CELLULAR DOMAINS

Edited by

IVAN R. NABI

A JOHN WILEY & SONS, INC. PUBLICATION

CELLULAR DOMAINS

CELLULAR DOMAINS

Edited by

IVAN R. NABI

A JOHN WILEY & SONS, INC. PUBLICATION

Copyright © 2011 by Wiley-Blackwell. All rights reserved

Published by John Wiley & Sons, Inc., Hoboken, New JerseyPublished simultaneously in Canada

No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http://www.wiley.com/go/permissions.

Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifi cally disclaim any implied warranties of merchantability or fi tness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profi t or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.

For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002.

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com.

Library of Congress Cataloging-in-Publication Data:

Cellular domains / edited by Ivan R. Nabi. p. ; cm. Includes bibliographical references and index. ISBN 978-0-470-59544-2 (cloth) 1. Cell membranes. I. Nabi, Ivan R. [DNLM: 1. Cell Membrane Structures. 2. Cell Physiological Phenomena. 3. Cytoplasmic Structures. QU 350] QH601.C435 2011 571.6'4–dc22 2010042298

Printed in Singapore

oBook ISBN: 9781118015759ePDF ISBN: 9781118015735ePub ISBN: 9781118013742

10 9 8 7 6 5 4 3 2 1

This book is dedicated to my wife Hakima and children Nessim and Zachary as well as to my parents Ruth and Jim who have supported me throughout my career. It is fi rst of all the work of the contributors, whom I thank enormously for their efforts. It is also the result of my own personal scientifi c journey that was shaped by my mentors, Avraham Raz and Enrique Rodriguez - Boulan, as well as by all the stimulating interactions I have enjoyed over the years with colleagues, collaborators, students and post - docs.

PREFACE ix

CONTRIBUTORS xi

PART I MEMBRANE DOMAINS

CHAPTER 1 CYTOSKELETON-INDUCED MESOSCALE DOMAINS 3

Ziya Kalay, Takahiro K. Fujiwara, and Akihiro Kusumi

CHAPTER 2 CLATHRIN-COATED PITS 23

James R. Thieman and Linton M. Traub

CHAPTER 3 CAVEOLAE 39

Dan Tse and Radu V. Stan

CHAPTER 4 LIPID RAFTS 61

Leonard J. Foster

CHAPTER 5 MODELING MEMBRANE DOMAINS 71

Daniel Coombs, Raibatak Das, and Jennifer S. Morrison

PART II ORGANELLAR DOMAINS

CHAPTER 6 MITOCHONDRIA 87

Michael Zick and Andreas S. Reichert

CHAPTER 7 THE ENDOPLASMIC RETICULUM 113

Jody Groenendyk and Marek Michalak

CHAPTER 8 THE GOLGI APPARATUS 133

James W. Dennis and Ivan R. Nabi

CONTENTS

CHAPTER 9 ENDOSOMES 147

Thierry Galvez and Marino Zerial

CHAPTER 10 LYSOSOMES AND PHAGOSOMES 165

Guillaume Goyette and Michel Desjardins

CHAPTER 11 ENDOPLASMIC RETICULUM JUNCTIONS 177

Jesse T. Chao and Christopher J.R. Loewen

PART III CYTOSKELETAL DOMAINS

CHAPTER 12 THE ACTIN CYTOSKELETON 197

Jonathan A. Kelber and Richard L. Klemke

CHAPTER 13 MICROVILLI 213

Florent Ubelmann, Sylvie Robine, and Daniel Louvard

CHAPTER 14 MICROTUBULES 229

Geoffrey O. Wasteneys and Bettina Lechner

CHAPTER 15 CILIA 245

Laura K. Hilton and Lynne M. Quarmby

CHAPTER 16 INTERMEDIATE FILAMENTS 267

Normand Marceau, Anne Loranger, Stéphane Gilbert, and François Bordeleau

PART IV ADHESIVE AND COMMUNICATING DOMAINS

CHAPTER 17 FOCAL ADHESIONS 285

Caitlin Tolbert and Keith Burridge

vii

viii CONTENTS

CHAPTER 18 THE ADHERENS JUNCTION 303

Christopher P. Toret and W. James Nelson

CHAPTER 19 SPECIALIZED INTERCELLULAR JUNCTIONS IN EPITHELIAL CELLS: THE TIGHT JUNCTION AND DESMOSOME 321

Keli Kolegraff, Porfi rio Nava, and Asma Nusrat

CHAPTER 20 GAP JUNCTIONS 339

Jared M. Churko and Dale W. Laird

PART V POLARIZED CELLULAR DOMAINS

CHAPTER 21 EPITHELIAL DOMAINS 351

Nancy Philp, Liora Shoshani, Marcelino Cereijido, and Enrique Rodriguez-Boulan

CHAPTER 22 NEURONAL DOMAINS 371

Jennifer S. Goldman and Timothy E. Kennedy

PART VI DOMAINS REGULATING GENE EXPRESSION

CHAPTER 23 NUCLEAR DOMAINS 393

Dale Corkery, Kendra L. Cann, and Graham Dellaire

CHAPTER 24 THE NUCLEAR PORE 415

Richard W. Wozniak, Christopher Ptak, and John D. Aitchison

CHAPTER 25 CYTOPLASMIC RNA DOMAINS 429

Henry Parker and Tom C. Hobman

INDEX 445

PREFACE

Cellular compartmentalization into and within organelles segregates biochemical reactions and increases local molecular concentrations, thereby promoting effi ciency of cellular processes. Within membranes, subdomains generate lateral heterogeneity that organizes the spatial distribution of glycoprotein receptors and membrane proximal effectors. Morphologically identifi able plasma membrane domains include not only clathrin - coated pits and caveolae but also lipid rafts that form a class of membrane domains that are poorly defi ned morphologically. Cellular organelles, such as mitochondria, the endoplasmic retic-ulum, the Golgi apparatus, endosomes, and lysosomes, also defi ne morphologically dis-tinct domains whose functionality depends, in large part, on the establishment of “ domains within domains. ” Cellular organization is determined by cytoskeletal elements, including the actin and microtubule cytoskeletons, that generate cell surface microvilli and cilia, respectively, as well as intermediate fi laments. Adhesive and communicating domains regulate interaction of the cell with the substrate through focal adhesions, as well as with other cells via adherens junctions, tight junctions, desmosomes, and gap junctions. The latter are particularly expressed in epithelial cells whose apical – basolateral polarization is critical to their transport function. Essentially, all cells are polarized, and the neuron rep-resents a prime example of how cellular polarization results in the formation of functional domains. Nuclear domains control genetic regulation and transcription, and nuclear – cytoplasmic exchange and transport is mediated by the nuclear pore that delivers RNA to cytoplasmic domains that regulate RNA translation and degradation. Molecular determi-nants of cellular domains therefore include essentially all molecular components of the cell, including DNA, RNA, proteins, lipid, and glycans. Defi ning domains and understand-ing the molecular basis of their formation is central to understanding cellular function.

Ivan R. Nabi

ix

CONTRIBUTORS

John D. Aitchison, PhD, Institute for Systems Biology, Seattle, WA 98103 - 8904

Fran ç ois Bordeleau, Centre de recherche en canc é rologie de l ’ Universit é Laval and Centre de Recherche du Centre Hospitalier de Qu é bec (CRCHUQ), Quebec City, Quebec, Canada

Keith Burridge, PhD, Department of Cell and Developmental Biology and Lineberger Cancer Center, University of North Carolina, Chapel Hill, NC 27599

Kendra L. Cann, Department of Pathology, Dalhousie University, Halifax, Nova Scotia, Canada

Marcelino Cereijido, MD, PhD, Center for Research and Advanced Studies, Department of Physiology, Biophysics and Neurosciences, Mexico City, Mexico

Jesse T. Chao, Department of Cellular and Physiological Sciences, University of British Columbia, Vancouver, British Columbia, Canada

Jared M. Churko, Department of Anatomy and Cell Biology, Dental Science Building, University of Western Ontario, London, Ontario, Canada

Daniel Coombs, PhD, Department of Mathematics and Institute of Applied Mathematics, University of British Columbia, Vancouver, British Columbia, Canada

Dale Corkery, Department of Pathology, Dalhousie University, Halifax, Nova Scotia, Canada

Raibatak Das, PhD, Department of Mathematics and Institute of Applied Mathematics, University of British Columbia, Vancouver, British Columbia, Canada

Graham Dellaire, PhD, Department of Pathology, Dalhousie University, Halifax, Nova Scotia, Canada

James W. Dennis, PhD, Samuel Lunenfeld Research Institute, Mount Sinai Hospital, Toronto, Ontario, Canada

Michel Desjardins, PhD, D é partement de pathologie et biologie cellulaire, Universit é de Montr é al, Montreal, Quebec, Canada

Leonard J. Foster, PhD, Centre for High - Throughput Biology and Department of Biochemistry & Molecular Biology, University of British Columbia, Vancouver, British Columbia, Canada

Takahiro K. Fujiwara, PhD, Center for Meso - Bio Single - Molecule Imaging (CeMI) and Institute for Integrated Cell - Material Sciences (iCeMS), Kyoto, Japan

Thierry Galvez, PhD, Max Planck Institute for Molecular Cell Biology and Genetics MPI - CBG, Dresden, Germany

St é phane Gilbert, PhD, Centre de Recherche en Canc é rologie de l ’ Universit é Laval and Centre de Recherche du Centre Hospitalier de Qu é bec (CRCHUQ), Quebec City, Quebec, Canada

Jennifer S. Goldman, Center for Neuronal Survival, Montreal Neurological Institute, McGill University, Montreal, Quebec, Canada

xi

xii CONTRIBUTORS

Guillaume Goyette, PhD, D é partement de pathologie et biologie cellulaire, Universit é de Montr é al, Montreal, Quebec, Canada

Jody Groenendyk, PhD, Department of Biochemistry, School of Molecular and Systems Medicine, University of Alberta, Edmonton, Alberta, Canada

Laura K. Hilton, Department of Molecular Biology & Biochemistry, Simon Fraser University, Burnaby, British Columbia, Canada

Tom C. Hobman, PhD, Department of Cell Biology, University of Alberta, Edmonton, Alberta, Canada

Ziya Kalay, PhD, Center for Meso - Bio Single - Molecule Imaging (CeMI), Institute for Integrated Cell - Material Sciences (iCeMS), Kyoto, Japan

Jonathan A. Kelber, PhD, UCSD School of Medicine, Department of Pathology and Moores Cancer Center, La Jolla, CA 92093 - 0612

Timothy E. Kennedy, PhD, Center for Neuronal Survival, Montreal Neurological Institute, McGill University, Montreal, Quebec, Canada

Richard L. Klemke, PhD, UCSD School of Medicine, Department of Pathology and Moores Cancer Center, La Jolla, CA 92093 - 0612

Keli Kolegraff, Department of Pathology and Laboratory Medicine, Emory University School of Medicine, Atlanta, GA 30322

Akihiro Kusumi, PhD, Center for Meso - Bio Single - Molecule Imaging (CeMI), Institute for Integrated Cell - Material Sciences (iCeMS) and Research Center for Nano Medical Engineering, Institute for Frontier Medical Sciences, Kyoto University, Kyoto, Japan

Dale W. Laird, PhD, Department of Anatomy and Cell Biology, Dental Science Building, University of Western Ontario, London, Ontario, Canada

Bettina Lechner, PhD, Department of Botany, University of British Columbia, Vancouver, British Columbia, Canada

Christopher J.R. Loewen, PhD, Department of Cellular and Physiological Sciences, University of British Columbia, Vancouver, British Columbia, Canada

Anne Loranger, PhD, Centre de recherche en canc é rologie de l ’ Universit é Laval and Centre de Recherche du Centre Hospitalier de Qu é bec (CRCHUQ), Quebec City, Quebec, Canada

Daniel Louvard, PhD, CNRS, Institut Curie, Paris, France

Normand Marceau, PhD, Centre de recherche en canc é rologie de l ’ Universit é Laval and Centre de Recherche du Centre Hospitalier de Qu é bec (CRCHUQ), Quebec City, Quebec, Canada

Marek Michalak, PhD, Department of Biochemistry, School of Molecular and Systems Medicine, University of Alberta, Edmonton, Alberta, Canada

Jennifer S. Morrison, Department of Mathematics and Institute of Applied Mathematics, University of British Columbia, Vancouver, British Columbia, Canada

Ivan R. Nabi, PhD, Department of Cellular and Physiological Sciences, Life Sciences Institute, University of British Columbia, British Columbia, Canada

Porfi rio Nava, Department of Pathology and Laboratory Medicine, Emory University School of Medicine, Atlanta, GA 30322

CONTRIBUTORS xiii

W. James Nelson, PhD, Department of Biology, Stanford University, Stanford, CA 94305

Asma Nusrat, PhD, Department of Pathology and Laboratory Medicine, Emory University School of Medicine, Atlanta, GA 30322

Henry Parker, PhD, Department of Cell Biology, University of Alberta, Edmonton, Alberta, Canada

Nancy Philp, PhD, Department of Pathology, Anatomy and Cell Biology, Jefferson Medical College, Thomas Jefferson University, Philadelphia, PA 19107

Christopher Ptak, PhD, Department of Cell Biology, University of Alberta, Edmonton, Alberta, Canada

Lynne M. Quarmby, PhD, Department of Molecular Biology & Biochemistry, Simon Fraser University, Burnaby, British Columbia, Canada

Andreas S. Reichert, PhD, CEF Makromolekulare Komplexe, Mitochondriale Biologie, Fachbereich Medizin, Goethe - Universit ä t Frankfurt am Main, Frankfurt am Main, Germany

Sylvie Robine, PhD, CNRS, Institut Curie, Paris, France

Enrique Rodriguez - Boulan, MD, Dyson Vision Research Institute, Departments of Ophthalmology and Cell Biology, Weill Medical College of Cornell University, New York, NY 10065

Liora Shoshani, PhD, Center for Research and Advanced Studies, Department of Physiology, Biophysics and Neurosciences, Mexico City, Mexico

Radu V. Stan, MD, Department of Pathology, Dartmouth Medical School, One Medical Center Drive, Lebanon, NH 03756

James R. Thieman, Department of Cell Biology, University of Pittsburgh School of Medicine, Pittsburgh, PA 15261

Caitlin Tolbert, Department of Cell and Developmental Biology, University of North Carolina, Chapel Hill, NC 27599

Christopher P. Toret, PhD, Department of Biology, Stanford University, Stanford, CA 94305

Linton M. Traub, PhD, Department of Cell Biology, University of Pittsburgh School of Medicine, Pittsburgh, PA 15261

Dan Tse, MD, Department of Pathology, Dartmouth Medical School, One Medical Center Drive, Lebanon, NH 03756

Florent Ubelmann, PhD, CNRS, Institut Curie, Paris, France

Geoffrey O. Wasteneys, PhD, Department of Botany, University of British Columbia, Vancouver, British Columbia, Canada

Richard W. Wozniak, PhD, Department of Cell Biology, University of Alberta, Edmonton, Alberta, Canada

Marino Zerial, PhD, Max Planck Institute for Molecular Cell Biology and Genetics MPI - CBG, Dresden, Germany

Michael Zick, Adolf - Butenandt - Institut f ü r Physiologische Chemie, Ludwig - Maximilians - Universit ä t M ü nchen, M ü nchen, Germany

PART I MEMBRANE DOMAINS

M EMBRANE BILAYERS form hydrophobic dividers within the aqueous environment that exists both within and without the cell. The plasma membrane surrounds the cell, segregating the intracellular milieu from the outside environ-ment. Domain organization of the plasma membrane is based not only on lipid - and protein - based interactions but also on the organization of the underlying actin cytoskeleton that impacts on molecular dynamics and function in the membrane (Chapter 1 ). Plasma membrane domains include not only clathrin - coated pits (Chapter 2 ), caveolae (Chapter 3 ), and less well - characterized lipid rafts (Chapters 4 and 5 ) but also cell - substrate adhesions (Chapter 17 ), cell – cell junctions (Chapters 18 – 20 ), and specialized polarized cellular domains (Chapters 13 , 15 , 21 , and 22 ). Molecular transport across the plasma membrane and exchange with the extracel-lular milieu is the key to cellular functionality. It is mediated in large part by endocytosis via clathrin - coated pits (Chapter 2 ) and also, as described in later sections, by Golgi secretion (Chapter 8 ), exosomes (Chapter 9 ), transporters (Chapter 21 ), and gap junctions (Chapter 20 ). How lipids organize domains is a subject of intense investigation that has focused primarily on the role of lipid rafts (Chapters 1 , 4 , and 5 ). The underlying cytoskeleton controls and contributes to molecular dynamics in the plane of the membrane and to downstream signaling (Chapter 1 ; also Chapter 12 ), and the tools available to study and model membrane domain organization in living cells (Chapters 1 and 5 ) represent key elements of future research. Importantly, membrane domain organization is relevant not only to the plasma membrane but also to membranes of intracellular organelles (Chapters 6 – 11 ).

Cellular Domains, First Edition. Edited by Ivan R. Nabi.© 2011 John Wiley & Sons, Inc. Published 2011 by John Wiley & Sons, Inc.

1

CHAPTER 1

DEFINITION

The cell is much more than just a bag of protein juice. Indeed, many mechanisms exist in a living cell to keep its contents well organized. One of the most important apparatuses that the cell utilizes to organize the cytoplasm is the cytoskeleton. Therefore, the infl uence of the cytoskeleton on the two - dimensional fl uid of the plasma membrane is an interesting subject. In this chapter, we will focus on this issue, and review the literature about the membrane domains delimited by the part of the actin - based cytoskeleton that is closely opposed to the cytoplasmic surface of the plasma membrane. This part of the cytoskeleton is referred to as the membrane skeleton . Due to its close association with the cytoplasmic surface of the plasma membrane, the membrane - skeleton meshwork directly infl uences the functions of the plasma membrane. As a consequence of the membrane - skeleton meshwork, the plasma membrane is effectively partitioned into mesoscale domains, or compartments, with sizes varying between 30 and 250 nm (with the exception of the larger 750 - nm domain in the doubly nested compartments in normal rat kidney [NRK] cells; Fujiwara et al. 2002 ).

We emphasize the characteristic size of these compartments, one to several hundred nanometers, that falls in the mesoscale, where collective dynamics of molecules play criti-cal roles. At this size scale, the number of molecules in the system is insuffi cient for thermodynamics to hold, but is still too big to be tractable for quantum mechanics. In the plasma membrane, there are three types of major mesoscale domains ( meso domains): (1) membrane compartments delineated by the actin - based membrane skeleton; (2) raft domains, where specifi c proteins, glycosphingolipids, and cholesterol are concentrated; and (3) the protein oligomer domains. In this chapter, we will concentrate on the membrane - skeleton - induced membrane compartments.

Membrane lipids and proteins are both temporarily trapped in these membrane compartments with residency times between 1 ms and 1 second (Kusumi et al. 2005 ). Namely, the two key functional elements of the membrane are both infl uenced by the membrane skeleton, and a growing number of fi ndings support the involvement of the membrane skeleton in many membrane processes (Gaidarov et al. 1999 ; Nakada et al. 2003 ; O ’ Connell et al. 2006 ; Lajoie et al. 2007 ; Chung et al. 2010 ; Treanor et al. 2010 ). In this respect, one of our main goals in this chapter is to provide an account of the various functions of these membrane - skeleton - based mesodomains.

CYTOSKELETON - INDUCED MESOSCALE DOMAINS

Ziya Kalay Takahiro K. Fujiwara

Akihiro Kusumi

Cellular Domains, First Edition. Edited by Ivan R. Nabi.© 2011 John Wiley & Sons, Inc. Published 2011 by John Wiley & Sons, Inc.

3

4 CHAPTER 1 CYTOSKELETON-INDUCED MESOSCALE DOMAINS

HISTORICAL PERSPECTIVE

One of the most important characteristics of the plasma membrane is that it is a two - dimensional liquid. The fl uid mosaic model proposed by Singer and Nicolson (1972) successfully accounted for many properties of the plasma membrane. In fact, this model is still widely believed to represent the basic structure of the plasma membrane (and intracellular membranes) of all living cells on earth. However, it fails to answer two basic questions, which have puzzled scientists for the past three decades:

1. Why do the membrane proteins and lipids diffuse faster in artifi cial membranes than in the cellular plasma membrane by a factor of ∼ 20 (ranging from 5 to 50) ? (Murase et al. 2004 )

2. How do molecular complexes become immobilized on the cell surface or diffuse at surprisingly lower rates, as compared with single molecules? (Kusumi et al. 2005 )

Early Fluorescence Recovery After Photobleaching ( FRAP ) Observations

Early measurements of diffusion coeffi cients using FRAP revealed signifi cantly slower diffusion in the plasma membrane, as compared with that in artifi cially reconstituted membranes, by a factor of ∼ 20 (ranging from 5 to 50) (summarized by Murase et al. 2004 ). Some of these fi ndings suggested that the discrepancy might be due to interactions between the cytoskeleton and the plasma membrane. In 1980, Sheetz et al. found that the diffusion coeffi cient of band 3 proteins in mouse erythrocyte mutants lacking the spectrin network is an order of magnitude higher than that in normal cells. In these cells, the membrane skeleton is formed by the spectrin network, and therefore, the results obtained with these mutant cells suggested that the membrane skeleton directly interferes with the diffusion of membrane proteins. For some other early examples of FRAP measurements that detected signifi cantly reduced diffusivity in the plasma membrane (but not rotational diffusion; Tsuji et al. 1988 ), see the following references: Axelrod et al. (1976) , Golan and Veatch (1980) , Chang et al. (1981) , Tsuji and Ohnishi (1986) , and Tsuji et al. (1988) .

Early Single - Molecule Observations: Membrane - Skeleton Fence Model for Transmembrane Proteins

The advent of single - molecule imaging methods, such as single - particle tracking (SPT) and single fl uorescent - molecule tracking (SFMT), allowed researchers to image and track single molecules in the plasma membrane of living cells. In 1994, by using SPT tech-niques, Sako and Kusumi observed that transferrin receptor and α 2 - macroglobulin recep-tor undergo a peculiar kind of motion, which is characterized by temporary confi nement of the molecule in a bounded region of the membrane, with an average area of 0.25 μ m 2 , interrupted by rare hops into adjacent, yet still temporarily confi ning, regions. This behav-ior was termed hop diffusion , as the membrane molecules seemed to be hopping between adjacent membrane compartments and diffusing freely within a compartment. Subsequent experiments showed that all of the transmembrane proteins in every cell type observed displayed hop diffusion as well.

As the single - molecule imaging technique became more popular, a wealth of data started to accumulate that needed to be properly analyzed. The single - particle trajectories that exhibit hop diffusion have been analyzed by fi tting appropriate models to them (see also Chapter 5 for further discussion of modeling single - particle trajectories). An early

HISTORICAL PERSPECTIVE 5

theoretical treatment by Powles et al. (1992) considered the diffusion of a particle trapped between partially permeable barriers arranged periodically in space. Here, exact mathe-matical expressions for some transport quantities in the system were obtained. In more recent treatments by Kenkre et al. (2008) and Kalay et al. (2008) , exact formulas for the time - dependent mean square displacement and diffusion coeffi cient for a particle moving in a similar compartmentalized space were acquired, and the effects of disordered fence strengths and compartment sizes were predicted. The results of these works have success-fully been used to deduce compartment sizes from single - particle trajectories recorded in the plasma membrane.

Based on these fi ndings, the membrane - skeleton fence model was proposed (Sako and Kusumi 1994 ; Kusumi and Sako 1996 ), in which the plasma membrane is effectively partitioned into mesoscale compartments. In this model, the membrane skeleton forms a meshwork near the cytoplasmic part of the plasma membrane with which the cytoplasmic domains of transmembrane proteins can interact. As a result of this interaction, transmem-brane proteins are temporarily confi ned in membrane compartments induced by the membrane - skeleton meshwork. Transmembrane proteins can hop between adjacent com-partments if the distance between the meshwork and the membrane becomes large enough, or if the meshwork temporarily and locally dissociates. See Figure 1.1 (top) for an illustra-tion of these ideas.

This new, compartmentalized view of the plasma membrane was further supported by an atomic force microscopy study of its cytoplasmic surface (Takeuchi et al. 1998 ), which produced similar estimates for the spectrin membrane - skeleton mesh size to those obtained by SPT of band 3 proteins in human erythrocyte ghosts (Tomishige et al. 1998 ).

Super - Speed Single - Molecule Imaging: Hop Diffusion of Phospholipids as well as Transmembrane Proteins and Anchored - Protein Picket Model

Another striking discovery was made in 2002. Fujiwara et al. (2002) demonstrated that even phospholipids, which are the most basic molecular species for membrane formation, undergo hop diffusion between compartments with sizes similar to those detected by protein hop diffusion. In previous single - molecule observations, images were obtained at video (30 frames/s) or slower rates. However, the characteristic time for lipid residency within a compartment is generally much shorter than the time between two consecutive image frames employed in those observations. Therefore, the detection of membrane compartments for lipids had to wait for the development of ultrafast methods for perform-ing SPT. Fujiwara et al. (2002) performed their measurements at a rate of 40,000 frames/s, the fastest single - molecule imaging ever made, and have increased the rate even further to 160,000 frames/s.

The discovery of lipid molecules undergoing membrane - skeleton - dependent hop diffusion in the outer leafl et of the plasma membrane was very surprising, since they only reach halfway through the membrane and lack the cytoplasmic domains of transmembrane proteins. To explain the hop diffusion of phospholipids in the outer leafl et, the anchored - protein picket model was proposed (Fujiwara et al. 2002 ; Kusumi et al. 2005 ), where lipids are envisaged to interact with the membrane skeleton indirectly through picket proteins that are attached to the membrane skeleton, as shown in Figure 1.1 (middle). Several dif-ferent mechanisms for the interaction between lipids and protein pickets were proposed. First, the picket can block the passage of a lipid molecule due to volume exclusion, causing steric hindrance. Second, the lipid molecules can be packed more in the immediate vicinity of the picket than the bulk membrane, and thus the free area available for a lipid molecule

6 CHAPTER 1 CYTOSKELETON-INDUCED MESOSCALE DOMAINS

Figure 1.1. Schematic illustrations of the membrane - skeleton fence and anchored - protein picket models, and the picket - induced slowdown effect. According to the membrane - skeleton fence model (top), transmembrane proteins with protrusions in the cytoplasm are temporarily confi ned in membrane - skeleton - based compartments formed by a dynamic meshwork of actin fi laments. Consequently, these proteins undergo hop diffusion, as illustrated by the color - coded trajectory. Lipids in the upper leafl et of the plasma membrane can also undergo hop diffusion even though they lack cytoplasmic domains. In the anchored - protein picket model (middle), the presence of pickets, proteins that are anchored to the boundaries of membrane - skeleton - based compartments, restricts the motion of all membrane molecules including lipids. As explained in the text, the presence of an immobile protein may locally reduce the diffusivity of lipids. This so - called picket - induced slowdown effect is illustrated in the bottom fi gure. Here, a lipid molecule in the close proximity of a picket, which we refer to as the low diffusivity domain, diffuses at a slower rate as compared with those far from the picket, as indicated by the smaller step size in the particle ’ s trajectory.

Anchored-protein picket model

Membrane-skeleton fence model

Picket-induced slowdown effect

Actin filament(membrane skeleton)

Anchored-protein picket

Actin filament Low diffusivity domain

Actin filament (membrane skeleton)

Transmembrane protein

START

START

Phospholipid

Anchored-protein pickets

Phospholipid

ELECTRON TOMOGRAPHY OF THE THREE-DIMENSIONAL STRUCTURE 7

to move into decreases, leading to diminishing mobility around the picket (Sperotto and Mouritsen 1991 ; Almeida et al. 1992 ). Third, lipid molecules can experience hydrody-namic slowing near an immobilized picket molecule, similar to a fl uid particle that moves close to the boundary of a container (Bussell et al. 1994, 1995 ; Dodd et al. 1995 ). Therefore, lipid diffusivity can be reduced in the vicinity of a picket protein for different reasons, a phenomenon we call the picket - induced slowdown effect . A schematic illustra-tion of the low diffusivity domain due to this effect is shown in Figure 1.1 (bottom). When many such immobilized picket proteins are aligned along the membrane - skeleton fence, the membrane molecules cannot easily pass through the compartment boundaries, and thus become temporarily confi ned within a compartment. Namely, in the anchored - protein picket model, the entire plasma membrane is partitioned into compartments by transmem-brane protein pickets, lining the membrane - skeleton fence. By performing Monte Carlo simulations that account for the reduction in free area, the picket density along the com-partment boundary necessary for reproducing the observed hop diffusion of lipids was estimated to be 20 – 30% (Fujiwara et al. 2002 ). Namely, the compartment boundaries do not have to be closed off by concentrating the transmembrane protein pickets there, but when only 1/5 – 1/3 of the boundary is occupied by picket proteins, it would be suffi cient to induce temporary trapping of lipids within a compartment.

It is important to note that the effect of anchored - protein pickets is not limited to that of mere protein crowding. Several studies have indicated that if the obstacles are mobile, then they do not lead to a signifi cant decrease in the diffusion coeffi cient of the rest of the mobile particles. Monte Carlo studies by Saxton (1987, 1990) , which did not consider hydrodynamic effects, demonstrated that mobile obstacles do not reduce the dif-fusivity as much as their immobile counterparts. Later on, Bussell et al. (1994, 1995) and Dodd et al. (1995) showed that the inclusion of hydrodynamic effects did not change this conclusion, provided that the lipids can be assumed to form a continuum. This assumption would not hold in cases where the protein diameter is comparable with the diameter of lipids (we will address this point later in this chapter), and one might consider the possibil-ity that the details of collective protein – lipid dynamics on the mesoscale may determine the properties of the system. However, since the diffusion coeffi cients of proteins and lipids within a compartment are similar to those measured in liposomes, reconstituted membranes, and membrane blebs (see fi gs. 1 and 2 in Fujiwara et al. 2002 and fi gure 8 in Murase et al. 2004 ), the immobilization of obstacles (pickets) seems to be a key factor in the picket - induced slowdown effect. In addition, the proteins do need not be continually anchored in order to function as effective pickets. If the proteins are immobilized for more than ∼ 10 μ s at a time, then this may suffi ce to produce the observed hindrance of lipid diffusion.

ELECTRON TOMOGRAPHY OF THE THREE - DIMENSIONAL STRUCTURE OF THE CYTOPLASMIC SURFACE OF THE PLASMA MEMBRANE

Perhaps the most direct evidence for the membrane - skeleton - based partitioning of the plasma membrane was obtained in 2006 by Morone et al. , who imaged the three - dimensional structure of the cytoplasmic surface of the plasma membrane by electron tomography. Here, electron tomography was fi rst applied to platinum - replicated samples: the three - dimensional structure of the cytoplasmic surface of the plasma membrane was reconsti-tuted from the platinum - coated membrane specimen, prepared with minimal intrusion, by using the freeze - etching technique. These images clearly demonstrated that the membrane

8 CHAPTER 1 CYTOSKELETON-INDUCED MESOSCALE DOMAINS

skeleton entirely covers the cytoplasmic surface of the plasma membrane, except for certain membrane domains, such as clathrin - coated pits (CCPs) (Chapter 2 ), caveolae (Chapter 3 ), and focal adhesions (Chapter 17 ), that the membrane - skeleton meshwork is primarily composed of actin fi laments since almost every fi lament exhibited a distinct striped pattern with a 5.5 - nm periodicity (see Fig. 1.2 for typical electron micrographs, although they are normal two - dimensional images), and that some of the meshwork is located as close as within a nanometer of the plasma membrane.

Based on these images, the size distribution of the actin skeleton mesh on the cyto-plasmic surface of the plasma membrane was obtained and found to agree well with the compartment sizes found by analyzing phospholipid hop diffusion as revealed by SPT. This result strongly supports the partitioning of the plasma membrane by membrane - skeleton fences and transmembrane protein pickets lining the fence.

Figure 1.2. Electron microscopic images of the membrane skeleton of NRK (upper left) and fetal rat skin keratinocytes (FRSKs) (lower left) cells, where the scale bars in the main fi gures correspond to 100 nm. The inset in the lower left image highlights the 5.5 - nm periodic striped pattern characteristic of actin fi laments, showing that the membrane skeleton is primarily composed of actin fi laments (scale bar 50 nm). Note that both of these images also contain clathrin - coated pits, which are distinguished by their lattice structure (from Morone et al. 2006 ). In the lower right corner, the trajectory of a gold - tagged phospholipid (DOPE) in an NRK cell, obtained by SPT at 40,000 frames/s, is displayed. The trajectory was color coded after performing a quantitative analysis that detects jumps between adjacent domains. The histograms in the upper right corner show that the size distribution of the membrane - skeleton meshes directly contacting the cytoplasmic surface of the plasma membrane, as determined by electron tomography, is very close to that of the compartments determined from the DOPE diffusion data in either NRK or FRSK cells, whereas the distributions for these two cell types are entirely different (in part from Fujiwara et al. 2002 ).

StartFinish

Finish

Start 1 µm

NRK cellPhospholipid (DOPE)

NRK

FRSK

Electron TomographySPT

ANSWERS TO THE TWO THREE-DECADE-OLD ENIGMAS 9

ANSWERS TO THE TWO THREE - DECADE - OLD ENIGMAS

The fence – picket model gives straightforward answers to the two three - decade - old ques-tions raised in the beginning of this section.

1. Why do membrane proteins and lipids diffuse faster in artifi cial membranes than in the cellular plasma membrane by a factor of ∼ 20 (ranging from 5 to 50)?

As stated in the previous paragraph, the diffusion coeffi cient within a compart-ment is not small, as compared with that in artifi cial membranes. However, if the diffusion coeffi cient is measured on much greater scales, for example, a FRAP spot size of ∼ 500 nm, or at slower observation rates, for example, single - molecule track-ing observed at a rate of 30 frames/s or slower, then the diffusion appeared to be slow because what was observed is the apparent diffusion coeffi cient, which is affected by the presence of compartment boundaries. It takes time to hop from one compartment to an adjacent one, which makes the macroscopic diffusion of lipids and proteins in the plasma membrane very slow.

2. How do molecular complexes become immobilized on the cell surface or diffuse at surprisingly lower rates, as compared with single molecules?

This can be explained by the “ oligomerization - induced trapping ” effect of the fences and pickets. Monomers of membrane molecules may hop across the inter-compartment boundaries with relative ease, but upon forming oligomers or molecu-lar complexes, the entire complex, rather than single molecules, has to hop across the picket – fence all at once, and therefore, these complexes are expected to hop across the boundaries at much slower rates. In addition, due to the avidity effect, molecular complexes are more likely to be bound to the membrane skeleton, perhaps temporarily, which also induces (temporary) immobilization or trapping of oligo-mers and molecular complexes. Such enhanced confi nement and binding effects induced by oligomerization or molecular complex formation are collectively termed oligomerization - induced trapping (Kusumi and Sako 1996 ; Iino et al. 2001 ; Kusumi et al. 2005 ).

One might argue that even in the absence of membrane - skeleton fences and anchored pickets lining the fences, the oligomerization of membrane proteins could greatly reduce the diffusion coeffi cient. However, experimental and theoretical studies clearly showed that this does not occur with transmembrane proteins that contain approximately three or more membrane - spanning α - helices (Saffman and Delbr ü ck 1975 ; Peters and Cherry 1982 ; Vaz et al. 1982 ; Liu et al. 1997 ; Gambin et al. 2006 ). Based on these studies, we predict that tetramer formation from monomers (a twofold increase in radius) will only decrease the diffusion coeffi cient by a factor of 1.1, and even 100mers (a 10 - fold increase in radius) will have a diffusion rate reduced by only a factor of less than 2 from that of monomers. Therefore, the large reductions of the diffusion coeffi cient, upon oligomeriza-tion or molecular complex formation, clearly indicate that the plasma membrane cannot be considered as a two - dimensional fl uid continuum, and that the fence – picket model is consistent with the experimental observations and theoretical predictions.

In artifi cial membranes without the membrane skeleton, Liu et al. (1997) and Gambin et al. (2006) reported decreases of the diffusion coeffi cient by a factor of ∼ 2 when the probe hydrodynamic diameter (in the cross section of the transmembrane domain paral-lel to the membrane surface) was increased by a factor of ∼ 2 from the original diameter of ∼ 0.5 nm. In this spatial scale, the probe (solute) size is comparable with the solvent

10 CHAPTER 1 CYTOSKELETON-INDUCED MESOSCALE DOMAINS

molecular size, and the continuum fl uid model is no longer correct. Therefore, the trans-lational diffusion coeffi cient of the test particle strongly depends on its diameter on this particular spatial scale.

However, this should not be confused with the reduction in the diffusion coeffi cient when single - pass transmembrane proteins, such as many receptor molecules, form oligo-mers in the plasma membrane . The size - dependent decrease of the diffusion coeffi cient observed in artifi cial membranes (Liu et al. 1997 ; Gambin et al. 2006 ) occurs only for test particles of ∼ 0.5 nm in diameter undergoing virtually simple diffusion, which exhibits a monomer diffusion coeffi cient of ∼ 10 μ m 2 /s. In the plasma membrane, the effective diffu-sion coeffi cients of single - pass transmembrane protein monomers are generally ∼ 0.2 μ m 2 /s, which are already slower by a factor of 50 than those found in artifi cial membranes. Oligomerization of such a single - pass transmembrane protein tends to decrease the diffu-sion coeffi cient by a factor of ∼ 2, but the weak resemblance of these reduction factors is merely incidental. The oligomerization - induced reduction of the diffusion coeffi cient in the plasma membrane was not due to changes in protein interaction with membrane lipids, as was found in artifi cial membranes, but rather to the compartment boundaries. Evidence for this statement comes from high - speed single - molecule tracking data. Single - pass transmembrane proteins generally exhibit hop diffusion between compartments, with a microscopic diffusion coeffi cient within a compartment of 5 – 10 μ m 2 /s and a residency time of 20 – 100 ms, and oligomerization lengthened the residency time without affecting the compartment size and the microscopic diffusion coeffi cient within a compartment (Murase et al. 2004 ; Kusumi et al. 2005 ). This clearly indicates the necessity for being careful in interpreting long - range, slow - speed diffusion measurements, for carrying out high - speed single - molecule tracking, and for careful attention to loosely applying the results of Liu et al. (1997) and Gambin et al. (2006) to the observations made in the plasma membrane.

In addition to the corralling effect of the membrane skeleton and associated trans-membrane protein pickets, three major factors are frequently discussed in the literature to account for the slower diffusion of membrane molecules in the cellular plasma membrane, observed by methods with low spatiotemporal resolutions: the crowding effect of trans-membrane proteins, the trapping or exclusion effect of raft domains, and the ordering effect of cholesterol. All three of these factors can lead to a decrease in molecular mobility in the plasma membrane, but even if all of these three factors are combined, a decrease in the diffusion coeffi cient by a factor of 20 cannot be explained (perhaps, a factor of 2 could be explained). Some of the phenomena including the immobilization of receptor com-plexes can only be explained by considering the effects of the membrane skeleton.

MEMBRANE - SKELETON - BASED MESODOMAINS ARE NOT DETECTED IN EVERY STUDY

A number of studies failed to detect the effect of the membrane skeleton on the diffusivity of membrane molecules. Two FRAP studies (Schmidt and Nichols 2004 ; Frick et al. 2007 ) found that disruption of the cortical actin did not lead to signifi cant changes in the dif-fusivity of some membrane proteins and a lipid analog. However, experiments involving drug - induced actin (de)polymerization revealed various effects, depending on the drug concentration, the treatment duration, and the cell type, and thus comparisons of the results are diffi cult (Vrljic et al. 2005 ; Umemura et al. 2008 ). For the intricate cell reactions to such treatments, see Suzuki et al. (2005) . Furthermore, Fujiwara et al. (2002) and Murase

MEMBRANE-SKELETON-BASED MESODOMAINS ARE NOT DETECTED IN EVERY STUDY 11

et al. (2004) found that without directly observing the compartment size and the residency time within a compartment, detecting drug - induced changes will be diffi cult.

In studies using fl uorescence correlation spectroscopy (FCS) (Wawrezinieck et al. 2005 ; Lenne et al. 2006 ) and stimulated emission depletion – FCS (STED - FCS) (Eggeling et al. 2009 ), phospholipid analogs were observed to diffuse freely. To detect hop diffusion in FCS experiments, the focal area of the laser should be less than the characteristic com-partment size. In STED - FCS (Eggeling et al. 2009 ), the diameter of the laser beam can be as small as 30 nm, which is comparable with the smallest compartment sizes reported in the literature (Murase et al. 2004 ). Therefore, a further decrease in the diameter of the laser beam seems to be necessary for determining the true nature of molecular diffusion by FCS techniques.

In addition to these results obtained by measuring the signal from many molecules simultaneously, several single - molecule/particle studies also did not fi nd hop diffusion. For instance, Wieser et al. (2007, 2008) found simple Brownian diffusion of lipids and a glycosylphosphatidylinositol (GPI) - anchored protein CD59, and Crane and Verkman (2008) reported the observation of freely diffusing aquaporin - 1 water channels. In these studies, the rate at which the images were acquired was 2000 Hz or less, resulting in determination of the positions of single molecules at relatively sparse time points such that it would simply be impossible to detect fast hop diffusion among mesoscale compartments.

In fact, careful attention must always be paid to the camera frame rates relative to the molecular hop frequency in each cell type when interpreting the single - molecule track-ing results, since this is a key factor in the ability to detect confi nement. The residency time of a molecule in a compartment may vary between one to hundreds of milliseconds, and will be shorter for lipids (Fujiwara et al. 2002 ; Murase et al. 2004 ) and longer for proteins (Tomishige et al. 1998 ; Tomishige and Kusumi 1999 ). Especially for lipids, to obtain statistically signifi cant results, single - molecule/particle tracking must be performed at very high speeds. At a camera speed of 40,000 Hz, which was used by Fujiwara et al. to detect the hop diffusion of lipids, the time between two consecutive frames is 25 μ s, thus allowing one to obtain 40 data points even for short residency times such as 1 ms (Fujiwara et al. 2002 ; Murase et al. 2004 ). Readers are referred to Murase et al. (2004) for the residency times of an unsaturated phospholipid, 1,2 - dioleoyl - sn - glycero - 3 - phosphoethanolamine (DOPE), in a number of different cells. The residency time divided by 40 essentially provides a good estimate of the inverse frame rate that is necessary to detect hop diffusion.

For example, Sahl et al. (2010) reported that they failed to detect the hop diffusion of a fl uorescent phospholipid analog, but instead found alternating temporary entrapment within a domain (30% of the observed duration) and simple Brownian diffusion (70%) in the plasma membrane of PtK2 cell, using an SFMT method with a time resolution of 0.5 ms. Using the same cell line, the average residency time of a phospholipid within a compartment was found to be ∼ 1 ms (Fujiwara and Kusumi, unpublished observations). This result indicates that, to detect the hop diffusion of phospholipids, the time resolution of the instrument must be better than 0.025 ms, suggesting that the results obtained by Sahl et al. (2010) are due to the lack of time resolution (on average, they made only two coordinate determinations during the average residency time of 1 ms): when a molecule stays in a compartment much longer than average, even the 0.5 - ms observation rate would be suffi cient to detect confi nement, but when the molecule stays in compartments for shorter periods, a 0.5 - ms resolution would be insuffi cient to detect confi nement, thus reporting simple Brownian diffusion.

12 CHAPTER 1 CYTOSKELETON-INDUCED MESOSCALE DOMAINS

MOLECULAR COMPOSITION OF THE MEMBRANE SKELETON

As shown previously (Heuser and Kirschner 1980 ; Hirokawa and Heuser 1981 ; Hirokawa et al. 1982; Morone et al. 2006 ; Hanson et al. 2008 ), the membrane skeleton mainly consists of actin fi laments (see also Chapter 12 ). Actin fi laments grow at their fast growing ends, which interact with the membrane (Small et al. 1978 ; Wang 1985 ). Therefore, the membrane compartments should depend on the proteins involved in actin polymerization, including the Arp2/3 complex, Wiskott – Aldrich syndrome protein (WASP), suppressor of cyclic AMP receptor (SCAR)/WASP family verprolin - homologous protein (WAVE), cortactin, and the formin family of proteins (Pollard 2007 ; Campellone and Welch 2010 ), as well as those involved in actin depolymerization such as the cofi lin family of proteins (Bernstein and Bamburg 2010 ). Actin fi laments and the plasma membrane are also linked by proteins that can laterally bind to actin, including ponticulin; the ezrin/radixin/moesin family of proteins; the villin – gelsolin superfamily proteins; the epithelial protein lost in neoplasm, fi lamin, dystrophin and utrophin, and tropomyosin; and the myosin family of proteins (Morone et al. 2008 ). Furthermore, actin fi laments bind to various other membrane - associated proteins and lipids with low affi nities, to facilitate dynamic associations with the membrane. This easy binding to and dissociation from the membrane would be needed for the dynamic regu-lation of the diffusivities of various molecules in the plasma membrane. Despite such low affi nities, the binding sites would be numerous, making the dynamic binding very effective, although rendering their binding undetectable in general pull - down assays. As indicated in a previous section, transmembrane proteins transiently bound to the membrane skeleton would serve well as a diffusion barrier, if the anchored durations are greater than ∼ 10 μ s.

In normal erythrocytes, the organization of the membrane skeleton has been studied extensively and is known to be different from those of other cell types. The erythrocyte membrane - skeleton mesh is composed of a network of spectrin tetramers, acting as fences, which are cross - linked by protein complexes formed by short actin fi laments, adducin, band 4.1, and the transmembrane protein glycophorin C, anchoring the entire network to the plasma membrane (Goodman et al. 1988 ; Bennett 1990 ; Anong et al. 2009 ). Spectrin tetramers form by the tail - to - tail linkage of spectrin dimers, and the dimers and the tetra-mers exist in a very dynamic equilibrium. For the passage of transmembrane proteins across the spectrin fences, a spectrin tetramer dissociation gate model (SPEQ gate model) was proposed. In the model, the dissociation of a spectrin tetramer into dimers entails transient gate opening, allowing transmembrane molecules to pass through the compart-ment boundaries (Tomishige et al. 1998 ). Malfunctions of the human erythrocyte mem-brane skeleton often lead to serious types of anemia, indicating the important roles played by the membrane skeleton in the functions, stability, and fl exibility of the plasma mem-brane (Bennett and Healy 2008 ; Kodippili et al. 2009 ).

FUNCTIONS OF MEMBRANE - SKELETON - INDUCED MESODOMAINS

Effects of the Membrane Skeleton on Signaling

B - Cell Receptor ( BCR ) Signal Transduction: The Membrane Skeleton May Be Involved in the Formation of Receptor Oligomers and/or Engaged - Receptor - Based Lipid Rafts Several different mechanisms for the relationship between the actin cytoskeleton and BCR signaling have been proposed. In the oligomeric BCR complex model (Reth et al. 2000 ;

FUNCTIONS OF MEMBRANE-SKELETON-INDUCED MESODOMAINS 13

Schamel and Reth 2000 ), disruption of BCR oligomers was considered to be required for signaling, and the role of the membrane skeleton in regulating the formation and mainte-nance of oligomers was discussed.

In another approach, the effect of lipid rafts was emphasized (see also Chapters 4 and 5 ). In this model, BCR is considered to enter rafts only upon antigen binding, which is accompanied by the coalescence of smaller rafts. This in turn enables BCR to react with raft - associated kinases to start signaling (Tolar et al. 2005 ; Gupta et al. 2006 ; Gupta and DeFranco 2007 ; Sohn et al. 2008 ). The membrane skeleton can participate in the regulation of signaling as it restricts the growth and coalescence of lipid rafts. Recent observations by Treanor et al. (2010) showed that the treatment of cells with actin - modulating chemicals resulted in up to a 10 - fold change in the diffusivity of BCR, suggesting that the membrane skeleton can control the mobility of these molecules. Furthermore, the expression level of ezrin, which is among the proteins that can form a link between the cytoskeleton and membrane proteins, also caused a change in BCR diffusivity by a factor greater than 3. This fi nding is also consistent with the anchored - protein picket model, since ezrin could act as a link between the protein pickets and the membrane skeleton. Moreover, the cross - linking of BCR and the dynamics of the membrane skeleton were correlated, and BCR signaling upon antigen binding leads to signifi cant cytoskeletal reorganization (Fleire et al. 2006 ; Arana et al. 2008 ; Lin et al. 2008 ). Interestingly, alteration of the actin cyto-skeleton also led to B - cell signaling comparable with that elicited by BCR cross - linking. Therefore, a clear link exists between the membrane skeleton and BCR signaling, implying that membrane mesodomains can play a critical role in this process, by controlling the distribution of membrane molecules.

Epidermal Growth Factor ( EGF ) Receptor ( EGFR ) and I g E - F c (F C portion of Immunoglobulin E) Receptor Signal Transduction: The Membrane Skeleton May Increase the Receptor Oligomerization Rate and Promote Receptor Activity Together with the Galectin Lattice Lajoie et al. (2007) raised the question of how the three mechanisms of regulating the dynamics, clustering, distribution, and function of the EGFR are coordinated: (1) Galectin - 3 is located on the extracellular surface of the plasma membrane and cross - links β 1,6GlcNAc - branched N - glycans on cell - surface glycoproteins, including EGFR, to form a heterogeneous lattice; (2) caveolin - 1 (Cav1), a major constituent of caveolae, acts as a negative regulator of growth factor signaling (Parton and Simons 2007 ), and in addi-tion, Cav1 forms noncaveolar microdomains, which are likely to contain at least 15 Cav1 molecules (Parton et al. 2006 ); (3) membrane - skeleton fences and pickets suppress the diffusion of EGFR, and particularly that of the signal - capable receptor oligomers. Lajoie et al. (2007) showed that the extracellular galectin - 3 lattice interacts with the N - glycans on EGFR and impedes its diffusion, which predominantly protects EGFR from loss to caveolae and Cav1 microdomains where EGFR signaling is suppressed. Disruption of the actin membrane skeleton with latrunculin A increased the mobile fraction of EGFR measured by FRAP, suggesting that the galectin - bound EGFR is further stabilized by the actin membrane skeleton. It is likely that galectin - 3 cross - links EGFR to other actin - associated membrane glycoproteins or pickets, thus generating actin - stabilized signaling domains.

Using quantum - dot - based SPT techniques, Chung et al. (2010) measured the dif-fusion coeffi cient of EGFR as a function of time. Assuming a relationship between dif-fusivity and molecular size, the authors inferred that EGFRs form transient oligomers, even in the absence of ligand, with lifetimes ranging from a few to a few tens of seconds. Interestingly, the probability of fi nding receptor dimers was reportedly higher at the

14 CHAPTER 1 CYTOSKELETON-INDUCED MESOSCALE DOMAINS

periphery of EGFR - overexpressing A431 and BT20 cells, in an actin - dependent manner. The abundance of dimers along the cell periphery might be a consequence of the increased collision frequency of receptors confi ned in the mesodomains delimited by the membrane skeleton. Previous observations provided evidence for the increased density of actin fi la-ments near the leading edge of migrating cells (Borisy and Svitkina 2000 ; Pollard et al. 2000 ). In this respect, another function of the membrane - skeleton - induced compartments could be promoting the formation of receptor complexes, and enabling the polarized response after ligand binding that is necessary in certain cellular processes such as chemotaxis.

The involvement of raft domains induced by receptor engagement in signaling, by enhancing the recruitment of Lyn kinase and reducing the collisions with phosphatases, was shown in the case of the IgE - Fc receptor, Fc ε RI (Field et al., 1997 ; Wu et al. 2004 ; Young et al. 2005 ). In addition, the involvement of membrane - skeleton - delimited compart-ments was found by simultaneous observations of the quantum - dot - labeled Fc ε RI and green fl uorescent protein (GFP) - tagged actin (Andrews et al. 2008 ). The diffusion rate of cross - linked receptor in cells with a disrupted actin cytoskeleton was signifi cantly higher than that in intact cells, in accordance with oligomerization - induced trapping. Immobilization of cross - linked receptors at the onset of signaling (Menon et al. 1986 ) also enables the cell to remember the position of the stimulus for short periods (on the order of 10 seconds) and to perform localized responses. This is essential in certain processes, such as chemo-taxis. Therefore, the actin meshwork is also involved in reliably responding to local changes in the environment (Kusumi and Sako 1996 ; Kusumi et al. 2005 ).

In fact, receptor redistribution and clustering are key steps in many signal transduc-tion pathways (Petrini et al. 2004 ; Minguet et al. 2007 ; Briegel et al. 2009 ; Nikolaev et al. 2010 ). Several reports have indicated the active roles played by the cytoskeleton in inhibiting (Wang et al. 2001 ; Boggs and Wang 2004 ) or enabling (Gomez - Mouton et al. 2001 ; Rodgers and Zavzavadjian 2001 ; Baumgartner et al. 2003 ) the redistribution/clustering of membrane molecules. In oligodendrocytes, stimulation of the cells induced actin depolymerization that was followed by coclustering of membrane molecules, includ-ing myelin basic protein and galactosylceramide (Boggs and Wang 2004 ). However, when the actin fi laments were artifi cially stabilized, coclustering was not observed. This observa-tion is consistent with the presence of membrane - skeleton - based mesodomains, since lateral diffusion, which is necessary for coclustering, is hindered by the presence of actin - based membrane - skeleton fences and pickets.

Kv2.1 Potassium Channels: The Membrane Skeleton Is Involved in Cluster Maintenance and Distribution FRAP and quantum - dot - based imaging revealed that Kv2.1 potassium channels form dynamic clusters with sizes and spatial distributions that are infl uenced by the actin - based membrane skeleton (O ’ Connell et al. 2006 ). Interestingly, the individual channels diffused freely within a cluster, suggesting that the clusters might actually be channels corralled by the membrane skeleton. This proposal was further supported by the results obtained upon disruption of the membrane skeleton by latrunculin A treatment, which resulted in a 10 - fold increase in the average cluster area and a decrease in the number of clusters, indicating that the smaller clusters merged during the treatment.

Furthermore, the spatial distribution of Kv2.1 channels in hippocampal neurons was also affected by the disruption of the membrane skeleton (O ’ Connell et al. 2006 ). In cells with an intact cytoskeleton, the channels were restricted to the cell body. However, after the latrunculin A treatment, the channels were found in the neurites as well as in the cell body. These observations suggest that the Kv2.1 clusters are maintained in membrane -