975) reported a potassium concentration of 137 mmoles/liter cell water. ' This is the average pH, but there can be a strong axial gradient. In the tobacco hornworm larva (Munducu sextu), midgut pH is 11.O in the midregion, but 9.5 at the anterior end and 8.5 at the posterior end. (From Dickenson, Cioffi, and Harvey, unpublished results.) f
7. INSECT ELECTROGENIC CATION PUMP
111
into the mechanisms of active electrogenic cation transport by epithelial membranes in general (Blankemeyer and Harvey, 1977). Both anatomically and physiologically the midgut epithelium is a convenient model tissue: It is one cell thick, is in direct contact with oxygen entering via the tracheal system, and is relatively independent of central homeostatic controls (Wigglesworth, 1972). Furthermore, the epithelium is easily isolated and readily prepared for electrophysiological measurements, ion flux studies, or ultrastructural analysis. In the discussion that follows, we shall describe the main physiological features of the midgut K + transport system, with an emphasis on its electrical properties. We shall then review selected flux and morphological studies focused on the anatomical localization of the pump and its ion pools. And finally, we shall turn to the question of the chemical nature of the transport system. The older literature on potassium transport by lepidopteran midgut has been reviewed comprehensively by Keynes (1969), Maddrell (1971), Harvey and Zerahn (1972), and more recently by Zerahn (1977, 1978) and Harvey (1982).
II. METHODS A. Experimental Material For studies on the midgut K + pump, at least half a dozen different species of lepidopteran larvae have been used, including Bombyx mori, Philosamia Cynthia, Macrothylatia rubi (Giordana and Sacchi, 1977), and Antheraea pernyi (Wood, 1972). The majority of work, however, has been carried out on the giant silkworm, H. cecropia, and on the tobacco hornworm, Manduca sexta, which has become the species of choice because of its commercial availability. Under comparable conditions of rearing and experimental manipulation, there are no significant qualitative differences in the physiological behavior of these different midgut preparations, but for all the species treated, the type of diet upon which the larva is reared can have a profound effect on the magnitude of the ionic pool accessible to the K + pump. This problem will be discussed in Section IV,C. Ordinarily, lepidopteran larvae are reared on either foliage or an artificial diet and are used in the fifth-instar stage, just prior to the onset of pupation (see Haskell et al., 1968). For the experiments, each larva is chilled in crushed ice and then opened longitudinally to expose the midgut. The midgut is excised and slit to form a flat sheet. After removal of the peritrophic membrane, the desired region (Section IV) of the epithelial sheet is mounted on the chamber diagramed in Fig. 2 (see Harvey and Wolfersberger, 1979). Although the bathing solutions used in the two por-
112
MICHAEL G. WOLFERSBERGER
et a/.
tions of the chamber may vary, depending upon experimental objectives, optimal transport (judged from measurements of short-circuit current) is usually obtained with solutions containing 32 m M KCl(70 m M for high-K+ medium; 8 m M for low-K* medium), 5 m M Tris buffer (or bicarbonate buffer), pH 8, 1 m M CaC12, 1 m M MgC12,and 166 mM sucrose; this solution is somewhat hypotonic to the larval blood, but it does not damage the epithelium and it can readily be supplemented with moderate amounts of additional salt. B. Epithelial Anatomy and Transport Nomenclature The lepidopteran midgut epithelium is composed mainly of two kinds of cells, as diagramed in Fig. 1: columnar cells (CC), whose apical (luminal) surface is greatly elaborated with microvilli; and goblet cells (GC) containing a large cavity into which the primary K + secretion process probably
Basal
Apical (lumen) side
0
FIG. 1. Diagram of the lepidopteran larval midgut. CC, Columnar cells; GC, goblet cells. Less numerous regenerative cells (not shown) are scattered at the base. Locations of the portasomes (T studdings), the K + pump (heavy arrow), and closed lateral junctions (heavy X’s) anticipate the discussions of Section I V .
7. INSECT ELECTROGENIC CATION PUMP
113
operates (see also Section IV,A and Fig. 8). Scattered along the basal (blood) side of the epithelium are small regenerative cells which are not included in Fig. 1. The lateral couplings between adjacent cells are normally inoperative in midguts from foliage-reared larvae. In order to avoid the obvious problems which can arise in trying t o apply the conventional physiological terms “influx” and “efflux” to transepithelial transport (Harvey and Nedergaard, 1964; Wood el al., 1975), this article will use a strictly anatomical nomenclature to refer to the different fluxes. Specifically, BA flux = C$BA = unidirectional flux from basal to apical side; AB flux = C$AB = unidirectional flux from apical to basal side; - C$ab = net flux from basal to apical side; A, a = apical JBA= (luminal side); B, b = basal (blood side); C , c = cell; uppercase letters = the larger of two fluxes; foreflux = flux through the pump in the same direction as net flux; backflux = flux through the pump in the direction opposing net flux; positive (current or flux) = from the basal to the apical side (Fig. 1) of the epithelium for cations and from the apical t o basal side for anions; influx = movement strictly into single cells (across one plasma membrane); and efflux = movement strictly out of single cells (across one plasma membrane).
C. The Short-Circuit Technique Because of the midgut’s low resistivity (150 s2 cm2, compared with 180 s2 cm2 for the standard bathing solution), the classic short-circuit procedure-developed by Ussing and Zerahn (195 1) for high-resistance membranes-can give very large errors due to the series resistance of the solutions, as pointed out by Rehm (1968) and others. To deal with this problem, we use the chamber and multiple-electrode arrangement diagramed in Fig. 2 (Wood, 1972). The short-circuit current (SCC, I,,), to oppose the spontaneous transepithelial potential difference, is passed through plate electrodes D and E. Three-voltage recording electrodes, A to C , are arranged with equal spacing, so that B-C monitors the voltage drop through the solution alone and A-B monitors this plus the transepithelial potential difference. The difference between these two pairs thus gives the true transepithelial voltage difference, which is used to control the current actually passed by the voltage clamp (Wood and Moreton, 1978). The utility of measuring SSCs across the lepidopteran midgut lies of course in the fact that under well-defined conditions the net flux of accounts for more than 90% potassium from the blood to the lumen (JBA) of the SCC (Nedergaard and Harvey, 1968; Harvey and Wolfersberger,
114
MICHAEL G. WOLFERSBERGER
et
6‘1.
-E
D
A
II
II
B
C
FIG.2. Midgut recording chamber. The opened and flattened midgut (M) is attached to the flange of the left-hand compartment with loops (t) of cotton thread (exposed tissue area = 0.5 cm’). The two halves of the chamber are clamped together and sealed with a rubber washer (W). Three measuring bridges (A, B, C) are equally spaced, 0.5 cm apart, along the chamber. Two Ag-AgC1 current electrodes (D, E) are placed symmetrically 3 cm from the midgut. The chamber is made of plexiglass. (From Wood and Moreton, 1978, with permission from the Journal of Experimental Biology.)
1979; Cioffi and Harvey, 1981). Difficulties can arise with nonphysiological ion concentrations, both because all other alkali metal ions appear to be transported from blood to lumen (Harvey and Zerahn, 1972) and because calcium and magnesium are transported toward the blood (Wood and Harvey, 1976; Wood et a f . , 1975), but these fluxes contribute little to the SCC as long as K + comprises at least 50% of the total alkali metal cations in the bathing solutions and both Ca2+and Mg2+are present at 1 mM.
Ill.
BEHAVIOR OF THE MIDGUT K + TRANSPORT SYSTEM
A. Time Dependence
In H . cecropia the net potassium flux JBA,measured by isotope movement, is maximal a few minutes after midgut isolation and falls with time, reaching about 50% of the peak value after 1 hour. At that time, the is nearly equal to JBA,since the opposing flux unidirectional BA flux 4BA
115
7. INSECT ELECTROGENIC CATION PUMP
4BA.As potassium in the basal solution rises, 4 B A saturates, showing a K , of about 10 mM and a maximal velocity of about 25 peq c m 2 hr-l. The SCC and the open-circuit potential difference (PD, Vo)decrease ~, the transepithelial resistance (R,) rises, roughly in parallel with c # J ~while over a period of more than 3 hours following isolation of the midgut. These functions are displayed in Fig. 3 for a midgut perfused with the normal (32 mM K + ) solution. The main component of decline in both SCC and Vo occurs with a half-time of 158 minutes, but the SCC falls much more rapidly at first, with an apparent half-time of 16 minutes. The average values (five experiments) of all three parameters at 60 minutes are as follows: SCC = 498 f 160 pA/cmZ; Vo = 98 f 11 mV; and R, = 150 f 26 s2 cm2. The origin of the initial fast component of the SCC decline is not yet resolved. The early suggestion by Wood (1972) that it might be related to net loss of tissue K + was challenged by the finding of Zerahn (1975) that
qjab is only about 2% of
b -
I20
x
cu I
E
0
a
80
=L
.C
cu
E
60
I20
I80
TIME, min FIG. 3. Time course of the electrical parameters of a typical midgut. VO,Potential difference; R,, transepithelial resistance; SCC, short-circuit current. (From Wood and Moreton, 1978, with permission from the Journal of Experimental Biology.)
116
MICHAEL G. WOLFERSBERGER et
a/.
total tissue K + remains constant in the interval from 15 to 75 minutes. However, much of the SCC decrease is in the first 15 minutes, and the small decrease in tissue K + expected by loss from the goblet cell pool might have been missed. Recently Cioffi (unpublished results) noted that the tracer-measured BA flux is only a fraction of the SCC during this initial period, and she has evidence that 42K+ washes out from tissue to lumen faster than it moves from blood side to lumen side. The proposal by Schultz and Jungreis (1977) that the SCC decline arises from a massive release of plugs from goblet cell cavities has been contradicted by Cioffi (1980) who found that the so-called goblet plugs are instead apocrine secretion droplets from the columnar cells. Our present working hypothesis is that the initial rapid decline in the SCC is caused by a decrease in the concentration of goblet cell K + as the tissue adjusts from the 284 mM K + concentration of midgut contents in vivo to the 32 mM K + concentration in vitro. In any case, no one has succeeded in controlling the fast phase, whereas the slow-phase decline of the SCC can be temporarily reversed by adding dibutyryl-CAMP or a protein-free extract of larval blood to the bathing solution (Wolfersberger and Giangiacomo, 1980).
B. Ionic Dependence Figure 4 demonstrates that the SCC declines steeply as potassium is reduced below about 10 m M in the basal solution. This contrasts with the behavior of the open-circuit PD, which is nearly insensitive to the basal potassium level. Taken together, these observations imply that the total epithelial resistance, V,/SCC, increases when the pump is slowed by low potassium levels (Harvey and Zerahn, 1972). Solution pH is also a factor in determining the magnitude of the SCC, which is maximal with a basal-side however, pH of 8.5 and an apical-side pH of 9.0 or above (at least to 113; at least for changes in the basal solution, the decline of the SCC away from the optimum pH is slight: no more than 10% for +2.5 pH units (Wood,
-
1972).
C. Metabolic Dependence Potassium transport, and the associated electrical parameters of midgut epithelium, are strongely dependent upon the rate of oxygen consumption. In fact, since oxygen pressures below 0.7 atm actually limit total oxygen consumption in isolated lepidopteran tissue, it is simple to demonstrate restriction of flux or the SCC by lowered oxygen pressure (Fig. 5 ) . As expected, furthermore, membrane resistance also increases with lowered
117
7. INSECT ELECTROGENIC CATION PUMP
4-
i
CI
m
.-C
e x LI
0
V
L
0
-
a2 Do
m
C
u
8
E
t
I I
-
II I
I
' II
c PD
I I
7
f f
scc
I
FIG. 4 Potassium dependence of electrical parameters in isolated lepidopteran midgut. PD, Open-circuit potential difference; SCC, short-circuit current. Values on the ordinate scale are percentages of the PD and SCC at the control K t concentration of 32 mM. Control values at the start of the experiment: 98 mV and 1600 FA; at the end of the experiment: 79 mV and 1480 p A . Potassium was reduced by replacement of KC1 with osmotically equivalent amounts of sucrose. (From Harvey and Zerahn, 1972.) ~~
oxygen consumption, to the extent that pure nitrogen gas (which halts the pump) causes R, to double. Because of this obligatory dependence of K transport upon oxidative metabolism, and because of the reciprocal interdependence of transport and oxygen consumption which characterizes other epithelial transport systems, (see, e.g., Zerahn, 1956), we were surprised to discover (Harvey et al., 1967) that lepidopteran midgut preparations showed no dependence of +
MICHAEL G. WOLFERSBERGER ef
118
a/.
500
400
<
-t
300
% rA
200
100
30
90
60
120
I50
Time. min
FIG.5 . Oxygen dependence of the short-circuit current (SCC). Control periods (before 60 minutes, after 120 minutes): bathing solutions stirred with 100% oxygen. Experimental tests (60-120 minutes): oxygen reduced to the figure given, by the addition of nitrogen. Dashed curve, expected time course of SCC in maintained 100% oxygen. (From Wood, 1972, with permission from Dr. J . L. Wood.)
O 2 consumption upon K + transport, whether transport was varied by substituting different ions or by imposing different electrical gradients. This phenomenon may be accounted for, however, by a low steady state phosphate potential such that respiration is maximally stimulated at all times (Mandel et al., 1980a,b). Supporting evidence for this hypothesis comes from the fact that, although the normal tissue levels of ATP and ADP (1.2 and 0.43 ymoles/gm tissue, respectively) are within the range for other epithelia, free inorganic phosphate (Pi) is unusually high, about 10 mM. The phosphate potential, then, can be calculated as ATP/(ADP x Pi) = 280 M-I. Since no significant change occurs in the cellular concentration of ATP, ADP, or Pi, when the SCC is reduced by 65% (by decreasing the basal K + concentration from 32 to 8 mM), the phosphate potential appears to be stabilized at a value slightly below the range (300-3000 M-I) over which intact cells and tissues generally exhibit respiratory control (Wilson et al., 1974). [It should be noted that important dynamic parameters of respiratory metabolism are
7. INSECT ELECTROGENIC CATION
PUMP
119
quite normal in isolated mitochondria from midgut tissue. Respiratory control ratios of 5 (state 3/state 4)have been found, along with P/O ratios of 3 for NADH-linked substrates and 2 for succinate (M. sexta mitochondria; Mandel et al., 1980a,b).] It is interesting to speculate upon the function served by such metabolic extravagance in larval midgut tissue. Why should a tissue continue to burn oxygen when one of its principal energy-requiring functions is eliminated? Consider that the larva is surrounded by leafy food, large amounts of which must be consumed for essential nutrients other than carbohydrates. As an alternative to gross obesity, metabolic turnover-rather than strict energy conservation-may be important; the energy which must be wasted is presumably dissipated as heat. Although energy-wasting processes or futile metabolic cycles have been identified in other animal tissues, the identity of such processes in midgut remains unknown. Potassium transport is inhibited rather rapidly by anoxia: t , for the change in the P D is 1.5 minutes and t , for the change in the SCC is 2.1 minutes, so that within 10 minutes of oxygen removal both V, and SCC are essentially zero, while R, has doubled (Haskell et al., 1965; Wood, 1972; Blankemeyer, 1976; Harvey and Wolfersberger, 1979). On the basis of these observations, plus the fact that larval midgut contains a peculiarly high concentration of extramitochondrial cytochrome bs (Shappirio and Williams, 1957), it was suggested some time ago (Haskell et al., 1968) that the midgut K + transport system might operate as a redox pump. This notion was rejected by Mandel et af. (1975), because during the restoration of oxygen to nitrogen-treated tissue, oxidation of cytochrome b, occurs much more slowly than either the recovery of SCC or the reoxidation of mitochondria1 cytochromes. Furthermore, during the onset of anoxia, reduction of all tissue cytochromes occurred 2.5- to 3.0-fold faster than the decline of the SCC, implying the existence of a distinct metabolic reservoir between respiration and transport (Mandel et af., 1980b). This reservoir was identified with ATP, since tissue ATP does in fact decline in parallel with the SCC, as demonstrated in Fig. 6. The properties of the membranebound ATPase whose existence is implied by this result will be discussed in Section V.
D. A Special Inhibitor Agents which directly interfere with oxidative metabolism, such as dinitrophenol and iodoacetate, correspondingly inhibit potassium transport (Haskell et af., 1965), but no inhibitor has yet been certified which is analogous to the Na+ pump inhibitor, ouabain, in being highly specific for
120
MICHAEL G. WOLFERSBERGER
0
Tisue
et a/.
ATP
' 80 O O h
\\
cy I
+
N2
1
I
I
2
1
3
4
-4 5
20
Time (minutes)
FIG.6 . Comparison of the time courses of short-circuit current (SCC), cytochrome oxidation (cyt), and tissue ATP levels (plotted points) during the onset of anoxia. At zero time (vertical arrow) nitrogen-saturated solutions were exchanged for the control solutions. (From Mandel et al., 1980b, with permission from the American Physiological Society.)
the lepidopteran K + transport system itself. In recent years, however, much attention has been given to the &endotoxin of Bacillus thuringiensis (Bt), which-when administered in vivo-both interferes with K + regulation in the intact larva (Faust et al., 1967) and reduces the SCC in isolated midguts (M. sexta; Griego et al., 1979). Applied simply to the isolated gut, Bt has no effect on transport, but the toxin does alter transport parameters of the midgut if it is first incubated for a while at strongly alkaline pH (PH > 10; Harvey and Wolfersberger, 1979). [The toxin is activated in the insect gut by alkaline or enzymatic hydrolysis (Faust el al., 1967, 1974a,b).] The action of alkaline-treated toxin, or of a polypeptide fragment, is complex. Oxygen consumption is stimulated by 30%, and both the SCC and the electrical resistance of the midgut are reduced by 50-55%. Potassium BA flux (+BA) is unaffected, but the reverse flux (+ab) is enhanced threefold by Bt (Harvey and Wolfersberger, 1979). The ratio of +ab in oxygen to dabin nitrogen or in Bt is 0.3 (Blankemeyer, 1978; Harvey and Wolfersberger, 1979). The absence of an effect on +BA rules out at least three possible mechanisms for Bt action: increased exchange diffusion, increased facilitated diffusion, and a generalized increase in permeability. On the basis of these facts, Harvey and Wolfersberger have suggested
121
7. INSECT ELECTROGENIC CATION PUMP
that there may be two distinct components of active K + transport in lepidopteran midgut. These are diagramed in Fig. 7. The more obvious which is blocked by anoxia but not by Bt; the component is (part of) more subtle component is (part of) +ab, which is enhanced by anoxia or Bt and inhibited by oxygen in the absence of Bt. In pure nitrogen the reduced and and enhanced +ab would become equal, making the net flux (JBA) the SCC zero. This hypothesis is consistent with the further observation (Blankemeyer, 1978) that, under conditions where K + (but not Cs+) is actively transported, the a b flux of K f only (not Cs') is enhanced by nitrogen. B
(a )
A /
02
Active tl u 1
?,.
I
* Decreased
Increased B A - flux
FIG. 7. Organization of K + transport pathways in lepidopteran midgut, based on the action of &endotoxin (Bt) from B. thuringiensis. (a) Solutions saturated with oxygen, n o Bt added; (b) solutions saturated with oxygen, but Bt added; (c) solutions saturated with nitrogen. In all three diagrams, the upper arrow signifies BA flux and the lower arrow signifies ab flux. Anoxia (nitrogen) reduces BA flux but increases a b flux. The only effect of Bt is to increase ab flux.
122
MICHAEL G. WOLFERSBERGER
et a/.
While it may be unwise at the present time to ignore the possibility that the Bt effect on potassium transport might be mediated via oxidative metabolism, the following observations militate against such an interpretation: (1)The entire active BA flux and 40% of SCC are insensitive to BT but are reversibly blocked by anoxia; (2) in vivo, general body movements and heartbeat are spared for a considerable time after Bt has inhibited glucose uptake and stimulated the ab flux of potassium; (3) only a large (30,000dalton) peptide of Bt uncouples oxidative phosphorylation, whereas a 5000-dalton fraction-are toxic smaller components-particularly without uncoupling oxidative phosphorylation (Travers et al., 1976). The whole collection of information about Bt suggests that the toxin, or one of its hydrolytic products, in fact acts directly on the potassium transport system in the lepidopteran larval midgut.
E. Electrogenicity The observations presented in Figs. 3-6 are representative of the principal data from which it has been concluded that the K + pump in lepidopteran larval midguts is in fact electrogenic, meaning that the energycoupled reaction per se drives net charges through at least one surface of the epithelium. The apparent electromotive force (EMF) of the system is approximately 200 mV, and its operating efficiency (based on the free energy of ATP hydrolysis in midgut) is about 40% if 1 mole of K + is transported per mole of ATP hydrolyzed, but nearly 80% if 2 moles of K + are transported per mole of ATP hydrolyzed (Harvey et al., 1981). However, it is more difficult to be secure in these judgments for epithelial membranes-even those like the insect midgut epithelium, which are structurally simple (Fig. 1)-than for single-cell plasma membranes. Sandwiching of membranes, intermembranal and intercellular reservoirs and junctions, and the associated uncertainties of ion distribution all complicate the interpretation of transport measurements, since the direction and magnitude of transepithelial PDs and SCCs are determined by the sum of all active and passive ion movements in the tissue. Nevertheless, it can be said that no other single ion seems required to be transported in order for the K + pump to operate, and the obvious kinds of electroneutral coupled ion transport thus do not occur in lepidopteran larval midgut (Nedergaard and Harvey, 1968). The simplest interpretation of all the physiological observations therefore is that a bona fide electrogenic pump exists in either the apical or the basal (basolateral) barrier of the epithelium, translocating K + in the basal-to-apical direction.
7. INSECT ELECTROGENIC CATION PUMP
IV.
123
MEMBRANE STRUCTURE AND LOCATION OF TRANSPORT FUNCTIONS
A. Morphology and Fine Structure The detailed structure of the lepidopteran midgut has been described for H . cecropia by Anderson and Harvey (1966), for Ephestis kuhniella by Smith el al. (1969), for B. mori by Akai (1969), and recently for M. sexta by Cioffi (1979). The latter investigator has found that both the gross morphology and the fine structure of the major cell types (Fig. 1) change along the length of the midgut, allowing its division into structurally distinct anterior, middle, and posterior regions. To the unaided eye, the anterior and posterior regions appear thicker than the middle region, but this difference is caused mainly by variations in the degree of epithelial folding. The gut is dominated by six corrugated strips of tissue running parallel from the anterior end to the posterior end. Each strip is separated from its neighbors by a thin unfolded area, on the basal side of which runs a large longitudinal muscle. While the corrugated strips themselves dominate the midregion, the arrangement is thrown into elaborate secondary folds anteriorly and posteriorly. Changes in the structure of the goblet cells are most dramatic and most relevant to the present discussion, and Fig. 8 illustrates the main structural features of goblet cells. In the anterior and middle regions of the gut, goblet cells have a rounded basal region which tapers to a narrow neck apically. The apical membrane is invaginated to reach almost to the base of the cell, forming a large cavity which opens into the gut lumen via a narrow neck. Projections of the apical membrane into the cavity and are each filled with an elongated mitochondrion (Anderson and Harvey, 1966; Cioffi, 1979). In the posterior region of the midgut, however, the goblet cell cavity is restricted t o the apical two-thirds of the cell, and there is no neck region. Also, mitochondria do not extend into the membranous projections but are restricted t o the surrounding bulk cytoplasm. The latter arrangement (in posterior midgut) violates the generally accepted dogma (Keynes, 1973) that mitochondria should be intimately associated with plasma membranes in actively transporting cells, since posterior midgut is clearly capable of pumping potassium and generating substantial SCCs (Cioffi and Harvey, 1981). Undoubtedly the most characteristic feature of membrane fine structure in the goblet cells is that particles approximately 100 A long are invariably found on the cytoplasmic surface of the apical membrane in all three regions of the midgut (Anderson and Harvey, 1966; Cioffi, 1979). Such particles are generally found (in insect cells) associated with membranes
7. INSECT ELECTROGENIC CATION PUMP
125
that are known or suspected to be involved in active ion transport; but they are never found on other cellular membranes, such as the basolateral membranes of goblet cells or membranes of columnar or replacement cells (Harvey, 1980; Harvey et al., 1981). This unique arrangement, plus the independent evidence that active K + transport is indeed located in the apical membrane of goblet cells (Section IV,B), clearly identifies the particles with causation of active transport. The particles have therefore been called portasomes (Harvey, 1980); their similarities to the F,-F, particles and their role in active cation transport is discussed by Harvey et a f . (1981).
B. Pump Location Association of active K + transport with the apical membrane of one of the two major cell types (columnar, goblet) in the midgut epithelium can be inferred from elementary considerations. Associated structures, such as muscle cells, regenerative cells, the peritrophic membrane, and the basement membrane, either lack the topologically required continuity throughout the midgut or lack any semblance of metabolic coupling devices. The basolateral membrane of both columnar and goblet cells can be eliminated as the site of the primary K + pump, because then--in vivo-potassium would need t o diffuse passively from the cell interior into the midgut lumen FIG. 8. Fine structure of goblet cells in larval midgut from Munducu sexlu. AP, Fingerlike projections of the goblet cell apical membrane; BI, infoldings of the basal membrane; GC, goblet cell cavity; M, mitochondrion; MV, microvilli; NC, nucleus of a columnar cell; NG, nucleus of a goblet cell. Arrowheads point to particles (portasomes) which stud the cytoplasmic side of the membrane enclosing the goblet cell cavity. (a) Anterior region of the midgut, showing a goblet cell flanked by two columnar cells. The apical membrane of the goblet cell, invaginated almost to the base of the cell, forms a large, basally located cavity which opens into the midgut lumen through a long neck. Scale bar = 10 pm. (b) Posterior region of the midgut. The cavity of the goblet cell is confined to the apical portion of the cell, and the basal portion of the cell is reduced to a narrow stalk of cytoplasm. Scale bar = 10pm. (c) Enlargement of the basal region of a goblet cell from the anterior midgut to show the projections of the apical membrane into the cavity. Note that each projection contains an elongated mitochondrion. Infoldings of the basal membrane are also evident. Scale bar = 2 pm. (d) Detailed structure of the apical membrane of a goblet cell (anterior midgut), showing the particles which stud the cytoplasmic surface of that membrane. Scale bar = 0.2 pm. (e) Enlargement of the basal region of a goblet cell from the posterior midgut, to show that the apical membrane projections are short and devoid of mitochondria. Scale bar = 1 pm. (f) Detailed structure of a goblet cell apical membrane to show the particles studding that membrane, in the posterior region of the midgut. Note the lack of associated mitochondria. Scale bar = 0.2 pm. (g) Enlargement of the basal region of a goblet cell from the posterior midgut, to show the basal infoldings. These infoldings are less elaborate than in the anterior and midregions of the larval midgut. Scale bar = 1 pm. (From Cioffi, 1979, with permission of Longman Group, Ltd.)
126
MICHAEL
G. WOLFERSBERGER et a/.
against a fourfold concentration gradient (Table I) and against an electric potential difference in excess of 180 mV (Wood et al., 1969). Moreover, an electrogenic pump located at the basolateral membrane and moving ions in the basal-to-apical direction should make the cell interiors positive with respect to the basal solution, whereas microelectrode measurements have shown the internal potential t o be negative. A definitive demonstration that the apical membrane of goblet cells contains the K + pump has been obtained from microelectrode experiments. Wood et al. (1969) studied the profile of electric potential through isolated cecropia midgut by passing microelectrodes progressively through the entire membrane, from the basal to the apical side. Two steps were observed: one of approximately - 25 mV as the microelectrode penetrated the cell interiors, and one of about + 180 mV as the electrode entered the apical solution. The first step could be increased by diminishing the K + concentration of the basal solution, suggesting that potassium entry from the blood side is passive. At that time, the electrical profiles for all impalements were similar. In more recent experiments, with fine-tipped microelectrodes (Blankemeyer, 1976; Blankemeyer and Harvey, 1977, 1978), however, a second electric profile was identified (the initial step being - 5 mV) and assumed to be associated with one of the two main cell types, designated operationally as the low-potential difference (LPD) cell. The key observation which followed was that the total cellular resistance of the LPD cell, measured from the cell interior to both basal and apical solutions increases when the pump is slowed by anoxia. The apical membrane/basal membrane resistance ratio increases dramatically in nitrogen, which strongly suggests the existence of an electrogenic ion pump in the apical membrane of the LPD cell. The other cell type, designated high-potential difference (HPD) displays no change in membrane resistance during anoxia. A comparison of the frequency histograms for HPD and LPD profiles with counts of cell types in the epithelium strongly suggests that the HPD profiles belong to the more numerous columnar cells, while the LPD profiles (and therefore the pump) belong to the goblet cells.
-
C. Pool Location Measurement of ion transport pools by the analysis of flux kinetics has had limited success when applied to frog skin or toad bladder (Hoshiko and Ussing, 1960), mainly because their sodium pools are too small a part of the total tissue sodium. Now the lepidopteran larval midgut, with its large cells and high K + content, seems to be a more favorable material, but the initial experiments of Harvey and Zerahn (1969) and the later ones of
7. INSECT ELECTROGENIC CATION PUMP
127
Zerahn (1973, 1975) identified a brief, constant isotope mixing time, which indicates that only a small K + pool is accessible to the pump. This suggested that the potassium pool must be extracellular, located in special intracellular channels, or confined to the goblet cells. At the same time other experiments (Harvey and Wood, 1972, 1973; Wood and Harvey, 1975, 1979) provided convincing evidence that the mixing time is long and that the kinetic pool size is approximately equal to the total cellular potassium. Blankemeyer’s discovery (1976) of a second electrical potential profile (LPD) implied that midgut cells are not always coupled together, which suggests the possibility that there might (under different conditions) actually be two different pools accessible to the K + transport system. To investigate this point further, intercellular communication was examined in isolated midgut preparations, using the methods of Loewenstein (1966). Four principal experimental variants were tested: open-circuited and shortcircuited preparations, and midguts removed from leaf-reared and from ar-
a
> €
0-
cn cn
0-
a
TIME, min FIG.9. Effect of membrane short-circuiting upon the size of the potassium pool in H . cecropia taken from larvae fed on an artificial diet. Left: Open-circuit conditions; the upper trace represents the transmembrane potential difference plotted against time. Right: Shortcircuit conditions; the upper trace represents the SCC plotted against time. Bottom: The curves represent the BA flux calculated from isotope transfer and plotted against time. The arrow heads indicate the time at which K + label was introduced into the basal solution. The quasi-steady state fluxes were extrapolated to the time of isotope injection, and pool sizes were calculated from the enclosed area, as described by Wood and Harvey (1975). Note that the K + pool is more than threefold larger under short-circuit conditions than under open-circuit conditions. (From Blankemeyer and Harvey, 1977.)
128
(-)
MICHAEL G. WOLFERSBERGER
Blood Side
et 6’1.
Midgut
FIG. 10. Equivalent circuit diagram of the lepidopteran larval midgut, deduced by comparison of electrical and anatomical data. CC, Columnar cells, assumed to have HPD profiles; GC, goblet cells, assumed to have LPD profiles. The K + pump is located in the apical (cavity) membrane of the goblet cell. Under open-circuit conditions potassium transport occurs just through the goblet cell. Under short-circuit conditions, in midguts from diet-fed larvae, the columnar cells are coupled to the goblet cells so that isotopic potassium to be transported exchanges throughout the epithelium. (From Blankemeyer and Harvey, 1977.)
tificial diet-reared larvae. The results were unequivocal: Only in shortcircuited midguts isolated from diet-reared larvae does electrical coupling exist between columnar and goblet cells (coupling ratio = 0.6). It follows, therefore, that large pool sizes should be found in short-circuited preparations from diet-fed larvae but in none of the other variants. This prediction has been precisely verified (Blankemeyer and Harvey, 1977, 1978), and one demonstrative set of flux analyses is given in Fig. 9 for open-circuited and short-circuited midguts isolated from larvae fed on artificial diets. In the two lower parts, the rising curves represent isotopic fluxes of K + flowing from the basal solution, tracer-labeled at zero time, into the apical solution. The quasi-steady state fluxes have been extrapolated to zero time, and the relevant K + pools calculated from the area of the enclosed triangle.
7. INSECT ELECTROGENIC CATION PUMP
129
It is evident that the K + pool in the short-circuited midgut is more than threefold the size of that in the open-circuited midgut. (Unstirred layers contribute to the calculation of pool size by this method, but independent experiments have shown their contribution t o be practically negligible.) We can now combine the results of the electrical studies with these new tracer kinetic results t o formulate the model for transport routes shown in Fig. 10. In open-circuited midguts and those taken from leaf-fed larvae, the electrogenic active transport of potassium occurs just through the goblet cells, with their small pool size. But in short-circuited midguts from dietfed larvae the columnar cells become coupled to the goblet cells, thereby adding their large amount of potassium to the transport pool. While the earlier controversy about pool size (see three paragraphs above) was sustained partly by technical errors in some of the experiments, the main facts which resolve the controversy are (1) that the experiments of Harvey and Wood (1972, 1973; Wood and Harvey, 1975, 1979) were all done on short-circuited midguts from diet-fed larvae, but (2) the experiments of Harvey & Zerahn (1969) and subsequently of Zerahn (1973, 1975) were done either on midguts from leaf-fed larvae or on open-circuited midguts from diet-fed larvae.
V.
A POTASSIUM TRANSPORT ATPase
Given the facts that the lepidopteran midgut K + pump is located in the apical (cavity) membrane of a minority cell type and that this membrane is elaborated into projections packed with mitochondria, it is not surprising that early searches failed to find a clear K+-stimulated ATPase that might be identified with the active transport system (Harvey and Zerahn, 1972; Keynes, 1973). Turbeck et al. (1968) identified an anion-stimulated MgATPase from H . cecropia midgut with properties similar to those of an ATPase described in gastric mucosa by Kasbekar and Durban (1965), but concluded that it was probably of mitochondria1 origin. Other important information emerged, too, from these searches, namely, that the almost ubiquitous Na+, K+-ATPase apparently does not exist in insect larval midgut preparations (Turbeck el al., 1968; Jungreis and Vaughan, 1977), and-accordingly-that midgut does not bind the cardiac glycosides (e.g., ouabain) which are specific inhibitors of Na+-K+ pumping. More favorable prospects for obtaining the expected K+-stimulated ATPase emerged with Cioffi's (1979) discovery that, in the posterior region of M . sexta midgut, the K+-pumping membrane of the goblet cells is not interlaced with mitochondria (Section IV,A and B). By taking advantage of this situation and using only the posterior midgut as starting material,
130
MICHAEL G. WOLFERSBERGER
et a/.
Wolfersberger and Cioffi (Harvey et al., 1981) prepared a subcellular fraction containing very few mitochondria and composed mainly of large fragments of plasma membrane. Wolfersberger (1979) showed that, compared to the whole homogenate, this fraction is depleted of succinic dehydrogenase and enriched in Mg-ATPase activity. As expected, the MgATPase activity in this crude plasma membrane fraction, like that in the homogenate, is insensitive t o ouabain. It is also insensitive to mitochondrial ATPase inhibitors, such as oligomycin, and is clearly stimulated by potassium, as shown in Table 11. Perhaps the most interesting property of the K+-ATPase is the manner in which potassium stimulates its activity. As shown in Fig. 11, added potassium decreases by threefold the enzyme's K , for ATP but has little effect (20%) on the apparent maximal velocity for ATP hydrolysis. This response is similar to that expected for a K-class, heterotropic, allosteric enzyme in which potassium acts as a positive modulator (Monod et al., 1965). Such behavior of the enzyme itself could help explain why the rate of K + transport in midgut tissue has no influence on the rate of oxygen consumption (Section 111,C): By increasing the substrate affinity of the TABLE 11 ENZYME ACTIVITIES IN SUBCELLULAR FRACTIONS OF Enzyme
Homogenate
Mg-ATPase Plus 50 mM KCI Plus 0.2 mM ouabain Plus 5 pg/ml oligomycin Plus 10 mM cysteine Alkaline phosphatase Plus 10 mM cysteine Succinate dehydrogenase
3.21 f 0.22 3.29 f 0.30 3.17 f 0.27 2.08 f 0.25 3.12 f 0.29 14.08 f 0.35 1.41 f 0.42 5.10 f 0.46
MID GUT"^
Plasma Membrane
4.58 f 6.33 f 4.71 f 4.53 f 4.48 f 11.58 f 1.74 f 1.89 f
0.11 0.55' 0.38 0.31 0.43 0.34 0.35 0.19
Mitochondria
6.47 f 0.95 6.15 f 0.60 6.13 f 0.22 3.49 f 0.61 6.25 f 0.48 14.58 f 0.36 0.41 f 0.55 12.23 f 1.05
'Phosphatase activities are expressed in micromoles of Pi per milligram of protein per hour. Succinate dehydrogenase was assayed according to Ackrell el al. (1978);activity is expressed in micromoles of 2,6-dichlorophenolindophenol(DCIP) reduced per milligram of protein per minute. Alkaline phosphatase was assayed according t o Gordon (1952).The assay system for Mg-ATPase activity consisted of 3 mM ATP, 5 mM MgCI,, and 0.3 mg of enzyme protein in 1.0 ml of 40 mMTris-HCI buffer, p H 8.1. ATPase assays were started by the addition of enzyme; after 12 minutes at 25"C,they were stopped by the addition of 4 ml 50% isobutanol in benzene. Specific activities, corrected for endogenous and nonenzymatically formed P,, are means (at least three experiments) f 1 SEM. bFrom M. G. Wolfersberger (1979,unpublished results). 'The difference between Mg-ATPase plus KCI and Mg-ATPase is statistically significant ( p
-
< 0.01).
131
7. INSECT ELECTROGENIC CATION PUMP
0.4
--
1 V
0.2.-
0.5
I .o
1
AT P FIG. 11. Effect of elevated potassium concentration on plasma membrane-catalyzed ATP hydrolysis.., ATPase activity without added K'; W , ATPase activity in 70 mM K+ . Lineweaver-Burk plots. Ordinate scale, I/velocity (v measured in micromoles of Pi liberated per milligram of membrane protein per hour); abscissa scale, I/ATP millimolar concentration. Assay conditions similar to those of Table 11. (M. G. Wolfersberger, 1979, unpublished results.)
plasma membrane ATPase, elevated tissue potassium would allow transport to compete more effectively for the energy available from a nearly constant ATPpool. To take a specific example, Fig. 11 shows that, at the normal cellular ATP concentration of 1.2 mM, added potassium almost doubles the rate of ATP hydrolysis. If this enzyme is indeed involved in potassium transport, a similar change could be expected in vivo without any effect on ATP production or oxygen consumption. The converse response would be expected when tissue potassium drops.
ACKNOWLEDGMENT This work was supported in part by research grant A1 09503 from the National Institute of Allergy and Infectious Diseases.
132
MICHAEL G. WOLFERSBERGER
et a/.
REFERENCES Ackrell, B. A. C., Kearney, E. B., and Singer, T. P. (1978). In “Methods in Enzymology” (S. Fleisher and L. Packer, eds.), vol. 53, pp. 466-483. Academic Press, New York. Akai, H. (1969). J. Insect Physiol. 15, 1623-1628. Anderson, E., and Harvey, W. R. (1966). J. Cell Biol. 31, 107-134. Blankemeyer, J. T. (1976). Ph.D. Thesis, Temple University, Philadelphia. Blankemeyer, J. T. (1978). Biophys. J. 23, 313-318. Blankemeyer, J . T., and Harvey, W. R. (1977). I n “Water Relations in Membrane Transport in Plants and Animals” (A. Jungreis, T. Hodges, A. Kleinzeller, and S. Schultz, eds.), pp. 161-182. Academic Press, New York. Blankemeyer, J. T., and Harvey, W. R. (1978). J. Exp. Biol. 77, 1-13. Cioffi, M. (1979). Tissue Cell 11, 467-479. Cioffi, M. (1980). A m . Zool. 20, 939. Cioffi, M., and Harvey, W. R. (1981). J . Exp. Biol. 91, 103-116. Faust, R. M., Adams, J., and Heimpel, A. M. (1967). J . Invertebr. Pathol. 9, 488-499. Faust, R. M., Hallam, G. M., and Travers, R. S. (1974a). J. Invertebr. Pathol. 24, 365-373. Faust, R. M . , Travers, R. S., and Hallam, G. M. (1974b). J . Invertebr. Pathol. 23,259-261. Florkin, M . , and Jeuniaux, C. (1973). I n “The Physiology of Insecta” (M. Rockstein, ed.), 2nd Ed., Vol. V, pp. 256-307. Academic Press, New York. Giordana, B., and Sacchi, F. (1977). Comp. Biochem. Physiol. 56A, 95-99. Gordon, J. J. (1952). Biochem. J. 51, 97-103. Griego, V. M., Moffett, D. F., and Spence, K. D. (1979). J. Insect Physiol. 25, 283-288. Harvey, W. R. (1980). In “Insect Biology in the Future:VBW 80” (M. Locke and D. S. Smith, eds.), pp. 105-124. Academic Press, New York. Harvey, W. R. (1982). In “Membrane Physiology of Invertebrates” (R. Podesta and S. Timmers, eds). Dekker, New York (in press). Harvey, W . R., and Blankemeyer, J . T. (1975). In “Invertebrate Immunity” (K. Maramorosch and R. Shope, eds.), pp. 3-23. Academic Press, New York. Harvey, W. R., and Nedergaard, S. (1964). Proc. Null. Acad. Sci. U.S.A. 51, 757-765. Harvey, W. R., and Wolfersberger, M. G. (1979). J . Exp. Biol. 83, 293-304. Harvey, W. R., and Wood, J. L. (1972). In “Role of Membranes in Secretory Processes” (L. Bolis, R. Keynes, and W. Wilbrandt, eds.), pp. 310-331. North-Holland Publ., Amsterdam. Harvey, W. R., and Wood, J. L. (1973). In “Transport Mechanisms in Epithelia” (H. H . Ussing and N. A. Thorn, eds.), pp. 342-357. Academic Press, New York. Harvey, W. R., and Zerahn, K. (1969). J. Exp. Biol. 50, 297-306. Harvey, W. R., and Zerahn, K. (1972). In “Current Topics in Membranes and Transport” (F. Bronner and A. Kleinzeller, eds.), Vol. 3, pp. 367-410. Academic Press, New York. Harvey, W. R., Haskell, J. A., and Zerahn, D. (1967). J. Exp. Biol. 46, 235-248. Harvey, W. R., Wood, J. L., Quatrale, R. P., and Jungreis, A. M. (1975). J . Exp. Biol. 63, 321-330. Harvey, W. R., Cioffi, M., and Wolfersberger, M. G. (1981). Am. Zool. (in press). Haskell, J. A., Clemons, R. D., and Harvey, W. R. (1965). J. Cel/. Comp. Physiol. 65,45-55. Haskell, J. A., Harvey, W. R., and Clark, R. (1968). J. Exp. Biol. 48, 25-37. Hoshiko, T., and Ussing, H . H. (1960). A c f a Physiol. Scand. 49, 74-81. Jungreis, A. M., and Vaughan, G. L. (1977). J. Insecf Physiol. 23, 503-509. Kasbekar, D. K., and Durbin, R. P. (1965). Biochim. Biophys. Acta 105, 472-482. Keynes, R. D. (1969). Q.Rev. Biophys. 2, 177-281.
7. INSECT ELECTROGENIC CATION PUMP
133
Keynes, R. D. (1973). In “Transport Mechanisms in Epithelia” (H. Ussing and N. Thorn, eds.), pp. 505-51 1. Munksgaard, Copenhagen. Loewenstein, W. R. (1966). Ann. N. Y . Acud. Sci. 137, 441-472. Maddrell, S. H. P. (1971). In “Advances in Insect Physiology” (J. Beament, J. Treherne, and V. Wigglesworth, eds.), Vol. 8, pp. 199-331. Academic Press, New York. Mandel, L. J., Moffett, D. F., and Jabsis, F. F. (1975). Eiochim. Eiophys. Actu408, 123-134. Mandel, L. J. Moffett, D. F., Riddle, T. G., and Grafton, M. M. (1980a). A m . J. Physiol. 238, Cl-C9. Mandel, L. J., Riddle, T. G., and Storey, J. M. (1980b). A m . J. Physiol. 238, CIO-Cl4. Monod, J., Wyman, J., and Changeux, J.-P. (1965). J. Mol. B i d . 12, 88-118. Nedergaard, S., and Harvey, W. R. (1968). J. Exp. Eiol. 48, 13-24. Rehm, W. S. (1968). J. Theor. B i d . 20, 341-354. Schultz, T. W., and Jungreis, A. M. (1977). Tissue Cell 9, 255-272. Shappirio, D. G., and Williams, C. M. (1957). Proc. R. SOC. London B 147, 218-232. Smith, D. S., Compher, K., Janners, M., Lipton, C., and Wittle, L. (1969). J. Morphol. 127, 41-72. Stobbart, R. H., and Shaw, J. (1974). In “The Physiology of Insecta” (M. Rockstein, ed.), 2nd Ed. Vol. V, pp. 361-446. Academic Press, New York. Thomas, R. C. (1972). Physiol. Rev. 52, 563-594. Travers, R. S., Faust, R. M., and Reichelderfer, C. F. (1976). J. Invertebr. Puthol. 28, 351-356. Turbeck, B., Nedergaard, S., and Kruse, H. (1968). Eiochim. Eiophys. Acta 163, 354-361. Ussing, H. H., and Zerahn, K. (1951). Acta Physiol. Scund. 23, 110-127. Wigglesworth, V. B. (1972). “The Principles of Insect Physiology,” 7th Ed. Chapman & Hall, London. Wilson, D., Stubbs, M., Oshino, N., and Erecinska, M. (1974). Biochemistry 13, 5305-5311. Wolfersberger, M. G. (1979). Fed. Proc. Fed. A m . SOC. Exp. Eiol. 38, 242. Wolfersberger, M. G., and Giangiacomo, K. M. (1980). A m . Zool. 20, 938. Wood, J. L. (1972). Ph.D. Thesis, Cambridge University. Wood, J. L., and Harvey, W. R. (1975). J. Exp. Eiol. 63, 301-311. Wood, J. L., and Harvey, W. R. (1976). J. Exp. B i d . 65, 347-360. Wood, J. L., and Harvey, W. R. (1979). J. Exp. Eiol. 82, 1-9. Wood, J. L., and Moreton, R. B. (1978). J. Ekp. B i d . 77, 123-140. Wood, J. L., Farrand, P. S., and Harvey, W. R. (1969). J. Exp. B i d . 50, 169-178. Wood, J. L., Jungreis, A. M., and Harvey, W. R. (1975). J. Exp. Biol. 63, 313-320. Zerahn, K. (1956). Actu Physiol. Scund. 36, 300-318. Zerahn, K. (1973). In “Transport Mechanisms in Epithelia” (H. H. Ussing and N. A. Thorn, eds.), pp. 360-367. Academic Press, New York. Zerahn, K. (1975). J. Exp. B i d . 63, 295-300. Zerahn, K. (1977). In “Transport of Ions and Water in Animals” (B. Gupta, R. Moreton, J. Oschman, and B. Wall, eds.), pp. 381-401. Academic Press, New York. Zerahn, K. (1978). In “Membrane Transport in Biology” (G. Giebisch, D. Tosteson, and H. Ussing, eds.), pp. 273-306. Springer-Verlag, Berlin and New York.
This Page Intentionally Left Blank
CURRENT TOPICS IN MEMBRANES AND TRANSPORT, VOLUME 16
Chapter 8 The ATP-Dependent Component of Gastric Acid Secretion G . SACHS, B. WALLMARK, G . SACCOMANI, E. RABON, H . B. STEWART, D . R . DiBONA', T. BERGLINDH Laboratory of Membrane Biology University of Alabama Birmingham, Alabama
I. Introduction ............................................................................. . 136 ................................ 136 11. Site of Acid Secretion ..................... ................................ 140 111. Energy Source for Acid Secretion ..... ................................ 142 IV. Location of the K+-Dependent ATPas 144 V. Nature of the ATPase ........................................................................... ...... ............................... 145 VI. Steps in ATP Hydrolysis ....... 145 A. Formation of Phosphoenzyme ................................................... 145 ................................. B. Breakdown of Phosphoenzyme ......... 146 C. Steady State Kinetic Aspects .......................................... ............................ 148 VII. H + Transport by Gastric ATPase ......... ............................ 150 VIII. K + Transport by Gastric ATPase .......... 150 A. Active Cation Transport ................................................................. 150 B. Passive Cation Transport 151 C. Effect of External Cations 153 IX . Electrogenicity of the Pump ................................................................... ........................ 154 X. pH Gradient and Stoichiometry . . 156 x1. Structural Aspects of the ATPase ........................................ 157 XII. Summary and Conclusions .................................................................... References ................................... ....................... 158
'
Nephrology Research and Training Center, University of Alabama, Birmingham, Alabama.
135
Copyrighl 0 1982 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12- 1533 16-6
136
G . SACHS
1.
et a/.
INTRODUCTION
The mammalian gastric mucosa secretes HC1 at an effective final concentration of 160 mM. This represents a pH gradient of at least 6.6 units and therefore the largest gradient in eukaryotic cells. It was also the first transport process to be established as being clearly energy-dependent and therefore active. Accordingly theories as to its mechanism antedate the discovery of ATP, and suggestions in the 1920s and 1940s (Lund, 1928) involved the separation of protons and electrons by vectorially arranged redox pumps (Conway, 1950). Such mechanisms, now prevalent in mitochondrial physiology (Mitchell, 1966), were electrogenic, and the gastric epithelium was the primary example of an electrogenic epithelium for many years (Rehm, 1972). Perhaps the classic experiment was the demonstration that in C1--free solutions the current generated by frog tissue was a linear function of the acid rate (Rehm and LeFerre, 1965). It seems appropriate therefore to include a discussion of the gastric proton pump in this symposium on electrogenic pumping. As will be seen, however, as far as we understand today, the isolated ATPase from gastric mucosa is an H+pump which does not conform to the expected electrogenic type. The nature of the complete gastric H + pump has not as yet been entirely determined, and in order to demonstrate the steps leading to our current view, we have to deal with various models, passing from intact tissue to isolated subcellular organelles. One of the key observations in gastric physiology could be made only on isolated tissue, namely, that acid secretion is K +-dependent (Harris et al., 1958). To explore this further, however, it has proved necessary t o develop subtissue models of acid secretion and to define the process at the level of isolated cells or gastric glands. First one must establish the type of cell and the structure within that cell which is responsible for acid secretion. We will see that the technologies developed t o solve this problem will then allow us to establish the probable primary energy source for the H + pump. From this we shall demonstrate that the K+-dependent ATPase of the tissue is appropriately localized in the acid-secreting structure. Its catalytic cycle allows us to predict some of the transport properties of this ATPase and then, studying the transport properties of the isolated enzyme, we shall show that the pump is electroneutral and define the role for K + stated at the beginning of this paragraph. II. SITE OF ACID SECRETION The gastric epithelium is a complex, folded tissue composed of a single layer of heterogeneous cells. Acid secretion is produced by the major infoldings of the epithelium, called gastric glands. The gastric glands are composed of two major cell types: chief or zymogen-containing cells, and
8.
ATP-DEPENDENT COMPONENT
OF GASTRIC ACID SECRETION
137
parietal cells. It is the latter cells that-because of location, mitochondria1 content, and morphological changes-have long been considered responsible for the elaboration of HCl. Until recently, however, direct proof was lacking and little insight had been gained into the meaning of the morphological transformation that occurs. Several lines of evidence now can be cited for the parietal cell’s role as the acid secretor. When rabbit gastric glands are isolated by collagenase digestion, they retain the ability to respond to gastric stimulation by compounds such as histamine or CAMP, among others (Berglindh and Obrink, 1976). When viewed in the living state by differential interference-contrast microscopy, the unstimulated cells show only the structural features of the mitochondria-rich parietal cells and the granule-loaded peptic cells. Upon stimulation the parietal cell develops characteristic vacuoles, presumably loaded with acid. These vacuoles are shown in Fig. 1. It is possible to prove that these structures are indeed acid-containing, by two further types of experiments. It has been shown that the weak base aminopyrine is accumulated by gastric glands even in the absence of added stimulus (Berglindh and Obrink, 1976). Indeed, the use of this weak base in the radioactive form provided the major breakthrough in the study of isolated gastric cells or glands. It has also been shown that, even in the absence of a stimulus, the elevation of medium K+ to levels as high as 100 mMresults in a significant increase in aminopyrine accumulation (Berglindh, 1978). Since the pK of this base is 5 , this implies the presence of a compartment somewhere in the gastric glands or cells with a pH much lower than 5 . Now if this in turn indicates the presence of H + pumping, then the addition of high concentrations of weak base should result in accumulation, in buffering of the acid compartment, and in hypertonicity of the same compartment. Hence one would expect t o see the entry of water and swelling of the compartment. The experiment is then to add, in the presence of normal or high K + , concentrations of aminopyrine in the range 0.1-1 mM. In the case of normal K + this results in swelling of the area of the cells continuing to elaborate acid in spite of the absence of stimulus, and in the case of high K + results in expansion of the compartments now elaborating acid at an increased gradient. Figure 2 illustrates that, when this is done, vacuolation (induced by histamine in Fig. 1) is induced in the absence of histamine, at least in some cells, when the medium K + is maintained at physiological levels. In high-K+ medium the transformation is induced in virtually every parietal cell. This allows us to conclude that the vacuolar structures induced by histamine do indeed contain acid. To understand the electron microscopic correlate of the vacuoles observed by light microscopy, it is necessary to discuss some general aspects of the ultrastructural morphology of the parietal cell (Sedar, 1965; Helander, 1962). The unstimulated cell contains an elaborate system of smooth-surfaced structures called tubulovesicles and an infolding of the
FIG. 1. These two micrographs, using Nomarski optics, illustrate the morphological transformation of living rabbit gastric glands upon histamine stimulation. (a) A nonstimulated gland; (b) vacuoles (V) formed upon histamine treatment. Parietal cells are labeled P; chief cells are labeled Z .
138
FIG.2. An illustration of the effect of Maminopyrine on gastric glands incubated in normal-K+ or high-K+ medium but in the absence of secretagogue. The initial appearance corresponds to that shown in Fig. l a . (a) The effect of normal K f solutions, with rare vacuoles (here only in the bottommost cell); (b) the effect of high-K+ medium, with vacuoles present in all the parietal cells of this particular gland. 139
140
G. SACHS
et a/.
apical plasma membrane known as the secretory canaliculus. This “resting” appearance is maintained even in the presence of high K + , although there is an increased pH gradient as discussed above. Upon stimulation, the tubulovesicles disappear and are replaced by microvilli lining a much expanded canaliculus. This expanded canaliculus obviously represents the vacuoles discussed above and visualized by Nomarski optics. Thus we can conclude that the site of acid secretion is into the lumen of the secretory canaliculus. It is the limited access of bathing medium into this lumen that allows the cells t o retain the polar property of acid accumulation. The formation of vacuoles, by the addition of aminopyrine to what morphologically are resting cells, has one further implication: The basis for the normal morphological transformation may be the initiation of secretion, hence swelling of the compartment into which secretion is occurring. One could then regard the tubulovesicles of the resting parietal cell as being collapsed microvilli. An alternative view is that these are in fact vesicles in the resting state and that fusion between vesicles occurs along with arrangement of microfilaments to produce the functional lining of the secretory canaliculus (Forte et al., 1977). The osmotic expansion concept is obviously the simpler of the two. 111.
ENERGY SOURCE FOR ACID SECRETION
Yet another way of demonstrating the site of acid secretion is to use the dye probe acridine orange. This weak base is accumulated across membranes according to the pH gradient; then, depending on its concentration and the availability of negative binding sites, it aggregates and shifts its fluorescence emission peak from 530 toward 660 nm (Dell’Antone et al., 1972; Dibona et af., 1979). This single dye therefore might be expected to mark regions of normal pH with green fluorescence and regions of low p H with red fluorescence. Figure 3a shows a stimulated parietal cell, viewed by combined Nomarski and fluorescence optics. The red-staining intracellular secretory canaliculus is easily visible, which argues strongly that the site of acid secretion in the mammalian gastric mucosa must be the secretory canaliculus of the parietal cell. The ability to monitor acid secretion microscopically in this unique cell FIG.3. The use of acridine orange in defining both the site of acid secretion and the probable energy source. (a) A stimulated gastric gland is shown, with the acid-secreting areas delineated as red, inside the green fluorescent cytoplasm; (b) a similar gland is shown after exposure t o CN- and high-voltage shock treatment. A lack of red fluorescence in the (still distended) vacuoles is evident; (c) the same group of glands-following high-voltage shocking and CN--but 5 minutes after addition of 5 m M Mg-ATP. The restoration of the red areas characteristic of acid secretion can be discerned; (d) the same field but with fluorescence only.
This Page Intentionally Left Blank
8.
ATP-DEPENDENTCOMPONENT
141
OF GASTRIC ACID SECRETION
led t o the use of permeable gastric glands t o determine the energy source, or at least a partial energy source for H transport. High-voltage shocking of gastric cells, generally in the presence of CaZ+, results in fenestration of their limiting membranes, just as has also been demonstrated in the adrenal medulla (Baker and Knight, 1978) and elsewhere. The application of four brief 3-kV shocks to gastric glands suspended in normal or high-K+ medium results in impaired ability t o accumulate either acridine orange or aminopyrine. When cell ionic content is measured following this protocol, it closely resembles the medium ionic composition, and about 80% of the gastric glands take up trypan blue, a dye marker with a molecular weight of 961. These cells could therefore be expected to be permeable to ATP. If M CN- ion is added and the cells are then shocked, the appearance in Fig. 3b is produced. The vacuoles are still evident as dark regions, indicating a low acridine orange concentration. In the presence of high K +, the addition of ATP restores the red acridine orange fluorescence in some of the cells, as shown in Fig. 3c. Thus it seems, at least under these special conditions (i.e., permeable cells, mitochondria1 inhibition, and high cytoplasmic K + ) , that ATP can act as the energy source for acid secretion. The acridine orange technique has been supplemented by examining the uptake of [I4C]aminopyrine, a technique which allows determination of the general ability of the gastric parietal cells to secrete acid even after the shock procedure has completely destroyed the ability of some cells to secrete. Figure 4 shows the time course of aminopyrine uptake (expressed +
40-
0 0 0
0
2
4
6 8 time (min)
10
FIG.4. The time course of aminopyrine accumulation in shocked gastric glands after either CN- or amytal treatment. Lower curves: In the absence of added ATP; upper curves: following the addition of ATP. As for Fig. 3c, these glands have been suspended in high-K+ medium in order to demonstrate the ATP effect.
G. SACHS et a/.
142
as the ratio of aminopyrine accumulated in the gland water t o the aminopyrine concentration in the medium), following ATP addition to glands shocked in CN-- or amytal-containing solutions. The accumulation ratio of about 40 can be compared to the control, unshocked, and uninhibited ratio of about 100, showing that some damage has indeed occurred. However, the difference in pH gradient is relatively small, since the pH achieved is a log function of the ratio of base accumulated. From these results we must conclude at least that ATP can serve as an energy source for H + accumulation in the gastric glands. Other energy sources could be present, but these would be independent of mitochondria1 respiration, so that only redox systems in the plasma membrane would appear as viable additional candidates for energization of the gastric H + pump. The ability to probe the secretory canaliculus membrane in these permeable gastric cells opens new avenues to determination of the properties of the gastric proton pump in situ. And one of the more intriguing results to be obtained concerns the effect of thiocyanate (SCN-), an ion found to inhibit acid secretion by gastric glands from every species (including rabbit) studied thus far. As would be expected from the above discussion, SCN- inhibits ATP-dependent accumulation of both acridine orange and aminopyrine (Berglindh et al., 1980). Yet, as will be detailed below, in membrane vesicles isolated from hog stomach or rabbit gastric glands, SCN- is not an inhibitor of the (potassium-requiring) H + secretion which occurs upon the addition of ATP. Our working hypothesis t o account for this striking observation is that the action of SCN- ions requires a close association between gastric ATPase and the anion channel of carbonic anhydrase.
IV.
LOCATION OF THE K+-DEPENDENT ATPase
A significant finding in gastric parietal cell biology was the presence of a K -dependent ATPase in gastric homogenates and membrane fractions (Ganser and Forte, 1973). This enzyme was later purified until essentially only a single band was observed on sodium dodecyl sulfate (SDS)-mercaptoethanol gel electrophoresis (Saccomani et al., 1977). Now for this enzyme to play a role in gastric secretion of HC1, it must be located at the acidsecreting site. In order to identify the ATPase-bearing membrane, monospecific antibody against purified ATPase was prepared (Saccomani et al., 1979a). By standard immunofluorescence methods only the parietal cells were found to bind significant antibody; and the peroxidase assay revealed an intense reaction only with the microvilli of the secretory canaliculus (Fig. 5 ) . Thus the K+-dependent ATPase is indeed located appropriately at +
FIG.5 . The staining of a lumen of gastric gland following treatment of the tissue section with monospecific anti-ATPase antibody and the coupled peroxidase procedure. The lumen (L) contains heavily stained microvilli on the apical surface of the parietal cell (horizontal arrow) and the lightly stained apical membrane of the chief cell (vertical arrow). (Thanks are due to Dr. H. F. Helander for this micrograph.)
G. SACHS et a/.
144
the secretory surface of the parietal cell. (Weak staining of the apical surface of the peptic cells was seen with peroxidase, so that the luminal surface of this cell may also have some secretory capacity. Because the luminal surfaces of both peptic cells and parietal cells are exposed to an extraordinary concentration of H + ions, the structure of these membranes must in some way be specialized to resist acid disruption. Phospholipid composition does not appear remarkable, and therefore the carbohydrate component of these membranes seems likely to be specialized for acid resistance.)
V.
NATURE OF THE ATPase
A variety of ATPases are found in nature. We can classify these into (1) types, such as mitochondria1 F, ATPase, that do not form a phosphorylated intermediate and seem to be membrane extrinsic proteins transporting protons only in association with a membrane intrinsic peptide, Fo (Kagawa and Racker, 1971); and (2) types, such as Ca2+-and Na+,K+ATPases, that form phosphorylated intermediates. This latter class of enzymes contains at least one subunit of molecular weight 100,000 and forms an acyl phosphate in the presence of Mg2+-ATPplus the appropriate cation (Skou, 1965; Hasselbach, 1978). The gastric K+-dependent ATPase is in this second category, since it contains a peptide of 100,000 molecular weight (Saccomani et al., 1977) and is phosphorylated by ATP. Like Na+,K+-ATPase,it is also discharged of phosphate by K+ and other activating cations (Saccomani et al., 1975). The classification of ATPases into these two types, based on the formation (or not) of a phosphorylated intermediate, can be extended to a consideration of mechanism. There are two evident requirements for a pump translocating ions: a change in the sidedness of binding sites and a change in the affinity of the sites. The sidedness condition can be met either by a translocating carrier portion of the protein or by a channel capable of closed-open, closed-closed, opened-closed configurations. This feature of pump structure remains mysterious. The change in affinity limits the gradient which can be generated and is itself restricted by the stoichiometry of coupling between the ion and ATP hydrolysis. Stoichiometry may be either fixed or variable, which has profound implications for reversibility. In addition, the mechanism may be a uniport, symport, or antiport in terms of the ions required for translocation. Defining the mechanism of proton translocation presents special problems, because proton jumping can probably occur along hydrogen-bonded chains either in organic structures or in water itself (Morowitz, 1978). Moreover, in contrast to other ions, protons cannot be added to or removed from aqueous solutions without affecting multiple parameters. The
-
8.
ATP-DEPENDENTCOMPONENT OF GASTRIC
ACID SECRETION
145
determination of stoichiometries also becomes more difficult, since protons can be released or bound as a function of experimental conditions and yet still be regarded as actually translocated. Subsequent sections of this article will briefly treat some of these problems, as illuminated by different techniques applied to gastric ATPase. VI.
STEPS IN ATP HYDROLYSIS
To study these reactions, at room temperature and above, a flow-quench machine is necessary (Mirdh and Post, 1977). The main questions to be answered are: What reactions steps are involved? What are their individual rates? How are they influenced by specific cations?
A. Formation of Phosphoenzyme When the enzyme is first mixed with [ Y ~ ~ - P ] A Tand P then Mg2+ is added, the pseudo-first-order rate constant for phosphoenzyme formation is 4400 min-I. On the other hand if the free enzyme is mixed with Mg2+ ATP, the rate constant for phosphoenzyme formation is 1400 min-I. Finally, if the enzyme-radioactive ATP mixture is added to Mg2+ plus excess cold ATP, the steady level of radiolabeled phosphoenzyme is about 400 pmoles/mg protein, compared with 1200 pmoles/mg in the previous two protocols (Wallmark and Mirdh, 1979;Wallmark el a/., 1980). From these data the reaction sequence must be Enz
+ ATP = E n z - A T P e Enz-P
No assignment of the status of ADP is possible from such experiments. Although an ATP-ADP exchange activity is present in the enzyme, no discharge of phosphoenzyme is observed upon the addition of ADP under a variety of conditions, a circumstance which contrasts with that for Na+, K + - or Ca2+-ATPase.Under the conditions of study ( 5 pM ATP, pH 7.4, broken membranes with highly purified ATPase), the overall rate constant corresponds to approximately 210 min-I, so that the formation of phosphoenzyme is certainly rapid enough to allow phosphoenzyme to be an intermediate. Under these K +-free conditions the formation of phosphoenzyme is not rate-limiting.
B. Breakdown of Phosphoenzyme When phosphoenzyme formation is carried to a steady state under the above conditions (about 130 mseconds), and trans-l,2-diaminocyclohexanetetraacetic acid (CDTA) is then added to prevent further phosphoryla-
G. SACHS e t a l .
146
tion, the phosphoenzyme decays with a rate equivalent to about 10 min-l. This corresponds to the overall reaction rate in the absence of K + . When K + is added along with CDTA, the breakdown of phosphoenzyme occurs in two stages. The fast stage increases in rate and amplitude with the K+ concentration (up to a maximum at 100 mM K + ) , accounts for about 50% of the dephosphorylation, and has a rate constant of 4000 min-l at the optimal K + level. The slower phase has a rate constant of 210 min-I and is independent of K + above 0.5 mM. These data allow the following reaction sequence to be postulated: Enz-PI
- Enz-P,, K' Enz + Pi
The initial fast K +-activated step is not rate-limiting and allows phosphoenzyme t o be an intermediate in the reaction. The rate-limiting step appears to be the slow phase of phosphoenzyme breakdown, which is activated by low K + concentrations but is not sensitive to further increases in K + concentration. It should be noted that no evidence has been obtained for any inhibitory action of K + in this step of the reaction. C. Steady State Kinetic Aspects Steady state kinetic analysis of Mg2+- or Mg2+,K+-ATPaseas a function of ATP concentration shows the presence of two distinct apparent K,,, values for ATP. In the presence of Mgz+ alone, the values found are 0.4 and 50 pM, and in the presence of K + , 3.5 and 30 pM. When the activating effect of K + is explored by varying the K + concentration, the data of Fig. 6 are obtained. Clearly, the action of K + is biphasic. There is an initial activating component with a K E of 200 pM that is independent of ATP concentration. There is also an inhibitory effect with a K , that depends on the ATP concentration but lies in the range of 5-15 mM. To investigate the inhibition, experiments can be carried out on tight, intact vesicles having a K + permeability which is low enough so that the sidedness of the activating and inhibiting effects can be determined. The experiments using the flow-quench technique, as described in Section IV,A and B, were carried out on lyophilized preparations, which display no stimulation of the ATPase upon addition of K +-selective ionophores. When unbroken membranes are used, so that K + stimulation of the ATPase is dependent on the presence of ionophores, experiments can be carried out to examine the effects of K + present only on the outside (ATP) surface of the enzyme. The rate of phosphorylation is normal in these vesicles when enzyme is mixed with Mg2+ and ATP, but allowing phosphorylation to reach a steady state level and then pulsing with CDTA plus K + gives only very slow dephosphorylation. This result suggests that the K + site responsible for initiating dephosphorylation is present only on
8.
ATP-DEPENDENT COMPONENT
147
OF GASTRIC ACID SECRETION
' */
0.05mM ATP
0.005 mM ATP
\\
\
%
FIG. 6 . Steady state hydrolysis of ATP by gastric ATPase at varying K + concentrations and levels of ATP. The biphasic nature of the interaction of K C with the ATPase is shown. The time of incubation has been chosen to provide less than 10% hydrolysis of substrate.
the inner surface of the vesicles. Moreover, since the addition of K + under these conditions does not result in dephosphorylation, one can now investigate the possible inhibitory component of K + interaction with the ATPase. Accordingly, if the enzyme is mixed with Mg2+, ATP, and K + simultaneously, there is significant inhibition of the rate of phosphorylation as the K + concentration increases. Evidently, binding of K + to the outside surface (i.e., the ATP-binding side) of the membranes results in the inhibition of phosphoenzyme formation. At sufficiently high K +/ATP ratios, the formation of phosphoenzyme becomes rate-limiting for the overall ATPase reaction. The experiment is illustrated in Fig. 7. We can conclude that there are two K+-binding sites on the enzyme: one, internal, of high affinity, that activates dephosphorylation; and one, external, of low affinity, that slows phosphorylation. The reaction scheme is then: EP-Kil- EP-Kixl:E-P-Kixt- E-K:x,+
Pi:E-K&, + ATP
- E-ATP + K+
Thus K + reacts on both surfaces of the enzyme and reacts with different affinities at the two surfaces. This fulfills the condition mentioned earlier for a transport reaction: change in sidedness and change in affinity. It also suggests, for everted gastric vesicles which are oriented with the A TP site outside, that internal K + should be required for demonstrating H + transport and-in addition-that K should be translocated during transport. +
148
G. SACHS
ef a/.
c N
20
40
60
80
100
120
time (msec)
FIG. 7. A rapid kinetic analysis of the formation and breakdown of the phosphoenzyme intermediate of gastric ATPase. Experiments have been carried out in relatively ionimpermeable vesicles isolated from hog gastric mucosa. The upper curves show the effect of K f addition subsequent t o the formation of phosphoenzyme in these tight vesicles. The lack of rapid dephosphorylation is obvious (0,choline; A, 20 mM KCI). The curve with the solid circles, showing increasing levels of phosphoenzyme, is the rate of phosphorylation in the absence of any K t . The lower two curves show the effect of 10 and 25 m M KCI on the outside of the vesicles. The inhibitory action of K + binding to the external surface of the vesicles is evident in terms of the phosphorylation reaction, thus accounting for the inhibitory component of K + action illustrated in Fig. 6 .
VII.
H + TRANSPORT BY G.ASTRIC ATPase
The addition of ATP to dog (Lee et a/., 1974), hog (Sachs et al., 1976), frog (Rabon e t a / . , 1979), rabbit (Berglindh et al., 1979), and human (Saccomani et a/., 1979b) gastric membrane vesicles results in an uptake of H + into an intravesicular compartment, provided K + is present. The requirement for K + is in fact intravesicular, as demonstrated in Fig. 8. In this experiment, gastric vesicles were preincubated various lengths of time in KCl solutions at 4"C, warmed to room temperature in a stirred cuvet containing a pH electrode, and then given ATP. Judging from the increased initial rate and magnitude of proton uptake with increased preincubation time, it seems safe to conclude that proton transport depends on the intravesicular K + concentration. The characteristic overshoot can be explained
8.
ATP-DEPENDENT COMPONENT
149
OF GASTRIC ACID SECRETION
too .-c 0)
c
2 80 a 0
E
\
-
2 60 0 a 3
-k
I 40 u)
OI -
0
E
c
20
30
60 90 Time (sec)
120
150
FIG.8. The effect on Ht transport of a progressive increase in internal K + . H + movement was determined with an extravesicular pH electrode, following addition of ATP to the suspension. Vesicles were incubated at 4°C in 150 rnM KCI for the time noted and ATP was added at zero time. The effect of increasing GK is shown with valinomycin addition at steady state, and the effect of increasing GH is shown by the addition of TCS at steady state. The enhancing effect of increasing Kgternalis evident, along with effects of the PK+/ P Ht ratio.
as follows. During the rise of the pH gradient, internal K + exchanges for external H +, but subsequently HCl leaks from the vesicles, generating the downward phase along with a loss of osmotically active solute and shrinkage of the vesicles. At lower internal K + concentrations accumulated protons probably do not exceed the buffer capacity of the vesicle interior, and little or no leakage is observed. The steady state pH gradient which can be maintained is determined by the relative leakage for K + (inward) and the leakage for H + (outward). The addition of valinomycin, which enhances K + permeability without affecting Hi- conductance in gastric mucosa, releases the limitations on H + uptake normally imposed by internal K + and a low permeability t o potassium. Valinomycin enhances both the initial rate of H + uptake induced by ATP (in vesicles that have not equilibrated K + during the preincubation period) and the steady state pH gradient in equilibrated vesicles. Finally, it is clear that the effect of valinomycin is related directly t o potassium, rather than t o a nonspecific change in membrane conductance, since lipid-permeable cations such as triphenylmethylphosphonium do not substitute for valinomycin.
G. SACHS ef a/.
150
From these very simple experiments, it is tempting to suggest that gastric ATPase acts as a K+-for-H+ antiport pump (Sachs et al., 1976).
VIII.
K + TRANSPORT BY GASTRIC ATPase
Two aspects of this process are of special significance: the predicted ATP-dependent K + efflux and the entry (or reentry) of KCL, which is essential for H transport and ATPase activity. +
A. Active Cation Transport When gastric vesicles are preequilibrated with K + and ATP then added, the ATPase activity retains valinomycin sensitivity. This suggests that K+ which is internal before ATP addition effluxes following substrate addition. The expected K + movements can be confirmed directly by preequilibrating the vesicles in radioactive solutions containing (for example) s6Rb or 204T1,then adding ATP, and measuring intravesicular isotope. ATP does indeed induce an efflux of isotope, and since this occurs against a concentration gradient, it must be primary or secondary active transport. While secondary active transport of potassium could result from a membrane potential generated by the H + pump process, the addition of lipidpermeable anions such as SCN- should shunt this potential and prevent cation efflux. Since SCN- does not have such an inhibitory effect, cation efflux must result from direct coupling to the pump mechanism. This insensitivity to membrane potentials generated or imposed during the transport process (Schackmann et al., 1977) further strengthens the conclusion that gastric ATPase functions as a K+-H+ antiport.
B. Passive Cation Transport Both from the above results and from data on intact gastric glands, it is clear that the gastric H + pump requires relatively high K + concentrations on the luminal (intravesicular) surface, and that this K + is recycled during pump activity. Hence a pathway must be provided for K + entry. This can be studied using isotopic uptake experiments or by examining the osmotic response of gastric vesicles to imposed osmotic gradients of the cation salts in question. When high concentrations of KCl are added t o vesicle suspensions, rapid shrinkage of the vesicles occurs, with water loss. From the acti-
8.
ATP-DEPENDENT COMPONENT
OF GASTRIC ACID SECRETION
151
vation energy for this process, we can conclude that the water diffuses across hydrophobic regions and that water-filled channels do not function in the vesicle membranes. This is appropriate for a membrane of a low H + permeability. Subsequent reswelling, due to KCI entry, has a half-time of about 50 minutes at room temperature, which implies a rather low permeability (about cm/sec) (Habon et al., 1980) to the salt. This KCl permeability seems t o be the rate-limiting step for ATP hydrolysis in intact gastric vesicles. Calculations of H + secretory rates (per unit membrane area) in the intact animal show that the KCl permeability is at least one order of magnitude too low to enable the gastric ATPase to function as the sole source of H + without the addition of a KCl permeation pathway. Thus far only one report concerning possible mechanisms of enhanced KCl permeability has been published: a claim that Ca2+ added to very crude gastric vesicle preparations enhances KCl uptake (Michelangeli and Proverbio, 1978). With the vesicles used in our laboratory, we have not been able to confirm the Ca2+findings for either hog or rabbit gastric mucosa. C. Effect of External Cations on H + Transport
From the kinetic data it appears that, although internal cation is essential for pump cycling and for H + uptake, external cation should inhibit both ATPase activity and H + rate. This can be established by equilibrating the vesicles at a given KCl concentration (for example 150 mM) and diluting them into solutions of choline or Na+ chloride, neither of which supports H transport. When this is done using acridine orange as a probe of internal pH, the data of Fig. 9 are produced. It can be seen that, as the external K + is progressively lowered by chofine substitution, the same magnitude of pH gradient is reached, but it is reached with a progressively faster time course. On the other hand, if Na+ is the substituting cation, the rate is progressively reduced, consistent with a K +-like inhibitory action of Na+ in the phosphorylation step of the enzyme reaction. The result further shows that the dilution effect is not simply due to the formation of an outward K + gradient. Transport and kinetic results are therefore internally consistent and allow us to postulate the general model of gastric H + , K + ATPase, and its transport capabilities, shown in Fig. 10. The pump protein is portrayed for convenience as a circle, but this is not intended to have any mechanistic significance. In the presence of K + the reaction is initiated by the displacement of external K + by ATP. This is followed by Mg2+-dependent protein phosphorylation to give the EP, form, which results in the translocation of H + to the vesicle interior. Internal K + is provided by the entry of KCl across a pathway which is passive in these vesicles +
Cho 140 d-Ko 150
O J
1
10
20
30
40
50
60
70
80
90
time (sec) FIG. 9. The effect of varying K t distribution across the membranes of gastric vesicles on the rate and magnitude of proton transport. Gastric vesicles were preequilibrated in 150 mM KCI and diluted into media with choline or Na' partially substituted for K' . Curve d shows the rate of H + transport following ATP addition (determined from quenching of acridine orange fluorescence), with K t concentrations equal on both sides of the vesicle membranes. The curve moving to the left shows the effect of fixed internal K i , and reduced external K + , namely, a more rapid H i uptake, but to the same final level. The curves to the right show inhibitory effects of external Na' as the concentration of this cation is increased.
FIG. 10. A diagram of the reaction steps of the ATPase and their relationships to H i and K + transport, showing the necessity for K ' entry into the vesicle lumen.
8.
ATP-DEPENDENT COMPONENT
OF GASTRIC ACID SECRETION
153
but may be more complicated in the intact cell; and binding of K + displaces the H + ,converting the EP, to the EP,, form. The outward movement of K + is accompanied by the release of Pi and formation of the E-K:x, complex.
IX.
ELECTROGENICITY OF THE PUMP
Although it is obvious that gastric H + ,K+-ATPase cannot be classified as a uniport, the fact that it is an antiport does not necessarily exclude electrogenic operation. Thus, if the K + were recycled internal to the membrane or if the H + / K + stoichiometry were not unity, potentials could appear during operation of the enzyme. Measurement of potentials generated under transporting conditions appears to be the most straightforward way of determining the electrical correlates of transport. For measurement of vesicular membrane potentials inside negative, one would use a lipidpermeable cation with a radioactive or fluorescent tag, such as tetraphenylphosphonium (in practice, combined with small quantities of lipidpermeable anion) or diethyloxocarbodicyanine,respectively. But neither of these is measureably accumulated during H + transport. To exclude that the technique is too insensitive to measure the possible membrane potential, a protonophore such as 3,5,3 ',4'-tetrachlorosalicylanilide (TCS) can be added. This should short-circuit any electrogenic pump, eliminating both pH and potential gradients. On the other hand, if the pump were electroneutral, then an H + diffusion potential could still develop, and the carbocyanine dye should detect this potential. This indeed happens, as illustrated in Fig. l l , where-although no potential forms in the absence of TCS-a potential does form in the presence of TCS. The results even allow calibration of the pH gradient in the vesicles (Rabon et al., 1978). Alternatively, the vesicles might develop a positive (interior) potential due to pump action, which can be checked either with an isotopically labeled lipidpermeable anion such as SCN- or with a fluorescent probe such as anilinonaphthosulfonic acid. Again no uptake is noted in the absence of ionophore. The addition of valinomycin, however, should and does stimulate accumulation of both anions, presumably because of the membrane potential which results from inward diffusion of K + (Schackmann et al., 1977; Lewin et al., 1977). Confirmatory data were obtained in a stopflow spectrophotometer having a time resolution of 20 msec. Thus all data with lipid-soluble anions or cations-of either the radioactive or fluorescent variety-demonstrate that the gastric mucosal (H' + K +) pump-ATPase functions nonelectrogenically, at least in everted vesicle preparations.
154
0
- .02 u)
c ._ c 3
8 -.04 c
0
e s
n
a
-.06
-.08
Seconds FIG. 11. An experiment using the lipid-permeable cation diethyloxocarbodicyanine (DOCC) and its absorbance changes to demonstrate the electroneutrality of transport by the gastric ATPase. The dye senses the potential due to the induction of proton conductance by TCS in the presence of the H + gradient developed by the pump after A T P addition. Curve 3 shows the signal developed without the added H t conductance, as well as that developed in the presence of permeant cations such as TPMP. Curve 1 shows the effect of TCS, and curve 2 shows the effect of a permeant buffer, imidazole (0.3 mM). It can be concluded that the TCSdependent DOCC signal is due to an H + diffusion potential. Curve 4 shows the effect of vesicle shrinkage at a fixed internal K t level prior to A T P addition. The inset shows that the DOCC signal can be related quanitatively to the prevailing p H gradient in the presence of TCS.
X.
pH GRADIENT AND STOICHIOMETRY
As mentioned in Section I, the intact gastric mucosa generates a pH gradient of almost 7 units. If the ratio of reduced to oxidized pyridine nucleotide can be taken as a reasonable measure of change in cell pH, the gastric parietal cell has a cytosolic pH of 7.7 during secretion (Sarau et al., 1975). This implies an actual gradient of very close to 7 units. The isolated gastric cells accumulate aminopyrine at a ratio of up to 300 under optimal
a. ATP-DEPENDENTCOMPONENT
155
OF GASTRIC ACID SECRETION
conditions. By measurement of the vacuolar space the actual accumulation ratio would be close to 3000. Since the pK of aminopyrine is 5 , the pH of the secretory canaliculi in the isolated glands must be near 1.5, not an unreasonable value given some leakage of buffer into the intracellular canaliculi. The highest accumulation ratio for aminopyrine produced by the addition of ATP to permeabilized glands was 50, in the presence of mitochondria1 inhibitor. We can estimate that the maximal accumulation volume under these circumstances was 5% of the gland volume, giving an actual accumulation ratio of 1000, hence an average pH of 2, which is a gradient of 5.4 units induced by ATP. With all the uncertainties of these calculations, it seems clear that gastric ATPase in isolated form should generate a very large pH gradient indeed, perhaps close t o the theoretical maximum. In order to measure this pH gradient in the gastric vesicles, various weak bases can be used, such as 9-aminoacridine, acridine orange, aminopyrine, or even imidazole. Alternatively the potential induced by the protonophore, TCS, and followed with cyanine dyes, can be used to estimate the gradient. From this, the largest pH gradient achieved by the vesicles is between 3 and 4 units, a considerable shortfall (Lewin et al., 1977). The problem is further complicated by the fact that the actual pH achieved internally is almost 2 units higher than would be expected from the measured H + deficit in the external medium, as determined by an indicator dye (bromcresol green) or a glass electrode. This kind of discrepancy might be explained by the presence of intravesicular buffer, or by H -binding sites on the protein, so that the H + disappearing from the medium does not really correspond to the H + appearing inside. However, the presence of a nonpermeable buffer cannot in fact explain the shortfall, since ATP is added in excess. A starting internal level of 150 nmoles K + per microliter of vesicle space should suffice for the transport of an equivalent quantity of protons (assuming 1:1 stoichiometry). The figure is somewhat larger than that measured with extravesicular pH electrodes but shows that considerable buffering capacity must be present. It seems likely that a limitation of the sustainable pH gradient in gastric vesicles arises from a limitation of K + entry, which-even in the presence of valinomycin-cannot compensate for the H exit rate. Whether excessive H + leakage is an artifact of the isolation procedure is not yet clear, although the leakage of H + in intact gastric cells appears considerably lower than that in isolated vesicles. It is not possible in vesicles therefore t o determine the maximal H + gradient that can be generated by the ATPase. Either some means must be found to isolate a more physiological system, or the H + leakage of the standard vesicles will need to be modified. (Another detail that is difficult to deal with is the fraction of vesicles +
+
156
G. SACHS et 6'1.
transporting. For example, if only 10% of the isolated vesicles were able to generate a pH gradient, then the pH gradient would increase by one order of magnitude from that calculated above.) In order to generate a pH gradient of 7 units the H + / A T P stoichiometry cannot exceed 1; and for a pH gradient of 3.5 units, it cannot exceed 2. Consequently, the stoichiometries measured in the gastric vesicles could be useful in determining the adequacy of the pump in generating the necessary acid gradient, There are, however, considerable technical difficulties in measuring stoichiometry. Without an exact knowledge of the internal buffering capability, only the H + disappearance from the medium is meaningful. Since internal K + is essential for transport, these measurements should be made at a high internal K + . Since there is also a finite leakage of H from the vesicle interior, only initial rate measurements can be used for calculation. The resultant stoichiometry is close to 2 H + / 1 ATP split at the obligatory pH of 6.1. However, it is now known that the first reaction under the equilibrated conditions used in these measurements is the displacement of K + from an external site, which may be accompanied by the uptake of nontransported H +. (Arguing against this possibility is the lack of a pH change in the presence of nigericin, which should not prevent the K +-dependent H -binding reaction.) If indeed the stoichiometry under initial conditions were 2:1, then even for a pH gradient of 4 units the stoichiometry would fall at longer time intervals; and if the pH gradient could be made to reach physiological levels, then the stoichiometry would reach l:l. We suggest therefore that the stoichiometry of this pump-in contrast to that of most others-may be variable. Variable stoichiometry has been claimed for bacterial transport systems (Ramos and Kaback, 1977). Other workers have also claimed very low stoichiometries, but based on measurements of internal pH, carried out in nonequilibrated, valinomycin-treated vesicles. In these cases the time t o reach the maximal gradient was about 10 minutes, compared to less than 60 seconds under the equilibrated conditions. +
+
XI.
STRUCTURAL ASPECTS OF THE ATPase
Although SDS- polyacrylamide gel electrophoresis shows the presence of a single molecular-weight group of peptides, tryptic digestion has suggested the presence of two or perhaps even three peptides (Saccomani et al., 1979~).The active groups involved in enzyme activity have been defined by reagents that are more or less site-specific. Amino groups have been implicated by the use of butanedione, for example (Bonting, 1980); Lee and Forte, 1979), and sulfhydryl groups by the use of p-chloromercuribenzene
8.
ATP-DEPENDENT COMPONENT
OF GASTRIC ACID SECRETION
157
(PCMB) or 5,5 ‘ -dithiobis-(2-nitrobenzoic acid) (DTNB). The former appear to be involved at the active site, since ATP protection has been documented. In addition to these, histidine appears to be involved, since diethyl procarbonate inhibits the enzyme and the activity is protected by ATP (Saccomani et al., 1980). Carboxyl groups are involved in the K +-activated pathway. There are probably several classes of such carboxyl groups, for example, an internal ethoxycarbonylethoxydihydroquinolinesensitive group, which is hydrophobic and binds K + , the latter protecting against inactivation. This corresponds presumably to the high-affinity K + site (Saccomani et al., 1981). Thus various amino acid residues seem to be involved in catalysis and transport.
XII.
SUMMARY AND CONCLUSIONS
Although relatively recently discovered, gastric H ,K+-ATPase has contributed considerable information toward our understanding of the overall mechanism of ion transport. Perhaps the enzyme will serve as a useful model for studying biochemical aspects of transport, given both the ease of measurement of its physiological parameters and the ease of chemical isolation. The parietal cell and its machinery may also be able to answer questions of general biological relevance, given present technical capability for simultaneously controlling cytoplasmic composition and measuring cellular function. From a comparative point of view, there are two outstanding facts about transport ATPases. First, plasma membrane ATPases and mitochondria1 ATPases differ greatly in both structure and reaction mechanism. Second, even among plasma membrane ATPases, reaction mechanisms appear to differ with regard to the degree of internal charge coupling-that is, the degree of electrogenicity. It is interesting to try to rationalize some of these differences. The phosphate group may be regarded as central to the transport mechanisms of these proteins, whether or not it is actually covalently bound. In all cases the phosphate group determines the affinity transitions during transport and also determines the sidedness of the transport “channels.” Two forms of enzyme-phosphate complexes exist: complex I, with a high affinity for cations such as H + , Na+, and Ca2+,and complex 11, with a low affinity for these cations. In the case of uniport pumps, such as the F,-Fo, conversion of complex I to complex I1 follows the addition of ATP, with release of the cation and a return of complex I1 to the free enzyme. In the case of plasma membrane-type ATPases, the return of complex I1 to the ground state appears t o require the binding of another cation. Common +
G. SACHS et a/.
158
to all is probably the binding of K + , well established in the case of gastric and Na+,K+-ATPases, but still disputed in the case of Ca*+-ATPase. In this picture, electrogenicity of the mitochondria1 ATPase is intrinsic to the mechanism, while electrogenicity of the second group of enzymes depends upon the stoichiometric requirements for the counter ion. The physiological value of electrogenicity is the ability of such pumps to maintain constant electrochemical gradients for the particular ionic species. The gradient for certain ions, chiefly H + and Na+, can then be put to use in secondary symport or antiport processes, for cellular accumulation of sugars, amino acids, etc. In contrast, the H + gradient developed across gastric ATPase is important for its own sake, not for driving cotransport. Thus there would be no obvious advantage for this pump to operate in an electrogenic mode. A similar situation may obtain for the CaZ+-ATPase of the sarcoplasmic reticulum. Na+,K+-ATPase, the plant membrane H+-ATPases, and mitochondrial-type ATPases all generate ion gradients that are subsequently used for coupled transport processes. Although the chemical gradients of these ions may vary widely under physiological conditions, the electrogenic component in the pump reaction should allow the total electrochemical gradient to be kept invariant, or at least large, during wide swings of the concentration gradient.
ACKNOWLEDGMENTS Supported by NIH grants AM-28459, NFS grants PCM 80-08625 and 78-09208, and the Urology Research and Rehabilitation Center, University of Alabama in Birmingham.
REFERENCES Baker, P . F., and Knight, D. F. (1978). Nature (London) 267, 620-622. Berglindh, T. (1978). Acta Physiol. Scand. Spec. Suppl. 55-68. Berglindh, T., and Obrink, K. J . (1976). Acta Physiol. Scand. 96, 150-159. Berglindh, T., Helander, H . F., and Sachs, G. (1979). Scand. J . Gastroenterol. Suppl. 55, 7-14. Berglindh, T., Dibona, D. R., Pace, C. S., and Sachs, G. (1980). J . Cell B i d . 85, 392-401. Bonting, S. L., de Pont, J. J. H. H. M., van Arnelsvoort, J. M. M., and Schrijen, J . J. (1980). Ann. N. Y . Acad. Sci. 341, 335-356. Conway, E. J. (1950). Science 113, 270-273. Dell’Antone, P., Colonna, R., and Azzone, G. F. (1972). Eur. J . Eiochem. 24, 553-565. Dibona, D. R., Ito, S., Berglindh, T., and Sachs, G. (1979). Proc. Nail. Acad. Sci. U.S.A. 76, 6689-6693. Forte, T. M., Machen, T. E., and Forte, J. G. (1977). Gastroenterology 73, 941-955. Ganser, A. L., and Forte, J. G. (1973). Eiochim. Biophys. Acta 307, 169-180.
8. ATP-DEPENDENT COMPONENT OF GASTRIC ACID SECRETION
159
Harris, J. B., Frank, H., and Edelman, 1. S. (1958). A m . J . Physiol. 195, 499-504. Hasselbach, W. (1978). Biochim. Biophys. Acta 515, 23-53. Helander, H . F. (1962). J . Ultrastruct. Res. Suppl 4, 1-123. Kagawa, V., and Racker, E. (1971). J. Biol. Chem. 246, 5477-5487. Lee, H. C., and Forte, J. G. (1979). Fed. Proc. Fed. A m . SOC. Exp. Biol. 38, 1041. Lee, J., Simpson, E., and Scholes, P . (1974). Biochem. Biophys. Res. Commun. 60, 825-834. Lewin, M., Saccomani, G., Schackmann, R., and Sachs, G. (1977). J . Membr. Biol. 32, 301-318. Lund, E. S. (1928). J . Exp. Zoo/. 51, 265-337. M%rdh,S., and Post, R. L. (1977). J . Biol. Chem. 252, 633-638. Michelangeli, F., and Proverbio, F. (1978). Acta Physiol. Scand. Spec. Suppl. 399-408. Morowitz, H. J. (1978). A m . J . Physiol. 4, R99-Rl14. Mitchell, P. (1966). Biol. Rev. 41, 445-502. Rabon, E., Chang, H. H., and Sachs, G. (1978). Biochemistry 17, 3345-3353. Rabon, E., Saccomani, G., Kasbekar, D. K., and Sachs, G. (1979). Biochim. Biophys. Acta 551, 432-447. Rabon, E., Takeguchi, N., and Sachs, G. (1980). J . Membr. Biol. 53, 105-117. Ramos, S., and Kaback, H. R. (1977). Biochemistry, 16, 4271-4275. Rehm, W. S. (1972). In “Metabolic Pathways” (L. E. Hokin, ed.), Vol. 6, pp. 187-241. Academic Press, New York. Rehm, W. S., and LeFerre, M. E. (1965). A m . J . Physiol. 208, 922-930. Saccomani, G., Shah, G., Spenney, J. G., and Sachs, G. (1975). J . Biol. Chem. 250, 48024809. Saccomani, G., Stewart, H . B., Shaw, D., Lewin, M., and Sachs, G. (1977). Biochim. Biophys. Acta 465, 31 1-330. Saccomani, G., Helander, H . F., Crago, S., Chang, H. H., Dailey, D. W., and Sachs, G. (1979a). J . Cell Biol. 83, 271-283. Saccomani, G., Chang, H. H., Mihas, A. A., Crago, S., and Sachs, G. (1979b). J . Clin. In vest. 64, 627-63 5. Saccomani, G., Dailey, D. W., and Sachs, G. (1979~).J . Biol. Chem. 254, 2821-2827. Saccomani, G., Barcellona, M. L., Rabon, E., and Sachs, G. (1980). In “Hydrogen Ion Transport in Epithelia” (I. Schulz, G. Sachs, J. G. Forte, and K. J . Ullrich, eds.). Elsevier, Amsterdam. Saccomani, G., Barcellona, M. L., and Sachs, G. (1981). J . B i d . Chem. (in press). Sachs, G., Chang, H . H., Rabon, E., Schackmann, R., Lewin, M . , and Saccomani, G. (1976). J . Biol. Chem. 251, 7690-7698. Sarau, H. M., Foley, J., Moonsamy, G., Wickelhaus, V. D., and Sachs, G. (1975). J . Biol. Chem. 250, 8321-8329. Schackmann, R., Schwartz, A., Saccomani, G., and Sachs, G. (1977). J . Membr. B i d . 32, 361-381. Skou, J. C. (1965). Physiol. Rev. 45, 596-617. Sedar, A. W. (1965). Fed. Proc. Fed. A m . SOC. Exp. Biol. 24, 136-1367. Wallmark, B., and M%rdh, S. (1979). J . Biol. Chem. 254, 11899-11902. Wallmark, B., Stewart, H . B., Rabon, E., Saccomani, G., and Sachs, G . (1980). J . Biol. Chem. 255, 5313-5319.
This Page Intentionally Left Blank
Part Ill
ReversibiIity: ATP Synthesis Driven by Electric Fields
This Page Intentionally Left Blank
C U R R E N T T O P I C S IN MEMBRANES AND TRANSPORT. VOLUME 16
Chapter 9 Effect of Electrochemical Gradients on Active H' Transport in an Epithelium QAIS AL-A WQATI A N D TROY E. DIXON' Departments of Medicine and Physiology Columbia University, College of Physicians and Surgeons New York, New York
I. 11. 111. IV. V. V1. VII.
Introduction ..................................... Proton Secretion by Turtle Bladder .......................................................... Efficiency of Energy Conversion ..................................... Reversibility ................................................................ Stoichiometry ..................................................................................... Ion Transport as a Pacemaker of Cellular Metabolism ........................ Conclusions ....................................................................
..............................................................
1.
163 164 167 168 171 172 173 174
INTRODUCTION
Unlike transport through passive channels, ion flow through pumps is determined not only by electrochemical gradients but also by the free energy of a coupled metabolic reaction. The direction of movement of the ion depends on the balance between the electrochemical gradient and the metabolic force. The study of the interaction between these two types of forces in epithelia poses special problems dictated largely by the complex cellular architecture of these tissues. What makes transport across an epithelium different from other modes of cellular transport is that ions have to traverse two membranes in series, each with its own composition Present Address: Department of Medicine, State University of New York at Stony Brook, Stony Brook, New York. 163
1982 by Academic Press. Inc. Copyright All righfc of reproduction in any form reserved. ISBN 0-12-153316-6
164
QAlS AL-AWQATI AND TROY E. DlXON
and conductance. Moreover, epithelial cells are joined together by junctions that might be tight or leaky to ion flow in different epithelia. Hence externally applied electrochemical gradients will be reflected at each of these barriers depending on the conductance of the membrane and the relation among the various conductive elements of the tissue as a whole.
II.
PROTON SECRETION BY TURTLE BLADDER
We studied the effect of transepithelial electrochemical gradients on H + transport in the turtle urinary bladder, a “tight” epithelium with transepithelial resistances in excess of 3 kQ/cm2. It transports several ions actively; sodium is absorbed from the lumen, protons are secreted into the lumen, and chloride is absorbed in strict exchange for secreted bicarbonate (Fig. 1) (Steinmetz, 1974). These transport processes are independent of each other. When sodium transport is inhibited by ouabain, amiloride, or removal of ambient sodium, proton transport is the only active currentcarrying flaw in this membrane (Fig. 2). The proton pump appears t o be located in the luminal membrane of the epithelium (Steinmetz, 1969). Since this membrane accounts for more than 70% of the total transcellular resistance (Hirschhorn and Frazier, 1971) (especially in the presence of amiloride), it follows that an applied CELL
LUMEN
H+
BLOOD
I
&
HCOS
FIG. 1. A model of the ion transport processes in the turtle urinary bladder. N a b is transported from the lumen by the amiloride-sensitive pathway and out of the cell by the ouabain-inhibitable Na+ pump. A HC0;-secreting process present, which is electroneutral, exchanged absorbed CI- for the secreted HCO;. This process is dependent on cellular metabolism. Although the site of this exchanger is probably the luminal border, the evidence for its location needs to be obtained by direct means. An electrogenic proton pump is located in the luminal border. (For reviews of these processes see Steinmetz, 1974, and Al-Awqati, 1978.)
9. EFFECT OF ELECTROCHEMICAL GRADIENTS ON
scc
Hf
TRANSPORT
165
10 20 (nonomotes min-'1
FIG.2. Relation between the rate of acidification measured by pH-stat titration and the short-circuit current in ouabain-treated turtle bladders.
transepithelial potential will drop across this barrier almost quantitatively. This has greatly facilitated analysis of the effect of transepithelial potential on active H + transport. The low-resistance basolateral border contains sites that are freely permeable to HC0,- (or OH-). When the membrane is shortcircuited and bathed by solutions whose pH is 7.4, it appears that the net proton electrochemical gradient across the luminal membrane is zero. Addition of a proton conductor such as dinitrophenol or amphotericin B to the luminal membrane does not result in any change in the rate of transport (Beauwens and Al-Awqati, 1976; Steinmetz and Lawson, 1970). Further, application of a graded lumen-positive potential reduces the net rate of H' transport in a manner identical to the effect of a net transepithelial pH difference (Fig. 3) (Al-Awqati et al., 1977). The electrochemical potential difference at which the rate of net H + transport is zero, the so-called proton motive force (PMF), is seen to be 180 mV. The identity of the values of the PMF, measured either by pH displacement or by voltage displacement, supports our conclusion that the turtle bladder behaves essentially as a single barrier. The results in Fig. 3 show that, as the luminal solution becomes more acid (or electrically positive), the net rate of H + movement into the lumen is reduced. Since the net rate is the pump rate minus the leakage it is not possible to tell from this study the extent, if any, of the effect of the gradient on the pump. To arrive at an independent measure of the pump rate we measured the rate of oxidative metabolism by the epithelium simultaneously with the rate
-
166
QAlS AL-AWQATI AND TROY E. DIXON
30 AJr
-O---O 0-0
2 5 1
5t I
ApH
xN
.\ N
'<
I
I
I
I
I
24
53
83
112
143
171
ApH (units) .4
.9
1.4
1.9
2.4
2.9
FIG.3. Effect of applying proton electrochemical gradients o n the rate of H f transport in turtle urinary bladders. The rate of transport (JH)was measured by pH-stat titration after applying graded lumen-positive potentials (A$) at an ambient p H of 7.4. In the same bladders JHwas measured in the short-circuited state as the short-circuit current after the luminal p H was reduced in several steps. (Drawn from data in Al-Awqati et a[., 1977.)
of net H + transport. The H + pump can be thought of as a vectorial metabolic reaction in which two flows, transport and chemical, are coupled. To measure the pump rate one needs to measure either one of these flows. We measured the I4CO, production rate as an index of the metabolic reaction that fuels the pump. We used an ionization chamber method that allowed us to measure the rate of I4CO2evolved from the oxidation of various I4C-labeled substrates simultaneously with the rate of transport and the transepithelial electrochemical gradient (Beauwens and Al-Awqati, 1976). Reducing the luminal pH reduced the rate of transport and the rate of I4CO, production (Fig. 4). In separate experiments we showed that the rate of I4CO2production is reduced when the epithelium is clamped at a lumen-positive potential in the absence of a pH difference (Kelly et al., 1980). These results demonstrate that the reduction in the rate of net transport by the electrochemical gradient, at least in part, is due to a reduction in the pump rate. To arrive at a more quantitative evaluation of this we measured the relation of H + transport to I4CO, production under two circumstances. In one the rate of transport was reduced by applying a pH difference, while in another the ambient p C 0 , was reduced in the absence of a p H difference. As seen in Fig. 4, reduction of the ambient p C 0 , results in qn inhibition of the rate of H + transport and in the rate of glucose oxidation. The slope of H + transport in 14C0, production in the absence of a transepithelial electrochemical gradient is a measure of the characteristic stoichi-
Hf
9. EFFECT OF ELECTROCHEMICAL GRADIENTS ON
167
TRANSPORT
4/%
u)
2
5
4
I
I
-3
3
2
Time ( h o u r s )
MpH
6.55 14.78 15.62
I 7.1 I I
6.93
FIG.4. Relation between H + transport ( J , ) . the lumenal p H (MpH), and the rate of I4CO2 production (J&2) from uniformly labeled [14C]glucose. The serosal p H was kept constant at 7.1. (From Beauwens and Al-Awqati, 1976.)
ometry of the pump. In the absence of electrochemical gradients net transport represents flow through the pump, while in the presence of gradients net transport is the pump rate minus the back flux through parallel leaks. Only when there is no leak will the two stoichiometries measured in the presence or absence of gradients be equal. In 14 experiments there was no significant difference between the two stoichiometries (14.7 versus 14.2 Eqlmole) (Beauwens and Al-Awqati, 1976). Several conclusions can be drawn from the finding that net transport in the presence of a gradient is identical to the rate of the pump. First, the leakage pathway in this epithelium is negligible. Second, the P MF that we measure as the electrochemical gradient at zero net transport is in fact the PMF of the pump. Third, thermodynamic analysis of these findings (see below) shows that the efficiency of energy conversion by the pump is very high.
111.
EFFICIENCY OF ENERGY CONVERSION
Using an irreversible thermodynamic model of active transport (Essig and Caplan, 1968) one can state the formal equation of coupling:
J,
=
L, AD,
+- L,,
AGr
J,
=
L,, A&
+ L,
AG,
(1)
168
QAlS AL-AWQATI AND TROY E. DIXON
where the J’s are fluxes of protons ( J H + and ) metabolic reactions (J,). Each flux is coupled t o two forces, the electrochemical gradient (AjlH) and the free energy of the driving metabolic reaction (AG,). Each flux is coupled to its conjugate driving force by a straight coefficient (and to the other force by a cross-coefficient) (L). By Onsager symmetry the two crowcoefficients are equal. The PMF is given by
( A j l ~JH ) = = (412) AGr
(2)
Inspection of Eq. (2) shows that the ratio of the two forces (at zero flux) is equal t o the stoichiometry between the two fluxes multiplied by some factor ( q ) which is a function of the efficiency of energy conversion (Kedem and Caplan, 1965). When q is 1 , the system is maximally efficient, and when it is 0, the two processes (transport and reaction) are uncoupled. To determine q it can be shown from Eq. (1) that
, 0 -~ - (LH,)Z 42 = ( a J ~ / a J ) A j i= ( a J H / a J , ) ~ jzl ~o
(3)
L H L,
Since we have shown that the two slopes needed to measure q are equal, it follows that q G 1 and the H+ pump behaves as a near perfect energy converter. Although the conversion of chemical to electroosmotic energy by the H+ pump is near perfect, it does not follow that there is no entropy production by this system during transport. Entropy is produced by the synthesis of ATP from intermediary metabolism. During transport the rate of entropy production by the pump is given by = JH Aji,
+ J,
AG, 2 0
The finding that the system is highly efficient means that the entropy production by the pump is at a minimum.
IV.
REVERSlBlLlTY
This high degree of efficiency suggests that the pump is reversible. We tested this by measuring the intracellular ATP content in poisoned turtle bladders (Dixon and Al-Awqati, 1979a). Exposing the bladder to iodoacetate and cyanide resulted in a progressive decline in ATP levels. When all the ATP synthetic capacities of the cell were abolished, a large transepithelial gradient resulted in an increase in intracellular ATP content (Fig. 5 ) . It is to be expected that, if the ATP is synthesized by the proton pump, then it should only be synthesized at gradients greater than the PMF. This is seen to be the case in Fig. 6 . At AjlH greater than 120 mV
9. EFFECT OF ELECTROCHEMICAL GRADIENTS ON
H t TRANSPORT
169
TIME (hours) FIG. 5. Effect of addition of iodoacetate (2 mM) and sodium cyanide (2 mM) to the serosal side at t = 0. ATP levels were measured at designated times. 0 , Experiments in which the transepithelial proton electrochemical gradient (Aii,) was -60 mV; 0 , A&, of 310 mV was applied for 40 minutes. The dashed line should not be taken to imply that ATP synthesis is linear over this period time. Indeed it appears to be highly nonlinear, with initial rates being much higher than later rates. (From Dixon and Al-Awqati, 1979a.)
the increase in cellular ATP levels was proportional to the gradient. Note here that the PMF is only 120 mV rather than the 180 mV seen in Fig. 3 . The cause of this decline can be seen from Eq. (2). The PMF depends on AG,,,, which should be lower in poisoned cells than in normal cells. Both a pH difference and a potential difference can lead to ATP synthesis, provided the gradient is greater than 120 mV (Fig. 7). Further evidence that the ATP is synthesized by a proton pump located at the luminal border is given in Table I. The ATP synthesis can be prevented by pretreatment of the luminal medium by the proton conductor dintrophenoi. Further, addition of dicyclohexylcarbodiimide (DCCD) to the luminal solution results in a rapid inhibition of H + transport, which does not return on washing the luminal medium. The ATP content of these cells was not changed, suggesting that this brief exposure (15 minutes) of the luminal surface did not result in the entry of DCCD into the cytoplasm (Table I). However, in identically treated bladders poisoned with iodoacetate and cyanide an electrochemical gradient larger than the P M F of the pump failed to result in ATP synthesis (Table I). These results indicate that the proton pump of the turtle bladder is a reversible ATPase.
I
-60
I
I
I
0
60
120
I
180
I
I
240
300
TRANSEPITHELIAL A p H (mV) FIG. 6. Effect of application of Ap,, on ATP levels in epithelial cells. ApH was applied for 40 minutes and was composed of a A$ and a A$ and a ApH in different combinations. (From Dixon and Al-Awqati, 1979a.)
A pH = 2 units
0
I00
A9mv
A$ = 180
200
ApH (units)
FIG.7. Effect of a ApH and a A$ on ATP in epithelial cells. (Left) The effect of increasing A$ was tested in bladders clamped at a ApH of 2 units. (Right) The ApH was increased in bladders clamped at 180 mV. The Aji, was applied for 40 minutes. From Dixon and Al-Awqati, 1979a.)
9. EFFECT OF ELECTROCHEMICAL GRADIENTS ON
H + TRANSPORT
171
TABLE I ATP SYNTHESIS I N POISONED CELLS~
Addition
Cyanide (2 mM) plus ApH iodoacetate (5 mM) (310 mV) for 80 minutes for 40 minutes
+
DNP (2 mM) DCCD (0.2 mM) DCCD (0.2 mM)
-
+
ATP (nmoles/mg protein)
n
Control
Experimental
+
5
-
4
10.7 32.6 3.1
0.7 30.9 0.5
+
6
A
f
SEM
10.0
& 4.0b 1.7 i 1.9 2.6 i 0.6b
Effect of various treatments on ATP synthesis in response to an electrochemical gradient (ApH). All additions were to the mucosal medium in the final concentrations shown. b p < 0.05.
V.
STQICHIQMETRY
A number of methods are available for evaluating the H + / A T P stoichiometry of the proton pump. The ratio of the two fluxes is the most rigorous method but, due to technical difficulties, we have not as yet been able to measure it by this method. Another method is the ratio of the two forces given in Eq. (2). In the presence of a very high efficiency of energy coupling the major ambiguity deals with measurement of the AGATp in the epithelial cell. This is given by ACATp
= AGIT,+ R T l n ([ATP]/[ADP][P,])
Simultaneous measurements of the PMF and the ATP, ADP, and P, concentrations in normal bladders are shown in Table 11. The stoichiometry is - 3 H+/ATP. The AGiTpused was 30.16 kJ/mole for a Mg2+activity of 1 m M a t pH 7.0 and 22°C (Guynn and Veech, 1973). Since we do not know whether the ATP and ADP concentrations measured in fact equal the free nucleotide activity, these results should be considered preliminary and in need of further verification. A third method is to measure the ratio of H + transport to the rate of cellular oxidative metabolism which yields a stoichiometry that, given some assumptions, should equal the real stoichiometry. We measured the cou-
TABLE !I
H+/ATP STOICHIOMETRY (2) JH
(nEq min-I cm-2) 5.96
f
3.25
PMF (mV) 178.1
* 7.0
ATP/ADP*P, (M-9 6705
f
561
AGATP (mV)
z
3.93 519 f 0.40 2.92
f
AGATp
(kJ/mole) 50.11
f
0.11
172
QAlS AL-AWQATI AND TROY E. DlXON
TABLE 111 H + /ATP STOICHIOMETRY" I4C-labeled substrate
n
AJ,/AJ&~
H+/ATP
Glucose 0-Hydroxybutyrate Butyrate Oleate
6 8 5
15.4 f 2.8 18.5 f 3.6 28.5 f 6.9 29.5 f 6.1
2.5 3.0 3.5 3.5
6
The calculation of the H + /ATP ratio assumes a P/O ratio of 6 and a respiratory quotient of 1 for glucose and P-hydroxybutyrate and 0.7 butyrate and oleate.
pling ratio between I4CO, production and H + transport (Fig. 4). Two critical and untested assumptions are involved. One is the identity of the intracellular and extracellular specific activities of the metabolites used, and the other is that the P / O ratio is 6. Given these untested assumptions Table I11 shows that the stoichiometry based on this method is reasonably near 3 H /ATP. The similarity of the two estimates of the stoichiometry is comforting and suggests that it may indeed be 3 H+/ATP. However, it should be kept in mind that the errors in the two methods tend to overestimate the stoichiometry. +
VI.
ION TRANSPORT AS A PACEMAKER OF CELLULAR METABOLISM
It is common knowledge that changes in ion transport lead to changes in oxidative metabolism in cells (Whittam, 1961). As seen in Fig. 3, this is clearly the case of H + transport. The mechanism by which the epithelium (or more accurately the mitochondria) senses the changes in ion transport must be through changes in ATP, ADP, or P,. While the well-studied phenomenon of respiratory control of mitochondria1 oxygen consumption is frequently thought of in terms of the supply of ADP, simple thermodynamic arguments predict that the oxygen consumption should be related to the free energy of ATP hydrolysis. Indeed, Thayer (1977) has recently shown that the oxygen consumption by isolated beef heart mitochondria decreases linearly as the AGATpof the medium increases from 30 to 55 kJ/mole. Reducing the luminal pH to pH 5.0 from 7.4 caused the AG,,, of epithelial cells to increase by 5 kJ/mole (Table IV). If the oxygen consumption of turtle bladder mitochondria responds in a manner similar to that of beef heart (Thayer, 1977), then a change in cytoplasmic AG of 5 kJ/mole would produce a reduction in oxygen con-
-
-
9. EFFECT OF ELECTROCHEMICAL GRADIENTS ON
H+
TRANSPORT
173
TABLE IV EFFECT OF MUCOSAL ACIDIFICATION (n = 7) pH 7.4 5.58 f 3.75 f 1.46 f 0.71 f 5.90 f 4.32 i 4.13 f 905 f 46.00 f
3.49 0.20 0.27 0.13 0.31 0.67 1.1 330 2.2
pH 5.0 0.47 f 5.26 f 0.61 0.83 h 6.70 f 2.75 f 11.26 f 6383 f 50.66 f
0.64” 0.47a 0.13a 0.27 0.30 0.74a 2.6” 1899a 2.8“
‘ p < 0.05;pH 7.5 versus pH 5.0.
sumption of -30-40’70. Inspection of Fig. 4 shows that this is of a range similar to that of the change in I4CO, production seen in the turtle bladder. Further studies should allow a more quantitative analysis. However, qualitatively it is clear that the signal to the mitochondria appears to be the cytoplasmic A G A T p . Recently we have shown that changes in the rate of H + transport induced by increasing the p C 0 2 and by inhibiting carbonic anhydrase lead to appropriate changes in the A G A T p . Increases in transport lead to a reduction in the A G A T p , and vice versa (Dixon and Al-Awqati, 1979b). The fact that these changes are large indicates that the pool of ATP, ADP, and Pi is small in relation to the rates of ATP synthesis and hydrolysis and that, if compartmentalization of nucleotides is present, it does not confound these conclusions to any large degree.
VII.
CONCLUSIONS
The evidence for the electrogenicity of the H + pump in the turtle bladder may be summarized as follows: 1. Transepithelial potential is lumen positive during active H + movement into the lumen (Steinmetz, 1974). 2. There is identity between short-circuited current and proton flux (Steinmetz, 1974). 3. Removal of ambient Na+,K + , and C1- has no effect on H + flux if care is taken to keep ambient CO, tension constant. 4. H secretion is inhibited by lumen-positive potential (Al-Awqati, 1977). 5 . Reversal of potential and ATP synthesis by H + pump occurs in response to a lumen-positive potential (Dixon and Al-Awqati, 1979a). +
174
QAlS AL-AWQATI AND TROY E. DIXON
6 . Transepithelial conductance decreases when H transport is inhibited by DCCD (Al-Awqati, unpublished observations). The present studies show that the proton pump of the turtle urinary bladder is a reversible proton-translocating ATPase. The efficiency of energy conversion by the proton pump is very high. The stoichiometry of the pump is tentatively assigned a value of 3 H + / A T P . Changes in the rate of transport result in changes in the cytoplasmic AG,,,. It is suggested that this is the signal by which the pump regulates the rate of cellular oxidative metabolism. +
REFERENCES Al-Awqati, Q . (1978). A m . J. Physiol. 235, F77-F88. Al-Awqati, Q., Mueller, A , , and Steinmetz, P. R. (1977). A m . J. Physiol. 233, F502-F508. Beauwens, R . , and Al-Awqati, Q. (1976). J . Gen. Physiol. 68, 421-439. Dixon, T. E., and Al-hwqati, Q . (1979a). Proc. Natl. Acad. Sci. U.S.A. 76, 3135-3138. Dixon, T. E., and Al-Awqati, Q . (1979b). Kidney Int. 16, 811. (Abstract) Essig, A., and Caplan, S. R. (1968). Biophys. J. 8, 1434-1457. Guynn, R. W . , and Veech, L. (1973). J . Biol. Chem. 248, 6966-6972. Hirschhorn, N . , and Frazier, H. S. (1971). A m . J . Physiol. 220, 1158-1161. Kedem, O., and Caplan, S. R . (1965). Trans. Faraday SOC. 61, 1897-1911. Kelly, S., Dixon, T. E . , and Al-Awqati, Q . (1980). J. Membr. B i d . 54, 237-243. Steinmetz, P. R. (1969). J. Clin. Invest. 48, 1258-1265. Steinmetz, P. R. (1974). Physiol Rev. 54, 890-956. Steinmetz, P. R., and Lawson, L. R. (1970). J. Clin. Invest. 49, 596-601. Thayer, W . S., Tu, Y.-S., and Hinkle, P. C. (1973). J. B i d . Chem. 252, 8455-8458. Whittam, R. (1961). Nature (London) 191, 603-604.
CURRENT TOPICS IN MEMBRANES AND TRANSPORT, VOLUME 16
Chapter 10 Coupling between H' Entry and ATP Synthesis in Bacteria PETER C. MALONEY Department of Physiology The Johns Hopkins University School of Medicine Baltimore, Maryland
I.
Introduction
........................................................................................ ......................................... ..............
175
11. Voltage-Driven Reversal
111. IV. V. VI.
Proton Entry Coupled to ATP Synthesis .................... .............. Stoichiometry of the Coupling between H + and ATP Rates of ATP Formation and the Nature of the Drivin ....................... Conclusions ......................................................................................... References ..............................................
1.
187 191 192
INTRODUCTION
The experiments discussed here focus on the coupling between proton movements and the synthesis of ATP, a reaction catalyzed by the protontranslocating ATPase of bacteria. The diagrams in Fig. 1 illustrate the reversible nature of this ATPase in the bacterial world (for recent reviews, see Harold, 1977; Rosen and Kashket, 1978). Figure 1A indicates that in some cells, such as the streptococci, the ATPase normally couples ATP hydrolysis to the extrusion of protons, establishing both the membrane potential and pH gradient needed for other work functions (Harold et al., 1970; Harold & Papineau, 1972). The alternative role is shown in Fig. 1B. In this case, as in Escherichia coli during aerobic growth, it is the reactions of electron transport that initiate proton extrusion and maintain the electrochemical proton gradient at about 200 mV or more and directed inward (Collins and Hamilton, 1976; Ramos and Kaback, 1977; Zilberstein el al., 1979). Subsequent to this, the ATPase links the reentry of protons, down 175
Copyright @J 1982 by Academic Press, Inc. All rights of reproduction in any form reserved ISBN 0-12-153316-6
176
PETER C. MALONEY
B
A
H' -
PROTON EXTRUSION
PROTON ENTRY
(ANAEROBIC)
(AEROBIC)
FIG. 1. The proton-translocating ATPase of bacteria. (A) Proton extrusion coupled to ATP hydrolysis occurs in anaerobes (e.g., S. luctis) and in facultative anaerobes grown in the absence of terminal electron acceptors. (B) Proton entry coupled to ATP synthesis during oxidative phosphorylation occurs in aerobes and in facultative organisms (e.g., Escherichiu coli) when grown in the presence of an electron acceptor.
their gradient, to the synthesis of ATP during oxidative phosphorylation. Under physiological conditions, this ATPase normally operates in only one of these two modes. Nevertheless, under appropriate experimental conditions, one may demonstrate both reactions in a single organism. In all the experiments to be presented, the strategy has been to examine the coupling between proton entry and ATP formation when an electrochemical proton gradient is artifically imposed, for in this way the effects of membrane potentials and pH gradients of known size can be studied directly. For several reasons, this work has been done using intact cells of Streptococcus lactis, an anaerobe (Fig. 1A). Such cells are easily depleted of metabolizable reserves, and in washed cells the electrochemical proton gradient falls to nearly zero. In addition, this gram-positive organism is sensitive to ionophores without special pretreatment, so that the size of the membrane potential can be manipulated by varying the ratio of internal to external potassium in the presence of the ionophore valinomycin. Finally, internal buffering power of the intact cell is high, so that net proton fluxes can be readily observed in response to applied electrical or chemical gradients.
II. VOLTAGE-DRIVEN REVERSAL The experiment shown in Fig. 2 serves as an introduction to such studies, since it illustrates the voltage sensitivity of this bacterial ATPase. In this experiment (Maloney et al., 1974), cells of S. lactis were washed free of their growth medium and placed in sodium phosphate buffered at pH 6.
10.
Hf
ENTRY AND
ATP
177
SYNTHESIS IN BACTERIA
VA LINOMYCW
5
10
I5
20
25
MINUTES FIG.2. Reversal of the proton-translocating ATPase by an imposed membrane potential. Washed cells of S . / a d s were suspended at 0.3 mg dry wt/ml in 100 m M sodium phosphate, p H 6 . After samples were removed for measurement of zero-time ATP levels, either glucose (25 m M final concentration) or valinomycin (10 pM final concentration) was added. Under these conditions, the ratio of internal to external potassium was about 1000. From the assumption that the ionophore allows the distribution of potassium to reach electrochemical equilibrium, the membrane potential established after treatment with valinomycin would have an initial value of - 180 mV. (From Maloney et a/., 1974.)
The experiment was performed in two parts. In one case, as shown by the open circles, glucose was added to test the capacity of cells to form ATP from substrate-level phosphorylations during glycolysis. In this case internal ATP rose from the low basal level characteristic of washed cells to a final stable level of about 2.5-3.0 mM. In the second part of the experiment the potassium ionophore valinomycin was added instead of glucose. The idea was to impose a large inward driving force for protons by bringing the chemical distribution of potassium close to equilibrium as the ionophore greatly increased passive permeability to this ion. From estimates of external and internal potassium, one may calculate that this imposed electric gradient (a potassium diffusion potential) can be as high as 180 mV, negative inside. The solid circles show that a transient net synthesis of ATP occurred in response to this newly imposed membrane potential. ATP levels rose rapidly at first, to a peak value equaling that found during
178
PETER C. MALONEY
glycolysis, and then showed a later slow decay, presumably reflecting an eventual decay in the driving force behind the synthetic reaction. Several observations suggest that this transient net synthesis of ATP is catalyzed by the membrane-bound ATPase, that it represents a response to the imposed electric field, and that the driving force behind ATP formation is the electrochemical potential difference for protons. Thus valinomycininduced synthesis of ATP is not found when cells are previously treated with an inhibitor of this ATPase, dicyclohexylcarbodiimide (DCCD) (Harold et al., 1969; Fillingame, 1976). Nor does ATP formation occur when external potassium is made equal t o internal potassium so that the membrane potential remains near zero after addition of the ionophore. Finally, ATP synthesis is also blocked when the cell membrane is made specifically permeable t o protons themselves. Nevertheless, under all these conditions, the capacity of cells to form ATP from substrate-level phosphorylations is unimpaired (Maloney et al., 1974; Maloney and Wilson, 1975; Maloney, 1977). The major experimental significance of these early studies was to identify the driving force behind such ATP formation as the difference in electrochemical potential for protons. This conclusion strongly supports the ideas set forth in Mitchell’s chemiosmotic theory (reviewed by Mitchell, 1979). It is now clear that in bacteria, as in chloroplasts and mitochondria, the terminal step in oxidative phosphorylation is catalyzed by a reversible proton-translocation ATPase.
Ill.
PROTON ENTRY COUPLED TO ATP SYNTHESIS
The goal of this next group of experiments was to identify the major pathways available for proton entry in response to such artifically imposed membrane potentials and in response t o imposed pH gradients of equivalent thermodynamic weight. It appears that under suitable conditions there are only two pathways of quantitative significance. One is represented by the proton-translocating ATPase itself, and when protons move inward in this way, there is an obligatory coupling to ATP formation. The second pathway presumably reflects the general physical and chemical properties of biological membranes, for when protons enter by this alternative route, they d o so in a passive manner and their movements are not coupled to the performance of work. The experiment shown in Fig. 3 describes proton entry into washed cells of S . lactis when a membrane potential was imposed, using conditions similar, but not identical, to those described in Fig. 2. In this case (Fig. 3),
10.
H+
ENTRY AND
ATP
179
S'fNTHESIS I N BACTERIA
somple no.
MINUTES
FIG. 3. Proton entry after the addition of valinomycin. Washed cells were suspended in a lightly buffered medium (0.3 mg dry wt/ml) containing 100 m M 2-aminoethanesulfonic acid, 50 rnM choline chloride, and 0.3 m M potassium hydroxide, at pH 6.5. Samples were then placed in the chamber of a pH stat. Immediately before the addition of valinomycin (10 pM final concentration) a brief titration lowered the external p H to either 6.0 (nos. 1-6) or 6.4 (no. 7). After addition of the ionophore, automatic titration by the pH stat maintained constant external p H , and proton entry was given by amount of acid required to keep the external p H at the desired set point. Samples were analyzed in the order indicated by the sample numbers (1-7). For nos. 6, 4 and 2, potassium chloride was also present to raise external potassium to 1 .O, 2.9, and 9.6 mM, respectively. Initial values for the electrochemical proton gradient were estimated from calculations of botn membrane potential and p H gradient. The membrane potential was calculated by assuming that addition of the ionophore allowed the distribution of potassium t o attain electrochemical equilibrium, using the known external levels of potassium and the average value for internal potassium (360 mM) found in parallel experiments. Contributions made by the p H gradient were calculated from known external pH, taking internal p H (6.2) as measured in separate experiments. (From Maloney, 1977.)
in order to restrict net ion movements to an overall exchange of H' and K', other permeant cations were omitted from the external medium. In addition, the outside medium was only lightly buffered so that net proton entry after the addition of valinomycin at zero time could be monitored by recording small changes in external pH. This particular experiment shows proton entry when the initial value of the imposed membrane potential was systematically varied. Seven samples were examined, and the sample numbers indicate the order of analysis. Samples 1, 3, and 5 are replicates of the control preparation. For these controls the initial value of the electrochemical proton gradient was about 200 mV, and toward this the membrane potential contributed about 180 mV. In the even-numbered experimental determinations (samples 6 , 4, and 2) this imposed driving force was reduced stepwise in units of about 30 mV, by
180
PETER C. MALONEY
continued threefold elevations of outside potassium. It is clear that these manipulations had a significant effect on proton entry. In addition, it appears that proton entry showed a “gated” response with respect to the size of the imposed driving force. For example, in the controls, the net inward driving force was initially about 200 mV and the cumulative proton entry after 6 minutes was about 110-120 pmoles H + per liter of cell water. The behavior of sample 6 shows that, when the membrane potential was lowered by only 30 mV, there was a dramatic reduction in both the rate and extent of proton inflow. Yet the subsequent depolarizations by this same 30-mV step, as in samples 4 and 2, diminished proton entry only to the extent expected from the fractional decrease in total driving force. Stated in another way, these observations show that systematic increases in membrane potential are paralleled by equivalent increases in proton entry until the total inward driving force enters the region between 170 and 200 mV. At this point, an unpredictable and substantial acceleration of proton entry can occur. The second conclusion allowed by this experiment is that the controlling element in such gated behavior correlates with the total driving force on protons and not with the absolute value of the membrane potential. This follows from a comparison of samples 6 and 7. In both cases, a 25-30 mV decrement was introduced. But in one case (sample 6 ) this was done by lowering the membrane potential, while in the other case (sample 7) the equivalent stepdown was achieved by lowering the contribution made by the pH gradient, without changing the membrane potential. The results shown in Fig. 4 clearly indicate that the elevated proton entry seen in such controls (Fig. 3, samples 1, 3, and 5 ) includes a component attributable to the coupling between proton movements and the synthesis of ATP, catalyzed by the membrane-bound ATPase. In this experiment, the behavior of control cells was compared to that found for cells previously exposed to DCCD, an inhibitor of the ATPase. In both instances, cells were suspended in a lightly buffered medium in the presence of 0.3 mMexterna1 potassium. And since both treated and control cells had about 350 m M internal potassium, the addition of valinomycin at zero time established a membrane potential of about 180 mV. Figure 4A shows that net synthesis of ATP occurred in control cells and that the inhibitor (DCCD) prevented this response. The tracings in Fig. 4B give the simultaneous measurements of proton entry, estimated from changes in external pH. It is clear that proton entry into control cells proceeded more rapidly than into cells in which the ATPase was blocked. At this point, it is assumed that proton entry into DCCD-treated cells represents passive inflow, at a rate determined only by the size of the electrochemical proton gradient and the properties of the membrane with regard to the balance between passive
10.
H+
ENTRY AND
ATP
181
SYNTHESIS IN BACTERIA
FIG. 4. Proton entry in control and DCCD-treated cells when the imposed electrochemical proton gradient is dominated by a membrane potential. A concentrated stock of washed cells (16 mg dry wt/ml) was exposed to 1 m M DCCD (in ethanol) or a comparable volume of ethanol for 40 minutes at room temperature, using the lightly buffered medium described in Fig. 3. After this pretreatment, cells were centrifuged, resuspended in the same medium (without additives), and analyzed as described in Fig. 3. Portions of the diluted cell suspensions were also analyzed for internal and external potassium. Initial values for the electrochemical proton gradient were 194 and 197 mV for control and DCCD-treated cells, respectively. (From Maloney, 1977.)
2
c
g
= 120-
4
I
6
A
MINUTES
" u
&
40-
+ DCCD , B 2
4
6
MINUTES
movements of H + and OH- ions. Of necessity, then, proton entry in excess of this must be coupled to the performance of work. And since the elevated proton entry in cells containing functional ATPase was also associated with synthesis of ATP, the simplest interpretation of these results is that the proton-translocating ATPase of bacteria catalyzes an obligatory coupling between proton inflow and the formation of ATP. The experiment illustrated in Fig. 5 allows a similar conclusion when a pH gradient dominates the imposed driving force for protons. In this study washed cells were suspended in 200 mM potassium, buffered at pH 8. Sulfuric acid was then added to lower the outside pH to pH 3.5, imposing an initial driving force of about 250 mV due to the chemical potential for H + . The inset shows ATP levels measured for each of the three samples examined. For these same samples, changes in internal pH were used to monitor proton entry, and the larger graph gives these measurements as calculated from the distributions of salicylic acid. The behavior of cells given only the pH jump is indicated by the triangles. In this case, it was expected that the inflow of H + , down the chemical gradient, could generate a membrane potential, positive inside. Other studies suggest a membrane potential of 2 90 mV, positive inside, under these conditions, since the thiocyanate anion can accumulate at least 30-fold after such a pH jump.
182
PETER C. MALONEY
7.0
0
6.5
c L W
c
C 6.0 U 0, c
0 3 0 0
0
5.5
5 .O
2
4
6
8
Minutes FIG.5. Proton entry when the electrochemical proton gradient is dominated by a p H gradient. Washed cells were suspended in 100 mM potassium phosphate, p H 8, at about 0.15 mg dry wt/ml, in the presence of 5 p M [14C]salicylic acid (pK = 3). At zero time a small volume of 2 N sulfuric acid was added to lower the external pH t o 3.5, after which samples were removed to estimate both internal ATP (inset) and accumulation of the weak acid probe of internal pH (Maloney, 1978). Where indicated, valinomycin (10 p M final concentration) was added 5 minutes before the sulfuric acid. For pretreatment with DCCD, the inhibitor was added (1 m M final concentration) to a concentrated stock of cells 60 minutes before analysis. (Maloney and Hansen, 1982.)
Because this would lower the total electrochemical proton gradient, one can understand why these cells showed a relatively slow net acidification and little increase in ATP over basal levels. This “back potential’’ was eliminated in the remaining samples, since the presence of valinomycin allowed compensatory movements of potassium. As indicated by the circles, when the ionophore was also present, there was a marked stimulation of proton entry, as well as net ATP synthesis. Finally, the third sample, shown by the squares, indicates that the effect of DCCD is to reduce both proton entry and ATP formation. Once again, if proton entry into DCCD-treated cells reflects passive events, then the greater rate at which protons move into the control cells shows that as they move across the membrane a substantial
10.
H+
ATP
ENTRY AND
183
SYNTHESIS IN BACTERIA
fraction must enter by a pathway that requires a coupling to the performance of work, as in the synthesis of ATP. The interpretation given these two kinds of experiments (Figs. 4 and 5 ) depends heavily upon assumptions about the properties of DCCD-treated cells, in particular upon the assumption that proton entry into such treated cells represents a passive net inflow of H + down the electrochemical gradient. This assumption is supported by the comparisons given in Table I. This table summarizes the behavior of DCCD-treated cells in the experimental systems just described. For example, the first line gives results from four experiments in which a potassium diffusion potential of about 180 mV dominated the imposed electrochemical proton gradient. In these cases, the initial rate of proton entry corresponded to 0.12 pmole H + sec-l gm-i dry wt of cells for a 60-mV or 1-pH-unit driving force. The second line shows data from nine experiments in which DCCD-treated cells were subjected to pH gradients of varying size in the presence of valinomycin, equivalent to driving forces between 120 and 250 mV. In these cases, initial rates of acidification of the cell, along with estimates of internal buffering power, indicated proton entry at 0.17 pmole H+/second in these same units. These effective conductances come from experiments in which large driving forces were imposed. But the net conductance measurements shown on the last line were obtained in “acid pulse’’ experiments where small driving forces were used. With this technique, passive membrane conductance to H + is measured by analyzing the rate at which protons enter after their equilibrium distribution is perturbed by a small (0.1-pH-unit) deflection of external pH (Mitchell and Moyle, 1967). Moreover, this latter
TABLE I MEMBRANE H + CONDUCTANCE OF Streptococcus lactis
Experimental system
Treatment ~
A$ dominantb ApH dominantc Acid pulsed
~~
H + Conductance (pmoles H + sec-l pH-’ gm-l dry wt)”
~
+ + f
DCCD DCCD DCCD
0.12 0.17 0.20
0.02 0.02 f 0.02 f f
0 H + conductances given as mean values i= SEM. Data are not corrected for the effect of ionic strength on the activity of H + (Keilland, 1937). Such correction yields conductance measurements of 0.10, 0.13, and 0.16 pmole H + sec-l pH-’ gm-’ dry wt for experimental systems 1, 2, and 3 respectively. From data reported in Maloney (1977); see also Fig. 3. From Maloney and Hansen (1982); see also Fig. 5 . From Maloney (1979).
184
PETER
C. MALONEY
method allows one to measure H + conductance in both untreated and DCCD-treated cells. The results of 14 such determinations, performed with or without DCCD, between pH 5 and 8.5, showed constant membrane conductance for the net flux of H + at 0.2 pmole H+/second in these units. It is important to note that all these conductance measurements give the net balance between H + and OH- movements, but that the individual ions can move across membranes at high rates (Nichols and Deamer, 1978). The p H independence of this balance could indicate a reciprocal change in the permeability coefficients for the individual ions. Alternatively, net H and OH- movements may reflect a special mechanism in which an effective “translocation” (rather than transport) occurs by rearrangements of hydrogen bonds within temporary, water filled channels (Nichols and Deamer, 1980). The three different kinds of experiments discussed above (Figs. 4 and 5, Table I) identify the two major routes available for proton entry into S. factis. One pathway, in effect a “leakage” pathway, is distinguished by quantitative arguments-in cells without functional ATPase protons clearly enter after imposed electrical or chemical gradients, but they do so at a rate predictable from the assumption of a passive flow. Other observations suggest that the ATPase itself represents the second major route of proton entry-the influx of protons by this second route must be coupled to the performance of work; such proton entry is paralleled by ATP synthesis; and finally, this alternative pathway is found only in cells containing functional ATPase. +
IV.
STOICHIOMETRY OF THE COUPLING BETWEEN H + AND ATP
This next group of experiments has focused on the stoichiometry of coupling between proton entry and ATP synthesis. Earlier work with S. factis, experiments of the kind shown in Fig. 2, indicated that net synthesis of ATP occurred only when the imposed electrochemical proton gradient exceeded about 180-200 mV (Maloney and Wilson, 1975; Wilson et al., 1976). This is without regard to the relative proportion of electrical and chemical gradients. The same is true of proton entry mediated by the ATPase. For example, the experiment illustrated by Fig. 3 shows that, when a membrane potential dominated :he electrochemical gradient, proton entry in excess of that due to passive events was found only when the initial gradient rose above about 175 mV. A similar result is obtained when a chemical gradient is dominant, as summarized in Fig. 6. Figure 6 gives data collected from experiments in which a pH gradient of varying size was
10.
Ht
ENTRY AND
ATP
185
SYNTHESIS IN BACTERIA
300
I-[
200
I00
0
0
I
2
3
4
Initial Rate of Acidification (pH unitslmin) FIG. 6. Rates of acidification after a p H jump. Cells were treated as described in Fig. 5 . Rates of acidification were calculated from changes in the distribution of salicylic acid during the first 20-90 seconds after a p H jump. The value of the imposed electrochemical proton gradient (ordinate) could be estimated directly from changes in external pH, since H + was distributed at equilibrium before the p H jump. Measured rates of acidification (abscissa) do not reflect proton entry directly, since the internal buffering power varies with the pH (Maloney, 1979). However, when acidification was about 0.5 pH unitdminute or less, measurements could be made over a range of internal p H where the buffering power varied by n o more than 30%. (Maloney and Hansen, 1982.)
established by the addition of sulfuric acid to cells suspended in 200 m M potassium at pH 8 in the presence of valinomycin. The abscissa gives the initial rates of acidification of the cell, determined by the distribution of salicylic acid. The initial size of the electrochemical proton gradient, given by the ordinate, could be calculated directly from the changes in external pH, since protons were distributed at equilibrium before the pH jump. Results obtained with DCCD-treated cells are shown by the open symbols, and for these cells rates of acidification were directly related to the size of the imposed gradient. Moreover, as indicated earlier, this relationship is
186
PETER C. MALONEY
that expected for a passive inflow of protons. The behavior of cells with functional ATPase was more complex, as shown by the solid symbols. In these cases, proton entry was the same as that expected of passive flow at driving forces below about 170-190 mV. However, above this threshold greater rates of proton entry were observed, and the arguments presented earlier suggest that this elevated proton entry is mediated by the ATPase itself. The apparent threshold (170-190 mV) is taken to represent the “reversal potential” for the ATPase under these conditions (see also Schonfeld and Neumann, 1977), and from this assumption it is possible to derive the required stoichiometry between protons and ATP, provided that measurements of the phosphate potential are also available. Data presented in Table I1 give the information needed for derivation of the coupling between protons and ATP. The phosphate potential AG’ATp was calculated by first assuming a standard free energy of ATP hydrolysis of 7.6 kcal/mole. This is equivalent to 330 mV when expressed in electrical units and is appropriate for about pH 7.4 when ATP and ADP participate as their magnesium salts (Guynn and Veech, 1973). To complete the calculation, measured values for internal ATP, ADP, and inorganic phosphate were used. The final calculation shows that the phosphate potential of washed cells was about 8.5 kcal/mole, or 370 mV. Thus, if ATP synthesis were coupled to proton entry, the Gibbs free energy required would become available if one proton moved inward down a gradient of 370 mV. But for a stoichiometry of two protons per ATP, the graTABLE I1 CALCULATION OF THE STOICHIOMETRY OF THE PROTON-TRANSLOCATING ATPASEOF Streptococcus lactis FROM A COMPARISON OF THE REVERSAL POTENTIAL FOR PROTON ENTRY AND THE PHOSPHATE POTENTIAL OF WASHED CELLS Phosphate potentialUA C ~ T /PF
=
A G i T p/ F + (RT/F)(ln[ATPl /[ADP][P,]) 330 mV + 37 mV 367 mV 180 + 1 0 m V n(H+ /ATP)
:
=
Reversal potentialh Stoichiometry
= =
(n)(reversal potential) = phosphate potential (n)(180 f 10) = 367 mV n = 2.0 f 0.1 H + / A T P Phosphate potential calculated assuming a standard free energy of A T P hydrolysis of 7.6 kcal/mole (Guynn and Veech, 1973) and measured values for A T P (0.18 x 10-3M, 0.78 x 10-3M) and inorganic phosphate (50 x 10-3M). Nucleotide contents were measured as described (Maloney, 1977); identical values for inorganic phosphate were obtained using two different assays (Ames, 1966; Martin et a/., 1971). Taken from data shown in Fig. 5 .
10.
Hf
ENTRY AND
ATP SYNTHESIS
187
IN BACTERIA
dient need only be 185 mV. And if the stoichiometry were three or more protons per ATP, then the reversal potential of the ATPase under these conditions would fall to 120 mV or less. Clearly, the measured reversal potential of 180 mV indicates a coupling ratio of two protons per ATP for the ATPase of bacteria. In bacterial chromatophores, ATP formation also shows a stoichiometry of 2 H + / A T P (Petty and Jackson, 1979). However, in other systems where this proton-translocating ATPase is found, there is a dispute over the coupling ratio; some experiments have indicated 2 H + / A T P (Thayer and Hinkle, 1973), while others have suggested 3 H + /ATP (Portis and McCarty, 1976; Brand and Lehninger, 1977). I think it unlikely that the stoichiometry of H + /ATP differs for bacteria, mitochondria, and chloroplasts, and the final resolution of such a controversy over H+/ATP, and the parallel questions concerning other stoichiometries (H /site, P/O ratios, etc.), has important implications for the mechanism of proton pumping and charge separation in oxidative and photosynthetic phosphorylation (see the articles by Wikstrom and by Dutton, et a/., this volume, and a recent summary by Mitchell, 1979). +
V.
RATES OF ATP FORMATION AND THE NATURE OF THE DRIVING FORCE
This final set of experiments attempted to decide whether the ATPase responds equally well, in an overall kinetic sense, t o electrical and chemical gradients of equivalent thermodynamic value. It seems possible to approach this question, for net synthesis of ATP in S. lactis can occur sufficiently slowly so that assays of the rate of its appearance can be made well before the phosphate potential comes into equilibrium with the newly imposed gradient. In exploring this topic, two kinds of studies were performed (Maloney and Schattschneider, 1980). In one, ATP synthesis in valinomycin-treated cells was driven solely by imposed pH gradients. In the other, a p H gradient was also imposed but, in addition, cells were diluted into medium containing choline in partial replacement of the original potassium. Thus the membrane potential was manipulated as well. In both kinds of experiments, samples were taken at 6-second intervals for the first 18-30 seconds t o estimate the initial rates of ATP formation. Data from two of these experiments are summarized in Fig. 7. In this graph, the open circles give data from the experiment in which only the pH gradient was varied. The open and closed triangles illustrate the experiment manipulating both the electrical and chemical gradients. For each of these two studies, the observed rates of appearance of ATP are plotted on the
188
PETER C. MALONEY External pH ( 0,A) 5.0 r
4.5
4.0
3.5 1
1
0
pH grodmnt voried
A A pH gradient ond rnsrnbrone potsntlol varlad
+I
-
0 -
-z 1
-I
160
1
1
200
240
1
1
280
Electrochemical H* Gradient (mV) (0,A)
FIG. 7. Rates of net ATP formation in response to imposed pH gradient or membrane potential. Two separate experiments are described. In both instances, valinomycin-treated cells were initially suspended (15 mg dry wt/ml) in 100 mMpotassium phosphate, p H 8 (internal pH = 7 . 6 ) . In the experiment illustrated by the circles ( O ) , cells were diluted 20-fold with the same buffer before the addition of graded amounts of sulfuric acid lowered the external pH to the values shown by the upper abscissa. In the experiment described by the triangles .(A A),cells were diluted into phosphate buffer in which potassium was partially replaced by choline; 10 seconds later graded amounts of sulfuric acid were added t o lower the external p H t o the levels indicated by the upper abscissa. After addition of the acid, samples were removed at 6-second intervals for assays of intracellular ATP. Open triangles (A) and solid triangles (A)represent the same data, plotted against pH, and against the total electrochemical H + gradient, respectively. (From Maloney and Schattschneider, 1980.)
ordinate using a logarithmic scale. These rates were then treated in one of two ways. With the upper abscissa, rates could be expressed as a function of decreasing external pH (increasing external proton concentration). With the lower abscissa, the effect of the membrane potential was taken into account by expressing the rates as a function of the total driving force. These data suggest several conclusions. The first is that there is no necessary correlation between the rate of ATP synthesis and the external concentration of H + ,even though H + may be considered one of the reactants in the overall process. This is shown by a comparison of the open and solid triangles. Thus, when external pH is con-
10.
Hf
ENTRY AND
ATP
SYNTHESIS IN BACTERIA
189
stant, but the membrane potential increasing, the vertical line AB is observed. Alternatively, when H + levels outside are increasing, but the membrane potential decreasing, the horizontal line BC is found. The necessary correlation, shown by line AC, is not obtained unless variations in both the electrical and chemical gradients are known. Thus one cannot arrive at a sensible statement regarding the rate of ATP synthesis unless the electrochemical activity of the proton is described explicitly. In principle, these studies might also give information about the stoichiometry of the ATPase. If the stoichiometry is 2 protons per ATP, then one might expect rates of synthesis to increase in proportion to the square of the effective proton concentration. This would predict a 100-fold increase in rate for every 60 mV increase in the electrochemical gradient. In fact, for both of the experiments shown here, the slope of line AC suggests a coupling ratio of 1.7-1.8 protons per ATP, in good agreement with the earlier calculation of stoichiometry (Table 11). However, this second estimate must be considered tentative, for it assumes that nonspecific ATP hydrolysis, if occurring at significant rates within the cell, will bias the measured rates by a constant factor. This assumption has not been tested. Finally, these data suggest that identical rates of ATP synthesis are found when equivalent chemical or electrical gradients are imposed, at least within the limited range so far explored. This is shown by a comparison between the open circles and the solid triangles. In the case shown by the circles, only the pH gradient was varied, but in the second case, shown by the solid triangles, the span between 180 and 225 mV arose from an imposed membrane potential (increasing in value from about - 14 to - 59 mV). Thus it appears that models for energy coupling must be able to account for such quantitative interconversion of these two different driving forces. In concluding this discussion, it is of interest to restate one speculation that accommodates such interconversions, a mechanism outlined by Mitchell (Mitchell, 1969). In a simple way, the diagram at the top of Fig. 8 shows some of what is known from biochemical studies of this ATPase in bacteria, mitochondria, and chloroplasts (see Kagawa, 1978, for a recent review). The enzyme is divisible into two major parts. One of these, the F, sector, displays ATPase activity when removed from the membrane. But the F,, sector, which remains embedded within the membrane, does not catalyze chemical transformation. Instead, it behaves as a proton carrier or proton channel. However, proton movements mediated by F, are seen readily only when F, is removed. Thus, in the complete enzyme, this proton channel is (in effect) plugged at one end and converted into what has been termed a “proton well” (Mitchell, 1969). One presumes that at some point within this well specific chemical groups interact with incoming protons to initiate ATP
190
PETER C . MALONEY
I
1
0
0.5
I 3
FRACTIONAL DISTANCE ( X I WITHIN Fo
FIG.8 . The effect of membrane potential on the concentration of protons within a proton well. Upward travel on the vertical axis indicates increasing H + concentration within the proton well, expressed using a logarithmic scale (decreasing pH) for convenience. The membrane potential (A*) is taken as a positive quantity. See text and Mitchell (1969) for further explanation.
synthesis and to reverse the ATPase activity associated with isolated F,. This is presumed since no other intermediates in the coupling process have been identified for this ATPase, in sharp contrast to the other ATPases discussed in this volume, in which phosphorylated protein serves as the link between ion movement and the primary chemical transformation. In this context, then, it is appropriate to consider the factors that influence the concentration of protons with the local compartment represented by F,. The lower half of Fig. 8 shows an idealized plot of the effective proton concentration within the proton well (F,) as a function of some distance x from the outer surface. The relationships shown are appropriate for simple
10.
H+
ENTRY AND
ATP
SYNTHESIS IN BACTERIA
191
cases in which one assumes a constant electric field spanning the full thickness of the membrane. The two examples, A and B, are ones in which the same total electrochemical gradient has been assumed. But in case A the membrane potential is taken as zero. Consequently, in this first instance, the concentration of protons is the same at any point within the well. However, in case B, where a membrane potential is assumed, the H + concentration within the well depends upon distance x. In case B, it is only when the full thickness of the membrane has been crossed that the proton concentration within the well is the same as in case A. Thus this idea allows one to understand how an electric gradient can be converted to a chemical signal, and why increasing the membrane potential might have the same kinetic effect as increasing the outside proton concentration (Fig. 7). In two instances such an analysis has proven of interest. In one case a channel was studied. From a more sophisticated treatment of this general problem and starting with a slightly different point of view, Woodhull (1973) concluded that the sodium channel of frog nerve contains a ratelimiting proton-binding site located about one-quarter of the way along the electric field. In the other example, a carrier was examined. Schwab and Komor (1978) have studied the kinetics of H+-hexose cotransport in Chlorella vulgaris. Their results suggest that an H+-reactivesite lies about halfway through the plasma membrane of this alga. It is not yet clear whether this simple analysis is appropriate in interpretating the data presented in Fig. 7. At the very least, the observed equivalence between the electrical and chemical gradients as rate-determining elements in ATP synthesis supports the idea that F, plays a permissive rather than instructive role in the overall coupling process. Thus it appears that the rate of ATP formation is determined by the rate at which protons are made available to the interface between F, and F,. If, in addition, one takes the view that the effect of the membrane potential should be understood in terms of interactions between H + and specific chemical groups, then these data also support the idea that energy coupling is rate-limited by reactions that occur at a point where the electric field has fallen to zero. In the simple case where a constant electric field spans the full thickness of the membrane, this reasoning suggests that energy coupling takes place at the inner surface of the cell membrane.
VI.
CONCLUSIONS
The experiments summarized here suggest the following conclusions with regard t o the relationship between proton movement and ATP synthesis in bacteria. Under suitable conditions, protons moving inward by
192
PETER C. MALONEY
way of the proton-translocating ATPase may be identified, and this pathway represents one major route by which protons enter the cell. When protons move inward in this way, a coupling to ATP synthesis occurs with an average stoichiometry of two protons per ATP. The rate at which ATP synthesis occurs is determined by the size of the gradient down which protons fall, but no sensible correlation exists between rates and the absolute value of the membrane potential, the pH gradient, or the concentration of (external) protons. Rather, the electrochemical activity of the proton must be described before realistic appraisals can begin. Finally, electrical and chemical gradients of equal thermodynamic value elicit the same rate of net synthesis of ATP. In any such coupled reaction, the final equilibrium attained must be the same for equivalent electrical and chemical gradients. However, the rate at which equilibrium is approached need not be the same in the two instances, and the expected correlation between rate and driving force depends upon the model chosen to interpret available data. A central problem in energy coupling is to decide how membrane proteins make chemical sense out of an electric field. One solution, proposed by Mitchell (1969), is to assign the initial stages of the process to a portion of the enzyme that acts as an ion (H+)well. In this way, when there is little pH differential across the membrane, but a substantial electric field, the chemical activity of H + within a proton well may be elevated so that H + might be at sufficiently high concentration to participate directly in the reaction (Mitchell, 1974). This idea of a proton well seems to be useful in analyzing other systems, both channels and carriers, and thus may be an appropriate base for more detailed studies of the coupling between proton movements and ATP synthesis. ACKNOWLEDGMENT Work described in this article has been supported by a grant from the Public Health Service (GM 24195).
REFERENCES Ames, B. N. (1966). Methods Enzymol. 8, 115-118. Brand, M. D., and Lehingner, A. L. (1977). Proc. Natl. Acad. Sci. U.S.A. 14, 1955-1959. Collins, S. H., and Hamilton, W . A. (1976). J. Bacteriol. 126, 1224-1231. Fillingame, R. H. (1976). J. B i d . Chem. 251, 6630-6637. Guynn, R. W., and Veech, R. L. (1973). J . Biol. Chem. 248, 6966-6972. Harold, F. M. (1977). Curr. Top. Bioenerget. 6, 83-149. Harold, F. M., and Papineau, D. (1972). J . Membr. Biol. 8, 27-44. Harold, F. M., Baarda, J . R., Baron, C . , and Abrams, A. (1969). J . Biol. Chem. 244, 2261-2268.
10.
H+
ENTRY AND
ATP
SYNTHESIS IN BACTERIA
193
Harold, F. M., Pavlasova, E., and Baarda, J. R. (1970). Biochim. Biophys. Acta 196, 235-244. Kagawa, Y . (1978). Biochim. Biophys. Acta 505, 45-93. Keilland, J. (1937). J . A m . Chem. SOC. 59, 1675-1678. Maloney, P. C. (1977). J. Bacteriol. 132, 564-575. Maloney, P. C. (1978). Biochem. Biophys. Res. Commun. 83, 1496-1501. Maloney, P. C. (1979). J . Bacteriol. 140, 197-205. Maloney, P. C., and Hansen, F. C., 111 (1982). J. Membr. Biol. (in press). Maloney, P. C., and Schattschneider, S. (1980). FEBS Lett. 110, 337-340. Maloney, P. C., and Wilson, T. H. (1975). J. Membr. Biol. 25, 285-310. Maloney, P. C., Kashket, E. R., and Wilson, T. H. (1974). Proc. Natl. Acad. Sci. U.S.A. 71, 3896-3900. Martin, R. G . , Berberich, M. A,, Ames, B. N., Davis, W. W., Goldberger, R. F., and Yourno, J. D. (1971). Methods Enzymol. 178, 3-39. Mitchell, P. (1969). Theor. Exp. Biophys. 2, 159-216. Mitchell, P. (1974). FEBS Lett. 43, 189-194. Mitchell, P. (1979). Eur. J . Biochem. 95, 1-20. Mitchell, P., and Moyle, J. (1967). Biochem. J . 104, 588-600. Nichols, J. W . , and Deamer, D. W. (1978). In “Frontiers of Biological Energetics” (L. P. Dutton, J. S. Leigh, and A. Scarpa, eds.), Vol. 2, pp. 1273-1283. Academic Press, New York. Nichols, J. W., and Deamer, D. W. (1980). Proc. Natl. Acad. Sci. U.S.A. 77, 2038-2042. Petty, K. M., and Jackson, J. B. (1979). FEBS Lett. 97, 367-372. Portis, A. R., Jr., and McCarty, R. E. (1976). J. Biol. Chem. 251, 1610-1617. Ramos, S., and Kaback, H. R. (1977). Biochemistry 16, 848-854. Rosen, B. P., and Kashket, E. R. (1978). In “Bacterial Transport” (B. P. Rosen, ed.), pp. 559-620. Dekker, New York. Schonfeld, M., and Neumann, J. (1977). FEBS Lett. 73, 51-54. Schwab, W. G . W., and Komor, E. (1978). FEBS Lett. 87, 157-160. Thayer, W. S., and Hinkle, P. C. (1973). J. Biol. Chem. 248, 5395-5402. Wilson, D. M., Alderete, J. F., Maloney, P. C., and Wilson, T. H. (1976). J. Bacteriol. 126, 327-337. Woodhull, A. M. (1973). J. Gen. Physiol. 61, 687-708. Zilberstein, D., Schuldiner, S., and Padan, E. (1979). Biochemistry 18, 669-673.
This Page Intentionally Left Blank
CURRENT TOPICS I N MEMBRANES AND TRANSPORT, VOLUME 16
Chapter I I
Net ATP Synthesis by H+ATPase Reconstituted into Liposomes Y A W 0 KAGA WA Department of Biochemistry Jichi Medical School Minamikawachirnachi, Tochigi-ken, Japan
Introduction ........................................................................................ Electrogenic Properties of H -ATPase ...................................................... A. Vectorial H + -ATPase Reaction ......................................................... B. Measurement of ApH ...................................................................... C. Measurement of All. ..... ....................................................... D. Steady State Level O f AjiH + and the H +/ATP Ratio ............................... 111. Net ATP Synthesis Driven by ApH+ ..... A. Ion Gradients Applied to H -ATPase Liposomes .................................. B. Electric Fields Applied to H +-ATPase Liposomes ......................... IV. Molecular Properties of H + -ATPase ......................................................... A. H + Pump and H + Gate Activity of Crystalline ATPase F, ........................ B. H + Channel and H + Filter: Chemical Structure of Fo V. Epilogue ... ............................................................. References ..........................................................................................
I.
11.
+
+
1.
195 197 197 197 198 200 20 1 20 1 202 207 207 210 21 1 212
INTRODUCTION1
According to the chemiosmotic theory, H -ATPase couples the flow of chemical energy of ATP to the translocation of protons from one side of a +
I Abbreviations used in this chapter: ANS, 1-anilinonaphthalene 8-sulfonate; DCCD, N, N'-dicyclohexylcarbodiimide;AgH+ , electrochemical potential difference of protons; ApH, pH difference across the membrane; A$, membrane potential; F,, catalytic portion of Fo-F,; Fo, proton channel portion of Fo-Fl; Fo-pFI,H -ATPase; proton-translocating adenosinetriphosphatase; FCCP, carbonylcyanide p-trifluoromethoxyphenylhydrazone;TF,, thermophilic Fl; TFo, thermophilic Fo; TFo-Fl, thermophilic H + -ATPase. +
195
Copyright v 1982 by Academic Press, Inc All rights of reproduction in any form reserved ISBN 0-12-153316-6
196
YASUO KAGAWA
membrane to the other (Mitchell, 1976). In fact, H+-ATPase was extracted from the inner membrane of mitochondria (Kagawa, 1967; Kagawa and Racker, 1966b,c) and shown to translocate protons in reconstituted H +-ATPase liposomes (Kagawa, 1972: Kagawa and Racker, 1971). H -ATPase is universally distributed, since it is the most fundamental energy-transforming machinery of living organisms. There are excellent reviews on this enzyme in mitochondria (Penefsky, 1979), chloroplasts (McCarty, 1979), and prokaryotic plasma membranes (Downie et al., 1979). On the other hand, net ATP synthesis was demonstrated in chloroplasts by applying a H + gradient (Jagendorf and Uribe, 1966) or electric field (Witt et al., 1976). Reconstituted systems containing crude H -ATPase and electron transport complexes (or bacterial rhodopsin) were shown to synthesize ATP by respiration (or illumination) (Racker, 1976). However, ATP synthesis by these crude systems is not suitable for studying the molecular mechanism of energy conversion, since any manipulation of these complex systems may have, in addition to direct effects, indirect effects through the electron transport system and through contaminating proteins. As reviewed by Boyer et al., (1977), conformational changes in some proteins may be transferred, by direct contact, to H+-ATPase; or local electrochemical changes near electron transport components may be transmitted directly to H+-ATPase, not through the difference in electrochemical potential for protons (A@,+) between the two bulk water phases across the membrane. For studies on the niolecular mechanism of energy conversion, it is thus necessary to synthesize ATP using purified H -ATPase in liposomes, by imposing an ion gradient or electric field, and to analyze the roles of the components of the ATPase. Preparations of H+-ATPase from mitochondria are unstable and impure even when obtained with recently developed methods. For example, H -ATPase of mitochondria prepared by different methods gives 13 bands on gel electrophoresis (Stiggall et al., 1978), but it is still uncertain whether all these components are essential or whether some are contaminants. From membranes of the thermophilic bacterium we have prepared H -ATPase that is pure and stable and contains only eight subunits (Sone et al., 1975; Kagawa and Sone, 1979). As described in later sections, the H +-ATPase has been reconstituted from its component subunits, and the functions of the eight subunits have been determined (Kagawa, 1978, 1980). This thermophilic H -ATPase synthesizes ATP in reconstituted proteoliposomes when ion gradients (Sone et al., 1977; Kagawa et al., 1977) and an external electric field (Rogner et al., 1979) are imposed. Similar H -ATPases, giving eight bands, have recently been obtained from prokar+
+
+
+
+
+
+
11. NET
ATP SYNTHESIS BY H+-ATPase
197
yotic cells (Foster and Fillingame, 1979; Babakov and Vasilov, 1979), but reconstitution studies on the subunits are not yet complete. The purpose of this article is to describe findings on the molecular mechanism of the electrogenic H + pump reconstituted from the pure, stable H -ATPase of thermophilic bacteria. +
II.
ELECTROGENIC PROPERTIES OF H+-ATPase
A. Vectorial H +-ATPase Reaction The electrogenic properties of isolated H -ATPase have been studied by incorporating this enzyme into lipid bilayers (Kagawa, 1972, 1978; Kozlov and Skulachev, 1977). The ATPase reaction is formulated as +
ATP
+ HZO + xH:=
ADP
+ Pi + x H ~+ y H +
(1)
where H,' and H: are protons outside and inside the reconstituted liposomes, x is the number of protons translocated per 1 mole of ATP hydrolyzed, and y is the number of protons liberated by the difference in Pi. In the early stages of chemiosmotic theory, y pK, of ATP and ADP was confused with x , but y is the scalar component and x is the vectorial component of this membrane reaction. The value of y is about 1 at pH 8, but it becomes 0 at pH 6.25 in the presence of Mg2+.At the latter pH, net proton uptake (xH +) driven by ATP hydrolysis is observed in reconstituted H+-ATPase liposomes (Kagawa, 1972). On the other hand, if the membrane structure is destroyed or an uncoupler (H+ carrier) is added at neutral pH, ATP hydrolysis causes a pH change by y H , and x becomes zero. Transport of protons across the membrane produces ApH+,which is related to the membrane potential (All.) and pH difference (ApH) by
+
+
APH+ = FA$ - 2.3RTApH
(2)
in which R and T are the gas constant and absolute temperature, respectively (Mitchell, 1966). B. Measurement of ApH When protons are accumulated in the liposomes, fluorescent amines, such as 9-aminoacridine (9AA), are also concentrated by protonation in the liposomes (Fig. 1A). The fluorescence (excitation at 365 nm and emission
198
YASUO KAGAWA
ANS-
A
B
FIG. 1 . Vectorial H+-ATPase reaction in liposomes. (A) Measurement of ApH by fluorescence quenching of 9AA accumulated in the liposome. (B) Measurement of A$ by enhancement of the fluorescence of ANS attracted to the liposome.
at 451 nm) of the concentrated 9AA is quenched by the mutual interaction of 9AA. Thus ApH is calculated by (Rottenberg, 1975) ApH
=
logQ/(l-Q)
+ log 1/V
(3)
where Q is the fraction of the total fluorescence that is quenched in response to H + transport and V is the volume of the osmotic compartment as a fraction of the total volume of the assay mixture (Kagawa and Sone, 1979). [When permeant anions such as NO, are added (Fig. lA), the A$ component of ApH+ is converted to ApH (as long as H + is still pumped in by H+-ATPase) until the total ApH+ reaches about 200 mV. Thus permeant anions cause quenching of 9AA.l The internal pH of the liposomes can also be measured by enclosing a hydrophilic pH indicator, such as chlorophenol red, in the lumen during liposome reconstitution (Sone et al., 1976).
C. Measurement of A$ Two separate methods have also been used for estimating A$ in liposomes. In the first, the fluorescent anionic dye, 8-anilinonaphthalene-lsulfonate (ANS) is added to the suspension and binds to the liposomes. In response to a membrane potential positive inside the liposomes, the dye fluorescence ( F )is enhanced (excitation at 365 nm and emission at 480 nm). In control experiments the relative fluorescence enhancement (AF/F) has been found to be proportional to the diffusion potential of potassium
11. NET
ATP
SYNTHESIS BY
H+-ATPase
199
established across the membrane (Sone et af., 1976), so the enhancement can be converted to A$ by calibration via valinomycin-mediated K + diffusion from the liposomes using the Nernst equation: A$
=
( R T / F ) In ([K+l,/[K+li)
(4)
where [K+], and [K+], are the activities of K + inside and outside the liposomes, respectively. [When the ApH component of A&+ is converted to A$ by the addition of a permeant weak base (e.g., Tris), A F / F increases (Fig. l.)] Under steady state conditions certain artificial lipid-soluble ions appear to be distributed passively across the liposomal membranes. Disappearance of such ions from the incubation medium, upon the addition of liposomes, can therefore give a measure of A$ for the liposomes, which is particularly convenient because the free concentration of lipid-soluble ions can be assayed continuously with specific electrodes (Muratsugu et af., 1977). A$ is actually calculated by a modification of the Nernst equation, provided the total liposome volume and the electrode potentials before and after the addition of liposomes are known. This method has been used with tetraphenylphosphonium ion in our recent experiments (Fig. 2).
5
FIG.2. Electrode for the measurement of A$ with tetraphenylphosphoniurn ion (TPP' ). 1, KCI-agar bridge; 2, TPP solution (10 mM); 3, TPP-permeable poly(viny1 chloride) mem-
brane; 4, HgCl reference electrode; 5 , electrometer and recorder; 6, liposome; 7, stirring bar.
200
YASUO KAGAWA
D. Steady State Level of ApH+ and the H+/ATP Ratio The reversible H -ATPase reaction in reconstituted liposomes [Eq. (l)] will reach chemical equilibrium if there is no H + leakage. Calculations by means of Eqs. (1) and (2) indicate that a ApH+of 204 mV is necessary if x is 2 and the reaction is carried out at an ADP/ATP ratio of 50 in 2 mM Pi (assuming that the standard free energy change of ATP hydrolysis is - 8.0 kcal/mole at pH 8.0). However, in actual experiments, the steady state level of ApH+ is not determined by the chemical equilibrium but by the balance between the H + influx through the H+-ATPase and the H + efflux through the lipid bilayer and H+-ATPase. For a detailed theoretical treatment of this problem, irreversible thermodynamics is useful (for review, see Stucki, 1978). When ATP is added to H+-ATPase liposomes, the steady state level of ApH+ is reached within about 2 minutes. Experimental values for A&+ obtained by the methods described in Fig. 1 are summarized in Table I: The first line corresponds to Fig. lA, the second line to Fig. lB, and the third line to Fig. 1A without NOj (Sone et al., 1976). The maximum steady state value of A&+ is about 250 mV in the absence of permeant ions and buffer [which of course lower the total ApH+(Table I)]. The H + influx is decreased by reduction of the ATP concentration below the K, (0.3 mM) of H+-ATPaseand by the addition of an ATPase inhibitor such as DCCD. The H + efflux by leakage through the lipid bilayer itself increases in proportion to ApH+ , according to Fick's law. In the presence of ADP and Pi, H + efflux by reversal of the H -ATPase takes place and can be evaluated by the Pi-ATP exchange reaction (Kagawa and Racker, 1971). Under conditions of low leakage, the steady level of A&+ should approximate the reversal potential for the ATPase. The observed value of 250 mV supports the conclusion that x in Eq. (1) must be 2. The following chemical explanation for a +
+
ESTIMATIONS OF AFH + , A$,
Tricine plus NOj Tris Tri ci ne
102 48
145 70
TABLE I AND ApH OF H -ATPAsE LIPOSOMES' +
2.45 0 0.67
3.5 0
2.9
22 1 145 253
Liposomes were reconstituted from phospholipids of PS3 containing 0.25 mg of H+-ATPase protein as already described (Sone et al., 1976). The V value in Eq. (3) was 0.83 pl/ml, and the amount of ATP added was 0.5 pmole.
11. NET
ATP
SYNTHESIS BY
H+-ATPase
201
stoichiometry of 2 H + / A T P was given by Mitchell (1976): 2 H + attack the oxygen of PO, in a complex with ADPO- and Mg2+ at the active site of H+-ATPase. H 2 0 is released, and P+O, remains while ADPO- makes a nucleophilic attack on the P + center, thereby producing ATP. Some experiments, however, have suggested a ratio of 3 (Brand and Lehninger, 1977) rather than 2, and the experiments on intact mitochondria (Mitchell, 1976) can be criticized as being complicated by anitport, symport, ionic leakage (for ATP, ADP, substrates, Pi, K + , Ca2+,and Na+), and many accompanying reactions of the endogenous substrates, catalyzed by matrix enzymes. For this reason, the ratio of H + transported to ATP split by purified H -ATPase liposomes should give a cleaner, less equivocal result. Experiments were executed on K -loaded liposomes simultaneously given valinomycin and [y-3ZP]ATP,and a H + / A T P ratio very close to 2 was obtained (Y. Kagawa, unpublished). +
+
111.
NET ATP SYNTHESIS DRIVEN BY ApH+
A. Ion Gradients Applied to H +-ATPase Liposomes Theoretically, net ATP synthesis driven by an artificially imposed ion gradient in H -ATPase liposomes is the reverse reaction of ATP-driven H translocation. As described in the previous section, a A&,+ value of 204 mV is required to synthesize ATP [from Eq. (2)]. Therefore, Eq. (2) gives a value of ApH 3.2 if A$ is 0. Jagendorf and Uribe (1966) were in fact able to demonstrate ApH-driven ATP synthesis in chloroplast vesicles which had been prepared (loaded) at pH 4.5 and were then shifted to alkaline media at pH values greater than 7.7. However, most liposome preparations become too leaky to synthesize ATP under these drastic conditions unless additional energy is continuously supplied, as from ATP hydrolysis (Pi-ATP exchange, Kagawa and Racker, 1971), electron transport, or illumination of added bacteriorhodopsin (Racker, 1976). Fortunately, liposomes made from the saturated phospholipids of thermophilic bacteria (PS3) are less fragile and are suitable for the pH jump experiment. These phospholipids are composed of phosphatidylethanolamine, phosphatidylglycerol, and cardiolipin, and their molecular species are mainly of the 1,15-methylhexadecanoy1-2, 13-methyltetradecanoyl-sn-glycerol-3-phosphoryl type and the 1,2-di-13methyltetradecanoyl-sn-glycerol-3-phosphoryltype (Kagawa and Ariga, 1977; Kagawa, 1980). Experiments are carried out by first incubating the H -ATPase liposomes with acidic malonate buffer and valinomycin, and then shifting to alkaline solution containing the (impermeable) buffer glycylglycine (Table 11) +
+
+
202
YASUO KAGAWA
TABLE I1 REVERSAL OF THE H + PUMPBY AN ARTIFICIALLY IMPOSEDIONGRADIENT" ATP formed (nmoles/mg H -ATPase)
Conditions
+
Complete system (PH 5.5-8.3) Complete system (PH 8.0-8.3) Without ADP With DCCD (inhibitor) With FCCP (H+ carrier)
53.3 5.4 3.9 3.6 1.9 ~
~
~~~
The reconstituted liposomes (0.055 mg of H+-ATPase, 4 mg of PS3 phospholipids) were first incubated in acidic medium (pH 5 . 5 , final volume 0.25 ml) containing 10 pmoles malonate, 1 pmole ADP, and 0.1 pg valinomycin at 40°C for 10 minutes. Then, 0.25 ml of an alkaline medium consisting of glycylglycine (40 pmoles, pH 8.5), 75 pmoles MgSO,, 2 pmoles [32P]phosphate (6 x 10' cpm, sodium salt), 25 pmoles glucose, and 10 units hexokinase was added. The final pH was 8.3. The reaction was carried out at 40°C and terminated after 5 minutes. The experimental details are described by Kagawa and Sone (1979). a
(Kagawa et al., 1977; Sone et al., 1977). This instantaneous transition should create a AFH+ equal to 275 mV and composed about equally of ApH (2.38 units acidic inside) and A$ (125 mV positive inside) across the liposome membrane. ATP synthesis occurs at a velocity of 650 nmoledmg H -ATPase min-I, which is faster than substrate oxidation in mitochondria. The primary role of H translocation in oxidative phosphorylation has thus been confirmed (Kagawa el al., 1977). The maximal level of ATP synthesis is about 100 nmoles/mg H + ATPase in the reconstituted liposomes, whereas less than 2.5 nmoles/mg protein is synthesized in submitochondrial particles (Thayer and Hinkle, 1975) and bacterial membranes (Tsuchiya and Rosen, 1976). Figure 3 shows the effects of different pH values in the acid stage and in the base stage. The decreases in yield of ATP at pH values below 5.5 (Fig. 3A) and above 8.5 (Fig. 3B) are due to inactivation of the H+-ATPase. In these experiments the optimal KCI concentration was 0.15 M at a ApH of 2.8 units. T o synthesize ATP, a A&,+ of 200 mV was necessary, irrespective of its components, A$ and ApH, which were changed artificially (Sone et al., 1977). +
+
B. Electric Fields Applied to H -ATPase Liposomes +
In studies on the mechanism of ATP synthesis by H+-ATPase, rapid energization of H+-ATPase is necessary. Analysis of the reaction by
11. NET
ATP
SYNTHESIS BY
H+-ATPase
203
+-a 100-400 I
E" + I
*a0
50-200
1
0
L
.- .-C 2
d
P
;j ij
0;
5
7
6
8
6
7
aOs.
Acid Stoge pH
8 stag*
9
1
0
pH
FIG. 3. pH-driven reversal of the H+-ATPase in liposomes. (A) Dependence of ATP yield on the acid stage pH. The base stage pH was 8.0. (B) Dependence of ATP yield on the base stage pH. The acid stage pH was 5 . 5 . Experiments A and B were both carried out in the base stage at 40°C for 5 minutes with reconstituted Hf-ATPase liposomes (34 pg of H+-ATPase in 2 mg of PS3 phospholipids) in a final volume of 2 ml containing 0.15 M KCI, 2 M A D P , 10 mM "Pi, 1 m M MgSO,, 0.2 pg valinomycin, and 80 mM glycylglycine buffer (impermeable) at the indicated pH.
acid-base treatment, as described in the previous section, has the disadvantage that time resolution is poor and the energy components are complex, i.e., both A$ and ApH. An alternate method is to drive ATP synthesis by an external electric field. This method has been used for chloroplast particles (Witt et d.,1976; Witt, 1979) and has given excellent time resolution. The method can be applied to H+-ATPase liposomes if a sufficient membrane potential, about 200 mV, is attained. For this purpose, either the electric field strength or the radius of the reconstituted liposomes must be large. The membrane potential (A$) can be calculated via the Laplace equation for a suspended spherical droplet if the following information is known: X , the applied electric field; A , , the conductivity of the medium in which the droplet is suspended; A2, the conductivity of the droplet interior; a , the radius of the droplet; r, the distance from the droplet center to any point of measurement; and 0, the direction normal for a region of the droplet surface calculated with respect to the direction of the electric field. Then, letting be the local electric potential,
'-
2
i a (sin 8 ad - 1 2 ( , 2 ? @ ) +-) = 0
ar ar r2sin e ae ae from which the potential in the medium just outside the droplet surface which faces the electrode (surface perpendicular to the electric field) can be r2
204
YASUO KAGAWA
calculated as
The corresponding potential inside the droplet is
4. = - (
2 + A2/A,
)Xrcose
Since A, for phospholipid is low (A, %-A2),the potential at the surface facing the electrode is X a (cos 8 = 1). Now if the droplet is replaced by a liposome with the medium (A,) in its lumen, there is practically no potential drop across the lumen. This means that the magnitude of potential difference (membrane potential) across the two surfaces facing the electrodes will be
=+
A$
= *Xa
(6)
Therefore, to obtain a A$ value of 200 mV with an electric field strength of 1000 V/cm ( = X ) , the radii of the H+-ATPase macroliposomes should be about 2.7 pm. The electric field strength cannot be raised much above 1000 V/cm when a uniform potential gradient between two parallel electrodes is used, as shown in Fig. 4, because of the production of Joule heat by the current, even at low electrolyte concentrations. [A block diagram of the electric circuit in these experiments is shown in Fig. 5; the rectangular voltage pulses (200-V amplitude and 20-msec dura-
Platinum plate 0.51nm thick 1.3 cm2 radius
Teflon ring U
Plastic discs
FIG.4. Diagram of the plastic cell containing platinum electrodes used to apply electric pulses to reconstituted Hf-ATPase liposomes.
1 1 . NET
ATP
SYNTHESIS BY
H+-ATPase
205 Highvoltage power
20 msec
H
Pulse \-
generator
Switching circuit
I
0
Pt elect rode
FIG.5 . Block diagram of the electric circuit used to apply pulses to the electrode arrangement in Fig. 4.
tion) are delivered by a transistor NAND gate generator (Rogner et al., 1979).] Macroliposomes are reconstituted by the dialysis method (Kagawa and Sone, 1979) in the presence of a Sephasorb suspension (50 mg/ml). The dialyzed mixture is centrifuged at 10,000 g for 10 minutes, and the pellet, suspended in the dialyzing solution, is used for macroliposomes. In order to demonstrate voltage-driven ATP synthesis, it has proven necessary to include an ATP-trapping system, such as glucose hexokinase, with the liposome suspension. Presumably, this is because ATP synthesis at the liposomal membrane facing one electrode is offset by ATP hydrolysis on the opposite face of each liposome. The circumstance contrasts with that in chloroplasts (Witt et al., 1976; Graber, this volume), where ATPase activity does not interfere. ATP synthesis by macroliposomes in this system increased with the number of pulses applied (Fig. 6 ) , but not in exact proportion to the number of pulses. This again differs from the result with chloroplasts, perhaps because the liposomes are somewhat unstable. When trains of very short pulses (2.5 msec) are used, ATP synthesis diminishes significantly (to 40-50%; see Table 111) by comparison with that observed from longer pulses of equal total energy. The reason for this diminution is not clear but may involve the fact that the average turnover period for the enzyme should be substantially longer than each pulse. As summarized in Table 111, ATP synthesis is inhibited by both DCCD and FCCP. H+-ATPase reconstituted in microliposomes has not shown ATP synthesis under conditions comparable to those used for macroliposomes, probably because an insufficient membrane potential is induced.
206
YASUO KAGAWA
5001
n
-t I
n
D
a L
I /
Micro I iposomes
4 Number of Pulse
B
12
(20
16
msec 200 V )
FIG. 6. Voltage-driven reversal of the H' -ATPase in reconstituted liposomes. The assay mixture (0.78 ml) in the electrical cell (Fig. 4) consisted of 400 pl of a macro- or microliposome suspension, 1 mM Pi (32Pi, 4 pCi/ml), 20 units of hexokinase (Sigma, lyophilized), 25 mM glucose, and 0.5 m M A D P (see ROgner et al., 1979).
Since the amount of 32Pesterified was small in H -ATPase liposomes, rat liver mitochondria and submitochondrial particles were irradiated with external electric pulses (760 V/cm, 30 msec, rectangular). The net [32P]ATP synthesized was increased by increasing the number, voltage, and duration (0.05 nmolelmg protein/pulse) and decreased by the specific H -ATPase +
+
TABLE I11 EFFECTSOF VESICLE SIZE,PULSE PARAMETERS, AND INHIBITORS ON NET ATP SYNTHESIS IN RECONSTITUTED H +-ATPAsELIPOSOMES
Liposomes Macroliposomes Macroliposomes Macroliposomes Macroliposomes Macroliposomes Microliposomes '2.5 msec x 8 x 8.
Inhibitor
+ 0.6 pmole DCCD +
20 pg FCCP
Number of 20-msec pulses
ATP synthesized (pmoles)
12 12 12 8 80 8
139 38 33 3 96 178 0
1 1 . NET
ATP
SYNTHESIS BY
H+-ATPase
207
inhibitor, aurovertin, and very high concentration of uncouplers (16 pg FCCP/mg protein) (T. Hamamoto, K. Ohno, and Y. Kagawa, in preparation).
IV.
MOLECULAR PROPERTIES OF H+-ATPase
A. H + Pump and H + Gate Activity of Crystalline ATPase F, H+-ATPase is readily split into a catalytic portion (F,) (Pullman et al., 1960) and a H + channel portion (F,), and when F, is recombined with F,, it becomes sensitive to energy transfer inhibitors (Kagawa and Racker, 1966 a,b) that block H + leakage through Fo. As shown in Fig. 1 , F, is a particle of 90-A diameter attached to the surface of the H+-translocating membrane, while F, plugs through the membrane (Kagawa and Racker, 1966~). F, is both a H + pump (i.e., an energy transformer between ATP and Ap,+) and a H + gate (i.e., a valve blocking H + leakage through the H + channel (Kagawa, 1978). F, has an aggregate molecular weight of about 380,000 (Yoshida et al., 1979) and is composed of five subunits in all ATPsynthesizing membranes: mitochondria, chloroplasts, chromatophores, and bacterial plasma membranes. While there are many complicated hypotheses concerning both nucleotide binding and conformational changes in F, during energy transformation (reviewed by Boyer et al., 1977), no species of F, other than thermophilic F, has ever been reconstituted from its five subunits in the absence of nucleotides. The complete reconstruction of thermophilic F, has revealed the roles of each subunit, as summarized in Table IV (Yoshida et al., 1977, Kagawa et al., 1979). None of the subunits alone shows ATPase activity, but ATP and ADP subunits, and the resulting nucleoare bound to isolated a and tide-subunit complexes have a tight conformation, as revealed by the relaxation spectrum of the lHJH exchange reaction (Ohta et al., 1980). The conformation change induced by the binding of nucleotide to complete F, has also been studied (Ohta et al., 1978). The difference circular dichroic spectrum of F, with a low Mg.AT(D)P concentration (10 p M ) is similar to that of the Mg.AT(D)P-a subunit complex, while that with a high Mg. AT(D)P concentration (100 ph’) approaches that of the Mg. AT(D)Psubunit complex. The conformational interaction between a and p subunits has been examined with deuterated subunits (a*and p*) reconstituted into hybrid complexes (a*P and @*a) and Fourier transform infrared spectroscopy has shown that the a subunit renders the conformation of tight only when ATP is present (Ohta et al., 1980).
208
YASUO KAGAWA
TABLE IV ROLESOF SUBUNITS OF F l u Function Net ATP synthesis H transport by ATPase 32Pi-ATP exchange ATPase activity H t gate activity Binding to F, N3 sensitivity Stabilization of F, H + channel activity ATP, ADP binding ITP, IDP binding CTP binding ATPase inhibitory activity +
(54,600)b
(51,000)
(30,200)
(21,000)
(16,000)
+ + + +c
+
-
+ +
-
' Data from Kagawa et al. (1979) and Kagawa and Nukiwa (1981). The molecular weights of subunits of thermophilic F, are shown in parentheses. The y subunit is necessary for stabilization of the a(3 complex. Clearly shown only in chloroplast F,.
The ligand specificity, dissociation constants, binding velocity, and conformational change of the nucleotide-subunit complex suggest that a is an allosteric nucleotide-binding site in F, and that /3 is a n isosteric site (Kagawa, 1978). The specific binding site of ATP in the p subunit has the following sequence: Ile-Met-Asp-Pro-Asn-Ile-Val-Gly-Ser-Glu-His-Tyr*Asp-Val-Ala-Arg, where Tyr* is the O-[14C]-sulfonylatedderivative of the tyrosine residue (Esch and Allison, 1978). The large negative ellipticity of the nucleotide-/3 subunit complex at 275 nm may be the result of stacking of the protonated tyrosine residue and the base of the bound nucleotide. It is interesting that this tyrosine residue is surrounded by an imidazole, a hydroxyl, and two carboxyl groups that may transfer H + during the interaction of ADP, Pi, and Mg2+.In the case of adenylate kinase, the synthetic nonapeptide corresponding to residues 32-40 of the enzyme, i.e., Tyr32Gly-Tyr-Thr-His-Leu-Ser-Thr-Gly-40, was shown to bind to the Mg. AT(D)P analog but not to free ADP. As in the case of F,, the anti form of the ATP molecule is connected to the peptide at Tyr32and Tyr34(Hamada et al., 1979). Further information is now being obtained by X-ray analysis of crystallized F, (Kagawa el al., 1976; Spitsberg and Haworth, 1977; Amzel and Pedersen, 1978). The computerized image reconstruction of thermophilic
1 1 . NET
ATP SYNTHESIS BY H+-ATPase
209
F, is shown in Fig. 7 (Wakabayashi et al., 1977), and a detailed analysis of such images is in progress. The H gate function of F, has been examined by measuring the inhibition of H + leakage through F, by the absorption of F1, or its components, to F,. Only combinations containing y& subunits can block the H leakage (Yoshida et al., 1977a). The connecting bridge between the aPy complex and F, is the & complex. In the case of the H+-ATPase of chloroplasts, a sharp 1000-fold increase in H + conductance has been observed when ApH +
+
. .
FIG. 7. A translationally filtered image of crystalline ATPase (F,). (From Wakabayashi et al., 1977.)
210
YASUO KAGAWA
exceeds 3 . The characteristics of the curve of H + flux versus ApH+are very similar to the current-voltage relationship for a Zenerdiode (Schonfeld and Newmann, 1977). The low H + conductance (0.1 p0-I cm-2) of the membrane allows formation of a high A&,+ for ATP synthesis; and high H + conductance, above the threshold, may protect the membrane from damage. There is some evidence that an electric field may be necessary to open the gate for ATP synthesis, but further studies are required on this possibility.
6. H + Channel and H
+
Filter: Chemical Structure of F,
Ever since crude F, was first extracted from mitochondria (Kagawa and Racker, 1966a,b), F, has been assumed to be an H + channel (Mitchell, 1967, 1976). When purified F, is incorporated into liposomes loaded with K , the addition of valinomycin-forming an inside negative A$ according to Eq. (4)-causes a rapid uptake of H + through F, (Okamoto et al., 1977). The uptake and release of H + from F, liposomes, and 32P,-ATP exchange in the presence of an equivalent amount of F,, are proportional to the amount of F, added to the liposomes. The observed passive leakage of K + is specifically blocked by the addition of F, (or H + gate) energy transfer inhibitors such as DCCD and anti-F, antibody. Since the only ion that can permeate through F, is H +,F, evidently is a highly specific filter structure. There are many reports on the composition of F,. Mitochondria1 preparations have proven difficult for such studies, because of both their complexity and their instability (Stiggall et al., 1978). In prokaryotic F,, only three peptides have been detected (Kagawa et al., 1976; Foster and Fillingame, 1979; Babakov and Vasilov, 1979). These three peptides are the DCCD sensitivity-conferring protein (19,000 MW), F,-binding protein (13,500 MW), and DCCD-binding protein (7500 MW) (Kagawa et al., 1979). Only in the case of thermophilic F, has the F,-binding protein been separated from the DCCD-binding protein in an active form (Sone et al., 1979). It is still not certain whether the DCCD sensitivity-conferring protein is essential for ATP synthesis by prokaryotic H+-ATPase, but it is essential for that of mitochondria1 H -ATPase. The DCCD sensitivityconferring protein of mitochondria has been called the oligomycin sensitivity-conferring protein (basic protein, 18,000 MW) and has been purified from F, (Kagawa and Racker, 1966a) and shown to be essential for ATP synthesis (Racker, 1976; Kagawa, 1972). The DCCD-binding protein is the most hydrophobic protein known, and its amino acid sequence has been determined in many H + -ATPases (Sebald et al., 1979). Figure 8 shows the hypothetical structure of this peptide +
+
11. NET
ATP
SYNTHESIS BY
H+-ATPase
211
FIG. 8. DCCD-binding protein of thermophilic bacterium PS3. The primary structure is based on the sequence determined by Sebald ef a/. (1979), and the hypothetical conformation-with hydrophobic residues outside and hydrophilic residues inside-is like that of valinomycin.
obtained from thermophilic bacterium PS3 (sequence data from Sebald et al., 1979). It is crucial that six copies of this peptide be present in one molecule of H+-ATPase (Kagawa et al., 1976); when only one glutamyl
residue (Fig. 8, lower right) of one of the peptides is chemically blocked by DCCD, or any other agent, the H + channel activity of F, is lost. As shown in Fig. 8, there is only one tyrosyl group in the DCCD-binding protein, but when one-third of the total tyrosyl groups of this protein are nitrated with tetranitromethane, H + conduction is lost (Sone et al., 1979). Chemical modification of arginine with glyoxal or phenylglyoxal has a similar effect (Sone et al., 1979). These observations suggest that specific proton translocation through F, may occur by protonation and deprotonation of the polar groups in the DCCD-binding protein (Sone et al., 1981).
V. EPILOGUE Voltage-driven reversal of purified H -ATPase results in the synthesis of ATP. The stepwise reconstitution of H+-ATPase from the H + channel, H + gate, and H + pump result in stepwise restoration of physiological functions such as facilitated diffusion, regulated transport, and active transport of H + through lipid bilayers (Kagawa, 1978). No other channels and gates known to physiologists have been studied biochemically in such detail as those of H+-ATPase, and no ion pumps except F, and bacteriorhodopsin have been studied crystallographically. Yet the molecular mechanism of energy transduction in H +-ATPase is still a mystery, and even its H' /ATP stoichiometry is controversial. The energy of H + transport is the product of an intensive factor (ApH+) and an extensive factor (equivalents of H + translocated). When the flux of +
212
YASUO KAGAWA
H + (JH+) is observed as a function of time in the H+-ATPase system, the energy is t ApH+JH+ dt (7)
so
as discussed in Section II,D, both theoretical and experimental difficulties (such as H+ leakage or ATP hydrolysis by partially detached F,) complicate evaluation of this function. Clearly, molecular events occurring during H+ transport and ATP synthesis must be studied by new methods, both theoretical and experimental. ACKNOWLEDGMENT This research was supported by grants from the Ministry of Education, Science and Culture of Japan. REFERENCES Amzel, L.M., and Pedersen, P . L. (1978). J . Biol. Chem. 253, 2067-2069. Babakov, A. V., and Vasilov, R. G. (1979). Bioorgan. Khim. 5, 119-125. Boyer, P. D., Chance, B., Ernster, L., Mitchell, P., Racker, E., and Slater, E. C. (1977). Annu. Rev. Biochem. 46, 955-1026. Brand, M. D., and Lehninger, A. L. (1977). Proc. Natl. Acad. Sci. U.S.A. 74, 1955-1959. Downie, J . A., Gibson, F., and Cox, G. R. (1979). Annu. Rev. Biochem. 48, 103-131. Esch, S. F., and Allison, W. S. (1978). J . Biol. Chem. 253, 6100-6106. Foster, D. L., and Fillingame, R. H. (1979). J. Biol. Chem. 254, 8230-8236. Hamada, M., Palmieri, R. H., Russell, G . A., and Kuby, S. A. (1979). Arch. Biochem. Biophys. 195, 155-177. Jagendorf, A. T., and Uribe, E. (1966). Proc. Natl. Acad. Sci. U.S.A. 55, 170-177. Kagawa, Y. (1967). Meihods Enzymol. 10, 505-510. Kagawa, Y. (1972). Biochim. Biophys. A d a 265, 297-338. Kagawa, Y. (1978). Biochim. Biophys. Acia 505, 45-93. Kagawa, Y. (1980). J . Membr. Biol. 55, 1-8. Kagawa, Y . , and Ariga, T. (1977). J . Biochem. (Tokyo) 81, 1161-1165. Kagawa, Y., and Nukiwa, N. (1981). Biochem. Biophys. Res. Commun. 100, 1370-1376. Kagawa, Y., and Racker, E. (1966a). J . Biol. Chem. 241, 2461-2466. Kagawa, Y . , and Racker, E. (1966b). J . Biol. Chem. 241, 2467-2474. Kagawa, Y., and Racker, E. (1966~).J . Biol. Chem. 241, 2475-2482. Kagawa, Y., and Racker, E. (1971). J. Bioi. Chem. 246, 5477-5487. Kagawa, Y., and Sone, N. (1979). Methods Enzymol. 55, 364-372. Kagawa, Y., Ohno, K., Yoshida, M., Takeuchi, Y., and Sone, N. (1977). Fed. Proc. Fed. A m . Soc. Exp. B i d . 36, 1815-1818. Kagawa, Y., Sone, N., Yoshida, M., Hirata, H. and Okamoto, H. (1976). J. Biochem. (Tokyo) 80, 141-151. Kagawa, Y . , Sone, N., Hirata, H., and Yoshida, M. (1979). J . Bioenerg. Biomembr. 11, 39-78. Kozlov, I . A,, and Skulachev, V. P. (1977). Biochim. Biophys. Acia 465, 29-89.
11. NET ATP SYNTHESIS BY H+-ATPase
213
McCarty, R. E. (1979). Annu. Rev. Plant Physiol. 30, 79-104. Mitchell, P . (1966). Biol. Rev. 41, 455-502. Mitchell, P. (1967). In “Comprehensive Biochemistry” (M. Florkin and E. M. Stotz, eds.) Vol. 22, pp. 167-197. Elsevier, Amsterdam. Mitchell, P. (1976). Biochim. SOC. Trans. 4, 399-430. Muratsugu, M., Kamo, N., Kurihara, K., and Kobatake, Y. (1977). Biochim. Biophys. Acfa 464, 613-619. Ohta, S., Nakanishi, M., Tsuboi, M., Yoshida, M., and Kagawa, Y. (1978). Biochem. Biophys. Res. Commun. 80, 929-935. Ohta, S., Tsuboi, M., Yoshida, M., and Kagawa, Y. (1980). Biochemistry 19, 2160-2168. Okamoto, H., Sone, N., Hirata, H . , Yoshida, M., and Kagawa, Y. (1977). J. Biol. Chem. 252, 6125-6131. Penefsky, H. S. (1979). Adv. Enzymol. 49, 223-280. Pullman, M. E., Penefsky, H. S., Datta, A., and Racker, E. (1960). J. Biol. Chem. 235, 3322-3329. Racker, E, (1976). “A New Look at Mechanisms in Bioenergetics, pp. 1-197. Academic Press, New York. Rogner, M., Ohno, K., Hamamoto, T., Sone, N., and Kagawa, Y. (1979). Biochem. Biophys. Res. Commun. 91, 362-367. Rottenberg, H. (1975). J. Bioenerg. 7, 61-74. Schonfeld, M., and Neumann, J. (1977). FEBS Lett. 73, 51-54. Sebald, W . , Hoppe, J., and Wachter, E. (1979). In “Function and Molecular Aspects of Biomembrane Transport,” pp. 63-74. Elsevier, Amsterdam. Sone, N., Yoshida, M., Hirata, H., and Kagawa, Y. (1975). J . Biol. Chem. 250, 7917-7923. Sone, N., Yoshida, M., Hirata, H., Okamoto, H., and Kagawa, Y. (1976). J . Membr. Biol. 30, 121-134. Sone, N., Yoshida, M., Hirata, H., and Kagawa, Y. (1977). J . Biol. Chem. 252,2956-2960. Sone, N., Yoshida, M., Hirata, H., and Kagawa, Y. (1978). Proc. Natl. Acad. Sci. U.S.A. 75, 4219-4223. Sone, N., Ikeba, K., and Kagawa, Y. (1979). FEBS Lett. 97, 61-64. Sone, N., Hamamoto, T., and Kagawa, Y. (1981). J. Biol. Chem. 256, 2873-2877. Spitsberg, V., and Haworth, R. (1977). Biochim. Biophys. Acta 492, 237-240. Stiggall, D. L., Galante, Y. M., and Hatefi, Y. (1978). J . Biol. Chem. 253, 956-967. Stucki, J . W. (1978). In “Energy Conservation in Biological Membranes” (G. Schafer and M. Klingenberg, eds.), pp. 264-287. Springer-Verlag, Berlin and New York. Thayer, W. P., and Hinkle, P. C . (1975). J . Biol. Chem. 250, 5336-5342. Tsuchiya, T., and Rosen, B. P. (1976). J . Bacteriol. 127, 154-161. Wakabayashi, T., Kubota, M., Yoshida, M., and Kagawa, Y. (1977). J . Mol. Biol. 117, 5 15-5 19. Witt, H. T. (1979). Biochim. Biophys. Acta 505, 355-427. Witt, H. T., Schlodder, E., and Graber, P. (1976). FEBS Letr. 69, 272-276. Yoshida, M., Okamoto, H., Sone, N., Hirata, H., and Kagawa, Y. (1977a). Proc. Null. Acad. Sci. U.S.A. 74, 936-940. Yoshida, M., Sone, N., Hirata, H., and Kagawa, Y. (1977b). J . Biol. Chem. 252, 3480-3485. Yoshida, M., Sone, N., Hirata, H., Kagawa, Y., and Ui, N. (1979). J. Biol. Chem. 254, 9525-9533.
This Page Intentionally Left Blank
CURRENT TOPICS IN MEMBRANES A N D TRANSPORT. VOLUME 16
Chapter 72
Phosphorylat ion in Chloroplasts: ATP Synthesis Driven by A$ and by ApH of Artificial or Light-Generated Origin PETER GRABER Max- Volrner-Institut fur Biophysikalische und Physikalische Chemie Technische Universitat Berlin Berlin, Federal Republic of Germany
1. Introduction ....................................................................................... 11. Background Information .............................. ................................ A. . General ....................................................................................... B. Electron Transport ........................................................................
215 216 216 217 217 219 220 225 IV. The Functional Unit for ATP Synthesis ..................................................... 228 229 V. The Kinetics of ATP Synthesis ........ ....................... A. Simplified Steady-State Kinetics ........................................................ 229 232 B. Kinetics of ATP Synthesis by Light-Induced ApH and A+ ....................... C. ATP Synthesis Kinetics Produced by an Artificial A$ ................ 234 D. lnfluence of an Artificial ApH on the Kinetics of ATP n ............. 236 E. Interpretation of Thresholds and Lag Times for ATP Synthesis ................ 237 VI. The Problem of Energetic Sufficiency ... ................... 239 VII. Epilogue: Conformational Changes Associated with Energization ................... 24 1 References ......................................................................................... 243
1.
INTRODUCTION
Among biological membranes, one of the most highly structured and physiologically complex is that of the chloroplast thylakoid, wherein at 215
Copyright @ 1982 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-153316-6
216
PETER GRABER
least five distinct ensembles of molecular events can be identified and studied: light capture, redox reactions, transmembrane charge separation, ionic transport, and ATP synthesis. All these are organized for the central purpose of converting light energy to the currency most convenient for biological functions: the energy of redox equivalents and the energy of phosphate anhydride bonds in ATP. Over the past two decades, powerful physical and chemical techniques have been combined for mechanistic dissection of the chloroplast membrane, and many of the constituent reactions are now reasonably well understood. A comprehensive review of photosynthetic mechanisms has recently been presented by Witt (1979), and the major redox events-particularly related to oxygen evolution and transfer of protons-are reviewed elsewhere in this volume (Junge, this volume). This article will focus on the process of ATP synthesis in the thylakoid membrane, and-in keeping with the general subject of this volume-will treat especially those properties of the membrane ATPase and ATP synthetase involved in the utilization of electric fields and proton gradients to accomplish ATP synthesis. It is hoped that this article will be useful to investigators working on other problems in membrane transport and energy conservation, sincedespite structural differences among membrane ATPases-the general mechanisms of ATP hydrolysis and synthesis seem to be similar in a variety of biological membranes (see Boyer et al., 1977; Racker, 1978).
II.
BACKGROUND INFORMATION
A. General A typical thylakoid is a disc-shaped vesicle (membrane thickness about 7 nm) with a diameter of about 500 nm and a short axis of about 20 nm. Approximately 200 electron transport chains with = lo5 chlorophyll molecules and -200 ATPase molecules are embedded in the membrane of each thylakoid. Multiple thylakoids are partially interconnected in chloroplasts, and the shape and volume of the internal aqueous space depend upon properties of the external solution such as ionic strength and osmolarity. Isolated thylakoid membrane systems (without the chloroplast envelope), called class I1 chloroplasts, have provided the experimental material for most of the work reported in this article. These isolated membranes carry out all the reactions involved in light capture, evolution of oxygen, reduction of NADP+ , and phosphorylation of ADP.
217
12. PHOSPHORYLATION IN CHLOROPLASTS
B. Electron Transport The photosynthetic electron transport chain, which performs the work
of transferring electrons from water to NADP+ (representing a step of more than 1 . 1 V in redox potential) at the expense of photon energy, contains six main building blocks: the antenna system, two distinct reaction centers (I and 11), the water-splitting enzyme complex, a pool of plastoquinone molecules, and the NADP -reducing enzyme. Turnover of the molecular machinery is initiated by the absorption of light quanta in the antenna system and channeling of these quanta via singlet-singlet migration to the two reaction centers. A special chlorophyll molecule in each reaction center becomes oxidized, transferring electrons outward through the thylakoid membrane, thus setting up an electric field. This field can be detected either optically, by electrochromic absorption changes of the carotenoid pigments intrinsic to the membrane (Junge and Witt, 1968), or electrically, using a capacitive electrode with chloroplasts spread at a heptane-water interface (Trissl and Graber, 1980). Oxidizing equivalents from reaction center I1 are transferred to the water-splitting enzyme, resulting in the oxidation of water, evolution of oxygen, and release of protons to the inside of the thylakoid. This proton release can be measured with various p H indicator dyes (Auslander and Junge, 1975; Tiemann et af., 1979a). Electrons from the reaction center I1 acceptor are transferred to plastoquinone (Stiehl and Witt, 1969), to be joined by equivalent protons from the outside of the thylakoid and translocated electroneutrally to the inner space. Plastohydroquinone is subsequently reoxidized by oxidizing equivalents from reaction center I in such a way as to release one further proton to the inside. Electrons from reaction center I are transferred to NADP' . The net results of a single turnover of the photosynthetic apparatus are: (1) production of a 0 , and tNADPH, (2) generation of a potential of about 50 mV (Junge and Witt, 1968; Witt, 1979), and (3) translocation of two H + ions across the membrane. A detailed description of these events is provided by Junge (this volume). +
C. Structure of the ATPase
Current structural knowledge of chloroplast ATPase is based on five main lines of research (reviewed by Nelson, 1976; Baird and Hammes, 1979): electron microscopy (Howell and Moudrianakis, 1967), biochemical characterization, functional characterization by use of specific antibodies and cross-linking reagents, small-angle X-ray scattering (Paradies et af.,
218
PETER GRABER
1978), and fluorescence resonance energy transfer measurements (Cantley and Hammes, 1976). Figure 1 shows a tentative geometrical arrangement, the CF, part, drawn approximately to scale (Nelson, 1976; Baird and Hammes, 1979). The ATPase is easily resolved into two main parts. The larger of these, which bears the catalytic site, is water-soluble and appears attached t o (rather than embedded in) the membrane. It has been called the chloroplast coupling factor 1 (CF,). The smaller part, designated CF,, is strongly hydrophobic and embedded in the membrane. It appears to facilitate proton transfer through the thylakoid membrane into the catalytic portion, CF,. The subunit composition and functional arrangement of the chloroplast ATPase resemble those of mitochondria1 and bacterial membranes (see especially Kagawa, this volume), and are summarized along with other information in Table I. The spatial arrangement of the subunits shown in Fig. 1 is based largely on cross-linking data, which indicate that both CY and /3 subunits of CF, are in contact with the smaller subunits (Baird and Hammes, 1976), while the 6 subunit is most important for the attachment of CF, to the membrane (Younis et al., 1977). The complete ATPase, meaning CF,-CF,, can be reconstituted into liposomes, where it will carry out ATP synthesis if either some kind of proton pump is also incorporated into the liposomes (Winget et al., 1977) or the liposomes are subjected to an acid-base transition (Pick and Racker, 1979).
a
FIG. 1. Scheme of the proposed structure of the chloroplast ATPase. Outer surface of the thylakoid membrane is upward.
219
12. PHOSPHORYLATION IN CHLOROPLASTS
TABLE I PHYSICAL PROPERTIES AND SUGGESTED FUNCTIONS OF THE CHLOROPLAST ATPASECF,-CF, AND ITS SUBUNITS Molecular weight
Unit
417,700
325,000
Properties and functions Can be reconstituted into liposomes; active in A T P synthesis and A T P hydrolysis (Winget et a/., 1977; Pick and Racker, 1979) Easily removable from the membrane; can be reconstituted, isolated soluble enzyme catalyzes A T P separable into five hydrolysis; spherical d = 100 distinct subunits with stoichiometry cr2, P2, 7, 6, t2 (Nelson, 1976; Binder ef a/., 1978); strongly hydrated, containi?g 10’ H,O (Paradies et a/., 1978) Contains sites which regulate proton permeability (identical with site for tightly bound nucleotides?); spherical d = 58 A (McCarty, 1979; Nelson ef a/., 1973) Contains sites for ATP hydrolysis (and synthesis), spherical d = 5 8 A (Deters et a / . , 1975) Catalyzes ATP hydrolysis (Deters el a/., 1975) Provides binding sites for the ATPase inhibitor ( 6 subunit) (Nelson et a / . , 1972) Required for binding of C F , to the membrane (Younis et a/., 1977) elongated shape 28 x 25 x 100 A (Schmidt and Paradies, 1977) Inhibitor of ATP hydrolysis (Nelson ef a / . , 1972) Proton channel; binding site for CF, (Sigrist-Nelson et a/., 1978) Contains six identical subunits (MW = 7700); they form a proton-conducting channel in liposomes; binding of one molecule of dicyclohexylcarbodiimide (DCCD) t o the hexamer inhibits proton conduction (Nelson et a/., 1977; Sigrist-Nelson el a/., 1978) Possibly required for assembly of the six identical subunits to a functional active proton channel and for binding of CF, (Pick and Racker, 1979; Nelson ef a/., 1980)
A
cy
62,000
P
57,000
Y
238,000 38,000
6
21,000
cy2P2
14,000 92,700 DCCD-binding protein (hexamer)
Three further polypeptides
Ill.
46,200
17,500 15,500 13,500
COUPLING OF PROTON TRANSPORT TO ATP SYNTHESIS
A. Experimental Evidence for an Ionic Coupling According to chemiosmotic theory (Mitchell, 1961, 1966), the essential intermediate beLween electron transport and ATP synthesis is an electro-
220
PETER GRABER
chemical gradient for protons. Application of chemiosmotic theory to photosynthetic ATP synthesis leads to the following predictions: Lightinduced electron transport should create a difference in electric potential (A$) and/or a difference in proton concentration (ApH) across the thylakoid membrane. Protons thus transferred to the inside of the thylakoid should flow back through the membrane via the (membranebound) ATPase, so that the free enthalpy available from proton back flux (down its electrochemical gradient) can drive the ADP-ATP reaction toward formation of ATP. As a corollary of these predictions, the proton back flux should be stoichiometrically related to ATP synthesis; i.e., the ratio H + / A T P should be constant. Many different types of experiments have been carried out which admit straightforward interpretation of photosynthetic ATP synthesis by means of such proton coupling. The major results are outlined beiow. 1 . ATP SYNTHESIS IN AN ACID-BASETRANSITION
This result was the first to demonstrate the feasibility of a chemiosmotic mechanism in chloroplasts (Jagendorf and Uribe, 1966). It can be obtained by incubating class I1 chloroplasts for about 30 seconds in buffered medium at pH 4-5 (to allow buffer entry and acidification of the internal space) and then injecting the acidified chloroplasts into another buffered medium at a higher pH. With alkaline pHs in the neighborhood of 8 or higher (giving a transmembrane pH of 3-4 units) net ATP synthesis is observed, in an amount depending on the exact magnitude of ApH and on the internal buffer capacity (quantity of stored protons). ATP synthesis under these conditions is not blocked by electron transport inhibitors, but is blocked by ATPase inhibitors, as summarized in Table I1 (Jagendorf and Uribe, 1966). More recently, kinetic analysis of phosphorylation driven by an artificial ApH has been carried out using rapid quenched-flow techniques, which will be discussed in Section V,D.
2. ATP SYNTHESIS BY
AN
EXTERNAL VOLTAGEPULSE
Membrane potentials capable of driving ATP synthesis can be imposed on suspended chloroplast thylakoids by creating an electric field across the suspending medium. The principle of this experiment is outlined in Fig. 2. The actual membrane potential developed depends upon solution conductivities inside (A,) and outside (A,) the thylakoids, upon the membrane conductivity (AM), and upon the size and shape of the thylakoid vesicles, as well as upon the electric field strength. Under the usual conditions, however, a membrane potential of 150 mV can be calculated at the vesicular poles (facing the electrode plates) for a field strength of 1000 V/cm. It sF -1lI.d be
TABLE I1 ATP SYNTHESIS IN AN ACID-BASETRANSITION AND IN EXTERNAL VOLTAGEPULSES
Conditions
ATP yieldu
Addition of electron transport inhibitorb
Acid-base transitiond PHout= 8 reaction complete after -4 sec ApH = 4.2
Without exogenous buffer
14.6
No inhibition
External voltage pulsese PHout = 8 pulse duration 30 msec A$ = 200 mV
1 pulse 10 pulses
Quenched-flow acid-base transitionf PHout= 8 one ApH pulse ApH = 3.0
30-msec pulse duration 300-msec pulse duration
Plus 15 mM succinate
Millimoles ATP per mole Chl.
* For example, 3-(3’ ,4’-dichlorophenyl)-l,l-dimethylurea. For example, triphenyltin chloride. Jagendorf and Uribe (1966). Witt el al. (1976). 1 Schatz el al. (1978).
Addition of ATPase inhibitor
155 0.38 3.8
No inhibition
100% inhibition
1.5
No inhibition
100% inhibition
15
a
b
C A* d50mV
FIG. 2. Schematic explanation for generation of a transmembrane electric potential difference, A$, by an external electric field. Potential distribution (a) in a homogeneous conducting medium; (b) a nonconducting sphere is placed in the conducting medium; and (c) a vesicle with a conducting inner phase and a nonconducting shell is placed in the conducting medium.
12. PHOSPHORYLATION IN CHLOROPLASTS
223
noted that the actual membrane potential thus developed must be of opposite sign at the two poles, so that only one half of the membrane can be electrically polarized in the same direction produced by light. Voltage pulses applied in this manner produce ATP synthesis in an amount dependent on the magnitude and duration of the pulse (Witt eta/., 1976). Once again, ATP generation is not blocked by electron transport inhibitors but is abolished by ATPase inhibitors. A comparison of ATP yields from the original experiments with acid-base transitions (Jagendorf and Uribe, 1966), from ApH pulse experiments (Schatz et a/., 1978), and from the voltage pulse experiments (Witt et a/., 1976) is provided in Table 11. When geometric differences (especially the nonuniform membrane polarization during voltage pulses) are allowed for, it is evident that both A$ and ApH pulse experiments give roughly the same yield of ATP for equivalent times and gradients.
3. ACCELERATED CHARGEEFFLUXWITH ATP SYNTHESIS Light flashes of several milliseconds duration are capable of generating membrane potentials near 200 mV across the thylakoid membrane, as indicated by the carotenoid absorption shift (measured at 515 nm) referred to above. Under standard basal conditions, they decay (following each flash) with a half-time of 150-200 msec, as illustrated in Fig. 3 (top); but when phosphorylation is permitted to occur, by admitting ADP to the suspensions, the half-time for decay can shorten to 40 msec or less (Rumberg and Siggel, 1968; Junge era/., 1970). In the time plots of Fig. 3 (top), the curve slopes directly reflect charge efflux (i.e., the transmembrane current). Therefore, differentiation of each curve and replotting of the slope versus the actual value of the curve (relative electric potential) at several time points yields complete current-voltage curves for the thylakoid membrane under basal and phosphorylating conditions (Fig. 3, bottom). Obviously, the basal curve is linear (ohmic), while the phosphorylating curve is strongly superlinear at larger values of A$. Taken together, these observations suggest that the total charge flux, during decay of the light-induced membrane potential, is the sum of two components-basal and phosphorylating-which have very different voltage dependences.
4. ACCELERATED PROTON EFFLUXWITH ATP SYNTHESIS Results wholly analogous to those just described have been obtained for actual proton fluxes by measuring the response of external pH (with glass electrodes and indicator dyes) to switching off of the light after periods of saturating continuous illumination. Quantitative analysis of these experiments must allow, however, both for homeostatic mechanisms within the
224
PETER GRABER
T
1.0 X
3 k
YY-
Q) Q)
Zl
b V
a3 .->
0.5 / O
CI
a E
d
/ -o&+o--"=-+o,
lo
7:
-ADP
'
I
relative electric potential difference FIG. 3. Top: Time course of the decay of electric potential after light pulse excitation under phosphorylating ( + ADP) and nonphosphorylating ( - ADP) conditions. Bottom: Charge efflux as a function of A$ (current-voltage relationship) under phosphorylating and nonphosphorylating conditions.
12. PHOSPHORYLATION I N CHLOROPLASTS
225
thylakoids, which could offset part of the photosynthetic proton fluxes, and for phosphorylation-consumption of protons due to the pK differences among P,, ADP, and ATP. When such allowances are made, and the resultant data are plotted similarly to the flux-versus-voltage data of Fig. 3 (bottom), essentially similar curves are obtained, and an analogous conclusion has been drawn: Total proton flux, during decay of the lightinduced ApH, is the sum of two components-basal and phosphorylating-which have very different ApH dependences (Schroder et al., 1972).
B. Competition between Basal and Phosphorylating Proton Flux While the four different experiments described above make a strong (qualitative) argument for the coupling of ATP synthesis to charge and proton movement, they still leave several critical questions unanswered: Are the phosphorylation-coupled charge flux and the phosphorylationcoupled proton flux identical? Does the basal efflux occur physiologically (in vivo) or is it the result of preparative damage? How does the occurrence of two different fluxes affect measurements of the H +/ATP ratio? 1. BASALAND PHOSPHORYLATING PROTONFLUXES in Vivo
In order to identify positively the ionic species responsible for accelerated charge efflux, the decay of A$ has been studied as a function of ApH. Intact cells of the unicellular alga Chlorella vulgaris are preilluminated for about 2 minutes at different light intensities in order to establish different steady state p H gradients across the thylakoid membrane. Then, after a brief interruption of the light, a single-turnover flash is fired and the decay of A$ is followed via the absorption shift at 515 nm. When the resultant data are plotted in the same manner as in Fig. 3 (bottom), the results of Fig. 4 (top) are obtained. Each curve in this figure represents the apparent current-voltage relationship, for charge efflux from Chlorella thylakoids, at a different value of ApH. The similarities between these curves and those of Fig. 3 (bottom) are obvious; and in addition, the degree of curvature-or superlinearity (in the dependence of charge flux on potential)-can be seen to diminish as the steady state ApH across the thylakoid membrane diminishes. Similar results have been obtained with Chlamydomonas (Joliot and Delosme, 1974), which also has provided a mutant (F-54) blocked in phosphorylation (Bennoun and Levine, 1967). This mutant shows no accelerated charge efflux (Joliot and Delosme, 1974)
226
PETER GRABER
0.5-
/'
i i
i nonphosphorylat ing mutant F S L A p H =1-2
/
,.o-O~O I
1
relative electric potential difference FIG.4. Top: Charge efflux as a function of A$ at different values of ApH, in Chlorellu. Bottom: Charge efflux as a function of A$ under phosphorylating (wild type) and nonphosphorylating (mutant F-54)conditions in Chlumydomonus.
and little sign of superlinearity in its current-voltage relationship, as shown in Fig. 4 (bottom). Thus it may be concluded that basal charge flux indeed occurs in vivo, but that it is probably not carried by protons, since it is not affected by increasing ApH. Furthermore, the accelerated charge efflux during
12. PHOSPHORYLATION IN CHLOROPLASTS
227
phosphorylation increases with ApH, making protons by far the most likely charge-carrying species. 2. BASALAND PHOSPHORYLATING PROTONFLUXES IN THE STEADY STATE Experiments analogous to that just described, but examining proton flux at various values of A$, can be carried out using repetitive-flash excitation with class I1 chloroplasts (Graber and Witt, 1974, 1976). Variable-length trains of single-turnover flashes (20-psec flashes at 2-msec intervals) are used to vary A$ across the thylakoid membrane; and repetition of these trains at frequencies of 0.1-10 Hz is used t o set different values of ApH. The p H in the internal space of the thylakoids can be estimated by fluorescence quenching of 9-aminoacridine (Schuldiner et al., 1972), average proton flux can be calculated from the known stoichiometry of protons displaced per single turnover of the redox chain (Graber and Witt, 1975), and ATP formation can simultaneously be determined as organic phosphate remaining after the removal of inorganic phosphate with molybdate reagent (Avron, 1960). Once again, the results show a linear dependence of basal proton efflux on the ApH across the thylakoid but a strongly superlinear dependence of flux stimulated by phosphorylation. The rate of ATP synthesis also displays a supralinear dependence on ApH. In semilogarithmic plots of either the rate of ATP synthesis or the stimulated proton flux against ApH, straight lines are obtained having slopes between 2 and 2.5, with a slight dependence on A$, over the range 50-125 mV. Similar functional relationships have also been found by other authors (Schroder, 1974; Portis and McCarty, 1976). 3. THEH /ATP STOICHIOMETRY +
The different dependences of basal and phosphorylation-stimulated proton flux upon ApH across the thylakoid membrane necessarily means that the observed ratio H + / A T P must vary with ApH if no distinction between the two fluxes is made. Figure 5 is summary plot of observed H + / A T P ratios from the above experiments. As expected, the apparent coupling ratio is variable: With higher gradients, of either potential or ApH, a greater fraction of protons is channeled through the phosphorylating pathway; consequently, H+/ATP decreases and the curves approach a limit at which practically all protons are phosphorylating. The experimental limiting value has been found to be 2.4, a figure which lies in the middle of the range of values (2-3) for the coupling ratio estimated by other methods (for review see Hauska and Trebst, 1977; Reeves and Hall, 1978).
228
PETER GRABER
20
LO
H+ AT P
curves calculated
15
30 AIY2=75mV
A
e
A T 20
10
10
5
1.5
2 .o
2.5
3.0
APH FIG. 5 .
IV.
H + / A T P ratio and e/ATP ratio as a function of ApH at different A$.
THE FUNCTIONAL UNIT FOR ATP SYNTHESIS
At least one important additional type of result can be obtained from these studies on phosphorylation-accelerated charge and proton flux. The basal rate for the decay of membrane potential following a flash (Fig. 3, top) is interpreted as the result of nonspecific ionic leaks which draw current away from the ATP-synthesizing enzyme. If this is true, then artificial leaks in the thylakoid membrane should both accelerate the decay of flashgenerated membrane potential and sharply reduce flash-driven ATP synthesis. By using the channel-forming ionophoric agent gramicidin, Boeck and Witt (1972) were able to show that only two molecules of gramicidin per lo5 chlorophyll molecules were required to double the rate of decay of A$ and, simultaneously, to halve the ATP yield per flash. Since two molecules of gramicidin are needed to form a transmembrane pore (Bamberg and Lauger, 1973) and since there are lo5chlorophyll molecules per thylakoid, it follows that a single leakage channel can half-short-circuit ATP synthesis by the whole thylakoid. In other words, the energy supply for ATP synthesis (in this case the flash-induced A$) is delocalized over the whole thylakoid, thus defining the physiologically functional unit for ATP synthesis.
-
229
12. PHOSPHORYLATION IN CHLOROPLASTS
V.
THE KINETICS OF ATP SYNTHESIS
Thus far, no thorough formal description of the kinetic behavior of the membrane-bound ATPase or ATP synthetase of the chloroplast thylakoid has been achieved, because of the very complex reaction pattern: binding of ATP, ADP, and Pi at the catalytic site; binding of all three adenine nucleotides at one other site (at least); influence of magnesium ions on catalysis and on binding of CF, to the membrane; proton transport and interaction with two different solution phases; participation of inhibitor subunits of the protein; and feedback interaction between the ATPase and the membrane electron transport chains. In order to make progress with the kinetic analysis at all, therefore, it is absolutely essential to simplify the formal relationships and to devise experiments which hold as many peripheral conditions constant as possible, compatible with the theoretical development. Furthermore, under nonenergized conditions in the presence of ATP, no ATP hydrolysis is observed, although the AG value favors ATP hydrolysis. Obviously, this reaction is kinetically inhibited; i.e., the enzyme is inactive under nonenergized conditions. After energization, ATP hydrolysis can be observed; however, since the energization necessary for activation favors ATP synthesis, high rates of ATP hydrolysis cannot be expected.
A. Simplified Steadystate Kinetics We have begun with the simplest reaction sequence which can describe the activation of ATPase and ATP synthesis. The enzyme is transformed from its inactive state, Ei, into its activated state, E,, if simultaneously exposed (on opposite faces of the membrane) to acidic and basic solutions: Ei
+
bHii
+ bOH;,,
kl +
E,
(1)
k- 1
This preceding activation equilibrium is followed by the catalytic process E,'+ ADP + Pi + nH$
k2
+
k-2
E,*
k3
+
E,'+ nH;",
+ ATP + (n + 1)H,O
(2)
k-3
where b is the number of H + and OH- in the activation process and n is the number of H + necessary for ATP synthesis (H+/ATP). In reaction (1) the influence of ADP, ATP, Pi, Mgz+,and other reagents, e.g., thiol reagents, has been omitted. For the catalytic reaction (2) all considerations of binding order of substrates and proton association-dissociation and recycling
230
PETER GRABER
of the enzyme have been omitted. A more detailed description includinp these steps has been given elsewhere (Graber and Schlodder, 1980). Furthermore, it is assumed that both processes-activation and ATP synthesis-occur independently of each other; i.e., the substrates (ADP, ATP, Pi, and nH;) can bind to the enzyme in the inactive and in the activated state with the same affinity. This assumption is not implausible, since both processes occur in different subunits of the enzyme. It follows from this assumption that the inactive enzyme species, Ei, is in equilibrium with the activated enzyme species, i.e., for the simplified reaction scheme (2), E, = E,’+ E.: At equilibrium it results from these assumptions and from Eq. (l), with E,= E i + E, (Eo- total amount of enzyme), that
-E,_ - k,[H~lb[OH,u,lb 1’ + k-I E, k ,[Hi: 1b[OH~ut
(3)
If a potential difference A$ exists across the membrane and the ATPase, then the proton concentration, [Hg], at the H+-binding site of the ATPase must differ from that of the thylakoid interior, [Hf] (Fig. 6). A transformation of A$ into a change in the proton concentration might occur in a so-called proton well (Mitchell, 1968). The proton well was defined as a narrow cleft in the membrane, being in electrochemical equilibrium with the internal phase; whereas the electric potential changes in about the same way as in the adjacent parts of the membrane. In general, the potential at the H+-bindingsite $E will represent some fraction ( a )of the total potential difference across the membrane. Then, assuming electrochemical equilibrium between the binding site and the thylakoid interior, we can write [HE+]= [Hf] exp(aFA$/RT)
(4)
and it is the left-hand expression here, rather than simply [Hf], which must be introduced into Eq. (3). A similar relationship might be written for hydroxyl ions, but we shall assume instead that there is no potential difference between this binding site on the enzyme and the outer aqueous medium; then OH- is in simple chemical equilibrium, so that [OH,]
= [OH,,,
1
(5)
Since, in aqueous medium the proton and hydroxyl ion concentrations are linked by the dissociation constant of water, P = [H+] [OH-], [OH;,,] in Eq. (3) can be replaced by P/[H&,], and the resulting constant k , P bcan be symbolized by k;. After making these substitutions and entering the electrochemical potential for protons [left-hand side of Eq. (4)], we obtain
231
12. PHOSPHORYLATION IN CHLOROPLASTS
0
c Q) +
+ L Q)
distance FIG.6. Top: Structural diagram of the ATPase, giving rise to kinetic equations (1)-(11). Proton-binding sites on the enzyme ( H a are in electrochemica/ equilibrium with the internal phase; hydroxyl-binding sites on the enzyme (OHE) are in chemical equilibrium with the external phase. Bottom: Suggested A$ profile within CF,-CF,.
_E, _- k,’([H;I /[H&,1 Ib exP(baF A$/R T ) k;([H$] /[H,:,])b exp(baF A$/RT) E, Finally, letting ApH = pH,,, - pH,, gives
_Ea _- kl’exp[2.3b ApH + (abF/RT) A$] E,
k,‘ exp[2.3b ApH + (abF/RT) A$]
+ kq + k-,
(6)
(7)
The actual rate of ATP production can be written as (if the concentration of ATP is zero, i.e., no back reaction)
232
PETER GRABER
If all the activated ATPases are in the state E: [i.e., E,=E,*; this occurs if the ATPase is saturated with the substrates ADP and Pi and if the binding constant for the “phosphorylating protons” (nH +) is somewhat higher than the binding constant for the “activating protons’’ (bH+)], we can introduce E, from Eq. (7) into Eq. (8) ( k ; / / k 1 )exp [2.3b ApH + (abF/RT) A$] V,= k3Eo (k;/k.J exp [2.3b ApH +(abF/RT) A$]
+1
(9)
With KO,=k;/k_,.K ; an equilibrium constant between E, and Ei under nonenergized conditions (i.e., at ApH = 0 and A$ = 0), and x = In Kg + 2.36 ApH + abF A$/RT, it results that V , = k,E, ex/(1 + ex)
(10)
For small values of ApH and A$ the second term in the denominator can be neglected so that
V, = k3EOeX
(1 1)
The salient results from this kind of modeling of ATP synthesis in the chloroplast thylakoid are: (1) There should be an exponential dependence of the rate of synthesis upon the energetic status of the membrane at low ApH and A$; (2) ApH and A$ are not equivalent kinetically, because only the fraction a of A$ influences the reaction rate. [This distinction, however, depends upon the choice of sites for protonation and hydroxylation; if both reactions occur at the same potential, then Eq. ( 5 ) (for hydroxyl ions) must be replaced by one similar to Eq. (4), with the result that a cancels out of Eq. (6). Equation (6) then takes the same form as shown above, but without a.] ( 3 ) The dependence of the rate of ATP synthesis on ApH (and A$) is determined by the dependence of the activation equilibrium on ApH. (This regulation mechanism is in accordance with corresponding experiments on the A$ dependence of activation and phosphorylation (Graber et al., 1977; Section VII.) Experiments examining the kinetics of ATP synthesis, in relation to this simplified theory, have been conducted with three methods of energization, by light, by artificial voltage pulses, and by artificial pH pulses.
B. Kinetics of ATP Synthesis by Light-Induced A p H and A$ The same experiment described above for the purpose of examining basal and phosphorylating proton fluxes-that is, repetitive-flash excita-
233
12. PHOSPHORYLATION IN CHLOROPLASTS
tion (Section III,B,2)-has yielded information on the kinetics of ATP synthesis by light-induced gradients. Specifically, as shown in Fig. 7, the rate of ATP synthesis displays an exponential dependence on ApH (linear in plots of the logarithm of rate versus ApH), with an apparent value of b lying between 2.2 and 2.6. One type of inverse experiment, yielding the basal and phosphorylating charge fluxes as functions of membrane potential, has also been partially described above (Section III,B,l), and the data plotted in Fig. 4 (top). In this figure, the phosphorylating flux for each ApH can be obtained by subtracting the basal flux (approximately equal to the flattest curve) from the total flux plotted. The resultant phosphorylating flux at constant ApH depends approximately exponentially on A$, as predicted by Eq. (1 1). With the logarithm of charge flux plotted against A$, slopes ( = ab) of about 1 are obtained. This result might be taken as evidence for kinetic dissimilarity of ApH and A$, but scattering of the data prevents a really definite conclusion on this point.
1.5
2 .o
2.5
3.0
APH FIG. 7 . Dependence of the rate of ATP synthesis on ApH, at different initial values of A$.
234
PETER GRABER
C. ATP Synthesis Kinetics Produced by an Artificial All, Again, the basic experimental technique for the experiments has already been described, in Section III,A,2 and Fig. 2. For technical reasons somewhat different conditions must be maintained for the class I1 chloroplasts in these experiments, compared with most others discussed in this article. In order to minimize Joule heating, the suspending medium is of rather low ionic strength and pulse lengths are normally limited to a maximal duration of 30 msec, spaced not closer than 30 seconds apart. The suspension is thermostated at 4°C. Each 30-msec pulse of a fixed amplitude generates a fixed amount of ATP (- 0.4 mmole ATP/mole of chlorophyll, for a field pulse of 1100 V/cm). When allowance is made for the polarization geometry of the vesicles (Section III,A,2), this figure is essentially identical with the yield from light pulses producing an equivalent A$. More importantly for the present argument, the relationship of the per-pulse yield of ATP to the amplitude of each voltage pulse is indeed superlinear. A typical set of data demonstrating this point is plotted in Fig. 8. However, it is obvious-by
I
06
I
*/
a,
25 d -
EE E
external electric field strength/%V FIG. 8. ATP generated by 30-msec field pulses as a function of the field strength (lower abscissa1 scale) and the transmembrane A$ (upper abscissal scale).
235
12. PHOSPHORYLATION I N CHLOROPLASTS
comparison with Fig. 3 or 4 (top), that the degree of superlinearity is low; it gives a value for ab of 0.3-0.4, rather than the 1 obtained in the light pulse experiments (Graber et al., 1977). The reason for this quantitative discrepancy is not yet clear, but two major factors may contribute: interaction between ApH and A$, there being no ApH in voltage pulse experiments and, more important, the strongly nonuniform polarization of the thylakoid membrane in voltage pulse experiments. For variable pulse lengths, at least down to 500 psec, there is a strict proportionality between the per-pulse ATP production and pulse duration at any fixed amplitude (Fig. 9). In order to amplify the sensitivity of ATP measurement in experiments with very short pulses, repetitive stimulation was used (Schlodder and Witt, 1981). An unexpected conclusion which can be drawn from the linear relationship in Fig. 9, therefore, is that the ATP synthetase can be activated well within the 500-psec pulse duration. Pulse stimulation by an applied field has a number of obvious technical advantages over light stimulation. The most important of these is the fact that brief pulses can generate a A$ without a ApH. This makes it possible to determine the dependence of (voltage-driven) ATP production on the external pH (pH,,,; see Fig. 6) without perturbing ApH, thereby providing a way to discriminate between the kinetic and thermodynamic effects of pH changes. Results of such a manipulation of pH,,,, for standardized field pulses, are plotted in Fig. 10 and are compared with similar ATP data obtained with periodic light flashes (Schlodder and Rogner, 1978). It is evident that ATP synthesis is independent of pH,,, over the range pH 6.5-9.0. These results justify the simplified description of the dependence of VATP on only one parameter, ApH, instead of on both pHin and pH,,, [see Eq. (7)-(1 l)]. They also indicate that no proton-involving reaction steps be-
0
2
’-
’c
/*’
0
-2 C
Q5-
El 0)
.->
/
external voltage pulse
$7.
FIG.9. Relative amount of ATP as a function of the duration of an external voltage pulse.
236
PETER GRABER
I a Y-
0
0 0
1.0-
c
C
3 0
E p)
* \
/
d
0.5-
> .c
o
0 d
2 01
light pulses voltage pulses
I
I
I
I
I
I
5
6
7
8
9
10
FIG. 10. Relative amount of ATP generated by light pulses and by external voltage pulses as functions of pH,,,.
tween the ATPase and the external and the internal medium are ratelimiting for ATP production.
D. Influence of an Artificial Aptc on the Kinetics of ATP Formation For purposes of kinetic analysis, it has been necessary to develop a quenched-flow version of the pH-jump experiment originally carried out by Jagendorf and Uribe (1966). This has been implemented (Smith el af., 1976) as diagramed in Fig. 11. A suspension of chloroplasts is incubated in an acidic medium for 30 seconds within syringe I . It is then injected, together with medium buffered at pH 8.2 (syringe 11), into the mixing chamber A and pumped along until it reaches mixing chamber B, where all reactions are quenched by trichloroacetic acid or ammonium chloride. By this device, the time between the alkaline shift and the quench can be reduced to several milliseconds, and in practice the ATP content of the quenched suspension is analyzed at intervals of about 20 msec. Synthesis is linear with time for at least several hundred milliseconds for ApH values greater than 2.3 and the slopes of such curves give directly the rate of ATP synthesis. The results of such measurements have been plotted in Fig. 12. The solid line has been calculated from Eq. (10) with Arl/=O, b= 1.5, In Kg = - 10.5, and k,E,,= 135 mmoles ATP mole-' Chl-' sec-I. At low ApH the function is exponential.
237
12. PHOSPHORYLATION IN CHLOROPLASTS
I
D
time
D
time
'R
I
1
0
tR
I
'
I
ApH =O incubation
I I
i
reaction
A
t
,
I
O P denaturation or uncoupling
B
FIG. 11. Diagram of the rapid acid-base transition experiment @H jump) in a quenchedflow system.
E. Interpretation of Thresholds and Lag Times for ATP Synthesis It has been known for more than 10 years that a critical level of ApH or A$ is required in thylakoids before any ATP synthesis can be observed (Schwartz, 1968; Schuldiner et al., 1973; Junge et al., 1970; Beyerle and Bachofen, 1978), and that the threshold for one gradient is diminished by increasing the other gradient (Graber and Witt, 1976). The expected temporal correlate of such a gradient threshold is a time lag, representing, at least, the period from the onset of a change (i.e., a light flash or pH jump)
238
PETER GRABER
until the energizing gradient reaches the required threshold. It is no surprise, therefore, that the quench-flow technique reveals a lag of 3-5 msec before ATP synthesis can be detected in the pH jump experiments. Figure 12 displays this result (Schatz et af., 1979). However, because the overall reaction of the ATP synthetase must have a finite cycle time, the quenching operation too must influence the apparent lag time. For pH jump experiments ATP synthesis is stopped by the quench, presumably with the same mixing time as synthesis is started by the jump. If the jump and quench have about the same time constant, therefore, the lag time should reflect the kinetic properties of the reaction cycle for ATP synthesis. In the case of the artificial voltage pulse experiment, no such lag occurs (Fig. 9), in part because the rise in potential can be very fast (= 10 psec), but also because pulse-off does not block ATP formation from molecules of the enzyme which have already formed the activated complex. The lag in ATP synthesis commonly observed for light-driven electron transport (Kahn, 1962; Ort and Dilley, 1976; Ort et al., 1976; Beyerle and Bachofen, 1978) must be considerably more complex, since it depends both upon A$ (which peaks in about 5 msec and then declines slowly) and upon ApH (which rises very slowly). Evidently, the temporal behavior of both gradients must be worked out very carefully before any conclusions can be drawn about the kinetics of ATP synthesis in light pulse experiments.
VI.
THE PROBLEM OF ENERGETIC SUFFICIENCY
Under the conditions used for most of the experiments discussed above, the free energy of hydrolysis of ATP, AG,,, can be estimated a t approximately 34.6 kJ/mole (360 mV), of which approximately 32 kJ/mole represents the standard free energy (for pH 8, 1 mM Mgz+;see Rosing and Slater, 1968) and the rest represents the ratio [ATP]/[ADP][PJ under starta I-
a c.
0 c
C
3 0
EJ
aJ
.->
c
0
2
duration of pulse /ms
FIG.12. Relative amount of ATP as a function of the duration of an external ApH pulse.
239
12. PHOSPHORYLATION IN CHLOROPLASTS
ing conditions. For proton-driven ATP synthesis to occur, the following thermodynamic relationship must hold:
~A,,~o n&H+ +AcA,p=n[RTln ( [ H ~ u , I / [ H ~ ~ I ) + ~ ( r l / ~ , , - r l / i , ) l + A (12) or -n(2.3 RTApH+FArl/)+AGA,,~O
(13)
in which n is the stoichiometric ratio H + / A T P for the enzyme, rl/,, . Thus the electrochemical potential ApH = pH,,, - pHin,and A$ = difference for protons (Ap,,) should exceed 180 mV if n = 2 or should exceed 120 mV if n = 3. In an acid-base transition or in continuous light a considerable rate of ATP synthesis is observed for ApH>2.5 (150 mV) as expected for n = 3 (Fig. 13). Voltage pulse experiments produce net ATP synthesis at average membrane potentials as low as 50 mV (Fig. 8) with zero ApH. The simplest way to account for this apparent conflict, given the likelihood that n lies between 2 and 3, is to suppose that at least some thylakoid vesicles are large enough to develop membrane potentials (at least at the pole facing the cathode) near the thermodynamically required value. Now the values of A$
I
150
A PH Rate of ATP synthesis in a rapid acid-base transition plotted as a function of the magnitude of ApH.
240
PETER GRABER
indicated on the top abscissa1 scale in Fig. 8 are average values for one-half the vesicle sphere, and the polarization geometry is such that the cathodal pole should have twice the average membrane potential. In addition, the size distribution of vesicles is 1 p m s radius I8 pm, with an average radius of 3-4 pm. Since A$ developed across the membrane by the external field is proportional to the vesicle radius, an average (half-sphere) A$ of 50 mV implies that some vesicles must have average potentials of 100 mV and maximal potentials at the cathodal pole of 200 mV. Such a value is quite sufficient t o drive net ATP synthesis with n = 3. Thus, the known morphological heterogeneity of the chloroplast preparations provides quite a reasonable way to account for ATP synthesis at apparently low values of A$ in the voltage pulse experiments. However, this argument absolutely requires some kind of mechanism for inhibiting the reverse-polarized (anodal) hemisphere of each vesicle from catalyzing ATP hydrolysis. [Such hydrolysis evidently occurs in comparable experiments carried out with the proton-driven ATP-synthesizing enzyme of thermophilic bacteria (Rogner et al., 1980).] Such inhibition may occur at two levels. First, removal of the ATPase inhibitor ( E subunit; see Section II,C and Table I) may require the same conditions as ATP synthesis, so that only those ATPase molecules which “see” a favorable gradient are actually able to carry out catalysis. Second, half-maximal hydrolysis of ATP is known to occur at an ATP concentration of about 0.6 mM, whereas half-maximal ATP synthesis occurs at an ADP concentration 20-fold lower (Buchholz, 1977); this means that under normal phosphorylating conditions the enzyme should always be saturated with ADP, not with ATP. Both these arguments of course are kinetic arguments, which allow the chloroplast ATPase or ATP synthetase to act as a kind of oneway valve, or rectifier. Morphological heterogeneity cannot account for the observation of Junge et al. (1970) that net synthesis of ATP from light pulses occurs at A$ z 70 mV; in this case energization of the membrane is homogeneous, regardless of vesicle size. Two explanations have been widely discussed, though as yet neither is completely convincing. The ApH associated with light flashes may be just large enough (1 pH unit; see Junge et al., 1979; Tiemann et al., 1979b) so that it sums with the A$ to approximate AG,,,, for n = 3. Or, at least some protons produced by the redox apparatus may be secreted into a restricted region of the ATPase rather than into the bulk internal phase of the thylakoid. Within such a region the effective ApH might develop faster and become higher than the volume of free aqueous phase could allow. Numerous arguments and observations have been cited in support of such preferential movement of protons (Ort and Dilley, 1976; Ort et al., 1976; Williams, 1961, 1978; Kell, 1979), but none of those experiments involved direct measurement of either A$ or ApH.
12. PHOSPHORYLATION IN CHLOROPLASTS
24 1
Although the issue of a restricted pathway for protons, from the redox system to the ATPase, must still be regarded as unresolved, it should be emphasized that A TP synthesis driven by an artificial ApH or an artificial A$, as described in Sections 111 and V, certainly takesplace with the rnovement of protons from the bulk inner aqueous phase to the bulk outer aqueous phase. Thus, regardless of whether or not a simple “Mitchellian” coupling always occurs during photosynthesis, experiments unequivocally demonstrate that it can occur. It remains to discuss an apparent contradiction: If, in the dark, the , critical level of actual ADP/ATP ratio is not equilibrated with A p H + the ApH and/or A$ can be interpreted t o reflect the minimal energetic requirement needed to overcome the actual AG,,, value. This has been done above. In Section V, however, the ApH dependence of the rate of ATP synthesis [Eq. (ll)] was attributed to the ApH dependence of the activation equilibrium, and the data shown in Fig. 13 have actually been fitted using Eq. (11). One of the main arguments for such a preceding activation is the observation that ATP hydrolysis requires a ApH+value of magnitude similar to that of ATP synthesis (ApH 2 2.5), and therefore it may be concluded that this value is connected with activation of the ATPase [Eq. (l)]. Which direction the subsequent reaction will take after activation (ATP synthesis or hydrolysis) depends on the sign of (nApH++ AGATp)[see Eq. (2)]. Thus, the threshold may be interpreted as a critical level of ApH+for activation of the ATPase. It is self-evident, in this case, that this critical level must be higher or equal t o the minimum ApH+necessary for ATP synthesis; i.e., bApH+ + AG,rnAp,,+ + AGATP(AGE free enthalpy change for activation). This might be an explanation that in chloroplasts under physiological condition ATP hydrolysis in the dark is inhibited. A quantitative description of the relation between activation and phosphorylation has been given elsewhere (Graber and Schlodder, 1980).
VII.
EPILOGUE: CONFORMATIONAL CHANGES ASSOCIATED WITH ENERGIZATION
The experiments described above really stand at the beginning of our quest into the fundamental molecular mechanism of proton transport-ATP synthesis mediated by membrane-bound F,-type ATPase. They have defined in detail the chemical conditions required for operation of the enzyme and have yielded considerable information about the kinetic behavior of the enzyme. And they have made unequivocally clear that-in the vector-
242
PETER GRABER
ially constrained environment of biological membranes-an electric field can be a very effective “chemical” reagent. With these things established, experimental attention has come to focus in the past few years upon the more difficult question of exactly what kind of structural changes the ATPase undergoes in the process of converting the energy of electrochemical gradients into the energy of phosphate anhydride bonds. Although no coherent picture has yet emerged in answer to this question, some provocative clues have been uncovered, which can be summarized here very briefly. That extensive conformational change must occur in the ATPase during the transition from energization to deenergization (or. vice versa) has been argued from observed large changes in chemical reactivity: H exchange between the enzyme and (tritiated) water occurs rapidly under energized conditions, but as many as 100 hydrogen atoms per CF, are not exchanged back under nonenergized conditions (Ryrie and Jagendorf, 1971, 1972); also, under energized conditions only, the sulfhydryl reagent N-ethylmaleimide reacts with a specific site on the y subunit of the ATPase (McCarty eta/., 1971; McCarty and Fagan, 1973). Furthermore, CF, is known to contain tightly bound adenine nucleotide (AMP, ADP, or ATP) at positions other than the catalytic site (Roy and Moudrianakis, 1971a,b; Strotmann et al., 1979). Although the role@) of this nucleotide in the mechanism of phosphorylation is not known, and it is not directly phosphorylated, energization of the chloroplast membrane allows rapid exchange with nucleotides in the external medium (Harris and Slater, 1975; Strotmann et a/., 1976). Most remarkably, this exchange tracks closely the kinetic characteristics of ATP synthesis (Section V) depending upon ApH and A$, whether these are generated artificially or by light pulses. Apparently only one nucleotide molecule is exchanged per molecule of CF,, by maximal energization (A$ or light); and the fraction actually released is an exponential function of the level of energization. These results have been interpreted (Graber et a/., 1977) on the basis of two plausible assumptions: (1) The amount of nucleotide released directly indicates the number of ATPase molecules which have changed conformation-presumably involving removal of the inhibitor subunit ( E ) ; (2) only these “active” ATPase molecules [E,; see Eqs. (8) and (lo)] can catalyze ATP synthesis. Then, referring to pH jump or artificial voltage pulse experiments, the ratio of ATP synthesized by the standard 30-msec pulse per molecule of E,*(that is, per adenine nucleotide molecule exchanged) is constant at about 6-independent of the magnitude of ApH or A$. The turnover rate and cycle time for each molecule of E,*can be calculated as -200 sec-1 and 5 msec, respectively. These results imply that the dependence of the overall rate of ATP synthesis upon ApH and A$ reflects primarily the recruitment of activated ATPases, which occurs under condi+
12. PHOSPHORYLATION IN CHLOROPLASTS
243
tions such that ATP synthesis, rather than hydrolysis, will occur (see Section V,A). Such rectifier or “gating” behavior is reminiscent of voltagedependent channel gating that has long been known in nerve membranes and has more recently been described in artificial membrane systems. In principle, it should be accompanied by intramembranal charge movement (gating currents). No such charge movements have yet been detected for chloroplast ATPase, but present results indicate a clear direction for future experiments. ACKNOWLEDGMENTS
I thank Prof. W . Junge, Dr. G. Renger, M. Rogner, Prof. B. Rumberg, G. H. Schatz, E. Schlodder, Dr. U. Siggel, R. Tiernann, and Ch. Underwood for helpful discussions and critical reading of the manuscript. I am most grateful to Prof. H. T. Witt for many stimulating discussions, valuable suggestions, and continuing support of this work, and to Prof. C. L. Slayman who has been a tremendous help in formulating this article.
REFERENCES Auslander, W., and Junge, W. (1975). FEBS Lett. 60, 310-315. Avron, M. (1960). Biochim. Biophys. Actu 40, 257-272. Baird, B. A., and Hamrnes, G. G. (1976). J. Biol. Chem. 251, 6953-6962. Baird, B. A,, and Hammes, G. G. (1979). Biochim. Biophys. Actu 549, 31-53. Bamberg, E., and Lauger, P. (1973). J. BioL Chem. 11, 177-194. Bennoun, P., and Levine, P. P. (1967). Plant Physiol. 42, 1284-1287. Beyerle, W., and Bachofen, F. (1978). J. Biochem. 88, 61-67. Binder, A,, Jagendorf, A. T., and Ngo, E. (1978). J. Biol. Chem. 253, 3034-3099. Boeck, H., and Witt, H. T. (1972). Proc. Int. Congr., 2nd Photosynthesis Res., Stresu pp. 903-911. Boyer, P. D., Chance, B., Ernster, L., Mitchell, P., Racker, E., and Slater, E. C. (1977). Annu. Rev. Biochem. 46, 955-1026. Buchholz, J. (1977). Thesis, Technische Universitat, Berlin. Cantley, L. C., and Hammes, G. G. (1976). Biochemistry 15, 1-8. Deters, D. W., Racker, E., Nelson, N., and Nelson, H. (1975). J. Biol. Chem. 250, 104I - 1047. Graber, P., and Schlodder, E. (1980). Proc. Int. Congr. 5th Photosynthesis, Kallithea, Greece. GrBber, P., and Witt, H. T. (1974). Proc. Int. Congr. 3rd, Photosynthesis, Rehovot, Israel. pp. 427-436. Graber, P., and Witt, H. T. (1975). FEBS Lett. 59, 184-189. Graber, P., and Witt, H. T. (1976). Biochim. Biophys. Actu 423, 141-163. Graber, P., Schlodder, E., and Witt, H. T. (1977). Biochim. Biophys. Actu 461, 426-440. Harris, D. A., and Slater, E. C. (1975). Biochim. Biophys. Actu 387, 335-348. Hauska, G., and Trebst, A. (1977). In “Current Topics in Bioenergetics” (D. R. Sanadi, ed.), Vol. 6, pp. 151-1220. Academic Press, New York. Howell, S., and Moudrianakis, E. (1967). Proc. Nut/. Acud. Sci. U.S.A. 58, 1261-1268.
244
PETER GRABER
Jagendorf, A. T., and Uribe, E. (1966). Proc. Natl. Acad. Sci. U.S.A. 55, 170-177. Joliot, P., and Delosme, R. (1974). Biochim. Biophys. Acfa 357, 267-284. Junge, W., and Witt, H. T. (1968). Z. Natarurforsch. 23b, 244-254. Junge, W., Rumberg, B., and Schroder, H. (1970). Eur. J . Biochem. 14, 575-481. Junge, W., Auslander, W., McGeer, A. J., and Runge, T. (1979). Biochim. Biophys. Acta 546, 121-141. Kahn, J. S. (1962). Arch. Biochem. Biophys. 98, 100. Kell, D. B. (1979). Biochim. Biophys. Acfa 549, 55-99. McCarty, R. E. (1979). TIBS 4, 28-30. McCarty, R. E., and Fagan, J. (1973). Biochemistry 12, 1503-1507. McCarty, R. E., Fuhrman, J. S., and Tsuchiya, Y. (1971). Proc. Nafl. Acad. Sci. U . S . A . 68, 2522-2526. Mitchell, P. (1961). Nature (London) 191, 144-148. Mitchell, P . (1966). Biol. Rev. 41, 445-502. Mitchell, P. (1968). “Chemiosmotic Coupling and Energy Transduction.” Glynn Research, Bodmin, England. Nelson, N. (1976). Biochim. Biophys. Acta 456, 314-338. Nelson, N., Nelson, H., and Racker, E. (1972). J . Biol. Chem. 247, 7657-7662. Nelson, N., Deters, D. W., Nelson, H., and Racker, E. (1973). J. Biol. Chem. 248, 2049-2055. Nelson, N., Eytan, E., Notsani, B., Sigrist, H., Sigrist-Nelson, K., and Gitler, C. (1977). Proc. Nafl. Acad. Sci. U.S.A 74, 2375-2378. Nelson, N., Nelson, H., and Schatz, G. (1980). Proc. Nafl. Acad. Sci. U . S . A . 77, 1361-1364. Ort, D. R., and Dilley, R. A. (1976). Biochim. Biophys. Acta 443, 95-107. Ort, D. R., Dilley, R. S., and Good, N. E. (1976). Biochim. Biophys. Acta 449, 108-124. Paradies, H., Zimmermann, J., and Schmidt, U. D. (1978). J. B i d . Chem. 253, 8372-8979. Pick, U . , and Racker, E. (1979). J. Biol. Chem. 254, 2793-2799. Portis, A. R., and McCarty, R. E. (1976). J. B i d . Chem. 251, 1610-1617. Racker, E. (1978). TIBS 1, 244-247. Reeves, S. G., and Hall, D. 0. (1978). Biochim. Biophys. Acfa 463, 275-297. Rogner, M., Ohno, K., Hamamoto, T., Sone, N., and Kagawa, Y. (1950). Biochem. Biophys. Res. Commun. 91, 362-367. Rosing, J., and Slater, E. C. (1972). Biochim. Biophys. Acfa 267, 275-286. Roy, H., and Moudrianakis, E. N. (1971a). Proc. Nufl. Acad. Sci. U.S.A. 68, 464-468. Roy, H., and Moudrianakis, E. N. (1971b). Proc. Nafl. Acad. Sci. U.S.A. 68, 2720-2724. Rumberg, B., and Siggel, U. (1968). Z. Naturforsch. 23b, 239-244. Rumberg, B., and Siggel, U. (1969). Nafurwissenschaften 56, 130-132. Ryrie, I., and Jagendorf, A. T. (1971). J. Biol. Chem. 246, 3771-3774. Ryrie, I., and Jagendorf, A. T. (1972). J. Biol. Chem. 247, 4453-4459. Schatz, G. H., Schlodder, E., and Graber, P . (1978). Biophysiktagung Ulm F, 12. Schatz, G. H., Schlodder, E., Rogner, M., and Graber, P. (1979). Annu. Meet. Dtsch. Ges. Biophys., Konsfanz B, 49. Schlodder, E., and Rogner, M. (1978). Biophysiktagung Ulm F, 3. Schlodder, E., and Witt, H. T. (1981). Biochim. Biophys. Acta, 635, 571-584. Schmidt, U. D., and Paradies, H . M. (1977). Biochem. Biophys. Res. Commun. 78, 1043- 1052. Schroder, H. (1974). Thesis, Technische Universitat, Berlin. Schroder, H., Muhle, H., and Runberg, B. (1972). Proc. Int. Congr., 2nd, Photosynthesis Res., Stresa 1971, pp. 919-930. Schuldiner, S . , Rottenberg, H., and Avron, M. (1972). Eur. J. Biochem. 25, 64-70.
12. PHOSPHORYLATION IN CHLOROPLASTS
245
Schuldiner, S., Rottenberg, H., and Avron, M. (1973). Eur. J. Biochem. 39, 455-463. Schwartz, M. (1968). Nature (London) 219, 915-919. Sigrist-Nelson, K., Sigrist, H., and Azzi, A. (1978). Eur. J. Biochem. 92, 9-14. Smith, D. J . , Stokes, B. O . , and Boyer, P. D. (1976). J . Biol. Chem. 251, 4165-4171. Stiehl, H. H., and Witt, H. T. (1969). Z . Naturforsch. 24b,1588-1598. Strotmann, H., Bickel, S., and Huchzermeyer, B. (1976). FEBS Lett. 61, 194-198. Strotmann, H., Bickel-Sandkotter, S., Edelman, K., Eckstein, F., Schlimme, E., Boos, K. S., and Liistorff, J. (1979). Biochim. Biophys. Acta 545, 122-130. Tiemann, R., Renger, G., and Graber, P. (1979a). Annu. Meet. Dtsch. Ges. Biophys., Konstanz, B, 49. Tiemann, R., Renger, G., Graber, P., and Witt, H. T. (1979b). Biochim. Biophys. Acta 546, 498-5 19. Trissl, H.-W., and Graber, P. (1980). Bioelectrochem. Bioenerg. 7, 167-186. Williams, R. J . P. (1961). J . Theor. Biol. 1, 1-13. Williams, R. J . P. (1978). Biochim. Biophys. Acta 505, 1-44. Winget, G . D., Kanner, N., and Racker, E. (1977). Biochim. Biophys. Acta 460, 430-499. Witt, H. T. (1979). Biochim. Biophys. Acta 505, 427. Witt, H. T., Schlodder, E., and Graber, P. (1976). FEBS Lett. 69, 272-276. Younis, H. M., Winget, G. D., and Racker, E. (1977). J . Biol. Chem. 252, 1814-1818.
This Page Intentionally Left Blank
Part IV
Some Theoretical Questions
This Page Intentionally Left Blank
CURRENT TOPICS IN MEMBRANES AND TRANSPORT. VOLUME 16
Chapter 13
Response of the Proton Motive Force to the Pulse of an Electrogenic Proton Pump ERICH HEINZ Department of Physiology Cornell Medical School New York, New York
1. 11.
Introduction ......................................................................................... Treatment in Terms of the Thermodynamics of Irreversible Processes ................. References ............................................................................................
I.
249 250 256
INTRODUCTION
An electrogenic proton pump builds up a “proton motive force” (PMF= - A p H + ) by two distinct contributions: ( 1 ) the direct or primary contribution resulting from the forced movement of H + ions through the barrier, ahead of the electroneutralizing passive ion flows, and leads to the separation of charges and, by charging the capacity of the membrane, to an electric potential difference (PD). (2) The indirect or secondary contribution which involves the electroneutral movement of ions, to build up concentration gradients, often associated with a membrane diffusion PD. As the primary contribution does not require appreciable net transport of ions, it is almost exclusively electric and therefore appears rapidly after the pump is turned on, and just as quickly disappears when the pump is turned off. The secondary contribution, on the other hand, requires the expenditure and storage of osmotic work and therefore appears much more slowly after the start of the pump; and its decay after stopping the pump is also delayed. 249
Copyright $ 1982 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-1533 16-6
250
ERlCH HElNZ
The question arises whether the initial rise in electric PD, hence in proton motive force, attributable to the primary contribution, is of such magnitude as to anticipate a major fraction of the maximum P M F ultimately generated by the pump, to make it available for useful work at an early stage. Mitchell (1968) has calculated on the basis of experimental values applied t o a model of constant pumping rate that the electric P D indeed rises rapidly after the onset of the pump, but to such a low value as would be insignificant. Hence the full rise of thk PMF is largely determined by the much slower generation of a H + gradient by the electroneutral secondary contribution. Whereas present calculations confirm this, they also show that with a different model, in which not the pumping rate but the affinity of the driving (redox) reaction is to remain constant during pumping activity, the rapid initial rise of the P D may under certain conditions reach a value which comes close to the final P MF (Fig. 1, see p. 252). This will be illustrated for an arbitrary model of a proton pump similar to that assumed for the membrane of mitochondria, akaryotic microorganisms, chloroplasts, etc. We assume that this pump can be suddenly turned on and off at will, so that a square wave-like pumping pulse can be produced during which either the affinity (Ach)or the rate (J,) of the driving (redox) reaction remains constant. For simplicity it is also assumed that (1) besides H + , only K’ is able to leak across the membrane, and (2) the concentrations of K + in both compartments are so high that they d o not significantly change during the pumping pulse.
II.
TREATMENT IN TERMS OF THERMODYNAMICS OF IRREVERSIBLE PROCESSES
The subsequent derivations, as far as the response of the proton motive force at the onset is concerned, are partly based on Mitchell’s procedure (Mitchell, 1968), though in line with thermodynamic usage the forces (X,) are expressed here in joules per mole rather than in volts. Accordingly, both the electric capacity and the buffer capacity have been converted here from the conventional units into moles squared per joule per gram of protein (Kedem and Caplan, 1965). In the derivation we make convenient use of the “quasi-chemical approach” (Heinz, 1974), which treats the coupled process like a chemical reaction vHH++ CviSi= v,H+
+ CvjPj
13. RESPONSE OF
PMF TO PULSE OF ELECTROGENIC PROTON PUMP
251
where vH, v i and vj are stoichiometric coefficients, and Si and Pj represent substrates and products, respectively, of the driving (redox) reaction. The superscripts ’ and ” refer to the outside and inside compartments, respectively. The rate of the coupled process is accordingly Jr=
Lr(VHXH-Ach)
(la)
in which
X,= -ApH+ = -RTAln[H+]-FA$
(1b)
the thermodynamic correlate of the PMF and A,, is the effective affinity of the driving (redox) reaction (C&- CvjPj). The negative sign indicates that A,, tends to move H + in the outward-by definition negative-direction. L, is the phenomenological coefficient of the overall process. The total flow of H + contains, besides the coupled flow (vHL~),also the leakage flows, which for simplicity are treated as proportional to XHwith an overall leakage coefficient L;: J H = vHJr+
LbXH
(2a)
The flow of K + , as it is assumed not to be coupled to the pumping process, occurs only by leakage: J K = LkXH
(2b)
and
XK=-ApK= -RTAln[K+] -FA$
(2c)
where L; is the leakage coefficient. If on the other hand, the driving affinity A&, rather than J,, were constant during the pulse, we would have to insert Eq. (la) for J, into Eq. (2a) to obtain JH=(vj$L,+ L#)XH-VHL,A,,
(3)
J K = LaXK
(4)
The change in X,, i.e., of the PMF with time ( t ) is the sum of the changes of the chemical and of the electrical term, respectively: dXH/dt= - (dRT Aln[H+]/dt) - (Fd A$/dt) R T d(Aln[H +I ) /dt = JH/B
F d(A$)/dt
= (JH
+ JK)/C
(5)
(64 (6b)
where B is the effective buffer capacity, which depends on the buffer capacities of the adjacent solutions (l/B= 1/B’+ l / B ” ) and C is the electrical capacity of the membrane, both referring to grams of protein.
H+-pump ''on"
H+-pump " o f f "
____ -RTAlnlH+I X,(PMF)
-D
Time
FIG. I . Response of the P M F t o a pulse of an electrogenic H + pump. The time courses of the P M F (dotted line) and its components-the electrical P D (solid line) and the chemical P D (dashed line)-are plotted in response to a square wave pulse of the H + pump. Abscissa: Time, in arbitrary units; time scales are discussed below. Ordinate: the P M F and its components, expressed as percentage of the maximal steady state value when the pump is on. (A) Constant pumping rate during the pulse. The rate coefficients for passive movements (leakage) of H + (LH)and K + ( L K ) ,buffer capacity (B), and electrical capacity ( C )are the same as those used by Mitchell (1968). The rate coefficient for K + movement is smaller, however, than would correspond to the sum of rate coefficients for all permeant ions (other than H'), as used by Mitchell. This change is made so that the initial rise in electric P D (and its terminal drop) will be visible in the graph. Hence the peaks of the electric P D could-in experiments-be considerably lower than depicted here. In this figure, the peak of electric P D occurs at 400 msec, and the half-time for maximal chemical P D is 100 seconds. (These values compare with Mitchell's values of 200 msec and 80 seconds, respectively). (B) Constant driving force during the pulse. Values of L H ,LK, B , and C a r e the same as those above. L,, the overall pump coefficient, is assumed t o be 3LK. In this model, the peak electric P D is rather independent of the ratio LK/LH,as long as L, is sufficiently greater than LK.The times for peak electric P D and for the half-maximal steady state P M F are somewhat shorter than above: 200 msec and 20 seconds, respectively. (It should be pointed out, however, that more than 90% of the peak electric P D is reached at a time far shorter than 100 msec.)
13. RESPONSE OF
PMF
TO PULSE OF ELECTROGENIC PROTON PUMP
253
Depending on whether the pumping rate (J,) or the driving affinity (Ach) is to be treated as constant, JH and JKhave to be replaced according to Eq. (2a) and (2b) or to Eq. (3a) and (3b), respectively. Thus we obtain for constant J, d(RTAln[H+])/dt= (1/B) (-vHJr+LhXH) (74 d(FAll/)/dt=(I/C) [(-vHJ,+L,UXH)+LEXK] and for constant A,, d(RTAln[H+])/dt= (l/B) [(v2,L,+Lh)XH- vHLrAch]
(7b) (84
d(FA$)/dt= (1/c) [(VhL,+ L&)XH+ &XK-VHL,A,.~]
(8b) For further derivation we express XHand X , by their chemical and electrical terms, Eqs. (lb) and (2c). Because in our model [K+]is assumed to be very large in both compartments (as is true for mitochondria in their cellular environment), the chemical form of X , , RT Aln[K+] has been neglected. Hence for either set we arrive at two differential equations with two dependent variables only, of the general form dx/dt= -(a'x+a'y-c')
(9a)
dy/dt = - (a"x+ b"y- c")
(9b)
in which x a n d y stand for Rt Aln[H+]and FAG, respectively, and a ' , a", b", c' , and C" are constants. Similar equations can be derived for the same variables after the pump is suddenly turned off at the end of the pulse. Complete integration of these equations, for constant pumping rate (k) only, has been carried out by Mitchell (1968); the integration for constant driving affinity (Ach)is quite analogous. The resulting equations in either case are very involved, but under the conditions postulated for our model some simplifying approximations are permitted so that the following solutions adequately describe the events. For the onset of the pumping pulse RTAln[H+]= -Al(l FA$
=
-e-x2')
-A2(e-Xi'- e-h')
where t is the time after the pumping has been turned on. The corresponding equations for termination of the pumping pulse are RTAln[H+]= -Ale-X' FA$
= A2(e-xir- e-A29
(1 la) (1 1b)
Here t'is the time after the pump has been turned off after reaching static head. The parameters A l , A 2 ,XI, and X2 are at least formally identical for both onset and termination of the pump if the notation of TIP is applied.
254
ERlCH HElNZ
Depending on whether Ach or J, is to be treated as constant during the pumping pulse, these parameters have the values shown in Table I. Inserting these parameters into the corresponding equations, we can draw the following conclusions: At either constant J, or constant Ach, the initial as well as the terminal change in electric P D is faster than the corresponding changes in chemical PD, especially if B$ C [for liver mitochondria B is estimated t o be about 50x C (Mitchell, 1968)l. In the extreme, the two changes may almost occur in two distinct subsequent phases, the maximum electric P D being reached before the chemical P D can change significantly. There are, however, fundamental differences between the two conditions, which may be significant biologically. At constant J, our equations confirm those of Mitchell (Mitchell, 1968). The primary contribution, i.e., the initial rise in electric PD, though fast, is transient and very small and may be insignificant, especially if L;;a L i , i.e., if the mobility of the passively permeant ions (here K + ) greatly exceeds that of H + . Hence the rate at which the PMF develops is in this case largely determined by the (slower) rise of the chemical P D due to the secondary contribution. We may further predict that, after turning off the pump, the initial change in electric PD occurs fast but is of insignificant magnitude so the dissipation of the PMF is delayed by the slower, electroneutral ion movements (Fig. 1A). At constant A,, the situation may be much different, especially if the pump coefficient Lr is much greater than L;+ Lf,. First, the difference in TABLE la J, = constant
L;;
x,
=
Ach= constant
+ L#
L#L#
B(L# + L#) A2 =
+
L;; L# C
(I The constants for constant Jr become identical with the corresponding ones of Mitchell (1968) if in analogy with his model LK+L#.
13. RESPONSE OF
PMF
TO PULSE OF ELECTROGENIC PROTON PUMP
255
rate between the two phases is more pronounced, and the initial rise in electric P D (primary contribution) may be completed in less than 100 msec. In the second phase, as [K+] is assumed to be high, the P D declines slowly toward zero, whereas at a similar rate, RT Aln[H+] rises (secondary contribution), reaching its static head value (A,) only after seconds. If Lr+ L,+ L,, it follows further that the peak of the first phase A , is not much lower than A,. This would be so even if the passive permeability of H + were much smaller than that of other ions, here of K + , as long as the latter is sufficiently exceeded by L,. As a consequence, most of the maximum PMF (X,) available from the pump is already “anticipated” as an electric PD, at a very early time, long before the H + gradient has been built up. Similarly, after turning off the pump a rapid change in electric PD, corresponding to the loading of the membrane capacity (C) in the opposite direction, and thus to a decrease in the PMF, precedes the much slower dissipation of the ion gradients. Hence an electrogenic pump may under the above conditions make much of the power of the pump rapidly available for useful work, and after termination just as rapidly withdraw it, in either case well prior to appreciable ion transport (Fig. 1B). The behavior of the electrical P D in response to a pumping pulse resembles the electric induction phenomenon: The onset as well as the interruption of the pump activity causes a transient appearance of an electric PD, which opposes the pump at the onset and favors it at the termination. [It should be kept in mind that the complete decline of the electrical P D after the first rise is a consequence only of our simplifying assumptions that K + is so much in excess over H + (in spite of effective buffer capacity) that concentration changes in K + can be neglected. This may not be quite true for mitochondria whose effective buffer capacity for H + is significant and may further be reinforced by a K + / H + antiport mechanism. So in reality the P D may still have a sizeable value at steady state, and the chemical P D of H + will accordingly be lower at steady state. This argument, however, is irrelevant with respect t o the major point of this treatment and is, therefore, neglected.] Presumably neither model perfectly describes reality. However, as the rapid availability of a PMF at the beginning of a pumping pulse, as well as a rapid withdrawal of this force at the end, appear to be advantageous for regulatory purposes, it is tempting to favor the constant A,, model over the constant J, model in this electrogenic pump. Whether L,, the coefficient of the intrinsic pumping rate, is as fast as is postulated above, can presumably not be decided from ion flows, which for the maintenance of electroneutrality are necessarily limited by the slower species. Only direct measurement of the change in electric P D might give the answer.
256
ERlCH HElNZ
ACKNOWLEDGMENTS These studies were supported by U S P H S NIH grant ROI GM 26554-01. I wish t o thank Dr. S. Rubinow of the Division of Biomathematics for his helpful criticism and review of the manuscript.
REFERENCES Heinz, E. (1974). In “Current Topics in Membranes and Transport” (F. Brown and A. Kleinzeller, eds.), Vol. 5, p. 137. Academic Press, New York. Kedem, O . , and Caplan, S. R. (1965). Trans. Furuduy SOC. 61, 1897. Mitchell, P. (1968). “Chemiosmotic Coupling and Energy Transduction” Glynn Research, Bodmin, England.
CURRENT TOPICS IN MEMBRANES AND TRANSPORT, VOLUME 16
Chapter 14 Reaction-Ki netic Ana Iys is of Current-Voltage Relationships for Electrogenic Pumps in Neurospora and Acetabularia DIETRICH GRADMANN,' ULF-PETER HANSEN,2 A N D CLIFFORD L . SLA YMAN Department of Physiology Yale School of Medicine New Haven, Connecticut
I. Introduction .... ......................... .................................... 11. Theory: Reducti Models ........................................................ A. A Five-State Model ......................................................................... B. The Two-State Model ...................................................................... C. Determination of Model Parameters from I- VCurves ............................. D. Interpretation of Parameters: Comparison of the Two-State Model ...................................................... with n-State Models ....... 111. Results ............................................................................................... A. Neurospora ................................................................................... B. Acetabularia ............................ ........... ... C . Discussion: Localization of the Energy Shift ......................................... 1V. Extensions of the Model ......................................................................... A. Relation of Gradient-Driven Transport to Active Transport ...................... B. Influence of Multiple Charges and Multiple Charge Transfer Limbs ........... C. Unstirred Layers and Asymmetric Potential s ............................... D. Approximation of Pumps by Ideal Sources ................................. References ....................................................................................
258 258 258 260 264 265 266 266 270 272 213 273 274 274 274 276
I Present address: Max-Planck-Institut fur Biochemie, Abteilung Membranbiochemie, Miinchen, Federal Republic of Germany. Present address: Institut fur Angewandte Physik, Universitat Kid, D2300 Kid, Federal Republic of Germany.
'
257
Copyright 0 1982 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0- 12- I533 16-6
258
DIETRICH GRADMANN
I.
et a/.
INTRODUCTION
Enzyme-mediated ion transport processes constitute a special class of reactions which are subject to two separate sets of constraints: those of enzyme kinetics, and those of electric circuit theory. Since the electrical properties of many biological membranes-and, under special conditions, of specific ion transport systems-are relatively easy to measure, the question naturally arises as to how the electrical behavior of a transport system can illuminate underlying reaction mechanisms. It is clear that electrical measurements are required in order to prove that a given transport system is electrogenic; but the question of what else can be learned from electrical measurements on intact biological membranes is largely unresolved, at least for active transport systems. While considerable progress, both theoretical and experimental, has been made in understanding the kinetics of charge flow through passive carriers and channels (Shamoo, 1975), most electrophysiological work on active transport has been restricted either to pure description or to thermodynamic analysis. The main purpose of this discussion therefore is to explore the functional dependence of current flow through electrogenic ion pumps upon the magnitude of the imposed membrane potential and then to relate this socalled current-voltage relationship (I- V relationship) to the predicted behavior of certain kinds of kinetic models for transport. For the sake of simplicity, we shall restrict consideration to what may be termed Class-I electrogenic models: those having a single limb in which charges are transferred across the membrane. Furthermore, since the extraction of kinetic information from I- V relationships depends strongly on observable nonlinearities, we shall restrict the presentation of experimental results to the two electrogenic ion pumps for which nonlinear I- Vrelationships have been documented: ATP-driven extrusion of protons across the plasmalemma of the ascomycete fungus Neurospora crassa (Gradmann et a f., 1978) and the electrogenic uptake of chloride by the unicellular green marine alga Acetabularia mediterranea (Gradmann, 1975).
II. THEORY: REDUCTION OF KINETIC MODELS A. A Five-State Model Figure 1A shows a somewhat arbitrary reaction scheme for a cyclic, ATP-driven proton extrusion pump, such as might exist in the plasmalemma of Neurospora and other fungi, or of freshwater algae and higher plants. It may be viewed as a kind of minimal model, having the smallest
14. KINETIC ANALYSIS OF PUMPS
A
B out
H
259
out
in
in
+-
+-
I
FIG.1. A five-state reaction- kinetic model for an electrogenic proton pump (Class-I). The reaction cycle, driven by the free energy of hydrolysis of ATP, normally occurs in a clockwise direction. Dephosphorylation of a highenergy phosphoprotein during the charge transport step is assumed for reasons discussed in Section III,A and B. In (A) the protein carrier is designated by X. In (B)the membrane density of each form of the carrier is given by N:, N;, NZ, etc. The designated rate constants kl2,k23,etc., represent the product of the true rate constants multiplied by the appropriate reactant concentrations [ATP, ADP, H+, and Pi (inorganic phosphate)].
number of steps likely to be consistent with the enzymatic properties of active transport systems. The “carrier” exists in five distinct states, connected by five reversible reaction steps, so that the normal pump cycle can be described as follows. The uncharged carrier at the inner surface of the membrane (X, top right, Fig. 1A) is first phosphorylated and then binds a proton; subsequently, it undergoes a conformational change o r a transit which moves charge across the membrane, simultaneously releasing the phosphate and a large amount of free energy. The charged carrier at the external surface of the membrane (XH +,lower left, Fig. 1A) then releases the proton and later becomes accessible for rephosphorylation at the membrane inner surface, so continuing the cycle. [The carrier concept in this reaction scheme is of course only a convenience. The kinetically important features of the model are the number of states, the number of distinct rate constants, and the order in which energy and electric charge are bound, transferred, and released. The whole scheme can be written and analyzed without reference to carriers per se (Lauger, 1979).] Figure 1B defines the specific rate constants ( k l z ,k21,k23,etc.) and the
260
DIETRICH GRADMANN et
a/.
carrier densities (N:, N;, N3,etc.) associated with each step in the reaction cycle. The explicit concentrations of chemical ligands [H+, ATP, ADP, and inorganic phosphate (Pi)] have been omitted, since they always appear-in formal relationships-as products with their associated rate constants. (Thus, the concentrations of protons are contained within k,, and k,,, the concentration of Pi within k Z l ,that of ATP within k45,and that of ADP within k54).As noted in Section I, charge transfer across the membrane is assumed to occur in only one pair of reactions, the transition between N; and N;. The voltage dependence of charge transfer is then obtained by considering the membrane to be a single energy barrier to the passage of charge (Lauger and Stark, 1970). If this barrier is symmetric in the membrane, then the first two rate constants in the diagram of Fig. 1B can be written as k,, = ky2eu/2
and
k,, = k&e-u/2
(1)
where the superscript zero designates the transition rate constants which should be observed with zero membrane potential, and u is the reduced membrane potential:
u = zFV,,,/R T
(2)
Current flow (i) through the Class-I transport system is just the difference between the outward flux of N;and the inward flux of N;:
i= zF(N,+k,,- NZ+k2,) = ZF(N;~P,~"/~N+ko 2 21e-u/2)
(3)
And if all the rate constants plus the total density of carrier N ( = N:+ N2++N, + N4+ N J were known, the steady state current-voltage relationship for the electrogenic pump could be written explicitly. Since, however, no active transport system has yet been described with all the rate constants evaluated (or even identified), the practical problem we are confronted with is the inverse one: to evaluate some of the reaction rate constants from measured current-voltage relationships. And for this purpose it is absolutely necessary to simplify the model mathematically (Lieb and Stein, 1974). 6. The Two-State Model
Provided that only a single reaction step in the cycle actually carries charge across the membrane, and that steady state conditions for the overall cycle are assumed, all n-state models ( n = 3 , 4, 5 , 6, . . . ) of the type depicted in Fig. 1 can be reduced to the algebraic form of a two-state model, at least for the purpose of current-voltage curve analysis. A formal proof of this fact will be given elsewhere (Hansen et af., 1981) but it follows
14. KINETIC ANALYSIS
261
OF PUMPS
mathematically from the fact that the total density of carrier in the membrane can be written as N = AN:+ BN,+
(4)
where A and B are algebraic functions of the voltage-insensitive reaction rate constants. Figure 2, then, depicts the reduction of the five-state model to a twostate model. All the voltage-insensitive steps are lumped together and assigned the psuedo-rate constants X . The k’s are retained for the voltagesensitive rate constants, but the mnemonic subscripts i (inside) and o (outside) are substituted for the carrier designators 1 and 2. [It should be noted that in general the X ’ S and the k’s in the two-state model contain terms with A and B (see above); the consequence of this will be dealt with in Section II1,C.l The differential equation for the two-state model can be written as dNi/dt= - (kio+ xi0)N;+ (koi+ xoi)N0f= 0
(steady state)
(5)
This, coupled with the relationship for the total carrier (N=1vI++NO+), can be combined with Eq. (3) t o give an explicit function for the membrane current in the two-state model: i = zFN
kioxoi - koixio = zFN k&xoieu’2 - kE,Xoie-u/2 k&e’/2 + k$e-u’2 + xio+ xOi kio+ koi+ x~~+ xOi
(6)
B
A out
out
in
in Ko i
?ATP
-
A H + ADP
-+
+I
koi
N,+-N+ kio
I
262
DIETRICH GRADMANN
Membrane current / A m-*
Internal potential / m V
et a/.
t
0.05
100
200
300
15
._
0
FIG.3. Current-voltage relationships for a Class-I electrogenic pump. Effects of changing magnitudes of voltage-sensitive rate constants ( k ) relative to the voltage-independent rate constants (k). All three curves were calculated from Eq. (6). Passive carrier system, with no net change in free energy. Conditions: N = mole/m2; k: = kEi = k o (in sec-I); xoi = xio = 102 sec-'.
A few properties of this function are illustrated in Figs. 3 and 4. [For the sake of numerical calculation, the total carrier density in the membrane ( N ) is assigned the specific value mo1e/m2 or 6000 sites/pm*, and the stoichiometric coefficient (2) is put equal to 1.1 Figure 3 shows some current-voltage relationships for purely passive processes which neither absorb nor release net energy in a complete cycle. For convenience, the rate constants are arranged symmetrically (k: = kEi, and xio= xoi), although the general condition for zero free-energy change is somewhat weaker (kyoxoi= kEixio).Since the membrane potential is the only net driving force, all curves pass through the origin, with zero current at zero voltage. The detailed shapes of the curves, however, depend upon the relative magnitudes of the two paris of rate constants. With k: and kzi substantially greater than xio and xoi, the overall I- V curve (ko= lo3, Fig. 3) is sigmoid and approaches a hyperbolic tangent function. At limiting steepness the voltage displacement (from the origin), which yields half-saturating current, is 28/2 mV.3 The overall shape of the current-voltage relationship becomes more complicated, with three inflection points, when k: and kEi fall significantly below xoiand xio, as shown by curve ko= 10 of Fig. 3. Also, as k: and kEi become smaller, saturating current is achieved only at progressively larger membrane potentials, in either direction. The voltage at half-saturating The similarity in shape, from the origin to either saturating current, between this curve and
a Michaelis function is obvious. A strict formal relationship between the two also becomes apparent when the exponential functions of voltage (e''2, etc.) are considered analogous to chemical activity coefficients.
14. KINETIC ANALYSIS
OF PUMPS
263 Membrane current / A m-'
A
[1
- - 0 15 L-0.15
-
-800
-600
zoot
J, ,
,
n 100
xx)
500
700
900
Two - s t a t e model N = lo-' mol m-2
Conditions:
E, ( k ' )
E , ( K ) = -460mV
ref = reference: k c = kzi =
K ~ =, ~i~
ko=asymmetry in k o ' s : k70 = Koi
K
=asymmetry in
K'S :
k7, Koi
= 102 s- 1
lo6,kzi Kio
=1 0'
= fo-'s-' 5.'
kzi = 10' S-I = lo6,Klo = IO-'S-' 2
FIG. 4. Current-voltage relationships for Class4 electrogenic pumps. (A) Effects of adding inergy either in the voltage-sensitive limb (curve k o ) or in the voltage-independent pathway ( x ) . (B and C) The currents at extreme positive and extreme negative membrane potentials, when the energy enters the voltage-independent pathway.
current can be considered a characteristic parameter for the shape of such I- V curves, and this value moves 58/z mV (the limit for small values of kO) further from the origin for each 10-fold reduction in the product kzkgi. All these features are consequences of the fact that a very small value of kg or k:i requires a very large value of efUI2in order for the cycle reaction rate (clockwise or counterclockwise) to be limited by the voltage-insensitive rate constant xioor xOi. For active transport (or its reverse, gradient-driven ATP synthesis),
264
DIETRICH GRADMANN
ef a/.
where there is a net change in free energy for a single cycle, the picture can change radically from that in Fig. 3, depending on which reaction step absorbs or releases the energy. For the examples in Fig. 4, a free energy change equivalent to - 470 mV is assumed, as might represent the behavior of an ATP-driven transport system having a stoichiometry of one charge per one ATP molecule split. The “ref” curve in Fig. 4 represents the passive case, with all rate constants set equal to 100 sec-I. When the energy change is introduced at the voltage-dependent step, by setting the ratio k&/k:, equal to los (with k&k$ still equal to 1@), the net result is to shift the curve leftward along the voltage axis, displacing the intercept (reversal potential, or equilibrium potential) to -470 mV (curve k o in Fig. 4). As can readily be seen by inspection of Eq. (6),the saturating currents, determined by xoi or xio, are unaffected. On the other hand, when the energy change is introduced somewhere along the chain of voltage-insensitive steps, represented in Fig. 4 by setting xoi/xioequal to lo*, the saturating currents are strongly affected (curve x). While the current must still intersect the voltage axis at - 470 mV (Fig. 4B), only minute currents can flow in one direction (strong hyperpolarization, as set up for this case), while enormous currents flow in the other direction (Fig. 4C). In other words, such a pump cannot be run backward by imposing a membrane potential, though it still could be by imposing a concentration gradient; voltage is kinetically incompetent to synthesize ATP, even though ATP hydrolysis can generate a potential difference through the transport system.
C. Determination of Model Parameters from I-V Curves
I
From a complete I- V curve, all parameters in the two-state model can be calculated except for the carrier density N. It acts as a scaling factor and must be assumed or known on the basis of other information. The strategy for extracting the four rate constants will necessarily vary, depending on the range of the I- V curve available and the noisiness of the data. In practical cases the most satisfactory strategy is to fit Eq. (6) to the I-V data using a nonlinear curve-fitting procedure. The labor of this operation can be reduced, and the security increased, by taking advantage of simple relationships for the saturation currents and equilibrium potential, as follows. With a large value for u or - u, Eq. (6) becomes
is,, + = zFNx,, or
265
14. KINETIC ANALYSIS OF PUMPS
The equilibrium potential (i= 0) for the transport system is related to the four apparent rate constants by e-’E
= ( k t / k $ )(
(8)
K , ~ / K ~ ~ )
Under special circumstances, other simple relationships emerge, for example, in the slope conductance at maximum steepness, in the axial displacement of the point of symmetry for the whole Z- V curve, and in the voltage for half-saturation, discussed above. When applicable, the additional equations together with Eqs. (7) and (8) yield analytic rather than merely statistical values for all parameters. D. Interpretation of Parameters: Comparison of the Two-State Model with +State Models Since the nature of Z-V data, for single-step charge transport, limits analysis to two-state models, it is important to understand the manner in which “real” parameters in higher-state models are compressed into the apparent parameters of the two-state model. Detailed derivations will be presented elsewhere (Hansen et al., 1981), but the main result can be stated succinctly in term-by-term rewriting of Eq. (6) as follows:
i= zFN
kioxoi - koixio
kin+ koi + x0i+ x io
N Bk12 Ak;, - A k2l Bk;2 =zFA B Bk,, + Ak,, + Ak;, + Bk;,
(9)
in which A and B describe the distribution of different carrier states in the whole system and are the same voltage-independent factors seen in Eq. (4), while k;, and k;, are lumped rate constants describing the entire series of reactions in the voltage-insensitive loop. The algebraic complexity of A , B, k12, and k ‘ 2 ,increases rapidly as the number of “real” steps in the transport cycle increases. For a three-state cycle, with one intermediate form of the carrier (N3)lying in the voltageindependent loop, A=l+
kl 3
k31
and
B=l+
+ k32
k23
k31
(10)
+ k32
while k;2
k13
= k32 k3 I
+ k32
and
k23
k;, = k3, k31
+ k32
Thus, the two-state model-that is, practical Z- Vcurve analysis-leaves the real reaction rate constants uncertain by the factors A and B. And the physical interpretation of A and B must come from other information.
266
DIETRICH GRADMANN
111.
f?f a/.
RESULTS
Current-voltage relationships have already been published for the principal electrogenic ion pumps in the plasma membranes of both Neurospora and Acetabularia, and some empirical analysis has been carried out, but there has been no systematic treatment via reaction-kinetic theory. In this section, therefore, some of the published data will be reanalyzed, using the formalism of the two-state model. A. Neurospora Procedures for obtaining membrane current-voltage relationships in Neurospora have been detailed previously (Gradmann et al., 1978). All curves can be resolved into two parallel (additive) components: one for the electrogenic pump itself, and one which is linear or nearly so and which can formally be ascribed to a leak. Unfortunately, for Neurospora the accessible voltage range under normal conditions lies entirely to the right (depolarizing) of the equilibrium potential for the electrogenic pump and does not even convincingly approach the saturation current in that direction. This means that the simple equations discussed in Section II,C cannot be applied directly, so that a nonlinear curve-fitting procedure must be used to estimate all four rate constants in the two-state model. However, by using Eq. (8) for boundary purposes, we can deduce that the stoichiometric coefficient (z) must be 1. The combined free energy available from ATP hydrolysis and the proton gradient across the membrane is about 450 mV (Slayman et al., 1973), while the extrapolated equilibrium potential for the pump exceeds - 300 mV, which forbids an integral stoichiometry greater than 1 between ATP hydrolyzed and H + ions pumped. Point plots of two membrane I- V curves from Neurospora are shown in Fig. 5 . As demonstrated previously, these plots can be well described empirically as segments of simple inverted parabolas, with three parameters. Obviously, then, fitting the two-state model, plus a parallel leak, to the same individual plots would leave most parameters uncertain. The situation improves considerably, however, if two different circumstances are analyzed jointly using the assumption that only a single parameter changes between conditions.4 For purposes of studying the behavior of the electrogenic pump, the most effective change in conditions would be blockade of the pump by a specific inhibitor (as, for example, ouabain for 4This assumption cannot be strictly true, since a change in one rate constant of the n-state model affects all rate constants in the two-state model through A and B. It can be shown, however, that under most conditions the major change occurs within the expected rate constant. An important exception is discussed in Section II1,C.
267
14. KINETIC ANALYSIS OF PUMPS
A
Change lc0i
c
B
Change kyo 10.4 N I
E
7
L-0.4
FIG.5 . Current-voltage relations for the plasma membrane of Neurospora crassa. Plotted points represent I- Vdata obtained from voltage pulse scans of a single hypha before (Control) and during (CN-) administration of 1 mM KCI. Fitted curves generated by Eq. (6), with an electrically parallel resistive leak. (A) All parameters are the same for both curves, except the rate constant for carrier recharging xoi. (B) All parameters are the same for both curves, except k&, as defined in Fig. 2B. Standard errors for the fits are given in the text. (Data from Gradmann eta/., 1978.)
the Na+/K+ pump in animal cells). Since no inhibitor has yet been found which is both specific and fast in its action on the proton pump of Neurospora, we made use of the fact that respiratory blockage by cyanide rapidly withdraws ATP, thereby slowing the proton pump (Slayman, 1973; Slayman et al., 1973).5 In Fig. 5, therefore, the solid curves are results obtained by jointly fitting Eq. (6), plus a linear leak, to the I-V data with cyanide (CN-) and without cyanide (Control) and allowing only one parameter to differ between the two curves. The best results are shown in Fig. 5A, in which the parameter xoiwas allowed to vary. A typical result with another parameter varied (in this case /ria) is shown for visual contrast in Fig. 5B. Fits were evaluated from standard errors [Ed2/@- 8]”, where d is the difference between the observed and model-predicted current at each voltage. The results are listed in Table I. Evidently, a single change in either of two parameters, N / AB or xoi can describe the I-V curves for the cyanideinhibited portion of the proton pump in Neurospora. Other parameters can essentially be ruled out by visual inspection, since they do not even approxOrthovanadate, which inhibits the isolated membrane ATPase from Neurospora with a K , of 0.4 p M , also blocks the proton pump in vivo, but only at high concentrations (0.1-1 mM) and after a delay of 2-4 minutes (Kuroda el a/., 1980).
268
DIETRICH GRADMANN
EFFECTOF
TABLE 1 QUALITY O F JOINTLY FITTED I-
PARAMETER CHOICE ON THE
et a/.
v CURVES"
Organism
N/AB
k:
kEi
Xoi
Xi0
gL
EL
Neurospora Acetabularia
0.012 0.008
0.022 0.021
0.085 0.024
0.010 0.025
0.022 0.019
0.064
0.090 -
-
a Equation (6) (plus a leakage term for the Neurospora data) was fitted to measured I - V curves for two different experimental conditions. The test parameter in each fitting run was found separately for the two conditions, while all other model parameters were found in common. Experimental conditions: respiring versus cyanide-inhibited, for Neurospora; lightadapted versus dark, for Acetabularia.
imate the observed shapes of both membrane I-Vcurves (e.g., see Fig. 5B). A change in N / A B was assumed in the earlier empirical analysis (Gradmann et al., 1978). That was equivalent to supposing that ATP withdrawal switches the pump off, producing only a reduction in amplitude, not a change in the shape of the pump I-V curve. From a mechanistic point of view, however, that is less satisfying than supposing that ATP withdrawal decreases the probability of a clockwise turn of the reaction cycle (Fig. 1) without altering the number of membrane carriers available. We therefore prefer to look upon the cyanide effect as diminishing the apparent rate constant xOiand have summarized this result in Fig. 6. Figure 6A shows the calculated current-voltage curves for the A
B
- 0.4
Parameter values out
in Control En 107mV
CN132mV
-415mV
1-04
FIG.6. Separated current-voltage relationships for the electrogenic proton pump and the parallel resistive leak. Same results as in Fig. 5 . Numerical values of the fitted parameters are shown in (B). N / A B was assumed to be lo-* mole/m2.
14. KINETIC ANALYSIS OF PUMPS
269
pump, in the presence and absence of cyanide, and for the fixed linear leak. (The two pump I- Vcurves sum with the leak to give the two membrane I- V curves in Fig. 5A.) Numerical results are arrayed in Fig. 6B and may be summarized by the following points. (1) The major energy shift associated with the transport cycle appears to occur simultaneously with charge transit through the membrane. (This can be deduced qualitatively by comparing the data of Fig. 5 with curves ko and x in Fig. 4. Strong asymmetry-energy change-in the electroneutral steps should yield a curve which is concave along the positive current axis.) The energy thus dissipated under short-circuit conditions (zero membrane potential) would be in excess of 40 kJ/mole of charge (Ech = - 400 mV), representing a reaction ratio (k&/kgi)of 1.6 x 10’. (2) Recharging of the transport system is an endergonic reaction, absorbing about 10 kJ/mole (En= + 107 mV) at normal ATP levels and about 13 kJ/mole (En= + 132 mV) when ATP is depleted by cyanide treatment. (3) Net release of energy by the pump is thus about 30 kJ/mole for an equilibrium potential of slightly more than - 300 mV. The latter figure is somewhat smaller than the value of -390 mV obtained earlier in the empirical analysis, which assumed merely scaling down of the pump. (4) Treatment with cyanide-and presumably the consequent ATP depletion-can be adequately described by a minor (2- to 3-fold) retardation of the recharging reaction. The fact that this retardation is smaller than the 8- to 10-fold drop in ATP with cyanide inhibition (Slayman et al., 1973) can be accounted for, at least in part, by the interaction of k,, (Fig. 1) with the other rate constants in the terms A and B [see Eqs. (4), (9), (lo)]. ( 5 ) The shapes of the pump I-V curves, as well as the numerical values for some parameters, confirm the general physical supposition that kinetic limitations must become more important than thermodynamic ones, as the system is pushed away from equilibrium. Thus, a 2.7-fold drop in xOiis reflected by a 2.6-fold drop in pump current (fortuitously, at both shortcircuit and the control resting potential), by a 100-mV depolarization due to the decreased pump current, but by only a 25-mV (8%) decrease in the equilibrium potential for the pump. (6) Finally, it should be added that the computer-fitted leak parameters, which must satisfy both control and CNdata, are intuitively reasonable. The leak conductance (slope) is, by definition, residual membrane conductance with the pump off or saturated, and it is approximated best by the measured conductance at strong depolarization in the presence of cyanide. In all the cases we have examined, the bestfit value of the leak EMF (EL)is within a few millivolts of zero, so in practice it can be constrained to zero.
270
DIETRICH GRADMANN
et
6'1.
B. Acetabularia The current-voltage relationshi? of the electrogenic C1- pump in Acetabularia reveals a striking time dependence (Gradmann, 1975). It can be represented formally by an electrical circuit having four elements: an EMF in series with two variable conductances, one having an incipient sigmoid characteristic (Pl) and the other having an N-shaped characteristic (P2); the latter conductance lies in parallel with a very large quasicapacitance (35 F/m2). This whole array is in parallel with the surface capacitance (- 50 mF/mz) and in parallel with the membrane leak conductance. The long-term steady state current-voltage relationship for the system is dominated by the N-shaped characteristic of P2, which tends to obscure any simple carrier phenomena. On the other hand, the initial I- I/ relationship-observable with voltage pulses of 10- to 100-msec duration superimposed on a steady state voltage clamp-can be assigned to P1. Initial current flow through P1 is much larger than flow through the parallel leak conductance, so the latter can in most cases be ignored, although the data for Fig. 1 have been corrected for this small leakage.
A
[I:
6
Parameter values
.......'..'..'..'
out
in
2-
NLight
A.6
1.0x~0%ol m-2
NDark
- = 0.76x10'*mol rn-2 A.6
FIG.7 . Current-voltage relationships for the electrogenic chloride pump in Acetabularia. Numerical values of the fitted parameters are given in (B). (Data recomputed from Gradmann, 1975.) All parameters are the same for both curves, except for N / A B .
14. KINETIC ANALYSIS OF PUMPS
271
The properties of P1 expressed in the initial I- V relationship (Fig. 7A) are rather different from those deduced previously (Gradmann, 1975) by steady state analysis, especially in two particulars: the saturation currents are about 10-fold larger, and the half-saturation voltage has increased from 3 to 26 mV. As in the case of Neurospora, the new kinetic analysis on Acetabularia has been carried out jointly with two different sets of conditions, allowing only one parameter at a time (in the two-state model) to vary between conditions. Even though, once more, no specific and rapidacting chemical inhibitor of the electrogenic pump is known, the condition of total darkness was found to cause partial inhibition (Fig. 7A, solid circles) and was convenient to use in the joint analysis. Numerical results are displayed in Fig. 7B. Again, the principal energy shift in the transport cycle seems to occur during the charge transit step and amounts to 40 kJ/mole cycles. Beyond this point, however, the chloride pump in Acetabularia is quite different from the proton pump in Neurospora. The most conspicuous difference is that the stoichiometric coefficient for charge transport must be at least 2. In the description of Eq. (6) it was noted that the limiting slope conductance for an electrogenic pump (with k: and k:i much greater than xoiand xio)gave a half-saturation voltage of 28/z mV. The value of 26 mV is just too small for a stoichiometry of 1 C1-/cycle; and in fact, better fits of the I- V data results with z = 2, rather than z = 1. A second aspect of the higher value of z, for the chloride pump, is that energy cannot be transferred from chemical bonds to the ionic gradient at as high potentials as could be developed by the proton pump of Neurospora. This is manifest in Fig. 7A and B as an equilibrium potential smaller than (-)200 mV for the electrogenic chloride pump. Another obvious difference between the two electrogenic pumps is the 10-fold larger saturation current of the chloride pump. This is to be expected from the difference in the ordinate scales of Figs. 6A and 7A and emerges from the calculations as a nearly 10-fold larger value of xoi for the chloride pump. At the same time the back reaction in Acetabularia, represented by xio,is reduced relative to that in Neurospora, so that the overall recharging process is nearly at equilibrium. Finally, the effect of darkness on the chloride pump of Acetabularia is definitely not to change the balance between xoi and xio but to reduce the total number of pumps available for transport (Table I): N / A B falls by 25% in darkness. A comparison of the chloride pump in Acetabularia and the proton pump in Neurospora suggests that the apparent endergonic character of the recharging reaction (xoi/xio<1) in Neurospora may arise from the asymmetry of proton distribution in this organism. The internal pH in Neurospora appears t o be in the range 7.0-7.4 (Sanders et al., 1981), while the external pH in these experiments was 5.8. The resultant gradient, representing
272
DIETRICH GRADMANN
et a/.
perhaps 90 mV, would tend to drive the transport cycle in the counterclockwise direction (Figs. 1 and 2) by biasing the protonation and deprotonation reactions at the membrane surfaces. In Acetabularia, on the other hand, the cytoplasmic chloride concentration is close to that of seawater, about 525 mM (Mummert, 1979; Saddler, 1970) and therefore can exert no net driving force on the transport reaction. In order to facilitate comparisons between the two pumps and the two organisms, the results in Figs. 6 and 7 are listed (somewhat rounded) in Table 11. C. Discussion: Localization of the Energy Shift The models in Figs. 1 and 2 are set up to make a clear kinetic distinction between voltage-sensitive and voltage-independent steps in the overall transport cycle. Phase boundary potentials (surface charges) have been ignored for simplicity, and all reactions not involving transit of charge across the membrane have been assumed to be independent of the transmembrane difference in electric potential. The reactions include binding and release of the transported ion, transit of the neutral carrier, and energization (phosphorylation) of the carrier. Application of the model t o data on electrogenic proton and chloride pumps has led t o the conclusion that the major energy shift-which we regard as synonymous with deenerTABLE I1 COMPARISON OF ELECTROGENIC PUMPCHARACTERISTICS CharacteristicU Pumped ion Direction Substrate/ATP ( z )ratio N / A B (mole/m2) xio (sec-l) xoi (sec-l) En
(mv)
&bstr.
(mv)
k: (sec-I) k:i (sec-I) Ech
Ep (mv)
Neurospora
H+ 1-0
1:l 10-8 20,000 200 f
+ 100 + 100
2 x 106 2 x 10-1 - 400 - 300
Acetabularia
c10-i 2: 1 = 10-8 2000 2000 0 0 4 x lo6 4 x 10-1 - 200 - 200
“Ep is the equilibrium potential for the whole transport system. Esubstr, is the diffusion potential for the pumped ion. Other symbols are defined in Figs. 2 and 6 . N / A B is arbitrarily set equal to lo-* mole/m2. If the actual pump density differs from this, all the apparent rate constants will change in inverse proportion to N / A B .
14. KINETIC ANALYSIS OF PUMPS
273
gization of the carrier-seems to occur during the actual step of transmembrane charge transfer. This result is in agreement with the current understanding of partial reactions in the best-known ion pump, the (Na+,K+)ATPase system (EC 3.6.1.3; Post et al., 1975; Karlish el al., 1978). However, as mentioned in Section II,B, a major limitation of this analysis of current-voltage relationships is that the zero-voltage rate constants, k$ and k& from the two-state model, contain factors A and B, which depend on the number, size, and arrangement of the rate constants in the voltage-independent pathway. Since the magnitude of energy shift in the voltage-sensitive limb is measured by the ratio of rate constants, k & / k ~ i = B k ~ 2 / A k[see q , e.g., Eqs. (9)-(ll)], it is necessary to ask whether B/A could distort the rate constants so much as to transfer the apparent locus of the energy shift from the voltage-independent pathway to the voltage-sensitive limb. A numerical survey of the three-, four-, and fivestate models has shown that in most cases the answer is no; but there is one class of cases in which large errors can arise: if the energy change for clockwise driving of the transport system immediately follows charge transit across the membrane. In the three-state model, for example, when k23 , is limited to the range 1.0-1.5, becomes large with respect to k32, k 3 1 k, I 3 A but B becomes large. In this circumstance, variations of external pH (Neurospora) or chloride concentration (Acetabularia)-by modifying the ratio k,,/k3,-could mimic changes in the membrane potential. Such an effect might go far toward accounting for the remarkable kinetic interchangeability of voltage and ionic gradients observed with physiologically reversible proton pumps (Al-Awqati, 1982; Maloney, 1982).
IV.
EXTENSIONS OF THE MODEL
A. Relation of Gradient-Driven Transport to Active Transport From a formal point of view, the models presented in this article apply to any charge transport system, regardless of the amount or source of energy. The rate constants, as defined in Figs. 1B and 2B, contain reactant concentrations, and it does not matter in principle whether the transition from N4 to N5 involves phosphate addition or, for example, binding of a hypothetical sugar for export, with the net energy coming from an outward sugar gradient. Thus, the I- V characteristics of co- and countertransport systems are described just as well as active transport systems by Eq. (6). Only in the case of additional charge transfer limbs do the relationships change.
274
DIETRICH GRADMANN f?t
a/.
B. Influence of Multiple Charges and Multiple Charge Transfer Limbs In analyzing the data from Acetabularia, we have already discussed the effects of different stoichiometric coefficients (2) on the Z- V relationships. It may be useful to summarize the major points here: When z equals 2, instead of 1, (1) the saturation current doubles, (2) the pump equilibrium potential is halved, and (3) the limiting conductance of the pump doubles. Additional charge transfer limbs add additional inflection points to the overall Z- V curve: up to three inflection points for each additional limb. Furthermore, if different limbs have different charge stoichiometries, completely different phenomena can appear, such as negative slope conductances. [It is possible that the N-shaped Z- V curve which characterizes the P2 component of chloride transport in Acetabularia can be accounted for by adding a second charge transfer limb (with z z 2 ) to the scheme in Fig. 7B, but we have not yet explored the idea to determine whether it is quantitatively satisfactory.]
C. Unstirred Layers and Asymmetric Potential Profiles Unstirred layers, which have the effect of distorting the concentrations of transported substrates near the membrane surface, should proportionately change the apparent rate constants for substrate-binding reactions. Since these reactions are assumed to be insensitive to the transmembrane voltage, the main effects will be to reduce saturation currents and to shift the Z- Vcurves toward the sigmoid shape, with a single inflection (Fig. 3). Unstirred layers could be incorporated explicitly into the model as additional rate constants in the voltage-independent loop. A quite different change in shape of the Z-Vrelationship is effected if the energy barrier that the transported ion must cross is asymmetric, as might happen if the membrane has dissimilar charges on its two surfaces (Liiuger and Neumcke, 1973). Then different fractions of the reduced membrane potential appear in the negative and positive exponentials, so that the Z- V relationship can become strongly rectifying in the vicinity of the equilibrium potential, with no change in the final saturating currents. D. Approximation of Pumps by Ideal Sources There are two customary formalisms in the transport literature for handling electrogenic ion pumps. The first, mainly used in the literature of excitable tissues, treats such pumps as ideal current sources, passing a fixed
14. KINETIC ANALYSIS OF PUMPS
275
amount of current through the membrane at all realizable membrane potentials and thus effectively having an infinite internal resistance. The pump velocity is assumed t o be independent of the driving force. Substantial experimental evidence supports this assumption.6 But most evidence comes from work with nerve and muscle cells, where the range of realizable membrane potentials is kept narrow either by dielectric breakdown or by strongly voltage-dependent changes in passive permeability. The second formalism, often used in the literature of epithelial membranes, treats electrogenic pumps as real (and linear) voltage sources, passing a current that is inversely proportional to the fraction of total membrane resistance which is in series with the pump. This formalism has less extensive experimental justification but is attractive because it allows electrogenic pumps to be handled by a simple extension of linear coupling theory (irreversible thermodynamics; Hoshiko and Lindley, 1967). Both formalisms lend themselves easily to the description of transport events via the simple theory of linear equivalent circuits. The clearest point to emerge from the above development of kinetic models, however, is that these two customary formalisms represent two extreme points on the total nonlinear current- voltage relationship for electrogenic pumps: saturation (far from equilibrium) for the current source, and near equilibrium for the voltage source. The most general formal representation of electrogenic ion pumps then is Eq. (6). In circumstances where it cannot be used, the choice between the more customary approximations should be based on proximity to either of the extreme conditions. For animal cells, the energy available from hydrolysis of ATP is normally quite large (400-600 mV) compared with realized membrane potentials (maximally ca. - 150 mV), so that unless multiple-charge stoichiometry is suspected, the current source approximation is likely to be good, whereas the voltage source approximation is likely to be very poor, at least for a pump with k o energization (Fig. 4). [Some caution is due here, however, since a pump with x-type energization could imitate a real, linear voltage source if investigated over a narrow range of membrane potentials (e.g., - 200 to - 100 mV in Fig. 4, or - 50 to + 50 mV.] For energy-conserving membranes, which normally use ion gradients to drive ATP synthesis but which can readily switch and use ATP to pump ions, the voltage source approximation is likely to be most useful. And for fungi, giant algae, and higher plants, where resting potentials range between about - 50 and - 300 mV or beyond, either approximation might be satisfactory, depending upon the particular circumstances. 6Such a result is consistent only with the ko type of energization (Fig. 4). For the x type, pump current and slope conductances must be very voltage-sensitive over the normal range of membrane potential.
276
DIETRICH GRADMANN
et a/.
ACKNOWLEDGMENTS This work was supported by research grant GM-15858 from the National Institute of General Medical Sciences, and by grant Gr409/7 (to D.G.) and grant Ha712/6 (to U.-P. H.) from the Deutsche Forschungsgemeinschaft. The authors are also indebted t o Drs. Dale Sanders and Carolyn Slayman for many helpful discussions.
REFERENCES Al-Awqati, Q. (1982). In “Electrogenic Ion Pumps” ( C . L. Slayman, ed.). Academic Press, New York. Gradmann, D. (1975). J. Membr. Biol. 25, 183-208. Gradmann, D., Hansen, U.-P., Long, W. S., Slayman, C. L., and Warncke, J. (1978). J. Membr. Biol. 39, 333-367. Hansen, U.-P., Gradmann, D., Sanders, D., and Slayman, C. L. (1981). J. Membr. Biol. 63, 165-190. Hoshiko, T., and Lindley, B. D. (1967). J . Gen. Physiol. 50, 729-758. Karlish, S. J. D., Yates, D. W., and Glynn, 1. M. (1978). Biochim. Biophys. Acta 525, 252-264. Kuroda, H., Warncke, J., Sanders, D., Hansen, U.-P., Allen, K. E., and Bowman, B. J. (1980). In “Plant Membrane Transport” (J. Dainty and R. Spanswick, eds.), pp. 507508. Elsevier, Amsterdam. Lauger, P. (1979). Biochim. Biophys. Acta 552, 143-161. Lauger, P., and Neumcke, B. (1973). In “Membranes” (G. Eisenman, ed.), Vol. 2, pp. 1-59. Dekker, New York. Lauger, P., and Stark, G. (1970). Biochim. Biophys. Acta 211, 458-466. Lieb, W . R., and Stein, W. D. (1974). Biochim. Biophys. Acta 373, 178-196. Maloney, P. C. (1982). In “Electrogenic Ion Pumps” (C. L. Slayman, ed.). Academic Press, New York. Mummert, H . (1979). Ph.D. Thesis, University of Tubingen, 1979. Post, R . L., Toda, G . , Kume, S., and Taniguchi, K. (1975). J . Supramol. Struct. 3,479-497. Saddler, H. D. W. (1970). J. Gen. Physiol. 55, 802-821. Sanders, D., Hansen, U.-P., and Slayman, C . L. (1981). Proc. Nail. Acad. Sci. U S A . 78, 5903-5907. Shamoo, A. D. (1975). Ann. N . Y . Acad. Sci. 264, 485 pp. Slayman, C. L. (1973). J . Bucteriol. 114, 752-766. Slayman, C. L., Long, W. S., and Lu, C. Y. -H. (1973). J. Membr. Biol. 14, 305-338.
CURRENT TOPICS IN MEMBRANES A N D TRANSPORT, VOLUME 16
Chapter 15
Some Physics of Ion Transport HAROLD J. MORO WITZ Department of Molecular Biophysics and Biochemistry J . W . Gibbs Research Laboratory Yale University New Haven, Connecticut
I. Free Ion and Ion Carrier Migration ............................................................ 11. Ion Conductance .................................................................................... References ............................................................................................
277 278 281
The transport of ions across membranes can occur in a limited number of ways (Onsager, 1968): (1) free ion migration, (2) ion carrier migration, (3) specific ion conductance or semiconductance, and (4) specific ion conductor coupled to an energy source. The first two are essentially liquid phase phenomena, while the last two are solid state.
1.
FREE ION AND ION CARRIER MIGRATION
First, let us consider free ion migration in a gradient of electrochemical potential. For the case of a one-dimensional gradient the flux of ions across a plane perpendicular to the gradient is given by Flux = - DdC/dX + E.!Fc/F
(1)
Where D is the diffusion coefficient, c the concentration, E the field s t r e n g t h , F the Faraday, and F the molar frictional coefficient. The first term is Fick’s law of diffusion, and the second is the migration in a uniform electrical field. For free diffusion D and F are related by
D= R T / F 277
(2) Copyright @ 1982 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-153316-6
HAROLD J. MOROWITZ
278
so that F is the reciprocal of the conventional ionic mobility. The electrical field is given by the gradient of the electric potential -d$/dx, so we may write
[
]
flux=-5 d(RTlnc+~$) F dx
=-
c dp
Fdx
(3)
In the dilute solution approximation the electrochemical potential p is given by the quantity in parentheses, thus giving rise to the second equality. The flow in this case depends strictly on the gradient of electrochemical potential. But it is completely dissipative, the electrical component going into ohmic heating and the diffusion component going into the entropy of mixing. Similar arguments hold for ion carrier migration; the thermodynamic formulation in terms of the electrochemical gradient provides a formalism for studying the kinetics. All such liquid phase systems are dissipative, as outlined above by transfer of energy from potential to thermal forms. Such mechanisms may be adequate for transport but fail where energy transduction is required. Liquid phase transport has been extensively treated, and we will focus attention here on solid state processes where four types of conductors might be considered: (1) electron, (2) proton, (3) large ion, and (4) ion pair. The first type of conductor is extensively discussed in works on solid state physics (Kittel, 1976) and is the best known. From our point of view, we should note that an electron-conducting wire and the associated electrodes might be regarded as a semipermeable membrane for electrons. All electrochemistry depends on the existence of such wires. Since surface electron activity equilibrates with oxidation-reduction potentials, we move back and forth between electromotive forces (EMFs) and these potentials. II.
ION CONDUCTANCE
Ion conductance in solids has usually been reported in crystals and proceeds by one of the following mechanisms (Farrington and Brant, 1979): (1) Proton migration followed by Bjerrum fault migration; this can also apply t o ions other than protons; (2) lattice vacancy (Schottky defect); (3) Frenkel defect-ion interstitial position; and (4) interstitiacy-cooperative motion where a lattice ion hops to an interstitial site and an interstitial ion fills the vacancy created. To illustrate the extent of these conductances consider two examples: (1) Crystalline sodium alumina has a sodium ion conductivity at 25°C comparable to that of aqueous NaCl (0.1 M ) . (2) Ice at 0°C has a proton conductance about equal t o the ionic conductance of water at the same temperature.
15. SOME PHYSICS OF ION TRANSPORT
279
All these mechanisms require a regular lattice, which in most of the cases studied has been three-dimensional. Transmembrane conductors are likely to be proteins. While the a helix is a regular lattice from the point of view of conductance, it is a one-dimensional lattice. This will also be true for higher aggregates such as 2a helices in the coiled configuration and a number of other structural models which have been proposed for proteins. Low-dimensional materials have been the subject of considerable study in recent years in an engineering context (Milles and Epstein, 1978). Most of the applications have been t o electron conductors. Proton conductance has been extensively reviewed by Glasser (1975), and most of the cases that appear to be of biological interest consist of chains of contiguous hydrogen bonds similar to the --O-H--H-O--H chains of ice. The ice case has been studied in detail and consists of ion formation or fault formation followed by migration. The net transport of an ion plus a fault restores the system t o a conducting state, and at the same time results in the transport of one proton. Biological applications of this proton conductance have been discussed in some detail (Morowitz, 1978), and experiments are in progress to evaluate this theory. Possible channels for proton conduction within proteins have been suggested (Nagle and Morowitz, 1978; Dunker et al., 1976), and the proton wire is emerging as an attractive possibility. The solid state transport of heavy ions has recently been reviewed by Farrington and Brant (1979). A number of model substances have been investigated, and some progress is being made on the theory. While no biological heavy ion channels have been identified, Onsager (1967) has suggested the possibility. Proton channels might be able to conduct other monovalent cations under appropriate conditions. In this article we shall present a model of how a redox reaction can be linked to transmembrane proton transport. Consider first a transmembrane chain of hydrogen bonds as diagramed in Fig. 1 . Focus on a single hydrogen bond as shown in Fig. 2. When the hydrogen is in the well on the left of
Rl
I
FIG. 1. Chain of hydrogen bonds.
280
HAROLD J. MOROWITZ
L
COORDINATE OF HYDROGEN BOND PROTON FIG. 2.
Hydrogen bond versus ion pair.
the chain, it is in the ground state. When it is in the well on the right, an ion pair has formed since the right-hand oxygen has an extra positive charge and the left-hand oxygen has a deficit of a positive charge. Parallel to the chain of hydrogen bonds consider two possible structures: (1) Two heme groups in a redox chain, and (2) a chromaphore. In a redox reaction, an electron moves from one heme group to the next. If the structure was originally electrically neutral, the redox reaction will generate a dipole moment of eL, where L is the distance between the heme groups. In the case of the chromaphore, the excited state may have a dipole moment parallel to the proton chain. The excited state of retinal, for example, has a large dipole moment parallel to the molecular axis. In either case, we have a component of an electrical field parallel to one or more hydrogen bonds. If we call this component E , then the following relations can be obtained. The probability of an ion pair formation is given by P = e-Ac’RT
In the presence of the dipole field p = e(-AG+Ea)/RT
15. SOME PHYSICS OF ION TRANSPORT
281
where a is the distance between the wells in Fig. 2 and is normally about 0.07 nm. Thus the initial dipole leads to charge separation and migration in the hydrogen chain. A dipole in this chain has a sign opposite that of the original dipole. Since the excited state may have a charge distribution radically different from the ground state (Mathies and Stryer, 1976), normally neutral groups such as CH, may become appreciably acidic and act as proton injectors. If a conducting channel is present, it can couple the energetic electron process to proton transport. What is central to all these effects is a more-or-less regular lattice that permits ion hopping: The present focus must therefore be on an experimental search for such lattices in molecular complexes and organelles.
REFERENCES Dunker, A. K., Marvin, D. S., and Zaleske, D. J. (1976). Biophys. J . 16, 1020. Farrington, G . C., and Brant, J. L. (1979). Science 204, 1371-1379. Glasser, L. (1975). Chem. Rev. 75, 21-64. Kittel, C. (1976). “Introduction to Solid State Physics.” Wiley, New York. Mathies, R., and Stryer, L. (1976). Proc. Nut/. Acad. Sci. U.S.A. 73, 2169-2173. Milles, J. S., and Epstein, A. J., eds. (1978). Ann. N. Y. Acud. Sci. 313. Morowitz, H. J. (1978). A m . J . Physiol. 235, R99-R114. Nagle, J. F., and Morowitz, H. J. (1978). Proc. Nut/. Acud. Sci. U . S . A . 75, 298-302. Onsager, L. (1967) “Thermodynamics and Some Molecular Aspects of Biology: The Neurosciences Quarton” (T. Melnechuk, and F. 0. Schmitt, eds.). Rockefeller Univ. Press, New York. Onsager, L., (1968). Nobel Lecture.
This Page Intentionally Left Blank
Part V
Molecular Mechanisms of Charge Separation
This Page Intentionally Left Blank
CURRENT TOPICS IN MEMBRANES AND TRANSPORT, VOLUME 16
Chapter 16
An H+ -ATP Synt hetase: A Substrate Translocation Concept I . A . KOZLOV AND V. P. SKULACHEV Isotope Department and Department of Bioenergetics A . N . Belozersky Laboratory of Motecular Biology and Bioorganic Chemistry Moscow State University Moscow, USSR
I. The Substrate Translocation Hypothesis ..................................................... Determination of the Equilibrium Constant for the Reaction ATP + H,O=ADP + Pi at the Active Site of H+-ATP Synthetase .................... 111. The Energy-Dependent Release of F,-Bound AMPPNP from the Membrane of Submitochondrial Particles ............... IV. Comparative Inhibitor Analysis of Solubilized and M Factor F, .......................................................... A. ATP-MC: An Inhibitor Modifying the &Subunit............................................ Catalytic Site of Factor F, ......... B. ADP Derivatives as Modifiers of the Noncatalytic Site on the ........................................................ a-Subunit of Factor F, ...... C. Butanol Treatment: Localization of the Factor F, Catalytic Site Close to the Outer Surface of the Mitochondria1 Membrane ...................... D. Treatment with Lithium Chloride: Shielding of the Active Site of H+-ATP Synthetase from the Matrix Side by the a Subunits of Factor F, .................................................................................. References ..................... .................................................................
285
11.
1.
288
290 292 292 294 297
299
300
THE SUBSTRATE TRANSLOCATION HYPOTHESIS’
When formulating and developing the chemiosmotic hypothesis, Mitchell postulated several alternative schemes for the mechanism utilizing A&,+ to synthesize ATP. In one of these schemes, discussed in 1973-1974, it is I The following abbreviations are used in this article: Factor F,, soluble mitochondria1 ATPase; ADP-MC, ATP-MC, GDP-MC, CDP-MC, CTP-MC, GTP-MC, mixed anhydrides
285
Copyright a 1982 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-1533 16-6
286
I. A. KOZLOV AND V. P. SKULACHEV
assumed that there is translocation of adenine nwleotides and phosphate between the mitochondrial matrix and the area of the membrane where a proteolipid system (factor F,) is localized, the latter organizing a proton transfer from the catalytic site of the ATP synthetase (part of factor F,) to the extramitochondrial medium (Mitchell, 1973; Mitchell and Moyle, 1974). Since the ionized molecules of ADP and PI have two more negative charges than ATP, electrogenic antiport of ATPn/(ADP + Pl)n-2between the mitochondrial matrix and the F, active site might occur, thus contributing to the generation of a membrane potential by the ATPase reaction. ATP/(ADP+P,) antiport, if it really exists, can be organized in one of the following ways, already described for the translocation of other substances ihrough biological membranes. One possible mechanism consists of transmembrane movement of the substrate-carrier complex; another requires formation of a channel penetrable by the transported substance; and a third is a relay rn which the substance i s translocated between two or more binding sites situated different distances from the start, i.e., from the surface of the membrane (for review, Fee Ovchinnikov el ul., E974) Our research group has put forward the hypothesis that factor F, carries out the translocation of A'TP, ADP, and P, according lo the third (reliay) principle QKozlov, 2975; Kozlov and Skulachev, 1977). The most ,mportant points of this scheme are me foilowing: 1 . The catarytic site of the membrane-linked factor F, is immersed in the hydrophobic jayer of the membrane and is not therefore directly acces
of corresponding nucleotides and mesitylenecarboxylic acid; oADP, 2'-O-[(R)-formyl(adenine-9-yl)methyIy-3'-diphosphate-~'-deoxy-(S)-glyceraldehyde jdialdehyde derivative of ADP); CMCD, Ncyciohexyl-N' -P-(rl-methylmorpholinebethylcarbodiimide; CCCP, carbonylcyanide-p-chloromethoxyphenylhydrazone; IJCCD, dicyclohexylcarbodiimide; AMPPNP, 5 '-adenylylimidoiYiuhosphate; SDS, sodium dodecyl sulfate.
16.
H+-ATP SYNTHETASE:
A SUBSTRATE TRANSLOCATION CONCEPT
287
electric fields due to protonation and deprotonation of some protolytic groups of the ATP synthetase, if there is a ApH across the membrane. A schematic presentation of the above hypothesis is given in Fig. 1. It is proposed that the first stage of ATP synthesis consists of deprotonation of the substrates, ADP and/or PI, and complete neutralization of the ADP3 and P;- anions by six positive ligands (Mg2+ may serve as two of these ligands). These processes are assumed to take place at the F, noncatalytic site exposed to the mitochondria1 matrix. Then the electroneutral complex of ADP, P,, and ligands is translocated to a catalytic site where ATP and H,O are formed from ADP, PI, and two H . The latter are transported to the catalytic site via an oligomycin- and DCCD-sensitive proton-conducting pathway. ATP4 (formed from ADP3 and PO: ) makes a complex with six positive ligands (L), so that the total charge is 2 + . This complex moves electrophoretically back to the matrix surface of the membrane, down the transmembrane electric field if there is a A$ across the membrane. If ApH drives ATP synthesis (low pH outside and high pH inside the mitochondrion), the proton-accepting group@) of the catalytic site, X, and the proton-donating group(s) of the noncatalytic site, YH, are postulated to be in their charged forms (XH+ and Y-, respectively) when ATP is bound to the catalytic site. There will then be a local electric field (see Kozlov and Skulachev, 1977) across the F, molecule, favorable for translocation of the (ATP*6LI2+complex from the catalytic to the noncatalytic site. +
2U+
FIG. I . A scheme for the substrate translocation hypothesis of the H+-ATP synthetase mechanism. ADP03-, OPO;., and ADPOPOj-, Ionized forms of ADP, Pi, and ATP, respectively; L, a cationic ligand: X and Y, proton-transferringgroups; F, and F,, coupling factors.
288
I. A. KOZLOV AND V. P. SKULACHEV
II. DETERMINATION OF THE EQUILIBRIUM CONSTANT FOR THE REACTION ATP+H,O +ADP+ Pi AT THE ACTIVE SITE OF H+-ATP SYNTHETASE One essential tenet of the above-mentioned hypotheses is that no input of the ApH+ energy is required to form ATP at the active site of the H+-ATP synthetase. Considering this point, we would first like to discuss the experiments of Boyer and co-authors (Choate et al., 1979), who studied the ATP dependent H,I80-P, isotope exchange, which can be catalyzed by factor F,. From the results obtained, it follows that, at low concentrations of ATP (< 10-5M), the ATPase reaction velocity is limited by release of ADP and P, from the active site of F,. A dynamic equilibrium for the reaction ATP + H,O * ADP + Pi is established at the active site, and the rate of ADP phosphorylation seems to be higher than the rate of ADP and Pi release into water. Consequently, one can observe several 1 8 0 atoms from H,I80 incorporated into the Pi formed as a result of ATP hydrolysis. As the ATP concentration is increased to 10-20 pM, the H,I80-Pi exchange is inhibited. Figure 2 shows the dependence of the rate of ATP hydrolysis, catalyzed by isolated F,, on the concentration of substrate-as found in our group by Chernyak et al. (1981) (see also Shuster et al., 1975, and Takeshige et al., 1976). On the basis of Fig. 2 and the data of Boyer's group (Choate et al., 1979), the following scheme can be suggested to describe the mechanism of the ATPase reaction: ATP
+
F,
k+l
~
[ATP*F,j
KH
k2
[ A D P * F , * P i ] -Fl
+
ADP
+
Pi
+
ADP
+
Pi
k-1 ATP - , / k * T P
F, [""P:;i
Pi]
+
ATP
k3
In the first stage of the reaction, ATP is bound to the active site of F, in a step characterized by the bimolecular rate constant k , ,. Reversible hydrolysis of ATP occurs next, followed by dissociation of ADP and Pi from the active site into water. However, at low concentrations of ATP, the rate of ADP phosphorylation by inorganic phosphate exceeds the rate of ADP and Pi release from the active site. Apparent kinetic constants for the overall reaction are Km(app) = 0.03 mM, and k,,, = 5 x lo3 min-' (Fig. 2). At high substrate concentrations, factor F, binds a second ATP molecule, which leads both to an increase in the rate of ADP and Pi release
16.
H+-ATP SYNTHETASE: A SUBSTRATE TRANSLOCATION CONCEPT
50
400
289
>
% Mg-ATPI ,mM-' FIG. 2. Dependence of the rate of the F, ATPase reaction on substrate concentration (Lineweaver-Burke plot). Inset shows an enlarged fragment of the same curve. The ATPase activity was measured in an ATP-regenerating system. (From Chernyak et al., 1981.)
from the active site ( k , > k , ) and to a decrease in the affinity of the active site for the substrate. (Km(app) increases from 0.03 M a t a low ATP level up to 0.3 mM at a high one; see Fig. 2). Thus, the two major changes in the kinetics of ATP hydrolysis at high substrate concentrations (increase in the rate of hydrolysis and increase in can be explained by the lowered affinity of the F, active site for ATP and ADP, consequent upon binding of a second ATP molecule. The equilibrium constant K H ,at the active site of F,, can be determined if the relevant binding constants for ATP, ADP, and Pi are known. On the basis of data from competitive inhibition of the ATPase reaction by ADP, we estimated KA,,=5 x 10-4M(Kozlov and Kononenko, 1975). According to Ting and Wang (1980), K, G lO-,M. T o calculate KATP,we first had to determine k,, and kI.The dissociation rate constant of the complex (F;ATF) was estimated from the measured release of a nonhydrolyzable ATP analog, AMPPNP, from the active site. k-,was found to be 2x lo-, min-I (Chernyak et af., 1981). From the reaction scheme and discussion above, it follows that k, = kcat/Km(app), at low ATP concentrations, and the stated values of k,,, and give k, I = 1.7 x los min-I. Therefore, KATp= k _ , / k +I GZ 1O-Io M . Comparing the data on the binding constants for Pi and ADP at the
,
290
I . A. KOZLOV AND V. P. SKULACHEV
active site of factor F,, with the values of the ATP-binding constant, we could conclude that the difference in the energy of binding of ATP and of the ATPase reaction products ensures the reversibility of ATP hydrolysis at the active site of the ATPase (KH is close to 10). A similar conclusion may be drawn on the basis of the following independent consideration. As already noted above, at low ATP concentrations the rate of ADP phosphorylation at the active site of F, is greater than the release rates for ADP and Pi (Choate et al., 1979). Assuming that release of ADP and Pi from the active site is the rate-limiting step of the ATPase reaction (Choate et al., 1979), and taking into account the data of the ATPase reaction rate (Fig. 2), we may conclude that the rate of ATP synthesis at the active site is more than lo3 turnovers/minute. On the other hand, the maximal rate of ATP hydrolysis, like that of any other hydrolase reaction occurring according to the mechanism of acid-base catalysis should not exceed lo5 turnovers/minute. [The given value (lo5 turnovers/minute) is the limit for substrate protonation at the active site of the enzyme by a proton-donating group during acid-base catalysis. In fact, none of the known enzymes possessing hydrolytic activity acts at a higher rate.] The ratio of the maximal possible rate of ATP hydrolysis at the active site to the minimal rate of ATP synthesis results in a value of KHof no more than 100. Thus, the results obtained are compatible with one of the theses of the aforementioned hypothesis (Fig. 1)-that ATP can be synthesized at the active site of factor F, in the absence of external sources of energy. Hence the energy of ApH+ should be utilized at the stage of ATP release from the active site of H+-ATP synthetase into the solution. This conclusion was confirmed by experiments on the A&, +-dependent release of the nonhydrolyzable ATP analog AMPPNP.
1111. THE ENERGY-DEPENDENT RELEASE OF F,-BOUND AMPPNP FROM THE MEMBRANE OF SUBMITOCHON DRlAL PARTICLES Low concentrations of AMPPNP specifically slow down the ATPase reaction in submitochondrial particles without affecting oxidative phosphorylation (Holland et al., 1974; Pedersen et al., 1974; Penefsky, 1974). The ATPase inhibition becomes much stronger when an uncoupler is added (Melnik et al., 1975). In terms of the scheme shown in Fig. 1, these facts can be accounted for in such a way that AMPPNP, like ATP, is removed from the factor F, catalytic site in a ApH+-dependentfashion. AMPPNP cannot be bound to
16.
H+-ATP SYNTHETASE:
A SUBSTRATE TRANSLOCATION CONCEPT
29 1
the catalytic site under conditions of oxidative phosphorylation since A & + is high in such a state. But AMPPNP can combine with factor F, in the uncoupled state when there is no A j i . H + . The hypothesis thus predicts that a short incubation of the AMPPNP-pretreated particles under energized conditions should remove AMPPNP from the F, catalytic site, hence reversing the ATPase inhibition. Experiments performed by our group (Chernyak and Kozlov, 1979) have confirmed this prediction, as shown by Fig. 3. The ATPase activity measured in uncoupler-treated submitochondrial particles preincubated with AMPPNP is very low for several minutes and then increases spontaneously (Fig. 3, lower curve). Without the AMPPNP pretreatment, the rate of the ATPase reaction is high and constant from the beginning of the assay. In Fig. 4 (curve 1) AMPPNP-pretreated particles were added to a mixture containing succinate. After 20 seconds the uncoupler CCCP was added. One can see that in this case ATP hydrolysis is fast and linear; but if CCCP is added to the reaction mixture before the submitochondrial particles (curve 2), the ATPase reaction again is slow initially but then accelerates. Such results demonstrate that energization of the membrane greatly facilitates removal of an ATP analog, AMPPNP, from the catalytic site of mitochondria1 ATPase.
/
- AMPPNP
FIG. 3. Spontaneous reactivation of submitochondrial particles preincubated with AMPPNP (Chernyak and Kozlov, 1979). Beef heart submitochondrial particles (80 mg proteinlml) were preincubated for 5 minutes at 20°C in a solution of 0.25 M sucrose, 10 mM HEPES (pH 7 . 9 , and 0.1 mMMgS04 with or without 0.1 mMAMPPNP. Then 0.2 ml of this mixture was added to an 8-ml solution of 0.25 M sucrose, 2 mMTris-HC1 @H 8.3), and 2 pV CCCP. The ATPase reaction was initiated by the addition of 2 mM MgeATP. The reaction was measured at 15°C.
292
I. A. KOZLOV A N D V. P. SKULACHEV
FIG.4. The effect of particle energization on the rate of the ATPase reactivation for particles preincubated with AMPPNP (Chernyak and Kozlov, 1979). Conditions as in Fig. 3, but 10 mM succinate was added to the reaction mixture.
IV.
COMPARATIVE INHIBITOR ANALYSIS OF SOLU BlLlZED AND M EM BRAN E-BOU N D FACTOR F,
The scheme proposed in Fig. 1 for the ATPase reaction is also in good agreement with the results of further inhibitor experiments on mitochondrial ATP synthetase, performed by our group.
A. ATP-MC: An Inhibitor Modifying the P-SubunitLinked Catalytic Site of Factor F, In the first series of experiments (Kozlov et a[., 1977, 1979; Drutsa et al., 1979), mixed anhydrides of nucleoside triphosphates and mesitylene carboxylate investigated for their effects on factor F,, either in solution or in everted submitochondrial particles. (These compounds were synthesized by Z. A. Shabarova, N. I. Sokolova, M. V. Shalamberidze, and V. L. Drutsa.) It was found (Fig. 5 ) that ATP-MC, E-ATP-MC,and CTP-MC all were powerful inhibitors of solubilized factor F, ATPase activity. Preincubation for 40 minutes with these compounds almost completely blocked the catalytic activity of solubilized factor F, (Fig. 5 , curves 1 and
16.
H+-ATP SYNTHETASE:
A SUBSTRATE TRANSLOCATION CONCEPT
20
40
60
293
min
FIG. 5 . Inhibition of the ATPase activity of soluble factor F, and particle-bound F, (V)by mixed anhydrides of nucleoside triphosphate and mesitylene carboxylic acid. Curves 1-3 show ATPase activity plotted against time of preincubation with the inhibitors. Curve 4 shows binding of the fluorescent label to factor F, plotted against time of treatment with €-ATP-MC. Curve 1: 1 mM ATP-MC (0),1 mM E-ATP-Mc (X). Curve 2: 2 mM GTP-MC (A), 2 mM CTP-MC (0) 2 mMATP-MC (H).Curve 3 (0):1 mMATP-MC+ 10 mMATP. In all these experiments, solubilized factor F, was studied. Curve 3(V):Under the same conditions submitochondria1 particles were treated with 3 mM ATP-MC. (From Kozlov ef al., 1979.)
2). But under the same conditions, factor F, proved to be ATP-MCresistant when integrated with the membrane of submitochondrial particles or when a high concentration of ATP was added to the preincubation mixture for solubilized F, (curve 3). In experiments with E-ATP-MC it was shown that inhibition was accompanied by incorporation of the analog into factor F,, the maximal stoichiometry being about one 6-ATP-MC per one F, unit (Fig. 5, curve 4).Again, ATP protected factor F, from modification by E-ATP-MC (not shown). In a subsequent experiment, solubilized factor F, was treated with [3H]ATP-MC and subjected to SDS gel electrophoresis (Fig. 6). One can see that radioactivity is localized in the peak corresponding to the ,6 subunits. h
A
Slice number
FIG. 6 . SDS electrophoresis of [3H]ATP-MC-modified factor F,. (From Drutsa ef al., 1979.)
294
I. A. KOZLOV AND V.
P. SKULACHEV
These data are in agreement with the observations of Budker et al. (1977) and Wagenvoord et al. (1977, 1979) who showed that both an alkylating ATP derivative and the analog 8-azido-ATP combine with the /3 subunit when inhibiting factor F, ATPase in solution. It should be stressed here that in nonaqueous solutions the mixed anhydrides of nucleotides and mesitylene carboxylic acid serve as phosphorylating (nucleotide-adding) agents in reactions with various compounds containing amino, thio, or other nucleophilic groups. As a result, the nucleotide moiety becomes attached to these groups. In aqueous solution at neutral pH, however, mixed anhydrides of nucleotides and mesitylene carboxylate are stable, so that reactions with nucleophilic groups require much more time than is required to observe inhibition of factor F, (about 30 minutes; see Fig. 5). Apparently, factor F, binds ATP-MC at its active site and has a catalytic effect, accelerating the reaction of the modifier with a nucleophilic group in this state. The fact that a substance as hydrophilic as ATP-MC can modify watersoiuhjlized F, but not membrane-bound F , (Fig. 5 ) can be accommodated within the framework of the above scheme for the H+-ATP-synthetaseby assuming the catalytic site to be immersed in the hydrophobic region of the membrane.
B. ADP Derivatives as Modifiers of the Noncatalytic Site on the a-Subunit of Factor F,
‘The results of experiments with ADP-MC (Kozlov et a/., 1977, 1979) are quite different from those obtained with ATP-MC. First, only submitochondria1 particles, not solubilized factor F,, are sensitive to ADPMC which reduces the ATPase activity to about 30% of the untreated control (Fig. 7, curve 1). in addition, ADP-MC activity shows much greater specificity for the heterocyclic base. GTP or CTP can be substituted for ATP in ATP-MC, but the corresponding diphosphate derivatives do not inhibit the particle-bound F, ATPase. UnfortunateIy, ADP-MC treatment does not result in the formation of a stable bond between ADP and the enzyme, so we were compelled to look for another AL>P derivative to identify the subunit responsible for the observed inhibition of the ATPase in particles. In further experiments (Kozlov and Milgrom, 1980), we tried ADP oxidized by periodate [2’-O-(R)-formyl(adenine-9-yl)methyl-3’-diphosphate3 ‘-deoxy-(9-glyceraldehyde] . This compound, which we have designated oADP, combines with factor F, in such a manner that the oADP-F, bond can be stabilized by subsequent borohydride treatment. Data from a typical
16.
H+-ATP SYNTHETASE:
A SUBSTRATE TRANSLOCATION CONCEPT
t
295
GDP-MC
100
COD-MC
.-.
80-
-
60-
4
I-
d
n
c .-
20-
I
,
,
40
20
, M
, 40
* 50
/ 60
,
b
min
70
Fro. 7. Inhibition of the ATPase activity of submitochondrial particles by mixed anhydrides of nucleoside diphosphates and mesitylenecarboxylic acid. ATPase activity is plot ted against time of preincubation of the particles with nucleotide derivative at 3 mM. Curve 1: ADP-MC ( 0 ) .Curve 2: GDP-MC (A)and CDP-MC (0).(From Kozlov et al., 1979.)
experiment are shown in Fig. 8. It is clear that oADP fails to inhibit the activity of soluble factor F, but-under the same conditions-does inhibit ATP splitting by submitochondrial particles. The inhibition is not complete, apparently because oADP binding in the absence of borohydride is reversible. Borohydride stabilizes this bond but simultaneously reduces oADP in solution, thus preventing further modification of the enzyme by a given portion of oADP. Accordingly, the extent of inhibition was found to be increased when particles treated with oADP and borohydride were washed (to remove borohydride) and retreated. None of these maneuvers affected the activity of solubilized factor F,. Subsequent experiments with tritiated oADP have demonstrated that the
c n + .. I
c
P
z
a
t I 0
1
10
2a
30
40
min
FIG.8. The effect of oADP on the ATPase activity of submitochondrial particles and factor F,. Particles or factor F, were preincubated with 0.2 m M oADP for the time indicated by the abscissa. Sodium borohydride (4 mM) was then added to stabilize the bond(s) between oADP and the enzyme. Small aliquots of the resulting mixtures were used to measure ATPase activity. (From Kozlov and Milgrom, 1980.)
296
I. A. KOZLOV AND V. P. SKULACHEV
11.0
d
B
3
ss
n
T-
8
E,
h
P
10
20
30
U
I\
Bil
04
40
50
02
J 60
70
Slice number
Slice number
FIG.9. SDS electrophoresis of [3H]oADP-modifiedfactor F, (Kozlov and Milgrom, 1980). (A) Modification of solubilized factor F , . (B) Modification of factor F, in submitochondrial particles.
inhibitor binds to the a subunit; and it does so both for membrane-bound F, and for the solubilized enzyme, even though it is actually inhibitory only in the former case. A 60-minute treatment of solubilized F, with 0.2 mM [3H]oADP gives incorporation of about one molecule of [3H]oADP per molecule of factor F,. Reversible binding, with a dissociation constant of 8x M (Fig. lo), precedes actual incorporation. [3H]oADP incorporation is markedly depressed when the mixture is supplemented with ADP. In particles, the binding stoichiometry, [3H]oADP/F,,proved lower than in solubilized F, but increased after repeated treatment.
-e 0
.-
6-
0
-OM
o
0.01
0.02
0.03
a04
0.05
>
0.06
C'H-OADPI-~, pM-'
FIG. 10. Factor F, modification as a function of oADP concentration. Solubilized factor F, was treated with oADP for 15 minutes. (From Kozlov and Milgrom, 1980.)
16.
Hf -ATP SYNTHETASE:
A SUBSTRATE TRANSLOCATION CONCEPT
297
Under the conditions used, no incorporation into subunits other than a subunits was obtained, either in the particles or in the solubilized factor F,. These results show clearly that modification of the nucleotide-binding site on an a subunit of factor F, inhibits the ATPase activity in the membrane but not in solution. This fact can readily be accommodated by the above hypothesis for the H +-ATP synthetase mechanism by assuming that the catalytic site of F,-immersed in the membrane-is localized on the /3 subunit, while the noncatalytic site-facing the intramitochondrial water phase-is situated on the a subunit. The latter should not be essential for ATP hydrolysis by solubilized factor F,, since the catalytic site is exposed to water. In particles, translocation of the substrates via the noncatalytic site would occur along the pathway to the catalytic site. Therefore inhibitors modifying the noncatalytic site should depress the ATPase activity of the particles. It is not surprising that derivatives of ATP can be inhibitors of the catalytic site without showing high specificity for the heterocyclic base, whereas ADP derivatives primarily attack the noncatalytic site but d o show great specificity for the structure of the heterocycle. This follows from the fact that the intact H + - A T P synthetase complex is adapted to phosphorylate ADP rather than to dephosphorylate ATP. With both the high affinity for ADP and the high specificity with respect to the heterocycle base located at the noncatalytic (a-subunit) site, the catalytic /3 subunit can be specialized in catalysis per se rather than in substrate selection. The necessary discrimination between ADP and ATP, as well as between ADP and other nucleoside diphosphates, is carried out by the noncatalytic a subunit.
C. Butanol Treatment: Localization of the Factor F, Catalytic Site Close to the Outer Surface of the Mitochondria1 Membrane If the catalytic site of the H+-ATP synthetase is situated closer to the outer surface of the mitochondrial membrane than to the inner one (Fig. l), it is possible to strip the parts of the membrane structure that separate the catalytic site from the extramitochondrial space without completely disrupting the hydrophobic barrier. Looking for such a system, we turned our attention to the decreased oligomycin sensitivity of mitochondrial ATPase, which follows treatment with small amounts of butanol (Lenaz et a/., 1975). Two inhibitors of factor F, ATPase were tried: ATP-MC and a water-soluble CMCD which, like ATP-MC, specifically modifies the
298
I. A. KOZLOV AND V. P. SKULACHEV
catalytic site of factor F, in solution but not in the membrane (Kozlov and Chernyak, 1976; Imedidze et af., 1978). As shown in Table I , butanol treatment of mitochondria decreases the oligomycin sensitivity of the ATPase and sensitizes the enzyme to CMCD. In submitochondrial particles, the same treatment only induces the former effect, so that the resulting particles hydrolyze ATP mainly in an oligomycin- and CMCD-insensitive manner. Quite similar data were obtained when ATP-MC was used instead of CMCD (Fig. 11). In terms of the above scheme (Fig. l), these results can easily be explained if we assume that butanol induces a rupture in the structure of the H -ATP synthetase complex so that hydrophobic proteins fail to protect the F, catalytic site from direct contact with the solutes of the extramitochondria1 water phase. As a result, (1) CMCD and ATP-MC become accessible to the catalytic site, hence inhibitory, and (2) hydrogen ions of the water phase equilibrate with this site without the involvement of oligomycin-sensitive hydrophobic proteins. The same butanol treatment does not sensitize ATPase of submitochondrial particles to CMCD and ATPMC. So one can conclude that a hydrophobic barrier still exists, preventing hydrophilic modifiers from penetrating into the water space inside the particles. In agreement with the last conclusion, it was found that CMCD becomes inhibitory after incubation of the butanol-treated particles in a hypotonic medium, which apparently “opens” the membranous vesicles. +
TABLE I CHANGES IN THE SENSITIVITY OF MITOCHONDRIAL ATPASETO INHIBITORSAS A RESULTOF TREATING MITOCHONDRIA AND SUBMITOCHONDRIAL PARTICLES WITH B U T A N O L ~ , ~
System studied Mitochondria not treated with butanol Mitochondria treated with 0.35 M butanol Mitochondria treated with 0.5 M butanol Submitochondrial particles not treated with butanol Submitochondrial particles treated with 0.5 M butanol
No additions
Oligomycin, 1.5 &ml
CMCD, 1 mM
CMCD, 1 mM, and oligomycin, 1 &ml
0.35 (14)
0.07 (14)
0.35 (12)
0.07 (5)
0.35 (7)
0.15 (7)
0.23 (4)
0.08 (4)
0.25 (12)
0.20 (12)
0.10 (8)
0.07 (8)
1.50 (18)
0.30 (8)
1.50 (4)
0.30 (4)
1.00 (6)
0.65 (6)
0.88 (4)
0.60 (2)
Kozlov and Chernyak, 1976. The ATPase activity is expressed in micromoles per minute per milligram of protein. The number of experiments is shown in parentheses. l n the case of CMCD inhibition, mitochondria and submitochondrial particles were pretreated with this inhibitor for 50 minutes (pH 6.2, 2OOC).
16.
H+-ATP SYNTHETASE: A
SUBSTRATE TRANSLOCATION CONCEPT
299
preincubation(min)
FIG. 11. Butanol treatment as a factor sensitizing mitochondria1 ATPase to ATP-MC inhibition. The ATP-MC concentration was 1 mM. Curve 1: Solubilized factor F,. Curve 2: Mitochondria treated with 0.5 mM butanol. Point 3: mitochondria not treated with butanol. (From Kozlov and Chernyak, 1976.)
D. Treatment with Lithium Chloride: Shielding of the Active Site of H +-ATPSynthetase from the Matrix Side by the (Y Subunits of Factor F, As mentioned in the previous section, treatment of mitochondria with butanol makes the active site of H+-ATP synthetase accessible to hydrophilic inhibitors. Further experiments have revealed that the catalytic site of H -ATP synthetase can be made accessible to hydrophilic inhibitors in another way, too, namely, by treatment of submitochondrial particles with LiCl (Kozlov et al., 1980). This treatment leads to the removal of a subunits from membranebound factor F, (Kozlov et al., 1980). The readdition of purified a subunits to the LiC1-treated particles results in reconstitution of the ATPase activity (Fig. 12, curve 1). ATP-MC which-as mentioned above-inhibits the ATPase activity of factor F, in solution but not in submitochondrial particles, suppresses ATP hydrolysis in LiC1-treated particles (Fig. 12, curve 2). This result indicates that the catalytic site [located on the 0 subunit(s) of factor F,] is shielded in the particles by the a subunits. Removal of (Y subunits as a result of LiCl treatment makes the active site accessible to the ATP-MC. The absence of complete inhibition of LiCl particles by ATPMC (Fig. 12) appears to result from incomplete removal of the a subunits. These interpretations of the LiCl data have been confirmed by reconstruction of the enzyme from separated a subunits and LiC1-treated particles. As can be seen from Fig. 12, the addition of a subunits to LiCl particles which have been preincubated with ATP-MC has little effect on the ATPase activity (Fig. 12, curve 2). But the addition of a subunits to ATP+
300
I. A. KOZLOV AND V. P. SKULACHEV
FIG. 12. Inhibition of the reconstruction of LiCl particles with a subunits of factor F, as a result of treatment of LiCl particles with ATP-MC. LiCl particles (20 mg protein/ml) were preincubated in 10 mM MOPS buffer, pH 7.5, containing 0.25 M sucrose with (curve 2) or without (curve 1) 0.5 mM ATP-MC. The moment when the a subunits (50 pg/mg of particles) were added is shown by the arrows.At the intervals indicated, small aliquots were taken from the preincubation medium to measure ATP activity. (From Kozlov el at., 1980.)
OT
i -
2
4
Time of preincubation , hours
MC-nontreated LiCl particles causes a fivefold increase in ATPase activity (Fig. 12, curve 1). The results thus indicate that the catalytic site of the ATP synthetase, located on the /3 subunit, is shielded from the mitochondria1 matrix by the (Y subunit. This conclusion is in good agreement with the scheme proposed in Fig. 1 for ATP synthetase action, which postulates that the noncatalytic site (on the (Y subunit) should mediate transfer of the substrates between the matrix and the catalytic site located in the depths of the membrane. In conclusion, the previously described results of a study of mitochondrial H + - A T P synthetase seem to confirm the main postulates of the substrate translocation concept introduced 6 years ago as a working hypothesis (Kozlov, 1975; Kozlov and Skulachev, 1977; Skulachev, 1980).
REFERENCES Budker, V. G., Kozlov, I. A,, Kurbatov, V. A., and Milgrom, Ya. M. (1977). FEBS Lett. 83, 11-14.
Chernyak, B. V . , and Kozlov, I. A., (1979). FEBS Lett. 104, 215-1219. Chernyak, B. V . , Chernyak, V . Ya., Gladysheva, T. B., Kozhanova, Z. E., and Kozlov, I. A. (1981). Biochim. Biophys. Acta 635, 552-570. Choate, C. L., Hutton, R. L., and Boyer, P. D. (1979). J. Biol. Chem. 254, 286-290. Drutsa, V. L., Kozlov, I. A., Milgrom, Ya. M., Shabarova, Z . A., and Sokolova, N. I. (1979). Biochem. J . 182, 617-619. Holland, P. C., LaBell, W. C., and Lardy, H. A. (1974). Biochemistry 13, 4549-4553. Imedidze, E. A., Kozlov, I. A., Metelskaya, V. A., and Milgrom, Ya. M. (1978). Biokhimiya 43, 1404-1412. I. A. (1975). Bioorg. Khim. 1, 1545-1569. I. A., and Chernyak, B. V . (1976). Dokl. Akad. Nauk SSSR 231, 222-225. I. A., and Kononenko, V. A. (1975). Bioorg. Khim. 1, 489-493. I. A., and Milgrom, Ya. M. (1980). Eur. J . Biochem. 106, 457-462.
Kozlov, Kozlov, Kozlov, Kozlov,
16.
H+-ATP SYNTHETASE: A SUBSTRATE TRANSLOCATION CONCEPT
301
Kozlov, 1. A., and Skulachev, V. P. (1977). Biochim. Biophys. Acta 463, 29-89. Kozlov, 1. A., Shalamberidze, M. V., Novikova, I. Yu., Sokolova, N. I . , and Shabarova, Z.A. (1977). Biokhimiya 12, 1704-1709. Kozlov, 1. A., Shalamberidze, M. V., Novikova, I. Yu., Sokolova, N. I., and Shabarova, Z . A. (1979). Biochem. J. 178, 339-343. Kozlov, I . A., Milgrom, Ya. M., and Tsybovski, 1. S. (1980). Biochem. J. 192, 483-488. Lenaz, G., Parenti-Castelli, G., and Sechi, A. M. (1975). Arch. Biochem. Biophys. 167, 72-79. Melnik, R. L., Tavares De Sousa J., Maguire, J., and Packer, L. (1975). Arch. Biochem. BiOphYS. 166, 139-144. Mitchell, P. (1973). FEBS Lett. 33, 267-274. Mitchell, P., and Moyle, J. (1974). Biochem. SOC. Spec. Publ. 4, 9IT111. Ovchinnikov, Yu. A., Ivanov, V. T., and Shkrob, A. M. (1974). “Membrane-Active Complexones.” Elsevier, Amsterdam. Pedersen, P. L., Le Vine, H., 111, and Cintron, N. (1974). I n “Membrane Proteins in Transport and Phosphorylation” (G. F. Azzone, M. E. Klingenberg, E. Quagliariello, and N . Siliprandi, eds.), pp. 43-54. Elsevier, Amsterdam. Penefsky, H. S. (1974). J . Biol. Chem. 249, 3579-3585. Shuster, S. M., Ebel, R. E., and Lardy, H. A. (1975). J. Biol. Chem. 250, 7848-7853. Skulachev, V. P. (1980). In “Soviet Scientific Reviews: Biology” (V. P. Skulachev, ed.), Vol. 1, pp. 239-312. Harwood Academic Publ., Chur. Takeshige, K., Hess, B., Bohm, M., and Limmer-Telschow, H. (1976). Hoppe-Seylers 2. Physiol. Chem. 357, 1605-1622. Ting, L. P., and Wang, J. H. (1980). Biochemistry 19, 5665-5670. Wagenvoord, R. J., Van der Kraan, l., and Kemp, A. (1977). Biochim. Biophys. Acta 460, 17-24. Wagenvoord, R. J., Van der Kraan, I., and Kemp, A. (1979). Biochim. Biophys. Acta 548, 85-95.
This Page Intentionally Left Blank
CURRENT TOPICS IN MEMBRANES AND TRANSPORT, VOLUME 16
Chapter 17
Proton Translocation by Cytoc hrome Oxidase MARTEN WIKSTROM Department of Medical Chemistry University of Helsinki Helsinki, Finland
Introduction ........................................................................................ A. Respiration-Linked H + Translocation-A Brief Historical Note ................ B. Properties of an Eiectron-Translocating Cytochrome Oxidase . 11. The Discovery of True Proton Pumping by Cytochrome Oxidase ..................... 111. Controversy over Proton Translocation by Cytochrome Oxidase ...................... IV. Molecular Principles and Mechanisms of Proton Translocation A. The Relation between “Membrane Bchr” Effects and a Pr €3. General Principles of a Kedox-Linked Proton Translocator ...................... C. Possible Molecular Mechanism of Proton ’Translocation by Cytochrome Oxidase ................................................................... References ...................... ..... ....... 1.
1.
303 304
305 307 3 10 312 312 313
316 3i8
INTRODUCTION
Cytochrome c oxidase (ferrocytochrome c:oxygen oxidoreductase, EC 1.9.3.1) is the oxygen-reducing terminal redox complex of the respiratory chains of mitochondria and certain bacteria (for reviews, see Lemberg, 1969; Capaldi and Briggs, 1976; Malmstrom, 1974, 1979; Caughey et a/., 2976; Nicholls and Chance, 1974; Erecinska and Wilson, 1978; Wikstrom et ul., 1976). In mammalian mitochondria the enzyme appears to consist of six to eight different subunits (Downer et af., 1976; Buse et af., 1978; Penttila et al., 1979; Carroll and Racker, 1977). It contains four different redox centers per functional unit, two hemes, a and a,, and two coppers, often termed Cu, and Cu,. The entire protein complex is “plugged through” the 303
Copyrrght 0 1982 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-153316-6
MARTEN
304
WIKSTROM
inner mitochondria1 membrane asymmetrically, presenting a different set of subunits on each side and protruding about 60 A out of the membrane on the cytoplasmic (C) side (Eytan et al., 1975; Blasie et al., 1978; Henderson et al., 1977; Ruben et al., 1976; Dockter et al., 1977, 1978; Frey et al., 1978). Cytochrome c , which binds to the oxidase on the C-side of the membrane (DePierre and Ernster, 1977) donates electrons to the oxidase, usually at a redox potential (E,,) of about 300 mV in aerobic steady states. Since the potential of the 0,-H,O couple is about 800 mV in air-equilibrated aqueous media at room temperature, electron transfer occurs across of about 500 mV. Hence the oxidase reaction is itself a potential span (A,!?,,) highly exergonic. However, it has long been known (Maley and Lardy, 1954) that electron transfer between cytochrome c and dioxygen is coupled to oxidative phosphorylation. Consequently, a large fraction of the released energy in the cytochrome oxidase segment must be conserved, this segment often being referred to as the third coupling site or, simply, site 3 of oxidative phosphorylation. The molecular principle by which cytochrome oxidase conserves the available redox energy is the subject of this article. According to Mitchell’s chemiosmotic theory (see e.g., Mitchell, 1976), the general postulates of which are largely accepted today, energy is conserved in respiration by coupling electron transfer to net translocation of H + across the mitochondrial (or bacterial) membrane. Because of this electrogenic proton translocation the energy will be initially stored as A & + , i.e., an electrochemical proton gradient, which may be secondarily utilized as the driving force for ATP synthesis. The latter reaction is catalyzed by a membrane-bound H+-translocating ATPase (see Kozlov and Skulachev, 1977, for a review) which is driven in reverse by AilH+. In intimate conjunction with these more general principles of the chemiosmotic theory, a very important molecular principle was also proposed explaining how H + translocation is linked to respiration. In fact, it seems that historically it was this very principle that led to the proposal of the chemiosmotic theory. A. Respiration-Linked H Translocation-A Historical Note +
Brief
Fifty years ago Lund (1928) drew attention to the possibility that electric events in living cells might be coupled to oxidation-reduction reactions. Lundegirdh (1939,1945) developed the idea further in connection with a theory of anion respiration in plants and suggested that electron transfer as catalyzed by the iron-containing respiratory pigments might be orientated vectorially across a cell layer (not a cellular membrane), so that respiration
305
17. PROTON TRANSLOCATION BY CYTOCHROME OXIDASE
might directly drive transport of anions and cations. Davies and Ogston (1950) developed this idea further, in connection with gastric hydrochloric acid secretion, and also recognized the possibility of coupling respiration with phosphorylation through vectorial H + translocation (see also Davies, 1951; and Conway, 1953). Although it may seem odd from our present perspective, it was not until 1960 that these proposals were explicitly formulated by Robertson (1960) in terms of the separation of electrical charges across the mitochondria1 membrane. Thus the oxidation of hydrogen carriers on one side of the membrane by dioxygen would result in the production of H + on one side and of OH- on the other, the membrane being regarded as electron permeable but proton impermeable (see Robertson, 1960, and Fig. 1). The net function of such a respiratory chain would be to translocate H+ electrogenically. The “folding” of the respiratory chain into several redox loops, each of which includes both electroneutral H translocation and electrogenic electron translocation, was subsequently proposed by Mitchell (1961, 1966) as part of the chemiosmotic theory (Fig. 2A). Figure 2B shows the proposal for cytochrome oxidase function in this theory, namely, as the electrontranslocating limb of the third of three redox loops. B. Properties of an Electron-Translocating Cytochrome Oxidase Two important points are worth noting. In addition to generating A$, the electron-translocating oxidase (Fig. 2B) also catalyzes uptake of one H per electron from the inside (M-side) of the mitochondrion, creating a ApH. Related to this, the ApH generated will, in practice, be as large (per electron transferred) as during the function of one complete redox loop. This is so because the only missing step, namely, electroneutral release of 1 +
membrane of low permeability to ions
1/2 02+ HO ,
FIG. 1. Vectorial electron transfer in the respiratory chain according to Robertson (1960).
/ -
20H-
MARTEN
306
A
WIKSTROM
6
2H’
2n+
2H’
FIG.2. Arrangement of the respiratory chain in redox loops according to Mitchell (1966). (A) The entire respiratory chain. The region inside the rectangle represents cytochrome oxidase. (€3) The cytochrome c oxidase segment of the respiratory chain according to Mitchell (1966).
H /e- t o the C-side of the membrane, will make no significant contribution to ApH because of the large buffering power of the C-phase. It follows that the “half-loop” cytochrome oxidase reaction, as proposed by Mitchell, is thermodynamically equivalent to a complete redox loop, even though no hydrogen ions are released on the C-side (coupled to electron transfer from cytochrome c t o dioxygen). Energetically (but not mechanistically), therefore, cytochrome oxidase functions in this model as a proton pump with a stoichiometry of one H + translocated for each electron transferred. The organization of cytochrome oxidase as such an electron translocator has been generally accepted up to now. Experimental support for this model (Hinkle and Mitchell, 1970; Hinkle, 1973; Hinkle et af., 1972; Papa el af., 1975; Papa, 1976) has been considered strong enough even to provide more general support for electron translocation by respiratory chains, as well as the general concept of vectorial metabolism (Mitchell, 1979). To avoid confusion, it should be mentioned that the earlier findings of Hinkle and collaborators, with proteoliposomes into which purified cytochrome oxidase was incorporated, have often (perhaps unfortunately) been referred to as showing “proton pumping” or “proton translocation” by cytochrome oxidase. In most of these experiments Hinkle used hydrogen donors to the cytochrome c, which was added on the outside of the vesicles. Hence, analogously with Robertson’s scheme (Fig. l), the release of H + into the external medium is expected, but for trivial reasons. In such systems Hinkle and collaborators reported the release of a maximum of +
17. PROTON TRANSLOCATION BY CYTOCHROME OXIDASE
307
one H + per electron transferred (Hinkle, 1973). Consequently, the results were considered important evidence in favor of the electron-translocating model proposed by Mitchell. (However, for revision of these results, see Wikstrom and Saari, 1977; Krab and Wikstrom, 1978; Casey et uf., 1979; Sigel and Carafoli, 1979; Coin and Hinkle, 1979 and below.)
It. THE DISCOVERY OF TRUE PROTON PUMPING BY CYTOCHROME OXIDASE In earlier work (for review, see Wikstrom and Krab, 1979), we found that application of ApH+across the mitochondria1 membrane, with positive polarity in the C-phase, quite specifically induced configurational changes in the heme uu3 system, which was poised such that no electron transfer could occur. This led us to the idea that electron transfer catalyzed by cytochrome uu3 might result in a strained configuration of the hemes and/or their immediate vicinity in the heme pocket, and that relaxation of this strain might be linked to translocation of H + all across the membrane. [As shown recently (Wikstrom, 1981), the configurational change in ferric heme u3 under oxidized “high energy” conditions is due to reversed electron transfer from water to ferricytochrome c with binding of an oxidation product of water to the heme’s 6th (axial) position.] In other words, we envisaged that the oxidase might function as a true redox-linked proton pump in which the electron transfer was “conformationally” linked to proton translocation, and that we had in fact reversed this pump, in part, in our experiments applying ApH+ . While this suggestion was purely speculative, it had the virtue of being testable by comparatively simple experiments. The finding by Jacobs and Sanadi (1 960) that ferrocyanide donates electrons directly to cytochrome c in intact mitochondria provides a method for studying possible H + translocation coupled to electron transport in the terminal region of the respiratory chain (between cytochrome c and oxygen) and thereby provides a simple experimental system for testing the working hypothesis described above. As shown in Fig. 3D, initiation of respiration by the addition of ferrocyanide results in an initial phase of acidification of the mitochondrial suspension (after an initial “alkalinization” artifact, which occurs also without mitochondria). The acidification phase turns off subsequently, and in the final steady state there is net alkalinization of the medium at a rate of - 1 H+/e(Wikstrom, 1977; Wjkstrom and Saari, 1977; Krab and Wikstrom, 1979). Steady state alkalinization is expected from the con-
MARTEN
308
WI KSTROM
C
A
I 1 min
B
l
D
r
1 min
FIG.3. H + translocation linked to ferrocyanide respiration by rat liver mitochondria. The medium, containing 110 mM KCI-I mM HEPES (pH 7.0), was supplemented with 5 . 5 pM rotenone, 0.1 18 pg/ml antimycin, 0.055 p g h l valinomycin, and 2.5 m g h l mitochondria1 protein. The temperature was 24°C. The reaction was started by the addition of 0.8 mM potassium ferrocyanide (at the arrow). In (A) and (B), 0.5 pit4 of carbonylcyanide p-trifluoromethoxyphenylhydraione (FCCP) was also present. (A and C) Oxygen consumption in micromoles electrons per minute. (B and D) pH changes, alkalinization downward (adjacent numbers, micromoles H + per minute). (From Wikstrom and Krab, 1978.)
-
sumption of protons in the overall reduction of oxygen to water by an electron donor:
-
Ferrocyanide + a 0,+ 1H ferricyanide + iHzO (1) The initial proton ejection phase is completely abolished in the presence of a proton-conducting uncoupling agent (Fig. 3B); under such conditions the system behaves simply according to Eq. (1) from the very beginning. Similarly, if the mitochondrial membrane is not rendered permeable to K + by valinomycin, none or very little of the proton ejection phase is observed. These properties of proton ejection suggest very strongly that it is the result of electrogenic proton translocation across the whole membrane. Net proton ejection is expected to cease in the steady state, where the developed pH gradient (alkaline inside) pulls protons inward as fast as they are ejected. Under such conditions only the overall consumption of protons should be seen, according to Eq. (1). It should be recalled that reduction of oxygen by electrons requires +
17. PROTON TRANSLOCATION BY CYTOCHROME OXIDASE
309
uptake of H + immediately upon initiation of oxygen consumption by the addition of ferrocyanide. Hence the initial net production of H + in the C-phase, which occurs at a rate of 1 H+/e- (see Wikstrom and Krab, 1979), suggests either that 2 H+/e- are translocated across the membrane, with 1 H+/e- consumed according to Eq. (1); or that 1 H+/e- is translocated, but 1 H /e- is taken up in addition from the M-phase, combining with the electron derived from cytochrome c, in the reduction of oxygen to water. These two cases cannot be distinguished at present, since their distinction would require full knowledge of the translocation mechanism. Recent results with liposomes incorporated with a cytochrome oxidase preparation lacking subunit 111, allow a distinction in favor of the second possibility (Penttila and Wikstrom, 1981). However, in both cases there will be translocation of two electrical charges across the membrane per transferred electron, in contrast to the situation with the electron-translocating cytochrome oxidase model (Fig. 2B). This prediction, which obviously has important thermodynamic implications, can be tested independently either by measuring the stoichiometry of the electrophoretic counterflux of K + or, in the presence of Ca2+ (which penetrates the membrane without an added ionophore), by determination of the Ca2+/e- stoichiometry for calcium uptake. In both cases the charge/electron stoichiometry has been verified t o be close to 2.0 (Wikstrom, 1978; Sigel and Carafoli, 1978; Krab and Wikstrom, 1979). Experiments with donors other than ferrocyanide, and with methods other than pulsing the mitochondria with reductant, have extensively confirmed the above results (for review, see Wikstrom and Krab, 1979). We may therefore draw a general scheme of the overall proton-translocating function of the oxidase (Fig. 4A), in which the underlying mechanism is fundamentally different from the electron-translocating mechanism of Mitchell (Fig. 2B)-different from thermodynamic, mechanistic, and structural points of view (see Wikstrom and Krab, 1979). It has turned out not to be intuitively easy to appreciate that a function such as that drawn in Fig. 4A would be expected to yield experimental results such as those shown in Fig. 3. Therefore mathematical modeling of the scheme has been carried out (Krab and Wikstrom, 1979), from which computer simulation traces have been drawn in Fig. 4B. These are in excellent qualitative and quantitative agreement with the experimental findings just reviewed (Fig. 3). The conclusions have been greatly strengthened by more recent data with proteoliposomes into which isolated and purified cytochrome oxidase has been incorporated. This system is very much simpler than intact mitochondria, thus greatly decreasing the risk of artifacts. In contrast to the initial findings of Hinkle and collaborators with such proteoliposomes (see Hinkle, 1973), cytochrome oxidase in this system has now been verified +
MARTEN
310
out
in (M)
(C)
-5-
-2H’
1/2H20
-10
-
-15
-
‘.
WIKSTROM
\
.‘ \.
‘
1/40,
FIG. 4. The proton pump model of cytochrome oxidase. (A) Schematic representation based on experimental data. (B) Computer simulation of protonic changes in the extramitochondrial medium upon ferrocyanide oxidation by mitochondria. Pump, the case in which the oxidase functions as in (A). Electr. transloc., the case in which the oxidase functions as in Fig. 2B; uncoupled, the case (with both models) in which the mitochondrial membrane has been rendered fully permeable to protons. (Computer simulation from Krab and Wikstrom, 1979.)
to function as a proton pump, in complete agreement with the results from intact mitochondria (Wikstrom and Saari, 1977; Krab and Wikstrom, 1978; Casey et al., 1979; Sigel and Carafoli, 1979). More recently, these findings have also been confirmed by Coin and Hinkle (1979). Additional confirmation of this view has come from studies with socalled submitochondrial particles, in which the mitochondrial membrane is inverted, making it possible to measure the translocation as uptake of H “from the other side” of the membrane (Wikstrom and Saari, 1977; Sorgato and Ferguson, 1978; Sorgato et al., 1978). +
111.
CONTROVERSY OVER PROTON TRANSLOCATION BY CYTOCHROME OXIDASE
Remarkably the experimental findings showing cytochrome oxidase to function as a redox-linked proton pump continue to be explained away on grounds that have been carefully excluded in published control experi-
17. PROTON TRANSLOCATION BY CYTOCHROME OXIDASE
311
ments. Both Mitchell (Moyle and Mitchell, 1978; Mitchell and Moyle, 1978a, 1979) and Lorusso et al. (1979) who studied only intact mitochondria, suggest that proton ejection observed with cytochrome oxidase proteoliposomes represents scalar net release of hydrogen ions, due to interactions of cytochrome c with the phospholipid membrane. Yet, such an artifact has been repeatedly excluded by demonstrations that the overall consumption of H , in the oxidation of ferrocytochrome c by oxygen as catalyzed by these vesicles, amounts to 1 H+/e- (Wikstrom and Saari, 1977; Krab and Wikstrom, 1978; Casey et al., 1979; Sigel and Carafoli, 1979; Coin and Hinkle, 1979). Furthermore, cytochrome oxidase in proteoliposomes has been shown to translocate two electrical charges across the membrane per transferred electron, as predicted from the earlier work with intact mitochondria and cytochrome oxidase vesicles (Sigel and Carafoli, 1979; Coin and Hinkle. 1979), thus strongly supporting the model in Fig. 4A. The objections and alternative explanations raised (Moyle and Mitchell, i978a,b; Mitchell and Moyle, 1978, 1979) in regard to experiments with intact mitochondria have been carefuily considered and rechecked experimentally. In each case, these objections have been found not t o be valid (see Wikstrom and Krab, 1978; Sigei and Carafoli, i978; Wikstriim and Krab, 1979). More recently, we have been unable to confirm the finding by Lorusso et u3. (1979) of a discrepancy between rates of oxygen consumption and generation of ferricyanide, during ferrocyanide oxidation by mitochondria, which would have suggested rereduction of formed ferricyanide by endogenous mitochondria1 substrates (see also Wikstrom and Krab, 1978). It seems possible that either redox changes in mitochondria1 respiratory chain components or light-scattering changes in the mitochondria caused the discrepancy observed by these authors. The blockage of net H translocation (linked to ferrocyanide oxidation) by 2-n-heptyl-4-hydroxyquinoiineN-oxide (HOQNO), which was observed hv Lorusso et al. (197911is due not to inhibition of electron transfer by this compound but rather to catalysis of ApH-driven H + leakage back across the mitochondria1 membrane (at high concentrations of the antibiotic; Krab and Wikstrom, 1980). Lehninger and his co-workers (Alexandre et al., 1978; Lehninger et al., 1978) have reported, in contrast to Fig. 4A, that 2 H+/e- are released on the C-side of the membrane during ferrocyanide oxidation by mitochondria, with translocation of three eiectrical charges per transferred electron. However, these superstoichiometries (and see Azzone et al., 1978, 1979; Pozzan et a/., 1979) seem likely to result from underestimation of the rates of electron transport (see Wikstrom and Krab, 1979). It may be remarked that hoxh the Azzone group and Lehninger and his collaborators obtained +
+
MARTEN
312
WIKSTROM
their results exclusively with intact mitochondria, and that these results disagree with the extensive quantitative data from cytochrome oxidase proteoliposomes as summarized above.
IV.
MOLECULAR PRINCIPLES AND MECHANISMS OF PROTON TRANSLOCATION
The functioning of cytochrome oxidase as a redox-driven proton pump has two major implications for energy conservation mechanisms. First, it converts twice as much of the redox energy into an electrochemical gradient as the simple electron translocator. And second, it demands fundamentally different molecular and structural arrangements of the enzyme (Wikstrom and Krab, 1979). It is clear, therefore, that questions must now be asked about the molecular details by which the redox reactions of cytochrome oxidase are coupled to the translocation of hydrogen ions.
A. The Relation between “Membrane Bohr” Effects and a Proton Pump While proton pump-type mechanisms have previously been suggested as an alternative t o the redox loop principle of proton translocation (Chance et al., 1970; Papa, 1976), experimental evidence for the existence of such pumps was previously meager. Previous discussions have centered on changes in pK values of acidic groups in the apoprotein, linked to oxidoreduction of the redox center. Such phenomena have been designated “membrane Bohr” effects by analogy with the Bohr effect in hemoglobin (Chance et al., 1970; Chance, 1972; Papa, 1976). Thus Papa et al. (1975, 1976) suggested a “vectorial Bohr mechanism” in explaining their finding that the stoichiometry of proton translocation in the cytochrome bc, segment of the respiratory chain was dependent on the prevalent pH on either side of the mitochondria1 membrane. As pointed out by Boyer (1975; see also Wikstrom and Krab, 1979), mere pK changes linked to oxidoreduction are not sufficient to explain the function of a redox-linked proton pump, though they may be required for kinetic reasons (see below). A membrane Bohr effect alone can account only for proton release or uptake on either side of the membrane, whereas true proton translocation must function cyclically. While the stoichiometry of the protonic shift (per electron transferred) is expected to be pH dependent for a Bohr effect, the cyclical nature of true transport leads to the prediction that transport stoichiometry may both be independent of pH and be a whole number (see below).
17. PROTON TRANSLOCATION BY CYTOCHROME OXIDASE
313
It is significant that membrane Bohr effects are indeed characteristic of the cytochrome aa, system (Wilson et a/., 1972; Van Gelder et af., 1977; Artzatbanov et a/., 1978) and of the 6-type cytochromes of the respiratory chain (see Wikstrom, 1973), as demonstrated by the pH dependence of midpoint redox potentials. Though it may be difficult to interpret such effects as due to redox-dependent pK shifts in the respective oxidoreductase protein-due to the complexities inherent in membraneous systems (see Walz, 1979)-the pH dependence of cytochrome aa3 redox potentials is almost identical in mitochondria and in the purified enzyme (Wilson el a f . , 1972; Van Gelder e t a / . , 1977). If the membrane Bohr effects observed for the 6-type cytochromes, as well as for the cytochrome aa, system, are intimately related to the mechanism of proton translocation (Wikstrom, 1973; Wikstrom and Krab, 1979), they can only represent partial reaction steps in overall proton translocation.
B. General Principles of a Redox-Linked Proton Translocator Figure 5 shows a simplified schematic representation of the function of a redox-linked proton pump in which the transfer of one electron from donor (D) to acceptor (A) is linked to the translocation of one H + across the membrane. To conform to the function of cytochrome oxidase (Fig. 4A), this more general scheme should be supplemented with the uptake of one additional H i from the M-side and combination of this proton with the electron, in the reduction of oxygen to water (cf. Wikstrom and Krab, 1979). The general function of a redox-linked proton pump must follow certain basic rules: 1 . At least one acidic group should be functionally linked to the oxidoreduction state of the redox center. This group must have the property of orientating itself in either of two positions or states, in which it equilibrates with H + either from the M- or the C-side of the membrane (M-state and C-state). 2. The functional linkage between the acidic group and the redox center should be expressed in at least three ways. First, reorientation of this group between C - and M-states, which constitutes the actual translocational steps of the pump, must be highly specific for certain states of the translocator in order to ensure minimal energy losses. In Fig. 5 , only the reduced protonated and oxidized unprotonated states of the translocator possess a significant probability of reorienting the acid group between the
MARTEN
314 CRISTAE
WIKSTROM
MATRIX
MEMBRANE
Dlredl
D lox1
Iox-
Red-
- H+
, H I _ _
Alred!
H+-Red
Aloxl
FIG. 5 . General reaction scheme for a redox-linked proton pump. Two configurations of the system are shown. Electrons are received from the donor couple (D) by the oxidized form of the redox center (Ox-) in a state of the enzyme in which there is protonic contact with the matrix (M) phase via a proton channel (dotted line on the right). Electrons are donated to an acceptor couple (A) by the reduced redox center in a state of the enzyme in which there is protonic contact with the other (C) side of the membrane via a proton channel (dotted line on the left). lnterconversion between the C- and M-states of the enzyme are depicted by the thick open arrows. (The scheme is simplified from Wikstrilm and Krab, 1979.)
C- and M-states (thick open arrows). Second, electrons should be donated to the translocator only in the M-state and received by the acceptor only from the C-state (Fig. 5 ) . Third, a change in the oxidoreduction state of the translocator should shift the pK of the acidic group. The necessity for this shift is discussed below. For all relevant oxidoreduction systems the membrane Bohr effects are such that the pK decreases on oxidation and increases on reduction. Thus H + release is expected to be linked to oxidation and H + uptake to reduction, as depicted in Fig. 5 . 3. Passive proton-conducting channels in the apoprotein (Nagle and Morowitz, 1978; Williams, 1975; Kayalar, 1979), connecting the acidic group with the aqueous C- and M-phases on each side of the membrane, would have the effect of shortening the distance over which reorientation of this group would need to occur during turnover. Note, however, that this “reorientation” need not necessarily be a physical movement of the
17. PROTON TRANSLOCATION BY CYTOCHROME OXIDASE
315
acidic group with respect to the protein or membrane lattice. It could also be effected by the gating of proton channel structures. 4. Transfer of one electron equivalent across the translocator releases FAE, of energy, where AE,, is the difference in oxidoreduction potential between the donor and acceptor redox couples. Since the magnitude of AE, across coupling sites of the respiratory chain is at least 200 mV, and since the efficiency of coupling must be considered to be good in this kind of biological energy converter, the redox center of the translocator should operate at two widely different oxidoreduction potentials, upon electron acceptance and electron donation, respectively. If such modulation were not possible, kinetic insufficiency would arise because the probability of occurrence of either the reduced or oxidized state of the center would become too low (see DeVault, 1971; Wikstrom and Krab, 1979). In other words, the C- and M-states of the translocator represent two redox couples, which should be in near equilibrium with the acceptor and donor, respectively, for optimal thermodynamic efficiency. Kinetic versatility can therefore be ensured only if the midpoint potentials of the two redox couples roughly match the Eh of donor and acceptor couples. It follows that the redox couple in the M-state must have a midpoint potential 200 mV more negative than the corresponding midpoint potential in the C-state. Although there may be many means of introducing this modulation in midpoint potential, one of the simplest is t o shift the pK of the acidic group during a redox change at the redox center (see rule 2 above). This is particularly attractive, since it allows the potentials to be modulated directly by the electrochemical activity of H + . Hence the presence of membrane Bohr effects may be important in conferring kinetic compatibility and control, even though they do not seem to be required for translocation as such. 5 . Since the pump generates ApH+across the membrane, A&+ should have the effect of slowing down either H + uptake from the M-side or H + release t o the C-side (or both), with consequent deceleration of electron flow through the converter. In mitochondria, this phenomenon, known as respiratory control, can slow respiration by a factor of 10 or more, the residual (state 4) rate usually being ascribed to unspecific leakage of H + inward, driven by the high electrochemical proton gradient. Even if this state-4 respiration could be ascribed entirely to lack of specificity in the electron transport or acid group translocation steps of the transducer (see rule 2), such “slipping” reactions should have an aggregate probability at least one order of magnitude smaller than the main pathways of the translocator. This would ensure that the intrinsic stoichiometry of proton translocation will be essentially independent of pH on either side of the membrane, at least in the physiological pH range. (As noted above,
-
316
MARTEN
WIKSTROM
however, pH changes should still affect the kinetics of the transducer.) The high specificity also ensures that the stoichiometry of H translocation is an integer (H+/e-= 1 in Fig. 5 ) . It goes without saying that actually observed stoichiometries of proton translocation may be much lower and variable because of (variable) proton leaks through the membrane (Wikstrom and Krab, 1979). +
C. Possible Molecular Mechanism of Proton Translocation by Cytochrome Oxidase In cytochrome oxidase it seems likely that the heme groups are more directly involved in proton translocation than the copper groups, since it is the former which exhibit pH-dependent midpoint potentials (see above). Although it is not possible a priori to predict whether the coupling of hemes with the acidic translocation group is direct or indirect, the former possibility is more easily amenable to experimental test. At present, therefore, we have focused on modes of coupling that would directly involve the heme or its immediate vicinity in the apoprotein, and some experimental progress has been made which allows a tentative proposal for a possible molecular mechanism of proton translocation. Wikstrom and Saari (1975) previously reported that Ca2+-without penetrating the membrane barrier-perturbs the spectrum of ferrocytochrome a when added to intact mitochondria. The spectral perturbation is very similar to that induced upon energization of the mitochondria (i.e., by a large electrochemical gradient for protons). Saari and Wikstrom (1976) have shown a competition between H + and Ca2+(but no other cations) for this effect, which can also be studied in the isolated enzyme. While the effect had seemed to be an indirect one, we have recently found that Ca2+ perturbs the spectrum of isolated reduced bisimidazole heme A in a very similar manner. The spectral perturbation disappears entirely upon esterification of the carboxylic residues of the propionate side chains of the heme group (Saari eta/., 1980), making it very likely that it is a direct interaction between the cation and these residues. Since spectral interaction between isolated hemes and divalent cations is very unspecific, but spectral interaction in the enzyme is highly specific for Ca2+and H +, the propionate side chain is probably buried in the apoprotein which confers ion specificity. Yet it is clear that Ca2+ can reach this site from the C-side of the mitochondria1 membrane without penetrating the membrane proper (Wikstrom and Saari, 1975). While this finding may have some important implications with respect to the location of the hemes in the enzyme in situ, the spectral shift upon “energization” (Wikstrom,
17. PROTON TRANSLOCATION BY CYTOCHROME OXIDASE
317
1972) could indicate that the propionate carboxyl residues become protonated, hence that they might function as the pump’s outlet to the C-side of the membrane. A few years ago Caughey et al. (1973) made the very interesting observation, based on nuclear magnetic resonance (NMR) studies of pyridine hemochromes, that there is an interaction between the methyl hydrogens of the methyl ester of the propionates and the pyridine axial ligands. Thus, at least in solution, the propionate side chains may be mobile and also long enough to “flip up,” from their regularly depicted position at the bottom edge of the heme, to make direct contact with nitrogenous axial ligands. This would provide an interesting possibility of proton transfer between the axial ligand and the propionate carboxyl. The axial ligands of cytochrome a are probably histidine imidazoles (Babcock et al., 1979; Blumberg and Peisach, 1979) whose protonation state is closely linked to the redox state of the heme iron. Figure 6 presents a very tentative scheme showing how proton translocation might occur in such a system. Reduction of the heme iron results in an increased pK of the hydrogen-bonded imidazole-carboxylate complex and subsequent uptake of H + from a proton channel connecting to the M-phase. The protonated carboxylic acid residue, not being attracted by the axial ligand, “flips” to its position below the heme edge. Oxidation of the heme iron and dissociation of the carboxylic acid with H + release into the C-side of the proton channel complete the sequence. By this mechanism, H + should be translocated in a direction parallel to the heme plane. Since overall H + translocation is perpendicular to the plane of the membrane, the heme involved might be expected to lie with its plane perpendicular to the plane of the membrane. It is interesting that the work of Erecinska et al. (1978) has established that this is indeed the orientation of the hemes of cytochrome oxidase. More recent data have strongly implicated the heme of cytochrome a as the redox center of the proton pump and the mitochondrially made subunit I11 as intimately involved with the proton translocating mechanism (see Penttila and Wikstrdm, 1981; Wikstrom et al., 1981). It is also interesting to note that, with these first stumbling steps into the realm of molecular mechanisms of redox-linked proton translocation, there already appears to be some analogy to the much better characterized proton translocation mechanism of bacteriorhodopsin (Honig, this volume), particularly with respect to the proximity of the center of proton translocation to that of light absorption-oxidoreduction. While bacteriorhodopsin is a light-driven proton pump, and as such quite unique in biological energy conservation, it seems that the discovery of cytochrome oxidase as a proton pump could be taken as an indication of the possibility that redox-linked
MARTEN
31 8
r-
WIKSTROM
1
-Fe"'-N
NH
*
lV coo-
L
A
FIG.6. Tentative mechanism of proton translocation linked to redox changes in a heme center. The heme is shown from the side with the iron (Fe) in the center and with one axial imidazole ligand (tbe other axial ligand is not shown) plus the propionate side chain with its carboxylic group. The latter is assumed to exist in one of two positions, either interacting with the imidazole by means of hydrogen bonding, or below the heme edge. Proton uptake from the M-phase is indicated by H& and proton release to the C-phase by Hb. Bracketed structures indicate unstable intermediates. Thick black arrows depict transitions between states in protonic contact with the C- and M-phases, respectively (cf. Fig. 5). The axial imidazole is assumed to be in protonic contact with the M-phase at all times, while the propionate carboxyl is in protonic contact with the C-phase only when it is not hydrogen-bonded to the imidazole (lower three intermediates). Fe"' and Fe" indicate ferric and ferrous states of iron, respectively.
proton translocation may occur also more generally by pump-type mechanisms rather than by redox loops. This might be so not only in other regions of the respiratory chain but also in bacterial and photosynthetic energy conservation. ACKNOWLEDGMENTS
I am grateful for the continuous support of the Sigrid Juselius Foundation in Helsinki. I also wish to acknowledge stimulating and helpful discussions with Drs. Klaas Krab and Clyde Barlow, who have contributed much to the experiments and ideas presented in this article.
REFERENCES Alexandre, A., Reynafarje, B., and Lehninger, A. L. (1978). Proc. Natl. Acad. Sci. U.S.A. IS, 5296-5300.
17. PROTON TRANSLOCATION BY CYTOCHROME OXIDASE
319
Artzatbanov, V. Yu., Konstantinov, A. A., and Skulachev, V. P. (1978). FEBS Lett. 87, 180- 185. Azzone, G. F., Pozzan, T., DiVirgilio, F., and Miconi, V. (1978). In “Frontiers of Biological Energetics” (P. L. Dutton, J. S. Leigh, and A. Scarpa, eds.), pp. 375-383. Academic Press, New York. Azzone, G. F., Pozzan, T., and DiVirgilio, F. (1979). J . B i d . Chem. 254, 10206-10212. Babcock, G. T., van Steelandt, J., Palmer, G., Vickery, L. E., and Salmeen, 1. (1979). In “Cytochrome Oxidase” (T. E. King, Y. Orii, B. Chance, and K. Okunuki, eds.), pp. 105-1 15. Elsevier, Amsterdam. Blasie, J. K., Erecinska, M., Samuels, S., and Leigh, J. S., Jr. (1978). Biochim. Biophys. Act0 501, 33-52. Blumberg, W. E., and Peisach, J. (1979). In “Cytochrome Oxidase” (T. E. King, Y. Orii, B. Chance, and K. Okunuki, eds.), pp. 153-159. Elsevier, Amsterdam. Boyer, P. D. (1975). FEBS Lett. 58, 1-6. Buse, G., Steffens, G. J., Steffens, G. C. M., and Sacher, R. (1978). In “Frontiers of Biological Energetics” (P. L. Dutton, J. S. Leigh, and A. Scarpa, eds.), pp. 799-807. Academic Press, New York. Capaldi, R. A., and Briggs, M. (1976). In “The Enzymes of Biological Membranes” (A. Martonosi, ed.), Vol. 4, pp. 87-102. Wiley, New York. Carroll, R. C., and Racker, E. (1977). J. Biol. Chem. 252, 6981-6990. Casey, R. P., Chappell, J. B., and Azzi, A. (1979). Biochem. J. 182, 149-156. Caughey, W. S., Barlow, C. H., O’Keeffe, D. H., and O’Toole, M. C. (1973). Ann. N. Y. Acad. Sci. 206, 296-299. Caughey, W. S., Wallace, W. C., Volpe, J. A., and Yoshikawa, S. (1976). In “The Enzymes” (P. D. Boyer, ed.), Vol. 13, pp. 299-344. Academic Press, New York. Chance, B. (1972). FEBS Lett. 23, 3-20. Chance, B., Crofts, A. R., Nishimura, M., and Price, B. (1970). Eur. J . Biochem. 13, 364-374. Coin, J. T., and Hinkle, P. C. (1979). In “Membrane Bioenergetics” (C. P. Lee, G. Schatz, and L. Eruster, eds.), pp. 405-412. Addison-Wesley, Reading, Massachusetts. Conway, E. J. (1953). “The Biochemistry of Gastric Acid Secretion.” Thomas, Springfield, Illinois. Davies, R. E. (1951). B i d . Rev. 26, 87-120. Davies, R. E., and Ogston, A. G. (1950). Biochem. J. 46, 324-333. DePierre, J. W., and Ernster, L. (1977). Annu. Rev. Biochem. 46, 201-262. DeVault, D. (1971). Biochim. Biophys. Acta 226, 193-199. Dockter, M. E., Steinemann, A., and Schatz, G. (1977). In “Structure and Function of Energy-Transducing Membranes” (K. Van Dam, and B. F. Van Gelder, eds.), pp. 169-176. Elsevier, Amsterdam. Dockter, M. E., Steinemann, A., and Schatz, G. (1978). J. Biol. Chem. 253, 311-317. Downer, N. W., Robinson, N . C., and Capaldi, R. A. (1976). Biochemistry 15, 2930-2935. Erecinska, M . , and Wilson, D. F. (1978). Arch. Biochem. Biophys. 188, 1-14. Ereciriska, M., Wilson, D. F., and Blasie, J. K. (1978). Biochim. Biophys. Acta 501, 53-62. Eytan, G. O., Carroll, R. C., Schatz, G., and Racker, E. (1975). J. B i d . Chem. 250, 8598-8603. Frey, T. G., Schatz, G., and Chan, S. H. P. (1978). In “Frontiers of Biological Energetics” (P. L. Dutton, J. S. Leigh, and A. Scarpa, eds.), pp. 817-824. Academic Press, New York. Henderson, R., Capaldi, R. A,, and Leigh, J. S. (1977). J. Mol. Biol. 112, 631-648. Hinkle, P . C. (1973). Fed. Proc. Fed. A m . SOC. Exp. B i d . 32, 1988-1992. Hinkle, P., and Mitchell, P. (1970). J . Bioenerg. 1, 45-60.
320
MARTEN
WIKSTROM
Hinkle, P. C., Kim, J. J., and Racker, E. (1972). J. Biol. Chem. 247, 1338-1339. Jacobs, E. E., and Sanadi, D. R. (1960). Biochim. Biophys. Actu 38, 12-34. Kayalar, C. (1979). J. Membr. Biol. 45, 37-42. Kozlov, 1. A., and Skulachev, V. P. (1977). Biochim. Biophys. Actu 463, 29-89. Krab, K., and Wikstrom, M. (1978). Biochim. Biophys. Actu 504, 200-214. Krab, K., and Wikstrom, M. (1979). Biochim. Biophys. Actu 548, 1-15. Krab, K., and Wikstrbm, M. (1980). Biochem. J. 186, 637-639. Kunze, V., and Junge, W. (1977). FEBS Lett. 80, 429-434. Lehninger, A. L., Reynafarje, B., and Alexandre, A. (1978). In “Frontiers of Biological Energetics” (P. L. Dutton, J. S. Leigh, and A. Scarpa, eds.), pp. 384-393. Academic Press, New York. Lemberg, M. R. (1969). Physiol. Rev. 49, 48-121. Lorusso, M., Capuano, F., Boffoli, D., Stefanelli, R., and Papa, S. (1979). Biochem. J. 182, 133-147. Lund, E. J. (1928). J. Exp. Zool. 51, 265-307. LundegPrdh, H. (1939). Nature (London) 143, 203. Lundegirdh, H. (1945). Ark. Bot. A . 32, (12), 1-139. Maley, G. F., and Lardy, H . A. (1954). J. Biol. Chem. 210, 903-909. Malmstrom, B. G. (1974). Q. Rev. Biophys. 6, 389-431. Malmstrom, B. G. (1979). Biochim. Biophys. Actu 549, 281-303. Massari, S., and Azzone, G. F. (1970a). Eur. J . Biochem. 12, 301-309. Massari, S., and Azzone, G. F. (1970b). Eur. J. Biochem. 12, 310-318. Mitchell, P. (1961). Nature (London) 191, 144-148. Mitchell, P. (1966). “Chemiosmotic Coupling in Oxidative and Photosynthetic Phosphorylation.” Glynn Research, Bodmin, U. K. Mitchell, P. (1976). Biochem. Soc. Trans. 4, 399-430. Mitchell, P. (1979). Eur. J. Biochem. 95, 1-20. Mitchell, P., and Moyle, J. (1978). In “Frontiers of Biological Energetics” (P. L. Dutton, J. S. Leigh, and A. Scarpa, eds.), pp. 342-350. Academic Press, New York. Mitchell, P., and Moyle, J. (1979). In “Cytochrome Oxidase” (T. E. King, Y. Orii, B. Chance, and K. Okunuki, eds.), pp. 361-372. Elsevier, Amsterdam. Moyle, J., and Mitchell, P. (1978a). FEBS Lett. 88, 268-272. Moyle, J., and Mitchell, P. (1978b). FEBS Lett. 90, 361-365. Nagle, J. T., and Morowitz, H. J. (1978). R o c . Nutl. Acud. Sci. U.S.A. 75, 298-302. Nicholls, P., and Chance, B. (1974). In “Molecular Mechanisms of Oxygen Activation” (0.Hayashi, ed.), p. 479. Academic Press, New York. Papa, S. (1976). Biochim. Biophys. Actu 456, 39-84. Papa, S., Lorusso, M., Guerrieri, F., and Izzo, G. (1975). In “Electron Transfer Chains and Oxidative Phosphorylation” (E. Quagliariello, S. Papa, F. Palmieri, E. C. Slater, and N. Siliprandi, eds.), pp. 317-327. North-Holland Publ., Amsterdam. Penttila, T., and Wikstrbm, M. (1981). In “Vectorial Reactions in Electron and Ion Transport in Mitochondria and Bacteria” (F. Palmieri, eds.). Elsevier, Amsterdam. In press. Penttila, T., Saraste, M., and Wikstrom, M. (1979). FEBS Lett. 101, 295-300. Pozzan, T., DiVirgilio, F., Bragadin, M., Miconi, V., and Azzone, G. F. (1979). Proc. Nutl. Acud. Sci. U.S.A. 76, 2123-2127. Robertson, R. N. (1960). Biol. Rev. 35, 231-264. Ruben, G. C., Telford, J. N., and Carroll, R. C. (1976). J. Cell Biol. 68, 724-739. Saari, H. T., and Wikstrom, M. (1976). Int. Congr. Biochem. Hamburg (Abstract). Saari, H., Penttila, T., and Wikstrom, M. (1980). J. Bioenerg. Biomembr. 12, 325-338. Sigel, E., and Carafoli, E. (1978). Eur. J. Biochem. 89, 119-123. Sigel, E., and Carafoli, E. (1979). J. Biol. Chem. 254, 10572-10574.
17. PROTON TRANSLOCATION BY CYTOCHROME OXIDASE
321
Sorgato, M. C., and Ferguson, S. J. (1978). FEBS Lett. 90, 178-182. Sorgato, M. C., Ferguson, S. J., Kell, D. B., and John, P. (1978). Biochem. J. 174,237-256. Van Gelder, B. F., van Rijn, J. L. M. L., Schilder, G. J. A., and Wilms. J. (1977). In “Structure and Function of Energy-Transducing Membranes” (K. Van Dam and B. F. Van Gelder, eds.), pp. 61-68. Elsevier, Amsterdam. Walz, D. (1979). Biochim. Biophys. Acta 505, 279-353. Wikstrdm, M. K. F. (1972). Biochim. Biophys. Acta 283, 385-390. Wikstrom, M. K. F. (1973). Biochim. Biophys Acta 301, 155-193. Wikstrom, M. (1977). Nature (London) 266, 271-273. Wikstrom, M. (1978). In “The Proton and Calcium Pumps,” (G. F. h o n e , M. Avron, J. C. Metcalfe, E. Quagliariello, and N. Siliprandi, eds.), pp. 215-226. Elsevier, Amsterdam. Wikstr6m, M. (1981). Proc. Natl. Acad. Sci. U.S.A. 78, 4051-4054. Wikstrom, M., and Krab, K. (1978). FEBS Lett. 91, 8-14. Wikstrom, M., and Krab, K. (1979). Biochim. Biophys. Acta 549, 177-222. Wikstrom, M., and Saari, H. T. (1975). Siochim. Biophys. Acta 408, 170-179. Wikstrom, M., and Saari, H. T. (1977). Biochim. Biophys. Acta 462, 347-361. Wikstrom, M., Harmon, H. J., Ingledew, W. J., and Chance, B. (1976). FEBS Left. 65, 259-277. Wikstrom, M.,Krab, K., and Saraste, M. (1981). Annu. Rev. Biochem. 50, 623-655. Williams, R. J. P. (1975). In “Electron Transfer Chains and Oxidative Phosphorylation,” (E. Quagliariello, S. Papa, F. Palmieri, E. C. Slater, and N. Siliprandi, eds.), pp. 417-422. North-Holland Publ., Amsterdam. Wilson, D. F., Lindsay, J. G., and Brocklehurst, E. S. (1972). Biochim. Biophys. Acta 256, 277-286.
This Page Intentionally Left Blank
CURRENT TOPICS IN MEMBRANES A N D TRANSPORT, VOLUME 16
Chapter 18
Electrogenic Reactions of the Photochemical Reaction Center and the Ubiquinone-Cytochrome blc, Oxidoreduct ase P . LESLIE DUTTON, PA UL MUELLER, * DANIEL P . 0 'KEEFE, NIGEL K. PA CKHAM, *,' ROGER C. PRINCE, A N D DAVID M . TIEDE Department of Biochemistry and Biophysics University of Pennsylvania Philadelphia, Pennsylvania *Department of Molecular Biology Eastern Pennsylvania Psychiatric Institute Philadelphia, Pennsylvania
I . Introduction ........................................................................................ 11. The Reaction Center Protein .......................... A . Light-Induced Electric Cu ....................................... B. Notes on the Structure of the Reaction Center ....................................... C. Free Energies of the Electron Transport Steps in the Cytochrome c,-Reaction Center Complex ............................................
B. C.
E.
Redox Coupling of the Reaction Center and the Q-b/c2 Oxidoreductase Electrogenic Events with Possible Sources of the Electrogenic Reaction ....................................... References ....................................................... ..............
324 325 326 331 333 335 335 337 339 339 340 342
Present address: Department of Botany, Imperial College of Science, London SW7 2BB, United Kingdom.
323
Copyright I:1982 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-153316-6
P. LESLIE DUTTON et
324
1.
a/.
INTRODUCTION*
The photosynthetic bacterium Rhodopseudomonas sphaeroides has a light-driven cyclic electron transport system organized to generate an electric potential and a pH gradient across the cytoplasmic membrane. Although the details are lacking, it is widely accepted that the electrochemical gradients so generated are harnessed for ATP production, NAD+ reduction, and solute transport. The electron transport cycle is composed of two major redox protein complexes within the chromatophore membrane. One is the reaction center (RC) protein which absorbs light energy and initiates electron transfer leading t o the oxidation of cytochrome c2 and the reduction of a ubiquinone; the other is the ubiquinone-cytochrome b/c, (Q-b/c2)oxidoreductase which in effect returns the reducing equivalent from ubiquinone to the ferricytochrome c,, thereby completing the cycle. Evidence is accumulating that both protein complexes contribute directly to the generation of membrane potential, and under optimal conditions two protons are incorporated into the membrane for the net transport of one electron around the cycle. The current view is that these two protons are released on the other side of the membrane and that their efflux through the ATPase complex results in the synthesis of a single molecule of ATP. In this article we shall summarize results from several approaches that
*
Discussions of reaction center chemistry inevitably require an efficient shorthand for denoting a large number of redox components, reactions, and reaction parameters. The central elements of that shorthand-as used in this article-are set out in Scheme 1 (described in Section 11) and in Fig. 6. Redox changes involving only electron flow are designated by the appropriate addition and subtraction of charge symbols: ( + ) and (-) or the unpaired electron (-) in a neutral species, and redox changes involving one or two protons are shown by H and H2. The slash (1)is used to separate a redox couple, whether forms of a single substance (e.g., cyt b / b + , as diagramed in Fig. 6) or separate redox elements (e.g., Q-b/c2, for the span from ubiquinone-cytochrome b to cytochrome c2).Various individual redox elements are identified by two different subscripting systems: historical and potentiometric. The historical names: e.g., cytochrome b, c, and their later-identified congeners cytochrome b, and cytochrome c2 are convenient but relatively uninformative. More specific information pertinent to the discussion is conveyed, for example, by the names cytochrome b,,, and cytochrome b_w, which denote b-type cytochromes having midpoint potentials of 155 mV and -90 mV, pH 7, respectively. (Although most of the redox elements were initially identified by their absorption spectra, which has given rise to a third subscripting nomenclature, that notation is avoided in this article for the sake of simplicity.) Finally, several symbols are used to denote electric potentials: Eh is the redox potential of a given substance under the prevailing conditions, referred to the standard hydrogen electrode; Em is the midpoint potential, that is, the redox potential at which the substance is exactly half-oxidized and half-reduced; AEh and AE, are differences in redox and midpoint potential, respectively; and A$ is the transmembrane difference in electric potential (membrane potential). It should also be noted that, by the standard convention-followed in this discussion, redox potentials positive to the standard hydrogen electrode (-0) are plotted downward but referred to as high potentials; and negative redox potentials are plotted upward but referred to as low potentials.
18. ELECTROGENIC REACTIONS OF
RC
AND Q-b/C, OXIDOREDUCTASE
325
together contribute to the view of the RC as a light-activatable redox protein that generates both a redox potential (A&) and a membrane potential (A$). We shall then go into the evidence which describes the Q-b/c, oxidoreductase as a protein capable of using the A E h generated by the RC to drive the formation of further A$ and the translocation of protons across the membrane.
II.
THE REACTION CENTER PROTEIN
The RC has been the subject of much research since its isolation in Clayton's laboratory in the mid-1960s (Reed and Clayton, 1968). A book (Clayton and Sistrom, 1978) and several comprehensive reviews (Blankenship and Parson, 1978; Dutton et a/., 1979; Olson and Thornber, 1979) on its composition, its redox chemical complement, and its function are now available. Scheme 1 summarizes the redox components known to be functional, and the forward kinetics of light-activated electron transport through them. In the scheme, ferri- or ferrocytochrome c is the watersoluble cytochrome c, that serves as an electron donor to the RC; in the RC (BChl), is a bacteriochlorophyll dimer, BPh is bacteriopheophytin, and Q, and QI1are the RC primary and secondary quinones, respectively. The times given are approximate halftimes. 1. Ferrocytochrome c (BChl), BPh QI Q,I
t
2. Ferrocytochrome c (BChl):
J J
BPh QI QI1
<10 psec
3. Ferrocytochrome c (BChl);
BPhr Qr
150 psec
4. Ferrocytochrome c (BChl);
BPh QI;
41
usually
'Ferricytochrorne
c (BChl), BPh Q,' QII
J
c
100 psec
(BChl), BPh QI Q1<
SCHEME 1
P. LESLIE DUTTON
326
et a/.
As already indicated in Section I, the reactions of Scheme 1 are considered to be directed across the photosynthetic membrane, cytochrome c, being on one side of the membrane (Prince et al., 1975) and proton binding (Chance et al., 1970; Halsey and Parson, 1974; Petty and Dutton, 1976)-to what is widely assumed to be Q,,r-occurring on the other. The electrogenic nature of the charge separation has been demonstrated by making use of several indirect indicators of the light-generated A$: the carotenoid band shift (Jackson and Crofts, 1969; Dutton, 1971; Jackson and Dutton, 1973; Packham et al., 1978; Takamiya and Dutton, 1977), oxidation-reduction poise shifts between (BChl), and cytochrome c, (Takamiya and Dutton, 1977), delayed light emission (Crofts et al., 1972), and externally added oxonol dyes (Chance and Baltscheffsky, 1975; Bashford et al., 1979a). A. Light-Induced Electric Current Measurements More direct A$ measurements have been made (Barsky et al., 1976; Drachev et al., 1976) across preformed planar phospholipid membranes to which were added liposomes incorporating RCs; however, the overall geometry of these membranes is not certain. Recently we have endeavored to characterize better the structural features of the RC in the planar membrane and to obtain an alternative view of the electrogenic reactions. This has been approached by forming planar black lipid bilayer membranes containing RC (Packham et al., 1980), from octane-phospholipid solutions ON
1
/OFF
A RC Membranes No additions
B RC Membranes plus
-
Ferrocyt 2 on one side
C RC-Ubiquinone Membranes plus Ferrocyt 2 on one side
D As C but plus g-phenanthroline added to side opposite to Ferrocyt c
Ir
I
-
10.' A
4 IS
k-
T
FIG. 1. Steady state light-induced electrical currents across the RC-phospholipid membranes. RC concentration in the RC-phospholipid-octane solution from which the membranes were formed was 8.3 pM. (A) RC membrane alone. (B) Similar to (A) but with 25 phi' ferrocytochrome c added to the aqueous phase of one side of the membrane. (C) Similar to (B) but with 180 phi' ubiquinone-10 added to the RC-phospholipid-octane solution before membrane formation. (D) Similar to (C) but with 5 mM o-phenanthroline added to the aqueous side opposite the ferrocytochrome c.
18. ELECTROGENIC REACTIONS OF
RC
A No additions
AND
Q-blC, OXIDOREDUCTASE
327
-.L
-Lddd
B Plus Ferrocyt 2 added to one side
C As B but with Ferricyonide added to side opposite Ferrocyt c
-&e&k ,
4 x
A T
10"O
A
I
1 3 0r n s k
FIG.2. Flash-induced electrical currents across RC-ubiquinone-phospholipid membranes. The RC concentration as in Fig. 1. Ubiquinone-10 was added at 180 pMto this solution before membrane formation. (A) RC-ubiquinone membrane alone. (B) Similar to (A) but with 25 pM ferrocytochrome c added to one side of the membrane. (C) Similar to (B) but with 20 mM potassium ferricyanide added to the side of the membrane opposite that containing ferrocytochrome c. Because of the time constant of the voltage-clamped circuit, the duration of the recorded flash-induced currents is 1-2 orders of magnitude larger than both the flash duration and the sum of the electron transfer rates from ferrocytochrome c to the ubiquinone.
(Kendall-Tobias and Crofts, 1979; Schonfeld et al., 1979), and measuring flash-induced electric currents under voltage-clamped conditions (Mueller and Rudin, 1969). The work has yielded the following results and working conclusions. 1. Reaction centers are deposited in the bilayer membrane as an equal mix of vectorially opposing populations. Support for this model comes from the following findings: (a) Currents are not detected in RC membranes alone (Fig. l A , 2A, and 3A). Figure 4A represents a schematic
A No Additions
B Ferricyonide on one side
7
C As B but with Ferrocytochrorne the other
c on side
lighton , -
I
FIG. 3. Steady state light-induced electrical currents across the RC-phospholipid membranes. Conditions as in Fig. 2. (A) RC-ubiquinone membrane alone. (B) Similar to (A) but with 20 mMpotassium ferricyanide added to one side of the membrane. (C) Similar to (B) but with 25 pM ferrocytochrome c added to the side of the membrane opposite to that containing the potassium ferricyanide.
A
hu
c-
60 ms
B
hv
hv
___c
18. ELECTROGENIC REACTIONS OF
RC
AND Q-b/C, OXIDOREDUCTASE
329
model showing the two opposing populations of RCs and the canceling effects of the electron transport generating no net current across the membrane. The 60-msec dark (back) reaction is characteristic of the (),--to(BChl),+ reaction in RCs not supplemented with Qr or ferrocytochrome c. (b) Currents are observed if the symmetry of the populations is disturbed. This can be done by adding ferrocytochrome c (Figs. 1B-D and 2B and C) or potassium ferricyanide (Fig. 3B) to one side of the membrane, or both these compounds to opposite sides of the membrane (Figs. 2C and 3C). The direction, but not the magnitude of the current response, is dependent upon the side to which the addition is made.
C
60 ms
D Ferricyanide
hv
__c
FIG.4. Schematic diagram of the symmetrical distribution of reaction centers in the membrane. (A) No net current flows unless symmetry is disturbed by addition of ferrocytochromec (B) or ferricyanide (C) to one side of the membrane, or by addition of both reagents, one to each side of a single membrane (D).
330
P. LESLIE DUTTON
et a / .
2. Currents elicited by steady light or by single-turnover flashes are understandable in terms of the known electron transport processes associated with the RC. Figure 1B shows that cytochrome c added to one side of the membrane supports a small transient current. RC phospholipid membranes supplemented with an excess of ubiquinone-10 show enhancement of light-induced currents, as in Fig. 1C. The typical photocurrent response comprises a transient peak current which relaxes in the light t o a steady state level. A transient discharging current in the opposite direction is observed when the light is extinguished. Presumably, ubiquinoneinduced enhancement of the current response arises (a) from reconstitution of electron transfer from QI to QII, since QII is entirely lost during preparative procedures (Q, is also depleted), and perhaps (b) from subsequent electron transfer to the excess ubiquinone “pool” (Qp) in the membrane. Such a conclusion is supported by the dramatic effect (Fig. 1D) caused by the addition of o-phenanthroline. This well-known inhibitor of electron transport from QI to QII(see Clayton and Sistrom, 1978) returns the lightinduced current response to the same level observed before the addition of ubiquinone (Fig. 1B). These processes are examined in more detail in Fig. 2. With ferrocytochrome c only added to one side of the RC ubiquinone membranes, single-turnover flash excitation induces a small current following the first flash, but subsequent flashes induce a progressively increasing transient. Thus, on the first turnover both sets of RC populations are activated, but their opposing directions of electron transport through the membrane result in little measured current. By the second and subsequent turnovers, delivered with a 25-msec periodicity, only the RC population accessible to ferrocytochrome c is returned to a functional state, while the other population is still essentially in the photochemically inactive (BChl),? BPh QI QIls(or Qp7)state which takes seconds to recover. The steps of Fig. 2B are presented schematically in Fig. 4B. The failure to obtain a full current transient on the first turnover is circumvented (Fig. 2C) by the addition of potassium ferricyanide to the side opposite the ferrocytochrome c. We interpret these results to mean that ferricyanide oxidizes the (BChl), population accessible to its side, thereby inactivating it and removing opposition to the RC-cytochrome c population. This would enable a net current transient to be seen following even the first turnover. Figure 4C shows a schematic diagram of the ferricyanide effect. 3. Electrogenic reactions are contained in the electron transfer steps between (BChl), and QI, and between ferrocytochrome c and (BChl),. Figure 3 shows an experiment with ferricyanide present, as discussed above, on one side of the RC membranes (no added ubiquinone or ferrocytochrome c, so the experiment is restricted to only one turnover of the RC). Under these conditions (Fig. 3A and see the schematic of Fig. 4C) a
18. ELECTROGENIC REACTIONS OF
RC
AND Q-b/C, OXIDOREDUCTASE
331
current transient is obtained, indicating that at least one of the steps involved in the (BChl),- BPh- Q1 electron transfer sequence is electrogenic. The addition of ferrocytochrome c to the opposite side increases the transient by about one-third (see Fig. 3B and the schematic of Fig. 4D), implying that this reaction also has some electrogenic character, though not as much as that observed within the RC proper. The electrogenic nature of this step is corroborated in Fig. 2B, where a small current transient is consistently detected on the first flash; in the absence of ferricyanide, this could be the net current between (BChl),'- BPh- Q? operating one way and cytochrome c+ (BChl),-BPh-Q7 operating the other. The result is qualitatively consistent with indirect measurements of A$ in natural membranes (using, e.g., the carotenoid band shift; see below), which has led to the view that the cytochrome c-to-(BChl), reaction is electrogenic and that (BChl), is located 40-45'70of the way into the membrane dielectric from the cytochrome c side (Jackson and Dutton, 1973). 4. Electron transfer is possible in the RC ubiquinone membranes through Q1 and QI1 to at least part of Qp. Single-turnover flashes of the kind shown in Fig. 1 elicit current transients for at least 10 turnovers before diminishing, so that an electron sink must exist. And the integrated current response to steady illumination should estimate the size of the sink. It is typically 10-fold larger than the integrated current for a one-turnover flash. 5. Sources of the steady-state currents (Fig. 1C) remain unclear. One possibility is that Qp recycles the reducing equivalents back to the ferricytochrome c as hydrogens, in effect translocating H + across the membrane in an electrically silent manner. Alternatively, it is possible that the QIIsor Qp7formed by light reacts with molecular oxygen in the medium. In support of this latter possibility, preliminary results show that lightinduced oxygen consumption occurs at a rate commensurate with the steady state current.
-
B. Notes on the Structure of the Reaction Center 1 . OVERALL STRUCTURE
The RC, of molecular weight 90,000-100,000, has three subunits, L, M, and H with molecular weights in the range of 28,000-35,000. L and M when associated together, contain (spectrally intact) the whole BChl-BPhQ1 complement (plus an iron atom). The H subunit appears devoid of redox carriers. Antibody work (Valkirs et af., 1976; G. Feher, personal communication) indicates that L and M are exposed to the aqueous phase on one side of the photosynthetic membrane, while H and M are exposed on the other, so that M must extend across the entire membrane. The iodine
332
P. LESLIE DUTTON et
a/.
labeling experiments of Takemoto and Bachman (personal communication) differ from the antibody work in one respect; they have found that the H subunit, too, extends across the membrane. Photoaffinity labeling experiments (Marinetti et a/., 1979) suggest that QI is bound to M, while cytochrome c may bind to both L and M. We have sought structural information on this hydrophobic protein from low-angle x-ray scattering (RCs oriented in phospholipid membrane multilayers; Pachence et al., 1979). The results, at 10-A resolution in the profile of the membrane, reveal a protein approximately 57A long, which is more than sufficient to span the membrane bilayer. The profile of the protein is that of a “T,” as indicated in the schemes of Fig. 4. More recent neutron diffraction work on deuterated RCs (Pachence et al., 1980) confirms the x-ray results, and further x-ray studies with cytochrome c show it bound to RC at the end with the smaller cross section. 2. DISTANCES BETWEEN REDOX COMPONENTS
The estimates of distance between the redox components of the cytochrome-RC-quinone complex have come either from measurement of magnetic interactions between the electron spins of the radicals on the redox components (Tiede et al., 1978) or from application of electron tunneling theory to kinetic determinations (Hopfield, 1974; Jortner, 1976; Okamura et al., 1979). Thus far, electron paramagnetic resonance (EPR) estimates of the cytochrome c-to-(BChl), distance have been done only on Chromatium vinosum, yielding an edge-to-edge distance of 12- 18 A (Tiede et al., 1978). If a similar distance exists between cytochrome c, and (BChl), of R . sphaeroides, then it would be consistent with previous evidence that this reaction spans a significant portion of the membrane profile (i.e., 1020A), as would be required for the reaction to contribute to a transmembrane charge separation. Similarly the (BChl),-to-Q, distance could be substantial; a lack of detectable EPR interactions between these components leads us to expect that the molecular edge-to-edge distance will be greater than 13 A , again corroborating the notion that these components must lie across a significant part of the membrane in order to contribute to the electrogenic reactions. In contrast, measurable magnetic exchange coupling between BPh and QI or BPh and (BChl), imply a closer proximity. While exact distance determinations are uncertain, because the analysis involves several assumptions, current estimates for attenuation of an orbital overlap (Hopfield, 1974; Jortner, 1976) suggest that exchange coupling of the magnitude obtained between BPh and QI can occur over an edge-to-edge separation of 10-1 1 A (Hopfield, 1974; Jortner, 1976). This coupling between the BPh’and QSFe is comparable to the overlap and separations predicted from kinetic measurements (Hopfield, 1974; Rentzepis, 1978). Considerations of triplet
18. ELECTROGENIC REACTIONS OF
RC AND
Q-b/C, OXIDOREDUCTASE
333
formation by (BChl), would allow somewhat greater separations between (BChl),? and BPh’ (Blankenship et al., 1977; Hoff et al., 1977; Werner et al., 1978). It is finally worth noting that the molecular size of the redox reactants can be significant in reactions which transport electrons over several tens of angstroms. For example, if the (BChl), and the BPh were touching within the reaction center, they could contribute about 24A to the distance traveled by an electron.
C. Free Energies of the Electron Transport Steps in the Cytochrome c,-Reaction Center Complex Figure 5 presents the main redox components, listed in Scheme 1, on a potential scale relative to the standard hydrogen electrode. Blocks drawn about the redox couples indicate that the midpoint potentials (Em)have been determined in vivo with reproducibility; the blocks extend over the Eh
-800
-
-600
-
-400
-
- 200
~
I
AEm mnfributes b
AS
J
7
I
Eh (mV)
0> opemles 0 -cyt c!/h
200
-
oxidorebctase
FIG. 5 . The redox components of the photochemical RC. A discussion of the measurements leading to this figure may be found in Clayton and Sistom (1978).
334
P. LESLIE DUTTON
et a/.
range needed to take the couple from 9-91070 oxidized. The reducing potential indicated for the excited (BChl), * is obtained from the long-wavelength absorption band of the (BChl), in the RC. An accurate determination of the in vivo Em for BPh is still required. The salient results from Fig. 5 are as follows: 1. The (BChl),*-to-Q, span has sufficient energy to drive electron transfer against A$. If, as we have discussed, the (BChl),-BpH-Q, sequence spans 59-75% of the low-dielectric part of the membrane and the RC-Qb/c, oxidoreductase cycle can operate to generate and sustain a stable A$ of up to 400 mV (Jackson and Crofts, 1969; Takamiya et a[., 1979; Bashford et af., 1979b), then more than 240-300 mV will be required to drive the electrogenic reaction between (BChl), and Q,. It appears from Fig. 5 that there is sufficient free energy in the (BChl),*-to-Q, redox span to drive this reaction to near completion, even at the prevailing high A$ values. Detailed discussion of the foregoing is necessarily limited, because there is still much to be established theoretically as well as experimentally. In particular the operating Emof BPh and its location in the membrane profile relative to the (BChl), and Q, in the squence has still to be determined. Without this infor-
R qhaeroides
IA Eh
B
n
0-
mV
200 -
400 -
I t
FIG.6 . The redox components involved, or possibly involved, in ubiquinone-cytochrome b/c2 oxidoreductase. Measurements described in Clayton and Sistrom (1978). (A) includes the species demonstrably involved in electron flow, while (B) shows those components which are present but whose role remains unclear. See description in text.
18. ELECTROGENIC REACTIONS OF
RC
AND
Q-b&
OXIDOREDUCTASE
335
mation we cannot know the contribution made by the (BChl),-BPh and the BPh-Q, steps to the overall electrogenic reaction directed across the membrane. 2. The cytochrome c-to-(BChl), reaction also can drive electron transfer against A$. If the cytochrome c-to-(BChl), electron transfer spans 25-45% of the low-dielectric part of the membrane, then 100-180 mV of free energy will be required to compete against a prevailing A$ of 400 mV. The Em difference between the (BChl), and cytochrome c is only approximately 150 mV, and with this limited free energy available the effects on the redox poise between the two redox couples are clearly visible as the A$ approaches 400 mV (Takamiya and Dutton, 1977). 3. The free energy from the Q,-to-cytochrome c, span is sufficient to drive electrons through the Q-b/c, oxidoreductase. The Emvalues of Q, and cytochrome c, shown in Fig. 5 and again in Fig. 6A encompass the range of Emvalues of the redox components of the Q-b/c2 oxidoreductase (Fig. 6A).
III.
T H E UBIQUIN0 N E- CYTOC H RO M E OXIDOREDUCTASE
b/C,
A. Hierarchy of Redox Components
In sharp contrast to research on the mitochondria1 redox system, work on the bacterial photosynthesis system has been slow to assimilate the notion of a Q-b/c, oxidoreductase. Much progress has been made in determining the kinetics and pathways of electron transport in what we now consider the Q-b/c, oxidoreductase and, accompanying this, progress has been made in revealing the events coupled to electron transport through this membrane-associated redox protein. However, in contrast to the RC, where forward electron transfer sequences appear to be relatively simple, we find more elaborate patterns of electron flow through the Q-b/c2 oxidoreductase; indeed there may be several alternative reactions and sequences in the Q-b/c, oxidoreductase depending on the state of reduction of the redox components at the time of activation. As with the RC, we cannot yet assign roles to all the redox components that appear to be present in the Q-b/c, oxidoreductase. Our exploratory work has revealed three groups of redox components: those that are present and demonstrably involved by kinetic and thermodynamic criteria; those that are clearly involved but can only be detected and characterized indirectly; and those that are clearly present but can neither be included nor eliminated from reactions on the basis of current experimental evidence. Cytochrome c, and cytochrome b,, fall into the first category; the compo-
336
P. LESLIE DUTTON t?f
a/.
nent formerly designated Z and recently identified as a quinone (Em= 150 mV at pH 7, n = 2; see below), which is vital to the operation of the cycle, falls into the second category, and cytochromes b,55and b-w, the Rieske iron-sulfur center, several other iron-sulfur centers, and the ubiquinone “pool” all fall into the third category (e.g., see Clayton and Sistrom, 1978). Recent studies have clarified and simplified this situation.
1. IDENTIFICATION AND ROLEOF UBIQUINONE Qz Extraction and reconstitution experiments (Takamiya et af., 1979) have + 155 mV; see Fig. 6A) as a ubiquinone. A identified component Z (Em= total of 25 f 3 ubiquinone molecules is present per RC; and nearly exhaustive extraction is required to remove Z. More specifically, Takamiya et af. (1979)have found that removal of all but 1 ubiquinone molecule is required to eliminate component Z; but readdition of only 2-3 quinones, at the upper limit, restores compound Z. Kinetic and potentiometric studies on the reduction of flash-oxidized cytochrome c2 have shown the reaction to be first-order in both ferricytochrome c2 and QzH, and have indicated the presence of 0.8f 0.1 moiety of Qz/QzH2per reaction center (Prince et al., 1978). 2. POOLUBIQUINONE IS NOTAN ESSENTIAL PARTICIPANT IN CYCLIC ELECTRON FLOW
The experiment identifying component Z as a tightly bound ubiquinone molecule shows also that extraction of the main quinone pool has little effect on cyclic flow. 3 . RAPIDINTERACTION OF THE RIESKEIRON-SULFUR CENTER WITH CYTOCHROME c2
Recent experiments in collaboration with John Bowyer and Tony Crofts have shown that the Rieske center normally reduces flash-induced ferricytochrome c2 with a submillisecond halftime in a reaction hitherto undetected. This reaction is inhibited by 5-n-undecyl-6-hydroxy-4,7-dioxobenzothiazole (UHDBT), so in its presence the extent of cytochrome oxidation is apparently much greater than in its absence. Indeed it now seems that there is probably only a single cytochrome c2 molecule functionally associated with each RC, although more cytochromes may be trapped inside the chromatophore lumen. A more detailed discussion, which is outside the scope of this article, may be found in Bowyer and Crofts (1978) and Bowyer et af. (1980).The reaction does not seem to be electrogenic.
18. ELECTROGENIC REACTIONS OF
RC
AND Q-blC2 OXIDOREDUCTASE
337
4. IRON-SULFUR CENTERS OF SUCCINIC DEHYDROGENASE ARE NOT IN THE Q-b/c, OXIDOREDUCTASE INVOLVED
The iron-sulfur centers shown in Fig. 6B with Em values of +50 and
- 200 mV, and the high-potential iron-protein center (HiPIP; Em= + 80 mV at pH 7) can be removed by a single alkaline wash of the chromatophore. Succinic dehydrogenase activity is lost concomitantly (Ingledew and Prince, 1977), but cyclic electron flow is unaffected. B. Redox Coupling of the Reaction Center and the Q-blc, Oxidoreductase As far as we know, there are no electrogenic reactions associated with the electron transfer steps that couple the RC-c, complex with the Q-b/c, oxidoreductase. We may therefore expect minimal differences in the Em values of the reacting components. This is borne out, at least at the highpotential coupling, by the close similarity of the Emvalues of cytochrome c2 and the Rieske iron-sulfur center (Fig. 6A). At the low-potential coupling, the situation is less clear. The most plausible redox components involved with the coupling are QI1 and cytochrome b5,,, and these are currently being studied in several laboratories. No consistent picture has yet emerged, but the work has yielded the following observations.
PROPERTIES OF QII 1. THETHERMODYNAMIC Rutherford and Evans (1980), using EPR, have measured the Em values of what seems to be QI1interacting with the RC iron molecule. They obtained an Em for QII/QII'Hof 100 mV (at pH 7), an Emfor Q'H/QH, of 20 mV, and a pK on Qll 'H of pH 9.5, indicating an Emfor QII/QIIT of - 50 mV. 2. PROTON BINDINGAND SPECTRAL OSCILLATIONS
Studies on chromatophores suspended in solutions of pH indicator dyes have shown that one proton (H:) is bound by the chromatophore membrane for each flash-driven turnover of the RC (Petty and Dutton, 1976). Binding occurs with a half-time of about 100 psec (pH 7.0) and has been thought to occur on QII. However, absorption measurements made in the quinone region during trains of flashes (typically delivered at 5- to 10-second intervals) give a rather surprising result: An absorption increase is observed at 450 nm on each odd-numbered flash, and an absorption decrease occurs on each even-numbered flash (see, e.g., Wraight, 1977;
338
P. LESLIE DUTTON ef
a/.
Vermeglio, 1977). By analogy with extracted ubiquinone (Bensasson and Land, 1973), the increased absorption should be due to formation of the ubisemiquinone anion (Qr17),while the absorption decrease on the subsequent flash could result from the disappearance of Qlli into the quinone anion (Q,12-)and/or the protonated quinone (QIIH2),along with a transfer of both reducing equivalents to the pool quinones. Such a model would predict the binding of zero protons on odd-numbered flashes and two protons on even-numbered flashes. Unfortunately, over a wide range of conditions we have been unable to detect any such binary oscillations of proton binding. It seems, therefore, that certain underlying assumptions, particularly that flash-induced proton binding occurs on a quinone molecule, must now be reevaluated. 3. THE REDUCTION OF CYTOCHROME b Under conditions where oscillations are observed at 450 nm, no oscillations in the reactions of cytochrome b can be detected. However, at a high Eh (400 mV) the amount of ferricytochrome b reduced on the first flash decreases, while that reduced on the second flash remains constant. This phenomenon does not continue as a sustained binary oscillation with subsequent flashes. The decrease in cytochrome b reduction on the first turnover is accompanied by an increase in the 450-nm change. The apparent midpoint potential of both these effects (400 mV) is pH-independent between p H 5.5 and 1 1 .
4. IMPAIRMENT OF CYTOCHROME b REDUCTION AT Low AMBIENT POTENTIALS At potentials where the secondary quinone is mostly Qrl'Hprior to flash activation, oxidation of the Q,,H, formed on the first turnover and cytochrome b reduction are inhibited. This is unexpected, since thermodynamically this reaction has a substantial negative AG. It is clear that our understanding of the redox coupling between the RC and the Q-b/c, oxidoreductase rests heavily on the elucidation of which redox species of the secondary ubiquinone are functional in the system. Some of the obvious conflicts could be resolved if the flash-induced proton binding were t o a molecule other than Q1,. Along these lines, Wraight (1979) has suggested that the proton may be bound to a nonchromaphoric acid-base group which undergoes a pK shift in response to the proximity of the QnT. If this were the case, the functional species of the secondary quinone should be the Qll/QIl?couple, which has an Em of -50 mV (pHindependent). This would be capable of almost completely reducing
18. ELECTROGENIC REACTIONS OF
RC
AND Q-b/C, OXIDOREDUCTASE
339
cytochrome b50.But an additional constraint would then be needed to effect attenuation of the first-turnover cytochrome b reduction at high E,,.
C. Electrogenic Events within the Q-blc, Oxidoreductase Evidence that an electrogenic reaction exists within the Q-b/c2 oxidoreductase itself is provided by the so-called phase I11 carotenoid band shift. Native carotenoids in chromatophore membranes display a red-shifted absorption spectrum when an electric field is imposed on the membrane, either via an applied current or via photoactivation (see also the discussions of Junge and of Graeber, this volume). If experiments are carried out at different fixed redox potentials, the flash-induced spectral shift can be split into three components: phase I (Eh 400 mV) is identified with photooxidation of (BChl),, and phase I1 (Eh 250 mV) appears due to cytochrome c oxidation. We have already shown both of these reactions to be electrogenic, by direct measurements on RC bilayer membranes (see above), and have found that the flash-induced currents (voltage-clamped bilayers) are proportional to the magnitude of the carotenoid band shift seen in chromatophores. The phase 111 band shift is seen (in addition to phases I and 11) at redox potentials negative to about 200 mV (pH 7) and can be associated with the oxidation of cytochrome b because of its ready blockage by antimycin (van den Berg et al., 1979; Jackson and Dutton, 1973). Phase I11 occurs with millisecond kinetics closely resembling the reduction of flash-generated ferricytochrome c, (Bashford et al., 1979b), and its formation is dramatically stimulated by prior reduction of Q Z .The magnitude of phase 111 is about equal to the sum of phase I plus phase 11, so that the underlying reaction could represent charge separation across the entire membrane bilayer . Further evidence for the electrogenic character of the Q-b/c, oxidoreductase has also been given by rather complicated effects of imposed potentials on the redox states of cytochromes b and c, (Matsuura and Nishimura, 1978).
-
D. Associated H+ Transport Two protons are incorporated into the chromatophore for each electron transported through the RC Q-b/c, oxidoreductase cycle. One is H:, as discussed above. The agent responsible for the binding of the other, designated HA, is less certain. The binding of HA, unlike that of Hh is inhibited by antimycin, so it clearly must arise from events that also lead to ferricytochrome c, reduction and carotenoid band shift phase 111. The
340
P. LESLIE DUTTON et
a/.
extent and kinetics of HA binding are variable (0.2 msec to several milliseconds) and are dependent, in a way not clearly understood, on the redox state of cytochrome c, (or perhaps the Rieske iron-sulfur center) and Qz (Petty et al., 1979). The pK values apparent for the groups responsible for H:and H; binding also appear dependent on the redox states of the electron transfer components in the system. The pK values apparent for the groups responsible for HTand HA binding also appear dependent on the redox states of the electron transfer components of the system. The highest pK values for HTand H; are 8.5 and 7.5, respectively (Petty et al., 1979), which seem remarkably low for a system geared to H+ translocation.
E. Possible Sources of the Electrogenic Reaction The view held these last few years regarding generation of the carotenoid band shift phase I11 is based on an electron transfer from the inside (the cytochrome c, side) to the outside of the chromatophore membrane. In this respect the view is similar to that of the electrogenic events apparent for the cytochrome c,-RC reactions. However, the organization in the Q-b/c, oxidoreductase needed to promote an electrogenic reaction in this way must be rather different. In the case of the cytochrome c,-RC reaction, charge separation is rooted in the light reaction (probably near the middle of the membrane profile) which starts the process forming the positively charged strong oxidant (BChlX’ and an adjacent negatively charged strong reductant BPhl. The negative charge is pulled into the aqueous phase on one side of the membrane, driven by Urn between BPh and Q,-QIland by the free energy associated with the binding of H&while the positive charge is pulled between into contact with the opposite aqueous phase, driven by the Urn cytochrome c2and (BChl), (Fig. 5). This pattern of charge separation can be readily appreciated by looking at Scheme 1. In contrast, the Q-b/c, oxidoreductase has to work with an oxidant on one side of the membrane (cytochrome c,) and a reductant near the other (QII),which will combine in the ensuing steps within the Q-b/c, oxidoreductase. If the reducing equivalent, for example, as an “H” moiety, simply returned to the ferricytochrome c, as diagramed in Fig. 7A, this would be the simplest cycle that could promote the generation of A$ and a pH gradient in a simple chemiosmotic manner. This model, however, would provide neither an electrogenic step(s) nor a second proton binding (Hi) associated with the Q-b/c, oxidoreductase. To accommodate these features, several working schemes, all based on the tenets of Mitchell’s chemiosmotic hypothesis, have been engineered and tested, mainly by ourselves and by Crofts’ laboratory. Two that have been most useful are show in Fig. 7B and C. As we have already discussed, the position of the Rieske iron-sulfur
18. ELECTROGENIC REACTIONS
OF
RC
341
AND Q-b/C, OXIDOREDUCTASE
B
A
C
-
HI
-r A0
HA
- -c A
AQ
--%-
HA
-F
HA
FIG.7 . Simple schematic “circuit diagrams” of possible reaction pathways for the RC ubiquinone-cytochrome b/c2 oxidoreductase. The membrane structures shown are as follows: The “T” shape represents the RC complex; the circle represents cytochrome c, bound to the RC at the inner surface of the chromatophore membrane; and the rectangle represents the Q-b/c2 oxidoreductase, which would contain, in functional order from right to left (outside to inside edges of the membrane), the “chain” elements for Q,to the Rieske iron-sulfur center in Fig. 6A.
center at the high-potential end interacting with cytochrome c,, and the position of cytochrome b at the low-potential end interacting with Ql,, can be understood in terms of these models. In addition Qz, playing an experimentally established central role in the operation of Q-b/c, oxidoreductase can also be understood (see Clayton and Sistrom, 1978). However, we are increasingly aware that such models, while accounting for some experimental observations, are unsatisfactory in accounting for others. Many of the discrepancies are kinetic and involve the protons (see Petty et al., 1979); and they have led us to an additional class of possibilities, which is diagramed in Fig. 7D. This proposes proton translocation by a mechanism entirely different from that of the basic chemiosmotic
P. LESLIE DUTTON ef 6’1.
342
schemes. The model suggests a simple cycle, in Fig. 7A, but assigns the electrogenic step (associated with the Q-b/c, oxidoreductase) to the translocation of H i as a charged cation rather than as a neutral reducing equivalent (H’). While the overall economics for this model are the same as for the others, the mechanism is distinctly different, involving what may be called “a proton pump” (Wikstrom and Krab, 1978). We should also add that, while protons are shown as if transported across the entire membrane (entering on one side and leaving on the other), we are not yet sure. Information of the kind now available for proton binding (single-turnover flash, rapid analysis, calibrated H+/e- ratios) is still badly needed for proton release. Clearly, while much has been done in characterizing the components functional in the Q-b/c, oxidoreductase and in describing the overall energetics of the system, much has yet to be learned at the mechanistic level. ACKNOWLEDGMENTS This research was supported by research grants GM-27309 (P.L.D.) and GM 12202 (P.M.) from the National Institutes of Health, and by research grant PCM 7909042 (P.L.D.) from the National Science Foundation. REFERENCES Barsky, E. L., Darcshazy, Z., Drachev, L. A,, Il’ina, M. D., Kondrashin, A. A., Samuilov, V. D., and Skulachev,,V. P . (1976). J. Biol. Chem. 251, 7066-7071. Bashford, C. L., Chance, B., and Prince, R. C. (1979a). Biochim. Biophys. Acta 545,46-57. Bashford, C . L., Prince, R. C., Takamiya, K., and Dutton, P. L., (1979b). Biochim. Biophys. Acta 545, 223-235. Bensasson, R., and Land, E. J . (1973). Biochim. Biophys. Acta 325, 175-181. Blankenship, R. E., and Parson, W. W. (1978). Annu. Rev. Biochem. 41, 635-654. Blankenship, R. E., Schaafsma, J. J., and Parson, W. W. (1977). Biochim. Biophys. Acta 461, 297-305. Bowyer, J. R., and Crofts, A. F. (1978). In “Frontiers of Biological Energetics” (P.L. Dutton, J. S. Leigh, and A. Scarpa, eds.), Vol. I, pp. 326-333. Academic Press, New York. Bowyer, J . R., Dutton, P. L., Prince, R. C., and Crofts, A. R. (1980). Biochim. Biophys. Acta 592, 445-460. Chance, B., and Baltscheffsky, M. (1975). In “Biomembranes” (H. Eisenberg, E. Katchalski-Kalzir, and L. A., Mason, eds.), Vol. 7, pp. 33-50. Plenum, New York. Chance, B., Crofts, A. R., Nishimura, M., and Price, B. (1970). Eur. J . Biochem. 13, 364-374. Clayton, R. K., and Sistrom, W. R. (1978). “The Photosynthetic Bacteria.” Plenum, New York. Crofts, A. R., Jackson, J. B., Evans, E. H., and Cogdell, R. J. (1972). Proc. Int. Congr., 2nd, Phytosynthesis pp. 873-902. Drachev, L. A., Frolov, V. N., Kavlen, A. D., Kondrashin, A. A., Samuilov, V. D., Semenov, A. Y., and Skulachev, V . P. (1976). Biochim. Biophys. Acta 440, 637-660.
18. ELECTROGENIC REACTIONS OF
RC
AND Q-b/C, OXIDOREDUCTASE
343
Dutton, P. L. (1971). Biochim. Biophys. Acta 226, 63-80. Dutton, P. L., Prince, R. C., and Tiede, D. M. (1979). Photochem. Phoiobiol. 28, 939-949. Halsey, Y. D., and Parson, W. W. (1974). Biochim. Biophys. Acta 347, 404-416. Hoff, A. J., Rademaker, H., van Grondelle, R., and Duysens, J . N. M. (1977). Biochim. Biophys. Acta 460, 547-554. Hopfield, J. J. (1974). Proc. Nail.Acad. Sci. U.S.A. 71, 3640-3644. Ingledew, W. J., and Prince, R. C. (1977). Arch. Biochem. Biophys. 178, 303-307. Jackson, J . B., and Crofts, A. R. (1969). FEBS Lett. 4, 185-189. Jackson, J. B., and Dutton, P. L. (1973). Biochim. Biophys. Acta 325, 102-113. Jortner, J. (1976). J. Chem. Phys. 64, 4860-4867. Kendall-Tobias, M., and Crofts, A. R. (1979). Biophys. J. 25, 54a. Marinetti, T., Okamura, M. Y., and Feher, G. (1979). Biophys. J. 25, 204a. Matsuura, K., and Nishimura, M. (1978). J. Biochem. 84, 539-546. Mueller, P., and Rudin, D. 0. (1969). Curr. Top. Bioenerg. 3, 157-249. Okamura, M. Y., Isaacson, R. A,, and Feher, G. (1979). Biochim. Biophys. Acta 546, 394-417. Olson, J. M., and Thornber, J. P. (1979). In “Membrane Proteins in Energy Transduction” (R. A. Capaldi, ed.). Dekker, New York. Pachence, J. M., Dutton, P. L., and Blasie, J. K. (1979). Biochim. Biophys. Acia 548, 348-373. Pachence, J . M., Dutton, P. L., and Blasie, J. K. (1980). Biochim. Biophys. Acta 635, 267-283. Packham, N. K., Berriman, J. A., and Jackson, J. G. (1978). FEBS Leii. 89, 205-210. Packham, N. K., Packham, C . , Mueller, P., Tiede, D. M., and Dutton, P. L. (1980). FEBS Lett., 110, 101-106. Petty, K. M., and Dutton, P. L. (1976). Arch. Biochem. Biophys. 172, 335-345. Petty, K. M., Jackson, J. B., and Dutton, P . L. (1979). Biochim. Biophys. Acta 546, 17-42. Prince, R. C., Baccarini-Melandri, A., Hauska, G. A., Melandri, B. A., and Crofts, A. R. (1975). Biochim. Biophys. Acta 387, 212-227. Prince, R. C., Bashford, C. L., Takamiya, K., van den Berg, W. H., and Dutton, P. L. (1978). J. Biol. Chem. 253, 4137-4142. Reed, D. W., and Clayton, R. K. (1968). Biochem. Biophys. Res. Commun. 30, 471-475. Rentzepis, P. M. (1978). I n “Frontiers of Biological Energetics” (P. L. Dutton, J. S. Leigh, and A. Scarpa, eds.), pp. 19-29. Academic Press, New York. Rutherford, A. W., and Evans, M. C. W. (1980). FEBS Lett. 110, 257-261. Schonfeld, M., Montal, M., and Feher, G. (1979). Biophys. J. 25, 203a. Takamiya, K . , and Dutton, P. L. (1977). FEBS Leii. 80, 279-294. Takamiya, K., Prince, R. C . , and Dutton, P . L. (1979). J. Biol. Chem. 254, 11307-11311. Tiede, D. M., Leigh, J. S., and Dutton, P. L. (1978). I n “Frontiers of Biological Energetics” (P. L. Dutton, J. S. Leigh, and A. Scarpa, eds.), pp. 45-53. Academic Press, New York. Valkirs, G., Rosen, D., Tokuyasu, K. T., and Feher, G. (1976). Biophys. J . 16, 223a. van den Berg, W. H . , Prince, R. C., Bashford, C. L., Takamiya, K., Bonner, W. D., and Dutton, P. L. (1979). J. Biol. Chem. 254, 8594-8604. Vermeglio, A. (1977). Biochim. Biophys. Acia 459, 5 16-524. Werner, H. J., Schulten, K., and Weller, A. (1978). Biochim. Biophys. Acta 502, 255-268. Wikstrom, M. K. F., and Krab, K. (1978). In “Frontiers of Biological Energetics” (P. L. Dutton, J. S. Leigh, and A. Scarpa, eds.), pp. 351-358. Academic Press, New York. Wraight, C. A. (1977). Biochim. Biophys. Acta 459, 525-531. Wraight, C. A. (1979). Biochim. Biophys. Acia 548, 309-327. Wraight, C. A., Cogdell, R. J., and Clayton, R. K. (1975). Biochim. Biophys. Acta 396, 242-249.
This Page Intentionally Left Blank
CURRENT TOPICS IN MEMBRANES AND TRANSPORT, VOLUME 16
Chapter 19
Proton- Membrane Interactions in Chloroplast Bioenergetics R . A . DILLEY, L . J. PROCHASKA,’ G . M . BAKER, N . E. TANDY, AND P. A . MILLNER Department of Biological Sciences Purdue University Biochemistry Program Purdue University West Lafayette, Indiana
Introduction ........................................................................................ A. Ion Movements in Chloroplasts ...................... ................ B. On the Electrogenic Aspects of Chloroplast Ion Fluxes ............................ C. Chloroplast Redox React tructure ..... 11. Methods and Rationale ....... ................ 111. Results and Discussion ........................................................................... A. Decrease in Derivatization of Thylakoid Membrane Proteins by ................. ................. Acetic Anhydride Caused by Light B. The Correlation of Uncoupler-Enhanced Acetic Anhydride Derivation with Inhibition of Water Oxidation ..................................................... C . Membrane Proteins Showing Differential Acetic Anhydride Reactivity in Light and Dar ............. ....... IV. Concluding Remarks ............. ....... References .......................................................................................... I.
1.
345 345 346 341 351 352 352 351 360 363 361
INTRODUCTION
A. Ion Movements in Chloroplasts One of the first experiments showing energy-linked ion movements (protons in this case) in chloroplasts was suggested and performed in 1963 by Geoffrey Hind, who was at that time a postdoctoral fellow in Andre JagenPresent address: Biological Chemistry Program, School of Medicine, Wright State University, Dayton, Ohio.
345
Copyright @ 1982 by Academic Press, Inc All rights of reproduction In any form reserved ISBN 0- 12-IS33 16-6
346
R. A. DILLEY
et a/.
dorf’s laboratory (Jagendorf and Hind, 1963). Hind was inspired by Mitchell’s (1961) brilliant insights into the possibilities of redox-linked proton transport in energy-transducing membrane systems (see also Williams, 1961, 1975). Hind and Jagendorf were looking for the physicochemical basis of the phenomena of light-induced light-scattering changes which were then being studied in several laboratories (Itoh et al., 1963; Packer et al., 1963; Dilley and Vernon, 1964). Jagendorf and Hind (1963) found that the kinetics of the light-scattering changes were quite similar to those of the H ion uptake. We showed that the energy-linked light-scattering changes were due to a H + influx-K+ + Mg2+efflux accompanied by water flow, all of which resulted in volume changes in the chloroplast thylakoid membrane (Dilley and Vernon, 1965; Dilley et al., 1967). Packer and colleagues made an extensive study of the light-induced ion-dependent shrinking and swelling reactions of chloroplasts (Deamer et al., 1967; Crofts et al., 1967). This work and the experiments of Dilley and Rothstein (1967) indicated that chloroplast membranes had properties resembling those of an ionexchange resin, with a pK near 4.7. +
6. On the Electrogenic Aspects of Chloroplast Ion Fluxes The early work on H cation exchange did not strongly support an electrogenic aspect of the ion fluxes, although it seemed clear that proton uptake was the primary ion movement (Dilley and Vernon, 1964; Dilley, 1971). The actual proton release mechanisms-water oxidation, Qox+ H,O 2H + iO,+ Qred,and plastohydroquinone oxidation, Cyt f o x + PQH2PQox+ 2H + Cyt f,,,-could have significant electrogenic character provided the structure of the redox chain allows separation of released protons from the electron acceptors Q and cytochrome f. In fact, the structure of the redox chain vis-a-vis this problem remains unsolved, but it is likely that the electrons and protons do become spatially separated, hence there is probably a strong, at least local, electrogenic character to the protolytic reactions. If we define an electrogenic flux as one that carries a net current in one direction and produces a transmembrane electrical potential A+, then the chloroplast membrane is not electrogenic in the steady state. The chloroplast thylakoid (inner) membrane maintains a steady-state transmembrane A+ of only about 10 mV, as indicated by measurements of the distribution of K + and C1- (Schroder et al., 1972; Rottenberg et al., 1972). Yet the well-known 515-nm absorption change elucidated by Witt (1971) and his co-workers (Schliephake et al., 1968) provides excellent experimental support for some sort of membrane electrical potential +
-
+
+
347
19. PROTON-MEMBRANE INTERACTIONS
generated by the electron transfer reactions. According to these measurements, a transmembrane A$ of 50-100 mV is maintained in the steady state. However, there is no direct proof that the charge separation responsible for the 515-nm signal is transmembrane, bulk phase to bulk phase, rather than within the membrane. This article will dwell at length on the possibility that proton release in the redox reactions may be an intramembrane phenomenon, hence that the protolytic reactions may have a significant local electrogenic character. The term “intramembrane” as used here is not meant to imply insertion of a proton into the lipid phase of the lipid bilayer, an energetically unfavorable event. One possible view is that there may be domains of polar groups associated with membrane proteins (and perhaps the SO, group of the diglycerodigalactosyl sulfolipid), which extend into or through the 80-100 membrane. Exact knowledge about such aspects of membrane structure is yet t o come. In any event, if it is permissible to define the proton release in the protolytic reactions as electrogenic, then it may well be that chloroplasts have a strong electrogenic (but local) “pump.” Before going into this, a further discussion of the known biochemistry of chloroplast redox and ion transport reactions is in order.
A
C. Chloroplast Redox Reactions, Ion Transport, and Chloroplast Structure Generally, it is accepted that the photosynthetic electron transport pathway consists of two photosystems in series: photosystem I1 (PS 11), which oxidizes water, and photosystem I (PS I), which takes electrons from PS I1 and provides the energy necessary to reduce ferridoxin (Fd) and NADP’ (Fig. 1). Proton accumulation linked to the redox steps is believed to occur via the protolytic reactions of water oxidation and plastohydroquinone oxidation. The two photosystems can be physically separated (by detergents) and recombined (Arntzen et al., 1972), to give H,O-NADP+ redox activity. PS I seems to be associated with the external part of the membrane, while PS I1 is largely but not completely, buried on the inner part of the membrane (Arntzen et al., 1969). Figure 2A shows a stained, thin-sectioned chloroplast with the membranes seen in cross section. Figure 2B shows the freeze fracture-revealed structure of thylakoid membranes. The large-particle ( 180 A) and small-particle ( 110 A ) arrays revealed are part of the interior structure of the membrane and have been shown to be associated with the two photosystems. The larger particle is associated with PS I1 activity, and the smaller with P S I activity. The particles are
-
-
348
Elecmn donm to P
R. A. DILLEY
et a/.
Membrane
INNER AQUEOUS PHASE
FIG. 1. A model of the Z scheme for electron transport in chloroplasts. The protolytic reactions, water oxidation, and the PQH, oxidation are shown as resulting in direct deposition of protons into the inner aqueous space. When water oxidation is inhibited by treatment which removes manganese, an alternative donor can be substituted, such as iodide (I-, an electrononly donor) or DPC (an electron plus a proton donor).
probably multiprotein complexes comprising the functional entities of each photosystem, as shown in Fig. 2C, a model developed from combined biochemical and ultrastructural studies (Arntzen et al., 1969). In this article we shall discuss biochemical evidence for an intramembrane domain associated with proton processing specific for protons released in the PS I1 water oxidation mechanism. The point to be made here is that there is already an experimentally firm structural basis on which to construct a model depicting site-specific energy transduction events. The chemiosmotic hypothesis espoused by Mitchell says that the two protolytic reactions directly deposit H + ions into the inner aqueous space (Fig. l), leading to a transmembrane proton motive force Ap; APE AjiH+/F= (RTIQ2.3 log ( [H+]in/[E+]o,,) + AlL,, = ZAPK-~+ A$i-o, where R , T, and F have their usual meanings. The magnitude of the A p has been measured to be near 180 mV in chloroplasts, comprised mostly of the ApH in the steady state (Jagendorf, 1975). The downhill flux of protons outward through the CF,-CF, coupling complex is postulated as the energy source for ATP formation, and considerable supportive evidence has been forthcoming for A p as the driving force for ATP synthesis (for a review, see Jagendorf, 1975). While the Mitchell view has strenuously rejected any idea that intramembrane proton processing occurs, Williams (1975) has developed a concept in which some sort of a local proton current within the membrane is the primary link between the redox protolytic reactions and
FIG.2. (A) An electron micrograph of a chloroplast from lettuce, prepared by C. J. ArntZen using standard positive-staining thin-section techniques. GM, Stacked grana thylakoid membranes; SM, stroma membranes; S, stroma phase; OM, outer chloroplast envelope; CW, cell wall. x-26,000. (B) A freeze-fracture replica of a maize mesophyll chloroplast. GM and SM are grana and stroma membranes, respectively. The particles revealed are on the interior of the membranes as discussed in the text (Arntzen et a/., 1969). ~-39,OOo. (C) A model showing the relationship of the two freeze-fracture revealed particles, photosystems I and 11, and thylakoid membrane surfaces (Arntzen et a/., 1969).
350
R. A. DILLEY
FIG.2 B . See page 349 for legend.
et a/.
19. PROTON-MEMBRANE INTERACTIONS
351
FIG.2C. See page 349 for legend.
energy-coupling reactions (ATP synthesis). Experimental results obtained with various energy-transducing membrane systems have been interpreted as evidence for a transmembrane gradient as the driving force for energy coupling (Leiser and Gromet-Elhanan, 1977; Portis and McCarty, 1974), while other results have been more consistent with the concept of local proton gradients as the driving force (Melandri el al., 1978; Van Dam and Westerhoff, 1977; Ort et af., 1976). This article will review work in our laboratory concerning possible intrarnembrane proton processing.
11.
METHODS AND RATIONALE
The approach used for these studies is the detection of membrane protein conformational changes and/or changes in the reactivity of amino acid functional groups with chemical modification reagents. Protein biochemists have used this kind of assay for detecting changes in protein “conformation” as discussed by Timasheff and Gorbunoff (1967). The usual experimental protocol is to add, say, 3-5 rnM [3H]acetic anhydride to a suspension of chloroplasts, either in the dark or the light, wait for 30 seconds, and then add an excess of quenching reagent, such as 50 m M
352
R. A. DILLEY
et a / .
N-glycylglycine. After the removal of excess quenched reagent by repeated centrifugation, and after decolorizing with H,O,, the amount of radioactivity incorporated is determined by scintillation counting. In some cases incorporation of label into membrane proteins is determined by sodium dodecyl sulfate-polyacrylamide (SDS-PAGE) electrophoresis. In this article, the only chemical modification reagent mentioned in detail is acetic anhydride. However, many of the important aspects of the results have been duplicated using either diazobenzene sulfonate or iodoacetic acid (Dilley and Prochaska, 1978). Under the conditions (pH 8.6) used here, only the a-NH, (the N-terminus) and the E-NH, (lysine) groups are thought to form stable, covalent adducts with acetic anhydride (Means and Feeney, 1971). Iodoacetate reacts mostly with sulfhydryl groups, and diazobenzene sulfonate reacts with histidine, tyrosine, lysine, and sulfhydryl groups. It is important to note that only the uncharged form ( - NH,) of an amine reacts with acetic anhydride. Protonation to the NHf form involves the unshared electron pair of the nitrogen atom, removing the possibility of nucleophilic attack on the carbonyl carbon of the anhydride.
111.
RESULTS AND DISCUSSION
This section will present background data leading to the hypothesis on intramembrane site-specific proton processing as deduced from use of the chemical modification assay. Each experiment will be described, with some discussion, followed by a general discussion of the hypothesis and its implications for bioenergetics. A. Decrease in Derivatization of Thylakoid Membrane Proteins by Acetic Anhydride Caused by Light 1.
PROTON
RELEASE BY
THE
Ps 11 WATER OXIDATION REACTION
Proton release by the PS II water oxidation reaction or by an alternative proton donor to PS II is required to observe the labeling difference. Chloroplasts having water oxidation activity and showing light-dependent decreases in anhydride derivatization (control treatment in Table I) can be depleted of water oxidation by NH,OH treatment (Izawa and Ort, 1974). Electron transport supported by an alternative PS I1 donor such as I-, an electron-only donor, does not elicit any change in label incorporation compared to that obtained in the dark. The addition of an electron-protondonating PS I1 donor, diphenylcarbazide (DPC), restores the light-
353
19. PROTON-M EMBRANE INTERACTIONS
TABLE I INCORPORATION OF ACETICANHYDRIDE INTO HYDROXYLAMINE-TREATED CHLOROPLASTS
Treatment Experiment A' Control, high light Control, low light Control, dark NH20H-treated, I-, light NH20H-treated, I-, dark Experiment Bb Control, light Control, dark NH20H-treated, DPC, light NH20H-treated, DPC, dark
Incorporation (nmoles acetate/ mg protein)
3.31 f 0.06 5.53 f 0.29 6.07 f 0.26 6.56 f 0.07 6.48 f 0.37 9.8 f 0.6 12.9 f 0.8 10.6 f 1.4 14.7 f 0.6
Electron transport rate (pEq electrons/ mg chlorophyll per hour) 700 80 110
276 293
(I Hydroxylamine-treated and control chloroplasts were modified with 1 mM acetic anhydride in 50 mMKCI, 50 mM HEPPS-NaOH, pH 8.6, 2 mMMgCI2, 0.5 mMmethyl viologen, 0.5 mM NaN,, I 5 pkf nigericin, and 5 pkf valinomycin. KI (40 mM) was present during the modification of the hydroxylamine-treated chloroplasts. Control chloroplasts were modified in the presence of 40 mM KI under both light and dark conditions, and no difference was observed in the amounts of incorporation for the two conditions. The rate of electron transport was measured prior to the chemical modification. All other conditions were as described in Prochaska and Dilley (1978a). Chloroplasts (0.3 mg chloroplyll/ml) were modified with 1 mM acetic anhydride (specific activity, 25 mCi/mmole) at 20°C in 50 mM KCI, 50 mM HEPPS-NaOH, pH 8.6, 2 mM MgCI,,O.5 a m e t h y l viologen, 0.5 mMNaN3, 15 pkfnigericin, and 5 pkfvalinomycin. DPC (1 mM) was present during the modification of the hydroxylamine-treated chloroplasts. The rate of electron transport in control chloroplasts was lowered by decreasing the light intensity in order to approximate that of the DPC-restored electron transport rate in the hydroxylamine-treated chloroplasts. Hydroxylamine-treated chloroplasts exhibited about 3% of the control rate of water oxidation.
dependent decrease in acetic anhydride derivatization (Prochaska and Dilley, 1978a; cf. Giaquinta et al., 1975, p. 4395, for data and a discussion of the PS I1 proton-dependent changes in diazobenzene sulfonate incorporation). The significant difference between the two donors is the lack of H + release upon oxidation of the I-; both donor systems send electrons through the PS I1 reaction center and on to PS I. One interpretation of these data is that proton release, upon water or DPC oxidation by PS 11, causes protonation of protein -NH2 groups and subsequently less reactivity with the anhydride reagent. An alternative explanation might be that the change in reactivity is due not to proton involvement but to some other aspect of the redox reactions, when water or DPC is the donor, giving a
R. A. DILLEY et
354
a/.
conformational change that blocks reaction of the target groups with the anhydride. The latter interpretation is more difficult to defend, partly because acetic anhydride is very soluble in organic solvents and is expected to permeate the membrane readily. We favor the first interpretation.
I PROTOLYTIC REACTIONS 2. PHOTOSYSTEM Photosystem I protolytic reactions do not give the decreased anhydride derivatization provided dichlorophenyldimethylurea (DCMU) blocks PS 11 function. With I- as the PS I1 donor, the plastoquinone step translocates protons and potentiates ATP formation with an overall ATP/2e ratio near 0.5, compared to an ATP/2e ratio in the range 1.0-1.2 with the normal water-oxidizing system intact or when DPC is the PS I1 donor (Izawa and Ort, 1974). In spite of the active endogenous plastoquinone protolytic reaction in the I--MV (methyl viologen) system, the Table I data show that there is no effect on anhydride labeling with I- as the donor, compared to the dark condition. Table I1 (experiment A) shows that PS I cyclic electron transfer and the associated protolytic reaction do not decrease the anhydride reaction either, provided DCMU blocks PS I1 (Table 11, experi-
TABLE I1 PHOTOSYSTEM I ELECTRON TRANSPORT EFFECTSON ACETICANHYDRIDE INCORPORATION INTO CHLOROPLASTS
Treatment Experiment A' Light, MV Light, PMS Light, PMS, DCMU Dark, MV Experiment Bb Light, MV Light, MV, DCMU, DCIP-Asc Dark, MV
Incorporation (nmoles acetatelmg protein)
6.36 f 0.20 6.25 f 0.70 8.48 f 0.96 8.19 f 0.13
10.8 16.5 17.1
f
f f
0.8 0.9 0.8
' Chloroplasts were modified in 50 mMKC1, 2 mMMgC12, 50 M H E P P S - N a O H , p H 8.6, 5 pM valinomycin, 15 pM nigericin, and 30 pM PMS (f12.5 pM DCMU). Methyl viologen (MV, 0.5 m M ) and 0.5 mM NaN, were substituted for PMS in some experiments. All other conditions were as described in Prochaska and Dilley, 1978a. Chloroplasts were modified in a medium as described in experiment A (MV) above, except that 12.5 pM DCMU, 5 pMsodium ascorbate (ASC), and 50 p M D C I P were added where indicated. The rates of electron transport for both H 2 0 - M V and DCIP-MV were 630 pEq electrons/mg chlorophyll per hour. The DCIP-Asc -MV electron transport rate was adjusted to an equal value by lowering the light intensity (Prochaska and Dilley, 1978a).
19. PROTON-MEMBRANE INTERACTIONS
355
ment A, line 3). When DCMU is omitted and PS I1 water oxidation is permitted to occur in addition to the phenazine methosulfate (PMS) cyclic reaction, the decrease in label incorporation (line 2) does occur. The noncyclic PS I partial redox reaction, ascorbate (Asc), dichlorophenolindophenol (DC1P)-MV+DCMU, does not support the decrease in label incorporation either (Table 11, experiment B). Both these PS I reactions result in proton release upon oxidation of the reduced redox donor, but plastoquinone redox reactions are not involved (Trebst, 1974). We interpret the data of Tables I and I1 as indicating that the protolytic reaction of PS I1 is the critical factor, a point which will be discussed further below. From this point of view, the site of action of the PS 11-linked protons-giving the anhydride labeling change-cannot be in the inner aqueous space, because both PS 11- and PS I-linked proton accumulation acidify the inner aqueous space (Kraayenhof et al., 1972; Rottenberg ef al., 1972; Portis and McCarty, 1974). The working hypothesis we present, then, is that water (or DPC) oxidation initially deposits protons into a domain or space that is not in rapid equilibrium with the inner bulk aqueous phase; i.e., a local pH effect seems to be indicated. Although so far the data do not rigorously exclude other interpretations, the model put forth seems the most likely.
3. UNCOUPLER EFFECTSON ACETICANHYDRIDE DERIVATION Additional evidence, favoring the interpretation that local pH determines the reactivity of thylakoid components with acetic anhydride, comes from the observation that uncouplers are necessary to sensitize the membranes to the maximum level of labeling in the dark. Figure 3 shows this effect, as well as the enhanced inhibition of water oxidation, as uncouplers are added. Three quite different uncouplers give the same effect, carbonylcyanide p-trifluoromethoxyphenylhydrazone (FCCP), gramicidin, and nigericin, so it is not a peculiar effect unique to one uncoupler. A simple interpretation of the result is that uncouplers cause the local pH, in a region where the anhydride reacts, to equilibrate with the external pH (8.6) and promote acetylation of the -NH, groups postulated to have formed by deprotonation of the -NH; groups normally present. The region postulated must maintain sufficient acidity in the dark (resting state) to protonate the putative - NH, groups. By themselves these data do not require a special domain apart from the inner aqueous space. However, the lack of effect of PS I-linked protolytic activity in decreasing anhydride derivation strongly implies that the pH in a space available only to PS 11-linked protolytic events is the critical factor. Protolytic activity in PS I1 reverses the uncoupler-induced extra incor-
356
R. A. DILLEY et
a/.
090 c
‘I 0
.-c .n .-
0) +
60
850
c
81.0
0 c
0
c
q 8
0
E
770
.0;5
.d5
.Ib
I / / / /
I
1
.50
730
[Gramicidin] pM FIG. 3. Uncoupler dependence of the inhibition of water oxidation by acetic anhydride, compared with uncoupler dependence of [3H]acetate incorporation from anhydride. Spinach chloroplasts at 20 pg Chl (chlorophyll) ml-’ were treated with 3 mM [,H]acetic anhydride, in the dark, for 45 seconds at the indicated gramicidin concentrations, followed by the addition of N-glycylglycine to quench the untreated reagent (Baker et al., 1981). Water oxidation was subsequently measured in the H,O-MV Hill reaction.
poration of acetic anhydride, and Table I11 indicates that very potent uncoupling conditions (15 pA4 nigericin, 5 pA4 valinomycin) do not diminish this PS 11-dependent decrease in acetic anhydride reactivity. The lack of uncoupler effect can be explained if the uncouplers do not diffuse in and out of the putative restricted domain rapidly enough to overcome the acidification reaction associated with water oxidation. Though this is a difficult argument, it is supported by observations on the influence of uncouplers on acetic anhydride inhibition of water oxidation (see below). TABLE I11 EFFECTS OF UNCOUPLERS ON THE INCORPORATION OF ACETICANHYDRIDE INTO CHLOROPLASTS
Treatment’
Incorporation nmoles a c e t a t e h g protein
Light Light, valinomycin, nigericin Dark, valinomycin, nigericin
8.14 f 0.26 8.50 f 0.57 13.8 f 1.0
‘Chloroplasts were modified for 15 seconds with 1 mM [3H]acetic anhydride in 50 mM KCI, 50 mM HEPPS-NaOH, pH 8.6, 2 mM MgCI,, 0.5 mM NaN,, and 0.5 mM methyl viologen. Valinomycin (5 p M ) and 15 pMnigericin were presented for uncoupler experiments. All other conditions were as described in Prochaska and Dilley (1978a).
19. PROTON-MEMBRANE INTERACTIONS
357
The ‘data are consistent with the view that-if the effects are due to proton release-it is not the inner aqueous phase from which the effects emanate, since uncouplers rapidly equilibrate the pH between the bulk inner phase and the external solution (Gaensslen and McCarty, 1971; Jagendorf, 1975). The experiments of Table I and I1 were done with uncouplers present in all the treatments. This condition was chosen so as to abolish the known, uncoupler-sensitive, conformational changes in the coupling factor (CF,) protein (Jagendorf, 1975). B. The Correlation of Uncoupier-Enhanced Acetic Anhydride Derivation with inhibition of Water Oxidation It is often found with simple enzyme systems that covalent derivatization of amino acid functional groups leads to the inhibition of enzyme action (Means and Feeney, 1971; Horiike and McCormack, 1979). When chloroplasts are reacted with either acetic anhydride or diazobenzene sulfonate (Baker et al., 1981; Giaquinta et al., 1975), there is progressively more inhibition of electron transport with increasing levels of derivation (Fig. 3). In the case of the anhydride (Fig. 3), both inhibition of electron transfer in the dark and the correlated incorporation of label require the action of an uncoupler. We interpret this to mean that uncouplers cause proton leakage out of the domain containing the - NH,+ groups, thus converting them to the sensitive - NHz form. This dark state inhibition of water oxidation by acetic anhydride can be completely reversed by maintaining electron transfer activity (H,O-MV) before and during exposure to the anhydride (Fig. 4). Such a result is consistent with the notion that the protolytic reactions “reacidify” the space containing the anhydride-reactive - NH, groups, converting them back to the unreactive -NH,+ form (Baker et al., 1981). It has already been mentioned that the light-dependent decrease in anhydride incorporation is specifically dependent on proton release from PS I1 oxidation. Experiments similarly showed that inhibition of water oxidation by acetic anhydride was protected against only by the PS I1 protolytic reaction (Table IV). Chloroplasts were exposed to acetic anhydride when electron transfer either through PS I1 plus PS I or only through PS I was occurring,z Operation of PS I1 alone was arranged with dimethylbenzoquinone as the electron acceptor and 2,5-dibromo-3-methyl-6-isopropylbenzoquinone (DBMIB) present to block electron flow to PS I. Operation of PS I alone was arranged with pyocyanine as the electron carrier, dithiothreitol as a reductant, and DCMU present to inhibit PS 11. Operation of both photosystems was allowed by simply using MV as the terminal electron acceptor.
R. A. DILLEY et a/.
358
30 sec
1 1193)
;
t' Ac,O
h
t
on
on
Ac,O
NGG
off
on
on
FIG. 4. Protection against acetic anhydride (Ac,O) inhibition of water oxidation by lightinduced electron flow. (A) Showing dark inhibition, compare to (B).Ac,O at 3.5 mM was added to a chloroplast suspension (20 pg chl ml-') in the dark for a 45-second treatment, followed immediately by addition of 50 m M N-glycylglycine (NGG) to quench the unreacted reagent. After a 2-minute wait, the rate of electron flow (H,O-MV) was measured, the rate being indicated by the number in parentheses. (B) Light protection as in (A), but the light was turned on before Ac,O was added. (C and D) Controls showing that N-glycylglycine is an effective quencher of the Ac,O whether added in dark (C) or light (D) conditions prior to the Ac~O.
359
19. PROTON-MEMBRANE INTERACTIONS
TABLE IV OF FAILURE
Ps 1 TO PREVENT INHIBITION OF OXIDATION BY ACETICANHYDRIDE Hill reaction (H20-MV) activity remaining after Ac20 treatment (pEq electrons/hour per mg chlorophyll)
Treatment conditionsa 1. Dark, DCMU 2. Light, PS I only, DCMU pyocyanine, reductant 3. Light, PS I + 11 pyocyanine, reductant
Acetic 0.25 f l 0.1 pM 0.025 f l 0.01 f l anhydride (mM) FCCP FCCP gramicidin gramicidin
3.5 3.5
177 253
245 247
88 186
150 181
3.5
517
546
444
464
a Chloroplasts were suspended in 40 ml of reaction medium containing 100 mM sucrose, 50 mM KCI, 2 mM MgCI,, 50 mM HEPPS-NaOH, pH 8.6, 30 pM pyocyanine, 1 mM dithiothreitol, 1.2 f l DCMU, 0.8 mg chlorophyll, the indicated concentration of uncoupler, and 3.5 m M acetic anhydride (Ac20). Two separate batches of market spinach were used to characterize the effect of FCCP and gramicidin for each set of uncoupler concentrations. Treatments were as follows: (1) dark plus DCMU plus acetic anhydride, conditions giving inhibition of water oxidation. Chloroplasts were incubated 45 seconds in the presence of acetic anhydride followed by the addition of 50 mM N-glycylglycine. The suspension was then centrifuged at 12,000 g for 1 minute, resuspended in 35 ml of 100 mMsucrose, 50 mMKC1.2 mM MgCI,, and 50 mM HEPPS-NaOH, pH 8.6, centrifuged again, and finally resuspended in 0.3 ml of this medium. (2) PS I light conditions: after incubation with acetic anhydride, chloroplasts were processed as in (1). The acetic anhydride in this case was added 30 seconds after the light was turned on. (3) PS I1 + PS I light conditions, DCMU omitted. The resuspended samples (1-3) were then assayed for H20-MV electron transfer activity at 20 pg chlorophyll/ml with 5 f l gramicidin added to ensure that all samples were uncoupled. Light intensities for (2) and (3) were adjusted to each uncoupler concentration to give the same net H + accumulation of -70 nmoles H+/mg chlorophyll. At the lower light intensity for PS I1 + PS I, 100% light protection against anhydride inhibition of water oxidation was observed at 0.25 pM FCCP, 0.1 pM FCCP, 0.025 f l gramicidin, and 0.01 f l gramicidin. The higher light intensity used to drive PS I (+DCMU) activity demonstrated 85% light protection at 0.25 pMFCCP, 97% at O.lOflFCCP, 97% at 0.025 flgramicidin, and 100% at 0.01 pM gramicidin. Dark inhibition of water oxidation by acetic anhydride at each uncoupler concentration was 72% at 0.25 pM FCCP, 48% at 0.1 f l FCCP, 62% at 0.025 pM gramicidin, and 39% at 0.01 f l gramicidin. These measures of light protection and dark inhibition by acetic anhydride were determined according to the procedure outlined in Fig. 2 of Baker et a/. (1981).
such as t o give the same level of H + accumulation in both cases (about 50 nmoles/mg chlorophyll). This was followed by centrifugation and washing, to remove the DCMU from the PS I treatment. Subsequent assay for H,O- MV electron transfer activity showed that, when the uncoupler concentration was low (0.1 pM FCCP or 0.01 pM gramicidin), the P S I-only electron and proton transfer functions provided no protection against inhibition of water oxidation but that PS I I plus PS I activity did protect.
360
R. A. DILLEY
et 6'1.
However, by raising the uncoupler concentrations (0.25 pM FCCP or 0.025 pM gramicidin) we observed some protection against inhibition of water oxidation when the PS I reaction was the sole source of electron and proton transfer (Table IV; cf. Baker et al., 1981). These data are consistent with the concept that proton release and the putative -NH, group@) associated with water oxidation occur in a buried domain not normally available to protons translocated by the PS I redox reactions. Once again, since the inner aqueous space is undoubtedly acidified by the PS I protolytic reactions, the buried domain cannot be in rapid equilibrium with the inner aqueous space unless artificial access is afforded-e.g., by a high concentration of uncoupler. It is to be expected that, if an uncoupler can dissipate acidity in the putative intramembrane domain under dark conditions (leading to the reaction NHf- NH, + H+), then a sufficient uncoupler concentration should allow the inner aqueous phase, acidified under PS I conditions, to deliver protons to the-normally PS 11-specific-intramembrane domain. C. Membrane Proteins Showing Differential Acetic Anhydride Reactivity in Light and Dark
Protons released in water oxidation specifically interact with the 8000MW CF, subunit of the energy-coupling complex. Separation of membrane proteins on SDS-PAGE gels after dark or light exposure to [3H]acetic anhydride permits a determination of the polypeptide units which experience different levels of reactivity with the chemical modifier. This should define the - NH, group-containing proteins which constitute the putative intramembrane, PS 11-specific, proton-processing domain. Figure 5 shows typical SDS-PAGE gel traces and the radioactivity counts recovered in 1-mm gel slices. Table V summarizes the dark-light differences in specific activity of the acetic anhydride labeling for PS I1 plus PS I or PS I plus DCMU under light conditions. As discussed in Prochaska and Dilley (1978b,c), the major protein showing the PS I1 protolysis-dependent labeling change (band V) is the 8000 MW subunit of the CF, complex. This is the dicyclohexylcarbodiimidesensitive protein thought to be involved in H + conduction to CF, . The CF, protein was isolated and purified by butanol extraction and ether precipitation (Nelson et al., 1977). In subsequent work, we employed a different purification, the Fillingame (1976) technique with CHC1,-MeOH extraction and DEAE and Sephadex LH-20 column chromatography (Tandy, 1979). The purified, homogeneous protein has been partially sequenced (up to 40 residues out of 81 total) giving exact agreement so far with the
361
19. PROTON-MEMBRANE INTERACTIONS
r
l2
0
5
10
15
20
25
30
35
40
45
50
55
60
GEL SLICE NO.
FIG.5. SDS-gel electrophoresis of chloroplast membranes labeled by acetic anhydride in the light. Eight percent SDS-PAGE gels were used. Comassie blue-stained protein bands are shown by the solid trace, counts per minute of 1-mm slices of the gel are given by the dashed line. See Fig. 1 of Prochaska and Dilley (1978b) for details.
reported CF,, sequence published by Sebald et af. (1979). The protein, from dark- and light-exposed membranes treated with acetic anhydride, was purified in this way and checked for identity. The samples showed the same type of labeling difference (Table VI) reported earlier using the butanol extraction technique (compare to Table V). The amount of acetate incorporated into the 8000-molecular-weight protein was about 0.5 mole/mole of protein (dark) and 0.2 mole/mole protein (light) (Tandy, 1979). The amino acid composition and sequence data show one lysine and one tyrosine residue in the 8000-molecular-weight CF, protein, the only two likely sites for acetylation by acetic anhydride; the N-terminus is an N-formylmethionine (Sebald et af., 1979; Tandy, 1979). The tyrosine acetylation adduct is unstable at alkaline pH, so it is quite likely that the only residue acetylated is the lysine. A protein of molecular weight near 14,000 shows about 25% of the total light-dark difference in acetic anhydride labeling (Table V, and Prochaska and Dilley, 1978b). Ellenson et al. (1978), showed that light-dark dif-
362
R. A. DILLEY
et a/.
TABLE V EFFECTS OF P s 1 AND P s 11 ELECTRON TRANSPORT UPON THE lNCORPORATlON OF ACETIC ANHYDRIDE INTO CHLOROPLAST PROTEINS AND THE ISOLATED CF, COMPONENT'
Peak
Apparent MW
Area Counts of per protein minute stainb
Light-treated (H,O-MV) 57,000-60,000 3424 I 36,000 1409 I1 111 22,000 12,922 IV 13,000 5158 V 7200 4842 6000 V(iso1ated) Dark-treated I I1 111
IV V V(isolated)
57,000-60,OOO 4858 36,000 1698 22,000 16,054 13,000 6329 8310 7200 6600 -
cpm/ unit area
A Specific A Specific activity, A dark PS I1 activity ("0)
-
-
530 350 380 590 460 430
8.7 4.9 39.6 9.2 12.4 -
560 350 405 680 670 690
30 0 25 90 210 -
6.5 4.0 33.8 8.8 10.5
-
-
-
-
8 0 7 25 59
-
(I Chloroplasts were modified as described in Prochaska and Dilley (1978b). Electrophoresis of the CF, component and chloroplast membrane proteins were performed concurrently. Both the isolated CF, component and the chloroplast membrane gels were corrected for background. The area of protein stain for each peak was integrated by weighing. The A specific activity column represents the observed change in specific activity in each protein peak as a function of treatment. For example, the A specific activity for the H,O-MV treatment represents the specific activity of the dark treatment minus that of the H,O-MV treatment for each individual protein peak. For a more complete discussion, see Prochaska and Dilley (1 978b). Relative measure.
ferences in fluorescamine incorporation occurred in a 14,000-molecularweight component also thought to be part of CF,. It is assumed that part of the light-dark anhydride labeling difference should appear in a protein(s) (other than the CF, proteins) associated with the water oxidation apparatus. Perhaps some of the measured differences in acetic anhydride label incorporated into bands I, 11, and 111 may represent such proteins, but this remains to be demonstrated. Figure 6 shows a model of PS I- and PS 11-specific domains connecting the protolytic reactions directly with the CF,-CF, coupling complex. The question mark in the PS I domain indicates our uncertainty about the directness of the connection between the protolytic site (PQH, oxidation) and the CF,.
363
19. PROTON- M EMBRANE INTERACTIONS
TABLE VI INCORPORATION OF ACETICANHYDRIDE INTO LIGHT-A N D DARKTREATED CHLOROPLASTS AND INTO THE ISOLATED 8000-MOLECULAR-WEIGHTPROTEIN' nmoles acetate Treatment Light Dark
mg chlorophyll (whole thylakoids)
cpm/mg protein
nmoles acetate/nmole protein
33+3 40+2
440 990
0.20
~
~~~
0.45
~~
a Thylakoid membranes were exposed to 15 seconds of dark or light in the presence of 1 mM [3H]acetic anhydride, followed by quenching of the reagent and purification of the 8000-molecular-weight CF, protein (Table 3 of Tandy, 1979, thesis). The protein from each treatment was purified to homogeneity as shown by agreement of the amino acid composition data with the earlier data in our hands and that of Sebald et al. (1979). The whole thylakoid incorporation data are from eight separate determinations, and the data for the incorporation into the protein are from a single determination.
IV.
CONCLUDING REMARKS
The work discussed in this article is consistent with the hypothesis that the initial release of protons in PS I1 water oxidation (and perhaps also in the PS I protolytic reaction)-is into a domain, presumably intramembranal, which does not receive protons from other protolytic reactions. Figure 6 shows a diagrammatic model of this concept, which can be compared with the traditional Mitchell concept shown in Fig. 1 . The data strongly support a direct interaction between PS II-linked protons and the 8000-molecular-weight part of the CF,, so the acetic anhydride-detected change in the CF,, specific to PS I1 proton release, seems to be directly involved in proton movements associated with the energy transduction (ATP formation) in chloroplasts. This scheme, or some variant of it, seems necessary because the data described above are not consistent with the transmembrane, bulk phase-tobulk phase, proton gradient model initially put forth by Mitchell (1966). Yet it is clear that a ApH near 3 units (inside pH z 5 , outside pH G 8) builds up across the thylakoid membrane in the light (Gaensslen and McCarty, 1971; Rottengerg et al., 1972) and is decreased by about 0.4 pH units when ATP formation is coupled to the system (Pick et al., 1973; Portis and McCarty, 1974). These facts are not necessarily explained only by the transmembrane H + gradient model but can be consistent with intramembrane H + processing (or local pH effects) if we postulate that protons in the CF, channel (or well) can equilibrate with the inner aqueous space under steady state conditions.
4
H+ Inner aqueous phase
FIG.6. A model depicting intramembrane, site-specific proton processing in a chloroplast thylakoid membrane. The &Fo-NH2 shown in the CF, subunit implies the one lysine amino group of a single 8000-molecular-weight polypeptide (Sebald eta/., 1979) interacting with protons released specifically by PS 11-dependent water or DPC oxidation (cf. Fig. 1). This lysine -NH2 group is acetylated by acetic anhydride when there are no protons (derived from water oxidation) available in the intramembrane domain, and the group is protected against derivatization when in the -NHf form. Protons from PQH, oxidation cannot reach the k F0 - N H 2group, as discussed in the text. The artwork for this model was prepared by Ms. Kathy Shuster.
19. PROTON-MEMBRANE INTERACTIONS
365
Figure 6 indicates a direct deposition of protons into the CF, “channel,” whence the protons can go out through the CF, or go into the inner aqueous phase. Either a transmembrane electrical potential (A$) or a more localized A$, positive toward the inner end of the proton channel, would tend to force protons toward the CF, end. The resulting proton accumulation, shown in Fig. 6, will occur so as to satisfy the Nernst equation, A$=RT/nFln ([H+Ii/[H+],),where [H+Iiand [H+], out are proton activities (concentrations) at the bottom and top of the well, respectively. We know from the work of Schuldiner et al. (1973), Ort and Dilley (1976), Graber et af. (1977), and Vinkler et al. (1978) that, under appropriate conditions, a A$ can contribute to the energy needed for ATP synthesis. Particularly in the initial 50 msec of an illumination period, the A$ is absolutely needed to allow ATP formation t o begin (Ort and Dilley, 1976). The A$ could then contribute toward orienting the flow of protons toward the CF,. A critical question concerns the rapidity of proton equilibration between the CF, well and the inner space. A slow equilibration, for whatever structural reason, would permit driving of ATP formation by the local proton gradient, such that permeant buffers in the inner aqueous space would not be able to buffer out the critical ApH. Ort et al. (1976) and Ort and Dilley (1976) tested for this situation and concluded that, indeed, permeant buffers in the inner aqueous space did not readily equilibrate with the protons thought to drive ATP formation. Similar experiments, but with the opposite type of permeant buffer effect, have been reported by Vinkler et af. (1980) and Davenport and McCarty (1980). If experimental conditions can vary the effectiveness of the direct utilization of protons (in the well) for driving ATP formation, or can vary the rate of equilibration of protons between the proton channel and the inner aqueous space, perhaps both types of data could be explained. The model we are proposing (Fig. 6) requires either a close association of the PS I1 reaction center with the CF,-CF, complex or a mechanism for long-distance transport of the PS II-linked protons within the putative PS II-specific proton-conducting domain. Recent ultrastructural evidence suggests that there are no CF, protein complexes within the partition regions of the stacked thylakoid arrays (Miller and Staehelin, 1976; Murakami and Kunieda, 1977). However, Oleszko and Moudrianakis’ (1974) ultrastructural data indicate that there may be CF, complexes in the stacked regions which are more deeply immersed in the membrane matrix than the CF, complexes on the nonoppressed regions and thus not readily visualized by the staining techniques normally used. Ultrastructural and biochemical evidence indicates that the partition regions of stacked thylakoids have a uniform distribution of PS I1 and PS I reaction centers (Arntzen et al., 1969, 1972), as indicated by the distribution of the large (PS I1 markers)
366
R. A. DILLEY et a/.
and small (PS I markers) freeze-fracture-revealed particles (Fig. 2b and c). It is expected, then, that protolytic reactions occur in the central regions of a stacked thylakoid disc. Further studies may show that the stacked partition regions of thylakoids do not, in fact, have CF, complexes, but rather CF, occurs on the end parts of the stacked disc, on the outer part of the end thylakoid, and on the stroma thylakoids. In that case long-distance proton translocation would be required for our hypothesis to be valid. This remains an important unsolved problem; speculation therefore can run rampant. Is it certain that membrane regions buried in the partitions have actively functioning electron transport chains? It seems a generally accepted point, but several lines of evidence actually call it into question. Hall et al. (1971), in a cytochemical study, showed that PS 11-dependent ferricyanide reduction formed a reaction product (cupric-ferrocyanide) primarily at the edges of thylakoid stacks. It is not clear in these experiments whether product migration to “crystallization centers” occurred from the partition regions to the edges, or whether the product was formed at the edges. Higher degrees of membrane stacking correlate with slower electron transport rates, expressed either on a chlorophyll basis or on a reaction center basis, (Boardmann, 1977). This is especially clear in sungrown plants, compared to shade-grown plants. Moreover, the apparent rate-limiting step of electron transport, plastohydroquinone oxidation, is significantly slower in stacked thylakoids, compared to unstacked thylakoids (Keck et a[., 1970). While we are not suggesting that the evidence is strongly for the location of electron transfer only near the edges of stacked regions, this issue should be examined more carefully than it has been to date. One speculation about proton transfer in the membrane is that protons might be conducted along hydrogen-bonding groups of proteins within the membrane by the “proton-hopping” mechanism postulated by Nagle and Morowitz (1978). This is analogous to the rapid movement of protons along hydrogen-bonding chains of water molecules in structured water or ice. An attractive feature of the proton-hopping model is that it can explain the failure of uncouplers to abolish the PS I1 proton-specific acetic anhydride labeling effects on the 8000-molecular-weight unit of the CF,. There is not enough information about chloroplast membrane structure to identify a particular protein(s) that might play such a role. However, the fact that there are at least four to six 8000-molecularweight CF, subunits per CF, (Sigrist-Nelson et al., 1978; Sigrist-Nelson and Azzi, 1979) means that this polypeptide could be a candidate for intramembrane proton conduction. Nelson et al. (1977) showed that the 8000molecular-weight proteolipid incorporation into liposomes permitted dicyclohexylcarbodiimide-sensitive proton accumulation, suggesting that
19. PROTON-MEMBRANE INTERACTIONS
367
the 8000-molecular-weight protein has proton conduction properties. The question of whether the 8000-molecular-weight part of CF, may be excluded from the stacked membrane regions, as indicated for the CF, protein, should be studied. It can be easily tested by analyzing the appropriate membrane fragments. Such work is underway in our laboratory. Earlier ultrastructural studies have revealed ribbonlike structures derived from both chloroplasts (Stiles et al., 1968; Arntzen et al., 1972) and mitochondria (Haworth et al., 1977). The structures are closely associated with the CF, or F,. Further work is necessary to determine whether the ribbonlike structures are related in any way to functional activity (rather than being artifacts of electron microscopy), but the possibility exists that such long, ribbonlike structures are manifestations of an ordered array of membrane proteins which provide a pathway for proton movement. In keeping with the spirit of this volume, we should end on a note concerning electrogenic aspects of proton fluxes. The concept we have put forth implies local electrogenic proton gradients within membrane domains which are specific for each of the photosystem protolytic reactions. Further work is necessary to inquire into such aspects as mobile counterion (e.g., K + , Mg2+)exchange in response to the proton release. ACKNOWLEDGMENTS The authors wish to thank Drs. William A. Cramer and Deepak Bhatnagar and Ms. Marilyn Abbot for fruitful discussions of this work, and Ms. Jan Vanderbilt for her help in the preparation of the manuscript. This work was supported in part by grants PCM 76-01640 from NSF and GM 19595-06 from NIH.
REFERENCES Arntzen, C. J., Dilley, R. A., and Crane, F. L. (1969). J . Cell B i d . 43, 16-31. Amtzen, C. J., Dilley, R. A., Peters, G. A., and Shaw, E. R. (1972). Biochim. Biophys. Acfa 256, 85-107. Baker, G. M., Bhatnager, D., and Dilley, R. A. (1981). Biochem. 20, 2303-2315. Crofts, A. R., Deamer, D. W., and Packer, L. (1967). Biochiom. Biophys. Acfu 131,97-118. Davenport, J. W., and McCarty, R. E. (1980). Biochim. Biophys. Acfu 589, 353-357. Deamer, D. W., Crofts, A. R., and Packer, L. (1967). Biochim. Biophys. Actu 131, 81-96. Dickerson, R. E., and Geis, I. (1969). “The Structure and Action of Proteins.” Harper, New York. Dilley, R. A. (1971). Curr. Top. Bioenerg. 4, 237-271. Dilley, R. A., and Prochaska, L. J. (1978). In “The Proton and Calcium Pumps” (G. F. Azzone, M. Evron, J. C. Metcalfe, E. Quagliariello, and N. Filiprandi, eds.), pp. 45-54. Elsevier, Amsterdam. Dilley, R. A., and Rothstein, A. (1967). Biochim. Biophys. Acfu 135, 427-443.
368
R. A. DILLEY et
a/.
Dilley, R. A., and Vernon, L. P. (1964). Biochemistry 3, 817-824. Dilley, R. A., and Vernon, L. P. (1965). Arch. Biochem. Biophys. 111, 365-375. Dilley, R. A., Park, R. B., and Branton, D. (1967). Photochem. Photobiol. 6, 407-412. Dunker, A. K., and Marvin, D. A. (1978). J. Theor. Biol. 72, 9-16. Ellenson, J. L., Pheasant, D. J., and Levine, R. P. (1978). Biochim. Biophys. Actu 504, 123- 135. Fillingame, R. H., (1976). J. Biol. Chem. 251, 6630-6637. Gaensslen, R. E., and McCarty, R. E. (1971). Arch. Biochem. Biophys. 147, 55-65. Giaquinta, R. T., Ort, D. R., and Dilley, R. A. (1975). Biochemistry 14, 4392-4396. Graber, P., Schlodder, E., and Witt, H. T. (1977). Biochim. Biophys. Actu 461, 426-440. Hall, D. O., Edge, H., and Kalina, M. (1971). J. Cell Sci. 9, 289-303. Haworth, R. A., Komar, H., Green, D. E., and Vail, W. J. (1977). J. Bioenerg. Biomembr. 9, 151-170. Hoare, D. G., and Koshland, D. E. (1967). J. Biol. Chem. 242, 2447-2453. Horiike, K., and McCormack, D. B. (1979). J. Theor. Biol. 79, 403-414. Itoh, M., Izawa, S., and Shibata, K. (1963). Biochim. Biophys. Acta 66, 319-327. Izawa, S., and Ort, D. R. (1974). Biochim. Biophys. Actu 357, 127-143. Jagendorf, A. T. (1975). In “Bioenergetics of Photosynthesis” (Govindjee, ed.), pp. 414-492. Academic Press, New York. Jagendorf, A. T., and Hind, G. (1963). In “Photosynthetic Mechanisms of Green Plants,” pp. 599-610. NAS-RNC Publ. NO. 1145. Keck, R. W., Dilley, R. A., and Ke, B. (1970). Plant Physiol. 46, 699-704. Kraayenhof, R., Izawa, S., and Chance, B. (1972). Plant Physiol. 50, 713-718. Leiser, M., and Gromet-Elhanan, Z. (1977). Arch. Biochem. Biophys. 178, 79-88. Means, G. E., and Feeney, R. E. (1971). “Chemical Modification of Proteins,” p. 69. Holden-Day, San Francisco, California. Melandri, If. A., De Santis, A., Venturoli, G., and Melandri, A. B. (1978). FEBS Lett. 95, 130- 134. Miller, K. R., and Staehelin, L. A. (1976). J. Cell Biol. 68, 30-47. Mitchell, P. (1961). Nature (London) 191, 144-148. Mitchell, P. (1966). Biol. Rev. Cambridge, Philos. SOC.41, 445-540. Murakami, S., and Kunieda, R. (1977). In “Photosynthetic Organelles,” Special Issue of Plant Cell Physiol. pp. 403-414. Nagle, J. F., and Morowitz, H. J. (1978). Proc. Natl. Acud. Sci. U.S.A. 75, 298-302. Nelson, N., Eytan, E., Notsani, B., Sigrist, H., Sigrist-Nelson, K., and Gitler, C. (1977). Proc. Natl. Acud. Sci. U.S.A. 74, 2375-2378. Oleszko, S., and Moudrianakis, E. N. (1974). J. Cell Biol. 63, 936-948. Ort, D. R., and Dilley, R. A. (1976). Biochim. Biophys. Actu 449, 95-107. Ort, D. R., Dilley, R. A., and Good, N. E. (1976). Biochim. Biophys. Actu 449, 108-124. Packer, L., Marchant, R. H., and Mukohata, Y. (1963). Biochim. Biophys. Actu 75, 23-30. Pick, U., and Racker, E. (1979). J. Biol. Chem. 254, 2793-2799. Pick, U., Rottenberg, H., and Avron, M. (1973). FEBS Lett. 32, 91-94. Portis, A. R., and McCarty, R. E. (1974). J. Biol. Chem. 249, 6250-6254. Prochaska, L. J., and Dilley, R. A. (1978a). Arch. Biochem. Biophys. 187, 61-71. Prochaska, L. J., and Dilley, R. A. (1978b). Biochem. Biophys. Res. Commun. 83, 664-672. Prochaska, L. J., and Dilley, R. A. (1978~).In “Frontiers of Bioenergetics ” (P. L. Dutton, J. S. Leigh, and A. Scarpa, eds.). Vol. 1, pp. 265-274. Academic Press, New York. Rottenberg, H., Grunwald, T., and Avron, M. (1972). Eur. J. Biochem. 25, 54-63. Schliephake, W., Junge, W., and Witt, H. T. (1968). Z. Naturforsch. B 23, 1571-1578. Schroder, H., Muhle, H., and Rumberg, B. (1972). Proc. Znt. Congr. Photosyn. Res., End., I971 Vol. 11, pp. 919-930.
19. PROTON-MEMBRANE INTERACTIONS
369
Schuldiner, S., Rottenberg, H., and Avron, M. (1973). Eur. J. Biochem. 39, 455-462. Sebald, W., Hoppe, J., and Wachter, E. (1979). In “Function and Molecular Aspects of Biomembrane Transport” (E. Quagliariello et a/., eds.), pp. 63-74. Elsevier, Amsterdam. Sigrist-Nelson, K., and Azzi, Z. (1979). J. B i d . Chem. 254, 4470-4474. Sigrist-Nelson, K., Sigrist, H., and Azzi, A. (1978). Eur. J. Biochem. 92, 9-14. Stiles, J. W., Wilson, J. T., and Crane, F. L. (1968). Biochim. Biophys. Acta 162, 631-634. Tandy, N. E. (1979). Evidence for an interaction between proton release in photosystem I1 and the 8 KD subunit of the energy coupling complex of spinach chloroplasts. M.S. Thesis, Purdue University. Timasheff, S. N., and Gorbunoff, M. J. (1967). Annu. Rev. Biochem. 36, 13-54. Trebst, A. (1974). Annu. Rev. Plant Physiol. 25, 423-458. Van Dam, K., and Westerhoff, H. V. (1977). In “Structure and Function of Energy Transducing Membranes” (Van Dam and Van Gelder, eds.), pp. 157-167. Elsevier, Amsterdam. Vinkler, C., Avron, M., and Boyer, P. D. (1978). FEBS Lett. 96, 129-134. Vinkler, C., Avron, M., and Boyer, P. D. (1980). J. Biol. Chem. 254, 10,654-10,656. Williams, R. J. P. (1961). J. Theor. B i d . 1, 1-17. Williams, R. J. P. (1975). FEBS Lett. 53, 123-125. Witt, H . T. (1971). Q. Rev. Eiophys. 4, 365-477.
This Page Intentionally Left Blank
CURRENT TOPICS IN MEMBRANES AND TRANSPORT. VOLUME 16
Chapter 20
Photochemical Charge Separation and Active Transport in the Purple Membrane BARRY HONIG' Department of Physiology and Biophysics The University of Illinois Urbana, Illinois
I. 11.
111. IV. V.
..........................................
.................................
ochemical Event ............................................................ A. The Relation of Bacteriorhodopsin to Visual Pigments ............................ B. Photoisomerization and Charge Separation ..................... ............ Mechanistic Implications of Steady State Kinetics ......................................... Relating Kinetic and Molecular Models ............................................... Summary ..................................................................................... References ..........................................................................................
1.
371 372 372 374 377 379 38 1 381
INTRODUCTION
The purple membrane of Halobacterium halobium contains a single protein, bacteriorhodopsin, which functions as a light-driven proton pump (for a review see Stoeckenius et al., 1978). Bacteriorhodopsin consists of a single polypeptide chain whose sequence has recently been determined (Ovchinikov et al., 1979; Khorana et al., 1979) and which folds into seven a-helices that traverse the cell membrane (Unwin and Henderson, 1975). It is expected that the continued application of a variety of diffraction techniques currently in progress in a number of laboratories will ultimately yield a high-resolution structure for bacteriorhodopsin. For this reason it is likely to be the first system in which the mechanism of active transport is Present address: Department of Biochemistry, Columbia University, New York, New York. 371
Copyright 0 1982 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-153316-6
372
BARRY HONIG
understood in molecular detail and as such may reveal insights that are relevant to other systems. The existence of a well-resolved protein structure is of course no guarantee that a mechanism will be easily determined. However, the retinal chromophore of bacteriorhodopsin adds an extra dimension to mechanistic studies, since it provides an intrinsic optical probe whose spectroscopic properties are well understood and whose photochemical transformation is the first step in the pumping mechanism. In order t o characterize fully the mechanism of active transport in the purple membrane it will be necessary to understand the nature of this primary photochemical event and to consider how it can initiate the vectorial flow of protons. To this end, a model of the primary photochemical event derived from available experimental data will first be presented. Some essential requirements for any proton pump will then be discussed in terms of a recent steady state kinetic analysis of active transport. Finally, the possible relation between the kinetic and molecular models will be considered. 11.
THE PRIMARY PHOTOCHEMICAL EVENT
A. The Relation of Bacteriorhodopsin to Visual Pigments, The name “bacteriorhodopsin” originates from the close similarities between this pigment and the visual pigment, rhodopsin, of vertebrate rod photoreceptors (see Honig, 1978a, for a review of the optical properties of visual pigments and bacteriorhodopsin). Since the model to be discussed below is derived partially from analogies to visual pigments, it is useful to review the structural and functional relationship between the two systems. The first and most important similarity is that both pigments use essentially the same chromophore; all-trans-retinal in bacteriorhodopsin and 11-cisretinal in visual pigments (Fig. 1). Both chromophores are covalently bound to the apoprotein via a Schiff base linkage to the +amino group of a lysine, and the Schiff base is protonated in both cases (Oseroff and Callender, 1974; Lewis et al., 1974; Aton et al., 1977). Both proteins traverse the membrane and are largely a-helical, particularly bacteriorhodopsin, which is almost all helical. There is no apparent functional relationship between rhodopsin and bacteriorhodopsin. Bacteriorhodopsin is a proton pump (Oesterhelt and Stoeckenius, 1973), but the specific role of rhodopsin in initiating visual excitation has not yet been determined. However, visual pigments are clearly not involved in active transport, since the time scale for regeneration of the pigment following light absorption is on the order of many minutes. In contrast, bacteriorhodopsin undergoes a photochemical cycle
20. ACTIVE TRANSPORT IN THE PURPLE MEMBRANE
373
FIG. 1. Conformations of various isomers of retinal (X=O), its Schiff base ( X = N ) , and its protonated Schiff base (X = NH’). (a) All-trans;(b) 1 3 4 ; (c) 11-cis.
X
(b)
CH,
CH,
“““XI, CH3
CH,
CH,
(C)
which is complete in about 10 msec (Lozier and Niederberger, 1977) and which is thus short enough to provide the fast regeneration time required by its function as a pump. Despite these functional differences, there are remarkable photochemical similarities between the two pigments (Fig. 2). In both cases the primary photochemistry produces a red-shifted photoproduct which is stable at 77 K. These species, K for bacteriorhodopsin and bathorhodopsin for visual pigments, may be photochemically reverted to the parent pigment. However, they decay thermally through a series of intermediates to a blue-shifted species in which the Schiff base has been deprotonated. At this stage the behavior of the two systems diverges; in the case of bacteriorhodopsin the parent pigment is regenerated, while in visual pigments the chromophore dissociates from the protein. The primary photoproducts appear to be closely related. As mentioned above, both are red-shifted relative to the parent pigment, and both initiate a sequence of events leading to deprotonation of the Schiff base. Bathorhodopsin and K are both formed in picosecond times at room temperature (Applebury et af., 1978; Peters et af., 1978). Their rate of formation is decreased at low temperature and is further inhibited by replacement of all the exchangeable protons on the pigment with deuterons. Finally, both species are high-energy forms in which a significant fraction of the photon’s energy has been stored (Rosenfeld et af., 1977; Honig et al., (1979a).
374
BARRY HONIG bocteriorhodopsin (570)
IT all-trans retinal
+ opsin
( b)
FIG. 2. (a) The photochemical cycle of bacteriorhodopsin. (b) The bleaching sequence of rhodopsin. A number of intermediates appear along the thermal pathways denoted by T. For bacteriorhodopsin the primary photoproduct, K, decays to a short-wavelength form, M, from which the parent pigment is regenerated. In visual pigments the primary photoproduct is bathorhodopsin and it decays, irreversibly, to all-trans-retinal and opsin. Absorption maxima appear in parentheses.
B. Photoisomerization and Charge Separation
There has been considerable interest in obtaining a model which accounts for these observations. In the case of visual pigments it was proposed many years ago by Wald and collaborators that the primary event was a photochemical isomerization of 11-cis-retinal to all-trans-retinal (see e.g., Yoshizawa and Wald, 1963). The strongest evidence for this process is the fact that bathorhodopsin may be produced from either rhodopsin or isorhodopsin, the artificial pigment formed from 9-cis-retinal. The three pigments form a photostationary state at low temperature [and in intense
20. ACTIVE TRANSPORT I N THE PURPLE MEMBRANE
375
light at room temperature (Goldschmidt et al., 1976)] which may be represented by the relationship: rhodopsin (1 1-cis) + bathorhodopsin 7’ isorhodopsin ( 9 4 9 . Since bathorhodopsin is intermediate between two cis isomers, by far the most probable conformation of its chromophore is alltrans. Raman scattering results do indeed suggest a transoid conformation which is similar but not identical to the conformation of the all-trans model compound in solution (Eyring and Mathies, 1979; Doukas et al., 1980). This, plus the fact that all-trans-retinal dissociates thermally from the protein at a later stage, leaves little doubt that the primary event is indeed a cis-trans isomerization. On the basis of photochemical similarities between bacteriorhodopsin and visual pigments, we first suggested that the primary photochemical event in bacteriorhodopsin was an isomerization as well (Rosenfeld et al., 1977). This conclusion has gained support from a number of other studies, and-in particular-a trans-cis photoisomerization about the 13,14-double bond has been implicated (Pettei et al., 1977). However, it is clear that isomerization alone cannot explain the spectral red shift, the energy storage, and the proton motion which characterize the photochemical transformation in both pigments. First, in one pigment the isomerization is cis-trans and in the other is presumably trans-cis, so that changes in isomer composition cannot be implicated in properties that are characteristic of both pigments. Second, isomerization is not associated with significant spectral changes or energy differences when the chromophores are free in solution (Hubbard, 1966) and not bound to their respective apoproteins. Recently, however, we showed that the characteristic properties of the primary photochemical event in both the visual pigment and bacteriorhodopsin are natural consequences of a photoisomerization within the binding site of the protein (Honig et af., 1979a). Protonation of the Schiff base implies that a charged species is buried in the interior of a membrane protein. Given the low-dielectric medium, we suggested a number of years ago that, in all retinal-based pigments, a salt bridge is formed between the Schiff base nitrogen and a negatively charged amino acid located approximately 3 A from the chromophore (Honig et al., 1976). Alternatively, it is conceivable that the Schiff base proton is “solvated” by buried water molecules or dipolar groups on the protein. In either case, the effect of a photoisomerization, either trans- cis or cis trans, is to separate charge in the interior of the protein (Honig et al., 1979a). It is this charge separation that distinguishes isomerization in the interior of a protein from that in solution (Fig. 3). A spectral red shift is an immediate consequence of charge separation at the Schiff base terminus of the chromophore. It has recently been shown
-
376
BARRY HONIG
FIG. 3. Model for the primary photochemical event in bacteriorhodopsin. The chromophore is depicted with its Schiff base forming a salt bridge with a negativecounterion.The additional charge pair near the ring represents the group that determines the absorption maxima of the pigments. The photochemical event is an isomerization (probably about the 13,14-double bond), but any isornerization in any direction will produce charge separation as shown. The pK values of the Schiff base and those of other groups on the protein, such as R, and R,, are strongly affected by photoisomerization because a salt bridge is broken, a positive charge has moved near R,, and R, is now a bare negative charge. Possible proton transfer steps resulting from charge separation are depicted.
that the red shift in the absorption maximum of rhodopsin (A,, 2 500 nm) and bacteriorhodopsin (Amu 570 nm) relative to that of the protonated Schiff base in solution (A,,,~450 nm) is due to charged or dipolar groups located at different positions along the polyene chain. At the rhodopsin binding site these are located near the 11,12-double bond (Honig et al., 1979a), while in bacteriorhodopsin they are close to the P-ionone ring (Nakanishi et al., 1981). The further red shift of the primary photoproduct arises from separation of the chromophore from the Schiff base counterion. Energy storage is also an immediate consequence of charge separation in a low-dielectric medium. A simple estimate based on experimental data yields a value near 30 kcal/mole (Honig et al., 1979a). Of course, actually to “store” the energy long enough so that the decay of an unstable intermediate is channeled in the desired direction, it is necessary to prevent back reaction. In the case at hand this is made possible by an intrinsic large activation energy (20-30 kcal/mole) for thermal cis trans isomerization about double bonds. While there is enough energy in a photon (-50 kcal/mole) to surmount this barrier, thermal isomerizations should be extremely slow. A final consequence of charge separation in the interior of the protein will be a large change in the pK of various groups in the vicinity of the chromophore. The Schiff base proton, having moved away from its counterion, will of course be more acidic; but since the entire charge balance is
=
-
377
20. ACTIVE TRANSPORT IN THE PURPLE MEMBRANE
now disrupted, other groups will be affected as well. We have argued (Honig et at., 1979a) that the flop of protons between such groups gives rise to the deuterium effects detected in picosecond measurements of the formation of both K (Applebury et al., 1978) and bathorhodopsin (Peters et at., 1978).
111.
MECHANISTIC IMPLICATIONS OF STEADY STATE KINETICS
Figure 4 shows a four-state kinetic model of the proton pumping cycle of bacteriorhodopsin. I have previously presented a steady state kinetic analysis of this model (Honig, 1978b). Here I review the major mechanistic conclusions which have resulted from that analysis and from a similar study reported recently by Lauger (1979). The minimum requirement for any pump is that it exist in two states in which the transported substrate is exposed, respectively, to one or the other side of the membrane (Stein and Honig, 1977; Honig and Stein, 1978). In Fig. 4 these are designated E, and E,, both of which may be either unprotonated or protonated (E,H or E,H). (The four forms depicted represent the energetically most stable states or a composite defined by a number of states of approximately equal free energy. It is possible, for example, that E,H releases a proton through the unstable E;H by E,H-EE;H-E,+ 1H. Here E;H is not a transition state but simply some unstable intermediate.) Transformations between these two states are essentially isomerizations in which the direction of proton release is determined by some conformational change in the protein. Such isomerization either could involve a large-scale conformational change, during which the substrate essentially traverses the membrane, or could involve movement through only a few angstroms in part of the protein, with the bulk of the protein remaining fixed in the membrane. In the latter case there must be channels both to and from the site that actually moves.
FIG. 4. Four-state model for light-driven proton
pumping
in
bacteriorhodopsin.
Forms with subscript 1 bind protons on the inside of the cell, while those with subscript 2 bind protons on the outside. All steps can in principle have light and dark components contributing to their respective rate constants.
hv7 k T
378
BARRY HONIG
To understand how Fig. 4 can be made to represent a pump, consider first the equilibrium situation in which the rates of all forward and back reactions are equal. Now any external pertubation which modifies the rate constants so that in any step one direction is favored over the other will drive a pump. If light is the driving force, as it is in bacteriorhodopsin, vectorial flow can be achieved simply by having the different forms affected differently by light. For example, if light drives the E,H-E,H step, the rate of the photochemical back reaction can be arbitrarily small if E,H does not absorb light of the proper wavelength or has a low quantum yield for reversion to E,H. A negligible photochemical back reaction has in fact been assumed previously to simplify the kinetic analysis of proton pumping in bacteriorhodopsin (Honig, 1978b; Lauger, 1979). (With regard to the back reaction, it is important to realize that photochemical reactions proceed in a different electronic state than thermal reactions. Thus, even if the proton pump in bacteriorhodopsin could be reversed, fluorescence would not be observed but rather the back reaction would proceed entirely in the ground electronic state. For this reason microscopic reversibility is not obeyed, and the application of irreversible thermodynamics to the bacteriorhodopsin problem should be carried out with extreme caution, if at all.) It is clear from Fig. 4 that light could act either by driving the isomerization between states (E,H h-u E,H or E, h-u E l , assuming the flow of protons is in the 1-2 direction) or by affecting pK values (for example, by increasing the rate of proton release on side 2 by driving the reaction Q H h,. E, + H. Since proton uptake precedes proton release in the purple membrane (Lozier et al., 1976), it is evident that E,H-E,H and E,H-E2+H are the only possible candidates for the light-driven event. Which of these kinetically defined steps is in fact effected by light could in principle be determined from titration or exchange measurements that distinguish one side of the membrane from the other, or from the three-dimensional structure of bacteriorhodopsin once it is available. While any unidirectional change in the intrinsic rate constants will produce a proton pump, it is necessary that the various rate constants and affinities be such that fast turnover is possible. Otherwise leaks both internal and external to the pump will minimize the size of the pH gradient that can be achieved. For example, it is clearly advantageous for E l to have a high pK and for E, to have a low pK. Such pK differences can be intrinsic properties of the carrier or can be influenced by light if, for example, a low-pKspecies ( K H ) is produced by the photochemical reaction E,H- K H or E,H-EE;H (Stein and Honig, 1977; Honig, 1978b; Kalisky et af., 1981). It is important to realize, however, that pK changes are not required for the functioning of the pump. In principle, it is perfectly possible that E, and E, both have the same intrinsic pK and that light simply drives the E,H-E,H transformation. On the other hand, it is likely that large pK
20. ACTIVE TRANSPORT IN THE PURPLE MEMBRANE
379
changes do in fact accompany the bacteriorhodopsin photocycle (Kalisky et al., 1981). Since protons are released in 1 msec (or less) following light absorption (Lozier et al., 1976), an upper limit for the pK of the group releasing the proton is about 7 (Honig, 1978b). If the pK of bacteriorhodopsin itself were so low, only half of the molecules absorbing light at pH 7 could actually transport a proton, since only half would be protonated. However, there is no evidence for such a heterogeneous distribution of carriers in the resting state, and indeed it is difficult to believe that such a poorly designed pump could exist.
-
IV.
RELATING KINETIC AND MOLECULAR MODELS
It is interesting to consider possible roles for the light-induced charge separation (described in Fig. 3) in driving the pumping cycle (discussed in Section 111). In fact, photoisomerization and charge separation could play a number of kinetic roles. The simplest possible model is that the isomerization itself changes the exposure of the Schiff base proton from one side of the membrane to the other by moving it a few angstroms (Honig, 1978a). Thus, trans-cis isomerization would presumably move the proton across the permeability barrier separating the inside and outside of the membrane. For this model to be correct it is necessary that, in the resting state, the Schiff base proton be exposed to the inside of the cell so that bacteriorhodopsin itself corresponds to E,H in Fig. 4. The form E2H would then have the Schiff base proton exposed to the outside of the cell. Its pK could be significantly lower than that of E , H if the charge on the Schiff base is not balanced by a counterion. A second type of model in which the Schiff base proton is itself transported assumes that the resting state of bacteriorhodopsin is the kinetic form E2H (Kalisky et al., 1981). Now the role of isomerization of the chromophore cannot be to change accessibility (producing E , H from E,H would be counterproductive) but rather to lower the p K o f the Schiff base. Here, the motion of the proton resulting from photoisomerization would not traverse a permeability barrier, but rather would solely separate the Schiff base from the counterion, thus reducing its pK. From a kinetic standpoint, the photochemical event would drive the transformation E2H-E2 + H . The two models considered so far implicitly assume that the Schiff base proton is actually translocated during a single bacteriorhodopsin photocycle. That is, the proton attached to the Schiff base after completion of a cycle is taken to be different than the original one. That the Schiff base of
380
BARRY HONIG
the chromophore is in fact deprotonated during the course of a photocycle provides indirect evidence for this intuitively pleasing assumption (Lewis et a/., 1974; Aton et a/., 1977). It is, however, possible that the same proton which dissociates from the Schiff base is later rebound and that the pathway for proton translocation across the membrane does not include the chromophore. Moreover, since there is now evidence that two protons are transported per photocycle (Becher and Ebrey, 1978), it is necessary to consider mechanisms for transporting protons other than one that is bound to the Schiff base. Figure 3 suggests how a charge separation resulting from photoisomerization of the chromophore could influence the displacement of protons bound to nearby amino acid residues. In the specific case outlined, the pK of the group R , would be greatly decreased by the presence of a protonated nitrogen, while the original counterion would become more basic. The flow of protons as indicated would then be driven by the photochemistry and would correspond to the kinetic step E,H -E,+ H . Alternatively, the displacement of charge could affect the barrier structure in the vicinity, say, of group R , . As a result, before isomerization its proton would be released to the inside of the cell, but after photoisomerization its proton would be released to the outside. In this case the photochemical reaction would, in effect, drive the E,H -E2H transformation. Finally, it is possible that the photochemistry induces major structural changes in the protein, so that the role of the chromophore will involve effects less obvious than those considered above. In this regard, it is important to reiterate that the model shown in Fig. 3 provides a potentially enormous driving force for ensuing reactions, because large-scale energy storage is a consequence of photoisomerization in a low-dielectric medium. An interesting speculation as to how a small geometric change in the chromophore could be amplified into a large conformational change in the protein is suggested by the amino acid sequence of bacteriorhodopsin (Orchinikov e t a / . , 1979; Khorana e t a / . , 1979). The distribution of charged amino acids is such that it is essentially impossible t o fold the chain into seven helices without burying charges in the interior of the bilayer. It is quite likely that these charges appear in pairs as salt bridges which, from the sequence, would have to connect different helices. Disruption of even one salt bridge, by photoisomerization of the charged chromophore, would severely perturb the charge balance inside the membrane and might cause the helices to “slide” against one another in order to relieve the electrostatic strain. This would constitute a relatively large conformational change whose analysis may require that the dynamic aspects of proton conformation be taken into account.
20. ACTIVE TRANSPORT I N THE PURPLE MEMBRANE
V.
381
SUMMARY
In this article I have tried to indicate how it may be possible to combine kinetic and molecular models in order t o elucidate the mechanism of proton pumping in bacteriorhodopsin. While it may eventually be possible to use diffraction techniques to detect conformational changes that occur during the photocycle, the first structure that will become available will be of bacteriorhodopsin itself. Thus in order to understand the transport mechanism it will first be necessary t o identify the primary photochemical event. The model given in Fig. 3, as well as the experimental evidence and theoretical considerations used in its derivation, should make it possible to define the photochemical transformation in some detail once the structure of the. binding site is known. The kinetic requirements discussed in Section I11 should also aid in defining the primary event, since identification of bacteriorhodopsin with either kinetic form E,H or kinetic form E,H places a constraint on the type of reaction that is possible (i.e., Does it change “exposure” or pK?) Once the structure and primary event are known, it is still not clear whether the entire pumping mechanism will be easily understood. Nevertheless, the possibility of replacing the chromophore with chemically altered retinals (see, e.g., Tokunaga et al., 1977), along with our basic understanding of the spectroscopic properties of retinal and its analogs, leads t o an optimistic assessment for eventually working out the mechanism of active transport in great detail. The possibility that the results may be applicable in some way t o visual pigment research adds additional excitement to this extremely interesting problem. REFERENCES Applebury, M., Peters, K., and Rentzepis, P. (1978). Biophys. J . 23, 275-382. Aton, B., Doukas, A. G., Callender, R . H., Becher, B., and Ebrey, T. G. (1977). Biochemistry, 16, 2995-2999. Becher, B., and Ebrey, T. G. (1978). Biochemisfry, 17, 2293-2300. Doukas, A,, Aton, B., Callender, R., Dinur, U . , and Honig, B. (1980). Biophys. J . (in press). Eyring, G., and Mathies, R. (1979). Proc. Nufl. Acud. Sci. U.S.A. 76, 33-37. Goldschmidt, C., Kalisky, O., Rosenfeld, J., and Ottolenghi, M. (1976). Nature (London) 263, 169-171. Honig, B. (1978a). Annu. Rev. Phys. Chem. 29, 31-57. Honig, B. (1978b). In “Energetics and Structure of Halophilic Microorganisms” (R. Kaplan and M. Ginzberg, eds.), pp. 109-121. Elsevier, Amsterdam. Honig, B., and Stein, W. D. (1978). J . Theor. Biol. 75, 299-305. Honig, B., Greenberg, A., Dinur, U., and Ebrey, T. (1976). Biochemisfry 15, 4593-4599. Honig, B., Ebrey, T. G., Callender, R., Dinur, U., and Ottolenghi, M. (1979a). Proc. Nutl. Acud. Sci. U.S.A. 76, 2503-2507.
382
BARRY HONIG
Honig, B., Dinur, U., Nakanishi, K., Balogh-Nair, V. Gowinowicz, M. A., Arnaboldi, M., and Motto, M. (1979b). J. A m . Chem. SOC.101, 7084-7086. Hubbard, R. (1966). J. Biol. Chem. 241, 1814-1818. Kalisky, O., Ottolenghi, M., Honig, B., and Korenstein, R. (1981). Biochemistry 20,649-655. Khorana, H. G., Gerber, G. E., Herlily, W. C., Gray, C. P., Anderegg, R. J., Nidel, K., and Biemann, K. (1979). Proc. Natl. Acad. Sci. U.S.A. 16, 5046-5050. Lauger, P. (1979). Biochim. Biophys. Acta. 552, 143-161. Lewis, A., Spoonhower, J., Bogomolni, R. A., Lozier, R. H., and Stoeckenius, W. (1974). Proc. Natl. Acad. Sci. U.S.A. 71, 4462-4466. Lozier, R. H., and Niederberger, W. (1977). Fed. Proc. Fed. A m . SOC.Exp. Biol. 36, 18051809. Lozier, R. H., Niederberger, W., Bogomolni, R. A , , Hwang, S., and Stoeckenius, W. (1976). Biochim. Biophys. Acta 440, 545-556. Nakanishi, K., Balogh-Nair, V., Arnaboldi, M., Tsujimoto, K., and Honig, B. (1981). J. A m . Chem. SOC.102, 7945-7947. Oesterhelt, D., and Stoeckenius, W. (1973). Proc. Natl. Acad. Sci. U.S.A. 70, 2853-2857. Orchinikov, Yu. A., Abdulaev, N. G., Feigina, M. Yu., Kiselev, A. V., and Lobanov, N. A. (1979). FEBS Lett. 100, 219-224. Oseroff, A. R., and Callender, R. G. (1974). Biochemistry 13, 4243-4348. Peters, K., Applebury, M. L., and Rentzepis, P. M. (1978). Proc. Natl. Acad. Sci. U.S.A. 14, 3119-3123. Pettei, M . J., Yudd, A. P., Nakanishi, K., Henselman, R., and Stoeckenius, W. (1977). Biochemistry 16, 1955-1959. Rosenfeld, T., Honig, B., Ottolenghi, M., and Ebrey, T. G. (1977). Pure Appl. Chem. 49, 341-351. Stein, W. D., and Honig, B. (1977). Mol. Cell. Biochem. 15, 27-52. Stoeckenius, W., Bogomolni, R., and Lozier, R. (1978). Biochim. Biophys. Acta 505, 21 5-278. Tokunaga, F., Govindjee, R., Ebrey, T. G., and Crouch, R. (1977). Biophys. J . 19, 191-198. Unwin, N., and Henderson, R. (1975). J . Mol. Biol. 94, 425-440. Yoshizawa, T., and Wald, G. (1963). Nature (London) 197, 1279-1286.
CURRENT TOPICS IN MEMBRANES A N D TRANSPORT. VOLUME 16
Chapter 21
Mitochondrial Transhydrogenase: General Principles of Functioning I. A. KOZLOV Isotope Department, A . N . Belozersky Laboratory of Molecular Biology and Bioorganic Chemistry Moscow State University Moscow, USSR
I. Introduction ........................................................................................ The Hypothesis on the Mechanism of AILH+ Generation by the Transhydrogenase Reaction .................................................................... 111. Known Facts and Forecasts ..................................................................... IV. Conclusion .......................................................................................... References ..........................................................................................
383
11.
1.
384 387 391 391
INTRODUCTION
Nicotinamide nucleotide transhydrogenase catalyzes the reversible transport of the hydride ion, H-, in the reaction NAD+ + NADPH + NADH + NADP'
This enzyme, first described by Kaplan et al. (1953, 1956), is widely distributed in nature and is particularly common in the inner membrane of mitochondria and of many bacteria. Danielson and Ernster (1963a,b) and Ernster (Lee and Ernster, 1964) have shown that the transhydrogenase reaction in mitochondria is energetically coupled with the respiratory chain and ATPase: Respiration or the hydrolysis of ATP shifts the equilibrium of the transhydrogenase reaction toward the formation of NAD+ and NADPH. According to the chemiosmotic theory of Mitchell (1966), the mechanism of such energy coupling consists of utilization (by the transhydrogenase reaction) of the difference in electrochemical potential for H+ 383
Copyright Q 1982 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-153316-6
384
I. A. KOZLOV
ions (APH +), which is generated across the mitochondrial membrane by respiration or ATP hydrolysis, According to Mitchell, the equation of the reaction for mitochondrial transhydrogenase can be written NAD’
+ NADPH + 2HL
F’
NADH + NADP’
+ 2H&
That is, the oxidation of NADPH by NAD+ is coupled with the transfer of two H+ ions from the matrix to the extramitochondrial space, and the oxidation of NADH by NADP+ is coupled with the transport of H + ions in the opposite direction. Thus, the “forward” transhydrogenase reaction (with substrates NAD+ and NADPH) generates ApH+of the usual direction (electrically negative and alkaline inside the mitochondria), whereas the “reverse” transhydrogenase reaction (with substrates NADH and NADP+) produces AyH+ of the opposite sign. These elements of Mitchell’s theory have been confirmed experimentally (Dontsov et af., 1972; Grinius et af., 1970; Mitchell and Moyle, 1965; Moyle and Mitchell, 1963), and several schemes (Blazyk el af., 1976; Mitchell, 1972; Skulachev, 1970) have previously been proposed to describe the mechanisms of A,EH+ generation coupled with the transhydrogenase reaction. In this article a new general principle for transhydrogenase functioning is suggested (Kozlov, 1979). The mechanism proposed explains the way in which transfer of the hydride ion between nicotinamide nucleotides, in the course of the forward or reverse transhydrogenase reaction, ensures generation of APH+ of opposite directions across the mitochondrial membrane. Moreover, this new mechanism-unlike other schemes described in the literature-makes it possible to explain (without any additional postulates) why the simultaneous presence of substrates for both the forward reaction and the reverse reaction is not accompanied by uncoupling of the membrane due to H+ leakage through the transhydrogenase.
II. THE HYPOTHESIS ON THE MECHANISM OF ApH+ GENERATION BY THE TRANS. HYDROGENASE REACTION The main point in the mechanism proposed is that the reduced nicotinamide rings of NADH and NADPH facilitate H+ transfer, in a transhydrogenase molecule plugged through a coupling membrane, whereas oxidized rings block H+ transfer. Our reasoning is based on two facts: (1) Since the nicotinamide rings of NAD+ and NADP+ contain quaternary nitrogens, they are not capable of being proton acceptors or donors; and since they bear a positive charge,
21. MITOCHONDRIAL TRANSHYDROGENASE
385
they may be used to block the H+-transporting pathway in the transhydrogenase. (2) The reduced nicotinamide rings of NADH and NADPH, on the other hand, are good proton acceptors, and in a protonated form they can be donors of H+ ions (Jardetzky and Wade-Jardetzky, 1966). Thus, the nicotinamide rings of NADH and NADPH may participate in the relay transfer of H + ions. The scheme for generation of A&+ by the transhydrogenase is given in Fig. 1. It is supposed that the active site of the transhydrogenase is located in the hydrophobic phase of the mitochondrial membrane, near its inner surface. A proton-conducting pathway, which crosses a large part of the membrane's hydrophobic barrier, connects the active site of the transhydrogenase with the extramitochondrial space. The second, shorter, H+ pathway creates proton conductivity between the active site and the matrix. Two XH groups are located at the active site of the transhydrogenase between the two H+-conducting pathways but are separated from the H+-conducting pathways by regions of the active site that do not conduct H + ions.
FIG.1. The mechanism of A&+ generation across the mitochondrial membrane in the forward and reverse transhydrogenase reactions (Kozlov, 1979). The fragment of mitochondrial membrane is shown by diagonal shading. Horizontal shading indicates the transhydrogenase in the membrane. Proton-transporting pathways in the transhydrogenase are not shaded. NAD+ -ring, NADH-ring, NADP+ -ring, and NADPH-ring represent the hydride and H+ -exchanging rings of the corresponding nicotinamide dinucleotides. The numbers indicate the stages of the reaction. Details are given in the text.
386
I. A. KOZLOV
It is supposed that the region for binding the nicotinamide ring of NAD+ or (NADH) at the active site is located between the XH groups and the short H+-conducting pathway connected with the matrix. The region for binding the nicotinamide ring of NADP+ (or NADPH) is located between the XH groups and the long H+-conducting pathway connected with the extramitochondrial space. The first stage in the forward transhydrogenase reaction is binding of NAD+ by the transhydrogenase (Fig. 1). This binding, which is determined by the affinity of the transhydrogenase for the NAD+ molecule as a whole, leads to a drawing together of the positively charged nicotinamide ring of NAD+ and the XH groups of the active site. Immersion of the positive charge nearby results in lowering of the pK of the XH groups, which do not, however have any way of dissociating. Furthermore, location of the positively charged nicotinamide ring (of NAD+) between the XH groups and the short H+-conducting pathway blocks transfer of H + ions from the XH groups to the matrix. The second stage in the reaction is binding of NADPH at the active site of transhydrogenase. This comprises immersion of the reduced nicotinamide ring into the region of the active site between the XH groups and the long H+-conducting pathway. The reduced nicotinamide ring of NADPH can readily accept the H+ ions, thus creating proton conductivity between the H+-transporting pathway and the XH groups. Such proton conductivity makes possible the release of two H+ ions from the XH groups into the extramitochondrial space (stage 3). This H+ ion transfer produces a membrane potential which is positive on the outer side of the mitochondria1 membrane. At the next stage in the reaction, transfer of the hydride ion H- occurs from the reduced nicotinamide ring of NADPH to the oxidized nicotinamide ring of NAD+, resulting in the formation of NADH and NADP+. The positively charged ring of NADP+ prevents the return of the H+ ions from the extramitochondrial space to the X- groups of the active site. On the other hand, the reduced nicotinamide ring of NADH creates proton conductivity between the deprotonated X- groups and the H+-transporting pathway connected to the matrix. This allows the H + ions from the matrix to protonate the X- groups (stage 5). Protonation of the X- groups is accompanied by the dissociation of NADP+ into the matrix. At the last stage in the reaction, the release of NADH into the matrix occurs. Thus, the entire cycle is completed, and its outcome is the reduction of NAD+ by NADPH and the transfer of two H+ ions from the matrix to the extramitochondrial space. By reversing all stages of the forward reaction in the scheme proposed, we obtain the reverse transhydrogenase reaction. This reaction begins with the binding of NADH at the active site of the enzyme (Fig. 1). Binding of
21. MITOCHONDRIAL TRANSHYDROGENASE
387
NADP+ by the transhydrogenase, at the next stage in the reaction, leads to a drawing together of the positively charged nicotinamide ring of NADP+ and the XH groups of the active site. As a result, the pK of the XH groups decreases. Location of the positively charged nicotinamide ring of NADP+ between the XH groups and the long H+-conducting pathway blocks the release of protons from the XH groups into the extramitochondrial space. On the other hand, the reduced nicotinamide ring of NADH, located between the XH groups and the short H+-conducting pathway, allows protons to be released from the XH groups into the matrix. In the subsequent stage of the reaction, the transfer of a hydride ion Hfrom NADH to NADP+ takes place, giving NAD+ and NADPH. Now the positively charged nicotinamide ring of NAD+ prevents the return of the protons from the matrix to the X- groups, but the reduced nicotinamide ring of NADPH creates proton conductivity between the X- groups and the long H+-conducting pathway. This allows the X- groups to be protonated by H+ ions from the extramitochondrial space. This protonation is accompanied by the dissociation of NAD+ and NADPH to the matrix. If, as believed by Earle and Fisher (1980), only one proton is translocated through the membrane for each hydride ion equivalent transferred between the substrates, one (instead of two) XH group could be located in the active site of the transhydrogenase between the two H+-conducting pathways. The general scheme of ApH+ generation by the transhydrogenase reaction thereby remains the same.
111.
KNOWN FACTS AND FORECASTS
The scheme proposed for the mechanism of generation of A&+ by the transhydrogenase is in good agreement with a number of well-known facts regarding the functioning of this enzyme. 1. It has been shown that in the transhydrogenase reaction the formation of a triple complex of the enzyme with both substrates does occur (Rydstrom et af., 1971; Teixeira da Cruz et al., 1971). Transfer of the hydride ion H- between substrates is ensured by direct contact of the nicotinamide rings, NAD(H) and NADP(H) (Lee et af., 1965; Teixeira da Cruz et al., 1971). 2. The substrate-binding sites of the transhydrogenase exhibit a very high specificity. The site responsible for the binding of NAD+ and NADH does not bind either NADP+ or NADPH. On the other hand, the NADP+and NADPH-binding site does not bind either NAD+ or NADH (Rydstrom, 1972; Rydstrom et al., 1971; Teixeira da Cruz et al., 1971). In accordance with the data of Rydstrom, NAD+ is the first of the two
388
I. A. KOZLOV
substrates to be bound in the forward transhydrogenase reaction, and NADH is the first of the two substrates to be bound in the reverse transhydrogenase reaction. 3. From the data on the stereospecificity of the transhydrogenase reaction with regard to the H-4 of the nicotinamide ring (Lee et al., 1965), it may be thought that NAD(H) is bound by the enzyme with an anti orientation of the nicotinamide ring, and that NADP(H) is bound with a syn orientation of the ring (You et al., 1978). The scheme proposed for the transhydrogenase reaction (Fig. 1) presupposes a location of the two XH groups between the nicotinamide rings of NADP(H) and NAD(H). The exchange of H + ions between the XH groups and the reduced nicotinamide rings of NADH or NADPH will be possible if the distance from the XH This means groups to the N-1 in each of the nicotinamide rings is 1.5-2 that the distance between the N-1 atoms of the two nicotinamide rings should not be more than 3-4 A.When this fact is taken into account, and also the circumstance that the transfer of the hydride ion H- between the nicotinamide rings in the transhydrogenase reaction requires drawing together the C-4 atoms of both rings, it may be suggested that there is considerable overlapping of the planes of the rings at the active site of the transhydrogenase; i.e., a stacking structure is formed. The different (syn and anti) orientations of the nicotinamide rings, in which the amide groups are at the greatest distance from one another, is an important condition for the formation of such a stacking structure. Thus, the particular stereospecificity of the transhydrogenase reaction (Lee et af., 1965; You et al., 1978) can be explained within the framework of the scheme proposed. 4. Rydstrom has shown that the NADP(H)-binding site is located in a more hydrophobic environment than the NAD(H)-binding site (Rydstrom, 1972). In the scheme proposed (Fig. 1) the NADP(H)-binding site is deeper down in the hydrophobic layer of the membrane than the NAD(H)-binding site. According to the data of Fisher et al. (Jacobs and Fisher, 1979; McFadden and Fisher, 1978), the transhydrogenase from Rhodospiriflum rubrum (an enzyme carrying out the same functions in chromatophores as the mitochondrial transhydrogenase) consists of two peptide components: one tightly bound to the membrane, and the other water-soluble and easily detached from the membrane. It has been shown that the water-soluble peptide contains both NAD(H)- and NADP(H)-binding sites, while the membrane-bound peptide has NADP(H)-binding site but no NAD(H)binding site. The latter result indicates that the transhydrogenase contains two NADP(H)-binding sites: one (the active site) located in the depths of the membrane, and another (a superficial site) which may possibly participate in NADP(H) transfer from the matrix to the active site. A similar mechanism has been proposed for mitochondrial ATPase (Kozlov, 1975; Kozlov and Skulachev, 1977), in which the noncatalytic adenine nucleotide-
A.
389
21. MITOCHONDRIAL TRANSHYDROGENASE
binding site seems to participate in substrate transfer from the matrix to the active site of the ATPase. 5 . Translocation of NADP(H) from the matrix to the NADP(H)binding area of the active site is probably accompanied by a major conformational change in the protein molecule, since extensive changes in conformation of the transhydrogenase have been observed (Blazyk et al., 1976; O'Neal and Fisher, 1977) upon NADP+ or NADPH binding (but not upon NAD+ or NADH binding). 6. Energization of the mitochondrial membrane by ATP or the respiratory chain weakens the binding of NAD+ by the transhydrogenase (K, increases) (Rydstrom, 1977). This result is in good agreement with the scheme put forward in Fig. 1 , since that scheme postulates that NAD+ binding with the transhydrogenase leads to immersion of a positively charged nicotinamide ring into the hydrophobic layer of the membrane. Thus, if the membrane is energized (negative inside the mitochondrion), NAD+ binding must occur against the electrical gradient. Unlike NAD+, NADP+ binds better with transhydrogenase upon energization of the membrane (Rydstrbm, 1977), and this result is also readily accounted for within the framework of Fig. 1. According to this scheme NADP+ binding by the transhydrogenase is accompanied by immersion (in the hydrophobic phase of the membrane) of both the positively charged nicotinamide ring and the negatively charged pyrophosphate residue. The latter gives a total negative charge to the NADP+ molecule, which facilitates NADP+ translocation to the active site of transhydrogenase when the membrane is energized (positive outside the mitochondria). The suggestion of H -conducting pathways connecting the active site of the enzyme to the water phases has something in common with the ideas developed in the literature on the mechanism of Ap,+ generation by the other enzyme systems of the coupling membrane. Thus, Mitchell (1968) postulated the existence of such an H+-conducting pathway in the H -ATPase. According to the mechanism of H -ATPase functioning proposed by us (Kozlov, 1975; Kozlov and Skulachev, 1977), an H+-conducting pathway connects the extramitochondrial space with the active site of this enzyme, which is located in the depths of the mitochondrial membrane. Similar H+-conducting pathways have also been postulated and experimentally substantiated for bacteriorhodopsin (Kozlov and Skulachev, 1977; Skulachev, 1977). The idea that one of the transhydrogenase substrates is translocated into the depths of the membrane was first put forward by Skulachev in 1970. He later proposed a mechanism of A,%"+ generation by the transhydrogenase, in which oxidation of the nicotinamide ring of NADPH (by NAD+)creates +
+
+
390
I. A. KOZLOV
a local electric field and causes repulsion between the positively charged NADP+ and a cationic group of the enzyme. This repulsion results in conformational changes in the transhydrogenase, coupled to transport of H across the hydrophobic barrier of the membrane (generation of A j i H + ) (Skulachev, 1975). A new variation in the scheme given in this article consists of the idea that the reduced nicotinamide ring takes part in the relay transfer of protons across the membrane and that oxidation of the nicotinamide ring stops this relay transfer. Thus, the scheme in Fig. 1 provides an explanation, without additional postulates, of the generation of A j i H + with different signs for the forward and reverse reactions, as well as of failure of the transhydrogenase to uncouple the membrane (by proton leakage at some intermediate stage in the reaction). Positively charged nicotinamide rings of NAD+ or NADP+ would block passive H + conductance through the transhydrogenase at any stage of the reaction when NADH and NADP+ were simultaneously present in the system. (There is only one combination of substrates bound to the active site of the transhydrogenase that could lead to the passive H + conductance: if NADH and NADPH were to bind simultaneously, each having proper orientation with respect to the XH groups. But one superfluous hydrogen atom located in the region of contact of the nicotinamide rings could disturb this correct orientation.) No obvious solution to this problem had previously been found. The new scheme also shows a number of additional predictions which can be checked experimentally: +
1. NADP+ or NADPH binding with the transhydrogenase, in the absence of NADH and NAD+, should result in a small, temporary, reverse polarization of the membrane (negative outside the mitochondria), due to the fact that NADP(H), penetrating the hydrophobic layer of the membrane carries a net negative charge. NAD+ binding with the transhydrogenase, which is accompanied by the entry of the positively charged NAD+ ring into the membrane should (in the absence of NADPH) lead to a small, temporary polarization in the normal direction (negative inside). 2. According to the scheme proposed, when NADH is present at the active site of the transhydrogenase, the extent of the protonation of the XH groups depends on the pH of the matrix. In the presence of NADPH the extent of protonation of the XH groups depends on the pH of the extramitochondria1 space. This means that in the presence of NADPH the rate of NAD+ release from the active site of the transhydrogenase should increase as the pH falls in the extramitochondrial space. Increased pH in the matrix should result in increased affinity of the transhydrogenase for NADP in the presence of NADH. 3. The scheme proposed envisages protonation of the nicotinamide +
21.
M ITOCHONDRIAL TRANSHYDROGENASE
391
rings of NADH and NADPH at intermediate stages in the transhydrogenase reaction. Therefore, NADH and NADPH analogs with lowered H + acceptor properties, or NAD+ and NADP+ analogs capable of accepting the H + ions, should be unable to generate AFH+ in the transhydrogenase reaction. 4. The membrane-bound fragment (of the transhydrogenase molecule), which is presumed to form the H+-conducting pathway, may by itself create proton conductivity across the membrane. NADPH binding by this fragment should facilitate H +-conductivity, whereas NADP + binding should block H + conductivity.
IV.
CONCLUSION
Unlike other energy-transducing enzymes of the coupling membranes, mitochondria1 transhydrogenase generates AiH + of opposite directions, depending on the concentrations of its substrates NAD' , NADH, NADP , and NADPH. The mechanism of transhydrogenase action suggested in this article explains (1) how the direction of the transhydrogenase reaction determines the sign of AFH+ generated across the membrane and (2) why both forward and reverse transhydrogenase reactions are not accompanied by uncoupling of the membrane due to H + leakage through the transhydrogenase. It is proposed for the transhydrogenase reaction that the reduced nicotinamide rings of NADH or NADPH take part in a relay transfer of H + ions across the enzyme molecule spanning the membrane, whereas the oxidized nicotinimide rings of NAD+ or NADP+ block this H -conducting pathway. +
+
ACKNOWLEDGMENTS I would like to express my sincere gratitude to Professor V. P . Skulachev for his contribution to the development of the ideas formulated in this work. I would also like to thank Drs. A. N. Glagolev, A. A. Kondrashin, A. A. Konstantinov, Ya. M. Milgrom, B. V. Chernyak, and V. Ya. Chernyak for discussion of the article and valuable criticism. I would like to thank my wife Glenys for translation and preparation of the manuscript. REFERENCES Blazyk, J. F., Lam, D., and Fisher, R. R. (1976). Biochemistry, 15, 2843-2848. Danielson, L., and Ernster, L. (1963a). Biochem. Biophys. Res. Commun. 10, 91-96. Danielson, L., and Ernster, L. (196313). Biochem. Z. 338, 188-205. Dontsov, A. E., Grinius, L. L., Jasaitis, A. A., Severina, I. I., and Skulachev, V. P. (1972). J. Bioenerg. 3, 277-303. Earle, S. R., and Fisher, R. R. (1980). J . Biol. Chem. 255, 827-830. Grinius, L. L., Jasaitis, A. A., Kadziauskas, J. P., Liberman, E. A., Skulachev, V. P.,
392
I. A. KOZLOV
Topali, V. P., Tsofina, L. M., and Vladimirova, M. A. (1970). Biochim. Biophys. Acfa 216, 1-12. Jacobs, E., and Fisher, R. R. (1979). Biochemistry 18, 4315-4322. Jardetzky, O., and Wade-Jardetzky, N. G. (1966). J. Biol. Chem. 241, 85-91. Kaplan, N. O., Colowick, S. P., and Neufeld, E. F. (1953). J. B i d . Chem. 205, 1-15. Kaplan, N. O., Swartz, M. N., Frech, M. E., and Ciotti, M. M. (1956). Proc. Nail. Acad. Sci. U.S.A. 42, 481-487. Kozlov, I. A., (1975). Bioorg. Khim. 1, 1545-1569. Kozlov, I. A. (1979). Biokhimiya 44, 1731-1737. Kozlov, I . A,, and Skulachev, V. P. (1977). Biochim. Biophys. Acta 463, 29-89. Lee, C. P., and Ernster, L. (1964). Biochim. Biophys. Acfa 81, 187-190. Lee, C. P., Simarol-Diequesne, N., Ernster, L., and Hoberman, H. D. (1965). Biochim. Biophys. Actn 105, 397-409. McFadden, B. J., and Fisher, R. R. (1978). Arch. Biochem. Biophys. 190, 820-828. Mitchell, P. (1966). “Chemiosmotic Coupling in Oxidative and Photosynthetic Phosphorylation.” Glynn Research, Bodmin, England. Mitchell, P. (1968). “Chemiosmotic Coupling and Energy Transduction.” Glynn Research, Bodmin, England. Mitchell, P. (1972). J. Bioenerg. 3, 5-24. Mitchell, P., and Moyle, J. (1965). Nature (London) 208, 1205-1206. Moyle, J., and Mitchell, P. (1973). Biochem. J. 132, 571-585. O’Neal, S. G., and Fisher, R. R. (1977). J . Biol. Chem. 252, 4552-4556. Rydstrom, J. (1972). Eur. J. Biochern. 31, 496-504. Rydstrom, J. (1977). Biochim. Biophys. Acta 463, 155-184. Rydstrom, J., Teixeira da Cruz, A., and Ernster, L. (1971). Eur. J. Biochem. 23, 212-219. Skulachev, V. P. (1970). FEES Lett. 11, 301-305. Skulachev, V . P . (1972). “Energy Transformation in Biomembranes.” Nauka Press, Moscow. Skulachev, V. P. (1975). In “Energy Transducing Mechanisms” (E. Racker, ed.), pp. 31-73. MTI. Skulachev, V. P. (1977). Proc. FEES Meet. 11th 45, 49-59. Teixeira da Cruz, A., Rydstrom, J., and Ernster, L. (1971). Eur. J . Biochem. 23, 210-211. You, K., Arnold, L. J., Jr., Allison, W. S., and Kaplan, N. 0. (1978). TIES 3, 265-268.
CURRENT TOPICS IN MEMBRANES AND TRANSPORT, VOLUME 16
Chapter 22 Membrane Vesicles, Electrochemical Ion Gradients, and Active Transport H . R . KABACK Laboratory of Membrane Biochemistry Roche Institute of Molecular Biology Nutley, New Jersey
Introduction ........................................................................................ Molecular Architecture of Escherichia coli Membrane Vesicles ........................ 111. Chemiosmotic Phenomena ...................................................................... A. The Proton Electrochemical Gradient and Active Transport ..................... B. Proton-Dependent Transport ............................................................ C. Sodium-Dependent Transport ........................................................... IV. Carrier Action ..................................................................................... A. Mechanistic Studies ........................................................................ B. Chemical Modification of Transport Activity ........................................ References .......................................................................................... I.
11.
I.
393 394 395 395 397 398 399 399 40 1 402
INTRODUCTION'
Plasma membrane vesicles isolated from bacteria (Kaback, 1970, 1971, 1974, 1976) and, more recently, eukaryotic cells (cf. Lever, 1979, for a recent review) provide important systems for studying certain aspects of transport. These vesicles are devoid of cytoplasm, and their metabolic IThe following abbreviations are used in this article: A&+, the proton electrochemical gradient; A$, membrane potential; ApH, pH gradient; TPP' , tetraphenylphosphonium (bromide salt); TPMP+ , triphenylmethylphosphonium (bromide salt). TMG, methyl- 1-thio-(3-D-galactopyranoside (TMG); DEPC, diethylpyrocarbonate. A$ and ApH can be varied reciprocally with little or no change in ApHt (Ramos et a/., 1976; Ramos and Kaback, 1977a; Tokuda and Kaback, 1977).
393
Copyright 0 1982 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-1 53316-6
394
H. R. KABACK
activities are restricted to those provided by the enzymes of the membrane itself. Thus, uphill transport by membrane vesicles per se is practically nil, and the driving force for the accumulation of a particular solute can be determined by studying which compounds or experimental manipulations stimulate accumulation. Moreover, the metabolism of the transported solute and the energy source is minimal, allowing a clear definition of the reactions involved. Finally, removal or disruption of the cell wall in certain bacteria allows the use of probes, inhibitors, and ionophores to which the plasma membrane is normally inaccessible. This article focuses mainly on phenomena related to respiration-linked active transport in membrane vesicles from Escherichia coli. Despite the emphasis here on a single membrane system, it should be stressed that the results are likely to be relevant to active transport and to energy transduction more generally. The first sections of the article review the structure of E. coli vesicle membranes and the organization of several distinct transport processes in these membranes. The final section treats more mechanistic questions of carrier symmetry, reactivity, and electric field dependence. II. MOLECULAR ARCHITECTURE OF Escherichia coli MEMBRANE VESICLES
Despite a considerable body of evidence (cf. Stroobant and Kaback, 1975, for a review) demonstrating that bacterial membrane vesicles prepared by osmotic lysis retain the same configuration as the membrane in the intact cell, this contention has been controversial. Recent studies (Owen and Kaback, 1978, 1979a,b) carried out collaboratively with Peter Owen of Trinity College in Dublin, Ireland, provide virtually incontrovertible support for this argument, however. In these experiments, the architecture of E. coli vesicles was studied using crossed immunoelectrophoresis, and a reference pattern of 52 discrete immunoprecipitates was established. Progressive immunoabsorption experiments conducted with untreated control vesicles and with physically disrupted vesicles demonstrate that the membrane-associated immunogens fall into two categories: (1) those whose expression is minimal unless the vesicles are disrupted, typified by D-lactate dehydrogenase ATPase, NADH dehydrogenase (two immunologically distinct immunoprecipitates stain for NADH dehydrogenase), dihydroorotate dehydrogenase, 6-phosphogluconate dehydrogenase, polynucleotide phosphorylase, and 0-galactosidase, and (2) immunogens that are expressed to similar extents in untreated and in disrupted vesicles, such as Braun’s lipoprotein (an outer membrane component) and three unidentified immunogens. A mathematical relationship between the peak area sub-
22. VESICLES, ION GRADIENTS, AND ACTIVE TRANSPORT
395
tended by an immunoprecipitate in the crossed immunoelectrophoresis system and the quantity of vesicles used in the absorption process has been derived, and the relationship allows quantitation of the degree to which specific membrane immunogens partition between exposed and unexposed surfaces of the vesicle membrane. The results demonstrate that at least 98% of the vesicles in the preparation are in the form of sealed sacculi with the same polarity as the intact cell. In addition, the findings provide a strong indication that dislocation of immunogens from the inner to the outer surface of the membrane during vesicle preparation does not occur to an extent exceeding 10%.
111.
CHEMIOSMOTIC PHENOMENA
A. The Proton Electrochemical Gradient and Active Transport
Chemiosmotic phenomena, as postulated by Mitchell (1961, 1966, 1968, 1973, 1979), play a central, obligatory role in the energetics and mechanism of active transport in the vesicle system (Kaback, 1976; Harold, 1976; Konings and Boonstra, 1977). Accordingly, it has been demonstrated that oxidation of certain electron donors via the membrane-bound respiratory chain is accompanied by the expulsion of protons into the external medium, leading to the development of an electrochemical gradient of protons (Ap,+) that is the immediate driving force for active transport. A & + is composed of electrical and chemical parameters according to the following relationship: A P H +=A$-(2.3 RT/F)ApH
(1)
where A$ represents the electrical potential across the membrane and ApH is the chemical difference in proton concentration across the membrane (2.3RT/F is equal to 58.8 mV at room temperature). Measurement of ApH and A$ in microscopic systems is frequently based on determination of the equilibrium distribution of permeant weak acids or bases and permeant lipophilic ions, respectively (Rottenberg, 1975, 1979). Traditionally, the internal concentration of the permeant species is measured after separation of the cells or membrane vesicles from the bathing medium by filtration or centrifugation techniques. However, the manipulations involved in the separation step may result in significant and sometimes complete loss of accumulation solute, leading to an underestimate of the concentration gradient and thus to artifactually low values of ApH+.Recently (Ramos et al., 1976), this problem has been resolved
396
H. R.
KABACK
through the use of flow dialysis, a technique devised originally to measure ligand binding to enzymes (Colowick and Womack, 1969). The technique allows continuous monitoring of the external concentration of any dialyzable solute under conditions that require no manipulation of the experimental system (Ramos et al., 1979a), and over the past few years, it has been utilized to measure A&+ and steady state levels of accumulation of a variety of transport substrates in a number of systems. Strong independent support for the quantitative validity of the measurements has been obtained recently. Electrophysiological techniques have been applied to E. coli giant cells induced by growth in 6-amidinopenicillanic acid (Felle et al., 1978), providing for the first time a means of verifying quantitatively results obtained with lipophilic ions. Using E. coli W1485 giant cells and recording intracellularly, Felle et al. (1980; Porter et al., 1979) has measured a A$ (interior negative) that varies directly with external pH, increasing from - 100 mV at pH 5.5 to - 141 mV at pH 8.0. Between pH 5.5 and 7.0, the increase is linear with a slope of - 22 mV/pH unit. With the lipophilic cations rH]TPP+ and PH]TPMP+,the values are very similar, increasing from about - 104 mV at pH 5.0 to - 151 mV at p H 8.0, with a slope of - 23 mV/pH unit between pH 5.0 and 7.0. In addition, it has been demonstrated that the distribution of T P P + can be used to determine A$ in cultured neuroblastoma/glioma NG108-15 hybrid cells (Lichtshtein et al., 1979a,b), isolated guinea pig synaptosomes (Ramos et al., 1979b), splenic lymphocytes (Kiefer et al., 1980), and chick embryo heart cells (Elsas et al., 1981). In a number of bacterial systems, it has been demonstrated that ApH is extremely sensitive to external p H (Ramos et al., 1976; Ramos and Kaback, 1977a; Padan et al., 1976; Zilberstein et al., 1979; Bakker et al., 1976; Tokuda and Kaback, 1977; Guffanti et al., 1978; Krulwich et al., 1978; Friedberg and Kaback, 1980). With intact E. coli (Padan et al., 1976; Zilberstein et al., 1979) and right-side-out vesicles from this organism (Ramos et al., 1976; Ramos and Kaback, 1977a) and Salmonella typhimurium (Tokuda and Kaback, 1977), in particular, ApH exhibits maximal values of about 2 pH units (i.e., - 120 mV) at pH 5.5 and decreases to zero at about pH 7.5. Corroborative support for these measurements in E. coli has come from high-resolution 31P nuclear magnetic resonance spectroscopy (Navon et al., 1977; Ogawa et al., 1978) which yields results that are very similar both qualitatively and quantitatively to those obtained from distribution studies with permeant weak acids. Inverted membrane vesicles from E. coli generate a transmembrane ApH+ that is of similar magnitude but opposite in polarity to that observed in right-side-out vesicles during oxidation of D-lactate, succinate, reduced phenazine methosulfate, or NADH or hydrolysis of ATP (Reenstra et al.,
397
22. VESICLES, ION GRADIENTS, AND ACTIVE TRANSPORT
1980). Using the distribution of the lipophilic anion thiocyanate to measure A$ (interior positive) and the permeant weak base methylamine to measure ApH (interior acid), maximal values for ApH+of approximately 160 mV are obtained. Many of the properties of ApH+ in inverted vesicles are similar to those described in right-side-out vesicles: (1) The magnitude of the A$ generated in the presence of D-lactate or reduced phenazine methosulfate is similar to that observed in right-side-out vesicles but of opposite polarity and independent of pH from 5.5 to 8.0; (2) plots of ApH versus internal pH in inverted vesicles and external pH in right-side-out vesicles are similar with D-lactate as electron donor; (3) as observed with right-side-out vesicles, dissipation of A$ or ApH leads to a concomitant increase in the other parameter without a change in the rate of respiration; (4)inverted vesicles catalyze Na+ accumulation, and it is apparent that the process can be driven by either A$ (interior positive) or ApH (interior acid).
+
6. Proton-Dependent Transport
Transport of substrates such as lactose or glucose 6-phosphate which are accumulated in relatively large amounts by the appropriate right-side-out vesicles causes partial collapse of A$ (Schuldiner and Kaback, 1975) and/or ApH (Ramos and Kaback, 1977b), providing direct support for the argument that ApH+ drives solute accumulation via coupled movements with protons. Titration studies with the ionophores valinomycin and nigericin demonstrate that A$ and ApH can be varied reciprocally with little or no change in ApH+(Ramos et al., 1976; Ramos and Kaback, 1977a; Tokuda and Kaback, 1977). Analogous studies with various solutes indicate that, at pH 5.5, there are two classes of transport systems (Ramos and Kaback, 1977b): those driven primarily by A,iiH+ (lactose, proline, serine, glycine, tyrosine, glutamate, leucine, lysine, cysteine, and succinate) and those driven primarily by ApH (glucose 6-phosphate, lactate, glucuronate, gluconate, inorganic phosphate; Patel and Kaback, unpublished experiments) and, in S. typhimurium vesicles, citrate (Patel and Kaback, unpublished experiments). Strikingly, however, at pH 7.5 and above, all these transport systems are driven by A$ which is the only component of A j i H + . In addition, when the steady state level of accumulation of transport substrates is examined as a function of pH, none of the profiles corresponds to those described for ApH+,ApH, or A$. Furthermore, at external pH values exceeding 5.5-6.0, ApH+ is insufficient apparently to account thermodynamically for the concentration gradients observed for most of the substrates if the stoichiometry between protons and substrates is 1:l (Ramos and Kaback, 1977b). This finding and the observation that the
398
H. R. KABACK
accumulation of certain organic acids is in equilibrium with ApH at acid pH and with A$ at alkaline pH, where ApH is absent, led to the suggestion that the stoichiometry between protons and transport substrates may vary as a function of external pH, and more direct evidence supporting this hypothesis was presented (Ramos and Kaback, 1977~). Recent studies with intact cells (Zilberstein et al., 1979; Booth et al., 1979; Felle el al., 1980), however, have cast doubt on the contention that there is a discrepancy between the steady state level of lactose accumulation and Ap,+ at alkaline pH. Specifically, it has been demonstrated that in intact E. coli, as opposed to membrane vesicles, A$ increases markedly with pH in such a manner as to compensate for the decrease in ApH. Thus, ApH+in intact cells does not decrease as drastically with increasing pH as observed in vesicles, and the steady state level of lactose accumulation at high pH can be accommodated without a change in proton/lactose stoichiometry. In addition, more direct studies on proton-lactose symport in deenergized cells are not indicative of a change in stoichiometry at high pH (Zilberstein et al., 1979; Booth et al., 1979). On the other hand, numerous studies with both intact E. coli (Felle et al., 1980; Padan et al., 1976; Zilberstein et al., 1979; Navon et al., 1977; Ogawa et al., 1978) and isolated membrane vesicles (Ramos et al., 1976; Ramos and Kaback, 1977a,b) demonstrate that ApH is absent at pH 7.5 and above. Thus, it is difficult to explain how the transport of certain organic acids can be coupled to ApH at acid pH and to A$ at alkaline pH without invoking a pH-dependent increase in protonlsubstrate stoichiometry (Ramos and Kaback, 1977b,c; Rottenberg, 1976), and direct measurements in both intact cells (Taylor and Essenberg, 1979) and membrane vesicles (LeBlanc et al., 1979) supporting this notion have been presented recently.
C. Sodium-Dependent Transport One attractive conceptual aspect of the chemiosmotic hypothesis for bacterial active transport is its analogy to the mechanism suggested for sugar and amino acid transport in many eukaryotic cells (Crane, 1977). In these systems, an electrochemical gradient of sodium rather than protons is generated through the action of the membraneous sodium-, potassiumdependent ATPase, and accumulation of sugars and amino acids occurs via coupled movements with sodium (this process is referred to traditionally as cotransport rather than symport). Although it is certain that many bacterial transport systems catalyze proton-substrate symport, several instances have been reported in which the
22. VESICLES, ION GRADIENTS, AND ACTIVE TRANSPORT
399
transport of a specific solute is dependent upon the presence of sodium or lithium ion (see Tokuda and Kaback, 1977, for a review). Moreover, some of these studies, in particular those of Stock and Roseman (1977) and Lanyi et al. (1976), indicate that symport or cotransport mechanisms may be operative. Since the basic energy-yielding process in bacteria is thought to be proton extrusion and bacteria apparently do not possess a sodium-, potassium-dependent ATPase or a primary sodium pump, the existence of such transport systems presents certain obvious problems, such as the relationship between A&+ and these transport systems and the mechanism by which the internal sodium concentration is maintained at a low level. Membrane vesicles isolated from S. typhimurium G-30 grown in the presence of melibiose catalyze TMG transport in the presence of sodium or lithium (Tokuda and Kaback, 1977). TMG-dependent sodium uptake is also observed, but only when a potassium diffusion potential (interior negative) is induced across the vesicle membrane. Cation-dependent TMG accumulation varies with the ApH+ generated as a result of D-lactate oxidation, and the vesicles catalyze D-lactate-dependent sodium efflux in a manner consistent with the operation of a proton-sodium exchange mechanism. The results are consistent with a model in which TMG-sodium (lithium) symport is driven by A f i H + ,which functions to maintain low intravesicular sodium and lithium concentrations through proton-sodium (lithium) antiport. A similar mechanism has been suggested for lightdependent glutamate transport in vesicles from Halobacterium halobium (Lanyi et al., 1976). IV.
CARRIER ACTION
A. Mechanistic Studies
Passive, carrier-mediated lactose efflux has proven to be a useful device for probing the mechanism of &galactose translocation in E. coli membrane vesicles, particularly with respect to lactose-proton symport (Kaczorowski and Kaback, 1979; Kaczorowski et al., 1979). When vesicles are equilibrated with high concentrations of [14C]lactoseand diluted 200-fold into media devoid of galactosides, efflux is first-order and the maximal rate is pH-dependent, increasing about threefold from pH 5.5 to pH 7.5 (tYi=45, 27, and 15 seconds at pH 5.5, 6.6, and 7.5, respectively). In contrast, experiments performed under identical conditions with equimolar lactose in the external medium (i.e., under exchange conditions) demonstrate that the exchange reaction is insensitive to pH and very fast relative to efflux (t,<2 seconds). Proton symport occurs during lactose efflux,
400
H. R. KABACK
resulting in the transient formation of a membrane potential (A$, interior negative), as demonstrated by efflux-dependent accumulation of rubidium (in the presence of valinomycin) and active transport of proline, both of which are abolished by the protonophore carbonylcyanide-m-chlorophenylhydrazone. Moreover, the magnitude of the A$ generated increases with pH in much the same manner as the rate of lactose efflux, suggesting tight coupling between the processes. Comparison of the efflux and exchange reactions suggests that the rate-determining step for efflux involves return of the unloaded carrier to the inner surface of the membrane, and that either loss of the symported proton from the carrier or translocation of the unloaded carrier may be limiting. Counterflow experiments conducted at various pH values reveal that external lactose affects proton loss from the carrier. When external lactose is present at concentrations below the apparent K, of the carrier, counterflow is pH-dependent and decreases from pH 5.5 to 7.5, indicating that deprotonation of the carrier occurs frequently under these conditions to limit the counterflow process. When the external lactose concentration is saturating, however, counterflow is unaffected by pH. Moreover, the transient formation of A$ observed during lactose efflux is abolished under these conditions. The observations are consistent with an ordered mechanism for efflux whereby lactose is released first, followed by loss of a proton. In addition, the data suggest that the loaded carrier recycles in the protonated form during counterflow and exchange. Imposition of a membrane potential (A$, interior negative) or a pH gradient (ApH, interior alkaline) across the membrane leads to a marked, transient increase in the fluorescence of 6’ -(N-dansy1)aminohexyl-1-thio-b-Dgalactopyranoside. The maximum increase in fluorescence appears to be a linear function of the magnitude of the imposed A$ or ApH, and the effect of each parameter is additive. Imposition of A$ or ApH also alters the rate of carrier-mediated lactose efflux from the intravesicular pool, and the effects are dependent upon the polarity of the imposed A$ or ApH. The rate of efflux is diminished when the vesicle interior is made electrically negative or alkaline and is enhanced when the vesicle interior is made electrically positive or acid. Here again, the effects of A$ and ApH are additive, and kinetic experiments demonstrate further that A$ and ApH alter the maximal velocity of efflux without significant effect on the apparent K, of the process. Strikingly, imposed A$, ApH, or A$+ApH-of either polarity-has no effect whatsoever on the exchange flux. The data support the idea that the rate-limiting step for carrier-mediated lactose efflux down a concentration gradient involves either deprotonation of the carrier or translocation of the unprotonated carrier to the inner surface of the membrane. In addition, since efflux but not exchange is influenced by the
22. VESICLES, ION GRADIENTS, AND ACTIVE TRANSPORT
401
imposed gradients, it seems likely that the loaded carrier (i.e., the ternary complex between the carrier, protons, and lactose) is neutral, while the unloaded carrier is negatively charged. In this context, it is interesting that recent experiments with the Na+-dependent melibiose transport system (Cohn and Kaback, 1981) suggest that the loaded carrier may be positively charged rather than neutral. Thus, in this case, rates of Na+-dependent efflux, as well as exchange, are influenced by the imposition of All., ApH, or All. + ApH. Moreover, as opposed to the lac system, where the imposition of A p H + (interior negative and alkaline) leads to a dramatic decrease in the apparent K, for the influx with a small increase in V,,, with the Na+dependent melibiose transport system, the imposition of A p H +has no effect on the apparent K , for influx but causes a marked increase in V,,,. Given the suggestion that the rate-limiting step for lactose efflux may be deprotonation of the carrier on the external surface of the membrane, one means of further investigating the mechanism is to search for a solvent deuterium isotope effect. Preliminary experiments (Kaczorowski and Kaback, unpublished experiments) have yielded promising results in this regard. At equivalent pH and pD (i.e., p D = p H +0.4),the rate of lactose efflux is approximately 2-2.5 times slower in deuterated media (with over 95% of the H 2 0 replaced by 4 0 ) relative to control conditions in H 2 0 , while the rate of exchange is identical in the presence of D 2 0 and H 2 0 . Furthermore, during counterflow with external lactose concentrations below the apparent high-affinity K, of the carrier, the magnitude of the overshoot is greater in 4 0 than in H 2 0 .
B. Chemical Modification of Transport Activity Although it is clear from the chemiosmotic hypothesis that proton symport and hydroxide antiport are thermodynamically equivalent and cannot be differentiated energetically, the functional groups responsible for proton binding and translocation would be expected to be radically different from those involved in hydroxide binding and translocation. With a view toward differentiating between these alternatives, a study on the effects of various group-specific reagents on the transport activity of E. coli membrane vesicles was undertaken, and it soon became apparent that reagents known to modify histidine residues produce interesting effects on ApH+-dependent transport in the vesicle system (Padan el a/., 1979). Exposure of E. coli membrane vesicles to the histidine-specific reagent diethylpyrocarbonate (DEPC), at pH 6.0, leads to concentration- and timedependent inactivation of each respiration-linked transport system, and the
402
H. R. KABACK
sensitivity to inactivation is enhanced when A&, + (interior negative and alkaline) is generated across the vesicle membrane. Although P-D-galactopyranosyl-1-thio-0-D-galactopyranoside, a high-affinity substrate of the lac transport system, and other carrier-specific substrates for other transport systems block DEPC inactivation, binding of p-nitropheynl-a-Dgalactopyranoside to the lac carrier is not significantly altered, indicating that DEPC does not react at the binding site of the lac carrier and presumably of other carriers. Strikingly, vesicles treated with DEPC exhibit an increased apparent K, for A&+ -driven transport and counterflow but no change in the V,,, of these reactions and no change in the apparent K, or V,,, of lactose-facilitated diffusion. Moreover, alkalinization of the external medium induced by the addition of lactose to membrane vesicles under nonenergized conditions is progressively blocked by the same concentrations of DEPC that alter the apparent K, for Ap,,.-driven active transport (Patel, Garcia, and Kaback, unpublished experiments). Although it cannot be stated categorically that the effects of DEPC are due specifically to modification of membrane histidyl residues, a number of indirect lines of evidence support this contention: (1) DEPC treatment at pH 6.0 specifically modifies histidyl residues in a number of soluble proteins (Heinridson and Kramer, 1974; Miles, 1977); (2) hydroxylamine displaces the ethoxycarbonyl moiety from the imidazole nitrogen of histidine (Heinridson and Kramer, 1974; Miles, 1977) and regenerates counterflow activity in DEPC-treated vesicles; (3) exposure of the vesicles to light in the presence of rose bengal, an operation that leads to photooxidation of histidyl residues, causes effects that are virtually identical to those observed with DEPC; and (4) when first-order rates of inactivation of lactose counterflow are studied as a function of pH during treatment with DEPC or rose bengal photooxidation, the curves generated are remarkably similar to those observed for the chemical modification of histidine (Garcia, Patel, and Kaback, unpublished experiments). Taken as a whole, the results suggest that a histidyl residue(s) in different carriers or in another protein in the translocation complexes (Hong, 1977) is involved either in the binding and translocation of protons or in a conformational transition that may occur upon protonation of various carrier proteins.
REFERENCES Bakker, E. P., Rottenberg, H . , and Caplan, S. R. (1976). Biochim. Biophys. Acta 40, 557. Booth, I. R., Mitchell, W. J., and Hamilton, W. A. (1979). Biochem. J. 182, 687. Cohn, D. E., and Kaback, X. R. (1980). Biochemistry 19, 4237. Colowick, S., and Womack, F. C. (1969). J. Biol. Chem. 244, 774.
22. VESICLES, ION GRADIENTS, AND ACTIVE TRANSPORT
403
Crane, R. K. (1977). “Reviews on Physiology, Biochemistry and Pharmacology.” SpringerVerlag, Berlin and New York. Elsas, L. J., Krick, J. H., Serravezzo, J., and Detlaan, R. L. (1981). In preparation. Felle, H., Stetson, D. L., Long, W. S., and Slayman, C. L. (1978). In “Frontiers of Biological Energetics” (P. L. Dutton, J. J. Leigh, and A. Scarpa, eds.), p. 1399. Academic Press, New York. Felle, H., Porter, J. S., Slayman, C. L., and Kaback, H. R. (1980). Biochemistry 19, 3585. Friedberg, I., and Kaback, H. R. (1980). J. Bacteriol. 142, 651. Guffanti, A. A., Susman, P., Blanco, R., and Krulwich, T. A. (1978). J. Biol. Chem. 253, 708. Harold, F. M. (1976). Curr. Top. Bioenerg. 6, 83. Heinridson, R. L., and Kramer, K. J. (1974). In “Progress in Bioorganic Chemistry” (E. T. Kaiser and F. J. Kezdy, eds.), Vol. 3, p. 141. Wiley, New York. Hong, J. - S . (1977). J. Biol. Chem. 252, 8582. Kaback, H. R. (1970). Annu. Rev. Biochem. 39, 561. Kaback, H. R. (1971). Methods Enzymol. 22, 99. Kaback, H. R. (1974). Science 186, 882. Kaback, H. R. (1976). J. Cell. Physiol. 89, 575. Kaczorowski, G. J., and Kaback, H. R. (1979). Biochemistry 18, 3691. Kaczorowski, G. J., Robertson, D. E., and Kaback, H. R. (1979). Biochemistry 18, 3967. Kiefer, H., Blume, A. J., and Kaback, H. R. (1980). Proc. Nutl. Acad. Sci. U.S.A. 77, 2200. Konings, W. N., and Boonstra, J. (1977). Curr. Top. Membr. Tramp. 9, 177. Krulwich, T. A., Davidson, L. F., Filip, S. J., Zuckerman, R. S., and Guffanti, A. A. (1978). J. Biol. Chem. 253, 4599. Lanyi, J., Renthal, R., and MacDonald, R. I. (1976). Biochemistry 15, 1603. LeBlanc, G., Rimon, G., and Kaback, H. R. (1979). Biochemistry 19, 2522. Lever, J . E. (1979). CRC Crit. Rev. Biochem. 7, 187. Lichtshtein, D., Kaback, H. R., and Blume, A. J. (1979a). Proc. Natl. Acad. Sci. U.S.A. 76, 650. Lichtshtein, D., Dunlop, K., Kaback, H. R., and Blume, A. J. (1979b). Proc. Natl. Acad. Sci. U.S.A. 76, 2580. Miles, E. W. (1977). Methods Enzymol. 47, 431. Mitchell, P. (1961). Nature (London) 191, 144. Mitchell, P. (1966). Biol. Rev. Cambridge Philos. SOC.41, 445. Mitchell, P. (1968). “Chemiosmotic Coupling in Oxidative and Photosynthetic Phosphorylation.” Glynn Research, Bodmin, England. Mitchell, P. (1973). J. Bioenerg. 4, 63. Mitchell, P. (1979). Eur. J. Biochem. 95, 1. Navon, G., Ogawa, S., Shulman, R. B., and Yamane, T. (1977). Proc. Null. Acad. Sci. U.S.A. 74, 888. Ogawa, S., Shulman, R. B., Glynn, P., Yamane, T., and Navon, G. (1978). Biochim. Biophys. Acta. 502, 45. Owen, P., and Kaback, H. R. (1978). Proc. Natl. Acad. Sci. U.S.A. 75, 3148. Owen, P., and Kaback, H. R. (1979a). Biochemistry 18, 1413. Owen, P., and Kaback, H. R. (197913). Biochemistry 18, 1422. Padan, E., Zilberstein, D., and Rottenberg, H. (1976). Eur. J. Biochem. 63, 533. Padan, E., Patel, L., and Kaback, H. R. (1979). Proc. Natl. Acad. Sci. U.S.A. 76, 6221. Porter, J. S., Slayman, C. L., Kaback, H. R., and Felle, H. (1979). Annu. Meet. A m . SOC. Microbiol., 79th 145, (Abstr. K2). Ramos, S., and Kaback, H. R. (1977a). Biochemistry 16, 848. Ramos, S., and Kaback, H. R. (1977b). Biochemistry 16, 854.
404
H. R. KABACK
Ramos, S., and Kaback, H. R. (1977~).Biochemistry 16, 4271. Ramos, S., Schuldiner, S., and Kaback, H. R. (1976). Proc. Natl. Acad. Sci. U.S.A. 73,1892. Ramos, S., Schuldiner, S., and Kaback, H. R. (1979a). Methods Enzymol. 54, 680. Ramos, S., Grollman, E. F., Lazo, P. S., Dyer, S. A., Habig, W. H., Hardegree, M. C., Kaback, H. R., and Kohn, L D. (1979b) Proc. Natl. Acad. Sci. U.S.A. 16, 4783. Reenstra, W. W., Patel, L., Rottenberg, H., and Kaback, H. R. (1980). Biochemistry 19, 1. Rottenberg, H. (1975). J. Bioenerg. 7, 61. Rottenberg, H. (1976). FEBS Lett. 66, 159. Rottenberg, H. (1979). Methods Enzymol. 54, 547. Schuldiner, S., and Kaback, H. R. (1975). Biochemistry 14, 5451. Stock, J., and Roseman, S. (1971). Biochem. Biophys. Res. Commun. 44, 132. Stroobant, P., and Kaback, H . R. (1975). Proc. Natl. Acad. Sci. U.S.A. 72, 3970. Taylor, D. J., and Essenberg, R. C. (1979). Int. Congr. Biochem., l l t h , Toronto Abstract, p. 460. Tokuda, H., and Kaback, H. R. (1977). Biochemistry 16, 669. Zilberstein, D., Schuldiner, S., and Padan, E. (1979). Biochemistry 18, 669.
Part VI
Biological Significance of Electrogenic Ion Pumps
This Page Intentionally Left Blank
CURRENT TOPICS IN MEMBRANES AND TRANSPORT, VOLUME 16
Chapter 23 The Role of Electrogenic Proton Translocation in Mitochondrial Oxidative Phosphorylation JANNA P . WEHRLEI Department of Physiological Chemistry The Johns Hopkins University School of Medicine Baltimore, Maryland
Introduction ....................................................................................... Respiration-Dependent Proton Pumping ........... ....................................... A. Structure of the Mitochondrial Respiratory Chain ................................. B. Models for Respiration-Dependent Proton Pumping .............................. C. Data on Respiratory Chain Proton Pumping ........................................ 111. Reversible Electrogenic Proton Translocation by the Fl-Fo ATPase ................. A. Structure of the Mitochondrial F,-F, ATPase Complex ......................... B. Models for Synthesis of ATP ............................................................ C. Electrogenic Proton Pumping by the Mitochondrial F,-Fo ATPase Complex .......................................................................... IV. The Role of Proton Translocation in Mitochondrial Oxidative Phosphorylation .................................................................................. A. Models for Energy Coupling ............................................................ B. Correlation between Phosphorylation Potential and AKH+ .................................................................................... C. Synthesis of ATP by Artificial Proton Gradients ................................... V. Electrophoretic Metabolite Transport ....................................................... VI. Summary ........................................................................................... References ............................................. ....................................... I. 11.
.
408 409 409 410 41 1 416 416 417 420 422 423 424 425 426 428 428
I Present address: Department of Chemistry, University of Maryland-Baltimore County, Catonsville, Maryland 21228
407 Copyright @ 1982 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-153316-6
408
JANNA P. WEHRLE
I.
INTRODUCTION
Oxidative phosphorylation in eukaryotes is accomplished through the simultaneous operation of no less than four electrogenic proton “pumps” embedded within the inner membrane of the mitochondrion. Three of these ordinarily operate in the forward direction, producing net translocation of protons from the mitochondrion during substrate oxidation. In contrast, the F,-F, ATPase functions in reverse, as a proton-driven ATP synthetase during oxidative phosphorylation. Like its counterpart in bacteria, the ATPase complex of mitochondria can also function as a proton-pumping ATPase, resulting in the accumulation of cations or reversal of electron flow through the respiratory chain. In addition to the proton pumps, mitochondria also contain a number of other transport sytems that catalyze net transfer of charge. These are electrophoretic systems, driven by the membrane potential created by the proton pumps, but do not function electrogenically under physiological conditions. The two major electrophoretic transport systems are the ATP4-ADP3- exchange translocator and the Ca2+ uniporter. Mitochondria also contain proton-driven translocators which are electrically neutral and therefore are driven solely by the proton concentration gradient. In the early 1960s Mitchell (1961) and Williams (1961) independently introduced the concept that proton movement stabilized by the special structure of the biological membrane might be responsible for storing energy released during substrate oxidation and transferring this energy to the synthesis of ATP. Although hypotheses about the precise nature of the proton movement vary and are currently the subject of much research and debate, it now seems likely that this general mechanism is involved not only in oxidative phosphorylation in mitochondria but in oxidative phosphorylation in bacteria and in photophosphorylation in chloroplasts as well. In comparing models for oxidative phosphorylation it is important to realize that there are at least three separate major problems involved: (1) the mechanism of respiratory chain proton pumping, (2) the mechanism of ATPase proton pumping and ATP synthesis, and (3) the mechanism of conversion of redox to phosphate bond energy. Not all models address all three problems, and more importantly most experimental data address only one of these problems. In this article I have attempted to isolate the three problems of oxidative phosphorylation insofar as possible, to present in each case several models currently under consideration, and to present the most pertinent data which may lead to an understanding of the molecular mechanism and the true role of electrogenic proton pumping in mitochondria.
23.
H
+
409
PUMPING IN MITOCHONDRIAL OXIDATIVE PHOSPHORYLATION
11.
RESPIRATION-DEPENDENT PROTON PUMPING
A. Structure of the Mitochondria1 Respiratory Chain An abbreviated view of the respiratory chain is shown in Fig. 1. A more detailed picture may be found in Hatefi and Galante (1978). The points marked “ATP” are the so-called sites of oxidative phosphorylation. This term refers to electron transfers which release sufficient energy for synthesis of one molecule of ATP, and not to the locution of ATP synthesis, which is the F,-F, ATPase. Thus substrates whose oxidations produce NADH cause electron transfer through all three sites and have P/O ratios of 3, because nearly three ADPs are phosphorylated per oxygen consumed, that is, per two electrons transferred. Electrons from succinate (and other FAD-linked substrates) cross sites 2 and 3 only, and so have P/O ratios near 2. Certain artificial substrates which reduce cytochrome c have P / O ratios of 1 or less. Experimental values are typically extrapolated to the next highest integer, but the existence of a precise stoichiometry relation between phosphorylation and electron transport has not been unequivocally established. The sequence of electron transfer components shown in Fig. 1 was developed on the basis of studies with site-specific inhibitors and a variety of electron donors and acceptors of known redox potential. Its physical reality can be demonstrated when the membrane is disrupted with detergent. Certain sections of the chain persist as well-defined, functional multipolypeptide complexes, indicating relative internal cohesiveness. These complexes, designated I-IV, correspond rather strikingly to the sites of oxidative phosphorylation demonstrated classically. (For a review, see Hatefi and Galante, 1978.) Complex V, isolated similarly, is the F,-F, ATPase. Complex I contains the proteins for the reduction of quinones by NADH (site I). Complex I1 contains succinic dehydrogenase and other proteins necessary for the reduction of quinone by succinate. This section is not a site of phosphorylation. Complex I11 (cytochromes b and c1and other
NADH+~Qp,&cyti DEHYDRO
.&02
succinato
FIG. 1. Organization of the mitochondria1 respiratory chain.
H20
410
JANNA P. WEHRLE
proteins) transfers electrons from quinones to cytochrome c and so includes site 2. Complex IV (cytochrome oxidase) catalyzes the oxidation of cytochrome c by oxygen (site 3). This complex includes cytochromes a and a, and copper.
B. Models for Respiration-DependentProton Pumping Figure 2 illustrates the orientation of proton pumping systems of the inner membrane. In intact mitochondria (Fig. 2A) the F,-F, ATPase faces the internal compartment, the matrix. The side of the membrane bearing the ATPase is referred to as the M (matrix)-side. The opposite side is the C (cytoplasmic)-side. These names are retained even in submitochondrial vesicles, which are “inside-out” (Fig. 2B). These vesicles, prepared by sonicating the inner membrane, played an important role in examinations of the spatial arrangement of the inner membrane. Models for proton pumping fall into two categories, the “direct chemiosmosis” model of Mitchell (1979) and the molecular proton pump model, which includes many variations. In Mitchell’s hypothesis the protons that are translocated are formed from redox hydrogens. Redox carriers of hydrogen and carriers of electrons alternate throughout the chain folded back and forth across the membrane so that only hydrogens move from the M-side to the C-side, and only electrons from the C-side to the M-side (Fig. 3A). Hydrogens donated at the M-face discard their protons at the C-face before the electron travels back across the membrane during its next transfer, and then an M-side proton must be absorbed before hydrogen can recross the membrane in the subsequent step. This model has several testable predictions. First, the number of protons translocated per electron must equal the number of complete “loops” traversed. Second, the structure of the respiratory chain must consist of alternating hydrogen
B
A
FIG.2. Orientation of proton translocation during oxidative phosphorylation in (A) mitochondria and (B) submitochondrial vesicles.
U
INTACT MITOCHONDRION
SUBMITOCHONDRIAL VESICLE
23.
H
+
PUMPING IN MITOCHONDRIAL OXIDATIVE PHOSPHORYLATION
41 1
FIG.3. Models for respiratory chain proton translocation. (A) Mitchell’s (1974) direct chemiosmosis. (B) Proton pumping due to protein conformational changes accompanying the redox transition. See text for details.
and electron carriers appropriately disposed at the M- and C-faces and suitable for transmembrane electron transfer. Specifically, regions of pure electron transfer are not candidates for proton pumping according to this scheme. In contrast to the direct formulation of Mitchell are the mechanisms described as molecular proton pumps (Fig. 3B). According to any of these relatively undefined models, redox reactions cause conformational changes in protein components of the chain, resulting in net proton translocation. Often invoked is the Bohr effect in hemoglobin, where binding of oxygen causes a change in the protonation state of the protein. In the case of mitochondria, a “vectorial Bohr effect’’ has been suggested, in which a proton is taken up at the M-face on reduction but released at the C-face upon oxidation. These models in general lack enough detail to permit complete experimental evaluation. However, one may expect a pH-dependence of certain redox transitions, and probably a stoichiometric relationship between protons pumped and electrons translocated, although the ratio could be any number. Structural features might resemble known molecular ion pumps or be similar at all three sites of proton pumping in the respiratory chain. However, the possibility of three unique molecular pumps is a real one. It should be noted that a chemiosmotic scheme for energy conservation does not depend upon the mechanism of proton pumping, but upon equilibration of the translocated protons with the aqueous phase before their use for ATP synthesis.
C. Data on Respiratory Chain Proton Pumping Electrogenic proton translocation during coupled respiration has been observed under a wide variety of conditions almost since the idea was first suggested (Mitchell, 1961). Although it has not yet been possible to observe a substantial respiration-dependent, interior negative transmembrane potential in mitochondria using microelectrodes (Bowman et al., 1978), a wide variety of other techniques have produced results consistent with the existence of such a potential. During respiration cations are accumulated
412
JANNA P. WEHRLE
by intact mitochondria when provided with either endogenous or exogenous pathways for electrophoretic transport. These include Ca*+, K+ in the presence of the ionophore valinomycin, and lipid-soluble cations such as tetraphenylphosphonium (Brand et al., 1976; Kamo et al., 1979). Respiration-dependent changes in certain potential-sensitive indicator dyes are identical to those observed in the presence of artificially imposed diffusion potentials which are inside negative (Jasaitis et al., 1971). The results obtained with these and other techniques, although not quantitatively identical (Azzone et al., 1978a,b), support the conclusion that respiration induces formation of a proton electrochemical gradient of between 180 and 220 mV, alkaline and negative inside (Nicholls, 1974; Kamo et al., 1979). Sorgato et al. (1978) measured a gradient of 185 mV positive and acid inside in submitochondrial vesicles. For the present it does not seem useful to disregard the existence of a respiration-dependent potential on the basis of the microelectrode data alone. For a discussion, see Tedeschi (1979) and Rottenberg (1979). The observation of respiration-dependent proton extrusion by mitochondria is dependent upon the presence of a permeant ion to provide charge compensation, indicating that development of a transmembrane electrical potential opposes further H+ translocation. (But for an alternative explanation see Kell, 1979.) Uncoupling, the state in which respiration is maximized and its ability to do work is abolished, occurs only when a path for proton recycling is created by protonophores or by a combination of ionophores which provide a complete electrophoretic proton return pathway (e.g., valinomycin plus nigericin) or by disruption of the membrane. Taken together, these results indicate that electrogenic proton translocation is a necessary part of energy-conserving respiration. Two approaches have been taken in the search for the location and molecular mechanisms of proton pumping by the respiratory chain. In the first approach inhibitors and artificial redox donors and acceptors are used to bracket isolated spans of the chain in mitochondria or submitochondrial vesicles. In the second approach the components of the chain are physically isolated, purified, and then reconstituted into phospholipid vesicles and assayed for respiration-dependent charge and proton transfer. The combination of these two approaches is yielding a more unified picture of the proton-translocating systems of the respiratory chain. OF PROTONS BY THE INTACT RESPIRATORY CHAIN 1. TRANSLOCATION
Knowledge of the location and stoichiometry of H+ and e- movements during respiration is a necessary prerequisite for an adequate model of electrogenic proton translocation. Unfortunately, knowledge of the specific
H
23.
+
413
PUMPING IN MITOCHONDRIAL OXIDATIVE PHOSPHORYLATION
redox transfers associated with H + translocation is only beginning to emerge, and there is no general agreement on proton stoichiometry at any of the energy-conserving sites. Recently determined stoichiometries from several laboratories are summarized in Table I. Early values for H+ transfer were very close to 2 H + / 2 e- per site for a variety of substrates in mitochondria (Mitchell and Moyle, 1967) and in submitochondrial particles (Hinkle and Horstman, 1971), in keeping with the hypothesis of Mitchell. The same ratio has continued to receive supporting data from Mitchell's laboratory (Moyle and Mitchell, 1978). Other laboratories have shown, however, that respiration-dependent uptake of even small amounts of phosphate (which is coupled with H+ uptake) can cause a substantial underestimation of the number of protons ejected by mitochondria during respiration in the presence of permeant cations. Measurements of H translocation where phosphate movement has been inhibited or where endogenous phosphate has been removed (Brand et al., 1976) give values of at least 3 H+/2 e- per site for all substrates tested. Indeed, steady state measurements, as contrasted with the usual oxygen pulse measurements, yield H+/2 e- per site ratios approaching 4 in experiments with a variety of techniques and substrates. Net charge transfer ratios of 4 K+ /2 e- per site were measured in the presence of valinomycin whether phosphate moved or not (Reynafarge and Lehninger, 1978). In many cases the observation of different values by different laboratories can be rationalized in terms of experimental conditions, but whether the highest values are the physiologically significant ones is not easy to say. Although observations of Hf per site ratios greater than 2 are now widely reported, it is not clear whether all three coupling sites actually function with the same ratio. Brand et al. (1978) have titrated mitochondria with low levels of substrates and determined the dependence of the resul+
TABLE I PROTONS TRANSLOCATED PER TWO ELECTRONS DURING RESPIRATION Site
Mitchell and Moyle"
Lehninger et at.6
Brand et al.=
Wikstrom et
Azzone et aLe
1 2 3
Two Two Two
Four Four Four
Two Two Four
Two or three Two Four
Four Four Four
' Mitchell (1979). Lehninger et al. (1979). Brand et at. (1978). Wikstrom and Krab (1979). Azzone et at. (1979a); Pozzan et al. (1979).
414
JANNA P. WEHRLE
tant membrane potential upon the rate of oxidation, using safranine dye to track the potential. Different substrates were used to bracket either all three sites (3-hydroxybutyrate), or site 2 plus site 3 (succinate), or only site 3 (isoascorbate). In all cases plots of potential versus the rate of oxygen consumption were linear, with slopes in the ratio 2.0: 1.5: 1.O, respectively. The result is consistent with the translocation of different numbers of protons at the three sites: e.g., 2, 2, and 4 at sites 1, 2, and 3, respectively. One broadly accepted result is the stoichiometry of Ca2+uptake to oxygen consumption. Both Reynafarge and Lehninger (1977) and Moyle and Mitchell (1977a) agree that two Ca2+are taken up per electron pair per site. However, disagreement over the net charge transfer associated with Caz+ uptake (Section V) has allowed this to be used as evidence for the ejection of either two or four H + per site (Moyle and Mitchell, 1977b; Fiskum et al., 1979). Perhaps the greatest controversy at present involves the possibility of proton translocation below cytochrome c. Because the redox reactions through this region are exclusively electron transfers they cannot form a Mitchell loop. Mitchell has positioned the third proton translocation as part of a stepwise reduction of ubiquinone (Q cycle, Mitchell, 1979). This would leave the last transmembrane electron transfer alone at the third site, but a number of laboratories (Wikstrom, 1975; Azzone et al., 1979; Alexandre et al., 1979) have found that electron transfer from ferrocyanide or reduced cytochrome c is in fact accompanied by translocation of protons. Although the stoichiometry varies from 2 to 4, depending on the experimental conditions, the results of Wikstrom and others appear clearly incompatible with a direct Mitchell loop at site 3. Some kind of protonpumping conformation change almost certainly is involved. More evidence on this subject is found in studies on purified cytochrome oxidase (see below). PUMPING BY PURIFIED RECONSTITUTED RESPIRATORY 2. PROTON CHAINCOMPLEXES Each respiratory chain complex associated with a site of oxidative phosphorylation has been shown to function as an oxidation-driven, electrogenic proton pump when purified and reconstituted into phospholipid vesicles. In contrast, the succinic dehydrogenase complex does not appear to pump protons during oxidation. The NADH dehydrogenase-containing complexes from rat liver and from beef heart mitochondria have been shown by Lawford and Garland (1972) to couple reduction of the ubiquinone analog Q, by NADH to an uncoupler-sensitive translocation of protons into the interior of phospho-
23.
H+
PUMPING IN MITOCHONDRIAL OXIDATIVE PHOSPHORYLATION
415
lipid vesicles. H +/2 e- ratios of 0.75 and 1.4 were obtained for the liver and heart enzyme, respectively. Ragan and Hinkle (1975) obtained similar results with beef heart complex I. Oxidation was stimulated threefold by uncoupling with a protonophore carbonylcyanide p-trifluoromethoxyphenylhydrazone (FCCP) or a combination of K+ plus valinomycin plus nigericin, suggesting that the development of A&, + inhibited oxidation. Proton uptake was dependent on charge compensation, in this case K+ plus valinomycin. Oxidation also drove the accumulation of a lipophilic anion, tetraphenylboron (TPB-), indicating a positive potential in the interior of the vesicles. (Operational asymmetry in this system is obtained because the substrate NADH cannot itself enter the vesicles.) The components responsible for proton translocation have not been identified. Complex I has both hydrogen and electron carriers, as required for a Mitchell loop. The pH dependence of component midpoint potentials in this region has not been thoroughly examined; it is known, however, that the electron acceptor ubiquinone has a pH-dependent midpoint potential (Dutton and Wilson, 1974), as would be necessary for a redox change in equilibrium with proton binding or release, and that two of the iron-sulfur centers appear to have midpoint potentials sensitive to the presence of ATP (Ohnishi and Pring, 1974). For a review on complex I, see Ragan (1976). Purified complex I11 has been shown by Leung and Hinkle (1975) to translocate protons when reduced Q2 is oxidized by exogenous cytochrome c. These authors observed outward translocation of approximately 2 H+/2 e-. In the same experiments oxidation was stimulated 10-fold by uncoupling with either protonophores or K+ plus valinomycin and nigericin. While the ubiquinone-to-cytochrome c step provides one site of Mitchelltype proton release in this region, the absence of any hydrogen carriers after ubiquinone has lead Mitchell to postulate the Q cycle (Mitchell, 1975), by which a two-stage oxidation of ubiquinone would produce net translocation of two protons per election, leaving only the final transmembrane electron transport to complex IV (cytochrome oxidase). Potential sites for proton pumping by other mechanisms are also limited. Cytochrome c, has a pH-insensitive midpoint potential, but one form of cytochrome b (b566) appears to have a midpoint potential which is changed by deprotonation (von Jagow, 1979) and is sensitive to the presence of ATP (Dutton and Wilson, 1974). In addition, evidence that the reduction induces a conformational change in complex 111 is substantial (e.g., Rieske et al., 1967). For a review on complex 111, see Rieske (1976). Extensive study of purified complex IV (cytochrome oxidase) confirms the results of studies with the intact respiratory chain, suggesting that protons are translocated during the reduction of cytochrome c by oxygen (Krab and Wikstrom, 1978; Sigel and Carafoli, 1979; Wrigglesworth and
416
JANNA
P. WEHRLE
Nicholls, 1979; Casey et al., 1979). Cytochrome oxidase catalyzes an inhibitor and uncoupler-sensitive H+ ejection equivalent to 2 H+/2 e-, but nearly 4 K+/2 e- are taken up, compensating both for protons consumed to form water (2 H +12 e-) and for protons translocated (apparently 2 H+12 e-). A wide variety of substrates have been used. Translocation is strictly proportional to the number of turnovers of the enzyme and is in excess of the amount of any prereduced components of complex I11 which might be present. (See the arguments of Moyle and Mitchell, 1978; and Lorusso et a[., 1979.) In the purified system stoichiometries higher than 2 H + / 2 ehave not been observed. Whether the higher stoichiometries observed in the laboratories of Lehninger and of Azzone in intact mitochondria represent artifacts or whether the coupling in intact mitochondria is much better than in the purified complexes is not known at the present time. Candidates for proton-releasing redox proteins in complex IV include cytochromes a and a,, both of which have pH-sensitive midpoint potentials. In contrast, the midpoint potential of the copper appears not to be influenced by pH (Dutton and Wilson, 1974). (For a review on cytochrome oxidase, see Wikstrom and Krab, 1979.) 111. REVERSIBLE ELECTROGENIC PROTON TRANSLOCATION BY THE F,-F, ATPase A. Structure of the Mitochondria1 F,-F, ATPase Complex The F,-F, ATPase of mitochondria, like its counterpart in bacteria and chloroplasts, has a structure much more complex than that of many other electrogenic ion pumps (Soper et al., 1979; Kagawa, 1978; Nelson, 1976). F,-F, ATPase complexes of mitochondria from a variety of sources appear to consist of at least 9 types of polypeptide chains and as many as 20 peptides per functional unit (Serrano et al., 1976; Stiggal et al., 1978; Soper et af., 1979). Located in the inner mitochondria1 membrane, the ATPase complex has traditionally been described as a “ball-and-stalk” arrangement (Fig. 2) where the “ball” represents the chemical catalytic portion of the complex, designated F, F, has been suggested to be attached to the membrane sector (designated F,) by an ill-defined protein “stalk.” The appropriateness of this well-used model has recently been confirmed experimentally. Soper et al. (1979) have shown electron micrographs of a dispersed preparation of ATPase complex (“oligomycin-sensitive ATPase” in that reference) from rat liver mitochondria in which the individual enzyme complexes bear an unmistakable resemblance to the artists’ versions published earlier. The F, portion of the ATPase complex can be dissociated from the membrane-bound peptides by a variety of techniques. The F, ATPase alone
-
23.
H
+
PUMPING IN MITOCHONDRIAL OXIDATIVE PHOSPHORYLATION
41 7
contains five types of polypeptide chains and has a molecular weight in the range 350,000-380,000 (Catterall and Pedersen, 1972). Isolated F, contains all the catalytic sites involved in the non-energy-coupled hydrolysis of ATP, but its activity is no longer inhibited by oligomycin or dicyclohexylcarbodiimide (DCCD), which inhibit both ATP synthesis and ATP hydrolysis when the enzyme is bound to its proper membrane sector. Less well characterized is the membrane or F, sector of the ATP synthetase complex, consisting of highly hydrophobic membrane proteins and coupling proteins which bind F, to the membrane. The major component of F, is the DCCD-binding protein. This small, CHC1,-MeOH-soluble protein is the binding site for both DCCD and oligomycin. F, appears to form a proton-conducting channel through the mitochondrial membrane. When reconstituted into phospholipid vesicles, it induces a proton-specific conductance which can be inhibited by DCCD or oligomycin or by rebinding F, (Shchipakin et al., 1976). In addition to the peptides of F, and F, the ATP synthetase complex appears to contain a small inhibitor or regulator peptide. First reported by Pullman and Monroy (1963), the peptide inhibits ATP hydrolysis under certain conditions and appears not to inhibit ATP synthesis under conditions which favor linear rates of synthesis (Cintron and Pedersen, 1979), but its mode of action is far from clear. For a review on the mitochondrial ATPase complex, see Pedersen et al., 1978. B. Models for Synthesis of ATP Models for ATP synthesis may discuss only catalysis or only coupling or may include both. Most models which consider only catalysis only invoke an energy-dependent conformational change of unspecified origin to drive phosphate bond formation or release of preformed ATP. Two such models are illustrated in Fig. 4. The work of Tiefert and Moudrianakis (1979) on the analogous chloroplast ATPase system has lead these authors to suggest that enzyme-bound AMP, rather than ADP, may be the acceptor of the new phosphodiester bond, this first step being followed by an adenylate kinase-type transfer of P to an acceptor ADP. The quite different model of Boyer et af. (1973) shares the idea that addition of Pi to ADP may not be the energy-requiring step in phosphorylation. According to the “alternating catalytic site” mechanism, tightly bound ADP is the primary acceptor but reacts nearly at equilibrium to form ATP. ATP is then expelled from its binding site in an energy-requiring reaction, promoting recycling of the enzyme. In fact Rosing et al. (1977) have reported isotope exchange data consistent with the notion that formation of the P,y-phosphodiester bond is relatively insensitive to protonophore uncouplers.
-
418
JANNA P. WEHRLE
FIG.4. Conformational models for catalysis of ATP hydrolysis and synthesis. (A) Boyer et al. (1974) have suggested an alternating catalytic site mechanism. (B) Tiefert and Moudrianakis (1979) have suggested that AMP rather than ADP may be the primary acceptor in phosphorylation by chloroplast F,-F, ATPase. See text for details.
Other models for synthesis describe explicitly a role for protons in ATP synthesis. These are therefore also models for ATP-dependent H+ pumping and ultimately must take some position on the nature of energy coupling. These models may be divided into two categories, described below. 1. CHEMIOSMOTIC MECHANISMS OF SYNTHESIS
Mitchell has suggested direct participation in the chemical reactions of ATP synthesis as the simplest mechanism by which translocation of protons can aid in the dehydration of ADP and Pi (Mitchell, 1973). As illustrated in Fig. 5A, his model designates the proton-conducting channel of the F, as a “proton well” in which the proton concentration increases as the thickness of the membrane is traversed, parallel to the reduction in A$. At the M-surface, where the concentration of protons corresponds to the total A,!iH+, the active site of F, is organized in space so that protonation of the oxygen departing from phosphate is favored, without permitting protonation of ADPO-. This is followed by an attack of ADP on the phosphorus, with phosphodiester bond formation and the release of H,O.
23.
H + PUMPING I N MITOCHONDRIAL OXIDATIVE PHOSPHORYLATION
A
41 9
B
2H*
FIG. 5 . Models for energy coupling to ATP synthesis. (A) Mitchell’s (1974) direct chemical involvement mechanism for proton-driven ATP synthesis. (B) Protonation of the protein may induce a conformational change resulting in ATP synthesis by mechanisms such as those shown in Fig. 4.See text for details.
ATP is thought to leave the catalytic site in a protonation state, which differs from the protonation of ADP + Pi by exactly the number of protons used in the protonation steps, so that release of ATP into solution on the M-side completes translocation of the protons. Mitchell has frequently suggested that two protons are involved (Moyle and Mitchell, 1973) but, as discussed below and as mentioned by Mitchell (1977), the number may be three, which does not eliminate the possibility of direct chemical involvement. Still consistent with a role for A,iiH+ as the driving force for ATP synthesis are the conformational model of Boyer (1975) and the magnesium complex model of Racker (1977). These are indirect, by comparison with Mitchell’s model, since the electrochemical gradient for protons would only serve to release energy (in the form of ATP) from storage. To distinguish further between direct and indirect chemiosmotic models would require detailed molecular information not currently available. Now that F,-F, ATPase complexes have been purified from several mitochondria1 sources (Serrano et al., 1976; Stigall et al., 1978; Soper et al., 1979), it is to be hoped that such necessary information will soon be forthcoming.
2. NONCHEMIOSMOTIC MECHANISMS OF SYNTHESIS Nonchemiosmotic models for energy coupling differ substantially in the role assigned to proton translocation. Few contain any specific predictions about the mechanism of ATP synthesis. R. J. P. Williams (1975) has suggested that a high local concentration of protons in some nonaqueous phase of the membrane assists in the dehydration by immediate protonation of the newly formed water, removing it from the active site. In this
420
JANNA P. WEHRLE
model protons in the bulk aqueous phase-and therefore proton pumping as it is commonly conceived-play no role. Direct protein-protein conformational interactions are still considered necessary to explain all the available data, according to some authors (Slater, 1977).
C. Electrogenic Proton Pumping by the Mitochondria1 F,-F, ATPase Complex As discussed earlier in connection with respiratory chain-linked activity (Section II,A), the generation of a substantial membrane potential due to ATP hydrolysis by the F,-F, ATPase complex has not been observed directly using microelectrodes. Nonetheless it is an absolute prediction of the chemiosmotic hypothesis that the ATP complex is an electrogenic H + pumping device. Therefore this proposal has been examined by other techniques. The evidence that the F,-F, ATPase translocates protons during ATP hydrolysis and that such translocation is electrogenic falls into two categories. Kinetic studies have measured the rate of appearance of protons during ATP hydrolysis. Other studies measure the establishment of a steady state transmembrane electrochemical potential difference in protons generated during ATP hydrolysis, using a wide variety of mobile ions or pH- or potential-sensitive dyes. Although the quantitative results are not identical, every technique used so far has produced data consistent with net electrogenic translocation of protons from the M-face aqueous phase to the C-face aqueous phase during coupled ATP hydrolysis in intact mitochondria, in submitochondrial particles, and in liposomes containing the purified F,-F, ATPase complex. 1. PROTONPUMPING BY F,-F, IN INTACT MITOCHONDRIA AND SUBMITOCHONDRIAL VESICLES
Hydrolysis of added ATP, by respiration-inhibited mitochondria in the presence of K+ plus valinomycin, is accompanied by the appearance of protons in the external medium (Mitchell and Moyle, 1968; Brand and Lehninger, 1977). The ejection is blocked by oligomycin (which also inhibits ATP hydrolysis) and also by protonophores, which actually stimulate ATP hydrolysis. This suggests that translocation rather than net production of protons is occurring. In the absence of a mobile cation neither proton ejection nor substantial ATP hydrolysis occurs, which suggests the need for charge compensation in order to obtain repeated cycles of ATP hydrolysis. Reynafarge and Lehninger (1978) report the inward movement of K+ equivalent to the outward movement of protons. The quantitative aspect of ATP-driven proton pumping is still the sub-
23.
H
+
PUMPING IN MITOCHONDRIAL OXIDATIVE PHOSPHORYLATION
42 1
ject of debate. Mitchell and Moyle (1968) and Brand and Lehninger (1977), using small pulses of ATP, observed a proton/ATP ratio of about 2. More recently Alexandre et al. (1978) have reported steady state rate ratios of three protons ejected per ATP hydrolyzed, as well as three K + taken up per ATP hydrolyzed in the presence of valinomycin (Reynafarge and Lehninger, 1978). This suggests fully electrogenic translocation of three H+ per ATP. Translocation ratios were equivalent when Ca2+was used for charge compensation (Alexandre et al., 1979). Ca2+has an endogenous system for rapid electrophoretic uptake, avoiding the sometimes criticized practice of using ionophores. Measurements of the steady state A,%”+generated by intact mitochondria hydrolyzing ATP are not as numerous as those measuring the respirationdependent state, although qualitative indications of the development of an internally negative transmembrane potential have been observed using a variety of methods. Nicholls (1974), using the distribution of Rb+ in the presence of valinomycin and HOAc, reported a A$ of 125 m V and a ApH corresponding to 85 mV in rat liver mitochondria hydrolyzing ATP. More recently, using trace amounts of the permanent cation tetraphenylphosphonium, Kamo et al. (1979) measured values of A$= 150 mV and ApH = 30 mV. This method, which does not employ valinomycin, demonstrates that a potential is generated by the low “state 4-like” ATP hydrolysis which occurs under tightly coupled conditions, as well as during hydrolysis facilitated by rapid charge compensation as with K+ plus valinomycin. ATP hydrolysis by the F,-F, ATPase in submitochondrial vesicles is accompanied by protonophore-sensitive proton uptake (Moyle and Mitchell, 1973; Thayer and Hinkle, 1973). Because submitochondrial vesicles are inverted relative to intact mitochondria, the ATPase has direct access to ATP without intervention of the electrogenic ATP-ADP exchange translocator. Measurements were carried out at pH 6.2-6.3 in both studies to avoid scalar proton release due to inorganic phosphate, which in these vesicles is released into the external medium. Under these conditions ratios of H+ uptake to ATP hydrolysis of 1.7 or slightly less were obtained in both laboratories. In submitochondrial vesicles protons leak through the membrane via F, sectors from which F, has been removed. Respiration rates can be depressed by blocking F,, with oligomycin (Thayer and Hinkle, 1973). For this reason the authors concluded that H+/ATP ratios were somewhat underestimated, and they suggested that a ratio of 2 H+/ATP might be the correct value. Because of the poor coupling in submitochondrial vesicles due to uncovered F,, values for A&+ developed by ATP hydrolysis alone are rarely reported. Qualitative indications for the development of a membrane
422
JANNA P. WEHRLE
potential during ATP hydrolysis (uptake of the lipid-soluble anion tetraphenylboron) have been reported by Grinius et al. (1970). As observed during the measurement of respiration-dependent H+ translocation, the movement of ions or metabolites may complicate interpretations of ATP-dependent proton transport data in intact mitochondria or submitochondrial vesicles. Study of the F,-F, ATPase in better resolved systems has therefore assisted substantially in clarifying the behavior of the complex. 2. PROTON PUMPING BY PURIFIED F,-F,
The most straightforward evidence that the F,-F, ATPase complex can function as an electrogenic proton pump has come from the study of the purified enzyme complex. The F,-F, ATPase purified from bovine heart mitochondria has been reconstituted into artificial phospholipid vesicles (Serrano et al., 1976). In the presence of valinomycin, to allow rapid charge compensation by K+ movement, uptake of protons during ATP hydrolysis has been demonstrated. Proton uptake is sensitive to rutamycin, a form of oligomycin, but also to the protonophore uncoupler FCCP, indicating that translocation of protons rather than scalar production of OH- is occurring. Enhancement of the fluorescence of 1-anilino-&naphthalene sulfonate can also be observed, indicating the development of a membrane potential. Now that several other preparations of ATPase complexes are available, confirmation of the results of Serrano et al. (1976) may be forthcoming. In the case of the purified, reconstituted thermophilic bacterium enzyme complex TF,-F,, Sone et al. (1976) have demonstrated that the hydrolysis of ATP leads to creation of a AilH+ of at least 280 mV, with substantial contributions from both A$ and ApH.
IV. THE ROLE OF PROTON TRANSLOCATION IN MITOCHONDRIAL OXIDATIVE PHOSPHORYLATION The last of the three problems of oxidative phosphorylation is the mechanism of energy coupling between respiration and ATP synthesis. It has been established that both the respiratory chain and the F,-F, ATPase can function as electrogenic proton pumps under certain conditions. What is the role of this translocation in energy transfer? This question can be addressed in two ways. One is to reexamine the relationship between AilH+ generated by respiration and ATP synthesis. The other is to assay directly for synthesis of ATP by an artificial proton gradient in intact mitochondria and in better resolved systems.
23.
H
+
PUMPING IN MITOCHONDRIAL OXIDATIVE PHOSPHORYLATION
423
A. Models for Energy Coupling 1. CHEMIOSMOTIC MECHANISMS The term “chemiosmosis” refers specifically to Mitchell’s hypothesis about the mechanism of energy coupling and not to his proposals for the mechanisms of proton pumping discussed above. Chemiosmosis proposes that energy is stored and used in the form of a true transmembrane gradient in the electrochemical potential of osmotically active protons. It is only protons in equilibrium with the bulk aqueous phase which can be described by the now familiar expression for the total electrochemical potential difference in protons (AFH+)across a membrane, or the proton motive force @P):
AFH+ = FAp
AP = A$ - ZApH
(1)
where A$ is the electrical potential difference, ApH is the pH difference across the membrane, and Z is 2.303RT/F, equal t o about 59 mV at 25°C (Mitchell, 1961). According to this formulation, the only effect of the membrane potential is to increase the potential energy difference between protons on the M-side and protons on the C-side of the membrane. All models in which ApH+ is considered the “high-energy intermediate” are chemiosmotic mechanisms, regardless of the mechanism proposed for generation of the proton gradient. The energy of formation of ATP under a given set of circumstances is given by the so-called phosphorylation potential AG,: AG,=AG,O +2.303RTlog ([ATP]/[ADPl [Pi])
(2)
In a chemiosmotic system the energy available to synthesize ATP is given by the number of protons moved multiplied by the potential energy difference through which they are moved; thus AG,=nAji,+
(3)
Both the prediction of linearity and the stoichiometry can be tested experimentally, as can the prediction that only total A,iiH+ is directly related to phosphorylation and not its separate components. 2. NONCHEMIOSMOTIC MECHANISMS In contrast to the hypothesis of Mitchell are a wide variety of other models, some of which involve protons and some of which do not. In the first category are the models in which protons play a role but are never in full equilibrium with either aqueous phase between relocation by respira-
424
JANNA P. WEHRLE
tion and utilization for ATP synthesis. The model of R. J. P. Williams (1975) conceives of protons as separated from their charge-compensating electrons but remaining within the membrane, later to be used to dehydrate ADP + Pi, still within the membrane, rather than equilibrating with either aqueous phase. Kell (1979) has suggested that protons released by respiration may be restricted to the membrane by the energy barrier associated with penetration of the first layer of organized water at the membrane surface. Other proton-driven but nonchemiosmotic models have been proposed by Azzone et al. (1977), Robertson and Boardman (1975), and Rottenberg (1978). Models for energy coupling in which high-energy protons are not considered the sole intermediate are still under consideration. Direct protein-protein interactions between respiration and phosphorylation are considered necessary by some authors in order to explain certain inhibitor data which suggests that the ATP synthetase and the respiratory chain are not completely independent (Slater, 1977).
B. Correlation between Phosphorylation Potential and ApH+ The AGp/AiH+ ratio has now been measured under a wide variety of conditions. Values of n, the number of protons translocated per ATP synthesized, may be compared to the direct proton flux measurements during ATP hydrolysis. Evaluations of ApH and A$ have proved difficult and strongly dependent upon the indicators used (Azzone et al., 1978a,b). Nonetheless many types of studies have yielded similar results. Nicholls (1974), using SCN- and OAc-, determined that respiring rat liver mitochondria maintained a AFH+ ranging from - 180 to - 230 mV, depending upon the medium, but that in all cases AGp was to approximately -7.9 kcal/mole, which would correspond to 270-290 mV if two protons were translocated per ATP synthesized. Kamo et al. (1979), who measured a ApH+ of 200 mV during respiration or ATP hydrolysis, calculated AGP/ApH+= 2 . 7 and suggested that the true value might be n = 3. In submitochondria1 vesicles Rottenberg measured a AG,/AFH+ of 2.9. Sorgato et al. (1978), who measured a AFH+ in submitochondrial vesicles of 185 mV, found AGp/AFH+ ratios of 3.1, 3.0, and 3.3 in various media. Where it has been measured, AFH+ has been found to decline during ATP synthesis, consistent with its proposed role as the energy source (Nicholls, 1974; Sorgato et al., 1978). Although the possibility remains that ApH+ is being seriously underestimated in all cases, it appears that a value of n = 3 for ATP synthesis is widely supported both by flux measurements during ATP hydrolysis (Sec-
23.
H
+
PUMPING IN MITOCHONDRIAL OXIDATIVE PHOSPHORYLATION
425
tion 111, C ) and by measurements of the ratio of A $ H + generated to the AG, maintained. This is incompatible with a pumping of 2 H + / 2 e- per site by the respiratory chain and so with Mitchell’s direct chemiosmosis hypothesis of how the proton gradient is generated; it is certainly not incompatible with a chemical role for protons in ATP synthesis, nor generally with the chemiosmotic hypothesis for energy coupling. A more serious problem for the chemiosmotic hypothesis is presented by the work of Azzone and co-workers (e.g., Azzone el al., 1978a,b), who have made an extensive study of the relationship between A P H + and AG, as A$H + is varied by different reagents (salts, respiratory inhibitors, protonophores, ionophores, and combinations of ionophores). According to their results, not all methods of changing A,iiH+ have the same effect on AGp (Azzone et al., 1977). Thus, using nigericin plus valinomycin-so that A$ and ApH are lowered independently-produces a simultaneous fall in AhH+ and AG,, with a constant ratio of n = 3 . In contrast, protonophore uncouplers such as FCCP, which lower A,iiH+ directly, give a value of n approaching infinity at a low A/ZH+, which suggests that AG, can be maintained by a very large number of low-energy protons. It is unfortunate that AG, and AFH+in these studies were not measured under identical conditions for each addition of special reagent. However, the results nevertheless require explanation if the chemiosmotic hypothesis is to be accepted.
C. Synthesis of ATP by Artificial Proton Gradients In order t o answer the question of whether proton pumping or some other function of respiration is responsible for ATP synthesis, a more direct physical approach has also been taken. This involves the use of artificial gradients of proton concentration and charge to synthesize ATP in the complete absence of respiration and ultimately in the absence of all respiratory chain components. As early as 1967 Cockrell et al. (1967) had obtained evidence that a small amount of ATP could be synthesized by mitochondria during K + efflux facilitated by valinomycin and driven by a K + concentration gradient. Glynn (1967) suggested that this might be due to the development of a membrane potential which could drive ATP synthesis. Rossi and Azzone (1970) demonstrated that, by increasing the internal K + , substantial amounts of ATP could be synthesized in respiration-inhibited mitochondria using valinomycin-facilitated K + efflux. They noted that the addition of acid t o the external medium increased the amount of ATP formed, and they attributed the ATP synthesis to reversal of a proton-cation exchange pump. Azzone and Massari (1971) showed that the effluent ion could be Rb+ as well as K + . In their study several ratios of ApH to A$ (described as
426
JANNA P. WEHRLE
ApK+) could synthesize ATP, provided that the total gradient exceeded 2.5 units (150 mV). The K+,,u,/ATP,, stoichiometry was found to vary from 2 to 4 as the pH gradient decreased at a constant K + gradient. In 1975 Thayer and Hinkle (1975a) reported the synthesis of ATP in response to additive pH and pK+ gradients in respiration-inhibited submitochondrial vesicles treated with valinomycin. They also demonstrated (Thayer and Hinkle, 1975b) that ATP synthesis using artificial gradients could proceed at a rate faster than ATP synthesis energized by respiration, indicating that proton gradient-driven ATP synthesis was kinetically competent t o transfer energy from respiration to synthesis. Varying the ratio of A$ to ApH at a constant Ah,+, they found similar rates of ATP synthesis, supporting the proposal of Mitchell that the effect of A$ is to increase Ah,+ rather than to produce some separate effect. Net synthesis of ATP by purified, reconstituted mitochondrial F1-Fo ATPase using only chemical gradients and ionophores has not yet been reported, although this has been accomplished for the more stable complex from thermophilic bacteria (Sone el al., 1977; Kagawa, this volume). The bovine heart mitochondrial F,-F, ATPase enzyme has been reconstituted with other purified respiratory chain complexes to give phosphorylating proteoliposomes showing P/O ratios as high as 0.5. A simpler protein, bacterial rhodopsin reconstituted into liposomes, can be shown to pump protons inward in response to light. Simultaneous incorporation of mitochondrial F,-F, ATPase plus bacteriorhodopsin into liposomes produces a light-dependent, uncoupler- and rutamycin-sensitive ATP synthesis, as summarized by Racker el al. (1975). Although the results are consistent with synthesis being dependent upon Afi, + generated by respiratory or photosynthetic pumps, the presence of other proteins in the liposomes leaves open the possibility of direct protein-mediated energy transfer. Thus, the situation with mitochondrial ATPase complexes cannot be considered as well defined as for the analogous system in thermophilic bacteria. Nonetheless, the evidence is quite convincing that A TP can be synthesized by an imposed AhH+ via the F,-Fo ATPase complex, independent of the redox activity or even of the presence of respiratory chain components. Although it is not impossible that the in vivo pathway may be different, certainly the burden of proof rests with those who suggest pathways which do not involve Ah, + .
V.
ELECTRO PH0 R ETlC METABO LlTE TRANSPORT
A large variety of metabolites and inorganic ions are transported both in and out of mitochondria. For a review, see LaNoue and Schoolwerth
23.
H
+
PUMPING IN MITOCHONDRIAL OXIDATIVE PHOSPHORYLATION
427
(1979). The transport of most of these substances has been shown to be dependent upon the mitochondrial energy state. Electrophoretic transport of cations or anions involves net transfer of charge and can be driven by the electrical potential gradient. Electroneutral transport by proton symport or antiport can be driven by ApH. Electroneutral exchange diffusion of substrates also occurs. Electrogenic transport of ions other than protons has not been demonstrated in coupled, respiring mitochondria. However, given the necessary concentration gradients, systems which function electrophoretically under energized conditions can be identified, because they will function electrogenically under deenergized conditions. The natural permeability of mitochondria to cations is quite limited, with the exception of Ca 2+,for which an extremely rapid endogenous electrophoretic carrier system exists (Bygrave, 1978). Ca2+ uptake appears to be fully electrogenic (Fiskum eta/., 1979; Akerman, 1978). It stimulates increased respiration and causes H ejection necessary to reestablish the membrane potential decreased by cation uptake. This permits Ca2+ to be used experimentally as a replacement for K + plus valinomycin, and for observation of the maximal events of H + translocation. Ca2+uptake in the absence of added permeant weak acid anion is clearly limited (40 ng. atoms/mg of protein), and Moyle and Mitchell (1977a,b) claim that all Ca2+is taken up by obligatory symport with Pi, thus reducing the effective charge transfer to C a + . However, direct measurements during Ca2+ uptake (Fiskum el a/., 1979) have not shown translocation of phosphate. Resolution of this conflict is critical t o an understanding of charge translocation in mitochondria. Another very active endogenous transport system that results in net charge transfer is the ATP4--ADP3- exchange system (Klingenberg et a/., 1977). This process, which promotes an antiport of ATP4- for ADP3-, is specific for free nucleotides (rather than Mg2+-boundnucleotides), in contrast t o virtually every other cell reaction. The energy cost of translocating one negative charge for ATP out of the matrix against the membrane potential is a substantial fraction of the cost of oxidative phosphorylation (Lehninger et al., 1979). There are a variety of mitochondrial transport systems which catalyze electroneutral ion flux. One of the most active is H,POa.H+ symport (Coty and Pedersen, 1975). Lipophilic weak acids, e.g., acetate and 6-hydroxybutyrate are accumulated in a similar manner, except that their transport occurs by unmediated diffusion of the neutral free acid (Chappell and Haarhoff, 1967). The Na +-H+ system and the less active K+- H+ exchange system may function in cation extrusion (Brierley, 1976). These systems are all driven under appropriate conditions by the proton concentration gradient alone. Another type of neutral exchange transporter responsive to +
428
JANNA
P. WEHRLE
neither All/ nor ApH is the anion-exchange mediator. These systems catalyze exchange of anionic metabolites for other metabolites of the same charge. (See LaNoue and Schoolwerth, 1979.) VI.
SUMMARY
The substantial evidence that electrogenic proton translocation is responsible for oxidative phosphorylation in mitochondria makes more important than ever elucidation of the molecular mechanisms involved. If direct chemiosmosis must be abandoned because of well-substantiated measurements of proton movements in excess of the number of electrons transferred, then the mechanism of proton pumping by the respiratory chain is completely unknown. The proton-pumping complexes bear no striking resemblance to simple ion-pumping ATPases, to the mitochondria1 F,-F, ATPase complex, or to each other. Knowledge of the precise pathway of electron transfer within each complex is essential. Experimental verification must be sought for each of the conformational changes so casually invoked. Indeed study of the proton pumps of oxidative phosphorylation has just begun.
ACKNOWLEDGMENTS The author would like to thank Dr. Gary Fiskum, for helpful discussions regarding ion translocations, and Dr. Peter L. Pedersen for critical reading of the manuscript. This article was written while the author was supported by National Science Foundation grant PCM 7813249 awarded to Dr. Peter L. Pedersen.
REFERENCES Akerman, K. E. 0. (1978). FEBS Lett. 93, 293-296. Alexandre, A., Reynafarge, B., and Lehninger, A. L. (1979). Proc. Natl. Acad. Sci. U.S.A. 75, 5296-5300. Azzone, G. F., and Massari, S. (1971). Eur. J. Biochem. 19, 97-107. Azzone, G. F., Massari, S., and Pozzan, T. (1977). Mol. Cell. Biochem. 17, 101-111. Azzone, G. F., Pozzan, T., Massari, T., and Bragadin, M. (1978a). Biochim. Biophys. Acta 501, 296-306. Azzone, G. F., Pozzan, T., and Massari, S. (1978b). Biochim. Biophys. Acta 501, 307-316. Azzone, G. F., Pozzan, T., and Di Virgilio, F. (1979). J. Biol. Chem. 254, 10206-10212. Bowman, C., Maloff, B. L., and Tedeschi, H. (1978). In “Frontiers of Biological Energetics: From Electrons to Tissues” (P. L. Dutton, J. S. Leigh, and A. Scarpa, eds.), pp. 413-421. Academic Press, New York. Boyer, P. D. (1975). FEBS Lett. 50, 91-94.
23. H
+
PUMPING IN MITOCHONDRIAL OXIDATIVE PHOSPHORYLATION
429
Boyer, P. D., Cross, R. L., and Momsen, W. (1973). Proc. Natl. Acad. Sci. U.S.A. 70,28372839. Brand, M. D., and Lehninger, A. L. (1977). Proc. Natl. Acad. Sci. U.S.A. 14, 1955-1959. Brand, M. D., Reynafarge, B., and Lehninger, A. L. (1976). J. Eiol. Chem. 251, 5670-5679. Brand, M. D., Harper, W. G., Nicholls, D. G., and Ingledew, W. J. (1978). FEES Lett. 95, 125- 129. Brierley, G. P. (1976). Mol. Cell. Eiochem. 10, 41-62. Bygrave, F. L. (1978). Eiol. Rev. 53, 43-79. Casey, R. P., Chappell, J. B., and Azzi, A. (1979). Eiochem. J. 192, 149-156. Catterall, W. A., and Pedersen, P. L. (1974). Eiochem. Soc. Spec. Publ. 4, 63-88. Chappell, J. B., and Haarhoff, K. N. (1967). In “Biochemistry of Mitochondria” (E. C. Slater, Z . Kaniuga, and L. Wojtczak, eds.), pp. 75-91. Academic Press, New York. Cintron, N. M., and Pedersen, P. L. (1979). J. Eiol. Chem. 254, 3439-3443. Cockrell, R. S., Harris, E. J., and Pressman, B. C. (1967). Nature (London) 215, 1487-1488. Coty, W. A., and Pedersen, P. L. (1975). Mol. Cell. Eiochem. 9, 109-124. Dutton, P. L., and Wilson, D. F. (1974). Eiochim. Eiophys. Acta 346, 165-212. Fiskum, G., Reynafarge, B., and Lehninger, A. L. (1979). J. Eiol. Chem. 254, 6288-6295. Glynn, I. M. (1967). Nature (London) 216, 1318-1319. Grinius, L. L., Jasaitis, A. A., Kadziakuskas, Yu. P., Liberman, E. A., Skulachev, V. P., Topali, V. P., Tsofina, L. M., and Vladimirova, M. A. (1970). Eiochim. Eiophys. Acta 216, 1-12. Hatefi, Y., and Galante, Y. M. (1978). In “Energy Conservation in Biological Membranes” (G. Schafer and M. Klingenberg, eds.), pp. 19-30. Springer-Verlag, Berlin and New York. Hinkle, P. C., and Horstman, L. L. (1971). J. Eiol. Chem. 246, 6024-6028. Jasaitis, A. A., Kuliene, V. V., and Skulachev, V. P. (1971). Eiochim. Eiophys. Acta 234, 177-181. Kagawa, Y. (1978). Eiochim. Eiophys. Acta 505, 45-93. Kamo, N., Muramatsugu, M., Hongoh, R., and Kobatake, Y. (1979). J. Membr. Eiol. 49, 105-121. Kell, D. B. (1979). Eiochim. Eiophys. Acta 549, 55-99. Klingenberg, M., Aquila, H., Kramer, R., Babel, W., and Feckl, J. (1977). In “Biochemistry of Membrane Transport” (G. Semenza and E. Carafoli, eds.), pp. 567-579. SpringerVerlag, Berlin and New York. Krab, K., and Wikstrom, M. (1978). Eiochim. Eiophys. Acta 504, 200-214. LaNoue, K. F., and Schoolwerth, A. C. (1979). Annu. Rev. Eiochem. 48, 871-922. Lawford, H. G., and Garland, P. B. (1972). Eiochem. J. 130, 1029-1044. Lehninger, A. L., Reynafarge, B., and Alexandre, A. (1979). In “Cation Flux Across Biomembranes” (Y. Mukohata and L. Packer, eds.), pp. 343-354. Academic Press, New York. Leung, K. H., and Hinkle, P. C. (1975). J. Eiol. Chem. 250, 8467-8471. Lorusso, M., Capuano, F., Boffoli, D., Stefanelli, R., and Papa, S. (1979). Eiochem. J. 182, 133-147. Mitchell, P. (1961). Nature (London) 191, 144-148. Mitchell, P. (1973). FEES Lett. 33, 267-274. Mitchell, P. (1975). FEES Lett. 50, 95-97. Mitchell, P. (1977). FEES Lett. 78, 1-20. Mitchell, P. (1979). Eur. J. Eiochem. 95, 1-20. Mitchell, P., and Moyle, J. (1967). Eur. J. Eiochem. 105, 1147-1162. Mitchell, P., and Moyle, J. (1968). Eur. J. Eiochem. 4, 530-539.
430
JANNA P. WEHRLE
Moyle, J., and Mitchell, P. (1973). FEBS Lett. 30, 317-320. Moyle, J., and Mitchell, P. (1977a). FEBS Lett. 73, 131-136. Moyle, J., and Mitchell, P. (1977b). FEBS Lett. 77, 136-140. Moyle, J., and Mitchell, P. (1978). FEBS Lett. 88, 268-272. Nelson, N. (1976). Biochim. Biophys. Acta 456, 314-338. Nicholls, D. G. (1974). Eur. J. Biochem. 50, 305-315. Ohnishi, T., and Pring, M. (1974). In “Dynamics of Energy Transducing Membranes” (L. Ernster, R. W. Estabrook, and E. C. Slater, eds.), pp. 169-180. Elsevier, Amsterdam. Pedersen, P. L., Amzel, L. M., Soper, J. W., Cintron, N. M., and Hullihen, J. (1978). In “Energy Conservation in Biological Membranes” (G. Schafer and M. Klingenberg, eds.), pp. 159-194. Springer-Verlag, Berlin and New York. Pullman, M. E., and Monroy, G. C. (1963). J. Biol. Chem. 238, 3762-3769. Racker, E. (1977). Annu. Rev. Biochem. 46, 1006-1013. Racker, E., Knowles, A. F., and Eytan, E. (1975). Ann. N.Y. Acud. Sci. 264, 17-33. Ragan, C. I. (1976). Biochim. Biophys. Acta 456, 249-290. Ragan, C. I., and Hinkle, P. C. (1975). J. Biol. Chem. 250, 8472-8476. Reynafarge, B., and Lehninger, A. L. (1977). Biochem. Biophys. Res. Commun. 77, 12731279. Reynafarge, B., and Lehninger, A. L. (1978). J. Biol. Chem. 253, 6331-6334. Rieske, J. S. (1976). Biochim. Biophys. Acta 456, 195-247. Rieske, J. S., Baum, H., Stoner, C. D., and Lipton, S. H. (1967). J. Biol. Chem. 242, 4854-4866. Robertson, R. N., and Boardman, N. K. (1975). FEBS Lett. 60, 1-6. Rosing, J., Kayalar, C., and Boyer, P. D. (1977). J. Biol. Chem. 252, 2478-2485. Rossi, E., and Azzone, G. F. (1970). Eur. J. Biochem. 12, 319-327. Rottenberg, H., (1978). FEBS Lett. 94, 295-297. Rottenberg, H. (1979). TIBS 4, N182-Nl85. Serrano, R., Kanner, B. I., and Racker, E. (1976). J. Biol. Chem. 251, 2453-2461. Shchipakin, V. Chuchlova, E., and Evtodienko, Y. (1976). Biochem. Biophys. Res. Commun. 69, 123-127. Sigel, E., and Carafoli, E. (1978). Eur. J. Biochem. 89, 119-123. Slater, E. C. (1977). Annu. Rev. Biochem. 46, 1015-1026. Sone, N., Yoshida, M., Okamoto, H., and Kagawa, Y. (1976). J. Membr. Biol. 30, 121-124. Sone, N., Yoshida, M., Hirata, H., and Kagawa, Y. (1977). J. Biol. Chem. 252, 2956-2960. Soper, J. W., Decker, G. L., and Pedersen, P. L. (1979). J . Biol. Chem. 254, 11170-11176. Sorgato, M. C., Ferguson, S. J., Kell, D. B., and John, P . (1978). Biochem. J. 174,237-256. Stigall, D. L., Galante, Y. M., and Hatefi, Y. (1978). J . Biol. Chem. 253, 956-964. Tedeschi, H. (1979). TIBS 4, N182-NI85. Thayer, W. S., and Hinkle, P. C. (1973). J. Biol. Chem. 248, 5395-5402. Thayer, W. S., and Hinkle, P. C. (1975a). J. Biol. Chem. 250, 5330-5335. Thayer, W. S., and Hinkle, P. C. (1975b). J. Biol. Chem. 250, 5336-5342. Tiefert, M. A., and Moudrianakis, E. N. (1979). J. Biol. Chem. 254, 9500-9508. von Jagow, G., Schagger, H., Engel, W. D., Hackenberg, H., and Kolb, H. J. (1978). In “Energy Conservation in Biological Membranes” (G. Schafer and M. Klingenberg, eds.), pp. 43-52. Springer-Verlag, Berlin and New York. Wikstrom, M. K. F. (1975). Nature (London) 266, 271-273. Wikstrom, M. K. F., and Krab, K. (1979). Biochim. Biophys. Acta 549, 177-222. Williams, R. J. P. (1961). J. Theor. Biol. 1, 1-13. Williams, R. J. P. (1975). FEBS Lett. 53, 123-125. Wrigglesworth, J. M., and Nicholls, P . (1979). Biochim. Biophys. Acta 547, 36-46.
CURRENT TOPICS IN MEMBRANES A N D TRANSPORT. VOLUME 16
Chapter 24
Electrogenic Reactions and Proton Pumping in Green Plant Photosynthesis WOLFGANG JUNGE Schwerpunkt Biophysik Universitat Osnabriick Osnabruck, Federal Republic of Germany
Introduction ............................................................ The Membrane .............................. ....................... A. Structure .................. c1ty ............................... B. Surface Charge Density and .............................. 111. Electrogenic Reaction Steps ..... A. Survey ..................... B. Chloroplast Electrochro C. Electric Generators ..... ................................. IV. Protolytic Reaction Steps ................... A. Survey ........................................................ B. Spectrophotometric Det pH-Indicating Dyes ...... C. Proton Pumps ...................................... V . Comments on the Pathway of Protons to the ATP S v1. Summary ..............................
I.
11.
....................,..................,..........
1.
..............
43 1 433 433 435 437 437 438 442 449 449 450 453 458 459 461
INTRODUCTION
Photosynthesis by green plants is the fundamental energy-providing process for terrestrial life. Energy from sunlight is used to synthesize carbohydrates from water and carbon dioxide, which serve as both fuel and substrates for plants, and indirectly for animals. The first relatively stable intermediates between light absorption and carbohydrate synthesis are re431
Copyright &, 1982 by Academic Press. Inc All rights of reproduction in any form reserved ISBN 0-12- I533 16-6
WOLFGANG JUNGE
432
duced nicotinamide adenine dinucleotide phosphate (NADPH) and adenosine triphosphate (ATP). The light-driven production of NADPH and of ATP, which is topologically separated from the subsequent “dark” reactions, occurs in the inner membrane system (thylakoids) of chloroplasts. Thylakoid membranes form disc-shaped vesicles (diameter 500 nm and thickness 10 nm) often arrayed in stacks (grana) with their internal phases extensively interconnected. The reduction of NADP+ occurs along a linear electron transport chain, driven by two photochemical reaction centers (in series), with water as the ultimate electron donor. Phosphorylation of ADP is indirectly coupled to electron transport, and-though the point was once highly disputedprotons are now generally accepted as the obligatory “high-energy” intermediate (see Boyer et al., 1977). Figure 1 illustrates the sequence of events in the thylakoid membrane: Absorption of light by chlorophylls and
-
-
1 THYLAKOID ( 2 lo5 chlorophylls I quanta
electrons
protons
products (NADPH,ATPl
AT P
loosely coupled antennae units
strongly coupled electron transport chains
300 chlorophylls
a total of at least 200, each wllh=105A2 and 600 chlorophylls
per photosystem I and U
common pool of electrochemical energy functional unit at least as big 0s one thylokoid
ATP synthases
at least 200, mobile on surface
FIG. 1 . Schematic representationof antennae function, electron transport, proton pumping, and phosphorylation in thylakoids from green plants. The rectangular arrangement of photosystems is arbitrary (Junge, 1977a).
24. ELECTRIC GENERATORS IN PHOTOSYNTHESIS
433
carotenoids-acting as antennae for two types of photochemical reaction centers (photosystems I and 11)-promotes the transfer of electrons from water to NADP+. Some of the redox reactions (e.g., the oxidation of water) are followed by proton release; others (e.g., the reduction of quinones) are followed by proton uptake. With electron transfer directed across the membrane as indicated, the thylakoid membrane becomes electrically charged, and protons are pumped into the internal space. The stored electrochemical energy is then used by a proton-translocating ATP synthetase whose properties and mode of action are summarized by Grilber in this volume. Under continuous illumination of thylakoid membranes, the linear electron transport chain gains approximately 1.2 eV as the free energy of the NADPH/NADP+ couple and another 0.2 eV at each of the two sites where protons are pumped into the internal space of thylakoids. The total gain of 1.6 eV requires the input of two light quanta. If these are derived from red light, which is sufficient to drive these processes (e.g., at a wavelength of 690 nm equivalent to 1.8 eV), the efficiency for the conversion of light energy into useful forms of energy is approximately 40% ,with one-quarter stored as the electrochemical potential of protons. (The normal practical efficiency, however, is an order of magnitude lower; Knox, 1979). The main purpose of this article is to review the molecular events-as they are currently understood-which connect redox reactions in the thylakoid membrane with the generation of an electrochemical potential difference for protons across this membrane. The subject has been reviewed previously by numerous authors, and the interested reader may refer, for example, to Witt (1975, 1979) and to Hauska and Trebst (1977).
II. THE MEMBRANE
A. Structure
In intact mature chloroplasts, the larger part of each thylakoid membrane is a disc-shaped bag having a typical diameter of 500 nm, an apparent membrane thickness of about 7 nm, and an internal aqueous phase which measures 10-20 nm thick in the electron microscope (see, e.g., Muehlethaler, 1977) but appears smaller-perhaps as thin as 5 nm-from distribution studies (Heldt et al., 1973) based on the surface density of chlorophyll (see Wolken and Schwertz, 1953; Thomas et al., 1956). In isolated broken chloroplasts the shape and internal volume of thylakoids are highly variable, depending on the salt composition and osmolarity of the suspending medium. Thicknesses of 15-120 nm can be calculated from the data of Gaensslen and McCarty (1971),
434
WOLFGANG JUNGE
Rottenberg et al. (1972), and Ort et al. (1976), assuming the basic disc shape to be preserved. With vesicles of such dimensions, the extent to which internal events can be treated as (averaged) bulk phase events depends upon the reach of electrostatic effects from the membrane surface into the inner aqueous phase. This reach diminishes in absolute size as the ionic strength of the solution increases and as the fixed charge density of the membrane falls; it diminishes in importance as the volume (thickness) of the inner phase increases. With chloroplasts suspended, for example, in 10 mMKC1 as the only salt and having a surface charge density of 1 pC/cm2 (Itoh, 1979a,b), the surface potential would be 39 mV and the Debye length would be 3 nm. If the internal phase were 15 nm thick, the average internal pH would be 0.2 units more acid than that of the medium, and the average internal K+ concentration would be 14 mM. Thus, the average ion concentrations in the internal phase would be reasonably close to those of the medium; but near the membrane-water interface much stronger deviations would occur. The thylakoid membrane, like the cristae membrane of mitochondria, has a relatively high protein content. Forty-seven percent of the total dry weight is ether-soluble (Wintermans, 1967), but this includes 10% pigments (chlorophylls and carotenoids), so the lipid content in a narrower sense must be less than 40%. And it could be much less, since only 10-15% of the total area (outer surface) was found susceptible to lipid antibodies (Radunz, 1979). This has the following consequences for the electrochemical behavior of the membrane: (1) Proteins will contribute to the surface charge density and to the proton-buffering capacity. (2) The membrane dielectric as well as the surface charge density will be highly granular. (3) The granularity will probably not be blurred by rapid lateral diffusion of the charge carriers, because the apparent microviscosity of the thylakoid membrane appears quite high, possibly caused by protein aggregation (see Wagner and Junge, 1980). The thylakoid membrane also appears complex in the direction normal to the plane, as indicated by two different compartmentation experiments. Auslander and Junge (1974) observed proton uptake from a region near the outer side of the membrane, which was shielded from the outer bulk phase by a proteinaceous barrier. And Quintanilha and Packer (1978) observed accumulation of an impermeant amphiphilic spin probe in a region of membrane only weakly influenced by the diffuse ionic double layer. Despite such complexities, we shall use the following simplified model of the thylakoid membrane throughout most of this article: A more-or-less homogeneous dielectric core (electric capacitance of 0.5- 1.O pF/cm2) is bounded by fixed negative charges (density 1-3 pC/cm2) which create a negative surface potential and a diffuse ionic double layer reaching into the
24. ELECTRIC GENERATORS IN PHOTOSYNTHESIS
435
adjacent water phases. There are five main compartments in such a model: bulk water, interface, membrane core, interface, and bulk water. In certain experiments, further subcompartments may need to be considered, as indicated above (see also Kell, 1979). B. Surface Charge Density and Buffering Capacity Because of their potential influence on surface reaction rates and on total energy storage in ionic gradients, two physical parameters which must be kept in mind during the discussions of electrogenesis in thylakoid membranes are the membrane surface charge and the proton-buffering capacity of the thylakoids. Table I summarizes recent determinations of surface charge density, calculated mostly via the Gouy-Chapman theory, from data on the influence of salt concentrations (in the internal or external bulk phases) on the apparent rates of reactions involving charged species [Fe(CN)a-, Fe(CN):-, and H +1. For a variety of reasons, the results must at present be taken as semiquantitative estimates, but they do make clear that the two surfaces of the thylakoid membrane are not widely different, both having values of 1-3 FC/cm2, or onenegative charge per 5-16nm2 of surface. (Theaverage area per chlorophyll molecule is 2.2 nm2). A comprehensive review of the role of surface potentials in photosynthesis has been presented by Barber (1981). Recent values of buffering capacity are summarized in Fig. 2. Curve 1, obtained from a slow dark titration, represents the internal and external phases taken together and agrees well with data obtained by other authors on a different chloroplast preparation (Junge and Auslander, 1974; Saphon and Crofts, 1977b). The lower three curves present results for the inner phase only, but obtained by three different laboratories using different methods and covering different p H ranges. There is no consensus yet as to whether curve 2 or 4 is the most satisfactory in the acid range. It is clear, however, that at all pH values the buffering capacity of the external phase dominates that of the internal phase. It is also clear that the buffer capacity in both phases is essentially constant above pH 6.5, which is most easily explained by the involvement of a large number of buffering groups with different pKvalues (i.e., protein buffering). The steep increase in buffering capacity at low p H indicates the existence of dissociable groups (lipids?) with pK values of about 5 (Walz el al., 1974; Mercer et al., 1955; Akerlund el al., 1979). It may be worthwhile to compare the absolute magnitudes of the electric and chemical storage capacities of the thylakoid membrane. In a single turnover the two proton pumps in the linear electron transport chain translocate two charge equivalents per lo3 nm2. Under the assumption of an electric capacitance of 0.5 pF/cm2 the generated voltage is 64 mV. How many turn-
TABLE I SURVEY OF DATAON THE SURFACE CHARGE DENSITY ON BOTHSmEs OF THE THYLAKOID MEMBRANE' Surface charge density bC/cm2) Outer Side
Inner Side
Locus
-1.1, -1.5
-
PS I, reducing side
- 1.3
-
PS 11, reducing side
-2.5
-
-0.77, -0.88
PS 11, P680-Q PS I, oxidizing side
-
- 0.59 -1.3
Same as above Average inside
-
- 3.2
Between PS I1 and PS I
(I
P S I , Photosystem I; PS 11, photosystem 11.
Experimental technique Rate of electron flow from PS I to ferricyanide Rate of electron flow from PS I1 to ferricyanide in the presence of DCMU Chlorophyll fluorescence Rate of electron flow from ferrocyanide to PS I in sonicated chloroplasts Same, but in PS I particles Indirect argument, based on dark-light redistribution of K+ and Cl- between inside and outside Reaction rate between PQH, and P700
Reference Itoh (1979b) Itoh (1978)
Barber (1 977) Itoh (1979b)
Itoh (1979b) Rumberg and Muhle (1976)
Huber and Rumberg (1979)
437
24. ELECTRIC GENERATORS IN PHOTOSYNTHESIS
FIG. 2. Proton-buffering capacity of thylakoids as functions of the pH. 1 , Lettuce chloroplasts, dark titration experiment, both phases (Walz et al., 1974); 2, lettuce chloroplasts, internal phase only (Walz et al., 1974); 3 , spinach chloroplasts, internal phase only (Junge et at., 1979); 4, spinach chloroplasts, internal phase only (Reinwald, 1970).
'.. 2 00
overs are required to create the energeticallyequivalent pH difference? With a specificinternal buffering capacity of 0.18 mole H+/mole chlorophyll per pH unit (Junge et al., 1979)and 5 0 0 chlorophylls per electron transport chain, a minimum of 49 turnovers is calculated. Hence, in dynamic situations, transients of the voltage occur much faster than those of the pH difference. As thylakoids operate their ATP synthesismainly on the pH difference, they are well buffered and protected against fluctuations of the illumination level in a time domain of several seconds. 111.
ELECTROGENIC REACTION STEPS
A. Survey
Although electron and proton transfer reactions may be intrinsically electrogenic, only under three conditions will they noticeably contribute to the electric potential difference across the thylakoid membrane: (1) The reaction path must not be directed parallel to the plane of the membrane; (2) the reaction must cross a considerable distance of the dielectriccore of the membrane; and (3) the transported charge must not be locally compensated (as occurs, for example, when electrons and protons are transported together on plastoquinone). Of the three electrogenic reaction steps (see below) for which there is evidence in the thylakoid membrane, two are linked to the primary
438
WOLFGANG JUNGE
photochemical reactions in photosystems I and 11, respectively, and the other seems to be linked to thermally activated electron transfer in the neighborhood of photosystem I. Rise of the membrane potential via the first two steps (or sites) is extremely rapid (at least nanoseconds) following brief flash excitation, but is much slower (milliseconds)via the third site. In a single turnover, each site contributes about 25 mV to the membrane potential. While the voltage is primarily generated by electron transfer, under steady operation the interplay of redox reactions with protolytic reactions makes it appear as if inwardly directed proton transfer were electrogenic. Most of the information available on these electrogenic reactions in thylakoids was obtained by studying electrochromic absorption changes of intrinsic membrane pigments, and it is therefore necessary to discuss this technique in some detail before passing on to the experimental results themselves.
6. Chloroplast Electrochromism as a Molecular Voltmeter In recent years a variety of dye substances have been developed and tested for their ability to measure biological membrane potentials via absorption or fluorescence shifts (Waggoner, 1976). As extrinsic probes, these substances respond to voltage changes with half-times ranging from tens of seconds to microseconds and function by several different molecular mechanisms. However, the thylakoid membranes and other photosynthetics organelles (see, e.g., Dutton et al., this volume) are intrinsically highly pigmented and offer the unique opportunity to monitor voltage-sensitiveabsorption changes in the resident native pigments. Observed changes occur with (yet instrumentlimited) half-times of nanoseconds and appear to be genuine electrochromic shifts, as originally defined by Platt (1961). Fundamentally, a chromophore oriented in an intense electric field [commonly in biological membranes, 100 mV/nm ( = lo7 V/m)] can respond to a variation in field strength by a variation in the resonant wavelength for light absorption. [For a theoretical discussion see Liptay (1969) and Reich and Sewe (1977).] Such effects tend to be small in relation to the normal bandwidth for the chromophore and so must ordinarily be followed with a differential instrument which can be arranged to show either the shift in absorption peak itself or a change in absolute absorbance at a given wavelength. Initially, the obvious advantages of this kind of technique were partly offset by difficulties in discriminating genuine electrochromic changes from apparent absorption changes of spurious origin. Accumulated evidence defining and confirming the electrochromic effects in chloroplast pigments is summarized below.
439
24. ELECTRIC GENERATORS IN PHOTOSYNTHESIS
1. KINETICS
Junge and Witt (1968) observed that, for certain absorption changes which appear when chloroplasts are excited by brief flashes, decay (following each flash) is accelerated by manipulations which increase the ionic conductivity of biological membranes: aging, osmotic shock, organic solvents, and iontransporting antibiotics. This is illustrated in Fig. 3 for absorption changes (A A ) measured at 520 nm and responding to a 10-psecflash of light (delivered at time zero). In all three traces the rise of A A occurred exceedingly rapidly, essentially vertically on this millisecond time scale. For chloroplasts suspended in the normal buffer solution, containing sodium and potassium, the absorption signal decayed by 25% in 20 msec following the flash. Decay was hardly affected by addition of the potassium carrier valinomycin, provided no potassium was present (Fig. 3B), but was greatly speeded by valinomycin in the presence of potassium (Fig. 3C). In similar experiments with the channelforming antibiotic gramicidin, a single molecule per lo5 chlorophyll molecules (the average number for spinach thylakoids) yielded substantial acceleration of the decay, confirming that the field to which the absorption change responds must be delocalized over the entire thylakoid (in the time domain of a millisecond). .Na*
a
l
.K'
l
20
0 time [msec] volinommrch +No'
U'
.No*
+K*
20
0 time [msec]
tim [mmsed
FIG.3. Time course of theelectrochromic absorption changes at 520nm, observed after excitation of a chloroplast suspension with a short flash at t = 0. Increased ionic conductance, as by potassium in the presence of valinomycin, accelerates decay of the electrochemic absorption change (Junge and Schmid, 1971).
440
WOLFGANG JUNGE
2. SPECTRUM
The difference spectrum of the absorption changes, which are sensitive to the ionic conductivity of the thylakoid membrane (Emrichet al., 1969), can be synthesized from the separate electrochromic difference spectra of carotenoids and chlorophylls, measured in microcapacitors (Schmidt et al., 1970). Two main peaks, at 428 and 520 nm, arise from a minor carotenoid fraction (<20%:DeGrooth et al., 1980; Schlodder and Witt, 1980), probably luteins complexed by chlorophyll b (Sewe and Reich, 1977a). Apparently it is formation of this complex which, by the prepolarization of lutein, produces a pseudolinear relation between the absorption shift and the membrane electric field. Pseudolinearity (at 520 nm) has been confirmed by several independent techniques (Reinwald et al., 1968; Witt and Zickler, 1974; Amesz and DeGrooth, 1975; Schapendonk and Vredenberg, 1977). 3. ARTIFICIALLY IMPOSED POTENTIALS
In bacterial chromatophores electrochromic absorption changes have been most convincingly demonstrated by exposing chromatophore vesicles to salt jumps in the presence of cation-specific ionophores (Jackson and Crofts, 1969). Despite the much greater light-scattering changes usually observed with thylakoids (Strichartz and Chance, 1972), recent experiments have demonstrated the same effect on thylakoids (Schapendonk and Vredenberg, 1977). In such experiments the absorption change is linearly proportional to the calculated (Nernst-Planck) diffusion potential, so that the technique yields a calibration of the flash-induced electrochromic change. Electric field transients across the thylakoid membrane generated via macroscopic electrodes, in thylakoid suspensions, can produce absorption shifts which duplicate those produced by light (Ellenson and Sauer, 1976; DeGrooth et al., 1980; Schlodder and Witt, 1980). Although, for avariety oftechnical reasons, this technique cannot be used to calibrate the light-induced absorption shifts, it can be used to extend the range of voltages applicable to the thylakoid up to about 500 mV (see Graber, this volume). 4. LIMITATIONS
It is evident that electrochromic absorption changes (mainly at 5 18 and 478 nm) are useful primarily as indicators of the electric field strength in the thylakoid membrane, with high time resolution. Limitations of this molecular voltmeter arise from four main sources. The most serious of these is scattered light, which can easily be mistaken for absorption transients. Even double-beam spectrophotometry cannot wholly correct this situation, which becomes especially severe in studies with long illumination times. Scattering changes do seem to be negligible, however, when the chloroplasts are illu-
24. ELECTRIC GENERATORS IN PHOTOSYNTHESIS
441
minated with brief flashes at a low repetition rate. Under certain conditions, particularly at high flash energies, large absorption changes not related to electrochromismcan occur in the wavelength domain around 520 nm. These can be distinguished from the electrochromic response by higher saturating light levels, more rapid decay (at least 1000-fold), and persistence in heatinactivated chloroplasts. It has been attributed to metastable carotenoid triplets (Mathis, 1970; Wolff and Witt, 1969; Witt and Wolf, 1970). Field nonuniformity, in either a lateral or a normal direction, may lead to the misinterpretation of electrochromic changes. Lateral nonuniformities can arise, for example, as a result of the membrane granularity (see above) and the fact that the detecting carotenoids are associated primarily with photosystem I1 (Sewe and Reich, 1977a;DeGrooth etal., 1980;Schlodder and Witt, 1980), whereas two of the three electrogenic steps are associated with photosystem I. If the charge separation process is punctate or regional, and distant from the detecting carotenoids, then the transmembrane potential difference will be attenuated between the source and the detector. This kind of nonuniformity could become serious either when the electrical conductivity of the internal and external bulk phases is very low (as under cryogenic conditions; see Amesz and DeGrooth, 1975; Vermeglio and Mathis, 1974; Conjeaud et al., 1976) or when the thylakoid is nonuniformly activated (e.g., by an extrinsic electric field. Field nonuniformity, in a direction normal to the membrane, must always exist to a certain extent, because of surface potentials. Presumably, the electrochromic pigments monitor the field strength within the membrane, and this must be related to the sum of the difference of the bulk phase potentials plus the difference of surface potentials. Electrochromic shifts, therefore, could be distorted by asymmetric changes in membrane surface potentials. Most of our experiments have been designed to minimize this kind of error, by using brief light flashes at a low repetition frequency. p h e change in surface potential occurring for displacement of two protons (a single-turnover flash with both photosystems active) should be less than 2mV, as long as bulk phase monovalent electrolytes are in excess of 3 mM.1Under continuous illumination, the problem can become serious, however (Witt, 1979). Finally, it should be mentioned that use of the electrochromic decay time course as a measure of the membrane’s ionic conductivity is subject to the usual restrictions of uniformity-in size and specific transport properties-which must be applied to any ion flux measurement for vesicular populations (see Schmid and Junge, 1975). 5 . ALTERNATIVE TECHNIQUES
In recent years several other techniques have been developed for estimating the membrane potential in thylakoids under various conditions. Although
442
WOLFGANG JUNGE
some quantitative discrepancies exist between results obtained by these other techniques and results obtained from electrochromic shifts, the qualitative pictures are generally in agreement. Fowler and Kok (1972, 1974) and Witt and Zickler (1973, 1974) used macroscopic electrodes, immersed in chloroplast suspensions, to detect bulk charge displacement resulting from nonuniform illumination. The latter authors were, in fact, able to use the technique to confirm that electrochromic signals in the thylakoids are linear indicators of the electric field strength. Microelectrodes have been used by Bulychev, Vredenberg, and their collaborators (Bulychev el al., 1972; Vredenberg and Tonk, 1975; Bulychev and Vredenberg, 1976) to record lightgenerated voltage transients-probably partially shunted (both in resistance and capacitance)-in giant chloroplasts from Peperomia metallica. Delayed fluorescence from thylakoids, following a flash, has been shown to depend on the membrane potential, as determined either by artificially induced diffusion potentials (Barber and Kraan, 1970; Wraight and Crofts, 1971) or by externally imposed fields (Ortoidze el al., 1979). Finally, Trissl (1980) and Trissl and Graber (1980) have observed rapid photovoltages (via macroscopic electrodes) on thylakoid membranes spread at a hexane- water interface. Though this technique is restricted to artificial membrane structures, it yields a very high signal-to-noise ratio and a very high time resolution.
C. Electric Generators 1 . ELECTROGENIC STEPS IN THE LINEAR ELECTRON TRANSPORT CHAIN
In both photosystems (Schliephake et al., 1968; Malkin, 1978) it appears to be the primary photochemical act-electron transfer from the donor molecule to the primary acceptor-which takes place across the dielectric core of the thylakoid membrane. At least this is the simplest conclusion which can account simultaneously for the speed and temperature insensitivity of the initial electrogenic process observed on flash excitation. Delayed fluorescence, which is believed to represent reversal of the photochemical reactions, has been shown by several different laboratories (Barber and Kraan, 1970; Wraight and Crofts, 1971; Ortoidze et al., 1979) to be sensitive to the potential difference imposed across the thylakoid membrane. Under flash stimulation by a Q-switched ruby laser, intact thylakoids show an instrument-limited risetime of 20 nsec for the electrochromic absorption change; and this can be shortened to 2 nsec by spreading the thylakoid membranes at a hexane-water interface (Trissl and Graber, 1980 ). Measurable shifts in electrochromic absorption survive at temperatures down at least to -50°C (Amesz and DeGrooth, 1975; Mathis and Vermeglio, 1975) and, although the signal following stimulation of photosynthesis I disappears at still lower
24. ELECTRIC GENERATORS IN PHOTOSYNTHESIS
443
temperatures (- 125OC; Conjeaud et al., 1976), this should probably be attributed to field attenuation (hindered delocalization) between photosystem I and the detecting carotenoids located near photosystem I1 (Section III,B,4). The polarity of the light-generated electric potential difference is positive toward the internal aqueous phase, as demonstrated both by macroscopic electrodes (Section III,B,5) and by ionic redistributions (Deamer and Packer, 1969; Schroder et al., 1972; Hind et al., 1974). All evidence presently available indicates that none of the other electron transfer steps in the linear electron transport chain contributes appreciably to the transmembrane electrogenesis; in particular, there is no component of the electrochromic response which kinetically matches the restoration (rereduction) of the primary electron donor at photosystem I, P700 (Junge, 1972). ELECTRON TRANSPORT 2. ELECTROGENIC STEPSIN CYCLIC Recently, several laboratories (Velthuys, 1978,1979; Slovacek et al., 1979; Crowther et a/., 1979;Horvath et al., 1979; Bouges-Bocquet, 1980)haveidentified a slowly rising (milliseconds)change in absorption which appears to be genuinely electrochromic but is associated with an electrogenic step other than electron transfer. Apparently this change is observed only in freshly isolated chloroplasts which have been prepared with the outer membrane and stroma intact, retaining soluble ferredoxin. Figure 4 illustrates the slow absorption change superimposed on the fast changes due to photosystems I and I1 (control). Blockage with 3’ -(3,4-dichlorophenyl)-l’,l’-dimethylurea (DCMU) abolishes all the light-induced absorption changes, but release of photosystem I with dithionite restores part of the fast component and all of the slow component. Thus, photosystem I must provide the driving force for the slow electrochromic change (Crowther et al., 1979). Further experiments have shown that the redox state of the plastoquinone pool-and, thereby, photosystem 11-has an important regulatory effect on the slow change. Although there is not yet complete agreement on the origin of the slow electrogenesis, it is possible to synthesize a tentative scheme based on the combined work from the laboratories of Velthuys, Crowther, and Horvath. One such scheme is shown in Fig. 5 (lower diagram), complementary to a simplified scheme for the fast electrogenesis. The main point is that the slow electrogenic reaction is part of an electron-hydrogen loop linked to linear electron flow via the plastoquinone pool. A prerequisite for the slow reaction is reduction of cytochromef(Velthuys, 1978),and it seems as if electron donation by the two-electron donor, plastohydroquinone (PQH,), to the oneelectron acceptor, cytochrome f,creates the driving force for an additional one-electron step, probably from plastosemiquinone (PQ-) to a 6-type cytochrome. It is this step which probably takes place across the dielectric core
444
WOLFGANG JUNGE
+
IJ I
DCMU
+
DITHIONITE
I
I
I
I
0
10
20
30
I
LO ms
time FIG.4. Time course of the electrochromic absorption changes at 518 nm, during linear and pseudocyclic electron transport. (Top) Both photosystems and the pseudocyclic path active. (Middle) Both photosystems blocked by DCMU. (Bottom) Reactivation of photosystem I with the artificial reducing agent, dithionite (Crowther et al., 1979).
of themembrane. Near theouter surfaceof themembrane, theelectron would again be transferred to plastoquinone to yield the semiquinone. The regulatory effect of photosystem I1 could be accounted for by the fact that only fully reduced and protonated plastoquinone molecules can pass inward through the membrane, so that a source of additional electrons is required for the system to continue running. As suggested by Slovacek e t a / . (1979), the acceptor chain of photosystem I might also supply the needed electrons in a reaction involving soluble ferredoxin. [The situation may be analogous to that in blue-green algae (Knaff, 1977), where cytochrome b6is reduced by ferredoxin and oxidized via an ADP-sensitive step that probably involves proton pumping via plastoquinone.] 3. THEMAGNITUDE OF THE ELECTRIC POTENTIAL DIFFERENCE
a. Excitation with Single-Turnover Flashes. Early estimates of the magnitude of the electric potential difference arising from a single turnover of
445
24. ELECTRIC GENERATORS IN PHOTOSYNTHESIS
Outer aqueous phase H'
I Fd-ACC
-R -P
y
e-~-b-P~30
~ P O H;PO ~
~-
S,) P 680
(So
/ l
PO H2-PC-P700
H20
Inner aqueous phase
1 H'
H' U E
H
c
in
:: c
r R
Outer phase
7 Td -ACC
POH,-CYT-F
P700
Inner phase 2H'
FIG.5 . Schematic presentationof proton pumping andgenerationof anelectric potential difference under linear electron transport (top) and cyclic electron transport (bottom). Double arrows, electrogenic transfer, outwardly directed; dashed arrow, electron path responsible for the slowly rising component of the electric potential (see Fig. 4). P680, P700, reaction center chlorophylls, designated by the characteristic absorption peaks (in nanometers); PQ, PQH2, pool plastoquinones in various oxidation states; Q-R, bound plastoquinone couple; PC, plastocyanin; Cyt-F, cytochromef; A, A2- P430: acceptor chain; Fd, ferredoxin, iron-sulfur center; ACC, terminal electron acceptor; So, . . . ,S3, redox states of the water-oxidizing enzyme.
-
the photosystems were based on assumptions about the average membrane area occupied by each reaction center and about the electric capacitance of the thylakoid membrane (Schliephake et al., 1968). Allowing 220 A2 per chlorophyll, 600chlorophylls per pair of reaction centers, and 0.5 pF/cm2 (the generally accepted value for ordinary planar lipid bilayer membranes) gives
446
WOLFGANG JUNGE
48 mV for the membrane potential from a single short flash. Microelectrode measurements on intact giant chloroplasts from P . rnetallica have recently given values near 40 mV (Bulychev and Vredenberg, 1976). Zickler et al. (1976) have discussed a possible underestimation due to impalement shunting by this technique. Although this is certainly no valid objection for kinetic arguments, it is still to be tested whether capacitive mismatching of the detector amplifier causes voltage readings that are too low (argument by Dr. Trissl). Attempts have also been made to calibrate the light-induced electrochromic changes by means of salt jumps in the presence of valinomycin (with bacterial chromatophores, Jackson and Crofts (1969); with chloroplasts, Schapendonk and Vredenberg (1977), taking care to minimize lightscattering changes. By this means an apparent maximal single-turnover voltage of 58.5 mV was obtained for chloroplasts, which reduced to 35 mV under certain assumptions about the permeability ratio for K + and C1-. Much higher estimates, 105-135 mV, have been reported by Zickler et al. (1976) based on the action of a voltage-gated, channel-forming antibiotic, alamethicin. In view of the indirect nature of all these estimates, it is difficult to favor one over another, but a figure near 50 rn Vseemsprobable. As already mentioned (Section 111,B,4), changes in membrane surface potential accompanying a single-turnover flash should not seriously compromise this value. b. Excitation with Continuous Light. Under conditions of steady, saturating illumination, the electric potential difference measured across the thylakoid membrane becomes a function of the method of measurement. Methods which indicate the bulk phase potential difference-such as redistribution of permeant ions (Rottenberg et al., 1972; Schroder et al., 1972) or microelectrodes (Bulychev and Vredenberg, 1976)-give values near 10 mV (thylakoid interior positive); but methods which monitor the intramembranal field-such as delayed fluorescence (Barber, 1972) or the electrochromic absorption changes (Graber and Witt, 1974)-indicate values 10-fold greater, 100 mV. This discrepancy, which is highly reproducible, can be taken as evidence of a strong asymmetry of membrane surface charge, produced as a result of sustained electron transport and proton pumping (Rumberg and Muhle, 1976; Rumberg, 1977; Witt, 1979). The following time course of events can be constructed following the onset of a step illumination. Initially, the bulk phase potential difference rises to a high value, two- to fourfold that observed with short-flash illumination. Under the influence of this membrane potential, a massive ion exchange begins, with cations driven outward and anions inward. As the internal phase becomes increasingly acidic, protons themselves take over an increasing proportion of the outward current driven by the field (see Graber and Witt, 1976), and the steady state is reached when proton efflux equals the active inward
-
24. ELECTRIC GENERATORS IN PHOTOSYNTHESIS
447
proton pumping. The high effective conductivity for protons, conjoined with a negative feedback effect upon the velocity of electron transport, caused by internal acidification (up to 3 pH units), leads to a progressive decrease in the electric portions of the bulk-phase potential difference. At the onset of illumination the rapid rise in the electric potential and the much slower rise in the chemical potential of the proton reflect the difference between the electrical and the chemical buffering capacities of the membranes, as pointed out in Section II,B. The specific proton uptake under steady illumination of thylakoids (in the absence of uncoupling agents) amounts to 0.2-0.5 H+ per chlorophyll. Since protons are the only actively pumped species, and since there is no evidence for coupled proton-cation exchange or proton-anion cotransport across the thylakoid membrane, all other ionic species can be assumed to adjust their respective concentrations in the bulk phases according to the Nernst-Planck The quasi-steady state value of the relationship: ZF A$ = RT In (Cin/Cout). electric potential difference between bulk phases, however, will depend on the concerted action of all permeant species. The most limiting ionic species in this process is likely to be that present at the largest internal concentration. As an example, in a typical experiment with isolated swollen chloroplasts (internal aqueous volume 50 liters/mole chlorophyll) the KCl concentration of the suspending medium is 30 mM. Under dark conditions the internal concentration of KCl will also be 30 mM, and under steady, saturating illumination, a maximum of 0.5 K+ ions per chlorophyll will be driven out, which-with 0.02 mole chlorophyll/liter-will decrease the internal K+ concentration at most by 10 mM. The Nernst-Planck potential for a K+ concentration ratio (inside/outside) of 20:30 is 10 mV, interior positive, in close agreement with the steady state values actually observed by methods which measure the bulk phase potential difference.
=
4. ELECTRICAL CONDUCTIVITY OF THE THYLAKOID MEMBRANE
A reasonable estimate for the electrical conductivity of the thylakoid membrane can be obtained from the decay of electrochromic absorption changes following a light flash (Fig. 3). Typically, the time constant for decay ( 7 ) ranges between 10 and 100 msec, depending on the quality of the chloroplast preparation. Assuming a specific capacitance of 0.5 pF/cm2, then, gives a to 2 x 1O-j S/cm2 for the membrane conductivity. range of 2 x Under flashing light or moderate illumination, the conductivity is dominated by K+ ,C1-, and Mgz+,not by protons, as is evident from the slow relaxation of pH differences across the thylakoid membrane ( = 5- 10seconds; Junge and Auslander, 1974). Isolated chloroplasts display only poor ion selectivity (Barber, 1972), so that ionic control of conductance depends
WOLFGANG JUNGE
448
strongly on the composition of the suspending medium (Dilley and Vernon, 1965; Hind et al., 1974; Schr6der et al., 1972). In intact chloroplasts, Mg2+-whichis the most abundant ion (2.33 pmoles/mgchlorophyll; Barber, 1977)-dominates conductivity of the thylakoid membrane and counterbalances most of the pumped protons (Barber et al., 1974; Krause, 1973; Portis and Heldt, 1976). With steady, saturating illumination and the consequent acidification, an outward proton current (probably flowing through the phosphorylation coupling factor) dominates the membrane conductivity (Graber and Witt, 1976). 5. STRUCTURAL ASPECTS OF THE ELECTROGENIC REACTIONS
Primary charge separation is directed across the thylakoid membrane. In photosystem I, the electron donor is a special pair of chlorophyll a molecules (absorbing at 700 nm and named P700) located near the internal membrane-water interface. This location has been inferred from the polarity of the light-generated electric field, plus the observation that the reduction of P700 by plastoquinone via plastocyanin is not electrogenic (Junge, 1972). (Plastocyanin is a water-soluble protein known to be located inside the thylakoid, near the inner surface of the membrane.) The fact that the potential difference indicated by electrochromic absorption shifts can occur in at least 2 nsec (Section III,C,l; Trissl and Graber, 1980a) suggests strongly that the primary electron acceptor must reside close to the outer surface of the membrane. The primary acceptor is possibly another pair of chlorophyll a molecules which, according to Shuvalov et al. (1979) can be reduced in less than 60 psec. At present, however, the time resolution of the electrochromic measurements is insufficient to permit spatial localization of the individual members of the acceptor chain (see Sauer et al., 1978). For photosystem 11, studies of the emission of delayed fluorescence have suggested that one of the later acceptors is located 25 from the primary electron donor (Ortoidze et al., 1979), within a single protein complex (Fig. 10). [The probable subunit composition of the protein which contains the primary electron donor in photosystem I, as described by Bengis and Nelson (1975), is also shown in Fig. 10, along with one possible spatial arrangement. They axes of the two chlorophyll a rings (Breton, 1976; Junge and Eckhof, 1974),and probably also thexaxes (Junge and Schaffernicht, 1979),are tilted only slightly out of the plane of the thylakoid membrane.] It is not yet clear in molecular terms how electrons can be rapidly channeled, or “tunneled,” across the thylakoid membrane. Since neither the chlorophyll rings themselves nor the &carotenes associated with the photosystem I complex are significantly inclined to the plane of the thylakoid membrane (Junge et al., 1977; Junge and Schaffernicht, 1979), there is no identifiable “molecular wire” which spans the thickness of the membrane. A
a
24. ELECTRIC GENERATORS IN PHOTOSYNTHESIS
449
proposal has been made (Tributsch, 1972)that the interior of the protein complex acts like an injection semiconductor, so that the force driving photoinjected electrons across the membrane is exerted by asymmetric surface potentials (Duniec and Thorne, 1979). Unfortunately, both the direction of charge asymmetry required for this mechanism and the predicted decrease in driving force with increased ionic strength of the medium are incorrect (Table I and Section III,C,3,b, and Fig. 7 of Trissl and Graber, 1980b). It is more reasonable at present to suppose that electron conduction occurs across a series of large n-electron systems, as is known from studies on photoprocesses in model membranes (Mange1et al., 1975)and monolayer assemblies (Kuhn, 1979; Moebius, 1979), but it must be admitted that experimental support for this picture is thus far inadequate.
6. LOCALIZED VERSUS DELOCALIZED ELECTRIC FIELD Physical considerations argue that the act of photochemical charge separation must generate localized dipole fields between electron-hole pairs on a picosecond time scale, and that ionic conduction in the adjacent aqueous phases must distribute this field more evenly over the membrane capacitance but on a longer time scale. By means of macroscopic electrodes (Section III,C,5), Witt and Zickler (1973) were able to show that only about 10 psec is required for the electric field to “homogenize” over the entire membrane of a single thylakoid. The process can be greatly slowed, however, by lowtemperature treatment of intact thylakoids (- 125”C, see Section III,C, 1; Vermeglio and Mathis, 1974; Conjeaud et al., 1976), because the ionic conductivity of the internal and external phases becomes exceedingly small at such temperatures.
IV.
PROTOLYTIC REACTION STEPS
A. Survey Figure 5 diagrams how protolytic reactions can be linked to electron transfer. With pure linear electron transport (upper diagram), two protons per electron are transferred from the external to the internal aqueous phase of the thylakoids. Of the two sites for proton uptake from theouter phase, oneis associated with the terminal electron acceptors (ACC) and the other is associated with the reduction of a special quinone which is probably also active during pseudocyclic electron transfer (lower diagram). Of the two sites for proton release into the internal phase, one is associated with the oxidation of water (i.e., the primary electron donor to photosystem 11), and the other is
450
WOLFGANG JUNGE
associated with oxidation of the PQH,, which again may be active during pseudocyclic electron transfer. Present evidence indicates that the linkage between proton uptake or release and the redox reactions is trivial, so that for chloroplasts-unlike mitochondria and halobacteria-there has not been a need to invoke special proton pumping action by the proteins. Single-turnover flashes at low frequency produce only very small p H changes inside the thylakoid ( - 0.05 p H units), and approximately 99.99% of the pumped protons are buffered away. With continuous illumination, however, the influx of protons and concomitant efflux of K+ and Mg2+ (driven by the potential difference) cause a pH drop of more than 3 units, which becomes the major driving force for photophosphorylation (see Graber, this volume). Still larger pH differences between the internal and external bulk phases are prevented by a negative feedback effect of the internal pH on the rate of electron transfer. At present, a major controversy concerns the pathway that protons take, from the “sources” on the internal side of the thylakoid membrane into the ATP synthetase. In order to resolve this controversy, it has been necessary to devise molecular probe strategies for proton subcompartmentation. Our most accurate information has come from spectrophotometry with pHindicating dyes, which have proved to be highly resolving both in kinetic and spatial studies. Once again, it is necessary to discuss some aspects of the technique before presenting the experimental results. B. Spectrophotometric Detection of pH Changes with pH-Indicating Dyes Dyes used to measure p H changes in suspensions of cell organelles, over the physiological pH range and with minimal disturbance of biological activity, include bromcresol purple (pK=6; Chance and Mela, 1966), neutral red (pK= 6.6; Lynn, 1968; Auslander and Junge, 1975), bromthymol blue (pK= 6.8; Schliephake et al., 1968), phenol red (pK= 7.3; Schroder el al., 1972), and cresol red (pK=7.9; Junge and Auslander, 1974). These dyes distribute over the aqueous phases and membrane spaces according to their solubility and respond to pH changes by easily measured absorption changes. However, straightforward interpretation of the absorption changes can be hampered by several complications, including overlapping absorption shifts in intrinsic pigments and unknown distribution among different internal compartments. The compartmentation problem has many possible ramifications: Individual compartments may show different p H changes in response to experimental manipulation; steady state dye concentrations in separate compartments may be unequal (they will be unequal if compartment p H
24. ELECTRIC GENERATORS IN PHOTOSYNTHESIS
45 1
values are dissimilar); apparent dye pK values may differ between compartments, as observed, for example, with dyes bound to membranes or proteins (Fernandez and Fromherz, 1977); and large pH changes will cause massive redistribution of the dye. A fortuitous combination of circumstances has allowed pH-indicating dyes to be used quite satisfactorily for short-flash studies, where the abovecomplications can be minimized. The technique to be described, however, does not carry over well either to studies with continuous illumination (Pick and Avron, 1976; Siefermann-Harms, 1978) or to experiments on organelles other than thylakoids (Schnetkamp et al., 1980). A dye such as cresol red, which is of low lipid solubility at the ambient pH, reports mainly the pH in the outer aqueous phase. The reason for this is probably not the impermeability of the thylakoid membrane to such indicators but the small size of the total internal aqueous volume of thylakoids relative to the outer volume (the ratio is smaller than in a typical experiment). On the other hand, a dye such as neutral red resides in the membrane to a large extent and reports p H changes from both phases. By combining such indicator dyes with the use of appropriate penetrating or nonpenetrating buffers, it is possible to define unequivocally the phase of predominant p H shift following flash excitation, as demonstrated in Fig. 6. In this case bovine serum albumin (BSA) has been used as a nonpenetrating buffer and imidazole as a penetrating buffer. In the absence of dye (Fig. 6A), the light flash produces a slight change in absorbance due to the intrinsic chloroplast pigments (Junge et al., 1979); this intrinsic response must be substracted from that observed in the presence of dye. When dye is added (Fig. 6B), the flash produces absorption changes both with neutral red and with cresol red. External buffering (with BSA, Fig. 6C) eliminates the response of cresol red, but full buffering (BSA plus imidazole, Fig. 6D) is required to abolish the response of neutral red. Thus, cresol red, without buffer, can be used to estimate p H changes in the external aqueous phase; and neutral red, with strong BSA buffering, can be used to estimate flash-induced p H changes in the internal aqueous phase of thylakoids (Auslander and Junge, 1975). The possibility that the dyes and buffers might have spurious biological effects (because of direct interaction with proteins or lipids) which could be misinterpreted has been eliminated by showing that 10 different buffers diminish the flash-induced change in internal pH in directproportion to their calculated buffering capacity (Junge et al., 1979). The same experiments have also indicated that the “internal” pH measured by neutral red is in fact that of the internal aqueous phase, not some restricted compartment within the membrane itself, since even very hydrophilic buffers-once allowed to penetrate the thylakoids-diminished the flash-induced absorption change in neutral red. An absolute calibration curve for the flash-induced p H changes inside
452
WOLFGANG JUNGE 575nm
52Lnm
A ..
+indicator neutral redlcresol red
T
m
+indicator + BSA
+indicator + BSA +imidazole
0
0
400ms
LWmS
time FIG.6. pH-indicating absorption shifts of dyes in a chloroplast suspension. (Left) Neutral red, which distributes to both internal and external phases. (Right) Cresol red, which is restricted to the external bulk phase. Bovine serum albumin (BSA) is used as a nonpermeating buffer, and imidizole as a permeating buffer. Valinomycin was added to virtually eliminate electrochromic absorption changes at 524 nm (Junge et at., 1978a).
thylakoids can be obtained in (Fig. 7) by plotting the magnitude of the absorption change in neutral red against the pH of the medium, under the condition that flashes are spaced far apart (here, 33 seconds) to avoid accumulation of protons. The resultant plot resembles the theoretical sensitivity of a pH indicator having an apparent pKof 7.25 (solid curve in Fig. 7). The alkaline shift in pK, from the 6.6 expected in aqueous solution, is consistent with the model studies of Fernandez and Fromherz (1977) and can be taken as evidence that the indicating molecules are actually bound to the negatively charged thylakoid membrane. The inset in Fig. 7 demonstrates incremental absorption changes in neutral red obtained when flashes are fired at short intervals
24.
453
PH FIG.7. Calibration of pH-indicating absorption changes in neutral red by means of flashinduced changes over a range of external pH values. The solid curve is a calculated fit assuming the apparent pK of internal (bound) dye to be 7.25. Inset: time course of absorption changes in neutral red during nine closely spaced (75-msec) cumulated pulses, from a starting pH of 8 (see superset diagram). The acidification per flash in this experiment is 0.05 units (Junge et al., 1978b. 1979).
(here, 75 msec) so that the pH changes accumulate. Comparison with the calibration curve yields 0.05 pH units of acidification for each flash (Junge et al., 1979).
C. Proton Pumps 1. PROTON RELEASE DURING WATER OXIDATION
Technically, the water-oxidizing enzyme system is perhaps the most interesting part of the photosynthetic apparatus. With the end of the petroleum era at hand, hydrogen gas produced at the expense of sunlight is an attractive alternative. The green plant has developed a technique for extracting electrons from water (with the possibility of producing hydrogen gas; see Hall, 1977) by using light quanta in the visible spectral region. The high efficiency
454
WOLFGANG JUNGE
of this naturally occurring process has not yet been technically simulated. To produce one molecule of oxygen from water requires the removal of four electrons and four protons from two molecules of water: 2H20(liq.)- O,(gas) + 4H+(aq.) + 4eIf this process occurred stepwise in water the abstraction of the first electron would be far more endothermic than to be driven by a quantum of red light. Apparently nature has coupled highly endothermic reactions with less endothermic ones to adapt the energetic requirements of each step to the free energy of red light. Experimentally this is apparent from the fact that oxygen is produced only after the accumulation of four oxidizing equivalents (after the input of four quanta of light). Actually, upon excitation of dark-adapted chloroplasts, a release of dioxygen occurs immediately after excitation with the third in a group of short flashes (for a review, see Joliot and Kok, 1975). This has been interpreted to indicate that the singly oxidized state of the enzyme system is the most stable state in the dark. Little is known about the chemical details of the enzyme system which stores the oxidizing equivalents. At least four manganese atoms are required for its function, which seem to change their oxidation states (Wydrzynski et al., 1976). However, it is premature to decide whether the four oxidizing equivalents are stored upon manganese proper, before reacting with two water molecules, or whether “bound” water complexed by manganese is successively oxidized and only finally released as 0, (Renger, 1977). In order to discriminate between these two alternatives, a number of different laboratories have examined the release of protons with successive flashdriven transitions between oxidation states. The results are summarized in Table 11, where the oxidation states are designated SO, S1, . . , S4, and the
.
TABLE I1 SURVEY OF STOICHIOMETRIES AND STEPS OF PROTON RELEASEBY THE WATER-SPLITTING ENZYME
Protons released per electron‘ ~
so-s1
~~
s1-s2 0 0 0 1 1
0
0.75 1 0 1 1 1 (I
0 0
SO, S1,
~~
S2-S3 0 1.25 1 1 1 1 1
~
s3-s4, 4 2 2 2 1 2 2
so
Reference Fowler and Kok (1974) Fowler (1 977 ) Saphon and Crofts (1977) Junge et al. (1977) Hope and Morland (1979) Velthuys (1980) Forster et al. (1981)
. . . , S4 designate the sequential discrete oxidation states of the water-splitting . . . , S4-S4 designate the corresponding transitions between oxidation states.
enzyme. SO-S1,
24. ELECTRIC GENERATORS IN PHOTOSYNTHESIS
455
transitions are designated SO-S1 , . . . ,S3-S4. While the disagreement among different laboratories and even within one laboratory is obvious, one major point of agreement is also clear: Except in the early experiments of Fowler and Kok (1974), at least three of the four transitions are accompanied by proton release. Superficially, this is in agreement with the suggestion of Renger (1977) that bound water is successivelyoxidized, but Sauer (1980) has pointed out that hydroxyl ions could be the specific counterion for oxidized maganese in the enzyme complex. Since OH- binding is equivalent to H+ release, the stoichiometric experiments would not then discriminate between the two alternative mechanisms. Further attempts to distinguish the two alternatives have been based on detailed analysis. Measurements (via neutral red) of successive flash-induced acidifications on a microsecond time scale have shown each proton ejection step to be multiphasic, with the initial rise having a time constant of 100psec (Junge and Auslander, 1978). The multiphasic rise has been analyzed, and components attributed to certain transitions between subsequent oxidation steps (Forster et al., 1981). It became clear that one portion of proton release precedes the turnover of a supposed precursor of water oxidation, which became apparent in electron paramagnetic resonance spectroscopy (“signal I1 vf”; Babcock et al., 1976).
-
2. PROTOLYTIC REACTIONS INVOLVING PLASTOQUINONE The known electron transfer sequence beginning in photosystem I1 is as follows: the red-absorbing chlorophyll unit designated P680 (Doring el al., 1968); pheophytin a (Klimov et al., 1980); a tightly bound plastoquinone molecule (Q), originally defined as the quencher of chlorophyll fluorescence; a second, closely associated quinone molecule designated R; (Stiehl and Witt, 1968; Bouges-Bocquet, 1973; Velthuys and Amesz, 1974); and then the “pool” plastoquinone molecules, which amount to at least six per pair of reaction centers and serve as a kinetic buffer. Upon transfer of one electron into the oxidized Q/R couple, the state Q/R- is formed rapidly and remains stable for several seconds (Pulles et al., 1976). After a second electron enters, to make Q-/R-, the couple dismutates to Q/Rz-, which is the first form that can transfer electrons onward to the quinone pool (Bouges-Bocquet, 1973; Velthuys and Amesz, 1974). From a comparison of the equilibrium constants for electron transfer within the Q/R system with the equilibrium constants for dismutation of duroquinone in aqueous ethanol solution, Diner (1977) proposed that protonation of plastoquinone occurred at the state of R2-, which would predict binding of zero protons on each odd-numbered flash and two protons on each even-numbered flash, beginning with fully oxidized Q/R. Experiments on proton uptake from the outer phase, associated with plastoquinone reduction, have shown that one proton is taken up per electron transferred
456
WOLFGANG JUNGE
(i.e., per single-turnover flash) during low-frequency, repetitive excitation (Schliephake et al., 1968; Junge and Auslander, 1974; Auslander and Junge, 1975). The predicted periodicity in proton uptake by plastoquinone under excitation of dark-adapted chloroplasts with a series of flashes has not been observed experimentally (Fowler, 1977a). The situation is very similar to that found in bacterial chromatophores (see Dutton et af., this volume), where a diffusion barrier for protons was postulated to cover the reduction sites of the special bound ubiquinones (Wraight, 1978). A demonstration of the time course for acidification of the internal aqueous phase, following a single-turnover flash, is given in Fig. 8. The rapid phase, represented by the upstroke (which is itself complex on a microsecond time scale; see Section IV,C,l) is associated with water oxidation. The slow phase can be attributed to plastoquinones, on the basis of the fact that 2,5-dibromo-3-methyl-6-isopropyl-p-benzoquinone (DBMIB), a quinone analog and antagonist, abolishes the phase (Auslander and Junge, 1975). The half-time for proton release via the plastoquinones is thus about 20 msec, whereas that for proton uptake-determined previously (Auslander and Junge, 1974)-was 60 msec. The simplest way to accommodate this discrepancy is to suppose that proton uptake occurs at an external site shielded from the aqueous bulk phase by a diffusion barrier (Auslander and Junge, 1975). The half-time for proton uptake can be reduced by mechanical disruption, by detergents, and by uncoupling agents, so that-in the extreme-half-times of 2 msec are obtained, tolerably close to the half-time of 0.6 msec observed for electron transfer (hydroquinone formation). Additional evidence for a diffu-
U
0
50 rns
0
50 rns
time FIG.8. Time course of flash-induced, pHin-indicatingabsorption changes in neutral red for discriminating protolysis associated with water oxidation and with plastoquinone oxidation. (Left) Both photosystems-and therefore both proton steps-functioning. (Right) Plastoquinone blocked by DBMIB, so that only the water oxidation step releases a proton.
457
24. ELECTRIC GENERATORS IN PHOTOSYNTHESIS
sion barrier, impeding the access of artificial electron acceptors to photosystem 11, has been presented by Renger (1979). The gross topological arrangement of proton systems, with their kinetic constants, in the thylakoid membrane is summarized in Fig. 9. [The right-hand portions (leakages and the ATP synthetase, which are not otherwise discussed here) are included for the sake of providing an overall picture. Details of the ATP synthetase are given by Graber (this volume).]
3. PROTOLYTIC REACTIONS INVOLVING THE TERMINAL ELECTRON ACCEPTOR Proton uptake at the terminal acceptor is determined entirely by the chemical properties of the acceptor. With the non-proton-binding acceptor ferricyanide, no proton uptake is observed upon reduction of photosystem I; in the case of oxygen, mediated by benzyl or methyl viologen, one proton per electron is taken up, and we can infer that reduction of NADP+ results in the uptake of 0.5 proton per electron (Junge and Auslander, 1974). Like the uptake site at plastoquinone, that at the terminal acceptor appears also to be shielded by a diffusion barrier (Auslander and Junge, 1974).
PUMPS
LEAKS
AT P SYNTHASE
ADPiP
H+
ATP
H+
SHIELD
CORE (low dielectric constant, low conductance 1
FIG.9. Kinetic parameters for electric potential generation (double arrows), proton uptake from the external phase and release into the internal phase (heavy downward arrows), proton leakage from thylakoids (heavy upward arrow), and useful proton flow through the ATP synthetase. On the outer side of the membrane, protons are taken up from buffering groups under a diffusion barrier or shield. These postulated buffering groups can be refilled from the outer aqueous phase only after a finite delay.
458
WOLFGANG JUNGE
4. OVERALL STOICHIOMETRYAND MAGNITUDE OF THE pH DIFFERENCE Under conditions in which the linear electron transport chain (but not the cyclic system) is operative, the net stoichiometry of proton pumping is two H + transported from the external aqueous phase to the internal aqueous phase for each electron transferred from water to the terminal acceptor (assuming benzyl or methyl viologen). This conclusion has also been reached by many other investigators (see, e.g., Karlish and Avron, 1971; Saphon and Crofts, 1977b; Chow and Hope, 1977; and the reviews of Trebst, 1974; or Dilley and Giacinta, 1975), using both flashing light excitation and continuous illumination. Under continuous illumination, very substantial bulk phase pH changes can result from proton-pumping activity. With a starting pH of 7-8, and ADP absent, the internal phase can become acidic by as much as 3 units (Rumberg and Siggel, 1968; Rottenberg et al., 1972) during linear electron transport, and by as much as 3.8 units during cyclic electron transport (with the artificial redox mediator pyocyanin; Pick and Avron, 1976). Some restriction of acidification may occur as the external pH is experimentally lowered, but Portis and McCarty (1973), using low pK amines, demonstrated a pH difference of 2.8 units with pH, = 6.5. Thus, the proton pumps and redox systems can still run at least down to ApH 3.7. V.
COMMENTS ON THE PATHWAY OF PROTONS TO THE ATP SYNTHETASE
Although it is not the purpose of this article to cover the mechanism of proton reaction during actual ATP synthesis (see the discussions by Graber, Kagawa, and Kozlov and Skulachev, this volume), some brief comments on the routing of protons from the redox pump to the ATP sites do seem in order. After the strenuous discussions on the nature of energy flow in photophosphorylation had faded away (for an afterglow, see the multiauthored review by Boyer et al., 1977), and the general tenets of proton coupling-as formulated by Mitchell (1961, 1966)-had been accepted, the question arose of whether the relevant protons arrive at the ATP synthetase via the internal bulk (osmotic) volume (Mitchell, 1961, 1966, 1977, 1978) or via special localized conducting subspaces either within the membrane (Williams, 1959, 1961, 1976, 1978) or at the membrane-water interface (Kell, 1979). Several observations suggest the latter alternative: (1) Ort et al. (1976), studying ATP synthesis under illumination by short light pulses (10-100 msec), found that permeating buffers had much less effect on synthesis than expected from the calculated buffer suppression of the bulk phase p H change. (For more recent aspects of these experiments, see Dilley et al., this volume.)
24. ELECTRIC GENERATORS IN PHOTOSYNTHESIS
459
(2) The existence of some kind of subspace seems to be confirmed by the time relationships of electron transport-driven proton uptake and release, as described in Section IV,C,2. (3) Especially in mitochondria, proton/ATP stoichiometries obtained from kinetic experiments are largely at variance with those obtained by static head experiments (reviewed by Kell, 1979), suggesting that the ATP synthetase does not see the average bulk proton motive force (Van Dam et al., 1978). But rejoiners to these observations have already been offered: (1) At least two other laboratories (Davenport and McCarty, 1980; Vinkler et al., 1980)have reached opposite conclusions in repeating the experiments of Ort et al. (1976). (2) The space or shield defined by the experiments of Auslander and Junge (1974) is on the outside of the thylakoid membrane, not the inside. (3) The stoichiometric problems are especially obvious in mitochondria and perhaps may not be extrapolated to chloroplasts. The issue has been further attacked experimentally by determining how fast released protons equilibrate with the internal aqueous volume of the thylakoids and how fast they disappear into the ATP synthetase, using flash spectrophotometry with neutral red. As already mentioned in Section IV,B, almost all distributed buffers-even the most hydrophilic ones-suppress flash-induced absorption changes in neutral red, so there seems little doubt that the dye estimates p H in the internal bulk phase. Furthermore, even the most rapid component (7%= 100 psec) of pH change after flash excitation of photosystem I1 is sensitive to water-soluble buffers in a manner dependent on the internal osmotic volume (Fig. 4 in Junge et al., 1978a). Therefore, exchange of protons among the water-oxidizing sites, the neutral redmeasuring space, and the osmotic space is at least as rapid as 100 psec. However, relaxation of the flash-induced pH change, during ATP synthesis, occurs with a half time of 10-12 msec (Fig. 5 in Junge et al., 1978a). Thus, at least with the relatively low-velocity ATP formation that occurs during flashing light, there is ample time for protons to equilibrate with the internal aqueous space before they enter the ATP synthetase.
VI.
SUMMARY
Figure lopresents, in an architectural form, our present view of molecular events involving proton transport in the thylakoid membrane. Photon excitation in either (or both) photosystems produces extremely rapid electron transfer from special chlorophyll a molecules on the inner side of the thylakoid membrane to acceptors on the outer side. The resultant charge shift is expressed both as a bulk phase potential difference and as an asymmetry of membrane surface potentials, the two together contributing to the electric field assayed by intrinsic membrane probes. In addition, intact chloroplasts
t IPMO-A)
Mn-COMPLEX
P680
D R
ICHL-A)
H'
I
PO
CVT-86 CVT-F FeS PC
I C h l - A1
P 700
H'
FIG. 10. Artist's view of the arrangement of electron transfer proteins and the ATP synthetase in the thylakoid membrane. The membrane is shown as a lipid bilayer with the known proteins sized according to their molecular weights. Double arrows, electrogenic electron transfer; thin arrows, nonelectrogenic electron transfer; heavy arrows, proton transfer steps. Certain groups active in electron transfer are specified: managnese in the water-oxidizing complex, chlorophyll a in both photosystems, pheophytin a in photosystem 11, plastoquinones (the dots near photosystem 11), copper in plastocyanin, the cytochrome hemes, and several iron-sulfur groups. The vertical serration locates a likely lateral separation between the serial components. (Photosystem I1 is preferentially located in the stacked inner portions of thylakoid membranes, whereas photosystem I and the ATP synthetase are preferentially in the nonstacked portions. Physiological connection between the photosystems is provided by lateral diffusion of phastocyanin and probably also plastoquinone.) Except for the shape of the ATP synthetase (CFI -CFo), the drawn shape of the proteins is arbitrary.
24. ELECTRIC GENERATORS IN PHOTOSYNTHESIS
461
carry out a third, slower electrogenic reaction (under conditions of pseudocyclic electron transport) which involves photosystem I and can be modulated by photosystem 11. Observable proton transfer seems directly linked to the redox reactions (of water, plastoquinone, and the terminal acceptor), and there is not yet convincing evidence for a special proton pump or proton-pumping protein. On the outer side of the membrane, proteins can act as shields and/or intermediate proton donors to the reduced quinones, so that relative alkalinization of an intramembranal (or surface) subcompartment is possible. On the inner side of the thylakoid membrane, protons are released rapidly by the wateroxidizing enzyme system, but comparatively slowly by plastohydroquinone. Both the electric potential difference between the bulk phases and the pH difference appear to be acollective property of a large functional unit containing at least lo5 chlorophyll molecules and corresponding broadly to the size of one thylakoid. It is well established that the electric potential difference and the pH difference are exchangeable as driving forces for photophosphorylation (see Griiber, this volume), but under steady illumination the electric component between the bulk phases appears negligible. Our current research in this field is being directed toward the molecular mechanisms of water oxidation, of the extremely rapid charge separation taking place in the photoreaction centers, and of the manner in which the ATP synthetase utilizes proton flow. ACKNOWLEDGMENTS The author has enjoyed many years of joint efforts with Professors Reich, Rumberg, and Witt at the Technische Universitlt Berlin. He wishes to thank his co-workers, Drs. Ausllnder, Emrich, Forster, Hong, McGeer, and Schmid. Thanks are due to I. Columbus and M. Offermann who prepared the graphs and to Professor Slayman for extensive editorial input. This work was financially supported by the Deutsche Forschungsgemeinschaft and the European Commission. REFERENCES Akerlund, H . E., Anderson, B. L., Person, B., and Albertson, P. A. (1979). Biochim. Biophys. Acta 552,238-246. Amesz, J., and DeGrooth, B. G. (1975). Biochim. Biophys. Actu 376,298-307. Ausllnder, W., and Junge, W. (1974). Biochim. Biophys. Actu 357,285-298. AuslSinder, W., and Junge, W. (1975). FEBSLett. 59,310-315. Babcock, G. T., Blankenship, R. E., and Sauer, K. (1976). FEBS Lett. 61,286-289. Barber, J. (1972). FEBSLett. 20,251-254. Barber, J. (1977). Proc. CON.Int. Potash Inst., 13th pp. 83-93. Barber, J. (1981). Biochim. Biophys. Actu, in press. Barber, J., and Kraan, G. P. B. (1970). Biochim. Biophys. Actu. 197,49-59. Barber, J., Telfer, A., and Nicolson, J. (1974). Biochim. Biophys. Actu 357, 161-165.
462
WOLFGANG JUNGE
Bengis, C., and Nelson, N. (1975). J. Biol. Chem. 250,2783-2788. Bouges-Bocquet, B. (1973). Biochim. Biophys. Acta 314,250-256. Bouges-Bocquet, B. (1980). FEBS Lett. 117, 54-58. Boyer, P . D., Chance, B., Ernster, L., Mitchell, P., Racker, E., andslater, E. C. (1977). Annu. Rev. Biochem. 46,955-1026. Breton, J . (1976). Biochim. Biophys. Acta 459, 66-75. Bulychev, A. A., and Vredenberg, W. J. (1976). Biochim. Biophys. Acta 423, 548-556. Bulychev, A. A., Andrianov, V. K., Kurella, G. A., and Litvin, F. F. (1972). Nature (London) 236, 175-176. Chance, B., and Mela, L. (1966). J. Biol. Chem. 241,4588-4596. Chow, W. S., and Hope, A. B. (1977). Aust. J. Plant Physiol. 4,647-665. Conjeaud, H., Michel-Villaz, G. M., Vermeglio, M. A., and Mathis, P. (1976). FEBSLett. 71, 138-144. Crowther, D., Mills, J. D., and Hind. G. (1979). FEBS Left. 98, 386-390. Davenport, J. W., and McCarty, R. E. (1980). Biochim. Biophys. Acta 589,353-357. Deamer, D. W., and Packer, L. (1969). Biochim. Biophys. Acta 172,539-545. DeGrooth, B. G., van Gorkom, H. J., and Meiburg, R. F. (1980). Biochim. Biophys. Acfa589, 299-314. Dilley, R. A., and Giacinta, R. T. (1975). In “Current Topics in Membranes and Transport” (F. Bronner and A. Kleinzeller, eds.), Vol. 2, pp. 49-107. Academic Press, New York. Dilley, R. A., and Vernon, L. P. (1965). Arch. Biochem. Biophys. 111,365-375. Dilley, R. A., Prochaska, L. J., Baker, G. M., Tandy, N. E., andMillner, P. A. (1981). In “Electrogenic Ion Pumps” (C. L. Slayman, ed.). Academic Press, New York. Diner, B. A. (1977). Biochim. Biophys. Acta 460,247-258. Doring, G., Bailey, J. L., Kreutz, W., Weikard, J., and Witt, H. T. (1968). Nafurwisseuschaften 55,219-220. Duniec, J. T., and Thorne, S. W. (1979). FEBS Lett. 105, 1-4. Dutton, P. L., Meuller, P., O’Keefe, D. P., Packham, N. K., Prince, R. C., and Tiede, D. M. (1981). In “Electrogenic Ion Pumps” (C. L. Slayman, ed.). Academic Press, New York. Ellenson, J . L., and Sauer, K. (1976). Photochem. Photobiol. 23, 113-123. Emrich, H . M., Junge, W., and Witt, H. T. (1969). Z . Naturforsch. 24B, 1144-1146. Fernandez, M. S., and Fromherz, P. (1977). J . Phys. Chem. 81, 1755-1761. Forster, V., Hong, Y. Q., and Junge, W. (1981). Biochim. Biophys. Acfa 863, in press. Fowler, C. F. (1977a). Biochim. Biophys. Acta 459,351-363. Fowler, C. F. (197713). Biochim. Biophys. Acta 462,414-421. Fowler, C. F., and Kok, B. (1972). Int. Congr. Photobiol., 6th, Bochum No. 417 (Abstr.). Fowler, C. F., and Kok, B. (1974). Biochim. Biophys. Acta 357,299-307. Fowler, C. F., and Kok, B. (1976). Biochim. Biophys. Acta 423,510-523. Gaensslen, R., and McCarty, R. (1971). Arch. Biochem. Biophys. 147,55-65. Grlber, P. (1981). In “Electrogenic Ion Pumps (C. L. Slayman, ed.). Academic Press, New York. GrBber, P., and Witt, H. T. (1974). Biochim. Biophys. Acta 333,389-392. Graber, P., and Witt, H. T. (1976). Biochim. Biophys. Acta 423, 141-163. Hall, D. D. (1977). In “Solar Powers and Fuels” (J. R. Bolton, ed.), pp. 27-52. Academic Press, New York. Hauska, G., and Trebst, A. (1977). In “Current Topics in Bioenergetics” (R. 0. Sanadi, ed.), Vol. VI, pp. 151-220. Academic Press, New York. Heldt, H. W., Werdan, K., Milovancev, M., and Geller, G. (1973). Biochim. Biophys. Acta314, 224-241. Hind, G., Nakatani, H. Y., and Izawa, D. (1974). Proc. Natl. Acad. Sci. U.S.A.71,1484- 1488. Hope, A. B., and Morland, A. (1979). Aust. J. Plant Physiol. 6, 1-16.
24. ELECTRIC GENERATORS IN PHOTOSYNTHESIS
463
Horvath, G., Niemi, H. A., Droppa, M., and Fahludi-Daniel, A. (1979). Plant Physiol. 63, 778-782. Huber, H. L., and Rumberg, B. (1979). Annu. Meet. Dsch. Ges. Biophy. p. 74 (Abstr.). Itoh, S. (1978). Biochim. Biophys. Acta 504,324-340. Itoh, S. (1979a). Biochim. Biophys. Acta 548, 579-595. Itoh, S. (1979b). Biochim. Biophys. Acta 548, 596-607. Jackson, J. B., and Crofts, A. R. (1969). FEBSLert. 4, 185-188. Joliot, P., and Kole, B. (1975). In “Bioenergetics of Photosynthesis” (Govindjee, ed.), pp. 388-412. Academic Press, New York. Junge, W. (1972). Proc. Int. Congr. Photosynthesis, 2nd. pp. 1065-1074. Junge, W. (1977a). Annu. Rev. Plant Physiol. 28, 503-536. Junge, W. (1977b). In “Photosynthesis I” (A. Trebst and M. Avron, eds.), pp. 59-93. SpringerVerlag, Berlin and New York. Junge, W., and Auslander, W. (1974). Biochim. Biophys. Acta 333,59-70. Junge, W., and Ausllnder, W. (1978). In “Photosynthetic Water Oxidation” (H. Metzner, ed.), pp. 213-228. Academic Press, New York. Junge, W., and Eckhof, A. (1974). Biochim. Biophys. Acta 357, 103-1 17. Junge, W., and Schaffernicht, H. (1979). In “Chlorophyll Organization and Energy Transfer in Photosynthesis/Ciba Symp. 61” (G. Wolstenholme and D. W. Fitzsimmons, eds.), pp. 127-146. Elsevier, Amsterdam. Junge, W., and Schmid, R. (1971). J. Membr. Biol. 4, 179-192. Junge, W., and Witt, H. T. (1968). Z . Naturforsch, Ser. B 23,244-254. Junge, W., Schaffernicht, H., and Nelson, N. (1977). Biochim. Biophys. Acta 462,73-85. Junge, W., McGeer, A., and AuslSinder, W. (1978a). In “Frontiers of BiologicalEnergetics” (P. L. Dutton, J . S. Leigh, and A. Scarpa, eds.), pp. 275-283. Academic Press, New York. Junge, W., McGeer, A., AuslSinder, W., and Kolia, J. (1978b). In “Energy Conservation in Biological Membranes” (G. Schafer and M. Klingenberg, eds.), pp. 113-127. SpringerVerlag, Berlin and New York. Junge, W., Ausllnder, W., McGeer, A. J., and Runge, T. (1979). Biochim. Biophys. Acta 546, 121-141. Kagawa, Y. (1981). In “Electrogenic Ion Pumps” (C. L. Slayman, ed.). Academic Press, New York. Karlish, S. J. D., and Avron, M. (1971). Eur. J . Biochem. 20,51-57. Kell, D. B. (1979). Biochim. Biophys. Acta 549, 55-99. Klimov, V. V., Dolan, E., and Ke, B. (1980). FEES Lett. 112,97-101. Knaff, D. B. (1977). Arch. Biochem. Biophys. 182, 540-545. Knox, R. S. (1979). In “Light Induced Charge Separation in Biology and Chemistry” (H. Gerischer and J . J. Katz, eds.), pp. 45-60. Verlag Chemie, Weinheim, Germany. Kozlov, I. A,, and Skulachev, V. P. (1981). In “Electrogenic Ion Pumps” (C. L. Slayman, ed.). Academic Press, New York. Krause, G. H. (1973). Biochim. Biophys. Acta 292,715-728. Kuhn, H. (1979). In “Light Induced Charge Separation in Biology and Chemistry” (J. J. Katz and H. Gerischer, eds.), pp. 151-170. Verlag Chemie, Weinheim, Germany. Liptay, W. (1969). Angew. Chem. Int. Ed. Engl. 8, 177-188. Lynn, W. S. (1968). J . Biol. Chem. 243, 1060-1064. Malkin, R. (1978). FEES Lett. 87,329-333. Mangel, M., Berns, D. S., and Ilani, A. (1975). J. Membr. Biol. 20, 171-180. Mathis, P. (1970). Acad. Sci. 271, 1094-1096. Mathis, P., and Vermeglio, A. (1975). Biochim. Biophys. Acta 369, 371-381. Mercer, F. V., Hedge, A. J., Hope, A. B., and McLean, J. D. (1955). Aust. J . Biol. Sci. 8,l-18. Mitchell, P . (1961). Nafure (London) 191, 144-148.
464
WOLFGANG JUNGE
Mitchell, P. (1966). Biol. Rev. 41, 445-502. Mitchell, P. (1977). FEBS Lett. 78, 1-20. Mitchell, P. (1978). TIBS 3,N 58-61. Moebius, D. (1979). I n “Light-Induced Charge Separation in Biology and Chemistry” (J. J . Katz and H. Gerischer, eds.), pp. 171-186. Verlag Chemie, Weinheim, Germany. Muehlethaler, F. (1977). I n “Photosynthesis I ”(A. Trebst and M.Avron, eds.),pp. 503-521. Springer-Verlag, Berlin and New York. Nakatani, H. Y., Barber, J., and Forrester, J. A. (1978). Biochim. Biophys. Actu504,215-225. Ort, D. R., Dilley, D. A., and Good, N . E. (1976). Biochim. Biophys. Actu 449,108-126. Ortoidze, T. V., Borisevitch, G. P., Venediktov, P. S., Kononenko, A. A., Matorin, D. N., and Ruhin, A. B. (1979). Biochem. Physiol. Pflunzen 174,85-91. Pick, U., and Avron, M. (1976). FEBSLett. 65,48-53. Platt, J. R. (1961). J. Chem. Phys. 346,862-863. Portis, A. R., and Heldt, H. W. (1976). Biochim. Biophys Actu 449,434-446. Portis, A. R., and McCarty, R. E. (1973). Arch. Biochim. Biophys. 156,621-625. Pulles, M. P. J., van Gorkom, H. J., and Willemsen, J. G. (1976).Biochim. Biophys. Actu 449, 536-540. Quintanilha, A. T., and Packer, L. (1978). Arch. Biochem. Biophys. 190,206-209. Radunz, A, (1979). Z . Nuturforsch. 34 C , 1199-1204. Reich, R., and Sewe, K. U. (1977). Photochem. Photobiol. 26, 11-16. Reinwald, E. (1970). Thesis, Technische Universitat, Berlin. Reinwald, E., Stiehl, H. H., and Rumber, B. (1968). Z . Nuturforsch. Ser. B 23, 1616-1617. Renger, G. (1977). FEBS Lett. 81,223-228. Renger, G. (1979). Z . Nuturforsch. 34 C , 1010-1014. Rottenberg, H., Grunwald, T., and Avron, M. (1972). Eur. J. Biochem. 25,54-63. Rumberg, B. (1977). In “Photosynthesis I” (A. Trebst and M. Avron, eds.), pp. 405-415. Springer-Verlag, Berlin and New York. Rumberg, B., and Muhle, H. (1976). Bioelectrochem. Bioenerg. 3, 393-403. Rumberg, B., and Siggel, U. (1968). Z . Nuturforsch. Ser. B 23,239-244. Saphon, S., and Crofts, A. R. (1977a). Z . Nuturforsch. 32 C , 617-626. Saphon, S., and Crofts, A. R. (1977b). Z . Nuturforsch. 32 C , 810-816. Sauer, K. (1980). Acc. Chem. Res., in press. Sauer, K., Mathis, P., Acker, S., and van Best, J. A. (1978). Biochim. Biophys. Actu 503, 120-134. Schapendonk, A. H. C. M., and Vredenberg, W. J. (1977). Biochim. Biophys. Actu 462, 613-621. Schliephake, W., Junge, W., and Witt, H. T. (1968). Z . Nuturforsch. 23 B , 1571-1578. Schlodder, E., and Witt, H. T. (1980). FEBS. Lett. 112, 105-113. Schmid, R., and Junge, W. (1975). Biochim. Biophys. Actu 394,76-92. Schmidt, S., Reich, R., and Witt, H. T. (1970). Nuturwissenschuften 58, 414-415. Schnetkamp, P. P . M., Kaupp, U. S., and Junge, W. (1980). In preparation. Schrdder, H., Muhle, M., and Rumberg, B. (1972). Proc. I n t . Congr. Photosynthesis, 2nd pp. 919-930. Sewe, K. U., and Reich, R. (1977a). Z . Nuturforsch. 32 C , 161-171. Shuvalov, V. A., Klevanik, A. V., Sharkov, A. V., Kryukov, P. G., and Ke, B. (1979). FEBS Lett. 107, 313-316. Siefermann-Harms, D. (1978). Biochim. Biophys. Actu 504,265-277. Slovacek, R. E., Crowther, D., and Hind, G. (1979). Biochim. Biophys. Actu 547, 138-148. Stiehl, H. H., and Witt, H. T. (1968). Z . Nuturforsch. 23 B, 220-224. Strichartz, G. R., and Chance, B. (1972). Biochim. Biophys. Actu 256,71-84. Thomas, J. B., Minnaert, K., and Elber, P. D. (1956). Actu Bot. Neerl. 5, 314-321.
24. ELECTRIC GENERATORS IN PHOTOSYNTHESIS
465
Trebst, A. (1974). Annu. Rev. Plant Physiol. 25,423-458. Tributsch, H. (1972). Photochem. Photobiol. 16, 261-269. Trissl, H. W. (1980). Biochim. Biophys. Acta 595, 82-95. Trissl, H. W., and Grilber, P. (1980a). Biochim. Biophys. Acta 595,96-108. Trissl, H. W., and Grilber, P. (1980b). Bioelectrochem. Bioenerget. 7, 167-186. Van Dam, K . , Wiechmann, A. H. C. A., Hellingwerf, K. J., Arents, J. C., and Westerhoff, K. (1978).In “MembraneProteins” (P. Nichollsetal., eds.), pp. 121-132. Pergamon, Oxford. Velthuys, B. R. (1978). Proc. Natl. Acad. Sci. U.S.A. 75,6031-6034. Velthuys, B. R. (1979). Proc. Natl. Acad. Sci. U.S.A. 76,2765-2769. Velthuys, B. R. (1980). FEBSLett. 115, 167-170. Velthuys, B. R., and Amesz, J. (1974). Biochim. Biophys. Acta 333, 85-94. Vermeglio, A., and Mathis, P. (1974). Biochim Biophys. Acta 368, 9-17. Vinkler, C., Avron, M., and Boyer, P. D. (1980). J. Biol. Chem. 255,2263-3366. Vredenberg, W. J., and Tonk, W. J. M. (1975). Biochim. Biophys. Acta 387,580-587. Waggoner, A. (1976). J. Membr. Biol. 27,317-337. Wagner, R., and Junge, W. (1980). FEBS Lett. 114,327-333. Walz, D., Goldstein, L., and Avron, M. (1974). Eur. J. Biochem. 47,403-407. Williams, R. J. P. (1959). In “Enzymes” (P. D. Boyer, H. Lardy, and K. Myrbaeck, eds.), Vol. 1 , p. 391. New York, Academic Press, New York. Williams, R. J. P. (1961). J. Theor. Biol. 1, 1-13. Williams, R. J. P . (1976). TIBS 1,222-224. Williams, R . J. P. (1978). FEBSLett. 85,9-19. Wintermans, J. (1967). In “Le Chloroplaste” (C. Sironval, ed.), pp. 86-90. Masson, Paris. Witt, H. T. (1975). In “Bioenergetics of Photosynthesis” (Govindjee, ed.), pp. 493-554. Academic Press, New York. Witt, H. T. (1979). Biochim. Biophys. Acta 505,355-327. Witt, H. T., and Wolff, C. (1970). Z. Naturforsch. Ser. B 25,387-388. Witt, H. T., and Zickler, A. (1973). FEBS Lett. 37,307-310. Witt, H. T., and Zickler, A. (1974). FEBSLett. 39, 205-208. Wolff, C., and Witt, H. T. (1969). Z. Naturforsch. Ser. B 24, 1031-1037. Wolken, J. J., and Schwertz, F. A. (1953). J. Gen. Physiol. 37, 111-120. Wraight, C. A. (1978).In “Frontiers of Biological Energetics” (P. L. Dutton, J. S. Leigh, and A. Scarpa, eds.), pp. 218-226. Academic Press, New York. Wraight, C. A., and Crofts, A. R. (1971). Eur. J. Biochem. 19, 386-397. Wydrzynski, T., Zumbulyadis, N., Schmidt, P. G., Gutowsky, H. S., and Godvindjee (1976). Proc. Natl. Acad. Sci. U.S.A. 73, 1 196-1 198. Yerkes, Chr. T., and Babcock, G. T. (1981). Biochim. Biophys. Acta 634, 19-29. Zickler, A., Witt, H. T., Boheim, G. (1976). FEBSLett. 66, 142-148.
This Page Intentionally Left Blank
CURRENT TOPICS IN MEMBRANES A N D TRANSPORT, VOLUME 16
Chapter 25 The Role of the Electrogenic Sodium Pump in Controlling Excitability in Nerve and Cardiac Fibers MARIO VASSALLE Department of Physiology State University of New York Downstate Medical Center Brooklyn, New York
................. ......................................
Introduction ............. The Excitation Process 111. The Sodium Pump and I.
11.
VI.
Concluding
1.
467 468 469 47 1 474 474 475 480 48 1 482
INTRODUCTION
In many tissues, excitation consists of a relatively brief depolarization (the action potential) caused by movement of ions down their electrochemical gradient. In order for the ionic homeostasis of the cells to be maintained, ions flowing during excitation must be returned to their original environment. This is accomplished by a metabolism-driven pump. The purpose of this presentation is to review briefly the events underlying excitability in certain tissues and the modifications of the excitability process by active ion 467
Copyright i~ 1982 by Academic Press, Inc All rights of reproduction in any form reserved ISBN 0-12-1533 16-6
468
MARIO VASSALLE
transport. The active ion transport to be discussed will be limited to the sodium-potassium pump, and particular emphasis will be laid on pump activity in cardiac tissues, since much information in this area is recent and not readily available in review form. In particular, evidence will be presented which shows that the sodium pump controls excitability, in many instances because it is electrogenic. Several reviews are available which cover various aspects of the activity of the sodium pump (Kernan, 1970; Kerkut and York, 1971; Thomas, 1972, this volume; Haas, 1972; Glitsch, 1979; Schwartz et al., 1975; Wallick et al., 1979). II. THE EXCITATION PROCESS The phenomena underlying excitability are complex and are not identical for different structures. However, certain common features are generally shared. One of these characteristics is that all excitable cells have a negative resting potential, an unequal distribution of ions across the cell membrane and the ability to undergo time- and voltage-dependent changes in ionic conductances. In many tissues, the resting potential is largely a potassium diffusion potential but is usually less negative than expected from the potassium concentration gradient. The resting potential is important for excitability in two respects. One is that it provides an inward driving force for sodium ions, and the other is that the increase in sodium permeability during excitation is conditioned by the value of the resting potential (availability of the sodium carriers) prior to excitation (Fig. 1; Weidmann, 1955a). Excitation is brought about when the membrane is depolarized to a critical value (the threshold), at which potential the inward sodium current begins to overpower the outward potassium current. The threshold is attained when the membrane potential is decreased either by a pacemaker potential, by local circuit current, or by external stimuli. Since rapid depolarization increases sodium conductance and sodium is more concentrated outside, this ion enters the fiber and brings about the upstroke of the action potential. The sodium conductance is then quickly inactivated, and the cell is now refractory to excitatory stimuli. An increase in potassium conductance as a consequence of depolarization brings about repolarization. These events occur in nerve and also in many cardiac cells, although in different tissues of the heart the events underlying excitation are far more complicated (see McAllister et al., 1975) than in nerve. While excitability of different tissues is the precondition for their function, clearly there must also be mechanisms by which excitation is prevented. The most obvious form of control is that resulting from the release of
25.
Na+ PUMP
469
CONTROL OF NERVE AND CARDIAC FIBERS
600-
” x
1
. > v
$
400 I
L
u
CL
200
+++-++-it--+ -
y+
-
L
d?/
-
d/’
0-
+ I
40
60
258 Na
80
I
I
100
I
,
,
120
Clamp potential (mV)
FIG.1. Maximal rate of rise of the action potential (upstroke) plotted as a function of membrane potential prior to excitation. The membrane potential was “clamped” to the values indicated on the abscissa for about SOmsec, and then the fibers were stimulated as shown in the inset. The maximal rate of rise elicited from each clamped potential is indicated on the ordinate. The open circles designate values recorded in Tyrode’s solution; crosses, those in a solution containing 25% of the normal sodium; and solid circles, the value on returning to Tyrode’s solution. (Reproduced with permission from Weidmann, 1955a.)
neuromediators. Another form of control of excitation results from the activation of an electrogenic sodium pump. Ill.
THE SODIUM PUMP AND CONTROL OF EXCITABILITY
The relationship between the sodium pump and excitability has several aspects. Electroneutral pumps maintain ionic homeostasis and therefore maintain the resting potential. In addition, the pump keeps the intracellular sodium concentration small and the inward chemical gradient for sodium large. Blockage of an electroneutral sodium pump causes a progressive, slow fall in the resting potential and a decline in excitability. With pumps blocked, the rate of decline of the resting potential should be related to the magnitude of the sodium leak at rest, the number of action potentials per unit of time, and the surface-to-volume ratio of the cells. Furthermore, the consequences of blocking the sodium pump also involve changes in the distribution of ions other than sodium and potassium. The accumulation of intracellular sodium leads to the accumulation of intracellular calcium (see Blaustein, 1974; Langer, 1977), which in turn increases potassium conductance (Meech, 1974; Isenberg, 1977a,b). One consequence of this in the heart is shortening of the action potential plateau and shortening also of the total refractory period. In fact, when calcium accumulation results from poisoning of the sodium pump
470
MARIO VASSALLE
by cardiac glycosides, a transient inward current develops at the end of the action potential, as seen particularly in Purkinje fibers (Fig. 2; Lederer and Tsien, 1976). This induces a transient depolarization and sometimes repetitive excitations (see Ferrier, 1977; and below). In addition to these effects, an electrogenic sodium pump can also modify excitability by creating an outward current. As a result, the resting potential becomes more negative, the depolarization needed to attain threshold increases, and the refractory period shortens due to shortening of the action potential. This is particularly important in cardiac tissues, since the membrane resistance is large during the plateau (and therefore the effect of an outward pump current on action potential duration is relatively large). Such a shortening allows an unimpaired excitation at a high rate of discharge. Also, even a small negative shift of the pacemaker potential may readily result in quiescence, since the threshold potential is no longer reached. Furthermore, the pump (either electroneutral or electrogenic) can modify excitability by decreasing potassium in confined spaces immediately outside the cell membrane (Cohen et af., 1976). Such a decrease in K may lead to an increased resting potential (and therefore conduction velocity). The decrease in potassium conductance resulting from a lower [K+], may allow an elec-
P
O[ -10
u 5 sec
FIG.2. Oscillatory potentials and transient inward current recorded from the same preparation. (Top) Two superimposed records of membrane potential. In both tests the stimulation was interrupted after the fourth action potential shown. In one test, oscillatory potentials followed the last action potential; in the other, the membrane potential was clamped at the maximum diastolic potential. (Bottom) The membrane current in the absence and in the presence of voltage clamp control. When the potential was clamped at the maximal diastolic potential, a transient inward current was recorded superimposed on the pacemaker current. The vertical line shows that the peak of the transient inward current occurred just before the first peak of the oscillatory potentials. (Reproduced with permission from Lederer and Tsien, 1976.)
25. Na+ PUMP CONTROL OF NERVE AND CARDIAC FIBERS
471
trogenic pump to hyperpolarize the membrane more by reducing the shortcircuiting effect of passive K+ movements (as long as pump activity itself is not reduced by the lowered extracellular potassium concentration).
IV.
EXCITABILITY IN NERVE AND THE ELECTROGENIC SODIUM PUMP
The existence of an electrogenic sodium pump in neural tissues has been demonstrated by loading the tissue with sodium, by direct intracellular injection, by exposure of the tissue to low temperatures and to low K+ , or by repetitive activation. Much of the evidence obtained by these methods is presented in the reviews by Thomas (1972; see also this volume), Kerkut and York (1971), and Kernan (1970). The evidence essentially consists of the demonstration that as a consequence of the procedures just mentioned there is a prolonged hyperpolarization, often to values negative to EK. Reduction or abolition of this hyperpolarization by cardiac glycosides, lithium, and metabolic inhibitors demonstrates that the hyperpolarization is mediated by an electrogenic sodium pump. The results, however, are often complicated by several other factors. For one, E, may vary as the result of a depletion of [K+1, by an electroneutral pump. Then, hyperpolarization can be brought about by several other mechanisms, such as an increased potassium conductance in the wake of an action potential. Such an increaseing, is relatively short, but it can summate (within limits) during successive action potentials. Another hyperpolarizing mechanism is an increase in potassium conductance induced either by neuromediator release or by intracellular calcium accumulation with activity. Furthermore, neuromediators can modify the activity of the sodium pump, and the hyperpolarization can result both from an increase in potassium conductance and from a pump current. In fact, in several tissues the inhibition of activity is brought about by a combination of different effects. One example is offered by the results of Akasu et al. (1978a) who found that the slow inhibitory postsynaptic potential (IPSP) in curarized and nicotinized sympathetic ganglia of the bullfrog consisted of two components. One component is due to an increased potassium conductance and is depressed or reversed by hyperpolarizing pulses in the presence of ouabain. The other component is due to a pump current: It is unaffected by calcium deficiency, is eliminated by ouabain or low temperature, and shows no reversal. One puzzling feature seen before is that the pump potential persists in a potassium-free solution: the authors suggest, however, that the potassium necessary to stimulate the pump would originate from the activated presynaptic terminals. Similar conclusions were reached by Holloway and Poppele (1978). They showed that the spontaneous fre-
472
MARIO VASSALLE
quency of cat stretch receptor (spindle) decreases after a train of antidromic impulses and then recovers gradually to the original frequency. Two components could be distinguished, with time constants of 250 msec and -2.6 seconds (28°C); the slower component was abolished by ouabain and diminished by lithium. The inhibition may represent the basis for the adaptation of the receptor to step inputs. In sympathetic preganglionic neurons of the cat, the irregular background spike activity is temporarily suppressed by a burst of action potentials (Mannard et al., 1977). The mechanism of this silent period could not be explained by the summation of the postspike increase in g,, since the inhibition was too long and showed no saturation. Although an increase in g, induced by an increase in calcium influx could not be ruled out, the findings could be accounted for by an electrogenic sodium extrusion. Whatever the mechanism, this finding shows that reduced excitability, following a burst of activity, is important in stabilizing the sympathetic tone. Of much interest in this regard are the results of Nicholls and collaborators. Baylor and Nicholls (1969a) found that the sensory neurons of the leech undergo a marked and prolonged hyperpolarization when the receptor fields are stimulated. This hyperpolarization could exceed the equilibrium potential for the inhibitory transmitter, thus actually reversing the inhibitory postsynaptic potentials. It was associated with increased input resistance, was not blocked by 20 mM magnesium (which blocks chemical synapses), and occurred in C1-free Ringer’s solution (in which solution the IPSPs are inverted). High K+ depolarized more effectively during the hyperpolarization, possibly through a decreased sodium-potassium pumping ratio. The hyperpolarization was eliminated by strophanthidin, was markedly reduced by low temperature, and was concluded to be due to an electrogenic extrusion of sodium. The hyperpolarization altered cell excitability in several ways: It changed the shape of the action potential, increased the threshold potential, sometimes caused conduction block, changed the polarity of synaptic potentials, and could influence the amount of neuromediators released by the terminals. Each of these changes was recognized to be of importance in changing the balance between activation and inhibition. A cell that had hyperpolarized following a burst of activity could be influenced by aneighboring cell through local accumulation of potassium or through synaptic release-since reversed IPSPs would actually decrease the membrane potential toward its resting value, thus facilitating excitation (Baylor and Nicholls, 1969a,b). Successive work by Jansen and Nicholls (1973) confirmed a role for the electrogenic pump in the hyperpolarization following fast activity. Injection of sodium caused hyperpolarization which was blocked by strophanthidin, and this identified the ion involved in the hyperpolarization. However, evidence for a second factor in the hyperpolarization-an increased g,-was
25.
Na+
PUMP CONTROL OF NERVE AND CARDIAC FIBERS
473
found, since high calcium augmented and prolonged the hyperpolarization and simultaneously decreased input resistance. In low “a+ I,, there still was hyperpolarization, and low [K+], increased it by shifting E K .The increased potassium conductance appeared to be more important at the beginning of the hyperpolarization, whenit could shunt the pumpcurrent. Also, theroleof the pump in the hyperpolarization seemed more significant for touch cells than for other sensory neurons. While the physiological function of hyperpolarization is clear from their work, Jansen and Nicholls note that the relative roles of the two components are difficult to separate. One recent report is of interest from the point of view of control of the pump. It involves the study of caffeine effects on electrical events at the cell membrane and on synaptic transmission in the superior cervical ganglion of the rabbit (Skok et al., 1978). Electrical recording revealed that caffeine could induce either rhythmic or sustained hyperpolarizations but, because caffeine has complex actions (e.g., releasing intracellular bound calcium and inhibiting phosphodiesterase), great care must be taken in order to prove any effect on the electrogenic Na+ pump per se. The rhythmic hyperpolarizations were shown, on several grounds, to be due to an increase in potassium conductance. The sustained hyperpolarizations (of about 5 mV), on the other hand, were associated with an increase in membrane resistance; were not affected by hexamethonium, atropine, or high [K+I,; and were blocked both by strophanthidin and by replacement of sodium with lithium. Also, caffeine stimulated the efflux of sodium, and this was blocked by strophanthidin. These findings show convincingly that the sustained hyperpolarizations were generated by an electrogenic pump. This result may be important for normal physiological function, because it suggests that norephinephrine may alter excitability by stimulating an electrogenic pump, as well as by its effect on the conductance of different membranes. Another field in which electrogenic sodium pumps may prove of interest is epilepsy. Tower (1969) includes an impairment of the sodium pump among the possible neurochemical mechanisms in epilepsy. Evidence is available that injection of ouabain into the central nervous system results in severe convulsions (Bignami and Palladini, 1966). While it is proposed that blockage of the pump acts through ionic derangements, the possibility exists that inhibition of an electrogenic pump, with the resulting depolarization and enhancement of excitability, may contribute to epileptic seizures (see Spencer and Kandel, 1 969). The few studies quoted above on the role of the electrogenic pump in nerve excitability confirm the concept that stimulation of the pump leads to an electrogenic ejection of sodium, although the hyperpolarization that follows a burst of activity is likely to include other factors as well. The resultant suppression of spontaneous and evoked activity now appears quite important as a means of controlling neural activity.
474
MARIO VASSALLE
V.
EXCITABILITY IN CARDIAC CELLS AND THE ELECTROGENIC PUMP
A. Evidence for an Electrogenic Sodium Pump in Cardiac Tissues While there are several demonstrations of an electrogenic sodium pump in other tissues, the evidence for cardiac tissue has not been very extensive, although recently several papers have appeared on this topic (e.g., Gadsby, this volume). Dtltze (1960) found that cooling ventricular and Purkinje fibers below about 20°C decreased the resting potential much more than expected for a diffusion potential. This fall was not due to an increased sodium influx, since it occurred in sodium-free solutions, and it was not due to potassium accumulation outside the cell membrane, since on rapid heating the resting potential increased in less than 1 second. In the presence of monoiodoacetate and anoxia, the decrease in resting potential was less at lower temperature. The conclusion of DCltze was that the fall in resting potential with low temperature was due to the slowing of an electrogenic sodium pump. Similar conclusions were reached by Page and Storm (1969, Glitsch (1969), Hiraoka and Hecht (1973), and Tamai and Kagiyama (1968). On returning to normal conditions, the membrane potential increased to above the calculated potassium equilibrium potential (Page and Storm, 1965; Glitsch, 1969; Tamai and Kagiyama, 1968) or above the original resting potential (Hiraoka and Hecht, 1973). This hyperpolarization was abolished by cardiac glycosides (Hiraoka and Hecht, 1973). In another approach, cardiac tissues have been loaded with sodium by exposure to potassium-free solutions (Noma and Irisawa, 1974, 1975; Lieberman et al., 1977), and hyperpolarization has been observed to occur immediately on reexposure t o normal solutions. Such hyperpolarization does not occur if the fibers are loaded with lithium in the absence of sodium (Hiraoka and Hecht, 1973; Noma and Irisawa, 1975). In order to distinguish hyperpolarization brought about by an electrogenic pump from that brought about by an electroneutral pump depleting extracellular potassium, Glitsch et al. (1978) decreased K+ to various levels, before and after hypothermia. The membrane potential increased whenever [K' 1, was lowered, but after hypothermia the effect showed a distinct maximum at moderate K+ concentrations, about 2.7 mM. Furthermore, the hyperpolarization after hypothermia in 2.7 m M K + was significantly larger than that produced by a potassium-free solution at normothermia. These findings support the concept that hyperpolarization is due to an electrogenic pump rather than to potassium depletion. Wiggins and Cranefield (1974) found that in fibers exposed to a sodium-
25.
Na+
PUMP CONTROL OF NERVE AND CARDIAC FIBERS
475
free solution, adding a small amount of sodium resulted in an increased resting potential. Similar findings were reported for guinea pig atria (Glitsch and Hare, 1977): Fibers perfused in a sodium-free solution hyperpolarized when exposed to Na+ ,and this effect was diminished by low temperature and digitoxigenin. The removal of K+ resulted in immediate depolarization.
B. The Effects of the Pump Current on Cardiac Excitability and Function The experiments reported above demonstrate the existence of an electrogenic sodium pump in different cardiac tissues, at least under non-steady state conditions. This elicits questions as to whether the pump contributes directly to the membrane potential under normal conditions and whether there are situations in which such a contribution varies. This, in turn, leads to consideration of how the excitability and function of cardiac cells may be affected by pump current. 1. EFFECTS ON THE RESTING POTENTIAL
The experiments of DCl6ze (1960) have shown that the resting potential falls according to a passive distribution of potassium for a decrease of 10°C from the normal (37°C) temperature. This implies that thecontribution of the pump to the resting potential during normothermia is negligible. However, cardiac cells are never at rest for any length of time, and even with a modest increase in discharge rates the maximal diastolic potential increases (see below). Thus, it appears that in heart tissues the contribution of the electrogenic pump to the resting potential is small or absent when a tissue is kept quiescent, but it is present when the cell discharges at physiological rates. And its contribution increases under non-steady state conditions, such as recovery from hypothermia or from repetitive activity. 2. EFFECTS ON THE ACTIONPOTENTIAL.
In the presence of metabolic inhibitors, cardiac tissues show a decrease in resting potential, a less negative threshold potential, a small upstroke, and a shortening of the action potential (see Haas, 1972). These changes are compatible with a blockage of the sodium-potassium pump and consequent changes in ionic distribution (including that of calcium). The contribution of blockage of the pump current, as such, to these changes is difficult to assess but has been demonstrated under certain conditions. Isenberg and Trautwein (1 974) applied long depolarizing clamps to Purkinje fibers and found that the late outward current was reduced by dihydroouabain. The voltage-current
476
MARIO VASSALLE
relation was shifted in a negative direction (more inward), while the shape of the curve was not affected (Fig. 3), at least for short exposures to the glycoside. The other currents appeared not to be affected. In the absence of voltage clamping, dihydroouabain prolonged the action potential until repolarization failed. Similar results were obtained with 4 &lithium, which is supposed to block the pump. Isenberg and Trautwein conclude that the electrogenic pump contributes to the membrane current over a wide range of potentials, and that this is important not only after loading but also for steady state conditions. The effect should be particularly marked on the plateau, since the effect of an electrogenic current is larger when the membrane resistance is larger, as shown by Adrian and Slayman (1966). In frog atrium, treatment with dinitrophenol (DNP) gave somewhat different results; the initial inward current decreased and the steady state outward current increased (Haas etal., 1970). The latter result was attributed to a possible electrogenicpotassium pump which would contribute an inward current during the plateau. Another interesting proposal assigns to the electrogenic sodium pump a rather substantial role in genesis of the action potential (Chapman et al., 1979). In a computer reconstruction, certain cardiac action potentials have been reproduced by assuming that (1) sodium and potassium are the only ions which contribute substantial currents through the membrane, (2) potassium permeability (and that of sodium, after the upstroke) is voltage-dependent, not time-dependent, and (3) an electrogenic sodium pump is responsible for
FIG.3. Effects of dihydroouabain on the current-voltage relation in cardiac Purkinje fibers. The scale for the clamp potential is on the abscissa, and that for the late membrane current is on the ordinate. In (A), open circles show the control I- Yrelation; and solid circles, that recorded after a 2-minute exposure to dihydroouabain. In (B), solid circles show the same data as in (A), and open circles and crosses show the recovery of the I- Yrelation 3 and lominutes, respectively, after returning to Tyrode’s solution. (Reproduced with permission from Isenberg and Trautwein, 1974.)
25.
Na+
PUMP CONTROL OF NERVE AND CARDIAC FIBERS
477
time-dependent changes in membrane current. This would make the electrogenic pump indispensable-and as important as are passive currents-for generation of the action potential.
3. EFFECTS ON PACEMAKER ACTIVITY Under normal conditions, the sinus node is the dominant pacemaker of the heart, and the latent pacemakers in the atria and ventricles are driven by conducted impulses. If sinus node activity or atrioventricular conduction is suddenly blocked (for example by vagal stimulation), the latent pacemakers do not take over immediately. In other words, the latent pacemakers are under some form of suppression. This suppression appears to involve, as its major mechanism, an electrogenic sodium extrusion. Sinus node activity is modulated mainly by the vagus and sympathetic nerves, although Noma and Irisawa (1975) have proposed that changes in pump activity may contribute to control of the rate of discharge. Ventricular pacemakers are accelerated somewhat by sympathetic discharge but are usually kept inhibited by a frequency-dependent mechanism (overdrive suppression). Vagal stimulation, by inhibiting the sinus node, merely reveals the suppression that the sinus node was exerting on the ventricular pacemakers by virtue of its faster rate. This has been demonstrated in several ways (see Vassalle, 1977). For example, in complete atrioventricular blockage the idioventricular rhythm is temporarily suppressed by a period of fast drive. Events underlying the frequency-dependent inhibition of idioventricular pacemakers have been studied in Purkinje fibers perfused in vitro. As shown in Fig. 4, fast drive (overdrive) of a spontaneously firing fiber causes an initial decrease, and a subsequent increase, in the maximal diastolic potential during the drive; the overdrive is followed by a pronounced suppression of spontaneous activity. The suppression is due to a marked flattening of diastolic depolarization. When spontaneous activity resumes, the maximal
7 0
n
1
I 12O/min
DR I V E
50
sec
FIG.4. Overdrive suppression in cardiac Purkinje fibers. Only the lower part of the action potential is shown. The Purkinje preparation was spontaneously active and was driven for 2 minutes at 120 min-'. During the drive, the maximum diastolic potential first decreased below, and then increased above, the predrive value. The cessation of drive was followed by aprolonged suppression of spontaneous activity. (Reproduced with permission from Vassalle, 1970.)
478
MARIO VASSALLE
diastolic potential is still more negative than the control, and several minutes are required for restoration of the initial rate of discharge. While the initial decline in maximum diastolic potential is due to potassium accumulation (see Vassalle, 1977), the subsequent increase in maximal diastolic potential seems to come from sodium pump current. Thus, the increase disappears, or even reverses, in the absence of sodium (replaced by lithium), or in the presence of inhibitors such as dinitrophenol (Vassalle, 1970), strophanthidin (Carpentier and Vassalle, 197 l), antimycin, and iodoacetate (Bhattacharyya and Vassalle, 1980). The increase in maximal diastolic potential is larger when [K+1, has been reduced from 5.4 to 2.7 mM (because membrane resistance doubles in low K + ) (Vassalle, 1970; see also Browning et al., 1979), in the presence of norepinephrine (Carpentier and Vassalle, 197 l), and during fast driving. When [K+], is rather low (Alanis and Benitez, 1967), or ouabain is present (Wittenberg et al., 1972, see below), overdrive does not necessarily cause suppression. The suppression does not result from the release of inhibitory substances, since it cannot be transmitted humorally (Vassalle et al., 1977). Kodama et al., (1 977) reported similar changes in potential (a decrease and then an increase) during overdrive, both in Purkinje fibers and in ventricular muscle fibers; the hyperpolarization decreased in amplitude but lasted longer at a lower (31°C) temperature. Browning et al. (1979) also argued for the action of an electrogenic pump in overdrive suppression, on the basis that hyperpolarization was reduced by low temperature, was abolished by ouabain, and was increased by cesium. The time constant for the decay of hyperpolarization did not depend on the sodium load (the duration or frequency of the drive), in agreement with Thomas’ (1969) earlier observation on neurons. In rat atria, Diacono (1979) found depolarization occurring at the beginning of the drive and again afterward; as usual, the magnitude of these changes was greater with faster or longer drives. The hyperpolarization was decreased in 50% sodium solution and was abolished by ouabain. The membrane resistance decreased during the drive and returned to the control level after the drive, in agreement with the results of Browning et al. (1979). Overdrive of the sinus node is also followed by a brief suppression which has been attributed mostly to the release of acetylcholine (Lu et al., 1965). Recently, Courtney and Sokolove (1979) found that atropine somewhat reduced the overdrive suppression of the sinus node, but the effect was reversibly abolished by ouabain. Exposure to a K+ -free solution accelerated the sinus node; and on return to 5 mM K + , the maximal diastolic potential increased, temporarily suppressing activity. With ouabain present, the effects of restoring normal K+ were abolished. These results resemble the earlier observations of Ito and Surawicz (1977) on Purkinje fibers, where the shift from low K+ to normal K+ produced hyperpolarization, cessation of spon-
25.
Na+
PUMP CONTROL OF NERVE AND CARDIAC FIBERS
479
taneous activity, and shortened action potentials upon resumption of activity. Again, the hyperpolarization was reduced by ouabain. These results, and those of Wiggins and Cranefield (1974), demonstrate that a suppression of spontaneous activity occurs when electrogenic sodium extrusion is stimulated by an increased sodium load, regardless of the method used to load the fiber with sodium. The question should be asked whether an increase in potassium conductance also participates in hyperpolarization after drive (Vick, 1969) or after low-potassium exposure (Ito and Surawicz, 1977). The influx of calcium (as well as that of sodium) should increase during overdrive, and elevated intracellular calcium is known to increase potassium conductance in Purkinje fibers (Isenberg, 1977a,b). Furthermore, on returning to normal K+ , the potassium conductance also should increase. We overdrove Purkinje fibers in normal, high, and low calcium. As expected, calcium changed overdrive suppression in several respects, modifying both the rate of discharge and the threshold (Weidmann, 1955b). However, the increase in maximum diastolic potential after longer drives was as large in low calcium as in high calcium (Vassalle and MUSSO,1978), indicating that the predominant mechanism for increasing membrane potential relates to the pump current rather than an increase in calcium-mediated K+ conductance. CONSEQUENCES 4. PHYSIOLOGICAL There are several distinct physiological implications of these experiments. A major one is that the sinus node, acting through the process of conduction, can inhibit all subsidiary pacemakers by virtue of its faster rate. This is accomplished by overloading the subsidiary pacemaker with sodium and making sodium extrusion electrogenic. When sinus node activity stops, the frequency-dependent suppression of latent pacemakers is removed, allowing latent pacemakers to take over the activation of the heart chambers. Normally, when the sinus node does act as the dominant pacemaker, electrogenic sodium extrusion may play an important role in excitability, because of the relationship between diastolic potential and conduction velocity. The rate of rise in the action potential depends rather steeply on the preceding membrane potential, in the range -60 to -90 mV, which is the range of diastolic depolarization in Purkinje fibers. The more negative the membrane is held by pump current-at the time the cell is activated by conduction-the greater its upstroke velocity and amplitude and (in turn)its conduction velocity. In addition, the pump current acts to shorten the action potential duration (Vassalle and Musso, unpublished experiments) and to prevent the occurrence of conduction blockage at faster rates. It should be stressed that, if an enhancement of pump current results in a
480
MARIO VASSALLE
suppression of spontaneous activity, a decrease in the function of the pump should have an opposite effect. Cardiac glycosides, used clinically, do cause arrhythmias. While this is usually due to the development of a transient inward current (Lederer and Tsien, 1975; Kass et al., 1978a,b), the simultaneous inhibition of the pump removes an inhibitory influence. As shown in Fig. 5 , overdrive in the presence of strophanthidin does not cause hyperpolarization; and it is not followed by suppression, but by an acceleration of the spontaneous discharge. This acceleration occurs because cellular calcium is allowed to rise as a consequence of poisoning of the sodium pump. C. The Measurement of Pump Current Recently the pump current has been measured in cardiac tissues, as Thomas (1969) did in neurons some years ago. Akasu et al. (1978b) found that during readmission of potassium, after a period of exposure to a K + -freesohtion, the membrane potential increased, or-under voltage clamp conditions-an outward current flowed which was eliminated by ouabain. Gadsby and Cranefield (1979) carried out a similar experiment on Purkinje fibers (Fig. 6 ) .The outward current increased with longer exposures to zero K+ .The decay of theoutward current was approximately exponential, with a time constant of approximately 75 seconds, independent of the magnitude of the peak DRIVE 90/min
-
.
CONTROL
STHOFH
FIG.5 . Effect of overdrive in the absence and in the presence of strophanthidin. (Top) The first trace shows two spontaneous action potentials and the beginning of a fast drive at 90min-I. When the 60-second drive was interrupted (at the beginning of the second top trace) a prolonged suppression followed. (Bottom) After the preparation had been perfused for 7 minutes in the presence of strophanthidin (l(r6A4)and its spontaneous rate was faster. The 60-second drive was repeated, and its cessation was followed not by suppression but by an actual acceleration of the rate of discharge. (Recorded by M. Vassalle.)
25.
Na+
481
PUMP CONTROL OF NERVE AND CARDIAC FIBERS
FIG.6 . Changes in membrane voltage and current in response to a short exposure to zero potassium. As shown by the top trace, potassium was reduced from 4 to 0 mM for 1 minute. During the exposure to zero potassium the membrane potential decreased below the control value, and during the recovery it increased above the control value. When the membrane voltage was kept at a constant value by voltage clamp control (third trace), an inward current was recorded during exposure to zero-K+ solution, and an outward current was recorded after K+ restoration (fourth trace). (Reproduced with permission from Gadsby and Cranefield, 1979.)
A 20mV[
I
Clamp off
-
-
ot
1min
0 Clamp on 5mV[
current. It was concluded that the rate of extrusion of sodium was proportional to sodium accumulation. Eisner and Lederer (1979) obtained similar results with sheep Purkinje fibers, where the time constant of the electrogenic current (3.8 minutes) was not too different from the time constant of the decrease in actual [Na+Iimeasured under similar (but not identical) conditions (Deitmer and Ellis, 1978). VI.
CONCLUDING REMARKS
While it is not simple t o dissect pump effects from those of other mechanisms, the role of an electrogenic pump in regulating membrane potential (and therefore excitability) is reasonably well established. The shift in membrane potential caused by the pump current under physiological conditions is in the hyperpolarizing direction and usually involves a few millivolts at rest. However, such a relatively small hyperpolarization can result in marked changes in the excitability process, since several important parameters are modified. These include the resting potential, the threshold potential, the maximal velocity and magnitude of the upstroke, the duration of the action potential, and the pacemaker potential. The role of the pump in excitability is enhanced by a burst of activity, and it is likely that the subsequent moderation of discharge in certain systems (nerve) is part of an overall physiological regulation. In the heart, the pump has an inhibitory influence on the latent pacemaker automaticity, an inhibition which is automatically removed when latent pacemaker automaticity needs to be brought into play. In addition, as the function of the electrogenic pump is enhanced by a period of supernormal activity, a period of reduced activity favors a restoration of the ionic balance. The electrogenic sodium pump would thus help repay the “sodium debt” contracted during the period of enhanced activity.
482
MARIO VASSALLE
ACKNOWLEDGMENTS Original work reported here has been supported by grants from the National Institutes of Health, Heart and Lung Institute and the New York Heart Association.
REFERENCES Adrian, R. H., and Slayman, C. L. (1966). J . Physiol. (London) 184,970-1014. Akasu, T., Omura, H., and Koketsu, K. (1978a). Life Sci. 23,2405-2410. Akasu, T., Ohta, Y., and Koketsu, K. (1978b). Experientia 34,488-490. Alanis, J., and Benitez, D. (1967). Jpn. J. Physiol. 17, 556-571. Baylor, D. A., and Nicholls, J. G. (1969a). J . Physio(. (London) 203, 555-569. Baylor, D. A., and Nicholls, J. G. (1969b). J . Physiol. (London) 203, 571-589. Bhattacharyya, M . L., and Vassalle, M. (1980). Arch. Int. Pharmacodyn. Ter., 246,28-37. Bignami, A., and Palladini, G. (1966). Nature (London) 209,413-414. Blaustein, M. P. (1974). Rev. Physiol. Biochem. Pharmacol. 70,33-82. Browning, D. J., Tiedeman, J. S., Stagg, A. L., Benditt, D. G., Scheinman, M. M., and Straws, H. C. (1979). Circ. Res. 44,612-624. Carpentier, R., and Vassalle, M. (1971). In “Research in Physiology. A Liber Memorialis in Honor of Professor Chandler McCuskey Brooks” (F. F. Kao, K . Koizumi, and M. Vassalle, eds.), pp. 81-98. Gaggi, Bologna. Chapman, J. B., Kootsey, J. M., and Johnson, E. A. (1979). J . Theor. Biol. 80,405-424. Cohen, I., Daut, J., and Noble, D. (1976). J . Physiol. (London) 260, 75-103. Courtney, K. R., and Sokolove, P. G. (1979). J. Mol. Cell. Cardiol. 11,787-794. Deitmer, J. W., and Ellis, D. (1978). J . Physiol. (London) 284, 241-254. DCleze, J. (1960). Circ. Res. 3, 553-557. Diacono, J. (1979). J . Mol. Cell. Cardiol. 11,5-30. Eisner, D. A., and Lederer, W. J. (1979). J . Physiol. (London) 294, 279-301. Ferrier, G. R . (1977). Prog. Cardiovas. Dis. 19,459-474. Gadsby, D. C., and Cranefield, P. F. (1979). Natl. Acud. Sci. U.S.A. 76, 1783-1787. Glitsch, H. G. (1969). Pfleugers Arch. 307, 29-46. Glitsch, H . G. (1979). A m . J . Physiol. 236, H189-HI99. Glitsch, H. G., and Klare, J. (1977). PfleugersArch. 368, Suppl. R3. Glitsch, H. G., Grabowski, W., and Thielen, J. (1978). J . Physiol. (London) 276, 515-524. Haas, H. G. (1972).In “Electrical Phenomenain the Heart” (W. C. deMello, ed.), pp. 163- 189. Academic Press, New York. Haas, H. G., Kern, R., and Einwachter, H. M. (1970). J. Membr. Biol. 3, 180-209. Hiraoka, M., and Hecht, N.H . (1973). Pfluegers Arch. 339,25-36. Holloway, S. F., and Poppele, R. E. (1978). Brain Res. 154, 144-147. Isenberg, G. (1977a). Pfluegers Arch. 371, 71-76. Isenberg, G. (1977b). Pfluegers Arch. 371, 77-85. Isenberg, G., and Trautwein, W. (1974). Pfluegers Arch. 350,41-54. Ito, S., and Surawicz, B. (1977). Circ. Res. 41,799-807. Jansen, J. K. S., and Nicholls, J . G. (1973). J . Physiol. (London) 229,635-655. Kass, R. S., Lederer, W. J., Tsien, R. W., and Weingart, R. (1978a). J . Physiol. (London) 281, 187-208. Kass, R. S., Tsien, R. W., and Weingart, R. (1978b). J . Physiol. (London) 281,209-226. Kerkut, G. A., and York, B. (1971). “The Electrogenic Sodium Pump.” Scientechnica, Bristol. Kernan, R. P. (1970). In “Membranes and Ion Transport” (E. E. Bittar, ed.), Vol. 1, pp. 395431. Wiley (Interscience), New York.
25.
Na+
PUMP CONTROL OF NERVE AND CARDIAC FIBERS
483
Kodama, I., Hirata, Y., Ando, S., Toyama, J., and Yamada, K. (1977). J. Mol. Cell. Cardiol. 9, Suppl. 3, 36. Langer, G. A. (1977). Fed. Proc. Fed. Am. SOC.Exp. Biol. 36,2231-2234. Lederer, W. J., and Tsien, R. W. (1976). J . Physiol. (London) 263, 73-100. Lieberman, M., Horres, C. R., Alton, J. F., and Johnson, E. A. (1977). Proc. Int. Union Physiol. Sci. 13,446. Lu, H. H., Lange, G., and Brooks, C. McC. (1965). Circ. Res. 17,460-471. McAllister, R . E., Noble, D. and Tsien, R. W. (1975). J. Physiol. (London) 251, 1-59. Mannard, A., Rajchgot, P., and Polosa, C. (1977). Brain Res. 126,243-261. Meech, R. W. (1974). J . Physiol. (London) 237,237-277. Noma, A., and Irisawa, H. (1974). Pfluegers Arch. 351, 177-182. Noma, A., and Irisawa, H. (1975). Pfluegers Arch. 358,289-301. Page, E., and Storm, S. R. (1965). J . Gen. Physiol. 48,957-972. Schwartz, A., Lindenmayer, G. E., and Allen, J. C. (1975). Pharmacol. Rev. 27,3-133. Skok, V. I . , Storch, N. N., and Nishi, S. (1978). Neuroscience 3,697-708. Spencer, W. A., and Kandel, E. R. (1969). In “Basic Mechanisms of the Epilepsies” (H. H. Jasper, A. A., Ward, and A. Pope, eds.), pp. 575-603. Little, Brown, Boston, Massachusetts. Tamai, T., and Kagiyama, S. (1968). Circ. Res. 22,423-433. Thomas, R. C. (1969). J. Physiol. (London) 201,495-514. Thomas, R. C. (1972). Physiol. Rev. 52,563-594. Tower, D. B. (1969). In “Basic Mechanisms of the Epilepsies” (H. H. Jasper, A. A. Ward, and A. Pope, eds.), pp. 61 1-638. Little, Brown, Boston, Massachusetts. Vassalle, M. (1970). Circ. Res. 27, 361-377. Vassalle, M. (1977). Circ. Res. 41,269-277. Vassalle, M., and Musso, E. (1978). Int. Congr. Pharmacol. 7th p. 42. (Abstr). Vassalle, M., Krellenstein, D. J., Pliam, M. B., and Brooks, C. McC. (1977). J. Mol. Cell. Card i d . 9,921-93 l . Vick, R. L. (1969). Am. J . Physiol. 271,451-457. Wallick, E. T., Lane, L. K . , and Schwartz, A. (1979). Ann. Rev. Physiol. 41, 397-41 1 . Weidmann, S. (1955a). J . Physiol. (London) 127,213-224. Weidmann, S. (1955b). J. Physiol. (London) 129,568-582. Wiggins, J . R., and Cranefield, P. F. (1974). J. Gen. Physiol. 64,473-493. Wittenberg, S. M., Gandel, P., Hogan, P. M., Kreuzer, W., and Klocke, F. J. (1972). Circ. Res. 30, 167-176.
This Page Intentionally Left Blank
CURRENT TOPICS IN MEMBRANES A N D TRANSPORT. VOLUME 16
Chapter 26 Pumps and Currents: A Biological Perspective’ FRANKLIN M . HAROLD Division of Molecular and Cellular Biology National Jewish Hospital and Research Center Denver, Colorado and Department of Biochemistry, Biophysics and Genetics University of Colorado Medical School Denver, Colorado
I. 11. 111.
IV. V. VI . VII.
Introduction ....................................................................................... The Role of Ion Currents in the Metabolic Economy of Bacteria ........................ Ion Currents and Energy Coupling in Eukaryotic Cells ................................... Cellular Homeostasis ............................................................................ .................................. Calcium Currents as Biological Signals ............................. Transcellular Currents and Morphoge A Sense of Direction ............................................................................. References .............................................
1.
485 487 492 495 501 505 510 513
INTRODUCTION
The electrical activities of living things were, until very recently, of interest chiefly to neurophysiologists and to the handful of students concerned with such apparent curiosities as electric eels and excitable plants. To them we owe the rigorous formulation of laws that govern the movement of ions across biological membranes and the potentials and currents that arise therefrom. In consequence, electrophysiology became one of the most sophisticated branches of biological science but rather an arcane one, seemingly remote from the 1 Dedicated to Peter Mitchell, as a token of respect and affection, on the occasion of his sixtieth birthday.
485
Copyright 8 1982 by Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 0-12-153316-6
486
FRANKLIN M. HAROLD
concerns of cell biologists, microbiologists, or biochemists. The recent melding of this specialized subject into the wider stream of molecular physiology affords a striking illustration of how science grows: It resulted, not from refined observations and advanced instrumentation in the tradition established by students of muscle and nerve, but from the impact of a novel theory-and one devised to deal with an unrelated issue, at that. The enormous expansion of the scope of electrophysiology during the past decade reflects primarily its anastamosis with bioenergetics. The central issue in molecular bioenergetics is embodied in the phrase ‘‘energy coupling.” All the familiar activities of living things entail the consumption of energy that must be supplied by metabolism. How, then, is the free energy released by fermentation, respiration, and photosynthesis captured and harnessed t o support such diverse cell functions as the biosynthesis of macromolecules, active transport of ions and nutrients, movement of muscles and flagella, and assembly of the cell itself? A partial answer came to hand with the recognition (Lipmann, 1941) that ATP serves as the central energy currency: The great metabolic highways were found to generate ATP which, in turn, drives the biosynthesis of molecules large and small, the contraction of muscle, and even the transport of Na+ and K + across animal cell membranes. Yet there remained two major lacunae to be filled. One was the mechanism of ATP generation during oxidative and photosynthetic phosphorylation, catalyzed by enzymes built into the membranes of mitochondria and chloroplasts, which defied the best efforts of a generation of biochemists. The other was the mechanism of those vectorial active transport processes which seemed but indirectly linked to ATP. As it turned out, both puzzles led t o recognition of the fundamental role that ion pumps and currents play in cell physiology. In 1961, Crane (Crane, 1962) proposed that transport of sugars across epithelial cell membranes is effected by cotransport with Na+ and energized by the electrochemical potential gradient of Na+. About the same time Peter Mitchell, building upon his earlier insights into the relationships between enzymic catalysis and transport (Mitchell and Moyle, 1958), proposed that oxidative and photosynthetic phosphorylation are energized by currents of protons generated, respectively, by the respiratory chain and by the photosynthetic apparatus (Mitchell, 1961). The chemiosmotic hypothesis was soon broadened with the suggestion that the accumulation of nutrients by bacteria is linked to a proton current across the cytoplasmic membrane (Mitchell, 1962a). With these proposals, ion pumps and ion currents ceased t o be the private preserve of neurophysiologists and entered the purview of anyone concerned with how living cells work. One consequence of the increasing interplay between ion transport and bioenergetics has been to transform our views of the functions of ion transport in cellular physiology. This article does not pretend to be a com-
26. PUMPS AND CURRENTS
487
prehensive review of the torrential literature, though I d o hope to provide some guidance for those who wish to delve deeper into the role of ion currents in the performance of chemical, osmotic, and mechanical work; in biological signals and triggers; and in the largely mysterious processes by which cells assemble themselves in space, grow, divide, and develop as integrated units. Rather, it is an attempt at synthesis, part of a continuing effort to define the meaning of a scientific revolution that has dominated my professional life and in which I have been privileged to play a minor part. And if I focus almost exclusively on microorganisms, it is not entirely because they are the organisms with which I am most familiar: The choice also reflects the Baconian maxim, that the Nature of Things is commonly better perceived in small than in great. A few terms should be explained, if not rigorously defined. The transport nomenclature follows Mitchell (1967) in distinguishing between two kinds of processes: primary transport, directly linked to a chemical reaction (pumps), and secondary transport, not connected with any chemical reaction but often linked to an ion gradient. “Energy” and “work” are used rather loosely, as is customary in biology: Endergonic reactions or processes (producing a positive change in free energy) cannot proceed spontaneously and are said to require an input of energy or the performance of work. Energy is defined as the capacity to do work.
II. THE ROLE OF ION CURRENTS IN THE METABOLIC ECONOMY OF BACTERIA Bacteria are in general much too small for microelectrodes; electrical potentials must therefore be inferred from indirect measures such as the distribution of permeant ions and the response of various fluorescent probes (Rosen and Kashket, 1978; Rottenberg, 1979; Waggoner, 1979). Ultrafine microelectrodes suitable for use with giant cells of Escherichia cofi are even now being developed (Felle et af., 1980), and it is reassuring t o find that the data fall within the expected range. Despite these technical handicaps, it is the bacteria which illustrate most clearly how ion currents link cellular metabolism to the performance of work. The discovery of these principles is the fruit of the past decade and stems very largely from the application of chemiosmotic theory. It is no longer possible to do justice to this flourishing field of research within the confines of a brief essay. I therefore propose merely to sketch our present understanding of the role of ion currents in the metabolic economy of the very simplest cells and to point out some fundamental issues that are now being addressed. Documentation will be found in major reviews that cover
488
FRANKLIN M. HAROLD
chemiosmotic theory in general (Mitchell, 1966, 1970, 1976; Harold, 1972, 1977a), its application to bacteria in particular (Haddock and Hamilton, 1977; Harold, 1972, 1977b), respiration and oxidative phosphorylation (Boyer et al., 1977; Haddock and Jones, 1977; Hinkle and McCarty, 1978), photosynthesis (Crofts and Wood, 1978; Dutton and Prince, 1978), and active transport (Eddy, 1978; Hamilton, 1975; Rosen, 1978; Simoni and Postma, 1975). Mitchell originally proposed, and subsequent research has amply verified, that bacteria expel H + across the cytoplasmic membrane by means of one or more electrogenic pumps that transduce chemical energy into the electrochemical potential of protons (Fig. 1). These are the chemiosmotic reactions that have an intrinsic direction in space and which lend the theory its name. The cytoplasmic membrane has a very low conductivity and capacitance. The transport of protons therefore generates an electrical potential, cytoplasm negative; under conditions such that the movement of another ion compensates for that of protons, the cells also generate a pH gradient, cytoplasm alkaline. The sum of these two gradients is designated the proton motive force, a term closely related to the electrochemical potential of H +: A p = A& + /F=$ - (2.3 R T/F)ApH = $ - ZAP H
(1)
where A p is the proton motive force, ApH+is the electrochemical potential of
FIG. 1. Energy transduction by proton circulation in bacteria. The diagram illustrates generation of a proton current by vectorial electrogenic H + extrusion through the respiratory chain and completion of the current loop by several paths that perform useful work: protontranslocating adenosine triphosphate (oxidative phosphorylation), pyridine nucleotide transhydrogenase, the flagellar motor, and several kinds of proton-linked transport systems. Also shown is the generation of a secondary sodium current by Na+ /H+ antiport and an example of sodium-linked substrate transport. S, Substrate. (Reproduced from Harold, 1977a. with permission of Annual Reviews, Inc.)
489
26. PUMPS AND CURRENTS
H +,F i s Faraday’s constant, I,5 is the membrane potential, ApH is the pH gradient, and Z= 2.3 RT/F=60mV. The protons expelled by the pump “seek to” return across the membrane, down their electrochemical potential gradient. The free energy potentially available from completion of the proton circuit is captured by channeling the flow of protons through diverse macromolecular devices cunningly articulated so as to link the passage of protons across the membrane to the performance of useful work (Fig. 1). Examples of such devices include numerous transport carriers that mediate symport (cotransport) or antiport of particular metabolites with protons; the driving force upon the proton determines the maximal concentration gradient the metabolite can attain. For example, for symport of a neutral substrate S with one proton, log( [S]j/[S],)=(II,-ZApH)/Z=Ap/Z
(2)
where [SIiand [S], are the concentrations of the substratein the cytoplasm and medium, respectively. Another example is the “motor” built into the cytoplasmic membrane at the base of each flagellum, which transduces the energy of the proton current into the rotary motion of the flagellum. Finally, there are devices that perform chemical work. The most important of these is the membrane-bound ATPase complex which reversibly couples the hydrolysis of ATP to the translocation of H + and ultimately accounts for oxidative phosphorylation. If, as now seems established, the ATPase transports two protons per cycle, the relationship between the cytoplasmic phosphorylation potential AGpand the proton motive force can be described by the formula AGp/F= 2Ap
(3)
where AG,= G,’+ R T In ([ATP]/[ADP][P,]). Another transducing device that performs chemical work at the expense of the proton current is the pyridine nucleotide transhydrogenase, which generates reducing power for biosynthesis in the form of NADPH (Rydstrom, 1977). Much effort has been devoted to measuring the electrical and thermodynamic parameters of the proton current, upon which the performance of useful work ultimately depends. Membrane potentials in bacteria range from - 90 mV to about - 180 mV (as high as - 250 mV in chromatophores from photosynthetic bacteria). Gradients of pH may be on the order of 1 or 2 units (60-120 mV). The total proton motive force depends greatly upon the organism and theconditions, but values of about -200mVarecommon. Therateof proton pumping is also variable; in E. coli, for example, the proton flux can be estimated from the rate of respiration to be about 100 peq/cm2 sec, or 10 pA/cm2. Ionophores of various kinds, as well as bacteriocins, dissociate metabolism from work in a manner that is generally consistent with their
490
FRANKLIN M. HAROLD
capacity to dissipate the proton motive force. Not always is the measured proton motive force sufficient t o perform the task assigned to it. Nevertheless, the weight of the evidence is now so great that the controversy over the validity of chemiosmotic coupling principles has largely subsided and attention has begun to shift to a new generation of issues. Let me briefly touch upon some of these.
1. Molecular mechanisms of primary and secondary ion transport systems: It will be seen from Fig. 1 that bacterial membranes must be studded with dozens of ion-translocating pathways, primary and secondary, electrogenic as well as electroneutral. Their existence is in many cases well established, but their molecular mechanisms remain virtually unknown. Consider, for example, the primary electrogenic pumps that make up the engine of the bacterial economy. We presently recognize four distinct kinds: the reversible proton-translocating ATPase; a variety of oxidation chains that link the extrusion of protons to the oxidation of reduced substrates (e.g., NADH to oxygen or to fumarate); the photosynthetic apparatus, based on bacteriochlorophyll and a redox chain; and bacteriorhodopsin, unique to the halobacteria (these remarkable organisms have been the subject of several recent reviews, e.g., Lanyi, 1978; Stoeckenius el al., 1978). The molecular architecture of these enzyme complexes is fairly well defined, but in no single case can we presently formulate a molecular mechanism that accounts satisfactorily for both the enzymological and vectorial characteristics of the process. Even so basic a parameter as the number of protons transported per ATP split, or per coupling site, has recently become controversial, and we are no longer certain whether cytochrome oxidase translocates only electrons, or protons as well. We are equally ignorant of the molecular basis of the secondary carriers that transmit protons back across the membrane and convert A&+ into useful work. The celebrated lac permease has been the subject of hundreds of papers (recently the protein has even been sequenced), yet we have little idea how it works. This applies afortiori to the flagellar motor and its modulation by attractants and repellants (for a provocative model see Lauger, 1977). In truth, even the path that the protons follow from source to sink may not always be quite as simple as elementary chemiosmotic theory supposes. Some of the discrepancies that trouble us may be resolved by admitting the existence of rather more intimate contacts than those afforded by a current of protons through the aqueous bulk phase; others may require us to invoke local electric fields. These caveats seem particularly pertinent in the case of oxidative and photosynthetic phosphorylation, where the measured proton motive force is consistently too small to sustain the phosphorylation potential (see, for exam-
26. PUMPS AND CURRENTS
49 1
ple, De Pierre and Ernster, 1977; Boyer et a/., 1977; Williams, 1978). These are burning issues that have come to dominate the attention of biochemists, but they fall outside the scope of this article, oriented as it is toward the cellular rather than the molecular level. 2. Energy coupling by currents of sodium and other ions: Although H + is the central coupling ion in bacterial metabolism, it is not the only one. In a number of cases, transport of sugars and amino acids is effected by symport with Na+ and is energized by the electrochemical potential of Na+ (Lanyi, 1979). Animal cells have long been known to expel Na+ by a primary pump, the Na+,K+-ATPase.Bacteria lack this enzyme; instead, Na+ is generally expelled by secondary antiport for protons, as indicated in Fig. 1 . Thus, even Na+-linked transport systems are ultimately driven by the proton circulation. This uniform and widely accepted pattern may, however, require revision. MacDonald and his associates have very recently discovered (MacDonald et a/., 1979) that halobacteria possess what appears t o be a primary electrogenic sodium pump energized by light. It remains to be seen whether this unprecedented capacity occurs in other bacteria as well or is confined to halobacteria. In any event, generalizations about “bacteria” must be tempered by the recognition that our concept of bacteria as one of the five taxonomic kingdoms is itself under challenge. Woese and Fox (1977) have argued very persuasively that certain bacteria (including halobacteria, methanogens, and others) differ so markedly from their conventional congeners as to warrant creation of a new kingdom, the archaebacteria, that diverged very early in the history of life. It will be most interesting to learn whether archaebacteria and eubacteria differ systematically in the way they generate and utilize ion currents. 3 . In search of new patterns: It is reasonable to ask whether bacterial proton circulation performs work functions that have hitherto escaped recognition and whether additional ion circulations remain to be discovered. There are, indeed, some straws in the wind. Two functions that clearly are dependent on proton circulation are the regulation of cytoplasmic pH and of osmotic turgor; both involve potassium transport as well and will be considered in Section IV. The suggestion has been advanced that chemotaxis is specifically controlled by calcium movements (Section V), but this remains to be established. The involvement of the proton motive force in nitrogen fixation also remains tenuous. Finally, since proton circulation is the engine of the cell, one would expect bacteria to monitor it as an index of their metabolic status, just as they appear to monitor the adenylate charge (Atkinson, 1977). The finding that reagents which dissipate the proton motive force often stimulate the phosphotransferase enzymes of glucose metabolism points in this direction. It would also be reasonable to look for
492
FRANKLIN M. HAROLD
control of adenylate cyclase by the proton motive force, and some evidence to this effect has been obtained. Nevertheless these are nebulous ideas badly in need of factual support. One can also ask whether ion currents are obligatory to the growth and division of bacterial cells, apart from their role in energy transduction. Apparently the answer here is no (Harold and Van Brunt, 1977). The fermentative bacterium Streptococcus faecalis will grow in the presence of the ionophore gramicidin under conditions so arranged that membrane potential, pH gradient, and proton motive force are all zero. To be sure, such cells are crippled and require a rich medium of alkaline pH and a high potassium content. But they do grow, divide, and are morphologically normal: Evidently, currents of H + , K + , and Na + are not required to synthesize macromolecules and to assemble the fabric of a bacterial cell in an orderly fashion.
Ill.
ION CURRENTS AND ENERGY COUPLING IN EUKARYOTIC CELLS
The plasma membrane of prokaryotic cells performs a number of functions (including oxidative phosphorylation and motility) that eukaryotes assign to intracellular organelles. The architectural complexity so apparent in any electron micrograph extends to the pattern of ionic currents, which features multiple independent circuits within a single cell (Fig. 2).
ATP
FIG.2. Multiple ion circulations in eukaryoticcells. (A) A fungal hypha with mitochondrion and ATP-driven proton circulation. (B) An animal cell with mitochondrion, Na+,Kf -ATPase, and a calcium ATPase.
26. PUMPS AND CURRENTS
493
After nearly two decades of disputation, researchers working with mitochondria and chloroplasts have now generally accepted the proposition (Mitchell, 1961,1966) that energy coupling in these organelles is based on proton circulation in the prokaryotic mode. The joint communique recently published by six leading investigators (Boyer et al., 1977) marks a welcome historic end point, at the same time serving as the point of departure for renewed debate over the molecular mechanics of electron transport and phosphorylation. Other internal organelles also possess private ion circulations, or at least electrogenic ion pumps. A proton-translocating ATPase is well documented in certain storage granules (Njus and Radda, 1977) and may also explain the low pH of lysosomal contents. Proton pumps may prove to be a feature of certain cytoplasmic vesicles of fungi and are likely to be found in the tonoplast of plant cells (Smith and Raven, 1979). Calcium pumps hold a special place and will be considered below. This article will center on the plasma membrane, for it is at this anatomical level that we can best appreciate the variety of physiological functions that ion currents have come to serve. Fifteen years ago, in a study that remains a landmark, Slayman (1965a,b) inserted microelectrodes into hyphae of the fungus Neurospora crassa and discovered that they maintained a membrane potential of approximately - 200 mV. Part of this potential could be attributed to K + diffusion, but most of it was independent of K + and extremely sensitive to inhibitors of mitochondrial respiration. Slayman proposed that the potential is generated by an electrogenic ion pump closely linked to respiratory metabolism; it was one of the first serious challenges to the view prevailing at the time, that ion pumps are electroneutral and that bioelectric potentials arise by diffusion of ions down their potential gradient. Subsequent research extended this discovery and linked it to our emerging understanding of energy coupling by ion currents. There is now strong evidence that the potential is generated by an electrogenic, protontranslocating ATPase plugged through the plasma membrane (Slayman et al., 1973; Scarborough, 1976; Gradmann et al., 1978). Although the plasma membrane ATPase is functionally analogous to that from mitochondria, the two enzymes differ in molecular constitution and probably in their mechanism of action as well (Scarborough, 1977; Gradmann et al., 1978; Bowman and Slayman, 1979). The proton current generated by the ATPase, on the order of 20pA/cm2 (200 peq H +/cm2 sec), in turn drives the accumulation of various nutrients (Fig. 2A). Symport of glucose with protons has been carefully characterized (Slayman and Slayman, 1974), and there is evidence implicating the proton current in the transport of other nutrients as well, including phosphate and ammonium ion (Slayman, 1977). Much the same pattern has emerged from studies with yeast. Extensive studies by Eddy and his associates (Eddy, 1978) have documented the role of
494
FRANKLIN M. HAROLD
proton circulation in the accumulation of amino acids, sugars, phosphate, and potassium (see also Harold, 1977a; Poole, 1978). An ATPase presumed to translocate protons has been isolated from yeast plasma membrane and shown to resemble that from Neurospora (Dufour and Goffeau, 1978; Willsky, 1979). The unicellular green alga Chlorella illustrates that photosynthetic protists also rely on proton circulation to couple metabolism to the accumulation of nutrients, sugars in this instance (Komor and Tanner, 1976; Poole, 1978). A proton-translocating ATPase has not yet been reported from Chlorella. This enzyme has, however, been thoroughly studied by electrophysiological methods in the giant cells of the green algae Chara and Nitella and shown to be an electrogenic, ATP-driven proton pump that accounts for the membrane potential of these cells (Hope and Walker, 1975; Poole, 1978; Shimmen and Tazawa, 1977; Tazawa, this volume). Evidently, the basic pattern is once again that of an ATP-driven proton circulation. At first sight the metabolic economy of eukaryotic protists appears to be homologous to that of bacteria, but this may be deceptive. Two major differences should be emphasized. First, there is no convincing evidence for the existence of ion-translocating redox pathways across the plasma membrane of any eukaryotic cells, high or low; at this time, only ATPases are known to generate ion currents. Second, the proton-translocating ATPase of eukaryotic plasma membranes appears to belong to a molecular family distinct from that made up by the F,-F, ATPases of bacteria, mitochondria, and chloroplasts. As the data now stand, they suggest that the eukaryotic ATPase consists of a single large polypeptide that spans the membrane and also bears the catalytic site, that the reaction involves a phosphorylated intermediate, and that it may mediate the translocation of a single proton rather than two (Bowman and Slayman, 1979; Scarborough, 1977; Gradmann et al., 1978; Dufour and Goffeau, 1978; Willsky, 1979). These fundamental differences appear to me t o be one more reason to reconsider the evolutionary origin of eukaryotic cells from prokaryotic ancestors. Are proton pumps the sole ionic mechanism of energy coupling at the plasma membrane of lower eukaryotes? Probably not. Gradmann (1976) has characterized an electrogenic chloride pump in the marine alga Acetabularia. This pump, whose biochemical basis is quite unknown, evidently makes a major contribution to the membrane potential and thus contributes to the electrochemical driving force across the membrane. A similar pump may occur in at least some fungi and plants (Poole, 1978). At this time we know very little about the transport of potassium, sodium, and calcium across protistal plasma membranes; primary cation pumps as well as secondary porters are certainly conceivable, particularly for sodium. Finally, I would point out that the genesis of ion gradients and electrical potentials by ciliated protozoa and
495
26. PUMPS AND CURRENTS
amebas is not well understood, even though they underlie the remarkable mechanisms for locomotor control and stimulus-response coupling that we shall outline in Section V. There appears to be no evidence for a proton pump; the resting potential, at least in Paramecium ( - 20 t o - 40 mV), is thought to arise primarily from the outward diffusion of potassium. Gradients of K + and Ca2+play as important a role in the life of these predatory animals as the proton circulation does in sedentary saprophytes, and their origin deserves serious attention. Turning very briefly to multicellular organisms, there is mounting evidence that green plants retain the ancestral pattern based on proton circulation (Poole, 1978). Animal cells, in contrast, rely primarily if not exclusively upon the circulation of Na+ generated by a primary electrogenic Na+,K+-ATPase (Fig. 2B); the functions of the sodium current in nutrient transport, osmotic control, and electrical activity are well known and fall, in any case, outside the scope of this article. The sodium circulation surely reflects the sodium-rich environment provided by the internal fluids. One would certainly like to know at what stage of evolution the sodium pump and sodium current arose and whether there is any evolutionary significance to the molecular resemblance between the eukaryotic proton-translocating ATPase and the sodium pump.
IV.
CELLULAR HOMEOSTASIS
Single cells, no less than multicellular organisms, must maintain a characteristic and relatively constant internal environment compatible with the entire range of physiological functions, including growth. Regulation of cytoplasmic volume and ionic composition is, of course, one of the essential duties of the cytoplasmic membrane; it provides new and instructive examples of ion transport as the link between metabolism and the performance of useful work. This section will touch upon the relationship of ion transport to osmotic water flow, the regulation of cytoplasmic pH, and other aspects of cellular homeostasis. The presence of macromolecules in the cytoplasm (not to speak of small osmolites) generates a colloid osmotic pressure across the plasma membrane, resulting in the influx of water and a tendency of the cell to swell. To cope with this, at least three distinct mechanisms arose early in evolution. First, water itself (more precisely, a dilute fluid) is exported with the aid of a contractile vacuole; ciliates and perhaps some fungal zoospores illustrate this solution. Second, a suitable solute is exported; animal cells limit swelling by rendering themselves effectively impermeable to sodium with the aid of the N a + , K+-ATPase. Finally, in most bacteria, fungi, and algae a strong cell wall compensates for the osmotic pressure, restricting osmotic water flow and swel-
496
FRANKLIN M. HAROLD
ling. The resulting turgor pressure keeps the plasma membrane tightly appressed to the cell wall; it is responsible for maintaining the form and rigidity of the organism, provides the driving force for expansion during growth, and may directly perform mechanical work (as in the opening of stomata). On the debit side, excessive turgor renders the organism vulnerable to osmotic bursting, especially at the growing tips where the wall is weakened by the insertion of new elements. All this implies the need for mechanisms that monitor cell volume or pressure and respond t o deviations in the appropriate direction. The brief summary that follows leans on recent reviews by Gutknecht et al. (1977), Gutknecht and Bisson (1977), and Zimmerman (1978). It is now generally agreed that the flow of water itself is strictly passive and occurs in response to a gradient of chemical activity or of pressure, or (because of solute-solvent interactions) to the flow of a specific solute. In microorganisms, at least, the first of these seems by far the most important, since one often finds that the concentration of internal solutes, both organic and inorganic, shifts in response to changes in the external osmotic pressure in the direction that tends t o restore the original turgor pressure. We shall focus here on inorganic ions and particularly on K + because it is the predominant cytoplasmic ion and appears to have a special relationship to osmotic control. This is certainly true of E. coli, which maintains a turgor pressure of about 3 atm. Most of this is accounted for by K + whose concentration ranges from 0.15 to 0.6 M , depending on the external osmolarity; the charge is balanced by a mixture of organic anions. The importance of K + is emphasized by the presence of two powerful transport systems that have been meticulously characterized by Epstein and his associates (Epstein and Laimins, 1979). One, designated Kdp, has a very high affinity and utilizes ATP as the energy source; recent studies indicate that it is a primary ion pump whose mechanism involves a phosphorylated intermediate (Epstein et a /., 1978; Laimins et al., 1978; Wieczorek and Altendorf, 1979). The other, labeled TrKA, is of modest affinity, but its rate is sufficient to replace the entire K + pool in minutes; its operation requires both ATP and the proton motive force (Rhoads and Epstein, 1977). Recent studies with S.faecalis suggest that it may be a secondary porter, modulated by ATP but energized by the proton motive force (Bakker and Harold, 1980). The relevant point here is that both systems quickly respond to an increase in the external osmolarity with a net uptake of potassium, until turgor has been restored. How the cell senses the shift is unknown, but it is reasonable to postulate a receptor in the plasma membrane that monitors turgor and controls the activity of the K + transport systems. The same signal should, directly or indirectly, elicit corresponding shifts in metabolism to provide organic osmolites, both anion and electroneutral (Measures, 1975), but not much is known about this either.
26. PUMPS AND CURRENTS
497
With reference now to eukaryotic cells, little can be said about molecular mechanisms of osmotic control in fungi and protozoa. We are, however, quite well informed concerning the control of turgor, volume, and ionic composition in the cells of giant algae, whose osmotic relations are dominated by a large central vacuole filled with a hyperosmolar salt solution. As an example, consider the marine alga Valonia which has been extensively studied, most recently in the laboratories of Gutknecht and of Zimmermann. The organism responds efficiently t o changes in the external osmolarity by adjusting the composition of the vacuolar sap, consisting chiefly of K + and CI-. There appears to be a primary electrogenic K + pump in the tonoplast whose activity is directly controlled by the turgor pressure: Low turgor stimulates the pump, high turgor inhibits it. Chloride apparently follows passively as the counterion to K + , while movements of Na+ make but a minor contribution (Gutknecht et al., 1977; Gutknecht and Bisson, 1977; Hastings and Gutknecht, 1976). Two other marine algae, Codium and Halicystis, likewise monitor the turgor pressure and respond to low pressure by increased transport of ions into the vacuole. In these, however, the primary pump controlled by the turgor pressure transports chloride rather than potassium. Excessive turgor can cause bursting of the wall; mechanisms for relieving excess turgor involve the controlled release of salts from the vacuole, both in marine algae and in those that grow in fresh water. It is increasingly likely that the long-known action potentials of some giant algal cells reflect just this phenomenon. In Chara and Nitella, for instance, a variety of stimuli have been found to elicit a spreading depolarization of the plasma membrane. For technical reasons it is convenient to trigger the action potential electrically, but the finding (Zimmermann and Beckers, 1978) that injection of fluid into the vacuole is also effective suggests that the physiological stimulus may be the turgor pressure. The action potential results from a transient rise in the chloride permeability; there ensues a substantial efflux of chloride, accompanied by smaller amounts of K + (Hope and Walker, 1975). Part of the chloride efflux is electrically uncompensated, and it is this net movement of charge that explains the depolarization of the membrane. A somewhat similar depolarization due to chloride efflux occurs in Acetabularia, but in this case the underlying mechanism is probably a change in the activity of the primary chloride pump (Gradmann, 1976). In both cases, the loss of K + and C1- is large enough to produce a significant reduction in the internal concentration, hence in the osmotic pressure (Hope and Walker, 1975; Mummert and Gradmann, 1976). The relationship is particularly striking in developing embryos of the marine brown algae Fucus and Pelvetia, which again maintain turgor by the uptake of K and C1-. In the course of their studies on the extracellular electric currents generated during development (Section VI), Nuccitelli and Jaffe ( 1976a) discovered that the embryos periodically emitted large current pulses; +
498
FRANKLIN M. HAROLD
the current was carried by the efflux of chloride from the growing tip, accompanied by a loss of K + both at the tip and elsewhere on the cell. Pulsing could be induced by small decreases in the external osmotic pressure, and the amounts of K + and C1- released were approximately such as to restore the original turgor pressure. Nuccitelli and Jaffe (1976b) suggested that the growing tip serves as a transducer that senses turgor pressure and responds by triggering chloride release. It appears, then, that many algae can monitor the turgor pressure (or a related parameter) and respond appropriately: Low turgor stimulates the active transport of ions into the vacuole, excessive turgor elicits controlled leakage of ions outward. To appreciate the physiological importance of these responses let us recall that, apart from enabling the organisms to cope with changes in the external osmolarity, they govern growth: Both extension of the cell wall and expansion of the huge vacuole are ultimately effected by these ion movements. How can cells measure turgor? The answer is not known, but it seems very likely that the sensor is located in the plasma membrane and registers changes in the geometric dimensions of the membrane resulting from compression, stretching, or the manner in which the flexible membrane is pressed against the rigid wall. Zimmermann (1978) has developed an electromechanical model explaining how such changes in dimension modify the electrical properties of the membrane; indeed, the observation (Zimmermann and Beckers, 1978) that injection of fluid into the vacuole of Chara is instantaneously reflected in the electrical potential across the plasma membrane can be taken as supporting this hypothesis. Alternatively, a localized distortion could be transmitted to adjacent ion pumps or leaks by altering the conformation of these macromolecular structures, by means of intracellular messengers such as cyclic nucleotides, or, as proposed by Nuccitelli and Jaffe (1976a), by eliciting a local current of calcium (Section V). Whatever the mechanism may be, economy of hypotheses suggests that it should also be capable of controlling the synthesis of organic osmolites. Let us now turn to quite another measure of cellular well-being, the cytoplasmic pH. Surprisingly little attention has been paid to the regulation of pH, even though it has long been known to hover near neutrality over a wide range of external pH values. In their valuable recent review, Smith and Raven (1979) cite data to this effect from both prokaryotes and eukaryotic algae. Cellular buffers are insufficient to account for this constancy, and so are metabolic mechanisms. Most likely, the control of cytoplasmic pH is another of the manifold functions of ion pumps, and particularly of proton pumps (Mitchell, 1966, 1970; Skulachev, 1978); indeed, Raven and Smith (1976) have argued that the need to control the pH of the internal environment in which metabolic enzymes function was the evolutionary origin of chemiosmotic coupling mechanisms. I hasten to add that the regulation of
499
26. PUMPS AND CURRENTS
cytosolic p H is not at all well understood and seems to call for rather more intervention by an invisible hand than is altogether comfortable. To illustrate both the problem and the line of inquiry that may lead to its solution, consider E. coli, which maintains the cytoplasmic pH near 7.8 over an external pH range from 5 to 9 (Padan et al., 1976; Zilberstein et al., 1979; Booth et al., 1979). This relatively constant internal pH is achieved in the face of the production of metabolic acids, several powerful proton pumps, and a host of processes that carry H + back into the cytosol (Fig. l), not to speak of the inevitable leakage of protons down the electrochemical gradient. It seems likely that movements of K + and Na+ play amajor role here, in the wider context of buffering that all-important energetic parameter, the proton motive force (Skulachev, 1978). Because the electrical capacitance of the membrane is small (about 1 pF/cm2),the movement of a few thousand protons across the membrane suffices to build up II, but will not affect the cytosolic pH. An increase in pH, whether for purposes of buffering Ap, + or of regulating the pH itself, may be achieved by an electrophoretic influx of K + . This is fine, but could easily be carried to excess, raising the pH beyond the capacity of cellular buffers to limit and of the cell to tolerate, so a second governor is required. This could be one function of a sodium/proton antiporter, allowing protons to flow inward in exchange for sodium (Fig. 3A). Sodium/proton antiporters have been shown to exist in many bacteria (Review: Lanyi, 1979) and, in one case at least, loss of the antiporter seems to entail difficulties in maintaining a constant cytoplasmic pH (Krulwich et al., 1979). However, it is likely that Fig. 3A, though correct in principle, is seriously oversimplified. Many Na+/ H + antiporters appear t o be electrogenic (Lanyi, 1979; Beck and Rosen, 1979); the K + transport system is not well understood but may possibly carry H + as well as K + (Bakker and Harold, 1980); and many bacteria grow happily in the absence of significant amounts of Na +. An alternative scheme, shown in Fig. A
Q:
K+
B
Na+ - t
H+
:Q+H
Na +
?
nH+
Fic. 3. Possible mechanisms for the control of cytoplasmic pH by movements of K + and Na+ . (A) Electrophoretic potassium uniport and electroneutral Na+ / H + antiport, as envisaged by Skulachev (1978). (B) Scheme showing the cation transport carriers currently thought t o exist in E. coli. The stoichiometry and molecular mechanism of K transport are not known. The proton pump is not specified.
500
FRANKLIN M. HAROLD
3B, is suggested by the recent discovery (Brey et a/., 1980) of an electroneutral K+ /H+ antiporter in E. coli; the activity of this porter is enhanced by alkaline pH, a requirement if it is to function as the scheme suggests. Still other mechanisms for controlling the cytoplasmic pH may well exist; a thoughtful study of the cytoplasmic pH in relation to ion movements is badly needed. The analogous problem in eukaryotic cells, also far from solution, has been considered by Smith and Raven (1979) who point out the probable importance of kinetic control over the primary proton pump by the cytosolic pH and other factors. The opportunity for direct electrophysiology is, so far at least, balanced by our ignorance of the mechanisms of Na+ and K + transport. Passing mention, at least, must be made of one additional function that depends on the accumulation of K + and extrusion of N a + : Both prokaryotic and eukaryotic ribosomes require a high concentration of K (near 0.1 M ) for proper translation of RNA. Many enzymes also appear to find K + a more compatible cation than Na+. Nevertheless, the fact that cells cease to grow when deprived of K + , even though their internal K + content is still very high, suggests that the critical roles of K + in cell physiology are related to the control of turgor and pH and perhaps to energy storage, rather than to macromolecule synthesis. T o conclude this section we may ask whether bacteria or eukaryotic cells monitor the membrane potential and endeavor to keep it constant. For bacteria the answer appears to be no: in E. coli, for instance, the value varies widely over a range of external conditions such as pH and K + content of the medium. It seems more likely that ApH+ is monitored, as mentioned in Section 11, although this is far from certain. As for eukaryotes, Slayman and his associates (Slayman, 1977; Gradmann et a/., 1978), in the course of their painstaking characterization of the proton-translocating ATPase of Neurospora, have found that the organism does possess homeostatic mechanisms which restore the potential after a metabolic downshift by altering both the activity of the pump and that of secondary porters. However, it appears that the teleonomic purpose of this is to regulate the rate of energy turnover (the ATPase of Neurospora consumes as much as a quarter of the ATP generated) rather than the potential per se (Slayman, 1977). Recently, Pall has found that reagents which depolarize the cytoplasmic membrane of Neurospora induce a rapid and dramatic increase in the level of cyclic adenosine 3'3'-monophosphate (CAMP) (Pall, 1977); the same is true of other fungi (Trevillyan and Pall, 1979). Whether this points to a role for CAMPin the regulation of eukaryotic energy metabolism or suggests, as Pall believes, that the organism monitors the membrane potential as an index of membrane integrity, remains to be determined. +
501
26. PUMPS AND CURRENTS
V.
CALCIUM CURRENTS AS BIOLOGICAL SIGNALS
Even the simplest living things respond to chemical and physical cues from the environment. Many of these impinge first upon the plasma membrane, and the mechanisms by which signals are received, transmitted, and transduced into a response often involve ion gradients and currents. To be sure, this need not always be so: Enzyme induction and repression are exerted at the level of the genome, and many chemical stimuli may be mediated by second messengers (including the adenylate cyclase-CAMP system) without the intervention of ion currents. Yet there can be no doubt that early in evolution there arose devices in which the energy cost of information is paid for in the currency of an electrochemical ion potential. Whether this assertion holds for bacteria is not yet entirely clear. I am not aware of any example of a metabolic response to an exogenous stimulus mediated by an ion flux, but there is reason to invoke such mechanisms in behavioral responses. Motile bacteria have quite an elaborate behavioral repertoire expressed primarily through their reactions to light, oxygen, nutrients, or pH; for example, bacteria tend to swim up gradients of attractants but away from repellants. Many aspects of locomotor physiology have now been clarified (reviews: Berg, 1975; Koshland, 1977; McNab, 1978): We know that chemical stimuli are recognized by receptor proteins external to the plasma membrane; that gradients are monitored, not in space but in time-the receptor compares, as it were, occupancy now with that a little while ago; and that the information is somehow transmitted to the flagellar motor and determines the frequency with which the motor spontaneously reverses its sense of rotation (tumbling). What we do not know is in what form the signal is transmitted. Since the motor is driven by the proton current (Fig. l), it is plausible that fluctuations in 1,5 or AiiH+ may modulate the sense of rotation, and some evidence for such a mechanism has been obtained. Ordal (1977) has proposed that a current of calcium controls the rotor, in keeping with the fact that bacteria (like eukaryotes) expel Ca2 and maintain a large inward electrochemical potential ApCa2+.His proposal has not been confirmed and remains under the cloud of the Scots verdict, not proven. In contrast, there is abundant evidence that the flagella and cilia of eukaryotic cells are controlled by ionic signals fundamentally analogous to those seen in the excitable cells of higher organisms. The paradigm for such research is the ciliated protozoan Paramecium whose virtues include both sexual recombination and robust cells that retain normal behavior when impaled by microelectrodes. Thanks t o the work of Eckert. Kung, and their associates in particular, one can look forward to a day when “Paramecium may become the first eukaryotic organism for which essentially all phases of sensory+
502
FRANKLIN M. HAROLD
motor behavior are understood in biophysical and molecular detail” (Eckert and Brehm, 1979). For present purposes there is no need to go into the details of the voltagecontrolled channels for Ca2+and K + that underlie the ionic control of movement and behavior (Eckert, 1972,1977; Eckert and Brehm, 1979; Kung et d., 1975), nor am I qualified to do so. Suffice it to sketch the outlines (Fig. 4). The coupling ion that links cellular metabolism to the work of information processing is calcium. Paramecium maintains a very low level of free cytosolic calcium, below M , by continuously pumping it out and (presumably) by sequestering CaZ+in mitochondria or in vesicles. Since the external calcium concentration is normally on the order of 10-3M, calcium is subject to a steep electrochemical potential gradient consisting of the sum of the concentration difference ( - 90 mV for the divalent cation) and the membrane potential (- 20 to - 40 mV). The rate of calcium influx (the calcium current) is a function of the conductance of specific calcium channels, primarily localized in the ciliary membrane, and the conductance in turn is controlled by the electrical potential across the plasma membrane: Depolarization of the membrane increases the conductance, to an extent dependent on the initial depolarization. For example, in the classic avoidance reaction, collision of the anterior end of the organism with a solid object stimulates local stretch receptors and initiates a local depolarization of the membrane; this spreads over the entire surface and causes the calcium channels to open. Calcium rushes into the ciliary compartment, raising the local calcium concentration above
Stimulus
FIG.4 . The role of a calcium current in the control of locomotor activity in Paramecium. An influx of calcium in response to a stimulus leads to reversal of the direction of ciliary beating. (After Eckert, 1977, with permission.)
C a influx
I
A t
$,
I Ciliary elevate.‘ I
-
F o r w a r d locomotion
B a c k w a r d l o co m o ti o n
Ca - a c t i v a t e d K+channels oDen
1
26. PUMPS AND CURRENTS
503
M . Now, the direction of the power stroke during ciliary beating is controlled, in a manner that is not fully understood, by the intraciliary calcium concentration; the rise in [Ca2+Iicauses the beat to reverse, and the organism swims backward. Concurrently, the rise in [Ca2+Iiprogressively inactivates the open calcium channels, allowing calcium expulsion to regain the upper hand and the organism to resume forward swimming (Fig. 4). Restoration of the membrane potential depends upon outward diffusion of K + , again through specialized channels that respond to electrical stimuli. The “escape reaction,” elicited by gently patting the cell’s posterior, results from an accelerated K + efflux which hyperpolarizes the membrane and makes the cilia beat forward faster. The mechanoreceptors, both front and aft, are in themselves devices that convert a local stimulus or deformation into a local “receptor current .” In the anterior set, at least, the receptor current is carried by calcium. Thus in Paramecium the initial sensing of the stimulus, its amplification, and the cell’s response all depend upon a calcium current. This makes it all the more surprising that so little is presently known about the molecular pumps that generate the gradients of Ca2+and K + in the first place. Paramecium is by far the best understood example of locomotor control by a calcium current but is by no means unique. Other ciliates behave similarly, and flagellated organisms likewise appear to favor calcium (Eckert, 1972; Schmidt and Eckert, 1976; Hyams and Borisy, 1978; Litvin et af., 1978). Nichols and Rickmenspoel (1978) have demonstrated a role for Mg2+ by microinjection experiments, but this seems to be a regulatory rather than a coupling function. There is strong evidence that ameboid motion is controlled by a localized influx of calcium across the plasma membrane; an introduction to this complex subject will be found in the recent paper of Nuccitelli et af. (1977) and in a review by Hitchcock (1977). Some of the best data come from work with the acellular slime mold Physarum, in which fluctuations of the cytosolic calcium level have been clearly implicated in the control of motility and cytoplasmic streaming (Ridgeway and Durham, 1976). A powerful calcium-stimulated ATPase, which is associated with a vesicular fraction from these organisms, is presumably part of the mechanism that maintains low cytosolic calcium (Kato and Tonomura, 1977). Finally, passing mention must be made of the role of calcium in the contraction of the stalk of vorticellid ciliates (Routledge and Amos, 1977): In this instance, cytosolic calcium not only triggers contraction of the spasmoneme but also powers it, by binding to the protein spasmin. Calcium, if not universal, is evidently the favored coupling ion for locomotor control in lower eukaryotes, circulating either across the plasma membrane or between cytosol and intracellular storage sites. Localized calcium currents serve not just in the control of motility but also
504
FRANKLIN M. HAROLD
as biological signals for a variety of physiological purposes. It is well known that, in animal cells, coupling of a stimulus to secretion by exocytosis is often mediated by calcium; the release of neurotransmitters at a presynaptic terminal is a familiar example. This, also, has microbial antecedents, and once again the best studied example comes from Paramecium. Many strains of this ciliate bear, just beneath the plasma membrane, structures called trichocysts. These are protein filaments, enclosed within membrane vesicles, that are discharged explosively when the organism comes in contact with certain kinds of prey and may assist in capturing the prey by entangling it. Be this as it may, trichocyst release depends upon local fusion of the trichocyst and plasma membranes and is a fine example of exocytosis. There have long been indications that this is triggered by a local influx of calcium across the plasma membrane (Plattner et al., 1977), presumably in response to the electrochemical potential gradient. Very recently, Satir and Oberg (1978) have neatly identified the calcium gate by the use of temperature-sensitive mutants: At the restrictive temperature these organisms lack the rosette of intramembrane particles normally present at the site of contact between trichocyst and membrane and also fail to discharge. The blockage can be overcome by addition of the ionophore A23187 which carries Ca2+across the membrane. Just how calcium promotes membrane fusion is still uncertain. Many microbial responses to environmental signals depend on oriented growth rather than movement. Researchers studying fungi and lower plants must wrestle with the mechanisms by which such organisms grow at the tip. It is well known that cell wall precursor vesicles, which may include chitosomes, accumulate at the tip and that new cell wall elements are preferentially inserted there (Stewart and Rogers, 1978). Repulsion by neighboring individuals and chemotropism toward nutrients modulate growth. Should we seek a role for calcium currents in these phenomena? In some cases, at least, the answer is clearly yes. By far the most compelling evidence comes from the work of Jaffe, Nuccitelli, Robinson et al. on the localization of growth in developing embryos of the brown algae Fucus and Pelvetia, which will be discussed in detail in Section VI. Unfortunately, little is yet known about cytosolic calcium and calcium fluxes in fungi, and the role of calcium should certainly not be taken for granted; here is a large and open field of research in cellular physiology. The biological functions performed by calcium currents are manifestly different from those of the H and Na+ currents discussed in earlier sections. Do they even belong to the same universe of discourse? At the risk of belaboring the obvious let it be said that the gradients of Ca2+and K + do not supply energy for ciliary beating; ATP does, through a sliding-filament mechanism. But the interplay of K + and Ca2+ currents makes possible a set of rapid, vectorial, and indeed intelligent responses to the surroundings that constitute a +
26. PUMPS AND CURRENTS
505
special kind of work. Metabolic energy pays for this by establishing the gradients of Caz+ and K + , by providing specific proteins for translocating calcium and binding it internally, and by localizing all the components where they will best serve. Just how a calcium current promotes vesicle fusion in exocytosis or tip growth is not known. Jaffe has emphasized the generation of local fields that can move particles by electrophoresis (Jaffe et al., 1974; Jaffe and Nuccitelli, 1977): straightforward work in the usual sense. But it is equally likely that the effects are mediated by calcium-binding proteins, as is the case with many enzymes whose activity is controlled by calcium, and that the work done is a case of providing information: instructions to go now, to grow here. In either case the mechanism depends critically on the fact that calcium, because of its propensity to bind to cytoplasmic constituents, has a low cytoplasmic mobility and is thus well fitted t o serve as the vehicle for local currents performing local work.
VI.
TRANSCELLULAR CURRENTS AND MORPHOGENESIS
This section will be devoted to what is, in my view, the most significant development in ion transport physiology. During the past decade it has been established that many (if not all) eukaryotic cells and organisms drive electrical currents, not just across the plasma membrane but clear across the cell. Such transcellular currents are related both to morphology and to developmental processes and are strongly implicated in the localization of growth in space. The notion that endogenous bioelectric fields are among the ultimate sources of morphogenetic information is far from new. Between 1921 and 1947, Elmer Lund carried out extensive research on the electrical potentials generated by a variety of plants and animals, and on the effects of external fields on growth and development. He was fully alive to the implications of his work: The existence of an electric polarity of the structurally polar cell, and the resulting bioelectric field, brings into relief the fact that there must exist in such polar cells an oriented electrochemical reaction system as a part of the . . . polar fine structure of certain components of the cell protoplasm. [The electrical pattern] is intimately related to the morphogenetic processes and polar or vector properties of cell and tissue functions. One function of this electrical field . . . is to act as a directive force in laying down of new structures (i.e., growth) and possibly the orderly transfer of various materials in morphogenesis. . . (Lund, 1947).
The reformulation of these concepts in modern idiom is chiefly the work of Lionel Jaffe and his associates and stems historically from the case of the Fucus egg (Jaffe et af., 1974; Jaffe and Nuccitelli, 1977; Jaffe, 1979).
506
FRANKLIN M. HAROLD
During the first day after fertilization the developing egg of the marine brown algae Fucus and Pelvetia elongates and then divides into two unequal halves, one destined to become the rhizoid (or holdfast) and the other the thallus (or frond). At this stage the rhizoid is the growing part, as theembryo’s first objective is to secure firm lodging. What makes this system so valuable is that, unlike animal eggs, the zygotes of Fucus and related algae have no predetermined axis or polarity. The site and direction of outgrowth are determined by one of a variety of external stimuli: light, electric fields, and gradients of pH and of certain ions (for example, the embryo elongates away from the direction of incident light). The various stimuli are thought to produce a localized membrane change that confers polarity upon the embryo; once outgrowth has begun, its direction is self-maintained. Studies on the mechanisms which determine the site and direction of growth were guided by the hypothesis that growth is localized by an ion current (Jaffe et al., 1974). It was proposed that the plasma membrane at the growing point is specifically permeable to a cation present externally at a higher electrochemical potential than in the cytoplasm (Ca2+,H + , or Na+); the potential gradient is generated by pumps located elsewhere in the cell. Passage of the cation inward completes a current loop across the cell, which depends on the segregation of pumps and leaks in separate locations; this current may then act as the “directive force” in morphogenesis. That developing Fucus zygotes d o generate endogenous transcellular currents was first demonstrated in 1966 by an ingenious method that, in effect, put about 100 eggs in series. The exploration of the current pattern around a single egg (diameter about 100 pm) had to await the development of the ultrasensitive vibrating probe (Nuccitelli and Jaffe, 1974), an instrument capable of detecting potential difference of 0.1 pV and less in the external medium. It is now known that, shortly after fertilization, the egg begins to drive an electrical current through itself such that positive charges enter the presumptive rhizoid and leave from elsewhere on the egg’s surface (Fig. 5A). Current begins to flow as early as 30 minutes after fertilization, preceding any visible signs of polarization such as local cortical clearing. The pattern is unstable at first but becomes progressively better defined as polarization proceeds; the site of stable current entry can be used to predict the site of outgrowth. Certain experimental manipulations that alter the current pattern, such as shining a second beam of light onto the egg, also alter the locus of outgrowth in a corresponding manner (Nuccitelli and Jaffe, 1974; Nuccitelli, 1978). We can therefore assert that the current precedes and predicts the position of growth and is likely to be causally involved in localizing it. The pattern of current flow persists for many hours while the embryo divides and begins to grow (Fig. 5B), but it changes progressively in intensity and in chemical composition. Two current components must be distin-
507
26. PUMPS AND CURRENTS
A
B
FIG.5 . The current pattern of the Pelvetiu egg. (A) A fertilized egg just before germination; outgrowth will occur at the top. (From Nuccitelli, 1978, with permission.) (B) Pattern of pulse current in a two-celled embryo. The arrow indicates the flow of positive charge. (From Nuccitelli and Jaffe, 1976a, with permission.)
guished, a steady current on the order of 1 pA/cm2 and periodic current pulses that are very much larger. The pulse current is carried largely by the efflux of chloride from the tip; it is apparently involved in regulating turgor pressure (Section IV) and need not detain us here. What matters for the polarization of growth is probably the steady current, particularly that portion which reflects the entry of calcium ions into the growing tip (Robinson and Jaffe, 1975): Evidence is mounting that it is this localized influx of calcium that determines the position of outgrowth. Direct evidence was recently obtained by Robinson and Cone (1979) who grew embryos in a gradient of the calcium ionophore A23 187 and found that the rhizoids formed predominantly on the side corresponding to maximal calcium influx. The polarization of outgrowth by external electrical fields (Peng and Jaffe, 1976)can also be rationalized in terms of this hypothesis. Granting that the transcellular current localizes growth, we must still determine how the symmetrical zygote generates a polarized current in the first place. It seems very likely that this reflects progressive amplification of a small asymmetry in calcium flow by a kind of positive feedback. One mechanism, outlined in some detail by Nuccitelli (1978), supposes that the calcium channels are mobile in the lipid phase of the plasma membrane and are somehow attracted to the patch of membrane that carries inward current. The initial stimulus, e.g., unilateral light, would effect some redistribution of the channels; continued current flow would progressively collect all the available channels and thus localize growth. Interactions that involve the cytoskeleton should be kept in mind here, as there is evidence indicating that the direction of growth is eventually fixed by the cell’s structure (Quatrano, 1978). Growth in Fucus, as in other algae, plants, and fungi, depends on cell wall precursor vesicles such as chitosomes, which accumulate at the growing tip and fuse locally with the plasma membrane. How might an electrical current
508
FRANKLIN M. HAROLD
into this region, particularly a calcium current, serve as a directive force to localize growth? From the beginning, Jaffe has favored the hypothesis that calcium entry generates an electric field across the cytoplasm, with the entry point positive. This would constitute a force that could direct cell constitutents, either by the electrophoresis of charged vesicles within the cytoplasm or of particles within the fluid phase of the membrane (Jaffe, 1977). Estimates of cytoplasmic conductivity would lead one to expect a field of about 1 mV across an egg, not a large field but demonstrably sufficient to effect redistribution of membrane particles when acting over a period of hours (Po0 and Robinson, 1977). However, the evidence for the existence of endogenous electric fields across the cytoplasm is, as yet, very meager (Jaffe and Nuccitelli, 1977; Jaffe, 1979); there is none for Fucuseggs or for any comparable system. A cytoplasmic field would perform work in a very different fashion, but there is no reason to exclude other mechanisms from consideration (Fig. 6 ) . Local calcium entry is presumably involved in the fusion of precursor vesicles with the membrane (Section V); it may well control actin-based contractile systems and thus direct precursor vesicles to the growing point by a mechanism related to that of transport in nerve axons; or it could influence cytoplasmic streaming. Some developmental processes, at least, are mediated by CAMP,and one may well envisage localized activation of a protein kinase by the rise in the local Ca2+level. At this stage of the investigation speculation is not much hampered by facts, and an open mind should be helpful. In particular, while it is clear that only Ca2+ and H + can generate significant cytoplasmic fields (Jaffe et al., 1974), it seems premature to dismiss the possibility that a current of a mobile ion, such as K ,may supply a “directive force” by a mechanism yet to be conceived. I have discussed Fucus and Pelvetia in some detail because of the wealth of +
CaZr /
FIG. 6 . How a calcium current may polarize a cell. The diagram indicates generation of a cytoplasmic field, electrophoresis of cytoplasmic and membrane components, polarization of cytoskeletal elements, and localized exocytosis. For discussion and caveats see text.
26. PUMPS AND CURRENTS
509
information that has been garnered. Data for other organisms are much more scattered (for review see Jaffe and Nuccitelli, 1977; Jaffe, 1979) but sufficient to support the generalization that most, if not all, eukaryotic cells and organisms drive transcellular currents through themselves and that these currents are involved in localizing growth and development in space. As a rule, the entry of positive charge, energetically downhill, occurs into the region of maximal growth. Examples from the microbial world include the marine alga Acetabularia in which regeneration of the cap is heralded by the establishment of a polarized current pattern and can be experimentally manipulated by applied electric fields (Christ-Adler and Bentrup, 1976). This may be the place to recall the formation of definite acid and alkaline bands along the internodal cells of giant algae (Hope and Walker, 1975). These certainly testify to the segregation of ion pumps at separate locations; current flows between them (Walker and Smith, 1977), but its relationship to cell growth is not clear. Germinating pollen grains establish a marked transcellular current pattern: Potassium enters the germ tube in response to protons pumped out of the grain (Weisenseel and Jaffe, 1976). Here again, local influx of calcium appears to direct the elongation of the germ tube. A very recent study from this laboratory (Stump, et al., 1980) has shown that the water mold Blastocladiella emersonii drives an electric current on the order of 1 pA/cm* through itself: In vegetative cells a positive current (protons?) enters the rhizoid and leaves from all over the thallus; when the cells enter the sporulation pathway the current reverses, a positive charge (calcium?) entering the thallus. This organism is particularly interesting because growth of the rhizoid can be polarized by gradients of proton-conducting ionophores (Harold and Harold, 1980). Could it be that, in this case, a proton current localizes growth? Examples could be multiplied by citing work with sea urchin eggs, frog embryos, muscle cells, and the outgrowth of neurites (Jaffe, 1979), but this would take us beyond the limits of this review without bringing us much closer to the molecular mechanisms by which endogenous transcellular currents act to localize growth. At the other biological extreme, bacteria are so small that external current patterns cannot be mapped with the instruments presently available. There are, however, indirect data to suggest that prokaryotic cells may differ fundamentally from eukaryotic ones with respect to ion currents. As mentioned above, under special conditions the fermentative organism S. faecalis can grow in the presence of ionophores that short-circuit currents of H ,K ,and Na+ (Harold and Van Brunt, 1977), producing cells that are indistinguishable from normal ones. The same has recently been found to be true for E. coli (H. R. Kaback, personal communication). We are thus encouraged to speculate that bacteria may not need a n ion current to localize growth but may rely primarily on the self-assembly of macromolecules to con+
+
510
FRANKLIN M. HAROLD
struct their cells. Be this as it may, there is good reason to believe that the study of transcellular ion currents will ultimately shed a flood of light on the mysterious processes by which cells and organisms assemble themselves in space.
VII.
A SENSE OF DIRECTION
A time-honored view of biological transport still quite widely held regards it as consisting of the processes required to supply the cell with the requisite nutrients, metabolites, or ions. The foregoing survey of the functions of ion transport in the life of unicellular organisms should show how limited this view is and how little it does justice to biological reality. To be sure, there do exist transport processes which serve primarily to augment the concentration of a particular ion; and there are important cellular functions that depend directly on such changes in concentration, including the regulation of cell volume and pH (Section IV). But it should be clear that the teleonomic significance of ion transport is far wider: The generation of ion currents across biological membranes is one of the fundamental mechanisms in cellular energetics. Ion currents are a versatile currency that allows cells to interconvert different kinds of energy and to support a wide variety of work functions, from the active transport of metabolites to adaptive behavior and the localization of growth in space. Even biological clocks, and with them the temporal organization of cells, may eventually prove to have an ionic basis. The recurrent theme is that of the coupling ion. An ion is pumped “uphill” across the membrane, at the expense of some source of energy, to generate a gradient of electrochemical potential; and the return of this ion (or even of another) down the gradient is drawn upon to perform work. Cellular bioenergetics rings the changes upon this theme, varying the species of ion used, the initial driving force, and the nature of the work accomplished. Coupling ions usually carry a positive charge, but anions may serve in special cases, as exemplified by the role of phosphate in linking together several transport systems in mitochondria. In principle even uncharged solutes could perform this role, but ions have the advantage that both electrical and concentration gradients contribute to the driving force. Besides, an electrical imbalance is rapidly transmitted over relatively large distances along the surface of a membrane and is thus particularly well suited both to energy transduction and to communication. Within limits it is useful to regard a cell as a special kind of electrical machine, operating on ion currents as befits a machine whose conducting phase is water. Before carrying the argument further, let us look a little more closely at two words that were freely employed in this article, “energy” and “work.” In
26. PUMPS AND CURRENTS
511
biochemistry the emphasis is traditionally placed on the Gibbs free energy function, which allows one to predict the direction and extent of chemical reactions: A process with a negative change in free energy (exergonic) can proceed spontaneously; one in which the free energy change is positive (endergonic) cannot, but may be coupled to (i.e., driven by) a second, exergonic one, and so on. The work done by ATP, for example, in the many biosynthetic reactions in which it participates, can be described as a shifting of the equilibrium position of the driven reaction, the energy source being the group transfer potential of the phosphoryl group. It is, on the surface at least, a scalar conception, appropriate to the description of chemical reactions in solution. The meaning of “work” in physics is quite different: We speak of work being done when an object on which a force acts moves through space and define it as the product of force and displacement. Similarly, work must be done to move a charged particle closer to a point charge of the same sign. Thus, the conception of work in physics has explicit vectorial connotations. The pertinence of these rather abstract reflections is made clearer by just looking at a cell under the microscope. It is true, but not very illuminating, to say that almost all cellular functions-movement, transport, biosynthesis, growth, morphogenesis-require an input of energy which is channeled through the pathways of metabolism. The point is that much of the work a cell performs is not merely done against an abstract gradient of free energy but is concretely specified in space and visibly vectorial. Transport of metabolites into cells or out is an elementary example of a biological function that has a direction in space. Growth and morphogenesis supply many others, more complicated but no less obvious. The point was well made years ago by Mitchell (1962b) in a little-known paper: The processes of growth and morphogenesis represent the movements of chemical groups . . . at certain rates and in certain definite directions in space, from nutrients initially present in the growth medium to polymers or metabolic intermediates (including inorganic ions and water) within the organism. Thus, growth and morphogenesis represent transport processes which, in common with the more popular membrane transport, must be described by vector quantities, having both magnitude and direction in space.
How can biochemical reactions, including those we identify as transport processes, have a direction in space? The riddle is formally expressed in Curie’s theorem which states that no process can be more asymmetric than its causes; it bedevilled bioenergetics for years, for it was held that enzymic reactions are scalar and therefore could not give rise to vectorial consequences. The paradox vanished with the recognition (Mitchell and Moyle, 1958;Mitchell, 1962a, 1979) that at the molecular level of resolution proteins are highly asymmetric structures which confer upon enzymic catalysis an inherent direction in space. The vectorial nature of metabolism is masked in solution but becomes apparent when enzymes are organized into anisotropic structures. It
512
FRANKLIN
M. HAROLD
is thus that an enzyme complex “plugged through” a membrane can generate a current of ions that has a direction perpendicular to the plane of the membrane. We have seen, in the preceding discussion, many cellular functions whose direction in space is at least partly specified by that of an ion current. The biological importance of this mechanism is most forcefully urged by the discovery that cellular growth and morphogenesis may be localized by transcellular ion currents. But the intrinsic asymmetry of proteins is equally fundamental to the assembly of vectorial structures in the cytoplasm, including the microtubules and microfilaments of the cytoskeleton; surely these must play a complementary but no less important role in specifying the position and direction of growth. Indeed, vectorial ligand conduction (Mitchell, 1979) should confer a degree of orientation upon the activities of many “soluble” enzymes loosely attached to the cell’s structural frame. Ultimately, the spatial organization of cells can be seen as an expression of the inherent asymmetry of the macromolecules of which they are made. Philosophical ramblings, unbecoming an experimental scientist? I do not think so, for it seems to me that these reflections touch the sense of uneasiness that many thoughtful persons feel toward the molecular view of life. The late Loren Eiseley, naturalist and poet, musing on the shapes of life in an autumn field, perceived the difficulty clearly: I have come to suspect that this long descent down the ladder of life, beautiful and instructive though it may be, will not lead us t o the final secret. In fact I have ceased to believe in the final brew or the ultimate chemical. I would not be understood to speak ill of scientific effort. . . . It is only that somewhere among these seeds and beetle shells and abandoned grasshopper legs I find something that is not accounted for very clearly in the dissections to the ultimate virus or crystal or protein particle (Eiseley, 1957).
It is not some mystical life force that Eiseley challenges us to produce, but a straightforward explanation for the spatial patterns of structure and function that every cell and organism displays, for all to see who would look. And his judgment, “not fully accounted for,” is still a fair one, even at the level of the simplest living things. The trouble is, of course, that they are not things at all but processes, structures that take their form from the flow of matter and energy, like a flame or an eddy. If we are to satisfy ourselves that we understand a whole cell, over and above the activities of its innumerable molecular parts, we must come to terms with this dimension of orderly, directional flow. To find the secret of life we shall have t o marry the study of macromolecules and genes with that of the vectorial flux of ions, molecules, and chemical groups in fields of force. ACKNOWLEDGMENTS 1 am indebted to the following sources for permission to reproduce material first published elsewhere: to Annual Reviews, Inc., for Figs. 1 and 2A; to Roger Eckert and the publishers of
513
26. PUMPS AND CURRENTS
Nature for Fig. 4; to Richard Nuccitelli and Academic Press for Fig. 5; to Random House Publishing Company for the quotation from “The Immense Journey” by Loren Eiseley; to the University of Texas Press for the quotation from Elmer Lund’s “Bioelectric Fields and Growth;” and to Peter Mitchell and Cambridge University Press for some lines from an article that first appeared in the Journalof GeneralMicrobiology. Last but not least, thanks aredue to my wife and to many friends and colleagues who were generous with preprints, advice, and constructive criticism. Original work from the author’s laboratory was supported by grants from the Institute of Allergy and Infectious Diseases, U.S. Public Health Service, and from the National Science Foundation.
REFERENCES Atkinson, D. E. (1977). ‘‘Cellular Energy Metabolism and Its Regulation.” Academic Press, New York. Bakker, E. P., and Harold, F. M. (1980). J . Biol. Chem. 255,433-440. Beck, J. C., and Rosen, B. P . (1979). Arch. Biochem. Biophys. 194,208-214. Berg, H. C. (1975). Sci. A m . 233, 36-44. Booth, I . R., Mitchell, W. J . , and Hamilton, W. A. (1979). Biochem. J. 182, 687-696. Bowman, B. J . , and Slayman, C. W. (1979). J . Biol. Chem. 254,2928-2934. Boyer, P. D., Chance, B., Ernster, L., Mitchell, P., Racker, E., and Slater, E. C. (1977). Annu. Rev. Biochem. 46,955-1026. Brey, R. N., Rosen, B. P., and Sorensen, E. N. (1980). J. B i d . Chem., 255,39-44. Christ-Adler, M., and Bentrup, F. W. (1976). Planta 129, 91-93. Crane, R. K. (1962). Proc. Fed. A m . SOC.Exp. Biol. 21, 891-895. Crofts, A. R., and Wood, P. M. (1978). Curr. Top. Bioenerg. 7, 175-244. DePierre, J. W., and Ernster, L. (1977). Annu. Rev. Biochem. 46,201-262. Dufour, J . P., and Goffeau, A. (1978). J . Biol. Chem. 253, 7026-7032. Dutton, L. P., and Prince, R. C. (1978). In “The Bacteria” (L. N. Ornston and J. R. Sokatch, eds.), Vol. V1, pp. 523-584. Academic Press, New York. Eckert, R., (1972). Science 176, 473-481. Eckert, R. (1977). Nature (London) 268, 104-105. Eckert, R., and Brehm, P . (1979). Annu. Rev. Biophys. Bioeng. 8, 353-383. Eddy, A. A. (1978). Curr. Top. Membr. Transp. 10,279-360. Eiseley, L. (1957). “The Immense Journey.” Random House, New York. Epstein, W., and Laimins, L. (1980). Trends Biochem. Sci. 5, 21-23. Epstein, W., Whitelaw, V., and Hesse, J. (1978). J . Biol. Chem. 253, 6666-6668. Felle, H . , Porter, J. S., Slayman, C. L., and Kaback, H. R. (1980). Biochemistry 19,3585-3590. Gradmann, D. (1976). J . Membr. Biol. 29,23-45. Gradmann, D., Hansen, U. P., Long, W. S., Slayman, C. L . , and Warncke, J . (1978). J . Membr. Biol. 39, 333-367. Gutknecht, J., and Bisson, M. A. (1977). In “Water Relations in Membrane Transport in Animals and Plants” (A. M. Jungreis, T. Hodges, A. M. Kleinzeller, and S. G. Schultz, eds.), pp. 3-14. Academic Press, New York. Gutknecht, J., Hastings, D. F., and Bisson, M. A. (1977). In “Transport Across Biological Membranes” (G. Giebich, D. C. Tosteson, and H. H. Ussing, eds.), Vol. 3, pp. 125-174. Springer-Verlag, Berlin and New York. Haddock, B. A , , and Hamilton, W. A,, eds. (1977). Symp. SOC.Gen. Microbiol. No. 27, pp. 1-442. Cambridge Univ. Press, London and New York. Haddock, B. A., and Jones, C. W. (1977). Bacteriol. Rev. 41,47-99. Hamilton, W. A. (1975). Adv. Microbiol. Physiol. 12, 1-53.
514
FRANKLIN M. HAROLD
Harold, F. M. (1972).Bacteriol. Rev. 36,172-230. Harold, F. M. (1977a).Annu. Rev. Microbiol. 31,181-203. Harold, F. M. (1977b).Curr. Top. Bioenerg. 6,84-149. Harold, F. M. (1978).In “The Bacteria” (L. N. Ornston and J. R. Sokatch, eds.), Vol. VI, pp. 463-521.Academic Press, New York. Harold, R. L., and Harold. F. M., (1980).J . Bacteriol. 144,1159-1167. Harold, F. M., and Van Brunt, J. (1977). Science 197,372-373. Hastings, D. F., and Gutknecht, J. (1976). J . Membr. Biol. 28,263-275. Hinkle, P . C., and McCarty, R. E. (1978).Sci. A m . 238, 104-123. Hitchcock, S. E. (1977). J . Cell. Biol. 74,1-15. Hope, B. A., and Walker, N. A. (1975). “The Physiology of Giant Algal Cells.” Cambridge Univ. Press, London and New York. Hyams, J . S., and Borisy, G . 0. (1978).J . Cell. Sci. 33,235-253. Jaffe, L. F. (1977).Nature (London) 265,600-602. Jaffe, L. F. (1979).In “MembraneTransduction Mechanisms” (R. A. Coneand J . E. Dowling, eds.), pp. 199-231.Society of General Physiologists Ser. Vol. 33,Raven, New York. Jaffe, L. F., and Nuccitelli, R. (1974).J . Cell. Biol. 63,614-628. Jaffe, L. F., and Nuccitelli, R. (1977). Annu. Rev. Biophys. Bioeng. 6,445-476. Jaffe, L. F . , Robinson, K . R., and Nuccitelli, R. (1974). Ann. N . Y . Acad. Sci. 238,372-389. Kato, T., and Tonomura, Y. (1977). J . Biochem. (Tokyo) 81,207-213. Komor, E., and Tanner, W. (1976).Eur. J. Biochem. 70,197-204. Koshland, D. E., J r . (1977).Science 196,1055-1063. Krulwich, T. A,, Mandel, K . G., Bornstein, R. F., and Guffanti, A. A. (1979). Biochem. Biophys. Res. Commun. 91,58-62. Kung, C., Chang, S.-Y., Satow, Y., Van Houten, J . , and Hansma, H. (1975). Science 188,
898-904. Laimins, L., Rhoads, D. B., Altendorf, K., and Epstein, W. (1978). Proc. Nut/. Acud. Sci. U.S.A. 75,3216-3219.
Lanyi, J . K. (1978).Microbiol. Rev. 42,682-706. Lanyi, J. K. (1979).Biochim. Biophys. Acta 559,377-398. Lauger, P. (1977). Nature (London) 268,360-362. Lipmann, F. (1941). A d v . Enzymol. 1,99-162. Litvin, F. F., Sineshchekov, V.,andSineshchekov, V. A. (1978).Nature(London)271,476-478. Lund, E. J. (1947). “Bioelectric Fields and Growth.” Univ. of Texas Press, Austin. MacDonald, R. E., Greene, R. V., Clark, R. D., and Lindley, E. V. (1979). J . Biol. Chem., 254,
11,831-11,838. McNab, R. M. (1978). C R C R e v . Biochem. 5,291-341. Measures, J . C. (1975).Nature (London) 257,398-400. Mitchell, P . (1961).Nature (London) 191,144-148. Mitchell, P. (1962a).Biochem. SOC. Symp. 22,142-168. Mitchell, P. (1962b).J . Gen. Microbiol. 29,25-37. Mitchell, P. (1966). Biol Rev. Cambridge Philos. SOC. 41,445-502. Mitchell, P. (1967).Adv. Enzymol. 29,33-87. Mitchell, P. (1970).In “Membranesand Ion Transport” (E. E. Bittar, ed.), Vol. 1, pp. 192-256. Wiley, New York. Mitchell, P. (1976). Biochem. SOC. Trans. 4,399-430. Mitchell, P. (1979).Eur. J . Biochem. 95,1-20. Mitchell, P., and Moyle, J. (1958).Narure (London) 182,372-373. Mummert, H., and Gradmann, D. (1976).Biochim. Biophys. Acta 443,443-450. Nichols, K. M., and Rickmenspoel, R. (1978).J . Cell Sci. 29,233-247.
26. PUMPS AND CURRENTS
515
Njus, D., and Radda, G. K . (1977). Biochim. Biophys. Acta 463,219-244. Nuccitelli, R. (1978). Dev. Biol. 62, 13-33. Nuccitelli, R., and Jaffe, L. F. (1974). Proc. Natl. Acad. Sci. U.S.A. 71,4855-4859. Nuccitelli, R., and Jaffe, L. F. (1976a). Dev. Biol. 49, 518-531. Nuccitelli, R., and Jaffe, L. F. (1976b). Planta 131, 315-320. Nuccitelli, R., Poo, M. M., and Jaffe, L. F. (1977). J. Gen. Physiol. 69, 743-763. Ordal, 0. W. (1977). Nature (London) 270,66-67. Padan, E., Zilberstein, D., and Rottenberg, H. (1976). Eur. J. Biochem. 63, 533-541. Pall, M. L. (1977). J . Biol. Chem. 252,7146-7150. Peng, H. B., and Jaffe, L. F. (1976). Dev. Biol. 53,277-284. Plattner, H., Reichel, K., and Matt, H. (1977). Nature (London) 267, 702-704. Poo, M.-M., and Robinson, K. R. (1977). Nature (London) 265,602-605. Poole, R. J. (1978). Annu. Rev. Plant Physiol. 29,437-460. Quatrano, R. S. (1978). Annu. Rev. Plant Physiol. 29,487-510. Raven, J. A., and Smith, F. A. (1976). J . Theor. Biol. 57,301-312. Rhoads, D. B., and Epstein, W. (1977). J . Biol. Chem. 252, 1394-1401. Ridgway, E. B., and Durham, A. C. H. (1976). J . Cell. Biol. 69,223-226. Robinson, K. R., and Cone, R. (1979). Science 207,77-78. Robinson, K. R., and Jaffe, L. F. (1975). Science 187,70-72. Rosen, B. P., ed. (1978). “Bacterial Transport,” pp. 1-684. Dekker, New York. Rosen, B. P., and Kashket, E. R. (1978). In “Bacterial Transport” (B. P. Rosen, ed.), pp. 559-620. Dekker, New York. Rottenberg, H. (1979). Methods Enzymol. 55,547-569. Routledge, L. M., and Amos, L. B. (1977). In “Calcium-binding Proteins and Calcium Function” (R. H. Wasserman, R. A. Corradino, E. Carafoli, R. H. Kretsinger, D. H. MacLennand, and F. L. Siegel, eds.), pp. 439-453. Elsevier, Amsterdam. Rydstrom, J. (1977). Biochim. Biophys. Acta 463, 155-184. Satir, B. H., and Oberg, S. G. (1978). Science 199, 536-538. Scarborough, G. A. (1976). Proc. Nail. Acad. Sci. U.S.A. 73, 1485-1488. Scarborough, G. A. (1977). Arch. Biochem. Biophys. 180,384-393. Schmidt, J. A., and Eckert, R. (1976). Nature (London) 262,713-715. Shimmen, T., and Tazawa, M. (1977). J . Membr. Biol. 37, 167-192. Simoni, R. D., and Postma, P. W. (1975). Annu. Rev. Biochem. 44,523-554. Skulachev, V. P. (1978). FEBS Lett. 87, 171-179. Slayman, C. L. (1965a). J . Gen. Physiol. 49,69-92. Slayman, C. L. (1965b). J . Gen. Physiol. 49,93-116. Slayman, C. L. (1977). In “Water Relations in Membrane Transport in Plants and Animals’’ (A. M. Jungreis, T. Hodges, A. M. Kleinzeller, and S. G. Schultz, eds.). Academic Press, New York. Slayman, C. L., and Slayman, C. W. (1974). Proc. Nafl. Acad. Sci. U.S.A. 71, 1935-1939. Slayman, C. L., Long, W. S., and Lu, C. Y. -H. (1973). J. Membr. Biol. 14,305-308. Smith, F. A., and Raven, J. A. (1979). Annu. Rev. Plant Physiol. 30,289-311. Stewart, P. R., and Rogers, P. J. (1978). In “The Filamentous Fungi” (J. E. Smith and D. R. Berry, eds.), pp. 164-196. Wiley, New York. Stoeckenius, W., Lozier, R. H., and Bogomolni, R. A. (1978). Biochim. Biophys. Acta 505, 215-278. Stump, R. F., Robinson, K . R., Harold, R. L.,andHarold, F. M. (1980). Proc. Natl. Acad. Sci. U.S.A. 77,6673-6677. Trevillyan, J . R., and Pall, M. L. (1979). J . Bacteriol. 138, 397-403. Waggoner, A. S. (1979). Annu. Rev. Biophys. Bioeng. 8,47-68.
516
FRANKLIN M. HAROLD
Walker, N. A , , and Smith, F. A. (1977). J . Exp. Bot. 28, 1190-1206. Weisenseel, M. H., and Jaffe, L. F. (1976). Planla 133, 1-7. Wieczorek, L., and Altendorf, K. (1979). FEBS Lett. 98, 233-236. Williams, R. J . P. (1978). Biochirn. Biophys. Acta 505, 1-44. Willsky, G. R. (1979). J . Biol. Chem. 254, 3326-3332. Woese, C. R., and Fox. G. E. (1977). Proc. Natl. Acad. Sci. U.S.A. 14, 5088-5090. Zilberstein, D., Schuldiner, S., and Padan, E. (1979). Biochemistry 18,669-673. Zimmermann, U. (1978). Annu. Rev. Plant'Physiol. 29, 121-148. Zimmermann, U., and Beckers, F. (1978). Planta 138, 173-179.
Index a
aftereffects of Na+ loading on action potentials, 27-29 Acetabularia, electrogenic pumps in, 270-272 aftereffects of Na+ loading on resting ATPase membrane potentials, 26-27 potassium transport, 129-131 direct measurement of changes in Na+ reconstituted into liposomes pump current in voltage clamp electrogenic properties of H + -ATPase experiments, 30-32 measurement of ApH, 197-198 Cardiac cells, excitability, electrogenic pump measurement of A$, 198-199 and, 474-475 steady-state level of All, t and the Chara electrogenesis H + / ATP ratio, 200-201 vectorial H + -ATPase reaction, 197 dependence on net H + efflux and Mg-ATP, 53-55 molecular properties of H t -ATPase pHi, pH,, and [K+],, 55-62 H' channel and H + filter: chemical modulation by light, 62-63 structure of Fo, 210-211 H + pump and H + gate activity of method for controlling intracellular crystalline ATPase F,, 207-210 environment by internal net ATPase synthesis driven by Ai, + perfusion, 51-53 Chloroplasts electric fields applied to H + -ATPase ion fluxes in, electrogenic aspects of, liposomes, 202-207 346-347 ion gradient applied to H t -ATPase ion movement in, 345-346 liposomes, 201-202 redox reactions, ion transport and structure, 347-351 B Chloroplast phosphorylation background information Bacteria electron transport, 217 coupling between H + entry and ATPase general, 216 synthesis structure of ATPase, 217-219 proton entry and ATP synthesis, conformational changes associated 178- 184 with energization, 241-243 rates of ATP formation and nature of coupling of proton transport to driving force, 187-191 ATP synthesis stoichiometry of coupling, 184-187 competition between basal and voltage-driven reversal, 176- 178 phosphorylating proton flux, 225-228 experimental evidence for ionic C coupling, 219-225 functional unit for ATP synthesis, 228 Canine cardiac Purkinje fibers kinetics of ATP synthesis hyperpolerization during enhanced artificial A$ and, 234-236 N+ -K+ exchange influence of artificial ApH on, 236-237 517
518
INDEX
Chloroplast phosphorylation (cont.) interpretation of threshold and lag times, 237-239 light-induced ApH and All., 232-233 simplified steady-state kinetics, 229-232 problem of energetic sufficiency, 239-241 Cytochrome oxidase, proton translocation by, 304-307 controversy over, 310-312 discovery of true proton pumping, 307-310 molecular principles and mechanisms of, 312-318
E Electrochemical gradient, effect on active H + transport in epithelium efficiency of energy conversion, 167-168 ion transport as pacemaker of cellular metabolism, 172-173 proton secretion by turtle bladder, 164-167 reversibility, 168- 17 1 stoichiometry, 171 -1 72 Electrogenesis dependence on net H + efflux and Mg-ATP H + efflux in relation to [ATPJ, 54-55 involvement of MgeATP in electrogenesis, 53-54 pH and [K'] analysis of results via linear equivalent-circuit model, 59-62 dependence of E, and Rn,on external K f concentration, 58-59 dependence of E, and R, on external pH in presence and absence of ATP, 56-58 dependence of En, and R,, on internal pH in presence and absence of internal ATP, 55-56 modulation by light, 62-63 Electrogenic alkali metal ion pump behavior of insect midgut K + transport system electrogenicity, 122 ionic dependence, 116 metabolic dependence, 116-1 19 special inhibitor, 119-122 time dependence, 114-1 16
membrane structure and location of transport functions morphology and fine structure, 123-125 pool location, 126-129 pump location, 125-126 methods experimental material, 1 1 1-1 12 insect midgut epithelial anatomy and transport nomenclature, 112-1 13 short-circuit technique, 113-1 16 potassium transport ATPase and, 129-131 Electrogenic proton pump, response of proton motive force to treatment in terms of thermodynamics of irreversible processes, 250-255 Electrogenic proton translocation, mitochondrial oxidative phosphorylation and electrophoretic metabolite transport, 424-426 respiration-dependent proton pumping, 407-414 reversible electrogenic proton translocation by F,-Fo ATPase, 414-420 role of proton translocation, 420-424 Electrogenic pump(s) extensions of model approximation of pumps by ideal sources, 274-275 influence of multiple charges and multiple charge transfer limbs, 274 relation of gradient-driven transport to active transport, 273 unstirred layers and asymmetric potential profiles, 274 in Nitella plasma membrane dependence of membrane potential on external and internal pH, 39-41 energy source for pump, 38-39 evidence for pump, 36-37 identity of pumped ions, 37-38 relationship between pump and membrane conductance, 42-44 results Acetabularia, 270-272 discussion: localization of energy shift, 272-273 Neurospora, 266-269 theory: reduction of kinetic models determination of model parameters from I-V curves, 264-265 five-state model, 258-260
INDEX
interpretation of parameters: comparison of two-state model with n-state models, 265 two-state model, 260-264 Electrogenic reactions and proton pumping in green plant photosynthesis comments on proton pathway to ATP synthetase, 458-459 electrogenic reaction steps chloroplast electrochromism as a molecular voltmeter, 438-442 electric generators, 442-449 survey, 437-438 membrane structure, 433-435 surface charge density and buffering capacity, 435-437 protolytic reaction steps proton pumps, 453-458 spectrophotometric detection of pH changes with dyes, 450-453 survey, 449-450 Electrogenic reactions of photochemical reaction center reaction center protein, 325-326 free energies of electron transport steps in cytochrome c2-reaction center complex, 333-335 light-induced electric current measurements, 326-331 notes on structure of reaction center, 331-333 ubiquinone-cytochrome b/c2 oxidoreductase associated H + transport, 339-340 electrogenic events within Q-b/c2 oxidoreductase, 339 hierarchy of redox components, 335-337 possible sources of electrogenic reaction, 340-342 redox coupling of reaction center and Q-b/c2 oxidoreductase, 337-339 Electrogenic sodium pump coupling between active Naf and K + transport active Na+ transport, 91-92 determination of coupling ratio by blocking K + channels, 103-106 K + flux, 92-97 use of filipin to determine coupling
519 between Na' and Kt transport, 97- 103 in mammalian tight epithelium basic transport properties of rabbit urinary bladder, 72-74 electrical measurements, 74-76 electrical structure of epithelium, 72 epithelial parameters, 76-79 pump properties, 79-85 Electrogenic sodium pump in controlling excitability in nerve and cardiac fibers excitability in cardiac cells effects of pump current on, 475-480 evidence for electrogenic pump in, 474-475 measurement of pump current, 480-481 excitability in nerve, 471-473 excitation process, 468-469 sodium pump and control of excitability, 469-47 1 Epithelium effect of electrochemical gradients on active H + transport in efficiency of energy conversion, 167-168 ion transport as pacemaker of cellular metabolism, 172- 173 proton secretion by turtle bladder, 164-167 reversibility, 168-171 stoichiometry, 171-172 electrical measurements, 74-76 electrical structure of, 72 membrane potentials, 77 membrane resistances, 76-77 membrane selectivity, 78-79 pump properties cell loading, 83-85 energy, 85 increasing Na' entry rate, 79-80 zero-gradient potentials, 80-83
F Filipin, use to determine coupling between Nat and K + transport, 97-103 Frog skeletal muscle fibers, hyperpolarization during enhanced Na+-K+ exchange K + depletion in T system, 21-25 pump hyperpolarization, 19-21
520
INDEX
Frog skin anatomy, 88 coupling between active Na+ and K f transport, 91-106 electrogenic Na+ pumps, 90-91 models, 88-90
G Gastric acid secretion electrogenicity of pump, 153-154 energy source for, 140-142 H + transport by gastric ATPase, 148- 150 K + transport by gastric ATPase active cation transport, 150 effect of external cations on H transport, 151-153 passive cation transport, 150-151 location of K + -dependent ATPase, 142-144 nature o f ATPase, 144-145 p H gradient and stoichiometry, 154-156 site of secretion, 136-140 steps in ATP hydrolysis breakdown of phosphoenzyme, 145-146 formation of phosphoenzyme, 145 steady state kinetic aspects, 146-148 structural aspects of ATPase, 156-157 Green plant photosynthesis, electrogenic reactions and proton pumping in comments on proton pathway to ATP synthetase, 458-459 electrogenic reaction steps, 437-449 membrane, 433-437 protolytic reaction steps, 448-458
I Insect midgut, electrogenic alkali metal ion Pump behavior of K t transport system, 114-122 membrane structure and location of transport functions, 123-129 methods, 111-114 potassium transport ATPase, 129-131 Ion transport, physics of free ion and ion carrier migration, 277-278 ion conductance, 278-281
M Membrane vesicles, electrochemical ion gradients and active transport carrier action chemical modification of transport activity, 401-402 mechanistic studies, 399-401 chemiosmotic phenomena proton-dependent transport, 397-398 proton electrochemical gradient and active transport, 395-397 sodium-dependent transport, 398-399 molecular architecture of Escherichiu coli membrane vesicles, 394-395 Mitochondrial oxidative phosphorylation electrophoretic metabolite transport, 424-426 respiration-dependent proton pumping data on respiratory chain proton pumping, 409-414 models for, 408-409 structure of respiratory chain, 407-408 reversible electrogenic proton translocation by F,-Fo ATPase complex electrogenic proton pumping by F,-Fo ATPase complex, 418-420 models for synthesis of ATP, 415-418 structure of F,-Fo ATPase complex, 414-415 role of proton translocation, 420 correlation between phosphorylation potential and A ;, + , 422-423 models for energy coupling, 421-422 synthesis of ATP by artificial proton gradients, 423-424 Mitochondrial transhydrogenase hypothesis on the mechanism of A ;, + generation, 384-387 known facts and forecasts. 387-391
N Nerve, excitability, electrogenic sodium pump and, 471-473 Neurospora, electrogenic pumps in, 266-269 Nitellu, electrogenic pump in plasma membrane dependence of membrane potential on external and internal p H , 39-41
INDEX
energy source for pump, 38-39 evidence for pump, 36-37 identity of pumped ions, 37-38 relationship between pump and membrane conductance, 42-44
P Phosphorylation in chloroplasts background information electron transport, 217 general, 216 structure of ATPase, 217-219 conformational changes associated with energization, 241-243 coupling of proton transport to ATP synthesis competition between basal and phosphorylating proton flux, 225-228 experimental evidence for ionic coupling, 219-225 functional unit for A T P synthesis, 228 kinetics of A T P synthesis artificial A$ and, 234-236 influence of artificial A p H on, 236-237 interpretation of threshold and lag times, 237-239 light-induced A p H and A$, 232-233 simplified steady-state kinetics, 229-232 problem of energetic sufficiency, 239-241 Photochemical charge separation and active transport in purple membrane mechanistic implications of steady-state kinetics, 377-379 primary photochemical event photoisomerization and charge separation, 374-377 relation of bacteriorhodopsin t o visual pigment, 372-374 relating kinetic and molecular models, 379-380 Proton-ATP synthetase comparative inhibitor analysis of solubilized and membrane-bound factor F, ADP derivatives as modifiers of a-subunit of factor F,, 294-297 ATP-MC: an inhibitor of catalytic site of factor F,, 292-294 butanol treatment: localization of factor F, catalytic site and, 297-299
lithium chloride: active site of H' ATPase and, 299-300 determination of equilibrium constant for ATP + H,O = ADP + P, at active site, 288-290 energy-dependent release of F,-bound AMPPNP from membrane of submitochondrial particles, 290-292 substrate translocation hypothesis, 285-287 Proton entry, coupled to A T P synthesis in bacteria proton entry and ATP synthesis, 178-184 rates of ATP formation and nature of driving force, 187-191 stoichiometry of coupling, 184-187 voltage-driven reversal, 176- 178 Proton-membrane interactions in chloroplast bioenergetics methods and rationale, 351-352 results and discussion correlation of uncoupler-enhanced acetic anhydride derivation with inhibition of water oxidation, 357-360 decrease in derivation of thylakoid membrane proteins by acetic anhydride caused by light, 352-357 membrane proteins showing differential acetic anhydride reactivity in light and dark, 360-363 Proton motive force, response to pulse of an electrogenic proton pump, treatment in in terms of thermodynamics of irreversible processes, 250-255 Proton translocation by cytochrome oxidase controversy over proton translocation, 310-312 discovery of true proton pumping by cytochrome oxidase, 307-310 molecular principles and mechanisms o f proton translocation general principles of a redox-linked proton translocator, 313-316 possible molecular mechanism, 316-318 relation between membrane Bohr effects and a proton pump, 312-313 properties of cytochrome oxidase, 305-307 respiration-linked H + translocation, 304-305
INDEX
Pumps and currents, a biological perspective calcium currents as biological signals, 501-505 cellular homeostasis, 495-500 ion currents and energy coupling in eukaryotic cells, 492-495 role of ion currents in metabolic economy of bacteria, 487-492 sense of direction, 510-512 transcellular currents and morphogenesis, 505-510 Purple membrane, photochemical charge separation and active transport in mechanistic implications of steady-state kinetics, 377-379 primary photochemical event, 372-377 relating kinetic and molecular models, 379-380
R Rabbit urinary bladder, basic transport properties of, 72-74
S Snail neuron, electrophysiology of sodium pump in
discussion, 14-15 methods, 5-6 results, 6-14 Sodium pump electrophysiology discussion coupling ratio of pump, 15 sources of error, 14-15 methods, 5-6 results comparison with pHi-regulating system, 13-14 effect of increasing Em o n p u q p current, 12-13 iontophoretic transport number for sodium injections, 9-12 measurement of current and charge generated by pump, 7-9 sodium injection and membrane potential, 6-7
T Transhydrogenase, mitochondria1 hypothesis on the mechanism of A ,,, generation, 384-387 known facts and forecasts, 387-391 Turtle bladder, proton secretion by, 164-167