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

Current Topics in Membranes and Transport VOLUME 36

Protein-Membrane Interactions

This Page Intentionally Left Blank

Current Topics in Membranes and Transport Edited by Joseph F. Hoffman

Gerhard Giebisch

Deparhnent of Cellular and Molecular Physiology Yale University School of Medicine New Haven, Connecticu

Deparmtenr of Cellular and Molecular Physwbgy

Yale University School of Medicine New Haven, Connecticui

VOLUME 36 Protein-Membrane Interactions Guest Editor Toni Claudio Department of Cellular and Molecular Physiology Yale University School of Medicine New Haven, Connecticut

Volume 36 is part of the series from the Yale Department of Cellular and Molecular Physiology

I#( San Diego

ACADEMIC PRESS, INC. Harcourt Brace Jovanovich, Publishers

New York Boston London Sydney Tokyo Toronto

This book is printed on acid-free paper.

@

COPYRIGHT 0 1990 BY ACADEMIC PRESS, INC. All Rights Reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher.

ACADEMIC PRESS, INC. San Diego, California 92101 United Kingdom Edition published by ACADEMIC PRESS LIMITED 24-28 Oval Road, London NW I 7DX

LIBRARY OF CONGRESS CATALOG CARD NUMBER:

ISBN

0-12-153336-0 (alk. paper)

PRINTED IN THE UNITED STATES OF AMERICA 9 0 9 1 9 2 9 3

9

8

7

6

5

4

3

2

1

70-1 17091

Contents

Contributors, ix Preface, xiii Yale Membrane Transport Processes Volumes, xv

PART 1.

PROTEIN INSERTION INTO AND ACROSS MEMBRANES

Chapter 1. Mitochondrial Protein Import: Specific Recognition and Membrane Insertion of Precursor Proteins NIKOLAUS PFANNER AND WALTER NEUPERT

I. 11. Ill. IV. V.

Introduction, 3 Receptor Proteins on the Mitochondria1 Surface, 5 The “General Insertion Protein,” 8 Role of Receptors and General Insertion Protein, 10 Summary and Perspectives, 11 References, I 1

Chapter 2. An Enzymological Approach to Membrane Assembly and Protein Secretion BILL WICKNER Text, 15

Chapter 3. Protein Translocation in Yeast JOANNE CROWE AND DAVID I. MEYER

I. Introduction, 19 11. Yeast as a Model System, 20 111. The Importance of Preprotein Conformation, 21 IV. Binding to the Membrane and Translocation Are Separable Events, 22 V. Analysis of pt l l - A Translocation Mutant, 24 References, 25

V

vi

CONTENTS

PART II.

STRUCTURE OF PROTEINS IN THE MEMBRANE

Chapter 4. Architectural Editing: Regulating the Surface Expression of the Multicomponent T-cell Antigen Receptor RICHARD D. KLAUSNER, JENNIFER LIPPINCOTT-SCHWARTZ, AND JUAN S. BONIFACINO I. 11. 111. IV. V. VI. VII.

Introduction, 31 Structure and Composition of the T-cell Antigen Receptor, 33 Subunit Interactions within the T-cell Receptor, 36 Stoichiometry within the Complex, 39 Assembly of the T-cell Receptor Complex, 40 Fate of Newly Synthesized T-cell Receptor Chains, 43 Summary, 47 References, 48

Chapter 5. The Photosynthetic Reaction Center from the Purple Bacterium Rhodopseudomonas viridis: Aspects of Membrane Protein Structure HARTMUT MICHEL AND JOHANN DEISENHOFER

I. Introduction, 53 11. Results and Discussion, 54 111. Conclusions, 68 References, 68

Chapter 6. Bacteriorhodopsin Folding in Membranes: A Two-Stage Process D. M. ENCELMAN, B. D. ADAIR, I. F. HUNT, T. W. KAHN, AND J.-L. POPOT I. Introduction, 71 11. Bacteriorhodopsin, 73 111. Bacteriorhodopsin Fragments Contain Stable Transbilayer Helices, 73

IV. V. VI. VII.

Links and Retinal Are Not Required for Folding, 74 Polar Interactions in Helix-Helix Associations, 75 Packing Effects, 76 Summary, 77 References. 77

PART 111.

PROTEIN MOBILITY IN MEMBRANES

Chapter 7. Molecular Associations and Membrane Domains MICHAEL EDIDIN I. Models of Membrane Organization, 81 11. The Study of Large-Scale Molecular Mobility in Cell Surface Membranes, 83

vi i

CONTENTS 111. A Basis for the Organization of Morphologically Polarized Cell Surfaces, 90 IV. Concluding Remarks, 93 References, 93

Chapter 8. Actin-Membrane Interactions in Eukaryotic Mammalian Cells THOMAS P. STOSSEL I. The Actin System and Membrane Function, 97 11. The Erythrocyte Cytoskeleton: Paradigm or Distraction in Approaching the Interaction between the Eukaryote Plasma Membrane and Actin System?, 98 111. The Actin System in Membrane Stabilization and Retraction, 99 IV. The Actin System in Membrane Propulsion, 101 V. The Actin System and Membrane Propulsion and Retraction, 104 References, 106

Chapter 9. Biogenesis and Cell Surface Distribution of Acetylcholine Receptors Stably Expressed in Fibroblasts TONI CLAUD10

I. Introduction, 109 11. Acetylcholine Receptor-Fibroblast Cell Lines, 110 111. Properties of Acetylcholine Receptors Expressed in Acetylcholine Receptor-Fibroblast Cells, 111 IV. Properties of Individually Expressed Subunits, 112 V. Posttranslational Modifications, 113 VI. Conclusions, 113 References. 114

Chapter 10. Control of Organelle Movements and Endoplasmic Reticulum Extension Powered by Kinesin and Cytoplasmic Dynein MICHAEL P. SHEETZ, SANDRA L. DABORA, ERIC STEUER, AND TRINA A. SCHROER I. Introduction, 117 11. Membranous Organelle Transport on Microtubules, 118 111. Interaction of Motors with Organelles to Produce Motility, 121 References, 126

PART IV. SIGNALING AND COMMUNICATION Chapter 11. G Protein-Coupled Receptors: Structure and :unction of Signal-Transducing Proteins ERIC M. PARKER AND ELLIOTT M. ROSS I. Introduction, 131 11. Mechanism of G-Protein Activation by Agonist-Liganded Receptors, 111. General Structure of G Protein-Coupled Receptors, 134

32

viii

CONTENTS

1V. Structure of the Ligand-Binding Domain, 136 V. Structure of the G Protein-Binding Domain, 138 VI. Receptor-Mimetic Peptides as Models for the G Protein-Binding Domain, 140 References, 141

Chapter 12. Mechano-Sensitive Ion Channels in Microbes and the Early Evolutionary Origin of Solvent Sensing CHING KUNG, YOSHIRO SAIMI, AND BORIS MARTINAC

I. Introduction, 145 11. A Stretch-Activated Ion Channel of Escherichia coli, 146 111. A Stretch-Activated Ion Channel in Yeast, 147 IV. Touch Receptors and Channels of Paramecium, 149 V. Mechano-Sensitive Channels and the Concept of Solvent Senses, 150 References. 152

Chapter 13. Selection of an afl T-cell Antigen Receptor in Vivo and Expression in Vitro in a Soluble Form M. M. DAVIS, B. FAZEKAS DE ST.GROTH, L. J. BERG, A. LIN, B. DEVAUX, C. SAGERSTROM, J. F. ELLIOTT, AND P. J. BJORKMAN

I. Introduction, 155 11. Positive Selection, 156 111. Negative Selection, 157 IV. Soluble T-cell Receptor Heterodimers, 158 References, 159

Chapter 14. Perforin and the Mechanism of Lymphocyte-Mediated Cytolysis ECKHARD R. PODACK AND MATHIAS G . LICHTENHELD I. Introduction, 161 11. Physicochemical and Functional Properties of Perforin, 162

111. Sequence of Murine and Human Perforin, 163 IV. Homology of Perforin to Complement Proteins: The Perforin Family, 167 V. Lack of Homologous Restriction of Perforin, 168 VI. Expression of Perforin mRNA in Vitro and in Vivo, 169 VII. The Contribution of Membrane Pores to DNA Degradation, 170 VIII. Conclusions, 171 References, 171 Note Added in Proof, 175

index, 177

Contributors Numbers in parentheses indicate the pages on which the authors’ contributions begin.

B. D. Adair, Department of Molecular Biophysics and Biochemistry, Yale University, New Haven, Connecticut 0651 1 (71)

L. J. Berg, Howard Hughes Medical Institute and Department of Microbiology and Immunology, Stanford University School of Medicine, Stanford, California 94305 (I 55)

P. J. Bjorkman, Department of Microbiology and Immunology, Stanford University School of Medicine, Stanford, California 94305 (155) Juan S. Bonifacino, Cell Biology and Metabolism Branch, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland 20892 (31)

Toni Claudio, Department of Cellular and Molecular Physiology, Yale University School of Medicine, New Haven, Connecticut 06510 (109)

Joanne Crowe, Department of Biological Chemistry and Molecular Biology Institute, School of Medicine, University of California, Los Angeles, Los Angeles , California 90024 ( 19) Sandra L. Dabora, Department of Cell Biology and Physiology, Washington University Medical School, St. Louis, Missouri 631 10 (117)

M. M. Davis, Howard Hughes Medical Institute and Department of Microbiology and Immunology, Stanford University School of Medicine, Stanford, California 94305 (155)

Johann Deisenhofer, Howard Hughes Medical Institute and Department of Biochemistry, University of Texas, Southwestern Medical Center, Dallas, Texas 75235 (53) B. Devaux, Howard Hughes Medical Institute and Department of Microbiology and Immunology, Stanford University School of Medicine, Stanford, California 94305 (155) ix

CONTRIBUTORS

X

Michael Edidin, Department of Biology, The Johns Hopkins University, Baltimore, Maryland 21218 (81)

J. F. Elliott, Department of Microbiology and Immunology, Stanford University School of Medicine, Stanford, California 94305 (155)

D. M. Engelman, Department of Molecular Biophysics and Biochemistry, Yale University, New Haven, Connecticut 0651 1 (71)

B. Fazekas de St. Groth, Howard Hughes Medical Institute and Department of Microbiology and Immunology, Stanford University School of Medicine, Stanford, California 94305 (155)

J. F. Hunt, Department of Molecular Biophysics and Biochemistry, Yale University, New Haven, Connecticut 0651 1 (71)

T. W. Kahn, Department of Molecular Biophysics and Biochemistry, Yale University, New Haven, Connecticut 0651 1 (7 1)

Richard D. Klausner, Cell Biology and Metabolism Branch, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland 20892 (31)

Ching Kung, Laboratory of Molecular Biology and Department of Genetics, University of Wisconsin-Madison, Madison, Wisconsin 53706 (145) Mathias G. Lichtenheld, Department of Microbiology and Immunology, University of Miami School of Medicine, Miami, Florida 33101 (161)

A. Lin, Howard Hughes Medical Institute and Department of Microbiology and Immunology, Stanford University School of Medicine, Stanford, California 94305 (155)

Jennifer Lippincott-Schwartz, Cell Biology and Metabolism Branch, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland 20892 (31)

Boris Martinac, Laboratory of Molecular Biology and Department of Genetics, University of Wisconsin-Madison, Madison, Wisconsin 53706 (145)

David I. Meyer, Department of Biological Chemistry and Molecular Biology Institute, School of Medicine, University of California, Los Angeles, Los Angeles, California 90024 (19)

Hartmut Michel, Max-Planck-Institut fur Biophysik, D-6000 Frankfurt/M 7 1, Federal Republic of Germany (53)

Walter Neupert, Institut fur Physiologische Chemie, Universitat Miinchen, D-8000 Munchen 2, Federal Republic of Germany (3)

CONTRIBUTORS

xi

Eric M. Parker, Department of Pharmacology, University of Texas Southwestern Medical Center, Dallas, Texas 75235 (131) Nikolaus Pfanner, Institut fur Physiologische Chemie, Universitat Miinchen, D-8000 Munchen 2, Federal Republic of Germany (3) Eckhard R. Podack, Department of Microbiology and Immunology, University of Miami School of Medicine, Miami, Florida 33101 (161) J.-L. Popot, Institut de Biologie Physico-Chimique, CollZge de France, 7523 1 Paris, France (71) Elliott M. Ross, Department of Pharmacology, University of Texas Southwestern Medical Center, Dallas, Texas 75235 (131) C . Sagerstrom, Department of Microbiology and Immunology, Stanford University School of Medicine, Stanford, California 94305 (155)

Yoshiro Saimi, Laboratory of Molecular Biology and Department of Genetics, University of Wisconsin-Madison, Madison, Wisconsin 53706 (145) Trina A. Schroer, Department of Cell Biology and Physiology, Washington University Medical School, St. Louis, Missouri 631 10 (1 17) Michael P. Sheetz, Department of Cell Biology and Physiology, Washington University Medical School, St. Louis, Missouri, 631 10 (1 17) Eric Steuer, Department of Cell Biology and Physiology, Washington University Medical School, St. Louis, Missouri 63110 (117) Thomas P. Stossel, Department of Medicine, Harvard Medical School, Massachusetts General Hospital, Boston, Massachusetts 02 114 (97) Bill Wickner, Molecular Biology Institute and Department of Biological Chemistry, University of California, Los Angeles, Los Angeles, California 90024 (15)

This Page Intentionally Left Blank

Membrane-associated proteins mediate the full range of communications between a cell and its environment. The aim of this book is to explore some of the unique properties of membrane proteins which allow them to be intermediaries in cell-cell interactions and to interact with their environment. The contributions have been organized into four major sections. Part I covers membrane protein synthesis, targeting, and translocation into and across membranes. Part I1 treats structures of proteins in membranes. Problems associated with heterooligomeric subunit assembly, the folding of membrane proteins within the lipid bilayer, and the first successful crystallization of a membrane protein are described. In Part 111 the ability of membrane proteins to move or be moved in the membrane as well as the control of organelle movements are discussed. Part IV describes the involvement of membrane proteins in trans-membrane signal transduction. The conference from which this book was derived was sponsored by the Department of Cellular and Molecular Physiology of the Yale University School of Medicine. Many distinguished scientists participated, among whom was Dr. Hartmut Michel who learned during the first coffee break that he had won the Nobel Prize in Chemistry. I would like to thank all the participants, especially those who contributed to this volume. I gratefully acknowledge the following organizations for their generous financial support: Biogen Research Corporation, Johnson & Johnson, Miles, Inc., G. D. Searle Research & Development, Smith Kline & French Laboratories, The Squibb Institute for Medical Research, and The Upjohn Company. TONICLAUDIO

xiii

This Page Intentionally Left Blank

Yale Membrane Transport Processes Volumes 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.). (198 1). “Molecular Mechanisms 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.). (1982). “Electrogenic Ion Pumps”: Volume 16 o f Current Topics in Membranes and Transport (A. Kleinzeller and F. Bronner, eds.). Academic Press, New York. Joseph F. Hoffman and Bliss Forbush 111 (eds.). (1983). “Structure, Mechanism, and Function of the Na/K Pump”: Volume 19 of Current Topics in Membranes and Transport (F. Bronner and A. Kleinzeller, eds.). Academic Press, New York . James B. Wade and Simon A. Lewis (eds.). (1984). “Molecular Approaches to Epithelial Transport”: Volume 20 of Current Topics in Membranes and Transport (A. Kleinzeller and F. Bronner, eds.). Academic Press, New York. Edward A. Adelberg and Carolyn W. Slayman (eds.). (1985). “Genes and Membranes: Transport Proteins and Receptors”: Volume 23 o f Current Topics in Membranes and Transport (F. Bronner and A. Kleinzeller, eds.). Academic Press, Orlando. Peter S. Aronson and Walter F. Boron (eds.). (1986). ‘“a+-H+ Exchange, Intracellular pH, and Cell Function”: Volume 26 of Current Topics in Membranes and Transport (A. Kleinzeller and F. Bronner, eds.). Academic Press, Orlando. Gerhard Giebisch (ed.). (1987). “Potassium Transport: Physiology and Pathophysiology”: Volume 28 of Current Topics in Membranes and Transport (F. Bronner and A. Kleinzeller, eds.). Academic Press, Orlando. William S. Agnew, Toni Claudio, and Frederick J. Sigworth (eds.). (1988). “Molecular Biology of Ionic Channels”: Volume 33 of Current Topics in Membranes and Transport (J. F. Hoffman and G. Giebisch, eds.). Academic Press, San Diego. xv

XVI

YALE MEMBRANE TRANSPORT PROCESSES VOLUMES

Stanley G. Schultz (ed.). (1989). “Cellular and Molecular Biology of Sodium Transport”: Volume 34 of Current Topics in Membranes and Transport (J. F. Hoffman and G. Giebisch, eds.). Academic Press, San Diego. Toni Claudio (ed.). (1990). “Protein-Membrane Interactions”: Volume 36 of Current Topics in Membranes and Transport (J. F. Hoffman and G. Giebisch, eds.). Academic Press, San Diego.

Part I

Protein Insertion into and across Membranes

This Page Intentionally Left Blank

CURRENT TOPICS IN MEMBRANES AND TRANSPORT, VOLUME 36

Chapter I

Mitochondrial Protein Import: Specific Recognition and Membrane Insertion of Precursor Proteins NIKOLAUS PFtWNER AND WALTER NEUPERT Institut fur Physiologische Chemie Universitiit Miinchen 0-8000Miinchen 2 , Federal Republic of Germany

I.

Introduction Receptor Proteins on the Mitochondria1 Surface A. Functional Characterization of Receptor Sites B. Identification of Receptors C. Import of Precursor Proteins Bypassing Receptor Sites 111. The “General Insertion Protein” IV. Role of Receptors and General Insertion Protein V. Summary and Perspectives References 11.

I. INTRODUCTION Eukaryotic cells are divided into numerous membrane-bounded compartments (“organelles”), each of which contains a unique and specific set of proteins. Most of the organellar proteins are synthesized as precursor proteins on cytosolic polysomes and thus have to be transported to their functional destination (Wickner and Lodish, 1985). The question of how proteins are directed to their specific target membrane and how they are translocated into and across organellar membranes poses a central theme of modem cell biology. The cytosolic precursor proteins carry specific targeting sequences that are assumed to bind to complementary structures (“receptors”) on the surface of organelles. The translocation

Copyright 0 1990 by Academic Press, Inc. All nghts of reproduction in any form reserved.

4

NIKOLAUS PFANNER AND WALTER NEUPERT

into and across the membranes may be mediated by proteinaceous components and/or lipids of the membranes. The biogenesis of organellar proteins is even more complex in case of mitochondria. Two membranes (outer and inner mitochondrial membranes) limit two soluble compartments, namely, the intermembrane space and the matrix. Whereas some mitochondrial proteins are coded for by mitochondrial genes and are synthesized in the matrix, >90% of the proteins are coded for by nuclear genes and are imported from the cytosol (Pfanner and Neupert, 1987a; Attardi and Schatz, 1988; Hart et al., 1989). It was first shown for mitochondrial protein import that translocation of precursor proteins across membranes is not mechanistically coupled to synthesis of the protein on ribosomes, both in vivu and in vitro (Hallermayer and Neupert, 1976; Harmey et al., 1976, 1977; Hallermayer et al., 1977). Many precursor proteins carry amino-terminal extension sequences ( “presequences”) of about 20-80 amino acid residues. Presequences contain signal information for translocation into mitochondria (Honvich et al., 1985; Hurt and van Loon, 1986). With some precursor proteins, targeting sequences were also found in nonamino-terminal (carboxyl-terminal) regions of the polypeptide (Pfanner et al., 1987b,c; Smagula and Douglas, 1988). The precursor proteins are recognized by specific receptors on the mitochondrial surface and are then inserted into the outer membrane (see later). Further translocation into the inner membrane occurs predominantly at sites of close contact between the mitochondrial outer and inner membranes (“contact sites”) (Schleyer and Neupert, 1985; Schwaiger et al., 1987; Rassow et a l . , 1989). Hydrophilic (proteinaceous) components apparently represent essential parts of contact sites (Pfanner et al., 1987a; Vestweber and Schatz, 1988). Presequences are proteolytically cleaved by the processing peptidase in the mitochondrial matrix (Hawlitschek et al., 1988), the proteins are sorted to their final intramitochondrial location (Hart1 et ul., 1987), and are often assembled into multisubunit complexes (Schmidt et al., 1983; Lewin and Norman, 1983). Protein import requires energy in (at least) two different forms. Adenosine triphosphate is involved in unfolding of precursor proteins in the cytosol and/or in release of precursor proteins from cytosolic cofactors (Pfanner et al., 1987d, 1988a; Eilers and Schatz, 1988; Murakami et al., 1988). The electrical potential (AT)across the inner mitochondrial membrane is needed for the initial transfer of precursors into or across the inner membrane (Pfanner and Neupert, 1985; Schleyer and Neupert, 1985). This article focuses on the problems of specific recognition and membrane insertion of mitochondrial precursor proteins. Harmey et al. (1977) had proposed that “some mechanism of selective recognition of precursor proteins by the mitochondria” exists. A detailed functional analysis in recent years has led to the characterization of receptor sites and of a membrane insertion site for precursors

1. MITOCHONDRIAL PROTEIN IMPORT RECEPTORS

5

and thereby provided the basis for identification of components of the mitochondria] protein import apparatus.

II. RECEPTOR PROTEINS ON THE MITOCHONDRIAL SURFACE A. Functional Characterization of Receptor Sites Pretreatment of isolated mitochondria with proteases diminished subsequent import of in vitro-synthesized precursor proteins (Gasser et al., 1982; Argan et al., 1983). A mild pretreatment with proteases inhibited not only import but also binding of precursors to mitochondria (Riezman et al., 1983; Zwizinski et al., 1984). The mitochondrial membranes were shown to remain intact under these conditions, suggesting that proteinaceous surface components are involved in import of precursor proteins. These components are assumed to perform the function of receptor sites. The import of various precursors exhibited a differential sensitivity toward pretreatment of mitochondria with specific proteases, supporting a model in which several distinct receptor sites exist on the mitochondrial surface (Zwizinski et al., 1984). Precursor proteins bind to the mitochondrial surface in the absence of a membrane potential across the inner membrane. After reestablishing a membrane potential, the precursor proteins are imported from the binding sites without prior release from the membranes (Zwizinski et al., 1983; Riezman et al., 1983; Pfanner and Neupert, 1987b; Pfanner et al., 1987d). This type of binding is termed specific (productive) binding and depends on the presence of surface proteins (“receptors”) (Zwizinski et al., 1984), whereas nonproductive (unspecific) binding also occurs to protease-pretreated mitochondria. Unspecific binding may occur to lipids of the outer membrane. The precursor of the mitochondrial outer membrane protein porin was prepared in large amounts (Pfaller et al., 1985) and bound to isolated mitochondria such that the mitochondrial import sites for porin and other precursor proteins (see Section 111) were saturated by the porin precursor. The precursor of the inner membrane protein ADP/ATP carrier could still bind to its proteinaceous surface sites (receptor); that is, binding of ADP/ATP carrier to its receptor was not competed for by porin (Pfaller et al., 1988). Since the affinities of those two precursors for interaction with mitochondria are in a similar range (Pfaller and Neupert, 1987; Pfaller et al., 1988), porin and ADP/ATP carrier seem to use distinct receptor sites. In summary, functional characterization suggests the following properties of specific mitochondrial import receptors. (i) Receptors are exposed on the mito-

6

NIKOLAUS PFANNER AND WALTER NEUPERT

chondrial surface. (ii) A specific receptor recognizes only a subset of precursor proteins; it is not involved in import of all precursor proteins. (iii) Receptors are required for specific binding of precursors to the mitochondrial surface.

B. Identification of Receptors In the past no receptor protein for import of mitochondrial precursor proteins could be identified despite numerous efforts in several laboratories. This led to speculations that the initial steps of mitochondrial protein import may not involve receptor proteins on the mitochondrial surface (Roise et al., 1986; Hurt and van Loon, 1986), although functional evidence clearly suggested the existence of specific receptor sites (summarized in Section &A; Pfanner et al., 1988b). We have started a systematic and basic approach with the aim of identifying the mitochondrial import receptors. Since the mitochondrial outer membrane contains only a relatively small number of proteins (-25), we tried to produce monospecific antibodies against each of these proteins and to test the effect of these antibodies on the import of precursor proteins. Outer membrane vesicles were isolated from mitochondria of the fungus Neurospora crassa. The proteins were separated on sodium dodecyl sulfate (SDS)-polyacrylamide gels and transferred to nitrocellulose. Twenty-five distinct bands were excised and used for generation of polyclonal antisera in rabbits. We obtained monospecific antisera against many of the outer membrane proteins (Sollner et al., 1989). In a first screening we investigated the effect of immunoglobulin G (IgG), which was prepared from the antisera, on the import of ADP/ATP carrier and porin in the assumption that these two precursor proteins use different receptor sites (see Section 11,A). Immunoglobulin G was prebound to mitochondria, the mitochondria were reisolated, and the import of in vitro-synthesized precursors was tested. Most of the IgG had no significant effect on import of precursor proteins, including IgG against porin, the major protein of the outer membrane. Immunoglobulin G directed against a mitochondrial outer membrane protein of 19 kDa (MOM19) inhibited the import of porin, but not of ADP/ATP carrier. Thus MOM19 which is exposed on the mitochondrial surface appeared to be a possible candidate for a mitochondrial import receptor (Sollner et al., 1989). Fab fragments directed against MOM 19 were prebound to mitochondria and the import of precursors to the four mitochondrial subcompartments was studied (Table I). The import of porin (outer membrane), cytochrome cl, and Fe/S protein of the bc, complex (intermembrane space side of the inner membrane), subunit 9 of F,F,-ATPase (inner membrane), and subunit p of the F,F,-ATPase (matrix side of the inner membrane) were strongly inhibited. The import of ADP/ATP carrier (inner membrane) was practically unaffected. A series of controls excluded unspecific effects of IgG and Fab fragments against MOM19 (antiMOM 19), such as inactivation of precursor proteins or cytosolic cofactors and

7

1. MITOCHONDRIAL PROTEIN IMPORT RECEPTORS

TARGETING PATHWAYS

Protein

OF

TABLE I MITOCHONDRIAL PRECURSOR PROTEINS

Functional destination

Present in N-terminal prokaryotic targeting ancestor sequence

Receptor

Membrane insertion

Cytochrome cl Intermembrane space Fe/S protein Intermembrane space Fo-ATPase subunit 9 Inner membrane Fl-ATPase subunit p Matrix Porin Outer membrane

Yes Yes Yes Yes

Yes

?

?

ADP/ATP carrier

Inner membrane

No

No

MOM72b GIP

Cytochrome c

Intermembrane space

Yes

?

No surEndogenous face reactivity of ceptor precursor

UMitochondrial outer membrane protein of 19 kDa. bMitochondria1 outer membrane protein of 72 kDa. General insertion protein.

inhibition of later transport steps (translocation from the outer into the inner membrane, the membrane potential-dependent step, and proteolytic processing of precursor proteins) (Sollner et al., 1989). As a further control, the import of cytochrome c was tested. The precursor, apocytochrome c, spontaneously inserts into the outer mitochondrial membrane (Rietveld et al., 1985; Stuart et al., 1990) and does not use a protease-accessible surface receptor (Nicholson et al., 1988). As expected, anti-MOM19 did not inhibit the import of apocytochrome c. Inhibition of protein import by anti-MOM19 occurred at the level of specific binding of precursors to the mitochondrial surface. Thus MOM19 fulfills all the functional criteria established for a mitochondrial import receptor: it is exposed on the mitochondrial surface; it is involved in the import of a subset of precursor proteins, and it is required for specific binding of precursors to mitochondria. We conclude that the outer membrane protein MOM19 is identical to (or closely associated with) a specific import receptor (Sollner et al., 1989). With similar procedures we found IgG and Fab fragments against a mitochondrial outer membrane protein of 72 kDa (MOM72). Anti-MOM72 selectively inhibited the import of ADP/ATP carrier, but not of other mitochondrial precursor proteins tested. The inhibition of import occurred at the level of specific binding of ADP/ATP carrier to the mitochondrial surface, whereas other import steps were unaffected (Sollner et al., 1990). We conclude that MOM72, which is exposed on the mitochondrial surface, represents a specific import receptor for ADP/ATP carrier and probably similar precursors (Table I).

a

NIKOLAUS PFANNER AND WALTER NEUPERT

C. Import of Precursor Proteins Bypassing Receptor Sites Blocking of receptor sites by specific antibodies or degradation of receptors by treatment of mitochondria with proteases strongly reduced the import rates of mitochondrial precursor proteins (Pfaller et al., 1989; Sollner et al., 1989). This suggests a crucial role of receptor sites for the efficiency of protein import. A residual import of precursors, however, can also occur when the surface receptors are blocked or degraded. This low efficient import still exhibits several basic features of mitochondrial protein import, including dependence on ATP and membrane potential A* and translocation via contact sites (Pfaller et al., 1989). Precursor proteins are obviously able to bypass surface receptors and enter the mitochondrial import pathways at a later stage. The very low efficiency suggests that bypass import does not significantly contribute to the import processes under physiological conditions (Pfanner et al., 1988~). The existence of bypass import, at least under certain experimental conditions (high amounts of precursor proteins), led to a series of very surprising findings. Nonmitochondrial “targeting” sequences, such as a chloroplast signal sequence (Hurt et al., 1986) or sequences of a cytosolic protein (Hurt and Schatz, 1987), could direct proteins into mitochondria albeit with a low efficiency (summarized in Pfanner er al., 1988~).The main common property of these nonmitochondrial signals was the abundance of positively charged amino acid residues, which appears to be an essential requirement for the A*-dependent insertion into the inner membrane. Import directed by nonmitochondrial signals was not affected by pretreatment of mitochondria with proteases. These “targeting” signals apparently bypass the surface receptors (Pfaller et al., 1989). Mitochondria1 import receptors therefore specifically interact with authentic mitochondrial targeting sequences; receptors are responsible for the selectivity of protein import. Yeast mitochondria with disrupted outer membrane are able to translocate precursor proteins directly across the inner membrane (Ohba and Schatz, 1987a; Rassow, Pfanner, and Neupert, in preparation). The nature and function of these inner membrane import sites is unknown. They might for example be related to the translocation sites that are (permanently or transiently) present in contact sites between both membranes (Schwaiger et al., 1987). Further studies are required to decide how specific these import sites are and if they contain components for recognition of mitochondrial precursor proteins.

111.

THE “GENERAL INSERTION PROTEIN”

After interaction with specific surface receptors, mitochondrial precursor proteins are inserted into the outer membrane. Studies on the import pathways of

1. MITOCHONDRIAL PROTEIN IMPORT RECEPTORS

9

ADP/ATP carrier and porin suggested the existence of a new functional component for protein translocation across membranes, a membrane insertion site (Pfanner and Neupert, 1987b; Pfaller and Neupert, 1987). Precursor proteins that are inserted into the outer membrane are not accessible to specific antibodies or to low concentrations of proteases added to the mitochondria, in contrast to precursors that are bound to the surface receptors (Sollner et ul., 1988), suggesting that the membrane insertion sight is buried in the outer membrane. The membrane insertion sites are saturable; the determined number of sites is practically identical for ADP/ATP carrier and porin and is in a similar range as the number of receptor sites (Pfaller and Neupert, 1987; Pfaller et al., 1988). Precursor proteins inserted into the outer membrane are extractable from the membranes by “hydrophilic perturbants” or “protein denaturants” such as carbonate ions or urea (Pfanner and Neupert, 1987b; Pfaller and Neupert, 1987). The precursor proteins thus may be inserted into a proteinaceous membrane environment. The precursor of porin competed for the import of nearly all other mitochondrial precursor proteins tested, including cytochrome cl, Fe/S protein, F,-ATPase subunit 9, F,-ATPase subunit p, and ADP/ATP carrier (Table I). Competition of import specifically occurred at the level of insertion of precursors into the outer membrane (Pfaller et al., 1988) and not for the interaction with receptor sites (see Section 11,A). We concluded that the various precursor proteins competed for interaction with the same component of the protein import apparatus, namely, a common membrane insertion site. The only precursor protein the import of which was not competed for was apocytochrome c (Table I); this fits well with the bulk of evidence suggesting that cytochrome c uses a very unique import pathway (Nicholson et al., 1988). The common membrane insertion site, termed the general insertion protein (GIP), has not been identified so far. The receptors MOM19 and MOM72 can form a high molecular weight complex that contains two other outer membrane proteins. One of these proteins, termed MOM38 (molecular weight 38 K), exhibits the properties expected of GIP (Pfaller, Sollner, Griffiths, Pfanner, and Neupert, in preparation). A recent finding on protein import into yeast mitochondria also may be of interest for the identification of GIP. Antibodies directed against 45-kDa mitochondrial proteins inhibit import of precursor proteins when bound to mitochondria that had been pretreated with proteases (Ohba and Schatz, 1987b). Where the antibodies bound to intact mitochondria, the inhibition of import was only marginal; the antibodies obviously do not block receptor proteins that are exposed on the mitochondrial surface. Since the inactivated component(s) appears to be protected against proteases, it may be buried in the outer membrane and has to be “freed” from other proteins by the pretreatment with proteases in order to be accessible to the antibodies. We speculate that a component that is recognized by the anti-45-kDa antibodies is related to GIP (Pfaller et

10

NIKOLAUS PFANNER AND WALTER NEUPERT

al., 1988). Vestweber el al. (1989) reported that these antibodies also recognized a protein of 42 kDa, termed ISP42 (import site protein). Inhibition of import was found to be caused by the anti-ISP42 antibodies. Moreover, a fraction of precursor proteins that was accumulated in contact sites was cross-linked to ISP42. MOM38 of Neurospora crassa and ISP42 of yeast thus may be related to GIP.

IV.

ROLE OF RECEPTORS AND GENERAL INSERTION PROTEIN

Receptor proteins on the mitochondrial surface are responsible for the specificity and selectivity of protein uptake (see Section I1,C). They recognize mitochondria] targeting sequences and strongly enhance the import rates of those precursors. Precursor proteins with amino-terminal targeting sequence interact with the receptor MOM19, whereas ADP/ATP carrier, a precursor with several internal targeting sequences (but no N-terminal signal), interacts with MOM72 (Table I). Treatment of mitochondria with the protease elastase generates an 17-kDa fragment of MOM19 that still mediates import of F,-ATPase subunit f3, whereas import of other MOM 19-dependent precursor proteins is inhibited (Sollner et al., 1989). Distinct segments of MOM19 may thus be responsible for interaction with the various precursor proteins. This offers the possibility for characterization of functional sites of this import receptor. Most mitochondrial proteins that were found to require MOM19 for import have equivalents in bacteria and thus were probably already present in the prokaryotic ancestors of mitochondria (Table I) (the evolutionary origin of porin is unknown). According to the endosymbiont hypothesis, after endocytosis of the prokaryotic cell, the (now) mitochondrial genes for these proteins were transferred to the nucleus (see Hart1 et al., 1987, for a discussion). An amino-terminal targeting sequence directed the proteins back to mitochondria. We propose that MOM19 was used as surface receptor for those precursor proteins. On the other hand, proteins exist that were most likely not present in the prokaryotic ancestor. The ADP/ATP carrier was probably established in the eukaryotic cell (Klingenberg, 1985). Its targeting sequences are not located at the amino terminus of the precursor protein (Pfanner et al., 1987b; Smagula and Douglas, 1988), and the precursor uses a different surface receptor, MOM72 (Sollner et al., 1990). The receptors MOM19 and MOM72 then transfer the precursor proteins to the GIP in the outer membrane. The receptors themselves may possess some activity for membrane insertion of precursor proteins and thereby facilitate the action of GIP. Alternatively, receptors may only be able to bind precursor proteins, and the insertion into the outer membrane is solely performed by GIP (possibly in cooperation with lipids of the outer membrane). As described earlier, MOM19, MOM72, and probably GIP can be detected in a protein complex in the outer membrane (Pfaller et al., in preparation). The receptors and GIP may not func-

1. MITOCHONDRIAL PROTEIN IMPORT RECEPTORS

11

tion as independent entities in the outer membrane. Their possible assembly into a multisubunit complex may help in coordination of their activities; it may even be a prerequisite for some of their functions. Beyond GIP, the import pathways diverge; some precursors assemble into the outer membrane (porin), whereas most precursors move on to contact sites between both membranes and then to the other mitochondrial subcompartments (Pfaller et al., 1988; Hart1 et al., 1989). Since translocation of proteins across the mitochondrial membranes occurs predominantly at contact sites, receptors and GIP may be concentrated in contact site regions of the outer membrane to ensure efficient and rapid translocation of precursor proteins. It might well be that, in addition, receptors and GIP are distributed over the entire mitochondrial surface in order to increase the probability for the initial high-affinity binding of precursor proteins and to collect the precursors for transfer to contact sites. This implies lateral diffusion of receptors and/or GIP in the outer membrane. V.

SUMMARY AND PERSPECTIVES

Functional characterization of initial steps of mitochondrial protein import provided the tools for identification of two mitochondrial outer membrane proteins, MOM 19 and MOM72, as specific receptors for precursor proteins. Bound precursor proteins are transferred to a common membrane insertion site in the outer membrane, the “general insertion protein” (GIP). Future research will address the role of functional domains of the receptors and the type and specificity of interaction with precursor proteins. The existence of (at least) two distinct membrane-bound receptors for precursor proteins appears to be of relevance for protein translocation across membranes in general; it may have implications on other organelles such as chloroplasts and the endoplasmic reticulum. A GIP for the entry of precursor proteins into a membrane is most likely not only present in mitochondria, but may also be found in several other biological membranes. For instance, the SecY (PrlA) protein, an integral protein of the cytoplasmic membrane of Escherichia coli that is involved in export of proteins (Watanabe and Blobel, 1989; Wickner, 1989), might have a similar role. ACKNOWLEDGMENTS We thank R . A. Stuart for critical reading of the manuscript. We are grateful to our colleagues for their many contributions to this work. REFERENCES Argan, C., Lusty, C. J., and Shore, G. C. (1983). Membrane and cytosolic components affecting transport of the precursor for ornithine carbamyltransferase into mitochondria. J . Biol. Chem. 258, 6667-6670. Attardi, G . , and Schatz, G. (1988). Biogenesis of mitochondria. Annu. Rev. Cell Biol. 4, 289-333.

12

NIKOLAUS PFANNER AND WALTER NEUPERT

Eilers, M., and Schatz, G. (1988). Protein unfolding and the energetics of protein translocation across biological membranes. Cell 52, 481-483. Gasser, S. M., Daum, G., and Schatz, G. (1982). Import of proteins into mitochondria: Energydependent uptake of precursors by isolated mitochondria. J . Biol. Chem. 257, 13034-13041. Hallermayer, G., and Neupert, W. (1976). Studies on the synthesis of mitochondrial proteins in the cytoplasm and on their transport into the mitochondrion. In “Genetics and Biogenesis of Chloroplasts and Mitochondria” (T. Biicher, W. Neupert, W. Sebald, and S. Werner, eds.), pp. 807-812. North-Holland Publ., Amsterdam. Hallermayer, G., Zimmermann, R., and Neupert, W. (1977). Kinetic studies on the transport of cytoplasmically synthesized proteins into the mitochondria in intact cells of Neurosporu crussu. Eur. J . Biochem. 81, 523-532. Harmey, M. A,, Hallermayer, G., and Neupert, W. (1976). In vitro synthesis and transport into mitochondria of cytoplasmically translated proteins. In “Genetics and Biogenesis of Chloroplasts and Mitochondria” (T. Biicher, W. Neupert, W. Sebald, and S . Werner, eds.), pp. 8138 18. North-Holland Publ., Amsterdam. Harmey, M. A,, Hallermayer, G., Korb, H., and Neupert, W. (1977). Transport of cytoplasmically synthesized proteins into the mitochondria in a cell free system from Neurosporu crussu. Eur. J . Biochem. 81, 533-544. Hartl, F.-U., Ostermann, J., Guiard, B., and Neupert, W. (1987). Successive translocation into and out of the mitochondrial matrix: Targeting of proteins to the intermembrane space by a bipartite signal peptide. Cell 51, 1027-1037. Hartl, F.-U., Pfanner, N., Nicholson, D. W., and Neupert, W. (1989). Mitochondria1protein import. Biochim. Biophys. Actu 988, 1-45. Hawlitschek, G., Schneider, H., Schmidt, B., Tropschug, M., Hartl, F.-U., and Neupert, W. (1988). Mitochondrial protein import: Identification of processing peptidase and of PEP, a processing enhancing protein. Cell 53, 795-806. Horwich, A. L., Kalousek, F., Mellman, I., and Rosenberg, L. E. (1985). A leader peptide is suficient to direct mitochondrial import of a chimeric protein. EMBO J . 4, 1129-1 135. Hurt, E. C., and Schatz, G. (1987). A cytosolic protein contains a cryptic mitochondrial targeting signal. Nature (London) 325, 499-503. Hurt, E. C., and van Loon, A. P. G. M. (1986). How proteins find mitochondria and intramitochondrial compartments. Trends Biochem. Sci. 11, 204-207. Hurt, E. C., Soltanifar, N., Goldschmidt-Clermont M., Rochaix, J.-D., and Schatz, G. (1986). The cleavable pre-sequence of an imported chloroplast protein directs attached polypeptides into yeast mitochondria. EMBO J . 5 , 1343-1350. Klingenberg, M. (1985). Principles of carrier catalysis elucidated by comparing two similar membrane translocators from mitochondria, the ADPIATP carrier and the uncoupling protein. Ann. N . Y . Acud. Sci. 456, 279-288. Lewin, A. S., and Norman, D. K. (1983). Assembly of F,-ATPase in isolated mitochondria. J . Biol. Chem. 258, 6750-6755. Murakami, H., Pain, D., and Blobel, G. (1988). 70-kD heat shock-related protein is one of at least two distinct cytosolic factors stimulating protein import into mitochondria. J . Cell B i d . 107, 205 1-2057. Nicholson, D. W., Hergersberg, C., and Neupert, W. (1988). Role of cytochrome c heme lyase in the import of cytochrome c into mitochondria. J . Biol. Chem. 263, 19034-19042. Ohba, M., and Schatz, G. (1987a). Disruption of the outer membrane restores protein import to trypsin-treated yeast mitochondria. EMBO J . 6, 21 17-2122. Ohba, M . , and Schatz, G. (1987b). Protein import into yeast mitochondria is inhibited by antibodies raised against 45-kd proteins of the outer membrane. EMBO J . 6, 2109-21 15.

1. MITOCHONDRIAL PROTEIN IMPORT RECEPTORS

13

Pfaller, R., and Neupert, W. (1987). High-affinity binding sites involved in the import of porin into mitochondria. EMBO J. 6, 2635-2642. Pfaller, R., Freitag, H., Harmey, M. A., Benz, R., and Neupert, W. (1985). A water-soluble form of porin from the mitochondrial outer membrane of Neurosporu crussu; Properties and relationship to the biosynthetic precursor form. J. Biol. Chem. 260, 8188-8193. Pfaller, R., Steger, H. F., Rassow, J., Pfanner, N., and Neupert, W. (1988). Import pathways of precursor proteins into mitochondria: Multiple receptor sites are followed by a common membrane insertion site. J . Cell Biol. 107, 2483-2490. Pfaller, R., Pfanner, N., and Neupert, W. (1989). Mitochondria1 protein import: Bypass of proteinaceous surface receptors can occur with low specificity and efficiency. J. Biol. Chem. 264, 34-39. Pfanner, N., and Neupert, W. (1985). Transport of proteins into mitochondna: A potassium diffusion potential is able to drive the import of ADP/ATP carrier. EMBO J. 4, 2819-2825. Pfanner, N., and Neupert, W. (1987a). Biogenesis of mitochondrial energy transducing complexes. Curr. Top. Bioenerg. 15, 177-219. Pfanner, N., and Neupert, W. (1987b). Distinct steps in the import of ADP/ATP carrier into mitochondria. J . Biol. Chem. 262, 7528-7536. Pfanner, N., Hartl, F.-U., Guiard, B., and Neupert, W. (1987a). Mitochondria1 precursor proteins are imported through a hydrophilic membrane environment. Eur. J. Biochem. 169, 289-293. Pfanner, N., Hoeben, P., Tropschug, M., and Neupert, W. (1987b). The carboxyl-terminal two thirds of the ADP/ATP carrier polypeptide contains sufficient information to direct translocation into mitochondria. J. Biol. Chem. 262, 14851-14854. Pfanner, N., Miiller, H.,Harmey, M. A., and Neupert, W. (1987~).Mitochondrial protein import: Involvement of the mature part of a cleavable precursor protein in the binding to receptor sites. EMBO J. 6, 3449-3454. Pfanner, N., Tropschug, M., and Neupert, W. (1987d). Mitochondrial protein import: Nucleoside triphosphates are involved in conferring import-competence to precursors. Cell 49, 8 15-823. Pfanner, N., Pfaller, R., Kleene, R., Ito, M., Tropschug, M., and Neupert, W. (1988a). Role of ATP in mitochondrial protein import: Conformational alteration of a precursor protein can substitute for ATP requirement. J. Biol. Chem. 263, 4049-4051. Pfanner, N., Hartl, F.-U., and Neupert, W. (1988b). Import of proteins into mitochondria: A multistep process. Eur. 1. Biochem. 175, 205-212. Pfanner, N., Pfaller, R., and Neupert, W. (1988~).How finicky is mitochondrial protein import? Trends Biochem. Sci. 13, 165-167. Rassow, J., Guiard, B., Wienhues, U., Herzog, V., Hartl,F.-U., and Neupert, W. (1989). Translocation arrest by reversible folding of a precursor protein imported into mitochondria. A means to quantitate translocation contact sites. J. Cell Biol. 109, 1421- 1428. Rietveld, A., Ponjee, G. A. E., Schiffers, P., Jordi, W., van de Coolwijk, P. J. F. M., Demel, R. A., Marsh, D., and de Kruijff, B. (1985). Investigations on the insertion of the mitochondrial precursor protein apocytochrome c into model membranes. Biochim. Biophys. Acru 818, 398409. Riezman, H., Hay, R., Witte, C . , Nelson, N., and Schatz, G. (1983). Yeast mitochondrial outer membrane specifically binds cytoplasmically synthesized precursors of mitochondrial proteins. EMEO J. 2, 1113-1118. Roise, D. Horvath, S. J., Tomich, J. M., Richards, J. H., and Schatz, G. (1986). A chemically synthesized pre-sequence of an imported mitochondrial protein can form an amphiphilic helix and perturb natural and artificial phospholipid bilayers. EMBO J . 5, 1327-1334. Scbleyer, M., and Neupert, W. (1985). Transport of proteins into mitochondria: Translocational intermediates spanning contact sites between outer and inner membranes. Cell 43, 339-350.

14

NIKOLAUS PFANNER AND WALTER NEUPERT

Schmidt, B . , Hennig, B., Zimmermann, R., and Neupert, W. (1983). Biosynthetic pathway of mitochondrial ATPase subunit 9 in Neurosporu crussa. J . Cell Biol. 96, 248-255. Schwaiger, M., Herzog, V., and Neupert, W. (1987). Characterization of translocation contact sites involved in the import of mitochondrial proteins. 1. Cell Biol. 105, 235-246. Smagula, C. S . , and Douglas, M. G. (1988). ADP-ATP carrier of Saccharomyces cerevisiae contains a mitochondrial import signal between amino acids 72 and 111. J . Cell. Biochem. 36,323-328. Sollner, T., Pfanner, N., and Neupert, W. (1988). Mitochondria1protein import: Differential recognition of various transport intermediates by antibodies. FEES Left. 229, 25-29. Sollner, T., Grifiths, G., Pfaller, R., Pfanner, N., and Neupert, W. (1989). MOM19, an import receptor for mitochondrial precursor proteins. Cell, in press. Sollner, T., Pfaller, R., Grifiths, G., Pfanner, N., and Neupert, W. (1990). A mitochondrial import receptor for the ADPlATP carrier. Submitted. Stuart, R. A,, Nicholson, D. W., and Neupert, W. (1990). Early steps in mitochondrial protein import: Receptor functions can be substituted by membrane insertion activity of apocytohchrome c. Celf 60, in press. Vestweber, D., and Schatz, G. (1988). Mitochondria can import artificial precursor proteins containing a branched polypeptide chain or a carboxyl-terminal stilbene disulfonate. J . Cell Biol. 107, 2045-2049. Vestweber, D., Brunner, J., Baker, A,, and Schatz, G. (1989). A 42k outer-membrane protein is a component of the yeast mitochondria1 protein import site. Nature (London) 341, 205-209. Watanabe, M., and Blobel, G. (1989). Site-specific antibodies against the PrlA (SecY) protein of Escherichia coli inhibit protein export by interfering with plasma membrane binding or preproteins. Proc. Natl. Acad. Sci. U.S.A. 86, 1895-1899. Wickner, W. (1989). Secretion and membrane assembly. Trends Biochem. Sci. 14, 280-283. Wickner, W. T., and Lodish, H. F. (1985). Multiple mechanisms of protein insertion into and across membranes. Science 230, 400-407. Zwizinski, C., Schleyer, M., and Neupert, W. (1983). Transfer of proteins into mitochondria: Precursor to the ADP/ATP carrier binds to receptor sites on isolated mitochondria. J . Biol. Chem. 258, 4071-4074. Zwizinski, C . , Schleyer, M., and Neupert, W. (1984). Proteinaceous receptors for the import of mitochondrial precursor proteins. J . Biol. Chem. 259, 7850-7856.

CURRENT TOPICS IN MEMBRANES AND TRANSPORT, VOLUME 36

Chapter 2 An Enzymological Approach to Membrane Assembly and Protein Secretion BILL WICKNER Molecular Biology Institute and Department of Biological Chemistry University of California, Los Angeles Los Angeles, California 90024

Despite almost two decades of study, little is known at a mechanistic level of how proteins assemble into, or across, biological membranes. This problem has been intensively studied for mitochondria, endoplasmic reticulum, chloroplasts, and the bacterial cell surface. In mammalian systems, in vivo studies have been difficult, and most work has focused on addition of intact organelles to in vitro protein synthesis reactions. Microorganisms offer several major advantages for the study of this fundamental process. They have advanced genetics, they can readily be grown in large culture for biochemistry, and their growth on minimal medium allows isotopic labeling for in vivo studies. Bacterial secretion studies have employed each of these major avenues, and today are the most advanced toward mechanistic understanding of protein translocation. Genetic identification of temperature-sensitive mutants that affect secretion (sec mutants) or of mutants that suppress leader sequence mutants, and are thereby more permissive for protein localization @rl mutants), has led to the identification of several genes that are essential for the membrane transit of most proteins. In vivo studies, in which growing cells are pulse-labeled and then “chased” with a chemical excess of nonradioactive amino acids, has established that translocation is not coupled to translation in this organism. Indeed, the “battle” over whether bacterial translocation is obligately cotranslational or not set the stage for subsequent studies in yeast and mammalian systems that have shown that translocation is, in general, never coupled to translation in Nature (or Science or Cell!). With the experimental resolution of the process into the distinct steps of translation and translocation, it became possible to study the energetics 15 Copyright 0 1990 by Academc Press, Inc. All rights of reproduction in any form reserved.

16

BILL WICKNER

of the process. Bacterial protein export, as with mitochondria1 protein import, requires a membrane electrochemical potential in vivo. After translocation, leader peptidase, an inner membrane protein, cleaves the amino-terminal leader sequence from exported proteins. Two years ago, the only available pure component of this system was leader peptidase. Thus, the biochemistry of this process was not yet studied. To begin this study, we have undertaken an enzymological approach, beginning with the in vivo translocation reactions of Tai, Blobel, and their colleagues, and moving toward a reconstitution of protein translocation that will employ defined, purified components. We began by isolating a precursor form of a secreted protein. We chose proOmpA, the precursor form of the major outer membrane protein A. The isolation of proOmpA was hampered by its limited solubility. Elliott Crooke, who at the time was a student in my laboratory, chose the bold approach of solubilizing the proOmpA in 8 M urea, reflecting a faith that the proOmpA would be renaturable for membrane assembly once it was purified and the urea was removed. When Crooke purified the proOmpA and removed the urea by dialysis, he was dismayed to find that the proOmpA was totally inactive for translocation. However, he found that the addition of a soluble protein extract to the proOmpA prior to dialysis allowed the dialyzed proOmpA to be active for membrane assembly. This provided the assay for “trigger factor,” a protein that stabilizes the proOmpA in an active confirmation for membrane assembly. Trigger factor is just one of the three known Escherichia coli “chaperone” proteins; the others are GroEL and SecB protein. The relative roles of each in protein export will require a genetic analysis; this should be facilitated by the recent isolation of the trigger factor gene (B. Guthrie and W. Wickner, unpublished). Biochemically, each is able to form a complex with proOmpA. This complex is isolable and stable, yet can readily release the proOmpA to assembly across the membrane. Studies are under way to try to determine the structure of proOmpA in these complexes, that is, whether the proOmpA has a unique tertiary structure or is unfolded at the tertiary-structure level. In addition, it will be important to determine the basis of specificity of these and other chaperone proteins. The ability to combine genetics and biochemistry in yeast and E . coli should make these the premier organisms for such studies. In our experiments, the only cytosolic proteins that are needed for translocation are the chaperones and the preproteins themselves. Proceeding to the membrane, the SecA protein is essential for translocation. Thanks to the studies of Don Oliver, this protein was available in large quantities and had been shown to be a peripheral membrane protein of 102,000 Da. We found that the pure SecA protein would restore activity to membranes that had been inactivated for translocation by N,-ATP, and proceeded from this observation to the finding that SecA itself is an ATPase. Most strikingly, this ATPase activity is subject to a remarkable group of allosteric controls that we are only just now sorting out.

2. AN ENZYMOLOGICAL APPROACH

17

SecA protein is a fully water-soluble protein, with little tendency to aggregate at even high protein concentrations. Nevertheless, SecA binds to membrane vesicles or even to liposomes as a peripheral membrane protein. We have not yet been able to demonstrate any high-affinity binding to a saturable site, although this may simply reflect current limitations on our assay conditions. The ATP hydrolytic activity of SecA is stimulated up to 100-fold by the simultaneous presence of proOmpA (or other precursor proteins) and inner membrane vesicles. We term this activity “translocation ATPase.” What are the characteristics of this translocation ATPase? Omission of any one of Mg, SecA protein, proOmpA, or inner membrane vesicles results in a complete loss of the translocation ATPase activity. The membranes contribute at least two essential components, the lipid and the SecY protein. The proOmpA itself must be in a “competent” state for translocation (either freshly diluted from urea or in complex with chaperone) in order to function for translocation. This assay is thus clearly coupled to each of the elements of protein transit across membranes. However, there is not a tight, stoichiometriccoupling of the moles of ATP hydrolyzed to the moles of proOmpA translocated, and in this sense we view our current translocation ATPase assay as “uncoupled.” We have been able to exploit this reaction to advance our understanding in two directions. Membrane vesicles have been solubilized in detergent and reconstituted to form proteoliposomes that support translocation ATPase. The detergent extract contains solubilized SecY, and this SecY is required for the reconstitution. The secY protein is being purified in a functional state based on this assay. A second manner in which we have exploited the translocation ATPase assay began with a humble Mg concentration curve. Roland Lill in the laboratory found that, at low Mg level, the SecA ATPase can be stimulated to high levels of activity by proOmpA and lipids, in the absence of SecY or other integral membrane proteins. This activity depends on the proOmpA as directly as the SecYdependent translocation ATPase. In addition, we have observed that the leader peptide plus the membrane protein can substitute for the intact precursor protein (R. Lill, unpublished observations)! Thus, the SecA protein can directly recognize the preprotein. Many challenges remain in the study of protein translocation. How is the energy of ATP hydrolysis and the membrane electrochemicalpotential coupled to the translocation event? What is the chemical role of the SecY protein, and of other integral membrane proteins? Does the precursor protein pass through the bilayer directly or through a proteinaceous transport pore? I am confident that the resolution and functional reconstitution of the integral membrane proteins that support translocation will be essential to answer these mechanistic questions. My laboratory and many others are devoted to this approach.

This Page Intentionally Left Blank

CLRRENT TOPICS IN MEMBRANES AND TRANSPORT, VOLUME 36

Chapter 3

Protein Translocation in Yeast JOANNE CROWE AND DAVID I . MEYER Department of Biological Chemistry and Molecular Biology Institute School of Medicine University of California, Los Angeles Los Angeles, California 90024

I.

Introduction

11. Yeast as a Model System 111. The Importance of Preprotein Conformation

IV. Binding to the Membrane and Translocation Are Separable Events A. Binding to the Membrane B. Translocation V. Analysis of ptll-A Translocation Mutant A. In Vivo Analysis B . In Vitro Analysis References

1.

INTRODUCTION

In eukaryotes, proteins that are destined for secretion, lysosomes, the plasma membrane, and organelles of the secretory pathway must be transported from their site of synthesis to their ultimate location in the cell. The first step in this process is the targeting of the nascent protein to the rough endoplasmic reticulum (ER), followed by translocation into or across the membrane. Analysis of this process in vitro was made possible by the development of the heterologous cellfree translocation assay (Blobel and Dobberstein, 1975), the use of which resulted in the characterization of several participating components. These include the signal recognition particle (SRP) (Walter et al., 1981; Walter and Blobel, 1981a,b), its receptor, the docking protein (Meyer et a l . , 1982; Gilmore et al., 1982), signal peptidase-purified as a complex (Evans 'et a l . , 1986)-the putative signal sequence receptor (SSR) (Wiedmann et al., 1987), and a protein 19 Copyright 6 1990 by Academc Press, Inc. All nghts of reproduction in any form reserved.

20

JOANNE CROWE AND DAVID I . MEYER

associated with the docking protein (Tajima et al., 1986). The binding of ribosomes to the ER appears to depend on an unidentified membrane protein (Hortsch et a l . , 1986). There are probably several other ER proteins involved in the translocation process; however, their isolation has not yet been possible using conventional biochemical techniques in an in vitro system. Proteins specific to the rough ER have not so far been efficiently fractionated, making identification of participants by reconstitution of translocation in vitro very difficult.

II. YEAST AS A MODEL SYSTEM The yeast Saccharomyces cerevisiae carries out secretion and membrane biogenesis in a fashion analogous to mammalian cells and possesses all the relevant organelles: rough ER, Golgi apparatus, and secretory vesicles. In addition, it has a well-characterized, easily manipulated genome, and a convenient methodology exists for the isolation of genes of interest. The combination of in vivo and genetic analyses of yeast translocation mutants with the in vitro analysis of these mutants using the yeast cell-free translocation assay (Rothblatt and Meyer, 1986; Hansen et a l . , 1986; Waters and Blobel, 1986) represents a powerful system for characterizing both membrane and cytosolic components involved in translocation. This approach has led to considerable understanding of the events in this process (summarized in Table I). An important advantage of the yeast cell-free assay, which has been exploited in many studies (reviewed in Zimmerrnann and Meyer, 1986), is the uncoupling of translocation from translation. The precursor of the yeast mating pheromone, prepro-a-factor, can be translocated across yeast rough microsomes in the total absence of ongoing protein synthesis, although certain cytosolic factors are required for this process (Rothblatt e f a l . , 1987; Sanz and Meyer, 1988; Chirico et TABLE I EVENTSIN PROTEIN TRANSLOCATION IN YEAST" Location Cytosol ER Cytosolic face ER Membrane ER Lumen

Process

Translation Conformation preservation Recognition Binding to the membrane Translocation Processing, glycosylation

OThe process of translocation in yeast can be divided into several separate events, each occurring at a precise intracellular location.

3. PROTEIN TRANSLOCATION IN YEAST

21

al., 1988; Deshaies et al., 1988). A more refined assay, in which affinity-purified precursor proteins-rapidly diluted out of urea-can cross yeast rough microsomes in the absence of cytosolic factors, has allowed further dissection of the translocation process in vitro (Sanz and Meyer, 1988; Crooke et al., 1988a). 111. THE IMPORTANCE OF PREPROTEIN CONFORMATION There is considerable evidence that preprotein conformation and its maintenance play an important role in translocation across a variety of different membranes (reviewed in Meyer, 1988). For certain bacterial preproteins, such as proOmpA, a cytosolic component known as trigger factor is needed for translocation in vitro across inverted vesicles (Crooke and Wickner, 1987). It has been shown to maintain the protein in a more protease-sensitive state, indicative of a more relaxed or “open” conformation. The product of the bacterial SecB gene appears to perform a similar function to trigger factor, albeit acting on different precursor proteins, both in vivo and in vitro (Randall and Hardy, 1986; Collier et al., 1988). Even in the absence of these cytosolic proteins, the presence of a signal sequence on the precursor of the bacterial maltose-binding protein has been shown to play a role in the retardation of folding (Park et al., 1988). Cytosolic factors identified as members of the hsp70 family of heat shock proteins have been postulated to act as preprotein “unfoldases” or preservers of conformation (Pelham, 1986; Rothman and Kornberg, 1986). These proteins, which bind ATP, have been shown in yeast, both in vitro and in vivo, to be necessary for translocation into the ER, and also to be involved in import into mitochondria and vacuoles (Chirico et al., 1988; Deshaies et al., 1988; Murakami et al., 1988). They have also been shown to stimulate translocation of preproteins across mammalian rough ER in vitro (Zimmermann et al., 1988). Several studies have examined the effect of precursor folding on translocation. Mitochondria1 membranes, which are able to translocate even large, branched molecules (Vestweber and Schatz, 1988a,b), show dramatically reduced translocation efficiencies when the preprotein is artificially held in a folded conformation (Eilers and Schatz, 1986, 1988). The use of denatured precursor proteins enhanced the efficiency of posttranslational translocation across mitochondrial, bacterial, and yeast ER membranes in vitro (Sanz and Meyer, 1988; Eilers et al., 1988; Crooke et al., 1988b). In this assay, precursors that have been denatured in 8 M urea are competent for translocation when the urea is rapidly diluted into a solution of membranes in buffer only. This competence is lost as the time between dilution and the addition of membranes increases. The translocation competence can be preserved over longer periods of time if the preproteins are diluted from the urea into a solution of appropriate cytosolic factors. This provides an

22

JOANNE CROWE AND DAVID I. MEYER

experimental system for the analysis of the requirements of posttranslational translocation. Using this simplified version of the yeast in vitro system, Sanz and Meyer (1988) have shown that SRP, when present during the refolding of urea-denatured proOmpA or prepro-a-factor, can replace all the cytosolic factors required for posttranslational translocation competence in yeast, as was also shown in bacteria (Crooke et al., 1988a). It can also stimulate the translocation of purified proOmpA across pancreatic microsomes. In addition, they demonstrated that proOmpA incubated with SRP showed enhanced sensitivity to protease, suggesting that it is in a more relaxed conformation than native proOmpA. These data imply that, among other things, SRP enhances the translocation competence of the preprotein by maintaining it in a more open or unfolded state. Since SRP is the only cytosolic factor required for translocation across mammalian microsomes, it is likely that it fulfills this function in addition to its well-established role in targeting proteins to the rough ER. These dual functions put SRP into the family of proteins described as “molecular chaperones” (Ellis, 1987), and it will be interesting to see which of its subunits is responsible for each activity. Although the yeast hsp70 heat shock proteins stimulate translocation in vitro (Chirico et al., 1988; Deshaies et al., 1988; Murakami et al., 1988) they require the presence of other cytosolic components to perform this function. Moreover, unlike SRP, neither the hsp70 proteins, SecB protein, nor trigger factor have yet been shown to interact with signal sequences or to be directly involved in the translocation process. A yeast equivalent of mammalian SRP has not yet been found. When these additional yeast factors are eventually isolated, the cloning of their genes will allow a more detailed analysis of their role in translocation in vivo .

IV. BINDING TO THE MEMBRANE AND TRANSLOCATION ARE SEPARABLE EVENTS Translocation has been shown to be a two-step process for both yeast (Sanz and Meyer, 1989) and Neurospora (Addison, 1988) microsomes, for bacterial membranes (Thom and Randall, 1988), and also for nuclear-encoded proteins crossing mitochondria1 membranes (Eilers and Schatz, 1988; Pfanner er al., 1988). Whether this extends to mammalian ER remains to be seen. The mammalian system, in particular, is difficult to elucidate because of the inability of microsomes to translocate most proteins posttranslationally. However, there is some evidence to suggest that it might also be a multistep process. An engineered protein with a C-terminal loop has been shown to bind to mammalian microsomes, and only to be translocated upon the reduction of the disulfide bonds maintaining the loop (Miiller and Zimmerman, 1988). In addition, a possible

3. PROTEIN TRANSLOCATION IN YEAST

23

SSR has been found in mammalian membranes (Wiedmann et al., 1987), but the protein and its function remain to be characterized. Similarly, a putative outer membrane receptor for transit peptides has been found in chloroplast membranes (Pain et al., 1988). Sanz and Meyer (1989) have used a further refinement of the posttranslational assay using purified precursors diluted out of 8 M urea to study translocation across yeast microsomal membranes. They found that this process could be divided into two sequential events in vitro,separable by the withholding of ATP, that were amenable to biochemical as well as genetic analysis. In this assay, the labeled precursor protein was diluted into buffer containing yeast microsomes in the absence of any cofactors. The membranes were pelleted by centrifugation through a sucrose cushion to remove any unbound precursor, and the pellet was resuspended and divided into two aliquots. From the first sample, the amount of specifically bound labeled precursor could be determined, and from the second, after addition of ATP and further incubation, the proportion of bound precursor that was translocated and processed to the mature form could be measured. A. Binding to the Membrane Prior to translocation, in the absence of ATP, both proOmpA and prepro-afactor bind with high affinity to specific receptors on the cytoplasmic face of the microsomal membranes. The binding is saturable, with a dissociation constant of 7.5 x 10- M, and can be competed for by other precursor proteins, but not by mature proteins, which suggests a possible involvement of signal sequences. It is strongly inhibited by predigestion of the membranes with papain, a process known to inhibit translocation. Moreover, proOmpA that had been incubated in buffer for 15 hr at O'C, and was consequently incompetent for translocation (Sanz and Meyer, 1988), was also unable to bind to the membranes. This indicates that conformations of the precursor is important in the binding step. Whether it is important for further steps in the translocation process remains to be determined. 8. Translocation Addition of ATP to the membrane aliquot that had bound precursor proteins allowed translocation and processing to the mature form. The translocation could be inhibited by pretreatment of the membranes with N-ethylmaleimide, urea, or papain. However, since only papain also inhibited binding of the precursor, translocation may be mediated by another protein or set of proteins distinct from that mediating binding. One of these must be an ATPase, since translocation requires the hydrolysis of ATP. Further anaIysis of these proteins, or other membrane components involved in

24

JOANNE CROWE AND DAVID 1. MEYER

translocation, is very difficult using conventional biochemical methods in vitro. Luminally disposed proteins, or those embedded in the lipid bilayer, would be difficult to modify chemically or enzymatically, and are probably inaccessible to antibodies. The study of yeast translocation mutants provides an alternate way to identify and characterize these components.

V.

ANALYSIS OF ptI7-A

TRANSLOCATION MUTANT

Translocation mutants in microbial systems can be selected by screening mutagenized cells for the incorrect localization of an artificial fusion protein (Beckwith and Ferro-Novick, 1986; Ito, 1986; Deshaies and Schekman, 1987). In wild-type yeast cells this protein is secreted, while in the mutant it accumulates in the cytoplasm. The temperature-sensitive protein translocation mutant ptll was isolated in this way by Toyn and co-workers (1988). They used a fusion of the prepro region of the MFcll gene to T R P l , a gene that encodes phosphoribosyl anthranilate (PRA) isomerase, a cytoplasmic enzyme essential for tryptophan biosynthesis. In a trpl strain this fusion protein is efficiently translocated out of the cytoplasm and growth is permitted only on tryptophan-supplemented media. A translocation mutant in this strain, however, will accumulate enough of the fusion protein in the cytosol to reverse the Trp- phenotype to Trp+ . The mutants can then be easily isolated by screening for growth in the absence of added tryptophan. The translocation phenotype of the ptll mutation, which has been localized to chromosome 15, has been examined by Toyn et al. (1988) both in vivo and in vitro.

A. In Viwo Analysis The mutant ptll exhibits a temperature-sensitive, pleiotropic accumulation of presecretory proteins in the cytoplasm. The behavior of three secretory proteins having different final destinations was examined in this mutant. These were prepro-a-factor, which is secreted into the medium, and the precursors of the vacuolar hydrolase carboxypeptidase Y (CPY) and a 33-kDa cell wall glycoprotein (Sanz et al., 1987). More than 95% of all three proteins remains unprocessed at the nonpermissive temperature of 37°C. Protease digestion after gentle lysis showed these unprocessed precursor forms to be accumulated in the cytoplasm. Pulse-chase experiments showed that the defect was at the level of entry into the ER, and that transport further along the secretory pathway was unaffected in this mutant. We have already referred to the ability of yeast to translocate full-length preproteins across microsomal membranes in vitro. It was unknown, however, if translocation could occur posttranslationally in vivo. The use of p t l l , which

3. PROTEIN TRANSLOCATION IN YEAST

25

effectively accumulates precursor proteins in the cytoplasm, provided a definitive way to answer this question. The mutant was pulse-labeled at the restrictive temperature of 37°C in order to establish a cytoplasmic accumulation of radiolabeled preproCPY. When the cells were shifted to the permissive temperature of 22"C, the labeled cytoplasmic form of CPY could be seen to chase into the glycosylated ER and Golgi forms. This indicates that in yeast, a protein at least as large as 50-60 kDa can be efficiently translocated posttranslationally across the ER in vivo.

6. In Who Analysis Studying a translocation mutant in a cell-free system allows the localization of the defect to either the membrane or the cytosolic compartment. Wild-type membranes, when combined with cytosol derived from either wild-type or mutant cells, translocate pure proOmpA when it is rapidly diluted from 8 M urea at both 22" and 37°C. When membranes from ptll are used, however, they show reduced translocation of proOmpA at 22"C, and are virtually translocation-incompetent when assayed at 37°C. This phenotype could be reversed simply by shifting the assay temperature back to 22"C, indicating that the defect is in the ptll membranes, and is both temperature-sensitive and reversible in vitro. When ptll membranes are analyzed using the two-step binding assay described previously, they were able both to bind and to translocate proOmpA at 22°C. However, while translocation was completely blocked when assayed at 37"C, binding of the preprotein to the mutant membranes was unaffected. Taken together these data indicate that the product of the PTLI gene is most likely a membrane component directly involved in translocation, but not preprotein binding. Isolation of the wild-type allele of this gene by complementation of the mutant phenotype, will permit the use of biochemical studies to characterize further the role played by its protein product, Ptllp, in the translocation process. Similar analyses of other mutants will hopefully lead us to an even deeper understanding of the whole translocation process. REFERENCES Addison, R. (1988). Translocation of a fragment of invertase across microsomal vesicles isolated from Neurospora crassa requires the hydrolysis of a nucleotide triphosphate. J . Biol. Chem. 263, 1 4281 - 14287. Beckwith, J., and Ferro-Novick, S . (1986). Genetic studies on protein export in bacteria. Cur. Top. Microbiol. Immunol. 125, 5-27. Blobel, G . , and Dobberstein, B. (1975). Transfer of proteins across membranes. 11. Reconstitution of functional rough microsomes from heterologous components. J . Cell Biol. 67, 852-862. Chirico, W. J., Waters, M. J., and Blobel, G. (1988). 70K heat shock related proteins stimulate protein translocation into microsomes. Nature (London) 332, 805-8 10. Collier, D. N., Bankaitis, V. A., Weiss, J-. B., and Bassford, P. J. (1988). The antifolding activity of SecB promotes the export of the E . coli maltose-binding protein. Cell 53, 273-283.

26

JOANNE CROWE AND DAVID I. MEYER

Crooke, E., and Wickner, W. (1987). Trigger factor: A soluble protein that folds proOmpA into a membrane-assembly competent form. Proc. Natl. Acad. Sci. U.S.A. 84, 5216-5220. Crooke, E., Guthrie, B., Lecker, S., Lill, R., and Wickner, W. (1988a). ProOmpA is stabilized for membrane translocation by either purified E. coli trigger factor or canine signal recognition particle. Cell 54, 1003-1011. Crooke, E., Brundage, L., Rice, M., and Wickner, W. (1988b). ProOmpA spontaneously folds into a membrane assembly competent state which trigger factor stabilizes. EMBO J. 7, 1831-1835. Deshaies, R.J., and Schekman, R. W. (1987). A yeast mutant defective at an early stage in import of secretory protein precursors into the endoplasmic reticulum. J. Cell B i d . 105, 633-645. Deshaies, R. J., Koch, B. D., Werner-Washburne, M., Craig, E. A , , and Schekman, R. W. (1988). 70kD stress protein homologues facilitate translocation of secretory and mitochondria1 precursor polypeptides. Nature (London) 332, 800-805. Eilers, M., and Schatz, G. (1986). Binding of a specific ligand inhibits import of a purified precursor protein into the mitochondria. Nature (London) 322, 228-232. Eilers, M., and Schatz, G. (1988). Protein unfolding and the energetics of protein translocation across biological membranes. Cell 52, 481-483. Eilers, M., Hwang, S., and Schatz, G. (1988). Unfolding and refolding of a purified precursor protein during import into isolated mitochondria. EMBO J. 7, 1139- 1145. Ellis, J. (1987). Proteins as molecular chaperones. Nature (London) 328, 378-379. Evans, E. A,, Gilmore, R., and Blobel, G. (1986). Purification of microsomal signal peptidase as a complex. Proc. Natl. Acad. Sci. U.S.A. 83, 581-585. Gilmore. R., Walter, P., and Blobel, G . (1982). Protein translocation across the endoplasmic reticulum. I. Detection in the microsomal membrane of a receptor for the signal recognition particle. J. Cell Biol. 95, 463-469. Hansen, W., Garcia, P., and Walter, P. (1986). In vitro protein translocation across the endoplasmic reticulum: ATP-dependent post-translational translocation of the prepro-a-factor. Cell 45, 397406. Hortsch, M., Avossa, D., and Meyer, D. I. (1986). Characterisation of secretory protein translocation: Ribosome-membrane interaction in the endoplasmic reticulum. J. Cell Biol. 103, 241253. Ito, K. (1986). Genetic control of protein secretion and localisation. Adv. Biophys. 21, 267-280. Meyer, D. I. (1988). Preprotein conformation: The year’s major theme in translocation studies. Trends Biochem. Sci. 13, 471-474. Meyer, D. I., Krause, E., and Dobberstein, B. (1982). Secretory protein translocation across membranes-the role of the “docking protein.” Nature (London) 297, 647-650. Miiller, G . , and Zimmerman, R. (1988). Import of honeybee prepromelittin into the endoplasmic reticulum: Energy requirements for membrane insertion. EMBO J. 7, 639-648. Murakami, H., Pain, D., and Blobel, G. (1988). 70-kD heat shock related protein is one of at least two distinct cytosolic factors stimulating protein import into mitochondria. J. Cell Biol. 107, 2051-2057. Pain, D., Kanwar, Y. S., and Blobel, G. (1988). Identification of a receptor for protein import into chloroplasts and its localization to envelope contact zones. Nuture (London) 331, 232-236. Park, S . , Liu, G . , Topping, T. B., Cover, W. H., and Randall, L. L. (1988). Modulation of the folding pathways of exported proteins by the leader sequence. Science 239, 1033-1035. Pelham, H. R. B. (1986). Speculations on the functions of the major heat shock and glucose-related proteins. Cell 46, 959-961. Pfanner, N., Hartl, F.-U., and Neupert, W. (1988). Import of proteins into mitochondria: A multistep process. Eur. J. Biochem. 175, 205-212. Randall, L. L., and Hardy, S. J. S. (1986). Correlation of competence for export with lack of tertiary

3. PROTEIN TRANSLOCATION IN YEAST

27

structure of the mature species: A study in vivo of maltose-binding protein in E. coli. Cell 46, 921-928. Rothblatt, J. A., and Meyer, D. I. (1986). Secretion in yeast: Reconstitution of the translocation and glycosylation of a-factor and invertase in a homologous cell-free system. Cell 44, 619-628. Rothblatt, J. A,, Webb, J. R., Ammerer, G., and Meyer, D. I. (1987). Secretion in yeast: Structural features influencing the post-translational translocation of prepro-a-factor in vifro. EMBO J. 6 , 3455-3464. Rothman, J. E., and Kornberg, R. D. (1986). An unfolding story of protein translocation. Nafure (London) 322, 209-210. Sanz, P., and Meyer, D. I. (1988). Signal recognition particle (SRP) stabilizes the translocationcompetent conformation of presecretory proteins. EMBO 1.7, 3553-3557. Sanz, P., and Meyer, D. I. (1989). Secretion in yeast: Preprotein binding to a membrane receptor and ATP-dependent translocation are sequential and separable events in vitro. J. Cell Biol. 108, 2101-2106. Sanz, P., Herrero, E., and Sentandreu, R. (1987). Secretory pattern of a major integral mannoprotein of the yeast cell wall. Biochim. Biophys. Actu 924, 93-103. Tajima, S., Lauffer, L., Rath, V. L., and Walter, P. (1986). The signal recognition particle receptor is a complex that contains two distinct polypeptide chains. J. Cell Biol. 103, 1167- 1178. Thorn, J. R., and Randall, L. L. (1988). Role of the leader peptide of maltose-binding protein in two steps of the export process. J. Bucteriol. 170, 5654-5661. Toyn, J., Hibbs, A. R., Sanz, P., Crowe, J., and Meyer, D. I. (1988). I n vivo and in vifro analysis of prll, a yeast rs mutant with a membrane-associated defect in protein translocation. EMBO J. 7, 4347-4353. Vestweber, D., and Schatz, D. (1988a). Mitochondria can import artificial precursor proteins containing a branched polypeptide chain or a carboxyl-terminal stilbene disulfonate. J. Cell Biol. 107, 2045-2049. Vestweber, D., and Schatz, G. (1988b). A chimeric mitochondrial precursor protein with internal disulfide bridges blocks import of authentic precursors into mitochondria and allows quantitation of import sites. J. Cell Biol. 107, 2037-2043. Walter, P., and Blobel, G. (1981a). Translocation of proteins across the endoplasmic reticulum. 11. Signal recognition protein (SRP) mediates the selective binding to microsomal membranes of in vitru-assembled polysomes synthesising secretory proteins. J. Cell Biol. 91, 55 1-556. Walter, P., and Blobel, G. (1981b). Translocation of proteins across the endoplasmic reticulum. 111. Signal recognition protein (SRP) causes signal sequence-dependent and site-specific arrest of chain elongation that is released by microsomal membranes. J . Cell Biol. 91, 557-561. Walter, P., Ibrahimi, I., and Blobel, G. (1981). Translocation of proteins across the endoplasmic reticulum. I. Signal recognition protein (SRP) binds to in virro assembled polysomes synthesizing secretory proteins. J . Cell Biol. 91, 545-550. Waters, M. G . , and Blobel, G. (1986). Secretory protein translocation in a yeast cell-free system can occur post-translationally and requires ATP hydrolysis. J . Cell Biol. 102, 1543-1550. Wiedmann, M., Kurzchalia, T. V.,Hartmann, E., and Rapoport, T. A. (1987). A signal sequence receptor in the endoplasmic reticulum membrane. Nature (London) 328, 830-833. Zimmermann, R., and Meyer, D. I. (1986). 1986: A year of new insights into protein translocation across membranes. Trends Btochem. Sci. 11, 512-515. Zimmermann, R., Sagstetter, M., Lewis, M. J., and Pelham, H. R. B. (1988). Seventy-kilodalton heat shock proteins and an additional component from reticulocyte lysate stimulate import of M13 procoat protein into microsomes. EMBO J . 7, 2875-2880.

This Page Intentionally Left Blank

Part II

Structure of Proteins in the Membrane

This Page Intentionally Left Blank

CURRENT TOPICS IN MEMBRANES AND TRANSFORT, VOLUME 36

Chapter 4

Architectural Editing: Regulating the Surface Expression of the Multicomponent T-cell Antigen Receptor RICHARD D . K U U S N E R , JENNIFER LIPPINCOTTSCHWARTZ, AND JUAN S . BONIFACINO Cell Biology and Metabolism Branch National Institute of Child Health and Human Developmen1 National Institutes of Health Bethesda, Maryland 20892

1. Introduction 11. Structure and Composition of the T-cell Antigen Receptor A. Clonotypic Subunits B. Nonpolymorphic Subunits 111. Subunit Interactions within the T-cell Receptor IV. Stoichiometry within the Complex V. Assembly of the T-cell Receptor Complex VI. Fate of Newly Synthesized T-cell Receptor Chains A. Getting to the Cell Surface B. Getting Newly Synthesized Chains Out of the Endoplasmic Reticulum VII. Summary References

1.

lNTRODUCTlON

As we learn more about the structure of membrane proteins, it is becoming clear that a growing number of proteins are complex assemblies of multiple chains. For many, these complexes are the result of the specific assembly of oligomers of identical subunits. These may be covalently linked, as in the case of the transferrin receptor (reviewed in Harford et al., 1990) or the insulin receptor (reviewed in Rosen, 1987) or, perhaps more commonly, represent tight non31

32

RICHARD D. KLAUSNER ET AL.

covalent associations as in a number of viral membrane glycoproteins. These include vesicular stomatitis virus (VSV) G protein (Kreis and Lodish, 1986), the influenza hemagglutinin (Wilson et al., 1981), and the Rous sarcoma virus (RSV) envelope glycoprotein (Einfeld and Hunter, 1988). Other membrane complexes, such as the T-cell antigen receptor (Samelson et al., 1985), the IgE receptor (Metzger et al., 1986), and surface major histocompatibility complex (MHC) molecules (reviewed in Kaufman et al., 1984), are formed by the association of two or more different subunits. How often what is considered to be a monomeric chain in fact exists as an oligomer is not clear. Apparently, the more investigators look for oligomers, the more they find them. One impediment to the identification of the oligomeric structure of integral membrane proteins is the tendency of oligomers to dissociate upon detergent solubilization when they are not covalently linked to each other. This can be avoided with techniques that maintain, as near as possible, the native oligomeric state of integral membrane proteins including the use of relatively nondenaturing, nonionic detergents, and/or the addition of lipids during solubilization to optimize lipid/detergent ratios. Over the past several years, as our knowledge of the number of membrane proteins that are a part of multichain complexes increases, a variety of observations has led to the realization that there is an intimate relationship between the intracellular assembly of these complexes and the expression on the surface of the cell. The ability of the cell to prevent the expression on the surface of incompletely assembled complexes implies the existence of an important underlying mechanism that can (1) recognize and distinguish assembled from unassembled complexes and ( 2 ) differentiate the fate of fully assembled complexes from either partially assembled, incorrectly assembled, or unassembled membrane proteins. We will refer to this overall cellular process as “architectural editing.” We use this term because our current understanding is that it is something about the structure of the complexes (i.e., their architecture) that the cell recognizes. The editing refers to the fact that once the architecture is recognized as being either correct or incorrect, the cell then can determine the subsequent fate of the complex. We have been studying the T-cell antigen receptor as a model for a multisubunit membrane complex, the intracellular fate of which is determined by its oligomeric structure. We now recognize that this receptor is an extremely complex hetero-oligomer. At least six different gene products encode the receptor, which is made up of at least seven transmembrane proteins (Samelson et al., 1985). The structural complexity of this receptor has provided us with the opportunity to examine the role of architectural editing in determining cell surface expression of fully assembled receptor complexes. Many of the observations that have been made with the T-cell antigen receptor have found support and/or precedent in simpler oligomeric structures, such as viral glycoproteins. In addi-

4. REGULATING EXPRESSION OF THE T-CELL RECEPTOR

33

tion, and perhaps because of the complexity of this receptor, it has enabled us to recognize a wider repertoire of editing possibilities than has been previously recognized. Before discussing how architectural editing determines the expression of correctly assembled cell surface complexes, it is necessary to introduce the structural components that make up this receptor.

II. STRUCTURE AND COMPOSITION OF THE T-CELL ANTIGEN RECEPTOR

A. Clonotypic Subunits The function of the T-cell antigen receptor is to recognize antigen and, through that recognition, to initiate the biochemical events that lead to T-cell activation. The T-cell antigen receptor is one of two types of antigen receptors that allow the immune system to recognize a myriad of different proteins. The other type, present on B cells, is surface immunoglobulin (Ig). Because of the ready availability of Ig in the serum as a soluble protein, the determination of Ig structure was one of the great advances of immunology (reviewed in Davies and Metzger, 1983; Amzel and Poljak, 1979). All Ig molecules have a characteristic domain structure referred to as the immunoglobulin or Ig domain. This domain is formed by the folding together of two sets of @pleated sheets. Immunoglobulins can be divided into several different types of Ig domains. The amino-terminal domains are called the variable regions, and the more carboxy-terminal domains are constant regions. This is a manifestation of the genetically determined diversity of the amino acid composition of the different domains. It is the variability of the amino acid composition of the variable domains that gives the immune system its astoundingly broad repertoire for the recognition of a very large number of different antigens. Any clone of B cells expresses on its surface an Ig molecule with a carboxy-terminal extension that allows it to cross the membrane and thereby be anchored as an integral membrane protein. The Ig receptor expressed on any particular clone of B cells has a unique amino acid composition in its variable region that gives it its antigen specificity. The recognition element of the T-cell antigen receptor is composed of two Iglike chains that have been referred to as Ti chains (see reviews in Davis and Bjorkman, 1988; Wilson et al., 1988; Allison and Lanier, 1987). As with the Bcell receptor, the T-cell antigen receptor is different for every clone of T cells. It has been calculated that the genetic mechanisms that give rise to the extraordinary diversity of Ti allows the T-cell population to recognize up to 10l8 determinants. There are two types of pairs of Ti chains that can provide the recognition components of T-cell antigen receptors. The most abundant and well studied are the Ti-a and -p chains. Subsequent to the recognition and discovery of the Ti-a@

34

RICHARD D. KLAUSNER ET AL.

heterodimers, two other Ig-like chains, Ti-y and -6 were described (reviewed in Brenneq er al., 1988). It is most likely that Ti-yS serves the function of recognizing antigen but its role has not been definitively determined. For the remainder of this chapter, we will be referring to the predominant and better studied T-cell receptor that contains Ti-a and -p chains. The a and p chains each have core protein molecular weights of -30,000. Multiple N-linked glycosylations raise their M , values to between 40,000 and 55,000. As with Ig, they contain both amino-terminal variable and carboxy-terminal constant domains. The a and p chains are linked to each other by a single disulfide. Each chain has a hydrophobic domain near the carboxy terminus that serves as the transmembrane domain. What is unusual about the transmembrane domains of the a and p chains is that they contain one or two positively charged groups in the region predicted to cross the lipid bilayer. Each of the chains has short charged cytoplasmic tails composed of 5-10 amino acid residues.

B. Nonpolymorphic Subunits Studies in the early 1980s pointed to the possibility of the physical association

on the surface of the T cell of a group of membrane proteins referred to as the CD3 complex. The connection between CD3 and the T-cell receptor was first suggested by the observation that monoclonal antibodies against one of the chains of CD3 were capable of either stimulating or blocking T-cell function (van Wauwe et al., 1980; Reinherz et al., 1980, 1982). In the intervening years, it has become clear that CD3 is, in fact, a collection of subunits of the T-cell receptor. The evidence that has led to this recognition includes (1) the ability to coimmunoprecipitate Ti chains using anti-CD3 antibodies and the ability to coprecipitate CD3 chains using anti-Ti antibodies (Reinherz er al., 1983; Borst er al., 1983a,b; Samelson et al., 1985); (2) the ability completely and stoichiometrically to comodulate CD3 and Ti from the surface of T cells using cross-linked antibody against either of the components (Reinherz et al., 1982; Meuer et al., 1983); (3) the finding that in variant or mutant T cells, failure to synthesize Ti chains resulted in the complete abrogation of surface expression of CD3 (Weiss and Stobo, 1984; Schmitt-Verhulst et al., 1987; Chen et al., 1988); and (4) the ability chemically to cross-link Ti to CD3 on the surface of intact T cells (Allison and Lanier, 1985; Brenner er al., 1985). For historical reasons, CD3 was thought of as one complex of membrane proteins fundamentally distinct from the T-cell receptor. Unfortunately, this view persists in the literature. We now know, however, that the functional T-cell receptor is made up of the clonotypic chains referred to as Ti and nonpolymorphic chains that include three CD3 components. This does not simply represent the association of two distinct sets of surface molecules, but rather the assembly of these components into a single macromolecular receptor complex.

4. REGULATING EXPRESSION OF THE T-CELL RECEPTOR

35

The CD3 complex was initially described biochemically in human T cells (Borst et al., 1983a,b; Kanellopoulos et al., 1983). It consists of three chains: y, 6, and E. The gene for each of these chains has been cloned so each chain is structurally well defined (van den Elsen et al., 1984; Krissansen et al., 1986; Gold et al., 1986). y and 6 are glycoproteins with core M, values of approximately 17,000-18,OOO. Glycosylation results in M, for 6 of -20,000 and for y of 26,OOO-28,000. The chain is a nonglycosylated 20-kDa protein. Each of these chains possesses relatively long intracellular and extracellular domains, the latter of which bear homology to members of the Ig gene superfamily. In addition, the three chains have significant sequence homology to each other. That they are structurally related is reflected in their genomic organization. In both human and mouse, the genes encoding these three chains are clustered within 30-300 kb of each other (Tunnacliffe et al., 1987; Clevers et al., 1988). The y and 6 genes are within 2 kb of each other and may, in fact, share transcriptional control elements (Georgopoulos et al., 1988). The murine CD3 chains are highly homologous to their human counterparts (van den Elsen et al., 1985; Krissansen et al., 1987; Gold et al., 1987; Haser et al., 1987). One distinction is that whereas the human y is more highly glycosylated than human 6, the opposite holds in murine T cells. In the latter, the M, of mature and completely processed 6 is 25,000-26,000, while the y chain has M,of approximately 20,000-21 ,OOO. In addition to the CD3 chains, there are two other nonpolymorphic chains that are part of the T-cell receptor complex. The 5 chain is a 16-kDa nonglycosylated transmembrane protein originally described in murine T cells (Samelson et al., 1985). Its structure is quite distinct from the other constant chains of the receptor (CD3). Whereas the CD3 chains contain large extracellular, Ig-like domains, 5 contains approximately eight amino acids for an extracellular domain (Weissman et al., 1988a). All of the nonpolymorphic chains are transmembrane proteins with single membrane-spanning regions (van den Elsen et al., 1984; Gold et al., 1986; Krissansen et al., 1986; Weissman el al., 1988a). Interestingly, all of them contain single negative-charge groups in their transmembrane regions. The 5 chain exists as a disulfide-linked homodimer in the vast majority of T-cell receptors. 6 bears no sequence or structural homology to the CD3 chains and, in fact, the gene is found on chromosome 1 in both mouse and human and is thus unlinked to the CD3 cluster (Weissman et al., 1988b; Baniyash et al., 1989). For these reasons, we do not consider the 5 chain part of the CD3 complex. We have subsequently described yet another chain of the T-cell receptor complex, which we have termed q (Baniyash et al., 1988; Orloff et al., 1989). AS of the time of this writing, the gene encoding has not been identified and thus we know significantly less about its structure than any of the other chains. The q chain is a 22-kDa protein that, like 5, is basic. Biochemical characterization has revealed that q is nonglycosylated and we have detected no posttranslational modifications of it (Orloff et al., 1989). It is immunologically related to 5 in that

36

RICHARD D. KLAUSNER ET AL.

certain antipeptide antibodies raised against peptides contained within 5 will directly recognize q. However, other anti-t; antibodies do not recognize q. In addition, amino acid composition studies have demonstrated that q and 5, despite their being immunologically related, are distinct proteins (Orloff et al., 1989). A peculiar characteristic of q is the fact that it exists as part of the T-cell receptor in a covalent linkage to a 5 monomer (Baniyash et al., 1988). Thus, even on cloned T cells, there seem to be two populations of receptors. A minority (-20% on many T cells) contain the (q heterodimer while the majority contain only the 55 homodimer. We do not know yet whether these two dimers are mutually exclusive or whether a single receptor complex can have both
111.

SUBUNIT INTERACTIONS WITHIN THE T-CELL RECEPTOR

We know very little about the overall structure of the T-cell receptor complex. It will undoubtedly be some time before we have a view of the three-dimensional structure of this extremely complicated integral membrane assembly. However, some insights are emerging from studies that attempt to look at the interactions that are allowable between the different components of this complex. These interactions give us at least an indirect picture of potential nearest neighbor relations within the final assembly. Four approaches have been used to gain insights into these relations: (1) studies with bifunctional cross-linking reagents (Brenner et al., 1985); (2) analysis of partial complexes observed by the selective dissociation of subunits during solubilization andlor immunoprecipitation (Bonifacino et al. , 1988a); (3) analysis of partial complexes observed in mutant or variant T cells (Bonifacino et al., 1988a); and (4) transfection of the genes and/or cDNAs encoding pairs or groups of subunits into non-T-cells (LippincottSchwartz et aE., 1988; Berkhout et al., 1988; Bonifacino et al., 1989). The first and most striking conclusion that derives from these studies is that stable partial complexes can be formed even in the absence of one or more of the remaining chains of the complex. The only information to be derived from cross-linking analysis comes from a study by Brenner and colleagues who demonstrated that

37

4. REGULATING EXPRESSION OF THE T-CELL RECEPTOR

rn 79-105 aa

25-27 aa

I aa

44-55 aa 13 aa

Human Chromosome 11 Mouse Chromosome 9

Human and Mouse Chromosome 1

FIG.1. Structure of the T-cell antigen receptor. The T-cell antigen receptor is composed of two clonotypic chains, referred to as Ti. In most T cells, these chains are the products of the cx and p gene loci. Like immunoglobulins, the Ti chains have N-terminal variable ( V )and C-terminal constant ( C ) domains. At least five nonpolymorphic chains are also part of the receptor complex. These are the 7 , 6, and E chains, referred to as CD3, and the 52 homodimer. Up to 20% of all surface receptors in a variety of T cells contain another invariant chain, called -q. In these receptors, the -q chain is disulfide-linked to a 5 monomer. The sizes of the extracellular, transmembrane, and cytoplasmic domains of the CD3 and 5 chains are shown. Symbol: (0-) N-linked oligosaccharide chain.

the human CD3-y chain could be specifically cross-linked to the Ti-@chain (Brenner et al., 1985). Unfortunately, no other neighbor relations have been defined by this type of procedure. We have attempted such studies but have found them to be unrevealing. In order to observe the entire complex as a unit, we and others have employed immunoprecipitation using antibodies directed at any of the subunits. Such studies reveal that all of the other subunits can be specifically coimmunoprecipitated by virtue of the ability to maintain the complex and its associations through solubilization and immunoprecipitation. This requires, at the minimum, the use of nondenaturing and preferably nonionic detergents. Although detergents such

38

RICHARD D. KLAUSNER ET AL.

as CHAPSwill allow one to observe the complex, Triton X-100 and digitonin are the most reliable. The complex cannot withstand even small amounts (>O. 1%) of ionic detergents such as deoxycholate or sodium dodecyl sulfate (SDS). If, however, immune precipitates are washed with progressively higher concentrations of denaturing ionic detergents, one observes a hierarchy of subunit dissociation. Using anti+ antibodies, for example, as one raises the concentration of ionic detergent, one sees the simultaneous loss of a,p, and 5 with the continued coprecipitation of the full CD3 complex. With increasing concentrations, one tends to lose 6 followed by y and finally the target antigen, in this case, the E chain (R. D. Klausner et al., unpublished observations). Using a particular anti-5 polyclonal antiserum, we have found that the complex can be dissociated such that the CD3 triplet is no longer coprecipitated but the Ti a p heterodimer is (Bonifacino et al., 1988a). This suggests that arc{ can exist as a partially stable complex, independent of CD3.

TABLE I PARTIAL T-CELLRECEPTOR COMPLEXES A. Murine T Cells

Complex Cell line

Characteristics

apy6e<2 2B4 ap6r<, 2B4 apt2 2B4 apy6e MA 5.8 a p 8 ~ MA 5.8 y6e 21.2.2 21.2.2 6c Ye BW5147

T-cell hybridoma T-cell hybridoma T-cell hybridoma c-Deficient 2B4 variant I-Deficient 2B4 variant p-Deficient 2B4 mutant p-Deficient 2B4 mutant <,&Deficient mouse thymoma

Antibody used to isolate the complex anti-a anti-6 anti-< anti-c anti-6 anti-• anti-6 anti-•

References Samelson et al. (1985) Bonifacino er al. (1988a) Bonifacino et a/. (1988a) Bonifacino et al. (1988a) Bonifacino er al. (1988a) Bonifacino et al. (1988a) Bonifacino et a / . (1988a) Bonifacino et a / . (1988a)

B. Transfected Fibroblasts

Complex

Cell line

Antibody used to isolate the complex

4

TT 5.8

Anti-a

Y6E 6E Ye

cos cos cos

Anti-6, anti-c Anti-6, anti-e Anti-c, anti-y

References Bonifacino et al. (1988a) Lippincott-Schwartz et al. (1988) Berkhout et al. (1988) Berkhout et al. (1988) Berkhout et al. (1988) Bonifacino et al. (1989)

4. REGULATING EXPRESSION OF THE T-CELL RECEPTOR

39

Studies using variants and mutants of T cells have allowed us to observe a wide range of partial complexes with stability, as determined by immunoprecipitation, that is comparable to that seen for the mature complete complex. In cells that fail to produce any 5, a complex containing a, f3, 7 , 6, and E is formed (Bonifacino et al., 1988a; Sussman et ul., 1988). In cells that fail to produce any p chain, one sees 5 dimerization as well as a stable association of y, 6, and E. However, in this situation, no free a chain can be coprecipitated with any other antibodies and the 5 dimer cannot be coprecipitated with the antibodies directed against CD3 components (Bonifacino et al., 1988a; Chen et al., 1988). When only y and E are produced, one sees stable complexes of these two chains (Bonifacino et al., 1988a). Reconstitution studies using transfected genes andlor cDNAs encoding different T-cell receptor subunits have likewise demonstrated that partial complexes are stable. These include the stable CD3 complex containing 6, E, and y (Berkhout er al., 1988), a stable complex containing only E and y (Berkhout et al., 1988; Bonifacino et al., 1989), and the formation of disulfidelinked af3 heterodimers in the absence of any other chains of the T-cell receptor (Bonifacino et al., 1988a). Transfecting fibroblasts with the cDNA encoding 5 results in the rapid and efficient dimerization of 5 to produce the 55 homodimer (Orloff et al., 1989). Thus, both types of covalently linked disulfide dimers observed in the receptor can be formed in the absence of any other T-cell receptor component or, for that matter, any other T-cell specific protein. A list of partial complexes is shown in Table I.

IV. STOICHIOMETRY WITHIN THE COMPLEX We do not know what the stoichiometry of subunits is within the T-cell receptor complex. Attempts have been made to address this question by steady-state metabolic labeling using a variety of amino acids, and comparing the relative intensities of all of the precipitated assembled bands, corrected for the known amino acid composition of the proteins. These studies are compatible with a ratio of 1 : 1 : 1 : 1 : 1 : 2 for a,p, y, 6 , E, and 5 (R.D. Klausner et al., unpublished observations). However, no studies have addressed the actual stoichiometry of subunits that makes up a functional receptor. Thus we do not know, even if the relative stoichiometry proves to be what the metabolic labeling studies point to, whether the functional receptor is a dimer, trimer, or higher oligomer of this basic set of subunits. One interesting observation derives from the study of several different &deficient cell lines. In these cells, we see the assembly of partial complexes that contain several of the subunits except 6 (Bonifacino et al., 1989, and unpublished observations). However, when one examines the relative intensity of different subunits determined either by metabolic labeling or by im-

40

RICHARD D. KLAUSNER ET AL.

munoblotting in assembled complexes, one sees a relative increase in the amount of y, as if it has replaced the deficient 6. It is intriguing to find this, in light of the high level of structural homology between y and 6. Despite the apparent replacement structurally of 6 with y, as we will see later, the cell is not “fooled” and can distinguish complexes containing 6 from complexes that lack 6.

V. ASSEMBLY OF THE T-CELL RECEPTOR COMPLEX Biosynthetic pulse-chase studies demonstrate that the newly synthesized Tcell receptor chains begin to assemble almost immediately after, or perhaps even during, their synthesis in the endoplasmic reticulum (ER). It is important to point out that in discussing the assembly of the T-cell receptor, we are using the term operationally. We observe assembly by virtue of coimmunoprecipitation of different subunits. Thus we can measure the kinetics and the pattern of development of complexes whose stability must allow them to withstand the rigors of solubilization and immunoprecipitation. Using this definition, it is clear that when we see coimmunoprecipitation, we have seen assembly. However, the opposite is not true. Thus, the failure to see coprecipitation does not mean that the chains are not assembled, but simply that they may not be stable enough to withstand the analysis. The most rapid event seen in the assembly of this complex is the dimerization of the chain. This is over within minutes of biosynthesis (Orloff et a l . , 1989). Although assembly begins very rapidly after synthesis, it takes -30 min, in T-cell hybridoma cell lines, for assembly to be completed. The CD3 chains seem to assemble more rapidly than the full complex. Some studies suggest that CD3-y first interacts with C D ~ - E followed , by the addition of CD3-6. However, all of these events overlap, and at this point it is difficult to be very precise about the order of addition of chains. It is, in fact, possible that there is no absolutely defined order of addition but that the chains can come together in a variety of orders. We can observe two distinct steps in the formation of the mature stable receptor complex. The first of these is what we refer to under the rubric of assembly: the coming together of the chains. The second process we can refer to as maturation of the structure. By this we do not mean carbohydrate processing, but rather structural changes in the assembled complex that take place after assembly. The most clear-cut of these processes is the disulfide linkage of the 01 and p chains. We have observed that this is a relatively slow process, some of which takes place after the complex has reached the Golgi apparatus. Our studies have demonstrated that we do not get cx stably assembled with the rest of the complex if the p chain is missing. However, during the course of assembly of the complete complex, there is a relatively prolonged period

<

4. REGULATING EXPRESSION OF THE T-CELL RECEPTOR

41

where one observes the precipitation of non-disulfide-linked cwp chains with antibodies directed against the nonpolymorphic subunits of the receptor. The second aspect of maturation involves a more vaguely defined conformational maturation. This is manifested by the observation that certain antibodies directed against the receptor will result in the dissociation of subunits for a period of time after those subunits have assembled (Bonifacino er al., 1988a). The way this is determined is by finding a disparity between the apparent state of assembly when using different antireceptor antibodies. For example, if newly assembled receptor complexes are immunoprecipitated with an anti-a-chain clonotypic monoclonal antibody, one can see that no CD3-y is recovered in the immunoprecipitate. This is not true if molecules against other subunits are used. It takes approximately 1-2 hr after biosynthesis for the y chain to withstand the conformationally caused stress by the use of the anti-a antibody (D. Antusch, unpublished observations). All of our current data points to the ER as the site of assembly. This is in contrast to these maturational events which can extend into and through the Golgi apparatus. All of the actual assembly occurs before the chains reach the cis-Golgi as determined by carbohydrate processing (Bonifacino et al., 1988a). In addition, a variety of manipulations that block the egress of the chains out of the ER (or at least to the cis-Golgi) have absolutely no effect on the extent or kinetics of assembly (Alarcon er al., 1988; J. Lippincott-Schwartz, unpublished observations). The converse is that the failure to assemble correctly or completely results in the failure to leave the ER and will be discussed later. In addition, several lines of evidence suggest that the assembly occurs among subunits derived from a relatively tight biosynthetic cohort. In other words, proteins synthesized at around the same time are capable of assembling with each other and there is very little mixing and assembly with proteins that have been synthesized at an earlier point. A great deal of recent interest has focused on the role of accessory proteins within the ER during the assembly of multichain or oligomeric complexes. These proteins can be conceptually divided as having two types of functions. One, exemplified by Ig heavy chain-binding protein, or BiP, appears to stabilize andlor “solubilize” unassembled or misfolded subunits (Pelham, 1986). Such proteins may allow the attainment of the correct conformation so that assembly and/or oligomerization can take place. The second group of proteins may more directly or specifically catalyze the assembly or transport of oligomeric structures. A protein that may perform some or all of these functions has been observed in both human and murine T cells in studies of biosynthesis and assembly of the T-cell receptor complex. This protein was first observed in human T cells and is referred to as the T3-p28 or CD3-w protein by Terhorst and colleagues (Pettey er al., 1987; Alarcon et d., 1988). We have studied what appears

42

RICHARD D. KLAUSNER ET AL.

to be the homologous protein in murine T cells (Bonifacino et al., 1988b). We have termed this protein TRAP for T-cell receptor-associated protein. Although what we know about it suggests that it functions during the early stages of assembly, we have yet to be able to define a precise role for TRAP. TRAP is a M, 26,000 basic, nonglycosylated protein. It can be observed by pulse-labeling cells and immunoprecipitation using antibodies directed against any of the CD3 components. It appears to be most tightly associated with CD3-E and is best seen using any of a number of antibodies directed against this chain. We do not observe the coprecipitation of TRAP using antibodies against a@ or 5. Perhaps its most intriguing characteristic is the transient nature of its association with the CD3 complex. We have not been able to observe the kinetics of association of TRAP with CD3-E. The maximum amount of TRAP that one sees associated with E is present at the end of the earliest pulse during biosynthetic labeling studies. TRAP remains associated for 10 min and then begins to be lost. What is the nature of this loss? In the absence of direct reagents to look at TRAP, we do not know what the fate of TRAP is once it is no longer coprecipitated along with the CD3 chains. It may simply dissociate or it may be degraded within the ER. Recent evidence from our laboratory suggests that TRAP may be proteolytically clipped to a M, 16,000 neutral protein that can still be precipitated with anti+ antibodies (Antusch et al., 1990). This species appears transiently and then is also lost. This may suggest that TRAP is being degraded within the ER and that we are observing a relatively stable proteolytic intermediate in this process. Studies from our laboratory using a variety of pharmacological manipulations of cells have examined some of the characteristics of the loss of TRAP (Bonifacino et al., 1988b; Antusch et al., 1990). First of all, it is clear that TRAP is lost within the ER system. Dissociation of TRAP precedes cis-Golgi processing of the T-cell receptor chains and occurs to the same extent and with the same kinetics in the absence of any movement to the Golgi (Bonifacino et al., 1988b). The loss of TRAP is exquisitely temperature-sensitive and fails to occur at temperatures <25"C. In addition, TRAP dissociation can be inhibited by a variety of agents that would be predicted to affect acidic compartments within the cell including chloroquine, ammonium chloride, monensin, and CCCP (Bonifacino et al., 1988b). Finally, the fungal antibacterial reagent, Brefeldin A (BFA), which totally blocks movement from the ER to the Golgi system and results in a dramatic alteration in the components of the ER due to the recycling of cis and medial Golgi components to this organelle (Lippincott-Schwartz et al., 1989), completely blocks the dissociation of TRAP. In the presence of this drug, TRAP remains associated with newly synthesized chains for hours and no cleavage of TRAP is seen (Antusch et al., 1990). Despite the inability to detect any significant dissociation of TRAP from CD3, this drug does not block assembly of the CD3 chains with a@ and 6 (J. Lippincott-Schwartz, unpublished observations).

-

4. REGULATING EXPRESSION OF THE T-CELL RECEPTOR

VI.

43

FATE OF NEWLY SYNTHESIZED T-CELL RECEPTOR CHAINS

A. Getting to the Cell Surface Several years ago we began studying the fate of newly synthesized T-cell receptor chains using a murine T-cell hybridoma. The most striking finding that emerged from these studies was that, in these cells, the vast majority of newly synthesized T-cell receptor chains were degraded within a few hours after synthesis (Minami et al., 1987). Pulse-chase studies followed by specific immune precipitation of the receptor complex in these cells demonstrated that two populations of newly synthesized T-cell receptor chains could be identified: (1) shortlived proteins (T 1/2 -2 hr) and (2) long-lived components (T 1/2 >15 hr). Between 80 and 95% of newly synthesized a,yP, y, 6, and E chains fell into the first category. The remainder were long-lived. In contrast, between 80 and 100% of the t; chains were long-lived. The difference between t; and the rest of the chains was mirrored in the rate of biosynthesis. 5 was synthesized at approximately one-tenth the rate of the other chains. This was reflected in an approximately 8- to 10-fold lower level of steady-state mRNA for 5 than for the other chains (D. Orloff, unpublished observations). The loss of the rapidly degraded chains was characterized by a delay of 1-2 hr from the time of synthesis before degradation was observed. During the early phase of this lag period, assembly of the receptor into a stable complex occurred. Following this lag, degradation occurred with a T 1/2 of approximately 30-60 min. The receptor chains that survive in the long term could be differentiated from chains with short half-lives by examining the state of assembly of the chains. Thus, when we examined the fate of a and/or the ayP heterodimer using antibodies directed against the a chain, we observed both the rapidly turning over population (-90%) and the chains destined to survive over the long term. If, on the other hand, we looked at the survival of a (and ayP heterodimers) during the pulse-chase by immunoprecipitating, not with an anti-a antibody, but rather, with an anti+ or anti-b antibody, we only precipitated the 10% of ayP that were going to survive over the long term. Similarly, if one looked at the fate of E with an anti+ antibody, again we observed both the majority of short-lived E chains as well as the long-lived E chains. When newly synthesized E chains were immunoprecipitated with an anti-a monoclonal antibody, now only the 10% of E that survived long term was precipitated. In other words, it appeared that only those complexes that assembled (as determined by stable coimmunoprecipitation)survived long term. Not surprisingly, considering that all of t; survived long term, we found that all of the t; chain assembled rapidly with the rest of the chains. Predictably, when anti-t; antibodies were used to coprecipitate non-t; subunits, and these were followed in a pulse-chase experiment, all of those chains were in the fraction that survived long term.

44

RICHARD D. KLAUSNER ET AL.

We can summarize these results quite simply. Assembly occurs in the ER. Those chains that assemble into complete receptor complexes will survive long term. The remainder of the chains that fail to form complete complexes are rapidly degraded. Because 5 is the limiting chain in these cells, all 5 assembles and immunoprecipitation with 5 quantitatively predicts the entire stable population of receptor subunits. Where are these short-lived chains degraded? When the fates of the newly synthesized glycoproteins are followed, we can readily recognize that degradation occurs after passage of the newly synthesized chains through the Golgi complex (Minami et al., 1987; Lippincott-Schwartz et al., 1988). This was determined by following the carbohydrate processing of these subunits. For example, when the a ( a p heterodimer) chain was followed in the pulse-chase experiment, virtually all of the a chain became resistant to cleavage with endoglycosidase H with -30 min of synthesis. This was followed by sialation of the entire cohort of the newly synthesized chains (Lippincott-Schwartz et al., 1988). From the time of biosynthesis it takes approximately li-2 hr to complete the carbohydrate processing of the newly synthesized chains. During that time there is little, if any, loss in the amount of newly synthesized chains in the cohort being followed, Thus, there is a lag of between 1 and 2 hr before any degradation takes place. After this lag phase, however, there is a rapid degradation of the newly synthesized excess chains, with a half-life during that time of approximately 3040 min. This degradation follows a single exponential, and by 4 hr only the fraction that will survive long term remains in the cell. Thus, the degradation is preceded by transport out of the ER through the cis, medial, and trans-Golgi. During the time of processing and transport through the Golgi, no significant degradation is taking place. This pattern is consistent with transport to lysosomes. Further evidence was sought to implicate this organelle as the site of degradation. Drugs that are known to inhibit lysosomal degradation were all found to be effective at inhibiting the degradation of these chains (Minami et al., 1987; Lippincott-Schwartz et al., 1988; Chen et al., 1988). These include weak bases such as ammonium chloride and chloroquine, methionine methyl ester, and lysosomal protease inhibitors such as leupeptin and EM. Percoll density gradient studies to assess subcellular fractionation produced results that were consistent with movement from these chains to dense lysosomes before they were degraded (R. D. Klausner et al., unpublished observations). Finally, fluorescence and electron-microscopic studies have demonstrated that the lysosomes represent the major site of labeling when antibodies against rapidly degraded T-cell receptor chains are used for immunomicroscopic localization (Chen et al., 1988; Lippincott-Schwartz et al., 1989; Bonifacino et al., 1989). As stated earlier, five of the chains in the T-cell hybridomas studied were made in excess and shared the fate of being transported out of the ER through Golgi to lysosomes. The limiting 5 chain was not degraded. One prediction of these

45

4. REGULATING EXPRESSION OF THE T-CELL RECEPTOR

observations was that if we were to study a T cell that lacked all <,essentially all of the chains would be transferred to lysosomes. We in fact have isolated and characterized several cell lines that fail to synthesize the chain. For reasons we do not currently understand, in T-cell hybridomas in culture, <-chain expression is readily lost. One particular cell line, termed MA5.8, had normal levels of synthesis of a, p, y, 6, and E but no detectable synthesis of (Sussman et al., 1988). was not present either at the level of protein or at the level of mRNA (Sussman et al., 1988; Weissman et al., 1988a). Despite normal levels of synthesis and assembly of the remaining five chains, these cells expressed only 35% of the level of surface receptor found on the parental cell type. When the fate of the newly synthesized chains in these cells were followed, it was apparent that virtually all of the chains moved out of the ER and through the Golgi, as determined by carbohydrate analysis in a manner that is indistinguishable from the parental 2B4 cell (Sussman et al., 1988). However, in contrast to 2B4, there is approximately a 20-fold drop in the fraction of a or 6 that survives long term down from -10% to <0.5% of newly synthesized chains (Sussman et al., 1988). The rest of these chains follow the same route as the degraded chains in the parental cell types. When these cells are treated with ammonium chloride, for example, the vast majority of the degradation is inhibited and now the survival of the chains in 2B4 and MA5.8 are indistinguishable (Sussman et al., 1988). We have isolated the cDNA encoding the chain (Weissman et al., 1988b). This cDNA has been inserted into an expression vector and the MA5.8 cells have been used as the target for stable transfections. As predicted, the reexpression of 5 allows the reexpression of full surface T-cell receptor complexes (Weissman et al., 1989). In addition, we have manipulated the level of expression of the transfected chain and this results in a proportional change in the fractional survival and fractional surface expression of the remaining chains. These studies confirm the original observations that point to the chain as the critical determinant of survival of the rest of the chains, allowing the cell to make the decision between lysosomal degradation and stable surface expression.

<

<

<

<

<

<

6. Getting Newly Synthesized Chains Out of the Endoplasmic Reticulum As was described elsewhere in this chapter, studies with viral integral membrane glycoproteins have pointed to the role of homo-oligomerization as a prerequisite for newly synthesized chains moving beyond the ER (Copeland et al., 1986; Gething et al., 1986; Kreis and Lodish, 1986). An analogous phenomenon is seen with this hetero-oligomeric complex. As we discussed before, the absence of the chain seems to have no effect on the ability of the rest of the chains to assemble and to move out of the ER (Sussman et al., 1988). In contrast, it appears that the lack of any other chain of the T-cell antigen receptor results in

<

46

RICHARD D. KLAUSNER ET AL.

failure of all of the remaining chains to leave the ER. This can be observed by following carbohydrate processing, subcellular fractionation, and immunoelectron and fluorescence microscopy. Studies using variant T cells, mutants of a particular T-cell hybridoma, and by transfecting combinations of T-cell receptor genes into fibroblasts have provided numerous examples of partial complexes, all of which fail to leave the ER (Lippincott-Schwartz et al., 1988, 1989; Chen et al., 1988; Bonifacino et al., 1989). Examples of this include the ap heterodimer, which dimerizes normally in the ER in the absence of any other chains but fails to leave that organelle (Bonifacino et al., 1988a; Lippincott-Schwartz et al., 1988). Other combinations include YE, 6yc, and others (Chen et al., 1988; Berkhout et al. , 1988; Bonifacino et al., 1989). Despite retention of these chains in the ER, we have failed to observe the assembly of these chains with any other component of the ER to explain this retention. What is the fate of T-cell receptor subunits that fail to reach the cis-Golgi? We have noted that there are two possible fates within this compartment: (1) some of the chains are retained for long periods within the ER (Chen et al., 1988; Bonifacino et al., 1989) and (2) some of the subunits are subject to rapid degradation (Lippincott-Schwartz et al., 1988; Chen et al., 1988; Bonifacino et al., 1989). In studies examining a variety of partial complexes that cannot negotiate ER-to-Golgi transport, the a, p, and 6 chains are seen to be rapidly degraded (Lippincott-Schwartz et al., 1988; Chen et al., 1988; Bonifacino et al., 1989). In contrast, the <, E, and y chains are stable for many hours within the ER (Chen er al., 1988; Bonifacino et al., 1989). All of our data point to either the ER or a subcompartment of this organelle as the site of this degradation. The pharmacology of degradation within the ER distinguishes it readily from lysosomal degradation. Thus, agents that inhibit lysosomal proteolysis all fail to affect ER degradation (Lippincott-Schwartz et al., 1988; Chen et al., 1988; Bonifacino et al., 1989). Two major questions confront us about the phenomenon of ER degradation. (1) What are the characteristics of a protein that determine whether or not it is subject to this degradation? (2) Where does this degradation actually take place? The answer to either of these questions will likely illuminate the answer to the other. Thus, if, in order to degrade proteins within the ER, the target proteins must be sorted to a degradative region of that organelle, then it is the signal for that sorting that will determine which chains are susceptible to this process. If, on the other hand, the degradation occurs throughout the ER, then it is most likely specific structural characteristics of a protein that determines whether or not it is a substrate for the ER proteases. One physical phenomenon that can influence whether or not a chain is degraded within the ER system is its state of assembly. We were initially surprised to observe the rapid degradation of the CD3-6 chain, while the CD3-y chain remained stable in the ER (Chen et al., 1988). This could be observed in cells that were expressing the y, 6, and E

4. REGULATING EXPRESSION OF THE T-CELL RECEPTOR

47

chains. Careful analysis of the fate of these chains within the EK revealed that the S chain apparently dissociated from the newly assembled complex in the absence of movement to the cis-Golgi. It appears that this disassembly is a prerequisite for the selective degradation of 6 (J. Lippincott-Schwartz, unpublished observations). In contrast, y remains stably associated with E. If, on the other hand, individual CD3 chains are transfected into fibroblasts, a different pattern of susceptibility to ER degradation is observed. As in the T cells, 6 is rapidly degraded. However, unlike T cells that express the E chain, when the y chain alone is expressed in fibroblasts, it too is rapidly degraded within the ER (Bonifacino et al., 1989). If, instead, E and y are cotransfected and simultaneously expressed in fibroblasts, they will assemble and neither the E nor the y chain is degraded (Bonifacino et al., 1989). When the E chain alone is expressed in fibroblasts, it is not subject to rapid ER degradation (Bonifacino et al., 1989). Thus it appears that some chains are intrinsically stable and are not targeted for ER degradation. If other subunits remain associated with the stable chains, then they too will survive for a long period of time within the ER. Subunits that cannot maintain a stable association with long-lived subunits are then subject to ER degradation.

VII. SUMMARY The assembly of multicomponent membrane complexes poses an interesting set of problems in cell biology. In addition, these complexes are also providing us a window into fundamental control pathways within the organellar system of the cell. From the time of the origin of components of multicomponent complexes in the EK, numerous quality control steps are imposed upon their maturation and their intracellular transport. We have chosen to refer to the general process whereby the cell ensures the expression of correct complexes in their appropriate ,subcellular localizations as “architectural editing.” The seven-chain T-cell antigen receptor complex has proved to be an excellent model system for exploring the characteristics of architectural editing. By examining the intracellular fate of partial or incomplete T-cell receptor complexes, we can observe the outline of four potential fates. Only complete and presumably correctly assembled complexes make it through the cell to the plasma membrane. Partial complexes are subject to a number of possible alternative fates. Certain near-complete complexes can make it from the ER through the Golgi system, at which point they are efficiently transferred to lysosomes and rapidly and completely degraded. More incomplete complexes remain within the ER system. Complexes that fail to be transported out of the ER apparently disassemble. Within the ER system, individual components may follow two fates. Certain components are stable within

48

RICHARD D. KLAUSNER ET AL.

this organellar system and remain for relatively long periods of time. Other components are subject to rapid degradation within this organelle. The T-cell receptor complex and the fate of its components have uncovered two pathways for membrane protein degradation, each of which takes place in a unique organellar system: the classical lysosome and the ER degradative pathway. Many aspects of architectural editing remain to be elucidated. The studies reported in this chapter on the T-cell antigen receptor from our laboratory represent the bare outlines of architectural editing afforded by this one system. Perhaps the most pressing area of ignorance concerns the unanswered questions about the signals that the cell uses to determine where proteins can or cannot go within the secretory pathway, We suspect, however, that in defining the possible routes of this editing process, we will at least formulate more exactly the questions for which molecular answers must be sought. REFERENCES Alarcon, B., Berkhout, B., Breitmeyer, J., and Terhorst, C. (1988). Assembly of the human T cell receptor-CD3 complex takes place in the endoplasmic reticulum and involves intermediary complexes between the CD3-gamma, delta, epsilon core and single T cell receptor alpha or beta chains. J. Eiol. Chem. 263, 2953-2961. Allison, J. P., and Lanier, L. L. (1985). Identification of antigen receptor associated structures on murine T cells. Nature (London) 314, 107-109. Allison, J. P., and Lanier, L. L. (1987). Structure, function and serology of the T-cell antigen receptor complex. Annu. Rev. Immunol. 5 , 503-540. Amzel, L. M., and Poljak, R. J. (1979). Three dimensional structure of immunoglobulins. Annu. Rev. Biochem. 48, 961-997. Antusch, D., Bonifacino, J. S., Burgess, W. H., and Klausner, R. D. (1990). The T-cell receptor associated protein (TRAP) is proteolytically cleaved in a pre-Golgi compartment. Submitted for publication. Baniyash, M., Garcia-Morales, P., Bonifacino, J. S., Samelson, L. E., and Klausner, R. D. (1988). Disulfide linkage of the zeta and eta chains of the T cell receptor. Possible identification of two structural classes of receptors. J. Eiol. Chem. 263, 9874-9878. Baniyash, M., Hsu, V. W., Seldin, M. F., and Klausner, R. D. (1989). The isolation and characterization of the murine T cell antigen receptor zeta chain gene. J . Biol. Chem. 204, 1325213257. Berkhout, B., Alarcon, B., and Ternorst, C. (1988). Transfection of genes encoding the T cell receptor-associated CD3 complex into COS cells results in assembly of the macromolecular structure. J . Eiol. Chem. 263, 8528-8536. Bonifacino, J. S., Chen, C., Lippincott-Schwartz, J., Ashwell, J. D., and Klausner, R. D. (1988a). Subunit interactions within the T cell antigen receptor: Clues from the study of partial complexes. Proc. Natl. Acad. Sci. U.S.A. 85, 6929-6933. Bonifacino, J. S . , Lippincott-Schwartz, J., Chen, C., Antusch, D., Samelson, L. E., and Klausner, R. D. (1988b). Association and dissociation of the murine T cell receptor associated protein (TRAP): Early events in the biosynthesis of a multisubunit receptor. J. Biol. Chem. 263, 89658971. Bonifacino, J. S., Suzuki, C. K., Lippincott-Schwartz, J., Weissman, A. M., and Klausner, R. D.

4. REGULATING EXPRESSION OF THE T-CELL RECEPTOR

49

(1989). Pre-Golgi degradation of newly synthesized T-cell antigen receptor chains: Intrinsic sensitivity and the role of subunit assembly. J. Cell. Biol. 109, 73-83. Borst, J., Alexander, S., Elder, J., and Terhorst, C. (1983a). The T3 complex on human T lymphocytes involves four structurally distinct glycoproteins. J. Biol. Chem. 258, 5135-5141. Borst, J., Prendiville, M. A., and Terhorst, C. (1983b). The T3 complex on human thymus-derived lymphocytes contains two different subunits of 20 kDa. Eur. J. Immunol. 13, 576-580. Brenner, M. B., Trowbridge, I. S., and Strominger, J. L. (1985). Cross-linking of human T cell receptor proteins: Association between the T cell idiotype beta subunit and the T3 glycoprotein heavy subunit. Cell 40, 183-190. Brenner, M. B., Strominger, J. L., and Krangel, M. S. (1988). The gamma-delta T cell receptor. Adv. Immunol. 43, 133-192. Chen, C., Bonifacino, J. S., Yuan, L., and Klausner, R. D. (1988). Selective degradation of T cell antigen receptor chains retained in a pre-Golgi compartment. J. Cell Biol. 107, 2149-2161. Clevers, H. C., Dunlap, S., Wileman, T. E., and Terhorst, C. (1988). Human CD3-Egene contains three miniexons and is transcribed from a non-TATA promoter. Proc. Natl. Acad. Sci. U.S.A. 85, 8156-8160. Copeland, C. A., Doms, R. W., Bolzau, E., Webster, R. G., and Helenius, A. (1986). Assembly of influenza hemagglutinin trimers and its role in intracellular transport. J. Cell Biol. 103, 11791191. Davies, D. R., and Metzger, H. (1983). Structural basis of antibody function. Annu. Rev. Immunol. 1, 87-117. Davis, M. M., and Bjorkman, P. J. (1988). T-cell antigen receptor genes and T-cell recognition. Nature {London) 334, 395-402. Einfeld, D., and Hunter, E. (1988). Oligomeric structure of a prototype retrovirus glycoprotein. Proc. Natl. Acad. Sci. U.S.A. 85, 8688-8692. Georgopoulos, K., van den Elsen, P., Bier, E., Maxam, A,, and Terhorst, C. (1988). A T-cellspecific enhancer is located in a DNase I-hypersensitive area at the B1 end of the CD3-6 gene. EMBO J. 7, 2401-2407. Gething, M. J., McCammon, K., and Sambrook, J. (1986). Expression of wild-type and mutant forms of influenza hemagglutinin: The role of folding in intracellular transport. Cell 46, 939950. Gold, D. P., Puck, J. M., Pettey, C. L., Cho, J., Coligan, J., Woody, J. N., and Terhorst, C. (1986). Isolation of cDNA clones encoding the 20 K non-glycosylated chain of the human T cell receptor~T3complex. Nature (London) 321, 43 1-434. Gold, D. P., Clevers, H., Alarcon, B., Dunlap, S., Novotny, J., Williams, A. F., and Terhorst, C. (1987). Evolutionary relationship between the T3 chains of the T-cell receptor complex and the immunoglobulin supergene family. Proc. Natl. Acad. Sci. U.S.A. 84, 7649-7653. Harford, J. B., Casey, J. L., Koeller, D. M., and Klausner, R. D. (1990). Structure, function and regulation of the transfemn receptor: Insights from molecular biology. In “IntracellularTrafficking of Proteins” (C. J. Steer and J. A. Hanover, eds.). Cambridge Univ. Press, London and New York, in press. Haser, W. G., Saito, H., Koyoma, T., and Tonegawa, S. (1987). Cloning and sequencing of murine T3 gamma cDNA from a subtractive cDNA library. J. Exp. Med. 166, 1186-1 191. Kanellopoulos, J. M., Wigglesworth, N. M., Owen, M. J., and Crumpton, M. J. (1983). Biosynthesis and molecular nature of the T3 Antigen of human T lymphocytes. EMBO J. 2, 18071814. Kaufman, J. E., Aufiay, C., Korman, A. J., Shackelford, D. A., and Strominger, J. (1984). The class I1 molecules of the human and murine major histocompatibility complex. Cell 36, 1-13. Kreis, T., and Lodish, H. F. (1986). Oligomerizationis essential for transport of vesicular stomatitis viral glycoprotein to the cell surface. Cell 46,929-937.

50

RICHARD D. KLAUSNER ET AL.

Krissansen, G. W., Owen, M. J., Verbi, W., and Crumpton, M. J. (1986). Primary structure of the T3-gamma subunit of the T3/T cell antigen receptor complex deduced from cDNA sequences: Evolution of the T3-gamma and delta subunits. EMBO J . 5, 1799-1808. Krissansen, G. W., Owen, M. J., Fink, P. J., and Crumpton, M. J. (1987). Molecular cloning of the cDNA encoding the T3-gamma subunit of the mouse T3/T cell antigen receptor complex. J . Immunol. 138, 3513-3518. Lippincott-Schwartz, J., Bonifacino, J. S., Yuan, L., and Klausner, R. D. (1988). Degradation from the endoplasmic reticulum: Disposing of newly synthesized proteins. Cell 54, 209-229. Lippincott-Schwartz, J., Yuan, L. C., Bonifacino, J. S., and Klausner, R. D. (1989). Rapid redistribution of Golgi proteins into the endoplasmic reticulum in cells treated with Brefeldin A: Evidence for membrane cycling from Golgi to ER. Cell 56, 801-813. Metzger, H., Alcaraz, G., Hohman, R., Kinet, J.-P., Pribluda, V.. and Quarto, R. (1986). The receptor with high affinity for immunoglobulin E. Annu. Rev. Immunol. 4, 419-470. Meuer, S . , Fitzgerald, K., Hussey, R., Hodgdon, J., Schlossman, S . , and Reinherz, E. (1983). Clonotypic structures involved in antigen-specific human T cell function. Relationship to the T3 molecular complex. J . Exp. Med. 157, 705-719. Minami, Y., Weissman, A. M., Samelson, L. E., and Klausner, R. D. (1987). Building a multichain receptor. Synthesis, degradation and assembly of the T-cell antigen receptor. Proc. Natl. Acad. Sci. U.S.A. 84, 2688-2692. Orloff, D. G., Frank, S. F., Robey, F. A , , Weissman, A. M., and Klausner, R. D. (1989). Biochemical characterization of the eta chain of the T cell antigen receptor. Evidence for its relationship to zeta. J . B i d . Chem. 264, 14812-14817. Pelham, H. R. B. (1986). Speculations on the functions of the major heat shock and glucose regulated proteins. Cell 46,959-961. Pettey, C . L., Alarcon, B., Malin, R., Weinberg, K., and Terhorst, C. (1987). T3-p28 is a protein associated with the delta and epsilon chains of the T cell receptor-T3 antigen complex during biosynthesis. J . Biol. Chem. 262, 4854-4859. Reinherz, E. L., Hussey, R. E., and Schlossman, S . F. (1980). A monoclonal antibody blocking human T cell function. Eur. J . Immunol. 10, 758-762. Reinherz, E. L., Meuer, S., Fitzgerald, K. A., Hussey, R. E., Levine, H., and Schlossman, S. F. (1982). Antigen recognition by human T lymphocytes is linked to surface expression of the T3 molecular complex. Cell 30, 735-743. Reinherz, E., Meuer, S . , Fitzgerald, K., Hussey, R., Hodgdon, J., Acuto, O., and Schlossman, S. ( 1983). Comparison of T3-associated 49- and 42-kilodalton cell surface molecules on individual human T-cell clones: Evidence for peptide variability in T-cell receptor smctures. Proc. Narl. Acad. Sci. U.S.A. 80, 4104-4108. Rosen, 0. M. (1987). After insulin binds. Science 237, 1452-1458. Samelson, L. E., Harford, J. B., and Klausner, R. D. (1985). Identification of the components of the murine T cell antigen receptor complex. Cell 43, 223-231. Schmitt-Verhulst, A.-M., Guimezanes, A., Boyer, C., Peonie, M., Tsien, R., Bufeme, M., Hua, C . , and Leserman, L. (1987). Pleiotropic loss of activation pathways in a T-cell receptor alphachain deletion variant of a cytolytic T cell clone. Nature (London) 325, 628-63 1 . Sussman, J. J., Bonifacino, J. S . , Lippincott-Schwartz, J., Weissman, A. M., Saito, T., Klausner, R. D., and Ashwell, J. D. (1988). Failure to synthesize the T cell CD3-zeta chain: Structure and function of a partial T cell receptor complex. Cell 52, 85-95. Tunnacliffe, A,, Buluwela, L., and Rabbits, T. H. (1987). Physical linkage of three CD3 genes on human chromosome 11. EMBO J . 6, 2953-2957. van den Elsen, P., Shepley, B.-A., Borst, J., Coligan, J. E., Markham, A. F., Orkin, S . , and Terhorst, C. (1984). Isolation of cDNA clones encoding the 20 K T3 glycoprotein of human T cell receptor complex. Nature (London) 312, 413-418.

4. REGULATING EXPRESSION OF THE T-CELL RECEPTOR

51

van den Elsen, P., Shepley, B.-A,, Cho, M., and Terhorst, C. (1985). Isolation and characterization of a cDNA clone encoding the murine homologue of the human 20 K T3/T-cell receptor glycoprotein. Nature (London) 314, 542-544. van Wauwe, F. P., DeMay, J. R., and Goossener, J. G. (1980).OKT3: A monoclonal anti-human T lymphocyte antibody with potent mitogenic properties. J . Immunol. 124, 2708-2713. Weiss, A , , and Stobo, J. D. (1984). Requirement for the co-expression of T3 and the T cell antigen receptor on a malignant human T cell line. J . Enp. Med. 160, 1284-1299. Weissman, A. M., Baniyash, M., Hou, D., Samelson, L. E., Burgess, W. H . , and Klausner, R. D. (1988a). Molecular cloning of the zeta chain of the T cell antigen receptor.Science 239, 10181021. Weissman, A. M., Hou, D., Orloff, D. G . , Modi, W. S . , Sevanez, H . , O’Brien, S . , and Klausner, R. D. (1988b). The molecular cloning and chromosomal localization of the human T cell receptor zeta chain: Distinction from the molecular CD3 complex. Proc. Nutf. A c Q ~ .Sci. U.S.A. 85, 9709-9713. Weissman, A. M., Rank, S . J., Orloff, D. G., Mercep, M . , Ashwell, J. D., and Klausner, R. D. (1989). Role of the zeta chain in the expression of the T cell antigen receptor: Genetic reconstitution studies. EMBO J . 8, 3651-3656. Wilson, I. A., Skehel, J. J., and Wiley, D. C. (1981). Structure of the haemagglutinin membrane glycoprotein of influenza virus at 3A resolution. Nature (London) 289, 366-373. Wilson, R. K., Lai, E., Concannon, P., Barth, R. K . , and Hood, L. E. (1988). Structure, organization and polymorphism of murine and human T-cell receptor alpha and beta chain gene families. Immunol Rev. 101, 149-172.

This Page Intentionally Left Blank

CURRENT TOPICS IN MEMBRANES AND TRANSPORT, VOLUME 36

Chapter 5

The Photosynthetic Reaction Center from the Purple Bacterium Rhodopseudomonas viridis Aspects of Membrane Protein Structure HARTMUT MICHEL" AND JOHANN DEISENHOFERf *Max-Planck-Institutfur Biophysik 0-6000 FrankjfurtlM 71, Federal Republic of Germany ?Howard Hughes Medical Institute and Department of Biochemistry University of Texas Southwestern Medical Center Dallas, Texas 75235

Introduction Results and Discussion A. Pigment Arrangement B. General Architecture and Protein Structure C. The Use of Only One Pigment Branch D. The Membrane Anchor of the Cytochrorne Subunit E. General Aspects of Membrane Protein Structure F. Distribution of Surface Atoms, Amino Acids, and Bound Water Molecules 111. Conclusions References I. 11.

1.

INTRODUCTION

The so-called photosynthetic reaction centers are complexes of integral membrane proteins, which form the binding sites for the photosynthetic pigments. These pigments are involved in the primary charge separation of photosynthesis and the subsequent electron transfer steps. The photosynthetic reaction centers from the purple photosynthetic bacteria 53 Copynght 0 19% by Academic Press, Inc. All rights of repduction in my form rerewed

54

HARTMUT MICHEL AND JOHANN DEISENHOFER

are easy to isolate and relatively stable. They are the first membrane proteins that could be obtained in a well-ordered crystalline form. As a consequence, an X-ray crystallographic analysis could be performed at high resolution. Therefore, they are the model membrane proteins that can be used to test the various methods of predicting the topology and the secondary structure of membrane proteins. Most of the reaction centers from purple bacteria contain three protein subunits that are called H (heavy), M (medium), and L (light) subunits according to their apparent molecular weights as determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (for reviews see Feher and Okamura, 1978; Hoff, 1982). Many reaction centers, including that from Rhodopseudomonas (Rps.) viridis, contain a tightly bound cytochrome subunit, which is involved in the rereduction of the photooxidized primary electron donors. The cytochrome subunit of the reaction center from Rps. viridis contains four heme groups. All four heme groups operate at different redox potentials (Dracheva et al., 1988). In Rps. viridis the photosynthetic pigments and cofactors are four bacteriochlorophyll b molecules, two bacteriopheophytin b molecules, one menaquinone as “primary acceptor” (QA), one ferrous nonheme iron, and one “secondary acceptor” (QB). Most of the reaction centers from the other purple bacteria contain bacteriochlorophyll a instead of bacteriochlorophyll b, bacteriopheophytin a instead of bacteriopheophytin b, and the primary acceptor is formed by another ubiquinone. The successful crystallization of the reaction center from Rps. viridis (Michel, 1982) was an important step. The reaction centers in the crystals retained their photochemical activity (Zinth et al. , 1983). The X-ray crystallographic analysis of these crystals allowed the calculation of an electron density map at 3 A resolution from which the arrangement of the chromophores could be deduced (Deisenhofer et al., 1984). Subsequently, the structure of the protein subunits (Deisenhofer et al . , 1985) and details of the pigment-chromophore interactions (Michel et al., 1986b) could be presented. The photosynthetic reaction center from another purple bacterium, Rhodobacter (Rb.) sphaeroides, could also be crystallized (Allen and Feher, 1984; Chang et al., 1985). Using the known structure of the reaction center from Rps. viridis, the crystal structure of the Rb. sphaeroides reaction center could be determined by molecular replacement. In the following sections we describe and discuss pigment arrangement, protein structure, and the membrane protein aspects of the reaction center structure with special emphasis on the membrane protein structure.

II. RESULTS AND DISCUSSIONS A. Pigment Arrangement

X-Ray crystallographic analysis yielded two highly surprising results. These are the linear arrangement of the four heme groups, which can be seen at the top of Fig. 1 (taken from Deisenhofer et al., 1984), and the symmetrical arrangement

5. PHOTOSYNTHETIC REACTION CENTER

55

i s’ HE

o . . . . . . . . . . . . . .

I . . . . . . . . . . . . . .

FIG. 1. Stereo pair: Arrangement of the pigments in the photosynthetic reaction center from Rhodopseudomonas viridis, showing, from top to bottom, four heme groups (HE), four bacteriochlorophylls (BC), on nonheme iron (Fe), and one menaquinone (MQ). The approximately 2fold symmetry axis relating the photosynthetic pigments runs vertically in the plane of the drawing. The approximate position of the periplasmic (0, outer) and cytoplasmic (I, inner) face of the photosynthetic membrane is indicated by dotted lines.

of the photosynthetic pigments, as shown in the lower half of Fig. 1. The four heme groups, which have the function of rereducing the primary electron donor, are related by an approximately 2-fold local rotation axis, which runs nearly perpendicular to the picture plane in Fig. 1. The other approximate local 2-fold rotation axis, which we call the central local rotation axis, runs vertically in the picture plane of Fig. 1 . In the center of Fig. 1 it relates the two bacteriochlorophyll molecules, which constitute the primary electron donor (“special pair”); at the bottom it runs through the ferrous nonheme iron atom. The existence of a special pair as primary electron donor had been postulated by Norris et al. (197 1) on the basis of EPR experiments. A number of proposals for its exact structure have been made, but not one proved to be correct. The ring systems of the two bacteriochlorophylls constituting the dimer are nearly parallel and overlap with their pyrrole rings I, which carry an acetyl side chain. The plane-plane distance

56

HARTMUT MICHEL AND JOHANN DEISENHOFER

of these two bacteriochlorophylls is -3.2 A. The ring systems of all chlorine pigments are also roughly related by the central local rotation axis. Since two “accessory” bacteriochlorophylls and the two bacteriopheophytins are also related by the central load rotation axis, two structurally nearly equivalent branches are formed that might be used for electron transfer across the membrane. The symmetry does not hold for the arrangement of the phytyl side chains, which is substantially different in both branches. Also, the original electron density map did show only the localization of the primary quinone, Q A , at the end of one of the pigment branches near the ferrous nonheme iron. The secondary quinone, QB, could not be localized, since most of it is lost during the isolation and crystallization of the reaction centers. Another hint for the inequivalence of the two branches is obtained from the fact that the two bacteriopheophytins absorb light of different wavelengths. It is known that only the one absorbing light at longer wavelength is used in the electron transfer. Optical absorbance spectra of crystals taken with polarized light (Zinth et al., 1983), together with the structural arrangements of the pigments, show that the bacteriopheophytin absorbing at the longer wavelength is the one closer to QA, shown in the right-hand side of Fig. 1. Electron density for the secondary quinone QB appeared during the refinement of the structure and the electron density at 2.3 A resolution (Deisenhofer et al., 1990). Its binding site is symmetrically related to the binding site of QA by the local 2-fold rotation axis. From the structure a hopping of the electron from the primary electron donor to the “accessory” chlorophyll and then to the bacteriopheophytin is quite feasible. However, the evidence published to date favors the bacteriopheophytin as first electron acceptor (e.g., Breton et al., 1986). The role of the accessory bacteriochlorophyll is under discussion as a “virtual” (see, e.g., Bixon et al., 1987; Creighton et al., 1988) intermediate in a “superexchange mechanism.” From the bacteriopheophytin the electron is transferred to QA and from there to QB parallel to the surface of the membrane. There is no evidence for the participation of the nonheme iron in this electron transfer, since it can be removed and replaced by other metal ions without substantial changes in the kinetics of the electron transfer from QA to QB. The binding site of QB is the target for a number of herbicides such as terbutryn (2-thiomethyl-4-ethylamino-6-tert-butylamino-~-~azine), whose binding site has also been determined by X-ray crystallography (Michel et al., 1986b). Terbutryn is closer to the entrance to the binding site than in the binding site. 6. General Architecture and Protein Structure The L and M subunits form the membrane-embedded core of the photosynthetic reaction center. Both possess five long membrane-spanning helices. Polypeptide segments, which are also partly helical, connect the transmembrane

5. PHOTOSYNTHETIC REACTION CENTER

57

helices and form flat surfaces parallel to the membrane surfaces. A comparison of the polypeptide chain folding in both subunits shows a high degree of similarity. The structurally similar segments include the transmembrane helices and a large fraction of their connections. These structurally similar segments can be superimposed by a rotation of 180" around the central local axis, which also relates the photosynthetic pigments (see Section 11,A). The photosynthetic pigments are bound exclusively to the L and M subunits. The special pair and the ferrous nonheme iron atom are found at the interface between both subunits near the outer (periplasmic) and inner (cytoplasmic) sites of the membrane. The accessory chlorophyll and the bacteriopheophytin involved in the light-driven electron transfer are more closely associated with the L subunit than with the M subunit. The primary quinone QA is located at the end of this L branch of pigments, but to complicate the picture, the binding site of its head group is made up by the helix connection of the fourth and fifth transmembrane helices of the M subunit. Vice versa, the head group of Qe is bound to the connection of the fourth and fifth transmembrane helices of the L subunit. This core of the reaction center has elliptical cross sections with axes of 70 and 30 A. The L and M subunits possess a sequence homology of 25%. Mainly, amino acids of structural importance and amino acids involved in pigment binding are identical in both subunits. These sequence identities are most likely explained by the existence of an ancestral entirely symmetrical reaction center (see Section 11,C). In the present reaction center the M subunit with 323 residues is 50 residues longer than the L subunit with its 273 residues (Michel et al., 1986a). The insertions in M are located near the N-terminus (20 residues), in the connections between the first and second transmembrane helices (7 residues), between the fourth and fifth transmembrane helices (7 residues) and at the N-terminus, where the M subunit possesses an appendage of 16 residues. The insertion between the fourth and fifth transmembrane helices is of importance for the binding of the nonheme iron by donating a glutamic acid as a bidentate ligand to the ferrous nonheme iron atom. Figure 2 shows a drawing of the polypeptide chains together with the arrangement of the cofactors. It can be seen that the H subunit is anchored to the membrane by a single membrane-spanning helix. The remainder mainly forms a globular domain that is bound to the LM core on the cytoplasmic side of the membrane. This domain contains an extended system of antiparallel and parallel p sheets. The H subunit with 258 amino acids, despite its name, is the smallest subunit of the complex (Michel et al., 1985). It is not involved in electron transfer. It may be of importance for the assembly of the whole complex, and it seems to play a role in the proton transfer to the doubly reduced secondary quinone QB. With 336 residues the cytochrome subunit is the largest subunit of the complex (Weyer et al., 1987b). Its complicated structure can be summarized as follows: it consists of an N-terminal segment, two pairs of heme-binding segments, and a

58

HARTMUT MICHEL AND JOHANN DEISENHOFER

-v FIG. 2. Stereo pair: Ribbon drawing of the polypeptide chains of the reaction center from Rhodopseudomoms viridis, together with the chromophore model (thin lines), showing the cytochrome (top), and subunits L (middle left), M (middle right), and H (bottom, with the N-terminal helix extending from the cytochrome).

segment connecting the two pairs. Each heme-binding segment consists of a helix followed by a turn and the covalent heme-binding sequence Cys-X-Y-CysHis, which is typical for c-type cytochromes. The heme planes are parallel to the helices. The cytochrome subunit is attached to the L and M subunits on the outer ( periplasmic) side. It is anchored to the membrane by two covalently bound fatty acids (see Section 11,D). C. The Use of Only One Pigment Branch

The overall symmetrical core of the reaction centers from the purple bacteria and the sequence homologies between the core-forming L and M subunits sug-

5. PHOTOSYNTHETIC REACTION CENTER

59

gests that the ancestors of the purple bacteria (and the ancestors of the photosystem II-containing organisms) possessed a completely symmetrical reaction center, which contained two symmetrically arranged copies of the same protein subunit leading to a “homodimeric” structure. Only after a gene duplication and subsequent mutations did the formation of the heterodimeric reaction center become possible. Obviously, these evolutionary relations suggest an advantage of the heterodimeric reaction center, where one of the electron-conducting pigment branches could be switched off. First, we will describe the structural differences of both pigment branches that are the consequence of the different protein environment in both branches.

1 . The bacteriochlorophyll ring systems of the special pair show a different degree of nonplanarity. The M-side bacteriochlorophyll of the special pair is considerably more deformed than the one on the L side. The different degree of nonplanarity can cause an unequal charge distribution between the two components of the special pair, which in turn can lead to the unidirectional electron transfer. 2. The symmetry operations to achieve optimal superposition of the specialpair bacteriochlorophylls, or of the accessory bacteriochlorophylls, or of the bacteriopheophytins are slightly different. Optimal superposition of the tetrapyrrole rings of the special pair is achieved by a rotation of 179.7”, of the accessory bacteriochlorophylls by - 175.8”, and of the bacteriopheophytins by - 173.2”. Because of this imperfect symmetry, interatomic distances and interplanar angles are different in both branches. The closest distance between atoms involved in double bonds in the special pair and in the h i d e bacteriopheophytin is 0.7 A less than the corresponding distance between the special pair and the M-side bacteriopheophytin. This and similar structural differences lead to different overlap of electronic orbitals in both branches. 3. The reaction center along the M-side branch is more disordered than along the h i d e branch. Both phytyl side chains of the M-side accessory bacteriochlorophyll and bacteriopheophytinare partially disordered at their ends. On the L side they have a different conformation and are well ordered. A carotenoid is present only near the M-side accessory bacteriochlorophyll. It may contribute to the difference in phytyl chain structure, since it prevents an identical arrangement of the phytyl side chain on both sides. 4. The structure along the L branch is more rigid than along the M branch. The crystallographic temperature factors (B values) are considerably lower on the L side than on the M side. 5. The protein environment influences the electronic properties. Around the special pair there are more electron-attractingpolar residues and hydrogen bonds on the L side than on the M side. There are also more aromatic residues along the L side. This fact could lead to the high rigidity of the structure on the L side,

60

HARTMUT MICHEL AND JOHANN DEISENHOFER

since the aromatic residues are bulkier than the aliphatic ones, but the presence of n electrons might help in electron transfer. A prominent residue is tryptophan M250, bridging the L-side bacteriopheophytin and Qa. Why was one of the branches for light-driven electron transfer switched off? As a rather trivial explanation, the asymmetry in both pigment branches and the protein environment can cause an asymmetry in the distribution of the electron in the special pair, especially in the excited state. Such an asymmetry can lead to a preferred release in one direction and result in a faster first electron transfer step, thereby minimizing competing reactions. A higher quantum yield for the electron transfer can be achieved. A clear advantage in the present day’s reaction centers resides on the electronaccepting side. QB is a two-electron carrier. Transfer to the QB of one electron only leads to a semiquinone radical. The electron is lost, and no energy is stored if a second electron is not transferred to QB within milliseconds to seconds. Only after (or during) double reduction is Qe protonated and energy stored in the form of the quinol. With two identical parallel electron transfer chains the probability for the second electron to be transferred to the same quinone is only 50%. Thus, frequently the absorption of two photons leads to the formation of two semiquinone radicals in the same reaction center, and energy is not stored in a stable way. This problem can be solved when the two quinones act in series, and protonation and release are only possible for the final quinone. This is possible if the two protein subunits become different by a gene duplication and subsequent mutations, thus leading to a heterodimeric reaction center. The necessijry steps are indicated in Fig. 3. It then becomes QB, as it is seen in the reaction centers of modem purple bacteria. A considerable increase in the efficiency of light energy conversion, especially under low light conditions, must result.

D. The Membrane Anchor of the Cytochrome Subunit The crystallographic analysis established firmly that the L and M subunits both possess five transmembrane helices, whereas the H subunit is anchored to the membrane by one transmembrane helix. The electron density map did not indicate any intramembranous part of the cytochrome subunit. However, its aggregation tendency in the absence of detergents suggested that it still may be a membrane protein. One hint for the possible existence of a membrane anchor was the observation that no N-terminal amino acid could be identified after the first step of Edman degradation, but the Edman degradation could be continued and an amino acid sequence could be obtained starting with the second amino acid from the N-terminus, thus suggesting a modified N-terminal amino acid. The phenylthiohydantoin derivative of such a modified amino acid could be isolated, and its structure could be identified by mass-spectroscopic analysis

5. PHOTOSYNTHETIC REACTION CENTER

61

FIG. 3 . Schematic drawing of the situation in an ancestral symmetrical reaction center, and the steps needed to convert it into the modem reaction center. These steps are shown in broken lines. It is most important to open a way for electron transfer from QI to QII, which then become QA and QB, and to switch off protonation and release of QI by appropriate mutations. PheoI and PheoII indicate the (bacteri0)pheophytins of pigment branches I and 11, respectively.

(Weyer ef al., 1987a,c): the N-terminal amino acid is a cysteine linked to a glyceryl residue via a thioether bridge. The fatty acids are covalently bound to the two OH groups of the glyceryl residue. The fatty acids consist of a statistical mixture of various singly unsaturated and singly hydroxylated C,, fatty acids. Thus, the cytochrome subunit also possess a membrane anchor, which is now of a lipid type and not of a peptide type. This membrane anchor is similar to that of the bacterial lipoproteins (see, e.g., Pugsley ef al., 1986; Yu et al., 1986).

E. General Aspects of Membrane Protein Structure The X-ray structure analysis indicated the existence of five transmembrane helices in both the L and M subunits as a major structural motif. With the availability of the amino acid sequences a variety of structural features, unique to membrane proteins, could be deduced. The transmembrane helices, which have a length of 21-28 residues, all contain a stretch of 19 uncharged, mainly hydrophobic amino acid residues. In most cases they have glycines or prolines at or very close to their ends (Michel et al., 1986a). These residues are conserved between the L and M subunits of the bacterial reaction centers. These amino acids allow the formation of necessary turns of the peptide chains at the beginnings and ends of the transmembrane helices. Frequently at the cytoplasmic end as many as five charged residues are observed. The function of these charged residues may be to stop the insertion into the membrane and transfer of the peptide chain. It is thus not surprising that the commonly used prediction meth-

62

HARTMUT MICHEL AND JOHANN DEISENHOFER

ods for membrane-spanninghelices (e.g., according to Kyte and Doolittle, 1982) correctly predict five membrane-spanning helices (if large “windows” of 19 or 21 amino acids are used), but that the ends of the helices are not correctly predicted because of the charge clusters at the ends of the helices (Michel et al., 1986a). However, there is already one example of an incorrect prediction. The anaiog to the L subunit in photosystem I1 reaction centers, the 32-K Da or Dl protein, has been predicted to contain seven transmembrane helices (Rao et al., 1983). Experimental evidence and sequence comparisons clearly show it contains only five membrane-spanning helices. Two long connections between transmembrane helices, between the first and second as well as between the fourth and fifth transmembrane helices, are predicted to be transmembrane helices. In line with these observations, the number of charged residues in the L and M subunits on the periplasmic side of the membrane is considerably lower than on the cytoplasmic side. The L subunit has 17 charged residues on the cytoplasmic side of the membrane and only 10 on the periplasmic side; the M subunit has 24 charged residues on the cytoplasmic side and only 14 on the periplasmic side. The reduced amount of charged residues on the periplasmic side makes good sense, since these residues certainly hinder the translocation of the extracellular parts of these proteins through the hydrophobic part of the bilayer. If one counts the net charges of the peptide chains on the periplasmic and on the cytoplasmic sides on the basis of the assumption that all glutamic acids, aspartic acids, and carboxy termini are negatively charged, whereas the arginines, lysines, and amino termini are positively charged, then one finds that the polar ends of the transmembrane helices and their connections on the periplasmic side are always more negatively charged than their counterparts on the cytoplasmic side. As a result of this unequal distribution of positive and negative charges the cytoplasmic side of the M subunit carries four positive net charges and the periplasmic side four negative net charges. The same consideration for the L subunit reveals two positive net charges on the cytoplasmic and four negative net charges on the periplasmic side. This asymmetric charge distribution across the membrane leads to an electric dipole moment perpendicular to the plane of the membrane. The electric polarity becomes even more pronounced if one considers that the reaction centers from the purple bacteria contain a firmly bound ferrous nonheme iron on the cytoplasmic side. Since the bacterium is negatively charged inside the cells (because of the action of electrogenic ion pumps), the L and M subunits are oriented in the cells with their electric dipoles in the energetically more favorable direction. Thus the combination of the electric field across the membrane and the unequal distribution of charges in the membrane proteins may be one of the factors that determine the orientation of the membrane proteins inside and outside of the cell. The electric field across the membrane may also be needed for the correct protein folding. If in a long hydrophobic stretch, a helix or a pair of helices has been formed during mRNA translation, it may become

63

5. PHOTOSYNTHETIC REACTION CENTER

associated with the hydrophobic part of the membrane. Then the electric field across the membrane and the net charges at the ends of the helix or helix pair may determine which end will be transferred to the periplasmic side of the membrane. In this respect it is interesting to note that the N-terminus of the H subunit that has to be translocated across the membrane also carries one negative net charge.

F. Distribution of Surface Atoms, Amino Acids, and Bound Water Molecules Since the interaction between the lipids and the proteins has to occur at the surface of the protein, a look at the distribution on the polar atoms (N, 0)and the C atoms at the surface of the reaction center may be very informative. Figure 4 shows the accessible surface area, occupied by C atoms, in sections parallel to the membrane. At the cytoplasmic side C atoms occupy -60% of the surface. Then an increase to 95% occurs-indicating the contact zone between the protein and the hydrophobic portion of the lipid bilayer-before the surface occupancy of C atoms declines to -57%, which is the standard value for watersoluble, globular proteins (Miller et al., 1987). Two important conclusions can be drawn from Fig. 4: (1) the thickness of the hydrophobic position is about 3031 A, which is smaller than expected, and (2) the special pair is located in the hydrophobic part of the lipid bilayer whereas the ferrous nonheme iron is found in the polar part or the cytoplasmic side. Figure 5 shows the distribution of those

Fe

100

SPP

80

n

al

$ 60

t al

s! a"

40

20

0

0

20 4( Central diad

100

FIG.4. Percentage (vertical axis) of the accessible surface area occupied by carbon atoms shown for 3-A-thick layers perpendicular to the noncrystallographic central rotation axis, which runs through the ferrous nonheme iron atoms (E) and the special pair (SpP).

64

HARTMUT MICHEL AND JOHANN DEISENHOFER

FIG. 5 . C,-backbone model (thin lines) of the photosynthetic reaction center from Rhodopseudornonas viridis and the location of the charged amino acid residues Asp, Glu, Are, and Lys (shown as bold atomic models).

5. PHOTOSYNTHETIC REACTION CENTER

65

FIG. 6 . C,-backbone model (thin lines) of the L and M subunits of the photosynthetic reaction center from Rhodopseudomoms viridis and the location of the tryptophan residues (shown as bold atomic models).

amino acids that have charged side chains under normal conditions. A central zone devoid of any charged residue can be seen. Its 25-A thickness is slightly less than that of the hydrophobic surface zone. This is due to two arginine residues and one glutamic acid, which are located in a hydrophobic environment close to the cytoplasmic end of the hydrophobic zone. There is no apparent countercharge nearby that would allow a classical charge neutralization. However, the positive charge of the arginine side chains seems to cancel the partial negative charge at the C-terminal ends of the short helices (which partly intrude into the membrane) in the connections of the long C and D transmembrane helices. These arginines

66

HARTMUT MICHEL AND JOHANN DEISENHOFER

0 0

0

8

0 0 0

00

0

0

FIG. 7. Position of the firmly bound water molecules in the photosynthetic reaction center from Rhodopseudomonas viridis.

certainly play an important structural role. They are conserved between all bacterial reaction centers and even photosystem I1 reaction centers. The glutamic acid (L104) seems to be protonated, thus neutral, and to form a hydrogen bond with one of the bacteriopheophytins (Michel et al., 1986b). In the L and M subunits the amino acid tryptophan also shows a remarkably uneven distribution. As can be seen in Fig. 6, about two-thirds of the tryptophans are found at the ends of the transmembrane helices or in the helix connections at the periplasmic side. Only a few tryptophan residues are in the hydrophobic zone, where they are in contact with pigments. The residual tryptophans are located in the transition zone from the hydrophobic to polar surface near the

5. PHOTOSYNTHETIC REACTION CENTER

67

cytoplasmic side. The indole rings of the tryptophans are preferentially oriented toward the hydrophobic zone of the membrane. At the present resolution of 2.3 8,we also could tentatively identify 201 water molecules. Their location is shown in Fig. 7. Only 5 water molecules are found in the hydrophobic intramembranous zone. They may play an important structural role, since a few of them cross-link transmembrane helices by hydrogen bonds. Figure 8 shows one of these water molecules and its probable hydrogen-

FIG. 8. A firmly bound water molecule (301 WAT) in the hydrophobic part of the membrane cross-linking two transmembrane helices by forming hydrogen bonds with the peptide oxygen atoms of leucine L180 and alanine M207.Another hydrogen bond with the side chain of asparagine L183 is possible.

68

HARTMUT MICHEL AND JOHANN DEISENHOFER

bonding pattern. It seems to form one hydrogen bond as a donor to a peptide oxygen of the M subunit, and another to a peptide oxygen of the L subunit. A third hydrogen bond with an asparagine side chain is possible. How much these water molecules contribute to the stability of the reaction center is an open question that has to be answered in the future.

111.

CONCLUSIONS

We have shown that membrane proteins can be crystallized and that an X-ray crystallographic analysis can lead to results of the same quality as with watersoluble proteins. Some important conclusions on the structure and biosynthesis of membrane proteins could be drawn, as well as on the function of the photosynthetic reaction center. There is good hope that other membrane proteins will follow, even though the progress made with other membrane proteins has been slow. Well-diffracting crystals have been obtained only with photosynthetic reaction centers from bacteria, and quite recently with bacterial light-harvesting complexes (R. J. Cogdell et al., unpublished observations; H. Michel, unpublished observations). When crystallizing membrane proteins the necessary finetuning with respect to the size of the detergent micelle and the size of the polar head group of the detergent is still a formidable task that has to be solved empirically for each individual membrane protein. REFERENCES Allen, J. P., and Feher, G. (1984). Crystallization of reaction center from Rhodopseudomonas sphaeroides: Preliminary characterization. Proc. Narl. Acad. Sci. U.S.A. 81, 4795-4799. Bixon, M., Jortner, J., Michel-Beyerle, M. E., Ogrodnik, A , , and Lersch, W. (1987). The role of the accessory bacteriochlorophyll in reaction centers of photosynthetic bacteria: Intermediate acceptor in the primary electron transfer? Chem. Phys. Lett. 140, 626-630. Breton, J. Martin, J.-L., Petrich, J., Migus, A., and Antonetti, A. (1986). The absence of a spectroscopically resolved intermediate state P + B - in bacterial photosynthesis. FEBS Leu. 209, 37-43. Chang, C.-H., Schiffer, M., Tiede, D., Smith, U . , and Noms, J. (1985). Characterization of bacterial photosynthetic reaction center crystals from Rhodopseudomonas sphaeroides R-26 by X-ray diffraction. J . Mol. Biol. 186, 201-203. Creighton, S . , Hwang, J.-K., and Warshel, A . (1988). Simulating the dynamics of the primary charge separation process in bacterial photosynthesis. Biochemist? 27, 774-78 I . Deisenhofer, J., Epp, O., Miki, K., Huber, R . , and Michel, H. (1984). X-ray structure analysis of a membrane protein complex. J . Mol. Biol. 180, 385-398. Deisenhofer, J., Epp, 0.. Miki, K . , Huber, R., and Michel, H. (1985). Structure of the protein subunits in the photosynthetic reaction centre of Rhodopseudomonas viridis at 3 A resolution. Nature (London) 318, 618-624. Deisenhofer, J . , Epp, O., Sinning, I . , and Michel, H. (1990). To be published. Dracheva, S. M., Drachev, L. A . , Konstantinov, A. A., Semenov, A. Y., Skulachev, V. P . , Arutjunjan, A. M., Shuvalov, V. A , , and Zaberezhnaya, S . M. (1988). Electrogenic steps in the redox

5. PHOTOSYNTHETIC REACTION CENTER

69

reactions catalyzed by photosynthetic reaction-centre complex from Rhodopseudomonas viridis. Eur. J . Biochem. 171, 253-264. Feher, G., and Okamura, M. Y. (1978). In “The Photosynthetic Bacteria’’ (R. Sistrom and R. K. Clayton, eds.), pp. 349-386. Plenum, New York. Hoff, A. J. (1982). Mol. Biol. Biochem. Biophys. 35, 81-151. Kyte, J.. and Doolittle, R. F. (1982). A simple method for displaying the hydropathic character of a protein. J. Mol. Biol. 157, 105-132. Michel, H. (1982). Three-dimensional crystals of a membrane protein complex. The photosynthetic reaction centre from Rhodopseudomonas viridis. J. Mol. Biol. 158, 567-572. Michel, H., Weyer, K. A,, Gruenberg, H., and Lottspeich, F. (1985). The “heavy” subunit of the photosynthetic reaction centre from Rhodopseudomoms viridis: Isolation of the gene, nucleotide and amino acid sequences. EMBO J . 4, 1667-1672. Michel, H., Weyer, K. A,, Gruenberg, H., Dunger, I., Oesterhelt, H., and Lottspeich, F. (1986a). The “light” and “medium” subunits of the photosynthetic reaction centre from Rhodopseudomonus viridist Isolation of the genes, nucleotide and amino acid sequence. EMBO J. 5 , 1149-1 158. Michel, H . , Epp, O., and Deisenhofer, J. (1986b). Pigment-protein interactions in the photosynthetic reaction centre from Rhodopseudomonas viridis. EMBO J . 5, 2445-245 1. Miller, S., Janin, J., Lesk, A. M., and Chothia, C. (1987). Interior and surface of monomeric proteins. J . Mol. Biol. 1%, 64-656. Noms, J. R., Uphaus, R. A , , Crespi, H. L., and Katz, J. 1. (1971). Electron spin resonance of chlorophyll and the origin of signal I in photosynthesis. Proc. Natl. Acad. Sci. U.S.A. 68,625628. Pugsley, A. P., Chapon, C., and Schwartz, M. (1986). Extracellular pullulanase of Klebsiella pneumoniae is a lipoprotein. J . Bacteriol. 166, 1083-1088. Rao, J. K. M., Hargrave, P. A,, and Avgos, P. (1983). Will the seven-helix bundle be a common structure for integral membrane proteins? FEES Lett. 156, 165-169. Weyer, K. A., Schafer, W., Lottspeich, F., and Michel, H. (1987a). The cytochrome subunit of the photosynthetic reaction center from Rhodopseudomonas viridis is a lipoprotein. Biochemistry 26, 2909-2914. Weyer, K. A., Lottspeich, F., Gruenberg, H., Lang, F., Oesterhelt, D., and Michel, H. (1987b). Amino acid sequence of the cytochrome subunit of the photosynthetic reaction centre from the purple bacterium Rhodopseudomonas viridis. EMBO J . 6, 2197-2202. Weyer, K. A,, Lottspeich, F., Schafer, W., and Michel, H. (1987~).The fatty acid-anchored four heme cytochrome of the photosynthetic reaction center from the purple bacterium Rhodopseudomonus viridis. In “Cytochrome Systems,” (S. Papa, B. Chance and L.Ernster, eds.), pp. 325-331. Plenum, New York. Yu, F., Inouye, S., and Inouye, M. (1986). Lipoprotein-28, a cytoplasmic membrane lipoprotein from Esrherichia roli. J . B i d . Chem. 261, 2284-2288. Zinth, W., Kaiser, W., and Michel, H. (1983). Efflcient photochemical activity and strong dichroism of single crystals of reaction centers from Rhodopseudomoms viridis. Biochim. Biophys. Acta 723, 128-131.

This Page Intentionally Left Blank

CURRENT TOPICS IN MEMBRANES AND TRANSPORT, VOLUME 36

Chapter 6 Bacteriorhodopsin Folding in Membranes: A Two-Stage Process D . M . ENGELMAN, * B . D . ADAIR,* J . F . HUNT, * T . W . KAHN, * AND J.-L. POPOTf *Department of Molecular Biophysics and Biochemistry Yale University New Haven, Connecticut 06511 TInstitut de Biologie Physico-Chimique Collkge de France, 75231 Paris, France

I. Introduction 11. Bacteriorhodopsin 111. Bacteriorhodopsin Fragments Contain Stable Transbilayer Helices IV. Links and Retinal Are Not Required for Folding V. Polar Interactions in Helix-Helix Associations

VI. Packing Effects VII. Summary References

1.

INTRODUCTION

The folding of many membrane proteins within lipid bilayers can be understood in terms of a two-stage process in which individually stable transmembrane helices assemble laterally to form tertiary structure. The idea of independently stable helices emerges from the structures of the photosynthetic reaction center (1-3) and bacteriorhodopsin (4,5), in which the transmembrane regions of the polypeptides are all in helical conformations. That these structures are well predicted by simple searches for hydrophobic stretches of 20 amino acids (6,7) suggests that each of the helices would be independently stable as a separate 71 Copynght 0 1990 by Acadermc Press. Inc. All rights of repmductlon m any form reserved

72

D. M. ENGELMAN ET AL.

STflGE

I

STFIGE I 1

FIG. 1. Two stages in the folding of a polytopic membrane protein are shown. In the first stage, transbilayer helices are formed as independently stable entities. The principal energies contributing to helical stability are the hydrogen bonding of the main chain and the hydrophobic effect arising from the presence of largely nonpolar side chains. In stage 11, the helices interact with each other to form a complete, functional tertiary structure. Some folding of the l w p regions also occurs. Stage I1 is thought to be driven largely by the preferred packing of helices against each other compared with the packing of helices against lipid. Other important contributions to stability arise from the tendency of lipid to self-associate compared with associations with helices, contributions from the polypeptide links between helices, polar interactions, and interactions with prosthetic groups if they are present. Experiments with bacteriorhodopsin have provided support for this model.

transbilayer structure interacting only with the nonpolar region of the lipid bilayer. Since the helices are found in the fully folded polytopic proteins, it seems that they have not been much altered from the conformation that would be expected for them as separate entities, and it follows that the interaction energies producing the tertiary structure of the folded molecule may not be as strong as those stabilizing the helices. Thus, it is proposed that the folding can be conceptually divided into two stages as shown in Fig. 1. It is interesting that the division of folding into a two-stage process allows a division of energies. The hydrophobic effect and hydrogen bonding account for the stability and conformation of transmembrane helices (5,7). To assemble the helices requires the participation of other effects, such as polar interactions, links between helices, links to extra membrane domains or other proteins, interactions with prosthetic groups, and packing effects. In the following, we summarize experimental approaches to examine some of the propositions advanced in the two-stage model, focusing on experiments with bacteriorhodopsin.

6. BACTERIORHODOPSIN FOLDING IN MEMBRANES

73

II. BACTERIORHODOPSIN Bacteriorhodopsin (BR) is a polytopic membrane protein found in Halobacterium halobium. Light energy absorbed by the BR prosthetic group, retinal, is used to pump protons across the plasma membrane, producing an electrochemical gradient that can serve as an energy source for the organism. Its transmembrane secondary structure was revealed in the mid-1970s to be a set of seven helices (AG ) traversing the lipid bilayer (4,5). Subsequent work has brought the structure to higher resolution (8). The location of the retinal has been established, and the identities of several of the helices with respect to the sequence are now known (9,lO). It appears that an early model for the assignment of helices to densities may prove to be correct (1 1). Chymotrypsin cleaves BR at a single site between residues 71 and 72 of the mature protein (248 residues total). The smaller of the two fragments, C2, corresponds to two helices in the folded structure while the large fragment (Cl) corresponds to five (12). The C l and C2 fragments have been denatured, purified, and reinserted into lipid vesicle bilayers. The reconstitution is done from a solution of one or the other of the fragments with lipid and dodecyl sulfate. Precipitation with potassium removes the dodecyl sulfate, and vesicles are formed that contain the individual fragments. The interactions of the fragments can then be studied by fusion of the vesicles (13). When the fragments are reconstituted in a mixture with each other and with retinal, the original absorption spectrum of BR is regained. It has been found that the molecules thus reconstituted can re-form crystal lattices very similar to those of the purple membrane (9). Thus, the molecules appear to re-fonn a native structure from the fragments, and it is on this observation that the folding experiments described herein are based.

111.

BACTERIORHODOPSIN FRAGMENTS CONTAIN STABLE TRANSBILAYER HELICES

When the two chymotryptic fragments of BR have been reinserted separately into lipid vesicles, they each exhibit circular-dichroism spectra, indicating that they have regained a highly a-helical structure (13). In principle, such structures may be present in arrangements that do not traverse the lipid bilayer, so experiments were conducted (in collaboration with the laboratory of Kenneth Rothschild, Boston University) to use infrared-dichroism measurements as a means of establishing helix orientation. The measurement is based on the fact that the principal infrared transition moments for the peptide bonds in a helix have characteristic orientations, one of which is parallel to and the other perpendicular

74

D. M. ENGELMAN ET AL.

to the helix axis. Thus, in an oriented system, measures of infrared dichroism give the orientation of helices. Lipid vesicles containing one or the other of the fragments were separately dried onto support films to produce oriented multilarnellar arrays of lipid molecules. Preparations of BR molecules were made in a similar way for comparison. It was found that the fragment corresponding to two helices in the native structure is oriented with its helical axes perpendicular to the plane of the lipid bilayer, and that the degree of orientation is about the same as for the average of the helices in the intact bacteriorhodopsin molecule. While the data for the C1 (five helix) fragment were less compelling, a suggestion of similar orientation was seen. Thus, the circular-dichroism and infrared-dichroism measurements establish that the formation of helices by the individual fragments is similar to that of the native molecule: sets of helices are formed that traverse the lipid bilayer even though only a part of the BR molecule is present.

IV.

LINKS AND RETINAL ARE NOT REQUIRED FOR FOLDING

The seven transmembrane helices of the BR structure are linked by segments of polypeptide chain that are organized outside the nonpolar region of the lipid bilayer. It is probable that these links contribute to the stability of the structure, in part by restricting the delocalization of helices in the plane of the bilayer and in part through the formation of organized structures in the link regions. The question that arises in the two-stage model is whether these factors are dominant in dictating the folding of the molecule in going from stage I to stage 11, or whether, perhaps, they are of secondary importance. In this section, we argue that the links, while adding to the stability of the molecule, are in many cases not essential for the molecule to reach a folded state in a lipid bilayer environment. The two fragments that result from chyrnotryptic cleavage of the link between the B and C helices can be separately incorporated into lipid bilayer vesicles, where they form sets of transmembrane helices as mentioned earlier. When such vesicles are mixed and fused by freeze-thawing, the fragments appear to reassociate and provide a binding site for retinal (13). Kinetic analysis of this event suggests that fragment reassociation takes place in the absence of retinal. Until recently, however, a direct demonstration of this association was not available. In recent work, labeling of the fragments with fluorescent groups has been used to show, using energy transfer measurements, that physical reassociation of the fragments occurs in the absence of retinal binding. This work establishes that the link between helix B and helix C is not needed for correct association of fragments. Furthermore, it establishes that such a reassociation is not driven principally by the binding of the retinal prosthetic group. While the detailed

6. BACTERIORHODOPSIN FOLDING IN MEMBRANES

75

structure of the fragments reassociated without retinal is not known, the data strongly suggest that some specific interactions between helices must be occurring, although the retinal may have a role in completing the folded structure (see later). In other work, we have explored the effect of a cleavage between helix E and helix F. Using the V8 protease, Sigrist’s laboratory has produced fragments comprising the first five and last two helices in the sequence, respectively (21). They have used the reconstitution methods developed in our laboratory, and produced a reconstituted molecule that binds retinal. We have examined this reconstituted material, and found conditions in which we can reconstitute the lattice of the purple membrane. X-Ray diffraction measurements from this material show that the structure and molecular interactions of the reconstituted molecule are very similar to those of the native molecule. Thus, like the link between helix B and helix C, the link between helix E and helix F does not appear essential for the folding of the molecule at room temperature. The observations described here support the notion that the second stage of folding is driven more by the tendency of helices to self-associatethan it is by the links between them or by the binding of retinal. This does not imply that retinal binding and helix-helix linkages have no structural role. Recent calorimetric studies conducted in collaboration with the laboratory of Julian Sturtevant show that removing the retinal or cleaving the linkage between helix B and helix C lowers the denaturation temperature of the molecule. At pH 9, removing the retinal lowers the denaturation temperature from 83” to 65°C; cleaving the linkage lowers the temperature from 83”to 78°C. Thus, there are important contributions to the stability of the molecule from retinal binding and from the interhelical link. But, these effects are not the dominant factors in helix-helix association.

V. POLAR INTERACTIONS IN HELIX-HELIX ASSOCIATIONS In earlier discussions, it has been suggested that polar interactions play an important role in producing helix-helix associations in bilayers (6,7,11). The basic argument is that the role of polar interactions in the low dielectric environment of the membrane interior will be greatly enhanced compared with such interactions in an aqueous milieu. These interactions may include factors such as the strongly polar ends produced by helix formation (14), the formation of hydrogen bonds between helices, and ion pairing or strong hydrogen bonding (15). While it is widely discussed, the “helical dipole” may prove to be a weak factor in helix-helix association. It appears that parallel helix association is

D.M. ENGELMAN ET AL.

76

involved in structures such as the dimer of glycophorin molecules (16), and theoretical arguments have been made that solvent exposure and counterion effects will vastly weaken the role of this polarity in helix interactions (15,17). As in the case of the hydrogen bonds stabilizing the helices themselves, hydrogen bonds made between helices in the nonpolar environment of the bilayer will be strong factors that may associate helices. In addition to side-chain groups, carbonyl groups that would be part of the main-chain hydrogen-bonding scheme will be liberated wherever prolines exist. Furthermore, very strong hydrogen bonds may be created when a protonated carboxyl group or a deprotonated amino group is found within the bilayer. These groups may form either strong hydrogen bonds with each other, or possibly ion pairs if the proton is transferred (15). It must be borne in mind, however, that the alternative of creating strong hydrogen bonds, such as the bond between a deprotonated amino group and a hydroxyl group or that between a protonated carboxyl group and a carbonyl group, may be present. Such bonds have much greater energies than the more common hydrogen bonds found in proteins (18). While polar interactions may contribute to the stability of bacteriorhodopsin, the findings from site-directed mutagenesis argue that no single polar interaction is likely to play a key role. Systematic substitutions of many of the polar moieties by nonpolar amino acid side chains have revealed that most can be replaced without serious alteration of the folding or function of the molecule (20). Thus, polar interactions need to be considered in any detailed model of the folding events leading from stage I to stage 11, but it is unclear whether they are essential.

VI.

PACKING EFFECTS

A source of interaction energy that is of some interest and that may prove to be the dominant factor in stage I1 involves packing effects. It is well known that (Y helices can pack well together, and there are a number of structures that show this in a variety of circumstances. The basic mode of such interactions is one in which the protruding side chains of one helix fit into the spaces between side chains on an adjacent helix. It is apparent that the hydrophobic chains of a lipid molecule cannot easily follow the contours of an ci helix in the way that another cx helix can. Thus, it is expected that either voids will be created at the interface between the lipid chains and a helix or side chains will adopt higher energy conformations, folding against the helix to minimize the voids. An estimate has been made that voids in the interior of protein structures can create large energy deficits if a void the size of a side chain is present (19). Thus, the tendency to fill the voids leads to the association of helices in a side-to-side manner, minimizing the lipid-helix interface. A related energy to consider is that the interface may also be unfavorable for a

6. BACTERIORHODOPSIN FOLDING IN MEMBRANES

77

lipid molecule compared to its situation in the bulk bilayer environment. When in a pure lipid environment, a lipid molecule has free diffusion in two dimensions. In contrast, placing such a lipid next to a helix prevents some directions of diffusion. Thus, at least from an entropic standpoint, the lipid molecules are expected to prefer to be associated with other lipid molecules rather than to be located at the protein boundary. Thus, there are two separate aspects of packing that tend to separate regions of transmembrane helix associations from regions of lipid bilayer. Working against these influences are the entropy of mixing, poor detailed packing between helices, and polar repulsion of lipid head groups or helix ends. To explore the possibility that packing may be a major contributor to the stability of a folded protein, we have begun experiments in which small nonpolar molecules are added to bilayers containing BR. At present our finding is that retinal is dissociated from BR by the presence of small molecules such as 2,2dimethylbutane or 2,3-dimethylbutane. When the perturbing small niolecule is removed, retinal is rebound and the correct absorption spectrum is recovered. We are now in the process of determining whether the failure to bind retinal is a result of an unfolding event in which helices come apart from each other or simply an interference through occupancy of the binding site by small inolecules.

VII.

SUMMARY

The idea that membrane protein folding can be considered in two distinct stages is considerably supported by our observations on bacteriorhodopsin. As predicted, substructures of the BR molecule form sets of transmembrane helices. Furthermore, factors in the association of those helices to form tertiary structure have been identified, and it appears that the dominant influence may he packing effects, with comparatively minor contributions from polar interactions, prosthetic group binding interactions, and extramembranous loops, each of which contributes to the detailed structure of the folded molecule or its overall stability. It is hoped that detailed understanding and refinement of these concepts may lead to theoretical notions that will permit the folding of a membrane protein to be predicted on the basis of sequence information. REFERENCES

1 . Deisenhofer, I.,Epp, O., Miki, K . , Huger, R. and Michel, H. (1987). In “The Light Reactions” (J. Barber ed.), pp. 447-493. Elsevier, Amsterdam. 2. Michel, H., Weyer, K . A, , Gmenberg, H . , Dunger, I . , Oesterhelt, D . , and Lootspeich, F. (1986). EMBO J. 5, 1149-1158. 3. Michel, H . , Epp, O., and Delsenhofer, J. (1986). EMBO J . 5, 244-2451. 4 . Henderson, R . , and Unwin, P. N . T. (1975). Narure (London) 257, 28-32. 5. Henderson, R. (1977). Annu. Rev. Biophys. Bioeng. 6, 87-109.

78

D. M. ENGELMAN ET AL.

6. Engelman, D. M., Goldman, A , , and Steitz, T. A. (1982). Merhods Enzymol. 88, 81-88. 7. Engelman, D. M., Steitz, T. A., and Goldman, A. (1986). Annu. Rev. Biophys. Biophys. Chem. 15, 321-353. 8. Tsygannik, I. N., and Baldwin, J. M. (1987). Eur. Biophys. J . 14, 263-272. 9. Popot, J.-L., Trewhella, J., and Engelman, D. M. (1986). EMBO J . 5, 3039-3044. 10. Popot, J.-L., Engelman, D. M., Gurel, O., and Zacca?, G. (1989). J. Mol. Biol. 11. Engelman, D. M., Henderson, R., McLachlan, A. D., and Wallace, B. A. (1980). Proc. Natf. Acad. Sci. U.S.A. 77, 2023-2027. 12. Huang, K.-S., Bayley, H., Liao, M.-J., London, E., and Khorana, H. G. (1981). J . Biol. Chem. 256, 3802-3809. 13. Popot, J.-L., Gerchman, S.-E., and Engelman, D. M. (1987). J . Mol. Biol. 198, 655-676. 14. Hol, W. G. J. (1985). Prog. Biophys. Mol. B i d . 45, 149-195. 15. Honig, B. H., Hubbell, W. L., and Flewelling, R. F. (1986). Annu. Rev. Biophys. Biophys. Chem. 15, 163-193. 16. Bormann, B. J., Knowles, W. J., and Marchesi, B. T. (1989). J. Biol. Chem. 264,4033-4037. 17. Rogers, N. K., and Sternberg, M. J. E. (1984). J. Mol. Biol. 174, 527-542. 18. Allen, L. C. (1975). Proc. Nu& Acud. Sci. U.S.A. 72, 4701-4705. 19. Rashin, A. A., Iofin, M., and Honig, B. H.(1986). Biochemistry 25, 3619-3625. 20. Khorana, H.G. (1988). J . B i d . Chem. 263, 7439-7442. 21. Sigrist, H.,Wenger, R. H., Kislig, E., and Wiithrich, M. (1988). Eur. J. Biochem. 177, 125133.

Part 111

Protein Mobility in Membranes

This Page Intentionally Left Blank

CURRENT TOPICS IN MEMBRANES AND TRANSPORT. VOLUME 36

Chapter 7 Molecular Associations and Membrane Domains MICHAEL EDIDIN Department of Biology The Johns Hopkins University Baltimore, Maryland 21218

I. Models of Membrane Organization A. Fluid Membranes with Component Molecules That Are Highly Mobile B. Spatially Differentiated Membranes-Constraints to Molecular Motion 11. The Study of Large-Scale Molecular Mobility in Cell Surface Membranes A. Fluorescence Photobleaching and Recovery, a Method for Measuring Lateral Diffusion B. Lateral Diffusion of Erythrocyte Band 3: A Model for Organization of Cell Surfaces by a Cytoskeleton C. The Model Suggested by the Erythrocyte Has Been Tested in Other Types of Cells D. Mobile Fraction as an Indicator of Cell Surface Organization 111. A Basis for the Organization of Morphologically Polarized Cell Surfaces IV. Concluding Remarks References

1.

MODELS OF MEMBRANE ORGANIZATION

A. Fluid Membranes with Component Molecules That Are Highly Mobile The fluid properties of membrane bilayer lipids have long dominated our thinking about the organization of cell membranes. The most widely cited model of this organization (Singer and Nicolson, 1972) is termed “fluid mosaic.” It emphasizes the mobility and autonomy of membrane lipids and proteins and implies a role for molecular mobility in membrane function. Indeed, this aspect of the model has been the basis for proposals that reactions between membrane proteins may be coupled by collision of diffusing species (Gupte et a1 , 1984; see Hochman et al., 1985, for a contrary view) and that signaling from the cell surface may require clustering of mobile receptors by antigens or hormones 81 Copynght 0 1990 by Academic Press, Inc All nghu of reprcducuon In any form reserved

82

MICHAEL EDlDlN

(Taylor et al., 1971; Schlessinger, 1988). However, these and other associations between membrane proteins either are transient or are induced by molecules external to the membrane: hormones, antigens, materials of the extracellular matrix.

B. Spatially Differentiated Membranes-Constraints Molecular Motion

to

Though never fully articulated, another model can be made from the large number of experiments that show, not autonomy, but interactions between and organization of membrane molecules on all scales of space and time, from microseconds to thousands of seconds. Membrane lipids interact to form ordered regions, to disrupt lipid order, or even to separate into immiscible regions differing in lipid composition (Small, 1986). Membrane proteins disorder lipids in their immediate vicinity, boundary lipids (Marsh, 1985) and the lipids in turn affect protein function (Stubbs and Smith, 1984; Smith and Stubbs, 1987). Lipids alone may associate preferentially, separating fluid from gel regions, or even creating immiscible regions of fluid lipids. Gel lipids may even be present in cell surface membranes at physiological temperatures (Brasitus and Schachter, 1980; Wolf et al., 1981; Karnovsky et al., 1982; Bearer and Friend, 1980; Edidin and Sessions, 1984; Metcalf et al., 1986; Weaver and Edidin, 1980). Protein-rich domains, with diameters of a micrometer or more, are found in fibroblasts and hepatoma cells (Yechiel and Edidin, 1987; Edidin ef al., 1990). These may be organized by lipids alone, or may be organized by protein-protein interactions. Such interactions play a significant role in constraining diffusion of proteins in the plane of a bilayer (Ryan et al., 1988; Benveniste et al., 1988). Protein complexes are a commonplace of cell surface cell membranes. Particularly good examples are the acetylcholine receptor (AChR) with its five subunits (Claudio et al., 1987), the T-cell receptor for antigen consisting of the antigenspecific Ti a and p chains complexed to five or six other proteins (Samuelson and Klausner, 1988), and the interleukin 2 (IL2) receptor, which has at least three different subunits (Szollosi et al., 1987), two of which are required for high-affinity binding of I L 2 (Smith, 1988). Such complex receptors could, in principle, be formed by aggregation of their components after arrival at the surface membrane. This is not generally the case. The AChR and T-cell receptors are assembled during biosynthesis and transported as completed complexes to the surface. The fluidity of the membrane bilayer and the potential for diffusion of membrane components seem to be resisted rather than utilized in the assembly of complex membrane structures. Cell surfaces are also organized by peripheral proteins. Membrane lipids, glycolipids, and integral proteins interact with, and are immobilized by, extracellular matrix (Spiegel et al., 1984; Foley et al., 1986; Wier and Edidin, 1986)

7. MOLECULAR ASSOCIATIONS AND MEMBRANE DOMAINS

83

and by proteins of the cytoskeleton (Devaux and Signoret, 1985; Marchesi, 1985; Bloch and Morrow, 1989; Rodriguez-Boulan and Nelson, 1989). Such immobilization results in very large-scale differentiation of membranes, creating regions of hundreds to thousands of nanometers in diameter, the composition and function of which differ from that of the remaining membrane. Extreme examples of such domains are the apical and basolateral surfaces of morphologically polarized epithelial cells. Like lipid domains and complex receptors, the large differentiated surfaces of epithelial cells indicate ways of resisting the randomizing and disorganizing effects of translational diffusion in a nominally fluid cell membrane.

II. THE STUDY OF LARGE-SCALE MOLECULAR MOBILITY IN CELL SURFACE MEMBRANES My laboratory has long been interested in the translational diffusion, Dlat,of membrane proteins, especially in the way in which this diffusion is constrained and controlled. In the remainder of this paper I will discuss results with a technique, fluorescence photobleaching and recovery (FPR), which was developed to quantitate Dlatin membranes. After describing FPR and some results on Dlat,I will discuss how this technique for measuring molecular mobility on a scale of micrometers per minute can be used to demonstrate molecular associations and molecular immobilization, that is, the way in which FPR can be used to investigate the spatial and temporal organization of cell surfaces.

A. Fluorescence Photobleaching and Recovery, a Method for Measuring Lateral Diffusion Po0 and Cone (1974) showed that Dlatof vertebrate rhodopsin could be measured by bleaching the rhodopsin in a defined region of disk membranes and following the change in concentration of the unbleached rhodopsin within and outside of the bleached region. By alternately measuring absorbance in bleached and unbleached halves of a stack of disks they could follow return of unbleached molecules to the bleached region and loss of these molecules from the other half of the disks. The Dlatderived from such measurements was 5 x cm2/sec, consistent with the estimated size of rhodopsin and the estimated viscosity of the lipid bilayer (- 100 times that of water). Similar results were obtained by Liebman and Entine (1974). The photobleaching method used to measure D,, of rhodopsin is attractively simple, but of low sensitivity (it measures small changes in absorbance) and restricted to cell membranes containing a bleachable pigment. It was generalized by using exogenous fluorescent labels for proteins and lipids, Peters and co-

84

MICHAEL EDlDlN

workers (1974) first attempted this by covalently labeling erythrocytes with fluorescein, a procedure that mainly labels band 3. An arc lamp was used to bleach half of a labeled ghost, and then an attenuated light source was used to detect recovery of fluorescence in the bleached half of the cell. In any event, no recovery of fluorescence could be detected over many minutes, probably because of a combination of factors, particularly the small Dlatof -3 X 10- l 1 cm2/sec (- 1/ 100 that of rhodopsin) and cross-linking and aggregation of labeled proteins by the rather long period of illumination required for bleaching a large area of the erythrocyte ghost. Despite the failure to measure Dlat,the experiment by Peters and colleagues was important in pointing the way to a generalizedfluorescence photobleaching and recovery method that could be generally applied. The successful FPR (or FRAP, fluorescence recovery after photobleaching) method was developed independently in three laboratories (Edidin et al., 1976; Jacobson et al., 1976; Schlessinger et a l . , 1976; Zagyansky and Edidin, 1976). All three laboratories used a brief pulse from a focused laser beam to bleach a spot in a uniformly fluorescent surface. The initial fluorescence in the spot and the recovery after bleaching are monitored with the laser light attenuated 1000- to 10,000-fold. The FPR method has been used for many different types of fluorescent labels, ranging from lipophilic dyes to covalently bound haptens, to antibody fragments, antibodies, and lectins (for a review of labeling methods see Edidin, 1989). This laser photobleaching method, using blue or green light from a continuous laser output, does not heat the surface (Axelrod, 1977), and at the concentrations of label present in membranes does not aggregate or otherwise modify membrane proteins and lipids (Wolf et al., 1980). The Dlatof fluorescent-labeled visual rhodopsin, measured by FPR, is the same as that found by Po0 and Cone: 5 X cm2/sec (Wey e t a l . , 1981). A scheme of the FPR experiment is shown in Fig 1. Note that the intensity profile of the laser beam is Gaussian, not uniform. This bleaches a Gaussian profile of fluorescence and complicates the quantitative analysis of recovery curves. The theory for the analysis has been worked out (Axelrod et a l . , 1976) and a good approximation to theory has been developed that can be readily implemented on a computer (Barisas and Leuther, 1977). The FPR recovery curve is normalized to the intensity of fluorescence of the spot before bleaching. Two parameters are derived from such a recovery curve. One, R or mobile fraction of label, is given by the ratio of fluorescence after recovery (the asymptote of the recovery curve) to the fluorescence before bleaching. The second parameter, t , l , , is the time for recovery of fluorescence to half its maximum intensity. If the size of the illuminated spot (i.e., the radius of the laser beam) and the depth of bleaching are known, then t,12 can be used to calculate D , , . In my view, R and Dlateach reflect different aspects of the mobility of mole-

7. MOLECULAR ASSOCIATIONS AND MEMBRANE DOMAINS

a5

BLEACHING ENERGY

0

I

r/w

I

-

FLUORESCENCE RECOVERY

FUSION

L= To = i 2 / 4 D

MOBILE FRACTION * A / B

0

2

4

6

8

-‘”

0 1 2

14

t/T

FIG. 1. Schematic diagram of the fluorescence photobleaching and recovery (FPR) experiment. A general view of the cell surface and the focused laser beam is shown at left. The intensity distribution of fluorescence across the spot after bleaching is shown at top right. Greater bleaching results in progressively steeper intensity profiles, but excessive bleaching changes the intensity profile from a Gaussian to a step. The latter distribution will yield incorrect estimates of Qa,.

cules in the plane of a membrane. The value of R is rarely 100% when measured in native rather than in reconstituted (artificial) membranes. Typically it ranges between 30 and 70%. Some part of the immobile label is contributed by cell autofluorescence and, when protein labels are used, by nonspecifically bound label. However, even when this is allowed for, it is clear that a fraction of most membrane molecules, including inositol-lipid-linked proteins and even lipid analogs, is immobile on the time scale of an FPR experiment. This fraction may change with changes in the biology of a cell, for example, differentiation (Salas et af., 1988; Madreperla et al., 1989) or secretion of extracellular materials (Spiegel et al., 1984; Wier and Edidin, 1986). The value of R appears to reflect anchorage of molecules or their confinement in bounded regions or domains (Yechiel and Edidin, 1987). Translational diffusion evaluates constraints to molecules that are not anchored or confined on the scale of an FPR experiment. Although DI,, for rhodopsin is that for a protein diffusing freely in a lipid bilayer, most measured Dlatvalues are smaller than (1/5 to 1/1OOO)this value. Again, some low values of Dlatmay be artifacts, for example, due to cross-linking of membrane proteins by aggregates of a nominally monovalent or divalent label. However, values remain low when

86

MICHAEL EDlDlN

such artifacts are carefully controlled. The variety of constraints to Dlatgives us some insights into the way in which cell surface membranes are organized.

B. Lateral Diffusion of Erythrocyte Band 3: A Model for Organization of Cell Surfaces by a Cytoskeleton The first FPR result (Peters et al., 1974) strongly suggested that translational diffusion of band 3 was highly constrained. Both FPR measurements (Sheetz et al . , 1980; Koppel and Sheetz, 1981) and experiments in which redistribution of fluorescence was followed after fusing labeled and unlabeled ghosts (Fowler and Branton, 1977; Koppel and Sheetz, 1981) estimated Dlatas - l o - ' * cm2/sec. Dlatwas two orders of magnitude higher in ghosts from spectrin-deficient mice (Sheetz et al., 1980). It also increased by this factor when proteins were dissociated from labeled normal ghosts (Fowler and Bennett, 1978; Golan and Veatch, 1980). These data implied that the movement of band 3 in the plane of the membrane was hindered by the (spectrin-containing) cytoskeleton of the erythrocyte, but that band 3 molecules were not immobilized by the cytoskeleton. Measurements of rotational diffusion of band 3 (Nigg and Cherry, 1979) showed that it appeared to be a freely rotating dimer. The two sets of data were reconciled by a model in which a spectrin meshwork caged molecules of band 3 into small regions, 100 nm on a side. The molecule can rotate freely within such regions, but can only move in the plane of the membrane where the spectrin net is open and incomplete. Such percolation would be the basis for the diffusion observed, which would in fact be a composite of a DIa,and the time taken for the spectrin barriers to open sufficiently. An analysis of this model by Saxton (1989) includes a good summary of the subject.

C. The Model Suggested by the Erythrocyte Has Been Tested in Other Types of Cells 1 . EVIDENCE AGAINST CYTOSKELETAL CONTROL OF

TRANSLATIONAL DIFFUSION The results just cited suggested that a cytoskeleton might hinder translation diffusion in cells other than erythrocytes. Two lines of evidence indicate that this is not so. First, disruption of the actin cytoskeleton with cytochalasins, or of the spectrin network by microinjection of antispectrin antibodies, does not alter Dlat (Schlessinger et al., 1976; Jacobson et al., 1987). Second, and apparently more compelling, DIatof three different mutant membrane proteins that have shortened cytoplasmic regions does not differ from that of the wild-type proteins (Edidin and Zuniga, 1984; Livneh et al., 1986; Scullion et al., 1987).

87

7. MOLECULAR ASSOCIATIONS AND MEMBRANE DOMAINS

This evidence requires that we look elsewhere for constraints to Dlat cells (fibroblasts), in which the cytoskeleton does not seem to interact strongly with diffusing molecules. We have shown that the extracellular bulk of membrane proteins may hinder their diffusion (Wier and Edidin, 1988). The D,,,of mutant membrane glycoproteins lacking one or more oligosaccharideunits, or otherwise reduced in bulk (by deletion of protein domains), approaches that found for rhodopsin. This effect is shown in Fig. 2. Inhibition of glycosylation by drugs also increases Dlat, though to a lesser extent (M. Wier and M. Edidin, unpublished observations). Scullion et al. (1987) have also found effects of glycosylation on Dlatof vesicular stomatitis virus (VSV) G proteins, although both deletion and addition of glycosylation sites produced increases in Dlat. These data, together with other information on the effects of surface protein concentration on mobility in membranes (Small et al., 1984; Ryan et al., 1988; Benveniste et al., 1988), give us a picture of the cell surface membrane as crowded with proteins, which impede each other's mobility across the surface. Such a picture does not accord with the usual cartoon of a membrane including large protein-free regions. However it is quite consistent with the known protein/lipid ratio of most membranes.

2. EVIDENCE FOR CYTOSKELETAL CONTROL OF TRANSLATIONAL DIFFUSION IN NUCLEATED CELLS Experiments on D,, in membrane blebs led to conclusions opposite to those implied in the previous section. Tank et al. (1982) showed that diffusion of both

i 0

1

2

3

Units Deleted

FIG.2. Effect of glycosylation on D,,, of class I MHC antigens. It can easily be seen that Dla, increases with loss of oligosaccharide units. (After Wier and Edidin, 1988.)

88

MICHAEL EDlDlN

proteins and lipid probes increased by orders of magnitude in membrane blebs that formed at the surfaces of cultured myotubes treated with dilute formaldehyde and a reducing agent. Similar results were found for blebs on lymphocytes (Wu et al., 1982). The blebs appear empty, and do not stain for F-actin. It would seem that diffusion changes because their membranes are no longer in contact with the cytoplasm. The results cited in Section II,C, 1 show that the bulk and crowding of proteins may dominate constraints to Dlat,but they should not be interpreted to mean that these constraints are the most important in all cell types. Note that they are for cells in which there is little or no functional coupling of surface molecules to cytoplasm. For example, the results on class I major histocompatibility complex (MHC) antigens of L cells describe the antigens’ behavior in a cell in which they are only slowly internalized and in which they do not respond to added phorbol esters (Capps et al., 1989). We have found that class I1 MHC antigens of B cells do interact with the cytoplasm. Class I1 MHC antigens lacking portions of their cytoplasmic domains have larger Dlatvalues than wild-type antigens (Wade et al., 1989b). The interaction here is functional, since cells expressing the mutant molecules are also defective in their responses to phorbol esters (Wade et al., 1989a). It will be worth looking at Dlatand the functional behavior of a variety of mutant, truncated, membrane proteins in cells in which they ought to be functional. For example, mutant class I MHC antigens ought to be studied, not in L cells as they have been, but in T cells in which they are functionally internalized and cycled back to the surface.

D. Mobile Fraction as an Indicator of Cell Surface Organization The mobile fraction, R , of a fluorescent label has proved to be an important parameter for characterizing cell surfaces. I will summarize our work on R in three different cases, each of which adds to our information about the way in which cell surfaces are arranged.

1. A FLUORESCENCE PHOTOBLEACHING AND RECOVERY METHOD FOR MEASURING MOLECULAR PROXIMITY We have commented already that many surface receptors are compounds of several different kinds of chains and that they are often further complexed with molecules such as class I MHC antigens, which have no obvious role in receptor function. While such compounds may be formed and have been shown by imrnunoprecipitation, the affinity of their component chains at times is not high enough to maintain the associations during the extensive washing of immunoprecipitates.

7. MOLECULAR ASSOCIATIONS AND MEMBRANE DOMAINS

89

Szollosi et al. (1987) found that a 96-kDa peptide, termed T27, was sometimes found in immunoprecipitates of the IL-2 receptor Tac peptide. Fluorescence measurements showed that T27 was in molecular proximity to the IL-2 receptor on the cell surface and also suggested that class I MHC antigens were near the receptor. We found that the anti-T27 monoclonal antibody, OKT27, immobilized the 95-kDa peptide. Among 26 cells labeled with OKT27, 21 showed no recovery of fluorescence at all. We reasoned that if T27 was in fact associated with Tac peptide, then immobilizing T27 ought to affect the mobility of Tac. This was in fact the case. Tac peptide, labeled with fluorescent antibody, was totally immobilized in nearly one-third the cells examined (Edidin er al., 1988). A further unexpected result was that monoclonal antibody against class I MHC antigens also affected translational diffusion of Tac, but this effect was seen as a reduction in Dlat,rather than in mobile fraction. Reductions in Dlatsuggest transient associations of HLA antigens with the IL-2 receptor, while the immobilization of T27 could be due to extensive aggregation or to aggregation-induced anchoring to the cytoskeleton. A precedent for the latter is found in the effects of antiglobulin on anchoring of slg to the actin cytoskeleton of B cells (Woda and Woodin, 1984). Preliminary work on the relationship between insulin receptors and class I MHC antigens of human B lymphoblasts also shows that cross-linking one species of molecule affects diffusion of the other. HLA antigens are efficiently cross-linked, and this effects a reduction in R of labeled insulin receptors. Insulin receptors are not as thoroughly cross-linked by our monoclonal antibody, and this is seen as a reduction in Dlatof HLA antigens (Reiland and Edidin, L987). The relationship is specific in that cross-linking of sIg does not affect Dlat or R of either insulin receptor or MHC antigens. A schematic view of this experiment is shown in Fig. 3. These two examples show that FPR can be used as a qualitative tool for detecting molecular associations in surfaces. A minimum of two cross-linking antibodies is required, and ideally a third, irrelevant monoclonal antibody should be available to show that effects of surface cross-linking are specific for the pair of molecules of interest. If the experiment is properly constructed, membrane organization can be resolved on the scale of pairs of interacting molecules. PHOTOBLEACHING AND RECOVERY 2. FLUORESCENCE EXPERIMENTS DETECTMICROMETER-SCALE MEMBRANE DOMAINS The geometry of a photobleaching experiment, the size of the spot bleached, will affect t,12, but should not alter R unless the spot is so large that a significant fraction of the whole membrane is bleached (see Fig. 1). However, Yechiel and Edidin (1987) showed that this is true for diffusion in large liposomes, but is not the case for diffusion in cell membranes. The R value of both lipid and protein

90

MICHAEL EDlDlN

FIG. 3. Cross-linking insulin receptors may effect diffusion of class I MHC antigens if these are associated with the insulin receptors. IR, Insulin receptor; Ab, antireceptorantibody cross-linking the insulin receptors; fl-MHC, class I MHC antigens labeled with a fluorescent Fab fragment. It can be seen that cross-linking receptors will affect MHC antigens to varying extents depending on the efficiency of cross-linking and the extent of association of IR and MHC.

labels of human fibroblasts is a function of bleaching spot size. Such data are shown in Fig. 4. These, together with data on Dlat,lead to a model of the fibroblast surface as divided into protein-rich domains, separated by a lipid-rich continuum. Figure 5 shows the way in which the geometry of photobleaching such a membrane would affect the mobile fraction. We can scale the size of these domains from the known sizes of the laser spots used to measure diffusion. We estimate that they are several micrometers in diameter. Photographs of large spots on cell surfaces labeled with a fluorescent lipid analog show patterns of bright and dark spots consistent with our model. The domains detected in these experiments may not be the same as those differentiated by the behavior of the lipid probe diI (Klausner and Wolf, 1980; Wolf et al., 1981, 1988; Weaver and Edidin, 1990). The mobile fraction of diI does not change with spot size, even in native membranes. The difference between the behavior of the two probes could reflect different domains that are differently coupled to the inner bilayer leaflet and to the cytoskeleton.

111. A BASIS FOR THE ORGANIZATION OF MORPHOLOGICALLY POLARIZED CELL SURFACES We have so far described membrane organization on two scales: that of molecular associations (nanometers to tens of nanometers) and that of regions of high lipid or protein concentration (hundreds to thousands of nanometers). A still

7. MOLECULAR ASSOCIATIONS AND MEMBRANE DOMAINS

91

100

n

P 0

l/e2 r a d i u s

(pm)

FIG. 4. Change in fraction of mobile molecules ( R ) as the size of the spot bleached is increased. Human skin fibroblasts were labeled with either NBD-6 PC (A), a lipid analog, or with fluorescent (Data replotted from Yechiel and Eididin, 1987.) Fab fragment-labeled proteins (0).

larger scale organization is seen in the surfaces of morphologically polarized epithelial cells (Ahlmers and Stirling, 1983; Le Grimellec er al., 1988; Rodriquez-Boulan and Nelson, 1989). Their apical surfaces are morphologically, functionally, and biochemically distinct from their basolateral surfaces. The lifetime of such cells ranges from days in the intestine to years in the liver. Translational

FIG. 5 . A domain model consistent with the results of Fig. 4. The open circles represent domains, regions closed to entry and exit of either lipids or proteins on a time scale of minutes. These domains are probably enriched for membrane proteins. The region between them is lipid-rich and relatively poor in proteins. Small bleaching spots will bleach a fraction of a domain or a very small fraction of the region between domains. Large spots may bleach entire domains, and such bleached domains cannot contribute to recovery of fluorescence.

92

MICHAEL EDlDlN

diffusion of surface membranes and lipid must rigorously be resisted if the surfaces of epithelial cells are to remain differentiated (Wolf, 1987). Though experiments on mechanisms for this control have only begun, we have some indications that different mechanisms may operate for apical and basal surfaces and that, in the latter, a spectrin-containing cytoskeleton is important. The first attempts to determine mechanisms of organization of epithelial cell surfaces showed that proteins of the apical surface redistributed after the cells were dissociated (Pisam and Ripoche, 1976; Ziomek et al., 1980). This suggested that the proteins were free to diffuse within the apical surface, but that they were kept from entering the basal surface by the barrier of the tight junctions. Experiments on the fate of fluorescent lipid labels added to the apical surface of cultured kidney epithelial cells reinforced this view. Labels shown by quenching experiments to localize in the outer leaflet of the bilayer did not redistribute after insertion in the apical surface. Labels in the inner leaflet were redistributed (Dragsten ef al., 1981). This and other experiments (e.g., van Meer and Simons,1986) are consistent with a role for the tight junctions in restricting both proteins and lipids to apical or basal domains, leaving them free to diffuse within a domain. Measurements by the FPR technique on marker antigens of differentiated MDCK cells show that indeed apical markers are free to diffuse, and are not anchored to the cytoskeleton, being readily extractable by detergents (Salas et al., 1988). In contrast, a significant fraction of basal marker antigens is immobile, and a similar fraction resists extraction in detergent. The shift in mobility and extractability of the basal markers can be followed as the MDCK cells polarize in culture. It appears that in the basal surface many molecules of a given protein are anchored to a cytoskeleton. Creation of a spectrin cytoskeleton is a major feature of epithelial cell differentiation, and this could have an important role in the anchoring observed. Spectrin certainly plays such a role in the surface differentiation of chicken photoreceptors. The surface Na ,K+ -ATPase of mature photoreceptors is localized to one region of the surface. It is uniformly distributed in the precursors. This surface distribution is matched precisely by the distribution of cytoplasmic spectrin. Most of the spectrin and most of the Na+ ,K -ATPase is extractable from unpolarized cells and most resists extraction from polarized cells. In FPR the mobile fraction of ATPase is also reduced to the same extent as the cells differentiate. Here, too, it seems that anchoring to or through spectrin keeps a surface protein appropriately localized (Madreperla et al., 1989). However, 100% immobile protein is never seen; on the average 30% of the Na+ ,K+ -ATPase is mobile in differentiated cells. Perhaps the small fraction of mobile molecules is in a metabolic pool of newly arriving or newly departing molecules. It may also be that anchoring is inefficient and a fraction of ATPase molecules diffuse away from the site of their insertion to a sink where they are internalized and metabolized. +

+

7. MOLECULAR ASSOCIATIONS AND MEMBRANE DOMAINS

93

IV. CONCLUDING REMARKS The FPR technique, which measures lateral mobility of membrane lipids and proteins, can be readily used to show that the cell surface is “grainy,” differentiated into regions of specialized composition and function. The scale of these regions runs from nanometers to tens of micrometers, and the mechanism for local differentiation seems to vary with scale. It will be interesting to study further the very largest domains, the specialized apical and basal surfaces of polarized cells, to see if they in turn are composed of smaller domains. It will be more interesting and more important to come to an understanding of the basis for and the function of these regions of a membrane, which are far from a random collection of proteins embedded in a random collection of lipids. REFERENCES Ahlmers, W., and Stirling, C. (1984). Distribution of transport proteins over animal cell membranes. J . Membr. Biol. 77, 169-186. Axelrod, D. (1977). Cell surface heating during fluorescence photobleaching recovery experiments. Biophys. J . 18, 129-131. Axelrod, D., Koppel, D. E., Schlessinger, J., Elson, E., and Webb, W. W. (1976). Mobility measurement by analysis of fluorescence photobleaching recovery kinetics. Biophys. J . 16, 1055-1069. Barisas, B. G., and Leuther, M. D. (1977). Fluorescence photobleaching recovery measurement of protein absolute diffusion constants. Biophys. Chem. 10, 221-229. Bearer, E. L., and Friend, D. S. (1980). Anionic lipid domains: Correlation with functional topography in a mammalian cell membrane. Proc. Natl. Acad. Sci. (I.S.A. 77, 6601-6605. Benveniste, M., Livneh, E., Schlessinger, J., and Kam, Z. (1988). Overexpression of epidermal growth factor receptor in NIH-3T3-transfected cells slows its lateral diffusion and rate of endocytosis. J . Cell Biol. 106, 1903-1909. Bloch, R. J., and Morrow, J. S. (1989). An unusual beta-spectrin associated with clustered acetylcholine receptors. J . Cell Biol. 108, 481-493. Brasitus, A., and Schachter, D. (1980). Lipid dynamics and lipid-protein interactions in rat enterocyte basolateral and microvillus membranes. Biochemistly 19, 2763-2769. Capps, G. G . , van Kampen, M., Ward, C. L., and Zuniga, M. C. (1989). Endocytosis of the class I major histocompatibility antigen via a PMA-inducible path is a lymphocyte-specific phenomenon and requires the cytoplasmic domain. J . Cell Biol. 108, 1317-1329. Claudio, T.,Green, W. N . , Hartman, D. S . , Hayden, D., Paulson, H. L., Sigworth, F. J., Sine, S . M., and Swedlund, A. (1987). Genetic reconstitution of functional acetylcholine receptor channels in mouse fibroblasts. Science 238, 1688-1694. Devaux, P. F., and Signoret, M. (1985). Specificity of lipid-protein interactions as determined by spectroscopic techniques. Biochim. Biophys. Acta 822, 63- 125. Dragsten, P., Blurnenthal. R., and Handler, J . S. (1981). Membrane asymmetry in epithelia: Is the tight junction a barrier to diffusion in the plasma membrane? Nature (London) 294, 718-722. Edidin, M. (1989). Fluorescent labeling of cell surfaces. Methods Cell Biol. 29, 87-102. Edidin, M . , and Sessions, A . (1984). Heterogeneity in the plasma membrane lipidos of eukaryotic cells. Ann. N.Y. Acad. Sci. 414. 8-18.

94

MICHAEL EDlDlN

Edidin, M., and Zuniga, M. (1984). Lateral diffusion of wild-type and mutant Ld antigen in L cells. J . Cell Biol. 99, 2333-2335. Edidin, M., Zagyansky, Y., and Lardner, T. J. (1976). Measurement of membrane protein lateral diffusion in single cells. Science 191, 466-468. Edidin, M., Soloski, M., and Stroynowski, I. (1988). Inositol-linked membrane proteins, class I MHC antigens, do not localize to membrane domains in the same manner as other MHC antigens. J . Cell Biol. 107, 783a. Foley, M., McGregor, A. N., Kusel, J. R . , Garland, P. B . , Downie, T., and Moore, I . (1986). The lateral diffusion of lipid probes in the surface membrane of Schistosoma mansoni. J . Cell Biol. 103, 807-818. Fowler, V., and Bennett, V. (1978). Association of spectrin with its membrane attachment site restricts lateral mobility of human erythrocyte integral membrane proteins. J . Supramol. Struct. 8, 215-221. Fowler, V., and Branton, D. (1977). Lateral mobility of human erythrocyte integral membrane proteins. Nature (London) 268, 23-26. Golan, D., and Veatch, W. (1980). Lateral mobility of band 3 in the human erythrocyte membrane studied by fluorescence photobleaching recovery: Evidence for control by cytoskeletal interactions. Proc. Natl. Acad. Sci. U.S.A. 77, 2537-2541. Gupte, S., Wu, E.-S., Hoechli, L., Hoechli, M., Jacobson, K., Sowers, A. E., and Hackenbrock, C. R. (1984). Relationship between lateral diffusion, collision frequency, and electron transfer of mitochondria1 inner membrane oxidation-reduction components. Proc. Natl. Acad. Sci. U.S.A. 81, 2606-2610. Hochman, J., Ferguson-Miller, S., and Schindler, M. (1985). Mobility in the mitochondria1 electron transport chain. Biochemistry 24, 2509-2516. Jacobson, K., Wu, E., and Poste, G. (1976). Measurement of the translational mobility of concanavalin A in glycerol-saline solutions and on the cell surface by fluorescence recovery after photobleaching. Biochim. Biophys. Acta 433, 215-222. Jacobson, K., Ishihara, A., and Inman, R . (1987). Lateral diffusion of proteins in membranes. Annu. Rev. Physiol. 49, 163-175. Karnovsky, M. J . , Kleinfeld, A. M., Hoover, R. L., and Klausner, R . D. (1982). The concept of lipid domains in membranes. J . Cell Biol. 94, 1-6. Klausner, R . D., and Wolf, D. E. (1980). Selectivity of fluorescent lipid analogues for lipid domains. Biochemistry 19, 6199-6203. Koppel, D. E., and Sheetz, M. F. (1981). Fluorescence photobleaching does not alter the lateral mobility of erythrocyte membrane glycoproteins. Nature (London) 293, 159- 161. Le Grimellec, C . , Friedlander, G., and Giocondi, M.-C. (1988). Lipid asymmetry and transport function in renal epithelial cells. NIPS 3, 227-229. Liebman, P. A , , and Entine, G. (1974). Lateral diffusion of visual pigment in photoreceptor disk membranes. Science 185, 457-459. Livneh, E., Benveniste, M., Prywes, R., Felder, S., Kam, Z., and Schlessinger, J. (1986). Large deletions in the cytoplasmic kinase domain of the epidermal growth factor do not affect its lateral mobility. J . Cell Biol. 103, 327-331. Madreperla, S. A., Edidin, M., and Adler, R . (1989). Na+ .K+-adenosine triphosphate polarity in retinal phosphoreceptors: A role for cytoskeletal attachments. J . Cell Biol. 109, 1483- 1493. Marchesi, V. T. (1985). Stabilizing infrastructure of cell membranes. Annu. Rev. Cell Biol. 1, 531561. Marsh, D. (1985). I n “Progress in Protein-Lipid Interactions” (A. Watts and J. J. H . H . M. de Pont, eds.), Vol. 1, pp. 143-172. Elsevier, Amsterdam. Metcalf, T. N., Ill, Wang, J. L., and Schindler, M. (1986). Lateral diffusion of phospholipids in the

7. MOLECULAR ASSOCIATIONS AND MEMBRANE DOMAINS

95

plasma membrane of soybean protoplasts: Evidence for membrane lipid domains. Proc. Natl. Acad. Sci. U.S.A. 83, 95-99. Nigg, E. A., and Cherry, R. I. (1979). Dimeric association of band 3 in the erythrocyte membrane demonstrated by protein diffusion measurements. Narure (London) 277, 493-494. Peters, R.,Peters, J., Tews, K. H., and Bahr, W. (1974). A microfluorimetric study of translation diffusion in erythrocyte membranes. Biochim. Biophys. Acra 367, 282-294. Pisam, M., and Ripoche, P. (1976). Redistribution of surface macromolecules in dissociated epithelial cells. J. Cell Biol. 71, 907-920. Poo, M. M., and Cone, R. A. (1974). Lateral diffusion of rhodopsin in the photoreceptor membrane. Nature (London) 247, 438-441. Reiland, J., and Edidin, M. (1987). Crosslinking of insulin receptor affects MHC antigen mobility. Fed. Proc., Fed. Am. Soc. Exp. Biol. 46, 1465. Rodriguez-Boulan, E., and Nelson, J. W. (1989). Morphogenesis of the polarized epithelial cell phenotype. Science 245, 718-725. Ryan, T. A,, Myers, J., Holowka, D., Baird, B., and Webb, W. W. (1988). Molecular crowding on the cell surface. Science 239, 61-64. Salas, P. J. I., Vega-Salas, D. E., Hochman, J., Rodriguez-Boulan, E., and Edidin, M. (1988). Selective anchoring in the specific plasma membrane domain: A role in epithelial cell polarity. J . Cell Biol. 107, 2363-2376. Samuelson, L. E., and Klausner, R. D. (1988). The T-cell antigen receptor. Structure and mechanism of activation. Ann N.Y. Acad. Sci. 540, 1-3. Saxton, M. (1989). The spectrin network as a barrier to lateral diffusion in erythrocytes. Biophys. J . 55, 21-28. Schlessinger, J. (1988). The epidermal growth factor receptor as a multifunctional allosteric protein. Biochemistry 27, 3119-3123. Schlessinger, J., Koppel, D. E., Axelrod, D., Jacobson, K., Webb, W. W., and Elson, E..L. (1976). Lateral transport on cell membranes: Mobility of concanavalin A receptors on myoblasts. Proc. Natl. Acad. Sci. U.S.A. 73, 2409-2413. Scullion, B. F., Jou, Y., Puddington, L., Rose, J. K., and Jacobson, K. (1987). Effects of mutations in three domains of the vesicular stomatisis viral glycoprotein on its lateral diffusion in the plasma membrane. J . Cell Biol. 105, 69-75. Sheetz, M. P., Schindler, M., and Koppel, D. E. (1980). Lateral mobility of integral membrane proteins is increased in spherocytic erythrocytes. Nature (London) 285, 5 10-5 12. Singer, S. J., and Nicolson, G. L. (1972). The fluid mosaic model of the structure of cell membranes. Science 175, 720-73 1. Small, D. M. (1986). The physical chemistry of lipids from alkanes to phospholipids. In “Handbook of Lipid Research” (D. Hanahan, ed.). Plenum, New York. Small, R., Blank, M., and F‘fenninger, K. H. (1984). Components of the plasma membrane of growing axons. 11. Diffusion of membrane protein complexes. J. Cell Biol. 98, 1434-1443. Smith, A. D., and Stubbs, C. D. (1987). Modulation of membrane protein by bilayer lipids. Basic Res. Cardiol. 82, 93-97. Smith, K. A. (1988). Interleukin-2 Inception, impact, and implications. Science 240, 1169-1 176. Spiegel, S . , Schlessinger, J., and F shman, P. (1984). Incorporation of fluorescent gangliosides into human fibroblasts: Mobility, fate and interaction with fibronectin. J . Cell Biol. 99, 699-704. Stubbs, C. D., and Smith, A. D. (1984). The modification of mammalian polyunsaturated fatty acid composition in relation to membrane fluidity and function. Biochirn. Biophys. Acra 779, 89137. Szollosi, J., Damjanovich, S., Goldman, C. K., Fulwyler, M., Aszalo, A. A,, Goldstein, G., Rao, P., Talk, M. A., and Waldmann, T. A. (1987). Flow cytometric resonance energy transfer

96

MICHAEL EDlDlN

measurements support the association of a 95-kDa peptide termed T27 with the 55-kDa Tac peptide. Pror Natl. Acad. Sci U.S.A. 84, 7246-7250. Tank, D. W., Wu, E.-S., and Webb, W. W. (1982). Enhanced molecular diffusibility in muscle membrane blebs: Release of lateral constraints. J . Cell Biol. 92, 207-212. Taylor, R. B . , Philip, W., Duffus, H., Raff, M. C., and de Petris, S . (1971). Redistribution and pinocytosis of lymphocyte surface immunoglobulin molecules induced by anti-immunoglobulin antibody. Nature (London), New Biol. 233, 225-229. van Meer, G . , and Simons, K . (1986).The function of tight junctions in maintaining differences in lipid composition between the apical and the basolateral cell surface domains of MDCU cells. EMBO J . 5 , 1455-1464. Wade, W. F., Chen, Z. Z . , Maki, R., McKercher, S . , Palmer, E., Cambier, J. C . , and Freed, J. H. (1989a). Altered I-A-mediated transmembrane signaling in B cells that express truncated [-Ak protein. Proc. Natl. Acad. Sci. U.S.A. 86, 6297-6301. Wade, W. F., Freed, J. H., and Edidin, M. A. (1989b). Translational diffusion of class I1 major histocompatibility complex molecules is constrained by their cytoplasmic domains. J . Cell. Biol. 109, 3325-3331. Weaver, F. E., and Edidin, M. (1990). Effects of temperature and lipid composition on the organization of sea urchin egg plasma membranes. In preparation. Wey, C.-L., Edidin, M. A,, and Cone, R. A. (1981). Lateral diffusion of rhodopsin in photoreceptor cells measured by fluorescence photobleaching and recovery (FPR). Biophys. J . 33, 225-232. Wier, M., and Edidin, M. (1986). Effects of cell density and extracellular matrix on lateral diffusion of MHC antigens in cultured fibroblasts. J . Cell Biol. 103, 212-222. Wier, M., and Edidin, M. (1988). Constraint of the translational diffusion of a membrane glycoprotein by its external domains. Science 242, 412-414. Woda, B . A., and Woodin, B. (1984). The interaction of lymphocyte membrane proteins with the lymphocyte cytoskeletal matrix. J. Immunol. 133, 2767-2772. Wolf, D. E. (1987). Overcoming random diffusion in polarized cells-corralling the drunken beggar. Bio. Essays 6, 116-121. Wolf, D. E., Edidin, M., and Dragsten, P. R. (1980). Effect of bleaching light on measurements of lateral diffusion in cell membranes by the fluorescence photobleaching and recovery method. Proc. Nurl. Acad. Sci. U.S.A. 77, 2043-2045. Wolf, D. E., Edidin, M., and Handyside, A. H. (1981). Changes in the organization of the mouse egg plasma membrane upon fertilization and first cleavage: Indications from the lateral diffusion rates of fluorescent lipid analogs. Dev. Biol. 85, 195-198. Wolf. D. E., Lipscomb, A. C., and Maynard, V. M. (1988). Causes of nondiffusing lipid in the plasma membrane of mammalian spermatozoa. Biochemistry 27, 861-865. Wu, E . - S . ,Tank, D. W., and Webb, W. W. (1982). Unconstrained lateral diffusion of concanavalin A receptors on bulbous lymphocytes. Proc. Natl. Acad. Sci. U.S.A. 79, 4962-4966. Yechiel, E., and Edidin. M. (1987). Micrometer-scale domains in fibroblast plasma membranes. J . Cell Biol. 105, 755-760. Zagyansky, Y., and Edidin, M. (1976). Lateral diffusion of concanavalin A receptors in the plasma membrane of mouse fibroblasts. Biochim. Biophys. Acra 433, 209-214. Ziomek, C. A., Schulman, S . , and Edidin, M. (1980). Redistribution of membrane proteins in isolated mouse intestinal epithelial cells. J . Cell Biol. 86, 849-857.

CURRENT TOPICS IN MEMBRANES AND TRANSPORT, VOLUME 36

Chapter 8 Actin-Membrane Interactions in Eukaryotic Mammalian Cells THOMAS P . STOSSEL Department of Medicine Harvard Medical School Massachusetts General Hospital Boston, Massachusetts 02114

I.

The Actin System and Membrane Function The Erythrocyte Cytoskeleton: Paradigm or Distraction in Approaching the Interaction between the Eukaryote Plasma Membrane and Actin System? 111. The Actin System in Membrane Stabilization and Retraction IV. The Actin System in Membrane Propulsion V. The Actin System and Membrane Propulsion and Retraction References 11.

1.

THE ACTIN SYSTEM AND MEMBRANE FUNCTION

Actin is one of the major cytoskeletal polymer systems, the others being the microtubules and intermediate filaments. Much evidence attests to the role of actin in the control of shape and functions such as locomotion, exocytosis, phagocytosis, and cytokinesis. Actin assembles reversibly into linear polymers, and a large number of actin-binding proteins regulate this assembly and influence the architecture of actin filaments to produce a wide variety of three-dimensional structures. Actin and this family of actin-binding proteins constitute “the actin system.” Actin tends to concentrate in the periphery of cells in general and is therefore close to the plasma membrane, the shape of which it presumably influences and where it in turn can be regulated by signals generated by interaction of membrane receptors with extracellular ligands. The question as to how 97 Copynghl 0 1990 by Academic Press. Inc. All rights of reproduction in m y form reserved

98

THOMAS P. STOSSEL

these interactions take place between the actin system and the plasmalemma is the basis of an interesting research program (Schliwa, 1986).

II. THE ERYTHROCYTE CYTOSKELETON: PARADIGM OR DISTRACTION IN APPROACHING THE INTERACTION BETWEEN THE EUKARYOTE PLASMA MEMBRANE AND ACTIN SYSTEM? The best characterized membrane cytoskeleton is the protein network that laminates the membrane of the red blood cell (Bennett, 1985). Although spectrin, a rodlike protein that assembles into tetramers, constitutes the principal mass of protein in this network, actin oligomers (along with other proteins) link the spectrin molecules, and therefore this cytoskeleton can be considered to be a member of the actin system. Relatively high-affinity associations exist between this network and transmembrane proteins such as band 3, thereby establishing direct linkage between the membrane bilayer and the underlying cytoskeleton. Proteins resembling the components of the erythroid cytoskeleton have been detected in nonerythroid cells. These molecules, which bear names such as fodrin and TW (for “terminal web”) 260/240, contain epitopes immunoreactive with spectrin and resemble erythroid spectrin in size and shape. These properties, as well as the presence in nonerythroid cells of proteins resembling ankyrin (the link between erythrocyte spectrin and band 3) in addition to other erythroid protein analogs, have led researchers to search for a “membrane skeleton” analogous to the spectrin-actin shell in eukaryotic cells. How applicable, however, is the erythrocyte model to other cell types? It has long been known that the physiology of red cell membranes differs from that of other cells. Furthermore, logic implies that considerable variation might accompany the mechanics underlying the variety of cell shapes, degree of motility, and particularity of surface specializations in different cell types. It is not surprising, therefore, that while one finds spectrin or spectrin analogs at the surface of red cells and certain epithelial cells in abundance (e.g., Nelson and Veshnock, 1986, 1987), a spectrin network resembling the submembrane lamina of the erythrocyte has not yet been observed morphologically, and the spectrinlike molecules of intestinal brush border epithelial cells evidently reside deep in the cytoplasm beneath the surface where they connect microvillus actin bundles (Mooseker, 1985). An important distinction between red cells and many eukaryotes is the presence in the latter of a three-dimensional actin network that extends for variable distances but as much as several hundred nanometers away from the plasma membrane and is presumed to confer stability on the cell cortex and to explain its organelle exclusion, and rearrangements of this actin network are proposed to generate cell shape changes and mobility.

8. ACTIN-MEMBRANE INTERACTIONS IN CELLS

111.

99

THE ACTIN SYSTEM IN MEMBRANE STABILIZATION AND RETRACTION

The construction of the erythrocyte cytoskeleton is suited for stabilizing a membrane that is subjected to passive deformation. The red cell mernbrane, so far as we know, however, is not usually influenced by internal retractive or propulsive events. Similar considerations may apply to certain tissue cells such as neurons, which do not move or divide but are subjected to external bending or compression and which feature cytoskeletons analogous to the erythrocyte. Many cells, on the other hand, demonstrate deformations that strongly imply internal contractility. Certainly the actin filaments of striated muscle cells and syncytia resist being tom from their insertions by powerful contractions; yet the molecules mediating the attachment of actin filaments at these insertions are not clearly identified. Cell division requires cleavage furrowing, a process proven to involve myosin-based contraction of actin filaments attached to the plasma membrane at the equator of the dividing cell, yet there is no clear evidence that spectrin or related proteins mediate this attachment. Many cells retract extended pseudopodia and some locomote actively and phagocytose, events associated with membrane puckering that resembles a contractile ring (Singer and Kupfer, 1986). When actin-binding myosin fragments are used to identify actin filaments in cells by the characteristic arrowhead configuration that they confer upon actin filaments, actin filaments with arrowhead polarity pointing away from the membrane have most frequently been reported. This finding was conceptually gratifying, because the direction of pull of myosin on such filaments would be away from the membrane, consistent with a mechanism resembling the muscle sarcomere for contraction of membranes toward the center of the cell. Such a contraction of the plasma membrane almost certainly requires that submembrane actin filaments acquire a purchase on the membrane, and reasonably high-affinity bonds between membrane molecules and actin filaments seem to be a foregone conclusion. If spectrin and its associated proteins do not mediate these attachments, what molecules do? Several avenues of investigation have sought the answer to this question (reviewed in Niggli and Burger, 1987). One has been to use immunocytochemistry to map the location of actin and other proteins with respect to membrane ligandreceptor complexes on the cell surface or to points where the cell adheres to its substrate. Such studies have repeatedly revealed close associations between the ends of actin filaments, especially with actin filament bundles (“stress” fibers) at the ventral surface of cultured cells, and extracellular ligands or adhesion-promoting molecules such as fibronectin. The drawback to this kind of study was its inability to differentiate a true molecular interaction from simply “being in the neighborhood.” Another approach was to precipitate ligand-receptor complexes from de-

100

THOMAS P. STOSSEL

tergent-solubilized cells and look for actin in the precipitates, and a large variety of such complexes were identified, including associations between actin and surface molecules such as immunoglobulins, lectins, hyaluronate, and fibronectin (e.g., Kalomiris and Bourguignon, 1988; Lacy and Underhill, 1987; Salisbury et a l . , 1988). A major problem with this strategy is that when ligation of receptors stimulates cells, actin assembly frequently takes place; the gellike properties of polymerized actin render it likely to precipitate nonspecifically and to trap other molecules. In addition, it has not been clear from these investigations whether actin was directly attached to the receptors or indirectly via other molecules. These difficulties notwithstanding, some evidence has emerged for the existence of different kinds of relatively stable linkages between membrane proteins and actin. In one case, Luna and colleagues have isolated a lectin-binding transmembrane glycoprotein, which they named ponticulin, from the amoeba Dictyostelium discoideum. Kinetic studies using ponticulin-containing membranes suggested that ponticulin promotes the aggregation of actin monomers into nuclei that can elongate bidirectionally, resulting in actin filaments zippered along their sides to the plasma membrane. Myosin molecules and actin-binding myosin fragments inhibit the binding of ponticulin to actin filaments, potentially explaining why when such fragments are used to identify intracellular actin filaments, few side-on associations are observed (Schwartz and Luna, 1988). Burridge and co-workers have amassed evidence that certain molecules immunohistochemically localized at cell adherence sites bind one another in a cascade of interactions that spans the membrane from extracellular ligand to intracellular actin. Fibronectin ligates its receptor, integrin. The cytoplasmic tail of integrin binds to a high molecular weight protein, talin, which in turn binds the pearshaped protein vinculin. Vinculin binds a-actinin, a known actin-binding protein (references in Moloney et al., 1987). The weak but not necessarily fatal link in this scheme is that binding of integrin to talin is of rather low affinity. However, in the dense environment of adhesion plaques found in cultured cells, the reactants may be sufficiently concentrated to mediate associations of adequate strength. Studies of the actin system in mammalian white blood cells and platelets have shown that ABP, a high-molecular-weight actin-binding protein, produces an orthogonal actin gel in this cellular domain (Hartwig and Shevlin, 1986). Okita et al., (1985) and Fox (1985) showed that platelet ABP, in addition to its actin gelation function, binds with high affinity to the transmembrane platelet glycoprotein in Von Willebrand’s factor receptor, glycoprotein Ib (GPIb). Fox’s results have been confirmed and the GPIb-binding site on the platelet ABP molecule has been localized (Ezzell et al., 1988). In summary, there is much evidence that nonerythroid cells in some instances :xpress actin-membrane interactions that resemble the stable high-affinity link3ges found in the erythrocyte. The specific proteins involved in these interac-

8. ACTIN-MEMBRANE INTERACTIONS IN CELLS

101

tions, however, may be quite variable for different types of cells, and the details of these differences are probably highly important in regard to the particular forces exerted on the membrane and the requirements for membrane stabilization in a particular situation.

IV. THE ACTIN SYSTEM IN MEMBRANE PROPULSION In contradistinction to the “pulling” function of actin emphasized previously, many cells “push” out processes such as filopodia and pseudopodia, and such protrusive activity is central to cell locomotion and phagocytosis. Since stimuli that promote locomotion (e.g., chemotactic peptides) and phagocytosis (opsonized particles) stimulate actin assembly, it is widely assumed that this polymerization drives the propulsive activity. The properties of actin assembly in vitro have been extensively catalogued. Briefly, actin assembles via a nucleation-condensation process in which nucleation is thermodynamically unfavored, whereas elongation is diffusion-limited, at least from the end of filaments that are “barbed” with respect to the arrowhead configuration conferred by binding of myosin head fragments to filamentous actin. The rate of exchange of monomers is also an order of magnitude faster at the “barbed” than at the “pointed” end of actin filaments. At steady state in physiological solution conditions in vitro, a “critical” concentration of actin monomers of 0.1 ph4 coexists with actin filaments in an exponential length distribution (Kom et al., 1987). This situation contrasts with actin inside of many cells such as leukocytes and platelets, of which near millimolar concentrations are unpolymerized (50% of the total cell actin) but which assemble in response to cellular activation by various agonists (Stossel, 1988). Thus, one challenge is to explain what signals generated by cell stimulation have what effects on actin or the family of actin-modulating proteins that affect actin nucleation or elongation. Given the large variety of three-dimensional configurations with respect to one another that actin filaments exhibit in cells, is it reasonable to focus on linear assembly in approaching the problem of how actin affects membrane and cell function? In vitro at least, one can polymerize actin in the presence of certain actin-binding proteins derived from cells that display bundles and in which such bundles contain these actin-binding proteins, and obtain bundles resembling those found in situ (Matsudaira and Janmey, 1988). When actin assembles from monomers in the presence of macrophage ABP in the absence of any other actinmodulating proteins, an orthogonal network forms with architectural features strikingly similar to those found in ABP-containing cortical cytoplasm (Niederman et al., 1983; Hartwig and Shevlin, 1986). It is possible, therefore, that the three-dimensional architecture of the assembling actin arises from a stochastic interaction between actin and actin-cross-linking proteins, and that the problem

102

THOMAS P. STOSSEL

of remodeling and its regulation can be simplified by considering it a matter of linear actin assembly. There is also a large number of actin-binding proteins that can influence linear actin assembly. The actions of these proteins, however, boil down to a few basic properties. One action is to bind monomers with sufficient avidity that they cannot nucleate and will not bind to the barbed ends of preexisting filaments. A variation on this theme is that the binding protein ligates monomeric actin with sufficient avidity to prevent nucleation but not growth onto the barbed ends of filaments. Both of these functions are subsumed by the protein profilin, first isolated from mammalian spleen by Lindberg and co-workers. Since its original discovery as a tight 1 : 1 complex with actin, subsequent workers isolated free profilin molecules with a relatively low (p.M kDa) affinity for actin. Studies in our laboratory recently reconciled these results by showing that platelet profilin can exist in both low- and high-affinity states with respect to actin binding (Lind el al., 1987). Recent evidence that the membrane polyphosphoinositides ( ppI) phosphatidylinositol 4,5-bisphosphate (PIP,) and phosphatidylinositol monophosphate (PIP) dissociate tight complexes of actin and profilin, has added to profilin’s attractiveness as a regulatory protein, since PIP and PIP, turn over during cell activation (Lassing and Lindberg, 1985). The mechanism by which profilin regains its high avidity for monomeric actin is not currently known. Profilin or profilinlike proteins could theoretically account for actin assembly in cells and be linked to phosphoinositide turnover. A limitation of such regulation is that the slow spontaneous nucleation rate of actin would place a lag between the stimulus for assembly and actual polymerization. In addition, the location of nascent filaments would not be determined, except insofar as lowaffinity profilin gives up monomers to preexisting actin filaments. It may be necessary for cells in some circumstances, however, to destroy old filaments and put new ones in different places. Another set of actin-binding proteins addresses these issues. The relevant properties are binding to barbed filament ends and blockade of exchange between the filaments and monomers, nucleation of actin filament growth, and even the severing of preformed filaments. In some cells, these different functions may be deployed over a number of different proteins (Stossel et al., 1985; Pollard and Cooper, 1986). In several mammalian cells, one protein displays all of these functions and distributes them in response to a complex set of instructions provided by cellular signaling systems. This protein is gelsolin, first discovered in our laboratory as the initial representative of an evolutionarily conserved family of proteins that reversibly shorten actin filaments by mechanically severing actin polymers (a reaction unique in polymer science), nucleating actin monomer assembly, and blocking exchange at the barbed ends of filaments. Gelsolin is activated for binding to actin by micromolar calcium concentrations, and its overall action is reversed by PIP and PIP, (Janmey and Stossel, 1987; Janmey et al., 1987). The removal of calcium only

8. ACTIN-MEMBRANE INTERACTIONS IN CELLS

103

partially reverses gelsolin-actin interaction unless ppI are present. In the absence of ppl, gelsolinfactin 1 : 1 complexes persist that are resistant to EGTA. Their presence can be taken as evidence that gelsolin, actin, and calcium were exposed to one another. Cells can dissociate these complexes, from which it is inferred that the complexes interacted with cellular ppI (Chaponnier et al., 1987; Lind et al., 1987). The genomic structure of gelsolin, its relationship to other similar proteins, and its mechanism of action have been reviewed elsewhere (Kwiatkowski ef al., 1986, 1988; Kwiatkowski, 1988, Matsudaira and Janmey, 1988). A hypothetical model for how cortical structure is regulated in response to perturbation of externally disposed receptors by agonists is presented in Fig. 1. Ligation of membrane receptors causes PIP, hydrolysis, which generates IP,, releasing calcium from intracellular stores. Extracellular calcium may also enter the cell through channels opened by receptor-mediated pathways. Calcium activates gelsolin, which severs the cortical actin network preexisting in the resting cell, and/or nucleates actin assembly. Both effects shorten the length and increase the number of actin filaments compared to actin polymerizing in the absence of activated gelsolin. Gelsolin-capped actin oligomers diffuse to the plasma membrane, where gelsolin interacts with membrane PIP and PIP,, which increases following their initial hydrolysis because of the action of kinases specific for PIP and PIP,. High-affinity profilin-actin complexes form during activation (the mechanism for this association is not presently understood) and diffuse to the membrane, where they bind to membrane PIP and PIP, and dissociate. Therefore, PIP and PIP, generate actin monomers and actin filament barbed ends for rapid assembly and cross-linking by other ABP. We refer to the shuttling

Actin Network

FIG. 1. Hypothetical model for how perturbation of membrane receptors by specific ligands may cause actin assembly. PI, Phosphatidylinositol; PIP, phosphatidylinositol phosphate; PIP,, phosphatidylinositol 4,5-bisphosphate; DAG, diacylglycerol; IP3, inositol hisphosphate.

104

THOMAS P. STOSSEL

of gelsolin and profilin between actin and the plasma membrane as the gelsolin and profilin cycles, respectively. The evidence underlying this model has primarily been obtained from work on platelets and macrophages. Specifically, platelet activation by thrombin or ADP results in the transient conversion of profilin from a low-affinity to a high-affinity state. It also causes the brief formation of EGTA-resistant gelsolin-actin complexes in platelet extracts, indicating that most gelsolin and propilin interacted with calcium and actin in the activating platelet. These changes precede the assembly of actin in response to the platelet agonists. Recent additional work has shown by ultrastructural immunocytochemistry that gelsolin molecules reside throughout the cytoplasm of resting platelets but migrate transiently to the plasmalemma during platelet activation, and this translocation follows the time course of complex formation documented in the biochemical studies (Hartwig et al., 1989a,b). An important conceptual aspect of this background is that the interface of the actin system with the plasma membrane in the context of propulsion may represent transient, metabolic, and dynamic rather than stable mechanical interactions that are required for retraction. The demonstration of ppI-gelsolin and ppIprofilin associations provides a biochemical basis for this interaction as well as a mechanism for localizing actin assembly to places where receptor perturbation is taking place. Our studies suggest that the physical state of the ppI is very important for their functional association with gelsolin (Janmey and Stossel, 1987, 1989; Janmey et al., 1987). Therefore, alterations in other membrane lipids might affect the way in which ppI affect profilin and gelsolin. Conversely, gelsolin and profilin or their complexes with actin may influence the cycles by having effects on the interconversion of ppI. It would be reasonable, for example, to have gelsolin-actin complexes activate PI, PIP, and/or PIP, kinase(s) in order to generate metabolites that dissociate the complexes.

V.

THE ACTIN SYSTEM AND MEMBRANE PROPULSION AND RETRACTION

How is actin assembly coupled to pseudopod extension? If actin filaments are tightly apposed to the plasmalemma, how can new monomers add on to the ends of these filaments and extend the membrane outward? Furthermore, it would be counterproductive to have actin firmly attached to a membrane that needs to expand outward. One attractive hypothesis is that the events accompanying cell activation such as calcium influx, PIP,, hydrolysis, and the subsequent dissociation of gelsolin-actin and profilin-actin complexes by regenerated ppl generate osmotically active particles that cause membrane protrusion because of water influx (Oster, 1984). The solation of a preexisting actin gel by gelsolin also can contribute to osmotic swelling by relieving the cross-linking of a gel that has an

A

severed actin

B

‘4

Water efflux

FIG.2 . (A) First step in membrane protrusion, involving propulsive forces. Calcium release and influx, PIP? hydrolysis, and profilin-actin and gelsolin-actin dissociations yield osmotic particles, which promote water influx and membrane expansion. Severing of the preexisting actin gel by gelsolin relieves the gel swelling pressure. Contractions elsewhere activated by calcium contribute through hydrostatic effects. (B) Second step in membrane protrusion, involving stabilization and contraction. Availability of barbed actin filament ends near the membrane and actin monomers leads to assembly and cross-linking of actin into an orthogonal network. Adsorption of osmotically active particles (actin monomers) onto giant molecules (actin filaments) reduces the osmotic drive generated previously.

106

THOMAS P. STOSSEL

intrinsic swelling pressure (Fig. 2A). According to this model, removal of calcium and restoration of ppI reverses the effects of gelsolin and profilin, and the subsequent polymerization and cross-linking of actin resulting in the formation of a new actin network stabilizes the osmotically extended pseudopod (Fig. 2B). Only the presence of a restored network permits myosin molecules activated by calcium (Warrick and Spudich, 1987) to contract the network and thereby retract membranes with actin filaments attached to transmembrane proteins. It is evident that the relationship of the actin system to the membrane of actively motile cells is necessarily much more complex than that of passively deformed cells or of cells that have fixed positions in tissues. Motile cells require interactions that are both fixed and dynamic, the distribution between them dependent on space, time, and specific signals. ACKNOWLEDGMENT Some of the work described in this review was supported by USPHS grants HL19429 and AI28465. REFERENCES Bennett, V. (1985). The membrane skeleton of human erythrocytes and its implications for more complex cells. Annu. Rev. Biochem. 54, 273-304. Chaponnier, C., Yin, H. L., and Stossel, T. P. (1987). Reversibility of gelsolinkactin interaction in macrophages. J. Exp. Med. 166, 97-106. Ezzell, R. M., Kenney, D. M., Egan, S., Stossel, T. P., and Hartwig, J. H. (1988). Localization of the domain of actin-binding protein that binds to membrane glycoprotein Ib and actin in human platelets. 1.Biol. Chem. 263, 13303-13309. Fox, J. E. B. (1985). Identification of actin-binding protein as the protein linking the membrane skeleton to glycoproteins on platelet plasma membranes. J. Biol. Chem. 260, 11970- 11977. Hartwig, J. H., and Shevlin, P. (1986). The architecture of actin filaments and the ultrastructural location of actin-binding protein in the periphery of lung macrophages. J. Cell B i d . 103, 10071020. Hartwig, J. H., Chambers, K . , and Stossel, T. P. (1989a). Association of gelsolin with actin filaments and cell membranes of macrophages and platelets. J . Cell Biol. 108, 467-480. Hartwig, J. H., Chambers, K. A,, Hopcia, K. L., and Kwiatkowski, D. J. (1989b). Association of prolifin with filament free regions of human leukocyte and platelet membranes and reversible membrane binding during platelet activation. J. Cell. B i d . 109, 1571-1579. Janmey, P. A , , and Stossel, T. P. (1987). Modulation of gelsolin function by phosphatidyiinositol 4.5-bisphosphate. Nature (London) 325, 362-364. Janmey, P. A., and Stossel, T. P. (1989). Gelsolin-polyphosphoinositide interaction. J. Biol. Chem. 264, 4825-483 1. Janmey, P. A., iida, K . , Yin, H. L., and Stossel, T. P. (1987). Polyphosphoinositide-containing vesicles dissociate endogenous gelsolin-actin complexes and promote actin assembly from the fast-growing end of actin filaments blocked by gelsolin. J. Biol. Chem. 262, 12228-12236. Kalomiris, E. L., and Bourguignon, L. Y. W. (1988). Mouse T lymphoma cells contain a transmembrane glycoprotein (GP85) that binds ankyrin. J. Cell Biol. 106, 319-327. Kom, E. D., Carlier, M.-F., and Pantaloni, D., (1987). Actin polymerization and ATP hydrolysis. Science 238, 638-644.

8. ACTIN-MEMBRANE INTERACTIONS IN CELLS

107

Kwiatkowski, D. J. (1988). Predominant induction of gelsolin and actin-binding protein during myeloid cell differentiation, J. Biol. Chem. 263, 8239-8243. Kwiatkowski, D. J., Stossel, T. P., Orkin, S. H., Mole, J. E., Colten, H. R.,and Yin, H. L. (1986). Plasma and cytoplasmic gelsolins are encoded by a single gene and contain a duplicated actinbinding domain. Nature (London) 323, 455-458. Kwiatkowski, D. J., Mehl, R., and Yin, H. L. (1988). Genomic organization and biosynthesis of secreted and cytoplasmic forms of gelsolin. J. Cell Biol. 106, 375-382. Lacy, B. E . , and Underhill, C. B. (1987). The hyaluronate receptor is associated with actin filaments. J. Cell Bioi. 105, 1395-1404. Lassing, I . , and Lindberg, U. (1985). Specific interaction between phosphatidylinositol 4,5bisphosphate and profilactin. Nature (London) 314, 472-474. Lind, S. E., Janmey, P. A,, Chaponnier, C., Herbert, T.-J., and Stossel, T. P. (1987). Reversible binding of actin to gelsolin and profilin in human platelet extracts. J . Cell Biol. 105, 833-842. Matsudaira, P., and Jamey, P. A. (1988). Pieces in the actin-severing puzzle. Cell 54, 139-140. Moloney, L., McCaslin, D., Abernathy, J., Paschal, B., and Burridge, K. (1987). Properties of talin from chicken gizzard smooth muscle. J. Biol. Chem. 262, 7790-7795. Mooseker, M. S. (1985). Organization, chemistry, and assembly of the cytoskeletal apparatus of the intestinal brush border. Annu. Rev. Cell Biol. 1, 209-241. Nelson, W. J., and Veshnock, P. J. (1986). Dynamics of membrane-skeleton (fodrin) organization during development of polarity in Madin-Darby canine kidney epithelial cells. J. Cell Biol. 103, 1751-1765. Nelson, W. J., and Veshnock, P. J. (1987). Ankyrin binding to (Na+ +K+)ATPase and implications for the organization of membrane domains in polarized cells. Nature (London) 328, 533-536. Niederman, R . , Amrein, P., and Hartwig, J. H. (1983). The three dimensional structure of actin filaments in solution and an actin gel made with actin-binding protein. J. Cell Biol. 96, 14001413. Niggli, V., and Burger, M. M. (1987). Interaction of the cytoskeleton with the plasma membrane. J . Membr. Biol. 100, 97-121. Okita, L. R.,Pidard, D., Newman, P. J., Montgomery, R. R.,and Kunicki, T. J. (1985). On the association of glycoprotein Ib and actin-binding protein in human platelets. J. Cell Biol. 100, 317-321. Oster, G. F. (1984). On the crawling of cells. J. Embryol. Exp. Morphol. 83, Suppl., 329-364. Pollard, T. D., and Cooper, J. A. (1986). Actin and actin-binding proteins: A critical evaluation of mechanisms and functions. Annu. Rev. Biochem. 55, 987-1035. Salisbury, J. L., Baron, A. T., Keller, B. G. A., and Skiest, D. (1988). Membrane IgM: Interactions with the cortical cytoskeleton in the human lymphoblastoid cell line WiL2. Cell Motii. Cytoskeleton 9, 140- 152. Schliwa, M. (1986). “The Cytoskeleton. An Introductory Survey.” Springer-Verlag, 1986. New York . Schwartz, M. A., and Luna, E. J. (1988). How actin binds and assembles onto plasma membranes from Dictyostelium discoideum. J. Cell Biol. 107, 201-209. Singer, S. J., and Kupfer, A., (1986). The directed migration of eukaryotic cells. Annu. Rev. Cell Biol. 2, 337-365. Stossel, T. P. (1989). From signal to pseudopod. How cells control cytoplasmic actin assembly. J. Biol. Chem. 264, 18261-18264. Stossel, T. P., Chaponnier, C., Ezzell, R. M., Hartwig, J. H., Janmey, P. A,, Kwiatkowski, D. J., Lind, S . E., et al. (1985). Nonmuscle actin-proteins. Annu. Rev. Cell Biol. 1, 353-402. Wanick, H. M . , and Spudich, J. A. (1987). Myosin structure and function in cell motility. Annu. Rev. Cell Biol. 3, 379-421.

This Page Intentionally Left Blank

CURRENT TOPICS IN MEMBRANES AND TRANSPORT, VOLUME 36

Chapter 9

Biogenesis and Cell Surface Distribution of Acetylcholine Receptors Stably Expressed in Fibroblasts TONI CLAUDIO Department of Cellular and Molecular Physiology Yale University School of Medicine New Haven, Connecticut 06510

I. Introduction 11. Acetylcholine Receptor-Fibroblast Cell Lines 111. Properties of Acetylcholine Receptors Expressed in Acetylcholine Receptor-Fibroblast Cells IV. Properties of Individually Expressed Subunits A . Antigenicity B. Stability C. Ability to Bind cu-Bungarotoxin V. Posttranslational Modifications A. Glycosylation B. Phosphorylation VI. Conclusions References

1.

INTRODUCTION

Although the most extensively investigated member of the ligand-gated family of receptor channels consists of the nicotinic acetylcholine receptors (AChRs), much still remains to be elucidated concerning subunit biosynthesis, processing, assembly, and transport. Having AChR subunit clones available has greatly aided in the analysis of this receptor. Transient expression systems (such as microinjecting cRNAs into Xenopus oocytes) have proved extremely useful as a method of quickly expressing wild-type and mutagenized AChRs for structure-function analysis (Mishina ef a l . , 1985; Claudio, 1987; Leonard et al.. 1988). However, 109 Copynght 0 19W by Academic Press. Inc All nghts of reproduction m any form reserved

110

TONI CLAUD10

stable expression systems may be more useful for investigating a variety of cell biological processes. The muscle (or musclelike) AChRs are heterologous multisubunit complexes composed of four different subunits in the stoichiometry a#$; each subunit undergoes several posttranslational modifications, and each subunit appears to span the membrane four times. In order to reconstitute all of the functional properties of this receptor, the four subunits must be present and at least some of the posttranslational modifications must be executed correctly. Although musclelike and neuronal AChR subunit genes and cDNAs have been isolated from a number of species (reviewed in Claudio, 1989), we have chosen to work with the AChR isolated from Torpedo californica electric organ (Claudio, 1987) for several reasons: (1) it is the most extensively studied and best characterized AChR; (2) it is the one for which the most structural information is available (Toyoshima and Unwin, 1988; see also reviews in Karlin, 1980; Conti-Tronconi and Raftery, 1982; Popot and Changeux, 1984; and (3) many tools are available for its analysis including libraries of monoclonal antibodies containing subunit- and conformation-specific antibodies.

II. ACETYLCHOLINE RECEPTOR-FIBROBLAST CELL LINES Several stable cell lines have been established that express functional cell surface Torpedo AChRs using two different gene transfer methods: calcium phosphate-DNA-mediated cotransfection (Claudio et ul., 1987) and viral infection (Claudio et ul., 1989). By cotransfection, we determined that 80% of Ltkuprt cells, which incorporated an adenine phosphoribosyltransferase (uprt) gene, also incorporated all four AChR subunit cDNAs (Claudio, 1987). Fully functional Torpedo AChRs have been expressed in Ltk-aprt- cells by cotransfecting with thymidine kinase ( t k ) (all-11 cells) and in NIH3T3 cells by cotransfecting with the neomycin resistance gene (neo') (all- 15 cells). We have also used helper-free recombinant retroviruses to produce cell lines expressing individual AChR subunits. Two mouse fibroblast (L, NIH3T3) and three rodent muscle (rat L6, mouse C2, mouse BC,H-1) cell lines have been successfully infected with murine retrovirus recombinants. Two types of AChR-expressing cell lines have been established using retroviruses: all-Torpedo AChR-fibroblast cell lines and Torpedo-endogenous rat hybrid AChR L6 cell lines. When t t k aprt- cells containing integrated copies of pSV2-p, -y, -6 cDNAs and the tk gene (all-6 cells) were infected with retroviral or-recombinants, fully functional AChRs were expressed on the surface of the cells (Claudio et ul., 1989). When rat L6 muscle cells were infected with the recombinant, three types of AChRs were expressed on the surface: AChRs containing two rat a subunits (normal endogenous AChRs), hybrid AChRs containing two Torpedo or subunits, and -

9. BlOGENESlS OF ACETYLCHOLINE RECEPTORS

111

hybrid AChRs containing one rat and one Torpedo ci subunit (Paulson and Claudio, 1988, 1990). The observation that an AChR molecule in this system can be composed of 01 subunits from Torpedo and rat proves that the two ci subunits in a muscle cell need not derive from the same polysome.

111.

PROPERTIES OF ACETYLCHOLINE RECEPTORS EXPRESSED IN ACETYLCHOLINE RECEPTORFIBROBLAST CELLS

We have shown that expression of Torpedo AChRs is acutely temperature sensitive. Subunit polypeptides are made in our cell lines at 37°C; however, they are not assembled into complexes at this temperature (Claudio et al., 1987; Paulson and Claudio, 1988, 1990). As the temperature is lowered below 37"C, the number of surface AChRs increases. Because the mammalian host cells function optimally at 37"C, we are unable to reduce the temperature sufficiently to obtain optimal Torpedo AChR expression in this system. However, a compromise temperature of -26°C permits sufficient host cell synthesis and processing to occur, while at the same time allowing Torpedo AChR complexes to form. The defect in expression of AChRs is not in transport or insertion of complexes into the plasma membrane, but rather appears to be incorrect folding of the polypeptides at higher temperatures (Paulson and Claudio, 1990). Thus, it is not surprising that Torpedo ci subunits are inefficiently assembled in their mammalian host cell environment, since we are attempting assembly at ill-adapted temperatures. Even with suboptimal temperatures, -40,000 surface AChR molecules are expressed per cell in all-1 1 cells, giving a site density of -5O/t.~m~.Only a single population of AChRs is expressed on the surface of these cells. Surface receptors are composed of all four subunits; they are assembled into azpyS pentamers (Hartman et al., 1989), and they migrate on sucrose gradients with the proper sedimentation coefficient of 9 S (Claudio et al., 1987). Amplitude and kinetic analyses of single-channel recordings made from AChR-fibroblasts support the observation that only a single population of cell surface AChR is expressed in these cell lines (Claudio et al., 1987; Sine et al., 1990). The receptors display proper physiological properties (similar desensitization kinetics, mean channel open time, single-channel conductance) and they display all of the pharmacological properties of Torpedo AChRs isolated from Torpedo membranes [same rank order of affinities for agonists and antagonists, dissociation constant for a-bungarotoxin (BuTx; a competitive antagonist) of 7.8 X 1 0 - l LM ,reversible antagonists exhibit different affinities for the two binding sites on the AChR]. The cell surface distribution of the AChRs was monitored using rhodamine-BuTx and fluorescence microscopy (Hartman et al., 1989). We found that every cell was expressing AChRs, the level of expression was virtually identical in each cell, and the receptors were fairly evenly distributed on the surface.

112

TONI CLAUD10

IV.

PROPERTIES OF INDIVIDUALLY EXPRESSED SUBUNITS

The availability of cell lines expressing individual uncomplexed subunits has allowed us to characterize subunit molecular masses, antigenicity, posttranslational processing, cell surface expression, stability in fibroblasts, stability in differentiated and undifferentiated muscle cells, ability (of a) to bind BuTx (Claudio e f al., 1989), and whether subunits self-associate (Paulson and Claudio, 1987). The level of expression of the different subunits vaned among the cell lines whether the lines were established by transfection or infection, with the highest level of expression being -270,000 molecules per cell.

A. Antigenicity Each of the individually expressed subunits is recognized by subunit-specific antisera and migrates on sodium dodecyl sulfate-polyacrylamide gels with the proper electrophoretic mobility except y, which has a slightly faster mobility. The altered mobility appears to be due to small differences in posttranslational modifications between Torpedo electrocytes and other expression systems (Claudio e f al., 1989). None of the individually expressed subunits is expressed on the cell surface of fibroblasts at 37°C or 28°C.

B. Stability Half-life studies conducted on subunit cell lines indicated that a,y, and 6 each had the same short half-life of -43 min at 37°C whether expressed in fibroblasts, undifferentiated muscle cells, or differentiated muscle cells. In contrast, the p subunit had an extremely short half-life of only -12 min at 37°C. When the cell lines were grown at 28"C, we found that although all the subunit half-lives increased, the p subunit was the most profoundly affected. It is intriguing to postulate that p might play some critical regulatory role in the process of subunit assembly and that the observed unique properties of f3 might be related to its role.

C. Ability to Bind a-Bungarotoxin It is only the a subunits in an AChR pentamer that bind BuTx. a Subunits assembled into proper pentamers have a dissociation constant for BuTx of - 10- lo M ; however, denatured and renatured Torpedo subunits or fragments of a polypeptides bind BuTx with an affinity of only -lo-' M (Lentz and Wilson, 1988). Torpedo AChRs expressed in fibroblasts bind BuTx with a Kd of 7.8 X 10- I * M (Claudio ef al., 1987). Although essentially 100% of individually expressed a subunits bind BuTx, binding occurs over four log units with a Kapp of -4 x 1 0 - 7 ~ .

9. BIOGENESIS OF ACETYLCHOLINE RECEPTORS

V.

113

POSTTRANSLATIONAL MODIFICATIONS

A. Glycosylation Experiments with tunicamycin and endoglycosidase H indicated that a,p, y, and 6 had incorporated one, one, two, and three units of oligosaccharide, respectively (Claudio et ul., 1989), consistent with values predicted from the carbohydrate mass of each subunit isolated from Torpedo electric organ (Nomoto et al., 1986). This later observation demonstrates that individually expressed subunits are being transported at least as far as the endoplasmic reticulum. B. Phosphorylation One posttranslational modification long thought to play a role in AChR function is phosphorylation (reviewed in Schuetze and Role, 1987; Huganir and Greengard, 1987). Phosphorylation of other membrane-bound receptors has been shown to cause diverse effects including a decrease in agonist binding, an increase in the rate of receptor internalization, or an increase in desensitization. The AChR-fibroblast system has allowed us to correlate direct measurements of AChR phosphorylation with in vivo measurements of AChRs. After treatment of all-1 1 cells with 32P, three classes of labeled AChR subunits can be isolated: (1) unassembled, cytoplasmic subunits; ( 2 ) assembled, cytoplasmic AChR complexes; and (3) assembled, cell surface AChR complexes. Only phosphorylation of the y and 6 subunits was observed in any of the AChR pools. The degree and pattern of phosphorylation, however, was found to be different among these three pools, suggesting a possible role of phosphorylation in assembly of AChR subunits (Green and Claudio 1988). The CAMP-dependent protein kinase stimulators forskolin and CAMP-S increased phosphorylation of the different pools, and in addition, we have found that treatment of all-1 1 cells with these agents results in an increase in the number of surface AChRs (Green et al., 1989). VI.

CONCLUSIONS

Using cotransfection, viral infection, or a combination of the two techniques, we have established cell lines expressing one, two, three, or four different AChR subunits. Having cell lines expressing individual subunits has been of great use in analyzing biosynthesis, posttranslational processing, transport, and whether subunits form stable self-associations. In cell lines expressing all four subunits, fully functional, cell surface AChRs are expressed that can be easily analyzed using biochemical, pharmacological, immunological, or electrophysiological techniques. Though tedious to establish, the stable expression system appears to be ideally suited for investigations of this and other heterologous multisubunit protein complexes.

114

TONI CLAUD10 ACKNOWLEDGMENTS

This work was supported by National Institutes of Health grants NS21714 and HL38156 REFERENCES Claudio, T. (1987). Stable expression of transfected Torpedo acetylcholine receptor a-subunits in mouse fibroblast L cells. Proc. Natl. Acad. Sci. U.S.A. 84, 5967-5971. Claudio, T. (1989). Molecular genetics of acetylcholine receptor-channels. In “Frontiers in Molecular Biology” (D. M. Glover and D. Hames, eds.), Vol. 2, pp. 9-88. IRL Press, London. Claudio, T., Green, W. N., Hartman, D., Hayden, D., Paulson, H. L., Sigworth, F. J., Sine, S. M., and Swedlund, A. (1987). Genetic reconstitution of functional acetylcholine receptor-channels in mouse fibroblasts. Science 238, 1688-1694. Claudio, T., Paulson, H. L., Green, W. N., Ross, A. F., Hartman, D. S . , and Hayden, D. (1989). Fibroblasts transfected with Torpedo acetylcholine receptor 0.7, and 6 subunit cDNAs express functional AChRs when infected with a retroviral a-recombinant. J. Cell Biol. 108,2277-2290. Conti-Tronconi, B. M., and Raftery, M. A. (1982). The nicotinic cholinergic receptor: Correlation of molecular structure with functional properties. Annu. Rev. Biochem. 51, 491-530. Green, W. N., and Claudio, T. (1988). Differences in the phosphorylation of unassembled, assembled but cytoplasmic, and surface acetylcholine receptors. SOC. Neurosci. Abstr. 14, 1045. Green, W. N., Ross, A. F., and Claudio, T. (1989). Cell surface ACh receptor expressed in mouse fibroblasts is increased by CAMP-dependent kinase stimulation. SOC. Neurosci. Abstr. 15, 1300. Hartman, D. S., Poo, M.-M., Green, W. N., Ross, A. F., and Claudio, T. (1989). Synaptic contact between embryonic neurons and acetylcholine receptor-fibroblasts. J. Physiol. (Paris) 83, 1-8. Huganir, R. L., and Greengard, P. (1987). Regulation of receptor function by protein phosphorylation. Trends Phurmacol. Sci. 8, 472-477. Karlin, A. (1980). Molecular properties of nicotinic acetylcholine receptors. I n “The Cell Surface and Neuronal Function” (C. W. Cotman, G. Poste, and G. L. Nicolson, eds.), Vol. 6, pp. 191260. Elsevier/North-Holland, Amsterdam. Lentz, T. L., and Wilson, P. T.(1988). Neurotoxin-binding site on the acetylcholine receptor. Int. Rev. Neurobiol. 29, 117-160. Leonard, R. J., Labarca, C. G., Chamet, P., Davidson, N., and Lester, H. A. (1988). Evidence that the M2 membrane spanning region lines the ion channel pore of the nicotinic receptor. Science 242, 1578-1581. Mishina, M., Tobimatsu, T., Tanaka, K., Fujita, Y . , Fukuda, K., Kurasake, M., Takahashi, H., Morimoto, Y., Hirose, T., Inayama, S., Takahashi, T., Kuno, M., and Numa, S. (1985). Location of functional regions of acetylcholine receptor a-subunit by site-directed mutagenesis. Nature (London) 313, 364-369. Nomoto, H., Takahashi, N., Nagaki, Y., Endo, S., Arata, Y., and Hayashi, K. (1986). Carbohydrate structures of acetylcholine receptor from Torpedo californicu and distribution of oligosaccharides among the subunits. Eur. J. Eiochem. 157: 233-242. Paulson, H. L., and Claudio, T. (1987). Analysis of Torpedo culifornica nicotinic acetylcholine receptor subunits expressed in mammalian fibroblast and muscle cell lines. J. Cell Biol. 105, 62a. Paulson, H. L., and Claudio, T. (1988). Temperature-sensitive expression of all-Torpedo and Torpedo-rat hybrid AChRs in mammalian cells. SOC.Neurosci. Absrr. 14, 1045. Paulson, H. L., and Claudio, T. (1990). Temperature-sensitive expression of all-Torpedo and Torpedo-rat hybrid AChR in mammalian muscle cells. J. Cell. Biol. (in press). Popot, J.-L., and Changeux, P.-J. (1984). Nicotinic receptor of acetylcholine: Structure of an oligomeric integral membrane protein. Physiol. Rev. 64, 1162-1239.

9. BIOGENESIS OF ACETYLCHOLINE RECEPTORS

115

Schuetze, S. M.,and Role, L. (1987). Developmental regulation of nicotinic acetylcholine receptors. Annu. Rev. Neurosci. 10, 403-457. Sine, S. M . , Claudio, T., and Sigworth, F. J. (1990). Activation of Torpedo acetylcholine receptors expressed in mouse fibroblasts: Single channel current kinetics reveal distinct agonist binding affinities. J . Gen. Physiof. (in press). Toyoshima, C . , and Unwin, N. (1988). Ion channel of acetylcholine receptor reconstructed from images of postsynaptic membranes. Nature (London) 336, 247-250.

This Page Intentionally Left Blank

CURRENT TOPICS IN MEMBRANES AND TRANSPORT, VOLUME 36

Chapter 10

Control of Organelle Movements and Endoplasmic Reticulum Extension Powered by Kinesin and Cytoplasmic Dynein MICHAEL P . SHEETZ, SANDRA L . DABORA, ERIC STEUER, AND TRINA A . SCHROER Department of Cell Biology and Physiology Washington Universig Medical School St. Louis, Missouri 63110

I.

Introduction Membranous Organelle Transport on Microtubules A. In Vitro Reconstitution of Organelle Motility B . Microtubule Differentiation and Cell Polarity C. Paradigms for Control of Motility from Actin-Based Systems 111. Interaction of Motors with Organelles to Produce Motility A. Model of the Organelle Translocation Complex B. Endoplasmic Reticulum Extension References 11.

1.

INTRODUCTION

Recent studies have demonstrated that the transport of vesicular organelles and endoplasmic reticulum (ER) components along microtubules requires the microtubule-based motors, kinesin and cytoplasmic dynein. Although these motors are clearly the enzymes that power organelle movements, they require additional cytoplasmic components for reconstitution of a motile organelle translocation complex. We can speculate on several sites for regulation of organelle motility; this regulation is likely to involve not only the motors and their accessory factors but also the organelle binding sites that are thought to determine directionality. In addition, other microtubule-associated proteins may affect motility by forming a 117 Copyright Q 1990 by Academic k s s . Inc. All nghts of reproducuon in any form reserved.

118

MICHAEL P. SHEETZ ET AL.

stable linkage between an organelle and a microtubule or by simply coating the microtubules to block the binding sites for motors. It is of obvious importance to understand how these different possible modes of regulation may operate in a variety of biological systems and as a case in point, we discuss the extension of the ER in vivo.

II. MEMBRANOUS ORGANELLE TRANSPORT ON MICROTUBULES The directed movements of membranous organelles on microtubules clearly are responsible for fast axonal transport (Allen el al., 1985; Vale et al., 1985a; Schnapp et al., 1985), the congregation of the Golgi apparatus (Matteoni and Kreis, 1987), and the extension of the ER (Dabora and Sheetz, 1988a; Lee and Chen, 1988). In addition, the inherent order of the cytoplasm created by the organization of the microtubules with their minus ends at the centrosome and plus ends at the periphery (Fig. 1) suggests that organelle transport on microtubules may be responsible for much of the intracellular traffic from the cell center outward and vice versa. The two microtubule motors, kinesin and cytoplasmic dynein, can logically carry material outward toward the plus ends of

FIG. 1. This diagram of a fibroblastic cell illustrates the normal organization of the microtubules with their minus ends anchored at the centrosome and plus ends at the periphery. Plus-end-directed motility (kinesin) would be expected to drive material to the cell surface (e.g., during secretion), whereas minus-end-directed motility (cytoplasmic dynein) would move material to the cell center. MTOC, Microtubule-organizing center.

10. CONTROL OF ORGANELLE MOVEMENTS

119

the microtubules (kinesin; Vale et al., 1985a) or inward toward the minus ends (cytoplasmic dynein; Lye et al., 1987; Paschal and Vallee, 1987). Although a variety of cellular functions, including endocytosis and protein and lipid synthesis and processing, continue in the absence of microtubules, the rates of these processes are often decreased [e.g., the movement of glycolipids to the apical surface of MCDK cells (G. van Meer and K. Simons, personal communication)l. In highly asymmetric cells such as neurons, active transport is necessary for materials to reach the end of processes, whereas in nonpolarized cells directed transport may simply serve to facilitate intracellular organelle movement (Vale, 1987). Our understanding of the molecular mechanism of organelle transport has come largely from studies in which the movement of purified organelles on microtubules has been reconstituted in vitro. In this review we will summarize the current understanding of the molecular basis of organelle motility and speculate on how organelle motility may be controlled.

A. In Witro Reconstitution of Organelle Motility We have been able to reconstitute the microtubule-dependent movement of small vesicular organelles from cultured fibroblasts (Dabora and Sheetz, 1988b; Schroer et al., 1989) and squid axons (Vale et al., 1985b; Schroer et al., 1988) as well as ER networks from fibroblasts (Dabora and Sheetz, 1988b) and neural tissue. The experimental paradigm used to determine both the frequency and direction of organelle movements is illustrated in Fig. 2. The movement of these organelles in vitro appears similar to organelles in vivo. In vivo and in vitro, the velocities of small organelles are the same and organelle movements are blocked by similar concentrations of inhibitors. The major distinguishing feature of organelle movement in vivo is the low velocity of the larger organelles (>0.5 pm) and the saltatory (oscillatory) nature of the organelle movements, both of which might be explained by the effect of the dense cytoskeleton in the cell. The in vitro assays for organelle motility thus appear to provide a valid means for studying the molecular mechanism of organelle motility and perhaps its regulation.

B. Microtubule Differentiation and Cell Polarity In locomoting cells, establishment of cell polarity can be correlated with the posttranslational modification of microtubules (Gundersen and Bulinski, 1988) and in many cases the movement of the microtubule-organizing center relative to the nucleus (Singer and Kupfer, 1986). This might induce asymmetry within the cell by way of specific transport of certain components to one portion of the cell. Perhaps modification of microtubules and their associated proteins may affect the transport of organelles to different sites in the cell. The role that microtubule

120 Organelles

MICHAEL P. SHEETZ ET AL. Centrosome (MTOC)

Soluble Factors

Cytosol Fraction or Purified Motor Protein

FIG. 2. In vifro organelle motility assays are performed by mixing the potassium iodide-washed organelles with purified microtubules (polymerized from centrosomes for directionality measurements) and an appropriate motor fraction (see Schroer et al., 1989, for details). The components are mixed on a glass coverslip, and microtubules and organelles are visualized by video-enhanced light microscopy. The number of individual organelle movements per unit time per video field are counted as the measure of the amount of motility (video field area and the concentration of microtubules are kept constant).

structure plays in the transport of organelles is unknown at present; however, our understanding of actin-based motility systems provides a basis for speculation about certain factors that might influence motility on microtubules.

C. Paradigms for Control of Motility from Actin-Based Systems In our studies of actin-based motility in vifro we found that motility could be affected by the binding of two kinds of actin-binding proteins. Certain proteins were found to coat actin filaments and prevent the movement of myosin-coated beads on the same filaments. These included the troponin-tropomyosin complex (Vale et al., 1985b) and N-ethylmaleimide (NEM)-inactivated myosin subfragment I (Meusen and Cande, 1979). A second class of proteins induced the formation of an ATP-insensitive linkage between the myosin-containing object and the filament and blocked movement. For example, beads coated with a mixture of myosin and brush border spectrin moved slowly or not at all (M.P.

10. CONTROL OF ORGANELLE MOVEMENTS

121

Sheetz and M. Mooseker, unpublished observations). Similarly, the presence of inactive myosin heads (i.e., by NEM) could slow movement (Jones and Sheetz, 1986). We have not examined the effects of analogous proteins on microtubulebased organelle movement.

111. INTERACTION OF MOTORS WITH ORGANELLES TO PRODUCE MOTILITY The in virro reconstitution of organelle movements on purified microtubules has led to a better understanding of the motile complex formed by an organelle and a motor. The simplistic notion that a motor enzyme binds directly to a membrane receptor and by itself causes movement appears to be incorrect, since we have found it impossible to produce organelle motility with purified motors alone (Schroer et al., 1988, 1989). Purified motors will stimulate the movement of impure organelles (Vale et al., 1985b); however, washing organelles with 0.6 M potassium iodide removes all latent organelle motor activity, and no motility is seen in the presence or absence of highly purified motors. Two similar models have been developed from our studies of kinesin- and dynein-dependent organelle motility. The specific inactivation of cytoplasmic dynein (a minus-enddirected motor) has been shown to inhibit organelle movement toward the minus ends of microtubules in a variety of in virro systems including fibroblast extracts (Schroer et al., 1989), squid axons (Schnapp and Reese, 1989), and fish chromatophores (Haimo et al., 1989). In the fibroblast studies the fraction of inactivated dynein closely correlated with the degree of inhibition of organelle motility. We do not know whether or not a single dynein molecule (a multimeric complex of heavy and light chains) on an organelle surface is sufficient for movement. However, we observed that the velocity of organelle movement in the presence of mixtures of active and inactivated dynein is considerably slower than normal, suggesting that multiple dynein molecules may contribute to movement. The movement of the organelles could not be restored using purified dynein alone but required additional soluble component(s) of cytoplasm. The “accessory factors” copurified with the microtubule motors in a microtubule affinity purification step, and further purification is currently under way. To study the role of kinesin (a plus-end-directed motor) in organelle motility, we have used an antikinesin antibody affinity column to subtract kinesin form a supernatant that supported organelle movement (Schroer et al., 1988). After kinesin was removed, organelle movement was inhibited by 70% and motility could not be restored by the readdition of kinesin. The degree of inhibition was greater than expected (less than 50% of the organelle movements driven by the starting supematant were toward the plus ends of microtubules), suggesting that dynein-dependent organelle motility was also inhibited. Organelle motor activity

122

MICHAEL P. SHEETZ ET AL.

could be recovered from the antibody affinity column by elution with 0.5 M potassium chloride and 10 mM ATP. Under these conditions only a small fraction of the kinesin and cytoplasmic dynein in the preparation were eluted, but both anterograde and retrograde organelle movements were seen. Our results suggested that additional soluble proteins were required for kinesin-dependent organelle motility and that these proteins might bind to kinesin and, therefore, bind to the antikinesin column (see Fig. 3). In addition, the results suggest that the accessory proteins required for cytoplasmic dynein motility also bind to kinesin and that the amount of the accessory factors and not the amount of the motors may be limiting in producing translocating organelles. ORGANELLE MOTOR COMPONENTS Kinesin [Antemgrade Motor ? I SALT-WASHED ORGANELLES

(A)

Retrograde Motor

Accessory Protein

BEADS

AXOPLASMIC 52 SUPERNATANT

[+I end

ICMAI

KINESIN DEPLETED SUPERNATANT

IU I

1+1 end

1+1 PURIFIED KINESlN

ICI

cc

end

Q+

FIG.3 . Summary diagram of the current in vitro motility experiments showing that the organelle translocation complex contains not only a motor protein but also at least one accessory component. In contrast, the latex beads will move with only the motors bound to their surface.

10. CONTROL OF ORGANELLE MOVEMENTS

123

A. Model of the Organelle Translocation Complex The motility of organelles is unlike the motility of latex spheres because highly purified motors are sufficient to drive the movement of inert particles but not organelles. Kinesin will bind to membranes without causing motility (Schroer and Sheetz, 1989), suggesting that the accessory factors play an important role with the motors in producing the motile organelle translocation complex. Although little is known about the interactions of motors and the accessory factors, we can certainly speculate about the molecular nature of the organelle translocation complex and how it might be regulated. We suggest the following testable model for organelle motility. The basic elements of this model are the two microtubule motors (kinesin and cytoplasmic dynein), the accessory protein(s), and the organelle membrane-binding sites (motor receptors). In this model the membrane-binding site determines the directionality of organelle transport, the accessory proteins regulate the activity of the motor proteins in the binding site, and the motors themselves provide the force for movement when assembled into an active complex. An additional component that may be present under certain circumstances is a stabilizing linkage, which would serve to attach the organelle to the cytoskeleton in order to maintain its position (synapsin possibly plays this role for synaptic vesicles: Benfenati et al., 1989). There are many possible sites for regulation and, therefore, it is best to discuss this model in terms of specific types of organelle movements.

6. Endoplasmic Reticulum Extension The extension of the ER within cultured cells to form a polygonal array appears to involve microtubules. In early electron micrographs of cultured cells there were reports of parallel alignment of the ER with microtubules (F'ranke, 1971). Terasaki et al., (1986) later observed a close correspondencebetween the paths of ER strands and microtubules in the periphery of cells. Using an in vitro system with membranes isolated from chick embryo fibroblasts, it has been possible to generate an extensive reticular array on a microtubule bed using a soluble fibroblast extract containing microtubule motors (Fig. 4). This membrane network can be stained with markers for the ER. The formation of tubular membrane strands is a general property of the focal application of force to a membrane (Hochmuth et al., 1982; Waugh, 1982). As might be expected for a general physical property of membranes, similar networks could be formed by fluid shear (Vale and Hotani, 1988). Alternatively, an active role might be played by actin-based motility (Kachar and Reese, 1988) and by kinesin-driven movements of microtubules to which membranes were statically bound (Vale and Hotani, 1988). Observations of ER movements in live cells have revealed that ER formation in vivo (Lee and Chen, 1988) is remarkably similar to the process

124

MICHAEL P.SHEETZ ET AL.

FIG. 4. Video micrograph of an ER network formed in vitro from a postnuclear supernatant fraction of CEF cells (Dabora and Sheetz, 1988b). Note that the microtubules are lower in contrast than the ER strands and that fingers of ER are seen moving out along the microtubules.

we have reconstituted in vitro (Dabora and Sheetz, 1988b). It would appear that the construction of the ER in vivo requires the movement of ER membrane strands along microtubules as opposed to the microtubule gliding (Lee et al., 1989). According to our current model, the movement of the ER to the cell periphery involves the kinesin-dependent movement of ER tubules on microtubules (see Fig. 4). This is the most obvious part of the process because those strands can be readily visualized using the fluorescent dye, DiOC6(3). It is possible that cytoplasmic dynein bound at other sites on the ER might serve to bring tubules toward the centrosome and the Golgi apparatus (see Fig. 5). After the network is extended, stabilizing proteins might serve to link the ER to the cytoskeleton until other signals are given for redistribution. Indeed, Lee and Chen (1988) found that only 10%of the cells in an unsynchronized culture contain actively spreading ER. Because microtubules are highly dynamic structures in interphase cells, it is likely that the ER would form linkages to more stable elements of the cytoskeleton such as the intermediate filaments. Another important aspect of the ER is the modulation of ER assembly during

10. CONTROL OF ORGANELLE MOVEMENTS

125

1 Extension

FIG. 5 . Diagram depicting the two major steps in the process of ER extension in vivo. They are depicted here in sequence, but it is expected that there will be considerable overlap of fusion and extension.

126

MICHAEL P. SHEETZ ET AL.

the cell cycle. The transition into mitosis involves the complete randomization of the intracellular membranes into a dispersed population. We might expect that any stabilizing linkages between membranes and cytoskeleton would also be disrupted. There are many candidate proteins (synapsins, microtubule-associated proteins or MAP) that could serve as stable linkages between organelles and the cytoskeleton, and the assembly and disassembly of the nuclear envelope provides a model of how stable attachments would organize the membranes. During mitosis the ER, Golgi, and nuclear membranes vesiculate and are randomly spread between the two daughter cells. For an interphase ER to form, the vesicles must fuse and then be extended on the cellular microtubules. The motility of both plus-end and minus-end-directed organelles appears to be inhibited during mitosis and reactivated in interphase (Tooze and Burke, 1987; Matteoni and Kreis, 1987). It is possible that organelle motility is modulated by regulating the activity of the accessory components required for transport. This is consistent with our finding (see Section 11) that organelle movement in vitro is limited not by the concentration of motors but by the amount of accessory factors (Schroer et al., 1988). According to this model, the organelles contain signals for either plus-end or minus-end-directed movement and the motors are present in excess in cytoplasm, but motile activity is regulated by the amount of the accessory components. An alternative possibility is that the membrane receptors themselves are modified to control motor binding and subsequent motility. The development of quantitative in vitro assays of organelle motility has opened the way to understanding the molecular components needed to create a motile organelle-motor complex. Our understanding of these in vivo processes involving organelle motility, such as ER extension, provide the phenomenological basis to propose testable models that can guide experiments to understand how the processes actually do occur. Further, the purification of additional molecular components involved in organelle motility will make it possible to probe their effects on in vitro motility and to approach gene deletion or other gene modification experiments to understand how these components interact in the in vivo process of motility. ACKNOWLEDGMENTS This work was supported by grants from NIH, the Muscular Dystrophy Association, and the Juvenile Diabetes Foundation. REFERENCES Allen, R. D., Weiss, D . G., Hayden, J. H., Brown, D. T.,Fujiwake, H., and Simpson, M. (1985). Gliding movement of a bidirectional transport along single native microtubules from squid axoplasm: Evidence for an active role of microtubules in axonal transport. J . Cell B i d . 100, 1736-1752. Benfenati, F., Bahler, M . , Jahn, R . , and Greengard, P. (1989). Interactions of synapsin I with small

10. CONTROL OF ORGANELLE MOVEMENTS

127

synaptic vesicles: Distinct sites in synapsin I bind to vesicle phospholipids and vesicle proteins. J. Cell Biol. 108, 1863-1872. Dabora, S. L., and Sheetz, M. P. (1988a). The microtubule-dependent formation of a tubulovesicular network with characteristics of the ER from cultured cell extracts. Cell 54, 27-35. Dabora, S. L., and Sheetz, M. P. (1988b). Cultured cell extracts support organelle movement on microtubules in vitro. Cell Motil. Cytoskeleton 10, 482-495. Franke, W. W. (1971). Cytoplasmic microtubules linked to endoplasmic reticulum with crossbridges. Exp. Cell Res. 66, 486-489. Gundersen, G. G., and Bulinski, J. C. (1988). Selective stabilization of microtubules oriented toward the direction of cell migration. Proc. Nutl. Acad. Sci. U.S.A. 85, 5946-5950. Haimo, L. T., Thaler, C. D., and knton, R. D. (1989). Mechanism of bidirectional organelle movements in melanophores. Cell Motil. Cytoskeleton 11, 197. Hochmuth, R. M., Wiles, H. C., Evans, E. A., and McCown, J. T. (1982). Extensional flow of erythrocyte membrane from cell body to elastic tether: 11. Biophys. J. 39, 83-89. Jones, R., and Sheetz, M. P. (1986). Inhibition of motility by inactivated myosin heads. In “The Cytoskeleton” (T. W. Clarkson, P. R. Sager, and T. L. M. Syversen, eds.), pp. 213-220. Plenum, New York. Kachar, B., and Reese, T. S. (1988). The mechanism of cytoplasmic streaming in Characean algal cells: Sliding of endoplasmic reticulum along actin filaments. J. Cell Biol. 106, 1545-1552. Lee, C., and Chen, L. B. (1988). Behavior of endoplasmic reticulum in living cells. Cell 54,36-42. Lee, C., krguson, M., and Bo Chen, L. (1989). Construction of the endoplasmic reticulum. J. Cell Biol. 109, 2045-2055. Lye, R. J., Porter, M. E., Scholey, J. M., and McIntosh, J. R. (1987). Identification of a microtubule-based cytoplasmic motor in the nematode C . elegans. Cell 51, 309-318. Matteoni, R., and Kreis, T. J. (1987). Translocation and clustering of endosomes and lysosomes depends on microtubules. J . Cell B i d . 105, 1253-1266. Meusen, R. L., and Cande, W. Z. (1979). N-Ethylmaleimide-modifiedheavy meromyosin. A probe for actomyosin interactions. J. Cell Biol. 82, 57-65. Paschal, B. M., and Vallee, R. B. (1987). Retrograde transport by the microtubule-associated protein MAP 1C. Nature (London) 330, 181-183. Schnapp, B. J., and Reese, T. S. (1989). Dynein is the motor for retrograde axonal transport of organelles. Proc. Nutl. Acad. Sci. U.S.A. 86, 1548-1552. Schnapp, B. J., Vale, R. D., Sheetz, M. P., and Reese, T. S. (1985). Single microtubules from squid axoplasm support bidirectional movement of organelles. Cell 40,455-462. Schroer, T. A., and Sheetz, M. P. (1989). The role of kinesin and kinesin-associated proteins in organelle transport. In “Cell Movement” (J. R. McIntosh and F. D. Warner, eds.), Vol. 2, pp. 295-306. Alan R. Liss, New York. Schroer, T. A., Schnapp, B. J., Reese, T. S., and Sheetz, M. P. (1988). The role of kinesin and other soluble factors in organelle movement along microtubules. J. Cell Biol. 107, 1785-1792. Schroer, T. A , , Steuer, E. R., Sheetz, M. P. (1989). Cytoplasmic dynein is a minus end-directed motor for membranous organelles. Cell 56, 937-946. Singer, S . J., and Kupfer, A. (1986). Directed migration of eukaryotic cells. Annu. Rev. Cell Biol. 2 , 337-365. Terasaki, M., Chen, L. B . , and Fujiwara, K. (1986). Microtubules and the endoplasmic reticulum are highly interdependent structures. J . Cell B i d . 103, 1557- 1568. Tooze, J., and Burke, B. (1987). Accumulation of adrenocorticotropin secretory granules in the midbody of telophase AtT20 cells: Evidence that secretory granules move anterogradely along microtubules. J . Cell Biol. 104, 1047-1057. Vale, R. D. (1987). Intracellular transport using microtubule-based motors. Annu. Rev. Cell Biol. 3, 347-378.

128

MICHAEL P. SHEETZ ET AL.

Vale, R. D., and Hotani, H. (1988). hnnation of membrane networks in vitro by kinesin-driven microtubule movement. J . Cell Biol. 107, 2233-2230. Vale, R. D., Reese, T. S., and Sheetz, M. P. (1985a). Identification of a novel force-generating protein, kinesin, involved in microtubule-based motility. Cell 41, 34-4 1. Vale, R. D., Szent-Gyorgyi, A., and Sheetz, M. P. (1985b). Troponin-tropornyosin complex confers C a + + control to myosin bead movement in vitro. Biophys. J . 45, 145a. Waugh, R. E. (1982). Surface viscosity measurements from large bilayer vesicle tether formation: 11. Biophys. J. 38, 29-37.

Part IV

Signaling and Communication

This Page Intentionally Left Blank

CURRENT TOPICS IN MEMBRANES AND TRANSPORT, VOLUME 36

Chapter I I G Protein-Coupled Receptors: Structure and Function of Signal-Transducing Proteins ERIC M . PARKER AND ELLIOTT M . ROSS Department of Pharmacology University of Texas Southwestern Medical Center Dallas, Texas 75235

I. 11. 111. IV. V. VI .

Introduction Mechanism of G-Protein Activation by Agonist-Liganded Receptors General Structure of G Protein-Coupled Receptors Structure of the Ligand-Binding Domain Structure of the G Protein-Binding Domain Receptor-Mimetic Peptides as Models for the G Protein-Binding Domain References

1.

INTRODUCTION

Every cell must interpret and respond to a wide variety of extracellular signals such as hormones, neurotransmitters, odors, and light. This task is often accomplished by three-component signal transduction systems based on GTP-binding regulatory proteins, or G proteins (see Gilman, 1987; Stryer and Bourne, 1986; Ross, 1989, for reviews). When a receptor’s extracellular binding site is occupied by a hormone, the receptor facilitates the exchange of bound GDP for GTP by a G protein on the inner face of the plasma membrane. The GTP-liganded G protein is thus activated such that it can regulate the activity of an effector protein that generates an intracellular chemical or electrical signal. Numerous receptors act on 10 G proteins to regulate perhaps a dozen effector proteins, including adenylyl cyclase, cyclic GMP phosphodiesterase, phospholipases A, and C, and ion channels. G proteins are heterotrimers of a distinct (Y subunit (39-52 ma),one of two p subunits (35 and 36 m a ) , and a y subunit (8-10 kDa) (See Gilman, 1987, for

-

131 Copynght 0 1990 by Academic R e s s . Inc A11 rights of reproduction in any form reserved.

132

ERIC M. PARKER AND ELLIOTT M. ROSS

review). The 01 subunit is unique to each G protein and is responsible for guanine nucleotide binding and hydrolysis. Several 01 subunits have been identified both biochemically and by molecular cloning, and all have been found to have highly homologous primary structures (Lochrie and Simon, 1988). Each G protein is defined by its unique a subunit andlor by the effector protein that it regulates. Thus, G, stimulates adenylyl cyclase and Gi was identified as mediating the inhibition of the enzyme. At least three closely related Gi forms have now been identified (Jones and Reed, 1987); G i l , Gi2, and G,, share several functions in addition to inhibition of adenylyl cyclase (e.g., regulation of phospholipases and ion channels). Two distinct but highly homologous P subunits (Fong et al., 1988; Gao et al., 1987) and at least three different y subunits have been identified (Sternweis and Robishaw, 1984; Hildebrandt et al., 1985; Hurley et al., 1984). The Py subunits modulate the binding of nucleotides to a subunits, and may also exert independent regulatory functions. It has not been possible to separate the P and y subunits under nondenaturing conditions, and their individual activities are thus unknown. With the exception of the retinal G protein transducin, there is no indication that a particular Py complex associates specifically with a particular 01 subunit; each a subunit copurifies with a heterogeneous mixture of fir cornplexes. GTP-activated 01 subunits regulate effector proteins independently of the Py subunits, at least in most cases. Cell surface receptors that utilize G proteins represent a large and diverse group of proteins. They include many neurotransmitter receptors (e.g., adrenergic, muscarinic cholinergic, serotonergic, and peptidergic receptors), receptors for pituitary protein hormones and eicosanoids, the rhodopsins, and pheromone receptors in yeast and slime molds. All of these receptors are integral membrane glycoproteins that display significant structural and functional homology. The application of molecular cloning techniques to the study of G proteincoupled receptors has led to the elucidation of the primary structure of several of these receptors. This article will review current knowledge of the structure of G proteincoupled receptors and the structural features that are responsible for the various functions ascribed to these receptors. Particular emphasis will be given to the (3adrenergic receptor because it, along with rhodopsin, is the most extensively studied of the G protein-coupled receptors.

II. MECHANISM OF G-PROTEIN ACTIVATION BY AGONIST-LIGANDED RECEPTORS G proteins are activated when they bind GTP and deactivated when the bound GTP is hydrolyzed to GDP. The mechanism of activation and how activation is

11. G PROTEIN-COUPLED RECEPTORS

133

mediated by cell surface receptors have been studied in detail using intact membranes, purified G proteins, and purified G proteins and receptors that have been co-reconstituted into unilamellar phospholipid vesicles (Asano el al., 1984; Asano and Ross, 1984; Brandt and Ross, 1986). The regulatory interactions of receptors, G-protein subunits, and their many ligands are extremely convoluted, but a general pattern of reactions has emerged from these studies (Fig. 1). G proteins exist predominantly in an inactive complex with a single molecule of tightly bound GDP. The dissociation of GDP from the G protein is slow and limits the rate of GTP binding and consequent activation. Agonist-liganded receptor decreases the affinity of the G protein for guanine nucleotitrles, essentially converting the nucleotide-binding site from a “closed” state to an “open” state that freely exchanges nucleotide. Because the GTP/GDP ratio in the cell is large, the net result of agonist binding is an increase in activated, G’I’P-boundG protein. The activated G protein regulates its effector protein until the bound GTP is hydrolyzed to GDP by an intrinsic GTPase activity, the rate of which is slow (k,,, = 4 min- I ) and has not been shown to be regulated. Because the rate of receptor-catalyzed guanine nucleotide exchange exceeds the rate of GTP hydrolysis, a receptor can maintain several G proteins in an active form. A single P-adrenergic receptor can activate >30 molecules of G, (Brandt and Ross, 1986), and a single rhodopsin molecule can activate 1000 molecules of transducin (Liebman et al., 1987). The catalytic ability of receptors to regulate multiple G proteins yields considerable amplification of hormonal signals. It also explains, in part, the pharmacologically defined phenomenon of spare receptors, wherein only a fraction of the total receptor population need bind agonist in order to activate an effector maximally.

FIG. 1. Regulatory GTPase cycle. The activated species is shown as G*-GTP. Hydrolysis of bound GTP (reaction 2) can occur in the presence or absence of effector, E. The receptor-hormone complex (R.H) catalyzes both release of GDP and binding of GTP, increasing both the steady-state GTPase rate and the fraction of G protein in the activated form.

134

ERIC M. PARKER AND ELLIOTT M. ROSS

Ill. GENERAL STRUCTURE OF G PROTEIN-COUPLED RECEPTORS Over the past 3 years, the primary sequences of 20 different G protein-coupled receptors have been determined (see Fig. 2). All are homologous to the visual opsins, which also signal via G proteins and are among the best studied integral membrane proteins. The most striking aspect of the primary structure of the G protein-coupled receptors is the presence of seven stretches of hydrophobic amino acids. In the case of rhodopsins, biophysical, proteolytic, immunocytochemical, and chemical modification experiments have shown that the seven hydrophobic regions are largely helical and span the lipid bilayer (see Findlay and Pappin, 1986, for review). It is presumed that the seven hydrophobic stretches in the other receptors also represent transmembrane a-helices, although direct evidence is lacking. By further analogy to bacteriorhodopsin, the structure of which is known at low resolution from electron-microscopic studies (Henderson and Unwin, 1975), these helices are presumed to form a bundle around the activating ligand. Such a structure is consistent with spectroscopic data on visual rhodopsin (Findlay and Pappin, 1986). Given the analogy with the rhodopsins, the predicted topology of G proteincoupled receptors is illustrated in Fig. 3 using the avian P-adrenergic receptor. The amino termini of the receptors typically have one or more consensus sites for N-linked glycosylation and, hence, are presumed to be oriented extracellularly. This orientation dictates the disposition of the putative connecting loops and the carboxy terminus relative to the membrane. As shown in Fig. 2, homology among the G protein-coupled receptors is concentrated in the putative membrane-spanning helices. Homology is minimal in the amino- and carboxy-terminal domains, in the largest intracellular loop, and in the extracellular connecting loops. Limited deletions or amino acid substitutions in the amino terminus and extracellular connecting loops of the hamster p,-adrenergic receptor do not affect its function and, hence, these regions may only be required for folding and processing of the protein (Dixon et a l . , 1987). The first two cytoplasmic loops are fairly short and generally display some homology among the various receptors, but the third cytoplasmic loop is typically longer and is quite divergent. Its length is also quite variable; it is -150 amino acids longer in the muscarinic cholinergic receptor than in the substance K receptor or rhodopsin. Variability in this cytoplasmic loop may indicate that it confers specificity to the interaction of the receptors with G proteins or other cytoplasmic structural or regulatory molecules. The carboxy terminus is also quite divergent among receptors and is of variable length, being relatively long in the p- and a,-adrenergic receptors and much shorter in most of the others. It may also be involved in a receptor-specific function. In some cases, most notably the P-adrenergic receptors and rhodopsin, the carboxy terminus is rich in serine and threonine residues. In the case of rhodop-

135

11. G PROTEIN-COUPLED RECEPTORS I1

I .BE1

34 E L L ~ A O ~ ~ L L U L ~ L I V A U ~ V I M I ~ ~ P I L O ~ L T ~ ~ F I I S L I U D L W W L ~ F W ~ L W R G ~ - U L W I S F L C E M S L D

BET1 BE12 ALFl ALFZ Hll. NllC

51 E P L O O O Y I ~ L U L L I V L L I V ~ V I V A I U l P R L o l L l M L F l * I L A U D L ~ L W P F ~ l l W - ~ f G S F F C E L ~ ~ ~ 26 EElDEAWVOUlL*NIVLAIVFOmVIlAIUFERL~lvlNYFllsUL~~FWSHlL*a-HlFUFKE~YTSI~ S8 P P L D W U I S V G L V L W F I L F A l W . l l L V I L W L C * R H L R l P I M Y F I V Y U I ~ L L L S F l m P F U l L E V L G f ~ W L C s I ~ ~ I U U ~ 26 I P f S L O W L l L W L f f i L L M L L l V F ~ VI Al V F I Y * L U P P * L F L V f L A U D l L V A l L V l P F S L A N E ~ f - W F ~ l K E l YLALD 50 DWWW- I l S L L L G l L I F C I V L C M C W M IALERSLDUVANYL I G S L A W D L M n V L P * L I L f W L M K ~ W L ~ l m L F l A L D 4 7 FOFPDEYPY*PILSIWII I ~ T I W I L V I M Y I * E K K L N ~ A ~ N Y F L ~ U I Y Y Y Y U L H D L S L U I L ~ Y ~ ~ P R I L C ~ I S ~ D M c 2 16 SPYKIFEWFIVLVAGSLSLVII l O l L H N f I ~ R ~ L 9 l ~ N f F L f S L A WI LG IY T O l L Y l L I l V I G Y - W L W W ~ L m A L D l U C 1 18 ~ ~ P U I V A F I G I ~ ~ ~ L L S U ~ W ~ L L ~ I S F ~ ~ E L IGlFOlLYllfLLffiII-WLGlLAmLmALD K ~ ~ ~ ~ F L L S L A ~ L ~ SUUK 2 6 F 0 4 J I P L A L Y T M f L A L V L V A ~ l V l U Il L A ~ R l V l N Y F l V Y L A U D L Q U F W J i F ~ F W A S H N.WFGRAFCYFO*LFP I RHW 30 YL~FWFPIL#AYUFLLIVLGFPlMFLlLlWWHK~LRlPLNf ILLNL~VMLF*~EGF~S~LY~SLHGY~FVFW~GCIILEGFFA

ME1 UEll BE12 ALFl ALF2 Nlla HllC

122 139 11b 126 114 117 136

M C Z lo(

106 WBK 114 lUCl

u r n iia

V .BE1 BET1 BE12 ALFl ALFZ Ulla Bllc

206

AYAlASSIISFYIPLLI~IFVYLRWREAKEOlRKlDRCEGRFYGsOE9~PP~PP~Pl-~--~-~--

.........L G Y G R l S K R K l S R W * R E I I U L K l L G

223 198

AfAIASSWSFYWLCIMFVILMVFREA~EIKlDSCERRFLGWARVPSPSPSPWAPAPPP~PRP

-~lAPLAYGRAGKRIIPSI)LVALREO~L~llG

202

105 193 21b M C Z 105 M c l 187 WUK 196 IKO 202

VI aUE1 )Ell BE12 ALF1 ALF2 111. HI1c

VII

2?4 323 277 298 3?3

IWOlFlL~PFflALPLGSLFSlLKPPDAVFK(---WFULGfFMKL~PIIfPCSSKEFKIAFUILGC9CRSGSRRR+ 1 4 1 .a's WIGVFWNFPFFFlYTLlAV~S~----WRlLFKFFFUFGYCNSSLMWlfl~FM~FRRAiKK~L~CRWRKRlV

w

llffilFlL~PFFlVALVLPF~SXH--~lLlWllN~Gf~SLLMWlfAYFN~F~AFKKllKCMFC~

II*eVFlLMPFFLVYIVYVF~lD----LWWLFVFF~GYAMUFMVIIlCRSP-DFRUFKRLL-CFPRUDRRL

r116 .a's

I 1 * e V F l L M P F F L W W F H ~ ~ - - - - L V P O R L F V F F ~ U L G f W U F ~ P I I Y C R S P P - D F R U F D 6 L L - C U R R M R+R R7 6 u ' s

I I f f i l F l L ~ P F F I V Y I V W V I ~ ~ - - - - L I ~ E W I L L ~ ~ G Y V Y U F ~ P L I Y C R S P P - D F R l A F O E L L - C L R R S+S S68U f.a's

517 I V F F V F L I * I C P F F I l ~ I L I C ~ ~ L ~ K L L N v F ~ l G Y ~ S G l M P L W l L F M K l f R M F ~ Y ~ R+~ 64 Y KI'..~ K K P V McZ 391 A I L L l F l I l W f y H N L I M l F c M l p l l - - - - w T I G f y L C l l M S I l M P A O A L D u r n l l r x I I I I K w * I * I l R U C l 369 A I L L A F l V T W P V ~ I M V f l F ~ ~ l ~ ~ - - L Y L 6 Y ~ E Y V Y S l I M P K Y M C ~ F l D l F R ~ L L L C+R5~1K.a's RR~K S W K 2% L W n F A I N L P Y H L f F l L G l F ~ l Y C N K F I O O Y l L A L F Y U * S S l M Y M P I I f C C L M * I I F R S G F R L A F R C C R N T s T E l+ 17s... R W 256

IWIAFLICYYPYANAFYIFlWOCt---~FGPIFUllPAFFUSMIlMPVIYIYtOFRNQLllI-CCW(MPLGDD

*

1 8 .a's

FIG. 2. Alignment of the amino acid (aa) sequences of several representative G protein-coupled receptors. Nonhomologous regions at the amino and carboxy termini and in the loop between spans V and VI are not shown, although the sizes of the omissions are indicated. The seven putative membrane-spanning domains are overlined and numbered. Dashes indicate gaps placed in the sequence to optimize the alignment. The sequences were taken from the following references: turkey p-adrenergic (quasi+,) receptor (aBET) from Yarden et al., (1986); human PI-adrenergic receptor (BET1) from Frielle er al., (1987); hamster pz-adrenergic receptor (BET2) from Dixon ef al. (1986); hamster a,-adrenergic receptor (ALFI) from Cotecchia et al. (1988); human a,-adrenergic receptor (ALF2) from Kobilka er al. (1987b); human 5HT-la receptor (HTla) from Kobilka et al. (1987a); rat 5HT-lc receptor (HTlc) from Julius er al. (1988); porcine muscannic cholinergic receptor (MAc2) from Peralta er al. (1987); porcine muscarinic cholinergic receptor (MAc1) from Kubo er af. (1986); bovine substance K receptor (SUBK) from Masu et al. (1987); human rhodopsin (RHOD) from Nathans and Hogness (1984).

136

ERIC M. PARKER AND ELLIOTT M. ROSS

EXTRACELLULAR SPACE

FIG. 3. Proposed topography of the avian erythrocyte f3-adrenergic receptor. Putative membranespanning regions are boxed and were assigned on the basis of hydropathy analysis and by analogy to rhodopsin. Those regions that can be proteolytically removed with no loss of function (Rubenstein et al.. 1987) are shown in dashed boxes. Positively charged residues are enclosed in squares and negatively charged residues are enclosed in circles. The single consensus site for N-linked glycosylation in the N-terminal region is shown. The arrangement of putative cytoplasmic and extracellular domains is arbitrary and of no significance.

sin, several of these serine and threonine residues are phosphorylated by a retinaspecific rhodopsin kinase upon bleaching. This light-induced phosphorylation is thought to be involved in light adaptation (Liebman et al., 1987). By analogy, phosphorylation of these residues by a receptor kinase may also be involved in desensitization of other G protein-coupled receptor systems (Sibley et al., 1988). Their removal by in vitro mutagenesis alters desensitization of the receptor (Bouvier et al., 1988). Regions rich in hydroxyl groups are also found in the third cytoplasmic loop of several receptors and may serve similar functions.

IV. STRUCTURE OF THE LIGAND-BINDING DOMAIN The strong homology among the G protein-coupled receptors in the putative membrane-spanning domains suggests that their hydrophobic core of these proteins carries out the same function in every case. One likely function for this region is the binding of ligands specific for the various receptors. In the case of

11. G PROTEIN-COUPLED RECEPTORS

137

rhodopsin, retinal is linked to a lysine residue located in the seventh membrane span and extends into a hydrophobic pocket formed by the remaining membranespanning helices (Findlay and Pappin, 1986). Interestingly, the biogenic amine and muscarinic cholinergic ligands of G protein-coupled receptors are chemically similar to the retinal-lysine conjugate in that they consist of a hydrophobic moiety linked to a cationic side chain. Hence, it is possible that these ligands are also bound within a hydrophobic pocket formed by the membrane-spanning domains. The hypothesis that the ligand-binding domain of G protein-coupled receptors lies in the hydrophobic core of the protein is supported by several lines of evidence. Site-directed mutagenesis studies by Dixon et al. (1987) have shown that large sections of the amino terminus, carboxy terminus, and the intracellular and extracellular hydrophilic loops of the hamster P,-adrenergic receptor are not required for ligand binding. Similarly, Rubenstein et al. (1987) found that proteolytic removal of most of the amino- and carboxy-terminal domains and a large portion of the third cytoplasmic loop left ligand binding intact. The two peptides obtained in a limited digest, which remained noncovalently associated in detergent solution, represented only the hydrophobic core and associated short loops. Wong et al. (1988) used two (3-adrenergic photoaffinity labels, [1251]iodocyanopindolol-diazirineand [ 1251]iodoazidobenzylpindolol, to study the ligandbinding site of the turkey erythrocyte P-adrenergic receptor. Both of the tryptic peptides described by Rubenstein et al. (1987) incorporated label upon photolysis in roughly equal amounts. Similar results have been obtained by Dohlman et al. (1987b, 1988, and personal communication). Further proteolysis followed by peptide sequencing localized one site to Trp330,which is located in the middle of the seventh membrane span. This is very close to the corresponding residue in rhodopsin to which retinal is attached ( L Y S * ~ The ~ ) . labeled site in the larger fragment has not been determined as precisely, but it is localized somewhere between the carboxy-terminal ends of spans I1 and V. The observation that label is incorporated into two regions that are widely separated in the primary sequence suggests that the P-adrenergic receptor folds such that span VII is closely apposed to a region in spans 11-V. A similar orientation of the membranespanning a helices has been proposed for rhodopsin by aligning the rhodopsin structure to the low-resolution crystal structure of bacteriorhodopsin (Findlay and Pappin, 1986). This is consistent with the idea that the membrane-spanning domains fold to form a hydrophobic pocket that comprises the ligand-binding domain. Furthermore, these results suggest that all G protein-coupled receptors are folded in a similar fashion and probably bind ligand, undergo ligand-induced conformational changes, and activate G proteins in a similar, if not identical, manner. By examining the ligand-binding specificity of a series of au,-adrenergicreceptor-P,-adrenergic receptor chimeras, Kobilka ef al., (1988) concluded that the

138

ERIC M. PARKER AND ELLIOTT M. ROSS

seventh membrane-spanning domain was a major determinant of the ligandbinding characteristics of a particular receptor. Relative a-versus P-adrenergic specificity seemed to reflect the relative amount of a-or P-adrenergic receptor sequence in the chimera, but sequence from span VII seemed to have a somewhat larger effect than did sequence from the other spans. Regardless, the observation that these chimeras can bind adrenergic ligands and activate G proteins reinforces the notion that all G protein-coupled receptors have a common structural arrangement and that agonist binding is coupled to G-protein activation by a common mechanism. If ligands bind in a hydrophobic pocket formed by the membrane-spanning domains, the cationic side chain of these ligands must be buried in this hydrophobic environment. This could be made energetically palatable by pairing the cationic side chain with acidic amino acid side chains in the membrane-spanning domains of the receptor. There are several aspartate residues in the second and third membrane-spanning domains that are conserved with only rare exception in all the G protein-coupled receptors sequenced to date (see Fig. 1). These aspartate residues in the P,-adrenergic receptor have been replaced by asparagine by side-directed mutagenesis (Strader et al., 1987b; Fraser et al., 1988). Of particular interest is the observation that replacement of Asp1I3 eliminates the ability of the receptor to bind iodocyanopindolol with high affinity and, based on measurements of adenylate cyclase activity, also decreases the affinity of the receptor for agonists by several orders of magnitude. Hence, this residue is a good candidate for the counterion that participates in an electrostatic interaction with the cationic amine moiety that exists in most ligands of G protein-coupled receptors.

V.

STRUCTURE OF THE G PROTEIN-BINDING DOMAIN

Despite the high degree of homology among their a subunits, each G protein must display selectivity for the receptor that regulates it. Similarly, the homologous G protein-coupled receptors presumably undergo a similar conversion to an activating conformation in response to the binding of agonist. Based on these considerations, it is likely that the structural components of receptors and G proteins that determine their interaction are generally similar, but yet sufficiently (and subtly) different to allow the required selectivity. Because G proteins behave as peripheral membrane proteins, it is reasonable to assume that one or more of the cytoplasmic domains of the receptors are responsible for regulating G proteins. Several recent experiments have yielded evidence supporting this assumption. Strader et al. (1987b) deleted several segments of the third putative cytoplasmic loop of the hamster P,-adrenergic receptor and found that deletion of residues immediately adjacent to the membrane abolished the ability of the receptor to activate adenylate cyclase. Deletion of

11.

G PROTEIN-COUPLED RECEPTORS

139

other regions of this loop, either by site-directed mutagenesis or by proteolysis (Rubenstein et al., 1987), did not alter the ability of the receptor to activate adenylate cyclase. The importance of the third cytoplasmic loop is also highlighted by the data of Kubo er al. (1988). These investigators constructed several chimeric Ml-M2 muscarinic cholinergic receptors and showed that MI sequence in the large cytoplasmic loop is sufficient to determine characteristic M1 function when these receptors are expressed in frog oocytes. No data on M2 responses were shown for any of the chimeras. In similar studies by Kobilka et al. (1988), a chimeric a,-P,-adrenergic receptor having p-receptor sequence only in the fifth and sixth transmembrane domains and in the third cytoplasmic loop was able to activate adenylate cyclase in response to a a,-adrenergic agonists. The G,-coupled a2functions were likewise not assayed. Hence, it appears that regions of the third cytoplasmic loop that are proximal to the membrane play a major role in the selectivity of receptors among G proteins. Mutations in the carboxy-terminal cytoplasmic domain of the human P,-adrenergic receptor immediately adjacent to the membrane also hinder the ability of the receptor to activate G, (O’Dowd er al., 1988). As is the case with the third cytoplasmic loop, deletions of the carboxy-terminal domain distal to the membrane have no effect on the ability of the receptor to activate G, (Strader et al., 1987b; Rubenstein et al., 1987). The possible roles of the first and second cytoplasmic loops in coupling to G proteins is still unclear, but it is certainly conceivable that all four putative cytoplasmic domains act in concert to form the G protein-binding domain. As discussed earlier, it is likely that the structural determinants of selectivity are subtle and not absolute. Indeed, data from our laboratory suggest that the ability of receptors to select among G proteins is not absolute. The P-adrenergic receptor acts predominantly via G, in most cells, and the receptor regulates G, with high efficiency when both purified proteins are reconstituted into phospholipid vesicles. In addition, however, when reconstituted in phospholipid vesicles, the avian P-adrenergic receptor can also activate three pertussis toxinsensitive forms of Gi, albeit with less efficiency (Asano et al., 1984; Rubenstein and Ross, unpublished data; Abramson and Molinoff, 1987; Abramson et al., 1988). Some selectivity among the Gi forms was noted, and there was no evidence for the receptors’ regulation of Go or transducin. Abramson et al. (1988) have also detected regulation of the P-adrenergic receptor by Gi in membranes of the cyc- mutant of S49 lymphoma cells. A G,-receptor complex was detected chromatographically, and Gi regulated the receptor’s affinity for agonists. Muscarinic receptors can couple to Gi, Go, and at least one other uncharacterized G protein that regulates phospholipase C and is insensitive to pertussis toxin (Florio and Sternweis, 1985; Masters er al., 1984; Peralta et al., 1988). Cerione et al. (1985) have demonstrated that reconstituted rhodopsin can stimulate not only its physiological target, transducin, but also Gi. In the converse experiment, a,-

140

ERIC M. PARKER AND ELLIOTT M. ROSS

adrenergic receptors, which normally couple to G,, could also activate transducin (Cerione et af.,1985). Careful analysis of the efficiency with which natural and mutant receptors activate various G proteins will probably be required to sort out the mechanisms of selectivity. Our laboratory is one of several that are constructing chimeric receptors that will test the importance of specific structural determinants for the regulation of G proteins and their selectivity. These studies should also clarify the extent to which divergence of a receptor’s signal to different G protein-mediated pathways is a common regulatory motif in animal cells.

VI.

RECEPTOR-MIMETIC PEPTIDES AS MODELS FOR THE G PROTEIN-BINDING DOMAIN

Higashijima et al. (1988) speculated that a group of peptide toxins known as mastoparans may serve as structural models for the G protein-regulatory domain of G protein-coupled receptors. Mastoparans are amphipathic, cationic, tetradecapeptides found in wasp venoms. These authors showed that mastoparans activate G proteins in a manner that resembles agonist-liganded receptors in several characteristic ways. Mastoparans increase the rate of GDP release and GTP binding by several G proteins, particularly Gi and Go. Activation of Go or Gi by mastoparans is completely blocked by pertussis toxin-catalyzed ADP ribosylation, an effect often diagnostic of receptor interaction with a pertussis toxinsensitive G protein. Although mastoparans can regulate nucleotide exchange in free a subunits, the ability of mastoparan to activate G proteins is greatest when the a p y trimer is reconstituted into phospholipid vesicles, a phenomenon that is also seen with receptors. Finally, mastoparans, like receptors, activate G proteins at micromolar concentrations of Mg2 . Thus, it seems possible that the mechanism by which mastoparans activate G proteins may provide clues to the mechanisms by which receptors activate G proteins. Mastoparans display no sequence similarity with the G protein-coupled receptors but can be hypothesized to have conformational similarity. Nuclear magnetic resonance analysis has shown that membrane-bound mastoparans are &-helices that are arranged such that the four positive charges present in the molecule are directed away from the membrane (Higashijima et al., 1983). Interestingly, all of the putative cytoplasmic loops of the G protein-coupled receptors are rich in basic amino acid residues (see Figs. 1 and 2). It is therefore possible that the G protein-binding face of receptors consists of some ordered array of positively charged amino acid residues that are either exposed or redistributed upon agonist binding. However, many integral membrane proteins have an asymmetric distribution of basic amino acids, with these residues being preferred on the cytoplasmic domains of these proteins (von Heijne and Gavel, 1988). It has been postulated that this asymmetric distribution of basic amino acids determines the +

11. G PROTEIN-COUPLED RECEPTORS

141

proper topography of the protein during insertion into the membrane. Further experimentation will be necessary to test the hypothesis that a specific structural arrangement of basic amino acid residues on the cytoplasmic face of G proteincoupled receptors and in mastoparans are critical for activation of G proteins. Synthetic mastoparans are currently being designed that should shed some light on the primary and secondary structural features of these molecules that are required for G-protein activation and that determine selectivity among G proteins. The ability of synthetic mastoparans and chimeric receptors to activate particular G proteins should give us a better understanding of the structure of the G protein-binding face of receptors. ACKNOWLEDGMENTS Studies from the authors’ laboratory have been supported by NIH grant GM30355 and postdoctoral fellowship GMI 1943 (E. M. P.) and by R. A. Welch Foundation grant 1-982. REFERENCES Abramson, S. N., and Molinoff, P. B. (1987). Interactions of P-adrenergic receptors with a membrane protein other than the stimulatory guanine-nucleotide binding protein. Biochem. Pharmacol. 36, 2263-2269. Abramson, S. N., Shorr, R. G. L., and Molinoff, P. B. (1988). Interactions of P-adrenergic receptors with a membrane protein other than the stimulatory guanine nucleotide-binding protein. Biochem. Pharmacol. 36, 2263-2269. Asano, T., and Ross, E. M. (1984). Catecholamine-stimulated guanosine 5‘-@(3-thiotriphosphate) binding to the stimulatory GTP-binding protein of adenylate cyclase. Biochemistry 23, 54675471. Asano, T., Pederson, S. E., Scott, C. W., and Ross, E. M. (1984). Reconstitution of catecholaminestimulated binding of guanosine 5’-0-(3-thiotriphospate)binding to the stimulatory GTP-binding protein of adenylate cyclase. Biochemistry 23, 5460-5467. Bouvier, M . , Hausdorff, W. P., DeBlasi, A., O’Dowd, B. F., Kobilka, B. K . , Caron, M. G., and Lefkowitz, R. J. (I 988). Removal of phosphorylation sites from the Pz-adrenergic receptor delays onset of agonist-promoted desensitization. Nature (London) 333, 370-373. Brandt, D. R., and Ross, E. M. (1986). Catecholamine-stimulated GTPase cycle. Multiple sites of regulation of P-adrenergic receptor and Mg2+ studied in reconstituted receptor-(;, vesicles. J . Biot. Chem. 261, 1656-1664. Cerione, R. A . , Staniszewski, C., Benovic, J. L., Lefkowitz, R. J . , Caron, M. C . , Gierschik, P., Somers, R., Speigel, A. M . , Codina, J . , and Birnbaumer, L. (1985). Specificity of the functional interactions of the P-adrenergic receptor with guanine nucleotide regulatory proteins reconstituted in phospholipid vesicles. J . Biol. Chem. 260, 1493- 1500. Cotecchia, S., Schwinn, D. A., Randall, R. A , , Lefkowtiz, R. J., Caron, M. G., and Kobilka, B. K. (1988). Molecular cloning and expression of the cDNA for the hamster a,-adrenergic receptor. Proc. Natl. Acad. Sci. U.S.A. 85, 7159-7163. Dixon, R. A. F., Kobilka, B. K., Strader, D. J., Benovic, J. L., Dohlman, H. K., Frielle, T., Bolanowski, M. A., Bennett, C . D., Rands, E., Diehl, R. E., Mumford, R. A , , Slater, E. E., Sigal, 1. S . , Caron, M. G . , Lefkowitz, R. J., and Strader, C. D. (1986). Cloning of the gene and cDNA for mammilian P-adrenergic receptor and homology with rhodopsin. Nature ( L o d o n ) 321, 75-79. Dixon, R. A. F., Sigal, I . , Candelore, M. R., Register, R. B., Scattergood, W., Rands, E., and

142

ERIC M. PARKER AND ELLIOTT M. ROSS

Strader, C. D. (1987). Structural features required for ligand binding to the P-adrenergic receptor. EMEO J. 6, 3269-3275. Dohlman, H. G., Bouvier, M., Benovic, J. L., Caron, M. G., and Lefkowitz, R. J. (1987). The multiple membrane spanning topography of the P2-adrenergic receptor. Localization of the sites of binding, glycosylation, and regulatory phosporylation by limited proteolysis. J. Biol. Chem. 262, 14282- 14288. Dohlman, H. G., Caron, M. G., Strader, C. D., Amlaiky, N., and Lefkowitz, R. J. (1988). Identification and sequence of a binding site peptide of the P2-adrenergic receptor. Biochemistry 27, 1813-1817. Findlay, J. B. C., and Pappin, D. J. C. (1986). The opsin family of proteins. Biochern. J . 238, 625642. Florio, V. A., and Sternweis, P. C. (1985). Reconstitution of resolved muscarinic cholinergic receptors with purified GTP-binding proteins. J. Eiol. Chern. 260, 3477-3483. Fong, H. W. K., Amatruda, T. T., Birren, B. W., and Simon, M. L. (1987). Distinct forms of the p subunit of GTP-binding regulatory proteins identified by molecular cloning. Proc. h'atl. Acud. Sci. U.S.A. 84, 3792-3796. Fong, H. W. K., Yoshimoto, K. K., Eversole-Cire, P. E., and Simon, M. L. (1988). Identification of a GTP-binding protein o subunit that lacks an apparent ADP-ribosylation site for pertussis toxin. Proc. Nutl. Acad. Sci. U.S.A. 85, 3066-3070. Fraser, C. M., Chung, F., Wang, C., and Venter, J. C. (1988). Site-directed mutagenesis of human padrenergic receptors: Substitution of aspartic acid-130 by asparagine produced a receptor with high-affinity agonist binding that is uncoupled from adenylate cyclase. Proc. Nuti. Acad. Sci. U.S.A. 85, 5478-5482. Frielle, T., Collins, S., Daniel, K. W., Caron, M. G., Lefkowitz, R. J., and Kobilka, B. K. (1987). Cloning of the cDNA far the human P1-adrenergic receptor. Proc. Nutl. Acud. Sci. U.S.A. 84, 7920-7924. Gao, B., Gilman, A. G., and Robishaw, J. D. (1987). A second form of the p subunit of signaltransducing G proteins. Proc, Nutl. Acud. Sci. U.S.A. 84, 6122-6125. Gilman, A. G. (1987). G proteins: Transducers of receptor-generated signals. Annu. Rev. Biochem. 56, 615-649. Henderson, R., and Unwin, P. N. (1975). Three-dimensional model of purple membrane obtained by electron microscopy. Nature (London) 257, 28-32. Higashijima, T., Wakamatsu, K., Takemitsu, M., Fujino, M., Nakajima, T., and Miyazawa, T. (1983). Cornformational change of mastoparan from wasp venom on binding with phospholipid membrane. FEBS Lett. 152, 227-230. Higashijima, T., Uzu, S., Nakajima, T., and Ross, E. M. (1988). Mastoparan, a peptide toxin from wasp venom, mimics receptors by activating GTP-binding regulatory proteins (G proteins). J. Biol. Chem. 263, 6491-6494. Hildebrandt, J. D., Codina, J., Rosenthal, W., Birnbaumer, L., Neer, E. J., Yamazaki, A., and Bitensky, M. (1985). Characterization by two-dimensional peptide mapping of the y subunits of N, and Ni, the regulatory proteins of adenylyl cyclase, and of transducin, the guanine nucleotide-binding protein of rod outer segments of the eye. J. Eiol. Chem. 260, 14867-14872. Hurley, J. B., Fong, H. K. W., Teplow, D. B., Dreyer, W. J., and Simon, M. I. (1984). Isolation and characterization of a cDNA clone for the y subunit of bovine retinal transducin. Proc. Nutl. Acad. Sci. U.S.A. 81, 6948-6952. Jones, D. T., and Reed, R. R. (1987). Molecular cloning of five GTP-binding protein cDNA species from rat olfactory neuroepithelium. J. Eiol. Chem. 262, 14241- 14249. Julius, D., MacDermott, A. B., Axel, R., and Jessel, T. M. (1988). Molecular characterization of a functional cDNA encoding the serotonin l c receptor. Science 241, 558-564. Kobilka, B. K., Frielle, T., Collins, S., Yang-Feng, T., Kobilka, T. S., Francke, U., Lefkowitz, R.

11. G PROTEIN-COUPLED RECEPTORS

143

J., and Caron, M. G. (1987a). An intronless gene encoding a potential member of the family of receptors coupled to guanine nucleotide regulatory proteins. Nature (London) 329, 75-79. Kobilka, B. K., Matsui, H., Kobilka, T. S., Yang-Feng, T. L., Francke, U., Caron, M. G., Lefkowitz, R. J., and Regan, J. W. (1987b). Cloning, sequencing, and expression of the gene coding for the human platelet a*-adrenergic receptor. Science 238, 650-656. Kobilka, B. K., Kobilka, T. S., Daniel, K., Regan, J. W., Caron, M. G., and Lefkowitz, R. J. (1 988). Chimeric a2-, P2-adrenergic receptors: Delineation of domains involved in effector coupling and ligand binding specificity. Science 240, 1310- 1316. Kubo, T., Fukuda, K., Mikami, A,, Maeda, A,, Takahashi, H., Mishina, M., Haga, T., Haga, K., Ichiyama, I . , Kangawa, K., Kojima, M., Matsuo, M., Hirose, T., and Numa, S. (1986). Cloning, sequencing, and expression of a complementary DNA encoding the muscarinic acetylcholine receptor. Nurure (London) 323, 41 1-416. Kubo, T., Bujo, H., Akiba, I., Nakai, J., Mishina, M.,and Numa, S. (1988). Location of a region of the muscarinic acetylcholine receptor involved in selective effector coupling. FEBS Lett. 241, 119- 125. Liebman, P. A., Parker, K. R., and Dratz, E. A. (1987). The molecular mechanism of visual excitation and its relation to the structure and composition of the rod outer segment. Annu. Rev. Physiol. 49, 765-791. Lochrie, M. A,, and Simon, M. L. (1988). G protein multiplicity in eukaryotic signal transduction systems. Biochemistry 27, 4958-4965. Masters, S. B., Harden, T. K., and Brown, J. H. (1984). Relationship between phosphoinositide and calcium response to muscarinic agonists in astrocytoma cells. Mol. Phurmucol. 26, 149-155. Masu, Y., Nakayama, K., Tamaki, H., Harada, T., Kuno, M., and Nakanishi, S. (1987). cDNA cloning of bovine substance K receptor through oocyte expression system. Nature (London) 329, 836-838. Nathans, J., and Hogness, D. S. (1984). Isolation and nucleotide sequence of the gene encoding human rhodopsin. Proc. Nutl. Acud. Sci. U.S.A. 81, 4851-4855. O’Dowd, B. F., Hnatowich, M., Regan, J. W., Leader, W. M., Caron, M. G., and Lefkowtiz, R. J. (1988). Site-directed mutagenesis of the cytoplasmic domains of the human P2-adrenergic receptor. Localization of regions involved in G protein receptor coupling. J . Biol. Chem. 263, 15985- 15992. Peralta, E. G., Winslow, J. W., Peterson, G. L., Smith, D. H., Ashkenazi, A,, Ramachandran, J., Schimerlik, M. I., and Capon, D. J. (1987). Primary structure and biochemical properties of an M2 muscarinic receptor. Science 236, 600-605. Peralta, E. G., Ashkenazi, A,, Winslow, J. W., Ramachandran, J., and Capon, D. J. (1988). Differential regulation of PI hydrolysis and adenylyl cyclase by muscarinic receptor subtypes. Nature (London). 334., 434-437. Ross, E. M. (1989). Signal sorting and amplification through G protein-coupled receptors. Neuron 3, 141-152. Rubenstein, R. C., Wong, S. K.-F., and Ross, E. M. (1987). The hydrophobic tryptic core of the padrenergic receptor retains G, regulatory activity in response to agonists and thiols. J. Biol. Chem. 262, 16655-16662. Sibley, D. R., Benovic, J. L., Caron, M. G., and Lefkowitz, R. J. (1988). Phosphorylation of cell surface receptors: A mechanism for regulating signal transduction pathways. Endocr. Rev. 9, 38-56. Sternweis, P. C . , and Robishaw, J. D. (1984). Isolation of two proteins with high affinity for guanine nucleotides from membranes of bovine brain. J. Biol. Chem 259, 13806- 13813. Strader, C.D., Sigal, I. S., Register, R. B., Candelore, M. R., Rands, E., and Dixon, R. A. F. (1987a). Identification of residues required for ligand binding to the P-adrenergic receptor. Proc. Nutl. Acud. Sci. U.S.A. 84, 4384-4388.

144

ERIC M. PARKER AND ELLIOTT M. ROSS

Strader, C. D., Dixon, R. A. F., Cheung, A. H., Candelore, M. R., Blake, A . D., and Sigal, I. S . (1987b). Mutations that uncouple the P-adrenergic receptor from G , and increase agonist affinity. J. Biol. Chem. 262, 16439-16443. Stryer, L., and Bourne, H. (1986). G. Proteins: A family of signal transducers. Annu. Rev. Cell B i d . 2, 391-419. von Heijne, G., and Gavel, Y. (1988). Topogenic signals in integral membrane proteins. Eur. J. Biochem. 174, 671-678. Wong, S. K.-F., Slaughter, C., Ruoho, A,, and Ross, E. M. (1988). The catecholamine binding site of the P-adrenergic receptor is formed by juxtaposed membrane-spanning domains. J. B i d . Chem. 263, 7925-7928. Yarden, Y., Rodriguez, H., Wong, S. K.-F., Brandt, D. R., May, D. C., Burnier, J., Harkins, R. N., Chen, E. Y., Ramachandran, J., Ullrich, A,, and Ross, E. M. (1986). The avian P-adrenergic receptor: Primary structure and membrane topology. Proc. Nutl. Acad. Sci. U.S.A. 83, 67956799.

CURRENT TOPICS IN MEMBRANES AND TRANSPORT, VOLUME 36

Chapter 12 Mechano-Sensitive Ion Channels in Microbes and the Early Evolutionary Origin of Solvent Sensing CHING KUNG, YOSHIRO SAIMI, AND BORIS MARTINAC Laboratory of Molecular Biology and Department of Genetics University of Wisconsin-Madison Madison, Wisconsin 53706

I. Introduction 11. A Stretch-Activated Ion Channel of Escherichiu coli 111. A Stretch-Activated Ion Channel in Yeast IV. Touch Receptors and Channels of Paramecium V. Mechano-Sensitive Channels and the Concept of Solvent Senses References

1.

INTRODUCTION

A room without a door is but a tomb. One could therefore argue on first principles that some kind of portals should emerge with, or soon after, the primordial cell membranes. Speculations aside, we have now shown that protozoans, yeast, and even bacteria all have one class of such portals, the ion channels. It therefore appears that all cellular forms of life have ion channels, and we are forced to conclude that they must have emerged and evolved very early. Ion channels are gated pores. A certain stimulus can increase the probability of a given channel being open. Such a stimulus (gating principle) can be an external ligand (e.g., acetylcholine for the nicotinic acetylcholine receptor or channel); an internal second messenger (e.g., Ca2+ for Ca2+-gated K + channel, cGMP for the cGMP-gated channel in the rod outer segment); or cross-membrane voltage 145 Copynghl Q 1990 by Acadernlc Press. Inc All nghts of reproduction in any form reserved

146

CHING KUNG ET AL.

(e.g., the voltage-gated Na+ channel and delayed rectifier K + channel of nerves). These classes of channels have been extensively studied and reviewed (Hille, 1984). Other ion channels have subsequently been found to be gated by GTP-binding proteins (Yatani et a l . , 1988) or by arachidonic acid (Kim and Clapham, 1989; Ordway et al., 1989). Last, but not least, is a class of channels that are gated by mechanical forces in the membrane. The activities of the last type of channels, those gated by mechanical forces, have been studied in the hair cells of the inner ear (Howard et al., 1988) and in ciliated protozoans (Machemer and Deitmer, 1985). The activities of individual stretch-activated channels were first demonstrated by Guharay and Sachs (1985) in chick skeletal muscle with a patch clamp. Since then, stretch-activated channels have been found in neurons, endothelial cells, blood cells, eggs, cultured plant cells, and guard cells, through patch-clamp examinations (see Sachs, 1988, for a review). Surprisingly, we encountered stretch-activated ion channels in both the fission yeast Saccharomyces cerevisiae and the bacterium Escherichia coli in patchclamp survey of their membranes. These findings led us to speculate that they might represent ancient devices for the detection of water concentrations through osmotic pressure. Reviews of microbial channels can be found in Saimi et al., (1988a,b) and Martinac et al. (1988).

II. A STRETCH-ACTIVATED ION CHANNEL OF ESCHERlCHlA COLl A typical bacillus, being < 1 pm in diameter, is too small for electrophysiology, even with a patch-clamp electrode. However, there are ways to generate giant cells or giant spheroplasts, some 5-10 pm in diameter, sufficient in size for patch-clamp experiments. Escherichia coli cells can be grown into filaments 50150 pm long by culturing with cephalexin, which prevents septation. Lysozyme, together with EDTA, can then be applied to nick the peptidoglycan wall, thereby converting the filaments into giant spheroplasts (Ruthe and Adler, 1985). The activities of ion channels that we found in these spheroplasts are not artifacts of cephalexin or EDTA-lysozyme, since four other methods gave similar results. Some of these methods exclude the use of one or both of these agents (Buechner et a l . , 1990). There is a stretch-activated channel on the surface of E. coli (Fig. 1). Suction equivalent to a few centimeters of mercury applied on a membrane patch in an on-cell mode recording opens these channels. The applied forces are physiological. The pressure of 5 crn Hg column is equivalent to the pressure generated by <3 mOsm differences across the membrane. Depolarization also facilitates gating of the stretch-sensitive channel. The unit conductance measured at 300 mM

147

12. MECHANO-SENSITIVEION CHANNELS

suction:

Ton

off

t

on

1

off

FIG. 1. Activities of mechano-sensitive channels of Escherichia coli. Channels from an on-cell patch of a giant spheroplast were activated when suction was applied to the recording pipette, and they closed after release of the suction (arrows). This caused large jumps in the measured conductance. The voltage imposed on the patch in this example was - 10 mV.(From Martinac et al., 1987.)

salt solution is very large, being 950 pS at hyperpolarizing and 630 pS at depolarizing voltages. It has very little selectivity although it prefers anions slightly. Large organic ions such as glutamate can pass through this channel. The steady-state opening probability of single channels plotted against the applied pressure showed sigmoidal dependence on tension along the plane of the membrane (Martinac et al., 1987). The envelopes of gram-negative bacteria have two membranes: a more conventional inner membrane, and a specialized outer membrane. The two membranes are separated by the periplasmic space, which contains, among other material, the peptidoglycan cell wall. Experiments using a double mutant, Lpp-OmpA - , lacking the major proteins that anchor the outer membrane to the peptidoglycan, suggested that this channel is located in the outer membrane. Experiments in which excised patches were treated further with lysozyme also support this view. This treatment made the channel even more sensitive to stretch, suggesting that the peptidoglycan might be the element that resists the stretch and therefore restrains channel opening (Martinac et al., 1989). There are at least two other types of ion channels found in E. coli giant spheroplasts or upon reconstitution of E. coli membranes into liposomes (Delcour et al., 1989a,b). They are voltage-gated. Suctions large enough to break the patches cannot open these two channels, further supporting the view that activations of the above channel by small suctions are not artifactual.

111.

A STRETCH-ACTIVATED ION CHANNEL IN YEAST

The plasma membrane of the budding yeast, Saccharomyces cerevisiae, can be patch-clamped after removing the glucan cell wall with zymolyase to yield spheroplasts some 5-7 pm in diameter. Ion channel activities are routinely

148

CHlNG KUNG ET AL.

observed and recorded most commonly in cell-attached mode or whole-cell mode. We also discovered a stretch-activated channel in the plasma membrane of yeast spheroplast (Fig. 2) (Gustin et al., 1988). In whole-cell mode, this channel is activated upon application of pressure equivalent to a few centimeters of mercury exerted through the pipette. A dilution of the bath solution, equivalent to increasing an outward osmotic pressure, also opens these channels, monitored in the whole-cell mode. These channels have a unit conductance of -40 pS. They are poorly ion selective, passing both cations and anions. Laplace's law states

PRESSURE OFF

ON

OFF

41 msec

FIG. 2. Activities of mechano-sensitive channels of the budding yeast, Saccharomyces cerevisiae. On applying 29 IcNlm2 suction to the external surface of an inside-out patch excised from the surface of a yeast spheroplast and held at -60 mV, fluctuations of inward current were observed (top trace). Viewed at higher time resolution (bottom three traces), the current fluctuations appear quantal. With this patch, the conductance typically changed in a two-step manner (*) with both steps approximately equal in size. (From Gustin et al., 1988.)

149

12. MECHANO-SENSITIVE ION CHANNELS

that tension on the thin wall of a sphere is directly proportional to the product of pressure and diameter of the sphere. We found that less pressure is needed to open the channels in larger cells. Thus, our results are consistent with the idea that the channel molecule is gated by tension along the plane of the membrane and not by pressure perpendicular to it. A sudden shift to high pressure causes a transiently high channel activity, which soon relaxes to a low steady state during sustained pressure. Interestingly, this adaptation is dependent on negative voltages. Positive voltages allow little adaptation. A voltage-gated K -specific channel is also found in the plasma membrane of yeast (Gustin et a l . , 1986). Recently, we have also patch-clamped the vacuolar membrane of yeast and discovered ion channels in it (Minorsky et al., 1989). +

IV. TOUCH RECEPTORS AND CHANNELS OF PARAMECIUM Paramecium is the most thoroughly studied microbe in terms of electrophysiology. It is not a simple cell (see Machemer, 1988; Preston and Saimi, 1989, for recent reviews). Currents through nine different ion channels can be recorded using the method of two-electrode whole-cell voltage clamp. Seven of these channels are gated either by voltages (de- or hyperpolarizations) or by internal Ca2 . The remaining two are gated by mechanical forces. A touch at the anterior end of Paramecium triggers a depolarization (Eckert et a l . , 1972). This “receptor potential” can then trigger the voltage-gated Ca2+ channel to let in Ca2+. The behavioral consequence is the CaZ -driven reversal of the ciliary beat direction and a transient backward swimming. A touch at the posterior triggers a hyperpolarization (Naitoh and Eckert, 1973), which eventually leads to an augmentation of ciliary beat at a near normal direction and a forward spurt by a mechanism that is not fully understood, although CAMP appears to be involved (Bonini et a l . , 1986). As studied under voltage clamp, the anterior receptor current of Paramecium is camed readily by divalent but poorly by monovalent cations (Satow et a l . , 1983). The posterior response is based on a K + flux and can be blocked by tetraethylammonium (Eckert et a l . , 1972; Naitoh and Eckert, 1973). Similar studies with a different ciliate, Sfylonychia, led to the same conclusions (de Peyer and Deitmer, 1980; Deitmer, 1982; Machemer and Deitmer, 1985). Unlike in E . culi and yeast, activities of individual mechanically sensitive channels in ciliates have not been reported to date. This lack may have resulted from the method used in the patch-clamp study of Paramecium membrane, in which the integrity of the membrane-cytoskeleton complex (the cell cortex) might have been disrupted (Saimi and Martinac, 1989). +

+

150

CHING KUNG ET AL.

V.

MECHANO-SENSITIVE CHANNELS AND THE CONCEPT OF SOLVENT SENSES

Mechano-sensitive ion channels must have emerged and evolved early, since they are found in all these very different microbes including a lower eukaryote (yeast) and a prokaryote (E. coli). Those in Paramecium seem to function as touch receptors that generate mechano-receptor potentials. What functions do they serve in yeast and E. coli? One possibility is that they are used to monitor water concentration in the environment by measuring the osmotic pressure. Water is of obvious concern to life. Animals and plants have to deal with dehydration or overhydration. Microbes are no exception. Escherichia coli, for example, pumps K , synthesizes osmoprotectants, makes membrane-derived oligosaccharides, remodels its outer membrane, and changes its density upon osmotic changes. We have even demonstrated a behavioral avoidance of dehydration (osmotaxis) in E. coli (Li et al., 1988). Despite the wealth of information on the various responses to hydrating or dehydrating stresses, the molecular mechanism(s) that senses the changes of osmolarity remains unknown. Whether the stretch-activated channels in yeast or E. coli are actually the sensors of osmotic stress has yet to be proved. Biophysically, water has to be measured by a mechanism different from those that detect other molecules in solution. Recall that pure water is 55.56 M in concentration. Yet, cells seem to be able to detect and respond to millimolar changes in water concentrations. It therefore stands to reason that our ingrained notion on how chemicals (solutes) are detected by cells does not apply to water (the solvent). A water-binding pocket with a K , of nanomolar to millimolar simply would not do. Rain or drought changes the osmolarity outside the cell immediately and the internal osmolarity eventually because water can go through lipid bilayers with or without protein channels (Finkelstein, 1987). These changes result in osmotic pressure on the cell membranes. Stretch-activated ion channels can indeed be opened by changing osmolarity (Gustin et al., 1988; B . Martinac et al., unpublished observations). We would like to speculate that primordial cells developed molecular devices to measure water concentration (including channels and other receptors) which are categorically different from devices evolved to detect other chemicals. Because water detectors are transducers of mechanical forces as discussed earlier, it is not difficult to imagine their eventually becoming the molecular mechanisms for touch, vibrations, hearing, balance, proprioception. We propose to call these senses collectively the solvent senses. Solute detectors, perhaps first used to detect nutrients and wastes, could be remodeled to serve in taste, smell, in the reception of pheromone, hormone, or neurotransmitter, and in vision. They can be collectively called the solute senses. Figure 3 represents this speculation. So categorized, it is interesting to note that many of the molecular mechanisms of +

151

12. MECHANO-SENSITIVEION CHANNELS

vision

hearing

\

transmitters

If

smell

balance proprioception

taste

hormones pheromones

touch

\J

vibration

nutrients

SOLUTE SENSES

SOLVENT SENSES

FIG.3. A scheme of evolution of senses. Solute senses use proteins with specific binding pockets. Solvent senses use different types of proteins that measure mechanical forces. The scheme emphasizes the completely separate origins of solute senses and solvent senses. The order and timing of the emergence of different senses are less critical here.

solute senses are worked out, including the now famous G-protein circuitry. On the other hand, the molecular mechanisms for the solvent senses are completely obscure. There is presently no structural information, from primary sequence to three-dimensional modeling, on any mechano-sensitive proteins. Two caveats need to be pointed out. First, mechano-sensitive channels may have another function, that of morphogenesis. They may well be devices for the cell to measure its own size and shape (Christensen, 1987; Falke and Misler, 1989). However, osmosensation and morphogenesis are not necessarily mutually exclusive. The same channels may serve both functions. Second, there might well be convergent evolution resulting in channels of different molecular designs serving the same function. Note that in cilated protozoans the anterior touch

receptor is selective for divalent cations, and the posterior touch receptor is K specific, while most known mechano-sensitive channels are poorly ion selective. Note also that only the mechano-sensitive channel in E . coli is known to have a huge conductance and can pass large ions. Models of most channels have the ion pathways lined by transmembrane a helices, each contributed by one subunit or repeat of the channel protein complex. On the other hand, sequences of bacterial porins (channels in the outer membrane of gram-negative bacteria) and voltagedependent anion channels (VDAC) in the outer membrane of mitochondria suggest that they are made of P-pleated sheets (Paul and Rosenbusch, 1985; Rosenbusch, 1986; Forte et al., 1987). The stretch-activated channel we found in the +

152

CHING KUNG ET AL.

outer membrane of E . coli may be structurally very different from those in yeast or other cells, for example. We have yet to learn the structural detail of the microbial channels reviewed here. ACKNOWLEDGMENT Work in our laboratory and collaborating laboratories headed by Professors Julius Adler and Michael Culbenson is supported in part by NIH GM22714, GM36386, GM37925, DK29121, and a grant from the Lucille P. Markey Trust. REFERENCES Bonini, N. M., Gustin, M. C., and Nelson, D. L. (1986). Regulation of ciliary motility by membrane potential in Paramecium: A role for cyclic AMP. Cell Motil. Cytoskeleton 6, 256-272. Buechner, M., Delcour, A. H., Martinac, B., Adler, J., and Kung, C. (1990). Ion channel activities in the Escherichia coli outer membrane. Biochim. Biophys. Acta (in press). Christensen, 0. (1987). Mediation of cell volume regulation by Ca*+ influx through stretch-activated channels. Nature (London) 330, 66-68. Deitmer, J. W. (1982). The effects of tetraethylammonium and other agents on the potassium mechanoreceptor current in the ciliate Stylonychia. J. Exp. Biol. 96, 239-249. Delcour, A. H.,Martinac, B., Adler, J., and Kung, C. (1989a). A modified reconstitution method used in patch-clamp studies of Escherichia coli ion channels. Biophys. J . 56, 631-636. Delcour, A. H., Martinac, B., Adler, J., and Kung, C. (1989b). A voltage-sensitive ion channel of Escherichia coli. J . Membr. Biol. 112, 267-275. de Peyer, J. E., and Deitmer, J. (1980). Divalent cations as charge carriers during two functionally different membrane currents in the ciliate Stylonychia. J. Exp. Biol. 88, 73-89. Eckert, R., Naitoh, Y., and Friedman, K. (1972). Sensory mechanisms in Paramecium. I. Two components of the electric response to mechanical stimulation of the anterior surface. J. Exp. Biol. 56, 683-694. Falke, L. C., and Misler, S. (1989). Activation of ion channels during volume regulation by clonal NlE115 neuroblastoma cells. Proc. Natl. Acad. Sci. U.S.A. 86, 3919-3923. Finkelstein, A. (1987). “Water Movement through Lipid Bilayers, Pores, and Plasma Membranes: Theory and Reality,” Distinguished Lecture Series of the Society of General Physiologists, Vol. 4. Wiley, New York. Forte, M., Guy, H. R., and Mannella, C. A. (1987). Molecular genetics of the VDAC ion channel: Structural model and sequence analysis. J . Bioenerg. Biomembr. 19, 341-349. Guharay, F., and Sachs, F. (1985). Mechanotransduction ion channels in chick skeletal muscle: The effects of external pH. J . Physiol. (London) 363, 119-134. Gustin, M. C., Martinac, B., Saimi, Y., Culbertson, M. R., and Kung, C. (1986). Ion channels in yeast. Science 233, 1195-1 197. Gustin, M. C., Zhou, X.-L., Martinac, B., and Kung, C. (1988). A mechanosensitive ion channel in the yeast plasma membrane. Science 242, 762-765. Hille, B. (1984). “Ion Channels of Excitable Membranes.” Sinauer, Sunderland, Massachusetts. Howard, J . , Roberts, W. M., and Hudspeth, A. J. (1988). Mechano-electrical transduction by hair cells. Annu. Rev. Biophys. Biophys. Chem. 17, 99-124. Kim, D., and Clapham, D. E. (1989). Potassium channels in cardiac cells activated by arachidonic acid and phospholipids. Science 242, 1174-1 176. Li, C.-Y., Boileau, A. J., Kung, C., and Adler, J. (1988). Osmotaxis in Escherichia coli. Proc. Natl. Acad. Sci. U.S.A. 85, 9451-9455. Machemer, H. (1988). Electrophysiology. In “Paramecium” (H. D. Gortz, ed.), pp. 186-215. Springer-Verlag, Berlin and New York.

12. MECHANO-SENSITIVE ION CHANNELS

153

Machemer, H., and Deitmer, J. W. (1985). Mechanoreception in ciliates. In “Progress in Sensory Physiology” (H. Autrum er a/., eds.), Vol. 5, pp. 81-118. Springer-Verlag, Berlin and New York. Martinac, B., Buechner, M., Delcour, A. H., Adler, J., and Kung, C. (1987). Pressure-sensitive ion channel in Escherichia coli. Proc. Natl. Acad. Sci. U.S.A. 84, 2297-2301. Martinac, B., Saimi, Y., Gustin, M. C., Culbertson, M. R., Adler, J., and Kung, C. (1988). Ion channels in microbes. Period. Biol. 90,375-384. Martinac, B., Buechner, M., Delcour, A. H., Adler, J., and Kung, C. (1989). A mechanosensitive ion channel in the bacterium Escherichia co/i. Proc. Intern. Union Physiol. Sci. 17, 451-452. Culbertson, M. R., and Kung, C. (1989). A patch clamp analysis of a Minorsky, P V., Zhou, X.-L., cation-current in the vacuolar membrane of the yeast Saccharomyces. Plant Physiol. Abstr. 89, Abstr. 882. Naitoh, Y., and Eckert, R. (1973). Sensory mechanisms in Paramecium. 11. Ionic basis of the hyperpolarizing mechanoreceptor potential. J. Exp. Biol. 59, 53-65. Ordway, R. W., Walsh, J. V., Jr., and Singer, J. S. (1989). Arachidonic acid and other fatty acids directly activate potassium channels in smooth muscle cells. Science 244, 1176-1 179. Paul, C., and Rosenbusch, J. P. (1985). Folding patterns of porin and bacteriorhodopsin. EMBO J. 4, 1593- 1597. Preston, R. R., and Saimi, Y. (1989). Calcium ions and the regulation of motility in Paramecium. In “The Structure and Function of Ciliary and Flagellar Surfaces” (R. Bloodgood, ed.). Plenum, New York. Rosenbusch, J. P. (1986). Three-dimensional structure of membrane proteins. In “Bacterial Outer Membranes as Model Systems” (M. Inoyue, ed.), pp. 141-162. Wiley, New York. Ruthe, H.-J., and Adler, J. (1985). Fusion of bacterial spheroplasts by electric fields. Biochim. Biophys. Acra 819, 105- 11 3. Sachs, F. (1988). Mechanical transduction in biological systems. CRC Crit. Rev. Biomed. Eng. 16, 141- 169. Saimi, Y., and Martinac, B. (1989). A calcium-dependent potassium channel in Paramecium studied under patch-clamp. J. Membr. Biol. 112, 79-89. Saimi, Y., Martinac, B., Gustin, M. C., Culbertson, M. R., Adler, J., and Kung, C. (1988a). Ion channels in Paramecium, yeast, and Escherichia coli. Trends Biochem. Sci. 13, 304-309. Saimi, Y., Martinac, B., Gustin, M. C., Culbertson, M. R., Adler, J., and Kung, C. (1988b). Ion channels in Paramecium, yeast, and Escherichia coli. Cold Spring Harbor Symp. Quanr. Biol. 53, 667-673. Satou., Y., Murphy, A. D., and C. Kung, (1983). The ionic basis of the depolarizing mechanoreceptor potential of Paramecium rerraurelia. J . Exp. Biol. 103, 253-264. Yatani, A., et al., (1988). The G protein-gated atrial K + channel is stimulated by three distinct G, Qsubunits. Nature (London)336, 680-682.

This Page Intentionally Left Blank

CURRENT TOPICS IN MEMBRANES AND TRANSPORT, VOLUME 36

Chapter 73

Selection of an ap T-cell Antigen Receptor in Vivo and Expression in Vitro in a Soluble Form M . M . DAVIS,"? B . FAZEKAS DE ST. GROTH,*f L . J . BERG, * f A . LIN, * f B . DEVAUX, * f C . SAGERSTROM, f J . F. ELLIOTT, f AND P . J . BJORKMANf *Howard Hughes Medical Institute and fDepartment of Microbiology and Immunology Stanford University School of Medicine Stanford, California 94305

I. 11. 111. IV.

Introduction Positive Selection Negative Selection Soluble T-cell Receptor Heterodimers References

1.

INTRODUCTION

Much progress has now been made in terms of identifying and characterizing the molecules important in T-cell recognition, both receptors and ligands. This work has established that T cells recognize fragments of antigens (Ag) embedded in major histocompatibility complex (MHC) molecules. Many issues remain unresolved, however, among which are the following: How and on what basis are T-cell receptors (TCR) selected in the thymus? What is the precise nature of the TCR-Ag-MHC recognition event? To address these issues, we have developed two types of experimental systems. The first involves mice transgenic for specific TCR genes crossed onto different MHC backgrounds. With suitable serological probes we can follow the expression of both a and p chains and the progress (or lack thereof) of T cells bearing these receptors in the animal. Together with the work of von Boehmer and colleagues (Kisielow et al., 155 Copynght 0 1990 by Academic Press, Inc All nghts of reproduction in any form reserved

156

M. M. DAVIS ET AL.

1988a,b) and Loh and colleagues (Sha et al., 1988a,b), the results of these studies give us some very clear ideas about the reality and efficiency of both positive and negative selection in the thymus. In addition, these transgenic systems should make possible the biochemical characterization of these phenomena in the near future. A second area we have been interested in concerns how best to make soluble forms of TCR heterodimers in order to move the bases for discussion from a cellular readout (e.g., T-cell activation assays) to a molecular one (such as binding and structural studies). To this end we have recently been successful in expressing both a mouse and a human a p heterodimer in a lipid-linked fashion, which can then be cleaved off the cell surface to produce soluble protein. Preliminary indications are that this type of expression of ap TCR (in the absence of CD3) is faithful to the original and can take us a long way toward the goal of better understanding the troika of T-cell recognition (TCR-Ag-MHC) at a molecular level.

II. POSITIVE SELECTION While the ligand for ap TCR in the periphery (Ag-MHC) seems clear, what T-cell receptors see in the thymus has been a very controversial issue. Kisielow er al. (1988b) have provided data indicating that positive selection for a class 1 + HY-specific TCR can be provided by the original restricting element (Db). Similarly, Sha et al. (1988b) have obtained evidence that the K molecule may be the positively selecting element in their transgenic system. In our own work, we have recently obtained convincing evidence that the original restricting element of the 2B4 TCR, namely, I-E, is necessary and sufficient for the efficient export of transgene-expressing T cells from the thymus. We have demonstrated this in two ways. (Berg et al., 1989b): (1) Peripheral T cells expressing both a- and pchain TCR of 2B4 are much more abundant in H-2k-bearing mice than in H-2b homozygotes (which lack functional I-E expression) as judged by surface staining. This effect can also be seen in H-2b mice that carry a functional E , transgene. (2) The frequency of moth cytochrome c-reactive T cells in the periphery of H-2k mice is at least 10 times higher than in H-2b homozygotes. Again, this can also be seen on b haplotype animals bearing E , transgenes. In addition, E , deletion mutants that express I-E predominantly in either the medullary or cortical regions of the thymus (van Ewijk er al., 1988) show clearly that positive selection can occur even with MHC expression only on cortical epithelial cells. In contrast, medullary MHC expression has no apparent effect (Berg et al., 1989b; Benoist and Mathis, 1989). The general thymic phenotype of

13. SELECTION OF AN ap T-CELL ANTIGEN RECEPTOR

157

ap-transgenic mice in the absence of an appropriate restricting element is that of an arrest at the double-positive (CD4+ 8 + ) stage of differentiation. That is, very few single-positive thymocytes are detectable. In addition, we see a strong disposition toward CD4 expression, especially in peripheral cells in the presence of the restricting element (1-E). This is consistent with the observations of Kisielow et a1. (1988a) and Sha et al. (1988a) in which class I MHC-specific TCR transgene expression promotes a pronounced skewing toward CD8 expression. Thus it seems that the MHC class specificity of a given TCR somehow determines its accessory molecule (CD4 or CD8) phenotype. It would be interesting to determine how this is accomplished.

111.

NEGATIVE SELECTION

In the 2B4 transgenic system we have also encountered negative selection based on Mls reactivity, which helps to resolve some of the current issues in the literature (Berg et al., 1989a). In particular, there is a quantitative removal of cells expressing a high level of the 2B4P chain because of the combination of its Vp3 component and the Mls 2A/3Agenotype (C3H/HeJ) of one of the parental strains (as defined by Pullen et al., 1988). We found deletion of mature 2B4Ppositive cells to be evident in the thymuses of both 2B4 cwp TCR mice and 2B4P mice. What was interesting, however, was the very different phenotype exhibited by these two types of mice, both of which were undergoing massive and efficient negative selection. The thymic CD4/CD8 profile indicates that the p mice had a very similar arrest in the double-positive stage as seen in the H-2b aP 2B4 transgenics or in normal mice undergoing Mls-mediated deletion (Kappler et al., 1988). In contrast, the ap mice that deleted V63 had a greatly reduced percentage of double-positive cells (2% of the total) and overall less than one-tenth of the normal number of thymocytes, exactly as reported by Kisielow et al. (1988a) for HY + Db, and as seen in the high-expressing mice of Sha et al. (1988b). We therefore conclude that the presence of up TCR transgenes greatly augments the effects of negative selection, perhaps by speeding up the kinetics of T-cell differentiation and selection in the thymus (Berg et al., 1989a). This is in fact predictable from the observations that the TCR a chain is the last TCR to be rearranged and expressed during thymic development, and thus the presence of a rearranged a chain is bound to have an effect on maturation of at least some T cells. In fact, current data (Fazekas de St. Groth et al., 1990) indicate that TCR a transgenics express ap TCR fully 2 days before normal (day 15 vs. day 17) in fetal mice and at inappropriately high levels. This effect may be due to the absence of normal controlling elements, especially the silencer regions 3' of C , recently described by A. Winoto and D. Baltimore (personal communication).

158

M. M. DAVIS ET AL.

Another important point is that this study of negative selection shows the essential equivalence of Mls-mediated deletion versus Ag + MHC (Kisielow et al., 1988a) and alloreactive (Sha et al., 1988b) deletion in @-transgenic mice. This indicates that Mls-mediated deletion is, in fact, a valid model for the developmental aspects of self-tolerance.

IV. SOLUBLE T-CELL RECEPTOR HETERODIMERS Because the TCR heterodimer is assembled with the CD3 polypeptides to form a complex of at least seven polypeptides before appearing on the surface (Minami et al., 1987), it may be one of the more difficult molecules to produce in a soluble form. On the other hand, the very Ig-like character of its V, D,J , and C region elements makes it very likely that (1) it can bind to Ag-MHC by itself and (2) that structurally it could exist in solution much like an antibody. Thus the challenge has been to find conditions in which TCR chains can be expressed and form heterodimers free of CD3 molecules, either secreted from cells in culture or in a membrane-associated form that could be easily cleaved from the surface. Initial attempts in our lab took the form of TCR(V)-lg(C) hybrids expressed in myeloma cell lines (Gascoigne et al., 1987). Interestingly, only V , (TCR) C, (Ig) chimeras could be assembled or secreted as apparently normal Ig molecules (being assembled and expressed with light chains). Neither V6 (Gascoigne et al., 1987), V y , nor V , (R. Wallich et al., unpublished observations; MacNeil et a / . , unpublished observations) gene segments could be expressed in that same context, suggesting that there is some structural barrier to proper folding. Recently, we became aware of the increasingly large list of surface proteins known to be lipid linked (as reviewed in Ferguson and Williams, 1988). All such proteins can be cleaved from the surface with a specific enzyme, PI-PLC, to produce soluble forms. Caras et al. (1987) had shown that the C-terminal 37 amino acids of decay-accelerating factor (DAF) could serve as the signal sequence for the lipidlinked expression of a Herpes simplex virus membrane protein. We decided to apply this finding to TCR in the hope that expressing TCR polypeptides as PIlinked molecules might allow them to associate in the plane of a membrane to maximize heterodimer formation and to be cleaved off the surface of expressing cells to produce a soluble form. Thus far we find that, in most cases, TCR-PI-signal sequence chimeras are expressed well in both COS cells (for transient transfections) and CHO cells (for long-term transfectants). Single chains of the TCR could be expressed with high efficiency and cleaved off the surface of expressing cells with PI-PLC (in early experiments, the kind gift of Dr. Martin Low). Expression could be demonstrated by either the immunoprecipitation of surface-iodinated cells or fluorescence tagging and analysis by microscopy or FACS . immunoprecipitation indicates both

13. SELECTION OF AN ap T-CELL ANTIGEN RECEPTOR

159

dimers and monomers, in most cases (Lin et a / ., 1990; Devaux et al., 1990). TO demonstrate that heterodimers between OL and p were being formed that juxtapose V, and Vrj determinants, we made use of the fact that preincubation of 2B4 TCR-bearing cells with saturating amounts of anti-V,,,, (A2B4-2 from Samelson et a/., 1983) antibody abolished staining with the anti-Vp, antibody (KJ25 from Pullen et a/., 1988). This same cross-blocking phenomenon can also be shown to occur in CHO cells expressing the 2B4 OL and p chains in a lipidlinked form (Lin et al., 1990). Current we are able to purify hundreds of micrograms of soluble receptor heterodimers. Perhaps the most crucial test is whether TCR expressed in this way can bind to Ag-MHC complexes. Preliminary indications are that iodinated soluble 2B4 TCR can bind to I-@-bearing cells in the presence of the appropriate moth cytochrome c polypeptide. Thus, we are very hopeful of being able to bring T-cell recognition into the biochemical and biophysical realm. ACKNOWLEDGMENTS We wish to thank Dr. lay Bangs for advice concerning PI-linked proteins, Dr. Martin Low for PIP E , Angela Nervi for excellent technical assistance, and Brenda Robertson for preparation of the manuscript. We also thank NIH for grant support. Mark M. Davis is a scholar of the P.E.W. Foundation, Leslie J. Berg is a fellow of the Leukemia Society of America, Augustine Lin was previously a fellow of the Cancer Research Institute and is now supported by the Howard Hughes Medical Institute. Barbara Fazekas de St. Groth was supported by an Irvington House postdoctoral fellowship and is now supported by a fellowship from the National Health and Medical Research Council of Australia. Pamela J. Bjorkman was supported by an American Cancer Society postdoctoral fellowship, and John F. Elliott is a centennial fellow of the Medical Research Council of Canada.

REFERENCES Benoist, C., and Mathis, D. (1989). Positive selection of the T cell repertoire: Where and when does it occur? Cell 58, 1027-1033. Berg, L. J., Fazekas de St. Groth, B., Pullen, A. M., and Davis, M. M. (1989a). Phenotypic differences between a p versus p T-cell receptor transgenic mice undergoing negative selection. Nature (London) 340, 559-562. Berg, L. I., Pullen, A. M., Fazekas de St. Groth, B., Mathis, D., Benoist, C., and Davis, M. M. (1989b). Antigen/MHC-specific T cells are preferentially exported from the thymus in the presence of their MHC ligand. Cell. 58, 1035-1046. Caras, I. W., Weddell, G. N., Davitz, M. A., Nussenzweig, V., and Martin, D. W. (1987). Signal for attachment of a phospholipid membrane anchor in decay accelerating factor. Science 238, 1280- 1283. Devaux, B. et al. (1990). In preparation. Fazekas de St. Groth, B. et at. (1990). In preparation. Ferguson, M. A. J., and Williams, A. (1988). Cell-surface anchoring of proteins via glycosylphosphatidylinositol structures. Annu. Rev. Biochem. 57, 285-320. Gascoigne, N. R. M., Goodnow, C., Dudzik, K . , Oi, V. T., and Davis, M. M. (1987). Secretion of a

160

M. M. DAVIS ET AL.

chimeric T-cell receptor-immunoglobulin protein. Proc. Narl. Acad. Sci. U.S.A. 84, 29362940. Kappler, J. W., Staerz, U., White, J., and Marrack, P. C. (1988). Self-tolerance eliminates T cells specific for Mls-modified products of the major histocompatibility complex. Nature (London) 3332, 35-40. Kisielow, P., Bluthmann, H., Staerz, U. D., Steinmetz, M., and von Boehmer, H. (1988a). Tolerance in T cell receptor transgenic mice involves deletion of nonmature CD4f8 + thymocytes. Nature (Londonj 333, 742-746. Kisielow, P., Teh, H. S . , Bluthmann, H., and von Boehmer, H. (1988b). Positive selection of antigen-specific T cells in thymus by restricting MHC molecules. Nature (London) 335, 730733. Lin, A. er al. (1990). In preparation. Minami, Y., Weissman, A. M . , Samelson, L. E., and Klausner, R. D. (1987). Building a rnultichain receptor: Synthesis, degradation, an assembly of the T-cell antigen receptor. Proc. Nutl. Acad. Sci. U.S.A. 84, 2688-2692. Pullen, A. M.,Marrack, P., and Kappler, J. W. (1988). The T cell repertoire is heavily influenced by tolerance to polymorphic self-antigens. Nature (London) 335, 796-801. Samelson, L. E., Germain, R. N., and Schwartz, R. H. (1983). Monoclonal antibodies against the antigen receptor on a cloned T-cell hybrid. Proc. Natl. Acad. Sci. U.S.A. SO, 6972-6976. Sha, W. C., Nelson, C. A., Newberry, R. D., Kranz, D. M., Russell, J. H., and Loh, D. Y. (1988a). Selective expression of an antigen receptor on CDI-bearing T lymphocytes in transgenic mice. Nature (London) 335, 27 1-274. Sha, W. C . , Nelson, C. A., Newberry, R. D., Kranz, D. M., Russell, J.H., andLoh, D. Y. (1988b). Positive and negative selection of an antigen receptor on T cells in transgenic mice. Nature (London) 336, 73-76. van Ewijk, W., Ron, R., Monaco, J., Kappler, J., Marrack, P., Le Mew, M., Gerlinger, P., Durand, B., Benoist, C., and Mathis, D. (1988). Compartmentalization of MHC class I1 gene expression in transgenic mice. Cell 53, 357-370.

CURRENT TOPICS IN MEMBRANES AND TRANSPORT, VOLUME 36

Chapter 14 Perforin and the Mechanism of Lymphocyte-Mediated Cytolysis ECKHARD R . PODACK AND MATHIAS G . LICHTENHELD Department of Microbiology and Immunology University of Miami School of Medicine Miami, Florida 33101 I.

Introduction Physicochemical and Functional Properties of Perforin Sequence of Murine and Human Perforin IV. Homology of Perforin to Complement Proteins: The Perforin Family V. Lack of Homologous Restriction of Perforin VI. Expression of Perforin mRNA in V i m and in Vivo VII. The Contribution of Membrane Pores to DNA Degradation VIII. Conclusions References Note Added in Proof 11. 111.

1.

INTRODUCTION

The mechanism of lymphocyte-mediated cytolysis is controversial (see recent reviews: Bleackley et al., 1988; Brunet et al., 1988; Clark et al., 1988; Jenne and Tschopp, 1988a; Mueller et a l . , 1988; Muller-Eberhard, 1988; Munger et al., 1988; Nagler-Anderson et al., 1988; Sitkovsky, 1988; Young et al., 1988; Podack et a l . , 1988a,b; Jenne and Tschopp, 1988b; Bleackley, 1988; Hershberger et al., 1988; Stevens et al., 1988). The main contention concerns the question whether secretory functions of the killer cell as proposed by the vectorial granule secretion model of cytotoxicity (Podack 1985; Henkart, 1985) are required for target cell lysis. Another open question is related to the role of DNA degradation in target cell lysis (Russel et al., 1982; Duke et a l . , 1983; Love11 and Martz, 1987) and how this process is mediated molecularly. Unfortunately, no molecular tools are available at the present time to assess the importance or even existence of the putative secretion-independent pathway of lymphocyte-mediated cytolysis. 161 Copynght 0 1990 by Academic Press. Inc. All nghrs of reproduction in any form IexNed.

162

ECKHARD R. PODACK AND MATHIAS G. LICHTENHELD

Undisputed, on the other hand, is the cytolytic activity of perforin isolated from the granules of cytolytic T cells and natural killer (NK) cells (Podack and Dennert, 1983; Dennert and Podack, 1983; Millard et al., 1984; Podack and Konigsberg, 1984; Henkart et al., 1984; Young et al., 1986a,b; Masson and Tschopp, 1985; Masson et al., 1985; Podack et al., 1985; Zalman et al., 1986; Podack, 1987; Tschopp et al., 1986). The mechanism of perforin-mediated lysis appears to require the vectorial secretion of perforin alone or of perforin associated with the cytoplasmic granules of the killer cell onto the target membrane following recognition and conjugate formation. Perforin has recently been isolated and characterized including sequence determination by cDNA cloning (Lichtenheld et al., 1988; Shinkai et al., 1988; Lowrey et al., 1989). The tools generated by this work are being used for a critical assessment of the role of perforin as opposed to other molecules in lymphocyte-mediated cytolysis. In this review, we will summarize the current state of knowledge of the structure and function of perforin. The properties of other granule proteins, granzymes and proteoglycan, have been reviewed (Jenne and Tschopp, 1988a,b; Stevens et al., 1988; Bleackley, 1988; Hershberger et al., 1988) and will not be described here.

II. PHYSICOCHEMICAL AND FUNCTIONAL PROPERTIES OF PERFORIN Perforin is a glycoprotein with M , 70,000 upon reduction and sodium dodecyl sulfate-polyacrylamide gel electrophoresis (Masson and Tschopp, 1985; Podack et al., 1985; Henkart, 1985; Young et al., 1986a). It binds to heparin with high affinity and to anion exchange resins at neutral pH. Perforin is localized in the cytoplasmic granules of cytolytic T cells and NK cells (Grosscurth et al., 1987), where it is in all likelihood associated with the granule proteoglycan chondroitin sulfate A (McDermott et al., 1985). Because they contain perforin, isolated cytoplasmic granules are highly cytolytically active in the presence of Ca (Millard et al., 1984; Podack and Konigsberg, 1984; Young et al., 1986a; Allbritton et al., 1988; Criado et al., 1985; Masson et al., 1985). In granule preparations lacking cytolytic activity despite their content of perforin, lytic activity can be generated by salt extraction (Lichtenheld et al., 1988), indicating that ionic interaction of perforin with granule proteoglycans. Isolated perforin is a cytolytic molecule. Cytolysis is mediated by polymerization of perforin in a Ca-dependent reaction within the attacked membrane to a transmembrane channel of varying diameter. According to electron-microscopic and electrophysiological evidence, the maximal channel size is 16 nm (Podack and Dennert, 1983; Young et al., 1986a; Blumenthal et al., 1984; Criado et al., 1985). This channel is formed by a tubular (cylindrical) complex of -20 perforin

-

14. PERFORIN AND LYMPHOCYTE-MEDIATED CYTOLYSIS

163

protomers. The polyperforin tubule is 16 nm long and inserted 4 nm deep into the lipid bilayer of the membrane (Fig. 1). Smaller transmembrane channels are formed when <20 perforin molecules form an incompletely closed polyperforin complex. Though not directly demonstrated so far, the theoretical limit is a perforin dimer with a functional channel diameter of 1-2 nm. Transmembrane channel formation by oligomeric complexes is explained by the putative structure of the membrane-spanning domain, which is postulated to be hydrophobic on one face and hydrophilic on the other. Such a structure could be formed by an amphipathic helix or an amphipathic P-pleated sheet. Perforin polymerization is mediated by Ca or Sr ions but not by Mg or Mn. In the presence of Ca, perforin at low temperature (4°C) binds reversibly to natural membranes without causing lysis (Young et al., 1987; Yue et al., 1987; Tschopp et al., 1989). Addition of EDTA allows the dissociation of perforin in active form from the membrane. At elevated temperature, membrane-associated perforin (via Ca) irreversibly inserts and polymerizes to a lytic transmembrane channel that cannot be dissociated by EDTA. These functional properties of perforin are quite analogous to those of C9, which reversibly binds to C5b-8 in the cold and at elevated temperature undergoes restricted unfolding and polymerization with the simultaneous appearance of hydrophobicity (Fig. 2) (for review see Podack, 1986; see also Podack and Tschopp, 1982; Tschopp et al., 1982, 1985; Podack et al., 1982). A unique property of perforin is its high affinity for phosphorylcholine (Tschopp et al., 1989; Yue et al., 1987), which acts as a Ca-dependent receptor and thereby initiates its membrane insertion and polymerization. In the case of C9, C5b-8 is required and the interaction is Ca-independent. Even though the polymerization reaction of C9 is Ca-independent and proceeds in the presence of EDTA, C9 binds one molecule of Ca with high affinity in the N-terminal part of the molecule (Thielens et al., 1988). It will be important to determine whether perforin also is capable of binding Ca, and whether within cytolytic granules it is complexed with Ca. The current procedures for perforin isolation use EGTA- or EDTA-containing buffers, which chelate any previously bound Ca.

111.

SEQUENCE OF MURINE AND HUMAN PERFORIN

Perforin cDNA from murine and human cytolytic lymphocytes was recently cloned and sequenced (Shinkai et al., 1988; Lowrey et al., 1989; Lichtenheld et al., 1988). The comparison of the murine and human sequence is shown in Fig. 3. On the protein level the two proteins are 67% identical. The mature proteins are 534 amino acids long and are preceded by a 20 (murine)- or 21 (human)amino acid-long leader peptide, consistent with the localization of perforin in membrane-enclosed granules. Perforin has two potential N-linked glycosylation

164

ECKHARD R. PODACK AND MATHIAS G. LICHTENHELD

14. PERFORIN AND LYMPHOCYTE-MEDIATEDCYTOLYSIS

165

FIG. 1. (A) Polyperforin membrane lesions as seen by negative-staining electron microscopy, The internal diameter of the circular complexes is 16 nm. The inset shows several polyperforin tubules in side view on the edge of a membrane vesicle. (B) Schematic view of a cross section of a membrane-inserted polyperforin tubule. The membrane and the tubular complex are drawn to scale.

sites, which however, are not in the same position in the murine and human proteins. The coding region of murine perforin is preceded by 148 untranslated nucleotides that do not contain typical ribosomal binding sequences (see Note Added in Proof). The stop codon is followed by 1100 nucleotides of 3'-untranslated

-

c 9

poly c 9

P1

poly P 1

FIG.2. Schematic drawing of the concerted polymerization-insertion process of C9 and perforin (Pl) in membranes.

166

ECKHARD R. PODACK AND MATHIAS G. LICHTENHELD

FIG. 3. Homology of the cDNA and deduced amino acid sequence of human and mouse perforin. Regions of high homology on the protein level are underlined. Dashed boxes: putative membranebinding region. Solid box: Cysteine-rich domain homologous to the EGF precursor.

167

14. PERFORIN AND LYMPHOCYTE-MEDIATED CYTOLYSIS

sequence. The total length of the sequenced cDNA is 2920 bp. This length is close to the observed length of the mRNA transcript of perforin of 2.9 kb detected in Northern blots of perforin-expressing cells.

IV.

HOMOLOGY OF PERFORIN TO COMPLEMENT PROTEINS: THE PERFORIN FAMILY

Even before the sequence of perforin was established, it was evident from functional (Podack and Dennert, 1983; Podack and Tschopp, 1982), morphological (Podack, 1984, 1986), and immunological comparisons (Podack, 1987; Young et al., 1986b; Zalman et al., 1986) that perforin and C9 may be related proteins. The sequence comparison now available fully confirmed these predictions. Figure 4 shows the composition of mosaic domains of perforin and of complement component C7, C8, and C9. The central part of perforin, containing the putative membrane-binding region and a cysteine-rich domain typical for the epidermal growth factor (EGF) precursor, is homologous to similar regions in C9 (DiScipio et al., 1984; Stanley et al., 1985), C8 (Rao et al., 1987; Haefliger et al., 1987; Howard et al., 1987), C7 (DiScipio et al., 1988), and presumably C6 (Podack, 1984). The N-terminal44 and the C-terminal 124 amino acids of perforin are different from those of the complement proteins. Compared to C9, the membrane-binding domain of perforin is thus shifted by 120 amino acids toward the N terminus. This situation makes the model for membrane insertion of C9 as proposed by Stanley (1988) rather unlikely for perforin. c7

Membrane-binding domain

EGF receptorlike sequence

Thrombospondinlike sequence

LDL receptorlike sequence C5-bindingsequence?

FIG. 4. Domain structure of the complement proteins C7, C8, C9, and petforin (Pl). The central domains encompassing the membrane-binding sites and the EGF receptorlike domain are homologous. Other domains are homologous only among the complement proteins.

168

ECKHARD R. PODACK AND MATHIAS G. LICHTENHELD

It is of interest to note that the complement proteins in addition to the two central domains of perforin have evolved three additional homology regions. The implication of that finding may entail a function of those domains for the interaction of the C7, C8, and C9 proteins to form the heteropolymeric tubule of the membrane attack complex of complement (Podack, 1984). Evolutionarily, it is likely that a perforinlike molecule gave rise to the complement proteins of the membrane attack complex of complement. We therefore propose to name the members of this protein family as the perforin family (Podack, 1987). Figure 5 shows a hypothetical pedigree of the current known members of that protein family.

V.

LACK OF HOMOLOGOUS RESTRICTION OF PERFORIN

Complement is under strict control to prevent lysis of host cells. Cytotoxic T cells, on the other hand, are required to lyse host cells when they become infected by virus, since cell lysis is the primary host defense against virus infection. For protection of host cells from complement lysis, two host membrane-associated proteins have evolved, named decay-accelerating factor (DAF) (for review see Medof, 1988) and homologous restriction factor (HRF) (for review see Hansch, 1988). Both proteins appear to be anchored via a glycolipid anchor in the host cell membrane. Whereas DAF acts at the level of C3b and decays or prevents formation of the C 3 , C5 convertase, HRF has not been (PRIMORDIAL PERFORIN) HUUORAL

CELLULAR

(C9 LIKE) -ORIN)

1

A

I

(PERFORIN)

c9

1

c9

(PERFORIN)

i" A T

C8a

C6

C7

CSP

1

PERFORIN

0

PROTZIMS IM BRACKETS ARE INFERRED

FIG. 5 .

A hypothetical pedigree for the pore-forming proteins of complement and of T cells

169

14. PERFORIN AND LYMPHOCYTE-MEDIATED CYTOLYSIS

extensively characterized. It is believed to prevent C9 polymerization by interacting on the host membrane with accidentally inserted CSb-8 complexes. The function is thus to prevent C9-mediated lysis. It has been shown by Haensch et al. (1981) that homologous erythrocytes are resistant to lysis by homologous but not to heterologous C9 (hence HW). Murine and human perforin, in contrast, can lyse homologous and heterologous erythrocytes (Lichtenheld et al., 1988; Jiang et al., 1988) and nucleated cells equally well. These findings indicate that perforin is not inhibited by HRF and disagree with reports by Zalman et al. (1987, 1988) suggesting self-protection of cytotoxic T lymphocytes (CTL) and of other target cells from perforin lysis by an HRF-like factor.

VI.

EXPRESSION OF PERFORIN mRNA in Vitro AND in Vivo

Analysis of perforin expression has been camed out on the mRNA level using perforin cDNA as specific hybridization probe. To date all lymphocytes expressing cytotoxic activity also express the message for perforin. In addition, the expression of perforin in mixed lymphocyte reactions, upon addition of interleukin 2 (IL-2) or after stimulation with concanavalin A (Con A), correlates with the expression of cytotoxicity generated by these regimens. Maximum expression of perforin precedes maximum expression of cytotoxicity by 12 hr. Table I lists the cell lines and clones expressing perforin mRNA. The expression of perforin under in vivo conditions has been controversial previously, and it was suggested that perforin expression is an artifact of in vitro culture of cells due to the high levels of lymphokines in the cell culture system (Berke and Rosen, 1987; Dennert et al., 1987). However, in two model systems in vivo induction of perforin expression could clearly be demonstrated. In the first system the expression of perforin in primary peritoneal exudate cells (PEL) was measured by Northern blot analysis and compared to the level of perforin mRNA in cloned CTL (Nagler-Anderson et al., 1990). Primary PEL were divided into CD8+ and CD8- cells by cell sorting. Only CD8+ cells contained cytotoxicity and perforin mRNA. In the second system, CTL were isolated from the liver of mice infected with a hepatotropic strain of lymphocytic choriomeningitis virus (LCMV) (Mueller et al. , 1990). The isolated lymphocytes had virusspecific, major histocompatibility complex (MHC)-restricted cytotoxic activity. They were predominantly CD8+ and contained perforin mRNA. In in situ hybridization experiments, perforin expression colocalized with the virus antigen and the infiltration of CD8 CTL that expressed both granzyme A and perforin.

-

+

ECKHARD R. PODACK AND MATHIAS G. LICHTENHELD TABLE I EXPRESSION OF PERFORINin Vitro A N D in Vivo Clone HY3-AG3 3F9 51 10-20 8/10-20 50.1 532 521 2c 21Cll

Cell type and specificities

% Perforin mRNAa

NK like CTL H-2Db-allospecific CTL AED-specific CTL H-2Kb-restricted CTL Three LCMV-specific, H-2-restricted CTL (various H-2 specificities) Allospecific CTL Class 11-restricted, ovalbumin-specific helper killer Primary MLC (BALB/c antikB6) Con A-stimulated splenocytes Con A-stimulated thymocytes Secondary MLC (anti-VSV splenocytes) Secondary MLC (anti-Vaccinia splenocytes) Primary in vivo CTL (anti-LCMV splenocytes) Primary in vivo CTL (anti-LCMV liver lymphocytes) Primary in vivo PEL (purified CD8 PEL CTL) +

EL4 P815 L929 MC57G K562

T-cell lymphoma Mastocytoma Fibrosarcoma

Methylcholanthrene-induced fibrosarcoma Erythroleukemia Liver Kidney Brain Pancreas Thymus Spleen CD8- (negatively selected) PEL

100 15 25 25 75 25 25 15 100

15 15 10 15 15 15 15 30 None None None None None None None None None None <5 None

“Relative to HY3-AG3.

VII. THE CONTRIBUTION OF MEMBRANE PORES TO DNA DEGRADATION Cytotoxic T lymphocyte-mediated lysis of target cells is accompanied or possibly even caused by the nuclear disintegration of the target cell and degradation and release of nuclear DNA in the form of integral multiples of 200-bp fragments (for reviews see Russel, 1983; Duke and Cohen, 1988). DNA degradation in susceptible cells is mediated by membrane pores (Hameed et al., 1989) and by

14. PERFORIN AND LYMPHOCYTE-MEDIATED CYTOLYSIS

171

factors such as lymphotoxin (Schmid and Ruddle, 1988), tumor necrosis factor or TNF (Kull, 1988), and NKCF (Bonavida and Wright, 1988). Factor-mediated DNA degradation is slow and requires 24-48 hr. Pore-mediated DNA degradation is rapid and detectable in 0.5-4 hr. The combination of pore formers with TNF results in the increase of DNA degradation within 4 hr, suggesting poreassisted uptake of TNF (Hameed et al., 1989). The molecular mechanism by which transmembrane pores cause nuclear disintegration is not known. The process is induced by Ca ionophores (Allbritton et al., 1988) and blocked by intracellular Ca chelators and strongly inhibited by lysosomotropic agents (Hameed et al., 1989). The mechanism by which pore formers induce DNA degradation may thus involve Ca influx through the transmembrane channel and activation of processes requiring acidic compartments. The enhancement of DNA degradation by TNF is interpreted as facilitated uptake through pore repair-endocytosis, engulfing TNF in endosomes and delivering it to the lysosomal compartment.

VIII.

CONCLUSIONS

Lymphocyte-mediated cytotoxicity by perforin apparently has three effects. The direct membrane damage by transmembrane channel formation can lead to target cell lysis by the effects of osmotic imbalance and the loss of the transmembrane potential. The influx of Ca ions through the transmembrane channel may lead in an unknown reaction pathway to the target cell suicide by nuclear disintegration and DNA degradation. Finally, the attempts of the target cell to repair the transmembrane pores by endocytosis of the perforated membrane can lead to the pinocytotic uptake of additional factors such as TNF that can also initiate target DNA degradation. The combined mechanism of membrane destruction and DNA degradation is an effective means to eliminate virus-infected and tumor targets by the immune system of the host. ACKNOWLEDGMENT This work was supported by grants from USPHS: AI-21999 and Ca-39201. REFERENCES Allbritton, N. L., Verret, C. R., Wolley, R. C . , and Eisen, H. N. (1988). Calcium ion concentrations and DNA fragmentation in target cell destruction by murine cloned cytotoxic T lymphocytes. J . Exp. Med. 167, 514. Berke, G . , and Rosen, D. (1987). Are lytic granules and perforin 1 thereof involved in lysis induced by in vivo primed, peritoneal exudate CTL? Trunsplant. Proc. 19, 412. Bleackley, R . C. (1988). The isolation and characterization of two cytolytic lymphocyte specific serine protease genes. Curr. Top. Microbiol. Immunol. 140, 6767. Bleackly, R . C . , Lobe, C. G . , Duggan, B . , Ehrman, N., Fregeau, C . , Meier, M., Letellier, M . ,

172

ECKHARD R. PODACK AND MATHIAS G. LICHTENHELD

Caliopi, H., Shaw, J., and Paetkau, V. (1988). The isolation and characterization of a family of serine protease genes expressed in activated cytotoxic lymphocytes. Immunol. Rev. 103, 5 . Blumenthal, R. P., Millard, P. J., Henkart, M. P., Reynolds, C. W., and Henkart, P. A. (1984). Liposomes as targets for granule cytolysin from cytotoxic large granular lymphocyte tumors. Proc. Natl. Acad. U.S.A. 81, 5551. Bonavida, B., and Wright, S . C. (1988). Natural killer cytotoxic factor (NKCF) as mediator in the lytic pathway of NK cell mediated cytotoxicity. In “Cytolytic Lymphocytes and Complement: Effectors of the Immune System” (E. R. Podack, ed.), Vol. 2, pp. 91-104. CRC Press, Boca Raton, Florida. Brunet, J.-F., Denizot, F., and Goldstein, P. (1988). A differential molecular biology search for genes preferentially expressed in functional T-lymphocytes: The CTLA genes. Immunol. Rev. 103, 21. Clark, W., Ostergaard, H., Gormann, K., and Torbett, B. (1988). Molecular mechanism of CTLmediated lysis: A cellular perspective. Immunol. Rev. 103, 37. Criado, M., Lindstrom, J. M . , Anderson, C. G . , and Dennert, G . (1985). Cytotoxic granules from killer cells: Specificity of granules and insertion of channels of defined sizes into target membranes. J . Immunol. 135, 4245. Dennert, G . , and Podack, E. R. (1983). Cytolysis by H2-specific killer cells. J . Exp. Med. 157, 1483. Dennert, G . , Anderson, C. G., and Psocharka, G . (1987). High activity of Na-benzyloxycarbonyl-Llysine thiobenzylester serine esterase and cytolytic perforin in cloned cell lines is not demonstrable in in vivo- induced cytotoxic effector cells. Proc. Natl. Acad. Sci. U.S.A. 84, 5004. DiScipio, R. G . , Gehring, M. R., Podack, E. R., Kan, C. C., Hugh, T. E., and Fey, G. H. (1984). Nucleotide sequence of cDNA and derived amino acid sequence of human complement component C9. Proc. Natl. Acad. Sci. U.S.A. 81, 7289. DiScipio, R. G . , Chakravarti, D. N., Miiller-Eberhard, H. J., and Fey, G . H. (1988). Structure of human C7 and the C5b-7 complex. J . Biol. Chem. 263, 549. Duke, R. C., and Cohen, 1. J. (1988). The role of nuclear damage in lysis of target cell by cytotoxic T lymphocytes. In “Cytolytic Lymphocytes and Complement: Effectors of the Immune System” (E. R. Podack, ed.), Vol. 2, pp. 35-64), CRC Press, Boca Raton, Florida. Duke, R. C., Chervenak, R., and Cohen, J. J. (1983). Endogenous endonuclease-induced DNA fragmentation: An early event in cell-mediated cytolysis. Proc. Natl. Acad. Sci. U.S.A. 80, 6361. Grosscurth, P., Qiao, B. Y . , Podack, E. R . , and Hengartner, H. (1987). Cellular localization of perforin 1 in mwine cloned cytotoxic lymphocytes. J . Immunol. 138, 2749. Haefliger, I. A., Tschopp, J., Nardelli, D., Wahli, W., Kocher, H. P., Tosi, M., and Stanley, K. K . (1987). Complementary DNA cloning of complement C8 beta and its sequence homology to C9. Biochemistry 26, 355 1. Hameed, A,, Olsen, K. J., Lichtenheld, M. G . , and Podack, E. R. (1989). Cytolysis by Capermeable transmembrane channels: Pore formation causes extensive DNA degradation and cell lysis. J . Exp. Med. 169, 765. Hansch, G. M. (1988). The homologous species restriction of the complement attack: Structure and function of the C8 binding protein. Curr. Top. Microbiol. Immunol. 140, 109. Hansch, G. M . , Hammer, C . , Vanguri, P., and Shin, M. L. (1981). Self versus nonself restriction in the lysis of erythrocytes by the terminal complement proteins. Proc. Narl. Acad. Sci. U . S . A . 78, 5118. Henkart, P. A. (1985). Mechanisms of lymphocyte mediated cytotoxicity. Annu. Rev. Immunol. 3, 31. Henkart, P. A , , Millard, P. J., Reynolds, C. W., and Henkart, M. P. (1984). Cytolytic activity of purified cytoplasmic granules from cytotoxic rat LGL tumors. J . Exp. Med. 160, 75.

14. PERFORIN AND LYMPHOCYTE-MEDIATED CYTOLYSIS

173

Hershberger, R. J., Mueller, C., Gershenfeld, H. K., and Weissman, I. L. (1988). A serine protease encoding gene that marks activated cytotoxic T-cells in vivo and in vitro. Curr. Top. Microbiol. Imrnunol. 140, 81. Howard, 0. M. Z., Rao, A. G., and Sodetz, J. M. (1987). Complementary DNA and derived amino acid sequence of the beta subunit of complement protein C8: Identification of a close structural and ancestral relationship to the alpha subunit and C9. Biochemistry 26, 3565. Jenne, D., and Tschopp, J. (1988a). Granzymes, a family of serine proteases released from granules of cytolytic T lymphocytes upon cell receptor stimulation. Immunol. Rev. 103, 53. Jenne, D. E., and Tschopp, J. (1988b). Granzymes: A family of serine esterases in granules of cytolytic lymphocytes. Curr. Top. Microbiol. Immunol. 140, 33. Jiang, S . , Persedini, P. M., Zychlinsky, A,, Liu, C. C., Perussia, B., and Young, J. D. E. (1988). Resistance of cytolytic lymphocytes to perforin mediated killing. J . Exp. Med. 168, 2207. Kull. F. C. (1988). Tumor necrosis factor/cachectin. In “Cytolytic Lymphocytes and Complement: Effectors of the Immune System” (E. R. Podack, ed.), Vol. 2, pp. 75-90. CRC Press, Boca Raton, Florida. Kwon, 9. S . , Wakulchik, M., Liu, C.-C., Peresechini, P. M., Trapani, J. A,, Haq, A. K., Kim, Y., and Young, J. D.-E. (1989). Biochem. Biophys. Res. Commun. 158, 1. Lichtenheld, M. G . , Lu, P., Olsen, K. J., Lowrey, D. M., Harneed, A,, Hengartner, H., and Podack, E. R. (1988). Structure and function of human perforin. Nature (LondonJ 335, 448. Lovell, D. M., and Martz, E. (1987). The degree of CTL induced DNA solubilization is not determined by the human vs. mouse origin of the target cell. J . Immunol. 138, 3695. Lowrey, D. M., Olsen, K. J., Lichtenheld, M. G., Aebischer, T., Rupp, F., Hengartner, H., and Podack, E. R. (1989). Cloning and analysis and expression of murine perforin 1 cDNA, a component of cytolytic T-cell granules with homology to complement component C9. Proc. Natl. Acad. Sci. U.S.A. 86, 247. McDermott, R. P., Schmidt, R. E., Caulfield, J. P., Hein, A., Banley, G. T., Ritz, J., Schlossman, S . F . , Austen, K. F., and Stevens, R. L. (1985). Proteoglycans in cell mediated cytotoxicity. Identification, localization and exocytosis of a chondroitin sulfate proteoglycan from human cloned natural killer cells during target cell lysis. J . Exp. Med. 162, 1771. Masson, D., and Tschopp, J. (1985). Isolation of a lytic, pore-forming protein (perforin) from cytolytic T-lymphocytes. J . Biol. Chem. 260, 9069. Masson, D., Corthesy, P., Nabholz, M., and Tschopp, J. (1985). Appearance of cytolytic granules upon induction of cytolytic activity in CTL-hybrids. EMBO J . 4, 2533. Medof, E. M. (1988). Decay accelerating factor and the defect of paroxysmal nocturnal hemoglobinuria. I n “Cytolytic Lymphocytes and Complement: Effectors of the Immune System” (E. R. Podack, ed.), Vol. 2, pp. 57-88. CRC Press, Boca Raton, Florida. Millard, P. J., Reynolds, C. W., and Henkart, M. P. (1984). Cytolytic activity of purified cytoplasmic granules from cytotoxic rat large granular lymphocyte tumors. J . Exp. Med. 160, 75. Mueller, C., Gershenfeld; H. K., and Weissman, I. L. (1988). Activation of CTL-specific genes during cell-mediated cytolysis in vivot Expression of HF gene analyzed by in situ hybridization. hrnunoi. Rev. 103, 73. Mueller, C., Kaegi, D., Aebischer, T., Odermatt, B., Podack, E. R., Zinkernagel, R. M., and Hengartner, H. (1989). Detection of perforin and granzyme AmRNAs in infiltrating cells during LCMV infection of mice. Eur. J . Immunol. 19, 1253. Miiller-Eberhard, H. J. (1988). The molecular basis of target cell killing by human lymphocytes and of killer self protection. Immunol. Rev. 103, 87. Munger, W. E., Berrebi, G. A , , and Henkart, P. A. (1988). Possible involvement of CTL granule proteases in target cell DNA breakdown. Immunol. Rev. 103, 99. Nagler-Anderson, C., Allbritton, N. L., Verret, C. R., and Eisen, H. N. (1988). A comparison of the

174

ECKHARD R. PODACK AND MATHIAS G. LICHTENHELD

cytolytic properties of CD8+ cytotoxic T lymphocytes and cloned cytotoxic T cell lines. Immunol. Rev. 103, 11 I . Nagler-Anderson, C., Lichtenheld, M., Eisen, H. N., and Podack, E. R. (1989). Perforin mRNA in primary peritoneal exudate cytotoxic T lymphocytes. J. Immunol. 143, 3440. Podack, E. R. (1984). Molecular composition of the tubular structure of the membrane attack complex of complement. J. Biol. Chem. 259, 8641. Podack, E. R. (1985). The molecular mechanism of lymphocyte mediated tumor lysis. Immunol. Today 6 , 12. Podack, E. R. (1986). Molecular mechanism of cytolysis by complement and by cytolytic lymphocytes. J. Cell Biochem. 30, 133. Podack, E. R. (1987). Perforins: A family of pore forming proteins in immune cytolysis. In “Membrane Mediated Cytotoxicity” (Collier and B. Bonavida, eds.), pp. 339-352. Alan R. Liss, New York. Podack, E. R. (1988). Granule mediated cytolysis of target cells. Curr. Top. Microbiol. Immunol. 140,l. Podack, E. R., and Dennert, G. (1983). Cell mediated cytolysis: Assembly of two types of tubules with putative cytolytic function by cloned natural killer cells. Nature (London) 302, 442. Podack, E. R., and Konigsberg, P. J. (1984). Cytolytic T-cell granules: Isolation, structural, biochemical and functional characterization. J . Exp. Med. 160, 695. Podack, E. R., and Tschopp, J. (1982). Polymerization of the ninth component of complement (C9): Formation of poly (C9) with a tubular ultrastructure resembling the membrane attack complex of complement. Proc. Natl. Acad. Sci. U.S.A. 79, 574. Podack, E. R., Tschopp, J., and Miiller-Eberhard, H. J. (1982). Molecular organization of C9 within the membrane attack complex of complement. J. Exp. Med. 156, 268. Podack, E. R., Young, J. D. E., and Cohn, Z. A. (1985). Isolation and biochemical and functional characterization of perforin 1 from cytolytic T-cell granules. Proc. Nutl. Acud. Sci. U.S.A. 82, 8629. Podack, E. R., Lowrey, D. M., Lichtenheld, M. G., Olsen, K. J., Aebischer, T., Binder, D., Rupp, F., and Hengartner, H. (1988a). Structure, function and expression of murine and human perforin 1 (Pl). Immunol. Rev. 103, 203. Podack, E. R., Olsen, K. J., Lowrey, D. M., and Lichtenheld, M. G. (1988b). Structure and function of perforin. Curr. Top. Microbiol. Immunol. 140, 11. Rao, A. G., Howard, 0. M. Z., Ng, S. C . , Whitehead, A. S., Colten, H. R., and Sodetz, J. M. (1987). Complementary DNA and derived amino acid sequence of the alpha subunit of human complement protein C8: Evidence for the existence of a separate alpha subunit mRNA. Biochemistv 26, 3556. Russel, J. H. (1983). Internal disintegration model of cytotoxic lymphocyte induced target damage. Immunol. Rev. 72, 97. Russel, J. H., Masakowski, V., Rucinsky, T., and Phillips, G.(1982). Mechanisms of immune lysis. 111. Characterization of the nature and kinetics of the cytotoxic T lymphocyte-induced nuclear lesion in the target. J. Immunol. ‘128, 2087. Schmid, D. S., and Ruddle, N. H. (1988). Production and function of lymphotoxin secreted by cytolytic T cells. In “Cytolytic Lymphocytes and Complement: Effectors of the Immune System” (E. R. Podack, ed.), Vol. 2, pp. 61-74. CRC Press, Boca Raton, Florida. Shinkai, Y.,Takio, K., and Okumura, K. (1988). Homology of perforin to the ninth component of complement (C9). Nature (London) 334, 525. Sitkovsky, M. V. (1988). Mechanistic, functional, and immunopharmacological implications of biochemical studies of antigen receptor triggered cytolytic T-lymphocyte activation. Immunol. Rev. 103, 127. Stanley, K. K. (1988). The molecular mechanism of complement component C9 insertion and polymerization in biological membranes. Curr. Top. Microbiol. Immunol. 140, 49.

14. PERFORIN AND LYMPHOCYTE-MEDIATED CYTOLYSIS

175

Stanley, K. K . , Kocher, M. P., Luzio, J. P., Jackson, P., and Tschopp, J. (1985). The sequence and topology of human complement components. EMBO J . 4, 375. Stevens, R. L., Karnada, M. N., and Serafin, W. E. (1988). Structure and function of the family of proteoglycans that reside in the secretory granules of natural killer cells and other effector cell of the immune response. Curr. Top. Microbiol. Immunol.140, 93. Thielens, N. M., Lohner, K., and Esser, A. (1988). Human complement protein C9 is a calcium binding protein. J. Biol. Chem. 263, 6665. Tschopp, J., Miiller-Eberhard, H. J., and Podack, E. R. (1982). Formation of transmembrane tubules by spontaneous polymerization of the hydrophilic complement protein C. Nature (London) 298, 534. Tschopp, J., Podack, E. R., and Miiller-Eberhard, H. J. (1985). The membrane attack complex of complement: C5b-8 complex as accelerator of C9 polymerization. Proc. Natl. Acad. Sci. U.S.A. 134, 495. Tschopp, J., Massan, D., and Stanley, K. K. (1986). Structurallfunctional similarity between proteins involved in complement and cytotoxic T-lymphocyte-mediated cytolysis. Nature (London) 322, 331. Tschopp, I.,Schafer, S., Masson, D., Peitsch, M. C., and Heusser, C. (1989). Phosphorylcholine acts as Ca2 -receptor molecule for lymphocyte perforin. Nature (London) 337, 272, Young, J. D. E., Hengartner, H., Podack, E. R., and Cohn, Z. A. (1986a). Purification and characterization of a cytolytic pore-forming protein from granules of cloned lymphocytes with natural killer cell activity. Cell 44,849. Young, 1. D. E., Cohn, Z. A , , and Podack, E. R. (1986b). The ninth component of cytotoxic T-cells: Structural and functional homologies. Science 233, 184. Young, J. D. E., Damiano, A . , DiNome, M. A . , Leong, L. G., and Cohn, 2. A. (1987). Dissociation of membrane binding and lytic activities of the lymphocyte pore-forming protein ( perforin). J. Exp. Med. 165, 1371. Young, 1. D. E., Liu, C.-C., Persechini, P. M., and Cohn. Z. A. (1988). Perforin-dependent and independent pathways of cytotoxicity mediated by lymphocytes. Immunol. Rev. 103, 161. Yue, C. C., Reynolds, C. W., and Henkart, P. A . (1987). Inhibition of cytolysin activity in large granular lymphocyte granules by lipids: Evidence for a membrane insertion mechanism of lysis. Mol. Immunol.24, 647. Zalman, L. S., Brothers, M. A., Chiu, F. J., and Miiller-Eberhard, H. J. (1986). Mechanism of cytotoxicity of human large granular lymphocytes: Relationship of the cytotoxic lymphocyte protein to C8 and C9 of human complement. Proc. Natl. Acad. Sci. U.S.A. 83, 562. Zalman, L. S., Wood, L. M., and Miiller-Eberhard, H. J. (1987). Inhibition of antibody dependent lymphocyte cytotoxicity by homologous restriction factor incorporated into target cell membranes. J . Exp. Med. 166, 947. Zalman, L. S., Brothers, M. A., and Miiller-Eberhard, H. J. (1988). Self protection of cytotoxic lymphocytes: A soluble form of homologous restriction factor in cytoplasmic granules. Proc. Natl. Acad. Sci. U.S.A. 85, 4827. NOTEADDED IN PROOF.Since submission of this article three murine perforin sequences have been published showing the following differences. The murine sequence published by Lowrey et al. (1989)differs in its C-terminal amino acids 523to 530 from the other two murine sequences (Shinkai et al., 1988; Kwon et al., 1989) and from the human sequence (Lichtenheld et al., 1988). The correct murine amino acid sequence 523 to 530 should probably read GDPPGNRS. Major differences are found in the 5’ untranslated sequences published by Lowrey et al. (1989) and Kwon et al. (1989). Our current analysis indicates that the 5’ untranslated sequence of Lowrey et al. (1989) 5‘ beyond nucleotide position 110 represents unspliced intron sequence. +

This Page Intentionally Left Blank

Index

A

ABP, see Actin-binding proteins Accessory factors, organelle movements and, 121, 123, 126 Acetylcholine receptor, 109, 110, 113 cell lines, 110, 111 expression, 111, 112 molecular associations and, 82 posttranslational modifications, 1 13 Actin, 97, 98 erythrocyte cytoskeleton, 98 n~olecularassociations and, 89 organelle movements and, 120, 121, 124 propulsion, 101- 106 stabilization, 99-101 Actin-binding proteins, 97, 99- 103 Adenylate cyclase, G protein-coupled receptors and, 138, 139 Adenylyl cyclase, G protein-coupled receptors and, 131, 132 ADP G protein-coupled receptors and, 140 mitochondria1 protein import and, 5-7, 9, 10

Adrenergic receptor, G protein-coupled receptors and, 132-134, 137-139 Amino acids bacteriorhodopsin and, 71, 76 enzymes and, 15 G protein-coupled receptors and, 134, 140, 141 mitochondrial protein import and, 4, 8 perforin and, 163, 167

photosynthetic reaction center and, 57, 6062,65, 66 T cell antigen receptor and, 33-35, 39 aPT cell antigen receptor and, 158 Ankyrin, actin and, 98 Antibodies acetylcholine receptor and, 110 mitochondrial protein import and, 8, 9 molecular associations and, 84, 86, 89 organelle movements and, 121, 122 T cell antigen receptor and, 34, 36-39, 4144 aPT cell antigen receptor and, 158, 159 yeast and, 24 Antigen acetylcholine receptor and, 112 molecular associations and, 8 1, 82, 88, 89 perforin and, 169 T cell receptor, see T cell antigen receptor Arachidonic acid, mechano-sensitive ion channels and, 146 Architectural editing, T cell antigen receptor and, 31-33,47,48 assembly, 40-42 composition, 33-36 stoichiometry, 39,40 subunit interactions, 36-39 synthesis, 43-47 ATP enzymes and, 17 mitochondrial protein import and, 4-10 organelle movements and, 120 yeast and, 2 1 , 2 3 ATPase enzymes and, 16, 17

177

178

INDEX

mltochondrial protein import and, 6, 7, 9, 10

molecular associations and, 92 yeast and, 23

B

B cells, T cell antigen receptor and, 33 Bacteria enzymes and, 15, 16 mechano-sensitive ion channels and, 145147, 151 mitochondrial protein import and, 10 photosynthetic reaction center and, 53-68 yeast and, 2 1, 22 Bacteriopheophytin, photosynthetic reaction center and, 56, 57, 59, 60,66 Bacteriorhodopsin, 71-73,77 G protein-coupled receptors and, 134, 137 packing effects, 76, 77 polar interactions, 75, 76 retinal, 74, 75 transbilayer helices, 73, 74 BFA, see Brefeldin A Brefeldin A, T cell antigen receptor and, 42, 46 a-Bungarotoxin, acetylcholine receptor and, 111,112

C

Chaperone proteins, enzymes and, 16, 17 Chloroplasts mitochondrial protein import and, 8 , 1 1 yeast and, 23 Chromophores, photosynthetic reaction center and, 54 Chromosomes T cell antigen receptor and, 35 yeast and, 24 Chymotrypsin, bacteriorhodopsin and, 73, 74 Clones acetylcholine receptor and, 109 G protein-coupled receptors and, 132 perforin and, 162, 169 T cell antigen receptor and, 33-36, 41 yeast and, 22 Complement protein, perforin and, 167, I68 Concanavalin A, perforin and, 169 Cyclic AMP, mechano-sensitive ion channels and, 149 Cysteine, perforin and, 167 Cystoskeleton, mechano-sensitive ion channels and, 149 Cytochalasins, molecular associations and, 86 Cytochromes mitochondrial protein import and, 6, 7, 9 photosynthetic reaction center and, 54, 58, 60,61 aPT cell antigen receptor and, 156, 159 Cytoplasm acetylcholine receptor and, 113 actin and, 98, 100, 101, 104 G protein-coupled receptors and, 134, 136141

Calcium actin and, 102-106 mechano-sensitive ion channels and, 149 perforin and, 162, 163, 171 Carbohydrate acetylcholine receptor and, 113 T cell antigen receptor and, 40, 41, 44, 46 Carboxypeptidase Y, yeast and, 24, 25 cDNA acetylcholine receptor and, 110 perforin and, 162, 163, 167 T cell antigen receptor and, 36, 39, 45 Centrosomes, organelle movements and, 118, 124 Cephalexin, mechano-sensitive ion channels and, 146

mitochondrial protein import and, 11 molecular associations and, 86, 88, 92 organelle movements and, 117, 118, 121124,126 perforin and, 162 photosynthetic reaction center and, 57, 6163,65,67 T cell antigen receptor and, 34 yeast and, 23, 25 Cytoskeleton actin and, 97-99 molecular associations and, 83, 86-90, 92 organelle movements and, 119, 123, 124, 126 Cytosol enzymes and, 16 mitochondrial protein import and, 3, 4, 6, 8

INDEX

179

yeast and, 20-22, 24, 25 Cytotoxic T lymphocytes, perforin and, 169, 170

D

Decay-accelerating factor perforin and, 168 aPT cell antigen receptor and, 158 Dialysis, enzymes and, 16 Dictyosfelium discoideum, actin and, 100 DNA, perforin and, 161, 170, 171 Dynein, organelle movements and, 117, 118, 121-124

protein secretion and, 15-17 a P T cell antigen receptor and, 158 yeast and, 24 Epidermal growth factor, perforin and, 167 Epithelial cells, molecular associations and, 83,91,92 Epitopes, actin and, 98 Erythrocytes actin and, 98- 100 molecular associations and, 84, 86-88 perforin and, 169 Escherichia coli enzymes and, 16 mechano-sensitive ion channels and, 146, 147, 149-152 mitochondrial protein import and, 11 Extracellular matrix, molecular associations and. 82

E

EDTA mechano-sensitive ion channels and, 146 perforin and, 163 EGTA actin and, 102, 104 perforin and, 163 Electron microscopy G protein-coupled receptors and, 134 perforin and, 162 T cell antigen receptor and, 44 Endocytosis mitochondrial protein import and, 10 organelle movements and, 119 perforin and, 171 Endoglycosidase H acetylcholine receptor and, 113 T cell antigen receptor and, 44 Endoplasmic reticulum acetylcholine receptor and, 113 mitochondrial protein import and, 11 organelle movements and, 117, 118, 123126 T cell antigen receptor and, 40-42, 44-48 yeast and, 19-22, 25 Endosomes, perforin and, 171 Endothelial cells, mechano-sensitive ion channels and, 146 Enzymes G protein-coupled receptors and, 132 organelle movements and, 121

F

Fibroblasts acetylcholine receptor and, 109- 113 molecular associations and, 82, 87, 90 organelle movements and, 119, 121, 123 T cell antigen receptor and, 39, 46, 47 Fibronectin, actin and, 99, 100 Fluid mosaic model, molecular associations and, 81 Fluorescence bacteriorhodopsin and, 74 organelle movements and, 124 T cell antigen receptor and, 44, 46 aPT cell antigen receptor and, 158 Fluorescence microscopy, acetylcholine receptor and, 11 1 Fluorescene photobleaching and recovery, molecular associations and, 83-86, 88-90, 92,93 Forskolin, acetylcholine receptor and, 1 13 FF'R, see Fluorescene photobleaching and recovery Fractionation T cell antigen receptor and, 43-46 yeast and, 20 FRAP, see Fluorescene photobleaching and recovery

180

INDEX G

G protein mechano-sensitive ion channels and, 146, 151 molecular associations and, 87 G protein-coupled receptors, 131-133 receptor-mimetic peptides, 140, 141 structure, 134-136 G protein-binding domain, 138-140 ligand-binding domain, 136- 138 GDP, G protein-coupled receptors and, 131133, 140 Gelsolin, actin and, 102-104 General insertion protein, mitochondrial protein import and, 8- 1 1 GIP, see General insertion protein Glycolipid, perforin and, 168 Glycoprotein actin and, 100 G protein-coupled receptors and, 132 molecular associations and, 87 T cell antigen receptor and, 32, 35, 44, 45 yeast and, 24 Glycoprotein Ib, actin and, 100 Glycosylation G protein-coupled receptors and, 134 molecular associations and, 87 perforin and, 163 T cell antigen receptor and, 34, 35 Golgi apparatus organelle movements and, 118, 124, 126 T cell antigen receptor and, 40-42, 44-47 yeast and, 20, 25 Granzymes, perforin and, 162, 169

mechano-sensitive ion channels and, 150 molecular associations and, 81, 82 Hyaluronate, actin and, 100 Hybridomas, T cell antigen receptor and, 38, 40,43-46 Hydrogen, bacteriorhodopsin and, 72, 7 5 . 76

1

Immunoglobulins actin and, 100 mitochondrial protein import and, 6, 7 T cell antigen receptor and, 33-35, 41 apT cell antigen receptor and, 158 Inhibition mitochondrial protein import and, 5-7. 9. 10 molecular associations and, 87 organelle movements and, 119, 121 perforin and, 169, 171 T cell antigen receptor and, 42, 44-46 yeast and, 23, 24 Inositol, molecular associations and, 85 Insulin molecular associations and, 89 T cell antigen receptor and, 3 I Integrin, actin and, 100 Interleukin-2 molecular associations and, 82, 89 perforin and, 169 Intermediate filaments actin and, 97 organelle movements and, 124 Ion channels, mechano-sensitive, 145- 152

K

H

Haptens, molecular associations and, 84 Herpes simplex virus, aPT cell antigen receptor and, 158 Heterodimers, aPT cell antigen receptor and, 158, 159 Homologous restriction factor, perforin and, 168, 169 Hormones G protein-coupled receptors and, 131- 133

Kidney, molecular associations and, 92 Kinesis, organelle movements and, 117-1 19, 121- 124

1

Lectins actin and, 100

181

INDEX molecular associations and, 84 Leukocytes, actin and, 101 Ligands acetylcholine receptor and, 109 actin and, 97, 99, 100 G protein-coupled receptors and, 131- 134, 136-138, 140 mechano-sensitive ion channels and, 145 aPT cell antigen receptor and, 155 Lipid bacteriorhodopsin and, 71-74, 76, 77 enzymes and, 17 G protein-coupled receptors and, 134 mitochondria1 protein import and, 4 , 5 molecular associations and, 81-83, 85, 8890,92,93 organelle movements and, 119 perforin and, 163 photosynthetic reaction center and, 61, 63 aPT cell antigen receptor and, 156, 158, 159 yeast and, 24 Liposomes enzymes and, 17 mechano-sensitive ion channels and, 147 Liver, molecular associations and, 91 Lymphoblasts, molecular associations and, 89 Lymphocyte-mediated cytolysis, perforin and, 161, 162, 171 DNA degradation, 170, 171 lack of homologous restriction, 168, 169 mRNA expression, 169, 170 properties, 162, 163 sequence, 163- 167 Lymphocytes, molecular associations and, 88 Lymphocytic choriomeningitis virus, perforin and, 169 Lymphokines, perforin and, 169 Lymphotoxin, perforin and, 171 Lysine, G protein-coupled receptors and, 137 Lysosomes, perforin and, 171 Lysozymes, mechano-sensitive ion channels and, 146, 147

M

Macrophages, actin and, I04 Magnesium enzymes and, 17

G protein-coupled receptors and, 140 Major histocompatibility complex molecular associations and, 88, 89 perforin and, 169 (rPT cell antigen receptor and, 155-159 Mastopmans, G protein-coupled receptors and, 140, 141 MDCK cells molecular associations and, 92 organelle movements and, 119 Mechano-sensitive ion channels, 145, 146 Escherichia coli, 146, 147 Paramecium, 149 solvent senses, 150- 152 yeast, 147-149 Microsomes, yeast and, 20-24 Microtubule-associated proteins, organelle movements and, 117, 126 Microtubules actin and, 97 organelle movements and, 117-126 Mitochondria enzymes and, 15, 16 mechano-sensitive ion channels and, 151 yeast and, 21,22 Mitochondria1 protein import, 3-5, 11 general inserion protein, 8- 1 1 receptors, 5-8 Mitosis, organelle movements and, 126 Molecular associations, 93 membrane organization, 8 1-83 mobility, 83 erythrocytes, 86-88 fluorescence photobleaching, 83-86 mobile fraction, 88-90 morphologically polarized cells, 90-92 Molecular chaperones, yeast and, 22 Monoclonal antibodies, molecular assoclations and, 89 Morphology, molecular associations and, 83, 90-92 mRNA perforin and, 167, 169, 170 photosynthetic reaction center and, 62 Mutagenesis acetylcholine receptor and, 109 bacteriorhodopsin and, 76 G protein-coupled receptors and, 138 Mutation enzymes and, 15 G protein-coupled receptors and, I40

182

INDEX

molecular associations and, 86-88 photosynthetic reaction center and, 59, 60 apT cell antigen receptor and, 156 yeast and, 24, 25 Myosin actin and, 99, 101, 106 organelle movements and, 120, 121

N Natural killer cells, perforin and, 162 Neomycin, acetylcholine receptor and, 110 Neurospora, yeast and, 22 Neurosporu crussa, mitochondria1 protein import and, 6, 10 Nucleotides G protein-coupled receptors and, 132, 133, 140 perforin and, 165

0

Oligosaccharides acetylcholine receptor and, 113 molecular associations and, 87 Organelle movements, 1 17, 118 microtubules, 118- I2 1 motility, 121, 122 ER extension, 123-126 translocation, 123

P Packing effects, bacteriorhodopsin and, 76, 77 Papain, yeast and, 23 Paramecium, mechano-sensitive ion channels and, 149, 150 Peptidoglycan, mechano-sensitive ion channels and, 147 Perforin, 161, 162, 171 DNA degradation, 170, 171 homology to complement proteins, 167, 168 lack of homologous restriction, 168, 169 mRNA expression, 169, 170

properties, 162, 163 sequence, 163-167 Peritoneal exudate cells, perforin and, 169 Pertussis toxin, G protein-coupled receptors and, 139, 140 Phagocytosis, actin and, 99, 101 Phenotype aPT cell antigen receptor and, 156, 157 yeast and, 24, 25 Pheromones, mechano-sensitive ion channels and, 150 Phorbol esters, molecular associations and, 88 Phosphoinosi tides actin and, 102-104 aPT cell antigen receptor and, 158 Phospholipases, G protein-coupled receptors and, 131, 132, 139 Phospholipids, G protein-coupled receptors and, 133, 139, 140 Phosphoribosyl anthranilate isomerase, yeast and, 24 Phosphorylation acetylcholine receptor and, 113 G protein-coupled receptors and, 136 Photosynthesis, bacteriorhodopsin and, 7 1 Photosynthetic reaction center, 53, 54, 68 architecture, 56-58 cytochrome subunit, 60,61 distribution, 63-68 pigment arrangement, 54-56 pigment branch, 58-60 protein structure, 61-63 Plasma membrane acetylcholine receptor and, 11 1 actin and, 97-100, 103 bacteriorhodopsin and, 73 G protein-coupled receptors and, 131 mechano-sensitive ion channels and, 148, 149 Plasmalemma, actin and, 98, 104 Platelets, actin and, 100, 104 PMN, see Polymorphonuclear lymphocytes Polarity, organelle movements and, 119, 120 Polymers actin and, 97, 100-104 perforin and, 162, 163, 169 Polymorphonuclear lymphocytes, actin and, 97 Polypeptides acetylcholine receptor and, 11 1 bacteriorhodopsin and, 71, 74

183

INDEX mitochondrial protein import and, 4 photosynthetic reaction center and, 57 aPT cell antigen receptor and, 158, 159 Polyperforin, 163 Ponticulin, actin and, 100 Porin mechano-sensitive ion channels and, 151 mitochondrial protein import and, 5-7, 9, 11 Posttranslational modification acetylcholine receptor and, 110, 112, 113 organelle movements and, 119 Potassium bacteriorhodopsin and, 73 mechano-sensitive ion channels and, 149151

Prepro-a-factor, yeast and, 20, 22 Preprotein enzymes and, 17 yeast and, 21, 22, 24, 25 Profilin, actin and, 102, 104 Proteases bacteriorhodopsin and, 75 mitochondrial protein import and, 5, 7-9 yeast and, 21, 22, 24 Protein actin-membrane interactions and, 97106 bacteriorhodopsin folding and, 7 1-77 enzymes and, 15-17 mechano-sensitive ion channels and, 147, 150, 151 mitochondrial, see Mitochondria1 protein imPo* molecular associations and, 81-93 organelle movements and, 117-126 perforin and, 162, 163, 165, 167, 168 photosynthetic reaction center and, 53-68 signal-transducing, 131- 141 T cell antigen receptor and, see T cell antigen receptor aPT cell antigen receptor and, 156, 158 translocation in yeast and, 19-25 Protein kinase, acetylcholine receptor and, 113 Proteoglycans, perforin and, 162 Proteolysis G protein-coupled receptors and, 134, 137, 138 mitochondrial protein import and, 4, 7

R Retinal bacteriorhodopsin and, 73-75,77 G protein-coupled receptors and, 132, 137 Retrovirus, acetylcholine receptor and, 110 Rhodamine, acetylcholine receptor and, 111 Rhodobucter sphaeroides, photosynthetic reaction center and, 54 Rhodopseudomoms viridis, photosynthetic reaction center and, 53-68 Rhodopsin G protein-coupled receptors and, 132-134, 136, 137, 139 molecular associations and, 83-85, 87 Ribosomes mitochondrial protein import and, 4 perforin and, 165 yeast and, 20

S

Sucrharomyces rerevisiae, mechano-sensitive ion channels and, 146, 147 Serine, G protein-coupled receptors and, 134, 136 Signal recognition particle, yeast and, 19, 22 Signal sequence receptor, yeast and, 19, 23 Signal-transducingproteins, 131- 141 Soluble heterodirners, aPT cell antigen receptor and, 158, 159 Solvent senses, mechano-sensitive ion channels and, 150-152 Spectrin actin and, 98 molecular associations and, 86, 92 organelle movements and, 120 SRP, see Signal recognition particle Synapsin, organelle movements and, 123, 126

T

T cell antigen receptor, 31-33, 47, 48 assembly, 4 C 4 2 composition, 33-36

184

INDEX

stoichiometry, 39, 40 subunit interactions, 36-39 synthesis, 43-47 a p T cell antigen receptor, 155, 156 negative selection, 157, 158 positive selection, 156, 157 soluble heterodimers, 158, 159 T cells molecular associations and, 82, 88 perforin and, 162, 168 Threonine, G protein-coupled receptors and, 134, 136 Thrombin, actin and, 104 Thymidine kinase, acetylcholine receptor and, 110 Thymus, a p T cell antigen receptor and, 155I57 Torpedo, acetylcholine receptor and, 110-1 13 Transcription, perforin and, 167 Transducin, G protein-coupled receptors and, 132, 133, 139 Translocation enzymes and, 15-17 mitochondria1 protein import and, 4, 8, 9,

Vesicles bacteriorhodopsin and, 74 enzymes and, 17 G protein-coupled receptors and, 133, 139, 140 organelle movements and, 117, 119, 123, 126 yeast and, 20, 21 Vesicular stomatitis virus, molecular associations and, 87 Vinculin, actin and, 100 VSV, see Vesicular stomatitis virus

X Xenopus, acetylcholine receptor and, 109

Y

11

organelle movements and, 117, 122, 123 photosynthetic reaction center and, 62, 63 Tryptophan photosynthetic reaction center and, 60, 66, 67 yeast and, 24 Tumor necrosis factor, perforin and, 171 Tunicamycin, acetylcholine receptor and, 113

U

Yeast enzymes and, 15, 16 G protein-coupled receptors and, 132 mechano-sensitive ion channels and, 147150,152 protein translocation and, 19, 20 binding, 22-24 model system, 20,21 mutation, 24, 25 preprotein conformation, 21, 22

Unfoldases, yeast and, 21

V Vacuoles, yeast and, 21

Zymolyase, mechano-sensitive ion channels and, 147