Current Topics in Membranes, Volume 38
Ordering the MembraneCytoskeleton Trilayer
Current Topics in Membranes, Volum...
22 downloads
465 Views
14MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Current Topics in Membranes, Volume 38
Ordering the MembraneCytoskeleton Trilayer
Current Topics in Membranes, Volume 38 Yale Series Editors
Joseph F. Hoffman and Gerhard Giebisch Department of Cellular and Molecular Physiology Yale University School of Medicine New Haven, Connecticirt
Murdoch Ritchie Department of Pharmacology Yale University School of Medicine New Haven, Connecticut
Series Editors
Amost Kleinzeller Department of Physiology University of Pennsylvania School of Medicine Philadelphia, Pennsylvania
Douglas M. Fambrough Department of Biology Johns Hopkins University Baltimore, Maryland
Current Topics in Membranes, Volume 38
Ordering the MembraneCytoskeleton Trilayer Guest Editors Mark S. Mooseker
Jon 5. Morrow
Department of Biology Yale University New Haven, Connecticut
Department of Pathology Yale University School of Medicine New Haven, Connecticut
Volume 38 is part of the series from the Yale Department of Cellular and Molecular Physiology.
ACADEMIC PRESS, INC. Harcourt Brace Jovanovich, Publishers San Diego New York Boston London Sydney Tokyo Toronto
This book is printed on acid-free paper. @
Copyright 0 1991 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 NWl 7DX
Library of Congress Catalog Card Number: 70-1 17091
ISBN 0-12-153338-7
(alk. paper)
PRINTED IN THE UNITED STATES OF AMERICA
91 92 93 94
9 8 7 6 5 4 3 2
1
Contents Contributors xi Foreword xv Preface xvii Previous Volumes in Series xix
CHAPTER 1 Molecular Genetic Analyses of Drosophila Kinesin Russell J . Stewart and Lawrence S . B . Goldstein 1. 11. 111. IV.
Introduction 1 Kinesin Structure 2 Kinesin Function 6 Conclusion 9 References 9
CHAPTER 2 Acanthamoeba Myosin I: Past, Present, and Future Edward D . Korn
I. Discovery
13
15 III. Heavy Chain Sequences 16 IV. Structural Basis of the Actin-Activated ATPase Activity 17 V. Mechanochemical Properties and Their Structural Basis 19 VI. Regulation 21 VII. Intracellular Localization 22 VIII. Concluding Remarks 25 References 27 11. Multiplicity of Isoforms
CHAPTER 3 Structural and Functional Dissection of a MembraneBound Mechanoenzyme: Brush Border Myosin I Mark S. Mooseker, Joseph S . Wolenski, Thomas R . Coleman, Steven M . Hayden, Richard E . Cheney, Enilza Espreajco, Matthew B . Heintzelman, and Michelle D . Peterson
V
vi
Contents I. Introduction: Is There a Superfamily of Myosin Genes? 32 11. The Structural and Functional Properties of Brush Border
Myosin I: An Overview 33 111. Probing the Function of Brush Border Myosin I Calmodulin
Light Chains 45 IV. The Interaction of Brush Border Myosin I with the Microvillar Membrane 48 V. Some Notions Regarding Brush Border Myosin I Function 51 References 53
CHAPTER 4 Protein Interactions Linking Actin t o the Plasma
Membrane in Focal Adhesions Keiko 0 . Simon, Carol A . Otey, Fredrick M. Pavalko, and Keith Burridge I. Introduction 57 11. Interaction between a-Actinin and PIIntegnn 111. Interaction between Talin and Actin 61
59
1V. Conclusion 61 References 63
CHAPTER 5 Ankyrins: A Family of Proteins That Link Diverse
Membrane Proteins To the Spectrin Skeleton Vann Bennett, Ed Otto, Jonathan Davis, Lydia Davis, and Ekaterini Kordeli
I. Introduction 65 11. Ankyrin Structure 66 111. Ankyrins Are a Multigene Family 68 IV. Functional Diversity of Ankyrin due to Alternative Splicing of
mRNA 71 V. Mapping the Binding Sites of Ankyrin 72 VI. Summary and Future Perspectives 75 References 76
CHAPTER 6 Contractile and Cytoskeletal Proteins in Drosophila
Em btyogenesis Daniel P . Kiehart 1. Introduction 79 11. Movements of Early Embryogenesis
80
vii
Contents 111, Cell Shape Change Requires Remodeling of the Actin
Cytoskeleton 83 IV. Non-Muscle Myosins 84 V. Spectrins 89 VI. Summary 92 References 93
CHAPTER 7 Dominant Mutations of Cytoskeletal Proteins in
Xenopus Embryos Jan L. Christian, Gregory M . Kelly, and Randall T. Moon
I. Introduction 99 11. The Role of Membrane Skeleton Protein 4.1 during
Embryogenesis
101
111. Dissecting the Functions of Vimentin during Embryonic Development 102 IV. Discussion 105
References
107
CHAPTER 8 The Animal Models of Duchenne Muscular
Dystrophy: Windows on the Pathophysiological Consequences of Dystrophin Deficiency Eric P . HofSman and Jose Rafael M . Gorospe
I. Introduction
113
11. The Animal Models of Duchenne Dystrophy: Biochemical and Genetic Homology 1 18 111. Human Duchenne Dystrophy: Progressive Histopathology
Leading to Muscle Wasting 121 1V. The mdr Mouse: Nonprogressive Histopathology Resulting in Hypertrophy 124 V. The xmd Dog: Rapidly Progressive Histopathology Leading to Muscle Wasting 125 VI. The Dystrophin-Deficient Cat: Semiprogressive Histopathology with No Loss of Muscle Fibers 126 VII. Phase I. The Primary Cellular Consequence of Dystrophin Deficiency: Generalized Leakage of the Plasma Membrane Buffered by the Syncytial Cytoplasm? 127 VIII. Phase I Conclusion: An Integrated Model Featuring Ca2+ 130 IX. Phase 11: Progressive Pathology and Clinical Weakness Specific to Humans and Dogs 134 X. The Development of Progressive Histopathology: Could Basic Fibroblast Growth Factor Have a Major Role? 143
viii
Contents
XI. Conclusion: An Integrated Model and Its Consequences on the Development of Therapeutics 144 References 148
CHAPTER 9 Mutant Cytoskeletal Proteins in Hemolytic Disease Sully L. Marchesi
I. Introduction: Membrane/Skeleton Symposium
155
11. Detection and Characterization of Spectrin Mutants in
Hereditary Elliptocytosis
157
111. Mutant Forms of Protein 4.1 in Hereditary Elliptocytosis IV. Hereditary Spherocytosis 169 References 17 1
167
CHAPTER 10 Dynamics of Intestinal Epithelial Tight Junctions James L. Madurn
I . Introduction
175
11. Tight Junctions (Zonula Occludens) as Potentially Regulated
Barriers
176
111. Intestinal Zonula Occludens as a Regulated Transport
Pathway 179 IV. Intestinal Zonula Occludens Function in Model Disease States 180 V. Conclusions 181 References 182
CHAPTER 1 1 Regulation of Actin a n d Myosin II Dynamics in
Living Cells John Kolrgu und D . Lansing Taylor
I. Introduction 187 It. Quiescent Fibroblasts as a Model System 188 111. The Multimode Approach 190 IV. Actomyosin Dynamics in Serum-Starved 3T3 Fibroblasts 192 V. Regulation of Stress Fiber Contraction 195 VI. Stress Fiber Movement and Cell Motility 198 VII. Some Future Prospects 199 References 202
ix
Contents
CHAPTER 12 Expression and Function of Genetically Engineered Actin-Binding Proteins in Dictyosteljum Walter Witke, Michael Schleicher, Helmut Einherger, Wolfgang F . Neuhert, and Angelika A . Noegel
I . The Microfilament System of Dictyostelium discoideum
208
11. F-Actin Cross-Linking Proteins in Dictyostelium
discoideum
209
111. Genetic Manipulation of a-Actinin in Escherichia coli and
Dictyostelium discoideum References 2 I4
CHAPTER 13
2 10
Interaction of Profilins with Membrane Lipids Thomas D . Pollard, Laura MacheJb,, and Pascal Goldschmidt-Clermont
I . Binding of Profilin to Phospholipids 217 11. Profilin Inhibits Soluble Phospholipase C 219 111. Interaction of Profilin with Actin 221 IV. What Does Profilin Actually Do in a Cell? 222
References
224
CHAPTER 14 Polarized Assembly of Spectrin and Ankyrin in Epithelial Cells Jon S . Morrow, Carol D . Cianci, Scott P . Kennedy, and Stephen L. Warren 1. Introduction: The Red Cell Paradigm 227 11. The Unexpected Spatial Polarization of the Nonerythroid
Spectrin Cytoskeleton
229
111. The Interaction of Spectrin and Ankyrin with Basolateral
Na+/K+-ATPase 231 IV. Does Polarized Membrane Skeletal Assembly Require Ankyrin? 233 V. Models for the Role of the Nonerythroid Spectrin-Actin Cytoskeleton 237 References 242
Index 245
This Page Intentionally Left Blank
Contributors Numbers in parentheses indicate the pages on which the authors’ contributions begin.
Vann Bennett (65), Department of Biochemistry and the Howard Hughes Medical Institute, Duke University Medical Center, Durham, North Carolina 277 10 Keith Burridge (57), Department of Cell Biology and Anatomy, The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599 Richard E. Cheney (31), Department of Biology, Yale University, New Haven, Connecticut 065 10 Jan L. Christian (99), Department of Pharmacology, University of Washington, School of Medicine, Seattle, Washington 98 195 Carol D. Cianci (227), Department of Pathology, Yale University School of Medicine, New Haven, Connecticut 065 10 Thomas R. Coleman (3 l), Department of Biology, Yale University, New Haven Connecticut 065 10 Jonathan Davis (65), Department of Biochemistry and the Howard Hughes Medical Institute, Duke University Medical Center, Durham, North Carolina 27710 Lydia Davis (65), Department of Biochemistry and the Howard Hughes Medical Institute, Duke University Medical Center, Durham, North Carolina 277 10 Helmut Einberger (207), Max-Planck-Institute for Biochemistry, D-8033 Martinsried, Germany Enilza Espreafico (3 l), Department of Biology, Yale University, New Haven, Connecticut 06510 and Departmento de Biochiumica, de Faculdade de Medicina de Ribeirao Preto, Universidade de Siio Paulo, 05508 Sio Paulo, Brazil. Pascal Goldschmidt-Clermont (2 17), Department of Cell Biology and Anatomy, The Johns Hopkins Medical School, Baltimore, Maryland 21205 Lawrence S. B. Goldstein (I), Department of Cell and Developmental Biology, Harvard University, Cambridge, Massachusetts 02 138 Jose Rafael M. Gorospe ( 1 13), Departments of Molecular Genetics and Bioxi
xii
Contributors
chemistry, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania 15261
Steven M. Hayden (31), Department of Biology, Yale University, New Haven, Connecticut 065 10 Matthew B. Heintzelman (3 1). Department of Biology, Yale University, New Haven, Connecticut 06510 Eric P. Hoffman ( I 13), Departments of Molecular Genetics and Biochemistry, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania 1526 1 Gregory M. Kelly (99), Department of Pharmacology, University of Washington, School of Medicine, Seattle, Washington 98 195 Scott P. Kennedy (227), Department of Pathology, Yale University School of Medicine, New Haven, Connecticut 065 10 Daniel P. Kiehart (79), Department of Cellular and Developmental Biology, Harvard University, The Biological Laboratories, Cambridge, Massachusetts 02138 John Kolega (1 87), Department of Biologial Sciences and Center for Fluorescence Research in Biomedical Sciences, Carnegie Mellon University, Pittsburgh, Pennsylvania 15213 Ekaterini Kordeli (65), Department of Biochemistry and the Howard Hughes Medical Institute, Duke University Medical Center, Durham, North Carolina 27710 Edward D. Korn (1 3), Laboratory of Cell Biology, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland 20892 Laura Machesky (217), Department of Cell Biology and Anatomy, The Johns Hopkins Medical School, Baltimore, Maryland 21205 James L. Madara (173, Department of Pathology, Brigham and Women’s Hospital, and Harvard Medical School, and the Harvard Digestive Disease Center, Boston, Massachusetts 021 15 Sally L. Marchesi (155), Department of Pathology and Laboratory Medicine, Yale Medical School, New Haven, Connecticut 065 10 Randall T. Moon (99), Department of Pharmacology, University of Washington, School of Medicine, Seattle, Washington 98 195 Mark S. Mooseker (3 l), Department of Biology, Yale University, New Haven, Connecticut 065 10
Contributors
xiii
Jon S. Morrow (227), Department of Pathology, Yale University School of Medicine, New Haven, Connecticut 065 10 Wolfgang F. Neubert (207), Max-Planck-Institute for Biochemistry, D-8033 Martinsried, Germany Angelika A. Noegel (207), Max-Planck-Institute for Biochemistry, D-8033 Martinsried, Germany Carol A. Otey (57), Department of Cell Biology and Anatomy, The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599 Ed Otto (65), Department of Biochemistry and the Howard Hughes Medical Institute, Duke University Medical Center, Durham, North Carolina 277 10 Fredrick M. Pavalko (57), Department of Cell Biology and Anatomy, The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599 Michelle D. Peterson (31), Department of Biology, Yale University, New Haven, Connecticut 06510 Thomas D. Pollard (217), Department of Cell Biology and Anatomy, The Johns Hopkins Medical School, Baltimore, Maryland 21 205 Michael Schleicher (207), Max-Planck-Institute for Biochemistry, D-8033 Martinsried, Germany Keiko 0. Simon (57), Department of Cell Biology and Anatomy, The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599 Russell J. Stewart (l), Department of Cell and Developmental Biology, Harvard
University, Cambridge, Massachusetts 02 138
D. Lansing Taylor (1 87), Department of Biologial Sciences and Center for Fluorescence Research in Biomedical Sciences, Carnegie Mellon University, Pittsburgh, Pennsylvania 152 13
Stephen L. Warren (227), Department of Pathology, Yale University School of Medicine, New Haven, Connecticut 065 10 Walter Witke (207), Max-Planck-Institute for Biochemistry, D-8033 Martinsried, Germany Joseph S. Wolenski (3 l), Department of Biology, Yale University, New Haven, Connecticut 065 10
This Page Intentionally Left Blank
The Yale Series of Conferences has been published as an integral part of Current Topics in Membranes and Transport since 1980. Before that time, Felix Bronner and Arnost Kleinzeller, who began the series, were the sole editors. In 1980, Joseph Hoffman and Gerhard Giebisch joined as the editors of the Yale Series, to be published as part of a joint enterprise. Since that time, the senior editors have changed such that Douglas Fambrough now co-edits with Arnost Kleinzeller, a yearly volume separate from the Yale Series. In addition, Murdoch Ritchie has been added as a member of the Yale Series editors. Coincident with the change of the senior editors was the recognition by them that the field of membrane transport has become too limited to encompass adequately the burgeoning field of membrane biology. We believe that the title should reflect this change in emphasis and, therefore, it was decided that the new title for the series should be Current Topics in Membranes. We anticipate that the coverage of future volumes will naturally mirror the expanding scope of membrane biology. JOSEPH F. HOFFMAN GERHARD GIEBISCH J. MURDOCH RITCHIE
xv
This Page Intentionally Left Blank
In the beginning there was a lipid membrane bilayer and three major cytoskeletal proteins: actin, tubulin, and intermediate filaments. The membrane was fluid, and the cytoskeleton static. Then membrane proteins were discovered that were not so fluid, and additional cytoskeletal proteins emerged that could rapidly and reversibly change the state of the major filament systems (such as the first actin binding proteins discovered, projilin, or macrophage actin binding protein). By the mid- 1970s detergent extracts of erythrocytes and other cells revealed the presence of an elaborate submembraneous protein scaffolding. To this was given the name membrane cytoskeleton. Other studies established clear molecular linkages between integral membrane proteins and components of this membrane cytoskeleton. Later there emerged evidence that such interactions bestowed global and local order in membrane. By the mid-1980s the generality of this concept was well established. All cells possessed some form of membrane cytoskeleton, and many different classes of membrane-associated cytoskeletons were identified. In addition, the association of cytoskeletal proteins with the membrane was revealed to be dynamic and subject to many levels of posttranslational control, ranging from fatty acylation to phosphorylation to proteolysis. Later, cytoskeletal proteins were appreciated as key constituents in processes regulating cell shape, stability, membrane receptor organization, vesicular transport, and membrane trafficking. Thus, by 1990, a more accurate concept of the plasmalemma was that of a structural and functional membranecytoskeleton trilayer. Key questions for the coming decade will center on the implications of the enormous (and increasingly too apparent) diversity of cytoskeletal proteins and the mechanisms by which they contribute to vesicle transport, membrane topography, and signal transduction. Also of interest will be the ways these interactions are regulated at both the transcriptional and post-translational level, and the molecular pathology of disorders with defective membrane-cytoskeletal function. Many of these questions are framed in the present volume. Individually, each author addresses a specific problem of identification, assembly, or regulation of a cytoskeletal protein or system. Collectively, these chapters highlight the rich diversity and daunting complexity of the membrane-cytoskeletal trilayer, and establish a clear framework within which general principles of motility, cytoskeletal membrane assembly, and control will emerge. xvii
xviii
Preface
The editors wish to extend their most sincere appreciation to the many individuals and institutions who contributed to the Symposium and this volume. Foremost credit goes to the authors, whose enthusiastic participation insured its success and merit. Equal credit is extended to our corporate sponsors: Boehringer Ingelheim Pharmauceticals, Bristol-Myers Squibb Co., Hoffman-LaRoche, Merck Sharp and Dohme Research Laboratories, and Miles Inc. Credit is also due the Department of Molecular and Cellular Physiology at Yale. Without sponsors, there would be no conference or book. MARKS. MOOSEKER JON S. MORROW
Previous Volumes in Series Current Topics in Membranes and Transport Volume 12 Carriers and Membrane Transport Proteins (1979) Edited by F. Bronner and A. Kleinzeller Volume 13 Cellular Mechanisms of Renal Tubular Ion Transport* (1980) Edited by Emile L. Boulpaep Volume 14 Carriers and Membrane Transport Proteins (1980) Edited by F. Bronner and A. Kleinzeller Volume 15 Molecular Mechanisms of Photoreceptor Transduction* (198 1) Edited by William H. Miller Volume 16 Electrogenic Ion Pumps* (1982) Edited by Clifford L. Slayman Volume 17 Membrane Lipids of Prokaryotes (1982) Edited by Shmuel Razin and Shlomo Rottem Volume 18 Membrane Receptors (1983) Edited by Arnost Kleinzeller and B. Richard Martin Volume 19 Structure, Mechanism, and Function of the Na/K Pump* (1983) Edited by Joseph F. Hoffman and Bliss Forbush 111 Volume 20 Molecular Approaches to Epithelial Transport* (1984) Edited by James B. Wade and Simon A. Lewis Volume 21 Ion Channels: Molecular and Physiological Aspects (1984) Edited by Wilfred D. Stein Volume 22 The Squid Axon (1984) Edited by Peter F. Baker Volume 23 Genes and Membranes: Transport Proteins and Receptors* (1985) Edited by Edward A. Adelberg and Carolyn W. Slayman Volume 24 Membrane Protein Biosynthesis and Turnover (1985) Edited by Philip A. Knauf and John S. Cook * Part of the series from the Yale Department of Cellular and Molecular Physiology. xix
xx
Previous Volumes in Series
Volume 25 Regulation of Calcium Transport across Muscle Membranes (1985) Edited by Adil E. Shamoo Volume 26 Na+ -H Exchange, Intracellular pH, and Cell Function* (1986) Edited by Peter S . Aronson and Walter F. Boron +
Volume 27 The Role of Membranes in Cell Growth and Differentiation (1986) Edited by Lazaro J. Mandel and Dale J. Benos Volume 28 Potassium Transport: Physiology and Pathophysiology* ( 1987) Edited by Gerhard Giebisch Volume 29 Membrane Structure and Function (1987) Edited by Richard D. Klausner, Christoph Kempf and Jos van Renswoude Volume 30 Cell Volume Control: Fundamental & Comparative Aspects in Animal Cells (1987) Edited by R. Gilles, Arnost Kleinzeller and L. Bolis Volume 31 Molecular Neurobiology: Endocrine Approaches (1987) Edited by Jerome F. Strauss 111 and Donald W. Pfaff Volume 32 Membrane Fusion in Fertilization, Cellular Transport, and Viral Infection ( 1 988) Edited by Nejat Duzgunes and Felix Bronner Volume 33 Molecular Biology of Ionic Channels* (1988) Edited by William S . Agnew, Toni Claudio, and Frederick J. Sigworth Volume 34 Cellular and Molecular Biology of Sodium Transport* (1989) Edited by Stanley G. Schultz Volume 35 Mechanisms of Leukocyte Activation (1990) Edited by Sergio Grinstein and Ori D. Rotstein Volume 36 Protein-Membrane Interactions* ( 1990) Edited by Toni Claudio Volume 37 Channels and Noise in Epithelial Tissues (1990) Edited by Sandy I. Helman and Willy Van Driessche
Current Topics in Membranes Volume 38 Ordering the Membrane Cytoskeleton Tri-Layer* (1991) Edited by Mark S. Mooseker and Jon S. Morrow Volume 39 Developmental Biology of Membrane Transport Systems (1991) Edited by Dale J. Benos (in preparation)
CHAPTER 1
Molecular Genetic Analyses of Drosophila Ki n e s i n Russell J. Stewart and Lawrence S. B. Goldstein Department of Cell and Developmental Biology Harvard University Cambridge, Massachusetts 021 38
I . Introduction 11. Kinesin Structure A . Electron Microscopy
B . Drosophila Kinesin Heavy Chain Sequence Analysis C. Molecular Genetic Dissection of the Drosophila Kinesin Heavy Chain D. Comparison of the Drosophila Kinesin Heavy Chain to Other Kinesin Heavy Chain and Kinesin-Like Sequences 111. Kinesin Function A. Cellular Function B. Kinesin Heavy Chain Motility in V i m C. Kinesin Mechanochemical Cycle IV. Conclusion References
1. INTRODUCTION
Eukaryotic cells use active and controlled intracellular movements of cytoplasmic components to organize and compartmentalize the cytoplasm. Intracellular movements are necessary for establishment of cytoplasmic order and also for communication between the specialized compartments of the cytoplasm. These needs are at least partly fulfilled by a growing list of cytoskeleton-associated mechanochemical proteins that includes myosins, dyneins, dynamin, and kinesin. These mechanochemical proteins hydrolyze ATP and move along cytoskeletal filaments-either actin or microtubules. Current Topics in Membranes, Volume 38 Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.
1
2
Russell J. Stewart and Lawrence S. B. Goldstein
Kinesin is a microtubule-associated mechanochemical protein complex. Although the exact cellular function of kinesin is not yet understood, the occurrence of kinesin in most eukaryotic cell types suggests that kinesin plays a fundamental role in the spatial organization of the eukaryotic cytoplasm. Recently, several kinesin-like proteins with more clearly defined intracellular motility functions have been discovered using genetic and molecular genetic methods. The Saccharomyces cerevisiae KAR3 gene (Meluh and Rose, 1990), the Asperigillus bimC gene (Enos and Morris, 1990), the Drosophila ncd gene (Endow et al., 1990; McDonald and Goldstein, 1990), and the nod gene of Drosophila (Zhang et al., 1990) all encode proteins related to kinesin. The discovery of these kinesin-like genes led to the realization that there probably exists a superfamily of kinesin-like proteins with perhaps dozens of members that perform highly specialized intracellular motility functions (Vale and Goldstein, 1990). The members of this superfamily are related by possessing a common microtubuleassociated mechanochemical domain. However, all of the members of this superfamily discovered thus far have distinct specialized domains that are likely to associate with specific cellular components. In this chapter, we briefly review the kinesin literature relevant to kinesin structure and summarize recent results concerning the structure and the in vitro motility properties of genetically manipulated forms of the Drosophila kinesin heavy chain. We present biochemical evidence that the kinesin mechanochemical head domain is a functional unit independent of additional domains to which it can be attached. These results strongly support the concepts regarding the design and implementation in the cell of kinesin-like proteins emerging from molecular genetic studies of intracellular motility.
II. KINESIN STRUCTURE Kinesin was originally identified in extracts of squid axoplasm by virtue of its ability to move microtubules along a glass surface, or carboxylated beads along microtubules (reviewed by Vale, 1987). Subsequently, kinesin has been isolated from many tissue types, usually by taking advantage of the tight association of kinesin with microtubules in the presence of AMPPNP (a nonhydrolyzable ATP analog). Native kinesin purified by microtubule affinity has been found to be a heterotetramer composed of two heavy chains (1 10- 130 kDa) and two light chains (60-80 kDa) (Kuznetsov et al., 1988; Bloom et al., 1988). The arrangement of the subunits and the broad overall structure of the kinesin heavy chain have been determined by electron microscopy of native kinesin and by sequence analysis of the Drosophila kinesin heavy chain gene.
1. Genetic Analyses of
Drosophilu Kinesin
3
A. Electron Microscopy
Native kinesin appeared in the electron microscope as an elongate molecule 80-100 nm in length, with a pair of globular domains 10 nm in diameter at one end, and a “feathered” or fan-shaped domain at the opposite end (Amos, 1987; Hirokawa et al., 1989; Scholey et al., 1989). The rodlike stalk connecting the globular head domains and the fan-shaped tail was 2-4 nm in diameter and often appeared bent as though it contained a flexible hinge region. The identity of the domains and the location of the light chains have been established by antibody labeling of native kinesin. Hirokawa et al. (1989) demonstrated that monoclonal antibodies against the kinesin light chains decorated the fan-shaped tail of the kinesin complex and that monoclonal antibodies recognizing heavy chain epitopes decorated the globular head domains. Quick-freeze, deep-etch electron microscopy of kinesin bound to microtubules suggested that the globular head domains associated with microtubules. In similar experiments, Scholey et al. (1989) decorated kinesin with a monoclonal antibody (designated SUK 4; Ingold et al., 1988) that recognized a 45-kDa proteolytic fragment of the kinesin heavy chain. The 45-kDa proteolytic fragment displayed microtubuleactivated ATPase activity, which suggested that it represented the mechanochemical domain of kinesin (Kuznetsov et al., 1989). The SUK 4 antibodies bound to the 10-nm globular heads of native kinesin, confirming that the globular heads were part of the heavy chain and further demonstrating that the heavy chains were arranged in parallel. Therefore, the globular heads correspond to the 45-kDa proteolytic fragment and most likely represent the mechanochemical domain of kinesin. In additional experiments, the location of the SUK 4 epitope was mapped using genetically truncated kinesin fragments. These experiments demonstrated that the mechanochemical domain was located within the N-terminal region of the heavy chain (see below). A model of our current understanding of kinesin ultrastructure is diagrammed in Fig. 1.
B. Drosophila Kinesin Heavy Chain Sequence Analysis
At the same time that the structure of kinesin was being determined by electron microscopy, a three-domain structure of the kinesin heavy chain was predicted by analysis of the DNA sequence of the Drosophilu kinesin heavy chain gene (Yang ef al., 1989). A Drosophila kinesin heavy chain cDNA clone was isolated (Yang et al., 1988) using an antibody against the Drosophila kinesin heavy chain purified by the conventional method of microtubule affinity in the presence of AMPPNP (Saxton et al., 1988). That the gene corresponded to bona fide Drosophila kinesin heavy chain was demonstrated by comigration of the gene product
Russell J. Stewart and Lawrence S. B. Goldstein
4
BEND?
FIG. 1 A model of the three-domain ultrastructure of native kinesin. The globular head domains, corresponding to the amino termini of two heavy chains, contain the microtubule and ATP binding sites. The stalk most likely consists of a coiled-coil structure formed by the interaction of the extended a-helical regions of two heavy chains. Two coiled-coil regions (C.C.1 and C.C.11) are separated by a region that often appears bent. The bend may correspond to a helix-disrupting proline at amino acid position 587. The kinesin light chains associate with the carboxy termini of the heavy chains to form a “feathered” domain.
with native Drosophila kinesin heavy chain at 115 kDa, by peptide mapping, by cross-reactivity of the gene product with antisera against both squid and sea urchin kinesin, by the nucleotide-dependent microtubule-binding properties of the gene product, and later by sequence comparison with kinesin heavy chain genes from squid (Kosik et al., 1990) and sea urchin (J. M. Scholey, personal communication). Examination of the predicted secondary structures of the kinesin heavy chain using the methods of Gamier et al. (1978) revealed three distinct regions in the predicted amino acid sequence. The N-terminal400 residues appeared to form a globular domain because of frequent alternations between short regions of ahelix, P-sheet, and p-turns. Within this potentially globular domain an ATPbinding consensus sequence (GXXXXGKTXXXXXXIN; Walker et al., 1982) was found, as would be expected for an ATP-dependent mechanochemical protein domain. In contrast to the N-terminal portion of the kinesin heavy chain, the center of the molecule, from the proline at position 399 to the proline at position 883, was predicted to be mostly a-helical, and therefore most likely forms an extended structure corresponding to the stalk observed by electron microscopy. The final 92 residues were characterized by frequent alternations in predicted secondary structure and therefore most likely also formed a globular domain. The structure predicted from the kinesin heavy chain sequence correlates well with the electron microscopic observations of kinesin structure described above. Additional information regarding the subunit organization of the kinesin complex came from analysis of the sequence of the extended a-helical domain (residues 399-883). This region was discovered to contain, from Leu-437 to Gln-830, an heptapeptide repeat pattern a,b,c,d,e,f,g, with enrichment of hydrophobic residues at positions a and d. A heptad repeat pattern with strong peaks of hydrophobicity at positions a and d has been found to be characteristic of an a-
1 . Genetic Analyses of Drosophila Kinesin
5
helical coiled-coil conformation (McLachlan and Karn, 1983). The extended ahelical domain of kinesin is therefore likely to be responsible for the dimerization of kinesin heavy chains by the formation of an extended a-helical coiled-coil. This prediction has been largely confirmed by physical characterization of a polypeptide produced in Escherichia coli by expression of a truncated kinesin heavy chain gene containing only the sequence of the a-helical domain. The purified recombinant stalk protein dimerized and had a circular dichroism spectrum consistent with a coiled-coil conformation (de Cuevas e f al., 1991).
C. Molecular Genetic Dissection of the Drosophila Kinesin Heavy Chain The functional domain structure of the kinesin heavy chain was further defined by Yang et al. (1989), who assayed nucleotide-dependent microtubule binding of kinesin heavy chain polypeptides transcribed and translated in vitro from truncated kinesin genes. The full-sized kinesin heavy chain demonstrated the same behavior as native kinesin in the microtubule-binding assay. In the presence of AMPPNP the kinesin heavy chain bound and sedimented with microtubules, but did not bind to microtubules and remained in the supernatant in the presence of ATP. Analysis of several kinesin heavy chain polypeptides with C-terminal truncations revealed that a protein consisting of as little as the N-terminal50% of the heavy chain displayed nucleotide-dependent microtubule binding identical to the full-sized kinesin heavy chain, confirming that the mechanochemical domain is at the N-terminus of the heavy chain. Analysis of additional truncated polypeptides established a C-terminal boundary for the microtubule-binding site near residue 390. A deletion from the N-terminus that removed the first 170 amino acids of the kinesin heavy chain, which included the ATP binding consensus sequence (residues 86- 106), resulted in a polypeptide that sedimented with microtubules in the presence of both AMPPNP and ATP. The nucleotide independence of microtubule binding by this polypeptide supported the suggestion that the N-terminal region of the heavy chain was involved in ATP binding and hydrolysis and also suggested that the microtubule-binding domain may fold independently of the ATP-binding region of the kinesin heavy chain. Other N-terminal truncations extending further into the heavy chain gave more ambiguous microtubule-binding results, and so a reliable N-terminal boundary of the microtubule-binding site was not established.
D. Comparison of the Drosophila Kinesin Heavy Chain t o Other Kinesin Heavy Chain and Kinesin-Like Sequences The Drosophila kinesin heavy chain gene shares striking overall homology with the squid kinesin heavy chain gene (Kosik et al., 1990). As with Drosophila
6
Russell J. Stewart and Lawrence S. B . Goldstein
kinesin, the secondary structure and physicochemical parameters predicted from the sequence of squid kinesin revealed the three-domain structure of the kinesin heavy chain. The three domains were also demarcated by the degree of conservation between Drosophila and squid kinesins. The highest sequence conservation occurs in the globular domains at the head and tail, -80 and -75% respectively, while the region between residues -350 and -750 shares only -50% identity. The breaks in similarity may represent the approximate junctions between the three domains. In contrast, much more limited sequence similarity was found in comparisons of the predicted amino acid sequence of the Drosophila kinesin heavy chain with the predicted amino acid sequences of the S. cerevisiae KAR3 gene (Meluh and Rose, 1990), the Aspergillus bimC gene (Enos and Morris, 1990), and the ncd gene of Drosophila (Endow et al., 1990; McDonald and Goldstein, 1990). All four genes shared about 40-45% identity over a 350-amino acid region corresponding to the mechanochemical domain of kinesin, with especially strong identities in the ATP- and microtubule-binding regions. However, outside of the mechanochemical domain, virtually no similarity was found, despite the suggestion that all four genes have three-domain structures. Interestingly, the KAR3 and ncd gene products are arranged with the kinesin-like mechanochemical domain at the C-terminus of the protein. Although it has not been demonstrated that any of the recently discovered kinesin-like proteins have motility properties similar to kinesin, the phenotypes of mutants in which the proteins are defective suggest that they are all involved in intracellular movements. If these proteins are kinesin-like motors, then the functional motor domain must comprise only 350 amino acids and must operate independently of additional domains to which it is attached.
111. KINESIN FUNCTION
A. Cellular Function
The cellular function of kinesin is apparently to move cytoplasmic components along microtubules from one location in the cell to another. Exactly what components are moved and how these movements are regulated are still incompletely understood. The presence of kinesin in most eukaryotic organisms and tissue types suggests that it may perform a general housekeeping function rather than a tissue or cell-cycle-specific function. Considerable evidence suggests that kinesin may be involved in transporting membranous components of the cell, e.g., anterograde vesicle movement in axons (Vale et al., 1985; Schroer et al., 1988), and may, perhaps, be involved in elaborating the endoplasmic reticulum (Dabora and Sheetz, 1988; Vale and Hotani, 1988; Hollenbeck, 1989).
1. Genetic Analyses of Drosophila Kinesin
7
Some of the cellular functions once proposed for kinesin, e.g., mitosis, are more likely performed by kinesin-like proteins. Genetic analysis suggests that the S. cerevisiae KAR3 gene product is involved in the migration of nuclei following mating and may have a mitotic role as well (Meluh and Rose, 1990). Similarly, the bimC gene product of Aspergillus appears to be involved in spindle pole body separation before the onset of mitosis (Enos and Morris, 1990), the ncd gene product of Drosophila is involved in female meiosis and early zygotic mitosis (Sturtevant, 1929; Lewis and Gencarella, 1952; Davis, 1969; Kimble and Church, 1983; Sequeira et al., 1989; Yamamoto et al., 1989), and the nod gene product of Drosophila is also involved in female meiosis and possibly in mitosis (Carpenter, 1973; Zhang and Hawley, 1990). These observations suggest that the mechanochemical head domain of the primordial kinesin has been repeatedly utilized to accomplish specific motility functions by fusing it with distinct tail domains specialized to associate with a particular cargo (Vale and Goldstein, 1990). Until the identification of the kinesin-like ncd and nod genes in Drosophila, there had been no evidence of a single organism possessing more than a single kinesin-like gene. Now, mounting evidence suggests that there exists a large superfamily of kinesin-like proteins within a single multicellular organism. Drosophila possesses at least five kinesin-like genes in addition to kinesin, ncd, and nod (Stewart et al., 1991). Characterization of these genes is likely to reveal several fascinating combinations of microtubule-based motility with other activities. B. Kinesin Heavy Chain Motility in Vitro
The mechanics of how kinesin and the kinesin-like proteins generate force against microtubules are almost entirely unknown. The recent isolation and subsequent expression in E . coli of the Drosophila kinesin heavy chain gene together with in vitro assays of kinesin function have opened new avenues to study the mechanism of kinesin-mediated motility. One of the first questions we have addressed with this capability concerns the minimum motile element of the kinesin heavy chain (Yang et al., 1990). Expression of the full-length kinesin heavy chain gene in E. coli from any of several expression vectors, followed by partial purification of kinesin from E. coli lysates by microtubule affinity in the presence of AMPPNP, resulted in a kinesin protein active in moving microtubules in an in vitro video microscopy assay. The rate and direction of microtubule movement generated by the recombinant kinesin were the same as the rate and direction generated by native kinesin isolated from Drosophila. These results demonstrated that the kinesin heavy chain, in the absence of the kinesin light chains and any other eukaryotic factors, was capable of generating microtubule movements.
8
Russell J. Stewart and Lawrence S. B. Goldstein
To determine the elements of the heavy chain that are essential for motility, several genetically manipulated versions of the heavy chain were expressed in E . coli, partially purified, and assayed for microtubule motility activity. Briefly, heavy chains truncated from the C-terminus up to residue 533 were active in moving microtubules at rates similar to full-sized and native kinesin, demonstrating that the C-terminal two-thirds of the kinesin stalk, including the potential bend region, was inessential for motility. The amino-terminal third of the stalk was then deleted, leaving only the apparently inessential C-terminal portion of the stalk. This construct also moved microtubules at rates comparable to native kinesin, suggesting that the entire stalk was inessential for motility. Previously, a 45-kDa proteolytic fragment of the kinesin heavy chain, corresponding to the N-terminal mechanochernical domain, had been shown to be inactive in moving microtubules in vitro (Kuznetsov et al., 1989). Likewise, when only the kinesin head (447 residues) was expressed and assayed it too failed to move microtubules in vitro. Two explanations were possible: either the stalk actively contributed to motility, or, in the absence of the stalk, the head failed to adhere to glass or to adhere in a suitable orientation to generate motility. This issue was resolved by constructing a chimeric kinesin molecule containing the N-terminal447 residues of the kinesin head and, in place of the kinesin stalk, 1280 residues of Drosophila a-spectrin. The chimeric protein moved microtubules at rates comparable to those of native kinesin, providing additional evidence that the first 447 amino acids of the kinesin head contained all of the elements necessary to generate force against microtubules (Yang et al., 1990). More recent evidence (R. J. Stewart and L. S . B. Goldstein, unpublished observations) suggests that as few as the first 340 amino acids are sufficient to generate microtubule motility in vitro and that the kinesin head domain is active with the artificial a-spectrin tail at either the C- or N-terminus.
C. Kinesin Mechanochemical Cycle Like the other filament-associated mechanochemical proteins (i.e., myosins and dyneins), the directional movement of kinesin along microtubules is powered by the hydrolysis of ATP (Kuznetsov and Gelfand, 1986). The basal rate of ATP hydrolysis by kinesin is enhanced as much as 1000-fold by the presence of microtubule (Hackney, 1988). Presumably, translocation along microtubules occurs as a result of the cyclical formation of crossbridges between the kinesin head and the microtubule surface coupled to conformation changes within the kinesin molecule resulting from ATP hydrolysis and product release. The structural and enzymatic similarities of kinesin and myosin invite comparisons between their potential mechanisms of force generation. Two major models have been proposed to account for force generation by myosin: the
1. Genetic Analyses of Drosophila Kinesin
9
rotating head model (Huxley, 1969), and the helix-coil transition model (Harrington, 1971). The first proposes that the major conformational change leading to generation of force against actin occurs within the head or at the junction between the head and stalk, while the second proposes that the conformational change occurs when the coiled-coil structure in the hinge region near the center of the tail melts into a shorter, less ordered structure. This issue has remained unresolved for nearly two decades. However, the recent demonstration that the myosin head, in the absence of the hinge region, was capable of in viho motility strongly supports the proposal that the conformational change occurs within or very near the head (Hynes et al., 1987). The possibility remains, however, that in vivo a dual mechanism pertains, involving both types of conformational change. Likewise, our demonstration that the kinesin head alone was capable of moving microtubules in vifro may preempt suggestions that the conformational change occurs in the stalk, and suggests that the conformational change occurs within the head. Also, our observation that the head domain is active whether it forms the N- or C-terminus of the protein suggests that there are no specific constraints on the nature of the junction between the head and stalk.
N. CONCLUSION Our biochemical analysis, taken in concert with the finding that there exists a superfamily of proteins with domains that are similar to the N-terminal domain of the kinesin heavy chain, suggests that the kinesin head constitutes a functionally autonomous, microtubule-associated mechanochemical domain that is completely independent of additional domains to which it has been attached and is even independent of its orientation within the linear sequence of the molecule. Furthermore, the conformational change that results in the generation of force against microtubules most likely occurs entirely within this conserved mechanochemical domain. The capability of expressing the kinesin heavy chain in a tractable organism like E . coli offers exciting possibilities for studying the nature of this conformational change. As examples, it should be possible to substitute specific amino acids or to engineer precisely located reporter groups by in v i m mutagenesis.
References Amos, L. A . (1987). Kinesin from pig brain studied by electron microscopy. J . Cell Sci. 87, 105111.
Bloom, G . S., Wagner, M . C., Pfister, K. K . , and Brady, S . T. (1988). Native structure and physical properties of bovine brain kinesin and identification of the ATP-binding subunit polypeptide. Biochemistry 27, 3409-34 16. Carpenter, A . T. C. (1973). A mutant defective in distributive disjunction in Drosophilu melunogaster. Genetics 73, 393-428.
10
Russell J. Stewart and Lawrence S. B. Goldstein
Dabora, S. L., and Sheetz, M. P. (1988). The microtubule-dependent formation of a tubulovesicular network with characteristics of the ER from cultured cell extracts. Cell (Cambridge, Mass.) 54, 27-35. Davis, D. G. (1969). Chromosome behavior under the influence of claret-nondisjunctional in Drosophila melanogusrer. Genetics 61, 577-594. de Cuevas, M., Tao, T., and Goldstein, L. S. B. (1991). In preparation. Endow, S. A., Henikoff, S., and Niedziela, L. S. (1990). Mediation of meiotic and early mitotic chromosome segregation in Drosophila by a protein related to kinesin. Naiure (London) 345, 81-83. Enos, A. P., and Moms, N. R. (1990). Mutation of a gene that encodes a kinesin-like protein blocks nuclear division in A. niduluns. Cell (Cambridge, Mass.) 60, 1019-1027. Gamier, J . , Osguthorpe, D. J., and Robson, B. (1978). Analysis of the accuracy and implications of simple methods for predicting the secondary structure of globular proteins. J. Mol. Biol. 120, 97- 120. Hackney, D. D. (1988). Kinesin ATPase: Rate-limiting ADP release. Proc. Nurl. Acud. Sci. U.S.A. 85, 6314-6318. Hanington, W. F. (197 I). A mechanochemical mechanism for muscle contraction. Proc. Nurl. Acud. Sci. U.S.A. 68, 685-689. Hirokawa, N., Pfister, K . K., Yorifuji, H., Wagner, M. C., Brady, S. T., and Bloom, G. S. (1989). Submolecular domains of bovine brain kinesin identified by electron microscopy and monoclonal antibody decoration. Cell (Cambridge, Muss.) 56, 867-878. Hollenbeck, P. J. (1989). The distribution, abundance and subcellular localization of kinesin. J. Cell Biol. 108, 2335-2342. Huxley, H. E. (1969). The mechanism of muscle contraction. Science 164, 1356-1366. Hynes, T. R., Block, S. M., White, B. T., and Spudich, J. A. (1987). Movement of myosin fragments in vitro: Domains involved in force production. Cell (Cambridge, Muss.) 48, 953963. Ingold, A. L., Cohn, S . A,, and Scholey, J. M. (1988). Inhibition of kinesin-driven microtubule motility by monoclonal antibodies to kinesin heavy chain. J. Cell Biol. 107, 2657-2667. Kimble, M., and Church, K. (1983). Meiosis and early cleavage in Drosophila melunogusrer eggs: Effects of the claret-nondisjunctional mutation. J. Cell Sci. 62, 301-3 18. Kosik, K. S., Orecchio, L. D., Schnapp, B., Inouye, H., and Neve, R. L. (1990). The primary structure and analysis of the squid kinesin heavy chain. J. Biol. Chem. 265, 3278-3283. Kuznetsov, S. A., and Gelfand, V. I. (1986). Bovine brain kinesin is a microtubule-activated ATPase. Proc. Natl. Acud. Sci. U.S.A. 83, 8530-8534. Kuznetsov, S. A., Vaisberg, E. A,, Shanina, N. A., Magretova, N. N., Chernyak, V. Y., and Gelfand, V. 1. (1988). The quaternary structure of bovine brain kinesin. EMBO J. 7, 353-356. Kuznetsov, S. A , , Vaisberg, Y. A,, Rothwell, S. W., Murphy, D. B., and Gelfand, V. I. (1989). Isolation of a 45-kDa fragment from the kinesin heavy chain with enhanced ATPase and microtubule-binding activities. J. Biol. Chem. 264, 589-595. Lewis, E. B., and Gencarella, W. (1952). Claret and nondisjunction in Drosophilu melanogusrer. Generics 37, 600-601. McDonald, H. B., and Goldstein, L. S. B. (1990). Identification and characterization of a gene encoding a kinesin-like protein in Drosophila. Cell (Cambridge, Mass.) 61, 991- 1000. McLachlan, A. D., and Karn, J. (1983). Periodic features in the amino acid sequence of nematode myosin rod. J. Mol. Biol. 164, 605-626. Meluh, P. B., and Rose, M. D. (1990). KAR3, a kinesin-related gene required for yeast nuclear fusion. Cell (Cambridge, Muss.) 60, 1029-1041. Saxton, W. M., Porter, M. E., Cohn, S. A,, Scholey, J. M., Raff, E. C., and McIntosh, J. R. (1988). Drosophila kinesin: Characterization of microtubule motility and ATPase. P roc. Nurl. Acad. Sci. U.S.A. 85, 1109-1113.
1. Genetic Analyses of Drosophila Kinesin
11
Scholey, J. M., Heuser, J., Yang, J. T., and Goldstein, L. S. B. (1989). Identification of globular mechanochemical heads of kinesin. Nature (London) 338, 355-357. 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. Sequeira, W., Nelson, C. R., and Szauter, P. (1989). Genetic analysis of the claret locus of Drosophila melanogaster. Genetics 123, 5 11-524. Stewart, R. J., Pesavento, P. A,, Woerpel, D. N., and Goldstein, L. S. B. (1991). Identification of a kinesin super family in Drosophila. Proc. Natl. Acad. Sci. U.S.A. (in press). Sturtevant, A. H. (1929). The claret mutant type of Drosophila sumulans. Z. Wiss. Zool. 135, 323356. Vale, R. D. (1987). Intracellular transport using microtubule-based motors. Annu. Rev. Cell Biol. 3, 347-378. Vale, R. D., and Goldstein, L. S. 8.(1990). One motor, many tails: An expanding repertoire of force-generating enzymes. Cell (Cambridge, Mass.J 60, 883-885. Vale, R. D., and Hotani, H. (1988). Formation of membrane networks in vitro by kinesin-driven microtubule movement. J. Cell Biol. 107, 2233-2241. Vale, R. D., Schnapp, B. J., Reese, T. S., and Sheetz, M. P. (1985). Organelle, bead, and microtubule translocations promoted by soluble factors from the squid giant axon. Cell (Cambridge, MUSS.J 40, 559-569. Walker, J. E.,Saraste, M., Runswick, M. J., and Gay, N. 1. (1982). Distantly related sequences in the a-and P-subunits of ATP synthase, myosin, kinases and other ATP-requiring enzymes and a common nucleotide binding fold. EMBO J. 1, 945-95 I . Yamamoto, A. H., Komma, D. J., Shaffer, C. D., Pirrotta, V., and Endow, S. A. (1989). The claret locus in Drosophila encodes products required for eyecolor and meiotic chromosome segregation. EMBO J. 8, 3543-3552. Yang, J. T., Saxton, W. M., and Goldstein, L. S. B. (1988). Isolation and characterization of the gene encoding the heavy chain of Drosophila kinesin. Proc. Natl. Acad. Sci. U . S . A . 85, 18641868. Yang, J. T., Laymon, R. A., and Goldstein, L. S. B. (1989). A three-domain structure of kinesin heavy chain revealed by DNA sequence and microtubule binding analyses. Cell (Cambridge, MUSS.) 56, 879-889. Yang, J. T., Saxton, W. M., Stewart, R. J., Raff, E. C., and Goldstein, L. S. B. (1990). Evidence that the head of kinesin is sufficient for force generation and motility in vitro. Science 249, 4247. Zhang, P., and Hawley, R. S. (1990). The genetic analysis of distributive segregation in Drosophila rnelanogaster. 11. Further genetic analysis of the nod locus. Genetics 125, 115-127. Zhang, P., Knowles, B. A,, Goldstein, L. S. B., and Hawley, R. S. (1990). A kinesin-like protein required for distributive chromosome segregation in Drosophila. Cell (Cambridge, Mass.)62, 1053- 1062.
This Page Intentionally Left Blank
CHAPTER 2
Acantharnoeba Myosin I: Past, Present, and Future Edward D. Korn Laboratory of Cell Biology National Heart, Lung, and Blood Institute National Institutes of Health Bethesda, Maryland 20892
I . Discovery 11. Multiplicity of Isoforms 111. Heavy Chain Sequences IV. Structural Basis of the Actin-Activated ATPase Activity
V. VI. VI1. VIII.
Mechanochemical Properties and Their Structural Basis Regulation lntracellular Localization Concluding Remarks References
This is an appropriate time to review the original experiments that led to the discovery of the myosin I isoforms in Acanthamoeba castellanii, the subsequent studies that led to their biochemical characterization, and the recent discoveries concerning their intracellular distribution, regulation, and possible functions. The myosin I isoforms of Dictyostelium and the intestinal brush border myosin I, which is the subject of another chapter in this volume (see Mooseker), are also briefly mentioned.
1. DISCOVERY
In the years 1969 to 1971, we demonstrated that the previously described cytoplasmic microfilaments in the soil amoeba Acanthamoeba castellanii were FCurrenf Topics in Membranes, Volume 38
13
14
Edward D. Kom
actin (Weihing and Kom, 1969, 1971; Pollard et a l . , 1970). The presence of these microfilaments in the cell cortex (Bowers and Kom, 1968), particularly in regions of phagocytic cup formation, provided preliminary evidence for a role of actin in motile events. On the assumption that where there was actin there also would be myosin (and, thus, actomyosin-dependent motile activity), Pollard and I set out to purify myosin from these cells. When purification methods based on procedures developed for the purification of muscle myosin (and which had been successfully applied to the purification of myosin from human platelets and Physarum polycephalum) failed, we tried a novel approach based on the unusual catalytic characteristics of skeletal muscle myosin. Muscle myosin has very low Mg2 -ATPase activity which is dramatically activated by F-actin. This is the physiologically relevant activity. But, muscle myosin also expresses a physiologically irrelevant Ca2 -ATPase activity, and, unique among ATPases, has very high catalytic activity in the presence of ethylenediaminetetraaceticacid (EDTA) and high concentrations of K + (or NH2) ions. Therefore, relying on the enzymatic rather than the structural properties of muscle myosin, we took a particulate-free supernatant of an amoeba homogenate, passed it over an agarose gel column, and assayed every fraction for ATPase activity in the presence of either Mg2+, Ca2+, or K+ and EDTA. The results were clear and surprising (Pollard and Kom, 1973a). There were about four peaks with low activity in the presence of Ca2 , but only the peak eluting at an approximate mass of 150 kDa also had both low ATPase activity in the presence of Mg2 and very high ATPase activity in the presence of K + and EDTA. This result was unexpected because the only myosins known at that time had masses of about 500 kDa, comprising a pair of heavy chains of about 200 kDa each and two pairs of light chains of about 20 and 16 kDa. When purified, the amoeba enzyme was shown to contain a single heavy chain of about 140 kDa and seemed to have two light chains of about 16 and 14 kDa (Pollard and Kom, 1973a). Although the amoeba enzyme was unable to form the thick filaments characteristic of all myosins known at that time, its Mg2+ATPase activity resembled that of myosin; it was substantially activated by Factin when a partially purified “cofactor” protein was also present (Pollard and Korn, 1973b). This cofactor protein was later found to be a myosin I heavy chainspecific kinase (Maruta and Korn, 1977b; Hammer et a l . , 1983). The Mg2+ATPase activity had a curious triphasic dependence on actin concentration, being first activated, then inhibited, and then activated again as the F-actin concentration was increased (Pollard and Korn, 1973b). Also, in those first papers, we reported that the amoeba enzyme seemed to cross-link filaments of F-actin (Pollard and Korn, 1973b). Despite its inability to form bipolar filaments analogous to the thick filaments of muscle myosins, Pollard and Kom (1973a,b) decided to call this amoeba enzyme “myosin” because of its actin-activated Mg2 -ATPase activity and its +
+
+
+
+
15
2. Acanthamoeba Myosin 1
ATP-reversible binding to F-actin. Although questioned at the time, this nomenclature was fully justified by later work (see below). Recently, the nomenclature was refined to use the term myosin I for all nonfilamentous, single-headed myosins (i.e., one heavy chain) and myosin I1 for the “conventional” filamentous, two-headed myosins (Korn and Hammer, 1988); Acanthamoeba also contains a myosin I1 isoform (Maruta and Korn, 1977a; Pollard et a l . , 1978). In their two papers published in 1973, Pollard and Korn (1973a,b) speculated that nonfilamentous myosin I might be capable of moving one actin filament relative to another by virtue of the postulated presence of two actin-binding sites in the same molecule, and, in addition, might serve to link actin filaments to the plasma membrane or other membranous structures. These two predictions have been verified by recent data and are now the major reasons for the growing interest in myosins of the type I class.
II. MULTIPLICITY OF ISOFORMS At the time of writing, three myosin I isozymes have been purified from Acanthamoeba castellanii and extensively characterized (Maruta et a l . , 1979; Gadasi et a l . , 1979; Gadasi and Korn, 1979; Lynch et a l . , 1989). By sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis, myosin IA has a heavy chain of 140 kDa and a light chain of 17 kDa, myosin IB a heavy chain of 125 kDa and a light chain of 27 kDa, and myosin IC a heavy chain of 130 kDa and a light chain of 14 kDa (Table I and Lynch et al., 1989; Albanesi et a l . , 1985b,c). Earlier indications that myosins IA and IB contained variable amounts of a 14-kDa light chain, in addition to the light chains that were unique to each, were almost certainly due to their contamination by the recently de-
TABLE I Properties of Acunrharnoeba Myosin I Isoenzyrnes
Molecular weight Native SDS-PAGE ATPase activity (sec-l) K + , EDTA Ca*+ Mgz+ Mgz+ + F-actin Mg2+ + F-actin + phosphorylated
Myosin 1A
Myosin IB
Myosin IC
159,000 140,000
150,000 125,000
+ 17,000
+27,000 21.4 4.4 0.3 0.5 17.4
162,000 130,000 + 14,000 15.3
22.2 2.1
0.3 0.6 18.1
3.3 0.2 1.5 20.3
16
Edward D. Korn
scribed myosin IC isozyme. Molecular genetic evidence suggests the presence of as many as three additional myosin I isozymes (Jung and Hammer, 1989b) for a possible total of six enzymes of the myosin I class in Acanthamoeba. A myosin I has also been purified from Dictyostelium discoideum (Cdtc et al., 1985; Fukui et al., 1989) and molecular genetic evidence, including sequence information, suggests the presence of at least four other myosin I isozymes in this organism (Jung and Hammer, 1990; Titus et a l . , 1989). We also now know that the 110K/calmodulin complex, originally identified as the protein that cross-links the microfilament bundle to the plasma membrane of intestinal brush border microvilli (Matsudaira and Burgess, 1979), is a myosin I (Collins and Borysenko, 1984; Coluccio and Bretscher, 1987; Conzelman and Mooseker, 1987).
111. HEAVY CHAIN SEQUENCES
Hammer and colleagues have cloned and sequenced the genomic DNAs corresponding to Acanthamoeba myosin IB (Jung and Hammer, 1990) and myosin IC (Jung et a l . , 1987) heavy chains. [The gene originally identified as myosin IB is, in fact, the myosin IC gene and that described originally as myosin IL is the myosin IB gene (Lynch et al., 1989; Brzeska et al., 1989a).] In addition, some partial protein sequences are available for myosin IA (Brzeska et a l . , 1989b; Lynch et a l . , 1986). The sequence data are represented schematically in Fig. 1A (Korn and Hammer, 1990), in which Acanthamoeba myosin IB is compared to a typical myosin 11, rat embryonic skeletal muscle myosin (Strehler et al., 1986), and in Fig. IB, in which Acanthamoeba myosin IC, one of the Dictystelium myosin I isozymes, and the bovine intestinal brush border myosin I are compared to myosin IB. Briefly summarized, residues 1 through 670 of myosin IB are 60% similar (allowing for conservative substitutions) to residues -80-800 of muscle myosin 11, the globular subfragment 1 that contains the catalytic activity of the muscle myosin. The remainder of the myosin IB sequence bears no similarity to the remainder of the myosin I1 sequence, i.e., the short carboxyl-terminal region of IB is totally dissimilar to the tail domains of the two heavy chains of myosin 11, which interact to form the a-helical coiled-coil rods that self-associate to form the bipolar filaments. In contrast, myosins IB and IC and the Dictyostelium isozyme sequenced by Jung et al. (1989a) can be aligned throughout almost their entire sequences. Interestingly, the carboxyl-terminal domains of all three have a glycine, proline, alanine (GPA)-rich region. Residues 902 to 1094 in myosin IB contain 48% glycine, 17% proline, and 12% alanine; residues 923 to 978 and 1034 to I168 in myosin IC are of similar composition; and residues 922 to 1059 in Dictyostelium are a GPQ-rich region (Q = glutamine). The rest of the carboxyl-terminal do-
17
2. Acanthamoeba Myosin I
A
a%-
ACANTHAMOEBA
B
1
0%
-670
ACANTHAMOEBA MYOSIN IB
902
1094 1147
I 'GPA
*]
1
ACANTHAMOEBA MYOSIN IC 1
mu690
922 10% 1111
DICTYOSTELIUM MYOSIN I 1
-720 -850 -1030 1043
BRUSH BORDER -46% MYOSIN I FIG. 1 Schematic comparisons of the sequences of the heavy chains of amoeba myosins I to each other and to mammalian myosin 11. (A) Comparison of Acanthamoeba myosin IB heavy chain to rat muscle myosin 11. The percent similarities are shown between the globular head domains (black areas) and carboxyl-terminal regions (cross-hatched area in myosin I1 and open area in myosin IB). (B) Comparisons of three other myosins I to the sequence of myosin IB. The dotted areas designate glycine/proline/alanine-richregions (glycine/proline/glutaminefor Dictyostelium myosin); the lined areas designate sequences unique to brush border myosin I; the percent similarities are shown for the remainder of the carboxyl-terminal regions (open areas). The precise locations of an ATP-binding site and the phosphorylation site and the approximate positions of two actin-binding sites and a membrane-binding site are indicated for Acanrhamoeba myosin IB. This figure is reproduced in modified form from Korn and Hammer (1990).
mains of these Acanthamoeba and Dictyostelium myosins I are almost as similar, as are their amino-terminal regions. Intestinal brush border myosin I does not have a GPA-rich region but does have a 130-residue region with strong sequence similarity to the rest of the carboxyl-terminal domains of the other myosins I (Hoshimaru and Nakanishi, 1987; Garcia er a l . , 1989). A second Dictyostelium isozyme, sequenced by Titus er al. (1989), is also similar to myosin IB except that it too lacks the GPA-rich region. All of the GPA-rich regions also have a net basic charge but little direct sequence similarity.
IV. STRUCTURAL BASIS OF THE ACTIN-ACTIVATED ATPase ACTIVITY As already mentioned, Acanthamoeba myosins I are actin-activated Mg2 ATPases with turnover numbers as high as those of any muscle myosin (Table I and Lynch er al., 1989) provided that the heavy chain has been phosphorylated +
Edward D. Kom
18
(see below). Early studies showed that full enzymatic activity was retained by the heavy chain of myosin IB when the light chains were removed by LiCl (Maruta et al., 1978). For this reason, subsequent structure-function studies were concentrated on the heavy chain. There are still no structural or sequence data for any of the myosin I light chains. The data for the heavy chains are summarized in Figs. 1 and 2. Direct photoaffinity labeling with radioactive UTP and ATP (Maruta and Korn, 1981a,b; Albanesi et al., 1984; Lynch et al., 1987) localized part of the ATP-binding site to a region about 11 kDa from the amino terminus (Fig. 1) that almost certainly, by analogy to Acanthamoeba myosin I1 (Atkinson et al., 1986), involved Glu- 108 (myosin IB sequence). The actin-binding site that activates the ATPase activity is located near, and possibly on both sides of, a trypsin-sensitive site about 64 kDa from the amino terminus (Figs. 1 and 2; Brzeska et al., 1988, 1989a). An 80-kDa amino-terminal fragment that is obtained by proteolysis in the presence of actin (Fig. 2) has full ach-activated Mg*+-ATPase activity (Brzeska et al., 1988), while a 64-kDa amino-terminal fragment obtained by proteolysis in the absence of actin (Fig. 2) has lost the coupling of ATPase activity to actin (Lynch et al., 1987; Brzeska et al., 1989a). In view of the
P
UNPHOSPHORY LATED OR PHOSPHORYLATED
-6 4 8 0
UNPHOSPHORY LATED + ACTIN C
6 4 8 0 PHOSPHORYLATED + ACTIN
-7
C
6 4 8 0 FIG. 2 Schematic representation of the major trypsin-sensitive sites of Acunrharnoebu myosin IA heavy chain and the effects of phosphorylation and F-actin on the susceptibility of these sites to trypsin cleavage. The numbers indicate the distance in kilodaltons of three trypsin-cleavage sites from the amino terminus. The 64-kDa site is protected and the 38-kDa site is partially protected by F-actin. Protection of the 38-kDa site is enhanced when the myosin is phosphorylated at a site about three residues from the cleavage site. This figure is reproduced from Brzeska et al. (1989a).
2 . Acanthamoeba Myosin I
19
sequence similarities between the Acanthamoeba myosins I and muscle myosins 11, it is not surprising that the ATP- and actin-binding sites of the myosins I are at the same positions as those of the muscle myosins (Mornet et al., 1989). The single phosphorylation sites of myosin IA, IB, and IC heavy chains (Figs. 1 and 2) have all been identified by partial sequencing of the phosphopeptides produced by controlled tryptic cleavage (Brzeska et al., 1989b): they are Ser-315 in myosin IB, Ser-311 in IC, and a corresponding Thr residue in IA (the heavy chain of which has not yet been sequenced). Interestingly, an adjacent trypsinsensitive site 38 kDa from the amino terminus (Fig. 2), which is partially protected from cleavage by F-actin in the absence of phosphorylation, is very strongly protected by F-actin when the heavy chain is phosphorylated (Brzeska et al., 1989a). Phosphorylation does not protect against tryptic cleavage in the absence of F-actin. The protective effects of actin on proteolysis led to the suggestion (Brzeska et al., 1989a) that binding of actin to unphosphorylated myosin I at sites about 64 kDa from the amino terminus directly protects against tryptic cleavage in that region and also induces a conformational change in the myosin heavy chain that confers partial protection against cleavage at the 38-kDa site (Fig. 2). Phosphorylation at a serine (myosins IB and IC) or threonine (myosin IA) immediately adjacent to this latter site probably induces a further conformational change that results in further protection against tryptic cleavage at the 38kDa tryptic site. This latter conformational change might then allow actin to activate hydrolysis of ATP by myosin I.
V. MECHANOCHEMICAL PROPERTIES A N D THEIR STRUCTURAL BASIS
The Acanthamoeba myosins I have two well-described mechanochemical properties: they can cause superprecipitation of cross-linked actomyosin complexes (Fujisaki et al., 1985) and they can support the movement of myosincoated beads (Albanesi et al., 1985a) or membrane vesicles (Adams and Pollard, 1986) along actin cables; both depend on ATP hydrolysis. The ability to crosslink actin filaments has been shown to be due to a second actin-binding site in the carboxyl-terminal domain (Fig. 1;Lynch et al., 1986). Speculatively, this second actin-binding site may be located in or near the GPA-rich regions of the carboxylterminal domains (Fig. l), but, as of now, there are no data on this point. The presence of this second (and ATP-insensitive) actin-binding site has been shown to be the reason for the triphasic behavior of the myosin I ATPase activity as a function of actin concentration (Albanesi et al., 1985a,b, 1986; Brzeska et al., 1988, 1989b; Lynch et al., 1986). The membrane-binding site of myosin I seems likely to be localized in the region of the carboxyl-terminal domain immediately adjacent to the subfragment 1-like domain of Acanthamoeba myosin I (Fig. 1 and Adams and Pollard, 1989)-a sequence that is shared by all of the myosins I for
20
Edward D. Korn
which relevant data have been published-but this point has also not yet been established unequivocally. Although the specific details of the molecular basis of the second actin-binding site and the membrane-binding site are not known, it seems clear that their presence in the carboxyl-terminal domain of myosin I permits these myosins to function analogously to myosins I1 (Fig. 3), even though they have only a single head and do not form bipolar filaments. By the simple expedient of having two actin-binding sites within their single heavy chains, Acanthamoeba myosins I are able to cross-link and move two actin filaments, one relative to the other, without the need for any higher degree of organization. Similarly, a membrane-binding site in the tail allows the monomolecular Acanthamoeba myosins I to support movement of membranes relative to actin filaments, or actin filaments relative to membranes, depending on which is anchored and which is free to move (Fig. 3). Recently, it has been noticed (Rodaway et al., 1989; Drubin et al., 1990) that the sequences of about 50 residues beginning at Ala-1097 of myosin IB and Ala-984 of myosin IC are strikingly similar to the previously described SH3 domain common to a number of proteins that bind to the membranecytoskeleton, including nonreceptor tyrosine kinases such as c-src and c-abl, phospholipase C, and spectrin. This sequence is also present in the Dictyostelium myosin I schematically shown in Fig. 1, but not in brush border myosin I. This region may be part of the ATP-insensitive F-actin-binding domain of the amoeba myosins I.
-
a >
C
Actin
1
Plasma Membrane
-
FIG. 3 Schematic representation of three different mechanisms by which myosins I might support movement: (A) movement of one actin filament relative to another, (B) movement of an actin filament relative to a fixed membrane or movement of a membrane relative to a fixed filament, (C) movement of a membrane-bound vesicle along an active filament.
21
2. Acanthamoeba Myosin I AGT*TYALNLNKMQAIGSRDALAKAMYSRIFD :: I : : : II 293TALLYRTITTGEQGRGRSS*VYSCPQDPLGAIYSRDALSKALYSRMFD33g
myosin
IA
myosin
IC
297NALLFRVLNTGGAGAKKMS*TYNVPQNVEQAASARDALAKTIYSRMVD3~3
myosin I B
IIIII:II:III:II
IlI:I:II
I : I I l l :
I :IIII:I::IIIII
FIG. 4 Comparison of the sequences around the phosphorylation sites of the heavy chains of Acantharnoeba myosin IA, IB, and IC. The phosphorylated amino acid in each heavy chain is
identified with an asterisk. Identical amino acid residues are connected by a line and conservative substitutions are connected by dots. The data are from Brzeska et al. (1989b).
VI. REGULATION
The actin-activated Mg2 -ATPase activities of the Acanthamoeba myosin I isozymes are regulated, as we have discussed above, by phosphorylation of a single serine or threonine in the heavy chain. The sequences around the phosphorylation sites of myosins IA, IB, and IC are shown in Fig. 4 (Brzeska et al., 1989b). Recent work has revealed considerably more about the specificity of the myosin I heavy chain kinase. As judged by the V,,, and K M for a number of synthetic peptides (Table II), there is a requirement for one or, preferably, two basic amino acids within a few residues on the amino-terminal side and for a tyrosine residue within a few residues on the carboxyl-terminal side of the phosphorylation site (Brzeska et al., 1990b). This unique specificity, and that for a tyrosine is particularly interesting, distinguishes this myosin I heavy chain kinase from any other kinase yet described. Of even greater interest is the discovery (Brzeska et al., 1990a) that myosin 1 heavy chain kinase is activated approximately 50-fold by autophosphorylation of up to eight sites and that the rate of autophosphorylation is enhanced at least 20+
TABLE I1 Substrate Specificity of Acanthamoeba Myosin I Heavy Chain Kinase
KM
v,,,
Peptidea
(pM)
(pmollmin x mg)
VrnAJKh4 ( x 103)
G B G B s s* V I s G _ L G B S S *V U S G B G_L S S* V l S G L G L S S* V U S G_RG_RS S* V L S G B G & S* S I V S
54 313 616 3520 I820 154
15 13 14 6 2 12
278 42 23 1.7 1 78
a Amino acids are identified by the single letter code: G , glycine; R, arginine: S. serine; V , valine; Y, tyrosine; L, leucine. The site of phosphorylation is indicated by an asterisk. Amino acids that significantly affect either K, or V,,,, are underlined.
22
Edward D.Korn
fold by the presence of phosphatidylserine (Brzeska er al., 1990a) or phosphatidylinositol (Brzeska et al., 1990b); neutral phospholipids, phosphatidylcholine and phosphatidylethanolamine, are without effects (Brzeska er al., 1990b). The acidic phospholipids affect only the rate of autophosphorylation of the kinase; they do not enhance the rate of phosphorylation of myosin I by the phosphorylated kinase. These results establish the existence of a regulatory cascade as shown in Fig. 5.
VII. INTRACELLULAR LOCALIZATION In 1980, Gadasi and Korn found by immunofluorescence studies that the myosin I isozymes are located near the plasma membrane of Acanrhamoeba. This was confirmed in similar studies by Hagen et al. (1986). The earlier studies also showed that myosin I was associated with highly purified preparations of plasma membranes (Gadasi and Korn, 1980). These observations were extended by Miyata er al. (1989), who showed that the myosin I was not linked to the plasma membrane through the F-actin that was also associated with the membranes, that the myosin I could be removed by elevated ionic strength, and that highly purified myosin I could rebind to the plasma membranes with a K,, of about 30-50 nM. In similar studies, Adams and Pollard (1989) found that myosin I could bind with equivalent affinity to a partially purified membrane fraction that included plasma membranes, even after the membranes were stripped of
FIG. 5 Regulatory cascade for the actin-activated Mg2+ -ATPase activity of Acanrhamoeba myosins I. Acidic phospholipids enhance the rate of autophosphorylation of myosin I heavy chain kinase about 20-fold; phosphorylation activates the kinase about 50-fold; phosphorylation of myosin I by the kinase enhances its actin-activated Mg2 + -ATPase activity at least 30-fold.
2. Acanthamoeba Myosin 1
23
most of their proteins by alkaline extraction. Adams and Pollard (1989) also showed that myosin I could bind to phospholipid vesicles. Very recently, these studies have been extended by Baines and Korn (1990) using immunoelectron microscopy. Whereas the studies by Gadasi and Korn (1980) and Miyata et al. (1989) had utilized polyclonal antibodies raised against myosins IA and IB, Baines and Korn used polyclonal antibodies raised against a 26-residue synthetic peptide corresponding to the sequence around the phosphorylation site of myosin IC. The antibodies were specific for myosin IC. In preliminary immunofluorescence studies, no reaction was detected unless the cells were first permeabilized with either saponin, Triton, or methanol. With gentle lipid extraction, only the plasma membrane was immunoreactive, but with more extreme pretreatment strong reaction was also seen at the contractile vacuole. By immunoelectron microscopy using gold-labeled second antibodies (Fig. 6), the myosin IC was found to be intimately associated with the plasma membrane and the membrane of the contractile vacuole. It is curious that both the plasma membrane and the contractile vacuole membrane appeared to be labeled predominantly on their external surfaces. As it is highly unlikely that the phosphorylation site of the myosin I heavy chain (the epitope against which the antibodies were raised) is exposed at the external surface of these membranes, the external orientation of the myosin IC was probably an artifact of the partial extraction of membrane lipids that was required for positive straining. Although it cannot be ruled out that the reactivity of the contractile vacuole membrane is due to another protein that shares the epitope [as, for example, was the case for nuclear reaction seen in the studies of Hagen er al. (1986)], this seems unlikely in light of the specificity of the primary antibody. An important advance was made in the studies by Fukui et al. (1989) of the distribution of myosin I in locomoting Dicfyosteliurn. They used an affinitypurified polyclonal antibody raised against the Dicfyosteliurn myosin I isozyme that had originally been purified by CBtt et al. (1985) and sequenced by Jung and Hammer (1990). Cells undergoing CAMP-induced chemotaxis, dividing cells, and phagocytosing cells all gave the same result: myosin I was strongly concentrated at the leading edges (Fig. 7) and at the phagocytic cup in addition to a diffuse cytoplasmic stain. In contrast, myosin I1 was found, as reported previously by others, only at the rear of locomoting cells and in the contractile ring (Fig. 7). F-actin was present at both the sites of myosin I and of myosin 11. Thus, it was proposed that actomyosin I is involved in motile activities at the leading edge of locomoting cells, where pseudopodal, lamellipodal, and phagocytic extensions form, while actomyosin I1 is involved in motile activities at the rear of locomoting cells and in the contractile ring of dividing cells. The roles of myosin I1 in movement and cytokinesis have been established by others (for recent references, see De Lozanne and Spudich, 1987; Knecht and Loomis, 1987; Wessels et a l . , 1988).
2 . Acanthamoeba Myosin I
25
The plasma membrane association of myosin I in Acanthamoeba and Dictyostelium and the phospholipid stimulation of Acanthamoeba myosin I heavy chain kinase make it very attractive to speculate that external ligands may regulate actomyosin I ATPase activity, and hence motile activity at the leading edge, through some modification of membrane lipids. Information on the distribution of phosphorylated and unphosphorylated kinase and phosphorylated and unphosphorylated myosin I is needed to test this hypothesis. The most recent attempts to establish the functions of myosin I have involved genetic engineering. To this end, Jung and Hammer (1990) have generated a Dictyostelium mutant missing the myosin I isozyme that had been localized by Fukui et al. (1989). Chemotactic streaming, aggregation, and phagocytosis were impaired in these cells but not abolished, indicating that this isozyme is not absolutely required for these processes. It seems likely that the four other myosin I isozymes now known to occur in Dictyostelium provide functional redundancy sufficient to obscure the role of any one of them. Thus, it may become necessary to study double, triple, or multiple myosin I deletion mutants to answer the question.
VIII. CONCLUDING REMARKS Other than a passing reference or two, this article has focused on the myosins I of Acanthamoeba castellanii, the organism in which they were discovered and the source of the best studied enzymes in the myosin I class. Acanthamoeba castellanii continues to be a very useful system, especially for biochemical studies, because of the relative ease of obtaining large quantities of cells. From presently available information, the biochemistry of the Dictyostelium myosins I
will be similar, but not necessarily identical, and this is a more favorable system for genetic manipulation. There is, however, yet more to the story. Brush border myosin I will probably be only the first of many examples of the myosin I class in vertebrate and mammalian cells. With the intriguing example of the ninaC gene of Drosophila (Monte11 and Rubin, 1988), which contains an amino-terminal domain similar to the catalytic domain of protein kinases coupled to a myosin I-like domain, and the very recent discovery by Horowitz and Hammer (1990) of a novel highmolecular-weight myosin I isozyme in Acanthamoeba, it seems highly likely that FIG. 6 Immunoelectron microscopic localization of myosin I in Acantharnoeba. Polyclonal antibodies specific for myosin IC were raised against a synthetic peptide with the sequence of the myosin IC phosphorylation site. Saponin-permeabilized cells were incubated with the antisera followed by gold-labeled second antibodies. Myosin IC was localized exclusively to the plasma membrane (A) and the contractile vacuole (cv) membrane (B) (Baines and Korn, 1990).
2. Acanthamoeba Myosin I
27
the business end of myosin (i.e., the subfragment 1 domain) will be found throughout nature linked to a variety of other protein domains that will adapt the “myosin” to specific functions.
References
Adams, R. J., and Pollard, T. D.(1986). Propulsion of organelles isolated from Acanfharnoeba along actin filaments by myosin 1. Nafure (London) 322, 754-756. Adams, R. J., and Pollard, T. D.(1989). Membrane-bound myosin I provides new mechanisms in cell motility. Nature (London) 340, 566-568. Albanesi, J. P., Fujisaki, H., and Korn, E. D.(1984). Localization of the active site and phosphorylation site of myosins IA and IB. J . Eiol. Chem. 259, 14184-14189. Albanesi, J. P., Fujisaki, H., Hammer, J. A., 111, Korn, E. D.,Jones, R., and Sheetz, M. P . (1985a). Monomeric Acanfhamoeba myosins I support movement of beads along actin cables. J . Eiol. Chem. 260, 8649-8652. Albanesi, J. P., Fujisaki, H., and Kom, E. D.(1985b). A kinetic model for the molecular basis of the contractile activity of Acanfhamoeba myosins IA and IB. J . Biol. Chem. 260, 11174-1 1179. Albanesi, J. P., Cout, M., Fujisaki, H.,and Korn, E. D.(1985~).Effect of actin filament length and filament number concentration on the actin-activated ATPase activity of Acanrhamoeba myosin I. J . Eiol. Chem. 260, 13276-13280. Albanesi, J. P., Lynch, T. J . , Fujisaki, H., and Korn, E. D.(1986). Regulation of the actin-activated ATPase activity of Acanfhamoeba myosin I by cross-linking actin filaments. J. Biol. Chem. 261, 10445- 10449. Atkinson, M. A. L., Robinson, E. A. Appella, E., and Kom, E. D.(1986). Amino acid sequence of the active site of Acantharnoeba myosin 11. J . Eiol. Chem. 261, 1844-1848. Baines, I., and Korn, E. D. (1990). Localization of myosin IC and myosin II in Acanthamoeba casfellanii by indirect immunofluorescence and immunogold electron microscopy. J . Cell Eiol. 111, 1895-1904. Bowers, B., and Korn, E. D. (1968). The fine structure of Acanfhamoeba casfellanii I. The trophozoite. J. Cell Eiol. 39, 95- 1 I 1. Brzeska, H., Lynch, T. J., and Kom, E. D. (1988). Localization of the actin-binding sites of Acanfhamoeba myosin 1B and effect of limited proteolysis on its actin-activated Mg* + -ATPase activity. J. Eiol. Chem. 263, 427-435. Brzeska, H., Lynch, T. J . , and Kom, E. D.(1989a). The effect of actin and phosphorylation on the tryptic cleavage pattern of Acanfhamoeba myosin IA. J . Eiol. Chem. 264, 10243-10250. Brzeska, H., Lynch, T. J., Martin, B., and Korn, E. D.(1989b). The localization and sequence of the phosphorylation sites of Acanfhamoeba myosin I. An improved method for locating the phosphorylated amino acid. J . Biol. Chem. 264, 19340-19348. Brzeska, H., Lynch, T. J . , and Korn, E. D.(1990a). Acanfhamoeba myosin I heavy chain kinase is activated by phosphatidylserine-enhancedphosphorylation. J . Eiol. Chem. 265, 3591-3594. Brzeska, H., Lynch, T. J., Martin, B., Corigliano-Murphy, A,, and Korn, E. D.(1990b). Substrate specificity of Acantharnoeba myosin I heavy chain kinase as determined by synthetic peptides. J . Eiol. Chem. 265, 16138-16144. Collins, J. H., and Borysenko, C. W. (1984). The 110,000-dalton actin-binding and calmodulin-
FIG. 7 Immunofluorescence localization of myosins 1 and I1 in a dividing Dictyostelium amoeba. (A) Phase-contrast image, (B) fluorescent image of the distribution of antibodies specific for myosin I, and (C) fluorescent image of the distribution of antibodies specific for myosin 11, all of the same cell. The micrographs were produced by Dr. Y. Fukui (Northwestern University) in collaboration with Drs. T. Lynch and E. D. Kom and are similar to those published in Fukui ef al. (1989).
28
Edward D. Korn
binding protein from intestinal brush-border is a myosin-like ATPase. J. Biol. Chem. 259, 14128- 14135. Coluccio, L. M., and Bretscher, A. (1987). Calcium-regulated cooperative binding of the microvillar 1 10 K calmodulin complex to F-actin--Formation of decorated filament. J. Cell Biol. 105,325333. Conzelman, K. A,, and Mooseker, M. S . (1987). The 110-kD protein calmodulin complex of the intestinal microvillus is an actin-activated MgATPase. J. Cell Biol. 105, 313-324. CBtC, G. P., Albanesi, J. P., Ueno, T., Hammer, J. A , , 111, and Korn, E. D. (1985). Purification from Dicryosrelium discoideum of a low-molecular weight myosin that resembles myosin I from Acanrhamoeba castellanii. J. Eiol. Chem. 260, 4543-4546. De Lozanne, A , , and Spudich, J. A. (1987). Disruption of the Dicryosrelium myosin heavy chain gene by homologous recombination. Science 236, 1086- 1091. Drubin, D. G., Mulholland, J., Zhu, Z., and Botstein, D. (1990). Homology of a yeast actin-binding protein to signal transduction proteins and myosin I. Nature (London) 343, 288-290. Fujisaki. H., Albanesi, J. P., and Korn, E. D. (1985). Experimental evidence for the contractile activities of Acanrhamoeba myosins IA and IB. J . Biol. Chem. 260, I 1182-1 1189. Fukui, Y.,Lynch, T. J., Brzeska, H., and Korn, E. D. (1989). Myosin I is at the leading edge of locomoting Dicryosrelium amoebae. Narure (London) 341, 328-33 1. Gadasi, H., and Korn, E. D. (1979). Immunochemical analysis of Acanthamoeba myosins IA, IB, and 11. J. Eiol. Chem. 254, 8095-8098. Gadasi, H., and Korn, E. D. (1980). Evidence for differential intracellular localization of the Acanthamoeha myosin isoenzymes. Narure (London) 286, 452-456. Gaddsi, H., Maruta, H., Collins, J. H., and Korn, E. D. (1979). Peptide maps of the myosin isoenzymes of Acanthamoeba casrellanii. J. Eiol. Chem. 254, 363 1-3636. Garcia, A , , Coudrier, E . , Carboni, J., Anderson, J., Vandekerckhove, J., Mooseker, M., Louvard, D., and Arpin, M. (1989). Partial deduced sequence of the 110-kD-calmodulin complex of the avian intestinal microvillus shows that this mechanoenzyme is a member of the myosin I family. J . Cell Eiol. 109, 2895-2903. Hagen, S. J., Kiehart, D. R., Kaiser, D. A,, and Pollard, T. D. (1986). Characterization of monoclonal antibodies to Acanthamoeha myosin I that cross-react with both myosin I1 low molecular weight nuclear proteins. J . Cell Eiol. 103, 2121-2128. Hammer, J. A., 111, Albanesi, J. P., and Korn, E. D. (1983). Purification and characterization of a myosin I heavy chain kinase from Acanrhamoeba castellanii. J. Eiol. Chem. 258, 1016810175. Horowitz, J., and Hammer, J. A., 111 (1990). A new Acanthamoeha myosin heavy chain gene: Cloning of the gene and immunological identitication of the polypeptide. J . Biol. Chem. 265, 20646-20652. Hoshimaru, M., and Ndkanishi, S . (1987). Identification of a new type of mammalian myosin heavy chain by molecular cloning. Overlap of its mRNA with preprotachykinin B mRNA. J. Eiol. Chem. 262, 14625- 14632. Jung, G., and Hammer, J. A , , I11 (1990). Generation and characterization of Dicryosrelium cells deficient in a myosin I heavy chain isoform. J . Cell Eiol. 110, 1955-1964. Jung, G . , Korn, E. D., and Hammer, J. A,, 111 (1987). The heavy chain of Acanthamoeba myosin IB is a fusion of myosin-like and non-myosin-like sequences. Proc. Natl. Acad. Sci. U.S.A. 84, 6720-6724. Jung, G., Saxe, C. L . , 111, Kimmel, A. R., and Hammer, J. A., I11 (1989a). Dictyosrelium discoideum contains a gene encoding a myosin I heavy chain. Proc. Natl. Acad. Sci., U.S.A. 86, 6 186-61 90. Jung, G., Schmidt, C. J., and Hammer, J. A , , 111 (l989b). Myosin I heavy chain genes of
2. Acanthamoeba Myosin I
29
Acanrhamoeba casrellanii: Cloning of a second gene and evidence for the existence of a third isofonn. Gene 82, 269-280. Knecht, D., and Loomis, W. F. (1987). Antisense RNA inactivation of myosin heavy chain gene expression in Dictyostelium discoideum. Science 236, 1081- 1086. Korn, E. D., and Hammer, J. A,, 111 (1988). Myosins of nonmuscle cells. Annu. Rev. Eiophys. Eiophys. Chem. 17, 23-45. Korn, E. D., and Hammer, J. A,, 111 (1990). Myosin I. Curr. Opin. Cell Eiol. 2, 57-61. Lynch, T. I., Albanesi, J. P., Korn, E. D., Robinson, E. A., Bowers, B., and Fujisaki, H. (1986). ATPase activities and actin-binding properties of sub-fragments of Acanfhumoeba myosin IA. J . Eiol. Chem. 261, 17156-17162. Lynch, T. J., Brzeska, H., and Korn, E. D. (1987). Limited tryptic digestion of Acanrhamoeba myosin IA abolishes regulation of actin-activated ATPase activity by heavy chain phosphorylation. J . Eiol. Chem. 262, 13842-13849. Lynch, T. J., Brzeska, H., Miyata, H., and Korn, E. D. (1989). Purification and characterization of a third isoform of myosin I from Acanrhamoeba casrellanii. J . Eiol. Chem. 264, 1933319339. Maruta, H., and Korn, E. D. (1977a). Acanrhamoeba myosin 11. J . Eiol. Chem. 252, 6501-6509. Maruta, H., and Korn, E. D. (1977b). Acanfhamoeba cofactor protein is a heavy chain kinase required for actin-activation of the Mg2+ -ATPase activity of Acanfhamoeba myosin I. J . Eiol. Chem. 252, 8329-8332. Maruta, H., and Korn, E. D. (1981a). Direct photoaffinity labeling by nucleotides of the apparent catalytic site on the heavy chain of smooth muscle and Acanfhumoeba myosins. J. Eiol. Chem. 256, 499-502. Maruta, H., and Korn, E. (1981b). Proteolytic separation of the actin-activatable ATPase from the phosphorylation site on the heavy chain ofAcanfhamoebamyosin IA. J . Eiol. Chem. 256,503506. Maruta, H., Gadasi, H., Collins, J. H., and Korn, E. D. (1978). The isolated heavy chain of an Acanthamoeba myosin contains full enzymatic activity. J . Biol. Chem. 253, 6297-6300. Maruta, H., Gadasi, H., Collins, J. H., and Korn, E. D. (1979). Multiple forms ofAcanthamoeba myosin I. J . Eiol. Chem. 254, 3624-3630. Matsudaira, P., and Burgess, D. (1979). Identification and organization of the components in the isolated microvillus cytoskeleton. J . Cell Eiol. 83, 667-673. Miyata, H., Bowers, B., and Korn, E. D. (1989). Plasma membrane association of Acanrhamoebn myosin I. J . Cell Eiol. 109, 1519-1528. Montell, C., and Rubin, G . M. (1988). The Drosophila minaC locus encodes two photoreceptor cell specific proteins with domains homologous to protein kinases and the myosin heavy chain. Cell (Cambridge, Mass.) 52, 755-772. Mornet, D., Bonet, A., Audemard, E., and Bonicel, J. (1989). Functional sequences on the myosin head. J . Muscle Res. Cell Motil. 10, 10-24. Pollard, T. D., and Korn, E. D. (1973a). Acanrhamoeba myosin. 1. Isolation from Acanthamoeba castellanii of an enzyme similar to muscle myosin. J . Eiol. Chem. 248, 4682-4690. Pollard, T. D., and Korn, E. D. (1973b). Acanthamoeba myosin. 11. Interaction with actin and with a new cofactor protein required for actin activation of Mg2+ -adenosine triphosphatase activity. J. Eiol. Chem. 248, 4691-4697. Pollard, T. D., Shelton, E., Weihing, R. R., and Korn, E. D. (1970). Ultrastructural characterization of F-actin isolated from Acanrhamoeba castellanii and identification of cytoplasmic filaments as F-actin by reaction with rabbit heavy meromyosin. 1. Mol. Eiol. 50, 91-97. Pollard, T. D., Stafford, W. F., 111, and Porter, M. E. (1978). Characterization of a second myosin from Acanthamoeba casrellanii. J. Eiol. Chem. 253, 4798-4808.
30
Edward D.Kom
Rodaway, A. R. F., Sternberg, M. J. E., and Bentley, D. L. (1989). Similarity in membrane proteins. Nature (London) 342, 624. Strehler, E. E., Strehler-Page, M.-A., Perriard, J.-C., Periasamy, M., and Nadal-Ginard, B. (1986). Complete nucleotide and encoded amino acid sequence of a mammalian heavy chain gene. J . Mol. Biol. 190, 291-317. Titus, M. A , , Wanick, H. M., and Spudich, J. A. (1989). Multiple actin-based motor genes in Dictyostelium. Cell Regul. 1, 55-63. Weihing, R. R., and Korn, E. D. (1969). Ameba actin: The presence of 3-methylhistidine. Biochem. Biophys. Res. Commun. 35, 906-912. Weihing, R. R., and Korn, E. D, (1971). Acanthamoeba actin: Isolation and properties. Biochemistry 10, 590-600. Wessels, D., Soll, D. R., Knecht, D., Loomis, W. F., De Lozanne, A., and Spudich, J. A. (1988). Cell motility and chemotaxis in Dictyostelium amebae lacking myosin heavy chain. Dev. B i d . 1, 164-177.
CHAPTER 3
Structural a n d Functional Dissection of a Membrane-Bound Mechanoenzyme: Brush Border Myosin I Mark S. Mooseker,* Joseph S. Wolenski,* Thomas R. Coleman,*J Steven M. Hayden,**2Richard E. Cheney,” Enilza Espreafico,*.t Matthew B. Heintzelman,* and Michelle D. Peterson* *Department of Biology Yale University New Haven, Connecticut 06510 tDepartmento de Biochiumica de Faculdade de Medicina de Ribeirao Preto Universidade de Sio Paulo 05508 Sio Paulo, Brazil
I. Introduction: Is there a Superfamily of Myosin Genes? 11. The Structural and Functional Properties of Brush Border Myosin 1: An Overview A. Subunit Composition and Domain Structure B. Actin Binding Properties C. ATPase Properties of Brush Border Myosin I D. Mechanochemistry of Brush Border Myosin I 111. Probing the Function of Brush Border Myosin 1 Calmodulin Light Chains A. The Role of Calmodulin Light Chains as Repressors Rather than Activators of Brush Border Myosin I Mg2 -ATPase B. Identification of Calmodulin-Binding “Neck” Domains in Other Unconventional Myosins IV. The Interaction of Brush Border Myosin I with the Microvillar Membrane A. Interaction of Brush Border Myosin I with Acidic Phospholipids B. Evidence for a Microvillar Membrane “Docking” Protein for Brush Border Myosin I V. Some Notions Regarding Brush Border Myosin I Function References +
‘Present address: Division of Biology, California Institute of Technology, Pasadena, California 91125. 2Present address: University of Colorado Health Sciences Center, Denver, Colorado 80262
Current Topics in Membranes, Volume 38 Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.
31
32 1. INTRODUCTION:
Mark S. Mooseker et al.
IS THERE A SUPERFAMILY OF M Y O S I N GENES?
A sense of excitement is growing in the cytoskeleton field regarding molecular motors. It seems, based on recent studies, that the array of mechanoenzymes for both microtubule- and actin-based motility may be far more diverse than once expected (see Chapter 1, by Goldstein and co-workers, for a discussion of microtubule-based motors). In the case of actin-based movements, the idea that a given cell may contain multiple myosins which perform distinct functions has been firmly established for two types of amoeboid cells, Acanthamoeba and Dicqostelium (see Chapter 2 , by Korn and co-workers). These cells contain a conventional two-headed and -tailed myosin (dubbed myosin 11; see Kom and Hammer, 1988). In addition, both of these amoeboid cells express multiple myosins which are structurally unconventional. To date, most of these “unconventional” myosins appear to belong to a single structural class of myosins termed myosin 1 (for recent reviews on amoeboid myosins I, see Korn and Hammer, 1990; Pollard er a l . , 1991). Based on information gleaned from the amoeboid myosins I that have been both sequenced and characterized as proteins, this class of actin-based motors has a single myosin head domain linked to a carboxy-terminal tail domain which is non-a-helical and segmented into distinct functional domains. For Acanthamoeba myosins I, these segments include a second actin-binding site that is ATP insensitive and a site for interaction with phospholipids (and presumably the plasma membrane; reviewed in Korn and Hammer, 1990; Pollard et a l . , 1991). The potential importance of the myosin I family has been underscored by a stunning series of experiments in which the phenotype of Dictyostelium amoebae was molecularly engineered to lack functional myosin I1 (reviewed in Spudich, 1989). From these studies it is clear that only a subset of the actin-based motile/contractile behaviors of the cells is compromised, notably, cytokinesis and generation of cortical tension (Spudich, 1989). Other key movements, such as phagocytosis and cell locomotion, persist in these cells. The obvious conclusion from these studies is that such phenomena are the venue of the other actinbased motors expressed in these cells, such as the myosins I. Not surprisingly, the above studies have generated considerable interest and effort to determine whether or not the expression of unconventional myosins is a ubiquitous feature of eukaryotic cells. The answer, based on a limited, but phylogenetically diverse, data set is yes (for review, see Pollard et a l . , 1991). Perhaps the most unusual example of an unconventional myosin described thus far is the pair of proteins encoded by the ninaC gene of Drosophila (Monte11 and Rubin, 1988). These proteins resemble myosin 1 in that they contain a myosin head domain and a non-a-helical tail domain which lies C-terminal to the head. However, these presumed myosins also have a “nose” domain, on the N-terminal side of the head; its sequence suggests that it is a kinase. Another example of
3. Brush Border Myosin I
33
a myosin that structurally does not fall into either the myosin I or I1 category is the presumed actin-based motor encoded by the M Y 0 2 gene of yeast. This unconventional myosin has a large tail domain that is predominantly non-ahelical, but it does contain a segment of a helix (with heptad repeats) that most likely would promote oligomerization (G. Johnston, Dalhousie University, personal communication; see also Section 111, below). This same structural organization is shared by a possible vertebrate member of the my02 family of myosins, p190, a calmodulin-binding protein in vertebrate brain (see Section 111). Given these few examples, it seems certain that our knowledge of the breadth of the myosin gene family will continue to enlarge. To date, however, only one example of a vertebrate myosin I has been thoroughly characterized. This myosin, termed brush border (BB) myosin I, is expressed within the microvilli of the intestinal epithelial cell. It is this myosin I, with its apparently unusual compliment of multiple calmodulin (CM) light chains, that is the focus of the research summary which follows.
II. THE STRUCTURAL A N D FUNCTIONAL PROPERTIES O F BRUSH BORDER MYOSIN I: AN OVERVIEW The actin bundle within the intestinal microvillus (MV) is tethered laterally to the plasma membrane by spirally arranged bridges (reviewed in Mooseker, 1985; see Fig. 1 for a schematic overview of the molecular constituents of the BB cytoskeleton). A series of studies (for reference, see Coluccio and Bretscher, 1989; Mooseker, 1985) has demonstrated that these bridges are composed of a protein complex consisting of a heavy chain of an apparent molecular mass of 110 kDa and multiple associated CM light chains (formerly termed 110 kDacalmodulin complex). The culmination of these studies was the reconstitution of bridges by addition of BB myosin I to synthetic MV actin bundles constructed from purified actin and the core proteins villin and fimbrin (Coluccio and Bretscher, 1989). Prompted by the initial studies of Collins and Borysenko (1984), who demonstrated that the 1 IOK-CM complex exhibited myosin-like ethylenediaminetetraacetic acid (EDTA) ATPase activity, several laboratories have now provided ample evidence that this complex is a single-headed myosin-the operational definition of a myosin I (see Korn and Hammer, 1988).
A. Subunit Composition and Domain Structure
1. Subunit Composition of Brush Border Myosin I The initial methods for isolation of BB myosin I (Howe and Mooseker, 1983; Collins and Borysenko, 1984; Conzelman and Mooseker, 1987) produced prepa-
34
Mark S . Mooseker et al.
FIG. 1 Schematic model for the brush border (BB) cytoskeleton. Note that the two classes of myosin present in the BB, myosin I and conventional myosin, are found in separate compartments of this actin-based cytoskeleton. BB myosin I interacts with the uniformly polarized array of filaments within the microvillus, while conventional myosin is confined to the terminal web region, where it can interact with filaments of opposing polarity (between filaments of adjacent microvillar core rootlets and within the actin bundle of the zonula adherens junction; not shown). This modality of myosin sorting may be a general feature for the distribution of these two classes of myosins [e.g., see Fukui er al. (1989) for the distribution of myosin I and I1 in Dicryosreliurn].
rations which we now know were depleted in CM content, with ratios of BB myosin I heavy chain : CM in the range of 1 : < 1-2. The most recent purification protocols for BB myosin I (Coluccio and Bretscher, 1987; Collins et al., 1990; Hayden et al., 1990) yield preparations with much higher CM content (heavy chain : CM of 1 : 3-4). Unlike most CM-binding proteins, the association of CM with the BB myosin I heavy chain appears to be most stable at low [Ca2+], since partial dissociation of CM light chains occurs in the presence of Ca2+ (Collins et al., 1990).
2. Protein Domain Structure of Brush Border Myosin I Electron microscopy of rotarj-shadowed preparations of BB myosin I (performed by J. Heuser, Washington University; see Conzelman and Mooseker, 1987) reveals this complex to be a tadpole-shaped molecule -30 nm in length; the presumed tail domain is substantially thicker than that of conventional myosin viewed using the same technique (8 versus 2 nm).This view of the molecule is entirely consistent with the domain structure of BB myosin I as determined by
35
3. Brush Border Myosin I
analysis of products generated by limited digestion with either trypsin (Carboni et af., 1988) or chymotrypsin (Coluccio and Bretscher, 1988). These initial studies demonstrated that the molecule consists of an N-terminal, myosin-like head domain (based on presence of an ATP-binding site, myosin head epitopes, actin binding, and enzymatic properties) linked to a tail domain (subsequently shown to be on the C-terminal side of the head; Garcia et af., 1989) which contains the binding sites for CM light chains (see Fig. 2). The products of chymotryptic digestion are dependent on the [Ca*+], indicating that this ion affects the conformation of the molecule (Coluccio and Bretscher, 1987). In the absence of Ca2 , a 90-kDa fragment, lacking the most C-terminal portion of its tail (see Hayden ef af., 1990) is generated which retains bound CM (Coluccio and Bretscher, 1988), while a shorter, CM-free head fragment of 78 kDa is generated in the presence of Ca2 . As noted below, these fragments have been quite useful in assessing the contribution of the tail domain, with and without its CM light chains, to the various functional properties of BB myosin I. +
+
3. Primary Structure of Brush Border Myosin I Heavy Chain In 1987, the primary structure of the first putative vertebrate myosin I was obtained by Hoshimaru and Nakanishi (1987). Termed bovine myosin I heavy chain gene, the product of this gene turned out to be the bovine form of BB myosin I heavy chain as determined by comparison with the deduced sequence of the heavy chain of chicken BB myosin I (Garcia er al., 1989; see also Hoshimaru et af., 1989). The primary structure of the heavy chain (bovine and avian) predicts a protein of 119 kDa, with a domain structure consistent with that determined from the protein studies above. That is, BB myosin I heavy chain consists of an N-terminal domain of -720 amino acids that encodes a myosin head domain, with highest degree of similarity to the head domains of amoeboid
-
FIG. 2 A model of BB myosin I depicting its subunit composition and functional domain structure. See text for details.
36
Mark S . Mooseker et al.
myosins I. Unfortunately, the sequence of the first -40 amino acids of the avian protein is still undetermined. The most striking feature of the heavy chain to be predicted from its sequence is the presence of several possible CM-binding domains just distal to the head domain (Hayden et al., 1990; Hoshimaru et al., 1989; see Figs. 3 and 4). The position of these sites is completely consistent with the mapping of the CM domains either by proteolysis (Coluccio and Bretscher, 1988; Carboni et al., 1988) or by CM-binding studies on bacterially expressed fusion proteins encoded by avian BB myosin I heavy chain cDNA (Garcia et al., 1989). In our view, it seems useful to refer to the CM-binding domains as a “neck” region of the molecule, since these sites are immediately adjacent to the head and, in fact, are in a position analogous to the region of the head domain of conventional myosin implicated in light chain interaction (Vibert and Cohen, 1988; Mitchell et al., 1989). Thus, the positioning of BB myosin I light chains may well be functionally equivalent to that for light chain-heavy chain contact sites in conventional myosins. More recently, Halsall and Hammer (1990) have described a splice variant of BB myosin I heavy chain that contains a 29-amino acid insert encoding yet another CM-binding domain (see Figs. 3 and 4). The segmentation of the heavy chain into functional domains is underscored by marked differences in the charge distribution among the three large domains discussed above. The overall calculated isoelectric point for BB myosin I heavy chain is PI -9.8; however, the myosin head domain has an isoelectric point of
700
800
I
I
716
717
900
1000
FIG. 3 The “neck” region of BB myosin I contains multiple potential CM-binding domains. This includes the expression of an apparently minor splice variant of BB myosin I that contains a 29amino acid insert between residues 716 and 717 encoding an additional CM-binding domain (Halsall and Hammer, 1990). The shaded regions represent possible CM-binding domains, the largest of which (striped domain) shares sequence homology with the CM-binding domain of erythrocyte membrane Ca* + -ATPase (Hayden er al., 1990).
37
3. Brush Border Myosin I calmodulin binding region from bovine neuromodulin (29-52) Dictyostelium myosin I1 (763-808) putative calmodulin binding region from chicken brain p190 (763-907)
yeast MY02 unconventional myosin (783-926)
DKAHKAATKIQASFRGHITRKLK
SEIIKAIQAATRGWIARKVYKQAREH TVAARIIQQLNRAYIDFKSW
DKLRAACIRIQKTIRGWLMRKYM PNRRAAITIQRYVRGHQARCYATFL RRTRAAIIIQKFQRMYVVRKRYQC MRDATIALQALLRGYLVRNKYQMML REHKSIIIQKHVRGWLARVHYH RTLKAIVYLQCCYRRMMAKRELKKL KMHNSIVMIQKKIRAKYYRKQYLQ ISQAIKYLQNNIKGFIIRQRVNDE MKVNCATLLQAAYRGHSIRANVFSVL
RTITNLQKKIRKELKQRQLKQE HEYNAAVTIQSKVRTFEPRSRFL RTKKDTVWQSLIRRRAAQRKLKQL
RVAELATLIQKMFRGWCCRKRYQL chicken brush border MRKSQILISAWFRGHMQRNRYKQ myosin I (654-716) plus a MKRSVLLLQAYARGWK fourth repeat from (SRRLLRELKVQRRRHLAASTISAYWKGYQ) an alternatively spliced form FIG. 4 Comparison of the putative CM-binding neck domains in the unconventional myosin of yeast, MY02 (G. Johnston, personal communication); chick p190, a myosin-like CM-binding protein in vertebrate brain (Espreafico CI d . , 1990); and BB myosin 1. The CM-binding domain of neuromodulin (GAP 43; P-57) is shown for comparison (Watkim et a / . . 1987). Note that there is a segment of similar sequence at the head-tail junction of conventional myosins such as Dicfyosreliurn myosin 11. These sequences were manually aligned with respect to conserved residues within each repeat; these are shown in bold.
only pi 6.3. The tail domain containing the CM-binding domains is extremely basic, with an isoelectric point of PI 12.4. Presumably, much of this charge is masked in the native molecule by the bound CM light chains. The C-terminal segment of the tail is also basic, with an isoelectric point of 9.8. Additional structural features of the region of the tail domain distal to the CM-binding neck region are discussed below in Section IV,A.
B. Actin Binding Properties
1. Stoichiometry and CaZ+Dependence Like other myosins, BB myosin I binds to F-actin with high affinity in the absence but not in the presence of ATP (Howe and Mooseker, 1983; Collins and Borysenko, 1984; Conzelman and Mooseker, 1987; Coluccio and Bretscher, 1987). This interaction results in the formation of “decorated” filaments similar in morphology to that observed for the binding of conventional myosin head
38
Mark S . Mooseker et al.
fragments (HMM or S1) to F-actin (Coluccio and Bretscher, 1987; see Fig. 5). However, important differences in the stoichiometry and CaZ dependence of BB myosin I-actin interaction have been reported, most probably due to intrinsic digerences in the preparations characterized; the molecular basis for most of these differences remains undetermined. For example, Coluccio and Bretscher +
FIG. 5 Actin filament cross-linking by BB myosin I. This is a negatively stained preparation of F-actin-BB myosin I combined at saturating ratios ( I : 1). Bar, 200 nm.
39
3. Brush Border Myosin I
(1987) observed that BB myosin I-actin interaction was cooperative in the absence but not the presence of Ca2 , and that this interaction was of relatively low affinity. In contrast, the most recent preparations of BB myosin I obtained in our laboratory, using essentially identical purification methods, exhibit highaffinity, noncooperative binding to actin that saturates at ratios of actin : BB myosin I of 1 : 1 in the absence of Ca2 and 1 : 1.25 in the presence of 10 pJ4 Ca2 (Wolenski er al., 1990). It is likely that the slightly higher ratio obtained in the presence of Ca2+ is the result of partial CM dissociation, which causes artifactual aggregation of BB myosin I. This explanation is based on the observation that our initial studies on BB myosin I preparations containing only -2 CM light chaindheavy chain exhibited actin : BB myosin I binding ratios, at saturation, of 1 :4-5 (Conzelman and Mooseker, 1987). +
+
+
2. Actin Filament Cross-Linking by Brush Border Myosin I In our early studies on BB myosin I, we observed that addition of BB myosin I to F-actin in the absence of ATP induces the formation of cross-linked arrays of filaments, as assayed by cosedimentation, and light and electron microscopy (Howe and Mooseker, 1983; Conzelman and Mooseker, 1987). As mentioned above, this cross-linking was probably in part due to artifactual aggregation of these CM-depleted preparations. However, our most recent, CM-replete preparations of BB myosin I, which show no evidence of aggregation, still cross-link actin filaments, even at saturating ratios of actin : BB myosin I (Hayden et a l . , 1990; Wolenski er a l . , 1990; see Fig. 5). We have recently demonstrated that the removal of the distal portion of the tail domain by chymotryptic digestion results in a loss of cross-linking activity (Wolenski et al., 1990). This observation is consistent with the notion that BB myosin I may be analogous to Acanthamoeba myosins I, which contain a second, ATP-independent actin-binding site on their tail domains (Korn and Hammer, 1988, 1990). Several lines of evidence suggest that this is nor the case. First, BB myosin I heavy chain does not contain a domain similar in sequence to the actin-binding site on the tail of Acanthamoeba myosin I. Second, unlike Acanthamoeba myosin I, the Mg2+-ATPase of BB myosin I is not “superactivated” at high ratios of myosin I : actin (Conzelman and Mooseker, 1987; Collins et a l . , 1990; Wolenski er al., 1990). This superactivation is a diagnostic property of myosins I containing an ATP-insensitive actinbinding site on their tails (reviewed in Korn and Hammer, 1988). Third, we can detect no binding of BB myosin I to actin in the presence of ATP, even at high concentrations of BB myosin I (J. Wolenski, unpublished observation). Thus, if a second site is present, it is either of very low affinity or, like the site on the head, is ATP sensitive. Finally, we have been unable to isolate a tail fragment from proteolytic digests of BB myosin I that exhibits actin-binding properties (J. Wolenski, unpublished observation). Our current working model for how BB myosin I effects filament cross-linking is that actin binding induces the self-
Mark S. Mooseker et al.
40
association of bound BB myosins I by their tail domains. This in vitro association may not be relevant to the function of BB myosin I in vivo, since within the MV the tail domain of this myosin is presumably affixed to the membrane (see Section IV). On the other hand, BB myosin I is also expressed in non-MV regions of the enterocyte where tail-tail interactions might come into play. During enterocyte differentiation, BB myosin 1 exhibits predominantly a cytoplasmic distribution (reviewed in Heintzelman and Mooseker, 1991). Moreover, BB myosin I is not exclusively localized to the MV in the differentiated cell. Low levels of immunostaining have been observed in association with the lateral membrane (Heintzelman and Mooseker, 1990) and within the terminal web region (Drenckhahn and Dermietzel, 1988; Heintzelman and Mooseker, 1990).
C. ATPase Properties of Brush Border Myosin I
Like other myosins, BB myosin I exhibits its highest ATPase activity under nonphysiological conditions that displace Mg2 from the protein. This includes high levels of activity in the presence of high [Ca2+]and in the presence of KEDTA (Collins and Borysenko, 1984; Conzelman and Mooseker, 1987; Krizek et al., 1987; Swanljung-Collins et af., 1987; see Table I). Conversely, BB myosin I exhibits low levels of activity in the presence of Mg2+. In buffers containing 50- 100 mM added KCI, the Mg2 -ATPase of BB myosin I is only modestly activated by actin (up to 3- to 5-fold) in an actin-concentration-dependentfashion (Conzelman and Mooseker, 1987; Krizek et al., 1987; Mooseker and Coleman, 1989; see Table I). Recently, Collins et al. (1990) have reported significantly higher levels of actin activation (up to -40-fold); these higher levels are the result of using lower ionic strength assay conditions (Wolenski et al., 1990; see Table I). The Mg2 -ATPase of BB myosin I is also activated, independently of actin, by Ca2+ in the 5-10 p,M range (Conzelman and Mooseker, 1987; Mooseker and Coleman, 1989; Wolenski et al., 1990; Collins et a!., 1990). Experiments from our laboratory to investigate the role of the CM light chains in this activation by Ca2+ are discussed in Section 111. +
+
+
D. Mechanochemistry of Brush Border Myosin I
As is summarized below, there is now ample in vitro evidence demonstrating that BB myosin 1 is an active mechanoenzyme. However, very important differences in both motility rate and Ca2+ sensitivity for motility have been reported for different preparations of BB myosin I (see Fig. 6). Fortunately, the rate differences can now be attributed to assay method. We have recently demon-
41
3. Brush Border Myosin I
TABLE 1 ATPase Activity of BB Myosin I Activity" (nmol PJmg min)
Mg'+ - actin
Mg22
pCa'+ 6 5
pCa2+
Ref.
K-EDTA
Ca*+
>9
6
5
3
+ actin
>9
3
v,,
K", (FM)
M g 2 + - A T P a s e assays with buffers containing >40 mM KCI
I. 2. 3.
-
-
30
-
-
160 500
400
180
-
90
-
-
3
4. 5. 10.
-330 -
40 50 10 20 10 4 0 -
-
-
-
130 - -
-
6.
-
7.
8.
170 480
610
9.
-
-
10.
-
-
-
360 -380 -
-
-
-
-
100
250 - -
-
0
-
9 0 -
-
-
280
-
1 m - - -
M g 2 + - A T P a s e assays with buffers containing < I 0 rnM KCI
170
20 - 30 - -5 2 0 - - 20 50 300
150 -
-
80 3 0 - - 2 0 - - 290 470 350 700
260
-
570
48
1 . Howe and Mooseker (1983). 2. Conzelman and Mooseker (1987). 3. Krizek e r a / . (1987). 4 . Coluccio and Bretscher (1988). 5. Mooseker and Coleman (1989). 6. Collins er a/. (1990). 7. Collins and Borysenko (1984). 8. Swanljung-Collins er a / . (1987). 9. Hayden er a / . (1990). 10. Wolenski PI a / . (1991). '1 The values reported for Mg2 -ATPase activity ( + actin) are based on assays (ref. nos. in parentheses) using the following concentrations of actin: 0.5 mg/ml (7,8). 0.9 mgiml ( 2 , 5 ) . I mgiml (9, lo), and 0.86 mgiml (6). The corresponding value for reference (3) was obtained from a graph of activity versus [actin] at I mgirnl actin. The effect of increasing concentrations of actin on BB myosin I Mgz+-ATPase has been assayed by three different laboratories using four different preparations (2, 3. 6, 10). Maximum reported actin-activatable MgZ+-ATPdse values in nmol PJmg min are 160 using 52 phf aclin (2). 40 using 4 pkf actin (3). 226 using 80 phf actin (6), and 350 using 40 phf actin (10). In each instance, the levels of Mgz+-ATPaseactivity in low concentrations of F-actin (1-10 p M ) were comparable. However, the BB myosin 1 preparations of Conzelinan and Mooseker (1987) exhibited significantly higher (40-50 nmol PJmg min) basal levels of Mgz+-ATPdsein the absence of F-actin than the preparations of Collins er a / . (1990) and Wolenski el a / . ( I Y Y I ) , who reported basal activities of -4 and 10-20 nmol P,/mg min, respectively. Such differences in basal activities can partly explain the wide variations [2 lo 3-fold (2) versus 30 to 40-fold (6, lo)] in actin-activatable Mgz+-ATPase reported by these investigators.
42
Mark S. Mooseker et al.
Nitrocellulosecoated coverslip
FIG. 6 In v i m motility of various preparations of BB myosin I. (A) Rates and calcium sensitivities of movements for four different preparations of BB myosin I assayed using the Nitella bead movement assay. Except for the MV disk preparation, note that similar rates but different calcium requirements were observed for these different preparations. See text for details. (B) Rates and calcium sensitivities for movement determined by the gliding filament assay of Kron and Spudich (1986). Depicted are the results of Collins et al. (1990) and Wolenski et al. (1991). The latter results were obtained using preparations that exhibited substantially slower movements when assayed by the Nitella assay (see A). Note that, unlike the original preparations of BB myosin 1 (Mooseker and Coleman, 1989). movement was observed over a broad range of [Ca*+]. The basis for these differences in calcium requirements among preparations has not been determined.
strated that the Nitellu bead movement assay (Sheetz et al., 1986) yields much slower rates of movement than the gliding actin filament assay of Kron and Spudich (1986). On the other hand, the variability in Ca2+ requirements must be due to as yet uncharacterized differences among the various preparations of BB myosin I used for these studies. Using the Nitella bead movement assay, we demonstrated that BB myosin I is a mechanoenzyme with the same directionality of movement determined for other myosins (pointed to barbed end; Mooseker and Coleman, 1989). In this initial study, very slow rates of bead movement were observed (-9 nm/sec).
43
3. Brush Border Myosin I
This preparation of BB myosin I promoted bead movement in calcium/EGTA buffers containing 10 p V free Ca2 (Portzehl et al., 1964); no movement was observed in the same buffer containing only EGTA. In the course of these experiments, we found that the requirement for Ca2 was lost if we coupled the BB myosin I to the bead using a monoconal antibody (mAb CX-7) reactive with the C-terminal end of the tail domain of BB myosin I (see Hayden et al., 1990; Garcia er al., 1989, for characterization of this mAb). Under these conditions, the same slow rate of movement was observed in both the absence and presence of Ca2+ (see Fig. 7). It should be noted that saturating amounts of the CX-7 had no effect on the actin-activated Mg2+-ATPase of BB myosin I in either the absence or presence of Ca2+ (T. Coleman, unpublished observation). The simplest explanation for the effect of the CX-7 mAb is that it somehow mimics the effect(s) of Ca2 on BB myosin I. Such effects of Ca2 that resulted in a motile preparation might include the following: (1) conformational activation of the motor, (2) inactivation of a second actin-binding site on the tail domain that blocks movement in the absence of Ca2 , and (3) dispersal of aggregates of BB myosin I that block motility. Unfortunately, we have been unable to determine the basis for mAb CX-7 action since, as discussed below, our most recent preparations of BB myosin I exhibit exactly the opposite Ca2+ sensitivity, even though enzymatically these preparations are comparable. Using the gliding actin filament assay of Kron and Spudich, Collins et al. (1990) have observed significantly faster rates of motility for BB myosin I (20 nm/sec at 22°C up to 90 nmisec at 37°C). In this assay, movement of single actin filaments over coverslips coated with a lawn of myosin is observed. Most importantly, these workers observed movement in buffers containing a broad range of [Ca2+] from 5-10 pA4 Ca2+, and, like the studies of Collins et al. (1990), this inhibition could be partially restored by addition of exogenous Ca2+. Interestingly, although movement was observed over a broad range of
-
+
+
+
+
+
44
Mark S . Mooseker el a / .
0
5
lo
15
20
25
30
Velocity (nrnlsec)
u-
0
L Q)
P
E
3
z
0
1 0
5
lo
15
20
25
30
Velocity (nrnlsec)
FIG. 7 A tail-binding monoclonal antibody (mAb CX-7) changes the calcium sensitivity but not rate of motility of a BB myosin I preparation that requires calcium for motility. Shown are histograms of velocities for beads coated with mAb CX-7-BB myosin I complexes in the absence (A) and
45
3. Brush Border Myosin I
[Ca2+], the fastest rates of movement (up to 70-80 nm/sec) were observed in the presence of 1-5 pM Ca2 , exactly the threshold concentration for activation of BB myosin I Mg2 -ATPase (see Section 111). Obviously, there is much work to be done to dissect the molecular basis for the variability we have observed regarding the role of Ca2iin regulating the mechanochemical activity of BB myosin I. This is a particularly troubling puzzle since essentially the same isolation methods were used to obtain these preparations (a modification of the protocol of Coluccio and Bretscher, 1987; see Hayden et al., 1990). Important variables to consider systematically include the following: ( 1 ) CM light chain content, (2) ratio of BB myosin I isoforms with and without the CM splice insert (see Halsall and Hammer, 1990), (3) phosphorylation state of BB myosin I heavy chain as the isolated BB contains kinase activity which phosphorylates residue(s) on the C-terminal end of the tail domain (J. S. Wolenski and M. S. Mooseker, unpublished observations), and (4) presence of trace but highly “active” contaminants, e.g., kinases, Ca2 -sensitive actinbinding proteins. This latter variable seems an unlikely candidate to explain the difference in Ca2+ sensitivity among BB myosin I preparations given the effect of the CX-7 mAb on our earlier preparations.
-
+
+
+
111. PROBING THE FUNCTION OF BRUSH BORDER MYOSIN I CALMODULIN LIGHT CHAINS
A number of potential activities associated with BB myosin I may be regulated by Ca2+ binding to its CM light chains. In our view, these can be grouped into two categories, only one of which may be relevant in vivo. The first category of Ca2 -dependent effects are those that may be effected by the CM light chains in the bound state. These might include the Ca2 activation of BB myosin I Mg2 ATPase (see Section II1,A) and the apparent increase in velocity of this mechanoenzyme observed at threshold levels of Ca2 . The second category, which we +
+
+
+
presence (B) of calcium (-10 p M ) . Note that comparable rates of movement were observed under both conditions (average rate, -8-10 nmlsec). These studies were conducted using preparations of BB myosin I that exhibited movements of the same velocity but only in the presence of buffers containing 10 pA4 Ca2+ (see Mooseker and Coleman, 1989). The MgZ+-ATPaseof BB myosin I (2 actin; Ca2+) in the presence of saturating concentrations of CX-7 was identical to that determined (Mooseker and Coleman, 1989) for BB myosin I (for those values see Table I). Methods: BB myosin I was isolated as described in Mooseker and Coleman (1989). The CX-7 mAb, which interacts with the C-terminal end of the BB myosin I tail (see text for refs.) was purified by anti-mouse IgG affinity chromatography. Mixtures of BB myosin I(0.5 mglml) and CX-7 (0.8 mglml) were incubated at room temperature for 10-20 min prior to adsorption onto covaspheres using identical buffer conditions to those in Mooseker and Coleman (1989). Note: The effect of CX-7 on our more recent preparations of BB myosin I that are inhibited by elevated Ca2+ (Wolenski et a / . , 1991) has not yet been determined.
*
46
Mark S . Mooseker et al.
feel may not be relevant in vivo, are those aspects of BB myosin I activity (assayed in vim) that result from Ca2 -dependent dissociation of CM light chains. The studies of Collins et al. (1990) clearly demonstrate that elevated Ca2 results in partial loss of CM from the complex. Numerous properties of BB myosin I appear to change when its CM content is reduced. As noted above, CMdepleted BB myosin I appears to aggregate. The actin-activated Mg2 -ATPase of BB myosin I low in CM content is lost in the presence of 10 pM Ca2+, a loss that can be blocked by adding exogenous CM (Conzelman and Mooseker, 1987). Similarly, as noted above, elevated Ca2 results in a loss of mechanochemical activity in vitro that can also be at least partially prevented by inclusion of exogenous CM (Collins et al., 1990; Wolenski et al., 1991). Taken together, such studies establish that variations in CM content can have a profound effect on the properties of this myosin. They also demonstrate that the relative affinity of the heavy chain for one or more of its light chains is reduced when those light chains bind Ca2+. On the other hand, it seems unlikely that the complete dissociation of one or more of its CM light chains in response to elevated intramicrovillar Ca2+ would actually occur in vivo. The concentration of CM within the MV is substantially higher (in the 1-10 mM range; Howe and Mooseker, 1983; Glenney and Glenney, 1985) than the concentration of exogenously added CM required in the above in vitro studies to counteract the effects of added Ca 2 + . Although time-consuming and expensive, it will be worthwhile to repeat many of the studies summarized above in the presence of high concentrations of exogenously added CM. +
+
+
+
A. The Role of Calmodulin Light Chains as Repressors Rather than Activators of Brush Border Myosin I Mgz+-ATPase
In order to determine whether the Ca2+ sensitivity of BB myosin I Mg2+ATPase is mediated by Ca2 binding to the head domain or to its tail-associated CM light chains, we have compared the effects of [Ca2+]on the actin-activated Mg2 -ATPase of the 78-kDa (CM-free) and 90-kDa (CM-retaining) chymotryptic peptides to that of intact BB myosin I (Wolenski et al., 1990; see Fig. 8). Beginning at -3-5 pM free Ca2 , a sharp increase in the Mg2+-ATPase of BB myosin I was observed, which is simply additive to that induced by actin. The same results were obtained using the CM-retaining 90-kDa peptide. In contrast, the tail-less, CM-free 78-kDa peptide exhibited high, actin-independent Mg2 ATPase activity that was insensitive to the [Ca2+]; this activity was further activated by actin, also in a Ca2 -insensitive fashion (Fig. 8). Thus, the tail-less, CM-free head domain of BB myosin I appears to be enzymatically equivalent to the intact protein in the presence of 10 pA4 Ca2 . These results clearly establish a role of one or more of the CM light chains in regulating the Mg2 -ATPase of +
+
+
+
+
+
+
47
3. Brush Border Myosin I
FIG. 8 Calcium regulation of BB myosin I Mg2+ -ATPase is mediated by its CM light chain(s). Shown are histograms of the Mg2+-ATPase activities for intact BB myosin I and the 90-kDa (CMretaining) and 78-kDa (CM-free) peptides. Activities in the absence and presence of F-actin, at low calcium (EGTA alone), and 10 phi Ca2+ are shown. Data are taken from Wolenski et al. (1990, 1991).
BB myosin I. The likely mode of Ca2+ regulation by the CM light chains revealed by these studies is unusual and exactly opposite of what we had expected from the action of other myosin light chains. Based on the high, Ca2+-insensitive Mg2+-ATPase of the CM-free 78-kDa head fragment, it appears that, in the absence of Ca2+, the CM light chains serve to suppress this activity. Binding of Ca2+ by one or more of the CM light chains results in a loss of this suppression, without affecting the actin-activatible Mg2 -ATPase of BB myosin I which appears to be, for the most part, a Ca2+-insensitive attribute of this myosin. +
B. Identification of Calmodulin-Binding “Neck” Domains
in Other Unconventional Myosins Until we have a better understanding of the function of BB myosin I in vivo, we will remain in the dark regarding the true regulatory functions of its multiple CM light chains. It has seemed logical to assume that the presence of multiple CM light chains is indicative of some property of BB myosin I peculiar to its
48
Mark S. Mooseker et al.
function within the enterocyte. A favorite speculation in our laboratory has been that this motor with its associated CM light chains is somehow involved in vitamin D-dependent uptake of Ca2+ by the enterocyte (see Section V). On the other hand, there appear to be other unconventional myosins with analogous “neck” domains containing multiple CM-binding sites. One of these is the presumed unconventional myosin encoded by the M Y 0 2 gene of yeast. Recently characterized by G. Johnston and colleagues (G. Johnston, Dalhousie University, personal communication), this myosin contains a series of 5-6 repeats that are likely to be CM-binding domains. Like BB myosin I, these repeats lie just Cterminal to the head domain. There appears to be an unconventional myosin similar to the M Y 0 2 gene product in vertebrates as well. In collaboration with the laboratory of R. Larson (University of Sao Paulo, at Ribeirao Preto), we have obtained partial deduced sequence of p190 (see Larson et al., 1990; Espreafico et al., 1990), a myosin-like CM-binding protein in vertebrate brain. Sequence comparison reveals this protein to be similar in both head and tail domains to the MY02 protein of yeast. This includes the presence of multiple repeats encoding potential CM-binding sites (see Fig. 4; R. Cheney and E. Espreafico, unpublished observations). Thus, the one feature of BB myosin I we once thought unique to this mechanoenzyme, its multiple CM light chains, may well be a common feature of certain members of the myosin gene family. Consequently, the past and future studies to elucidate the role of CM light chains in BB myosin I function should be of general as well as specific impact. For example, are there cooperative interactions among the multiple CMs? Does each CM bind to the heavy chain with equivalent avidity and topography relative to the N- and Cterminal domains of the dumbbell-shaped CM molecule? Does each CM serve distinct regulatory functions? Is there significant CM exchange between BB myosin I heavy chain and other CM-binding proteins in the cell? Does the splice variant of BB myosin I, which contains an additional CM-binding site, differ in its properties and subcellular localization?
N. THE INTERACTION OF BRUSH BORDER M Y O S I N I W I T H THE MICROVILLAR MEMBRANE
Once we have dissected the molecular basis for the interaction of BB myosin I with the MV membrane, we will have made major strides in understanding its function(s) in vivo. As for any peripherally associated membrane protein, several modes of interaction must be considered. For example, some proteins can interact directly with phospholipids, while others are inserted into the bilayer via covalently linked hydrocarbon chains. The other obvious mode of interaction is through the presence of an integral membrane protein receptor which targets BB
3. Brush Border Myosin I
49
myosin I to the MV membrane. Based on the limited data set presently available, both modes of interaction must be considered.
A. Interaction of Brush Border Myosin I with Acidic Phospholipids
Studies by both Adams and Pollard (1989) and Miyata el a1. (1989) have demonstrated that purified Acanrhamoeba myosins I bind to plasma membrane vesicles and that this binding appears to be independent of endogenous membrane proteins. Moreover, Adams and Pollard (1989) could obtain similar highaffinity, saturable binding to liposomes containing acidic phospholipids. The domain responsible for phospholipid binding is most likely on the tail of Acanthamoeba myosin I, just C-terminal to the head-tail junction (reviewed in Pollard et al., 1991). In an independent series of studies (Hayden et al., 1990), we have demonstrated a similar interaction of BB myosin I with acidic phospholipids in vitro. In brief, we observed high-affinity, saturable binding to liposomes of varied acidic phospholipid content. No binding to neutral or basic phospholipids was observed. The region of the BB myosin I molecule involved in this interaction is located on the distal, C-terminal half of the tail (see Fig. 2). This was determined by a combination of proteolysis protection assays and the loss of phospholipid binding on removal of the C-terminal domain by chymotrypsin. In addition, we demonstrated that phospholipid-bound BB myosin I retains a number of key biological properties, including actin binding and actinactivated Mg2 -ATPase. These findings suggest that the topography of BB myosin I-vesicle interaction is equivalent to that in the MV, with the head domain of this motor free to interact with actin filaments. Obviously, such in vitro binding studies should be taken with a grain of salt. Like Acanthamoeba myosin I, the domain of the BB myosin I tail implicated in phospholipid binding is rich in basic amino acid residues whose actual contribution to tail function may have nothing to do with phospholipid binding. On the other hand, the presence of such basic residue-rich domains on the tails of myosins I seems to be a common feature among many of the unconventional myosins for which primary structure has been obtained (reviewed in Pollard et al., 1991). Indeed, the tail domain of BB myosin I heavy chain contains a -210amino acid region which has limited sequence homology to the presumed phospholipid-binding domain of Acanthumoeba myosins I (Jung et al., 1989; Hayden et al., 1990). A common viewpoint held among many workers in the field is that, while phospholipid binding may well drive the association of myosin 1 with the cytoplasmic surface of the membrane, specificity of localization must require additional input, e.g., by specified myosin I receptor proteins or additional components of the membrane-skeletal network. +
50
Mark S . Mooseker et al.
8. Evidence for a Microvillar Membrane “Docking”Protein for Brush Border Myosin I A number of years ago, Coudrier et al. (1983) identified a -140-kDa glycoprotein (generated by lumenal protease cleavage from a 200-kDa protein) in preparations of porcine MV which, unlike the bulk of MV membrane proteins, remained tightly associated with the cytoskeleton fraction after extraction with nonionic detergent. Using an overlay technique, these workers showed that this glycoprotein (GP-140) binds to the heavy chain of BB myosin I; no binding to other MV core proteins was observed. Taken together, these observations support the idea that GP-I40 is the linker protein which tethers BB myosin I to the MV membrane. In our own studies on BB myosin I-membrane interaction in the chicken BB, we have obtained evidence consistent with the above studies that also suggests the presence of a specific glycoprotein linker in the MV membrane. We have characterized an unusual detergent-resistant membrane fraction purified from ATP extracts of Triton-extracted MV that contains tightly bound BB myosin I (Mooseker et al., 1989). This fraction consists of small, circular sheets of membrane which we have termed MV membrane disks. In addition to BB myosin I, these MV disks contain a single glycoprotein band at -130-140 kDa that is present in stoichiometric quantities with BB myosin I. We have conducted an extensive characterization of the structural organization and functional properties of the BB myosin I present in these disks. Electron microscopy revealed that the disk-bound BB myosin I molecules are tethered to the membrane by their tails. Consistent with this conclusion is that the bound BB myosin I retains essentially all the activities associated with the free molecule, including actin binding and mechanochemical activity (see Fig. 6). It is worth noting that the MV disks move with a faster rate than purified BB myosin I (up to -80-100 nmisec using the Nitella assay). The critical question that remains is to determine if the “GP-140” band present in this preparation is indeed a specific receptor for BB myosin I. The problem lies in the fact that the majority of the numerous glycoproteins present in the avian MV membrane are all in approximately the same molecular weight range as the GP-140 present in the disks. The obvious concern we have is that the GP-140 band is simply a “residue” of various MV membrane proteins. Experiments currently in progress (e.g., protein sequence analysis and generation of mAbs to the GP- 140 protein) should resolve this issue in the near future. Assuming that the GP-140 is indeed a single protein, the next step will be to reconstitute this protein into liposomes and assess whether it promotes the interaction of BB myosin I with membranes of varied phospholipid content. Another tack will be to use GP-140 antibodies to compare its pattern of expression during BB assembly to that of BB myosin I. For example, do these two protein colocalize during the redistribution of BB myosin I that occurs late in enterocyte differentiation (see Fig. 9)?
3. Brush Border Myosin I
51
FIG. 9 Late assembly of BB myosin I into the BB during enterocyte differentiation in the adult chicken intestine. Light micrographs depicting the colocalization of F-actin (a) and BB myosin I (b) along the villus-crypt axis. (c) Phase-contrast image of this cryosection. The apical localization of Factin from the base of the crypt is also seen for villin and fimbrin (not shown). Note the sharp change in distribution of BB myosin I from a diffuse to an apical localization. Bar, 30 pm. Reprinted from Heintzelman and Mooseker (1990).
V. S O M E N O T I O N S REGARDING BRUSH BORDER M Y O S I N I FUN CTlON
There are obvious speculations one can pose for BB myosin I action within the enterocyte based on its presumed motor functions. BB myosin I might participate in the biogenesis and recycling of the apical membrane by transporting vesicles up through the terminal web. The polarity of the microvillar core filaments is appropriate to support such movements. Moreover, the late “delivery” of BB myosin I to the MV cytoskeleton during enterocyte differentiation (reviewed in Heintzelman and Mooseker, 1991) is certainly consistent with a role for BB myosin I in the delivery of newly synthesized membrane proteins to the apical domain. However, it is equally plausible that BB myosin I must passively await the late synthesis and apical insertion of its docking protein in the MV mem-
52
Mark S. Mooseker et al.
brane. In this scenario, apical delivery of vesicles to the BB membrane would not require an active role for BB myosin I. The above studies on BB assembly have provided a comfortable base upon which to speculate regarding BB myosin I function in enterocyte differentiation (e.g., see Shibayama et al., 1987; Heintzelman and Mooseker, 1990, 1991). However, the striking precision of the organization of BB myosin I within the MV seems to be more consistent with roles for this motor within the assembled, fully operational BB. One can imagine that BB myosin I could cause a relative downward, or perhaps rotational, movement of the core within the MV membrane. Similarly, BB myosin I could promote the bulk movement of membrane upward along the MV axis, perhaps as part of an active vesicular shedding mechanism. Such shedding of right-side-out MV vesicles into the gut lumen has been shown to be a normal aspect of enterocyte biology (e.g., see Misch et al., 1983). A variant of the above membrane movement would be the selective movement of BB myosin I-bound integral membrane proteins within the plane of the membrane (some of these notions are schematically summarized in Fig. 10). Such models, based on relative translocation events along the MV axis, are easy to imagine, but they seemingly accomplish little for the enterocyte. In our view, the most likely function for BB myosin I will be an involvement in the spec@c functions carried out by the enterocyte, i.e., nutrient and electrolyte absorptiodsecretion. For example, this motor may act as a mechanochemical regulator of membrane proteins involved in solute movement across the membrane. One
Move microvlllar core filaments
FIG. 10 Speculations for the function of BB myosin I within the intestinal microvillus. Arrows indicate direction of movement either of membrane relative to a stationary filament or vice versa. See text for details. IMR, Integral membrane proteins; X, any solute molecule moved across the BB membrane.
3. Brush Border Myosin I
53
obvious candidate transport system that might be regulated mechanochemically by BB myosin I is the movement of Ca2+ across the MV membrane (for review, see Wasserman and Fullmer, 1989). In this scheme, BB myosin I would affect the permeability of the putative vitamin D-dependent “channel” in the membrane through direct tension applied at the cytoplasmic side of the MV membrane. We must be vague regarding mechanism here because there is both disagreement and little information regarding the molecular basis for the inward movement of Ca2+ across the membrane (Wasserman and Fullmer, 1989). Nevertheless, if the GP-140 protein is indeed the specific docking protein for BB myosin I, its molecular and enzymatic properties should tell us whether such “mechanoregulation” models are worth exploring experimentally.
Acknowledgments The studies summarized here from the Mooseker laboratory were supported by NIH grant DK 25387.
References Adams, R. J., and Pollard, T. D. (1989). Membrane-bound myosin-I provides new mechanisms in cell motility. Cell Moril. Cyroskel. 14, 178-182. Carboni, J. M., Conzelman, K. A., Adams, R. A., Kaiser, D. A , , Pollard, T. D., and Mooseker, M. S. (1988). Structural and immunological characterization of the myosin-like 1 10-kD subunit of the intestinal microvillar I IOK-calmodulin complex: Evidence for discrete myosin head and calmodulin-binding domains. J. Cell Eiol. 107, 1749-1757. Collins, J. H., and Borysenko, C. W. (1984). The 110,000-dalton actin- and calmodulin-binding protein from intestinal brush-border is a myosin-like ATPase. J. Eiol. Chem. 259, 1412814135. Collins, K., Sellers, J. R., and Matsudaira, P. T. (1990). Calmodulin dissociation regulates brush border myosin-I (1 IOK-calmodulin) activity in vitro. J. Cell Eiol. 110, 1137-1 147. Coluccio, L. M., and Bretscher, A. (1987). Calcium-regulated cooperative binding of the microvillar 1 IOK-calmodulin complex to F-actin-Formation of decorated filaments. J. Cell Eiol. 105, 325-333. Coluccio, L. M., and Bretscher, A. (1988). Mapping of the microvillar 1 l0K-calmodulin complex: Calmodulin-associated or calmodulin-free fragments of the 1lOkD polypeptide bind F-actin and retain ATPase activity. 1. Cell Eiol. 106, 367-373. Coluccio, L. M., and Bretscher, A. (1989). Reassociation of microvillar core proteins: Making a microvillar core in vitro. J. Cell Eiol. 108, 495-502. Conzelman, K. A., and Mooseker, M. S. (1987). The 110-kD protein-calmodulin complex of the intestinal microvillus is an actin-activated MgATPase. J . Cell Eiol. 105, 3 13-324. Coudrier, E., Reggio, H . , and Louvard, D. (1983). Characterization of an integral membrane glycoprotein associated with microfilaments of pig intestinal microvilli. EMEO J. 2, 469-474. Drenckhahn, D., and Dermietzel, R. (1988). Organization of the actin filament cytoskelton in the intestinal brush border: A quantitative and qualitative immunoelectron microscopic study. J . Cell Eiol. 107, 1037-1048. Espreafico, E., Cheney, R., Spindola, F., Coelho, M., Pitta, D., Mooseker, M.,and Larson, R. (1990). Characterization of p190, a myosin like protein in vertebrate brain and other tissues. J . Cell Eiol. 111, 167a.
54
Mark S. Mooseker et al.
Fukui, Y., Lynch, T. J., Brzeska, H., and Korn, E. D. (1989). Myosin I is located at the leading edges of locomoting Dicfyosfeliumamoebae. Nature (London) 341, 328-33 1. Garcia, A,, Coudrier, E., Carboni, J., Anderson, J., Vandekerhove, I . , Mooseker, M., Louvard, D., and Arpin, M. (1989). Partial deduced sequence of the 1 10-kD-calmodulin complex of the avian intestinal microvillus shows that this mechanoenzyme is a member of the myosin I family. J. Cell Eiol. 109, 2895-2903. regulated events in the intestinal brush Glenney, J. R., and Glenney, P. ( I 985). Comparison of Ca border. J. Cell Eiol. 100, 754-763. Halsall, D. J., and Hammer, J. A., I11 (1990). A second isoform of chicken brush border myosin I contains a 29-residue inserted sequence that binds calmodulin. FEES Lett. 267, 126-130. Hayden, S. M., Wolenski, J. S., and Mooseker, M. S . (1990). Binding of brush border myosin I to phospholipid vesicles. J. Cell Eiol. 111, 443-45 1. Heintzelman, M. B., and Mooseker, M. S . (1990). Assembly of the brush border cytoskeleton: Changes in the distribution of microvillar core proteins during enterocyte differentiation in adult chicken intestine. Cell Motil. Cytoskel. 15, 12-22, Heintzelman, M. B., and Mooseker M. S. (1991). Assembly of the intestinal brush border cytoskeleton. In “The Cytoskeleton and Development (E. Bearer, ed.) (in press). Hoshimaru, M., and Nakanishi, S. (1987). Identification of a new type of mammalian myosin heavy chain by molecular cloning: Overlap of its mRNA with preprotachykinin B mRNA. J. Eiol. Chem. 262, 14625-14632. Hoshimaru, M., Fujio, Y., Sobue, K., Sugimoto, T., and Nakanishi, S . (1989). Immunochemical evidence that myosin-I heavy chain-like protein is identical to the 1 10-kilodalton brush-border protein. J. Eiochem. (Tokyo) 206, 445-459. Howe, C. L., and Mooseker, M. (1983). Characterization of the 1 10-kdalton actin-calmodulin-and membrane-binding protein from microvilli of intestinal epithelial cells. J. Cell Eiol. 97, 974985. Jung, G., Schmidt, C. J., and Hammer, J. A., 111 (1989). Myosin-I heavy chain genes of Acanthamoeba castellanii: Cloning of a second gene and evidence for the existence of a third isofonn. Gene 82, 269-280. Korn, E. D., and Hammer, J. A., 111 (1988). Myosins of nonmuscle cells. Annu. Rev. Eiophys. Chem. 17, 23-45. Korn, E. D., and Hammer, J. A. (1990). Small myosins. Curr. @in. Cell Eiol. 2, 57-61, Krizek, J., Coluccio, L. M., and Bretscher, A. (1987). ATPdse activity of the microvillar 1 IOkDa polypeptide-calmodulin complex is activated in Mg2+ and inhibited in K + EDTA by F-actin. FEES Left. 225, 269-272. Kron, S. I.,and Spudich, J. A. (1986). Fluorescent actin filaments move on myosin fixed to a glass surface. Proc. Nafl. Acad. Sci. U.S.A. 83, 6272-6276. Larson, R. E., Espinodola, F. S., and Espreafico, E. (1990). Calmodulin binding protein and calciumlcalmodulin regulated enzyme activities associate with brain actomyosin. J . Neurochem. 54, 1288-1294. Misch, P., Giebel, P., and Faust, R. (1983). Intestinal microvilli responses to feeding and fasting. Eur. J. Cell Eiol. 21, 264-279. Mitchell, E. J., Karn, J., Brown, D. M., Newman. A., Jakes, R., and Kendrick-Jones, J. (1989). Regulatory and essential light chain binding sites in myosin heavy chain subfragment-I mapped by site directed mutagenesis. J. Mol. Eiol. 208, 199-205. Miyata, H., Bowers, B., and Korn, E. (1989). Plasma membrane association of Acanthamoeba myosin I. J . Cell Eiol. 109, 1519-1528. Montell, C., and Rubin, G. (1988). The Drosophila ninaC locus encodes two photoreceptor cell specific proteins with domains homologous to protein kinases and the myosin heavy chain head. Cell (Cambridge, Mass.) 52, 757-772. +
+
3. Brush Border Myosin I
55
Mooseker, M. S. (1985). Organization, chemistry and assembly of the cytoskeletal apparatus of the intestinal brush border. Annu. Rev. Cell Eiol. 1, 261-293. Mooseker, M. S . , and Coleman, T. R. (1989). The 110-kD protein-calmodulin complex of the intestinal microvillus (brush border myosin I) is a mechanoenzyme. J . Cell Eiol. 108, 23952400. Mooseker, M . S . , Conzelman, K. A., Coleman, T. R., Heuser, J. E., and Sheetz, M. P. (1989). Characterization of intestinal microvillar membrane disks: Detergent-resistant membrane sheets enriched in associated brush border myosin I ( I 10K-calmodulin).J. Cell Eiol. 109, 1 153- 1161. Pollard, T. D . , Doberstein, S. K., and Zot, H. G. (1991). Myosin I . Annu. Rev. Physiol. 53, 653681. Portzehl, H., Caldwell, P. C., and Ruegg, J. C. (1964). The dependence of contraction and relaxation of muscle fibres from the crab Mura squinado. Eiochim. Eiophys. Acfu 79, 581-591. Sheetz, M. P., Block, S. M., and Spudich, J. A. (1986). Myosin movement in vitro: A quantitative assay using oriented actin cables for Nitella. In “Methods in Enzymology” (R.B. Vallee, ed.), Vol. 134, pp. 531-543. Academic Press, Orlando, Florida. Shibayama, T., Carboni, J., and Mooseker, M. S. (1987). Assembly of the intestinal brush border: Appearance and redistribution of microvillar core proteins in developing chick enterocytes. J. Cell Biol. 105, 335-344. Spudich, J. A. (1989). In pursuit of myosin function. Cell Regul. 1, 1-1 1 . Swanljung-Collins, H., Montibeller, J., and Collins, J. H. (1987). Purification and characterization of the 1 10-kDa, actin- and calmodulin-binding protein from intestinal brush border: A myosinlike ATPase. In “Methods in Enzymology” (A. R. Means and P. M. Conn, eds.), Vol. 139, pp. 137- 148. Academic Press, Orlando, Florida. Vibert, P., and Cohen, CI. (1988). Domains, motions and regulation in the myosin head. J. Muscle Res. Cell Motil. 9, 296-305. Wakim, B. T., Alexander, K. A,, Masure, H. R., Cimler, B. M., Storm, D. R., and Walsh, K. A. (1987). Amino acid sequence of P-57, a neurospecific calmodulin binding protein. Biochemistry 26, 7466-7470. Wasserman, R. H., and Fullmer, C. S. (1989). On the molecular mechanism of intestinal calcium transport. Adv. Exp. Med. Eiol. 249, 45-66. Wolenski, J. S., Hayden, S., and Mooseker, M. S. (1990). Contributions of the C-terminal tail domain to BB myosin I function. J . Cell Eiol. 111, 168a. Wolenski, J. S., Hayden, S., Forscher, P., and Mooseker, M. S. (1991). Characterization of the actin binding, enzymatic and mechanochemical properties of brush border myosin I: Contributions of the “tail” domain. Submitted for publication.
This Page Intentionally Left Blank
CHAPTER 4
Protein Interactions Linking Actin to the Plasma Membrane in Focal Adhesions Keiko 0. Simon, Carol A. Otey, Fredrick M. Pavalko, and Keith Burridge Department of Cell Biology and Anatomy The University of North Carolina at Chapel Hill Chapel Hill, North Carolina 27599
I. Introduction 11. Interaction between a-Actinin and PI Integrin 111. Interaction between Talin and Actin IV. Conclusion References
1. INTRODUCTION
The discrete regions of the ventral plasma membrane where cells adhere tightly to the substratum are known as focal adhesions, focal contacts, or adhesion plaques. In these areas, stress fibers are anchored to the cytoplasmic face of the plasma membrane at the sites where the cell interacts directly with the extracellular matrix. In sifu there are structures that appear to be homologous with focal adhesions, such as the dense plaques of smooth muscle, the myotendinous junctions of skeletal muscle, and the adhesions made by platelets and many other cells interacting with the extracellular matrix. During development, cell interactions with the extracellular matrix provide important cues for cellular differentiation, migration, and growth characteristics (for reviews, see McClay and Ettensohn, 1987; and Otey and Burridge, 1990). Cells grown in tissue culture, such as fibroblasts, endothelial cells, and epithelial cells, develop focal adhesions and provide a model system to study the interaction of cells with the extracellular matrix. Much of the research on focal adhesions is directed toward Current Topics in Membranes, Volume 38
Copyright 8 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.
57
58
Keiko 0.Simon et al.
understanding the proteins involved in linking actin filaments to the plasma membrane. The purpose of this chapter is to describe briefly our current knowledge of the organization of focal adhesions and to present recent data which suggest two alternative mechanisms for attachment of actin to the membrane at these sites. Formation of focal adhesions requires adsorption of extracellular matrix components (e.g., fibronectin, vitronectin) to the substrate. In tissue culture, the serum protein vitronectin is often the predominant extracellular matrix component of focal adhesions (Neyfakh et al., 1983; Hayman et al., 1985; Baetscher et al., 1986; Fath and Bumdge, 1989). Extracellular matrix components are bound by specific receptors in the plasma membrane, many of which belong to the integrin family (for reviews, see Hynes, 1987; Ruoslahti and Pierschbacher, 1987; Albelda and Buck, 1990). Integrins are transmembrane, heterodimeric proteins composed of CY and p subunits. Many of the integrins have been cloned and sequenced. The p subunits, which are highly conserved across many species, may be divided into at least seven classes according to slight differences in their sequences (Albelda and Buck, 1990). The (Y subunit sequences are more varied. Typically, a single type of p chain may complex with multiple different CY chains, forming heterodimers of different ligand specificity. These C Y combina~ tions function to create diversity among different integrins. Cells express specific integrins on their surface which enable them to bind selectively to different extracellular matrix components. Many extracellular matrix proteins involved in cell adhesion include a domain which contains an Arg-Gly-Asp sequence (Ruoslahti and Pierschbacher, 1987). This sequence in fibronectin and vitronectin is recognized by distinct integrins. In addition, both proteins contain heparinbinding domains that may bind to cell surface heparan sulfate proteoglycans. There is evidence that both of these domains may be important in forming focal adhesions (Woods et al., 1986). For example, when human embryo fibroblasts were plated on either the integrin-binding domain or the heparin-binding domain of fibronectin, the cells adhered but failed to form focal adhesions. However, plating these cells on both domains (either contained in one fibronectin fragment or as two separate fragments) did result in the development of focal adhesions (Woods et al., 1986). In general, focal adhesions appear to contain the same cytoskeletal proteins even when cells are plated on different extracellular matrix substrates. This may be due to the similarity of the cytoplasmic domains of the various integrins. Using a variety of in vitro binding assays and purified proteins, a potential protein-protein bridge connecting the membrane to actin filament bundles has been suggested. A direct interaction of low affinity between integrin and talin was observed using an equilibrium gel filtration assay (Horwitz et al., 1986). Talin has been shown to bind vinculin, another focal adhesion protein (Otto, 1983; Wilkins et al., 1983; Burridge and Mangeat, 1984), which in turn binds to a-actinin, an actin-binding protein (Wachsstock et al., 1987). Thus, actin stress
4. Attachment of Actin to Plasma Membrane
59
fibers may be linked to the cytoplasmic face of the plasma membrane sequentially via integrin, talin, vinculin, and a-actinin (Burridge et al., 1988). Additional linkages to the membrane at focal adhesions almost certainly exist. This is suggested by the existence of other, as yet incompletely characterized, focal adhesion components which have been detected by different methods (Turner el al., 1990; Glenney and Zokas, 1989; Beckerle, 1986), and also by the low affinity of some of the interactions described above, which may require additional stabilization in the cell (Burridge et al., 1988). Also, the associations demonstrated in vitro have not been confirmed in vivo. Recently, two other interactions have been identified in vitro which suggest alternative connections to the membrane.
II. INTERACTION BETWEEN a-ACTININ AND P I INTEGRIN In a study designed to identify cytoplasmic integrin-binding proteins, a synthetic peptide corresponding to the cytoplasmic domain of P I integrin was used in affinity chromatography experiments (Otey et al., 1990). The peptide was synthesized with an additional N-terminal cysteine, which was used to couple the peptide to thiopropyl Sepharose. This should mimic the orientation of the intact integrin as it emerges from the membrane. Extracts of cultured chick embryo fibroblasts were passed over this column and the bound proteins were eluted with 300 mM salt. The column eluates contained a number of unidentified proteins, as well as two proteins which were identified as a-actinin and vinculin by immunoblotting. 12sI-Labeleda-actinin also bound to the integrin peptide in a solidphase binding assay. In the same assay, 12sI-labeled vinculin did not bind, suggesting that vinculin does not bind directly to integrin and may have been retained on the peptide column as part of a larger complex of proteins. When aactinin was cleaved proteolytically into two fragments, the 53-kDa rod domain bound to the integrin peptide but the 27-kDa globular head domain (which contains the actin-binding site) did not bind. These results suggest that a-actinin contains a binding site for integrin which is located somewhere other than in the actin-binding domain. In future experiments, it should be possible to map the interactive sites on both a-actinin and P, integrin by using smaller proteolytic fragments of a-actinin and shorter integrin peptides. In order to determine if a-actinin could bind to intact integrins and to integrins belonging to subfamilies other than P I , the smooth muscle (PI) integrin was purified from chicken gizzard and the glycoprotein IIb/IIIa (P3) was purified from human platelets. 1251-Labeleda-actinin bound to both of the purified integrins when these were adsorbed to microtiter wells. However, the binding of aactinin to the synthetic peptide was of higher affinity than the binding to intact integrins, perhaps because the integrins undergo a conformational change when
60
Keiko 0.Simon et al.
they are removed from a membrane environment or because the a chain interferes with the binding, To determine if a-actinin could bind to integrins in a more native conformation, purified IIb/IIIa was reinserted into synthetic phospholipid vesicles. lZ5I-Labeled a-actinin bound to vesicles containing incorporated IIb/IIIa, but did not bind appreciably to plain phospholipid vesicles. Taken together, these results demonstrate that a-actinin binds to a synthetic integrin peptide either on an affinity matrix or in a microtiter well, and that a-actinin also binds to purified integrin which is adsorbed to plastic or incorporated into a phospholipid membrane (Otey et al., 1990). One interpretation of these results is that a-actinin may play a role in the attachment of actin filaments to the membrane in focal adhesions as well as in bundling actin filaments in the stress fiber. Finding an interaction between aactinin and integrin was unexpected because earlier electron microscopy results suggested that a-actinin is located farther from the membrane than vinculin (Chen and Singer, 1982). Several approaches are being employed to investigate the relationship between a-actinin and integrin in vivo. One of these has been to study the behavior of a-actinin and its fragments after their microinjection into living cells. Since the binding site for integrin on a-actinin was found to lie within the 53kDa rod domain of a-actinin, it seemed possible that this fragment might coloc a k e with integrin in focal adhesions upon its introduction into live cells. This was tested by microinjecting the 53-kDa fragment into fibroblasts, after it had been fluorescently labeled. Consistent with its interaction with integrin in vitro, this fragment of a-actinin was indeed found to concentrate initially in integrincontaining focal adhesions (Pavalko and Burridge, 1991). However, following localization of the microinjected 53-kDa a-actinin fragment at these sites, disruption of stress fibers and ultimately of most of the focal adhesions was observed. This result was somewhat surprising since proteolytic fragments of talin that were microinjected into cells had been shown previously to concentrate in focal adhesions without disrupting these structures or their associated stress fibers (Hock et al., 1989; Nuckolls et al., 1990). There are several possible explanations for the disruptive effect that the 53kDa fragment of a-actinin has on the actin cytoskeleton. Vinculin binds to aactinin (Wachsstock et d . , 1987), which could potentially provide a binding site for the rod domain of a-actinin in focal adhesions. The 53-kDa fragment might interfere with the vinculin-a-actinin interaction by blocking such a binding site. This seems unlikely, however, since in Western blot assays 1251-labeledvinculin did not bind to the 53-kDa fragment but bound to the 27-kDa actin-binding domain of a-actinin (Pavalko and Burridge, 1991). Since the 53-kDa fragment of a-actinin does not bind actin, it is unlikely that stress fiber disruption results from the 53-kDa fragment interfering directly with the interaction between actin filaments and endogenous a-actinin molecules in the cell. Other proteins may,
4. Attachment of Actin to Plasma Membrane
61
however, interact with the rod domain of a-actinin, and it is possible that the 53kDa fragment competes with intact a-actinin for binding to another, as yet unidentified, protein. Displacement of a protein that binds to the rod domain of a-actinin could result in disruption of stress fibers if this protein were involved in organizing actin filaments into stress fibers. Alternatively, the 53-kDa fragment may compete with intact a-actinin for binding to integrin cytoplasmic domains and thereby dislocate an important attachment of stress fibers to focal adhesions. This could result in disassembly of the stress fibers. Support for this latter possibility is provided by the initial incorporation of the 53-kDa fragment into focal adhesions and by the inability to detect endogenous a-actinin within the adhesions of injected cells, suggesting that the a-actinin has been displaced by the 53-kDa fragment (Pavalko and Burridge, 1991).
111. INTERACTION BETWEEN TALlN AND ACTIN
Recently, a direct binding between talin and actin has been identified (Muguruma et d.,1990). This interaction involves the binding of talin to polymerized actin as detected by cosedimentation and other assays. In these assays, the binding of talin to actin appears to be of high stoichiometry. Talin can be proteolytically digested by a calcium-dependent protease (CDP 11) into two fragments, a 47-kDa fragment and a 190-kDa fragment (Fox et al., 1985; O’Halloran et al., 1985; Beckerle et a!., 1987). The actin-binding site in talin has been localized to the 190-kDa fragment. The regulation and exact nature of this interaction have yet to be clarified. Using similar cosedimentation assays, we have confirmed the interaction between talin and actin (K. Simon, unpublished observations). This evidence indicates that talin may provide a direct link between actin and integrin in focal adhesions. Previously, talin was reported to contain an actin-severing activity (Collier and Wang, 1982). In our hands, however, highly purified talin did not contain any actin-severing activity (K. Simon and K. Burridge, unpublished observations). Impure preparations of talin did sever actin filaments but this activity could be separated from talin during subsequent purification steps. We conclude that the reported actin-severing activity is due to a contaminant in the talin preparations (K. Simon and K. Burridge, unpublished observations).
N. CONCLUSION These recent observations imply that single proteins, either a-actinin or talin, form a bridge between actin and integrin (illustrated in Fig. 1). The interaction of actin with these two proteins may serve different functions in focal adhesions.
62
Keiko 0.Simon et al.
FIG. 1 A diagram illustrating some of the protein interactions in focal adhesions which have been determined in vifro. Vitronectin (VN) or fibronectin (FN) is depicted adsorbed to a glass or plastic substratum. Integrins (ap dimers) binding to extracellular matrix components are shown spanning the plasma membrane (PM). On the cytoplasmic side of the membrane, the integrin cytoplasmic domains are shown binding to either talin or a-actinin (a-A). In turn, each of these proteins binds actin. Vinculin (V) is shown binding to both talin and a-actinin. Paxillin (Px) is drawn binding to vinculin (Turner era/., 1990). Other proteins have been identified in focal adhesions, but their interactions have not been established. This working model is modified from Burridge et al. (1988).
Besides providing an attachment site for actin filaments, focal adhesions may also play a role in regulating the assembly and disassembly of the associated stress fibers. In addition, exchange of actin subunits occurs in focal adhesions as evidenced by studies of fluorescently labeled actin microinjected into cells (Kreis et al., 1982; Wang, 1984). If actin subunit exchange occurs within stress fibers at the same time that tension is being transmitted to the membrane, then it would not be surprising to have more than one mode of linkage to the membrane. One set of links between actin and the membrane may be involved in the initial formation of focal adhesions, whereas other proteins may function primarily in maintaining attachment in mature focal adhesions. Undoubtedly, focal adhesions are complex structures. Many proteins have been identified in these regions and yet little is known about the functions of most of these (Burridge et al., 1988). Determining how these proteins interact and their role in anchoring actin filaments to the membrane will continue to be a major goal for this and other laboratories.
4. Attachment of Actin to Plasma Membrane
63
Acknowledgments We are grateful to Drs. Muguruma, Matsumura, and Fukuzawa for sharing and allowing us to cite their unpublished results. Work in this laboratory is supported by NIH grant GM-29860 to K.B. Support from an MDA Postdoctoral fellowship to C.A.O. and NIH National Research Service Award CA-08493 to F.M.P. are gratefully acknowledged.
References Albelda, S. M., and Buck, C. A. (1990). lntegrins and other cell adhesion molecules. FASEE J . 4, 2868-2880. Baetscher, M.. Pumplin, D. W., and Bloch, R. J. (1986). Vitronectin at sites of cell-substrate contact in cultures of rat myotubes. J. Cell Eiol. 103, 369-378. Beckerle, M. C. (1986). Identification of a new protein localized at sites of cell-substrate adhesion. J. Cell Eiol. 103, 1679-1687. Beckerle, M. C., Burridge, K., DeMartino, G . N., and Croall, D. E. (1987). Colocalization of calcium-dependent protease I1 and one of its substrates at sites of cell adhesion. Cell (Cambridge, Mass.) 51, 569-577. Burridge, K., and Mangeat, P. (1984). An interaction between vinculin and talin. Narure (London) 308, 744-746. Burridge, K., Fath, K., Kelly, T., Nuckolls, G . , and Turner, C. (1988). Focal adhesions: Transmembrane junctions between the extracellular matrix and the cytoskeleton. Annu. Rev. Cell Eiol. 4, 487-525. Chen, W. T., and Singer, S. J. (1982). Immunoelectron microscopic studies of the sites of cellsubstratum and cell-cell contacts in cultured fibroblasts. J . Cell Eiol. 95, 205-222. Collier, N. C., and Wang, K. (1982). Human platelet P235: A high Mr protein which restricts the length of actin filaments. FEES LE7T. 143, 205-210. Fath, K. R., Edgell, C. J. S . , and Burridge, K. (1989). The distribution of distinct integrins in focal contacts is determined by the substratum composition. J. Cell. Sci. 92, 67-75. Fox, J. E. B., Goll, D. E., Reynolds, C. C., and Phillips, D. R. (1985). Identification of two proteins (actin-binding protein and P235) that are hydrolyzed by endogenous Ca2+ dependent protease during platelet aggregation. J. Eiol. Chem. 260, 1060-1066. Glenney, J. R., Jr., and Zokas, L. (1989). Novel tyrosine kinase substrates from Rous sarcoma virustransformed cells are present in the membrane skeleton. J. Cell Eiol. 108, 2401-2408. Hayman, E. G., Pierschbacher, M. D., Suzuki, S., and Ruoslahti, E. (1985). Vitronectin-A major cell attachment-promoting protein in fetal bovine serum. Exp. Cell Res. 160, 245-258. Hock, R. S., Sanger, J. M., and Sanger, J. W. (1989). Talin dynamics in living microinjected nonmuscle cells. Cell Moril. Cyroskel. 14, 271-287. Honvitz, A., Duggan, K., Buck, C., Beckerle, M. C., and Burridge, K. (1986). Interaction of plasma membrane fibronectin receptor with talin-A transmembrane linkage. Narure (London) 320, 531-533. Hynes, R. 0. (1987). Integrins: A family of cell surface receptors. Cell (Cambridge, Mass.) 48,549554. Kreis, T. E., Geiger, B., and Schlessinger, J. (1982). Mobility of microinjected rhodamine actin within living chicken gizzard cells determined by fluorescence photobleaching recovery. Cell (Cambridge. Mass.) 29, 835-845. McClay, D. R., and Ettensohn, C. A. (1987). Cell adhesion in morphogenesis. Annu. Rev. Cell Eiol. 3, 319-45. Muguruma, M., Matsumura, S . , and Fukuzawa, T. (1990). Direct interactions between talin and actin. Eiochem. Eiophys. Res. Commun. 171, 1217-1223. Neyfakh, A. A., Tint, I. S., Svitkina, T. M., Bershadsky, A. D., and Gelfand, V. I. (1983).
64
Keiko 0.Simon et al.
Visualization of cellular focal contacts using a monoclonal antibody to the 80 kD serum protein adsorbed on the substratum. Exp. Cell Res. 149, 387-396. Nuckolls, G. H., Turner, C. E., and Bumdge, K. (1990). Functional studies of the domains of talin. J. Cell Biol. 110, 1635-1644. O’Halloran, T., Beckerle, M. C., and Burridge, K. (1985). Identification of talin as a major cytoplasmic protein implicated in platelet activation. Nature (London) 317, 449-45 1 . Otey, C. A,, and Bumdge, K. (1990). Patterning of the membrane cytoskeleton by the extracellular matrix. Semin. Cell Biol. 1, 391-399. Otey, C. A,, Pavalko, F. M., and Bumdge, K. (1990). An interaction between a-actinin and the PI integrin subunit in virro. J . Cell Biol. 111, 721-729. Otto, J. J. (1983). Detection of vinculin-binding proteins with an 1251-vinculingel overlay technique. J . Cell Biol. 97, 1283-1287. Pavalko, F. M., and Burridge, K. (1991). Disruption of the actin cytoskeleton following microinjection of proteolytic fragments of a-actinin. J. Cell Biol. 114. Ruoslahti, E., and Pierschbacher, M. D. (1987). New perspectives in cell adhesion: RGD and integrins. Science 238, 491-497. ’hrner, C. E., Glenney, J. R., Jr., and Bunidge, K. (1990). Paxillin: A new vinculin-binding protein present in focal adhesions. J . Cell Biol. 111, 1059-1068. Wachsstock, D. H., Wilkins, J. A,, and Lin, S. (1987). Specific interaction of vinculin with aactinin. Biochem. Biophys. Res. Commun. 146, 554-560. Wang, Y.-L. (1984). Reorganization of actin filament bundles in living fibroblasts. J . Cell Biol. 99, 1478- 1485. Wilkins, J. A., Chen, K. Y.,and Lin, S. (1983). Detection of high molecular weight vinculin binding proteins in muscle and nonmuscle tissues with an electroblot-overlay technique. Biochem. Biophys. Res. Commun. 116, 1026-1032. Woods, A., Couchman, J. R., Johansson, S., and Hook, M. (1986). Adhesion and cytoskeletal organization of fibroblasts in response to fibronectin fragments. EMBO J. 5, 665-670.
CHAPTER 5
Ankyrins: A Family of Proteins That Link Diverse Membrane Proteins to t h e Spectrin Skeleton Vann Bennett, Ed Otto, Jonathan Davis, Lydia Davis, and Ekaterini Kordeli Department of Biochemistry and the Howard Hughes Medical Institute Duke University Medical Center Durham, North Carolina 27710
I. Introduction 11. Ankyrin Structure 111. Ankyrins Are a Multigene Family IV. Functional Diversity of Ankyrin due to Alternative Splicing of mRNA V. Mapping the Binding Sites of Ankyrin
VI. Summary and Future Perspectives References
1. INTRODUCTION
Spectrin is the principal component of a system of structural proteins known as the membrane skeleton that is associated with the plasma membranes of many cells. Spectrin was first characterized in human erythrocytes, but now is known to be a family of closely related polypeptides present in most vertebrate cells, Drosophila, echinoderms, and lower eukaryotes such as Dictyostelium. Spectrins are especially prominent in vertebrate brain tissue, where they comprise 3% of the total membrane protein. The structure of spectrin and the organization of a spectrin-actin network were first elucidated in the human erythrocyte membrane. Spectrin is a flexible rodshaped molecule comprising two subunits aligned side to side to form heterodimers and head to head into tetramers that are capable of cross-linking actin Current Topics in Membranes, Volume 38 Copyright 0 1991 by Academic Press, Inc. All rights of reprcductlon in any form reserved.
65
66
Vann Bennett et al.
with binding sites for actin on both ends. Two classes of protein interactions have been identified as essential for assembly of spectrin tetramers into a membraneassociated network: (1) Linkage of spectrin to the membrane. Spectrin molecules have two distinct sites of interaction with integral membrane proteins that are likely to occur simultaneously. A major membrane attachment is provided by high-affinity association of the p subunit of spectrin with ankyrin at a site located in the midregion of spectrin tetramers. Ankyrin is a peripheral membrane protein which, in turn, is associated with the cytoplasmic domain of the anion exchanger. Another association of spectrin with the membrane is mediated by protein 4.1, which is located at the ends of spectrin molecules and recognizes membrane sites that may include glycophorin C and the anion exchanger. (2) Association of multiple spectrin molecules with actin to form a two-dimensional meshwork. Spectrin molecules are cross-linked at their ends by association with actin and several accessory proteins into a regular lattice-like organization with 5-6 spectrin molecules attached to short actin filaments 30-50 nm in length to form a sheet of 5- and 6-sided polygons. Spectrin-actin junctions also contain additional proteins, including protein 4.1 and protein 4.9. Other proteins are candidates to participate in spectrin-actin interactions, including adducin, tropomyosin, and a tropomyosin-binding protein named tropomodulin. The general areas of spectrin and the erythrocyte membrane have been the subject of several recent reviews (Bennett, 1990; Coleman et al., 1989; Mangeat, 1988; Goodman et al., 1988). In spite of the high abundance of spectrin, understanding of the structure and function of the spectrin skeleton outside of the human erythrocyte still is in an embryonic stage. One general principal that can be derived from the erythrocyte membrane is that spectrin alone does not polymerize or assemble into higher order structures without the aid of accessory proteins. The focus of this chapter is on ankyrins, which comprise an important class of spectrin-organizing proteins that are candidates to link a number of integral membrane proteins to the spectrin skeleton in specialized plasma membrane domains.
II. ANKYRIN STRUCTURE Ankyrin was first discovered in human erythrocytes, where is provides a highaffinity linkage between spectrin and the cytoplasmic domain of the anion exchanger (Bennett and Stenbuck, 1979a,b, 1980a,b). Ankyrin is a large protein with a mass of 206 kDa that contains three independently folded domains: (1) an N-terminal 89-kDa domain that binds to the anion exchanger (L. Davis and Bennett, 1990); (2) a 62-kDa domain (apparent MW of 72,000 on SDS gels) that binds to spectrin (Bennett, 1978); and (3) a C-terminal 55-kDa regulatory region comprising at least two domains that modulates activity of the binding domains
67
5 . Ankyrins
(Hall and Bennett, 1987). The complete sequence of human erythrocyte ankyrin has been deduced from analysis of cDNA which encodes a protein of 1881 amino acids from a mRNA of 7 kb (Lux et al., 1990; Lambert el al., 1990). A striking feature of the 89-kDa domain of ankyrin is the presence of 22 repeats containing 33 residues that occur in tandem (Fig. 1). The repeats contain 15 highly conserved and 18 variable residues. The 33-residue periodicity is rigorously maintained, with the exception of the fourth repeat, which is 22 residues in length. Related 33-residue motifs are present in a number of apparently unrelated proteins of broad phylogenetic distribution (see Lux et al., 1990, for references): ( 1) membrane proteins involved in cell differentiation, including Linl2 and Glp-1 of Caenorhabditis elegans and Notch protein of Drosophila; (2) cytosolic proteins involved in cell-cycle regulation such as SW16 and SW14 of Saccharomyces cerevisiae and CDC 10 of Schizosaccharomyces pombe. The functional basis for these homologous sequences is not clear, although a shared interaction with a common class of molecules is a reasonable guess. The 89-kDa domain is globular and has a circular dichroism spectrum consistent with 30% (Y helix (L. Davis and Bennett, 1990). These physical properties provide some boundary conditions for predictions of the organization and folding of the 33-amino acid repeating sequences (Fig. 2). The value of 30% (Y helix implies a single helix of 10 residues per repeat if it is assumed that each repeat is folded in a quasi-equivalent manner. The repeats may be packed as independent
-1
regulatory domain
t
t
anion spectrin-binding domain exchangerbinding domain
I
89K
X
62K
H
FIG. 1 Schematic model for the domain structure of human erythrocyte ankyrin. The presence of 33-residue repeats in the 89-kDa domain is demonstrated in the dot-plot (right) of ankyrin sequence compared with itself (Lux et al., 1990).
68
Vann Bennett et al. A. EXCHANGER BINDING
SPECTRIN BINDING
REGULATORY alt. splice (2.2)
33-aa repeats
. A
A. EXCHANGERBINDING DOMAIN N
827 403
779
retains activity
repeats 12-22 FIG. 2 Linear representation of the sequence of human erythrocyte ankyrin and its binding domains. The regions of the sequence implicated in association with the anion exchanger and spechin are denoted by thin lines and are drawn to scale in the model of the entire ankyrin molecule (top). This information is based on the study by L. Davis and Bennett (1990).
units, or, alternatively, participate in a larger structure involving several repeats. The repeats or combination of repeats presumably are organized such that each unit has similar regions exposed to solvent with other portions confined to the interior of the 89-kDa domain. The shape of the 89-kDa domain requires that the repeats be packed either into a sphere or a compact helix but not into an extended rod. A helical configuration would accommodate an unlimited number of repeats while a sphere would have some upper limit. It is of interest that another form of ankyrin from brain also has 22 repeats of 33 amino acids (E. Otto, unpublished data; see below). The maximum number of ankyrin repeats thus may be 22 if the repeats are packed into a sphere.
111. ANKYRINS ARE A MULTIGENE FAMILY
Ankyrin has been detected associated with the membranes of a number of tissues in addition to erythrocytes by radioimmunoassay, immunoblots of sodium dodecyl sulfate gels, and immunofluorescence (Bennett, 1979; Davis and Bennett, 1984a,b; Drenckhahn and Bennett, 1987). An isoform of ankyrin has been purified from brain (Davis and Bennett, 1984a,b). Brain ankyrin has properties in common with erythrocyte ankyrin, although it is the product of a distinct gene (see below). Brain and erythrocyte ankyrin share physical properties (asymmetric monomers of MW of approximately 200,000), have a similar domain structure, associate with the p subunit of spectrin at a site close to the midregion of spectrin tetramers, and both proteins bind to the cytoplasmic domain of the erythrocyte anion exchanger. The two ankyrins also share the property of binding to tubulin
69
5. Ankyrins
via the 89-kDa domain. These features support the idea that brain and erythrocyte ankyrin are closely related isoforms. Recent work indicates that erythrocyte and brain ankyrins are prototypes of two families of ankyrin that have major differences in cellular expression and localization (Table I) (J. Davis et al., 1989; Kordeli et al., 1990). Ankyrin, forms or restricted ankyrins react better with antibodies against red blood cell (RBC) ankyrin, are expressed in a limited number of cells in brain and kidney, and have a highly polarized distribution within these cells. Ankyrin, in the nervous system is expressed primarily in neurons, and in kidney is present in high concentrations in distal tubule cells and intercalated cells of the collecting duct (Drenckhahn et al., 1985; J. Davis et al., 1989; Koob et al., 1988). Ankyrin, forms are localized at specialized cell domains such as the node of Ranvier (Kordeli et al., 1990), postsynaptic membrane of the neuromuscular junction (Flucher and Daniels, 1989), and basolateral surfaces of epithelial cells (Drenckhahn and Bennett, 1987; Koob et al., 1988). Ankyrin, also is present in nerve cell bodies, at the initial segment of axons in dendrites, and in unmyelinated axons (Kordeli et al., 1990). Several ion channels colocalize with ankyrin, in specialized membrane domains and interact with erythrocyte ankyrin in in vitro assays: the anion exchanger of kidney collecting ducts (Drenckhahn et al., 1985), the a,isoform of the Na+ /K+-ATPase of kidney (Nelson and Veshnock, 1987; Koob et al., 1988; Morrow et al., 1989), and the voltage-dependent sodium channel of brain (Srinivasan et al., 1988). A current working hypothesis is that ankyrin plays a role in either initial targeting of these ion channels to specialized areas of the cell or in maintaining them once the membrane domains have assembled. Members of the ankyrin, group cross-react better with antibodies against the major form of ankyrin in brain and are expressed in most cells of brain and kidney. Ankyrin, is present in kidney in glomeruli, proximal and distal tubules, and loops of Henle (J. Davis et al., 1989). Ankyrin, in brain is present in glial as
TABLE I Two Families of Ankyrin in Kidney and Brain 1. Ankyrin, isoforms (cross-react with RBC ankyrin)
Expressed only in certain cells, e . g . , neurons and distal tubules and collecting ducts Localized in specialized membrane domains, e.g., node of Ranvier, neuromuscularjunction, and basolateral domains of kidney cells Likely membrane protein sites: Na+/K+-ATPase,distal tubule, anion exchanger, collecting duct, voltage-dependent sodium channel, nerve 2. Ankyrin, isoforms (cross-react with brain ankyrin) Expressed in most cell types Membrane protein site@) may include GP85iCD44 and ABG 205; other sites not known
70
Vann Bennett et al.
well as neuronal cells and does not exhibit a high concentration at the nodes of Ranvier. Neither form of ankyrin is present in internodal regions of myelinated axons, at least as detected with available antibodies, even though these areas of the membrane contain spectrin. The membrane attachment sites for ankyrin, forms are not yet established. Current candidates include a broadly distributed membrane glycoprotein, termed Pgp- 1, gp-85, or CD44 antigen (Kalomiris and Bourguignon, 1988), which is likely to participate in intercellular adhesion (Stamenkovic et al., 1989). Another ankyrin-binding protein, termed AGP-200, has been detected in brain (Treharne et al., 1988). Recent experiments in our laboratory using ankyrin,-affinity columns suggest that multiple membrane proteins recognize this form of ankyrin (J. Davis and E. Otto, unpublished data). The differences in cellular localization of ankyrins may be due to variations in relative affinities for target proteins. Functional differences between the ankyrin, and ankyrin, isoforms have been demonstrated in in vitro assays employing erythrocyte and brain ankyrin as prototypes of these families. Each type of ankyrin binds preferentially to a distinct type of spectrin: brain ankyrin binds better with the general form of spectrin and erythrocyte ankyrin better with erythrocyte spectrin (Bennett et al., 1982; Davis and Bennett, 1984b). The specificity of ankyrins for spectrin isoforms is likely to be important in differential targeting of ankyrins at least in brain, where both erythroid and general spectrin are coexpressed in the same cells but localized in different domains of certain neurons (Lazarides and Nelson, 1983; Goodman et al., 1988). Membrane-binding sites for ankyrins also are distinct in brain (Davis and Bennett, 1986) and in kidney (J. Davis et al., 1989). The specificity of membrane sites for ankyrins in kidney was not absolute, but reflected 2.5 to 10-fold differences in relative affinities. Further evidence for common features in binding sites was that the cytoplasmic domain of the erythrocyte anion exchanger displaced membrane binding of both ankyrins. A view supported by these experiments is that the system of ankyrins and their membrane sites is formally analogous to that of catecholamines, where chemically related but distinct ligands interact with a family of membrane receptors. Important information required to understand the complex system of ankyrins and their membrane sites is the sequence of different members of the ankyrin family, cDNA encoding ankyrin, from human brain has recently been cloned and sequenced (Otto et al., 1991). The open reading frame encodes a protein of 204 kDa, which is close to the size of erythrocyte ankyrin. Two major regions of high homology to erythrocyte ankyrin are present in the sequence of brain ankyrin: one involves the 33-residue repeats which are present in 22 tandem copies in both proteins, and the other area of homology is located within the spectrinbinding domains. Conservation within the repeat domains of these ankyrins includes preservation of the number of repeats and the feature that the fourth repeat in both proteins is the only repeat to deviate from the 33-residue peri-
5 . Ankyrins
71
odicity. Moreover, each repeat in brain ankyrin is more closely related to the corresponding repeat in erythrocyte ankyrin such that brain repeat number one is most homologous with erythrocyte repeat number one, repeat number two is most homologous with corresponding repeat number two, etc. Two areas of almost complete divergence between the two ankyrins are at the connection between the 33-residue repeat domain and the spectrin-binding domain, and in the regulatory domains. These areas of sequence differences are candidate sites to explain the functional differences between brain and RBC ankyrins. How many different genes encode ankyrins? The information to answer this question still is incomplete, although a minimum number is at least five distinct ankyrin genes. Brain and RBC ankyrins are each products of distinct genes based on multiple differences in codon usage, and have been mapped to distinct chromosomes in mice (personal communication from J. Barker and colleagues at the Jackson Laboratories) and humans (personal communication from B. Tse and B. Forget, Yale University School of Medicine). Northern blots and immunoblots with antibody raised against recombinant proteins indicate that the two ankyrins from kidney are distinct from both brain and RBC ankyrin, and that liver has a unique form of ankyrin not present in brain, RBCs, or kidney (Otto er al., 1990). Moreover, in brain tissue, ankyrin, at the node of Ranvier may differ from ankyrin, in neuronal cell bodies based on differential reactivities with antibodies (K. Kordeli and V. Bennett, unpublished data). A rational nomenclature to describe these different ankyrins has not been formulated, and probably will require additional information. Simply using the tissue origin to designate ankyrins is not sufficient. For example, the same gene that encodes erythrocyte ankyrin also is responsible for a form of ankyrin in the cerebellum and forebrain. Evidence for expression of erythrocyte ankyrin in brain is that mutant mice missing erythrocyte ankyrin also are missing a form of ankyrin in the cerebellum, as detected by Northern blot analysis and immunoblots (Lambert et al., 1990; J. Barker, personal communication). An arbitrary system would be to refer to the ankyrins by number according to the order in which they were sequenced. A nomenclature based on function would be preferable, but will require more knowledge of the structures of different ankyrins and their roles in cells.
IV. FUNCTIONAL DIVERSITY OF ANKYRIN DUE TO ALTERNATIVE SPLICING OF mRNA
Erythrocyte ankyrin has a regulatory region comprising several domains which are located at the C-terminal end of the polypeptide. One of these domains, located near the C-terminus, is cleaved by calpain and results in an ankyrin with
72
Vann Bennett et al.
reduced binding to the anion exchanger in erythrocyte membranes (Hall and Bennett, 1987). Another regulatory domain is located N-terminal to the calpainsensitive domain. Deletion of this domain results in a lower molecular weight form of ankyrin present in human erythrocyte membranes known as protein 2 . 2 . Protein 2.2 results from differential processing of mRNA, since the missing region, as identified by antibodies, lies internally within the sequence (Lux et a l . , 1990). Moreover, cDNA clones have been isolated that contain an in-frame deletion of 163 amino acids that include the portion of sequence missing in protein 2.2 (Lux et af., 1990; Lambert et af., 1990). The alternatively spliced protein 2.2 is an activated ankyrin with an increased affinity for spectrin and increased association with the anion exchanger in erythrocyte membranes (Hall and Bennett, 1987). Protein 2.2 also expresses a binding site for a major class of unidentified protein sites in kidney microsomes that do not recognize the larger form of ankyrin (L. Davis et af., 1989). The regulatory region thus defines specificity in binding to membrane sites as well s modulates affinities. The explanation for increased activities of protein 2.2 is now known. One possibility is a conformational difference between 2.2 and intact ankyrin. Another alternative is that the domain missing in 2 . 2 occupies binding sites as a pseudosubstrate, as occurs with regulatory domains of several protein kinases (Kemp et al., 1987). The phenomenon of alternative splicing of ankyrin mRNA is likely to involve regions in addition to the region missing in protein 2.2 and to be a feature of other members of the ankyrin family. In the of erythrocyte ankyrin, a highly basic stretch of 32 residues (pf greater than 10) located at the C-terminus of the regulatory domain also is alternatively spliced (Lambert et af., 1990). Another candidate site that would have a significant functional consequence would be in the region interconnecting the 89-kDa and spectrin-binding domains. A potential example of splicing in brain involves ankyrin,, which exhibits two sizes of mRNA on Northern blots (Otto et al., 1991).
V. MAPPING THE BINDING SITES OF ANKYRIN
How does ankyrin interact selectively with the anion exchanger, Na /K ATPase, and voltage-sensitive sodium channel as well as other membrane proteins that remain to be identified? Two general possibilities for association of ankyrin with multiple membrane proteins are either ( 1 ) many different proteins contain a conserved sequence that is sufficient to bind to ankyrin, or ( 2 ) multiple proteins have independently developed ability to associate with ankyrin through an evolved fit, a process analogous in principle to antigen-antibody interactions. In order to approach this question, we have examined the site of the anion +
+
73
5 . Ankyrins
exchanger that interacts with ankyrin and the sites of ankyrin involved in binding to the anion exchanger and the Na+ / K + -ATPase (L. Davis et al., 1989; J. Davis and Bennett, 1990; L. Davis and Bennett, 1990). The available evidence supports the view that ankyrin-binding activity evolved independently at least in the case of the small set of proteins currently known to associate with ankyrin. An important area of contact between the erythrocyte anion exchanger and ankyrin is located between residues 174 and 186 within the cytoplasmic domain (L. Davis et al., 1989). However, these residues alone are not sufficient for binding. A high-affinity interaction requires extended noncontiguous segments distributed along a stretch of at least 100 amino acids and is likely to involve several sites of contact (L. Davis et al., 1989; Willardson et al., 1989). Thus, a simple linear sequence that defines an ankyrin-recognition site is not likely, as is also suggested by lack of obvious sequence homology between the anion exchanger, Na+ /K -ATPase, and voltage-dependent sodium channel. The ankyrin-binding site of the anion exchanger could have similar folding to the site of the Na /K -ATPase, even though these proteins do not exhibit obvious similarity in primary sequence. A likely possibility, if these proteins evolved through divergent evolution, is that they bind to closely related sites on ankyrin. Alternatively, if ankyrin-binding activity evolved independently, target proteins would recognize different sites and possibly different domains of ankyrin. The binding site for the anion exchanger is completely contained within the 89-kDa domain of human erythrocyte ankyrin (J. Davis and Bennett, 1990). Ability of the 89-kDa domain to bind to the anion exchanger is retained by proteolytic fragments containing only 33-residue repeats (J. Davis and Bennett, 1990). The 33-residue repeats thus play a major role in association of ankyrin with the anion exchanger. The 33-amino acid repeats of erythrocyte ankyrin contain the following consensus sequence: -G-TPLH-AA-GH--V-LLGA-N-, where the 15 designated residues are highly conserved and the 18 undesignated positions are variable between different repeats (Lux et al., 1990). The specificity in interaction between the anion exchanger and ankyrin could be provided either by conserved portions of the 33-amino acid repeat or by the variable portions of this motif. If the conserved portions of the repeats are involved, then the repeats would be equivalent with respect to binding activity. However, if the variable segments are required, then association of the anion exchanger would be restricted to some but not all of the repeats. A consequence of association mediated by the variable regions of sequence would be that ankyrin could interact with a number of different proteins. In this case, the repeats would be analogous to antibodies where diversity results from variable domains that couple different antigens to the same molecules through a constant domain. The potential for diversity of associations of ankyrin is not exploited in the erythrocyte which has a single predominant ankyrin-binding protein, but could +
+
+
74
Vann Bennett et al.
be important in more complex cell types that may have multiple proteins interacting with ankyrin. The binding site for the anion exchanger could be mediated by either a single repeat, the interface between adjacent repeats, or involve a surface requiring participation of multiple repeats. A site involving several repeats is consistent with the observation that the anion exchanger is an elongated molecule that requires an extended sequence to interact with ankyrin, presumably through several sites of contact (L. Davis et al., 1989; Willardson et al., 1989). This issue as well as the question whether binding is mediated by variable or constant regions of repeats is currently under investigation using portions of the 89-kDa domain expressed in bacteria (E. Otto and L. Davis, unpublished data). The anion exchanger and Na /K -ATPase associate with distinct sites on ankyrin. The anion exchanger binds exclusively to the 89-kDa domain as discussed above, while the Na /K -ATPase binds only weakly to the 89-kDa domain and also associates with the spectrin-binding domain (J. Davis and Bennett, 1990). The Na+/K+-ATPase thus may require contacts with two and perhaps more domains of ankyrin to form the high K, complex observed with intact ankyrin. The fact that at least two domains of ankyrin participate in the high-affinity binding to the Na /K -ATPase while the 89-kDa domain alone mediates the binding of the anion exchanger provides boundary conditions for consideration of possible models of evolution for the ankyrin-binding activity of these molecules. These considerations suggest that the Na+ /K+ -ATPase and anion exchanger acquired high-affinity ankyrin-binding activities by convergent evolution of complex recognition sites involving tertiary and possibly quaternary structure. These proteins could have developed ankyrin-binding activity by completely independent pathways. Alternatively, precursors of the anion exchanger and the ATPase may have shared a common motif capable of a low-affinity association with the 89-kDa domain, and the high-affinity interactions subsequently evolved along separate pathways. In either case, the capacity of the Na+ /K+-ATPase for high-affinity binding must have developed following fusion of the 89-kDa and spectrin-binding domains to form modem ankyrin. The anion exchanger could have developed its high-affinity interaction either before or after emergence of modem ankyrin structure. The fact that ankyrin interacts with many proteins suggests that ankyrin remained relatively constant, at least to the extent of not losing key recognition sites. Considerable attention has been focused on the role of relatively simple signature sequences in mediating protein-protein interactions, with prime examples including the nuclear localization signals (Dingwall and Laskey, 1986) and the RGD sequences of a variety of extracellular matrix molecules (Ruoslahti and Pierschbacher, 1987). In the case of ankyrin and its binding proteins, a different mechanism apparently has operated that requires an evolved fit between the +
+
+
+
+
+
5 . Ankyrins
75
proteins and is analogous in principle to antigen-antibody interactions. Predicated features of this type of recognition are complex binding sites involving multiple contacts between the interacting proteins and little sequence similarity among different proteins interacting with the same target protein. A process requiring individualized binding sites may appear inelegant compared to utilization of a targeting sequence. However, in the case of molecules such as the Na /K -ATPase that may have already evolved functions such as ion channel activity, the fusion with a gene encoding an appropriate targeting sequence could be less probable than gradual evolution of a binding site. +
+
VI. SUMMARY AND FUTURE PERSPECTIVES Ankyrins are a family of structural proteins associated with the plasma membrane that are candidates to link several and perhaps many integral membrane proteins to the spectrin skeleton. Functional diversity among the ankyrins is provided by multiple genes, which currently include at least five, as well as alternative splicing of mRNAs. Currently known ankyrin-binding proteins include three different ion channels: the anion exchanger of erythrocytes and kidney collecting ducts, the Na+ /K+ -ATPase of kidney distal tubules, and the voltage-dependent sodium channel of brain. Each of these channels is colocalized with a member of the ankyrin, family in specialized membrane domains that include the basolateral domains of kidney cells, nodes of Ranvier in nervous tissue, and the postsynaptic area of neuromuscular junctions. The correct spatial positioning of these channels within cells is essential for their physiological functions. Association of ion channels with ankyrin and subsequent immobilization by the spectrin skeleton may be important for either initial targeting and/or maintaining the appropriate location of these protein in cells. Ability of these ion channels to interact with ankyrin may have evolved prior to or in parallel with development of specialized tissues early in metazoan evolution. At least two independent evolutionary pathways are likely: one utilized by the anion exchanger which recognizes an ancient 33-residue repeat motif confined to one domain of ankyrin, and another, exemplified by the N a + / K + ATPase, which involves contacts with at least two domains of ankyrin. Future questions include elucidation of the exact role of ankyrin in the assembly and/or maintenance of specialized membrane domains. It also will be important to discover additional ankyrin-binding proteins, both within the ion channel group and, possibly, other types of membrane proteins. Examples of potential candidate ion channels would be the voltage-dependent calcium channel thought to be confined at presynaptic endings and potassium channels which also must have a precisely defined localization for their activity. Cell-adhesion molecules such as CD44 and NCAM are another group of proteins that may require interac-
76
Vann Bennett et al.
tion with intracellular structural proteins and will be of interest to evaluate for possible ankyrin-binding activity.
Acknowledgments The research of this laboratory was supported in part by grants from NIH to V.B. and from the Multiple Sclerosis Society to E.K. Brenda Sampson is thanked for her help in preparing the manuscript.
References Bennett, V. (1978). Purification of an active proteolytic fragment of the membrane attachment site for human erythrocyte spectrin. J . Biol. Chem. 253, 2292-2299. Bennett, V. (1979). Immunoreactive forms of human erythrocyte ankyrin are present in diverse cells and tissues. Narure (London) 281, 597-599. Bennett, V. (1990). The spectrin-based membrane skeleton: A multipotential adaptor between the plasma membrane and cytoplasm. Physiol. Rev. (in press). Bennett, V., and Stenbuck, P. J. (1979a). Identification and partial purification of ankyrin, the high affinity membrane attachment site for human erythrocyte spectrin. J . Biol. Chem. 254, 25332541. Bennett, V., and Stenbuck, P. J. (1979b). The membrane attachment protein for spectrin is associated with band 3 in human erythrocyte membranes. Nature (London) 280, 468-473. Bennett, V., and Stenbuck, P. I. (l980a). Human erythrocyte ankyrin. Purification and properties. J . Biol. Chem. 255, 2540-2548. Bennett, V., and Stenbuck, P. J. (1980b). Association between ankyrin and the cytoplasmic domain of band 3 isolated from the human erythrocyte membrane. J . Biol. Chem. 255, 6424-2432. Bennett, V., Davis, J., and Fowler, W. (1982). Brain spectrin, a membrane-associated protein related in structure and function to erythrocyte spectrin. Narure (London) 299, 126-131. Coleman, T. R., Fishkind, D. J., Mooseker, M. S . , and Morrow, J. S . (1989). Functional diversity among spectrin isoforms. Cell Motil. Cytoskel. 12, 225-247. Davis, J., and Bennett, V. (1984a). Brain ankyrin-Purification of a 72,000 M, spectrin-binding domain. J . Biol. Chem. 259, 1874-1881. Davis, J., and Bennett, V. ( I984b). Brain ankyrin-A membrane associated protein with binding sites for spectrin, tubule and the cytoplasmic, domain of the erythrocyte anion channel. 1. Biol. Chem. 259, 13550-13559. Davis, J., and Bennett, V. (1986). Association of brain membranes and isolation of proteolytic fragments of membrane-associated ankyrin-binding protein(s). J . Biol. Chem. 264, 1619816206. Davis, J., and Bennett, V. (1990). The anion exchanger and Na+ K ATPase interact with distinct sites on ankyrin. J . Biol. Chem. 265, 17252-17256. Davis, J., Davis, L., and Bennett, V. (1989). Diversity in membrane binding sites of ankyrins: Brain ankyrin, erythrocyte ankyrin, and processed erythrocyte ankyrin associate with distinct sites in kidney microsomes. J . Biol. Chem. 264, 6417-6426. Davis, L., and Bennett, V. (1990). Mapping the binding sites of human erythrocyte ankyrin for the anion exchanger and spectrin. J . Biol. Chem. 265, 10589- 10596. Davis, L., Lux, S . E., and Bennett, V. (1989). Mapping the ankyrin-binding site of the human erythrocyte anion exchanger. J . Biol. Chem. 264, 9665-9672. Dingwall, C., and Laskey, R. A. (1986). Protein import into cell nucleus. Annu. Rev. Cell B i d . 2, 367-390. Drenckhahn, D., and Bennett, V. (1987). Polarized distribution of M, 210,000 and 190,000 analogs of erythrocyte ankyrin along the plasma membrane of transporting epithelia, neurons and photoreceptors. Eur. J . Cel/ Biol. 43, 479-486. +
77
5. Ankyrins
Drenckhahn, D . , Schluter, K., Allen, D., and Bennett, V. (1985). Colocalization of band 3 with ankyrin and spectrin at the basal membrane of intercalated cells in the rat kidney. Science 230, 1287-1289. Flucher, B., and Daniels, M. (1989). Distribution of Na+ channels and ankyrin in neuromuscular junctions is complementary to that of acetylcholine receptors and the 43 kd protein. Neuron 3, 163- 175. Goodman, S . R., Keith, K. E . , Whitfeld, C. F., Riederer, B. M., and Zagon, 1. S. (1988). Spectrin and related molecules. CRC Crii. Rev. Eiochem. 23, 171-234. Hall, T. G . , and Bennett, V. (1987). Regulatory domains of erythrocyte ankyrin. 1.B i d . Chem. 262, 10537- 10545. Kalomiris, E. L . , and Bourguignon, L. Y. (1988). Mouse T lymphoma cells contain a transmembrane glycoprotein (GP85) that binds ankyrin. J . Cell Eiol. 106, 319-327. Kemp, B., Pearson, R., Guemero, V., Bagchi, I., and Means, A. (1987). The calmodulin binding domain of chicken smooth muscle myosin light chain kinase contains a pseudosubstrate sequence. J. Eiol. Chem. 262, 2542-2548. Koob, R., Zimmerman, M., Schoner, W., and Drenckhahn, D. (1988). Colocalization and coprecipitation of ankyrin and Na + ,K -ATPase in kidney epithelial cells. Eur. J . Cell Eiol. 45, 230-237. Kordeli, K.. Davis, J., Trapp, B., and Bennett, V. (1990). An isoform of ankyrin is localized at nodes of Ranvier in myelinated axons of central and peripheral nerves. J. Cell Eiol. 110, 1341- 1352. Lambert, S., Yu, H., Prchal, J., Lawler, J., Ruff, P., Speicher, D., Cheung. M . , Kan, Y., and Palek, P. (1990). cDNA sequence for human erythrocyte ankyrin. Proc. Nail. Acad. Sci. U.S.A. 87, 1730-1 734. Lazarides, E., and Nelson, W. (1983). Erythrocyte and brain forms of spectrin in cerebellum: Distinct membrane-cytoskeletal domains in neurons. Science 220, 1295- 1296. Lux, S . E . , John, K., and Bennett, V. (1990). Analysis of cDNA for human erythrocyte ankyrin indicates a repeated structure with homology to tissue-differentiation and cell-cycle control proteins. Nature (London) 344, 36-42. Mangeat, P. H. (1988). Interaction of biological membranes with the cytoskeleton framework of living cells. Eiol. Cell 64, 262-28 I . Morrow, J. S., Cianci, C. D . , Ardito, T., Mann, A. S., and Kashgarian, M. (1989). Ankryin links fodrin to the alpha subunit of Na,K-ATPase in Madin-Darby canine kidney cells and in intact renal tubule cells. J. Cell Eiol. 108, 455-465. Nelson, W. J., and Veshnock, P. I. (1987). Ankyrin binding to the (Na+ + K + ) ATPase and implications for the organization of membrane domains in polarized cells. Nature (London)328, 533-535, Otto, E., Kunimoto, M., McLaughlin, T., and Bennett, V. (1991). J. Cell B i d . (in press). Ruoslahti, E., and Pierschbacher, M. (1987). New perspectives in cell adhesion: RGD and integrins. Science 238, 49 1-497. Srinivasan, Y., Elmer, L., Davis, J., Bennett, B., and Angelides, K. (1988). Ankyrin and spectrin associate with voltage-dependent sodium channels in brain. Nature (London) 333, 177- 180. Stamenkovic, I., Amiot, M., Pesando, J., and Seed, B. (1989). A lymphocyte molecule implicated in lymph node homing is a member of the cartilage link protein family. Cell (Cambridge, Mass.) 56, 1057-1062. Treharne, K. J., Rayner, D., and Baines, A. J. (1988). Identification and partial purification of ABGP205, an integral membrane glycoprotein from brain that binds ankyrin. Eiochem. J. 253, 345-350. Willardson, B. M . , Thevinin, B. J.-M., Harrison, M. L . , Kuster, W. M., Benson, M. D., and Low, P. S. (1989). Localization of the ankyrin-binding site or erythrocyte membrane protein, band 3. J. Eiol. Chem. 264, 15893-15899. +
This Page Intentionally Left Blank
CHAPTER 6
Contractile and Cytoslteletal Proteins in Drosophilil Embryogenesis Daniel P. Kiehart Department of Cellular and Developmental Biology Harvard University The Biological Laboratories Cambridge, Massachusetts 02138
I. Introduction 11. Movements of Early Embryogenesis A. Pole Cell Formation B. Cellularization C. Gastrulation 111. Cell Shape Change Requires Remodeling of the Actin Cytoskeleton IV. Non-Muscle Myosins A. Native Myosin 11 B. Myosin 11 Heavy Chains C. Myosin I1 Light Chains V. Spectrins VI. Summary References
1. INTRODUCTION
My laboratory has focussed on the interactions among the plasma membrane, the membrane skeleton, and the cytoskeleton that allow force production and transmission for cell shape changes during development. In Drosophila and other organisms, such movements are an integral part of homeostasis and development. Our long-range goal is to provide a molecular description of how forces are produced by components of the cytoskeleton, then transmitted to membranes. Current Topics in Membranes. Volume 38 Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.
79
80
Daniel P. Kiehart
We seek to generate a complete catalog of the molecular players that contribute to cell shape changes and to investigate the signals that effect such changes at precise times and locations in a developing embryo. Drosophila offers new hope for solutions to these problems as a consequence of powerful genetic and molecular genetic approaches that are uniquely developed in this organism among metazoans. The following details our progress in examining the primary structure of a cellular motor protein, conventional, non-muscle myosin, and our examination of the role it plays in cell shape change in Drosophila. In addition, I review our analysis of other key components of the cytoskeleton and membrane skeleton, the spectrins, and describe a new isoform of P-spectrin, PH. A rich intellectual background has focused on those genes that regulate pattern formation and the mechanism by which they precisely orchestrate development in Drosophila (e.g., French et al., 1988; Ingham, 1988). Because development is characterized by cell locomotions, shape changes, and rearrangements, such genes surely modulate directly, or indirectly, cytoskeletal and contractile protein function. In concert with more traditional cell biological approaches, molecular genetics in Drosophila promises new insight into membrane, membrane skeleton, and cytoskeleton function in development. 11. MOVEMENTS OF EARLY EMBRYOGENESIS Changes in cell surface architecture and membrane shape during Drosophila development have been studied by light microscopy of histological sections (reviewed in Campos-Ortega and Hartenstein, 1985) and of living developing embryos (e.g., Ede and Counce, 1956; Bownes, 1975; Warn and Magrath, 1982; Foe and Alberts, 1983; Wieschaus and Nusslein-Volhard, 1986; Kiehart et al., 1990; D. P. Kiehart, P. Young, and S. Inoue, unpublished observations). Higher resolution views have been provided by both scanning (Turner and Mahowald, 1976, 1977) and transmission electron microscopy of fixed material (Mahowald, 1963a,b; Fullilove and Jacobson, 1971; Rickoll, 1976; Zalokar and Erk, 1977; Rickoll and Counce, 1980; Stafstrom and Staehelin, 1984a,b; Callaini and Anselmi, 1988; Katoh and Ishikawa, 1989; K. McDonald and D. P. Kiehart, unpublished observations). Early movements of primary interest include cell shape changes that occur during polar body formation, cellularization, gastrulation, and the morphogenetic movements that characterize later embryogenesis. A. Pole Cell Formation
Following fertilization, nuclei proliferate rapidly, without accompanying cytokinesis. Late in mitotic cycle 8, most nuclei begin to migrate outward toward the embryo surface. Nuclei reach the posterior pole first, and a unique set of
6. Contractility in Drosophilu Development
81
cells, the cells of the presumptive germ line or so-called pole cells, are the first elaborated (during nuclear cycles 9 and 10). Membrane bulges outward, a nucleus is drawn into the bulge, and membrane pinches off behind the nucleus to form a cell. Approximately 1-5 nuclei from the posterior end of the embryo participate, and subsequent divisions increase the number of pole cells to 15-30. B. Cellularization
During cycles 9 and 10, when most of the remaining nuclei reach the embryo surface, an intimate relationship between the dividing nuclei, the membrane skeleton, and the shape of the surface of the embryo is established. During interphase, a bulge forms on the embryo surface; at mitosis, the bulge relaxes or collapses. The electron microscope offers dramatic views of the changes in the shape of the cell surface (Turner and Mahowald, 1976, 1977) and documents an intimate relationship between nucleus, microtubules, microfilaments, and the surface bulge (Mahowald, 1963a,b; Fullilove and Jacobson, 1971; Rickoll, 1976; Zalokar and Erk, 1977; Rickoll and Counce, 1980; Stafstrom and Staehelin, 1984a,b; Callaini Anselmi, 1988; Katoh and Ishikawa, 1989). Following nuclear division 13, nuclei at the cell surface are each packaged into a single, columnar epithelial cell during a complex cytokinetic event called cellularization. As in earlier cycles, a bulge in the cell surface becomes prominent during this interphase. However, nuclei begin to elongate, and the bulge fails to relax. Instead, membranes plunge down between adjacent nuclei. Cellularization occurs in two distinct phases. Initially, the rate of membrane invagination is slow. When the leading edge of the membrane (called the furrow canal because of its tear-drop shape) reaches the base of the elongated nucleus, a second, fast phase of membrane invagination begins. Mutants can disrupt either the slow, initial phase, or the fast, final phase of cellularization, suggesting that the two cytokinetic processes are mechanistically, as well as kinetically, distinguishable (Merrill er d., 1988). During the second phase of cellularization, the furrow canals plunge inward to a final depth of -30 pm from the cell surface. Concomitant with this inward movement, the furrow canals elongate laterally (in a plane parallel to the cell surface) to nearly, but not quite completely, occlude the connection between the yolk and the newly formed columnar cells. This small bridge between the yolk and the columnar cells remains until the movements of early gastrulation. The changes in the morphology of the membrane that occur during cellularization are summarized in Fig. 1A and B. C. Gastrulation
Early in gastrulation, the plate of columnar epithelial cells that underlies the pole cells begins to differentiate by thickening, then puckers inward and migrates
82
C
Daniel P. Kiehart
D
FIG. 1 Schematic of membrane architecture and forces for shape change during cytokinesis. (A and B) Successive stages of Drosophilu embryogenesis. M in all panels shows the location of myosin; pm is the plasma membrane, lipid bilayer. (C and D) Cytokinesis in a “generic” cell type, for example, a sea urchin egg. Dense bodies on the plasma membrane anchor actin (A) filaments. Myosin filaments mediate contractility and shortening of the distance between dense bodies, giving rise to a net displacement of the membrane toward the center of the embryo. Arrows represent force vectors. In each case, a net displacement of membrane in, toward the middle of the embryo, is the result of forces that are predominantly directed tangential to the membrane surface. In the Drosophila embryo, the net displacement of membrane in later stages is tangential to the surface and results in the formation of the basal margins of the forming, columnar epithelial cells.
anteriorly, along the dorsal midline. Finally, the plate, with the overlying pole cells, invaginates to form the amnioproctodeal invagination, an extension of the ventral furrow (Sonneblick, 1950). Later in gastrulation, cell shape changes and cell sheet morphogenesis continue. Remodeling of the anterior end of the ventral furrow and of the amnioproctodeal invagination forms the endodermal primordium of the anterior midgut and the posterior midgut, respectively. The germ band retracts, and two sheets of cells from the ventral epithelium growth dorsally along either side of the embryo to meet and seal at the dorsal midline, a process called dorsal closure. Complex morphogenetic movements continue to characterize the development of Drosophilu through later stages of embryogenesis and the formation of larvae, then adults (e.g., Fristrom, 1988). In summary, changes in the shape of the embryo surface begin early in devel-
6. Contractility in Drosophilu Development
83
opment and occur surprisingly quickly. Cellularization is complete by about 3 hr following fertilization; gastrulation ensues and is complete (marked by full germ band extension) by approximately 6 hr; dorsal closure is complete by about 13 hr (Wieschaus and Nusslein-Volhard, 1986). As described below, changes in the shape of the embryo are coincident with profound alterations in the form and function of both the membrane skeleton and the cytoskeleton. Such alterations provide the molecular basis for the cell shape change.
111. CELL SHAPE C H A N G E REQUIRES REMODELING OF THE ACTIN CMOSKELETON
Actin is a prominent component of the Drosophila membrane skeleton and cytoskeleton (the membrane skeleton can be viewed as that subset of the cytoskeleton that is closely apposed to the plasma membrane; essentially, it is the cytoskeleton of the cell or precellular embryo cortex; Kiehart, 1990a) throughout development. Moreover, it changes dynamically with the cell cycle and through early morphogenetic events (Warn and Magrath, 1983; Warn et al., 1984, 1985, 1990; Wam, 1986; Karr and Alberts, 1986; Kellogg et al., 1988; Pesacreta et al., 1989; reviewed in Kiehart, 1990a). Upon arrival of the nuclei at division cycles 9 and 10 (Warn et a f . , 1985; Karr and Alberts, 1986), actin is concentrated in a disklike, cortical cap between the nucleus and the plasma membrane. These cortical actin caps form prior to an observable bulge in the cell surface. Moreover, as nuclear divisions proceed and the surface bulges and relaxes, the morphology of the actin caps also changes cyclically, from dome to flat disk, precisely paralleling the bulged, then flattened, form of the cell surface. During cellularization, the membrane, membrane skeleton, and cytoskeleton are completely remodeled, and again the array of actin filaments in the cortex keeps pace with these complex changes. As membranes advance between adjacent nuclei, actin is a conspicuous component of the membrane skeleton, from the tip of the advancing furrow canal, along the lateral margins of the forming cells, to their apices (in Fig. 1, actin would be localized to all regions of the cortex). Electron microscopy of Drosophila embryos shows a dense meshwork of microfilaments in the embryos that corresponds to actin localization by immunofluorescent methods (e.g., Rickoll, 1976; Katoh and Ishikawa, 1989). The importance of the integrity of actin filaments for a variety of early movements, including the migration of nuclei to the poles and cellularization, comes from drug microinjection experiments. Cytochalasins and/or phalloidin can block these movements when microinjected into living embryos (Zalokar and Erk, 1976; Foe and Alberts, 1983; Warn et al., 1990). Cytochalasins have also implicated actin function in gastrulation in another dipteran, Heteropeza pygm e a (Kaiser and Went, 1987), and imaginal disc morphogenesis in Drosophila
84
Daniel P. Kiehart
(Fristrom and Fristrom, 1975; Fristrom, 1988). Unequivocal interpretation of these experiments is complicated by the complexity of the effect of cytochalasin on actin filaments in vivo (Cooper, 1987). Aphidicolin, a drug that inhibits DNA synthesis and nuclear division, demonstrates that it is not the nuclei per se, but the centrosomes and presumably the array of astral microtubules that they organize, that initiate changes in the actin membrane skeleton. When microinjected into embryos, aphidicolin allows centrosomes to migrate, in the absence of nuclei, to the embryo surface, where they reorganize cortical actin in a pattern comparable to uninjected controls (Raff and Glover, 1989).
N. NON-MUSCLE MYOSINS Myosins are chemomechanical force producers that interact with actin to power cell movements (reviewed in Spudich, 1989; Kiehart, 1990b). Nonmuscle myosin localization in developing embryos suggests that this protein plays a key role in powering changes in the architecture of the cytoskeleton and membrane skeleton during development. Prior to cellularization, myosin is concentrated in the embryo cortex and is localized diffusely throughout the periplasm. As cells are formed, membrane is drawn down, between adjacent nuclei, and myosin is concentrated around the leading edge of the furrow canal (Warn et al., 1980; Young et al., 1991a). In Fig. 1, myosin is schematically shown as localized to the base of furrows (M). Force production for cellularization is likely to occur through the contractility of the interconnected rings of actomyosin and the meshwork of actomyosin between them. While we do not understand the geometry of myosin assemblies on the membrane, in principle, the organization is likely to be similar to that proposed for contractile rings in other cells (Fig. 1C and D; reviewed in Conrad and Schroeder, 1990). During gastrulation, myosin is concentrated at the apical ends of cells that change shape to cause epithelial sheet invagination. In other regions of the embryo, a diffuse, low level of myosin is conspicuous, but enhanced apical staining is absent (Young et al., 1991a). This distribution is consistent with a role for myosin in force production for cell sheet morphogenesis: in computer modeling studies, apical contractions have been shown sufficient to explain cell sheet morphogenesis (e.g., Odell et al., 1981). Moreover, apical constrictions are also concomitant with invaginations of cell sheets in the ventral furrow (Leptin and Grunewald, 1990). A. Native Myosin II
A conventional, non-muscle myosin was characterized structurally and functionally and was identified as the major (possibly the only) conventional (myosin
6 . Contractility in Drosophilu Development
85
II), non-muscle myosin isoform in flies (Kiehart and Feghali, 1986). The native myosin has a long tail, two heads, forms filaments, and moves beads in the Nitellu in vitro motility assay (T.-1. Chen, S . Block, and D. P. Kiehart, unpublished observations). Peptide mapping experiments and immunochemical analysis of the myosin heavy chains distinguished the non-muscle myosin from thoracic muscle isoforms.
B. Myosin II Heavy Chains The gene that encodes the 205-kDa non-muscle myosin heavy chain was cloned and sequenced (Kiehart et al., 1989, 1990a,b; Ketchum et al., 1990). Polyclonal antibodies, specific for the non-muscle isoform, were used to recover a fragment of the non-muscle myosin gene from a Drosophilu genomic DNA library in hgtl 1 , then hybridization methods were used to recover cDNAs that represented a full-length transcript. By a number of structural and functional criteria, the gene cloned represents the non-muscle myosin heavy chain polypeptide that has been characterized, is located at polytene chromosome band 60E9, and is distinct from the sarcomeric myosin heavy chain gene at 36B. Complete sequence analysis of a full-length cDNA reveals a predicted, non-muscle myosin polypeptide that shares extensive homology with a vertebrate non-muscle myosin isoform [72% identity in the myosin head region and 51% identity in the tail, with overall conservation of 61% identity (77% overall conservation allowing conservative substitutions; Ketchum et al., 1990)]. Developmental Northems indicate a peak of transcript accumulation at 4- 12 hr, at early third instar larval and early pupal stages (Kiehart et al., 1989). Sequence analysis of the 5 ’ end of a number of cDNAs revealed that the 5 ‘ end of the gene is differentially spliced to make two transcripts that differ in size by 64 base pairs (Ketchum et al., 1990). An interesting feature is that the longer transcript has a translation start that is 5’ and in frame with the translation start for the major non-muscle myosin heavy chain isoform found in cells. As a consequence, the extended transcript can make an amino-terminal extended polypeptide (Kozak, 1989). Immunoprecipitation experiments using antibodies elicited to a synthetic peptide (that corresponds to the amino-terminal extension) demonstrate that the higher molecular weight isoform is made and is stable in flies. The extension is extremely basic (calculated pf is 10.45) and shows no similarity to protein sequences recorded in the protein data bases. This peptide extension does not seem to target myosin to a particular location in the cell, because localization patterns with the antipeptide antibody are, to a first approximately, identical to localization studies performed with antibodies against whole myosin. It is possible that this sequence modulates myosin ATPase activity or contractility in some way.
86
Daniel P. Kiehart
Affinity-purified antibodies were microinjected into embryos in an attempt to block myosin function and to test for myosin function during early developmental stages (Kiehart et al., 1990; D. Lutz and D. P. Kiehart, unpublished observations). Antibodies fail to diffuse freely throughout the embryo cytoplasm, and, as a consequence, cause local effects at the site of injection. This proves fortuitous, because normal development of uninjected regions of the embryo serves as an important control for embryo viability. Injections of anti-cytoplasmic myosin during preblastoderm stages prevent a normal number of nuclei from forming and from reaching the embryo surface. This suggests a role for myosin in nuclear proliferation and migration. The anti-myosin response is specific (control IgG fractions or buffer alone do not cause the effect) and phenocopies (i.e., mimics) injections of cytochalasin and antibodies directed against a high-molecularweight actin-binding protein (Zalokar and Erk, 1976; Foe and Alberts, 1983; Edgar et al., 1987; Miller et a l . , 1988; K. Miller, personal communication). Injections of anti-myosin at syncytial blastoderm stages disrupt the organization of nuclei at the cell surface and cause the clear periplasm, which normally extends -30 pm inward from the plasma membrane, to bulge farther into the yolk mass. As the embryo develops and cellularization proceeds in regions of the embryo away from the site of injection, cellularization of the region at the site of injection is retarded or fails to occur. These observations suggest a role for myosin in maintaining cortical organization and are consistent with a role for myosin in cellularization. Again, anti-myosin injections phenocopy microinjections of cytochalasin and antibodies against the high-molecular-weight actinbinding protein. To test more directly myosin’s role in cellularization, antimyosin was injected during cellularization. In contrast to injections at earlier stages, the injections had no effect. Similarly, anti-myosin injected prior to the onset of cytokinesis in echinoderm eggs readily blocks cleavage, but, after cytokinesis has begun, comparable injections have no effect on cytokinesis until the next division cycle (Mabuchi and Okuno, 1977; Kiehart et a l . , 1982). In addition, cellularization (cytokinesis) in Drosophila is sensitive to cytochalasin injected prior to its onset, but much higher concentrations of the drug are required once cellularization has begun (Zalokar and Erk, 1976; Lutz and Kiehart, 1990). Observations that antibodies localize myosin to the furrow canals and that antibodies injected prior to the onset of cellularization inhibit cytokinesis are consistent with a role for myosin in cellularization. The antibodies’ inability to affect myosin function when injections are after initiation of cellularization can be explained if myosin was assembled into a complex structure that is no longer sensitive to the effect of antibodies. Such antibody-insensitive states have been observed in vitro (Kiehart et al. 1984a,b; Kiehart and Pollard, 1984). Finally, it is not surprising that antibodies directed against a myosin, injected early on,
6. Contractility in Drosophila Development
87
block cytokinesis: genetic methods, designed to compromise myosin function in Dictyostelium, also block cytokinesis (reviewed in Spudich, 1989). To understand better myosin’s role in powering movements during early development, we have embarked on a genetic analysis of non-muscle myosin heavy chain function (Young et al., 1991b). A standard F2 screen was used to recover recessive, lethal mutations uncovered by a deficiency that removed the gene for cytoplasmic myosin and adjacent loci. Three thousand mutagenized chromosomes were scored by placing them in trans to a 200-kb deficiency (ESI; Cote et al., 1987), and three lethal complementation groups were recovered. One likely encodes non-muscle myosin: two alleles have phenotypes distinguishable on immunoblots of whole-fly proteins probed for non-muscle myosin heavy chain. One allele (1.3)encodes an electrophoretic variant of myosin: a 195-kDa heavy chain band, in addition to the wild-type 205-kDa band, is observed. The second allele (zipttFIO7; Nusslein-Volhard et al., 1984) encodes a myosin heavy chain polypeptide that is either not synthesized or is very unstable: in homozygotic embryos (selected by morphological phenotype, see below), the level of myosin is greatly reduced or absent. These data suggest strongly that the complementation group that includes I .3 and zipttFIO7 encodes the non-muscle myosin heavy chain. Additional data confirm that this complementation group encodes myosin: a cDNA construct whose expression is driven by a heat-shock promoter rescues the mutant phenotype. Analysis of the embryonic phenotypes of this mutation is underway. Because wild-type myosin is initially present in all mutants (the mother fly was heterozygous for the mutation), defects appear only on dilution and/or turnover of the maternally encoded myosin proteins and failure of the zygotically (homozygous) mutant myosin to function properly. As a consequence, early movements that depend on myosin proceed normally. Only when the amount of myosin drops below a certain level are contractile events compromised. The zipper myosin null phenotype (henceforth called zipper because of historical precedence) is pleiotropic and includes two sets of epibolic or cell sheet movements (NussleinVolhard et al., 1984; Cote et al., 1987; Zhao et al., 1988), dorsal closure, and head involution. During normal dorsal closure, epidermal cells from the ventral side of the embryo grow dorsally to enclose the amnioserosa, uncovered during germ band shortening. Head involution is complicated, but also requires changes in the shape of cell sheets. The cellular basis for these changes remains unknown, and may involve several distinct mechanisms, including cell shape changes and rearrangements (Ettensohn, 1985; Fristrom, 1988). Preliminary analysis of cell shape on the ventral side of the embryo during the course of dorsal closure shows a change from isometric, polygonal cells to highly elongate ones. In addition, myosin is particularly concentrated in cells at the leading edge of the cell sheet during dorsal closure. We do not know how myosin functions
88
Daniel P. Kiehart
during dorsal closure, but it is interesting to speculate that myosin contributes to cell shape change either directly, or indirectly, as contractility at the leading edge of the cell sheet during dorsal closure drags the epithelial cell sheet dorsally. Whatever the molecular mechanism of myosin function during morphogenesis, the apical distribution of myosin in the invaginating furrows during early gastrulation events, computer modeling studies that suggest apical contraction can drive cell sheet morphogenesis, and the failure of later cell sheet movements in embryos with mutations in myosin, collectively support the notion that myosin plays a key role for cell sheet morphogenesis during development. A third, neurological phenotype involves failure of proper axon path finding and is consistent with the high concentration of myosin seen in vertebrate axon growth cones and in the Drosophila central nervous system (Zhao et al., 1988; Forscher and Smith, 1988; Goldberg and Burmeister, 1989; Young et al., 1991a). It is likely that these three processes fail in flies mutant for non-muscle myosin because they are the first that require zygotically encoded myosin. The challenge will be to explain the cellular basis of these and other phenotypes that are due to the absence of myosin.
C. Myosin II Light Chains
Two non-muscle myosin light chains have been cloned and partially characterized (Kiehart and Feghali, 1986; Karess et al., 1991). Each of the putative cytoplasmic myosin light chains (M,18 and 16 kDa) was purified from native myosin by preparative sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and was characterized as summarized below (Karess et al., 1991). Two peptides from each light chain were partially sequenced, and the data indicate that the 18- and 16-kDa polypeptides are indeed cytoplasmic myosin light chains and that they are distinct from muscle myosin light chains in Drosophila. Peptide sequence data confirm that LC-I is a regulatory light chain with best sequence match to chicken smooth muscle myosin (Maita et al., 1981; Parker et al., 1985) and that LC-2 is an alkaline light chain with best sequence match to chicken cardiac myosin (Matsuda et al., 1981; Falkenthal et al., 1984). Based on these amino acid sequence data and considerations of Drosophila codon biases (Cherbas et al., 1986), oligonucleotide probes for each of the myosin light chains were constructed and the polymerase chain reaction (PCR) was used to generate product using various cDNAs and cDNA libraries as templates. The PCR product was subcloned, then used to probe on developmental Northern blots. LC-1 - and LC-2-specific probes recognize messages of 1 .O and 1.2 kb, respectively. Message accumulates in a pattern that generally mimics that observed for the myosin heavy chain (see above, Kiehart et al., 1989). Subcloned PCR product was used to screen appropriate Drosophila genomic and
6. Contractility in Drosophiln Development
89
cDNA libraries. Sequences from cDNA clones establish that the genes encode the peptides sequenced, further confirm their identity by sequence similarity to vertebrate light chains, and confirm that they are distinct from the corresponding Drosophila muscle isoforms. Genomic clones are being used to establish a map of the transcription unit and have established the polytene location of the genes that encode LC-I and LC-2 at 5D6 and 5A6, respectively. Single genes appear to encode each light chain. A mutation in the gene that encodes the regulatory light chain, LC-1, was recovered independently in a screen for mitotic defects (Karess et a l . , 1991). In larvae homozygous for a defect in this gene, tentatively called spaghettisquash (sqh), many of the normally diploid imaginal cells become highly polyploid, due to repeated failures of cytokinesis. Other aspects of the mitotic cycle seem unaffected by this mutation. Homozygotes die around the time of pupation, probably due to a failure in the development of imaginal tissues. The only allele recovered to date is caused by the insertion of a P element into an upstream, untranslated region of the transcript. Careful inspection of the phenotype of mutant embryos suggests that the mutant phenotype is not null: the mutant allele over a genetic deficiency that uncovers the region is a pupal lethal with qualitatively, but not quantitatively, the same polyploid mitotic phenotype as the homozygote. The discrepancy between the phenotype of arrest of animals homozygous for a mutation in their heavy chain versus their regulatory light chains can be explained because the P-element allele is not a null and/or the perdurance of maternal light chains is longer than the perdurance of the heavy chains. An essential maternal contribution of LC-I to embryogenesis has been demonstrated, as embryos derived from mutant sqh maternal germ line clones do not hatch, regardless of their genotype. Satisfactory explanation of this discrepancy awaits better characterization of both heavy and light chain mutations and a more complete understanding of the stability of the heavy and light chain polypeptides. Characterization of both the regulatory and alkali light chains is in progress. The function of LC-2 has not been investigated, but, by analogy to vertebrate systems, likely plays a role in stabilizing heavy chain conformation, may contribute to LC- I binding, and likely modulates native myosin function. Because the light chains modulate and stabilize (LC-2) and regulate (LC-1) native myosin function, these studies will extend our analysis to allow a more complete understanding of the mechanism of non-muscle myosin function in Drosophila.
V. SPECTRINS
In Drosophila spectrin polypeptides were first identified in an enriched fraction of high-molecular-weight actin-binding proteins (Dubreuil et al., 1987). By
90
Daniel P. Kiehart
electron microscopy of platinum-shadowed specimens, a conspicuous species in this heterogeneous fraction were molecules with a morphology comparable to vertebrate spectrins. Subsequent characterization (see below) has shown that, like nonerythroid spectrin from other sources (Bennett, 1985, 1990; Marchesi, 1983, they appear to contribute to an anastomosing meshwork that provides mechanical support for the plasma membrane. The distribution of a-spectrin changes in concert with the complex changes in embryo form that occur during early development (Pesacreta et al., 1989). As shown using antibodies specific for the a-spectrin subunit, the spectrin distribution is a subset of the pattern of actin distribution, and, generally, the pattern appears to complement that of myosin. In early embryos, a-spectrin is found concentrated in the cortex (apposed to the inner surface of the plasma membrane), in cytoplasmic islands that surround the rapidly dividing nuclei, and, at a low level and diffusely, throughout the interior of the embryo. As the embryo develops, spectrin accompanies the migrating nuclei to the embryo surface and is further concentrated in the embryo cortex. Like myosin and actin, the spectrin is localized in the supranuclear caps that form on arrival of the nuclei at the cell surface. Interestingly, while double-labeling studies suggest that actin and myosin are recruited to the caps essentially simultaneously, it appears that spectrin is localized to the caps only after the actin and myosin. During mitotic divisions of nuclei at the cell surface, the cytoskeletal caps that include spectrin elongate (during interphase and prophase) and divide (as metaphase and anaphase progress). During cellularization, the pattern of spectrin localization changes dramatically. Spectrin continues to be localized in the embryo cortex (the apical ends of the forming cells) but also concentrates in the forming, lateral margins of the columnar epithelial cells. This cortical staining extends to a position just apical of the furrow canals. During the final phases of cellularization, redistribution of the spectrin is rapid and nearly complete: spectrin leaves the lateral margins of the cells and localizes almost exclusively to the apical boundaries. Thus, in detail, the distribution of a-spectrin contrasts the distribution of cortical myosin, which is focused at the leading edge of the furrow canal, and the distribution of cortical actin overlaps the distribution of both myosin and spectrin. During gastrulation, spectrin remains associated primarily with the apical ends of cells. Especially high concentrations of spectrin are present in the apical ends of cells in the complex folds of the gastrulating embryo. During later stages (germ band extension, ca. 4 hr), the total amount of spectrin in the embryo increases and becomes more uniformly distributed in the cortex of all margins of the cells. Interestingly, there is a significant difference in the overall amount and distribution of spectrin among different tissue types: respiratory tract cells appear to have the highest levels, whereas low levels characterize the intestinal epithelium. This contrasts the high concentration of nonerythroid spectrins found in vertebrate intestinal epithelial cells (reviewed in Mooseker, 1985; Coleman et al., 1989),
6. Contractility in Drosophila Development
91
suggesting that a second isoform of a-spectrin may be present in the intestinal epithelial cells of the fly. The patterns of spectrin localization throughout early embryogenesis, particularly during cellularization and subsequent gastrulation, suggest that a-spectrin in particular, and spectrins in general, are likely to play a key role in stabilizing existing membrane morphologies. Active changes in membrane and cell shape are likely to require interactions between actin and myosin. Drosophila spectrin polypeptides and the genes that encode them have been characterized. Antiserum was made to this heterogeneous constellation of proteins and was subsequently shown to react predominantly with two polypeptides, a-spectrin and a high-molecular-weight, P-spectrin isoform (PH, see below). A conventional spectrin tetramer consisting of 234-kDa a and 226-kDa P heterodimers was purified to homogeneity and characterized (Dubreuil et al., 1987). These polypeptides are spectrins by a number of structural and functional criteria: they form an equimolar complex, have molecular morphology comparable to vertebrate spectrins (two intertwined, elongated strands with a contour length of 180 nm), react with antibodies directed against vertebrate spectrins, and have actin binding activity. The a subunit binds calmodulin in a free calcium ion concentration-dependent fashion (Dubreuil et al., 1987; Byers et al., 1987). Potential spectrin clones, from a Xgtl 1 library of Drosophila head cDNAs, were identified with the antiserum to the heterogeneous actin-binding fraction. Subsequently, cDNA sequences that encode the entire a-spectrin polypeptide were recovered (Byers et al., 1987; Dubreuil et al., 1989). These a-spectrin polypeptides have been characterized further by our colleagues: proteins were made in Escherichia coli, and the calmodulin-binding domain was identified. In addition, calcium binds to a region of a-spectrin whose sequence has a typical EF band motif, suggesting that Drosophila a-spectrin function may be regulated by calcium independently of calrnodulin (Dubreuil et al., 1991a). Standard hybridization methods were used to recover genomic DNA for the locus and the gene was localized to polytene chromosome band 62B. It is likely that this gene is single copy: P-spectrin overlays of Western blots of whole-fly cell homogenates suggest that, if another a-spectrin gene does exist, its product precisely comigrates with the a-spectrin polypeptides that were characterized. Further, at standard and reduced stringency, both genomic Southerns and in situ hybridization to polytene chromosomes suggest that only a single a-spectrin gene exists. However, similar genes may well encode structurally similar molecules that perform parallel function in specific cell types or regions of the cell. Two or more P-spectrins are present in Drosophila. Our colleagues have cloned a conventional P-spectrin, using novel, a-spectrin overlay methods (Byers et al., 1989). The gene for this conventional P-spectrin is at polytene chromosome location 16C. A second, high-molecular-weight P-spectrin isoform (PH) was cloned in the original screen for a-spectrin using the antisera raised against the Drosophilu
92
Daniel P. Kiehart
high-molecular-weight actin-binding protein fraction isolated from non-muscle cells (Byers et al., 1987; Dubreuil et al., 1990b). The cDNAs we recovered encode PH: antibodies, affinity-purified on the PH fusion protein, recognize a single, 405-kDa polypeptide in immunoblots of whole Drosophila cell lysates (human dystrophin was used as a size standard) and the 5-kb cDNA that was recovered recognizes a 13-kb message. Subsequent characterizations using both protein biochemical and molecular biological methods reveal that this protein is a high-molecular-weight isoform of P-spectrin (Dubreuil et al., 1990b). It forms an equimolar complex with, and copurifies to homogeneity with, a-spectrin. The complex forms a spectrin-like tetramer with a contour length significantly longer (250 versus 180nm) than the conventional a-P tetramer. The tetramer binds actin with high affinity. The predicted protein sequence of PH (the cDNA recovered encodes the 5' end of the message) shows sequence identity throughout its length with the conventional 6-spectrin (34% identity). Partial sequence analysis of the amino-terminal end of both the conventional (Byers et al., 1989) and P,-spectrin (Dubreuil et al., 1990b) polypeptides shows that this region of the two Drosophila P-spectrin isoforms shares 56% identity and confirms the striking homology of a 240-amino acid domain with amino-terminal domains of aactins and dystrophin that had been reported elsewhere (Davison and Critchley, 1988). This domain contributes to actin binding in a-actinin and may function in a similar capacity in the other proteins. Lack of sequence conservation (among other actin-binding proteins) downstream of this domain suggests that it functions at least somewhat independently of more carboxy-terminal regions. This pspectrin isoform, or closely related isoforms are present in both muscle and nonmuscle tissues. Together with a recent localization of embryonic transcripts to the visceral mesoderm and to muscle attachment sites ( G . Thomas and D. Kiehart, unpublished observation), and the high molecular weight of this protein, the data suggest that this isofonn of P-spectrin (or closely related isoforms, encoded by a single gene) may be a dystrophin-like molecule. Its true relationship to the dystrophin subfamily of proteins has yet to be established. The gene appears to be single copy and maps to polytene chromosome location 63C,D; a genetic analysis is underway.
VI. SUMMARY
Actin, myosin, and spectrins are key components of the membrane skeleton and cytoskeleton. Their intimate interaction with the plasma membrane in Drosophila is suggested by the immunofluorescent staining patterns described by ourselves and others and by analogy to other systems where the molecular basis of membrane skeleton function has been investigated in detail (reviewed in Bennett, 1985, 1990).
6. Contractility in Drosophila Development
93
Many questions remain. Among them are two of particular interest. Recent experiments suggest that conventional, non-muscle myosins are linked to the cortex and membrane independently of actin (Schroeder and Otto, 1988) and myosins 1 (motor proteins related to the conventional myosin described here) bind phospholipids directly (see other contributions to this symposium and reviews by Adams and Pollard, 1989, and Korn and Hammer, 1990). The molecular basis of the linkage between myosin, actin, and the plasma membrane needs to be elucidated. Targeting of myosin to specific cell margins in specific cells needs to be temporally regulated in a precise fashion. The regulatory light chain of myosin is phosphorylated at a minimum of two distinct sites. Myosin light chain kinase phosphorylates one (serine) residue and stimulates myosin activity (Sellers and Adelstein, 1987). Both protein kinase C and cyclin-p34cdc2 kinase phosphorylate other (serine and/or threonine) residues that suppress myosin activity, probably by inhibiting phosphorylation at the myosin light chain kinase site (e.g., Ikebe et al., 1986, 1987; Pollard er al., 1990; Satterwhite, personal communication). These kinases are in turn subject to regulation by a variety of factors, including the concentration of free Ca2 , CAMP,and cyclins (a function of time in the cell cycle). Clearly, regulation of just myosin function alone is complex. Deciphering how the function of myosin, spectrin, and actin, along with the function of a myriad of other elements of the cytoskeleton are integrated will prove a formidable problem. In Drosophilu, classical genetic and modem molecular genetic methods promise to provide key insight into these problems. Analysis of second-site suppressor mutations, either intragenic or intergenic, promise new insight into the proteins that contribute to contractile function during cell shape changes. Conditional alleles of myosin and spectrin, either recovered naturally or specifically engineered, will allow insight into the function of these proteins during later stages in the life cycle of the fly. Finally, the implementation of transgenic methods will enable us to ask critical questions about structure/function relationships in these proteins in the context of living cells. +
Acknowledgments I thank members of my laboratory group for their help on this manuscript and Jane Salant for her expertise in the preparation of this manuscript. Portions of this paper are excerpted from previously published work (Kiehart and Feghali, 1986; Kiehart el al., 1989, 1990; Kiehart, 1990b). Funding was from the National Institutes of Health (GM33830), the Muscular Dystrophy Association, and the March of Dimes.
References Adams, R. J., and Pollard, T. D. (1989). Membrane-bound myosin-I provides new mechanisms in cell motility. Cell Motil. Cyfoskel. 14, 178-182.
94
Daniel P. Kiehart
Bennett, V. (1985). The membrane skeleton of human erythrocytes and its implications for more complex cells. Annu. Rev. Eiochem. 54, 273-304. Bennett, V. (1990). Spectrin: A structural mediator between diverse plasma membrane proteins and the cytoplasm. Curr. Opin. Cell Eiol. 2, 51-56. Bownes, M. (1975). A photographic study of development in the living embryo of Drosophila melanogaster. J. Embryol. Exp. Morphol. 33, 789-801. Byers, T. J., Dubreuil, R.. Branton, D., Kiehart, D. P., and Goldstein, L. S. B. (1987). Drosophila spectrin. 11. Conserved features are revealed by analysis of cDNA clones and fusion proteins. J . Cell Eiol. 105, 2103-21 10. Byers, T. J., Husain-Chishti, A., Dubreuil, R. R . , Branton, D., and Goldstein, L. S. B. (1989). Sequence similarity of the amino-terminal domain of Drosophilu beta spectrin to alpha actinin and dystrophin. J. Cell Eiol. 109, 1633-1641. Callaini, G., and Anselmi, F. (1988). Centrosome splitting during nuclear elongation in the Drosophila embryo. Exp. Cell Res. 178, 415-425. Campos-Ortega, J. A , , and Hartenstein, V. (1985). “The Embryonic Development of Drosophilu mehogaster, ” pp. 1-227. Springer-Verlag, New York. Chang, X.-J., Edwards, K., Bukata, S., and Kiehart, D. P. (1990). Drosophilu cytoplasmic myosin: Isolation and characterization of cDNA encoding regulatory myosin light chains. In preparation. Cherbas, P., Schultz, R. A,, Koehler, M. M. D., Savakis, C., and Cherbas, P. (1986). Structure of the Eip28/29 gene, an ecdysone-inducible gene from Drosophilu. J. Mol. Eiol. 189, 617-631. Coleman, T. R., Fishkind, D. J., Mooseker, M. S., and Morrow, J. S. (1989). Functional diversity among spectrin isoforms. Cell Motil. Cytoskel. 12, 225-247. Conrad, G. W., and Schroeder, T. E. (1990). “Cytokinesis. Mechanisms of Furrow Formation During Cell Division.” New York Academy of Sciences, New York. Cooper, J. A. (1987). Effects of cytochalasin and phalloidin on actin. J. Cell Biol. 105, 1473-1478. Cote, S., Preiss, A., Haller, J., Schuh, R., Kienlin, A., Seifert, E., and Jackle, H. (1987). The gooseberry-zipper region of Drosophilu: Five genes encode different spatially restricted transcripts in the embryo. EMEO J. 6, 2793-2801. Davison, M. D., and Critchley, D. R. (1988). Alpha-actinins and the DMD protein contain spectrinlike repeats. Cell (Cambridge, Mass.) 52, 159- 160. Dubreuil, R. R., Byers, T. J., Branton, D., Goldstein, L. S. B., and Kiehart, D. P. (1987). Drosophila spectrin. I. Characterization of the purified protein. J . Cell Eiol. 105, 2095-2102. Dubreuil, R. R., Byers, T. J., Sillman, A. L., Bar-Zvi, D., Goldstein, L. S. B., and Branton, D. ( I 989). The complete sequence of Drosophila alpha-spectrin: Conservation of structural domains between alpha-spectrins and alpha-actinin. J. Cell Eiol. 109, 2197-2205. Dubreuil, R. R., Byers, T. J., Stewart, C. T., and Kiehart, D. P. (1990). A p spectrin isoform from Drosophila (PH)is similar in size to vertebrate dystrophin. J . Cell. Eiol. 111, 1849-1858. Dubreuil, R. R., Brandin, E., Sun-Reisberg, J., Goldstein, L. S. B., and Branton, D. (1991). Structure, calmodulin-binding, and calcium-binding properties of recombinant a spectrin polypeptides. J. Eiol. Chem. 266, 7189-7193. Ede, D. A,, and Counce, S. 1. (1956). A cinematographic study of the embryology of Drosophila melanogaster. Wilhelm R o d Arch. Entwicklungsmech. Org. 148, 402-415. Edgar, B. A,, Odell, G. M., and Schubiger, G. (1987). Cytoarchitecture and the patterning offushi faruzu expression in the Drosophila blastoderm. Genes Dev. 1, 1226- 1237. Ettensohn, C. A. (1985). Mechanisms of epithelial invagination. Q. Rev. Eiol. 60, 289-307. Falkenthal, S., Parker, V. P., Mattox, W. W., and Davidson, N. (1984). Drosophila melanogaster has only one myosin alkali light-chain gene which encodes a protein with considerable amino acid sequence homology to chicken myosin alkali light chains. Mol. Cell Eiol. 4, 956-965. Foe, V. E., and Alberts, B. (1983). Studies of nuclear and cytoplasmic behavior during the five mitotic cycles that precede gastrulation in Drosophila embryogenesis. J . Cell Sci. 61, 31-70.
6. Contractility in Drosophila Development
95
Forscher, P., and Smith, S. J. (1988). Actions of cytochalasins on the organization of actin filaments and microtubules in a neuronal growth cone. J. Cell Biol. 107, 1505-1516. French, V., Ingham, P., Cooke, J., and Smith, J., eds. (1988). “Mechanisms of Segmentation Development,” Suppl. 104. Fristrom, D. (1988). The cellular basis of epithelial morphogenesis. A review. Tissue Cell 20, 6 4 690. Fristrom, D., and Fristrom, J. W. (1975). The mechanism of evagination of imaginal discs of Drosophila melanogasrer. I , General considerations. Dev. Biol. 43, 1-23. Fullilove, S. L., and Jacobson, A. G. (1971). Nuclear elongation and cytokinesis in Drosophila montana. Dev. Biol. 26, 560-577. Goldberg, D. J., and Burmeister, D. W. (1989). Looking into growth cones. Trends Neurol. 12,503506. Ikebe, M., Hartshorne, D. J., and Elzinga, M. (1986). Identification, phosphorylation, and dephosphorylation of a second site for myosin light chain kinase on the 20,000-dalton light chain of smooth muscle myosin. J. Biol. Chem. 261, 36-39. Ikebe, M., Hartshorne, D. J., and Elzinga, M. (1987). Phosphorylation of the 20,000-dalton light chain of smooth muscle myosin by the calcium-activated, phospholipid dependent protein kinase: Phosphorylation sites and effects of phosphorylation. J . Biol. Chem. 262, 9569-9573. Ingham, P. W. (1988). The molecular genetics of embryonic pattern formation in Drosophila Nature (London) 335, 25-34. Kaiser, J., and Went, D. F. (1987). Early embryonic development of the dipteran insect Heteropeza pygmaea in the presence of cytoskeleton-affecting drugs. Row’s Arch. Dev. Biol. 196, 356366. Karess, R . E., Chang, X.-J., Edwards, K. A,, Kulkami, S., Aquilera, I . , and Kiehart, D. P. (1991). The regulatory light chain of non-muscle myocin is encoded by spaghetti-squash, a gene required for cytokinesis in Drosophila. Cell (Cambridge, MA) 65, 1177-1 189. Karr, T. L., and Alberts, B. (1986). Organization of the cytoskeleton in early Drosophila embryos. J. Cell Biol. 102, 1494-1509. Katoh, K., and Ishikawa, H. (1989). The cytoskeletal involvement in cellularization of the Drosophila melanogaster embryo. Protoplasma. 150, 83-95. Kellogg, D. R . , Mitchison, T. J., and Alberts, B. M. (1988). Behavior of actin filaments and microtubules in living Drosophila embryos. Development (Cambridge, U . K . ) 103, 675-686. Ketchum, A . S . , Stewart, C. T.. Stewart, M., and Kiehart, D. P. (1990). The complete sequence of the Drosophila non-muscle myosin heavy chain transcript: Conserved sequences in the myosin tail and differential splicing in the 5’ untranslated sequence. Proc. Natl. Acad. Sci. U.S.A. 87, 63 16-6320. Kiehart, D. P. (1990a). The actin membrane skeleton in Drosophila development. Semin. Cell Biol. 1, 325-339. Kiehart, D. P. (1990b). Molecular genetic dissection of myosin heavy chain function. Cell (Cambridge, Mass.) 60, 347-350. Kiehart, D. P., and Feghali, R. (1986). Cytoplasmic myosin from Drosophila. J . Cell Biol. 103, 15 17-1525. Kiehart, D. P., and Pollard, T. D. (1984). Inhibition of Acanthamoeba actomyosin-ll ATPase activity and mechanochemical function by specific monoclonal antibodies. J. Cell Biol. 99, 1024- 1033. Kiehart, D. P., Mabuchi, I., and Inoue, S. (1982). Evidence that myosin does not contribute to force production in chromosome movement. J . Cell Biol. 94, 165-178. Kiehart, D. P., Kaiser, D. A,, and Pollard, T. D. (1984a). Monoclonal antibodies demonstrate limited structural homology between myosin isozymes from Acanthamoeba. J. Cell Biol. 99, 1002-1014. Kiehart, D. P., Kaiser, D. A., and Pollard, T. D. (1984b). Direct localization of monoclonal
96
Daniel P. Kiehart
antibody-binding sites on Acanthamoeba myosin-I1 and inhibition of filament formation by antibodies that bind to specific sites on the myosin-11 tail. J. Cell Biol. 99, 1015-1023. Kiehart, D. P., Lutz, M. S., Chan, D., Ketchum, A. S., Laymon, R., Nguyen, B., and Goldstein, L. S. B. (1989). Identification of the gene for fly non-muscle myosin heavy chain: Drosophila myosin heavy chains are encoded by a gene family. EMBO J. 8, 913-922. Kiehart, D. P., Ketchum, A., Young, P., Lutz, D., Alfenito, M. R., Chang, X.-J., Awobuluyi, M., Pesacreta, T. C., Inoue, S., Stewart, C. T., and Chen, T.-L. (1990). Contractile proteins in Drosophila development. Ann. N . Y . Acad. Sci. 582, 233-251. Kom, E. D., and Hammer, J. A. (1990). Myosin I . Curr. Opin. Cell Biol. 2, 57-61. Kozak, M. (1989). The scanning model for translation: An update. J. Cell Biol. 108, 229-241. Leptin, M.. and Grunewald, B. (1990). Cell shape changes during gastrulation in Drosophila. Development (Cambridge, U . K . ) 110, 73-84. Lutz, D. A., and Kiehart, D. P. (1990). Disruption of cortical organization and nuclear position in Drosophila embryos by microinjection of antimyosin. J. Cell Biol. (submitted for publication). Mabuchi, I., and Okuno, M. (1977). The effect of myosin antibody on the division of starfish blastomeres. J. Cell Biol. 74, 251-263. Mahowald, A. P. (1963a). Ultrastructural differentiation during formation of the blastoderm in the Drosophila melanogaster embryo. Dev. Biol. 8, 186-204. Mahowald, A. P. (1963b). Electron microscopy of the formation of the cellular blastoderm in Drosophila melanogaster. Exp. Cell Res. 32, 457-463. Maita, T., Chen, 1.. and Matsuda, G. (1981). Amino-acid sequence of the 20000-molecular-weight light chain of chicken gizzard-muscle myosin. Eur. J. Biochem. 117, 417-424. Marchesi, V. (1985). Stabilizing infrastructure of cell membranes. Annu. Rev. Cell Biol. 1,531-561. Matsuda, G., Maita, T., and Umegane, T.(1981). The primary structure of LI light chain of chicken fast skeletal muscle myosin and its genetic implication. FEBS Lett. 126, I I 1-113. Menill, P. T., Sweeton, D., and Wieschaus, E. (1988). Requirements for autosomal gene activity during precellular stages of Drosophila melanogaster. Development (Cambridge, U . K . ) 104, 495-509. Miller, K. G., Field, C. M., and Alberts, B. M. (1988). Subsets of actin filaments have different roles in embryonic organization: The distribution of actin-associated proteins in the cortex of early Drosophila embryos. J. Cell Biol. 107, 252a. Mooseker, M. S. (1985). Organization, chemistry, and assembly of the cytoskeletal apparatus of the intestinal brush border. Annu. Rev. Cell Biol. 1, 209-241. Nusslein-Volhard, C., Wieschaus, E., and Kluding, H. (1984). Mutations affecting the pattern of the larval cuticle in Drosophila melanogaster. Wilhelm Roux's Arch. Dev. Biol. 193, 267-282. Odell, G. M., Oster, G., Alberch, P., and Burnside, B. (1981). The mechanical basis of morphogenesis. I. Epithelial folding and invagination. Dev. Biol. 85, 446-462. Parker, V. P., Falkenthal, S . , and Davidson, N. (1985). Characterization of the myosin light-chain-2 gene of Drosophila melanogaster. Mol. Cell Biol. 5, 3058-3068. Pesacreta, T., Byers, T. J., Dubreuil, R. R . , Kiehart, D. P., and Branton, D. (1989). Localization of spectrin and actin during development of the Drosophila embryo. J. Cell Biol. 108, 1697- 1709. Pollard, T. D., Sattenvhite, L., Cisek, L., Corden, J., Sato, M., and Maupin, P. (1990). Actin and myosin biochemistry in relation to cytokinesis. Ann. N . Y . Acad. Sci. 582, 120-146. Raff, J. W., and Glover, D. M. (1989). Centrosomes and not nuclei initiate pole cell formation in Drosophila embryos. Cell (Cambridge, Muss.) 57, 61 1-619. Rickoll, W. L. (1976). Cytoplasmic continuity between embryonic cells and the primitive yolk sac during early gastrulation in Drosophila melanogaster. Dev. B i d . 49, 304-3 10. Rickoll, W. L., and Counce, S. J. (1980). Morphogenesis in the embryo of Drosophila melanogaster-Germ band extension. Wilhelm Rowr's Arch. Dev. Biol. 188, 163- 177.
6. Contractility in Drosophila Development
97
Schroeder, T. E., and Otto, J. J. (1988). lmmunofluorescent analysis of actin and myosin in isolated contractile rings of sea urchin eggs. 2001.Sci. 5 , 713-725. Sellers, I. R., and Adelstein, R. S . (1987). Regulation of contractile activity. In “Enzymes” (P, D. Boyer, ed.), 3rd ed., Vol. 18, pp. 381-418. Academic Press, Orlando, Florida. Sonneblick, B. P. (1950). The early embryology of Drosophilu melanoguster. In “Biology of Drosophila” (M. Demerec, ed.), pp. 62-167. Wiley, New York. Spudich, J. (1989). In pursuit of myosin function. Cell Regul. 1, 1 - 1 1 . Stafstrom, J. P., and Staehelin, L. A. (1984a). Dynamics of the nuclear envelope and of nuclear pore complexes during mitosis in the Drosophila embryo. Eur. J . Cell Biol. 34, 179-189. Stafstrom, J. P., and Staehelin, L. A. (1984b). Are annulate lamellae in the Drosophilu embryo the result of overproduction of nuclear pore components? J. Cell Biol. 98, 699-708. ’hrner, F. R., and Mahowald, A. P. (1976). Scanning electron microscopy of Drosophilu embryogenesis. 1 . The structure of the egg envelopes and the formation of the cellular blastoderm. Dev. Biol. 50, 95-108. Turner, F. R., and Mahowald, A. P. (1977). Scanning electron microscopy of Drosophilu embryogenesis. 2. Gastrulation and segmentation. Dev. Biol. 57, 403-416. Warn, R. M. (1986). The cytoskeleton of the early Drosophila embryo. J. Cell Sci. Suppl. 5 , 311328. Warn, R. M., and Magrath, R. (1982). Observations by a novel method of surface changes during the syncytial blastoderm stage of the Drosophila embryo. Dev. Biol. 89, 540-548. Warn, R. M., and Magrath, R. (1983). F-actin distribution during the cellularization of the Drosophila embryo visualized with FLphalloidin. Exp. Cell Res.143, 103-1 14. Warn, R. M., Bullard, B., and Magrath, R. (1980). Changes in the distribution of cortical myosin during the cellularization of the Drosophilu embryo. J. Ernbryol. Exp. Morphol. 57, 167- 176. Warn, R. M., Magrath, R., and Webb, S. (1984). Distribution of F-actin during cleavage of the Drosophila syncytial blastoderm. J . Cell B i d . 98, 156- 162. Warn, R. M., Smith, L., and Warn, A. (1985). Three distinct distributions of F-actin occur during the division of polar surface caps to produce pole cells in Drosophilu embryos. J. Cell Biol. 100, 1010- 1015.
Warn, R. M., Warn, A,, Planques, V., and Robert-Nicoud, M. (1990). Cytokinesis in the early Drosophilu embryo. In “Cytokinesis: Mechanisms of Furrow Formation During Cell Division” (G. W. Conrad and T. E. Schroeder, eds.), pp. 260-272. New York Academy of Sciences, New York . Wieschaus, E., and Nusslein-Volhard, C. (1986). Looking at embryos. In “Drosophila: A Practical Approach” (D. B. Roberts, ed.), pp. 199-227. IRL Press, Oxford. Young, P., Pesacreta, T., and Kiehart, D. P. (1991a). Dynamic changes in the distribution of the cytoplasmic myocin during Drosophila embryogenesis. Development (Cambridge, U . K .J 111, 1-14. Young, P., Ketchum, A, , and Kiehart, D. P. (1991b). Morphogenesis requires non-muscle myocin heavy chain function. Submitted for publication. Zalokar, M., and Erk, 1. (1976). Division and migration of nuclei during early embryogenesis of Drosophilu melanoguster. J. Microsc. Biol. Cell. 25, 97- 106. Zalokar, M., and Erk, I. (1977). Phase-partition fixation and staining of Drosophilu. Stain Technol. 52, 89-95. Zhao, D.-B., Cote, S., pdhnig, F., Haller, J., and Jackle, H. (1988). Zipper encodes a putative integral membrane protein required for normal axon patterning during Drosophilu neurogenesis. EMBO J . 7 , 1115-1 119.
This Page Intentionally Left Blank
CHAPTER 7
Dominant Mutations of Cytoskeletal Proteins in Xenopus Embryos Jan L. Christian, Gregory M. Kelly, and Randall T. Moon Department of Pharmacology University of Washington School of Medicine Seattle, Washington 98195
1. 11. 111. IV.
Introduction The Role of Membrane Skeleton Protein 4.1 during Embryogenesis Dissecting the Functions of Virnentin during Embryonic Development Discussion References
1. INTRODUCTION During embryonic development, the cytoskeleton is believed to play an essential role in mediating cellular morphogenesis. For example, cellular shape change and the motility of embryonic cells are brought about primarily by two filament systems, the microtubules and microfilaments. In contrast, the morphogenetic role of other cytoskeletal components, including intermediate filaments and membrane skeleton proteins, is poorly understood. Intermediate filaments comprise a complex multigene family of more than 30 unique proteins, all of which have the capacity to assemble into 8-10 nm filaments (Steinert and Roop, 1988). Individual proteins are expressed in a developmentally regulated and cell-type-specific pattern, suggesting that each type of subunit fulfills a unique role in cellular differentiation. Vimentin exhibits the widest distribution of all cytoplasmic intermediate filaments, being the predomiCurrent Topics in Membranes, Volume 38 Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.
99
100
Jan L. Christian er al.
nant subunit in many differentiated cells of mesenchymal origin, in most cultured cells regardless of origin, as well as in some differentiating cells prior to the appearance of the cell-type-specific subunit (reviewed in Steinert and Roop, 1988; Lazarides, 1982). For example, vimentin is the major subunit found in certain cells of the developing and mature frog retina (Szaro and Gainer, 1988) and in all fibroblastic cells (Bennet et al., 1978). In the developing central nervous system, neurofilament expression is preceded by that of vimentin, and both types of intermediate filaments coexist within the same cell for a short period of time (Szaro and Gainer, 1988). In addition, vimentin is coexpressed with the cell-type-specific intermediate filament protein in some fully differentiated cell types, such as with glial acid fibrillary protein in astrocytes (Schnitzer et al., 1981). During embryogenesis, vimentin is among the earliest of intermediate filament subunits to be expressed. In Xenopus laevis, these filaments are reported to be a component of the germ plasm of oocytes and have been detected in cells of early embryos (Godsave et ul., 1984; P. Tang et al., 1988). This remains a matter of debate, since others have found no evidence for vimentin in the frog oocyte or in embryos prior to neurulation, at which time expression is restricted to a population of cells lining the lateral margins of the neural tube (Franz et al., 1983; Henmann et al., 1989; Dent et al., 1989). A large body of data exists regarding the expression and biochemical characteristics of intermediate filaments, yet the in vivo functions of these cytoskeletal elements remain elusive. In cultured cells, disruption of vimentin filaments by microinjection of antibodies has no effect on cell division, morphology, or motility (Gawlitta et al., 1981; Klymkowsky, 1981; Lin and Feramisco, 1981; Tolle et al., 1986), leading to speculation that intermediate filaments may function at the level of tissues, organs, or whole animal rather than at the level of the single cell (Klymkowsky et al., 1989). The membrane skeleton, consisting of a meshwork of proteins associated with the inner surface of the plasmalemma, may also play some role in the cellular interactions required for normal development. Band 4.1 is one of the major structural proteins of the erythrocyte membrane skeleton, along with ankyrin and spectrin. In erythroid cells, protein 4.1 stabilizes spectrin-actin interactions, binds phospholipids and integral membrane proteins, and is responsible, in part, for providing structural integrity to the plasma membrane (for reviews, see Bennett, 1985; Marchesi, 1985). Numerous immunologically cross-reactive forms of protein 4.1 have been identified in a variety of erythroid and nonerythroid tissues, including lens, brain, granulocytes, fibroblasts, and platelets (Aster et al., 1986; Cohen et al., 1982; Granger and Lazarides, 1984; Goodman et al., 1984; reviewed in Moon and McMahon, 1985). These forms, a result of tissue-specific alternative mRNA processing (Ngai et al., 1987; Conboy et al., 1988; T. K. Tang et al., 1988a,b, 1990), differ not only in size and relative abundance, but also in their intracellular localization, being found in the nucleus
7. Mutations of Xenopus Cytoskeletal Proteins
101
(Leto et al., 1986) and along stress fibers (Cohen et al., 1982) as well as in the membrane skeleton. In X . laevis, a single species of protein 4.1 transcript is expressed in eggs and this transcript persists at a fairly constant level throughout early development (Giebelhaus et al., 1988). Two isoforms of protein 4.1 can be detected on Western blots of proteins extracted from X . laevis tadpoles. Immunoreactive peptides exist in a variety of embryonic neural tissues, in skeletal muscle, and in some epithelial tissue (Giebelhaus et al., 1988), but the highest level of expression is seen in the developing retina (Spencer et al., 1990). Although the function of protein 4.1 is well documented for human erythrocytes, its roles in nonerythroid and/or embryonic cells remain obscure. As with vimentin, protein 4.1 may affect interactions at the tissue or organ level, making it difficult to decipher its functions by studying individual cells. The ability to disrupt cytoskeletal elements within cells of X . laevis embryos provides a means for studying the function(s) of these proteins within the context of the intact, developing organism.
II. THE ROLE OF MEMBRANE SKELETON PROTEIN 4.1 DURING EMBRYOGENESIS Protein 4.1, along with other membrane skeleton proteins, is believed to be involved in establishing and maintaining specialized membrane-cytoskeletal domains in nonerythroid cells (reviewed by Moon and McMahon, 1985). The importance of differential localization of membrane components for proper functioning and interaction of embryonic cells has not previously been investigated. As a step toward examining the role of protein 4.1 during development, antisense RNA was used to perturb the expression of this cytoskeletal element in X . laevis embryos (Giebelhaus et al., 1988). Recombinant plasmids containing an antisense gene were microinjected into fertilized eggs, leading to expression of RNA complimentary to endogenous 4.1 transcripts at the mid-blastula transition. This results in a dramatic reduction in expression of protein 4.1, which is apparent at the level of mRNA as well as protein. Embryos receiving the antisense plasmid are smaller, have malformed organs, and demonstrate a dose-dependent increase in mortality. Eggs injected with sublethal amounts of plasmid develop into tadpoles exhibiting a specific retinal defect in that the normal interdigitation of the photoreceptor outer segment with the pigment epithelium is lacking, leading to a greatly expanded interphotoreceptor space. This phenotype, which is not seen in embryos expressing an analogous sense plasmid, correlates with a reduced level of immunoreactive protein 4. I in the retina relative to control levels. Subsequent to these experiments, an investigation of the normal pattern of protein 4.1 expression in X . laevis embryos revealed that immunoreactive protein
102
Jan L. Christian et al.
4.1 is expressed during specific periods of development (Spencer et al., 1990). Retinal levels of protein 4.1 are observed to increase sharply at a developmental timepoint coinciding with the initial elongation and interdigitation of photoreceptor outer segments with pigment epithelial cells, and the 4.1 is initially restricted to the photoreceptor outer segments. These data, taken together with the antisense phenotype, lend credence to the hypothesis that protein 4.1 plays a critical role in initiating and/or stabilizing the interaction of these cell types in the developing retina. Several alternative approaches are currently being used to study the role of protein 4.1 during early amphibian development. First, synthetic protein 4.1 RNA is being microinjected into fertilized eggs in order to examine the developmental consequences of increasing the endogenous pool. A second approach involves the introduction of mutant protein 4.1 polypeptides into cells of embryos in an attempt to inhibit the function of the resident wild-type protein (Herskowitz, 1987). Protein 4.1 is an attractive candidate for this approach since distinct functional domains of the protein have been identified, including membrane-binding and spectrin-actin-binding regions (Leto and Marchesi, 1984). Mutant proteins lacking, for example, the membrane-binding domain may continue to interact with spectrin and actin but in a dysfunctional manner. Because the amount of spectrin and actin within a cell is limited, such mutant proteins would competitively inhibit the essential constructive interactions between endogenous protein 4.1 and these membrane skeleton proteins, thus generating a dominant mutant phenotype which may provide clues to the normal functions of this protein.
111. DISSECTING THE FUNCTIONS OF VlMENTlN DU RING EMBRYONIC DEVELOPMENT Although the role of vimentin during embryogenesis is not clear, biochemical and morphological studies support a number of hypotheses. In most cell types, the vimentin filament network stretches from the plasma membrane to the nucleus, leading to the postulate that vimentin is involved in centering and/or anchoring the nucleus (Lehto et al., 1978). Another hypothesis, based on the observed association of vimentin and desmin with the Z lines of assembling muscle fibers, states that these filaments are required for proper lateral organization and registration of myofibrils during myogenesis (Gard and Lazarides, 1980; Lazarides, 1982). Finally, immunoreactive vimentin is reported to be a component of the germ plasm in oocytes of X. laevis (Godsave et al., 1984; P. Tang et al., 1988), supporting a role for this cytoskeletal element in the aggregation or migration of germ plasm during early cleavage stages. To test the hypothesis that vimentin is involved in one or more of these processes during embryogenesis,
7. Mutations of Xenopus Cytoskeletal Proteins
103
several independent approaches were used to create dominant mutations of vimentin in embryos of X. faevis. Following the successful use of antisense RNA to inhibit embryonic expression of protein 4.1, an identical approach was used in initial attempts to perturb the expression of vimentin. A partial-length X. laevis vimentin cDNA was cloned, in both orientations, into the expression vector pEMSV-CAT, which employs the murine sarcoma virus long terminal repeat to drive expression of downstream sequences (Giebelhaus et af., 1988). Antisense constructs injected into fertilized eggs were expressed following the mid-blastula transition, but had no effect on the level of endogenous vimentin mRNA. New constructs were made employing more efficient promoters, such as the cytoskeletal actin promoter (Harland and Misher, 1988), to increase the concentration of antisense RNA within cells. In addition, multiple 3' and 5' restriction fragments of the vimentin cDNA were tested as templates in order to minimize the potential for formation of unfavorable secondary structure in the antisense RNA. Microinjection of these constructs into fertilized eggs resulted in high-level expression of the plasmid. The amount of antisense RNA was clearly in excess of the endogenous vimentin message and yet expression of vimentin was not noticeably reduced. As an alternative, independent approach to studying the functions of vimentin during embryogenesis, we have bypassed the normal temporal and spatial regulation of vimentin expression by microinjecting synthetic vimentin RNA into fertilized frog eggs. In addition, two mutant forms of vimentin which lack domains proposed to be required for filament formation were expressed in X. faevis embryos by in vitro transcription of deletion mutant cDNAs and injection of the resulting transcripts into zygotes. These mutant polypeptides were designed to interact with, and inhibit the assembly of, endogenous subunits when overexpressed in vimentin-containing cells. The first mutant, termed VAN49, lacks 68% of the non-a-helical aminoterminal domain, including the majority of sites phosphorylated by protein kinase C and protein kinase A (Evans, 1988a,b, 1989; Geisler et af., 1989; Ando et a f . , 1989). Previous studies implicate the amino terminus as being a requirement for filament formation in vitro (Traub and Vorgias, 1983, 1984; Moon and Lazarides, 1983). In addition, site-specific phosphorylation of amino-terminal serine and threonine residues is believed to regulate filament disassembly (Inagaki et al., 1987, 1988; Chou et al., 1989) or redistribution during mitosis (Lamb et af., 1989; Escribano and Rozengurt, 1988). Thus, these mutant subunits may be assembly incompetent and/or may inhibit the assembly or disassembly of endogenous vimentin subunits during the cell cycle. The second vimentin mutant, designated V5AC-myc, is missing 19 amino acids from the a-helical core of the protein. The deleted region includes an amino acid sequence (TYRKLLEGE) which is present within the carboxyterminal rod domain of nearly all intermediate filament proteins studied to date.
104
Jan L. Christian et al.
This highly conserved region is believed to be a critical determinant of intermediate filament structure, a view supported by the recent deletion analyses of a human epidermal keratin (Albers and Fuchs, 1987, 1989) and a neurofilament subunit (Chin et al., 1989). In creating the VSAC-myc cDNA, deleted sequence was replaced with an in-frame oligonucleotide encoding a 10-amino acid epitope of human c-myc. This peptide is the minimal sequence recognized by the monoclonal antibody 9E10 (Evan et al., 1985; Munro and Pelham, 1987), thus allowing one to distinguish immunologically between endogenous vimentin subunits, which are recognized by vimentin antiserum, and the myc-tagged mutant, which is recognized by both anti-vimentin and the monoclonal anti-myc antibody. Microinjection of synthetic wild-type or deletion mutant vimentin RNAs leads to high-level expression of the respective polypeptides. Translation products of each RNA can be detected on immunoblots within 30 min after injection and the resultant polypeptides persist at levels greater than endogenous throughout the tailbud (VSAC-myc) or tadpole (V5 and V5AN49) stages. Whole-mount analysis (Dent et al., 1989) of endogenous vimentin expression, using an anti-frog vimentin antiserum, shows that this intermediate filament is first detectable during neurulation, at which time expression is restricted to a subpopulation of cells located at the lateral margins of the neural tube. In contrast, following injection of wild-type or mutant RNAs, expression of immunoreactive polypeptides is detected within most cells of early blastula stage embryos. When embryos overexpressing either wild-type (V5) or amino-terminal mutant (V5AN49) subunits are dissociated into single cells at the gastrula stage, extensive vimentin filament networks can be visualized throughout the cytoplasm by indirect immunofluorescence. A similar analysis of cells derived from uninjected gastrulae demonstrates a complete absence of vimentin staining at this stage. Thus, exogenously introduced wild-type vimentin is capable of assembling into higher order structures within cell types which do not normally express this protein. More surprisingly, an intact amino terminus is not essential for filament formation in vivo. Staining of dissociated cells expressing V5AC-myc, using either anti-vimentin or anti-myc antibodies, was dramatically different, appearing as punctate dots throughout the cytoplasm. When V5 and VSAC-myc synthetic RNAs are mixed and coinjected into fertilized eggs, an identical staining pattern is observed, despite evidence that the quantity of wild-type protein introduced into these cells is sufficient to form filaments which are detectable by indirect immunofluorescence when expressed in the absence of mutant subunits. Thus, VSAC-myc subunits are capable of acting in a dominant fashion to inhibit assembly of wildtype vimentin polypeptides. Embryos overexpressing either wild-type or dominant mutant vimentin subunits exhibit grossly normal development. To address the question of whether this overexpression affects the morphology of developing muscle, skeletal mus-
7. Mutations of Xenopus Cytoskeletal Proteins
105
cle was examined in thick sections of tailbud and tadpole stage embryos expressing either V5 or VSAC-myc. Immunostaining demonstrates a clear overexpression of each protein in localized regions of differentiating muscle, yet, in each case, myofibers show a normal striated appearance and are arranged in parallel arrays as in control embryos. To test the hypothesis that the presence of an extensive vimentin cytoskeleton in cell types normally devoid of these filaments influences nuclear positioning or morphology, Hoechst-stained nuclei within intact V5-expressing gastrulae were examined. No apparent differences in nuclear centration or shape were observed between cells expressing immunoreactive vimentin and those lacking detectable filaments. As previously stated, endogenous vimentin expression was first observed during neurulation, at which time the small size of individual cells precludes resolution of cellular detail by whole-mount immunocytochemical methods. In addition, vimentin expression is restricted to a select group of cells so that identification of these cells by immunofluorescent staining following dissociation of embryos was not feasible. For these reasons, it was not possible to discern unambiguously whether mutant subunits were coexpressed in cells along with endogenous vimentin subunits and, if so, whether endogenous subunits assembled into filaments in these cells. The observation that overexpression of V5AC-myc in most cells of early embryos has no effect on rates of cell division, on cell morphology, or on cell survival suggests that vimentin filaments, if present during these early stages, as detected by other researchers, do not perform essential functions during the initial stages of embryogenesis.
N. DISCUSSION The inability to directly mutate and thus inactivate, specific genes in higher eukaryotes has been a major stumbling block to dissecting their functions. In X . laevis, a variety of approaches have been described for disrupting gene function at the level of RNA or protein. One approach involves the introduction of antisense molecules into cells (Green et a l . , 1986). In immature frog oocytes, injection of antisense RNA has been used to interfere with expression of injected genes (Melton, 1985; Harland and Weintraub, 1985) and with endogenous ribosomal protein Ll (Wormington, 1986). A more effective strategy has since been developed in which antisense oligodeoxynucleotides are introduced into cells, mediating RNase H-like cleavage and destruction of transcripts. This approach has been used to eliminate a variety of endogenous RNAs in X . laevis oocytes, including those encoding histone H4, Vgl, the 70-kDa heat-shock protein, c-mos, calmodulin, and xglv7 (Shuttleworth and Coleman, 1988; Sagata et a l . , 1988; Dash et a l . , 1987; Kloc et
106
Jan L. Christian et al.
al., 1989). The ability to mature and fertilize frog oocytes in vitro, following ablation of specific RNAs, makes this a feasible approach for studying the function of maternally encoded information during postfertilization development (Shuttleworth et al., 1988; Kloc et al., 1989). Unfortunately, the use of antisense RNA to inhibit gene function later in development has for the most part been unsuccessful. This failure was initially attributed to the presence of a doublestranded RNA unwinding activity, which exists at high levels in fertilized eggs and embryos prior to late blastula (Rebagliati and Melton, 1987; Bass and Weintraub, 1987). The subsequent discovery that this activity not only unwinds but also covalently modifies RNA hybrids, changing adenosine residues to inosine, complicates the issue (Bass and Weintraub, 1988; Wagner et al., 1989). Substitution of inosine (the coding equivalent of guanine) for adenosine would alter the coding capacity of transcripts, most likely resulting in the generation of a variety of nonfunctional proteins. In addition, this modifying activity may be involved in targeting RNAs for rapid degradation in vivo (Kimmelman and Kirschner, 1989). Thus, the modification of double-stranded RNA would be expected to enhance, rather than block, the antisense effect and the cause for failure of some antisense experiments remains a mystery. Injection of recombinant plasmids encoding antisense genes into fertilized eggs offers an alternative means of introducing inhibitory RNAs into the developing organism (Giebelhaus et al., 1988). In X . laevis, injected DNA sequences replicate to varying degrees, are partitioned into the nuclei of most (up to 85%) of the descendant blastomeres, and the majority are expressed at the midblasula stage of development (Etkin and Pearman, 1987; Shiokawa et al., 1989). Unlike the situation in transgenic mice, in which expression of foreign genes can be regulated temporally or spatially by relying on tissue-specific or inducible promoters (Palmiter and Brinster, 1985), in the frog, expression of injected DNA is highly mosaic and strict tissue-specific expression is generally not observed (Etkin and Pearman, 1987). To date, mosaicism in DNA expression has been monitored only at the level of protein expression, for example, by injection of plasmids encoding reporter genes such as P-galactosidase into fertilized eggs, following which embryos are allowed to develop to various stages and are then analyzed for the distribution of enzyme activity. Deficiencies in posttranscriptional processing or transport of RNA from the nucleus, as well as cell-typespecific differences in RNA/protein stability may play some role in the observed mosaicism, but the relative contribution of these factors, if any, has not yet been assessed. In the future, it would be valuable to determine whether the plasmid is transcribed more uniformly than is apparent at the level of protein expression. If so, the observed mosaic expression of plasmids in frogs may not be a major deterrent to their use in antisense experiments, as has previously been assumed, since one mechanism by which antisense RNA has been shown to act involves the formation of nuclear RNA hybrids which are unable to enter the cytoplasm
7 . Mutations of Xenopus Cytoskeletal Proteins
107
(Kim and Wold, 1984). Thus, transcription is required for an antisense effect, but processing, transport, and stability of transcripts may not be essential factors in some experiments. A second approach to identifying the function of gene products involves disruption of gene activity at the protein level. The cloned gene is altered so that it encodes a mutant polypeptide capable of inhibiting the wild-type gene product when overexpressed in a cell (Herskowitz, 1987).This approach has been used to confer a dominant negative phenotype in a variety of cultured cells (Albers and Fuchs, 1987, 1989; Clegg et al., 1987), but has not yet been highly exploited in embryos of X. laevis. The ability to introduce mutant polypeptides into embryos by microinjection of synthetic RNA into fertilized eggs (Moon and Christian, 1989) increases the appeal and likely success of this approach, since injected RNA tends to segregate fairly uniform and thus can be overexpressed throughout the embryo during early development. Ribozymes, which are RNAs capable of catalyzing RNA cleavage reactions, provide a final potential tool with which to inhibit the expression of specific genes. General rules have been established allowing the design of RNA enzymes which can be targeted to virtually any transcript (Haseloff and Gerlach, 1988). Such RNA molecules have been shown to be effective in mediating destruction of specific RNAs in the test tube (Cotten et al., 1989) and in cultured cells (Sarver et al., 1990), but the high ratio of ribozyme to substrate required for cleavage in a complex cellular environment has impeded their use in vivo with the following exception. Cotten and Birnstiel (1989) employed a modified tRNA gene to drive high-level expression of a specifically designed ribozyme in frog oocytes. The target transcript, u7snRNA, was efficiently cleaved, thus demonstrating that ribozymes can provide a feasible approach to gene inactivation in vivo in selected cases. The advantages of X. laevis as an experimental model system include its easy husbandry, rapid morphogenesis, and the accessibility and large size of eggs and embryos, which facilitate physical manipulations such as microinjections. While the inapplicability of classical genetics in X. laevis may have limited its use in the past, the ability to introduce dominant mutations at the level of RNA or protein, rather than at the level of the gene, demonstrates the potential of this organism as a tool for studying the role of various proteins in vertebrate development.
References Albers, K., and Fuchs, E. (1987). The expression of mutant epidermal keratin cDNAs transfected in simple epithelial and squamous cell carcinoma lines. J . Cell B i d . 105, 791-806. Albers, K . , and Fuchs, E. (1989). Expression of mutant keratin cDNAs in epithelial cells reveals mechanisms for initiation and assembly of intermediate filaments. J . Cell B i d . 108, 14771493. Ando, S . , Kazushi, T., Gonda, Y., Sato, C . , and Inagaki, M. (1989). Domain- and sequence-specific
108
Jan L. Christian et al.
phosphorylation of vimentin induces disassembly of the filament structure. Biochemistry 28, 2974-2979. Aster, J. C., Brewer G. 1.. and Maisel, H. (1986). The 4.1-like proteins of bovine lens: Spectrinbinding proteins closely related in structure to the red blood cell protein 4.1. J. Cell Biol. 103, 115-122. Bass, B. L., and Weintraub, H. (1987). A developmentally regulated activity that unwinds RNA duplexes. Cell (Cambridge. Mass.) 48, 607-613. Bass, B. L., and Weintraub, H. (1988). An unwinding activity that covalently modifies its doublestranded RNA substrate. Cell (Cambridge, Mass.) 55, 1089- 1098. Bennet, G. S., Fellin, S. A,, and Holtzer, H. (1978). Immunofluorescent visualization of 100 8, filaments in different cultured chick embryo cell types. Diflerentiation 12, 71-82. Bennett, V. (1985). The membrane skeleton of human erythrocytes and its implications for more complex cells. Annu. Rev. Biochem. 54, 273-304. Chin, S . S . M., Hashim, G. A,, and Liem, R. K. H. (1989). Expression of truncated neurofilament proteins in transiently transfected cells. J. Cell Biol. 109, 69a. Chou, Y.,Rosevear, E., and Goldman, R. D. (1989). Phosphorylation and disassembly of intermediate filaments in mitotic cells. Proc. Natl. Acad. Sci. U.S.A. 86, 1885-1889. Clegg, C. H., Corell, L. A,, Cadd, G. G . , and McKnight, G. S . (1987). Inhibition of intracellular CAMP dependent protein kinase using mutant genes of the regulatory type I subunit. J. Biol. Chem. 262, 13111-13119. Cohen, C. M., Foley, S. F., and Korsgren, C. (1982). A protein immunologically related to erythrocyte band 4. I is found on stress fibers of non-erythroid cells. Nature 299, 648-650. Conboy, J. G., Chan, J., Mohandas, N., and Kan, Y. W. (1988). Multiple protein 4. I isoforms produced by alternative splicing in human erythroid cells. Proc. Narl. Acad. Sci. U.S.A. 85, 9062-9065. Cotten, M., and Birnstiel, M. L. (1989). Ribozyme, antisense RNA and antisense DNA inhibition of U7 small nuclear ribonucleoprotein-mediatedhistone pre-mRNA processing in vitro. Mol. Cell. Biol. 9, 4479-4487. Dash, P., Lotan, I . , Knapp, M., Kandel, E. R., and Goelet, P. (1987). Selective inhibition of mRNAs in vivo: Complementary oligodeoxynucleotides promote RNA degradation by an RNase H-like activity. Proc. Natl. Acad. Sci. U.S.A. 84, 7896-7900. Dent, J. A,, Poulson, A. G., and Klymkowsky, M. W. (1989). A whole-mount immunocytochemical analysis of the expression of the intermediate filament protein vimentin in Xenopus. Development (Cambridge, U . K . J l 0 5 ,61-74. Escribano, J., and Rozengurt, E. (1988). Cyclic AMP increasing agents rapidly stimulate vimentin phosphorylation in quiescent cultures of Swiss 3T3 cells. J. Cell. Physiol. 137, 223-234. Etkin, L. D., and Pearman, B. (1987). Distribution, expression and germ line transmission of exogenous DNA sequences following microinjection into Xenopus laevis eggs. Development (Cambridge, U . K . ) 99, 15-23. Evan, G. I., Lewis, G . K., Ramsay, G., and Bishop, J. M. (1985). Isolation of a monoclonal antibody specific for human c-myc proto-oncogene product. Mol. Cell Biol. 5 , 3610-3616. Evans, R. M. (1988a). The intermediate filament proteins vimentin and desmin are phosphorylated in specific domains. Eur. J. Cell Biol. 46, 152-160. Evans, R. M. (1988b). Cyclic-AMP dependent protein kinase-induced vimentin filament disassembly involves modification of the N-terminal domain of intermediate filament subunits. FEES Lett. 234, 73-78. Evans, R. M. (1989). Phosphorylation of vimentin in mitotically selected cells. In vitro cyclic-AMP independent kinase and calcium-stimulated phosphatase activities. J. Cell Biol. 108, 67-78. Franz, J. K., Gall, L., Williams, M. A,, Picheral, B., and Franke, W. W. (1983). Intermediate sized
7. Mutations of Xenopus Cytoskeletal Proteins
109
filaments in a germ cell: Expression of cytokeratins in oocytes and eggs of the frog Xenopus. Proc. Natl. Acad. Sci. U.S.A. 80, 6254-6258. Gard, D. L., and Lazardies, E. (1980). The synthesis and distribution of desmin and vimentin during myogenesis in vitro. Cell (Cambridge, Mass.) 19, 263-275. Gawlitta, W., Osborn, M., and Weber, K. (1981). Coiling of intermediate filaments induced by microinjection of a vimentin-specific antibody does not interfere with locomotion and mitosis. Eur. J. Cell Biol. 26, 83-90. Geisler, N., Hatzfeld, M., and Weber, K. (1989). Phosphorylation in vitro of vimentin by protein kinases A and C is restricted to the head domain. Eur. J. Eiochem. 183, 441-447. Giebelhaus, D. H., Eib, D. W., and Moon, R. T. (1988). Antisense RNA inhibits expression of membrane skeleton protein 4. I during embryonic development of Xenopus. Cell (Cambridge, Mass.) 53, 601-615. Godsave, S. J. Anderton, B. H., Heasman, J., and Wylie, C. C. (1984). Oocytes and early embryos of Xenopus laevis contain intermediate filaments which react with anti-mammalian vimentin antibodies. J. Embryol. Exp. Morphol. 83, 169-187. Goodman, S. R., Casoria, L. A., and Coleman, D. B. (1984). Identification and location of brain protein 4.1. Science 224, 1433-1436. Granger, B., and Lazarides, E. (1984). Membrane skeleton protein 4.1 of avian erythrocytes is composed of multiple variants that exhibit tissue-specific expression. Cell (Cambridge, Mass.) 37, 595-607. Green, P. J., Pines, O., and Inouye, M. (1986). The role of antisense RNA in gene regulation. Annu. Rev. Biochem. 55, 569-597. Harland, R.,and Misher, L. (1988). Stability of RNA in developing Xenopus embryos and identification of a destabilizing sequence in TF IIIA messenger RNA. Development (Cambridge, U.K.) 102, 1-12. Harland, R., and Weintraub, H. (1985). Translation of mRNA injected into Xenopus oocytes is specifically inhibited by antisense RNA. J. Cell Biol. 101, 1094-1099. Haseloff, J., and Gerlach, W. L. (1988). Simple RNA enzymes with new and highly specific endoribonuclease activities. Nature (London) 334, 585-591. Hemnann, H., Fouquet, B., and Franke, W. W. (1989). Expression of intermediate filament proteins during development of Xenopus laevis. I. cDNA clones encoding different form so vimentin. Development (Cambridge, U . K . ) 105, 279-298. Herskowitz, I. (1987). Functional inactivation of genes by dominant negative mutations. Nature (London 329, 219-222. Inagaki, M., Nishi, Y.,Nishizawa, K., Matsuyama, M., and Sato, C. (1987). Site-specific phosphorylation induces disassembly of vimentin filaments in vifro. Nature (London) 328, 649652. Inagaki, M., Gonda, Y.,Matsuyama, M., Nishizawa, K., Nishi, Y.,and Sato, C. (1988). Intermediate filament reconstitution in vifro. J. Biol. Chem. 263, 5970-5978. Kim, S. K., and Wold, B. J. (1984). Stable reduction of thymidine kinase activity in cells expressing high levels of antisense-RNA. Cell (Cambridge, Mass.) 42, 129-138. Kimmelman, D., and Kirschner, M. (1989). Antisense mRNA directs the covalent modification of the transcript encoding fibroblast growth factor in Xenopus oocytes. Cell (Cambridge, Mass.) 59, 687-696. Kloc, M., Miller, M. Carrasco, A. E., Eastman, E., and Etkin, L. (1989). The maternal store of the xlgv7 mRNA in full grown oocytes is not required for normal development in Xenopus. Development (Cambridge, U.K.)107, 899-907. Klymkowsky, M. W. (1981). Intermediate filaments in 3T3 cells collapse after the intracellular injection of a monoclonal anti-intermediate filament antibody. Nature (London)291, 249-25 1.
110
Jan L. Christian et al.
Klymkowsky, M. W., Bachant, J. B., and Domingo, A. (1989). Functions of intermediate filaments. Cell Motil. Cytoskel. 14, 309-322. Lamb, N. J. C . , Femandez, A,, Feramisco, J. R., and Welch, W. J. (1989). Modulation of vimentin containing intermediate filament distribution and phosphorylation in living fibroblasts by the cyclic-AMP dependent protein kinase. J. Cell Biol. 108, 2409-2422. Lazarides, E. (1 982). Intermediate filaments: A chemically heterogeneous, developmentally regulated class of proteins. Annu. Rev. Biochem. 51, 219-250. Lehto, V. P., Virtanen, I., and Kurki, P. (1978). Intermediate filaments anchor the nuclei in monolayers of cultured human fibroblasts. Nature (London 272, 175-177. Leto, T. L., and Marchesi, V. T. (1984). A structural model of human erythrocyte protein 4. I . J. Biol. Chem. 259, 4603-4608. Leto, T. L., Pratt, B. M., and Madri, J. A. (1986). Mechanisms of cytoskeletal regulation: Modulation of aortic endothelial cell protein band 4. I by the extracellular matrix. J. Cell. Physiol. 127, 423-43 I . Lin, J. J. C., and Feramsico, J. R. (1981). Disruption of the in vivo distribution of intermediate filaments in fibroblasts through the microinjection of a specific monoclonal antibody. Cell (Cambridge, Mass.) 24, 185-193. Marchesi, V. T. (1985). Stabilizing infrastructure of cell membranes. Annu. Rev. CellBiol. 1, 531561. Melton, D. A. (1985). Injected anti-sense RNAs specifically block mRNA translation in vivo. Proc. Natl. Acad. Sci. U.S.A. 82, 144-148. Moon, R. T., and Christian, 1. L. (1989). Microinjection and expression of synthetic mRNAs in Xenopus embryos. Technique 1, 76-89. Moon, R. T., and Lazarides, E. (1983). Canavanine inhibits vimentin assembly but not its synthesis in chicken embryo erythroid cells. J. Cell Biol. 97, 1309-1314. Moon, R. T., and McMahon, A. P. (1985). Composition and expression of spectrin-based membrane skeletons in non-erythroid cells. BioEssays 7, 159- 164. Munro, S., and Pelham, H. R. B. (1987). A carboxy-terminal signal prevents secretion of luminal ER proteins. Cell (Cambridge, Mass.) 48, 899-907. Ngai, J., Stack, J. H., Moon, R. T., and Lazarides, E. (1987).Regulated expression of multiple chicken erythroid membrane skeletal protein 4. I variants is governed by differential RNA processing and translational control. Proc. Natl. Acad. Sci. U.S.A. 84, 4432-4436. Palmiter, R. D., and Brinster, R. L. (1985). Transgenic mice. Cell (Cambridge, Mass.) 41, 343345. Rebagliati, M. R., and Melton, D. A. (1987). Antisense RNA injections in fertilized frog eggs reveal an RNA duplex unwinding activity. Cell (Cambridge, Mass.) 48, 599-605. Sagata, N., Oskarsson, M., Copeland, T., Brumgaugh, J., and Vande Woude, G. F. (1988). Function of c-mos proto-oncogene product in meiotic maturation in Xenopus oocytes. Nature (London) 335, 519-525. Sarver, N., Cantin, E. M., Chang, P. S., Zaia, J. A,, Ladne, P. A., Stephens, D. A,, and Rossi, J. J. (1990). Ribozymes as potential anti-HIV- 1 therapeutic agents. Science 247, 1157- 1264. Schnitzer, J., Franke, W. W., and Schachner, M. (1981). Immunocytochemical demonstration of vimentin in astrocytes and ependymal cells of developing and adult mouse nervous system. J. Cell Biol. 90, 435-437. Shiokawa, K., Fu, Y., Nakakura, N., Tashiro, K., Sameshima, M., and Hosokawa, K. (1989). Effects of injection of exogenous DNAs on gene expression in early embryos and coenocytic egg cells of Xenopus laevis. Roux's Arch. Dev. Biol. 198, 78-84. Shuttleworth, J., and Coleman, A. (1988). Antisense oligonucleotide directed cleavage of mRNA in Xenopus oocytes and eggs. EMBO J. 7 , 427-434. Shuttleworth, J., Matthew, G . , Dale, L., Baker, C . , and Coleman, A. (1988). Antisense deox-
7. Mutations of Xenopus Cytoskeletal Proteins
111
yribonucleotide-directedcleavage of maternal mRNAs in Xenopus oocytes and embryos. Gene 12, 267-275. Spencer, M. B., Giebelhaus, D. H., Kelly, G. M., Bicknell, J., Florio, S. K., Milam, A. H., and Moon, R. T. (1990). Membrane skeleton protein 4.1 in developing Xenopus: Expression in post mitotic cells of the retina. Dev. Biol. (in press). Steinert, P. M., and Roop, D. R. (1988). Molecular and cellular biology of intermediate filaments. Annu. Rev. Biochem. 57, 593-625. Szaro, B. G., and Gainer, H. (1988). Immunocytochemical identification of non-neuronal intermediate filament proteins in the developing Xenopus luevis nervous system. Dev. Brain Res. 43, 207-224. Tang, P.,Sharpe, C. R., Mohan, T. I., and Wylie, C. C. (1988). Vimentin expression in oocytes, eggs and early embryos of Xenopus luevis. Development (Cambridge, U . K . ) 103, 279-287. Tang, T. K., Leto, T. L., Correas, I., Alonso, M. A., Marchesi, V. T., and Benz, E. I., Jr. (1988a). Selective expression of an erythroid-specific isoform of protein 4.1. Proc. Nurl. Acud. Sci. U.S.A. 85, 3713-3717. Tang, T. K.,Leto, T. L., Marchesi, V. T., and Benz, E. J., Jr. (1988b). Expression of specific isoforms of protein 4.1 in erythroid and non-erythroid tissues. In “Molecular Biology of Hemopoiesis” (M. Tavassoli, E. D. Zanjani, J. L. Ascensao, N. G. Abraham, and A. S . Levine, eds), pp. 81-95. Plenum, New York. Tang, T. K., Qin, Z., Leto, T., Marchesi, V. T., and Benz, E. J., Jr. (1990). Heterogeneity of mRNA and protein products arising from the protein 4.1 gene in erythroid and non-erythroid tissues. J. CellBiol. 110, 617-624. Tolle, H. G., Weber, K., and Osborn, M. (1986). Microinjection of monoclonal antibodies to vimentin, desmin and glial fibrillary acidic protein in cells which contain more than one intermediate filament type. Exp. Cell Res. 162, 462-474. Traub, P, and Vorgias, C. E. (1983). Involvement of the N-terminal polypeptide of vimentin in the formation of intermediate filaments. J. Cell Sci. 63, 43-67. Traub, P., and Vorgias, C. E. (1984). Differential effect of arginine modification with 1,2-cyclohexanedione on the capacity of vimentin and desmin to assemble into intermediate filaments and to bind to nucleic acids. J. Cell Sci. 65, 1-20. Wagner, R. W., Smith, J. E.. Cooperman, B. S . , and Nishikura, K.(1989). A double-stranded RNA unwinding activity introduces structural alterations by means of adenosine to inosine conversions in mammalian cells and Xenopus eggs. Proc. Natl. Acad. Sci. U.S.A. 86, 2647-2651. Wormington, W. M. (1986). Stable repression of ribosomal protein L1 synthesis in Xenopus oocytes by microinjection of antisense RNA. Proc. Nutl. Acud. Sci. U.S.A. 83, 8639-8643.
This Page Intentionally Left Blank
CHAPTER 8
The Animal Models of Duchenne Muscular Dystrophy: Windows o n t h e Pathophysiological C o n s e q u e n c e s of Dystrophin Deficiency Eric P. Hoffman and Jose Rafael M. Gorospe Departments of Molecular Genetics and Biochemistry University of Pittsburgh School of Medicine Pittsburgh, Pennsylvania 15261
I . Introduction The Animal Models of Duchenne Dystrophy: Biochemical and Genetic Homology Human Duchenne Dystrophy: Progressive Histopathology Leading to Muscle Wasting The mdr Mouse: Nonprogressive Histopathology Resulting in Hypertrophy The xmd Dog: Rapidly Progressive Histopathology Leading to Muscle Wasting The Dystrophin-Deficient Cat: Semiprogressive Histopathology with No Loss of Muscle Fibers VII. Phase I. The Primary Cellular Consequence of Dystrophin Deficiency: Generalized Leakage of the Plasma Membrane Buffered by the Syncytial Cytoplasm'? VIIl. Phase I Conclusion: An Integrated Model Featuring Ca2+ IX. Phase 11: Progressive Pathology and Clinical Weakness Specific to Humans and Dogs X . The Development of Progressive Histopathology: Could Basic Fibroblast Growth Factor Have a Major Role? XI. Conclusion: An Integrated Model and Its Consequences on the Development of Therapeutics References 11. 111. IV. V. VI.
I. INTRODUCTION Duchenne muscular dystrophy is considered one of the most important human genetic diseases for two reasons. First, the Duchenne muscular dystrophy gene Currenr Topics in Membranes, Volume 38 Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.
113
114
Eric P. Hoffman and Jose Rafael M.Gorospe
has a sporadic mutation rate an order of magnitude higher than most other diseases. The high mutation rate results in a large proportion of new, sporadic cases, with no previous family history of the disorder, and, coupled with the hemizygous expression of the disease, makes Duchenne dystrophy one of the most common genetic diseases. Second, the clinical progression of Duchenne dystrophy is devastating; affected boys may appear normal until about age 4 when they begin to have trouble being as active as their friends. Progressive muscle wasting and weakness ensue, leaving patients wheelchair-bound by age 11 years. Respiratory or cardiac failure results in death before age 30. The cloning of the Duchenne muscular dystrophy gene and the identification of the normal protein product have led to a thorough understanding of the genetic and primary biochemical basis of the disease (Monaco et af., 1986; Burghes et al., 1987; Hoffman et al., 1987b). The high mutation rate seems to be the result of the extraordinarily large size of the gene; the gene is simply a very large target for mutation (Koenig et al., 1987; Burmeister et al., 1988). Sixty-five percent of patients have deletion mutations of the gene (Koenig et af., 1987), while an additional 6% have duplication mutations (Hu et al., 1990). Many of the deletion mutations originate in just 2 or 3 of the more than 70 introns; these apparent “hot spots” may reflect only the large size of the introns rather than sequence-specific mutation sites, but this needs more thorough investigation. The 30% of patients who do not have detectable deletion/duplication mutations are presumed to have point mutations that affect mRNA splicing or protein translation, but these putative point mutations have not been identified. All mutations apparently disrupt the gene in a way that precludes production of stable dystrophin molecules; 100% of Duchenne patients lack dystrophin (< 3% normal levels) regardless of the type of mutation (Hoffman et al., 1988a) (Fig. 1). Becker muscular dystrophy, a clinically milder form of Duchenne dystrophy, is characterized by dystrophin abnormalities rather than total lack of the protein (Monaco et af.,1988; Hoffman et al., 1989). Becker-type abnormalities are due to amino acid deletions/duplications within the dystrophin protein, reduced quantities of dystrophin, or both (Hoffman et af., 1989; Beggs et al., 1991). The dystrophin protein has been most extensively studied in skeletal muscle, where it is thought to form a network on the cytoplasmic face of the plasma membrane, and possibly the contiguous t-tubules (see Hoffman and Kunkel, 1989; Salviati et al., 1989; Bornemann and Schmalbruch, 1991) (Fig. 2). While best studied in skeletal muscle, dystrophin seems to be present in most excitable and contractile cells (Miyatake et al., 1990; Chang et al., 1989). Currently, there are two general hypotheses about the cellular function of dystrophin, neither necessarily exclusive of the other. Dystrophin may serve to stabilize the plasma membrane, or it may anchor or modulate the integral membrane proteins to
8. Duchenne Muscular Dystrophy
115
FIG. 1 Dystrophin deficiency in human Duchenne muscular dystrophy. Shown are cryosections of muscle biopsies from an adult with a mild type of limb-girdle muscular dystrophy (A, B) and a 5 year-old boy with Duchenne muscular dystrophy (C, D). B and D are the DIC photographs corresponding to the immunofluorescence in A and C. Dystrophin is visualized at the periphery of each muscle fiber in the limb-girdle muscle (A), but is completely deficient in the Duchenne dystrophy muscle (C). Bar, 500 pm. The muscle biopsy used in A/B was kindly referred by Corrado Angelini, Clinica Neurologica, Padova, Italy, and the biopsy in C / D by David VanDyke, Blodgett Memorial Medical Center, Grand Rapids, Michigan.
which it binds. Both hypotheses fit the earlier “membrane hypothesis” of Duchenne dystrophy (Rowland, 1980). Several features of the genetics and biochemistry of Duchenne dystrophy have facilitated the identification of animal models (Fig. 3). First, the dystrophin gene seems to have been very highly conserved through vertebrate evolution. By using
Eric P. Hoffman and Jose Rafael M. Gorospe
116
FIG. 2 Hypothetical subcellular organization of the dystrophin protein. The dystrophin protein molecule is composed of four domains, marked as A,B,C, and D. It is thought that the functional dystrophin molecule is an antiparallel homodimer, as shown in the diagram; however, this has yet to be shown directly. The central domain, B , is thought to form an extended rod composed of triplehelical coiled coils, similar to analogous domains in the spectrins and a-actinins. The functional length of the dystrophin molecule has been shown to be 125 nm by immunoelectron microscopy studies (Watkins er al., 1988; Cullen era/. , 1990). and these same studies have shown that dystrophin is probably organized as a meshwork beneath the plasma membrane. Additional studies have shown that dystrophin is not itself a transmembrane protein (Koenig er a/., 1988), yet is very tightly associated with the plasma membrane (Hoffman ef a / ., 1 9 8 7 ~Knudson ; er a/., 1988). most likely via strong interactions with other transmembrane proteins, as shown in the figure. Recent strides have been made in identifying these dystrophin-associated transmembrane proteins (Ervasti efa/., 1990).
antibodies or cloned genes from either mouse or human, the dystrophin gene or protein, or both, have been identified in mouse (Hoffman et al., 1987a; Chamberlain et al., 1987; Sugita et al., 1988), chicken (Hoffman et al., 1988b; Lemaire et al., 1988), rabbit (Hoffman et al., 1987c), torpedo (Chang et al., 1989), dog (Cooper et al., 1988a), cat (Carpenter et al., 1989), guinea pig, frog,
-Dystrophin is very highly conserved through at least vertebrate evolution. -Dystrophin deficiency appears completely specific for Duchenne muscular dystrophy in humans. -The dystrophin gene is probably X-linked in all placental mammals.
Thus, any animal which 1. Shows dystrophin deficiency in its muscle. 2. Inherits dystrophin deficiency as an X-linked trait.
is likely an animal model of Duchenne muscular dystrophy. FIG. 3 Rationale used for identification of the animal models of Duchenne dystrophy. Listed are the expected characteristics of animal models for Duchenne dystrophy. This rationale has allowed the identification of true animal models for Duchenne dystrophy based only on biochemical data and inheritance patterns. The finding of animal models has been further facilitated by the apparent high frequency of mutations of the dystrophin gene in placental mammals, similar to the high mutation rate seen in humans.
8. Duchenne Muscular Dystrophy
117
and rat (E. P. Hoffman, unpublished). Second, dystrophin deficiency is specific for Duchenne muscular dystrophy in humans (Hoffman et al., 1987a; Arahata et al., 1988; Pate1 et al., 1988; Bonilla et al., 1988; Zubrzycka-Gaarn et al., 1988). It is therefore probably safe to assume that dystrophin deficiency is similarly specific for “Duchenne dystrophy” in animals. Third, X-linked genes in humans have been found to be X-linked in all placental mammals (Ohno’s hypothesis) (Ohno et al., 1964). Therefore, a true animal model of Duchenne dystrophy should also exhibit X-linked inheritance due to mutations within the X-linked dystrophin gene. Given these three observations, it can be assumed that any animal that shows dystrophin deficiency in muscle and inherits dystrophin deficiency as an X-linked recessive trait is a true animal model of human Duchenne dystrophy (Fig. 3). Identification of the actual dystrophin gene mutation in a possible model is the most compelling evidence of genetic homology of the animal and human diseases. It is true that the mutations responsible for 30% of human Duchenne patients are still unknown. Nevertheless, it is safely assumed that these 30% of human patients do indeed have Duchenne dystrophy based on biochemical (and clinical) criteria. Therefore, the identification of the specific gene mutation in a putative animal model cannot yet be a prerequisite for classification as an animal model. Identification of animal models of Duchenne dystrophy has also been facilitated by the likelihood that the dystrophin gene is as large in most animals as it is in humans (Hoffman et al., 1987a; Meng, et al., 1991). The sporadic mutation rate of animal dystrophin genes should therefore be as high as it is in humans, and there might be a similarly high frequency of “Duchenne dystrophy” in animals. This in fact appears to be the case because all of the animal models have been picked up “off the street”; all were sporadic cases that were identified without any systematic search. With all these assumptions and tools at hand, three species of dystrophindeficient mammals have been identified since the discovery of dystrophin 3 years ago, namely, dogs (Cooper et al., 1988a), cats (Carpenter et al., 1989), and mice (Bulfield et al., 1984; Hoffman et al., 1987b; Sicinski et al., 1989). Additional alleles of the dog disease (B. Cooper, personal communication; J. R. Gorospe, E. Hoffman, and G. Cardinet, unpublished observations) and cat disease (F. Gaschen and E. Hoffman, unpublished observations) have been identified since the initial reports, and three additional alleles of the original sporadic mouse mutation have been induced by mutagens (Chapman et al., 1989). The mutageninduced mouse alleles appeared at a frequency considerably higher than expected (Chapman et al., 1989), supporting the hypothesis that mutability of the mouse gene may be as high as that of the human gene, as outlined above. Within each species, the disease appears relatively uniform; for example all mdx alleles show similar clinical phenotypes (Chapman et al., 1989), and the
118
Eric P. Hoffman and Jose Rafael M . Gorospe
phenotype is not dependent on the genetic background of the mouse (E. Hoffman, unpublished). However, in different species the disease is not uniform. Indeed, the clinical differences between the mouse and human diseases led early authors to believe that the conditions were probably not genetically or biochemically homologous (Heilig et al., 1987; Brockdorff et al., 1987), a view that was later disproved by molecular analysis. The clinical differences between the several animal models and the human disease have created a state of discomfort among neuromuscular investigators. How could the same genetic and biochemical abnormality result in the drastic differences of clinical phenotype? This review first describes how each animal model was identified, and then presents the similarities and dissimilarities of the dystrophin-deficient species at the genetic, biochemical, cellular, histopathological, and clinical levels. The differences between species can be rationalized in light of the similarities.
11. THE ANIMAL MODELS OF DUCHENNE DYSTROPHY: BIOCHEMICAL AND GENETIC HOMOLOGY To qualify as a true animal model of human Duchenne dystrophy, the animal must have an X-linked myopathy associated with dystrophin deficiency in skeletal muscle. A possible mouse model was first identified as an X-linked myopathy in England by Bulfield et al., (1984), who called it mak. The mak mutation appeared sporadically and was found quite fortuitously in a search for genetic abnormalities of red blood cells. In this search, the authors measured serum enzyme activities, including creatine kinase (CK), in an inbred C57B110 breeding colony. They found that some animals inherited very high CK values in a pattern consistent with X-linkage. Dystrophin deficiency in mak skeletal muscle was later found and reported with the first description of the dystrophin protein (Hoffman et al., 1987b) (Fig. 4), a finding verified many times (Sugita et al., 1988; Hoffman et al., 1988b; Bonilla et al., 1988; Zubrzycka-Gaarn et al., 1988; Nicholson et al., 1989; Miyatake et al., 1989). The specific gene mutation in the mdr mouse is a “nonsense” type of point mutation that causes premature termination of translation, which is predicted to result in a severely truncated and apparently nonfunctional dystrophin (Sicinski et al., 1989). The mutant protein seems to be rapidly degraded by the muscle cell as there has been no truncated forms of dystrophin found in these animals (Hoffman et al., 1988b). The first cat model also appeared sporadically; in this case on Nantucket Island off the coast of Massachusetts. The affected cats were two male littennates; they were identified and characterized by J. Carpenter of the Angel1 Memorial Hospital (MSPCA) in Boston (Carpenter et al., 1989). Disease transmission in the cat family was consistent with X-linked inheritance, and muscle showed histological
8. Duchenne Muscular Dystrophy
119
Dystrophin
Fast ATPase
+
1
2
1
2
FIG. 4 Dystrophin deficiency in the mdr mouse. Shown is immunoblot analysis of dystrophin in membrane preparations from normal mouse muscle (B6; lane I ) and mdr mouse muscle (mdr; lane 2). Dystrophin is seen to be completely deficient in the mdr muscle. This figure is taken from Hoffman et al. (1988b) with permission of the publisher.
changes of myopathy associated with dystrophin deficiency (Fig. 5). The specific mutation has not been identified, but these cats also fulfill the basic criteria for being a true animal model of human Duchenne dystrophy. Both affected cats had been neutered, making it impossible to carry out exhaustive clinical and genetic studies. However, a second litter of fertile dystrophin-deficient cats has been identified in Florida, and these are currently being bred (F. Gaschen and E. Hoffman, unpublished data). Both litters of cats were not derived from any specific breed; they were simply outbred domestic short-hair cats. The Florida litter was traced by Gaschen back to a supermarket parking lot where they were being given away as kittens. The dog model appeared sporadically on a golden retriever background, and
120
Eric P. Hoffman and Jose Rafael M. Gorospe
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
FIG. 5 Dystrophin deficiency in the cat model and xmd dog model. Shown is dystrophin immunoblot analysis of skeletal muscle from the cat and dog models of Duchenne muscular dystrophy. Both the Nantucket cats and the xmd dogs show complete dystrophin deficiency. This figure is taken from the original experiment shown in both Carpenter et al. (1989) and Cooper eta!. (1988a).
was first identified in upstate New York by De La Hunta (1983), and then characterized in North Carolina by Kornegay et al. (1988) and at Cornell University by Cooper et al. (1988a). The canine myopathy is associated with dystrophin deficiency (Fig. 5), with X-linked inheritance (Cooper et al., 1988b). The mutation causing the myopathy in these retrievers has been mapped within the dystrophin gene by linkage analysis (Cooper et al., 1988b), but the mutation has not yet been identified. Thus, the mdx mouse, and the dystrophin-deficient cats and dogs are all true animal models that are genetically and biochemically homologous to human Duchenne dystrophy. In contrast, several other possible animal models have normal dystrophin and an autosomal inheritance pattern so they are not animal models of Duchenne dystrophy. Among the excluded models are dystrophic
8. Duchenne Muscular Dystrophy
121
chickens (Hoffman et al., I988b), dystrophic mice (dy/dy), cardiomyopathic/myopathic hamsters, and muscular dysgenic mice (mdg/mdg) (E. Hoffman, unpublished). There has been a report of an X-linked myopathy in rabbits (Engel, 1976). The clinical and pathological description of the rabbits seemed consistent with the other true animal models of Duchenne dystrophy; however neither live animals nor tissue are remaining for dystrophin gene or protein studies, making verification impossible (E. Hoffman, unpublished). There are many pathological and clinical similarities and dissimilarities between each of the animal models and human Duchenne dystrophy. Most of the similarities appear early in the disease, while many of the dissimilarities occur consequent to the progressive aspects of the disease. It is therefore possible to compare the models in two “phases” of the disease. Phase I includes the primary consequences of dystrophin deficiency, and is similar in all dystrophin-deficient species at an early age. Phase I1 represents the progressive aspects of the disease and differs in the different dystrophin-deficient species. We will first describe the pathology of each model.
111. HUMAN DUCHENNE DYSTROPHY: PROGRESSIVE HISTOPATHOLOGY LEADING T O MUSCLE WASTING
Normal human skeletal muscle contains dystrophin from fetal life onward, while in Duchenne dystrophy dystrophin is lacking from fetal life onward. Dystrophin-deficient fetal human muscle shows only occasional eosinophilic hypercontracted fibers in an otherwise normal muscle (Emery, 1977; Bertorini et al., 1984). These hypercontracted fibers become more frequent and obvious in neonatal muscle (Hudgson et a/.,1967; Guibaud et al., 1981; Barmada, 1990), but the pathology is still minimal (Fig. 6A,B). Despite the mild histopathology of Duchenne muscle in early life, the fibers leak massive amounts of cytoplasmic contents into the extracellular space. As a Duchenne patient gets older, histopathology changes (Fig. 6). Degeneration and regeneration of fibers become progressively more evident. There is an active proliferation of connective tissue (“fibrosis”) between muscle fibers. Fibrosis is a generic reaction of most tissues to injury. However, in Duchenne dystrophy, the proliferation of connective tissue between each muscle fiber (endomysial fibrosis) is more active than seen in other neuromuscular disorders (Hantai et al., 1985). Paralleling the progressive fibrotic replacement of the muscle, more and more fibers are seen in some stage of the degenerationregeneration cycle (Fig. 6). Moreover, there is a gradual failure of regeneration, and a gradual loss of muscle fibers. By the time a patient is 10 or 11 years old, most of the muscle tissue has been replaced by fibro-fatty connective tissue.
122
Eric P. Hoffman and Jose Rafael M. Gorospe
FIG. 6 Progressive histopathology of human Duchenne muscular dystrophy. Shown are micrographs of hematoxylin-eosin-stained frozen sections of muscle biopsies from patients as described (bars, 500 pm): A and B show the histopathology of two 2-year-old boys, one with Duchenne muscular dystrophy (B) and the other with an unrelated myopathy (A). Despite the limited pathology in Duchenne muscular dystrophy at this age (B), the serum CK levels of the patient shown in B were grossly elevated. Note the occasional hypertrophied fibers in the Duchenne muscle (B). As the vast majority of the muscle fibers appear intact and functional, the fiber hypertrophy seen is unlikely to be compensatory, and instead possibly represents a primary consequence of dystrophin deficiency. The muscle biopsy shown in A was kindly referred by T. Cabone, Springfield, Massachusetts, and the biopsy in B was kindly referred by C. Angelini, of University of Padova, Italy. C shows the histopathology of muscle from a 5-year-old patient with Duchenne muscular dystrophy. The pathology is much more dramatic than that seen in younger patients (B), with marked variation in fiber size and marked proliferation of connective tissue between fibers (endomysial fibrosis). This muscle biopsy was kindly referred by D. VanDyke, Blodgett Memorial Medical Center. D is the histopathology of a 9-year-old patient with Duchenne dystrophy. Note the extensive fibrotic replacement of much of the muscle. Many of the remaining muscle fibers are atrophic, and appear to have aborted attempts at regeneration. Other fibers are markedly hypertrophic. The fiber hypertrophy seen at this later stage is probably compensatory hypertrophy. This biopsy was kindly referred by C. Greco, Children’s Hospital, San Francisco.
FIG. 7 Dynamic, though nonprogressive, histopathology of the mdx mouse. Shown are agematched micrographs from normal mice gastrocnemii (A, C, E, G ) , and dystrophin-deficient mdx mice gastrocnemii (B, D, F. H). All panels are from hematoxylin-eosin stained frozen sections (bars, 500 pn). A, B, C, and D are from 3-week-old mice. A and C are from the same normal mouse muscle, with A showing a higher magnification. Similarly, B and D are from the same mdx mouse muscle, with B showing a higher magnification of typical early “prenecrotic” muscle, while D shows a region of the muscle experiencing its first round of synchronous grouped necrosis. Note the hypertrophy of the prenecrotic mdx muscle fibers (B)relative to the normal muscle (A). As with early Duchenne dystrophy, this hypertrophy might represent one of the first visible signs of dystrophin deficiency. Most, if not all of the muscle fibers will experience the initial round of grouped necrosis, as shown in D, within a few weeks after it is first seen at this 3-week time point. There is no detectable pathology before this time point. E, F, G, and H are from age-matched adult mice, with E and G representing different magnifications of the same normal mouse gastrocnemius, while F and G are from the same mdx gastrocnemius. F shows the typical rndx pathology at older age, with prominent internal nuclei and marked hypertrophy of muscle fibers relative to normal muscle (E). While most of older mdx muscle appears to have “successfully regenerated,” isolated foci of regenerating fibers (group of smaller, darker fibers in F) and necrotic fibers (H)can be readily found. Thus, while the grouped degenerationlregeneration of rndx muscle is most dramatic in the 3- to 5 week age range, the mdr myopathy is chronic and continues throughout the life of the mouse. Note the paucity of fibrotic tissue in the mdr mouse (F,H)relative to human Duchenne dystrophy muscle (Fig. 6).
Eric P. Hoffman and Jose Rafael M.Gorospe
124
FIG. 7 (con!.)
N. THE m d MOUSE: ~ NONPROGRESSIVE HISTOPATHOLOGY RESULTING IN HYPERTROPHY mdx mouse muscle shows no hypercontracted eosinophilic fibers or any other signs of pathology until 3 to 4 weeks postpartum (Bridges, 1986; Coulton et al., 1988a) (Fig. 7B). At this relatively late time, there is extensive grouped necrosis (Fig. 7D),where areas of as many as 100 fibers can be seen undergoing synchronous degeneration or regeneration. Most of the skeletal muscles of the mouse seem to experience at least one “round” of necrosis within 2-3 weeks. Subsequent rounds of degeneration and regeneration occur over a much longer
8. Duchenne Muscular Dystrophy
125
time frame, giving the muscle an appearance of full recovery (Anderson et al., 1988). However, there is ample evidence for a chronic myopathy even in old rndx mice (Fig. 7F,H); the muscle never fully recovers. The most striking difference between the histopathology of muscle in the mouse and humans is the paucity of progressive fibrosis in the mouse. Even in aged mdx mice, the muscle fibers are generally tightly packed, with little evidence of the endomysial connective tissue proliferation so characteristic of the late stages of human Duchenne dystrophy (compare Fig. 6 and 7). While fibrosis is not an overt characteristic of mdx muscle, there is an accumulation of collagen compared to normal muscle (Marshall et al., 1989), but this is much less than in human Duchenne muscle. Another difference is the preponderance of central nucleation in rndx mouse muscle fibers; the failure of rndx myonuclei to migrate to the periphery of muscle fibers may be a consequence of the late onset of necrosis in the mouse (Morgan et al., 1990). However, central nucleation appears to be a general feature of muscle regeneration in rodents.
V. THE xmd DOG: RAPIDLY PROGRESSNE HISTOPATHOLOGY LEADING TO MUSCLE WASTING
At birth, dystrophin-deficient dogs show massive elevation of CK levels (Valentine et al., 1988), associated with histopathological changes similar to those of neonatal humans (B. Valentine and B. Cooper, personal communication). Specifically, eosinophilic hyaline fibers are seen frequently, with no overt degeneration or regeneration, as in humans. Within a few weeks, the pathology changes to the kind of florid myopathy seen in a 2- to 5-year-old human Duchenne patient (Fig. 8). As with humans, proliferation of connective tissue is an early and chronic feature (Valentine et al., 1986, 1990). Similarly, with the gradual increase in fibrosis there is more and more loss of muscle fibers. Thus, the canine histopathology is the same as that of humans, only sped up; the changes take years to develop in humans, only weeks in the dogs. While the above description of the histopathology holds true for the most often studied muscles, recent studies in the dog suggest that this may be a generalization. Systematic histopathological studies of neonatal dog muscles have shown that those muscles which experience the greatest activity in puppies do in fact show widespread necrosis at birth, namely, the tongue, diaphragm, trapezius, deltoideus, extensor carpi radialis, and sartorius muscles (B. Valentine and B. Cooper, personal communication). The pathological lesions in the diaphragm appear to be responsible for the high frequency of neonatal death in dystrophindeficient dogs (B. Valentine and B. Cooper, personal communication).
126
Eric P. Hoffman and Jose Rafael M. Gorospe
FIG. 8 Histopathology of the dystrophin-deficient xmd dog. Shown is a modified trichromestained frozen section from a muscle biopsy of a 6-month-old xmd dog. This photograph represents the pathology representative of the mid-stage disease, analogous to the 5-year-old human Duchenne muscle shown in Fig. 6C. Like the human muscle, the dog muscle is characterized b y progressive histopathology, with the eventual replacement of the muscle by fibrotic tissue and the gradual failure of the muscle to regenerate. The early (neonatal) histopathology of the dog is also strikingly similar to that of humans. This photograph was kindly supplied by Dr. Bany Cooper of Cornell University.
VI. THE DYSTROPHI N - DEFICIENT CAT: SEMIPROG RESSNE HISTOPATHOLOGY WITH NO LOSS OF MUSCLE FIBERS The first two affected males were studied relatively late in development (6 months and later), were neutered, and did not breed (Carpenter er al., 1989). There has been no systematic study of the pathology at different ages, and we await breeding of the new litter in Florida (F. Gaschen and E. Hoffman, unpublished). The histopathology of the cat is similar to that of the mdr mouse, where the most striking feature is variation of fiber size, indicative of a chronic but mild myopathy (Fig. 9). Like the mouse, there is only mild fibrosis and no apparent loss of muscle fibers (Fig. 9). A striking feature of the cat muscle is focal calcium deposits.
8. Duchenne Muscular Dystrophy
127
FIG. 9 Histopathology of the dystrophin-deficient cat. Shown is the typical muscle histopathology of a 6-month-old dystrophin-deficient cat. Note the variation in fiber size, with many hypertrophied myofibers. While the cat muscle shows a relative paucity of endomysial connective tissue proliferation relative to older Duchenne dystrophy human muscle (Fig. 6). occasional small foci of fibrotic replacement of the muscle can be seen. Interestingly, the myofibers within these foci often appear atrophic and halted in regeneration, much like those in late stage Duchenne dystrophy (Fig. 6D), an observation in keeping with a prominent role for fibrosis in the progressive histopathology of the human and dog. A striking and consistent feature of the cat histopathology is the foci of mineralization (very dark regions). The limited number of dystrophin-deficient cats identified to date has precluded extensive studies of the histopathology over development, as has been done for other dystrophin-deficient species. The necropsy specimen used for this micrograph was kindly supplied by Dr. Frederic Gaschen, University of Florida, Gainesville. Bar, 500 pm.
VII. PHASE 1. THE PRIMARY CELLULAR CONSEQUENCE OF DYSTROPHIN DEFICIENCY: GENERALIZED LEAKAGE OF THE PLASMA MEMBRANE BUFFERED BY THE SYNCYTIAL CYTOPLASM? Dystrophin is found normally at equal levels in fetal (second half of gestation), neonatal, and adult muscle, and it is absent at all ages in all dystrophin-deficient organisms (Hoffman el al., 1988b; Patel et al., 1988; Bieber et al., 1989). The earliest and most consistent evidence of dystrophin deficiency in all organisms is the high serum levels of CK and other muscle enzymes, attributed to release of
128
Eric P. Hoffman and Jose Rafael M. Gorospe
soluble cytoplasmic enzymes from skeletal muscle fibers (Rowland, 1980). After release from the muscle fibers, these muscle-specific enzymes appear in large quantities in peripheral blood, where they can be easily and accurately detected and quantitated. The leakage of these enzymes from muscle decreases the levels of these enzymes in muscle tissue (Dreyfus et al., 1954; Rowland, 1964). CK is the enzyme most commonly assayed because the assay is simple and sensitive, and the high serum levels are particularly striking. However, CK measurement is infamous for variability between laboratories, between patients, and at different times in the same patient. It is therefore difficult to compare CK values quantitatively. However, to give the reader an idea of the magnitude of the elevations of serum CK in dystrophin-deficient species, CK values in Duchenne dystrophy patients are at least 50 times the upper limit of normal; cat CK values may reach 140,800 U/liter (normal values are < 200 U/liter) (Carpenter et al., 1989), the dog as high as 162,100 U/liter (normal -500 U/liter) (Valentine et al., 1989a), and the mdr mouse as high as 4,400 U/liter (normal 100 Ulliter) (Quinlan et al., 1990). Many physiological or biochemical insults on normal muscle may lead to high serum CK levels, including vigorous exercise (Fowler et al., 1968) and ischemic or other metabolic damage (Hearse, 1979). The general mechanism responsible for the changes in serum CK is easily visualized: When muscle cells are sufficiently damaged they undergo segmental necrosis, releasing cellular contents into the extracellular space. This results in the appearance of muscle CK in circulating blood. As might be expected, the magnitude of serum CK elevation is often proportional to the amount of muscle fiber necrosis taking place (Rowland, 1980), and serum CK values are used as an indicator of heart muscle damage after myocardial infarction. Very high values of serum CK are often accompanied by widespread myofiber necrosis and consequent clinical weakness. All the dystrophin-deficient organisms manifest extremely high levels of serum CK from birth onward. If the gross elevations of CK in the dystrophindeficient species (humans, dogs, cats, and mice) were solely the result of widespread necrosis, there should be evidence of this both histologically and clinically. This is not the case. While neonates with Duchenne dystrophy have tremendously elevated serum CK levels, the histopathology of the muscle may be subtle, showing only occasional eosinophilic “hypercontracted” fibers, with little or no overt cell necrosis (Hudgson et al., 1967; Guibaud et al., 1981; Barmada, 1990) (Fig. 7). Similarly, neonatal dystrophin-deficient dogs show tremendous elevations of CK (Valentine et al., 19881, accompanied by the generally subtle pathology similar to that seen in human neonates (B. Cooper, personal communication). Clinical weakness is also usually associated with major elevation of serum CK. However, consistent with the minimal pathology,
-
129
8. Duchenne Muscular Dystrophy
neonatal humans and dogs usually appear clinically normal. Thus, the pathological and clinical pictures seem inconsistent with the extremely elevated CK values. It is therefore necessary to derive a mechanism for the massive elevation of CK in dystrophin-deficient muscle which is not dependent on myofiber necrosis. This mechanism is most likely chronic transient myofiber leakage; dystrophindeficient muscle fibers show a generalized membrane instability due to the lack of part of their membrane cytoskeleton. This instability is exacerbated by the torsional stresses on the fiber membrane. Indeed, physical breaks of the plasma membrane were observed in Duchenne dystrophy muscle over 15 years ago (delta lesions; Mokri and Engel, 1975). While such membrane instability would be lethal to most types of cells, myofibers are unique in two respects. First, muscle fibers are syncytial, so any transient localized membrane permeability problem will be buffered by a functionally infinite cytoplasm. Second, as described in more detail below, muscle is uniquely equipped to deal with toxic Ca2+ influx. Some recent important papers reinforce the hypothesis of a generalized membrane instability and consequent leakage problem. Morandi et al. (1990) studied female carriers of Duchenne muscular dystrophy, and found that dystrophinnegative fibers showed both increased intracellular Ca2 and albumin despite a normal morphological appearance. Adjacent dystrophin-positive myofibers did not show increased Ca2+ and albumin. Menke and Jockusch (1991) studied dystrophin-deficient mclx muscle fibers in culture, and showed that they were highly sensitive to osmolality changes in the medium, supporting a weak membrane problem. To the contrary, some authors have argued against a generalized instability problem, and instead hypothesize that the increased intracellular Ca2 in human and mouse dystrophin-deficient muscle fibers is due to entrance of Ca2+ through specific “Ca2+ leak channels” (Fong et al., 1990). These same authors have more recently written that the normal concentrations of sodium in dystrophin-deficient fibers argues against a generalized membrane leakage (Turner et al., 1991). However, the large efflux of CK and other large polypeptides out of morphologically normal dystrophin-deficient myofibers, and the influx of albumin, another large peptide, both argue against a specific channel-mediated Ca2+ influx. While the primary cellular manifestation of dystrophin deficiency in skeletal muscle may be transient myofiber leakage, this leakage does not necessarily lead to myofiber necrosis. Neonatal humans, dogs, and mice all show little evidence of myofiber death, despite the striking elevations of CK. As described above, the extent of myofiber death and its exact timing in specific muscle groups seem variable. What then is the relationship between leakage and necrosis in dystrophin-deficient muscle fibers? +
+
Eric P. Hoffman and Jose Rafael M. Gorospe
130
VIII. PHASE I CONCLUSION: AN INTEGRATED MODEL FEATURING Ca2+ Cell death has been intensively studied in many different systems. While the cascade of events leading to the necrosis of a cell is often complex, the terminal event in most, if not all, cell death systems seems to be the influx of Ca2+ into the cell past a threshold; the elaborate intracellular Ca2 buffering and regulatory mechanisms are pushed beyond capacity (Carafoli, 1987; Schanne et al., 1979; Farber, 1982). All cells make extensive use of Ca2+ as an intracellular messenger, with elaborate systems of regulation, but muscle fibers are unique in the large amount of Ca2+ shuttled with each depolarization, contraction, and relaxation (Cannell and Allen, 1984). To direct this traffic, there are highly specialized organelles and proteins which the muscle utilizes (Carafoli, 1987) (Fig. 10). Despite the apparent high buffering capacity of the Ca2+ regulatory proteins in muscle, there is a threshold of Ca2+ concentration, beyond which the regulatory systems are overwhelmed and the fiber dies. Influx of extracellular Ca2+ seems to play a prominent role in the myofiber necrosis of experimental muscle damage (Duncan, 1978; Jones et al., 1984; Jackson et al., 1984) and in genetic myopathies (Wrogemann and Pena, 1976; Ebashi and Sugita, 1979). Is there evidence for a role for Ca2+ in the pathophysiology of dystrophindeficient muscle fibers? If there is leakage of CK out of myofibers, smaller molecules, such as Ca2+ ions, probably enter more readily than normal. In fact, the eosinophilic hyaline fibers, the earliest histopathological manifestations of dystrophin deficiency, contain increased amounts of Ca2 (Bodensteiner and Engel, 1978; Maunder-Sewry et al., 1980; Bertorini et al., 1984). Moreover, both human Duchenne and mouse mdx muscle contain higher levels of Ca2+ than normal muscle (Turner et al., 1988; Mongini et al., 1988; Bertorini el a l ., 1982). Pathological levels of Ca2 are also evident in the calcium deposits of cat skeletal muscle (Carpenter et al., 1989) (see Fig. 9) and the calcification of heart muscle in dog (Valentine et al., 1989b) and cat (Carpenter et a!., 1989). Thus, it is probably safe to assume that dystrophin deficiency results in Ca2 influx. When the Ca2 influx overwhelms the Ca2 -regulatory capacity of the cell (i.e., passes the lethal threshold) then segmental necrosis ensues. It is important to speculate as to why the timing and rate of necrosis seem so variable. For example, why does cardiac muscle not experience the same extensive necrosis seen in skeletal muscle, despite the fact that cardiac muscle is as “dystrophin-deficient” as skeletal muscle? The process leading to dystrophin deficiency-induced necrosis in any specific cell is probably best visualized as a delicate balance between physiological and biochemical events that control Ca2 influx and the intracellular regulatory mechanisms that remove Ca2+ from the cytoplasm to maintain homeostatic levels. +
+
+
+
+
+
+
8. Duchenne Muscular Dystrophy
Normal muscle
131 Dvstrophin-deficient muscle
Working for Fiber:
Working against Fiber:
(ironsporls Ca Into SR) (exlcnaive In hear(, not in skelelal muscle) (binds Ca, not in smoolli muscle) Cnlmadulln (high in smoolh muscle, low in rkelelnl) PlnSma membranc (binds Ca Via anlonle phorphollpids) Pnrvnlbumln (less In mammulim muscle) Oihers??
Dyslrophln dclieirncy (causes lrnkngc d libcr) -Augmented by sircss an fiber? (cxercltc, inerrnscd weight olorgnnism) 4 u g m r n l e d by lnerroscd size or fibers? (lower surlaer area or membmnr) -Augmcnlcd by malurollon o r rilmrs? I.ncnllzcd irehrmin (causes influx or cm) -Augmented by dyrirophin deneieney In smoolh muscle? -Augmenlrd by dense rlilrolic lissuc? ~ c i i v i i yor musoio ribtr -Ca rclcarc by snrcqililsmlc reliculum (rynnodinr receptor) -Cn influx rrom menrbronc dcpolariultion (DRP reccpior)
CaMg ATPase NoCn cxlinnger T r ~ p o n l iC i
Milocliondrln
(very imporlonl in palhologlcol ~ I i u ~ t l c m s , malrix iondlng provides rxlrnslve CPburner)
FIG. 10 The delicate balance of Ca2+-induced myofiber death initiated by dystrophin deficiency. Listed are some of the better studied components of the Ca2+ regulatory machinery of muscle cells (left). To the right are listed physiological events, both normal and pathological, which might serve to increase Ca2 influx in myofibers, the most prominent being dystrophin deficiency-induced membrane leakage. This model proposes that the penultimate event leading to cell necrosis in dystrophin-deficient fibers is a tipping of the balance toward intracellular Ca2+ concentrations past a threshold, where the cell can no longer compensate and reestablish homeostasis. The Ca2+ regulatory mechanisms listed on the left were gleaned from the excellent review by Carafoli (1987). +
The mechanisms are poorly understood, but simple models can be derived from what is known (Fig. 10).To begin to construct such models, one must first identify the cellular mechanisms which are responsible for Ca2 homeostasis in different muscle types, and also identify the factors which work against homeostasis by raising intracellular Ca2+ levels (Fig. 10). Different mechanisms in +
132
Eric P. Hoffman and Jose Rafael M. Gorospe
different cell types could shift the threshold toward leakage-induced necrosis, or away from it. To take the example of the relative insensitivity of heart muscle to dystrophin deficiency-induced cell necrosis, there are a number of characteristics of cardiocytes that would be expected to shift the balance in favor of Ca2 homeostasis relative to skeletal muscle (Fig. 10). Specifically, heart muscle has an active Na Ca2 exchanger system that is not found in skeletal muscle. The Na Ca2 exchanger serves as an efficient bulk Ca2+ extrusion system for the heart (Carafoli, 1987). Heart muscle has a much greater surface area of plasma membrane per milligram of tissue, which might serve to lower stress on the membrane, thereby decreasing the amount of leakage due to dystrophin deficiency-induced membrane instability. The increased membrane surface area might also remove Ca2 from the cytoplasm via binding to anionic lipids. Cardiac muscle contains more mitochondria than skeletal muscle, and mitochondria are known to serve as life-saving Ca2 buffers when intracellular Ca2 levels reach pathological concentrations (matrix loading). Finally, eccentric contractions of skeletal muscle fibers are known to increase leakage of CK out of fibers (and presumably Ca2 into the fiber) (Newham et al., 1983; McCully and Faulkner, 1985). Cardiac muscle uses concentric contractions, which are thought to be much less damaging to the fibers. These different mechanisms by which cardiac muscle and limb muscle regulate CaZ provide an explanation for their different sensitivity to dystrophin deficiency-induced cell death; cardiac muscle has a greater margin of safety before reaching its Ca2 -induced necrosis threshold. Thus, the delicate balance of Ca2+ influx versus homeostasis in heart muscle is tilted in favor of homeostasis (Fig. lo), despite dystrophin deficiency. Similar differences may be seen in different types of skeletal muscle. For example, fast-twitch glycolytic skeletal muscle contains fewer mitochondria than slow-twitch oxidative muscle. Since mitochondria serve as vital Ca2 buffers in pathological situations, one might expect that fast-twitch muscle experiences more necrosis than slow-twitch muscle due to a tilting of the balance toward necrosis. However, fast-twitch skeletal muscle contains a more efficient Ca2 Mg2 -ATPase system in the sacroplasmic reticulum than does slow muscle, which might efficiently mop up Ca2+ influx due to dystrophin deficiencyinduced leakage. This characteristic of fast-twitch muscle should tilt the balance toward homeostasis; fibers might less frequently reach pathological levels of Ca2 where they would require the buffering effects of mitochondria. In support of this “Ca2 balance” hypothesis, the skeletal muscle group least affected in Duchenne dystrophy is the extraocular muscles. Normal extraocular fibers are capable of reaching Ca2 homeostasis consequent to a contraction I0 times faster than fibers in other muscle groups; such a fast and efficient Ca2 regulatory system should tilt the balance toward homeostasis and away from Ca2+ influx-induced necrosis in dystrophin-deficient extra-ocular muscles. +
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
133
8. Duchenne Muscular Dystrophy
One of the mechanisms listed in the “working against fiber” category probably plays a prominent role in much of the necrosis of dystrophin-deficient muscle of all types, namely, ischemia. Ischemia increases Ca2+ influx into cells, a major factor in ischemia-induced cell death in all types of cells (Farber, 1982). To understand the probable role of ischemia in dystrophin deficiency, it is necessary to consider the pathology of dystrophin-deficient muscle. The definition of “myopathy” is a disorder intrinsic to the muscle fiber. In most necrotizing myopathies, fibers show segmental necrosis singly, in isolation from neighboring fibers. However, in the mdx mouse, muscle fibers do not show necrosis in isolation, but die in large groups, often of hundreds of fibers (Fig. 7). This kind of “grouped necrosis” is more characteristic of vascular disease (ischemia) than it is of a typical myopathy, implying that the fibers within a vascular bed are subject to cell death simultaneously. Grouped necrosis is also observed in human Duchenne dystrophy (Engel, 1977; Mendel et al., 1971, 1972), and is even more striking in Becker dystrophy (Grim 1986). Indeed, Engel and associates (Engel, 1977; Mendel et al., 1971, 1972) proposed that the vasculature has a primary role in the etiology of Duchenne dystrophy, because of this appearance and because experimentally induced local ischemia in rabbit muscle mimicked the human histopathology (and, in retrospect, even more faithfully mimics mdx histopathology!) (Engel, 1977; Mendel et al., 1971, 1972). In addition, all dystrophindeficient species share a characteristic and unique cardiomyopathy, which is localized to the basolateral free wall of the left ventricle (Bridges, 1986; Carpenter et d . , 1989; Valentine et al., 1989b; Perloff et d . , 1966, 1967, 1984; Frankel and Rosser, 1976). This region of the heart is most susceptible to ischemic insult (see Hoffman et a f . , 1988b), though ischemic lesions of the heart are primarily subendocardial, while those in the dystrophin-deficient species are subepicardial. When the ischemia-induced Ca2 influx becomes excessive in normal fibers, then necrosis ensues. However, dystrophin-deficient fibers are stressed in regulating Ca2+ because of the plasma membrane leakage. Thus, any additional insult due to localized ischemia could push a group of myofibers within a particular vascular bed beyond the Ca2 threshold and into necrosis. Localized ischemia, sublethal in normal myofibers, may be lethal in dystrophin-deficient fibers because it makes an already bad situation worse. Moreover, vascular smooth muscle normally contains dystrophin, and this is also lacking in all the dystrophin-deficient species (Hoffman et al., 1988b; E. P. Hoffman, unpublished). Vascular smooth muscle is probably also functionally defective (Hoffman et a f . , 1988b; Miyatake et al., 1989). If the regulation of blood flow to muscle is compromised by malfunction of dystrophin-deficient vascular smooth muscle, a vicious cycle could lead to more and more overt necrosis of groups of muscle fibers. In support of this hypothesis, human Duchenne muscle shows progressive 31P-NMR changes consistent with progressive metabolic +
+
134
Eric P. Hoffman and Jose Rafael M. Gorospe
deterioration (Younkin et al., 1987) that could result from vascular insufficiency (ischemia). In conclusion, dystrophin deficiency is the primary biochemical defect which directly or indirectly results in plasma membrane (and probably t-tubule) instability. Cell death (necrosis) results when localized leakage exceeds a threshold, with local Ca2 concentrations the threshold determinant. Many biochemical and physiological factors influence the degree of myofiber leakage or the threshold (Fig. 10); these variables determine when and how extensively a muscle group becomes necrotic. +
IX. PHASE li: PROGRESSIVE PATHOLOGY AND CLINICAL WEAKNESS SPECIFIC TO HUMANS AND DOGS All dystrophin-deficient organisms share the same genetic and biochemical defect, and all share similar early cellular lesions. However, with advancing age, humans and dogs show histopathology that is progressive, while cats and mice do not show these progressive features. The first clinical manifestation of human dystrophin deficiency seems to be muscle hypertrophy, most noticeably of the calf muscles (Fig. 11). By the age of 5 years there is difficulty climbing stairs, rising from the floor, and generally keeping up with peers. Muscle wasting and weakness progresses and the patient is usually wheelchair-bound by age 12 years. Respiratory or cardiac failure is usually evident by age 20 or soon thereafter. The dog shares a similar progressive clinical phenotype, but over a shorter time frame. The dogs show little or no overt weakness until 2 months of age, whereupon a stiff-limbed waddling gait is noticed. Muscle wasting and weakness then rapidly ensue, leaving dogs quite disabled by 6 months of age (Fig. 12) (Komegay et al., 1988; Valentine et al., 1988). Most of the dogs die of cardiac failure, or are put to sleep when their disability gets too severe. While most affected dogs follow the described clinical progression, a minority of dogs die at birth for reasons which are currently not clear. In contrast, both the cat (Fig. 13) and mouse (Fig. 14) show no overt clinical weakness at any point. Rather, the primary clinical expression of dystrophin deficiency in the mouse and cat seems to be continued hypertrophy of muscle (Fig. 13-16). The hypertrophy becomes quite impressive (Fig. 13-16), with many of the muscle groups of the cats weighing twice as much as those of normal cats (Carpenter et al., 1989). The hypertrophy of the cat muscle is accompanied by stiffness. The stiffness does not seem to be due to progressive weakness, but because of uncoordinated contractions of myofibers, as shown by the psuedomyotonic electromyograph (Carpenter et al., 1989). The hypertrophy may itself
8. Duchenne Muscular Dystrophy
135
FIG. 11 The clinical progression of Duchenne muscular dystrophy. Shown are patients at the early, preclinical stage of the disease (A), and the later stage of the disease (B). The only overt clinical characteristic seen in the 4-year-old patient to the left is the hypertrophy of muscles, particularly of the calves. Weakness and muscle atrophy begins around this age, and then progresses, leaving patients wheelchair-bound around the age of 11 years. The patient to the right is 13 years old, and much of the patient's skeletal muscle has wasted. These photographs are taken from Emery (1988). with the kind permission of Dr. Emery, Dr. Victor Dubowitz, and the publisher.
be life-threatening; one of the Florida cats had extremely hypertrophic diaphragm muscles that occluded the esophagus, making it necessary to put the cat to sleep to avoid starvation (F. Gaschen, personal communication) (Fig. 15). Hypertrophy can be equally impressive in the mouse (Fig. 16). In fact, old mdx mice have been shown to be physically stronger than normal mice (Coulton et al., 1988b). Why does muscle in mice and cats enlarge, while dog and human muscle
136
Eric P. Hoffman and Jose Rafael M. Gorospe
FIG. 12 Dystrophin-deficient dog (xmd). Shown is a 6-month-old dystrophin-deficient golden retriever. These dogs have an X-linked dystrophin deficiency, and have a progressive histopathological and clinical disease very similar to that of human Duchenne dystrophy. The clinical course is over a shorter time period compared to humans: dogs appear normal until about 2 months of age, when a stiff-limbed swaying gait is observed. Weakness and muscle atrophy is then rapidly progressive. At 6 months of age (dog shown in photograph), there is significant weakness of the proximal musculature, as seen in the hunching of the hips and the laying flat of the distal portions of the rear legs. The electromyography of the dog shows "pseudo-myotonia," indicative of hypersensitivity of the plasma membranes of myofibers (Valentine el al.. 1989a), an observation in keeping with the leaky membrane hypothesis. Photograph through the generosity of Dr. Barry Cooper, Cornell School of Veterinary Medicine.
wastes? First, humans do indeed hypertrophy at a young age, most noticeably in the calf muscles (Fig. 11). In certain rare cases, the hypertrophy in Duchenne patients can be quite impressive, as shown in Fig. 17. The hypertrophy in Duchenne dystrophy has traditionally been considered pseudohypertrophy, due to the extensive fibrotic replacement of the muscle which occurs at later stages of the disease. However, as shown in the histopathology data presented above, such fibrotic replacement can be minimal even at 2 years of age, yet hypertrophy is clearly present at this time point. Thus, the hypertrophy in young Duchenne patients is almost certainly true hypertrophy, rather than pseudohypertrophy. Apparently, all dystrophin-deficient species show hypertrophy as an early, and probably primary manifestation of the disease. Cats and mice continue to hypertrophy, while dogs and humans begin to atrophy instead.
8. Duchenne Muscular Dystrophy
137
FIG. 13 Dystrophin-deficient cat. Shown is one of the original dystrophin-deficient cats, found on Nantucket Island off the coast of Massachusetts. The cats are born with no overt clinical phenotype. They show progressive hypertrophy of most musculature, and, by 6 months of age, most muscle groups weigh double those of bone size-matched normal cats (note hypertrophy of the neck in the cat shown). While the cats show little overt weakness, the hypertrophy can become somewhat debilitating, causing stiffness. The electromyogrdphy of the cat shows pseudomyotonia, similar to that of the dog and mouse, suggestive of hypersensitivity of the muscle fibers to membrane damage. Photograph through the kind generosity of the cat’s owner, Ms. Margo Bond, and Dr. James Carpenter, Angel1 Memorial Hospital of the MSPCA, Boston.
138
Eric P. Hoffman and Jose Rafael M. Gorospe
FIG.14 The dystrophin-deficient mdr mouse. Shown is an mdr mouse of approximately I year of age. mdr mice show no overt clinical phenotype at any point in their normal life span. They do show progressive hypertrophy, though this is not debilitating. The electromyography of the mouse is very similar to that of the cat and dog, showing hyperirritability of fibers, evidenced by pseudomyotonia (Chapman ef al., 1989). though such high spontaneous activity has also been found in normal control mice (B. Valentine, personal communication).
What causes the hypertrophy, and what then causes the switch to wasting? Hypertrophy in dystrophin-deficient muscle is reflected in the histopathology by the hypertrophy of individual muscle fibers (nearly 200% larger than normal in the mouse) (Anderson et al., 1988) (see Fig. 6 and 7). It is difficult to determine if the actual number of muscle fibers also increases because there is also extensive fiber-splitting. The enlargement of individual muscle fibers in Duchenne dystrophy has been attributed to compensatory hypertrophy, rather than a primary manifestation of the disease. This concept must be questioned, given the hypertrophy seen in preclinical Duchenne patients and in the mouse and cat models, all of which show no loss of muscle fibers and no apparent need for compensatory hypertrophy. Thus, hypertrophy seems to be a primary manifestation of dystrophin deficiency. Muscle atrophy, on the other hand seems to be part of the progressive phase I1 (Fig. 18), which is presumably a secondary consequence of dystrophin-deficiency. The clinical progression of both the human and dog disease is directly reflected in the progressive histopathology. There are three features of the dog and human
8. Duchenne Muscular Dystrophy
139
FIG. 15 Hypertrophy in the dystrophin-deficient cat: the diaphragm at 6 months. Shown is the diaphragm muscle of one of the original dystrophin-deficient cats. Extreme hypertrophy is evident, The hypertrophy of the diaphragm of the cats can be life-threatening: the diaphragm of one of the recent Florida cats occluded the esophagus, bringing the cat to the brink of starvation (F. Gaschen, personal communication). Photograph through the kind generosity of Dr. James Carpenter, Angel1 Memorial Hospital of the MSPCA, Boston.
140
Eric P. Hoffman and Jose Rafael M . Gorospe
FIG. 16 Hypertrophy in the dystrophin-deficient rndx mouse. Shown are skinned forelegs from four 1.5-year-old male mice of different genotypes. To the left is shown the mdr mutation on its original background (C57B10 mdx, lower left, note incorrect labeling), matched with a normal control mouse of the same inbred strain (C57B10, upper left). The hypertrophy of the mdr muscle is obvious. To ensure that the hypertrophy of mdx muscle was not simply the result of the genetic background of the particular inbred strain, an mdr C57BIO homozygous female was crossed with a BalbC male mouse (one of the inbred strains of smallest size). Shown is the skinned foreleg of one of the hemizygous mdr male progeny of this cross (C57BIOIBalbC mdx, lower right), matched with an age-matched normal male of a different inbred strain (C57B6, upper right). As with the other mice, the rndr mouse leg is clearly hypertrophied relative to the normal mouse. The hypertrophy of mdx muscle occurs in the absence of severe fibrotic replacement of the muscle, indicating that it represents true hypertrophy, rather than pseudohypertrophy.
histopathology which seem completely correlated with the clinical expression (Fig. 18), namely, progressive fibrosis, progressive failure of regeneration, and progressive fiber loss. These three features lead to gross muscle wasting, and the wasting is then correlated with the muscle weakness. Cats and mice, on the other hand, never develop any of these histopathological features, and therefore never atrophy and become weak. Thus, it seems that the disease seen in Duchenne dystrophy can be considered
8. Duchenne Muscular Dystrophy
141
FIG. 17 Extreme example of hypertrophy in human Duchenne dystrophy. Muscle hypertrophy (traditionally called pseudohypertrophy) is one of the first overt clinical manifestations of Duchenne muscular dystrophy. Shown is a particularly dramatic example of hypertrophy in Duchenne dystrophy. Note the nonuniform distribution of the hypertrophy, with some muscle groups showing extreme hypertrophy (gastrocnemius, deltoid), while other groups show atrophy (quadriceps, bicepsltriceps). It is proposed that this nonuniform distribution provides graphic evidence of different muscle groups in different stages of the disease process, with the hypertrophied muscles representing phase I of the disease, and the atrophied muscles representing phase 11. Photograph taken from Emery (1988) and from Bundey (1972) with the kind permission of Dr. A.E.H. Emery, Dr. Sarah Bundey, and the publishers.
Eric P. Hoffman and Jose Rafael M. Gorospe HUMAN
DOG
MOUSE
X-linked Disorder Genetic Biochemical Homology
Mutations Localized in Dystrophin Gene Dystrophin Deficiency in Muscle
PHASE I:
Muscle Fiber Necrosis
The primary consequences
Fiber Size Variation
or dystmphin deficiency
Marked Fiber Hypertrophy Muscle Group Hypertrophy At Young Age At Old Age
PHASE 11:
Progressive Fibrosis
The secondary consequences
Progressive Failure of Reeeneration
or dystmphin deficiency
Progressive Loss of Muscle Fibers Progressive
weakness Deficient Cardiomyo-
Yes
Yes
I
I
Yes
Yes
+Yes
Yes
I Yes
I
yes
No
I
No
Specific Involvement of Basolateral Free Wall?
FIG. 18 Rationalization of the pathophysiology of dystrophin deficiency as a two-phase process. Shown are the apparently invariant features of all dystrophin-deficient species, which are proposed to represent the primary consequences of dystrophin deficiency in skeletal muscle (phase I, top of figure). Intrinsic to the pathophysiology of phase I are all the factors which might influence the Ca2+ threshold of individual fibers, as outlined in Fig. 10. Features of the disease seen only in humans and dogs are hypothesized to represent the secondary, progressive consequences of dystrophin deficiency (phase 11). It is the progressive features of phase I1 which seem to lead to weakness, and ultimately death, in humans with Duchenne dystrophy and the xmd dogs.
8. Duchenne Muscular Dystrophy
143
a two-phase disease (Fig. 18). Phase I is the primary consequence of dystrophin deficiency; phase I1 expresses the secondary consequences. Muscle wasting leading to weakness and a premature death in humans and dogs is therefore not the direct result of dystrophin deficiency, it is the consequence of secondary, progressive histopathological changes.
X. THE DEVELOPMENT OF PROGRESSIVE HISTOPATHOLOGY: COULD BASIC FIBROBLAST GROWTH FACTOR HAVE A MAJOR ROLE?
What causes the progressive histopathology in humans and dogs? One possible culprit is the proliferation of fibrotic tissue, which has long been implicated in the pathogenesis of Duchenne muscular dystrophy (Bourne and Golarz, 1959; Duance et al., 1980). In fact, Duchenne (1868) himself originally described the disease as “paralysie myosclerosique.” However, fibrotic replacement of any tissue is a generic reaction to any injury, making if difficult to believe that fibrosis has a primary role in the etiology of Duchenne dystrophy. Nevertheless, the fibrosis in dog and human dystrophin-deficient muscle seems more severe than in other neuromuscular disorders. Proliferation of connective tissue is part of the complex process of wound repair, and occurs whenever there is substantial cell death. The fibrosis seen in Duchenne dystrophy is reminiscent of that seen consequent to chronic tissue damage (Duance et al., 1980), and is probably no more than internal scarring of muscle. However, chronic tissue damage is usually equated with chronic cell necrosis, and the proliferation of connective tissue in human Duchenne muscle appears before there is much cell necrosis. There seems to be an aggressive wound repair response in human and dog muscle in the absence of a visible wound. What could be responsible for the aggressive wound repair in the absence of large-scale cell necrosis? It seems certain that cell necrosis is aphysiological trigger for the activation of focal wound repair mechanisms. Growth factors and mediation from mast cells and immune system have all been implicated as possible biochemical triggers for wound repair (Claman, 1990; Rothe and Kerdel, 1991). One intensively studied potent and wide-acting growth factor is basic Jibroblast growth factor (bFGF) (Rifkin and Moscatelli, 1989; D’Amore, 1991). bFGF is found in most cells; receptors for bFGF also appear on the surface of most cells. While bFGF has been intensively studied for many years, it is not clear what mechanisms exist for its release from cells. Indeed, the only known mechanism for its release is cell necrosis; the bFGF contained in a dying cell is simply released into the extracellular space and quickly bound by receptors on other cells, or by the surrounding connective tissue (Rifkin and Moscatelli, 1989; D’Amore, 1991). This release mechanism for bFGF seems to be quite “logical” and effective: when a cell dies, bFGF is released, and wound repair ensues.
144
Eric P. Hoffman and Jose Rafael M. Gorospe
Could bFGF be involved in the fibrosis of muscle in human and dog dystrophin deficiency? bFGF has been purified from human adult skeletal muscle (Vaca et al., 1989), and skeletal muscle precursor cells (myoblasts) have been shown to have bFGF receptors (Olwin and Hauschka, 1986). Thus, the role of bFGF in skeletal muscle wound repair is probably similar to other tissues. Release of bFGF from skeletal muscle probably also follows cell necrosis. However, dystrophin-dejcient muscle may also release bFGF through a novel mechanism, cell leakage. If bFGF is chronically released from leaking dystrophindeficient cells along with CK, then the surrounding connective tissue might be triggered to enter into wound repair in the absence o f a sign$cant wound. Thus, the chronic leakage of myofibers may result in the chronic leakage of bFGF, which in turn induces a chronic wound repair resulting in extensive fibrosis. Chronic “pseudo-wound repair” in dog and human muscle could rationalize the complex cellular and physiological findings of Duchenne dystrophy (Hoffman et al., 1988b; Engel, 1977) (Fig. 10). For example, extensive proliferation of connective tissue might compromise the ability of the muscle to revascularize and reinnervate regenerating muscle fibers, leading to “ischemic” histopathology (grouped necrosis) (Engel, 1977; Mendel et a l . , 1971) and perhaps “neurogenic” features (McComas et al., 1971) as well. The ability of the vasculature and nerve fibers to regenerate in dystrophin-deficient muscle may be even further compromised by abnormalities intrinsic to the vascular smooth muscle cells and neurons, both of which normally contain dystrophin and are lacking dystrophin in Duchenne dystrophy and in all the animal models (Hoffman et al., 1988b; H. P. Hoffman, unpublished). Additional work must be done before bFGF can be implicated in the pathogenesis of Duchenne dystrophy. Moreover, if bFGF is responsible for the fibrosis in humans and dogs, it is not clear why mice and cats do not respond similarly, particularly because mdr mice show elevated levels of bFGF bound to the extracellular matrix (DiMario et al., 1989). Finally, mast cells, the immune system, and their mediators may have important roles in the development of fibrosis (Claman, 1990; Rothe and Kerdel, 1991). All these need to be studied in muscle of dystrophin deficient animals. XI. CONCLUSION: AN INTEGRATED MODEL AND ITS CONSEQUENCES ON THE DEVELOPMENT OF THERAPEUTICS
Many human genetic diseases, such as cystic fibrosis and sickle cell anemia, are the result of one or a few mutations which have embedded themselves in a particular population for some unknown reason. If the responsible mutations can be detected early, it is possible to develop programs for carrier-detection and prenatal diagnosis. Widespread and accurate genetic screening programs could
8. Duchenne Muscular Dystrophy
145
eradicate the disease by genetic counseling. For instance, random screening for camers of cystic fibrosis is being considered. However, genetic screening for Duchenne dystrophy will never be complete because the spontaneous mutation rate is too high. Even if it were possible to screen for all female carriers (including those 35% having a single point mutation of the 2.5 million base pair gene!), over one third of current cases of Duchenne dystrophy would still occur simply because 30% of cases arise from de n o w mutations in eggs or sperm. Completely accurate genetic screening, unless done on all fetuses for all 2.5 million base pairs of the gene, could decrease the incidence of Duchenne dystrophy only to about half of what it is today. If genetic screening cannot effectively lower the incidence of Duchenne dystrophy, it becomes even more important to attempt to develop a rational therapy. The animal models of Duchenne dystrophy might seem to be excellent vehicles for the development of treatment. Or are they? The mdr mouse cannot be used for therapies aimed to slow the progression of human Duchenne dystrophy because the mdr mouse does not share the progressive characteristics. To visualize how the animal models might be best utilized, it is useful to again refer to the two phases of the human disease: rational therapies could be directed at either phase I or phase 11, or both. Phase I is the result of dystrophin deficiency, so the human disorder could be cured by replacing dystrophin where it is deficient, halting the disorder before it starts. Muscle (myoblast) cell-based, dystrophin gene-based, and dystrophin protein-based therapies are all directed toward dystrophin replacement in myofibers, and therefore directed against phase I of the disease. Successful replacement of dystrophin should halt all the primary consequences of dystrophin deficiency (Figs. 18 and 19), preventing the onset of phase I1 and curing the disorder. There are two important corollaries to this logic regarding therapies directed toward phase I. First, both the mouse and cat diseases should be excellent therapeutic models because they fully manifest dystrophin deficiency, and because dystrophin replacement can be accurately monitored in treated animals by immunoblotting or immunofluorescence. Indeed, mice and cats are probably better for studying phase I than are human Duchenne patients themselves, because the effect of the dystrophin replacement therapy will not be complicated by the progressive secondary effects of phase 11. The second important corollary is that dystrophin replacement in an 1l-yearold Duchenne muscular dystrophy patient may not be therapeutic, even if 100% efficiency were accomplished; the secondary effects of dystrophin deficiency (progressive fibrosis, fiber loss, and failure of regeneration) may have damaged the muscle so that no myofiber can be sustained, independent of dystrophin content (Fig. 19). It may be necessary to establish neonatal screening programs for Duchenne muscular dystrophy for any realistic replacement of dystrophin (Greenberg et al., 1991); by age 5 , it may be too late to slow progression of the disease. For similar reasons, the dog, unless used in the neonatal period, may be
146
Eric P. Hoffman and Jose Rafael M. Gorospe
FIG. 19 Histopathological cycles initiated by dystrophinopathies. Part of this figure was taken from Rowland (1980).
147
8. Duchenne Muscular Dystrophy
a poor model for dystrophin replacement because of the progressive secondary features shared with humans. Dystrophin replacement using a cell-based approach (myoblast implantation) has already been effective in the mdx mouse (Partridge et al., 1989; Morgan et al., 1990; Karpati et al., 1989). After dystrophin-positive muscle precursor cells (myoblasts) have been injected into dystrophin-deficient muscle, the protein is expressed. Human trials are already underway. While the limited mouse results are encouraging, several obstacles must be overcome before myoblast injections can be applied as a rational therapy for humans (Partridge et al., 1989; Partridge, 1990). For instance, there must be an adequate supply of myoblasts to inject. Also, there are delivery problems; injection sites may have to be spaced 500 pm apart in three dimensions. Nevertheless, the mouse model (and the cat or young dogs) should prove to be invaluable for addressing these problems. It may be possible to inject purified DNA for dystrophin directly into muscle, instead of transplanting myoblasts (Wolff et al., 1990). The rndx mouse is an inexpensive yet superb model to test these therapeutic strategies. While dystrophin replacement addresses phase I of the disease, and thereby should cure it, the longevity and overall fitness of the dystrophin-deficient cats and mice suggest that developing therapies directed toward phase 11 may also be devised to treat human Duchenne dystrophy. As illustrated by the cats and mice, muscle seems to survive and function quite well without dystrophin, as long as the progressive aspects of the disease are kept from occumng. If the fibrotic replacement of the muscle could be slowed or halted in humans, there might be little or no weakness. However, mdr mice and the cats cannot serve as useful models for therapies directed against phase 11, at least in the near future. It is conceivable, however, that we will fully understand the pathophysiology of the progression of Duchenne dystrophy, and that this would permit the manipulation of mdr mice in a way that would replicate phase I1 in the mice. We cannot wait for this unlikely achievement. In the meantime, the mdx dogs provide an excellent model for therapies directed against phase 11. The dogs are probably a better model than the humans because the progression of the canine disorder is so rapid that efficacy of a treatment can be assessed rapidly. If the homeostasishecrosis equilibrium outlined in Fig. 10 is indeed active, then one can predict that pharmacological agents directed toward features of the “working against the fiber” category could very likely have therapeutic value. For instance, pharmacological agents that improve vascular flow to the muscle may help prevent transient ischemia. Drugs that alter the Ca2 concentrations in favor of homeostasis should be beneficial. No matter what agents are used, the timing of treatment relative to the progression of the disease will be critical. For example, drugs that increase regional blood flow should prevent focal transient ischemia early in the disease, and thereby slow the rate of necrosis, but fibrotic +
148
Eric P. Hoffman and Jose Rafael M. Gorospe
replacement later in the disease could so disrupt vascularization that the drug might be ineffective. In conclusion, rapidly increasing knowledge of the primary genetic and biochemical basis of Duchenne muscular dystrophy has illuminated a complex disorder that is best viewed as a dynamic process, rather than a simple and static biochemical defect. The complex nature of the disease frustrates attempts to understand it. However, this same complexity provides many therapeutic possibilities, and tools to address the many questions are available. The most important of these tools are the animal models.
Acknowledgments The authors are grateful for the “wild red pen” of Lewis Rowland, and the sage wisdom behind it. The authors are indebted to the many clinicians who have patiently explained what human Duchenne muscular dystrophy is at the histopathological, clinical. and human levels. The authors would also like to express thanks to Terence Partridge, James Carpenter, Frederic Gaschen, Bany Cooper, and Beth Valentine, for collaborations on the animal models, for offering valuable comments on the manuscript, and for permitting the inclusion of previously unpublished figures in this review [specifically Fig. 8 and 12 (Dr. Cooper), Fig. 13 and 15 (Dr. Carpenter), and the necropsy specimen used for Fig. 9 (Dr. Gaschen)]. The potential role for bFGF evolved through many conversations with Patricia D’Amore, Children’s Hospital, Boston, and the author (E.P.H.) is indebted to her for these stimulating discussions. Finally, most of the work described in this review would not have been possible without the generous financial support of the Muscular Dystrophy Association to the authors and other researchers in the field.
References Anderson, J. E., Bressler, B. H., and Ovalle, W. K. (1988). Functional regeneration in the hindlimb skeletal muscle of the mdx mouse. .I. Muscle Res. Moril. 9, 499-515. Arahata, K., Ishiura, S., Ishiguro, T., ef a / . (1988). lmmunostaining of skeletal and cardiac muscle surface membrane with antibody against Duchenne muscular dystrophy peptide. Nature (London) 333, 861-863. Barmada, M. A. (1990). Histopathology of neonatal Duchenne muscular dystrophy. In preparation. Beggs, A. H., Hoffman, E. P., Snyder, J. R., Arahata, K., Specht, L., Shapiro, F., Angelini, C., Sugita, H.,and Kunkel, L. M. (1991). Exploring the molecular basis for variability among patients with Becker muscular dystrophy: Dystrophin gene and protein studies. Am. J . Hum. Genet. (in press). Bertorini, T. E., Bhattacharya, S. K.. Palmieri, G. M. A,, Chesney, C. M., Pifer, D., and Baker, B. (1982). Muscle calcium and magnesium content in Duchenne muscular dystrophy. Neurology 32, 1088-1092. Bertorini, T. E., Cornelio, F., Bhattacharya, S. K . , Palmieri, G . M. A , , Dones, I., Dworzak, F., and Brambati, B. (1984). Calcium and magnesium content in fetuses at risk and prenecrotic Duchenne muscular dystrophy. Neurology 34, 1436-40. Bieber, F. R., Hoffman, E. P., and Amos, J. (1989). Dystrophin analysis in Duchenne muscular dystrophy: Use in fetal diagnosis and in genetic counseling. Am. J . Hum. Genet. 45, 362-367. Bodensteiner, J. B., and Engel, A. G. (1978). Intracellular calcium accumulation in Duchenne dystrophy and other myopathies: A study of 567,000 muscle fibers in 114 biopsies. Neurology 28, 439-446. Bonilla, E . , Samitt, C. E., Miranda, A. F., Hays, A. P., Salviati, G . , DiMauro, S., Kunkel, L. M.,
8. Duchenne Muscular Dystrophy
149
Hoffman, E. P., and Rowland, L. P., and Rowland, L. P. (1988). Duchenne muscular dystrophy: Deficiency of dystrophin at the muscle cell surface. Cell (Cambridge. Mass.) 54,447-452. Bornemann. A,, and Schmalbruch, H. (1991). Anti-dystrophin stains triadic junctions in regenerating rat muscles. Muscle Nerve, in press. Bourne, G. H., and Golarz, M. N. (1959). Human muscular dystrophy as an aberration of the connective tissue. Nature (London) 183, 1741- 1743. Bridges, L. R. (1986). The association of cardiac muscle necrosis and inflammation with the degenerative and persistent myopathy of mdx mice. J. Neurol. Sci. 72, 147-157. BrockdorfT, N., Cross, G . S., Cavanna, J. S., Fisher, E. M. C., Lyon, M. F., Davies, K. E., and Brown, S. D. M. (1987). The mapping of a cDNA from the human X-linked Duchenne muscular dystrophy gene to the mouse X chromosome. Nature (London) 328, 166-168. Bulfield, G., Siller, W. G., Wight, P. A,, and Moore, K. J. (1984). X Chromosome-linked muscular dystrophy (mdx) in the mouse. Proc. Natl. Acad. Sci. U.S.A. 81, 1189-1 192. Bundey, S. E. (1972). Extreme muscle hypertrophy in Duchenne muscular dystrophy. In “The Fifth Conference on the Clinical Delineation of Birth Defects” (D. Bergsma, ed.), Vol. X, p. 341. Williams & Wilkins, Baltimore, Maryland. Burghes, A. H. M., Logan, C., Hu, X., Belfall, B., Worton, R. G., and Ray, P. N. (1987). A cDNA clone from the DuchenndBecker muscular dystrophy gene. Nature (London) 328, 434-437. Burmeister, M., Monaco, A. P., Gillard, E. F., van Ommen, G. J. B., Affara, N. A,, FergusonSmith, M. A,, Kunkel, L. M., and Lehrach, H. (1988). A lomegabase physical map of human Xp21, including the Duchenne muscular dystrophy gene. Genornics 2, 189-202. Cannell, M. B., and Allen, D. G. (1984). Model of calcium movements during activation in the sarcomere of frog skeletal muscle. Eiophys. 1. 45, 913-925. Carafoli, E. (1987). Intracellular calcium homeostasis. Annu. Rev. Eiochem. 56, 395-433. Carpenter, J. L., Hoffman, E. P., Romanul, F. C. A., Kunkel, L. M., Rosales, R. K., Ma, N. S. F., Dasbach, J. J., Rae, J. F., Moore, F. M., McAfee, M. B., and Pearce, L. K. (1989). Feline muscular dystrophy with dystrophin deficiency. Am. J. Pathol. 135, 909-919. Chamberlain, J. S., Grant, S. G., Reeves, A. A,, Mullins, L. J., Stephenson, D. A,, Hoffman, E. P., Monaco, A. P., Kunkel, L. M., Caskey, C. T., and Chapman, V. M. (1987). Regional localization of the murine Duchenne muscular dystrophy gene on the mouse X chromosome. Somatic Cell Mol. Genet. 13, 671 -678. Chang, H. W., Bock, E., and Bonilla, E. (1989). Dystrophin in electric organ of Torpedo californica homologous to that in human muscle. J. Biol. Chem. 264, 20831-20834. Chapman, V. M., Miller, D. R., Armstrong, D., and Caskey, C. T. (1989). Recovery of induced mutations for X chromosome-linked muscular dystrophy in mice. Proc. Natl. Acud. Sci. U.S.A. 86, 1292-1926. Claman, H. N. (1990). Mast cells and fibrosis. The relevance to scleroderma. Rheum. Dis. Clin. N . Am. 16, 141-151. Cooper, B. J., Winand, N. J., Stedman, H., Valentine, B. A., Hoffman, E. P., Kunkel, L. M., Scott, M. O., Fischbeck, K. H., Kornegay, J. N., Avery, R. H., Williams, J. R., Schmickel, R. D., and Sylvester, J. E. (1988a). The homologue of the Duchenne locus is defective in X-linked muscular dystrophy of dogs. Nature (London) 334, 154-156. Cooper, B. J., Valentine, B. A., Wilson, S., Patterson, D. F., and Concannon, P. W. (l988b). Canine muscular dystrophy: Confirmation of X-linked inheritance. 3. Hered. 79, 405-408. Coulton, G. R., Morgan, J. E., Partridge, T. A., and Sloper, J. C. (1988a). The mdx mouse skeletal muscle myopathy: 1. A histological, morphometric and biochemical investigation. Neuropathol. Appl. Neurobiol. 14, 53-70. Coulton, G. R., Curtin, M. A., Morgan, J.E., and Partridge, T. A. ( I 988b). The mdx skeletal muscle myopathy: 11. Contractile properties. Neuropathol. Appl. Neurobiol. 14, 299-3 14. Cullen, M . J., Walsh, J., Nicholson, L. V. B., and Hams, J. B. (1990). Ultrastructural localization of
150
Eric P. Hoffman and Jose Rafael M. Gorospe dystrophin in human muscle using gold immunolabelling. Proc. R. SOC.London, Ser. B 240, 197-2 10.
D’Amore, P. A. (1990). Modes of FGF release in vivo and in virro. Cancer Merasrasis Rev. 9, 227238.
delahunta, A. (1983). “Veterinary Neuroanatomy and Clinical Neurology.” Saunders, Philadelphia, Pennsylvania. DiMario, J. Buffinger, N., Yamada, S., and Strohman, R. C. (1989). Fibroblast growth factor in the extracellular matrix of dystrophic (mdx) mouse muscle. Science 244, 688-690. Dreyfus, J. C., Schapira, G., and Schapira, F. (1954). Biochemical study of muscle in progressive muscular dystrophy. J. Clin. Invest. 33, 794-797. Duance, V. C., Stephens, H. R., Dunn, M., Bailey, A. J., and Dubowitz, V. (1980). A role for collagen in the pathogenesis of muscular dystrophy? Nature (London) 284, 470-472. Duchenne, G. B. (1968). Recherches sur la paralysie musculaire pseudo-hypertrophique ou paralysie myo-sclbrosique. Arch. Gen. Med. 6, 5 . Duncan, C. J. (1978). Role of intracellular calcium in promoting muscle damage: A strategy for controlling the dystrophin condition. Experientia 34, 1531-1535. Ebashi, S., and Sugita, H. (1979). The role of calcium in physiological and pathological processes of skeletal muscle. In “Current Topics in Nerve and Muscle Research” (A. J. Aguayo and G. Karpati, eds.), pp. 73-84. Excerpta Medica, Amsterdam. Emery, A. E. H. (1977). Muscle histology and creatine kinase levels in fetuses in DMD. Nature (London) 266, 472-473. Emery, A. E. H. (1988). “Duchenne Muscular Dystrophy” Oxford Monographs on Medical Genetics, Vol. 15. Oxford Univ. Press. Engel. D. (1976). A muscular dystrophy-like condition in rabbits. Res. Exp. Med. 169, 123-131. Engel, W. K. (1977). Integrative histochemical approach to the defect of Duchenne muscular dystrophy. In “Pathogenesis of Human Muscular Dystrophies” (L. P. Rowland, ed.), pp. 277-309. Excerpta Medica, Amsterdam. Ervasti, J. M., Ohlendieck, K., Kahl, S. D., Gaver, M. G., and Campbell, K. P. (1990). Deficiency of a glycoprotein component of the dystrophin complex in dystrophic muscle. Nature (London) 345, 315-319. Farber, J. L. (1982). Membrane injury and calcium homeostasis in the pathogenesis of coagulative necrosis. Lab. Invest. 47, 114- 123. Fond, P., ’nune.r, P. R., Denetclaw, W. F., and Steinhardt, R. A. (1990). Increased activity of calcium leak channels in myotubes of Duchenne human and mdx mouse origin. Science 250, 673-676.
Fowler, W. M., Gardner, G. W., Kazerunian, H. H., and Lauvstad, W. A. (1968). The effect of exercise on serum enzymes. Arch. Phys. Med. Rehabil. 49, 554-565. Frankel, K. A., and Rosser, R. J. (1976). The pathology of the heart in progressive muscular dystrophy: Epimyocardial fibrosis. Hum. Parhol. 7 , 375-386. Greenberg, C. R., Jacobs, H. K., Halliday, W., and Wrogemann. K. (1991). Three years’ experience with neonatal screening for Duchenne/Becker muscular dystrophy: Gene analysis, gene expression and phenotype prediction. Am. J. Hum. Genet. 39, 68-75. Grim, T. (1986). Becker muscular dystrophy. In “Myology: Basic and Clinical” (A. G . Engel and B. Q. Banker, eds.), pp. 1241-1250. McGraw-Hill, New York. Guibaud, P., Canier, H. N., Planchu, H., et al. (1981). Manifestations musculaires precoces, cliniques et histopathologiques, chez 14 garcons presentant dans la premiere annee une activite serique elevee de creatine phosphokinase. J . Genet. Hum. 29, 71-84. Hantai, D., Labat-Robert, J., Grimaud, J. A., and Fardeau, M. (1985). Fibronectin, laminin, type I, 111, and IV collagens in Duchenne’s muscular dystrophy, congential muscular dystrophies and congenital myopathies: An immunocytochemical study. Connect. Tissue Res. 13, 273-28 1.
8. Duchenne Muscular Dystrophy
151
Hearse, D. J. (1979). Cellular damage during myocardial ischaemia: Metabolic changes leading to enzyme leakage. In “Enzymes in Cardiology; Diagnosis and Research” (D. J. Hearse and J. de Leiris, eds.), pp. 1-20. Wiley, Chicester. Heilig, R., Lemaire, C., Mandel, J. L., Dandolo, L., Amer, L., and Avner, P. (1987). Nature (London) 328, 168-170. Hoffman, E. P., and Kunkel, L. M. (1989). Dystrophin abnormalities in Duchenne/Becker muscular dystrophy. Neuron 2, 1019-1029. Hoffman, E. P., Monaco, A. P., Feener, C. C., and Kunkel, L. M. (1987a). Conservation of the Duchenne muscular dystrophy gene in mice and humans. Science 238, 347-530. Hoffman, E. P., Brown, R. H., and Kunkel, L. M. (1987b). Dystrophin: The protein product of the Duchenne muscular dystrophy locus. Cell (Cambridge, Muss.) 51, 919-928. Hoffman, E. P., Knudson, C. M., Campbell, K. P., and Kunkel, L. M. (1987~).Subcellular fractionation of dystrophin to the triads of skeletal muscle. Nature (London) 330, 754-758. Hoffman, E. P., Fischbeck, K. H.,Brown, R. H., Johnson, M. Medori, R., Loike, J. D., Harris, J. B., Waterson, R., Brooke, M., Specht, L., Kupsky, W., Chamberlain, J., Caskey, C. T., Shapiro, F., and Kunkel, L. M. (1988a). Dystrophin characterization in muscle biopsies from Duchenne and Becker muscular dystrophy patients. N. Engl. J. Med. 318, 1363-1368. Hoffman, E. P., Hudecki, M., Rosenberg, P., Pollina, C., and Kunkel, L. M. (1988b). Cell and fiber-type distribution of dystrophin. Neuron 1, 41 1-420. Hoffman, E. P.,Kunkel, L. M., Angelini, C., Clarke, A., Johnson, M., and Hams, J. B. (1989). Improved diagnosis of Becker muscular dystrophy by dystrophin testing. Neurology 39, 10111017. Hu, X., Ray, P. N., Murphy, E. G . , Thompson, M. W., and Worton, R. G . (1990). Duplicational mutation at the Duchenne muscular dystrophy locus: Its frequency, distribution, origin, and phenotype/genotype correlation. Am. J. Hum. Genet. 46, 682-695. Hudgson, P., Pearce, G. W., and Walton, J. N. (1967). Pre-clinical muscular dystrophy: Histopathological changes observed on muscle biopsy. Brian 90, 565-576. Jackson, M. J., Jones, D. A., and Edwards, R. H. T.(1984). Experimental skeletal muscle damage: The nature of the calcium-activated degenerative processes. Eur. J. Clin. fnvesr. 14, 369374. Jones, D. A., Jackson, M. J., McPhail, G., and Edwards, R. H. T. (1984). Experimental mouse muscle damage: The importance of external calcium. Clin. Sci. 66, 317-322. Karpati, G., Pouilot, Y., Aubrzycka-Gaam, E., Carpenter, S., Ray, P. N., Worton, R. G., and Hollaned, P. (1989). Dystrophin is expressed in mdx skeletal muscle fibers after normal myoblast implantation. Am. J. Pathol. 135, 27-32. Knudson, C. M., Hoffman, E. P.! Kahl, S. D., Kunkel, L. M., and Campbell, K. P. (1988). Evidence for the association of dystrophin with the transverse tubular system in skeletal muscle. J. Biol. Chem. 263, 8480-8484. Koenig, M., Hoffman, E. P., Bertelson, C. J., Monaco, A. P., Feener, C., and Kunkel, L. M. (1987). Complete cloning of the Duchenne muscular dystrophy (DMD) cDNA and preliminary genomic organization of the DMD gene in normal and affected individuals. Cell (Cambridge, MUSS.)50, 509-517. Koenig, M., Monaco, A. P., and Kunkel, L. M. (1988). The complete sequence of dystrophin predicts a rod-shaped cytoskeletal protein. Cell (Cambridge, Mass. J 53, 219-228. Kornegay, J. N., ’hler, S. M., Miller, D. M., and Levesque, D. C. (1988). Muscular dystrophy in a litter of golden retriever dogs. Muscle Nerve 11, 1056-1064. Lemaire, C., Heilig, R., and Mandel, J. L. (1988). The chicken dystrophin cDNA: Striking conservation of the C-terminal coding and 3’ untranslated regions between man and chicken. EMBO J. 7, 4157-4162. Marshall, P. A., Williams, P. E., and Goldspink, G. (1989). Accumulation of collagen and altered
152
Eric P. Hoffman and Jose Rafael M. Gorospe fiber-type ratios as indicators of abnormal muscle gene expression in the mdx dystrophic mouse. Muscle Nerve 12, 528-537.
Maunder-Sewry, C.A., Corodetsky, R., Yaron, R., and Dubowitz, V. (1980). Elemental analysis of skeletal muscle in Duchenne muscular dystrophy. Muscle Nerve 3, 502-508. McComas, A. J., Sica, R. E., and Campbell, R. E. (1971). “Sick” motor neurones. A unifying concept of muscle disease. Lancet 1, 321-325. McCully, K. K., and Faulkner, J. A. (1985). Injury to skeletal muscle fibers of mice following lengthening contractions. J . Appl. Physiol. 59, 119-126. Mendel, J. R., Engel, W. K., and Derrer, E. C. (1971). Duchenne muscular dystrophy: Functional ischemia reproduces its characteristic lesions. Science 172, 1143-1 145. Mendel, J. R., Engel, W. K., and Derrer, E. C. (1972). Increased plasma enzyme concentrations in rats with functional ischaemia of muscle provide a possible model of Duchenne muscular dystrophy. Narure (London) 239, 522-524. Meng, G.,Kress, W., Scherpf, S., Bettecken, T., Feichtinger, W., Schempp, W., Schmid, M., and Muller, C. R. (1991). A comparison of the dystrophin gene structure in primates and lower vertebrates. In “Muscular Dystrophy Research: From Diagnosis to Therapy” (C. Angelini, G. A. Damieli, and D. Fontanari, eds.), pp. 23-30. Excerpts Medica, Amsterdam. Menke, A., and Jockusch, H. (1991). Decreased osmotic stability of dystrophin-less muscle cells from the mdx mouse. Nature (London) 349, 69-71. Miyatake, M., Miike, T.,Zhao, J., Yoshioka, K., Uchino, M., and Usuku, G. (1989). Possible systemic smooth muscle layer dysfunction due to a deficiency of dystrophin in Duchenne muscular dystrophy. J. Neurol. Sci. 93, 11-17. Miyatake, M., Miike, T., Zhao, J., Yoshioka, K . , Uchino, M., and Usuku, G.(1991). Dystrophin: Localization and presumed function. Muscle Nerve 14, 113- 119. Mokri, B., and Engel, A. G.(1975). Duchenne dystrophy: Electron microscopic findings point to a basic or early abnormality in the plasma membrane of muscle fibers. Neurology 25, 1 I 1 1- 1 120. Monaco, A. P., Neve, R. L., Colletti-kener, C . , Bertelson, C. J., Kurnit, D. M., and Kunkel, L. M. (1986). Isolation of candidate cDNAs for portions of the Duchenne muscular dystrophy gene. Nature (London) 323, 646-650. Monaco, A. P., Bertelson, C. J., Liechti-Gallati, S., Moser, H., and Kunkel, L. M. (1988). An explanation for the phenotypic differences between patients bearing partial deletions of the DMD locus. Genomics 2, 90-95. Mongini, T., Ghigo, D., Doriguzzi, C., Bussolino, F., Pescarmona, G . , Pollo, B., Schiffer, D., and Bosia, A. (1988). Free cytoplasmic Ca+ + at rest and after cholinergic stimulus is increased in cultured muscle cells from Duchenne muscular dystrophy patients. Neurology 38, 476-480. Morandi, L., Mora, M., Gussoni, E., Tedeschi, S., and Cornelio, F. (1990). Dystrophin analysis in Duchenne and Becker muscular dystrophy carriers: Correlation with intracellular calcium and albumin. Ann. Neurol. 28, 674-679. Morgan, J. E., Hoffman, E. P., and Partridge. T. A. (1990). Normal myogenic cells from newborn mice restore normal histology to degenerating muscles of the mdx mouse. J. Cell Biol. 111, 2437-2749. Newham, D. J., McPhail, G . , Mills, K. R., and Edwards, R. H. T. (1983). Ultrastructural changes after concentric and eccentric contractions of human muscle. J. Neurol. Sci. 61, 109-122. Nicholson, L. V. B., Davison, K., Falkous, G . , Harwood, C., O’Donnell, E., Slater, C. R., and Harris, J. B. (1989). Dystrophin in skeletal muscle. I. Western blot analysis using a monoclonal antibody. J. Neuro/. Sci. 94, 125-136. Ohno, S., Becak, W., and Becak, M. L. (1964). X-Autosome ratio and the behavior pattern of individual X-chromosomes in placental mammals. Chromosoma 15, 14-30. Olwin, B. B., and Hauschka, S. D. (1986). Identification of the fibroblast growth factor receptor of Swiss 3T3 cells and mouse skeletal muscle myoblasts. Biochemistry 25, 3487-3492.
8. Duchenne Muscular Dystrophy
153
Partridge, T. A. (1991). Myoblast transplantation: A possible therapy for inherited myopathies? Muscle Nerve 14, 197-212. Partridge, T. A,, Morgan, J. E., Coulton, G . R., Hoffman, E. P., and Kunkel, L. M. (1989). Conversion of mdx myofibers from dystrophin-negative to dystrophin-positive by injection of normal myoblasts. Nature (London) 337, 176-179. Patel, K.,Voit, T., Dunn, M. J., Strong, P. N., and Dubowitz, V. (1988). Dystrophin and nebulin in the muscular dystrophies. J. Neurol. Sci. 87, 315-326. Perloff, J. K., de Leon, A. C., and O’Doherty, D. (1966). The cardiomyopathy of progressive muscular dystrophy. Circulation 33, 625-648. Perloff, J. K., Roberts, W. C., de Leon, A. C., and O’Doherty, D. (1967). The distinctive electrocardiogram of Duchenne’s progressive muscular dystrophy. An electrocardiographic-pathologic correlative study. Am J. Med. 42, 179-188. Perloff, J. K., Henze, E., and Schelbert, H. R. (1984). Alterations in regional myocardial metabolism, perfusion, and wall motion in Duchenne muscular dystrophy studied by radionuclide imaging. Circulation 69, 33-42. Quinlan, J. G., Johnson, S. R., and Samaha, F. J. (1990). Dantrolene normalizes serum creatine kinase in mdx mice. Muscle Nerve 13, 268-269. Rifkin, D. B., and Moscatelli, D. (1989). Recent developments in the cell biology of basic fibroblast growth factor. J. Cell Biol. 109, 1-6. Rothe, M. J., and Kerdel, F. A. (1991). The mast cell in fibrosis. Int. J . Dermatol. 30, 1316. Rowland, L. P. (1964). Muscular dystrophies and related diseases: Metabolic aspects. Manitoba Med. Rev. 44,540-545. Rowland, L. P. (1980). Biochemistry of muscle membranes in Duchenne muscular dystrophy. Muscle Nerve 3, 3-20. Salviati, G., Betto, R., Ceoldo, S., Biasia, E., Bonilla, E., Miranda, A. F., and DiMauro, S . (1989). Cell fractionation studies indicate that dystrophin is a protein of surface membranes of skeletal muscle. Biochem. J. 258, 837-841. Schanne, F. A. X., Kane, A. B., Young, E. E., and Farber, J. L. (1979). Calcium dependence of toxic cell death: A final common pathway. Science 206, 700-702. Sicinski, P., Geng, Y.,Ryder-Cook, A. S., Barnard, E. A., Darlison, M. G., and Barnard, P. 1. (1989). The molecular basis of muscular dystrophy in the mdx mouse: A point mutation. Science 244, 1578-1580. Sugita, H., Arahata, K., Ishiguro, T., Suhara, Y.,Tsukahara, T.,Ishiura, S . , Eguchi, C., Nonaka, I., and Ozawa, E. (1988). Negative immunostaining of Duchenne muscular dystrophy and mdx muscle surface membrane with antibody against synthetic peptide fragment predicted from DMD cDNA. Proc. Jpn. Acad. 64, 37-39. Turner, P. R., Westwood, T., Regen, C. M., and Steinhardt, R. A. (1988). Increased protein degradation results from elevated free calcium levels found in muscle from mdx mice. Nature (London) 335, 735-738. Turner, P. R., Fong, P., Denetclaw, W. F., and Steinhardt, R. A. (1991). Increased calcium influx in dystrophic muscle. In press. Vaca, K., Stewart, S . S . , and Appel, S. H. (1989). Identification of basic fibroblast growth factor as a cholinergic growth factor from human muscle. J. Neurosci. Res. 23, 55-63. Valentine, B. A., Cooper, B. J., Cummings, 1. F., and De La Hunta, A. (1986). Progressive muscular dystrophy in a golden retriever dog: Light microscope and ultrastructural features at 4 and 8 months. Acta Neuropathol. 71, 301-310. Valentine, B. A., Cooper, B. J., De La Hunta, A , , O’Quinn, R., and Blue, J. T. (1988). Canine Xlinked muscular dystrophy; an animal model of Duchenne muscular dystrophy: Clinical studies. J. Neurol. Sci. 88, 69-81.
154
Eric P. Hoffman and Jose Rafael M. Gorospe
Valentine, B. A., Komegay, J. N., and Cooper, B. J. (1989a). Clinical electromyographic studies of canine X-linked muscular dystrophy. Am. J. Vet. Res. 50, 2145-2147. Valentine, B. A., Cummings, J. F., and Cooper, B. J. (1989b). Development of Duchenne-type cardiomyopathy: Morphologic studies in a canine model. Am. J. Parhol. 135,671-678. Valentine, B. A., Cooper, B. J., Cummings, J. F., and delehunta, A. (1990). Canine X-linked muscular dystrophy: Morphologic lesions. J . Neurol. Sci. 97, 1-23. Watkins, S . C., Hoffman, E. P., Slayter, H. S . , and Kunkel, L. M. (1988). Immunoelectron microscopic localization of dystrophin in myofibres. Nature (London) 333, 863-866. Wolff, J. A., Malone, R. W., Williams, P., Chong, W., Acsadi, G., Jani, A,, and Felgner, P. L. (1990). Direct gene transfer into mouse muscle in viva Science 247, 1465-1468. Wrogemann, K.,and Pena, S . D. J. (1976). Mitochondria1 calcium overload: A general mechanism for cell-necrosis in muscle diseases. Lancet 1, 672-674. Younkin, D. P., Berman, P., Sladky, J., Chee, C., Bank, W., and Chance, B. (1987). 31P NMR studies in Duchenne muscular dystrophy: Age-related metabolic changes. Neurology 37, 165169. Zubrzycka-Gaam, E. E., Bulman, D. E., Karpati, G., Burghes, A. H. M., Belfall, B., Hajklamut, H., Talbot, J., Hodges, R. S . , Ray, P. N., and Worton, R. G. (1988). The Duchenne muscular dystrophy gene product is localized in the sarcolemma of human skeletal muscle fibres. Nature (London) 333, 466-469.
CHAPTER 9
Mutant Cytoskeletal Proteins in Hemolytic Disease Sally L. Marchesi Department of Pathology and Laboratory Medicine Yale Medical School New Haven, Connecticut 06510
1. Introduction: Membrane/Skeleton Symposium 11. Detection and Characterization of Spectrin Mutants in Hereditary Elliptocytosis 111. Mutant Forms of Protein 4.1 in Hereditary Elliptocytosis IV. Hereditary Spherocytosis References
1. INTRODUCTION: MEMBRANEEKELETON SYMPOSIUM
During the years 1980-1990, substantial progress was made in determining the molecular basis of inherited hemolytic anemias known as hereditary spherocytosis (HS) and elliptocytosis (HE). The observed changes in red cell shape in these disorders from biconcave disc to spheres, ovalocytes, elliptocytes, and, in some cases, fragments suggested missing or abnormal membrane components. However, there was little direct evidence to support this concept until recently. During the past 5 years, HE has been shown to be the result of a variety of mutations in spectrin and protein 4.1, two major components of the erythrocyte cytoskeleton (Palek, 1987; McGuire et al., 1988). Deficiency of ankyrin, the third major component of the erythrocyte skeleton, has recently been shown to cause severe spherocytic hemolytic anemia (Lux ef al., 1990). A recessive form of HS, less common than dominantly inherited spherocytosis but with more severe hematological consequences, has been shown to be associated with a point mutation in the ninth repeat unit of the spectrin ci subunit (Marchesi et al., 1989), but the pathophysiological consequences of this mutation are still obscure. Current Topics in Membranes, Volume 38 Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.
155
156
Sally L. Marchesi
157
9. Mutant Cytoskeletal Proteins
The most abundant protein of the erythrocyte cytoskeleton is spectrin, so named because it was the first protein to be purified from red cell “ghosts.” Spectrin in composed of a and P subunits, long filamentous proteins of 280 and 246 kDa, respectively, which interact with one another noncovalently head to head and side to side to form heterodimers, tetramers, and higher oligomers. Attachment of this “matrix” of spectrin filaments to the overlying bilipid layer is mediated by ankyrin, a globular 210-kDa protein which also binds to the transmembrane anion channel (band 3). Spectrin also associates with protein 4.1, an 80-kDa protein, in the presence of short actin filaments. The binding site for protein 4.1 and actin is near the N-terminus of P-spectrin and the C-terminus of the a subunit at a site homologous with the actin-binding region of a-actinin. Triton-extracted cytoskeletons from normal erythrocytes form a hexagonal latticework composed of junctional complexes (actin and 4.1 ), cross-linked by spectrin tetramers and occasional higher spectrin multimers, with ankyrin attached near the spectrin self-association site (Liu et al., 1987) (Fig. 1). The function of the cytoskeleton is to give the red cell the stability and deformability required for its passage intact through the smallest capillaries of the circulatory system. Mutation of any cytoskeletal component that impairs its deformability and mechanical stability may significantly decrease red cell survival.
II. DETECTION AND CHARACTERIZATION OF SPECTRIN MUTANTS IN HEREDITARY ELLIPTOCYTOSIS During the period 1980-1983, Speicher el al. (1980, 1982) developed a chemical domain map of the spectrin a and P subunits comprised of five tryptic “domains” in the a subunit (aI-aV) and four “domains” in the P subunit (PIPIV). These large tryptic peptides of spectrin and their daughter peptides were mapped in two dimensions according to O’Farrell (1975) (Fig. 2). Maps of this ~
~~
FIG. 1 Spread membrane skeleton examined by negative-staining electron microscopy. Membrane skeletons derived from Triton-treated (2.0%) normal red cell ghosts were isolated by sucrose density gradient centrifugation under hypotonic conditions. The skeletons were applied to thin carbon-coated grids, fixed with glutaraldehyde, stained with uranyl acetate, air dried, and examined by transmission electron microscopy. (a) A large area of spread meshwork is shown, revealing the marginal region of the exposed bottom layer of the skeleton. (b) A higher magnification of the spread meshwork reveals the hexagonal lattice of junctional complexes, presumably containing short F-actin and band 4. I , cross-linked by spectrin tetramers (Sp4). three-armed spectrin molecules (Sp6), and double spectrin filaments (two Sp4). Globular structures of ankyrin (or ankyrin-containing complexes) are attached to spectrin filaments at the ankyrin-binding site, i.e., 80 nm from the distal end of spectrin. (c) A tentative assignment of these structural elements. Reproduced from Liu et a l., (1987) with permission.
158
Sally L. Marchesi
type, both one- and two-dimensional, have been used extensively by ourselves and other laboratories for detection of mutations in spectrin from red cells of individuals with HE and HS. The first test of reproducibility and usefulness of the spectrin tryptic map for detection of abnormal spectrins came during a study of a large family with dominantly inherited HS (Knowles ef al., 1984). Spectrin maps from some members of this family showed changes in isoelectric point and molecular weight of aII-domain peptides which did not correlate with the presence of spherocytosis. However, the aII-domain variants were inherited according to simple Mendelian genetics. Eventually, three variants of the aII-domain peptides were recognized (Fig. 3) involving a small basic shift in isolectric point (a11 type 4), a 4-kDa increase in
FIG. 2 Two-dimensional map (IEF/SDS) of normal spectrin peptides after tryptic digestion for 90 min at 2°C at an enzyme : protein ratio of I : 20. Five tryptic domains of the a subunit and four domains of the p subunit are identified. Subdomain peptides of all, alII, and a l V and PI1 and PIV are also shown.
9. Mutant Cytoskeletal Proteins
159
FIG. 3 Polymorphisms of the spectrin all domain. “Wild-type” (type 1) a11 domain peptides at -46 and 35 kDa are shown on the tryptic map at the upper left. Polymorphic forms of a11 may have an apparent increase in molecular mass of -4 kDa (type 3). a basic shift in isolectric point (type 4),or both molecular mass and IEP shift (type 2). The drawing (lower right panel) shows the position of all four variants of a11 46- and 35-kDa peptides. Reproduced from Knowles et al.. (1984). with permission of the American Society of Clinical Investigation.
apparent molecular mass (a11 type 3), or shifts in both isoelectric point and molecular mass (a11 type 2). (The “wild type” a11 is called type 1). These polymorphisms are common in the Black population and do not, as the designation “polymorphism” suggests, appear to have any biological consequence. The amino acid substitutions responsible for these polymorphisrns have recently been determined by D. W. Speicher (personal communication). Despite the fact that the first abnormalities detected on spectrin maps in a
160
Sally L. Marchesi
family with hemolytic disease proved to be non-disease-related polymorphisms, their reproducibility gave us confidence that the maps would be useful in detecting disease-related mutations in other families, and, indeed, that turned out to be true. Tryptic mapping of spectrin from families with HE has identified a large number of mutations in the a1 domain which cause increased susceptibility to tryptic digestion of the N-terminal domain of the OL subunit (aI). The presence of these mutations is detected on the maps as a quantitative
FIG. 4 Spectrin tryptic maps of subjects with HE and HPP and mutations in the a1 80-kDa domain. Open triangles identify aI/80. (A) Normal control. (B) Subject with HPP and two mutations which destabilize the a1 80-kDa domain, resulting in the appearance of new a 1 peptides at 74 and 50 kDa (aI/74 and aI/SOa), which are circled. (C) Subject with HE and a mutation in the a 1 domain giving rise to new basic a 50-kDa peptide (circled) called aI15Ob. (D) Subject with H E and a mutation in the a1 domain resulting in the appearance of a new 68-kDa peptide (circled) called aI/68. Polymorphism of the a11 domain and a111 domains are indicated by arrows. Reproduced from Marchesi er al., (1987). with permission of the American Society of Clinical Investigation.
9. Mutant Cytoskeletal Proteins
161
decrease in the a1 80-kDa parent domain peptide, and appearance of new peptides at 78, 74, 65-68 kDa, or one of two 50-kDa peptides with different isolectric points (50a and 50b) (Knowles et al., 1983, Marchesi el al., 1986, 1987), all derived from the a1 domain. Figure 4 shows spectrin maps of three unrelated subjects with HE, one of whom has a clinically severe variant of HE called hereditary pyropoikilocytosis (see below). The appearance of new tryptic peptides on the mutant spectrin maps is thought to be the result of regional unfolding of the molecule as a consequence of a nonconservative amino acid substitution, e.g., substitution of proline for serine in a region of helical conformation. Figure 5 shows maps of affinity-purified a1 peptides from normal and mutant spectrins. Figure 6 gives the N-terminal sequences of the normal spectrin
FIG. 5 Affinity-purified a1 domain tryptic peptides of spectrin from a normal subject and from three subjects with HE and a mutant (unstable) spectrin a1 domain. Spectrin was digested in all cases at 4°C at an enzyme : substrate ratio of of I : 20. (A) Normal aU80-kDa is present and intact except for a small mount of aI/74. (B) Mutant aI/80-kDa domain is cleaved by trypsin to peptides of 50.30, and 21 kDa (mutation a1/50a). (C) Mutant aI/BO-kDa domain is cleaved by trypsin to a basic 50-kDa peptide (50b) and smaller peptides at 19 and 17 kDa (mutation a/50b). (D) Mutant aI/80-kDa domain is partially cleaved to a 68-kDa fragment (mutation (~1168).Reproduced from Marchesi e t a / ., (1987). with permission of the American Society of Clinical Investigation.
162
Sally L. Marchesi
N A
AR
Ts
N B
HB
I
q YEQxLDl!&,
DF
QYEQxLDJUL, dl80 +T68 m
N C
SG
HP
[ [ [
, TBO
hlLWDLLLELTLEKGDQLLRALKFQQYVQEC xLWDLLLELtLEKGDQLaRALKfQQYME T68
I I
HLWDLLLELTLEKGWLLORALWQ~YVQE.T~~,
FIG. 6 Partial amino acid sequences of aI/8O-kDa peptides from normal and mutant spectrins. A proline is substituted for the normal residue in three of the mutations shown. The mutant aI/68kDa peptide is the result of insertion of a leucine 17-19 residues distant from the tryptic cleavage site. Reproduced from Marchesi et al., (1987), with permission of the American Society of Clinical Investigation.
a1 domain and of a1 peptides of -68* and 50 kDa (50a and 50b) which appear on maps of subjects with HE and unstable a1 domains. Figure 7 shows a model of the a1 domain proposed by Speicher (see Marchesi er al., 1987) which places the mutant sites at homologous points within the first,
second, and fourth complete repeat units. Mutations in the spectrin a1 domain are characteristically associated with impairment of spectrin dimer self-association to form higher oligomers (Fig. 8), which is associated with decreased membrane stability as measured by ektacytometry (Knowles et al., 1983). Hemolytic disease resulting from mutations in the a1 domain ranges from mild ovalocytosis without significant hemolysis to hereditary pyropoikilocytosis (HPP), characterized by striking fragmentation of red cells and severe hemolytic anemia requiring splenectomy to normalize the hematocrit (Fig. 9). Recent studies (Iarocci el al., 1988) suggest that HPP is the consequence of compound *Molecular masses given for a1 peptides are estimates from the tryptic maps. aI/68 is called a1/65 in other studies.
9. Mutant Cytoskeletal Proteins
163
FIG. 7 Model of repeat units of the a1 domain according to Speicher (13). 0, Helical segments; r\ , joining s e g m e n t s ; m , pleated sheet. Sites of the mutations described in Fig. 6 are indicated by an X . The locations of new tryptic peptides produced as a result of these mutations are indicated by arrows. Reproduced from Marchesi ei a / . , (1987). with permission of the American Society of Clinical Investigation.
FIG. 8 Nondenaturing gel electrophoresis of spectrin from normal, HE, and HPP subjects after concentration of spectrin and incubation at the indicated concentrations for 3 hr at 37°C ( 1 1). D, Dimers; T, tetramers; 0, higher oligomers. The broad band near the origin in the HPP patient is hemoglobin bound noncovalently to the membrane and extracted from the ghosts with spectrin. Reproduced from Knowles et al., (1983), with permission of the American Society of Clinical Investigation.
164
Sally L. Marchesi
FIG. 9 Peripheral blood smears of (A) normal subject, (B) subject with mild sphero- and ovalocytosis without significant hemolysis, (C) subject with hereditary elliptocytosis (HE) and compensated hemolysis, (D) subject with hereditary pyropoikilocytosis, a variant of HE associated with brisk hemolysis and anemia.
heterozygosity for two discrete mutations in the spectrin a subunit. This condition is illustrated in Fig. 4B, which shows aI/5Oa and aI/74 in spectrin of a boy with HPP. Table I, Marchesi et al. (1987), Morlt ef al. (1989, 1990) Lecomte et al. (1989a), Garbarz er al. (1990), Floyd et al. (1990), Tse et al. (1990), Roux er al. (l989), Gallagher et al. (1990a), and T. Coetzer (unpublished), describes 15 point mutations in the a1 domain and one in the PI domain of spectrin from subjects with HE, all of which destabilize the normally trypsin-resistant a1 domain, resulting in the appearance of new tryptic peptides on the spectrin maps. It is interesting to note that, while most of the new tryptic cleavage sites in the mutant spectrins are near (one or two residues away from) the mutational site, others may be separated by more than 50 residues (Gallagher et al., 1990). While
165
9. Mutant Cytoskeletal Proteins
TABLE I Spectrin Mutations Associated with the Appearance of New a1 65/68 and 50-kDa Tryptic Peptideso Spectrin subunit a a (I
a a a a a
a
P a
~~~~
Codon
Substitution
CGG-, TGG Arg+ Trp AGG+ AGT 45 Arg+ Ser GGT-, GIT 46 Gly+ Val c T T + TIT 49 Leu-, Phe CGT+ CAT 28 Arg-, His CGT-, TGT 28 Arg+ Cys AAG+ AGG 48 Lys-, Arg CGT+ CIT 28 Arg+ Leu CGT+ AGT 28 Arg+ Ser GCT-, CCT 2053 Ala+ Pro (C-term) TTG repeat 155 + Leu 41
a
155
TTG repeat
a
260
a
26 1
a
207
a
47 1
CTG+ LeuTCC+ Ser+ CTG-, Leu-, CAG+ Gln+
New tryptic cleavage site
New tryptic peptide
Spectrin name
Lys 16
a1178
Tunis
Lys 16
a1178
> 45
(11174
Culoz
Lys 45 > 48
a1174
Lyon
Lys 48
a1174
Lys 48
a1174 a1174 a1174 a1174 Lys 48
a1174
Arg 138
a1165
(11165
+ Leu
CCG Pro CCC Pro CCG Pro CCG Pro
Arg 256
(1115Oa
Lys 258
aI15Oa
Lys 258
aI15Oa
Arg 468
aI15Ob
Ref. MorlC et al. (1989) Lecomte et al. (1998a) MorlC et al. ( 1990) MorlC et al. ( 1990) Garbarz et al. ( 1990) T. Coetzer (unpublished) Floyd er al. ( 1990) Floyd er al. ( 1990) Floyd et al. ( 1990) Tse et al. ( I 990) Marchesi et al. (1987), Roux et al. (1989) Roux et al. ( 1989) Marchesi et al. (1987) Marchesi ez al. (1987) Galagher et al. ( 1990a) Marchesi er al. (1987)
__________
Assignments of ccdon number and tryptic cleavage sites are based on the recently completed sequence of the N-terminus of the 01 subunit, beginning with the N-terminal methionine. a
detailed information on the tertiary structure of the spectrin ci subunit is incomplete, it is thought that the extensive folding and stacking of helices within the repeat units of the ci subunit results in physical proximity of a mutant site and a tryptic cleavage site which may be 50 residues apart by sequence analysis. Furthermore, abnormal tryptic cleavage of the a1 80-kDa peptide near its N-
166
Sally L. Marchesi
terminus to form a 74-kDa fragment results not only from mutation on the a subunit but from mutations and deletions near the C-terminus of the p subunit (Tse et al., 1990; Pothier et al., 1987). Truncated forms of a a and p subunits of spectrin are associated with HE. (Table 11) The sites of deletion in three of the four shortened p subunits reported to date (Lecomte et al., 1989b; Gallagher et al., 1990b; Pothier et al., 1987; Dhermy et al., 1982; Gallagher er al., 1990~;Ohanian et al., 1985; Eber et al., 1988) are all near the C-terminus in exons X or Y. As mentioned above, the truncated p subunit of spectrin Nice (Pothier et al., 1987; Gallagher et al., 1990c) destabilizes the adjacent a subunit, evidenced by the increase in a1/74 peptide on spectrin maps. The sites of deletion of the two shortened ct subunits (Lane et al., 1987; S . L. Marchesi and J. Winkelmann, unpublished observations) are not known, although spectrin ctM (Lane et al., 1987) demonstrates decreased dimer-dimer association, suggesting that the deletion is near the Nterminus of the ct subunit. From the foregoing, it is evident that characterization of mutant spectrins found in families with HE has not only provided information on the molecular mechanisms of hemolytic disease, but has given us new insights into the essential elements of spectrin conformation required for the formation of a normal cytoskeleton.
TABLE I1 Truncated Spectrin Q and P Subunits Associated with HE Spectrin subunit
Decrease in masso (ma)
P
-2
P
-4
B
-6
P
-16 -5- 10
ct Q
-5-10
Spectrin name
New tryptic peptide
Site of mutation
Rouen
a1174
Exon Y (deleted) Exon X Exon X (deleted) ?
Lehy
? ?
Nice
QM
a'
Ref. Lecomte er al. (1989b), Gallagher et al. (1990b) Pothier et al. (1987). Gallagher et al. ( 1 9 9 0 ~ ) Dhermy el al. (1982). Gallagher et al. ( 1 9 9 0 ~ ) Ohanian et al. (1985) Lane et al. (1987) S. L. Marchesi and J . Winkelmann (unpublished observations) ~
Apparent changes in mass of (I and p subunits are based on old estimates from gel electrophoresis of 240 and 220 kDa for (I and p subunits, respectively. Actual masses of a and p subunits derived from complete sequence are 279 and 246 kDa. respectively. Thus, deletions may be larger than the earlier estimates. a
9. Mutant Cytoskeletal Proteins
167
111. MUTANT FORMS OF PROTEIN 4.1 IN HEREDITARY ELLIMOCYTOSIS Normal protein 4.1 is a doublet of 80- and 78-kDa peptides which, in the presence of short actin filaments, binds spectrin near the N-terminus of the p subunit (Bennett, 1985). Protein 4.1 is also thought to aid in binding of the cytoskeleton to the bilipid layer via its interaction with glycophorins A and C (Anderson and Lovrien, 1984). Deficiency of protein 4.1 was first discovered in a consanguineous Algerian family with HE (Fe6 et af., 1980). Erythrocytes of one of the three affected children were entirely lacking in protein 4.1, resulting in severe hemolytic anemia. We were given the opportunity to study the role of 4.1 in hemolytic disease by P. Agr6, who provided us with blood from two families, one of which (family N) is heterozygous for a high-molecular-mass form of 4.1 at -95 kDa, compared with -80 kDa for normal 4.1; affected members of the other family ( G ) are heterozygous for normal 4.1 and two low-molecular-mass species of 4.1 at -68 and 65 kDa (McGuire et al., 1988). The presence of the high-molecular-mass form of 4.1 in family N resulted in mild ovalocytosis without significant shortening of red cell survival, while the presence of low-molecular-mass 4.1 of family G resulted in severe HE. In collaboration with J. Conboy and N. Mohandas at University of California, San Francisco, the mutant 4.1 species were characterized at both protein and DNA levels (Marchesi ef al., 1990; Conboy et al., 1990). Immunoblots using antibodies to synthetic peptides of protein 4.1 showed that patient N’s membranes reacted with antibodies to all four chymotryptic domains of 4.1 (38), while patient G’s membranes reacted with all anti-4.1 antibodies except for antibody to the 8-kDa spectriniactin-binding domain (Fig. 10; Let0 and Marchesi, 1984; Correas et al., 1986). We evaluated the abnormal 4.1 molecules from these families using two-dimensional peptide mapping methods developed by Let0 and Marchesi (1984). The N-terminal sequence of two chymotryptic peptides of high-molecularmass 4.1 suggested that the insertion resulting in 4. 195 might be repeated sequence. Amplification of DNA from this region of the molecule by the polymerase chain reaction (PCR) led to the conclusion that 4.195 is the result of a three-exon repeat which includes the spectridactin-binding domain (Fig. 11; Conboy et al., 1990). Studies of patient G’s low-molecular-mass 4.1 (4. 165) by similar methods showed a two-exon deletion which included the 8-kDa spectrin/actin-binding domain (Fig. 11, Conboy et al., 1990). Thus, the patient is haploid with respect to the spectrin/actin-binding domain of 4.1, resulting in his severe hemolysis. The deformability index (by ektacytometry) of red cell ghosts containing 4. 195 was normal, but deformability of ghosts containing 4. 16* was markedly abnor-
--b
30
16
10
___f
24
FIG. 10 Western blots of red cells ghosts electrophoresed in 8% acrylamide and immunoblotted with antibodies to synthetic peptides of the four chymotryptic domains of 4.1. C, Control; G, subject with low-molecular-mass 4.1; N, subject with high-molecular-mass 4.1. Reproduced from Marchesi er al., (1990), with permission of the American Society of Clinical Investigation.
FIG. 11 Protein structure and partial gene map of normal 4. IS0 and of high-molecular-mass 4.1 (4.195) and low-molecular-mass 4. I (4.168/65)(Marchesi e t a / . , 1990; Conboy er a/.,1990). Boxes represent exons; numbers above each box indicate amino acids encoded by that exon. Three exons are duplicated in 4.195 and two exons are deleted in 4.168165. Reproduced from Conboy er al., (1990). with permission of the American Society of Clinical Investigation.
9. Mutant Cytoskeletal Proteins
169
ma1 (Marchesi et al., 1990). The abnormal deformability index of 4.168 was partially reversed by addition of normal 4.1 by exchange hemolysis.
N. HEREDITARY SPHEROCYTOSIS Hereditary spherocytosis is usually dominantly inherited and classically results in mild to moderate hemolysis and anemia, occasionally requiring splenectomy to normalize the hematocrit and reduce the reticulocytosis. The disorder is readily diagnosed by family history, peripheral blood smear, and by the increase in osmotic fragility of red cells incubated overnight. Despite the ready availability of blood samples from subjects with dominant HS for studies of red cell membranes, the molecular basis of HS has remained obscure. In 1982, independent studies of two families with HS suggested that spherocytosis may be the result of an abnormality in binding of spectrin to protein 4.1 (Goodman et al., 1982; Wolfe cr al., 1982). One of these families has recently been restudied in detail (Becker et al., 1987), with the finding of increased oxidant sensitivity and abnormal tryptic cleavage of the PIV domain associated with impaired binding to protein 4.1. It is not clear yet how generally this abnormality will be found in other families with HS. In 1985, Agr6 et al. described nondominant HS in five families from North Carolina. The index patients were two sisters presenting with a severe hemolytic anemia in infancy. Peripheral blood smears showed marked macro- and microspherocytosis. The patients were transfusion dependent and splenectomy was required early in childhood, in contrast to the relatively mild course of dominant HS. Spectrin tryptic maps of four of these families showed a subtle but reproducible change in the a11 domain peptides (Fig. 12; Marchesi, 1989). In the “heterozygous” condition, the normal a11 domain peptides are present, but in addition there is a new acidic peptide at the same molecular mass which was shown to be derived from the a11 domain by Iz5I mapping according to Elder et al. (1977). In the “homozygous” condition, the acidic shift in isoelectric point of d 4 6 - k D a peptides is again present and the most basic of the normal a11 domain peptides is absent. The acidic shift in isoelectric point is present in the daughter a11 peptides at 35 kDa, but not at 30 kDa (Fig. 12). We have now studied over 20 families in which one or more members shows the “aIIa” trait (Marchesi et al., 1989). In most cases, the parents are normal hematologically, while affected children have moderate to severe hemolytic anemia associated with macro- and microspherocytosis. With approximate knowledge of the site of tryptic cleavage producing the 35- and 30-kDa peptides, deduced from N-terminal sequence (D. w. Speicher, personal communication), Forget and colleagues were able, by use of PCR, to identify the amino acid substitution responsible for the acidic shift in
170
Sally L. Marchesi
FIG. 12 Spectrin tryptic maps of subjects with recessive spherocytosis showing an acidic shift in a11 peptides at 46, 35, and 30 kDa (circled). In the map designated alla, normal a11 peptides at 46 and 35 kDa are present but are joined by an acidic all peptide (poorly resolved at 35 kDa). In the map designated aIIaa, the acidic 46-kDa peptide is prominent and the most basic 46-kDa peptide is lost. The a11135 peptide but not a11130 is also shifted in the acidic direction. Reproduced from Marchesi, (1989, pp. 102-103) by courtesy of Marcel Dekker, Inc.
spectrin a11 peptides as a11 309 GCT + GAT, which substitutes an aspartic acid for alanine (Marchesi et al., 1989). Slot blots of DNA from patients and family members with recessive HS and “spectrin aIIa” using normal and mutant probes in the region of the GCT + GAT mutation confirmed the information derived from protein maps. Models of spectrin place the mutation within the ninth repeat unit of the a subunit, adjacent to the nonhomologous tenth repeat unit. We have no information on specific functions or interactions of spectrin in this region, and thus we do
9. Mutant Cytoskeletal Proteins
171
not know how this substitution affects the structure of the spectrin a subunit and its behavior within the cytoskeleton. Finally, deficient or abnormal ankyrin has been recently implicated in several individuals and families with spherocytosis. This topic is reviewed in detail in chapter 5 of this volume. In summary, it is clear that HS and HE may be the result of inadequate production of, or mutations and deletions in spectrin, ankyrin, and protein 4.1. It is possible that deficiencies or mutations of less abundant skeletal proteins, e.g., 4.2 and 4.9, may also cause abnormal red cell shape and life span (Rybicki et al., 1988). We can speculate that mutational events in transmembrane proteins (band 3, the anion channel, and the glycophorins) may be responsible for changes in red cell volume, shape, and survival, as observed in stomatocytosis and the rarer disorders of cell volume known as dessicocytosis and xerocytosis (Glader et af., 1974; Snyder et al., 1978). As in many other areas of research medicine, characterization of mutant red cell membrane proteins causing hemolytic states has enhanced our understanding of the structure and function of the normal erythrocyte cytoskeleton.
References Agr6, P., Casella, J. F., Zinkham, W. H., McMillan, C., and Bennett, V. (1985). Partial deficiency of erythrocyte spectrin in hereditary spherocytosis. Nature (London) 314, 380-383. Anderson, R . A,, and Lovrien, R. E. (1984). Glycophorin is linked by band 4.1 protein to the human erythrocyte membrane skeleton. Nature (London) 307, 655. Becker, P. S.. Morrow, J. S., and Lux, S. E. (1987). Abnormal oxidant sensitivity and P-chain structure of spectrin in hereditary spherocytosis associated with defective spectrin-protein 4.1 binding. J. Clin. Invest. 80, 557-565. Bennett, V. (1985). The membrane skeleton of human erythrocytes and its implication for more complex cells. Annu. Rev. Biochem. 54, 273. Conboy, J., Marchesi, S., Kim, R., Agr6, P., Kan, Y. W., and Mohandas, N. (1990). Molecular analysis of insertion/deletion mutations in protein 4.1 in elliptocytosis. Part 11. Determination of molecular genetic origins of rearrangements. J. Clin. Invest. 86, 524-530. Correas, I., Leto, T. L., Speicher, D. W., and Marchesi, V. T. (1986). Identification of the functional site of erythrocyte protein 4. I involved in spectrin-actin associations. J. Biol. Chem. 7, 33103315. Dhermy, D., Lecomte, M. C., Garbarz, M., Bournier, O., Galand, C., Gautero, C., Feo, C., Alloisio, N., Delaunay, J., and Boivin, P. (1982). Spectrin P-chain variant associated with hereditary elliptocytosis. J. Clin. Invest. 70, 707-7 15. Eber, S. W., Moms, S. A , , Schroter, W., and Gratzer, W. B. (1988). Interactions of spectrin in hereditary elliptocytes containing truncated spectrin (3 chains. J. Clin. Invest. 81, 523-530. Elder, J. H., Pickett, R. A., Hampton, J., and Lerner, R. (1977). Radioiodination of proteins in single polyacrylamide gel slices. J. Biol. Chem. 252, 6510-6515. Fe6, C., Fischer, S., Piau, J. P., Grange, M. I., and Tchernia, G. (1980). Premitre observation de I’absence d’une prot6in de la membrane 6rythrocytaire (bande 4 , ) dans un cas d ’ a n h i e elliptocytaire familiale. Nouv. Rev. F r . 22, 315-325. Floyd, P., Gallagher, P. G.. Marchesi, S. L., and Forget, B. G. (1990). Heterogeneity in the
172
Sally L. Marchesi molecular basis of aI/74 hereditary elliptocytosis and hereditary pyropoikilocytosis (HPP). Blood 76, 31a (abstr.).
Gallagher, P. G., Garbarz, M., Tse, W., Coetzer, T., Lawlor, J., Zarkowsky, H. S., Lecomte, M. C., Dhermy, D., Baruchel, A , , Marchesi, S . L., Garbarz, M., Palek, J., and Forget, B. G. (1990a). A common mutation in 1x1146-50a kD in hereditary elliptocytosis (HE) and hereditary pyropoikilocytosis (HPP). Blood 76, (ahstr.). Gallagher, P. G.. Garbarz, M., Tse, Picat, C., Lecomte, M. C . , Dhermy, D., and Forget, B. G. (1990b). Exon skipping due to a splice site mutation causes hereditary elliptocytosis (HE) associated with shortened p chain of spectrin Rouen. Clin. Res. 38, 266A (abstr.). Gallagher, P. G., Tse, W. T., Pothier, B., Costa, F., Boivin, P., Delauney, I., and Forget, B. G. (1990~).Molecular basis of shortened p spectrin chains in hereditary elliptocytosis (HE): A splice site mutation and exon skipping in spectrin LePuy and a frame shift due to a two base insertion in spectrin Nice. Blood 76, 8a (abstr.). Garbarz, M., Lecomte, M. C., Feo, C., Devaux, I., Picat, C., Lefebvre, C., Galibert, F., Gautero, Galand, C., Forget, B. G., Boivin, P., and Dhermy, D. (1990). The DNA’s H., Bournier, 0.. from HE and HPP patients in a Caucasian French family with the spectrin a1174 kd variant contain a CGT ---* CAT codon change (Arg + His) at position 22 of the a1180 kd domain. Blood 75, 1691. Glader, B. E., Fortier, N., Albala, M. M., and Nathan, D. G. (1974). Congenital hemolytic anemia associated with dehydrated erythrocytes and increased potassium loss. N. EngI. 1.Med. 291, 491 -496. Goodman, S. R., Shiffer, K A,, Casoria, L. A,, and Eyster, M. E. (1982). Identification of the molecular defect in the erythrocyte membrane skeleton of some kindreds with hereditary spherocytosis. Blood 60,772-784. Iarocci. T. A., Wagner, G. M., Mohandas, N., Lane, P. A,, and Mentzer, W. C. (1988). Hereditary poikilocytic anemia associated with the co-inheritance of two alpha spectrin abnormalities. Blood 71, 1390-1396. Knowles, W. J., Morrow, J. S., Speicher, D. W., Zarkowsky. H. S . , Mohandas, N., Mentzer, W. C., Shohet, S. B., and Marchesi, V. T. (1983). Molecular and functional changes in spectrin from patients with hereditary pyropoikilocytosis. J. Clin. Invest. 71, 1867- 1877. Knowles, W. I . , Bologna, M. L., Chasis, J. A., Marchesi, S. L., and Marchesi, V. T. (1984). Common structural polymorphisms in human erythrocyte spectrin. J. Clin. Invest. 73, 973979. Lane, P. A,, Shew, R. L., Iarocci, T. A., Mohandas, N., Hays, T., and Mentzer, W. C. (1987). Unique alpha-spectrin mutant in a kindred with common hereditary elliptocytosis. J. Clin. Invest. 79, 989-996. Lecomte, M. C., Garbarz, M., Grandchamp, B., Fe6, C., Gautero, H., Devaux, I., Bournier, O., Galand, C., d’Auriol, L., Galibert, F., Sahr, K. E., Forget, B. G., Boivin, P., and Dhermy, D. (1989a). A mutation of the a1 spectrin domain in a White kindred with HE and HPP phenotypes. Blood 74, 1126-1 133. Lecomte, M. C., Lahary, A,, Vannier, J. P., Gautero, H., Galand, C., Boumier, O., Monconduit, A new M., Thron, P. H., Boivin, P., and Dhermy, D. (1989b). Spectrin Rouen (p220/218): shortened spectrin (Sp) p chain variant in a kindred of hereditary elliptocytosis (HE). Br. J. Haematol. Leto, T., and Marchesi, V. T. (1894). A structural model of human erythrocyte 4.1. J. Biol. Chem. 259, 4603-4608. Liu, S . C., Derick, L. H., and Palek, J. (1987). Visualization of the hexagonal lattice in the erythrocyte membrane skeleton. J. Cell Biol. 104, 527-536. Lux, S. E., Tse, W. T., Menninger, J. C., John, K. M., Harris, P., Shalev, O., Chilcote, R. R., Marchesi, S . L., Watkins, P. C., Bennett, V., Mclntosh, S., Collins, F. S . , Francke, U., Ward,
9. Mutant Cytoskeletal Proteins
173
D. C., and Forget, B. G. (1990). Hereditary spherocytosis associated with deletion of human erythrocyte ankyrin gene on chromosome 8. Nature (London) 345, 746. Marchesi, S. L., (1989). Red blood cell membranes. In “Red Blood Cell Membranes” (P. AgrC and J. Parker, eds.), p. 77. Dekker, New York. Marchesi, S. L., Knowles, W. J., Morrow, J. S., Bologna, M., and Marchesi, V. T. (1986). Abnormal spectrin in hereditary elliptocytosis. Blood 67, 141-151. Marchesi, S. L., Letsinger, J. T., Speicher, D. W., Marchesi, V. T., AgrC, P., Hyun, B., and Gulati, G . (1987). Mutant forms of spectrin a-subunits in hereditary elliptocytosis. J. Clin. Invest. 80, 191-19s. Marchesi, S. L., AgrC, P. A,, Speicher, D. W., Tse, W. T., and Forget, B. G. (1989). Mutant spectrin a11 domain in recessively inherited spherocytosis. Blood 74 (7). 182a. Marchesi, S. L., Conboy, J.. Angre, P., Letsinger, J. T., Marchesi, V. T., Speicher, D. W., and Mohandas, N. (1990). Molecular analysis of insertion/deletion mutations in protein 4.1 in elliptocytosis. Part I . Biochemical identification of rearrangements in the spectrinlactin binding domain and functional characterization. J . Clin. Invest. 85, 516-523. McGuire, M., Smith, B. L., and Agri, P. (1988). Distinct variants of erythrocyte protein 4.1 inherited in linkage with elliptocytosis and Rh type in three white families. Blood 72, 287-293. MorlC, L., MorlC, F., Roux, A. F., Godet, J . , Forget, B. G., Denoroy, L., Garbarz, M., Dhermy, ~), variant, D., Kastally, R., and Delaunay, I. (1989). Spectrin Tunis ( S P ~ ~ ’an~ elliptocytogenic is due to the CGG + TGG codon change (Arg + Trp) at position 35 of the a1 domain. Blood 74, 828-832. MorlC, L., Roux, A. F., Allosio, N., Pothier, B., Starck, J., Denoroy, J., Morl6, F., Rudigoz, R. C . , Forget, B. G., Delaunay, J., and Godet, J. (1990). Two elliptocytogenic a1174 variants of the spectrin a1 domain. Spectrin Culoz and spectrin Lyon, mutated at position 40 (GGT + G l T ; GLY + VAL) and at position 43 ( C l T + TIT; LEU + PHE), respectively. J Clin. Invest. 86, 548-553. O’Farrell, P. H. (1975). High resolution two-dimensional electrophoresis of proteins. J . Biol. Chem. 250, 4007-402 1. Ohanian, V., Evans, I. P., and Gratzer, W. B. (1985). A case of elliptocytosis associated with a truncated spectrin chain. Br. J . Haematol. 61, 31-39. Palek, J. (1987). Hereditary elliptocytosis, spherocytosis, and related disorders: Consequences of a deficiency or a mutation of membrane skeletal proteins. Blood Rev. 1, 147. Pothier, B., MorlC, N., Allosio, N., Ducluzeau, M. T., Caldam, C., Feo, C., Garbarz, M., Chaveroche, I . , Dhermy, D., Lecomte, M. C., Bolvin, P., and Delaunay, J. (1987). Spectrin A shortened p chain variant associated with an increase of the a1174 fragment in Nice (p220’216): a case of elliptocytosis. Blood 69, 1759-1765. Roux, A. F., MorlC, F., Guetarni, D., Colonna, P., Sahr, K., Forget, B. G., Delauney, J., and Godet, J. (1989). Molecular basis of aIP5 hereditary elliptocytosis in North Africa: Insertion of a TTG triplet between codons 147 and 149 in the a-spectrin gene from five unrelated families. Blood 73, 2196-2201. Rybicki, A. C., Heath, R., Wolf, J. L., Lubin, B., and Schwartz, R. S. (1988). Deficiency of protein 4.2 in erythrocytes from a patient with a Coombs negative hemolytic anemia. J. Clin. Invest. 81, 893-901. Snyder, L. M., Lutz, H. U., Sauberman, N., Jacobs, J., and Fortier, N. L. (1978). Fragmentation and myelin formation in hereditary xerocytosis and other hemolytic anemias. Blood 52, 750-761. Speicher, D. W., Morrow, J. S., Knowles, W. J., Hsu,C. J., and Marchesi, V. T. (1980). Identification of proteolytically resistant domains of human erythrocyte spectrin. Proc. Natl. Acad. Sci. U.S.A. 77, 5673-5677. Speicher, D. W., Morrow, 1. S., Knowles, W. J., and Marchesi, V. T. (1982). A structural model of human erythrocyte spectrin. J. Biol. Chem. 257, 9093-9101.
174
Sally L. Marchesi
Tse, W. T., Lecomte, M. C., Costa, F. F., Garbarz, M . , Fed, C., Boivin, P., Dhermy, D . , and Forget, B. G . (1990). Point mutation in the P-spectrin gene associated with aU74 hereditary elliptocytosis (HE).J . Clin Invest. 86, 909-916. Wolfe, L. C . , John, K. M . , Falcone, J. C . , Byrne, A. M . , and Lux, S. E. (1982). A genetic defect in the binding of protein 4.1 to spectrin in a kindred with hereditary spherocytosis. N. Engi. J . Med. 307, 1367-1374.
CHAPTER 10
Dynamics of Intestinal Epithelial Tight Junctions James L. Madara Department of Pathology Brigham and Women’s Hospital and Harvard Medical School and the Harvard Digestive Disease Center Boston, Massachusetts 021 15
I. Introduction 11. Tight Junctions (Zonula Occludens) as Potentially Regulated Barriers 111. Intestinal Zonula Occludens as a Regulated Transport Pathway
IV. Intestinal Zonula Occludens Function in Model Disease States V. Conclusions References
1. INTRODUCTION
The intestinal epithelium normally absorbs the vast quantity of water, ions, and nutrients presented daily but, in addition, aids in preventing the free mixing of lumenal contents with underlying interstitial and vascular fluids. A crucial element in this latter “barrier function” of the intestinal epithelium is the intercellular tight junction-the focus of this chapter. Tight junctions in this epithelium are not only crucial to barrier function, but may be regulated, and such regulation ironically plays a substantive role in the uptake of nutrients. Given the plasticity of the tight junction in health, it is not surprising that studies of models suggest that this structure may be functionally altered in disease states, even when the epithelium remains confluent.
Current Tupics in Membranes, Volume 38 Copyright 0 1991 by Academic Press. Inc. All rights of reproduction in any form reserved.
175
176
James L. Madara
II. TIGHT JUNCTIONS (ZONULA OCCLUDENS) AS POTENTIALLY REGULATED BARRIERS
As shown in the Fig. 1, there exist two potential pathways for passive permeation across intestinal epithelia: the transcellular pathways or the paracellular. The paracellular pathway is composed of the tight junction, more properly termed the zonula occludens (ZO), and the underlying intercellular space. As recently reviewed by Gumbiner (1987), and first described by Farquhar and Palade (1963), the ZO consists of a narrow belt which wraps epithelial cells at the apical pole (Fig. 1). Within this belt the lateral membranes of adjacent cells are focally fused and these fusion points circumferentially belt the apex of the cell in an anas-
FIG. 1 Schematic illustrations depicting tight junction (ZO) location and structure in intestinal epithelial cells. On the left is shown the brush border region of a transected cell. Fragments of the lateral membranes of flanking, neighbor cells are shown. The membrane face of the neighboring cell in the foreground is shown to highlight the freeze-fracture appearance of the 20.The inset displays a speculative model of the molecular substructure of the ZO (see text). Note that not only might the cytoskeleton modulate the ZO indirectly by tensile forces within the perijunctional actomyosin ring, which circumferentially wraps the cell and inserts on the lateral membrane just below the ZO (left), but direct cytoskeletal-ZO interactions also appear to occur (inset). Reproduced from Madara (l989), with permission.
10. Dynamics of Epithelial Tight Junctions
177
tomosing fashion. Such fusions or “kisses” appear on replicas of freeze-fracw e d membranes as an interconnected series of strands and grooves of unknown composition (Kachar and Reese, 1982; Griepp et al., 1983; Stevenson and Goodenough, 1984). While a barrier, the ZO displays substantial permeability to small molecules [although under baseline conditions and in the absence of convective water flow it is a relatively effective seal to molecules 11.5A and greater in Stokes radius (Madara and Dharmsathaphorn, 1985)l. Epithelia vary substantially in resistance to passive ion permeation, and variation in ZO resistance is largely the basis for this variability (for review, see Powell, 1981). For example, while biomembrane resistance is high (1,000-10,000 R cm2) (Powell, 1981), the resistance of mammalian small intestinal epithelium is approximately 50-90 R cm2 (Madara et al., 1986; Schultz and Zalusky, 1964; Field et al., 1971). Such observations lead one to suspect that the paracellular pathway is the major site for passive transepithelial permeation in this epithelium and this has been verified by Frizzell and Shultz (1972), who showed that at least 85% of passive ion flow across mammalian small intestine takes this paracellular route. In contrast to the ZO, which restricts macromolecular flow, free diffusion occurs in the intercellular space below the ZO of macromolecules (Phillips et al., 1987). Thus, under most conditions, the major barrier in the paracellular pathway is the ZO. In summary, the ZO, while displaying substantial permeability to ions, is the ratelimiting barrier in the major permeation pathway (paracellular) of the intestine. ZO structure often varies in a cell-type-specific fashion (Madara and Trier, 1982; Pricam et al., 1974; Schneeberger, 1980; Ernst et al., 1980), and such variability may contribute to epithelial transport events. For example, in the small intestine, ZOs of crypt cells are structurally irregular and, due to the narrow apex of the cells, ZO density in the crypt is very high (80 m/cm2 of crypt luminal surface area) (Marcial et al., 1984). In contrast, villus absorptive cells in fasted animals have ZOs with more structural subunits and with junctional density only a quarter of that of the crypt (Marcial et al., 1984). Since it appears that intestinal secretion originates from the crypt and consists of the net electrogenic transfer of C1- across cells followed by subsequent passive transjunctional movement of Na+ (Field, 1981), such ZO structural modifications in the crypt may facilitate secretion. How ZO structure relates to ZO barrier function is not yet certain. A highly speculative working model, shown in Fig. 1 , takes into consideration available data, such as ZO structure-function correlations (Madara and Dharmsathaphorn, 1985; Claude and Goodenough, 1973; Claude, 1978), ZO ion selectivity sequences (Powell, 1981; Diamond and Wright, 1969; Cereijido et al., 1978), ZO sieving characteristics (Madara and Dharmsathaphorn, 1985), and ZO charge selectivity (Smyth and Wright, 1966; Madara, 1983). In this model, ZO subunits (strands by freeze-fracture/kisses by thin section) are viewed as impermeable structures interrupted by discontinuities or channels. As with channels of bio-
178
James L. Madara
membranes, these channels may open and close (Claude, 1978) and, furthermore, biophysical data suggest the interior of the channel is highly hydrated and contains fixed negative charges presumably due to carboxyl groups. This model predicts that there may exist numerous ways to modify ZO barrier function: the number of ZO subunits could change, the probability of open-state channels could alter, or the radius, hydrophilicity, or surface charge of the channel could change. Recent data tend to support the above prediction that there are multiple ways to modify ZO function. Examples include alteration in ZO subunit number (Madara et al., 1986; Madara, 1983; Bentzel et al., 1988; Palant et al., 1983), and alteration in ZO charge selectivity (Krasney et al., 1983). Recent work using two strains of MDCK cells which have either high (-300 R cm2) or low (-100 R cm2) resistance suggests that the basis of the difference in resistance observed is not due to strain-related differences in 20 subunit number (Stevenson et ul., 1991). Rather, here it appears that variation in the density, probability of being in the open state, or other physical characteristic of the ZO pores determines the strain-related variation in resistance (Stevenson et al., 1991). A variety of observations suggest that ZOs can be regulated by intracellulur events. First, intracellular mediators can alter ZO function and/or structure in some epithelia. For example, Duffey et al. (1981) showed that the ZOs of amphibian gallbladder epithelium display enhanced resistance to passive ion flow as intracellular cAMP is elevated and, paralleling this response, ZOs gained strand subunits. cAMP also alters ZO function in goldfish epithelia (Krasney et al., 1983; Bakker and Groot, 1984). Exposure of amphibian gallbladder epithelium to Ca2+ ionophore enhances ZO resistance, and this functional response is again accompanied by alterations in ZO charge selectivity and structure (Palant et al., 1983). Lastly, in a renal cell line, activation of protein kinase C, as mimicked pharmacologically with phorbol esters, lessons ZO resistance (Ojakian, 1981; Mullin and O’Brien, 1986). How do signals, such as the mediators listed above, lead to ZO alterations? Only indirect data are available to address this issue, but they suggest the cytoskeleton may be involved in transducing such messages. Functional links between the cytoskeleton and the ZO were first described in cultured renal epithelium (Meza et al., 1980) and in gallbladder epithelium (Bentzel et a l . , 1988) in studies which pharmacologically manipulated the cytoskeleton and assessed the impact of this manipulation on ZO function. This putative ZOcytoskeletal link has also been explored in the intestine. One characteristic feature of the absorptive cell brush border (for review, see Mooseker, 1985) which has been noted is a circumferential ring of actin and myosin which wraps each absorptive cell brush border (Rodewald et al., 1976; Keller and Mooseker, 1982; Burgess, 1982). This preijunctional actomyosin ring inserts on the lateral membrane and can be elicited to contract with ATP and divalent cations
10. Dynamics of Epithelial Tight Junctions
179
(Rodewald et al., 1976; Burgess, 1982). During this contractile response, the brush border rounds (Rodewald et al., 1976; Burgess, 1982) and myosin is phosphorylated (Keller and Mooseker, 1982). Since mechanically applied lateral tension can alter ZO structure (Pitelka and Taggart, 1983), it was speculated (Mooseker, 1985) that ring contraction might change ZO structure and permeability. Using the actin-binding agent cytochalasin D, we demonstrated that absorptive cell ZOs became perturbed in structure, permeability, and charge selectivity (Schultz and Zalusky, 1964) as the perijunctional ring segmented and condensed and the brush borders rounded-features suggesting that a contraction-like effect had occurred segmentally in the ring. Additionally, these effects of cytochalasin D were energy dependent (Madara et al., 1987). Similar data were obtained from T,, monolayers-a model intestinal epithelium (Madara et al., 1988). Subtle direct associations also appear to exist between the cytoskeleton and the ZO (Madara, 1987; Drenckhahn and Dermietzal, 1988). Detergentextracted preparations of intestinal absorptive cells display plaque-like condensations of electron-dense material immediately adjacent to the cytoplasmic face of the ZO (Madara, 1987). This material often specifically localizes at the sites of kisdstrands within the ZO. It is possible that this material in part represents ZO-1 (Stevenson and Goodenough, 1984; Anderson et al., 1991)-a ZO-specific peripheral membrane phosphoprotein-or cingulin (Citi et al., 1988), which are candidate molecules for linking the cytoskeleton with the ZO. Actin microfilaments also intimately associate with the ZO-associated peripheral plaque-like condensations (Madara, 1987; Drenckhahn and Dermietzal, 1988), raising the possibility that the ZO could be more directly affected by the cytoskeleton.
111. INTESTINAL ZONULA OCCLUDENS AS A REGULATED TRANSPORT PATHWAY
Until recently, the predominant view of glucose absorption by the intestine could be summarized as follows (Alpers, 1987; Gray, 1978): glucose is cotransported across the apical membrane with Na , and, via the Na -K ATPase pump and by basolateral glucose-facilitated transport, these solutes are subsequently deposited into the paracellular space. Absorption of water across the ZO is driven by the deposition of these osmotically active solutes in the paracellular space. However, recent data suggest that nutrient exposure alters ZO permeability and, in so doing, provides an additional (paracellular) pathway for nutrient absorption (Atisook et al., 1990; Pappenheimer, 1987; Madara and Pappenheimer, 1987; Pappenheimer and Reiss, 1987). Exposure to mucosal glucose elicits a change in ZO permeability as assessed by direct (Atisook et af., 1990) or alternating (Pappenheimer, 1987; Madara and Pappenheimer, 1987) current techniques or by flux (Atisook et al., 1990) studies. These changes in ZO +
+
+
180
James L. Madara
permeability occur in parallel to changes in ZO strand architecture and coincide with condensation of the preijunctional ring of actin and myosin (Madara and Pappenheimer, 1987). Lastly, these effects are energy dependent (Atisook et al., 1990) and appear to be triggered by turnover of the Na+ -glucose 'cotransporter present on the apical membrane of absorptive cells (Atisook et al., 1990). As a result of these changes, the aforementioned absorptive solvent (water) flow across the ZO sweeps lumenal solutes, including nutrients, into the paracellular space by solvent drag (Pappenheimer and Reiss, 1987). Such paracellular nutrient movement, by solvent drag across ZO, as a result of regulated ZO permeability, may account for a large fraction of net nutrient absorption (Atisook et al., 1990; Pappenheimer and Reiss, 1987). Indeed, since the cotransporters of the apical membrane are saturated at relatively low nutrient concentrations but, after a meal, nutrient concentrations in the proximal intestine often exceed 200 mM (Borgstrom et al., 1957; Fordtran and Ingelfinger, 1968), the ZO may be one, if not the, major pathway of nutrient uptake. In agreement with this hypothesis, glucose uptake by the intestine increases with escalating luminal nutrient concentration past the point at which saturation of the transcellular pathway occurs (MacLeod et al., 1930; Cummins, 1952). Lastly, another type of physiological regulation of the ZO occurs at the villus tip in the small intestine. Here, sloughing epithelial cells are extruded into the intestinal lumen following their migration from the crypt to this site. However, such extrusion zones are not sites at which macromolecular leaks occur (Madara, 1990). Rather, as extruding cells lift from the epithelium, shouldering cells migrate beneath them and new ZO strands are formed both between the extruding cell and its neighbors and between the newly abutted cells (Madara, 1990). As the extruding cell lifts further, the ZO subunits sweep over the lateral aspect of the membrane and the extruding cell finally lifts from an epithelium which maintains its ability to resist penetration by macromolecules (Madara, 1990).
N. INTESTINAL ZONULA OCCLUDENS FUNCTION IN MODEL DISEASE STATES It is intuitively obvious that, when epithelial cells are grossly separated, as would occur with ulcers or erosions, the permeability barrier of the epithelium is disrupted. However, there also exist diseases such as celiac sprue (Trier, 1978) in which ZOs leak large molecules even though the epithelium remains confluent. Using T,, monolayers of model intestinal epithelia (Dharrnsathaphorn et al., 1984), recent studies have been carried out in hopes of modeling disease-related phenomena for mechanistic studies. T,, cells grow as confluent monolayers, display high baseline resistance, have ZOs with subunit structure-function correlates not dissimilar to those seen in native intestinal epithelium (Madara and
10. Dynamics of Epithelial Tight Junctions
181
Dharmsathaphorn, 1985), and harbor actin-rich perijunctional rings (Madara et al., 1988) which, like those of natural intestinal epithelial cells, segment and condense on exposure to cytochalasin D-an event accompanied by enhanced ZO permeability (Madara et al., 1988). Intestinal inflammation of the intestine is often characterized by movement of polymorphonuclear leukocytes (PMN) across the epithelium to form so-called crypt abscesses (Kumar et al., 1982). This process may be modeled in T,, or other (Cramer et al., 1980; Evans et al., 1983) monolayers using isolated PMN transepithelial chemotactic gradients. Under such conditions, PMN move across T,, monolayers by crossing the ZO (Nash et al., 1987). With large numbers of transmigrating PMN, ZO permeability to ions and inert solutes is reversibly increased (Nash et al., 1987). Since PMN secretory products do not appear necessary for ZO impalement by the PMN (Nash et al., 1991) but PMN-epithelial cell adhesion may be (Nash et al., 1991), we speculate that the opening of the ZO during PMN transmigration is produced by external (by PMN pseudopod) mechanical force just as internal mechanical force may underlie physiological regulation of the ZO. Additional, in vitro models of intestinal disease promise to yield insights into cytoskeletal-ZO relationships. Toxin A, a protein exotoxin of Clostridium difficile, causes a severe enterocolitis which, in part, may be due to its effects on inflammatory cells (Triadafilopoulos et al., 1987). However, toxin A disrupts the barrier function of T,, monolayers such that transepithelial resistance is nearly abolished within 6-8 hr (Hecht et al., 1988). Surprisingly, the monolayers remained confluent, and cytotoxicity does not parallel this initial dramatic increase in ZO ion permeability (Hecht et al., 1988). Furthermore, this toxin-elicited increase in permeability was restricted to molecules less than 58, in Stokes radius. A prominent effect of toxin A in this model is diminution of F-actin staining in the perijunctional ring. Lastly, the inflammatory cytokine, interferon-y also affects ZO permeability in this model epithelium and elicits changes in the perijunctional actin myosin ring (Madara and Stafford, 1989). Given the apparent plasticity of ZOs and the putative relationships between ZOs and the cytoskeleton, it will not be surprising if various other disease-related challenges substantially alter intestinal epithelial barrier function even if epithelial continuity is maintained. Further analyses of such systems should enhance understanding of the factors involved in eliciting abnormal epithelial permeability in disease states.
V. CONCLUSIONS As modeled by intestinal epithelia, ZOs behave quite unlike the static gaskets they were assumed to be in the not-too-distant past. It appears this cell surface
182
James L. Madara
structure intimately associates with cytoskeletal components and that such ZOcytoskeletal associations may be functionally relevant. Not only can ZO structure and function be altered by pharmacological manipulation of the cytoskeleton, but intracellular messengers, such as CAMP, can also affect the ZO-putatively by generating cytoskeletal alterations. More than that, physiological signals, such as the activation of Na-coupled nutrient transporters on the apical membrane of intestinal villus absorptive cells, can substantially alter ZO structure and function. Indeed, it appears the intestine would be unable to harvest the vast quantity of nutrients presented daily in the absence of this regulated ZO pathway. .Evidence suggests that such ZO regulation may ultimately be tied to tensile force emanating from an intracellular sphincter-like actomyosin ring. Lastly, the plasticity of the ZO is also manifest in several model disease states. These latter findings not only serve as a basis for understanding how epithelial permeability can change so greatly with disease, even in the presence of retained epithelial confluency, but also provide models to dissect ZO-cytoskeletal relationships further. It is becoming clear that modulation of intestinal epithelial ZOs is important in health and disease. It is now time to attempt to understand with more precision the nature of intracellular pathways which modulate this cell surface structure which is so crucial a part of intestinal epithelial barrier function.
Acknowledgments I particularly want to thank and acknowledge postdoctral associates and collaborators who participated in many of the experiments described here: Drs. Manuel Marcial, David Barenberg, Gail Hecht, Shirin Nash, Ronda Moore, Kanit Atisook, Jerry Trier, and John Pappenheimer. I also acknowledge and thank the Journal of Clinical Invesrigarion for permission to use some phrasing which appeared in a perspectives article written by the author (Madara 1989). Supported by NIH grants DK35932 and DK33506.
References Alpers, D. H. (1987). Digestion and absorption of carbohydrates and proteins. In “Physiology of the Gastrointestinal Tract” (L. R. Johnson, ed.), 2nd ed. pp. 1469-1488. Raven Press, New York. Anderson, J. M., Stevenson, B. R., Jesaitis, L. A,, Goodenough, D. A,, and Mooseker, M. S. (1988). Characterization of ZO-1, a protein component of the tight junction from mouse liver and MDCK cells. J . Cell Biol. 106, 1141-1 149. Atisook, K., Carlson, S., and Madara, J. L. (1990). Effects of phlorizin and sodium on glucoseelicited alterations of cell junctions in intestinal epithelia. Am. J . Physiol. 258, C77-C85. Bakker, R., and Groot, J. A. (1984). CAMP-mediated effects of ouabain and theophylline on paracellular ion selectivity. Am J . Physiol. 246, G213-G217. Bentzel, C. J., Hainan, B., Ho, S . , Hui, S. W., Edelman, A,, Anagnostopoulos, T., and Benedetti, E. R. (1988). Cytoplasmic regulation of tight-junction permeability: Effects of plant cytokinins. Am. J . Physiol. 239, C75-C89. Borgstrom, B., Dahlquist, A , , Lundh, G . , and Sjovall, J. S . (1957). Studies of intestinal digestion and absorption in the human. J . Clin. Invest. 36, 1521-1536. Burgess, D. R. (1982). Reactivation of intestinal epithelial brush border motility: ATP-dependent contradiction via a terminal web contractile ring. J . Cell Biol. 95, 853-863.
10. Dynamics of Epithelial Tight Junctions
183
Cereijido, M., Robbins, E. S., Dolan, W. I., Rotunno, C. A,, and Sabatini, D. D. (1978). Polarized monolayers formed by epithelial cells on a permeable and translucent support. J. Cell Biol. 77, 853-906. Citi, S., Sabanay, H., Jakes, R., Geiger, B., and Kendrick-Jones, J. (1988). Cingulin, a peripheral component of tight junctions. Nature (London) 333, 272-276. Claude, P. (1978). Morphologic factors influencing transepithelial permeability: A model for the resistance of the zonula occludens. J. Membr. Biol. 39, 219-232. Claude, P., and Goodenough, D. A. (1973). Fracture faces of zonulae occludents from “tight” and “leaky” epithelia. J. Cell Biol. 58, 390-400. Cramer, E. B., Milks, L. C., and Ojakain, G. K. (1980). Transepithelial migration of human neutrophils: An in vitro model system. Proc. Nurl. Acud. Sci. U.S.A. 77, 4069-4073. Cummins, A. J. (1952). Absorption of glucose and methionine from the human intestine: The influence of glucose concentration in the blood and in the intestinal lumen. J. Clin. Invest. 31, 928-937. Dharmsathaphorn, K., Mandel, K. G., McRoberts, J. A,, Tisdale, L. D., and Masui, H. (1984). A human colonic tumor cell line that maintains vectorial electrolyte transport. Am. J . Physiol. 246, G204-G208. Diamond, I. M., and Wright, E. M. (1969). Biological membranes: The physical basis of ion and non-electrolyte permeability. Annu. Rev. Physiol. 31, 58 1-646. Drenckhahn, D., and Dermietzal, R. (1988). Organization of the actin filament cytoskeleton in the intestinal brush border: A quantitative and qualitative immunoelectron microscope study. J. Cell B i d . 107, 1037-1048. Duffey, M. E., Hainan, B., Ho, S., and Bentzel, C. J. (1981). Regulation of epithelial tight junction permeability by cyclic AMP. Nature (London) 294, 451-453. Emst, S. A,, Dodson, W. C., and Kamaky, K. J. (1980). Structural diversity of occluding junctions in the low-resistance chloride secreting opercular epithelium of seawater-adapted Killifish (Fundulus heteroclitus). J. Cell Biol. 87, 488-497. Evans, C. W., Taylor, J. E., Walker, J. D., and Simmons, N. L. (1983). Transepithelial chemotaxis of rat peritoneal exudate cells. Br. J. Exp. futhol. 64, 644-654. Farquhar, M. G., and Palade, G. E. (1963). Junctional complexes in various epithelia. J. Cell Biol. 17, 375-412. Field, M. (1981). Secretion by the small intestine. I n “Physiology of the Gastrointestinal Tract” (L. R. Johnson, ed.), pp. 963-982. Raven Press, New York. Field, M., Fromm, D., and McCall, I. (1971). Ion transport in rabbit ileal mucosa. 1. Na and CI fluxes and short circuit current. Am. J. Physiol. 220, 1388-1396. Fordtran, I. S., and Ingelfinger, F. J. (1968). Absorption of water, electrolytes and sugars from the human gut. In “Handbook of Physiology” (C. F. Cook, ed.), Sect. 6, Vol. 3, pp. 1457-1490. Williams & Wilkins, Washington, D.C. Frizzell, R. A., and Schultz, S. G. (1972). Ionic conductance of extracellular shunt pathway in rabbit ileum. J. Gen. Physiol. 59, 318-346. Gray, G. M. (1978). Mechanisms of digestion and absorption of food. In “Gastrointestinal Disease” (M. Sleisenger and J. Fordtran, ed.), pp. 241 -250. Saunders, Philadelphia, Pennsylvania. Griepp, E. B., Dolan, W. J., Robbins, E. S. and Sabatini, D. D. (1983). Participation of plasma membrane proteins in the formation of tight junctions by cultured epithelial cells. J. Cell Biol. 96, 693-702. Gumbiner, B. (1987). The structure, biochemistry, and assembly of epithelial tight junctions. Am. J. Physiol. 253, C749-C758. Hecht, G., Pothoulakis, C., LaMont, J. T., and Madara, J. L. (1988). Clostridium difficile toxin A perturbs cytoskeletal structure and tight junction permeability of cultural human intestinal epithelial monolayers. J. Clin. Invest. 82, 1516-1524.
184
James L. Madara
Kachar, B., and Reese, T. S. (1982). Evidence for the lipidic nature of tight junction strands. Nature (London) 296, 464-466. Keller, T. C. S., and Mooseker, M. S. (1982). Caf2-calmodulin dependent phosphorylation of myosin, and its role in brush border contraction in vitro. J. Cell Biol. 95, 943-959. Krasney, E., Madara, I. L., DiBona, D., and Frizzell, R. (1983). Cyclic AMP regulates tight junction permselectivity in flounder intestine. Fed. Proc., Fed. Am. SOC. Exp. Biol. 42, 1100 (abstr.). Kumar, N. B., Nostrant, T. T. and Appelman, H. D. (1982). The histopathologic spectrum of acute self-limited colitis (acute infectious-type colitis). Am. J. Surg. Pathol. 6, 523-529. MacLeod, J. J. R., Magee, H. E., and Purves, C. B. (1930). Selective absorption of carbohydrates. J. Physiol. (London) 70, 404-416. Madara, J. L. (1983). Increases in guinea pig small intestinal transepithelial resistance induced by osmotic loads are accompanied by rapid alterations in absorptive-cell tight junction structure. J. Cell Biol. 97, 125-136. Madara, J. L. (1987). Intestinal absorptive cell tight junctions are linked to cytoskeleton. Am. J . Physiol. 253, C17l-Cl75. Madara, J. L. (1989). Loosening tight junctions: Lessons from the intestine. J. Clin. Invest. 83, 1089- 1094. Madara, J. L. (1990). Maintenance of the macromolecular barrier at cell extrusion sites in intestinal epithelium: Physiological rearrangement of tight junctions. J. Membr. B i d . 116, 177- 184. Madara, J. L., and Dharmsathaphorn, K. (1985). Occluding junction structure-function relationships in a cultured epithelial monolayer. J. Cell Biol. 101, 2124-2133. Madara, J. L., and Pappenheimer, J. R. (1987). Structural basis for physiological regulation of paracellular pathways in intestinal epithelia. J. Membr. Biol. 100, 149- 164. Madara, J. L., and Stafford, J. (1989). Interferon-y directly affects barrier function of cultured intestinal epithelial monolayers. J. Clin.Invest. 83, 724-727. Madara, J. L., and Trier, J. S. (1982). Structure and permeability of goblet cell tight junctions in rat small intestine. J . Membr. Biol. 66, 145-157. Madara, J. L., Barenberg, D., and Carlson, S. (1986). Effects of cytochalasin D on occluding junctions of intestinal absorptive cells: Further evidence that the cytoskeleton may influence paracellular permeability. J. Cell Biol. 97, 2125-2135. Madara, J. L., Moore, R., and Carlson, S. (1987). Alteration of intestinal tight junction structure and permeability by cytoskeletal contraction. Am J. Physiol. 253, C854-C861. Madara, J. L., Stafford, J., Barenberg, D., and Carlson, S. (1988). Functional coupling of tight junctions and microfilaments in Ts4 monolayers. Am. J. Physiol. 254, G 4 1 6 4 4 2 3 . Marcial, M., Carlson, S. L., and Madara, J. L. (1984). Partitioning of paracellular conductance along the ileal crypt-villus axis: A hypothesis based on structwal analysis with detailed consideration of tight junction structure-function relationships. J. Membr. Biol. 80, 59-70. Meza, I., Obarra, G., Sabanero, M., Martinez-Palomo, A,, and Cereijido, M. (1980). Occluding junctions and cytoskeletal components in a cultured transporting epithelium. J . Cell Biol. 87, 146-154. Mooseker, M. S . (1985). Organization, chemistry, and assembly of the cytoskeletal apparatus of the intestinal brush border. Annu. Rev. Cell Biol. 1, 209-242. Mullin, J. E., and O’Brien, T. G. (1986). Effects of tumor promoters on LUJ-PK, renal epithelial tight junctions and transepithelial fluxes. Am. J . Physiol. 251, C597-C602. Nash, S., Stafford, J., and Madara, J. L. (1987). Effects of polymorphonuclear leukocyte transmigration on the barrier function of cultured intestinal epithelial monolayers. J . Clin. Invest. 80, 1 104- 1 113. Nash, S . , Stafford, J., and Madara, J. L. (1988). The selective and superoxide-independent disrup-
10. Dynamics of Epithelial Tight Junctions
185
tion of intestinal epithelial tight junctions during leukocyte transmigration. Lab. Invest. 59, 531-537. Nicholson, G. L. (1976). Transmembrane control of the receptors on normal and tumor cells. 1. Cytoplasmic influence over cell surface components. Biochim. Biophys. Acta 457, 57-108. Ojakian, G . (1981). Tumor promoter-induced changes in the permeability of epithelial cell tight junctions. Cell (Cambridge, Mass.) 23, 95- 103. Palant, C. E., Duffey, M. E., Mookerjee, B. K., Mo, S., and Bentzel, C. D. (1983). Ca+ + regulation of tight junction permeability and structure in Necturus gallbladder. Am. J . Physiol. 245, C203-C212. Pappenheimer, J. R. (1987). Physiological regulation of transepithelial impedance in the intestinal mucosa of rat and hamsters. J. Membr. Biol. 100, 137-148. Pappenheimer, J. R., and Reiss, K . Z. (1987). Contribution of solvent drag through intercellular junctions to absorption of nutrient by the small intestine of the rat. J. Membr. Biol. 100, 123136. Phillips, T. E., Phillips, T. L., and Neutra, M. R. (1987). Macromolecules can pass through occluding junctions of rat ideal epithelium during cholinergic stimulation. Cell Tissue Res. 247, 547-554. Pitelka, D. R., and Taggart, B. N. (1983). Mechanical tension induces lateral movement of intramembrane components of the tight junction: Studies on mouse mammary cells in culture. J . Cell Biol. 96, 606-612. Powell, D. (1981). Barrier function of epithelia. Am. J . Physiol. 241, G 2 7 5 4 2 8 8 . Pricam, C., Humbert, F., Perredet, A., and Orci, L. (1974). A freeze-etch study of the tight junction of the rat kidney tubules. Lab. Invest. 30, 286-291. Rodewald, R., Newman, S . B., and Karnovsky, M. 1. (1976). Contraction of isolated brush borders from the intestinal epithelium. J. Cell Biol. 70, 541-545. Schneeberger, E. F. (1980). Heterogeneity of tight junction morphology in extrapulmonary and intrapulmonary airways of the rat. Anat. Rec. 198, 193-208. Schultz, S. G., and Zalusky, R. (1964). Ion transport in isolated rabbit ileum. I. Short circuit current and Na fluxes. J. Gen. Physiol. 47, 567-584. Smyth, D. H., and Wright, E. M. (1966). Streaming potentials in the rat small intestine. J. Physiol. (London) 182, 591-602. Stevenson, B. R., and Goodenough, D. A. (1984). Zonulae occludentes in junctional complexenriched fractions from mouse liver. Preliminary morphological and biochemical characterization. J. CellBiol. 98, 1209-1221. Stevenson, B. R., Anderson, J. M., Goodenough, D. A , , and Mooseker, M. S. (1991). Tight junction structure and ZO-I content are identical in two strains of Madin-Darby canine kidney cells which differ in transepithelial resistance. J . Cell Biol. (in press). Triadafilopoulos, C., Pothoulakis, C., O’Brien, M. J., and LaMont, J. T. (1987). Differential effects of Closrridium difJicile toxins A and B on rabbit ileum. Gastroenterology 93, 273-279. Trier, J. S. (1978). Celiac sprue disease. In “Gastrointestinal Disease” (M. Sleisenger and J. Fordtran, eds.), pp. 1029- 105I . Saunders, Philadelphia, Pennsylvania.
This Page Intentionally Left Blank
CHAPTER 11
Regulation of Actin and Myosin I1 Dynamics in Living Cells John Kolega and D. Lansing Taylor Department of Biological Sciences and Center for Fluorescence Research in Biomedical Sciences Carnegie Mellon University Pittsburgh, Pennsylvania 15213
I. Introduction 11. Quiescent Fibroblasts as a Model System 111. The Multimode Approach
IV. V. VI. VII.
Actomyosin Dynamics in Serum-Starved 3T3 Fibroblasts Regulation of Stress Fiber Contraction Stress Fiber Movement and Cell Motility Some Future Prospects References
1. INTRODUCTION
In this volume it is hardly necessary to introduce the intimate association between the plasma membrane and the underlying actin-based cytoskeleton, nor to dwell on the far-reaching consequences of this association on things ranging from cell shape to the distribution and mobility of hormone receptors and ion channels. We instead focus on the regulation of the structure and dynamics of that actin-based cytoskeleton. Actin, myosins, and associated proteins exhibit complex and varied behavior in muscle and non-muscle cells alike: assembling and disassembling, and forming fibers, bundles, and isotropic gel networks. In turn, the resulting actin-based structures display yet another level of behavior as they move about the cell. How these various formations arise, interconvert, and translocate are questions of critical importance to understanding not only the Current Topics in Membranes, Volume 38 Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.
187
188
John Kolega and D. Lansing Taylor
structural organization of the membrane-cytoskeletaltrilayer, but a much broader range of cellular functions, particularly those involving cell shape, intracellular transport, cytoplasmic compartmentalization, and cell motility.
II. QUIESCENT FIBROBLASTS AS A MODEL SYSTEM The system in which we have elected to study the dynamics of actomyosin structures is the quiescent fibroblast. When Swiss 3T3 fibroblasts are deprived of serum they become quiescent, displaying markedly reduced mitotic and locomotive activity (Brooks, 1976; McNeil and Taylor, 1987), as well as low and constant pH (Bright et al., 1987, 1989) and free calcium ion concentration (McNeil et al., 1985). They also undergo characteristic morphological changes, becoming much flatter and distinctly less polarized than unstarved cells (Fig. 1). In attaining their flattened morphology, serum-starved cells also develop an extensive array of actin- and myosin-containing stress fibers (Fig. 2). These fibers contain numerous other cytoskeletal components in addition to actin and myosin, including tropomyosin, a-actinin, filamin, and caldesmon, and each component has a characteristic distribution within the stress fiber network. For example, tropomyosin and myosin have been observed in periodic, punctate patterns along fibers, with a-actinin having a complementary, antiperiodic distribution (Gordon, 1978; Sanger et al., 1983). Meanwhile, actin is present uniformly along the fiber, and talin and vinculin are associated primarily with ends of fibers (Geiger et al., 1980; Burridge and Connell, 1983). This complex organization is severely disrupted when serum or serum-derived growth factors are restored to starved cells. Serum- or growth factor-stimulated cells lose their stress fibers and display new locomotive behavior, such as spreading and ruffling of cell margins. This change in motility is only one of many rapid physiological changes initiated by stimulation, and the literature addressing the intracellular signalling pathways associated with growth factor stimulation is extensive. Stimulation induces reentry into the cell cycle (Brooks, 1976; Rozengurt and Mendoza, 1980), pH (Cassell et al., 1983; Moolenaar, 1986; Ives and Daniel, 1987; Bright et al., 1989) and calcium ion (Moolenaar et al., 1984; Morris et al., 1984; McNeil et al., 1985; Byron and Villereal, 1989; Tucker and Fay, 1990) transients, increased phosphorylation of specific intracellular proteins (Seuwen et al., 1989), and 'modulation of cyclic nucleotide levels (Nishizuka, 1984). Which, if any, of these signals is necessary and sufficient for the observed cytoskeletal changes is unknown, but this system seems ideally suited for addressing just such questions. The highly organized structure found in quiescent cells provides an excellent barometer of the cytoskeletal weather, since changes and disruptions are readily observed and the structural organization lends itself well to quantitation. Fiber lengths, rates of transport, and inter- or intrafiber
FIG. 1 Serum-starvation of Swiss 3T3 fibroblasts. (a) Cells cultured for 48 hr in normal growth medium (Dulbecco’s modified Eagle’s medium + 10%calf serum) are highly motile. They extend broad lamellar protrusions (arrows) from one or more of their edges and display extensive spreading and ruffling activity. (b) Cells serumstarved by culturing for 48 hr in Dulbecco’s modified Eagle’s medium + 0.2% calf serum flatten dramatically, with a resultant increase in spread area (note that the magnification is the same in a and b). Ruffling and protrusive activity are absent, and numerous linear elements (arrowheads), which correspond to actin-containing stress fibers, become visible throughout the cytoplasm.
190
John Kolega and D. Lansing Taylor
FIG. 2 Stress fibers in serum-starved Swiss 3T3 fibroblasts. Cell that were starved as in Fig. 1 were fixed, permeablized, and stain for F-actin with fluorescein-phalloidin (Molecular Probes, Eugene, OR) and for myosin with a rabbit polyclonal antibody against myosin I1 (courtesy of J. Sellers, National Institutes of Health) followed by rhodamine-conjugated goat anti-rabbit IgG antibody. Shown are (a) fluorescein (actin) fluorescence and (b) the rhodamine (myosin) image of the same cell. Serum-starved cells have extensive arrays of actin-containing stress fibers (a). Myosin is distributed in a periodic, punctate pattern along these fibers (b, arrows).
spacings are just a few quantitative indicators available. Measurements of such features are enhanced by two additional qualities of the serum-starved fibroblast: First, the extreme flatness of the cells renders much of the fibrous structure visible by light microscopy. Using video-enhanced contrast (VEC) or fluorescence microscopy of microinjected fluorescent analogs, fibers can be tracked in living cells with minimal out-of-focus noise. Second, because quiescent fibroblasts do not translocate, fiber motion is not complicated by the superposition of bulk displacement of the whole cell. When the accessibility of the cytoskeleton to observation and quantitation is coupled with the extensive biochemical manipulations made possible by the cell’s serum sensitivity, one is graced with a potent experimental system.
111. THE MULTIMODE APPROACH
In order to dissect the regulation of actomyosin in living cells, one should ideally be able to determine precise temporal and spatial relationships among the
1 1. Regulation of Actin and Myosin I1
191
various effecting and responding elements, as well as to manipulate the effectors while observing the responders. Light microscopy, especially fluorescence microscopy, has proved to be a very powerful tool for doing just this, in that it permits direct visualization of macromolecular behavior under conditions that can be substantially controlled by the investigator (Taylor and Wang, 1989; Wang and Taylor, 1989; Giuliano et al., 1990). The technique has been made even more useful by ongoing developments in optics, electronics, and chemistry. Increasingly sensitive detectors have pushed back the limits of what can be seen, while the advent of electronic image enhancement makes it possible to extract information that is often hidden from normal human senses. Electronic image processing and analysis have also greatly increased the rate at which quantitative information can be collected from data sets. At the chemical front, brighter and more photostable fluorophores have permitted the use of lower levels of potentially harmful illumination, the capture of rapidly occuring events through the use of shorter exposure times, and the acquisition of more images from a single subject. Not only has the quality of probes been improved, but also the range of parameters that can be studied. Calcium- and pH-sensitive probes open windows onto the ionic milieu within a cell (Tsien, 1989), and the now-routine construction of fluorescent analogs of biologically active molecules (Wang, 1989) permits determination of the precise location and dynamic behavior of important proteins in living cells. Finally, with the proper choice of fluorophores, many of these probes can be made with nonoverlapping excitation and emissions spectra so that it is now possible to follow as many as five parameters at once (DeBiasio er al., 1987). Taken together, these advances allow one to view the dynamic behavior of multiple intracellular variables simultaneously or in very rapid succession. This needs to be done at two levels. First, we still know relatively little about what has been called the morphodynamics of cytoskeletal proteins. Quite simply, which components go where, and when? For example, despite the large body of experiments indicating centripetal flow of actin and surface-associated receptors, we have no knowledge of where these materials go once they reach the proximal reaches of the cell, nor how they arrive at the distal regions to begin with. And, in the course of this cycle, when and where do the assorted constituent molecules assemble and disassemble? Multimode microscopy is capable of revealing not only the precise dynamics of important structural elements, but also the order of these movements relative to that of other dynamic macromolecules. The second level of cellular dynamics with which we are concerned is the temporal relationship between cytoskeletal changes and potential regulatory signals. The serum-stimulation of 3T3 fibroblasts is a prime case in point: How is serum-stimulation of motility related temporally to serum-induced pH changes? Is the disruption of the actomyosin stress fibers preceding, following, or independent of calcium fluxes? Do these events occur at different times in different
192
John Kolega and D.Lansing Tay!or
regions of the cell, and what are the spatial relationships between signal and responder? With calcium- and pH-sensitive fluorescent probes available alongside fluorescent analogs of cytoskeletal proteins, all the tools exist to answer these questions (and countless adjunct queries) definitively simply by watching the events as they happen in the cell. Furthermore, photodetectors, radio imaging, and automated or semi-automated image analysis make precise, quantitative answers possible. Such quantitation seems crucial, given the complex and often subtle interconnections among regulatory pathways.
N. ACTOMYOSIN DYNAMICS IN SERUM-STARVED 3T3 FIBROBLASTS Stress fibers are generally regarded as essentially static structures. There is a strong correlation between the development of an extensive array of stress fibers and a decline in the rate of cell translocation (Couchman and Rees, 1979; Herman et d , 1981; Lewis et al., 1982). On short time scales of seconds or minutes, stress fibers often appear immobile, while movement is readily detected in ruffles, arcs, and blebs, the forward and backward fluctuations of the cell margin, and the retrograde transport of particles on the cell surface. However, we find that, when viewed over longer time periods (tens of minutes), stress fibers can be very dynamic even in quiescent cells. In serum-starved Swiss 3T3 fibroblasts, a large proportion of the stress fibers, particularly those lying parallel to the cells’ distal margins, are in continuous motion (Giuliano and Taylor, 1990). They move centripetally, i.e., from the cell periphery toward the nucleus, until they reach the perinuclear cytoplasm, where they disappear. The speed of this movement varies from cell to cell and from location to location within a single cell, but averages from 10 to 20 p.m/hr. In comparison, the movements of ruffles, arcs, and other centripetally directed motions that have been observed in translocating fibroblasts, are generally 5- to 20-fold faster (Table I). As old fibers move centripetally and disappear, new fibers form at the distal edges of the cell and join the parade. This continuous retrograde movement of fibers necessitates that there also be a continuous flow of actin and myosin subunits to the distal regions of the cell in order to supply material for the formation of new fibers. Such flow is probably also necessary to supply material in the distal tip of an extending protrusion, and to resupply material during other centripetal motions. How this flow is driven (in either direction) is unknown, but understanding the mechanism is likely to constitute a large step in unraveling the mysteries of cell locomotion. One possible driving force is actin assembly itself. Evidence from a number of other systems suggests that assembly of filaments
1 1 . Regulation of Actin and Myosin I1
193 TABLE I
Rates of Transport for Some Selected Objects
Object
Rate (prn/min)
Cell type
Reference
Lectin-conjugated gold particles and plastic spheres
15-30
Fish keratocytes
Kucik ct al. (1989)
Polystyrene beads
13
Swiss 3T3 fibroblasts
Fisher et al. (1988)
Anion exchange resin particles 3.3 Carbon particles Arcs
I .69 2-3 1.5-3
Chick embryo fibroblasts Hams and Dunn (1972) Chick embryo fibroblasts Abercrombie ef al. (1970) Human lung fibroblasts Soranno and Bell (1982) Chick embryo fibroblasts Heath (1981) Fisher et al. (1988)
Transverse fibers
16
Swiss 3T3 fibroblasts
Antibody- or lectin-induced patches
1-3
Chick embryo fibroblasts Heath (1983a)
“Retrograde waves’’
3-6
Aplysia neurons
Forscher and Smith (1988)
Intracellular vesicles
16
Swiss 3T3 fibroblasts
Fisher et al. (1988)
Actin, a-actinin (along stress fibers)
0.24, 0.29 Chick embryo fibroblasts McKenna and Wang (1986)
Stress fibers
0.2-0.3
Serum-starved Swiss 3T3 Giuliano and Taylor (1990) fibroblasts
may be polarized and localized at the cell surface. In S1-decorated fibroblast cytoskeletons, filaments in the leading edge are polarized with their barbed ends toward the membrane (Small et al., 1978), the barbed end being the preferred end for addition of monomers in v i m (Pollard and Cooper, 1986). In intact cells, microinjected biotin-labeled actin incorporates first at the distal ends of actin bundles, suggesting preferential polymerization at the membrane-fiber interface (Okabe and Hirokawa, 1989). Moreover, Wang (1985) and McKenna and Wang (1986) followed the motion of photobleached spots in lamellipodia and stress fibers labeled by microinjection of a fluorescent analog of actin. The motion of such bleached spots is retrograde, as is the transport of cytoplasm observed by video-enhanced contrast (Fisher et al., 1988). Forscher and Smith (1988) have also observed actin-based centripetal flow in living cells; in this case, the centripetal streaming of material in growth cones of Aplysia neurons. This streaming is reversibly inhibited by cytochalasin and, when cytochalasin is removed, is reinitiated at the cell’s edge and proceeds centripetally, suggesting polarized assembly of actin as a possible component of the driving force. We find that fiber formation and transport in serum-starved fibroblasts are also inhibited by cyto-
194
John Kolega and D. Lansing Taylor
chalasin, and that this inhibition is reversible; i.e., transport immediately resumes when the drug is washed out (Kolega e f al., 1991). If fiber transport is driven by assembly of microfilaments, it should be sensitive to other effectors of actin assembly. Proteins that cap filament assembly, such as cap42, brain capping protein (Fuchtbauer et al., 1983), and the calciumindependent fragment of gelsolin (Chaponnier et al., 1986) should block fiber transport when injected into cells. Transport should also be slowed or halted by tying up actin monomers with the actin-binding drug latrunculin (Spector el al., 1989) or by microinjection of monomer-binding proteins such as DNase 1 or excess profilin. Both DNase I and vitamin D-binding protein, which bind actin monomers in vifro,and the calcium-independent gelsolin fragment CT-40, which caps and severs actin filaments in vitro,can disrupt stress fibers when injected in non-muscle cells (Sanger et al., 1990; Cooper et al., 1987). In both cases, the resultant morphology is similar to that induced by cytochalasin. However, these authors do not report on fiber dynamics during the process. Of course, none of these experiments distinguishes between transport that is actually driven by actin polymerization and transport that merely requires the formation of an organized structure in order for motion to occur. Making this very difficult distinction might require more direct control over actin polymerization, using, for example, caged capping or severing agents or finely tuned laser microdissection. The reagents for such manipulations are not currently available “off-the-shelf,” but the technology certainly is. The questions are sufficiently important that this application should not be far off. In addition to their constitutive centripetal translocation, stress fibers may also display contractile activity. That stress fibers can contract has been shown in numerous extracted cell model systems (Kreis and Birchmeier, 1980) and in fibers isolated from cells by laser microdissection (Isenberg et al., 1976). In intact cells, Sanger e f al. (1986) have observed spacing along stress fibers of a fluorescent analog of a-actinin microinjected in chick cardiac fibroblasts. This spacing, which is highly reminiscent of the sarcomeric spacing in muscle fibers, changes over time as fibers lengthen and shorten in living cells. In quiescent 3T3 fibrobroblasts, stress fibers shorten dramatically when the cells are stimulated by addition of serum of growth factors (Giuliano and Taylor, 1990). When these fibers are prelabeled by microinjecting the cells with fluorescent analogs of actin or myosin, it can be seen that shortening is accompanied by thickening of the fiber and an increase in fluorescence intensity along the fiber. This indicates that shortening involves real fiber contraction, and not merely dissolution of the fiber ends. Furthermore, in fibers that have incorporated fluorescent myosin, the spacing of pseudosarcomeric structures and other irregularities along the fibers is observed to shorten as the fibers shorten, again indicating that the fibers are contracting.
11.
Regulation of Actin and Myosin I1
195
V. REGULATION OF STRESS FIBER CONTRACTION
What controls the change in cytoskeletal organization of quiescent 3T3 fibroblasts from one in which fibers transport in steady and apparently orderly fashion to one in which large and irregular regions of the cytoplasm contract into tight foci? Much attention has been given to one likely contractile motor, myosin 11. In extracted cell model systems, it has been shown that fiber contraction requires calcium and ATP (Isenberg et al., 1976; Kreis and Birchmeier, 1980; Masuda et al., 1984) and is accompanied by phosphorylation of the regulatory light chain of myosin (Masuda et al., 1984; Wysolmerski and Lagunoff, 1990). Furthermore, Ehrlich et al. (1986) found that contraction of glycerol-extracted fibroblasts was blocked by treating the cytoskeletons with alkaline phosphatase, and that this inhibition could be relieved by subsequent treatment with myosin light chain kinase. Wysolmerski and Lagunoff (1 990) report similar findings for permeabilized endothelial cells. In intact smooth muscle cells, Itoh et al. (1989) were able to activate contraction by microinjecting a calcium-independent fragment of the myosin light chain kinase, which phosphorylates the myosin regulatory light chain in vitro. Moreover, injection of inhibitors of the endogenous myosin light chain kinase, i.e., the pseudosubstrate peptide SM-1, inhibited the fiber contraction that is normally induced by extracellular potassium. Also in support of myosin light chain phosphorylation as a regulator of fiber contraction, Bockus and Stiles (1984) have reported an increase in the phosphorylation of a 20-kDa protein in response to serum stimulation in 3T3 cells, although the protein has not been positively identified as the regulatory light chain of myosin. In contrast to these results, decreases in myosin light chain phosphorylation have been observed when stress fibers are disrupted by elevation of CAMPlevels (Lamb et al., 1988; Kreisberg et al., 1985) or by trypsin- or EGTA-induced rounding (Bayley and Rees, 1986). These studies do not indicate the intermediate events in fiber disruption, but it is assumed that fiber dissolution is brought about by relaxation of the actomyosin complex. There are numerous avenues to the loss of stress fibers, including viral transformation, treatment with tumor promotors, and growth factor stimulation. Given the broad range of biochemical pathways that these manipulations affect, it would not be surprising to find multiple mechanisms of fiber disassembly. It should be noted that studies of these systems have generally looked at a minimum number of timepoints pre- and posttreatment, so that the dynamics of fiber disruption are not precisely known. It seems essential that these details be determined if potential regulatory mechanisms are to be compared. Unfortunately, most of the published studies have used the static tool of immunofluorescence microscopy to interpret the dynamics that occurred. The use of multimode light microscopy with the plethora of reagents now available will permit direct analysis of the dynamics.
196
John Kolega and D. Lansing Taylor
Another possible explanation for the apparently disparate effects of myosin light chain phosphorylation is that regulation of myosin may not be solely responsible for control of stress fiber organization. According to Kenney et al. (1990), the level of phosphorylation of myosin light chain required for maximal tension in smooth muscle fibers is relatively low (less than 20%). The absolute levels of light chain phosphorylation have not been determined in any of the studies mentioned above. It may well be the case that baseline levels of phosphorylation are sufficient for contractility, and that activity is regulated at some other point. For example, we have found that fiber shortening in serum-starved cells can also be induced by cytochalasin treatment, in the absence of any stimulus (Fig. 3; Kolega et al., 1991). When starved cells are treated with cytochalasin, fibers display shortening, thickening, and decreased myosin spacing in a manner very similar to that observed on serum-stimulation. These observations argue for a solation-induced decrease in the actin gel structure of stress fibers (see Taylor and Fechheimer, 1982). It is interesting to note that microinjection of the actin-binding proteins cap42, brain capping protein, and the calcium-independent fragment of gelsolin, which cap and/or sever microfilaments in vitro, produce effects nearly identical to cytochalasin treatment in normal, actively locomoting cells (Fuchtbauer et al., 1983; Cooper et al., 1987). However, without a record of the dynamic behavior of stress fibers during disruption, it is not known if the morphological changes observed in these studies involve contraction or direct dissolution. An additional caveat is that cytochalasin (Weber et al., 1976; Kreis and Birchmeier, 1980; Schliwa, 1982), gelsolin (Kanno and Sasaki, 1989), and fragmin (Avnur et al., 1983) fail to induce contraction in extracted cell model systems. This indicates that the contractile activity is probably not simple elastic recoil of a network already under excess tension. Rather, the contractile process requires at least some additional metabolic activity, as originally reported by Miranda et al. (1974). These observations also raise the question of whether serum-induced contraction is brought about primarily through activation of the contractile machinery, e.g., upregulation of myosin activity, or release of the fiber network from anchors, e.g., by partial solation of actin filaments, or both. A very promising hypothesis that appears in many different contexts is that the cytoskeleton exists in a delicate balance between the force exerted by the contractile machinery and the integrity of the structures which it is pulling against (Ingber and Jamieson, 1985; Buxbaum and Heidemann, 1988; Danowski, 1989). Contraction could then be initiated in two ways. The contractile force might increase to the point where it is greater than the resistive forces (such as the tensile strength of filaments, actin-binding proteins, or their attachment points) or the compressive strength of any components acting as struts. Alternatively, attachments, filaments, or struts may be weakened or severed, permitting existing contractility to condense the network (see Taylor and Fechheimer, 1982).
FIG. 3 Cytochalasin-induced shortening of stress fibers. Cells were starved as in Fig. 1, then microinjected with a rhodamine analog of myosin. After 1.5 hr, during which the injected cells were allowed to recover and the analog permitted to distribute through the cell, fluorescence images were acquired at regular intervals using a cooled CCD camera as the cells were perfused with 0.5 phf cytochalasin D. The time in min : sec before (-) and after (+) perfusion is indicated in the lower left comer of each panel. Arrows mark slight thickenings along a stress fiber. These distinctive points can be followed tluough the sequence of images as the fiber shortens. Note also the accumulation of myosin fluorescence into bright foci (arrowheads) as fibcrs shortcn.
198
John Kolega and D. Lansing Taylor
VI. STRESS FIBER MOVEMENT AND CELL MOTlLIlY At the most rudimentary level, any understanding of how actomyosin structures move within the cell and how this movement is regulated provides potentially valuable insight into the mechanisms a cell uses to produce “normal” cell movements. More specifically, the quiescent fibroblast model addresses two critically important processes. The first is the retrograde translocation of cytoskeletal structures. Similar centripetal movement, termed “cortical flow” by Bray and White (1988), has been observed for particles on the cell surface (Abercrombie et al., 1970; Harris and Dunn, 1972; Fisher et al., 1988), an assortment of cell-surface molecules (Taylor et al., 1971; Heath, 1983a), actin fibers (Soranno and Bell, 1982; Heath, 1983b; Fisher et al., 1988), and other intracellular material (Abercrombie et al., 1971; Bray and White, 1988; Forscher and Smith, 1988). These motions have been detected in a wide variety of cells, including the amoeba (Taylor et al., 1980), and always in regions where the cell has extended some form of locomotive protrusion. Like the various forms of successful and abortive locomotive activity that occur at the cell’s free edge (ruffling, protrusion, forward and backward fluctuations in cell spreading, etc.), this transport is constitutive. It seems likely that this constant flow of material is part of the machinery for optimally making and moving a protrusion; a cytoskeletal motor with the clutch in. Fixation of only a small portion of this material at the cell’s edge, or repolarization of a modest proportion of the flow, could produce rapid forward extension. One need only determine the means by which the rate and direction of this flow are controlled. Quiescent fibroblasts also provide a venue for studying stress fiber contractility. The shortening of centripetally moving stress fibers in serum-starved cells suggests a number of possible roles for myosin I1 in cell locomotion. Recent demonstrations that myosin 11 is not necessary for producing cell translocation (DeLozanne and Spudich, 1987; Knecht and Loomis, 1987) have quashed many long-standing conjectures for how myosin I1 could act as the motor for protrusive activity and cell locomotion. Nonetheless, there remain several functions for which it may still be a critical component. Myosin I1 may be involved in force generation over large intracellular distances, for example, during fibroblast tail retraction (Chen, 1981) and the retraction of cells from neighbors during wound healing (Conrad et al., 1989; DeBiasio et a!. , 1988). Such long-range forces may also be important in the maintenance of normal shape and cell polarity. In this regard, it is noteworthy that cells in which myosin I1 has been eliminated or inhibited from assembly develop abnormal shapes (DeLozanne and Spudich, 1987; Knecht and Loomis, 1987; Honer et al., 1988). The cells tend to become highly elongate along an axis perpendicular to the direction of movement. This is the same axis along which actomyosin stress fibers contract and along which transverse fibers are observed in lamellar protrusions. Ordinarily, these fibers
1 1. Regulation of Actin and Myosin I1
199
might be acting like a partially potent cytokinetic contractile ring: pulling in the sides of the cell, but not going so far as to divide the cell in two. If they are rendered noncontractile by the removal of myosin 11, then elongation along this axis would be expected. This supposes that myosin 11-based contraction is part of the normal cycling of actin during fibroblast translocation. If this is so, then myosin I1 activity might also be an important factor in determining the rate of this cycle, and myosin distribution might influence the location and length of the cycle’s path. In other words, it would help in polarizing protrusive activity by (1) contracting filaments at the base or rear of a protrusion (and thereby releasing actin for recycling to the front of the cell?) and (2) shortening fibers so as to confine the region of activity to a defined region of the cell rather than the maximal available width. Myosin I1 has been shown to be excluded from the initial protrusions of a fibroblast’s leading edge (Conrad et al., 1989; DeBiasio et al., 1988), presumably because its large size inhibits its translocation through a finely meshed cytoplasmic gel network (Luby-Phelps et al., 1986; Luby-Phelps and Taylor, 1988). In addition, fluorescence recovery after photobleaching (FRAP) indicates that myosin is more mobile at the base of a protrusion (where fibers shorten) than in the mid-regions (where fibers are transporting) (DeBiasio et al., 1988). This fits well with a cycle in which contraction is associated with filament solation and subsequent release of actin monomers for transport to the cell’s fore. It will be of great interest to determine the relative mobilities of the actin and myosin in and around stress fibers during the course of their traverse of the serum-starved cell. It will also be important to determine if myosin I is colocalized with myosin I1 in stress fibers and other regions of the cell. Can either myosin function in the same processes?
VII. SOME FUTURE PROSPECTS The behavior of stress fibers in serum-starved cells involve multiple transitions in the state of actomyosin. Not only do actin and myosin cycle between polymerized and unpolymerized forms as fibers form and dissolve, but the direction of their transport shifts from centripetal to centrifugal. In addition, fibers can be growing, transporting, dissolving, or contracting, depending on the physiological state of the cell and the location of the fiber within the cell. The physiological responses of stress fibers to serum-stimulation suggest many possible pathways by which some of these transitions might be controlled. At the same time, they provide significant opportunities to test the suggested mechanisms. We have already discussed the potential role of phosphorylation of the myosin light chain, as well as the possibility that contraction might be initiated by microfilament solation. Do either of these events actually occur in response to stimulation, and do they occur with the appropriate kinetics to explain the observed fiber dynam-
200
John Kolega and D. Lansing Taylor
ics? Serum-stimulation does trigger many fundamental cellular signals and messengers, several of which are known to interact directly or indirectly with the regulation of actin and myosin. For example, a release of intracellular calcium can activate the capping and severing activities of gelsolin, leading to breakdown of the actin filament network. It can also activate calcium/calmodulin-dependent myosin light chain kinase, leading to phosphorylation of the myosin light chain. These are only two of the many branchings of the calcium-regulated pathways in the cell. In addition to these effects, one must consider the regulatory consequences of the other serum-stimulated pathways: pH transients, other kinase activation, elevation of CAMP, phosphatidylinositol metabolism, etc. Modulation of primary intracellular signals with specific agents, such as ionophores, lipid metabolites, and cyclic nucleotide analogs, should help in determining which of these signals are necessary, sufficient, or superfluous. Given the complex branchings and interactions of the regulatory pathways, the preceding approach can be substantially enhanced by introducing later intermediates directly into the cell via microinjection. This latter technique has been elegantly exploited in recent studies of myosin light chain phosphorylation (Lamb et al., 1988; Fernandez et al., 1990). However, an even more powerful tool for investigating molecular regulatory events would be direct fluorescent indicators of specific modifications. For example, a myosin light chain analog whose fluorescence changes with phosphorylation of serine 19 (the site preferred by myosin light chain kinase) would permit direct visualization of this putative activation step in real time in the living cell. This would reveal spatial as well as temporal variations, and permit very precise correlation between changes in phosphorylation and variation in motility. Precedent for such an indicator has been set by Hahn et al. (1990), who recently developed a fluorescent calmodulin that changes fluorescence upon binding calcium. This was achieved by placing an environmentally sensitive fluorophore in a region of the molecule shown to be affected by calcium binding. We believe this strategy, applied to other molecules and other molecular modifications, holds great potential for generating an entire new class of fluorescent probes: indicators of potential regulatory events that occur beyond the traditional second messengers. Another advancement that will greatly facilitate studies of cytoskeletal regulation is the availability of putative regulatory molecules in photoactivatable forms. Photoactivatable, or “caged,” molecules give the investigator precise control of both when and how much of the caged molecule is released within the cell. Using appropriately masked or focused illumination, one can even control very precisely where release occurs. But, perhaps more importantly, photoactivation can be done with substances, such as calcium, that are not otherwise easily controlled from outside the cell, or which change too rapidly to be monitored following microinjection. To date, photoactivation has been performed in living cells with only a handful of agents (reviewed by Gurney and Lester, 1987;
1 I . Regulation of Actin and Myosin I1
201
McGray and Trentham, 1989), but these include caged ATP (Horiuti et al., 1989), cyclic AMP (Nerbonne et al., 1984), and inositol triphosphate (Walker et al., 1987), as well as photolabile calcium chelators (Kao et al., 1989). A much wider variety of agents is likely to become available in the future. For example, photocleavable moieties could be attached to critical sites on small regulatory peptides, such as the SM-1 inhibitor of myosin light chain kinase, or peptide drugs, such as phalloidin. The power of these tools makes them well worthy of pursuit: consider the ultimate possibility of being able to turn a given cellular process on or off literally at the flip of a switch. Finally, we must be wary of the limitations of doing chemistry through a microscope. One needs to consider the ability of an injected or perfused molecule to reach its “normal” location in the cell. How effectively does it exchange with its endogenous counterpart? Does it have access to all the locations it may have had access to had it been made in situ? How well is its activity maintained during purification, modification, and reintroduction? Some of these difficulties can be overcome by introducing altered cytoskeletal or regulatory molecules genetically, and asking what functions are disturbed by the genetic alterations (DeLozanne and Spudich, 1987; Knecht and Loomis, 1987; Andre et al., 1989; Friederich et al., 1989). This offers the added advantages that (1) the cell is not subjected to the perturbation of microinjection and (2) it is considerably easier to produce cells in sufficient quantity for more detailed biochemical analyses. Unfortunately, this approach can be confounded by redundancy both in terms of the gene for the particular molecule in question (consider the number of actin genes in the typical mammalian genome, for example), but also in terms of the ability of other molecules in the cell to perform the same function. For example, can myosin I substitute for myosin 11, and vice versa? Null mutants have been constructed for a number of cytoskeletal proteins with little, or very subtle, effect on cell behavior and morphology. Are these molecules simply unnecessary for cytoskeletal structure and motility, or can the cell use other molecules to perform some of the same functions when a single component is knocked out? In other words, one must keep in mind the possibility of compensatory regulation in a genetically deficient cell. Here, microinjection studies offer the advantage that they directly and immediately alter cellular concentration of a given component, albeit at the expense of introducing it in a “less natural” state. Thus, both genetic engineering of cytoskeletal components and direct chemical perturbation of the cytoskeleton will ultimately be needed, one confirming the other, in future dissection of the functions of the cytoskeletal constituents.
Acknowledgments We thank P. L. Post, K. M. Hahn, and K. A. Giuliano for their critical reading of this chapter during its preparation. Our research for this chapter, both in the library and the laboratory, was supported by grants #AR-32461 and GM-34639 from the National Institutes of Health.
202
John Kolega and D. Lansing Taylor
References Abercrombie, M., Heaysman, J. E. M., and Pegrum, S. M. (1970). The locomotion of fibroblasts in culture. 111. Movements of particles on the leading lamella. Exp. Cell Res. 62, 389-398. Abercrombie, M., Heaysman, J. E. M., and Pegrum, S. M. (1971). The locomotion of fibroblasts in culture. IV. Electron microscopy of the leading lamella. Exp. Cell Res. 67, 359-367. Andre, E., Brink, M . , Gerisch, G., Isenberg, G . , Noegel, A,, Schleicher, M . , Segall, J. E., and Wallraff, E. (1989). A Dictyostelium mutant deficient in severin, an F-actin fragmenting protein, shows normal motility and chemotaxis. J. Cell Eiol. 108, 985-995. Avnur, Z., Small, J. V., and Geiger, B. (1983). Actin-independent association of vinculin with the cytoplasmic aspect of the plasma membrane in cell-contact areas. J. Cell Biol. 96, 1622-1630. Bayley, S. A., and Rees, D. A. (1986). Myosin light chain phosphorylation in fibroblast shape change, detachment and patching. Eur. J. Cell Eiol. 42, 10-16. Bockus, B. J., and Stiles, C. D. (1984). Regulation of cytoskeletal architecture by platelet-derived growth factor, insulin and epidermal growth factor. Exp. Cell Res. 153, 186-197. Bray, D., and White, J. G. (1988). Cortical flow in animal cells. Science 239, 883-888. Bright, G. R., Fisher, G . W., Rogowska, J., and Taylor, D. L. (1987). Fluorescence ratio imaging microscopy: Temporal and spatial measurements of cytoplasmic pH. J. Cell B i d . 104, 10191033. Bright, G . R., Whitaker, J. E., Haugland, R. P., and Taylor, D. L. (1989). Heterogeneity of the changes in cytoplasmic pH upon stimulation of quiescent fibroblasts. J. Cell. Physiol. 141, 410-4 19. Brooks, R. F. (1976). Regulation of fibroblast cell cycle by serum. Nature (London) 260, 248-250. Burridge, K., and Connell, L. (1983). A new protein of adhesion plaques and ruffling membranes. J. Cell Eiol. 91, 359-367. Buxbaum, R. E., and Heidemann, S. R. (1988). A thermodynamic model for force integration and microtubule assembly during axonal elongation. J . Theor. Eiol. 134, 379-390. Byron, K. L . , and Villereal, M. L. (1989). Mitogen-induced [Ca2+Iichanges in individual human fibroblasts. J. Biol. Chem. 264, 18234-18239. Cassell, D., Rothenberg, P., Zhuang, X. Y., Deuel, T. T., and Glaser, L. (1983). Platelet-derived growth factor stimulates Na+ / H + and induces cytoplasmic alkalinization in NR6 cells. Proc. Natl. Acad. Sci. U.S.A. 80, 6224-6228. Chaponnier, C., Janmey, P. A . , and Yin, H. L. (1986). The actin filament-severing domain of plasma gelsolin. J. Cell Eiol. 103, 1473- 1481. Chen, W.-T. (1981). Mechanism of retraction of the trailing edge during fibroblast movement. J. Cell Eiol. 90, 187-200. Conrad, P. A., Nederlof, M. A., Herman, I. M., and Taylor, D. L. (1989). The correlated distribution of actin, myosin, and microtubules at the leading edgeof migrating Swiss 3T3 fibroblasts. Cell Moril. Cytoskel. 14, 527-543. Cooper, J. A., Bryan, J., Schwab, B., 111, Frieden, C., Loftus, D. J., and Elson, E. L. (1987). Microinjection of gelsolin into living cells. J. Cell Biol. 104, 491-501. Couchman, J. R., and Rees, D. A. (1979). The behaviour of fibroblasts migrating from chick heart explants: Changes in adhesion, locomotion and growth, and in the distribution of actomyosin and fibronectin. J . Cell Sci. 39, 149-165. Danowski, B. (1989). Fibroblast contractility and actin organization are stimulated by microtubule inhibitors. J . Cell Sci. 93, 255-266. DeBiasio, R., Bright, G . R., Ernst, L. A., Waggoner, A. S., and Taylor, D. L. (1987). Fiveparameter fluorescence imaging: Wound healing of living Swiss 3T3 cells. J . Cell Eiol. 105, I613- 1622. DeBiasio, R. L . , Wang, L.-L., Fisher, G. W., and Taylor, D. L. (1988). The dynamic distribution of
1 I . Regulation of Actin and Myosin I1
203
fluorescent analogs of actin and myosin in protrusions at the leading edge of migrating Swiss 3T3 fibroblasts. J. Cell Biol. 107, 2631-2645. DeLozanne, A,, and Spudich, J. A. (1987). Disruption of theDicryostelium myosin heavy chain gene by homologous recombination. Science 236, 1086- 1091. Ehrlich, H. P., Rajaratnam, J. B. M., and Griswold, T. (1986). ATP-induced cell contraction in dermal fibroblasts: Effects of CAMPand myosin light-chain kinase. J. Cell. Physiol. 128, 223230. Fernandez, A,, Brautigan, D. L., Mumby, M., andLamb, N. I. C. (1990). Protein phosphatase typeI, not type-2A, modulates actin microfilament integrity and myosin light chain phosphorylation in living nonmuscle cells. J. Cell Biol. 111, 103-1 12. Fisher, G. W., Conrad, P. A , , Debiasio, R. L., and Taylor, D. L. (1988). Centripetal transport of cytoplasm, actin, and the cell surface in lamellipodia of fibroblasts. Cell Motil. Cytoskel. 11, 235-247. Forscher, P., and Smith, S. J. (1988). Actions of cytochalasins on the organization of actin filaments and microtubules in a neuronal growth cone. J. Cell Biol. 107, 1505-1516. Friederich, E., Huet, C . , Arpin, M., and Louvard, D. (1989). Villin induces microvilli growth and actin redistribution in transfected fibroblasts. Cell (Cambridge, Mass.) 59, 461-475. Fuchtbauer. A,, Jockusch, B. M., Maruta, H., Kilimann, M. W., and Isenberg, G. (1983). Disruption of microfilament organization after injection of F-actin capping proteins into living tissue culture cells. Nature (LondonJ304,361-364. Geiger, B.,Tokuyasu, K. T., Dutton, A. H., and Singer, S. J. (1980). Vinculin, an intracellular protein localized at specialized sites where microfilament bundles terminate at cell membranes. Proc. Natl. Acad. Sci. U.S.A. 77, 4127-4131. Giuliano, K. A., and Taylor, D. L. (1990). Formation, transport, contraction and disassembly of stress fibers in fibroblasts. Cell Motil. Cytoskel. 16, 14-21. Giuliano, K. A,, Nederlof, M. A., DeBiasio, R.,Lanni, F., Waggoner, A. S., and Taylor, D. L. (1990). Multi-mode light microscopy. In “Optical Microscopy for Biology” (B. Herman and K. Jacobson, eds.), pp. 543-557. Wiley-Liss, New York. Gordon, W. E., 111 (1978). Immunofluorescence and structural studies of “sacomeric” units in stress fibers of cultured non-muscle cells. Exp. Cell Res. 117, 253-260. Gurney, A. M., and Lester, H. A. (1987). Light-flash physiology with synthetic photosensitive compounds. Physiol. Rev. 67, 583-617. Hahn, K. M., Waggoner, A. S., and Taylor, D. L. (1990). A calcium-sensitive fluorescent analog of calmodulin based on a novel calmodulin-binding fluorophore. J. Biol. Chem. 265, 2033520345. Harris, A. K., and Dunn, G. A. (1972). Centripetal transport of attached particles on both surfaces of moving fibroblasts. Exp. Cell Res. 73, 519-523. Heath, J. P. (1981). Arcs: Curved microfilament bundles beneath the dorsal surface of the leading lamellae of moving chick embryo fibroblasts. Cell Biol. fnt. Rep. 5, 975-980. Heath, J. P. (1983a). Direct evidence for microfilament-mediated capping of surface receptors in crawling fibroblasts. Nature (London), 302, 532-534. Heath, J. P. (1983b). Behavior and structure of the leading lamella in moving fibroblasts. I. Occurrence and centripetal movement of arc-shaped microfilament bundles beneath the dorsal cell surface. J. Cell Sci. 60, 331-354. Herman, I. M., Crisona, N. J., and Pollard, T. D. (1981). Relation between cell activity and the distribution of cytoplasmic actin and myosin. J . Cell B i d . 90, 84-91. Honer, B., Citi, S., Kendrick-Jones, J., and Jockusch, B. M. (1988). Modulation of cellular morphology and locomotory activity by antibodies against myosin. J. Cell Biol. 107, 2181-2189. Horiuti, K., Somlyo, A. V., Goldman, Y.E., and Somlyo, A. P. (1989). Kinetics of contraction
204
John Kolega and D. Lansing Taylor
induced by flash photolysis of caged adenosine triphosphate in tonic and phasic muscles. J . Gen. Physiol. 94, 769-781. Ingber, D. E., and Jamieson, J. D. (1985). Cells as tensegrity structures: Architectural regulation of histodifferentiation by physical forces transduced over basement membrane. In “Gene Expression During Normal and Malignant Differentiation” (L. C. Anderson, C. G. Gahnberg, and P. Ekblom, eds.), pp. 13-32. Academic Press, Orlando, Florida. Isenberg, C., Rathke, P. C., Hulsman, N., Franke, W. W., and Wohlfarth-Botterman, K. E. (1976). Cytoplasmic actomyosin fibrils in tissue culture cells. Cell Tissue Res. 166, 427-443. Itoh, T., Ikebe, M., Kargacin, G. J., Hartshorne, D. J., Kemp, B. E., and Fay, F. S. (1989). Effects of modulators of myosin light-chain kinase activity in single smooth muscle cells. Nature (London) 338, 164-167. Ives, H. E., and Daniel, T. O., (1987). Interrelationship between growth factor-induced pH changes and intracellular Ca2+. Proc. Natl. Acad. Sci. U.S.A. 84, 1950-1954. Kanno, K., and Sasaki, Y. (1989). Smooth muscle gelsolin and a Ca2+-sensitive contractile cell model. J. Cell. Physiol. 139, 58-67. Kao, J. P. Y., Harootunian, A. T., and Tsien, R. Y. (1989). Photochemically generated cytosolic calcium pulses and their detection by fluo-3. J. Eiol. Chem. 264, 8179-8184. Kenney, R. E., Hoar, P. E., and Kemck, W. G. L. (1990). The relationship between ATPase activity, isometric force and myosin light-chain phosphorylation. J . Biol. Chem. 265, 8642-8649. Knecht, D., and Loomis, W. F. (1987). Antisense RNA inactivation of myosin heavy chain gene expression in Dicryostelium discoideum. Science 236, 1081- 1086. Kolega, J., Janson, L. W., and Taylor, D. L. (1991). The role of solation-contraction coupling in regulating stress fiber dynamics in non-muscle cells. J. Cell Eiol. (in press). Kreis, T. E., and Birchmeier, W. (1980). Stress fiber sacromeres of fibroblasts are contractile. Cell (Cambridge, Mass.) 22, 555-561. Kreisberg, J. I., Venkatachalam, M. A., Radnik, R. A., and Patel, P. Y. (1985). Role of myosin light-chain phosphorylation and microtubules in stress fiber morphology in cultured mesangial cells. Am. J . Physiol. 249, F227-F235. Kucik, D. F., Elson, E. L., and Sheetz, M. P. (1989). Forward transport of glycoproteins on leading lamellipodia in locomoting cells. Nature (London) 340, 3 15-3 17. Lamb, N. J. C . , Fernandez, A,, Conti, M. A., Adelstein, R., Glass, D. B., Welch, W. J., and Feramisco, J. R. (1988). Regulation of actin microfilament integrity in living nonmuscle cells by the CAMP-dependent protein kinase and the myosin light chain kinase. J . Cell Eiol. 106, 19551971. Lewis, L., Verna, J. M., Levinstone, D., Sher, S., Malek, L., and Bell, E. (1982). The relationship of fibroblast translocation to cell morphology and stress fiber density. J. Cell Sci. 53, 21-36. Luby-Phelps, K., and Taylor, D. L. (1988). Subcellular compartmentalization by local differentiation of cytoplasmic structure. Cell Moril. Cytoskel. 10, 28-37. Luby-Phelps, K., Lanni, F., and Taylor, D. L. (1986). Probing the structure of the cytoplasm. J. Cell Eiol. 102, 2015-2022. Masuda, H., Owaribe, K., Hayashi, H., and Hatano, S. (1984). Ca2+ dependent contraction of human lung fibroblasts treated with TX-100: A role of Ca2+ -calmodulin-dependent phosphorylation of myosin 20,000-dalton light chain. Cell Motil. Cyroskel. 4, 315-331. McCray, J. A., and Trentham, D. A. (1989). Properties and uses of photoreactive caged compounds. Annu. Rev.Eiophys. Eiophys. Chem. 18, 239-270. McKenna, N. M., and Wang, Y.-L. (1986). Possible translocation of actin and alpha-actinin along stress fibers. Exp. Cell Res. 167, 95-105. McNeil, P. L., and Taylor, D. L. (1987). Early cytoplasmic signals and cytoskeletal responses initiated by growth factors in cultured cells. In “Cell Membranes” (E. Elson, W. Frazier, and L. Glaser, eds.), Vol. 3, pp. 365-405. Plenum, New York.
1 1. Regulation of Actin and Myosin I1
205
McNeil, P. L., McKenna, M. P., and Taylor, D. L. (1985). A transient rise in cytosolic calcium follows stimulation of quiescent cells with growth factors and is inhibitable with phorbol myristate acetate. J. Cell Eiol. 104, 372-379. Miranda, A. F., Godman, G. C., Deitch, A. D., and Tanebaum, S. W. (1974). Action of cytochalasin D on cells of established lines. I. Early events. J . Cell Eiol. 61, 481-500. Moolenaar, W. H. (1986). Effects of growth factors on intracellular pH regulation. Annu. Rev. Physiol. 48, 363-376. Moolenaar, W. H., Tertoolen, L. G. I . , and de Laat, S. W. (1984). Growth factors immediately raise cytoplasmic free Ca+ + in human fibroblasts. J. Eiol. Chem. 259, 8066-8069. Moms, J. D. H., Metcalfe, J. C., Smith, G. A,, Hesketh, T. R., and Taylor, M. V. (1984). Some mitogens cause rapid increases in free calcium in fibroblasts. FEES Leff. 169, 189-193. Nerbonne, J. M., Richard, S., Nargeot, J., and Lester, H. A. (1984). New photoactivatable cyclic nucleotides produce intracellular jumps in cyclic AMP and cyclic GMP concentrations. Nature (London) 310, 74-76. Nishizuka, Y. (1984). The role of protein kinase C in cell surface signal transduction and tumor production. Nufure (London) 308, 694-698. Okabe, S . , and Hirokawa, N. ( I 989). Incorporation and turnover of biotin-labelled actin microinjected into fibroblastic cells: An irnmunoelectron microscopic study. J. Cell Eiol. 109, 15811595. Pollard, T. D., and Cooper, J. A. (1986). Actin and actin-binding proteins. A critical evaluation of mechanism and function. Annu. Rev. Eiochem. 55, 987- 1035. Rozengurt, E., and Mendoza, S . (1980). Monovalent ion fluxes and the control of cell proliferation in cultured fibroblasts. Ann. N . Y . Acad. Sci. 339, 175-190. Sanger, J. M., Mittal, B., Pochapin, M., and Sanger, J. W (1986). Observations of microfilament bundles in living cells microinjected with fluorescently labelled contractile proteins. J. Cell Sci. Suppl. 5, 17-44. Sanger, J. M., Dabiri, G., Mittal, B . , Kowalski, M. A,, Haddad, J. G., and Sanger, J. W. (1990). Disruption of microfilament organization in living nonmuscle cells by microinjection of plasma vitamin D-binding protein or DNase I. Proc. Nurl. Acad. Sci. U.S.A. 87, 54745478. Sanger, J. W., Sanger, J. M., and Jockusch, B. M. (1983). Differences in the stress fibers between fibroblasts and epithelial cells. J . Cell Eiol. 96, 961-969. Schliwa, M. (1982). Action of cytochalasin D on cytoskeletal networks. J . Cell Eiol. 92, 79-91. Seuwen, K., Chambard, J. C., L‘Allemain. G., Magnaldo, I., Pans, S., and Pouyssegur, J. (1989). Thrombin as a growth factor: Mechanism of signal transduction. Afheroscler. Rev. 19, 217232. Small, J. V., Isenberg, G . , and Celis, J. E. (1978). Polarity of actin at the leading edge of cultured cells. Nature (London) 212, 638-539. Soranno, T., and Bell, E. (1982). Cytoskeletal dynamics of spreading and translocating cells. J. Cell Ei01. 95, 127-136. Spector, I., Schochet, N. R., Elasberger, D., and Kashman, Y. (1989). Latrunculins-novel marine macrolides that disrupt microfilament organization and affect cell growth: I. Comparison with cytochalasin D. Cell Mofil. Cyfoskel. 13, 127-144. Taylor, D. L., and Fechheimer, M. (1982). Cytoplasmic structure and contractility: The solationcontraction hypothesis. Philos. Trans. R. SOC. London. Ser. E 299, 185-197. Taylor, D. L., and Wang, Y.-L. (1989). “Fluorescence Microscopy of Living Cells in Culture,” Part B. Academic Press, San Diego, California. Taylor, D. L., Wang, Y.-L., and Heiple, J. M. (1980). Contractile basis of ameboid movement VII. The distribution of fluorescently labeled actin in living amebas. J. Cell Eiol. 86, 590-598. Taylor, R. B., Duffins, W. P. H., Raff, M. C., and DePetris, S. (1971). Redistribution and pi-
206
John Kolega and D. Lansing Taylor
nocytosis of lymphocyte surface immunoglobulin molecules induced by anti-immunoglobulin antibody. Nature (London), New Biol. 233, 225-229. Tsien, R. Y. (1989). Fluorescent indicators of ion concentrations. Methods Cell Biol. 30, 127-156. lhcker, R. W., and Fay, F. S. (1990). Distribution of intracellular free calcium in quiescent BALB/c 3T3 cells stimulated with platelet-derived growth factor. Eur. J. Cell Biol. 51, 120-127. Walker, J. W., Somlyo, A. V., Goldman, Y. E., Somlyo, A. P., and Trentham, D. A. (1987). Kinetics of smooth muscle activation by laser pulse photolysis of caged inositol I ,4,5-triphosphate. Nature (London) 327, 249-251. Wang, Y.-L. (1985). Exchange of actin subunits at the leading edge of living fibroblasts: Possible role of treadmilling. J. Cell Biol. 101, 597-602. Wang, Y.-L. (1989). Fluorescent analog cytochemistry: Tracing functional protein components in living cells. Merhods Cell Biol. 29, 1-12. Wang, Y.-L., and Taylor, D. L. (1989). “Fluorescence Microscopy of Living Cells in Culture,” Part A. Academic Press, San Diego, California. Weber, K., Rathke, P. C., Osborn, M., and Franke, W. W. (1976). Distribution of actin and tubulin in cells and in glycerinated cell models after treatment with cytochalasin B (CB). Exp. Cell Res. 102, 285-297. Wysolmerski, R. B., and Lagunoff, D. (1990). Involvement of myosin light-chain kinase in endothelial cell retraction. Proc. Narl. Acad. Sci. U.S.A. 87, 16-20.
CHAPTER 12
Expression and Function of Genetically Engineered Actin-Binding Proteins in Dictyostelium Walter Witke, Michael Schleicher, Helmut Einberger, Wolfgang F. Neubert, and Angelika A. Noegel Max-Planck-Institute for Biochemistry D-8033 Martinsried, Germany
I. The Microfilament System of Dictyostelium discoideum 11. F-Actin Cross-Linking Proteins in Dicfyostelium discoideum 111. Genetic Manipulation of a-Actinin in Escherichia coli and Dicfyostelium discoideum A. Investigation of the Actin-Binding Domain B. Investigation of the EF-Hand Domains C. Investigation of the in Vivo Function of a-Actinin References
Dictyostelium discoideum cells are highly motile cells. As free living amoebae they feed on bacteria, under starvation conditions they respond chemotactically to CAMP, form an aggregate, and differentiate into spore and stalk cells in the fruiting body. Dictyosteliurn cells are easily cultivated; they grow axenically or on bacteria. Dictyostelium discoideum is a haploid organism. This allows direct isolation of mutants. A parasexual genetic system has been established (Newell, 1978) and it has been possible to use molecular genetic methods since the description of a transformation system (Nellen et al., 1984.) These features have made D . discoideum a favorable organism for the investigation of various aspects of cell motility and the cytoskeleton.
Current Topics in Membranes, Volume 38 Copyright 6 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.
207
208
Walter Witke et al.
1. THE MICROFILAMENT SYSTEM OF Dictyostelium discoideum As in other nonmuscle cells, in D.discoideurn the majority of actin filaments are located underneath the membrane. Several proteins which are thought to regulate polymerization and depolymerization of the filament network have been isolated. A brief summary of actin-binding proteins from D . discoideurn is shown in Fig. 1. The interactions of these proteins with G- or F-actin are similar to the ones observed for actin-binding proteins from other organisms, indicating that they belong to classes of proteins with conserved function during evolution (Schleicher et al., 1988a). Isolation of the corresponding genes and determination of the primary structures showed that they are not only conserved with regard to their function but also on the level of the primary structure.
P24 (CYPPqP)
C a p l o o ~ b
4 Q4 4
*
profilin 1/11
a
Cap32/34
\dl
severin
crosslinking proteins
FIG. 1 Actin-binding proteins in Dicryostelium discoideum. Profilins (Kaiser et al., 1989; M. Haugwitz and M. Schleicher, unpublished) interact with G-actin, thereby inhibiting polymerization. Cap32134 (Schleicher er al., 1984) and severin (Brown et al., 1982) decrease the viscosity of the cytoplasm by capping and severing actin filaments. Cross-linking proteins increase the stage of gelation; myosins produce contractile forces (Warrick and Spudich, 1987; Korn and Hammer, 1988). Ponticulin (Wuestehube and Luna, 1987), hisactophilin (Scheel er al., 1989) and p24 (Stratford and Brown, 1985; Noegel er al., 1990) are membrane-associated actin-binding proteins; Cap100 seems to be a nucleation inhibitor (M.Schleicher, unpublished).
12. Actin-Binding Proteins in Dictyostelium
209
II. F-ACTIN CROSS-LINKING PROTEINS IN Dictyosteliurn discoideurn The 97-kDa protein a-actinin (Condeelis and Vahey, 1982; Fechheimer et al., 1982) and the 120-kDa gelation factor (Condeelis et al., 1981) are responsible for the two main cross-linking activities in a D. discoideum homogenate. Both proteins are homodimers in which the subunits are arranged in antiparallel fashion (Wallraff et al., 1986; Brink et al., 1991). In the electron microscope they appear as elongated rodlike molecules with a length of approximately 40 and 35 nm, respectively (Condeelis et al., 1984). a-Actinin was first isolated from the rabbit skeletal muscle (Ebashi and Ebashi, 1965) and later also found in non-muscle cells (Lazarides and Burridge, 1975). In D. discoideum it represents about 1% of soluble protein. The crosslinking activity is observed in the presence of EGTA and is completely inhibited by the addition of calcium ions. a-Actinin of D. discoideum is encoded by a single gene (Witke et al., 1986). The genomic structure has been elucidated and the cDNA and genomic sequences are known (Noegel et al., 1987; Witke and Noegel, 1990). Analysis of the cDNA-derived sequence revealed the presence of three structural domains which appear to be important for the F-actin crosslinking function of a-actinin: (1) two complete EF-hands are responsible for the Ca2 regulation; (2) internal repeats with a high a-helical potential render the aactinin a spectrin-like molecule; (3) an amino-terminal domain of about 250 amino acids is highly conserved between a-actinins from various organisms (Baron et al., 1987; Arimura et al., 1988). Because of the high degree of similarity, this region was proposed to represent the actin-binding domain of aactinin. In Dictyostelium a-actinin these domains are organized in exons. The coding region consists of three exons separated by two short introns. Exon 2 contains the putative actin-binding site and rod-forming domain of a-actinin, whereas exon 3 includes the two EF-hand domains (Witke and Noegel, 1990). A comparable arrangement has also been noted in spectrin, an actin crosslinking molecule consisting of two different subunits (a,p) which form a tetramer (a2, p2). The putative actin-binding domain of spectrin is located in the amino-terminal region of the P-subunit and has been found to be highly homologous to the actin-binding domain of a-actinin (Byers et al., 1989). A similar sequence was also observed in dystrophin, a protein which is lacking or altered in patients suffering from Duchenne or Becker muscular dystrophy (Koenig et al., 1988), and in the gelation factor of Dictyostelium (Noegel et al., 1989) (Fig. 2). The gelation factor, although in its shape similar to a-actinin, does not contain ahelical regions. Instead, it is composed of the N-terminal domain and a sixfold repeat structure. These repeats are built from about 100 residues each, have a high content of glycine and proline residues, and most likely fold into a cross-f3 conformation. +
210
Walter Witke et al.
CONSENSUS:
V
LV 1
I G
* *
DIVDGN * *
KLTLGLIWTI I LR *
I
* *
*
Q
I
FIG. 2 Consensus sequence between actin-binding proteins. Sequence comparisons between D. discoideum a-actinin (DA), chicken a-actinin (CA), D . discoideum gelation factor (DG), chicken dystrophin (CD), human dystrophin (HD), and Drosophila P-spectrin (DS) showed the strongest homology in a 31-amino acid consensus region. Identical residues present in at least three of the proteins are boxed. The consensus sequence lists positions where residues among the six proteins are identical. Residues that are present in all listed sequences are indicated by a star.
111. GENETIC MANIPULATION OF a-ACTININ IN Escherichia coli AND Dictyostelium discoideum A. Investigation of the Actin-Binding Domain
The actin-binding activity of the conserved amino-terminal domain was investigated using a bacterially expressed a-actinin peptide. This 50-kDa peptide, which is encoded by a 1.2-kb EcoRI fragment of the a-actinin gene (Witke et al., 1986), contained most of the conserved region as well as an epitope recognized by the monoclonal antibody 47-60-8 (Schleicher et al., 1988b), enabling us to follow the purification of the peptide and to assay its actin-binding activity. The EcoRI fragment was cloned into the expression vector pIMS6 (Simon et al., 1988) and led to a polypeptide that contained only two additional amino acids after the starting methionine. F-actin binding activity of the bacterially expressed polypeptide was tested in a cosedimentation assay in which actin filaments are sedimented under high speed and binding proteins are enriched in the pellet. Figure 3 shows the cosedimentation of the fusion protein encoded by the 1.2-kb EcoRI fragment. A bacterially expressed polypeptide comprising the region of highest similarity (amino acid residues 97- 172 of D . discoideum a-actinin) could also be cosedimented with F-atin and confirmed the actin-binding activity of this region. The 50-kDa peptide discussed above is now expressed in D . discoideum under the control of a Dictyostelium actin promoter and investigated with regard to its localization in the cell. Overexpression of this actin-binding protein may interfere with a-actinin function and may give hints about aactinin’s in vivo function.
12. Actin-Binding Proteins in Dictyostelium
211
FIG. 3 Cosedimentation of the bacterially expressed fusion protein encoded by the 1.2-kb EcoRI fragment with F-actin. Equal amounts of the original soluble fraction (lanes I / I ’), the supernatant (lanes 2/2’), and the pellet (lanes 3/3’), after centrifugation were separated by SDS-PAGE and the gels either stained with Coomassie blue (lanes 1-3) or blotted onto nitrocellulose, probed with directly iodinated monoclonal antibody 47-60-8, and subjected to autoradiography (lanes 1’-3’),
B. hves tigation of the EF-Hand Domains
The elucidation of the a-actinin sequence in D. discoideum led to the development of a model of how a-actinin could interact with F-actin in a Ca2 -sensitive manner (Noegel et al., 1987). In the a-actinin monomer, the actin-binding domain is located at the amino terminus and the EF-hand domains are close to the carboxy terminus. Because of the physical separation of these domains - aactinin has a rodlike structure with amino and carboxy termini separated in space - it is difficult to conceive a model of their interaction with the same subunit. In a dimer, actin- and Ca2 -binding sites would be juxtaposed and an interaction between both of them can be imagined. In analogy to studies on the crystal structure of troponin C (Herzberg and James, 1985) and calmodulin (Babu et al., 1985), on binding of Ca2 a conformational change of the EF-hands in a-actinin +
+
+
212
Walter Witke et al.
could be transmitted to the neighboring actin-binding site and, thereby, the actinbinding activity of a-actinin changed. In order to test this model, the Ca*+-binding activity of the molecule was altered by introducing a short sequence into the second EF-hand of D. discoideum a-actinin. This insert codes for a viral epitope for which a monoclonal antibody is available (Einberger et al., 1990) (Fig. 4). The modified gene was then introduced by calcium phosphate-mediated transformation into D. discoideum strain HG 1 130, a nitrosoguanidine-generated mutant producing only trace amounts of a-actinin (Wallraff er al., 1986; Schleicher et al., 1988b; Witke and Noegel, 1990). The transformation vector carried a selection marker and the in vitro-modified a-actinin cDNA without promoter sequences. Positive transformants were tested for expression of the viral epitope and could only arise if the cDNA had inserted behind the endogenous a-actinin promoter via homologous recombination. The mutated a-actinin was purified from transfomants and tested in vitro for F-actin cross-linking activity and 45Ca2+-binding. In a viscometric assay (MacLean-Fletcher and Pollard, 1980), it was no longer able to increase the viscosity of the F-actin solution and binding of Ca2+ was greatly reduced. These results indicate that the alteration of the second EF-hand in the described way leads to a conformational change mimicking a state in which Ca2+ is bound and leaves the molecule inactive. C. Investigation of the in Vivo Function of a-Actinin
The conservation of domains among cytoskeletal proteins points to an important function of these proteins. However, from D. discoideum several strains have been isolated both after conventional mutagensis and after disruption of the corresponding genes via homologous recombinations that lack either a-actinin (Wallraff et al., 1986; Witke et al., 1987) or the gelation factor (Brink et al., 1991). Lack of either one of the two proteins did not affect viability or development, and motility and chemotaxis were not altered as compared to the parent
2
X Y Z -Y-X
EF-hand
-Z
LTEEQLNQVISKIDTDGNGTISFEEFI DYMVSSR GTISFEEFIDYMVSSR LTEEQLNQV I SK I
/
/
Sendai epitope
\
DGSLGDIEPYDSs
FIG. 4 Genetic alteration of the second EF-hand of a-actinin. In position 2418 of the a-actinin cDNA a 42-bp oligonucleotide was inserted which codes for a peptide of 13 amino acids. The functionally important amino acid residues which provide the liganding oxygens for Ca2+ -binding are indicated above the EF-hand sequence.
12. Actin-Binding Proteins in Dictyostelium
213
strain. These results led to the concept of redundancy, which suggests that proteins with similar domains can compensate for the lack of another protein with comparable function (Schleicher et al., 1988b; Bray and Vasiliev, 1989). The data emerging from determination of the primary structures of cross-linking proteins seem to support this hypothesis. The isolation of a-actinin and gelation factor negative cells was facilitated because of the availability of nitrosoguanidine-generatedmutants that lacked either a-actinin (strain HGll30: Wallraff et al., 1986; Schleicher et al., 1988b) or the gelation factor (strain HG1264: Brink et al., 1990). By disruption of the aactinin gene in strain HG1264 and the gelation factor gene in strain HGll30, we generated cells that lacked two F-actin cross-linking proteins. The double mutans were still motile and chemotactically active. Development was initiated normally as tested with antibodies directed towards stage-specific proteins and cDNA probes, and spores and stalk cells were formed. However, spores and stalk cells did not undergo the final morphogenetic movements and were only occasionally organized into a normal fruiting body with a stalk and a spore head. Most of the cells remained in aggregates on the agar surface (Fig. 5). As proof that the inability of these strains to undergo morphogenesis was due to the lack of the two F-actin cross-linking proteins, a vector was introduced into one of the double mutants which allowed synthesis of a-actinin again. In all rescued transformants the ability to complete development was restored and fruiting bodies indis-
wildtype
GA mutant GA mutant rescued
FIG. 5 Late developmental states of wildtype and mutant strains of Dicryosreliurn. Wildtype strain AX2, the a-actinin- and gelation factor-deficient mutant CAI. I and a transformant of GAI. 1 expressing a-actinin again (GA mutant rescued) were grown on agar plates containing Klebsiellu uerogenes. Wildtype and rescued mutant formed fruiting bodies with stalk and spore head whereas the GA mutant was not able to culminate.
214
Walter Witke et al.
tinguishable from wildtype were formed. These data indicate that (a) the two cross-linking proteins a-actinin and gelation factor have a distinct function during development, and (b) that they are components of a redundant system and can replace each other with regard to this specific function.
References
Arimura, C., Suzuki, T., Yanagisawa, M., Imamura, M., Hamada, Y.,and Masaki, T. (1988). Primary structure of skeletal muscle and fibroblast a-actinins deduced from cDNA sequences. Eur. J . Biochem. 177, 649-655. Babu, Y. S., Sack, J. S . , Greenhough, T. J., Bugg, C. E., Means, A. R., and Cook, W. J. (1985). Three-dimensional structure of calmodulin. Narure (London) 315, 37-40. Baron, M. D., Davison, M. D., Jones, P., and Critchley, D. R. (1987). The sequence of chick aactinin reveals homologies to spectrin and calmodulin. J . B i d . Chem. 262, 17623-17629. Bray, D., and Vasiliev, J. (1989). Networks from mutants. Nature (London) 338, 203-204. Brink, M., Gerisch, G., Isenberg, G., Noegel, A. A., Segall, J. E., Wallraff, E., and Schleicher, M. (1990). A Dictyostelium mutant lacking an F-actin cross-linking protein, the 120-kd gelation factor. J. Cell Biol. 111, 1477-1489. Brown, S. S., Yamamoto, K., and Spudich, J. A. (1982). A 40,000-dalton protein from Dictyosrelium discoideum affects assembly properties of actin in a Ca2+ -dependent manner. J. Cell Biol. 93, 205-210. Byers, T. J., Husain-Chisthi, A,, Dubreuil, R. R., Branton, D., and Goldstein, L. S. B. (1989). Sequence similarity of the amino-terminal domain of Drosophila beta spectrin to alpha actinin and dystrophin. J. Cell Biol. 109, 1633-1641. Condeelis, J., and Vahey, M. (1982). A calcium- and pH-regulated protein from Dictyostelium discoideum that cross-links actin filmanets. J . Cell Biol. 94, 466-471. Condeelis, J., Salisbury, J., and Fujiwara, K. (1981). A new protein that gels F-actin in the cell cortex of Dictyosrelium discoideum. Nature (London) 292, 16I - 163. Condeelis, J., Vahey, M., Carboni, 1. M., DeMey, I., and Ogihara, S., (1984). Properties of the 120,000- and 95,000-dalton actin-binding proteins from Dictyosrelium discoideum and their possible functions in assembling the cytoplasmic matrix. J. Cell Biol. 99, 119s-126s. Ebashi, S., and Ebashi, F. (1965). a-Actinin, a new structural protein from striated muscle. J. Biochem. (Tokyo) 58, 7-12. Einberger, H., Mertz, R., Hofschneider, P.H., and Neubert, W. J. (1990). Purification, renaturation and reconstituted protein kinase activity of the Sendai virus large (L) protein: L protein phosphorylates the NP and P proteins in vitro. J . Virol. 64, 4274-4280. Fechheimer, M., Brier, J . , Rockwell. M., Luna, E. J., and Taylor, D. L. (1982). A calcium- and pHregulated actin binding protein from D . discoideum. Cell Moril.2, 287-308. Herzberg, 0..and James, M. N. G. (1985). Structure of the calcium regulatory muscle protein troponin C at 2.8A resolution. Narure (London) 313, 653-659. Kaiser, D. A., Goldschmidt-Clermont, P. J.. Levin, B. A., and Pollard, T. D. (1989). Characterization of renatured profilin purified by urea elution from poly-lproline agarose columns. Cell Moiil. Cytoskel. 14, 25 1-262. Koenig, M., Monaco, A. P., and Kunkel, L. M. (1988). The complete sequence of dystrophin predicts a rod-shaped cytoskeletal protein. Cell (Cambridge, Mass.) 53, 219-228. Korn, E. D., and Hammer, I. A,, 111, (1988). Myosins of nonmuscle cells. Annu. Rev. Biophys. Biophys. Chem. 17, 23-45. Lazarides, E., and Burridge, K. (1975). a-Actinin: Immunofluorescent localization of a muscle structural protein in nonmuscle cells. Cell (Cambridge, Mass.) 6 , 289-298.
12. Actin-Binding Proteins in Dictyostelium
215
MacLean-Fletcher, S., and Pollard, T. D. (1980). Viscometric analysis of the gelation of Acanrhamoeba extracts and purification of two gelation factors. J. Cell Eiol. 85, 414-428. Nellen, W., Silan, C., and Firtel, R. A. (1984). DNA-mediated transformation in Dictyosteliurn discoideum: Regulated expression of an actin gene fusion. Mol. Cell. Eiol. 4, 2890-2898. Newell, P. D. (1978). Genetics of the cellular slime molds. Annu. Rev. Genet. 12, 69-93. Noegel, A. A , , Witke, W., and Schleicher, M. (1987). Calcium-sensitive non-muscle a-actinin contains EF-hand structures and highly conserved regions. FEES Leu. 221, 391-396. Noegel, A. A , , Rapp, S., Lottspeich, F., Schleicher, M., and Stewart, M. (1989). The Dictyosteliurn gelation factor shares a putative actin binding site with a-actinins and dystrophin and also has a rod domain containing six 100-residue motifs that appear to have a cross-beta conformation. J. Cell Eiol. 109, 607-618. Noegel, A. A,, Gerisch, G., Lottspeich, F., and Schleicher, M. (1990). A protein with homology to the C-terminal repeat sequence of Octopus rhodopsin and synaptophysin is member of a multigene family in Dictyostelium discoideurn. FEES Lett. 266, 118-122. Scheel, J., Ziegelbauer, K., Kupke, T., Humbel, B. M., Noegel, A. A., Gerisch, G., and Schleicher, M. (1989). Hisactophilin, a histidine-rich actin-binding protein from Dictyostelium discoideum. J. Eiol. Chem. 264, 2832-2839. Schleicher, M., Gerisch. G., and Isenberg, G. (1984). New actin binding proteins from Dictyostelium discoideum. EMBO J . 3, 2095-2100. Schleicher, M., Andri, E., Hartmann, H., and Noegel, A. A. (1988a). Actin-binding proteins are conserved from slime mold to man. Dev. Genet. 9, 521-530. Schleicher, M., Noegel, A,, Schwarz, T., Wallraff, E., Brink, M., Faix, J., Gerisch, G., and Isenberg, G. (1988b). A Dicfyostelium mutant with severe defects in a-actinin: Its characterization using cDNA probes and monoclonal antibodies. J. Cell Sci. 90,59-71. Simon, M.-N., Mutzel, R., Mutzel, H., and Veron, M. (1988). Vectors for expression of truncated coding sequences in Escherichia coli. Plasmid 19, 94-102. Stratford, C. A,, and Brown, S. S . (1985). Isolation of an actin-binding protein from membranes of Dicfyostelium discoideum. J. Cell Biol. 100, 127-735. Wallraff, E., Schleicher, M., Modersitzki, M., Rieger, D.. Isenberg, G., and Gerisch, G. (1986). Selection of Dictyostelium mutants defective in cytoskeletal proteins: Use of an antibody that binds to the ends of alpha-actinin rods. EMBO J. 5 , 61-67. Warrick, H. M., and Spudich, 1. A. (1987). Myosin structure and function in cell motility. Annu. Rev. Cell Biol. 3, 379-421. Witke, W., and Noegel, A. A.(1990). A single base exchange in an intron of the Dicryostelium discoideum a-actinin gene inhibits correct splicing of the RNA but allows transport to the cytoplasm and translation. J. Biol. Chem. 265, 34-39. Witke, W. Schleicher, M., Lottspeich, F., and Noegel, A. (1986). Studies on transcription, translation, and structure of a-actinin in Dicfyosrelium discoideum. J. Cell Biol. 103, 969-975. Witke, W., Nellen, W., and Noegel, A. (1987). Homologous recombination in the Dictyostelium aactinin gene leads to an altered mRNA and lack of protein. EMBO J. 6, 4143-4148. Wuestehube, L. I., and Luna, E. J. (1987). F-actin binds to the cytoplasmic surface of ponticulin, a 17-kD integral glycoprotein from Dictyostelium discoideum plasma membranes. J . Cell Eiol. 105, 1741-1751.
This Page Intentionally Left Blank
CHAPTER 13
Interaction of Profilins with Membrane Lipids Thomas D. Pollard, Laura Machesky, and Pascal Goldschmidt-Clermont Department of Cell Biology and Anatomy The Johns Hopkins Medical School Baltimore, Maryland 21205
I. Binding of Profilin to Phospholipids 11. Profilin Inhibits Soluble Phospholipase C 111. Interaction of Profilin with Actin IV. What Does F’rofilin Actually Do in a Cell? References
Following the initial discovery by Lassing and Lindberg (1985) and by Janmey and Stossel (1987) that polyphosphoinositides could dissociate the actin-binding proteins profilin and gelsolin from actin, we have carried out a detailed analysis of the binding of several isoforms of profilin with phospholipids (Machesky et al., 1990). Our initial findings (Goldschmidt et al., 1990) indicated that human platelet profilin binds to small clusters of phosphatidylinositol 4,5-bisphosphate (PIP,) and in doing so inhibits the hydrolysis of the PIP, by a soluble phosphoinositide-specific phospholipase C (PLC). These observations suggest that profilin might participate in the regulation of both actin polymerization and the phosphoinositide (PI) signaling pathway. In this chapter we summarize our recent observations that confirm these initial results.
1. BINDING OF PROFlLlN TO PHOSPHOLIPIDS We used small zone and equilibrium gel filtration (Hummel and Dreyer, 1962) to analyze the binding of profilin isoforms to various phospholipids. Our results Currenr Topics in Membranes, Volume 38 Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any Form reserved.
217
218
Thomas D. Pollard et al.
are summarized in Table I. These results demonstrate that this interaction is specific for both the protein and the phospholipid. The experiments have been carried out on lipids in a variety of different physical forms, including micelles, small unilamellar vesicles, and large unilamellar vesicles made by an extrusion technique (LUVETs). Human platelet profilin binds to PIP, and PIP with a dissociation constant in the range of 1p.M. The stoichiometry of the complex is about one profilin molecule per four or five phospholipids. Acanthamoeba profilin I1 binds to PIP, with a similar affinity and stoichiometry. On the other hand, the acidic isoform of Acantharnoeba profilin, profilin I, only binds weakly to PIP,, with a dissociation constant greater than 50 to 100 p.M. This remarkable difference between profilin I and profilin I1 provides evidence for the specificity of binding because these two profilins have very similar amino acid sequences. They differ at only 13 out of 125 residues. Most importantly, profilin I1 has a net positive charge greater by +2 than profilin I. Although we have no direct evidence that the additional basic residues are directly responsible for the binding, we note that, in experiments with peptides binding to acidic phospholipid head groups (McLaughlin, 1989), a single positive charge contributes about 1.4 kcal per mole of binding energy and therefore one order of magnitude to the equilibrium constant. The small number of differences between these two isoforms will provide an opportunity through in vitro mutagenesis to establish which amino acid residues contribute to the affinity of profilin I1 for PIP,. Platelet profilin binds with approximately equal affinity to PIP, and PIP but only very weakly to phosphatidylinositol (PI) and phosphatidylserine (PS), and not at all to phosphatidylethanolamine (PE) or phosphatidylcholine (PC). These observations provide evidence that the 4-phosphate on the inositol is essential for high-affinity binding of the lipid head group to profilin. Although we suspect that the interaction is electrostatic, we note that a negatively charged phospholipid head group is not sufficient to account for the high-affinity binding of profilin. All
TABLE I Dissociation Equilibrium Constants for the Binding of Profilins to Phospholipids~ PIP,
(PM) Acantharnoeba profilin I Acantharnoeba profilin I1 Human platelet profilin 0 PIP,, Phosphatidylinositol phosphatidylserine.
- 100 -5 -1
PS Not detectable Not detectable Weak
4,5-bisphosphate;
PS,
13. Interactions of Profilins with Lipids
219
of our results indicate that four to five phospholipid head groups contribute to the binding site for one profilin molecule. In keeping with this conclusion, based primarily on small zone gel filtration experiments, we find that individual PIP, head groups in the form of inositol trisphosphate (IP,) do not bind with high affinity to profilin in small zone gel filtration experiments. The actual mechanism of binding of the profilin to PIP or PIP, in a phospholipid bilayer is likely to be much more complex than suggested by the designation of a single equilibrium constant for this reaction. For example, it seems likely that the formation of such a complex must involve the primary, low-affinity association of profilin with a single phospholipid head group followed by trapping of additional phospholipids that diffuse into the region of the primary complex. These initial complexes are likely to have very low affinities and therefore very short life times. Additional, much more sophisticated experimentation will be necessary to reconstruct the pathway of binding and dissociation, even in a highly simplified system reconstituted from purified lipids and profilin. The situation is undoubtedly even more complicated in the context of a cell where the mole fraction of PIP and PIP, is low in the lipid bilayer and additional integral and peripheral membrane proteins may provide competing or enhancing reactions. Given the concentrations of PIP, (1OOpM) and profilin (50pM) in platelets, the 4-5 to 1 molar ratio of lipid to profilin in the complex and the micromolar dissociation constant, a large fraction of the PIP, may be bound to profilin at equilibrium in the cell.
II. PROFlLlN INHIBITS SOLUBLE PHOSPHOLIPASE C Our initial experiments showed that platelet profilin inhibits the hydrolysis of PIP, by soluble platelet phospholipase C-11 (Goldschmidt er al., 1990). In these experiments and new experiments (Machesky er al., 1990) with the two Acanthamoeba profilin isoforms, we find that the affinity of these profilins for the substrate PIP, accounts to a first approximation for the profilin concentration dependence of the inhibition of these enzymes. Platelet profilin is the strongest inhibitor of the enzyme, followed by Acanrhamoeba profilin I1 and Acanthamoeba profilin I, which is only a very weak inhibitor of the enzyme. We have confirmed these observations with purified brain PLC-y, another soluble phospholipase C. These biochemical experiments combined with our knowledge about the concentrations of profilin and PIP, in cells suggest that the profilins can influence the PI signaling pathway by binding to PIP,, the substrate for PIX, which gives rise to the second messengers IP, and diacylglycerol (DAG). Thus, profilin is a candidate for the mysterious inhibitor of PJX in resting cells (Rhee et al., 1989).
220
Thomas D. Pollard et al.
In activated cells there is a large increase in the flux of molecules through the inositol phosphate pathway leading to a depletion of PI at the expense of the net production IP, and DAG. At least in platelets, the concentrations of PIP and PIP, remain approximately constant even in strongly stimulated cells (Wilson et al., 1985), although there can be a small drop in the concentration of PIP, up to about 50% of the total. This flux requires the hydrolysis of PIP, by PLC, so there must be some way of overcoming the sequestration of PIP, by the high concentrations of profilin that are found in the cells. In recent studies (Goldschmidt-Clermont et al., 1991a) we found that phosphorylation of PLC-y-1on tyrosines by the epidermal growth factor receptor gives this PLC full activity even in the presence of profilin (Fig. 1). This suggests profilin may be a component of the biochemical machinery that couples this class of growth factor receptors to the production of IP,. On the other hand, profilin
FIG. 1 A schematic model for the regulation of IP3 production in cells responsive to epidermal growth factor (EGF). (A) In the resting cell phospholipase C-y-1 is inhibited by profilin binding to membrane PIP2. (B) When EGF binds to its receptor on the cell surface, the receptor dimerizes and activates the tyrosine kinase activity in the cytoplasmic domain. This activated kinase phosphorylates phospholipase C-y-1 on specific tyrosine residues. The phosphorylated phospholipase C-y-1 is not inhibited by profilin, so IP3 is produced. See Goldschmidt-Clermont ef al.. (1991a) for details.
221
13. Interactions of Profilins with Lipids
does not inhibit the membrane associated PLC-p (Goldschmidt-Clermont et al., 1991a). This class of phospholipases appears to be regulated by a different mechanism involving a novel G-protein (Smrcka et al., 1991). These two mechanisms appear likely to connect IP, production to different classes of receptors on the cell surface.
111. INTERACTION OF PROFlLlN WITH ACTIN
We have characterized the effects of three profilin isoforms on actin and find that all three have similar mechanisms and affinities (Table 11). Both of the Acanthamoeba profilins and platelet profilin bind to actin monomers with a
TABLE I1 Reactions Involving Phosphoinositides, Actin, Platelet Profilin, and ATP" PI metabolism PI + T = PIP + D PIP + T = PIP, + D PIP, = IP, + DAG
Variable enzyme reaction Variable enzyme reaction Variable enzyme reaction
k+ ( p M - I sec-1)
k(sec-I)
Profilin binding to lipids P + SPIP = P(PIP), P + SPIP, = P(PIP,),
-1
1
Profilin binding to actin P + AT = PAT
1
Actin nucleotide exchange A+T=AT A+D=AD PA + T = PAT
10 10
Actin filament elongation AT + BEND = A.BEND AT + PEND = A.PEND
10
Actin filament capping PAT + BEND = P.BEND P + BEND = P.BEND
Kd
(pM)
10
I
0.01 0.1 10
1 I
0.001 0.01
I 0. I 1 -10 -1
a PI, Phosphatidylinositol; PIP. phosphatidylinositol 4-phosphate; PIP,. phosphatidyl 4.5-bisphosphate; IP,, inositol I ,45trisphosphate; DAG. diacylglycerol: P. profilin; A, nucleotide-free actin monomer; T . ATP; D. ADP; BEND, the barbed end of actin filaments; PEND, pointed end of actin filaments. Order of magnitude estimates of the rate and equilibrium constants come from Machesky er a / . (1990) and Goldschmidt-Clermont ef a / . (1991b).
222
Thomas D. Pollard et al.
dissociation constant in the micromolar range. Profilin alters the polymerization properties of actin. Complexes of actin with profilin appear incapable of forming nuclei, thereby suppressing the already high unfavorable reactions that lead to the formation of dimers and of the trimeric nuclei that initiate the growth of longer actin filaments. Complexes of actin with profilin probably cannot bind to the pointed end of actin filaments and support elongation at that end. On the other hand, all three profilins are weak inhibitors of elongation at the barbed end. This led us to hypothesize that complexes can bind to the barbed end of the actin filament and, therefore, that profilin has some affinity for the barbed end of the actin filament (Pollard and Cooper, 1984). The affinity of profilin for the barbed end of the filament appears to be lower by at least an order of magnitude than its affinity for free actin monomers. Consequently, any profilin that binds on its own or as part of an actin profilin complex to the barbed end of the actin filament will rapidly dissociate from the filament. During the time that the profilin is associated with the end of the filament it may act as a capping protein, preventing subunit addition at that end. This mechanism explains how profilin can bind to actin monomers with a much higher affinity than expected from its relatively weak effects on the growth of the barbed end of the filament, and how actin filaments can grow even in the presence of high concentrations of profilin. Profilin also affects the exchange of nucleotide (Mockrin and Kom, 1980; Nishida, 1985) and divalent cations bound to actin monomers. Profilin acts catalytically, hopping on a subsecond time scale from one actin monomer to the next. When bound to an actin monomer, profilin increases the rate constant for nucleotide dissociation by a factor of 1000 (Goldschmidt-Clermont et al., 1991b). Since the dissociation of nucleotide is the rate-limiting step in the exchange of nucleotide, profilin can catalyze the exchange of nucleotide. Our analysis of divalent cation exchange is not compete, but the process may be similar to the nucleotide exchange and may, in fact, be linked to it.
N. W H A T DOES PROFlLlN ACTUALLY DO IN A CELL? Over the years, there have been many interesting schemes suggested to explain the roles of profilin in cells. The rapidly evolving information on the many associations and effects of profilin have made many of these preliminary suggestions obsolete and opened up many new regulatory opportunities for this small ubiquitous protein. Profilin null mutations in yeast show that profilin is essential for normal cellular function (Magdolen et al., 1988; Haarer er al., 1990). The yeast cells without profilin are extremely sick, do not grow at normal temperatures, and have large bizarre shapes. One way of assessing our insight into this regulator system is simply to write
13. Interactions of Profilins with Lipids
223
down all of the known reactions involving profilin, actin, and polyphosphoinositides (Table 11). This is a minimum list of reactions that need to be considered. Undoubtedly, there are additional reactions that we do not yet know about. It is encouraging that we already know the rate constants or the equilibrium constants for a number of these reactions. Given the concentrations of profilin, actin, and phosphoinositides in platelets, we can calculate that a large fraction of the PIP, and PIP would be bound to profilin in the absence of any other reactions. Some, but certainly not all, of the unpolymerized actin will be bound to the remaining profilin. These reactions alone cannot explain the entire pool of unpolymerized actin in cells. Consequently, other proteins, such as actobindin (Lambooy and Korn, 1986), the actophoriddepactidADF class (Cooper et al, 1986; Mabuchi, 1983; Bamburg et al., 1980) of actin monomerbinding proteins, and the recently discovered 5-kDa actin monomer-binding protein (Safer et al, 1990) must account for the rest of the unpolymerized actin. It is obvious from this list of reactions that changes in the concentrations of PIP and PIP, might have an important influence on the assembly of actin. Others have postulated that a rise in the concentration of PIP, during cell activation might release actin from profilin and provide a pool of monomers to elongate actin filaments (Lassing and Lindberg, 1988). The available data on phosphoinositide turnover in platelets (Wilson et al., 1985) are inconsistent with this idea. Although there is an increased flux of molecules through the system from PI to IP, and DAG, the concentrations of PIP and PIP, are either stable or fall slightly after cell activation. This raises an important and complex question of how a j u x through the PI-PIP-PIP, pathway affects the binding of profilin and therefore the availability of profilin for interaction with actin. If the formation of a profilin-PIP, complex is a slow multistep process, flux through the pathway might reduce profilin binding even if the bulk concentration of PIP and PIP, were relatively stable. Since profilin does not bind to PI, it might be possible in the future to reconstitute in vitro such a flux through the system with PIP kinase and PLC and to measure the binding of profilin under conditions of flux. There is evidence for transient formation of actin-profilin complexes immediately after activation of platelets (Lind et al., 1987) that could be caused by profilin dissociation from PIP and PIP, during the flux through the PI pathway. This transient association of profilin with actin might have a positive effect on actin assembly through its catalysis of nucleotide exchange on actin monomers. For example, if a part of the unpolymerized actin exists as ADP-actin, the release of profilin into the cytoplasm and its association with these ADP-actin monomers might catalyze the exchange of ADP for ATP on the actin subunits and thus promote actin assembly. However attractive these speculations might be, most of them await testing in biochemical reconstitutions and essentially all of them await testing in the context of a live cell.
224
Thomas D. Pollard et al.
References Bamburg, J. R., Harris, H. E., and Weeds, A. G. (1980). Partial purification and characterization of a depolymerizing factor from brain. FEBS Lerr. 121, 178--182. Cooper, J. A , , Blum, J. D., Williams, R. C . , Jr., and Pollard, T. D. (1986). Purification and characterization of actophorin, a new 15,000-dalton actin-binding protein from Acanrhamoeba casrellanii. J. Cell Biol. 261, 477-485. Goldschmidt-Clermont, P. J., Machesky, L. M., Baldassare, J. J., and Pollard, T. D. (1990). The actin-binding protein profilin binds to PIP2 and inhibits its hydrolysis by phospholipase C. Science 247, 1575-1578. Goldschmidt-Clermont, P. J., Kim, J. W., Machesky, L. M., Rhee, S. G., and Pollard, T. D. (1991a). Regulation of phospholipase C-y- 1 by profilin and tyrosine phosphorylation. Science 251, 1231-1233. Goldschmidt-Clermont, P. J., Machesky, L. M., Doberstein, S . K., and Pollard, T. D. (1991b). Mechanism of the interaction of human platelet profilin with actin. J . Cell B i d . 113, 10811089. Haarer, B. K., Lillie, S . H., Adams, A. E. M., Magdolen, V., Bandlow, W., and Brown, S . S . (1990). Purification of profilin from Saccharomyces cerevisiae and analysis of profilin-deficient cells. J. Cell Biol. 110, 105-114. Hummel, J. P., and Dreyer, W. J. (1962). Measurement of protein-binding phenomena by gel filtration. Biochem. Biophys. Acfa 63, 530-532. Janmey, P. A., and Stossel, T. P. (1987). Modulation of gelsolin function by phosphatidylinositol 4,5-bisphosphate. Nature (London) 325, 3362-364. Lambooy, P. K., and Korn, E. D. (1986). Purification and characterization of actobindin, a new actin monomer-binding protein from Acanrhamoeba casrellanii. J. Biol. Chem. 261, 71507155.
Lassing, I., and Lindberg, U. (1985). Specific interaction between phosphatidylinositol 4 3 bisphosphate and the profilactin. Nature (London) 314, 472-474. Lassing, I., and Lindberg, U. (1988). Specificity of the interaction between phosphatidylinositol4,5bisphosphate and the profi1in:actin complex. J. Cell Biochem. 37, 255-267. 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, 833842. Mabuchi, 1. (1983). An actin-depolymerizing protein (depactin) from starfish oocytes: Properties and interaction with actin. J. Cell Biol. 97, 1612-1621. Machesky, L.. Goldschmidt-Clermont, P. J., and Pollard, T. D. (1990). The affinities of human platelet and Acanrhamoeba profilin isoforms for polyphosphoinositides account for their relative abilities to inhibit phospholipase C. Cell Regul. 1, 937-950. Magdolen, V., Oeschner, U.,Muller, G . , and Bandlow, W. (1988). The intron containing gene for yeast profilin (PFY) encodes a vital function. Mol. Cell Biol. 8, 5108-5115. McLaughlin, S. (1989). The electrostatic properties of membranes. Annu. Rev. Biophys. Biophys. Chem. 18, 113. Mockrin, S. C., and Korn, E. D. (1980). Acanrharnoeba profilin interacts with G-actin to increase the exchange of actin bound adenosine 5’-triphosphate. Biochemistry 19, 5359-5362. Nishida, E. (1985). Opposite effects of cofilin and profilin from porcine brain on rate of exchange of actin-bound adenosine 5’-triphosphate. Biochemistry 24, 1160- 1164. Pollard, T. D., and Cooper, J. A. (1984). Quantitative analysis of the effect of Acanrhamoeba profilin on actin filament nucleation and elongation. Biochemistry 23, 6631-6641, Rhee, S. G., Suh, P.-G., Ryu, S.-H., and Lee, S. Y. (1989). Studies of inositol phospholipid-specific phospholipase C. Science 244, 546-550.
13. Interactions of Profilins with Lipids
225
Safer, D., Golla, R., and Nachmias, V. (1990). Isolation of a 5-kilodalton actin-sequesteringpeptide from human blood platelets. Proc. Nurl. Acud. Sci. U.S.A. 87, 2536-2540. Smrcka, A. V., Hepler, J. R . , Brown, K . 0..and Stemweis, P. C. (1991). Regulation of polyphosphoinositide-specific phospholipase C activity by purified Gq. Science 251, 804-807. Wilson, D. B., Neufeld, E. I., and Majerus, P. W. (1985). Phosphoinositide interconversion in thrombin-stimulated human platelets. J . B i d . Chem. 260, 1046.
This Page Intentionally Left Blank
CHAPTER 14
Polarized Assembly of Spectrin a n d Ankyrin in Epithelial Cells Jon S. Morrow, Carol D. Cianci, Scott P. Kennedy, and Stephen L. Warren Department of Pathology Yale University School of Medicine New Haven, Connecticut 06510
I. Introduction: The Red Cell Paradigm 11. The Unexpected Spatial Polarization of the Nonerythroid Spectrin Cytoskeleton 111. The Interaction of Spectrin and Ankyrin with Basolateral Na+/K+-ATPase IV. Does Polarized Membrane Skeletal Assembly Require Ankyrin? V. Models for the Role of Nonerythroid Spectrin-Actin Cytoskeleton References
1. INTRODUCTION: THE RED CELL PARADIGM
Our detailed understanding of the spectrin cytoskeleton in the mammalian erythrocyte is the result of two decades of concerted effort by many laboratories, and stands as a triumph of modem molecular cell biology. Composed predominantly of spectrin, actin, protein 4.1, protein 4.9, adducin, tropomyosin, and ankyrin, the approximate disposition and role of each of these proteins in the red cell skeleton is now comprehensible (Fig. 1) (for reviews, see Bennett, 1989; Marchesi, 1985; Coleman et al., 1989; Morrow, 1989). The predominant component is spectrin, which is an antiparallel heterodimer of two subunits (a,P), with calculated molecular masses of 280,000 Da (Sahr et al., 1990) and 246,000 Da (Winkelmann et al., 1990), respectively. Spectrin heterodimers self-associate through a to P subunit interactions at the amino terminus of the a-subunit (Fig. 2) (Morrow et al., 1980), and bind F-actin by an interaction near the amino terminus of the P-subunit (Becker el al., 1990; Karinch et al., 1990). These interactions allow the formatian of a stoichiometric and approximately penCurrent Topics in Membranes, Volume 38 Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.
227
228
Jon S. Morrow et al.
FIG. 1 Model of the red cell spectrin membrane skeleton. The spectrin cytoskeleton in mammalian erythrocyte is composed of an anastomosing, approximately planar and uniform array of spectrin tetramers and oligomers cross-linking short filaments of F-actin. The major point of attachment of this array to the plasma membrane is via an ankyrin-mediated linkage to the cytoplasmic domain of band 3. Secondary attachments form between protein 4.1 and glycoproteins (GP). Other accessory proteins act to stabilize and order the short actin filaments (e.g., adducin, protein 4.9, and tropomyosin).
FIG. 2 Schematic structure of the spectrin heterodimer. Each subunit is composed of 106 residue repeating units. The degree of sequence identify between repeats is 10 to 30%. Areas of specialization (shaded) are marked by nonrepetitive structure. The sites of protein binding are as indicated. For erythrocyte spectrin, calmodulin binds near the amino terminus of P-spectrin (Anderson and Morrow, 1987). For the nonerythroid spectrins, calmodulin binds in a region of nonrepetitive sequence in the eleventh unit of a-spectrin (Harris er al.. 1988). The region of homology to the SH3 domain of the tyrosine kinases is also indicated (Wasenius et al, 1989).
14.
Assembly of Spectrin and Ankyrin in Cells
229
tagonal to hexagonal planar array that is linked by ankyrin to the cytoplasmic domain of band 3, and by protein 4.1 and possibly other proteins to the cytoplasmic domains of band 3 and other transmembrane glycoproteins. Spectrin also binds a number of accessory proteins which modify its interactions with actin or the bilayer, and all interactions within the spectrin skeleton appear subject to several levels of posttranslational control (see above reviews, also Mische and Morrow, 1988). The major role of the spectrin skeleton in erythrocytes is to enhance the structural stability and deformability of the plasma membrane. Many observations support this notion. When spectrin is extracted from erythrocytes by lowionic-strength buffer, the plasma membrane vesiculates and fragments. Inbred strains of mice with deficiencies in erythrocyte spectrin suffer profound and lethal hemolysis (Bodine et al., 1984), and, as reviewed in chapter 9 of this volume, single point-mutations in spectrin that affect its ability to self-associate or bind actin lead to inherited hemolytic disease (Marchesi, this volume). In contrast, there is much less evidence that erythrocyte spectrin plays a significant role in the organization of transmembrane proteins. The distribution of both the spectrin skeleton and intramembranous particles is uniform in the mature erythrocyte. Except for a small fraction (10-20%) that is directly linked to spectrin via ankyrin (Bennett and Stenbuck, 1979), band 3 is only minimally constrained by the spectrin-actin meshwork (Fowler and Branton, 1977) and by the association state of spectrin (Tsuji and Ohnishi, 1986). Perhaps reflecting this limited role of the spectrin skeleton in erythrocytes, erythroid spectrins are the most divergent members of the spectrin protein family. Compared to their nonerythroid analogs, the erythroid spectrins display the greatest degree of evolutionary drift (Winkelmann et al., 1990; Sahr et al., 1990; Moon and McMahon, 1990).
II. THE UNEXPECTED SPATIAL POLARIZATION OF THE N O N ERYTH ROI D SPECTRIN CYTOSKELETON With the recognition of a spectrin cytoskeleton in nonerythroid cells have come many unexpected observations. Collectively, these suggest that the primary role of nonerythroid spectrin is not global structural support of the membrane. Perhaps the most striking of these observations is the temporal and spatial polarization exhibited by the nonerythroid spectrins in a variety of cells. For example, nonerythroid spectrin can undergo patching and capping, without apparent injury or instability of the plasma membrane (Levine and Willard, 1983). In some cells, it may even exist as an intracellular pool which is recruited to the membrane after receptor activation (Black et al., 1988; Lee et al., 1988). Other studies demonstrate no membrane instability even after the microinjection of anti-spectrin antibodies, which result in a clearing of most spectrin from the plasma membrane
230
Jon S. Morrow et al.
FIG 3 Immunofluorescent analysis of the distribution of nonerythroid spectrin in MDCK cells during the development of tight cell-cell contact. MDCK cells (type I) were subcloned and then grown on glass coverslips in normal medium (a-e) or in medium containing only 5 phf Ca2+ (f). Cells were harvested at the single cell (a, b), small colony (c, d), and confluent monolayer (e, f ) stages. Prior to indirect immunofluorescent staining with affinity-purified anti-human brain spectrin antibodies, the cells in b and d were extracted with 0.5% Triton X-100.All cells were then fixed for 15 min at 4°C in 1.75% buffered formalin. Final magnification of the negative was 200X.
231
14. Assembly of Spectrin and Ankyrin in Cells
(Mangeat and Burridge, 1984). When the various isoforms of spectrin and its associated proteins are considered, perhaps the most dramatic examples of polarization of the spectrin cytoskeleton occur in epithelial cells and the nervous system. Unique isoforms of spectrin and/or ankyrin are differentially distributed within neurons (Lazarides and Nelson, 1983; Zagon et al., 1986); at the motor end plate of skeletal muscle (Bloch and Morrow, 1989), at the nodes of Ranvier (Kordeli et al., 1990); with the voltage-gated sodium channel (Srinivasan et al., 1988); and at the brush border of avian enterocytes. In kidney cells, spectrin and certain isoforms of ankyrin are confined to the basolateral margins of the cell (Koob et al., 1987; Nelson and Veshnock, 1987; Morrow, 1989; Davis et al., 1989). Significantly, in cultured MDCK cell monolayers, this polarized distribution of spectrin and ankyrin at the basolateral margin is only attained after productive cell-cell contact is established (Fig. 3) (Morrow, 1989; Nelson and Veshnock, 1986). Thus, while the spectrin skeleton in nonerythroid cells is clearly associated with certain classes of membrane receptors, it seems unlikely that its role is simply one of supporting the bilayer. The challenge of many current investigations is to unravel the factors that activate and direct the polarized assembly of the spectrin cytoskeleton, and to understand its meaning. 111. THE INTERACTION OF SPECTRIN A N D ANKYRIN W I T H BASOLATERAL N A + /K+-ATPase
The identification of isoforms of spectrin and ankyrin in a pattern coincident with that of basolaterally polarized integral membrane proteins such as Na /K -ATPase in kidney cells immediately suggested that these proteins were directly linked. If this linkage were similar to the linkage of band 3 to spectrin in the red cell, it would require that Na+/K+-ATPase bind ankyrin. Such binding has now been independently demonstrated by several laboratories using complimentary techniques, beginning with the pioneering studies of Nelson and Veshnock (1987) (also see Koob et al., 1987; Morrow, 1989; Davis et al., 1989). Kidney membranes enriched for Na /K -ATPase are able to bind erythrocytetype ankyrin in a saturable fashion with good affinity and with a meaningful stoichiometry of one ankyrin molecule for about every four Na /K -ATPase molecules in the membrane (Fig. 4). The actual affinity of ankyrin for Na+ / K + ATPase appears to be somewhat variable, depending on the state of ankyrin and possibly the ATPase as well. Interestingly, even mild proteolysis of ankyrin appears to enhance its affinity for kidney membranes and alter is specificity, suggesting an important role for a regulator domain in ankyrin that modulates this function (Davis et al., 1989.) The interaction of ankyrin with Na /K -ATPase appears to involve primarily the a-subunit of Na+/K+-ATPase, as judged from the ability of lZ51-labeled +
+
+
+
+
+
+
+
300
.....A . ............
........... . .I
200
100
r
0
1.o
0.0
2.0
3.0
free ankyrin (JIM)
b 200 - n
100
0
0
100
200
300
400
bound (pg/mg) FIG. 4 Dog kidney membranes enriched in Na+ ,K+-ATPase bind human erythrocyte ankyrin avidly. (a) The binding of 1251-labeled ankyrin to either stripped human erythrocyte membranes (A) or to kidney membranes ( 0 , 0 ) was measured by cosedimentation assay at pH 7.6. The best and worst of five dog kidney membrane preparations are shown. The binding isofoms were fitted by nonlinear regression analysis to a single class of binding sites. The apparent Kd for ankyrin binding to erythrocyte membranes was 0.1 M, with a Em,, of one molecule of ankyrin for every three molecules of band 3. The apparent Kd for the binding of ankyrin to Na+ ,K+-ATPase membranes was 2.5 pkf, with a Em,, of one molecule of ankyrin for every three to five molecules of Na+ ,K + ATPase. (b) Data in a were replotted according to Scatchard. Although by this analysis there is a hint of nonlinearity, which could be interpreted as indicative of a low-capacity but high-affinity binding site, rigorous error analysis indicates that this apparent deviation is not significant in these experiments (see Leatherbarrow, 1990). Adapted from Morrow (1989) with permission.
233
14. Assembly of Spectrin and Ankyrin in Cells
FIG. 5 Areas of homology between the sequences of Na+ IK + -ATPase a subunits and band 3 at the putative band 3-ankyrin-binding site. Identity, dark shading; homology, light shading. HEB3, human erythrocyte hand 3; HKB3, human kidney band 3; Sheep A l , Rat A l , A2,and A3 are (INa+ ,K+-ATPase isoforms.
ankyrin to react specifically with this subunit in solid-phase binding assays (Morrow, 1989). Inhibition studies also indicate that the binding of ankyrin to Na /K -ATPase is blocked by the cytoplasmic domain of erythrocyte band 3, suggesting that the sites of binding of both band 3 and ATPase in ankyrin must be physically close if not identical. A comparison of the sequences of band 3 and Na /K -ATPase reveals only one region in the cytoplasmic domain with any significant shared homology. The region involved is residues 646-665 of Na /K -ATPase (Morrow, 1989) (Fig. 5). Interestingly, this region is also highly conserved in other isoforms of a-Na+ IK -ATPase (Schneider et al., 1985). Studies using antibodies and synthetic peptides to block binding indicate that sites near this region, and perhaps overlapping it, participate in the binding of ankyrin to band 3 (Davis et al., 1989), although a simple linear sequence alone is insufficient for binding. Collectively, these observations suggest the possibility that the aforementioned sequence (Fig. 5) must be presented in the context of a proper secondary and tertiary conformation in order for binding activity to be realized. A similar requirement that a unique linear seguence be presented in the context of a properly folded peptide unit has recently been found to characterize the spectrin-binding site for ankyrin (Kennedy er al., 1991; also see below). +
+
+
+
+
+
+
N. DOES POLARIZED MEMBRANE SKELETAL ASSEMBLY REQUIRE ANKYRIN? Given the codistribution of Na+ /K -ATPase, ankyrin, and spectrin in mature epithelia, and the demonstrated capacity of ankyrin to bind both N a + / K + ATPase and spectrin in vitro, an obvious and reasonable conclusion would be +
234
Jon S . Morrow et al.
that the polarized assembly of the spectrin skeleton at the membrane in epithelial cells is mediated by ankyrin linkages and guided to the basolateral domain by Na+ /K -ATPase (Nelson and Veshnock, 1987). However, several recent experiments challenge this view. When MDCK cells are examined for the distribution of erythroid-type ankyrin at various stages of cell density, as in Fig. 3, one finds the persistence of a large cytoplasmic pool of ankyrin that can be easily extracted with Triton X-100 even when most of the stainable spectrin is apparent in a relatively insoluble pool at the membrane surface. The differences between spectrin and ankyrin in cultured MDCK cells are most apparent at intermediate cell densities, as shown in Fig. 6. Early in the course of monolayer formation, as shown in Fig. 6, most of the spectrin is localized to regions of cell-cell contact, while ankyrin remains largely cytoplasmic. +
FIG. 6 Ankyrin remains largely cytoplasmic in partially confluent MDCK cells, while spectrin sorts to the regions of cell-cell contact. Sparse cultures of MDCK cells were seeded onto glass coverslips and allowed to grow to confluence. At various times after plating, samples were examined by indirect immunofluorescent microscopy for the distribution of nonerythroid spectrin (a) and erythroid-type ankyrin (b). An intermediate state of confluence is shown. At earlier time points, before cell-cell contact is established, both proteins are cytoplasmic in distribution (cf. Fig. 3). After a confluent monolayer has been well established, both proteins are distributed at the basolateral margin of the cell (cf. Fig. 3; also Morrow, 1989). Note that, at this intermediate state of confluence, ankyrin remains largely cytoplasmic, while the spectrin is largely confined to the membrane of the cell, particularly along the zones of cell-cell contact.
14. Assembly of Spectrin and Ankyrin in Cells
235
To explore more directly the temporal sequence of spectrin and ankyrin assembly in MDCK cells, advantage was taken of the ability of Ca2+ to regulate the formation of productive cell-cell contacts in cell monolayers and, subsequently, the establishment of cell polarity (e.g., see Nelson and Veshnock, 1986). MDCK cells are plated at high density in either normal Ca2+ containing cell culture medium or medium that contains only 1-5 pM (estimated) free Ca2+. This system is shown diagrammatically in Fig. 7. At time zero, Ca2+ is then returned to the medium, and the progress of polarized cytoskeletal assembly is monitored at various times by examining the distribution, Triton extractability, and metabolic stability of spectrin and ankyrin. Using this system, the changes in extractability and half-life of nonerythroid spectrin and erythroid-type ankyrin observed in MDCK cells as they approach confluence mimic the immunofluorescence results. Although the establishment of confluence leads to an early stabilization of nonerythroid spectrin, this is not the case for ankyrin (Fig. 8). Experiments show the half-life of these proteins, using pulse labeling with [35S]methionine,also indicate that spectrin is stabilized before ankyrin (data not shown). The kinetic studies described above suggest that spectrin sorted to the lateral membrane of MDCK cells and was stabilized there independently of ankyrin.
FIG. 7 Protocol for comparing the rates of assembly of spectrin and ankyrin in MDCK cells. Cells are plated at high density in medium containing 1-5 phf Ca2+. At time zero, cells are returned to normal Ca2+ medium, after which the production of effective cell-cell contacts and polarization of the cell (and its cytoskeleton) are markedly accelerated. Cells maintained in the minimal Ca2+ medium are also examined at various times as controls. The use of an initial plating at high density to achieve "instant confluence" is an experimental manipulation developed by Nelson and Veshnock (1986).
236
Jon S. Morrow er al. 30
"High Shear" Extraction Conditions
-.-
0
10
18
Tlme In N o m l Caklum Contalnlng Medlum (Hours)
B
e
r
FIG. 8 The establishment of tight Ca*+-mediated cell-cell contacts in confluent MDCK cells leads to the stabilization of nonerythroid spectrin (B) but not ankyrin (.).(A) The extractability of both proteins with 0.5% Triton X-100was compared as function of time in confluent monolayers of MDCK cells plated in the presence of low Ca2 (1-5 @ and then I) grown in the presence of normal medium (approx. 1-2 mM Ca2+) (cf. Fig. 7). Note that Ca*+ caused a many-fold stabilization of spectrin, but not ankyrin, after 18 hr in culture. (B) lmmunoblots of erythroid-type ankyrin (a) and nonerythroid spectrin (f) in the Triton-soluble extract (e) and the Triton-insoluble residue (r) after vigorous extraction. Note that after 18 hr, over half of the nonerythroid spectrin is insoluble, while most of the ankyrin remains soluble. The apparent differences in the intensity of the ankyrin bank in the early time points compared to the 18-hr point are a technique artifact, and do not reflect a significant change in the abundance of ankyrin over this time period. +
However, these studies could not exclude the participation of a small pool of ankyrin in this process, which would not be detectable against a large background of unstabilized ankyrin. As a way to address this concern directly, the ability of an ankyrin-binding competent recombinant P-spectrin peptide which lacked most of the other functions of spectrin (e.g., actin binding, calmodulin
14. Assembly of Spectrin and Ankyrin in Cells
237
binding, protein 4.1 binding) to sort to the plasma membrane of MDCK cells was examined. Previously, we had identified the ankyrin-binding domain of P-spectrin and P-nonerythroid spectrin as residing in the fifteenth repetitive unit, residues 1768-1898 (Kennedy et al., 1991). A recombinant protein produced from the P-28 clone of erythroid spectrin (Winkelmann et al., 1990) displayed full ankyrin-binding activity in vitro (Kennedy et al., 1991). This clone, which encoded approximately half of the (3-spectrin subunit (residues 1397 to 2 137), was expressed in MDCK cells using a retroviral vector called pHIPPO-42 (Tn5 neo-r gene), which contains a unique BamHI restriction site downstream of the 3' splice site for insertion of a BglII-BamHI P-28 spectrin restriction fragment (Warren et al., 1988). The coding sequence in this vector is under the transcriptional control of the Moloney long-terminal repeat promoter/enhancer, and the vector contains a dominant selectable marker gene (neo-r) under the control of the internal SV40 early region promoter/enhancer. MDCK cells infected with this construct expressed a normal amount of endogenous native nonerythroid spectrin, as well as abundant recombinant protein (Fig. 9). Several different stable infected MDCK cell lines were examined, each expressed a different amount of the recombinant protein. Antibodies raised to Cterminal sequences in erythroid P-spectrin which did not cross-react with nonerythroid P-spectrin allowed the distribution of both the recombinant protein and the wild-type endogenous P-spectrin to be individually monitored. The recombinant protein bound ankyrin avidly in vitro, and was expressed in high abundance and appeared relatively stable in MDCK cells. However, its pattern of distribution was dramatically different from that of native nonerythroid spectrin in confluent monolayers (Fig. 10). While endogenous spectrin assembled at the lateral margins of the cell in regions of cell-cell contact, the recombinant protein remained cytoplasmic and extractable, in a distribution similar to that previously noted for ankyrin. Thus, it may be that ankyrin-binding ability alone is insufficient to assure the sorting of spectrin to the basolateral domain of epithelial cells.
V. MODELS FOR THE ROLE OF THE NONERMHROID
SPECTRIN-ACTIN CMOSKELETON Many models can be envisioned which would explain the colocalization of spectrin, ankyrin, and Na+ / K + -ATPase in a polarized distribution in epithelial cells. From these models, one can also postulate testable hypotheses about the larger role of the spectrin skeleton in nonerythroid cells. Three of these models are summarized in Fig. 11. The data discussed above suggest that a passive model of the spectrin skeleton, whereby it assembles via ankyrin-mediated linkages just at sites where Na /K -ATPase has already been directed, is probably incorrect. Likewise, we find it relatively difficult to reconcile our results with the +
+
238
Jon S . Morrow et al.
a
0
a .-a A I a
a 2 d#. 0 I
00
cu
m
$ 5 :
a
MDCK Clones
b
MDCK cell clone
FIG. 9 The p-28 cDNA clone of erythroid p-spectrin was expressed in MDCK cells using a retroviral expression system. Several stable lines were established, each expressing a different level of recombinant protein. In virro studies have established that the recombinant protein produced from this clone is fully competent to bind erythroid ankyrin, the isoform believed to be found at the basolateral margin of MDCK cells. (a) Western blot analysis of several MDCK cell lines demonstrating the expression of the recombinant p-spectrin. (b) Quantitative Western blot analysis for native nonerythroid spectrin indicates that there is no significant effect of the recombinant P-spectrin on the level of wild-type spectrin expression.
14. Assembly of Spectrin and Ankyrin in Cells
239
FIG. 10 A recombinant p-spectrin peptide (p-28) which is competent for ankyrin binding does not sort to the plasma membrane of confluent MDCK cells. (a) Immunofluorescent micrograph demonstrating the distribution of endogenous native spectrin in MDCK cells expressing rec-p-28 spectrin. (b) Immunoflourescent micrograph of the same clone of MDCK cells stained for the distribution of rec-p-28 spectrin. Note that cytoplasmic distribution of the recombinant protein, under conditions in which the native protein (which contains all of the binding functions of spectrin, not just its ankyrin-binding ability) sorts to the lateral membranes.
model shown in Fig. 1 l B , in which spectrin-ankyrin-Na+ /K+-ATPase complexes assemble first on an internal membrane compartment, and then move collectively to the lateral membrane. We favor the model shown in Fig. 11C. In this scenario, spectrin would sort and bind to the lateral membrane (or other topographically defined regions) by processes independent of ankyrin. A likely candidate molecule which could guide the assembly of a nascent spectrin skeleton would be a homotypic cell-cell adhesion molecule such as LCAM (also known as E-cadherin or uvomorulin). Indeed, when one follows their assembly along the points of cell-cell contact, one finds near coincidence of LCAM and spectrin sorting (Fig. 12), and complexes containing LCAM (uvornorulin) and
240
Jon S . Morrow et al.
1-
Q
ankyrin
\ spectrin
BrVlnrVvl
C
FIG. 11 Three possible models of the way that spectrin and ankyrin may participate in the topographic assembly of Na+,K+-ATPase. (A) All assemble at the lateral membrane; ATPase is targeted by other mechanisms. (B) All assemble on internal vesicles; targeted by other mechanisms as a preassembled complex (Nelson and Hamerton, 1989). (C) L C A M assembles spectrin at lateral membrane. Ankyrin-ATPase complexes bind to spectrin via a high-affinity ankyrin-spectrin interaction, leading to topographic assembly. This latter pathway suggests that spectrin may play a general role in guiding the assembly of several proteins (for transport and signaling?) at topographically defined regions of the cell.
spectrin have been recently identified in MDCK cell extracts (Nelson et al., 1990). An important conceptual feature of this latter model is the idea that the nascent spectrin skeleton, topographically assembled by either direct or indirect LCAM interactions, becomes a “nucleating” center about which other proteins needed for efficient transport or signal transduction are gathered. Thus, the key and most fundamental role of the spectrin skeleton may be that of an organizer of membrane transport and signaling complexes (Morrow, 1989). Indeed, several previous reports, in addition to the information presented above, are consistent with this interpretation. In the avian erythrocyte, it is believed that band 3 (a transport protein) is recruited (via ankyrin) into a preassembled spectrin-actin lattice (Cox et al., 1987). Ankyrin-independent associations of spectrin with the plasma membrane also exist (e.g., Cox et al., 1987; Howe et al., 1985; Steiner and Bennett, 1988), suggesting that the spectrin-ankyrin interaction may not be primarily involved with the assembly of the spectrin skeleton at the membrane. A critical role of spectrin in the organization of membrane proteins and signaling complexes might also explain why this protein is rapidly recruited to the membrane in activated lymphocytes (Black et al., 1988) and why it has been so strongly conserved throughout evolution (e.g., see Wasenius et a / . , 1989; Sahr et al., 1990). In future studies, it will be important to test this hypothesis of the spectrin skeleton as a key organizer of transport and signal transduction complexes in the membrane by identifying other putative transport and/or signaling proteins that may associate with the spectrin cytoskeleton, and the mechanisms that control their binding.
14. Assembly of Spectrin and Ankyrin in Cells
241
FIG. 12 Immunofluorescent localization of (a) LCAM and (b) spectrin in cultured MDCK cells under sparse culture conditions at the earliest states of cell-cell contact. Note the localization of both proteins along the zone of cell-cell contact. Other skeletal proteins which seem to assemble simultaneously with the CAMS are 4.1 and adducin.
242
Jon S. Morrow et al.
Acknowledgments The authors with to thank Drs. J. Anderson (Yale), B. Forget (Yale), B. Cunningham (Rockerfeller), A. Hams (Genetics Institute), and J. Winkelmann (U. of Minnesota) for help with various cDNA probes and vectors, antibodies, and ideas. This research was supported by a grant from the March of Dimes Foundation (#I-982) (J.S.M), and by grants from the National Institutes of Health: POI-DK38979 (J.S.M), R01-HL28560 (J.S.M), and F32-DKO8434 (S.P.K).
References Becker, P. S., Schwartz, M., Morrow, J. S., and Lux, S . E. (1990). Radiolabel-transfer cross-linking demonstrates that protein 4.1 binds to the N-terminal region of beta spectrin and to actin in binary interactions. Eur. J. Biochem. 193, 827-836. Bennett, V. (1989). The spectrin-actin junction of erythrocyte membrane skeletons. Biochem. Biophys. Acta. 988, 107-121. Bennett, V., and Stenbuck, P. I. (1979). The membrane attachment protein for spectrin is associated with band 3 in human erythrocyte membranes. Nature (London) 280, 468-473. Black, J. D., Koury, S. T., Bankert, R. B., and Repasky, E. A. (1988). Heterogeneity in lymphocyte spectrin distribution: Ultrastructural identification of a new spectrin-rich cytoplasmic structure. J . Cell Biol. 106, 97- 109. Bloch, R. J., and Morrow, J. W. (1989). An unusual b-spectrin associated with clustered acetylcholine receptors. J. Cell Biol. 108, 481-493. Bodine, D. M., Birkenmeier, C. S., and Barker, J. E. (1984). Cell (Cambridge, Mass.) 37, 721729. Coleman, T. R., Fishkind, D. J., Mooseker, M. S., and Morrow, J. S. (1989). Functional diversity among spectrin isoforms. Cell Moiil. 12, 225-247. Cox, J. V., Stack, J. H., and Lazarides, E. (1987). Erythroid anion transporter assembly is mediated by a developmentally regulated recruitment onto a preassembled membrane cytoskeleton. J. Cell Biol. 105, 1405-1416. Davis, J., Davis, L., and Bennett, V. (1989). Heterogeneity in membrane binding sites of ankyrin: Brain ankyrin, erythrocyte ankyrin, and processed erythrocyte ankyrin associate with distinct sites in kidney microsomes. J. Biol. Chem. 264, 6417-6426. Fowler, V., and Branton, D. (1977). Lateral mobility of human erythrocyte integral membrane proteins. Naiure (London) 268, 23-26. Howe, C. L., Sacramone, L. M., Mooseker, M. S., and Morrow, J. S. (1985). Mechanisms of cytoskeletal regulation: Modulation of membrane affinity in avian brush border and erythrocyte spectrins. J . Cell Biol. 101, 1379-1385. Karinch, A. M., Zimmer, W. E., and Goodman, S . R. (1990). Theidentification and sequence of the actin-binding domain of human red blood cell beta-spectrin. J. Biol. Chem. 265, 11833-1 1840. Kennedy, S. P.,Warren, S. L., Forget, B., and Morrow, J. S. (1991). The ankyrin binding domain of erythroid and nonerythroid P-spectrin is in the lYh repetitive unit. J. Cell Biol., in press. Koob, R., Zimmermann, M., Schoner, W., and Drenckhahn, D. (1987). Colocalization and coprecipitation of ankyrin and Na+ ,K+-ATPase in kidney epthelial cells. Eur. J . CellBiol. 45, 230-237. Kordeli, E., Davis, J., Trapp, B., and Bennett, V. (1990). An isoform of ankyrin is localized at nodes of Ranvier in myelinated axons of central and peripheral nerves. J . Cell Biol. 110, 1341-1352. Lazarides, E., and Nelson, W. J. (1983). Erythocyte and brain forms of spectrin in cerebellum: Distinct membrane-cytoskeletal domains in neurons. Science 220, 1295- 1296. Leatherbarrow, R. J. (1990). Using linear and non-linear regression to fit biochemical data. Trends Biochem. Sci. 15, 455-458.
14. Assembly of Spectrin and Ankyrin in Cells
243
Lee, J. K., Black, J. D., Repasky, E. A , , Kubo, R. T., and Bankert, R. B. (1988). Activation induces a rapid reorganization of spectrin in lymphocytes. Cell (Cambridge. Mass.) 55, 807-8 16. Levine, J., and Willard, M. (1981). Fodrin: axonally transported polypeptides associated with the internal periphery of many cells. J . Cell Biol. 90, 631-643. Mangeat, P. H.,and Bunidge, K. (1984). Immunoprecipitation of nonerythroid spectrin within live cells following microinjection of specific antibodies: Relation to cytoskeletal structures. J. Cell B i d . 98, 1363-1377. Marchesi, V. T. (1985). Stabilizing infrastructure of cell membranes. Annu. Rev. Cell Biol. 1, 531561.
Mische, S. M., and Morrow, J. S. (1988). Post-translational regulation of the erythrocyte cortical cytoskeleton. Proloplasma 145, 167- 175. Moon, R. T., and McMahon, A. P. (1990). Generation of diversity in nonerythroid spectrins. J . Biol. Chem. 265, 4427-4433. Morrow, 1. S. (1989). The spectrin membrane skeleton: Emerging concepts. Curr. Opin. Cell Biol. 1, 23-29. Morrow, J. S., Speicher, D. W., Knowles, W. J., and Marchesi, V. T. (1980). Identification of functional domains of human erythrocyte spectrin. Proc. Natl. Acad. Sci. U.S.A. 77, 65926596. Morrow, I. S . , Cianci, C., Ardito, T., Mann, A., and Kashgarian, M. T. (1989). Ankyrin links fodrin to alpha Na/K ATPase in Madin-Darby canine kidney cells and in renal tubule cells. J. Cell Biol. 108, 455-465. Nelson, J. W., Shore, E. M., Wang, A. Z., and Hamerton, R. W. (1990). Identification of a membrane-cytoskeletal complex containing the cell adhesion molecule uvomorulin (Ecadherin), ankyrin, and fodrin in Madin-Darby canine kidney epithelial cells. J . Cell B i d . 10, 349-357. Nelson, W.I., 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- 1766. 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. Rodriques-Boulan, E., and Nelson, W. J. (1989). Morphogenesis of the polarized epithelial cell phenotype. Science 245, 718-725. Sahr, K. E . , Laurila, P., Kotula, L., Scarpa, A. L., Coupal, E., Leto, T. L., Linnenbach, A. J., Winkelmann, J. C., Speicher, D. W., Marchesi, V. T., Curtis, P. J., and Forget, B. G. (1990). The complete cDNA and polypeptide sequences of human erythroid a-spectrin. J. B i d . Chem. 265,4434-4443. Schneider, J. W., Mercer, R. W., Caplan, M., Emanuel, J. R., Sweadner, K. J., Benz, E. J., Jr., and Levenson, R. (1985). Molecular closing of rat brain Na,K-ATPase alpha-subunit cDNA. Proc. Natl. Acad. Sci. U.S.A. 82, 6357-6361. Srinivasan, Y.,Elmer, L., Davis, J., and Bennett, V. (1988). Ankyrin and spectrin associate with voltage-dependent sodium channels in brain. Nature (London) 333, 177- 180. Steiner, J. P., and Bennett, V. (1988). Ankyrin-independent membrane protein binding sites for brain and erythrocyte spectrin. J. Biol. Chem. 263, 14417-14425. Tsuji, A., and Ohnishi, S.-C. (1986). Restriction of the lateral motion of bank 3 in the erythrocyte membrane by the cytoskeletal network: Dependence on spectrin association state. Biochemistry 25, 6133-6139. Warren, S. L., Handel, L. M., and Nelson, W. J. (1988). Elevated expression of pp6oC-src alters a selective morphogenetic property of epithelial cells in vitro without a mitogenic effect. Mol. Cell Biol. 8, 632-646.
244
Jon S. Morrow et al.
Wasenius, V.-M. et a / . (1989). Primary structure of the brain a-spectrin. J . Cell Eiol. 108, 79-93. Winkelmann, J. C. era/. (1990). Full-length sequence of the cDNA for human erythroid P-spectrin. J . Eiol. Chem. 265, 11827-11832. Zagon, I. S . , Higbee, R . , Riederer, B. M . , andGoodman, S.R. (1986). Specific spectrin subtypes in mammalian brain: An Immunoelectron microscopic study. J . Neurosci. 6, 2977-2986.
Index A Acanrhamoeba castellanii myosin I actin-activated ATPase activity, 17- 19 discovery, 13- 15 heavy chain kinase substrate specificity, 21 heavy chain sequences, 14, 16-17 intracellular localization, 22-25 isoenzymes, 15- 16, 22-23 mechanochemical properties, 19-21 purification, 13- I4 regulation, 2 1-22 Actin cytoskeleton remodeling during embryogenesis (Drosophila), 83-84 dynamics in serum-starved 3T3 fibroblasts, 192- 194 filament cross-linking by brush border myosin 1, 37-40 microfilaments associated with zona occludens peripheral plaque-like condensations, 179 profilin-actin interaction, 22 1-222 regulation, multimode approach, 190-192 role in stress fiber movement and cell motility, 198-199 spectrin association to form two-dimensional meshworks, 66 subunits, flow towards distal cell regions, 192- 193 talin binding, 61 Actin-activated MgZ+-ATPases, Acanthamoeba myosin I, 17-19 a-Actinin consensus sequences, 210 EF-hands domains, 2 I 1-2 12 53-kDa fragment effects on actin cytoskeleton, 60-61 genetic manipulation, 210-21 I p, integrin binding, 59-61
in vivo function of, 212-214
isolation and characterization, 209 Actomyosin, regulation, multimode approach, 190-192 Adhesion plaques, see Focal adhesions Alkaline phosphatase, effect on glyceroltreated fibroblast contraction, 195 Amino acid sequences Acanthamoeba myosin I , 16-17, 21 Drosophila melanogaster kinesin, 5-6 spectrin, normal and mutant, 162 Animal models, Duchenne muscular dystrophy, 113-118, 144-148 biochemical and genetic homology, 1 18- 124 dystrophin-deficient cat, 126- 127 mdr mouse, 124-125 xmddog, 125-126, 134-143 Anion exchanger-binding domain ankyrin, 73-74 spectrin, 66-68 Ankyrins binding sites, 72-75 brain, 68-69 cellular localizations, 70 comparison with spectrin in MDCK cells, 234-237 89-kDa domain, 67-68 erythrocyte, 66-68 genes encoding, 7 I human erythrocyte, domain structure, 67-68 interaction with basolateral Na+/K+ATPase, 231-233 isoforms, 68-69 mRNA, alternative splicing, 71-72 mutants, in hereditary spherocytosis, 171 nomenclature for, 7 I polarized membrane skeletal assembly and, 233-237 sequence differences, 70-71 and spherocytosis, 171 structure, 66-68 245
246
Index
Antisense molecules, gene function disruption with, 103-106 Aphidicolin, effect on early embryogenesis (Drosophila), 84 Apical membrane, brush border myosin I function in, 51 Aspergillus bimC gene, kinesin-related proteins encoded by, 2, 6 ATP, requirements for fiber contraction, 195 ATPase activity, see also Mg+-ATPase; Na+IK+ -ATPase Acanrhamoeba myosin I , 17-19 brush border myosin I , 40-41, 46-47 role of calmodulin light chains, 46-47
B Barrier function, zona occludens, 176-177 Basic fibroblast growth factor, role in Duchenne muscular dystrophy histopathology, 143-144 Brush border myosin I actin binding properties, 37-40 actin filament cross-linking, 39-40 ATPase activity, 40-41 calmodulin light chains, 37, 45-47 functions, 51-53 heavy chain primary structure, 35-36 sequences, 17 mechanochemistry, 40-45 microvillar membrane interactions, 48-5 1 subunit composition, 33-34
C Caenorhabdiris elegans Linl2 and Glp-I, 67 Calcium ion and brush border myosin I-actin interaction, 37-38 and dystrophin-deficient muscle fiber pathophysiology, 130- I34 movement across microvillar membrane, brush border myosin I role in, 53 regulation of brush border myosin I ATPase activity, 46-47 requirements for fiber contraction, 195
role in brush border myosin I mechanochemical activity, 40-45 Calcium ionophore, effect on zona occludens function, 178 Calmodulin binding, brush border myosin I heavy chain, 36-37 Calmodulin-binding neck domains, in unconventional myosins, 37, 47-48 Calmodulin light chains, repressor role in brush border myosin I MgZ+-ATPase,4647 Cap42, effect on fiber transport, 194 Carboxy-terminal domains actin-binding site, Acanthamoeba myosin I, 19-20 brush border myosin I, 49 spectrin, 66-67 Cats, dystrophin-deficient, 126- 127 CDCIO, Schizosaccharomyces pombe, 67 Celia sprue, 180 Cell leakage, basic fibroblast growth factor release and, 144 Cell motility, role of actomyosin structures, 198-199 Cell shape, changes during early embryogenesis (Drosophila), 82-84 Cellularization, Drosophila during embryogenesis, 81 myosin I1 heavy chain role, 86 myosin role in, 86 Centripedal translocation, stress fibers in serum-starved 3T3 fibroblasts, 192-194, 198 Charge selectivity, zona occludens, 178 Clinical progression, Duchenne muscular dystrophy, 114, 134-143 Consensus sequences, actin-binding proteins, 210 Contractile vacuole membranes, Acanthamoeba myosin I localization, 23-25 Contraction, stress fibers in serum-starved 3T3 fibroblasts, 194, 198-199 regulation, 195- 197 Creatine kinase serum levels and X-linkage, 118 serum levels in Duchenne muscular dystrophy, 127- 128 Cross linking actin filaments by Acanrhamoeba myosin I, 19
Index
247
actin filaments by brush border myosin I, 39-40 Crypt abscesses, 181 Crypt cells, zona occludens structure, 177 CT-40,effect on stress fibers, 194, 196 Cyclic AMP, effect on zona occludens function, 178 Cytochalasin D, effect on zona occludens, 179 Cytochalasins, effect on early embryogenesis (Drosophila), 83-84 fiber formation and transport in serumstarved 3T3 fibroblasts, 193-194 serum-starved 3T3 fibroblasts, 196 Cytoskeleton actin a-actinin 53-kDa fragment effects, 60-61 remodeling during embryogenesis (Dmsophila), 83-84 brush border myosin I, 34 during Drosophila early embryogenesis myosin 11 heavy chain role, 85-88 myosin 11 light chain role, 88-89 native myosin I1 role, 84-85 remodeling, 83-84 spectrins, 89-92 link with zona occludens, 178-179 mutant proteins in hereditary elliptocytosis, 157- 169 in hereditary spherocytosis, 169-171 Xenopus, 102-107 nonerythroid spectnn, spatial polarization of, 229-23 1 retrograde translocation of structures, 198199 spectrin-actin, role of, 237-241 during Xenopus embryogenesis protein 4.1 role, 101-102 vimentin role, 102- 105
D Deformability index, red cell ghosts containing protein 4.1, 167-169 Deletion mutations, in Duchenne muscular dystrophy, 114 Dictyosielium disroideum a-actinin genetic manipulation actin-binding domain, 210-21 1
EF-hands domains, 21 1-212 in vivo function of, 212-214 calmodulin-binding domain, 37 F-actin cross-linking proteins in, 209-210 heavy chain sequence, 17 locomoting, myosin I distribution in, 23 microfilament system in, 208-209 mutant missing myosin I isozyme, 25 Disease states, zona occludens function in, 180- 181 DNase, effect on fiber transport, 194 Docking proteins, microvillar, for brush border myosin I, 50 Dogs, dystrophin-deficient, 125- 126, 134-143 Domain structure brush border myosin I, 34-35 human erythrocyte ankyrin, 67-68 kinesin heavy chain (Drosophila), 34-35 spectrin mutants, 157-166, 169-170 Drosophila melanogaster early embryogenesis actin cytoskeleton remodeling, 83-84 cellularization during, 81 gastrulation, 81-83 myosin 11 heavy chain role, 85-88 myosin I1 light chain role, 88-89 native myosin I1 role, 84-85 pole cell formation, 80-81 spectnn polypeptides and encoding genes, 89-92 kinesin cellular function, 6-7 electron microscopy, 3 heavy chain comparisons, 5-6 molecular genetic dissection, 5 motility in virro, 7-8 sequence analysis, 3-4 mechanochemical cycle, 8-9 mitosis-related functions, 6-7 structure, 2-6 ninaC gene, 32 Notch protein, 67 Duchenne muscular dystrophy animal models, 113-118, 144-148 biochemical and genetic homology, 118124 dystrophin-deficient cats, 126-127 mak mouse, 124- 125 xmd dog, 125-126, 134-143
248
Index
basic fibroblast growth factor role, 143-144 clinical progression, 114, 124-143 human, histopathological changes, 121-124, 134-143 membrane hypothesis of, 114-1 15, 129 phase I, 121, 127-134 leakage of plasma membrane, 127- I34 therapies, 144- 147 phase 11, 121, 134-143 progressive histopathology and clinical weakness, 134- 143 therapies, 147-148 role of Ca2+, 130-134 therapeutics, 144- 148 Duplication mutations, in Duchenne muscular dystrophy, 114 Dystrophin cellular function, 114-1 15 consensus sequences with actin-binding proteins, 210 protein subcellular organization, 1 16 replacement, 144- I48 Dystrophin deficiency in cats, 126 cellular manifestation, 128- I29 in humans phase I, 127-134 phase 11, 134-143 progressive histopathology, basic fibroblast growth factor role, 143-144 invariant features in human, dog, cat, and mouse, 142 ischemia role, 133-134 in the mdx mutation, 118 in the xmd mutation, 119-120
E E-cadherin, role in spectrin skeleton assembly, 239-240 EDTA Acanthamoeba myosin I ATPase activity, 14-15 brush border myosin I ATPase activity, 4041 EF-hands domains, a-actinin, 21 1-212 Electron microscopy brush border myosin I, 34-35
cell changes during embryogenesis (Drosophila), 80 kinesin, 3 triton-extracted cytoskeletons (normal erythrocytes), 156-157 Elliptocytosis protein 4.1 mutants, 167-169 spectrin mutants, 157- 166 Embryogenesis Drosophila actin cytoskeleton remodeling, 83-84 cellularization, 8 I gastrulation, 80-81 myosin I1 heavy chain role, 85-88 myosin I1 light chain role, 88-89 native myosin I1 role, 84-85 pole cell formation, 80-81 spectrin polypeptides and encoding genes, 89-92 Xenopus laevis protein 4.1 role, 101-102 vimentin role, 102-105 Enterocyte differentiation, brush border myosin function in, 51-53 Epithelial cells ankyrin interaction with basolateral NA /K -ATPase, 23 1-233 LCAM role in spectrin skeleton assembly, 239-240 nonerythroid spectrin-actin cytoskeleton, 237-241 nonerythroid spectrin cytoskeleton, spatial polarization of, 229-23 I requirements for polarized membrane skeletal assembly, 233-237 spectrin interaction with basolateral Na+/ K + - ATPase, 23 1-233 Erythrocytes, ankyrin sequence (human), 67-68 Escherichia coli a-actinin actin-binding domain, 210-21 1 EF-hands domains, 21 1-212 in vivo function, 212-214 Ethylenediaminetetraaceticacid, see EDTA +
+
F F-actin, Acanthamoeba myosin I ATPreversible binding to, 14-15
Index
249
Fiber studies quantitative indicators for, 188- 190 techniques for, 190 Fibroblasts, quiescent, actomyosin dynamics in, 188-190, 192-194 Fibrolast growth factor, role in Duchenne mus. cular dystrophy histopathology, 143- 144 Fluorescence microscopy, for actomyosin dynamics, 191 Fluorescence recovery after photobleaching, myosin mobility, 199 Fluorescent indicators, for myosin light chain phosphorylation, 200 Focal adhesions a-actinin-p, integrin interactions, 59-61 formation, 58-59 protein interactions in virro, 62 talin-actin interactions, 6 I Fragmin, effect on stress fibers, 196
G Gastrulation, Drosophila, 81 Gelation factor complement to a-actinin, 21 1-214 consensus sequences, 210 isolation and characterization, 209 Gelation factor-deficient mutants, 21 3 Gelsolin, effect on stress fibers, 194, 196 Gene activity, disruption at protein level (Xenopus), 103- 107 at RNA level (Xenopus), 103-106 Gene function, disruption at RNA and protein levels (Xenopus), 103- 107 Gene localization, spectrin polypeptides (Drosophila), 91-92 Gene mapping, protein 4. I , 167- 168 Genetic engineering actin-binding proteins in Dicryostelium a-actinin, 2 10-2 14 F-actin cross-linking proteins, 209-210 Dicryostelium mutant missing myosin I isozyme, 25 Genetic screening, for Duchenne dystrophy, 144- 145 Gliding actin filament assay, brush border myosin 1. 43-45
Glp-l , Caenorhabditis elegans, 67 Glucose, intestinal uptake, zona occludens role, 180 GP-140 linker protein, for brush border myosin 1. 50
H Heart muscle, insensitivity to dystrophin deficiency-induced necrosis, 132 Heavy chain sequences Acanrhamoeba myosin I , 14, 16-17, 21 brush border myosin I, 17, 35-36 Drosophila kinesin, 3-5 myosin I1 heavy chain (Drosophila), 8588 squid kinesin, 5-6 Hemolytic disease hereditary elliptocytosis, 157- 166 hereditary spherocytosis, 169- 17 1 protein 4.1 mutants, 167-169 spectrin mutants, 157-166 Hereditary elliptocytosis protein 4.1 mutants, 167-169 spectrin mutants, 157-166 Hereditary pyropoikilocytosis, 162 Hereditary spherocytosis, mutant cytoskeletal proteins in, 169-171 Histopathology Duchenne muscular dystrophy, 121 - 124 basic fibroblast growth factor role, 143I44 cycles of, 146 phase I1 in humans and dogs, 134-143 in the dystrophin-deficient cat, 126 in the xmd dog, 125-126, 134-143 Hypertrophy in the mdr mouse, 124-125 in phase I1 Duchenne muscular dystrophy, 134-143
1
Immunoelectron microscopy, Acanfhamoeba myosin 1 localization, 23-25 Immunofluorescence
Index Acanrhamoeba myosin I intracellular localization, 22-23 L-CAM localization in MDCK cells, 241 spectrin in MDCK cells expressing rec-p-28 spectrin, 237 Immunoprecipitation, myosin I1 heavy chains (Drosophila), 85 p, Integrin, binding to a-actinin, 59-61 Interferon-?, effect on zona occludens permeability, 18 I Intestinal epithelial tight junctions, see Zona occludens Intracellular events, effect on zona occludens, 178 Ischemia, role in dystrophin deficieny, 133134 Isoenzymes, Acanthamoeba myosin I , 15- 16, 22-23 Isoforms Acanthamoeba myosin I , 15- 16 ankyrins, 68-69
K Kinesin Aspergillus bimC gene, 2, 6 Drosophila cellular function, 6-7 electron microscopy, 3 heavy chain comparisons, 5-6 molecular genetic dissection, 5 motility in virro, 7-8 sequence analysis, 3-4 mechanochemical cycle, 8-9 mitosis-related functions, 6-7 structure, 2-6 ultrastructure, three-domain, 4 S.cerevisiae Kar3 gene, 2, 6 squid, homology with Drosophila kinesin, 5-6
L Latrunculin, effect on fiber transport, 194 LCAM, role in spectrin skeleton assembly, 239-240
LC-I regulatory light chain (Drosophila), 8889 LC-2 regulatory light chain (Drosophila), 8889 Leakage, basic fibroblast growth factor release and, 144 Light chains calmodulin, brush border myosin I, 45-47 myosin I1 (Drosophila), 88-89 phosphorylation, myosin I1 fluorescent indicators of, 200 stress fibers and, 195-196 Light microscopy for cell changes during embryogenesis (Drosophila), 80 for fiber studies, 190 Linl2, Caenorhabditis elegans, 67
M MDCK cells comparison of spectrin and ankyrin in, 234237 L-CAM immunofluorescent localization, 241 native spectrin expressing rec-p-28 spectrin, immunofluorescence, 237 temporal sequence of spectrin and ankyrin assembly in, 235-237 mdr mutation dystrophin deficiency in, 118 dystrophin replacement in, 144- 148 non-progressive histopathology, 124- 125 Mechanochemical properties Acanthamoeba myosin I, 19-21 brush border myosin I, 40-45 Drosophila kinesin, 8-9 Membrane hypothesis, of Duchenne muscular dystrophy, 114-1 15, 129 Membrane skeleton protein 4.1, see Protein 4.1 Membrane vesicles, movement along actin cables supported by Acanthamoeba myosin I, 19-20 Messenger RNA, ankyrin, alternative splicing, 7 1-72 MgZ+-ATPase Acanrhamoeba myosins I , 15, 17- 19 brush border myosin I, 40-41, 46-47
Index Mice, dystrophin-deficient, 124- 125 Microfilament systems, Dicryosrelium discoideum, 208-209 Microvillar membranes brush border myosin 1 interactions, 48-53 docking protein for brush border myosin I, 50
Mitosis, kinesin cellular functions, 6-7 Monoclonal antibodies, mAb CX-7, effect on brush border myosin motility, 44-45 Monolayers, Ta4, 180- I8 I Motility brush border myosin I, 40-45 in virro, kinesin heavy chain, 7-8 Muscle fibers hypertrophy, in the mdr mouse, 124-125 necrosis, dystrophin deficiency-induced, 130- 134 release of cytoplasmic enzymes from, 128129 Muscle wasting in human Duchenne muscular dystrophy, I 2 1- 124 in phase I1 Duchenne muscular dystrophy, 134-143 in the xmd dog, 125-126 Mutant proteins ankyrin, 171 in Duchenne muscular dystrophy, 114-1 18 gelation factor-deficient GA I . 1, 2 13 in hereditary elliptocytosis, 157-169 in hereditary spherocytosis, 169-171 microinjection into embryos, 107 myosin I isozyme (Dicryosrelium), 25 protein 4.1, 167-169 spectrin, 157-166, 169-171 VSLTAC-myc, 103-105 VLTAN49, 103-105 vimentin, 103-104 Xenopus, 102-107 Myoblast implantation, dystrophin replacement with, 145-147 MY02 gene unconventional myosin encoded by, 48 yeast, 33 Myopathies, X-linked with dystrophin deficiency, 118 Myosin genes, superfamily of, 32-33
subunits, flow towards distal cell regions, 192-193 zipper phenotype, 87-88 Myosin-coated beads, movement along actin cables supported by Acanrhamoeba myosin I, 19-20 Myosin I Acanrhamoeba actin-activated ATPase activity, 17- 19 discovery, 13- I5 heavy chain kinase substrate specificity, L1
heavy chain sequences, 14, 16-17 intracellular localization, 22-25 isoenzymes, 15- 16 isoforms, 15-16 mechanochemical properties, 19-21 purification, 13-14 regulation, 2 1-22 brush border actin binding properties, 37-40 actin filament cross-linking, 39-40 ATPase properties, 40-41 calmodulin light chains, 45-47 functions, 51-53 heavy chain, primary structure, 35-36 mechanochemistry, 40-45 microvillar membrane interactions, 48-5 1 subunit composition, 33-34 Dicfyosrelium (locomoting), distribution, 23 Myosin I1 and cell motility, 198-199 dynamics in serum-starved 3T3 fibroblasts, 192-194 heavy chain function, genetic analysis, 87 genetic analysis (Drosophila), 87 role in cytoskeleton development during embryogenesis (Drosophila), 85-88 sequence analysis (Drosophila), 85-88 light chain phosphorylation fluorescent indicators of, 200 stress fibers and, 195-196 polymerase chain reaction (Drosophila), 88-89 role in cytoskeleton development during embryogenesis (Drosophila), 88-89 SDS-PAGE (Drosophila), 87-88
Index
252 native, role in cytoskeleton development during embryogenesis (Drosophila), 84-85 regulation multimode approach to, 190- 192 and stress fiber movement. 198-199
N Na+ -glucose cotransporter, zona occludens permeability and, 180 Na+ /K+-ATPase ankyrin-binding site, 72-75 basolateral, interaction with spectrin and ankyrin, 23 1-233 Necrosis, dystrophin deficieny-induced, 130134 Neonatal screening, for Duchenne muscular dystrophy, 145 Neuromodulin, calmodulin-binding domain (bovine), 37 ninaC gene (Drosophila), 32 Nitella assays bead movement, brush border myosin I , 4243 in vitro motility, native myosin I1 (Drosophila), 85 Nomenclature, for ankyrins, 7 1 Nondenaturing gel electrophoresis, normal, hemolytic, and pyropoikilocytosis spectrin, I63 Notch protein, Drosophila, 67 Nutrients, effect on zona occludens permeability, 179-1 80
P p190 (chicken brain), calmodulin-binding domain, 37 Peptide mapping, spectrin mutants, 157- 166 Permeability, zona occludens, 176-177 Phalloidin, effect on early embryogenesis (Drosophila), 83-84 Phosphatidylinositol 4, 5-bisphosphate, human platelet prolifin binding, 217-219 Phosphoinositide signaling pathway, profilin effects, 219-220 Phosphoinositide-specific phospholipase C, inhibition by profilin, 219-221
Phospholipids acidic, interaction with brush border myosin I, 49 profilin binding, 2 17-2 19 Phosphorylation sites, Acanthamoeba castellanii myosin I heavy chain, 21 Photoactivatable molecules, for cytoskeletal regulation studies, 200-201 Plaque-like condensations, zona occludens, 179 Plasma membrane Acanthamoeba myosin 1 localization, 22-25 Dictyostelium myosin I localization, 23, 25 focal contacts, see Focal adhesions leakage in Duchenne muscular dystrophy, 127- I34 Plasmids, recombinant, injection into fertilized eggs, 106 Platelet profilin, see Profilin Point mutations, spectrin peptides, 164- 165 Polarization membrane skeletal assembly, ankyrin requirements, 233-237 nonerythroid spectrin cytoskeleton, 229-23 1 Pole cells, formation during early embryogenesis (Drosophila), 80-81 Polymerase chain reaction, myosin I1 light chains, 88-89 Polymorphisms, spectrin subunits, 158-160 Polymorphonuclear leukocytes, transmigration, zona occludens opening and, 181 Polypeptides, mutant, microinjection into embryos, 107 Potassium ion Acanthamoeba myosin I ATPase activity, 14-15 brush border myosin I ATPase activity, 4041 Primary structure ankyrin, 66-68 brush border myosin I heavy chain, 35-36 spectrin, 66-68 Profilin Acanthamorba, 219, 221 actin-profilin interaction, 221-222 binding to phospholipids, 217-219 effect on fiber transport, 194 inhibition of phosphoinositide-specificphospholipase C, 219-221 reactions involving, 221
Index
253
Protein 4. I deformability index of red cell ghosts, 167169 mutant forms in hereditary elliptocytosis, 167- 169 protein structure and partial gene map, 169 role in embryogenesis (Xenopus), 101-102 Protein kinase C, effect on zona occludens function, 178-179 Pyropoikilocytosis, hereditary, I62
R Recombinant plasmids encoding antisense genes, injection of, 103, 106 injection into fertilized eggs, 106 Redundancy concept, in protein function, 213 Retrograde translocation, cytoskeletal structures. 198- 199 Ribozymes for gene expression inhihition, I07 gene expression inhibition with, 107 RNA antisense, inhibition of protein 4.1 (Xtnopus), 102- 103 gene function disruption methods, 105- 107 vimentin, synthetic, 103-104 vimentin mutants, microjection, 103- 104
S Saccharoniyces cerevisiae KAR3 gene, kinesin-related proteins encoded by, 2, 6 SW14 and SW16, 67 Schizosaccharomyces pomhe CDC 10, 67 SDS-PAGE Acanthamoeba myosin I isoenzymes, 15- 16 myosin I1 light chains (Drosophila), 88-89 Sequence analyses ankyrins, 70-71 erythrocyte ankyrin (human), 67-68 kinesin heavy chain (Drosophila), 3-4 myosin I1 heavy chain (Drosophila), 8588 myosin 11 light chain (Drosophila), 88-89 spectrin isoforms (Drosophila), 91-92
Small intestine, crypt cell zona occludens structure, 177 SM- 1 pseudosubstrate peptide, effect on stress fiber contraction, 195 Sodium dodecyl sulfate-polyacrylamide gel electrophoresis, see SDS-PAGE Spatial polarization, nonerythroid spectrin cytoskeleton, 229-23 I Spectrin actin binding domain, 209 alpha forms during Drosophila development, 90 ankyrin binding domain, 66-68 beta forms during Drosophila development, 9 1-92 consensus sequences with actin-binding proteins, 210 cytoskeleton structure, 227-228 hemolytic disease-associated truncated subunits, 166 heterodimer structure, 228 interaction with basolateral Na+/K+ATPase, 23 1-233 mutants detection and characterization, 157-166 in hereditary elliptocytosis, 157- 166 in hereditary spherocytosis, 169- 17 1 nonerythroid cytoskeleton models for role of, 237-241 spatial polarization, 229-231 polypeptides during Drosophila embryogenesis, 89-92 skeleton assembly, LCAM role, 239-240 subunits, 157 tetramer assembly into membrane-associated network, 66 Spectrin-actin cytoskeleton, nonerythroid, 235-237 Spherocytosis ankyrin mutant proteins in. 171 spectrin mutant proteins in, 169-171 Squid, kinesin, 5-6 Stoichiometry, brush border myosin I-actin interaction, 38 Stress fibers centripedal translocation, 192- 194. 198-199 contractile activity, 194 contraction regulation, 195-197 disruption, myosin light chain phosphorylation and, 195
Index dynamics in serum-starved 3T3 fibroblasts, 192-194 mechanisms of disassembly, 195 shortening induced by cytochalasin, 196I97 Substrate specificity, Acaniharnoeba myosin I heavy chain kinase, 21 Subunits actin, flow towards distal cell regions, 192193 brush border myosin I, composition, 33-34 myosin, flow towards distal cell regions, 192- 193 spectrin domain map, 157-166 truncated forms associated with hemolytic disease, 166 zona occludens, 178 Superprecipitation, cross-linked actomyosin complexes by Acanihamoeba myosin I, 19-20 Swiss 3T3 fibroblasts, serum-starved, actomyosin dynamics in, 188- 190, 192I94
T Talin, actin binding, 61 Temporal sequences, spectrin and ankyrin assembly in MDCK cells, 235-237 Therapeutics, Duchenne muscular dystrophy, 144-148 Tight junctions, see Zona occludens T,, monolayers, modeling intestinal epithelia with, 179-181 Toxin A, effect on zona occludens ion permeability, 18 I Translocation, retrograde, of cytoskeletal structures, 198-199 Tryptic mapping spectrin mutants in hereditary spherocytosis, 169-171 spectrin peptides, 157- 162
Uvomorulin, role in spectrin skeleton assembly, 239-240
V
MAC-myc mutant, 103-105 VAN49 mutant, 103-105 Video-enhanced contrast, for fiber studies, 190 Villus absorptive cells, zona occludens structure, 177 Villus tip, zona occludens regulation at, 180 Vimentin role in embryonic development (Xenopus), 102- 105 WAC-myc mutant, 103-105 VAN49 mutant, 103-105 Vitamin D-binding protein, effect on stress fibers. 194
W Wound repair, basic fibroblast growth factor and, 143-144
X Xenopus laevis disruption of gene function at RNA level, 102-107 embryogenesis, protein 4.1 role, 101-102 function of gene products, 105-107 X-linked myopathy, 118 xmd mutation, progressive histopathological changes in, 125-126
Y Yeast MY02 gene, 33 calmodulin binding, 37 unconventional myosin encoded by, 48
U Ultrastructure, Drosophila kinesin, 4
zipper myosin null phenotype, 87-88 Zona occludens
Index alteration of function, 178- 179 barrier function, 176-177 cytoskeletal link, 178- 179 function in model disease states, 180-181
255 intestinal, 179- 180 kisses of, 177 regulation of, 178- I80 structure-function correlations, 177- 178
This Page Intentionally Left Blank