Actin-based Motility: Cellular, Molecular and Physical Aspects

Actin-based Motility Marie-France Carlier Editor Actin-based Motility Cellular, Molecular and Physical Aspects 123 ...

Author: Marie-France Carlier

17 downloads 1321 Views 10MB 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

DOWNLOAD PDF

Actin-based Motility

Marie-France Carlier Editor

Actin-based Motility Cellular, Molecular and Physical Aspects

123

Editor Dr. Marie-France Carlier CNRS Laboratory of Structural Enzymology and Biochemistry 91198 Gif-sur-Yvette Cedex UPR 3082 CNRS - Bât. 34 France [email protected]

ISBN 978-90-481-9300-4 e-ISBN 978-90-481-9301-1 DOI 10.1007/978-90-481-9301-1 Springer Dordrecht Heidelberg London New York Library of Congress Control Number: 2010935449 © Springer Science+Business Media B.V. 2010 No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work. Printed on acid-free paper Springer is part of Springer Science+Business Media (www.springer.com)

Preface

Since the discovery of actin by Straub in the 1950’s and the pioneering work of Oosawa on actin self-assembly in helical filaments in the 1960’s, many books and conference proceedings have been published. As one of the most essential proteins in life, essential for movement in organisms rangingfrom bacteria to higher eukaryotes, it is no surprise that actin has fascinated generations of scientists from many different fields. Actin can be considered as a “living treasure” of biology; the kinetics and thermodynamics of self-assembly, the dissipative nature of actin polymerization, the molecular interactions of monomeric and polymerized actin with regulators, the mechanical properties of actin gels, and more recently the force producing motile and morphogenetic processes organized by the actin nanomachine in response to signaling, are all milestones in actin research. Discoveries that directly derive from and provide deeper insight into the fundamental properties of actin are constantly being made, making actin an ever appealing research molecule. At the same time, the explosion in new technologies and techniques in biological sciences has served to attract researchers from an expanding number of disciplines, to study actin. This book presents the latest developments of these new multiscale approaches of force and movement powered by self-assembly processes, with the hope to opening our perspectives on the many areas of actin-based motility research. Three sections, focusing on the cellular, molecular and physical aspects of actin-based motility, provide examples of the main cellular processes in which actin self-assembly drives movement and changes the shape of cells. Emphasis is given on how the analysis of molecular interactions of actin with regulators of filament assembly enabled the reconstitution of self-organized motile systems. Physical and mathematical models give a formal description of force production by such actin assemblies. In various chapters, insight is provided into novel cell imaging technologies, high resolution structural methods applied to large protein assemblies, single molecule measurements and computer simulations. Self-assembly is an essential feature of morphogenesis, and actin interacts not only with itself but with a large variety of proteins to reconstitute simple modules. Most certainly, actin also integrates into multicomponent systems containing lipids or nucleic acids, to generate new activities of higher complexity that drive, for instance, the highly tuned and coordinated

v

vi

Preface

morphogenetic events. Recent technological advances make it possible to monitor the detailed underlying molecular interactions in vivo, in real time. Future books will have to reveal more elaborated works of actin in physiological and pathological situations or in organism development. The present book restricts its coverage of actin-based motility processes to topics at the interface of molecular and cellular aspects that may have seminal bearings in future studies. Gif-sur-Yvette Cedex, France

Marie-France Carlier

Contents

Part I

Cellular Aspects

1 Elementary Cellular Processes Driven by Actin Assembly: Lamellipodia and Filopodia . . . . . . . . . . . . . . . . . . . . . . J. Victor Small and Klemens Rottner

3

2 Coupling Membrane Dynamics to Actin Polymerization . . . . . . Shiro Suetsugu and Tadaomi Takenawa

35

3 Endocytic Control of Actin-based Motility . . . . . . . . . . . . . . Andrea Disanza, Emanuela Frittoli, Chiara Giuliani, Francesca Milanesi, Andrea Palamidessi, Flavia Troglio, and Giorgio Scita

59

4 Actin in Clathrin-Mediated Endocytosis . . . . . . . . . . . . . . . Marko Kaksonen

85

5 Actin Cytoskeleton and the Dynamics of Immunological Synapse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Viveka Mayya and Michael L. Dustin

103

6 Actin-based Motile Processes in Tumor Cell Invasion . . . . . . . . Matthew Oser, Robert Eddy, and John Condeelis

125

7 Actin-based Chromosome Movements in Cell Division . . . . . . . Rong Li

165

8 Roles for Actin Dynamics in Cell Movements During Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Minna Roh-Johnson, Jessica Sullivan-Brown, and Bob Goldstein Part II

187

Molecular Aspects

9 Regulation of the Cytoplasmic Actin Monomer Pool in Actin-based Motility . . . . . . . . . . . . . . . . . . . . . . . . . Pekka Lappalainen, Maarit Makkonen, and Hongxia Zhao

213

vii

viii

10

Contents

From Molecules to Movement: In Vitro Reconstitution of Self-Organized Actin-based Motile Processes . . . . . . . . . . . Marie-France Carlier and Dominique Pantaloni

237

11

The WASP-Homology 2 Domain and Cytoskeleton Assembly . . . Roberto Dominguez

255

12

Formin-Mediated Actin Assembly . . . . . . . . . . . . . . . . . . David R. Kovar, Andrew J. Bestul, Yujie Li, and Bonnie J. Scott

279

13

Visualization of Individual Actin Filament Assembly . . . . . . . . Emmanuèle Helfer

317

14

Movement of Cargo in Bacterial Cytoplasm: Bacterial Actin Dynamics Drives Plasmid Segregation . . . . . . . . . . . . . Dyche Mullins

335

Part III Physical Aspects 15

Protrusive Forces Generated by Dendritic Actin Networks During Cell Crawling . . . . . . . . . . . . . . . . . . . . . . . . . Ovijit Chaudhuri and Daniel A. Fletcher

359

Mathematical and Physical Modeling of Actin Dynamics in Motile Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Anders E. Carlsson and Alex Mogilner

381

Force Production by Actin Assembly: Simplified Experimental Systems for a Thorough Modeling . . . . . . . . . . C. Sykes, J. Prost, and J.F. Joanny

413

Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

433

16

17

Contributors

Andrew J. Bestul Department of Molecular Genetics and Cell Biology, Department of Biochemistry and Molecular Biology, The University of Chicago, 920 East 58th Street, Chicago, IL 60637, USA, [email protected] Marie-France Carlier CNRS, Gif-sur-Yvette, Paris, France, [email protected] Anders E. Carlsson Department of Physics, Washington University, St. Louis, Mo 63130, USA, [email protected] Ovijit Chaudhuri Department and Biophysics Program, University of California, Berkeley, Berkeley, CA 94720, USA; Physical Biosciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA, [email protected] John Condeelis Department of Anatomy and Structural Biology; Gruss Lipper Biophotonics Center, Albert Einstein College of Medicine of Yeshiva University, Bronx, NY 10461, USA, [email protected] Andrea Disanza IFOM, The FIRC Institute for Molecular Oncology Foundation, at the IFOM-IEO Campus, Via Adamello 16, 20139 Milan, Italy, [email protected] Roberto Dominguez Department of Physiology, University of Pennsylvania School of Medicine, 3700 Hamilton Walk, Philadelphia, PA 19104-6085, USA, [email protected] Michael L. Dustin Skirball Institute of Biomolecular Medicine, New York University School of Medicine, 540 First Avenue, New York, NY 10016, USA, [email protected] Robert Eddy Department of Anatomy and Structural Biology, Albert Einstein College of Medicine of Yeshiva University, Bronx, NY 10461, USA, [email protected] Daniel A. Fletcher Department of Bioengineering and Biophysics Program, University of California, Berkeley, Berkeley, CA 94720, USA; Physical

ix

x

Contributors

Biosciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA, [email protected] Emanuela Frittoli IFOM, The FIRC Institute for Molecular Oncology Foundation, at the IFOM-IEO Campus, Via Adamello 16, 20139 Milan, Italy, [email protected] Chiara Giuliani IFOM, The FIRC Institute for Molecular Oncology Foundation, at the IFOM-IEO Campus, Via Adamello 16, 20139 Milan, Italy, [email protected] Bob Goldstein Biology Department, UNC Chapel Hill, Chapel Hill, NC 27599-3280, USA, [email protected] Emmanuèle Helfer Laboratoire d’Enzymologie et Biochimie Structurales, CNRS UPR3082, Bât. 34 Avenue de la Terrasse, F-91198, Gif-sur-Yvette Cedex, France, [email protected] J.F. Joanny Physicochimie Curie (CNRS-UMR168), Institut Curie, Section de Recherche, 26 rue d’Ulm, 75248 Paris Cedex 05, France, [email protected] Marko Kaksonen EMBL, Heidelberg, Germany, [email protected] David R. Kovar Department of Molecular Genetics and Cell Biology, Department of Biochemistry and Molecular Biology, The University of Chicago, 920 East 58th Street, Chicago, IL 60637, USA, [email protected] Pekka Lappalainen Institute of Biotechnology, University of Helsinki, P.O. Box 56, 00014, Helsinki, Finland, [email protected] Rong Li Stowers Institute for Medical Research, 1000 East 50th street, Kansas City, MO 64110, USA, [email protected] Yujie Li Department of Molecular Genetics and Cell Biology, Department of Biochemistry and Molecular Biology, The University of Chicago, 920 East 58th Street, Chicago, IL 60637, USA, [email protected] Maarit Makkonen Institute of Biotechnology, University of Helsinki, Helsinki, Finland, [email protected] Viveka Mayya Skirball Institute of Biomolecular Medicine, New York University School of Medicine, 540 First Avenue, New York, NY 10016, USA, [email protected] Francesca Milanesi IFOM, The FIRC Institute for Molecular Oncology Foundation, at the IFOM-IEO Campus, Via Adamello 16, 20139 Milan, Italy, [email protected] Alex Mogilner Department of Neurobiology, Physiology and Behavior, and Department of Mathematics, University of California, Davis, CA 95616, USA, [email protected]

Contributors

xi

Dyche Mullins UCSF School of Medicine, N312F Genentech Hall, 600 16th Street, San Francisco, CA, USA, [email protected] Matthew Oser Department of Anatomy and Structural Biology, Albert Einstein College of Medicine of Yeshiva University, Bronx, 10461, NY, USA, [email protected] Andrea Palamidessi IFOM, The FIRC Institute for Molecular Oncology Foundation, at the IFOM-IEO Campus, Via Adamello 16, 20139 Milan, Italy, [email protected] Dominique Pantaloni CNRS, Gif-sur-Yvette, Paris, France, [email protected] J. Prost Physicochimie Curie (CNRS-UMR168), Institut Curie, Section de Recherche, 26 rue d’Ulm, 75248 Paris Cedex 05, France; E.S.P.C.I, 10 rue Vauquelin, 75231 Paris Cedex 05, France, [email protected] Minna Roh-Johnson Biology Department, UNC Chapel Hill, Chapel Hill, NC 27599-3280, USA, [email protected] Klemens Rottner Cytoskeleton Dynamics Group, Helmholtz Centre for Infection Research, Inhoffen Straße 7, D-38124, Braunschweig, Germany, [email protected] Giorgio Scita IFOM, The FIRC Institute for Molecular Oncology Foundation, at the IFOM-IEO Campus, Via Adamello 16, 20139 Milan, Italy; Dipartimento di Medicina, Chirurgia ed Odontoiatria, Universita’ degli Studi di Milano, 20122 Milan, Italy, [email protected] Bonnie J. Scott Department of Molecular Genetics and Cell Biology, Department of Biochemistry and Molecular Biology, The University of Chicago, 920 East 58th Street, Chicago, IL 60637, USA, [email protected] J. Victor Small Institute of Molecular Biotechnology, Austrian Academy of Sciences, Dr. Bohr Gasse 3, A-1030 Vienna, Austria, [email protected] Shiro Suetsugu Laboratory of Membrane and Cytoskeleton Dynamics, Institute of Molecular and Cellular Biosciences, The University of Tokyo, Tokyo, Japan: PRESTO, Japan Science and Technology Agency, Saitama, Japan, [email protected] Jessica Sullivan-Brown Biology Department, UNC Chapel Hill, Chapel Hill, NC 27599-3280, USA, [email protected] C. Sykes Physicochimie Curie (CNRS-UMR168), Institut Curie, Section de Recherche, 26 rue d’Ulm, 75248 Paris Cedex 05, France, [email protected]

xii

Contributors

Tadaomi Takenawa Department of Lipid Biochemistry, Kobe University Graduate School of Medicine, Kobe, Japan, [email protected] Flavia Troglio IFOM, The FIRC Institute for Molecular Oncology Foundation, at the IFOM-IEO Campus, Via Adamello 16, 20139 Milan, Italy, [email protected] Hongxia Zhao Institute of Biotechnology, University of Helsinki, Helsinki, Finland, [email protected]

Part I

Cellular Aspects

Chapter 1

Elementary Cellular Processes Driven by Actin Assembly: Lamellipodia and Filopodia J. Victor Small and Klemens Rottner

Abstract Cell migration is initiated by the extension of thin cytoplasmic sheets, termed lamellipodia and finger-like rods, termed filopodia. Both structures are composed of actin filaments, organized in networks in lamellipodia and bundles in filopodia and protrusion is driven by the polymerization of actin, with monomer insertion at the tips of these processes, between the filament plus ends and the membrane. The formation of lamellipodia and filopodia is induced via different but overlapping pathways, but the extent of their structural and functional interrelationships remains controversial. We propose that lamellipodia and filopodia are each driven by core machineries, partly overlapping, which are poised for engagement with upstream signalling pathways. Protrusion in each case is associated with the recruitment of protein complexes to the cell membrane that nucleate actin filaments, regulate their elongation and tether their ends to the membrane. Other factors are required to stabilise, crosslink and regulate the turnover of the actin filament assemblies. In addition to their role in protrusion, lamellipodia and filopodia seed filaments for the construction of the contractile regions of the cytoskeleton required for retraction.

Contents 1.1 1.2 1.3 1.4 1.5 1.6 1.7

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . Choosing Your Protrusion . . . . . . . . . . . . . . . . . . . Signalling to Lamellipodia and Filopodia: Rho-GTPases and Beyond Organization of the Pushing Machinery in Lamellipodia . . . . . . Actin Nucleators and Elongators in Lamellipodia . . . . . . . . . Determinants of Actin Turnover in Lamellipodia . . . . . . . . . Determinants of Filopodia Formation . . . . . . . . . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

4 7 7 8 12 15 20

J.V. Small (B) Institute of Molecular Biotechnology, Austrian Academy of Sciences, Dr. Bohr Gasse 3, A-1030 Vienna, Austria e-mail: [email protected]

M.-F. Carlier (ed.), Actin-based Motility, DOI 10.1007/978-90-481-9301-1_1, C Springer Science+Business Media B.V. 2010

3

4

J.V. Small and K. Rottner

1.8 Actin Filament Recycling: From Protrusion to Retraction . . . . . . . . . . . . 1.9 Moving Forward . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

22 24 25

1.1 Introduction In order to move, eukaryotic cells have developed subtle variations in strategy, all based on remodelling the actin cytoskeleton. Net translocation is achieved by protrusion at the cell front, followed by retraction at the rear. Pulling up the rear entails recruitment of actin and myosin filaments (myosin II) into contractile arrays that interconnect traction points on the cell surface. Protrusion relies on the extension of pseudopodia, in one of three forms: finger like “filopodia”, sheet-like “lamellipodia” or bulbous “blebs”. The extension of filopodia and lamellipodia, often seen in combination, is driven by the polymerization of actin and is not myosin II dependent. Blebs on the other hand result from membrane bulging through locally induced flexibility of the cell cortex, coupled with cytoplasmic flow driven by actomyosindependent intracellular pressure (Charras and Paluch, 2008). The archetypal form of blebb-based movement is displayed by Amoeba proteus, but is also exhibited by primordial germ cells in teleosts (Blaser et al., 2006; Trinkaus, 1984), by Walker carcinoma cells (Wicki and Niggli, 2001) and other tumour-derived cells (Sahai and Marshall, 2003). Dictyostelium amoebae can switch from a blebbing to a lamellipodia/filopodia mode of protrusion when myosin II activity is suppressed (Yoshida and Soldati, 2006) suggesting, together with other data (Lammermann and Sixt, 2009) the ability of cells to adopt alternative and perhaps combined strategies for advancement, depending on cell type and the intrinsic and extrinsic factors at play. For many eukaryotic cells, the extension of lamellipodia and filopodia appears to be the preferred mode of protrusion (Fig. 1.1). These structures are morphologically better suited than blebs for probing and contacting other cells, for penetrating tissue spaces as well as for interacting with the extracellular matrix. Lamellipodia and filopodia were already evident in drawings of the early histologists, but under different names. In his treatise on the nervous system Ramon y Cajal makes various observations on the neuronal growth cone: “Finally, one occasionally observes highly flattened cones that resemble a membrane reinforced with elongated ridges; all in all they remind one of a webbed foot” (Swanson and Swanson, 1995). Cajal’s images, in a Golgi stain (Fig. 1.2A), match contemporary fluorescence labelling for actin (Fig. 1.2B) and clearly show filopodia interspersed by lamellipodia veils. In the context of pseudopodia, the term “filopodia” appeared rather late on the scene. Using phase contrast microscopy and 3000 feet of 16 mm cine film, Nakai (Nakai, 1956) described the movement of “filopodia” but even then the term was competing with “fine pseudopodia”, “microfibrils” and “microspikes” (AlbrechtBuehler, 1976; Gey, 1954; Hughes, 1953; Taylor and Robbins, 1963). Abercrombie suggested “lamellipodia” by analogy to filopodia to describe respectively sheet-like and rod-like extensions (Abercrombie et al., 1970; also Albrecht-Buehler, 1976), a terminology now generally accepted. Depending on cell type, lamellipodia can

1

Elementary Cellular Processes Driven by Actin Assembly

5

Fig. 1.1 Protruding and contractile domains of a moving cell. (A) Schematic representation of the actin cytoskeleton in a polarised fibroblast. The different organisational forms of actin filaments are depicted: Diagonal actin filament meshwork in the lamellipodium, with associated radial bundles (Filopodium and Microspike); contractile bundles of actin (Stress fiber) in the cell body and at the cell edge; and a loose actin network throughout the cell. Arc-shaped bundles are sometimes observed that move inwards under the dorsal cell surface (Arc). The diagram shows an idealized cell; in reality the actin arrays are interconnected in various combinations and geometries. Sites of adhesion of the cell with the substrate are also indicated, in red. The flat region behind the lamellipodium and in front of the nucleus (N) has been termed the Lamella. At the cell front, in lamellipodia and filopodia, actin filaments are all polarized in one direction, with their fast polymerizing ends forwards (for pushing); in the body of the cytoskeleton, actin filaments form bipolar assemblies with myosin to form contractile arrays (for retracting); (B and C) Video frames of a living goldfish fibroblast transfected with mCherry-actin and GFP-myosin-II light chain (as in Nemethova et al., 2008): (B) actin channel; (C) superimposition of actin and myosin channels. Myosin II is mainly excluded from the protruding front. Images in (B) and (C) were provided by Maria Nemethova

6

J.V. Small and K. Rottner

Fig. 1.2 Lamellipodia, filopodia and microspikes. (A) Image of a neuronal growth cone (marked “c”) stained by the Golgi method produced by S. Ramon Y. Cajal (From Swanson and Swanson, (1995), with permission). (B and C) live cell images of goldfish fibroblast (B) and a mouse melanoma cell (C) expressing fluorescent actin. Abbreviations: lp, lamellipodium; fp, filopodium; ms, microspike

be traversed by bundles of actin filaments that do not project beyond the tips of the lamellipodium boundary (Fig. 1.2B, C). To distinguish these from filopodia, we have adopted the term “microspike” (Faix et al., 2009; Small et al., 2002). Video analysis shows that microspikes can either remain as bundles spanning the breadth of the lamellipodium or can transform into filopodia by extending beyond the lamellipodium boundary (Small et al., 2002; Svitkina et al., 2003). In this context, the sheet-like extensions in fibroblasts originally referred to as lamellipodia by Abercrombie (Abercrombie et al., 1970) are structurally heterogeneous; they generally contain variable mixtures of interwoven actin networks and bundles with transition structures between (Small et al., 1982). The current understanding of a lamellipodium is a sheet-like segment of protruding cytoplasm composed of an actin network devoid of tightly bundled actin. Before considering the protrusion machinery in more detail we should briefly reflect on the evidence that protrusion of lamellipodia and filopodia is driven by actin polymerization. Tilney was the first to propose that actin can push by polymerization from his findings on the acrosomal reaction of Thyone sperm (Tilney et al., 1973). He showed that actin was the major protein of the acrosomal process, that the actin bundle of the process emerged from a concentrated pool of profilactin around the actomere and that actin filaments in the bundle were polarized, with exogenously added myosin heads pointing away from the tip (Tilney and Kallenbach, 1979) corresponding to the actin filament end already shown to be favoured for growth (Woodrum et al., 1975; now referred to as the plus end). At about the same time fibroblast lamellipodia were found to be composed of a network of actin filaments with the filaments polarized uniformly with their plus ends towards the direction of protrusion (Small et al., 1978), suggesting that here too actin polymerization could be the driving force. Subsequently, Wang demonstrated by photobleaching

1

Elementary Cellular Processes Driven by Actin Assembly

7

in fibroblasts that actin monomer is incorporated at the tips of lamellipodia and microspikes with actin subunits in the filaments moving rearwards in a treadmilling mode (Wang, 1985), a process predicted earlier by Wegner (1976). As already indicated, myosin II-based contractility does not play a direct role in protrusion since myosin-II null Dictyostelium amoeba (Yoshida and Soldati, 2006) and vertebrate cells treated with contractility inhibitors (Koestler et al., 2008; Lammermann et al., 2008; Rottner et al., 1999b) develop or maintain both lamellipodia and filopodia. Protrusion is therefore achieved as a result of actin polymerization and elongation at the network/membrane interface. Independent, direct evidence for the ability of actin filaments to push by polymerization has come from the seminal development of in vitro models of actin-based motility (Carlier et al., 2003; Loisel et al., 1999) as elaborated elsewhere in this volume. In the present chapter, we consider current concepts of the molecular regulation and mechanics of protrusion based on lamellipodia and filopodia of eukaryotic cells as well as aspects of coupling protrusion at the front with retraction at the rear.

1.2 Choosing Your Protrusion The relative expression of lamellipodia versus filopodia varies considerably between cell types. Epidermal keratocytes from fish and amphibia in primary culture develop extensive lamellipodia for migration that are devoid of filopodia. At the other extreme, nerve growth cones are enriched in filopodia, although not devoid of lamellipodia. So far, the complete lack of lamellipodia in a migrating cell has not been described, but the possibility that cells can move with filopodia alone has yet to be excluded. Filopodia appear to serve several specific functions: (1) as guidance structures, sensing the matrix and growth factor environment ahead (Bentley and Toroian-Raymond, 1986; Chien et al., 1993; Gerhardt et al., 2003); (2) as initiators of adhesion with the extracellular matrix (DePasquale and Izzard, 1987; Letourneau, 1981; Nemethova et al., 2008); (3) as intercellular communicators aligning epithelial cell interfaces (Jacinto et al., 2002; Wood and Martin, 2002); (4) and as we shall see, as precursor building blocks for construction of the contractile part of the cytoskeleton (Nemethova et al., 2008). Lamellipodia and ruffles, on the other hand are well tailored to produce cup like structures for phagocytosis and to engulf liquid nutrients during macropinocytosis. In this latter activity, ruffles commonly gain structural support from microspikes and filopodia integrated into the actin networks. In short, the ability of a cell to generate sheets and rods, with all combinations in between allows for a structural flexibility in protrusions that can be tailored according to demand and subject to feedback responses to environmental cues.

1.3 Signalling to Lamellipodia and Filopodia: Rho-GTPases and Beyond The formation of leading protrusions such as lamellipodia and filopodia can be stimulated by various treatments, typically by engagement of soluble ligands by growth

8

J.V. Small and K. Rottner

factor or hormone receptors or by interaction of extracellular matrix components with integrins (DeMali et al., 2003; Disanza et al., 2005). These events cause a complex set of intracellular responses, both short-term and long-term, with some of them inducing the formation of lamellipodia and filopodia. Notably, plasma membrane blebbing has not yet been linked to signal transduction pathways (Charras and Paluch, 2008). The short-term events implicated in signal transduction to actin polymerization in lamellipodia and filopodia range from phosphorylations by protein and lipid kinases to activation of heterotrimeric or small GTPases, the latter of which are thought to serve as key links to actin binding proteins (Burridge and Wennerberg, 2004; Ladwein and Rottner, 2008; Ridley, 2006). The signalling pathways induced by soluble ligands or extracellular matrix components to drive protrusions are complex and diverse, but appear to invariably converge on some common mediators of signal transduction, such as Src or PI3-Kinases (Cain and Ridley, 2009; Frame, 2004). Similiarly, Rho-GTPases and their direct activators (such as guanine-nucleotide exchange factors, GEFs) have emerged as central regulators of protrusion, with Cdc42 and Rac1 inducing filopodia and lamellipodia, respectively (Heasman and Ridley, 2008; Monypenny et al., 2009; Nobes and Hall, 1995; Ridley et al., 1992). However, several unresolved issues remain, including whether lamellipodia can form in the absence of any Rac activity, for instance downstream of Cdc42 (Heasman and Ridley, 2008), or how filopodia can form in the absence of Cdc42 (Czuchra et al., 2005). Likewise, the precise target(s) of a given Rho-GTPase effecting the respective protrusion, especially for those not yet extensively studied (like Rif or Tc10 or RhoG) remain to be characterized.

1.4 Organization of the Pushing Machinery in Lamellipodia Since lamellipodia are not crystalline in their organization, electron microscopy is the only method currently available to resolve lamellipodia ultrastructure. We are here confronted with the problem, not generally acknowledged, that actin filaments are readily distorted by some of the conventional procedures used for electron microscopy. As a result opinions are currently divided about how actin filaments are organized in lamellipodia, specifically whether they are branched (Svitkina and Borisy, 1999) or not branched (Small et al., 2008). We will outline the origin of the discrepancy and then mention new data that serve to resolve the issue. Following the first visualization of actin filaments in ultrathin plastic sections of muscle (Huxley, 1957) it took more than a decade to positively identify actin filaments in non-muscle cells in plastic sections, namely through the decoration of cytoplasmic actin with myosin heads (Ishikawa et al., 1969). Further work showed that actin filaments could be otherwise identified from their diameter in sections of non-muscle cells, but only where they occurred in bundles, as in stress fibres, in filopodia (Goldman and Knipe, 1973) or in the acrosomal process (Tilney et al., 1973). But even in plastic sections of filopodia, actin was often poorly preserved: in a first report on filopodia ultrastructure, Taylor (Taylor, 1966) described

1

Elementary Cellular Processes Driven by Actin Assembly

9

an ill-defined “fascicle of strands” which he mistakenly identified as labile microtubules! And in the first plastic sections of growth cone lamellipodia only a reticular pattern of irregular fibres was observed (Yamada et al., 1971). The poor preservation of lamellipodia actin in embedded material was subsequently explained by the degrading effect of both osmium tetroxide fixation and dehydration in organic solvents on actin filaments (Maupin-Szamier and Pollard, 1978; Small, 1981). Only where actin filaments are complexed with certain proteins, in particular myosin and tropomyosin as in muscle cells and in stress fibre bundles of non-muscle cells, are they stabilized against these agents. After a conventional dose of osmium tetroxide as a post-fixative, followed by dehydration and embedding, lamellipodia can appear in plastic cross sections as an empty membrane envelope, almost devoid of structure (Abercrombie et al., 1971) indicating the more or less complete destruction of actin. These findings are cited to highlight the differential sensitivity of actin filaments to some of the conventional procedures used in electron microscopy, according to whether they occur in networks or bundles. We are fortunate that lamellipodia are thin enough to be penetrated by an electron beam in a conventional microscope, and can therefore be visualized directly in cells grown on EM support films, without the need of plastic embedding and sectioning. Current methods used for processing cell monolayers to visualize lamellipodia ultrastructure are summarized in Fig. 1.3. Ideally, one would prefer to freeze cells directly and perform cryo electron tomography, avoiding all processing steps that could potentially introduce artifacts. Although this method has been applied to Dictyostelium amoeba and actin filaments were filtered out of the tomograms of peripheral zones of frozen cells (Medalia et al., 2002) there was no supporting evidence to show that the region(s) imaged were lamellipodia. More work is needed to exploit the potential of this approach. Alternative methods rely on permeabilising the cell membrane to expose the cytoskeleton, using detergent in the presence of glutaraldehyde (Flanagan et al., 2001; Hoglund et al., 1980; Koestler et al., 2008; Korobova and Svitkina, 2008; Small, 1981) or in the presence of a polymer (e.g. polyethylene glycol; Svitkina et al., 1995) to limit extraction of protein. From thereon, processing strategies for electron microscopy diverge, with markedly different results. In the critical point drying method championed by Svitkina (Fig. 1.3), fixed cytoskeletons are subjected to multiple processing steps: hardening with tannic acid, dehydration in ethanol, critical point drying in C02 and platinum coating. The images obtained of keratocyte lamellipodia show an amalgamated network of filaments that appear branched at intervals of 20-50nm in regions close to the lamellipodium front, whereby the Arp2/3 complex is thought to generate the branch points (Svitkina and Borisy, 1999). However, the same critical point procedure produces branches in gels of pure actin filaments, without the Arp2/3 complex (Resch et al., 2002). A second approach (Flanagan et al., 2001; Heuser and Kirschner, 1980; Small et al., 1994; Fig. 1.3) entails rapid freezing of cytoskeletons, freeze drying and coating with platinum. After this procedure, keratocyte lamellipodia show a variable appearance, with individual filaments following a more or less undulating course, but with nanoscale kinks and irregularities (Lee et al., 1993; Small et al., 1994)

10

J.V. Small and K. Rottner

Fig. 1.3 Scheme of alternative processing steps to visualize the actin cytoskeleton in the electron microscope. To expose the cytoskeleton, cells are extracted and fixed simultaneously in a mixture of Triton X-100 and glutaraldehyde. They can then be taken through alternative steps to the electron microscope, involving drying by different means: dehydration in an organic solvent and critical point drying; drying in negative stain or freeze drying. Alternatively, cells can be rapidly frozen without fixation or extraction and transferred to a cryo-electron microscope. See text for details

likely arising from uncontrolled ice crystal damage. In the third approach, fixed cytoskeletons are simply dried in an aqueous, neutral heavy metal salt (Hoglund et al., 1980; Koestler et al., 2008; Small, 1981; Small et al., 1982; Fig. 1.3). In the latter case actin filaments are noticeably linear in lamellipodia and are more clearly resolved in filopodia than by other methods (Fig. 1.4). These different appearances of the lamellipodium (see also Small et al., 2008) highlight again the delicate nature of actin networks. In short, it requires only minor distortions to convert a network of originally overlapping filaments into one in which filaments appear fused or branched. In the study by Koestler et al., mouse melanoma cells transfected with RFPactin and GFP-VASP were grown on EM support films and single cells imaged by high-resolution fluorescence microscopy. After fixation the same cells, contrasted by negative stain, were relocated in the EM and images recorded of lamellipodia at different locations along the cell edge. The results showed that the organization of the actin network changed according to whether the lamellipodium was continuously protruding, slowing or pausing at the time of fixation. In particular, it was demonstrated that the angular distribution of filaments changed towards lower angles to the lamellipodium tip during slowing and pause. There was no regular angle of around 70◦ between adjacent filaments, as would be expected from the dendritic branching model of lamellipodium protrusion (Pollard and Borisy,

1

Elementary Cellular Processes Driven by Actin Assembly

11

Fig. 1.4 Lamellipodia (lp) and filopodia (fp) as visualized in a cytoskeleton of a B16 melanoma cell after negative staining. The micrograph was provided by Stefan Koestler: conditions as in Koestler et al. (2008)

2003). Because of the high density of actin filaments in lamellipodia networks it was however not possible in two-dimensional projections obtained by conventional transmission electron microscopy (Koestler et al., 2008), to unambiguously distinguish overlapping from branching events. Resolution of this problem requires 3D imaging of actin networks by EM tomography (Small et al., 2008). More recent studies using this approach have directly demonstrated that actin filaments do not branch in lamellipodia (Urban et al., 2010) substantiating the conclusion that actin branches described in critical point dried preparations of cytoskeletons, from keratocytes and other cells (Korobova and Svitkina, 2008; Svitkina and Borisy, 1999; Yang et al., 2007) are artifacts of the preparation procedure. A central feature of the branching idea is that the Arp2/3 complex, an essential co-factor in lamellipodia protrusion (see further below), induces the branching of actin filaments. This idea comes originally from the finding of branched actin structures in in vitro mixtures of Arp2/3 and gelsolin-capped actin seeds (Mullins et al., 1998). Subsequent work on in vitro mixtures has confirmed the existence of branched actin filaments in the presence of the Arp2/3 complex, but in the invariable presence of phalloidin, which has been shown to promote branching (Mahaffy and Pollard, 2008). The relevance of actin branches in vitro needs to be re-evaluated in light of these findings. We return to the role of Arp2/3 in lamellipodia protrusion below. The idea that actin networks are structured by the Arp2/3-complex and its associated proteins has resulted in the inadvertent oversight of the role of actin cross-linkers in lamellipodia organization. In particular, lamellipodia contain filamin (Hartwig and Shevlin, 1986; Langanger et al., 1984) a potent cross-linker of actin filaments. Certain melanoma cell lines lack filamin and mainly bleb at their periphery, producing lamellipodia only transiently (Flanagan et al., 2001). In the electron microscope, these lamellipodia appear less three-dimensional than those in the same

12

J.V. Small and K. Rottner

cells transfected with filamin, consistent with a role of filamin in stabilizing the actin network (Flanagan et al., 2001). More attention to filamin and other cross-linkers is required to understand their role in the generation, stabilization and turnover of actin networks in lamellipodia.

1.5 Actin Nucleators and Elongators in Lamellipodia Actin nucleation in the lamellipodium, usually transduced via Rac (Nobes and Hall, 1995; Ridley et al., 1992) is currently considered to be driven largely by filament generation through the Arp2/3-complex (Goley and Welch, 2006; Machesky and Insall, 1998; Steffen et al., 2006). However, Arp2/3-complex is inactive in isolation, thus has to be activated by so-called nucleation promoting factors (NPFs), such as Scar/WAVE (Takenawa and Suetsugu, 2007). At the earliest detectable times of lamellipodia protrusion, the WAVE complex, comprising in mammals Scar/WAVE and Abi isoforms (three each), Sra1/PIR121, Nap1 (or haematopoietic Hem1) and Brick1/HSPC300, is recruited to the lamellipodium tip (Derivery et al., 2008; Hahne et al., 2001; Steffen et al., 2004; Stradal et al., 2001, 2004), together with other components, including VASP/Mena, lamellipodin/PREL1, IRSp53 and heterodimeric capping protein (Jenzora et al., 2005; Krause et al., 2004; Lai et al., 2008; Mejillano et al., 2004; Nakagawa et al., 2003; Rottner et al., 1999a; Small et al., 2002). When lamellipodia arrest or retract, canonical tip components such as VASP or WAVE delocalize instantly from the lamellipodium tip (Rottner et al., 1999a; Small et al., 2002), indicating that the polymerization machinery is only assembled at membrane target sites when required. In accordance with these findings, injection of the small GTPase Rac1 into cells lacking lamellipodia causes immediate protrusion of lamellipodia and the coincident recruitment and accumulation of the WAVE-complex at the protruding tips (Steffen et al., 2004; Stradal et al., 2004). Indeed, the presence of WAVE complex components or VASP at the cell edge can be considered as diagnostic of lamelliopodia formation; so far no physiological situations have been described in which the two do not go hand in hand. The mechanisms or factors that restrict the localization of the polymerization machinery to specific sites and that determine protrusion in the form of a sheet remain to be elucidated. In addition to these canonical tip components, Yang et al. described a similar accumulation in the tip region of the lamellipodium of a member of the formin family of actin nucleators (Goode and Eck, 2007), more specifically an active variant of the diaphanous related formin mDia2, also known as Drf3 (Yang et al., 2007). However, more recent data using similarly activated versions of this formin indicated this accumulation to be restricted to structures harbouring not only lamellipodial (like Arp2/3-complex), but also filopodial markers (such as fascin), indicating that the protrusions observed corresponded to intermediate hybrid structures, rather than bona fide lamellipodia (Block et al., 2008). Further work will be required to establish the functional relevance of actin nucleation mediated by formins in the lamellipodium (if it exists) relative to Arp2/3-complex (see also below).

1

Elementary Cellular Processes Driven by Actin Assembly

13

The family of canonical nucleation promoting factors in human mediating Arp2/3-complex activation comprises three WAVE isoforms (1, 2 and 3), haematopoietic WASP, the more ubiquitous but neuronally enriched N-WASP, the recently discovered WASH (Linardopoulou et al., 2007) and the closely related WHAMM (Campellone et al., 2008) and JMY proteins (Zuchero et al., 2009). The productive interaction between all these factors and Arp2/3-complex depends on a so-called WA-domain (also called VCA or WCA), which is essential for actin and Arp2/3-complex binding. When expressed in cells as a single domain the WA-region abrogates lamellipodium protrusion, presumably due to Arp2/3-complex sequestration (Machesky and Insall, 1998). Based on both their subcellular location and results derived from genetic depletion in mouse, neither WASP nor N-WASP are involved in lamellipodium protrusion (Lommel et al., 2001; Snapper et al., 2001), but instead in vesicle trafficking and podosome/invadopoda formation (Benesch et al., 2002, 2005; Jones et al., 2002; Merrifield et al., 2004; Mizutani et al., 2002; Yamaguchi et al., 2005). The claim that N-WASP cooperates in protrusion with IQGAP-1 (Bensenor et al., 2007; Le Clainche et al., 2007) is inconsistent with the finding that lamellipodia formation occurs normally in the total absence of N-WASP (Lommel et al., 2001; Stradal et al., 2004). The key role for Arp2/3complex-dependent actin assembly in trafficking has recently been reinforced by the elucidation of WHAMM functions in maintaining Golgi structure and promoting anterograde membrane transport (Campellone et al., 2008). Although WAVEcomplex is currently thought to constitute the key activator of Arp2/3-complex in the lamellipodium in vertebrates and invertebrates (Innocenti et al., 2004; Kunda et al., 2003; Rogers et al., 2003; Steffen et al., 2004), other NPFs have been implicated in the process as well. For example, WASH was concluded to localize to ruffles in Drosophila cells (Linardopoulou et al., 2007), although a clear involvement in Arp2/3-complex activation in the lamellipodium remains to be confirmed. The discovery that the p53-cofactor JMY can promote both Arp2/3-dependent and independent actin assembly in vitro, has recently received significant attention. Both activities have been suggested to contribute to actin assembly in lamellipodium protrusion and efficient cell migration (Roadcap and Bear, 2009; Zuchero et al., 2009). Whether or not JMY can indeed activate actin assembly in the lamellipodium independently of the WAVE/Arp2/3-complex signalling axis remains to be clarified, but would be surprising in light of the essential function proposed for WAVE-complex in diverse cell types (Stradal and Scita, 2006). Arp2/3-complexindependent actin assembly induced by JMY is mediated by three consecutive WH2-domains, similar to the mechanism described for another class of actin nucleators, including Spire and Cobl (Ahuja et al., 2007; Chesarone and Goode, 2009; Quinlan et al., 2005). As opposed to Spire, the largely neuronal protein Cobl was also shown to localize to ruffles and lamellipodia, although its contribution relative to other Arp2/3-dependent or -independent actin assembly events in lamellipodiaor ruffle-like structures remains to be established. Yet another class of Arp2/3-complex activators implicated in protrusion comprises in mammals the Arp2/3- and actin filament-binding protein cortactin, and its haematopoietic counterpart HS1 (Ammer and Weed, 2008). Cortactin co-localizes

14

J.V. Small and K. Rottner

in lamellipodia with Arp2/3-complex, and was proposed to activate Arp2/3dependent nucleation, as well as to stabilize the Arp2/3-complex in filament branch points (Cai et al., 2008; Cosen-Binker and Kapus, 2006). Irrespective of the existence of branches in the lamellipodium (see above), thorough analysis of Racinduced lamellipodium protrusion and migration in cells genetically depleted for cortactin has recently revealed cortactin to be dispensible for Arp2/3-complexdependent actin assembly in the lamellipodium (Lai et al., 2009). Cortactin was also shown to exhibit different turnover kinetics from Arp2/3 in lamellipodia, inconsistent with it being involved in Arp2/3-complex activation (Lai et al., 2008). We conclude that cortactin modulates lamellipodium protrusion by means other than Arp2/3-complex activation and/or stabilization. It is presently unclear if it is at all possible to generate cells genetically deficient for Arp2/3-complex and amenable to tissue culture studies (Vauti et al., 2007). Other approaches will be required to probe for Arp2/3-independent mechanisms. Cell permeable chemical inhibitors of Arp2/3-complex show interesting potential (Nolen et al., 2009), but as with RNA interference (Kunda et al., 2003; Steffen et al., 2006) they will per definition not suffice to achieve full blockade of Arp2/3-complex activity in cells. We may surmise that while WAVE-and Arp2/3-complex appear to play a major role in the nucleation of actin filaments in the lamellipodium (see below), the discovery and initial characterization of multiple additional factors potentially involved in generating actin filaments hint at further complexities yet to be unveiled. The identification and characterization of factors mediating or promoting actin filament elongation in the lamellipodium has emerged as an equally challenging issue, as illustrated by studies on the lamellipodial tip component VASP. Ena/VASP proteins can accelerate actin assembly when recruited to Listeria (Geese et al., 2002; Loisel et al., 1999; Niebuhr et al., 1997), coated on beads (Breitsprecher et al., 2008; Samarin et al., 2003) or targeted to the lamellipodium (Bear et al., 2002). The question of whether Ena/VASP proteins promote protrusion as simple “elongators” of actin filaments, or by employing additional, proposed biochemical activities, ranging from nucleation to inhibition of Arp2/3-dependent branching or anti-capping or a combination of them, has remained controversial (Bear and Gertler, 2009; Trichet et al., 2008). When VASP family members were sequestered on mitochondria and largely depleted from the cell periphery, lamellipodia still formed, but the actin filaments in lamellipodia appeared shorter and more branched (Bear et al., 2002). This observation (which requires corroboration by electron tomography) is consistent with the observation that VASP can processively elongate actin filaments in vitro (Breitsprecher et al., 2008). VASP clustered on a bead elongates actin from the bead surface without the filaments falling from the bead. This can be explained by the collaboration of VASP molecules in tetramers, taking turns in holding the filament and adding actin monomers to the end by direct interaction of their WH2-domains with actin monomers (Breitsprecher et al., 2008). Interestingly, this VASP activity did not even require profilin. By holding and elongating actin, VASP oligomers could operate as a type of “actoclampin”, as envisioned by Dickinson (2009), although this activity could in theory be exerted by any factor capable of interacting with both the filament barbed end, and with protein complexes residing at the membrane/actin

1

Elementary Cellular Processes Driven by Actin Assembly

15

network interface. Indeed, the idea of such membrane-associated complexes that nucleate and elongate actin filaments while tethering them at the same time seems more tenable than a model in which the gap between the filament and the membrane opens according to Brownian motion to allow addition of actin monomer (Mogilner, 2009). In cells Ena/VASP proteins are presumably themselves targeted to proteinaceous components at the lamellipodium tip. Candidates that could mediate this activity include lamellipodin/PREL family proteins (Jenzora et al., 2005; Krause et al., 2004), the WAVE-complex subunit Abi-1, in the presence or absence of association with other WAVE-complex subunits (Stradal and Scita, 2006), or IRSp53, another tip component (Nakagawa et al., 2003). Both IRSp53 and Abi-1 were previously shown to interact with the Ena/VASP family member Mena (Krugmann et al., 2001; Tani et al., 2003). The precise relevance of all these interactions remains to be clarified. Interestingly, a mouse model lacking all mammalian family members (VASP, Mena, Evl) has been established (Kwiatkowski et al., 2007), but the consequences for lamellipodium protrusion remain to be explored. Based on in vitro work (Breitsprecher et al., 2008; Brieher et al., 2004; Loisel et al., 1999; Samarin et al., 2003) and studies on Ena/VASP function in Listeria motility in the cytoplasm of cells (Sechi and Wehland, 2004), VASP is considered an elongator rather than a nucleator of actin filaments. Although the nucleation of actin filaments by Ena/VASP proteins in the lamellipodium tip has not been formally disproven, this appears currently unlikely. Brieher et al. have recently demonstrated that Ena/VASP can indeed continue to elongate actin filaments on Listeria surfaces upon Arp2/3-complex inhibition (Brieher et al., 2004). This suggests VASP and potentially other factors sharing barbed end binding capability (such as formins) could take over filaments nucleated by other factors, a scenario that could be envisaged in lamellipodia. The lateral flow of polymerizing filament ends at the lamellipodium front that results as a consequence of their oblique orientation to the membrane (Koestler et al., 2008) could promote or at least facilitate the clustering of polymerization complexes. We can expect that cooperation between nucleation and elongation factors will become a general theme in concepts of actin driven processes (Chesarone and Goode, 2009).

1.6 Determinants of Actin Turnover in Lamellipodia The lamellipodial actin network is not static, but continuously turned over, which occurs through highly coordinated actin filament assembly and disassembly mechanisms. The turnover of actin in a treadmilling mode produces a retrograde flow in the lamellipodium, readily detected by different techniques, including fluorescence recovery after photobleaching (FRAP: Lai et al., 2008; Wang, 1985), fluorescence speckle microscopy (FSM: Waterman-Storer et al., 1998), fluorescence localization after photobleaching (FLAP: Lai et al., 2008; Zicha et al., 2003) and spatio-temporal image correlation spectroscopy (STICS Brown et al., 2006). Receptors on the cell

16

J.V. Small and K. Rottner

surface link to the cytoskeleton and likewise move rearwards, taking with them surface bound beads, as shown in the early studies of Harris, which were the first to reveal the phenomenon of retrograde flow in lamellipodia (Harris, 1973). In B16-F1 mouse melanoma cells, the average retrograde flow rate of actin approximates 4 μm/min, so for a lamellipodium width of 3 μm, the network is turned over in less than a minute (Lai et al., 2008). Fibroblast lamellipodia polymerize actin less rapidly (2 μm/min) but overall turnover was comparable to B16-F1 owing to reduced average lamellipodial width (Lai et al., 2009). The rate of net protrusion of the cell depends on the efficiency with which the retrograde flow is coupled to the substrate (Mitchison and Kirschner, 1988). Efficient substrate coupling of the network is observed for instance in fish keratocyte lamellipodia (Anderson and Cross, 2000; Small et al., 1995; Theriot and Mitchison, 1991), allowing a more or less direct conversion of actin polymerization into protrusion, whereas fibroblasts plated on poly-lysine spread in all directions and polymerize large lamellipodia continuously slipping rearwards without effective protrusion (Flinn and Ridley, 1996; Small et al., 1999). In addition, lamellipodial actin polymerization can be largely uncoupled from substrate association, as when lamellipodia are lifted up during ruffling (see also below). The seminal experiments by Yu-Li Wang demonstrated for the first time that actin filaments in the lamellipodium are assembled at the front, although alternative concepts known as “nucleation release” or “basal polymerization” have also been put forward (Theriot and Mitchison, 1991; Watanabe and Mitchison, 2002). A common feature of the latter models is the existence of a significant amount of actin polymerization throughout the lamellipodium (Miyoshi et al., 2006; Tsuji et al., 2009). We recently analyzed lamellipodial actin turnover in more detail by comparing fluorescence recovery in the front and back halves of bleached lamellipodia. The resulting average difference between the two curves provided an unbiased, quantifiable number, the socalled “treadmilling factor”, which correlated well with mathematical simulation, assuming actin assembly to occur essentially at the front (Lai et al., 2008). The “basal polymerization”-model derived from single molecule tracking by fluorescence speckle microscopy (Tsuji et al., 2009; Watanabe and Mitchison, 2002) was not confirmed by the FRAP data. More recent findings by electron microscopy (Urban et al., 2010) indicate that the majority of actin filament plus ends are located at the lamellipodium front, consistent with the FRAP data and not with the “basal polymerization”-model. We therefore assume that actin assembly occurs at the front (Lai et al., 2008; Wang, 1985). Using the “treadmilling factor” as an indicator of lamellipodium dynamics, it was also possible to investigate the turnover of other lamellipodial components, relative to actin. FRAP experiments of cells expressing an EGFP-tagged Arp2/3-complex subunit revealed the exclusive incorporation of Arp2/3 at the lamellipodium tip (Lai et al., 2008), where it is presumably activated by WAVE (Stradal and Scita, 2006). Further experiments with cells co-expressing actin and Arp2/3-complex demonstrated that the temporal and spatial turnover of both are closely linked (see Fig. 1.5), excluding any major reorganization in the network of either component during retrograde flow, as was recently proposed (Cai et al., 2008; Miyoshi et al., 2006; Tsuji et al., 2009). In accord with earlier

1

Elementary Cellular Processes Driven by Actin Assembly

17

Fig. 1.5 Identical rearward flow rates of actin and Arp2/3-complex in the lamellipodium. B16-F1 cell co-expressing mCherry-actin (red) and the Arp2/3-complex subunit p16B (green) subjected to FRAP experiment. After bleaching a rectangular region of the lamellipodium (panels under FRAP), both Arp2/3 and actin re-incorporate into the lamellipodium tip as expected (Lai et al., 2008) with identical dynamics (see merge at the bottom and double-headed arrows). Note also that microspikes (red asterisks in actin and merged panels) are devoid of Arp2/3-complex (green asterisks in Arp2/3-channel), as shown previously (Svitkina et al., 1999), indicating a mechanism of formation distinct from lamellipodia. Time is in seconds after bleaching. The experiment was provided by Malgorzata Szczodrak

18

J.V. Small and K. Rottner

suggestions (Mullins et al., 1998), the Arp2/3 complex likely resides at the actin filament pointed ends, acting as a pointed end capper after serving its nucleation role at the lamellipodium tip. Network turnover is controlled by the activities of additional components, some known and likely others yet to be discovered. From studies on actin-driven pathogens (Rosenblatt et al., 1997) and biomimetic motility systems (Loisel et al., 1999; Pantaloni et al., 2001) we can assume that actin disassembly in the lamellipodium is mediated, at least in part, by ADF/cofilin. RNAi-mediated ADF/cofilin knockdown promoted lamellipodia formation in Jurkat T cells (Nishita et al., 2005) although not in fibroblasts (Hotulainen and Lappalainen, 2006). In insect S2 cells, the knockdown of cofilin (Iwasa and Mullins, 2007) or of the cofilin activator slingshot (Rogers et al., 2003) induced wider lamellipodia, which would be consistent with reduced de-polymerization of actin at the lamellipodium rear. Such a cofilin activity would be in full accord with both its localization and turnover in the lamellipodium relative to actin (Lai et al., 2008). However, cofilin can also nucleate actin filaments in vitro, at least at high concentrations (Andrianantoandro and Pollard, 2006), and – in one model – synergizes with Arp2/3-complex by severing filaments, thereby generating barbed ends to serve as templates for Arp2/3-dependent branching (van Rheenen et al., 2009). This mechanism is proposed to operate for instance in tumor cells during growth factor-induced formation of both lamellipodia (DesMarais et al., 2004; Ichetovkin et al., 2002) and invadopodia (Oser et al., 2009). More work is required to harmonize all these observations concerning ADF/cofilin functions and activity at the cell periphery. Heterodimeric capping protein has also been attributed a role in lamellipodium turnover. As opposed to pointed end binding by Arp2/3-complex, which could influence network turnover by protecting bound filaments from depolymerisation, capping protein binds to barbed ends (Schafer et al., 1998). In a generalized scheme of events capping protein stochastically blocks barbed end assembly in the lamellipodium front, and resides at capped ends while the network advances forward (Pollard, 2007; Pollard and Borisy, 2003). However, as capping protein localization is quite closely restricted to the front (Mejillano et al., 2004), with turnover rates in this location virtually identical to actin (Lai et al., 2008), we speculate it could mediate the capping of short or unproductive filaments (until they depolymerise) at the lamellipodium tip (Lai et al., 2008). It is also surprising that the functions of capping protein and Arp2/3-compley are so tightly connected. Mejillano et al. have shown capping protein knockdown to abolish both lamellipodia formation and Arp2/3-complex recruitment to the cell periphery (Mejillano et al., 2004). On the basis of more recent in vitro work, Akin and Mullins have suggested that capping protein is essential for Arp2/3-mediated nucleation because it counteracts the competition of elongating filaments for actin monomer (Akin and Mullins, 2008). However, this “Monomer Gating Model” would require monomer concentrations to become limiting in the lamellipodium, which is inconsistent with the presence in lamellipodia of actin monomer concentrations more than a 1000-fold in excess of the critical assembly concentration in vitro (Koestler et al., 2009). As an interactor of the Arp2/3 complex (see above), cortactin was suggested to play a key

1

Elementary Cellular Processes Driven by Actin Assembly

19

role in lamellipodium turnover (Ammer and Weed, 2008; Lapetina et al., 2009), although FRAP studies have shown cortactin and Arp2/3 to have different turnover dynamics (Lai et al., 2008). A more subtle role for cortactin is suggested from recent experiments employing conditional gene inactivation. Somewhat surprising, a modest increase in actin and Arp2/3-complex flow was observed in the lamellipodium of cortactin-deficient cells as compared to their respective controls (Lai et al., 2009), but lamellipodium dynamics and structure appeared otherwise normal. We presume that actin turnover in the lamellipodium is modulated by multiple factors, and could well be subject to tissue- or cell-type specific tuning of lamellipodial activity and thus migratory performance in different physiological and pathological conditions. How is filament elongation regulated at the lamellipodium tip? Protrusion rate and thus actin assembly correlates linearly with VASP concentration in the tip (Rottner et al., 1999a). This together with functional data (e.g. Loisel et al., 1999) had suggested Ena/VASP to serve key functions in promoting actin assembly (see above), which was confirmed, at least in part, by genetic deletion of Ena/VASP proteins or their sequestration (Bear and Gertler, 2009; Bear et al., 2002). Indeed, VASP and potentially other factors could reside in the lamellipodium tip to control elongation, as proposed from experiments on beads (Breitsprecher et al., 2008), but this would require lamellipodial tip components to reside at these sites long enough to productively exert their function. Recent FRAP data shows that VASP has a residence time at the lamellipodium tip (t1/2 of recovery of about 14 s) consistent with promoting filament elongation in the growing actin network. Similar turnover rates were obtained for the WAVE-complex subunit Abi-1 (Lai et al., 2008). But which parameters define the turnover of tip components such as VASP or WAVE? Exciting recent studies provided potential hints in this direction, as they established that actin polymerization itself could be an important determinant of the turnover of its regulators at the tip. One study exploited as a model system Vaccinia virusinduced actin tail formation induced at the plasma membrane, a signalling cascade ignited by tyrosine phosphorylation events on viral A36R, leading to recruitment of a protein complex containing the Arp2/3-complex activator N-WASP. Interestingly, the continuous Arp2/3-complex-dependent actin tail motility was accompanied by N-WASP turnover at the actin tail tip (Weisswange et al., 2009), reminiscent of WAVE turnover in the lamellipodium tip (Lai et al., 2008). However, N-WASP variants incapable of Arp2/3-complex activation failed to turn over at the viral surface, and inhibition of actin polymerization by Cytochalasin D also largely blocked turnover of wildtype N-WASP (Weisswange et al., 2009). These data suggest that active actin polymerization might provide a feedback signal to the turnover of certain actin assembly regulators exemplified by N-WASP in the vaccinia system. Such a scenario is consistent with the abrupt and reversible accumulation of VASP in the lamellipodium tip of B16-F1 lamellipodia upon local inhibition of protrusion by Cytochalasin B (see Fig. 1.6), which could again be caused by reduced turnover. Similar observations were made with the WAVE-complex component Abi-1 (our unpublished data). Likewise, the haematopoietic WAVE-complex subunit Hem-1 was dramatically increased in intensity at the periphery of neutrophils treated with latrunculin, and this again coincided with complete failure of Hem-1 to recover in

20

J.V. Small and K. Rottner

Fig. 1.6 Cytochalasin treatment causes enhanced accumulation of VASP at the cell periphery. The lamellipodium of a B16-F1 cell expressing EGFP-VASP was imaged by video microscopy and treated locally and transiently (+, with; –, without) by needle application of Cytochalasin B (needle concentration: 25 μg/ml). Note the reversible halt of protrusion with cytochalasin and the accumulation of VASP during arrest (arrowheads). Numbers correspond to video frames relative to 1st Cytochalasin B application

FRAP experiments (Weiner et al., 2007), as observed for N-WASP in the Vaccinia system (Weisswange et al., 2009). It is thus tempting to speculate that the accumulation of VASP in Cytochalasin B-treated B16-F1 cells (Fig. 1.6) shown here is caused by the selective block of its release from the lamellipodium tip, which might be the decisive parameter for turnover. The relevance of these turnover events remains elusive, as VASP- or Arp2/3-activator-coated beads cannot mimic this activity, and yet are capable of persistent motility (Breitsprecher et al., 2008; Samarin et al., 2003). Future studies should shed mechanistic insights into the intriguing dependence of actin assembly and the turnover of actin regulators, and how this relates to in vitroand in vivo- models of actin-based motility.

1.7 Determinants of Filopodia Formation Filopodia are rod-like, protrusive and membrane-enclosed bundles of actin filaments formed at the cell periphery, frequently in combination with lamellipodia (see also above). The molecular determinants of filopodia formation have been extensively discussed in recent reviews (Faix et al., 2009; Gupton and Gertler, 2007; Mattila and Lappalainen, 2008), so will not be repeated in detail. However, we should emphasize that the mechanisms of filopodia formation are even less clear than in the case of lamellipodia. The convergence of filaments from lamellipodia into microspike bundles and filopodia observed by electron microscopy formed the basis of the first suggestions (Small, 1981; Small et al., 1982) revived two decades later (Svitkina

1

Elementary Cellular Processes Driven by Actin Assembly

21

et al., 2003) that filopodia can form via recruitment of filaments from lamellipodia. Independent biochemical and cell biological evidence led to a scheme whereby filopodia form by Arp2/3-complex-dependent actin assembly, downstream of Cdc42 through the activation of Arp2/3 by N-WASP (Carlier et al., 1999; Martinez-Quiles et al., 2001; Miki et al., 1998). Subsequent work showed however that N-WASP and Arp2/3-complex are not essential for filopodia formation (Lommel et al., 2001; Nicholson-Dykstra and Higgs, 2008; Snapper et al., 2001; Steffen et al., 2006) and that N-WASP was instead involved in vesicle trafficking and endocytosis (Benesch et al., 2005; Innocenti et al., 2005; Merrifield et al., 2004). In Dictyostelium filopodia formation relies on the diaphanous related formin dDia2, which can nucleate and elongate actin filaments independently of Arp2/3-complex, and associates with the tips of filopodia (Faix and Rottner, 2006; Schirenbeck et al., 2005). Similar observations were made for mammalian orthologues of this formin, especially mDia2 (Pellegrin and Mellor, 2005; Yang et al., 2007), also known as Drf3 (Faix and Grosse, 2006), although its importance relative to other formins in the mammalian system remains to be established. Recent work has shown that active variants of mDia2/Drf3 can indeed cause the formation of exaggerated, club-shaped filopodia (Yang et al., 2007), built by actin filaments generated through de novo nucleation (Block et al., 2008). Filopodia formation presumably involves the clustering of filament plus ends, the zipping up of filaments into a bundle and filament elongation at the tip. The ability of VASP to form tetramers (Zimmermann et al., 2002) and to cross-link actin filaments (Schirenbeck et al., 2006) suggests that VASP, strategically localized at lamellipodia and filopodia tips (Rottner et al., 1999a; Svitkina et al., 2003), could contribute to the clustering of filament plus ends (Breitsprecher et al., 2008). The disruption of VASP in Dictyostelium abrogates filopodia formation (Faix et al., 2009; Han et al., 2002), and in vertebrate cells depleted of VASP family proteins filopodia are poorly developed (Dent et al., 2007). The actin binding protein fascin localises to filopodia bundles (Adams, 2004; Nemethova et al., 2008; Otto et al., 1979), and depletion of fascin in B16 melanoma cells compromises filopodia formation (Vignjevic et al., 2006), suggesting that fascin is a major bundling component in filopodia. Fascin also localizes to lamellipodia where it may contribute to the formation of parallel filament arrays observed by electron microscopy to merge into filopodia/microspike bundles (Small et al., 2002; Urban et al., 2010). Noteworthy is also the presence of tailless myosins in filopodia: Myosin VII in Dictyostelium and myosin X in vertebrates (Faix et al., 2009). In Dictyostelium, genetic deletion of myosin VII compromises filopodia formation (Tuxworth et al., 2001) and over expression of myosin X in vertebrate cells potentiates the formation of filopodia (Berg and Cheney, 2002). One function of these myosins is suggested to be in the delivery of cargo such as integrins to the filopodium tip (Zhang et al., 2004), but the issue of whether or not these myosins contribute to the protrusion process remains open. Much remains to be learned about the molecular mechanisms of filopodia formation. A pressing question relates to the definitions of lamellipodia and filopodia (see above), namely to what extent these two structures can be considered as distinct. More specifically, are all filaments born within the lamellipodial network

22

J.V. Small and K. Rottner

nucleated by an Arp2/3-dependent mechanism or by an alternative route involving, for example, a formin. It was recently suggested that a formin may be required to produce the “mother filaments” from which actin filaments are branched by the Arp2/3 complex (Yang et al., 2007), but the absence of actin branching in lamellipodia (Urban et al, 2010) would exclude this possibility. Nevertheless, formins could still potentially contribute to the nucleation of filaments in lamellipodia and if so might generate a filament subset destined for bundle formation. In the latter case, filopodia could in principle form independently of simultaneous protrusion of lamellipodia, consistent with the protrusion of Cdc42-induced filopodia upon interference with WAVE- or Arp2/3 complex function (Steffen et al., 2006). New experimental approaches will be required to clarify the interrelationships between filaments in lamellipodia and filopodia assemblies (Faix et al., 2009; Ladwein and Rottner, 2008).

1.8 Actin Filament Recycling: From Protrusion to Retraction When fluorescent actin is injected into fibroblasts it is rapidly incorporated, within a few minutes into two structures, lamellipodia/filopodia and focal adhesions; only later is fluorescence detected along actin bundles behind the lamellipodium (Glacy, 1983; Wang, 1985). Techniques that detect the flow of cytoplasmic actin monomer in living cells likewise demonstrate a major and rapid incorporation of the cytoplasmic pool into lamellipodia and filopodia, coincident with the continuous treadmilling of actin (Lai et al., 2008; Zicha et al., 2003). The protrusion machinery at the cell front thus appears to be the major generator of actin filaments in moving cells. Behind the lamellipodium boundary, in the region generally termed the lamella in cultured cells actin occurs as single filaments or in contractile stress fibre bundles and arcs (Heath and Holifield, 1991), the size and abundance of the bundles varying considerably between different cell types. In rapidly migrating cells, such as neutrophils, dendritic cells and amoeba, stress fibre bundles are absent. However, in these cells fine actin bundles can be seen to delimit the “inactive” cell edge or to form an interconnected network throughout the cytoplasm (Diez et al., 2005; Riedl et al., 2008; Yumura and Kitanishi-Yumura, 1990; Yuruker and Niggli, 1992). Hence, independent of cell type the actin cytoskeleton provides, in one way or another structural connections between the protruding front and the trailing rear. As already noted the front is protrusive and the rear contractile, with protrusion at the front effected by unidirectional polymerisation of actin and contraction behind by the sliding of antiparallel actin arrays, powered by myosin II. According to earlier considerations (Small and Resch, 2005; Small et al., 1998) recent findings suggest that the protruding front provides actin filaments for seeding assembly of the contractile assemblies at the rear. Live cell imaging of cells expressing fluoresent actin analogues has shown that a sub-population of microspikes and filopodia bundles are recycled, after protrusion,

1

Elementary Cellular Processes Driven by Actin Assembly

23

Fig. 1.7 Contribution of filopodia to the construction of the lamella. Video frame of living Goldfish fibroblast transfected with mCherry-actin and GFP-fascin. Some filopodia, marked with fascin, move as complete bundles or as fragments into the region behind the lamellipodium and contribute to stress fibre assembly. From Nemethova et al. (2008), with permission

into the lamella behind (Anderson et al., 2008; Koestler et al., 2008; Nemethova et al., 2008; Fig. 1.7). Bundles often traverse laterally and can fold sideways into the lamella via the base of the lamellipodium, they can fold backwards with a ruffle or kink and break and move with retrograde flow into the lamella. It is common to observe bundles moving in opposite directions along the protruding front, to see them recruited into the lamella, where they become populated by myosin (Koestler et al., 2008). Some filopodia become immobilized by forming an adhesion site and seed with their basal region the formation of a stress fibre bundle (Anderson et al., 2008; Nemethova et al., 2008). All these re-organisations of bundles serve to generate anti-parallel arrangements of actin filaments in the lamella that can form contractile assemblies with myosin (Nemethova et al., 2008). By a similar means, actin filaments generated by Arp2/3 nucleation in the lamellipodium contribute to the formation of actin arcs (Hotulainen and Lappalainen, 2006) that move rearward on the dorsal surface of moving or spreading cells (Heath and Holifield, 1991). Those filaments seeded in the lamellipodium that flow into the lamella to contribute to contractile arrays or arcs have mostly turned over before entering the lamella by treadmilling (Koestler et al., 2008; Lai et al., 2008). Protrusion at the front does not simply occur in a single plane, but also features ruffles; ruffles correspond to lamellipodia, including also filopodia protruding upwards and backwards, eventually merging into the cytoplasm several microns behind the lamellipodium (Abercrombie et al., 1970; Harris, 1973). Ruffles are especially well portrayed in scanning EM images of migrating cells (Heath and Holifield, 1991). Ruffles perform important function in phagocytosis and pinocytosis and no doubt in cell–cell recognition. We contend that ruffling activity also adds a further complement of filaments of flipped polarity to the lamella for integration into contractile arrays

24

J.V. Small and K. Rottner

Fig. 1.8 Contribution of ruffling to actin cytoskeleton assembly. Schematic illustration of the generation of contractile arrays in a hypothetical cell lacking stress fibre bundles (e.g. a neutrophil). Ruffles, including lamellipodia and filopodia fold back into the cell and contribute actin filament seeds for the formation of antiparallel assemblies with myosin that interconnect throughout the cell body. The contractile bundle around the periphery of the trailing part of the cell is formed via prior ruffling activity at those locations. Actin filaments are in green, myosin filaments in blue

(Fig. 1.8). The mechanisms underlying the lifting up and rearward movement of ruffles remain to be elucidated: since ruffling occurs in the presence of Rho kinase inhibitor (Rottner et al., 1999b) and blebbistatin (Nemethova et al., 2008), it is not directly dependent on myosin II. We should not be left with the misconception that filaments generated at the front of a moving cell end up at the rear. It is the prehistory of protrusion, substrate adhesion and contractile activity at specific regions of the cell periphery that determines the local organization of the actin cytoskeleton. Cells do not move in straight lines, even during chemotaxis, but they shuffle along in zig-zag trajectories. This is due to spontaneous protrusive activity around the entire cell periphery, with the general direction determined by the most persistent, actively protruding region. Cultured cells, such as fibroblasts and melanoma cells move randomly sometimes reversing their direction; and during contact inhibition the cessation of protrusive activity at regions of cell–cell contact are followed by activation of protrusion from the previously trailing cell edge (Abercrombie, 1980). In short, the contractile assemblies of actin at the immediate rear of a cell are seeded by previous regional and transient lamellipodia and filopodia activity (Fig. 1.8).

1.9 Moving Forward Mimetic models of motility continue to contribute valuable insights into the basic machinery of actin-based protrusion (Carlier, this volume). But in our quest to understand how protrusion takes place in the living cell we are faced with additional levels of complexity. How many nucleators or nucleation-promoting factors play a role in lamellipodia and filopodia and what determines their activity, in collaboration or alone, in one or the other structure? What is the cascade of events that leads to protrusion, from signalling to the assembly of the polymerization machinery at

1

Elementary Cellular Processes Driven by Actin Assembly

25

the membrane – to the recruitment of actin? What constraints limit the thickness and length of the lamellipodium and the size of filopodia and what roles do actin cross-linkers play in this process? And how do lamellipodia and filpodia lift up to produce ruffles? There is still much to learn about the first step that a cell takes to move. Acknowledgements We thank Malgorzata Szczodrak, Maria Nemethova and Stefan Koestler for kindly providing unpublished images and Tibor Kulcsar for assistance with the graphics. This work was supported in part by generous grants from the Austrian Science Research Foundation (FWF, to JVS) and from the German Research Council (DFG, to KR).

References Abercrombie, M. 1980. The Croonian Lecture, 1978: The crawling movement of metazoan cells. Proc Royal Soc London B. 207:129–47. Abercrombie, M., J.E. Heaysman, and S.M. Pegrum. 1970. The locomotion of fibroblasts in culture. II. “Ruffling”. Exp Cell Res. 60:437–44. Abercrombie, M., J.E. Heaysman, and S.M. Pegrum. 1971. The locomotion of fibroblasts in culture. IV. Electron microscopy of the leading lamella. Exp Cell Res. 67:359–67. Adams, J.C. 2004. Roles of fascin in cell adhesion and motility. Curr Opin Cell Biol. 16:590–6. Ahuja, R., R. Pinyol, N. Reichenbach, L. Custer, J. Klingensmith, M.M. Kessels, and B. Qualmann. 2007. Cordon-bleu is an actin nucleation factor and controls neuronal morphology. Cell. 131:337–50. Albrecht-Buehler, G. 1976. The function of filopodia in spreading 3T3 mouse fibroblasts. In Cell Motility. D. Pollard, R. Goldman, and J. Rosenbaum, editors. Cold Spring Harbor Laboratory, New York. 247–264. Ammer, A.G., and S.A. Weed. 2008. Cortactin branches out: roles in regulating protrusive actin dynamics. Cell Motil Cytoskeleton. 65:687–707. Anderson, T.W., A.N. Vaughan, and L.P. Cramer. 2008. Retrograde flow and myosin II activity within the leading cell edge deliver F-actin to the lamella to seed the formation of graded polarity actomyosin II filament bundles in migrating fibroblasts. Mol Biol Cell. 19:5006–18. Andrianantoandro, E., and T.D. Pollard. 2006. Mechanism of actin filament turnover by severing and nucleation at different concentrations of ADF/cofilin. Mol Cell. 24:13–23. Bear, J.E., and F.B. Gertler. 2009. Ena/VASP: Towards resolving a pointed controversy at the barbed end. J Cell Sci. 122:1947–53. Bear, J.E., T.M. Svitkina, M. Krause, D.A. Schafer, J.J. Loureiro, G.A. Strasser, I.V. Maly, O.Y. Chaga, J.A. Cooper, G.G. Borisy, and F.B. Gertler. 2002. Antagonism between Ena/VASP proteins and actin filament capping regulates fibroblast motility. Cell. 109:509–21. Benesch, S., S. Lommel, A. Steffen, T.E. Stradal, N. Scaplehorn, M. Way, J. Wehland, and K. Rottner. 2002. Phosphatidylinositol 4,5-biphosphate (PIP2)-induced vesicle movement depends on N-WASP and involves Nck, WIP, and Grb2. J Biol Chem. 277:37771–76. Benesch, S., S. Polo, F.P. Lai, K.I. Anderson, T.E. Stradal, J. Wehland, and K. Rottner. 2005. N-WASP deficiency impairs EGF internalization and actin assembly at clathrin-coated pits. J Cell Sci. 118:3103–15. Bensenor, L.B., H.M. Kan, N. Wang, H. Wallrabe, L.A. Davidson, Y. Cai, D.A. Schafer, and G.S. Bloom. 2007. IQGAP1 regulates cell motility by linking growth factor signaling to actin assembly. J Cell Sci. 120:658–69. Bentley, D., and A. Toroian-Raymond. 1986. Disoriented pathfinding by pioneer neurone growth cones deprived of filopodia by cytochalasin treatment. Nature. 323:712–15. Berg, J.S., and R.E. Cheney. 2002. Myosin-X is an unconventional myosin that undergoes intrafilopodial motility. Nat Cell Biol. 4:246–50.

26

J.V. Small and K. Rottner

Blaser, H., M. Reichman-Fried, I. Castanon, K. Dumstrei, F.L. Marlow, K. Kawakami, L. SolnicaKrezel, C.P. Heisenberg, and E. Raz. 2006. Migration of zebrafish primordial germ cells: a role for myosin contraction and cytoplasmic flow. Dev Cell. 11:613–27. Block, J., T.E. Stradal, J. Hanisch, R. Geffers, S.A. Kostler, E. Urban, J.V. Small, K. Rottner, and J. Faix. 2008. Filopodia formation induced by active mDia2/Drf3. J Microsc. 231:506–17. Breitsprecher, D., A.K. Kiesewetter, J. Linkner, C. Urbanke, G.P. Resch, J.V. Small, and J. Faix. 2008. Clustering of VASP actively drives processive, WH2 domain-mediated actin filament elongation. EMBO J. 27:2943–54. Brieher, W.M., M. Coughlin, and T.J. Mitchison. 2004. Fascin-mediated propulsion of Listeria monocytogenes independent of frequent nucleation by the Arp2/3 complex. J Cell Biol. 165:233–42. Brown, C.M., B. Hebert, D.L. Kolin, J. Zareno, L. Whitmore, A.R. Horwitz, and P.W. Wiseman. 2006. Probing the integrin-actin linkage using high-resolution protein velocity mapping. J Cell Sci. 119:5204–14. Burridge, K., and K. Wennerberg. 2004. Rho and Rac take center stage. Cell. 116:167–79. Cai, L., A.M. Makhov, D.A. Schafer, and J.E. Bear. 2008. Coronin 1B antagonizes cortactin and remodels Arp2/3-containing actin branches in lamellipodia. Cell. 134:828–42. Cain, R.J., and A.J. Ridley. 2009. Phosphoinositide 3-kinases in cell migration. Biol Cell. 101:13–29. Campellone, K.G., N.J. Webb, E.A. Znameroski, and M.D. Welch. 2008. WHAMM is an Arp2/3 complex activator that binds microtubules and functions in ER to Golgi transport. Cell. 134:148–61. Carlier, M.F., S. Wiesner, C. Le Clainche, and D. Pantaloni. 2003. Actin-based motility as a selforganized system: mechanism and reconstitution in vitro. C R Biol. 326:161–70. Charras, G., and E. Paluch. 2008. Blebs lead the way: How to migrate without lamellipodia. Nat Rev Mol Cell Biol. 9:730–36. Chesarone, M.A., and B.L. Goode. 2009. Actin nucleation and elongation factors: mechanisms and interplay. Curr Opin Cell Biol. 21:28–37. Chien, C.B., D.E. Rosenthal, W.A. Harris, and C.E. Holt. 1993. Navigational errors made by growth cones without filopodia in the embryonic Xenopus brain. Neuron. 11:237–51. Cosen-Binker, L.I., and A. Kapus. 2006. Cortactin: The gray eminence of the cytoskeleton. Physiology (Bethesda). 21:352–61. Czuchra, A., X. Wu, H. Meyer, J. van Hengel, T. Schroeder, R. Geffers, K. Rottner, and C. Brakebusch. 2005. Cdc42 is not essential for filopodium formation, directed migration, cell polarization, and mitosis in fibroblastoid cells. Mol Biol Cell. 16:4473–84. DeMali, K.A., K. Wennerberg, and K. Burridge. 2003. Integrin signaling to the actin cytoskeleton. Curr Opin Cell Biol. 15:572–82. Dent, E.W., A.V. Kwiatkowski, L.M. Mebane, U. Philippar, M. Barzik, D.A. Rubinson, S. Gupton, J.E. Van Veen, C. Furman, J. Zhang, A.S. Alberts, S. Mori, and F.B. Gertler. 2007. Filopodia are required for cortical neurite initiation. Nat Cell Biol. 9:1347–59. DePasquale, J.A., and C.S. Izzard. 1987. Evidence for an actin-containing cytoplasmic precursor of the focal contact and the timing of incorporation of vinculin at the focal contact. J Cell Biol. 105:2803–9. Derivery, E., J. Fink, D. Martin, A. Houdusse, M. Piel, T.E. Stradal, D. Louvard, and A. Gautreau. 2008. Free Brick1 is a trimeric precursor in the assembly of a functional wave complex. PLoS One. 3:e2462. Dickinson, R.B. 2009. Models for actin polymerization motors. J Math Biol. 58:81–103. Diez, S., G. Gerisch, K. Anderson, A. Muller-Taubenberger, and T. Bretschneider. 2005. Subsecond reorganization of the actin network in cell motility and chemotaxis. Proc Natl Acad Sci U S A. 102:7601–6. Disanza, A., A. Steffen, M. Hertzog, E. Frittoli, K. Rottner, and G. Scita. 2005. Actin polymerization machinery: the finish line of signaling networks, the starting point of cellular movement. Cell Mol Life Sci. 62:955–70.

1

Elementary Cellular Processes Driven by Actin Assembly

27

Faix, J., D. Breitsprecher, T.E. Stradal, and K. Rottner. 2009. Filopodia: Complex models for simple rods. Int J Biochem Cell Biol. 41:1656–64. Faix, J., and R. Grosse. 2006. Staying in shape with formins. Dev Cell. 10:693–706. Faix, J., and K. Rottner. 2006. The making of filopodia. Curr Opin Cell Biol. 18:18–25. Flanagan, L.A., J. Chou, H. Falet, R. Neujahr, J.H. Hartwig, and T.P. Stossel. 2001. Filamin A, the Arp2/3 complex, and the morphology and function of cortical actin filaments in human melanoma cells. J Cell Biol. 155:511–17. Frame, M.C. 2004. Newest findings on the oldest oncogene; how activated src does it. J Cell Sci. 117:989–98. Geese, M., J.J. Loureiro, J.E. Bear, J. Wehland, F.B. Gertler, and A.S. Sechi. 2002. Contribution of Ena/VASP proteins to intracellular motility of listeria requires phosphorylation and proline-rich core but not F-actin binding or multimerization. Mol Biol Cell. 13: 2383–96. Gerhardt, H., M. Golding, M. Fruttiger, C. Ruhrberg, A. Lundkvist, A. Abramsson, M. Jeltsch, C. Mitchell, K. Alitalo, D. Shima, and C. Betsholtz. 2003. VEGF guides angiogenic sprouting utilizing endothelial tip cell filopodia. J Cell Biol. 161:1163–77. Gey, G.O. 1954. Some aspects of the constitution and behavior of normal and malignant cells maintained in continuous culture. Harvey Lect. 50:154–229. Glacy, S.D. 1983. Subcellular distribution of rhodamine-actin microinjected into living fibroblastic cells. J Cell Biol. 97:1207–13. Goldman, R.K., and D.M. Knipe. 1973. Functions of cytoplasmic fibers in nonmuscle cell motility. In Cold Spring Harbor Symposium on Quantitative Biology. Cold Spring Harbor Laboratory, Cold Spring Harbor. 523–34. Goley, E.D., and M.D. Welch. 2006. The ARP2/3 complex: an actin nucleator comes of age. Nat Rev Mol Cell Biol. 7:713–26. Goode, B.L., and M.J. Eck. 2007. Mechanism and function of formins in the control of actin assembly. Annu Rev Biochem. 76:593–627. Hahne, P., A. Sechi, S. Benesch, and J.V. Small. 2001. Scar/WAVE is localised at the tips of protruding lamellipodia in living cells. FEBS Lett. 492:215–20. Han, Y.H., C.Y. Chung, D. Wessels, S. Stephens, M.A. Titus, D.R. Soll, and R.A. Firtel. 2002. Requirement of a vasodilator-stimulated phosphoprotein family member for cell adhesion, the formation of filopodia, and chemotaxis in dictyostelium. J Biol Chem. 277: 49877–87. Harris, A.K. 1973. Cell surface movements related to cell locomotion. In Locomotion of Tissue Cells, CIBA Foundation Symposium. Vol. 14. R. Porter and D.W. Fitzsimmons, editors. Elsevier, Amsterdam. 3–26. Hartwig, J.H., and P. Shevlin. 1986. The architecture of actin filaments and the ultrastructural location of actin-binding protein in the periphery of lung macrophages. J Cell Biol. 103:1007–20. Heasman, S.J., and A.J. Ridley. 2008. Mammalian Rho GTPases: new insights into their functions from in vivo studies. Nat Rev Mol Cell Biol. 9:690–701. Heath, J.P., and B.F. Holifield. 1991. Cell locomotion: New research tests old ideas on membrane and cytoskeletal flow. Cell Motil Cytoskeleton. 18:245–57. Heuser, J.E., and M.W. Kirschner. 1980. Filament organization revealed in platinum replicas of freeze-dried cytoskeletons. J Cell Biol. 86:212–34. Hoglund, A.S., R. Karlsson, E. Arro, B.A. Fredriksson, and U. Lindberg. 1980. Visualization of the peripheral weave of microfilaments in glia cells. J Muscle Res Cell Motil. 1:127–46. Hotulainen, P., and P. Lappalainen. 2006. Stress fibers are generated by two distinct actin assembly mechanisms in motile cells. J Cell Biol. 173:383–94. Hughes, A. 1953. The growth of embryonic neurites; a study of cultures of chick neural tissues. J Anat. 87:150–62. Huxley, H.E. 1957. The double array of filaments in cross-striated muscle. J Biophys Biochem Cytol. 3:631–48.

28

J.V. Small and K. Rottner

Innocenti, M., S. Gerboth, K. Rottner, F.P.L. Lai, M. Hertzog, T.E.B. Stradal, E. Frittoli, D. Didry, S. Polo, A. Disanza, S. Benesch, P.P. Di Fiore, M.F. Carlier, and G. Scita. 2005. Abi1 regulates the activity of N-WASP and WAVE in distinct actin-based processes. Nat Cell Biol. 7:969–76. Innocenti, M., A. Zucconi, A. Disanza, E. Frittoli, L.B. Areces, A. Steffen, T.E. Stradal, P.P. Di Fiore, M.F. Carlier, and G. Scita. 2004. Abi1 is essential for the formation and activation of a WAVE2 signalling complex. Nat Cell Biol. 6:319–27. Ishikawa, H., R. Bischoff, and H. Holtzer. 1969. Formation of arrowhead complexes with heavy meromyosin in a variety of cell types. J Cell Biol. 43:312–28. Iwasa, J.H., and R.D. Mullins. 2007. Spatial and temporal relationships between actin- filament nucleation, capping, and disassembly. Curr Biol. 17:395–406. Jacinto, A., S. Woolner, and P. Martin. 2002. Dynamic analysis of dorsal closure in Drosophila: From genetics to cell biology. Dev Cell. 3:9–19. Jenzora, A., B. Behrendt, J.V. Small, J. Wehland, and T.E. Stradal. 2005. PREL1 provides a link from Ras signalling to the actin cytoskeleton via Ena/VASP proteins. FEBS Lett. 579: 455–63. Jones, G.E., D. Zicha, G.A. Dunn, M. Blundell, and A. Thrasher. 2002. Restoration of podosomes and chemotaxis in Wiskott-Aldrich syndrome macrophages following induced expression of WASp. Int J Biochem Cell Biol. 34:806–15. Koestler, S.A., S. Auinger, M. Vinzenz, K. Rottner, and J.V. Small. 2008. Differentially oriented populations of actin filaments generated in lamellipodia collaborate in pushing and pausing at the cell front. Nat Cell Biol. 10:306–13. Koestler, S.A., K. Rottner, F. Lai, J. Block, M. Vinzenz, and J.V. Small. 2009. F- and G-actin concentrations in lamellipodia of moving cells. PLoS ONE. 4:e4810. Korobova, F., and T. Svitkina. 2008. Arp2/3 complex is important for filopodia formation, growth cone motility, and neuritogenesis in neuronal cells. Mol Biol Cell. 19:1561–74. Krause, M., J.D. Leslie, M. Stewart, E.M. Lafuente, F. Valderrama, R. Jagannathan, G.A. Strasser, D.A. Rubinson, H. Liu, M. Way, M.B. Yaffe, V.A. Boussiotis, and F.B. Gertler. 2004. Lamellipodin, an Ena/VASP ligand, is implicated in the regulation of lamellipodial dynamics. Dev Cell. 7:571–83. Krugmann, S., I. Jordens, K. Gevaert, M. Driessens, J. Vandekerckhove, and A. Hall. 2001. Cdc42 induces filopodia by promoting the formation of an IRSp53:Mena complex. Curr Biol. 11:1645–55. Kunda, P., G. Craig, V. Dominguez, and B. Baum. 2003. Abi, Sra1, and Kette control the stability and localization of SCAR/WAVE to regulate the formation of actin-based protrusions. Curr Biol. 13:1867–75. Kwiatkowski, A.V., D.A. Rubinson, E.W. Dent, J. Edward van Veen, J.D. Leslie, J. Zhang, L.M. Mebane, U. Philippar, E.M. Pinheiro, A.A. Burds, R.T. Bronson, S. Mori, R. Fassler, and F.B. Gertler. 2007. Ena/VASP is required for neuritogenesis in the developing cortex. Neuron. 56:441–55. Ladwein, M., and K. Rottner. 2008. On the Rho’d: the regulation of membrane protrusions by Rho-GTPases. FEBS Lett. 582:2066–74. Lai, F.P., M. Szczodrak, J. Block, J. Faix, D. Breitsprecher, H.G. Mannherz, T.E. Stradal, G.A. Dunn, J.V. Small, and K. Rottner. 2008. Arp2/3 complex interactions and actin network turnover in lamellipodia. EMBO J. 27:982–92. Lai, F.P., M. Szczodrak, J.M. Oelkers, M. Ladwein, F. Acconcia, S. Benesch, S. Auinger, J. Faix, J.V. Small, S. Polo, T.E. Stradal, and K. Rottner. 2009. Cortactin promotes migration and PDGF-induced actin reorganization by signaling to Rho-GTPases. Mol Biol Cell. 20:3209–23. Lammermann, T., B.L. Bader, S.J. Monkley, T. Worbs, R. Wedlich-Soldner, K. Hirsch, M. Keller, R. Forster, D.R. Critchley, R. Fassler, and M. Sixt. 2008. Rapid leukocyte migration by integrinindependent flowing and squeezing. Nature. 453:51–5. Lammermann, T., and M. Sixt. 2009. Mechanical modes of “amoeboid” cell migration. Curr Opin Cell Biol. 21:636–44.

1

Elementary Cellular Processes Driven by Actin Assembly

29

Langanger, G., J. de Mey, M. Moeremans, G. Daneels, M. de Brabander, and J.V. Small. 1984. Ultrastructural localization of alpha-actinin and filamin in cultured cells with the immunogold staining (IGS) method. J Cell Biol. 99:1324–34. Le Clainche, C., D. Schlaepfer, A. Ferrari, M. Klingauf, K. Grohmanova, A. Veligodskiy, D. Didry, D. Le, C. Egile, M.F. Carlier, and R. Kroschewski. 2007. IQGAP1 stimulates actin assembly through the N-WASP-Arp2/3 pathway. J Biol Chem. 282:426–35. Lee, J., A. Ishihara, and K. Jacobson. 1993. The fish epidermal keratocyte as a model system for the study of cell locomotion. In Symp. Soc. Expl Biol. XLVII, Cell Behaviour: Adhesion and Motility. G. Jones, C. Wigley, and R. Warn, editors. The Company of Biologists 73–89. Letourneau, P.C. 1981. Immunocytochemical evidence for colocalization in neurite growth cones of actin and myosin and their relationship to cell – substratum adhesions. Dev Biol. 85:113–22. Linardopoulou, E.V., S.S. Parghi, C. Friedman, G.E. Osborn, S.M. Parkhurst, and B.J. Trask. 2007. Human subtelomeric WASH genes encode a new subclass of the WASP family. PLoS Genet. 3:e237. Loisel, T.P., R. Boujemaa, D. Pantaloni, and M.F. Carlier. 1999. Reconstitution of actin-based motility of Listeria and Shigella using pure proteins. Nature. 401:613–16. Lommel, S., S. Benesch, K. Rottner, T. Franz, J. Wehland, and R. Kuhn. 2001. Actin pedestal formation by enteropathogenic Escherichia coli and intracellular motility of Shigella flexneri are abolished in N-WASP-defective cells. EMBO Rep. 2:850–57. Machesky, L.M., and R.H. Insall. 1998. Scar1 and the related Wiskott-Aldrich syndrome protein, WASP, regulate the actin cytoskeleton through the Arp2/3 complex. Curr Biol. 8:1347–56. Mahaffy, R.E., and T.D. Pollard. 2008. Influence of phalloidin on the formation of actin filament branches by Arp2/3 complex. Biochemistry. 47:6460–67. Maupin-Szamier, P., and T.D. Pollard. 1978. Actin filament destruction by osmium tetroxide. J Cell Biol. 77:837–52. Medalia, O., I. Weber, A.S. Frangakis, D. Nicastro, G. Gerisch, and W. Baumeister. 2002. Macromolecular architecture in eukaryotic cells visualized by cryoelectron tomography. Science. 298:1209–13. Mejillano, M.R., S. Kojima, D.A. Applewhite, F.B. Gertler, T.M. Svitkina, and G.G. Borisy. 2004. Lamellipodial versus filopodial mode of the actin nanomachinery: pivotal role of the filament barbed end. Cell. 118:363–73. Merrifield, C.J., B. Qualmann, M.M. Kessels, and W. Almers. 2004. Neural Wiskott Aldrich Syndrome Protein (N-WASP) and the Arp2/3 complex are recruited to sites of clathrinmediated endocytosis in cultured fibroblasts. Eur J Cell Biol. 83:13–8. Mitchison, T., and M. Kirschner. 1988. Cytoskeletal dynamics and nerve growth. Neuron. 1: 761–72. Mizutani, K., H. Miki, H. He, H. Maruta, and T. Takenawa. 2002. Essential role of neural WiskottAldrich syndrome protein in podosome formation and degradation of extracellular matrix in src-transformed fibroblasts. Cancer Res. 62:669–74. Mogilner, A. 2009. Mathematics of cell motility: have we got its number? J Math Biol. 58:105–34. Monypenny, J., D. Zicha, C. Higashida, F. Oceguera-Yanez, S. Narumiya, and N. Watanabe. 2009. Cdc42 and Rac family GTPases regulate mode and speed but not direction of primary fibroblast migration during platelet-derived growth factor-dependent chemotaxis. Mol Cell Biol. 29:2730–47. Mullins, R.D., J.A. Heuser, and T.D. Pollard. 1998. The interaction of Arp2/3 complex with actin: nucleation, high affinity pointed end capping, and formation of branching networks of filaments. Proc Natl Acad Sci U S A. 95:6181–86. Nakagawa, H., H. Miki, M. Nozumi, T. Takenawa, S. Miyamoto, J. Wehland, and J.V. Small. 2003. IRSp53 is colocalised with WAVE2 at the tips of protruding lamellipodia and filopodia independently of Mena. J Cell Sci. 116:2577–83. Nakai, J. 1956. Dissociated dorsal root ganglia in tissue culture. Am J Anat. 99:81–129. Nemethova, M., S. Auinger, and J.V. Small. 2008. Building the actin cytoskeleton: Filopodia contribute to the construction of contractile bundles in the lamella. J Cell Biol. 180:1233–44.

30

J.V. Small and K. Rottner

Nicholson-Dykstra, S.M., and H.N. Higgs. 2008. Arp2 depletion inhibits sheet-like protrusions but not linear protrusions of fibroblasts and lymphocytes. Cell Motil Cytoskeleton. 65:904–22. Niebuhr, K., F. Ebel, R. Frank, M. Reinhard, E. Domann, U.D. Carl, U. Walter, F.B. Gertler, J. Wehland, and T. Chakraborty. 1997. A novel proline-rich motif present in ActA of Listeria monocytogenes and cytoskeletal proteins is the ligand for the EVH1 domain, a protein module present in the Ena/VASP family. EMBO J. 16:5433–44. Nobes, C.D., and A. Hall. 1995. Rho, rac, and cdc42 GTPases regulate the assembly of multimolecular focal complexes associated with actin stress fibers, lamellipodia, and filopodia. Cell. 81:53–62. Nolen, B.J., N. Tomasevic, A. Russell, D.W. Pierce, Z. Jia, C.D. McCormick, J. Hartman, R. Sakowicz, and T.D. Pollard. 2009. Characterization of two classes of small molecule inhibitors of Arp2/3 complex. Nature. 460:1031–34. Oser, M., H. Yamaguchi, C.C. Mader, J.J. Bravo-Cordero, M. Arias, X. Chen, V. Desmarais, J. van Rheenen, A.J. Koleske, and J. Condeelis. 2009. Cortactin regulates cofilin and N-WASp activities to control the stages of invadopodium assembly and maturation. J Cell Biol. 186: 571–87. Otto, J.J., R.E. Kane, and J. Bryan. 1979. Formation of filopodia in coelomocytes: localization of fascin, a 58,000 dalton actin cross-linking protein. Cell. 17:285–93. Pellegrin, S., and H. Mellor. 2005. The Rho family GTPase Rif induces filopodia through mDia2. Curr Biol. 15:129–33. Pollard, T.D., and G.G. Borisy. 2003. Cellular motility driven by assembly and disassembly of actin filaments. Cell. 112:453–65. Quinlan, M.E., J.E. Heuser, E. Kerkhoff, and R.D. Mullins. 2005. Drosophila Spire is an actin nucleation factor. Nature. 433:382–88. Resch, G.P., K.N. Goldie, A. Hoenger, and J.V. Small. 2002. Pure F-actin networks are distorted and branched by steps in the critical-point drying method. J Struct Biol. 137:305–12. Ridley, A.J. 2006. Rho GTPases and actin dynamics in membrane protrusions and vesicle trafficking. Trends Cell Biol. 16:522–29. Ridley, A.J., H.F. Paterson, C.L. Johnston, D. Diekmann, and A. Hall. 1992. The small GTPbinding protein rac regulates growth factor-induced membrane ruffling. Cell. 70:401–10. Riedl, J., A.H. Crevenna, K. Kessenbrock, J.H. Yu, D. Neukirchen, M. Bista, F. Bradke, D. Jenne, T.A. Holak, Z. Werb, M. Sixt, and R. Wedlich-Soldner. 2008. Lifeact: A versatile marker to visualize F-actin. Nat Methods. 5:605–7. Roadcap, D.W., and J.E. Bear. 2009. Double JMY: making actin fast. Nat Cell Biol. 11:375–76. Rogers, S.L., U. Wiedemann, N. Stuurman, and R.D. Vale. 2003. Molecular requirements for actinbased lamella formation in Drosophila S2 cells. J Cell Biol. 162:1079–88. Rottner, K., B. Behrendt, J.V. Small, and J. Wehland. 1999a. VASP dynamics during lamellipodia protrusion. Nat Cell Biol. 1:321–22. Rottner, K., A. Hall, and J.V. Small. 1999b. Interplay between Rac and Rho in the control of substrate contact dynamics. Curr Biol. 9:640–48. Sahai, E., and C.J. Marshall. 2003. Differing modes of tumour cell invasion have distinct requirements for Rho/ROCK signalling and extracellular proteolysis. Nat Cell Biol. 5: 711–19. Samarin, S., S. Romero, C. Kocks, D. Didry, D. Pantaloni, and M.F. Carlier. 2003. How VASP enhances actin-based motility. J Cell Biol. 163:131–42. Schirenbeck, A., R. Arasada, T. Bretschneider, T.E. Stradal, M. Schleicher, and J. Faix. 2006. The bundling activity of vasodilator-stimulated phosphoprotein is required for filopodium formation. Proc Natl Acad Sci U S A. 103:7694–99. Schirenbeck, A., T. Bretschneider, R. Arasada, M. Schleicher, and J. Faix. 2005. The Diaphanousrelated formin dDia2 is required for the formation and maintenance of filopodia. Nat Cell Biol. 7:619–25. Sechi, A.S., and J. Wehland. 2004. ENA/VASP proteins: multifunctional regulators of actin cytoskeleton dynamics. Front Biosci. 9:1294–310.

1

Elementary Cellular Processes Driven by Actin Assembly

31

Small, J.V. 1981. Organization of actin in the leading edge of cultured cells: influence of osmium tetroxide and dehydration on the ultrastructure of actin meshworks. J Cell Biol. 91: 695–705. Small, J.V., S. Auinger, M. Nemethova, S. Koestler, K.N. Goldie, A. Hoenger, and G.P. Resch. 2008. Unravelling the structure of the lamellipodium. J Microsc. 231:479–85. Small, J.V., M. Herzog, M. Haner, and U. Abei. 1994. Visualization of actin filaments in keratocyte lamellipodia: negative staining compared with freeze-drying. J Struct Biol. 113:135–41. Small, J.V., G. Isenberg, and J.E. Celis. 1978. Polarity of actin at the leading edge of cultured cells. Nature. 272:638–39. Small, J.V., and G.P. Resch. 2005. The comings and goings of actin: coupling protrusion and retraction in cell motility. Curr Opin Cell Biol. 17:517–23. Small, J.V., G. Rinnerthaler, and H. Hinssen. 1982. Organization of actin meshworks in cultured cells: the leading edge. Cold Spring Harb Symp Quant Biol. 46 Pt 2:599–611. Small, J.V., K. Rottner, I. Kaverina, and K.I. Anderson. 1998. Assembling an actin cytoskeleton for cell attachment and movement. Biochim Biophys Acta. 1404:271–81. Small, J.V., T. Stradal, E. Vignal, and K. Rottner. 2002. The lamellipodium: Where motility begins. Trends Cell Biol. 12:112–20. Snapper, S.B., F. Takeshima, I. Anton, C.H. Liu, S.M. Thomas, D. Nguyen, D. Dudley, H. Fraser, D. Purich, M. Lopez-Ilasaca, C. Klein, L. Davidson, R. Bronson, R.C. Mulligan, F. Southwick, R. Geha, M.B. Goldberg, F.S. Rosen, J.H. Hartwig, and F.W. Alt. 2001. N-WASP deficiency reveals distinct pathways for cell surface projections and microbial actin-based motility. Nat Cell Biol. 3:897–904. Steffen, A., J. Faix, G.P. Resch, J. Linkner, J. Wehland, J.V. Small, K. Rottner, and T.E. Stradal. 2006. Filopodia formation in the absence of functional WAVE- and Arp2/3- complexes. Mol Biol Cell. 17:2581–91. Steffen, A., K. Rottner, J. Ehinger, M. Innocenti, G. Scita, J. Wehland, and T.E. Stradal. 2004. Sra1 and Nap1 link Rac to actin assembly driving lamellipodia formation. Embo J. 23:749–59. Stradal, T., K.D. Courtney, K. Rottner, P. Hahne, J.V. Small, and A.M. Pendergast. 2001. The Abl interactor proteins localize to sites of actin polymerization at the tips of lamellipodia and filopodia. Curr Biol. 11:891–95. Stradal, T.E., K. Rottner, A. Disanza, S. Confalonieri, M. Innocenti, and G. Scita. 2004. Regulation of actin dynamics by WASP and WAVE family proteins. Trends Cell Biol. 14:303–11. Stradal, T.E., and G. Scita. 2006. Protein complexes regulating Arp2/3-mediated actin assembly. Curr Opin Cell Biol. 18:4–10. Svitkina, T.M., and G.G. Borisy. 1999. Arp2/3 complex and actin depolymerizing factor/cofilin in dendritic organization and treadmilling of actin filament array in lamellipodia. J Cell Biol. 145:1009–26. Svitkina, T.M., E.A. Bulanova, O.Y. Chaga, D.M. Vignjevic, S. Kojima, J.M. Vasiliev, and G.G. Borisy. 2003. Mechanism of filopodia initiation by reorganization of a dendritic network. J Cell Biol. 160:409–21. Svitkina, T.M., A.B. Verkhovsky, and G.G. Borisy. 1995. Improved procedures for electron microscopic visualization of the cytoskeleton of cultured cells. J Struct Biol. 115:290–303. Swanson, N., and L.W. Swanson. 1995. Histology of the nervous system of man and vertebrates by S. Ramon Y. Cajal. Oxford University Press, Oxford. Takenawa, T., and S. Suetsugu. 2007. The WASP-WAVE protein network: Connecting the membrane to the cytoskeleton. Nat Rev Mol Cell Biol. 8:37–48. Tani, K., S. Sato, T. Sukezane, H. Kojima, H. Hirose, H. Hanafusa, and T. Shishido. 2003. Abl interactor 1 promotes tyrosine 296 phosphorylation of mammalian enabled (Mena) by c-Abl kinase. J Biol Chem. 278:21685–92. Taylor, A.C. 1966. Microtubules in the microspikes and cortical cytoplasm of isolated cells. J Cell Biol. 28:155–68. Taylor, A.C., and E. Robbins. 1963. Observations on microextensions from the surface of isolated vertebrate cells. Dev Biol. 7:660–73.

32

J.V. Small and K. Rottner

Tilney, L.G., S. Hatano, H. Ishikawa, and M.S. Mooseker. 1973. The polymerization of actin: its role in the generation of the acrosomal process of certain echinoderm sperm. J Cell Biol. 59:109–26. Tilney, L.G., and N. Kallenbach. 1979. Polymerization of actin. VI. The polarity of the actin filaments in the acrosomal process and how it might be determined. J Cell Biol. 81:608–23. Trichet, L., C. Sykes, and J. Plastino. 2008. Relaxing the actin cytoskeleton for adhesion and movement with Ena/VASP. J Cell Biol. 181:19–25. Trinkaus, J.P. 1984. Cells into organs: the forces that shape the embryo. Prentice Hall, Upper Saddle River, NJ. Tuxworth, R.I., I. Weber, D. Wessels, G.C. Addicks, D.R. Soll, G. Gerisch, and M.A. Titus. 2001. A role for myosin VII in dynamic cell adhesion. Curr Biol. 11:318–29. Urban, E., S. Jacob, M. Nemethova, G.P. Resch, and J.V. Small. 2010. Electron tomography reveals unbranched networks of actin filaments in lamellipodia. Nat Cell Biol. 12:429–35. van Rheenen, J., J. Condeelis, and M. Glogauer. 2009. A common cofilin activity cycle in invasive tumor cells and inflammatory cells. J Cell Sci. 122:305–11. Vauti, F., B.R. Prochnow, E. Freese, S.K. Ramasamy, P. Ruiz, and H.H. Arnold. 2007. Arp3 is required during preimplantation development of the mouse embryo. FEBS Lett. 581: 5691–97. Vignjevic, D., S. Kojima, Y. Aratyn, O. Danciu, T. Svitkina, and G.G. Borisy. 2006. Role of fascin in filopodial protrusion. J Cell Biol. 174:863–75. 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. Waterman-Storer, C.M., A. Desai, J.C. Bulinski, and E.D. Salmon. 1998. Fluorescent speckle microscopy, a method to visualize the dynamics of protein assemblies in living cells. Curr Biol. 8:1227–30. Wegner, A. 1976. Head to tail polymerization of actin. J Mol Biol. 108:139–50. Weiner, O.D., W.A. Marganski, L.F. Wu, S.J. Altschuler, and M.W. Kirschner. 2007. An actinbased wave generator organizes cell motility. PLoS Biol. 5:e221. Weisswange, I., T.P. Newsome, S. Schleich, and M. Way. 2009. The rate of N-WASP exchange limits the extent of ARP2/3-complex-dependent actin-based motility. Nature. 458: 87–91. Wicki, A., and V. Niggli. 2001. The Rho/Rho-kinase and the phosphatidylinositol 3-kinase pathways are essential for spontaneous locomotion of Walker 256 carcinosarcoma cells. Int J Cancer. 91:763–71. Wood, W., and P. Martin. 2002. Structures in focus--filopodia. Int J Biochem Cell Biol. 34:726–30. Woodrum, D.T., S.A. Rich, and T.D. Pollard. 1975. Evidence for biased bidirectional polymerization of actin filaments using heavy meromyosin prepared by an improved method. J Cell Biol. 67:231–37. Yamada, K.M., B.S. Spooner, and N.K. Wessells. 1971. Ultrastructure and function of growth cones and axons of cultured nerve cells. J Cell Biol. 49:614–35. Yamaguchi, H., M. Lorenz, S. Kempiak, C. Sarmiento, S. Coniglio, M. Symons, J. Segall, R. Eddy, H. Miki, T. Takenawa, and J. Condeelis. 2005. Molecular mechanisms of invadopodium formation: the role of the N-WASP-Arp2/3 complex pathway and cofilin. J Cell Biol. 168: 441–52. Yang, C., L. Czech, S. Gerboth, S. Kojima, G. Scita, and T. Svitkina. 2007. Novel roles of formin mDia2 in lamellipodia and filopodia formation in motile cells. PLoS Biol. 5:e317. Yoshida, K., and T. Soldati. 2006. Dissection of amoeboid movement into two mechanically distinct modes. J Cell Sci. 119:3833–44. Yumura, S., and T. Kitanishi-Yumura. 1990. Fluorescence-mediated visualization of actin and myosin filaments in the contractile membrane-cytoskeleton complex of Dictyostelium discoideum. Cell Struct Funct. 15:355–64. Yuruker, B., and V. Niggli. 1992. Alpha-actinin and vinculin in human neutrophils: reorganization during adhesion and relation to the actin network. J Cell Sci. 101(Pt 2):403–14.

1

Elementary Cellular Processes Driven by Actin Assembly

33

Zhang, H., J.S. Berg, Z. Li, Y. Wang, P. Lang, A.D. Sousa, A. Bhaskar, R.E. Cheney, and S. Stromblad. 2004. Myosin-X provides a motor-based link between integrins and the cytoskeleton. Nat Cell Biol. 6:523–31. Zicha, D., I.M. Dobbie, M.R. Holt, J. Monypenny, D.Y. Soong, C. Gray, and G.A. Dunn. 2003. Rapid actin transport during cell protrusion. Science. 300:142–45. Zimmermann, P., K. Meerschaert, G. Reekmans, I. Leenaerts, J.V. Small, J. Vandekerckhove, G. David, and J. Gettemans. 2002. PIP(2)-PDZ domain binding controls the association of syntenin with the plasma membrane. Mol Cell. 9:1215–25. Zuchero, J.B., A.S. Coutts, M.E. Quinlan, N.B. Thangue, and R.D. Mullins. 2009. p53-cofactor JMY is a multifunctional actin nucleation factor. Nat Cell Biol. 11:451–59.

Chapter 2

Coupling Membrane Dynamics to Actin Polymerization Shiro Suetsugu and Tadaomi Takenawa

Abstract WASP/WAVE family proteins are important regulators of the Arp2/3 complex, which causes exponential growth of actin filaments. WASP/WAVE proteins mediate actin polymerization for both cellular protrusions, such as filopodia and lamellipodia, and invaginations, such as coated pits for endocytosis. However, it had been unclear how the direction of actin polymerization for these topologically different structures is precisely regulated. Recently, the BAR domain superfamily members, which contain membrane-deforming or membrane-adaptor domains, were found to interact with the WASP/WAVE family proteins. These membranedeforming or membrane-adaptor domains contain BAR, EFC/F-BAR, and IMD/IBAR domains, which induce membrane invaginations or membrane protrusions. Due to the various geometries of the membranes bound by the BAR domain superfamily members, these proteins could connect specific membrane structures to actin filaments, mediated by the WASP/WAVE family proteins and the Arp2/3 complex.

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A Variety of Plasma Membrane Invaginations . . . . . . . . . . . . . . . . . Membrane Deformation by the BAR and EFC/F-BAR Domains . . . . . . . . . The BAR and EFC/F-BAR Domain-Containing Proteins are Connected to the actin Cytoskeleton Through N-WASP . . . . . . . . . . . . 2.5 IMD/I-Bar Domain Induces Outward Protrusions . . . . . . . . . . . . . . . . 2.6 Direction of Actin Polymerization Beneath the Membrane . . . . . . . . . . . . 2.7 Membrane Curvature-Dependent Actin Polymerization . . . . . . . . . . . . . 2.8 Conclusions: Activation of Signal Transduction Cascades by Membrane Curvature . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.1 2.2 2.3 2.4

36 36 39 43 46 48 50 51 52

S. Suetsugu (B) Laboratory of Membrane and Cytoskeleton Dynamics, Institute of Molecular and Cellular Biosciences, The University of Tokyo, Tokyo, Japan; PRESTO, Japan Science and Technology Agency, Saitama, Japan e-mail: [email protected] M.-F. Carlier (ed.), Actin-based Motility, DOI 10.1007/978-90-481-9301-1_2, C Springer Science+Business Media B.V. 2010

35

36

S. Suetsugu and T. Takenawa

2.1 Introduction The plasma membrane contains various membrane microstructures, ranging from ∼10 nm to several microns, that are supported by the cytoskeleton and/or membrane binding proteins. These fine structures of the plasma membrane include invaginations, such as caveolae, clathrin-coated pits, and others, as well as protrusions, such as filopodia, lamellipodia, and so on. These fine micro-membrane structures have their own specific morphological characteristics, which have been mainly defined by extensive electron micrographic studies. In most cases, actin filaments accompany these structures. However, it has been unclear how these characteristic structures are formed by the dynamic collaboration between the actin cytoskeleton and the plasma membrane. Recently, increasing numbers of membrane-deforming proteins connecting the actin cytoskeleton and the plasma membrane have been found. These proteins contain the BAR, EFC/F-BAR and IMD/I-BAR domains, which not only deform membranes but also bind to N-WASP/WAVE proteins. In this chapter, we will discuss the roles of these membrane-deforming proteins and the N-WASP/WAVE proteins in the formation of such structures (Table 2.1).

2.2 A Variety of Plasma Membrane Invaginations The clathrin-coated pit (CCP) is one of the most well characterized membrane invaginations at the plasma membrane. Clathrin-coated pits are characterized by an invagination surrounded by a dense lattice of clathrin, and are approximately 100–200 nm in diameter (Doherty and McMahon, 2009; Heuser, 1980; Perrais and Merrifield, 2005). Clathrin-mediated endocytosis plays an important role in receptor internalization, synaptic vesicle recycling, and somatic nutrient uptake. The formation of clathrin-coated vesicles involves three steps (Kaksonen et al., 2005; Merrifield et al., 2005). First, the clathrin coat assembles on the flat membrane with other proteins, such as AP-2, and captures the cargo to form a hemispherical clathrin-coated pit. Second, the CCP slowly invaginates, and the actin polymerizes. Third, scission proteins are recruited to the neck of the invaginated CCP, for the separation of the newly formed clathrin-coated vesicles from the plasma membrane. Epsins participate in the initial clathrin assembly and CCP formation steps (Ford et al., 2002), while dynamins play a role in the scission step (Praefcke and McMahon, 2004). Epsins also have membrane deforming ability, through the insertion of an alpha-helix into the membrane (Doherty and McMahon, 2009; Itoh and De Camilli, 2006). In clathrin-coated pits, the power of actin polymerization, mediated by the Arp2/3 activation, is required for the scission of the clathrin-coated pit from the plasma membrane. The actin polymerization is observed when the clathrin-coated pit detaches from the plasma membrane. The Bin-AmphiphysinRvs167 (BAR)-domain proteins are involved in multiple steps of endocytosis, and among these steps, their involvement in the scission step may be the most common function in clathrin-mediated endocytosis (Kaksonen et al., 2005; McMahon and Gallop, 2005) (Fig. 2.1a).

Form of actin filament

Branched

Branched Bundled Branched Branched Branched

Cellular structure

Clathrin-coated pit

Caveolae Filopodia Lamellipodia Phagocytosis Podosome

? N-WASP? WAVE WAVE N-WASP

N-WASP

WASP/WAVE

? Yes and No Yes Yes Yes

Yes

Arp2/3 involvement

IRSp53 IRSp53 IRSp53 FBP17

FBP17, CIP4, Toca-1 Amphiphysin SNX-9

BAR domain protein involvement

Table 2.1 Cellular structures and actin filaments

Cdc42 Rac Rac

Small GTPase

Invagination Protrusion Protrusion Protrusion Protrusion

Invagination

Protrusion or Invagination

2 Coupling Membrane Dynamics to Actin Polymerization 37

38

S. Suetsugu and T. Takenawa

A 1: Protein assembly

2: Invagination

3: Neck formation

4: Scission

BAR domain protein Amphiphysin dynamin Membrane Cytosol

Clathrin

EFC/F-BAR protein FBP17/Toca-1/CIP4

Actin polymerization N-WASP-WIP complex Arp2/3 complex

B

3-1: Protrusion with bundled filaments

1: Tiny Protrusion?

3-2: Protrusion with branched filaments ?

2: Protrusion

Membrane Cytosol ?

IMD/I-BAR family protein

Actin polymerization N-WASP-WIP complex WAVE complex Arp2/3 complex Bundling proteins

Fig. 2.1 Schematic representation of the formation of the plasma membrane structures. (a) The assembly of clathrin forms the invaginations that may be recognized by the EFC/F-BAR domain. The binding of EFC/F-BAR domain proteins could further induce the invagination/tubulation of the plasma membrane, as well as actin polymerization to constrict the tubules. The narrower tubules are eventually bound to the BAR domain. Actin polymerization and the dynamin recruited by the EFC/F-BAR and BAR domain proteins induce the scission of the vesicles from the plasma membrane. (b) In filopodia and lamellipodia formation, the first step equivalent to the clathrincoat assembly is unclear. However, the expression or the local increase of an IMD domain protein induces the membrane protrusions without actin filaments. Subsequently, the actin fills the tubules to form filopodia and lamellipodia. If the actin filaments in the membrane protrusions are bundled, then filopodia may be formed. If the branched actin filaments are formed by the Arp2/3 complex, then lamellipodia may be formed

Caveolae are also well-known structures at the plasma membrane, and are localized in the vicinity of the actin filaments that run beneath the plasma membrane. A detailed electron micrographic analysis revealed that one end of the branched filament is attached to caveolae (Richter et al., 2008). Caveolae are characterized by a flask-shaped invagination with a diameter of approximately 50–100 nm (Doherty and McMahon, 2009; Palade and Bruns, 1968; Parton and Simons, 2007; Rothberg

2

Coupling Membrane Dynamics to Actin Polymerization

39

et al., 1992; Yamada, 1955). It is important to note that the structures of caveolae are more stable than those of clathrin-coated pits, and that caveolae can function as scaffolds for the modulation of intracellular signal transduction, transporters, and/or channels (Fujimoto et al., 1992; Lisanti et al., 1994; Pani and Singh, 2009; Parton and Simons, 2007). Caveolin and Cavin family members are the major proteins in caveolae biogenesis. Caveolin is an integral membrane protein essential for caveolae formation, as determined by knockout studies (Drab et al., 2001; Rothberg et al., 1992). Cavin family proteins and caveolin are involved in caveolae biogenesis (Aboulaich et al., 2004; Bastiani et al., 2009; Hansen et al., 2009; Hill et al., 2008; Liu and Pilch, 2008; McMahon et al., 2009; Rothberg et al., 1992). There are four Cavin members, including polymerase I and transcription release factor (PTRF)/Cavin1, serum deprivation response (SDR) gene product (SDPR)/Cavin-2, SDR-related gene product that binds to C kinase (SRBC)/Cavin-3, and muscle-restricted coiledcoil protein (MURC)/Cavin-4. These proteins bind to negatively-charged lipids, especially phosphatidylserine. SDPR and the truncated form of PTRF can reportedly form tubular structures in cells (Hansen et al., 2009; Liu and Pilch, 2008). However, only PTPF was suggested to promote caveolae formation, among the four Cavin members (Bastiani et al., 2009). The mechanism of the formation of invaginations linked to the actin cytoskeleton has thus remained unclear. The depletion of cholesterol results in the flattening of caveolae. This characteristic of caveolae is common to the lipid raft, and the raft fraction contains caveolin-1. The actin cytoskeleton in the vicinity of caveolae is therefore considered to stabilize caveolae. The treatment of cells with an actin depolymerizing drug results in the mislocalization of caveolae (Mundy et al., 2002). A small percentage of caveolae are engaged in endocytosis, such as of SV40, in an actin-dependent manner (Pelkmans et al., 2001; Pelkmans et al., 2002). There are several less characterized invaginations in cells, such as those of the CLIC-GEEC pathway, muscle T-tubules, and others (Doherty and McMahon, 2009). The CLIC-GEEC pathway has deeper invaginations, in which the function of the BAR domain protein GRAF1 has been indicated (Lundmark et al., 2008). T-tubules are essential for calcium signaling in muscles, and the involvement of caveolin-3 and the BAR domain protein amphiphysin was demonstrated (Galbiati et al., 2001; Lee et al., 2002; Razzaq et al., 2001).

2.3 Membrane Deformation by the BAR and EFC/F-BAR Domains The structure of the Bin-Amphiphysin-Rvs167 (BAR) domain provided great advances in understanding how these finely organized membrane microstructures are generated (Peter et al., 2004). All of the domains that belong to the BAR domain superfamily form homodimers. BAR domain superfamily proteins deform membranes to a geometry that corresponds to the structures of the membrane-binding

40

S. Suetsugu and T. Takenawa

surface of the protein, and/or bind to the membrane that fits their structures, and thus function to generate specific membrane geometries (Doherty and McMahon, 2009; Itoh and De Camilli, 2006; Takenawa and Suetsugu, 2007). The structure of the BAR domain from amphiphysin was the first to be solved, and the protein was demonstrated to be important for membrane deformation. The BAR domain structure is a crescent-shaped dimer, in which the concave surface is positively charged (Fig. 2.2) (Peter et al., 2004). The positively-charged surface of the domain binds to the negatively-charged inner surface of the plasma membrane, mostly through phosphatidylserine. Furthermore, the BAR domains from endophilin and amphiphysin have hydrophobic amino acids on the concave surface or dimer ends (Gallop et al., 2006; Masuda et al., 2006; Peter et al., 2004). The hydrophobic amino acids on the concave surface are inserted into the membrane, thereby strengthening the interaction between the membrane and the BAR domain. The BAR domain of amphiphysin deforms the membrane into narrow tubular invaginations, which are considered to correspond to the location of amphiphysin function at the late stage of endocytic vesicle formation. The BAR domain of SNX9 has a BAR domain closely connected to the PX domain, but the BAR and PX domain units of SNX9 have broad phosphoinositide specificity (Yarar et al., 2008). The above mentioned BAR domains, such as in the amphiphysin, endophilin, and SNX9 containing proteins, have SH3 domains that bind to dynamin and N-WASP, providing the connection between the membrane shape and the actin polymerization machinery (Fig. 2.3) (Otsuki et al., 2002; Yamada et al., 2007; Yarar et al., 2008). The structure of the extended FER-CIP4 homology (EFC) or FCH and BAR (F-BAR) domain (hereafter, we refer to this as the EFC/F-BAR domain or F-BAR domain) from CIP4 and FBP17 was the first to be solved among the EFC/F-BAR domains (Fig. 2.2) (Shimada et al., 2007). Toca-1 has almost identical EFC/F-BAR domains to those of CIP4 and FBP17 (Itoh et al., 2005; Shimada et al., 2007; Tsujita et al., 2006). The EFC/F-BAR domain forms a crescent-shaped dimer, in which the concave surface binds to the membrane (Shimada et al., 2007). As for the BAR domain, the overexpression of the EFC/F-BAR domain protein fragment alone induces the tubular membrane invaginations in cells. Overexpression of fulllength FBP17 also induces tubulation. The curvature generated by the EFC/F-BAR domain of CIP4 and FBP17 was much larger than that induced by the BAR domains from amphiphysin or endophilin, and appears to correspond to the curvature of the initial stages of clathrin-coated pits. FBP17 and CIP4 function in the endocytosis of clathrin-coated pits, presumably by recruiting dynamin and actin polymerization machinery to the membrane curvature for clathrin-coated pits (Hartig et al., 2009; Itoh et al., 2005; Shimada et al., 2007; Tsujita et al., 2006). FBP17 and Toca-1 were shown to induce N-WASP and the Arp2/3 complex-mediated actin polymerization dependent on membrane curvature in a reconstitution assay (Fig. 2.3) (Takano et al., 2008). Pacsin/Syndapin forms one branch of the F-BAR domain protein family (Itoh et al., 2005; Takenawa and Suetsugu, 2007). The crystal structure of the pacsin F-BAR domain has a shallower concave surface, and the membrane tubules induced by the pacsin F-BAR are narrower than those induced by the F-BAR domains of FBP17, CIP4, and Toca-1 (Fig. 2.2) (Wang et al., 2009). The narrower diameter of

2

Coupling Membrane Dynamics to Actin Polymerization

A

EFC/F-BAR FBP17, Toca-1

EFC/F-BAR pacsin1, pacsin2

41

N-BAR IMD/RCB/I-BAR Amphiphysin IRSp53

K55 K56 K33 R35

K108

R113 K114

K137 R140

K130 K171 K142 K143 R145 K147

K161 K163

K166

B

EFC /FBA

C R

Invaginations such as endocytosis: BAR and EFC/F-BAR domain

Protrusions such as filipodia and lamellipodia IMD/RCB/I-BAR domain IMD/I-B AR

Cytosol Membrane

Fig. 2.2 (a) Structures of the BAR, EFC/F-BAR, and IMD/I-BAR domains. The structures are of FBP17, pacsin2, Drosophila amphiphysin, and IRSp53, respectively. The amino acids responsible for lipid binding are colored blue. Dashed lines indicate putative membrane contacts, as determined by amino-acid substitution studies. (b and c) The possible mechanism of membrane deformation, based on the findings of the EFC/F-BAR study and the IMD study. The spiral formed by the EFC/F-BAR protein is observed on the surface of the tubulated liposome. The EFC/F-BAR domain displays end to end and lateral homo-interactions between the dimers. These homo-interactions are considered to facilitate the sensing of membrane curvature and membrane deformation by the EFC/F-BAR domain. The binding of IMD to the inner surface of the protruding membrane was demonstrated. However, the implementation of such mechanisms in BAR-domain- and IMDinduced membrane deformations has not been demonstrated so far. The domains associated with the tubulated membrane are in the cytosol

42

S. Suetsugu and T. Takenawa

EFC/F-BAR domain

A

PACSIN3 (Hs) PACSIN2 (Hs) PACSIN1 (Hs) Synd (Dm) F45E1.7 (Ce) K08E3.3 (Ce) Imp2 (Sp) F09E10.8 (Ce) PSTPIP1 (Hs) CIP4 (Dm) Hypothetical protein (Dd) CIP4 (Hs) Nostrin (Hs) Toca-1 (Hs) FBP17 (Hs) Bzz1p (Sc) BZZ1 (Sp) srGAP1 (Hs) WRP (Hs)

IRTKS (Hs) IRSp53 (Hs)

IMD/I-BAR domain

Tuba (Hs) CG32082 (Dm) Hypothetical protein (Dd) EEN-B1/endophilin (Hs) HOB3 (Sp) EEN (Hs) Rvs167p (Sc) EEN-B2 (Hs) F58G6.1 (Ce) unc-57 endoA Amph (Ds) (Ce) (Dm) amphiphysin (Hs)

BAR domain

B

BAR or EFC/F-BAR or IMD/I-BAR domain

SH3

• membrane-binding • deformation • actin filament

• WASP/WAVE family • dynamin • Ena/VASP etc

C Amphiphysin1, 2 Bin1 Endophilin1, 2, 3 Rvs167p

BAR

SH3

BAR

SH3

Toca-1, FBP17, CIP4

EFC/F-BAR

HR1

Syndapin/Pacsin1, 2, 3

EFC/F-BAR

SH3

SH3

Bin2

BAR

PSTPIP1, 2

EFC/F-BAR

SH3

Bin3, Rvs161p

BAR

Nostrin

EFC/F-BAR

SH3

Nadrin, SH3BP1

BAR

Nwk1, 2, Bzz1p

EFC/F-BAR

srGAP1, FBP3, WRP, p115RhoGAP4

EFC/F-BAR

RhoGAP

Fes, Fer

EFC/F-BAR

SH2

FCHO1, 2

EFC/F-BAR

BAR

PX

SNX9, 18 SH3

ASAP1/Ddef1 Centaurin b1, 2, 5

SH3

SH3

SH3 SH3 Tyr-kinase

BAR PX

BAR

BAR PH

PTB

IRSp53 (BAIAP2) IMD/I-BAR

BAR

BAR

PH

ArfGAP

IRTKS(BAIAP2L1) IMD/I-BAR

SH3

CRIB

WH2-like

PDZ

Ank Ank Ank

PICK1

RhoGAP

BAR

SNX1, 2, 4, 5, 6, 7, 8

APPL1, 2

PH

SH3 WH2-like SH3 WH2-like

BAR

Tuba SH3 SH3 SH3

PH ArfGAP SH3

FLJ22582(BAIAP2L2) IMD/I-BAR

Ank Ank Ank

Oligopherenin1-l GRAG2 Arfaptin1, 2 ICA69

RhoGAP

RhoGEF

BAR

SH3

SH3

MIM-ABBA IMD/I-BAR

SH3

WH2

Fig. 2.3 WASP/WAVE binding proteins with BAR, EFC/F-BAR or IMD/I-BAR domains. (a) Phylogenetic analysis of BAR-, EFC/F-BAR- or IMD/I-BAR -domain-containing proteins that reportedly interact with WASP/WAVE. The yeast, Dictyostelium, C.elegans, Drosophila, and human homologues are shown. Proteins on blue, red or green lines reportedly bind to WASP/NWASP, WAVEs, or both, respectively. Sc: Saccharomyces cerevisiae. Sp: Schizosaccharomyces pombe. Dd: Dictyostelium discoideum. Ce: Caenorhabditis elegans. Dm: Drosophila melanogaster. Hs: Homo sapiens. (b) Typical domain structures of BAR, EFC/F-BAR and IMD/I-BAR domain containing proteins. The BAR, EFC/F-BAR or IMD/I-BAR domain is at the N terminus, and the SH3 domain is at the C-terminus. (c) Diagram of the domain structures of BAR-, EFC/F-BAR- and IMD/I-BAR -domain-containing proteins

2

Coupling Membrane Dynamics to Actin Polymerization

43

the tubules induced by pacsin2 suggests that pacsin2 is recruited to the clathrincoated pit at the late stage of clathrin coated vesicle fission, where narrower membrane tubules connect the plasma membrane to the coated vesicles (Fig. 2.1a). However, pacsin was reportedly not colocalized with clathrin (Modregger et al., 2000). Therefore, the F-BAR domain of pacsin2 forms a clearly distinct subfamily in the F-BAR family, and pacsin2 appears to be involved in different biogenesis of cellular organelles than the BAR or the F-BAR domains. The F-BAR domains of pacsin1 and pacsin2 have a hydrophobic insertion loop, as also found in the endophilin BAR domain and the MIM IMD/I-BAR domain (described later). Importantly, in contrast to the canonical F-BAR proteins, such as FBP17 and Toca1, pacsin2 overexpression did not cause membrane tubule formation in cells. The SH3 domain of pacsin/Syndapin also binds to N-WASP and dynamin (Kessels and Qualmann, 2002). These EFC/F-BAR domains bind to phosphatidylserine and PI(4,5)P2 , and deform artificial liposomes and cell membranes into tubules. The binding of the EFC/F-BAR domains to phosphatidylserine was confirmed by many laboratories, and is supported by the crystal structure (Henne et al., 2007; Itoh et al., 2005; Shimada et al., 2007; Tsujita et al., 2006; Wang et al., 2009). The binding of the EFC/F-BAR domain to PI(4,5)P2 is reasonable, based on its strong negative charge and its involvement in endocytosis, because many proteins involved in endocytosis bind to PI(4,5)P2 (Tsujita et al., 2006). The F-BAR domain of FCHo2 also forms a crescent-shaped dimer, but the curvature of its membrane binding concave surface is larger than that of EFC/F-BAR of CIP4 and FBP17 (Henne et al., 2007; Shimada et al., 2007). In contrast to the EFC/F-BAR of CIP4 and Toca-1, the lateral surface of F-BAR of FCHo2 is curved (Henne et al., 2007). The location of FCHo2 function in cells is still unclear. No functional domains have been assigned besides the F-BAR domain in FCHo1 or FCHo2.

2.4 The BAR and EFC/F-BAR Domain-Containing Proteins are Connected to the actin Cytoskeleton Through N-WASP Several kinds of invaginated structures are formed at the plasma membrane. The most representative structures are those of clathrin coated-pits and caveolae. The clathrin-coated pit is rapidly internalized into cells, by forming clathrin-coated vesicles. In the scission of the vesicles, actin polymerization appears to play important roles. Actin polymerization is also suggested to be involved in subsequent vesicle movement inside the cell. The endocytosis machinery includes many BAR superfamily proteins, and most of them bind to N-WASP as well as membrane scissor dynamin (Doherty and McMahon, 2009; Gundelfinger et al., 2003; Itoh and De Camilli, 2006; Itoh et al., 2005; Peter et al., 2004; Takenawa and Suetsugu, 2007; Tsujita et al., 2006). These membrane-binding proteins include proteins with BAR domains, such as amphiphysin, endophilin, SNX9, and PICK1, and with EFC/F-BAR domains, such as FBP17, CIP4, and Toca-1.

44

S. Suetsugu and T. Takenawa

Initial studies of syndapin II/pacsin2 revealed that the overexpression of the SH3 domain inhibits the endocytosis of transferrin (Kessels and Qualmann, 2002; Qualmann et al., 2000). This inhibition was apparently a consequence of titrating out dynamin or N-WASP from the invagination structures, because the C-terminal SH3 domain of pacsin2/Syndapin II binds to N-WASP and dynamin (Kessels and Qualmann, 2006). Among the proteins listed above, amphiphysin, endophilin, SNX9, FBP17, and Toca-1 reportedly bind to N-WASP through their SH3 domains, and activate N-WASP in a membrane-dependent manner. The other signaling cascades, such as those mediated by small GTPases, are integrated on N-WASP to regulate the Arp2/3 complex-mediated actin polymerization (Fig. 2.4) (Otsuki et al., 2002; Takano et al., 2008; Yamada et al., 2007; Yarar et al., 2007). PICK1 does not have an SH3 domain, but contains a PDZ domain. Interestingly, the BAR domain of PICK1 reportedly interacts with the Arp2/3 complex to suppress the nucleation of actin filaments, thereby suppressing the endocytosis of neurotransmitter receptors (Rocca et al., 2008; Suh et al., 2008). The yeast WASP homologue, Las17p, was identified in a screen of mutants defective in endocytosis (Kaksonen et al., 2005; Naqvi et al., 1998; Qualmann and Kelly, 2000). Las17p and Vrp1p, a yeast homologue of WIP, are recruited to clathrincoated pits in the early stage of endocytosis with Bzz1p and Rvs167p, the yeast F-BAR and BAR-containing proteins (Kaksonen et al., 2005; Naqvi et al., 1998; Soulard et al., 2002). Toca1 in mammalian cells also forms a protein complex with N-WASP and WIP (Ho et al., 2004). The recruitment of N-WASP and the involvement of the actin cytoskeleton in endocytosis have also been found in mammalian cells (Kaksonen et al., 2005; Merrifield et al., 2005). Therefore, the roles of WASP and N-WASP in endocytosis are well conserved from yeast to mammals. In mammalian cells, a detailed analysis of the time course of protein recruitment to the clathrin coated pit and subsequent internalization into cells, i.e. endocytosis, was performed (Merrifield et al., 2004; Shimada et al., 2007). N-WASP, the Arp2/3 complex, and FBP17 were maximally recruited to the clathrin coated pit when the clathrin disappeared or was endocytosed from the membrane. Interestingly, dynamin recruitment to clathrin was faster than that of N-WASP or the Arp2/3 complex, and was maximal when the amount of clathrin reached its maximum (Merrifield et al., 2004). Dynamin recruitment should be faster than that of FBP17, because N-WASP and FBP17 were recruited simultaneously to the clathrin coated pit (Shimada et al., 2007). Interestingly, the tubules induced by FBP17 were antagonized by the expression of dynamin, but this dynamin-mediated antagonism of tubulation was dependent on actin filaments (Itoh et al., 2005), suggesting that the actin filaments mediated by N-WASP are essential for the dynamin-mediated fission of tubulated membranes or clathrin-coated pits. The yeast homolog of dynamin, Vps1p, interacts with the actin filament binding protein Sla1p, and functions in actin cytoskeletal organization (Yu and Cai, 2004). Caveolae are cholesterol-rich invaginations of the plasma membrane. Caveolae are known to be linked to the actin cytoskeleton, but the linking mechanism is unclear. The treatment of cells with latrunculin B leads to the mislocalizaton of caveolin-1 and the accumulation of caveolin-1 at the center of the cell (Mundy

2

Coupling Membrane Dynamics to Actin Polymerization

45

direct binding to small GTPase

SH3 domain

WIP/ CR16/ WICH

N-WASP

WH1/EVH1 B

CRIB

PI(4,5)P2

G- Arp2/3 actin complex

Pro-rich

V V C A

Cdc42

binding to phosphoinositide (localization or optimization of Arp2/3 activation)

activation of the Arp2/3 complex (induction of actin polymerization)

binding to SH3 domains (including that of BAR domain superfamily) and profilin

formation of the protein complex (localization and stability)

PIP3 WHD/SHD

WAVE2 HSPC 300

Nck

B

Pro-rich

Abi

IRSp53

Nap1

Rac

V

C A

G- Arp2/3 actin complex

PIR121/Sra1

Rac

indirect binding to small GTPase

interaction with membrane or membrane binding proteins actin polymerization

Fig. 2.4 Domains and basic binding partners of N-WASP and WAVE2. N-WASP and WAVE2 are the ubiquitous isoforms of the WASP and WAVE families, respectively. The general function of each domain is shown. N-WASP has an N-terminal WASP-homology 1 (WH1) domain, where WIP, CR16, or WICH binds. WAVE2 has an N-terminal WAVE/SCAR homology domain (WHD/SHD) that mediates protein complex formation with HSPC300, Abi, Nap1, and PIR121/Sra1. The basic region (B) is common to both N-WASP and WAVE2, where phosphoinositides (PIP2 or PIP3) bind for protein localization or activation of the Arp2/3 complex. N-WASP has the Cdc42-Rac-interactive binding region (CRIB) for Cdc42 binding. WAVE2 binds to Rac through PIR121/Sra1 in the WAVE2 complex, and through IRSp53, which binds to the proline-rich (Pro-rich) region of WAVE2. N-WASP, WAVE2 and other WASP/WAVE family proteins have a proline-rich region for binding to SH3 domain proteins and profilin. The binding of the SH3 domains to N-WASP or WAVE2 contributes to the optimization of the Arp2/3 activation. The adaptor protein Nck also binds to Nap1. The C-terminal region is called the VCA region. The actin monomer (G-actin) binds to the Verproline-homology region (V). The Arp2/3 complex binds to the cofilin-homology-acidic region (CA). The simultaneous binding of G-actin and the Arp2/3 complex to the VCA region contributes to the activation of the Arp2/3 complex-mediated actin polymerization

46

S. Suetsugu and T. Takenawa

et al., 2002). The involvement of the BAR domain superfamily has not been clarified yet, but the raft fraction contains endophilin and N-WASP (Otsuki et al., 2002), suggesting its role in rafts or caveolae.

2.5 IMD/I-Bar Domain Induces Outward Protrusions There are several types of protrusive structures in cells. Most of these structures contain actin filaments, presumably for their stability and mechanical strength required for the force execution for cell motility or the uptake of extracellular materials. The best-characterized cellular protrusive structures are filopodia and lamellipodia (Koestler et al., 2008; Mattila and Lappalainen, 2008; Pollard and Borisy, 2003; Takenawa and Suetsugu, 2007) (also described in detail in the chapters by JV Small, K Rottner, and G Borisy). Filopodia are cellular protrusions with spike- or needle-like morphology, and contain bundled actin filaments. Lamellipodia are relatively flat cellular structures that protrude in the direction of cell movement. Branched actin filaments fill the inside of the lamellipodia. Protrusive structures driven by actin polymerization are also observed at phagocytosis. During phagocytosis, the protruding lamellipodialike structure surrounds the material that is incorporated into the cell. For a long time, the local cellular structural elements of these morphologies, filopodia and lamellipodia, were considered to be generated solely as a consequence of actin polymerization. WAVE2-mediated activation of the Arp2/3 complex is essential for the formation of branched actin filaments at lamellipodia (Takenawa and Suetsugu, 2007). The actin bundle formation at filopodia appears to be generated by several pathways, which are classified as Arp2/3 dependent and Arp2/3 independent (Mattila and Lappalainen, 2008). In the Arp2/3 dependent mechanisms, the branched filament is bundled by additional factors. In the Arp2/3 independent mechanisms, the unbranched filaments are bundled. These mechanisms appear to be sufficient for the generation of protrusions. Therefore, the discovery of membrane protrusions lacking actin filaments was surprising, and their existence in vivo is still in question. The second clarified BAR domain superfamily structure is from IRSp53 and MIM, which contain the IRSp53-MIM homology domain (IMD)/inverse-BAR (I-BAR) domain. The IMD/I-BAR domain (hereafter, we refer to this as IMD) binds to the membrane through its convex surface. Due to the inverted geometry of the membrane binding surface, the IMD domain is involved in the plasma membrane protrusions of filopodia and lamellipodia (Mattila et al., 2007; Millard et al., 2005; Scita et al., 2008; Suetsugu et al., 2006b). The binding of the IMD to the membrane on the inner surface of the tubules was confirmed by cryo-electron microscopy (Saarikangas et al., 2009). Most of the interaction occurred through phosphatidylserine, but a preference for PI(4,5)P2 and PI(3,4,5)P3 was observed for the IMD of IRSp53 (Saarikangas et al., 2009; Suetsugu et al., 2006b). The IMD from MIM also has a helix for insertion into the membrane (Saarikangas et al., 2009).

2

Coupling Membrane Dynamics to Actin Polymerization

47

Some IMDs reportedly interact with Rac, a small GTPase. The IMD of IRSp53 binds to the active form of Rac, whereas the IMD of MIM binds to the inactive form (Bompard et al., 2005; Miki et al., 2000; Suetsugu et al., 2006b). These differences in the affinities to the small GTPases may modulate the membrane binding of the IMDs. It is clear that the overexpression of the IMD fragment alone induces highly dynamic membrane protrusions that persist even in the presence of an actin polymerization inhibitor (Mattila et al., 2007; Suetsugu et al., 2006b; Yamagishi et al., 2004; Yang et al., 2009). When the full-length proteins are overexpressed, the induced protrusions contain actin filaments, presumably because the SH3 domain recruits proteins that bundle and/or induce the formation of actin filaments. However, several regions without actin filaments were observed in these protrusions. Interestingly, the SH3 domain of IRSp53 binds to the Arp2/3 activator, WAVE2, which plays essential roles in lamellipodium formation, and also to N-WASP, which is considered to function in filopodium formation and endocytosis (Lim et al., 2008; Miki et al., 2000; Suetsugu et al., 2006a). MIM has the Verproline/WH2 domain, and directly binds to actin (Machesky and Johnston, 2007; Mattila et al., 2003). The IMD itself has been proposed to have an actin-filament bundling activity, although the bundling activity is still controversial. IMD binding to the actin filament has been confirmed by several laboratories (Mattila et al., 2007; Millard et al., 2005; Suetsugu et al., 2006b). The IRSp53 SH3 domain also reportedly binds to dynamin (Lim et al., 2008), but the significance of dynamin in membrane protrusions is still unclear. IRSp53 also has a PDZ binding motif, and it binds to several proteins with the PDZ domain, which may be important for the assembly of some cellular structures (Hori et al., 2003, 2005; Massari et al., 2009; Soltau et al., 2004). Currently, only the IMDs are reportedly involved in membrane deformation for protrusions. Interestingly, IRSp53 is involved in both filopodium and lamellipodium formation, as suggested from the localization and also from the binding of WAVE2 and N-WASP, VASP, and Mena (Govind et al., 2001; Krugmann et al., 2001; Lim et al., 2008; Miki et al., 2000) (Fig. 2.4). An analysis with N-WASP knockout cells indicated that the IRSp53-mediated formation of filopodium-like protrusions requires N-WASP, but its Arp2/3 complex activating ability was not involved in the protrusion formation (Lim et al., 2008). The siRNA mediated knockdown of IRSp53 also revealed its role in lamellipodia formation (Suetsugu et al., 2006a). The membrane protrusions with cylinder-like shapes, induced by the IMD, may be extended by the dendritic Arp2/3-mediated branched actin filaments for the lamellipodia (Fig. 2.1b). However, it is very hard to observe these transient structures in cells. The lamellipodia-like structures induced by WAVE2 and IRSp53 are involved in phagocytosis (Abou-Kheir et al., 2008; Shi et al., 2005). Interestingly, the EFC/FBAR protein, FBP17, was shown to be involved in both phagocytosis and the formation of podosomes, invasive structures that degrade the extracellular matrix (Tsuboi et al., 2009).

48

S. Suetsugu and T. Takenawa

2.6 Direction of Actin Polymerization Beneath the Membrane Although actin polymerization is essential for endocytosis, the direction of actin polymerization, i.e., whether the barbed end is facing toward the endocytic vesicles or toward the plasma membrane, is not clear. In lamellipodia and filopodia, it is clear that the barbed ends are directed to the plasma membrane (Kaksonen et al., 2006; Pollard and Borisy, 2003; Svitkina and Borisy, 1999). Actin comet formation on vesicles with endocytic characteristics has been observed in cultured cells and Xenopus eggs after the expression of type I phosphatidylinositol phosphate 5-kinase, a treatment with pervanadate, a tyrosine phosphatase inhibitor, or a treatment with phorbol myristate acetate (PMA) (Rozelle et al., 2000; Taunton et al., 2000). N-WASP is localized at the vesicles of these actin comets. In addition, similar actin comets are induced by pathogens, such as Listeria and Shigella. Thus, actin polymerization appears to occur inwardly from the plasma membrane; the barbed ends are directed toward the vesicles travelling away from the plasma membrane (Taunton et al., 2000) (Fig. 2.5, Model A). A detailed electron micrograph of reconstituted actin comets revealed the barbed end facing toward plastic beads coated with the actin regulatory protein, Listeria ActA (Cameron et al., 2001). In this model, actin polymerization directly generates the force required for vesicle movement. Interestingly, there is a mutant strain of yeast that shows the accumulation of actin filaments at the site of endocytosis. The photo-bleaching analysis of actin in this yeast Sla2 mutant suggested that actin polymerization occurs toward the plasma membrane, i.e., the barbed end faces toward the plasma membrane, rather than the endocytic vesicles (Kaksonen et al., 2003). Therefore, the direction of actin polymerization for endocytosis seems to be the same as the direction of actin polymerization for protrusive lamellipodia formation (Fig. 2.5, Models B and D). In this case, the force generated by the actin polymerization may work for vesicle movement or for vesicle fission. For vesicle movement, the polymerizing pointed ends of the actin filaments may generate the direct force required for the comet-like vesicle movement (Kaksonen et al., 2005, 2006) (Fig. 2.5, Model B). For vesicle fission, the pushing of the plasma membrane by the actin polymerization may generate the force required for the fission of the vesicles from the plasma membrane (Fig. 2.5, Model B). The models of the direction of actin polymerization described above hypothesize that the polymerization pulls the vesicles toward or pushes them away from the plasma membrane. When we consider the tubulation ability of the BAR domain superfamily, the direction of actin polymerization toward the tubulated membrane may be more reasonable, because tubulation is mediated by the BAR superfamily proteins, and N-WASP should be recruited to the tubulated membrane surface (Fig. 2.5, Model C). The polymerizing actin filament may compress the arc-shaped membrane at the neck to form the narrower tubules as well as to cause the tension to separate the vesicles, in a similar manner as the arc-shaped membrane at lamellipodia. Alternatively or simultaneously, the polymerizing actin filament may twist the vesicles and tubules, thereby constricting the tubules and eventually pinching

2

Coupling Membrane Dynamics to Actin Polymerization Model for “inward” deformation such as endocytosis

49 Possible model for “outward” protrusion by analogy with inward deformation.

Force direction by actin polymerization New filaments

Cytosol

Arp2/3 Arp2/3 WASP/N-WASP complex Dimeric BAR/EFC/RCB SH3 Arp2/3 protein WASP/N-WASP complex Arp2/3 SH3 SH3 SH3 Dynamin etc

membrane

Arp2/3 WASP/N-WASP /WAVE complex

Arp2/3 WASP/N-WASP /WAVE VASP etc complex SH3 SH3

Dimeric BAR/EFC/RCB protein

Force generated by actin polymerization and dynamin causes fission of the deformed membrane.

B

C W A coSP Ar /W p2 mple A /3 x VE

Arp2/3

WASP/WAVE complex Arp2/3

SH3 WASP/WAVE complex

Ar p2

/3 Dynamin etc

WASP/WAVE complex

SH3

/3

WASP/WAVE complex

D

Arp2/3

SH3

2 Arp

SH3

SH3

Dynamin etc

/3

SH3

Dynamin etc

Arp2/3

/3

Arp2/3

2 Arp

Arp2/3

SH3

WASP/WAVE complex

Arp2/3 SH3

Arp2/3

2/3

Arp

Arp2/3

Arp2/3 Arp2/3

Ar p2

Arp2/3

Force generated by actin polymerization may extend deformed membrane.

WA coSP/W mp A lex VE

A

Arp2/3

Arp2/3

WASP/WAVE complex WASP/N-WASP /WAVE Arp2/3 complex

VASP etc

WASP/N-WASP /WAVE complex

Fig. 2.5 Functional models of the BAR, EFC/F-BAR or IMD/I-BAR-containing proteins and of WASP/WAVE proteins in inward or outward deformation of the membrane. WASP/WAVE proteins and the BAR, EFC/F-BAR or IMD/I-BAR-containing domains act as functional units that organize both the membrane and cytoskeleton at membrane locations with a specific curvature. In endocytosis, actin polymerization directs the barbed end toward the endocytic vesicles (A), toward the plasma membrane (B), or toward the tubulation membrane (C). In all cases, the BAR or EFC proteins recruit both N-WASP and dynamin to induce membrane fission. In model (A), the elongating barbed end pushes the vesicles for vesicle movement. In model (B), the elongating pointed end pushes the vesicles or the elongating barbed end toward the plasma membrane, generating the force required for vesicle fission. In model (C), the elongating barbed end pushes or twists the membrane tubules at the neck in a similar manner as the actin filaments at lamellipodia. Simultaneously, the force for vesicle fission may be applied by the tension caused by actin polymerization. The actin polymerization toward the tubules can be easily converted to the polymerization toward the vesicles or the membrane, by a small rotation of the direction of actin polymerization. For outward protrusions such as lamellipodia and filopodia, actin polymerization occurs outwardly: the barbed end is oriented toward the plasma membrane (model D), and may possibly be initiated by the IMD domain-containing proteins that can recruit N-WASP, WAVE, VASP and other proteins. If model (B) occurs, then it is still unclear how the endocytosis (model B) and the outward protrusion (model C) are differentially regulated in actin polymerization. Membrane-binding domains, such as BAR, EFC/F-BAR and RCB/IMD, may sense the curvature of the membrane to determine the direction of actin cytoskeleton reorganization. Small GTPases and protein kinases are not illustrated in the figure, but they regulate the activities of the WASP/WAVE family members to induce actin polymerization through the Arp2/3 complex

off the vesicles. It is interesting to note that dynamin causes twisting for scission of the tubules into vesicles. The longitudinal tension along the tubule caused by the anchoring of the membrane is required for dynamin-mediated vesicle fission (Roux et al., 2006). Thus, the actin filaments induced on the membrane tubules may provide the scaffold for dynamin-dependent fission of tubules. In this model (C), the ongoing formation of branched actin filaments can be easily adapted to the direction of vesicle movement or the plasma membrane after the scission of the vesicles.

50

S. Suetsugu and T. Takenawa

2.7 Membrane Curvature-Dependent Actin Polymerization The BAR domain superfamily structure determines the diameter of the tubulated membrane or liposomes. The BAR domain superfamily members have certain preferences for the membrane curvatures that they deform into tubules. The EFC/F-BAR domain of FBP17 prefers liposomes with a large diameter of more than approximately 500 nm, rather than small liposomes. In contrast, the BAR domain of amphiphysin has a preference for smaller liposomes (Shimada et al., 2007). Consistently, the induction of actin polymerization by FBP17 or Toca-1 is dependent on the size of the liposomes added to the actin polymerization assay system (Takano et al., 2008). However, neither the exact diameter of the liposome on which actin polymerization occurs nor the orientation of the EFC/F-BAR domain on the membrane upon the induction of actin polymerization was determined. The EFC/FBAR domain forms tubulated membranes by creating laterally contacting spirals of the EFC/F-BAR domain on the surface of the tubulated membrane (Frost et al., 2008; Shimada et al., 2007). The concave surface of the EFC/F-BAR domain binds to the tubules. However, at the initial stage of tubule formation, another side-lying interface of the EFC/F-BAR domain contacts the membrane (Frost et al., 2008). Full length Toca-1 and FBP17 generated tubulated membranes in a manner similar to that accomplished by the EFC/F-BAR domain alone. Therefore, full-length Toca-1 or FBP17 appears to interact with the membrane in a manner similar to that of the EFC/F-BAR domain fragment proteins. Actin polymerization reached a maximum within a few minutes. The tubulation of the membrane was observed within 2 min of an incubation in the presence or absence of the N-WASP-WIP complex, the Arp2/3 complex, and actin. This evidence indicates that actin polymerization and membrane tubulation apparently occur simultaneously. Therefore, there are two possible modes of actin filament formation. The first involves actin polymerization becoming induced upon the preferential binding of the EFC/F-BAR domain proteins to relatively large liposomes. In this case, the sidelying interface of the EFC/F-BAR domain may bind to the membrane. The second involves actin polymerization occurring on the tubulated liposomes decorated with Toca-1 or FBP17 proteins. In the latter case, the small liposome vesicles with a diameter of 150 nm or less are poor substrates for Toca-1 or FBP17, and thus actin polymerization was not induced. Importantly, the curvature-dependent actin polymerization was not achieved by the EFC/F-BAR domain alone. The membrane curvature dependence of actin polymerization was only examined for the FBP17 and Toca-1 proteins. The acidic amino acids adjacent to the SH3 domain are required for rapid actin polymerization by FBP17 or Toca-1. These acidic amino acids are conserved among FBP17, CIP4, and Toca1, but not among pacsin/syndapin, suggesting the weak actin polymerization induced by pacsin/syndapin in the presence of membranes. Furthermore, the curvature-dependent actin polymerization requires the direct binding of N-WASP to the membrane, through the basic region of N-WASP (Takano et al., 2008). Therefore, the curvature-dependent actin polymerization requires the association of both EFC/F-BAR proteins and the N-WASP-WIP complex with the

2

Coupling Membrane Dynamics to Actin Polymerization

51

membrane, in a spatially ordered manner. Cryo-electron microscopy revealed the existence of membrane surfaces between the spirals of the EFC/F-BAR domain (Frost et al., 2008), which may be utilized for the interaction with the basic region of N-WASP. When actin polymerization in cells was blocked by latrunculin treatment, the EFC/F-BAR-domain-containing proteins became localized to tube-like structures, indicating that N-WASP-mediated actin polymerization is important for the fission and subsequent movement of vesicles or for supporting the membrane structures to prevent excess invagination (Itoh et al., 2005; Tsujita et al., 2006). Interestingly, full-length pacsin/syndapin did not induce membrane invagination when overexpressed in cells (Itoh et al., 2005). However, the treatment of cells with latrunculin B induced membrane tubulation with pacsin/syndapin localization. Therefore, the actin filament appears to suppress the tubulation to maintain the plasma membrane and to prevent the plasma membrane structures from undergoing excess tubulation or invagination. Although the mechanism of the suppression of tubulation by actin polymerization is unknown, the actin filament beneath the plasma membrane appears to be converted, to support the tubular structures where pacsins or other EFC/F-BAR domain fragments are localized. The protrusions induced by the overexpression of the IMD are also highly dynamic, and sometimes rapidly extend without actin polymerization (Suetsugu et al., 2006b; Yang et al., 2009; Saarikangas, 2009, #1181). These structures may not be observed without the overexpression of the IMD domain. The coupling of membrane protrusions and actin dynamics in cells remains to be clarified in the future.

2.8 Conclusions: Activation of Signal Transduction Cascades by Membrane Curvature Molecules that regulate the actin cytoskeleton have been studied extensively in the past decade. The identification of the WASP/WAVE family proteins and of the Arp2/3 complex in the 1990s greatly enriched our understanding of how extracellular stimuli trigger rearrangements of the actin cytoskeleton. The exponential growth of actin filaments that are linked to each other by branching seems to generate the forces required for cell shape alterations. The same molecules, the WASP family proteins and the Arp2/3 complex, are involved in actin polymerization for both protrusions, such as lamellipodia and filopodia, and invaginations, such as endocytic vesicles. The generation of curvature and the production of phosphoinositides involved in signal transduction appear to be correlated in events such as endocytosis and other morphological changes. Therefore, the activation of signal molecules, such as small GTPases, and the generation of membrane curvature should occur simultaneously, especially during the formation of clathrin-coated pits following receptor activation (Itoh and De Camilli, 2006). N-WASP is known to bind to the SH3 domains of adaptor proteins, such as

52

S. Suetsugu and T. Takenawa

Grb2, Src, and Nck (Takenawa and Suetsugu, 2007). These adaptor proteins also accumulate at the regions of receptor activation. However, these adaptor molecules or small G proteins, such as Cdc42, have no spatial order of signal transduction, i.e. the direction of actin polymerization could be both inward or outward from the plasma membrane. Therefore, the mechanisms underlying the regulation of the direction of actin polymerization have been totally unclear. The discovery of the membrane-deforming ability of BAR domain-related membrane-binding proteins, including the BAR, EFC/F-BAR and IMD/I-BAR domains, suggests another means of achieving shape changes in cellular membranes. Interestingly, the BAR domainrelated proteins are apparently coupled to the WASP/WAVE family proteins. The protein complexes of the WASP/WAVE and BAR domain-related proteins seem to synergistically regulate the cytoskeleton and the membrane shape. The putative functional units of the WASP/WAVE proteins and the BAR domain-related proteins are involved in both outward protrusion and inward vesicle trafficking. Therefore, these two types of cell shape changes might have an identical origin. Currently, most of the BAR domain-related proteins have been implicated as functioning in endocytosis and associating with WASP proteins. Analyses of the EFC/F-BAR domain suggested that the actin polymerization induced by the BAR– WASP coupling has various degrees of output. Fewer molecules are associated with the WAVE proteins. The diverse characteristics of the membrane curvature induction, the membrane association, and the speed of actin polymerization provide the connections between various plasma membrane structures as well as cellular organelles to the cytoskeleton. The membrane curvature-dependent actin polymerization by FBP17 or Toca-1 and the N-WASP-WIP complex was dependent on the direct N-WASP binding to the membrane, achieved by the finely tuned spatial organization of their SH3 domains. Therefore, the curvature-dependent actin polymerization and signal transduction are not achieved by the BAR domain or the EFC/F-BAR domain alone. However, there are several examples where the membrane curvature regulates the signal transduction or the location of cellular events, without utilization of the BAR domain. ArfGAP1 was shown to be activated by the small diameter of vesicles (Bigay et al., 2003). Recently, membrane geometry has been demonstrated to provide important cues for the localization of the peripheral membrane protein SpoVM in Bacillus subtilis (Ramamurthi et al., 2009). Thus, the membrane-binding proteins with curvature-sensing modules may control the spatial direction of the results of signal transduction, such as actin polymerization.

References Abou-Kheir, W., Isaac, B., Yamaguchi, H., and Cox, D. (2008). Membrane targeting of WAVE2 is not sufficient for WAVE2-dependent actin polymerization: a role for IRSp53 in mediating the interaction between Rac and WAVE2. J Cell Sci 121: 379–390. Aboulaich, N., Vainonen, J.P., Stralfors, P., and Vener, A.V. (2004). Vectorial proteomics reveal targeting, phosphorylation and specific fragmentation of polymerase I and transcript release factor (PTRF) at the surface of caveolae in human adipocytes. Biochem J 383: 237–248.

2

Coupling Membrane Dynamics to Actin Polymerization

53

Bastiani, M., Liu, L., Hill, M.M., Jedrychowski, M.P., Nixon, S.J., Lo, H.P., Abankwa, D., Luetterforst, R., Fernandez-Rojo, M., Breen, M.R., Gygi, S.P., Vinten, J., Walser, P.J., North, K.N., Hancock, J.F., Pilch, P.F., and Parton, R.G. (2009). MURC/Cavin4 and cavin family members form tissue-specific caveolar complexes. J Cell Biol 185: 1259–1273. Bigay, J., Gounon, P., Robineau, S., and Antonny, B. (2003). Lipid packing sensed by ArfGAP1 couples COPI coat disassembly to membrane bilayer curvature. Nature 426: 563–566. Bompard, G., Sharp, S.J., Freiss, G., and Machesky, L.M. (2005). Involvement of Rac in actin cytoskeleton rearrangements induced by MIM-B. J Cell Sci 118: 5393–5403. Cameron, L.A., Svitkina, T.M., Vignjevic, D., Theriot, J.A., and Borisy, G.G. (2001). Dendritic organization of actin comet tails. Curr Biol 11: 130–135. Doherty, G.J., and McMahon, H.T. (2009). Mechanisms of Endocytosis. Annu Rev Biochem. Drab, M., Verkade, P., Elger, M., Kasper, M., Lohn, M., Lauterbach, B., Menne, J., Lindschau, C., Mende, F., Luft, F.C., Schedl, A., Haller, H., and Kurzchalia, T.V. (2001). Loss of caveolae, vascular dysfunction, and pulmonary defects in caveolin-1 gene-disrupted mice. Science 293: 2449–2452. Ford, M.G., Mills, I.G., Peter, B.J., Vallis, Y., Praefcke, G.J., Evans, P.R., and McMahon, H.T. (2002). Curvature of clathrin-coated pits driven by epsin. Nature 419: 361–366. Frost, A., Perera, R., Roux, A., Spasov, K., Destaing, O., Egelman, E.H., De Camilli, P., and Unger, V.M. (2008). Structural basis of membrane invagination by F-BAR domains. Cell 132: 807–817. Fujimoto, T., Nakade, S., Miyawaki, A., Mikoshiba, K., and Ogawa, K. (1992). Localization of inositol 1,4,5-trisphosphate receptor-like protein in plasmalemmal caveolae. J Cell Biol 119: 1507–1513. Galbiati, F., Engelman, J.A., Volonte, D., Zhang, X.L., Minetti, C., Li, M., Hou, H., Jr., Kneitz, B., Edelmann, W., and Lisanti, M.P. (2001). Caveolin-3 null mice show a loss of caveolae, changes in the microdomain distribution of the dystrophin-glycoprotein complex, and t-tubule abnormalities. J Biol Chem 276: 21425–21433. Gallop, J.L., Jao, C.C., Kent, H.M., Butler, P.J., Evans, P.R., Langen, R., and McMahon, H.T. (2006). Mechanism of endophilin N-BAR domain-mediated membrane curvature. EMBO J 25: 2898–2910. Govind, S., Kozma, R., Monfries, C., Lim, L., and Ahmed, S. (2001). Cdc42Hs facilitates cytoskeletal reorganization and neurite outgrowth by localizing the 58-kD insulin receptor substrate to filamentous actin. J Cell Biol 152: 579–594. Gundelfinger, E.D., Kessels, M.M., and Qualmann, B. (2003). Temporal and spatial coordination of exocytosis and endocytosis. Nat Rev Mol Cell Biol 4: 127–139. Hansen, C.G., Bright, N.A., Howard, G., and Nichols, B.J. (2009). SDPR induces membrane curvature and functions in the formation of caveolae. Nat Cell Biol 11: 807–814. Hartig, S.M., Ishikura, S., Hicklen, R.S., Feng, Y., Blanchard, E.G., Voelker, K.A., Pichot, C.S., Grange, R.W., Raphael, R.M., Klip, A., and Corey, S.J. (2009). The F-BAR protein CIP4 promotes GLUT4 endocytosis through bidirectional interactions with N-WASp and Dynamin-2. J Cell Sci 122: 2283–2291. Henne, W.M., Kent, H.M., Ford, M.G., Hegde, B.G., Daumke, O., Butler, P.J., Mittal, R., Langen, R., Evans, P.R., and McMahon, H.T. (2007). Structure and analysis of FCHo2 F-BAR domain: a dimerizing and membrane recruitment module that effects membrane curvature. Structure 15: 839–852. Heuser, J. (1980). Three-dimensional visualization of coated vesicle formation in fibroblasts. J Cell Biol 84: 560–583. Hill, M.M., Bastiani, M., Luetterforst, R., Kirkham, M., Kirkham, A., Nixon, S.J., Walser, P., Abankwa, D., Oorschot, V.M., Martin, S., Hancock, J.F., and Parton, R.G. (2008). PTRFCavin, a conserved cytoplasmic protein required for caveola formation and function. Cell 132: 113–124.

54

S. Suetsugu and T. Takenawa

Ho, H.Y., Rohatgi, R., Lebensohn, A.M., Le, M., Li, J., Gygi, S.P., and Kirschner, M.W. (2004). Toca-1 mediates Cdc42-dependent actin nucleation by activating the N-WASP-WIP complex. Cell 118: 203–216. Hori, K., Konno, D., Maruoka, H., and Sobue, K. (2003). MALS is a binding partner of IRSp53 at cell-cell contacts. FEBS Lett 554: 30–34. Hori, K., Yasuda, H., Konno, D., Maruoka, H., Tsumoto, T., and Sobue, K. (2005). NMDA receptor-dependent synaptic translocation of insulin receptor substrate p53 via protein kinase C signaling. J Neurosci 25: 2670–2681. Itoh, T., and De Camilli, P. (2006). BAR, F-BAR (EFC) and ENTH/ANTH domains in the regulation of membrane-cytosol interfaces and membrane curvature. Biochim Biophys Acta 1761: 897–912. Itoh, T., Erdmann, K.S., Roux, A., Habermann, B., Werner, H., and De Camilli, P. (2005). Dynamin and the actin cytoskeleton cooperatively regulate plasma membrane invagination by BAR and F-BAR proteins. Dev Cell 9: 791–804. Kaksonen, M., Sun, Y., and Drubin, D.G. (2003). A pathway for association of receptors, adaptors, and actin during endocytic internalization. Cell 115: 475–487. Kaksonen, M., Toret, C.P., and Drubin, D.G. (2005). A modular design for the clathrin- and actinmediated endocytosis machinery. Cell 123: 305–320. Kaksonen, M., Toret, C.P., and Drubin, D.G. (2006). Harnessing actin dynamics for clathrinmediated endocytosis. Nat Rev Mol Cell Biol 7: 404–414. Kessels, M.M., and Qualmann, B. (2002). Syndapins integrate N-WASP in receptor-mediated endocytosis. EMBO J 21: 6083–6094. Kessels, M.M., and Qualmann, B. (2006). Syndapin oligomers interconnect the machineries for endocytic vesicle formation and actin polymerization. J Biol Chem 281: 13285–13299. Koestler, S.A., Auinger, S., Vinzenz, M., Rottner, K., and Small, J.V. (2008). Differentially oriented populations of actin filaments generated in lamellipodia collaborate in pushing and pausing at the cell front. Nat Cell Biol 10: 306–313. Krugmann, S., Jordens, I., Gevaert, K., Driessens, M., Vandekerckhove, J., and Hall, A. (2001). Cdc42 induces filopodia by promoting the formation of an IRSp53: Mena complex. Curr Biol 11: 1645–1655. Lee, E., Marcucci, M., Daniell, L., Pypaert, M., Weisz, O.A., Ochoa, G.C., Farsad, K., Wenk, M.R., and De Camilli, P. (2002). Amphiphysin 2 (Bin1) and T-tubule biogenesis in muscle. Science 297: 1193–1196. Lim, K.B., Bu, W., Goh, W.I., Koh, E., Ong, S.H., Pawson, T., Sudhaharan, T., and Ahmed, S. (2008). The Cdc42 effector IRSp53 generates filopodia by coupling membrane protrusion with actin dynamics. J Biol Chem 283: 20454–20472. Lisanti, M.P., Scherer, P.E., Tang, Z., and Sargiacomo, M. (1994). Caveolae, caveolin and caveolinrich membrane domains: a signalling hypothesis. Trends Cell Biol 4: 231–235. Liu, L., and Pilch, P.F. (2008). A critical role of cavin (polymerase I and transcript release factor) in caveolae formation and organization. J Biol Chem 283: 4314–4322. Lundmark, R., Doherty, G.J., Howes, M.T., Cortese, K., Vallis, Y., Parton, R.G., and McMahon, H.T. (2008). The GTPase-activating protein GRAF1 regulates the CLIC/GEEC endocytic pathway. Curr Biol 18: 1802–1808. Machesky, L.M., and Johnston, S.A. (2007). MIM: a multifunctional scaffold protein. J Mol Med 85: 569–576. Massari, S., Perego, C., Padovano, V., D Amico, A., Raimondi, A., Francolini, M., and Pietrini, G. (2009). LIN7 mediates the recruitment of IRSp53 to tight junctions. Traffic 10: 246–257. Masuda, M., Takeda, S., Sone, M., Ohki, T., Mori, H., Kamioka, Y., and Mochizuki, N. (2006). Endophilin BAR domain drives membrane curvature by two newly identified structure-based mechanisms. EMBO J 25: 2889–2897. Mattila, P.K., and Lappalainen, P. (2008). Filopodia: molecular architecture and cellular functions. Nat Rev Mol Cell Biol 9: 446–454.

2

Coupling Membrane Dynamics to Actin Polymerization

55

Mattila, P.K., Pykalainen, A., Saarikangas, J., Paavilainen, V.O., Vihinen, H., Jokitalo, E., and Lappalainen, P. (2007). Missing-in-metastasis and IRSp53 deform PI(4,5)P2-rich membranes by an inverse BAR domain-like mechanism. J Cell Biol 176: 953–964. Mattila, P.K., Salminen, M., Yamashiro, T., and Lappalainen, P. (2003). Mouse MIM, a tissuespecific regulator of cytoskeletal dynamics, interacts with ATP-actin monomers through its C-terminal WH2 domain. J Biol Chem 278: 8452–8459. McMahon, H.T., and Gallop, J.L. (2005). Membrane curvature and mechanisms of dynamic cell membrane remodelling. Nature 438: 590–596. McMahon, K.A., Zajicek, H., Li, W.P., Peyton, M.J., Minna, J.D., Hernandez, V.J., Luby-Phelps, K., and Anderson, R.G. (2009). SRBC/cavin-3 is a caveolin adapter protein that regulates caveolae function. EMBO J 28: 1001–1015. Merrifield, C.J., Perrais, D., and Zenisek, D. (2005). Coupling between clathrin-coated-pit invagination, cortactin recruitment, and membrane scission observed in live cells. Cell 121: 593–606. Merrifield, C.J., Qualmann, B., Kessels, M.M., and Almers, W. (2004). Neural Wiskott Aldrich Syndrome Protein (N-WASP) and the Arp2/3 complex are recruited to sites of clathrinmediated endocytosis in cultured fibroblasts. Eur J Cell Biol 83: 13–18. Miki, H., Yamaguchi, H., Suetsugu, S., and Takenawa, T. (2000). IRSp53 is an essential intermediate between Rac and WAVE in the regulation of membrane ruffling. Nature 408: 732–735. Millard, T.H., Bompard, G., Heung, M.Y., Dafforn, T.R., Scott, D.J., Machesky, L.M., and Futterer, K. (2005). Structural basis of filopodia formation induced by the IRSp53/MIM homology domain of human IRSp53. EMBO J 24: 240–250. Modregger, J., Ritter, B., Witter, B., Paulsson, M., and Plomann, M. (2000). All three PACSIN isoforms bind to endocytic proteins and inhibit endocytosis. J Cell Sci 113 Pt 24: 4511–4521. Mundy, D.I., Machleidt, T., Ying, Y.S., Anderson, R.G., and Bloom, G.S. (2002). Dual control of caveolar membrane traffic by microtubules and the actin cytoskeleton. J Cell Sci 115: 4327–4339. Naqvi, S.N., Zahn, R., Mitchell, D.A., Stevenson, B.J., and Munn, A.L. (1998). The WASp homologue Las17p functions with the WIP homologue End5p/verprolin and is essential for endocytosis in yeast. Curr Biol 8: 959–962. Otsuki, M., Itoh, T., and Takenawa, T.T. (2002). N-WASP is recruited to rafts and associates with endophilin A in response to EGF. J Biol Chem 278: 6461–6469. Palade, G.E., and Bruns, R.R. (1968). Structural modulations of plasmalemmal vesicles. J Cell Biol 37: 633–649. Pani, B., and Singh, B.B. (2009). Lipid rafts/caveolae as microdomains of calcium signaling. Cell Calcium 45: 625–633. Parton, R.G., and Simons, K. (2007). The multiple faces of caveolae. Nat Rev Mol Cell Biol 8: 185–194. Pelkmans, L., Kartenbeck, J., and Helenius, A. (2001). Caveolar endocytosis of simian virus 40 reveals a new two-step vesicular-transport pathway to the ER. Nat Cell Biol 3: 473–483. Pelkmans, L., Puntener, D., and Helenius, A. (2002). Local actin polymerization and dynamin recruitment in SV40-induced internalization of caveolae. Science 296: 535–539. Perrais, D., and Merrifield, C.J. (2005). Dynamics of endocytic vesicle creation. Dev Cell 9: 581–592. Peter, B.J., Kent, H.M., Mills, I.G., Vallis, Y., Butler, P.J., Evans, P.R., and McMahon, H.T. (2004). BAR domains as sensors of membrane curvature: the amphiphysin BAR structure. Science 303: 495–499. Pollard, T.D., and Borisy, G.G. (2003). Cellular motility driven by assembly and disassembly of actin filaments. Cell 112: 453–465. Praefcke, G.J., and McMahon, H.T. (2004). The dynamin superfamily: universal membrane tubulation and fission molecules? Nat Rev Mol Cell Biol 5: 133–147.

56

S. Suetsugu and T. Takenawa

Qualmann, B., and Kelly, R.B. (2000). Syndapin isoforms participate in receptor-mediated endocytosis and actin organization. J Cell Biol 148: 1047–1062. Qualmann, B., Kessels, M.M., and Kelly, R.B. (2000). Molecular links between endocytosis and the actin cytoskeleton. J Cell Biol 150: F111–F116. Ramamurthi, K.S., Lecuyer, S., Stone, H.A., and Losick, R. (2009). Geometric cue for protein localization in a bacterium. Science 323: 1354–1357. Razzaq, A., Robinson, I.M., McMahon, H.T., Skepper, J.N., Su, Y., Zelhof, A.C., Jackson, A.P., Gay, N.J., and O Kane, C.J. (2001). Amphiphysin is necessary for organization of the excitation-contraction coupling machinery of muscles, but not for synaptic vesicle endocytosis in Drosophila. Genes Dev 15: 2967–2979. Richter, T., Floetenmeyer, M., Ferguson, C., Galea, J., Goh, J., Lindsay, M.R., Morgan, G.P., Marsh, B.J., and Parton, R.G. (2008). High-resolution 3D quantitative analysis of caveolar ultrastructure and caveola-cytoskeleton interactions. Traffic 9: 893–909. Rocca, D.L., Martin, S., Jenkins, E.L., and Hanley, J.G. (2008). Inhibition of Arp2/3-mediated actin polymerization by PICK1 regulates neuronal morphology and AMPA receptor endocytosis. Nat Cell Biol 10: 259–271. Rothberg, K.G., Heuser, J.E., Donzell, W.C., Ying, Y.S., Glenney, J.R., and Anderson, R.G. (1992). Caveolin, a protein component of caveolae membrane coats. Cell 68: 673–682. Roux, A., Uyhazi, K., Frost, A., and De Camilli, P. (2006). GTP-dependent twisting of dynamin implicates constriction and tension in membrane fission. Nature 441: 528–531. Rozelle, A.L., Machesky, L.M., Yamamoto, M., Driessens, M.H., Insall, R.H., Roth, M.G., LubyPhelps, K., Marriott, G., Hall, A., and Yin, H.L. (2000). Phosphatidylinositol 4,5-bisphosphate induces actin-based movement of raft-enriched vesicles through WASP-Arp2/3. Curr Biol 10: 311–320. Saarikangas, J., Zhao, H., Pykalainen, A., Laurinmaki, P., Mattila, P.K., Kinnunen, P.K., Butcher, S.J., and Lappalainen, P. (2009). Molecular mechanisms of membrane deformation by I-BAR domain proteins. Curr Biol 19: 95–107. Scita, G., Confalonieri, S., Lappalainen, P., and Suetsugu, S. (2008). IRSp53: crossing the road of membrane and actin dynamics in the formation of membrane protrusions. Trends Cell Biol 18: 52–60. Shi, J., Scita, G., and Casanova, J.E. (2005). WAVE2 Signaling Mediates Invasion of Polarized Epithelial Cells by Salmonella typhimurium. J Biol Chem 280: 29849–29855. Shimada, A., Niwa, H., Tsujita, K., Suetsugu, S., Nitta, K., Hanawa-Suetsugu, K., Akasaka, R., Nishino, Y., Toyama, M., Chen, L., Liu, Z.J., Wang, B.C., Yamamoto, M., Terada, T., Miyazawa, A., Tanaka, A., Sugano, S., Shirouzu, M., Nagayama, K., Takenawa, T., and Yokoyama, S. (2007). Curved EFC/F-BAR-domain dimers are joined end to end into a filament for membrane invagination in endocytosis. Cell 129: 761–772. Soltau, M., Berhorster, K., Kindler, S., Buck, F., Richter, D., and Kreienkamp, H.J. (2004). Insulin receptor substrate of 53 kDa links postsynaptic shank to PSD-95. J Neurochem 90: 659–665. Soulard, A., Lechler, T., Spiridonov, V., Shevchenko, A., Li, R., and Winsor, B. (2002). Saccharomyces cerevisiae Bzz1p is implicated with type I myosins in actin patch polarization and is able to recruit actin-polymerizing machinery in vitro. Mol Cell Biol 22: 7889–7906. Suetsugu, S., Kurisu, S., Oikawa, T., Yamazaki, D., Oda, A., and Takenawa, T. (2006a). Optimization of WAVE2-complex-induced actin polymerization by membrane-bound IRSp53, PIP3, and Rac. J Cell Biol 173: 571–585. Suetsugu, S., Murayama, K., Sakamoto, A., Hanawa-Suetsugu, K., Seto, A., Oikawa, T., Mishima, C., Shirouzu, M., Takenawa, T., and Yokoyama, S. (2006b). The RAC binding domain/IRSp53MIM homology domain of IRSp53 induces RAC-dependent membrane deformation. J Biol Chem 281: 35347–35358. Suh, Y.H., Pelkey, K.A., Lavezzari, G., Roche, P.A., Huganir, R.L., McBain, C.J., and Roche, K.W. (2008). Corequirement of PICK1 binding and PKC phosphorylation for stable surface expression of the metabotropic glutamate receptor mGluR7. Neuron 58: 736–748.

2

Coupling Membrane Dynamics to Actin Polymerization

57

Svitkina, T.M., and Borisy, G.G. (1999). Arp2/3 complex and actin depolymerizing factor/cofilin in dendritic organization and treadmilling of actin filament array in lamellipodia. J Cell Biol 145: 1009–1026. Takano, K., Toyooka, K., and Suetsugu, S. (2008). EFC/F-BAR proteins and the N-WASPWIP complex induce membrane curvature-dependent actin polymerization. EMBO J 27: 2817–2828. Takenawa, T., and Suetsugu, S. (2007). The WASP-WAVE protein network: connecting the membrane to the cytoskeleton. Nat Rev Mol Cell Biol 8: 37–48. Taunton, J., Rowning, B.A., Coughlin, M.L., Wu, M., Moon, R.T., Mitchison, T.J., and Larabell, C.A. (2000). Actin-dependent Propulsion of Endosomes and Lysosomes by Recruitment of N-WASP. J Cell Biol 148: 519–530. Tsuboi, S., Takada, H., Hara, T., Mochizuki, N., Funyu, T., Saitoh, H., Terayama, Y., Yamaya, K., Ohyama, C., Nonoyama, S., and Ochs, H.D. (2009). FBP17 mediates a common molecular step in the formation of podosomes and phagocytic cups in macrophages. J Biol Chem 284: 8548–8556. Tsujita, K., Suetsugu, S., Sasaki, N., Furutani, M., Oikawa, T., and Takenawa, T. (2006). Coordination between the actin cytoskeleton and membrane deformation by a novel membrane tubulation domain of PCH proteins is involved in endocytosis. J Cell Biol 172: 269–279. Wang, Q., Navarro, M.V., Peng, G., Molinelli, E., Lin Goh, S., Judson, B.L., Rajashankar, K.R., and Sondermann, H. (2009). Molecular mechanism of membrane constriction and tubulation mediated by the F-BAR protein Pacsin/Syndapin. Proc Natl Acad Sci USA 106: 12700–12705. Yamada, E. (1955). The fine structure of the renal glomerulus of the mouse. J Biophys Biochem Cytol 1: 551–566. Yamada, H., Ohashi, E., Abe, T., Kusumi, N., Li, S.A., Yoshida, Y., Watanabe, M., Tomizawa, K., Kashiwakura, Y., Kumon, H., Matsui, H., and Takei, K. (2007). Amphiphysin 1 is important for actin polymerization during phagocytosis. Mol Biol Cell 18: 4669–4680. Yamagishi, A., Masuda, M., Ohki, T., Onishi, H., and Mochizuki, N. (2004). A novel actin bundling/filopodium-forming domain conserved in insulin receptor tyrosine kinase substrate p53 and missing in metastasis protein. J Biol Chem 279: 14929–14936. Yang, C., Hoelzle, M., Disanza, A., Scita, G., and Svitkina, T. (2009). Coordination of membrane and actin cytoskeleton dynamics during filopodia protrusion. PLoS One 4: e5678. Yarar, D., Surka, M.C., Leonard, M.C., and Schmid, S.L. (2008). SNX9 activities are regulated by multiple phosphoinositides through both PX and BAR domains. Traffic 9: 133–146. Yarar, D., Waterman-Storer, C.M., and Schmid, S.L. (2007). SNX9 couples actin assembly to phosphoinositide signals and is required for membrane remodeling during endocytosis. Dev Cell 13: 43–56. Yu, X., and Cai, M. (2004). The yeast dynamin-related GTPase Vps1p functions in the organization of the actin cytoskeleton via interaction with Sla1p. J Cell Sci 117: 3839–3853.

Chapter 3

Endocytic Control of Actin-based Motility Andrea Disanza, Emanuela Frittoli, Chiara Giuliani, Francesca Milanesi, Andrea Palamidessi, Flavia Troglio, and Giorgio Scita

Abstract Endocytosis and recycling are emerging as essential components of the wiring enabling cells to perceive extracellular signals and resolve them in a temporally and spatially controlled fashion, directly influencing not only the duration and intensity of the signaling output, but also its correct location. One process, which requires the precise resolution of spatial information, is actin-based cell motility. This is achieved by coordinating membrane traffic, cell substrate adhesion, and actin remodeling in order to generate propulsive forces responsible for the formation of a diverse set of polarized migratory protrusions, the first steps of cell locomotion. Here, we will discuss how prototypical endocytic molecules control actin dynamics, frequently by linking the core machinery of actin polymerization to the plasma membrane. We will further discuss how endocytosis and recycling ensure spatial restriction of signaling to actin dynamics, thus enabling cells to migrate in response to different extracellular stimuli and in diverse microenvironments adopting diverse motile strategies, which have important implications in relevant physiological and pathological processes, first and foremost cancer cell invasion and dissemination.

Contents 3.1 Introduction: Mechanisms of Endocytic Control of Actin Based Cell Migration . . 3.1.1 Endocytosis Extinguishes the Signal at the Proper Time and Location . . . . 3.1.2 The Route Taken by Internalized Motogenic Receptor Influence Cell Migration . . . . . . . . . . . . . . . . . . . . 3.1.3 Endocytic/Recycling Cycle (EEC) Ensures Spatial Restriction of Signaling and Powers Directional Motility . . . . . . . . . . . . . . . . 3.2 Endocytic Molecules that Regulate Actin Dynamics: the case of Dynamin . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

60 60 62 63 65

G. Scita (B) IFOM, the FIRC Institute for Molecular Oncology Foundation, at the IFOM-IEO Campus, Via Adamello 16, 20139, Milan, Italy; Dipartimento di Medicina, Chirurgia ed Odontoiatria, Universita’ degli Studi di Milano, 20122 Milan, Italy. e-mail: [email protected] M.-F. Carlier (ed.), Actin-based Motility, DOI 10.1007/978-90-481-9301-1_3, C Springer Science+Business Media B.V. 2010

59

60

A. Disanza et al.

3.2.1 Physical and Functional Interactions of Dynamin and Dynamin-Interacting Proteins with the Actin Cytoskeleton . . . . . . . . . 3.2.2 Dynamin-Interacting Proteins and the Actin Cytoskeleton . . . . . . . . . 3.3 EEC Control of Signalling Circuitries that Orchestrate Polarized Cell Motility and Invasion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.1 Endocytosis and Recycling Ensure Spatial Restriction of Signaling Controlling Migratory Programs in Three Dimensions . . . . . 3.3.2 Endo/Exocytic Cycles of Adhesion Receptors in the Regulation of Migratory and Invasive Programs . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

66 68 71 73 77 79

3.1 Introduction: Mechanisms of Endocytic Control of Actin Based Cell Migration One essential feature of cells is their ability to respond to spatial information. This is generally achieved by adapting the architectural organization of their cytoskeletal and membrane components so as to maintain an asymmetric and polarized distribution of molecules, whose signaling output, thus, become spatially restricted. Accordingly, spatial restriction of signals has emerged as a critical device for the execution of a number of polarized cellular processes (reviewed in Disanza et al. 2009), including directed cell migration, growth cone movement, tissues morphogenesis during development, and cell invasion into the surrounding tissues of metastatic cells, which all depends on the resolution in a tight spatio-temporal control manner of signals that regulate the actin machineries.

3.1.1 Endocytosis Extinguishes the Signal at the Proper Time and Location A large body of evidence has defined the existence of a complex bi-univocal relationship between the components of membrane trafficking and the machineries controlling actin dynamics. It has become clear for instance (see for details Kaksonen et al. in this book), that forces generated by the actin tread milling are critical for the execution and completion of various initial steps of membrane internalization, endosomal vesicle formation, intracellular vesicle fusion and motility, and endomembrane recycling and exocytosis (Kaksonen 2008). On the other hand, however, there are multiple examples and mechanisms that can be envisioned through which the internalization and trafficking of membrane and membranebound molecules may control the localized output of signaling cascades, specifically controlling where and how actin polymerization and depolymerization occurs in cells, ultimately affecting not only directional motility but also the modes through which cell moves.

3

Endocytic Control of Actin-based Motility

61

For instance, by efficient internalization and intracellular transport of plasma membrane proteins and their associated signaling molecules, the endocytic process is commonly regarded as a critical mean to attenuate extracellular ligand-induced response, particularly those promoting directed cell migration. This is intuitively achieved by direct clearance of motogenic receptors, which function as the firstline sensors of the cellular microenvironment, from the plasma membrane limiting the intensities of the ensuing signaling events. In addition to this, internalized activated receptors are often destined to trafficking routes that lead to their degradation, thus hampering the ability of cells to respond to the continuous presence of extracellular cues by directly reducing the number of responsive receptors. A number of plasma membrane receptors, for example, including Receptor tyrosine kinases and G protein couple receptor (GPCR), function as motogenic sensors, which especially during developmental morphogenesis, respond to gradients of chemotactic factors that guide the migration of cells to their final destination. These cells, in addition to migrate directionally, must also be capable to arrest at their target sites, where the concentration of the chemotactic factors is the highest (Raz 2004). Thus, a ligand-dependent internalization/sorting mechanism that drives a motogenic receptor toward a degradative pathway may be critical to switch off the migratory signal and the ensuing actin polymerization/depolymerization cycles when appropriate. A scenario of this kind has been, for instance, demonstrated to operate during the migration of primordial germ cells toward the gonads in zebrafish development (Raz 2004). PGCs express the chemokine receptor CXCR4b and directionally migrate toward sites in the embryo at which the ligand SDF-1a is expressed (Raz 2004). Binding of SDF-1 elicit also internalization of the receptor promoting its spatial redistribution and restricting its signalling, two factors that are required for proper chemotaxis. Indeed an internalization defective receptor led to aberrantly elevated signals, increased time spent “running”, preventing cells to reach their final target, while promoting ectopic cell migration (Minina et al. 2007). Recently, an additional non-cell-autonomous endocytic mechanism has been shown to contribute to the directed migration of PGC (Boldajipour and Raz 2007). The somatic cells surrounding the germ cells were found to express another SDF-1 receptor, CXCR7. Binding of the “somatic” receptor to SDF-1 was shown not much to enhance cell signaling, rather to clear the ligand from the extracellular environment ensuring that SDF-1 protein does not spread too far from the source. In other words, CXCR7 would clear SDF-1 protein from areas that are no longer needed as sources of attractant. In the absence of CXCR7, SDF-1 protein would spread further and be maintained for longer, resulting in the aberrant guidance of germ cells. How does SDF-1 control cell-autonomously polarized cell migration? An obvious mechanism for controlling cell polarization and migration involves increased actin polymerization at the leading edge of migrating cells [reviewed in (Pollard 2003)]. An additional mechanism has also been proposed suggesting that hydrostatic pressure in the cytoplasm constitutes the driving force for deformations of the cell surface by the generation of protrusions or blebs (Charras et al. 2005). According to this model, myosin-dependent contractility at the cell cortex generates local hydrostatic pressure (Charras et al. 2005) or ruptures in the cortex

62

A. Disanza et al.

(Sheetz et al. 2006) that lead to membrane detachment from the cytoskeleton and flow of cytoplasm that expands the cellular protrusion. It was also shown that actomyosin contraction could be regulated by an influx of calcium ions into the cell presumably by activating calcium-dependent kinases that mediate myosin contractility [reviewed in (Matsumura 2005)]. Thus, the precise region where a bleb forms could depend on the level of local hydrostatic pressure (Charras et al. 2005) or on the position where contractility generates breaks in the cell cortex as well as on the local strength of the connection between the membrane and the actin-based cortical cytoskeleton. Consistent with these notions, in migrating PGC polymerized actin is not enriched underneath the expanding cell membrane. Additionally, their mode of motility resembles the one typical observed in amoeboid moving cells, such Dictyostelium Discoideum or human leukocytes (Garcia and Parent 2003), and appear to be driven by the extension of blebbing-like protrusion powered by cytoplasmic flow, rather than canonical flat lamellipodia (see also below). Finally, the blebbing protrusions are formed at sites of higher levels of free calcium where activation of myosin contraction occurs. Thus, one reasonable scenario is that polarized activation of the receptor CXCR4 leads to a rise in free calcium that in turn activates myosin contraction in the part of the cell responding to higher levels of the ligand SDF-1. The biased formation of new protrusions in a particular region of the cell in response to SDF-1 defines the leading edge and the direction of cell migration (Blaser et al. 2006).

3.1.2 The Route Taken by Internalized Motogenic Receptor Influence Cell Migration A large body of emerging evidence points to the fact that signaling is not restricted to the plasma membrane. As internalization proceeds, activated transmembrane molecules, with their tails exposed toward the cell cytoplasm, are confined into endomembrane organelles, which thus become bona fide signaling platform influencing not only the time and amplitude of the resulting signal, but also its specificity (Sorkin and Goh 2009). Consistent with this view, more and more signal transduction pathways are reported to require an active endocytic machinery, or strikingly to originate from various types of endosomes. A variation and an extension of this latter concept has further emerged with the recognition that plasma membrane receptors are internalized through different pathways, e.g. Clathrin-mediated endocytosis (CME) or non-Clathrin endocytosis (NCE), which have been shown to control directly the biological outcome of their signaling (Le Roy and Wrana 2005). During the endocytic transport, molecules undergo a discrete set of route-dependent posttranslational modifications, such as phosphorylation/dephosphorylation and ubiquitination, that directly influence the composition of the signaling cascade that is being activated (Sorkin and Goh 2009). Thus, the biological output might be controlled not only through compartmentalization of signaling into endosomal platforms, but also by the routes through which molecules reach the different

3

Endocytic Control of Actin-based Motility

63

compartments. In keeping with this latter notion, it has been recently shown that the stimulation of one of the most potent fibroblastic motogenic factor, PDGF and its cognate receptors, can elicit either proliferation or cell migration depending on the ligand concentrations and the routes taken by the activated receptor. Cells stimulated with low doses of PDGF responded by rearranging their cytoskeleton and acquiring migratory properties in a CME-dependent manner. Mitogenesis and cell proliferation were, instead, preferentially induced by higher concentration of PDGF, a response that required a functional raft/caveolin pathway (De Donatis et al. 2008). In this case, the precise mechanisms and molecular targets that control actin-based machineries responsible for directional motility are yet to be fully defined. It is, however, tempting to speculate that a low dose of PDGF stimulation, in analogy to EGF (Sigismund et al. 2008), may preferentially promote recycling of PDGFR or of its key signal transducers back to the plasma membrane, triggering positive feedback loops that ensure localized and sustain signaling to actin-based dynamics. Within this context, the members of the RhoGTPases, and in particular Rac proteins, which are known to control multiple critical events (e.g. de novo actin nucleation through the Nucleation Promoting Factors (NPF), WAVEs, Capping proteins, and F-actin depolymerization, through ADF/cofilin (Stradal and Scita 2006) concomitantly required for actin dynamics, may represent ideal cargos that hitchhike a trafficking ride back to the plasma membrane resulting in spatial-restriction and polarization of their signaling output.

3.1.3 Endocytic/Recycling Cycle (EEC) Ensures Spatial Restriction of Signaling and Powers Directional Motility Endocytic internalization of membrane and membrane-associated proteins is frequently accompanied by recycling of these factors back to the plasma membrane (Fig. 3.1). This process may function to replenish ligand-free receptor for the next round of signaling and transport. Alternatively, the internalization/recycling cycle can also serve either as a mean to redirect and confine signaling molecules to specialized and distinct areas of the plasma membrane, such as the apical and basal membrane of polarized epithelial monolayer (Bryant and Mostov 2008), or as a positive feedback mechanism capable of maintaining the polarization state of critical signaling molecules, such the small GTPase Cdc42 during the formation of polarized buds in budding yeast (Birtwistle and Kholodenko 2009). Finally, endocytosis of molecules is accompanied by the internalization and recycling of plasma membrane generating a constant membrane flow. In analogy to the actin tread milling cycle, this flow of membrane had early been proposed to either generate forces that support the extension of migratory protrusions (Bretscher and Aguado-Velasco 1998), or to promote the rearward movement of molecules bound to the surface of many cell types, as it may occur in cell protrusions of motile cells. Results consistent with membrane flow have been obtained from HeLa cells (Bretscher 1983) and fibroblasts (Schmoranzer et al. 2003), where membrane

64

A. Disanza et al. Clathrin mediated endocytosis (CME)

Non-clathrin endocytosis (NCE) LR

Dynamin accessory proteins

Polarized protrusions

Fast Recycling

Polarized protrusions

Exocyst Rab4 Rab11

Arf6

Rab5

Sl

ow

Re

MHC-I

EE

cy

cli

ng

Ligand

Rab8

Lipid Rafts (LR) Rab7

RTKs AP-2 Clathrin

LE/MVB

Lysosome

Ub

Ubiquitin

Fig. 3.1 The endocytic trafficking map. Signaling receptors, like other integral membrane proteins, enter the endocytic pathway and undergo sorting to various membrane compartments. Endocytosis regulates cell signaling by controlling the number of specific receptors available for ligand-induced activation in the plasma membrane, and the activation of receptors or downstream effectors often stimulates rapid endocytosis of the receptors (Doherty and McMahon 2009; Sorkin 2004). There is mounting evidence that virtually every family of signaling receptors enter through a Clathrin-mediated pathway (CME). Consistently, the mechanisms by which proteins involved in this process recruits cargos into newly formed clathrin pit (CCP) and subsequently form clathrincoated vesicels (CCV) have been subject of intense scrutiny. AP-2 and Dynamin act to promote vesicle scission aided by actin polymerization. Receptors can also interact with accessory proteins, which, in turn, bind to Clathrin heavy chain and/or AP-2 (Doherty and McMahon 2009). There is also evidence for Non-Clathrin-Mediated endocytosis (NCE) of certain receptors including GPCRs, RTKs, and TGF band Notch receptors. These endocytic entry routes include also raft-mediated pathways. Notably, ubiquitination of receptors (e.g TGF-b and EGFR) mediated by the E3 ligases preferentially direct these cargos toward NCE and degradative pathways, albeit the underlying molecular mechanisms responsible for this routing are ill-defined (Sorkin and Goh 2009; Di Guglielmo et al. 2003). Both CDE and NCE converge into Rab5-positive early endosomes (EE), which represent the first endosomal sorting station. From EE cargos can be redirected through a fast recycling, Rab4-dependent routes back to the plasma membrane, or enter, via Rab8, into a Rab11-endocytic recycling compartment (ERC) before being retargeted to the PM (Stenmark 2009). While recycling has been considered a default cargo route from endosomes, it is now emerging as a mechanims for polarized delivery of regulatory molecules, ensuring spatial restriction of actin dynamics (Disanza et al. 2009). Alternatively, cargo can traffic to a Rab7-dependent lysosomal degradative route, by entering into multivesicular body/late endosome (MVB/LE) before being degraded into lysosome. In addition to these “canonical recycling pathways”, cargos, such as MHC-1 or the interleukin receptor that do not enter through CME, can be sorted into Arf6positive recycling pathways (Arf6-RE) and redelivered back to the PM (Donaldson 2005). Fusion of recycling vesicle from Rab11- and Arf6-dependent pathways is in part mediated by the Exocyst

3

Endocytic Control of Actin-based Motility

65

vesicles can be seen approaching and fusing with the leading edge. However, experiments with leukocytes and keratocytes failed to detect any significant rearward membrane flow (Kucik et al. 1990), suggesting that this is not a universal property of moving cells, although it may be important for some of them. Recently, the requirement of a continuous flow of membranes propelled by key endocytic molecules, such as Clathrin, was shown to be essential for the extremely dynamic changes of cell shape that occur during directional, chemotactic migration of the amoeba Dyctiostelium Discoideum, a professional mover (Wessels et al. 2000; Traynor and kay 2007). A similar role of CME in controlling cell shape is also at play in a variety of migratory cells, from mammals to zebrafish (Palamidessi et al. 2008), supporting the existence of a tight coupling between membrane dynamics, cell shape changes and polarized signaling. Thus, endocytosis plays an important role in spatially-restricting signaling controlling actin dynamics, ultimately affecting cell locomotion in response to chemotactic gradient in two dimensions or the different motility programs utilized by cell when moving in three dimensional matrices. In this chapter, we will describe prototypical examples of endocytic molecules, which exert a direct or indirect control on actin dynamics. We will than illustrate how endocytic and recycling cycles and polarized exocytosis of membrane and membrane associated actin regulators ensure spatial restriction of actin-based signaling, thus orchestrating cell motility.

3.2 Endocytic Molecules that Regulate Actin Dynamics: the case of Dynamin Eukaryotic cells employ endocytosis to internalise plasma membrane, surface receptors, and various extracellular soluble molecules, including nutrients. The complexity of endocytosis is underscored by the existence of multiple entry routes, Clathrin-mediated endocytosis (CME) (Doherty and McMahon 2009) being the most extensively characterised, but increasing attention being dedicated to several non-Clathrin endocytosis (NCE) mechanisms (Mayor and Pagano 2007) (Fig. 3.1). Additionally endocytosis of membrane receptors is governed by a core machinery (Doherty and McMahon 2009), which is in turn controlled by a wealth of accessory regulatory proteins (Doherty and McMahon 2009). Core proteins, such as Clathrin or Dynamin, exert specific role mediating key steps of the endocytic process that include membrane bending, vesicle formation, vesicle scission, and movement of the newly formed vesicle into the cell cytoplasm (Doherty and McMahon 2009). Conversely accessory proteins, (e.g. Epsins, Eps15, AP-2) either assist the core proteins in the execution of their tasks or provide the specificity and diversification to the various routes and modes of internalization and are frequently emerging as cargo specific factors (Sorkin and von Zastrow 2009). Finally, GTPase of the Rab family, control distinct trafficking steps along the endocytic routes and between various vesicular organelle (such as Early, late and recycling endosomes and multivesicular body and lysosomes) (Fig. 3.1) [reviewed in (Stenmark 2009)].

66

A. Disanza et al.

In performing these various tasks, endocytic proteins must deal with and often harness the actin cytoskeleton, and to this end they have developed various ways to directly or indirectly interact with the actin machinery. It is therefore not surprising that an increasing number of prototypical endocytic or accessory molecules are reported to directly associate with actin, influencing its polymerization and depolymerization cycles, ultimately contributing to couple membrane and actin dynamics. In the following chapter, We will outline few exemplar cases supporting this notion focusing primarily on Dynamin and its interacting proteins. It is worth noting that there are numerous, additional emerging examples of proteins, which are not only capable to associate directly with membranes, but also to sense and promote their curvatures, such as the large family of BAR -containing proteins (Gallop and McMahon 2005). Remarkably a number of these molecules in addition to deform membrane into invaginating pits and tubules through their BAR domain are also capable of binding, via diverse protein:protein interaction surfaces, actin remodelling complexes (Frost et al. 2008). These molecules have been the subjects of recent, excellent review to which the reader is referred to (Gallop and McMahon 2005; Frost et al. 2008). Additionally they are described in details in other chapters of this book, and therefore will not be illustrated here. It is sufficient, here, to point out that the unique ability of a number of this class of proteins to connect and control the activity of actin regulatory complexes (Gallop and McMahon 2005) coupled with their membrane deforming properties render them ideal candidates to orchestrate actin and membrane dynamics, predicting that they play an essential role in promoting actin-base cell migration and motility.

3.2.1 Physical and Functional Interactions of Dynamin and Dynamin-Interacting Proteins with the Actin Cytoskeleton Dynamin is a GTPase that plays a critical role in endocytosis, by determining the fission of vesicles from the plasma membrane (Praefcke and McMahon 2004). A number of disparate observations functionally links Dynamin to the actin cytoskeleton (reviewed in (Schafer 2004; Kruchten and McNiven 2006)). In addition, expression of a dominant negative mutant of Dynamin-1 was shown to cause a redistribution of actin stress fibers to the cell cortex (Damke et al. 1994). Further links were obtained with studies of podosomes and actin comet tails. In the former case, a Dynamin isoform, Dynamin-2aa, was found to colocalize with filamentous actin at membrane ruffles (Cao et al. 1998) and at podosomes. A temperature sensitive mutant of Dynamin-2aa disrupted podosome formation (Ochoa et al. 2000). Interestingly, another Dynamin mutant (K44A), which is impaired in GTP hydrolysis and potently inhibits endocytosis, only delayed actin dynamics at podosomes, but did not prevent their formation (Ochoa et al. 2000). This latter result suggests that the interference of Dynamin with actin remodeling is not due to secondary effects of an endocytic block, but rather to a direct functional role on the actin cytoskeleton.

3

Endocytic Control of Actin-based Motility

67

A similar scenario may also be at play during the internalization of enteropathogenic Escherichia coli (EPEC) into cells. Dynamin2, in this latter case, has been implicated in the formation of actin-based pedestal that are utilized by these bacteria to attach and enter into to human intestinal epithelial cells (Unsworth et al. 2007). Dynamin2 is found in EPEC pedestal, where it is required for the recruitment of N-WASP, Arp3 and cortactin, indicating that Dynamin2 is an integral component of the signaling cascade leading to actin polymerization in these structures. Two reports identified Dynamin 1 and 2 as components of actin comet tails generated by Listeria, an intracellular pathogen known to utilize an actin tail for movement within the cytoplasm, and by type I PIP kinase. An inactive (K44A) Dynamin mutant significantly reduced comet number, length, velocity, and efficiency of movement, suggesting that Dynamin is part of a protein network that controls nucleation of actin from the membranes (Orth et al. 2002; Lee and Camilli; 2002). Despite the variety of examples that implicate Dynamin in actin remodellingbased processes, the molecular details of the physical links of Dynamin to the actin cytoskeleton is unknown. An emerging view is that Dynamin, through direct and indirect interactions with proteins that stimulate de novo actin filament assembly, recruits the cellular machinery for assembling actin filaments to membranes undergoing remodelling. Alternatively, Dynamin may function as a physical link between actin filaments and membranes through interactions with proteins like mAbp1 (Kessels et al. 2001) or cortactin (McNiven et al. 2000; Schafer et al. 2002) (see also below). Such a linkage might be particularly relevant for endomembranes that lack an elaborate cortical filament network like that underlying the plasma membrane. Furthermore, Dynamin assembled on membranes could be a structural component that stabilizes or organizes the actin filaments associated with membranes. Finally, recent work implicates Dynamin in the regulation of the localization and activity of Rac1 (Palamidessi et al. 2008; Schlunck et al. 2004). If Dynamin2 function is linked with that of Rho-family small GTPases, Dynamin2 could globally influence the dynamics of actin filaments during cell migration, membrane traffic and as cells maintain their polarity. Within this context a further intriguing “signaling connection” emerged from the functional characterization of the putative tumor suppressor protein nm23H1. Nm23 belongs to a set of structurally conserved nucleoside diphosphatase kinases (NDK), a family of enzymes that synthesize triphosphates from their respective nucleotide diphosphates (Wagner and Vu 1995). In Drosophila, nm23, regulates synaptic vesicle internalization at a stage where the function of the Dynamin GTPase activity is required (Krishnan et al. 2001), suggesting a model whereby the NDK activity of nm23 is required locally to increase the concentration of GTP, in order to facilitate the loading of Dynamin, which has an unusually low affinity for GTP. Thus, interference with the function of nm23 should impair receptor internalization, resulting in prolonged exposure of the receptor in the plasma membrane and sustained signaling, a possibility compatible with its postulated metastasis suppressor role. This effect may be further strengthened through an additional activity of nm23H1, which

68

A. Disanza et al.

also inhibits Tiam-1-induced production of GTP-bound Rac-1 (Otsuki et al. 2001). Tiam-1 is a Rac specific GEF originally identified as an invasion and metastasisinducing gene (Habets et al. 1994). Nm23h1 interacts directly with Tiam-1, and blocks its GEF activity in vivo (Otsuki et al. 2001). Thus, given the inhibitory effects exerted by GTP-loaded Rac in receptor internalization (see below), the lack of nm23h1 may result in enhanced Rac activation contributing to the prolonged permanence of receptors on the plasma membrane. Concomitantly, Rac activation may enhance actin-based cell migration, thereby further contributing to the onset of the phenotypic features typical of metastatic cells. One important, but unanswered question with respect to the role of Dynamin in actin dynamics is whether its GTPase activity also has a critical and more direct role in modulating the actin assembly machinery. Some clues in this direction came from a recent report supporting the notion that Dynamin GTPase activity promotes actin filament remodelling in conjunction with the actin binding protein Cortactin. Cortactin comprises an amino-terminal acidic domain, mediating Arp2/3-complex binding, followed by the F-actin binding region containing six and a half 37 amino acid repeats, an α-helical domain, a proline-rich region, and a Src homology (SH3) domain at the distal carboxy terminus (Cosen-Binker and Kapus 2006). Biochemically, Cortactin was shown to promote actin assembly by simultaneously binding to Arp2/3-complex and actin filaments (Cosen-Binker and Kapus 2006). Furthermore, through its SH3 domain Cortactin can associate with Dynamin (McNiven et al. 2000) promoting its GTPase activity and regulating actin assembly and filament organization (Schafer et al. 2002). Importantly, the GTPase activity of Dynamin appears essential to execute this latter function (Mooren et al. 2009). Tightly associated actin filaments crosslinked by Dynamin2 and cortactin became loosely associated after GTP addition when viewed by transmission electron microscopy. Moreover, real-time monitoring by Total Internal Reflection Microscopy (TIRF) revealed that filaments unraveled and fragmented after addition of GTP to Dynamin2/Cortactin-containing solution, indicating a potent filament remodeling activity exerted in concert by the two proteins. Filaments remodeled by Dynamin2/Cortactin also displayed enhanced sensitivity to the severing activity of ADF/cofilin, suggesting that Dynamin may favor the accessibility of proteins to actin filaments. These observations were accompanied by the findings that removal of Dynamin2 by RNAi caused a global remodeling of F-actin structures, further suggesting that the Dynamin/Cortactin complex through its ability to reorganize actin filament may orchestrate a variety of actin-based processes.

3.2.2 Dynamin-Interacting Proteins and the Actin Cytoskeleton Clues as to the molecular mechanisms through which actin dynamics might influence the function of Dynamin are also emerging from studies of Dynamininteracting proteins. Syndapin, for example, binds to numerous proteins involved in the endocytic process (Dynamin, synaptojanins and synapsins) and in the

3

Endocytic Control of Actin-based Motility

69

regulation of actin dynamics (N-Wasp, Sos-1) (Kessels and Qualmann 2006). Accordingly, Syndapin affects both endocytosis and actin cytoskeleton by a mechanism that coordinates vesicle fission and actin nucleation, thereby triggering a burst of actin polymerization around the vesicle neck which favours detachment from the plasma membrane and propulsion into the cytoplasm. Amphiphysin I and II represent another class of Dynamin-binding proteins with a dual involvement in endocytosis and actin cytoskeleton. Amphiphysins are thought to target Dynamin to the plasma membrane (reviewed in (Zhang and Zelhof 2002)). In addition, expression of their SH3 domains reduces receptor endocytosis, probably by preventing the formation of the Dynamin ring (Wigge and McMahon 1998). Treatment of primary hippocampal neurons with amphyphysin I antisense oligonucleotides inhibits neurite outgrowth, an effect similar to that exerted by suppression of Dynamin expression in the same system (Mundigl et al. 1998). While the mechanism remains to be determined, it is tempting to speculate that suppression of neurite outgrowth might reflect a role of Amphiphysin in actin cytoskeleton, possibly exerted through cooperation with the Amhiphysin-interacting protein synaptojanin, an inositol 5-phosphatase that cleaves phosphoinositides, which are in turn able to regulate various aspects of actin remodeling (Micheva et al. 1997). A further example is provided by Abp1, which binds to F-actin and it is associated to dynamic actin structures in lamellipodia and filopodia, in a fashion controlled by the Rac1 GTPase (Kessels et al. 2000). Abp1 binds to Dynamin both in vitro and in vivo, via its SH3 domain. Ectopic expression of the Abp1-SH3 domain leads to a reduction in receptor-mediated endocytosis, an effect reverted by the simultaneous overexpression of Dynamin. Mammalian Abp1 is also enriched at sites where both Dynamin and actin are present (Kessels et al. 2000). This supports the possibility that it may serve as a link between the endocytic machinery, via its SH3 domain, and the actin cytoskeleton, through its N-terminal F-actin binding region. The discovery that Abp1 modulates actin polymerization through its binding and regulation of the activity of the Arp2/3 actin nucleation complex (Goode et al. 2001) further strengthens the possibility that actin dynamics may play an active role in driving the completion of Clathrin-coated vesicle budding. Two intriguing protein scaffolds, which bind Dynamin and N-WASP and possess biochemical activities that activate Cdc42 are Intersectin-1 (Hussain et al. 2001) and Tuba (Salazar et al. 2003). Association of N-WASP with Intersectin-1, coupled with activation of Cdc42, could promote focal actin assembly at clathrin-coated pits (Hussain et al. 2001). Alternatively, association of the phosphoinositide phosphatase, Synaptojanin, with Intersectin at Clathrin-coated pits could either suppress or enhance actin filament formation, depending on which other actin-regulatory proteins were present, by modulating the local concentration of polyphosphoinositides (Hussain et al. 2001). Both these possibilities suggest that Intersecting may mediate the crosstalk between membrane and actin dynamics. It remains unclear, however, whether Intersectins through their binding to Dynamin may also coordinate the function of the latter proteins and link it to the actin polymerization machinery. The protein scaffold, Tuba, shares many biochemical activities and binding partners with intersectin. Tuba is a multidomain scaffolding protein that is

70

A. Disanza et al.

ubiquitously expressed and consists of four N-terminal SH3 domains that bind directly to Dynamin1, an internal BAR (for ‘Bin, Amphiphysin, Rvs’) domain, a GTPase exchange factor (GEF) domain specific for the Rho-family GTPase Cdc42, and two C-terminal Src-homology 3 (SH3) domains (Salazar et al. 2003). In keeping with the ability of Tuba to associate with Dynamin, the expression of its SH3 domain was reported to inhibit transferring internalization presumably interfering with the recruitment and/or the activation of Dynamin itself (Salazar et al. 2003). The extreme C-terminal SH3 domain binds multiple actin-regulatory proteins directly, including N-WASP and Ena/vasodilator stimulated phosphoprotein (Ena/VASP) (Salazar et al. 2003). These latter proteins are though to be important in the generation of various actin-based protrusion, including filopodia (Dent et al. 2007) and lamellipodia (Mejillano et al. 2004). This is achieved through their multiple actin related activities that include antagonizing capping proteins, capturing barbed ends and cross-linking actin filaments (reviewed in (Bear and Gertler 2009)). Thus, Tuba may function as a scaffold that link Dynamin to the actin cytoskeleton machinery. Consistent with this view, targeting of Tuba’s C-terminal SH3 domain to mitochondria, for example, resulted in accumulation of F-actin on this organelle presumably by recruiting actin filaments directly or promoting de novo actin polymerization at the mitochondrial surface. Additionally, Tuba was shown to be recruited to PIP2-rich lipid vesicles and to promote dorsal ruffling in a N-WASP-dependent manner (Kovacs et al. 2006). Finally, similar to the tandem DH-PH domains of Intersectin-1, the Tuba DH-BAR domain specifically catalyzes guanine nucleotide exchange activity on Cdc42; thus concerted interactions of Tuba with Cdc42 and N-WASP or Ena/WASP proteins and Dynamin could link membrane dynamic and actin nucleation and filament organization through mechanisms that remain yet to fully identified. One final link, whose functional consequences are still obscure, is the interaction between Dynamin and a Profilin2 (an neuronal specific isoform) (Witke et al. 1998). Profilins are critical regulator of the dynamics of actin assembly. This is achieved through multiple mechanisms: actin monomer binding, actin nucleotide exchange activity and interactions with proline-rich sequences in proteins (Pollard and Borisy 2003). As ligands of proline-rich sequences of many different proteins, profilins can stabilize specific conformations of their interacting partners (Yang et al., 2000) or deliver actin monomer (as profilin-actin) to sites where profilin is recruited (Le Clainche and Carlier 2008). In the case of Dynamin, the speculative function of Profilin2 could be that of limiting the accessibility to Dynamin Proline-rich region of other endocytic interactors due the elevated supramicromolar concentration of profiling in cells (Gareus et al. 2006). Alternatively, and more likely, Dynamin may serve as scaffold facilitating the recruitment of profiling or the profiling:actin complex to NPFs such as N-WASP. In summary, it has become clear that the Dynamin, the critical pincher of vesicle scission in endocytosis, possesses more functional roles than its mechanobiochemical activities on membrane vesicle and tubule suggest. A plethora of evidence that support Dynamin involvement in the control of actin polymerization and in a diverse set of actin-dynamics based motile processes. While the precise molecular

3

Endocytic Control of Actin-based Motility

71

mechanisms though which Dynamin integrate actin and membrane dynamics remains to be firmly established, it seems reasonable to propose that this protein is a converging node that coordinates membrane deformation and actin-generated forces. The orchestration of these processes, may, in turn, be required not only to promote endocytic events, but also to impose spatial constrains needed for the generation of polarized actin-base protrusions, essential for directed cell migration and invasion. A somewhat similar role appear to be played also by endocytic accessory proteins, such as CIN85/CD2AP or the members of the Bar-containing family of proteins, which through different mechanisms and activities appears to integrate membrane trafficking processes with the actin cytoskeleton. This dual function reflects the need to exploit the forces generated by the actin treadimilling to deformed membrane into vesicle and tubule and initiate vesicle trafficking. It further highlights the importance of ensuring proper spatial and temporal coordination to the actin dynamic-based forces in order to elicit effective and directional motility, bringing support to the emerging notion that endocytosis (or better endo/exocytic processes) is an integrated and pervasive network embedded in the cellular masterplan.

3.3 EEC Control of Signalling Circuitries that Orchestrate Polarized Cell Motility and Invasion One process that depends on the intertwined connection between signalling and localization is cell motility, where the precise perception of extracellular cues in three-dimensional setting is remarkably complex, particularly when cells move toward chemo-attractants in a polarized fashion. During chemotactic cell migration, cells must reorient by polarising PM sensors according to the direction of travel, and coordinate membrane traffic, cell substrate adhesion, and actin remodelling to generate the propulsive forces. One mechanism that has been proposed to orchestrate these processes is by maintaining the polarized state of critical motogenic sensors and their signalling effectors through a flow of endocytic internalization of membranes accompany by redelivery to of vesicles and cargos to the cell front (Bretscher and Aguado-Velasco 1998). Originally, this idea was prompted by the observations that vesicles-containing recycling cargos, such as the receptors for low density lipoprotein (LDL) and transferrin, are concentrated toward the periphery of the cells (Bretscher 1983). To explain these non-uniform distributions, the recycling of receptors was invoked as a way to return them to the cell surface at the cell’s leading edge. Since the distribution of coated pits (the internalization structures) on cells is uniform, Bretscher and Thomson (Bretscher 1983) proposed that there is a bulk membrane flow toward the cell centre that drive cell locomotion. This possibility was further supported by the finding that exocytic delivery of recycling membrane is essential for EGF-mediated ruffling (Bretscher and Aguado-Velasco 1998). While bulk membrane flow as a mechanisms to propel cell motility has been questioned by single particle tracking

72

A. Disanza et al.

experiments in fish keratocytes where the membrane current is likely to be greatest (Kucik et al. 1990), the possibility remains that polarized endocytic/recycling may act, at least in some cell types, as a device to spatially confine actin dynamics regulatory molecules. Consistently, recent studies in tractable model organisms provided unequivocal genetic evidence that, indeed, endocytic/recycling is essential for migration induced by chemotactic gradient. In Drosophila melanogaster, the migration of border cells is regulated by gradients of motogenic growth factors, such as EGF and PV (PDGF/VEGF), toward which cells directionally move (Jekely et al. 2005). Remarkably, genetic removal of CBL, a prototypical E3 endocytic ligase, or Sprint, a Drosophila homologue of the Rab5 GEFs Rin1, or the expression of a dominant negative Shibire/Dynamin mutant impaired in GTP hydrolysis and thus unable to promote vesicle scission, completely disrupted border cell migration (Jekely et al. 2005). Thus, endocytic pathways, particular those impinging on Rab5, are required to ensure spatial resolution of chemotactic signalling emanating from different RTKs, regulating actin-based, polarized protrusive activity and motility. It has to be pointed out that genetic evidence in support of the requirement of polarized recycling for Drosophila border cells to migrate directionally has not yet been provided. However, there are several studies in mammalian cells that support a specific role of endosomal recycling in cell migration. For example, it has been shown that in KB cells, surface ruffles may arise from exocytosis of internal membrane from endosomal cycles (Bretscher and Aguado-Velasco 1998), a process that is coupled to the polarized distribution toward the cell front of recycling membrane receptors (Bretscher and Aguado-Velasco 1998). There is also some evidence that inhibition of the slow recycling pathway by expressing dominant negative Rab11 or the truncated Myosin Vb or Rab11-FIP, an effector of Rab11, impaires HeLa cell migration (Mammoto et al. 1999) and chemotaxis of basophilic leukemia (RBL)-2H3 cells (Fan et al. 2004). This latter results have been recently extended also in epithelial PtK1 cells, where, however, the interference with Rab11 recycling pathway increased random motility possibly as a consequence of delocalized formation of protrusive lamellipodia, but concomitantly impaired directional and persistent migration (Prigozhina and Waterman-Storer 2006), supporting the notion that polarized endosomal recycling is not required for cell locomotion per se, but it is critical for the maintenance of cell migration polarity, which when disrupted leads to disorganized motility. So what might be the additional Rab11 effector(s) that mediates efficient directional motility. Likely candidate are the members of the evolutionary conserved exocysts (or Sec6/8) complex (Sec3, Sec5, Sec6, Sec8, Sec10, Sec15, and Exo70 and Exo84) that were first identified in yeast as being essential for polarized exocytosis of secretory vesicles (Novick et al. 1980). In yeast the exocyst is involved in post-Golgi trafficking via an interaction with the Rab GTPase Sec4p and is involved in tethering the vesicles to the plasma membrane before SNARE (SNAP and NSF Attachment REceptors)-mediated fusion. In mammalian cells, however, the Exocyst is also involved in intracellular vesicle transport. Consistently, Sec15, an Exocyst component, is a well-established interactor of Rab11 (Langevin et al. 2005), while Sec10 is an effector of the recycling GTPases, ARF6 (Prigent et al. 2003).

3

Endocytic Control of Actin-based Motility

73

Collectively these findings suggest that the Exocyst mediates the delivery and fusion to the PM of vesicle containing cargos important to promote actin-based polarized protrusions. It was shown, for example that RalA and RalB, members of the Ras family of small GTP-binding proteins, regulate exocyst function (Brymora et al. 2001; Moskalenko et al. 2002), and that the Ral-exocyst interaction induces filopodia formation through a mechanism that is independent of exocytosis (Sugihara et al. 2002). The exocyst is involved in directional cell motility (Zuo et al. 2006; Rosse et al. 2006; Spiczka and Yeaman 2008), presumably by coordinating the delivery of secretory and/or recycling vesicles to sites of dynamic plasma membrane expansion. This later possibility is further supported by the finding that Exo70 directly interacts with the Arpc1 subunit of the Arp2/3 complex in EGF dependent manner on membrane protrusions, whose extension depends on both the Arp2/3 complex and Exo70 (Zuo et al. 2006). How this association results in promotion of membrane protrusions is still unclear. One possibility is that Exo70 locally activates the Arp2/3 complex promoting branched filament elongation. Alternatively, Exo70 may mark microdomains on the PM, where the Arp2/3 complex is recruited and subsequently activated by PM-localized NPFs of the N-WASP and WAVE family. In support of the latter possibility, recently another component of the Exocyst, Sec3, was found in yeast to bind to phosphatydilinostol 4,5-biphopshate and the small GTPase, Cdc42 (Liu et al. 2009). Disruption of these interactions inhibited Exocysts polarization and affected cell morphogenesis. Given the conserved role of Cdc42 in cellular polarity and directional migration, a plausible yet hypothetical scenario is that PI4,5P and Cdc42-rich membrane domain may serve as targeting sites for Exocyst-mediated Arp2/3 complex localization, resulting in spatiallyrestricted actin polymerization and membrane extension (Liu et al. 2009). A further corollary to this model is that other cargos might be transported through Exocystmediated exocytosis to site of high actin dynamics. This may be particularly relevant for the generation of invasive/adhesive structures, such as invadopodia, which are actin-rich protrusions specialized in localized delivery of metalloproatease, MMPs (Gimona et al. 2008). Indeed, interference with the Exocycts prevented the formation of invadopodia and inhibited invasion of breast cancer cells by blocking MMPs secretion (Liu et al. 2009), suggesting that the Exocist can couple spatiallyrestricted actin polymerization with pericellualr proteolysis, promoting invasion in three dimensions.

3.3.1 Endocytosis and Recycling Ensure Spatial Restriction of Signaling Controlling Migratory Programs in Three Dimensions EEC has also been proposed to account for the generation of different kinds of migratory protrusions induced after stimulation with Receptor tyrosine kinases (RTK), such as peripheral lamellipodia (PL) and dorsal surface circular dorsal ruffles (CDR) (Buccione et al. 2004). Notably, lamellipodia are the first obligatory

74

A. Disanza et al.

step in two-dimensional cell motility. The function of circular dorsal ruffles is less established, but have been shown to be sites of internalization events, such as fluid phase endocytosis (Buccione et al. 2004), and required for migration in three-dimensional matrix (Suetsugu et al. 2003). These latter combined features suggest that circular ruffles may be sites where integration between actin dynamics and endocytosis take place. In keeping with this notion, initial dissection of the molecular machinery responsible for the formation of these protrusions revealed that endocytic proteins, such as Dynamin, and actin dynamics regulators, such as cortactin, function coordinately in the generation of dorsal ruffles (also called wave) (Krueger et al. 2003; Orth and McNiven 2003). Moreover, a tripartite signaling cascade, originating from Rab5, phosphatidylinositol-3-OH kinase, and Rac must simultaneously operate to allow the formation of circular dorsal ruffles (Lanzetti et al. 2004). Remarkably, within this context Rab5 appears to control through its negative regulator/effector molecule RN-tre, the localization on CDR of the actin crosslinker, Actinin-4, suggesting the existence of a linear pathway Rab5-RN-treActinin-4 required to remodel filamentous actin for the formation of specialize protrusions (Lanzetti et al. 2004). Conversely, a pathway, linearly connecting, RasPI3K-Rac to actin remodelling is sufficient to generate lamellipodia protrusions in response to RTK activation (Lanzetti et al. 2004). How these three-pronged signalling pathways cross-talk and are co-regulated in space and time had remained unexplored. Recently, however, we showed that Rab5 and Clathrin-mediated endocytic trafficking of Rac, and its recycling to the plasma membrane is specifically required for the generation of CDR (Palamidessi et al. 2008). Consistently, inhibition of endocytosis in HeLa cells by depletion of Clathrin-coated pits or expression of a GTPase-deficient form of Dynamin inhibited Rab5-dependent Rac activation and CDR formation. Additionally, Rac trafficking from the plasma membrane to endocytic vesicles containing Rab5 and Tiam1 were required for spatially-confined Rac activation, although it remains unclear whether the active Rac on Rab5 endosomes comes exclusively from cell surface-internalized Rac or whether there is also recruitment of cytoplasmic Rac from Rac-GDI (GDP dissociation inhibitor) complexes. This notwithstanding, to enable localized execution of Rac function, the internal pool of activated Rac GTP must be transported back to the plasma membrane to specific sites of high actin activity. Blockade of recycling by temperature switch to 16◦ C completely abrogated the formation of CDR in response to Rab5 activation and HGF stimulation. Furthermore, by exploiting Rac-fused to photoactivable GFP (paGFP) coupled with two-photon activation excitation that allows targeting of single internalized vesicles, confining the activation of paGFP molecules in three dimensions, while completely excluding the plasma membrane compartment, a direct visualization of Rac polarized transport toward ruffles in the plasma membrane could be evidenced in real time. The targeting of vesicles back to the plasma membrane is accomplished by the small GTPase Arf6, but not by Rab4 or Rab11 (Fig. 3.2), indicating that the ARF6-recycling compartment is the one utilized by Rac for its redelivery to the plasma membrane. Notably, Arf6 has previously been implicated in Rac trafficking (Radhakrishna et al. 1999), moreover Arf6-mediated trafficking of vesicles is known to be responsible for recycling of certain receptors,

3

Endocytic Control of Actin-based Motility

75

A

B Mesenchymal-like

PL

Rab25

CDR

Rab5

MAT

AMT

Rac Tiam1

Amoeboid-like

GTP

HGF Rac GTP

c-Met Arf6 Integrin

Rac GTP

Clathrin

Fig. 3.2 EEC spatially restricts signalling to actin dynamics. (a). Hepatocyte growth factor (HGF) receptor activation induces the formation of circular dorsal ruffles (CDR, arrow) and peripheral lamellipodia (PL, arrowhead). This is accompanied by Clathrin-mediated internalization of the cognate receptor tyrosine kinase, c-Met. Clathrin-and Rab5-dependent endocytosis is required for Tiam-1-mediated, a guanine-nucleotide exchange factor, Rac activation in endosomal vesicles. Recycling of Rac back to plasma membrane areas is essential to promote the reorganization of actin into CDR. The small GTPase Arf6 targets activated Rac to specialized plasma membrane areas (such as ruffles) with high actin activity. The integrin α5β1 can associate with the small GTPase Rab25, a member of the Rab11 family, involved in vesicle recycling (Caswell et al. 2007). Both Rab25 and α5β1 integrin colocalize in intracellular vesicular compartments within the distal tips of pseudopods (the lamellipodia equivalent in a 3D setting) (Caswell et al. 2007). Integrin α5β1 can traffic bidirectionally between intracellular Rab25 vesicles and the plasma membrane within the confines of the pseudopodial tips, this promotes the compartmentalization of a spatially restricted subpopulation of cycling α5β1 within the tip regions of extending pseudopods (Caswell et al. 2007). A HeLa cell stimulated with HGF and stained with phalloidin to detect F-actin (white) is shown. b. Mode of motility of individual cells in 3D. Two modes of single-cell 3D migration in extracellular matrices have been described (Wolf and Friedl 2002). The elongated-lamellipodia movement (mesenchymal mode) begins with the formation of flat, adherent protrusions driving directional motility. Conversely, amoeboid migration depends on Rho/Rock-dependent actomyosin contractility, driving blebbing-like movements of loosely adherent cells. Conversion between these migratory modes, amoeboid-to-mesenchymal (AMT) and mesenchymal-to-amoeboid (MAT) transition, confers plasticity to metastatic tumor cell migration (Wolf and Friedl 2002). The endocytic/recycling of Rac and α5β1 may control the migration into 3D matrices by promoting a mesenchymal mode of migration. GFP-expressing, HeLa cells embedded in matrigel and treated (top) or not (bottom) with the Rock inhibitor Y27632 or expressing Rab5 (not shown) undergo an amoeboid-to-mesenchymal-like cell shape transition in 3D. [Images are reprinted from (Disanza et al. 2009) Copyright (2009), with permission from Elsevier].

such as the β1 integrins, to the plasma membrane (Donaldson 2005), suggesting the intriguing possibility that these vesicles may be the site where different kinds of cargo are sorted and co-ordinately delivered to ensure spatial restriction of signalling required for polarized motility.

76

A. Disanza et al.

Mechanistically, one important question raised by these findings concerns how signalling molecules are recycled to specific regions of the plasma membrane (as opposed to the bulk plasma membrane), to execute spatially restricted signaling. In the case of Rac, one possible answer may come form recent studies connecting localized Rac activation with integrin-mediated adhesion and lipid raft internalization. These studies suggested that Rac positioning at, and trafficking from and to plasma membrane specific locations may be regulated through raft endocytosis, (Schlunck et al. 2004; Radhakrishna et al. 1999; del Pozo et al. 2004, 2005). Thus, upon activation of integrin, sites of high Rac affinity becomes available on the plasma membrane preventing its internalization, which instead, occurs following cell detachment in a Dynamin and Caveolin-dependent manner (del Pozo et al. 2005). Notably disruption of Dynamin function also results in segregation of active Rac in aberrant membrane invaginations, away from the plasma membrane, preventing the formation of regularly shaped lamellipodia, likely by blocking macropinocytic Rac internalization (Schlunck et al. 2004). It is thus tempting to speculate that Rac, activated through a process that requires CME (Palamidessi et al. 2008), is then delivered to regions of the plasma membrane, enriched in lipid rafts (del Pozo et al. 2004, 2005) whose non-clathrin internalization is prevented by integrin signalling. Such a circuitry would require tight coordination between different internalization routes, a concept that also emerged from recent genomic studies of the clathrin and non-clathrin pathways of internalization (Pelkmans et al. 2005). Within this context, Arf6-dependent recycling appears to be the critical route controlling not only the redelivery of Rac (Palamidessi et al. 2008) and integrins (del Pozo et al. 2004, 2005; Del Pozo et al. 2002), but also of lipid raft, back to the plasma membrane, ultimately coordinating Rac signalling and directional migration with adhesion-dependent cell growth (Balasubramanian et al. 2007). What are the precise molecular determinants linking and regulating the activities of all these proteins is not entirely clear. One clue in this direction is provided by the observations that Arf6-polarized recycling of cholesterol-rich membranes requires microtubule and motors for directed delivery (Balasubramanian et al. 2007), suggesting integration between membrane trafficking and microtubules-based transport. Alternatively, the multimolecular unit composed of GIT1, a member of a family of GTPase-activatingproteins for ARF6, paxillin, a focal adhesion protein, and a complex including the Rac/Cdc42 exchanging factors PIX/Cool and the kinase PAK, which assembles all the critical elements controlling vesicle recycling, focal adhesion turnover, and Racand Cdc42 -directional migration (Hoefen and Berk 2006), may serve as a platform for intergration of diverse signaling pathways and traffiking routes. The importance of spatial restriction of Rac signalling in controlling migratory protrusions is underscored by the discovery that the formation of CDR correlates with the ability of cells to acquire a mode of motility which is typical of mesenchymal moving cell, and to migrate in 3D (Suetsugu et al. 2003), properties that are defining features of metastatic invading cancer cells. Recent experimental evidence mainly based on the use of intravital two-photon analysis of invading

3

Endocytic Control of Actin-based Motility

77

cancer has revealed that individual tumor cells detaching from the original mass and adventuring into the surrounding tissue can adopt diverse modes of migration whose molecular determinants are only just starting to be elucidated (for review see ref. (Wolf and Friedl 2006). Amoeboid motility is characterized by the formation of uniformly distributed fast-extending and -retracting blebs, and relies on efficient actomyosin contractility (Fukui 2002). This type of motility is conserved throughout evolution and has been shown to have physiological relevance at the organismal level, such as during primordial cell migration in Zebrafish (Blaser et al. 2006). Mesenchymal-type migration, conversely, relies on the formation of persistent and extended Rac-dependent lamellipodia protrusions, and on the activation of metalloproteases (Wolf and Friedl 2006). Remarkably, tumor cells can plastically switch from one mode of motility to the other dramatically increasing their chance to invade regardless of the variety, but rather taking advantage of the diversity of the microenvironmental conditions. Rac signalling has been shown to be critical for mesenchymal migration (SanzMoreno et al. 2008). This together with the importance of endocytic/recyling of suggesting that endocytic/recycling by ensuring Rac spatial restriction and polarization may also control the amoeboid-to-mesenchymal transition. Indeed, the expression of Rab5 in cultured melanoma cells, a model for amoeboid motility, induced a change from amoeboid to a more mesenchymal-like morphology and movement. Conversely, inhibition of Rab5 in cultured colon carcinoma cells, a model for mesenchymal motility, resulted in a switch to a more amoeboid-like morphology and motility (Palamidessi et al. 2008). Most intriguingly, Rab5 was shown to play a role also in normal cell migration in the guidance of primordial germ cell migration during zebrafish development (Palamidessi et al. 2008). These findings suggest that spatially-regulated membrane trafficking, and trafficking of key signalling molecules, such as Rac, is an evolutionary conserved mechanism to orchestrate localized actin dynamics, polarized protrusive activity, directed cell motility, ultimately affecting the modes of motility of normal and cancer cells.

3.3.2 Endo/Exocytic Cycles of Adhesion Receptors in the Regulation of Migratory and Invasive Programs Cell migration proceeds by cycles of edge protrusion, adhesion, and retraction whose precise coordination in space and time is critical to promote cell locomotion. An optimal organization of actin filaments, myosin, and adhesion sites is, thus, required for fast migration, indicating that actin-filament assembly, force generation, and adhesion are interdependent functions (Giannone et al. 2007). Integrins, which are the major cell surface adhesion receptors for ligands in the extracellular matrix (Hynes 2002), play a critical role in regulating cell migration. Integrins are heterodimeric, transmembrane proteins consisting of an α and a β-chain and are involved in the transmission and interpretation of signals from the extracellular environment into various signalling cascades (Hynes 2002). Several different mechanisms, including expression and subunit heterodimerization

78

A. Disanza et al.

patterns, clustering and lateral diffusion in the plane of the plasma membrane, and interaction with the actin cytoskeleton and the inside of cells control their activity (Calderwood 2004). In addition to this, many integrins are continually internalized from the plasma membrane into endosomal compartments and are subsequently recycled (Caswell and Norman 2006; Pellinen and Zvaska 2006), prompting the proposal that the endo/exocytic cycle of these adhesion receptors is essential to control various aspect of cell locomotion. A number of excellent reviews have recently covered the molecular details and mechanisms through which various integrins by being internalized either through CME and CIE are subsequently recycled through Rabs -or Arf6-dependent routes toward the cell front, thus maintaining a spatially-restricted, polarised distribution of cycling extracellular matrix receptors within the confine locale of an advancing cell protrusion (reviewed in (Caswell and Norman 2008)). Remarkably, the integrin endo/exocytic cycle, similar to what has been proposed for motogenic receptors of the RTK and GPCR family, appears to be critical to promote directional migration in 2D and was recently shown to mediate the motility behaviour of cells in 3D, ultimately controlling cell invasion. Clear examples of this kind are the studies on α V β 6, whose role in promoting haptotaxis and invasive migration in oral carcinoma is firmly established (Thomas et al. 2007). Disruption of α V β 6 CME-internalization prevented haptotactic cell migration toward α V β 6 ligands and invasion through matrigel in organotypic invasion assays, which more closely model the tumor stroma (Ramsay et al. 2007a, b). This provided the first demonstration that endocytic trafficking of adhesion receptor directly contributes to invasive migration, similar to what is observed during tumor metastatization. Evidence, instead, that recycling routes control the ability of integrinmediated migration into fibronectin-containing 3D-extracellular matrix has been provided for α5β1, that associates with the small GTPase Rab25, a member of the Rab11 family, involved in vesicle recycling (Caswell et al. 2007). Rab25 was shown to promote the formation of long pseudopodial extensions as cells migrate in 3D contexts, and both Rab25 and α5β1 integrin co-localise in relatively static intracellular vesicular compartments within the distal tips of pseudopods (the lamellipodia equivalent in a 3D setting) (Caswell et al. 2007). Using a photoactivatable GFP-α5 unit to analyse the dynamics of integrin trafficking in real-time permitted to reveal that α5β1 integrin traffics bidirectionally between intracellular Rab25 vesicles and the plasma membrane within the confines of the pseudopodial tips. Additionally, Rab25 promoted the compartmentalisation of a spatially restricted subpopulation of cycling α5β1 within the tip regions of extending pseudopods as cancer cells move forward through a 3D matrix (Caswell et al. 2007). Collectively these observations indicate that polarized delivery of vesicles containing key cell migratory adhesive molecules is critical for their restricted localization and to ensure confined signalling output directly influencing the efficiency and modes of 3D cell migration and invasion. The scenario that is emerging from all these studies is that there is a “rush-hourlike” intense trafficking of vesicles and cargo molecules as a cell becomes motile, whose precise regulation is required to ensure polarized migratory response both in 2D and 3D setting. One mechanism that may enable to guide different cargos

3

Endocytic Control of Actin-based Motility

79

to the right itinerary among this intricate road network is to coordinate their navigation so that they can reach the same destination in a temporal and spatially controlled fashion. Some evidence in support of the latter notion is beginning to be provided. For instance, it has been shown that the there is a complex level of signal and trafficking crosstalk between α V β 3 and α5β1 integrins and the EGFR. Blockade of the adhesive function of α V β 3 promoted Rabs-dependent polarized recycling of α5β1 leading to the acquisition of rapid/random movement on two-dimensional substrates and to a marked increase in fibronectin-dependent migration of tumor cells into three-dimensional matrices. Remarkably, the enhance recyling of α5β1 had no effect on its adhesive properties, rather it facilitated the association of EGFR with the recycling machinery promoting the coordinated recycling of these two receptors, increasing EGFR autophosphorylation and activation of the proinvasive kinase PKB/Akt (Caswell et al. 2008). How general is the trafficking coupling of diverse receptors remains to be investigated, nevertheless the possibility that co-navigation not only of membrane-bound receptors, but also of some of the key signalling effectors may have evolved as an effective way to achieve polarization and spatial restriction will surely receive increasing attention. Acknowledgments We apologise to all those colleagues whose primary references or important discoveries could not be properly acknowledged for lack of space. The authors of this review are supported by grants from: AIRC (Associazione Italiana Ricerca sul Cancro), European Community (VI Framework) and PRIN2007 (progetti di ricerca di interesse nazionale). F.T and A.D. are supported by FIRC fellowship.

References Balasubramanian N, Scott DW, Castle JD, Casanova JE, Schwartz MA (2007) Arf6 and microtubules in adhesion-dependent trafficking of lipid rafts. Nat Cell Biol 9: 1381–1391. Bear JE, Gertler FB (2009) Ena/VASP: towards resolving a pointed controversy at the barbed end. J Cell Sci 122: 1947–1953. Birtwistle MR, Kholodenko BN (2009) Endocytosis and signalling: a meeting with mathematics. Mol Oncol. 3(4): 308–320. Blaser H, Reichman-Fried M, Castanon I, Dumstrei K, Marlow FL, et al. (2006) Migration of zebrafish primordial germ cells: a role for myosin contraction and cytoplasmic flow. Dev Cell 11: 613–627. Boldajipour B, Raz E (2007) What is left behind – quality control in germ cell migration. Sci STKE 2007: pe16. Bretscher MS (1983) Distribution of receptors for transferrin and low density lipoprotein on the surface of giant HeLa cells. Proc Natl Acad Sci USA 80: 454–458. Bretscher MS, Aguado-Velasco C (1998) EGF induces recycling membrane to form ruffles. Curr Biol 8: 721–724. Bretscher MS, Aguado-Velasco C (1998) Membrane traffic during cell locomotion. Curr Opin Cell Biol 10: 537–541. Bryant DM, Mostov KE (2008) From cells to organs: building polarized tissue. Nat Rev Mol Cell Biol 9: 887–901. Brymora A, Valova VA, Larsen MR, Roufogalis BD, Robinson PJ (2001) The brain exocyst complex interacts with RalA in a GTP-dependent manner: identification of a novel mammalian Sec3 gene and a second Sec15 gene. J Biol Chem 276: 29792–29797.

80

A. Disanza et al.

Buccione R, Orth JD, McNiven MA (2004) Foot and mouth: podosomes, invadopodia and circular dorsal ruffles. Nat Rev Mol Cell Biol 5: 647–657. Calderwood DA (2004) Integrin activation. J Cell Sci 117: 657–666. Cao H, Garcia F, McNiven MA (1998) Differential distribution of dynamin isoforms in mammalian cells. Mol Biol Cell 9: 2595–2609. Caswell P, Norman J (2008) Endocytic transport of integrins during cell migration and invasion. Trends Cell Biol 18: 257–263. Caswell PT, Chan M, Lindsay AJ, McCaffrey MW, Boettiger D, et al. (2008) Rab-coupling protein coordinates recycling of alpha5beta1 integrin and EGFR1 to promote cell migration in 3D microenvironments. J Cell Biol 183: 143–155. Caswell PT, Norman JC (2006) Integrin trafficking and the control of cell migration. Traffic 7: 14–21. Caswell PT, Spence HJ, Parsons M, White DP, Clark K, et al. (2007) Rab25 associates with alpha5beta1 integrin to promote invasive migration in 3D microenvironments. Dev Cell 13: 496–510. Charras GT, Yarrow JC, Horton MA, Mahadevan L, Mitchison TJ (2005) Non-equilibration of hydrostatic pressure in blebbing cells. Nature 435: 365–369. Cosen-Binker LI, Kapus A (2006) Cortactin: the gray eminence of the cytoskeleton. Physiology (Bethesda) 21: 352–361. Damke H, Baba T, Warnock DE, Schmid SL (1994) Induction of mutant dynamin specifically blocks endocytic coated vesicle formation. J Cell Biol 127: 915–934. De Donatis A, Comito G, Buricchi F, Vinci MC, Parenti A, et al. (2008) Proliferation versus migration in platelet-derived growth factor signaling: the key role of endocytosis. J Biol Chem 283: 19948–19956. del Pozo MA, Alderson NB, Kiosses WB, Chiang HH, Anderson RG, et al. (2004) Integrins regulate Rac targeting by internalization of membrane domains. Science 303: 839–842. del Pozo MA, Balasubramanian N, Alderson NB, Kiosses WB, Grande-Garcia A, et al. (2005) Phospho-caveolin-1 mediates integrin-regulated membrane domain internalization. Nat Cell Biol 7: 901–908. Del Pozo MA, Kiosses WB, Alderson NB, Meller N, Hahn KM, et al. (2002) Integrins regulate GTP-Rac localized effector interactions through dissociation of Rho-GDI. Nat Cell Biol 4: 232–239. Dent EW, Kwiatkowski AV, Mebane LM, Philippar U, Barzik M, et al. (2007) Filopodia are required for cortical neurite initiation. Nat Cell Biol 9: 1347–1359. Di Guglielmo GM, Le Roy C, Goodfellow AF, Wrana JL (2003) Distinct endocytic pathways regulate TGF-beta receptor signalling and turnover. Nat Cell Biol 5: 410–421. Disanza A, Frittoli E, Palamidessi A, Scita G (2009) Endocytosis and spatial restriction of cell signaling. Mol Oncol. 3(4): 280–296. Doherty GJ, McMahon HT (2009) Mechanisms of endocytosis. Annu Rev Biochem 78: 857–902. Donaldson JG (2005) Arfs, phosphoinositides and membrane traffic. Biochem Soc Trans 33: 1276–1278. Fan GH, Lapierre LA, Goldenring JR, Sai J, Richmond A (2004) Rab11-family interacting protein 2 and myosin Vb are required for CXCR2 recycling and receptor-mediated chemotaxis. Mol Biol Cell 15: 2456–2469. Frost A, Perera R, Roux A, Spasov K, Destaing O, et al. (2008) Structural basis of membrane invagination by F-BAR domains. Cell 132: 807–817. Fukui Y (2002) Mechanistics of amoeboid locomotion: signal to forces. Cell Biol Int 26: 933–944. Gallop JL, McMahon HT (2005) BAR domains and membrane curvature: bringing your curves to the BAR. Biochem Soc Symp: 223–231. Garcia GL, Parent CA (2008) Signal relay during chemotaxis. J Microsc 231: 529–534. Gareus R, Di Nardo A, Rybin V, Witke W (2006) Mouse profilin 2 regulates endocytosis and competes with SH3 ligand binding to dynamin 1. J Biol Chem 281: 2803–2811.

3

Endocytic Control of Actin-based Motility

81

Giannone G, Dubin-Thaler BJ, Rossier O, Cai Y, Chaga O, et al. (2007) Lamellipodial actin mechanically links myosin activity with adhesion-site formation. Cell 128: 561–575. Gimona M, Buccione R, Courtneidge SA, Linder S (2008) Assembly and biological role of podosomes and invadopodia. Curr Opin Cell Biol 20: 235–241. Goode BL, Rodal AA, Barnes G, Drubin DG (2001) Activation of the Arp2/3 complex by the actin filament binding protein Abp1p. J Cell Biol 153: 627–634. Habets GG, Scholtes EH, Zuydgeest D, van der Kammen RA, Stam JC, et al. (1994) Identification of an invasion-inducing gene, Tiam-1, that encodes a protein with homology to GDP-GTP exchangers for Rho-like proteins. Cell 77: 537–549. Hoefen RJ, Berk BC (2006) The multifunctional GIT family of proteins. J Cell Sci 119: 1469–1475. Hussain NK, Jenna S, Glogauer M, Quinn CC, Wasiak S, et al. (2001) Endocytic protein intersectin-l regulates actin assembly via Cdc42 and N-WASP. Nat Cell Biol 3: 927–932. Hynes RO (2002) Integrins: bidirectional, allosteric signaling machines. Cell 110: 673–687. Jekely G, Sung HH, Luque CM, Rorth P (2005) Regulators of endocytosis maintain localized receptor tyrosine kinase signaling in guided migration. Dev Cell 9: 197–207. Kaksonen M (2008) Taking apart the endocytic machinery. J Cell Biol 180: 1059–1060. Kessels MM, Engqvist-Goldstein AE, Drubin DG (2000) Association of mouse actin-binding protein 1 (mAbp1/SH3P7), an Src kinase target, with dynamic regions of the cortical actin cytoskeleton in response to Rac1 activation. Mol Biol Cell 11: 393–412. Kessels MM, Engqvist-Goldstein AE, Drubin DG, Qualmann B (2001) Mammalian Abp1, a signalresponsive F-actin-binding protein, links the actin cytoskeleton to endocytosis via the GTPase dynamin. J Cell Biol 153: 351–366. Kessels MM, Qualmann B (2006) Syndapin oligomers interconnect the machineries for endocytic vesicle formation and actin polymerization. J Biol Chem 281: 13285–13299. Kovacs EM, Makar RS, Gertler FB (2006) Tuba stimulates intracellular N-WASP-dependent actin assembly. J Cell Sci 119: 2715–2726. Krishnan KS, Rikhy R, Rao S, Shivalkar M, Mosko M, et al. (2001) Nucleoside diphosphate kinase, a source of GTP, is required for dynamin- dependent synaptic vesicle recycling. Neuron 30: 197–210. Kruchten AE, McNiven MA (2006) Dynamin as a mover and pincher during cell migration and invasion. J Cell Sci 119: 1683–1690. Krueger EW, Orth JD, Cao H, McNiven MA (2003) A dynamin-cortactin-Arp2/3 complex mediates actin reorganization in growth factor-stimulated cells. Mol Biol Cell 14: 1085–1096. Kucik DF, Elson EL, Sheetz MP (1990) Cell migration does not produce membrane flow. J Cell Biol 111: 1617–1622. Langevin J, Morgan MJ, Sibarita JB, Aresta S, Murthy M, et al. (2005) Drosophila exocyst components Sec5, Sec6, and Sec15 regulate DE-Cadherin trafficking from recycling endosomes to the plasma membrane. Dev Cell 9: 365–376. Lanzetti L, Palamidessi A, Areces L, Scita G, Di Fiore PP (2004) Rab5 is a signalling GTPase involved in actin remodelling by receptor tyrosine kinases. Nature 429: 309–314. Le Clainche C, Carlier MF (2008) Regulation of actin assembly associated with protrusion and adhesion in cell migration. Physiol Rev 88: 489–513. Le Roy C, Wrana JL (2005) Clathrin- and non-clathrin-mediated endocytic regulation of cell signalling. Nat Rev Mol Cell Biol 6: 112–126. Lee E, De Camilli P (2002) Dynamin at actin tails. Proc Natl Acad Sci USA 99: 161–166. Liu J, Yue P, Artym VV, Mueller SC, Guo W (2009) The Role of the Exocyst in MMP Secretion and Actin Dynamics during Tumor Cell Invadopodia Formation. Mol Biol Cell. 20(16): 3763–3771. Mammoto A, Ohtsuka T, Hotta I, Sasaki T, Takai Y (1999) Rab11BP/Rabphilin-11, a downstream target of rab11 small G protein implicated in vesicle recycling. J Biol Chem 274: 25517–25524. Matsumura F (2005) Regulation of myosin II during cytokinesis in higher eukaryotes. Trends Cell Biol 15: 371–377.

82

A. Disanza et al.

Mayor S, Pagano RE (2007) Pathways of clathrin-independent endocytosis. Nat Rev Mol Cell Biol 8: 603–612. McNiven MA, Kim L, Krueger EW, Orth JD, Cao H, et al. (2000) Regulated interactions between dynamin and the actin-binding protein cortactin modulate cell shape. J Cell Biol 151: 187–198. Mejillano MR, Kojima S, Applewhite DA, Gertler FB, Svitkina TM, et al. (2004) Lamellipodial versus filopodial mode of the actin nanomachinery: pivotal role of the filament barbed end. Cell 118: 363–373. Micheva KD, Kay BK, McPherson PS (1997) Synaptojanin forms two separate complexes in the nerve terminal. Interactions with endophilin and amphiphysin. J Biol Chem 272: 27239–27245. Minina S, Reichman-Fried M, Raz E (2007) Control of receptor internalization, signaling level, and precise arrival at the target in guided cell migration. Curr Biol 17: 1164–1172. Mooren OL, Kotova TI, Moore AJ, Schafer DA (2009) Dynamin2 GTPase and cortactin remodel actin filaments. J Biol Chem. 284(36): 23995–24005. Moskalenko S, Henry DO, Rosse C, Mirey G, Camonis JH, et al. (2002) The exocyst is a Ral effector complex. Nat Cell Biol 4: 66–72. Mundigl O, Ochoa GC, David C, Slepnev VI, Kabanov A, et al. (1998) Amphiphysin I antisense oligonucleotides inhibit neurite outgrowth in cultured hippocampal neurons. J Neurosci 18: 93–103. Novick P, Field C, Schekman R (1980) Identification of 23 complementation groups required for post-translational events in the yeast secretory pathway. Cell 21: 205–215. Ochoa GC, Slepnev VI, Neff L, Ringstad N, Takei K, et al. (2000) A functional link between dynamin and the actin cytoskeleton at podosomes. J Cell Biol 150: 377–389. Orth JD, Krueger EW, Cao H, McNiven MA (2002) The large GTPase dynamin regulates actin comet formation and movement in living cells. Proc Natl Acad Sci USA 99: 167–172. Orth JD, McNiven MA (2003) Dynamin at the actin-membrane interface. Curr Opin Cell Biol 15: 31–39. Otsuki Y, Tanaka M, Yoshii S, Kawazoe N, Nakaya K, et al. (2001) Tumor metastasis suppressor nm23H1 regulates Rac1 GTPase by interaction with Tiam1. Proc Natl Acad Sci USA 98: 4385–4390. Palamidessi A, Frittoli E, Garre M, Faretta M, Mione M, et al. (2008) Endocytic trafficking of Rac is required for the spatial restriction of signaling in cell migration. Cell 134: 135–147. Pelkmans L, Fava E, Grabner H, Hannus M, Habermann B, et al. (2005) Genome-wide analysis of human kinases in clathrin- and caveolae/raft-mediated endocytosis. Nature 436: 78–86. Pellinen T, Ivaska J (2006) Integrin traffic. J Cell Sci 119: 3723–3731. Pollard TD (2003) The cytoskeleton, cellular motility and the reductionist agenda. Nature 422: 741–745. Pollard TD, Borisy GG (2003) Cellular motility driven by assembly and disassembly of actin filaments. Cell 112: 453–465. Praefcke GJ, McMahon HT (2004) The dynamin superfamily: universal membrane tubulation and fission molecules? Nat Rev Mol Cell Biol 5: 133–147. Prigent M, Dubois T, Raposo G, Derrien V, Tenza D, et al. (2003) ARF6 controls post-endocytic recycling through its downstream exocyst complex effector. J Cell Biol 163: 1111–1121. Prigozhina NL, Waterman-Storer CM (2006) Decreased polarity and increased random motility in PtK1 epithelial cells correlate with inhibition of endosomal recycling. J Cell Sci 119: 3571–3582. Radhakrishna H, Al-Awar O, Khachikian Z, Donaldson JG (1999) ARF6 requirement for Rac ruffling suggests a role for membrane trafficking in cortical actin rearrangements. J Cell Sci 112: 855–866. Ramsay AG, Keppler MD, Jazayeri M, Thomas GJ, Parsons M, et al. (2007) HS1-associated protein X-1 regulates carcinoma cell migration and invasion via clathrin-mediated endocytosis of integrin alphavbeta6. Cancer Res 67: 5275–5284. Ramsay AG, Marshall JF, Hart IR (2007) Integrin trafficking and its role in cancer metastasis. Cancer Metastasis Rev 26: 567–578.

3

Endocytic Control of Actin-based Motility

83

Raz E (2004) Guidance of primordial germ cell migration. Curr Opin Cell Biol 16: 169–173. Rosse C, Hatzoglou A, Parrini MC, White MA, Chavrier P, et al. (2006) RalB mobilizes the exocyst to drive cell migration. Mol Cell Biol 26: 727–734. Salazar MA, Kwiatkowski AV, Pellegrini L, Cestra G, Butler MH, et al. (2003) Tuba, a novel protein containing bin/amphiphysin/Rvs and Dbl homology domains, links dynamin to regulation of the actin cytoskeleton. J Biol Chem 278: 49031–49043. Sanz-Moreno V, Gadea G, Ahn J, Paterson H, Marra P, et al. (2008) Rac activation and inactivation control plasticity of tumor cell movement. Cell 135: 510–523. Schafer DA (2004) Regulating actin dynamics at membranes: a focus on dynamin. Traffic 5: 463–469. Schafer DA, Weed SA, Binns D, Karginov AV, Parsons JT, et al. (2002) Dynamin2 and cortactin regulate actin assembly and filament organization. Curr Biol 12: 1852–1857. Schlunck G, Damke H, Kiosses WB, Rusk N, Symons MH, et al. (2004) Modulation of Rac localization and function by dynamin. Mol Biol Cell 15: 256–267. Schmoranzer J, Kreitzer G, Simon SM (2003) Migrating fibroblasts perform polarized, microtubule-dependent exocytosis towards the leading edge. J Cell Sci 116: 4513–4519. Sheetz MP, Sable JE, Dobereiner HG (2006) Continuous membrane-cytoskeleton adhesion requires continuous accommodation to lipid and cytoskeleton dynamics. Annu Rev Biophys Biomol Struct 35: 417–434. Sigismund S, Argenzio E, Tosoni D, Cavallaro E, Polo S, et al. (2008) Clathrin-mediated internalization is essential for sustained EGFR signaling but dispensable for degradation. Dev Cell 15: 209–219. Sorkin A (2004) Cargo recognition during clathrin-mediated endocytosis: a team effort. Curr Opin Cell Biol 16: 392–399. Sorkin A, Goh LK (2009) Endocytosis and intracellular trafficking of ErbBs. Exp Cell Res 315: 683–696. Sorkin A, von Zastrow M (2009) Endocytosis and signaling: intertwining molecular networks. Nat Rev Mol Cell Biol 10(9): 609–622. Spiczka KS, Yeaman C (2008) Ral-regulated interaction between Sec5 and paxillin targets Exocyst to focal complexes during cell migration. J Cell Sci 121: 2880–2891. Stenmark H (2009) Rab GTPases as coordinators of vesicle traffic. Nat Rev Mol Cell Biol 10: 513–525. Stradal TE, Scita G (2006) Protein complexes regulating Arp2/3-mediated actin assembly. Curr Opin Cell Biol 18: 4–10. Suetsugu S, Yamazaki D, Kurisu S, Takenawa T (2003) Differential roles of WAVE1 and WAVE2 in dorsal and peripheral ruffle formation for fibroblast cell migration. Dev Cell 5: 595–609. Sugihara K, Asano S, Tanaka K, Iwamatsu A, Okawa K, et al. (2002) The exocyst complex binds the small GTPase RalA to mediate filopodia formation. Nat Cell Biol 4: 73–78. Thomas GJ, Nystrom ML, Marshall JF (2006) Alphavbeta6 integrin in wound healing and cancer of the oral cavity. J Oral Pathol Med 35: 1–10. Traynor D, Kay RR (2007) Possible roles of the endocytic cycle in cell motility. J Cell Sci 120: 2318–2327. Unsworth KE, Mazurkiewicz P, Senf F, Zettl M, McNiven M, et al. (2007) Dynamin is required for F-actin assembly and pedestal formation by enteropathogenic Escherichia coli (EPEC). Cell Microbiol 9: 438–449. Wagner PD, Vu ND (1995) Phosphorylation of ATP-citrate lyase by nucleoside diphosphate kinase. J Biol Chem 270: 21758–21764. Wessels D, Reynolds J, Johnson O, Voss E, Burns R, et al. (2000) Clathrin plays a novel role in the regulation of cell polarity, pseudopod formation, uropod stability and motility in Dictyostelium. J Cell Sci 113 (Pt 1): 21–36. Wigge P, McMahon HT (1998) The amphiphysin family of proteins and their role in endocytosis at the synapse. Trends Neurosci 21: 339–344.

84

A. Disanza et al.

Witke W, Podtelejnikov AV, Di Nardo A, Sutherland JD, Gurniak CB, et al. (1998) In mouse brain profilin I and profilin II associate with regulators of the endocytic pathway and actin assembly. Embo J 17: 967–976. Wolf K, Friedl P (2006) Molecular mechanisms of cancer cell invasion and plasticity. Br J Dermatol 154 Suppl 1: 11–15. Yang C, Huang M, DeBiasio J, Pring M, Joyce M, et al. (2000) Profilin enhances Cdc42-induced nucleation of actin polymerization. J Cell Biol 150: 1001–1012. Zhang B, Zelhof AC (2002) Amphiphysins: raising the BAR for synaptic vesicle recycling and membrane dynamics. Bin-Amphiphysin-Rvsp. Traffic 3: 452–460. Zuo X, Zhang J, Zhang Y, Hsu SC, Zhou D, et al. (2006) Exo70 interacts with the Arp2/3 complex and regulates cell migration. Nat Cell Biol 8: 1383–1388.

Chapter 4

Actin in Clathrin-Mediated Endocytosis Marko Kaksonen

Abstract In eukaryotic cells the actin cytoskeleton provides mechanical force for many processes in which the plasma membrane is reshaped. These processes include cell migration, cell division and the formation of endocytic vesicles. Here I will focus on the role of the actin cytoskeleton in clathrin-mediated endocytosis.

Contents 4.1 Actin and Endocytosis in Yeast . . . . . . . . . . . . . . . . . . . 4.1.1 Stages of Yeast Endocytosis . . . . . . . . . . . . . . . . . 4.1.2 Visualizing Actin in Yeast Endocytosis . . . . . . . . . . . . 4.1.3 Regulation of Actin Polymerization at Yeast Endocytic Sites . . . 4.1.4 Layers of Activation and Inhibition of Actin Polymerization . . . 4.1.5 Further Links Between the Membrane and the Actin Cytoskeleton 4.2 Actin and Clathrin-Mediated Endocytosis in Mammals . . . . . . . . 4.2.1 Actin Dynamics in Living Mammalian Cells . . . . . . . . . . 4.2.2 Regulation of Actin Polymerization at Endocytic Sites . . . . . 4.2.3 Multiple Roles of Actin at the Plasma Membrane . . . . . . . . 4.3 Open Questions in Actin-Mediated Endocytosis . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

86 86 87 88 89 90 91 92 92 95 96 97

The classical view of clathrin-mediated endocytosis did not include a direct role for the cytoskeleton in the formation of the clathrin-coated vesicles. However, we now have good evidence that at least under some conditions dynamic actin filaments are critical for clathrin-mediated endocytosis. The first functional indications for actin’s role in clathrin-mediated endocytosis came from inhibitor studies that showed that actin has an important but variable role in this process in mammalian

M. Kaksonen (B) EMBL, Heidelberg, Germany e-mail: [email protected]

M.-F. Carlier (ed.), Actin-based Motility, DOI 10.1007/978-90-481-9301-1_4, C Springer Science+Business Media B.V. 2010

85

86

M. Kaksonen

cells (Fujimoto et al., 2000; Gottlieb et al., 1993; Lamaze et al., 1997; Salisbury et al., 1980). Genetic studies showed that actin is an essential player in endocytosis in budding yeast (Kübler and Riezman, 1993). I will begin this chapter by describing work done in yeast, Saccharomyces cerevisiae, which has turned out to be a good experimental system for studying actin-dependent endocytosis. I will then review the role of actin in endocytosis in mammals and finish by discussing common themes emerging from these studies.

4.1 Actin and Endocytosis in Yeast A lot of work on actin’s role in clathrin-mediated endocytosis has been done using the yeast Saccharomyces cerevisiae as a model organism. Yeast offers several important advantages for studies of the interplay of actin cytoskeleton and clathrinmediated endocytosis. Yeast cells lack the dense cortical actin cytoskeleton, which in mammalian cells gives structural support to the plasma membrane and maintains the cell shape. In yeast this supportive function is provided by the cell wall that lies outside of the plasma membrane. Furthermore, yeast cells are immotile and devoid of the extensive membrane-associated actin structures that power cell migration in animal cells. The yeast plasma membrane is thus mostly free of actin filaments, which makes it easy to microscopically detect local and transient polymerization of actin filaments at the endocytic sites. Indeed, these endocytic actin accumulations called actin patches are the most prominent actin structures in yeasts and were discovered long before the link to endocytosis was understood (Adams and Pringle, 1984). Only one endocytic pathway has been described in yeast, further simplifying the interpretation of experiments compared to mammalian cells that are known to use multiple different endocytic routes (Conner and Schmid, 2003). Most of the yeast proteins implicated in endocytic uptake are homologous to mammalian proteins involved in clathrin-mediated endocytosis (Engqvist-Goldstein and Drubin, 2003). Importantly, yeast endocytosis is strictly actin-dependent, facilitating the analysis of actin’s role. These features, together with the powerful genetics, make yeast an attractive model organism for studying the basic mechanisms of actin-driven endocytosis.

4.1.1 Stages of Yeast Endocytosis Although it was thought earlier that yeast endocytosis was fundamentally different from the process in mammals, it now seems that the overall stages of clathrinmediated endocytosis are very similar in these organisms. Live-cell imaging of fluorescently-tagged proteins has revealed that the process is initiated in yeast, as in mammalian cells, by recruitment of conserved endocytic adaptors, accessory proteins and clathrin to the plasma membrane (Carroll et al., 2009; Kaksonen et al., 2005; Newpher et al., 2005; Reider et al., 2009; Toshima et al., 2006). These early arriving proteins are likely to be involved in the selection of the endocytic site and in recruitment of cargo to the forming endocytic vesicle. After the

4

Actin in Clathrin-Mediated Endocytosis

87

initial assembly of the clathrin coat, a second wave of assembly brings in more coat-associated proteins including epsins, Sla2 (homolog of mammalian proteins Hip1 and Hip1R) and the Pan1/Sla1/End3-complex, which shares similar domain composition with mammalian Eps15 and intersectin (Kaksonen et al., 2003, 2005; Toret et al., 2008). About 1–2 min after the initiation of the coat assembly the membrane under the coat starts to form a tubular invagination. The coat proteins move inward at the tip of this invagination, which is about 50 nm wide and up to 200 nm long (Idrissi et al., 2008; Mulholland et al., 1994). Within 10 s after the formation of the invagination, a vesicle is pinched off from its tip. This primary endocytic vesicle is then rapidly uncoated and the proteins involved in the formation of the vesicle return to the soluble cytosolic pool to be recruited for another round of endocytosis. These steps of the endocytic internalization process can be visualized microscopically, by following the movement of fluorescently tagged coat proteins (Galletta et al., 2008; Idrissi et al., 2008; Kaksonen et al., 2005; Newpher et al., 2005).

4.1.2 Visualizing Actin in Yeast Endocytosis Live-cell imaging has shown that actin filaments polymerize locally at the endocytic sites and form a tiny actin patch during the final stages of endocytosis (Kaksonen et al., 2003). The actin patches are very transient, having lifetimes of only about 15 s. Imaging of actin patches together with coat proteins showed that the coat structure starts moving just after the actin polymerization has started and the amount of actin increases continuously during the movement of the coat (Kaksonen et al., 2003). The peak actin accumulation is reached in less than 10 s after which the actin signal starts rapidly decreasing signifying depolymerization. Soon after the actin signal has started to drop the actin patch is released from the membrane and starts moving faster and in relatively random manner corresponding to the scission and release of a new endocytic vesicle (Huckaba et al., 2004). Simultaneously the signal intensity of coat-associated proteins starts decreasing indicating that the vesicle is being uncoated (Toret et al., 2008). Immuno-electron microscopy studies have shown that the actin patches have a diameter of about 150 nm and they enclose the endocytic invaginations at the plasma membrane (Idrissi et al., 2008; Mulholland et al., 1994). The actin patches are composed of short (∼50 nm), cross-linked and branched actin filaments (Mulholland et al., 1994; Rodal et al., 2005; Young et al., 2004). However, due to the difficulty of visualizing individual actin filaments in electron microscopy, the exact organization of the filaments in situ in actin patches is still unknown. The movement, but not the assembly, of the coat can be reduced or blocked by altering actin polymerization dynamics by drugs or by mutations in cytoskeletal proteins such as actin nucleating, capping, cross-linking or depolymerizing proteins (Galletta et al., 2008; Kaksonen et al., 2003, 2005; Kim et al., 2006; Martin et al., 2006; Okreglak and Drubin, 2007; Sun et al., 2006). Actin polymerization is thus needed for the formation of the membrane invagination and possibly for the following vesicle scission step, but not for the assembly of the clathrin coat.

88

M. Kaksonen

4.1.3 Regulation of Actin Polymerization at Yeast Endocytic Sites The formation of the actin network in the yeast actin patches is dependent on the Arp2/3 complex (Winter et al., 1997). The Arp2/3 complex nucleates an actin filament and remains bound to its minus-end, while it also forms a branch by binding to another actin filament (Mullins et al., 1998). The Arp2/3 complex is not highly active alone but is controlled by nucleation promoting factors (NPFs), which bind to the Arp2/3 complex and stimulate its nucleation activity. Yeast has four different types of protein with NPF activities: Las17, Pan1, Myo3/5 and Abp1 (Engqvist-Goldstein and Drubin, 2003). Interestingly, each of these NPF proteins has a specific localization and assembly dynamics. Las17, a homolog of mammalian WASP family proteins, and Pan1, an Eps15-like protein, are the first NPFs to localize to the endocytic site (Kaksonen et al., 2003). They appear about 20 s before actin polymerization starts. Type I myosins Myo5 and its redundant homolog Myo3, and Abp1, an actin filament binding protein, arrive to the endocytic site when actin starts polymerizing (Jonsdottir and Li, 2004; Kaksonen et al., 2003; Sun et al., 2006). Immuno-electron microscopy studies have shown that these different NPFs have distinct localizations at the endocytic site (Fig. 4.1) (Idrissi et al., 2008). Las17 localizes to the neck of the endocytic invagination. Most of the Myo5 protein localizes to the base of the invagination, although in longer invaginations some myosin is also found at the tip. Pan1 is associated with the clathrin coat at the tip of the invagination. Abp1 localizes with the actin filaments surrounding the invagination and is less intimately associated with the membrane than the other NPFs. The spatial relationships of the different NPFs may change during endocytosis as the membrane bends and forms the invagination. Myo5, Las17 and Pan1 may initially be in close proximity, but when the membrane starts bending Pan1 moves away at the tip of the invagination (Kaksonen et al., 2003). The NPFs function in complex co-operation that is still incompletely understood. Las17 and Myo5 exhibit significantly higher NPF activities compared to Pan1 and Abp1 (Sun et al., 2006). Deletion of LAS17 or MYO3/5 gene pair causes a strong block in endocytosis (Geli and Riezman, 1996; Goodson et al., 1996; Madania et al., 1999). Mutants of Pan1 also lead to block of endocytosis but its complete deletion

Myo5

Myo5

Las17

Las17

Abp1

Abp1

Fig. 4.1 Localization of different Arp2/3 complex activators and actin at the endocytic site in yeast. The depicted endocytic invagination represents a stage briefly before vesicle scission

Plasma membrane Abp1

Abp1

Actin network

Abp1 Pan1 Abp1

Pan1

Pan1 Pan1

Abp1

4

Actin in Clathrin-Mediated Endocytosis

89

is lethal (Wendland et al., 1996). On the other hand, the lack of Abp1 leads only to a mild effect on endocytosis (Kaksonen et al., 2005). Studies with mutant versions of Las17, Myo5 and Pan1 that lack their NPF activities have provided some understanding about their individual roles (Galletta et al., 2008; Sun et al., 2006). These mutants cause much milder phenotypes than the complete gene deletions suggesting that the NPF activity is only one of the roles these proteins have. The NPF functions are partially overlapping, so that inhibition of any single protein’s NPF function does not block endocytic internalization. Only when NPF mutants are combined can endocytosis be prevented. When the NPF activity of Las17 was inhibited, the actin polymerization was significantly delayed, suggesting that Las17 is important in the initiation of the actin polymerization (Galletta et al., 2008; Sun et al., 2006). The myosins Myo3/5 appear to be critical later during the formation of the invagination and possibly during the vesicle scission (Galletta et al., 2008; Sun et al., 2006). Pan1’s NPF role is apparent only in mutants in which the NPF activity of Las17 is also inhibited. This combination led to further delay in the initiation of actin polymerization, while the internalization movement was still quite normal (Galletta et al., 2008; Sun et al., 2006). Although Abp1 also has an NPF activity in vitro, in cells lacking Abp1 the actin patches have more actin filaments and their disassembly takes longer (Kaksonen et al., 2005). The increased actin content of the patches may be due to Abp1’s inhibitory action toward Las17 (D’Agostino and Goode, 2005). The different activators of the Arp2/3 complex are thus co-operating in complicated and partially redundant manner. The spatially and temporally regulated recruitment of the NPFs to the endocytic site is still poorly understood, but some studies have revealed an interesting interplay between the different NPFs. The mechanisms that trigger the localization of the first arriving NPFs Las17 and Pan1 are not known. Las17, however, is needed for the recruitment of Vrp1, a homolog of mammalian WIP (Sun et al., 2006). Vrp1 then together with actin filaments recruits Myo5. Direct interaction between Vrp1 and Myo5 is also needed for Myo5’s NPF activity. Las17 and Pan1 localize independently of actin filaments, but the fourth NPF Abp1 localizes by binding to actin filaments at the endocytic site (Kaksonen et al., 2005; Sun et al., 2006). The NPFs thus form an assembly cascade where the Las17 initiates the recruitment of the later arriving Myo5 and Abp1. This NPF recruitment cascade has also been studied in detail in fission yeast Schizosaccharomyces pombe (Sirotkin et al., 2005).

4.1.4 Layers of Activation and Inhibition of Actin Polymerization The NPFs regulating actin filament nucleation via the Arp2/3 complex are themselves subject to further regulation by other endocytic factors. Especially Las17, the initiator of actin polymerization, can be either inhibited or activated by several other proteins that localize to the endocytic sites. Sla1 protein, that forms a complex with Pan1, can interact via its SH3 domains with Las17 to inhibit its NPF activity in vitro (Rodal et al., 2003). Cells lacking the gene for Sla1 showed increased accumulation of actin at the endocytic sites, but the coat internalization appeared otherwise

90

M. Kaksonen

normal (Kaksonen et al., 2005). However, these mutant cells also exhibited a long delay in initiation of actin polymerization, which suggests that Sla1 may also somehow facilitate the initiation of actin polymerization in vivo. Another regulator of Las17 is Bbc1, which in vitro inhibits Las17 by binding to it via an SH3 domain (Rodal et al., 2003). Deletion of Bbc1 led to accelerated coat internalization, possibly due to enhanced actin nucleation (Kaksonen et al., 2005). Another SH3 domain protein Bzz1 (a homolog of mammalian syndapin) was shown to be able to relieve the inhibition of Las17 by Sla1 (Sun et al., 2006). Abp1, although it can activate the Arp2/3 complex, may in fact act as a negative factor in cells. Abp1 binds tightly to the Arp2/3, but activates it only weakly and can thus inhibit Las17 by competing with it for Arp2/3 binding (D’Agostino and Goode, 2005). Pan1, the NPF that localizes to the vesicle coat, has been shown to be inhibited by another coat associated protein Sla2 (Toshima et al., 2007). Pan1 is also a target for Ark1 and Prk1 kinases (Zeng and Cai, 1999). This redundant pair of kinases is recruited late to the endocytic site by an interaction with Abp1 (Cope et al., 1999). These kinases promote the disassembly of several proteins from the endocytic site by phosphorylating them (Huang et al., 2003; Sekiya-Kawasaki et al., 2003; Watson et al., 2001; Zeng et al., 2001). The Ark1/Prk1 kinase pair thus forms an important negative feedback loop in the regulation of the endocytic machinery. These examples illustrate the complexity of regulation of the actin polymerization at endocytic sites, and underline the need for combined in vitro and in vivo approaches in unraveling the regulatory mechanisms. The complex loops of positive and negative regulation are probably responsible for the robustness of the spatial and temporal accuracy of the assembly and disassembly of the endocytic machinery.

4.1.5 Further Links Between the Membrane and the Actin Cytoskeleton In order to form an invagination, and to pinch off a vesicle, the force from the actin cytoskeleton has to be transmitted to the membrane. This probably requires direct physical links with the membrane via proteins in the coat and/or in the neck of the invagination. Among the coat-associated proteins, Pan1 and Sla2 are the best candidates for linking the coat and the actin cytoskeleton. Pan1 can, in addition to its Arp2/3 binding and stimulating activities, also directly bind to actin filaments (Toshima et al., 2005). Pan1 forms a complex with Sla1 and End3 and can interact with other coat proteins such as epsins via its EH domains (Tang et al., 2000; Wendland et al., 1999). Pan1 could thus link the membrane to the actin network. However, a mutant of Pan1 lacking the filament binding activity does not give a strong phenotype (Toshima et al., 2005). Another potential linker is Sla2, which directly interacts with the membrane by binding to PIP2 , via its ANTH domain located in the N-terminus of the protein. In its C-terminal part, Sla2 has a talinlike domain that binds to actin filaments. In addition, Sla2 can bind to clathrin and it also forms dimers. A deletion of SLA2 gene gives an intriguing phenotype that

4

Actin in Clathrin-Mediated Endocytosis

91

is consistent with the suggestion that Sla2 is a key linker between the membrane and the actin cytoskeleton. In Sla2 deficient cells endocytosis is strongly inhibited. Live-cell imaging showed that there is continuous actin polymerization at the endocytic sites, which leads to the formation of elongated actin patches, which are not able to move the coats (Kaksonen et al., 2003). However, a mutant version of Sla2 that lacks the actin binding domain can support an almost normal rate of endocytosis, indicating that Sla2 cannot be the only protein that can provide the connection (Wesp et al., 1997). The link between the coat and the actin network is thus likely to be maintained redundantly by several different proteins. In addition to the coat, actin may also need to be linked to the proteins in the neck of the invagination either to provide force to pinch off the vesicle or to couple actin polymerization dynamics to the timing of vesicle scission. One candidate for this connection is the yeast amphiphysin homolog Rvs167, which has been functionally implicated in scission (Kaksonen et al., 2005). Rvs167 has a membrane binding BAR-domain and a C-terminal SH3-domain. The SH3-domain of Rvs167 can interact with Abp1 and Las17 and thus could link the membrane at the invagination neck to the surrounding actin network (Colwill et al., 1999; Lila and Drubin, 1997). Finally, the myosins Myo5 and Myo3, which are located at the base of the endocytic invagination, have motor domains, which can exert force on the actin filaments or anchor them to the membrane. Myosin’s exact role is not understood, but it is essential for endocytosis in yeast (Jonsdottir and Li, 2004; Sun et al., 2006). In summary, there are multiple proteins that can form connections between the actin cytoskeleton and the membrane, and these proteins can regulate actin filament nucleation and physically couple the coat and the actin network. These connections are highly dynamic and evolve over the short lifetime of the endocytic machinery.

4.2 Actin and Clathrin-Mediated Endocytosis in Mammals Endocytosis has been studied in mammalian cells intensively for decades. The classical view of the process maintained that the vesicle coat together with some accessory proteins is sufficient to form the vesicle. The idea that actin has an important role in endocytosis in mammalian cells has emerged much more recently. The first functional indications that actin is important in mammalian cells came from drug studies. Experiments with drugs such as latrunculin A and cytochalasin D showed that blocking actin polymerization inhibited endocytosis in certain cell lines (Fujimoto et al., 2000; Gottlieb et al., 1993; Lamaze et al., 1997; Salisbury et al., 1980; Toshima et al., 2005). However, the importance of actin for clathrin-mediated endocytosis varied widely depending both on the cell line used and on whether the cells were adhering to a substrate or grown in suspension (Fujimoto et al., 2000). These studies showed that in mammalian cells actin is not always an essential component of the clathrin-mediated endocytosis machinery, but is required under certain conditions.

92

M. Kaksonen

4.2.1 Actin Dynamics in Living Mammalian Cells Live-cell imaging approaches have brought key advances in understanding of actin’s role in clathrin-mediated endocytosis. In a seminal study Merrifield and colleagues used a combination of TIRF (total internal reflection fluorescence) and epifluorescence microscopy to follow the behavior of clathrin at the endocytic sites in mouse 3T3 fibroblasts (Merrifield et al., 2002). Transient actin bursts were detected at the clathrin-coated pits (CCPs) in the bottom surface of adherent cells. TIRF illumination excites fluorescence only in a very thin slice (∼100 nm) at the bottom of the cell that is in contact with the cover glass. This facilitates detection of actin polymerization at the endocytic sites by excluding the strong background fluorescence from the cytoskeletal elements deeper inside the cell. Combined TIRF and epifluorescence imaging of clathrin allowed the investigators to determine the movement of the vesicle coat away from the cell surface, as the ratio of signals from TIRF and epifluorescence depends on the distance of the object from the cover glass surface. These experiments showed that transient actin plumes associated with ∼80% of internalizing CCPs (Merrifield et al., 2002). Actin polymerization started briefly before the clathrin coat started moving in and the peak actin intensity coincided with the maximum distance traveled by the coat. Further studies used an ingenious method to detect the timing of vesicle scission in living cells (Merrifield et al., 2005). A pH-sensitive GFP was fused to the extracellular domain of an endocytic cargo transferrin receptor. The cargo was then followed by TIRF microscopy while the observed cells were alternately perfused with medium of neutral or low pH to cyclically turn on and of the GFP signal. The cargo accumulated at the forming endocytic sites and became insensitive to the pH cycling when it was sealed in a newly formed vesicle. This method could detect vesicle scission with a temporal resolution of a few seconds (Merrifield et al., 2005). This approach revealed that the scission takes place when actin polymerization is at its maximum. Thus, similar to yeast, the accumulation of actin filaments starts just before the coat moves in. The disassembly of the actin network starts when the vesicle is pinched off, suggesting that there is a tight coupling between actin polymerization and the scission event.

4.2.2 Regulation of Actin Polymerization at Endocytic Sites The Arp2/3 complex localizes to the sites of clathrin-mediated endocytosis with dynamics that are very similar to the dynamics of actin (Benesch et al., 2005; Merrifield et al., 2004). Sequestering the Arp2/3 complex in cells with overexpressed Arp2/3 binding domain led to complete inhibition of actin plume formation at CCPs (Benesch et al., 2005). These results suggest that the Arp2/3 complex is responsible for nucleating the actin filaments at the endocytic sites. N-WASP and cortactin, two Arp2/3 activators, are recruited to the CCPs at the time of coat internalization (Benesch et al., 2005; Cao et al., 2003; Merrifield

4

Actin in Clathrin-Mediated Endocytosis

93

et al., 2004, 2005). N-WASP is a ubiquitously expressed member of the mammalian WASP protein family of Arp2/3 activators (Takenawa and Suetsugu, 2007). Cortactin, in addition to its Arp2/3 stimulating activity, can bind to actin filaments and to many other endocytic proteins, including dynamin, using its SH3 domain (Ammer and Weed, 2008). Inhibition of N-WASP function by gene deletion or RNA knockdown caused a reduction in the uptake of endocytic cargoes (Benesch et al., 2005; Innocenti et al., 2005). The N-WASP deficiency also reduced the frequency of recruitment of actin and the Arp2/3 complex to the CCPs (Benesch et al., 2005). Interestingly, N-WASP deletion did not completely prevent the formation of actin plumes at the CCPs like sequestering the Arp2/3 complex did, suggesting that the activation of the Arp2/3 complex at the endocytic sites is controlled redundantly by different NPFs (Benesch et al., 2005). Inhibition of cortactin’s function has also been shown to reduce the internalization of endocytic cargoes (Cao et al., 2003; Zhu et al., 2005). N-WASP, like its yeast homolog Las17, is under control of multiple other factors that can regulate its activity (Takenawa and Suetsugu, 2007). For example, N-WASP regulating proteins Nck1, Nck2, WIP and Abi1 are recruited to CCPs with similar timing compared to N-WASP (Benesch et al., 2005; Innocenti et al., 2005). BAR-domain proteins are another interesting group of endocytic factors that can regulate N-WASP. BAR-domains dimerize to form a concave interaction surface, which binds to membranes that have a specific curvature (Frost et al., 2008). BAR-domain proteins may also form larger oligomers that can drive membrane tubulation at endocytic sites and help in vesicle scission (Frost et al., 2009). Endophilin, amphiphysin, FBP17, CIP4, Toca-1, SNX9 and syndapin, which are BAR-domain proteins implicated in clathrin-mediated endocytosis, have recently been shown to use their SH3-domains to bind to N-WASP and stimulate N-WASPArp2/3-mediated actin polymerization (Dharmalingam et al., 2009; Ho et al., 2004; Otsuki et al., 2003; Tsujita et al., 2006; Yamada et al., 2009; Yarar et al., 2007). Interestingly, this stimulation of N-WASP activity by the BAR-domain proteins may depend on a specific membrane curvature at the endocytic site (Takano et al., 2008). This could be a mechanism that couples the regulation of actin nucleation to the changes in the membrane curvature during endocytosis. A crucial factor in clathrin-mediated endocytosis in mammalian cells is dynamin, a large GTPase that oligomerizes at the neck of the endocytic invagination and then constricts it, thus driving the final scission step. Dynamin interacts with many endocytic proteins that can link it to the actin cytoskeleton (Schafer, 2004). For example, dynamin can directly bind to cortactin, and can indirectly interact with N-WASP via binding to Nck or Intersectin (Schafer, 2004; Zhu et al., 2005). Furthermore, dynamin interacts with many BAR-domain proteins, such as amphiphysin, endophilin, syndapins that can regulate N-WASP activity (Takei et al., 2005). Over-expression of F-BAR domain proteins in cultured cells led to extensive tubulation of the plasma membrane (Itoh et al., 2005). The tubulation was inhibited by expressing dynamin and enhanced by blocking actin polymerization, suggesting a functional interplay between dynamin and actin in vivo. A very recent study has provided further understanding of this interplay (Ferguson et al., 2009). The

94

M. Kaksonen

authors used conditional knockout technology to generate mouse cell lines deficient of two redundant dynamin genes. Surprisingly, in these dynamin deficient cells N-WASP, cortactin, the Arp2/3 complex and actin strongly accumulated at CCPs. In addition, several BAR-domain proteins such as endophilin, amphiphysin and SNX9 accumulated at CCPs. Interestingly, the localization of the BAR-domain proteins was dependent on actin filaments. The lack of dynamin blocked vesicle scission, but also led to actin dependent formation of long BAR-domain protein covered membrane tubules that connected the CCPs to the plasma membrane. In wild type cells dynamin-mediated vesicle scission probably quickly ends the actin and BARdomain protein driven membrane invagination thus preventing the formation of long tubules and strong accumulation of actin and the BAR-domain proteins. These data suggest that actin, the BAR-domain proteins and dynamin are recruited in a sequential manner forming a functional cascade (Fig. 4.2). Cortactin, the other Arp2/3 activator at the mammalian CCPs, is regulated by Hip1R, a clathrin-binding protein, which is homologous to yeast Sla2. Hip1R was shown to bind directly to cortactin and counteract its actin polymerization activity by inhibiting actin barbed end elongation (Le Clainche et al., 2007). RNAi depletion of Hip1R blocked endocytosis and caused an accumulation of dynamic actin filaments at endocytic sites, a phenotype

time scission coat initiation & maturation

membrane invagination

vesicle release & uncoating

1. 2. 1) N-WASP/Las17 and Arp2/3 are recruited. Actin polymerization starts. 2) Actin & membrane bending recruit BAR proteins. 3) Dynamin is recruited. 4) Actin polymerization is turned off.

3. 4.

clathrin and other coat proteins

BAR-domain proteins

actin cytoskeleton

dynamin

Fig. 4.2 Stages of clathrin- and actin-mediated endocytosis. The model presents schematically the key stages of actin-mediated endocytis in relation to the changing membrane shape. The processes shown in the model are shared between mammals and yeast, except that in yeast the requirement for dynamin in scission has not been demonstrated

4

Actin in Clathrin-Mediated Endocytosis

95

closely resembling the SLA2 deletion phenotype in yeast (Engqvist-Goldstein et al., 2004). The Arp2/3, and two of its activators, N-WASP and cortactin, are thus the key factors in the regulation of actin polymerization at the sites of clathrin-mediated endocytosis. N-WASP and cortactin are themselves subject to regulation by a complex network of protein interactions that probably guarantees the tight spatial and temporal control of actin polymerization and couples it to the changes in the shape of the endocytic membrane.

4.2.3 Multiple Roles of Actin at the Plasma Membrane Local polymerization of actin at the endocytic site may be only one of the mechanisms by which actin affects clathrin-mediated endocytosis in mammalian cells. First, the actin cytoskeleton is important for the generation and regulation of membrane tension. Perturbation of the actin cytoskeleton reduces the tension of the plasma membrane (Sheetz and Dai, 1996). High membrane tension can potentially increase the force needed for the endocytic membrane invagination, and lowering the tension could reduce the force requirement. Indeed, in yeast reducing the membrane tension by altering the osmolarity of the growth medium was recently shown to help cells to endocytose when their actin cytoskeleton was genetically compromised (Aghamohammadzadeh and Ayscough, 2009). Secondly, the cortical actin cytoskeleton can participate in organizing the plasma membrane into different domains that could have different potential for supporting endocytosis. In addition, a dense cortical actin network may form a barrier that locally prevents endocytic vesicles from forming (Qualmann et al., 2000). Global perturbations of actin dynamics could thus have multiple effects on endocytosis. For example, inhibiting actin polymerization may reduce membrane tension and remove actin barriers, which would promote endocytosis, and simultaneously block the local actin polymerization at the endocytic site, which would inhibit endocytosis. The overall effect of actin perturbation on endocytosis could thus depend on multiple parameters that may be different in different cell types or under different growth conditions. A quantitative analysis of clathrin-coated structures in mammalian cells that had been treated with low levels of latrunculin A showed that multiple aspects of CCP dynamics were altered (Merrifield et al., 2005; Yarar et al., 2005). In addition to the reduction in the frequency of scission events, the formation of new CCPs and their lateral movements on the plasma membrane were also inhibited. However, inhibiting just the Arp2/3 dependent actin polymerization by sequestering the Arp2/3 complex prevented actin polymerization at the CCPs but did not appear to affect the formation of new CCPs (Benesch et al., 2005). The effect of latrunculin A on the formation of CCPs may thus be due to perturbation of other actin dependent processes such as cell adhesion, which has recently been shown to have a local effect on CPP dynamics and actin requirement (Liu et al., 2009; Saffarian et al., 2009).

96

M. Kaksonen

4.3 Open Questions in Actin-Mediated Endocytosis Our understanding of actin’s role during endocytic internalization is still quite incomplete, but some common themes are emerging from studies done in different experimental organisms. Actin polymerization, in both yeast and mammals, takes place during the final membrane invagination and vesicle scission steps and contributes to these processes functionally (Fig. 4.2). The regulation of actin polymerization at the endocytic sites appears to be very complex involving a large number of different interacting proteins. However, the Arp2/3 complex together with its key activator, N-WASP in mammals and Las17 in yeast, constitute the core machinery regulating actin polymerization. In addition to N-WASP and Las17, both yeast and mammals have other less conserved Arp2/3 activators and a large number of potential regulators of the Arp2/3 activators. An important challenge will be to find out how this regulatory complexity gives rise to the spatially and temporally precise stages of the endocytic process. Another major open question is the mechanism by which actin filaments drive endocytosis. Polymerization of actin filaments alone is thought to be able to push the membrane in migrating animal cells or propel pathogens such as Listeria monocytogenes in the cytoplasm of infected cells (Pollard and Borisy, 2003). Filament polymerization coupled to the thermal fluctuations of the propelled object or the filaments themselves is in principle enough to provide force for the movement via a mechanism called the Brownian ratchet (Peskin et al., 1993). When an actin filament grows into contact with the membrane it would stop polymerizing because there is no space for monomer addition. However, the Brownian motion of the membrane or the filaments would create the space needed for monomer addition thus forming the Brownian ratchet mechanism. In yeast the motor activity of the membrane associated type I myosins is also essential for endocytosis. Mutations in the motor domain of Myo5 block endocytic internalization, without preventing actin polymerization (Sun et al., 2006). The myosin motors could act instead or in addition to the Brownian motion to facilitate the actin filament-based ratchet mechanism by pushing the filament plus-ends away from the membrane to allow further addition of actin monomers to the filaments. Simplified in vitro systems together with mathematical modeling approaches are likely to be helpful in trying to solve these mechanistic questions. Another unresolved problem, related to the mechanism of actin-driven endocytosis, is the organization of the actin filaments surrounding the endocytic invagination. Two opposite models have been suggested (Fig. 4.3). The first model is analogous to the movement of Listeria bacteria or endosomes in the cytoplasm by actin “rocketing” mechanism, where the Arp2/3 complex nucleates actin filaments on the surface of the bacterium or endosome, and the growing filament plus-ends point toward the bacterium or endosome pushing it forward (Pollard and Borisy, 2003). In endocytic internalization the filament plus-ends would be growing toward the forming vesicle driving its internalization (Galletta et al., 2008; Galletta and Cooper, 2009) (Fig. 4.3a). The other model assumes an opposite orientation for the actin filaments where they would grow toward the cell surface and the minus-ends or the sides of the

4

Actin in Clathrin-Mediated Endocytosis

A

97

B Actin filament plus ends

Actin-membrane linker Plasma membrane

Fig. 4.3 Two simple models of actin filament organization at the endoctic sites. (a) Actin filament plus-ends point toward the forming vesicle and the filament polymerization takes place at the vesicle surface. (b) The growing actin filament plus-ends point toward the cell surface. The filament minus-ends are linked to the forming vesicle

filaments would be connected to the endocytic coat (Fig. 4.3b). This second model was based on experiments with mutant yeasts where two Las17 inhibitors, Bbc1 and Sla1, had been deleted. In these mutant cells actin patches are highly enlarged and photobleaching of actin-GFP revealed that actin filaments are being polymerized at the cell surface, not on the vesicle surface (Kaksonen et al., 2005). However, due to the small size of the endocytic actin structures in wild-type cells, the organization of actin filaments under normal conditions has remained unresolved. It is possible that the actin filaments are organized in a more complex manner than these simple models suggest (Suetsugu, 2009). There could, for example, be several populations of filaments that are oriented differently, and whose nucleation could be controlled by different Arp2/3 activators. Advanced methods of electron microscopy, such as cryo-electron microscopy, may shed light on this question in the future. Although a lot has been learned about actin’s role in endocytosis during the recent years, there is still a lot of work to be done before we understand the mechanistic details of the process. Combination of multiple approaches providing different types of data will likely be critical for future advances. Live-cell fluorescence imaging has proven to be a powerful method for obtaining dynamic information about the endocytic machinery at high temporal resolution. Electron microscopy will most likely still be the key method for acquiring information at high spatial resolution, especially when correlated with live-cell imaging data. In addition, biochemical approaches and in vitro experiments with reconstituted systems will be essential for achieving quantitative understanding of the molecular mechanisms of actin-driven endocytosis.

References Adams AE, Pringle JR (1984) Relationship of actin and tubulin distribution to bud growth in wildtype and morphogenetic-mutant Saccharomyces cerevisiae. J Cell Biol 98:934–45 Aghamohammadzadeh S, Ayscough KR (2009) Differential requirements for actin during yeast and mammalian endocytosis. Nat Cell Biol 11:1039–1042

98

M. Kaksonen

Ammer AG, Weed SA (2008) Cortactin branches out: roles in regulating protrusive actin dynamics. Cell Motil Cytoskeleton 65:687–707 Benesch S, Polo S, Lai FPL, et al. (2005) N-WASP deficiency impairs EGF internalization and actin assembly at clathrin-coated pits. J Cell Sci 118:3103–15 Cao H, Orth JD, Chen J, et al. (2003) Cortactin is a component of clathrin-coated pits and participates in receptor-mediated endocytosis. Mol Cell Biol 23:2162–70 Carroll SY, Stirling PC, Stimpson HEM, et al. (2009) A yeast killer toxin screen provides insights into a/b toxin entry, trafficking, and killing mechanisms. Dev Cell 17:552–60 Colwill K, Field D, Moore L, et al. (1999) In vivo analysis of the domains of yeast Rvs167p suggests Rvs167p function is mediated through multiple protein interactions. Genetics 152: 881–93 Conner SD, Schmid SL (2003) Regulated portals of entry into the cell. Nature 422:37–44 Cope MJ, Yang S, Shang C, et al. (1999) Novel protein kinases Ark1p and Prk1p associate with and regulate the cortical actin cytoskeleton in budding yeast. J Cell Biol 144:1203–18 D’Agostino JL, Goode BL (2005) Dissection of Arp2/3 complex actin nucleation mechanism and distinct roles for its nucleation-promoting factors in Saccharomyces cerevisiae. Genetics 171:35–47 Dharmalingam E, Haeckel A, Pinyol R, et al. (2009) F-BAR proteins of the syndapin family shape the plasma membrane and are crucial for neuromorphogenesis. J Neurosci 29:13315–27 Engqvist-Goldstein AE, Drubin DG (2003) Actin assembly and endocytosis: from yeast to mammals. Annu Rev Cell Dev Biol 19:287–332 Engqvist-Goldstein AEY, Zhang CX, Carreno S, et al. (2004) RNAi-mediated Hip1R silencing results in stable association between the endocytic machinery and the actin assembly machinery. Molecular Biology of the Cell 15:1666–79 Ferguson S, Raimondi A, Paradise S, et al. (2009) Coordinated actions of actin and BAR proteins upstream of dynamin at endocytic clathrin-coated pits. Dev Cell 17:811–22 Frost A, Perera R, Roux A, et al. (2008) Structural basis of membrane invagination by F-BAR domains. Cell 132:807–17 Frost A, Unger VM, De Camilli PV (2009) The BAR domain superfamily: membrane-molding macromolecules. Cell 137:191–96 Fujimoto LM, Roth R, Heuser JE, et al. (2000) Actin assembly plays a variable, but not obligatory role in receptor-mediated endocytosis in mammalian cells. Traffic 1:161–71 Galletta BJ, Chuang DY, Cooper JA (2008) Distinct Roles for Arp2/3 Regulators in Actin Assembly and Endocytosis. PLoS Biol 6:e1 Galletta BJ, Cooper JA (2009) Actin and endocytosis: mechanisms and phylogeny. Curr Opin Cell Biol 21:20–7 Geli MI, Riezman H (1996) Role of type I myosins in receptor-mediated endocytosis in yeast. Science 272:533–35 Goodson HV, Anderson BL, Warrick HM, et al. (1996) Synthetic lethality screen identifies a novel yeast myosin I gene (MYO5): myosin I proteins are required for polarization of the actin cytoskeleton. J Cell Biol 133:1277–91 Gottlieb TA, Ivanov IE, Adesnik M, et al. (1993) Actin microfilaments play a critical role in endocytosis at the apical but not the basolateral surface of polarized epithelial cells. J Cell Biol 120:695–710 Ho H-YH, Rohatgi R, Lebensohn AM, et al. (2004) Toca-1 mediates Cdc42-dependent actin nucleation by activating the N-WASP-WIP complex. Cell 118:203–16 Huang B, Zeng G, Ng AYJ, et al. (2003) Identification of novel recognition motifs and regulatory targets for the yeast actin-regulating kinase Prk1p. Mol Biol Cell 14:4871–84 Huckaba TM, Gay AC, Pantalena LF, et al. (2004) Live cell imaging of the assembly, disassembly, and actin cable-dependent movement of endosomes and actin patches in the budding yeast, Saccharomyces cerevisiae. J Cell Biol 167:519–30 Idrissi FZ, Grötsch H, Fernández-Golbano IM, et al. (2008) Distinct acto/myosin-I structures associate with endocytic profiles at the plasma membrane. J Cell Biol 180:1219–32

4

Actin in Clathrin-Mediated Endocytosis

99

Innocenti M, Gerboth S, Rottner K, et al. (2005) Abi1 regulates the activity of N-WASP and WAVE in distinct actin-based processes. Nat Cell Biol 7:969–76 Itoh T, Erdmann KS, Roux A, et al. (2005) Dynamin and the actin cytoskeleton cooperatively regulate plasma membrane invagination by BAR and F-BAR proteins. Dev Cell 9: 791–804 Jonsdottir GA, Li R (2004) Dynamics of yeast Myosin I: evidence for a possible role in scission of endocytic vesicles. Curr Biol 14:1604–9 Kaksonen M, Sun Y, Drubin DG (2003) A pathway for association of receptors, adaptors, and actin during endocytic internalization. Cell 115:475–87 Kaksonen M, Toret CP, Drubin DG (2005) A modular design for the clathrin- and actin-mediated endocytosis machinery. Cell 123:305–20 Kim K, Galletta BJ, Schmidt KO, et al. (2006) Actin-based motility during endocytosis in budding yeast. Mol Biol Cell 17:1354–63 Kübler E, Riezman H (1993) Actin and fimbrin are required for the internalization step of endocytosis in yeast. EMBO J 12:2855–62 Lamaze C, Fujimoto LM, Yin HL, et al. (1997) The actin cytoskeleton is required for receptormediated endocytosis in mammalian cells. J Biol Chem 272:20332–35 Le Clainche C, Pauly BS, Zhang CX, et al. (2007) A Hip1R-cortactin complex negatively regulates actin assembly associated with endocytosis. EMBO J 26:1199–210 Lila T, Drubin DG (1997) Evidence for physical and functional interactions among two Saccharomyces cerevisiae SH3 domain proteins, an adenylyl cyclase-associated protein and the actin cytoskeleton. Mol Biol Cell 8:367–85 Liu AP, Loerke D, Schmid SL, et al. (2009) Global and local regulation of clathrin-coated pit dynamics detected on patterned substrates. Biophys J 97:1038–47 Madania A, Dumoulin P, Grava S, et al. (1999) The Saccharomyces cerevisiae homologue of human Wiskott-Aldrich syndrome protein Las17p interacts with the Arp2/3 complex. Mol Biol Cell 10:3521–38 Martin AC, Welch MD, Drubin DG (2006) Arp2/3 ATP hydrolysis-catalysed branch dissociation is critical for endocytic force generation. Nat Cell Biol 8:826–33 Merrifield CJ, Feldman ME, Wan L, et al. (2002) Imaging actin and dynamin recruitment during invagination of single clathrin-coated pits. Nat Cell Biol 4:691–98 Merrifield CJ, Perrais D, Zenisek D (2005) Coupling between clathrin-coated-pit invagination, cortactin recruitment, and membrane scission observed in live cells. Cell 121:593–606 Merrifield CJ, Qualmann B, Kessels MM, et al. (2004) Neural Wiskott Aldrich Syndrome Protein (N-WASP) and the Arp2/3 complex are recruited to sites of clathrin-mediated endocytosis in cultured fibroblasts. Eur J Cell Biol 83:13–18 Mulholland J, Preuss D, Moon A, et al. (1994) Ultrastructure of the yeast actin cytoskeleton and its association with the plasma membrane. J Cell Biol 125:381–91 Mullins RD, Heuser JA, Pollard TD (1998) The interaction of Arp2/3 complex with actin: nucleation, high affinity pointed end capping, and formation of branching networks of filaments. Proc Natl Acad Sci USA 95:6181–86 Newpher TM, Smith RP, Lemmon V, et al. (2005) In vivo dynamics of clathrin and its adaptor-dependent recruitment to the actin-based endocytic machinery in yeast. Dev Cell 9: 87–98 Okreglak V, Drubin DG (2007) Cofilin recruitment and function during actin-mediated endocytosis dictated by actin nucleotide state. J Cell Biol 178:1251–64 Otsuki M, Itoh T, Takenawa T (2003) Neural Wiskott-Aldrich syndrome protein is recruited to rafts and associates with endophilin A in response to epidermal growth factor. J Biol Chem 278:6461–69 Peskin CS, Odell GM, Oster GF (1993) Cellular motions and thermal fluctuations: the Brownian ratchet. Biophys J 65:316–24 Pollard TD, Borisy GG (2003) Cellular motility driven by assembly and disassembly of actin filaments. Cell 112:453–65

100

M. Kaksonen

Qualmann B, Kessels MM, Kelly RB (2000) Molecular links between endocytosis and the actin cytoskeleton. J Cell Biol 150:F111–16 Reider A, Barker S, Mishra S, et al. (2009) Syp1 is a conserved endocytic adaptor that contains domains involved in cargo selection and membrane tubulation. EMBO J 28:3103–3116 Rodal AA, Kozubowski L, Goode BL, et al. (2005) Actin and septin ultrastructures at the budding yeast cell cortex. Mol Biol Cell 16:372–84 Rodal AA, Manning AL, Goode BL, et al. (2003) Negative regulation of yeast WASp by two SH3 domain-containing proteins. Curr Biol 13:1000–8 Saffarian S, Cocucci E, Kirchhausen T (2009) Distinct dynamics of endocytic clathrin-coated pits and coated plaques. PLoS Biol 7:e1000191 Salisbury JL, Condeelis JS, Satir P (1980) Role of coated vesicles, microfilaments, and calmodulin in receptor-mediated endocytosis by cultured B lymphoblastoid cells. J Cell Biol 87:132–41 Schafer DA (2004) Regulating actin dynamics at membranes: a focus on dynamin. Traffic 5:463–69 Sekiya-Kawasaki M, Groen AC, Cope MJTV, et al. (2003) Dynamic phosphoregulation of the cortical actin cytoskeleton and endocytic machinery revealed by real-time chemical genetic analysis. J Cell Biol 162:765–72 Sheetz MP, Dai J (1996) Modulation of membrane dynamics and cell motility by membrane tension. Trends Cell Biol 6:85–9 Sirotkin V, Beltzner CC, Marchand JB, et al. (2005) Interactions of WASp, myosin-I, and verprolin with Arp2/3 complex during actin patch assembly in fission yeast. J Cell Biol 170:637–48 Suetsugu S (2009) The direction of actin polymerization for vesicle fission suggested from membranes tubulated by the EFC/F-BAR domain protein FBP17. FEBS Lett 583:3401–3404 Sun Y, Martin AC, Drubin DG (2006) Endocytic internalization in budding yeast requires coordinated actin nucleation and myosin motor activity. Dev Cell 11:33–46 Takano K, Takano K, Toyooka K, et al. (2008) EFC/F-BAR proteins and the N-WASP-WIP complex induce membrane curvature-dependent actin polymerization. EMBO J 27:2817–28 Takei K, Yoshida Y, Yamada H (2005) Regulatory mechanisms of dynamin-dependent endocytosis. J Biochem 137:243–47 Takenawa T, Suetsugu S (2007) The WASP-WAVE protein network: connecting the membrane to the cytoskeleton. Nat Rev Mol Cell Biol 8:37–48 Tang HY, Xu J, Cai M (2000) Pan1p, End3p, and S1a1p, three yeast proteins required for normal cortical actin cytoskeleton organization, associate with each other and play essential roles in cell wall morphogenesis. Mol Cell Biol 20:12–25 Toret CP, Lee L, Sekiya-Kawasaki M, et al. (2008) Multiple pathways regulate endocytic coat disassembly in Saccharomyces cerevisiae for optimal downstream trafficking. Traffic 9:848–59 Toshima J, Toshima JY, Duncan MC, et al. (2007) Negative regulation of yeast Eps15-like Arp2/3 complex activator, Pan1p, by the Hip1R-related protein, Sla2p, during endocytosis. Mol Biol Cell 18:658–68 Toshima J, Toshima JY, Martin AC, et al. (2005) Phosphoregulation of Arp2/3-dependent actin assembly during receptor-mediated endocytosis. Nat Cell Biol 7:246–54 Toshima JY, Toshima J, Kaksonen M, et al. (2006) Spatial dynamics of receptor-mediated endocytic trafficking in budding yeast revealed by using fluorescent alpha-factor derivatives. Proc Natl Acad Sci USA 103:5793–98 Tsujita K, Suetsugu S, Sasaki N, et al. (2006) Coordination between the actin cytoskeleton and membrane deformation by a novel membrane tubulation domain of PCH proteins is involved in endocytosis. J Cell Biol 172:269–79 Watson HA, Cope MJ, Groen AC, et al. (2001) In vivo role for actin-regulating kinases in endocytosis and yeast epsin phosphorylation. Mol Biol Cell 12:3668–79 Wendland B, McCaffery JM, Xiao Q, et al. (1996) A novel fluorescence-activated cell sorter-based screen for yeast endocytosis mutants identifies a yeast homologue of mammalian eps15. J Cell Biol 135:1485–500 Wendland B, Steece KE, Emr SD (1999) Yeast epsins contain an essential N-terminal ENTH domain, bind clathrin and are required for endocytosis. EMBO J 18:4383–93

4

Actin in Clathrin-Mediated Endocytosis

101

Wesp A, Hicke L, Palecek J, et al. (1997) End4p/Sla2p interacts with actin-associated proteins for endocytosis in Saccharomyces cerevisiae. Mol Biol Cell 8:2291–306 Winter D, Podtelejnikov AV, Mann M, et al. (1997) The complex containing actin-related proteins Arp2 and Arp3 is required for the motility and integrity of yeast actin patches. Curr Biol 7: 519–29 Yamada H, Padilla-Parra S, Park S, et al. (2009) Dynamic interaction of amphiphysin with n-wasp regulates actin assembly. J Biol Chem 284:34244–34256 Yarar D, Waterman-Storer CM, Schmid SL (2005) A dynamic actin cytoskeleton functions at multiple stages of clathrin-mediated endocytosis. Mol Biol Cell 16:964–75 Yarar D, Waterman-Storer CM, Schmid SL (2007) SNX9 couples actin assembly to phosphoinositide signals and is required for membrane remodeling during endocytosis. Dev Cell 13:43–56 Young ME, Cooper JA, Bridgman PC (2004) Yeast actin patches are networks of branched actin filaments. J Cell Biol 166:629–35 Zeng G, Cai M (1999) Regulation of the actin cytoskeleton organization in yeast by a novel serine/threonine kinase Prk1p. J Cell Biol 144:71–82 Zeng G, Yu X, Cai M (2001) Regulation of yeast actin cytoskeleton-regulatory complex Pan1p/Sla1p/End3p by serine/threonine kinase Prk1p. Mol Biol Cell 12:3759–72 Zhu J, Zhou K, Hao JJ, et al. (2005) Regulation of cortactin/dynamin interaction by actin polymerization during the fission of clathrin-coated pits. J Cell Sci 118:807–17

Chapter 5

Actin Cytoskeleton and the Dynamics of Immunological Synapse Viveka Mayya and Michael L. Dustin

Abstract T lymphocytes traffic through tissues by different modes of amoeboid motility while scanning for antigenic peptides presented on the surface of AntigenPresenting Cells (APCs). Effective T-Cell Receptor (TCR) signaling, upon recognition of a peptide, leads to the formation of a cell-cell junction between the T lymphocyte and the APC. This junction has been termed as the immunological synapse (when it is radially symmetric and stable) or as the immunological kinapse (when it is asymmetric and mobile). Well-regulated coordination between the migratory behavior of T cells and the dynamics of immunological synapse is important for all the major facets of adaptive immune response. In this chapter we discuss how the actin cytoskeleton mediates this coordination by serving as a crucial platform for integrating and influencing signals from the TCR, adhesion and chemokine receptors. We also highlight key experimental approaches that have shaped our current understanding on the role of the actin cytoskeleton in the dynamics of the synapses and kinapses. Abbreviations APC DC IS IK SMAC MHC MHCp S1P GEF CTL

Antigen Presenting Cell Dendritic cell Immunological Synapse Immunological Kinpase supramolecular activation cluster Major Histocompatibility Complex peptide-loaded MHC Sphingosine-1-phosphate Guanine nucleotide exchange factor Cytotoxic T lymphocyte

M.L. Dustin (B) Skirball Institute of Biomolecular Medicine, New York University School of Medicine, 540 First Avenue, New York, NY 10016, USA e-mail: [email protected]

M.-F. Carlier (ed.), Actin-based Motility, DOI 10.1007/978-90-481-9301-1_5, C Springer Science+Business Media B.V. 2010

103

104

SMAC SPB TIRFM MC LBPA DIC LN TPM

V. Mayya and M.L. Dustin

Supra Molecular Activation Cluster Supported Planar Bilayer Total Internal Reflection Fluorescence Microscopy Microcluster Lyso bis phosphatidic acid Differential Interference Contrast Microscopy Lymph node Two-photon Microscopy

Contents 5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2 TCR Signaling, Adhesion and the Actin Cytoskeleton . . . . . . . . . . . . . . 5.2.1 Overview of TCR Signaling . . . . . . . . . . . . . . . . . . . . . . 5.2.2 Actin-Remodeling and -Binding Proteins Involved in TCR Signaling . . . . 5.2.3 Role of Actin Cytoskeleton in Integrin-Mediated Adhesion Upon TCR Engagement . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3 Immunological Synapse . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3.1 Organization of a Stable IS . . . . . . . . . . . . . . . . . . . . . . . 5.3.2 Peripheral TCR Microclusters are Involved in Sustained TCR Signaling . . . 5.3.3 SMACs Correspond to Modules of the Motility Apparatus . . . . . . . . . 5.4 Immunological Kinapse . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4.1 T Cell Activation Occurs Even in the Absence of a Stable IS . . . . . . . . 5.4.2 Immunological Kinapse is a Mobile and Asymmetric Counterpart to IS . . . 5.5 Role and Significance of Immunological Synapse and Kinapse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.5.1 Directed Secretion . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.5.2 Assymetric Cell Division . . . . . . . . . . . . . . . . . . . . . . . . 5.5.3 Balance Between IS and IK in Tuning Immune Responses . . . . . . . . . 5.6 Issues and Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . 5.6.1 TCR Triggering and Clustering . . . . . . . . . . . . . . . . . . . . . 5.6.2 Diversity in IS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.6.3 Symmetry Breaking of IS . . . . . . . . . . . . . . . . . . . . . . . 5.6.4 Balance Between IS and IK . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

104 106 106 106 107 109 109 110 111 114 114 115 115 115 116 116 117 117 118 119 120 120

5.1 Introduction The adaptive immune system is a collection of highly mobile cells that move between strategically placed vascular organs that evolved to sense and act against dangerous foreign agents that have penetrated the physical barriers between the environment and body of vertebrates (Abbas et al., 2000). A “healthy” adaptive immune system is also geared towards eliminating cancerous cells, controlling

5

Actin Cytoskeleton and the Dynamics of Immunological Synapse

105

resident microbiota, and avoiding self-injury (Abbas et al., 2000). All phases of the immune response depend on leukocyte migration and cell-cell interactions in both lymphoid organs and peripheral tissues (Friedl and Weigelin, 2008). Let us take the case of T lymphocytes (or T cells) in some detail, which is the subject of this chapter. Naive T cells in circulation home into the secondary lymphoid organs under the influence of endothelial presented adhesion molecules and chemokines (Cyster, 2005). Within the T-cell area of the secondary lymphoid tissues, they move about extensively by amoeboid motility, while scanning for the presence of antigens on the surface of dendritic cell (DC) networks (Dustin, 2004). In most cases, T cells do not detect their agonistic antigens and after an average period of 12–18 h, migrate out into circulation by developing responsiveness to Sphingosine-1-phosphate (S1P) gradients near cortical lymphatics (Cyster, 2005). But upon antigen encounter, the migration patterns change drastically in the secondary lymphoid organs to initiate an immune response. Activated DCs enter lymphatics and move to lymph nodes that drain the affected tissues and present peptide antigens loaded on to Major Histocompatibility Complexes (MHC). Specific T cell clones recognize these antigens through the T Cell Receptor (TCR). An immediate response of TCR signaling is to transiently shut down the exit program mediated by S1P (Pham et al., 2008). The longer term outcomes of TCR signaling are clonal expansion of antigen specific T cells and secretion of cytokines (Abbas et al., 2000). Expanded clones continually interact with antigen bearing DCs and finally differentiate into effector T cells. During this process, some of the primed T cells migrate in response to chemokines of the B-cell follicles and position themselves at the boundary between the T and B cell zones. There, they recognize and activate specific antigen-presenting B cells to secrete antibodies. After several rounds of division, the effector T cells reacquire their responsiveness to S1P and also become equipped to home into affected tissues. These effector T cells, depending on the sub-type, either engage antigen-bearing macrophages to activate them or recognize affected cells presenting the specific antigens to kill them (Abbas et al., 2000). It has recently been proposed that effector T cells interact with dendritic cells in tissue to maintain or adjust effector phenotype at the infection site. Thus, periods of migratory behavior are punctuated by cell-cell interactions between a T cell and an antigen-presenting cell (APC) throughout the immune response. As explained in later sections, the actin cytoskeleton is mainly responsible for the coordination between these (alternating) phases of the T cell and also for helping the rare antigen specific T-cells respond to antigens in the midst of copious amounts of self-peptides in a context-dependent manner (Dustin, 2008a). The junction mediating T cell-APC interactions has been termed as “Immunological Synapse” (IS) or “Immunological Kinapse” (IK), depending on its internal organization and dynamics (Dustin, 2008b). In this chapter we discuss how these cell junctions are manifestations of the actin cytoskeleton apparatus that drives amoeboid motility. We have also highlighted key experimental approaches that have shaped our current understanding of the role of the actin cytoskeleton in the dynamics of the IS. We hope the reader gets to appreciate how the actin cytoskeleton serves as a crucial platform for integrating and influencing signals from the TCR, adhesion and chemokine receptors.

106

V. Mayya and M.L. Dustin

5.2 TCR Signaling, Adhesion and the Actin Cytoskeleton 5.2.1 Overview of TCR Signaling The TCR complex recognizes the peptide-loaded MHC on the surface of an APC along with a coreceptor, either CD4 or CD8, which also mark the two major subtypes of T cells (Abbas et al., 2000). A tremendous amount of molecular information has been gathered on TCR signaling from extensive studies over the last two decades using transformed T cell lines and antigen-specific mouse T-cells from transgenic mice (Abraham and Weiss, 2004). There are several intricate and emergent features to the TCR signaling network (Das et al., 2009; Prasad et al., 2009). We refer to multiple reviews and only mention salient points here (Abbas et al., 2000, Weiss and Littman, 1994; Huse, 2009). The TCR complex has multiple CD3 chains that link to tyrosine kinases upon recognition of agonistic peptide-loaded MHC (MHCp). Activation and association of multiple non-receptor tyrosine kinases with the CD3 chains sets in motion hundreds of discrete tyrosine phosphorylation events (Mayya et al., 2009). Multiple T-cell specific or hematopoietic lineage-specific adapters are engaged in complexes mediated by SH2/SH3 domains, leading to the activation of multiple phospholipases, guanine-nucleotide exchange factors (GEFs), and small GTPases. Calcium (Ca2+ ) influx is a major feature of effective TCR signaling. All these events lead to activation of PKC and MAPK isoforms that then activate ‘Fos/Jun’ transcription factors. Isoforms of NF-kB and NFAT family of transcription factors also play a major role in tuning the transcriptional profile in activated T cells that eventually modulate the immune response. TCR signaling also brings about many other intracellular changes, most of which involve the actin cytoskeleton and have important immunological consequences (Mayya et al., 2009; Penninger and Crabtree, 1999; Billadeau et al., 2007).

5.2.2 Actin-Remodeling and -Binding Proteins Involved in TCR Signaling TCR signaling leads to increased polymerization and turn-over of actin, resulting in polarization of cellular F-actin towards the IS. Preferential activation of Rac and Cdc42 small GTPases at or near the IS through GEFs like Vav, Dock2 and PIXα is mainly responsible for the polarization of F-actin (Mayya et al., 2009; Billadeau et al., 2007). Pharmacological inhibition studies have showed that the dynamic nature of actin cytoskeleton is important for multiple processes in an engaged T cell: triggering, clustering and endocytosis of the TCR, patterning of surface proteins in the IS, activation of integrins, influx of Ca2+ , and other downstream signaling events necessary for cytokine transcription (Mayya et al., 2009; Penninger and Crabtree, 1999; Billadeau et al., 2007).1 Actin-remodeling proteins that show

1 Mayya et al. (2009) includes lists of journal articles on hundreds of functional studies relevant to TCR signaling and T cell activation. Please look into the references there in for details.

5

Actin Cytoskeleton and the Dynamics of Immunological Synapse

107

preferential expression in T lymphocytes or hematopoietic cells like WASp, Wave2, HS1, EVL, and most recently coronin 1A have garnered more attention (Billadeau et al., 2007; Burkhardt et al., 2008). Wave2 appears to coordinate the activity of both the Arp2/3 and formin type actin polymerization nucleators down stream of Rac. HS1 is a homolog of cortactin and is a prominent tyrosine phosphorylated protein in T cells that contributes to actin based protrusions. Up to 18 actin-regulatory and/or actin-binding proteins have been implicated in the above mentioned processes by studies using small-molecule inhibitors, variants with domain truncations and variable expression strategies coupled with imaging and/or wide variety of assays (Mayya et al., 2009; Billadeau et al., 2007). As explained in later sections, multiple roles in T cell activation have been ascribed to some of these proteins as per the context of the experiments. There is evidence to suggest that some proteins have functions in TCR signaling independent of regulating actin cytoskeleton as in the case of Vav, WAVE2 and WIP/WASp complexes. Tyrosine phosphorylation of HS1, WASP (both by Lck), and Vav (by Zap70) during TCR signaling increases their activity (Mayya et al., 2009; Billadeau et al., 2007). Serine phosphorylation of cofilin (at S3) and L-plastin (at S5) have also been demonstrated during TCR signaling (Mayya et al., 2009). Kinases such as Itk, Pak1, and PKCθ have been implicated in regulating actin cytoskeleton during TCR signaling (Mayya et al., 2009; Billadeau et al., 2007). A recent phosphoproteomic analysis conducted on Jurkat leukemic T cells revealed that the class of non-motor actin binding proteins is the most over-represented class of proteins with TCR-responsive phosphorylation sites (Mayya et al., 2009). In all, 37 actin-binding proteins were found to contain TCR-responsive phosphorylation site(s), most of which were novel. The majority of these phosphorylation sites are conserved across phyla, and onethird are within the known actin-binding domain(s) of these proteins. Proteins with novel TCR-responsive phosphorylation sites cover the full spectrum of remodeling properties on actin. Further, 11 modulators of Rho family of GTPases were found to harbor TCR-responsive phosphorylation sites (Mayya et al., 2009). These results bring to light previously unanticipated extent of influence of phosphorylation in remodeling actin cytoskeleton upon initiating TCR signaling.

5.2.3 Role of Actin Cytoskeleton in Integrin-Mediated Adhesion Upon TCR Engagement T cell adhesion to the APC is necessary to bring the two membranes at close apposition and allow the physical interaction between surface-bound MHCp and TCR (Shaw and Dustin, 1997). Adhesive interactions are also needed to overcome the barrier presented by the negatively charged glycocalyx around the cells (Shaw and Dustin, 1997). Adhesive mechanisms increase the sensitivity to MHCp by ∼100-fold, which is important for physiological detection of peptide antigens and mounting an effective immune response (Bachmann et al., 1997). Studies in the 1970s showed that adhesion of cytotoxic T lymphocytes (CTLs) to target cells

108

V. Mayya and M.L. Dustin

is antigen-specific, implying TCR mediated adhesion. The discovery of lymphocyte adhesion molecules led to some confusion as to how antigen specificity could be reconciled with expression of large numbers of active adhesion molecules that lacked antigen specificity. This paradox was resolved by the demonstration that TCR crosslinking increased the avidity of LFA1, a heterodimer of αL and β2 integrins subunits, and the major integrin on T cells, to ICAM1 on APCs (Dustin and Springer, 1989). Time-course and pharmacological inhibition experiments suggested that TCR signaling events increase the integrin avidity by “inside-out” mechanisms. Further, it was shown that motile T cells on an ICAM1 substratum are arrested by TCR engagement, leading to a proposal that TCR signaling to LFA1 is a “stop signal” for a motile T cell (Dustin et al., 1997). LFA1 is also important for T lymphocyte homing into tissues under the influence of chemokines. These early studies have prompted extensive investigations on integrins over the last two decades. The unifying theme emerging from these studies is that, in the physiological setting, the actin cytoskeleton collaborates with chemokine receptors and LFA1 to ensure effective TCR signaling as will be elaborated below. Resting T cells in the circulation are spheroid with submicron microvilli projections (Burkhardt et al. 2008; Dustin and Cooper, 2000). At this stage, LFA-1 displays low avidity. LFA-1 also displays poor lateral mobility as its bridging molecule, talin, is bound to the cortical actin shell (Dustin and Cooper, 2000). Under the influence of chemokine signaling, T cells become polarized. They acquire a leading edge with active membrane protrusions and a uropod with the collapsed and shrunken intermediate filament and cortical actin shell (Dustin, 2004; Dustin and Cooper, 2000). Chemokine receptor signaling leading to the activation of Rac, Rho and myosin IIA have been shown to be important for these morphological changes (Xu et al., 2003). Simultaneously, LFA-1 is released from inhibitory cytoskeletal interaction, making it freely mobile and available for binding to ICAM-1 (Dustin and Cooper, 2000). The leading edge with dynamic membrane protrusions, driven by new rounds of actin polymerization, has been shown to be more sensitive (responsive) to both MHCp and ICAM1 (Wei et al., 1999). Weak adhesion contributes to the amoeboid motility of T cells as well as in initiating cell-cell junctions (Dustin, 2008b). Triggering of TCR culminates in inside-out signaling that further increases the avidity of LFA-1 and contributes to the arrest of the T cell (Burbach et al., 2007). Several proteins involved in TCR-mediated inside-out signaling have been discovered and characterized in the last decade. Small GTPase Rap1 is the central mediator in inside-out signaling. TCR signaling leads to activation and plasma membrane localization of Rap1 by multiple means (Menasche et al., 2007). Most important among them is the activation of C3G, a GEF for Rap1 in T cells, by a pathway involving Vav and Wave2 (Nolz et al., 2008). Rap1 has multiple effectors for integrin activation in T cells. Predominant among them is RIAM, which is coupled to upstream TCR signaling complex containing LAT and SLP-76 by adaptors ADAP and SKAP-55 (Burbach et al., 2007; Menasche et al., 2007; Peterson et al., 2001). RIAM binds talin and induces conformational changes in the β2 chain of LFA-1 (Burbach et al., 2007; Tadokoro et al., 2003). Wave2 has also been implicated in

5

Actin Cytoskeleton and the Dynamics of Immunological Synapse

109

recruiting vinculin and talin to the IS and thereby contribute to inside-out signaling (Nolz et al., 2009). RIAM and ADAP are also known to bind EVL (Menasche et al., 2007). Thus, multiple actin-remodeling and actin-binding proteins are involved in inside-out integrin activation. CD2 is another receptor on the T cells involved in adhesion to APCs by binding to CD58 in humans and CD48 in rodents (Springer, 1990). Both LFA1 and CD2 have been shown to contribute to TCR signaling events and thus attributed costimulatory roles (Dustin, 2007).

5.3 Immunological Synapse 5.3.1 Organization of a Stable IS T cell-B cell conjugates and CTL conjugates with target cells have been studied by different microscopy approaches since the 1980s (Dustin, 2009). Initial studies highlighted features such as adhesion, polarization of the cytoskeleton and the secretory apparatus. Springer predicted that adhesion receptors CD2 and LFA1, and TCR get self-segregated into different regions in the contact area based on the molecular sizes of the ligand-receptor pairs (Springer, 1990). He proposed further that large glycoproteins such as CD43 and CD45 are excluded from the contact area. Kupfer and colleagues acquired fluorescence image stacks of fixed and stained T cell-B cell conjugates (Monks et al., 1998). They performed deconvolution of image stacks to resolve the contact area and revealed micron-scale organization patterns. The noncovalent ensembles containing hundreds to thousands of the ligand-receptor pairs were called Supra-Molecular Activation Clusters (SMACs) as some of the downstream signaling proteins also localized to these assemblies. The bull’s eye cluster of MHCp-TCR was termed the central SMAC (cSMAC) and the surrounding ring of ICAM1-LFA1 was termed the peripheral SMAC (pSMAC) (Fig. 5.1). Kupfer’s group later defined distal SMAC (dSMAC) as the outermost ring structure which, is enriched in CD45 (Freiberg et al., 2002). We used the supported planar bilayer (SPB) system to present laterally mobile, fluorescent dye-conjugated and GPI-anchored MHCp and ICAM1 for recognition by primary T-cells (Dustin et al., 2007). This system allowed time-lapse imaging of the formation of the IS (Grakoui et al., 1999). Initially there is a symmetrical spreading and adhesion with concomitant formation of LFA-1 adhesion zones at the center and TCR signaling clusters in the periphery. But gradually over 10 min, the TCR clusters coalesce at the center with the peripheral LFA-1 ring, recapitulating the pattern observed in fixed T cell-B cell conjugates. This organization of the IS can remain stable for over an hour. Thus, the early studies reinforced the notion of the IS being an organized stable junction responsible for sustained signaling in priming of naïve T cells and cell polarization and directed secretion in effector T cells. The role of the cSMAC as an activation cluster has been controversial and delving into this issue has revealed significant complexity of this sub-cellular region

110

V. Mayya and M.L. Dustin

Fig. 5.1 Supramolecular organization in a stable immunological synpase. Schematic side-on and front views of an IS formed by a T cell with a SPB presenting ICAM1 and MHCp. The interface can be divided into three supra molecular activation clusters (SMACs) based the segregation of certain membrane proteins, as denoted in the figure. TIRFM on SPBs provides better resolution and contrast allowing the visualization of microclusters (MCs) of TCR (and LFA-1)

(Lee et al., 2003; Varma et al., 2006). The central region of cSMAC contains a large cluster of inactive TCRs destined for internalization and degradation (Section 5.3.2). The annular region of cSMAC shows enrichment of CD28 and PKCθ, forming a compartment for costimulatory signaling (Fig. 5.1) (Yokosuka et al., 2008; Tseng et al., 2008). CD2 is also organized as an annular ring on the inner side of pSMAC (Dustin et al., 1998). Thus, the pSMAC and cSMAC can be further subdivided into domains with distinct compositions and functions. Formation of the IS is dependent on the actin cytoskeleton, as treatment with cytochalasin D or latrunculin A inhibits the initial contact formation (Grakoui et al., 1999). Bridging molecules CD2AP and talin link the adhesion receptors CD2 and LFA1 respectively to the actin cytoskeleton (Mayya et al., 2009; Dustin et al., 1998). These play an important role in the patterning observed at the IS. Large glycoproteins such as CD43 and PSGL1 are excluded from the IS primarily due to their linkage to the actin cytoskeleton through moesin (Mayya et al., 2009; Delon et al., 2001). Though multiple adaptors link the TCR with the actin cytoskeleton, phosphorylation of Y153 on CD3ζ has been implicated in direct association of TCR with F-actin (Mayya et al., 2009). Thus, the actin cytoskeleton and multiple actin-binding proteins are involved in the patterning observed at the IS.

5.3.2 Peripheral TCR Microclusters are Involved in Sustained TCR Signaling Studies using anti-MHCp antibodies to stop new MHCp-TCR interactions indicated that new MHCp-TCR interactions are needed continuously for sustained signaling (Valitutti et al., 1995). But MHCp-TCR interaction in the cSMAC region appeared to be stable based on contact-area FRAP experiments (Grakoui et al., 1999). This conflicted with the notion of cSMAC being the region of sustained TCR signaling in a stable IS. Further, phosphorylated forms of relevant kinases and adapters showed more punctate staining in the pSMAC and dSMAC (Lee et al., 2002). Mechanism of sustained TCR signaling was revealed by observing the dynamics of IS on SPBs

5

Actin Cytoskeleton and the Dynamics of Immunological Synapse

111

using Total Internal Reflection Fluorescence Microscopy (TIRFM) (Varma et al., 2006; Campi et al., 2005; Yokosuka et al., 2005). In TIRFM the light is reflected off of the cell-contact area due to the acute angle of the incident laser beam and the large difference in refractive index. Thus fluorescence excitation takes place due to an evanescent wave up to a depth of ∼200 nm into the cell, increasing the signal-to-noise ratio and contrast. This mode of fluorescence microscopy revealed that microclusters (MCs) of TCR are formed at the initial contact areas and later on, in peripheral contact areas continuously. These MCs are small with on the order of 10 TCR molecules depending upon the MHCp density and the time of stimulation, moved centripetally and reached the cSMAC in ∼2 min (Varma et al., 2006). The addition of anti-MHCp antibodies immediately stopped the formation of new MCs. The rate of decay in Ca2+ influx matched that of the loss of extant TCR MCs from the dSMAC and pSMAC areas when new MC formation was blocked by addition of anti-MHCp antibodies. In contrast, the cSMAC structure takes 20–30 min to dissolve after the cessation of Ca2+ influx. Further evidence for these peripheral MCs being the source of sustained signaling comes from co-localization of phosphorylated forms of the adaptors with peripheral MCs in fixed cells (Campi et al., 2005). TIRFM of live cells also revealed that Zap70 and SLP76 are co-localized and transported with the peripheral TCR MCs, but dissociate away as the MCs reach the cSMAC (Campi et al., 2005; Yokosuka et al., 2005). Interestingly, the cSMAC region stains for tyrosine phopshorylation in CD4 T cells from CD2AP deficient mice, in contrast to those from the wild-type mice (Lee et al., 2003). CD2AP is also involved in the internalization and degradation of TCR (Mayya et al., 2009). Further, the cSMAC area stains for Lysobisphosphatidic acid (LBPA), which is abundant in the multivesicular bodies (MVBs) (Varma et al., 2006). MVBs are the sites for sorting internalized receptors that are marked for degradation. All these results have lead to the proposal that TCR proximal signaling is initiated and sustained by peripheral MCs and terminated at the cSMAC. The formation of MCs is also dependent on constituent F-actin (Varma et al., 2006; Campi et al., 2005). But once formed, the MCs, especially those closer to the cSMAC, do not dissolve (or dissipate) with latrunculin A treatment. In contrast, the signaling activity and transport of MCs is continuously dependent on the actincytoskeleton (Varma et al., 2006). What is the basis for the centripetal movement of MCs in a stable IS?

5.3.3 SMACs Correspond to Modules of the Motility Apparatus An amoeboid cell can be thought of as consisting three functional actin cytoskeleton modules: lamellipodium at the front is the sensory module; lamella in the middle is the drive module; and uropod at the rear end is the detachment (-cum-retraction) module (Dustin, 2008a). The lamellipodium is rich in F-actin. The lamella is enriched in adhesion sites, whereas the uropod is rich in myosin II (Xu et al., 2003). The lamellipodium is the site of new polymerization and turnover of F-actin. Active membrane protrusions make the lamellipodium a highly sensitive region for forming

112

V. Mayya and M.L. Dustin

receptor-ligand interactions or maybe better locations for forming such interactions (Dustin and Cooper, 2000). Actin polymerization pushes against the membrane tension and contractile activity of myosin II leading to force generation and retrograde flow of F-actin from the lamellipodium (Dustin, 2004; Kruse et al., 2006). The cell is propelled forward when the retrograde flow of F-actin is coupled to the formation of adhesion sites on an immobile substratum (Kruse et al., 2006). During cellular movement, the adhesion sites themselves exhibit a cyclical pattern of formation at the probing lamellipodium, maturation in the lamella coupled with backward movement due to the retrograde flow of F-actin, and disengagement near the uropod (Dustin, 2004). Several pieces of evidence suggest that the cSMAC, pSMAC, and dSMAC of the IS correspond to uropod, lamella, and lamellipodium of a motile cell, respectively (Fig. 5.2a).

Fig. 5.2 Schematic of the organization and dynamics of actin cytoskeleton modules in the immunological synapse (IS) and kinapse (IK). (a) Representation of F-actin layers (colored red and green) in IS and IK. Note the difference in the orientation of the microtubule array, microtubule organizing center, and vesicles of the secretory apparatus between a T cell with IS and IK. Equivalence of the three SMACs with the three F-actin motility modules is also represented. IS represents the radially symmetrical manifestation of the motility module. When the symmetry is broken, the cell moves in the direction of lamellar sheet. (b) Illustration of contractile oscillations in the lamellipodium. Protrusion occurs by F-actin polymerization in the lamellar sheet (red). The centripetal flow of F-actin is shown by the arrow. A second F-actin layer (green) also starts to form on top of the first. The top layer is subjected to myosin II contraction at its rear end (denoted by the blue arrow). This causes the lower F-actin layer to buckle (3rd panel from top). This in turn helps stabilize nascent adhesion sites. The buckling continues until the upper layer is separated from its nascent growing end at the front, which then starts another cycle of extension. Weaker adhesion during contraction results in a membrane ruffle as the actin layers are lifted off the surface (bottom three panels). The protrusion and retraction cycle also propagates as a wave around the periphery of the cell

5

Actin Cytoskeleton and the Dynamics of Immunological Synapse

113

Sheetz and colleagues have used symmetric spreading of fibroblasts on fibronectin coated glass as an accessible model system to study the coordination between lamella and lamellipodium (Giannone et al., 2007). They also manipulated the activity of myosin II by pharmacological and genetic means to get insights into the role of myosin contraction. They compared and contrasted time-lapse data from Differential Interference Contrast (DIC), epiflourescence, TIRFM, and electron microscopy. The main conclusion of the study is that the contractile force due to myosin II activity on the top layer of growing F-actin is responsible for the dynamics of the lamellipodium (Fig. 5.2b). This dynamic behavior of the actin cytoskeleton at the periphery was therefore termed “contractile oscillations” (Giannone et al., 2007). The contractile oscillations are important for mechano-sensing of the substratum by the lamellipodium (Giannone et al., 2007). IS of naïve T cells formed on SPBs also exhibit contractile oscillations as revealed by edge-tracking algorithms and correlation analysis (Sims et al., 2007). T cells interacting with APCs also generate active membrane protrusions and dynamic ruffles at the periphery that seem to scan the surface of the APC for new MHCp (Valitutti et al., 1995; Tskvitaria-Fuller et al., 2003). The lower layer of F-actin corresponds to the lamella, whereas the upper layer combined with its portion of the lower layer corresponds to the lamellipodium (Fig. 5.2) (Giannone et al., 2007). The lamellar actin layer contains more tropomyosin and talin, where as the upper layer in the lamellipodium is enriched in cofilin. Correspondingly, tropomyosin and cofilin stain the pSMAC and dSMAC regions, respectively (Sims et al., 2007). Talin has been a classical marker of pSMAC (Monks et al., 1998). Just as the lamella is rich in integrin adhesion sites, the pSMAC is also the region in which LFA-1 clusters accumulate (Kaizuka et al., 2007). The Uropod is depleted in F-actin. The cSMAC is also relatively free of F-actin based on phalloidin staining of fixed primary T cells, speckle microscopy in Jurkat T cells expressing GFP-actin, and electron micrographs of CTL conjugates (Sims et al., 2007; Kaizuka et al., 2007). We have demonstrated the centripetal flow of actin in IS by speckle microscopy in Jurkat T cells (Kaizuka et al., 2007). We presented monovalent anti-CD3ε antibodies on SPBs by bridging biotinylated anti-CD3 to streptavidin molecules linked to the bilayer through biotinylated lipids to form IS in Jurkats. Just as the adhesion complexes form at the lamellipodium and move backwards with retrograde actin flow, both TCR and LFA-1 MCs form in the dSMAC region and move towards the centre in IS (Kaizuka et al., 2007). The role of myosin II in the IS also appears to be analogous to that in a motile cell. Myosin activity is necessary for forming a mature, stable and adherent IS (Ilani et al., 2009). Myosin activity is also necessary for the centripetal movement of TCR MCs presumably by regulating the centripetal actin flow (Ilani et al., 2009). All these results support the model that pSMAC and dSMAC are radial lamella and lamellipodium, respectively. Thus, radial “symmetrization” of the motility apparatus presents an elegant way of maintaining the dynamic features of IS and yet achieve stability by cancelling out forces from all directions. It is important to recognize that “symmetrization” of the motility apparatus is the actual “stop signal” resulting in prolonged contact and signaling (Dustin, 2008b).

114

V. Mayya and M.L. Dustin

5.4 Immunological Kinapse 5.4.1 T Cell Activation Occurs Even in the Absence of a Stable IS Much of the studies on the properties, formation and function of stable IS were performed using primed T cells on SPBs or as conjugates with B cells/target cells. In contrast to primed T cells, naïve T cells take part in transient interactions with DCs and SPBs in many scenarios. Evidence has been mounting to support the view that TCR signaling continues to occur in naïve T cells without arresting (Dustin, 2004). Preliminary evidence came from a study wherein both T cells and mature DCs were allowed to interact in a collagen gel matrix (Gunzer et al., 2000). Antigenspecific T cells continually and rapidly migrated within the gel that eventually lead to their proliferation. Suspension culture system also provided evidence for short-lived interactions with antigen bearing DCs leading to T cell proliferation (Benvenuti et al., 2004). Conclusive evidence for TCR signaling in a motile state has come from deep-tissue, intravital imaging of lymph nodes (LNs) by two-photon microscopy (TPM) (Dustin, 2004). In TPM, fempto second-pulsed laser is used to excite the fluorophores with two infra-red photons simultaneously instead of a single photon of visible light. The power of the laser is adjusted such that enough energy for two-photon excitation is attained only at the focal point of a high numerical aperture objective. Also, the power of the laser is adjusted depending on the depth of the focal point in the tissue. The use of infra-red laser allows one to image greater tissue depth with minimal phototoxicity. The laser is raster scanned across the tissue to provide 2-D imaging and then successive 2D images at incrementally increasing depth allow 3D rendering. This process is then repeated over time to provide information about dynamics, which distinguishes TPM from classical histology. TPM imaging of explant LNs by Cahalan and colleagues revealed that naïve T cells exhibit apparently random migration with an average speed of ∼10–15 μm/min in the DC network of the T cell area (Miller et al., 2002). This motility behavior is central to their “stochastic repertoire scanning” strategy in the search of agonist MHCp. However, upon antigen exposure, the specific T cells stopped their random migratory behavior and either swarmed around or stopped completely, which were also fixed in location. Von Andrian and colleagues have further extended TPM to intact LNs in live animals (Mempel et al., 2004). They observed that antigen-specific interactions of T cells with DCs occur in three distinct phases: transient for the first 2–8 h, stable between 8–24 h, which then reverted back to transient later on (by 36 h). The authors also noted that CD69 surface expression occurred within the first 8 h, thus providing direct evidence for TCR signaling during the early, transient and serial interactions (Mempel et al., 2004). It is likely that antigen dependent down-regulation of CCR7 in the early phase is also due to initiation of TCR signaling from mobile T cell junctions with DCs, which is discussed in Section 5.5.3 (Sallusto et al., 1999). Naïve T cells receive homeostatic signals through the TCR recognition of self-MHC during the steady state migration and repertoire scanning (Fischer et al., 2007). We have noted that naive T cells continue to exhibit steady state motility in response to low potency MHCp that induce

5

Actin Cytoskeleton and the Dynamics of Immunological Synapse

115

tolerance by anergy (Skokos et al., 2007). All these support the view that T cells integrate TCR signals through serial contacts while on the move.

5.4.2 Immunological Kinapse is a Mobile and Asymmetric Counterpart to IS We have extended the use of SPBs to study the dynamics of IS formation in naïve T cells (Sims et al., 2007). When ICAM1, CD80, and MHCp are presented to naïve T cells, they undergo alternating phases of stable IS formation, autonomous migration and relocation on the bilayer every 20–30 min on average. Initiation of migration is preceded by the loss of symmetry of the pSMAC, which is visualized as the opening of the flourophore-tagged ICAM1 ring. The direction of movement is oriented towards the convex edge of the crescent-shaped focal zone of LFA1 due to the net vectorial sum of the forces (Fig. 5.2a). Consistent with our model, arrest of a naïve T cell and stabilization of the IS occurs with the closure of the ICAM-1 ring (Sims et al., 2007). Also, the speed of naïve T cell movement on the bilayer correlated well with the extent of polarity of the pSMAC at any instance of time, consistent with the function of lamella acting as the “drive” module of an amoeboid cell. Interestingly the dSMAC maintained its radial symmetry and continued to exhibit “contractile oscillations” irrespective of the status of pSMAC. These observations on SPBs provide a conceptual basis to explain the intravital observations on the dynamics of naïve T cell interactions with antigen-bearing DCs. These observations have also lead to the coining of the term “Immunological Kinapse” (IK), to signify the mobile and asymmetric aspects of the cell junction that distinguish it from the IS (Dustin, 2008b). We have further showed that PKCθ promotes symmetry breaking (Sims et al., 2007). The PKCθ null, naïve T cells form stable IS on SPBs and display reduced motility in the spleen in vivo by intravital microscopy upon encountering antigen. In contrast, WASp is needed for reestablishment of the symmetry of the IS in naïve T cells (Sims et al., 2007). Interestingly, a specific inhibitor of PKCθ rescues the reestablishment defect of WASp deficient T cells. These findings provide a molecular basis for the regulation of the balance between IS and IK and also clues on how the transition between the two forms of cell junctions takes place.

5.5 Role and Significance of Immunological Synapse and Kinapse 5.5.1 Directed Secretion The structure, organization, and dynamics of the IS serve different functions in the immune responses. We first discuss the role of stable IS in directed secretion. Directed secretion is important for the effector function of differentiated T cells. It is especially appreciated in the function of CD8 CTLs to efficiently and

116

V. Mayya and M.L. Dustin

exclusively kill the target cell by releasing lytic granules into a protected synaptic cleft (Dustin, 2008b). Cortical F-actin inhibits granule secretion in mast cells. Accordingly the cSMAC region, beneath which the cytolytic graules gets positioned in CTL conjugates, is depleted in F-actin (Stinchcombe et al., 2006). There is support for considering the cSMAC as the predominant region for TCR internalization (Lee et al., 2003; Varma et al., 2006). TCR recycling has been shown to be important for establishing polarity by aligning the secretory apparatus (Arkhipov and Maly, 2007). Poenie and colleagues have observed that pSMAC region provides anchoring scaffolds for cytoplasmic dynein and help the vectorial movement of MTOC towards the IS by microtubule sliding (Combs et al., 2006; Kuhn and Poenie, 2002). The adhesive zones in pSMAC along with the centripetal movement of receptor complexes form a barrier for diffusive leakage of the cytolytic content and thus protect bystander cells (Dustin, 2008b). The functional significance of a stable pSMAC ring has recently been demonstrated by the use of PKCθ inhibitor to increase the efficiency of lysis in CD4+ CTLs (Beal et al., 2008). Recent studies have demonstrated directed secretion of multiple cytokines (IL-2, IFN-gamma) in helper T cells, but also the surprising finding that the same T cells release other cytokines (IL-4, TNF) and chemokines in a non-directional manner (Huse et al., 2008).

5.5.2 Assymetric Cell Division Assymetric cell division is a frequently used mechanism by stem cells to maintain their “stemness” and also to give rise to daughter cells with differentiated fate. Assymetric cell division is often guided by a specific cellular niche. Stable interaction with a DC mediated by the IS can be thought of as a niche for the rare naïve T cells to allow for asymmetric division to eventually result in memory and effector T cell populations (Dustin and Chan, 2000). Recently, it was demonstrated that CD8 T cells undergo asymmetric cell division in the first round during the response to Listeria mnoncytogenes, an intracellular bacterial pathogen (Chang et al., 2007). The isolated T cell blasts were found to have polarized distribution of surface proteins involved in TCR signaling and T cell differentiation as well as of polarity proteins. This polarized distribution allowed the investigators to mark the proximal and distal ends of the T cell blast with respect to the IS. The investigators observed that the larger proximal and smaller distal daughter cells gave rise to effector and memory cells, respectively. Using mice lacking ICAM1, it was shown that stable interaction with DCs is necessary for the initial establishment of polarized distribution of the putative fate determinants. However, the mechanism by which the “memory” of the location of the IS on the T cell surface is maintained and perpetuated after the loss of contact was not addressed in the study.

5.5.3 Balance Between IS and IK in Tuning Immune Responses T cells in both lymphoid and non-lymphoid tissues are under the constant influence of chemokines, which drive and guide their motile behavior. Naïve T cells

5

Actin Cytoskeleton and the Dynamics of Immunological Synapse

117

experience a continuous competition between the “stop” signals by the MHCp and “go” signals by the chemokines that are on the surface of DCs and stromal cells. We showed, almost a decade back, through a transwell assay that TCR is able to overcome the “go” signals by certain chemokines but not others, especially the ligands for CCR7 (Bromley et al., 2000). Further, the dominance of CCR7 over TCR signaling was dose dependent: transwell migration of T cells occurred only at low antigen dose, but not at higher doses. This result proposes a model wherein the chemokinetic factors in the local microenvironment and the nature of antigens will have a profound influence on the duration of stable IS (Dustin, 2004). Indeed, recent observations from intravital imaging support this proposal. Rapid steady state migration of naïve T cells continues to occur in the presence of low-potency tolerogenic MHCp, as the dominant chemokinetic signals are not overcome (Skokos et al., 2007) Similarly naïve T cells continue to be motile during the early stages of interaction with DCs when presented with low dose agonistic TCR stimuli (Henrickson et al., 2008). Sustained signaling or integration through serial contacts, mediated by IK, allows the T cells to scan through many DCs presenting few agonistic MHCp. This is thought to allow the T cells to reset their signaling thresholds, overcome the chemokinetic signals likely by the downregulation of CCR7 (Sallusto et al., 1999), and eventually to symmetrize the synapse. Thus, the primed T cells can form stable IS with the DCs presenting the same level of MHCp that could not arrest them earlier. Several aspects of antigen presentation seem to be crucial parameters in controlling the duration of initial signaling via kinapses, as it varies from a few minutes to hours in different model systems (Dustin, 2008b). Results from intravital imaging and in vitro culture experiments point out that balance between the duration of signaling from IK and IS appears to be important for tolerance vs. immune response induction (Dustin, 2004, 2005). Components of the innate immunity may alter the balance favoring IS and thus promoting autoimmunity (Zehn and Bevan, 2006). Direct evidence for the importance of balance between IS and IK comes from the studies of naïve T cell dynamics on SPBs (Sims et al., 2007). Naïve T cells that made multiple contacts on SPBs were found to produce more IL2 than those that maintained a single contact, as per single cell cytokine capture assays. Further, PKCθ deficient cells, which are biased towards the Th1 lineage form hyperstable IS. In contrast, WASp deficient cell, which are biased towards the Th2 lineage form unstable IS (Sims et al., 2007). These findings suggest that the balance between IS and IK is also important for tuning the effector cell differentiation.

5.6 Issues and Considerations 5.6.1 TCR Triggering and Clustering Optical and electron microscopy observations as well as inferences from functional studies and other model systems support the concept of clustering of TCRs in the nanometer scale prior to triggering (Dustin, 2009). TCRs are organized in “confinement zones” or “protein islands” that are mostly less than 100 nm in size (Dustin,

118

V. Mayya and M.L. Dustin

2009; Lillemeier et al., 2006). Single TCR complexes diffuse rather freely within these domains (islands) and also occasionally move into neighboring domains (islands) by “hop diffusion” (Dustin, 2009). The above studies also present evidence for the central role of actin cytoskeleton in the formation and/or maintenance of the “protein islands” and in corralling groups of TCRs within the actin meshwork (Lillemeier et al., 2006; Andrews et al., 2008). Thus, actin cytoskeleton may play an important role in TCR triggering within the pre-clusters. Davis and colleagues have recently used PALM, dcFCCS, and TEM2 in conjunction to demonstrate that 5–15 individual protein islands of TCR concatenate to form larger clusters that may be equivalent to MCs observed by TIRFM (Lillemeier et al., 2009). It is known that formation of MCs and signaling by MCs is dependent on F-actin (Varma et al., 2006; Campi et al., 2005). F-actin scaffold that supports the protein islands of TCR may then bring them together to form MCs. Nascent F-actin nucleation driven by TCR triggering is likely to be very important as well. Though many biochemical events and molecules that link TCR triggering to actin nucleation have been characterized, the actual events in the nanometer scale that lead to concatenation of protein islands into MCs is not known (Billadeau et al., 2007; Dustin, 2009). Lipid domains or membrane rafts have also been implicated in TCR triggering and/or signaling (Dustin, 2009). F-actin scaffolds may also help in the movement of TCR and other proteins into or out of these compartments (Dustin, 2009). Modeling studies and examination of other signaling systems have revealed that high degree of spatial ordering in signaling complexes leads to increased gain or amplification in the signaling response (Mello et al., 2004). Thus F-actin may facilitate amplification of TCR signals from MCs. Also, there are unanswered questions regarding the cSMAC and F-actin. For example, what supports the formation of large coalesced TCR cluster at cSMAC that is depleted in F-actin? How are the MCs transported to the cSMAC once they become independent of the F-actin scaffold? What prevents TCR clusters from recruiting downstream signaling components and nucleating new F-actin at the cSMAC?

5.6.2 Diversity in IS Much of our understanding comes from the notion of IS as an organized and symmetrical structure promoted by studies on SPBs. However, studies with cell conjugates reveal many other forms of organization (Friedl et al., 2005). The secretory synapse has a region within the cSMAC where the vesicles and granules for exocytosis are segregated. IS with DCs show multiple foci of TCR and LFA1 clusters (Dustin et al., 2006). One hypothesis is that membrane domains and cytoskeletal dynamics on the DC may act as barriers and prevent the formation

2 The acronyms stand for Photo-Activation Localization Microscopy, dual color Fluorescence Cross Correlation Spectroscopy, and Transmission Electron Microscopy respectively.

5

Actin Cytoskeleton and the Dynamics of Immunological Synapse

119

of monocentric IS (as in the case of SPBs). To formally test this hypothesis, we have used electron beam lithography to “write” chrome lines on the glass surface and prevent free diffusion and transport of presented ligands on SPBs (Mossman et al., 2005). The scaling of clusters in the micropatterned surfaces was similar to those observed in the IS with DCs, thus supporting the original hypothesis. But live cell imaging by spinning disk confocal microscopy and sub-diffraction PALM are necessary to understand the dynamics of IS in cell conjugates. It is notable that TCR MCs have been observed in the IS with APC only recently using spinning disk confocal imaging (Oddos et al., 2008). Such live cell imaging approaches are also important to gain insights into synapses formed by immature T cells with thymic epithelial cells, B cells with DCs, and NK cells with the target cells.

5.6.3 Symmetry Breaking of IS Symmetry breaking of IS leads to the formation of a mobile junction (Dustin, 2008a,b, 2007). While we know that PKCθ destabilizes and that WASp stabilizes IS, the actual mechanism that underlies symmetry breaking of IS is currently not known. It may be worthwhile to consider lessons from studying symmetry breaking in other well established systems. Myosin II is a central player in symmetry breaking in these systems. It has been proposed that contraction by myosin II activity is responsible for rupturing the lamellar actin network as in other systems (Yam et al., 2007). A set of protein-protein interactions that link PKCθ, WIP and WASp to myosin II has been described in NK cells (Krzewski et al., 2006). Here in, PKCθ dependent phosphorylation of WIP results in the dissociation of WASp, which promotes binding of WIP to myosin II. But it has been clarified recently that PKCθ dependent phosphorylation of WIP does not result in dissociation of WASp and that WIP and WASp stay as a complex during TCR signaling. Thus how PKCθ and WASp impinge on myosin II activity is still debatable. Phenomenological and mathematical modeling initiatives have provided common principles that underlie symmetry breaking (Sohrmann and Peter, 2003; Li and Bowerman, 2009). Local positive feedback loop coupled to long-range inhibition that incorporate mechanochemical aspects of the actin cytoskeleton are important for symmetry breaking (Xu et al., 2003; Sohrmann and Peter, 2003; Millius et al., 2009). But a positive feedback alone establishes the asymmetry indefinitely as in the case fish keratocyte fragments (Yam et al., 2007; Sohrmann and Peter, 2003). Incorporation of a locally acting negative feedback along with the positive feedback allow for temporal modulation of polarization (Sohrmann and Peter, 2003; Millius et al., 2009). As inter-conversions between IS and IK happen continuously the machinery for symmetry breaking in the IS is expected to contain negative feedback components as well. A puzzling aspect of the IK on SPBs is the apparent radial symmetry of dSMAC in spite of the asymmetric nature of the pSMAC. However, this may not apply in a physiological context as the external chemokinetic signals are going to be as important as the internal dynamics in the T cell.

120

V. Mayya and M.L. Dustin

5.6.4 Balance Between IS and IK The concept of IK is new and many aspects require investigation. One is not sure of the experimental approaches to understand the nature of signal integration enabled by serial contacts of IK. Secondly, it may be worthwhile studying the differences between the cSMAC and the uropod to understand the differences in signaling from the IS and IK. Studying the dynamics of rho family GTPases at the IS and IK may also lead to insights. It is well established that sensitivity to chemokines is modulated depending on the activation and differentiation state of the T cell. This occurs mostly by up- or down- regulation of surface expression of the chemokine receptors. Although it is well established that chemokine “go” signals compete with TCR “stop” signals in real time, the mechanisms of “cross-talk” have not been studied (Dustin, 2008a,b; Bromley et al., 2000). It is imperative that the actin cytoskeleton forms the platform for the crosstalk between the two signaling systems, but nonetheless systematic studies are necessary if one is to better understand the factors responsible for the balance between the duration of IS and IK. Transition between IS and IK of naïve T cells on SPBs is autonomous, dictated by interplay of endogenous factors (Sims et al., 2007). To date only two molecules, PKCθ and WASp, that control the transition between IS and IK are known. It is likely that many more proteins modulate this process, as has been found by siRNA based screening for regulators of focal adhesion complex formation and cell migration (Simpson et al., 2008; Winograd-Katz et al., 2009). Having a full catalogue of proteins that modulate the balance between IS and IK are important to better understand the decisions of tolerance vs. immune response (Dustin, 2008b). Acknowledgments The preparation of this chapter was supported by the NIH Director’s Fund through a Nanomedicine development center PN2 EY016586.

References Abbas AK, AH Lichtman and JS Pober (2000) Cellular and Molecular Immunology, 4 edn (Philadelphia, PA: W.B. Saunders). Abraham RT and A Weiss (2004) Jurkat T cells and development of the T-cell receptor signalling paradigm. Nat Rev Immunol 4: 301–308. Andrews NL, KA Lidke, JR Pfeiffer et al (2008) Actin restricts FcepsilonRI diffusion and facilitates antigen-induced receptor immobilization. Nat Cell Biol 10: 955–963. Arkhipov SN and IV Maly (2007) A model for the interplay of receptor recycling and receptormediated contact in T cells. PLoS One 2: e633. Bachmann MF, K McKall-Faienza, R Schmits et al (1997) Distinct roles for LFA-1 and CD28 during activation of naive T cells: adhesion versus costimulation. Immunity 7: 549–557. Beal AM, N Anikeeva, R Varma et al (2008) Protein kinase C{theta} regulates stability of the peripheral adhesion ring junction and contributes to the sensitivity of target cell lysis by CTL. J Immunol 181: 4815–4824. Benvenuti F, S Hugues, M Walmsley et al (2004) Requirement of Rac1 and Rac2 expression by mature dendritic cells for T cell priming. Science 305: 1150–1153. Billadeau DD, JC Nolz and TS Gomez (2007) Regulation of T-cell activation by the cytoskeleton. Nat Rev Immunol 7: 131–143.

5

Actin Cytoskeleton and the Dynamics of Immunological Synapse

121

Bromley SK, DA Peterson, MD Gunn et al (2000) Cutting edge: hierarchy of chemokine receptor and TCR signals regulating T cell migration and proliferation. J Immunol 165: 15–19. Burbach BJ, RB Medeiros, KL Mueller et al (2007) T-cell receptor signaling to integrins. Immunol Rev 218: 65–81. Burkhardt JK, E Carrizosa and MH Shaffer (2008) The actin cytoskeleton in T cell activation. Annu Rev Immunol 26: 233–259. Campi G, R Varma and ML Dustin (2005) Actin and agonist MHC-peptide complex-dependent T cell receptor microclusters as scaffolds for signaling. J Exp Med 202: 1031–1036. Chang JT, VR Palanivel, I Kinjyo et al (2007) Asymmetric T lymphocyte division in the initiation of adaptive immune responses. Science 315: 1687–1691. Combs J, SJ Kim, S Tan et al (2006) Recruitment of dynein to the Jurkat immunological synapse. Proc Natl Acad Sci USA 103: 14883–14888. Cyster JG (2005) Chemokines, sphingosine-1-phosphate, and cell migration in secondary lymphoid organs. Annu Rev Immunol 23: 127–159. Das J, M Ho, J Zikherman et al (2009) Digital signaling and hysteresis characterize ras activation in lymphoid cells. Cell 136: 337–351. Delon J, K Kaibuchi and RN Germain (2001) Exclusion of CD43 from the immunological synapse is mediated by phosphorylation-regulated relocation of the cytoskeletal adaptor moesin. Immunity 15: 691–701. Dustin ML (2004) Stop and go traffic to tune T cell responses. Immunity 21: 305–314. Dustin ML (2005) A dynamic view of the immunological synapse. Semin Immunol 17: 400–410. Dustin ML (2007) Cell adhesion molecules and actin cytoskeleton at immune synapses and kinapses. Curr Opin Cell Biol 19: 529–533. Dustin ML (2008a) Hunter to gatherer and back: immunological synapses and kinapses as variations on the theme of amoeboid locomotion. Annu Rev Cell Dev Biol 24: 577–596. Dustin ML (2008b) T-cell activation through immunological synapses and kinapses. Immunol Rev 221: 77–89. Dustin ML (2009) The cellular context of T cell signaling. Immunity 30: 482–492. Dustin ML and AC Chan (2000) Signaling takes shape in the immune system. Cell 103: 283–294. Dustin ML and JA Cooper (2000) The immunological synapse and the actin cytoskeleton: molecular hardware for T cell signaling. Nat Immunol 1: 23–29. Dustin ML and TA Springer (1989) T cell receptor cross-linking transiently stimulates adhesiveness through LFA-1. Nature 341: 619–624. Dustin ML, MW Olszowy, AD Holdorf et al (1998) A novel adapter protein orchestrates receptor patterning and cytoskeletal polarity in T cell contacts. Cell 94: 667–677. Dustin ML, SK Bromley, Z Kan et al (1997) Antigen receptor engagement delivers a stop signal to migrating T lymphocytes. Proc Natl Acad Sci USA 94: 3909–3913. Dustin ML, SY Tseng, R Varma et al (2006) T cell-dendritic cell immunological synapses. Curr Opin Immunol 18: 512–516. Dustin ML, T Starr, R Varma et al (2007) Supported planar bilayers for study of the immunological synapse. Curr Protoc Immunol Chapter 18: Unit 18 13. Fischer UB, EL Jacovetty, RB Medeiros et al (2007) MHC class II deprivation impairs CD4 T cell motility and responsiveness to antigen-bearing dendritic cells in vivo. Proc Natl Acad Sci USA 104: 7181–7186. Freiberg BA, H Kupfer, W Maslanik et al (2002) Staging and resetting T cell activation in SMACs. Nat Immunol 3: 911–917. Friedl P and B Weigelin (2008) Interstitial leukocyte migration and immune function. Nat Immunol 9: 960–969. Friedl P, AT den Boer and M Gunzer (2005) Tuning immune responses: diversity and adaptation of the immunological synapse. Nat Rev Immunol 5: 532–545. Giannone G, BJ Dubin-Thaler, O Rossier et al (2007) Lamellipodial actin mechanically links Myosin activity with adhesion-site formation. Cell 128: 561–575. Grakoui A, SK Bromley, C Sumen et al (1999) The immunological synapse: a molecular machine controlling T cell activation. Science 285: 221–227.

122

V. Mayya and M.L. Dustin

Gunzer M, A Schafer, S Borgmann et al (2000) Antigen presentation in extracellular matrix: interactions of T cells with dendritic cells are dynamic, short lived, and sequential. Immunity 13: 323–332. Henrickson SE, TR Mempel, IB Mazo et al (2008) T cell sensing of antigen dose governs interactive behavior with dendritic cells and sets a threshold for T cell activation. Nat Immunol 9: 282–291. Huse M (2009) The T-cell-receptor signaling network. J Cell Sci 122: 1269–1273. Huse M, EJ Quann and MM Davis (2008) Shouts, whispers and the kiss of death: directional secretion in T cells. Nat Immunol 9: 1105–1111. Ilani T, G Vasiliver-Shamis, S Vardhana et al (2009) TCR signaling and immunological syanpse stability requires myosin IIA. Nat. Immunol 10(5): 531–539. Kaizuka Y, AD Douglass, R Varma et al (2007) Mechanisms for segregating T cell receptor and adhesion molecules during immunological synapse formation in Jurkat T cells. Proc Natl Acad Sci USA 104: 20296–20301. Kruse K, JF Joanny, F Julicher et al (2006) Contractility and retrograde flow in lamellipodium motion. Phys Biol 3: 130–137. Krzewski K, X Chen, JS Orange et al (2006) Formation of a WIP-, WASp-, actin-, and myosin IIA-containing multiprotein complex in activated NK cells and its alteration by KIR inhibitory signaling. J Cell Biol 173: 121–132. Kuhn JR and M Poenie (2002) Dynamic polarization of the microtubule cytoskeleton during CTLmediated killing. Immunity 16: 111–121. Lee KH, AD Holdorf, ML Dustin et al (2002) T cell receptor signaling precedes immunological synapse formation. Science 295: 1539–1542. Lee KH, AR Dinner, C Tu et al (2003) The immunological synapse balances T cell receptor signaling and degradation. Science 302: 1218–1222. Li R and B Bowerman (2009) Symmetry breaking in biology. CSH Persp Biol doi:10.1101/cshperspect.2(3):a003475 Lillemeier BF, JR Pfeiffer, Z Surviladze et al (2006) Plasma membrane-associated proteins are clustered into islands attached to the cytoskeleton. Proc Natl Acad Sci USA 103: 18992–18997. Lillemeier BF, MA Mortelmaier, MB Forstner et al (2009) TCR and LAT occur in separate domains, and concatenate during activation. Nat Immunol Accepted 11(1): 90–96. Mayya V, DH Lundgren, SI Hwang et al (2009) Quantitative phosphoproteomic analysis of T cell receptor signaling reveals system-wide modulation of protein-protein interactions. Sci Signal 2: ra46. Mello BA, L Shaw and Y Tu (2004) Effects of receptor interaction in bacterial chemotaxis. Biophys J 87: 1578–1595. Mempel TR, SE Henrickson and UH Von Andrian (2004) T-cell priming by dendritic cells in lymph nodes occurs in three distinct phases. Nature 427: 154–159. Menasche G, S Kliche, N Bezman et al (2007) Regulation of T-cell antigen receptor-mediated inside-out signaling by cytosolic adapter proteins and Rap1 effector molecules. Immunol Rev 218: 82–91. Miller MJ, SH Wei, I Parker et al (2002) Two-photon imaging of lymphocyte motility and antigen response in intact lymph node. Science 296: 1869–1873. Millius A, SN Dandekar, AR Houk et al (2009) Neutrophils establish rapid and robust WAVE complex polarity in an actin-dependent fashion. Curr Biol 19: 253–259. Monks CR, BA Freiberg, H Kupfer et al (1998) Three-dimensional segregation of supramolecular activation clusters in T cells. Nature 395: 82–86. Mossman KD, G Campi, JT Groves et al (2005) Altered TCR signaling from geometrically repatterned immunological synapses. Science 310: 1191–1193. Nolz JC, LP Nacusi, CM Segovis et al (2008) The WAVE2 complex regulates T cell receptor signaling to integrins via Abl- and CrkL-C3G-mediated activation of Rap1. J Cell Biol 182: 1231–1244.

5

Actin Cytoskeleton and the Dynamics of Immunological Synapse

123

Nolz JC, RB Medeiros, JS Mitchell et al (2007) WAVE2 regulates high-affinity integrin binding by recruiting vinculin and talin to the immunological synapse. Mol Cell Biol 27: 5986–6000. Oddos S, C Dunsby, MA Purbhoo et al (2008) High-speed high-resolution imaging of intercellular immune synapses using optical tweezers. Biophys J 95(10): L66–68. Penninger JM and GR Crabtree (1999) The actin cytoskeleton and lymphocyte activation. Cell 96: 9–12. Peterson EJ, ML Woods, SA Dmowski et al (2001) Coupling of the TCR to integrin activation by Slap-130/Fyb. Science 293: 2263–2265. Pham TH, T Okada, M Matloubian et al (2008) S1P1 receptor signaling overrides retention mediated by G alpha i-coupled receptors to promote T cell egress. Immunity 28: 122–133. Prasad A, J Zikherman, J Das et al (2009) Origin of the sharp boundary that discriminates positive and negative selection of thymocytes. Proc Natl Acad Sci USA 106: 528–533. Sallusto F, E Kremmer, B Palermo et al (1999) Switch in chemokine receptor expression upon TCR stimulation reveals novel homing potential for recently activated T cells. Eur J Immunol 29: 2037–2045. Shaw AS and ML Dustin (1997) Making the T cell receptor go the distance: a topological view of T cell activation. Immunity 6: 361–369. Simpson KJ, LM Selfors, J Bui et al (2008) Identification of genes that regulate epithelial cell migration using an siRNA screening approach. Nat Cell Biol 10: 1027–1038. Sims TN, TJ Soos, HS Xenias et al (2007) Opposing effects of PKCtheta and WASp on symmetry breaking and relocation of the immunological synapse. Cell 129: 773–785. Skokos D, G Shakhar, R Varma et al (2007) Peptide-MHC potency governs dynamic interactions between T cells and dendritic cells in lymph nodes. Nat Immunol 8: 835–844. Sohrmann M and M Peter (2003) Polarizing without a c(l)ue. Trends Cell Biol 13: 526–533. Springer TA (1990) Adhesion receptors of the immune system. Nature 346: 425–434. Stinchcombe JC, E Majorovits, G Bossi et al (2006) Centrosome polarization delivers secretory granules to the immunological synapse. Nature 443: 462–465. Tadokoro S, SJ Shattil, K Eto et al (2003) Talin binding to integrin {beta} tails: a final common step in integrin activation. Science 302: 103–106. Tseng SY, JC Waite, M Liu et al (2008) T cell-dendritic cell immunological synapses contain TCRdependent CD28-CD80 clusters that recruit protein kinase Ctheta. J Immunol 181: 4852–4863. Tskvitaria-Fuller I, AL Rozelle, HL Yin et al (2003) Regulation of sustained actin dynamics by the TCR and costimulation as a mechanism of receptor localization. J Immunol 171: 2287–2295. Valitutti S, M Dessing, K Aktories et al (1995) Sustained signaling leading to T cell activation results from prolonged T cell receptor occupancy. Role of T cell actin cytoskeleton. J Exp Med 181: 577–584. Varma R, G Campi, T Yokosuka et al (2006) T cell receptor-proximal signals are sustained in peripheral microclusters and terminated in the central supramolecular activation cluster. Immunity 25: 117–127. Wei X, BJ Tromberg and MD Cahalan (1999) Mapping the sensitivity of T cells with an optical trap: polarity and minimal number of receptors for Ca(2+) signaling. Proc Natl Acad Sci USA 96: 8471–8476. Weiss A and DR Littman (1994) Signal transduction by lymphocyte antigen receptors. Cell 76: 263–274. Winograd-Katz SE, S Itzkovitz, Z Kam et al (2009) Multiparametric analysis of focal adhesion formation by RNAi-mediated gene knockdown. J Cell Biol 186: 423–436. Xu J, F Wang, A Van Keymeulen et al (2003) Divergent signals and cytoskeletal assemblies regulate self-organizing polarity in neutrophils. Cell 114: 201–214. Yam PT, CA Wilson, L Ji et al (2007) Actin-myosin network reorganization breaks symmetry at the cell rear to spontaneously initiate polarized cell motility. J Cell Biol 178: 1207–1221. Yokosuka T, K Sakata-Sogawa, W Kobayashi et al (2005) Newly generated T cell receptor microclusters initiate and sustain T cell activation by recruitment of Zap70 and SLP-76. Nat Immunol 6: 1253–1262.

124

V. Mayya and M.L. Dustin

Yokosuka T, W Kobayashi, K Sakata-Sogawa et al (2008) Spatiotemporal regulation of T cell costimulation by TCR-CD28 microclusters and protein kinase C theta translocation. Immunity 29: 589–601. Zehn D and MJ Bevan (2006) T cells with low avidity for a tissue-restricted antigen routinely evade central and peripheral tolerance and cause autoimmunity. Immunity 25: 261–270.

Chapter 6

Actin-based Motile Processes in Tumor Cell Invasion Matthew Oser, Robert Eddy, and John Condeelis

Abstract Tumor cell invasion and metastasis is a multi-step process where tumor cells escape from the primary tumor and seed at distant sites in the body. The invasion of tumor cells through the surrounding extracellular matrix (ECM) and intravasation through the endothelium to enter the bloodstream requires several highly regulated actin-based motile processes including: 1. Chemotaxis 2. Formation of invasive protrusions 3. Active cell migration, and 4. Active focal degradation of the ECM. The invasive tumor cell generates highly specialized actinbased invasive protrusions known as lamellipodia to initiate active cell migration and invadopodia to actively degrade through dense areas of ECM. The formation of both lamellipodia and invadopodia requires precise control of the activities of several proteins that function to regulate the actin cytoskeleton including: cortactin, cofilin, the Arp2/3 complex, diaphanous-related formins (DRFs), and myosin. In this chapter, we will discuss the molecular mechanisms that are known to control the actin cytoskeleton during the formation and maturation of both lamellipodia and invadopodia, and the contribution of these actin-based motile processes to the invasive tumor cell phenotype during metastasis in vivo. Abbreviations ECM MMP DRF EGF

extracellular matrix matrix metalloproteinase diaphanous-related formins epidermal growth factor

M. Oser (B) Department of Anatomy and Structural Biology, Albert Einstein College of Medicine of Yeshiva University, Bronx, NY 10461, USA e-mail: [email protected]

M.-F. Carlier (ed.), Actin-based Motility, DOI 10.1007/978-90-481-9301-1_6, C Springer Science+Business Media B.V. 2010

125

126

M. Oser et al.

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chemotaxis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Lamellipodia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Invadopodia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.4.1 Invadopodium Maturation Requires Actin Polymerization . . . . . . . . . 6.4.2 Regulation of Actin Polymerization in Invadopodia . . . . . . . . . . . . 6.4.3 The Stages of Invadopodium Maturation . . . . . . . . . . . . . . . . . 6.4.4 Function and Regulation of Cofilin During Invadopodium Maturation . . . . 6.4.5 Pathways Leading to Activation of the Arp2/3 Complex During Invadopodium Maturation . . . . . . . . . . . . . . . . . . . . . . . 6.5 Formin-Dependent Actin Nucleation During Tumor Cell Invasion . . . . . . . . 6.6 Nonmuscle Myosin II and Contractile Force Regulation During Tumor Cell Invasion . . . . . . . . . . . . . . . . . . . . 6.7 Coordination of Pathways Linking Actin Polymerization to the Matrix Degradation Activity of Invadopodia . . . . . . . . . . . . . . . . 6.8 The Relative Contribution of Chemotaxis, Invadopodium Formation, and Lamellipodium Formation to the Invasive Tumor Cell Phenotype In Vivo . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

6.1 6.2 6.3 6.4

126 129 132 134 136 140 145 147 150 152 153 155 156 158

6.1 Introduction Tumor metastasis is a multi-step process where tumor cells escape from the primary tumor and seed at distant sites in the body. Metastasis requires a tumor cell to invade through the surrounding extracellular matrix (ECM), intravasate through the endothelium, enter the bloodstream, and extravasate and colonize at a distant site in the body (Condeelis and Segall 2003). The high positive correlation between the quantity of tumor cells that intravasate through the endothelium and the number of lung metastases during mammary carcinoma cell metastasis demonstrates that entry into the bloodstream is a critical step during metastasis (Wyckoff et al. 2000a). Coordinated cell migration is essential during the early stages of metastasis, specifically during invasion and intravasation. Invasive tumor cells can either migrate autonomously as single cells or migrate collectively as a cell stream of unattached tumor cells through the extracellular matrix at relatively high velocities (Condeelis and Segall 2003) or, again, migrate collectively as a group of attached cells at significantly slower velocities (Friedl and Gilmour 2009). Although collective cell migration involving attached cells is potentially important for invasion at the tumor margin and local recurrence, the rapid migration of invasive tumor cells, both as individuals and as cell streams, appears to be more involved in systemic dissemination instead of local recurrence (Wyckoff et al. 2007; Robinson et al. 2009). Furthermore, individual invasive tumor cells are more difficult to detect during clinical imaging and pathological examination, and are often missed during the initial diagnosis (Pantel et al. 2008). Therefore, we will focus this chapter on

6

Actin-based Motile Processes in Tumor Cell Invasion

127

the actin-based motile processes involved in the invasion of individual tumor cells during progression to distant metastasis. Intravital imaging of metastatic breast carcinoma cells in living animals have shown that invasive tumor cells compared to non-invasive tumor cells migrate as single cells at high velocities along ECM collagen fibers toward blood vessels (Condeelis and Segall 2003), have increased ability to orient their migration toward blood vessels, and are more adept at crossing the endothelial cell barrier to enter the circulation (Wyckoff et al. 2000a, 2007). To accomplish these steps, the invasive tumor cell requires precise coordinated rearrangement of the actin cytoskeleton (Olson and Sahai 2009). At least 4 actin-based motile processes are required and highly regulated during tumor cell invasion and intravasation including: 1. Chemotaxis (Kay et al. 2008), 2. Formation of invasive protrusions (Yamaguchi and Condeelis 2007), 3. Active cell migration (Machesky 2008), and 4. Active focal degradation of the ECM (Weaver 2008). The invasive tumor cell generates highly specialized actin-based invasive protrusions known as lamellipodia to initiate active cell migration (Condeelis et al. 2001; Machesky 2008), and invadopodia to actively degrade through dense areas of ECM (Weaver 2008; Poincloux et al. 2009) (Fig. 6.1a, b). Precise temporal and spatial regulation of these actin-based motile processes is critical for successful tumor cell invasion and intravasation (Mouneimne et al. 2006; Wang 2006). Gene expression profile microarray studies from invasive tumor cells compared to the average cells of the primary tumor in both xenograft and transgenic models of mammary adenocarcinomas (Wang et al. 2004, 2007b) have revealed that invasive cancer cells use the same pathways involved in chemotaxis (Mouneimne et al. 2006), lamellipodium formation (Machesky 2008), and invadopodium formation (Yamaguchi et al. 2005) for successful tumor cell invasion and intravasation in vivo (Wang et al. 2007a). These microarray studies specifically identified that signaling pathways involved in activation of the actin severing and nucleating protein, cofilin, and the actin nucleation factor, the Arp2/3 complex, as critical regulators of the invasive tumor cell phenotype (Wang et al. 2004, 2007b) (Fig. 6.2) suggesting that the regulation of cofilin and the Arp2/3 complex during chemotaxis, and lamellipodium and invadopodium formation is essential for tumor cell invasion in vivo. Direct evidence for the importance of the cofilin pathway comes from a xenograft study using orthotopic injection of mammary carcinoma cells showing that increasing the output of cofilin activity by expressing a kinase-dead LIM Kinase leads to increased metastasis, and decreasing the output of cofilin activity by expressing a full-length LIM Kinase leads to decreased metastasis (Wang et al. 2006). Thus, understanding mechanisms that regulate cofilin in lamellipodia and invadopodia are important for understanding how actin-based processes are controlled during tumor cell invasion. The formation of lamellipodia and invadopodia have been studied extensively in invasive tumor cell lines in many different cancer cell types in culture (Machesky 2008; Buccione et al. 2009; Caldieri et al. 2009; Linder 2009). This has allowed detailed dissection of the signaling pathways and protein binding interactions that control proper maturation of both lamellipodia and invadopodia. We will focus this chapter on the molecular mechanisms that regulate chemotaxis and the maturation of lamellipodia and invadopodia in invasive tumor cells.

128

M. Oser et al.

Fig. 6.1 Invadopodia and lamellipodia are distinct actin-based subcellular structures formed by invasive tumor cells. (a) Cartoon of an invasive tumor cell plated on ECM showing actin-based subcellular compartments used during tumor cell invasion. Lamellipodia form at the leading edge of the tumor cell and produce actin-based protrusions in the XY direction. Invadopodia form on the ventral cell surface and produce actin-based protrusions in the Z direction and also function to degrade the ECM using MMP activity. Red represents the actin filament network in both subcellular compartments. (b) Image of an MTLn3 mammary carcinoma cell stimulated with EGF stained for actin free barbed ends showing the presence of barbed ends in both lamellipodia (top right inset outlined in blue) and in invadopodia (bottom left inset outlined in yellow). Bar = 10 μm. (c) Image of an MDA-MB-231 mammary carcinoma cell plated on Alexa488 gelatin thin matrix expressing TagRFP-cortactin fixed and stained for Tks5 showing invadopodium precursors (bottom left inset outlined in white) and mature matrix-degrading invadopodia (top right inset outlined in yellow). Bar = 10 μm

6

Actin-based Motile Processes in Tumor Cell Invasion

129

Fig. 6.2 The signaling pathways upregulated in invasive tumor cells leading to the activation of cofilin and the Arp2/3 complex. Invasive tumor cells were collected using an in vivo invasion assay (Wyckoff et al. 2000b) from rat mammary tumors and the gene expression profiles were analyzed by microarray analysis (Wang et al. 2004). Diagram shows the motility pathways that are upregulated in the invasive tumor cell subpopulations compared to the cells of the primary tumor. Numbers show the fold increase in gene expression compared to the average tumor cells of the primary tumor. The pathways leading to the activation of cofilin severing activity and activation of the Arp2/3 complex are highly upregulated leading to efficient actin polymerization. By upregulating cofilin and Arp2/3 signaling pathways, invasive tumor cells have increased capacity to polymerize actin for invasion

6.2 Chemotaxis One of the early steps of tumor invasion involving the actin cytoskeleton is the directed cell migration, ie. chemotaxis, of tumor cells toward an extracellular chemical stimulus. Precise mechanisms have evolved to control chemotaxis throughout the evolution of eukaryotic cells. The model organisms traditionally used to study the mechanisms of eukaryotic cell chemotaxis include Dictyostelium discoideum and neutrophils, and more recently invasive tumor cells (Soon 2007). Although Dictyostelium discoideum, neutrophils, and tumor cells have evolved unique mechanisms to regulate chemotaxis, some of the basic principles that allow a cell to sense and move toward an extracellular gradient have been conserved (Condeelis et al. 1992; Kay et al. 2008). An extracellular signal is “sensed” and amplified at the front of the cell, while dampened, or inhibited, at the back of the cell. This allows the chemotactic cell to sense and move toward the chemoattractant. This polarity of response is accomplished by the asymmetry of actin polymerization although other mechanisms, such as myosin-based contraction, are also important (Kay et al. 2008).

130

M. Oser et al.

Invasive tumor cells, but not the non-invasive cells from the primary tumor, have acquired the unique ability to orient their migration and chemotax towards blood vessels for intravasation (Wyckoff et al. 2000a; Sidani et al. 2007) suggesting that mechanisms that control chemotaxis are important for the invasive tumor cell phenotype. Growth factors in the tumor microenvironment, such as EGF (Wyckoff et al. 2004) or SDF-1 (Hernandez et al. 2009), that are secreted by macrophages (Wyckoff et al. 2004) or other stromal cells near blood vessels initiate tumor cell chemotaxis in vivo leading to tumor cell invasion and intravasation (Wyckoff et al. 2004). Interestingly, invasive tumor cells coordinately upregulate signaling pathways, downstream of the EGF receptor, that control chemotaxis and actin polymerization (Soon 2007) allowing for enhanced ability for a tumor cell to sense and migrate toward blood vessels. When invasive tumor cells are stimulated with EGF, several signaling pathways become activated leading to tumor cell chemotaxis (Bailly et al. 1998). One of the pathways activated downstream of EGF leads to the activation of the actin severing protein cofilin (Chan et al. 2000). A study using local activation of a caged (inactive when caged) cofilin in metastatic tumor cells demonstrated that local activation of cofilin alone is sufficient to initiate protrusion and determine the direction of tumor cell migration (Ghosh et al. 2004). Thus, the mechanisms that regulate the initial activation of cofilin are important for controlling tumor cell chemotaxis. Cofilin has two general functions during cell motility: 1. to depolymerize actin filaments from the pointed end (– end) to supply a pool of actin monomers for steady state actin polymerization (Carlier et al. 1997), 2. to sever actin filaments to create free barbed ends (+ ends) used for actin polymerization (Ichetovkin et al. 2000). In metastatic mammary carcinoma cells, the initial activation of cofilin leads to actin filament severing, which increases free barbed ends for actin polymerization (Chan et al. 2000; Mouneimne et al. 2004). This initiates a protrusion toward an extracellular chemoattractant signal, such as EGF (Ghosh et al. 2004; Mouneimne et al. 2006). When cofilin is depleted from mammary carcinoma cells, either by knockdown via siRNA (Mouneimne et al. 2004) or by microinjection of functional blocking antibodies (DesMarais et al. 2004), tumor cells are unable to sense an EGF gradient to form an actin-based protrusion toward the EGF gradient demonstrating that cofilin is critical for initiating chemotaxis in invasive tumor cells. The primary mechanisms responsible for the initial activation of cofilin during chemotaxis include: 1. Dephosphorylation (Eiseler et al. 2009) and 2. Release from phosphatidylinositol 4,5-bisphosphate (PI(4, 5)P2 ) binding (van Rheenen et al. 2007). In addition, it was recently demonstrated that the cortactin regulates cofilin activity in invadopodia (Oser et al. 2009) and that cortactin is also critical for chemotaxis toward EGF in the same mammary carcinoma cell type (Desmarais et al. 2009). Thus, it is possible that cortactin also may regulate the initial activation of cofilin leading to chemotaxis, but this hypothesis has yet to be tested experimentally. Cofilin activity is blocked upon phosphorylation of serine 3 and restored when cofilin is dephosphorylated (Bamburg and Wiggan 2002). Phosphorylation of cofilin at serine 3 inhibits cofilin’s ability to bind to actin and thereby blocks cofilin’s actin

6

Actin-based Motile Processes in Tumor Cell Invasion

131

severing and depolymerization activity (Arber et al. 1998). However, dephosphorylation is not the primary mechanism responsible for cofilin activation during tumor cell chemotaxis in mammary carcinoma cells since the phosphorylation of cofilin and cofilin activity both increase in response to EGF stimulation and are uncoupled (Song et al. 2006). An alternative mechanism to dephosphorylation for initiating cofilin activity involves binding and release from PI(4, 5)P2 . When cofilin is bound to PI(4, 5)P2 , its actin binding activity and hence severing activity, is inhibited (Gorbatyuk et al. 2006). The hydrolysis of PI(4, 5)P2 by PLCγ1 decreases the amount of PI(4, 5)P2 at the plasma membrane and leads to the initial activation of cofilin (van Rheenen et al. 2007). Thus, upon PLCγ1-mediated PI(4, 5)P2 hydrolysis, cofilin is released in its active, dephosphorylated form, from the plasma membrane and can initiate chemotaxis. Regulation of cofilin activity by PI(4, 5)P2 appears to be the primary mechanism for cofilin activation at the leading edge in mammary carcinoma cells (Song et al. 2006; van Rheenen et al. 2007). There are now several studies that support the model that tumor cell chemotaxis is initiated by a PLCγ1- PI(4, 5)P2 -cofilin pathway. Depletion of cofilin or PLCγ1 with siRNA, or inhibition of PLCγ1 activity using a chemical inhibitor, each resulted in impaired tumor cell chemotaxis and failure of tumor cells to generate free barbed ends in response to growth factor stimulation (Mouneimne et al. 2004; Meira et al. 2009). Furthermore, cell turning and the formation of new protrusions in response to external stimuli are inhibited in cells expressing a cofilin mutant, D122K, that has a 10-fold increased affinity for PI(4, 5)P2 (Leyman et al. 2009) demonstrating that release of cofilin from PI(4, 5)P2 is required for tumor cell chemotaxis. The precise mechanism by which cofilin is regulated by PI(4, 5)P2 -binding at the plasma membrane was elucidated using FRET and FLIP approaches with live mammary carcinoma cells (van Rheenen et al. 2007). In resting cells, cofilin is directly bound to PI(4, 5)P2 at the plasma membrane. Upon stimulation with EGF, cofilin is released from the membrane via PLCγ1-mediated PI(4, 5)P2 hydrolysis and binds to actin. Cofilin then severs actin filaments locally to create free barbed ends for actin polymerization. Scaffolding activators can enhance the activation of the PLCγ1/ PI(4, 5)P2 / cofilin pathway during tumor cell invasion. Recently, Memo, a scaffolding protein that interacts with specific phosphotyrosines on the ErbB2 receptor and is important for cell migration (Marone et al. 2004), was shown to enhance both the depolymerization and severing activity of cofilin in vitro (Meira et al. 2009). Memo directly binds to cofilin and interacts with a cofilin/PLCγ1/ErbB2 complex in vivo thereby increasing the output of the PLCγ1/cofilin pathway (Meira et al. 2009). Thus, the local activation of cofilin at the leading edge by PLCγ1 is amplified in the presence of Memo. Future studies will reveal whether additional unidentified proteins may also enhance or inhibit the initial activation of cofilin by PLCγ1 in invasive tumor cells. Cofilin activity is also regulated by pH (Van Troys et al. 2008). At the plasma membrane, the binding of cofilin to PI(4, 5)P2 is weakened by increases in pH that occur after activation of the Na+/H+ exchanger (Frantz et al. 2008). Upon

132

M. Oser et al.

growth factor stimulation, the Na+/H+ exchanger is activated leading to local increases in pH resulting in decreased cofilin-PI(4, 5)P2 binding at the membrane and thus increased cofilin-actin binding. Interestingly, high expression of the Na+/H+ exchanger is positively correlated with tumor cell invasion and is thought to enhance tumor metastasis (For a review see (Cardone et al. 2005)). Future studies will determine whether the Na+/H+ exchanger-dependent cofilin activation can enhance tumor cell metastasis. For efficient tumor cell chemotaxis, the chemoattractant signal must be inhibited at the back of the cell (Soon 2007). As mentioned above, cofilin activity is blocked when it is phosphorylated on serine 3 by specific kinases that include Limfamily Kinases (LIM 1 and 2) (Arber et al. 1998) and TES-family kinases (Scott and Olson 2007). Interestingly, in mammary carcinoma cells, total cofilin phosphorylation levels increase in response to EGF stimulation (Song et al. 2006) suggesting that phosphorylation and inactivation of cofilin at the back of the cell may allow for amplification of cofilin activation at the front of the cell leading to directional sensing and tumor cell chemotaxis. In fact, global increases in cofilin phosphorylation throughout the cell are necessary to generate a gradient of high cofilin activity at the front of the cell facing the chemotactic source (Mouneimne et al. 2006). Expression of a kinase-dead dominant negative LIM Kinase mutant that blocks increases in global cofilin phosphorylation in response to EGF, inhibited tumor cell chemotaxis both in vitro and in vivo (Mouneimne et al. 2006; Wang et al. 2006). The gradient of cofilin activity at the front of the cell controlled by the PLCγ1/ PI(4, 5)P2 pathway is sharpened by global inhibition of cofilin activity mediated by LIM Kinase activity. Thus, the activation of both PLCγ1 and LIM Kinase pathways are necessary for efficient tumor cell chemotaxis. This coordinated upregulation of cofilin and LIM Kinase activities has been observed in invasive mammary carcinoma cells isolated from xenograft tumors in rats and PyMT tumors from mice (Wang et al. 2004; Wang et al. 2007b) suggesting that the precise spatial control of cofilin activation is critical for efficient tumor cell invasion in vivo (reviewed in (Wang et al. 2007a)). Together, these findings support a LEGI (local excitation/global inhibition) model for cofilin activation during tumor cell chemotaxis. LEGI models have been used to explain chemotactic phenomenon in other eukaryotic organisms, such as Dictyostelium discoideum. In mammary tumor cells, local excitation of cofilin at the plasma membrane is controlled by PLCγ1-mediated PI(4, 5)P2 hydrolysis and the global inhibition of cofilin is controlled by phosphorylation via LIM Kinase. Together, these mechanisms sharpen the chemotactic gradient so the cell can quickly respond to a chemoattractant allowing for efficient tumor cell chemotaxis.

6.3 Lamellipodia As described in the previous section, pathways that activate cofilin at the front of the cell and inhibit cofilin activity at the back of the cell are critical for the efficient chemotaxis of mammary carcinoma cells. After the initial direction for cell

6

Actin-based Motile Processes in Tumor Cell Invasion

133

migration is determined using the chemotactic mechanisms described above, an invasive tumor cell must initiate actin-based cell motility to actively migrate through the surrounding tissue stroma. The process of cell migration can be simplified as a motility cycle beginning with the initiation of actin polymerization, which results in an actin-based protrusion of the plasma membrane, and is required for cell migration (Pollard and Borisy 2003; DesMarais et al. 2005). For invasive tumor cells, the overall result of the motility cycle is net translocation of an invasive tumor cell toward the blood or lymphatic vessels (Condeelis and Segall 2003). It is thought that a tumor cell forms specialized actin-based structures for protrusion known as lamellipodia in order to migrate toward blood vessels. Lamellipodia are formed at the leading edge of the cell and contain a dense branched actin network directly beneath the plasma membrane (Machesky 2008). Many motile cell-types form lamellipodia to generate actin-based protrusions and therefore the lamellipodium is not a unique feature of invasive tumor cells. However, it has been demonstrated that invasive tumor cells upregulate the expression of proteins, such as the Arp2/3 complex, that are involved in signaling pathways that produce lamellipodia, thus allowing invasive tumor cells to have enhanced migration ability (Wang et al. 2005). Lamellipodium formation requires actin polymerization at the leading edge of the tumor cell (Pollard and Borisy 2003). To accomplish this, cells must increase the number of free barbed ends (+ ends) for filament growth. In invasive tumor cells, as well as other eukaryotic cells such as Dictyostelium discoideum and fibroblasts, the amplification of free barbed ends occurs in two temporal transients: an early and a late transient. It is now clear that the early transient in invasive tumor cells involves activation of cofilin through the PLCγ1 pathway discussed above (Mouneimne et al. 2004). The net result is increased cofilin severing of actin filaments leading to the generation of free barbed ends. Interestingly, there is little productive protrusion generated by the first transient. The cofilin-generated barbed ends during the first transient determine the direction of tumor cell movement (Sidani et al. 2007) and supply barbed ends for the second transient to generate a productive protrusion (DesMarais et al. 2004; Mouneimne et al. 2004). The second barbed end transient requires the Arp2/3 complex and results in a productive protrusion (Mouneimne et al. 2004). When active, the Arp2/3 complex binds to the side of recently polymerized mother actin filaments and initiates a branched daughter actin filament resulting in the formation of a branched actin network. The Arp2/3 complex is a seven-subunit complex consisting of two actin-related proteins, Arp2 and Arp3, and five other subunits that function to maintain the Arp2/3 complex inactive (Pollard 2007). The Arp2 and Arp3 subunits are similar in structure to actin monomers and as a result mimic two actin monomers when the Arp2/3 complex is activated. Thus, when the Arp2/3 complex is bound to the side of a preexisting actin filament, addition of only one actin monomer, typically supplied by the upstream activators from the WASp and WAVE family, is required to initiate nucleation resulting in elongation of the daughter filament (Pollard 2007). The branched actin network, generated by the Arp2/3 complex, produces force to push the plasma membrane forward resulting in a protrusion-the functional output of the lamellipodium.

134

M. Oser et al.

Similar to cofilin, the Arp2/3 complex can be regulated by multiple signaling pathways. The most potent activators of the Arp2/3 complex are WASp or WAVE family proteins (Kurisu and Takenawa 2009), although other weak activators, such as cortactin, have been identified (Weaver et al. 2001). Interestingly, lamellipodium formation in tumor cells requires the Arp2/3 complex activator WAVE2, but not N-WASp (Sarmiento et al. 2008) suggesting that WAVE2 is the primary activator of the Arp2/3 complex at lamellipodia. WAVE2 is overexpressed in invasive melanoma cells and depletion of WAVE2 in these cells inhibits cell motility, invasion, and experimental metastasis in mice (Kurisu et al. 2005). Thus, WAVE2mediated Arp2/3 activation leading to lamellipodium formation may be critical for tumor cell invasion. Cofilin and the Arp2/3 complex have been shown to function synergistically both in vitro (Ichetovkin et al. 2002) and in vivo (DesMarais et al. 2004) and this synergy is required to produce a lamellipodial protrusion that forms at the front of the cell facing the chemotactic gradient during tumor cell migration (Sidani et al. 2007). When cofilin severs actin filaments to create free barbed ends, these newly formed barbed ends elongate to produce the newly polymerized actin filaments that are preferred sites for Arp2/3 binding (Ichetovkin et al. 2002). This is the basis for the synergy between the activities of cofilin and Arp2/3 dendritic nucleation leading to the efficient formation of dense branched actin networks located at sites of cofilin activation. Tumor cells with suppressed cofilin protein expression have mislocalized Arp2/3 complex at one end of the cell (Sidani et al. 2007) suggesting that cofilin severing activity is important for proper localization of the Arp2/3 complex and providing more evidence that cofilin and the Arp2/3 complex cooperate during tumor cell migration. As described earlier, invasive tumor cells upregulate pathways leading to cofilin and Arp2/3 activities allowing for enhanced ability for chemotaxis and lamellipodium formation (Wang et al. 2004). Together, coordinating chemotaxis and lamellipodium formation allows invasive tumor cells to sense a chemoattractant gradient near blood vessels and migrate toward that gradient. Both of these processes are important for tumor cell invasion.

6.4 Invadopodia In addition to coordinating chemotaxis and lamellipodium formation, the invasive tumor cell must migrate through basement membrane barriers compromised of thick crossed-linked ECM to successfully invade through the surrounding tissue stroma and intravasate through the endothelium into the bloodstream (Sabeh et al. 2009). Migrating tumor cells specifically require proteolysis to invade through a crosslinked collagen matrix (Packard et al. 2009). The basement membrane surrounding the epithelial tumor and the endothelium lining the blood vessels consists of dense, cross-linked extracellular matrix comprised of collagen IV, laminin, and heparan sulfate proteoglycans and presents a barrier for tumor cell invasion (Condeelis and Segall 2003; Poincloux et al. 2009). To cross these barriers, metastatic carcinoma

6

Actin-based Motile Processes in Tumor Cell Invasion

135

cells form specialized subcellular structures, known as invadopodia, that use matrix metalloproteinase (MMP) activity to degrade through dense areas of cross-linked ECM. Studies using highly invasive tumor cells expressing the Mena-invasive isoform that have dramatically increased intravasation ability also exhibit enhanced stability of invadopodia with increased matrix degradation activity (Philippar et al. 2008) supporting the hypothesis that matrix degradation activity by invadopodia is critical for tumor cell invasion and intravasation in vivo. In the next section, we will discuss the molecular mechanisms that regulate invadopodium formation and maturation. Invadopodia are defined as subcellullar structures uniquely formed by metastatic carcinoma cells that function to focally degrade ECM using MMP activity and have been observed in various metastatic carcinoma cell lines (for a review see (Buccione et al. 2009)). In addition, invadopodia also form in primary head and neck squamous carcinoma cells from aggressive human tumors (Clark et al. 2007). The most commonly used invasive tumor cell lines to model the mechanisms involved in invadopodium maturation include mammary carcinoma cells (Yamaguchi et al. 2005; Artym et al. 2006; Chan et al. 2009; Lizarraga et al. 2009), head and neck squamous carcinoma cells (Clark et al. 2007), and melanoma cells (Ayala et al. 2008). Interestingly, invadopodia form in a metastatic mammary carcinoma cell line (MTLn3 cells), but fail to form in a non metastatic cell line derived from the primary tumor of the same animal (MTC cells) (Yamaguchi et al. 2005) suggesting that invadopodia are subcellular structures formed uniquely by carcinoma cells that have high invasive potential. Invadopodia are studied in vitro by plating metastatic carcinoma cells on thin cross-linked gelatin matrix (Artym et al. 2009) or thick cross-linked gelatin matrix overlayed with fibronectin (FN) (Yamaguchi et al. 2005; Alexander et al. 2008). This allows researchers to experimentally investigate the mechanisms responsible for regulating invadopodium formation, actin polymerization, and matrix degradation activity. The gelatin (for the thin matrix assay) or FN (for the thick matrix assay) is conjugated with fluorescent dyes (e.g. as Alexa488 or Alexa568) in order to visualize the degradation by fluorescence microscopy (Fig. 6.1c). It is clear that the matrix degradation activity of invadopodia is dependent on the activity of MMPs. Endogenous MMP inhibitors, TIMP-1 and TIMP-2, block matrix degradation by invadopodia in mammary carcinoma cells (Artym et al. 2006). In addition, blocking MMP activity using the chemical inhibitors, GM6001 or BB-94, block the matrix degradation activity of invadopodia in several cancer cell lines including mammary carcinoma cells (Ayala et al. 2008; Chan et al. 2009), head and neck squamous cells (Clark et al. 2007), and melanoma cells (Ayala et al. 2008). Knocking down MT1MMP, via siRNA, results in complete loss of matrix degradation by invadopodia in many cancer cell types (Artym et al. 2006; Sakurai-Yageta et al. 2008; Lizarraga et al. 2009) showing that MT1-MMP is an essential MMP for matrix degradation by invadopodia. MT1-MMP is thought to be the master regulator of matrix degradation in invadopodia (Poincloux et al. 2009) since it can activate pro-MMP-2 (an MMP also enriched in invadopodia (Clark and Weaver 2008)) by cleaving it into its active form (Deryugina et al. 2001) and can also directly cleave ECM components.

136

M. Oser et al.

Washout studies of the chemical MMP inhibitor BB94 in melanoma cells demonstrate that MMP inhibitors are useful tools to synchronize invadopodia providing a time zero for experimental manipulations (Ayala et al. 2008). When invasive tumor cells are plated on ECM, invadopodia form and first appear as punctate structures, or precursors, enriched in F-actin and comprised of several other protein markers including: cortactin, cofilin, Tks5, MT1-MMP, N-WASp, the Arp2/3 complex, and Fgd1 (Yamaguchi et al. 2005; Artym et al. 2006; Ayala et al. 2009; Chan et al. 2009; Oser et al. 2009). These invadopodium precursors mature through time to focally degrade the ECM underlying the location where the precursor has formed (Artym et al. 2006; Oser et al. 2009). Thus, an invadopodium precursor is defined as a structure that contains the same proteins and morphological appearance as a mature invadopodia, but has not yet acquired the ability to efficiently degrade matrix (Yamaguchi et al. 2005; Chan et al. 2009; Desmarais et al. 2009; Oser et al. 2009) (Fig. 6.1c, bottom left inset). Inhibition of MMP activity did not significantly inhibit invadopodium precursor formation in two mammary carcinoma cells lines (Artym et al. 2006; Chan et al. 2009) suggesting that matrix degradation occurs after the initial formation of the invadopodium precursor in mammary carcinoma cells. However, another study using head and neck squamous carcinoma cells observed that MMP activity was required for invadopodium precursor formation (Clark et al. 2007) suggesting that matrix degradation activity can promote the formation of invadopodium precursors in some cancer cell types. Once maturation has taken place, degradation is visualized as a loss in fluorescence of the ECM and matrix degradation is colocalized with markers of invadopodia when the structure is actively degrading matrix (Fig. 6.1c, top right inset). Many of the proteins that localize to invadopodia and have a critical role for invadopodium function are involved in regulating actin assembly (Table 6.1). In fact, the polymerization of actin in invadopodia is essential for proper invadopodium maturation (Yamaguchi et al. 2005; Oser et al. 2009).

6.4.1 Invadopodium Maturation Requires Actin Polymerization The formation of invadopodia and matrix degradation activity requires the coordination of many cell biological processes including integrin and growth factor receptor signaling (Buccione et al. 2009), actin cytoskeletal reorganization (Yamaguchi et al. 2005; Ayala et al. 2009), membrane trafficking (Sakurai-Yageta et al. 2008), and matrix metalloproteinase localization, activation, and secretion (Clark et al. 2007; Sakurai-Yageta et al. 2008). Depletion of individual proteins necessary for regulating these biological processes in invadopodia can result in inhibition of invadopodium formation suggesting that invadopodium maturation is a highly regulated process (Table 6.1). The regulation of the actin cytoskeleton at invadopodia is essential for proper invadopodium maturation (Yamaguchi et al. 2005; Oser et al. 2009; Stylli et al. 2009). Signaling pathways leading to increased number of free barbed ends for

yes (Yamaguchi et al. 2005)

yes (Yamaguchi et al. 2005)

yes (Yamaguchi et al. 2005)

Cofilin

N-WASp

Arp2/3 complex

yes (Oser et al. 2009)

Required for actin polymerization in invadopodia

no (Yamaguchi yes (Oser et al. et al. 2005; 2009) Desmarais et al. 2009) yes ND (Yamaguchi et al. 2005) yes yes (Oser et al. (Yamaguchi 2009) et al. 2005)

yes (Artym et al. yes (Artym 2006; Ayala et al. 2006; et al. 2008; Ayala et al. Oser et al. 2008; Oser 2009) et al. 2009)

Localization to invadopodia

Cortactin

Protein

Required for precursor formation

Required for MMP recruitment to invadopodia

Overexpression in metastatic Important for cancers tumor metastasis (evidence)

ND

ND

ND

ND

yes (Yamaguchi ND et al. 2005)

yes (Yamaguchi ND et al. 2005)

yes (Wang et al. 2004; Wang et al. 2007; Otsubo et al. 2004; Semba et al. 2006

yes (Chuma et al. 2004; Huang et al. 2003; Hui et al. 1997; Rodrigo et al. 2000; Zhang et al. 2006) yes (Yamaguchi yes (Wang et al. yes (Wang et al. et al. 2005) 2006) 2004; Gunnersen 2000) yes (Yanagawa yes (Yamaguchi ND et al. 2006) et al. 2005)

Required for ECM degradation

yes (Artym et al. yes (Artym et al. yes (Artym et al. yes (Li et al. 2006; Oser 2006; Clark 2006; Ayala 2001) et al. 2009) et al. 2007) et al. 2008; Oser et al. 2009; Clark et al. 2007)

Required for stabilization of invadopodia

Table 6.1 The known functions of actin regulatory proteins during the various stages of invadopodium maturation. Specifically, the table shows whether the protein localizes to invadopodia and at which stage during invadopodium maturation (precursor formation, actin polymerization, stabilization, MMP recruitment, and ECM degradation) each protein is required. Lastly, it is shown whether each protein is important for tumor metastasis in animal models and whether there is evidence of overexpression in metastatic cancers

6 Actin-based Motile Processes in Tumor Cell Invasion 137

yes (Ayala et al. 2009) yes (Lizarraga et al. 2009)

Fgd1

DRFs

ND

yes (Yamaguchi et al. 2005) yes (Philippar et al. 2008) yes (Seals et al. 2005)

Localization to invadopodia

Cdc42

Tks5

Mena

Nck1

Protein

Required for actin Required for polymerization stabilization of in invadopodia invadopodia

yes (Ayala et al. ND 2009) yes (Lizarraga ND et al. 2009)

yes (Yamaguchi ND et al. 2005)

ND

ND

ND

yes (Yamaguchi yes (Oser et al. ND et al. 2005) 2009) ND ND yes (Philippar et al. 2008) yes (Seals et al. ND ND 2005)

Required for precursor formation

ND

ND

ND

ND

ND

ND

Required for MMP recruitment to invadopodia

Table 6.1 (continued)

yes (Yamaguchi et al. 2005; SakuraiYageta et al. 2008) yes (Ayala et al. 2009) yes (Lizarraga et al. 2009)

yes (Yamaguchi et al. 2005) yes (Philippar et al. 2008) yes (Diaz et al. 2009; Stylli et al. 2009)

Required for ECM degradation

ND

ND

ND

yes (Ayala et al. 2009) yes (Di Vizio et al. 2009; Zhu et al. 2008)

yes (Philippar yes (Wang et al. et al. 2008) 2004, 2007) Not important for yes (Seals et al. experimental 2005) metastasis(Blouw et al. 2008) ND yes (Wang et al. 2004, 2007)

ND

Overexpression in metastatic Important for cancers tumor metastasis (evidence)

138 M. Oser et al.

yes ND (Baldassarre et al. 2003) yes (SakuraiND Yageta et al. 2008) ND yes (SakuraiYageta et al. 2008)

Dynamin II

ND = not determined

RhoA

IQGAP1

yes (Alexander no (Alexander et al. 2008) et al. 2008) (surrounding)

Nonmuscle myosin II

Protein

Required for Localization to precursor invadopodia formation

ND

ND

ND

ND

Required for actin polymerization in invadopodia

ND

ND

ND

ND

Required for stabilization of invadopodia

yes (SakuraiYageta et al. 2008) ND

ND

ND

Required for MMP recruitment to invadopodia

Table 6.1 (continued)

yes (SakuraiYageta et al. 2008) yes (SakuraiYageta et al. 2008)

yes (Baldassarre et al. 2003)

yes (Alexander et al. 2008)

Required for ECM degradation

ND

ND

ND

ND

yes (Jadeski et al. 2008; Miyoshi et al. 2005 ) yes (Wang et al. 2004; Zhang et al. 2009)

yes (Croft et al. 2004; Kamai et al. 2003; Minamiya et al. 2005; Somlyo et al. 2000) ND

Overexpression Important in metastatic for tumor cancers metastasis (evidence)

6 Actin-based Motile Processes in Tumor Cell Invasion 139

140

M. Oser et al.

actin polymerization by cofilin and the Arp2/3 complex (Yamaguchi et al. 2005; Oser et al. 2009), or increased rates of actin monomer addition to free barbed ends by Mena (Philippar et al. 2008) and diaphanous-related formins (DRFs) (Lizarraga et al. 2009) are critical for invadopodium maturation and efficient matrix degradation activity (Table 6.1). Interestingly, an invadopodium precursor can form and never acquire the ability to polymerize actin within the structure (Oser et al. 2009). Invadopodia deficient in polymerizing actin also exhibit reduced matrix degradation activity (Yamaguchi et al. 2005; Oser et al. 2009; Stylli et al. 2009) suggesting that actin polymerization is critical for the acquisition of matrix degradation activity by invadopodia. Specific examples include cells depleted of cofilin or expressing cortactin with mutated tyrosine phosphorylation sites show normal formation of invadopodium precursors, but fail to polymerize actin in invadopodia (Oser et al. 2009) (Table 6.1 and 6.2). The result is inefficient ECM degradation by invadopodia (Yamaguchi et al. 2005; Oser et al. 2009) and inhibition of tumor cell metastasis in vivo (Li et al. 2001; Wang et al. 2006). These studies strongly suggest that the mechanisms that control actin polymerization in invadopodia are critical for tumor cell invasion during metastasis. The actin cytoskeleton in invadopodia is thought to provide the physical structure to spatially localize MMP activity and allow protrusion of invadopodia through the ECM (Albiges-Rizo et al. 2009). Recent evidence suggests that, in addition their roles in ECM degradation, invadopodia may also function to regulate tumor cell chemotaxis and cellular pathfinding. Cortactin and N-WASp, proteins that are critical for invadopodium formation (Yamaguchi et al. 2005; Artym et al. 2006; Cortesio et al. 2008), but not the initial lamellipodial protrusion (Sarmiento et al. 2008; Desmarais et al. 2009), are required for tumor cell chemotaxis toward EGF (Desmarais et al. 2009) suggesting that the regulation of actin polymerization in invadopodia is linked to chemotaxis. Leukocytes have been shown to form “invasive podosomes”, a related structure to invadopodia formed in myelocytic cells. These “invasive podosomes” function in vivo to probe the endothelial cells lining blood vessels for appropriate locations to initiate transcellular diapedesis. Once the location for diapedesis has been determined, the “invasive podosomes” elongate and allow the leukocyte to cross the endothelium transcellularly (Carman et al. 2007). It is conceivable that invadopodia have a similar pathfinding function in tumor cells that allow for intravasation during metastasis. However, it is unclear whether tumor cells use invadopodia for pathfinding and chemotaxis in vivo, and how the mechanisms used to regulate actin polymerization in invadopodia are linked to tumor cell chemotaxis as observed in vitro.

6.4.2 Regulation of Actin Polymerization in Invadopodia The proper control of cofilin and the Arp2/3 complex activities are critical for actin polymerization in both invadopodia (Yamaguchi et al. 2005; Oser et al. 2009) and lamellipodia (Mouneimne et al. 2004). As described earlier, these same pathways are also important for chemotaxis and are highly upregulated in the invasive

ND

Phosphorylation of serines 113, 405, and 418

ND = not determined

no (Oser et al. 2009)

ND

yes (Oser et al. 2009)

ND

yes (Oser et al. 2009)

Phosphorylation of tyrosines 421, 466, and 482

ND

yes (Oser et al. 2009)

NTA domain (Arp2/3 binding domain) SH3 domain

Important for actin polymerization in invadopodia

Important for invadopodium precursor formation

Protein domain or phosphorylation site on cortactin

yes (Oser et al. 2009) (dephosphorylation stabilizes invadopodia) ND

ND

ND

Important for stabilization of invadopodia

yes (Ayala et al. 2008)

yes (Ayala et al. 2008; Oser et al. 2009) yes (Ayala et al. 2008; Oser et al. 2009) yes (Ayala et al. 2008; Oser et al. 2009)

Required for ECM degradation

ND

yes (Li et al. 2001)

ND

ND

Important for tumor metastasis

Table 6.2 The table shows which protein domains or phosphorylation sites on cortactin are important during each stage (precursor formation, actin polymerization, stabilization, and ECM degradation) of invadopodium maturation, and which functions of cortactin are known to be important for tumor metastasis in animal models

6 Actin-based Motile Processes in Tumor Cell Invasion 141

142

M. Oser et al.

Fig. 6.3 Comparison of the signaling pathways activated after EGF stimulation that induce actin polymerization in invadopodia and lamellipodia. Although actin polymerization in both invadopodia and lamellipodia requires cofilin and the Arp2/3 complex, the upstream activators are unique in each subcellular compartment. The primary mechanism responsible for regulating cofilin is the binding and release from cortactin in invadopodia (left) or PI(4, 5)P2 in lamellipodia (right). The primary mechanism that controls the Arp2/3 complex is activation by N-WASp in invadopodia (left) or WAVE2 in lamellipodia (right). In invadopodia (left), N-WASp activity is thought to be controlled by either an Fgd1/Cdc42 pathway and/or a phosphorylated cortactin/Nck1 pathway. Yellow highlights unique pathways in invadopodia and blue highlights unique pathways in lamellipodia

mammary tumor cell population. Although cofilin and the Arp2/3 complex are both critical for the formation of free barbed ends in both invadopodia and lamellipodia (DesMarais et al. 2004; Mouneimne et al. 2004; Oser et al. 2009), the regulation of cofilin and the Arp2/3 complex appears to differ in the two compartments in the same metastatic mammary carcinoma cell type (Yamaguchi et al. 2005; Sarmiento et al. 2008; Oser et al. 2009) (Fig. 6.3). The unique regulation of cofilin and the Arp2/3 complex in invadopodia involves the scaffolding protein cortactin (Desmarais et al. 2009; Oser et al. 2009). 6.4.2.1 Role of Cortactin During the Stages of Invadopodium Maturation and Tumor Cell Invasion Cortactin is overexpressed (most commonly by chromosome amplification of 11q13) in many human cancers including head and neck squamous carcinomas, breast, and ovarian cancers, and is important for the metastasis of breast or esophageal cancers in mouse models (for a review see (Weaver 2008)). Specifically,

6

Actin-based Motile Processes in Tumor Cell Invasion

143

the tyrosine phosphorylation of cortactin is important for the metastasis of breast tumors to bone in xenograft mouse models (Li et al. 2001). These in vivo mouse studies and human correlates suggest that cortactin function is critical for tumor cell metastasis. Studies using tumor cells in culture have demonstrated that cortactin is required for the formation of invadopodia in many cell types (Artym et al. 2006; Clark et al. 2007; Ayala et al. 2008), but is not essential for the formation of lamellipodial protrusions in at least two cancer cell types (Bryce et al. 2005; Kempiak et al. 2005; Desmarais et al. 2009). In the lamellipodia of fibrosarcoma cells, cortactin is important for lamellipodium persistence and the formation of new adhesions after the initial stage of lamellipodium formation (Bryce et al. 2005). These results suggest that cortactin may have essential regulatory functions in tumor cells during invadopodium maturation that are not required of cortactin during lamellipodium initiation. Recently, a study using cortactin knockout mice showed that cortactin function is not essential for the migration of mouse embryonic fibroblasts (Tanaka et al. 2009). Together, the findings that cortactin is required for tumor cell metastasis, but not for the migration of non-malignant cells makes cortactin an appealing candidate for anti-metastatic therapy with potentially minimal side effects. We will now discuss the biochemical properties of cortactin that allow it to regulate multiple processes during invadopodium maturation. Cortactin is a 63–65 kDa protein that was identified as a substrate of Src-family kinases almost 20 years ago (Wu and Parsons 1993). Cortactin contains an Nterminal acidic domain (NTA) that binds Arp3 and weakly activates the Arp2/3 complex (Weed et al. 2000; Uruno et al. 2001; Weaver et al. 2001), followed by Factin binding motifs that function to bind to the branch points of actin filaments and stabilize them (Bryce et al. 2005). Together, the NTA and F-actin binding motifs are important for activating the Arp2/3 complex and stabilizing the actin filament branches generated by the Arp2/3 complex (Weaver et al. 2003). The Arp2/3 binding domain of cortactin is important for invadopodium formation and function in melanoma and mammary carcinoma cells (Ayala et al. 2008; Oser et al. 2009) (Table 6.2). At its C-terminus, cortactin contains an SH3 domain that has scaffolding functions to regulate actin assembly. The SH3 domain binds directly to N-WASp (Martinez-Quiles et al. 2004), dynamin II (Schafer et al. 2002), and WIP (Kinley et al. 2003). The binding of cortactin to N-WASp can activate N-WASp resulting in Arp2/3 activation (Kowalski et al. 2005). The SH3 domain of cortactin and pathways leading to N-WASp activation are both critical for invadopodium formation in mammary carcinoma and melanoma cells (Yamaguchi et al. 2005; Ayala et al. 2008; Oser et al. 2009) (Table 6.1 and 6.2). Thus, cortactin can activate the Arp2/3 complex either directly through its NTA domain, or indirectly by binding to N-WASp via its SH3 domain, and both of these mechanisms of Arp2/3 activation are important for proper invadopodium function (Ayala et al. 2008; Oser et al. 2009). Cortactin also binds to dynamin II via its SH3 domain and stimulates dynamin II’s GTPase activity (Schafer et al. 2002; Mooren et al. 2009). In turn, dynamin II’s GTPase activity is important for the global organization of the actin cytoskeleton (Mooren et al. 2009) and for the formation and function of invadopodia (Baldassarre et al.

144

M. Oser et al.

2003) (Table 6.1) suggesting that cortactin can indirectly regulate the organization of the actin cytoskeleton in invadopodia through binding interactions with dynamin II. In summary, the NTA and SH3 domains of cortactin allow cortactin to interact with and regulate many proteins that control invadopodium function including the Arp2/3 complex, N-WASp, and dynamin II. The C-terminal domain of cortactin also contains both tyrosine and serine phosphorylation sites. Here, we will focus on the functional consequences of tyrosine phosphorylation of sites 421, 466, and 482 since they have been identified as substrates of Src-family kinases, are known to have a critical role during actin assembly both in vitro (Tehrani et al. 2007) and in invadopodia (Oser et al. 2009) (Table 6.2), and during tumor cell metastasis in vivo (Li et al. 2001). In addition to Src, the tyrosine residues 421, 466, and 482 on cortactin are phosphorylated by numerous other kinases including: Arg, Abl, Fer, Fyn, Syk, and Met (for a review see (Ammer and Weed 2008)). It has been proposed that the Abl-family kinases (Abl and Arg) are the kinases that directly phosphorylate cortactin in vivo since Abl and Arg phosphorylate cortactin with a lower Km (Boyle et al. 2007) and the SH2 domain of Arg binds to phosphorylated cortactin with a dramatically higher affinity in vitro compared to Src (Christopher Mader, personal communication). Cortactin tyrosine phosphorylation induced by PDGF stimulation was completely eliminated in Abl-/- Arg-/- or Src-/- Yes-/- Fyn-/- (SYF) knockout fibroblasts suggesting that Abl and Src-family kinases are in the same signaling pathway leading to cortactin tyrosine phosphorylation. Based on the relative binding affinities and enzyme-substrate interactions for Abl-family kinases compared to Src-family kinases, it is reasonable to hypothesize that Src-family kinases function upstream to activate Abl-family kinases, and that Abl-family kinases phosphorylate cortactin directly. Although it is known that cortactin tyrosine phosphorylation is important for cell migration and chemotaxis (Luo et al. 2006), invadopodium function (Ayala et al. 2008) and for breast cancer metastasis in mice (Li et al. 2001), the precise mechanisms responsible for these phenotypes during tumor cell invasion are only beginning to be understood. In vitro studies have demonstrated that cortactin tyrosine phosphorylation is important for the recruitment of SH2 domain proteins involved in the activation of pathways of actin polymerization (Tehrani et al. 2007). Specifically, cortactin tyrosine phosphorylation is important for the recruitment and activation of the upstream N-WASp activator Nck1. A complex of tyrosine phosphorylated cortactin, Nck1, N-WASp, and the Arp2/3 complex leads to maximal generation of free barbed ends in vitro independent of cortactin’s ability to bind to the Arp2/3 complex via its NTA domain or to bind to N-WASp via its SH3 domain (Tehrani et al. 2007). Similar to this in vitro phosphorylated cortactin pathway, cortactin tyrosine phosphorylation is specifically important for the recruitment of Nck1 and the spatial enrichment of N-WASp activity in invadopodia leading to the amplification of free barbed ends in invadopodia (Oser et al. 2009). Together, these studies provide strong evidence that cortactin tyrosine phosphorylation functions during tumor cell invasion to recruit and localize SH2 domain proteins involved in actin assembly to specific subcellular compartments.

6

Actin-based Motile Processes in Tumor Cell Invasion

145

Cortactin also binds directly and inhibits the activity of cofilin in invadopodia (Oser et al. 2009). When cortactin is tyrosine phosphorylated, it can no longer inhibit cofilin’s severing activity. The result is dramatic amplification of free barbed ends leading to actin polymerization in invadopodia (Oser et al. 2009). Thus, apart from regulating the recruitment of SH2 domain proteins to invadopodia, cortactin tyrosine phosphorylation can also function to regulate the activities of effectors that do not have SH2 domains (e.g. cofilin) involved in actin polymerization. Together, the tyrosine phosphorylation of cortactin leads to activation of both cofilin and the Arp2/3 complex in invadopodia resulting in efficient actin polymerization (Oser et al. 2009).

6.4.3 The Stages of Invadopodium Maturation There is evidence that the maturation to a degradation-competent invadopodium requires a stepwise order of temporal and spatial events, referred to as the stages of invadopodium maturation, that lead to a fully functional invadopodium that efficiently degrades ECM. Interestingly, many of these maturation events are regulated by cortactin (Artym et al. 2006; Clark et al. 2007; Oser et al. 2009). The pioneer study that identified stages of invadopodium maturation (Artym et al. 2006) showed that the formation of a cortactin- and actin-containing invadopodium precursor precedes MT1-MMP recruitment and ECM degradation by invadopodia in mammary carcinoma cells. Using live cell imaging, they demonstrated clear temporal separation from the localization of cortactin and actin at invadopodium precursors to ECM degradation. Blocking any step of the maturation process (e.g. knocking down of cortactin or MT1-MMP) prevented matrix degradation by invadopodia suggesting that distinct temporal maturation events are required for ECM degradation to take place. Another study showed that cortactin is essential for the recruitment of MMPs (MT1-MMP, MMP2, and MMP9) to invadopodia in head and neck squamous carcinoma cells (Clark et al. 2007). Interestingly, Arp3- and F-actin-positive invadopodium precursors still formed (although the number of precursors/cell were significantly reduced) in cells depleted for cortactin (Clark et al. 2007). However, the precursors that formed failed to recruit MMPs and therefore had markedly reduced ECM degradation activity compared to the reduction in the number of F-actin- and Arp3-positive precursors/cell. These studies provide evidence that there is a temporal ordering to the biological processes that are required for proper invadopodium maturation and that these processes must be precisely coordinated in time and space. Recently, some of the molecular mechanisms involved in regulating the stages involved in invadopodium maturation (Artym et al. 2006) have been elucidated in mammary adenocarcinoma cells (Oser et al. 2009) (Fig. 6.4). The earliest stage, termed invadopodium precursor formation, specifically requires the Arp2/3 and N-WASp binding domains of cortactin. Following the formation of the precursor, cortactin is tyrosine phosphorylated, resulting in the activation of signaling pathways leading to free barbed end formation for actin polymerization. The

146

M. Oser et al.

Fig. 6.4 The stages of invadopodium maturation. During precursor formation (stage 1), cortactin, Tks5, N-WASp, cofilin, and Arp2/3 form a complex involving cortactin’s Arp2/3 and N-WASp binding domains. Tks5 is thought to promote the formation of invadopodium precursors when binding to PI(3, 4)P2 on the ventral cell surface plasma membrane. In addition, Fgd1 activation of Cdc42 signaling is also necessary for invadopodium precursor formation. Following precursor formation, cortactin is tyrosine phosphorylated, which activates cofilin’s severing activity to generate free barbed ends and the Arp2/3 complex can use these cofilin-generated barbed ends for efficient actin polymerization (stage 2). Cortactin is then dephosphorylated, which stabilizes the invadopodium precursor for maturation (stage 3). The presence of Mena enhances the stabilization of invadopodia while calpain 2 promotes the disassembly of invadopodia at stage 3. Box at stage 3+4 indicates stabilization. MT1-MMP can be recruited at stages 2–4. Stages 1–3 are required for a precursor to become a mature invadopodium that efficiently degrades ECM (stage 4). Modified from Figure 9, © Oser et al. (2009). Originally published in /J. Cell Biol./10.1083/jcb.200812176. Aug 24; 186(4):571–87

tyrosine phosphorylation of cortactin marks the beginning of the second stage of invadopodium maturation, termed activation of actin polymerization. Initially during precursor formation, cortactin directly binds to cofilin and inhibits cofilin’s severing activity. Following stimulation with EGF to induce the maturation of invadopodia, cortactin is tyrosine phosphorylated releasing its inhibitory effect on cofilin, and cofilin severs actin filaments to create free barbed ends for actin polymerization in invadopodia. In addition to its regulation of cofilin, cortactin tyrosine phosphorylation also regulates Nck1 recruitment to invadopodia resulting in the

6

Actin-based Motile Processes in Tumor Cell Invasion

147

local activation of an Nck1/N-WASp/Arp2/3 signaling pathway. Together, cofilin and the Arp2/3 complex cooperate for efficient actin polymerization in invadopodia. Cofilin is first released from cortactin to generate barbed ends by its severing activity and Arp2/3 is indirectly activated by cortactin phosphorylation via Nck1 and N-WASp pathway. As discussed earlier, cofilin and the Arp2/3 complex synergize to efficiently polymerize actin both in vitro (Ichetovkin et al. 2000) and in vivo (DesMarais et al. 2004) and this recent study (Oser et al. 2009) suggests that they synergize to polymerize actin in invadopodia. After the initial polymerization of actin in invadopodia, cortactin is dephosphorylated allowing cortactin to re-bind to cofilin preventing excessive cofilin severing activity, which stabilizes invadopodia (stage 3). Calpain 2, a calcium-dependent intracellular protease, increases the disassembly of invadopodia and therefore decreases the stability of invadopodia (Cortesio et al. 2008). Interestingly, cortactin is a specific calpain 2 substrate (Perrin et al. 2006) and Src-dependent tyrosine phosphorylation of cortactin increases calpain-induced proteolysis (Huang et al. 1997). This suggests that the dephosphorylation of cortactin may also protect cortactin from calpain-dependent proteolysis resulting in the stabilization of invadopodia. In addition to cortactin dephosphorylation, high expression of the anti-capping protein mena has also been shown to increase the stabilization of invadopodia (Philippar et al. 2008). Together, these studies demonstrate that multiple mechanisms can contribute to the stabilization of invadopodia. There is evidence that during the stabilization stage, MMPs are being recruited and targeted to invadopodia (Artym et al. 2006; Clark et al. 2007; Sakurai-Yageta et al. 2008; Liu et al. 2009). Finally, the invadopodium has matured and is now degradation-competent (stage 4).

6.4.4 Function and Regulation of Cofilin During Invadopodium Maturation Similar to lamellipodia, cofilin is important for the generation of free barbed ends in invadopodia (Oser et al. 2009). Studies using mammary carcinoma cells depleted of cofilin demonstrated that cofilin is also important for the stabilization of invadopodia (Yamaguchi et al. 2005) suggesting that the cofilin-generated free barbed ends contribute to invadopodium stabilization. Furthermore, cofilin is not important for invadopodium precursor formation (Desmarais et al. 2009) (Table 6.1). Thus, it appears that cofilin activity is required at a narrow time window during invadopodium maturation suggesting that the turning on and off of cofilin activity must be tightly regulated in invadopodia. In both lamellipodia and invadopodia of invasive tumor cells, the initial activation of cofilin is regulated by binding to an inhibitor: PI(4, 5)P2 in lamellipodia and cortactin in invadopodia (Fig. 6.5). As described above, in the plasma membrane in resting mammary carcinoma cells, cofilin is inhibited by PI(4, 5)P2 binding. After EGF stimulation, PI(4, 5)P2 is hydrolyzed by PLCγ1 and cofilin is released from the membrane, can bind to actin, and sever actin filaments to create free barbed

Fig. 6.5 (continued)

148 M. Oser et al.

6

Actin-based Motile Processes in Tumor Cell Invasion

149

ends for actin polymerization (van Rheenen et al. 2007). Cofilin must then be recycled back to the plasma membrane through a phosphorylation/dephosphorylation cycle to re-bind to PI(4, 5)P2 . In contrast to lamellipodia, the ventral cell surface plasma membrane where invadopodia form is depleted for PI(4, 5)P2 (Oser and Condeelis 2009; Oikawa et al., 2008) and enriched for PI(3, 4)P2 (Oikawa et al. 2008). Tks5, a scaffolding protein enriched in invadopodia (Oser et al. 2009) and required for invadopodium formation (Stylli et al. 2009), binds selectively to PI(3, 4)P2 (Abram et al. 2003), and the accumulation of Tks5 at PI(3, 4)P2 -enriched membrane locations is essential for podosome formation (Oikawa et al. 2008). Thus, the cofilin activity cycle in invadopodia requires a distinct regulatory mechanism from lamellipodia. As described above, the inhibitor in invadopodia that sequesters cofilin is cortactin (Oser et al. 2009). Cofilin bound to cortactin inhibits cofilin’s severing activity in resting mammary carcinoma cells. Following stimulation with EGF to induce the maturation of invadopodia, cortactin is tyrosine phosphorylated releasing its inhibitory effect on cofilin, and cofilin severs actin filaments to create free

Fig. 6.5 The cofilin activity cycle in lamellipodia and invadopodia. The PM at lamellipodia is enriched with PtdIns(4, 5)P2 and the on/off binding of cofilin to PtdIns(4, 5)P2 is the primary mechanism used to regulate cofilin activity. Invadopodia form in PtdIns(3, 4)P2 -enriched PM areas through a Tks5-PtdIns(3, 4)P2 binding interaction. In invadopodia, the on/off binding of cofilin to cortactin is the primary mechanism used to regulate cofilin activity. White and black arrows indicate pathways used to regulate cofilin activity in lamellipodia and invadopodia, respectively. (left) Cofilin cycles through three compartments near lamellipodia: the cytosol, F-actin and the plasma membrane (PM). Cofilin remains inactive at the PM by binding to PtdIns(4, 5)P2 , and in the cytosol when it is serine phosphorylated. When activated by either PLCγ1-mediated PtdIns(4, 5)P2 hydrolysis or dephosphorylation by SSH, cofilin translocates to the F-actin compartment where it binds and severs actin filaments resulting in the generation of free barbed ends and the formation of cofilin–G-actin complexes. These free barbed ends are amplified by WAVE2-dependent Arp2/3 activation resulting in efficient actin polymerization and the formation of cellular protrusions in lamellipodia. The release of cofilin from PtdIns(4, 5)P2 at the PM is amplified by an increase in pH (mediated by the Na+ -H+ exchanger NHE1), which reduces the affinity of cofilin for PtdIns(4, 5)P2 . Cofilin is then phosphorylated by LIM Kinase to inactivate cofilin. The cycle repeats when cofilin is dephosphorylated by SSH to either recycle cofilin to the PM or directly initiate actin filament severing by cofilin. “+” indicates pH increase. (Right) Cofilin cycles through two compartments near invadopodia: the cytosol and the F-actin and the PM. Cofilin remains inactive in the F-actin compartment by binding to cortactin, and in the cytosol when it is serine phosphorylated. When cortactin is tyrosine phosphorylated by either Abl or Src-family kinases, cortactin no longer inhibits cofilin’s severing activity and cofilin binds and severs actin filaments resulting in the generation of free barbed ends and the formation of cofilin–G-actin complexes. These free barbed ends are amplified by N-WASp-dependent Arp2/3 activation resulting in efficient actin polymerization in invadopodia. In addition, cortactin tyrosine phosphorylation activates Dynamin II’s GTPase activity, which remodels actin filaments making them more accessible to cofilin. Cofilin is then phosphorylated by LIM Kinase to inactivate cofilin. The cycle repeats when both cofilin and cortactin are dephosphorylated allowing the re-binding of cofilin to cortactin and inhibition of cofilin severing activity. Black arrows indicate primary pathways that regulate cofilin activity, yellow arrows indicate indirect pathways, and blue arrows indicate pathways downstream of cofilin severing activity. Modified from Figure 4A, Oser and Condeelis (2009 108: 1252–1262)

150

M. Oser et al.

barbed ends for actin polymerization in invadopodia. Cofilin must then be recycled back to re-bind cortactin possibly through a phosphorylation/dephosphorylation cycle. However, it is unknown whether cofilin is also regulated by phosphorylation/dephosphorylation in invadopodia. Both cortactin and cofilin are involved in many other cellular processes involving actin polymerization (For reviews see (Ammer and Weed 2008; Van Troys et al. 2008). Future studies will elucidate whether cortactin regulates cofilin in other subcellular compartments beside invadopodia. Apart from cortactin’s direct regulation of cofilin severing activity, cortactin may indirectly regulate cofilin’s activity in invadopodia. Cortactin regulates dynamin II’s GTPase activity resulting in the reorganization of actin filaments and making the filaments more accessible to cofilin (Mooren et al. 2009) (Fig. 6.5). Interestingly, when cortactin is tyrosine phosphorylated, it has increased affinity for dynamin II (Zhu et al. 2007) resulting in increased dynamin II GTPase activity, which may further enhance cofilin’s severing activity. Invadopodia are enriched in cortactin (Bowden et al. 2006), cofilin (Yamaguchi et al. 2005), and dynamin II (Baldassarre et al. 2003) and thus increased tyrosine phosphorylation of cortactin upon growth factor stimulation may initiate cofilin activity both through a direct binding interaction and indirectly through dynamin II.

6.4.5 Pathways Leading to Activation of the Arp2/3 Complex During Invadopodium Maturation The proper maturation of invadopodia requires signaling pathways leading to the activation of the Arp2/3 complex (Yamaguchi et al. 2005; Sakurai-Yageta et al. 2008; Ayala et al. 2009). The Arp2/3 complex is specifically required for invadopodium precursor formation (Yamaguchi et al. 2005), actin polymerization in invadopodia (Oser et al. 2009), and the targeting of MMP-containing vesicles to the plasma membrane where invadopodia form (Liu et al. 2009) (Table 6.1). Although both lamellipodia (DesMarais et al. 2004) and invadopodia (Oser et al. 2009) require the Arp2/3 complex for actin polymerization, distinct signaling pathways initiate Arp2/3 activation in the two subcellular compartments (Yamaguchi et al. 2005; Sarmiento et al. 2008) (Fig. 6.3). In the lamellipodium of mammary carcinoma cells, the dominant mechanism for activation of the Arp2/3 complex is dependent on WAVE2 (Sarmiento et al. 2008). Interestingly, WAVE2 does not localize to invadopodia (Yamaguchi et al. 2005). In the same mammary carcinoma cell type, N-WASp localizes to invadopodia and is essential for the formation and maturation of invadopodium (Yamaguchi et al. 2005). Furthermore, N-WASp-dependent Arp2/3 activation is not important for lamellipodium formation (Sarmiento et al. 2008). These results provide strong evidence that the Arp2/3 complex is regulated by mechanisms involving N-WASp or WAVE2 in invadopodia and lamellipodia, respectively.

6

Actin-based Motile Processes in Tumor Cell Invasion

151

Several components of signaling pathways leading to N-WASp-dependent activation of the Arp2/3 complex are important for invadopodium maturation (Yamaguchi et al. 2005; Sakurai-Yageta et al. 2008; Ayala et al. 2009) (Fig. 6.3 and Table 6.1) providing strong evidence that N-WASp-dependent Arp2/3 activation is essential for actin polymerization in invadopodia. Specifically, Cdc42 activation is required for invadopodium precursor formation and matrix degradation in both mammary carcinoma cells (Yamaguchi et al. 2005; Sakurai-Yageta et al. 2008) and melanoma cells (Ayala et al. 2009) (Table 6.1). Fgd1, a Cdc42-specific GEF that activates Cdc42, is important for invadopodium formation and function in melanoma cells (Ayala et al. 2009). Fgd1 colocalizes with F-actin in invadopodia at times preceding matrix degradation suggesting that Fgd1 plays an important role early during invadopodium maturation. In summary, current evidence suggests that an Fgd1/Cdc42/N-WASp pathway leading to activation of the Arp2/3 complex is essential for invadopodium precursor formation and matrix degradation activity by invadopodia. In addition to Cdc42, N-WASp can also be activated by Nck1 (Rohatgi et al. 2001). Nck1 is important for actin polymerization and matrix degradation activity by invadopodia in both melanoma and mammary carcinoma cells (Oser et al. 2009; Stylli et al. 2009) (Table 6.1). Two recent studies have demonstrated that Nck1 recruitment to invadopodia involves the interaction of Nck1’s SH2 domain with phosphorylated tyrosines on either cortactin (Oser et al. 2009) or Tks5 (Stylli et al. 2009). Cortactin tyrosine phosphorylation is important for the recruitment of Nck1 and enrichment of N-WASp activity in invadopodia in mammary carcinoma cells (Oser et al. 2009). The proposed signaling pathway involves tyrosine phosphorylated cortactin/Nck1/N-WASp/Arp2/3 and is supported by in vitro data demonstrating that tyrosine phosphorylated cortactin binds directly to Nck1 and that a complex of phosphorylated cortactin/Nck1/N-WASp/Arp2/3 maximally stimulates barbed end formation (Tehrani et al. 2007). In melanoma cells, it was demonstrated that Tks5 phosphorylation by Src is important for the recruitment of Nck1 to invadopodia (Stylli et al. 2009). Both studies support a model in which Nck1 is recruited to invadopodia through tyrosine phosphorylation. The two studies draw different conclusions as to which scaffolding protein, cortactin or Tks5, recruits Nck1 to invadopodia. A recent study shows that Tks5 recruits cortactin to podosomes (Crimaldi et al. 2009) suggesting that the Tks5 recruitment of Nck1 may be indirect and mediated by cortactin since cortactin and Nck1 are known to directly bind in vitro (Tehrani et al. 2007). It is also possible that both cortactin and Tks5 recruit Nck1 to invadopodia or that the differences are cell-type dependent. Future studies will be required to determine the specificity of Nck1 for phosphorylated tyrosine residues in invadopodia. Regardless of the phosphorylated protein responsible for recruiting Nck1 to invadopodia, it appears that Nck1 functions in invadopodia to initiate N-WASp activity leading to actin polymerization after invadopodium precursor formation (Oser et al. 2009). The current evidence suggests that Cdc42-dependent N-WASp activation leads to invadopodium precursor formation (Yamaguchi et al. 2005; Ayala et al. 2009) and Nck1-dependent N-WASp activation leads to actin polymerization in invadopodia at a stage following precursor formation. It will be interesting

152

M. Oser et al.

to determine the regulation that temporally segregates Cdc42 and Nck1 activities during the early stages of invadopodium maturation.

6.5 Formin-Dependent Actin Nucleation During Tumor Cell Invasion In addition to the Arp2/3 complex, additional mechanisms for controlling actin polymerization comprises the formin family (referred to hereafter as diaphanousrelated formins (DRFs)), characterized by the presence of highly conserved formin homology domains, FH1 and FH2 (Higgs 2005). Although the Arp2/3 complex and formins both nucleate actin filaments, each produces an actin filament network with unique geometries. The Arp2/3 complex generates a dendritic array of filaments by nucleating filaments at various branch points off the parent filament. Formins, however, nucleate the polymerization of linear actin filaments from actin monomers alone. DRFs generate linear actin filaments essential for the formation of stress fibers (Hotulainen and Lappalainen 2006) and filopodia (Schirenbeck et al. 2005). DRFs are targets of Rho GTPases and have been implicated in cytokinesis, cell motility and polarity (Faix and Grosse 2006; Gomez et al. 2007), which are often deregulated during cell transformation and tumor metastasis (Sahai 2005). Evidence supporting a role for DRFs during tumor cell protrusion has been reported in a variety of cancer cell models. Analysis of primary and metastatic human prostate tumors revealed a significantly higher frequency of deletion of DRF3 locus in metastatic tumors compared with primary tumors (Di Vizio et al. 2009). Stable overexpression of the DRF Formin Homology 2 Domain containing 1 (FHOD1) which can interact with Rac, increased the migration of WM35 melanoma cells on a collagen-coated Transwell assay (Koka et al. 2003), while FHOD2 overexpression is correlated with human metastatic colorectal cancer and increased invasion in vitro (Zhu et al. 2008). siRNA knockdown of the DRF1 but not DRF2 reduced 3D motility of human breast cancer MDA-MB-435 cells in a Transwell assay (Kitzing et al. 2007). In addition, only DRF1 siRNA knockdown altered the bleb-associated protrusions produced during cancer MDA-MB-435 cell invasion in Matrigel. However, impaired Matrigel invasion of MDA-MB-231 breast cancer cells was observed when each individual DRF isoform (DRF1-DRF3) is silenced by siRNA (Lizarraga et al. 2009). Interestingly, the EGF-induced increase in barbed ends at the lamellipodium and subsequent plasma membrane protrusion of MTLn3 cells was unaffected by knockdown of either DRF1 or DRF2 in the presence of the Arp2/3 regulators N-WASp and WAVE-2 (Sarmiento et al. 2008), suggesting a more complex role for DRF activity during the distinctive motility cycles of various tumor cells. Both DRF1 and DRF2 have recently been shown to play an important role during invadopodium maturation in human breast cancer MDA-MB-231 cells (Table 6.1). DRF1-3 siRNA silencing resulted in inhibition of both invadopodium precursor formation and matrix degradation activity of invadopodia (Lizarraga et al. 2009).

6

Actin-based Motile Processes in Tumor Cell Invasion

153

Interestingly, these isoform-specific knockdown phenotypes were not additive suggesting that each DRM isoform may have the capacity to form heterodimers (Copeland et al. 2007) and/or play unique roles in the nucleation of F-actin during the various stages of invadopodium maturation and tumor cell invasion (Lizarraga et al. 2009). Potential crosstalk between DRF- and Arp2/3-dependent mechanisms of actin nucleation was uncovered when N-WASp and WAVE2, positive upstream regulators of Arp2/3 activity, were simultaneously knocked down by siRNA in the mammary carcinoma cells (Sarmiento et al. 2008). The reduction of N-WASp and WAVE2 activity caused an increase in RhoA activity that was correlated with an increased formation of filopodia. siRNA knockdown of DRF1, but not DRF2 inhibited the enhanced filopodia formation phenotype establishing a link between the Arp2/3and DRF-dependent pathways for actin nucleation in breast tumor cells (Sarmiento et al. 2008). DRFs may also regulate tumor invasion through regulation of adherens junctions. Disruption of E-cadherin-based adherens junctions in cancer cells has been associated with enhancement of tumor development, invasion and metastasis (Gumbiner 2005). In MCF-7 breast cancer cells, DRF1 localizes to adherens junctions upon RhoA activation and siRNA silencing of DRF1 reduces E-cadherin-dependent adherens junctional integrity (Carramusa et al. 2007) suggesting that DRF1 plays a role in regulating tumor cell adherens junction formation as well as cell protrusion. Loss of adherens junction stability could enhance tumor invasion and metastasis by driving epithelial-to-mesenchymal transition (EMT) in tumor cells. Alternatively, loss of endothelial cell junctions would increase endothelial permeability and subsequent tumor cell extravasation and metastasis.

6.6 Nonmuscle Myosin II and Contractile Force Regulation During Tumor Cell Invasion The generation of contractile forces can play an important role in various phases of tumor cell migration. In conjunction with the actin cytoskeleton, nonmuscle myosin II is generally understood to generate contractile forces that contribute cell body translocation, deadhesion and retraction of the cell posterior during migration (Ridley et al. 2003). In addition to its role in contractile events, protrusive events at the leading edge of the cell lamella that occur during migration can also be linked to nonmuscle myosin II activity (Gupton et al. 2005). The involvement of nonmuscle myosin II was initially supported initially by morphological and behavioral responses of tumor cells to the potent non muscle myosin II inhibitor blebbistatin. Blebbistatin treatment significantly reduced spreading of human breast MDA-MB-231 carcinoma cells on fibronectin (Betapudi et al. 2006), invasion of 3-D collagen matrices by BE colorectal carcinoma cells (Wilkinson et al. 2005), and metastatic mammary carcinoma cells (Wyckoff et al. 2006). Isoform-specific deletion of nonmuscle myosin II by siRNA has revealed that both myosin IIA

154

M. Oser et al.

and IIB are both involved in migration of the metastatic human breast cancer cell MDA-MB-231 on fibronectin-coated surfaces and Boyden chambers (Betapudi et al. 2006). Interestingly, the myosin IIB isoform was found to play a greater role in lamellar protrusive events, indicative of isoform diversity during various stages of MDA-MB-231 cell motility. ROCKs (Rho-kinases), immediate downstream effectors of Rho GTPases, participate in regulating cytoskeletal signaling events critical to tumor cell motility (Riento and Ridley 2003). One important aspect of Rho/ROCK signaling is regulating myosin II activity by phosphorylation of the regulatory subunit of MLC phosphatase or MLC directly. Myosin light chain kinase (MLCK) can also activate MLC and increased MLCK mRNA expression is related to higher rates of non-small cell lung metastasis (Minamiya et al. 2005). ROCK protein expression is elevated in testicular and bladder cancers (Somlyo et al. 2000; Kamai et al. 2003) and deregulation of Rho/ROCK pathway has been implicated in tumor progression and, particularly, tumor metastasis. Expression of conditionally-activated ROCK in colon cancer cells resulted in aggressive dissemination of tumor cells into the surrounding stroma, demonstrating that enhanced ROCK signaling is important in promoting invasion from solid tumors (Croft et al. 2004). Inhibition of ROCK activity using the inhibitor Y-27632 has been used extensively to investigate the role of ROCK during in different aspects of tumor cell behavior. Y-27632 treatment is sufficient to abolish chemotactic migration in pancreatic cancer cell lines (Somlyo et al. 2000), suppress actomyosin activity, peritoneal invasion, and intrahepatic metastasis in rat and SCID mouse models (Genda et al. 1999; Itoh et al. 1999; Takamura et al. 2001). ROCK inhibition by Y-27632 and has been correlated with increased cell spreading on 2D surfaces in MDA-MB231 cells (Torka et al. 2006) and increased cell area in response to EGF stimulation in MTLn3 cells (El-Sibai et al. 2008). Despite this increase in cell spreading, ROCK inhibition led to a decrease in MTLn3 cell motility on 2D surfaces (El-Sibai et al. 2008), impaired deformation of model collagen matrices and decreased 3D invasion into these matrices (Wyckoff et al. 2006). Conversely, ROCK inhibition by Y-27632 accelerated motility of MDA-MB-231 cells on 2D surfaces while 3D invasion as measured using a Transwell assay was unaffected (Torka et al. 2006). Matrix degradation and adhesion was not affected by ROCK inhibition suggesting that myosin II-dependent force generation through ROCK is required for the invasive behavior of breast carcinoma cells. Since use of Y-27632 does not discriminate between the different isoforms of ROCK, recent studies have focused on detailing the role of ROCK isoforms (ROCK1 and ROCK2) in contractile force regulation by nonmuscle myosin II. Although both ROCK isoforms were identified as the downstream targets of RhoA (Leung et al. 1995; Matsui et al. 1996) evidence is mounting that ROCK1 and ROCK2 are differentially regulated and are not functionally redundant (Coleman et al. 2001; Yoneda et al. 2005). ROCK2 is linked with greater invasion and metastasis in several tumor models including bladder (Kamai et al. 2003), colon cancer (Vishnubhotla et al. 2007) and hepato-cellular carcinoma (HCC) (Wong et al. 2009). Experimental ROCK2 overexpression contributed to enhanced HCC motility and

6

Actin-based Motile Processes in Tumor Cell Invasion

155

invasion in vitro during wound healing and Transwell assays as well as in vivo. Knockdown of ROCK2 by shRNA was correlated with decreased motility, invasion, and reduced myosin II activation as measured by myosin phosphatase target subunit (MYPT1) phosphorylation. In addition, reduced filopodia and lamellipodia formation as well as diminished HCC directional movement was also observed in the absence of ROCK2 (Wong et al. 2009). ROCK2 knockdown by siRNA also decreased the invasion of several colon cancer cells lines in vivo (Vishnubhotla et al. 2007). Taken together, these studies suggest ROCK2 plays a significant role during tumor cell motility and invasion. Recent work has focused on the requirement for nonmuscle myosin II-generated contractile force during invadopodium maturation. In MCF10-CA1d breast cancer cells, blebbistatin inhibited matrix degradation activity of invadopodia (Alexander et al. 2008). While the fraction of mature invadopodia with matrix degradation activity was dramatically reduced in the presence of blebbistatin, smaller and more numerous actin- and cortactin-positive invadopodium precursors were observed, suggesting myosin II generated force is more important during invadopodium maturation rather than the initial formation of the precursor (Table 6.1). Similar reduction in the number of mature invadopodia with matrix degradation activity was also observed in the presence the ROCK inhibitor, Y-27632. In colon cancer cell models, ROCK2 has been localized to invadopodia and siRNA knockdown of ROCK2 led to a significant reduction in both MMP-2 and MMP-13 activity suggesting a regulatory role for ROCK2 in both invasion and matrix metalloprotease secretion at invadopodia (Vishnubhotla et al. 2007).

6.7 Coordination of Pathways Linking Actin Polymerization to the Matrix Degradation Activity of Invadopodia The proper regulation of the actin cytoskeleton is essential for maturation to a degradation-competent invadopodium. However, the mechanisms that link actin assembly at invadopodia to matrix metalloproteinase activity are poorly understood. As described earlier, in head and neck squamous carcinoma cells, cortactin has been shown to play a critical role in recruiting MMPs to invadopodia following invadopodium precursor formation (Clark et al. 2007). Cortactin tyrosine phosphorylation is important for actin polymerization in invadopodia at a stage that follows invadopodium precursor formation resulting in reduced matrix degradation activity (Oser et al. 2009). Based on these studies, one can hypothesize that actin polymerization in invadopodia, initiated by the tyrosine phosphorylation of cortactin, is required for stable MMP targeting to invadopodia. A recent study demonstrated that the exocyst complex components, Exo70 and Sec8, are critical for the formation and matrix degradation activity of invadopodia in MDA-MB-231 cells (Liu et al. 2009). The exocyst complex is involved in linking vesicles from the Golgi to the plasma membrane for exocytosis. The exo70 component of the exocyst complex directly interacts with the Arp2/3 complex and this

156

M. Oser et al.

interaction is important for cell migration (Zuo et al. 2006). Blocking the interaction of the exocyst complex with the Arp2/3 complex at invadopodia led to decreases in invadopodium formation, MMP secretion at invadopodia, and actin polymerization by the Arp2/3 complex. These findings shows that the components required for actin polymerization (in this case the Arp2/3 complex) directly interact with postGolgi vesicles trafficking MMPs to sites of invadopodia and these interactions are required for matrix degradation. In support of this, IQGAP1, a polarity protein that binds to the exocyst complex and is regulated by Cdc42 and RhoA, is important for the targeting of MMPs to invadopodia through its interactions with the exocyst complex (Sakurai-Yageta et al. 2008). Thus, the pathways involved in actin polymerization, such as Cdc42, also regulate MMP recruitment to invadopodia. In summary, Arp2/3 activity and the exocyst complex are coordinated directly through a binding interaction between Arp2/3 and exo70 (Liu et al. 2009) and indirectly through the common regulation of IQGAP1 and Arp2/3 by Cdc42 (Sakurai-Yageta et al. 2008) in invadopodia. Together, these findings provide strong evidence that the processes of actin polymerization and MMP delivery to invadopodia are precisely coordinated to allow for efficient ECM degradation by invasive tumor cells.

6.8 The Relative Contribution of Chemotaxis, Invadopodium Formation, and Lamellipodium Formation to the Invasive Tumor Cell Phenotype In Vivo We will close by discussing what is known about the contribution of pathways controlling tumor cell chemotaxis, invadopodium maturation, and lamellipodium formation to the invasive tumor cell phenotype in animal models. First we must ask, what are the common signaling pathways involved in regulating chemotaxis, lamellipodium formation, and invadopodium maturation and how do they differ? Although the upstream regulation of signaling pathways involved in actin polymerization appears to vary for each process, the output of each pathway leading to chemotaxis, lamellipodium formation, and actin polymerization during invadopodium maturation are similar. Signaling pathways that lead to activation of cofilin’s severing and nucleation activity, and dendritic nucleation by the Arp2/3 complex are required for tumor cell chemotaxis, invadopodium maturation, and the formation of lamellipodial protrusions. As we have discussed extensively, the current evidence suggests that the upstream regulation of both cofilin and Arp2/3 activation is different between invadopodia and lamellipodia. This implies that distinct signaling pathways are activated in spatially segregated subcellular compartments. It also suggests that the relative contribution of invadopodia compared to lamellipodia during tumor cell invasion in vivo can be determined by studying the effects of blocking subcellular compartment-specific mechanisms of cofilin or Arp2/3 activation. In vivo animal studies have demonstrated that blocking the output of these pathways can block tumor metastasis. Specifically, blocking cofilin inhibits invasion, intravasation, and metastasis of mammary carcinoma cells (Wang

6

Actin-based Motile Processes in Tumor Cell Invasion

157

et al. 2006). Inhibition of N-WASp, by shRNA or expression of dominant-negative N-WASp, also blocks intravasation and metastasis of mammary carcinoma cells (Wyckoff et al., In Preparation). Lastly, expression of a cortactin tyrosine phosphorylation mutant in mammary carcinoma cells leads to inhibition of bone metastasis (Li et al. 2001). Both cortactin tyrosine phosphorylation and N-WASp are essential for invadopodium maturation (Yamaguchi et al. 2005; Ayala et al. 2008; Oser et al. 2009), but are not required for lamellipodium formation (Bryce et al. 2005; Desmarais et al. 2009). In addition, Cdc42 is highly upregulated in the invasive subpopulation of mammary tumor cells (Wang et al. 2004) (Fig. 6.2) supporting the idea that invadopodium function correlates with the invasive tumor cell phenotype in vivo. Together, these studies provide strong evidence that blocking pathways that specifically regulate invadopodium maturation inhibits tumor cell invasion and metastasis in animal models. A recent report showed that CyfipY1/WAVE signaling pathways leading to Arp2/3 activation and lamellipodium formation may function to suppress, rather than promote, tumor cell invasion (Silva et al. 2009). CyfipY1 is commonly deleted in many human cancer types and downregulated during tumor cell invasion in vivo. Knockdown of CyfipY1 or WAVE proteins failed to form lamellipodia in 2D cultures and led to defects in cell-cell and cell-ECM adhesion. These changes in epithelial architecture resulted in increased tumor cell invasion in vivo. This study challenges the idea that the pathways leading to lamellipodium formation enhance tumor cell invasiveness and suggests that the formation of lamellipodia may be important for maintaining normal epithelium tissue architecture. However, a previous study (Kurisu et al. 2005) reported that knockdown of WAVE2 leads to inhibition of melanoma cell invasion and experimental metastasis concluding that WAVE2 is important during tumor cell invasion. Thus, more work is needed to make a definitive conclusion regarding the role of lamellipodia during tumor cell invasion. Measuring the activity of the Cdc42/N/WASp pathway in CyfipY1 knockdown cells may provide useful insights as to whether the balance between WAVEand N-WASp-mediated Arp2/3 activation is shifted towards N-WASp to promote tumor cell invasion. It is interesting to consider the possibility that the pathways that activate WAVE-dependent Arp2/3 activation (e.g. CyfipY1), as in lamellipodia, suppress tumor cell invasion, while pathways that activate N-WASp-dependent Arp2/3 activation, as in invadopodia, enhance tumor cell invasion. Alternatively, inhibition of WAVE2 and N-WASp leads to increased carcinoma cell lamellipodium protrusion in response to EGF involving DRF1-dependent actin polymerization (Sarmiento et al. 2008). These results suggest that the interaction of multiple pathways all leading to actin polymerization are coordinated to produce the invasive phenotype. Therefore, measurement of cell protrusion and motility behavior both in vitro and in vivo in response to each of these experimental manipulations is necessary to understand how they affect the invasive and metastatic phenotype. In conclusion, the emergence of the invadopodium as a subcellular structure critical for tumor cell invasion has allowed researchers to gain many novel insights into the actin-based mechanisms involved in tumor cell migration and chemotaxis. From our discussion on tumor cell chemotaxis, invadopodium maturation,

158

M. Oser et al.

and lamellipodium formation, it is clear that the invasive tumor cell must integrate extracellular signals from the microenvironment and coordinate chemotaxis, cell migration, and ECM degradation. During tumor cell invasion, the precise coordination of these actin-based processes can lead to tumor cell metastasis. Future studies must focus on how the various pathways leading to cell motility and invasion are coordinated, and how invadopodia, in particular, integrate both extracellular and intracellular signals to coordinate tumor cell chemotaxis, migration, and degradation. Acknowledgements This work was funded by CA113395, CA126511, CA150344 (JC).

References Abram, C. L., D. F. Seals, et al. (2003). “The adaptor protein fish associates with members of the ADAMs family and localizes to podosomes of Src-transformed cells.” J Biol Chem 278(19): 16844–51. Albiges-Rizo, C., O. Destaing, et al. (2009). “Actin machinery and mechanosensitivity in invadopodia, podosomes and focal adhesions.” J Cell Sci 122(Pt 17): 3037–49. Alexander, N. R., K. M. Branch, et al. (2008). “Extracellular matrix rigidity promotes invadopodia activity.” Curr Biol 18(17): 1295–9. Ammer, A. G. and S. A. Weed (2008). “Cortactin branches out: roles in regulating protrusive actin dynamics.” Cell Motil Cytoskeleton 65(9): 687–707. Arber, S., F. A. Barbayannis, et al. (1998). “Regulation of actin dynamics through phosphorylation of cofilin by LIM-kinase.” Nature 393(6687): 805–9. Artym, V. V., K. M. Yamada, et al. (2009). “ECM degradation assays for analyzing local cell invasion.” Methods Mol Biol 522: 211–9. Artym, V. V., Y. Zhang, et al. (2006). “Dynamic interactions of cortactin and membrane type 1 matrix metalloproteinase at invadopodia: defining the stages of invadopodia formation and function.” Cancer Res 66(6): 3034–43. Ayala, I., M. Baldassarre, et al. (2008). “Multiple regulatory inputs converge on cortactin to control invadopodia biogenesis and extracellular matrix degradation.” J Cell Sci 121(Pt 3): 369–78. Ayala, I., G. Giacchetti, et al. (2009). “Faciogenital dysplasia protein Fgd1 regulates invadopodia biogenesis and extracellular matrix degradation and is up-regulated in prostate and breast cancer.” Cancer Res 69(3): 747–52. Bailly, M., L. Yan, et al. (1998). “Regulation of protrusion shape and adhesion to the substratum during chemotactic responses of mammalian carcinoma cells.” Exp Cell Res 241(2): 285–99. Baldassarre, M., A. Pompeo, et al. (2003). “Dynamin participates in focal extracellular matrix degradation by invasive cells.” Mol Biol Cell 14(3): 1074–84. Bamburg, J. R. and O. P. Wiggan (2002). “ADF/cofilin and actin dynamics in disease.” Trends Cell Biol 12(12): 598–605. Betapudi, V., L. S. Licate, et al. (2006). “Distinct roles of nonmuscle myosin II isoforms in the regulation of MDA-MB-231 breast cancer cell spreading and migration.” Cancer Res 66(9): 4725–33. Blouw, B., et al. (2008). “A role for the podosome/invadopodia scaffold protein Tks5 in tumor growth in vivo.” Eur J Cell Biol 87(8–9): 555–67. Bowden, E. T., E. Onikoyi, et al. (2006). “Co-localization of cortactin and phosphotyrosine identifies active invadopodia in human breast cancer cells.” Exp Cell Res 312(8): 1240–53. Boyle, S. N., G. A. Michaud, et al. (2007). “A critical role for cortactin phosphorylation by Ablfamily kinases in PDGF-induced dorsal-wave formation.” Curr Biol 17(5): 445–51.

6

Actin-based Motile Processes in Tumor Cell Invasion

159

Bryce, N. S., E. S. Clark, et al. (2005). “Cortactin promotes cell motility by enhancing lamellipodial persistence.” Curr Biol 15(14): 1276–85. Buccione, R., G. Caldieri, et al. (2009). “Invadopodia: specialized tumor cell structures for the focal degradation of the extracellular matrix.” Cancer Metastasis Rev 28(1–2): 137–49. Caldieri, G., I. Ayala, et al. (2009). “Cell and molecular biology of invadopodia.” Int Rev Cell Mol Biol 275: 1–34. Cardone, R. A., V. Casavola, et al. (2005). “The role of disturbed pH dynamics and the Na+/H+ exchanger in metastasis.” Nat Rev Cancer 5(10): 786–95. Carlier, M. F., V. Laurent, et al. (1997). “Actin depolymerizing factor (ADF/cofilin) enhances the rate of filament turnover: implication in actin-based motility.” J Cell Biol 136(6): 1307–22. Carman, C. V., P. T. Sage, et al. (2007). “Transcellular diapedesis is initiated by invasive podosomes.” Immunity 26(6): 784–97. Carramusa, L., C. Ballestrem, et al. (2007). “Mammalian diaphanous-related formin Dia1 controls the organization of E-cadherin-mediated cell-cell junctions.” J Cell Sci 120(Pt 21): 3870–82. Chan, A. Y., M. Bailly, et al. (2000). “Role of cofilin in epidermal growth factor-stimulated actin polymerization and lamellipod protrusion.” J Cell Biol 148(3): 531–42. Chan, K. T., C. L. Cortesio, et al. (2009). “FAK alters invadopodia and focal adhesion composition and dynamics to regulate breast cancer invasion.” J Cell Biol 185(2): 357–70. Chuma, M., et al. (2004). “Overexpression of cortactin is involved in motility and metastasis of hepatocellular carcinoma.” J Hepatol 41(4): 629–36. Clark, E. S. and A. M. Weaver (2008). “A new role for cortactin in invadopodia: regulation of protease secretion.” Eur J Cell Biol 87(8–9): 581–90. Clark, E. S., A. S. Whigham, et al. (2007). “Cortactin is an essential regulator of matrix metalloproteinase secretion and extracellular matrix degradation in invadopodia.” Cancer Res 67(9): 4227–35. Coleman, M. L., E. A. Sahai, et al. (2001). “Membrane blebbing during apoptosis results from caspase-mediated activation of ROCK I.” Nat Cell Biol 3(4): 339–45. Condeelis, J., J. Jones, et al. (1992). “Chemotaxis of metastatic tumor cells: clues to mechanisms from the Dictyostelium paradigm.” Cancer Metastasis Rev 11(1): 55–68. Condeelis, J. and J. E. Segall (2003). “Intravital imaging of cell movement in tumours.” Nat Rev Cancer 3(12): 921–30. Condeelis, J. S., J. B. Wyckoff, et al. (2001). “Lamellipodia in invasion.” Semin Cancer Biol 11(2): 119–28. Copeland, S. J., B. J. Green, et al. (2007). “The diaphanous inhibitory domain/diaphanous autoregulatory domain interaction is able to mediate heterodimerization between mDia1 and mDia2.” J Biol Chem 282(41): 30120–30. Cortesio, C. L., K. T. Chan, et al. (2008). “Calpain 2 and PTP1B function in a novel pathway with Src to regulate invadopodia dynamics and breast cancer cell invasion.” J Cell Biol 180(5): 957–71. Crimaldi, L., S. A. Courtneidge, et al. (2009). “Tks5 recruits AFAP-110, p190RhoGAP, and cortactin for podosome formation.” Exp Cell Res 315(15): 2581–92. Croft, D. R., E. Sahai, et al. (2004). “Conditional ROCK activation in vivo induces tumor cell dissemination and angiogenesis.” Cancer Res 64(24): 8994–9001. Deryugina, E. I., B. Ratnikov, et al. (2001). “MT1-MMP initiates activation of pro-MMP-2 and integrin alphavbeta3 promotes maturation of MMP-2 in breast carcinoma cells.” Exp Cell Res 263(2): 209–23. DesMarais, V., M. Ghosh, et al. (2005). “Cofilin takes the lead.” J Cell Sci 118(Pt 1): 19–26. DesMarais, V., F. Macaluso, et al. (2004). “Synergistic interaction between the Arp2/3 complex and cofilin drives stimulated lamellipod extension.” J Cell Sci 117(Pt 16): 3499–510. Desmarais, V., H. Yamaguchi, et al. (2009). “N-WASP and cortactin are involved in invadopodiumdependent chemotaxis to EGF in breast tumor cells.” Cell Motil Cytoskeleton 66(6): 303–16. Diaz, B., et al. (2009). “Tks5-dependent, nox-mediated generation of reactive oxygen species is necessary for invadopodia formation.” Sci Signal 2(88): ra53.

160

M. Oser et al.

Di Vizio, D., J. Kim, et al. (2009). “Oncosome formation in prostate cancer: association with a region of frequent chromosomal deletion in metastatic disease.” Cancer Res 69(13): 5601–9. Eiseler, T., H. Doppler, et al. (2009). “Protein kinase D1 regulates cofilin-mediated F-actin reorganization and cell motility through slingshot.” Nat Cell Biol 11(5): 545–56. El-Sibai, M., O. Pertz, et al. (2008). “RhoA/ROCK-mediated switching between Cdc42- and Rac1dependent protrusion in MTLn3 carcinoma cells.” Exp Cell Res 314(7): 1540–52. Faix, J. and R. Grosse (2006). “Staying in shape with formins.” Dev Cell 10(6): 693–706. Frantz, C., G. Barreiro, et al. (2008). “Cofilin is a pH sensor for actin free barbed end formation: role of phosphoinositide binding.” J Cell Biol 183(5): 865–79. Friedl, P. and D. Gilmour (2009). “Collective cell migration in morphogenesis, regeneration and cancer.” Nat Rev Mol Cell Biol 10(7): 445–57. Genda, T., M. Sakamoto, et al. (1999). “Cell motility mediated by rho and Rho-associated protein kinase plays a critical role in intrahepatic metastasis of human hepatocellular carcinoma.” Hepatology 30(4): 1027–36. Ghosh, M., X. Song, et al. (2004). “Cofilin promotes actin polymerization and defines the direction of cell motility.” Science 304(5671): 743–6. Gomez, T. S., K. Kumar, et al. (2007). “Formins regulate the actin-related protein 2/3 complexindependent polarization of the centrosome to the immunological synapse.” Immunity 26(2): 177–90. Gorbatyuk, V. Y., N. J. Nosworthy, et al. (2006). “Mapping the phosphoinositide-binding site on chick cofilin explains how PIP2 regulates the cofilin-actin interaction.” Mol Cell 24(4): 511–22. Gumbiner, B. M. (2005). “Regulation of cadherin-mediated adhesion in morphogenesis.” Nat Rev Mol Cell Biol 6(8): 622–34. Gunnersen, J.M., et al. (2000). “Growth and migration markers of rat C6 glioma cells identified by serial analysis of gene expression.” Glia 32(2): 146–54. Gupton, S. L., K. L. Anderson, et al. (2005). “Cell migration without a lamellipodium: translation of actin dynamics into cell movement mediated by tropomyosin.” J Cell Biol 168(4): 619–31. Hernandez, L., T. Smirnova, et al. (2009). “The EGF/CSF-1 paracrine invasion loop can be triggered by heregulin beta1 and CXCL12.” Cancer Res 69(7): 3221–7. Higgs, H. N. (2005). “Formin proteins: a domain-based approach.” Trends Biochem Sci 30(6): 342–53. Hotulainen, P. and P. Lappalainen (2006). “Stress fibers are generated by two distinct actin assembly mechanisms in motile cells.” J Cell Biol 173(3): 383–94. Huang, C., N. N. Tandon, et al. (1997). “Proteolysis of platelet cortactin by calpain.” J Biol Chem 272(31): 19248–52. Huang, J., et al. (2003) “Cooperative roles of Fyn and cortactin in cell migration of metastatic murine melanoma.” J Biol Chem 278(48): 48367–76. Hui, R., et al. (1997) “EMS1 amplification can occur independently of CCND1 or INT-2 amplification at 11q13 and may identify different phenotypes in primary breast cancer”. Oncogene 15(13): 1617–23. Ichetovkin, I., W. Grant, et al. (2002). “Cofilin produces newly polymerized actin filaments that are preferred for dendritic nucleation by the Arp2/3 complex.” Curr Biol 12(1): 79–84. Ichetovkin, I., J. Han, et al. (2000). “Actin filaments are severed by both native and recombinant dictyostelium cofilin but to different extents.” Cell Motil Cytoskeleton 45(4): 293–306. Itoh, K., K. Yoshioka, et al. (1999). “An essential part for Rho-associated kinase in the transcellular invasion of tumor cells.” Nat Med 5(2): 221–5. Jadeski, L., et al. (2008). “IQGAP1 stimulates proliferation and enhances tumorigenesis of human breast epithelial cells.” J Biol Chem 283(2): 1008–17. Kamai, T., T. Tsujii, et al. (2003). “Significant association of Rho/ROCK pathway with invasion and metastasis of bladder cancer.” Clin Cancer Res 9(7): 2632–41. Kay, R. R., P. Langridge, et al. (2008). “Changing directions in the study of chemotaxis.” Nat Rev Mol Cell Biol 9(6): 455–63.

6

Actin-based Motile Processes in Tumor Cell Invasion

161

Kempiak, S. J., H. Yamaguchi, et al. (2005). “A neural Wiskott-Aldrich Syndrome proteinmediated pathway for localized activation of actin polymerization that is regulated by cortactin.” J Biol Chem 280(7): 5836–42. Kinley, A. W., S. A. Weed, et al. (2003). “Cortactin interacts with WIP in regulating Arp2/3 activation and membrane protrusion.” Curr Biol 13(5): 384–93. Kitzing, T. M., A. S. Sahadevan, et al. (2007). “Positive feedback between Dia1, LARG, and RhoA regulates cell morphology and invasion.” Genes Dev 21(12): 1478–83. Koka, S., C. L. Neudauer, et al. (2003). “The formin-homology-domain-containing protein FHOD1 enhances cell migration.” J Cell Sci 116(Pt 9): 1745–55. Kowalski, J. R., C. Egile, et al. (2005). “Cortactin regulates cell migration through activation of N-WASP.” J Cell Sci 118(Pt 1): 79–87. Kurisu, S., S. Suetsugu, et al. (2005). “Rac-WAVE2 signaling is involved in the invasive and metastatic phenotypes of murine melanoma cells.” Oncogene 24(8): 1309–19. Kurisu, S. and T. Takenawa (2009). “The WASP and WAVE family proteins.” Genome Biol 10(6): 226. Leung, T., E. Manser, et al. (1995). “A novel serine/threonine kinase binding the Ras-related RhoA GTPase which translocates the kinase to peripheral membranes.” J Biol Chem 270(49): 29051–4. Leyman, S., M. Sidani, et al. (2009). “Unbalancing the PI(4, 5)P2-Cofilin Interaction Impairs Cell Steering.” Mol Biol Cell 20: 4509–4523. Li, Y., M. Tondravi, et al. (2001). “Cortactin potentiates bone metastasis of breast cancer cells.” Cancer Res 61(18): 6906–11. Linder, S. (2009). “Invadosomes at a glance.” J Cell Sci 122(Pt 17): 3009–13. Liu, J., P. Yue, et al. (2009). “The Role of the Exocyst in MMP Secretion and Actin Dynamics during Tumor Cell Invadopodia Formation.” Mol Biol Cell 20: 3763–3771. Lizarraga, F., R. Poincloux, et al. (2009). “Diaphanous-related formins are required for invadopodia formation and invasion of breast tumor cells.” Cancer Res 69(7): 2792–800. Luo, C., H. Pan, et al. (2006). “CXCL12 induces tyrosine phosphorylation of cortactin, which plays a role in CXC chemokine receptor 4-mediated extracellular signal-regulated kinase activation and chemotaxis.” J Biol Chem 281(40): 30081–93. Machesky, L. M. (2008). “Lamellipodia and filopodia in metastasis and invasion.” FEBS Lett 582(14): 2102–11. Marone, R., D. Hess, et al. (2004). “Memo mediates ErbB2-driven cell motility.” Nat Cell Biol 6(6): 515–22. Martinez-Quiles, N., H. Y. Ho, et al. (2004). “Erk/Src phosphorylation of cortactin acts as a switch on-switch off mechanism that controls its ability to activate N-WASP.” Mol Cell Biol 24(12): 5269–80. Matsui, T., M. Amano, et al. (1996). “Rho-associated kinase, a novel serine/threonine kinase, as a putative target for small GTP binding protein Rho.” EMBO J 15(9): 2208–16. Meira, M., R. Masson, et al. (2009). “Memo is a cofilin-interacting protein that influences PLCgamma1 and cofilin activities, and is essential for maintaining directionality during ErbB2-induced tumor-cell migration.” J Cell Sci 122(Pt 6): 787–97. Minamiya, Y., T. Nakagawa, et al. (2005). “Increased expression of myosin light chain kinase mRNA is related to metastasis in non-small cell lung cancer.” Tumour Biol 26(3): 153–7. Miyoshi, T., et al. (2005) “Possible mechanism of metastasis in lung adenocarcinomas with a micropapillary pattern.” Pathol Int 55(7): 419–24. Mooren, O. L., T. I. Kotova, et al. (2009). “Dynamin2 GTPase and cortactin remodel actin filaments.” J Biol Chem 284: 23995–24005. Mouneimne, G., V. DesMarais, et al. (2006). “Spatial and temporal control of cofilin activity is required for directional sensing during chemotaxis.” Curr Biol 16(22): 2193–205. Mouneimne, G., L. Soon, et al. (2004). “Phospholipase C and cofilin are required for carcinoma cell directionality in response to EGF stimulation.” J Cell Biol 166(5): 697–708.

162

M. Oser et al.

Oikawa, T., T. Itoh, et al. (2008). “Sequential signals toward podosome formation in NIH-src cells.” J Cell Biol 182(1): 157–69. Olson, M. F. and E. Sahai (2009). “The actin cytoskeleton in cancer cell motility.” Clin Exp Metastasis 26(4): 273–87. Oser, M., H. Yamaguchi, et al. (2009). “Cortactin regulates cofilin and N-WASp activities to control the stages of invadopodium assembly and maturation.” J Cell Biol 186(4): 571–87. Oser, M. and J. Condeelis (2009). “The cofilin activity cycle in lamellipodia and invadopodia.” J Cell Biochem 108: 1252–1262. Otsubo, T., et al. (2004). “Involvement of Arp2/3 complex in the process of colorectal carcinogenesis.” Mod Pathol 17(4): 461–7. Packard, B. Z., V. V. Artym, et al. (2009). “Direct visualization of protease activity on cells migrating in three-dimensions.” Matrix Biol 28(1): 3–10. Pantel, K., R. H. Brakenhoff, et al. (2008). “Detection, clinical relevance and specific biological properties of disseminating tumour cells.” Nat Rev Cancer 8(5): 329–40. Perrin, B. J., K. J. Amann, et al. (2006). “Proteolysis of cortactin by calpain regulates membrane protrusion during cell migration.” Mol Biol Cell 17(1): 239–50. Philippar, U., E. T. Roussos, et al. (2008). “A Mena invasion isoform potentiates EGF-induced carcinoma cell invasion and metastasis.” Dev Cell 15(6): 813–28. Poincloux, R., F. Lizarraga, et al. (2009). “Matrix invasion by tumour cells: a focus on MT1-MMP trafficking to invadopodia.” J Cell Sci 122(Pt 17): 3015–24. Pollard, T. D. (2007). “Regulation of actin filament assembly by Arp2/3 complex and formins.” Annu Rev Biophys Biomol Struct 36: 451–77. Pollard, T. D. and G. G. Borisy (2003). “Cellular motility driven by assembly and disassembly of actin filaments.” Cell 112(4): 453–65. Riento, K. and A. J. Ridley (2003). “Rocks: multifunctional kinases in cell behaviour.” Nat Rev Mol Cell Biol 4(6): 446–56. Ridley A. J., M. A. Schwartz, K. Burridge, R. A. Firtel, M. H. Ginsberg, G. Borisy, J. T. Parsons, A. R. Horwitz (2003) Cell migration: integrating signals from front to back. Science 302: 1704–709. Robinson, B. D., G. L. Sica, et al. (2009). “Tumor microenvironment of metastasis in human breast carcinoma: a potential prognostic marker linked to hematogenous dissemination.” Clin Cancer Res 15(7): 2433–41. Rodrigo, J. P., et al. (2000) “EMS1 gene amplification correlates with poor prognosis in squamous cell carcinomas of the head and neck.” Clin Cancer Res 6(8): 3177–82. Rohatgi, R., P. Nollau, et al. (2001). “Nck and phosphatidylinositol 4,5-bisphosphate synergistically activate actin polymerization through the N-WASP-Arp2/3 pathway.” J Biol Chem 276(28): 26448–52. Sabeh, F., R. Shimizu-Hirota, et al. (2009). “Protease-dependent versus -independent cancer cell invasion programs: three-dimensional amoeboid movement revisited.” J Cell Biol 185(1): 11–9. Sahai, E. (2005). “Mechanisms of cancer cell invasion.” Curr Opin Genet Dev 15(1): 87–96. Sakurai-Yageta, M., C. Recchi, et al. (2008). “The interaction of IQGAP1 with the exocyst complex is required for tumor cell invasion downstream of Cdc42 and RhoA.” J Cell Biol 181(6): 985–98. Sarmiento, C., W. Wang, et al. (2008). “WASP family members and formin proteins coordinate regulation of cell protrusions in carcinoma cells.” J Cell Biol 180(6): 1245–60. Schafer, D. A., S. A. Weed, et al. (2002). “Dynamin2 and cortactin regulate actin assembly and filament organization.” Curr Biol 12(21): 1852–7. Schirenbeck, A., T. Bretschneider, et al. (2005). “The Diaphanous-related formin dDia2 is required for the formation and maintenance of filopodia.” Nat Cell Biol 7(6): 619–25. Scott, R. W. and M. F. Olson (2007). “LIM kinases: function, regulation and association with human disease.” J Mol Med 85(6): 555–68. Seals, D.F., et al. (2005). “The adaptor protein Tks5/Fish is required for podosome formation and function, and for the protease-driven invasion of cancer cells.” Cancer Cell 7(2): 155–65.

6

Actin-based Motile Processes in Tumor Cell Invasion

163

Semba, S., et al. (2006). “Coexpression of actin-related protein 2 and Wiskott-Aldrich syndrome family verproline-homologous protein 2 in adenocarcinoma of the lung.” Clin Cancer Res 12(8): 2449–54. Sidani, M., D. Wessels, et al. (2007). “Cofilin determines the migration behavior and turning frequency of metastatic cancer cells.” J Cell Biol 179(4): 777–91. Silva, J. M., E. Ezhkova, et al. (2009). “Cyfip1 is a putative invasion suppressor in epithelial cancers.” Cell 137(6): 1047–61. Somlyo, A. V., D. Bradshaw, et al. (2000). “Rho-kinase inhibitor retards migration and in vivo dissemination of human prostate cancer cells.” Biochem Biophys Res Commun 269(3): 652–9. Song, X., X. Chen, et al. (2006). “Initiation of cofilin activity in response to EGF is uncoupled from cofilin phosphorylation and dephosphorylation in carcinoma cells.” J Cell Sci 119(Pt 14): 2871–81. Soon, L. L. (2007). “A discourse on cancer cell chemotaxis: where to from here?” IUBMB Life 59(2): 60–7. Stylli, S. S., S. T. I, et al. (2009). “Nck adaptor proteins link Tks5 to invadopodia actin regulation and ECM degradation.” J Cell Sci 122: 2727–2740. Stylli, S. S., T. T. Stacey, et al. (2009). “Nck adaptor proteins link Tks5 to invadopodia actin regulation and ECM degradation.” J Cell Sci 122(Pt 15): 2727–40. Takamura, M., M. Sakamoto, et al. (2001). “Inhibition of intrahepatic metastasis of human hepatocellular carcinoma by Rho-associated protein kinase inhibitor Y-27632.” Hepatology 33(3): 577–81. Tanaka, S., M. Kunii, et al. (2009). “Generation of cortactin floxed mice and cellular analysis of motility in fibroblasts.” Genesis 47: 638–646. Tehrani, S., N. Tomasevic, et al. (2007). “Src phosphorylation of cortactin enhances actin assembly.” Proc Natl Acad Sci U S A 104(29): 11933–8. Torka, R., F. Thuma, et al. (2006). “ROCK signaling mediates the adoption of different modes of migration and invasion in human mammary epithelial tumor cells.” Exp Cell Res 312(19): 3857–71. Uruno, T., J. Liu, et al. (2001). “Activation of Arp2/3 complex-mediated actin polymerization by cortactin.” Nat Cell Biol 3(3): 259–66. van Rheenen, J., X. Song, et al. (2007). “EGF-induced PIP2 hydrolysis releases and activates cofilin locally in carcinoma cells.” J Cell Biol 179(6): 1247–59. Van Troys, M., L. Huyck, et al. (2008). “Ins and outs of ADF/cofilin activity and regulation.” Eur J Cell Biol 87(8–9): 649–67. Vishnubhotla, R., S. Sun, et al. (2007). “ROCK-II mediates colon cancer invasion via regulation of MMP-2 and MMP-13 at the site of invadopodia as revealed by multiphoton imaging.” Lab Invest 87(11): 1149–58. Wang, W., R. Eddy, et al. (2007a). “The cofilin pathway in breast cancer invasion and metastasis.” Nat Rev Cancer 7(6): 429–40. Wang, W., S. Goswami, et al. (2004). “Identification and testing of a gene expression signature of invasive carcinoma cells within primary mammary tumors.” Cancer Res 64(23): 8585–94. Wang, W., S. Goswami, et al. (2005). “Tumor cells caught in the act of invading: their strategy for enhanced cell motility.” Trends Cell Biol 15(3): 138–45. Wang, W., G. Mouneimne, et al. (2006). “The activity status of cofilin is directly related to invasion, intravasation, and metastasis of mammary tumors.” J Cell Biol 173(3): 395–404. Wang, W., J. B. Wyckoff, et al. (2007b). “Coordinated regulation of pathways for enhanced cell motility and chemotaxis is conserved in rat and mouse mammary tumors.” Cancer Res 67(8): 3505–11. Weaver, A. M. (2008). “Cortactin in tumor invasiveness.” Cancer Lett 265(2): 157–66. Weaver, A. M., A. V. Karginov, et al. (2001). “Cortactin promotes and stabilizes Arp2/3-induced actin filament network formation.” Curr Biol 11(5): 370–4. Weaver, A. M., M. E. Young, et al. (2003). “Integration of signals to the Arp2/3 complex.” Curr Opin Cell Biol 15(1): 23–30.

164

M. Oser et al.

Weed, S. A., A. V. Karginov, et al. (2000). “Cortactin localization to sites of actin assembly in lamellipodia requires interactions with F-actin and the Arp2/3 complex.” J Cell Biol 151(1): 29–40. Wilkinson, S., H. F. Paterson, et al. (2005). “Cdc42-MRCK and Rho-ROCK signalling cooperate in myosin phosphorylation and cell invasion.” Nat Cell Biol 7(3): 255–61. Wong, C. C., C. M. Wong, et al. (2009). “Rho-kinase 2 is frequently overexpressed in hepatocellular carcinoma and involved in tumor invasion.” Hepatology 49(5): 1583–94. Wu, H. and J. T. Parsons (1993). “Cortactin, an 80/85-kilodalton pp60src substrate, is a filamentous actin-binding protein enriched in the cell cortex.” J Cell Biol 120(6): 1417–26. Wyckoff, J., W. Wang, et al. (2004). “A paracrine loop between tumor cells and macrophages is required for tumor cell migration in mammary tumors.” Cancer Res 64(19): 7022–9. Wyckoff, J. B., J. G. Jones, et al. (2000a). “A critical step in metastasis: in vivo analysis of intravasation at the primary tumor.” Cancer Res 60(9): 2504–11. Wyckoff, J. B., S. E. Pinner, et al. (2006). “ROCK- and myosin-dependent matrix deformation enables protease-independent tumor-cell invasion in vivo.” Curr Biol 16(15): 1515–23. Wyckoff, J. B., J. E. Segall, et al. (2000b). “The collection of the motile population of cells from a living tumor.” Cancer Res 60(19): 5401–4. Wyckoff, J. B., Y. Wang, et al. (2007). “Direct visualization of macrophage-assisted tumor cell intravasation in mammary tumors.” Cancer Res 67(6): 2649–56. Yamaguchi, H. and J. Condeelis (2007). “Regulation of the actin cytoskeleton in cancer cell migration and invasion.” Biochim Biophys Acta 1773(5): 642–52. Yamaguchi, H., M. Lorenz, et al. (2005). “Molecular mechanisms of invadopodium formation: the role of the N-WASP-Arp2/3 complex pathway and cofilin.” J Cell Biol 168(3): 441–52. Yanagawa, R., et al. (2001). “Genome-wide screening of genes showing altered expression in liver metastases of human colorectal cancers by cDNA microarray.” Neoplasia 3(5): 395–401. Yoneda, A., H. A. Multhaupt, et al. (2005). “The Rho kinases I and II regulate different aspects of myosin II activity.” J Cell Biol 170(3): 443–53. Zhang, L.H., et al. (2006). “Dominant expression of 85-kDa form of cortactin in colorectal cancer.” J Cancer Res Clin Oncol 132(2): 113–20. Zhang, S., et al. (2009). “RhoA regulates G1-S progression of gastric cancer cells by modulation of multiple INK4 family tumor suppressors.” Mol Cancer Res 7(4): 570–80. Zhu, J., D. Yu, et al. (2007). “Receptor-mediated endocytosis involves tyrosine phosphorylation of cortactin.” J Biol Chem 282(22): 16086–94. Zhu, X. L., L. Liang, et al. (2008). “Overexpression of FMNL2 is closely related to metastasis of colorectal cancer.” Int J Colorectal Dis 23(11): 1041–7. Zuo, X., J. Zhang, et al. (2006). “Exo70 interacts with the Arp2/3 complex and regulates cell migration.” Nat Cell Biol 8(12): 1383–8.

Chapter 7

Actin-based Chromosome Movements in Cell Division Rong Li

Abstract Although microtubules are well-studied players moving chromosomes during mitosis and meiosis, recent work in mammalian oocytes has revealed intricate interactions between actin and chromosomes that directly control the positioning and extrusion of chromosomes during asymmetric meiotic cell divisions. New evidence also suggests that actin and actin-based motor proteins play interesting roles in the assembly and orientation of both meiotic and mitotic spindles. In this chapter we review what is known to date in this emerging area of actin-based motility and discuss outstanding questions and key mechanistic issues for future study.

Contents 7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2 Actin-Driven Chromosomal Movement During Meiosis I in Mouse Oocytes . . . . 7.2.1 Meiotic Chromosome Migration as a Key Step in Cellular Symmetry Breaking . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2.2 Early Evidence That Actin Powers Meiotic Chromosome Migration in Mouse Oocytes . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2.3 Pushing or Pulling – How Does Actin Move Chromosomes? . . . . . . . . 7.2.4 Questions Ahead . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.3 Interactions Between Actin and Chromatin to Establish the Site for Polar Body Extrusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.3.1 Chromatin-Induced Assembly of Cortical Actin and Myosin-II for Polar Body Extrusion . . . . . . . . . . . . . . . . . 7.3.2 The Ran-GTP Gradient: the Oocyte’s Tape Measure for Polar Body Extrusion? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.3.3 How Might Ran Regulate the Assembly of Cortical Actin and Myosin? . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

166 167 167 169 170 172 175 176 176 177

R. Li (B) Stowers Institute for Medical Research, 1000 East 50th street, Kansas City, MO 64110, USA e-mail: [email protected]

M.-F. Carlier (ed.), Actin-based Motility, DOI 10.1007/978-90-481-9301-1_7, C Springer Science+Business Media B.V. 2010

165

166 7.4 Does Actin Have a Hand in the Spindle? . . . . . . . . . . . . . . . . . . . . 7.5 Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

R. Li 179 181 182

7.1 Introduction Chromosomes exhibit a wide range of motions in both dividing and quiescent cells. The best known examples are the movement of chromosomes during the formation of a mitotic spindle and subsequent segregation of duplicated sister chromosomes to opposite poles of a dividing cell. These movements have been studied extensively and have contributed, to a large extent, to our knowledge about the forces and the molecular machinery that move chromosomes. In the spindle, chromosomes move mostly through two mechanisms: sliding along microtubules, riding on kinesin and dynein family motor proteins; moving with microtubule tips as a result of microtubule polymerization or depolymerization coupled to the action of motor proteins either associated with the microtubule organizing centers at the spindle poles or with the kinetochore, a highly complex structure assembled around the centromere region of each chromosome (McIntosh et al. 2002; Walczak and Heald 2008). Actin has mostly stayed on the sideline in chromosome movement, probably because disruption of actin has not been found to cause major defects in spindle assembly and anaphase progression in well studied models. In popular systems of actin-based motility, such as cell movement, actin generates protrusive (or pushing) forces through polymerization at their barbed ends, as opposed to pulling by depolymerization. Furthermore, most of the known regulators of actin polymerization are associated with the membrane, especially the plasma membrane. For example, Cdc42 and Rac, two members of the Rho family GTPases, localize primarily on the cell cortex where they activate effector proteins, most notably the Wiskott Aldrich Syndrome Protein (WASP) family proteins that control the nucleation of actin filaments by the Arp2/3 complex (Takenawa and Suetsugu 2007). Extension of Arp2/3-nucleated dendritic actin network is thought to provide the force that powers lamellipodia protrusion at the leading edge of motile cells (Pollard and Borisy 2003). The small GTPase Rho, which also localizes to the plasma membrane, stimulates assembly of cortical contractile networks of actin and type II myosin through several downstream effectors (Piekny et al. 2005). As such, actin is often linked to dynamics of the cell cortex or motions that associate with membrane-based structures, such as endocytic vesicles (Carlier and Pantaloni 2007; Galletta and Cooper 2009). That is why recent studies on direct involvements of actin in chromosome motion during meiotic cell division have attracted much attention and has, to some degree, brought actin to a more center stage in moving and positioning chromosomes. In this chapter, we will examine recent findings from experiments mostly in meiotic systems and discuss the possible mechanism by which actin interacts with the chromosomes and influence their positioning and segregation. As we are far from any

7

Actin-based Chromosome Movements in Cell Division

167

clear mechanistic explanation of actin’s role in chromosome motility, the focus of this chapter is not to draw any definitive conclusion, but instead to highlight important questions for future investigation. Even more enigmatic is perhaps actin’s involvement in the dynamics and function of interphase chromosomes, which has been discussed in several excellent review articles (e.g. (Kumaran et al. 2008; Louvet and Percipalle 2009)) and will not be a topic for discussion in this chapter.

7.2 Actin-Driven Chromosomal Movement During Meiosis I in Mouse Oocytes The most convincing evidence that actin, but not microtubules, plays a direct role in moving chromosomes has come from studies of mouse oocytes undergoing meiotic divisions. Meiotic maturation of oocytes is a dynamic sequence of events that lead to successive extrusion of two polar bodies during meiosis I (MI) and meiosis II (MII) and generate a female gamete with 1 N genome content and ready for fertilization. Though much focus has been devoted to understanding the control of meiotic cell cycle and mechanisms of homolog and sister chromosome segregation during each round of meiosis, it was only recent that attention has been turned to the fact that oocyte meiotic divisions are an exaggerated case of asymmetric cell division (Brunet and Maro 2005), referring to those divisions that give rise to two progeny cells with different fates and, in most cases, sizes and morphologies. Each round of meiotic division gives rise to a small polar body, just large enough to engulf the extruded complement of the genome and destined to apoptotic death, and a large maturing gamete possessing full developmental potential upon fertilization. The size asymmetry ensures that the egg inherits as much valuable cytoplasm as possible to support early embryonic development. Immature mouse oocytes arrested in meiotic prophase isolated from the ovary can be induced to undergo two rounds of meiotic divisions in vitro. Their excellent optical quality and size (80–100 μm in diameter), coupled with the genetic amenability in mouse, have made these large cells attractive for molecular cell biological studies of meiotic cell divisions.

7.2.1 Meiotic Chromosome Migration as a Key Step in Cellular Symmetry Breaking As a general principle in asymmetric cell division, the asymmetry originates from cell polarity established prior to cell division (Li 2007). In asymmetrically dividing mitotic cells, cell polarity serves to position and orient the mitotic spindle, which in turn determines the plane of cytokinesis (Siller and Doe 2009). An exception to this rule is the case of budding yeast, a unicellular eukaryote that divides asymmetrically during mitotic cell cycles to give rise to a new cell (bud) from an older cell (mother). The mother polarizes for both the growth of the new cell and specification of the cell division site, which correlates with the site of initial bud emergence

168

R. Li

and later the bud neck (Pearson and Bloom 2004; Segal and Bloom 2001). Cues and cytoskeleton structures that result from cell polarity orient the mitotic spindle along the mother-bud axis and position it such that spindle elongation segregates the duplicated genomes to opposing sides of the bud neck. Oocyte meiotic divisions represent yet a third scheme of asymmetric cell division, where the genetic material ultimately dictates spindle position and the orientation of cell polarity, which in turn sets up the site of cell division, or polar body extrusion (Longo and Chen 1985; Maro et al. 1986) (Fig. 7.1). When immature oocytes are cultured to initiate meiotic maturation in vitro, the nucleus, or germinal vesicle (GV), assumes a more or less central position in the oocyte. After GV breaks down, a meiotic spindle forms around the chromosomes, which align along the spindle equator. The spindle axis forms randomly with no influence from the cortex. Subsequently, the meiotic spindle moves from the oocyte center toward the cortex until one of the poles of the spindle is butting against the cortex (Verlhac et al. 2000). Not until after this migration does anaphase initiates along an axis perpendicular to the proximal cortex while a small polar body buds out engulfing a set of the segregating chromosomes. Cytoplasmic abscission then occurs to complete the extrusion of the first polar body. The center-to-cortex movement of the chromosomes along with the meiotic spindle breaks the oocyte symmetry not just on a geometric level, but also plays a direct role in establishing a polar cortical domain for polar body extrusion. Meiotic chromosomes signal the cortex to form an actin and myosin-II rich domain that is instrumental for both spindle orientation, polar body budding and cytokinesis (see below). As the effectiveness of this signal is inversely related to the distance between the chromosomes and cortex with the maximum allowable distance much shorter than the radius of the oocyte, the chromosomes must move toward the cortex in order to signal the formation of the polarized actomyosin domain for polar body

1 GV

2

3

4N

Meiosis I (MI)

2N

4

5

6

1N

Meiosis II (MII)

Fig. 7.1 Successive stages of meiotic maturation of mouse oocytes. Step 1: germinal vesicle (GV) break down and assembly of meiosis I (MI) spindle. Step2: movement of chromosomes and spindle from the oocyte center to a subcortical location and subsequent induction of a cortical actin cap (yellow) and myosin-II ring (red) by the chromatin (black). Step 3: extrusion of the first polar body, reducing genome ploidy to 2 N through segregation of homologs. Step 4: assembly of the second meiotic spindle and induction of cortical actin cap and myosin-II ring. Note at this stage the spindle is oriented in parallel to the actin cap. Step 5: the onset of MII anaphase is triggered by sperm entry during fertilization. Concomitant with sister chromosome segregation, the spindle rotates to orient in the direction of polar body extrusion. Step 6: cytoplasmic abscission to complete the extrusion of the second polar body, leaving behind a female pronucleus with 1 N genome content

7

Actin-based Chromosome Movements in Cell Division

169

extrusion. It is important to note that this cortical movement may not be required in vivo, as GV is reported to be attached to the cortex rather being in a central position when oocytes of several mammalian species were observed in situ (Albertini and Barrett 2004). As the chromosome movement in ex vivo cultured oocytes is a robust and highly stereotyped process, it is possible that in vivo this is a safe-keeping mechanism for re-establishment of cortical polarity should the GV detach from the cortex.

7.2.2 Early Evidence That Actin Powers Meiotic Chromosome Migration in Mouse Oocytes While the phenomenon has been long documented, the mechanism of chromosome movement to establish oocyte polarity has remained elusive. Studies in non-vertebrate organisms such as C. elegans had revealed a microtubule and kinesin motor-based system in moving chromosomes during oocyte meiosis (Yang et al. 2003, 2005); however, several studies in mouse oocytes reported the surprising finding that depolymerizing microtubules with different drugs had no inhibitory effect on chromosome movement (Longo and Chen 1985; Maro et al. 1986; Van Blerkom and Bell 1986). In fact, a more recent study using time-lapse 2-photon microscopy and 3-dimensional tracking found that in the absence of microtubules, meiotic chromosomes moved at an even faster speed than that in the unperturbed situation (Li et al. 2008). In contrast, actin poisons, such as latrunculin A, which inhibits actin polymerization, or jasplakinolide, which suppresses actin depolymerization, completely blocked chromosome migration (Dumont et al. 2007a; Li et al. 2008; Maro et al. 1986; Verlhac et al. 2000). Another important clue came from the finding that deletion of the mouse formin-2 (Fmn2) gene, previously thought to be involved in limb development, led to a female sterile phenotype (Leader et al. 2002). Observation of the meiotic process in mutant oocytes found that the failure in polar body extrusion was due to a lack of chromosome translocation to the cortex during meiosis I (Dumont et al. 2007a; Leader et al. 2002; Li et al. 2008). Around the same time, it was found in separate lines of research that formin family proteins nucleate the formation of actin filaments through their conserved formin-homology 2 (FH2) domain (Pruyne et al. 2002; Sagot et al. 2002b). The formins and the Arp2/3 complex are two highly conserved families of actinnucleating proteins. The discovery and characterization of these nucleating proteins and their upstream regulatory pathways have significantly improved our understanding of the regulation of actin dynamics. The Arp2/3 complex is considered a key player in actin polymerization-driven motility owning to its ability to nucleate dendritic actin structures that power the movement of such structures as pathogenic bacteria and viruses, endosomal vesicles, and the leading of migrating cells (Carlier and Pantaloni 2007; Galletta and Cooper 2009; Pollard and Borisy 2003). Formins, on the other hand, nucleate actin filaments without forming branches and stay

170

R. Li

processively at the barbed ends of actin filaments to facilitate their sustained elongation (Higgs 2005; Kovar 2006). The best studied role of formin in actin-based motility is in yeast, where formin proteins nucleate the formation of actin cables to serve as highways for type V myosin to transport vesicles and organelles toward the growing bud (Evangelista et al. 2002; Sagot et al. 2002a). These yeast formins belong to the Dia (as in the Drosophila formin Diaphanous) type whose mechanism of action and regulation by Rho GTPases are relatively well understood. Formin-2 does not belong to the Dia subtype but it contains the FH2 domain, as well as the FH1 domain that binds profilin, while lacking an N-terminal Rho GTPase binding domain that confers regulation of the actin nucleation activity (Higgs 2005).

7.2.3 Pushing or Pulling – How Does Actin Move Chromosomes? An obstacle in understanding how actin mediates meiotic chromosome movement was the inability to visualize filamentous actin in the vicinity of the migrating chromosomes. Staining chemically fixed oocytes with fluorescent phalloidin, which binds selectively filamentous (F)-actin, showed only actin associated with oocyte cortex (Dumont et al. 2007a; Maro et al. 1986; Simerly et al. 1998). Electron microscopy revealed some intriguing 8–10 nm filament bundles with regular striations in oocyte cytoplasm, some of which were clearly in the vicinity of the chromosomes, but immuno-gold labeling suggests that these structures contained mostly keratin filaments and some actin (Capco et al. 1993; Li et al. 2008). Whether and how these structures are involved in chromosome motility remains unclear. An inability of phalloidin to stain all types of actin structures in fixed cells is not unexpected, however, as highly dynamic actin filaments may be difficult to preserve with chemical fixation especially in a large cell as the oocyte. Exciting new insights into the mechanism by which actin moves chromosomes have come from three recent studies that employed protein probes for visualization of F-actin in live oocytes. Two of these used a probe made from the N-terminal 261 amino acids of utrophin, a member of the Dystrophin family actin binding proteins (Azoury et al. 2008; Schuh and Ellenberg 2008). The utrophin probe contains two calponin homology (CH) domains that confer binding specifically to F-actin but not globular (G)-actin (Burkel et al. 2007). Expressing this probe, which also contains a GFP tag, in mouse oocytes revealed, in addition to cortical actin, an abundance of actin bundles that form networks in oocyte cytoplasm, in particular surrounding and as part of the meiotic spindle. If the spindle is eliminated with nocodazol, the actin bundles can be seen in direct association with the chromosomes. Both studies showed that these actin networks are high dynamic and their assembly relies on Fmn-2. Both works demonstrated that the actin filaments network connects the spindle to the cortex, and the work by Schuh and Ellenberg showed that movement of the spindle toward the cortex correlated with an inward local curvature of the cortex, suggesting that the actin network generate a pulling force on the spindle pole and the proximal cortex (Fig. 7.2a). This possibility is corroborated by the observation

7

Actin-based Chromosome Movements in Cell Division

171

A pulling

B pushing

Fig. 7.2 Two models of how actin-based forces move MI chromosomes/spindle from oocyte center to periphery. (a) A model proposed by Schuh and Ellenberg (2008) where a contractile actin network (red) pulls the spindle toward the cortex. (b) A model proposed by Li et al. (2008) where actin polymerization in the vicinity of the chromosomes pushes the spindle toward the cortex. In both cases, the net translocation of the chromosome/spindle requires a symmetry breaking event that results in an imbalance of the actin-based forces

of phosphorylated myosin light chain, a marker for activated myosin-II, on spindle poles and inhibition of spindle movement by ML-7, a myosin light chain kinase inhibitor (Azoury et al. 2008; Schuh and Ellenberg 2008). In the third study (Li et al. 2008), a different probe, called Lifeact, was used for visualizing F-actin in live oocytes. Lifeact is a 17 aa peptide from the yeast actin binding protein Abp140 that is necessary and sufficient for this protein to bind F-actin (Riedl et al. 2008). Using Lifeact labeled with FITC, Li et al. observed a cloud of actin surrounding meiotic chromosomes after GV breakdown. During chromosome movement, the actin cloud distributes asymmetrically around the chromosomes with the majority trailing behind the chromosome mass. This spatial relationship between actin and the moving cargo resembles that in actin polymerization-driven movement of the pathogenic bacteria Listeria monocytogene, where actin forms a comet tail behind the bacterium. Based on this and their observation that inhibition of myosin II using blebbistatin (an inhibitor of nonmuscle myosin II ATPase activity) or ML-7 did not prevent chromosome migration while blocking polar body extrusion, they concluded that actin polymerization pushes the chromosomes toward the cortex (Fig. 7.2b). One observation that fueled this thinking is that Jasplakinolide, which suppresses actin turnover, did not block random chromosome motion but prevented persistent migration toward the oocyte cortex. A previous model of actin polymerization driven motility postulated that actin turnover is critical for the ability of actin polymerization to break symmetry

172

R. Li

in order to bias force generation on the cargo surface to support persistent motion in a certain direction (van Oudenaarden and Theriot 1999). Fmn2 was found to also localize around the chromosomes and in its absence the actin cloud was not observed. How do we reconcile the different conclusions drawn from above studies? First, it is important to evaluate the experimental observations separately from their interpretations. The main difference between the observations, which led to different conclusions, was the types and location of actin structures observed with the use of the two different probes. An ideal F-actin probe should be able to recognize all types of actin structures, but this is hardly possible using probes derived from actin-binding domains of cellular proteins, as no actin-binding protein known to date associates with every type of actin structures present in the cell. For example, although Abp140, from which Lifeact was derived, associate with two of the three types of actin structures in yeast – actin cables and cortical patches, it does not bind the actomyosin contractile ring (Asakura et al. 1998). CH domains from different actin-binding proteins associate with specific actin structures in the cell (Washington and Knecht 2008). Therefore, it is likely that Lifeact and the utrophin CH domains have recognized different actin structures in oocytes. Another factor to consider is the potential effect that an F-actin probe can have on native actin dynamics or organization either directly or by altering the way endogenous proteins interact with actin. While phalloidin has outstanding specificity toward F-actin, it is not useful for probing F-actin in live cells as the compound stabilizes and promotes the polymerization of F-actin even in substoichiometric amounts (Dancker et al. 1975). Although Lifeact was carefully tested for its neutrality on actin polymerization and stability, it is possible that at high enough concentrations it competes with native actin-binding proteins that share similar sites on the filaments. Likewise, this possibility holds for the utrophin probe, given the fact that the CH domains contact the subunit interface in F-actin where the actin depolymerizing factor ADF/cofilin also binds (Moores et al. 2000).

7.2.4 Questions Ahead Beyond identifying whether one or more kinds of actin structures are involved in meiotic chromosome movement, the next question is how the force is generated to drive chromosome motility. If myosin-II is indeed directly involved in the movement, a contractile force pulling chromosomes toward the cortex is an attractive model. The observation of localization of active myosin-II on spindle poles is intriguing given that myosin-II is known to generate contractile force by assembling along actin filaments to form a contractile network, rather than exerting force from a focal spot. It is also possible that other myosin motors, such as myosinX (see Section 7.4), plays a role in moving the chromosomes. If, instead, actin pushed the chromosomes toward the cortex, it is somewhat expected that actin barbed ends, where filament elongation occurs, to be facing the chromosomes.

7

Actin-based Chromosome Movements in Cell Division

173

There is currently no high resolution observation of the organization of actin filaments or the site of actin polymerization within the actin cloud that surround the chromosomes. Another question of interest is how a contractile or pushing force might be exerted on the chromosomes. In unperturbed oocytes, actin, observed either with the utrophin probe or with Lifeact, appears to be largely outside or at the periphery of the spindle, which suggests that the force may be exerted on spindle components as opposed to on the chromosomes directly. However, as chromosome movements still occurs, and at an even higher speed, when spindle microtubules were removed with nocodazole, the structure upon which the force is exerted should remain with the chromosomes. A possible candidate is the endoplasmic reticulum (ER). Brefeldin A, a drug that blocks ER-to-Golgi trafficking and profoundly changes the structure and composition of both Golgi and ER (Nebenfuhr et al. 2002), was found to prevent chromosome movement to the cortex during meiosis, which resulted in either a lack of polar body extrusion or symmetric cell division to generate two equal-sized MII oocytes (Wang et al. 2008). In MI oocytes, the Golgi is known to distribute as puncta dispersed through out the cytoplasm, and Brefeldin A treatment led to further dispersion of Golgi. The ER, in contrast, is concentrated around the meiotic chromosomes after GVBD and form predominantly cortical networks in mature eggs (Mehlmann et al. 1995). It may be interesting to investigate if the ER provides a platform for actin polymerization and force production during chromosome movement. Mechanics aside, even less is known about the temporal regulation of chromosome movement during MI. In cultured oocytes the center-to-cortex movement of chromosomes/spindle as a whole occurs in a well coordinated temporal sequence with chromosome movements within the spindle (Dumont et al. 2007a; Li et al. 2008; Longo and Chen 1985; Maro et al. 1986). The migration to the cortex does not occur until all chromosomes are aligned along the spindle equator. Since chromosome movement still occurs with roughly normal timing when the spindle is disrupted, it is unlikely that there is a checkpoint policing the order between metaphase spindle assembly and migration. Shortly after chromosomes reach the cortex, anaphase initiates and homologous chromosomes segregate within the meiotic spindle. Again, since in fmn2–/– oocytes, where chromosomes do not migrate, anaphase still occurs with roughly normal timing (Dumont et al. 2007a), there is unlikely a checkpoint monitoring the failure in chromosome movement before anaphase onset, but rather these events may be controlled by a common master regulator. However, it was intriguing to find that although chromosomes are segregated to spindle poles in Fmn2–/– oocytes, they rapidly congress to eventually arrest aligned at the equator of what appears to be a recovered metaphase spindle (Dumont et al. 2007a). This suggests that there may indeed be a checkpoint monitoring the position of the spindle or chromosomes or the onset of cytokinesis prior to the exit from meiosis I. A kinase that profoundly impacts many of the meiotic processes, especially asymmetric meiotic division is c-Mos, an upstream regulator of the cytostatic factor (CSF) that maintains meiosis II arrest in vertebrate oocytes (Tunquist and Maller

174

R. Li

2003). Oocytes from Mos–/– female mice fail to undergo chromosome migration, however, unlike in fmn2–/– oocytes cell division often occurs in a roughly symmetric manner, generating two large progeny cells (Choi et al. 1996). In a small fraction of Mos–/– oocytes, the cleavage plane is skewed toward one side, resulting in a somewhat asymmetric division. Careful observation of these oocytes found that the chromosome movement associated with this asymmetry is different from chromosome migration in wild-type oocytes because it is a result of unequal spindle elongation during anaphase (Verlhac et al. 2000). Mos is a mitogen activated protein (MAP) kinase kinase kinase. Several other studies have implicated the MAPK pathway in meiotic chromosome movement. A recent paper reported that dominant negative inhibition of the small GTPase Arf1, prevented chromosome migration (Wang et al. 2009). In Arf1-inhibited oocytes, MAPK kinase activation was absent. These studies implicate the involvement of a MAP kinase pathway in the regulation of meiotic chromosome movement. Identifying the relevant target(s) of active MAP kinase may also help revealing the molecular machinery underlying chromosome movement. Finally, we return to the issue of establishing asymmetry in the context of asymmetric meiotic cell division. It remains unclear whether there is a pre-exiting asymmetry or cue before and dictates the direction of chromosome movement, or whether chromosome migration, at least in cultured oocytes, truly reflects a case of spontaneous symmetry breaking. Timelapse observations indicate that the orientation of the spindle dictates, to a large degree, the direction of chromosomes migration. This could either be explained by forces directly concentrated on the spindle poles or the tendency to minimize drag force when moving a non-spherical object (the spindle) through the cytoplasm. Thus, in the presence of the spindle, only a bilateral symmetry needs to be broken for the chromosomes to migrate to one side of the oocyte. These observations led to the proposal of a tug-of-war model of an imbalance of pulling forces on the spindle poles to break the bilateral symmetry (Maro and Verlhac 2002; Schuh and Ellenberg 2008). While the spindle clearly can bias the axis of the movement, the fact that in its absence chromosomes still move to a single random position below the cortex (Longo and Chen 1985; Maro et al. 1986; Van Blerkom and Bell 1986) suggest that there is a mechanism for spontaneous symmetry breaking. Only successful movement of all chromosomes to a single unique position accomplishes cellular symmetry breaking to establish a single site for polar body extrusion. The presence of a spontaneous symmetry breaking mechanism and the ability of the spindle or even a pre-existing off-centered location of the GV to bias the direction of chromosome movement are not conflicting notions. In fact in most systems of cell polarization, spontaneous mechanisms collaborate with spatial cues to bring about robust symmetry breaking that responds accurately to even minute physiological signals. An exciting expectation is that akin to symmetry breaking during cell locomotion, symmetry break in oocytes is mechanistically linked to actin dynamics. But unlike cell movement, where actin-based forces are exerted primarily at cell periphery, in the meiotic system actin works within the cell interior to bring about asymmetry through chromosome positioning, which then leads to establishment of cortical polarity (see below).

7

Actin-based Chromosome Movements in Cell Division

175

7.3 Interactions Between Actin and Chromatin to Establish the Site for Polar Body Extrusion While it may seem a one-sided service during chromosome migration, upon arrival at the cell periphery, the chromosomes send a signal, a diffusible one as no contact is required, toward the cortex to initiate the assembly of an actin and myosin-II rich cortical domain that is instrumental for polar body extrusion. This cortical domain is characterized by a ring of myosin II (Deng et al. 2007), which also contains F-actin (referred to as actomyosin ring hereafter), surrounding a cap that is rich in actin filaments (referred to as the actin cap hereafter) (Deng et al. 2003; Longo and Chen 1985; Maro et al. 1984) (Fig. 7.3). Unlike in mitotic cell division where a contractile cytokinetic ring assembles during telophase (Glotzer 2001; Li 2007), the meiotic actomyosin ring forms prior to anaphase onset and constricts at the base of the polar body as the spindle elongates during anaphase. These cortical actin and

D

E

chromatin

RCC1 RanGDP RanGDP

RanGTP

RanGAP

Fig. 7.3 Induction of a cortical domain for polar body extrusion by meiotic chromatin. (a–c) cortical actin caps (red) and myosin-II rings (green) induced by MII chromosomes or injected DNA beads. Scale bar: 20 μm. Adapted from Supplemental figure 1 of Deng et al. (2007) with permission. (d, e) Schematic diagrams showing how differential localization of Ran GEF (RCC1) and GAP (RanGAP) to the chromatin or cytoplasm, respectively (d), generates a Ran-GTP gradient spreading from the chromatin that signals assembly of an actin cap and a myosin ring on the proximal but not distal cortex (e)

176

R. Li

myosin structures also play an important role in maintaining the position and the orientation of the meiotic spindle (see below), thus dictating the axis along which meiotic chromosomes segregate.

7.3.1 Chromatin-Induced Assembly of Cortical Actin and Myosin-II for Polar Body Extrusion The notion that the chromatin but not microtubules provides the signal for meiotic actomyosin ring formation came from experiments where chromosomes, either maternal or experimentally introduced, can robustly induce the actin cap and actomyosin ring even after the meiotic spindle is completely disrupted with microtubule inhibitors (Longo and Chen 1985; Maro et al. 1984). A useful experimental system for investigating the chromatin-induced assembly of the actomyosin ring is one where microinjected nanobeads, coated with plasmid DNA, can induce the formation of an actomyosin ring around an actin cap at virtually any location on the cortex that is in the vicinity of the DNA beads (Deng et al. 2007) (Fig. 7.3). Since the beads can be positioned at different distances from the cortex, it was shown that the ability of the chromatin to induce cortical actomyosin assembly is inversely related to the distance between the chromatin and the closest cortex, with the maximally allowable distance to be ∼20 μm, considerably smaller than the radius of the oocytes (∼ 40 μm). This spatial limitation perhaps helps to ensure that cell division occurs asymmetrically during oocyte maturation. In addition, the size of the cortical domain induced was found to be positively related to the number of the DNA beads injected. This may be indicative of an elegant system where the polar body is tailored according to the size of the genome to be extruded so as to accomplish the process with minimal consumption of the cytoplasm and plasma membrane. Similarly, a new study found that the chromatin also determines the size and shape of the spindle assembled in Xenopus egg extracts (Dinarina et al. 2009).

7.3.2 The Ran-GTP Gradient: the Oocyte’s Tape Measure for Polar Body Extrusion? The above observations raised the question of how distance and size may be measured in maturing oocytes. A similar problem of size and proximity determination has been studied in chromatin-induced formation of bipolar spindle (Heald et al. 1996), where the small GTPase Ran plays a crucial role (Goodman and Zheng 2006; Kalab and Heald 2008). Ran also has a well known interphase function in nuclear transport through regulation of importin-cargo interactions (Harel and Forbes 2004). The various functions of Ran all rely on a spatial separation between the opposing sides of the Ran GTPase cycle (Kalab and Heald 2008) (Fig. 7.3). The guanine nucleotide exchange factor (GEF) for Ran, called RCC1, and Ran-GDP associate with the chromatin through cooperative interactions, which, coupled with guanine

7

Actin-based Chromosome Movements in Cell Division

177

nucleotide exchange, releases Ran-GTP at chromosome vicinity (Hao and Macara 2008; Kalab and Heald 2008). The GTPase activating protein (GAP) is diffuse in the cytoplasm and converts Ran-GTP to Ran-GDP at locations away from the chromatin. The spatially separated actions of the GEF and GAP generate a RanGTP gradient that spreads from the chromatin (Caudron et al. 2005; Dumont et al. 2007b; Kalab et al. 2002, 2006). The size of this gradient and how it controls the polymerization of spindle microtubules around the chromosomes are determined by complex molecular interactions (Athale et al. 2008; Bastiaens et al. 2006; Kalab and Heald 2008). The basic mechanism in the case of spindle formation is that Ran-GTP activates factors required for microtubule nucleation, stabilization, and spindle pole formation by freeing these factors from the bound importin family nuclear transport receptors (Goodman and Zheng 2006; Kalab and Heald 2008). In this way, the RanGTP gradient generates a gradient of spindle assembly activities that are highest around the chromatin and decline as the distance from the chromatin increases. Although the Ran-GTP gradient had been demonstrated using fluorescent resonance energy transfer (FRET)-based probes in a variety of cell types, it was only recent that this gradient was observed in mouse oocytes during both meiosis I and II (Dumont et al. 2007b). Surprisingly, active Ran did not appear to be required for the formation of meiosis I spindle; however, the spread of this gradient, on the order of 10–30 μm, was roughly in the same range as the allowable distance for chromatininduced assembly of cortical actin and myosin (Deng et al. 2007; Dumont et al. 2007b). Supporting an involvement of Ran-GTP in the chromatin-induced cortical actin and myosin assembly, this process is prevented by injection of a dominant negative Ran mutant, RanT24N , which inhibits RCC1, into the oocyte. A constitutively active form of Ran also inhibits the process, which is often interpreted as a requirement for the Ran-GTP gradient, however, this may also reflect a requirement for the Ran GTPase cycle in recycling Ran from the complex with importin. It has yet to be demonstrated directly that a Ran-GTP gradient governs the distance constrain underlying the chromatin-induced cortical response.

7.3.3 How Might Ran Regulate the Assembly of Cortical Actin and Myosin? While Rho family GTPases are well known regulators of the actin cytoskeleton, there is no blueprint on a pathway connecting Ran and actin. As mentioned above, many of the factors that control microtubule nucleation and stabilization contain nuclear localization sequences (NLS) that bind importin family members, which keep these factors in an inactive state. Ran-GTP activates these factors by dissociating them from the bound importin (Goodman and Zheng 2006; Kalab and Heald 2008). It remains to be tested if an analogous mechanism applies to the way in which Ran regulates actin and myosin assembly. There are indications that Ran may ultimately connect with Rho family GTPases to regulate actin and myosin. Meiotic chromatin induces localization of YFP-tagged p21-binding domain of p21activated kinases (PAK-PBD), which was used as a probe for activated Rac, to the

178

R. Li

same region as cortical actin accumulation; however, immunofluorescence staining of Rac itself did not show polarized localization (Halet and Carroll 2007). Dominant negative inhibition of Rac did not affect the formation of the actomyosin domain overlying meiotic chromosomes in either MI or MII, but, surprisingly, resulted in abnormally long spindle with misaligned chromosomes in MI and a defect in spindle orientation/position in MII (see the following section for more discussion). Cdc42 perturbation with either dominant negative or constitutively active Cdc42 mutant also resulted in abnormally long MI spindles, though chromosomes appeared to be well aligned on the metaphase plate (Na and Zernicka-Goetz 2006). This defect is accompanied by a failure in chromosome migration to the cortex and a lack of cortical actin polarization. However, it is unclear if Cdc42 plays a direct role in chromosome movement or cortical actin assembly because the failure in these processes might be consequences of a highly elongated spindle with chromosomes stuck at the spindle equator and are thus too far away from the cortex on either side. A recent study found that IQGAP1, a protein involved in diverse cytoskeletal based functions and binds active Rac and Cdc42, localizes to the actomyosin ring in mouse oocytes (Bielak-Zmijewska et al. 2008). This localization, as well as the formation of the actin cap, was inhibited in oocytes injected with toxin B, a member of the clostridium toxins that glucosylate Rho family GTPases and hereby block their interaction with effector proteins (Aktories et al. 2000). However, as toxin B is effective toward all members of the Rho family, it is unclear which Rho GTPases is required for IQGAP localization and actin cap formation. In Xenopus oocytes, active Cdc42, detected by using a fluorescently tagged, Cdc42 binding domain of a Cdc42 effector (in this case WASP), localizes to the polar cortical domain in a manner that is dependent on asymmetric spindle position (Ma et al. 2006; Zhang et al. 2008). Dominant negative Cdc42 prevented dynamic actin assembly in the cortical zone inside the contractile ring during polar body extrusion. Work in Xenopus oocytes also indicated that Rho is crucial for the assembly of the actomyosin ring for polar body extrusion. Although this has not been reported for mouse oocytes (Ma et al. 2006; Zhang et al. 2008), it was interesting to find that while LatA treatment did not prevent myosin-II from localizing to a polar cap, inhibition of myosin light chain kinase by ML-7 blocked assembly of both myosin-II and actin to the cortex (Deng et al. 2007). This suggests that activation and localization of myosin-II may be one of the first steps in the establishment of the cortical actomyosin domain. It also appears that localization of myosin to the cortical cap and formation of a myosin ring require different levels of Ran activity: inhibition of the former requires a higher concentration of RanT24N than the latter (Deng et al. 2007). In addition, treatment with blebbistatin did not affect myosin localization to the cortical cap but prevented myosin ring formation. These results suggest that the initial localization of myosin-II to the cortical cap and formation of the myosin-II ring are distinct events, with the latter requires myosin-II contractility and a higher level Ran activity. It will be interesting to find out if one or both of these processes are regulated by Rho.

7

Actin-based Chromosome Movements in Cell Division

179

7.4 Does Actin Have a Hand in the Spindle? The effects of actin on chromosome positioning and the reverse impact of chromatin on cortical actomyosin assembly, as discussed above, may seem at best a meiotic anomaly, as microtubules appear to have all but monopolized the role in moving chromosomes and the nucleus during cell division and interphase. Even the widely studied role for actin and myosins in spindle orientation, a process crucial to both symmetric and asymmetric cell divisions, seem to be mediated through interactions of astral (non-spindle but being part of an aster) microtubules with the cortical actin cytoskeleton (Segal and Bloom 2001; Siller and Doe 2009). During polar body extrusion, however, actin’s role in spindle orientation may be more direct, since meiotic spindles are largely devoid of astral microtubules. During MI, once the spindle moves to the cortex, one of its poles is tightly attached to the cortex and the spindle axis is perpendicular to the actin cap. Actin is likely to be directly responsible for this attachment, as suggested by a strong accumulation of actin filaments between the spindle pole and the adjacent cortex and the indentation observed precisely in this region of the cortex (Schuh and Ellenberg 2008). In Xenopus oocytes, the connections between the spindle and the cortex appear to be prominently microtubules, but an actin-based motor protein, myosin-X, associates with these microtubules and its inhibition, using a dominant negative construct consisting of the tail domain of myosin X, dissociates the spindle from the cortex (Weber et al. 2004). The myosin tail homology domain 4 (MyTH4) of myosin-X binds microtubules and is sufficient for targeting myosin-X to microtubules in oocytes. In MII mouse oocytes, the metaphase spindle lies just below and in parallel to the cortex prior to the onset of anaphase induced by fertilization. During polar body extrusion, the spindle rotates after anaphase onset to enable extrusion of one set of the segregating sister chromosomes. This process is also actin-dependent as cytochalasin prevented spindle rotation, even though anaphase occurs normally following fertilization (Zhu et al. 2003). In cytochalasin B-treated oocytes, the spindle was eventually released from the cortex into the cytoplasm. Similar experiments also suggested a requirement for actin in spindle rotation in Xenopus oocytes (Gard et al. 1995). If the requirement for actin in spindle orientation seems an acceptable partnership, an involvement of actin in spindle assembly may be an outright encroachment on microtubule’s autonomy in this classically microtubule-based process. The notion that actin does not have any prominent role in spindle assembly or chromosome movement within the spindle was reinforced by studies that utilized cytochalasin D-treated Xenopus egg extracts to assemble functional spindles in vitro (Mitchison 2005) and the observation that staining with phalloidin, a compound that binds with high affinity and specificity to F-actin, did not reveal F-actin presence in the spindle (Barak et al. 1981). Evidence that actin and its associated proteins, such as myosin, are components of the spindle came from observation of these proteins, especially actin, in certain cell types and by using certain staining methods (Forer et al. 1979; Schloss et al. 1977; Silverman-Gavrila and Forer 2003). Experiments using actin inhibitors performed mostly in insect spermatocytes revealed some roles

180

R. Li

for actin in kinetochore microtubule dynamics and chromosome movement during anaphase (Fabian and Forer 2007; Forer and Pickett-Heaps 1998; Forer et al. 2007; Silverman-Gavrila and Forer 2000, 2001), but it remains unclear precisely what actin’s roles are and whether they are conserved across cell types and species. Below we discuss several recent studies in oocytes and early embryos that have shed new light on the involvements actin in specific aspects of spindle assembly. In starfish oocytes, because of the large size of the cell (170 μm) and its nucleus (80 μm) relative to the length of astral microtubules (maximally 15–20 μm), meiotic chromosomes scattered in the nucleoplasm rely on a cytoplasmic actin network for an unusual phase of congression toward the microtubule aster at the animal pole, to a point where they are close enough to be captured and congress to spindle equator in a more conventional microtubule-dependent manner (Lenart et al. 2005). Polymerization of the actin network occurs after GVBD near the nuclear rim, which subsequently disassembles but leaves a dense actin patch around each chromosome. High-resolution movies showed that contraction of the residual cytoplasmic actin network correlated with chromosome movement toward the animal pole. Interestingly, injection of DNA-coated beads can also trigger the formation of actin patches, revealing a signal elicited by the chromatin that induces actin polymerization in its vicinity. Further molecular dissection should shed light on whether the chromatin-induced actin assembly in starfish oocytes follows a similar mechanism to those observed at different stages of maturation of mouse oocytes discussed earlier. The involvement of actin in chromosome congression might have evolved to deal with a large nuclear size in cells such as oocytes. In addition to this more specialized requirement, a number of studies have suggested a role for actin in spindle formation. As mentioned earlier, dominant negative inhibition of either Cdc42 or Rac resulted in abnormally long MI spindle in mouse oocytes (Halet and Carroll 2007; Na and Zernicka-Goetz 2006). Given that actin nucleating proteins are major targets of these GTPases, a role for actin polymerization during spindle formation may underlie these findings, although it is possible that these GTPases regulate spindle length independently of actin. It is interesting that there is a subtle difference between the effects of inhibition of Rac and Cdc42: while Cdc42 inhibition only affected spindle length, Rac inhibition also resulted in chromosomes spreading along the entire length of the abnormally long spindle, suggesting that Rac, but not Cdc42, plays a role in chromosome congression. More direct evidence that actin is a functional component of the spindle came from a study testing the requirement for myosin-X in spindle assembly during mitotic divisions in Xenopus early embryos (Woolner et al. 2008). Knocking down myosin-X with morpholinos generated two obvious spindle defects: pole fragmentation and abnormally long spindle. Similar defects can also be produced with an anti-myosin-X antibody in the in vitro spindle assembly assay using Xenopus egg extract. Interestingly, the long spindle phenotype can be rescued by an N-terminal construct of myosin-X that contains the motor domain, IQ motifs and a coiledcoil region, whereas a construct lacking the motor domain but retaining the rest of the molecule, including the microtubule-binding MyTH4 region, suppressed pole

7

Actin-based Chromosome Movements in Cell Division

181

fragmentation in morpholino-treated embryos. Moreover, inhibition of actin polymerization with latrunculin B also rescued the spindle length abnormality but not spindle pole fragmentation. The authors also observed prominent fibers in the spindle that are labeled with the utrophin probe and interpreted these as actin cables. These results suggest the possibility of multiple roles for actin and myosin-X in spindle assembly. Mechanistic implications as well as the generality of these observations await further investigation. Clearly, functional spindles with the ability to segregate chromosomes can assemble in the absence of actin even in meiotic systems where actin is thought to be more extensively involved (Dumont et al. 2007a). Possibly, actin is a wildcard for spindle function and helps solve cell type or speciesspecific problems through its versatile assembly pathways and many accessory factors.

7.5 Perspectives Above we have highlighted intriguing findings, obtained mainly in oocytes undergoing meiotic maturation, which have opened up an exciting area of research on the direct interactions between the actin cytoskeletal system and chromosomes. Even though many of the phenomena, such as chromosome migration during meiosis I and assembly of a specialized cortical domain for polar body extrusion, have been known for some time, live-cell imaging and molecular genetic tools developed in recent years have brought new interests to the investigation on these phenomena and have begun to allow elucidation of mechanistic issues. Compared to many other areas of actin-based motility, the study of actin-driven chromosomal motility still faces the task of generating a workable molecular parts list, but the extensive knowledge on the conserved molecular machineries and pathways that control actin polymerization will enable rapid progress through hypothesis-driven experiments using various molecular manipulations such as gene knockout, RNAi or morpholinos, in suitable oocyte meiotic systems. There are a number of key issues that need to be elucidated in the near future. First, are there different pathways for chromatin-induced actin assembly even in a single system such as the mouse oocyte. So far, the evidence suggests so. For example, while Ran is required for the chromatin-induced formation of cortical actomyosin cap, Ran inhibition does not appear to affect the earlier event of chromosome movement from oocyte center to the cortex. If some fundamental mechanisms underlying chromatin-actin interactions can be discovered in the meiotic models, it will be interesting to explore the potential application or variation of these mechanisms in mitotic systems or even in interphase cells where actin’s role in the regulation of chromosome structure and gene expression remains elusive. In the general field of actin-based motility, chromosome movement may be one of the few frontiers left where there may still be some novel mechanisms waiting to be unraveled. Acknowledgement The author would like to thank Marie-Helene Verlhac for her corrections and suggestions on the manuscript.

182

R. Li

References Aktories K, Schmidt G, Just I (2000) Rho GTPases as targets of bacterial protein toxins. Biol Chem 381: 421–426. Albertini DF, Barrett SL (2004) The developmental origins of mammalian oocyte polarity. Semin Cell Dev Biol 15: 599–606. Asakura T, Sasaki T, Nagano F et al. (1998) Isolation and characterization of a novel actin filamentbinding protein from Saccharomyces cerevisiae. Oncogene 16: 121–130. Athale CA, Dinarina A, Mora-Coral M et al. (2008) Regulation of microtubule dynamics by reaction cascades around chromosomes. Science 322: 1243–1247. Azoury J, Lee KW, Georget V et al. (2008) Spindle positioning in mouse oocytes relies on a dynamic meshwork of actin filaments. Curr Biol 18: 1514–1519. Barak LS, Nothnagel EA, DeMarco EF, Webb WW (1981) Differential staining of actin in metaphase spindles with 7-nitrobenz-2-oxa-1,3-diazole-phallacidin and fluorescent DNase: is actin involved in chromosomal movement? Proc Nat Acad Sci USA 78: 3034–3038. Bastiaens P, Caudron M, Niethammer P, Karsenti E (2006) Gradients in the self-organization of the mitotic spindle. Trends Cell Biol 16: 125–134. Bielak-Zmijewska A, Kolano A, Szczepanska K et al. (2008) Cdc42 protein acts upstream of IQGAP1 and regulates cytokinesis in mouse oocytes and embryos. Dev Biol 322: 21–32. Brunet S, Maro B (2005) Cytoskeleton and cell cycle control during meiotic maturation of the mouse oocyte: integrating time and space. Reproduction 130: 801–811. Burkel BM, von Dassow G, Bement WM (2007) Versatile fluorescent probes for actin filaments based on the actin-binding domain of utrophin. Cell Motil Cytoskeleton 64: 822–832. Capco DG, Gallicano GI, McGaughey RW et al. (1993) Cytoskeletal sheets of mammalian eggs and embryos: a lattice-like network of intermediate filaments. Cell Motil Cytoskeleton 24: 85–99. Carlier MF, Pantaloni D (2007) Control of actin assembly dynamics in cell motility. J Biol Chem 282: 23005–23009. Caudron M, Bunt G, Bastiaens P, Karsenti E (2005) Spatial coordination of spindle assembly by chromosome-mediated signaling gradients. Science 309: 1373–1376. Choi T, Fukasawa K, Zhou R et al. (1996) The Mos/mitogen-activated protein kinase (MAPK) pathway regulates the size and degradation of the first polar body in maturing mouse oocytes. Proc Nat Acad Sci USA 93: 7032–7035. Dancker P, Low I, Hasselbach W, Wieland T (1975) Interaction of actin with phalloidin: polymerization and stabilization of F-actin. Biochim Biophys Acta 400: 407–414. Deng M, Suraneni P, Schultz RM, Li R (2007) The Ran GTPase mediates chromatin signaling to control cortical polarity during polar body extrusion in mouse oocytes. Dev Cell 12: 301–308. Deng M, Kishikawa H, Yanagimachi R et al. (2003) Chromatin-mediated cortical granule redistribution is responsible for the formation of the cortical granule-free domain in mouse eggs. Dev Biol 257: 166–176. Dinarina A, Pugieux C, Corral MM et al. (2009) Chromatin shapes the mitotic spindle. Cell 138: 502–513. Dumont J, Million K, Sunderland K et al. (2007a) Formin-2 is required for spindle migration and for the late steps of cytokinesis in mouse oocytes. Dev Biol 301: 254–265. Dumont J, Petri S, Pellegrin F et al. (2007b) A centriole- and RanGTP-independent spindle assembly pathway in meiosis I of vertebrate oocytes. J Cell Biol 176: 295–305. Evangelista M, Pruyne D, Amberg DC et al. (2002) Formins direct Arp2/3-independent actin filament assembly to polarize cell growth in yeast. Nat cell biol 4: 32–41. Fabian L, Forer A (2007) Possible roles of actin and myosin during anaphase chromosome movements in locust spermatocytes. Protoplasma 231: 201–213. Forer A, Pickett-Heaps JD (1998) Cytochalasin D and latrunculin affect chromosome behaviour during meiosis in crane-fly spermatocytes. Chromosome Res 6: 533–549.

7

Actin-based Chromosome Movements in Cell Division

183

Forer A, Jackson WT, Engberg A (1979) Actin in spindles of Haemanthus katherinae endosperm. II. Distribution of actin in chromosomal spindle fibres, determined by analysis of serial sections. J Cell Sci 37: 349–371. Forer A, Spurck T, Pickett-Heaps JD (2007) Actin and myosin inhibitors block elongation of kinetochore fibre stubs in metaphase crane-fly spermatocytes. Protoplasma 232: 79–85. Galletta BJ, Cooper JA (2009) Actin and endocytosis: mechanisms and phylogeny. Curr Opin Cell Biol 21: 20–27. Gard DL, Cha BJ, Roeder AD (1995) F-actin is required for spindle anchoring and rotation in Xenopus oocytes: a re-examination of the effects of cytochalasin B on oocyte maturation. Zygote 3: 17–26. Glotzer M (2001) Animal cell cytokinesis. Annu Rev Cell Dev Biol 17: 351–386. Goodman B, Zheng Y (2006) Mitotic spindle morphogenesis: Ran on the microtubule cytoskeleton and beyond. Biochem Soc Trans 34: 716–721. Halet G, Carroll J (2007) Rac activity is polarized and regulates meiotic spindle stability and anchoring in mammalian oocytes. Dev Cell 12: 309–317. Hao Y, Macara IG (2008) Regulation of chromatin binding by a conformational switch in the tail of the Ran exchange factor RCC1. J Cell Biol 182: 827–836. Harel A, Forbes DJ (2004) Importin beta: conducting a much larger cellular symphony. Mol Cell 16: 319–330. Heald R, Tournebize R, Blank T et al. (1996) Self-organization of microtubules into bipolar spindles around artificial chromosomes in Xenopus egg extracts [see comments]. Nature 382: 420–425. Higgs HN (2005) Formin proteins: a domain-based approach. Trends Biochem Sci 30: 342–353. Kalab P, Heald R (2008) The RanGTP gradient – a GPS for the mitotic spindle. J Cell Sci 121: 1577–1586. Kalab P, Weis K, Heald R (2002) Visualization of a Ran-GTP gradient in interphase and mitotic Xenopus egg extracts. Science 295: 2452–2456. Kalab P, Pralle A, Isacoff EY et al. (2006) Analysis of a RanGTP-regulated gradient in mitotic somatic cells. Nature 440: 697–701. Kovar DR (2006) Molecular details of formin-mediated actin assembly. Curr Opin Cell Biol 18: 11–17. Kumaran RI, Thakar R, Spector DL (2008) Chromatin dynamics and gene positioning. Cell 132: 929–934. Leader B, Lim H, Carabatsos MJ et al. (2002) Formin-2, polyploidy, hypofertility and positioning of the meiotic spindle in mouse oocytes. Nat cell biol 4: 921–928. Lenart P, Bacher CP, Daigle N et al. (2005) A contractile nuclear actin network drives chromosome congression in oocytes. Nature 436: 812–818. Li H, Guo F, Rubinstein B, Li R (2008) Actin-driven chromosomal motility leads to symmetry breaking in mammalian meiotic oocytes. Nat cell biol 10: 1301–1308. Li R (2007) Cytokinesis in development and disease: variations on a common theme. Cell Mol Life Sci 64: 3044–3058. Longo FJ, Chen DY (1985) Development of cortical polarity in mouse eggs: involvement of the meiotic apparatus. Dev Biol 107: 382–394. Louvet E, Percipalle P (2009) Transcriptional control of gene expression by actin and myosin. Int Rev Cell Mol Biol 272: 107–147. Ma C, Benink HA, Cheng D et al. (2006) Cdc42 activation couples spindle positioning to first polar body formation in oocyte maturation. Curr Biol 16: 214–220. Maro B, Verlhac MH (2002) Polar body formation: new rules for asymmetric divisions. Nat cell biol 4: E281–E283. Maro B, Johnson MH, Pickering SJ, Flach G (1984) Changes in actin distribution during fertilization of the mouse egg. J Embryol Exp Morphol 81: 211–237. Maro B, Johnson MH, Webb M, Flach G (1986) Mechanism of polar body formation in the mouse oocyte: an interaction between the chromosomes, the cytoskeleton and the plasma membrane. J Embryol Exp Morphol 92: 11–32.

184

R. Li

McIntosh JR, Grishchuk EL, West RR (2002) Chromosome-microtubule interactions during mitosis. Annu Rev Cell Dev Biol 18: 193–219. Mehlmann LM, Terasaki M, Jaffe LA, Kline D (1995) Reorganization of the endoplasmic reticulum during meiotic maturation of the mouse oocyte. Dev Biol 170: 607–615. Mitchison TJ (2005) Mechanism and function of poleward flux in Xenopus extract meiotic spindles. Philos Trans R Soc Lond B Biol Sci 360: 623–629. Moores CA, Keep NH, Kendrick-Jones J (2000) Structure of the utrophin actin-binding domain bound to F-actin reveals binding by an induced fit mechanism. J Mol Biol 297: 465–480. Na J, Zernicka-Goetz M (2006) Asymmetric positioning and organization of the meiotic spindle of mouse oocytes requires CDC42 function. Curr Biol 16: 1249–1254. Nebenfuhr A, Ritzenthaler C, Robinson DG (2002) Brefeldin A: deciphering an enigmatic inhibitor of secretion. Plant Physiol 130: 1102–1108. Pearson CG, Bloom K (2004) Dynamic microtubules lead the way for spindle positioning. Nat Rev Mol Cell Biol 5: 481–492. Piekny A, Werner M, Glotzer M (2005) Cytokinesis: welcome to the Rho zone. Trends Cell Biol 15: 651–658. Pollard TD, Borisy GG (2003) Cellular motility driven by assembly and disassembly of actin filaments. Cell 112: 453–465. Pruyne D, Evangelista M, Yang C et al. (2002) Role of formins in actin assembly: nucleation and barbed-end association. Science 297: 612–615. Riedl J, Crevenna AH, Kessenbrock K et al. (2008) Lifeact: a versatile marker to visualize F-actin. Nat Methods 5: 605–607. Sagot I, Klee SK, Pellman D (2002a) Yeast formins regulate cell polarity by controlling the assembly of actin cables. Nat cell biol 4: 42–50. Sagot I, Rodal AA, Moseley J et al. (2002b) An actin nucleation mechanism mediated by Bni1 and profilin. Nat cell biol 4: 626–631. Schloss JA, Milsted A, Goldman RD (1977) Myosin subfragment binding for the localization of actin-like microfilaments in cultured cells. A light and electron microscope study. J Cell Biol 74: 794–815. Schuh M, Ellenberg J (2008) A new model for asymmetric spindle positioning in mouse oocytes. Curr Biol 18: 1986–1992. Segal M, Bloom K (2001) Control of spindle polarity and orientation in Saccharomyces cerevisiae. Trends Cell Biol 11: 160–166. Siller KH, Doe CQ (2009) Spindle orientation during asymmetric cell division. Nat cell biol 11: 365–374. Silverman-Gavrila RV, Forer A (2000) Evidence that actin and myosin are involved in the poleward flux of tubulin in metaphase kinetochore microtubules of crane-fly spermatocytes. J Cell Sci 113 (Pt 4): 597–609. Silverman-Gavrila RV, Forer A (2001) Effects of anti-myosin drugs on anaphase chromosome movement and cytokinesis in crane-fly primary spermatocytes. Cell Motil Cytoskeleton 50: 180–197. Silverman-Gavrila RV, Forer A (2003) Myosin localization during meiosis I of crane-fly spermatocytes gives indications about its role in division. Cell Motil Cytoskeleton 55: 97–113. Simerly C, Nowak G, de Lanerolle P, Schatten G (1998) Differential expression and functions of cortical myosin IIA and IIB isotypes during meiotic maturation, fertilization, and mitosis in mouse oocytes and embryos. Mol Biol Cell 9: 2509–2525. Takenawa T, Suetsugu S (2007) The WASP-WAVE protein network: connecting the membrane to the cytoskeleton. Nat Rev Mol Cell Biol 8: 37–48. Tunquist BJ, Maller JL (2003) Under arrest: cytostatic factor (CSF)-mediated metaphase arrest in vertebrate eggs. Genes Dev 17: 683–710. Van Blerkom J, Bell H (1986) Regulation of development in the fully grown mouse oocyte: chromosome-mediated temporal and spatial differentiation of the cytoplasm and plasma membrane. J Embryol Exp Morphol 93: 213–238.

7

Actin-based Chromosome Movements in Cell Division

185

van Oudenaarden A, Theriot JA (1999) Cooperative symmetry-breaking by actin polymerization in a model for cell motility. Nat cell biol 1: 493–499. Verlhac MH, Lefebvre C, Guillaud P et al. (2000) Asymmetric division in mouse oocytes: with or without Mos. Curr Biol 10: 1303–1306. Walczak CE, Heald R (2008) Mechanisms of mitotic spindle assembly and function. Int Rev Cytol 265: 111–158. Wang L, Wang ZB, Zhang X et al. (2008) Brefeldin A disrupts asymmetric spindle positioning in mouse oocytes. Dev Biol 313: 155–166. Wang S, Hu J, Guo X et al. (2009) ADP-ribosylation factor 1 regulates asymmetric cell division in female meiosis in the mouse. Biol Reprod 80: 555–562. Washington RW, Knecht DA (2008) Actin binding domains direct actin-binding proteins to different cytoskeletal locations. BMC Cell Biol 9: 10. Weber KL, Sokac AM, Berg JS et al. (2004) A microtubule-binding myosin required for nuclear anchoring and spindle assembly. Nature 431: 325–329. Woolner S, O Brien LL, Wiese C, Bement WM (2008) Myosin-10 and actin filaments are essential for mitotic spindle function. J Cell Biol 182: 77–88. Yang HY, McNally K, McNally FJ (2003) MEI-1/katanin is required for translocation of the meiosis I spindle to the oocyte cortex in C elegans. Dev Biol 260: 245–259. Yang HY, Mains PE, McNally FJ (2005) Kinesin-1 mediates translocation of the meiotic spindle to the oocyte cortex through KCA-1, a novel cargo adapter. J Cell Biol 169: 447–457. Zhang X, Ma C, Miller AL et al. (2008) Polar body emission requires a RhoA contractile ring and Cdc42-mediated membrane protrusion. Dev Cell 15: 386–400. Zhu ZY, Chen DY, Li JS et al. (2003) Rotation of meiotic spindle is controlled by microfilaments in mouse oocytes. Biol Reprod 68: 943–946.

Chapter 8

Roles for Actin Dynamics in Cell Movements During Development Minna Roh-Johnson, Jessica Sullivan-Brown, and Bob Goldstein

Abstract Actin-dependent cellular movements and rearrangements are crucial for development. Studies in vitro have contributed much to the knowledge of actin biology. However, interesting environmental influences common in developing systems can differentially regulate actin dynamics and organization. In this chapter, we highlight several selected examples of directed cell migration during morphogenesis, in which actin dynamics have been observed directly in live-imaging studies. We discuss similarities and differences between collective cell and single cell migration during development, and we compare what has been learned from in vivo studies in developmental systems with in vitro studies of single cells.

Contents 8.1 Movements of Cell Sheets During Morphogenesis . . . . . . . . . 8.1.1 C. elegans Ventral Enclosure – Closing Both Ends . . . . . . 8.1.2 Drosophila Dorsal Closure – Multiple Actin-based Forces Contribute to a Single Morphogenetic Process . . . . . . . . 8.1.3 Neural Crest Cell Migration – Delamination and then Cell Contact-Dependent Migratory Behaviours Position Cells . . . 8.2 Single Cell Migration During Morphogenesis . . . . . . . . . . . . 8.2.1 Zebrafish Primordial Germ Cell Migration – Single Cells Come Together to Form Cell Clusters and Migrate Together to Their Final Destination . . . . . . . . . . . . . . . . . 8.2.2 C. elegans Axon Guidance – Using a Genetic System to Identify Proteins Required for Single Cell Migration In Vivo . . . . . . 8.3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3.1 Collective Cell Migration . . . . . . . . . . . . . . . . .

. . . . . . . . . . . .

190 190

. . . . . .

192

. . . . . . . . . . . .

194 198

. . . . . .

198

. . . . . . . . . . . . . . . . . .

200 202 202

M. Roh-Johnson (B) Biology Department, UNC Chapel Hill, Chapel Hill, NC 27599-3280, USA e-mail: [email protected]

M.-F. Carlier (ed.), Actin-based Motility, DOI 10.1007/978-90-481-9301-1_8, C Springer Science+Business Media B.V. 2010

187

188

M. Roh-Johnson et al.

8.3.2 Single Cell Migration – Amoeboid Versus Mesenchymal Migration . . . . 8.3.3 What Can We Learn About Actin Dynamics in a Model Developmental System? . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

203 204 205

Actin is integral to the dynamic cellular movements and rearrangements that occur during morphogenesis, a process critical for an organism to develop its final shape. Actin filaments have structural roles in addition to roles in producing forces that can drive cell movements. There are many types of cell movements that occur during morphogenesis, including ingression (single cell migration out of an epithelium, often from the surface to the interior of the embryo), epiboly (spreading and thinning of an epithelial sheet, often to enclose the interior layers of an embryo), invagination (inward folding of a cell sheet into an embryo), involution (inward rolling of an epithelial sheet across an opening), and delamination (separation of two sheets of cells or separation of a cell from a sheet). All of these cell movements involve remodeling of the actin cytoskeleton. Studies in vitro have contributed much to the knowledge of actin biology, from the discovery of actin in muscle extracts to the observation of the delicate architecture of actin networks at the leading edge of a cell (Svitkina and Borisy, 1999). During development, there are significant variations in extracellular milieu, for example a variety of intercellular signals as well as forces exerted by cells in moving tissues, that can differentially regulate actin dynamics and organization. In this chapter, we will highlight several examples of actin-based cell migrations in morphogenesis during development. These models of cell migration are commonly used as paradigms for understanding actin dynamics while taking into account the microenvironment of the cell. Morphogenetic processes often require multiple, redundant actin-based mechanisms. Dissecting the respective contribution of each mechanism is essential to understanding the forces that drive a morphogenetic process. Cell movements require cell shape changes that are dependent on remodeling of the cytoskeleton. One example of a simple change in cell shape is apical constriction, a process in which cells narrow their apical surfaces, generally by contraction of an apical actomyosin network (Sawyer et al., 2010). Apical constriction can drive cell movements during the processes of ingression or invagination (Lee and Harland, 2007; Harris et al., 2009). For example, in C. elegans, the endodermal precursor cells Ea and Ep (referred to collectively here as Ea/p), are born on the surface of the embryo. The Ea/p cells apically constrict, driving their movement to the embryonic interior, and this movement marks the initiation of gastrulation (Lee and Goldstein, 2003; Lee et al., 2006) (Fig. 8.1a–f). Pharmacological inhibition of actin polymerization or depletion of actin regulators, such as the Arp2/3 complex, results in cell internalization defects, supporting a role for actin architecture and/or dynamics in gastrulation (Severson et al., 2002; Lee and Goldstein, 2003). As Ea/p cells internalise, neighbouring cells fill in a gap that is left behind. Observations of F-actin dynamics in vivo, using an F-actin-binding domain of moesin fused to GFP (Edwards et al., 1997), have revealed that specific neighbouring cells form

8

Roles for Actin Dynamics in Cell Movements During Development

189

Fig. 8.1 C. elegans endodermal precursor cells internalise by apical constriction. Active cell migration might also contribute to the associated cell movements. Gastrulation stage embryos are shown with Ea/p cells colored purple and marked by asterisks, and the names of neighbouring cells are indicated. (a–c) A lateral view. Ea/p cells shorten their apical surfaces through actomyosin contraction, moving toward the embryonic interior, and neighbouring cells fill in the gap (arrows). (d–f) A ventral view. A ring of six cells fill in the gap (arrows) that is left behind by internalizing Ea/p cells. (g, h) A ventral view of embryos expressing the F-actin marker GFP::MOE. F-actin is enriched in mesodermal precursors specifically where they meet Ea/p (white arrows), and not at the other neighbouring cell boundaries (black arrowheads). The germline cell, P4 , also has actin accumulation (X), which is not dynamic nor in an extension. (i) A ventral view of an embryo expressing PH domain::mCherry to visualise cell membranes. Membrane protrusions form only where mesodermal descendants contact Ea/p cells. Adapted from Lee and Goldstein (2003). Adapted from Roh-Johnson and Goldstein (2009)

dynamic, Arp2/3-dependent, F-actin-enriched extensions at their borders with Ea/p cells (Roh-Johnson and Goldstein, 2009). Interestingly, the neighbours that form these extensions comprise one side of a closing ring of cells – three of the six cells that form the ring. The role that these extensions play in gastrulation is not well understood. It is possible that the extensions are specializations for cell crawling or cell rolling, or that they participate in sealing the ring upon closure (Roh-Johnson and Goldstein, 2009; Fig. 8.1 g–i). Endoderm internalization in C. elegans involves very few cells, with only two cells internalizing and a ring of just six cells closing the gap left, yet it provides one of many examples in which multiple types of cell movements participate together in morphogenesis. The role that actin plays in these developmental processes is under active exploration. We will highlight several selected examples of directed cell migration during morphogenesis, from movement of a sheet and/or groups of cells to single cell

190

M. Roh-Johnson et al.

migration. We discuss similarities and differences between concerted cell movements and single cell migration during development, and we compare what has been learned in vivo in developmental systems with in vitro studies of single cells. We focus on examples in which actin dynamics have been observed directly in live-imaging studies, and we discuss key signalling pathways that regulate actin dynamics in actively migrating cells during morphogenesis.

8.1 Movements of Cell Sheets During Morphogenesis 8.1.1 C. elegans Ventral Enclosure – Closing Both Ends Cells can move as a sheet in dramatic rearrangements of the germ layers of an animal. In C. elegans, epidermal cells are born on the dorsal side of the animal as two rows of cells (Chisholm and Hardin, 2005). These cells intercalate (referred to as dorsal intercalation), forming a single row on the dorsal midline. After dorsal intercalation, the epidermal sheet undergoes epiboly, spreading and fully enclosing the animal as the two edges of the sheet meet on the ventral side. Ventral enclosure occurs in two phases (Williams-Masson et al., 1997) (Fig. 8.2). In the first phase, two anterior pairs of cells, termed the “leading cells”, extend long, actinrich protrusions, making contact with each other on the ventral side. In the second phase, the cells posterior to the leading cells, termed the “pocket cells”, close the remaining gap. Both the leading cells and the pocket cells are important for ventral enclosure, as perturbing either cell population by laser irradiation of individual cells results in ventral enclosure defects (Williams-Masson et al., 1997). Both the leading cells and the pocket cells form F-actin-rich structures. Live imaging of adhesion complexes shows protrusions similar to filopodia, as well as broad lamellae, from the leading cells (Raich et al., 1999). Phalloidin staining reveals that the protrusions from the leading cells are enriched with F-actin (Williams-Masson et al., 1997; Sawa et al., 2003). In addition to proposed roles for filopodia in cell motility during ventral enclosure, these actin-rich fingers may play a role in facilitating strong cell-cell adhesion after cell contact is established (Raich et al., 1999). In a process termed “filopodial priming”, α-catenin is rapidly recruited to sites where contralateral filopodial tips first make contact. This loading of a cell adhesion complex member into the tips may facilitate rapid cell-cell adhesion as the epithelium seals on the ventral side. The ventral pocket cells accumulate a continuous belt of F-actin along the edge of each cell facing the pocket. The formation of this F-actin belt suggests that a purse-string mechanism may be driving the closure of the ventral pocket, a mechanism analogous to pulling closed a drawstring bag, except that each cell’s portion of the drawstring acts as a contractile unit (Williams-Masson et al., 1997). This observation leads to a model where the leading cells that seal at the midline produce a tension that pulls the ventral pocket cells around the embryo toward the ventral side. Once the pocket cells are pulled close enough to form a ring, ventral enclosure completes by an actin purse-string mechanism (Fig. 8.2).

8

Roles for Actin Dynamics in Cell Movements During Development

191

Fig. 8.2 Schematic of C. elegans ventral enclosure. Ventral cells are drawn in pink. The first 2 pairs of cells, the leading cells, extend long protrusions and make contact with their contralateral neighbours (arrow). After the leading cells make contact, the remaining cells, termed the pocket cells, extend around the embryo and meet along the midline. Figure adapted from Chisholm and Hardin (2005)

As morphogenetic events involve multiple actin-dependent processes acting in concert, different cells must simultaneously employ different mechanisms for actin regulation. Many actin regulators are involved in ventral enclosure. Several components of the Rac signalling pathway have been implicated in this process. These include homologs of the GTPase Rac, a Rac1-associating protein (Sra), and a Nckassociating protein (HEM2/NAP1/Kette) (Lundquist et al., 2001; Soto et al., 2002; Patel et al., 2008). The ventral enclosure defects observed in Rac signalling mutants may be due to disruption of function of the Arp2/3 complex, a complex that nucleates new actin branches off pre-existing actin filaments. Indeed, the Arp2/3 complex, as well as one of its upstream activators, WASP, have been shown to regulate ventral enclosure (Severson et al., 2002; Sawa et al., 2003). Several of the Rac signalling components as well as Arp2/3 and WASP, have been shown to localise to the leading edge of the leading cells, suggesting a role for these proteins in the protrusive activity (Sawa et al., 2003). Ena/Vasp also regulates ventral enclosure, presumably through its effects on dynamics at the plus end of actin filaments (Withee et al., 2004; Sheffield et al., 2007). Thus, key actin regulators play roles in ventral enclosure

192

M. Roh-Johnson et al.

and have predictable roles in ventral enclosure. Little is known yet about the precise effects of these proteins on actin dynamics during ventral enclosure. Improving microscopy techniques for visualization of actin architecture and dynamics may allow for a greater understanding of how these key actin regulators function in this system.

8.1.2 Drosophila Dorsal Closure – Multiple Actin-based Forces Contribute to a Single Morphogenetic Process The combination of actin-based cell protrusions and actin purse-string mechanisms to drive morphogenesis is not restricted to C. elegans. In Drosophila, a process known as dorsal closure also requires both actin-rich protrusions and an actin cable. During the final phases of Drosophila embryogenesis, there is a large hole in the epidermis on the dorsal side that is covered by a squamous epithelium, the amnioserosal cells (Fig. 8.3a–e). Forces from the migrating epidermal sheet combine with the forces from the contracting amnioserosal cells to drive closure. Amnioserosal cells apically constrict, pulling the leading edge cells toward the ventral midline, and the leading edge of the migrating epidermal sheet forms a supracellular F-actin pursestring that shortens by more than 25% as the hole closes (Kiehart et al., 2000; Hutson et al., 2003). Additionally, the leading edge cells form long filopodial protrusions, approximately 10 μm long. These protrusions are thought to participate in completing dorsal closure by zipping the two edges of the epidermal sheet (Jacinto et al., 2000; Kiehart et al., 2000; Hutson et al., 2003). Zipping occurs with great precision, with cells of the same segmental position meeting on each side of the opening, and

Fig. 8.3 Drosophila dorsal closure occurs through actin-based contributions from multiple tissues. (a–d) SEMs of dorsal closure. The epidermal sheet migrates by actin-based movements, covering the hole that is filled with amnioserosal cells. (e) GFP-actin expressing embryo during dorsal closure. An actin-rich cable and filopodia form at the leading edge. (f) GFP-actin expressing embryo that has been wounded with a laser. As in the embryo in (e), an actin-rich cable and filopodia form along the epithelial front. (a–d) Images from Jacinto et al. (2002)

8

Roles for Actin Dynamics in Cell Movements During Development

193

the closed seam eventually matures into a continuous epithelium. The process of dorsal closure provides an excellent model for teasing apart the forces contributed by multiple tissue types to drive a single morphogenetic process, combining tools of genetics, live microscopic imaging of fluorescently-labelled proteins, and precise laser cuts to assess the directions and relative strengths of forces deriving from each contributing tissue. Both the actin cable and the filopodia contribute to the migration and sealing of the epidermal sheet during dorsal closure. GFP- labelled moesin or actin shows enrichment continuously along the leading edge of the epidermal sheet (Jacinto et al., 2000; Kiehart et al., 2000; Hutson et al., 2003; Reed et al., 2004). Myosin II also colocalises with actin along the leading edge and is thought to provide the force necessary for the contractile purse string mechanism (Franke et al., 2005). When a laser is used to cut the supracellular actin purse-string, the leading edge recoils from the site of injury, revealing that this cable is under tension (Kiehart et al., 2000). In Rho or myosin II mutants, the F-actin cable disassembles part way through dorsal closure (Harden et al., 1999; Bloor and Kiehart, 2002). Observing GFP-labelled actin in these mutants reveals that the leading edge is less taut, and there is an increase in the number of filopodia, which can often coalesce into broad lamellipodia (Jacinto et al., 2002). Excessive filopodial protrusions were also observed when Rac signalling was depleted (Woolner et al., 2005). Thus, in addition to the role for the actin cable as a purse string, the cable may also have a structural role, maintaining epithelial integrity as well as restraining the formation of excess protrusions. F-actin rich filopodia can act as sensory processes that investigate the environment (Mattila and Lappalainen, 2008). During dorsal closure, there is evidence that filopodia may sense their contralateral partners. This phenomenon is best visualised when GFP-actin is expressed only in 4 cell wide stripes across the embryo (Jacinto et al., 2000). GFP expressing filopodia on one epithelial front will contact filopodia on the other epithelial front, making contacts with non-GFP expressing filopodia until filopodia reach GFP-expressing filopodia. Once filopodia find their contralateral partner, they appear to draw the epithelial sheets together and align the GFP-expressing stripes (Jacinto et al., 2000). There are two pieces of evidence that suggest that filopodia tug one another (Jacinto et al., 2000). First, the rate of epithelial front movement is slower prior to filopodial engagement: 0.11 ± 0.02 μm/min (average ± SD) before filopodial contact, and 0.24 ± 0.07 μm/min after contact. Second, the actin cable can appear kinked toward the sites of filopodial contact, suggesting that a force is being exerted on the actin cable. These filopodial tethers also pull the epithelial sheet into alignment with their correct neighbours (Millard and Martin, 2008). Depleting filopodia by dominant-negative Cdc42 expression or by blocking Jun N-terminal kinase signalling reveals that dorsal closure can still proceed, but the epithelial sheet is misaligned during sealing (Jacinto et al., 2000). Similar to what is observed during C. elegans ventral enclosure, the filopodia during dorsal closure are speculated to participate in α-catenin-based filopodial priming (Jacinto et al., 2000). During Drosophila dorsal closure, rather than forming nascent adhesion complexes when the two tips of filopodia meet as in ventral enclosure,

194

M. Roh-Johnson et al.

filopodia interdigitate during dorsal closure, and adhesion occurs along the two epithelial fronts. The regulation of F-actin dynamics in this system has been investigated by dissecting the phenotypes of mutants of several actin regulators. Filopodia in tissue culture cells are known to be regulated by WASP and Scar proteins through activation of the Arp2/3 complex (Zallen et al., 2002; Pollard and Borisy, 2003). In Drosophila, SCAR is the primary activator of Arp2/3 in morphogenesis (Zallen et al., 2002); however, it is unknown whether SCAR plays a role in dorsal closure. There are several upstream activators that do play roles in dorsal closure. Four small GTPases have been shown to be involved in the enrichment of cytoskeletal machinery at the leading edge: Rho1, Rac1, Cdc42 and Ras1. Dominant negative studies suggest that these proteins have overlapping roles in regulating myosin and in actin localization to the actin cable (Harden et al., 1999). Expressing a dominant negative Rac specifically in the epidermis results in defects in myosin and actin localization along the leading edge, whereas dominant negative Cdc42 results in subtle actin and myosin localization defects (Harden et al., 1999). Cdc42 also plays a role in the formation of filopodia (Jacinto et al., 2000). Mutations in Cdc42 abolish filopodia formation, affecting the ability of the leading edge cells to sense their neighbours. Mutants of Abelson kinase (Abl) also exhibit defects in dorsal closure. In embryos expressing a constitutively active Abl kinase, filopodia are absent and replaced with broad lamellae, the actin cable is disorganised, and the cells in the two sheets do not precisely align with one another (Stevens et al., 2008). One known target of Abl is the anti-capper Ena (Gertler et al., 1990). Overexpression of Ena can rescue defects caused by Abl mutations, indicating that the roles of Abl in dorsal closure are mediated by Ena (Gates et al., 2007; Stevens et al., 2008). Furthermore, Ena localises to filopodial tips and affects filopodial dynamics. Ena mutants slow dorsal closure timing and interfere with the ability of cells to match correctly with their neighbours (Gates et al., 2007).

8.1.3 Neural Crest Cell Migration – Delamination and then Cell Contact-Dependent Migratory Behaviours Position Cells Neural crest cells are highly migratory, travelling long distances through the embryo, and they are multipotent, giving rise to many tissue types including peripheral neurons, glia, connective tissue, bone, melanocytes, and the outflow tract of the heart (Gammill and Bronner-Fraser, 2003). These “explorers of the embryo” are unique to vertebrates, arising at the border between the neural and non-neural ectoderm during closure of the neural tube (Fig. 8.4a) (Gammill and Bronner-Fraser, 2003). Although the induction and migration patterns of the neural crest have been well studied, the cues that guide cytoskeletal rearrangements important for neural crest cell migration are only beginning to be revealed. Before neural crest cells begin their migration, they segregate from the neuroepithelium by an epithelial to mesenchymal transition (EMT). During EMT in

8

Roles for Actin Dynamics in Cell Movements During Development

195

Fig. 8.4 Neural crest cells delaminate from the neural epithelium and then migrate to their final destination. (a) The neural plate border (purple) is located between two cell types, the neuroectoderm (purple) and the non-neuroectoderm (blue). During neurulation, the neural folds elevate and some neural plate cells apically constrict, forming the neural tube. The neural crest cells (grey) delaminate from the neural tube. (b) Contact inhibition of locomotion when two neural crest cells meet, and the signalling pathway regulating contact inhibition

zebrafish embryos, neural crest cells display a sequence of protrusive activities, forming blebs and then filopodial protrusions. Blebbing occurs as delamination begins, followed by the translocation of the cell soma in the direction of the bleb (Berndt et al., 2008). Actin-rich filopodia and lamellipodia then form as neural crest cells exit the neuroepithelium. In vivo imaging of actin dynamics confirms that the blebs observed on the neural crest cells are similar to blebs of other cell types, with bleb formation initiated by separation of the membrane from the F-actin network, and with actin filaments accumulating beneath the membrane as the bleb retracts (Fig. 8.5a) (Berndt et al., 2008). Similar bleb dynamics are seen, for example, in mammalian tumour cells (Wolf et al., 2003; Sahai, 2005). When the myosin inhibitor blebbistatin is added to zebrafish embryos, actin accumulation to the bleb is delayed and the blebs fail to retract, but interestingly, lamellipodia and filopodia are not affected. Thus, actomyosin contractility may regulate the dynamics of membrane blebbing in neural crest cells (Berndt et al., 2008). What signals regulate actin dynamics during EMT? Bmp signalling and Wnt signalling have been implicated in neural crest delamination and migration and have been shown to regulate key actin regulators such as the Rho GTPases (BurstynCohen et al., 2004; De Calisto et al., 2005; Groysman et al., 2008). BMP4 triggers the downregulation of N-cadherin. N-cadherin normally maintains the neural crest in a premigratory state by two mechanisms: by increasing cell adhesion and by repressing canonical Wnt signalling (Shoval et al., 2007). BMP4 also induces expression of RhoB in the dorsal midline of the neural tube, in a region where

196

M. Roh-Johnson et al.

Fig. 8.5 Cells can migrate by membrane blebs and by actin-dependent protrusions at a leading edge. (a) Neural crest bleb formation. The membrane bleb expands past the actin cortex. Then actin accumulates beneath the bleb as the bleb retracts. (b) Primordial germ cell bleb formation. The upper panel is the merge of actin in red and membrane in green. The bottom panel is actin only. Much like neural crest bleb formation, the PGC bleb is not enriched with actin near the membrane. Actin then accumulates near the membrane beneath the bleb during retraction. (c) Neural crest EMT. Actin-rich protrusions (white arrowhead) form at the leading edge of migrating neural crest cells. (d) A C. elegans HSN neuron expressing labelled actin. HSN neurons form actin-rich filopodia (white arrows) on the growth cone. (a, c) Images from Berndt et al. (2008)

the neural crest forms (Liu and Jessell, 1998). Blocking Rho activity with the C3 exotoxin in chick neural tube explants inhibits neural crest cell delamination and disrupts formation of actin stress fibers (Liu and Jessell, 1998). Pharmacological agents that block Rho kinase (ROCK) or myosin II can also decrease the number of cells undergoing EMT in zebrafish embryos (Berndt et al., 2008). These studies suggest that Rho signalling may positively regulate EMT in the neural crest. However, a recent study has shown that both in explants and in vivo, inhibition of Rho signalling enhanced emigration of the neural crest rather than preventing EMT (Groysman et al., 2008). Blocking Rho signalling with a membrane-permeable C3 enzyme in chick neural tube explants enhances cell emigration from the explants. The membrane-permeable C3 enzyme is effective at much lower concentrations than

8

Roles for Actin Dynamics in Cell Movements During Development

197

the C3 exotoxin used in earlier studies, and it is possible that the increased specificity could account for the differing result (Liu and Jessell, 1998; Groysman et al., 2008). Consistent with the newly proposed role for RhoB in preventing migration, disrupting RhoB activity by another means, with a dominant negative RhoB GTPase construct, results in fewer stress fibers and increased emigration from the neural epithelium (Groysman et al., 2008). Inhibiting ROCK activity with Y27632 also results in a similar effect: more cells emigrate, and vinculin-containing focal contacts are reduced, suggesting that Rho/ROCK is required to maintain F-actin stress fibers in neural crest progenitors before EMT (Groysman et al., 2008). Interestingly, blocking Rho or ROCK activity by either pharmacological experiments or dominant negative constructs also results in the downregulation of N-cadherin in ovo in chick, suggesting that Rho/ROCK is also involved in maintaining neural cell adhesion (Groysman et al., 2008). These studies indicate that Rho and ROCK activity have important roles in neural crest cell emigration, but directly conflicting results leave unsettled the issue of whether Rho and ROCK promote or inhibit emigration (Berndt et al., 2008; Groysman et al., 2008). After the neural crest cells undergo EMT, they follow specific migratory patterns to multiple destinations. In general, neural crest cells from the cranial region migrate in three streams from the rhombomeres to the branchial arches. Neural crest cells from the trunk regions migrate along a medial route, through the somites, or a dorsolateral route, between the ectoderm and somites (Kuriyama and Mayor, 2008). Several cytoskeletal regulators, including N-cofilin, Nedd9, Syndecan-4, and Myosin-X, affect the migratory behaviour of neural crest cells (Gurniak et al., 2005; Matthews et al., 2008; Aquino et al., 2009; Hwang et al., 2009; Nie et al., 2009). In vivo imaging of migratory patterns of various populations of the neural crest reveal that these cells can arrange in chain-like formations, with cells contacting each other through filopodia-like processes (Fig. 8.5c) (Teddy and Kulesa, 2004; Young et al., 2004; Kasemeier-Kulesa et al., 2005). Contacts made by these processes to neighbouring cells can vary from short-range contacts (10–20 μm) to remarkably long-range contacts (up to 100 μm) (Teddy and Kulesa, 2004). When a neural crest cell becomes separated from a filopodial contact in the stream, the cell appears to move in an undirected manner (Kasemeier-Kulesa et al., 2005). Although direct observation alone cannot resolve the functions of these contacts, it raises the possibility the filopodial extensions may play roles in the collective and directional migration of the neural crest. Neural crest cells have been shown to display contact inhibition of locomotion in vivo (Fig. 8.4b) (Carmona-Fontaine et al., 2008; Goldstein and Hamada, 2009). Contact inhibition of locomotion is a long standing hypothesis by which cell contacts influence the direction of cell movements: at sites where a cell contacts another cell, protrusions involved in cell migration cease forming, and protrusions form at other sites instead. Contact inhibition of locomotion was first observed in fibroblasts in vitro (Abercrombie and Heaysman, 1954; Abercrombie et al., 1957). It has since been shown that when two neural crest cells come into contact in vivo, their protrusions collapse at the site of contact, and they can change their direction of migration. This behaviour appears to be

198

M. Roh-Johnson et al.

regulated by a PCP signalling pathway, as inhibition of Dishevelled (Dsh) or classic PCP genes (Wnt11, strabismus or prickle1) prevents the collapse of lamellipodia, and these cells fail to significantly change the direction of migration upon contact (Carmona-Fontaine et al., 2008). Furthermore, Dsh is enriched at sites of cell-cell contact (Carmona-Fontaine et al., 2008). Signalling appears to work through RhoA, as RhoA is required for filopodia retraction (Rupp and Kulesa, 2007), and RhoA is active at sites of cell-cell contact (Carmona-Fontaine et al., 2008). Consistent with this, PCP signalling has been shown to activate RhoA, and this activation has an inhibitory effect on Rac activity in neural crest cells (Matthews et al., 2008). It has been proposed the contact inhibition may account for the directional migration of a stream or sheet of neural crest cells, as only the exposed end of a cell at the leading edge can extend protrusions when other sides are in contact with other cells. Other studies have shown neural crest cells with extensions at both the leading and trailing end, making simultaneous contacts in lines of cells (Teddy and Kulesa, 2004; Rupp and Kulesa, 2007). It is possible that some filopodia-like extensions at the trailing ends of cells may be retraction fibers – contacts left behind that are progressively retracted – rather than filopodia extended in this direction. Differences in neural crest cells migratory mechanisms between frog, mouse and chick are also possible. Further studies examining the formation of filopodial-like protrusions at specific times and locations and comparing experimental systems may shed more light on this issue.

8.2 Single Cell Migration During Morphogenesis 8.2.1 Zebrafish Primordial Germ Cell Migration – Single Cells Come Together to Form Cell Clusters and Migrate Together to Their Final Destination Studies in vitro predominantly examine the migration of single cells. In development, although cells often migrate as sheets or groups of cells (Friedl and Gilmour, 2009), there are also examples of individual cell migration. Primordial germ cell (PGC) migration involves both single and collective cell migration. In zebrafish, PGCs are specified at four different regions in the embryo. The four populations of PGCs migrate to the site of gonad formation within the first day of development. The fidelity of this process is demonstrated when ectopic PGCs are transplanted randomly in the embryo: Transplanted cells efficiently migrate to the appropriate location (Ciruna et al., 2002). PGCs transition to migration in three stages (Reichman-Fried et al., 2004; Blaser et al., 2005). First, the PGCs appear rounded and morphologically indistinguishable from their somatic neighbours. Second, PGCs extend protrusions in all directions but do not actively migrate. Third, approximately 1 h later, the PGCs become sensitive to directional cues provided by somatic cells secreting the chemokine SDF-1a. The PGCs begin sending out polarised protrusions (Doitsidou et al., 2002; Weidinger et al., 2002; Raz,

8

Roles for Actin Dynamics in Cell Movements During Development

199

2004), and transition into a migratory phase that requires the downregulation of E-cadherin. PGCs migrate as individual cells until they form two clusters on each side of the body axis. At these sites, the PGCs send out small protrusions and remain at the same location for approximately 3 h (Reichman-Fried et al., 2004). They then migrate as a cluster to the site where the gonad will develop. High resolution imaging of GFP driven specifically in the PGCs reveals that the clusters move by individual cell migrations, with a lack of coordinated movement within the cluster, each cell exhibiting variably-directed short-range migrations. Furthermore, close cell-cell contacts are not observed. Consistent with cells in the cluster moving independently, each cell can spend a portion of its time at the front of the cluster. The cells at the front, which may be exposed to the highest levels of SDF-1a, exhibit directed migration toward the cue. During PGC migration, PGCs cycle between two phases: A “run” phase, when they actively migrate, and a “tumble” phase, when they lose their polarity and stay stationary (Reichman-Fried et al., 2004). The tumble phase has been interpreted as a pause, in which cells may resample the environment, and reorient their polarity, which may allow cells to more readily and precisely reach their target. Phalloidin staining of fixed PGCs shows F-actin enriched in the cell cortex (Blaser et al., 2005). Live imaging of EGFP-actin fusion protein reveals that there is an enrichment of actin at the cell front during directed cell migration. However, when the cells form protrusions, each protrusion extends past the belt of actin and is not itself enriched with actin. Thus, PGCs also form membrane blebs during their migration. During the tumbling phase, the cells lose their polarity yet still continue to form membrane blebs (Fig. 8.5b). Similar to blebs observed on neural crest cells, once the bleb is in its expanded state, F-actin accumulates beneath the bleb, and the bleb retracts. Experimentally disrupting actomyosin contractility by treating embryos with the myosin inhibitor blebbistatin, or expressing dominant negative constructs to prevent the phosphorylation and activation of myosin light chain, leads to loss of membrane bleb formation, and PGC migration is impaired (Blaser et al., 2005). These results are consistent with the hypothesis that a process dependent on actomyosin contraction, perhaps cytoplasmic flow, is required for bleb formation. PGCs differ from neural crest cells in that neural crest cells form blebs only during delamination from the neural epithelium and not during long distance migration. Interestingly, among the proteins that have been found to modulate PGC migration is a viral protein. Nef is a myristoylated HIV-1 protein abundant at early stages of infection, and Nef is known to disrupt cell migration when expressed in fibroblasts. Nef affects migration by interacting with the P21-activated kinase Pak2 and down-regulating the actin filament severing activity of cofilin. Fibroblast cells expressing Nef have disorganised F-actin. Nef can also inhibit SDF-1-induced chemotaxis of T-lymphocytes (Stolp et al., 2009). PGCs expressing Nef also have altered migration patterns (Stolp et al., 2009). Whether Nef blocks the migration of PGCs by similar mechanisms as in other cell types is currently unknown. However, expression of Nef without the Pak2-interacting domain in zebrafish has no effect on PGC migration, suggesting that Nef’s interaction with Pak2 is critical in inhibiting PGC migration.

200

M. Roh-Johnson et al.

A central player in many migrating cells is the phosphoinositide 3-kinase (PI3 K) family of proteins. In Dictyostelium, phosphatidylinositol (3,4,5)-triphosphase (PIP3 ) accumulates at the leading edge in response to receptor activation (Kolsch et al., 2008). This accumulation recruits several downstream proteins that regulate actin dynamics. During zebrafish PGC migration, loss of PIP3 results in slower PGC motility and reduced filopodial-like protrusions (Dumstrei et al., 2004). However, in contrast to Dictyostelium, PIP3 is uniformly localised around the cell periphery in PGCs. Thus, although PIP3 is required for overall PGC migration, PIP3 is unlikely play a role in directional PGC migration.

8.2.2 C. elegans Axon Guidance – Using a Genetic System to Identify Proteins Required for Single Cell Migration In Vivo Axon outgrowth is a classic example of single cell migration during morphogenesis. Axon outgrowth is led by the guidance of the growth cone. Growth cone guidance in vivo is an excellent paradigm to study how a cell responds to cues in its extracellular environment, and specifically how the cell remodels its actin cytoskeleton to respond appropriately to this cue. Growth cone guidance is studied in a variety of systems. Among these, C. elegans is an ideal genetic system to tease apart the signalling pathways that regulate the cytoskeleton in axon guidance because C. elegans lends itself readily to genetics and RNAi, and loss of many of the worm’s 302 neurons produces behavioural phenotypes in otherwise viable, fertile strains of worms. Growth cone guidance is mediated by filopodial and lamellipodial dynamics that are driven by actin dynamics. Growth cones produce these protrusions, which make contact with substrates and in turn function in propelling the growth cone forward. Many actin regulators control the formation of these actin-based structures (Fig. 8.6). In C. elegans, Arp2/3 activation, abLIM/UNC-115, and Ena/UNC-34 directly regulate actin dynamics (Lundquist et al., 1998; Yu et al., 2002; Struckhoff and Lundquist, 2003; Norris et al., 2009). Ena/UNC-34 also genetically and biochemically interacts with the single C. elegans lamellipodin (Lpd) homolog, MIG-10 (Chang et al., 2006). RhoG/MIG-2, Rac/RAC-2, and Rac/CED-10 act redundantly for axon guidance, and the Nck-interacting kinase (NIK) MIG-15 functions in all three Rac signalling pathways (Shakir et al., 2006). Thus, NIK/MIG-15 is a core component of each signalling pathway. The Rho GTPases and their upstream activators act as modulators for specificity. For example, the Rho GTPases RhoG/MIG-2 and Rac/CED-10 are regulated by the guanine nucleotide exchange factors (GEFs) Trio/UNC-73 and DOCK180/CED-5, respectively (Steven et al., 1998; Lundquist et al., 2001; Wu et al., 2002). Furthermore, genetic analysis indicates that RhoG/MIG-2 is in the same pathway as the upstream activator WASP/WSP-1, while Rac/CED-10 is in the same pathway as the upstream activator Wave/WVE-1 (Shakir et al., 2008). Both Rac GTPases converge on Sra-1/GEX2 and Kette/GEX-3 and regulate Arp2/3 function (Shakir et al., 2008). Thus, taken together, there are three pathways that regulate Arp2/3 (Fig. 8.6). The components of

8

Roles for Actin Dynamics in Cell Movements During Development

201

Fig. 8.6 Several pathways regulate actin during C. elegans axon outgrowth

these pathways again highlight the idea that there are several core components that are used in each pathway to elicit a response (e.g. Sra-1/GEX-2 and Kette/GEX-3), and the specificity of each pathway is then dictated by specific Rho GTPases and upstream regulators. There is also crosstalk between major pathways, as Rac/CED10 can function through abLIM/UNC-115 (Struckhoff and Lundquist, 2003). The Arp2/3 complex itself has also been shown to have roles in neuronal migration. Recently it was shown that depleting C. elegans of Arp2/3 results in defects in mechanosensory neuron migration (Schmidt et al., 2009). Growth cones respond to signals in their extracellular environment and alter actin dynamics in response. One such signal in C. elegans is the Netrin homolog UNC-6. Netrin/UNC-6 is a conserved axon guidance cue. A motor neuron, HSN, responds to Netrin/UNC-6 by asymmetrically localizing the receptor DCC/UNC-40 toward the direction of the signal (Adler et al., 2006). Lpd/MIG-10 also localises asymmetrically in the growth cone in response to Netrin/UNC-6, through the activity of Rac/CED-10 (Chang et al., 2006; Quinn et al., 2008). This asymmetric Lpd/MIG-10 localization is also coincident with asymmetric F-actin accumulation (Quinn et al., 2008). Plasma membrane markers can reveal projections, or neurites, from the cell body of developing HSN neurons (Fig. 8.5d) (Adler et al., 2006). These neurites are F-actin rich, and form toward the Netrin/UNC-6 cue. The HSN neuron has a clear leading edge, and filopodia and lamellipodia grow and retract in the direction of the signal (Adler et al., 2006). Defects caused by increased Netrin/UNC-6 signalling are suppressed in loss-of-function mutations in Rac/CED-10, Ena/UNC-34, and abLIM/UNC-115, suggesting that growth cone outgrowth and turning by Netrin/UNC-6 signals are mediated by these cytoskeletal regulators (Gitai et al., 2003). Interestingly, although filopodia are present on all

202

M. Roh-Johnson et al.

growth cones, suggesting that the formation of these F-actin rich structures is likely to be critical for axon guidance, lack of filopodia in ena/unc-34 mutants is consistent with proper HSN guidance (Chang et al., 2006). In these cells, lamellipodia still form. Thus, in vivo, it appears that dynamic filopodia form, but at least in some contexts are dispensable for guidance, and perhaps other cues in the extracellular milieu stimulate alternative migratory mechanisms, most likely via lamellipodia. The growth of an axon is important for guidance, but the inhibition of outgrowth is also important for precision. Although many proteins function intracellularly to promote axon outgrowth, there are few proteins yet known to negatively regulate this process. CRML-1, the C. elegans CARMIL homolog, was identified in C. elegans to inhibit axon outgrowth by affecting Trio/UNC-73 activity, although mammalian CARMIL acts to promote glioblastoma migration (Yang et al., 2005; Vanderzalm et al., 2009). CRML-1 and Trio/UNC-73 physically interact, and together control the direction of growth cone migration by altering the levels of a guidance receptor, Robo/SAX-3. Thus, through the inhibition of Rac signalling, CRML-1 can negatively regulate neuronal migration.

8.3 Conclusions 8.3.1 Collective Cell Migration This chapter discusses three different modes of collective migration: migration of epithelial sheets, cell clusters, and cell streaming. Interestingly, actin-based cell migrations during morphogenesis often occur through collective cell migration instead of single cell migration (Friedl and Gilmour, 2009). Why do cells in development so often migrate in groups? One hypothesis is that cell clusters can in general generate more force than can single cells (Kolega et al., 1982). C. elegans ventral enclosure and Drosophila dorsal closure both involve multiple actin-dependent cell movements. Different force-generating mechanisms are evident during Drosophila dorsal closure including actomyosin contraction of the amnioserosal cells, a supracellular purse string at the leading edge, and dynamic filopodia. These are coordinated spatially and temporally, regulating a single morphogenetic process (Hutson et al., 2003). This process is similar to C. elegans ventral enclosure, which involves a combination of forces from actively migrating leading cells and an actin purse string-like mechanism in the pocket cells. In both C. elegans ventral enclosure and Drosophila dorsal closure, filopodia aid in closing a ring. Actin purse-string mechanisms and filopodia formation are involved in both of these epibolic movements, and key molecular components may be conserved. The Arp2/3 complex, a major actin regulator, is required for C. elegans ventral enclosure (Sawa et al., 2003), but no role for Arp2/3 has been described for Drosophila dorsal closure. While upstream Arp2/3 activators such as Wave may play a role during Drosophila dorsal closure, it is possible that other actin nucleators, such as formins or Spire, may play a role in dorsal closure, and that Arp2/3 may be

8

Roles for Actin Dynamics in Cell Movements During Development

203

acting redundantly with these players. Indeed, a formin, Diaphanous, localises to the actin cable, and embryos expressing constitutively active diaphanous exhibit dorsal closure defects, mainly through defects in amnioserosal cell apical constriction (Homem and Peifer, 2008). The differences observed between C. elegans ventral enclosure and Drosophila dorsal closure suggest that there is some plasticity in mechanisms of regulation through evolution, with different inputs acting on a common outcome. Factors that are required to prevent filopodia formation and migration are also important for dorsal closure and possibly ventral enclosure. During the last stages of dorsal closure, the two epithelial leading edges must recognise each other and cease active migration. It is possible that apposition of the two edges of the migrating epithelium during dorsal closure results in contact inhibition, preventing the overmigration of the leading edges. Contact inhibition in neural crest cells is regulated by PCP/non-canonical Wnt signalling. When two neural crest cells contact each other, Dishevelled becomes localised to the membrane at areas of cell-cell contact, and RhoA becomes active (Carmona-Fontaine et al., 2008). RhoA is thought to direct the collapse of filopodia at the cell contact zones and aid in the change of migratory direction. However, during dorsal closure, when filopodial tips touch, the filopodia do not retract immediately (Jacinto et al., 2000). Rather they appear to contact each other, perhaps tethering the edges of the epithelial sheet and also aligning them. It will be interesting to determine if contact inhibition does occur during Drosophila dorsal closure and C. elegans ventral enclosure, perhaps via a different mechanism than in neural crest cells.

8.3.2 Single Cell Migration – Amoeboid Versus Mesenchymal Migration The mechanism of bleb formation appears to be different between neural crest cells and PGCs. PGCs require local actomyosin contraction, which produces cytoplasmic flow, a flow that may contribute to formation of a membrane bleb. When PGCs are treated with blebbistatin, the membrane blebs do not form (Blaser et al., 2005). Neural crest cells, on the other hand, can still form membrane blebs when treated with blebbistatin, suggesting that actomyosin contraction is not required for bleb formation, but is required for bleb retraction (Berndt et al., 2008). It is possible that this difference in bleb formation may account for the difference in long distance migration mechanisms between neural crest cells and PGCs. Neural crest cells exhibit blebs during delamination from the neural epithelium. They then adopt characteristics of a mesenchymal cell with a clear leading edge, actively migrating over longer distances. PGCs, on the other hand, form membrane blebs during long-range active migration. Thus, unlike what is seen in many other systems where actin polymerization produces the force to form and extend a pseudopod during migration, zebrafish PGCs have adopted a different form of motility.

204

M. Roh-Johnson et al.

Why do PGCs actively migrate by membrane bleb formation rather than actin polymerization-induced protrusions? The fact that cells can convert from amoeboid to mesenchymal forms of movement, and vice versa, suggests that a cell can change its migratory behaviour in response to its environment. It has been shown that blebdependent motility occurs as a result of changes in cell contacts or cell-cell adhesion (Shook and Keller, 2003). Rather than making contacts with an underlying substratum, bleb-dependent motility without attachment to a substrate allows cells to squeeze past obstacles and navigate through matrices (Hegerfeldt et al., 2002; Gadea et al., 2007; Tournaviti et al., 2007). This form of motility is similar to amoeboid motility. Cancer cells have also adopted this amoeboid form of motility, bypassing requirements for matrix metalloproteases in migration (Friedl, 2004; Sahai, 2005; Wyckoff et al., 2006). It is possible that PGCs have also adopted this form of motility to bypass a requirement for adhesion-based mechanisms.

8.3.3 What Can We Learn About Actin Dynamics in a Model Developmental System? Many actin regulators play conserved roles between cell migrations in morphogenesis and cell migrations in vitro. Components of the Rac signalling pathway, as well as key actin regulators such as Ena, are involved in actin dynamics in diverse systems. There are, however, some clear differences between in vivo and in vitro studies. Notably, filopodia in C. elegans growth cones are dispensable for axon outgrowth in vivo. This is markedly different than the proposed function for filopodia during axon outgrowth in vitro (Drees and Gertler, 2008). It is possible that there are other factors that are present in the extracellular milieu of an animal that could be providing a redundant role or providing an alternative mechanism for axon guidance. The strength of analyzing actin dynamics during morphogenesis is that one can understand the role of actin in its native environment. Morphogenetic processes seldom involve a single actin-based force-generating mechanism. More often, a morphogenetic process requires multiple, redundant actin-based mechanisms. Dissecting the contribution of each actin-dependent process is an important step for which developmental systems provide invaluable models. Drosophila dorsal closure, for example, is a powerful system for measuring the force contributions of each actin-dependent mechanism in a single morphogenetic process. Laser cut experiments have revealed that that the supracellular purse-string at the leading edge and contraction of the amnioserosa contribute to most of the forces required for dorsal closure (Hutson et al., 2003). The forces provided by the filopodia at the leading edge appear essential only for the late stages of closure. Thus, analyzing actin-dependent forces during morphogenesis allows for the understanding of how cells and tissues coordinate their forces and how these forces are regulated in space and time.

8

Roles for Actin Dynamics in Cell Movements During Development

205

These types of force studies need not be limited to dorsal closure. C. elegans is an attractive model for applying laser microsurgery to analyze the contributions of multiple actin-dependent processes during endodermal internalization and ventral enclosure. Similar cell movements and actin-based structures can also be found during wound healing. When Drosophila or Xenopus embryos are wounded with a needle, the leading edge cells surrounding the wound form a supracellular actin cable as well as filopodia (Wood et al., 2002; Clark et al., 2009; Martin and Parkhurst, 2004) (Fig. 8.3f). The wound heals in part by an actin purse-string mechanism. When the wound size is sufficiently decreased, filopodia can reach across the wound. The filopodia then form tethers with one another and appear to facilitate wound closure. Teasing apart the forces in these processes is an important step to understanding the mechanisms of cell and tissue movements. Actin dynamics have only recently been analyzed in real time in several model systems. Given the optical clarity and low light scattering in embryos of some model systems like zebrafish and C. elegans, actin dynamics can be imaged readily during a number of morphogenetic events even inside embryos. Furthermore, with the development of technology to image cells deep within an animal while minimizing toxic effects of intense laser energy (Condeelis and Segall, 2003), actin dynamics in a host of cells can be imaged in their native environments. The future of this research will likely involve interdisciplinary approaches combining in vitro and in vivo studies. Acknowledgements We thank Mark Peifer, David Reiner, Stephanie Nowotarski, Erik Lundquist, and members of the Goldstein lab for comments on the manuscript. Work in the Goldstein lab on actin dynamics in morphogenesis is supported by NIH R01 GM083071.

References Abercrombie M, Heaysman JE. 1954. Observations on the social behaviour of cells in tissue culture. II. monolayering of fibroblasts. Exp Cell Res 6:293–306. Abercrombie M, Heaysman JE, Karthauser HM. 1957. Social behaviour of cells in tissue culture. III. mutual influence of sarcoma cells and fibroblasts. Exp Cell Res 13:276–291. Adler CE, Fetter RD, Bargmann CI. 2006. UNC-6/netrin induces neuronal asymmetry and defines the site of axon formation. Nat Neurosci 9:511–518. Aquino JB, Lallemend F, Marmigere F, Adameyko II, Golemis EA, Ernfors P. 2009. The retinoic acid inducible Cas-family signaling protein Nedd9 regulates neural crest cell migration by modulating adhesion and actin dynamics. Neuroscience 162:1106–1119. Berndt JD, Clay MR, Langenberg T, Halloran MC. 2008. Rho-kinase and myosin II affect dynamic neural crest cell behaviors during epithelial to mesenchymal transition in vivo. Dev Biol 324:236–244. Blaser H, Eisenbeiss S, Neumann M, Reichman-Fried M, Thisse B, Thisse C, Raz E. 2005. Transition from non-motile behaviour to directed migration during early PGC development in zebrafish. J Cell Sci 118:4027–4038. Bloor JW, Kiehart DP. 2002. Drosophila RhoA regulates the cytoskeleton and cell-cell adhesion in the developing epidermis. Development 129:3173–3183. Burstyn-Cohen T, Stanleigh J, Sela-Donenfeld D, Kalcheim C. 2004. Canonical Wnt activity regulates trunk neural crest delamination linking BMP/noggin signaling with G1/S transition. Development 131:5327–5339.

206

M. Roh-Johnson et al.

Carmona-Fontaine C, Matthews HK, Kuriyama S, Moreno M, Dunn GA, Parsons M, Stern CD, Mayor R. 2008. Contact inhibition of locomotion in vivo controls neural crest directional migration. Nature 456:957–961. Chang C, Adler CE, Krause M, Clark SG, Gertler FB, Tessier-Lavigne M, Bargmann CI. 2006. MIG-10/lamellipodin and AGE-1/PI3K promote axon guidance and outgrowth in response to slit and netrin. Curr Biol 16:854–862. Chisholm AD, Hardin J. 2005. Epidermal morphogenesis. WormBook:1–22. Ciruna B, Weidinger G, Knaut H, Thisse B, Thisse C, Raz E, Schier AF. 2002. Production of maternal-zygotic mutant zebrafish by germ-line replacement. Proc Natl Acad Sci USA 99:14919–14924. Clark AG, Miller AL, Vaughan E, Yu HY, Penkert R, Bement WM. 2009. Integration of single and multicellular wound responses. Curr Biol 19:1389–1395. Condeelis J, Segall JE. 2003. Intravital imaging of cell movement in tumours. Nat Rev Cancer 3:921–930. De Calisto J, Araya C, Marchant L, Riaz CF, Mayor R. 2005. Essential role of non-canonical Wnt signalling in neural crest migration. Development 132:2587–2597. Doitsidou M, Reichman-Fried M, Stebler J, Koprunner M, Dorries J, Meyer D, Esguerra CV, Leung T, Raz E. 2002. Guidance of primordial germ cell migration by the chemokine SDF-1. Cell 111:647–659. Drees F, Gertler FB. 2008. Ena/VASP: proteins at the tip of the nervous system. Curr Opin Neurobiol 18:53–59. Dumstrei K, Mennecke R, Raz E. 2004. Signaling pathways controlling primordial germ cell migration in zebrafish. J Cell Sci 117:4787–4795. Edwards KA, Demsky M, Montague RA, Weymouth N, Kiehart DP. 1997. GFP-moesin illuminates actin cytoskeleton dynamics in living tissue and demonstrates cell shape changes during morphogenesis in Drosophila. Dev Biol 191:103–117. Franke JD, Montague RA, Kiehart DP. 2005. Nonmuscle myosin II generates forces that transmit tension and drive contraction in multiple tissues during dorsal closure. Curr Biol 15:2208–2221. Friedl P. 2004. Dynamic imaging of cellular interactions with extracellular matrix. Histochem Cell Biol 122:183–190. Friedl P, Gilmour D. 2009. Collective cell migration in morphogenesis, regeneration and cancer. Nat Rev Mol Cell Biol 10:445–457. Gadea G, de Toledo M, Anguille C, Roux P. 2007. Loss of p53 promotes RhoA-ROCK-dependent cell migration and invasion in 3D matrices. J Cell Biol 178:23–30. Gammill LS, Bronner-Fraser M. 2003. Neural crest specification: migrating into genomics. Nat Rev Neurosci 4:795–805. Gates J, Mahaffey JP, Rogers SL, Emerson M, Rogers EM, Sottile SL, Van Vactor D, Gertler FB, Peifer M. 2007. Enabled plays key roles in embryonic epithelial morphogenesis in Drosophila. Development 134:2027–2039. Gertler FB, Doctor JS, Hoffmann FM. 1990. Genetic suppression of mutations in the Drosophila abl proto-oncogene homolog. Science 248:857–860. Gitai Z, Yu TW, Lundquist EA, Tessier-Lavigne M, Bargmann CI. 2003. The netrin receptor UNC40/DCC stimulates axon attraction and outgrowth through enabled and, in parallel, Rac and UNC-115/AbLIM. Neuron 37:53–65. Goldstein B, Hamada H. 2009. Shape meets polarity in Japan. Development 136:2487–2492. Groysman M, Shoval I, Kalcheim C. 2008. A negative modulatory role for rho and rho-associated kinase signaling in delamination of neural crest cells. Neural Dev 3:27. Gurniak CB, Perlas E, Witke W. 2005. The actin depolymerizing factor n-cofilin is essential for neural tube morphogenesis and neural crest cell migration. Dev Biol 278:231–241. Harden N, Ricos M, Ong YM, Chia W, Lim L. 1999. Participation of small GTPases in dorsal closure of the Drosophila embryo: distinct roles for Rho subfamily proteins in epithelial morphogenesis. J Cell Sci 112 ( Pt 3):273–284. Harris TJ, Sawyer JK, Peifer M. 2009. How the cytoskeleton helps build the embryonic body plan: models of morphogenesis from Drosophila. Curr Top Dev Biol 89:55–85.

8

Roles for Actin Dynamics in Cell Movements During Development

207

Hegerfeldt Y, Tusch M, Brocker EB, Friedl P. 2002. Collective cell movement in primary melanoma explants: plasticity of cell-cell interaction, beta1-integrin function, and migration strategies. Cancer Res 62:2125–2130. Homem CC, Peifer M. 2008. Diaphanous regulates myosin and adherens junctions to control cell contractility and protrusive behavior during morphogenesis. Development 135:1005–1018. Hutson MS, Tokutake Y, Chang MS, Bloor JW, Venakides S, Kiehart DP, Edwards GS. 2003. Forces for morphogenesis investigated with laser microsurgery and quantitative modeling. Science 300:145–149. Hwang YS, Luo T, Xu Y, Sargent TD. 2009. Myosin-X is required for cranial neural crest cell migration in Xenopus laevis. Dev Dyn 238:2522–2529. Jacinto A, Wood W, Balayo T, Turmaine M, Martinez-Arias A, Martin P. 2000. Dynamic actinbased epithelial adhesion and cell matching during Drosophila dorsal closure. Curr Biol 10:1420–1426. Jacinto A, Woolner S, Martin P. 2002. Dynamic analysis of dorsal closure in Drosophila: from genetics to cell biology. Dev Cell 3:9–19. Kasemeier-Kulesa JC, Kulesa PM, Lefcort F. 2005. Imaging neural crest cell dynamics during formation of dorsal root ganglia and sympathetic ganglia. Development 132:235–245. Kiehart DP, Galbraith CG, Edwards KA, Rickoll WL, Montague RA. 2000. Multiple forces contribute to cell sheet morphogenesis for dorsal closure in Drosophila. J Cell Biol 149: 471–490. Kolega J, Shure MS, Chen WT, Young ND. 1982. Rapid cellular translocation is related to close contacts formed between various cultured cells and their substrata. J Cell Sci 54:23–34. Kolsch V, Charest PG, Firtel RA. 2008. The regulation of cell motility and chemotaxis by phospholipid signaling. J Cell Sci 121:551–559. Kuriyama S, Mayor R. 2008. Molecular analysis of neural crest migration. Philos Trans R Soc Lond B Biol Sci 363:1349–1362. Lee JY, Goldstein B. 2003. Mechanisms of cell positioning during C. elegans gastrulation. Development 130:307–320. Lee JY, Harland RM. 2007. Actomyosin contractility and microtubules drive apical constriction in xenopus bottle cells. Dev Biol 311:40–52. Lee JY, Marston DJ, Walston T, Hardin J, Halberstadt A, Goldstein B. 2006. Wnt/Frizzled signaling controls C. elegans gastrulation by activating actomyosin contractility. Curr Biol 16:1986–1997. Liu JP, Jessell TM. 1998. A role for rhoB in the delamination of neural crest cells from the dorsal neural tube. Development 125:5055–5067. Lundquist EA, Herman RK, Shaw JE, Bargmann CI. 1998. UNC-115, a conserved protein with predicted LIM and actin-binding domains, mediates axon guidance in C. elegans. Neuron 21:385–392. Lundquist EA, Reddien PW, Hartwieg E, Horvitz HR, Bargmann CI. 2001. Three C. elegans Rac proteins and several alternative Rac regulators control axon guidance, cell migration and apoptotic cell phagocytosis. Development 128:4475–4488. Martin P, Parkhurst SM. 2004. Parallels between tissue repair and embryo morphogenesis. Development 131:3021–3034. Matthews HK, Marchant L, Carmona-Fontaine C, Kuriyama S, Larrain J, Holt MR, Parsons M, Mayor R. 2008. Directional migration of neural crest cells in vivo is regulated by Syndecan4/Rac1 and non-canonical Wnt signaling/RhoA. Development 135:1771–1780. Mattila PK, Lappalainen P. 2008. Filopodia: molecular architecture and cellular functions. Nat Rev Mol Cell Biol 9:446–454. Millard TH, Martin P. 2008. Dynamic analysis of filopodial interactions during the zippering phase of drosophila dorsal closure. Development 135:621–626. Nie S, Kee Y, Bronner-Fraser M. 2009. Myosin-X is critical for migratory ability of Xenopus cranial neural crest cells. Dev Biol 335:132–142.

208

M. Roh-Johnson et al.

Norris AD, Dyer JO, Lundquist EA. 2009. The Arp2/3 complex, UNC-115/abLIM, and UNC34/Enabled regulate axon guidance and growth cone filopodia formation in Caenorhabditis elegans. Neural Dev 4:38. Patel FB, Bernadskaya YY, Chen E, Jobanputra A, Pooladi Z, Freeman KL, Gally C, Mohler WA, Soto MC. 2008. The WAVE/SCAR complex promotes polarized cell movements and actin enrichment in epithelia during C. elegans embryogenesis. Dev Biol 324:297–309. Pollard TD, Borisy GG. 2003. Cellular motility driven by assembly and disassembly of actin filaments. Cell 112:453–465. Quinn CC, Pfeil DS, Wadsworth WG. 2008. CED-10/Rac1 mediates axon guidance by regulating the asymmetric distribution of MIG-10/lamellipodin. Curr Biol 18:808–813. Raich WB, Agbunag C, Hardin J. 1999. Rapid epithelial-sheet sealing in the Caenorhabditis elegans embryo requires cadherin-dependent filopodial priming. Curr Biol 9:1139–1146. Raz E. 2004. Guidance of primordial germ cell migration. Curr Opin Cell Biol 16:169–173. Reed BH, Wilk R, Schock F, Lipshitz HD. 2004. Integrin-dependent apposition of Drosophila extraembryonic membranes promotes morphogenesis and prevents anoikis. Curr Biol 14: 372–380. Reichman-Fried M, Minina S, Raz E. 2004. Autonomous modes of behavior in primordial germ cell migration. Dev Cell 6:589–596. Rupp PA, Kulesa PM. 2007. A role for RhoA in the two-phase migratory pattern of post-otic neural crest cells. Dev Biol 311:159–171. Sahai E. 2005. Mechanisms of cancer cell invasion. Curr Opin Genet Dev 15:87–96. Sawa M, Suetsugu S, Sugimoto A, Miki H, Yamamoto M, Takenawa T. 2003. Essential role of the C. elegans Arp2/3 complex in cell migration during ventral enclosure. J Cell Sci 116: 1505–1518. Sawyer JM, Harrell JR, Shemer G, Sullivan-Brown J, Roh-Johnson M, Goldstein B. 2010. Apical constriction: A cell shape change that can drive cell morphogenesis. Dev Biol 341:5–19. Schmidt KL, Marcus-Gueret N, Adeleye A, Webber J, Baillie D, Stringham EG. 2009. The cell migration molecule UNC-53/NAV2 is linked to the ARP2/3 complex by ABI-1. Development 136:563–574. Severson AF, Baillie DL, Bowerman B. 2002. A Formin Homology protein and a profilin are required for cytokinesis and Arp2/3-independent assembly of cortical microfilaments in C. elegans. Curr Biol 12:2066–2075. Shakir MA, Gill JS, Lundquist EA. 2006. Interactions of UNC-34 Enabled with Rac GTPases and the NIK kinase MIG-15 in Caenorhabditis elegans axon pathfinding and neuronal migration. Genetics 172:893–913. Shakir MA, Jiang K, Struckhoff EC, Demarco RS, Patel FB, Soto MC, Lundquist EA. 2008. The Arp2/3 activators WAVE and WASP have distinct genetic interactions with Rac GTPases in Caenorhabditis elegans axon guidance. Genetics 179:1957–1971. Sheffield M, Loveless T, Hardin J, Pettitt J. 2007. C. elegans Enabled exhibits novel interactions with N-WASP, Abl, and cell-cell junctions. Curr Biol 17:1791–1796. Shook D, Keller R. 2003. Mechanisms, mechanics and function of epithelial-mesenchymal transitions in early development. Mech Dev 120:1351–1383. Shoval I, Ludwig A, Kalcheim C. 2007. Antagonistic roles of full-length N-cadherin and its soluble BMP cleavage product in neural crest delamination. Development 134:491–501. Soto MC, Qadota H, Kasuya K, Inoue M, Tsuboi D, Mello CC, Kaibuchi K. 2002. The GEX-2 and GEX-3 proteins are required for tissue morphogenesis and cell migrations in C. elegans. Genes Dev 16:620–632. Steven R, Kubiseski TJ, Zheng H, Kulkarni S, Mancillas J, Ruiz Morales A, Hogue CW, Pawson T, Culotti J. 1998. UNC-73 activates the Rac GTPase and is required for cell and growth cone migrations in C. elegans. Cell 92:785–795. Stevens TL, Rogers EM, Koontz LM, Fox DT, Homem CC, Nowotarski SH, Artabazon NB, Peifer M. 2008. Using Bcr-Abl to examine mechanisms by which abl kinase regulates morphogenesis in Drosophila. Mol Biol Cell 19:378–393.

8

Roles for Actin Dynamics in Cell Movements During Development

209

Stolp B, Reichman-Fried M, Abraham L, Pan X, Giese SI, Hannemann S, Goulimari P, Raz E, Grosse R, Fackler OT. 2009. HIV-1 Nef interferes with host cell motility by deregulation of Cofilin. Cell Host Microbe 6:174–186. Struckhoff EC, Lundquist EA. 2003. The actin-binding protein UNC-115 is an effector of Rac signaling during axon pathfinding in C. elegans. Development 130:693–704. Svitkina TM, Borisy GG. 1999. Arp2/3 complex and actin depolymerizing factor/cofilin in dendritic organization and treadmilling of actin filament array in lamellipodia. J Cell Biol 145:1009–1026. Teddy JM, Kulesa PM. 2004. In vivo evidence for short- and long-range cell communication in cranial neural crest cells. Development 131:6141–6151. Tournaviti S, Hannemann S, Terjung S, Kitzing TM, Stegmayer C, Ritzerfeld J, Walther P, Grosse R, Nickel W, Fackler OT. 2007. SH4-domain-induced plasma membrane dynamization promotes bleb-associated cell motility. J Cell Sci 120:3820–3829. Vanderzalm PJ, Pandey A, Hurwitz ME, Bloom L, Horvitz HR, Garriga G. 2009. C. elegans CARMIL negatively regulates UNC-73/Trio function during neuronal development. Development 136:1201–1210. Weidinger G, Wolke U, Koprunner M, Thisse C, Thisse B, Raz E. 2002. Regulation of zebrafish primordial germ cell migration by attraction towards an intermediate target. Development 129:25–36. Williams-Masson EM, Malik AN, Hardin J. 1997. An actin-mediated two-step mechanism is required for ventral enclosure of the C. elegans hypodermis. Development 124:2889–2901. Withee J, Galligan B, Hawkins N, Garriga G. 2004. Caenorhabditis elegans WASP and Ena/VASP proteins play compensatory roles in morphogenesis and neuronal cell migration. Genetics 167:1165–1176. Wolf K, Mazo I, Leung H, Engelke K, von Andrian UH, Deryugina EI, Strongin AY, Brocker EB, Friedl P. 2003. Compensation mechanism in tumor cell migration: mesenchymal-amoeboid transition after blocking of pericellular proteolysis. J Cell Biol 160:267–277. Wood W, Jacinto A, Grose R, Woolner S, Gale J, Wilson C, Martin P. 2002. Wound healing recapitulates morphogenesis in Drosophila embryos. Nat Cell Biol 4:907–912. Woolner S, Jacinto A, Martin P. 2005. The small GTPase Rac plays multiple roles in epithelial sheet fusion – dynamic studies of Drosophila dorsal closure. Dev Biol 282:163–173. Wu YC, Cheng TW, Lee MC, Weng NY. 2002. Distinct rac activation pathways control Caenorhabditis elegans cell migration and axon outgrowth. Dev Biol 250:145–155. Wyckoff JB, Pinner SE, Gschmeissner S, Condeelis JS, Sahai E. 2006. ROCK- and myosindependent matrix deformation enables protease-independent tumor-cell invasion in vivo. Curr Biol 16:1515–1523. Yang C, Pring M, Wear MA, Huang M, Cooper JA, Svitkina TM, Zigmond SH. 2005. Mammalian CARMIL inhibits actin filament capping by capping protein. Dev Cell 9:209–221. Young HM, Bergner AJ, Anderson RB, Enomoto H, Milbrandt J, Newgreen DF, Whitington PM. 2004. Dynamics of neural crest-derived cell migration in the embryonic mouse gut. Dev Biol 270:455–473. Yu TW, Hao JC, Lim W, Tessier-Lavigne M, Bargmann CI. 2002. Shared receptors in axon guidance: SAX-3/Robo signals via UNC-34/Enabled and a Netrin-independent UNC-40/DCC function. Nat Neurosci 5:1147–1154. Zallen JA, Cohen Y, Hudson AM, Cooley L, Wieschaus E, Schejter ED. 2002. SCAR is a primary regulator of Arp2/3-dependent morphological events in Drosophila. J Cell Biol 156:689–701.

Part II

Molecular Aspects

Chapter 9

Regulation of the Cytoplasmic Actin Monomer Pool in Actin-based Motility Pekka Lappalainen, Maarit Makkonen, and Hongxia Zhao

Abstract Actin-based motility is driven by spatially and temporally coordinated rapid actin polymerization. This relies on the precise regulation of the size, localization and dynamics of the cytoplasmic actin monomer pool by proteins that control different aspects of actin dynamics. In dynamic actin filament arrays, the “old” actin filaments are severed and depolymerized by ADF/cofilin and gelsolin family proteins, and the newly depolymerized actin monomers are subsequently recycled for new rounds of polymerization with the aid of actin monomer-binding proteins such as profilin and cyclase-associated protein. Furthermore, actin monomer sequestering proteins, such as twinfilin and β-thymosins control the size, nucleotide status, and dynamics of the actin monomer pool in cells. Actin polymerization at filament barbed ends is also controlled by capping proteins that “funnel” polymerizationcompetent actin monomers to a subset of filament ends at the sites of rapid actin assembly. In this chapter, we discuss the mechanisms by which these proteins control the cytoplasmic actin monomer pool and how their activities are regulated in cells to promote various actin-based motile processes.

Contents 9.1 Introduction . . . . . . . . . . . . . . 9.2 Actin Filament Disassembly . . . . . . 9.2.1 ADF/Cofilin Family Proteins . . . 9.2.2 Gelsolin Family Proteins . . . . . 9.3 Actin Monomer-Binding Proteins . . . . 9.3.1 Profilin . . . . . . . . . . . . . 9.3.2 CAP (Cyclase-Associated-Protein)

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

214 215 215 219 219 220 221

P. Lappalainen (B) Institute of Biotechnology, University of Helsinki, P.O. Box 56, 00014 Helsinki, Finland e-mail: [email protected]

M.-F. Carlier (ed.), Actin-based Motility, DOI 10.1007/978-90-481-9301-1_9, C Springer Science+Business Media B.V. 2010

213

214 9.3.3 Twinfilin . . . . . . . . . . . . . . 9.3.4 β-thymosins . . . . . . . . . . . . 9.4 Actin Filament Capping Proteins . . . . . . 9.4.1 Heterodimeric Capping Protein . . . . 9.4.2 Eps8 . . . . . . . . . . . . . . . . 9.4.3 Tropomodulin . . . . . . . . . . . . 9.4.4 Other Actin Filament Capping Proteins 9.5 Conclusions and Future Prospectives . . . . References . . . . . . . . . . . . . . . . . .

P. Lappalainen et al.

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

222 223 223 224 226 226 227 227 228

9.1 Introduction The actin cytoskeleton has fundamental role in a number of cellular processes involving membrane dynamics. These include for example cell morphogenesis, motility, endocytosis, and phagocytosis. The force required to generate membrane protrusions or invaginations in these processes is provided by the coordinated polymerization of actin filaments against cellular membranes. In motile cells, the rapidly polymerizing barbed ends of actin filaments at the lamellipodial network face towards the plasma membrane. Coordinated elongation of the filaments pushes the membrane forward, and consequently induces the formation of pseudopodia at the leading edge of the cell (reviewed in Pantaloni et al., 2001; Pollard and Borisy, 2003). Similarly, polymerization of endosomal vesicle-attached actin filaments at the plasma membrane drives the internalization of endosomes (reviewed in Kaksonen et al., 2006). Actin is a highly conserved (MW ∼42 kDa) protein that exists in two different forms. The globular actin (G-actin) is the monomeric form that can assemble into the filamentous form (F-actin). Actin filaments are polar structures and therefore have two structurally and biochemically distinct ends. Based on the arrowhead pattern created when myosin binds to actin filaments, these ends are called the “barbed end” and the “pointed end”. Under steady state conditions, actin filament elongation primarily occurs at the barbed end, and it favors ATP-bound actin monomers. As the filament ages, the ATP is hydrolyzed and the phosphate is released. This results in the formation of ADP-actin filaments, where ADP-actin monomers dissociate from the pointed end. In the monomeric form, actin can undergo relatively rapid nucleotide exchange to re-charge the monomer with ATP. The ATP-actin monomer is subsequently ready for a new round of polymerization. Consequently, under steady-state conditions actin filaments elongate at the barbed end and simultaneously shrink at the pointed end. For each round of polymerization one molecule of ATP is used. This actin filament dynamics that is driven by ATP-hydrolysis is called “treadmilling”, and it is the driving-force behind the above-mentioned actindependent motile processes (reviewed in Pantaloni et al., 2001). In addition to treadmilling, actin filaments may also undergo other types of dynamics, including rapid depolymerization of newly formed actin filaments at the barbed ends (Kueh

9

Regulation of the Cytoplasmic Actin Monomer Pool in Actin-based Motility

215

et al., 2008). However, the biological significance of this “dynamic instability of actin filaments” is currently unknown. The critical concentration for the polymerization of ATP-actin monomers at the filament barbed and pointed ends are approximately 0.1 and 0.6 μM, respectively (reviewed in Pollard et al., 2000). In the absence of actin-modulating proteins, the concentration of ATP-actin monomers is only slightly higher than the critical concentration of barbed ends. However, many cellular processes require very rapid and accurately regulated elongation of actin filaments and therefore cells typically contain a significantly higher cytoplasmic concentration of ATP-actin monomers associated with various actin-binding proteins. This ensures rapid actin filament polymerization for example during cell migration and endocytic internalization. Furthermore, the nucleation, elongation and disassembly of actin filaments are controlled by a large variety of proteins. As a result, the size, localization and nucleotide-status of the cytoplasmic actin monomer pool are accurately regulated in cells (reviewed in Paavilainen et al., 2004). In this chapter, we discuss how proteins enhancing actin filament disassembly, proteins interacting with actin monomers, as well as proteins inhibiting filament elongation at filament barbed and pointed ends, control the cytoplasmic actin monomer pool in actin-based motility.

9.2 Actin Filament Disassembly The spontaneous dissociation of ADP-actin monomers from filament pointed ends is a slow process (koff ∼0.2 s−1 ) that is considered to be the rate-limiting step in the actin filament treadmilling cycle (Pollard and Cooper, 1986). To ensure rapid actin filament treadmilling in cells, the dissociation of actin monomers from filaments must be accelerated either by increasing the number of filament pointed ends through severing of actin filaments and/or by accelerating the dissociation of actin monomers from filament pointed ends (Fig. 9.1). A number of studies have demonstrated that proteins of the actin-filament depolymerizing factor (ADF)/cofilin and gelsolin families play central roles in promoting actin dynamics in cells by severing actin filaments and by depolymerizing them at the pointed ends.

9.2.1 ADF/Cofilin Family Proteins Actin-depolymerizing factor (ADF) was initially identified as a small protein from chick brain that induced depolymerization of actin filaments in vitro (Bamburg et al., 1980). Subsequently, similar proteins were purified from a number of organisms and tissues, and named cofilin, destrin and actophorin. As the amino acid sequences of the proteins were determined, they turned out to be members of the same protein family that is currently known as ADF/cofilins (reviewed in Bamburg et al., 1999). Unicellular organisms, such as yeasts have typically one ADF/cofilin, whereas

216

P. Lappalainen et al. AC

ADP

ADP

ADP

AC

ADP

ADP

ADP

ADP

ADP

ADP

ATP

D-Pi

AC

AC

ADP

1

AC

AC

ADP

5

ADP

ADP

ADP

ADP

ADP

ADP

ADP

Pf

D-Pi

ATP

ATP

AC

ADP

AC

AT P AT P

AT P

4

AC

2

AD P

C

N

P CAf

ATP

β-thym ATP

3

P

ATP

Pf

Pf

Arp2/3

ATP

AC Pf

AD

P

ADF/cofilin Profilin

N

C

CAP

twinfilin Cyclase associated protein

Actin monomer Capping protein

β-thym

β-thymosin

Fig. 9.1 Regulation of the cytoplasmic actin monomer pool by actin-binding proteins. Actin filament disassembly is accelerated by ADF/cofilins (1) which sever actin filaments and depolymerize them at the pointed end. (2) Cyclase-associated protein interacts with actin monomers, ADF/cofilin and profilin. It is believed to recycle actin monomers for new rounds of polymerization and ADF/cofilins for new rounds of filament depolymerization. (3) Profilin accelerates the nucleotide exchange of actin monomers and their assembly into filament barbed ends. (4) Twinfilin and β-thymosins sequester ADP- and ATP-actin monomers, respectively. (5) Polymerization of a subset of actin filaments is also inhibited by filament barbed end capping proteins

multicellular organisms such as animals and plants often have several ADF/cofilin isoforms. Mammals express three ADF/cofilins, which display small biochemical differences to each other. From these, cofilin-1 is ubiquitously expressed, whereas cofilin-2 is expressed mainly in skeletal and heart muscles, and ADF in epithelial and neuronal cells (Ono et al., 1994; Vartiainen et al., 2002). ADF/cofilins bind both filamentous and monomeric actin with high affinity, with preferential affinity for ADP-actin (Maciver and Weeds, 1994; Carlier et al., 1997). When bound to an actin monomer, ADF/cofilins inhibit the spontaneous nucleotide exchange (Hawkins et al., 1993; Hayden et al., 1993). Importantly, ADF/cofilins increase actin dynamics in vitro by accelerating the dissociation of actin monomers from the filament pointed end by ∼25-fold (Carlier et al., 1997) and by severing actin filaments (Andrianantoandro and Pollard, 2006). The relative contributions of these two ADF/cofilin activities on actin filament treadmilling are not clear and may also vary between different ADF/cofilin isoforms. In addition to the pointed end depolymerization and severing activities, ADF/cofilins accelerate the dissociation of phosphate from ADP-Pi filaments and promote the debranching of

9

Regulation of the Cytoplasmic Actin Monomer Pool in Actin-based Motility

217

Arp2/3-nucleated actin filament network (Blanchoin et al., 2000; Chan et al., 2009). In cells, ADF/cofilins induce actin filament disassembly during motility (Hotulainen et al., 2005; Kiuchi et al., 2007), endocytosis (Lappalainen and Drubin, 1997) and cytokinesis (Abe et al., 1996). In addition to driving actin treadmilling through their filament disassembly activity, ADF/cofilins were reported to promote actin filament assembly by generating new polymerization-competent actin filament barbed ends at the leading edge of cells undergoing directional motility (reviewed in DesMarais et al., 2005). However, recent work suggested that ADF/cofilins contribute also to stimulus-induced actin filament assembly and lamellipodium extension primarily by supplying an abundant pool of cytoplasmic actin monomers through their actin filament severing and depolymerization activities (Kiuchi et al., 2007). All ADF/cofilins are small (MW = 13−19 kDa) proteins composed of a single structurally conserved ADF homology (ADF-H) domain. The structure consists of a four-stranded mixed β-sheet sandwiched between two pairs of α-helices (Fig. 9.2). Site-directed mutagenesis studies on various ADF/cofilins, together with structural analysis of the related twinfilin ADF-H domain/actin monomer complex, revealed

pointed end 4

2

3

1

barbed end Fig. 9.2 The mechanisms of ADF/cofilin interaction with monomeric and filamentous actin. ADF/cofilins interact with two adjacent subunits in an actin filament to alter the filament twist and to destabilize intersubunit interactions with the filament. This leads to filament severing and depolymerization at the pointed end. After monomer dissociation from the filament, ADF/cofilins stay bound to the cleft between actin subdomains 1 and 3. The models for ADF/cofilin interactions with actin filaments and monomers are based on the structure of the twinfilin ADF-H domain – actin monomer complex (Paavilainen et al., 2008)

218

P. Lappalainen et al.

that ADF/cofilins bind to the groove between actin subdomains 1 and 3 (Fig. 9.2). The G-actin binding site of ADF/cofilins consists of the N-terminal half of the long α-helix and elements surrounding the helix (Yonezawa et al., 1991; Lappalainen et al., 1997; Paavilainen et al., 2008). When bound to an actin filament, ADF/cofilins use the “G-actin binding site” to interact with the “barbed end” of one actin subunit, whereas another region located around the C-terminus of the protein interacts with the subdomain 2 of an adjacent actin monomer in the filament (Fig. 9.2). ADF/cofilins affect the actin filament conformation by stabilizing a filament state with a mean twist of 162◦ (McGough et al., 1997; Galkin et al., 2001). This changes the thermodynamic stability of the filament and leads to its depolymerization and/or severing (McGough and Chiu, 1999; Bobkov et al., 2002). The activities of ADF/cofilin are regulated by phosphorylation, pH, phosphoinositides, and through interactions with other proteins. Phosphorylation of ADF/cofilins at a serine positioned close to the N-terminus of the protein (Ser3 in animals) efficiently inactivates its actin binding and disassembly activities (Agnew et al., 1995; Ressad et al., 1998). The N-terminal serine is located at the G- and F-actin binding interface of ADF/cofilin, explaining why phosphorylation of this residue inhibits actin monomer and filament binding of ADF/cofilins (Paavilainen et al., 2008). ADF/cofilin phosphorylation in animals is catalyzed by LIM and TES family kinases, and its activation through depohosphorylation is induced by Slingshot and chronophin family phosphatases (Arber et al., 1998; Toshima et al., 2001; Niwa et al., 2002; Gohla et al., 2005). All ADF/cofilins bind phosphoinositides (mainly PIP2 and PIP3 ), and their activities in vitro and in cells are down-regulated by phosphoinositide-binding (Yonezawa et al., 1990; van Rheenen et al., 2007; Leyman et al., 2009). However, whether ADF/cofilins interact with PIP2 through a specific binding-pocket or if they bind phosphoinositides through a larger interface is currently controversial (Gorbatyuk et al., 2006; Ojala et al., 2001). The actin and phosphoinositide binding activities of some ADF/cofilins are also inhibited at elevated pH, but the physiological role(s) of the pH-regulation is unknown (Yonezawa et al., 1985; Frantz et al., 2008). ADF/cofilins also interact with other proteins such as Aip1 and coronin. Aip1 enhances the actin filament severing/depolymerization activity of ADF/cofilins and caps the barbed ends of ADF/cofilin-decorated filaments (reviewed in Ono, 2003). Coronin, on the other hand, inhibits the severing of newly polymerized ATP- or ADP-Pi-actin filaments, but enhances the ADF/cofilin-mediated disassembly of ADP-actin filaments (Cai et al., 2007; Gandhi et al., 2009). Taken together, the actin filament disassembly by ADF/cofilins is regulated by a complex interplay involving phosphorylation-dephosphorylation, interactions with PIP2 -rich membranes and with other proteins. This provides accurate tools to control actin filament stability and the availability of actin monomers at specific regions of cells in response to various intracellular or extracellular stimuli.

9

Regulation of the Cytoplasmic Actin Monomer Pool in Actin-based Motility

219

9.2.2 Gelsolin Family Proteins Gelsolin, villin, adseverin, capG, advillin, supervillin and flightless I form a family of actin filament severing and barbed end capping proteins, which contain either three or six copies gelsolin-like (G) domains (reviewed in Silacci et al., 2004). The founding member of the family was identified as a Ca2+ -activated protein that promotes actin filament disassembly (gel-to-solution transition) and thus named gelsolin (Yin and Stossel, 1979). The members of the family display also some important biochemical differences to each other, because for example capG lacks the actin filament severing activity, whereas villin additionally harbors actin filament bundling activity (Southwick and DiNubile, 1986; Glenney and Weber, 1981). Gelsolin is composed of six gelsolin-like domains that display structural homology to ADF-H domains. In a calcium-free environment, these domains are packed into a compact inert structure (Burtnick et al., 1997). Calcium-binding triggers a conformational change in gelsolin to expose the actin-binding surfaces of the domains. This leads to actin filament severing and subsequent barbed end capping by gelsolin (Burtnick et al., 2004; Nag et al., 2009). While the filament severing activity of gelsolin is activated by calcium, phosphoinositides, especially PI(4, 5)P2 , inhibit the interactions of gelsolin with actin and dissociate gelsolin from actin filaments (Janmey and Stossel, 1987). At least in N-cadherin adhesions of rat fibroblasts, a local increase in the PI(4, 5)P2 concentration regulates actin assembly by uncapping gelsolin from the filament barbed ends (El Sayegh et al., 2007). The exact roles of gelsolin family proteins in regulating actin-based motility are not fully understood. Studies on knockout mice revealed that gelsolin contributes to osteoclast motility through its involvement in adhesion (podosome) formation (Chellaiah et al., 2000). Gelsolin also contributes to neuronal growth cone filopodia retraction (Lu et al., 1997). On the other hand, CapG knockout mice displayed defects in phagocytosis and membrane ruffling in macrophages (Witke et al., 2001a,b). Thus, unlike ADF/cofilins that play a fundamental role in all eukaryote cells by promoting actin treadmilling in actin-based motility, the gelsolin family proteins appear to be involved in stimulus-induced (Ca2+ ) remodeling of actin filament structures during more specific cellular processes.

9.3 Actin Monomer-Binding Proteins Cytoplasmic concentration of actin monomers is typically much higher than the critical concentration for the polymerization of ATP-actin monomers. This is accomplished by the presence of actin monomer binding proteins that control the size, nucleotide status and subcellular localization of the cytoplasmic actin monomer pool. These proteins can either sequester actin monomers (e.g. twinfilin

220

P. Lappalainen et al.

and β-thymosins) or accelerate the nucleotide exchange of actin monomers and/or stimulate actin filament assembly by localizing monomers to the sites of rapid actin filament assembly (e.g. profilin and cyclase-associated protein) (see also Fig. 9.1).

9.3.1 Profilin Profilin is a small (MW = 12−16 kDa) actin binding protein that was first identified from calf spleen as a protein that forms a complex with actin monomers to inhibit their polymerization (Carlsson et al., 1977). Since then, profilins have been found from all eukaryotes studied, indicating that profilin is a fundamental actin-binding protein that is present in all eukaryotes and their cells (Witke, 2004). The number of genes encoding profilin varies between species. Budding yeast has only one profilin gene (Magdolen et al., 1988), while in mammals five isoforms named profilin I, IIa, IIb, III and IV exist. Profilin I is expressed in most tissues, whereas profilin II expression is restricted to neuronal cells (Witke et al., 1998, 2001a,b). Profilin III and IV are expressed in testis (Braun et al., 2002; Obermann et al., 2005). Profilins interact with actin monomers, polyproline-rich proteins and phosphatidylinositol-(4, 5)-bisphosphate (PIP2 ) (reviewed in Witke, 2004). They bind ATP-actin monomers with a higher affinity (Kd = 0.1 μM) than ADP-actin (Kd = 0.5 μM). Most profilins also catalyze the nucleotide exchange on actin monomers (reviewed in Pollard et al., 2000). In the absence of free actin filament barbed ends, profilin sequesters actin monomers. When barbed ends are available, profilin promotes the assembly of ATP-actin monomers to filaments (Pantaloni and Carlier, 1993). Thus, profilin inhibits spontaneous actin filament nucleation and directs actin filament assembly to the pre-existing barbed ends. Furthermore, through its interactions with polyproline-rich proteins, such as formins and Ena/VASP, profilin localizes ATP-actin monomers to the sites of rapid actin filament assembly in cells (Reinhard et al., 1995; Evangelista et al., 1997). Profilin is a central component of the actin-based motility reconstituted in vitro (Loisel et al., 1999). It is also an important regulator of cytoskeletal dynamics in vivo, because in cultured mammalian and Drosophila cells profilin is enriched at the dynamic actin-rich structures and is essential for lamellipodia formation (Buss et al., 1992; Rogers et al., 2003). Disruption of profilin gene leads to embryonic lethality in mice and in flies (Witke et al., 2001a,b; Cooley et al., 1992). Furthermore, deletion of profilin gene in budding yeast causes severe growth defects, changes in cell morphology and actin disorganization (Haarer et al., 1990; Wolven et al., 2000). The interaction of profilin with actin is regulated by binding to phosphoinositides, especially PI(4, 5)P2 (Lassing and Lindberg, 1985). Furthermore, profilins appear to regulate lipid metabolism, since profilin inhibits PIP2 -hydrolysis by phospholipase C (Goldschmidt-Clermont et al., 1990). However, the biological roles of profilin inhibition by PIP2 and profilin’s possible role in PIP2 -metabolisim in cells are largely unknown. The three-dimensional structure of profilin consists of a seven-stranded anti-parallel β-sheet with α-helices located on both sides of the β-sheet (Schutt et al.,

9

Regulation of the Cytoplasmic Actin Monomer Pool in Actin-based Motility

221

1993). The polyproline- and actin-binding sites are located at the opposite sides of the molecule, explaining how profilin can simultaneously interact with these two ligands (Mahoney et al., 1997). Mutagenesis studies also suggest that profilin has two distinct PIP2 -binding sites; one located close to the polyproline-binding site, and the other one overlapping with the actin binding site (Lambrechts et al., 2002; Skare and Karlsson, 2002). This explains how PIP2 inhibits the actin and polyproline interactions of profilin.

9.3.2 CAP (Cyclase-Associated-Protein) Cyclase-associated protein (CAP), also known as Srv2 in budding yeast, was first identified from Saccharomyces cerevisiae as a protein that forms a complex with adenylyl cyclase. Subsequently, Srv2/CAP was also identified as a suppressor of activated Ras allele in yeast (Field et al., 1988, 1990; Fedor-Chaiken et al., 1990). Since then, highly conserved CAP homologues have been identified in all eukaryotes studied (Hubberstey and Mottillo, 2002). While budding yeast has only one CAP protein, two tissue-specific CAP isoforms are present in mammals (Swiston et al., 1995; Bertling et al., 2004). Importantly, interaction between CAP and adenylyl cyclase detected in yeasts appears not to be conserved to other species such as mammals. Instead, all CAPs analyzed to date bind actin monomers and regulate cytoskeletal dynamics in cells (reviewed in Hubberstey and Mottillo, 2002). CAP consists of two domains: the N-terminal domain composed of a bundle of six anti-parallel α-helices (Ksiazek et al., 2003), and a C-terminal domain consisting of β-strands forming a sheet-like structure (Dodatko et al., 2004). These two domains are separated by a region containing two proline-rich regions and a putative actin-binding WH2 domain. Full-length CAP forms oligomers, and the predicted coiled-coil domain at the N-terminal end plays an important role in the oligomerization (Moriyama and Yahara, 2002; Quintero-Monzon et al., 2009). CAP interacts with actin monomers and with other actin-binding proteins. The N-terminal domain of CAP binds ADF/cofilin – actin monomer complex, and is essential for the capability of CAP to accelerate ADF/cofilin-mediated actin turnover in vitro (Moriyama and Yahara, 2002; Quintero-Monzon et al., 2009). The C-terminal domain of CAP binds both ADP-actin monomers (Kd ∼0.02 μM) and ATP-actin monomers (Kd ∼1.9 μM) through separate sites (Freeman et al., 1995; Mattila et al., 2004). The first polyproline region (P1) of CAP binds profilin (Bertling et al., 2007), whereas the second polyproline region (P2) is responsible for the correct subcellular localization of CAP at least in budding yeast (Freeman et al., 1996; Lila and Drubin, 1997). In vitro, CAP promotes actin filament turnover by dissociating the ADF/cofilin – G-actin complex, recharging the ADP-actin monomers with ATP, and thus recycling actin monomers from ADF/cofilin for new rounds of polymerization (Balcer et al., 2003; Moriyama and Yahara, 2002; Mattila et al., 2004). However, the exact mechanism by which the various protein interactions and oligomerization of CAP contribute to actin dynamics is not known.

222

P. Lappalainen et al.

CAP plays an important role in cytoskeletal dynamics in cells. In budding yeast, CAP displays partial co-localization with cortical actin patches and its deletion lead to defects in actin organization (Freeman et al., 1996). Deletion of CAP from Dictyostelium, Drosophila or plants resulted in severe defects in actin organization, endocytosis, and cell and organism morphology (Noegel et al., 1999; Baum et al., 2000; Barrero et al., 2002). In mammalian cells, CAP1 localizes to dynamic regions of the actin cytoskeleton and its depletion by RNAi lead to defects in cell motility, morphogenesis, and endocytosis (Bertling et al., 2004).

9.3.3 Twinfilin Twinfilin was originally identified from budding yeast as a homologue of ADF/cofilins. However, instead of a single ADF-H domain, twinfilins are composed of two ADF-H domains connected by a short linker region and followed by a C-terminal tail region (Goode et al., 1998). Unlike ADF/cofilins, twinfilins do not bind to the sides of actin filaments or induce their depolymerization from pointed ends. Instead, twinfilins efficiently prevent actin filament assembly by sequestering ADP-actin monomers and by capping filament barbed ends (Ojala et al., 2002; Helfer et al., 2006). At least in yeast, twinfilin also induces actin filament severing at low pH (Moseley et al., 2006). Furthermore, all twinfilins studied to date bind heterodimeric capping protein through their C-terminal tail region, and at least in budding yeast this activity is essential for twinfilin’s correct subcellular localization and function in actin filament turnover in vivo (Palmgren et al., 2001; Falck et al., 2004). However, the exact mechanism by which combination of the various activities of twinfilin contributes to actin-based motile processes in cells remains to be determined. Unicellular organisms such as yeasts have one twinfilin protein, whereas several twinfilin isoforms are present in many multicellular organisms such as mammals (Nevalainen et al., 2009). Deletion of twinfilin gene in budding yeast results in defects in the organization of cortical actin patches, which are the sites of endocytosis in yeast (Goode et al., 1998). Similarly, depletion of twinfilin isoforms from mammalian cells by RNAi results in defects in endocytosis (Pelkmans et al., 2005). These data suggest that twinfilin is important in regulation of actin dynamics during endocytic internalization. However, twinfilin appears to be also involved in other actin-dependent processes, because hypomorphic twinfilin mutant flies display rough eye phenotype and defects in bristle morphogenesis (Wahlström et al., 2001). The actin monomer binding and filament severing activities of twinfilin are inhibited by phosphoinositides in vitro, but the physiological significance of this regulation remains to be determined (Palmgren et al., 2001; Moseley et al., 2006). Collectively, twinfilin appears to be a central regulator of cytoplasmic actin monomer pool during actin-based motile processes such as endocytosis. However, the exact role of twinfilin during different actin-dependent cellular processes as well as how its activity is regulated in cells provide important challenges for future research.

9

Regulation of the Cytoplasmic Actin Monomer Pool in Actin-based Motility

223

9.3.4 β-thymosins Thymosin β4 was first identified as a potential thymic hormone involved in T-cell maturation (Low et al., 1981). However, subsequent studies revealed that thymosin β4 is a major actin monomer sequestering peptide in platelets (Safer et al., 1991). Thymosin β4 and its homologues (three β-thymosins are found in humans) are small (MW ∼5 kDa) peptides that are mostly unstructured in solution (reviewed in Mannherz and Hannappel, 2009). However, in the presence of actin β-thymosins form two α-helices that interact with the cleft between actin subdomains 1 and 3 at the “barbed end” and with the cleft between actin subdomains 2 and 4 at the “pointed end” of the monomer, respectively (Hertzog et al., 2004). β-thymosins display both sequence and structural homology to WH2 domains (Chereau et al., 2005; see also chapter XXX in the book by Dominguez et al.,). Like most WH2 domains, β-thymosins bind ATP-G-actin with higher affinity compared to ADP-G-actin (Hertzog et al., 2002; Mattila et al., 2003). However, unlike most WH2 domains that promote actin polymerization, β-thymosins are efficient actin monomer sequestering proteins. This is due to the presence of the C-terminal α-helix that binds to actin subdomain 2 to block the “pointed end” of the monomer. This helix is absent from WH2 domain structures and its inactivation in β-thymosin results in a switch from monomer sequestering to promotion of actin filament assembly (Hertzog et al., 2004). In mammals, thymosin β4 is especially abundant in platelets and leukocytes, whereas another isoform thymosin β15 is abundant in developing neurons and certain cancer cell-types. Interestingly, β-thymosins are not present in most invertebrates or unicellular eukaryotes (reviewed in Mannherz and Hannappel, 2009). This suggests that β-thymosins are not crucial regulators of actin-based motility. However, sequestration of a large cytoplasmic ATP-actin monomer pool by β-thymosins may have a central role in stimulus-responsive actin polymerization e.g. during lymphocyte and platelet migration and morphogenesis.

9.4 Actin Filament Capping Proteins The cytoplasmic actin monomer pool is also regulated by proteins that bind filament ends to inhibit actin monomer exchange (both association and dissociation). Most actin filament capping proteins bind to and regulate filament dynamics at the barbed ends (Fig. 9.3). These include the heterodimeric capping protein, Eps8, gelsolin, and twinfilin. Tropomodulins, on the other hand, cap actin filament pointed ends. In addition to these genuine actin filament capping proteins, also formins and Ena/VASP family proteins associate with actin filament barbed ends to promote actin dynamics. However, rather than inhibiting monomer exchange at filament barbed ends, formins and Ena/VASP proteins promote actin polymerization at filament barbed ends in cells. In this chapter, we discuss the functions of actin filament capping proteins. The biochemical activities and cellular roles of formins and Ena/VASP family proteins are discussed in other chapters of the book.

224

P. Lappalainen et al.

S2

C

S3

S6 S5 S4

N

S1

barbed end

capping protein

S1

S2

N

C

S3

S4

S5

S6

gelsolin

twinfilin

Fig. 9.3 Models for the mechanisms of barbed end capping by CapZ, gelsolin and twinfilin. Capping protein is a mushroom-shaped heterodimeric protein, which interacts with filament barbed ends mainly through the C-terminal amphipathic α-helices of the α- and β-subunits. Gelsolin is a Ca2+ induced actin filament severing protein composed of six structurally homologous domains. Gelsolin also caps actin filament barbed ends and the minimal region capable of capping filament barbed ends consists of domains G1-G2. Twinfilin is an actin monomer sequestering protein composed of two ADF-H domains. Twinfilin also caps filament barbed ends, and the presence of two functional ADF-H domains is required for barbed end capping by this protein

9.4.1 Heterodimeric Capping Protein Capping protein was first identified from muscle, and was initially named β-actinin and subsequently CapZ (Maruyama and Obinata, 1965). Conserved homologues of CapZ have been since then identified in a variety of organisms. It seems that nearly all eukaryotic cells express at least one capping protein isoform (reviewed in Cooper and Sept, 2008). Capping proteins are heterodimers composed of α- and β-subunits, which each have a molecular mass of ∼30 kDa. Although the α- and β-subunits

9

Regulation of the Cytoplasmic Actin Monomer Pool in Actin-based Motility

225

subunits lack detectable sequence homology to each other, they display very similar secondary and tertiary structures (reviewed in Cooper and Sept, 2008). In the stable heterodimer, these two structurally homologous subunits fold into a mushroomshaped structure. The C-terminal amphipathic α-helices of each subunit locate at the “top” of the mushroom and are believed to form the main actin-binding sites of the molecule (Yamashita et al., 2003; Wear et al., 2003; Narita et al., 2006)(Fig. 9.3). Heterodimeric capping proteins bind actin filament barbed ends with subnanomolar activity, and prevent the association and dissociation of actin monomers to the capped filament ends (reviewed in Cooper and Sept, 2008). Heterodimeric capping protein is an essential component of the reconstituted actin-based motility (Loisel et al., 1999), although it can be replaced by other barbed end capping proteins such as twinfilin, Eps8 or gelsolin (Disanza et al., 2004; Helfer et al., 2006). Capping protein is proposed to function in actin-based motility by preventing actin polymerization at undesired locations, and therefore “funneling” actin monomers to the sub-population of filament barbed ends that are not capped (Loisel et al., 1999). However, a recent study provided evidence that at least in vitro, capping protein may also contribute to actin-based motility by promoting more frequent actin filament nucleation by the Arp2/3 complex (Akin and Mullins, 2008). Heterodimeric capping protein has a central role in actin dynamics in cells. Deletion of capping protein subunits in budding yeast results in defects in the organization of cortical actin patches. In Drosophila, loss of capping protein leads to embryonic lethality due to problems in actin organization (Kim et al., 2006; Hopmann et al., 1996). RNAi knockdown studies revealed that in cultured mammalian and Drosophila cells, heterodimeric capping protein is important for lamellipodia formation (Mejillano et al., 2004; Iwasa and Mullins, 2007). Finally, the muscle cell specific capping protein isoform CapZ is essential for assembly and/or maintenance of myofibrils (Hart and Cooper, 1999). The activities of heterodimeric capping protein are regulated through interactions with phosphoinositides. PI(4, 5)P2 inhibits the barbed end capping by interacting with the same region of the protein that is also important for actin-binding (Kim et al., 2007). However, the exact mechanism by which phosphoinositides regulate capping protein is still controversial. It was proposed that PI(4, 5)P2 efficiently removes capping protein from the filament barbed ends in a concentrationdependent manner (Schafer et al., 1996; Kim et al., 2007). However, another study provided evidence that although PI(4, 5)P2 prevents capping protein binding to filament barbed ends, it does not efficiently dissociate capping protein from filament ends (Kuhn and Pollard, 2007). Nevertheless, these studies suggest that the activity of heterodimeric capping protein is inhibited at the PI(4, 5)P2 -rich membranes to enhance actin polymerization at these sites. The filament barbed end capping activity of the heterodimeric capping protein is also inhibited through interactions with other proteins such as CARMIL, V1/myotropin, CKIP-1 and CD2AP (Yang et al., 2005; Bhattacharya et al., 2006; Canton et al., 2005; Bruck et al., 2006). In the future it will be important to examine how capping protein, together with its interacting proteins and membrane phosphoinositides, regulates actin filament assembly during various actin-based motile processes in cells.

226

P. Lappalainen et al.

9.4.2 Eps8 Eps8 is a multifunctional scaffolding protein that interacts with many proteins involved in cytoskeletal dynamics, such as Abi1 and IRSp53. Eps-8 binds and bundles actin filaments through its C-terminal region and promotes filopodia formation in cells (Scita et al., 2001; Disanza et al., 2006; Menna et al., 2009). Importantly, Eps8 also displays a strong actin filament barbed end capping activity in vitro and can replace heterodimeric capping protein in a reconstituted actin-based motility assay. The filament barbed end capping and bundling activities appear to require different sites at the C-terminal region of the protein (Disanza et al., 2004). Thus, Eps8 can control actin filament assembly and organization by at least two distinct and direct mechanisms. In the future, it will be interesting to examine the mechanism(s) by which the various biochemical activities of Eps8 (barbed end capping, filament bundling, Abi1- and IRSp53-interactions) contribute to actin-based motile processes in cells. Furthermore, the structural mechanism by which Eps8 caps filament barbed ends remains to be elucidated.

9.4.3 Tropomodulin Tropomodulin was initially identified from erythrocytes as a tropomyosin-binding protein (Fowler, 1987). Whereas majority of actin filament capping protein identified to date interact with filament barbed ends, tropomodulin was shown to cap filament pointed ends (reviewed in Kostyukova, 2008). In the absence of tropomyosin, the affinity of tropomodulin to filament pointed ends is relatively low (Kd ∼0.2 μM). However, tropomodulin contains two tropomyosin-binding sites at the N-terminal region and consequently its affinity for tropomyosindecorated filaments is very high (∼50 pM) (Weber et al., 1999; Fowler et al., 2003). The cellular role of tropomodulin has been most extensive studied in muscle cells, where tropomodulin localizes to the pointed ends of sarcomeric actin filaments. Deletion of a certain tropomodulin isoform (Tmod1) leads to heart defects in mice (Fritz-Six et al., 2003). In cardiomyocytes, reduced Tmod1 expression leads to the formation of abnormally long actin filaments in myofibrils, suggesting that tropomodulin may regulate actin filament length through its pointed end capping activity (Littlefield et al., 2001). Tropomodulins play an important role in actin dynamics also in other tissues than muscles. Two tropomodulin isoforms, Tmod2 and Tmod3, are mainly expressed in non-muscle tissues (reviewed in Kostyukova, 2008). Over-expression of Tmod3 in endothelial cells inhibits their motility, and Tmod2 has been linked to certain pathological conditions in the brain (Fischer et al., 2003; Chen et al., 2007). However, it is important to note that in addition to its filament pointed end capping activity, Tmod3 also sequesters actin monomers (Fischer et al., 2006). Thus, the importance of actin filament pointed end capping by tropomodulins during actin-based motile processes, such as cell migration, remains to be examined in the future.

9

Regulation of the Cytoplasmic Actin Monomer Pool in Actin-based Motility

227

9.4.4 Other Actin Filament Capping Proteins Also gelsolin family proteins (see Section 9.2.2.) and twinfilin (see Section 9.3.3.) display actin filament barbed end capping activity. Both proteins can also substitute heterodimeric capping protein in a reconstituted actin-based motility assay (Helfer et al., 2006). In the case of gelsolin, the barbed end capping follows filament severing and thus ensures efficient disassembly of actin filaments (reviewed in Silacci et al., 2004). The biological role of filament barbed end capping by twinfilin is currently not known. Both gelsolin and twinfilin are composed of structurally homologous domains. Despite the lack of detectable sequence identity, these domains interact with a similar interface of actin (Burtnick et al., 2004; Paavilainen et al., 2008). Interestingly, recent structural and mutagenesis studies suggested that these proteins cap actin filament barbed ends at least through partially homologous mechanism (Paavilainen et al., 2007). The presence of the two functional ADF-H domains is essential for the barbed end capping activity of twinfilin, and these domains appear to interact with the same regions of the actin filament barbed ends as the gelsolin domains G1 and G2. Thus, at least in the case of heterodimeric capping protein, gelsolin and twinfilin, filament barbed end capping requires the presence of two structurally homologous domains/subunits that interact with distinct regions of the filament barbed ends (Fig. 9.3).

9.5 Conclusions and Future Prospectives Landmark studies by the Carlier laboratory where actin-based motility was reconstituted in vitro revealed that, in addition to an actin filament nucleating protein and its activator, actin based motility depends on proteins inducing filament disassembly, proteins promoting monomers recycling for new rounds of polymerization, and proteins regulating actin polymerization at filament barbed ends (Loisel et al., 1999). While ADF/cofilin family proteins appear to be the only proteins capable of depolymerizing filaments during actin based motility, other barbed end capping proteins, such as gelsolin, Eps8 or twinfilin, can replace heterodimeric capping protein in this assay. However, the situation is much more complex during various actin-based motile processes in cells, because additional cytoskeletal regulators such as actin monomer sequestering proteins and cyclase-associated protein play important roles at least during cell migration and endocytosis. Furthermore, the activities of various actin-binding proteins are regulated by phosphorylation as well as through interactions with membrane phosphoinositides and other proteins. Thus, future studies on cells and on more complex in vitro systems are required to elucidate how various central actinregulating proteins together with their interaction partners regulate complex motile processes.

228

P. Lappalainen et al.

References Abe H, Obinata T, Minamide LS, Bamburg JR (1996) Xenopus laevis actin-depolymerizing factor/cofilin: a phosphorylation-regulated protein essential for development. J Cell Biol. 132:871–885. Agnew BJ, Minamide LS, Bamburg JR (1995) Reactivation of phosphorylated actin depolymerizing factor and identification of the regulatory site. J Biol Chem. 270:17582–17587. Akin O, Mullins RD (2008) Capping protein increases the rate of actin-based motility by promoting filament nucleation by the Arp2/3 complex. Cell. 133:841–851. Andrianantoandro E, Pollard TD (2006) Mechanism of actin filament turnover by severing and nucleation at different concentrations of ADF/cofilin. Mol Cell. 24:13–23. Arber S, Barbayannis FA, Hanser H, Schneider C, Stanyon CA, Bernard O, Caroni P (1998) Regulation of actin dynamics through phosphorylation of cofilin by LIM-kinase. Nature. 393:805–809. Balcer HI, Goodman AL, Rodal AA, Smith E, Kugler J, Heuser JE, Goode BL (2003) Coordinated regulation of actin filament turnover by a high-molecular-weight Srv2/CAP complex, cofilin, profilin, and Aip1. Curr Biol. 13:2159–2169. Bamburg JR, Harris HE, Weeds AG (1980) Partial purification and characterization of an actin depolymerizing factor from brain. FEBS Lett. 121:178–182. Bamburg JR, McGough A, Ono S (1999) Putting a new twist on actin: ADF/cofilins modulate actin dynamics. Trends Cell Biol. 9:364–370. Barrero RA, Umeda M, Yamamura S, Uchimiya H (2002) Arabidopsis CAP regulates the actin cytoskeleton necessary for plant cell elongation and division. Plant Cell. 14:149–163. Baum B, Li W, Perrimon N (2000) A cyclase-associated protein regulates actin and cell polarity during Drosophila oogenesis and in yeast. Curr Biol. 10:964–973. Bertling E, Hotulainen P, Mattila PK, Matilainen T, Salminen M, Lappalainen P (2004) Cyclase-associated protein 1 (CAP1) promotes cofilin-induced actin dynamics in mammalian nonmuscle cells. Mol Biol Cell. 15:2324–2334. Bertling E, Quintero-Monzon O, Mattila PK, Goode BL, Lappalainen P (2007) Mechanism and biological role of profilin-Srv2/CAP interaction. J Cell Sci. 120:1225–1234. Bhattacharya N, Ghosh S, Sept D, Cooper JA (2006) Binding of myotrophin/V-1 to actin-capping protein: implications for how capping protein binds to the filament barbed end. J Biol Chem. 281:31021–31030. Blanchoin L, Pollard TD, Mullins RD (2000) Interactions of ADF/cofilin, Arp2/3 complex, capping protein and profilin in remodeling of branched actin filament networks. Curr Biol. 10:1273–1282. Bobkov AA, Muhlrad A, Kokabi K, Vorobiev S, Almo SC, Reisler E (2002) Structural effects of cofilin on longitudinal contacts in F-actin. J Mol Biol. 323:739–750. Braun A, Aszodi A, Hellebrand H, Berna A, Fassler R, Brandau O (2002) Genomic organization of profilin-III and evidence for a transcript expressed exclusively in testis. Gene. 283:219–225. Bruck S, Huber TB, Ingham RJ, Kim K, Niederstrasser H, Allen PM, Pawson T, Cooper JA, Shaw AS. (2006) Identification of a novel inhibitory actin-capping protein binding motif in CD2associated protein. J Biol Chem. 281:19196–19203. Burtnick LD, Koepf EK, Grimes J, Jones EY, Stuart DI, McLaughlin PJ Robinson RC (1997) The crystal structure of plasma gelsolin: implications for actin severing, capping, and nucleation. Cell. 90:661–670. Burtnick LD, Urosev D, Irobi E, Narayan K, Robinson RC (2004) Structure of the N-terminal half of gelsolin bound to actin: roles in severing, apoptosis and FAF. EMBO J. 23: 2713–2722. Buss F, Temm-Grove C, Henning S, Jockusch BM (1992) Distribution of profilin in fibroblasts correlates with the presence of highly dynamic actin filaments. Cell Motil Cytoskeleton. 22: 51–61. Cai L, Marshall TW, Uetrecht AC, Schafer DA, Bear JE (2007) Coronin 1B coordinates Arp2/3 complex and cofilin activities at the leading edge. Cell. 128:915–929.

9

Regulation of the Cytoplasmic Actin Monomer Pool in Actin-based Motility

229

Canton DA, Olsten ME, Kim K, Doherty-Kirby A, Lajoie G, Cooper JA, Litchfield DW (2005) The pleckstrin homology domain-containing protein CKIP-1 is involved in regulation of cell morphology and the actin cytoskeleton and interaction with actin capping protein. Mol Cell Biol. 25:3519–3534. Carlier MF, Laurent V, Santolini J, Melki R, Didry D, Xia GX, Hong Y, Chua NH, Pantaloni D (1997) Actin depolymerizing factor (ADF/cofilin) enhances the rate of filament turnover: implication in actin-based motility. J Cell Biol. 136:1307–1322. Carlsson L, Nystrom LE, Sundkvist I, Markey F, Lindberg U (1977) Actin polymerizability is influenced by profilin, a low molecular weight protein in non-muscle cells. J Mol Biol. 115:465–483. Chan C, Beltzner CC, Pollard TD (2009) Cofilin dissociates Arp2/3 complex and branches from actin filaments. Curr Biol. 19:537–545. Chellaiah M, Kizer N, Silva M, Alvarez U, Kwiatkowski D, Hruska KA (2000) Gelsolin deficiency blocks podosome assembly and produces increased bone mass and strength. J Cell Biol. 148:665–678. Chen A, Liao WP, Lu Q, Wong WS, Wong PT (2007) Upregulation of dihydropyrimidinase-related protein 2, spectrin alpha II chain, heat shock cognate protein 70 pseudogene 1 and tropomodulin 2 after focal cerebral ischemia in rats – a proteomics approach. Neurochem Int. 50:1078–1086. Chereau D, Kerff F, Graceffa P, Grabarek Z, Langsetmo K, Dominguez R (2005) Actinbound structures of Wiskott-Aldrich syndrome protein (WASP)-homology domain 2 and the implications for filament assembly. Proc Natl Acad Sci. 102:16644–16649 http://www.ncbi.nlm.nih.gov/pubmed/16275905?ordinalpos=6&itool=EntrezSystem2. PEntrez.Pubmed.Pubmed_ResultsPanel.Pubmed_DefaultReportPanel.Pubmed_RVDocSum. Cooley L, Verheyen E, Ayers K (1992) Chickadee encodes a profilin required for intercellular cytoplasm transport during Drosophila oogenesis. Cell. 6:173–184. Cooper JA, Sept D (2008) New insights into mechanism and regulation of actin capping protein. Int Rev Cell Mol Biol. 267:183–206. DesMarais V, Ghosh M, Eddy R, Condeelis J (2005) Cofilin takes the lead. J Cell Sci. 118:19–26. Disanza A, Carlier MF, Stradal TE, Didry D, Frittoli E, Confalonieri S, Croce A, Wehland J, Di Fiore PP, Scita G (2004) Eps8 controls actin-based motility by capping the barbed ends of actin filaments. Nat Cell Biol. 6:1180–1188. Disanza A, Mantoani S, Hertzog M, Gerboth S, Frittoli E, Steffen A, Berhoerster K, Kreienkamp HJ, Milanesi F, Di Fiore PP, Ciliberto A, Stradal TE, Scita G (2006) Regulation of cell shape by Cdc42 is mediated by the synergic actin-bundling activity of the Eps8-IRSp53 complex. Nat Cell Biol. 8:1337–1347. Dodatko T, Fedorov AA, Grynberg M, Patskovsky Y, Rozwarski DA, Jaroszewski L, AronoffSpencer E, Kondraskina E, Irving T, Godzik A, Almo SC (2004) Crystal structure of the actin binding domain of the cyclase-associated protein. Biochemistry. 43:10628–10641. El Sayegh TY, Arora PD, Ling K, Laschinger C, Janmey PA, Anderson RA, McCulloch CA (2007) Phosphatidylinositol-4,5 bisphosphate produced by PIP5KIgamma regulates gelsolin, actin assembly, and adhesion strength of N-cadherin junctions. Mol Biol Cell. 18: 3026–3038. Evangelista M, Blundell K, Longtine MS, Chow CJ, Adames N, Pringle JR, Peter M, Boone C (1997) Bni1p, a yeast formin linking cdc42p and the actin cytoskeleton during polarized morphogenesis. Science. 276:118–122. Falck S, Paavilainen VO, Wear MA, Grossmann JG, Cooper JA, Lappalainen P (2004) Biological role and structural mechanism of twinfilin-capping protein interaction. EMBO J. 23: 3010–3019. Fedor-Chaiken M, Deschenes RJ, Broach JR (1990) SRV2, a gene required for RAS activation of adenylate cyclase in yeast. Cell. 61:329–340. Field J, Nikawa J, Broek D, MacDonald B, Rodgers L, Wilson IA, Lerner RA, Wigler M (1988) Purification of a RAS-responsive adenylyl cyclase complex from Saccharomyces cerevisiae by use of an epitope addition method. Mol Cell Biol. 8:2159–2165.

230

P. Lappalainen et al.

Field J, Vojtek A, Ballester R, Bolger G, Colicelli J, Ferguson K, Gerst J, Kataoka T, Michaeli T, Powers S (1990) Cloning and characterization of CAP, the S. cerevisiae gene encoding the 70 kd adenylyl cyclase-associated protein Cell. 61:319–327. Fischer RS, Fritz-Six KL, Fowler VM (2003) Pointed-end capping by tropomodulin3 negatively regulates endothelial cell motility. J Cell Biol. 161:371–380. Fischer RS, Yarmola EG, Weber KL, Speicher KD, Speicher DW, Bubb MR, Fowler VM (2006) Tropomodulin 3 binds to actin monomers. J Biol Chem. 281:36454–36465. Fowler VM (1987) Identification and purification of a novel Mr 43,000 tropomyosin-binding protein from human erythrocyte membranes. J Biol Chem. 262:12792–12800. Fowler VM, Greenfield NJ, Moyer J (2003) Tropomodulin contains two actin filament pointed end-capping domains. J Biol Chem. 278:40000–40009. Frantz C, Barreiro G, Dominguez L, Chen X, Eddy R, Condeelis J, Kelly MJ, Jacobson MP, Barber DL (2008) Cofilin is a pH sensor for actin free barbed end formation: role of phosphoinositide binding. J Cell Biol. 183:865–879. Freeman NL, Chen Z, Horenstein J, Weber A, Field J (1995) An actin monomer binding activity localizes to the carboxyl-terminal half of the Saccharomyces cerevisiae cyclase-associated protein. J Biol Chem. 270:5680–5685. Freeman NL, Lila T, Mintzer KA, Chen Z, Pahk AJ, Ren R, Drubin DG, Field J (1996) A conserved proline-rich region of the Saccharomyces cerevisiae cyclase-associated protein binds SH3 domains and modulates cytoskeletal localization. Mol Cell Biol. 16: 548–556. Fritz-Six KL, Cox PR, Fischer RS, Xu B, Gregorio CC, Zoghbi HY, Fowler VM (2003) Aberrant myofibril assembly in tropomodulin1 null mice leads to aborted heart development and embryonic lethality. J Cell Biol. 163:1033–1044. Galkin VE, Orlova A, Lukoyanova N, Wriggers W, Egelman EH (2001) Actin depolymerizing factor stabilizes an existing state of F-actin and can change the tilt of F-actin subunits. J Cell Biol. 153:75–86. Gandhi M, Achard V, Blanchoin L, Goode BL (2009) Coronin switches roles in actin disassembly depending on the nucleotide state of actin. Mol Cell. 34:364–374. Glenney JR Jr, Weber K (1981) Calcium control of microfilaments: uncoupling of the F-actinsevering and -bundling activity of villin by limited proteolysis in vitro. Proc Natl Acad Sci. 78:2810–2814. Gohla A, Birkenfeld J, Bokoch GM (2005) Chronophin, a novel HAD-type serine protein phosphatase, regulates cofilin-dependent actin dynamics. Nat Cell Biol. 7: 21–29. Goldschmidt-Clermont PJ, Machesky LM, Baldassare JJ, Pollard TD (1990) The actin-binding protein profilin binds to PIP2 and inhibits its hydrolysis by phospholipase C. Science. 247:1575–1578. Goode BL, Drubin DG, Lappalainen P (1998) Regulation of the cortical actin cytoskeleton in budding yeast by twinfilin, a ubiquitous actin monomer-sequestering protein. J Cell Biol. 142:723–733. Gorbatyuk VY, Nosworthy NJ, Robson SA, Bains NP, Maciejewski MW, Dos Remedios CG, King GF (2006) Mapping the phosphoinositide-binding site on chick cofilin explains how PIP2 regulates the cofilin-actin interaction. Mol Cell. 24:511–522. Haarer BK, Lillie SH, Adams AE, Magdolen V, Bandlow W, Brown SS (1990) Purification of profilin from Saccharomyces cerevisiae and analysis of profilin-deficient cells. J Cell Biol. 110:105–114. Hart MC, Cooper JA (1999) Vertebrate isoforms of actin capping protein beta have distinct functions in vivo. J Cell Biol. 147:1287–1298. Hawkins M, Pope B, Maciver SK, Weeds AG (1993) Human actin depolymerizing factor mediates a pH-sensitive destruction of actin filaments. Biochemistry. 32:9985–9993. Hayden SM, Miller PS, Brauweiler A, Bamburg JR (1993) Analysis of the interactions of actin depolymerizing factor with G- and F-actin. Biochemistry. 32:9994–10004.

9

Regulation of the Cytoplasmic Actin Monomer Pool in Actin-based Motility

231

Helfer E, Nevalainen EM, Naumanen P, Romero S, Didry D, Pantaloni D, Lappalainen P, Carlier MF (2006) Mammalian twinfilin sequesters ADP-G-actin and caps filament barbed ends: implications in motility. EMBO J. 25:1184–1195. Hertzog M, Yarmola EG, Didry D, Bubb MR, Carlier MF (2002) Control of actin dynamics by proteins made of beta-thymosin repeats: the actobindin family. J Biol Chem. 277:14786–14792. Hertzog M, van Heijenoort C, Didry D, Gaudier M, Coutant J, Gigant B, Didelot G, Préat T, Knossow M, Guittet E, Carlier MF (2004) The beta-thymosin/WH2 domain; structural basis for the switch from inhibition to promotion of actin assembly. Cell. 117:611–263. Hopmann R, Cooper JA, Miller KG (1996) Actin organization, bristle morphology, and viability are affected by actin capping protein mutations in Drosophila. J Cell Biol. 133:1293–1305. Hotulainen P, Paunola E, Vartiainen MK, Lappalainen P (2005) Actin-depolymerizing factor and cofilin-1 play overlapping roles in promoting rapid F-actin depolymerization in mammalian nonmuscle cells. Mol Biol Cell. 16:649–664. Hubberstey AV, Mottillo EP (2002) Cyclase-associated proteins: CAPacity for linking signal transduction and actin polymerization. FASEB J. 16:487–499. Iwasa JH, Mullins RD (2007) Spatial and temporal relationships between actin-filament nucleation, capping, and disassembly. Curr Biol. 17(5):395–406. Janmey PA, Stossel TP (1987) Modulation of gelsolin function by phosphatidylinositol 4,5bisphosphate. Nature. 325:362–364. Kaksonen M, Toret CP, Drubin DG (2006) Harnessing actin dynamics for clathrin-mediated endocytosis. Nat Rev Mol Cell Biol. 7:404–414. Kim K, Galletta BJ, Schmidt KO, Chang FS, Blumer KJ, Cooper JA (2006) Actin-based motility during endocytosis in budding yeast. Mol Biol Cell. 17:1354–1363. Kim K, McCully ME, Bhattacharya N, Butler B, Sept D, Cooper JA. (2007) Structure/function analysis of the interaction of phosphatidylinositol 4,5-bisphosphate with actin-capping protein: implications for how capping protein binds the actin filament. J Biol Chem. 28:5871–5879. Kiuchi T, Ohashi K, Kurita S, Mizuno K (2007) Cofilin promotes stimulus-induced lamellipodium formation by generating an abundant supply of actin monomers. J Cell Biol. 177: 465–476. Kostyukova AS (2008) Tropomodulins and tropomodulin/tropomyosin interactions. Cell Mol Life Sci. 65:563–569. Ksiazek D, Brandstetter H, Israel L, Bourenkov GP, Katchalova G, Janssen KP, Bartunik HD, Noegel AA, Schleicher M, Holak TA (2003) Structure of the N-terminal domain of the adenylyl cyclase-associated protein (CAP) from Dictyostelium discoideum. Structure. 11: 1171–1178. Kueh HY, Brieher WM, Mitchison TJ (2008) Dynamic stabilization of actin filaments. Proc Natl Acad Sci. 105:16531–16536. Kuhn JR, Pollard TD (2007) Single molecule kinetic analysis of actin filament capping. Polyphosphoinositides do not dissociate capping proteins. J Biol Chem. 282:28014–28024. Lambrechts A, Jonckheere V, Dewitte D, Vandekerckhove J, Ampe C (2002) Mutational analysis of human profilin I reveals a second PI(4, 5)-P2 binding site neighbouring the poly(L-proline) binding site. BMC Biochem. 3:12. Lappalainen P, Drubin DG (1997) Cofilin promotes rapid actin filament turnover in vivo. Nature. 388:78–82. Lappalainen P, Fedorov EV, Fedorov AA, Almo SC, Drubin DG (1997) Essential functions and actin-binding surfaces of yeast cofilin revealed by systematic mutagenesis. EMBO J. 16: 5520–5530. Lassing I, Lindberg U (1985) Specific interaction between phosphatidylinositol 4,5-bisphosphate and profilactin. Nature. 314:472–474. Leyman S, Sidani M, Ritsma L, Waterschoot D, Eddy R, Dewitte D, Debeir O, Decaestecker C, Vandekerckhove J, van Rheenen J, Ampe C, Condeelis J, Van Troys M (2009) Unbalancing the PI(4, 5)P2-Cofilin Interaction Impairs Cell Steering. Mol Biol Cell. 20: 4509–4523.

232

P. Lappalainen et al.

Lila T, Drubin DG (1997) Evidence for physical and functional interactions among two Saccharomyces cerevisiae SH3 domain proteins, an adenylyl cyclase-associated protein and the actin cytoskeleton. Mol Biol Cell. 8:367–385. Littlefield R, Almenar-Queralt A, Fowler VM (2001) Actin dynamics at pointed ends regulates thin filament length in striated muscle. Nat Cell Biol. 3:544–551. Loisel TP, Boujemaa R, Pantaloni D, Carlier MF (1999) Reconstitution of actin-based motility of Listeria and Shigella using pure proteins. Nature. 401:613–616. Low TL, Hu SK, Goldstein AL (1981) Complete amino acid sequence of bovine thymosin beta 4: a thymic hormone that induces terminal deoxynucleotidyl transferase activity in thymocyte populations. Proc Natl Acad Sci. 78:1162–1166. Lu M, Witke W, Kwiatkowski DJ, Kosik KS (1997) Delayed retraction of filopodia in gelsolin null mice. J Cell Biol. 13:1279–1287. Maciver SK, Weeds AG (1994) Actophorin preferentially binds monomeric ADP-actin over ATPbound actin: consequences for cell locomotion. FEBS Lett. 347:251–256. Magdolen V, Oechsner U, Muller G, Bandlow W (1988) The intron-containing gene for yeast profilin (PFY) encodes a vital function. Mol Cell Biol. 8:5108–5115. Mahoney NM, Janmey PA, Almo SC (1997) Structure of the profilin-poly-L-proline complex involved in morphogenesis and cytoskeletal regulation. Nat Struct Biol. 4:953–960. Mannherz HG, Hannappel E (2009) The beta-thymosins: intracellular and extracellular activities of a versatile actin binding protein family. Cell Motil Cytoskeleton. 66:839–581. Maruyama K, Obinata T (1965) Presence of beta-actinin in the soluble fraction of the muscle cells of the chick embryo. J Biochem. 57:575–577. Mattila PK, Salminen M, Yamashiro T, Lappalainen P (2003) Mouse MIM, a tissue-specific regulator of cytoskeletal dynamics, interacts with ATP-actin monomers through its C-terminal WH2 domain. J Biol Chem. 278:8452–8459. Mattila PK, Quintero-Monzon O, Kugler J, Moseley JB, Almo SC, Lappalainen P, Goode BL (2004) A high-affinity interaction with ADP-actin monomers underlies the mechanism and in vivo function of Srv2/cyclase-associated protein. Mol Biol Cell. 15:5158–5171. McGough A, Pope B, Chiu W, Weeds A (1997) Cofilin changes the twist of F-actin: implications for actin filament dynamics and cellular function. J Cell Biol. 138:771–781. McGough A, Chiu W (1999) ADF/cofilin weakens lateral contacts in the actin filament. J Mol Biol. 291:513–519. Mejillano MR, Kojima S, Applewhite DA, Gertler FB, Svitkina TM, Borisy GG (2004) Lamellipodial versus filopodial mode of the actin nanomachinery: pivotal role of the filament barbed end. Cell. 118:363–373. Menna E, Disanza A, Cagnoli C, Schenk U, Gelsomino G, Frittoli E, Hertzog M, Offenhauser N, Sawallisch C, Kreienkamp HJ, Gertler FB, Di Fiore PP, Scita G, Matteoli M (2009) Eps8 regulates axonal filopodia in hippocampal neurons in response to brain-derived neurotrophic factor (BDNF). PLoS Biol. 7:e1000138. Moriyama K, Yahara I (2002) Human CAP1 is a key factor in the recycling of cofilin and actin for rapid actin turnover. J Cell Sci. 115:1591–1601. Moseley JB, Okada K, Balcer HI, Kovar DR, Pollard TD, Goode BL (2006) Twinfilin is an actinfilament-severing protein and promotes rapid turnover of actin structures in vivo. J Cell Sci. 119:1547–1557. Nag S, Ma Q, Wang H, Chumnarnsilpa S, Lee WL, Larsson M, Kannan B, Hernandez-Valladares M, Burtnick LD, Robinson RC (2009) Ca2+ binding by domain 2 plays a critical role in the activation and stabilization of gelsolin. Proc Natl Acad Sci. 106:13713–13718. Narita A, Takeda S, Yamashita A, Maéda Y (2006) Structural basis of actin filament capping at the barbed-end: a cryo-electron microscopy study. EMBO J. 25:5626–5633. Nevalainen EM, Skwarek-Maruszewska A, Braun A, Moser M, Lappalainen P (2009) Two biochemically distinct and tissue-specific twinfilin isoforms are generated from the mouse Twf2 gene by alternative promoter usage. Biochem J. 417:593–600.

9

Regulation of the Cytoplasmic Actin Monomer Pool in Actin-based Motility

233

Niwa R, Nagata-Ohashi K, Takeichi M, Mizuno K, Uemura T (2002) Control of actin reorganization by Slingshot, a family of phosphatases that dephosphorylate ADF/cofilin. Cell. 108:233–246. Noegel AA, Rivero F, Albrecht R, Janssen KP, Kohler J, Parent CA, Schleicher M (1999) Assessing the role of the ASP56/CAP homologue of Dictyostelium discoideum and the requirements for subcellular localization. J Cell Sci. 112:3195–3203. Obermann H, Raabe I, Balvers M, Brunswig B, Schulze W, Kirchhoff C (2005) Novel testisexpressed profilin IV associated with acrosome biogenesis and spermatid elongation. Mol Hum Reprod. 11:53–64. Ojala PJ, Paavilainen V, Lappalainen P (2001) Identification of yeast cofilin residues specific for actin monomer and PIP2 binding. Biochemistry. 40:15562–15569. Ojala PJ, Paavilainen VO, Vartiainen MK, Tuma R, Weeds AG, Lappalainen P (2002) The two ADF-H domains of twinfilin play functionally distinct roles in interactions with actin monomers. Mol Biol Cell. 13:3811–3821. Ono S, Minami N, Abe H, Obinata T (1994) Characterization of a novel cofilin isoform that is predominantly expressed in mammalian skeletal muscle. J Biol Chem. 269:15280–15286. Ono S (2003) Regulation of actin filament dynamics by actin depolymerizing factor/cofilin and actin-interacting protein 1: new blades for twisted filaments. Biochemistry. 42:13363–13370. Paavilainen VO, Bertling E, Falck S, Lappalainen P (2004) Regulation of cytoskeletal dynamics by actin-monomer-binding proteins. Trends Cell Biol. 14:386–394. Paavilainen VO, Hellman M, Helfer E, Bovellan M, Annila A, Carlier MF, Permi P, Lappalainen P (2007) Structural basis and evolutionary origin of actin filament capping by twinfilin. Proc Natl Acad Sci. 104:3113–3118. Paavilainen VO, Oksanen E, Goldman A, Lappalainen P (2008) Structure of the actindepolymerizing factor homology domain in complex with actin. J Cell Biol. 182:51–59. Palmgren S, Ojala PJ, Wear MA, Cooper JA, Lappalainen P (2001) Interactions with PIP2, ADPactin monomers, and capping protein regulate the activity and localization of yeast twinfilin. J Cell Biol. 155:251–260. Pantaloni D, Carlier MF (1993) How profilin promotes actin filament assembly in the presence of thymosin beta 4. Cell. 75:1007–1014. Pantaloni D, Le Clainche C, Carlier MF (2001) Mechanism of actin-based motility. Science. 292:1502–1506. Pelkmans L, Fava E, Grabner H, Hannus M, Habermann B, Krausz E, Zerial M (2005) Genomewide analysis of human kinases in clathrin- and caveolae/raft-mediated endocytosis. Nature. 436:78–86. Pollard TD, Cooper JA (1986) Actin and actin-binding proteins. A critical evaluation of mechanisms and functions. Annu Rev Biochem. 55:987–1035. Pollard TD, Blanchoin L, Mullins RD (2000) Molecular mechanisms controlling actin filament dynamics in nonmuscle cells. Annu Rev Biophys Biomol Struct. 29:545–756. Pollard TD, Borisy GG. (2003) Cellular motility driven by assembly and disassembly of actin filaments. Cell. 112:453–465. Quintero-Monzon O, Jonasson EM, Bertling E, Talarico L, Chaudhry F, Sihvo M, Lappalainen P, Goode BL (2009) Reconstitution and dissection of the 600-kDa Srv2/CAP complex: roles for oligomerization and cofilin-actin binding in driving actin turnover. J Biol Chem. 284: 10923–10934. Reinhard M, Giehl K, Abel K, Haffner C, Jarchau T, Hoppe V, Jockusch BM, Walter U (1995) The proline-rich focal adhesion and microfilament protein VASP is a ligand for profilins. EMBO J. 14:1583–1589. Ressad F, Didry D, Xia GX, Hong Y, Chua NH, Pantaloni D, Carlier MF (1998) Kinetic analysis of the interaction of actin-depolymerizing factor (ADF)/cofilin with G- and F-actins. Comparison of plant and human ADFs and effect of phosphorylation. J Biol Chem. 273:20894–20902. Rogers SL, Wiedemann U, Stuurman N, Vale RD (2003) Molecular requirements for actin-based lamella formation in Drosophila S2 cells. J Cell Biol. 162:1079–88.

234

P. Lappalainen et al.

Safer D, Elzinga M, Nachmias VT (1991) Thymosin beta 4 and Fx, an actin-sequestering peptide, are indistinguishable. J Biol Chem. 266:4029–4032. Schafer DA, Jennings PB, Cooper JA (1996) Dynamics of capping protein and actin assembly in vitro: uncapping barbed ends by polyphosphoinositides. J Cell Biol. 135:169–179. Schutt CE, Myslik JC, Rozycki MD, Goonesekere NC, Lindberg U (1993) The structure of crystalline profilin-beta-actin. Nature. 365:810–816. Silacci P, Mazzolai L, Gauci C, Stergiopulos N, Yin HL, Hayoz D (2004) Gelsolin superfamily proteins: key regulators of cellular functions. Cell Mol Life Sci. 61:2614–2623. Scita G, Tenca P, Areces LB, Tocchetti A, Frittoli E, Giardina G, Ponzanelli I, Sini P, Innocenti M, Di Fiore PP (2001) An effector region in Eps8 is responsible for the activation of the Rac-specific GEF activity of Sos-1 and for the proper localization of the Rac-based actin-polymerizing machine. J Cell Biol. 154:1031–1044. Skare P, Karlsson R (2002) Evidence for two interaction regions for phosphatidylinositol(4, 5)bisphosphate on mammalian profilin I. FEBS Lett. 522:119–124. Southwick FS, DiNubile MJ (1986) Rabbit alveolar macrophages contain a Ca2+-sensitive, 41,000-dalton protein which reversibly blocks the “barbed” ends of actin filaments but does not sever them. J Biol Chem. 261:14191–14195. Swiston J, Hubberstey A, Yu G, Young D (1995) Differential expression of CAP and CAP2 in adult rat tissues. Gene. 165:273–277. Toshima J, Toshima JY, Amano T, Yang N, Narumiya S, Mizuno K (2001) Cofilin phosphorylation by protein kinase testicular protein kinase 1 and its role in integrin-mediated actin reorganization and focal adhesion formation. Mol Biol Cell. 12:1131–1145. van Rheenen J, Song X, van Roosmalen W, Cammer M, Chen X, Desmarais V, Yip SC, Backer JM, Eddy RJ, Condeelis JS (2007) EGF-induced PIP2 hydrolysis releases and activates cofilin locally in carcinoma cells. J Cell Biol. 179:1247–1259. Vartiainen MK, Mustonen T, Mattila PK, Ojala PJ, Thesleff I, Partanen J, Lappalainen P (2002) The three mouse actin-depolymerizing factor/cofilins evolved to fulfill cell-type-specific requirements for actin dynamics. Mol Biol Cell. 13:183–194. Wahlström G, Vartiainen MK, Yamamoto L, Mattila PK, Lappalainen P, Heino TI (2001) Twinfilin is required for actin-dependent developmental processes in Drosophila. J Cell Biol. 155: 787–795. Wear MA, Yamashita A, Kim K, Maéda Y, Cooper JA (2003) How capping protein binds the barbed end of the actin filament. Curr Biol. 13:1531–1537. Weber A, Pennise CR, Fowler VM (1999) Tropomodulin increases the critical concentration of barbed end-capped actin filaments by converting ADP.P(i)-actin to ADP-actin at all pointed filament ends. J Biol Chem. 274:34637–34645. Witke W, Podtelejnikov AV, Di Nardo A, Sutherland JD, Gurniak CB, Dotti C, Mann M (1998) In mouse brain profilin I and profilin II associate with regulators of the endocytic pathway and actin assembly. EMBO J. 17:967–976. Witke W, Li W, Kwiatkowski DJ, Southwick FS (2001a) Comparisons of CapG and gelsolinnull macrophages: demonstration of a unique role for CapG in receptor-mediated ruffling, phagocytosis, and vesicle rocketing. J Cell Biol. 154:775–784. Witke W, Sutherland JD, Sharpe A, Arai M, Kwiatkowski DJ (2001b) Profilin I is essential for cell survival and cell division in early mouse development. Proc Natl Acad Sci. 98: 3832–3836. Witke W (2004) The role of profilin complexes in cell motility and other cellular processes. Trends Cell Biol. 14:461–469. Wolven AK, Belmont LD, Mahoney NM, Almo SC, Drubin DG (2000) In vivo importance of actin nucleotide exchange catalyzed by profiling. J Cell Biol. 150:895–904. Yamashita A, Maeda K, Maéda Y (2003) Crystal structure of CapZ: structural basis for actin filament barbed end capping. EMBO J. 22:1529–1538. Yang C, Pring M, Wear MA, Huang M, Cooper JA, Svitkina TM, Zigmond SH (2005) Mammalian CARMIL inhibits actin filament capping by capping protein. Dev Cell. 9:209–221.

9

Regulation of the Cytoplasmic Actin Monomer Pool in Actin-based Motility

235

Yin HL, Stossel TP (1979) Control of cytoplasmic actin gel-sol transformation by gelsolin, a calcium-dependent regulatory protein. Nature. 28:583–586. Yonezawa N, Nishida E, Sakai H (1985) pH control of actin polymerization by cofilin. J Biol Chem. 260:14410–14412. Yonezawa N, Nishida E, Iida K, Yahara I, Sakai H (1990) Inhibition of the interactions of cofilin, destrin, and deoxyribonuclease I with actin by phosphoinositides. J Biol Chem. 265: 8382–8386. Yonezawa N, Homma Y, Yahara I, Sakai H, Nishida E (1991) A short sequence responsible for both phosphoinositide binding and actin binding activities of cofilin. J Biol Chem. 266: 17218–17221.

Chapter 10

From Molecules to Movement: In Vitro Reconstitution of Self-Organized Actin-based Motile Processes Marie-France Carlier and Dominique Pantaloni

Abstract This chapter shows how a bottom-up approach has been instrumental in understanding and reconstituting actin-based movement from a minimum set of 5 individual purified proteins. We delineate the essential thermodynamic features that support actin-based movement as follows. We demonstrate first that the regulated treadmilling of actin filaments fuels the growth of barbed ends to develop protrusive or propulsive force against a functionalized soft or solid particle. Second, global inhibition of actin assembly in the medium and local stimulation of actin assembly restricts filament growth to specific sites and controls the filament size and its transient growth period against the particle; hence, stationary propulsion of the particle results from a balance between creation of new filaments and death of these filaments. Third, transient or permanent links between the particle-immobilized actin-nucleating machinery play an important role in maintaining the morphology of the force-producing actin meshwork and the velocity of the propulsion. These principles allow reconstitution of propulsive branched actin networks and forminbased processive growth of actin bundles. Finally upgrading the complexity of the bottom-up approach offers perspectives in analyzing more complex actin-based processes that occur in axis patterning and morphogenesis, as well as understanding the physical-chemical basis for the coordinated turnover of various actin structures during cell migration.

Contents 10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.2 Actin Assembly Dynamics: From the Simple In Vitro Assays to the Complex Physiological Context . . . . . . . . . . . . . . . . . . . . . . . 10.2.1 Intrinsic Properties of Actin Self-Assembly . . . . . . . . . . . . . .

238 239 239

M.-F. Carlier (B) CNRS, Gif-sur-Yvette, France e-mail: [email protected]

M.-F. Carlier (ed.), Actin-based Motility, DOI 10.1007/978-90-481-9301-1_10, C Springer Science+Business Media B.V. 2010

237

238

M.-F. Carlier and D. Pantaloni

10.2.2 Regulated Treadmilling . . . . . . . . . . . . . . . . . . . 10.2.3 Global Inhibition of Actin Assembly Restricts Filament Growth to Specific Sites . . . . . . . . . . . . . . . . . . . . . . 10.2.4 Reconstitution of Listeria and Shigella Propulsion by Assembly of a Dendritic Branched Filament Array . . . . . 10.2.5 From Listeria to Actin-based Motility Assays Using Chemically and Geometrically Controlled Functionalized Particles . . . . . 10.2.6 Propulsive Movement of Formin-Coated Particles by Processive Assembly of Actin Parallel Bundles . . . . . . . 10.2.7 Functional Role of Permanent or Transient Links Between the Particle-Immobilized Actin Nucleating Machinery and the Growing Actin Meshwork . . . . . . . . . . . . . . 10.3 Perspectives of Reconstituted Motility Assays . . . . . . . . . . . . . 10.3.1 Upgrading Motility Assays to a Higher Level of Biological Complexity . . . . . . . . . . . . . . . . . . 10.3.2 Structural Analysis and Dynamic Imaging of Reconstituted Motile Actin Arrays . . . . . . . . . . . . . . . . . . . . 10.3.3 Toward Coordinated Actin Dynamics in Cell Movement . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . .

241

. . . . . 241 . . . . . 244 . . . . . 245 . . . . . 247

. . . . . 248 . . . . 249 . . . . . 249 . . . . . 250 . . . . 251 . . . . 251

10.1 Introduction The living cell is often viewed as a factory in which modular nanomachines work in a spatially and temporally coordinated fashion to organize various intra muros lifemaintenance activities as well as extra muros communication with environment, mediated by changes in shape and movements. These nanomachines are highly complex large assemblies, in which protein-protein but also protein-nucleic acid or protein-lipid interactions are essential in the production of work. The directional nature of work production, like synthesis or degradation of biological materials, intracellular traffic, chromosome or plasmid segregation, cell migration, cytokinesis, implies that dissipation of energy regulates the interfaces between the components of the machineries. Motile processes that result from self-assembly of actin filaments are demonstrative of such dissipative reactions that produce force and directional movement. The other chapters of this book focus on the description and analysis of the many dynamic characteristics and morphological aspects of actin-based motile processes in cell function or organism development. The present article, in contrast, shows how a reductionist approach can identify the minimum number of biochemical functions needed to reconstitute directional movement and the cell components that are required and sufficient to fulfil these functions. The in vitro reconstitution endeavor is not new in biology. The concept was already at the heart of research in metabolism and gene replication and has experienced renewed interest with cell biology advances in identification and live monitoring of large macromolecular

10

From Molecules to Movement

239

assemblies in action. Based on Richard Feynman’s word “What I cannot create, I do not understand”, the bottom-up reconstitution has given birth to synthetic biology (Schwille and Diez, 2009). Reconstitution of a minimum system in which actin self-assembly occurs in a consitutive fashion to move a particle is an important primary step because nature used variations of this elementary theme to generate a large number of morphogenetic and motile processes of higher complexity involved in shaping organisms, organizing immune response, wound healing or synaptic plasticity. Both the robustness of the intrinsic assembly properties of actin and the structural versatility of this fascinating protein make this possible. The notion that spatially restricted and polarized self-assembly of actin was able by itself, in the absence of myosin, to generate a force sufficient to deform membranes and cause cellular protrusions, invaginations or vesicle scission was not immediately accepted, at a time when molecular motors were the main acknowledged actors in cell motility.

10.2 Actin Assembly Dynamics: From the Simple In Vitro Assays to the Complex Physiological Context 10.2.1 Intrinsic Properties of Actin Self-Assembly Actin is one of the most abundant highly conserved globular protein in eukaryotes, and has bacterial ancestors that display similar motile functions (see Chapter 14). The first main property of all actins is to self-assemble into helical, polar filaments. Monomeric (G for globular) actin predominates at very low ionic strength, and self-assembles spontaneously into filaments (F-actin for filamentous actin) upon increasing ionic strength. F-actin predominates in physiological ionic conditions. Actin filaments form, with microtubules and intermediate filaments, the cytoskeleton or internal architecture of the cell. ATP is tightly bound to monomeric or globular (G) actin and is hydrolyzed following self-assembly, in a single turnover reaction, leading to a polar helical filament made of ADP subunits (F-ADP-actin), inorganic phosphate ion being released in solution. Hence actin is an ATPase, its second main property. The fact that ATP hydrolysis is truly irreversible and subsequent to, but not required for polymerization creates an energetic difference between the two ends of the filament. In the presence of ATP in the medium, ADP-Factin filaments at a steady state co-exist with monomeric ATP-G-actin maintained at a stationary concentration, which feeds a flux of subunits through the polymer (Fig. 10.1). Net assembly of ATP-G-actin at one end balances net disassembly of ADP-F-actin at the other end, in a process called treadmilling or “head-to-tail polymerization”, conceptually discovered by Wegner (1976). The dissipative cycle of actin assembly-disassembly at steady state is pivotal in all actin-based motile processes. Like all energy-consuming reactions in living organisms, it is highly regulated, in rate and efficiency.

240

M.-F. Carlier and D. Pantaloni

Fig. 10.1 Actin filament treadmilling and its regulation by ADF and profilin. Top panel shows the basic slow treadmilling of pure actin assembled at a steady state in ATP. Monomer-polymer exchange events take place at both ends of the filament, only the net excess of dissociation events of ADP-actin at the pointed ends (k= 0.1 s−1 ) and association events of ATP-actin at the barbed ends are represented by arrows. Nucleotide exchange takes place on monomeric actin only. Bottom panel shows how the treadmilling is enhanced in a synergistic fashion by ADF/cofilin and profilin. ADF (represented in blue) binds to ADP-F-actin and causes a structural destabilization of the actin-actin interactions in the filament, which results, at steady state, in a large increase in the rate of dissociation of ADP-actin from the pointed ends. Nucleotide exchange is blocked on ADF-ADP-G-actin, but can occur upon dissociation of ADF. At subsaturating amounts of ADF, the increased flux of depolymerization leads to a higher stationary level of ATP-G-actin. Hence the turnover rate of the filament is enhanced. Profilin binds preferentially ATP-G-actin, accelerates nucleotide exchange on G-actin and makes with ATP-G-actin a complex that associates productively to barbed ends only. Addition of profilin to F-actin at steady state in the presence of ADF thus shifts all binding equilibria toward the predominant profilin-ATP-G-actin, thus facilitating the vectorial turnover of actin filaments

10

From Molecules to Movement

241

10.2.2 Regulated Treadmilling The turnover of filaments assembled from pure actin is very slow, a 3 μm long filament being renewed by treadmilling in about 2 h (Wegner, 1982). FRAP experiments in paradigm actin-based processes like extending lamellipodia strongly suggested, 25 years ago (Wang, 1985), that the turnover of filaments, although 100-fold faster than treadmilling of pure actin, was mediated by a treadmilling mechanism, which was fully confirmed later (Small, 1995; Small et al., 2002; Lai et al., 2008). In vitro functional analysis of several actin regulatory proteins enlighten the mechanisms by which treadmilling can be enhanced in vitro up to values observed in vivo. Accelerating the rate limiting step in the ATPase cycle of actin and improving the efficiency (or vectoriality) of treadmilling are the key targets. These effects are mediated by the synergistic effects of two actin regulatory proteins, ADF/cofilin and profilin (Fig. 10.1). ADF/cofilin is a small protein that binds prefentially the ADP-bound form of F- or G-actin. Its binding to ADP-F-actin destabilizes actin-actin contacts in the filament, thus enhancing, at steady state of assembly, the rate of pointed end depolymerization of ADP-F-actin that governs the turnover rate (Carlier et al., 1997). Profilin binds specifically ATP-actin in a complex that has the unique property to productively associate with barbed but not with pointed ends (Pollard and Cooper, 1984; Pantaloni and Carlier, 1993). Thus profilin promotes the vectoriality in actin association events to barbed ends in the treadmilling process. In vitro assays verify that in the presence of both ADF/cofilin and profilin, filaments turnover in vitro as fast as in vivo, and that the enhanced dynamics of actin filaments are associated with a large steady-state amount of polymerizable G-actin (Didry et al., 1998). Like in most regulated biological systems, the enhancement in treadmilling provided by ADF and profilin is balanced by antagonists such as tropomyosin, which stabilize filaments in a conformation that is incompatible with ADF binding.

10.2.3 Global Inhibition of Actin Assembly Restricts Filament Growth to Specific Sites Evidence coming from live cell imaging of the turnover of actin filaments in motile processes indicates that the nucleation and transient growth (turnover) of filaments is spatially restricted. For instance, filaments are constantly initiated at the leading edge (Iwasa and Mullins, 2007), or at the surface of endocytic vesicles (Taunton, 2001) or of intracellular pathogens (Cossart, 2000; Egile et al., 1999; Welch et al., 1997, 1998; Jeng et al., 2004). Filopodia are initiated and elongate from a tipassociated macromolecular complex (Medalia et al., 2007; Block et al., 2008). Researchers were puzzled for a long time by the mechanisms and multiple regulators apparently needed to promote local stimulation of actin assembly while inhibiting assembly in other regions of the cell. The concept of treadmilling in fact solves this puzzle easily using the activity of capping proteins. Capping proteins block barbed end assembly, thus cooperate with ADF and profilin to increase the

242

M.-F. Carlier and D. Pantaloni

concentration of polymerizable monomeric actin. The newly created barbed ends thus grow faster due to the higher concentration of available monomers (Carlier and Pantaloni, 1997). However they grow only transiently, because capping proteins also block spontaneously formed barbed end nuclei. Local stimulation of actin polymerization by site-bound nucleators cannot be effective unless the newly formed growing barbed ends escape barbed end capping at least transiently. This condition may be fulfilled theoretically in several ways. Either the nucleating machinery interacts with the newly created filament barbed ends in a manner that competes out cappers but does not inhibit actin or profilinactin association to the barbed end (this is what processive motors of actin assembly do); or the nucleating machinery induces a long lived, hysteretic, structure of the barbed end that is incompatible with binding of cappers; or – for instance in the case of filament branching by WASP-Arp2/3 machinery – the branched filaments are incorporated in an intricate meshwork that mechanically maintains the barbed ends in the vicinity of clustered, membrane-immobilized filament branching machineries, thus facilitating branching versus capping. In conclusion capping proteins enhance the rate of barbed end growth and control the length or life-time of newly created filaments (Fig. 10.2). In summary, the concept of treadmilling and its regulation by proteins known to play an important role in a large number of motile processes guided the general strategy for reconstitution of actin-based propulsion of a functionalized particle (Loisel et al., 1999; Pantaloni et al., 2001). Directional movement resulted from local stimulation of actin assembly at the surface of a particle. Constant creation (“birth”) of filaments is balanced by capping (“death”) that arrests their growth. Newly created filaments grow transiently in a medium that acts as a chemostat by maintaining a stationary supply of ATP-G-actin at a concentration above the critical concentration of barbed ends. In this medium all pointed ends depolymerize, thus maintaining a treadmilling cycle that supports a stationary perpetuated meshwork. Treadmilling organizes the polarity of movement. ATP is the source of energy. The chemostat consists of a solution of F-actin in ATP in the presence of profilin, ADF/cofilin and capping proteins (Fig. 10.3). Force is produced by site-directed polarized growth of filaments. Because the reactions of actin assembly develop in a medium of low Reynolds number, movement stops as soon as force – i.e. filament growth - is arrested. The molecular mechanism of initiation and polarized growth of filaments by the particle-immobilized protein machinery strongly affects the morphology and mechanical properties (elasticity) of the actin meshwork that drives particle movement. Remarkably, force and movement seem to strongly depend on the rate at which filaments elongate, independently of the structure of the assembled network. Two main protein machineries are known so far to operate in vivo to produce force by site-directed actin assembly: the WASP family proteins which use Arp2/3 complex to multiply filaments by branching, and the formins that act as processive motors of actin assembly (Pollard, 2007; Goode and Eck, 2007; Pollard and Cooper, 2009). Both use similar principles of regulated treadmilling. Reconstitution of actin-based movement has been performed with both.

10

From Molecules to Movement

243

Fig. 10.2 Principle in reconstitution of actin-based motility: barbed end capping funnels treadmilling and balances the production of new growing filaments by site-directed branching. Capping proteins (yellow circles) synergize with ADF to establish a high stationary concentration of actin monomers feeding rapid barbed end growth. Barbed ends created in an autocatlytic fashion by site-directed branching by immobilized N-WASP grow transiently until being capped. Antagonism between branching and capping maintains a stationary number of growing filaments in the vicinity of the surface (bead or membrane). Purple ellipse: N-WASP; orange ellipse: signaling molecule connecting N-WASP to the membrane (yellow). Red circles: ATP/ADP-Pi-actin; green: Arp2/3 complex

244

M.-F. Carlier and D. Pantaloni

Fig. 10.3 Actin meshwork assembled at the surface of a functionalized giant unilamellar vesicle. Toppanel: two-color fluorescence microscopy image of the actin tail (rhodamine-actin) assembled at the surface of a 10 μm diameter GUV functionalized with N-WASP (green fluorescence). Bottom panel: scheme of the assembly of a dendritic filament array within the treadmilling mechanism, as described in Figs. 10.1 and 10.2

10.2.4 Reconstitution of Listeria and Shigella Propulsion by Assembly of a Dendritic Branched Filament Array A small number of intracellular pathogens such as the bacteria Listeria monocytogenes (Tilney and Portnoy, 1989; Theriot et al., 1992; Kocks et al., 1992) or Shigella flexneri (Bernardini et al., 1989), Rickettsia (Heinzen et al., 1993) and viruses like vaccinia virus (Frischknecht et al., 1999) and other viruses (Dodding and Way, 2009), harness the actin cytoskeleton of the host cell and induce constitutive site-directed actin assembly at their surface, thus mimicking reactions that occur in the physiological context at the leading edge or in other actin-based motile processes. The Listeria/Shigella systems were very popular in the 1990’s as tools to unravel the molecular mechanism of actin-based movement and identify the cell components that support it, because the movement of Listeria could be monitored

10

From Molecules to Movement

245

in acellular extracts of Xenopus egg or non stimulated platelets (Theriot et al., 1994; Welch et al;, 1997; Carlier et al., 1997; Laurent et al., 1999). Pathogens turned out to initiate the cascade of reactions leading to actin assembly either by expressing at their surface a protein that behaves as a functional homolog of WASP family proteins (ActA for Listeria (Welch et al., 1998), RickA for Rickettsia (Jeng et al., 2004) or that mimics Cdc42 to activate N-WASP, as does IcsA protein for Shigella (Egile et al., 1999). Listeria or Shigella propel by constantly multiplying filaments by branching. Filament branching is an autocatalytic reaction performed by the WASP family proteins. These “filament branching enzymes”use Arp2/3 complex to generate a branch junction in a “mother”filament, giving birth to a “daughter”filament. The barbed ends of mother and daughter filaments grow simultaneously following branching, forming a “y”structure. At a sufficiently high surface density of ActA or N-WASP, continuous branching and transient growth of branched filaments creates an intricate dendritic meshwork which is initially non polarized, and which following symmetry breaking, displays high enough friction against the solvent for insertional polymerization to push the bacteria forward in a directional fashion (Van Oudenaarden and Theriot, 1999; Cameron et al., 1999, 2000). Propulsive movement of Listeria and Shigella thus could be observed by placing the bacteria in a reconstituted motility assay containing F-actin assembled in the presence of Arp2/3, ADF/cofilin, profilin, a capping protein and a large enough supply of ATP (Loisel et al., 1999). This medium buffered the concentration of polymerizable actin at a high enough concentration to support transient rapid barbed end growth of filaments created at the surface of the bacteria.

10.2.5 From Listeria to Actin-based Motility Assays Using Chemically and Geometrically Controlled Functionalized Particles Listeria and Shigella were useful but uncontrolled mimickers of the physiological actin-based motile processes, which are orchestratted by the mechanical and chemical interplay of membrane and cytoskeleton dynamics. To get closer to the cellular context, solid (polystyrene microbeads) or soft (giant unilamellar vesicles) were functionalized with N-WASP (Fig. 10.4) and their propulsion was monitored in the reconstituted motility assays (Pantaloni et al., 2001; Wiesner et al., 2003; Upadhyaya et al., 2003; Upadhyaya and van Oudenaarden, 2003; Heuvingh et al., 2007; Delatour et al., 2008). The full control of the coating density, of the concentrations of all components in the medium opened avenues for physical studies such as force production or velocity, that are fully described in chapters written by Fletcher and Sykes (Giardini et al., 2003; Marcy et al., 2004; Parekh et al., 2005; Shaevitz and Fletcher, 2007) and trajectory analysis (Schmidt et al., 2008). Use of fluorescently labeled actin and /or Arp2/3 complex, or N-WASP and analysis of their distribution in the actin meshwork during movement (Fig. 10.5) provided

246

M.-F. Carlier and D. Pantaloni

Fig. 10.4 Images of reconstituted actin-based propulsion of various functionalized particles. (a) Phase contrast image of Listeria in reconstituted motility solution of F-actin, ADF/cofilin, Arp2/3,

10

From Molecules to Movement

247

insight into molecular mechanisms responsible for force production (Wiesner et al., 2003; Delatour et al., 2008). Modeling studies, fully described in chapter written by Mogilner and Carlsson, have been carried out to correlate polymerization, mechanical properties of filaments and force production (Mogilner and Oster, 2003; Mogilner and Rubinstein, 2005; Mogilner, 2006 for review; Carlsson, 2003; Dayel et al., 2009; Burroughs and Marenduzzo, 2007)

10.2.6 Propulsive Movement of Formin-Coated Particles by Processive Assembly of Actin Parallel Bundles The discovery that formins nucleated and catalyzed rapid processive assembly of actin filament bundles (Pollard, 2007; Goode and Eck, 2007, for reviews) prompted the development of actin-based motility assays to challenge the view that formin-based motile processes in vivo were also mediated by regulated teadmilling. Propulsion of formin-coated polystyrene microspheres was well reconstituted in a motility medium similar to the one used for propulsion of N-WASP-coated beads, containing F-actin in the presence of ATP, profilin and ADF/cofilin, thus validating the principle of treadmilling (Romero et al., 2004). Profilin was required for constitutive in vitro formin-based movement, as it is for all known in vivo functions of formins. The major difference between the motility mediated by formation of a dendritic array and by processive assembly of linear actin bundles resides in the different regulatory effect of capping proteins: capping proteins are required for the formation and maintenance of dendritic actin arrays whereas they accelerate forminbased propulsive movement for a transient period of time, but eventually compete with formins for barbed end binding, causing arrest of movement. These observations account for in vivo observations that cappers favor lamellipodia and inhibit filopodia formation (Mejillano et al., 2004; Ditlev et al., 2009; Le Clainche and Carlier, 2008 for review).

Fig. 10.4 (continued) profilin and gelsolin. (b) Phase contrast image of the lamellar propulsive network assembled at the surface of a glass rod (1 μm diameter, 30 mm inlength), functionalized with N-WASP, mimicking a lamellipodial actin array. Note disassembly at the rear of the meshwork. The medium is as in a. (c) Phase contrast image of N-WASP-functionalized polystyrene beads of 0.5, 1 and 3 μm diameter propelling in reconstituted motility assay (conditions as in a and b). (d) Double fluorescence microscopy image of a N-WASP-functionalized GUV (10 μm in diameter, green fluorescence) propelling in the presence of rhodamine labeled actin (red fluorescence). Conditions are as in a, b, c. (e) TIRFM time lapse observation of the growth of a single filament mediated by processive assembly catalyzed by FH1-FH2 at the surface of a bead, placed in the pprsence of F-actin, profilin and ADF/cofilin. Time is in minutes. Note the detachment of the filament after 22 min. f: time lapse luorescence microscopy images of the propulsion of beads functionalized with the FH1-FH2 (constitutively active) domain of formin mDia1 and placed in a medium consiting of F-actin in the presence of ADF/cofilin and profilin. Processive insertional barbed end assembly of profilin-actin catalyzed by FH1-FH2 at the surface of the bead generates rapid propulsion, using treadmilling. Time is in minutes

248

M.-F. Carlier and D. Pantaloni

Fig. 10.5 Incorporation of Arp2/3 complex into the dendritic actin msehwork formed at the surface of a N-WASP functionalized bead. (a, b) double fluorescence image showing the labeled actin (red) and Arp2/3 complex (green) during stationary propulsion. (c, d) Time lapse phase contrast and fluorescence images of the incorporation of Arp2/3 complex into the actin meshwork during propulsive motion. Following initiation of movement in the presence of unlabelled Arp2/3, a pulse of fluorescently labeled Arp2/3 (green) was introduced. Arp2/3 binds instantly to N-WASP at the bead surface and is gradually incorporated into the growing actin meshwork, at the same rate as actin

10.2.7 Functional Role of Permanent or Transient Links Between the Particle-Immobilized Actin Nucleating Machinery and the Growing Actin Meshwork Actin-based movement is directional. Even though the trajectories of N-WASP functionalized solid or soft particles or of Listeria are not straight over distances of several tens of microns and display some curvature, the persistence length of

10

From Molecules to Movement

249

actin tails is much larger than the size of the propelling particles. Maintenance of direction and of the morphology of the filament array depends at least in part on the life-time of the connections made between growing actin filaments and the particle-bound protein machine that organizes filament nucleation and growth. In the simple case of formins, these links are quasi-permanent, due to the processive nature of the assembly. In a processively assembled actin bundle made of hundreds of filaments, the stochastic unfrequent detachment of some filaments does not compromise the integrity of the growing bundle. In another putative extreme case, if nucleated filaments detached from the particle immediately following the nucleation step, building of a cohesive meshwork growing in a defined direction over long distances and producing directional force might be difficult to achieve. In actin-based propulsion driven by formation of a dendritic branched filament array, early evidence for robust attachment of the growing actin tail to Listeria (Gerbal et al., 2000) indicated that at least transient links were formed between the growing filaments and the particle surface. The frequency of these links, mediated by Arp2/3-N-WASP functional interaction, control the morphology of the meshwork, the distribution of WASP machineries associated with the fluid lipid bilayer, thus determining motile behavior (Delatour et al., 2008; Weisswange et al., 2009). The dynamics of these attachment-detachment of filaments are thought to generate some processive behavior at the filament collective level, which results from the cycles of filament nucleation by branching at the membrane, dissociation and growth, according to a tethered ratchet model (Mogilner and Oster, 2003). The chemistry and kinetics of the elementary steps of this cycle are not fully understood. The structure of the transient large complexes formed during filament branching and dissociation at the particle surface is not elucidated either (Bugyi and Carlier, 2010 for review). A full control of the chemistry of these molecular interactions would allow a control of movement.

10.3 Perspectives of Reconstituted Motility Assays 10.3.1 Upgrading Motility Assays to a Higher Level of Biological Complexity The primary reconstituted motility assays were designed to demonstrate the minimum requirements for motility, and identify the essential players, “the bare bones of the cytoskeleton”(Machesky and Cooper, 1999). In vivo, a wide variety of actin-based motile and morphogenetic processes are generated by increasing the complexity of the minimum system, in many ways. For instance, the force developed by growth of dendritic arrays is used to deform membranes in concave or convex geometries, which is likely due to the association of WASP family proteins with proteins possessing BAR domains like ABBA, MIM, IRSp53 (see Chapter 2). In vitro assays might be designed using these additional ingredients to understand actin-mediated crucial steps in endocytosis or intracellular trafficking.

250

M.-F. Carlier and D. Pantaloni

To quote another example, refined regulation of the morphology and plasticity of dendritic arrays is exerted via proteins that affect the rate of filament debranching (dissociation of daughter filaments), like coronin, or by proteins that antagonize the action of ADF/cofilin, like tropomyosin. Specifically, tropomyosin, which binds filaments in competition with ADF, appears to bind and stabilize debranched filaments preferentially (Bugyi et al., 2010). The rate of debranching thus acts as a timer that generates a gradient in localization of tropomyosin and subsequent stabilization of filaments at the rear of the dendritic lamellipodial array. This property explains at least in part the existence of a lamellar array formed of non branched linear filaments at the rear of the lamellipodium. Third, promising future developments of reconstituted motility assays are offered by the possibility to understand the molecular basis of genetic interactions in motile processes. Genetic interactions have frequently been reported among actin-binding proteins involved in brain morphogenesis or early development processes like axis patterning or cellularization. Often such genetically interacting proteins are involved in a complex system in which they regulate actin assembly dynamics in a coupled or synergistic fashion. Reconstituted assays are key in understanding the molecular mechanisms underlying such processes. For instance formins of the formin 2 family (Drosophila Cappuccino, Fmn2) genetically interact with profilin and Spire to elicit massive transient actin assembly in oocyte polarity or assymetric cell division (Manseau and Schüpbach, 1989; Dahlgaard et al., 2007). Addition of Spire to reconstituted formin-based motility assays helps understand how formin, profilin and Spire act in synergy (Bosch et al., 2007). It is likely that in other similar instances reconstituted motility assays may be used to mimic specialized processes and elucidate the underlying mechanism. Fourth, reconstitution of biomimetic membranes functionalized with two or more actin-assembly promoting machineries would constitute a step forward in complexity, which would allow to address, for instance, the coordinated activities of formins and WASP machineries at the leading edge. The possibility to engineer patterned functionalized lipid bilayers further expands the range of feasible experiments to understand how filaments assembled in filopodia and dendritic arrays functionally and spatially segregate in vivo, while cooperating in global motility.

10.3.2 Structural Analysis and Dynamic Imaging of Reconstituted Motile Actin Arrays Cryo-electrotomography is a choice method to analyze the organization of intracellular space, in particular the cytoskeleton (Medalia et al., 2002, 2007). The results are not always unambiguous however, due to the complex cellular environment and the possibly mingled actin arrays in a single region of the cell. In vitro the method allows to visualize the structural organization of Arp2/3-branched filaments (Rouiller et al., 2008). Further applying the method to simpler in vitro reconstituted

10

From Molecules to Movement

251

motile actin meshworks in which structural parameters can be controlled by adjusting at will the chemical composition of the medium would be extremely useful in understanding the in vivo context. Similarly, live cell imaging of the dynamics of cytoskeletal arrays like fluorescence speckle microscopy often fails to provide conclusive interpretations of the processes underlying the birth and movement of speckles. Contributions from treadmilling, adhesion and contractility combine in the overall movement of a patch of molecules that exhibit coherent motion. The simpler composition of motile dendritic arrays offers a unique opportunity to characterize the dynamic parameters of a pure branched actin array, and quantitatively measure the effects of regulators such as capping proteins, Arp2/3 complex, or tropomyosin.

10.3.3 Toward Coordinated Actin Dynamics in Cell Movement The reconstitution of movement by isolated simple modules is a first step toward reconstitution of an ensemble of actin arrays undergoing coordinated movement in a confined environment, like in cell migration. This objective may not be achieved very soon, it is clearly a challenging experiment at several levels. What is the physical-chemical basis for the coordinated turnover of diverse actin arrays organizing protrusion, adhesion and contractility in the same cell? What extent of diversity in motile behaviors can be generated by combining them in different ratios? How is polarity of cell migration established and maintained? How is the crosstalk between actin filaments and microtubules orchestrated, in more complex self-organized processes like mitosis, assymetric division, cytokinesis? Building up reconstituted systems of increasing complexity by adding to simple modules individual bricks whose function is well understood is anticipated to provide answers to these burning issues.

References Bernardini ML, Mounier J, d’Hauteville H, Coquis-Rondon M, Sansonetti PJ. (1989) Identification of icsA, a plasmid locus of Shigella flexneri that governs bacterial intra- and intercellular spread through interaction with F-actin. Proc Natl Acad Sci USA. 86, 3867–71. Block J, Stradal TE, Hänisch J, Geffers R, Köstler SA, Urban E, Small JV, Rottner K, Faix J. (2008) Filopodia formation induced by active mDia2/Drf3. J Microsc. 231, 506–17. Bosch M, Le KH, Bugyi B, Correia JJ, Renault L, Carlier MF. (2007) Analysis of the function of Spire in actin assembly and its synergy with formin and profilin. Mol Cell. 28, 555–68. Bugyi B and Carlier MF. (2010) Control of actin filament treadmilling in cell motility. Annu Rev Biophys. 39, 449–70. Bugyi B, Didry D, Carlier MF. (2010) How tropomyosin regulates lamellipodial actin-based motility: a combined biochemical and reconstituted motility approach. EMBO J. 29, 14–26. Burroughs NJ, Marenduzzo D. (2007) Nonequilibrium-driven motion in actin networks: comet tails and moving beads. Phys Rev Lett. 98, 238302. Carlier MF, Laurent V, Santolini J, Melki R, Didry D, Xia GX, Hong Y, Chua NH, Pantaloni D. (1997). ADF /cofilin enhances the rate of filament turnover: implication in actin-based motility. J Cell Biol. 136, 1307–22.

252

M.-F. Carlier and D. Pantaloni

Carlier MF, Pantaloni D. (1997) Control of actin dynamics in cell motility. J Mol Biol. 269, 459–67. Carlsson AE. (2003) Growth velocities of branched actin networks. Biophys J. 84, 2907–18. Cameron LA, Footer MJ, van Oudenaarden A, Theriot JA. (1999) Motility of ActA protein-coated microspheres driven by actin polymerization. Proc Natl Acad Sci USA. 96, 4908–13. Cameron LA, Giardini PA, Soo FS, Theriot JA. (2000) Secrets of actin-based motility revealed by a bacterial pathogen. Nat Rev Mol Cell Biol. 1, 110–9. Review. Cossart P. (2000) Actin-based motility of pathogens: the Arp2/3 complex is a central player. Cell Microbiol. 2, 195–205. Review. Dahlgaard K, Raposo AA, Niccoli T, St Johnston D. (2007) Capu and Spire assemble a cytoplasmic actin mesh that maintains microtubule organization in the Drosophila oocyte. Dev Cell. 13, 539–53. Dayel MJ, Akin O, Landeryou M, Risca V, Mogilner A, Mullins RD. (2009) In silico reconstitution of actin-based symmetry breaking and motility. PLoS Biol. 7, e1000201 Delatour V, Helfer E, Didry D, Lê KH, Gaucher JF, Carlier MF, Romet-Lemonne G. (2008) Arp2/3 controls the motile behavior of N-WASP-functionalized GUVs and modulates N-WASP surface distribution by mediating transient links with actin filaments. Biophys J. 94, 4890–905. Didry D, Carlier MF, Pantaloni D. (1998) Synergy between actin depolymerizing factor and profilin in increasing filament turnover. J Biol Chem. 273, 25602–25611. Ditlev JA, Vacanti NM, Novak IL, Loew LM. (2009) An open model of actin dendritic nucleation. Biophys J. 96, 3529–42. Dodding MP, Way M. (2009) Nck- and N-WASP-dependent actin-based motility is conserved in divergent vertebrate poxviruses. Cell Host Microbe. 2009 Dec 17;6(6), 536–50. Egile C, Loisel TP, Laurent V, Li R, Pantaloni D, Sansonetti PJ, Carlier MF. (1999) Activation of the CDC42 effector N-WASP by the Shigella flexneri IcsA protein promotes actin nucleation by Arp2/3 complex and bacterial actin-based motility. J CellBiol. 146, 1319–32. Frischknecht F, Moreau V, Röttger S, Gonfloni S, Reckmann I, Superti-Furga G, Way M. (1999) Actin-based motility of vaccinia virus mimics receptor tyrosine kinase signalling. Nature. 401, 926–9. Gerbal F, Laurent V, Ott A, Carlier MF, Chaikin P, Prost J. (2000) Measurement of the elasticity of the actin tail of Listeria monocytogenes. Eur Biophys J. 29, 134–40. Giardini PA, Fletcher DA, Theriot JA. (2003) Compression forces generated by actin comet tails on lipid vesicles. Proc Natl Acad Sci USA. 100, 6493–8. Goode BL, Eck MJ. (2007) Mechanism and function of formins in the control of actin assembly. Annu Rev Biochem. 76, 593–627. Heinzen RA, Hayes SF, Peacock MG, Hackstadt T. (1993) Directional actin polymerization associated with spotted fever group Rickettsia infection of Vero cells. Infect Immun. 61, 1926–35. Heuvingh J, Franco M, Chavrier P, Sykes C. (2007) ARF1-mediated actin polymerization produces movement of artificial vesicles. Proc Natl Acad Sci USA. 104, 16928–33. Iwasa JH, Mullins RD. (2007) Spatial and temporal relationships between actin-filament nucleation, capping, and disassembly. Curr Biol. 17, 395–406. Jeng RL, Goley ED, D’Alessio JA, Chaga OY, Svitkina TM, Borisy GG, Heinzen RA, Welch MD. (2004) A Rickettsia WASP-like protein activates the Arp2/3 complex and mediates actin-based motility. Cell Microbiol. 6, 761–9. Kocks C, Gouin E, Tabouret M, Berche P, Ohayon H, Cossart P. (1992) L. monocytogenes-induced actin assembly requires the actA gene product, a surface protein. Cell. 68, 521–31. Lai F, Szczodrak M, Block J, Faix J, Breitsprecher D, Mannherz HG, Stradal TE, Dunn GA, Small JV, Rottner K (2008) Arp2/3 complex interactions and actin network turnover in lamellipodia. EMBO J. 27, 982–892. Laurent V, Loisel TP, Harbeck B, Wehman A, Gröbe L, Jockusch BM, Wehland J, Gertler FB, Carlier MF. (1999) Role of proteins of the Ena/VASP family in actin-based motility of Listeria monocytogenes. J Cell Biol. 144, 1245–58.

10

From Molecules to Movement

253

Le Clainche C, Carlier MF. (2008) Regulation of actin assembly associated with protrusion and adhesion in cell migration. Physiol Rev. 88, 489–513. Review. Loisel TP, Boujemaa R, Pantaloni D, Carlier MF. (1999) Reconstitution of actin-based motility of Listeria and Shigella using pure proteins. Nature. 401, 613–6. Machesky LM, Cooper JA. (1999) Cell motility. Bare bones of the cytoskeleton. Nature. 401, 542–3. Manseau LJ, Schüpbach T. (1989) cappuccino and spire: two unique maternal-effect loci required for both the anteroposterior and dorsoventral patterns of the Drosophila embryo. Genes Dev. 3, 1437–52. Marcy Y, Prost J, Carlier MF, Sykes C. (2004) Forces generated during actin-based propulsion: a direct measurement by micromanipulation. Proc Natl Acad Sci USA. 101, 5992–7. Medalia O, Weber I, Frangakis AS, Nicastro D, Gerisch G, Baumeister W. (2002) Macromolecular architecture in eukaryotic cells visualized by cryoelectron tomography. Science. 298, 1209–13. Medalia O, Beck M, Ecke M, Weber I, Neujahr R, Baumeister W, Gerisch G. (2007) Organization of actin networks in intact filopodia. Curr Biol.17, 79–84. Mogilner A, Oster G. (2003) Force generation by actin polymerization II: the elastic ratchet and tethered filaments. Biophys J. 84, 1591–605. Mogilner A, Rubinstein B. (2005) The physics of filopodial protrusion. Biophys J. 89, 782–95. Mogilner A. (2006) On the edge: modeling protrusion. Curr Opin Cell Biol. 18, 32–9. Mejillano MR, Kojima S, Applewhite DA, Gertler FB, Svitkina TM, Borisy GG. (2004) Lamellipodial versus filopodial mode of the actin nanomachinery: pivotal role of the filament barbed end. Cell. 118, 363–73. Pantaloni D, Carlier MF (1993) How profilin promotes actin filament assembly in the presence of thymosin ß4. Cell 75, 1007–1014. Pantaloni D, Le Clainche C, Carlier MF. (2001) Mechanism of actin-based motility. Science. 292, 1502–6. Review. Parekh SH, Chaudhuri O, Theriot JA, Fletcher DA. (2005) Loading history determines the velocity of actin-network growth. Nat Cell Biol. 7, 1219–23. Pollard TD, Cooper JA (1984) Quantitative analysis of the effect of Acanthamoeba profilin on actin filament nucleation and elongation. Biochemistry 23, 6631–664. Pollard TD. (2007) Regulation of actin filament assembly by Arp2/3 complex and formins. Annu Rev Biophys Biomol Struct. 36, 451–77. Pollard TD, Cooper JA. (2009) Actin, a central player in cell shape and movement.Science. 326, 1208–12. Romero S, Le Clainche C, Didry D, Egile C, Pantaloni D, Carlier MF. (2004) Formin is a processive motor that requires profilin to accelerate actin assembly and associated ATP hydrolysis. Cell. 119, 419–29. Rouiller I, Xu XP, Amann KJ, Egile C, Nickell S, Nicastro D, Li R, Pollard TD, Volkmann N, Hanein D. (2008) The structural basis of actin filament branching by the Arp2/3 complex. J Cell Biol. 180, 887–95. Schmidt S, van der Gucht J, Biesheuvel PM, Weinkamer R, Helfer E, Fery A. (2008) Non-Gaussian curvature distribution of actin-propelled biomimetic colloid trajectories. Eur Biophys J. 37, 1361–6. Schwille P, Diez S. (2009) Synthetic biology of minimal systems. Crit Rev Biochem Mol Biol. 44, 223–242. Shaevitz JW, Fletcher DA. (2007) Load fluctuations drive actin network growth. Proc Natl Acad Sci USA. 104, 15688–92. Small JV (1995) getting the actin filaments straight: nucleation-release or treadmilling? Trends in Cell Biol. 5, 52–55. Small JV Stradal T, Vignal E, Small JV (2002). The lamellipodium: where motility begins. Trends Cell Biol. 12, 112–120. Taunton J. (2001) Actin filament nucleation by endosomes, lysosomes and secretory vesicles. Curr Opin Cell Biol.13, 85–91. Review.

254

M.-F. Carlier and D. Pantaloni

Theriot JA, Mitchison TJ, Tilney LG, Portnoy DA. (1992) The rate of actin-based motility of intracellular Listeria monocytogenes equals the rate of actin polymerization. Nature. 357, 257–60. Theriot JA, Rosenblatt J, Portnoy DA, Goldschmidt-Clermont PJ, Mitchison TJ. (1994) Involvement of profilin in the actin-based motility of L. monocytogenes in cells and in cell-free extracts. Cell. 76, 505–17. Tilney LG, Portnoy DA. (1989) Actin filaments and the growth, movement, and spread of the intracellular bacterial parasite, Listeria monocytogenes. J Cell Biol. 1989 Oct;109(4 Pt 1), 1597–608. Upadhyaya A, van Oudenaarden A. (2003) Biomimetic systems for studying actin-based motility. Curr Biol. 13, R734–44. Review. Upadhyaya A, Chabot JR, Andreeva A, Samadani A, van Oudenaarden A. (2003) Probing polymerization forces by using actin-propelled lipid vesicles. Proc Natl Acad Sci USA. 100, 4521–6. van Oudenaarden A, Theriot JA. (1999) Cooperative symmetry-breaking by actin polymerization in a model for cell motility. Nat Cell Biol. 1, 493–9. Wang YL (1985) Excahnge of actin subunits at the leading edge of living fibroblasts: possible role of treadmilling. J. Cell Biol. 101, 597–602. Wegner A. (1976) Head to tail polymerization of actin. J Mol Biol. 108, 139–150. Wegner A. (1982) Treadmilling of actin at physiological salt concentrations. An analysis of the critical concentrations of actin filaments. J. Mol. Biol. 161, 607–615. Weisswange I, Newsome TP, Schleich S, Way M. (2009) The rate of N-WASP exchange limits the extent of ARP2/3-complex-dependent actin-based motility. Nature.458, 87–91. Welch MD, Iwamatsu A, Mitchison TJ. (1997) Actin polymerization is induced by Arp2/3 protein complex at the surface of Listeria monocytogenes. Nature. 385, 265–9. Welch MD, Rosenblatt J, Skoble J, Portnoy DA, Mitchison TJ. (1998) Interaction of human Arp2/3 complex and the Listeria monocytogenes ActA protein in actin filament nucleation. Science. 281, 105–8. Wiesner S, Helfer E, Didry D, Ducouret G, Lafuma F, Carlier MF, Pantaloni D. (2003) A biomimetic motility assay provides insight into the mechanism of actin-based motility. J Cell Biol. 160, 387–98.

Chapter 11

The WASP-Homology 2 Domain and Cytoskeleton Assembly Roberto Dominguez

Abstract One of the most abundant and functionally diverse actin-binding folds is the WASP homology 2 (WH2 or W) domain. The W domain is found in proteins involved in actin monomer sequestration and cytoskeleton scaffolding, but is particularly abundant among proteins that mediate the de novo formation of actin filaments, which includes actin filament nucleation and elongation factors. Known filament nucleators include the Arp2/3 complex and its large family of Nucleation Promoting Factors (NPFs), formins, Spire, Cobl, VopL/VopF, TARP and Lmod. These molecules are generally unrelated, but with the exception of formins they all use the W domain for interaction with actin. A common architecture, found in Spire, Cobl and VopL/VopF, consists of tandem W domains that bind three to four actin subunits to form a nucleus. Structural considerations suggest that NPFs-Arp2/3 complex can also be viewed as a specialized form of tandem W-based nucleator. Ena/VASP proteins are distantly related to WASP-family NPFs, and function as dedicated filament elongation factors among W-based nucleators.

Contents Introduction . . . . . . . . . . . . . . . . . . . . . . Filament Nucleation and the W Domain . . . . . . . . . Tandem W Domain-Based Filament Nucleators . . . . . . Oligomerization of W-Based Nucleators . . . . . . . . . Role of the W Domain in Filament Nucleation by NPFs-Arp2/3 Complex . . . . . . . . . . . . . . . . 11.6 The Arp2/3 Complex and NPFs as a Specialized Form of Tandem W Nucleator . . . . . . . . . . . . . . . . . 11.7 Leiomodin (Lmod) and the Nucleation of Actin Filaments in Muscle Cells . . . . . . . . . . . . . . . . . . . .

11.1 11.2 11.3 11.4 11.5

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

256 256 258 260

. . . . . . . . . .

261

. . . . . . . . . .

265

. . . . . . . . . .

266

R. Dominguez (B) Department of Physiology, University of Pennsylvania School of Medicine, 3700 Hamilton Walk, Philadelphia, PA 19104-6085, USA e-mail: [email protected]

M.-F. Carlier (ed.), Actin-based Motility, DOI 10.1007/978-90-481-9301-1_11, C Springer Science+Business Media B.V. 2010

255

256

R. Dominguez

11.8 The W Domain and Filament Elongation . . . . . . . . . . . . . . . . . . . 11.9 Other Functions of the W Domain . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

268 270 272

11.1 Introduction Many cellular functions, including cell motility, endocytosis and intracellular trafficking, involve dynamic remodeling of actin filament networks under the control of a myriad of actin-binding proteins (ABPs), along with signaling and scaffolding proteins. The functions of ABPs are diverse, but their interactions with actin involve a relatively limited number of folding motifs. For instance, the gelsolin domain, implicated in actin filament severing and barbed end capping (Silacci et al., 2004), and the ADF/cofilin domain, implicated in filament pointed end depolymerization (Bamburg, 1999), are structurally related and interact with actin in a similar manner (McLaughlin et al., 1993; Paavilainen et al., 2008). Another example is the calponin-homology (CH) domain, which is found in actin crosslinking proteins such as α-actinin, filamin and fimbrin, and in proteins that link the actin cytoskeleton to other cellular structures or filament networks (known as cytolinkers), such as dystrophin, utrophin and plectin (Gimona et al., 2002). But perhaps the most abundant and functionally diverse actin-binding fold is the WASP homology 2 (WH2 or W) domain. The W domain is found in proteins involved in actin monomer sequestration and cytoskeleton scaffolding, but has become particularly prominent among proteins that mediate the de novo formation of actin filaments, including actin filament nucleation and elongation factors (Dominguez, 2009).

11.2 Filament Nucleation and the W Domain In eukaryotic cells the concentration of actin monomers is much higher than the critical concentrations for monomer addition at the barbed (0.1 μM) and pointed (0.7 μM) ends of the actin filament (Pollard and Borisy, 2003). Under such conditions, most of the cellular actin would be expected to form filaments. However, cells use diverse mechanisms to control actin polymerization, maintaining a large fraction (∼50%) of the actin in the monomeric state. One such mechanism is the control of the actin monomer pool available for polymerization through actin monomer sequestering proteins such as thymosin-β4 (Tβ4). Another mechanism is the control of the nucleation step of actin polymerization, which is kinetically unfavorable and thereby rate limiting (Sept and McCammon, 2001). Thus, cells use filament nucleators to stabilize actin polymerization nuclei (or seeds). Known filament nucleators include the Arp2/3 complex and its large family of Nucleation Promoting Factors (NPFs), formins, Spire, Cobl, VopL/VopF, TARP and Lmod. Filament nucleators control the time and location for actin polymerization, and additionally influence the structures of the actin networks that they generate. The actin filament can be described as either a single left-handed short-pitch helix, where consecutive lateral subunits are staggered with respect to one another

11

The WASP-Homology 2 Domain and Cytoskeleton Assembly

257

by half a monomer length, or two right-handed long-pitch helices of head-to-tail bound actin subunits (Holmes, 2009; Holmes et al., 1990; Oda et al., 2009). Filament nucleators are generally unrelated, have markedly different nucleation activities and work by different mechanisms, stabilizing small actin oligomers (dimers, trimers and tetramers) along either the long- or the short-pitch helices of the actin filament. With the exception of formins, known filament nucleators use the W domain for interaction with actin, although in Lmod and the Arp2/3 complex other domains contribute as well. The W domain was first identified as a conserved motif in several WASP-family proteins that had been implicated in the regulation of actin cytoskeleton dynamics (Symons et al., 1996). It was noted that in these proteins the W domain occurred C-terminal to profilin-binding Pro-rich sequences. Subsequently, it was shown that the W domain had actin monomer binding activity (Miki and Takenawa, 1998), and that when combined with the acidic region at the C-terminus of WASP-family proteins, it induced actin assembly through the Arp2/3 complex (Machesky and Insall, 1998). The W domain has a short length (17–27aa) and is generally poorly conserved, making it difficult to identify based on sequence analysis alone (Dominguez, 2007). Lappalainen and colleagues were the first to recognize the existing relationship between the W domain and the thymosin fold involved in actin filament sequestration (Paunola et al., 2002). The N-terminal portion of the W domain forms a helix that binds in the hydrophobic cleft, or target-binding cleft (Dominguez, 2004), between subdomains

Fig. 11.1 Structure of the W domain of missing-in-metastasis (MIM, red) bound to actin (blue). Actin subdomains 1–4 are labeled. The W domain consists of an N-terminal amphiphilic helix that binds in the hydrophobic (or target-binding) cleft (Dominguez, 2004) between subdomains 1 and 3 of actin, followed by a C-terminal extended region featuring the conserved LKKT(V) motif (yellow carbon atoms). Hydrophobic amino acids of the N-terminal helix are colored greed. This and other W-actin structures were determined as ternary complexes with DNase I (not shown), which prevents actin polymerization (Chereau et al., 2005)

258

R. Dominguez

1 and 3 at the barbed end of the actin monomer (Chereau et al., 2005; Hertzog et al., 2004) (Fig. 11.1). After this helix, the W domain presents an extended region that climbs toward the pointed end of the actin monomer. This region has variable length and sequence but comprises the conserved four residue motif LKKT(V). The N-terminal helix and LKKT(V) motif constitute the conserved core of the W domain, but as discussed below the W domain displays remarkable functional and sequence diversity.

11.3 Tandem W Domain-Based Filament Nucleators The W domain often occurs in tandem repeats (Fig. 11.2), which is the most common architecture among known actin filament nucleators, observed in Spire (Quinlan et al., 2005), Cobl (Ahuja et al., 2007), and VopL/VopF (Liverman et al., 2007; Tam et al., 2007). The protein TARP (translocated actin recruiting phosphoprotein) from Chlamydia trachomatis contains a single W domain, but forms large oligomers (Jewett et al., 2006). The actin monomers bound to the W domains of these proteins are thought to come together to form an actin filament-like nucleus to initiate polymer assembly. However, the specific nucleation mechanism of each protein appears to be different (Fig. 11.2c), as reflected by dramatic differences in their nucleation activities. For instance, Spire with the largest number of W domains (four) has relatively weak nucleation activity (Quinlan et al., 2005), whereas Cobl and VopL/VopF with just three W domains are strong nucleators. At least in part, the explanation may lie in the variable linkers between W domains, in particular linker-2 between the second and third W domains. Differences in the linkers may dictate the relative arrangement of actin subunits in the polymerization nucleus, and thereby the nucleation activities of each protein. When the linkers are short, as in Spire, only actin subunits along the long-pitch helix of the actin filament can be connected (Rebowski et al., 2008). However, the brain-specific nucleator Cobl has strong nucleation activity and presents a long, Pro-rich linker-2 (Ahuja et al., 2007). Shortening Cobl’s linker-2 reduces dramatically its nucleation activity, whereas replacing this linker with an unrelated sequence of similar length restores most of the endogenous activity. Therefore, the length of the linker, but not necessarily its specific sequence, appears to be crucial for Cobl’s activity. Because a longer linker may allow successive W domains to connect actin subunits laterally, it has been proposed that Cobl stabilizes a trimeric short-pitch actin nucleus (Ahuja et al., 2007) (Fig. 11.2c). However, the exact arrangement of actin subunits in Cobl’s nucleus is unknown and two possibilities must be considered, with the actin subunit bound to either the first or the second W domain staggered forward at the pointed end. In any case, the example of Cobl, as well as the Arp2/3 complex and Lmod (discussed below), suggest that stabilization of a short-pitch actin trimer is a more effective way to promote nucleation than stabilization of a larger nucleus of four actin subunits along the long-pitch helix of the actin filament. A recent study, additionally suggests that some inter-W linkers present actin monomer-binding activity, and can as a result boost the nucleation activity of tandem W constructs (Zuchero et al., 2009). Thus, for example, a fragment consisting of the two W domains of N-WASP

11

The WASP-Homology 2 Domain and Cytoskeleton Assembly

259

Fig. 11.2 Tandem W domain filament nucleators and their relationship with NPFs and Ena/VASP (a) Tandem W domain-based filament nucleators, NPFs and Ena/VASP are all multidomain (modular) proteins, sharing a number of common features. The W domains and the C region of NPFs are both colored red to highlight their relationship. Pro-rich regions (magenta) abound among these proteins, and bind regulatory proteins and profilin-actin, the main source of polymerization competent actin in cells. Coiled coils and other oligomerization domains (both known and predicted from sequence analysis) are also common (green). In Ena/VASP, the WASP-Homology domain 1 (WH1), W and C regions are respectively known as EVH1 (Ena/VASP homology 1), GAB and FAB (G- and F-actin binding) domains, but despite their different names these domains are related to their N-WASP counterparts. Ena/VASP also has an acidic region after the C motif, albeit less acidic than in NPFs and lacking the key tryptophan residue. An ubiquitin-like segment in Cobl (a potential Ras-binding site), a snare-like helix in WAVE2 (a potential multi-protein association site), a spectrin-like antiparallel dimerization domain in JMY and a dimerization domain in VopL/VopF are all predicted from sequence analysis. The third W domain of Spire is noncanonical (red/magenta); it contains various Pro residues. Interestingly, this W domain occupies the position of the Pro-rich linker-2 of Cobl and VopL/VopF. (c) Sequence alignment of the W and CA regions of the proteins shown in part A. Conserved residues are colored according to their chemical properties. Uniprot codes: Drosophila Spire, Q9U1K1; mouse Cobl, Q5NBX1; Vibrio parahaemolyticus VopL, Q87GE5; mouse JMY, Q9QXM1; mouse N-WASP, Q91YD9; human WAVE2, Q9Y6W5; human VASP, P50552; human Evl, Q9UI08. (b) Proposed nucleation mechanisms of Spire (Quinlan et al., 2005) and Cobl (Ahuja et al., 2007). The lengths of the linkers determine whether neighboring W domains bridge actin subunits along a single strand (Spire) or across strands (Cobl) of the actin filament (i.e. long- vs short-pitch nuclei), which may in turn determine the nucleation activities of these proteins.\VopL has three W domains and, like Cobl, it is a strong nucleator, but linker-2 of VopL (and VopF) is too short to allow for the stabilization of a short-pitch nucleus

260

R. Dominguez

had no nucleation activity, but a modest increase in nucleation was observed when the naturally occurring inter-W linker was replaced by Spire’s linker-3 (Zuchero et al., 2009). Microbial pathogens often disrupt or hijack the host cell cytoskeleton for infection (Bhavsar et al., 2007; Gouin et al., 2005). A well known example is Listeria monocytogenes, whose surface protein ActA mimics eukaryotic NPFs and recruits both the filament elongation factor VASP and the Arp2/3 complex polymerization machineries at the surface of the parasite to propel its movement within and between cells (Cossart and Toledo-Arana, 2008). Vibrios are Gram-negative rod-shaped bacteria, comprising human pathogens that cause wound infections, gastro-intestinal disease and diarrhea, and are often associated with infection from consumption of raw seafood. Vibrio parahaemolyticus and Vibrio cholerae were nearly simultaneously shown to produce the type III secretion system (T3SS) virulence factors VopL (Liverman et al., 2007), and VopF (Tam et al., 2007), respectively. VopL and VopF display ∼57% overall sequence identity. Similar proteins are also found among other Vibrio species. VopL/VopF disrupt actin homeostasis, and appear to be required for infection (Liverman et al., 2007; Tam et al., 2007). Both proteins present three W domains and Pro-rich sequences, and like Cobl, both are strong filaments nucleators. It is, therefore, tempting to propose that like Cobl these two proteins stabilize a short-pitch polymerization nucleus. However, linker-2 in VopL/VopF is significantly shorter than in Cobl (Fig. 11.2b), and because the length of the linker is such a critical factor for Cobl’s activity (Ahuja et al., 2007), the reasons for the strong nucleation activities of VopL/VopF remain a mystery. A potential explanation is given next.

11.4 Oligomerization of W-Based Nucleators In addition to the inter-W linkers, oligomerization may influence the nucleation activities of tandem W-based filament nucleators. For instance, Spire interacts with the formin Cappuccino (Quinlan et al., 2007; Quinlan and Kerkhoff, 2008; Renault et al., 2008; Rosales-Nieves et al., 2006), and the two proteins appear to synergize to assemble actin filaments both in vitro (Bosch et al., 2007) and in vivo (Rosales-Nieves et al., 2006), where they may be involved in maintaining microtubule organization (Dahlgaard et al., 2007). The interaction, which involves the kinase non-catalytic C-lobe domain (KIND) of Spire (Fig. 11.2a) and the formin homology 2 (FH2) domain of Cappuccino, enhances the nucleation activity of Spire (Quinlan et al., 2007). It is likely that the increased activity results from Spire dimerization mediated by the FH2 dimer. Another possibility for Spire to function as a dimer is through its C-terminal FYVE zinc-finger domain (Fig. 11.2a), which in some proteins has been shown to dimerize (Dumas et al., 2001). A recent report additionally finds that WASH, a WASP-family NPF (Linardopoulou et al., 2007), interacts directly with Spire and synergizes with both Spire and Cappuccino to control actin and microtubule dynamics during Drosophila oogenesis (Liu et al., 2009). WASH also appears to dimerize through its N-terminal WASH homology domain 1 (WHD1) (Liu et al., 2009), providing yet another potential

11

The WASP-Homology 2 Domain and Cytoskeleton Assembly

261

mechanism for Spire dimerization. Whether Spire dimerizes directly through its FYVE zinc-finger domain or indirectly through interaction with Cappuccino or WASH, dimerization is likely a contributing factor in Spire’s nucleation activity in cells. Indeed, an optimally assembled Spire dimer could stabilize the formation of a nucleus consisting of eight actin subunits, four on each side of the filament (or long-pitch filament helix), potentially resulting in a very powerful nucleator. Another example is the T3SS protein TARP from Chlamydia trachomatis. Despite having a single W domain, TARP nucleates actin filaments, but this activity depends on the presence of the central Pro-rich domain, which in this protein appears to mediate oligomerization (Jewett et al., 2006). The existing relationship between NPFs-Arp2/3 complex and tandem W-based nucleators is discussed below. In this regard, it is interesting to note that a recent study finds that the dimerization of WASP by external factors increases its affinity for the Arp2/3 complex and enhances its nucleation activity (Padrick et al., 2008). It is still unknown whether WASP dimerization plays a role in vivo. However, the WASP-related protein WASH appears to dimerize by itself. Although WASH dimerization does not seem to enhance Arp2/3 complex-mediated nucleation (as proposed for WASP), it is likely to play a critical role in vivo, notably by mediating the bundling and crosslinking of actin filaments and microtubules under the control of the GTPase Rho1 (Liu et al., 2009). Finally, sequence analysis identifies potential oligomerization domains among other nucleators and NPFs (Fig. 11.2a). Oligomerization is thus emerging as an important factor modulating the activities of W-based nucleators, which is analogous to formins (Copeland et al., 2004). Cobl, as well as VopL/VopF, contain predicted coiled-coil dimerization regions, but whether oligomerization contributes to their nucleation activities remains to be demonstrated.

11.5 Role of the W Domain in Filament Nucleation by NPFs-Arp2/3 Complex The W domain also participates in filament nucleation through the NPF family of Arp2/3 complex activators, which can have between one and three W domains. The Arp2/3 complex was first purified ∼15 years ago (Machesky et al., 1994). The complex consists of seven proteins, including two actin-related proteins, Arp2 and Arp3, and subunits ARPC1 to 5 (Fig. 11.3a). By itself, the Arp2/3 complex has very low nucleation activity (Mullins et al., 1998). The discovery of ActA as an NPF at the surface of Listeria monocytogenes first revealed the strong nucleation capacity of NPFs-Arp2/3 complex (Welch et al., 1998). Around the same time, eukaryotic NPFs belonging to the WASP/WAVE-family of proteins were identified (Machesky and Insall, 1998; Machesky et al., 1999; Rohatgi et al., 1999; Winter et al., 1999; Yarar et al., 1999). NPFs are themselves regulated by multiple factors, in particular Rhofamily GTPases. Thus, WASP and N-WASP function under the control of Cdc42, whereas WAVE forms part of a large complex that is regulated by Rac (Bompard

262

R. Dominguez

Fig. 11.3 Actin filament nucleation by the Arp2/3 complex and NPFs. (a) The Arp2/3 complex consists of seven proteins, including the actin related proteins Arp2 and Arp3 and subunits ARPC1–5 (labeled 1–5). NPFs are large multi-domain proteins. The figure illustrates a prototypical NPF protein, characterized by the presence of regulatory/localization domains, a Pro-rich region, and a WCA region that can have between one and three W domains. Two of the W domains are colored red/gray (striped) to indicate that their presence is not absolutely necessary for nucleation. Combined, the Arp2/3 complex and NPFs have strong nucleation activity. WCA is the smallest fragment capable of catalyzing the formation of a polymerization nucleus, consisting of the two Arps and one to three actin subunits, as well as a conformational change in Arp2/3 complex that stimulates monomer addition to the branch filament and binding of the nucleus to the side of a preexisting filament (mother filament). The branch grows from the barbed ends of the Arps at a 70◦ angle with respect to the mother filament. (b) Structure of inactive Arp2/3 complex (Nolen and Pollard, 2007; Robinson et al., 2001). Subdomains 1 and 2 of Arp2 are disordered in the structures, but were added here by analogy with actin. The Arp2/3 complex subunits are colored according to the diagram of part A. (c) SAXS-derived model of WCA-actin-Arp2/3 complex (Boczkowska et al., 2008). The orientation is the same as in part B. The mother and branch filaments are shown for reference, although this work did not address branch assembly. Note that Arp2 moves up

11

The WASP-Homology 2 Domain and Cytoskeleton Assembly

263

and Caron, 2004; Eden et al., 2002; Goley and Welch, 2006; Hall, 2005; Kim et al., 2000; Ma et al., 1998). Classical NPFs, including WASP/N-WASP/WAVE (Goley and Welch, 2006; Pollard, 2007), WASH (Linardopoulou et al., 2007), WHAMM (Campellone et al., 2008) and JMY (Zuchero et al., 2009), present a C-terminal WCA region, which constitutes the shortest polypeptide necessary for activation of nucleation with the Arp2/3 complex (Machesky et al., 1999). This region consists of three distinct segments: W, C and A (Fig. 11.3a). The W domain (of which there can be between one and three) recruits the first actin subunit(s) of the new filament branch. The C (central or connecting) and A (acidic) motifs interact with various subunits of the Arp2/3 complex, helping to stabilize the activated conformation. However, the mechanism by which CA participates in Arp2/3 complex activation, and the specific interactions with subunits of the complex remain a mystery. The actin monomer(s) bound to the W domain(s), together with Arp2 and Arp3, are thought to form a filament-like seed for the nucleation of a filament branch that emerges at a 70◦ angle from the side of a preexisting filament (Fig. 11.3a). According to this model (Robinson et al., 2001), Arp2 and Arp3 are the first two subunits at the pointed end of the new filament branch, and are expected to adopt a short-pitch filament-like conformation. The crystal structure of Arp2/3 complex was first determined in the absence of nucleotide and NPF (Robinson et al., 2001) (Fig. 11.3b). In the structure, Arp2 and Arp3 are separated (i.e. not in a filament-like conformation) and the nucleotide cleft of Arp3 is wide open, whereas subdomains 1 and 2 of Arp2 are disordered. Thus, this structure was described as the inactive conformation of the complex (Robinson et al., 2001). Subsequently, Arp2/3 complex was crystallized in the presence of ATP or nucleotide analogs (Nolen and Pollard, 2007). Nucleotide binding favors closure of the nucleotide cleft of Arp3 and marginally stabilizes subdomains 1 and 2 of Arp2. However, the relative position of Arp2 and Arp3 was unchanged, indicating that, although necessary, ATP binding alone is insufficient to activate Arp2/3 complex. It is believed that the binding of nucleotide and WCA are thermodynamically coupled and that these two factors contribute together to activating Arp2/3 complex (Dayel et al., 2001; Goley et al., 2004; Le Clainche et al., 2001). Pre-existing filaments may help shift the equilibrium in favor of an activated complex (Pollard,

Fig. 11.3 (continued) compared to its location in part B. This study placed the first actin subunit of the branch at the barbed end of Arp2. The position of the W domain, which is known precisely from the crystal structure of its complex with actin (Fig. 11.1), imposes constraints on the location of the C motif. Thus, the hydrophobic cleft of Arp2 is in the path of the CA polypeptide as it progresses toward the Arp2/3 complex. The helical portion of the C motif may thus bind in this cleft, as supported by sequence similarity with W (see Fig. 11.2b). The position of the A motif (pink) is less well constrained, but this model would be consisting with it binding at the interface between Arp3 and ARPC3, as suggested by biochemical studies (Kelly et al., 2006; KreishmanDeitrick et al., 2005; Pan et al., 2004; Weaver et al., 2002; Zalevsky et al., 2001). (c) Model of filament nucleation by NPFs-Arp2/3 complex

264

R. Dominguez

2007). However, Arp2/3 complex can bind to and cap filament pointed ends with high affinity outside the branch (Mullins et al., 1998). Because pointed end binding requires an activated conformation (Boczkowska et al., 2008; Robinson et al., 2001; Rouiller et al., 2008), binding to the side of preexisting filaments may not be necessary for activation, although it is probably favored by activation. Considering the high concentration of actin monomers in cells and typical affinities of the W-actin interaction ranging from ∼0.05 to ∼0.25μM (Chereau et al., 2005; Marchand et al., 2001; Mattila et al., 2003), it is likely that NPFs are actin-loaded prior to encountering the Arp2/3 complex. Thus, actin-loaded NPFs and nucleotide are probably the most important factors shifting the equilibrium in favor of an activated complex in vivo. The structure of Arp2/3 complex in the branch and with bound WASP has been studied using electron microscopy (Egile et al., 2005; Rodal et al., 2005; Rouiller et al., 2008). These studies agree in that a major conformational change takes place upon activation, bringing Arp2 and Arp3 into a filament-like arrangement at the pointed end of the branch. Additionally, electron tomography of the branch junction reveals conformational changes in the mother filament at the interface with the Arp2/3 complex, and suggests that all seven subunits of the Arp2/3 complex contact the mother filament (Rouiller et al., 2008). None of the existing structures, however, resolves the location and interactions of the CA activator region of NPFs with subunits of the Arp2/3 complex. This question has mainly been addressed by crosslinking and NMR solution studies, showing that CA can be crosslinked to Arp2, ARPC1, Arp3 and ARPC3 (Kelly et al., 2006; Kreishman-Deitrick et al., 2005; Weaver et al., 2002; Zalevsky et al., 2001). Because of the short length of the WCA polypeptide (∼73 aa), and considering that both the C (Panchal et al., 2003) and W (Chereau et al., 2005) domains comprise regions of helical structure, it is difficult to rationalize how CA can span these four subunits in the complex. A recent study attempts a different approach to address this question. Actin has a highly reactive cysteine residue at position 374. The structures of W-actin complexes (Fig. 11.1) revealed that the N-terminal portion of the W domain faces directly actin Cys-374 (Chereau et al., 2005). Based on this observation, a Cys residue was introduced by mutagenesis at the N-terminus of WCA, which was then crosslinked to actin Cys-374 (Boczkowska et al., 2008). [In support of this approach, a crystal structure of crosslinked W-actin is now available and is nearly undistinguishable from the uncrosslinked structures (Rebowski et al., in preparation).] Contrary to WCA alone, crosslinked WCA-actin forms a stable high affinity complex with the Arp2/3 complex, while also capping its barbed end so that the nucleus cannot elongate by addition of actin monomers. The stoichiometry of WCAactin:Arp2/3 complex determined by various methods is precisely 1:1, indicating that there is only one binding site for WCA-actin on the Arp2/3 complex. This approach produced a stable WCA-actin-Arp2/3 complex particle, whose solution structure was studied by Small Angle X-ray Scattering (SAXS). The SAXS study indicated that the first actin subunit binds at the barbed end of Arp2, which additionally constrains the binding site of the C motif to subunit Arp2, near the interface with ARPC1 (Fig. 11.3c). Less can be said about the location and interactions of the

11

The WASP-Homology 2 Domain and Cytoskeleton Assembly

265

A region, except that it probably lies near the interface between subunits Arp3 and ARPC3, which is consistent with most of the biochemical evidence (Kelly et al., 2006; Kreishman-Deitrick et al., 2005; Weaver et al., 2002; Zalevsky et al., 2001). This study offers testable hypotheses and a new way to address the problem of activation, but because of its limited resolution it leaves unresolved the exact nature of the conformational change leading to activation and the precise role of WCA in this process.

11.6 The Arp2/3 Complex and NPFs as a Specialized Form of Tandem W Nucleator Structural considerations suggest that NPFs-Arp2/3 complex can be conceptually viewed as a specialized form of tandem W-based nucleator (Boczkowska et al., 2008). According to this view, the actual nucleators are the NPFs, and not the Arp2/3 complex as it has been traditionally described. The distinction is not merely semantic, but rather stems from a different structure-function understanding of how these proteins work. In isolation, neither the Arp2/3 complex nor the NPFs nucleate; they need each other for this activity. There is only one known exception to this rule, which actually reinforces the proposed relationship between NPFs and tandem W-based nucleators. It is the newly discovered NPF protein JMY, which presents three W domains N-terminal to its CA region and, in addition to activating the Arp2/3 complex, has some nucleation activity of its own (Zuchero et al., 2009). More importantly, there is the undisputable fact that the newly discovered filament nucleators (Spire, Cobl, and VopL/VopF) share far more in common with NPFs than they do with the Arp2/3 complex (Fig. 11.2a), including the presence of tandem W repeats and Pro-rich regions. The number of bona fide W domains in NPFs varies from 1 to 3, whereas the newly discovered nucleators contain between 1 and 4 W domains. However, as it has been pointed out by various investigators (Aguda et al., 2005; Boczkowska et al., 2008; Chereau et al., 2005; Hertzog et al., 2004), the C motif of NPFs is also related to the W domain, a relationship that can be further extended to the F-actin-binding (FAB) motif of Ena/VASP proteins (Ferron et al., 2007) (Fig. 11.2). Based on the location of the first actin subunit in the SAXS structure of WCA-actin-Arp2/3 complex, it was proposed that the C motif binds Arp2 (Boczkowska et al., 2008). Like the W domain, the N-terminal portion of the C motif consists of an amphiphilic helix (Panchal et al., 2003), which according to this proposal binds in the hydrophobic cleft of Arp2 (Fig. 11.3c), somewhat analogous to the binding of W to actin (Fig. 11.1). As for the Arp2/3 complex itself, it can be thought of as an actin dimer that upon activation adopts a short-pitch conformation. The association of the Arps with five other proteins in the Arp2/3 complex probably emerged from a need to integrate nucleation and branching within a single system. Based on these considerations, NPFs can be described as tandem W-based filament nucleators, whose function is to recruit and realign the Arp2-Arp3 short-pitch heterodimer and one to three actin monomers to form a polymerization nucleus.

266

R. Dominguez

Based on the proposed relationship between NPFs and tandem W-based filament nucleators, a model of filament nucleation by NPFs-Arp2/3 complex is proposed in Fig. 11.3d. According to this model, the conserved Trp in the A region of NPFs, which contributes the most to the binding affinity of WCA to the Arp2/3 complex (Marchand et al., 2001; Weaver et al., 2002), works as a “hook”, linking actin-loaded NPFs to the Arp2/3 complex. The SAXS study of WCA-actin-Arp2/3 complex suggests that after this initial encounter the first actin subunit binds at the barbed end of Arp2 (Boczkowska et al., 2008). Arp2, which is partially disordered in the inactive structure (Nolen and Pollard, 2007; Robinson et al., 2001), may transition between active/inactive states, but is stabilized in the activated structure by interaction with the C motif and the first actin subunit of the branch (bound to the W domain of NPFs). This model predicts that Arp2 moves mostly alone during activation, with minimal energetic cost, such as to occupy a filament-like position next to Arp3 (Aguda et al., 2005; Boczkowska et al., 2008). This is supported by flexibility of Arp2 in the inactive structure and the fact that it can be moved with minimal steric clashes. Yet, it is reasonable to expect that, in addition to movement of Arp2, other changes will occur in the complex during activation, including changes at the interface with the mother filament (Rouiller et al., 2008). Activation and branching (i.e. binding to the side of pre-existing filaments) may occur nearly simultaneously. Steric hindrance (discussed below) of the W domain with the actin subunits that begin joining the branch after activation (and possibly of the A motif with the mother filament) may help release NPFs after activation.

11.7 Leiomodin (Lmod) and the Nucleation of Actin Filaments in Muscle Cells The actin “thin” filaments in cardiac and striated muscle sarcomeres display regular length and spacing and are uniformly decorated with muscle-specific proteins such as the troponin complex, tropomyosin (TM) and the barbed and pointed end capping proteins CapZ and Tmod, respectively. Toward the center of sarcomeres, the actin filaments overlap with the myosin “thick” filaments, forming a tight hexagonal lattice. The appearance is that of a rigid structure, and it is not surprising that it has been traditionally thought that the actin filaments in sarcomeres are less dynamic than in non-muscle cells. This view is evolving (Gunst and Zhang, 2008; Sanger and Sanger, 2008; Skwarek-Maruszewska et al., 2009; Wang et al., 2005). The sarcomere may undergo constant dynamic remodeling (or repair), and actin filament nucleators may play a critical role in this process. Leiomodin (Lmod) is a tropomodulin (Tmod)-related protein expressed almost exclusively in muscle cells (Fig. 11.4a). mRNA expression analysis indicates there are three Lmod isoforms (Conley et al., 2001): Lmod1 expressed at low levels in most tissues and at high levels in smooth muscle, Lmod2 expressed exclusively in heart and skeletal muscles and the fetal isoform Lmod3. The first ∼340 amino acids of Lmod are ∼40% identical to Tmod, a pointed end capping protein in muscles (Fischer and Fowler, 2003; Fowler et al., 2003; Kostyukova et al., 2007). In Tmod,

11

The WASP-Homology 2 Domain and Cytoskeleton Assembly

267

Fig. 11.4 Domain organization of Lmod and nucleation mechanism. (a) Domain organization of Lmod compared to Tmod. The first ∼340aa of Lmod are ∼40% identical to Tmod, a pointed end capping protein in muscles. The N-terminal portion of Tmod is unstructured, except for three helical segments involved in tropomyosin (TM) and actin binding. Tmod has a second actin-binding site within its Leu-rich repeat (LRR) domain (Fowler et al., 2003; Krieger et al., 2002). Lmod shares this domain organization, but has only one of the two TM-binding sites. More importantly, Lmod has a ∼150aa C-terminal extension featuring a third actin-binding site, a W domain, as well as a basic patch, which is a predicted NLS. (b) Proposed nucleation mechanism; Lmod stabilizes a trimeric actin seed that grows from the barbed end. Importantly, Lmod’s nucleation activity and cellular localization are both modulated by interaction with TM (Chereau et al., 2008)

the N-terminal portion is unstructured, except for three helical segments involved in binding TM and actin. Tmod has a second actin-binding site within the C-terminal Leu-rich repeat (LRR) domain (Fowler et al., 2003; Krieger et al., 2002). Lmod shares this domain organization, except for one important difference; only one of the two TM-binding sites of Tmod appears to be conserved in Lmod. More importantly, Lmod has a ∼150 amino acid C-terminal extension featuring a third actin-binding site in the form of a W domain. With three actin-binding sites, Lmod could hypothetically recruit three actin monomers to form a trimeric polymerization nucleus, which led to the identification of Lmod as a potential filament nucleator (Chereau et al., 2008). Consistent with this idea, initial characterization of Lmod revealed a powerful nucleator, whose over- or down-expression had dramatic effects on sarcomeric structure and organization (Chereau et al., 2008). Compared to other nucleators, Lmod has one distinctive and important property, it interacts directly with TM. TM is a coiled coil dimer that associates head-to-tail to form long helical strands that wind symmetrically along the two long-pitch helices of the actin filament (Holmes and Lehman, 2008). At the pointed end of the actin filament in muscle sarcomeres TM interacts with Tmod via two helical segments

268

R. Dominguez

located within the N-terminal flexible domain of Tmod (Kostyukova et al., 2007). As mentioned above, only one of these helices is conserved in Lmod. Yet, TM not only modulates the nucleation activity of Lmod, but more importantly it appears to determine Lmod’s localization to filament pointed ends. Thus, Lmod162−495 , lacking the N-terminal flexible domain, retains ∼1/3 of the nucleation activity of full-length Lmod in vitro, but displays nuclear localization. A basic patch located within the long (and probably flexible) linker connecting the second and third actinbinding sites of Lmod is a predicted Nuclear Localization Signal (NLS) and may be responsible for the nuclear localization of Lmod162−495 . While it is unknown whether trafficking through the nucleus forms part of Lmod’s endogenous function, such an activity has been reported for Tmod (Kong and Kedes, 2004). Perhaps reflecting its uniqueness as a muscle-specific cell nucleator, Lmod shares little resemblance with other filament nucleators. With the presence of three actinbinding sites, Lmod is predicted to stabilize a trimeric short-pitch actin seed for nucleation (Fig. 11.4b). However, the actual organization of actin subunits in the Lmod nucleus is a mystery. The W domain in Lmod seems to play an auxiliary role, somewhat analogous to its role in NPFs where the W domain contributes actin subunit(s) to complete a nucleus with the Arps. However, it is unknown which of the two actin subunits of the short-pitch dimer in the Lmod nucleus is staggered forward. In other words, it is unknown whether the actin subunit bound to the Nterminal flexible domain of Lmod is staggered forward with respect to the one bound to the LRR domain or vise versa. This question also applies to Tmod, whose pointed end arrangement is still unknown. Lmod’s linker-2 is also much longer than Cobl’s linker-2, conferring significant freedom with respect to the relative positioning of the third actin subunit. Thus, the third actin subunit could be at the barbed end of either the first or the second subunit. Because Lmod contains a single TM-binding site, it can be predicted that in cells the Lmod nucleus is associated with a single TM dimer, but this has not been formally demonstrated. Finally, one of the most intriguing questions about Lmod concerns the interplay with Tmod. The two proteins are clearly related and appear to have similar localization, but despite this similarity Lmod and Tmod probably have well-separated roles and may be associated with different subsets of actin filaments.

11.8 The W Domain and Filament Elongation Other than nucleation, cells also control the rates of actin filament elongation, mainly through filament capping proteins and elongation factors. Formins are unique in that they not only promote nucleation, but also processive barbed end elongation (or depolymerization). However, formins are the only proteins that do not use the W domain for nucleation or elongation. Instead, they use the dimeric forminhomology 2 (FH2) domain for interaction with actin which, as often observed with the W domain, occurs C-terminal to Pro-rich regions (named FH1 domain in formins). The reasons why the FH2 domain can sustain both nucleation and elongation are diverse, and have been discussed extensively (Chesarone and Goode,

11

The WASP-Homology 2 Domain and Cytoskeleton Assembly

269

2009; Faix and Grosse, 2006; Goode and Eck, 2007; Higgs, 2005; Paul and Pollard, 2009; Pollard, 2007). In contrast, none of the known W-based filament nucleators sustains processive barbed end elongation, which appears to be due to a single and simple reason; steric hindrance of the W domain with intersubunit contacts in the actin filament. Indeed the crystal structure of a long-pitch actin dimer stabilized by a tandem repeat of two W domains has just been determined in our lab (Rebowski et al., in preparation). The structure shows that although the two actins subunits adopt a filament-like arrangement, they are somewhat more separated than in the actin filament. The separation occurs because the second W domain, bound in the hydrophobic cleft of the second actin subunit, interferes with filament-like contacts between the two actins. The implication is that tandem W-based nucleators cannot stay bound to filaments after nucleation, and therefore are unlikely to play a role in elongation. This is also likely to explain why NPFs are ejected from the Arp2/3 complex once the branch filament begins to grow. Part of the binding interface of the W-domain remains exposed in F-actin. So, tandem W domains may weakly (and non-specifically) co-sediment with F-actin and bind to (or cap) filament barbed ends, as discussed in a recent review (Renault et al., 2008). But this also means that minor changes in the sequence of the W domain may give rise to an F-actin-binding domain. This appears to be the solution that nature has found to produce a filament elongation factor that is compatible with the W domain. Thus, the elongation function among W-based filament nucleators has been “outsourced” to a dedicated family of proteins, Ena/VASP, which have a similar domain organization and may be evolutionarily related to WASPfamily NPFs (see Fig. 11.2 and legend for details). Ena/VASP and WASP/N-WASP both contain an N-terminal EVH1 (or WH1) domain, a central Pro-rich region and W-related sequences. A trace of their relationship can still be found in the acidic region C-terminal to the W-related sequences (albeit in Ena/VASP this region is less acidic than in NPFs and lacks the important tryptophan involved in binding to the Arp2/3 complex). The G-actin-binding (GAB) domain of Ena/VASP is not only related to the W domain of NPFs, but has also been shown to interact with actin in a similar manner (Ferron et al., 2007). Immediately C-terminal to the GAB domain is the F-actin-binding (FAB) domain, which is also related to the W domain. However, the FAB domain is more closely related to the C region of NPFs (Fig. 11.2b). Indeed, the FAB domain has evolved minor differences compared to the W domain, which probably reflect adaptation to bind F-actin in a way compatible with intersubunit contacts in the filament. A similar necessity may have arisen at the interface between Arp2 and the first actin subunit of the branch (Fig. 11.3c), which probably explains the resemblance between the C and FAB domains (Fig. 11.2b). Another event in Ena/VASP’s adaptation for processive barbed end elongation is tetramerization, which is mediated by the C-terminal coiled-coil domain (Bachmann et al., 1999; Kuhnel et al., 2004). Tetramerization may allow Ena/VASP to work cooperatively, by sequentially allowing each subunit of the tetramer to release and advance during monomer addition to the barbed end while the other subunits remain attached to the growing filament.

270

R. Dominguez

11.9 Other Functions of the W Domain Actin monomer sequestration – The W domain is found in many proteins that are not associated with either nucleation or elongation. One such function is actin monomer sequestration by proteins that contribute to maintaining a large fraction (∼50%) of the cellular actin in the unpolymerized pool. Tβ4 (Fig. 11.5), a short 43-aa polypeptide, is the best-known representative of this group (Safer et al., 1990). The N-terminal 20-aa portion of Tβ4 is related to the W domain (Dominguez, 2007; Hertzog et al., 2004; Irobi et al., 2004; Paunola et al., 2002), including an amphiphilic helix that binds in the hydrophobic cleft at the barbed end of the actin monomer and an LKKT sequence. The C-terminal 23-aa amino acid portion of Tβ4 consists of an extended loop that reaches the pointed end of the actin monomer and a C-terminal helix that binds atop actin subdomains 2 and 4 (Hertzog et al., 2004; Irobi et al., 2004). The combination of N- and C-terminal helices that block monomer addition to both the barbed and pointed ends of the actin monomer and an optimal-length LKKT-containing linker make Tβ4 a remarkably simple but extremely effective actin monomer sequestering protein. As a result, Tβ4-actin complexes cannot participate in actin filament nucleation or elongation. Instead, Tβ4 is though to function as an actin buffer, losing actin in competitive equilibrium to profilin, which has higher affinity for actin monomers and functions as the main reservoir of actin monomers for filament assembly (Pollard and Borisy, 2003). Regulation of barbed end filament growth – The presence of multiple copies of the W domain does not automatically mean that a protein is a filament nucleator. For

Fig. 11.5 Model of the Tβ4-actin complex. This model was generated by combining the structures of W-actin (Chereau et al., 2005) and that of a complex of actin with the C-terminal half of Tβ4 crystallized as a hybrid construct with gelsolin segment 1 (Irobi et al., 2004). The two structures overlap in the middle section, corresponding to the LKKT motif. The sequence and secondary structure assignment of Tβ4 are shown at the bottom for reference

11

The WASP-Homology 2 Domain and Cytoskeleton Assembly

271

example, amoeba actobindin (Hertzog et al., 2002), Drosophila ciboulot (Hertzog et al., 2004) and C. elegans tetrathymosin (Van Troys et al., 2004) present twoand-a-half, three and four copies of the W fold, respectively, but do not nucleate actin filaments. Evolutionarily, these three proteins share more in common with Tβ4, in particularly the presence of Tβ4-related sequences C-terminal to the LKKT(V) motif, than with shorter classical W domains of the kind found in WASP family proteins (Chereau et al., 2005). However, ciboulot and actobindin do not sequester actin monomers like Tβ4, but rather promote filament barbed end growth in a way analogous to profilin (Carlier et al., 2007; Hertzog et al., 2002). These two proteins form 1:1 complexes with actin, suggesting that only one of their actin-binding sites is fully functional (Hertzog et al., 2002, 2004). In contrast, tetrathymosin appears to bind multiple actin monomers and has both monomer sequestering and filamentbinding properties (Van Troys et al., 2004). The W domain as an accessory domain in formin function – Although formins use the FH2 domain for interaction with actin during nucleation and elongation, there is at least one formin, INF2 (Chhabra and Higgs, 2006), that contains a W domain. IFN2 appears to use its W domain for filament depolymerization and as a diaphanous autoregulatory domain (DAD) (Chhabra et al., 2009). Cytoskeleton scaffolding – The W domain is also found in modular, scaffolding proteins that link membrane and cytoskeleton dynamics (Dominguez, 2007). A tight spatial and temporal coordination of actin polymerization and cellular membrane remodeling is a characteristic feature of many cellular processes, including endocytosis, exocytosis, cell motility and intracellular trafficking (Engqvist-Goldstein and Drubin, 2003; Kaksonen et al., 2006; Le Clainche and Carlier, 2008; Ryan, 2006; Scita et al., 2008). These processes require the formation of different types of actin filament networks (branched networks, cross-linked bundles, stress fibers) and different types of membrane structures (tubules, vesicles, protrusions, invaginations). In these processes, inverted Bin/Amphiphysin/Rvs (I-BAR) proteins are emerging as key regulators, linking signaling pathways to actin cytoskeleton and membrane dynamics (Futterer and Machesky, 2007; Machesky and Johnston, 2007; Saarikangas et al., 2009; Scita et al., 2008). MIM (missing-in-metastasis), IRSp53 (insulin receptor tyrosine kinase substrate p53) and ABBA (actin-bundling protein with BAIAP2 homology) are the best characterized members of this family (Scita et al., 2008). These are modular, multidomain proteins containing, in addition to the dimerization/membrane binding I-BAR domain, protein-protein interaction modules that link to signaling proteins and to the actin cytoskeleton. One of these modules is the W domain, present in MIM (Lee et al., 2007), ABBA (Saarikangas et al., 2008) and certain isoforms of IRSp53. The W domains of these proteins tend to be slightly longer than those of the proteins involved in nucleation and elongation, displaying interactions with actin beyond the LKKT sequence (Chereau et al., 2005; Lee et al., 2007). The complex of actin with the W domain of MIM is shown in Fig. 11.1. These W domains also tend to occur as single copies (not tandem repeats), and are found at the extreme C-termini of I-BAR proteins, which is also different from other W-containing proteins (Fig. 11.2). Another distinction is that contrary to WASP- and VASP-family proteins, the W domains of I-BAR proteins

272

R. Dominguez

are not immediately preceded by profilin-binding Pro-rich sequences. This unique domain organization points to a role of the W domain in mediating the recruitment of I-BAR proteins and their binding partners to specific cytoskeletal networks (Dominguez, 2007; Lee et al., 2007; Scita et al., 2008). In summary, because of its remarkable simplicity and short length the W domain has become the most abundant and functionally diverse actin-binding motif. Tβ4, actobindin, ciboulot, tetrathymosin, IFN2 and I-BAR proteins are only some examples of how changes in the sequence of the W domain as well as the modular organization of the proteins in which it is found give rise to diverse functions in the regulation of actin cytoskeleton dynamics (Dominguez, 2007). Other functions of the W domain are likely to emerge in the future. Acknowledgement Supported by NIH grants GM073791.

References Aguda, A.H., Burtnick, L.D. and Robinson, R.C. (2005) The state of the filament. EMBO Rep 6, 220–226. Ahuja, R., Pinyol, R., Reichenbach, N., Custer, L., Klingensmith, J., Kessels, M.M. and Qualmann, B. (2007) Cordon-Bleu Is an Actin Nucleation Factor and Controls Neuronal Morphology. Cell, 131, 337–350. Bachmann, C., Fischer, L., Walter, U. and Reinhard, M. (1999) The EVH2 domain of the vasodilator-stimulated phosphoprotein mediates tetramerization, F-actin binding, and actin bundle formation. J Biol Chem, 274, 23549–23557. Bamburg, J.R. (1999) Proteins of the ADF/cofilin family: essential regulators of actin dynamics. Annu Rev Cell Dev Biol, 15, 185–230. Bhavsar, A.P., Guttman, J.A. and Finlay, B.B. (2007) Manipulation of host-cell pathways by bacterial pathogens. Nature, 449, 827–834. Boczkowska, M., Rebowski, G., Petoukhov, M.V., Hayes, D.B., Svergun, D.I. and Dominguez, R. (2008) X-ray scattering study of activated Arp2/3 complex with bound actin-WCA. Structure, 16, 695–704. Bompard, G. and Caron, E. (2004) Regulation of WASP/WAVE proteins: making a long story short. J Cell Biol, 166, 957–962. Bosch, M., Le, K.H., Bugyi, B., Correia, J.J., Renault, L. and Carlier, M.F. (2007) Analysis of the function of Spire in actin assembly and its synergy with formin and profilin. Mol Cell, 28, 555–568. Campellone, K.G., Webb, N.J., Znameroski, E.A. and Welch, M.D. (2008) WHAMM is an Arp2/3 complex activator that binds microtubules and functions in ER to Golgi transport. Cell, 134, 148–161. Carlier, M.F., Hertzog, M., Didry, D., Renault, L., Cantrelle, F.X., van Heijenoort, C., Knossow, M. and Guittet, E. (2007) Structure, function, and evolution of the beta-thymosin/WH2 (WASPHomology2) actin-binding module. Ann NY Acad Sci, 1112, 67–75. Chereau, D., Boczkowska, M., Skwarek-Maruszewska, A., Fujiwara, I., Hayes, D.B., Rebowski, G., Lappalainen, P., Pollard, T.D. and Dominguez, R. (2008) Leiomodin is an actin filament nucleator in muscle cells. Science, 320, 239–243. Chereau, D., Kerff, F., Graceffa, P., Grabarek, Z., Langsetmo, K. and Dominguez, R. (2005) Actinbound structures of Wiskott-Aldrich syndrome protein (WASP)-homology domain 2 and the implications for filament assembly. Proc Natl Acad Sci U S A, 102, 16644–16649. Chesarone, M.A. and Goode, B.L. (2009) Actin nucleation and elongation factors: mechanisms and interplay. Curr Opin Cell Biol, 21, 28–37.

11

The WASP-Homology 2 Domain and Cytoskeleton Assembly

273

Chhabra, E.S. and Higgs, H.N. (2006) INF2 Is a WASP homology 2 motif-containing formin that severs actin filaments and accelerates both polymerization and depolymerization. J Biol Chem, 281, 26754–26767. Chhabra, E.S., Ramabhadran, V., Gerber, S.A. and Higgs, H.N. (2009) INF2 is an endoplasmic reticulum-associated formin protein. J Cell Sci, 122, 1430–1440. Conley, C.A., Fritz-Six, K.L., Almenar-Queralt, A. and Fowler, V.M. (2001) Leiomodins: larger members of the tropomodulin (Tmod) gene family. Genomics, 73, 127–139. Copeland, J.W., Copeland, S.J. and Treisman, R. (2004) Homo-oligomerization is essential for F-actin assembly by the formin family FH2 domain. J Biol Chem, 279, 50250–50256. Cossart, P. and Toledo-Arana, A. (2008) Listeria monocytogenes, a unique model in infection biology: an overview. Microbes Infect, 10, 1041–1050. Dahlgaard, K., Raposo, A.A., Niccoli, T. and St Johnston, D. (2007) Capu and Spire assemble a cytoplasmic actin mesh that maintains microtubule organization in the Drosophila oocyte. Dev Cell, 13, 539–553. Dayel, M.J., Holleran, E.A. and Mullins, R.D. (2001) Arp2/3 complex requires hydrolyzable ATP for nucleation of new actin filaments. Proc Natl Acad Sci U S A, 98, 14871–14876. Dominguez, R. (2004) Actin-binding proteins – a unifying hypothesis. Trends Biochem Sci, 29, 572–578. Dominguez, R. (2007) The beta-thymosin/WH2 fold: multifunctionality and structure. Ann N Y Acad Sci, 1112, 86–94. Dumas, J.J., Merithew, E., Sudharshan, E., Rajamani, D., Hayes, S., Lawe, D., Corvera, S. and Lambright, D.G. (2001) Multivalent endosome targeting by homodimeric EEA1. Mol Cell, 8, 947–958. Eden, S., Rohatgi, R., Podtelejnikov, A.V., Mann, M. and Kirschner, M.W. (2002) Mechanism of regulation of WAVE1-induced actin nucleation by Rac1 and Nck. Nature, 418, 790–793. Egile, C., Rouiller, I., Xu, X.P., Volkmann, N., Li, R. and Hanein, D. (2005) Mechanism of filament nucleation and branch stability revealed by the structure of the Arp2/3 complex at actin branch junctions. PLoS Biol, 3, e383. Engqvist-Goldstein, A.E. and Drubin, D.G. (2003) Actin assembly and endocytosis: from yeast to mammals. Annu Rev Cell Dev Biol 19, 287–332. Faix, J. and Grosse, R. (2006) Staying in shape with formins. Dev Cell 10, 693–706. Ferron, F., Rebowski, G., Lee, S.H. and Dominguez, R. (2007) Structural basis for the recruitment of profilin-actin complexes during filament elongation by Ena/VASP. Embo J, 26, 4597–4606. Fischer, R.S. and Fowler, V.M. (2003) Tropomodulins: life at the slow end. Trends Cell Biol, 13, 593–601. Fowler, V.M., Greenfield, N.J. and Moyer, J. (2003) Tropomodulin contains two actin filament pointed end-capping domains. J Biol Chem, 278, 40000–40009. Futterer, K. and Machesky, L.M. (2007) “Wunder” F-BAR domains: going from pits to vesicles. Cell, 129, 655–657. Gimona, M., Djinovic-Carugo, K., Kranewitter, W.J. and Winder, S.J. (2002) Functional plasticity of CH domains. FEBS Lett, 513, 98–106. Goley, E.D., Rodenbusch, S.E., Martin, A.C. and Welch, M.D. (2004) Critical conformational changes in the Arp2/3 complex are induced by nucleotide and nucleation promoting factor. Mol Cell, 16, 269–279. Goley, E.D. and Welch, M.D. (2006) The ARP2/3 complex: an actin nucleator comes of age. Nat Rev Mol Cell Biol, 7, 713–726. Goode, B.L. and Eck, M.J. (2007) Mechanism and function of formins in the control of actin assembly. Annu Rev Biochem, 76, 593–627. Gouin, E., Welch, M.D. and Cossart, P. (2005) Actin-based motility of intracellular pathogens. Curr Opin Microbiol, 8, 35–45. Gunst, S.J. and Zhang, W. (2008) Actin cytoskeletal dynamics in smooth muscle: a new paradigm for the regulation of smooth muscle contraction. Am J Physiol Cell Physiol, 295, C576–587.

274

R. Dominguez

Hall, A. (2005) Rho GTPases and the control of cell behaviour. Biochem Soc Trans 33, 891–895. Hertzog, M., van Heijenoort, C., Didry, D., Gaudier, M., Coutant, J., Gigant, B., Didelot, G., Preat, T., Knossow, M., Guittet, E. and Carlier, M.F. (2004) The beta-thymosin/WH2 domain; structural basis for the switch from inhibition to promotion of actin assembly. Cell, 117, 611–623. Hertzog, M., Yarmola, E.G., Didry, D., Bubb, M.R. and Carlier, M.F. (2002) Control of actin dynamics by proteins made of beta-thymosin repeats: the actobindin family. J Biol Chem, 277, 14786–14792. Higgs, H.N. (2005) Formin proteins: a domain-based approach. Trends Biochem Sci, 30, 342–353. Holmes, K.C. (2009) Structural biology: actin in a twist. Nature, 457, 389–390. Holmes, K.C. and Lehman, W. (2008) Gestalt-binding of tropomyosin to actin filaments. J Muscle Res Cell Motil, 29, 213–219. Holmes, K.C., Popp, D., Gebhard, W. and Kabsch, W. (1990) Atomic model of the actin filament. Nature, 347, 44–49. Irobi, E., Aguda, A.H., Larsson, M., Guerin, C., Yin, H.L., Burtnick, L.D., Blanchoin, L. and Robinson, R.C. (2004) Structural basis of actin sequestration by thymosin-beta4: implications for WH2 proteins. Embo J, 23, 3599–3608. Jewett, T.J., Fischer, E.R., Mead, D.J. and Hackstadt, T. (2006) Chlamydial TARP is a bacterial nucleator of actin. Proc Natl Acad Sci U S A 103, 15599–15604. Kaksonen, M., Toret, C.P. and Drubin, D.G. (2006) Harnessing actin dynamics for clathrinmediated endocytosis. Nat Rev Mol Cell Biol, 7, 404–414. Kelly, A.E., Kranitz, H., Dotsch, V. and Mullins, R.D. (2006) Actin binding to the central domain of WASP/Scar proteins plays a critical role in the activation of the Arp2/3 complex. J Biol Chem, 281, 10589–10597. Kim, A.S., Kakalis, L.T., Abdul-Manan, N., Liu, G.A. and Rosen, M.K. (2000) Autoinhibition and activation mechanisms of the Wiskott-Aldrich syndrome protein. Nature, 404, 151–158. Kong, K.Y. and Kedes, L. (2004) Cytoplasmic nuclear transfer of the actin-capping protein tropomodulin. J Biol Chem 279, 30856–30864. Kostyukova, A.S., Hitchcock-Degregori, S.E. and Greenfield, N.J. (2007) Molecular basis of tropomyosin binding to tropomodulin, an actin-capping protein. J Mol Biol, 372, 608–618. Kreishman-Deitrick, M., Goley, E.D., Burdine, L., Denison, C., Egile, C., Li, R., Murali, N., Kodadek, T.J., Welch, M.D. and Rosen, M.K. (2005) NMR analyses of the activation of the Arp2/3 complex by neuronal Wiskott-Aldrich syndrome protein. Biochemistry, 44, 15247–15256. Krieger, I., Kostyukova, A., Yamashita, A., Nitanai, Y. and Maeda, Y. (2002) Crystal structure of the C-terminal half of tropomodulin and structural basis of actin filament pointed-end capping. Biophys J, 83, 2716–2725. Kuhnel, K., Jarchau, T., Wolf, E., Schlichting, I., Walter, U., Wittinghofer, A. and Strelkov, S.V. (2004) The VASP tetramerization domain is a right-handed coiled coil based on a 15-residue repeat. Proc Natl Acad Sci U S A, 101, 17027–17032. Le Clainche, C. and Carlier, M.F. (2008) Regulation of actin assembly associated with protrusion and adhesion in cell migration. Physiol Rev, 88, 489–513. Le Clainche, C., Didry, D., Carlier, M.F. and Pantaloni, D. (2001) Activation of Arp2/3 complex by Wiskott-Aldrich Syndrome protein is linked to enhanced binding of ATP to Arp2. J Biol Chem, 276, 46689–46692. Lee, S.H., Kerff, F., Chereau, D., Ferron, F., Klug, A. and Dominguez, R. (2007) Structural Basis for the Actin-Binding Function of Missing-in-Metastasis. Structure, 15, 145–155. Linardopoulou, E.V., Parghi, S.S., Friedman, C., Osborn, G.E., Parkhurst, S.M. and Trask, B.J. (2007) Human subtelomeric WASH genes encode a new subclass of the WASP family. PLoS Genet, 3, e237. Liu, R., Abreu-Blanco, M.T., Barry, K.C., Linardopoulou, E.V., Osborn, G.E. and Parkhurst, S.M. (2009) Wash functions downstream of Rho and links linear and branched actin nucleation factors. Development, 136, 2849–2860.

11

The WASP-Homology 2 Domain and Cytoskeleton Assembly

275

Liverman, A.D., Cheng, H.C., Trosky, J.E., Leung, D.W., Yarbrough, M.L., Burdette, D.L., Rosen, M.K. and Orth, K. (2007) Arp2/3-independent assembly of actin by Vibrio type III effector VopL. Proc Natl Acad Sci U S A, 104, 17117–17122. Ma, L., Rohatgi, R. and Kirschner, M.W. (1998) The Arp2/3 complex mediates actin polymerization induced by the small GTP-binding protein Cdc42. Proc Natl Acad Sci U S A, 95, 15362–15367. Machesky, L.M., Atkinson, S.J., Ampe, C., Vandekerckhove, J. and Pollard, T.D. (1994) Purification of a cortical complex containing two unconventional actins from Acanthamoeba by affinity chromatography on profilin-agarose. J Cell Biol, 127, 107–115. Machesky, L.M. and Insall, R.H. (1998) Scar1 and the related Wiskott-Aldrich syndrome protein, WASP, regulate the actin cytoskeleton through the Arp2/3 complex. Curr Biol, 8, 1347–1356. Machesky, L.M. and Johnston, S.A. (2007) MIM: a multifunctional scaffold protein. J Mol Med, 85, 569–576. Machesky, L.M., Mullins, R.D., Higgs, H.N., Kaiser, D.A., Blanchoin, L., May, R.C., Hall, M.E. and Pollard, T.D. (1999) Scar, a WASp-related protein, activates nucleation of actin filaments by the Arp2/3 complex. Proc Natl Acad Sci U S A, 96, 3739–3744. Marchand, J.B., Kaiser, D.A., Pollard, T.D. and Higgs, H.N. (2001) Interaction of WASP/Scar proteins with actin and vertebrate Arp2/3 complex. Nat Cell Biol, 3, 76–82. Mattila, P.K., Salminen, M., Yamashiro, T. and Lappalainen, P. (2003) Mouse MIM, a tissuespecific regulator of cytoskeletal dynamics, interacts with ATP-actin monomers through its C-terminal WH2 domain. J Biol Chem, 278, 8452–8459. McLaughlin, P.J., Gooch, J.T., Mannherz, H.G. and Weeds, A.G. (1993) Structure of gelsolin segment 1-actin complex and the mechanism of filament severing. Nature, 364, 685–692. Miki, H. and Takenawa, T. (1998) Direct binding of the verprolin-homology domain in N-WASP to actin is essential for cytoskeletal reorganization. Biochem Biophys Res Commun, 243, 73–78. Mullins, R.D., Heuser, J.A. and Pollard, T.D. (1998) The interaction of Arp2/3 complex with actin: nucleation, high affinity pointed end capping, and formation of branching networks of filaments. Proc Natl Acad Sci U S A, 95, 6181–6186. Nolen, B.J. and Pollard, T.D. (2007) Insights into the influence of nucleotides on actin family proteins from seven structures of Arp2/3 complex. Mol Cell, 26, 449–457. Oda, T., Iwasa, M., Aihara, T., Maeda, Y. and Narita, A. (2009) The nature of the globular- to fibrous-actin transition. Nature, 457, 441–445. Paavilainen, V.O., Oksanen, E., Goldman, A. and Lappalainen, P. (2008) Structure of the actindepolymerizing factor homology domain in complex with actin. J Cell Biol, 182, 51–59. Padrick, S.B., Cheng, H.C., Ismail, A.M., Panchal, S.C., Doolittle, L.K., Kim, S., Skehan, B.M., Umetani, J., Brautigam, C.A., Leong, J.M. and Rosen, M.K. (2008) Hierarchical regulation of WASP/WAVE proteins. Mol Cell, 32, 426–438. Pan, F., Egile, C., Lipkin, T. and Li, R. (2004) ARPC1/Arc40 mediates the interaction of the actin-related protein 2 and 3 complex with Wiskott-Aldrich syndrome protein family activators. J Biol Chem, 279, 54629–54636. Panchal, S.C., Kaiser, D.A., Torres, E., Pollard, T.D. and Rosen, M.K. (2003) A conserved amphipathic helix in WASP/Scar proteins is essential for activation of Arp2/3 complex. Nat Struct Biol, 10, 591–598. Paul, A.S. and Pollard, T.D. (2009) Review of the mechanism of processive actin filament elongation by formins. Cell Motil Cytoskeleton, 66, 606–617. Paunola, E., Mattila, P.K. and Lappalainen, P. (2002) WH2 domain: a small, versatile adapter for actin monomers. FEBS Lett, 513, 92–97. Pollard, T.D. (2007) Regulation of actin filament assembly by Arp2/3 complex and formins. Annu Rev Biophys Biomol Struct, 36, 451–477. Pollard, T.D. and Borisy, G.G. (2003) Cellular motility driven by assembly and disassembly of actin filaments. Cell, 112, 453–465. Quinlan, M.E., Heuser, J.E., Kerkhoff, E. and Mullins, R.D. (2005) Drosophila Spire is an actin nucleation factor. Nature, 433, 382–388.

276

R. Dominguez

Quinlan, M.E., Hilgert, S., Bedrossian, A., Mullins, R.D. and Kerkhoff, E. (2007) Regulatory interactions between two actin nucleators, Spire and Cappuccino. J Cell Biol, 179, 117–128. Quinlan, M.E. and Kerkhoff, E. (2008) Actin nucleation: bacteria get in-Spired. Nat Cell Biol, 10, 13–15. Rebowski, G., Boczkowska, M., Hayes, D.B., Guo, L., Irving, T.C. and Dominguez, R. (2008) X-ray scattering study of actin polymerization nuclei assembled by tandem W domains. Proc Natl Acad Sci U S A, 105, 10785–10790. Renault, L., Bugyi, B. and Carlier, M.F. (2008) Spire and Cordon-bleu: multifunctional regulators of actin dynamics. Trends Cell Biol, 18, 494–504. Robinson, R.C., Turbedsky, K., Kaiser, D.A., Marchand, J.B., Higgs, H.N., Choe, S. and Pollard, T.D. (2001) Crystal structure of Arp2/3 complex. Science, 294, 1679–1684. Rodal, A.A., Sokolova, O., Robins, D.B., Daugherty, K.M., Hippenmeyer, S., Riezman, H., Grigorieff, N. and Goode, B.L. (2005) Conformational changes in the Arp2/3 complex leading to actin nucleation. Nat Struct Mol Biol, 12, 26–31. Rohatgi, R., Ma, L., Miki, H., Lopez, M., Kirchhausen, T., Takenawa, T. and Kirschner, M.W. (1999) The interaction between N-WASP and the Arp2/3 complex links Cdc42-dependent signals to actin assembly. Cell, 97, 221–231. Rosales-Nieves, A.E., Johndrow, J.E., Keller, L.C., Magie, C.R., Pinto-Santini, D.M. and Parkhurst, S.M. (2006) Coordination of microtubule and microfilament dynamics by Drosophila Rho1, Spire and Cappuccino. Nat Cell Biol, 8, 367–376. Rouiller, I., Xu, X.P., Amann, K.J., Egile, C., Nickell, S., Nicastro, D., Li, R., Pollard, T.D., Volkmann, N. and Hanein, D. (2008) The structural basis of actin filament branching by the Arp2/3 complex. J Cell Biol, 180, 887–895. Ryan, T.A. (2006) A pre-synaptic to-do list for coupling exocytosis to endocytosis. Curr Opin Cell Biol, 18, 416–421. Saarikangas, J., Hakanen, J., Mattila, P.K., Grumet, M., Salminen, M. and Lappalainen, P. (2008) ABBA regulates plasma-membrane and actin dynamics to promote radial glia extension. J Cell Sci, 121, 1444–1454. Saarikangas, J., Zhao, H., Pykalainen, A., Laurinmaki, P., Mattila, P.K., Kinnunen, P.K., Butcher, S.J. and Lappalainen, P. (2009) Molecular Mechanisms of Membrane Deformation by I-BAR Domain Proteins. Curr Biol, 19, 95–107. Safer, D., Golla, R. and Nachmias, V.T. (1990) Isolation of a 5-kilodalton actin-sequestering peptide from human blood platelets. Proc Natl Acad Sci U S A, 87, 2536–2540. Sanger, J.M. and Sanger, J.W. (2008) The dynamic Z bands of striated muscle cells. Sci Signal, 1, pe37. Scita, G., Confalonieri, S., Lappalainen, P. and Suetsugu, S. (2008) IRSp53: crossing the road of membrane and actin dynamics in the formation of membrane protrusions. Trends Cell Biol, 18, 52–60. Sept, D. and McCammon, J.A. (2001) Thermodynamics and kinetics of actin filament nucleation. Biophys J, 81, 667–674. Silacci, P., Mazzolai, L., Gauci, C., Stergiopulos, N., Yin, H.L. and Hayoz, D. (2004) Gelsolin superfamily proteins: key regulators of cellular functions. Cell Mol Life Sci 61, 2614–2623. Skwarek-Maruszewska, A., Hotulainen, P., Mattila, P.K. and Lappalainen, P. (2009) Contractilitydependent actin dynamics in cardiomyocyte sarcomeres. J Cell Sci, 122, 2119–2126. Symons, M., Derry, J.M., Karlak, B., Jiang, S., Lemahieu, V., McCormick, F., Francke, U. and Abo, A. (1996) Wiskott-Aldrich syndrome protein, a novel effector for the GTPase CDC42Hs, is implicated in actin polymerization. Cell, 84, 723–734. Tam, V.C., Serruto, D., Dziejman, M., Brieher, W. and Mekalanos, J.J. (2007) A Type III Secretion System in Vibrio cholerae Translocates a Formin/Spire Hybrid-like Actin Nucleator to Promote Intestinal Colonization. Cell Host and Microbe, 1, 95–107. Van Troys, M., Ono, K., Dewitte, D., Jonckheere, V., De Ruyck, N., Vandekerckhove, J., Ono, S. and Ampe, C. (2004) TetraThymosinbeta is required for actin dynamics in Caenorhabditis elegans and acts via functionally different actin-binding repeats. Mol Biol Cell, 15, 4735–4748.

11

The WASP-Homology 2 Domain and Cytoskeleton Assembly

277

Wang, J., Shaner, N., Mittal, B., Zhou, Q., Chen, J., Sanger, J.M. and Sanger, J.W. (2005) Dynamics of Z-band based proteins in developing skeletal muscle cells. Cell Motil Cytoskeleton, 61, 34–48. Weaver, A.M., Heuser, J.E., Karginov, A.V., Lee, W.L., Parsons, J.T. and Cooper, J.A. (2002) Interaction of cortactin and N-WASp with Arp2/3 complex. Curr Biol, 12, 1270–1278. Welch, M.D., Rosenblatt, J., Skoble, J., Portnoy, D.A. and Mitchison, T.J. (1998) Interaction of human Arp2/3 complex and the Listeria monocytogenes ActA protein in actin filament nucleation. Science, 281, 105–108. Winter, D., Lechler, T. and Li, R. (1999) Activation of the yeast Arp2/3 complex by Bee1p, a WASP-family protein. Curr Biol, 9, 501–504. Yarar, D., To, W., Abo, A. and Welch, M.D. (1999) The Wiskott-Aldrich syndrome protein directs actin-based motility by stimulating actin nucleation with the Arp2/3 complex. Curr Biol, 9, 555–558. Zalevsky, J., Grigorova, I. and Mullins, R.D. (2001) Activation of the Arp2/3 complex by the Listeria acta protein. Acta binds two actin monomers and three subunits of the Arp2/3 complex. J Biol Chem, 276, 3468–3475. Zuchero, J.B., Coutts, A.S., Quinlan, M.E., Thangue, N.B. and Mullins, R.D. (2009) p53-cofactor JMY is a multifunctional actin nucleation factor. Nat Cell Biol, 11, 451–459.

Chapter 12

Formin-Mediated Actin Assembly David R. Kovar, Andrew J. Bestul, Yujie Li, and Bonnie J. Scott

Abstract Formins are remarkable large multi-domain proteins that utilize a novel mechanism to rapidly assemble actin filaments for diverse cellular processes such as division, motility, polarity, stress fibers and focal adhesions. The formin homology 1 and 2 domains (FH1 and FH2) nucleate actin assembly and then remain processively associated with the elongating barbed end while directing the rapid addition of profilin-actin as much as 10-fold faster than the theoretical diffusion limit. Regulatory domains flank the FH1FH2 domains. Many formins are autoinhibited through association of their N- and C-terminal regulatory domains, and are partially activated by Rho GTPases. Most cells express multiple formin isoforms. We are beginning to understand that differences in both the regulatory and actin assembly properties tailor diverse formins for specific cellular processes.

Contents 12.1 Introduction and Historical Perspective . . . . . . . . . . . . . . . 12.1.1 Actin Assembly for Diverse Cellular Processes . . . . . . . 12.1.2 Formin: Right Name, Wrong Gene? . . . . . . . . . . . . 12.1.3 Formin: From “Scaffolding Protein” to Direct Stimulator of Actin Assembly . . . . . . . . . . . . . . . . . . . . 12.2 Domain Organization and Regulation of Formins . . . . . . . . . . 12.2.1 Autoregulation of the Diaphanous-Related Formins . . . . . 12.2.2 Non-Canonical Formin Regulatory Mechanisms . . . . . . . 12.3 Formin Actin Assembly Properties and Mechanisms . . . . . . . . . 12.3.1 Formin-Mediated Nucleation . . . . . . . . . . . . . . . 12.3.2 Formin-Mediated Processive Elongation . . . . . . . . . . 12.3.3 Physiological Considerations of Formin-Mediated Processive Elongation . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . .

280 280 282

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

284 284 285 286 287 287 288

. . . . . . 290

D.R. Kovar (B) Department of Molecular Genetics and Cell Biology and Department of Biochemistry and Molecular Biology, The University of Chicago, 920 East 58th Street, Chicago, IL 60637, USA e-mail: [email protected] M.-F. Carlier (ed.), Actin-based Motility, DOI 10.1007/978-90-481-9301-1_12, C Springer Science+Business Media B.V. 2010

279

280 12.3.4 Structure of the FH2 Domain . . . . . . . . . . . . . . . . 12.3.5 Mechanism of Processive Elongation by Formin FH2 Domains . 12.3.6 The FH1 Domain Allows Formin to Utilize Actin Bound to Profilin . . . . . . . . . . . . . . . . . . . . . . . . . 12.3.7 Energy for Formin-Mediated Processive Elongation . . . . . . 12.3.8 Interaction of the Formin FH2 Domain with the Side of Actin Filaments . . . . . . . . . . . . . . . . . . . . . 12.3.9 Effect of other Actin-Binding Proteins on Formin-Mediated Actin Assembly . . . . . . . . . . . . . . . . . . . . . . . . . 12.4 Cellular Roles of Formins . . . . . . . . . . . . . . . . . . . . . . 12.4.1 Cytokinesis . . . . . . . . . . . . . . . . . . . . . . . . 12.4.2 Cell/Tissue Morphogenesis . . . . . . . . . . . . . . . . . 12.4.3 Filopodia/Cell Motility . . . . . . . . . . . . . . . . . . . 12.4.4 Stress Fibers/Cell Adhesion . . . . . . . . . . . . . . . . . 12.4.5 Polarity . . . . . . . . . . . . . . . . . . . . . . . . . . 12.4.6 Regulation of Microtubule Dynamics by Formins . . . . . . . 12.4.7 Formins and Disease . . . . . . . . . . . . . . . . . . . . 12.5 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

D.R. Kovar et al.

. . . . . . . .

291 292

. . . . . 295 . . . . 298 . . . . . 299 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

300 301 302 305 305 306 307 307 308 308 310

12.1 Introduction and Historical Perspective 12.1.1 Actin Assembly for Diverse Cellular Processes Eukaryotic cells assemble actin filaments with precise architectures at the correct time and place for a variety of fundamental processes including division, adhesion, motility, establishing and maintaining polarity, coordination of cell shape, endocytosis and intracellular trafficking (Chhabra and Higgs, 2009). Filaments at the leading edge of a migrating cell and those that drive endocytic vesicles are short and “branched” (Pollard et al., 2000; Pollard and Borisy, 2003), whereas the contractile ring of dividing cells consist of bundles of anti-parallel linear filaments (Kamasaki et al., 2007). Actin assembly and organization are regulated by a plethora of actin-modulating proteins with complementary biochemical properties, from monomer-binding to filament capping, crosslinking and severing (Pollard et al., 2000). We are beginning to understand that the final arrangement of actin filaments also depends largely upon the nature of their initial genesis. Actin structures with different filament architectures are assembled by an expanding list of mechanistically and functionally diverse nucleating factors that respond to different stimuli (Chhabra and Higgs et al., 2009). Cells maintain a large reserve of actin monomers bound to profilin (Witke, 2004; Carlsson et al., 1977; Jockusch et al., 2007). Profilin-actin cannot nucleates on its own or add to the slow growing pointed end, so actin filaments with free barbed ends are required for assembly (Fig. 12.1a). Either severing or uncapping of pre-existing filaments can create free barbed ends, but most new filaments are produced by

Formin-Mediated Actin Assembly

A

profilinactin

actin

profilinactin PPPP

‘Barbed’ ends

‘Pointed’ ends

PPPP

PPP

PPPP

PPP

?

PPP

Multiple cycles of monomer addition

PPPP

P PPP

FH1

PPP

FH2

PPP

Formin-mediated assembly

elongation

P PPP

Spontaneous assembly

nucleation

B

PPP P

PPP PPPP

PPP

PPP

PP PP

PPP

PPP

ring complex

PP PP

PP PP

PPP

Barbed

PPP

PPPP

PP P

One cycle of monomer addition

PP PP

C

281

PPP

12

PP PP

Pointed

Fig. 12.1 Formin-mediated assembly of profilin-actin. (a) Actin and profilin-actin assembly in the absence of nucleation factors. The spontaneous assembly of actin is slow because actin dimers and trimers are energetically unfavorable. Furthermore, the primary source of unassembled actin in cells is bound to profilin, which inhibits nucleation. Once a stable filament is formed, profilin-actin can add to the barbed end but not the pointed end. Arrow sizes indicate relative reaction rates. (b) Working model for profilin-actin assembly in the presence of formin. Formins nucleate actin filaments with varying efficiency by stabilizing the dimer. Profilin reduces the nucleation efficiency of formins by ∼10-fold. However, upon filament nucleation or association with a pre-assembled filament, the formin FH2 domain dimer remains processively associated with the elongating barbed end while the associated FH1 domains direct the rapid addition of actin monomers bound to profilin. Actin and profilin-actin can add directly to a formin-associated barbed end, but at significantly slower rates than profilin-actin binds to the FH1 domain and is transferred to the barbed end. Multiple proline-rich stretches in the FH1 domain facilitate the rapid addition of profilin-actin, but stretches farther from the FH2 domain contribute less. Arrow sizes indicate relative reaction rates. (c) Pathway for formin-mediated elongation of profilin-actin. Profilin-actin binds to prolinerich stretches in the FH1 domain. FH1-bound profilin-actin forms a “ring complex” by binding to the FH2-associated barbed end. The red-dashed circle indicates a favorable interaction of profilin and/or profilin-actin with the FH2 domain. Profilin then dissociates from the barbed end and the FH1 domain for subsequent rounds of elongation

282

D.R. Kovar et al.

nucleation factors de novo. Three general classes of nucleation factors have been identified that have unique actin assembly mechanisms. The seven-polypeptide Arp2/3 complex initiates the assembly of a branched daughter filament at a 70◦ angle from a mother filament (Goley and Welch., 2006; Machesky et al., 1994; Pollard et al., 2007). Because capping protein rapidly binds free barbed ends, the combination of Arp2/3 complex and capping protein produces a “dendritic network” of short-branched filaments ideal for pushing (Pollard et al., 2000). Some higher eukaryotic cells express a recently identified second class of actin nucleation factors that may stabilize an actin filament nucleus by binding to multiple actin monomers through a series of repeated WH2 domains (Qualmann and Kesseles, 2009). The WH2 domain-containing nucleation factors include Spire (Kerkhoff, 2006; Quinlan et al., 2005), Cordon-bleu (Ahuja et al., 2007), and Leiomodin (Chereau et al., 2008). Formins are a third major family of nucleation factors that have a conceptually novel mechanism that allows for the efficient utilization of profilin-actin (Fig. 12.1b) (Kovar, 2006; Wallar and Alberts, 2003; Faix and Grosse, 2006; Goode and Eck, 2007). Formin nucleates new filaments and binds to the barbed end of pre-existing filaments, and then remains continually (processively) associated with the elongating barbed end. Processive barbed end association allows formin to rapidly assemble linear filaments by protecting against capping protein and stimulating the addition of profilin-actin at rates 5 to 10-fold faster than free actin. In cells, formin-assembled filaments are typically bundled and pulled by myosin motors to produce contractile forces such as the cytokinetic contractile ring and stress fibers (Glotzer, 2005; Naumanen et al., 2008).

12.1.2 Formin: Right Name, Wrong Gene? The formin family was named based upon what ultimately turned out to be the incorrect gene. However, this has proven to be a fitting identity for a gene family that is important for the assembly of numerous fundamental cellular structures. In the early 1990s the original family member was named Formin-1 because it was hypothesized to be the gene whose disruption was responsible for limb deformity defects in mice (Maas et al., 1990; Woychik et al., 1990). Almost 15 years later it was discovered that limb deformity defects likely arose from disruption of the neighboring Gremlin gene (Zuniga et al., 2004), a member of the bone morphogenic protein antagonist family. However, disruption of Formin-1 activity in a manner that does not effect Gremlin expression also causes limb development defects (Zhou et al., 2009). In the decade following the naming of mouse Formin-1, orthologous family members were discovered in fruit flies (Castrillon and Wasserman, 1994; Manseau et al., 1996), budding yeast (Evangelista et al., 1997; Imamura et al., 1997; Kohno et al., 1996), fission yeast (Chang et al., 1997; Petersen et al., 1995, 1998), filamentous fungi (Harris et al., 1997; Marhoul and Adams, 1995), nematode worm (Swan et al., 1998) and vertebrates (Watanabe et al., 1997). All members of the emerging formin family shared two important characteristics. First, they contain

12

Formin-Mediated Actin Assembly

A

Activation

283 Autoinhibition

G DID GBD

DD

CC

FH1

FH2

DAD

FH3 Regulation & Localization

B

Profilin-Actin Assembly

Regulation

Autoinhibited/Inactive (Intradimer)

GTP-Rho

Partially Active

(Interdimer)

Fig. 12.2 Formin domain organization and regulation. (a) Domain organization of a representative DRF formin homodimer such as mouse mDia1. The DD and FH2 domains facilitate dimerization. Actin assembly FH1 and FH2 domains are flanked by regulatory domains. Association of the C-terminal DAD with the N-terminal DID inhibits the FH2. Rho-GTPase partially activates FH2-mediated actin assembly by associating with the GBD region overlapping the G and DID domains, resulting in dissociation of the DAD. Domains: G, region necessary for binding RhoGTPase; GBD, GTPase-binding region; DID, diaphanous inhibitory domain; DD, dimerization domain; CC, coiled-coil; FH1, FH2 and FH3, formin homology domains; DAD, diaphanous autoregulatory domain. (b) Cartoon models of formin autoinhibition via DAD-DID association. The formin homodimer may form an intradimer by self-association, or one homodimer may interact with another homodimer to form a tetrameric interdimer. Association of Rho-GTPase to the GBD region has been shown to “partially” relieve autoinhibition in vitro, suggesting that other factors are required for full activation

extensive sequence similarity in three regions defined as formin homology domains (Fig. 12.2a; FH1, FH2 and FH3) (Castrillon and Wasserman, 1994; Petersen et al., 1998). We now know that the FH1 and FH2 domains cooperate to stimulate the rapid processive assembly of profilin-actin, whereas the FH3 is a less conserved domain important for the regulation and localization of formin. Second, the phenotypes of formin mutants suggested involvement in cellular functions dependent upon the cytoskeleton, including cytokinesis and the establishment and maintenance of polarity. Mutants of budding yeast formin bni1 fail to grow in a polarized manner and correctly polarize the underlying actin cytoskeleton (Evangelista et al., 1997). Mutants of the fission yeast formin cdc12 fail to assemble the acto-myosin contractile ring (Chang et al., 1997). In addition to genetic evidence, several additional findings indicated that formins were involved in regulating the actin cytoskeleton. First, overexpression of truncated vertebrate formin mDia1 and budding yeast formin Bni1 caused abnormal

284

D.R. Kovar et al.

organization of the actin cytoskeleton including an unusual enrichment of actin filaments (Evangelista et al., 1997; Nakano et al., 1999; Watanabe et al., 1999). Second, formins were discovered to be downstream effectors of the Rho family of small GTPases, well-known “master regulators” of the actin cytoskeleton (Evangelista et al., 1997; Imamura et al., 1997; Watanabe et al., 1997, 1999). Third, formins directly interact with the actin monomer-binding protein profilin (Evangelista et al., 1997; Imamura et al., 1997; Chang et al., 1997; Watanabe et al., 1997).

12.1.3 Formin: From “Scaffolding Protein” to Direct Stimulator of Actin Assembly By the turn of the century, the specific role of formins in actin-mediated assembly was not known. Given that formins are multi-domain proteins that interact with key upstream (Rho GTPase) and downstream (profilin) regulators of actin assembly, the leading hypothesis was that formins serve as scaffolds to localize actin-building machinery (Frazier and Field, 1997; Tanaka, 2000). However, in 2002 it was discovered that the budding yeast formin Bni1 is directly responsible for stimulating the assembly of actin cables, which are required for the establishment and maintenance of cell polarity (Evangelista et al., 2002; Sagot et al., 2002a). In vitro, nanomolar concentrations of the Bni1 FH2 and FH1FH2 domains directly stimulate actin filament nucleation in the absence of other factors (Pruyne et al., 2002; Sagot et al., 2002b). These pioneering studies demonstrated that formin directly stimulates actin assembly and suggested a unique mechanism whereby formin allows barbed end elongation despite interacting with the barbed end (Pruyne et al., 2002). FH2 and FH1FH2 fragments from a range of evolutionarily diverse formins have now been characterized that stimulate actin assembly. In this chapter we focus primarily on formin’s roles in regulating actin assembly, including mechanisms of formin regulation, actin nucleation and elongation. However, it should be noted that formins have also been implicated in several other cellular activities such as regulating microtubule functions and microtubule-actin crosstalk reviewed in (Bartolini and Gundersen, 2009), as well as cell signaling events including transcription regulation through serum response factor and Src signaling reviewed in (Young and Copeland, 2008). We also recommend two reviews on formin-mediated actin assembly (Goode and Eck, 2007; Paul and Pollard, 2009a).

12.2 Domain Organization and Regulation of Formins Most eukaryotes contain multiple formin genes, including budding yeast S. cerevisiae (two), fission yeast S. pombe (three), fruit fly D. melanogaster (six), nematode worm C. elegans (six), vertebrates (fifteen) and flowering mustard weed A. thaliana (twenty-one) (Chalkia et al., 2008; Higgs and Peterson, 2005; Rivero et al., 2005).

12

Formin-Mediated Actin Assembly

285

Based upon sequence comparisons and predicted secondary and tertiary protein folds, prokaryotes do not appear to contain genes homologous to formins (Chalkia et al., 2008). Formins are large proteins (1000–2000 residues) with multiple domains that control actin assembly and regulation (Fig. 12.2a). The defining feature of all formins are the formin homology 2 (FH2) domain and adjacent FH1 domain, which are situated near the C-terminus of most formin genes. The FH2 is necessary and sufficient to stimulate actin assembly and FH1 interacts with profilin. The FH2 domain is the most highly conserved region of formins, which was originally identified as a ∼130 residue sequence in the D. melanogaster gene diaphanous by Castrillon and Wasserman (Castrillon and Wasserman, 1994). The FH1 domain was also identified as a region that contains a variable number of short tracks of proline residues that lie immediately upstream of the FH2 domain of diaphanous and other formin genes (Castrillon and Wasserman, 1994). Based upon the comparison of multiple formin sequences and solving several crystal structures (Higgs and Peterson, 2005; Rivero et al., 2005; Xu et al., 2004), we now know that the FH2 domain spans ∼400 residues. Animal formins have been divided into seven groups (Higgs and Peterson, 2005; Rivero et al., 2005), including the most well-studied diaphanousrelated formin group (DRFs). Nonmetazoan FH2 domains, such as from plants and fungi, have similar structural and functional properties but are phylogenetically distinct.

12.2.1 Autoregulation of the Diaphanous-Related Formins Formins contain less conserved regulatory regions flanking the actin assembly FH1 and FH2 domains (Fig. 12.2a). The DRFs regulatory regions control auto-inhibition through interaction of their N- and C-termini (Fig. 12.2a, b). Structural and biochemical analyses of the regulatory regions of DRF formins, particularly mouse Dia1, have revealed important mechanistic insights (Fig. 12.2, (Li and Higgs, 2005; Otomo et al., 2005; Rose et al., 2005)). The diaphanous autoregulatory domain (DAD) is a stretch of 20–30 amino acids found immediately C-terminal of the FH2 domain (Wallar and Alberts, 2003). The DAD domain binds with sub-μM affinity to the diaphanous inhibitory domain (DID), a ∼250 residue region found in the N-terminal half of DRFs (Li and Higgs, 2005; Lammers et al., 2005). Interaction of the DAD and DID domains is necessary and sufficient for auto-inhibition. RhoGTPases “partially” activate DRFs by binding with low micromolar affinity to a GTPase binding domain region (GBD) that comprises a G region (residues upstream of DID that are required for the GTPase to bind) and a portion of the DID domain that overlaps with the DID-DAD interaction (Li and Higgs, 2005; Rose et al., 2005; Nezami et al., 2006; Lammers et al., 2005). Therefore, although association of Rho and DAD with the N-terminal GBD-DID domains are mutually exclusive, their binding sites only partially overlap, a common theme for autoinhibitory regulation of Ras GTP effectors (Otomo et al., 2005). Immediately downstream of the DID domain are a dimerization domain (DD) and a coiled-coil (CC) domain (Fig. 12.2).

286

D.R. Kovar et al.

The FH3 domain defines a region of sequence similarity between DRFs that includes the DID and DD domains, and is thought to be important for proper localization (Petersen et al., 1998). The DID domain consists of four complete armadillo repeats and a fifth truncated repeat (Otomo et al., 2005; Rose et al., 2005). The DID armadillo repeats form a highly conserved concave surface that creates a binding site for the amphipathic DAD helix through hydrophobic interaction. The DAD domain was first discovered by sequence comparison of the C-terminal regions of DRFs (Alberts, 2001). Sequence alignment revealed a highly conserved DID binding motif “MDXLLXL” and a downstream basic region of typical sequence “RRKR.” Both regions are essential for regulating formin activity in vivo (Alberts, 2001; Wallar et al., 2005). Rho makes critical contacts with the N-terminus of formin at two sites (Rose et al., 2005). The switch I of Rho binds to the G domain while the switch II binds to the DID, which ultimately displaces the DAD from the DID via a two-step mechanism. First, Rho binds the G domain with low affinity, which induces the G domain to undergo a conformational change that brings the G domain and associated Rho closer to the DID. Second, Arg68 in the Rho switch II hydrogen bonds with Asn217 in the DID, destabilizing the DID-DAD interaction. Repulsion of DAD by acidic residues in Rho also contributes to dissociation of DAD. Rho-bound GBD also blocks the concave surface of the DID such that the DAD cannot bind (Nezami et al., 2006; Lammers et al., 2005). Given that structural and biochemical studies suggest that Rho is fully capable of displacing the DAD (Li and Higgs, 2005), it remains a puzzle as to why Rho does not fully activate DRFs.

12.2.2 Non-Canonical Formin Regulatory Mechanisms All formins that contain DID and DAD domains may have a similar mechanism of autoinhibiton and partial activation by Rho GTPases, such as the animal formin classes Dia (Diaphanous and Diaphanous-like), DAAM (disheveled-associated activators of morphogenesis), FRL (formin-related proteins identified in leukocytes), FHOD (formin homology domain-containing proteins), as well as non-metazoan formins such as budding yeast formins Bni1 and Bnr1 or the fission yeast formin For3 (Evangelista et al., 1997; Imamura et al., 1997; Kohno et al., 1996; Dong et al., 2003; Martin et al., 2007; Schulte et al., 2008). However, given that autoinhibition of these canonical DRFs can only be “partially” relieved by Rho-GTPase, there is much to learn about their regulation and it is likely that mechanisms for full activation will vary widely. In addition to Rho-GTPase, autoinhibited budding yeast Bni1 is activated through phosphorylation by the actin-regulating kinase Prk1 (Wang et al., 2009). Although the human formin FHOD1 is autoinhibited like mDia1, the structure of the N-terminal regions of mDia1 and FHOD1 are substantially different and FHOD1 is activated by phosphorylation (Takeya et al., 2008; Schulte et al., 2008). Furthermore, Daam1 is not significantly activated by Rho, but is activated by Dishevelled (Dvl) (Liu et al., 2008).

12

Formin-Mediated Actin Assembly

287

Other formins contain canonical DID and DAD regulatory domains, but do not appear to be regulated by conventional auto-inhibition and GTPase activation. The fission yeast formin Cdc12 contains putative DID and DAD sequences but current evidence suggests that Cdc12 is not regulated through these domains (Yonetani et al., 2008). Mammalian formins INF2 and FRL2 contain apparently conventional DID and DAD domains that directly interact, but their association does not inhibit the formin’s ability to stimulate actin assembly (Chhabra et al., 2009; Vaillant et al., 2008). It is important to be aware that other formins have distinct regulatory domain organizations completely different than the DRFs (Wallar and Alberts, 2003; Goode and Eck, 2007; Chalkia et al., 2008; Higgs, 2005), so their mode of regulation is certain to be different. The animal formin group FMN lacks obvious domains in their putative regulatory regions. Furthermore, some formins that do not contain DID and DAD domains have other recognizable domains. Plants express at least twenty-one formins that fall into three distinct classes that are different than DRFs (Chalkia et al., 2008; Grunt et al., 2008). Class I plant formins contain transmembrane sequences, class II plant formins carry an N-terminal lipid phosphatase PTEN-like domain and class III plant formins contain catalytically-inactive RhoGAP-like domains. It will be interesting to resolve how these “unconventional” formins are regulated.

12.3 Formin Actin Assembly Properties and Mechanisms 12.3.1 Formin-Mediated Nucleation Upon activation, the primary function of formin is to stimulate nucleation of new filaments, or associate with the barbed end of pre-existing filaments, and drive the rapid processive elongation of unbranched filaments. Purified fragments containing the FH2 or FH1FH2 domains from evolutionarily diverse formins have similar qualitative effects, but significantly different quantitative effects on the assembly of actin monomers in vitro The FH2 domain is necessary and sufficient to nucleate actin assembly. The precise mechanism of nucleation is not known. Kinetic simulation/mathematical models and structural studies suggest that nucleation is achieved by stabilizing dimers and/or trimers, which are energetically unfavorable nucleation intermediates (Fig. 12.1b) (Xu et al., 2004; Otomo et al., 2005; Pring et al., 2003; Shimada et al., 2004). Given that formin eliminates the “nucleation lag” at the outset of a spontaneous assembly reaction (Pruyne et al., 2002; Sagot et al., 2002b; Li and Higgs, 2003; Pring et al., 2003; Kovar et al., 2003), it is reasonable to hypothesize that formins predominantly stabilize the actin dimer. Like most actin assembly properties of formins, the nucleation efficiency (number of formin dimers required to initiate the assembly of a new filament) varies between formin isoforms. Under optimal conditions the Arp2/3 complex has a maximum nucleation yield of 100%, one new filament per Arp2/3 complex (Higgs et al., 1999). The reported nucleation

288

D.R. Kovar et al.

efficiencies of formins are less than 100%, and vary significantly. For example, the cytokinesis formins from fission yeast (Cdc12) and nematode worms (CYK-1) produce a new filament per ∼2.0 and ∼25 dimers (Kovar et al., 2003; Neidt et al., 2008). Human Daam1 stimulates actin assembly ∼10-fold less than mDia1 and mDia2 (Lu et al., 2007). Similarly, mouse formins mDia1, mDia2 and INF2 are 100-fold more potent actin assembly stimulators than mouse formin FRL (Chhabra and Higgs, 2006; Harris et al., 2004, 2006). Formins with high nucleation efficiencies, such as fission yeast Cdc12, likely generate new filaments in cells (Vavylonis et al., 2008). However, there is no experimental evidence demonstrating that any formin is required to nucleate actin filament assembly for their cellular function. Alternatively, it is possible that formins associate with the barbed end of pre-existing filaments. Formins with both low and high nucleation efficiency bind to the barbed end of preassembled filaments tightly, with low nanomolar affinity (Pring et al., 2003; Neidt et al., 2008; Harris and Higgs, 2004; Kovar et al., 2005; Moseley et al., 2004). Formin isoforms with low nucleation efficiencies may bind to filaments assembled by other actin nucleation factors such as the Arp2/3 complex (Chesarone and Goode, 2009). The “convergent elongation model,” originally proposed to explain how vasodilator-stimulated phosphoprotein (VASP) uses Arp2/3 complex-nucleated filaments at the leading edge of motile cells to initiate the assembly of filopodia (Svitkina et al., 2003), might also explain how formins function in particular actin-based cellular structures like filopodia (Schirenbeck et al., 2005a, b; Yang et al., 2007).

12.3.2 Formin-Mediated Processive Elongation Following nucleation of a new filament or binding to the barbed end of a preassembled filament, formins (with only one reported exception, the plant formin FORMIN1 (Michelot et al., 2005)) perform their most remarkable and functionally important “stunt”: the ability to remain processively associated with the elongating barbed end (Fig. 12.1b). Like many innovative hypotheses, the idea that formin binds to the barbed end and allows subunit addition (Pruyne et al., 2002) was initially met with considerable skepticism. Proteins that bind to the actin filament barbed end have been well established, such as CapZ (capping protein) and gelsolin, but these proteins completely inhibit subunit addition and loss (Pollard, 2000). Given that there is only one position on the actin filament barbed end where the next actin monomer can add (Holmes et al., 1990), the ability to remain bound to the barbed end as new subunits are added would be quite remarkable. Proteins/complexes that remain continually associated with the elongating barbed end and modify the rate of monomer addition had been theorized (Dickinson and Purich, 2002; Dickinson et al., 2002), but not established. Evidence that formins associate with the barbed end included discoveries that (A) formins prevent the end-to-end annealing of actin filaments (Kovar et al., 2003), (B) fission yeast formin Cdc12 and mDia3 increases the barbed end

12

Formin-Mediated Actin Assembly

289

critical concentration towards that of the pointed end (Shimada et al., 2004; Kovar et al., 2003), (C) some formins allow barbed ends to elongate in the presence of capping proteins (Harris and Higgs, 2004; Kovar et al., 2005; Zigmond et al., 2003), and (D) formins reduce the rate that pre-assembled filaments polymerize their barbed ends (Harris and Higgs, 2004; Kovar et al., 2004; Zigmond et al., 2003). All four of these findings indicated that formin not only binds to the actin filament barbed end, but also reduces access to the barbed end by other proteins including actin monomers, capping protein and actin filament pointed ends. Although consistent with processive association, all of these observations could be explained by a mechanism whereby formin rapidly comes on and off the barbed end. Formins were originally hypothesized to be processive because electron micrographs showed gold-labeled formin molecules near the actin filament barbed end (Pruyne et al., 2002), and because formin-dependent actin cables in budding yeast appeared to be elongating out of fixed points at the bud tip where formin is localized (Tony Bretscher, personal communication). Direct observation by fluorescence microscopy has provided the most convincing evidence that formins remain processively associated with elongating actin filament barbed ends. First, purportedly individual fluorescent formin dimer GFP-mDia1(FH1FH2) spots in transfected Xenopus fibroblasts were observed to translocate in linear 10 μm paths at a rate of 2.0 μm/sec, the same rate as actin assembly (Higashida et al., 2004). The fluorescent GFP-mDia1 spots were interpreted to be processively associated with an individual elongating actin filament barbed end. Second, individual actin filament barbed ends were observed by Total Internal Reflection Fluorescence microscopy (TIRFM) to elongate from fixed points on glass microscope slides coated with four different types of purified formins; Budding yeast Bni1, mouse mDia1 and mDia2, and fission yeast Cdc12 (Kovar et al., 2006; Kovar and Pollard, 2004). The addition of new actin monomers to the barbed end attached to formin immobilized on the slide surface was verified by flowing in brighter monomers (Kovar and Pollard, 2004), and by observations of filament buckling upon capturing the pointed end as it is “pushed” away from the elongating barbed end (Kovar et al., 2006; Kovar and Pollard, 2004). Third, two heterogeneous populations of filaments are simultaneously present in spontaneous actin assembly reactions visualized by TIRFM that elongate at significantly different rates (control filaments and formin-associated filaments) (Kovar et al., 2003, 2006; Neidt et al., 2008; Paul and Pollard, 2008, 2009b; Neidt et al., 2009). Fourth, individual actin filaments were observed to elongate from mouse formin mDia1-associated beads for > 10 mins without detachment, producing pushing forces that propelled the bead (Romero et al., 2004). Fifth, budding yeast formin Bni1 labeled with quantum dots were observed by TIRFM to move with the elongating barbed end of individual actin filaments (Paul and Pollard, 2009b). Because both profilin and FH1 were found to be required for the continuous propulsion of mDia1 functionalised beads (Romero et al., 2004), a mechanism for processive association was proposed whereby profilin molecules bound to the FH1 domain interacts with the ultimate and penultimate actin subunits and therefore allows formin to walk along the end of the actin filament like “climbing a rope”

290

D.R. Kovar et al.

(Romero et al., 2004, 2007). However, based upon TIRFM observation of individual filaments, it is now largely accepted that the FH2 domain is necessary and sufficient for processive association (Fig. 12.1b). First, diverse formin FH1FH2 or FH2 constructs including mouse mDia1 and mDia2, budding yeast Bni1, fission yeast Cdc12 and nematode worm CYK-1, are processive in the absence of profilin (Neidt et al., 2008, 2009; Kovar et al., 2006; Kovar and Pollard, 2004; Paul and Pollard, 2009b; Skau et al., 2009). Second, FH1FH2 constructs of mouse mDia1 and mDia2 are processive in the presence of profilin mutants that either cannot bind to formin or cannot bind to actin (Kovar et al., 2006).

12.3.3 Physiological Considerations of Formin-Mediated Processive Elongation All well-characterized formins facilitate the addition of thousands of subunits before dissociating from the elongating barbed end in vitro, but the specific dissociation rate varies by two orders of magnitude depending upon the particular formin (Neidt et al., 2008; Kovar et al., 2006; Paul and Pollard, 2009b; Romero et al., 2007). For example, the dissociation rates of the cytokinesis formins from fission yeast (Cdc12) and nematode worm (CYK-1) are ∼7.0 x 10−5 s−1 and ∼4 x 10−3 s−1 (Neidt et al., 2008). Differences in the length of the flexible linker joining the two halves of the FH2 dimer (Fig. 12.3b) correlate with the dissociation rate, with longer linkers leading to faster dissociation rates (Paul and Pollard, 2009b). The intrinsic dissociation rate might be extremely important for tailoring particular formin isoforms for their cellular functions (Paul and Pollard, 2009a). Given the relatively slow dissociation rates of formins, filaments become hundreds of microns long in vitro, much longer than filaments found in cells for many formin-dependent cellular processes (Kamasaki et al., 2005, 2007). Formin-dependent elongation might be terminated by re-activated auto-inhibition (Wallar and Alberts, 2003; Higgs, 2005), or the action of other factors such as capping protein (Romero et al., 2004). It was recently discovered that the budding yeast formin Bnr1 is displaced from growing barbed ends by the cell polarity factor Bud14 (Chesarone et al., 2009). Dia-interacting protein DIP is another negative regulator of formins that was found to specifically modulate the actin assembly properties of the mouse formin mDia2 (Eisenmann et al., 2007). Another functional consequence of processive elongation is that formins are able to establish and maintain proper polarity by positioning actin filament barbed ends at specific intracellular locations. For example, by localizing to the bud neck and bud tip, budding yeast formins Bnr1 and Bni1 assemble polarized actin cables whose barbed ends are oriented for myosin-V directed delivery of vesicles and organelles to the division site and the expanding bud tip (Evangelista et al., 1997; Imamura et al., 1997; Bretscher, 2003; Ozaki-Kuroda et al., 2001; Pruyne et al., 2004; Pruyne et al., 1998). Alternatively, by localizing to discrete foci at the division site in fission yeast, formin Cdc12 initiates the assembly of anti-parallel oriented filaments that

12

Formin-Mediated Actin Assembly

pointed-end

B

st po

A

291

3 coiled-coil

2 post

1

las so

ob

linker

kn

knob barbed-end

Fig. 12.3 Structure of the FH2 domain. (a) Ribbon diagram of the Bni1 FH2 domain crystal structure (residues 1350–1760; PDB 1Y64) as seen in the co-crystal with actin (Otomo et al., 2005). The FH2 domain forms a tethered homodimer in the shape of a donut. The two subunits are shown in green and purple and labels indicate approximate positions of the lasso, flexible linker, knob, coiled-coil, and post regions of the green subunit. (b) Bni1 FH2 domain in complex with three contiguous subunits of muscle actin. Actin subunits are shown as space filling representations in orange, blue, and grey and numbered 1–3 from the barbed end. Ribbon diagrams of two FH2 subunits are colored as in (a). Each FH2 monomer contains two actin-binding sites, where each site interacts with a different actin subunit. This view shows the two actin contacts made by the green FH2 subunit, where the knob binds in the hydrophobic groove between subdomains 1 and 3 of actin subunit 2 and the lasso/post makes electrostatic interactions along subdomain 1 of actin subunit 1. The partial transparency of the actin subunits reveals the symmetrical contacts of the purple FH2 subunit to actin subunits 2 and 3. Together, when all knob and post sites are engaged each FH2 dimer makes contact to three actin subunits

are ultimately “pulled” past each other by the molecular motor myosin II to drive contractile ring assembly and constriction (Vavylonis et al., 2008).

12.3.4 Structure of the FH2 Domain Crystal structures of three FH2 domains have been solved; mouse mDia1 (Shimada et al. 2004), budding yeast Bni1 (Xu et al. 2004; Otomo et al. 2005), and human DAAM1 (Lu et al. 2007; Yamashita et al., 2007). The structure of the Bni1 FH2 domain has been determined in both the actin-free and actin-bound states (Xu et al. 2004; Otomo, 2005), and is key to our current models for the mechanisms of forminmediated actin assembly (Fig. 12.3). In the actin-free form the FH2 domain forms a torus-shaped dimer (Xu et al., 2004), which suggested that the FH2 dimer encircles the growing barbed end of the actin filament. Each FH2 monomer is oriented in a head-to-tail fashion and composed of bundles of alpha helices that fall into four subregions from the N- to C-terminus: the lasso, knob, coiled-coil, and post (Fig. 12.3a).

292

D.R. Kovar et al.

Linking the lasso and knob regions is a mostly unstructured 17 amino acid peptide. To dimerize, the N-terminal lasso region of each monomer wraps around to the C-terminal post region of the other, which results in a “tethered dimer” architecture. The extensive lasso/post dimerization interface leads to a relatively stable dimer, is highly conserved among all known FH2 domains, and includes the core G-N-Y/FM-N motif. Comparing the crystal structures of wild-type and linker region mutant FH2 domains, suggested the linker provides flexibility to the FH2 dimer. By combining 3-D structural information with amino acid conservation, it is expected that the flexible “tethered dimer” architecture is characteristic of all FH2 domains. Subsequently, a remarkable co-crystal of the Bni1 FH2 domain with actin (Fig. 12.3b) revealed an FH2 dimer wrapped around an actin filament (Otomo et al., 2005). The fold and relative orientations of the knob, coiled-coil, and post regions are similar to the actin-free homodimer (Fig. 12.3b). This suggests that the donutlike shape is the relevant actin-binding form of the dimer. However, the spacing between the two FH2 monomers was dramatically wider in the actin-bound form than the actin-free form due to differences in the relative orientation of the linkers. This demonstrates that the linker provides the flexibility for the FH2 dimer to accommodate the width of an actin filament through the center. Each FH2 monomer contains two actin-binding sites, where each site interacts with a different actin subunit. One actin-binding site is located in the knob region and the second is located in the lasso/post region (Fig. 12.3). The knob site binds in a hydrophobic groove between subdomains one and three of actin, in which the invariant FH2 Ile1431 plays a key role. The post site makes primarily electrostatic interactions along subdomain one of actin in which the highly conserved FH2 Lys1359 and Lys1601 make critical contacts. Together, when all knob and post sites are engaged, each FH2 dimer makes contact to three actin subunits (Fig. 12.3b). More recently, a structure of human DAAM1 revealed a similar conserved helical core and “tethered dimer” architecture (Lu et al., 2007; Yamashita et al., 2007). However, the relative orientations of the knob, coiled-coil, and post regions, as well as dimerization contacts are different than the Bni1 structures. This observation supports the idea that FH2 dimers are inherently flexible. Interestingly, the DAAM1 structure revealed a novel β–strand interaction formed by the two ends of the linker segment. This interaction stabilizes the two monomers in an orientation that occludes the actin binding surfaces. Mutation within the β-strands increased the actin assembly activity of DAAM1 in vitro (Lu et al., 2007).

12.3.5 Mechanism of Processive Elongation by Formin FH2 Domains How FH2 domains remain processively associated with the elongating barbed end is a fascinating biophysical puzzle that has not been conclusively solved. In cases where direct visualization of elongating formin-associated actin filaments by

12

Formin-Mediated Actin Assembly

293

Table 12.1 Formin-Mediated Actin Assembly Rates Barbed end association rate constant (μM−1 s−1 ) Formin (FH1FH2)

Without Profilin

With Profilin

References

Actin only Fission yeast Cdc12

10 0.2

10 12.5

Mouse Dia2 Budding yeast Bni1

2 5–7.5

15 25

Moss For1 Worm CYK-1 Mouse Dia1

NDa 6 9

25 60 50–100

Moss For2A

NDa

125

Pollard (1986) Kovar et al. (2003), Kovar and Pollard (2004), Paul and Pollard (2009b), Skau et al. (2009) Kovar et al. (2006) Kovar et al. (2006), Kovar and Pollard (2004), Paul and Pollard (2008, 2009b) Vidali et al. (2009) Neidt et al. (2008, 2009), Kovar et al. (2006), Romero et al. (2004) Vidali et al. (2009)

a ND,

not determined

TIRFM has been utilized, formin FH2 domains reduce the rate of barbed end elongation of free G-actin. However, the magnitude of the effect varies significantly between different formins, from 10 to 99% (Table 12.1) (Li and Higgs, 2003; Otomo et al., 2005; Pring et al., 2003; Kovar et al., 2003, 2004, 2006; Neidt et al., 2008; Harris et al., 2004; Moseley et al., 2004; Zigmond et al., 2003; Romero et al., 2004). Differences in elongation rates (Kovar et al., 2006), structural studies (Xu et al., 2004; Otomo et al., 2005) and theoretical considerations (Vavylonis et al., 2006), suggest a “gating” model whereby the FH2 dimer on the elongating barbed end is in equilibrium between a “closed” state, which does not allow subunit addition, and an “open” state, in which a monomer can be added to the barbed end (Otomo et al., 2005). The model postulates that the two halves of the FH2 dimer, where each half consists of the lasso of one subunit attached to the post, coiled coil, and knob of the other (also called the hemidimer or bridge element), move as two separate units, flexibly connected by the linkers. Transition from closed to open states is thought to occur as one half of the FH2 dimer moves towards the barbed end, causing the two halves of the FH2 dimer to exchange positions and roles (Otomo et al., 2005). Variance in elongation rates between barbed ends associated with different formins suggests that the equilibrium between open and closed (Ko/c ) differs. The idea is that slowly elongating formins prefer the closed state, whereas faster elongating formins prefer the open state (Table 12.1). Fission yeast Cdc12 inhibits elongation by 99% (Ko/c = 0.01; (Kovar et al., 2003; Neidt et al., 2008; Paul et al., 2009b; Skau et al. 2009)), whereas mouse mDia1 slows elongation by less than 10% (Ko/c = 0.9; (Kovar et al., 2006; Romero et al., 2004)). In addition to variations in the transition rate between closed and open states, the rate of actin addition/dissociation from

294

D.R. Kovar et al.

the barbed end in the open state might also contribute to elongation differences of barbed ends associated with diverse formins. Based on this “two-state” hypothesis, two models of processive formin elongation have been proposed: the “stair stepping” model and the “stepping second” model (Paul and Pollard, 2009a; Xu et al., 2004; Higgs, 2005; Otomo et al., 2005; Moseley et al., 2004; Zigmond et al., 2003). In both cases, the FH2 dimer must undergo a conformational change from “closed” to the “open” states (Otomo et al., 2005), and satisfy the requirements that formins: (A) distinguish between the actin at the barbed end versus the interior of the filament, and (B) translocate onto the barbed end with the addition of each new actin subunit. In the stair stepping model, the formin translocates first followed by actin monomer addition, whereas in the stepping second model the actin monomer adds first followed by formin translocation (Paul and Pollard, 2009a). The stair stepping model proposes that in the closed state, the formin dimer binds three actin subunits at the barbed end with all four actin-binding sites (Fig. 12.3b; two knobs and two post sites). This interaction would be strong and sterically occlude addition of actin subunits. For a new monomer to add, one of the FH2 dimer halves dissociates from the barbed end (termed the leading subunit) and may reform one of its two actin contacts, where the knob site is bound and the post site is free in solution. The other half of the dimer remains bound (termed the trailing subunit). In doing so, the formin weakens its attachment to the barbed end but allows an incoming monomer to make contact with both the terminal actin subunit and the exposed leading FH2 domain, a conformation called the open state. Once the actin subunit has bound, the formin re-establishes its four actin contacts and is in the closed state. The trailing formin subunit is now poised to dissociate and continue the cycle in a “stair stepping” fashion. In the stair stepping model, it was first proposed that the FH2 dimer discriminates the terminal actin subunits by anti-cooperativity of the two actin binding halves, where one FH2 half binds with high affinity and the other half binds with low affinity (Xu et al., 2004). This model was revised upon solving the FH2-actin co-crystal, which showed the FH2 domain making four ideal actin contacts when the underlying actin monomers were over-rotated by 180◦ (Otomo et al., 2005). Therefore, there might be two local actin filament environments, a strained conformation at the barbed end (over rotated/180◦ ) and an unstrained conformation in the interior of the filament (167◦ rotation). The FH2 domain can discriminate the barbed end by preferring the strained actin conformation. As the formin half transitions from binding actins 2 and 3 in the closed state to the terminal actins 1 and 2 in the unfavorable open state, it releases actin 3 from the strained conformation and allows it to adopt a more optimal geometry (to 167◦ rotation). Therefore, the unfavorable open state is in equilibrium with the relief of strain energy of actin 3 and propels the FH2 dimer towards the barbed end. The stepping second model proposes that in the closed state, the formin dimer makes all four strong actin contacts at the barbed end only if the terminal actin subunits are strained. A new actin subunit must make contact first (possibly in a strained conformation (Otomo et al., 2005)), which allows the formin-bound actin subunits to relax their twist due to neighbor effects of interior actin subunits. This

12

Formin-Mediated Actin Assembly

295

causes the formin dimer to adopt a more strained or open state where the linker between the FH2 subunits distorts and both FH2 domains cannot fully engage the actin filament. This unfavorable, high-energy state of the formin dimer is relieved by translocation to the terminal strained actin subunits where it can revert to a closed state and re-establish the four strong actin contacts. All models for FH2 processive elongation involve one or more steps where the FH2 dimer is not fully engaged through all four of its strong actin-binding sites. Since the models differ on whether the leading FH2 subunit dissociates before monomer addition (stair stepping) or after monomer addition (stepping second), they can be distinguished by determining whether the rate of formin dissociation from the barbed end (assumed to occur in the “open” state) is coupled to the filament elongation rate. Paul and Pollard (2008) demonstrated by TIRFM that the dissociation rate of FH2 domains is proportional to the actin elongation rate (Paul and Pollard, 2008). This result is consistent with the stepping second model, where the rate at which formins dissociate from the barbed end is expected to be proportional to the number of cycles of actin monomer addition. Biochemical analyses of different formin FH2 domains show drastically different dissociation rates. For example, C. elegans cytokinesis formin CYK-1 dissociates from barbed ends ∼50 times faster that S. pombe cytokinesis formin Cdc12 (Neidt et al., 2008). Interestingly, these formins also differ in the length of their linker regions (22 aa for CYK-1 versus 14 aa for Cdc12). It has been postulated and experimentally confirmed that longer linker regions proportionally increase the formin dissociation rate (Paul and Pollard, 2009b). Another aspect of all models describing processive formin elongation is the “rotation paradox” (Kovar and Pollard, 2004; Shemesh et al., 2005), where the FH2 dimer must rotate around the filament to maintain its actin contacts as the helix grows. Surprisingly, formins immobilized to glass slides showed no rotation of the actin filament (Kovar and Pollard, 2004), which suggested an untested hypothesis that the FH2 dimer can move backwards around the filament axis to relieve strain built up from actin monomer addition (Shemesh et al., 2005).

12.3.6 The FH1 Domain Allows Formin to Utilize Actin Bound to Profilin Unassembled actin in cells is primarily associated with profilin, a small ∼14 kDa protein that can simultaneously bind actin monomers and poly-L-proline (Pollard et al., 2000; Witke, 2004; Carlsson et al., 1977; Kaiser et al., 1999; Carlier et al., 2003). Profilin is abundantly expressed in almost all cell types, binds to actin tightly and can “pilfer” monomers away from other proteins such as Thymosin beta 4 and cofilin (Didry et al., 1998; Pantaloni and Carlier, 1993). Profilin strongly inhibits nucleation, but profilin-actin readily adds to the barbed end once a filament is formed (Fig. 12.1a) (Kang et al., 1999; Pollard and Cooper, 1984). Formin is a remarkable molecule that is able to stimulate the addition of profilin-actin up to

296

D.R. Kovar et al.

10-fold faster than profilin-actin adds to a free barbed end. Figure 12.1b, c present a working hypothesis for how the formin FH1 and FH2 domains cooperate to drive the rapid processive assembly of profilin-actin. Fragments containing both the FH1 and FH2 domains, but not the FH2 domain alone, can stimulate the assembly of profilin-actin (Pruyne et al., 2002; Li and Higgs, 2003; Kovar et al., 2003). FH1 domains contain a variable number of distinct tracks of contiguous proline residues (Figs. 12.1b, c and 12.2a) (Higgs and Peterson, 2005; Rivero et al., 2005). For example, the fission yeast formin Fus1 has a single proline-rich track, whereas the nematode worm formin CYK-1 has six and the mouse formin mDia1 has fourteen. These proline-rich tracks are thought to form rigid type-II polyproline helices, whereas the sequences between the proline-rich tracks are not well conserved and expected to be flexible. The FH1 domain binds to profilin with low micromolar affinity (Evangelista et al., 1997; Chang et al., 1997; Watanabe et al., 1997; Neidt et al., 2009; Kursula et al., 2008), on the opposite side of profilin’s actin-binding site (Kursula et al., 2008). Therefore, profilin can bind simultaneously to actin and the formin FH1 domain (Fig. 12.3a). Profilin stimulates formin-mediated actin assembly (Sagot et al., 2002a; Li and Higgs, 2003; Kovar et al., 2003), and mutations in both the poly- L-proline binding and actin-binding surface of profilin diminish utilization by formin (Pruyne et al., 2002; Li and Higgs, 2003; Pring et al., 2003; Kovar et al., 2003; Moseley et al., 2004; Kovar et al., 2006). A confusing initial finding was that low concentrations of profilin stimulate formin-mediated actin assembly, whereas higher concentrations of profilin inhibit formin-mediated actin assembly (Sagot et al., 2002a; Kovar et al., 2003). The biphasic dependence of the “bulk” actin assembly rate on profilin is because, (A) profilin has opposite effects on the nucleation efficiency of formin and the elongation rate of formin-associated barbed ends (Kovar et al., 2003), and (B) excess free profilin competes with profilin-actin for the formin FH1 domain (Kovar et al., 2006; Vavylonis et al., 2006). Profilin significantly reduces the nucleation efficiency of formins by ∼10-fold. Fission yeast Cdc12 produces a new filament per ∼2.5 and ∼25 FH1FH2 dimers in the absence and presence of profilin (Kovar et al., 2003; Neidt et al., 2008). Nematode worm CYK-1 produces a new filament per ∼25 and ∼225 FH1FH2 dimers in the absence and presence of profilin (Neidt et al., 2008). The reduced nucleation efficiency is likely because formins cannot nucleate profilin-actin. Even in the presence of high profilin concentrations, it is likely that formin nucleates free actin dimers following the dissociation of profilin (Paul and Pollard, 2008). However, profilin dramatically increases the elongation rate of formin-associated barbed ends (Neidt et al., 2008; Kovar et al., 2006; Kovar and Pollard, 2004; Romero et al., 2004; Paul and Pollard, 2008; Vidali et al., 2009). Therefore, the net result of profilin on formin-mediated actin assembly is the production of fewer filaments that elongate significantly faster. In the absence of formin, barbed ends elongate at ∼10 μM−1 s−1 (Pollard, 1986). Formin FH1FH2 constructs are able to elongate profilin-actin at rates as high as 10-fold faster. However, the specific increase in elongation rate varies widely depending upon the particular formin (Table 12.1). At two extremes are fission yeast

12

Formin-Mediated Actin Assembly

297

Cdc12 (12.5 μM−1 s−1 ) and moss For2A (125 μM−1 s−1 ) (Kovar et al., 2003; Vidali et al., 2009). Given such a wide range of elongation rates, it has been hypothesized that in addition to regulation (activating a formin at the right time and place) the specific barbed end elongation rate is important for cellular functions (Neidt et al., 2008; Neidt et al., 2009). In agreement, it was recently discovered that a fast elongating formin can support polarized cellular growth in moss, but a slow elongating formin cannot (Vidali et al., 2009). Actin assembly is limited by the diffusion rate of monomers, and a high percentage of monomers are not oriented correctly upon colliding with the barbed end (Drenckhahn and Pollard, 1986). Formins enhance the elongation rate of profilinactin through a combination of increasing the local actin concentration and steering correctly-oriented profilin-actin monomers onto the barbed end (Fig. 12.1b, c) (Kovar et al., 2006; Romero et al., 2004; Paul and Pollard, 2008; Vavylonis et al., 2006). Based upon experimental studies and theoretical considerations, the direct transfer mechanism has been proposed to explain formin-mediated elongation of profilin-actin (Fig. 12.1b, c) (Kovar, 2006; Neidt et al., 2009; Paul and Pollard, 2008; Vavylonis et al., 2006). Multiple profilin-actins bind to the various prolinerich stretches in the formin FH1 domain. Flexibility allows FH1-bound profilin-actin to interact with and bind to the FH2 domain-associated barbed end. This has been described as the “ring complex” composed of FH2::barbed-end::profilin-actin::FH1. Following addition of the actin monomer to the filament, profilin and the FH1 domain must dissociate prior to additional rounds of assembly. Consistent with the “direct transfer hypothesis,” isolated FH1 and FH2 domains cannot cooperate to stimulate rapid assembly of profilin-actin in trans (as part of different molecules) (Neidt et al., 2009), and simulations of the direct transfer reactions are consistent with experimental data of four diverse formins (Paul and Pollard, 2008). Profilin-actin that is not bound to the FH1 domain can add directly to an FH2-associated barbed end at only approximately one-third the rate of actin alone (Fig. 12.1b) (Kovar, 2006). Therefore, the FH1-dependent pathway vastly exceeds the rate of the FH1-independent pathway (Fig. 12.1b) (Paul and Pollard, 2008; Vavylonis et al., 2006). As predicted, the elongation rate increases with the total number of proline-rich profilin-binding sites, and proline-rich sites further from the elongating FH2-associated barbed end contribute less (Fig. 12.1b) (Paul and Pollard, 2008). The emerging model is that the FH1 and FH2 domains work in concert to rapidly assemble unbranched actin filaments. The role of the FH2 domain is to nucleate the assembly of free actin dimers dissociated from profilin (Otomo et al., 2005; Pring et al., 2003; Paul and Pollard, 2008), or bind with high affinity to pre-assembled filaments (Neidt, 2008), and then to remain processively associated with the elongating barbed end (Otomo et al., 2005). The role of the FH1 domain is to bind profilin-actin and transfer the actin monomer to the FH2-associated barbed end at exceedingly fast rates (Kovar et al., 2006; Romero et al., 2004; Vavylonis et al., 2006). Given that many cells express multiple isoforms of formin and profilin, it is possible that particular pairs of formin and profilin are optimally matched to drive rapid actin assembly. It was recently discovered that particular formin isoforms do prefer

298

D.R. Kovar et al.

actin bound to specific profilin isoforms (Neidt et al., 2009; Ezezika et al., 2009). The nematode worm cytokinesis formin CYK-1 assembles actin bound to the three worm profilin isoforms CePFN-1 (∼60 μM−1 s−1 ), CePFN-2 (∼40 μM−1 s−1 ) and CePFN-3 (∼15 μM−1 s−1 ) at significantly different rates (Neidt et al., 2009). It is possible that these differences are physiologically relevant given that CePFN-1 is the only profilin required for cytokinesis in the nematode worm early embryo (Severson et al., 2002). A high formin-mediated elongation rate is critical for cell polarization in the moss Physcomitrella patens (Vidali et al., 2009), and it is becoming apparent that profilin isoforms expressed in the same cell can have different functions (Khadka et al., 2009). It might be assumed that the molecular basis for formin-profilin isoform selectivity is entirely due to the FH1 domain binding with higher affinity to particular profilin isoforms. However, by analyzing FH1 and FH2 domain chimeras between formins that prefer different profilins, it was discovered that both the FH1 domain and the barbed end-associated FH2 domain contribute to profilin specificity (Neidt et al., 2009). Given that isolated FH1 domains bind with similar low micromolar affinity to both “preferred” and “less-preferred” profilin isoforms, the role of the FH1 domain may be to bind specific profilin isoforms more tightly when bound to actin (Neidt et al., 2009). How the proline-rich stretches of the FH1 domain have a higher affinity for profilin bound to actin than to free profilin is not clear. A prolinerich peptide of Ena/VASP has ∼10-fold higher affinity for profilin-actin than for profilin without inducing a major conformational change in profilin-actin (Ferron et al., 2007). It has been suggested that an increased affinity of the formin FH1 domain for profilin-actin helps ensure that non-productive associations with free profilin do not occur (Dominguez, 2009). A role for the FH2 domain in selecting particular profilins was not anticipated, but suggests that during transfer of FH1bound profilin-actin to the elongating barbed end there is an important interaction between profilin or profilin-actin and the FH2 domain (Fig. 12.1c) (Neidt et al., 2009). In addition to FH1- and actin-binding sites, it is possible that profilins have a region important for interacting with the formin FH2 domain. It has also been argued that the FH1 domain and profilin are essential in processive association of formin with the elongating barbed end (Romero et al., 2004, 2007). The simple view that the FH1 and FH2 domains have complimentary but independent roles in facilitating rapid processive barbed end elongation is not the complete story.

12.3.7 Energy for Formin-Mediated Processive Elongation The energy source for formin-mediated processive elongation has been the focus of several investigations, and has led to two competing ideas that the energy comes from either profilin-promoted ATP hydrolysis and/or phosphate release or the free energy released upon actin monomer addition to the filament (Kovar et al., 2006; Romero et al., 2004, 2007; Paul and Pollard, 2009). The ATP hydrolysis model was proposed by Romero and colleagues (Romero et al., 2004), and is appealing because it simultaneously explains both the source of energy and how profilin

12

Formin-Mediated Actin Assembly

299

is removed from the barbed end. After profilin-actin associates with the barbed end, either attached to the formin FH1 domain or free in solution, profilin must dissociate to allow subsequent profilin-actin monomers to add. In the ATP hydrolysis model profilin dissociates because the ultimate actin subunit hydrolyzes ATP, and/or increases the rate of γ-phosphate release, and profilin has a lower affinity for ADP-Pi-actin and/or ADP-actin than ATP-actin (Pantaloni and Carlier, 1993). The energy from ATP hydrolysis can then support processive elongation (Romero et al., 2004; Romero et al., 2007). In agreement with this model, it was reported that the mouse formin Dia1 (A) requires both FH1 and profilin for processive elongation, (B) ATP-hydrolysis keeps up with the rate of FH1FH2-mediated assembly in the presence of profilin and (C) ADP-actin and non-hydrolyzable ATP-actin analogs inhibit formin-mediated assembly (Romero et al., 2004; Romero et al., 2007). Alternatively, it has been proposed that the free energy released from actin monomer addition can support processive elongation (Kovar et al., 2006; Paul and Pollard, 2009b). In support of this model, it has been found that the mouse formins Dia1 and Dia2, fission yeast Cdc12 and budding yeast Bni1 (A) do not require profilin for processive elongation, and (B) processively elongate ADP-actin (Kovar et al., 2006). Furthermore, in both the presence and absence of budding yeast Bni1, it has been reported that phosphate release lags well behind actin assembly with profilin (Paul and Pollard, 2009b; Blanchoin and Pollard, 2002). Discerning between these models is hampered by the difficulty in experimentally measuring ATP-hydrolysis. ATP-hydrolysis may not be required per se, but contributes to the faster rate of processive elongation in the presence of profilin. Another possibility is that ATP-hydrolysis and phosphate release deeper into the filament contribute to profilin dissociation and processive elongation (Yarmola and Bubb, 2009).

12.3.8 Interaction of the Formin FH2 Domain with the Side of Actin Filaments Almost all formins nucleate actin filament assembly and remain processively associated with elongating barbed end, and facilitate the rapid addition of profilin-actin. Additionally, some formin FH2 domains can bind to the side of actin filaments and induce bundling, severing and/or depolymerization (Vaillant et al., 2008; Chhabra and Higgs, 2006; Harris and Higgs, 2004; Michelot et al., 2005; Esue et al., 2008; Harris et al., 2006; Michelot et al., 2006; Moseley and Goode 2005). The physiological relevance of these activities has not been demonstrated, although it should be noted that most formin-dependent cellular structures such as the contractile ring and filopodia are composed of bundled actin filaments. A particularly interesting case with respect to additional actin-associated properties is the mouse formin INF2 (Chhabra and Higgs, 2006). INF2 stimulates actin assembly like other formins, but also contains an N-terminal WASP homology 2 WH2 domain that helps facilitate actin filament depolymerization and severing. INF2 localizes to the cytoplasmic face

300

D.R. Kovar et al.

of the endoplasmic reticulum (Chhabra et al., 2009). The cellular function of INF2 is not known, but given its localization it might be involved in ER function where a formin that produces filaments that rapidly turnover might be important (Chhabra et al., 2009).

12.3.9 Effect of other Actin-Binding Proteins on Formin-Mediated Actin Assembly In addition to profilin, other actin-binding proteins such as Bud6, Spire and tropomyosin also effect formin-mediated actin assembly. In both budding and fission yeasts, the polarity factor Bud6/Aip3 interacts with and positively regulates formin-mediated actin assembly for polarized cell growth (Martin et al., 2007; Kovar et al., 2004; Moseley and Goode, 2005). Of the two budding yeast formins Bni1 and Bnr1, which have overlapping roles in establishing cell polarity (Evangelista et al., 2002; Sagot et al., 2002a), Bud6 specifically interacts with and stimulates the actin assembly properties of Bni1 (Kovar et al., 2004; Moseley and Goode, 2005). In this case, Bud6 binds to the DAD region immediately downstream of the FH2 domain and is thought to behave functionally similar to profilin by binding actin monomers and “transferring” them to the FH2-domain associated barbed end (Kovar et al., 2004; Moseley and Goode, 2005). In fission yeast, cell biological studies are consistent with a model whereby Bud6 binds directly to the polarity formin For3 near its C-terminal “DAD-like” domain, and promotes For3mediated actin assembly by correctly localizing For3 and relieving autoinhibition (Martin et al., 2007). These different models for the role for Bud6 in promoting formin-mediated actin assembly for polarized cell growth in yeast are not mutually exclusive. It will be interesting to determine whether purified fission yeast Bud6 can stimulate the actin assembly properties of For3 in vitro similarly to budding yeast Bud6 and Bni1. There is an established connection between the actin nucleation factor Spire and formin, including their co-expression in neuronal tissues and synergistic involvement in establishing polarity in Drosophila oocytes (Schumacher et al., 2004; Quinlan et al., 2007), but the consequence of this interaction is disputed (Quinlan and Kerkhoff, 2009; Renault et al., 2008). In one scenario, Spire and fly formin Cappuccino directly bind and the interaction is thought to inhibit formin while stimulating the actin assembly properties of Spire (Quinlan et al., 2007; Pechlivanis et al., 2009). Alternatively, Spire has been proposed to indirectly stimulate formin-mediated actin assembly by increasing the concentration of polymerization-competent profilin-actin by binding to the actin filament barbed end and inhibiting profilin-actin addition (Bosch et al., 2007). Interestingly, the actin filament stabilizing protein tropomyosin typically binds to cellular structures composed of actin filaments assembled by formin (Gunning et al., 2005, 2008). For example, tropomyosin is absolutely required for contractile ring assembly in fission yeast (Balasubramanian et al., 1992). Tropomyosin inhibits

12

Formin-Mediated Actin Assembly

301

nucleation by the Arp2/3 complex in a competitive manner (Bugyi et al., 2009; Blanchoin et al., 2001), preferentially binds to Arp2/3 complex nucleated actin filaments following debranching (Bugyi et al., 2009) and stabilizes the filaments from cofilin- and gelsolin-mediated severing (Cooper, 2002). Tropomyosin might therefore be important for synergizing with formin to make stable unbranched filaments. Two recent studies found that tropomyosin has diverse effects on formin-mediated actin assembly in vitro (Skau et al., 2009; Wawro et al., 2007). The fission yeast cytokinesis formin Cdc12 may recruit tropomyosin to actin filaments by increasing the affinity of tropomyosin for ADP-loaded actin filaments (Skau et al., 2009). Once bound to formin-nucleated filaments, both vertebrate and fission yeast tropomyosin enhance their elongation rate (Skau et al., 2009; Wawro et al., 2007). Fission yeast tropomyosin also increases the length of formin nucleated filaments by allowing them to anneal end-to-end (Skau et al., 2009). On the other hand, fission yeast tropomyosin ultimately “turns off” formin-mediated actin assembly by displacing formin from the elongating barbed end or by “trapping” formin in annealed segments (Skau et al., 2009). How proteins that interact with different parts of the actin filament, for example formin on the barbed end and tropomyosin on the side of the filament, influence each other’s properties is an intriguing question. It is possible that formins influence the actin filament conformation by remaining processively associated with the barbed end. This hypothesis is supported by recent findings that mDia1 formin increases the flexibility of actin filaments (Bugyi et al., 2006; Papp et al., 2006; Kupi et al., 2009; Ujfalusi et al., 2009a). Conversely, tropomyosin binding to “flexible” forminassociated filaments seems to stabilize the filament (Ujfalusi et al., 2009b). Filament stabilization might alter formin-associated barbed end properties such as the elongation rate and annealing. Alterations in filament properties induced by different actin-binding proteins might be a general mechanism to regulate which actinbinding proteins bind to specific actin-based structures and collectively influence actin architecture and dynamics.

12.4 Cellular Roles of Formins A prevalent model is that eukaryotes express multiple formins because each assembles actin for different cellular processes (Fig. 12.4). Given that formin isoforms vary considerably in their (A) mode and mechanism of regulation, (B) actin assembly properties and rate constants and (C) ability to interact with the microtubule cytoskeleton and/or regulate signaling pathways, it would not be surprising if formins are tailored for particular cellular roles. The best example of isoform specialization is in fission yeast where each of the three isoforms has a distinct cellular role (Fig. 12.4b): Cdc12 assembles actin filaments for cell division (Fig. 12.5) (Chang et al., 1997), For3 assembles actin cables for polarized growth (Fig. 12.6) (Feierbach and Chang, 2001; Nakano et al., 2002), and Fus1 is required for conjugation during mating (Petersen et al., 1995, 1998). Examination of the 15 mammalian

302

D.R. Kovar et al.

A

B

division

division Formin Cdc12 contractile ring

Formin I

polarization Formin For3

contractile ring lamellipodium

Arp2/3 complex

actin cables

chemotaxis

stress fibers & adhesions

Formin III

mating Formin Fus1 h− h+

filopodia

Formin II

shmoo

endocytosis Arp2/3 Complex

motility

actin patches

Fig. 12.4 Multiple formin isoforms are required for diverse cellular roles. (a) Simplified schematic representation of dividing and crawling animal cells, indicating that specific formin isoforms are required for different cellular processes including division, filopodia assembly and stress fibers/adhesion. Both the regulatory (localization and activation) and actin assembly (nucleation efficiency, elongation rate, dissociation rate, bundling, severing) properties may tailor formins for their particular role. (b) Fission yeast S. pombe contains a simplified actin cytoskeleton. It has been established that four nucleation factors are specifically required for a different cellular processes. Three formin isoforms are required for division (Cdc12, (Chang, et al., 1997)), polarity (For3, (Feierbach and Chang, 2001; Nakano et al., 2002)) and mating (Fus1, (Petersen et al., 1998)). The Arp2/3 complex is required for endocytic actin patches (Balasubramanian et al., 1996; McCollum et al., 1996)

formins has revealed that particular isoforms are responsible for an ever-growing, diverse set of cellular roles (Fig. 12.4a). Determination of the cellular roles of formins might be greatly facilitated by the recent discovery of “general” and “isoform specific” small molecule inhibitors of formin (Gauvin et al., 2009; Rizvi et al., 2009), however the usefulness of these drugs has not been rigorously tested in cells. In this section we summarize several of formins’ cellular roles including cytokinesis, cell morphogenesis, cell motility, cell adhesion/stress fiber formation and polarity.

12.4.1 Cytokinesis The final step of cell division is cytokinesis, the physical separation of a mother cell into two daughter cells by a contractile ring of anti-parallel actin filaments (Glotzer, 2005; Eggert et al., 2006). Formin was first linked to cytokinesis when mutants of

12

Formin-Mediated Actin Assembly

303

P

Actin Filaments

B

Formin B B

P

P

P

Myosin-II B

Fig. 12.5 Formin-mediated contractile ring assembly in fission yeast. The fission yeast contractile ring is composed of anti-parallel actin filaments, which are assembled from ∼60 pre-ring nodes that include at least seven proteins including the formin Cdc12 and the motor myosin II (Coffman et al., 2009; Wu et al., 2003, 2006). Cdc12-mediated nucleation and processive barbed end elongation coupled with myosin II-mediated actin filament pulling links pre-ring nodes and drives their coalescence into a mature contractile ring consisting of anti-parallel filaments (Kamasaki et al., 2007; Vavylonis et al., 2008). Barbed (B) and pointed (P) ends are indicated

the Drosophila formin Diaphanous showed a failure in cell division during spermatogenesis (Castrillon and Wasserman, 1994). A large number of studies from yeast to animals, have subsequently shown that formins are essential for assembling and maintaining the contractile ring in evolutionarily diverse organisms (Imamura et al., 1997; Chang et al., 1997; Harris et al., 1997; Swan et al., 1998; Severson et al., 2002; Peng et al., 2007; Tolliday et al., 2002). The mechanism of formin-mediated contractile ring assembly is best understood in fission yeast (Fig. 12.5), where two models have been proposed. In the first model, the fission yeast cytokinesis formin Cdc12 is localized to a large progenitor spot from which a “leading cable” of actin filaments curves around the cell to guide assembly of the contractile ring (Kamasaki et al., 2007; Chang et al., 1997; Yonetani et al., 2008; Arai and Mabuchi, 2002; Chang, 1999). In this scenario it is difficult to appreciate how filaments ultimately become arranged in an anti-parallel fashion. More recently it has been proposed that the fission yeast contractile ring assembles from ∼60 pre-ring nodes (Fig. 12.5), which are composed of at least seven proteins including the actin filament motor protein myosin II (Balasubramanian et al., 1998; Kitayama et al., 1997; May et al., 1997), and the formin Cdc12 (Coffman et al., 2009). The coordinated activities of formin-mediated actin assembly with myosin IImediated actin filament pulling connects pre-ring nodes and drives their coalescence into a mature contractile ring (Vavylonis et al., 2008; Coffman et al., 2009; Wu et al., 2003, 2006). Pre-ring node associated Cdc12 nucleates the assembly of a filament

304

D.R. Kovar et al.

barbed

actin cables

pointed

barbed

nucleus microtubules

plus

minus

plus

Inactive For3 Active For3 Partially Active For3

a b

+TIP

microtubule a b

polarisome complex

+TIP a

1 2

b

a

Actin filament elongation

6

b

3

b

4

a

a b

5

Fig. 12.6 Formin-mediated actin cable assembly in fission yeast. An interphase fission yeast cell contains ∼four dynamic microtubules organized into anti-parallel bundles with plus ends facing and periodically interacting with the cell tip. Microtubules maintain the nucleus at a central position and contribute to polarity establishment by delivering polarity factors to the cell tip (Piel and Tran, 2009). Actin cables provide polar tracks for myosin motors to deliver vesicles and organelles towards the bud tip of budding yeast cells and the ends of fission yeast cells (Bretscher, 2003). Actin cables are composed of bundles of short actin filaments assembled by the formin For3 (Kamasaki et al., 2005; Feierbach and Chang, 2001; Nakano et al., 2002; Karpova et al., 1998). (1) For3 diffuses to the cell tip in an inactive state (For3 a and b). (2) The Tea1-Tea4 complex (+TIP) is deposited by microtubules at cell ends where it recruits the polarisome complex, which includes For3 and activators Rho-GTPase Cdc42 and Bud6. (Pollard and Borisy, 2003) Following activation by the Cdc42 and the polarity factors Bud6 and Pob1 (Martin et al., 2007; Rincon et al., 2009), For3 transiently mediates processive actin filament assembly (For3 a). (4) After a few seconds, For3 is partially inactivated and releases from the cortex along with its associated short filament (For3 a; remains bound to filaments but does not mediate further assembly). (4 and 5) Activation and transient processive actin filament assembly by a neighboring For3 (b) pushes partially active For3 molecules and associated filaments inward as the cable grows. (6) Distal to the tip, actin cables are disassembled and inactive formin dimers and actin monomers are recycled back to the tip

whose pointed end is pushed away by processive barbed end elongation. Myosin II on an adjacent node captures the filament, and “walking” towards the forminassociated barbed end pulls the pre-ring nodes together (Vavylonis et al., 2008). It is unclear whether animal cells utilize a similar mechanism, although myosin II is

12

Formin-Mediated Actin Assembly

305

found in discrete foci in nematode worm early embryos that appear to condense into the cleavage furrow (Munro et al., 2004; Werner et al., 2007). Determining which formin(s) is required for cytokinesis in mammalian cells has been difficult, and the underlying mechanisms of contractile ring assembly are not known. Recent studies have identified mDia2 as the mammalian formin predominantly responsible for cytokinesis. Upon knockdown of each of the three DRFs in NIH 3T3 fibroblast cells, only mDia2 disruption caused cytokinesis defects (Watanabe et al., 2008). mDia2 localizes to the cleavage furrow and intercellular bridge during cytokinesis. Knockdown of mDia2 in erythroblasts blocked contractile ring formation and erythroblast enucleation, while knockdown of mDia1 did not (Ji et al., 2008).

12.4.2 Cell/Tissue Morphogenesis Recently a considerable number of studies have demonstrated that formins play a critical role in morphogenesis and tissue development. Most work on morphogenesis has concentrated on the Daam1 formin, which is required for gastrulation in Xenopus and development of the tracheal cuticle in Drosophila (Habas et al., 2001; Matusek et al., 2006). Daam1 is a downstream ligand of Disheveled and part of the noncanonical Wnt signaling cascade (Habas et al., 2001). Daam1 also has multiple roles in early Xenopus embryos, through differential interactions with profilin 1 and profilin 2 (Khadka et al., 2009; Sato et al., 2006). Profilin 1 interacts with Daam1 and controls blastomere regulation, whereas profilin 2 is required for convergent extension movement during gastrulation. Similarly, in zebrafish formin zDia2 and profilin 1, but not profilin 2, are required for gastrulation in early embryos (Lai et al., 2008). These studies suggest that formins can be tailored for various cellular roles by interacting with different profilins or other binding partners. Formins have also been found to play a role in neuronal development. Drosophila Daam localizes to neurites and growth cones of developing neurons and loss of Daam leads to neuronal defects (Matusek et al., 2008). Interestingly, this defect was partially rescued by expression of mammalian Daam, demonstrating that the role of Daam in neuron development is conserved from flies to mammals.

12.4.3 Filopodia/Cell Motility Cell motility requires rapid remodeling of the actin cytoskeleton to form lamellipodia and filopodia at the leading edge (Fig. 12.4a) (Pollard and Borisy, 2003; Carlier et al., 2003). Filopodia are finger-like membrane protrusions comprised of bundles of linear actin filaments (Mallavarapu and Mitchison, 1999). The involvement of formin in filopodia assembly was first detected in Dictyostelium, where dDia2 localizes to the filopodial tips and is required for filopodia formation (Schirenbeck et al., 2005a, b). Formin is thought to induce filopodia formation by allowing rapid elongation of unbranched bundled filaments that are protected from capping protein.

306

D.R. Kovar et al.

The vasodilator-stimulated phosphoprotein (VASP) can also induce filopodia formation through the rapid quasi-processive assembly of unbranched long filaments (Svitkina et al., 2003; Bear et al., 2002; Breitsprecher et al., 2008). In addition to filopodia assembly, the mouse formin mDia2 was also found to have a critical role in lamellipodia formation where the Arp2/3 complex was thought to be the primary actin nucleation factor (Yang et al., 2007). This suggests that longer unbranched filaments are also important in the establishment and maintenance of lamellipodia, and formin-dependent lamellipodia filaments may be the precursors for filopodia formation. In diverse cell types different formin isoforms are important for cell motility. The cell motility of natural killer cells is disrupted by knockdown of human Dia1 but not Dia2 or FHOD1 (Butler and Cooper, 2009). Dia1 also is required for RAGE ligand-stimulated cell migration and activation of Rac1 and Cdc42 (Hudson et al., 2008). Formins are also utilized for movement by cellular invaders. The malaria parasite Plasmodium does not encode the Arp2/3 complex, but does express two formin isoforms. Formin 1 (PfFormin1) stimulates actin assembly and localizes to the moving tight junction between the parasite and host cell, the site of predicted movement for the parasite (Baum et al., 2008). Additionally, mammalian formins Dia1 and Dia2 are required for the pathogenic bacteria Shigella to spread in HeLa and PtK2 cells (Heindl et al., 2009).

12.4.4 Stress Fibers/Cell Adhesion Stress fibers are contractile bundles composed of anti-parallel linear actin filaments and myosin II motor, and are essential for cell adhesion and retraction during cell migration (Cramer et al., 1997). In cell culture, both Rho activated mDia1 and active FHOD fragments stimulate stress fiber assembly (Nakano et al., 1999; Kanaya et al., 2005). In addition to formin, Rho can also induce stress fiber formation through activation of Rho kinase I (ROCK I). Rho kinase I (ROCK I) activity stimulates stress fiber formation, whereas ROCK I down-regulation inhibits stress fiber formation (Leung et al., 1996; Uehata et al., 1997). The formin FHOD appears to be a downstream target of ROCK I that is responsible for inducing stress fiber assembly. Rho-dependent ROCK I phosphorylates three C-terminus residues on FHOD1. This phosphorylation disrupts FHOD autoinhibition and leads to stress fiber formation (Takeya et al., 2008). Formins are also important for cell-cell adhesions. Formin1 is recruited to adhesion junctions and assembles radial actin cables required for cell-cell adhesions (Dettenhofer et al., 2008). The role of Dia1 in cell-cell adhesion has been recently investigated in epithelial cells. Knockdown of Dia1 disrupts adherens junctions and decreased localization of E-cadherin to cell-cell contacts (Carramusa et al., 2007). GFP-Dia1 localizes to cell-cell contacts and over expression of Dia1 increases both adhesion zone width and the accumulation of E-cadherin and actin at junctions. More recently, Dia1 was found to maintain cell-cell junctions through recruitment by Abi1 (Ryu et al., 2009).

12

Formin-Mediated Actin Assembly

307

12.4.5 Polarity The mechanisms by which formins mediate cell polarity have been extensively studied in budding and fission yeast (Fig. 12.6). Yeast actin cables are composed of bundles of short parallel filaments (Kamasaki et al., 2005; Karpova et al., 1998), whose barbed ends are oriented towards the cell cortex to facilitate myosin-directed delivery of vesicles and organelles to the expanding cell tip (Bretscher, 2003). The budding yeast formins Bni1 and Bnr1, and the fission yeast formin For3, are absolutely required for the assembly of polarized actin cables (Evangelista et al., 2002a; Sagot et al., 2002a; Feierbach and Chang, 2001; Nakano et al., 2002). Genetic, cellular and theoretical considerations have led to a working model for For3-mediated actin cable assembly in fission yeast (Fig. 12.6) (Martin et al., 2007; Martin and Chang, 2006; Wang and Vavylonis, 2008). Polarity is initially established by microtubule plus ends, which interact with the cell tip and deposit factors required to recruit and activate the formin For3 (Martin et al., 2005; Terenna et al., 2008; Martin, 2009; Minc et al., 2009). At the cell tip, For3 is activated by the RhoGTPase Cdc42 and polarity factors Bud6 and Pob1, where it initiates processive actin assembly (Martin et al., 2007; Rincon et al., 2009). For3 fused to three copies of GFP (For3-3xGFP) was found to transiently associate with the cell cortex for only a few seconds and then move inwards with the elongating actin cable at the same rate as actin assembly (Martin et al., 2007). Therefore For3 may initiate actin filament assembly for only a few seconds before being “turned off,” released from the cortex and carried into the cell interior by subsequent actin filament assembly mediated by additional active For3 molecules back at the cortex. It is likely that a similar mechanism occurs in budding yeast. During bud formation, Bni1 localizes to the bud tip and Bnr1 to the bud neck and promote assembly of polarized actin cables (Ozaki-Kuroda et al., 2001; Yang and Pon, 2002). A set of four Rho-GTPases differentially regulate Bni1 and Bnr1 (Dong et al., 2003), and Bni1-3xGFP exhibits a similar transient cortical localization as fission yeast For3, as well as detachment and retrograde flow on elongating actin cables (Buttery et al., 2007). Formins also establish polarity in mammals and plants. In cardiac muscle, knockdown of FHOD3 led to rounded cardiac myocytes and loss of the normal striated actin/myosin pattern (Taniguchi et al., 2009). In the moss, Physcomitrella patens, silencing of both For2 isoforms led to small, round cells with no polarized extensions and large-scale disruption of F-actin organization (Vidali et al., 2009).

12.4.6 Regulation of Microtubule Dynamics by Formins Formins have also been shown to regulate microtubules and microtubule-dependent processes, which has been extensively reviewed (Bartolini and Gundersen, 2009). Mutations in a range of diverse formins result in defects in microtubules, including the orientation of interphase and spindle microtubules during motility and cell division (Bartolini and Gundersen, 2009). Mouse formin mDia3 localizes to the kinetochore and is required for bipolar microtubule attachment and correct

308

D.R. Kovar et al.

chromosome segregation (Yasuda et al., 2004), and mouse formin mDia1 localizes to the mitotic spindle and regulates spindle position (Ishizaki et al., 2001; Kato et al., 2001). Loss of mDia1 or FMNL1 in activated T-cells does not affect actin structures but does disrupt polarization of the microtubule organizing center/centrosome (Gomez et al., 2007). Whether interactions between formin and microtubules in a dividing cell are direct or an indirect consequence of formins role in regulating the actin cytoskeleton is not known. It is possible that their role in microtubulebased processes is direct because formins have been found to regulate microtubule stability in both cells and in vitro. Active mDia1 and mDia2 associate with stable microtubules through a Rho GTPase-dependent pathway (Palazzo et al., 2001). One proposed mechanism is that formin serves as a scaffold for the plus-end tracking proteins, EB1 and APC, and that the interactions among these components are required for formin-induced microtubule stabilization (Wen et al., 2004). However, a recent study demonstrated that a mouse formin mDia2 mutant construct, that cannot stimulate actin assembly, was able to directly stabilize microtubules in vitro and bind to EB1 and APC (Bartolini et al., 2008). Therefore, the actin stimulating and microtubule stabilizing activities of mDia2 are separable functions.

12.4.7 Formins and Disease Given that formin is required for numerous fundamental cellular processes, it is not surprising that formin defects would be associated with diseases. However, directly linking formins to disease pathogenesis has been difficult (for a comprehensive review see (DeWard et al., 2009)). Genetic studies of formin have provided the most insight into formin-related diseases. Knockout of mDia1 in mice led to development of an age-dependent myeloproliferative disorder (Peng et al., 2007). When the knockout was expanded to both mDia1 and RhoB, the mice showed severe myelodysplasia, which is usually a precursor to acute myelogenous leukemia (DeWard et al., 2009). Recent studies have revealed a role for formins in metastatic tumor formation. Formin-like 2 (FMNL2) expression was found to be elevated in metastatic colorectal cancer cell lines when compared to non-metastatic colorectal cancer cell lines (Zhu et al., 2008). Depletion of mDia2 in MDA-MB-231 breast cancer cells reduced their invasive ability (Lizarraga et al., 2009). Formins have also been linked to non-cancerous diseases. A progressive deafness disorder in humans called DFNA1 has been mapped to a frameshift mutation in the DIAPH1 gene (encodes hDia1) (Lynch et al., 1997). A breakpoint in the X chromosome that is associated with premature ovarian failure, was mapped to a disruption in the DIAPH2 gene (encodes mDia3) (Bione et al., 1998).

12.5 Concluding Remarks Since the discovery that formins are auto-regulated proteins that stimulate actin filament assembly, remarkable progress has been made in determining the underlying mechanisms. It has been established that the FH2 domain nucleates

12

Formin-Mediated Actin Assembly

309

actin assembly and remains processively associated with the elongating barbed end while the FH1 domain drives the rapid addition of profilin-actin. It has also been established that many formins are auto-inhibited through direct association of their N- and C-terminal regulatory regions and are subsequently activated, at least partially, by Rho-GTPases. However, we are very far from fully understanding the biochemical and biophysical mechanisms that drive processive and rapid elongation of profilin-actin by formins. Excellent progress has been made studying formins domain-by-domain, but it is particularly important to determine the actin assembly and regulatory properties of a wide range of full-length formin proteins. Full-length formin proteins might have significantly different properties than isolated domain fragments. Given the large number of formin isoforms, it will be extremely interesting to elucidate the specific role of each formin and the mechanistic basis for their involvement in a particular process. There are certain to be many additional surprises in the number of diverse mechanisms by which formins are regulated and localized to particular cellular locations. Given the large range of actin assembly properties exhibited by diverse formins it will also be important to determine to what extent, if any, the specific actin assembly properties are functionally important. Diverse cell types and particular cellular processes vary considerably in physical parameters, which might require specialized actin assembly properties. We have just begun to explore formin biology in numerous areas. For example, we have been so preoccupied with understanding how formins stimulate actin assembly that we hardly understand the mechanisms by which formin-mediated actin assembly is “turned-off.” Filament length control is extremely important. In vitro formins dissociate slowly from the elongating barbed end, resulting in filaments that are hundreds of microns in length. In vivo many formin-dependent actin structures, such as the contractile ring and actin cables in yeast, are typically composed of filaments as short as a micrometer long. Actin cables in yeast are assembled by formins that are only transiently active before detaching from the cortex and flowing with actin assembly into the cell. Understanding crosstalk between formin-mediated actin assembly and other actin nucleation factors such as Spire and Arp2/3 complex has only just begun. In the simple yeast system, each actin-based structure may be exclusively dependent upon a specific nucleation factor. However, in more complex animal cells it is becoming increasingly evident that actin nucleation factors work in concert to produce actin filaments for various cellular processes such as lamellipodia and filopodia. Furthermore, numerous types of actin-binding proteins with diverse biochemical properties are present in the complex cellular environment. Yet, very specific subsets of actin-binding proteins associate with different actin-dependent structures. How particular populations of actin-binding proteins specifically sort to formindependent or Arp2/3 complex-dependent actin filaments and collectively influence actin filament dynamics and architecture is poorly understood. In vitro reconstitution of actin-based processes in real time with single molecule imaging, coupled with cellular and genetic analysis will be critical to explore the collective actions of multiple actin-binding proteins. Finally, elucidating the roles that formins play in crosstalk between actin and microtubules is still in its infancy. We have little idea

310

D.R. Kovar et al.

of the functional consequences of formin’s involvement in microtubule dynamics or the effect that the microtubule cytoskeleton has on formin-mediated actin assembly. Acknowledgments We thank Erin Neidt for helpful suggestions. Work on formins in the Kovar lab is funded by a National Institutes of Health grant RO1GM079265.

References Ahuja, R., Pinyol, R., Reichenbach, N., Custer, L., Klingensmith, J., Kesseles, M. M., and Qualmann, B. (2007) Cell 131, 337–350 Alberts, A. S. (2001) J Biol Chem 276(4), 2824–2830 Arai, R., and Mabuchi, I. (2002) J Cell Sci 115, 887–898 Balasubramanian, M. K., Feoktistova, A., McCollum, D., and Gould, K. L. (1996) Embo J 15(23), 6426–6437 Balasubramanian, M. K., Helfman, D. M., and Hemmingsen, S. M. (1992) Nature 360(6399), 84–87 Balasubramanian, M. K., McCollum, D., Chang, L., Wong, K. C., Naqvi, N. I., He, X., Sazer, S., and Gould, K. L. (1998) Genetics 149(3), 1265–1275 Bartolini, F., and Gundersen, G. G. (2009) Biochem Biophys Acta Bartolini, F., Moseley, J. B., Schmoranzer, J., Cassimeris, L., Goode, B. L., and Gundersen, G. G. (2008) J Cell Biol 181, 523–536 Baum, J., Tonkin, C. J., Paul, A. S., Rug, M., Smith, B. J., Gould, S. B., Richard, D., Pollard, T. D., and Cowman, A. F. (2008) Cell Host Microbe 3, 188–198 Bear, J. E., Svitkina, T. M., Krause, M., Schafer, D. A., Loureiro, J. J., Strasser, G. A., Maly, I. V., Chaga, O. Y., Cooper, J. A., Borisy, G. G., and Gertler, F. B. (2002) Cell 109(4), 509–521 Bione, S., Sala, C., Manzini, C., Arrigo, G., Zuffardi, O., Banfi, S., Borsani, G., Jonveaux, P., Philippe, C., Zuccotti, M., Ballabio, A., and Toniolo, D. (1998) Am J Hum Genet 62, 533–541 Blanchoin, L., and Pollard, T. D. (2002) Biochemistry 41(2), 597–602 Blanchoin, L., Pollard, T. D., and Hitchcock-DeGregori, S. E. (2001) Curr Biol 11(16), 1300–1304 Bosch, M., Le, K. H., Bugyi, B., Correia, J. J., Renault, L., and Carlier, M. F. (2007) Mol Cell 28, 555–568 Breitsprecher, D., Kiesewetter, A. K., Linkner, J., Urbanke, C., Resch, G. P., Small, J. V., and Faix, J. (2008) EMBO J 27, 2943–2954 Bretscher, A. (2003) J Cell Biol 160(6), 811–816 Bugyi, B., Didry, D., and Carlier, M. F. (2009) EMBO J 29, 14–26 Bugyi, B., Papp, G., Hild, G., Lorinczy, D., Nevalainen, E. M., Lappalainen, P., Somogyi, B., and Nyitrai, M. (2006) J Biol Chem 281, 10727–10736 Butler, B., and Cooper, J. A. (2009) Curr Biol In press Buttery, S. M., Yoshida, S., and Pellman, D. (2007) Mol Biol Cell 18, 1826–1838 Carlier, M. F., Le Clainche, C., Wiesner, S., and Pantaloni, D. (2003) Bioessays 25(4), 336–345 Carlsson, L., Nystrom, L. E., Sundkvist, I., Markey, F., and Lindberg, U. (1977) J Mol Biol 115(3), 465–483 Carramusa, L., Ballestrem, C., Zilberman, Y., and Bershadsky, A. D. (2007) J Cell Sci 120, 3870–3882 Castrillon, D. H., and Wasserman, S. A. (1994) Development 120(12), 3367–3377 Chalkia, D., Nikolaidis, N., Makalowski, W., Klein, J., and Nei, M. (2008) Mol Biol Evol 25, 2717–2733 Chang, F. (1999) Curr Biol 9(15), 849–852 Chang, F., Drubin, D., and Nurse, P. (1997) J Cell Biol 137(1), 169–182 Chereau, D., Boczkowska, M., Skwarek-Maruszewska, A., Fujiwara, I., Hayes, D. B., Rebowski, G., Lappalainen, P., Pollard, T. D., and Dominguez, R. (2008) Science 320, 239–243 Chesarone, M. A., and Goode, B. L. (2009) Curr Opin Cell Biol 21, 28–37

12

Formin-Mediated Actin Assembly

311

Chesarone, M., Gould, C. J., Moseley, J. B., and Goode, B. L. (2009) Dev Cell 16, 292–302 Chhabra, E. S., and Higgs, H. N. (2006) J Biol chem 281, 26754–26767 Chhabra, E. S., and Higgs, H. N. (2007) Nat Cell Biol 9, 1110–1121 Chhabra, E. S., Ramabhadran, V., Gerber, S. A., and Higgs, H. N. (2009) J Cell Sci 122, 1430–1440 Coffman, V. C., Nile, A. H., Lee, I. J., Liu, H., and Wu, J. Q. (2009) Mol Biol Cell 20, 5195–5210 Cooper, J. A. (2002) Curr Biol 12, R523–R525 Cramer, L. P., Siebert, M., and Mitchison, T. J. (1997) J Cell Biol 136, 1287–1305 Dettenhofer, M., Zhou, F., and Leder, P. (2008) PLoS ONE 3, e2497 DeWard, A. D., Eisenmann, K. M., Matheson, S. F., and Alberts, A. S. (2009) Biochem Biophys Acta. 1803, 226–233 Dickinson, R. B., and Purich, D. L. (2002) Biophys J 82, 605–617 Dickinson, R. B., Southwick, F. S., and Purich, D. L. (2002) Arch Biochem Biophys 406, 296–301 Didry, D., Carlier, M. F., and Pantaloni, D. (1998) J Biol chem 273, 25602–25611 Dominguez, R. (2009) Crit Rev Biochem Mol Biol 44, 351–366 Dong, Y., Pruyne, D., and Bretscher, A. (2003) J Cell Biol 161(6), 1081–1092 Drenckhahn, D., and Pollard, T. D. (1986) J Biol Chem 261(27), 12754–12758 Eggert, U. S., Mitchison, T. J., and Field, C. M. (2006) Annu Rev Biochem 75, 543–566 Eisenmann, K. M., Harris, E. S., Kitchen, S. M., Holman, H. A., Higgs, H. N., and Alberts, A. S. (2007) Curr Biol 17, 579–591 Emmons, S., Phan, H., Calley, J., Chen, W., James, B., and Manseau, L. (1995) Genes Dev 9, 2482–2494 Esue, O., Harris, E. S., Higgs, H. N., and Wirtz, D. (2008) J Mol Biol 384, 324–334 Evangelista, M., Blundell, K., Longtine, M. S., Chow, C. J., Adames, N., Pringle, J. R., Peter, M., and Boone, C. (1997) Science 276(5309), 118–122 Evangelista, M., Pruyne, D., Amberg, D. C., Boone, C., and Bretscher, A. (2002) Nat Cell Biol 4(3), 260–269 Ezezika, O. C., Younger, N. S., Lu, J., Kaiser, D. A., Corbin, Z. A., Nolen, B. J., Kovar, D. R., and Pollard, T. D. (2009) J Biol chem 284, 2088–2097 Faix, J., and Grosse, R. (2006) Dev Cell 10(6), 693–706 Feierbach, B., and Chang, F. (2001) Curr Biol 11(21), 1656–1665 Ferron, F., Rebowski, G., Lee, S. H., and Dominguez, R. (2007) EMBO J 26, 4597–4606 Frazier, J. A., and Field, C. M. (1997) Curr Biol 7(7), R414–417 Gauvin, T. J., Fukui, J., Peterson, J. R., and Higgs, H. N. (2009) Biochemistry 48, 9327–9329 Glotzer, M. (2005) Science 307(5716), 1735–1739 Goley, E. D., and Welch, M. D. (2006) Nat Rev Mol Cell Biol 7, 713–726 Gomez, T. S., Kumar, K., Medieros, R. B., Shimizu, Y., Leibson, P. J., and Billadeau, D. D. (2007) Immunity 26, 177–190 Goode, B. L., and Eck, M. J. (2007) Annu Rev Biochem 76, 593–627 Grunt, M., Zarsky, V., and Cvrckova, F. (2008) BMC Genomics 8, 115 Gunning, P. W., O’Neill, G., and Hardeman, E. C. (2008) Physiol Rev 88, 1–35 Gunning, P. W., Schevzov, G., Kee, A. J., and Hardeman, E. C. (2005) Trends Cell Biol 15, 333–341 Habas, R., Kato, Y., and He, X. (2001) Cell 107, 843–854 Hannemann, S., Madrid, R., Stastna, J., Kitzing, T., Gasteier, J., Schonichen, A., Bouchet, J., Jimenez, A., Geyer, M., Grosse, R., Benichou, S., and Fackler, O. T. (2008) J Biol chem 283, 27891–27903 Harris, E. S., Li, F., and Higgs, H. N. (2004) J Biol Chem 279, 20076–20087 Harris, E. S., Rouiller, I., Hanein, D., and Higgs, H. N. (2006) J Biol Chem 281(20), 14383–14392 Harris, S. D., Hamer, L., Sharpless, K. E., and Hamer, J. E. (1997) EMBO J 16, 3474–3483 Heindl, J. E., Saran, I., Yi, C. R., Lesser, C. F., and Goldberg, M. B. (2009) Infect Immun 78, 193–203

312

D.R. Kovar et al.

Higashida, C., Miyoshi, T., Fujita, A., Oceguera-Yanez, F., Monypenny, J., Andou, Y., Narumiya, S., and Watanabe, N. (2004) Science 303(5666), 2007–2010 Higgs, H. N. (2005) Trends Biochem Sci 30(6), 342–353 Higgs, H. N., and Peterson, K. J. (2005) Mol Biol Cell 16(1), 1–13 Higgs, H. N., Blanchoin, L., and Pollard, T. D. (1999) Biochemistry 38(46), 15212–15222 Holmes, K. C., Popp, D., Gebhard, W., and Kabsch, W. (1990) Nature 347, 44–49 Hudson, B. I., Kalea, A. Z., Del Mar Arriero, M., Harja, E., Boulanger, E., D’Agati, V., and Schmidt, A. M. (2008) J Biol chem 283, 34457–34468 Imamura, H., Tanaka, K., Hihara, T., Umikawa, M., Kamei, T., Takahashi, K., Sasaki, T., and Takai, Y. (1997) Embo J 16(10), 2745–2755 Ishizaki, T., Morishima, Y., Okamoto, M., Furuyashiki, T., Kato, T., and Narumiya, S. (2001) Nat Cell Biol 3, 8–14 Ji, P., Jayapal, S. R., and Lodish, H. F. (2008) Nat Cell Biol 10, 314–321 Jockusch, B. M., Murk, K., and Rothkegel, M. (2007) Rev Physiol Biochem Pharmacol 159, 131–149 Kaiser, D. A., Vinson, V. K., Murphy, D. B., and Pollard, T. D. (1999) J Cell Sci 112 (Pt 21), 3779–3790 Kamasaki, T., Arai, R., Osumi, M., and Mabuchi, I. (2005) Nat Cell Biol 7(9), 916–917 Kamasaki, T., Osumi, M., and Mabuchi, I. (2007) J Cell Biol 178, 765–771 Kanaya, H., Takeya, R., Takeuchi, K., Watanabe, N., Jing, N., and Sumimoto, H. (2005) Genes Cells 10, 665–678 Kang, F., Purich, D. L., and Southwick, F. S. (1999) J Biol Chem 274(52), 36963–36972 Karpova, T. S., McNally, J. G., Moltz, S. L., and Cooper, J. A. (1998) J Cell Biol 142(6), 1501–1517 Kato, T., Watanabe, N., Morishima, Y., Fujita, Y., Ishizaki, T., and Narumiya, S. (2001) J Cell Sci 114, 775–784 Kerkhoff, E. (2006) Trends Cell Biol 16, 477–483 Khadka, D. K., Liu, W., and Habas, R. (2009) Dev Biol 332, 396–406 Kitayama, C., Sugimoto, A., and Yamamoto, M. (1997) J Cell Biol 137(6), 1309–1319 Kohno, H., Tanaka, K., Mino, A., Umikawa, M., Imamura, H., Fujiwara, T., Fujita, Y., Hotta, K., Qadota, H., Watanabe, T., Ohya, Y., and Takai, Y. (1996) EMBO J 15, 6060–6068 Kovar, D. R. (2006) Curr Opin Cell Biol 18, 11–17 Kovar, D. R., and Pollard, T. D. (2004) Proc Natl Acad Sci U S A 101(41), 14725–14730 Kovar, D. R., Harris, E. S., Mahaffy, R., Higgs, H. N., and Pollard, T. D. (2006) Cell 124, 423–435 Kovar, D. R., Kuhn, J. R., Tichy, A. L., and Pollard, T. D. (2003) J Cell Biol 161(5), 875–887 Kovar, D. R., Wu, J. Q., and Pollard, T. D. (2005) Mol Biol Cell 16(5), 2313–2324 Kupi, T., Grof, P., Nyitrai, M., and Belagyi, J. (2009) Biophys J 96, 2901–2911 Kursula, P., Kursula, I., Massimi, M., Song, Y. H., Downer, J., Stanley, W. A., Witke, W., and Wilmanns, M. (2008) J Mol Biol 375(1), 270–290 Lai, S. L., Chan, T. H., Lin, M. J., Huang, W. P., Lou, S. W., and Lee, S. J. (2008) PLoS ONE 3, e3439 Lammers, M., Rose, R., Scrima, A., and Wittinghofer, A. (2005) Embo J 24(23), 4176–4187 Leung, T., Chen, X. Q., Manser, E., and Lim, L. (1996) Mol Cell Biol 16, 5313–5317 Li, F., and Higgs, H. N. (2003) Curr Biol 13(15), 1335–1340 Li, F., and Higgs, H. N. (2005) J Biol Chem 280(8), 6986–6992 Liu, W., Sato, A., Khadka, D., Bharti, R., Diaz, H., Runnels, L. W., and Habas, R. (2008) Proc Natl Acad Sci U S A 105, 210–215 Lizarraga, F., Poincloux, R., Romao, R., Montagnac, G., Le Dez, G., Bonne, I., Rigaill, G., Raposo, G., and Chavrier, P. (2009) Cancer Res 69, 2792–2800 Lu, J., Meng, W., Poy, F., Maiti, S., Goode, B. L., and Eck, M. J. (2007) J Mol Biol 369, 1258–1269 Lynch, E. D., Lee, M. K., Morrow, J. E., Welcsh, P. L., Leon, P. E., and King, M. C. (1997) Science 278(5341), 1315–1318 Maas, R. L., Zeller, R., Woychik, R. P., Vogt, T. F., and Leder, P. (1990) Nature 346, 853–855

12

Formin-Mediated Actin Assembly

313

Machesky, L. M., Atkinson, S. J., Ampe, C., Vanderkerckhove, J., and Pollard, T. D. (1994) J Cell Biol 127, 107–115 Mallavarapu, A., and Mitchison, T. J. (1999) J Cell Biol 146, 1097–1106 Manseau, L., Calley, J., and Phan, H. (1996) Development 122, 2109–2116 Marhoul, J. F., and Adams, T. H. (1995) Genetics 139, 537–547 Martin, S. G. (2009) Trends Cell Biol 19, 447–454 Martin, S. G., and Chang, F. (2006) Curr Biol 16(12), 1161–1170 Martin, S. G., McDonald, W. H., Yates, J. R., 3rd, and Chang, F. (2005) Dev Cell 8(4), 479–491 Martin, S. G., Rincon, S. A., Basu, R., Perez, P., and Chang, F. (2007) Mol Biol Cell 18, 4155–4167 Matusek, T., Djiane, A., Jankovics, R., Brunner, D., Mlodzik, M., and Mihaly, J. (2006) Development 133, 957–966 Matusek, T., Gombos, R., Szecsenyi, A., Sanchez-Soriano, N., Czibula, A., Pataki, C., Gedai, A., Prokop, A., Rasko, I., and Mihaly, J. (2008) J Neurosci 28, 13310–13319 May, K. M., Watts, F. Z., Jones, N., and Hyams, J. S. (1997) Cell Motil Cytoskeleton 38(4), 385–396 McCollum, D., Feoktistova, A., Morphew, M., Balasubramanian, M., and Gould, K. L. (1996) Embo J 15(23), 6438–6446 Michelot, A., Derivery, E., Paterski-Boujemaa, R., Guerin, C., Huang, S., Parcy, F., Staiger, C. J., and Blanchoin, L. (2006) Curr Biol 16(19), 1924–1930 Michelot, A., Guerin, C., Huang, S., Ingouff, M., Richard, S., Rodiuc, N., Staiger, C. J., and Blanchoin, L. (2005) Plant Cell 17(8), 2296–2313 Minc, N., Bratman, S. V., Basu, R., and Chang, F. (2009) Curr Biol 19, 83–94 Moseley, J. B., and Goode, B. L. (2005) J Biol Chem 280(30), 28023–28033 Moseley, J. B., Sagot, I., Manning, A. L., Xu, Y., Eck, M. J., Pellman, D., and Goode, B. L. (2004) Mol Biol Cell 15(2), 896–907 Munro, E., Nance, J., and Priess, J. R. (2004) Dev Cell 7(3), 413–424 Nakano, K., Imai, J., Arai, R., Toh, E. A., Matsui, Y., and Mabuchi, I. (2002) J Cell Sci 115(Pt 23), 4629–4639 Nakano, K., Takaishi, K., Kodama, A., Mammoto, A., Shiozaki, H., Monden, M., and Takai, Y. (1999) Mol Biol Cell 10, 2481–2491 Naumanen, P., Lappalainen, P., and Hotulainen, P. (2008) J Microsc 231, 446–454 Neidt, E. M., Scott, B. J., and Kovar, D. R. (2009) J Biol Chem 284, 673–684 Neidt, E. M., Skau, C. T., and Kovar, D. R. (2008) J Biol Chem 283, 23872–23883 Nezami, A. G., Poy, F., and Eck, M. J. (2006) Structure 14, 257–263 Otomo, T., Otomo, C., Tomchick, D. R., Machius, M., and Rosen, M. K. (2005) Mol Cell 18(3), 273–281 Otomo, T., Tomchick, D. R., Otomo, C., Panchal, S. C., Machius, M., and Rosen, M. K. (2005) Nature 433(7025), 488–494 Ozaki-Kuroda, K., Yamamoto, Y., Nohara, H., Kinoshita, M., Fujiwara, T., Irie, K., and Takai, Y. (2001) Mol Cell Biol 21(3), 827–839 Palazzo, A. F., Cook, T. A., Alberts, A. S., and Gundersen, G. G. (2001) Nat Cell Biol 3, 723–729 Pantaloni, D., and Carlier, M. F. (1993) Cell 75(5), 1007–1014 Papp, G., Bugyi, B., Ujfalusi, Z., Barko, S., Hild, G., Somogyi, B., and Nyitrai, M. (2006) Biophys J 91, 2564–2572 Paul, A. S., and Pollard, T. D. (2008) Curr Biol 18(1), 9–19 Paul, A. S., and Pollard, T. D. (2009a) Cell Mot Cytoskel 66, 606–617 Paul, A. S., and Pollard, T. D. (2009b) J Biol chem 284, 12533–12540 Pechlivanis, M., Samol, A., and Kerkhoff, E. (2009) J Biol chem 284, 25324–25333 Peng, J., Kitchen, S. M., West, R. A., Sigler, R., Eisenmann, K. M., and Alberts, A. S. (2007) Cancer Res 67(16), 7565–7571 Peng, J., Wallar, B. J., Flanders, A., Swiatek, P. J., and Alberts, A. S. (2003) Curr Biol 13, 534–545 Petersen, J., Nielsen, O., Egel, R., and Hagan, I. M. (1998) J Cell Biol 141(5), 1217–1228 Petersen, J., Weilguny, D., Egel, R., and Nielsen, O. (1995) Mol Biol Cell 15, 3697–3707

314

D.R. Kovar et al.

Piel, M., and Tran, P. T. (2009) Curr Biol 19, R823–R827 Pollard, T. D. (1986) J Cell Biol 103(6 Pt 2), 2747–2754 Pollard, T. D. (2007) Annu Rev Biophys Biomol Struct 36, 451–477 Pollard, T. D., and Borisy, G. G. (2003) Cell 112(4), 453–465 Pollard, T. D., and Cooper, J. A. (1984) Biochemistry 23(26), 6631–6641 Pollard, T. D., Blanchoin, L., and Mullins, R. D. (2000) Annu Rev Biophys Biomol Struct 29, 545–576 Pring, M., Evangelista, M., Boone, C., Yang, C., and Zigmond, S. H. (2003) Biochemistry 42(2), 486–496 Pruyne, D. W., Schott, D. H., and Bretscher, A. (1998) J Cell Biol 143(7), 1931–1945 Pruyne, D., Evangelista, M., Yang, C., Bi, E., Zigmond, S., Bretscher, A., and Boone, C. (2002) Science 297(5581), 612–615 Pruyne, D., Gao, L., Bi, E., and Bretscher, A. (2004) Mol Biol Cell 15, 4971–4989 Qualmann, B., and Kesseles, M. M. (2009) Trends Cell Biol 19, 276–285 Quinlan, M. E., and Kerkhoff, E. (2009) Nat Cell Biol 10, 13–15 Quinlan, M. E., Heuser, J. E., Kerkhoff, E., and Mullins, R. D. (2005) Nature 433(7024), 382–388 Quinlan, M. E., Hilgert, S., Bedrossian, A., Mullins, R. D., and Kerkhoff, E. (2007) J Cell Biol 179, 117–128 Renault, L., Bugyi, B., and Carlier, M. F. (2008) Trends Cell Biol 18, 494–504 Rincon, S. A., Ye, Y., Villar-Tajadura, M. A., Santos, B., Martin, S. G., and Perez, P. (2009) Mol Biol Cell 20, 4390–4399 Rivero, F., Muramoto, T., Meyer, A. K., Urushihara, H., Uyeda, T. Q., and Kitayama, C. (2005) BMC Genomics 6(1), 28 Rizvi, S. A., Neidt, E. M., Cui, J., Feiger, Z., Gardel, M. L., Kozmin, S. A., and Kovar, D. R. (2009) Chem and Biol 16, 1158–1168 Romero, S., Didry, D., Larquet, E., Boisset, N., Pantaloni, D., and Carlier, M. F. (2007) J Biol Chem 232, 8435–8445 Romero, S., Le Clainche, C., Didry, D., Egile, C., Pantaloni, D., and Carlier, M. F. (2004) Cell 119(3), 419–429 Rose, R., Weyand, M., Lammers, M., Ishizaki, T., Ahmadian, M. R., and Wittinghofer, A. (2005) Nature 435(7041), 513–518 Ryu, J. R., Eckharri, A., Li, R., and Pendergast, A. M. (2009) Mol Cell Biol 29, 1735–1748 Sagot, I., Klee, S. K., and Pellman, D. (2002a) Nat Cell Biol 4(1), 42–50 Sagot, I., Rodal, A. A., Moseley, J., Goode, B. L., and Pellman, D. (2002b) Nat Cell Biol 4(8), 626–631 Sato, A., Khadka, D. K., Liu, W., Bharti, R., Runnels, L. W., Dawid, I. B., and Habas, R. (2006) Development 133, 4219–4231 Schirenbeck, A., Arasada, R., Bretschneider, T., Schleicher, M., and Faix, J. (2005a) Biochem Soc Trans 33(Pt 6), 1256–1259 Schirenbeck, A., Bretschneider, T., Arasada, R., Schleicher, M., and Faix, J. (2005b) Nat Cell Biol 7, 619–625 Schulte, A., Stolp, B., Schonichen, A., Pylypenko, O., Rak, A., Fackler, O. T., and Geyer, M. (2008) Structure 16, 1313–1323 Schumacher, N., Borawski, J. M., Leberfinger, C. B., Gessler, M., and Kerkhoff, E. (2004) Gene Expr Patterns 4, 249–255 Severson, A. F., Baillie, D. L., and Bowerman, B. (2002) Curr Biol 12(24), 2066–2075 Shemesh, T., Otomo, T., Rosen, M. K., Bershadsky, A. D., and Kozlov, M. M. (2005) J Cell Biol 170(6), 889–893 Shimada, A., Nyitrai, M., Vetter, I. R., Kuhlmann, D., Bugyi, B., Narumiya, S., Geeves, M. A., and Wittinghofer, A. (2004) Mol Cell 13(4), 511–522 Skau, C. T., Neidt, E. M., and Kovar, D. R. (2009) Mol Biol Cell 20, 2160–2173 Svitkina, T. M., Bulanova, E. A., Chaga, O. Y., Vignjevic, D. M., Kojima, S., Vasiliev, J. M., and Borisy, G. G. (2003) J Cell Biol 160(3), 409–421

12

Formin-Mediated Actin Assembly

315

Swan, K. A., Severson, A. F., Carter, J. C., Martin, P. R., Schnabel, H., Schnabel, R., and Bowerman, B. (1998) J Cell Sci 111 (Pt 14), 2017–2027 Takeya, R., Taniguchi, K., Narumiya, S., and Sumimoto, H. (2008) EMBO J 27, 618–628 Tanaka, K. (2000) Biochem Biophys Res Commun 267, 479–481 Taniguchi, K., Takeya, R., Suetsugu, S., Kan-O, M., Narusawa, M., Shiose, A., Tominaga, R., and Sumimoto, H. (2009) J Biol chem 284, 29873–29881 Terenna, C. R., Makushok, T., Velve-Dasquillas, G., Baigl, D., Chen, Y., Bornens, M., Paoletti, A., Piel, M., and Tran, P. T. (2008) Curr Biol 18, 1748–1753 Tolliday, N., VerPlank, L., and Li, R. (2002) Curr Biol 12(21), 1864–1870 Uehata, M., Ishizaki, T., Satoh, H., Ono, T., Kawahara, T., Morishita, T., Tamakawa, H., Yamagami, K., Inui, J., Maekawa, M., and Narumiya, S. (1997) Nature 389, 990–994 Ujfalusi, Z., Barko, S., Hild, G., and Nyitrai, M. (2009a) J Photochem Photobiol B 98, 7–11 Ujfalusi, Z., Vig, A., Hild, G., and Nyitrai, M. (2009b) Biophys J 96, 162–168 Vaillant, D. C., Copeland, S. J., Davis, C., Thurston, S. F., Abdennur, N., and Copeland, J. W. (2008) J Biol chem 283, 33750–33762 Vavylonis, D., Kovar, D. R., O’Shaughnessy, B., and Pollard, T. D. (2006) Mol Cell 21(4), 455–466 Vavylonis, D., Wu, J. Q., Hao, S., O’Shaughnessy, B., and Pollard, T. D. (2008) Science 319(5859), 97–100 Vidali, L., van Gisbergen, P. A. C., Guerin, C., Franco, P., Li, M., Burkart, G., Augustine, R. C., Blanchoin, L., and Bezanilla, M. (2009) Proc Natl Acad Sci U S A 106, 13341–13346 Wallar, B. J., and Alberts, A. S. (2003) Trends Cell Biol 13(8), 435–446 Wallar, B. J., Stropich, B. N., Schoenherr, J. A., Holman, H. A., Kitchen, S. M., and Alberts, A. S. (2005) J Biol Chem 281, 4300–4307 Wang, H., and Vavylonis, D. (2008) PLoS ONE 3, e4078 Wang, J., Neo, S. P., and Cai, M. (2009) Traffic In press Watanabe, N., Kato, T., Fujita, A., Ishizaki, T., and Narumiya, S. (1999) Nat Cell Biol 1(3), 136–143 Watanabe, N., Madaule, P., Reid, T., Ishizaki, T., Watanabe, G., Kakizuka, A., Saito, Y., Nakao, K., Jockusch, B. M., and Narumiya, S. (1997) EMBO J 16, 3044–3056 Watanabe, S., Ando, Y., Yasuda, S., Hosoya, H., Watanabe, N., Ishizaki, T., and Narumiya, S. (2008) Mol Biol Cell 19(5), 2328–2338 Wawro, B., Greenfield, N. J., Wear, M. A., Cooper, J. A., Higgs, H. N., and Hitchcock-DeGregori, S. E. (2007) Biochemistry 46, 8146–8155 Wen, Y., Eng, C. H., Schmoranzer, J., Cabrera-Poch, N., Morris, E. J., Chen, M., Wallar, B. J., Alberts, A. S., and Gundersen, G. G. (2004) Nat Cell Biol 3, 723–729 Werner, M., Munro, E., and Glotzer, M. (2007) Curr Biol 17(15), 1286–1297 Witke, W. (2004) Trends Cell Biol 14(8), 461–469 Woychik, R. P., Maas, R. L., Zeller, R., Vogt, T. F., and Leder, P. (1990) Nature 346, 850–853 Wu, J. Q., Kuhn, J. R., Kovar, D. R., and Pollard, T. D. (2003) Dev Cell 5(5), 723–734 Wu, J. Q., Sirotkin, V., Kovar, D. R., Lord, M., Beltzner, C. C., Kuhn, J. R., and Pollard, T. D. (2006) J Cell Biol 174(3), 391–402 Xu, Y., Moseley, J. B., Sagot, I., Poy, F., Pellman, D., Goode, B. L., and Eck, M. J. (2004) Cell 116(5), 711–723 Yamashita, M., Higashi, T., Suetsugu, S., Sato, Y., Ikeda, T., Shirakawa, R., Kita, T., Takenawa, T., Horiuchi, H., Fukai, S., and Nureki, O. (2007) Genes Cells 12, 1255–1265 Yang, C., Czech, L., Gerboth, S., Kojima, S., Scita, G., and Svitkina, T. (2007) 5, e317 Yang, H. C., and Pon, L. A. (2002) Proc Natl Acad Sci U S A 99(2), 751–756 Yarmola, E. G., and Bubb, M. R. (2009) Bioessays 31, 1150–1160 Yasuda, S., Oceguera-Yanez, F., Kato, T., Okamoto, M., Yonemura, S., Terada, Y., Ishizaki, T., and Narumiya, S. (2004) Nature 428, 767–771 Yonetani, A., Lustig, R. J., Moseley, J. B., Takeda, T., Goode, B. L., and Chang, F. (2008) Mol Biol Cell 19, 2208–2219

316

D.R. Kovar et al.

Young, K. G., and Copeland, J. W. (2008) Biochem Biophys Acta 1803, 183–190 Zhou, F., Leder, P., Zuniga, A., and Dettenhofer, M. (2009) Hum Mol Genet 18, 2472–2482 Zhu, X.-L., Liang, L., and Ding, Y.-Q. (2008) In J Colorectal Dis 23, 1041–1047 Zigmond, S. H., Evangelista, M., Boone, C., Yang, C., Dar, A. C., Sicheri, F., Forkey, J., and Pring, M. (2003) Curr Biol 13(20), 1820–1823 Zuniga, A., Michos, O., Spitz, F., Haramis, A. P., Panman, L., Galli, A., Vintersten, K., Klasen, C., Mansfield, W., Kuc, S., Duboule, D., Dono, R., and Zeller, R. (2004) Genes Dev 18, 1553–1564

Chapter 13

Visualization of Individual Actin Filament Assembly Emmanuèle Helfer

Abstract Actin is a major protein involved in many motile processes essential for embryonic development, immune response, endo/exocytosis, cytokinesis. These processes are based on the rapid reorganization of the actin cytoskeleton in response to extracellular signaling. Indeed, in vivo observation of motile cells showed the coexistence of different actin structures: branched networks in the lamellipodium, parallel bundles in filopodia, fibers in focal adhesion complexes. Though these structures can be located in the same region of the cell, their dynamics can be very different. Real-time cell imaging showed that the filament turnover differs from one structure to the other. The molecular mechanisms that drive actin assembly/disassembly in these networks are not yet fully deciphered. How the cell controls in a concerted way the various actin motile machineries is a crucial question that is to be answered using a multidisciplinary approach. Actin dynamics was initially studied in bulk: biochemical assays were developed allowing to follow polymerization of actin filaments, and to discriminate the dynamics at each end of the filament. These studies led to an understanding of the treadmilling process of the actin filament: in presence of ATP, the net polymerization at the barbed-end balances the net depolymerization at the pointed-end. These assays are currently used to decipher the function of new actin regulating proteins. To complement the solution studies microscopy assays were developed to observe individual actin filaments. However, the first observations were done with phalloidin-stabilized filaments unable to sustain any dynamics. Only recently the microscopy technique was improved, by using excitation by an evanescent wave, allowing observation of a dynamic filament. Dynamic imaging and solution studies lead to the measurement of filament turnover dynamics. The combination with biomimetic assays consisting in in vitro reconstitution of actin structure assembly provides mechanistic insight into actin-based motile processes.

E. Helfer (B) Laboratoire d’Enzymologie et Biochimie Structurales, CNRS UPR3082, Bât. 34 Avenue de la Terrasse, F-91198 Gif-sur-Yvette Cedex, France e-mail: [email protected]

M.-F. Carlier (ed.), Actin-based Motility, DOI 10.1007/978-90-481-9301-1_13, C Springer Science+Business Media B.V. 2010

317

318

E. Helfer

Contents 13.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . 13.1.1 Light Microscopy As a Complement to Biochemical Assays 13.1.2 Principle of TIRFM . . . . . . . . . . . . . . . . . . . 13.1.3 Instrumental Approaches of TIRFM . . . . . . . . . . . 13.1.4 Optimal Parameters of the TIRF Setup . . . . . . . . . . 13.1.5 Development of TIRFM Assays . . . . . . . . . . . . . 13.2 Actin Filaments Assembly Dynamics . . . . . . . . . . . . . . . 13.2.1 Assembly of Individual Filament . . . . . . . . . . . . . 13.2.2 Branched Actin Structures Generated by the Arp2/3 Complex 13.3 Conclusion and Perspectives . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

318 318 319 320 321 321 324 324 327 331 332

13.1 Introduction 13.1.1 Light Microscopy As a Complement to Biochemical Assays Actin has been studied for several decades now, using spectroscopy and microscopy techniques. Actin filaments (F-actin) are polar helical polymers resulting from selfassembly of globular actin monomers (G-actin). Spectroscopic solution studies allow following actin assembly at barbed and pointed ends, monitoring the interactions between actin and actin-binding proteins (ABPs), and deriving their molecular functions (Hertzog and Carlier, 2005). To visualize actin filament organization during the bulk measurements, microscopy assays were developed. The first assay was derived from the in vitro acto-myosin motility assay where few actin filaments, stabilized by rhodamine-phalloidin, can be observed using standard epifluorescence. Indeed, to be able to observe individual filaments the actin solution was diluted below the critical concentration of actin (a few 10 nM << [A]c = 0.1 μM). In these conditions, filaments would depolymerize if not stabilized with phalloidin. However, the binding of phalloidin to the actin filaments blocks their dynamics. These experiments allowed to acquire “snapshots” at varying times during polymerization and to derive from a statistical image analysis the kinetics of reactions, without direct observation. Direct visualization of the dynamics of individual actin filaments required thus an improvement of the imaging techniques. Recently, the wide-field fluorescence microscopy technique was modified into Total Internal Reflection Fluorescence Microscopy (TIRFM) allowing to work at concentrations above [A]c and to observe dynamic filaments in real-time without stabilization. The principle of TIRFM is described in the following section. Using this new technique, actin filament assembly was observed in real-time for the first time by Amann and Pollard in 2001 (Amann and Pollard, 2001). This chapter enlights how improvement of light microscopy analysis of actin filament dynamics provides mechanistic insight

13

Visualization of Individual Actin Filament Assembly

319

into actin assembly during cellular processes. After a brief description of TIRF principle we will show how the combination of light microscopy with dynamic systems allows to decipher molecular mechanisms of actin regulation.

13.1.2 Principle of TIRFM To improve signal-to-noise ratio of fluorescent molecules, an evanescent wave is used to illuminate only a thin layer near the surface (depth 100–200 nm). Figure 13.1 schematically describes the principle of Total Internal Reflection Fluorescence (TIRF). When a light beam encounters an interface between 2 media of different refractive indexes, the first one with a high refractive index (oil and glass: n1 = 1.518) and the second one with a lower refractive index (aqueous specimen: n2 = 1.33), it is refracted according to Snell’s law (Fig. 13.1a). When the incidence angle is greater than a certain angle called critical angle θc = Arcsin(n2 /n1 ), the light beam is totally reflected, and generates an exponentially decaying evanescent wave inside the specimen (Fig. 13.1b). Fluorescent particles lying close to the surface will be excited by the evanescent wave whereas objects lying above 200 nm from the surface will not be illuminated, as schematized in Fig. 13.2a: a bead in Brownian motion is fluorescent only when moving in the evanescent field. Fluorescent objects in bulk that usually contribute to fluorescence background in wide-field fluorescence will not emit light out of the evanescent illumination leading to an appreciable reduction of the background. This

A

B θr

n2(water) ~ 1.33

n2(water) ~ 1.33

n1(glass) ~ 1.5

n1(glass) ~ 1.5

θi At θi < θc : transmission Snell’s law Critical angle (~61° for water/glass interface) Evanescent wave intensity as function of distance from interface Penetration depth

θc At θi > θc : total reflection n 1·sin(θi) = n2·sin(θr) θc= Arcsin(n 2/n1) I(z) = I(0) e(-z/l p) lp = l/[4π(n12 sin2θi-n22)½]

Fig. 13.1 Principle of Total Internal Reflection Fluorescence Microscopy. (a) For θi < θc the incident beam is transmitted inside the sample at angle θr according to Snell’s law: n1 sin(θi ) = n2 sin(θr ). (b) For θi < θc the light beam is totally reflected and an evanescent wave is generated inside the sample that decays exponentially. The penetration depth depends on the beam wavelength

320

E. Helfer

Fig. 13.2 (a) A fluorescent bead moving in and out of the evanescent field emits (green color) only when it enters the illuminated region. (b) Actin filaments observed in wide-field fluorescence (left) and TIRFM (right): the filaments are barely distinguished with standard illumination. Bar = 10 μm

therefore enables a higher signal-to-noise ratio as compared to standard epifluorescence. Figure 13.2b shows a population of actin filaments observed using either wide-field fluorescence (left) or TIRF (right). Although the TIRFM technique is limited to the observation of the interface between 2 media it has a great number of applications: we describe in Section 13.2 how it allows real-time visualization of dynamic actin filaments.

13.1.3 Instrumental Approaches of TIRFM There are two basic configurations for TIRFM: the prism method and the objective lens method (Fig. 13.3). In the first setup the laser beam hits a prism in contact with A

B Prism

Oil

Glycerol Immersion objective

Objective rear aperture

Fig. 13.3 Schematics of TIRFM setups. (a) Prism method (b) Objective lens method

13

Visualization of Individual Actin Filament Assembly

321

the sample slide with an angle tuned to maintain the right incident angle (θi > θc ) at the glass/water interface. In this configuration, the evanescent wave occurs at the top interface of the sample, and the observation must be done through the entire sample leading to a decreased intensity of the emitting signal of the fluorophores. The sample thickness cannot be too large (∼30 μm) so that the objective working distance enables observation at the top surface. This setup was of general use before the design and manufacturing of objectives with high numerical apertures (NA). In this most recent setup, the laser beam is tilted before entering the objective so as to go at the periphery of the objective rear focal plane and hit the interface with the right θi , the illuminated region and the observed one are the same and are closer to the objective. This objective-based setup is the one supplied by the main microscopy companies.

13.1.4 Optimal Parameters of the TIRF Setup Several crucial factors govern the optimal use of evanescent wave in microscopy. • To work in total reflection conditions the refractive index of the medium of illumination incidence (oil and glass) must be higher than that of the sample (typically aqueous medium). • The numerical aperture (NA) of the objective is of great incidence on the evanescent wave efficiency. The relationship between NA and the maximum achievable illumination incidence angle θi is NA = n sin(θi ) where n is the refractive index of the medium between the cover glass and the objective front lens (oil). Combining this relationship with the total internal reflection conditions (with nwater = 1.33) leads to an objective of NA greater than nwater , ideally 1.45 or higher. The laser beam is focused at the periphery of the objective rear aperture to reach the maximum θi (Fig. 13.3b). • The penetration depth depends on the incident illumination wavelength (see Fig. 13.2), allowing observation in a layer of varying thickness depending on the laser used. One must choose appropriate fluorophores to be excited by the laser beam. When working with 2 distinct fluorophores the excitation laser wavelengths must be carefully chosen so as to have evanescent waves of similar penetration depths.

13.1.5 Development of TIRFM Assays Since the first experiments in 2001, the TIRFM assays have been implemented in various ways to optimize the visualization of actin filaments dynamics. • As the observation is limited to a thin layer (100–200 nm) from the surface, micrometer-long filaments move in and out of the evanescent field due to Brownian motion and cannot be tracked if they are not 2D-constrained. To

322

E. Helfer

maintain them close to the surface the polymerization medium is complemented with methycellulose (typically 0.2% w/v), a polymer which strongly reduces the Z-motion of the filaments. • Imaging and analyzing softwares were developed (Metamorph, ImageJ, Matlab. . .) in order to track automatically the filaments in Brownian motion and facilitate statistical analysis of data. However this automatization is mostly limited to experiments where filaments are located far enough from each other: if too crowded they can cross accidently and induce false length measurement. Recently, Vavylonis et al. developed a tracking routine using active contours that allows to discriminate the filaments even when they are crossing each other (Li et al. 2009). Another way to facilitate single filament observation consisted in attaching them to the surface either using ABPs (myosin, filamin. . .) or biochemical glu (biotin-streptavidin, biotin-antibiotin. . .). Depending on the surface density of the “linkers” filaments are more or less immobilized on the surface. When they are strongly anchored and do not fluctuate laterally, analyzing methods like kymograph can be applied to measure automatically the filament elongation. • Finally, efforts were done to develop new probes to visualize actin assembly. Apart from standard fluorophores that generate homogeneously fluorescent filaments, Alexa-532 is a new dye which bleaches faster than conventional ones (Alexa-488, Oregon Green, rhodamine. . .). Due to its rapid photobleaching it

Fig. 13.4 Alexa-532 bleaches faster than standard fluorophores like Rhodamine. (a) Snapshots of growing actin filaments labeled with Rhodamine: the filaments are homogeneously fluorescent and can be detected over their total length. (b) Snapshots of growing actin filaments labeled with Alexa532: as the dye bleaches rapidly one can observe only the newly polymerized parts of the filaments. (c) Decrease of fluorescence intensity of Rhodamine (top curve) and Alexa-532 (bottom curve) under TIRF illumination. From exponential decay fitting the photobleaching rate of Alexa-532 is 10 times faster than that of Rhodamine. (from (Michelot et al. 2006))

13

Visualization of Individual Actin Filament Assembly

323

induces comet-like appearance of actin filaments (Fig. 13.4). Alexa-532 thus appeared very useful for monitoring incorporation of new subunits in actin filaments and tracking the position of the growing end.

13.1.5.1 Limitations of TIRFM Assays Though the TIRF technique offers new ways to answer issues that were not solvable using simple bulk measurements, one cannot neglect drawbacks that limit the range of feasible experiments. • Labeling on actin can affect its polymerization properties: Pollard and coll. studied the polymerization kinetics of actin filaments using TIRFM, with varying amount of labeled monomers. They showed that the actin, labeled on cysteine 374, does not take part to polymerization (Amann and Pollard, 2001; Kuhn and Pollard, 2005). When not considering the amount of labeled actin, the authors derived kinetics constants of actin filament assembly within the range of the values deduced from bulk solution experiments (Gutsche-Perelroizen et al., 1999) and electron microscopy (Pollard and Cooper, 1986). Another important point is the binding of the fluorophore itself to actin: labeling on cysteine (via maleimide or iodoacetamide reactive groups) prevents interaction of actin with profilin whereas labeling on surface lysines (via succinimidyl ester reactive group) does not interfere. As a result, filaments nucleated by formins that elongate processively due to profilin-mediated interaction cannot incorporate easily cysteine-labeled monomers which bind poorly to profilin and therefore cannot interact efficiently with formin (Romero et al., 2004; Paul and Pollard, 2008). This results in dim filaments in contrast with bright filaments nucleated in absence of formins that can incorporate more often fluorescent monomers. The reduced signal-to-noise can be a problem to measure the filament elongation. On the other hand, it allows to distinguish filaments with and without formins bound to their end. • As written above attaching the filaments to the surface helps tracking their length. However it also increases the amount of filament crowding on the surface and reduces even more the concentration range of actin for analyzable data. Strong attachment to the surface can also induce artefacts in the observed kinetics. Kuhn and Pollard indeed noticed pauses during filaments growth which they correlate with unspecific interactions with the surface (Kuhn and Pollard, 2005). However, it was observed that discrete anchoring via NEM-myosin (binds to filament side) or anti-biotin (binds to biotinylated subunits in filaments) interferes with disassembly when the filament end comes to the anchoring point (personal communication). On the other hand, Kueh et al. used filamin to anchor the actin filaments which seemed not to interfere with the filament depolymerization (Kueh et al., 2008). Control experiments are thus critical to determine if the nature of the anchor affects or not the filament dynamics.

324

E. Helfer

• Rapid reactions after addition of a component to the medium cannot be studied using this technique. Flow cells are used to passivate or functionalize the sample surface before injecting the actin solution: observation typically starts ∼1 min after mixing. This dead-time prevents study of transient reactions.

13.2 Actin Filaments Assembly Dynamics Cell imaging has shown that the actin structures are highly dynamic. A striking example is the movie of a melanocyte crawling on a surface in which one can see GFP-actin turnover at the leading edge of the lamellipodium: as the cell moves the actin is constantly renewed at the front generating a fluorescence region of constant width always located at the front of the cell. In vivo studies show that actin filament arrays are regulated by diverse machineries with different dynamics, but cannot lead to the detailed molecular mechanisms of regulation. Combination of bulk solution measurements and analysis of individual filament dynamics allows a more complete understanding of assembly/disassembly of actin filament structures. We show in this section different TIRFM assays that were developed to measure the dynamics of actin filaments.

13.2.1 Assembly of Individual Filament 13.2.1.1 Filament Elongation from Pure Actin Let us start with the simplest TIRFM assay consisting in real-time observation of filament elongation from pure actin. It is known from bulk solution measurements that pure actin polymerizes as soon as actin concentration is higher than the critical concentration of the barbed end [A]c = 0.1μM. Once nucleation has started the filaments elongate at a rate that depends on the monomers concentration. Filaments grow initially at constant rate then slow down as the monomer pool is depleted by the polymerization of the filaments. Figure 13.5a shows the timelapse of actin filaments growing at 0.5 μM actin observed using TIRFM imaging. In this assay, filaments do not stick to the surface (coated with BSA to prevent non-specific interaction). Most of the filaments move in Brownian motion, 2D-restricted in the vicinity of the surface by the presence of methylcellulose (this thickening agent restricts large objects motion but not monomer diffusion). There is no way to distinguish the barbed and pointed ends and their respective elongation. However from time to time, one filament can attach accidently to the surface (red arrow), allowing to discriminate the two ends. One clearly sees that one part of the filament elongates whereas the other one barely changes during the acquisition. Measuring the distances between the anchoring point and the two ends showed that the total length corresponds actually to the sole elongation at the barbed end, and that the length change at the pointed end is negligible over the 25-min movie, at the low concentration of actin used here. Figure 13.5b

13

Visualization of Individual Actin Filament Assembly

Fig. 13.5 Measurement of filament elongation. (a) Time-lapse of growing actin filaments polymerized at 0.5 μM, 15% Alx-488 labelled actin (2 min between the images). The red arrow points to a filament which is stuck to the surface after a few mins allowing to discriminate barbed and pointed ends. Scale bars = 10 μm. (b) Length as function of time of the stuck filament (thick curve) and another filament. The growth rate is ∼ 0.8 μm/min. Time zero corresponds to the first image of the movie

325

A

B

shows that the total length of two filaments, the one stuck to the surface and another one moving freely at the surface, behave similarly: both have an elongation rate of 0.8 μm/min. The sharp length decreases occurring from time to time (arrows) correspond to images in which the filament end has moved out of the evanescent wave, leading to a shorter apparent length. The plot of the total length shows that the growth rate is constant during the time of observation (at least 25 min here) indicating that the observed growth is still in the beginning of the elongation regime before the depletion of the reservoir of G-actin can affect the growth rate. It must be noted that the TIRFM assay induces spontaneous nucleation. Indeed, at such low concentration of 0.5 μM, the nucleation lag in solution is longer than 1 h, whereas in TIRFM filaments appear at the surface within a few min. The elongation rate measured from TIRFM images is in agreement with bulk results. It seems the vicinity of the surface accelerates the nucleation process by an unknown mechanism without affecting the assembly dynamics. 13.2.1.2 Association/Dissociation Rate Measurement The TIRFM assays can then be used to visualize directly the assembly/disassembly at the two ends of actin bound to different nucleotides and derive the corresponding rate constants for comparison with values established from EM and bulk kinetic assays.

326

E. Helfer

Fig. 13.6 Filament elongation rate as a function of actin concentration. The slope is the association rate constant kon: Ve = kon ([A] − [A]c ), where [A] is the monomer concentration and [A]c = 0.1μM is the barbed end critical concentration

As stated earlier however, biochemical conditions cannot be the same in bulk and in TIRFM. The range of actin concentrations that can be explored is much broader in solution studies, as the spectroscopic measurements are averaging on a large population of filaments, than the one analyzable in TIRFM. In the latter case it is rather difficult to work at actin concentrations higher than 1 μM. By neglecting the pointed end dynamics measurement of total elongation rates (i.e. barbed-end elongation) at varying actin concentrations allows to deduce the kon rate of actin association at barbed end. Figure 13.6 shows a measurement series of filament growth rate at varying actin concentration. In agreement with biochemical studies the elongation rate increases linearly with the actin concentration. We measure a slope of kon = 9.6 μM−1 .s−1 , coherent with bulk studies (k+ = 10 μM−1 .s−1 , Gutsche-Perelroizen et al. (1999)) as well as with microscopy measurements of stabilized filaments (k+ = 9.5 μM−1 .s−1 , Romero et al. (2004)). Using similar TIRFM assays, Kuhn and Pollard tracked elongation and shrinkage of actin filaments in different conditions (actin concentration, type of nucleotide bound to actin) and deduced the association and dissociation rates at barbed and pointed ends (Table 1, in Kuhn and Pollard (2005)). The derived values were coherent with values measured using solution assays and slightly smaller than EM measurements (Pollard and Cooper, 1986). 13.2.1.3 Regulation of Actin Filament Assembly The effect of actin regulating proteins on actin assembly can as well be measured using TIRFM assays. Some examples are shown below. Profilin Profilin plays a major role in actin assembly at barbed end. It is known to inhibit spontaneous actin nucleation (Pollard and Cooper, 1984). It specifically binds ATP-G-actin in a complex that associates exclusively to barbed end. Gutsche-Perelroizen et al. showed that the barbed end association rate of profilin-actin is 30% lower than that of pure actin (Gutsche-Perelroizen et al., 1999).

13

Visualization of Individual Actin Filament Assembly

327

Fig. 13.7 Elongation rate as function of actin concentration, in absence (solid circles) and presence (open squares) of profilin. The association rate constant decreases from 7.4 to 5.6 μM−1 .s−1 when actin is bound to profilin. In the second assay, nucleation inhibition by profilin was compensated by addition of spectrin-actin seeds that immediately start to elongate

TIRFM measurements of elongation rate of non-attached filaments in absence and presence of profilin led to the same conclusion, confirming that direct observation of individual filaments, even close to a surface, gives the same result as averaged measurement on a solution of filaments (Fig. 13.7). Additionally, we observed that profilin inhibits nucleation but not totally, certainly due to the competing nucleation induced by surface vicinity. To observe growing filaments immediately, spectrinactin seeds were added to the mixture of actin and profilin that elongate from barbed ends exclusively. Actin Nucleators: Formins Formins are nucleators of single filaments and are involved in motile processes such as the formation of actin cables that contribute to the polarity of cells (Evangelista et al., 1997; Sagot et al., 2002) and filopodia formation (Peng et al., 2003). It was shown that some formins elongate actin filaments processively, and requires or not profilin for processivity (Romero et al., 2004; Kovar and Pollard, 2004). Using TIRFM, Kovar and Pollard observed the buckling of filaments anchored by NEMmyosin and assembled from formins immobilized on the surface (Fig. 13.8): they derived a minimal value of 1.3 pN of the force generated by actin polymerization (Kovar and Pollard, 2004).

13.2.2 Branched Actin Structures Generated by the Arp2/3 Complex In vivo imaging showed that different actin filament structures coexist in cells: branched networks in the lamellipodium, parallel bundles in filopodia, fibers in

328

E. Helfer

Fig. 13.8 TIRFM observation of the buckling of an actin filament. Formins and NEM-myosins are immobilized on the glass surface. The filament grows at its barbed end (green wedge) attached to a formin, its side is attached to a NEM-myosin (white ring), the pointed end is indicated by the red triangle. (from Kovar and Pollard (2004))

focal adhesion complexes. It was shown that the filament turnover differs from one structure to the other, even though they can be in the same region of the cell, like lamellipodium and filopodia both located at the leading edge of a moving cell. The coordinated regulation of the different actin machineries is of major interest, but requires first that molecular mechanisms that drive actin assembly/disassembly in each network are fully understood. The improved TIRFM technique opens a new approach to solve issues that could not be answered using simple bulk solution measurements and wide-field microscopy. 13.2.2.1 Formation of Actin Filament Branches The lamellipodium extension is induced by a protrusive force generated by the growth of a branched actin array against the cell membrane. Proteins of the WASP/Scar family activate the Arp2/3 complex to branch actin filaments and generate a polarized dendritic array in which actin filament barbed ends grow and push against the membrane. Though many studies have been done using diverse approaches (bulk solution assays, structural studies, microscopy assays, in vivo imaging) the filament multiplication by branching is still under investigation. It is agreed that the “activated” ternary complex, formed by the Arp2/3 complex, its activator and an actin monomer, binds to a “mother” filament and generates a “daughter” filament. Subsequent branching events therefore increase autocatalytically the number of growing filaments. Indeed, kinetic assays show an increase of actin polymerization in presence of Arp2/3 and the activator. However, whether branching occurs at the barbed end or at the side of the filament is still under debate. Contradictory bulk measurements indicate that free barbed ends of mother filaments are required (Pantaloni et al., 2000; Falet et al., 2002) or not (Mullins et al., 1998) and led to two different proposed mechanisms of branching: the barbed-end branching and the side-branching mechanisms. In the first model, the ternary complex binds to the barbed-end of the mother filament and the two resulting daughter branches have the same length (Pantaloni et al., 2000). In the second model, the

13

Visualization of Individual Actin Filament Assembly

329

Fig. 13.9 Timelapse of dendritic structure growth (195 s between images): 1 μM actin is polymerized in presence of 25 nM Arp2/3 complex and 250 nM VCA

ternary complex binds to the side of the filament, initiating new filaments from the sides of older filaments (Mullins et al., 1998). To confront the two models, branched structures were observed in microscopy. The first assay was done using phalloidin-stabilized filaments and showed sidebranching events (Blanchoin et al., 2000). In the following, Amann and Pollard in 2001 developed a TIRFM assay to observe, for the first time, branching events in real time (Amann and Pollard, 2001). Their results showed daughter branches shorter than the mother ones, supporting thus the side-branching model (Fig. 13.9). So far, TIRFM measurements of filament branching do seem to agree with sidebranching model. However, the results could be affected by the 2-dimensional nature of the assay. Indeed, in the published experiments, filaments have always been anchored to surface via NEM-myosin: the strong attachment of the helical filaments to the surface could hinder the growth of a daughter branch in the direction of the surface and delay its apparition. Also, side branching should generate an increasing density of branches on the older regions of mother filaments leading to structures more densely branched towards their point of origin. On the contrary, the structures observed in TIRFM look denser towards the growing ends of the arrays (Fig. 13.8) and different microscopy studies indeed show that filaments generate more branches on newly polymerized regions than on older regions (Amann and Pollard, 2001; Blanchoin et al., 2000; Ichetovkin et al., 2002; Mahaffy and Pollard, 2006). The branching mechanism therefore still requires more investigations to combine all these apparently contradictory observations and get a consistent understanding.

13.2.2.2 Dissociation of Actin Filament Branches The actin meshwork located at the leading edge of the lamellipodium has a constant width, indicating a constant renewal of the actin array. Whereas new filaments are generated by branching at the front, disassembly must occur at the rear to recycle the actin meshwork. Though there is no physiological evidence of debranching in the lamellipodium, branch dissociation has been proven to be essential in budding yeast for efficient endocytic internalization by a similar actin meshwork: this is the unique in vivo observation of the importance of debranching (Martin, 2006).

330

E. Helfer

As stated above, the Arp2/3 complex nucleates filaments by branching. Le Clainche et al., using a combined kinetic and microscopy study, showed that the branch dissociates with a half-time of 13 min (Le Clainche et al., 2003). However, the lifetime of the branched junction was derived from indirect measurements and was not observed in real-time. TIRFM should allow such a direct observation. Using evanescent wave microscopy, Mahaffy et al measured the branch frequency per μm along the mother filament and derived a similar debranching half-time but did not observe debranching events (Mahaffy and Pollard, 2006). It seems the process of branch release in standard conditions (actin polymerized in presence of Arp2/3 complex and its activator) is barely observed in TIRFM assays, even when filaments are not attached to the surface (personal communication). However, when enhanced, the dissociation of the branch can be observed using TIRFM. Recently, Cai et al. indeed showed that Coronin1B, known as a regulator of cell motility, destabilizes filament branches (Cai et al., 2008): they succeeded observing debranching events and could calculate a debranching frequency increased in presence of Coronin1B (Fig. 13.10). In absence of Coronin1B, they noted that most branches remain stable and observed very few debranching events. Chan et al. as well observed branch release in presence of ADF/cofilin and concluded that cofilin dissociates Arp2/3 complex and therefore the branch from the mother filament (Chan et al., 2009). The apparent stabilization of the branched junction is in contradiction with the debranching process suggested by bulk assays and branch density measurements. Moreover, it is also not coherent with the results of Mahaffy et al who suggested that branches are less stable on ADP-filaments than on newly polymerized ATPsegments (Mahaffy and Pollard, 2006). Branches grown from mother filaments remain attached for tens of min, i.e. bound to subunits on which the nucleotide

Fig. 13.10 Debranching frequency measured from TIRFM imaging of growth and detachment of branches. (from Cai et al. (2008))

13

Visualization of Individual Actin Filament Assembly

331

has been for sure hydrolyzed to ADP. Therefore, whether the stabilization of the branches is an artifact due to TIRFM experimental conditions or not is still unknown and needs more investigation. For example, in TIRFM assays, methylcellulose is required to maintain the filaments in the evanescent wave. It also inhibits the 2Dmotion of the actin filaments in the plane parallel to the surface and prevents strong mechanical constraints on the branched junctions due to Brownian motion of the two branches. In kinetic assays, the branches can move freely in bulk and the procedure before microscopy observation of the branched structures can induce mechanical breakage of the branched junctions. The two distinct methods could explain the contradictory results on debranching.

13.3 Conclusion and Perspectives Controlled assembly and disassembly of actin arrays are crucial for motile cellular processes. The recent improvements in imaging techniques allow now direct observation in real-time of individual actin filaments. TIRFM is a powerful technique that allows in vitro study of the function of proteins involved in the regulation of actinbased processes (Breitsprecher et al., 2009). TIRFM observations combined with bulk solution studies will lead to a better global understanding of actin structure dynamics in motile cells. The role of the ATP hydrolysis in the dynamics of a single actin filament has been studied theoretically. Three different models of hydrolysis (random, vectorial, and cooperative) have been proposed theoretically (Carlier et al., 1987; Pollard and Weeds, 1984; Li et al., 2009) and need experimental validation. The TIRFM technique is a powerful tool to investigate the filament behavior (polymerization/depolymerization) as function of the nucleotide composition at the barbed end cap, like the length fluctuations of treadmilling actin filaments observed by Fujiwara et al. (2002). The branching mechanism is still under debate and requires more investigations to be fully elucidated. A close time analysis of the branching distribution along the mother filaments should allow to discriminate the side- and end-branching mechanisms. A very challenging experiment is the tracking of labeled Arp2/3 complex during the branching event, whiwh would provide the localization of the branching complex on the mother filament. TIRFM, like standard fluorescence microscopy, is diffraction-limited to a spatial resolution of about 200 nm. On the other hand, localization-base techniques like PALM (Betzig et al., 2006) or STED (Willig et al., 2006) allow much better resolution, however with a low acquisition rate or a limited field of view. The technique of structured-illumination microscopy (SIM) allows to go beyond the diffraction limit by illumination with multiple interfering beams (Gustafsson, 2000). Though much faster than PALM and with fields of view larger than that observed with STED, it appeared to be still too slow for live 3D-imaging (Schermelleh et al., 2008). However, in a recent paper, the combination of SIM and TIRFM proved to reach

332

E. Helfer

100-nm resolution and acquisition frame rate of 11 Hz (Kner et al., 2008). One can expect this new method will be further developed in the following years to answer biological questions currently investigated using conventional TIRFM. Acknowledgements The author thanks L. Blanchoin for careful manuscript reading and helpful discussions.

References Amann, K.J. and T.D. Pollard, Direct real-time observation of actin filament branching mediated by Arp2/3 complex using total internal reflection fluorescence microscopy. PNAS, 2001. 98(26): 15009–13. Betzig, E., et al., Imaging intracellular fluorescent proteins at nanometer resolution. Science, 2006. 313(5793): 1642–45. Blanchoin, L., et al., Direct observation of dendritic actin filament networks nucleated by Arp2/3 complex and WASP/Scar proteins. Nature, 2000. 404(6781): 1007–11. Blanchoin, L., T.D. Pollard, and R.D. Mullins, Interactions of ADF/cofilin, Arp2/3 complex, capping protein and profilin in remodeling of branched actin filament networks. Curr Biol, 2000. 10(20): 1273–82. Breitsprecher, D., et al., Analysis of actin assembly by in vitro TIRF microscopy. Methods Mol Biol, 2009. 571: 401–15. Cai, L., et al., Coronin 1B antagonizes cortactin and remodels Arp2/3-containing actin branches in lamellipodia. Cell, 2008. 134(5): 828–42. Carlier, M.F., D. Pantaloni, and E.D. Korn, The mechanisms of ATP hydrolysis accompanying the polymerization of Mg-actin and Ca-actin. J Biol Chem, 1987. 262(7): 3052–59. Chan, C., C.C. Beltzner, and T.D. Pollard, Cofilin dissociates Arp2/3 complex and branches from actin filaments. Curr Biol, 2009. 19(7): 537–45. Evangelista, M., et al., Bnilp, a yeast formin linking cdc42p and the actin cytoskeleton during polarized morphogenesis. Science, 1997. 276(5309): 118–22. Falet, H., et al., Importance of free actin filament barbed ends for Arp2/3 complex function in platelets and fibroblasts. Proc Natl Acad Sci U S A, 2002. 99(26): 16782–87. Fujiwara, I., et al., Microscopic analysis of polymerization dynamics with individual actin filaments. Nat Cell Biol, 2002. 4(9): 666–73. Gustafsson, M.G., Surpassing the lateral resolution limit by a factor of two using structured illumination microscopy. J Microsc, 2000. 198(Pt 2): 82–7. Gutsche-Perelroizen, I., et al., Filament assembly from profilin-actin. J Biol Chem, 1999. 274(10): 6234–43. Hertzog, M. and M.F. Carlier, Functional characterization of proteins regulating actin assembly. Curr Protoc Cell Biol, 2005. Chapter 13: p. Unit 13 6. Ichetovkin, I., W. Grant, and J. Condeelis, Cofilin produces newly polymerized actin filaments that are preferred for dendritic nucleation by the Arp2/3 complex. Curr Biol, 2002. 12(1): 79–84. Kner, P., et al., Super-resolution video microscopy of live cells by structured illumination. Nat Meth, 2008. 6(5): 339–42. Kovar, D.R. and T.D. Pollard, Insertional assembly of actin filament barbed ends in association with formins produces piconewton forces. Proc Natl Acad Sci U S A, 2004. 101(41): 14725–30. Kueh, H.Y., W.M. Brieher, and T.J. Mitchison, Dynamic stabilization of actin filaments. Proc Natl Acad Sci U S A, 2008. 105(43): 16531–36. Kuhn, J.R. and T.D. Pollard, Real-time measurements of actin filament polymerization by total internal reflection fluorescence microscopy. Biophys J, 2005. 88(2): 1387–1402. Le Clainche, C., D. Pantaloni, and M.-F. Carlier, ATP hydrolysis on actin-related protein 2/3 complex causes debranching of dendritic actin arrays. PNAS, 2003. 100(11): 6337–42.

13

Visualization of Individual Actin Filament Assembly

333

Li, H., et al., Automated actin filament segmentation, tracking, and tip elongation measurements based on open active contour models. Proceeding of the IEEE International Symposium on Biomedical Imaging: from Nano to Macro, 2009. Li, X., J. Kierfeld, and R. Lipowsky, Actin polymerization and depolymerization coupled to cooperative hydrolysis. Phys Rev Lett, 2009. 103(4): 048102. Mahaffy, R.E. and T.D. Pollard, Kinetics of the formation and dissociation of actin filament branches mediated by Arp2/3 complex. Biophys J, 2006. 91(9): 3519–28. Martin, A.C., M.D. Welch, and D.G. Drubin, Arp2/3 ATP hydrolysis-catalysed branch dissociation is critical for endocytic force generation. Nature Cell Biology, 2006. 8(8): 826–33. Michelot, A., et al., A novel mechanism for the formation of actin-filament bundles by a nonprocessive formin. Curr Biol, 2006. 16(19): 1924–30. Mullins, R.D., J.A. Heuser, and T.D. Pollard, The interaction of Arp2/3 complex with actin: Nucleation, high affinity pointed end capping, and formation of branching networks of filaments. PNAS, 1998. 95(11): 6181–86. Pantaloni, D., et al., The Arp2/3 complex branches filament barbed ends: functional antagonism with capping proteins. Nature Cell Biology, 2000. 2(7): 385–91. Paul, A.S. and T.D. Pollard, The role of the FH1 domain and profilin in formin-mediated actinfilament elongation and nucleation. Curr Biol, 2008. 18(1): 9–19. Peng, J., et al., Disruption of the Diaphanous-related formin Drf1 gene encoding mDia1 reveals a role for Drf3 as an effector for Cdc42. Curr Biol, 2003. 13(7): 534–45. Pollard, T.D. and A.G. Weeds, The rate constant for ATP hydrolysis by polymerized actin. FEBS Lett, 1984. 170(1): 94–8. Pollard, T.D. and J.A. Cooper, Quantitative analysis of the effect of Acanthamoeba profilin on actin filament nucleation and elongation. Biochemistry, 1984. 23(26): 6631–41. Pollard, T.D. and J.A. Cooper, Actin and actin-binding proteins. A critical evaluation of mechanisms and functions. Annu Rev Biochem, 1986. 55: 987–1035. Romero, S., et al., Formin Is a Processive Motor that Requires Profilin to Accelerate Actin Assembly and Associated ATP Hydrolysis. Cell, 2004. 119(3): 419–29. Sagot, I., S.K. Klee, and D. Pellman, Yeast formins regulate cell polarity by controlling the assembly of actin cables. Nat Cell Biol, 2002. 4(1): 42–50. Schermelleh, L., et al., Subdiffraction multicolor imaging of the nuclear periphery with 3D structured illumination microscopy. Science, 2008. 320(5881): 1332–36. Willig, K.I., et al., STED microscopy reveals that synaptotagmin remains clustered after synaptic vesicle exocytosis. Nature, 2006. 440(7086): 935–39.

Chapter 14

Movement of Cargo in Bacterial Cytoplasm: Bacterial Actin Dynamics Drives Plasmid Segregation Dyche Mullins

Abstract Actin was once considered an exclusive hallmark of eukaryotic cells. In 2001, however, a filament-forming actin homolog was discovered in eubacteria as a regulator of cell shape (Jones, 2001). Since that time, other bacterial actin filaments have been shown to segregate DNA (Moller-Jensen, 2002; Becker, 2006; Derman, 2009) and position organelles (Komelli, 2006). Recent bioinformatic analysis has identified dozens of families of Actin Like Proteins (ALPs) encoded on prokaryotic chromosomes, plasmids, and phages. While the overwhelming majority of these proteins remains uncharacterized, two plasmid-segregating bacterial actins, ParM and AlfA, have been studied in vitro. The architecture and assembly dynamics of these filaments and their mechanisms of force generation and cargo translocation are examined in this chapter.

Contents 14.1 Filament Architecture and Assembly Dynamics . . . . . 14.2 ParM Assembly Dynamics . . . . . . . . . . . . . . 14.2.1 Fast Spontaneous Nucleation . . . . . . . . . . 14.2.2 Bidirectional Elongation . . . . . . . . . . . . 14.2.3 Dynamic Instability . . . . . . . . . . . . . . 14.3 Harnessing Polymerization for Movement . . . . . . . 14.4 Using a Brownian Ratchet to Find the Long Axis of a Cell 14.5 Insertional Polymerization . . . . . . . . . . . . . . 14.5.1 Segrosome Structure and Function . . . . . . . 14.5.2 Ring Model . . . . . . . . . . . . . . . . . 14.5.3 Clamp Model . . . . . . . . . . . . . . . . 14.6 Multiple Roles for Dynamic Instability . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

338 339 339 339 340 342 343 344 345 345 347 348

D. Mullins (B) UCSF School of Medicine, N312F Genentech Hall, 600 16th Street, San Francisco CA, USA e-mail: [email protected]

M.-F. Carlier (ed.), Actin-based Motility, DOI 10.1007/978-90-481-9301-1_14, C Springer Science+Business Media B.V. 2010

335

336 14.6.1 Plasmid Counting . . . . . . . . . . . . . . 14.6.2 Search and Capture . . . . . . . . . . . . . 14.6.3 Energetics of ParM-Mediated DNA Segregation 14.7 AlfA: DNA Segregation Without Dynamic Instability . 14.8 Summary and Conclusions . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . .

D. Mullins

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

348 349 350 351 353 354

Since its discovery in the 1940s (Straub, 1942) actin has succeeded in slipping every taxonomic noose thrown around it. Initially thought to be a muscle-specific protein, the discovery of actin in non-muscle cells (Yang, 1972) and in protozoa (Hatano, 1966) in the 1960s demonstrated that it was not limited to a single tissue or a single eukaryotic phylum. After its primary sequence was determined in the 1970s (Elzinga, 1971; Collins, 1975) actin was, for many years, thought to be evolutionarily isolated, with no remaining, close relatives. This splendid isolation ended in the genome sequencing projects of the early 1990s with the discovery of at least ten widely expressed and highly conserved Actin Related Proteins (Poch, 1997). By 2000, the only frontier remaining for actin was the boundary between eukaryotic and prokaryotic domains of life, and there appeared to be good reasons to imagine that actin was strictly limited to eukaryotes: (1) the small size of most eubacteria and archaea suggested that cellular components move by diffusion and do not require active transport; (2) bacterial cell shape is generally maintained by rigid cell walls with no apparent need for cytoskeletal scaffolds; (3) eubacteria and archaea generally do not maintain complex, internal compartments; and (4) fifty years of cytology through the electron microscope failed to reveal obvious cytoskeletal networks in bacteria. These reasons are no longer compelling: (1) Some cellular components, such as large DNA molecules, diffuse so slowly that, on the time scale of cell division, they barely move at all (Campbell, 2007). (2) Cells with non-trivial shapes (rods, helices, etc.) must asymmetrically localize their cell wall synthesis machinery, a process that implies the existence of long range order in the cytoplasm (Jones, 2001; Cabeen, 2009). (3) Several bacterial phyla are now known to contain internal, membrane-bounded structures (Shively, 2006). (4) Modern cryo-electron microscopy and image analysis now reveal organized filament networks inside bacterial cells (Kürner, 2004; Jensen, 2009). The Errington lab first demonstrated that relatives of eukaryotic actin form cytoskeletal filaments in eubacteria and that these filaments help determine cell shape (Jones, 2001). This was followed by the discovery of eubacterial actins that segregate DNA (Moller-Jensen, 2002; Becker, 2006; Derman, 2009) and align subcellular organelles (Komelli, 2006). Genome analysis has now identified more than forty classes of Actin Like Proteins (ALPs) from bacterial and archaeal sources (Derman, 2009). Only seven of these classes have been studied to date and all have been found to form filaments (Fig. 14.1). It is, therefore, likely that many of the uncharacterized ALPs will also form actin-like filaments. Some ALPs are encoded on mobile elements, including plasmids and phages, while others are encoded by chromosomal genes. The level of similarity used to define an ALP family was 20%, meaning that there is substantial sequence diversity both within and between families. The ParM (a.k.a. ALP3) family contains members that are only 40% identical

14

Movement of Cargo in Bacterial Cytoplasm

Fig. 14.1 Bacterial actin family tree. Twenty six of forty-one bacterial families of actin-like proteins identified by Derman et al. (2009). The distance from the branch points is inversely related to the level of sequence similarity but the low similarity between families (20%) makes evolutionary relationships difficult to establish. The seven families marked with a red asterisk have been shown to form actin-like filaments

337 Alp34 Alp8* Alp37 Alp6* Alp41 Alp40 Alp23 Alp9 Alp24 Alp20 Alp15 Alp14 Alp13 Alp22 Alp7* Alp30 Alp16 Alp27 Alp5 (AlfA)* Alp32 Alp4 (MamK)* Alp1 (MreB)* Alp3 (ParM)*

to each other. By comparison, conventional actins from widely separated species (e.g. human and Acanthamoeba) are more than 90% identical. The architectures of filaments formed by ALPs are also remarkably diverse and differ dramatically from that of eukaryotic actin (Fig. 14.2).

Fig. 14.2 Comparison of the structures of actin, ParM, and AlfA filaments. Top: envelopes of protein density derived from electron microscopy and 3D image reconstruction (Polka, 2009). Crystal structures of actin and ParM have been fitted into the envelope to help establish the position and orientation of individual subunits. Bottom: schematic representation of adjacent subunits in the same long-pitch helical strand of each filament. The relative orientation of the nucleotide binding pocket and the degree of rotation required to move from monomer to monomer are indicated

338

D. Mullins

The most well-studied bacterial actins are MreB, ParM, and AlfA (Amos, 2004; Löwe, 2004; Carballido-López, 2006; Needleman, 2008; Polka, 2009). MreB controls cell shape by asymetrically localizing cell wall synthesis machinery but, despite work from several laboratories, the assembly properties of MreB filaments are not well understood. Recent work (Mayer, 2009), in fact, suggests that MreB filaments from Bacillus do not assemble spontaneously under physiological concentrations and may require additional, as yet unknown, factors to promote assembly in vivo. Both ParM and AlfA function to segregate plasmid DNA and both have been studied in vitro. For the remainder of this chapter I will focus on the architecture and assembly properties of these two bacterial actins and the mechanisms by which they generate force and move cargo through cytoplasm.

14.1 Filament Architecture and Assembly Dynamics Like eukaryotic actin, both ParM and AlfA form two-stranded, helical filaments in the presence of ATP. When purified away from accessory factors, actin, ParM, and AlfA all self-assemble into filaments in two phases: (1) a slow, nucleation phase during which monomers self-associate to form stable filament nuclei; and (2) a fast phase in which filaments rapidly elongate. Apart from these basic properties, however, the architecture and assembly dynamics of the three filaments are remarkably different. The two strands of the actin filament wind around each other to form a right-handed helix while the ParM and AlfA helices are left-handed. Despite the difference in handedness, the contact surfaces between subunits in a ParM filament are similar to those in a conventional actin filament and the pitch of the helices is similar. Standing on one subunit of an actin filament, we step to the next subunit on the same strand by moving 5.4 nm along the filament axis and rotating around this axis 27◦ to the right. We hop from subunit to subunit along one strand of a ParM filament by moving 5.4 nm and rotating 30◦ to the left (Orlova, 2007; Galkin, 2009). The strands of an AlfA filaments are much more tightly wound than those of either actin or ParM. The rotation required to step along an AlfA filament is 46◦ , 50% greater than that required for ParM or actin. In addition, electron microscopic analysis of AlfA filaments indicates that the individual AlfA protomers are rotated inward by approximately 45◦ toward the filament axis compared to both actin and ParM. One important implication of this rotation is that the subunit contact sites in an AlfA filament are unique among actin like proteins studied to date. Also unique is the fact that AlfA filaments spontaneously assemble into higher-order bundles. The filaments in an AlfA bundle have a mixture of polarities but, within bundles, the filaments are remarkably well aligned, suggesting that bundling may be due to specific interactions between binding sites on the filaments rather than non-specific, charged residue interactions (Polka, 2009). In addition to ultrastructural differences, the pathway by which AlfA filaments assemble is different from those of actin and ParM. For both actin and ParM, three monomers must associate to form a filament nucleus (the smallest oligomer with a

14

Movement of Cargo in Bacterial Cytoplasm

339

higher probability of growing than shrinking). In contrast, assembly kinetics reveal that AlfA filament nuclei are composed of four subunits. Because nucleus size depends on the number of inter-subunit contacts required to form a stable structure, the larger AlfA nucleus is likely a consequence of the unique architecture of the AlfA filament and the unique nature of its subunit contacts. Together, these results indicate that bacterial actins are not only diverse at the level of primary sequence but that they also assemble into polymers with remarkably diverse architectures. More structural studies will be required to understand the extent and nature of this diversity.

14.2 ParM Assembly Dynamics In eukaryotes, the location, architecture and stability of cytoskeletal structures are determined by accessory factors that control filament nucleation, crosslinking, and disassembly. Plasmid-encoded actins appear to have only one function and do not rely on a large number of accessory factors to modulate their assembly. Because of the stripped-down nature of plasmid segregation machinery, information required to form plasmid-segregating structures must be encoded in the actin-like polymer itself. Three properties differentiate ParM filament assembly from that of conventional actin and appear to be essential for its cellular function: (1) fast spontaneous nucleation, (2) bidirectional elongation, and (3) dynamic instability.

14.2.1 Fast Spontaneous Nucleation Spontaneous nucleation of ParM filaments is more than 300-fold faster than that of conventional actin. For comparison, adding ATP to 5 μM ParM induces filament formation that reaches steady state in less than 10 s (Garner, 2004). Under similar conditions, 5 μM actin requires more than an hour to reach steady state. Cellular concentrations of ParM are estimated to be around 15 μM (Moller-Jensen, 2002) so, in vivo, ParM filaments will rapidly partition into steady state pools of monomer and polymer. One consequence of fast nucleation is that, unlike actin filaments, ParM filaments may not require a dedicated nucleation factor to carry out their cellular function. It is still unclear whether ParM’s DNA-protein cargo promotes nucleation or simply captures the spontaneously formed filaments.

14.2.2 Bidirectional Elongation Electron microscopy (Orlova, 2007) and x-ray fiber diffraction (Popp, 2008) reveals that ParM filaments are structurally polarized: one end is analogous to the fast-growing barbed end of an actin filament while the other is similar to the slowgrowing pointed end. Unlike eukaryotic actin, however, the elongation of ParM

340

D. Mullins

filaments is not polarized (Garner, 2004). Both ends of a ParM filament elongate at the same diffusion-limited rate (similar to that of the fast-growing end of an actin filament). This difference in kinetic polarity probably reflects functional differences between actin and ParM. Conventional actin filaments in eukaryotic cells construct polarized networks responsible for directional movement of cargo and protrusion of the leading edge. The polarized filament assembly that underlies construction of these networks has either removed the evolutionary pressure for both ends of the filament to elongate at the same rate or has selected against rapid growth from the pointed end. In contrast, to ensure inheritance by both daughter cells, individual ParM filaments must interact with two plasmids (Choi, 2008; Salje, 2008), one at each end of the filament. This bivalent attachment serves as a counting mechanism to ensure that two plamsids are pushed in opposite directions (see below). Since the two ends of the filament perform equivalent functions, there is either a lack of evolutionary pressure selecting for polarized assembly or active selection against it. The functional significance of polarized filament assembly would be difficult to test in eukaryotes, but potential deleterious effects of polarized assembly could be tested in a plasmid segregation system.

14.2.3 Dynamic Instability Probably the most interesting property of ParM filaments is that they switch between two biochemical states: either steadily elongating, or rapidly shortening. This behavior was first described in eukaryotic microtubules by Mitchison and Kirschner who called it dynamic instability (Mitchison, 1984). Dynamic instability probably contributes to several aspects of ParM function in vivo and I will discuss these potential roles in a later section. First, I will focus on what we know about the molecular mechanism underlying dynamic instability of ParM filaments. The switch between growth and shortening of ParM filaments is driven by hydrolysis of ATP bound to ParM protomers. Non-hydrolyzable ATP analogs and point mutations in ParM that ablate ATP hydrolysis lock filaments in the stable state (Garner, 2004). Two pieces of evidence suggest that the most important determinant of stability is the nucleotide state at the end of the filament: (1) Small amounts of hydrolysis-defective ParM mutant can stabilize filaments composed mainly of wildtype ParM. Maximum stabilization occurs when 20% of the total ParM is nonhydrolyzing mutant. (2) Depolymerizing ParM filaments that contain small amounts of non-hydrolyzing mutant occasionally switch back to stable elongation, a phenomenon known as rescue. Since rescue is observed at fractions of non-hydrolyzing mutant as low as 2%, the interpretation of this result is that sites of rescue represent clusters of non-hydrolyzing mutant ParM protomers that cap filaments and stabilize them against further depolymerization. We are beginning to understand the structural basis of dynamic instability in ParM filaments. Work from the Lowe lab has produced atomic structures of two conformations of ParM in two nucleotide states (van den Ent, 2002). In one

14

Movement of Cargo in Bacterial Cytoplasm

341

structure, ParM lacks nucleotide altogether (apo-ParM) and the “nucleotide binding pocket” (a gap between the two major domains of the protein) is open. In the other structure, the two domains of the protein are closed around a nucleotide diphosphate (ADP-ParM). The opening and closing of the nucleotide binding pocket appears to be the major conformational change common to members of the actin superfamily (Bork, 1992) and it is likely to control the assembly dynamics of all actin like filaments. Evidence that this conformational change controls dynamic instability of ParM filaments comes from electron microscopy. By binning electron microscope images based on morphology, the Egelman group (Galkin, 2009) discovered that ParM filaments also display two distinct conformations: “smooth” and “ragged”. The ParM crystal structure (van den Ent, 2002) with a closed nucleotide binding pocket fits into the “smooth” structure, while an atomic structure of ParM with an open nucleotide binding pocket is a better fit for the “ragged” conformation. The Egelman group, therefore, referred to the two filament conformations as “closed” and “open”. Further evidence that the different ParM filament conformations reflect different conformational states of the nucleotide binding pocket comes from the fact that the fraction of filaments found in each conformation depends on which nucleotide is bound. Filaments assembled with a non-hydrolyzable ATP analog (AMPPNP) are almost entirely in the the closed state while filaments polymerized in the presence of ATP and partially stabilized by addition of the phosphate analog BeF3 contain more regions in the open state. At first glance it is difficult to reconcile results from electron microscopy with those derived from x-ray crystallography. Crystallography suggests that the nucleotide binding cleft is open in apo ParM and closed in ADP ParM while electron microscopy suggests that the conformation of ADP (and/or ADP · Pi) ParM is open while ATP ParM is closed. One parsimonious explanation of these data is that ParM monomers are partitioned between the open and closed states with an equilibrium that depends on bound nucleotide. In this model, ATP strongly favors the closed state while the absence of nucleotide produces a strong preference for the open state. ADP ParM may partition more evenly between the two conformations. This partitioning is consistent with the electron microscopy data. During crystallization of ADP ParM, packing conditions may favor incorporation of molecules in the closed conformation. This is consistent with crystal structures of the Arp2/3 complex in which the Actin Related Protein, Arp3, exists in a closed conformation when bound to ATP but, depending on crystallization conditions, can adopt either a closed or open conformation when bound to ADP (Nolen, 2007). Hydrolysis-dependent opening of the nucleotide binding pocket provides a simple explanation for dynamic instability. Opening of the cleft would significantly reduce the surface are buried in subunit contacts and decrease the strength of protomer-protomer interactions (Fig. 14.3). The effect of this decreased binding energy would be most pronounced on subunits at the end of the filament which make the smallest number of contacts with adjacent subunits. Additional loss of binding energy at this position would significantly decrease affinity and produce rapid, hydrolysis-dependent, endwise depolymerization.

342

D. Mullins

Fig. 14.3 Possible structural basis of dynamic instability of ParM filaments. Top: envelope of ParM filaments determined from electron microscopy (Galkin, 2009). On the left is the “closed” conformation and on the right is the “open” conformation. The ratio of filament segments found in each conformation depends on the nucleotide state of the filament. Non-hydrolyzable nucleotides favor the closed conformation while the presence of hydrolyzable ATP correlates with an increase in the open conformation. Crystal structures of ParM monomers in open and closed conformations (van den Ent, 2003) are fitted into the EM structures. Opening of the nucleotide binding pocket appears to promote a significant rotation of subdomains 3 and 4 away from the filament axis. Bottom: schematic representation of the consequences of the nucleotide-dependent conformational change. Rotation of the mobile subdomains of the protein significantly reduces the surface area buried in protein–protein contacts that stabilize the filament. The consequences of this loss of binding energy are most pronounced when the subunit sits at the end of the filament

14.3 Harnessing Polymerization for Movement The Gerdes group first proposed that ParM assembly drives segregation of the R1 plasmid (Moller-Jensen, 2002, 2003; Gerdes, 2004). Imaging of polymers and plasmids in cells (Campbell, 2007; Salje, 2009) and the reconstitution of plasmid-polymer interaction in vitro (Garner, 2007; Salje, 2008) have demonstrated more-or-less conclusively that this model is correct. In the simplest mode of segregation, one plasmid binds to each end of a ParM filament. The filament elongates in both directions and, because the two plasmids remain associated with the growing filament ends, they are pushed apart (Fig. 14.4). Coupling between the plasmid and the ParM polymer is mediated by a DNA-protein complex composed of a cetromeric

14

Movement of Cargo in Bacterial Cytoplasm ParM monomer

343 plasmid DNA

ParR/parC complex

ParM filament

Fig. 14.4 Type II plasmid segregation driven by components encoded by the par operon: ParM, ParR, and parC. Assembly of ParM monomers into filaments drives bidirectional transport of plasmids. A protein-DNA complex (the ParR/parC or segrosome complex) binds to both ends of single, polar ParM filament (Choi, 2008; Salje, 2008) and converts the free energy of polymerization into useful work

DNA sequence, parC, and a protein, ParR (see below). The parC locus and the genes encoding ParR and ParM form an operon (the partitioning or par operon) that is necessary and sufficient for segregation.

14.4 Using a Brownian Ratchet to Find the Long Axis of a Cell Plasmid R1 is generally found in rodshaped enteric pathogens (Ebersbach, 2005) and cell shape appears to be an important determinant of its ability to segregate DNA. In general, DNA segregation systems must position two pieces of DNA on either side of the plane of cell division. Animal cells solve this problem during chromosome segregation by using the position and orientation of the mitotic spindle to specify the location and orientation of the division plane. In eubacteria the position of the chromosome can influence the site of division (Woldringh, 1991) but generally this is determined by the geometry of the cell wall and the assembly

344

D. Mullins

of cytoskeletal elements. In rod-shaped bacteria, for example, the division plane is perpendicular to the long axis of the cell and its orientation is determined by circumferential assembly of filaments composed of a tubulin homolog, FtsZ. The division site is generally located at the midpoint of the long axis and its position is controlled by factors that promote or disrupt FtsZ ring assembly (e.g. MinCDE in E. coli or MreB in C. crescentus). To place plasmids on opposite sides of the division plane, the par system must push them in opposite directions along the long axis of the cell. How does the plasmid segregation machinery align with the long axis? Either: (1) specific molecular landmarks align the spindle; or (2) alignment is a consequence of spindle elongation confined to a tube. The existence of molecular landmarks is unlikely because, in vivo, initiation of ParM spindle formation can occur anywhere in the cell, from the pole to the midpoint (Campbell, 2007). In addition, spindles often begin life orthogonal to the long axis. When elongation brings them into contact with the edges of the cell, further elongation appears to drive re-orientation parallel to the long axis. These observations argue that orientation of ParM spindles perpendicular to the plane of cell division is a consequence of the rod-like shape of the host cells and elongation-dependent alignment of the spindle. Alignment is consistent with a Brownian ratchet-type mechanism in which ParM polymerization biases small, random fluctuations in spindle position in a direction that accommodates maximum polymer incorporation (Hill, 1982).

14.5 Insertional Polymerization The coupling between plasmid and polymer is mediated by a DNA-protein complex composed of a set of 10 DNA sequence repeats in the parC locus of the plasmid bound to 10 copies of a dimeric repressor protein, ParR. To carry out its cellular function the ParR/parC complex must possess several properties whose molecular origins are poorly understood: (1) It must mediate processive insertional polymerization. In other words, the complex must interact sequentially with many different ParM protomers at the end of an elongating filament. For example, if a ParM filament elongates at 13 subunits per seconds (the rate of elongation measured in vivo; Campbell, 2007) and the ParR/parC complex remains attached for 45 s (the time requred for the filament to grow 1.5 μm from the attached end) the ParR/parC complex must interact sequentially with 550 newly added subunits without detaching from the filament. This activity resembles that of other cytoskeletal polymerases, including formin family proteins which promote insertional polymerization of actin filaments and the kinetochore proteins that mediate attachment of eukaryotic chromosomes to growing and shortening microtubules. Compared to a formin, which can remain attached to an actin filament for more than 1000 s and mediate incorporation of many thousands of actin monomers, the processivity required of the ParR/parC complex is modest. (2) Binding of the ParR/parC complex must stabilize ParM filaments and prevent their catastrophic disassembly. This property has been observed in vitro by demonstrating that ParR/parC coated particles can stabilize

14

Movement of Cargo in Bacterial Cytoplasm

345

ParM filaments and enable them to grow many times longer than unattached filaments (Garner, 2004; Garner, 2007). (3) Most remarkably, perhaps, the ParR/parC complex must be able to interact productively with both ends of a polarized ParM filament. This implies that two ParR/parC complexes move in opposite directions at the two ends of the same elongating filament. Formins, in contrast, interact very specifically with the barbed end of an actin filament and do not bind at all to the pointed end. Generally speaking, two models could account for the unique polymrase activity of the ParR/parC complex. I will call them the “clamp” model and the “ring” model, and they are based, in part, on two different interpretations of the structure of the ParR/parC complex.

14.5.1 Segrosome Structure and Function Because of its role in DNA segregation, the ParR/parC complex is sometimes called the “segrosome”. Experiments with purified components have revealed that the the segrosome inhibits the dynamic instability of the filament to which it is bound. Dilution experiments indicate that segrosome-bound filaments have the same critical concentration as filaments bound to non-hydrolyzable nucleotide analogs or non-hydrolyzing ParM mutants bound to ATP. This result suggests that the complex may stabilize filaments by simply preventing hydrolysis of ATP on the subunits to which it is bound. We have three structures of isolated segrosomes: two based on x-ray crystallography and one on atomic force microscopy. The atomic structure of ParR/parC from plasmid pSK41 by the Schumacher group demonstrates that ten dimers of ParR bind cooperatively to the parC locus of the plasmid. Interaction between ParR dimers bound to the DNA bends the structure into a right-handed helix or “lock washer” shape, with ParR dimers forming the inside of the helix and DNA running around the outside (Fig. 14.5). A similar structure of ParR from plasmid pB171, solved by the Lowe laboratory in the absence of DNA, further demonstrates the role of ParR dimer-dimer interactions in determining the helical shape of the complex. In atomic force microscope images, however, segrosomes from plasmid R1 appear to adopt a more “U” shaped morphology (Hoischen, 2008). The ring and “U” structures suggest that the segrosome is somewhat flexible.

14.5.2 Ring Model In the ring model the ParR/parC complex encircles the ParM filament, similar to the way the budding yeast Dam1 complex is thought to encircle the end of a microtubule (Salmon, 2005). The simplest way for the complex to track the end of a growing filament in this configuration would be for it to interact preferentially with ATP-bound regions of the filament. This could occur in one of two ways: (1) The complex could encircle both ADP- and ATP-bound regions of the filament but have a much higher affinity for ATP-ParM. On ADP-bound regions it would undergo one

346

D. Mullins

Fig. 14.5 Structure of the ParR/parC complex determined by electron microscopy and x-ray crystallography. Data and figures adapted from Moller-Jensen et al. (2007). Left panels: electron micrographs of DNA containing the parC seqeunce in the presence of purified ParR protein. Contrast generated by rotary shadowing with platinum. DNA and the segrosome complex are marked in the upper panel. Upper right: “side” and “end” views of the helical ParR oligomer structure determined by x-ray crystallography. Lower right: space filling model of helical ParR oligomer with electrostatic potential displayed in color. Red: negative. Blue: positive. The DNA winds around the outside, along the electropositive surface

dimensional diffusion until it encountered an ATP-bound region. Continual incorporation of ATP ParM at the end of the filament and subsequent hydrolysis following a short time lag would act to rectify the one-dimensional diffusion of the complex in the direction of polymerization. (2) The conformational change in ParM induced by ATP hydrolysis could prevent the complex from fitting around the filament. In this scheme, the segrosome would undergo one dimensional diffusion on the ATP-bound regions of the filament near the growing end. This diffusion would be rectified by the hydrolysis-dependent conformational change that blocks its access to the older parts of the filament. In this scheme, if hydrolysis occurs between the complex and the end of the filament, the segrosome will become trapped and no longer track the filament end. This situation would be resolved when hydrolysis catches up with the end of the filament and induces depolymerization. Depolymerization would proceed until it reaches the ATP-protomers encircled by the segrosome. At this point the segrosome would either fall off the filament or the filament would be “rescued” and begin to elongate again. One important feature of the ring model is that it can easily account for the ability of the segrosome to bind either end of a ParM filament since a ring could fit around the filament in the same orientation at either end (Fig. 14.6a).

14

Movement of Cargo in Bacterial Cytoplasm

Fig. 14.6 Models for processive filament attachment of ParR/parC complex. (a) Ring model. In this model the helical segrosome fits around the ParM filament. Differential affinity for ATP- and ADP-bound ParM protomers keeps the segrosome positioned near the end of the filament. (b) Clamp model. Here the segrosome interacts only with the terminal subunits of the filament. Differential affinity for ATP and ADP protomers causes the complex to “step” from subunit to subunit on the growing filament end

347 ParR/parC

A

ATP ParM

ADP ParM

Ring Model ParR/parC

B

ATP ParM

ADP ParM

Clamp Model

14.5.3 Clamp Model The clamp model, proposed by Salje and Lowe, is somewhat analogous to the way in which formin-family proteins track the growing ends of conventional actin filaments. In this model the segrosome forms a cap that interacts with two or three monomers at the end of a ParM filament. The segrosome would bind tightly to ATP ParM protomers and dissociate from them upon ATP hydrolysis. In this model the hydrolysis-dependent dissociation of the segrosome from one of the terminal protomers would uncap it and allow a new, ATP-bound monomer to incorporate into the filament (Fig. 14.6b). Our current understanding of ParM filament assembly poses several problems for this model. (1) The model requires ADP-bound ParM protomers be exposed at the end of the filament while previous studies of ParM dynamics suggest that filaments with ADP protomers at the end undergo catastrophic disassembly. This could be resolved if hemi-attachment of the segrosome (to only one strand of the filament) is sufficient for stabilization, even when the unattached, terminal protomer is ADP-bound. (2) The rate of filament elongation in

348

D. Mullins

this model is strictly limited by the rate of ATP hydrolysis. The measured rate of ATP hydrolysis on ParM filaments is 0.2 s−1 while the rate of monomer incorporation at steady state (both in vivo and in vitro) is 60 times faster (13 s−1 ). If the clamp model is correct then attachment of the segrosome must significantly accelerate ATP hydrolysis on the terminal subunits of the filament. (3) To cap both ends of the filament the segrosome must be able to interact with the filament in two different orientations, rotated by 180◦ from each other. One potential solution to this problem would be for the domains that interact with the filament to be capable of rotating freely with respect to the rest of the complex. Additional structural and biophysical studies will be required before we completely understand how the segrosome maintains processive attachment to both ends of the ParM filament. At the moment it appears that the ring model poses fewer conceptual problems than the clamp.

14.6 Multiple Roles for Dynamic Instability The key property of ParM filaments that enables them to assemble DNA-segregating spindles with a minimal number of accessory factors is dynamic instability. Dynamic instability of mircotubules is thought to be important for their ability to rapidly reorganize from an interphase array into a bipolar, DNA-segregating spindle during mitosis (Verde, 1990) and to efficiently search space and find chromosomes (Holy, 1994). The enhanced ability to search space comes from the fact that a growing microtubule that misses its target because it is pointing in the wrong direction eventually becomes unstable and falls apart. The subunits liberated by the disassembly of this microtubule can now incorporate into a new microtubule and, thus, promote multiple microtubule excursions with multiple chances of hitting the target. Quantitative models suggest that dynamic instability can increase the rate at which the ends of microtubules search space by as much as 1000-fold (Holy, 1994). Dynamic instability underlies three fundamental features of ParM-mediated DNA segregation: (1) It ensure that pairs of plasmids are pushed in opposite directions – in effect, counting plasmids. (2) It promotes search and capture by driving individual plasmids on randoms walks through the cytoplasm and increasing the rate at which they encounter other plasmids. (3) Finally, and somewhat counter-intuitively, it provides the energy for pushing plasmids through the cytoplasm.

14.6.1 Plasmid Counting Counting is a simple consequence of bipolar dynamic instability and the linear nature of ParM filaments. In addition to symmetrical elongation, ParM filaments are also symmetrical with respect to catastrophic disassembly. This was demonstrated

14

Movement of Cargo in Bacterial Cytoplasm

349

B

A

0 plasmids (least stable) 1 plasmid (intermediate stability) 2 plasmids (most stable) plasmid counting

search and capture

Fig. 14.7 Two functions of dynamic instability in ParM-mediated plasmid segregation: (a) Plasmid counting: the fact that free filament ends undergo catastrophic disassembly means that ParM filaments are maximally stable only when bound to two plasmids. This ensures that maximal polymer-driven plasmid movements are used to push two plasmids in opposite directions. (b) Search and capture: monovalent attachment of a ParM filament to a single plasmid produces randomly oriented, transient movements of the plasmid. Reaction forces driving these movements are described in the text. The active, random walk produced by monovalent attachment increases the apparent diffusion coefficient of the plasmid by more than 10-fold and enables single plasmids to find partners for active segregation

directly using laser irradiation to sever ParM spindles connecting ParR/parC-coated particles (Garner, 2007). After severing, both newly exposed ends of the cut spindle depolymerize completely. The spindles in this experiment are mixed-polarity bundles of ParM filaments and each exposed end contains an equal number of “barbed” and “pointed” ends. If catastrophic disassembly were more likely at one end or proceeded more rapidly in one direction, the fluorescence of the two halves of the severed spindle would initially decrease by a factor of two before disappearing completely. The fact that filaments require bivalent attachment for maximum stability means that ParM assembly drives maximum plasmid translocation when two plasmids are attached. That is, filament assembly has the greatest chance of driving a plasmid to the pole of a rod-shaped cell when another plasmid is being driven to the opposite pole (Fig. 14.7a).

14.6.2 Search and Capture Filaments attached to a single plasmid are less stable than bivalently attached filaments but they can still elongate for a finite distance before undergoing catastrophic disassembly. If the free end of the filament interacts with an obstacle in the cell, elongation can produce transient movement of the plasmid. If, for example, the free

350

D. Mullins

end of a ParM filament contacts the plasma membranes two reaction forces could be transmitted to the plasmid, one based on the friction of the contact between the filament and the membrane and the other based on the force required to translate the filament sideways through the cytoplasm (Fig. 14.7b). Both forces are nonzero but difficult to calculate based on our current understanding of the nature of bacterial cytoplasm. The plasma membrane is fluid but studded with obstacles including transmembrane proteins anchored to the cell wall and ribosomes carrying out translationally coupled membrane insertion (transertion) (Woldringh, 2002). The force required to translate a filament through the cytoplasm will depend on several factors, including the local density of chromosomal DNA. The effect of these random “kicks” can be seen by comparing the motion of single, fluorescently labeled plasmids with and without the par operon. The random nature of these kicks gives plasmid motion the character of a random walk and, at the level of light microscopy, makes their motion indistinguishable from constrained diffusion (Fig. 14.7b). Specifically, in the absence of the par operon, a labeled mini-F plasmid (5 kbp) undergoes confined diffusion with an apparent average diffusion coefficient of 5 × 10−5 μm2 /s. The same plasmid containing the par operon moves much more rapidly through the cytoplasm, with an average diffusion coefficient that is more than ten times higher (Campbell, 2007). In addition, the distribution of apparent diffusion coefficients is much wider in the presence of the par operon and individual plasmids can move more than 100 times faster than plasmids lacking a ParM-based segregation system. Because of their extremely low diffusion coefficient plasmids lacking the par operon explore only a fraction of the cytoplasmic volume on the time scale of cell division (20–60 min). On the same time scale plasmids containing the R1 par operon explore the entire cell volume. In vivo, bipolar segregation appears to require plasmids to start out in close proximity so ParM-driven random walks appear to be essential for plasmids to encounter each other and initiate the segregation process.

14.6.3 Energetics of ParM-Mediated DNA Segregation The low barrier to spontaneous nucleation and the high cellular concentration of ParM suggest that, in vivo, the monomer-to-filament ratio will always be at steady state, regardless of the presence of the ParR/parC complex. Measurements of filament elongation in vivo are consistent with this idea (Campbell, 2007). The steady state condition poses a challenge to plasmid segregation, namely, how do plasmidattached filaments elongate quickly under conditions where the average rates of monomer addition and dissociation are equal? The answer is that selective elongation of plasmid-attached filaments is driven by dynamic instability of the unattached filaments. One way to see this is by comparing the steady state monomer concentrations in the absence and presence of dynamic instability. Elongation and disassembly of polymers that are not dynamically unstable is governed by a single

14

Movement of Cargo in Bacterial Cytoplasm

351

set of association and dissociation rate constants (k+ and k− ), and the concentration of free monomers at steady state is given by mss =

k− k+

At steady state the lengths of all these filaments will fluctuate slowly because the probabilities of adding and subtracting monomers are equal and are the same on all filaments. A dynamically unstable polymer, however, can exist in one of two states, either growing and shortening like the single-state polymer (stable) or shrinking catastrophically at a constant rate, kcat (unstable). If, at steady state, a fraction (f) of filaments is stable and the rest (1 − f) unstable, the steady-state monomer concentration becomes 1 k− kcat + −1 mss = k+ f k+ Dynamic instability, therefore, adds a term to the equation that increases the steady state monomer concentration above the critical concentration of the stable filaments (Fig. 14.7c). This increase has been measured directly for ParM. When dynamic instability is suppressed by non-hydrolyzable ATP analogs or point mutations, the steady state monomer concentration of ParM is 0.6 μM while, in the presence of dynamic instability, it rises to 2.4 μM (Garner, 2004). Dynamic instability is suppressed on plasmid-attached ParM filaments and they appear to have the same critical concentration as mutants that cannot hydrolyze ATP (Garner, 2007). ATPase-driven dynamic instability of unattached ParM filaments, therefore, increases the steady state monomer concentration above the critical concentration of the attached filaments and provides the energy to drive their preferential elongation.

14.7 AlfA: DNA Segregation Without Dynamic Instability It would be convenient if our detailed understanding of ParM-mediated DNA segregation could be used as a template for understanding the molecular mechanisms underlying all DNA-segregating, bacterial actins. Alas, this does not appear to be possible. The only other DNA segregating actin studied in vitro, AlfA, lacks a key property of ParM filaments, dynamic instability. AlfA filaments assemble in the presence of ATP and GTP as well as ADP and GDP (Polka, 2009). Similar to conventional actin, the critical concentrations of AlfA assembly in the presence of nucleotide diphosphates are only around fourfold higher than in the presence of nucleotide triphosphates. In addition, the overall morphologies of AlfA filaments and bundles are the same in ATP and ADP. This lack of dynamic instability is also evident in vivo. In E. coli, fluorescent derivatives of ParM form ephemeral filaments that grow to a maximum length and then rapidly shorten (Campbell, 2007). In

352

D. Mullins

B. subtilis, similar fluorescent derivatives of AlfA form stable bundles that run from pole to pole and do not appear to undergo catastrophic disassembly (Becker, 2006). Photobleaching experiments suggest that AlfA filaments are dynamic in the sense that new subunits are continually incorporated into one end and lost from the other. Interestingly, this “treadmilling” appears to occur in both directions. That is, an AlfA bundle appears to be composed of multiple filaments treadmilling in opposite directions. These in vitro assembly properties and in vivo behaviors are inconsistent with our picture of ParM-mediated DNA segregation and argue strongly that AlfA operates by a different mechanism. The mechanism of AlfA-mediated DNA segregation is not understood but might be driven directly by the bidirectional treadmilling of AlfA bundles observed in vivo. Because the bundle is aligned with the long axis of the cell, attachment of a plasmid to the side of a treadmilling filament would carry it to one pole or the other (Fig. 14.8). This could dramatically increase the frequency with which plasmids are placed in developing spores and promote their stability during sporulation. To increase stability during vegetative growth, however, the movement of one plasmid must be coupled to the movement of another in the opposite direction. The mechanism of this coupling is, at present, a complete mystery (Polka, 2009).

A

B

no dynamic instability mss =

k−

f

1−f

with dynamic instability

k+ mss =

k− k+

+(

1 f

− 1)

k cat k+

Fig. 14.8 Energetics of filament elongation and plasmid segregation. (a) The steady-state monomer concentration of a single state polymer. If free and plasmid-attached filaments have the same stability then, at steady state, the average rate of plasmid segregation will be zero. (b) If unattached filaments can exist in one of two states, either stable or rapidly disassembling, the steady state concentration of monomers will be higher than the critical concentration of the stable filaments. If the fraction of unstable filaments is f and the fraction of stable filaments (1–f), then the steady state monomer concentration is given by the expression in (b). If plasmid-attached filaments are as stable as unattached filaments that are, at a given moment, not shrinking, the plasmid-attached filaments will grow steadily even at steady state

14

Movement of Cargo in Bacterial Cytoplasm AlfA monomer

353 plasmid DNA

Segrosome complex

treadmilling AlfA filament

Fig. 14.9 Possible scheme for AlfA-mediated DNA segregation (Polka, 2009)

14.8 Summary and Conclusions Apart from the fact of their existence, the fundamental surprise that has come from the study of bacterial actins is the remarkable diversity in their structures and in the ways they self-assemble. At the electron microscope level, conventional actin and the three best understood bacterial actins, MreB, ParM, and AlfA, have significantly different filament architectures. They also assemble with different kinetics. The bacterial actins assemble rapidly and, unlike conventional actin, may not require nucleation factors. Bacterial actins also exhibit spontaneous self-association (AlfA) and dynamic instability (ParM), properties that may be essential for their cellular function. No stepping motors have been discovered in bacteria but ParM-mediated DNA segregation represents a well-established example of a polymerization-driven motor. Plasmids couple to ParM filaments via a DNA-protein complex that promotes insertional polymerization and surfs on the growing end of the filament. Analogs of this activity in eukaryotes include formin-mediated polymerization of actin filaments and coupling of the Dam1 complex to growing and shortening ends of microtubules. The ability of growing ParM filaments to align with the long axis of rod-shaped cells and to processively push plasmids can be explained by

354

D. Mullins

Brownian ratchet-type mechanisms proposed to underlie the ability of conventional actin filaments to push membranes in eukaryotes. In summary, the specific cellular functions and biochemical activities of bacterial actins appear to differ dramatically from each other and from those of conventional actin. The physical and conceptual tools developed from our study of the eukaryotic cytoskeleton (e.g. Brownian ratchets, insertional polymerization, dynamic instability), however, can be deployed to help us understand the structure and function of prokaryotic actins.

References Amos LA, van den Ent F, Löwe J. (2004) Structural/functional homology between the bacterial and eukaryotic cytoskeletons. Curr Opin Cell Biol. 16(1):24–31. Becker E, Herrera NC, Gunderson FQ, Derman AI, Dance AL, Sims J, Larsen RA, Pogliano J. (2006) DNA segregation by the bacterial actin AlfA during Bacillus subtilis growth and development. EMBO J. 25(24):5919–31. Bork P, Sander C, Valencia A. (1992) An ATPase domain common to prokaryotic cell cycle proteins, sugar kinases, actin, and hsp70 heat shock proteins. Proc Natl Acad Sci U S A. 89(16):7290–94. Cabeen MT, Charbon G, Vollmer W, Born P, Ausmees N, Weibel DB, Jacobs-Wagner C. (2009) Bacterial cell curvature through mechanical control of cell growth. EMBO J. 28(9):1208–19. Epub 2009 Mar 12. Campbell CS, Mullins RD. (2007) In vivo visualization of type II plasmid segregation: Bacterial actin filaments pushing plasmids. J Cell Biol. 179(5):1059–66. Carballido-López R. (2006) The bacterial actin-like cytoskeleton. Microbiol Mol Biol Rev. 70(4):888–909. Choi CL, Claridge SA, Garner EC, Alivisatos AP, Mullins RD. (2008) Protein-nanocrystal conjugates support a single filament polymerization model in R1 plasmid segregation. J Biol Chem. 283(42):28081–86. Collins JH, Elzinga M. (1975) The primary structure of actin from rabbit skeletal muscle. Completion and analysis of the amino acid sequence. J Biol Chem. 250(15):5915–20. Derman AI, Becker EC, Truong BD, Fujioka A, Tucey TM, Erb ML, Patterson PC, Pogliano J. (2009) Phylogenetic analysis identifies many uncharacterized actin-like proteins (Alps) in bacteria: regulated polymerization, dynamic instability and treadmilling in Alp7A. Mol Microbiol. 73(4):534–52. Ebersbach G, Gerdes K. (2005) Plasmid segregation mechanisms. Annu Rev Genet. 39:453–79. Elzinga M. (1971) Amino acid sequence around 3-methylhistidine in rabbit skeletal muscle actin. Biochemistry. 10(2):224–99. Galkin VE, Orlova A, Rivera C, Mullins RD, Egelman EH. (2009) Structural polymorphism of the ParM filament and dynamic instability. Structure. 17(9):1253–64. Garner EC, Campbell CS, Weibel DB, Mullins RD. (2007) Reconstitution of DNA segregation driven by assembly of a prokaryotic actin homolog. Science. 315(5816):1270–4. Garner EC, Campbell CS, Mullins RD. (2004) Dynamic instability in a DNA-segregating prokaryotic actin homolog. Science. 306(5698):1021–25. Gerdes K, Møller-Jensen J, Ebersbach G, Kruse T, Nordström K. (2004) Bacterial mitotic machineries. Cell. 116(3):359–66. Hatano S, Oosawa F. (1966) Extraction of an actin-like protein from the plasmodium of a myxomycete and its interaction with myosin A from rabbit striated muscle. J Cell Physiol. 68(2):197–202.

14

Movement of Cargo in Bacterial Cytoplasm

355

Hill TL, Kirschner MW. (1982) Subunit treadmilling of microtubules or actin in the presence of cellular barriers: possible conversion of chemical free energy into mechanical work. Proc Natl Acad Sci USA. 79(2):490–94. Hoischen C, Bussiek M, Langowski J, Diekmann S. (2008) Escherichia coli low-copy-number plasmid R1 centromere parC forms a U-shaped complex with its binding protein ParR. Nucleic Acids Res. 36(2):607–15. Holy TE, Leibler S. (1994) Dynamic instability of microtubules as an efficient way to search in space. Proc Natl Acad Sci U S A. 91(12):5682–85. Jensen GJ. Cell biology. (2009) Protein filaments caught in the act. Science. 323(5913): 472–3. Jones LJ, Carballido-López R, Errington J. (2001) Control of cell shape in bacteria: helical, actinlike filaments in Bacillus subtilis. Cell. 104(6):913–22. Komeili A, Li Z, Newman DK, Jensen GJ. (2006) Magnetosomes are cell membrane invaginations organized by the actin-like protein MamK. Science. 311(5758):242–45. Kruse T, Gerdes K. (2005) Bacterial DNA segregation by the actin-like MreB protein. Trends Cell Biol. 15(7):343–45. Kürner J, Medalia O, Linaroudis AA, Baumeister W. (2004) New insights into the structural organization of eukaryotic and prokaryotic cytoskeletons using cryo-electron tomography. Exp Cell Res. 301(1):38–42. Löwe J, van den Ent F, Amos LA. (2004) Molecules of the bacterial cytoskeleton. Annu Rev Biophys Biomol Struct. 33:177–98. Mayer JA, Amann KJ. (2009) Assembly properties of the Bacillus subtilis actin, MreB. Cell Motil Cytoskeleton. 66(2):109–18. Mitchison T, Kirschner M. (1984) Dynamic instability of microtubule growth. Nature. 312(5991):237–42. Møller-Jensen J, Borch J, Dam M, Jensen RB, Roepstorff P, Gerdes K. (2003) Bacterial mitosis: ParM of plasmid R1 moves plasmid DNA by an actin-like insertional polymerization mechanism. Mol Cell. 12(6):1477–87. Møller-Jensen J, Jensen RB, Löwe J, Gerdes K. (2002) Prokaryotic DNA segregation by an actinlike filament. EMBO J. 21(12):3119–27. Møller-Jensen J, Gerdes K. (2004) Dynamic instability of a bacterial engine. Science. 306(5698):987–89. Møller-Jensen J, Ringgaard S, Mercogliano CP, Gerdes K, Löwe J. (2007) Structural analysis of the ParR/parC plasmid partition complex. EMBO J. 26(20):4413–22. Needleman DJ. (2008) Plasmid segregation: is a total understanding within reach? Curr Biol. 18(5):R212–14. Nolen BJ, Pollard TD. (2007) Insights into the influence of nucleotides on actin family proteins from seven structures of Arp2/3 complex. Mol Cell. 26(3):449–57. Orlova A, Garner EC, Galkin VE, Heuser J, Mullins RD, Egelman EH. (2007) The structure of bacterial ParM filaments. Nat Struct Mol Biol. 14(10):921–26. Poch O, Winsor B. (1997) Who’s who among the Saccharomyces cerevisiae actin-related proteins? A classification and nomenclature proposal for a large family. Yeast. 13(11): 1053–58. Polka JK, Kollman JM, Agard DA, Mullins RD. (2009) The structure and assembly dynamics of plasmid actin AlfA imply a novel mechanism of DNA segregation. J Bacteriol. 191(20): 6219–30. Popp D, Narita A, Oda T, Fujisawa T, Matsuo H, Nitanai Y, Iwasa M, Maeda K, Onishi H, Maéda Y. (2008) Molecular structure of the ParM polymer and the mechanism leading to its nucleotidedriven dynamic instability. EMBO J. Feb 6;27(3):570–79. Salje J, Löwe J. (2008) Bacterial actin: architecture of the ParMRC plasmid DNA partitioning complex. EMBO J. 27(16):2230–38. Salje J, Zuber B, Löwe J. (2009) Electron cryomicroscopy of E. coli reveals filament bundles involved in plasmid DNA segregation. Science. 323(5913):509–12.

356

D. Mullins

Salmon ED. (2005) Microtubules: a ring for the depolymerization motor. Curr Biol. 15(8): R299–302. Schumacher MA, Glover TC, Brzoska AJ, Jensen SO, Dunham TD, Skurray RA, Firth N. (2007) Segrosome structure revealed by a complex of ParR with centromere DNA. Nature. 450(7173):1268–71. Shively ed. (2006) “Complex Intracellular Structures in Prokaryotes” Springer, Berlin. Straub FB. (1942) Actin. “Studies from the Institute of Medical Chemistry, University Szeged. Vol II.” ed. A. Szent-Gyorgyi. S. Karger, Basel, Switzerland. van den Ent F, Møller-Jensen J, Amos LA, Gerdes K, Löwe J. (2002) F-actin-like filaments formed by plasmid segregation protein ParM. EMBO J. 21(24):6935–43. Verde F, Labbé JC, Dorée M, Karsenti E. (1990) Regulation of microtubule dynamics by cdc2 protein kinase in cell-free extracts of Xenopus eggs. Nature. 343(6255):233–38. Woldringh CL, Mulder E, Huls PG, Vischer N. (1991) Toporegulation of bacterial division according to the nucleoid occlusion model. Res Microbiol. 142(2–3):309–20. Woldringh CL. (2002) The role of co-transcriptional translation and protein translocation (transertion) in bacterial chromosome segregation. Mol Microbiol. 45(1): 17–29. Yang YZ, Perdue JF. (1972) Contractile proteins of cultured cells. I. The isolation and characterization of an actin-like protein from cultured chick embryo fibroblasts. J Biol Chem. 247(14):4503–9.

Part III

Physical Aspects

Chapter 15

Protrusive Forces Generated by Dendritic Actin Networks During Cell Crawling Ovijit Chaudhuri and Daniel A. Fletcher

Abstract Cell protrusions during crawling motility involve growth of a dendritic actin network at the leading edge. Numerous studies over the last several decades have revealed the identity of key proteins involved in dendritic actin networks and elucidated how these proteins interact with each other during network assembly and growth, enabling reconstitution in vitro. While it is clear that growth of these networks can displace loads, the mechanism and dynamics of force generation continue to be a subject of investigation. In this chapter we describe current theories for the underlying mechanism of force generation by dendritic actin networks and discuss experimental measurements that quantify the magnitude and dynamics of the forces. Reconstitution studies have played a central role in measurements of force generation by dendritic actin networks and demonstrated how the combination of a small set of proteins leads to complex history dependent growth. Measurements of protrusion in crawling cells are now beginning to test the predictions of these reconstitution studies. We conclude by discussing how dendritic actin network growth can be integrated into an overall understanding of cell protrusions during crawling motility.

Contents 15.1 Introduction to Cell Motility and Dendritic Actin Networks . . . . . . . . 15.1.1 Cell Motility Driven by the Growth of Dendritic Actin Networks . 15.1.2 Assembly of Dendritic Actin Networks Through In Vitro Reconstitution . . . . . . . . . . . . . . . . . . . . 15.1.3 Open Questions on Force Generation by Dendritic Actin Networks 15.2 Models of Force Generation for Dendritic Actin Networks . . . . . . . .

. . . . . .

360 360

. . . . 361 . . . 362 . . . 362

D.A. Fletcher (B) Department of Bioengineering and Biophysics Program, University of California, Berkeley, CA 94720, USA; Physical Biosciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA e-mail: [email protected]

M.-F. Carlier (ed.), Actin-based Motility, DOI 10.1007/978-90-481-9301-1_15, C Springer Science+Business Media B.V. 2010

359

360

O. Chaudhuri and D. Fletcher

15.2.1 Force-Velocity Relationships in Biology and for Actin Networks 15.2.2 Force-Velocity Relationship for a Single Actin Filament . . . . 15.2.3 Predictions for the Force-Velocity Relationship of Dendritic Actin Networks . . . . . . . . . . . . . . . . 15.3 Measurements of Force Generation by Dendritic Actin Network Growth . 15.3.1 Methods for Measuring Force Velocity Relationship for Actin Network Growth . . . . . . . . . . . . . . . . . 15.3.2 Growth Velocity is Loading History Dependent . . . . . . . . 15.3.3 Side-View AFM Reveals that Network Density Increases Against an Increasing Force . . . . . . . . . . . . . . . . . . . . . 15.3.4 Loading History Dependence of Actin Networks Exemplifies Emergent Dynamics in Multi-Molecular Systems . . . . . . . 15.3.5 Feedback Between Network Architecture and Force Generation . 15.4 Tying In Vitro Measurements to Cell Motility . . . . . . . . . . . . . 15.4.1 In Vivo Measurements of Protrusive Forces . . . . . . . . . . 15.4.2 From Dendritic Actin Network Growth to Cell Crawling . . . . 15.5 Future Directions . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . .

362 363

. . . . . 365 . . . . 367 . . . . . 367 . . . . 369 . . . . . 371 . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

371 372 373 373 375 376 377

15.1 Introduction to Cell Motility and Dendritic Actin Networks 15.1.1 Cell Motility Driven by the Growth of Dendritic Actin Networks Cell crawling along a substrate involves a coordinated cycle of leading edge protrusion, substrate adhesion, trailing edge retraction, and substrate de-adhesion (Abercrombie, 1980). The first stage of cell crawling occurs when the cell extends a leading edge in the direction of movement in what is known as a lamellipodial protrusion. The role of actin in these protrusions was first identified by Theriot and Mitchison, who observed that the existing actin network in a crawling fish keratocyte cell remained fixed relative to the substrate and that the nucleation and polymerization of new actin filaments at the leading edge was directly coupled to forward movement of the keratocyte (Theriot and Mitchison, 1991). In other cell types, movement is also accompanied by dendritic actin network growth at the leading edge, but the coupling between the network and cell movement may be reduced, leading to significant retrograde flow of the network toward the cell body. In either case, growth of the network generates directed forces and causes elastic compression of the existing network, leading to changes in network density and organization. This section discusses measurements of force generation by the actin network at the leading edge in both reconstitution and whole cell studies.

15

Protrusive Forces Generated by Dendritic Actin Networks During Cell Crawling

361

15.1.2 Assembly of Dendritic Actin Networks Through In Vitro Reconstitution Studies of the bacterial pathogen Listeria monocytogenes have provided fundamental insight into the biochemical details of dendritic actin network assembly and growth. A study of Listeria, which can move through the host cell cytoplasm and rocket into neighboring cells, showed that actin in the tail developing behind the bacterium remained stationary, while new actin was polymerized at the surface of the bacterium (Theriot et al., 1992). This was similar to the actin dynamics earlier observed in keratocytes, suggesting a similarity in mechanism. Only one bacterial surface protein, ActA, was found to be necessary for actin-based motility of Listeria inside of a cell (Vazquez-Boland et al., 1992). Subsequent studies of Listeria motility and in vitro characterization of proteins identified to be important in Listeria motility, combined with investigations of the localization and activity of these proteins in the lamellipodia of crawling cells led to the dendritic model of actin based motility (Welch et al., 1997; Mullins et al., 1998; Svitkina and Borisy, 1999; Zalevsky et al., 2001; Pollard et al., 2000). This model proposes that new filaments are nucleated when signaling pathways, in response to chemotactic or other signals, activate the nucleation promotion factor WASP or one of several homologs. The activated WASP binds to and activates the Arp2/3 complex, which nucleates the formation of a growing filament as a branch oriented at an angle of 70◦ with respect to the preexisting filament. The collective growth of these filaments in the network pushes the membrane, or bacterium in the case of Listeria, forward. Capping protein terminates elongation of individual filaments to create short filaments unlikely to buckle under compressive loads and to prevent depletion of the pool of available monomeric actin (G-actin). Importantly, this process localizes growth to just the nucleation surface. Continued growth of the network relies on continued nucleation of new filaments by activated WASP and the Arp2/3 complex. Dendritic actin network growth is consequently tightly regulated through control of WASP activation state. After assembly and force generation, the actin networks are disassembled and reused. ADF/cofilin is involved in severing and debranching the network, which is critical for recycling of the actin monomers. Since there is a limited supply of monomers in cells at any given moment in time, rapid protrusion requires both efficient polymerization and depolymerization of the actin. In vitro reconstitutions of actin-based motility have supported the basic elements of the dendritic nucleation model and served as a useful platform for investigations of individual proteins as well as overall network properties. Loisel and colleagues presented the first reconstitution of actin-based motility on dead Listeria with a set of purified proteins, showing that only the Arp2/3 complex, capping protein, ADF/cofilin, profilin, and actin in solution are required for sustained movement of the bacterium (Loisel et al., 1999). At the same time, Cameron and colleagues reconstituted actin based motility on a synthetic surface, Polystyrene beads in this

362

O. Chaudhuri and D. Fletcher

case, coated in ActA and immersed in cytoplasmic extract, demonstrating that dendritic actin network growth can be reconstituted on virtually any surface coated in the nucleation promotion factor (Cameron et al., 1999).

15.1.3 Open Questions on Force Generation by Dendritic Actin Networks Displacing beads or bacteria in reconstitution experiments like those described above requires only a small force generated by the dendritic actin networks to overcome viscous drag. However, movement of a crawling cell through a crowded extracellular matrix, or the passage of a Listeria pathogen through a host cell’s cytoskeleton network and cell membrane into a neighboring cell, must involve the generation of significant forces. Until recently, the range of forces generated, the mechanism of force generation, and the response of network growth to an external force were poorly understood. The following section describes recent experimental progress in vitro and in vivo towards elucidating the force generation characteristics of dendritic actin networks, and point out some interesting unanswered questions within this field.

15.2 Models of Force Generation for Dendritic Actin Networks 15.2.1 Force-Velocity Relationships in Biology and for Actin Networks Force generation in biological systems, or the conversion of chemical energy into mechanical work, is typically characterized by force-velocity (f-v) relationships. Force-velocity relationships were first measured for muscle contracting against an external load (Hill, 1938). Since then, these relationships have been theoretically predicted and experimentally measured for the movement of molecular motors such as kinesin and myosin along a cytoskeletal filament track, growth of individual microtubules, packaging by a DNA motor, and transcription of DNA into RNA by RNA polymerase, to name just a few examples (Dogterom and Yurke, 1997; Finer et al., 1994; Smith et al., 2001; Svoboda and Block, 1994; Wang et al., 1998). The f-v relationship refers to how the velocity of motion in the case of molecular motors or velocity of growth in the case of a growing filamentous polymer such as a microtubule is affected by an external force. These f-v relationships offer insight into the underlying fundamental mechanochemical cycles involved in the particular process. Initial ideas to explain the growth of dendritic actin networks focused on the single filament level, assuming that networks could be modeled as a simple superposition of single filament behavior. In this perspective, the load on the network is distributed evenly among the filaments, and the velocity of filament growth is governed by a universal force-velocity relationship for individual actin filaments.

15

Protrusive Forces Generated by Dendritic Actin Networks During Cell Crawling

363

15.2.2 Force-Velocity Relationship for a Single Actin Filament Hill and Kirschner first proposed that mechanical work could be generated by the growth of a polymer such as actin or a microtubule (Hill and Kirschner, 1982a, b). They argued that the tread-milling of monomer through an individual polymer, and the hydrolysis of ATP to ADP after addition that this treadmilling involves, could theoretically provide the energy necessary for useful work. Generation of work would occur if the filament is constrained and growing against a resisting force, but only if the exchange of monomers was still possible at both ends of the filament. They predicted that the growth velocity to decay exponentially with force (Fig. 15.1b). While Hill and Kirchners’ work provided thermodynamic arguments that actin filament growth should be able to generate force, it did not offer insight into the specific mechanism of force generation, or how monomers could add onto a polymer that is in physical contact at both ends with an external surface that exerts a load on the filament. Peskin and colleagues offered the first mechanistic explanation of this process, proposing a Brownian ratchet mechanism for actin filament growth against a load (Peskin et al., 1993). They considered the thermal motion of the load surface and the resulting fluctuation in distance between the filament end and load surface. When this distance fluctuated greater than the distance necessary for an actin monomer to intercalate in the filament (∼2.7 nm), an actin monomer could add onto the filament end, rectifying the thermal motion of the surface so as to effectively displace it, generating mechanical work in the process. Due to evidence that suggested thermal motions of the load surface in certain biologically relevant situations may not be sufficient to allow for monomer intercalation, Mogilner and Oster later revised the model of actin filament growth to include the concept of an elastic ratchet (Mogilner and Oster, 1996). This model proposed that thermal bending of the filament itself away from the load surface allowed intercalation of an actin monomer, and addition of the monomer onto the filament, which then pushes the load forward when the filament fluctuates back against the load. This is particularly relevant for growth of individual filaments within a growing network, such as for filopodia or lamellipodia, as other filaments must support the load surface to prevent it from moving back with the elastic fluctuations of an individual filament. More recent experimental evidence, however, suggests that fluctuations between the network and load still play an important role for the growth of dendritic actin networks, though likely in concert with an elastic ratchet type mechanism at the single filament level (Shaevitz and Fletcher, 2007). There have only been a few experimental measurements of force generation by the growth of individual actin filaments to date. Kovar and Pollard looked at the growth of filaments bound by the pointed end to a microscope slide, with the nucleating protein forming on the barbed end immobilized on the slide as well. Filaments as short as 0.7 μm were observed to buckle, demonstrating that actin filaments can generate more than 1 pN of force (Kovar and Thomas, 2004). More recently, Footer and colleagues directly measured the forces generated by actin filaments with an

364

O. Chaudhuri and D. Fletcher

Fig. 15.1 Different predictions for force generation by dendritic actin networks. (a) Cartoon of dendritic actin network growth, and the interactions and factors that could affect the force velocity curve. Modified from a similar cartoon that appeared in Mogilner and Oster (2003). (b) This graph shows our interpretation of different theoretical predictions for the force-velocity relationship for dendritic actin networks. The force and velocity are in arbitrary units for each curve, so only the shapes of the curves should be compared

optical trap resisting the growth of a bundle of actin filaments growing against a barrier. They argue that only one filament is in contact with the barrier at any given time and observe growth of this single filament to be stalled at forces on the order of 1 pN (Footer et al., 2007). While these studies have offered important experimental

15

Protrusive Forces Generated by Dendritic Actin Networks During Cell Crawling

365

validation of force generation by single filament growth, the exact nature of the f-v relationship for an individual actin filament has yet to be established experimentally.

15.2.3 Predictions for the Force-Velocity Relationship of Dendritic Actin Networks Models of the growth of dendritic actin networks have advanced significantly beyond considering the network as a collection of single filaments acting independently. Experimental results, in close collaboration with models and simulations, have revealed that various factors from the microscopic to mesoscopic scale can influence the measured force-velocity relationship (Fig. 15.1a). Beyond the basic f-v relationship for an individual actin filament, tethering of the network to the load can affect the effective force any individual filament supports. Filament density and branching rate may be force dependent, and network elasticity becomes important in converting network growth into displacement of the leading edge membrane rather than network deformation. The following sections detail the theoretical role these factors are thought to play. A wide range of predictions of the f-v relationship for dendritic actin networks have been made based on different sets of assumptions about the importance of tethering of the network to the load surface, the autocatalytic nature of dendritic actin network growth, and the elasticity of the network in addition to the basic single filament f-v relationship (Fig. 15.1a). A summary of the predictions for the shape of the force velocity curve discussed in this section is shown in Fig. 15.1b. While we focus on three models here, we note that there are other models of actin network growth that we do not cover in this paper but recommend (Dickinson and Purich, 2002; Dickinson et al., 2004; Lee and Liu, 2009; Alberts and Odell, 2004). 15.2.3.1 The Elastic-Tethered Ratchet Model Gerbal and colleagues conducted an experiment in which they attempted to separate a Listeria bacterium from its actin tail using an optical trap, but the attachment forces were too great (Gerbal et al., 2000b). This experiment revealed the adhesion between the network and bacterium could be greater than 10 piconewtons, and it challenged previous theories of actin based motility which did not consider there to be attachments between the network and the nucleation surface. This motivated the development of the elastic-tether ratchet model of dendritic actin network growth by Mogilner and Oster (2003). They posited that a transient attachment between the nucleation promotion factor and the Arp2/3 complex occurs when a filament is nucleated, and this attachment resists forward progress of the load. As a result, transient attachments oppose the protrusive force generated by growth of individual filaments in the network. By using an assumption for the force dependent offrate of these links, and some assumptions for nucleation and branching rates, they numerically simulated the force-velocity relationship. Mogilner and Oster found the

366

O. Chaudhuri and D. Fletcher

relationship to be biphasic with the velocity decreasing quickly at small loads, but then decreasing slowly at higher loads (Fig. 15.1b). They explained that as the force is increased at low loads, transient links remain attached longer, which increases the drag on the bacterium resulting in positive feedback that slows the bacterium down rapidly. At higher loads, however, the bacterium moves so slowly that these transient links between the network and bacterium detach before they are stressed, so that the velocity decreases more slowly with increasing force. 15.2.3.2 The Autocatalytic Model The autocatalytic model was proposed by Carlsson and was based on the idea that new filaments must be nucleated from existing filaments leading to an intrinsic mechanism for feedback (Carlsson, 2001, 2003). Specifically, he assumes there to be a branching region, or a certain distance from the load surface in which the nucleation promotion factor can act. In other words, since the nucleation promotion factor is attached to the surface and can only extend a few to tens of nanometers, new branches can only be nucleated off of filaments within the branching region. As a result of this assumption, Carlsson predicts the growth velocity of the network to be load independent at steady state. He argues that if a growing network pushes against an increasing force, the network could slow down transiently due to the increased force per filament. However, if the load surface slows down, more filaments within the network grow within the branching region, allowing a greater number of branches to be nucleated. This results in an increase in the number of newly branched filaments, reducing the force per filament, resulting in a return of the growth velocity to its original value. This argument also holds when the external force on the network is reduced, so that the velocity of network growth is load independent for a range of forces at steady state. From his model, the velocity of growth could change instantaneously following a change in force, but then the velocity should exponentially decay back to the original steady state velocity. However, we would expect there to be a maximum density of actin due to steric and volume constraints. Given this limitation, we would predict there to be a force above which the growth velocity decreases, potentially following the single filament f-v curve, until growth is stalled (second part of autocatalytic prediction shown in Fig. 15.1b). 15.2.3.3 The Elastic Gel Model A third model is the elastic gel model, which predicts the f-v relationship based on the elastic properties of the actin network (Gerbal et. al., 2000a). For the case of a network growing against a flat surface, this model predicts the growth or displacement velocity at a given force to be reduced due to elastic deformation of the network under compression. In other words, as the network polymerizes under a load, the load is displaced but also the network is compressed elastically. As a result, a lower growth velocity should be measured relative to the velocity that would be

15

Protrusive Forces Generated by Dendritic Actin Networks During Cell Crawling

367

expected treating the network as a rigid collection of individual filaments with a given f-v relationship. However, for the case of a network growing against a curved surface, like that of a spherical bead, this model predicts a fundamentally different mechanism of motion. In this situation, polymerization at the load surface pushes the existing elastic network outwards, increasing the stress in the surrounding network. The stress is then continually relaxed by the forward movement of the bead. There has been significant experimental support for the applicability of this model to curved surfaces in general, particularly for the motion of Listeria. However, lamellipodial extensions in cells like fish keratocytes are essentially flat, or if anything, exhibits the opposite curvature, so that we would not expect this mechanism of motion to be directly applicable for cell crawling (Boukellal et al., 2004; Marcy et al., 2004; van der Gucht et al., 2005). This dramatic difference in mechanism of movement between flat and curved surfaces, highlights the importance of selecting appropriate boundary conditions when developing model systems in vitro.

15.3 Measurements of Force Generation by Dendritic Actin Network Growth 15.3.1 Methods for Measuring Force Velocity Relationship for Actin Network Growth Since the original in vitro reconstitutions of actin based motility of Listeria and polystyrene microspheres, various methods have been used to probe the force generating characteristics of these reconstituted networks (Fig. 15.2). 15.3.1.1 Viscosity Based Approaches The first measurements of the force velocity relationship for actin network growth involved increasing the drag force against Listeria or polystereyne microspheres undergoing actin based motility by increasing the viscoelasticity of the medium with methylcellulose (Fig. 15.2a). Wiesner and colleagues modulated the drag force on beads undergoing actin-based motility from a mix of purified proteins by adding methylcellulose into the medium (Wiesner et al., 2003). They found that the bead velocities stayed constant against drag forces below 1 pN, but that the velocity dropped by ∼25% at a drag force of 50 pN. Alternatively, McGrath and colleagues added methylcellulose to brain extract to modulate the drag force against Listeria which had been placed in the extract (McGrath et al., 2003). They found the forcevelocity relationship to be highly curved, decaying sharply at low forces, and more gently at high forces, potentially supporting the tethered-elastic Brownian ratchet model for actin-based motility. However, methylcellulose may bundle filaments,

368

O. Chaudhuri and D. Fletcher

Method Viscosity Microfiber AFM cantilever

Surface geometry Curved Curved Flat

Max. Force (nN) 0.2 5 300

Pressure (Pa) 5 400 1000

Fig. 15.2 Experimental geometries for probing force generation by dendritic actin networks in vitro. (a) The drag force against a bead exhibiting actin based motility can be modulated by changing the viscosity of the surrounding medium. (b) By attaching the bead to a flexible fiber and grabbing onto the actin network with a pipette, Marcy and colleagues were able to exert larger forces against network growth. Cartoon is a modification of a similar cartoon in Marcy et al. (2004). Forces were measured by quantification of the displacement of the fiber via optical imaging, and could be controlled by translation of the micropipette. (c) By reconstituting the growth of dendritic actin networks off of the end of an AFM cantilever, we were able to exert compressive forces against actin network growth up through stall. The cantilever behaves like a Hookean spring for small deflections, and the deflection is measured with an optical lever. Forces are controlled through translation of the surface via a piezoelectric stage (Table). A summary of the forces imposed and geometries probed by the different techniques is shown

increase spontaneous nucleation, and decrease depolymerization in the network (Wiesner et al., 2003; Frederick et al., 2008). Also, methylcellulose modulates the viscoelasticity of the medium, leading to some potentially non-trivial timedependent and structural effects on the load resisting actin-driven movement. More importantly, the maximum force applied of 200 pN was well below that needed to stall network growth.

15

Protrusive Forces Generated by Dendritic Actin Networks During Cell Crawling

369

15.3.1.2 Flexible Microfiber More recently, Marcy and colleagues measured the force-velocity relationship of actin network growth in a novel micropipette based assay in which they were able to directly apply forces against network growth (Fig. 15.2b) (Marcy et al., 2004). A bead coated with WASP is glued onto the end of a glass microfiber and is immersed in a mix of purified proteins. The microfiber behaves like a spring for small deflections and thus acts as a force transducer in the experiment, and the deflection of the microfiber is measured via optical imaging. The actin network that grows from the bead attached to the microfiber is then held by a micropipette so that growth deflects the fiber. Translation of the microfiber enables constant force experiments in tension and compression. Marcy and colleagues found the network growth velocity to decay exponentially with force. These results are fully explained within the context of the elastic gel model for a network growing off of a spherical surface. Consequently, the measured force-velocity relationship does not seem to be directly applicable to a network growing off of a flat surface. Maximum compressional forces applied by the microfiber were 4.3 nN, but this was not high enough to stall network growth. 15.3.1.3 Atomic Force Microscopy A final technique that has been used to measure forces generated by actin network growth is atomic force microscopy (Fig. 15.2c). Atomic force microscopy (AFM) essentially involves probing a sample with a soft, micron-scale cantilever that behaves like a Hookean spring for small deflections (Binnig et al., 1986). Cantilever position is detected via an optical lever (bouncing a focused laser on the end of the cantilever and detected the position of the reflected beam) enabling sub-nm resolution of cantilever position and a force resolution of ∼10 pN, in addition to high temporal resolution (>100 Hz). We reconstituted dendritic actin network growth off of an AFM cantilever by coating the end of the cantilever with the nucleation promotion factor ActA (Choy et al., 2007), and were able to measure and exert forces on the network by manipulating the position of the surface from which the network was growing. We were able to measure the force velocity relationship for dendritic actin networks growing against an increasing force and then stall network growth (Parekh et al., 2005). We determined the stall force for these networks to be on the order of 50–300 nN, but this number is dependent on ActA density and the overall network area (which is larger in our experiments than in others), and so cannot be directly compared with other studies.

15.3.2 Growth Velocity is Loading History Dependent In our experiments probing the force-velocity relationship of actin networks with AFM, we let the network growth freely deflect the cantilever until the compressive force exerted by the cantilever against the network stalled further growth (Parekh et al., 2005). These experiments showed that the actin network growth was load

370

O. Chaudhuri and D. Fletcher

independent against increasing loads for a certain range of forces, but it exhibited a concave decrease in velocity with force until growth was stalled above this range (Fig. 15.3a). We hypothesized that this load independence arose from an increase in filament density against increasing in load such that the force per filament remained constant, as predicted by Carlsson’s autocatalytic model. To test this hypothesis, we performed experiments in which we reduced the force on the network instantaneously for a network growing within the load independent regime (Fig. 15.3b). We found that the velocity increased, relative to the velocity of growth previously at that load, as would be expected for an increase in filament density. Surprisingly, we found that this increased growth velocity was sustained over the course of the measurement and did not exponentially decay back to the

Fig. 15.3 Measurements of force generation by actin networks in an AFM. (a) The force velocity relationship for a dendritic actin network growing against increasing loads. (b) Force reduction experiments show velocity of network growth is loading history dependent. In this experiment, the force on the network is held constant as the network grows. Then the network is allowed to freely deflect the cantilever with further growth. Finally, the force is reduced to the original force clamp value, but the growth velocity is observed to be much higher. Cartoons above graphs depict progression of network during experiment, although bends in cantilever curvature is not drawn to scale and enhanced to emphasize increased application of force by cantilever. Reprinted with Permission from Nature Cell Biology (Parekh et al., 2005)

15

Protrusive Forces Generated by Dendritic Actin Networks During Cell Crawling

371

original velocity as would be predicted from Carlsson’s model. Thus, in contrast to well-defined force velocity relationships seen in molecular motors or single filaments, and predicted for dendritic actin networks, the force velocity relationship for dendritic actin network growth is loading history dependent. Preliminary results suggest that this is also seen in actin networks reconstituted in a mix of purified proteins or just from actin, the Arp 2/3 complex, and capping protein, and we have determined that this behavior is independent of the specific nucleation promotion factor (unpublished).

15.3.3 Side-View AFM Reveals that Network Density Increases Against an Increasing Force The recent development of a side-view AFM, which incorporates an imaging path providing imaging of a sample along a cross section that spans between the cantilever and the surface as in the cartoon in Fig. 15.2c, has enabled further insight into the basis for loading history dependence (Chaudhuri et al., 2009). An example experiment shows that the increased growth velocity seen in the “force-reduction” experiments is correlated with increased network density (Fig. 15.4). This suggests that an increased network density does explain the load-independent regime, but the question of how the increased density is sustained following a decrease in load remains unanswered. Further experiments with the side-view AFM using networks reconstituted from purified proteins, and correlating the Arp2/3 complex density with actin network density could offer further insight into this. One possible mechanism that deserves further investigation is whether filament bending can influence or enhance filament branching.

15.3.4 Loading History Dependence of Actin Networks Exemplifies Emergent Dynamics in Multi-Molecular Systems The observation of loading history dependence demonstrates that complex dynamics can emerge in relatively simple multi molecular systems. We observed that dendritic actin network growth can be different under identical external loading and biochemical conditions. This type of behavior has not been seen, nor is expected to be, in single molecule systems such as a molecular motor, or a single growing filament. However, the interaction of proteins involved in dendritic actin network growth leads to this nontrivial unexpected history dependent behavior. The structure of the network itself plays a role in network growth. This suggests the idea that there is a limit to which biological systems can be reduced to their component parts, and that in the limit of single molecules, the complexity that may define biological systems can be lost. The challenge in reconstitution biology then becomes identifying the minimal set of components that can capture these emergent dynamics.

372

O. Chaudhuri and D. Fletcher

Fig. 15.4 Force reduction experiments conducted in side-view AFM reveal that higher network growth velocity under the same force are correlated with increased network density. (a) Force and velocity are shown during a force reduction experiment, and the average fluorescence intensity along a line 2 μm away from the cantilever surface (as indicated by the green line in b) behind the tip is shown by the squares. (b) Below are images taken at the indicated times with the sideview AFM. Note that a reflection of the triangular cantilever tip and network off of the surface (indicated by the white dotted line) is seen. Non-uniformity of network cross section, or increased density of network just below cantilever, is not typical of all experiments. Scale bar is 10 m

15.3.5 Feedback Between Network Architecture and Force Generation The in vitro force generation results collected to date draw attention to the potential role and importance of network architecture in transmitting stresses. Some of our work investigating the deformability of these dendritic actin networks suggests that actin network architecture itself may be important in defining the mechanical properties of the network (Chaudhuri et al., 2007). By studying the nonlinear elasticity of these dendritic actin networks, we found that the resistance of filaments to compression, bending, and tension appear to influence the overall mechanical response to loading. Since the individual filament response to compression, characterized

15

Protrusive Forces Generated by Dendritic Actin Networks During Cell Crawling

373

by buckling, and tension, characterized by the entropic elasticity of individual filaments resisting extension, is so nonlinear and anisotropic, this suggests an importance of filament architecture in determining network elasticity. Simulations of crosslinked actin networks support this interpretation (Pronk and Geissler, personal communication). Taken together, these works suggest that there is a feedback between network growth and the mechanical properties of the resulting network. In this scenario, as loading conditions change, so does the structure of the actin network. This in turn affects the mechanical properties of the network, which then modulates the efficiency of load displacement.

15.4 Tying In Vitro Measurements to Cell Motility 15.4.1 In Vivo Measurements of Protrusive Forces While much of the work investigating force generation of lamellapodial protrusions has been done in the reconstituted in vitro systems, due to the ease in interpretation, there have been several studies that have looked at forces generated by actively crawling cells. These experiments provide important benchmarks for comparison with reconstituted studies, though with the caveat that various other cellular processes could contribute to the measurements. Using a conventional AFM, Brunner and colleagues were able to estimate the forces generated by a keratocyte as it crawled under and past a bead attached to the cantilever (Brunner et al., 2006) (Fig. 15.5a). In the study, the authors positioned the cantilever-mounted bead directly in the path of a crawling keratocyte and measured the range of forward forces that the keratocyte could exert and the range of distances between the bead and the surface that the keratocyte could squeeze through. Forward forces were measured by either treating the cell as an elastic wedge and calculating how the forward force of the cell translated into vertical deflection of the cantilever, or by measuring the torsion of the cantilever for cells crawling perpendicular to the orientation of the cantilever. Interestingly, Brunner and colleagues find that the velocity of protrusion did not seem to change for forces ranging from 0.4 to 2.4 nN against the lamellipodium. They also found the keratocyte to be able to compress up to 80% vertically to squeeze through gaps as small as 500 nm. A second study involved the use of a variation of AFM to directly measure the force-velocity relationship of a migrating keratocyte (Prass et al., 2006) (Fig. 15.5b). The group built a modified atomic force microscope in which an AFM cantilever was placed perpendicular to the sample surface and positioned directly in the path of a crawling cell. Cantilever deflection here is monitored via optical imaging of the cantilever tip. They measured how the cell growth velocity changed as the cell crawled into the cantilever, deflecting the cantilever further and further, thus crawling against higher forces until movement was eventually stalled. The resulting

374

O. Chaudhuri and D. Fletcher

Fig. 15.5 Experimental geometries for the measurement of protrusive forces in cells. (a) Cartoon of the experimental geometry of the measurement of protrusive forces as a cell squeezed under the bead in Brunner et al. (2006). (b) Experimental geometry for Prass et al. (2006), in which the forcevelocity relationship for a cell crawling directly into a cantilever was measured. (c) Experimentally measured force velocity curves from (b) from Prass et al. (2006) show a load independent regime. Panel C is taken from Prass et al., (2006). Reprinted with Permission from The Journal of Cell Biology (Prass et al., 2006)

force-velocity relationship showed a load independent regime, followed by a concave decrease in velocity with force until growth was stalled. However, the group also notes that the measured velocity was almost an order of magnitude less than the velocity of free cell crawling, suggesting a sharp decline in velocity at very low forces. They discuss a variety of mechanisms which could give rise to this initial decrease in velocity. While both of these measurements are qualitatively consistent with some of the force-velocity measurements of dendritic actin network growth reconstituted in vitro, care must be taken in making comparisons between the studies. These studies all reveal a load independence growth velocity, and the Prass study also finds a concave decrease in velocity after this regime until growth is stalled, similar to the force-velocity relationship found for dendritic actin networks reconstituted in vitro measured with AFM. While this behavior most likely reflects the behavior of the dendritic actin network at the leading edge of the crawling cell, osmotic imbalances, slipping of adhesions, or membrane undulations could also contribute to the measured relationship in cells.

15

Protrusive Forces Generated by Dendritic Actin Networks During Cell Crawling

375

15.4.2 From Dendritic Actin Network Growth to Cell Crawling The overall process of cell crawling is a multi-step multi-system process involving more than just the growth of a dendritic actin network. To form a more complete picture of cell crawling., critical factors and components that must be considered at the cell level include connections between the adhesions to the surface and the protrusive dendritic actin network, interaction between the membrane and filaments, coordination of global response, and also the nature of the mechanical environment itself that the cell is crawling through. 15.4.2.1 Connection Between Adhesion Machinery and Protrusive Machinery For dendritic actin network growth to be converted into a productive protrusion, the network has to be coupled directly to adhesions to the surface. The importance of this coupling was strikingly demonstrated in a study by Bohnet and colleagues (Bohnet et al., 2006). By flowing fluid through a micropipette against the leading edge of a crawling keratocyte, they were able to arrest flow locally at very small forces (∼few pN/μm), and observe upward folding of the leading edge. Through image analysis, they conclude that the fluid flow interferes disrupts nascent adhesions at the front of the cell and stalls the protrusion at forces much lower than those generated by the actin network. This strength of coupling between the dendritic actin network and adhesions seems to vary significantly between cells, such that significant retrograde flow of the actin network is observed in many types of cells, at the expense of protrusion of the leading edge (Lin and Forscher, 1995). Recent work using speckle microscopy to follow the polymerization and depolymerization dynamics of actin and traction force microscopy to measure stresses exerted by the adhesions on the substrate have helped elucidate this interaction and the proteins involved further (Watanabe and Mitchison, 2002; Ponti et al., 2004; Gardel et al., 2008). 15.4.2.2 Interaction Between the Membrane and Filaments Recent work has also implicated the role of membrane tension in modulating network growth dynamics. From studying the natural variation of motile epithelial keratocytes, Keren and colleagues suggest an important role for membrane tension and the resulting force on the network in modulating actin assembly and disassembly (Keren et al., 2008). In another study, Ji and colleagues used speckle microscopy to track the flow and turnover of actin and vinculin in protrusion-retraction cycles of epilethial cells, and based on this, propose that protrusion is limited by membrane tension (Ji et al., 2008). 15.4.2.3 Nature of Mechanical Microenvironment Several experiments have pointed to the important influence of the mechanical microenvironment of cells in affecting crawling of cells. Lo and colleagues

376

O. Chaudhuri and D. Fletcher

demonstrated that fibroblast cells exhibit durotaxis, or move towards stiffer substrates during crawling (Lo et al., 2000). More recently, Lammermann and colleagues found that the mechanism for leukocyte movement in a three dimensional environment is dramatically different than for leukocyte movement on a two dimensional substrate (Lämmermann et al., 2008). Specifically, they found that integrins were not necessary for leukocyte movement in a three dimensional gel with a large enough pore size, so that adhesion to the environment does not seem to play a role. These findings suggest a different mechanism of movement for cells in three dimensions, and could mean that dendritic actin network growth plays a different role in migration in three dimensions. 15.4.2.4 Adaptive Response of Cells to a Dynamic Microenvironment Another important area that requires further modeling and exploration is the coordination of actin network growth and adhesion globally in response to external perturbations. When a keratocyte cell crawls into a wall or rigid obstacle, it reorients and continues moving away from the wall or obstacle (Van Duijn, personal communication). Other cell types, such as white blood cells crawling through the extracellular matrix, must deal with a much more complicated mechanical environment. How do the dendritic actin network growth dynamics under load, in coordination with other intracellular systems, lead to the reorientation of the whole cell body?

15.5 Future Directions In this chapter we have described measurements of force generation by dendritic actin network growth in vitro as well as measurements of forces generated by cell crawling. We have highlighted the finding that dendritic actin network growth is loading history dependent – maintaining a memory of its previous loading interactions in its network architecture. Higher growth velocities under the same force appear to be correlated with increased network densities. This serves as a demonstration of how complex non-trivial behaviors arise due to the interactions of several different proteins within a defined module of interacting components. The important question remaining with this result is a sound theoretical explanation of how this history dependence arises. The bigger challenge in the near future for the cell crawling field will be to link some of these in vitro findings within the broader context of cell crawling. How are these growing dendritic actin networks coupled to adhesions? How do these results come together and explain the observed movement of cells bouncing off of a wall? We look forward to the further exploration of these questions in the near future. Acknowledgements We apologize to the many authors whose work we were not able to describe and cite due to space limitations. We thank the Fletcher lab for general discussions on the topic. OC acknowledges the support of an NSF graduate fellowship, and DAF acknowledges the support of NIH RO1 grants and the Cell Propulsion Lab, an NIH Nanomedicine Development Center.

15

Protrusive Forces Generated by Dendritic Actin Networks During Cell Crawling

377

References Abercrombie, M., 1980. The Croonian lecture, 1978: The crawling movement of metazoan cells. Proceedings of the Royal Society of London. Series B, Biological Sciences, 207(1167), 129–147. Alberts, J.B. & Odell, G.M., 2004. In silico reconstitution of Listeria propulsion exhibits nanosaltation. PLoS Biology, 2(12), e412. Binnig, G., Quate, C. & Gerber, Ch., 1986. Atomic force microscope. Physical Review Letters, 56(9), 930–933. Bohnet, S. et al., 2006. Weak force stalls protrusion at the leading edge of the lamellipodium. Biophysical Journal, 90(5), 1810–1820. Boukellal, H. et al., 2004. Soft Listeria: actin-based propulsion of liquid drops. Physical Review. E, Statistical, Nonlinear, and Soft Matter Physics, 69(6 Pt 1), 061906. Brunner, C.A. et al., 2006. Cell migration through small gaps. European Biophysics Journal: EBJ, 35(8), 713–719. Cameron, L.A. et al., 1999. Motility of ActA protein-coated microspheres driven by actin polymerization. Proceedings of the National Academy of Sciences of the United States of America, 96(9), 4908–4913. Carlsson, A.E., 2001. Growth of branched actin networks against obstacles. Biophysical Journal, 81(4), 1907–1923. Carlsson, A.E., 2003. Growth velocities of branched actin networks. Biophysical Journal, 84(5), 2907–2918. Chaudhuri, O., Parekh, S.H. & Fletcher, D.A., 2007. Reversible stress softening of actin networks. Nature, 445(7125), 295–298. Chaudhuri, O. et al., 2009. Combined atomic force microscopy and side-view optical imaging for mechanical studies of cells. Nature Methods, 6(5), 383–387. Choy, J.L. et al., 2007. Differential force microscope for long time-scale biophysical measurements. The Review of Scientific Instruments, 78(4), 043711. Dickinson, R.B., Caro, L. & Purich, D.L., 2004. Force generation by cytoskeletal filament endtracking proteins. Biophysical Journal, 87(4), 2838–2854. Dickinson, R.B. & Purich, D.L., 2002. Clamped-filament elongation model for actin-based motors. Biophysical Journal, 82(2), 605–617. Dogterom, M. & Yurke, B., 1997. Measurement of the force-velocity relation for growing microtubules. Science (New York, N.Y.), 278(5339), 856–860. Finer, J.T., Simmons, R.M. & Spudich, J.A., 1994. Single myosin molecule mechanics: piconewton forces and nanometre steps. Nature, 368(6467), 113–119. Footer, M.J. et al., 2007. Direct measurement of force generation by actin filament polymerization using an optical trap. Proceedings of the National Academy of Sciences of the United States of America, 104(7), 2181–2186. Frederick, K.B., Sept, D. & De La Cruz, E.M., 2008. Effects of solution crowding on actin polymerization reveal the energetic basis for nucleotide-dependent filament stability. Journal of Molecular Biology, 378(3), 540–550. Gardel, M.L. et al., 2008. Traction stress in focal adhesions correlates biphasically with actin retrograde flow speed. The Journal of Cell Biology, 183(6), 999–1005. Gerbal, F., Chaikin, P. et al., 2000a. An elastic analysis of Listeria monocytogenes propulsion. Biophysical Journal, 79(5), 2259–2275. Gerbal, F., Laurent, V. et al., 2000b. Measurement of the elasticity of the actin tail of Listeria monocytogenes. European Biophysics Journal: EBJ, 29(2), 134–140. Hill, AV, 1938. The heat of shortening and the dynamic constants of muscle. Proceedings of the Royal Society of London Series B-Biological Sciences, 126, 136–195. Hill, T.L. & Kirschner, M.W., 1982a. Bioenergetics and kinetics of microtubule and actin filament assembly-disassembly. International Review of Cytology, 78, 1–125.

378

O. Chaudhuri and D. Fletcher

Hill, T.L. & Kirschner, M.W., 1982b. Subunit treadmilling of microtubules or actin in the presence of cellular barriers: possible conversion of chemical free energy into mechanical work. Proceedings of the National Academy of Sciences of the United States of America, 79(2), 490–494. Ji, L., Lim, J. & Danuser, G., 2008. Fluctuations of intracellular forces during cell protrusion. Nature Cell Biology, 10(12), 1393–1400. Keren, K. et al., 2008. Mechanism of shape determination in motile cells. Nature, 453(7194), 475–480. Kovar, D.R. & Pollard, T.D., 2004. Insertional assembly of actin filament barbed ends in association with formins produces piconewton forces. Proceedings of the National Academy of Sciences of the United States of America, 101(41), 14725–14730. Lämmermann, T. et al., 2008. Rapid leukocyte migration by integrin-independent flowing and squeezing. Nature, 453(7191), 51–55. Lee, K. & Liu, A.J., 2009. Force-velocity relation for actin-polymerization-driven motility from brownian dynamics simulations. Biophysical Journal, 97(5), 1295–1304. Lin, C.H. & Forscher, P., 1995. Growth cone advance is inversely proportional to retrograde F-actin flow. Neuron, 14(4), 763–771. Lo, C.M. et al., 2000. Cell movement is guided by the rigidity of the substrate. Biophysical Journal, 79(1), 144–152. Loisel, T.P. et al., 1999. Reconstitution of actin-based motility of Listeria and Shigella using pure proteins. Nature, 401(6753), 613–610. Marcy, Y. et al., 2004. Forces generated during actin-based propulsion: a direct measurement by micromanipulation. Proceedings of the National Academy of Sciences of the United States of America, 101(16), 5992–5997. McGrath, J.L. et al., 2003. The force-velocity relationship for the actin-based motility of Listeria monocytogenes. Current Biology: CB, 13(4), 329–332. Mogilner, A. & Oster, G., 1996. Cell motility driven by actin polymerization. Biophysical Journal, 71(6), 3030–3045. Mogilner, A. & Oster, G., 2003. Force generation by actin polymerization II: the elastic ratchet and tethered filaments. Biophysical Journal, 84(3), 1591–1605. Mullins, R.D., Heuser, J.A. & Pollard, T.D., 1998. The interaction of Arp2/3 complex with actin: nucleation, high affinity pointed end capping, and formation of branching networks of filaments. Proceedings of the National Academy of Sciences of the United States of America, 95(11), 6181–6186. Parekh, S.H. et al., 2005. Loading history determines the velocity of actin-network growth. Nature Cell Biology, 7(12), 1219–1223. Peskin, C.S., Odell, G.M. & Oster, G.F., 1993. Cellular motions and thermal fluctuations: the Brownian ratchet. Biophysical Journal, 65(1), 316–324. Pollard, T.D., Blanchoin, L. & Mullins, R.D., 2000. Molecular mechanisms controlling actin filament dynamics in nonmuscle cells. Annual Review of Biophysics and Biomolecular Structure, 29, 545–576. Ponti, A. et al., 2004. Two distinct actin networks drive the protrusion of migrating cells. Science (New York, N.Y.), 305(5691), 1782–1786. Prass, M. et al., 2006. Direct measurement of the lamellipodial protrusive force in a migrating cell. The Journal of Cell Biology, 174(6), 767–772. Shaevitz, J.W. & Fletcher, D.A., 2007. Load fluctuations drive actin network growth. Proceedings of the National Academy of Sciences of the United States of America, 104(40), 15688–15692. Smith, D.E. et al., 2001. The bacteriophage straight phi29 portal motor can package DNA against a large internal force. Nature, 413(6857), 748–752. Svitkina, T.M. & Borisy, G.G., 1999. Arp2/3 complex and actin depolymerizing factor/cofilin in dendritic organization and treadmilling of actin filament array in lamellipodia. The Journal of Cell Biology, 145(5), 1009–1026.

15

Protrusive Forces Generated by Dendritic Actin Networks During Cell Crawling

379

Svoboda, K. & Block, S.M., 1994. Force and velocity measured for single kinesin molecules. Cell, 77(5), 773–784. Theriot, J.A. & Mitchison, T.J., 1991. Actin microfilament dynamics in locomoting cells. Nature, 352(6331), 126–131. Theriot, J.A. et al., 1992. The rate of actin-based motility of intracellular Listeria monocytogenes equals the rate of actin polymerization. Nature, 357(6375), 257–260. van der Gucht, J. et al., 2005. Stress release drives symmetry breaking for actin-based movement. Proceedings of the National Academy of Sciences of the United States of America, 102(22), 7847–7852. Vazquez-Boland, J.A. et al., 1992. Nucleotide sequence of the lecithinase operon of Listeria monocytogenes and possible role of lecithinase in cell-to-cell spread. Infection and Immunity, 60(1), 219–230. Wang, M.D. et al., 1998. Force and velocity measured for single molecules of RNA polymerase. Science (New York, N.Y.), 282(5390), 902–907. Watanabe, N. & Mitchison, T.J., 2002. Single-molecule speckle analysis of actin filament turnover in lamellipodia. Science (New York, N.Y.), 295(5557), 1083–1086. Welch, M.D., Iwamatsu, A. & Mitchison, T.J., 1997. Actin polymerization is induced by Arp2/3 protein complex at the surface of Listeria monocytogenes. Nature, 385(6613), 265–269. Wiesner, S. et al., 2003. A biomimetic motility assay provides insight into the mechanism of actinbased motility. The Journal of Cell Biology, 160(3), 387–398. Zalevsky, J., Grigorova, I. & Mullins, R.D., 2001. Activation of the Arp2/3 complex by the Listeria acta protein. Acta binds two actin monomers and three subunits of the Arp2/3 complex. The Journal of Biological Chemistry, 276(5), 3468–3475.

Chapter 16

Mathematical and Physical Modeling of Actin Dynamics in Motile Cells Anders E. Carlsson and Alex Mogilner

Abstract Mathematical modeling has been very instrumental in aiding traditional experimental methods in uncovering the mysteries of actin dynamics. Here we review recent quantitative models of actin dynamics focusing on ATP hydrolysis effects, force generation by single actin filaments and networks, self-organization and dynamics of actin networks, dynamics of lamellipodia, filopodia and lamella, and integrative mechanochemistry of whole motile cells. We discuss both modeling methods and specific insights from modeling that helped answering biological questions. Abbreviations ADP ATP BR CCA CP F-actin G-actin

adenosine diphosphate adenosine triphosphate Brownian ratchet cofilin, coronin, and Aip1 capping protein filamentous or polymerized actin globular or unpolymerized actin

Contents 16.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . 16.2 Polymerization and Force Generation by Single Filaments . . . 16.2.1 Effects of ATP Hydrolysis on Polymerization Dynamics 16.2.2 Force Generation by Polymerizing Actin Filaments . . 16.2.3 Barbed-End Tethering . . . . . . . . . . . . . . . . 16.2.4 Disassembly of Actin Filaments . . . . . . . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

. . . . . .

382 383 384 385 386 387

A.E. Carlsson (B) Department of Physics, Washington University, St. Louis, MO 63130, USA e-mail: [email protected]

M.-F. Carlier (ed.), Actin-based Motility, DOI 10.1007/978-90-481-9301-1_16, C Springer Science+Business Media B.V. 2010

381

382

A.E. Carlsson and A. Mogilner

16.3 Structure and Force Generation in Dendritic Actin Networks . . . . . . . . . 16.3.1 Force-Velocity Relation . . . . . . . . . . . . . . . . . . . . . . 16.3.2 Dendritic Network Structure . . . . . . . . . . . . . . . . . . . . 16.3.3 Symmetry Breaking and Hopping of Listeria and Actin-Propelled Beads 16.3.4 Spontaneous Waves of F-Actin . . . . . . . . . . . . . . . . . . . 16.4 Dynamics of Cell Protrusions and the Cell Periphery . . . . . . . . . . . . . 16.4.1 Filopodia . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.4.2 Lamellipodium and Lamellum . . . . . . . . . . . . . . . . . . . 16.5 Shape and Movements of the Whole Cell . . . . . . . . . . . . . . . . . . 16.5.1 Boundary Mechanics Models . . . . . . . . . . . . . . . . . . . 16.5.2 Actin-based Models . . . . . . . . . . . . . . . . . . . . . . . . 16.6 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . .

388 388 391 392 394 394 395 398 401 401 403 406 407

16.1 Introduction Mathematical modeling plays two major roles in biology. The first one is based on the reductionist agenda (Pollard, 2003): after biochemical, genetic and microscopy research is complete, a mathematical in silico reconstitution serves as a quantitative test for hypotheses about the workings of biological systems. The second role is synthetic, helping to understand how these systems work together. Involvement of modelers in a number of biological areas has varied for the last few decades; it has been unusually intense in studies of actin dynamics. The initial reason for that was the need to understand treadmilling of actin filaments, so the first modeling efforts (Oosawa and Asakura, 1962; Wegner, 1976; Hill and Kirschner, 1982) attempted to quantify the actin treadmill and use thermodynamics to understand the nature and magnitude of the polymerization force. These early studies introduced fundamental ideas that are still used in developing increasingly complex models. A crucial event for modeling was the measurement of the rates of actin polymerization and disassembly (Pollard, 1986) that provided parameter values for the previously conceptual models. In the last 20 years, these initial efforts have led to highly sophisticated models of bulk actin polymerization and its dependence on actin binding proteins. These have treated the effects of mechanisms such as severing and capping (Bindschadler and McGrath, 2007; Matzavinos and Othmer, 2007), and actin-binding proteins such as profilin (Yarmola et al., 2008). Here, we review some modern mathematical, physical and computational models of actin dynamics. There are so many of them already that we have had to limit ourselves to relatively few aspects of actin dynamics. Thus, we do not review models of bulk actin polymerization. We first review recent models of ATP hydrolysis effects on actin filament dynamics, continue with discussion of theories of force generation by polymerization of single actin filaments, and then treat models of dynamic structure and force generation in actin networks. We also briefly consider actin depolymerization, symmetry breaking,

16

Mathematical and Physical Modeling of Actin Dynamics in Motile Cells

383

and wave phenomena in actin networks. Then, we turn to reviewing models of the “organelles” of the actin machinery in motile cells, such as lamellipodia, filopodia, and lamella, in which actin interacts with myosin, membrane and adhesion systems (Fig. 16.1). We conclude by discussing integrative models of whole motile cells. Throughout this review, we discuss modeling methods such as analytical mathematical theories, computational approaches, continuum versus discrete models, and deterministic versus stochastic approximations. We also attempt to illuminate specifically how modeling has been helpful in solving key biological puzzles.

Fig. 16.1 Motile cycle of a crawling cell. The locomotion is based on the actin array treadmill in the “membrane bag”. One long filament growing in the direction of locomotion is shown for clarity. ATP-G-actin assembles onto its barbed end; hydrolysis occurs rapidly in the actin subunits. At the leading edge, the dendritic branching actin network (lamellipodium) forms so that nascent filaments branch off the existent ones where activated Arp2/3 complexes bind. The polymerization is terminated by capping proteins, while “aging” filaments disassemble with the help of ADF/cofilin. Behind the lamellipodial network, thicker and more disorganized lamellum forms, in which the actin mesh is being contracted by myosin motors. Protrusion (1) is generated by actin polymerization coupled with formation and turnover of adhesion complexes, firm at the front and weak at the rear. Retraction (2) depends on action of the contractile lamellar network that pulls the cell body (oval) and “tail” forward

16.2 Polymerization and Force Generation by Single Filaments An understanding of the polymerization properties of individual actin filaments, in the presence of actin-binding proteins, opposing force from the cell membrane, and (in some cases) attachments to the cell membrane, is a key underpinning for any computational model of cell motility. Recent modeling work in this field has focused

384 Fig. 16.2 Schematic of actin filament generating force against obstacle. Black, grey, and open circles denotes different hydrolysis states of actin. Dashed line labeled “Tether” denotes a membrane protein connecting actin to the membrane, which is present in some force-generation models. “F” denotes downward force acting on obstacle, opposing polymerization

A.E. Carlsson and A. Mogilner

i

on the effects of hydrolysis on filament-tip dynamics, the effects of opposing force on barbed-end polymerization, and the effects of barbed-end tethering. Figure 16.2 schematically illustrates these effects acting on a single filament.

16.2.1 Effects of ATP Hydrolysis on Polymerization Dynamics While we have a general understanding of the equilibrium thermodynamics of protein polymerization, the situation for actin is complicated by hydrolysis of ATP in filament subunits (Fig. 16.2). This process changes the polymerization properties of the filament ends in ways that are not completely understood. Building on classic theories (Hill, 1986; Keiser et al., 1986), three recent models have explored the effects of ATP hydrolysis on the dynamics of barbed ends near steady state (Vavylonis et al., 2005; Stukalin and Kolomeisky, 2006), and on the dynamics of actin depolymerization (Li et al., 2009). Vavylonis et al. treated a single filament polymerizing in a solution of ATP-actin. Polymerized subunits progressed from the ATP state to ADP-Pi to ADP. The phosphate release rate (for the second step) was taken much slower than the hydrolysis rate, and both processes occurred randomly along the filament. The off-rate of ADP-actin was taken to be about six times larger than those of ATP-actin and ADP-Pi actin. The effect on the off-rate was taken to depend on the chemical state of the subunit itself, but not on that of neighboring subunits. Consistent with earlier models of the effects of hydrolysis on actin polymerization, and with experimental results, a “corner” at the critical concentration was seen in a plot of polymerization rate vs free-actin concentration, with rapid depolymerization setting in below the critical concentration as the tip became mainly ADP-like. The “tip diffusion coefficient” Dtip , a measure of the fluctuations of the filament tip, was found to be greatly

16

Mathematical and Physical Modeling of Actin Dynamics in Motile Cells

385

enhanced slightly below the critical concentration. These fluctuations are consistent with measurements of single-filament tip dynamics (Fujiwara et al., 2002; Kuhn and Pollard, 2005), except that the theory revealed much smaller fluctuations than the experiments above the barbed-end critical concentration. Stukalin and Kolomeisky explicitly treated the existence of two protofilaments in the F-actin structure, and assumed vectorial hydrolysis, in which a front between entirely ATP and ADP regions propagates along the filament. As in the random-hydrolysis model, an elbow was found in the polymerization rate vs. free-actin concentration plot, and enhanced tip fluctuations were found. However, in this case the fluctuations were seen above the critical concentration, in closer agreement with the experiments. Most theoretical models of single-filament tip dynamics have assumed that ATP-actin tips polymerize fast and depolymerize slowly, while ADP-actin tips depolymerize much faster. This assumption has been challenged by recent experiments in which, after aging, actin filaments become more stable, suggesting a “dynamic stabilization” process (Kueh et al., 2008a). A recent model of actin filament assembly/disassembly (Li et al., 2009) has treated these phenomena using a model that generalizes the previous random- hydrolysis and vectorial-hydrolysis models by including cooperativity of both hydrolysis and phosphate release. Here cooperativity means that hydrolysis and/or release are strongly enhanced when subunits neighboring a given subunit have hydrolyzed or released their phosphate groups. The model includes both the random (no cooperativity) and vectorial (maximal cooperativity) hydrolysis limits. The parametrization incorporated the slowness of ADP-Pi actin depolymerization relative to that of either ATP-actin or ADP-actin (Fujiwara et al., 2007). It was found that when both hydrolysis and phosphate release are strongly cooperative, disassembly of individual filaments follows a time course with characteristic features corresponding to the uncovering of different chemical states of actin. At first, the tip is ATP-like, and disassembly is reasonably rapid. Then the ADP-Pi piece of the filament is exposed, leading to much slower disassembly. Finally the ADP-actin region is reached, and disassembly once again occurs rapidly. This sequence of events is consistent with the observed alternation between fast and slow disassembly. However, it remains to be seen whether this model can reproduce the experimental observation that almost all of the actin filaments are in the slowly depolymerizing phase at long times.

16.2.2 Force Generation by Polymerizing Actin Filaments Recent work exploring the effects of force on polymerization of individual actin filaments builds on the “Brownian-Ratchet” (BR) model (Peskin et al., 1993) in which the growth velocity of a filament perpendicular to an obstacle surface is: V = α exp(−F/F0 ) − β

(16.1)

Here, F0 = kT/a = 1.5 pN is the characteristic force scale over which the growth velocity decreases, k is Boltzmann’s constant, and T is temperature, a is

386

A.E. Carlsson and A. Mogilner

the monomer size, and α and β are on-and off-rate parameters. The stall force in this model is Fmax = (kT/a)ln[A/Ac ],

(16.2)

where A is the free-actin concentration and Ac is the critical concentration, consistent with a rigorous thermodynamic result for equilibrium polymerization (Hill and Kirschner, 1982). The BR model treats polymerization occurring by passive diffusion of monomers to the filament tip (Fig. 16.2). Monomer addition is assumed to occur in a simplified step-function fashion, where the obstacle effects vanish as soon as the distance from the obstacle to the tip exceeds a. The validity of this simplifying assumption has been tested by polymerization simulations for a filament growing in two dimensions, with free monomers moving via explicit Brownian dynamics under the influence of forces from the filament tip and the membrane (Carlsson, 2000). The results showed that for most orientations of the filament relative to the membrane, Eq. (16.1) is a good approximation to the force-velocity relation. However, at certain special orientations, diffusion of monomers to the filament tip is greatly slowed by the obstacle even when it is far enough away to admit the monomer. Recent calculations employing a simplified treatment of monomer addition have elucidated other aspects of force generation by polymerization (Burroughs and Marenduzzo, 2007a). An extension of the Brownian-ratchet model (Carlsson, 2008) to include two chemical states, corresponding to ATP- or ADP-actin, indicated that the stall force can be greatly reduced by hydrolysis. The model treated small numbers of parallel filaments growing against a hard obstacle. An ATP cap at the end of an actin filament formed when polymerization was fast enough, but slowing of polymerization by external force uncovered ADP-like subunits, and caused a transition to a depolymerizing state. Switches between the differing tip states led to large dynamic fluctuations in the obstacle position. The reduction in the stall force was about a factor of five for plausible values of the polymerization and hydrolysis rates. Measurement of the force-velocity relation of individual actin filaments is very difficult because even submicron actin filaments buckle under piconewton forces. However, force generation by small numbers of filaments has been explored by using actin filament bundles grown from acrosomes (Footer et al., 2007), attached to fluorescent beads, in combination with the measurement of deflections in an optical trap. These experiments showed that the stall force of bundles containing 6–8 filaments is much smaller than that expected from rate parameters based entirely on ATP-actin. In addition, the bead position displayed large dynamic fluctuations. These results are consistent with the extended Brownian-ratchet model (Carlsson, 2008) mentioned above. However, because of uncertainties in parameterization of such models, more detailed studies are needed to confirm the relationship between the stall force and ATP hydrolysis.

16.2.3 Barbed-End Tethering When a network of actin filaments generates force against a surface, at least some of the filaments are attached (see Fig. 16.2). Then propulsion takes place either by

16

Mathematical and Physical Modeling of Actin Dynamics in Motile Cells

387

the free filaments fighting tethering forces due to the attached filaments (Mogilner and Oster, 2003), or by the growth of the attached filaments themselves. The earliest work exploring propulsion by tethered actin filaments (Dickinson and Purich, 2002; Dickinson et al., 2002, 2004) assumed a coupling between hydrolysis of ATP-actin and filament elongation, where the release from the tether, required for continued elongation, occurs only after hydrolysis. Because the free energy released by hydrolysis is used for elongation, the stall force produced by this mechanism can be as high as 10 pN or more. The force-velocity relation in this model is very different from that in the BR model. It is constant at small forces and then drops rapidly at a certain critical force. The passive variant of this model – in which tether release is required for elongation but hydrolysis does not aid the process – was explored in a Browniandynamics simulation model (Zhu and Carlsson, 2006). A stall force comparable to that in the BR model was obtained, but the force-velocity relation had a sigmoidal form rather than showing exponential decay. A closely related set of calculations (Shemesh et al., 2005; Shemesh and Kozlov, 2007) have explored the phenomenon of actin polymerization during processive capping by formins. Here actin filament barbed ends grow despite being capped by a formin, which may be attached to a fixed surface. During this polymerization process, it would appear that the formin must rotate, but such rotation is difficult to reconcile with the observation of cross-linked bundles of actin filaments in yeast growing with formins at their ends (Yang and Pon, 2002). The calculations showed that an “open screw” mechanism, in which the formin turns relative to the actin filament, generates a rotation which cancels that resulting from the addition of actin monomers. In models of processive formin capping, the elastic distortion energy of the formin-barbed end complex plays a key role. It was shown that variation of this elastic energy can account for order-of-magnitude variations in the polymerization rate when different formins are present. A calculation focusing on the effects of profilin on elongation (Vavylonis et al., 2006) showed that, because formins have multiple profilin binding sites, they can act as “antennae” that capture profilin-actin complexes and then guide them onto actin filaments.

16.2.4 Disassembly of Actin Filaments One of the most pressing current issues in the modeling of single-filament polymerization/depolymerization dynamics is the rapidity of actin disassembly when exposed to key actin binding proteins found in cells. The combination of proteins cofilin, coronin, and Aip1 (CCA) leads to rapid depolymerization of Listeria tails (Brieher et al., 2006), which is insensitive to the free-actin concentration. Total-internal-reflection (TIRF) microscopy studies of individual filaments in the presence of the CCA proteins showed that disassembly occurs in rapid bursts of about 250 subunits, interspersed with quiescent periods (Kueh et al., 2008b). This is inconsistent with existing subunit-by-subunit models of depolymerization. Since the presence of multiple conformational substates is important in rapid microtubule disassembly, and such substates have been observed in electron-microscopy studies of actin filaments, these substates may well have a role in bursting actin

388

A.E. Carlsson and A. Mogilner

depolymerization. It is to be hoped that such effects are included in future models of actin filament dynamics.

16.3 Structure and Force Generation in Dendritic Actin Networks Actin in cells often occurs in the form of networks held together by branches and crosslinks. The branches are induced by Arp2/3 complex, and crosslinks are caused by several types of crosslinking proteins found in cells. The general scenario of branch induction, as assumed by the “dendritic nucleation model” (Mullins et al., 1998) is that Arp2/3 complex is activated by physical contact with the cell membrane. It then attaches to an existing filament and forms a new branch at a well-defined angle of 70 degrees. The structure and properties of such dendritic actin networks have been the subject of several computational investigations using differing methodologies. These investigations attempt to establish how the growth velocity, density, and filament orientation/length distributions of the network depend on external force, membrane curvature, and the rates of key biophysical processes. There have also been a very limited number of investigations of the mechanisms of network disassembly. We will focus on three issues that have been addressed by computations, although there is considerable overlap between the papers treating these issues.

16.3.1 Force-Velocity Relation A baseline for understanding the detailed growth simulations of the actin network force-velocity relation performed over the past decade is given by a simplified calculation of the lamellipodium growth velocity based on a fixed number of semiflexible actin filaments whose tip thermal fluctuations are sufficient to allow network growth (Mogilner and Oster, 1996). These calculations suggested that an orthogonal network of filaments results from optimization of force generation at a certain filament angle relative to the growth direction, and a growth velocity that decays rapidly as a function of opposing force, with a force-generation capacity of a few pN per filament. The effects of varying filament number, three-dimensional network structure, and stochastic growth/branching were included in a Monte Carlo type model in which an explicit actin network was grown on the computer (Carlsson, 2001). These simulations began with a single filament near a hard obstacle, and assumed that branching occurred only very near the obstacle. The filament generated a dendritic cluster which grew into a network with a well-defined velocity and structure. The most surprising result of these simulations was that the growth velocity is essentially independent of the opposing force. This occurs when branching is limited to the region near the obstacle, because an increase in opposing force causes more filament subunits to be in the region of the membrane. This causes increased branching,

16

Mathematical and Physical Modeling of Actin Dynamics in Motile Cells

389

so that the number of filaments grows with opposing force in self-regulating fashion. A filament number proportional to opposing force would lead to a force per filament (and thus growth velocity) independent of opposing force. The simulations found that the filament number increases nearly linearity with opposing force. A simplified mathematical treatment of branching and nucleation (Carlsson, 2003) showed that when nucleation of filaments occurs by non-branching mechanisms, the velocity depends very strongly on opposing force. The independence of the velocity from opposing force in branching nucleation was confirmed by later stochastic-growth simulations based on autocatalytic branching (Schaus et al., 2007). A model having a more detailed treatment of the interaction between filaments and the membrane, as well as filament bending (Schaus and Borisy, 2008), showed that “work-sharing” can optimize propulsion. In work-sharing, the distribution of filament-load distances self-organizes in such a way that an entering monomer can attach at an optimal location. In this way effective size of a monomer is decreased. This effect is enhanced by a flexible membrane. In this model, certain values of the parameters controlling the detachment rate cause the force-velocity relation to be flat at small forces, and then drop at higher forces. This behavior, and those of other recent force-velocity curves, are shown in Fig. 16.3a. Earlier force-velocity curves are given in a recent review (Mogilner, 2006). The force-independence of the velocity occurs when the force scale characterizing the detachment kinetics is large. a

b

Fig. 16.3 (a) Calculated force-velocity relations. Solid line: stochastic simulation (Schaus and Borisy, 2008). Dashed lines: passive-processive model with approximate inclusion of branching and deformation (Zhu and Carlsson, 2006). Solid circles: liquid-cluster model (Lee and Liu, 2009). (b) Measured force-velocity relations. Solid line: cantilever measurement of moving bead (Marcy et al., 2004), fitted to a smooth curve. Dashed line: cantilever experiment with actin network grown on flat substrate (Parekh et al., 2005)

While the treatments described above have regarded the dendritic network as mechanically rigid, a complementary approach has viewed it as a liquid of dendritic clusters (Lee and Liu, 2008). This work treats actin-based propulsion of a rigid disk through a viscous fluid. The simulation begins with a few seed filaments, and new actin filaments are nucleated as branches on existing filaments, in a region

390

A.E. Carlsson and A. Mogilner

near the disk. The clusters grow to a finite size and diffuse in the liquid, exerting contact forces on the disk when they impinge it. The propulsion speed obtained by this method varies nonmonotonically with the concentrations of capping protein and Arp2/3 complex. Calculation of the force-velocity relation within this model (Lee and Liu, 2009) showed that the force is constant at low forces, but drops rapidly beyond a critical force (Fig. 16.3a). The force-independence arises because the velocity is set by the net polymerization rate; since polymerization is spread out through the actin network, rather than being concentrated at the leading edge, the polymerization is not very sensitive to the opposing force. The point at which the velocity turns down is related to the pressure required to deform the actin network. Measurements of the force-velocity relation of actin networks have treated both propulsion of flat surfaces, and propulsion of smaller obstacles such as Listeria or biomimetic beads. The force-velocity relation of an actin network nucleated by Arp2/3 complex, growing against a flat surface, was measured via a cantilever method using several different force protocols (Parekh et al., 2005). When the cantilever base was held fixed, so that the force varied during polymerization, eventually reaching a stall force, it was found that the growth velocity was constant at small forces but dropped rapidly at higher forces (see Fig. 16.3b) as in the liquid-cluster model (Lee and Liu, 2009). However, a pronounced loading-history dependence of the force was also observed. When the force was ramped up to a certain level, then reduced to a lower level, the velocity was about twice as large as it was the first time the lower force was reached. The most plausible explanation of this effect is a change in the structure of the network caused by the opposing force. A subsequent cantilever-based study treated force generation of actin filaments nucleated by a nonbiological surface, where nonspecifically interacting filaments rather than a branched network are expected (Greene et al., 2009). A surprising result of this study was a “negative hysteresis” – the force opposing compression of the gel was larger during the retraction phase than during the compression phase. In addition, oscillations of the obstacle position, similar in magnitude and frequency to those seen for acrosome-propelled beads (Footer et al., 2007) were observed. The origin of these phenomena is not clear, but it was speculated that the negative hysteresis could result from stiffening of actin filaments as a results of a structural change caused by compression. The most readily interpretable experiment of the force-velocity relation of an actin-propelled bead (Marcy, 2004) was also performed using a cantilever setup. The bead was coated with the actin polymerization activator N-WASP, and the measurements were performed in a pure-protein medium. The measurements showed an initial rapid decay with opposing force, followed by a region of slower decay of velocity (Fig. 16.3b). This experiment is not explained by any of the theories described above (see, however, the Chapter 17 in this book), but Zhu and Carlsson (2006) showed that the results could be fitted by a passive-processive polymerization model (Fig. 16.3a) assuming autocatalytic filament generation. Some of the assumptions made in the above theories of force generation have been challenged by recent experiments. The theories assume that space for new monomer insertion is created by the fluctuations of the tips of individual filaments away from the membrane. This assumption has been challenged by experiments

16

Mathematical and Physical Modeling of Actin Dynamics in Motile Cells

391

measuring the velocities of actin-propelled beads in confined geometries (Shaevitz and Fletcher, 2007), which showed that the beads slow considerably when they are close to a wall. This result was attributed to the reduced diffusion constant of the bead near the wall. If the change in the bead diffusion constant is responsible for the reduction in speed, then it must be diffusive motions of the bead, rather than fluctuations of individual filament tips, which allow monomer insertion. Another mechanism that could slow the bead is the effect of the walls on the free-actin concentration around the bead, which has not yet been evaluated. In addition, most current theories of force generation assume that capping proteins aid network growth by maintaining an increased supply of free actin (the “funneling hypothesis”). A recent study of actin-propelled beads in a pure-protein motility medium showed that CP increases the ratio of Arp2/3-induced branches to F-actin in the actin tail (Akin and Mullins, 2008). Furthermore, CP reduced the concentration of free actin. These results suggest that CP acts synergistically with Arp2/3 complex to produce new branches, rather than increasing the free-monomer concentration.

16.3.2 Dendritic Network Structure Calculations of the filament orientation distribution using first a rate-equation model (Maly and Borisy, 2001), and subsequently a Monte Carlo type model (Schaus and Borisy, 2007), found a sharp preference for filaments oriented at 35◦ away from the direction of growth. Typical network structures are shown in Fig. 16.4a. The 35◦ preference results from the preferred dendritic branching angle of 70◦ and a competition between filament creation by branching, and capping which removes filaments from the actively growing population. As the orientation distribution evolved from a homogeneous starting state, it was found that at most orientations capping exceeded branching, so few filaments were found having these orientations. However, at the preferred orientations, branching balanced capping, and a strong peak in the distribution was seen at these orientations. The formation of the peaked distribution required protection from capping near the leading edge. Atilgan et al. (2005) used a stochastic-growth method to study the relationship between the molecular details of branch formation by Arp2/3 complex and the network’s orientation distribution. As Fig. 16.4b shows, when the orientation of Arp2/3 complex is restricted during each branching event, the orientation has a two-peaked structure similar to those seen in the other simulations; however, when this orientation is isotropically distributed, the two-peaked structure vanishes. The two-peaked structure is in agreement with several experimental studies (Small et al., 1995, Small and Celis, 1978; Maly and Borisy, 2001), with the preferred angle varying from 25 to 45◦ . However, this is a very limited dataset for comparison, and a fully three-dimensional calibration of the simulations would be helpful. Such a calibration would include the statistics of crosslinking and the topology of the network as defined by branching and crosslinking. Part of the difficulty is that there is no simple, physically consistent model to use as a starting point. Existing

392

A.E. Carlsson and A. Mogilner

Fig. 16.4 (a) Three-dimensional actin network, grown by stochastic simulations (Schaus and Borisy, 2007). Red dots denote barbed ends, blue denote branches. Viewed region is 0.5 × 0.5 μm. (b) Solid lines: filament orientation distribution obtained from stochastic simulation including no orientation restriction of branching (left frame) or only allowing upward branching (right frame) (Atilgan et al., 2005). Dashed lines: experimental orientation distribution (Maly and Borisy, 2001)

models of branched networks, such as Cayley trees, have a very unrealistic dependence of density on system size, and regular lattices found in three dimensions also seen to be inappropriate.

16.3.3 Symmetry Breaking and Hopping of Listeria and Actin-Propelled Beads There are several phenomena specific to Listeria and actin-propelled beads whose understanding may shed light on the processes underlying cell migration.

16

Mathematical and Physical Modeling of Actin Dynamics in Motile Cells

393

Spherically symmetric protein-coated beads, when placed in a motility medium, can break the initial symmetry and start moving persistently in a well-defined direction; furthermore, the asymmetry of the ActA distribution on Listeria is believed to be closely related to the motion of the bacterium. Both Listeria and beads can move in a “hopping” fashion, with periodic variations of the velocity. These phenomena have been treated using the “elastic” theory of Listeria/bead propulsion. However, we will not treat this methodology here because it is reviewed in the Chapter 17 in this book. The first treatment of symmetry breaking of actin-propelled beads (Van Oudenaarden and Theriot, 1999) focused on the nonlinearities in the force-velocity relation of single filaments. When these nonlinearities were treated in a multifilament model of bead propulsion, a symmetry-breaking effect was found. Another possible mechanism is suggested by the “competition for resources” model: if two oppositely oriented F-actin arrays compete for a common pool of resources, i.e. Arp2/3 complexes responsible for F-actin branching, then as a result of a spontaneous fluctuation, more than half of these resources could attach to filaments from one array. Then more than half of the daughter filaments that grow in the same direction as the mother filaments would belong to that array. In the next generation of filaments, the greater number of filaments from the winning array would get a proportionately larger share of the resources, and sprout a yet greater number of filaments. Such a competition for resources causes winner-takes-all dynamics that leads to symmetry breaking (Sambeth and Baumgaertner, 2001). A subsequent study of a very small bead propelled by actin polymerization (Burroughs and Marenduzzo, 2007b) found hopping motion resulting from detachment of the bead from the tail. However, experimental studies (van der Gucht et al., 2005) have shown that tearing of the actin gel on the bead occurs during symmetry breaking, and most subsequent computational treatments have focused on this effect. A method intermediate between the elastic theory and the detailed dendritic network simulations described above, called the Accumulative Particle-Spring (APS) model (Dayel et al., 2009), has been used to study both symmetry breaking and hopping motion of proteincoated beads. In this model, the actin network is represented as a network of nodes held together by links (although the links do not correspond to individual actin filaments). New nodes and links are created at the bead surface, corresponding to actin polymerization. The links are linearly elastic until they reach a finite breaking strain, and there is a repulsive force between links. This model confirmed a tearing mechanism of symmetry breaking for spherical beads. It also reproduced several observed features of the actin gel during the symmetry-breaking process, including geometrical parameters such as the time dependence of the crack expansion, the circumferential stretching at different points in the actin gel, and radial stretching of the gel. It predicted that uniformly protein-coated ellipsoidal beads should move sideways – perpendicular to the long axis. This prediction was confirmed by experiments on such beads. Finally, the simulations showed that hopping motion tends to occur as the network becomes denser, by a process of repeated tearing of the gel. A detailed three-dimensional model of actin network growth on Listeria has been used to study the dynamics of its motion and the dependence of speed on

394

A.E. Carlsson and A. Mogilner

the polarity of the ActA distribution (Rafelski et al., 2009). In this model, new filaments are generated stochastically by branching on existing filaments. The effects of tethering to the bacterial surface are included, spatial dependence of the free-actin concentration, and ATP hydrolysis on actin filaments are included. This method has been shown to produce a realistic actin-tail structure, and intermittent motion of the bacterium (Alberts and Odell, 2004). This method yielded a positive correlation between ActA asymmetry and propulsion speed, consistent with experimental results obtained in the same work (Rafelski et al., 2009), but only if a cooperative restraining mechanism for gel sliding along the surface was included.

16.3.4 Spontaneous Waves of F-Actin A growing number of observations, both by confocal fluorescence microscopy (Vicker, 2002) and total-internal reflection fluorescence microscopy (Bretschneider et al., 2004; Gerisch et al., 2004; Weiner et al., 2007; Bretschneider et al., 2009) have shown that actin in cells can form self-organized traveling waves and/or moving patches. These have been treated by recent theoretical models based on an activatorinhibitor scenario. In this scenario, the concentration of an activator builds up by positive feedback, a wave front moves by a combination of this feedback with diffusion, and the activator concentration is reduced by inhibitor buildup farther back. These models assume the existence of a nucleation-promoting factor (NPF) that activates Arp2/3 complex to generate new filaments. The models assume that the positive feedback and diffusion driving the waves are either in the F-actin (“actinfirst”) or in membrane-bound nucleation-promoting factors (“NPF-first”), which act upstream of actin polymerization. A coarse-grained model treated an actin-first mechanism with delayed inhibition from a membrane-bound agent (Whitelam et al., 2009). Assuming a cubic positive feedback of F-actin, spontaneous polarization of actin filaments, and diffusion-like spreading of F-actin caused by new filament generation at the patch edge, a scenario of patches growing into waves was seen. NPF-first mechanisms have been treated by Weiner et al. (2007) and Doubrovinski and Kruse (2008), where the NPF is Hem-1 and F-actin inhibits Hem-1 by detaching it from the membrane. The two approaches differ in their treatment of Hem-1 cooperativity and F-actin dynamics, but both yield spontaneous waves of Hem-1 and F-actin; Doubrovinski and Kruse (2008) also found regimes corresponding to actin patches. Thus there are plausible theories of actin waves at this point, but as yet no way of definitely establishing what is the correct theory. Progress in this direction might be made by more detailed modeling including an explicit network structure, such as those that are currently being used to study force generation and network structure.

16.4 Dynamics of Cell Protrusions and the Cell Periphery We now turn to the role of actin dynamics in the behavior of protruding and retracting organelles around the cell periphery. As described in recent reviews (Flaherty

16

Mathematical and Physical Modeling of Actin Dynamics in Motile Cells

395

et al., 2007; Carlsson and Sept, 2008; Mogilner, 2009), the details of interactions of (i) actin with (ii) adhesions, (iii) myosin, and (iv) the cell membrane are crucial in determining cell movements. The web made up by these interactions is very complex and poorly understood in its entirety (Ridley et al., 2003). For this reason, relatively abstract, conceptual models of the interactions of these four components (Veksler and Gov, 2007; Shlomovitz and Gov, 2007) are useful for delineating spatial-temporal patterns at the cell edge. These models are based on coupling the mechanical free energy of an undulating membrane with an actin network that grows and pushes on the membrane, and, in turn, is pulled on by myosin contractile forces. In addition, adhesion molecules undergoing biased diffusion in the membrane plane (these molecules are assumed to be curved and tend to prefer certain bent configurations of the membrane) affect growth and movements of the actin network. In spirit, the equations of these models follow the traditional theoretical physical approach, in which variables’ dynamics are surmised from symmetry and conservation laws, rather than being data-driven. Therefore, the resulting partial differential equations do not relate to any specific cell, but instead offer very general concepts and are amenable to linear stability analysis, which demonstrates that a featureless and uniform cell edge can spontaneously become unstable. Specifically, the edge can become phase-separated, with either periodically spaced large aggregates of actin/adhesions structures accompanied by membrane protrusions, or one large lump of actin/adhesions. Lateral waves of such structures/protrusions along the cell edge are also possible. These models’ predictions are useful for a more detailed examination of the actin structures at the cell edge, such as lamellipodia, lamellae and filopodia.

16.4.1 Filopodia These organelles (reviewed in Gupton and Gertler, 2007) are dynamic bundles of actin filaments protruding forward from the motile cell leading edge (Fig. 16.5). Cells use this structure either as a sensor (to probe the environment for chemical cues) or as a mechanical device (to penetrate the surroundings and pull the cell body up), or both. While filopodial function is not ripe for mathematical examination, two fundamental questions – how filopodia emerge and elongate – have been addressed vigorously in the last few years. Two mechanistic hypotheses about filopodial emergence were suggested. One of them posits that the actin bundle at the leading edge is nucleated by a specialized molecular structure (Steketee et al., 2001). Another hypothesis is that barbed ends of Arp2/3-induced filaments of the growing dendritic network at the leading edge converge, stick together and elongate as a proto-bundle normal to the edge (Fig. 16.5), subsequently capturing and bundling together a few other filaments (Svitkina et al., 2003). This highly explicit and mechanistic nature of the second mechanism makes it amenable to mathematical analysis, which can test the hypothesis’s physical feasibility. The filaments have to bend in order to get bundled (Fig. 16.5), but the increase in the elastic energy can be compensated by a decrease in binding energy of bundling proteins. Elegant theoretical examination of this energy balance supplemented with

396

A.E. Carlsson and A. Mogilner

Fig 16.5 Filopodial dynamics. (a) Convergence-elongation model for the initiation of the filopodium. (b) The actin bundle grows by some filaments advancing forward into space created by local membrane undulation, and then propping up the membrane allowing other filaments to polymerize. Buckling of the growing actin bundle is prevented by an increase of the bending energy of the membrane “wrapping” the bundle up. Capping of the filaments and their subsequent retraction due to the retrograde flow effect leads to a finite lifetime of the filopodium

Monte Carlo simulations and in-vitro reconstitution (Ideses et al., 2008) showed that the bundles’ thickness and length is controlled by the filaments’ lengths, spacing, and ultimately by the initial concentration of Arp2/3 complex. Besides the feasibility of the dendritic filament bundling process, an important question is how fast this could occur at realistic filament spacings. By performing a complex statistical physical analysis of stochastic dynamics of normal modes of the filaments’ thermal writhing, Yang et al. (2006) found that the bundling time can be surprisingly short. Another relevant recent analysis (Kraikivski et al., 2008) revealed that the filaments can converge not just tip-to-tip, but also in a tip-to-side fashion. Modeling also suggests that filopodial emergence can be membrane-induced (or assisted) (Atilgan et al., 2006; Liu et al., 2008): local membrane deformations arise as actin filaments polymerize against the membrane. Two such proximal deformations can merge to create a larger deformation that gathers additional filaments. Deformations which fail to gather additional filaments are stalled and diminish, but if enough filaments are bundled to overcome the membrane resistance to membrane tube formation, a “proto-filopodium” can elongate without further physical constraint until the free actin concentration is depleted. Again, these theoretical conclusions are supported by in vitro experiments (Liu et al., 2008). Another side of the filopodium dynamics – their elongation and maintenance – was first examined quantitatively by Mogilner and Rubinstein (2005), who demonstrated that if too few filaments are bundled together, then the filopodium will buckle, but too many filaments will deplete the G-actin pool. The optimal filopodial thickness predicted as a balance between these two factors (by estimating buckling mechanics and solving analytically the differential equations of G-actin transport)

16

Mathematical and Physical Modeling of Actin Dynamics in Motile Cells

397

turned out to be a few tens of filaments, as observed. This first model ignored subtle membrane fluctuations. Subsequent modeling efforts (Atilgan et al., 2006; Lan and Papoian, 2008) included the membrane dynamics. By performing stochastic simulations of the membrane deformations and filaments’ polymerization, they found that some leading filaments can prop up the membrane locally allowing other filaments to catch up, hence easing filopodial protrusion (Fig. 16.5). Another insightful physical analysis using energy estimates and Monte Carlo simulations, counter-intuitively pointed out that the presence of an enclosing membrane stabilizes the actin bundle against buckling (stretching and bending the membrane tube costs energy), rather than causing it (Pronk et al., 2008). Indeed, the filament bundle does not remain straight. In a narrow filopodium, the membrane will accommodate the filaments’ deflection by deforming congruently. Since any bending of the filament will lead to bending – rather than shortening – of the membrane, the compressive force that acts to shorten the membrane tube will be exerted along the contour of the filament (Fig. 16.5). Though compressive, this force will in effect counteract buckling by adding membrane bending energy to the filament bending energy. Last, but not least, two recent papers took into account the stochastic nature of filopodial protrusion (Lan and Papoian, 2008; Zhuravlev and Papoian, 2009). Stochastic effects are crucial in any serious modeling of actin-cell migration phenomena, as the systems of interest are usually mesoscopic: the copy numbers of the principal molecules involved are too large to allow practical use of moleculardynamics methods, but too small for reliable use of mean-field methods and statistical mechanical averaging. By simulating exact stochastic dynamics (using the Gillespie method) of coupled membrane deformations, actin monomer diffusion, fast filament growth, retraction modulated by slow capping (Fig. 16.5), and formin kinetics, the authors discovered that the discrete noise of small number of regulatory proteins is amplified, inducing macroscopic instability into filopodial dynamics. Namely, due to these stochastic effects, the filopodial length fluctuates greatly, and the filopodial lifetime becomes finite. It is noteworthy that the spatially resolved stochastic simulations employed in these models are very computationally intensive – they run in parallel on more than a hundred processors for days – pointing out future challenges and suggesting that when we attempt realistic multiscale modeling of migration of even individual cells, computerrelated issues will be formidable. Much creative thinking and many approximations will probably be needed for decades to come, and modelers are unlikely to be unemployed. Finally, interactions of filopodia with adhesions and myosin were recently modeled, and predictions confirmed experimentally by Chan and Odde (2008). This model treated the adhesion complexes as molecular clutches and their dynamics as “frictional slippage”, a behavior that has previously been observed on stiff glass substrates. Solving simple differential equations derived from force-velocity balances, the authors showed that on stiff substrates, molecular clutches engage the actin bundle but abruptly disengage as a rapid building of tension within engaged clutches drastically shortens actin/clutch interaction lifetimes. In this case, the filopodium

398

A.E. Carlsson and A. Mogilner

is continuously slipping backward at a roughly constant velocity. However, on soft substrates, the substrate compliance slows the rate at which tension builds along individually engaged clutches, prolonging actin/clutch interaction times. During this time, there is little relative motion between the actin bundle and the substrate as tension slowly develops within the substrate. Because of this lack of resistance, myosin motors work near their unloaded sliding velocity, leading to high rates of actin retrograde flow. As the substrate strains and greater tension builds, the clutches largely remain engaged, providing considerable resistance to the myosin force, substantially slowing the retrograde flow. Eventually, the load becomes so great that the tension developed along the engaged clutches results in lower mean traction forces and a higher failure rate, in which the unsupported load shifts progressively to remaining adhesions, further destabilizing the actin/clutch interaction. This quickly leads to an abrupt coupling failure where all clutches rapidly disengage, unloading the substrate and causing it to snap back to its initial rest position. Thus, computational modeling of myosin-adhesion motility on soft substrates predicts the natural emergence of an oscillatory “load-and-fail” filopodial dynamic. Similar effects in the stick-slip adhesions at lamellipodial protrusions predicted by Wolgemuth (2005) can be relevant to observed periodic extension-retraction cycles at the cell leading edge (Giannone et al., 2007). Another potential mechanism for oscillation of either the cell leading edge or objects propelled by actin networks, also in the same spirit, is suggested by the analytical model of Gholami et al. (2008). In this model, actin filaments transiently attached to the edge or obstacle break their attachments simultaneously, when pushing force by detached filaments increases above a certain threshold. Then the protrusion is rapid for a while, until enough filaments attach again and stall the protrusion, after which the cycle starts anew. An effect similar in spirit was previously predicted computationally (Alberts and Odell, 2004).

16.4.2 Lamellipodium and Lamellum Most motile cells have lamellipodia leading the way in migration, but between the lamellipodium and cell body, another actin network – the lamellum – is usually located (Fig. 16.6a). The coupled geometry, biophysics, and biochemistry of these two motile appendages are vigorously debated (Koestler et al., 2008; Vallotton and Small, 2009; Danuser, 2009). One contested view is that the lamellipodium and lamellum are different biochemically and have very different dynamics so that the lamellipodium is simply “riding” on top of the lamellum and not even necessary for protrusion. Another view is that lamellipodium and lamellum are not overlapping, and that there is a boundary between them defined by the set of points where myosin becomes engaged. This debate is too recent for modeling to offer a significant insight; besides, the questions are experimental and rather technical having to do with accurate tracking of fluorescent actin speckles. However, a few recent models have made predictions about the actin architecture relevant to the lamellipodium/lamellum question.

16

Mathematical and Physical Modeling of Actin Dynamics in Motile Cells

399

Fig 16.6 (a) Lamellipodial actin network at the leading edge (see Fig. 16.1) treadmills rapidly. According to the model of Huber et al. (2008), this network consists of short filaments that turn over rapidly due to the cofilin action. To the rear from it, there is the lamellar network of “surviving” long filaments associated with myosin and stabilized by tropomyosin. (b) According to Shemesh et al. (2009), growth of the lamellipodial network against the membrane at the leading edge (arrows) and against the focal adhesions at the lamellipodium/lamellum interface generates elastic stresses (illustrated with shadowing, dark corresponds to higher stress value). When a threshold stress is reached at the wavy boundary around the focal adhesion, the network breaks and disassembles

One of these models is rather unique (Ditlev et al., 2009): it uses the Virtual Cell computational framework (Slepchenko et al., 2003) to couple together in minute detail almost all actin dynamic processes including activation of Arp2/3 at the cell membrane by N-WASP, nucleation and branching on preexisting F-actin by activated Arp2/3, dissociation of Arp2/3 branches, inter-conversion between the ATP, ADP, and ADP-Pi forms of both G-actin and F-actin, acceleration of nucleotide exchange on G-actin by profilin, addition of G-actin associated with profilin to barbed ends, capping of the barbed ends, annealing and fragmentation of actin filaments, buffering of G-actin by thymosin, and severing and accelerated disassembly of actin filaments by ADF/cofilin. The resulting 3D model consists of 48 partial differential equations and 12 algebraic equations, the latter corresponding to fast reactions for which a pseudo-steady-state approximation was applied. This tremendous level of detail has the advantage of reducing the number of assumptions and simplifications that early models of the actin network dynamics (Mogilner and Edelstein-Keshet, 2002) had to make, but it also has a price: calculations require days of computer time, and the results are not always easy to interpret and explain in words. One commendable feature of this model is that it is “open” for the community to change, adapt and test various hypotheses. Simulations of this model

400

A.E. Carlsson and A. Mogilner

produced many results, one of which is the prediction of a sharp boundary between polymerization and depolymerization regions emerging from the interplay of rearward flow of the actin network, barbed end capping, and dissociation of Arp2/3 branches to expose free pointed ends. The authors suggest that these regions, seen in some speckle microscopy experiments, may correspond to the lamellipodium and lamellum, respectively. Another recent model (Huber et al., 2008) posits a simplified system of a small number of partial differential equations, sensibly ignoring minute details of the hydrolysis cycle and exact mechanisms of actin turnover, but instead adding interactions with tropomyosin, a protein which stabilizes actin filaments. Numerical analysis of these equations, that can be readily interpreted, shows that the actin network can separate into a frontal part (virtual lamellipodium) where the filaments are largely associated with ADF/cofilin and a rear part (virtual lamellum) where the F-actin is associated with tropomyosin (Fig. 16.6a). The model makes the testable prediction that a necessary condition for this separation is a significant difference between ADF/cofilin and tropomyosin unbinding rates. Furthermore, the model makes two predictions that compare well with prior experimental observations: (i) The lamellipodium is made of shorter filaments: depolymerization causes predominant disappearance of short filaments farther from the front, and annealing additionally leads to a majority of long filaments in the lamellum. (ii) The bulk of depolymerization takes place at the boundary between the lamellipodium and lamellum. The models reported in Huber et al. (2008) and Ditlev et al. (2009) are based on conservation laws for the numbers of molecular species involved, and thus take into account only transport and chemical turnover in the actin networks. Another recent model (Shemesh et al., 2009), in contrast, suggests a mechanical explanation for the lamellipodium/lamellum boundary (Fig. 16.6b). This model treats the lamellipodium and lamellum as elastic shells. According to the authors’ experimental observations, the boundary between the lamellipodium and the lamellum is demarcated by periodically spaced adhesions. The model assumes that growing filaments at the leading edge push against the cell membrane and thus drive the whole lamellipodial shell to the rear. The contacts between the adhesions and the lamellipodium resist this push, and as a result the lamellipodial shell deforms. Numerical solution of elasticity equilibrium equations demonstrated that the resulting elastic stress reaches a maximum at the boundary, displaying a characteristic concave arc shape between focal adhesions. The key feature of this model is a threshold stress at this boundary. When this stress is reached, the actin network ruptures and disintegrates, as if the rapid retrograde flow of the lamellipodial actin network, created by polymerization at the membrane, breaks against the adhesions that act like jetties and arches between them. A new boundary between the lamellipodium and lamellum forms along a new row of adhesions appearing in the lamellipodium. As a result, the lamellipodium/lamellum interface (and the whole cell with it) advances in cyclic leaps between the equilibrium and nonequilibrium shapes. A different mechanical scenario is suggested by Giannone et al. (2007), according to which the lamellipodial actin network lies above the lamellar network, and

16

Mathematical and Physical Modeling of Actin Dynamics in Motile Cells

401

myosin clusters periodically pull the lamellipodium rearward relative to the lamellum, which leads to buckling upward of the whole leading edge. The resulting breaking and re-growth of the lamellipodium lead to cycles of protrusion-retraction. Numerical modeling of this qualitative scenario would be informative. In the future, increasingly realistic modeling will have to take into account not only interactions between the actin, membrane, adhesion and myosin subsystems, but also the regulatory dynamics of G-proteins: a recent study (Machacek et al., 2009) revealed that local spatio-temporal oscillations of Cdc42, Rac1 and RhoA govern the leading edge dynamics, so that RhoA has a role in the initial events of protrusion, while Rac1 and Cdc42, antagonistic to RhoA, activate pathways implicated in stabilization of nascent protrusions. Also, there is an open question whether there is a connection between various investigations of mechanisms causing leading edge oscillations (Giannone et al., 2007; Shemesh et al., 2009; Machacek and Danuser, 2009), and those causing actin waves (Weiner et al., 2007; Doubrovinski and Kruse, 2008; Bretschneider et al., 2009; Whitelam et al., 2009).

16.5 Shape and Movements of the Whole Cell 16.5.1 Boundary Mechanics Models The shape of motile cells is defined by local balances between expansion and retraction at their dynamic boundaries. Phenomenological description of such boundary dynamics is a problem to which mathematical modeling has contributed significantly. For simple shaped cells like fish keratocytes, the phenomenology is based on a simple geometric principle posited in the graded radial extension model (Fig. 16.7a) (Keren et al., 2008). This model describes a steady state situation in which local advancement of the cell boundary is perpendicular to the cell’s boundary. To maintain the shape, the magnitude of the extension must be graded, from a maximum at the center of the leading edge to zero at the sides, while at the rear the retraction rate increases from the sides to the maximal value at the midpoint. A simple trigonometric formula expresses the cell shape through the ratio of the graded local extension rate and cell speed. The molecular mechanisms generating the graded extension rates responsible for maintaining this keratocyte geometry were recently explained by the observation of a graded distribution of actin filament density along the leading edge (Keren et al., 2008), which could result from a self-organization process. Namely, along the edge, growing filaments compete for resources (the molecular identity of which is unknown, but Arp2/3 complex may be involved) to branch out nascent filaments, while existing filaments become capped and lag behind the protruding edge. At the rear corners of the cell, the density of actin filaments is damped, perhaps by the large adhesion complexes there. Filaments at the center of the leading edge can outcompete filaments at the sides because they are not inhibited by the adhesions at the sides, so the actin density is peaked at the center. Analytical solution of partial

402

A.E. Carlsson and A. Mogilner

Fig. 16.7 Dynamics of the cell boundary and interior. (a) The graded radial extension model: before capping, the growing actin filaments abutting the membrane at various angles (dashed segments) effectively push the leading edge membrane outward in the locally normal direction. Graded rate of the local actin network growth determines the leading edge shape according to simple trigonometric formula. (b) In the models of Satulovsky et al. (2008) and Stéphanou et al. (2008), the cell boundary expands and retracts radially (dashed lines) from the cell center (black dot) according to the protrusion and retraction signals resulting from reaction-diffusion kinetics (Satulovsky et al., 2008) or actin growth and contractile mechanics (Stéphanou et al., 2008) illustrated with shadowing (dark/light corresponds to protrusion/retraction signal, respectively; arrows show the direction and magnitude of the local protrusion/retraction). (c) In the cell interior, actin (segments) – myosin (double arrow) network is viscoelastic (illustrated with spring and dashpot in series), contractile, and coupled to the substratum by viscous-like adhesions

differential equations describing these actin dynamics yielded an inverted parabolic actin filament distribution, as observed experimentally. These actin filaments push the cell membrane from within. At the center of the leading edge, where the actin filament density is high, the membrane resistance per filament is small, allowing filaments to grow rapidly and generate protrusion. As the filament density gradually decreases toward the cell sides, the load force per filament due to membrane tension increases. As a result, local protrusion rates decrease smoothly from the center toward the sides of the leading edge, causing the leading edge to become curved as observed. At the rear of the cell, membrane tension, assisted by myosin contraction, crushes the weakened disassembling actin network, thereby retracting the cell rear. The geometry of more complex, not necessarily constant, movements and shapes, can be described by a rule-based model (Fig. 16.7b), in which a cell is modeled by its perimeter (Satulovsky et al., 2008), and the local rates of protrusion and retraction are regulated by a model that incorporates local stimulation – global inhibition of protrusive activity. The evolution of local protrusion is calculated from lateral propagation and decay of protrusion signals, with a positive feedback increasing protrusion in the already protruding region. The retraction is governed by a rule, in which the retraction rate is constant and proportional to the total protrusive activity.

16

Mathematical and Physical Modeling of Actin Dynamics in Motile Cells

403

In this model, the points along the cell perimeter either protrude or retract along imaginary spokes radiating from the centroid of the cell. This model, by solving numerically simple reaction-diffusion equations along the cell boundary, reproduced the characteristic shape and trajectory of amoeboid Dictyostelium cells and, by simply varying the parameters, mimicked other motile cell shapes. For example, addition of focal adhesions resulted in wedge-shaped cells with tails, characteristic of fibroblasts. The model of Satulovsky et al. (2008) illustrates that a few changes in the distribution of protrusion/retraction along the boundary are sufficient to recapitulate the observed diversity in cell shapes. To advance beyond such phenomenology, the underlying biophysics and biochemistry of the actin cytoskeleton has to be modeled. A very insightful qualitative analysis in (Lämmermann and Sixt, 2009) points out that various spatial-temporal combinations of three processes – protrusion, contraction and adhesion – can be used by the cell to adapt to any 2D or 3D environment and to produce any mode of locomotion that has been observed. Detailed mathematical modeling of these processes is in its infancy. One of the early examples is the work of Stéphanou et al. (2008) that, in a sense, can be considered as a biophysical extension of the abstract model of Satulovsky et al. (2008). Specifically, Stéphanou et al. (2008) suggested that membrane protrusions are induced by hydrostatic pressure (see below) and are opposed by tension from the actin filaments linked to the membrane and actomyosin contractility. The model includes simple dynamics of actin turnover, and formation and maturation of adhesions, and takes into account cytoplasm viscosity, filament and membrane elasticity, and actomyosin contraction tonicity. All these detailed sub-processes are mathematically encoded in partial differential equations for variables on the free radially moving cell boundary, conceptually similar to the description of Satulovsky et al. (2008). Satisfyingly, numerical simulations of the detailed model predict a diverse gallery of cell shapes including spontaneously pulsating cells.

16.5.2 Actin-based Models All models of the cell shape described above treat the empirically obtained, effective mechanics of the cell boundary. To advance further, the assumed dynamics of the cell edge must be derived from models of the actin cytoskeleton (2D or 3D, depending on the type of motility). Mathematically, such models use either the Lagrangian or Eulerian formalisms. In the former, one follows trajectories of the material points of the cytoskeleton. In the latter, evolution of the actin velocity field at a spatial grid is computed. In principle, the first methodology is more suitable for an elastic cytoskeleton that deforms, rather than flows, under stress, while the second approach is a better fit for a viscous actin network that flows when a force is applied. A perfect example of the latter is the reactive interpenetrating flow theory of Dembo and collaborators (Herant et al., 2003). In this theory, the actin polymer network is considered to be a very viscous Newtonian fluid moving through the fluid cytoplasm. These two fluids can squeeze through each other, with an effective

404

A.E. Carlsson and A. Mogilner

friction, hence the name of the model. A crucial aspect of the theory is that the actin fluid is active: besides passive viscous stresses, there is an active contractile stress produced by myosin, which generates forces that drive cytoskeleton deformations and cytoplasmic flows. This theory has been very successful in describing amoeboid cell movements. The simulations are very involved and require highly nontrivial numerical solution of large systems of partial differential equations. Examples of the Lagrangian description of the cytoskeleton are purely elastic 2D models of lamellipodia (Rubinstein et al., 2005; Shemesh et al., 2009), in which the equations of linear elasticity theory have to be solved numerically to compute small deformations and stresses of the 2D elastic shells. Technically, such simulations are much simpler than the flow computations. The mechanical properties of the deformable actin cytoskeleton depend on the relevant time and spatial scales (Howard, 2001; Boal, 2002): if the fast (order of seconds) local mechanics of the cytoskeleton is of interest, then it is likely that actin filaments are firmly crosslinked by actin binding proteins, and the relevant behavior is more elastic than viscous. The situation is opposite if slow (order of minutes) movements of the whole cell are of interest, on whose scale individual crosslinks unbind and bind many times, and actin filaments flow relative to each other. Thus, perhaps the most accurate description is a viscoelastic one (Fig. 16.7c). In fact, there is the whole spectrum of viscoelastic behaviors. Some (such as Kelvin-like models) are closer to the elastic, others (such as Maxwell-like models) – closer to viscous. There are fine examples of all such theories, both in 1D (Gracheva and Othmer, 2004; Larripa and Mogilner, 2006), and in 2D (Kim and Sun, 2009; Rubinstein et al., 2009), and even in 3D (Stolarska et al., 2009). As a result of numerical solution of the continuous mechanics equations, all these models predict inward flow (or deformations) of the cytoskeleton generated by the balance between the myosin contraction and passive viscoelastic resistance. At the cell boundaries, this inward flow geometrically adds to the actin network growth to cause net protrusion at the front and retraction at the rear. Because of very complex membrane mechanics, the boundary conditions are often very difficult to implement (Herant et al., 2003). In order to simulate the motile cell, the cell edges have to be moved (in the most complex cases by using level-set methods (Yang et al., 2008)) according to this net protrusion-retraction balance, and the continuous mechanics equation has to be solved on the free boundary domain (Herant et al., 2003; Stolarska et al., 2009). Future models will need to treat the complexity of cells’ mechanical properties using models more complete than viscoelastic rheology. Recently, other theories have been proposed, for example, the poroelastic model (Charras et al., 2008), in which the actin cytoskeleton is assumed to behave like an elastic solid, with a fluid cytoplasm squeezing through the actin mesh. This theory predicts subtle differences from the viscoelastic description, for example slow propagation of pressure and deformations in the cell. A recent experimental test (Rosenbluth et al., 2008) favors the poroelastic over the viscoelastic description. One has to keep in mind, however, that there is no single model that is applicable to all motile phenomena; almost certainly any reasonable model is adequate for at least one process, and frequent

16

Mathematical and Physical Modeling of Actin Dynamics in Motile Cells

405

categorical statements about the novelty and groundbreaking success of one model and the invalidity of others should not be taken seriously. Besides being visco- or poro-elastic in general, the active actin- myosin cytoskeleton is also highly nonlinear (Kang et al., 2009): its material properties are not constant but depend on the applied stresses, and in fact on the history of such stresses. To the best of our knowledge, no model has yet attempted to examine these nonlinearities in the context of cell motility. Last, but not least, perhaps the most formidable future challenge is to account for the well known (Tang et al., 2001) but largely ignored polyelectrolyte nature of the cytoskeleton. The difficulty of the model equations in this case is evident from some early initial attempts (Wolgemuth et al., 2004). Mechanics is hardly the only factor governing the cell shape and movements: spatial-temporal biochemical signaling participates in the maintenance of cell locomotion (Ridley et al., 2003). Mathematically, Maree et al. (2006) demonstrated recently that reactions and diffusion of Rho GTPases and their interactions with the actin/Arp2/3 system characterized by nonlinear feedbacks and rapid diffusion of the inactive forms of the small G-proteins in the cytoplasm can be responsible for robust maintenance of steady rapid motility. This model employed an original approach to the anisotropic actin cytoskeleton by discretizing the cell interior into hexagons representing local actin arrays. Another rule-based model, similar in spirit, assuming interactions of actin with regulating cortical factors (Nishimura et al., 2009), was able to reproduce the keratocyte and amoeboid cell shapes. A significant barrier in the future for modelers will be the coupling of mechanics and regulation models and derivation and simulation of unified mechanochemical theories of the cell movements. So far, we have described continuous approaches to the actin cytoskeleton mechanics. The continuum can be considered as a limit of the discrete system consisting of nodes, springs and dashpots. Such a pioneering computational approach in 1D – modeling the cell as a 1D chain of springs (contractile actin-myosin units) connecting nodes (material points of actin cytoskeleton) – was proposed by DiMilla et al. (1991). The nodes are linked to the substrate through dashpots (adhesion complexes). In these models, at the front and rear of the cell there are two possibilities – either to advance the leading node and pull up the rear node preserving their identity, or to mimic actin polymerization at the front and depolymerization at the rear by adding nascent nodes at the front and deleting nodes at the rear, respectively. These models’ simulations involve solving large systems of ordinary differential equations, which is often simpler than the continuous numerical analysis. One of the great achievements of these models was prediction of biphasic cell speed as function of adhesion strength (DiMilla et al., 1991; Sarvestani and Jabbari, 2009) observed in experiments: when the adhesions are weak, the inward myosin-generated contraction largely cancels the actin protrusion, but when the adhesions are strong, the retraction of the cell rear is hindered, and resulting tension slows the protrusion down. Discrete mechanical models do not necessarily have to simulate a nodesand-springs network. For example, recently the lamellipodium was modeled as

406

A.E. Carlsson and A. Mogilner

overlapping elastic rods of actin filaments crosslinked by sliding crosslinks (Oelz et al., 2009). Another example is the work of Paul et al. (2008), where the lamellipodial network was described as a flat contractile cable (resisting stretching but not generating force when compressed) network adhering to the substrate at the periphery. Simulations of this model showed that both keratocyte- and fibroblastlike shapes can be reproduced by varying only one parameter. In the model, the peripheral adhesions were assumed to break when a critical force was exceeded. Interestingly, when the authors started with an ellipsoidal cell with a broad leading edge and narrow sides, the adhesions at the broad rear edge broke, sharp rear corners emerged, and the cell assumed the characteristic keratocyte-like shape. On the other hand, in an ellipsoidal cell with a narrow leading edge and broad sides, the adhesions at the rear halves of the sides broke, but the adhesions at the very rear survived, and a triangular tailed fibroblast-like shape emerged, with concave sides.

16.6 Discussion After infrequent (but very insightful) early theoretical actin studies, quantitative modeling has exploded in the last 15 years or so. The models reviewed here represent only a small part of the whole picture. These models certainly have become a significant part of mathematical biology and soft-matter physics, but the important question is: have they made an impact on biological research? The answer is a definite “yes”, though to different extents and in various ways. For example, the reviewed models of actin hydrolysis clearly exposed the problems with our current simplistic understanding of this process and proposed some possibilities for updating the existing knowledge. The models of actin force generation and actin network self-organization have become an indispensable part of biological discovery, equal in significance to the experiments, as they allow interpretation and integration of the data. The models of filopodia are good examples of feasibility studies: qualitatively, many biological hypotheses sound reasonable, but physical/mathematical investigation is needed to see whether they are consistent with the laws of nature. Last, but not least, the models of whole cells serve as precursors for future systemslevel efforts to integrate a tremendous amount of quantitative data generated by experiments. The significant successes of modeling of actin dynamics to date do not mean that efforts in this direction can relax. Quite the contrary – with rare exceptions, the existing models are conceptual, rather than detailed and predictive. The future effort has to be three-pronged: first, new conceptual models have to be suggested, but only if they are essentially different from the existing ones, and if there are qualitative observations supporting such models. Second, and more importantly, conceptual models have to be made detailed and predictive by quantifying actual biochemical pathways and physical mechanisms suggested by data and inferring values of model parameters from these data. Third, there are usually a number of alternative explanations for any phenomenon. Therefore experimental data must be used to limit the

16

Mathematical and Physical Modeling of Actin Dynamics in Motile Cells

407

number of possible hypotheses used in modeling, with the caveat that redundant mechanisms are often at work in motile cells. We have some very successful models of simple sub-systems of the actin machinery, but as far as systems-level modeling of the actin dynamics is concerned, it is in its infancy. However, valuable lessons can already be learned from other areas where systems approaches are used. One of these lessons is that it will be important to bridge micro- and macro-scales, so the models will become multi-scale. Some software for numerical analysis of multi-scale models already exists, but normally it requires formidable computer power and consumes a great deal of computer time. Perhaps combining small-scale “caricature” models with detailed large-scale ones and going back-and-forth between different levels to build intuition and ensure correctness will help with this challenge. Addressing this challenge will also require the development of physico-chemical mechanisms for modules of motility, and the addition of layers of complexity of spatial-temporal regulation, as well as analysis of systems-level features of sensitivity, robustness and redundancy of the motile actin networks. Ultimately, the modeling will be complete when not only relatively simple qualitative diagrams, like the dendritic-nucleation model, but also comprehensive mechanochemical “interactomes” involving tens of essential molecular players, are quantified and simulated. Acknowledgements This work was supported by NIH GLUE Grant “Cell Migration Consortium” (NIGMS U54 GM64346), by NSF Grant DMS-0315782 to A.M., and NIH Grant R01 GM086882 to A.E.C.

References Akin O, Mullins RD (2008) Capping protein increases the rate of actin-based motility by promoting filament nucleation by the Arp2/3 complex. Cell 133:841–851 Alberts JB, Odell GM (2004) In silico reconstitution of listeria propulsion exhibits nano-saltation. PLoS Biol 2:2054–2066 Atilgan E, Wirtz D, Sun SX (2005) Morphology of the lamellipodium and organization of actin filaments at the leading edge of crawling cells. Biophys J 89:3589–3602 Atilgan E, Wirtz D, Sun SX (2006) Mechanics and dynamics of actin-driven thin membrane protrusions. Biophys J 90:65–76 Bindschadler M, McGrath JL (2007) Relationships between actin regulatory mechanisms and measurable state variables. Ann Biomed Eng 35:995–1011 Boal D (2002) Mechanics of the Cell. Cambridge University Press, New York Bretschneider T, Anderson K, Ecke M et al. (2009) The three-dimensional dynamics of actin waves, a model of cytoskeletal self-organization. Biophys J 96:2888–2900 Bretschneider T, Diez S, Anderson K et al. (2004) Dynamic actin patterns and Arp2/3 assembly at the substrate-attached surface of motile cells. Curr Biol 14:1–10 Brieher WM, Kueh HY, Ballif BA et al. (2006) Rapid actin monomer-insensitive depolymerization of Listeria actin comet by cofilin, coronin, and Aip1. J Cell Biol 175:315–324 Burroughs NJ, Marenduzzo D (2007a) Three-dimensional dynamic Monte Carlo simulation of elastic actin-like ratchets. J Chem Phys 123:174908-1–174908-11 Burroughs NJ, Marenduzzo D (2007b) Nonequilibrium-driven motion in actin networks: Comet tails and moving beads. Phys Rev Lett 98:238302-1–238302-4

408

A.E. Carlsson and A. Mogilner

Carlsson AE (2000) Force-velocity relation for growing biopolymers. Phys Rev E 62:7082–7091 Carlsson AE (2001) Growth of branched actin networks against obstacles. Biophys J 81:1907–1923 Carlsson AE (2003) Growth velocities of branched actin networks. Biophys J 84:2907–2918 Carlsson AE (2008) Model of reduction of actin polymerization forces by ATP hydrolysis. Phys Biol 5:1–9 Carlsson AE, Sept D (2008) Mathematical modeling of cell migration. Methods Cell Biol 84: 911–937 Chan CE, Odde DJ (2008) Traction dynamics of filopodia on compliant substrates. Science 322(5908):1687–1691 Charras GT, Coughlin M, Mitchison TJ et al. (2008) Life and times of a cellular bleb. Biophys J 94(5):1836–1853 Danuser G (2009) Testing the lamella hypothesis: the next steps on the agenda. J Cell Sci 122(Pt 12):1959–1962 Dayel MJ, Akin O, Landeryou M et al. (2009) In silico reconstitution of actin-based symmetry breaking and motility. PLoS Biol 7(9):e1000201 Dickinson RB, Purich DL (2002) Clamped-filament elongation model for actin-based motors. Biophys J 82:605–617 Dickinson RB, Southwick FS, Purich DL (2002) A direct-transfer polymerization model explains how the multiple profilin-binding sites in the actoclampin motor promote rapid actin-based motility. Arch Biochem Biophys 406:296–301 Dickinson RB, Caro L, Purich DL (2004) Force generation by cytoskeletal filament end-tracking proteins. Biophys J 87:2838–2854 Dickinson RB, Purich DL (2006) Diffusion rate limitations in actin-based propulsion of hard and deformable particles. Biophys J 91(4):1548–1563 Dimilla PA, Barbee K, Lauffen burger DA (1991) Mathematical model for the effects of adhersion and mechanics on cell migration speed. Biophys J 60:15–37 Ditlev JA, Vacanti NM, Novak IL et al. (2009) An open model of actin dendritic nucleation. Biophys J 96(9):3529–3542 Doubrovinski K, Kruse K (2008) Cytoskeletal waves in the absence of molecular motors. EPL 83:18003 Flaherty B, McGarry JP, McHugh PE (2007) Mathematical models of cell motility. Cell Biochem Biophys 49(1):14–28 Footer MJ, Kerssemakers JW, Theriot JA et al. (2007) Direct measurement of force generation by actin filament polymerization using an optical trap. PNAS 104:2181–2186 Fujiwara I, Takahashi S, Tadakuma H et al. (2002) Microscopic analysis of polymerization dynamics with individual actin filaments. Nat Cell Bio 4:666–673 Fujiwara K, Vavylonis D, Pollard TD (2007) Polymerization kinetics of ADP-and ADP-Pi –actin determined by fluorescence microscopy. Proc Natl Acad Sci USA 104:8827–8832 Gerisch G, Bretschneider T, Muller-Taubenberger A et al. (2004) Mobile actin clusters and traveling waves in cells recovering from actin depolymerization. Biophys J 87:3493–3503 Gholami A, Falcke M, Frey E (2008) Velocity oscillations in actin-based motility. New J Phys 10:033022 Giannone G, Dubin-Thaler BJ, Rossier O et al. (2007) Lamellipodial actin mechanically links myosin activity with adhesion-site formation. Cell 128(3):561–575 Gracheva ME, Othmer HG (2004), A continuum model of motility in ameboid cells. Bull Math Biol 66:167–193 Greene GW, Anderson TH, Zeng H et al. (2009) Force amplification response of actin filaments under confined compression. Proc Natl Acad Sci 106:445–449 Gupton SL, Gertler FB (2007) Filopodia: the fingers that do the walking. Sci STKE 2007(400):re5 Herant M, Marganski WA, Dembo M (2003) The mechanics of neutrophils: synthetic modeling of three experiments. Biophys J 84:3389–3413 Hill TL (1986) Theoretical study of a model for the ATP cap at the end of an actin filament. Biophys J 49:981–986

16

Mathematical and Physical Modeling of Actin Dynamics in Motile Cells

409

Hill TL, Kirschner MW (1982) Bioenergetics and kinetics of microtubule and actin filament assembly- disassembly. Int Rev Cytol 78:1–125 Howard J (2001) Mechanics of Motor Proteins and the Cytoskeleton. Sinauer, Sunderland, MA Huber F, Käs J, Stuhrmann B (2008) Growing actin networks form lamellipodium and lamellum by self-assembly. Biophys J 95(12):5508–5523 Ideses Y, Brill-Karniely Y, Haviv L et al. (2008) Arp2/3 branched actin network mediates filopodialike bundles formation in vitro. PLoS ONE 3(9):e3297 Kang H, Wen Q, Janmey PA et al. (2009) Nonlinear elasticity of stiff filament networks: strain stiffening, negative normal stress, and filament alignment in fibrin gels. J Phys Chem B 113(12):3799–3805 Keiser T, Schiller A, Wegner A (1986) Nonlinear increase of elongation rate of actin filaments with actin monomer concentration. Biochemistry 25:4899–4906 Keren K, Pincus Z, Allen GM et al. (2008) Mechanism of shape determination in motile cells. Nature 453(7194):475–480 Kim JS, Sun SX (2009) Continuum modeling of forces in growing viscoelastic cytoskeletal networks. J Theor Biol 256(4):596–606 Koestler SA, Auinger S, Vinzenz M et al. (2008) Differently oriented populations of actin filaments generation in lamellipodia collaborate in pushing and pausing at the cell front. Nat Cell Biol 10(3):306–313 Kraikivski P, Slepchenko BM, Novak IL (2008) Actin bundling: initiation mechanisms and kinetics. Phys Rev Lett 101(12):128102 Kueh HY, Brieher WM, Mitchison TJ (2008a) Dynamic stabilization of actin filaments. Proc Natl Acad Sci 105:16531–16536 Kueh HY, Charras GT, Mitchison TJ et al. (2008b) Actin disassembly by cofilin, coronin, and Aip1 occurs in bursts and is inhibited by barbed-end cappers. JCB 182:341–353 Kuhn JR, Pollard TD (2005) Real-time measurements of actin filament polymerization by total internal reflection fluorescence microscopy. Biophys J 88:1387–1402 Lämmermann T, Sixt M (2009) Mechanical modes of “amoeboid” cell migration. Curr Opin Cell Biol. 21, 636–644 Lan Y, Papoian GA (2008) The Stochastic dynamics of filopodial growth, Biophys J 94:3839–3852 Larripa K, Mogilner A (2006). Transport of a 1D viscoelastic actin-myosin strip of gel as a model of a crawling cell. Physica A – Stat Mech Appl 372: 113–123 Lee KC, Liu AJ (2008) New proposed mechanism of actin-polymerization-driven motility. Biophys J 95:4529–4539 Lee KC, Liu AJ (2009) Force-velocity relation for actin-polymerization-driving motility from Brownian dynamics simulations. Biophys J 97:1295–1304 Li X, Kierfeld J, Lipowsky R (2009) Actin polymerization and depolymerization coupled to cooperative hydrolysis. Phys Rev Lett 103:048102 Liu AP, Richmond DL, Maibaum L et al. (2008) Membrane-induced bundling of actin filaments. Nat Phys 4:789–793 Machacek M, Hodgson L, Welch C et al. (2009) Coordination of Rho GTPase activities during cell protrusion. Nature 461(7260):99–103 Maly IV, Borisy GG (2001) Self-organization of a propulsive actin network as an evolutionary process. PNAS 98:11324–11329 Marcy Y, Prost J, Carlier MF et al. (2004) Forces generated during actin-based propulsion: A direct measurement by micromanipulation. Proc Natl Acad Sci USA 101:5992–5997 Marée AF, Jilkine A, Dawes A et al. (2006) Polarization and movement of keratocytes: a multiscale modeling approach. Bull Math Biol 68(5):1169–1211 Matzavinos A, Othmer HG (2007) A stochastic analysis of actin polymerization in the presence of twinfilin and gelsolin. J Theor Biol 249(4):723–736 Mogilner A (2006). On the edge: modeling protrusion. Curr Op Cell Biol 18:32–39 Mogilner A (2009) Mathematics of cell motility: Have we got its number? J Math Biol 58: 105–134

410

A.E. Carlsson and A. Mogilner

Mogilner A, Edelstein-Keshet L (2002) Regulation of actin dynamics in rapidly moving cells: a quantitative analysis. Biophys J 83(3):1237–1258 Mogilner A, Oster GF (2003) Force generation by actin polymerization II: The elastic ratchet and tethered filaments. Biophys J 84:1591–1605 Mogilner A, Oster GF (1996) The physics of lamellipodial protrusion. Eur Biophys J 25:47–53 Mogilner A, Rubinstein B (2005) The physics of filopodial protrusion. Biophys J 89(2): 782–795 Mullins RD, Heuser JA, Pollard TD (1998) The interaction of Arp2/3 complex with actin: Nucleation, high affinity pointed end capping, and formation of branching networks of filaments. Proc Natl Acad Sci USA 95:6181–6186 Nishimura SI, Ueda M, Sasai M (2009) Cortical factor feedback model for cellular locomotion and cytofission. PLoS Comput Biol 5(3):e1000310 Oelz D, Schmeiser C, Small JV (2008) Modeling of the actin-cytoskeleton in symmetric lamellipodial fragments. Cell Adh Migr 2(2):117–126 Oosawa F, Asakura S (1962) A theory of linear and helical aggregations of macromolecules. J Mol Biol 4:10–21 Parekh SH, Chaudhuri O, Theriot JA et al. (2005) Loading history determines the velocity of actin-network growth. Nat Cell Biol 7: 1219–1223 Paul R, Heil P, Spatz JP et al. (2008) Propagation of mechanical stress through the actin cytoskeleton toward focal adhesions: model and experiment. Biophys J 94(4):1470–1482 Peskin CS, Odell GM, Oster GF (1993) Cellular motions and thermal fluctuations: The Brownian ratchet. Biophys J 65:316–324 Pollard TD (1986) Rate constants for the reactions of ATP- and ADP-actin with the ends of actin filaments. J Cell Biol 103:2747–2754 Pollard TD (2003) The Cytoskeleton, Cellular motility, and the reductionist agenda. Nature 422:741–745 Pronk S, Geissler PL, Fletcher DA (2008) Limits of filopodium stability. Phys Rev Lett 100(25):258102 Rafelski SM, Alberts JB, Odell GM (2009) An experimental and computational study of the effect of acta polarity of the speed of listeria monocytogens actin-based motility. PLoS Comput Bio 5:1–14 Ridley AJ, Schwartz MA, Burridge K et al. (2003) Cell migration: integrating signals from front to back. Science 302(5651):1704–1709 Rosenbluth MJ, Crow A, Shaevitz JW et al. (2008) Slow stress propagation in adherent cells. Biophys J 95(12):6052–6059 Rubinstein B, Fournier MF, Jacobson K et al. (2009) Actin-myosin viscoelastic flow in the keratocyte lamellipod. Biophys. J. 97, 1853–1863 Rubinstein B, Jacobson K, Mogilner A (2005) Multiscale two-dimensional modeling of a motile simple-shaped cell. Multiscale Model Simul 3(2):413–439 Sambeth R, Baumgaertner A (2001) Autocatalytic polymerization generates persistent random walk of crawling cells. Phys Rev Lett 86:5196–5199 Sarvestani AS, Jabbari E (2009) Analysis of cell locomotion on ligand gradient substrates. Biotechnol Bioeng 103(2):424–429 Satulovksy J, Lui R, Wang YL (2008) Exploring the control circuit of cell migratin by mathematical modeling. Biophys J 94(9):3671–3683 Schaus TE, Borisy GG (2008) Performance of a population of independent filaments in lamellipodial protrusion. Biophys J 85:1393–1411 Schaus TE, Taylor EW, Borisy, GG (2007). Self-organization of actin filament orientation in the dendritic-nucleation/array-treadmilling model. Proc. Natl. Acad. Sciences 104, 7086–7091 Shaevitz JW, Fletcher DA (2007) Load fluctuations drive actin network growth. PNAS 104: 15688–15693 Shemesh T, Otomo T, Rosen MK, Bershadsky AD, Kozlov MM (2005) A Novel mechanism of actin filament processive capping by formins solution of the rotation pardox. J Cell Biol 170:889–893

16

Mathematical and Physical Modeling of Actin Dynamics in Motile Cells

411

Shemesh T, Kozlov MM (2007) Actin polymerization upon processive capping by formin: a model for slowing and acceleration. Biophys J 92:1512–1521 Shemesh T, Verkhovsky AB, Svitkina TM et al. (2009) Role of focal adhesions and mechanical stresses in the formation and progression of the lamellum interface. Biophys J 97(5):1254–1264 Shlomovitz R, Gov NS (2007) Membrane waves driven by actin and myosin. Phys Rev Lett 98(16):168103 Slepchenko BM, Schaff JC, Macara IG et al. (2003) Quantitative cell biology with the virtual cell. Trends Cell Biol 13:570–576 Small V and Celis JE (1978) Filament arrangements in negatively stained cultured cells. Eur J Cell Biol 16:308–325 Small V, Herzog M, and Anderson K (1995) Actin filament organization in the fish keratocyte lamellipodium. J Cell Biol 129:1275–1286 Steketee M, Balazovich K, Tosney KW (2001) Filopodial initiation and a novel filament-organizing center, the focal ring. Mol Biol Cell 12(8):2378–2395 Stéphanou A, Mylona E, Chaplain M et al. (2008) A computational model of cell migration coupling the growth of focal adhesions with oscillatory cell protrusions. J Theor Biol 253(4):701–716 Stolarska MA, Kim Y, Othmer HG (2009) Multi-scale models of cell and tissue dynamics. Philos Transact A Math Phys Eng Sci 367(1902):3525–3553 Stukalin EB, Kolomeisky AB (2006) ATP hydrolysis stimulates large length fluctuations in single actin filaments. Biophys J 90:2673–2685 Svitkina TM, Bulanova EA, Chaga OY et al. (2003) Mechanism of filopodia initiation by reorganization of a dendritic network. J Cell Biol 160(3)409–421 Tang JX, Käs JA, Shah JV et al. (2001) Counterion-induced actin ring formation. Eur Biophys J 30(7):477–484 Vallotton P, Small JV (2009) Shifting view on the leading role of the lamellopodium in cell migration: speckle tracking revisited. J Cell Sci 122(Pt 12):1955–1958 van der Gucht J, Paluch E, Plastino J et al. (2005) Stress release drives symmetry breaking for actin-based movement. Proc Natl Acad Sci USA 102:7847–7852 van Oudenaarden A, Theriot JA (1999). Cooperative symmetry-breaking by actin polymerization in a model for cell motility. Nat Cell Biol 1, 493–499. Vavylonis D, Yang Q, O’Shaughnessy B (2005) Actin polymerization kinetics, cap structure, and fluctuations. PNAS 102:8543–8548 Vavylonis D, Kovar DR, O’Shaughnessy B et al. (2006) Model of formin-associated actin filament elongation. Mol Cell 21:455–466 Veksler A, Gov NS (2007) Phase transitions of the coupled membrane-cytoskeleton modify cellular shape. Biophys J 93(11):3798–3810 Vicker MG (2002) F-actin assembly in Dicytostelium cell locomotion and shape oscillations propagates as a self-organized reaction-diffusion wave. FEBS Lett 510:5–9 Wegner A (1976) Head to tail polymerization of actin. J Mol Biol 108:139–150 Weiner OD, Marganski WA, Wu LF et al. (2007) An actin-based wave generator organizes cell motility. PLoS Bio 5:2053–2063 Whitelam S, Bretschneider T, Burroughs NJ (2009) Tranformation from spots to waves in a model of actin pattern formation. Phys Rev Lett 102:198103 Wolgemuth CW, Mogilner A, Oster GF (2004) The hydration dynamics of polyelectrolyte gels with applications to cell motility and drug delivery. Eur Biophys J 33(2):146–158 Wolgemuth CW (2005) Lamellspodral contractions during crawling and spreading. Biophys J 89:1643–1649 Yang L, Effler JC, Kutscher BL et al. (2008) Modeling cellular deformations using the level set formalism. BMC Syst Biol 2:68 Yang HC, Pon LA (2002) Actin cable dynamics in budding yeast. Proc Natl Acad Sci USA 99: 751–756 Yang L, Sept D, Carlsson AE (2006) Energetics and dynamics of constrained actin filament bundling. Biophys J 90(12):4295–4304

412

A.E. Carlsson and A. Mogilner

Yarmola ET, Dranishnikov DA, Bubb MR (2008) Effect of profilin on actin critical concentration: a theoretical analysis. Biophys J 96: 1232–1233 Zhu J, Carlsson AE (2006) Growth of attached actin filaments. Eur Phys J 21:209–222 Zhuravlev PI, Papoian GA (2009) Molecular noise of capping protein binding induces macroscopic instability in filopodial dynamics. Proc Natl Acad Sci USA 106(28):11570–11575

Chapter 17

Force Production by Actin Assembly: Simplified Experimental Systems for a Thorough Modeling C. Sykes, J. Prost, and J.F. Joanny

Abstract Cellular biopolomers have the specific property of transforming chemical energy into mechanical work. This specificity is the basis of cellular movements in general, from cell division to cell motility. We describe here how actin assembly, in the absence or in the presence of molecular motors, can generate forces that eventually lead to cell movements. We present the state-of-the-art of single filament and many filament force generation. We show that in all cases, generic models can be built that are independent of the detailed molecular mechanism, in agreement with experimental results and quantifications based on stripped-down systems designed using purified proteins or cell extracts. Molecular features play a role only in the details of the mechanics.

Contents 17.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . 17.2 Force Production by a Single Actin Filament . . . . . . . . 17.2.1 Experiments . . . . . . . . . . . . . . . . . . . 17.2.2 Theory . . . . . . . . . . . . . . . . . . . . . 17.2.3 Actin Polymerization in Stereocilia . . . . . . . . . 17.3 Non-contractile Polymerizing Actin Gels and Force Production 17.4 Contractile Gels . . . . . . . . . . . . . . . . . . . . . 17.4.1 Active Gel Theory . . . . . . . . . . . . . . . . 17.4.2 Experimental Systems . . . . . . . . . . . . . . . 17.4.3 Lamellipodium Motion of Keratocytes . . . . . . . 17.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

. . . . . . . . . . . .

414 414 414 416 419 421 423 423 425 427 429 429

C. Sykes (B) Physicochimie Curie (CNRS-UMR168), Institut Curie, Section de Recherche, 26 rue d’Ulm, 75248 Paris Cedex 05, France e-mail: [email protected]

M.-F. Carlier (ed.), Actin-based Motility, DOI 10.1007/978-90-481-9301-1_17, C Springer Science+Business Media B.V. 2010

413

414

C. Sykes et al.

17.1 Introduction Cell movement is based on the growth and assembly of biopolymers. Actin is one of the major components of cells which dynamic assembly is necessary for the three important steps of cell motility (Abercombie 1980): (i) lamellipodium protrusion, due to growing actin filaments organized in a dense branched structure underneath the plasma membrane and pushing it forward, (ii) formation of adhesion patches that link the surface or the extracellular matrix to the cell and especially to the actin cytoskeleton, which reacts in return by strengthening the link, (iii) contractility of the acto-myosin cortex at the back of the cell to push the cell forward. Whereas the role of actin at each step is qualitatively known, it is not easy to have a quantitative characterization of the force necessary to create movement, or of the dependence of the whole cell velocity on actin dynamics and actin mechanical properties. In cells, it is difficult to break the system into its elementary parts and to pin down quantitatively the origin of forces and stress distribution. Therefore, simplified in vitro systems are great tools to decompose cell movements into stripped-down modules. In this review, we use the expressions “force production”, “polymerization force” a.s.o in accordance with the leading terminology used in the field. One should however keep in mind that at low Reynolds numbers momentum conservation implies force conservation in the absence of external fields such as gravity or electric and magnetic field gradients. In practice, this means that force can only come from optical and magnetic traps, or from boundaries for instance from friction on a substrate, confinement between two walls etc. The polymerization process in itself does not generate forces, but does generate force dipoles the strength of which in general depends on boundary conditions potentially far from the polymerization region. This is a general feature common to all processes involving internal forces only (Julicher and Prost 2009). We discuss first here force production by single actin filaments. We then introduce the notion of actin gels and mechanics and show how the mechanism of force production by gel growth is generic, and independent of the molecular details of force production. Ultimately, we develop a description of contractile gels, and discuss how their contractility is evidenced in cells, and in in vitro systems.

17.2 Force Production by a Single Actin Filament 17.2.1 Experiments Whereas in vitro, actin filaments polymerize naturally in the presence of adenosine triphosphate (ATP) and salt, in cells, actin polymerization is enhanced or limited by numerous actin-binding proteins. For example, actin polymerization activation can be classified in two main groups: nucleation-promoting factors of actin filament growth can be either from the formin family and produce parallel bundles of actin

17

Force Production by Actin Assembly

415

filaments, seen for example in filopodia, or from the WASP (Wiskot Aldrich Syndrom Protein) family and activate the Arp2/3 complex to generate a dense, branched actin network such as the one of lamellipodia of cells (Alberts et al. 2002). Actin filament growth therefore occurs either spontaneously, or is enhanced in the presence of a nucleation-promoting factor. The first experiment that directly shows that a single growing actin filament is able to generate a force is due to Kovar and Pollard (2004): formin proteins are grafted on a surface together with dead head myosin motors. Dead head myosin motors serve as fixed attachment points for actin filaments that are polymerizing from a formin. Since the two points of attachment are fixed and the filament is growing from the formin site, the filament buckles in between the two attachment points. Buckling depends on the filament length and rigidity. The force F necessary to buckle a filament of length L and radius a, with an elastic modulus E can be estimated simply through dimensional arguments by considering that this force F is proportional to the persistence length of order Ea4 . It must also decrease as a power law of L. Dimensional analysis imposes then that F ∼ Ea4 /L2 . As a consequence, the length of the shortest filament that buckles reveals the maximum observed force produced by actin polymerization, which is 0.25–0.56 pN for filaments of 0.8–1.2 μm long. A thorough shape analysis of the buckled filaments allows to extract characteristic numbers such as attachment spring stiffness and the details of filament attachment to formin proteins (Berro et al. 2007). The force exerted on the filament is expected to transiently slow down the rate of assembly, but this effect could not be observed in this set of experiments due to technical limitations since the buckling force is too small compared to the stall force. Note that indeed the measured force is smaller than the theoretical maximal force, which as shown in the next section, reads (Hill 1981): F=

kB T log(c/ccrit ) δ

(17.1)

where kB is Boltzmann’s constant, T is temperature, δ the elongation distance by addition of a single protein subunit (2.7 nm for actin), c the concentration of actin monomers in solution, and ccrit is the critical concentration for polymerization (Fig. 17.1).

Fig. 17.1 Diagram of filament attachment and assembly on slides coated with NEM-myosin II (1), formin or GST-formin (2), and NEM-myosin II and formin. (From Kovar and Pollard (2004))

416

C. Sykes et al.

Another set of experiments addressed the issue of force generation during actin filament growth in the absence of any activator of actin polymerization using an experimental setup inspired from studies of the force generated by microtubule growth (Footer et al. 2007): actin filaments growing from an acrosome are put into contact with a wall. In this case, barbed ends of actin filaments are free, contrarily to the formin experiments where barbed ends are attached to the formins. The advantage of the acrosome is that it provides a rigid polarized bundle of actin filaments terminated by an average number of eight filaments, forming the growing end of the acrosome. A bead is attached to the acrosome and trapped in optical tweezers. In order to concentrate actin growth at the end of the acrosome and not in the bulk of the experiment, profilin is added in the mixture. At the micromolar concentration of actin, one would expect from Eq. 17.1 a maximal force of about 0.5–1 pN. Force measurement of actin growth from the acrosome was done with the use of specific optical tweezers, a “keyhole trap”. It consists in trapping both the bead and keeping the acrosome oriented perpendicularly to the wall with a delocalized optical trap. The acrosome is thus facing a wall against which actin polymerization is pushing, like a battering ram on a medieval castle gate. In the presence of actin monomers, actin filament elongation results in displacement of the bead within the optical trap. In these conditions, actin filaments elongate over a several nanometer distance to reach a stable force plateau corresponding to the growth stall. Further elongation of the filaments is made possible by displacing the wall away from the acrosome. This plateau corresponds to a load force of about 1 pN, which is the force expected to stall the growth of a single filament. Since the growing end of the acrosome is made of an average of eight filaments, one could expect that the total acrosome stall force should be the sum of the eight stall force on each filament. Hence the data suggest that either a single filament abutted the wall, or that among many filaments, a single one was stalled while others might buckle or depolymerize.

17.2.2 Theory The simplest model to describe the polymerization of an actin filament under an external force f considers polymerization and depolymerization at the barbed end of the filament with respective rates kp and kd which depend on force (Howard 2001; Oosawa and Akasura 1975). The probability distribution p(n, t) that the polymer comprises n actin monomers at time t satisfies the master equation ∂p(n, t) = kp p(n − 1) + kd p(n) − (kp + kd )p(n). ∂t

(17.2)

After summation of this equation over all filament lengths, one finds that the average number of monomers n¯ in the filament is such that ddtn¯ = kp − kd . Within the framework of Kramers rate theory, both the polymerization and depolymerization rates vary exponentially with the external force (Hill 1981). Experiments on microtubules

17

Force Production by Actin Assembly

417

suggest that the best fit is obtained by assuming that the depolymerization rate is independent of the force (Dogterom and Akasura 1975). We make this simplifying assumption and therefore assume that kd is a constant and that kp = kp0 exp −f /f0 where the prefactor kp0 is proportional to the local concentration c of actin monomers and the characteristic force f 0 is f0 = kB T/δ. In this simple model, the stall force fs1 defined as the force where the average number of monomers is constant (dn¯ /dt = 0) is fs1 = f0 log kp0 /kd .

(17.3)

It thus increases logarithmically with the actin monomer concentration. This simplified model seems to describe well the experimental results on microtubules (Dogterom and Akasura 1975). Models providing more details on how monomers can be added to the filament comfort the general structure of the simplified model and provide numbers for the coefficients (Mogilner and Oster 2003). This approach can be generalized in two ways, by considering several actin filaments in parallel (SandervanDoorn et al. 2000) and by taking into account the hydrolysis of the actin monomers after incorporation in the actin filament (Hill 1986; Ranjith et al. 2009; Stukalin and Kolomeisky 2006). When two actin filaments are polymerizing under force, only the leading filament carries the force and has a slower polymerization rate. A simple model for the growth of filaments against a force f is to suppose that the two filaments are shifted by half a monomer, and polymerize and depolymerize independently (Ranjith, 2009, Unpublished). The polymerization rate of the leading filament which carries the force is kp = kp0 exp −f /f0 and the polymerization rate of the lagging filament which does nos carry any force is kp0 . However, if the lagging filament is only half a monomer behind the leading filament, the two filaments share the force and a naive guess is that each filament feels a force f/2. The polymerization rate is then k1 = kp0 exp −f /(2f0 ). As in the one-filament model, we assume that for both filaments, depolymerization occurs at a force-independent rate kd . This model can be solved exactly by considering the master equation which generalizes Eq. 17.2 for the probability p(n, m, t) that at time t the longest filament has n monomers and that the difference between the two filaments is m monomers. In the limit of long times, the two variables decouple and the probability q(m) of the variable m becomes stationary. In particular one finds that q(1) = (kp0 − kp )/(kp0 + kd ). The average growth rate of the actin filament is then k = kp + kp0 /2 − kd + q(1) k1 − kp0 /2.

(17.4)

where the last term takes into account the fact that if m =1 the polymerization rate is k1 and not kp0 . One can check explicitly that the stall force where the growth rate vanishes is fs2 = 2fs1 . Note that this result is strongly related to the assumption that when m =1, the two filaments equally share the force.

418

C. Sykes et al.

This model neglects the lateral interactions between filaments. It is generally accepted that the two proto-filaments of an actin filament are never distant by more than half a monomer and that due to these lateral interactions, one always have m = 1. For an actin bundle containing N actin filaments, the model has been generalized in (SandervanDoorn et al. 2000). The growth rate of the bundle could not be calculated explicitly but the stall force is shown to be proportional to the number of filaments fsN = N fs1 . Another important feature of actin polymerization is that the actin monomers polymerize in an ATP form but are then hydrolyzed when incorporated into the filament. The hydrolysis occurs in two steps, first the hydrolysis reaction itself that transforms an actin-ATP molecule into an actin-adenosine diphosphate (ADP) molecule and a phosphate and then the release of the phosphate. This second step is the rate limiting step. However, we do not distinguish the two steps here and consider that there is an effective hydrolysis rate r. There are two extreme models for hydrolysis. In the random hydrolysis model (Ohm and Wegner 1991; Vavylonis et al. 2005), the actin monomers are incorporated in an ATP form and then hydrolyze randomly independently of their environment. In a sequential model (Ranjith et al. 2009; Stukalin and Kolomeisky 2006; Pantaloni et al. 1985), there is an ATP cap at the tip of the growing filament and the rest of the filament is made of ADP actin. Hydrolysis can only occur at the interface between these two regions. An intermediate model where the hydrolysis is not fully cooperative has been introduced in (Li et al. 2009). The sequential hydrolysis model has been studied in details in (Ranjith et al. 2009; Stukalin and Kolomeisky 2006). Actin polymerizes in an ATP form with a rate kp and depolymerizes from the barbed end with a rate kdT for actin-ATP and kdD for actin-ADP. The depolymerization rate is larger for actin-ADP kdD > kdT . There are three dynamical growth regimes for the actin filament. If the polymerization rate is large, the filament grows on average at constant rate, the actin-ATP cap at the end of the filament is large and its size increases linearly with time. The monomer at the tip of the filament is almost always an actin-ATP monomer and the growth rate of the filament is k = kp − kdT ; the growth rate of the actin-ATP cap is kcap = kp −(kdT +r). When the external force increases, the growth rate of the actin-ATP cap decreases and vanishes at the transition to the intermediate regime if kp = (kdT + r) In the intermediate regime, the ATP cap is finite and the monomer at the tip has a probability q = kp /(kdT + r) to be an actin-ATP monomer. The growth rate of the filament is k = kp − qkdT − (1 − q) kdD .

(17.5)

When the force is further increased, the growth rate vanishes at the stall force where both the filament and the actin-ATP cap become finite. The stall force is then fsh = f0 log

kp0 kdD + r kdD + kdT + r

.

(17.6)

17

Force Production by Actin Assembly

419

The stall force decreases with the hydrolysis rate and still increases logarithmically with the concentration of actin monomers. The alternative model is the random hydrolysis model. There are only two dynamical polymerization regimes in this model as the number of actin-ATP monomers always remains finite. The transition between the growing regime and the finite filament regime occurs at the stall force. There is no simple analytic expression of the stall force in the random hydrolysis model but the effect of hydrolysis is similar to what is predicted in the sequential model and leads to a reduction of the stall force. In fact it is argued in (Ranjith et al. 2010) that measurements of the growth rate only hardly allow to distinguish between the two models.

17.2.3 Actin Polymerization in Stereocilia The detection of sound in the cochlea of many animals is made by hair cells. Each of these cells possesses a bundle of cilia. One cilium, the kinocilium, has a classical microtubule structure, but all the other cilia, called stereocilia, are actin structures made of parallel actin filaments organized in a crystal like structure (Tilney et al. 1992; Holton and Hudspeth 1983). The actin filaments are parallel and are tightly cross-linked by a protein called espin. The stereocilia are surrounded by a phospholipid membrane which is continuous with the apical membrane of the hair cell. In the bundle, the stereocilia are organized in rows of increasing sizes. Each stereocilium in a row is connected to the neighboring stereocilium in the next taller row by a tip-link. The tip link is connected to an ion channel which opens when it is put under tension by the relative motion of the two connected cilia, allowing for the transformation of the mechanical signal into a chemical signal. Along the cochlea, the bundles of the haircells have growing sizes and each size codes for the frequency that this cell detects. The typical size of a stereocilium is in the range 10 – 100 μm. A very precise regulation of the size of the stereocilia is therefore needed to achieve the extreme frequency sensitivity of the ear. Recently the group of Kachar (Rzadzinska et al. 2004; Schneider et al. 2002) has obtained two spectacular results concerning actin in stereocilia by labeling the actin filaments. Actin is treadmilling in the stereocilia and in each stereocilium, the whole actin fascicle is moving down to the apical surface of the cell very slowly at a velocity vT . Moreover, the treadmilling velocity is proportional to the length of the stereocilium. In turn, this means that the treadmilling time corresponding to the complete turnover of an actin filament is the same for all stereocilia and it is found on the order of 48 h for mice or rat stereocilia. At the tip of the stereocilia, the actin filaments interact with the membrane via the so-called tip complex which is a protein complex including in particular the molecular motor myosin XV. The membrane is under tension and exerts a force on the actin filaments which slows down their polymerization. The polymerizing actin filaments exert a finite pressure Peff on the membrane as shown in

420

C. Sykes et al.

Fig. 17.2 Sketch of the structure of stereocilia (From Rzadzinska et al. J.Cell.Biol.2004)

(Prost et al. 2007). The force balance on the membrane at the tip of the stereocilium relates the polymerizing pressure Peff to the membrane tension σ via Laplace’s law Peff

2σ r0

(17.7)

where r0 is the radius of the cilium. The treadmilling velocity is then smaller than the polymerizing velocity of actin in the absence of tension vT = vp − 2λa σ/r0

(17.8)

where λa is a phenomenological positive mobility (Fig. 17.2). The presence of capping proteins preventing depolymerization at the pointed ends of the actin filaments could explain the independence of the treadmilling time with the length of the stereocilium (Prost et al. 2007). Crosslinking proteins such as espin play a similar role as they also stop depolymerization. In the limit where they bind only at the tip of the filament and unbind much slower than depolymerization, their effective role is equivalent to that of a capping protein. The treadmilling time is then inversely proportional to the uncapping rate ku of the capping protein and the length of the filament is L ∼ vT ku−1 . The shape of the stereocilia can then be calculated from the treadmilling dynamics of the individual filaments inside the stereocilium. The obtained shape is less elongated than the experimental shape of the native stereocilia; it is more similar to that of mutant stereocilia where myosin

17

Force Production by Actin Assembly

421

XV has been inhibited (Rzadzinska et al. 2004). An alternative model that fits better the experimental results is that the role of the capping protein is played by the crosslinking espin protein (Lenz, 2009, Unpublished). This model leads to a more elongated shape of the stereocilia that can be fitted to the observed shape. Note that many other proteins are involved in stereocilia (Salles et al. 2009). An alternative model for the shape of stereocilia, assumes that one of these proteins diffuses from the cell body and provoques the filament depolymerization (Naoz et al. 2008).

17.3 Non-contractile Polymerizing Actin Gels and Force Production One experimental approach used for addressing the question of how the assembly of actin monomers at a load surface can generate a force has been to use stripped-down systems, simpler than a cell, which reproduce movement via polymerization. The first system broadly studied has been the bacteria Listeria Monocytogenes, which moves at a velocity of the order of several μm/min within cells by polymerizing actin at its surface, actin filaments being oriented with their barbed ends mainly in the direction of movement. Endocytic vesicles display the same type of motility. No myosin is required for this movement, and the mechanism of actin polymerization is the same as for cells except that the recruitment of Arp2/3 is performed by a bacterial transmembrane protein, called ActA (Theriot et al. 1992). Such bacteria move within cells (Welch et al. 1998), by generating a comet tail made of actin filaments crosslinked by proteins from the cytosol, conferring to the comet the elastic properties of a polymer network characterized by an elastic modulus in the range of 103 –104 Pa (Gerbal et al. 2000; Marcy et al. 2004). Listeria movement has been extensively studied by using cell extracts instead of cells, which allows for adjustment of the motility mixture and led to the identification of the essential components of actin-based motility (Loisel et al. 1999). The logical extension of work on Listeria is the replacement of the bacterium with non biological cargos, since the only bacterial contribution to comet formation is the ActA protein on the surface. Indeed, spherical latex beads or oil droplets coated with variants of ActA or WASP develop comets and move once placed in cell extracts or pure protein mixes (Cameron et al. 1999; Boukellal et al. 2004; Bernheim-Groswasser et al. 2002). In all cases, there is no need for sophisticated protein grafting, since simple adsorption of the protein or protein fragment on the bead does not compromise protein activity. Comets form either from an initial assymmetry in coating of the bead with the actin nucleation promoting factor, or through spontaneous symmetry breaking (Van der Gucht et al. 2005). Actin gels have elastic properties that depend on the biochemical conditions under which the actin network is assembled (Gardel et al. 2004). These elastic properties are believed to play an important role in cell motility. At the mesocopic scale (larger than a molecule but smaller than the load), the elastic properties of the actin

422

C. Sykes et al.

Fig. 17.3 Photobleaching shows deformations in the gel during comet growth. (a, b) Lines bleached perpendicular to the comet axis are deformed when the lines are less than a bead-diameter away from the bead surface. Lines further away are not deformed (a). Protein concentrations: 8.1 mM F-actin (10% labeled with Alexa Fluor 488), 0.1 mM Arp2/3, 3 mMADF/cofilin, 1 mM profilin, and 0.3 mM gelsolin. Scale bar is 5 micrometers in all figures. (From Paluch et al. (2006))

network, synthesized by monomer addition at the surface of the bead or Listeria, generate stress (Noireaux et al. 2000; Gerbal et al. 2000). The relaxation of the stress results in propulsion as the actin gel squeezes the bead forward. As a consequence, comets undergo visible deformations as shown in Fig. 17.3 (Paluch et al. 2006). The use of deformable loads provides a tangible demonstration of stress buildup and squeezing. Liposomes (Upadhyaya et al. 2003; Giardini et al. 2003) or oil droplets (Boukellal et al. 2004) are deformed into a pear-like shape when propelled by an actin comet. The actin gel compresses at the sides of the object and pulls at the rear. The analysis of the contour of the deformed objects provides a qualitative distribution of the normal stresses along the load surface. In the case of oil droplets, where the interfacial tension is measurable, a precise value can be calculated for the normal stress (Boukellal et al. 2004). Hard objects present the advantage of allowing for experiments in which external force dipoles are exerted on the growing comet. One can thus perform quantitative measurements of the force-velocity relation in well defined geometries (Marcy et al. 2004; Parekh et al. 2005). The force that actin polymerization exerts on a hard bead is in the nanonewton range, as directly measured by micromanipulation experiments consisting of attaching the bead to a flexible microfiber that acts as a force sensor (Marcy et al. 2004). This force is many orders of magnitude higher than the viscous drag that opposes moving objects in solution, explaining why moving Listeria or beads are hardly sensitive to viscosity enhancers (Wiesner et al. 2003). Indeed, the stall force for a 2 μm size propelled bead is estimated at 7 nN from a force-velocity curve obtained for pulling to pushing forces in the (–1.7) to 4.3 nN range, fitted with an elastic and friction force (Marcy et al. 2004). This is in keeping with the nN range forces measured in cells by using substrates acting as force sensors (Tan et al. 2003). Using an AFM cantilever geometry, Parekh et al. [45] find a velocity independent of stress in the

17

Force Production by Actin Assembly

423

regime investigated in (Marcy et al. 2004). It is thus often concluded that the two experiments are incompatible. This is however not the case. The theoretical interpretation of (Marcy et al. 2004), involves a polymerization rate independent of stress locally as observed in (Parekh et al. 2005). This results in a non trivial force-velocity relation after a complete mesoscopic analysis. The same analysis allows to extract the friction between the comet and the bead. It is the so-called protein friction due to the attachment-detachment cycles of proteins that bind the surface of the bead to the actin gel. This leads to dissipation of the stored elastic energy. The interplay between the squeezing force and the internal friction force may result in a stick-slip phenomenon depending on the relative value of each effect as observed with Listeria mutants (Lasa et al. 1995; Kocks et al. 1992; Delatour et al. 2008) and expected theoretically (Gerbal et al. 2000). Indeed, cycling between high and low velocities with a spatial periodicity on the order of the size of the object is observed experimentally (Bernheim-Groswasser et al. 2002), and can be understood as follows: after a fast moving period, the elastic energy initially stored in the actin gel is almost entirely relaxed, the elastic propulsive force is close to zero, and the velocity is minimal. During that period however, the polymerization is at its fastest since there is little normal stress to impede it. The actin gel accumulates and the stress grows until a critical value is reached, where a large proportion of the connections between the bead and the gel break. This causes a sharp velocity increase to the maximum value. As soon as the stress is relaxed, the velocity drops sharply back to its minimum value. This mechanism leads to a non-monotonic friction force and to the so-called stick-slip phenomenon. Note that the same apparent behavior occurs with soft droplets or liposomes propelled by actin polymerization, but rely on a different mechanism based on stress release and diffusion of the nucleating activating factors on the fluid surface (Trichet et al. 2007). Bead motility can also be achieved by activating actin polymerization through formins (Berro et al. 2007; Romero et al. 2004). In this case, the actin comet is not a gel but rather a bundle of a few actin filaments. The mechanics is then different and based on filament bundles pushing the wall (Romero et al. 2004). Buckling of the growing filaments and bundles attached at the end of the coverslip can occur (Berro et al. 2007).

17.4 Contractile Gels 17.4.1 Active Gel Theory Many actin structures in cells are made of actin filaments interacting with myosin II molecular motors; this is the case for example in the cytoskeleton or in the actin cortical layer. The properties of the actin gel are then completely different from the passive gels described above. The activity of molecular motors provides them with contractility, a unique feature of active systems. A completely new theoretical framework is thus needed to describe the properties of acto-myosin gels. They are intrinsically non-equilibrium systems since they permanently consume energy in the

424

C. Sykes et al.

form of ATP (Kruse et al. 2005, 2004; Julicher et al. 2007). ATP is consumed both by actin polymerization and depolymerization and by molecular motors. A natural approach to the description of active actin gels is a microscopic theory which has to combine the properties of molecular motors often described at a mesosopic scale and the properties of the actin gel obtained from polymer physics. This has been done to some extent but it often leads to complex theories requiring many simplifying assumptions (Kruse and Julicher 2000; Doubrovinski and Kruse 2007; Liverpool et al. 2009; Aranson and Tsimring 2006). One can also use a more macroscopic hydrodynamic description (Julicher et al. 2007). The hydrodynamic theory only considers the properties of the gel at long length scales and over long time scales. It is based on a few constitutive equations which relate fluxes, such as the stress or the rate of change of the actin polarization, to generalized thermodynamic forces, such as the velocity gradient in the gel or the orientational field which acts locally on the filaments. Once the fluxes and forces are identified, the structure of the constitutive equations is obtained by using only symmetry arguments and conservation laws. This type of theory has been extensively used to describe complex fluids. The new feature of active gels is that one must explicitly take into account the energy consumption due to molecular motors by introducing a flux which is the rate of consumption of ATP molecules r and the corresponding thermodynamic force which is the free energy gain upon hydrolysis of an ATP molecule, μ. The constitutive equations involve macroscopic transport coefficients which can either be calculated from a specific microscopic theory or measured experimentally. A typical transport coefficient in a simple fluid is the shear viscosity. An important feature of the actin-myosin cytoskeleton is its viscoleasticity. The actin gel behaves as a solid with a finite shear modulus at short time scales or high frequency and as a liquid at long time scales or low frequency. Recent in vivo experiments (Fabry et al. 2001; Balland et al. 2006) show that the actin cytoskeleton has a broad distribution of relaxation times so that its complex elastic modulus increases as a power law of frequency G(ω) ∼ ωp where the exponent p is in the range 0.1, 0.2 for many cell types, (Fabry et al. 2001; Balland et al. 2006). For simplicity we use here the Maxwell model with a single relaxation time τ. In vitro experiments suggest that this relaxation time is in the range 10 – 100s. In the following we focus mostly on the long time behavior so that we consider the cytoskeleton as a liquid of viscosity η. In the absence of activity μ = 0, the hydrodynamic constitutive equations are the same as those obtained for a nematic liquid crystal or a ferroelectric liquid (DeGennes and Prost 1993) at long time scales and those of anisotropic gels at short time. They describe the viscous dissipation in the gel, the dynamics of orientation of the actin filaments and the coupling between flow and orientation of the filaments. The most important effect of the activity of molecular motors is an active contribution in the stress-strain relation. The active contribution has both an isotropic and a deviatoric term. The deviatoric term must be proportional to the local quadrupolar tensor which represents the average local orientation of the actin filaa = ζ μqαβ. In the absence of motion it can generate a normal ments qαβ : σαβ

17

Force Production by Actin Assembly

425

stress difference, whereas the isotropic term can generate a pressure. The deviatoric activity coefficient ζ is in general positive. This corresponds to a contractile stress in the direction of the actin filaments and to a dilative stress in the direction perpendicular to the actin filaments. In the hydrodynamic theory, the active stress is obtained by imposing the symmetries of the problem without any reference to the microscopic properties of molecular motors and their interactions with actin filaments. A simple microscopic picture is to associate this stress to myosin motor clusters crosslinking two different filaments and thus contracting the actin gel (Liverpool et al. 2009). Active systems exist at very different length scales and systems as different as bacterial colonies (Dombrowski et al. 2004), fish shoals, or bird flocks (Toner and Ramaswamy 2005) can be considered as active systems for which hydrodynamic theories have been constructed. The symmetries of these systems are similar to those of the cytoskeleton and the corresponding hydrodynamic theories have all the same structure with very different values for the transport coefficients (Simha and Ramaswamy 2002). The hydrodynamic theory of active gels has been used recently to study various properties of the cytoskeleton and in particular the actin cortical layer (Salbreux et al. 2007; 2009). It gives natural explanations for various instabilities of cells such as the formation of blebs (Tinevez et al. 2009; Paluch et al. 2005) or the existence of spontaneous oscillations. We now briefly discuss lamellipodium motion within this framework.

17.4.2 Experimental Systems The contractility of the acto-myosin network in cells has been evidenced recently by an elegant experiment using microtubules as reporters of contraction (Brangwynne et al. 2006). Microtubules are long (the cell size) and stiff (the persistence length is a few millimeters) polymers, which in a living cell display short-wavelength buckling. This effect is illustrated in cardiac myocytes where microtubules are shown to buckle and unbuckle with each wave of contraction and relaxation. In vitro, the actomyosin contration has been reproduced either with the use of cell extracts (Pollard 1976) or purified proteins. With purified proteins, it was known long ago (Takiguchi 1991) that acto-myosin solutions undergo contraction alternatively called superprecicipitation. However, contraction at physiological ATP concentrations requires the presence of an actin crosslinker. The maximal forces exerted by contraction in purified Xenopus extracts are on the order of 100 piconewton per F-actin bundle (Bendix et al. 2008). This corresponds to an active contractile stress of about 10 Pa. In cells, the contractile stress can reach larger values of the order of 100 – 100 Pa. Actomyosin contractility can drive cortical instability and lead to spontaneous cortex rupture (Paluch et al. 2005). As long as the cortex forms a closed shell, the tension that builds up owing to the presence of myosin II motors cannot be released. Such a closed geometry thus becomes unstable, and tension by itself can induce

426

C. Sykes et al.

Fig. 17.4 Consequences of enhanced cortical contractility. (a) Under control conditions, actomyosin cortical contraction is limited, which leads to two stable regions separated by a slight constriction: the cell body and a protrusion at the leading edge. If contractility is enhanced, there are two ways to release cortical tension: (b) the entire cell body can be propelled through the constriction; if the constriction region adheres to the surrounding matrix (black lines), the contraction wave that then crosses the cell body can result in net cell movement, allowing the cell to squeeze itself through gaps in the matrix; and (c) tension-induced ruptures of the cortex, followed by bleb formation. Green, actin; red, myosin II motors, blue, microtubules, grey, nucleus.

breakage of the cortex at places where contractility is locally higher or where the gel thickness decreases (Fig. 17.4)). The pierced actomyosin shell can then contract, leading to the expulsion of a membrane bulge. Actin and myosin accumulate at the edge of the hole that then appears in the cortex, creating a line tension limiting expansion of the hole. When cortical contractility is enhanced by depolymerization of the microtubules and when the cell is in suspension, the contraction of the cortex can expel the entire cell body, nucleus included, through the hole (Paluch et al. 2005). While the cell body contracts, its acto-myosin cortex partially depolymerizes or solates. An actomyosin cortex then reforms in the growing bulge, and once the entire cell body has passed through the hole, the cell returns to its initial shape (Fig. 17.4). A higher concentration of myosin motors at the place where the contraction ended favors new gel rupture close by. This results in cortical oscillations in which the constricting ring at the base of the bulge travels back and forth over the cell surface. Such cortical oscillations could be rectified in motility of cells adhering on a substrate (Paluch et al. 2006).

17

Force Production by Actin Assembly

427

17.4.3 Lamellipodium Motion of Keratocytes The motion of keratocyte cells provide a good example of cell motion driven by the acto-myosin cytoskeleton and has been studied experimentally in great details by several authors and in particular in the seminal work of (Verkhovsky et al. 1999; Vallotton et al. 2005). Keratocyte cells have a very flat lamellipodium and thicker cell body at their back. The lamellipodium is essentially made of an actomyosin cytoskeleton surrounded by a cell membrane. Keratocytes crawl on surfaces at velocities of the order of 10 μm per minute which is a rather large velocity for crawling cells. One of the spectacular results obtained by Verkhovsky is that circular cell fragments “cut” from the lamellipodium and for which the circular symmetry is broken by external mechanical action move essentially like the whole cell: with similar velocities and with the same canoe shape as the whole moving cell (Verkhovsky et al. 1999). Spontaneous symmetry breaking may also occur as shown by (Yam et al. 2007). In addition, the motion of keratocyte cells has been studied very quantitatively. The whole actin velocity field has been mapped using speckle microscopy (Vallotton et al. 2005) and shows the existence of a retrograde flow: at the leading edge of the lamellipodium, the actin velocity is in the direction opposite to the cell motion. The retrograde velocity is smaller than the cell advancing velocity and of the order of 1 μm per minute. The interaction of the cell with the substrate has also been studied in details by Oliver et al. (1999): the shear stress profile at the substrate has been determined and is of the order of σxz ∼ 103 Pa. For keratocytes the gel is fairly homogeneous, and its interaction with the substrate can be reasonably described as a viscous friction σxz = ξ v where v is the local advancing velocity on the substrate. The friction constant per unit area is on the order of ξ ∼ 1010 Pa.s/m. The geometry of the advancing keratocyte is rather complex and constructing a detailed theory of the motion is difficult. In (Kruse et al. 2006), a very simple one dimensional theory is proposed. In this theory only the coordinates x along the motion and z perpendicular to the substrate are retained. This is an oversimplification which implicitly assumes that the properties of the lamellipodium do not vary along the third direction y but even with this strong assumption, the active gel theory seems to capture some of the physical properties of the advancing cell. In this description the actin filaments are polarized along the direction x of the motion. They polymerize close to the tip of the advancing lamellipodium. We consider simply that actin polymerization proteins are localized in the membrane around the tip of the lamellipodium and that their density decays exponentially over a length λ of the order of 1 μm. The actin polymerization velocity decays then from the tip as vp = v0p exp −x/λ. Depolymerization occurs in the bulk of the cytoskeleton at the rear of the lamellipodium; for simplicity, we consider that depolymerization at the rear is dominant with a depolymerization velocity vd . The actin gel has a Poisson modulus roughly equal to 0.5 and in a first approximation, we consider it as incompressible. For a lamellipodium moving in the negative x direction at a constant velocity (−u) one can write

428

C. Sykes et al.

d h(u + v) = vp (x) dx

(17.9)

where h is the local thickness of the lamellipodium at position x. The lamellipodium is a very thin actin layer and its hydrodynamics can be studied using thin film equations. In this limit, the longitudinal stress σ and the velocity v only weakly vary over the thickness. The constitutive equation relating the stress to the velocity gradient can be written as 2η

dv = σ − ζ μ dx

(17.10)

We have assumed here that the lamellipodium behaves as a liquid and the last term is the active stress described in the last section. In the thin film approximation, the force balance is an equilibrium between the viscous stress and the friction on the substrate. dσ = ξv dx

(17.11)

The three equations 17.9, 17.10 and 17.11 for the three variables h, v and σ can be solved numerically. The results are plotted on Fig.17.5 for typical numbers corresponding to keratocytes. The plot of the thickness h shows a very flat region corresponding to the experimental lamellipodium profile. The velocity is retrograde at the leading edge of the lamellipodium and can be anterograde or not at the back depending on how strongly the bulk of the cell pulls. Around the tip, the velocity decays exponentially over a length d = 4ηh/ξ of the order of 6 μm: v (dζ μ/4η) exp −x/d. A comparison of the velocity at the tip x = 0 with the experimental value gives an active stress ζ μ ∼ 103 Pa. The third curve shows the integrated stress over the gel layer σh in the lamellipodium. In this model the

Fig. 17.5 Physical properties of an advancing lamellipodium. The three curves show the thickness profile of the lamellipodium h, the velocity profile v and the force profile F

17

Force Production by Actin Assembly

429

advancing velocity u is given at steady state by the depolymerization velocity vd and the motion is mostly related to actin polymerization. There is a small correction in the cell advancing velocity due to the retrograde flow. A more appropriate description considers actin as a visco-elastic fluid with a visco-elastic relaxation time τ. One then finds that the cytoskeleton behaves as a liquid except in a very thin region close to the tip of size l = uτ where it has a solid behavior. In this region the solid-like lamellipodium moves rearward at a velocity v(x = 0). The problem is very similar to the tip of stereocilia described in Section 17.2.3 and the interaction with the membrane regulates the polymerization velocity. The analysis presented here is made possible by the existence of precise experimental data on keratocytes. Similar data are available in the motion of neural growth cones (Betz et al. 2006) and lead to a smaller value of the active stress ∼ 50 Pa. The main merit of this analysis is that it obtains naturally the velocity profile and the hight profile. It allows for the estimate of the contractility coefficient and comparisons between different lamellipodia. For instance, the small value of the contractility coefficient might be due to the absence of the Arp 2/3 complex in growth cones. Other descriptions have been proposed that succeed in describing the polymerization process and the “top view” of the shape of the advancing cells (Keren et al. 2008). They however require more assumptions and are in this sense less generic.

17.5 Conclusion The recent knowledge acquired on proteins that are involved in cellular actinand actomyosin-based movements has allowed for the development of simplified experimental systems that enable quantitative characterizations in a controlled environment. In parallel, experiments on cells have benefited from improved optical techniques with the use of sophisticated biochemical markers. Based on this experimental work, generic models have been developed. They allow for a thorough understanding of many different situations in cells, from contractility to polymerization-based movements. The molecular details of how actin and myosin motors are assembled and interact are effectively taken into account in the models but do not play a critical role, since the models are generic. However, molecular models and mesoscopic models use different approaches that still need to be united in a single framework. The crossover between the single filament scale that is described in this chapter, and the active gel theory is one of the next scientific challenges for the cell biophysics community.

References M. Abercombie, Proc R Soc Lond B Biol Sci 207, 129 (1980). B. Alberts, et al., Molecular Biology of the Cell (Garland New York, NY, 2002). I. Aranson and L. Tsimring, Phys Rev E Stat Nonlin Soft Matter Phys. 74, 031915 (2006).

430

C. Sykes et al.

M. Balland, N. Desprat, D. Icard, S. Fereol, A. Asnacios, J. Browaeys, S. Henon, and F. Gallet, Phys Rev E Stat Nonlin Soft Matter Phys 74, 021911 (2006). P. M. Bendix, G. H. Koenderink, D. Cuvelier, Z. Dogic, B. N. Koeleman, W. M. Brieher, C. M. Field, L. Mahadevan, and D. A. Weitz, Biophys J 94, 3126 (2008). A. Bernheim-Groswasser, S. Wiesner, R. M. Golsteyn, M.-F. Carlier, and C. Sykes, Nature 417, 308 (2002). J. Berro, A. Michelot, L. Blanchoin, and J. Martiel, Biophys J 92, 2546 (2007). T. Betz, D. Lim, and J. A. Kaes, Phys Rev Lett 96, 098103 (2006). H. Boukellal, O. Campas, J. Joanny, J. Prost, and C. Sykes, Phys Rev E 69, 061906 (2004). C. Brangwynne, F. MackIbtosh, S. Kular, N. Geisse, J. Talbot, L. Mahadevan, K. Parker, D. Ingber, and D. Weitz, J Cell Biol 173, 733 (2006). L. A. Cameron, M. J. Footer, A. van Oudenaarden, and J. A. Theriot, Proc Natl Acad Sci USA 96, 4908 (1999). P. DeGennes and J. J. Prost, The Physics of Liquid crystals (Oxford University Press, Oxford, 1993). V. Delatour, E. Helfer, D. Didry, K. H. D. Le, J.-F. Gaucher, M.-F. Carlier, and G. Romet-Lemonne, Biophys J 94, 4890 (2008). M. Dogterom and B. Yurke, Science 278, 856 (1997). C. Dombrowski, L. Cisneros, S. Chatkaew, R. E. Goldstein, and J. O. Kessler, Phys Rev Lett 93, 098103 (2004). K. Doubrovinski and K. Kruse, Phys Rev Lett 99, 228104 (2007). B. Fabry, G. N. Maksym, J. P. Butler, M. Glogauer, D. Navajas, and J. J. Fredberg, Phys Rev Lett 87, 148102 (2001). M. Footer, J. Kerssemakers, J. Theriot, and M. Dogterom, Proc Nat Acad Sci 104, 2181 (2007). M. Gardel, J. Shin, F. MacKintosh, L. Mahadevan P. Matsudaira, and D. Weitz, Science 304, 1301 (2004). F. Gerbal, P. Chaikin, Y. Rabin, and J. Prost, Biophys J 79, 2259 (2000). F. Gerbal, V. Laurent, A. Ott, M. F. Carlier, P. Chaikin, and J. Prost, Eur Biophys J 29, 134 (2000). P. A. Giardini, D. A. Fletcher, and J. A. Theriot, Proc Natl Acad Sci USA 100, 6493 (2003). T. L. Hill, Biophys J 49, 981 (1986). T. L. Hill, Proc Nat Acad Sci 78, 5613 (1981). T. Holton and A. J. Hudspeth, Science 222, 508 (1983). J. Howard, Mechanics of Motor Proteins and the Cytoskeleton (Sinauer Associates, Sunderland, 2001). F. Julicher and J. Prost, Eur Phys J E Soft Matter 29, 27 (2009). F. Julicher, K. Kruse, J. Prost, and J.-F. Joanny, Phys Rep 449, 3 (2007), ISSN 0370-1573, nonequilibrium physics: From complex fluids to biological systems III. Living systems. K. Keren, Z. Pincus, G. Allen, E. Barnhart, A. Marriott, A. Mogilner, and J. Theriot, Nature 453, 475 (2008). C. Kocks, E. Gouin, M. Tabouret, P. Berche, H. Ohayon, and P. Cossart, Cell 68, 521 (1992). D. R. Kovar and T. D. Pollard, Proc Nat Acad Sci 101, 14725 (2004). K. Kruse and F. Julicher, Phys Rev Lett 85, 1778 (2000). K. Kruse, J. F. Joanny, F. Julicher, and J. Prost, Phys Biol 3, 130 (2006). K. Kruse, J. F. Joanny, F. Julicher, J. Prost, and K. Sekimoto, Phys Rev Lett 92, 078101 (2004). K. Kruse, J. Joanny, F. Julicher, J. Prost, and K. Sekimoto, Eur Phys J E 16, 5 (2005). I. Lasa, V. David, E. Gouin, J. B. Marchand, and P. Cossart, Mol Microbiol 18, 425 (1995). X. Li, J. Kierfeld, and R. Lipowsky, Phys Rev Lett 103, 048102 (2009). T. B. Liverpool, M. C. Marchetti, J.-F. Joanny, and J. Prost, EPL (Europhysics Letters) 85, 18007 (6 pp) (2009). T. P. Loisel, R. Boujemaa, D. Pantaloni, and M. Carlier, Nature 401, 613 (1999). Y. Marcy, J. Prost, M.-F. Carlier, and C. Sykes, Proc Natl Acad Sci USA 101, 5992 (2004). A. Mogilner and G. Oster, Biophys J 84, 1591 (2003). M. Naoz, U. Manor, H. Sakaguchi, B. Kachar, and N. Gov, Biophys J 95, 5706 (2008).

17

Force Production by Actin Assembly

431

V. Noireaux, R. M. Golsteyn, E. Friederich, J. Prost, C. Antony, D. Louvard, and C. Sykes, Biophys J 78, 1643 (2000). T. Ohm and A. Wegner, Biochemistry 30, 11193 (1991). T. Oliver, M. Dembo, and K. Jacobson, J Cell Biol 145, 589 (1999). F. Oosawa and S. Akasura, Thermodynamics of the Polymerization of proteins (Academic, London, 1975). E. Paluch, C. Sykes, J. Prost, and M. Bornens, Trends Cell Biol 16, ((2006). E. Paluch, J. van der Gucht, J.-F. Joanny, and C. Sykes, Biophys J 91, 3113 (2006). E. Paluch, M. Piel, J. Prost, M. Bornens, and C. Sykes, Biophys J 89, 724 (2005). D. Pantaloni, T. L. Hill, M. F. Carlier, and E. D. Korn, Proc Nat Acad Sci 82, 7207 (1985). S. H. Parekh, O. Chaudhuri, J. A. Theriot, and D. A. Fletcher, Nat Cell Biol 7, 1219 (2005). T. Pollard, J. Cell Biol 68, 579 (1976). J. Prost, C. Barbetta, and J.-F. Joanny, Biophys J 93, 1124 (2007). P. Ranjith, D. Lacoste, K. Mallick, and J.-F. Joanny, Biophys J 96, 2146 (2009). P. Ranjith, K. Mallick, J. F. Joanny, and D. Lacoste, Biophys J 98, 1418 (2010). S. Romero, C. L. Clainche, D. Didry, C. Egile, D. Pantaloni, and M.-F. Carlier, Cell 119, 419 (2004). A. Rzadzinska, M. Schneider, C. Davies, G. Riordan, and B. Kachar, J Cell Biol 164, 887 (2004). G. Salbreux, J. F. Joanny, J. Prost, and P. Pullarkat, Phys Biol 4, 268 (2007). G. Salbreux, J. Prost, and J. F. Joanny, Phys Rev Lett 103, 058102 (pages 4) (2009). F. T. Salles, R. C. Merritt, U. Manor, G. W. Dougherty, A. D. Sousa, J. E. Moore, C. M. Yengo, A. C. Dose, and B. Kachar, Nat Cell Biol 11, 443 (2009). G. SandervanDoorn, C. Tanase, B. Mulder, and M. Dogterom, Eur Biophys J 29, 2 (2000). M. Schneider, I. Belyantseva, R. Azevedo, and B. Kachar, Nature 418, 837 (2002). R. Simha and S. Ramaswamy, Phys Rev Lett 89, 058101 (2002). E. B. Stukalin and A. B. Kolomeisky, Biophys J 90, 2673 (2006). K. Takiguchi, J Biochem 109, 520 (1991). J. Tan, J. Tien, D. M. Pirone, D. Gray, K. Bhadriraju, and C. Chen, Proc Natl Acad Sci USA 100, 1484 (2003). J. Theriot, T. Mitchison, L. Tilney, and D. Portnoy, Nature 357, 257 (1992). L. Tilney, M. Tilney, and D. DeRosier, Annu Rev Cell Biol 8, 257 (1992). J. Tinevez, U.Schultze, G. Salbreux, J. Roensch, J. Joanny, and E. Paluch, Proc Nat Acad Sci 106, 18581 (2009). J. Toner, Y. Tu, and S. Ramaswamy, Adv Phys 318, 170 (2005). L. Trichet, O. Campas, C. Sykes, and J. Plastino, Biophys J 92, 1081 (2007). A. Upadhyaya, J. Chabot, A. Andreeva, A. Samadani, and A. van Oudenaarden, Proc Natl Acad Sci USA 100, 4521 (2003). P. Vallotton, G. Danuser, S. Bohnet, J.-J. Meister, and A. B. Verkhovsky, Mol Biol Cell 16, 1223 (2005). J. van der Gucht, E. Paluch, J. Plastino, and C. Sykes, Proc Nat Acad Sci 102, 7847 (2005). D. Vavylonis, Q. Yang, and B. O’Shaughnessy, Proc Natl Acad Sci USA 102, 8543 (2005). A. B. Verkhovsky, T. M. Svitkina, and G. G. Borisy, Curr Biol 9, 11 (1999). M. Welch, J. Rosenblatt, J. Skoble, D. Portnoy, and T. Mitchison, Science 281, 105 (1998). S. Wiesner, E. Helfer, D. Didry, G. Ducouret, F. Lafuma, M. Carlier, and D. Pantaloni, J Cell Biol 160, 387 (2003). P. T. Yam, C. A. Wilson, L. Ji, B. Hebert, E. L. Barnhart, N. A. Dye, P. W. Wiseman, G. Danuser, and J. A. Theriot, J Cell Biol 178, 1207 (2007).

Subject Index

A Acrosomal process, 6, 8 Actin regulatory proteins, 69–70, 137, 241 Active gels, 424–425 Actobindin, 271–272 ADF/cofilin, 18, 63, 68, 172, 215–219, 221–222, 227, 240–242, 245–247, 250, 256, 330, 361, 383, 399–400, 422 Adhesion, 5, 7, 22–24, 71, 76–79, 95, 105–113, 120, 143, 153–154, 157, 190, 193–195, 197, 204, 219, 251, 280, 302, 306, 328, 360, 365, 374–376, 383, 395, 397–403, 405–406, 414 AFM (atomic force microscopy), 368–374, 422 AlfA, 335, 339, 351–353 Amphiphysin, 36–44, 50, 69–70, 91, 93–94, 271 Arf6, 64, 72, 74–76, 78 Assymetric cell division, 116, 250 ATP hydrolysis, 214, 239, 298–299, 331, 340, 346–348, 382, 384–386, 394 B Bacterial actin, 335–354 Barbed and pointed ends of filaments, 270 BAR domain, 36–46, 48, 50–52, 66, 70, 91, 93–94, 249, 271 Blebbing, 4, 8, 62, 75, 195 Blebbistatin, 24, 153, 155, 171, 178, 195, 199, 203 Bottom-up reconstitution, 239 Brownian Ratchet, 96, 343–344, 363, 367, 385–386 Bundles, 5–6, 8–9, 20–24, 37–38, 46–47, 170, 221, 226, 247–249, 271, 280, 282, 291, 299, 304–307, 327, 338, 349, 351–352, 364, 367, 386, 387, 395–398, 414, 416, 418–419, 423, 425

C Caenorhabditis Elegans, 42 Capping proteins, 12, 18, 63, 70, 147, 216, 219, 222–227, 241–243, 245, 247, 251, 266–268, 282, 288–290, 305, 361, 371, 383, 390–391, 420–421 Caveolae, 36–39, 43–44, 46 Cdc8, 21–22, 37, 42, 45, 52, 63, 69–70, 73, 76, 106, 138, 142, 146, 151–152, 156–157, 166, 178, 180, 193–194, 245, 261, 304, 306–307, 401 Cellularization, 250 Chemokines, 105, 108, 116–117, 120 Chemotaxis, 24, 61, 72, 127, 129–132, 134, 140, 144, 156–158, 199, 302 Chromosome, 142, 165–181, 238, 308, 343–344, 348 Ciboulot, 271–272 Clathrin, 36–40, 43–44, 51, 62, 64–65, 69, 74–76, 85–97 Contractility, 7, 61–62, 75, 77, 178, 195, 199, 251, 403, 414, 423, 425–426, 429 Cordon-bleu, 282 Cortactin, 13–14, 18–19, 67–68, 74, 92–95, 107, 128, 130, 134, 136–137, 140–147, 149–151, 155, 157 Crawling, 189, 302, 324, 359–376, 383, 427 Cryo-electron microscopy, 46, 51, 97 Cytochalasin (B or D), 19–20, 91, 110, 179 D Dendritic cells, 22, 105 Dictyostelium discoideum, 42, 62, 129, 132–133 Differential Interference Contrast (DIC), 113 Drosophila melanogaster, 42, 72

M.-F. Carlier (ed.), Actin-based Motility, DOI 10.1007/978-90-481-9301-1, C Springer Science+Business Media B.V. 2010

433

434 Dynamic instability, 215, 339–342, 345, 348–353 Dynamin, 36, 38, 40, 42–44, 47, 49, 64–72, 74, 76, 93–94, 139, 143–144, 149–150 E Elastic tethered ratchet, 365–366 Elongation, 7, 14–15, 19, 21, 73, 94, 133, 168, 170, 172, 174, 214–215, 256, 260, 268–271, 281, 284, 287–290, 292–299, 301–305, 309, 322–327, 339–340, 344, 347–352, 361, 387, 396, 415–416 Embryo, 61, 143, 167, 180–181, 188–196, 198–199, 203, 205, 220, 225, 298, 305 Ena/VASP, 14–15, 19, 42, 70, 191, 220, 223, 259, 265, 269, 298 Exocyst, 64, 72–73, 155–156 Extra-cellular matrix (ECM), 4, 7–8, 47, 77–78, 126, 134, 362, 376, 414 F Fibroblasts, 5–7, 16, 18, 22–24, 63, 92, 113, 133, 143–144, 197, 199, 219, 289, 305, 376, 403, 406 Filament branching, 242, 245, 249, 329, 371 Filamin, 11–12, 256, 322–323 Force generation, 77, 112, 154, 172, 360–373, 382–394, 416 FRAP (fluorescence recovery after photobleaching), 15–17, 19–20, 110, 241 FSM (fluorescence speckle microscopy), 15–16, 251 Functionalized particles, 242, 245–247 G Gastrulation, 188–189, 305 Gelsolin, 11, 215, 219, 223–225, 227, 247, 256, 270, 288, 301, 422 I Immunological synapse, 103–120 Ingression, 188 Insertional polymerization, 245, 344–348 Integrins, 8, 21, 75–79, 106–109, 113, 136, 376 Intercalation, 190, 363 Intravasation, 126–127, 130, 135, 140, 156–157 Invagination, 36–41, 44, 51, 76, 87–91, 93–96, 188, 214, 239, 271 Invasion, 60, 68, 71–79, 125–158 J Jasplakinolide, 169, 171

Subject Index K Keratocytes, 7, 9, 11, 16, 65, 72, 119, 360–361, 367, 373, 375–376, 401, 405–406, 427–429 L Latrunculin (A or B), 19, 44, 51, 91, 95, 110–111, 169, 181 Leading edge, 61–65, 71, 108, 128, 131, 133, 153, 166, 188, 191–194, 196, 198, 200–205, 214, 217, 241, 244, 250, 280, 288, 305, 324, 328–329, 340, 360, 365, 374–375, 383, 390–391, 395, 398–402, 406, 426–428 Liposomes, 41, 43, 50, 422–423 Listeria monocytogenes, 96, 171, 244, 260–261, 361, 421 M Mammary carcinoma cells, 126–128, 130–132, 135–136, 142–143, 145, 147, 149–151, 153, 156–157 Mechano-sensing, 113 Metastasis, 67–68, 126–127, 132, 134, 137–144, 152–154, 156–158, 257, 271 Microscopy, 4, 8–11, 15–16, 20–21, 46, 51, 68, 87–88, 92, 97, 109, 111, 113–115, 117, 119, 135, 169–170, 192, 244, 247, 251, 264, 289, 317–319, 321, 323, 326, 328–331, 336–337, 339, 341–342, 345–346, 350, 369, 375, 382, 387, 394, 400, 427 Microspikes, 4–7, 17, 20–22 Mitotic spindle, 166–168, 308, 343 Molecular motors, 239, 291, 362, 371, 419, 423–425 Morphogenesis, 60–61, 73, 188–202, 214, 222–223, 250, 286, 302, 305 MreB, 337–338, 344, 353 N Neural growth cone, 429 Newtonian fluid, 403 Numerical modeling, 401 O Oocytes, 167–174, 176–181, 250, 300 P PALM, 118–119, 331 ParM, 336–353 PDGF (platelet-derived growth factor), 63, 72, 144 Phagocytosis, 7, 23, 37, 46–47, 214, 219

Subject Index Phalloidin, 11, 75, 113, 170, 172, 179, 190, 199, 317–318, 329 Phase contrast, 4, 246–248 Phosphoinositides, 45, 51, 69, 218–220, 222, 225 Plasmid segregation, 238, 335–354 Polarized helical polymer, 318 Processive filament attachment, 347 R Rafts, 39, 46, 63–64, 76, 118, 127, 132, 143, 415, 421 Reaction-diffusion equations, 403 Retraction, 4, 7, 22–24, 77, 111–112, 153, 196, 198, 203, 219, 306, 360, 375, 383, 390, 396–398, 401–405 Rho, 7–8, 24, 67, 70, 75, 107–108, 120, 152, 154, 166, 170, 177–178, 193, 195–197, 200–201, 283–286, 304, 306–309, 405 Rho GTPases, 7–8, 152, 154, 170, 178, 195, 200–201, 285–286, 307, 309, 405 Ruffles, 7, 13, 23–25, 66, 72–75, 112–113 Ruffling, 16, 23–24, 70–71, 219 S Saccharomyces Cerevisiae, 42, 86, 221 SAXS (small angle X-ray scattering), 262, 264–266 Self-assembly, 238–340 Self-organized systems, 237–251 Spire, 13, 202, 250, 256, 258–261, 265, 282, 300, 309, 416 Steady state, 114, 117, 130, 214, 239–241, 339, 348, 350–352, 366, 384, 399, 401, 429 Stereocilia, 419–421, 429 Stick-slip adhesion, 398

435 Stress fibers, 5, 66, 152, 196–197, 271, 282, 302, 306 Symmetry breaking, 115, 119, 167–169, 171, 174, 254, 382, 392–394, 421, 427 T T-cells, 105–106, 109, 223, 308 TCR (T-cell receptor), 105–111, 113–120 Thermal bending, 363 β-Thymosins, 216, 220, 223 Total Internal Reflection Fluorescence Microscopy (TIRF), 68, 92, 111, 289, 318–323, 387, 394 Treadmilling, 7, 15–16, 22–23, 214–217, 219, 239–244, 247, 251, 352–353, 363, 419–420 Tropomodulin, 223, 226, 266 Twinfilin, 216–217, 219, 222–225, 227 Two photon microscopy, 114 V Vesicle fission, 43, 48–49, 69 Vesicles, 13, 21, 36, 38, 40, 43, 48–52, 60, 64–67, 69–76, 78, 85–97, 112, 118, 150, 155–156, 166, 168–170, 214, 239, 241, 244–245, 271, 280, 290, 304, 307, 421 Viscoelasticity, 367–368 Viscous drag, 362, 422 W WASH, 13, 260–261, 263 WH2 (WASP-homology 2) domain, 13–14, 42, 47, 221, 223, 255–272, 282, 299 X X-ray crystallography, 341, 345–346 Z Zebrafish, 61, 65, 77, 195–196, 198–200, 203, 205, 305