The Role of a Trimeric Coiled Coil Protein in WASH Complex Assembly

The role of a trimeric coiled coil protein in WASH complex assembly

Thèse de doctorat de l'Université Paris-Saclay, préparée à l’Université Paris-Sud NNT: 2017SACLS291 2017SACLS291 NNT:

École doctorale n°577: Structure et Dynamique des Systèmes Vivants (SDSV) Spécialité de doctorat: Sciences de la Vie et de la Santé

Thèse présentée et soutenue à Palaiseau, le 22 septembre 2017, par 06DL3UDVDQQD9,6:(6+:$5$1

Composition du Jury :

'U0DUF0,5$1'( Directeur de Recherche CNRS, I2BC, Gif-sur-Yvette Président

'U3KLOLSSH&+$95,(5 Directeur de Recherche CNRS, Institut Curie, Paris Rapporteur

'U(PPDQXHO'(5,9(5< Chercheur, Laboratoire LMB du MRC, Cambridge, UK Rapporteur

'U5DSKDsO*8(52,6 Chercheur, CEA, I2BC, Saclay Examinateur

'U(YJHQ\'(1,629 Chercheur, Tomsk State University, Examinateur Tomsk Cancer Research Institute, Tomsk, Fédération de Russie

3U$OH[LV*$875($8 Directeur de Recherche CNRS, Professeur associé à l’Ecole Polytechnique, Palaiseau Directeur de thèse

I dedicate this thesis to my father and mother as a token of gratitude for their immense support and trust in me

This page intentionally left blank

2 Table of contents

Acronyms 05

Summary in French 07

Introduction 21

I Branched actin network and its regulation in the cell 23 Overview 23 1. Actin and actin filaments 25 1.1 The dynamic nature of the actin filament polymerization 25 1.2 In vivo actin and actin filaments regulation 27 1.2.1 In vivo regulators of actin 27 1.2.2 Regulators of actin filament nucleation and elongation 29 2. Branched actin network by the Arp2/3 complex 31 2.1 Characteristics of the Arp2/3 complex 33 2.2 Molecular function of the Arp2/3 complex 33 2.3 Arp2/3 complex activation and nucleation mechanism 33 2.4 Inhibitors of Arp2/3 complex and debranching factors 37 3. Nucleation promoting factors (NPFs) 38 3.1 N-WASP protein 38 3.1.1 Function of the N-WASP in cells 39 3.1.2 Molecular characteristics of the N-WASP 39 3.1.3 Activation of the N-WASP 41 3.2 The WAVE complex 41 3.2.1 Function of the WAVE complex in cells 43 3.2.2 Molecular characteristics of the WAVE complex 43 3.2.3 Activation of the WAVE complex 45 3.2.3.1 Activation by Rac 45 3.2.3.2 Activation by molecular interactions 45 3.2.3.3 Activation by phosphorylations 46 3.3 The WASH complex 45 3.3.1 Function of the WASH complex in cells 46

3 3.3.2 Molecular characteristics of the WASH complex 49 3.3.3 Regulation of the WASH complex 51 3.4 WHAMM/JMY 53 II The role of the WASH complex in endosomal system 57 Overview 57 1. A glance at endocytic pathway 57 2. Endosomal fusion and fission processes 58 3. Membrane scission mediated by the WASH complex 61 4. Endosomal sorting mediated by the WASH complex 62 5. Pathologies of defective WASH complex 65 5.1 Its role in neurodegenerative diseases 65 5.2 Its role in tumor progression 68 III Assembly of multi-protein complexes 69 1. Molecular machines in nature 69 2. Specific assembly mechanisms of molecular machines 69 3. The WAVE complex assembly 73

Objectives 77

Results 79

Discussion 87

1. HSBP1 dissociate CCDC53 trimers to promote WASH complex assembly 89 2. HSBP1 function is conserved and is necessary for CCDC53-WASH sub-complex 91 formation during WASH assembly pathway 3. HSBP1 inactivation phenocopies WASH complex inactivation in the cells 92 4. HSBP1 putative role at the centrosome 96 5. HSBP1 putatively drive cancer progression through WASH complex assembly 98

References 101

Acknowledgements 121

4 Acronyms

Arp Actin Related Protein AMPA -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid

AMPAR αAMPA-type glutamate Receptor BAR Bin-Amphisin-Rvs CRIB Cdc42 Rac Interactive Binding Domain DAD Diaphanous Autoregulatory Domain DID Diaphanous Inhibited Fomin GBD GTPase Binding Domain HSP70 Heat Shock Protein 70 HSF1 Heat Shock Factor 1 HSE Heat Shock Element HSBP1 Heat Shock Binding Protein 1 IRSp53 Insulin Receptor Tyrosine Kinase Substrate p53 JMY Junction Mediating and regulatory protein MTOC The Microtubule-Organizing Center

NPF Nucleation Promoting Factor PI(3)P Phosphatidylinositol-3-Phosphate PIP2 Phosphatidylinositol (4,5) Diphosphate PIP3 Phosphatidylinositol (3, 4, 5) triphosphate PRD Proline Rich Domain PDZ Derived from the names of first proteins in which this domain is found - Post-synaptic density protein 95 (PSD-95), Disks large homolog (Dlg1) and Zona occludens 1 (ZO-1) Ras Ras Sarcoma onco-proteins

Rac Ras-related C3 botulinum toxin substrate

Rho Ras Homologous proteins

Rab Ras-like proteins in Brain SCAR Suppressor of cAMP Receptor SH3 Src Homology 3

5 SHD SCAR homology domain TOCA-1 Transducer of Cdc42 Activity-1 WASH Wiskott Aldrich Syndrome Protein and Scar Homolog WASP Wiskott Aldrich Syndrome Protein WAVE WASP family Verprolin Homologous Protein WCA WH2, Connecting Region, Acidic domain WH2 WASP Homology Domain 2 WHAMM WASP Homologue associated with Actin, Membranes and Microtubules WIP WASP Interacting Protein

Summary in French

This page intentionally left blank

8 Introduction

Les cellules utilisent les réseaux d’actine branché pour contrôler leur forme, pour migrer et pour remodeler ses membranes dans le trafic intracellulaire (Rotty et al. 2013a). Le complexe Arp2/3 est le complexe multiprotéique qui génère ces réseaux d’actine branché. Il contient deux protéines associées à l’actine, Arp2 et Arp3, et cinq autres sous-unités qui maintiennent les deux sous-unités Arp2 et Arp3 associées. Lorsqu’il est activé par le domaine WCA des nucléateurs, appelés « Nucleation Promoting Factors » (NPF), le complexe Arp2/3 induit une branche formée d’actine (Pollard 2007) : il s’associe à un filament d’actine préexistant et nucléé une nouveau filament à partir des 2 sous-unités Arp2 et Arp3, qui sont mises en contact et miment l’extrémité d’un nouveau filament (Rouiller et al. 2008). Un tel complexe multiprotéique peut être définit comme une machine moléculaire pour mettre en évidence les fonctions coordonnées qu’il exerce (Alberts 1998). Aucune fonction n’a été assignée à Arp2 ou Arp3 seule, en dehors du complexe Arp2/3.

Les NPFs activent le complexe Arp2/3 à différentes localisations cellulaires : WAVE aux lamellipodes, où les réseaux d’actine branché génèrent la force nécessaire à la formation de protrusions membranaires (Rotty et al. 2013b)(Figure I), et WASH à la surface des endosomes, où la force générée par les réseaux d’actine branchés contribue à la scission des intermédiaires de transport (Derivery et al. 2009b, Gomez and Billadeau 2009)(Figure II). Ces intermédiaires de transport suivent la route rétrograde vers le Golgi (Gomez and Billadeau 2009, Harbour et al. 2010) ou recyclent des récepteurs internalisés vers la membrane plasmique (Temkin et al. 2011, Piotrowski et al. 2013). Les intégrines 5- 1 sont des cargo qui prennent les deux routes dépendantes de WASH puisqu’elles αsontβ recyclées à la membrane plasmique à la fois via les endosomes et après un passage par le Trans-Golgi (Zech et al. 2011, Duleh and Welch 2012, Shafaq-Zadah et al. 2015, Nagel et al. 2016). WAVE et WASH sont tous les deux associés à quatre autres protéines, qui contrôlent l’exposition du domaine WCA (Derivery and Gautreau 2010a, Rotty et al. 2013b). Le recrutement endosomal du complexe WASH dépend de la reconnaissance du complexe formé avec le cargo du rétromère (Harbour et al. 2012, Jia et al. 2012, Helfer et al. 2013, Gautreau et al. 2014)(Figure III). La formation de réseaux d’actine branché implique donc des cascades de machines moléculaires.

Figure I. Localisation de WAVE au lamellipode. Séquences d’images de cellules de mélanomes B16F1 exprimant GFP-WAVE en microscopie à fluorescence (haut) et en contraste de phase (bas). WAVE est localisé au lamellipode et disparait lorsque la cellule se rétracte. Adapté (Hahne et al. 2001).

10 La façon dont ces machines moléculaires sont assemblées à partir de sous-unités néosynthétisées n’est, dans la plupart des cas, pas connue. En effet, les machines moléculaires ne sont pas un simple assemblage induit par l’association spontanée des sous-unités. La simple addition, pas à pas, des sous-unités ou sous-complexes conduit à une complexe WAVE dans lequel le domaine WCA n’est pas proprement masqué (Innocenti et al. 2004, Derivery et al. 2009a). La reconstitution du complexe WAVE natif a été un tour de force, qui a requis une décennie de travail (Chen et al. 2014). En effet, dans la cellule, le protéasome exerce un contrôle qualité et dégrade jusqu’à 30% des protéines néosynthétisées (Yewdell et al. 2000). Lorsqu’une sous-unité des complexes WAVE, WASH ou Arp2/3 est déplétée, les autres sous-unités d’un même complexe sont généralement dégradées par le protéasome (Kunda et al. 2003, Steffen et al. 2006, Derivery et al. 2009b, 2009a, Jia et al. 2010). A l’inverse, quand une sous-unité exogène, généralement taguée, est surexprimée, la sous-unité endogène est dégradée, car ses sous-unités partenaires ont été complexées avec la protéine exogène plus abondante (Derivery et al. 2009b, 2009a). Ces observations suggèrent que les sous-unités doivent s’assembler avec leur sous-unités partenaires pour atteindre leur niveau natif et devenir stable (Derivery and Gautreau 2010a). Dans le cas du complexe WAVE, une sous-unité, Brk1 forme un homotrimère précurseur, bien qu’une seule sous-unité Brk1 soit présente dans le complexe natif (Derivery et al. 2008, Linkner et al. 2011)(Figure IV). Le turnover de Brk1 est plus rapide que celui du complèxe WAVE, ce qui suggère que le deux molécules de Brk1 qui subsistent après la dissociation du trimère sont aussi dégradées (Derivery and Gautreau 2010a, Wu et al. 2012).

Les grandes machines moléculaires, telles que les protéasomes, requièrent de multiples facteurs pour leur assemblage (Sahara et al. 2014, Budenholzer et al. 2017). Les facteurs d’assemblage s’associent de façon transitoire avec une ou plusieurs sous- unités, mais peuvent aussi éventuellement se dissocier avant que la machine soit complète. Si ces facteurs d’assemblage restaient associés, ce seraient des sous-unités. Il n’est pas établit si les facteurs d’assemblage sont systématiquement requis pour la formation de petites machines moléculaires, telles que les complexes Arp2/3 ou les NPFs. Jusqu’à maintenant, seul un facteur d’assemblage a été identifié pour le complexe WAVE : la protéine Nudel, qui interagit transitoirement avec 2 sous-complexes, est critique pour maintenir le niveau du complexe WAVE et donc pour former des

Figure II. WASH est impliqué dans la scission des endosomes. a, Une internalisation de la transferrine fluorescente a été réalisée sur des cellules fibroblastiques de souris transfectées avec un siRNA ctrl ou un siRNA ciblant WASH. Les cellules ont été observées par microscopie à épifluoresence. Lorsque WASH est déplété, les endosomes forment des tubules (flèches rouges). b, Représentation d’un bourgeon émanant d’un endosome. Une tension de membrane est créée par deux forces opposées: d’une part par les moteurs microtubulaires qui tirent le bourgeon pour former le tube (flèche droite), d’autre part la formation d’un réseau d’actine branché autour du cou du bourgeon, initié par le complexe WASH, et qui, à l’aide de la Dynamine, favorise la scission du tube (flèche gauche). Adapté de (Derivery et al. 2009b)

12 lamellipodes (Wu et al. 2012). Les facteurs d’assemblage représentent un moyen de contrôler les niveaux de complexes assemblés. Par exemple, la starvation induit l’expression de facteurs d’assemblage des protéasomes et donc favorise l’assemblage de nouveaux protéasomes pour promouvoir la dégradation des vieilles protéines, permettant ainsi la biosynthèse de nouvelles protéines dans des conditions restrictives en acides aminés (Rousseau and Bertolotti 2016).

Les complexes Arp2/3 et WAVE sont surexprimés dans de nombreux cancers (Molinie and Gautreau 2017). Ces surexpressions sont généralement associées à un haut grade, à l’invasion des ganglions lymphatiques et à un faible pronostic pour les patients (Semba et al. 2006, Iwaya et al. 2007, Molinie and Gautreau 2017). Puisque la majorité des sous-unités de ces machines moléculaires ne sont stables qu’à l’intérieur du complexe entier, cela signifie que les cellules tumorales invasives réussissent à assembler plus de ces complexes, mais le mécanisme impliqué n’est pas connu. Le complexe WASH, qui permet la distribution ciblée des métalloprotéases et le recyclage des intégrines, est critique pour l’invasion des cellules tumorales (Zech et al. 2011, Monteiro et al. 2013), mais il n’a pas été montré que son expression était dérégulée dans les tumeurs. Dans cette étude, nous avons identifié le premier facteur d’assemblage du complexe WASH, HSBP1, et caractérisé comment il facilite l’assemblée du complexe WASH. Nous avons trouvé que HSBP1 est surexprimé dans le cancer du sein et que sa surexpression est associée à une augmentation du niveau du complexe WASH et à une faible survie pour les patientes.

Résultats

Nous avons identifié HSBP1 comme un facteur d’assemblage critique du complexe WASH, grâce à un screen protéomique, où il se lie seulement avec la formé précurseur de la sous-unité CCDC53 et pas avec le complexe entier.

Par des expériences in vitro et de la modélisation structurale, nous avons démontré que HSBP1 s’associe avec le précurseur CCDC52 trimérique, le dissocie, et forme un trimère hétérogène qui va éventuellement libérer une seule molécule de CCDC53 pour l’assemblage du complexe WASH. Nous avons également montré que la déplétion de HSBP1 déstabilise le complexe WASH dans ces cellules de mammifère mais

Figure III. Régulation du complexe WASH. La protéine WASH est associée en un complexe stable avec FAM21, Strumpellin, SWIP, Ccdc53 et à l’hétérodimère CapZ et .

La sous-unité FAM21 lui permet d’interagir avec le complexe rétromère surα lesβ endosomes. Le complexe E3 ubiquitine ligase, TRIM27-MAGE-L2, également recruté par le complexe rétromère, poly-ubiquitine le domaine flexible de la sous-unité WASH qui expose alors son motif VCA. Le complexe Arp2/3 peut alors être activé et initie la formation d’un réseau d’actine branché à la surface de l’endosome. Adapté de (Gautreau et al. 2014)

14 également chez l’amibe, ce qui en fait un facteur d’assemblage conservé entre les espèces.

En cohérence avec un assemblage du complexe WASH déficient, la déplétion de HSBP1 bloque le développement d’adhésions focales et l’invasion de carcinomes mammaires, qui sont tous deux dus à un trafic des intégrines défectueux. Nous avons trouvé que HSBP1 est localisé aux centrosomes et est requis pour la polarisation cellulaire associée à la migration. De plus, le knock-out de HSBP1 phénocopie le knock- out de WASH chez Dictyostelium amoebae.

De plus, en analysant les niveaux d’ARN et de protéines dans des tumeurs mammaires, nous avons trouvé que HSBP1 est surexprimé dans ce cancer et que cette dérégulation corrèle avec des niveaux supérieurs du complexe WASH. Enfin, les patientes qui surexpriment HSBP1 ont une survie sans métastase faible. Ainsi, toutes nos observations, démontrent que HSBP1 est un facteur d’assemblage conservé qui contrôle le niveau du complexe WASH.

Discussion

Nous avons identifié HSBP1 en tant que partenaire spécifique de CCDC53 grâce à de la protéomique et avons montré son implication dans l’assemblage du complexe WASH. HSBP1 est le premier facteur d’assemblage identifié du complexe WASH. Cette petite protéine de 8,5 kDa est un régulateur négatif du facteur de transcription HSF1 (Satyal et al. 1998). HSBP1 réfère à HSF1 Binding Protein-1. HSF1 est le facteur de transcription majeur activé en réponse à un choc thermique. L’activation de HSF1 implique sa trimérisation par un coiled coil. HSBP1 achève la transcription de HSF1 dans le noyau en dissociant le trimère actif de HSF1 (Satyal et al. 1998). Dans cette étude, nous rapportons une fonction cytosolique de HSBP1 : HSBP1 peut dissocier la forme précurseur trimérique de CCDC53 pour favoriser l’assemblage du complexe WASH. Les deux fonctions de HSBP1 impliquent donc un mécanisme structurel identique. HSPB1 avec son variant plus grand F27 peut être désigné comme un trimère stable de protéines à domaine coiled coil, qui promeut la dissociation d’autres protéines à domaines coiled coil. Les phénotypes que nous rapportons ici suite à la déplétion de HSBP1 n’apparaissent pas corrélé à la terminaison de la transcription du facteur de transcription en réponse au choc thermique.

Figure IV. Modèle d’assemblage du complexe WAVE. Adapté de (Derivery and Gautreau 2010a)

16 L’inactivation de HSBP1 dans ces cellules humaines et chez Dyctostellium amoebae a révélé que sa fonction dans l’assemblage du complexe WASH a été conservée au cours de l’évolution. Dans les deux organismes, certaines sous-unités, mais pas toute, i.e. CCDC53 et WASH, mais pas Strumpellin, SWIP et FAM21, sont déstabilisées suite à la déplétion de HSBP1. Cela concorde avec le fait que seulement CCDC53 est déstabilisé quand WASH est déplété, et seulement WASH est déstabilisé quand CCDC53 est déplété (Jia et al. 2010). De plus, nous avons trouvé que la surexpression de WASH Halotaggé ou en le déstabilisant sélectivement par un traitement à HaloPROTAC3, seule la sous-unité CCDC53 reflète le niveau de WASH. Mis en communs, ces observations suggèrent que le sous-complexe Strumpelin-SWIP-FAM21 existe tel quel dans les cellules. Savoir si ce complexe incomplet représente un intermédiaire stable ou s’il a une fonction à part entière, et donc ses propres partenaires, est une question qu’il faudra résoudre dans de prochaines études.

L’inactivation de HSBP1 entraine des phénotypes similaires à ceux obtenues suite à la déplétion du complexe WASH. Chez Distyostelium amoeba, le Knock-out de WASH induit un recrutement d’Arp2/3 défectueux à la surface des vésicules lysosomales et une exocytose défectueuse de dextrane ingérable (Carnell et al. 2011) [42]. Dans les cellules humaines, la déplétion de WASH conduit à une réduction des niveaux des intégrines … à la surface et un nombre réduit d’adhésions focales (Zech et al. 2011, Duleh and Welch 2012). LA déplétion de WASH réduit également les capacités invasives de cellules cancéreuses (Zech et al. 2011, Monteiro et al. 2013). Nous avons observé tous ces phénotypes dans les cellules et amibes déplétées de HSBP1, confirmant donc que HSBP1 est requis pour l’assemblage d’un complexe WASH fonctionnel.

Les cellules déplétées de HSBP1 déploraient un défaut majeur dans l’établissement de leur polarité qui les empêche de migrer même si elles étaient capables de générer des protrusions membranaires. Ce phénotype, n’avait, à notre connaissance, jamais été rapporté suite à l’inactivation du complexe WASH. Tout de même, le recyclage des intégrines …, à travers leur association à Rab25 (Caswell et al. 2007a), et la route indirecte qu’elles prennent par le Golgi en suivant la voix de signalisation rétrograde (Shafaq-Zadah et al. 2015), est critique pour la persistance de migration. L’établissement et le maintien de la polarité cellulaire dépend également de la contractilité médiée par Rho (Lomakin et al. 2015), qui est elle-même régulée par le

17 recyclage des intégrines (White et al. 2007). Donc même si un défaut de polarité n’avait jamais été rapporté suite à la déplétion du complexe WASH, ce phénotype de HSBP1 est cohérent avec notre compréhension d’un défaut de recyclage des intégrines. De plus, nous avons trouvé que HSBP1 est associé aux centrosomes à la fois dans des cellules humaines et chez Dictyostelium amoebae. Le centrosome représente un marqueur de polarité orienté vers le lamellipode (Euteneuer and Schliwa 1992, Ueda et al. 1997). Il dicte la localisation du Golgi et donc peut réorienter la délivrance des intégrines au front de la cellule et contribuer à la persistance de migration (Théry et al. 2006, Wakida et al. 2010, Shafaq-Zadah et al. 2015). Les centrosomes ont été récemment identifiés comme déterminants dans l’invasion tumorale des cellules (Godinho et al. 2014).

La localisation de HSBP1 aux centrosomes suggère que les centrosomes sont des sites d’assemblage du complexe WASH. Le complexe WASH e été reporté pour nucléer des réseaux d’actine branché au centrosome, en plus de sa localisation majeure à la surface des endosomes (Farina et al. 2016). Ces réseaux d’actine branché aux centrosomes sont impliqués dans l’ancrage des centrosomes au noyau (Obino et al. 2016). Le centrosome pourrait ainsi être un site privilégié pour l’assemblage d’autres complexes multiprotéiques. En effet, Nudel, le facteur d’assemblage idéntifié pour le complexe WAVE, se concentre aussi au centrosome, dans le sens de son activité régulatrice de dynéine, le moteur des microtubules orientés vers leur extrémité négative (Guo et al. 2006). Le centrosome pourrait aussi être un site privilégié pour la dégradation, qui est souvent associé à l’assemblage des complexes multiprotéiques (Derivery and Gautreau 2010a, Wu et al. 2012). En effet, les protéines mal repliées s’accumulent au centrosome à cause du transport dépendant de la dynéine dans de structure appelée aggrésome, lorsque le protéasome est submergé par des substrats à dégrader (Johnston et al. 1998, Wan et al. 2012).

HSBP1 est surexprimé dans des tumeurs mammaires et sa surexpression est associée à un faible pronostic pour les patientes. Nous avons également montré que HSBP1 est surexprimé dans le tissu cancéreux par rapport au tissu normal de la moitié des patientes ayant un cancer du sein. En totale cohérence avec le rôle de HSBP1 dans l’assemblage du complexe WASH, cette surexpression est associée à des niveaux supérieurs des autres sous-unités du complexe WASH. La surexpression du complexe WASH dans les cancers n’avait jusque-là jamais été rapportée, bien que son rôle dans

18 l’invasion tumorale était établit (Zech et al. 2011, Monteiro et al. 2013). FAM21 est la seule sous-unité du complexe WASH dont les niveaux ne corrèlent pas avec ceux de HSBP1 et qui n’est pas surexprimé. FAM21 semble avoir ses fonctions propres, puisqu’il a été décrit pour réguler la transcription par NF-KB dans le noyau indépendamment du complexe WASH (Deng et al. 2015). Chez Dictyostelium amoebae, le phénotype de la déplétion de FAM21 est également assez différent de celui observé lors de la déplétion de WASH dans l’amibe (Park et al. 2013).

En conclusion, notre étude a identifié HSBP1 comme le premier facteur d’assemblage du complexe WASH, caractérisé structurellement l’étape à laquelle il contribue et établit que l’assemblage du complexe WASH médié par HSBP1 fournit un mécanisme à la surexpression du complexe WASH dans les tumeurs.

This page intentionally left blank

Introduction

This page intentionally left blank

22 Chapter I: Branched actin network and its regulation in the cell

Overview

The cytoskeletons are dynamic three-dimensional structures responsible for the mechanical properties and shapes of the cells. These structures help in the movement and stability of the cells. Cytoskeletons are made up of a mesh of long fibers which are polymers of subunits. Based on the fiber type, there are three types of cytoskeletons which are composed of actin filaments, microtubules, and intermediate filaments.

Actin filaments are responsible for the mechanical structure and motility of the cells. Microtubules are involved in processes like chromosome segregation and long- range cargos transportation inside the cells. Intermediate filaments generally provide mechanical support for plasma membrane where it comes in contact with the other cells or with the extracellular matrix. These three cytoskeletal polymers reinforce the cells cytoskeleton and their ability to turn over on a timescale of seconds to minutes that helps the cytoplasm to remodel dynamically.

The actin cytoskeleton is a part of elaborate system that governs the intracellular transport, cellular structure, motility and cytokinesis along with other cytoskeletal fibers. They are also involved in sensing the cell microenvironment, their mechanical properties and the external force that is applied to the cell. These systems are widely regulated by multi-protein complexes where major processes are carried out by assemblies of ten or more protein molecules or multi-protein complexes.

Numerously reported pathologies, spanning from neurological diseases to cancers has attributed to the defects in the regulation of this cytoskeletal system. Hence cytoskeleton proves to be a key player in the cell physiology which in turns translates to organisms physiology.

Figure1. Structure of the actin molecule and actin filament. Modified from Pollard (2016). a, Ribbon diagram of the actin molecule with space-filling ATP. N-amino terminus; C- carboxyl terminus. Numbers 1, 2, 3 and 4 label the four subdomains. b, Actin space-filling model depecting the nucleotide-binding cleft with ATP in situ and barbed-end groove. c, Electron micrograph of a negatively stained actin filament saturated with myosin heads. d, Cartoon of the actin filament showing the position of the pointed and barbed ends.

24 1/ Actin and actin filaments

Actin molecules are one of the most evolutionarily ancient, highly conserved molecules that assemble reversibly into filaments. All eukaryotes have one or more genes for actin and the sequence analysis has established that they vary only by a few amino acids between algae, amoeba, fungi, and animals (Gunning et al. 2015). The available crystal structures of eukaryotic and prokaryotic actins elucidate that they are globular in structure and posses an adenine nucleotide (ATP or ADP) binding site localized at the center (Figure 1a,b). Eukaryotic actin polypeptide which is composed of 375 residues tends to form a structure that is described to have four subdomains, starting with an amino terminus in 1st subdomain continuing to 2nd, 3rd, 4th and back to the 1st domain where the carboxyl terminus ends.

These actin monomers polymerize spontaneously into long, stable filaments under physiological conditions. Studies of X-ray diffraction and electron microscopy on isolated filaments revealed that these filaments are helicoidal polymers having a diameter of 6-7nm (Huxley 1963). As the subunits of the actin filaments point in the same direction, filaments show polarity in structure. The electron micrograph of the negatively stained filaments saturated with myosin heads reveals that they have polarity by forming an arrowhead shaped complex with each helical turn. The arrowheads of the myosin defines the “pointed” and “barbed” ends (Figure 1c,d) (D A Begg, R Rodewald 1978, Pollard 2016).

1.1 The dynamic nature of the actin filament polymerization

The actin filaments are highly dynamic in nature and they are made up of globular actin bound to either ATP or ADP. The assemblies of the monomers are powered by ATP hydrolysis and nucleation of the filament happens spontaneously in vitro. This spontaneous polymerization begins with a lag period that depends strongly on the concentration of the actin monomers. The affinity between two monomers is low but the addition of third one stabilizes the nucleus (Sept and McCammon 2001). This nucleation phase is followed by a rapid elongation phase where the ATP-bound actin addition happens mostly at the barbed end thus leading to the filament elongation. The elongation at the barbed end is highly favorable and is mainly limited by the diffusion of monomers (Drenckhahn and Pollard 1986) whereas the pointed end elongation is one

Figure2. Overview of actin and actin filament regulations. Modified from Pollard (2016). The image depicts the regulation of actin free pool by monomer binding proteins - Profilin and thymosin- filaments by CP proteins, filament severingβ4, elongation by cofilin of and filaments filament by branching formins, cappingby Arp2/3 of complex.

26 order of magnitude slower. This elongation of filament will continue while the ADP- actins disassembly rate at the pointed end is less than that of the ATP-actins binding at the barbed end. Since actin elongation is a bimolecular reaction between monomers and filaments, the speed of filament elongation is directly proportional to the actin concentration and the speed decreases as the ATP-actin monomers get depleted. Ultimately, this leads to an equilibrium state of the system where the rate of association of ATP-actin and the rate of disassociation of ADP-actin are balanced.

1.2 In vivo actin and actin filaments regulation

The in vivo actin behavior is dramatically different from the purified in vitro actin (Pollard 2016). At the given estimated actin concentration found in the cells (50-200 µm), almost all purified actin polymerizes in seconds whereas in the cells, almost 50% are maintained as soluble un-polymerised pool. Moreover filament assembly and turnover happens in a timescale of a few tens of seconds, which is much faster than that observed in vitro. These differences can be explained in the light of actin-binding proteins which virtually regulates all aspects of the actin assembly in the cells (Figure 2). These proteins provide machinery to the cell to produce and dissociate actin filaments when required. This is possible only when the spontaneous polymerization, elongation and dissociation are controlled which in-turn is done by regulating the actin monomer free pool and filaments. Profilin, thymosin and cofilin are key players in

regulating monomeric actin free pool whereas Formin β4and Arp2/3 complex controls the elongation and nucleation of actin filaments. Furthermore capping proteins blocks the filament growth.

1.2.1 In vivo regulators of actin

Profilin binds to the actin monomer in 1:1 stoichiometry and inhibits spontaneous nucleation (Pollard and Cooper 1984). In addition, profilin-bound actin monomer binds to the barbed end as efficient as free actin whereas in the pointed end, it doesn’t bind at all. Soon after the assembly into filament, profilin being a weak interacting partner to ATP-actin that is bound to the barbed-end, dissociates from the filament. This dissociation of profilin releases the barbed-end for further elongation. The thymosin blocks all the assembly reaction by competing with

the profilin to bindβ4, on to the actin other monomer. hand Hence, from these, a fundamental understanding

Figure3. Formin structure and activity. Modified from Chesarone et al. (2009a). a, crystal structures and a schematic of formin domains. b, Processive elongation by formin. c, Proposed model of regulatory points in the formin activity cycle.

28 can be achieved wherein the presence of profilin, actin polymerization is restricted to the barbed ends and rest of the ATP-actin monomers are inactive due to the binding of the cells, a way to precisely control actin thymosinpolymerisation β4. to Hence occur this only provides at the barbed ends. When the barbed ends become available, profilin -actin towards the barbed end due its competitive bindingshuttles nature thymosin (Pantaloni β4 bound and ATPCarlier 1993). In a given physiological condition,low concentration the actin of monomer free actin. is either bound to profilin or thymosin β4 thus leaving a

On the other hand, availability of barbed end filament is controlled by the cell with the assistance of regulatory protein called capping protein which contains two homologous subunits (Edwards et al. 2014). If the barbed ends are free in the cytoplasm, they would, α rapidlyand β grow in an uncontrolled fashion which in-turn would deplete the pool of actin monomers. Thus capping proteins controls and regulates the availability of free barbed ends by blocking the filaments barbed ends. The capping - -subunits that tightly proteinbinds to isthe a heterodimeractin filament consisting barbed ends. structurally Its presence similar in eukaryotic α and β cells in micromolar concentration makes it bind to the barbed ends in seconds and remains capped for about 30 minutes.

Efficient disassembly of actin filament is necessary for the actin monomer turnover for new filament assembly. This process is regulated by the cofilin which severs the actin filament by binding to the lateral side of the filaments, with a high affinity towards the ADP-actin rather than the ATP-actin or ADP-Pi actins (Cao et al. 2006). Nevertheless, weak binding of cofilin to ADP-Pi subunits in filaments promotes -phosphate, producing ADP-actin polymers in seconds rather than dissociationin minutes of(Blanchoin the γ and Pollard 1999), thus making this turnover timescale reasonable for the filaments in the cells.

1.2.2 Regulators of actin filament nucleation and elongation

As the above described proteins tightly regulate the free actin pool in the cells, spontaneous filament nucleation is intrinsically unfavorable. Also, cells rely on specific regulatory proteins to perform the actin filament polymerization in a controlled manner. This process is attributed to nucleation and elongation factors like Arp2/3

29 complex and Formins. Arp2/3 complex is used by the cell to produce branched actin whereas formins initiate linear filaments growth and elongation. A detailed description of the Arp2/3 complex is made in following sub-section 2.

Formins are implicated in a variety of cellular processes like focal adhesion, stress fibers generation, filopodia development, and contractile ring formation in cytokinesis (Goode and Eck 2007). Formins also associate with the plasma membrane or intracellular membranes such as endoplasmic reticulum. Most formins nucleate actin filaments by stabilizing actin dimers (Pring et al. 2003) and then remain bound to filaments as they elongate, specifically to the barbed end. Typically, formins are characterized by formin homology 1 (FH1) domain and formin homology 2 (FH2) domain. Formins are auto-inhibited by the interaction between diaphanous auto- inhibitory domain (DAD) near the N-terminus and carboxyl DAD-interacting domain (DID) (Figure 3a).

Formins are recruited to the plasma membrane and activated by Rho GTPases in the cell. Rho protein binds to the DID domain and displaces the DAD domain that leads to the formin FH2 domain activation. This facilitates the actin filament nucleus formation. Formin mediated actin filament elongation by incorporating just the actin monomer is a slow process as it is limited by the diffusion of monomers at that specific time and place in the cells. This limitation is surmounted by the FH1 domain which binds to profilin through various binding sites. As profilin is bound to almost all actin monomer in the cells, it concentrates multiple profilin-actin complexes near the filaments barbed end. Thus, FH1 transfers actin very rapidly to the elongating barbed end and therefore, filament growth occurs 5 to 6 times faster in the presence of profilin (Figure 3b,c). The elongation is processive as formin displaces itself to the newly added actins of the filament. This process is also called “stepping” where formin steps onto the newly added actin of the filaments for thousands of cycles without failure (Pollard 2016). This kind of polymerase activity inhibits the barbed end capping by capping proteins thus allowing the actin filaments to grow quickly.

Formins assemble most of the in vivo actin arrays that consist of short filaments (0.3-2.3 microns), whereas purified formins remain bound to and elongate the barbed ends in vitro for several minutes without dissociation (Chesarone et al. 2009a). These

30 observations suggest that formins should be regulated to prevent uncontrolled filament elongation. This could be achieved by inactivation or formin displacement from the barbed ends (Figure 3c). Such an displacement factor Bud14 is identified in S. cerevisiae for its displacement effect on the Bnr1 formin which is stably anchored to the cell membrane (Chesarone et al. 2009b). After the Bud14 mediated displacement of Bnr1 formin from the actin filament occurs, the Bnr1 formin remains associated with the cell membrane where it must be recycled for a new round of actin assembly. There are also other formins which are dynamically recruited to the membrane and released after a brief period of activity. These activities of different formins are regulated in a distinct temporal manner and their mechanisms are still elusive.

Hence, at a given time and space, formins produce short actin filaments in cells which are linear, not branched which in-turn contribute to the protrusive or tensile forces at a specific region, example - Filopodia formation during migration and contractile ring formation during cytokinesis. Alongside cells depend greately on branched actin networks which generate a uniform pushing force helping the cell to remodel its membrane.

2/ Branched actin network by the Arp2/3 complex

Arp2/3 complex is the only multi-protein complex that generates branched actin networks in the cell. Branched actin network is major cytoskeleton structure that plays crucial roles in the cell physiology. These networks are created as a result of new filament nucleation on the side of a pre-existing actin filament thus making a branched structure forming a mesh of actin filaments. These mesh-like structures help the cells to remodel the membrane due to the pushing force it generates. This can be observed at the surface of clathrin coated pits during endocytosis, at the endosomal surface during fission and on the cell lamellipodia. Hence branched actin networks play important roles in cellular processes such as cell locomotion, phagocytosis and intercellular vesicle motility.

Figure4. Arp2/3 complex localises at the junction of actin branch. Modified from Egile et al. (2005). a, Arp2/3 crystal structure b, Electron micrograph of branched actin filament showing the localisation of Arp2/3 complex at the branch junction.

32 2.1. Characteristics of the Arp2/3 complex

The first identification of Arp2/3 complex was made in Acanthamoeba using profilin affinity chromatography where it was characterized as the multi-protein complex that contains actin-related proteins, Arp2 and Arp3 (Machesky et al. 1994). Arps are characterized to share between 17% and 52% of sequence identity with actin and are structurally similar to actin monomer (Robinson et al. 2001; Muller et al. 2005; Pollard 2016). Arp2/3 complex is composed of two Arps – Arp2 and Arp3 along with other 5 subunits – ArpC1, ArpC2, ArpC3, ArpC4, and ArpC5; making it a 7 subunit complex, with a total mass of about 250 kDa (Figure 4a). This complex has been found in yeast (Winter et al. 1997) as well as in the vertebrates (Welch et al. 1997a). All the 7 subunits of Arp2/3 complex are well conserved as the set of genes encoding these subunits were found conserved across the genome of the eukaryotes.

2.2. Molecular function of the Arp2/3 complex

First observation of the Arp2/3 complex as an actin-nucleating machine was done by Welch and colleagues, where they documented the actin comet tail formation mediated by Arp2/3 at the surface of Listeria to propel the bacterium through cytoplasm of the infected cells (Welch et al. 1997b). Thus, it quickly emerged as a major actin nucleator in cells which was later shown to be localized in the cell at the actin-rich cortex found in the lamellipodial edge (Machesky et al. 1997, Welch et al. 1997a). The main attribute of Arp2/3 complex is to form branched actin network in cells by nucleating a new filament (daughter filament) from the side of pre-existing filament (mother filament), forming an angle of ~70 degrees (Mullins et al. 1998, Blanchoin et al. 2000). Thus the Arp2/3 complex anchors the pointed end of daughter filament on the mother filament while its barbed end keeps elongating (Egile et al. 2005). Hence, structurally Arp2/3 is localized in vivo and in vitro at the level of the branches’ Y- junction which also resembles ‘twigs on a bush’ (Figure 4b).

2.3. Arp2/3 complex activation and nucleation mechanism

The Arp2/3 complex exhibits 2 conformations – an inactive conformation, which is revealed by crystal structure of the complex where Arp2, Arp3 are maintained far apart (Robinson et al. 2001). In the active confirmation, Arp2/3 complex is at the

Figure5. Conformational changes of Arp2/3 complex and involved regulators. Modified from Molinie and Gautreau (2017). The scheme depicts the Arp2/3 conformational changes that occur due to the activation by NPFs that leads to actin branch formation. This process of activation is antagonised by inhibitors like Arpin, PICK1 and Gadkin. Established actin branches are destabilised by the factors like GMF and Coronin that releases Arp2/3 complex from the branch.

Figure6. Proposed pathway of actin branch formation by the Arp2/3 complex. Modified from Pollard and Cooper (2009).

34 branching junction having Arp2 and Arp3 close to each other becoming first two subunits of the daughter filament (Volkmann et al. 2001, Rouiller et al. 2008). Hence, Arp2/3 complex has to undergo a large conformation change from an intrinsically inactive form to active form where it initiates new filament formation. This new filament formation occurs as the closed Arp2 and Arp3 subunits mimic actin dimer, to which the new actin monomer binds forming a structure similar to actin trimer. This trimer acts as a stable nucleus for the filaments rapid elongation. This activation of the Arp2/3 complex requires binding to a pre-existing actin filament and a Nucleation Promoting Factor – NPF (Figure 5) (Welch et al. 1998; Higgs et al. 1999; Machesky et al. 1999).

There are several Arp2/3 activating NPFs identified till date. These NPFs are characterized to have VCA domain - Verprolin homology domain (this also known as WH2), Cofilin-homology domain (also known as central domain) and Acidic domain. V- motif of the VCA domain binds to monomeric actin whereas the CA motif binds to Arp2/3 complex. The activation is initiated by the binding of Arp2/3 complex to pre- existing actin filament and by an NPF through CA-motif of its VCA domain. This binding should bring Arp2 and Arp3 subunits close to each other that is facilitated by the dynamic conformational changes of all seven Arp2/3 complex subunits. The V-motif of NPFs VCA domain brings monomeric actin to the closed conformation of Arp2 and Arp3 thus forming a structure that mimics actin trimer nucleus to initiate rapid branch filament elongation.

The sequence of these events during the activation mechanism was suggested by measuring the different association/dissociation constants between the involved species. This also helped in proposing a kinetic model of branch formation (Beltzner and Pollard 2007). In this model, the first step involves NPF binding to actin monomer and the second favorable step where actin-NFP binds to Arp2/3 complex. At the third step, actin-NPF-Arp2/3 intermediate binds to the side of a pre-existing actin filament. As the last step, the activation of Arp2/3 complex occurs which leads to the nucleation of new actin filament branch (Figure 6).

Figure7. Domain organization of Arp2/3 activatory and inhibitory proteins. Modified from Molinie and Gautreau (2017). In this scheme, the respective localization of these regulators is indicated. For each localization, one can notice that there is usually a pair of activator and inhibitor.

36 2.4. Inhibitors of Arp2/3 complex and debranching factors

Branched actin networks formed due to Arp2/3 activation should be regulated and debranched at a certain point of time by the cell so as to dynamically modulate these cytoskeletal structures for its cellular processes. It is also mandatory for the cell to regulate these cytoskeletal structures to have the turnover of involved protein species like actin and Arp2/3 complex. This turnover in-turn helps the cell to initiate new actin branched networks.

Regulation of Arp2/3 complex by means of its inhibition from getting activated was performed majorly by three recently found proteins. These proteins are – Arpin, Gadkin and PICK1 (Figure 5). The common mechanism of their Arp2/3 inhibition is by antagonizing the NPFs. They all are characterized to bind towards the Arp2/3 complex via their acidic motif, similar to the motif found in NPFs (Figure 7) (Rocca et al. 2008, Maritzen et al. 2012, Dang et al. 2013).

These Arp2/3 inhibitors are found in specific regions of the cell. The basic assumption is that these proteins should diffuse at the specific region to control local signaling pathway that mediates Arp2/3 branched actin networks. Arpin protein is found to be localized at the migrating tip – lamellipodia and the Arp2/3 complex is sequestered into an inactive form by Arpin at this location (Dang et al. 2013). Gadkin, also known as -BAR localizes at the surface of endosomes and is observed to regulate the trans-Golgi γnetwork-endosomal traffic by getting into a complex with kinesin KIF5 and AP-1 adaptor of clathrin (Schmidt et al. 2009). In the absence of Gadkin, it is observed that the Arp2/3 mediated actin patches on the surface of endosomes are increased in size (Schachtner et al. 2015). Hence, Gadkin is expected to antagonize the NPF that activates Arp2/3 at the endosomes surface but this mechanism is still elusive. PICK1 is a protein that is characterized to contain PDZ domain and a BAR domain. It found to be enriched in brain tissues (Li et al. 2016). It is shown to inactivate Arp2/3 at the surface of clathrin-coated pits as it regulates the AMPA receptor endocytosis (Rocca et al. 2008).

On the other hand, debranching of actin network is essential for its spontaneous remodeling. During the branching, ATP hydrolysis by Arp2 occurs and is shown to be involved in debranching by destabilizing the Arp2/3 branch (Le Clainche et al. 2003,

37 Martin et al. 2006). This means of debranching is slow and hence this process is enhanced by the activity of several proteins such as Cofilin (Chan et al. 2009), Coronin (Cai et al. 2007) and GMF (Gandhi et al. 2010). Thus by facilitating efficient debranching, these factors assist actin branch turnover in cells (Figure 5).

3/Nucleation Promoting Factors (NPFs)

Various NPF proteins play a key role in activating the Arp2/3 complex at different region of the cell. Inactive Arp2/3 complex is diffused through cytosol of the cell. NPFs function is to activate the Arp2/3 complex at a specific region of the cell as various NPFs are found to be bound to specific cell membranes. Hence, various NPFs help in the local activation of Arp2/3 complex. In nature, NPFs are auto-inhibited with VCA domain being masked by intramolecular interactions or they exist as a multi- protein complex with the VCA domain being embedded in its subunits. Both scenarios exist to make NPFs intrinsically inactive in the cell. Upon signaling events, different subunits get modified to expose the VCA domain which activates Arp2/3 complex. Thus, NPFs provide temporal and spatial Arp2/3 complex regulation in the cell.

Complete VCA domain carrying NPFs can be classified as four families (Figure 7). They all share the C-terminal VCA domain which enables the activation of Arp2/3 complex and the N-terminal domain which defines different families of NPFs. Also, this N-terminus of the NPF is the key player in structuring the multi-protein complexes that contains it. The first characterized NPFs were WASP ("Wiskott-Aldrich Syndrom Protein")(Machesky and Insall 1998) and neural WASP (N-WASP) (Miki et al. 1996), later the WAVE ("WASP family Verprolin Homologous Protein" also known as SCAR) (Miki et al. 1998b) and its three isoforms were discovered. Over the last decade, WASH proteins (WASP and SCAR homologue) (Linardopoulou et al. 2007), WHAMM ("WASP homologous with actin, membranes and microtubules") (Campellone et al. 2008) and JMY (“Junction-mediating regulatory protein") (Zuchero et al. 2009) has been identified.

3.1 N-WASP protein

WASP was first identified as a mutated gene in the Wiskott-Aldrich syndrome, also known as WAS (Derry et al. 1994). The human genome contains two paralogous WASP protein – the ubiquitous N-WASP and the hematopoietic WASP. The WAS

38 syndrome is characterized by immune deficiency, low platelet count and eczema which are the causal effects of host cells migratory defects, phagocytosis defects and T-cell signaling defects due to WASP mutation. Due to the homology of N-WASP and WASP, a detailed description of N-WASP suffices its understanding which is done below.

3.1.1 Function of the N-WASP in cells

The N-WASP involved signaling pathway leads to endocytosis (Merrifield et al. 2004, Benesch et al. 2005). This internalization is mediated by Arp2/3 branch actin polymerization which in turn is activated by N-WASP. Branched actin network also facilitates the fission of endocytosed vesicles from the plasma membrane. In addition to the fission role, N-WASP helps the endosome to move away from the membrane by forming an actin comet tail on the endosome which gives the propulsion motion (Benesch et al. 2002). The N-WASP is also characterized to have a role in the formation of podosomes and invadopodia which are protrusive structures formed by macrophages and cancer cells respectively (Mizutani et al. 2002, Yamaguchi et al. 2005).

3.1.2 Molecular characteristics of the N-WASP

The N-WASP is composed of N-terminal WH1 domain (Wasp Homology Domain 1) followed by a basic region (B), CRIB domain (Cdc42/Rac – Interactive Binding region) also known as GBD domain (GTPase Binding Domain), PRD domain (Proline- Rich Domain) ending with a VCA domain at the C-terminal (Figure 8).

Cell membrane component PIP2 interacts with the basic region whereas small Cdc42 GTPase interacts with the CRIB domain (Rohatgi et al. 1999, 2000, Higgs and Pollard 2000). In its native form, N-WASP is stably associated with the WASP- interacting protein (WIP) (Ho et al. 2001, Kato et al. 2002, Aspenström 2002, 2004). It is well characterized that stable interactions of N-WASP with WIP is facilitated by the WH- 1 domain. This domain also maintains the stability of N-WASP (Sawa and Takenawa 2006, Krzewski et al. 2006). Also, WH1 domain is the region that is frequently mutated in WAS patients (Stewart et al. 1999, Imai et al. 2003). These observations emphasizes the physiological importance of N-WASP-WIP interaction for the normal cell functioning.

Figure8. N-WASP architecture and activation. Modified from Campellone and Welch (2010).

40 3.1.3 Activation of the N-WASP

In vivo N-WASP prefers to be in a conformational folding state where C-terminal VCA tail , N-terminal CRIB domain and basic region (B) are close together by intramolecular interactions (Rohatgi et al. 1999, Kim et al. 2000, Prehoda et al. 2000). Due to this interaction, the VCA domain is masked. Apart from this, the WIP interaction also promotes the inhibition of N-WASP (Ho et al. 2004).

In the activation of N-WASP, the CRIB domain also known as GTPase binding domain (GBD domain) plays a crucial role as it binds mutually exclusive with Cdc42 GTPase (Rohatgi et al. 1999, 2000, Kim et al. 2000, Buck et al. 2001). This binding releases the VCA domain which readily activates the Arp2/3 complex (Figure 8). The basic region (B) also contributes to the self-inactivation of N-WASP. Binding of PIP2 to this basic region works in synergy with Cdc42 for N-WASP activation (Miki et al. 1998a, Kim et al. 2000, Prehoda et al. 2000). The Toca-1 protein and SH3 domain proteins have also been demonstrated to be involved in N-WASP activation (Carlier et al. 2000, Rohatgi et al. 2001, Kowalski et al. 2005). WASP protein is also activated by phosphorylation. Two phosphorylation sites in its VCA domain at serines 483 and 484 are identified. Phosphorylation of these residues increases the affinity of the VCA domain to the Arp2/3 complex by 7-fold and this allows WASP protein to perform in vitro actin polymerization efficiently (Cory et al. 2003). Also, it has been shown that Rho GTPase Cdc42 and the Src family kinase Lck cooperation enhances WASP activation (Torres and Rosen 2006).

3.2 The WAVE complex

The WASP-family, Verprolin-homologous protein 1 – WAVE1 was identified by sequence homology with WASP protein and also found to regulate actin cytoskeleton through Arp2/3 complex (Machesky and Insall 1998, Miki et al. 1998b). In cells, WAVE1 has its homologs WAVE2 and WAVE3 (Suetsugu et al. 1999). WAVE2 is ubiquitously expressed whereas WAVE1 and WAVE3 expressions are tissue specific (Sossey-Alaoui et al. 2003).

Figure9. Native WAVE complex architecture. Modified from Derivery and Gautreau (2010a). The native WAVE complex is inactive in the cytosol and is auto-inhibited by intermolecular interactions between the subunits of complex that mask WAVE protein VCA tail.

42 3.2.1 Function of the WAVE complex in cells

The WAVE complex is primarily characterized to produce branched actin networks at the migrating tip of the cell hence playing an essential role in cell motility. Ubiquitously expressed WAVE2 regulates the formation of lamellipodia whereas WAVE1 promotes the formation of "dorsal ruffles" and stabilizes lamellipodia (Yan et al. 2003, Yamazaki et al. 2003, 2005). Explicit functioning of WAVE3 is not clear but it has been shown to be involved in lamellipodia formation (Sossey-Alaoui et al. 2007).

Recent studies show that plasma membrane protrusions are coordinated with the intracellular traffic. Clathrin heavy chain has been demonstrated to interact with the WAVE complex and promote its activation at the lamellipodium tip in an endocytosis- independent manner (Gautier et al. 2011). It is also shown that Sra1 subunit of the WAVE complex has a major role to play in the biogenesis of transporters which involves a transport pathway between Golgi apparatus and endosomes (TGN-to-endosomes pathway)(Anitei and Hoflack 2011). The Sra1 interacts with clathrin and adaptor Ap1 allowing the local recruitment of Rac that activates N-WASP leading to the branched actin networks necessary for tubulation of the carriers.

The lamellipodia generated by WAVE2 is also shown to facilitate the cell-cell contacts of epithelial cells and in the formation of adherent junctions mediated by cadherin (Yamazaki et al. 2007). Also it was shown to favor the separation of daughter cell during cytokinesis (King et al. 2010).

3.2.2 Molecular characteristics of the WAVE complex

The WAVE protein family is characterized by domain constituency of N-terminal SHD (“SCAR homology domain”), basic domain (B), PRD domain (“Proline-rich domain”) followed by C-terminal VCA domain. All the WAVE proteins are embedded into a multiprotein complex which is composed of ABI (“ABL interactor 1” or its paralogs Abi1 and Abi3), CYFIP/SRA (Cytoplasmic FMR1-interacting protein or Specifically Rac1- associated protein), NAP1 (NCK-associated protein 1) and BRK1 (also known as HSPC300) making them a 5 subunit complex of 400 kDa mass (Eden et al. 2002, Gautreau et al. 2004, Innocenti et al. 2004). Due to the number of paralogous subunits excepting BRK1, there are different isoforms of the WAVE complex (Derivery and Gautreau 2010a). To increase the complexity, some subunits like ABI1 are alternatively

Figure10. Activation cycle of the WAVE complex. Modified from Derivery and Gautreau (2010a). The intrinsically inactive WAVE complex is recruited to the membrane by different mechanisms: clathrin (CHC, "clathrin heavy chain") interacts with the WAVE complex and brings it to the plasma membrane; subunit Sra1 interacts with Rac; WAVE subunit interacts with PIP3; The proline-rich domain of WAVE interacts with the IRSp53 protein. All these interactions contribute to the activation of WAVE complex by releasing its VCA domain that activates Arp2/3 complex. The dimerization of IRSp53 containing an inverted BAR domain helps to bend and protrude the plasma membrane. This protrusion is enhanced by the branch actin polymerisation to form lamellipodial projections during migration.

44 spliced. Even though there is the existence of isoforms of WAVE complex, the ubiquitous complex of WAVE is composed of WAVE2, Brk1, Abi1, Sra1 and Nap1. The native WAVE complex is intrinsically inactive in cells similar to WASP proteins thus facilitating a temporal and spatial regulation of Arp2/3 complex (Derivery et al. 2009a) (Figure 9). In recent years, there are several observations regarding the WAVE complex and their subunit regulation which makes it clear that the WAVE complex undergoes a tightly regulated assembly pathway. A detailed description of the WAVE complex assembly has been made in chapter III.

3.2.3 Activation of the WAVE complex

The WAVE complex is activated at the plasma membrane of cells by several mechanisms which release the masked VCA domain to bring Arp2/3 and activate the branching of actin fibers. The WAVE complex was classically characterized to be an effector of Rac GTPases but in recent years, it has been shown that there are other proteins which interact with the WAVE complex to activate it. Also numerous phosphorylations are important to control the WAVE complex activity.

3.2.3.1 Activation of the WAVE complex by Rac

Lamellipodia formation is controlled by the WAVE complex regulation through Rac GTPase (Miki et al. 1998b). However, the WAVE exists as a stable complex and WAVE subunit does not contain any GBD domain. This helped to realize that the actual binder of Rac is Sra subunit because years before this particular subunit was independently characterized as Rac effector (Kobayashi et al. 1998, Kunda et al. 2003, Derivery and Gautreau 2010a, Chen et al. 2010).

3.2.3.2 Activation of the WAVE complex by molecular interactions

The WAVE2 subunit has been characterized to interact with the product of PI3K kinase, PIP3 (Oikawa et al. 2004) and with SH3 domain of IRSp53 protein ("Insulin receptor substrate protein of 53 kDa") via its PRD domain (Takenawa et al. 2000, Suetsugu et al. 2006, Scita et al. 2008, Abou-Kheir et al. 2008). Clathrin heavy chain (CHC) interacts with Sra subunit and helps with the recruitment of WAVE complex to the plasma membrane (Anitei and Hoflack 2011). The inactivation of CHC leads to the loss of localization of WAVE complex to plasma membrane which leads to protrusion

45 defects and a reduction in migration speed (Gautier et al. 2011). Recruitment of the WAVE complex to plasma membrane facilitates its interaction with different partners, Rac, PIP3, and IRSp53 thus leading to the unmasking of the WAVE2 VCA domain. The unmasking of VCA enables the complex to recruit and activate Arp2/3 complex (Figure 10). Furthermore, the IRSp53 contains a modified BAR domain which dimerizes itself as inverted curvature against the plasma membrane and promotes the formation of outward membrane protrusions by branched actin polymerization .

3.2.3.3 Activation of the WAVE complex by phosphorylations

Once the WAVE complex has been recruited to the plasma membrane, ABL- mediated phosphorylation of WAVE and ABI proteins on tyrosine residues may stabilize the active conformation of the WAVE complex. The tyrosine kinase ABL localizes at the lamellipodial edge and binds to the WAVE via SH3 domain. ABL tyrosine Kinase phosphorylation of WAVE1 and WAVE3 at Tyr151 and WAVE2 at Tyr150 exposes the masked VCA domain (Krause and Gautreau 2014). Thus, the motile cells spontaneously regulate the lamellipodial extensions that are necessary for rapid response to the changes in their environment.

3.3 The WASH complex

The WASH (“Wiskott-Aldrich Syndrome Protein and SCAR Homolog”) gene was discovered in the year 2007 and is found in all eukaryotes except yeast and Arabidopsis (Linardopoulou et al. 2007, Derivery and Gautreau 2010b). In human genome, WASH family genes are present in variable number, from 15 to 20 depending on the individual. This is the consequence to the sub-telomeric location of WASH gene, as this region is prone to recombination. Even though the WASH genes were duplicated, in practice it is possible to deplete all WASH proteins with single siRNA sequence like any other gene product. In mice and rats, only a single copy of WASH gene is found. Like WAVE complex, WASH also exists in a multi-protein complex with VCA domain masked by its subunits making it intrinsically inactive (Jia et al. 2010).

3.3.1 Function of the WASH complex in cells

Primarily, WASH protein was characterized to localize at the endosomes surface, predominantly on sorting and recycling endosomes (Derivery et al. 2009b, Gomez and

46 Billadeau 2009). On the surface of endosomes, WASH is recruited through its subunit FAM21 which contains multiple binding sites for the retromer complex (Harbour et al. 2012, Jia et al. 2012, Helfer et al. 2013). Interestingly, retromer is a well characterized and conserved cargo-protein recognizing complex which is shown to orchestrate multiple cargo-sorting events within the endosomal network (Burd and Cullen 2014). The sorting of endosomal cargoes by retromer are either destined to follow the retrograde route to trans-Golgi network or are recycled to the plasma membrane (Seaman et al. 2013).

In cells, WASH complex and retromer do not cover the whole surface of endosomes but rather they are confined to micro-domains, whose area is controlled by actin polymerization itself (Derivery et al. 2012). Depletion of WASH by interfering RNAs leads to actin network disappearance on the endosomes, suggesting that WASH recruits and activates the Arp2/3 complex in this region (Derivery et al. 2009b, Gautreau et al. 2014). In the cultured cells, loss of WASH results in the dis-functioning of the endosomal system leading to tubulation of endosomes and in some cases, enlarged endosomal compartments which are strong phenotypes of defective fission and missorting of endosomes (Derivery and Gautreau 2010b, Duleh and Welch 2010, Gomez et al. 2012, Piotrowski et al. 2013). Hence, the WASH complex performs endosomal fission and sorting, making it one of the key players in receptor trafficking in cells. There are numerous neurodegenerative diseases that have been linked to the WASH complex. As well, the WASH complex has been shown to be responsible for the invasiveness of cancer cells. The detailed description of WASH functions and abnormalities due to its defects are done in chapter II.

Recently, Drosophila WASH has been characterized to localize in the nucleus by the interaction with lamin where it plays a key role in the global nuclear organization (Verboon et al. 2015b). Following this study, WASH complex subunit FAM21 has been characterized to participate in NF- B-dependent gene regulation in pancreatic cancer cells (Deng et al. 2015). As NF- B isκ a key apoptotic regulator, destabilization of FAM21 sensitizes pancreatic cancer cellsκ to anticancer drug-induced apoptosis. Hence, WASH complex potential role in nuclear activities are yet to unfold. It is tempting to speculate the link between WASH nuclear activity to various identified neurological pathologies

Figure11. WASH and WAVE complexes are structurally similar. Inspired from Jia et al. (2010), Derivery and Gautreau (2010a) and Gautreau et al.(2014). Profile-against- profile search in the HHPred package identified distant homology between the WASH and WAVE complexes, except FAM21 and Abi subunit.

48 and cancer where some of their effects might reside in the nucleus and not on endosomes.

Interestingly, WASH can also bind to tubulin (Derivery et al. 2009b) and recently it has been characterized to localize at the centrosome which is a well-known microtubule organizing center (MTOC) and it plays a role in Arp2/3 mediated branched actin network formation at the centrosome (Farina et al. 2016). This actin nucleation at the centrosome is required to regulate lymphocyte polarity (Obino et al. 2016). Hence, WASH complex proves to be a versatile NPF thus performing regulatory functions in different cell compartments making a putative bridge between microtubule network and actin network.

3.3.2 Molecular characteristics of the WASH complex

The WASH NPF is a stable multi-protein complex, distantly related to the WAVE complex with one-to-one correspondence of the 5 subunits (Figure 11). The WASH protein of the WASH complex is characterized to possess N-terminal WAHD1 domain ("WASH homology domain 1") followed by a proline-rich domain (PRD) and then a C- terminal VCA domain. In the complex, WASH protein is embedded around other interacting subunits – FAM21 (Family with sequence similarity 21), Strumpellin, SWIP (Strumpellin and WASH Interacting Protein), and CCDC53 (Coiled-coil Domain Containing protein 53) (Derivery et al. 2009b, Jia et al. 2010). Most of the subunits are encoded by single gene and thus the WASH complex is less diversified than the WAVE complex (Derivery and Gautreau 2010b). Interestingly, in contrast to the WAVE complex, WASH complex additionally recruits a pre-existing complex – Capping Proteins (CP).

CP proteins are composed of heterodimer

actin filaments elongation and exist on their own ofin CapZα the cytosol. and CapZβ The recruitment which blocks of theCP complex to the WASH complex happens by the interaction that takes place via CP- interacting (CPI) motif in the C-terminal domain of FAM21 subunit (Derivery and Gautreau 2010b, Hernandez-Valladares et al. 2010). Interaction of CP dimer towards the WASH protein has shown to be important for the WASH function in amoeba Dictyostelium discoideum (Park et al. 2013) but still, it’s unclear why these two activities

Figure12. Regulation of WASH within a stable multiprotein complex and associated activities. Modified from Gautreau et al. (2014). Long, unstructured arm of FAM21 allows it to interact with retromer complex on the endosomes. Retromer also recruits E3 ubiquitin ligase, TRIM27-MAGEL2 that poly-ubiquitinates the flexible region of WASH leading to the VCA domain exposure. Thus, the Arp2/3 complex is activated and branch actin network formation is initiated on the surface of endosomes.

50 of actin nucleation and actin capping has to be coordinated in the same multi-protein complex.

Since, WAVE and WASH complex show structural similarities, like the WAVE complex, WASH complex stability also depends on its subunits. Degradation of one of the subunits including the CapZ heterodimer will lead to destabilization of the entire complex (Derivery et al. 2009b). To another end, like WAVE protein, WASH protein can perform in vitro Arp2/3 mediated actin nucleation but in the native complexed form, WASH is inactive similar to WAVE complex. Thus, these two distantly related complexes should be controlled by analogous structurally related mechanisms and undergo similar kind of complex assemblies that leads to intrinsically inactive complexes.

3.3.3. Regulation of the WASH complex

As the VCA domain of the WASH protein is masked by the subunits of its complex similar to WAVE complex, the WASH complex has to undergo activation mechanism which leads to the exposure of VCA domain. The activation of the WASH complex was studied to happen via ubiquitination (Hao et al. 2013). It was shown that WASH undergoes K63-linked poly-ubiquitination at K220 residue inside its flexible region. Usually, proteins that are targeted for degradation are marked by K48-linked ubiquitination process (Bochtler et al. 1999). However in contrast, K63-linked ubiquitination typically acts as a signalling event to modify function, such as altering protein-protein interactions or protein conformations (Sun and Chen 2004).

The replacement of endogenous WASH with a K220R mutant that cannot be ubiquitinated prevents the endosomal branched actin formation. Similarly, treatment of the purified wild-type WASH complex with AMSH enzyme (Associated Molecule with the SH3-domain of STAM), a K63 deubiquitinase inhibits its in vitro ability to nucleate actin filaments via Arp2/3 activation. Interestingly, the WASH region which harbours K220 falls in the putative ‘meander’ region that is predicted to regulate the VCA exposure.

The K63-linked ubiquitination of WASH protein is driven by E3 ubiquitin ligase TRIM27. The functionality of TRIM27 is enhanced by MAGE-L2 which is recruited to endosomes through an interaction with the VPS35 subunit of retromer complex. It is to be noted that retromer, specifically VPS35 subunit also recruits the WASH complex at

Figure13. WHAMM and JMY associates with Golgi. Modified from Campellone et al. (2008), Schlüter et al. (2014). a, Cos-7 cells stained with antibodies to WHAMM, Golgi marker GM130 and DAPI. b, Cos-7 cells expressing GFP-JMY were fixed and stained for the Golgi markers ERGIC53 and GM130. Scale bar: 10 µm.

52 the surface of endosomes via multiple binding sites present in the FAM21 long arm (Harbour et al. 2012, Jia et al. 2012, Derivery et al. 2012). Therefore, retromer fetches both WASH complex and its major regulator thereby connecting two types of machinery on the surface of endosomes which leads to the WASH complex activation (Figure 12).

Recently, in addition to this ubiquitination mechanism of WASH complex activation, endosomal PI4P also shows to play a role in recruitment and activation of the WASH complex (Dong et al. 2016). However, signalling pathways that regulate the activity of WASH complex are not well known. Studies have also shown that in Drosophila, WASH protein can function independently by not being in a complexed form and act as an effector of the small Rho1 GTPase (Verboon et al. 2015a). However, previous studies showed that in Drosophila, the presence of Rho didn’t increase the WASH complex NPF activity (Liu et al. 2009).

3.4 WHAMM/JMY

WHAMM and JMY are the two homologous proteins that are found only in vertebrates (Veltman and Insall 2010). Their domain organization is similar to that of other NPFs containing a family specific N-terminal domain, a proline-rich region and a C-terminal VCA domain (Campellone et al. 2008, Zuchero et al. 2009). Specifically, JMY has been characterized to possess three WH2 motifs that bind to the three actin monomers. Thus, unlike all other NPFs, interaction with three actin monomers allows it to nucleate filament alone in an Arp2/3 independent manner (Zuchero et al. 2009). WHAMM has the N-terminal WMD domain (WHAMM Membrane-interacting Domain) required for its interaction with the membranes and a central "coiled-coil" (CC) domain that interacts with the microtubules. Unlike other NPFs, WHAMM is not self-inhibited and does not appear to be a part of a multi-protein complex. Recent studies have revealed that, WHAMM binds to microtubules. This interaction with MT makes WHAMM VCA domain to be masked while its N-terminal domain is exposed and this allows it to interact with the vesicles to remodel them (Shen et al. 2012). Thus, the WHAMM activity is directly controlled by its interaction with microtubules.

WHAMM is well expressed in epithelial tissue such as colon, kidney, and lungs (Campellone et al. 2008) whereas JMY is enriched in the brain, testis and lymphoid (Firat-Karalar et al. 2011, Adighibe et al. 2014). Both WHAMM and JMY are localized at

53 the Golgi, specifically WHAMM at cis-Golgi (Campellone et al. 2008) and JMY at the trans-Golgi (Schlüter et al. 2014) (Figure 13). WHAMM regulates the transport of vesicles from the endoplasmic reticulum to the Golgi apparatus where it interacts with the actin cytoskeleton and microtubules to control membrane tubulation and dynamics (Campellone et al. 2008) whereas JMY facilitates the vesicular trafficking in the trans- Golgi region and at the ER-membrane contact sites (Schlüter et al. 2014).

This page intentionally left blank

Figure14. Scheme depicting the endocytic pathway. Modified from Derivery et al. (2009). Internalized molecules and receptors reach the early/sorting endosomes. From sorting endosomes, internalised cargos are sorted to major three routes – Recycling route to plasma membrane, degradation route to lysosomes or retrograde route to the Golgi. WASH localizes predominantly on early and recycling endosomes. Different Rab effectors localization on various compartments is depicted.

56 Chapter II: The role of the WASH complex in endosomal system

Overview of endosomal system

The endocytotic system is composed of vesicles that have mosaic of structural and functional regions. The regions consist of specialized protein-lipid domains within the plane of membrane or protein complexes associated with the specific membrane lipids. These regions also act as functionally distinct compartments communicate by tubule-vascular intermediates. Usually these compartments are characterized by their localization, acidity, lipid composition and the molecular markers they carry, specifically several Rab GTPases and their effectors (Galvez et al. 2012). Interestingly, Rab effectors assemble into multi-protein machines that control almost all the essential endosomal functions which include processes like vesicle tethering and fusion, cargo sorting, vesicular transport and recruitment of signalling molecules.

1/ A glance at endocytic pathway

Endocytosis begins at the plasma membrane where selected molecules and receptors are internalized by the invagination of plasma membrane into cytoplasm that pinches off into endocytic vesicles. These vesicles which are released by the complete scission from the plasma membrane fuse together, forming the early endosome. The early endosomes are characterized by the presence of Rab5 and EEA1 (Early Endosome Antigen 1) marker proteins. There are several types of endocytotic vesicles that deliver its content to the early endosomes. In that, the best-characterized endocytosis mechanism involves formation of clathrin-coated pits (CCP). Other characterized mechanisms include caveolae, phagocytosis, macropinocytosis and several CCP- independent pathways (Doherty and Mcmahon 2009). Early endosomes also receive vesicles that arrive from Trans-Golgi Network (TNG). Early endosomes act as a major sorting station where the fate of endocytosed proteins are decided (Huotari and Helenius 2011).

57 From early endosomes, endocytosed proteins can either be sorted into tubular extensions that separate from endosomes and deliver the proteins back to cell surface or get funnelled into endosomal lumen that leads to multi-vesicular bodies (MVBs) formation. During this formation, the endosomes mature into late endosomes and then fuse with lysosome for degradation of their content or can be recycled back either to the plasma membrane or to TGN (Burd and Cullen 2014). Thus recycling of endosomes occurs in two modes - ‘fast’ mode, where the materials transported directly from early endosomes to plasma membrane that is characterized to be regulated by Rab4 and Rab35. In contrast, during ‘slow’ mode, the cargo material transits through Rab11 positive recycling endosomes (Granger et al. 2014). Similarly, Rab7 regulates early-to- late endosomal transition and Rab9 controls retrograde traffic from the late endosomes to TGN (Figure 14).

2/ Endosomal fusion and fission processes

The whole endocytotic system greatly depends on endosomal fission and fusion processes. Endosomal fusion mechanism has been characterized well and is shown to be driven by two classes of tethering factors. The first class is composed of long coiled- coil molecules that capture vesicles at early endosomes. Such identified tether proteins are EAA1 and Rabenosyn-5. The second class is composed of multi-protein complexes called CORVET (Class C core vacuole/endosome tethering) and HOPS (homotypic fusion and protein sorting). CORVET being a Rab5 effector, functions on the early endosomes whereas HOPS are found on the late endosomes, lysosomes and are Rab7 effector (Gautreau et al. 2014).

Fission process is more complicated as there is interplay between microtubule and actin cytoskeletal network. Fission process is also more easily monitored in the context of endocytosis as the pinching of clathrin-coated pits that gets internalised is observed within the plane of plasma membrane (Merrifield and Kaksonen 2014) whereas, in the context of an endosome, it is more complicated as endosomes move fast along the microtubule tracks in a three-dimensional environment within the cells.

58 Role of microtubules in membrane fission -

Microtubules (MTs) are important structures that are instrumental for the endosomal movement and shape. It has been shown important for the generation of endosomal tubular structures (Soldati and Schliwa 2006). Cargo containing endosomes are transported along the MTs by molecular motors – dynein and kinesins. Due the polarity of MTs, kinesins, with a few exceptions move towards the plus-ends of MT which in most cells grow towards the periphery (Kull and Endow 2013). Contrarily the cytoplasmic dynein move towards the minus-end anchored to the MT-organising centre which is located near the nucleus (Allan 2011).

Occasionally, endosomes moving along the MTs can switch their direction, this suggests that few endosomes may carry both plus-end and minus-end directed motor proteins. This kind of switching is rare event for the early endosomes (Flores-Rodriguez et al. 2011, Zajac et al. 2013) but common for the late endosomes (Yi et al. 2011). Evidences that endosomes could switch directions are well documented in Dictyostelium and in HeLa cells (Soppina et al. 2009). Mallik and his colleagues observed that early endosomes which are moving in a certain direction can pause and undergo intermittent elongation as if they are being subjected to the opposing forces, before reversing direction. This effect is termed as “tug-of-war”. Hence, “tug-of-war” is a way for the endosomes to get tubulated. It also helps in the pinching of endosomal membrane due to tensions produced along with other effector proteins that leads to fission.

Role of actin cytoskeleton in membrane fission -

The role played by microtubules in endosmal fission brings its contrast towards endocytotic fission process. The endocytotic fission is independent of microtubules whereas endosomal fission is highly dependent on it. Actin cytoskeleton also plays a major role in both processes. N-WASP activates the Arp2/3 complex at the clathrin- coated pit (CCP) surface whereas WASH activates Arp2/3 at the endosomes surface (Suetsugu and Gautreau 2012). The branched actin network produced by N-WASP during endocytosis helps to stabilize and elongate the newly formed neck of invaginated clathrin pits. This elongation happens due to the actin polymerisation, which leads to barbed end pointed towards the membrane, thus exerting force against the membrane

Figure15. Role of WASH in endosomal scission. Modified from Derivery et al. (2009, Gautreau et al. (2014). a, WASH inhibition by RNAi induces tubulation of endosomes in 3T3 cells. Endosomes are labelled with fluorescent transferrin. Scale bar: 10µm. b, Representation of a tubule emanating from an endosome. Membrane tensions on endosome are created by two opposing forces: on one side, by the MT motors that pull the endosomal membrane to form a tubular extension and on other side, by the formation of WASH mediated actin networks around the neck of tubular extension. WASH also recruits the dynamin to the neck region which promotes the splitting of the tubular extensions leading to endosomal scission.

60 (Collins et al. 2011, Mooren et al. 2012). By this way, pinching of CCP from the membrane is mediated by actin cytoskeleton. A similar mechanism of force exertion happens on the surface of endosome by WASH mediated branched actin networks.

The WASH complex gets enriched in early and late endosomes (Derivery et al. 2009b, Gomez and Billadeau 2009). The WASH labels specific regions of the endosomes that are termed as micro-domains and they are associated with the actin patches. In all forms of endosomes, whether tubulated-or-not, the WASH keeps its association intact with the micro-domains on the endosomal surface.

3/Membrane scission mediated by the WASH complex

The strong evidence that WASH complex is involved in membrane fission came from the observation made by Derivery and his colleagues. When they tried to knock- down the complex they observed exaggerated tubules of endosomes as if their fission was impaired (Figure 15a) (Derivery et al. 2009b). Similar observation was also done by other researchers (Gomez and Billadeau 2009). This exaggerated tubule formation was observed to occur in all WASH positive vesicles which are characterized to be Rab4, Rab5, Rab7 and Rab11 positive compartments.

The identification of a membrane scission factor - dynamin along with the WASH (Derivery et al. 2009b) gave an insight regarding the synergy mechanism of these two complexes in performing fission. Interestingly, dynamin recruitment can also be facilitated by numerous additional interactions. For example, the branched actin networks generated by Arp2/3 complex are recognized by protein cortactin that contains an SH3 domain binding directly to the dynamin (McNiven et al. 2000, Cai et al. 2008). Moreover, inhibition of dynamin phenocopies the WASH depletion where the endosomes are tubulated.

With all these observations, the mechanism of endosome fission process became clear. Fission of the endosomes should be assisted by molecular motors that move along the MT tracks. These motors generate a pulling force on the endosomal membrane whereas the WASH-Arp2/3 pathway mediated branched actin networks generate a pushing force on the endosome. Due to these opposite forces, endosomes should form tubular budding of membranes that would undergo fission due to the high tension

61 (Figure 15b). Also, actin network helps in stabilizing the budding neck that facilitates dynamin to act and the membrane tension facilitates dynamin-mediated fission (Roux et al. 2006). Moreover, these push and pull forces on the membrane by MT motors and actin networks result in lipid sorting which might contribute to its own fission (Roux et al. 2005, Liu and Fletcher 2006). Altered lipid composition of such membrane which is under stress will be a perfect substrate for the dynamin-mediated membrane fission. Hence it can be concluded that, Arp2/3-mediated actin-polymerization driven by WASH around the endosomal tubule results in neck construction and favours scission.

4/Endosomal sorting mediated by the WASH complex

The endosomal sorting requires MT associated machinery as well as the localized actin polymerization. Endosomal sorting assisted by MT motor proteins (dynein and kinesin) creates the “tug-of-war” effect on endosomes thus leading to their tubulation and segregation of endosomal membrane domains (Soppina et al. 2009). Indeed, tubulation doesn’t occur when the cells were treated with MT-depolymerizing drug - nocodazole and endosomes became bigger in size as the receptor sorting was inhibited. Similarly, the inhibition of filamentous actin production has caused defects in the sorting of multiple membrane proteins and endosomal maturation (Ohashi et al. 2011).

Interestingly, defective WASH has been characterized to be associated with the defects of cargo transport from three major routes: (1) recycling of vesicles containing transferrin receptor (Derivery et al. 2009b), Integrins (Zech et al. 2011, Duleh and Welch 2012) (Puthenveedu et al. 2010), GLUT1 and TCR

(Piotrowski et, β2al. 2013); adrenergic (2) retrograde receptor transport of vesicles containing mannose-6- phosphate receptor (Gomez and Billadeau 2009, Harbour et al. 2010); (3) in the derivative pathway of the epidermal growth factor receptor (EGRF) (Duleh and Welch 2010) as well as in the v-ATPase retrieval from the lysosome in Dictyostelium (Carnell et al. 2011).

When the downstream function of WASH complex is stopped i.e., prevention of actin polymerization by inhibiting Arp2/3 complex results in enlarged endosomes with no tubules (Derivery et al. 2012). This observation didn’t add up with the previous observation of exaggerated tubulation of endosomes when WASH was silenced

62 (Derivery et al. 2009b, Gomez and Billadeau 2009, Duleh and Welch 2010). Things became clear when WASH knock-out in mice was generated. The complete knock-out of WASH induces enlarged endosomes with no tubules at all (Gomez et al. 2012). These distinct phenotype observations in defective WASH (silenced) vs. no WASH (knock-out) are like two sides of the same coin. In these observations, the presence of numerous endosomal tubules in WASH silenced cells may be due to the force generated by partial WASH mediated actin networks and defective endosomal membrane scission. Also, a defect in dynamin recruitment might be another reason for its inability to pinch off endosomal membrane. On the other hand, knock-out of WASH results in enlarged endosomes which reflects their inability to form endosomal tubules. This happens because there are no forces on endosomal membrane due to the complete lack of branched actin networks that leads to endosomal material accumulation. Another explanation for this phenomenon is endosomal actin networks has shown to prevent endosomes from clumping (Drengk et al. 2003) and thus the absence of actin branches might lead to the merge of endosomes.

Endosomal sorting mechanism greatly relies on the specialization of membrane micro-domains that cluster cargoes before they get transported. Branched actin is likely instrumental in defining the micro-domains by promoting lipid sorting (Liu and Fletcher 2006) and there are several WASH micro-domains observed at the surface of endosomes. These micro-domains likely possess a lipid composition that is distinct from the surrounding membrane as they coalesce easily when actin polymerization is impaired (Derivery et al. 2012). This study suggests that endosomal actin patches, prevent the micro-domains from coalescing which is likely to be important for clustering.

After the sorting of membrane micro-domain, tubules are generated by sorting- nexins (SNXs), a family of membrane-deforming proteins. These proteins are characterized to be required for transport of many proteins to the plasma membrane or TGN. They often work in cooperation with the VPS26/29/35 cargo-binding retromer complex (Burd and Cullen 2014) where each SNX acts in a subset of sorting events and defines a destination of the endosome.

Figure16. Endosome clumping in patients affected with mental retardation. Modified from (Gautreau et al. 2014). Endosomes are labelled with fluorescent transferrin in lymphoblastoid cells derived from patients carrying point mutation P1019R in SWIP. Scale bar: 10µm.

64 At the surface of endosomes, WASH binds to retromer and recruits Arp2/3 to generate branched actin networks, that facilitates the generation and scission of recycling tubules formed by the SNXs. Specifically, sorting into either endosome-to- Golgi pathway or endosome-to-cell surface pathway may be provided by the retromer or WASH complex-associated proteins such as SNX27 (Seaman et al. 2013). To add up to this point, it has been shown recently that FAM21 subunit of the WASH complex directly interacts with SNX27. This interaction is required for the precise localization of SNX27 at the endosomal micro-domains. SNX27-retromer-WASH complex directs the cargo to plasma membrane by blocking their transport to lysosomes and Golgi (Lee et al. 2016). Hence, WASH plays a critical role in endosomal fission and sorting. Defective WASH has been documented to cause many pathological disorders.

5/ Pathologies of defective WASH complex

5.1 Role of the WASH complex in neurodegenerative diseases

In general, neurons are sensitive to the defective endosomal trafficking. This might be due to the critical traffic along the long axons of motor neurons that extends up to 1 meter in length. Also, neuronal system development greatly depends on the endosomal trafficking (Mestres and Sung 2017)

WASH complex’s two large conserved subunits, Strumpellin and SWIP are mutated in diseases that lead to neuronal defects. Several mutations in Strumpellin are observed to cause hereditary spastic paraplegia (HSP) which is characterized by motor neurons degeneration. Several mutations are observed to be causing HSP apart from the locus of Strumpellin gene. Other HSP loci are recognised to encode proteins that are relevant for MT physiology. Such identified examples are KIF5A kinesin motor protein and spastin which is a MT severing protein (Dion et al. 2009). Three specific Strumpellin mutations that cause HSP are V626F, L619F and N471D. These mutations are found in the highly conserved region of Strumpellin (Freeman et al. 2013). A recent study showed that, defect in Spastin and Strumpellin has also caused defects in the endosome- Golgi traffic and lysosome functioning in mouse primary neurons and human stem cell– derived neurons. These neuronal cells phenocopied the lysosomal abnormalities in

65 human Spastin defective-HSP patients and Strumpellin knock-down phenotype in HeLa cells thus giving an insight into the HSP pathogenesis at cellular level (Allison et al. 2017).

Point mutation in the gene encoding SWIP has shown to involve in a genetic disease called non-syndromic mental retardation. Strikingly, this single point mutation P1019R in SWIP protein destabilizes the entire WASH complex that leads to endosomal clumping (Figure 16) (Ropers et al. 2011). Also, defective WASH complex due to SWIP mutation would impair the sorting mechanism of important receptor proteins. This might be the key reason for the development of non-syndromic mental retardation.

FAM21, another WASH complex large subunit plays a crucial role in normal cell physiology. It bears the domain that is responsible for WASH complex interaction with retromer complex, specifically to the VPS35 subunit. Recently, a point mutation D620N in VPS35 subunit was found to be involved in hereditary Parkinson’s disease (PD) (Vilariño-Güell et al. 2011, Zimprich et al. 2011). This mutation decreases the affinity binding between VPS35 and FAM21 by twofold. PD is the second most common neurodegenerative disease after Alzheimer’s disease. The key characteristics of PD are tremors, akinesia, rigidity, and postural inability. PD is mainly caused due to the progressive loss of dopaminergic neurons within midbrains substantia nigra pars compacta (SNpc). (McMillan et al. 2017). However, PD is also known to be caused due to the multifactorial disorder and hence, dissecting the molecular pathways are crucial to therapeutic intervention. Interestingly, another WASH accessory protein DNAJC13/RME8 that interacts with FAM21 subunit is also mutated. This mutation of DNAJC13/RME8 is also linked to the familial forms of PD (Vilariño-Güell et al. 2011, Freeman et al. 2014) thus proving the importance of retromer-WASH role in the disease.

Several phenotypes of defective WASH are also mirrored by cells having VPS35 subunit D620N mutation. The loss of WASH complex affects the endosomal sorting of cation-independent Mannose 6-Phosphate receptor (CI-MPR) (Gomez and Billadeau 2009) and similar defective receptor sorting is observed in cells carrying D620N mutation of VPS35 (Follett et al. 2014, McGough et al. 2014). Defective WASH complex also inhibits the autophagosome formation. Similarly, the cells carrying VPS35 subunit

66 D620N mutation exhibits a defective trafficking of autophagy-related protein 9A (ATG9A) which is a characteristics of inhibited autophagosome formation (Zavodszky et al. 2014). Also, in VSP35 subunit D620N mutated cells, glucose trasporter-1 (GLUT-1) is observed to be mis-localised. This is a well observed phenotype of the defective WASH complex.

In addition to the above-stated defects, trafficking of the cell surface localized dopamine receptor-1 is affected by D620N mutation in VPS35 in the mouse dopaminergic neurons which in-turn results in abnormal dopamine signalling (Wang et al. 2016). The D620N mutation of VPS35 is also found to inhibit the recycling of AMPA- type glutamate receptors (AMPARs). This recycling of AMPARs is crucial as it mediates the majority of excitatory transmission. Such misregulation of the synaptic function leads to chronic pathophysiological stress upon neuronal circuits which contributes to neurodegeneration (Munsie et al. 2015).

Recently, the point mutation K297X in VPS26A has demonstrated to affect the GLUT-1 endosomes sorting mediated by WASH-retromer complex. This happens due to the reduced binding of cargo adaptor called sorting nexin 27 (SNX27) (McMillan et al. 2016). SNX27 binds to the cargo proteins through its PDZ domain. SNX27 association with a VPS26 subunit of the retromer is mediated -hairpin of

SNX27 PDZ domain. This domain also engages arrestinthrough-like an structure exposed βof VPS26A retromer subunit (Gallon et al. 2014). Association of VPS26A increases the SNX27 affinity towards its cargo by more than 10-fold, revealing an allosteric relationship between retromer, SNX27 and cargo recognition. Interestingly, this study leads to the recognition of a minimal amino acid sequence of PDZ binding motif of the cargo proteins which are necessary for their binding to SNX27. With the help of this specific sequence, it is recognized that over 400 potential cargo proteins undergo SNX27- retromer mediated endosomal sorting (Clairfeuille et al. 2016).

Hence with all these observations it has become clear that retromer along with the WASH complex and SNX27 are very critical for the normal synaptic plasticity, nutrient uptake, metabolism, development and signal transduction (Steinberg et al. 2013, Lee et al. 2016, Clairfeuille et al. 2016, McMillan et al. 2017). Any defect among these three leads to perturbed neuronal activity and viability.

67 5.2 Role of the WASH complex in tumor progression

As WASH complex plays a crucial role in endosomal recycling and sorting, it has been demonstrated to orchestrate invasive migration of ovarian carcinoma cells through recycling of integrins to the plasma membrane (Zech et al. 2011). Also, in

breast carcinoma cells,α5β1 the WASH complex interacts with the exocyst complex and contributes to the focal delivery of trans-membrane type 1 matrix metalloproteinase (MT1-MMP) at the invadopodia tip (Monteiro et al. 2013). In this process, WASH helps in making tubular extensions from the late endosomes carrying MT1-MMP to plasma membrane. In cancer patient biopsies, there has been no examination of the WASH complex so far. However, it has been observed that the WASH complex subunit Strumpellin has overexpressed in high-grade prostate carcinomas due to genetic amplifications (Porkka et al. 2004).

68 Chapter III: Assembly of multi-protein complexes

1/ Molecular machines in nature

Multi-protein complexes has been termed as ‘molecular machines’ by Bruce Alberts to draw the comparison between human-made machines to complex assemblies of proteins in cells (Alberts 1998). Like a machine, multi-protein complexes are composed of several parts that work together in coordination to perform a specific task. It has been well observed in cells that, as the biological process is carried out there are assemblies of several proteins which in turn interacts with several other large complexes of proteins. This gave the notion of viewing the whole cell as a factory with a robust regulated system containing interlocking assembly lines leading to an elaborate network of “protein molecular machines”.

Generally, these machines have characteristics to operate as an obligate complex. Even if one of the parts of these machines is disturbed, the whole complex loses the function hence they are degraded by cells. During the lifetime of a molecular machine, subunits spend most of their time together in contrast to proteins transiently forming signaling complexes.

In nature, the best examples of molecular machines are Ribosomes which are also termed as RNA machines that involve more than 300 proteins and RNAs that plays a crucial role in protein translation (Nerurkar et al. 2015). Another good example would be proteasomes which is a large and highly conserved molecular machine. These machines are designated to eliminate the unwanted proteins which are marked by poly- ubiquitin chains for the degradation (Gu and Enenkel 2014).

2/ Specific assembly mechanisms of molecular machines

How do these machines get assembled? Indeed, molecular machines are not simple assemblies that happen due to the spontaneous association of subunits. Within each protein assembly, intermolecular collisions has been restricted to a small set of

Figure17. Assembly of the proteasome molecular machine. Modified from Murata et al. (2009), Sahara et al. (2014). a, Schematic diagram of the 26S proteasome containing 20S core particle (CP) and 19S regulatory particle (RP), followed by its subunit compositions. RPN10 is coloured yellow because it is supposed to be located at the base–lid interface. b, Assembly scheme of proteasome. The assembly of the CP and the base are regulated by proteasome-dedicated assembly factors. Assembly factors of CP - PAC1-4, and Ump1. Assembly factors of base - p27, p28, S5b and Rpn14 or n14.

70 possibilities made by the irreversible assemblage of sequential subunits that has better stability and energy state rather than being un-associated. In other words, during the reactions of A-B-C, reaction C depends on reaction B which in turn depends on reaction A and each reaction results in an assemblage that is stable than the previous. This stability is achieved by the underlying highly ordered conformational changes in proteins subunits. Such sequential conformational changes that are driven like this lead to a more stable form of protein complex, that tends to proceed in one direction.

Usually, these conformational changes of the subunits during complex assembly are facilitated by other proteins which are called ‘assembly factors’. These factors transiently associate with one or several subunits but eventually dissociate before the machine is completed. The proteasome molecular machine could be exemplified to illustrate the role of such assembly factors.

In eukaryotes, the 26S proteasome is a large complex that consists of 33 different subunits forming two sub-complexes, the 20S core particle (CP) and the 19S regulatory particle (RP) (Murata et al. 2009). The RP is attached to one or both ends of the CP. The Core complex of the proteasome is a barrel-shaped enzyme complex formed by axial stacks of four heteroheptameric rings (Figure 17a). These rings consist of two outer - rings made of seven - – -rings consistingα homologous- α subunits– (α1 α7)- ringsand two that inner are close β to each other ofform seven the homologouscatalytic cavity β subunits which presents(β1 β7). the The CP two proteolytic β active sites. The catalytic have a caspase-like, trypsin-like activityand chemostripsin is conveyed- likeby β1,activity. β2 and Hence, β5 subunits CP complex which is crucial for the proteasome machinery (Borissenko and Groll 2006, Sahara et al. 2014).

CP complex assembly is mediated by at least five conserved extrinsic assembly factors. These assembly factors are PAC1 to PAC4 and , POMP/UMP1 in mammals and , Pba1 to Pba4 and Ump1 in yeast (Ramos et al. 1998, Hirano et al. 2005, 2006, Le Tallec et al. 2007, Kusmierczyk et al. 2008, Hoyt et al. 2008, Murata et al. 2009). The CP -ring formation assisted by two heterodimer proteins assemblyPAC1-2 and initiates PAC3- with the α -ring formation as the knock- 4. These assembly factors are important for the α -ring structures. Further, thedown PAC1 ofand PAC3 PAC2 and knock PAC4- causes significant loss in the-ring α dimers which is down causes accumulation of α

71 -ring that would impair the CP complex formation (Hirano et al. an2005, undesirable 2006) form of α -ring thus regulating its intermediate. Hence, form PAC1 that isand suitable PAC2 forprevent assembly dimerization whereas, of precise the α mechanism of PAC3 -ring assembly. and PAC4 is unknown but they are crucial for the -αring from binding to RP while they are being assembled.These assembly This factorsis realized also by prevent the structural the α analysis of Pba1-Pba2-CP complex where it is -ring thus preventing theseen attachment that the of assembly RP (Hirano factors et al. bind 2005, to Stadtmueller the outer surface et al. 2012). of α After the - - - -subunit is assemblyrecruited onto of α thering, -ring β subunits in a proper are sequential incorporated order into of the– α ring. Each β (Hirano et al. 2008).α This assembly is initiated by the Ump1β2, β3, assembly β4, β5, factorβ6, β1 (Figureand β7 17b). Likewise the part of RP complex called base complex is also assembled by various assembly factors called p27, p28, S5b, Rpn14 in mammals and Nas2, Nas6, Hsm3, Rpn16 in yeast (Funakoshi et al. 2009, Le Tallec et al. 2009, Saeki et al. 2009, Kaneko et al. 2009, Park et al. 2009, Roelofs et al. 2009). Upon CP maturation, RP would bind to it that is driven by ATP hydrolysis. The assembly of both complexes are necessary. In the absence of sufficient amount of CP, fully assembled RP is unstable in vivo suggesting the CP-RP complex reach better stability than being un-associated (Glickman et al. 1998, Kriegenburg et al. 2008, Gu and Enenkel 2014). Hence, with the help of assembly factors, proteasome realizes the functional native state.

The robustness and sophistication of molecular machines are achieved by assembly factors by making the complex assembly sequential and unidirectional. This is not possible in a scenario of random spontaneous subunits association that greatly depend on the collisions of subunits due to Brownian motion. Importantly, these factors also offer a means to control the levels of assembled complexes. For example, starvation of yeast cells induces the expression of proteasome assembly factors thus favoring the assembly of new proteasomes which promotes degradation of old or damaged proteins. This process of assembly factor-mediated complex assembly is favorable for the cell as it allows the biosynthesis of new proteins in an amino acid restricted condition like starvation (Rousseau and Bertolotti 2016).

72 3/ The WAVE complex assembly

The WAVE complex is a molecular machine that activates another molecular machine – Arp2/3 complex to produce branched actin networks. WAVE complex localizes at the migrating tip of the cell hence playing a key role in cellular migration. The WAVE complex machine is characterized to be a stable pentameric complex composed of subunits - WAVE2, Brk1, Abi1, Sra1 and Nap1. The purification of WAVE2 subunit from mammalian cells co-purified the whole complex with its other subunits thereby establishing that subunits remain tightly bound to the complex (Gautreau et al. 2004).

Unlike WASP, WAVE protein does not contain GBD domain in their N-termini. Consequently, no direct association of the WAVE protein with the GTPases has been detected, still WAVE proteins work downstream of Rac GTPases (Miki et al. 1998b). Additionally, the in vitro full-length WASP is auto-inhibited and it can’t polymerize the actin whereas WAVE protein is active with the exposed VCA domain (Machesky and Insall 1998, Miki and Takenawa 2002, Eden et al. 2002, Innocenti et al. 2004). NPFs are deleterious to cells when they are not properly regulated. Indeed, overexpression of the NPFs result in exposed VCA domain which would activate Arp2/3 complex everywhere in the cytosol. This prevents the normal Arp2/3 recruitment and activation at the surface of specific membrane (Machesky and Insall 1998). Hence, in the case of WAVE NPF, there is necessity of the VCA domain to be masked within the complex.

Simple in vitro reconstitutions of the complex by stepwise addition of subunits or sub-complex yields a WAVE complex which has its VCA domain not masked whereas the assembled in vivo WAVE complex has its VCA domain masked by its subunits specifically by the Sra1 subunit (Stradal et al. 2004, Chen et al. 2010). In other words, in vitro reconstituted WAVE is active whereas in vivo WAVE is inactive suggesting that in the cell, WAVE complex doesn’t undergo assembly via spontaneous association of its subunits.

In the cell, subunits of WAVE complex are usually degraded when one subunit is missing (Blagg et al. 2003, Kunda et al. 2003, Le et al. 2006, Derivery et al. 2008). This effect is documented for the WAVE complex in mammalian cells, Drosophila cells, in amoeba Dictyostelium discoidum and in plants. Also, when the cytosol of HeLa cells was

Figure18. Putative pathway of WAVE complex assembly. Modified from Derivery and Gautreau (2010a). The Brk1 molecule is the only subunit having a large free pool. This pool is made of trimers and it is the precursor for assembly of WAVE complex. Brk1 exist as a single molecule in the native WAVE complex, demonstrating that precursor Brk1 trimer must dissociate to be incorporated in the WAVE complex. In the assembly, single Brk1 molecule binds to Abi and WAVE forming trimetic intermediate. Nap and Sra are proposed to be recruited sequentially to the Brk1/Abi/WAVE intermediate to mature into assembled WAVE complex.

74 examined for the distribution of WAVE complex subunit, majority of the WAVE complex subunits existed as complex. A notable exception was Brk1 which also exists on its own in the cytosol (Gautreau et al. 2004). Hence, most of the WAVE complex subunit and assembly intermediates should get assembled or else get degraded. In fact in the cells, proteasome machinery imposes a quality control by degrading neo-synthesized protein as much as 30% within minutes until they are protected by Hsp70 or Hsp70 family chaperones. This protection offered by Hsp70 in a way helps the proteins to attain their native confirmation (Yewdell et al. 2000, Qian et al. 2006).

In the light of all these observations, WAVE complex is an obligate complex and it should undergo a well-regulated assembly pathway that is mandatory to produce intrinsically inactive WAVE complex. The proteasome degradation machinery which works in synergy allows the stable species like a free-form-of-Brk1 or assembled WAVE complex to survive and leaves no intermediates in the cell.

Interestingly, electroporation of purified version of tagged Brk1 has allowed to follow the fate of protein inside the living cell (Derivery et al. 2008). This assay revealed that free pool Brk1 is the precursor for the WAVE complex assembly. This precursor form is a trimer which has to associate with other newly synthesized subunits in order to build a functional WAVE complex that is composed of a single molecule of each subunit. The crystal structure of precursor Brk1 and WAVE complex reconfirmed these observations that, the free form of Brk1 is a trimer and it is present as a single molecule in the assembled WAVE complex (Chen et al. 2010, Linkner et al. 2011). Hence, this precursor Brk1 has to undergo dissociation of its trimer form during the assembly process to end up as a single molecule in the assembled WAVE complex (Figure 18). It is expected that the small trimeric coiled-coil in the C-terminus of Brk1 molecule might be instrumental for the assembly process as it could easily interact and get associated with N-terminal coiled-coil of Abi and WAVE subunits during the assembly.

Recently, an assembly factor for the WAVE complex has been identified. It is called Nudel protein (Wu et al. 2012). Nudel transiently interacts with the sub-complex subunits Sra1, Nap1 and Abi1. It stabilizes them against degradation thus facilitating the assembly. Nudel has been demonstrated to be critical for the maintenance of WAVE complex level and is shown to be required for lamellipodia formation. Such assembly

75 factors are not established to be required for other small molecular machines like Arp2/3 complex or NPFs like WASH complex. On the other hand, it has been shown that Arp2/3 complex and WAVE complexes are up-regulated in many types of cancers and up-regulation of these complexes is associated with the poor prognosis of the patients (Semba et al. 2006, Iwaya et al. 2007, Molinie and Gautreau 2017). Most of these complex subunits don’t exist apart from being a fully-assembled-functional molecular machine. Hence assembly factors over-expression could be a hijacking mechanism for these carcinoma cells to over-produce molecular machines. This in-turn might help in carcinoma cells invasion and dissemination during cancer progression but such mechanisms are still elusive and requires further studies.

76 Objectives

The Arp2/3 molecular machine generates branched actin networks when activated by the Nucleation Promoting Factors (NPFs). NPFs are intrinsically inactive in cells due to the masking of their VCA domain, preventing them from activating the Arp2/3 complex. The VCA domain masking is facilitated by intermolecular interactions or it being embedded within its multi-subunit complex.

WASH protein similarly like WAVE protein is embedded in a multi-subunit complex that makes it intrinsically inactive. These molecular machines act as obligate complexes and lose function even if one of the subunits is destabilized. Moreover, such intrinsically inactive complexes can only be obtained by a proper assembly mechanism and not by random subunit association. Hence, these molecular machines have to undergo regulated assemblies in the cell which are usually driven by extrinsic assembly factors as observed for big molecular machines like proteasomes. Recently, one such assembly factor was reported for the WAVE complex. However, for the WASH complex, such assembly factors remained elusive.

Recently, WASH complex has been implicated for its role in the invasiveness of the cancer cells during tumor progression but its expression status in the tumors is not clear. On the other end, molecular machines like Arp2/3 complex and WAVE complex are shown to be often up-regulated in cancers and the subunits of such complexes always exist in an assembled form and don’t exist alone, giving an insight that they might be tightly regulated. Therefore, there is less possibility that these complexes are up-regulated due to the systematic overexpression of all their subunits in the cancer cells. Hence, the reason how cancer cells augment such molecular machines is a conundrum. Could it possibly by the overexpression of assembly factors in the cancer cells?

During my thesis, we identified the first assembly factor for the WASH complex which is called HSBP1. My first goal was to decipher the role of HSBP1 in the WASH complex assembly. With the help of biochemical reconstitutions and molecular modeling, we demonstrated HSBP1 mediated molecular mechanism in the WASH complex assembly. Also, we characterized the implications of HSBP1 loss in cellular

77 functions due to its role in WASH complex assembly. As my second goal, we checked the expression levels of HSBP1 and WASH complex in breast tumors. In line with HSBP1 being the assembly factor for WASH complex, we have demonstrated that in breast tumors, HSBP1 is overexpressed and its overexpression correlates with the up- regulation of WASH complex levels.

Results

This page intentionally left blank

80 Article A trimeric coiled coil protein promotes WASH complex assembly

This article is in peer review process to get published in EMBO journal:

Sai P. Visweshwaran, Peter A. Thomason, Raphael Guerois, Sophie Vacher, Evgeny V. Denisov, Lubov A. Tashireva, Maria E. Lomakina, Christine Lazennec-Schurdevin, Goran Lakisic, Sergio Lilla, Nicolas Molinie, Veronique Henriot, Yves Mechulam, Antonina Y. Alexandrova, Nadezhda V. Cherdyntseva, Ivan Bièche, Emmanuelle Schmitt, Robert H. Insall, Alexis Gautreau

The WASH complex induces actin polymerization by activating Arp2/3 molecular machine at the surface of endosomes and allows endocytosed vesicles scission containing transport intermediates such as integrins. Thus, WASH complex promotes integrin trafficking in cell migration and invasion. The WASH complex is a multi-subunit molecular machine like the WAVE complex and is required to get assembled first to perform its function. Here we identify HSBP1 as a critical assembly factor for the WASH complex through proteomic screening, where it only binds to the precursor form of the subunit CCDC53 and not to the whole complex.

With the help of in vitro experiments and structural modeling, we demonstrate that HSBP1 associates with the precursor CCDC53 trimer, dissociates it and forms a mixed trimer that will eventually contribute a single CCDC53 molecule to the assembling WASH complex. We also showed that depletion of HSBP1 destabilizes the WASH complex not only in mammalian cells but also in amoeba, making it a conserved assembly factor across species.

In line with the defective assembly of WASH complex, HSBP1 depletion impairs the development of focal adhesions and invasion of mammary carcinoma cells which is a causality of defective integrins trafficking. We found that HSBP1 localizes at the centrosome and is required for the cell polarity associated with the migration. Furthermore, HSBP1 knock-out closely resembled the WASH knock-out in Dictyostelium amoebae.

81 On the other end, by analyzing RNA levels and protein levels in breast tumors, we found HSBP1 to be overexpressed in cancers and its up-regulation correlates with the increased levels of WASH complex. Moreover, the patients overexpressing HSBP1 had a poor metastasis-free survival, thus HSBP1 up-regulation is associated with a poor prognosis in breast cancer. Hence, with all these observations, we demonstrate that HSBP1 is a conserved assembly factor that controls levels of the WASH complex.

This page intentionally left blank

84 A trimeric coiled coil protein promotes WASH complex assembly

Sai P. Visweshwaran1, Peter A. Thomason2, Raphael Guerois3, Sophie Vacher4, Evgeny V. Denisov5,6, Lubov A. Tashireva7, Maria E. Lomakina8, Christine Lazennec-Schurdevin1, Goran Lakisic1, Sergio Lilla2, Nicolas Molinie1, Veronique Henriot1, Yves Mechulam1, Antonina Y. Alexandrova8, Nadezhda V. Cherdyntseva5, Ivan Bièche4, Emmanuelle Schmitt1, Robert H. Insall2, Alexis Gautreau1*

1 Ecole Polytechnique, Université Paris-Saclay, CNRS UMR7654, Palaiseau 91120, France.

2 Beatson Institute for Cancer Research, Switchback road, Bearsden G61 1BD, United Kingdom

3 Institute for Integrative Biology of the Cell (I2BC), CEA, CNRS, Université Paris-Saclay, CEA-Saclay, Gif-sur-Yvette 91190, France.

4 Pharmacogenomics Unit, Department of Genetics, Institut Curie, 26 rue d'Ulm, 75005 Paris, France.

5 Laboratory of Molecular Oncology and Immunology, Cancer Research Institute, Tomsk National Research Medical Center Kooperativny pereulok, 5 Tomsk 634050, Russian Federation.

6 Laboratory for Translational Cellular and Molecular Biomedicine, Tomsk State University, 36 Lenin avenue, Tomsk 634050, Russian Federation.

7 Department of General and Molecular Pathology Cancer Research Institute, Tomsk National Research Medical Center Kooperativny pereulok, 5 Tomsk 634050, Russian Federation.

8 Institute of Carcinogenesis, N.N. Blokhin Cancer Research Center, Kashirskoe shosse 24, Moscow 115478, Russian Federation

Correspondence and material requests should be addressed to AG: [email protected]

Running title: Assembly of the WASH complex

1 ABSTRACT

The Arp2/3 complex generates branched actin networks that exert pushing forces onto different cellular membranes. The WASH complex activates the Arp2/3 complex at the surface of endosomes and thereby fission transport intermediates containing endocytosed receptors, such as 51 integrins. How the WASH complex is assembled in the cell is unknown. Here we identify the small coiled coil protein HSBP1 as a factor that specifically associates with the homotrimeric precursor form of the CCDC53 subunit to deliver a single molecule to the assembling WASH complex. In line with a conserved role in WASH complex assembly, HSBP1 depletion in human cancer cell lines and in

Dictyostelium amoebae phenocopies WASH depletion. HSBP1 is required for integrin recycling to the plasma membrane, for the development of focal adhesions and for tumor cell invasion. Overexpression of HSBP1 in breast tumors is associated with increased levels of the WASH complex and with poor prognosis for patients.

Key words: Actin cytoskeleton, Arp2/3 complex, WASH complex, HSBP1, multiprotein complex assembly, cell migration and invasion, centrosome.

2 INTRODUCTION

Cells use branched actin networks to control their shape, to power cell migration and to drive membrane remodelling in intracellular traffic (Rotty et al, 2013). The Arp2/3 complex is the stable multiprotein complex that generates branched actin networks. It contains two actin related proteins,

Arp2 and Arp3, and five other subunits that maintain the two actin related proteins associated. When activated by the WCA domain of so-called Nucleation Promoting Factors (NPFs), the Arp2/3 complex creates an actin branch (Pollard, 2007): it associates with a pre-existing actin filament and nucleates a new filament from its two subunits, Arp2 and Arp3, brought into close contact and thus mimicking a filament end (Rouiller et al, 2008). Such a multiprotein complex can be referred to as a molecular machine to highlight the coordinated work it performs (Alberts, 1998). No function has been ascribed to Arp2 or Arp3 in isolation, outside of the Arp2/3 complex.

NPFs activates the Arp2/3 complex at different subcellular locations: WAVE at the lamellipodium edge, where the branched actin network provides the force for membrane protrusion

(Rotty et al, 2013), and WASH at the surface of endosomes, where the force generated by the branched actin network contributes to the scission of transport intermediates (Derivery et al, 2009b;

Gomez & Billadeau, 2009). These transport intermediates either follow the retrograde route toward the

Golgi (Gomez & Billadeau, 2009; Harbour et al, 2010) or recycle internalized receptors to the plasma membrane (Temkin et al, 2011; Piotrowski et al, 2013). The 51 integrin is a cargo that takes the two WASH dependent routes, since it is recycled to the plasma membrane both from endosomes and after a detour through the trans-Golgi network (Zech et al, 2011; Duleh & Welch, 2012; Nagel et al,

2017; Shafaq-Zadah et al, 2016; De Franceschi et al, 2015). WAVE and WASH are both stably associated with four other proteins, which integrate inputs to control WCA exposure as an output

(Derivery & Gautreau, 2010b; Rotty et al, 2013). The endosomal recruitment of the WASH complex depends on the cargo-recognition complex of the retromer (Jia et al, 2012; Harbour et al, 2012; Helfer et al, 2013; Gautreau et al, 2014). Formation of branched actin networks thus involves cascades of molecular machines.

3 How cells assemble these molecular machines from neosynthesized subunits is not known in most cases. Indeed, molecular machines are not just simple assemblies driven by the spontaneous association of subunits. The simple stepwise addition of subunits or subcomplexes yields a WAVE complex, where the WCA domain is not properly masked (Innocenti et al, 2004; Derivery et al,

2009a). The reconstitution of a native WAVE complex from recombinant subunits was a tour-de- force, which required a decade of work (Chen et al, 2014). In fact, in the cell, proteasomes exert a quality control and degrade up to 30 % of neosynthesized proteins (Schubert et al, 2000). When one subunit of WAVE, WASH or Arp2/3 complexes is depleted, remaining subunits of the same molecular machine are usually degraded by proteasomes (Kunda et al, 2003; Steffen et al, 2006;

Derivery et al, 2009a; 2009b; Jia et al, 2010). Conversely, when an exogenous, usually tagged, subunit is overexpressed, the endogenous subunit is degraded, because its partner subunits have been titrated by the more abundant exogenous protein (Derivery et al, 2009a; 2009b). These observations suggest that subunits need to assemble with their partner subunits to reach their native state and become stable

(Derivery & Gautreau, 2010b). In the case of the WAVE complex, one subunit, Brk1, forms a homotrimeric precursor, even though only a single Brk1 subunit is present in the native complex

(Derivery et al, 2008; Linkner et al, 2011). Brk1 turns over more rapidly than the WAVE complex, suggesting that the two Brk1 molecules that should remain after dissociation of the trimeric precursor are also degraded (Derivery & Gautreau, 2010b; Wu et al, 2012).

Large molecular machines, like proteasomes, require many factors for their assembly (Sahara et al, 2014; Budenholzer et al, 2017). Assembly factors transiently associate with one or several subunits, but eventually dissociate before the machine is completed. If they were to remain associated, they would be subunits. It is not yet established whether assembly factors are systematically required for the formation of smaller molecular machines, like the Arp2/3 or NPF complexes. So far, an assembly factor was only identified in the case of the WAVE complex: the Nudel protein, which transiently interacts with two subcomplexes, is critical to maintain WAVE complex levels and thus to form lamellipodia (Wu et al, 2012). Assembly factors offer a means to control the levels of assembled complexes. For example, starvation induces the expression of proteasome assembly factors and hence

4 favors the assembly of new proteasomes to promote the degradation of old proteins, thereby allowing biosynthesis of new proteins in amino-acid restricted conditions (Rousseau & Bertolotti, 2016).

The Arp2/3 and the WAVE complexes are overexpressed in a variety of cancers (Molinie &

Gautreau, 2017). This overexpression is usually associated with a high grade, lymph node invasion and poor prognosis for patients (Semba et al, 2006; Iwaya et al, 2007; Molinie & Gautreau, 2017).

Since most subunits of these molecular machines are only stable within the whole complex, it means that invasive tumor cells managed to assemble more of these machines, but the mechanism involved is not known. The WASH complex, which allows focal delivery of metalloproteases and integrin recycling, is critical for tumor cell invasion (Zech et al, 2011; Monteiro et al, 2013), but whether its expression is deregulated in tumors is not known. Here we identify the first assembly factor of the

WASH complex, HSBP1, and characterize how it promotes WASH assembly. We found that HSBP1 is overexpressed in breast cancer and that its overexpression is associated with increased levels of

WASH complex and poor survival of patients.

5 RESULTS

Identification of HSBP1 as an assembly factor of the WASH complex

To look for putative assembly factors of the WASH complex, we derived stable 293 cell lines expressing tagged subunits of the WASH complex. We noticed a small molecular weight protein in the immunoprecipitates of CCDC53, but not of WASH, nor of Strumpellin (Figure 1A). We identified this protein as HSBP1 by mass spectrometry. The gene encoding this small protein of 76 amino acids

(8.5 kDa of predicted mass) was cloned and transiently co-expressed with each subunit of the WASH complex in 293T cells. Indeed, out of the 5 subunits of the WASH complex, CCDC53, WASH,

Strumpellin, SWIP and FAM21, only CCDC53 interacted with HSBP1 (Figure 1B). HSBP1 is thus a possible assembly factor, because it binds to a single subunit, and not to the whole WASH complex.

To examine whether HSBP1 directly binds to CCDC53, we produced and purified these two recombinant proteins in E. coli (Figure 1C). HSBP1 was previously crystallized and the HSBP1 core is formed by a trimeric coiled coil (Liu et al, 2009). HSBP1 was, however, described as a hexamer, because the asymmetric unit of the crystal contains two trimeric coiled coils associated end-to-end. In this structure, the two trimers differ by the conformation of a bulky residue in their coiled coil, phenylalanine 27 (F27; Figure S1A). The end-to-end association of two HSBP1 trimers is a crystal- induced contact, because we found that HSBP1 in solution behaves as a homotrimer, when its mass is evaluated by Size Exclusion Chromatography coupled to Multi Angle Light Scattering (SEC-MALS;

Figure 1D). The structure of CCDC53 was not previously characterized. We found by SEC-MALS that free CCDC53 also behaves as a homotrimer (Figure 1D). The situation is reminiscent of the distantly related Brk1 subunit of the WAVE complex (Jia et al, 2010): Free Brk1 exists as a homotrimer, even though a single subunit of Brk1 is present in the WAVE complex (Derivery et al,

2008). When CCDC53 and HSBP1 were mixed, we detected a new peak by SEC-MALS, which surprisingly corresponded to a mixed heterotrimer, containing two HSBP1 for a single CCDC53 (Fig.

1D). So not only HSBP1 directly binds to CCDC53, but also this interaction remodels the quaternary structure of each component.

Trimeric HSBP1 displays a distorted coiled coil. This distortion most likely arises from the accommodation of the bulky F27, which creates a steric hindrance (Figure S1A). Such a bulky residue

6 at the d position of a coiled coil heptad is highly unusual in trimeric coiled coils (Woolfson, 2005).

However, F27 is strictly conserved in all HSBP1 homologs (Figure S1B). To better understand the reaction occurring between HSBP1 and CCDC53, we modeled the structure of free CCDC53 according to the crystal structure of free Brk1 (Linkner et al, 2011). CCDC53, whose name stands for

Coiled Coil Domain Containing protein 53, can be modeled as a trimeric coiled coil protein, like Brk1

(Figure 2A). Then we used the two previous structural models of HSBP1 and CCDC53 to model the mixed HSBP1-CCDC53 heterotrimer. In heptads 4 and 5 of the mixed heterotrimer, we noticed that the single strand of CCDC53 brings two optimal electrostatic couples with the two strands contributed by HSBP1 (Figure 2B). These salt-bridges, formed between residues E(e)-K/R(g’) of heptads, were identified as the most stabilizing motifs in experimental studies of model trimeric coiled-coils (Roy &

Case, 2009). The residues involved are strongly conserved in the sequence alignments of both HSBP1 and CCDC53 (Figure S1B). These two stabilizing salt-bridges probably contribute to the spontaneous assembly of the mixed HSBP1-CCDC53 heterotrimer.

To evaluate the potential role of HSBP1 in the assembly of the WASH complex, we isolated

MDA-MB-231 clones stably depleted of HSBP1 and analyzed their content in WASH complex subunits. Upon HSBP1 knock-down, steady state levels of CCDC53 and WASH, but not those of

FAM21 and Strumpellin, were significantly decreased (Figure 3A). The effect of HSBP1 depletion was the same in HeLa cells upon transient transfection of siRNAs (Figure S2). The subunits of stable multiprotein complexes usually depend on each other for their stability. For the WASH complex, the depletion of a 'large' subunit, such as Strumpellin, SWIP or FAM21, induces the degradation of all remaining subunits, whereas the depletion of a 'small' subunit, such as WASH or CCDC53, induces the degradation of the other small subunit, leaving intact levels of large subunits (Jia et al, 2010).

Upon HSBP1 depletion, the co-dependent CCDC53 and WASH subunits are thus specifically down- regulated.

Further evidence for the relative independence of a WASH-CCDC53 subcomplex was provided by a stable MDA-MB-231 cell line that over-expresses tagged WASH. The increased WASH levels were associated with increased levels of CCDC53, but not of FAM21 and Strumpellin (Figure

3B). When tagged WASH was overexpressed, endogenous WASH was down-regulated, as observed

7 previously (Derivery et al, 2009b), suggesting that the levels of partner subunits control the total amount of WASH. In this experiment, WASH was tagged with both FLAG and HaloTag sequences.

The HaloTag covalently reacts with haloalkane reactive ligands. A chemical compound called

HaloPROTAC3 induces the specific degradation of HaloTagged proteins through the recruitment of the VHL E3 ubiquitin ligase (Buckley et al, 2015). When cells expressing HaloTagged WASH were incubated with 1 M HaloPROTAC3 for 24 h, levels of tagged WASH and CCDC53 significantly decreased, while endogenous WASH reappeared (Figure 3B).

To monitor the assembly of the WASH complex, we washed out the HaloPROTAC3 compound to resume expression of HaloTagged WASH and to follow its assembly into the WASH complex by immunoprecipitation. Using this procedure, we were able to detect the build up of WASH and associated subunits over several hours (Figure 3C). Upon HSBP1 depletion, HaloTagged WASH associated with a reduced amount of CCDC53, but with normal amounts of FAM21 and Strumpellin.

This experiment suggests that HSBP1, which dissociates the free homotrimer of CCDC53, is required for its assembly into the WASH complex. A structural model of the WASH complex, based on the crystal structure of the analogous WAVE complex (Chen et al, 2010), suggests a molecular scenario compatible with all the above observations: CCDC53 must be dissociated from a precursor homotrimeric form to contribute a single subunit in the assembling complex. HSBP1 would favor this dissociation by forming a mixed heterotrimer with a single CCDC53 and would then deliver this single CCDC53 to the WASH complex, where a new coiled coil can form between WASH, CCDC53 and FAM21 (Figure 3D). It should be stated, however, that the model of the WASH complex is not as robust as the one of CCDC53-HSBP1 heterotrimer, because apart from a predicted coiled coil segment in its N-terminus, the structure of FAM21 cannot be accurately predicted.

Phenotype associated with HSBP1 depletion in MDA-MB-231 cells

Because WASH is involved in recycling 51 integrins from intracellular compartments to the plasma membrane, we checked the localization of this particular integrin in the stable MDA-MB-

231 clones depleted or not of HSBP1. We found that 51 focal adhesions from HSBP1 depleted cells

8 were strongly reduced in numbers and slightly reduced in size (Figure 4A,B). HSBP1 depleted cells, indeed, display decreased levels of 51 integrin at their surface (Figure 4C) and this defect was associated with reduced adhesion to fibronectin, the major ligand of the α5β1 integrin, and to collagen type I (Figure 4D). Similar defects were previously documented in WASH depleted cells (Zech et al,

2011; Duleh & Welch, 2012).

MDA-MB-231 is an invasive mammary carcinoma cell line. A simple and classical assay in cancer research is the migration through pores of 'transwell filters'. Coating the transwell filter with a thick layer of extracellular matrix force cancer cells to invade the gel to reach the other compartment.

HSBP1 depleted cells were drastically impaired in their ability to migrate and invade through transwell filters (Figure 5A). We next examined how the HSBP1 depleted cells migrate on a 2D fibronectin substrate. When trajectories of individual cells were examined, tracks of HSBP1 depleted cells were much shorter than control ones (Figure 5B, Movie S1). The mean square displacement highlights that HSBP1 depleted cells explore a much smaller surface area over time than controls, a direct consequence of the reduced speed and decreased directional persistence of HSBP1 depleted cells. In 3D collagen gels supplemented with fibronectin, similar defects of HSBP1 depleted cells were observed (Figure S3, Movie S2).

The HSBP1 depleted cells were sufficiently adherent to form lamellipodia and to spread on fibronectin, but they were poorly polarized, as their lamellipodium usually covers their whole periphery (Figure 6A). This defect in cell polarization was captured by their circularity index, which remains higher over several hours than the one of control cells (Figure 6B). In sharp contrast, control cells break symmetry and elongate perpendicular to the migration direction, resulting in an increased aspect ratio, the ratio of the long over short axis. From time to time, control cells transiently lose polarity and change migration direction. The associated fluctuation in the aspect ratio is captured by the 'volatility' of this parameter, an index borrowed from stock market analysis. The volatility of the aspect ratio of HSBP1 depleted cells is much lower than the one of control cells, since the lack of polarity of the latter yield a relatively constant aspect ratio close to 1 (Figure 6B). These results suggest that defective polarity of HSBP1 depleted cells can account for their impaired migration and invasion.

The role of HSBP1 is conserved in the Dictyostelium discoideum amoeba

HSBP1 is a well conserved protein. We decided to knock-out the orthologous gene in the amoeba Dictyostelium discoideum, where we had previously studied the role of the WASH complex

(Carnell et al, 2011). To this end, we replaced the whole HSBP1 coding sequence by a selectable marker (Figure S4A). Using a specific WASH antibody, we found that WASH is down-regulated in

HSBP1 KO amoebae (Figure 7A). Re-expression of HSBP1 in the KO background restored WASH levels. We then used mass spectrometry to examine the expression of the other subunits of the WASH complex. We were able to detect and quantify specific peptides encoded by the orthologous WASH,

FAM21, Strumpellin and SWIP genes in complex extracts of amoebae (Figure 7B). The smallest subunit, CCDC53, which generates few peptides, was unfortunately not detected. HSBP1 KO induced down-regulation of WASH, but not of FAM21, Strumpellin and SWIP, as in mammalian cells. In the amoeba, WASH recruits the Arp2/3 complex at the surface of vesicles of lysosomal origin. As in

WASH KO, HSBP1 KO displayed a pronounced defect in Arp2/3 recruitment at the surface of vesicles (Figure 7C).

In the presence of fluorescent dextran, which is not digestible, HSBP1 KO amoebae become

'constipated' and that prevents them from proliferating (Figure S4B). In fact, HSBP1 KO amoebae continue to accumulate dextran well after the 2 h required for the WT amoebae to reach a steady state, when exocytosis compensates for endocytosis (Figure S4C). When fluorescent dextran was washed out, the exocytosis defect of HSBP1 KO amoebae became obvious (Figure 7D). This phenotype of

HSBP1 KO amoebae was also rescued by HSBP1 re-expression. Exocytosis involves post-lysosomal vesicles inthe amoeba and requires WASH mediated retrieval of the V-ATPase from lysosomes

(Carnell et al, 2011). The HSBP1 phenotype is thus the same as the one displayed by WASH KO amoebae. Together these experiments indicate that the role of HSBP1 in assembling functional WASH complexes is conserved in two distant species, human and Dictyostelium amoeba.

HSBP1 localizes to the centrosome

10 To examine the localization of HSBP1 in Dictyostelium amoeba, we generated GFP fusion proteins at both HSBP1 ends (Figure S4D). In both cases, HSBP1 localizes to central dot-like structures, which were identified as centrosomes using the -tubulin marker. We then examined

HSBP1 localization by immunofluorescence of mammalian cells. In MDA-MB-231 cells, HSBP1 antibodies also brightly stained centrosomes (Figure 8A). This staining is specific, because it was lost upon HSBP1 depletion (Figure S5). HSBP1 staining was clearly associated with -tubulin, a specific marker of the pericentriolar material, but did not completely overlap with it (Figure 5C). Then we stained cryosections of normal breast and mammary carcinoma. HSBP1 was associated with centrosomes in both cases. Centrosomes are at the base of cilia in normal breast, whereas these structures regress during cancer progression (Menzl et al, 2014). HSBP1 staining was always partially overlapping with -tubulin, but appeared slightly more elongated in the presence of cilia in normal breast (Figure 8B). In conclusion, HSBP1 is enriched in centrosomes in both amoebae and human cells.

HSBP1 expression in breast cancer

Since the WASH complex is critical for tumor cell invasion, we examined the putative involvement of HSBP1 in the progression of breast cancer. To this end, the levels of HSBP1 mRNA was quantified in the mammary tumors of a large retrospective cohort of 446 patients, whose long- term outcome was known. HSBP1 mRNA expression significantly increased with the Scarff-Bloom-

Richardson grade, which is a histological score used to evaluate cancer prognosis (Table S1). HSBP1 was also more expressed in tumors exhibiting overexpression of ERBB2. Most importantly, patients that carried tumors with a high level of HSBP1 had poor metastasis-free survival (Figure 9A).

In order to check the role of HSBP1 in WASH complex assembly in vivo, we examined the expression of HSBP1 and of WASH complex subunits from paired samples, which correspond to mammary carcinomas and adjacent normal tissue from the same patient. About half of tumors overexpressed HSBP1 at the protein level compared to normal tissue (Figure 9B). Most importantly, these tumors also exhibited increased levels of the WASH complex subunits CCDC53, WASH and

11 Strumpellin, but surprisingly not of FAM21. We verified by immunofluorescence of formalin-fixed paraffin-embedded sections that mammary carcinoma cells, stained by cytokeratin 7, and not to stromal cells, overexpress HSBP1 and WASH complex subunits, CCDC53 and WASH (Figure S6). In sharp contrast, the tumors that did not overexpress HSBP1 did not show overexpression of WASH complex subunits. These results obtained in patient samples suggest that HSBP1 overexpression is a means, by which in mammary carcinoma cells overexpress the WASH complex.

DISCUSSION

We have identified HSBP1 as a specific partner of CCDC53 using proteomics, and showed its involvement in WASH complex assembly. HSBP1 is the first assembly factor identified for the

WASH complex. This small protein of 8.5 kDa is a negative regulator of the HSF1 transcription factor (Satyal et al, 1998). HSBP1 stands for HSF1 Binding Protein-1. HSF1 is the major transcription factor that is activated in response to heat shock. HSF1 activation involves its trimerization through a coiled coil. HSBP1 terminates HSF1 transcription in the nucleus by dissociating the active HSF1 trimer (Satyal et al, 1998). Here we report a cytosolic function of HSBP1: HSBP1 can dissociate the precursor trimeric form of CCDC53 to favor WASH complex assembly. The two functions of HSBP1 thus involves a similar structural mechanism. HSBP1 with its invariant bulky F27 may be designed as a marginally stable trimeric coiled coil protein, which promotes the dissociation of other trimeric coiled coil proteins. The phenotypes we report here upon HSBP1 depletion appear unrelated to the termination of the heat-shock transcriptional response.

HSBP1 inactivation in human cells and Dictyostelium amoebae revealed that its function in

WASH complex assembly has been conserved across evolution. In both organisms, some subunits, but not all, i.e. CCDC53 and WASH, but not Strumpellin, SWIP and FAM21, are destabilized upon

HSBP1 depletion. This is reminiscent of the fact that only CCDC53 is destabilized, when WASH is depleted, and only WASH is destabilized, when CCDC53 is depleted (Jia et al, 2010). Furthermore, we found that, by overexpressing Halotagged WASH or by selectively destabilizing it through

HaloPROTAC3 treatment, only the CCDC53 subunit mirrors the amount of WASH. Taken together, these observations suggest that the Strumpellin-SWIP-FAM21 subcomplex exists in the cell. Whether this incomplete complex represents a stable assembly intermediate or whether it has a function of its own, and thus partners of its own, is an open question for future studies.

HSBP1 inactivation produced phenotypes similar to the ones obtained upon WASH complex depletion. In Dictyostelium amoeba, WASH knock-out produced defective Arp2/3 recruitment at the surface of lysosomal vesicles and defective exocytosis of non-digestible dextran (Carnell et al, 2011).

In human cells, WASH depletion led to reduced surface levels of 51 integrins and a decreased

13 number of focal adhesions (Zech et al, 2011; Duleh & Welch, 2012). WASH depletion also reduced invasive abilities of tumor cells (Zech et al, 2011; Monteiro et al, 2013). We have observed all these

WASH phenotypes in HSBP1 depleted human cells and Dictyostelium amoebae, thus confirming that

HSBP1 is required to assemble functional WASH complexes.

HSBP1 depleted cells displayed a striking defect in polarity establishment that prevented them from migrating, even though they were able to generate membrane protrusions. This phenotype was, to our knowledge, not previously reported upon WASH complex inactivation. However, the recycling of 51 integrins, through their association with Rab25 (Caswell et al, 2007), and when they take an indirect route through the Golgi by following the retrograde pathway (Shafaq-Zadah et al, 2016), is critical for migration persistence. Establishment and maintenance of cell polarity also rely on Rho- mediated contractility (Lomakin et al, 2015), which is itself regulated by the recycling of integrins

(White et al, 2007). So even if a polarity defect was not previously reported upon WASH depletion, this HSBP1 phenotype is consistent with our understanding of defective integrin recycling.

Furthermore, we found that HSBP1 is associated with centrosomes in both human cells and

Dictyostelium amoebae. The centrosome represents a polarity marker oriented toward the lamellipodium (Euteneuer & Schliwa, 1992; Ueda et al, 1997). It dictates the localization of the Golgi and as such, can reorient integrin delivery to the leading edge and contribute to migration persistence

(Théry et al, 2006; Wakida et al, 2010; Shafaq-Zadah et al, 2016). Centrosomes were recently identified as a determinant of tumor cell invasion (Godinho et al, 2014).

The localization of HSBP1 at centrosomes suggests that centrosomes are sites of WASH complex assembly. The WASH complex was recently reported to nucleate branched actin networks at the centrosome, in addition to its main localization at the surface of endosomes (Farina et al, 2016).

These centrosomal branched actin networks were implicated in anchoring centrosomes to the nucleus

(Obino et al, 2016). The centrosome might also be a privileged site for the assembly of other multiprotein complexes. Indeed, Nudel, the assembly factor identified for the WAVE complex, also accumulates at centrosomes, in line with its regulatory activity toward dynein, the microtubule minus- end directed motor (Guo et al, 2006). The centrosome might also be a privileged site for degradation, which is often associated with multiprotein complex assembly (Derivery & Gautreau, 2010b; Wu et

14 al, 2012). Indeed, misfolded proteins accumulate at the centrosome through dynein dependent transport in a structure called aggresome, when proteasomes are overwhelmed by substrates to degrade (Johnston et al, 1998; Wan et al, 2012).

HSBP1 is overexpressed in breast tumors and its overexpression is associated with poor prognosis for patients. We documented HSBP1 overexpression in tumors compared to normal tissue in about half of patients. Fully consistent with the role of HSBP1 in assembling WASH complexes, this overexpression was associated with increased levels of WASH complex subunits. The overexpression of the WASH complex in cancer was not previously reported, even if its role in tumor cell invasion was established (Zech et al, 2011; Monteiro et al, 2013). FAM21 was the only WASH complex subunit, whose levels were not correlated with the ones of HSBP1 and which was not overexpressed.

FAM21 appears to have moonlighting functions, since it was found to regulate transcription by NF-B in the nucleus independently of the WASH complex (Deng et al, 2015). In Dictyostelium amoebae, the

FAM21 knock-out phenotype is also quite different from the one observed in the WASH knock-out amoeba (Park et al, 2013).

In conclusion, our study identified HSBP1 as the first assembly factor of the WASH complex, structurally characterized the step in which it functions and established that HSBP1 mediated assembly of the WASH complex provides a mechanism for overexpressing the WASH complex in tumors.

MATERIALS AND METHODS

Information on patients, biopsies and qRT-PCR analyses can be found in supplementary methods

Plasmids

All the human Open Reading Frames (ORFs) were flanked by FseI and AscI sites for easy shuttling between compatible plasmids. Human HSBP1 was amplified by PCR from cDNA clones IMAGE:3010737. The WASH complex subunits CCDC53, Strumpellin, and SWIP were amplified by PCR using the respective cDNA clone IMAGE:4272385, IMAGE:30532658 and IMAGE:8143997. WASH was previously cloned similarly (Derivery et al, 2009b; Helfer et al, 2013). All PCR amplified genes were verified by sequencing. ORFs were subcloned into the following plasmids:

- a modified pET28b for the expression in E. coli. This plasmid tags the proteins with a His tag and a TEV cleavage site (MGSSHHHHHHSSGENLYFQGRP); - pcDNAm FRT PC GFP Blue (Derivery et al, 2009b) and pcDNAm FRT PC mCherry Blue (Helfer et al, 2013) for expression of GFP or mCherry tagged proteins at their N-terminus and the easy isolation fo Flp-In cell lines (Derivery et al, 2009a); - MXS EF1Flag HaloTag Blue SV40pA PGK PuroR bGHpA for the expression of proteins tagged at their N-terminus with FLAG and HaloTag in a plasmid. The shRNA-expressing pGIPZ plasmids, non-targeting control (RHS4346), shHSBP1 #1 (V3LHS_368660, TTATCTTGCATCTGCTGCA, ORF), shHSBP1 #2 (V3LHS_409720, TGAATAGCACAACTGACCA, 3'UTR) were obtained from Dharmacon.

Protein purification and SEC-MALS His tagged CCDC53 and HSBP1 were purified from E. coli BL21 Rosetta pLacI-Rare (Merck, Novagen), grown in 2xTY containing 25 µg/mL of kanamycin and 34 µg/mL of chloramphenicol. Cultures were induced overnight at 37°C with 1 mM IPTG, and growth was continued for 4 hours at 18°C. Tagged CCDC53 and HSBP1 were retained on a Talon affinity resin (Clontech) equilibrated in buffer A (10 mM HEPES pH 7.5, 500 mM NaCl, 3 mM 2-mercaptoethanol, 0.1 mM PMSF, 0.1 mM benzamidine), eluted with buffer A containing 125 mM imidazole and further purified by gel filtration on a Superdex 200 column (10/300; GE Healthcare) equilibrated in buffer B (10 mM HEPES pH 7.5, 200 mM NaCl, 3 mM 2-mercaptoethanol, 0.1 mM PMSF, 0.1 mM benzamidine). Pure fractions were pooled and concentrated to 10 mg/mL. To purify the HSBP1-CCDC53 complex, an extract from His tagged CCDC53 expressing E. coli cells was mixed with an extract from E. coli expressing untagged HSBP1 and purified as described above, except that the mixed heterotrimer was eluted with only 10 mM imidazole. Superdex 200 increase column (10/300; GE Healthcare) coupled to multi-angle static

16 light scattering (SEC-MALS, Wyatt). Light scattering and refractive index measurements were performed with a MiniDAWN TREOS device (Wyatt technology) and a RID-20A (Shimadzu), respectively. Molecular mass and hydrodynamic radius calculations were performed with the ASTRA 6.1 software (Wyatt Technology) using a dn/dc value of 0.183 mL.g−1.

Antibodies, immunoprecipitations and western blot Polyclonal antibodies targeting CCDC53 and FAM21 were obtained by immunizing rabbits with full length CCDC53 and the FAM21 (1196-1341) fragment and affinity purified on a column coupled to the immunogen. Polyclonal antibodies directed against human and Dictyostelium WASH were previously described (Derivery et al, 2009b; Carnell et al, 2011).

Strumpellin polyclonal antibodies (C-14) were from Santa Cruz Biotechnology. HSBP1 pAb (Prestige HPA028940), HSBP1 mAb (clone 2C3), anti-Flag mAb (Clone M2) and Anti-γ-Tubulin mAb (Clone GTU-88) antibodies were from Sigma-Aldrich. Integrin α5 mAb (Clone VC5), p150Glued (clone 1) were from BD Biosciences. Integrin β1 mAb (Clone 12G10) was from Abcam. Cytokeratin 7 mAb (clone N-20) was from Santa Cruz Biotechnology. Tubulin monoclonal antibody (clone E7) was obtained from Developmental Studies Hybridoma Bank. The Protein-C monoclonal antibody recognizing the PC tag (clone HPC4) was from Roche.

Immunoprecipitations were performed using resin directly coupled to HPC4 Protein-C monoclonal antibody (Roche), FLAG M2 beads (Sigma) or GFP-trap beads (Chromotek). Beads were incubated in the lysates prepared in RIPA buffer (10 mM HEPES, 50 mM NaCl, 1% NP-40, 0.5% DOC, 0.1% SDS, pH 7.7) and washed with the same buffer. HSBP1 was identified by nano LC- MS/MS as previously described (Schlüter et al, 2014). SDS-PAGE was performed using 4-12 % Bis- Tris Nupage gels (ThermoFischer scientific) and proteins were transferred onto nitrocellulose membranes (Protran BA85; 0.45µm; Sigma) using liquid transfer (Criterion blotter, BioRad) with sodium carbonate transfer buffer (6.25mM NaHCO3, 4.3 mM Na2CO3, 20% Ethanol, pH 9.5). To reveal the small protein HSBP1, we optimized the western blot protocol: It involved 30 min of in gel renaturation in (20% glycerol, 50 mM Tris-HCl, pH7.4); after the transfer, proteins were cross-linked to the membrane using a 30 min incubation in PBS containing 0.4 % paraformaldehyde. Blots were developed with alkaline phosphatase coupled antibodies (Promega) and NBT-BCIP as substrates (Promega) or Horseradish peroxidase (HRP) coupled antibodies (Sigma) and SuperSignal West Pico chemiluminescent substrate (Thermo Fisher Scientific). Signals were then quantified by densitometry of product deposition on membranes or by photon counting (ChemiDoc imaging system; Bio-Rad).

Mammalian cells and transfections

17 293T, HeLa and Flip-In 293T cells (Thermo Fisher Scientific) were grown in DMEM. MDA- MB-231 parental cells and MDA-MB-231 shRNA clones were grown Leibovitz’s L-15 medium. It was later found that MDA-MB-231 cells grew much faster in MEM medium and cells expressing HaloTagged WASH were grown and selected in MEM medium. All media were supplemented with

10% FBS. Media and serum were from Thermo Fisher Scientific. Cells were incubated in 5% CO2 except cells in cultured in Leibovitz’s L-15 medium, which were grown with no CO2. All cells and stable clones were regularly tested for mycoplasma infection using a sensitive PCR assay and found to be negative.

For transient HSBP1 knockdown, HeLa or MDA-MB-231 cells were transfected with non- targeting siRNA (siZerosense; Sigma) or HSBP1 targeting siRNA (siHSBP1# 1 - SASI_Hs01_00232421, siHSBP1# 2 - SASI_Hs01_00232417; Sigma) using lipofectamine RNAiMax (Thermo Fisher Scientific), and examined after 2 days. 293T cells were transiently transfected using calcium phosphate precipitation. Stable transfectants of Flip-In 293T cells were obtained by homologous recombination at the FRT site as previously described (Derivery & Gautreau, 2010a). To isolate stable clones depleted of HSBP1 or expressing HaloTagged WASH, MDA-MB-231 cells were transfected with appropriate plasmids using Lipofectamine 2000 (Thermo Fisher Scientific) and selected with 2 μg.mL−1 Puromycin (Invivogen). Clones were isolated with cloning rings and expanded.

Mammalian cell assays Cell adhesion assays are performed with cells detached with 2mM EDTA in PBS and plated in 96 well plates previously coated with 10 µg.mL-1 of collagen-I (BD Bioscience), fibronectin (Sigma), or heat-denatured BSA (Sigma) as a negative control for background substraction. Cells were allowed to adhere for 45 min at 37˚C, non-adherent cells were washed using PBS, and adherent cells were estimated using the MTS reagent (Promega).

2D cell migration was performed on glass bottomed µ-Slide (Ibidi) coated with 10 µg.mL-1 fibronectin. 3D cell migration, cells were sandwiched between two collagen gels (2 mg.mL-1 collagen- I, 10 µg.mL-1 fibronectin, 25 mM Hepes, 10% FBS, in DMEM) polymerised in glass bottomed µ-Slide at 37°C for 1 h.

Transwell assays were performed using Transwell inserts (FluoroBlok, Corning) either coated with 10 µg.cm-2 fibronectin (migration) or with a collagen gel containing fibronectin as described above. Cells were incubated in serum free medium and attracted to the other side of the filters by 10 % serum. Cells were stained with calcein AM (Thermofischer Scientific) and cells were imaged without fixation.

18 Immunofluorescence and FACS Cells were fixed with 2% paraformaldehyde and permeabilized with 0.25 % Triton X-100 for integrin staining. Cells were fixed with cold (-20°C) methanol-acetone (1:1) for HSBP1 and -tubulin staining. For immunofluorescence on tissue section displayed in Fig. 8, biopsies were frozen in liquid nitrogen and serially sectioned. Sections were fixed with methanol-acetone. For immunofluorescence on tissue section displayed in Fig.S7, 7 μm-thick sections were prepared from FFPE tumor samples, deparaffinized, rehydrated, processed for heat-induced epitope retrieval in PT Link (Dako) with EDTA buffer (pH 8.0), and blocked with 3% BSA in PBS. For FACS analysis, cells were fixed with 2% paraformaldehyde and analyzed on a Guava easyCyte system (Millipore).

Microscopy and image analysis Widefield imaging was performed on an Axio Observer microscope (Zeiss) equipped with a Plan-Apochromat 100x/1.40 or 63×/1.40 oil immersion objective, an EC Plan-Neofluar 40×/1.30 oil immersion objective and a Plan-Apochromat 20×/0.80 air objective, a Hamamatsu camera C10600 Orca-R2 and a Pecon Zeiss incubator XL multi S1 RED LS (Heating Unit XL S, Temp module,

CO2 module, Heating Insert PS and CO2 cover). Live cell analysis was performed as described (Dang et al, 2013). Confocal images were acquired on a commercial confocal laser scanning microscope (TCS SP8, Leica) equipped with an inverted frame (Leica), a high NA oil immersion objective (HC PL APO 63x/1.40, Leica) and a white light laser (WLL, Leica). Image analysis was performed with ImageJ and FIJI. Analyses of focal adhesions, cell shape, including volatility of the aspect ratio, and cell trajectories were previously described (Horzum et al, 2014; Lomakin et al, 2015; Zhang et al, 2016; Gorelik & Gautreau, 2014).

Computational modelling Modeling of the CCDC53 homotrimer and of the HSBP1-CCDC53 heterotrimer was performed using the RosettaCM protocol to build and relax the trimeric structures (Song et al, 2013). The CCDC53 sequence was aligned on the sequence and structure of the homotrimeric Brk1 paralog (PDB code 3PP5) (Linkner et al, 2011) following the alignment generated by the HHpred profile-profile comparison method (Söding, 2005; Alva et al, 2016). The probability of a correct match between CCDC53 and HSBP1 Hidden Markov Model (HMM) profiles was 99.9%. In the heterotrimer, the register of the single CCDC53 strand with respect to the two HSBP1 strands was provided by the HMM profile of CCDC53, which matches the one of HSBP1 with a probability of 93%.

The WASH complex has been modeled based on the structure of the WAVE complex (Chen et al, 2010). Sequence alignments for every subunit were obtained using Hhpred (Söding, 2005) and structural models, restricted to the reliable parts of the alignments, were generated with RosettaCM (Song et al, 2013). Only for the FAM21 subunit, a significant alignment could not be obtained with

19 any of the WAVE complex subunits. The domain of FAM21 spanning residues 58-128 having the highest probability to fold as a coiled-coil was used to model FAM21 as a heterotrimeric coiled-coil in the WASH complex context.

Dictyostelium discoideum

The HSBP1 gene was disrupted in Ax2 strain by homologous recombination (Supplementary Fig. 4a). Blasticidin resistant colonies were picked and tested to lack the entire HSBP1 coding region by PCR (Supplementary Fig.4b). The HSBP1 coding region was amplified from cDNA by PCR and cloned into Dictyostelium expression vectors to generate pPT606 (GFP-HSBP1), pPT607 (HSBP1- GFP), pPT608 (untagged HSBP1).

Amoebae were incubated in HL5 medium in the presence of 0.5mg/ml Tetramethylrhodamine isothiocyanate (TRITC) dextran (average MW 40kDa; Sigma) overnight in sterile flasks shaking at 100 rpm. Total fluorescence levels were acquired as previously described (Carnell et al, 2011). Cells expressing GFP-ARPC4 were incubated in SIH medium overnight with 2% dextran to enlarge vesicles for ease of imaging in glass bottom dishes (Mattek). Images were captured on a Zeiss 880 LSM inverted confocal microscope with Airyscan detector, using a 63x 1.4NA objective and deconvolved using the Zen software (Zeiss). For mass spectrometry measurements of WASH complex subunits, amoebae were boiled in 1% SDS. Lysates were reduced using dithiothreitol and subsequently alkylated with iodoacetamide. Trichloroacetic acid precipitated pellets were washed with water, resuspended in 8M urea buffer, diluted 1 to 10, and digested with trypsin. Digested peptides solutions were labelled on a Sep-Pak reverse phase C18 cartridge with light and medium dimethyl. Desalted peptides were then mixed, and separated by a 2.1 mm RP-HPLC column. Fractions were dried down and injected into an Orbitrap Velos mass spectrometer and acquired for 90 minutes each.

Statistics Statistical analysis and plotting of the results were carried out with GraphPad software (Prism v6.01). Data are expressed as mean ± sem with respect to the number of independent experiments, unless mentioned otherwise. ANOVA followed by post hoc Dunnett’s multiple-comparison tests is performed to analyze significant differences between more than 2 groups. To analyze the difference between two groups, a paired Student’s t-test was performed. Differences were considered significant at confidence levels greater than 95% (two-tailed). Four levels of statistical significance are distinguished: *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

20 Acknowledgments

We thank David Cornu from the proteomics platform (P3S-SICAPS), Christophe Velours from the protein-protein interaction platform. We thank Prof. Craig Crews from Yale University for the kind gift of HaloPROTAC3. We also thank Emmanuel Derivery and Philippe Chavrier for critical reading of the manuscript. This work was supported in AG's group by grants from the Agence Nationale de la Recherche (ANR ANR-15-CE13-0016-01), from the fondation ARC pour la Recherche sur le Cancer (PGA120140200831), from Institut National du Cancer (INCA_6521). MEL acknowledges the support from RFBR number 16-34-01316. Analyses of tumor samples was supported by the Russian Science Foundation grant 14-15-00318 with an equipment funded by the Tomsk regional common use center. EVD was supported by Tomsk State University Competitiveness Improvement Program. SPV was supported by PhD fellowships from Ministère de l'Enseignement Supérieur et de la Recherche for the first 3 years and from Ligue Nationale contre le Cancer for the 4th year.

Author contributions SPV performed most experiments and drafted the manuscript. AG supervised the work and wrote the manuscript. RG contributed structural models. PAT and RHI contributed experiments with Dictyostelium. SL performed the quantitative mass spectrometry experiment using amoeba extracts. GL performed the initial proteomics experiment that identified HSBP1 as a CCDC53 partner. VH performed DNA cloning, obtained and purified FAM21 and CCDC53 antibodies. NM performed the FACS analysis. CLS, YM and ES produced recombinant protein and performed SEC-MALS analysis. SV and IB contributed qRT-PCR measurements of HSBP1 in the retrospective cohort of breast cancer patients. EVD and NVC provided tumor extracts. LAT performed immunofluorescence on tumor biopsies. MEL and AYA contributed HSBP1 immunofluorescence in normal breast tissue.

Conflict of interest The authors declare that they have no conflict of interest.

21 References

Alberts B (1998) The cell as a collection of protein machines: preparing the next generation of molecular biologists. Cell 92: 291–294

Alva V, Nam S-Z, Söding J & Lupas AN (2016) The MPI bioinformatics Toolkit as an integrative platform for advanced protein sequence and structure analysis. Nucleic Acids Research 44: W410–5

Buckley DL, Raina K, Darricarrere N, Hines J, Gustafson JL, Smith IE, Miah AH, Harling JD & Crews CM (2015) HaloPROTACS: Use of Small Molecule PROTACs to Induce Degradation of HaloTag Fusion Proteins. ACS Chem. Biol. 10: 1831–1837

Budenholzer L, Cheng CL, Li Y & Hochstrasser M (2017) Proteasome Structure and Assembly. J Mol Biol: 1–70

Carnell M, Zech T, Calaminus SD, Ura S, Hagedorn M, Johnston SA, May RC, Soldati T, Machesky LM & Insall RH (2011) Actin polymerization driven by WASH causes V-ATPase retrieval and vesicle neutralization before exocytosis. J Cell Biol 193: 831–839

Caswell PT, Spence HJ, Parsons M, White DP, Clark K, Cheng KW, Mills GB, Humphries MJ, Messent AJ, Anderson KI, McCaffrey MW, Ozanne BW & Norman JC (2007) Rab25 associates with alpha5beta1 integrin to promote invasive migration in 3D microenvironments. Dev Cell 13: 496–510

Chen B, Padrick SB, Henry L & Rosen MK (2014) Biochemical reconstitution of the WAVE regulatory complex. Methods Enzymol 540: 55–72

Chen Z, Borek D, Padrick SB, Gomez TS, Metlagel Z, Ismail AM, Umetani J, Billadeau DD, Otwinowski Z & Rosen MK (2010) Structure and control of the actin regulatory WAVE complex. Nature 468: 533–538

Dang I, Gorelik R, Sousa-Blin C, Derivery E, Guérin C, Linkner J, Nemethova M, Dumortier JG, Giger FA, Chipysheva TA, Ermilova VD, Vacher S, Campanacci V, Herrada I, Planson A-G, Fetics S, Henriot V, David V, Oguievetskaia K, Lakisic G, et al (2013) Inhibitory signalling to the Arp2/3 complex steers cell migration. Nature 503: 281–284

De Franceschi N, Hamidi H, Alanko J, Sahgal P & Ivaska J (2015) Integrin traffic - the update. J Cell Sci 128: 839–852

Deng Z-H, Gomez TS, Osborne DG, Phillips-Krawczak CA, Zhang J-S & Billadeau DD (2015) Nuclear FAM21 participates in NF-κB-dependent gene regulation in pancreatic cancer cells. J Cell Sci 128: 373–384

Derivery E & Gautreau A (2010a) Assaying WAVE and WASH complex constitutive activities toward the Arp2/3 complex. Methods Enzymol 484: 677–695

Derivery E & Gautreau A (2010b) Generation of branched actin networks: assembly and regulation of the N-WASP and WAVE molecular machines. Bioessays 32: 119–131

Derivery E, Fink J, Martin D, Houdusse A, Piel M, Stradal TE, Louvard D & Gautreau A (2008) Free Brick1 is a trimeric precursor in the assembly of a functional wave complex. PLoS ONE 3: e2462

Derivery E, Lombard B, Loew D & Gautreau A (2009a) The Wave complex is intrinsically inactive. Cell Motil Cytoskeleton 66: 777–790

22 Derivery E, Sousa C, Gautier JJ, Lombard B, Loew D & Gautreau A (2009b) The Arp2/3 activator WASH controls the fission of endosomes through a large multiprotein complex. Dev Cell 17: 712–723

Duleh SN & Welch MD (2012) Regulation of integrin trafficking, cell adhesion, and cell migration by WASH and the Arp2/3 complex. Cytoskeleton (Hoboken) 69: 1047–1058

Euteneuer U & Schliwa M (1992) Mechanism of centrosome positioning during the wound response in BSC-1 cells. J Cell Biol 116: 1157–1166

Farina F, Gaillard J, Guérin C, Couté Y, Sillibourne J, Blanchoin L & Théry M (2016) The centrosome is an actin-organizing centre. Nat Cell Biol 18: 65–75

Gautreau A, Oguievetskaia K & Ungermann C (2014) Function and regulation of the endosomal fusion and fission machineries. Cold Spring Harb Perspect Biol 6:

Godinho SA, Picone R, Burute M, Dagher R, Su Y, Leung CT, Polyak K, Brugge JS, Théry M & Pellman D (2014) Oncogene-like induction of cellular invasion from centrosome amplification. Nature 510: 167–171

Gomez TS & Billadeau DD (2009) A FAM21-containing WASH complex regulates retromer- dependent sorting. Dev Cell 17: 699–711

Gorelik R & Gautreau A (2014) Quantitative and unbiased analysis of directional persistence in cell migration. Nature Protocols 9: 1931–1943

Guo J, Yang Z, Song W, Chen Q, Wang F, Zhang Q & Zhu X (2006) Nudel contributes to microtubule anchoring at the mother centriole and is involved in both dynein-dependent and - independent centrosomal protein assembly. Mol Biol Cell 17: 680–689

Harbour ME, Breusegem SY & Seaman MNJ (2012) Recruitment of the endosomal WASH complex is mediated by the extended ‘tail’ of Fam21 binding to the retromer protein Vps35. Biochem. J. 442: 209–220

Harbour ME, Breusegem SYA, Antrobus R, Freeman C, Reid E & Seaman MNJ (2010) The cargo- selective retromer complex is a recruiting hub for protein complexes that regulate endosomal tubule dynamics. J Cell Sci 123: 3703–3717

Helfer E, Harbour ME, Henriot V, Lakisic G, Sousa-Blin C, Volceanov L, Seaman MNJ & Gautreau A (2013) Endosomal recruitment of the WASH complex: active sequences and mutations impairing interaction with the retromer. Biol. Cell 105: 191–207

Horzum U, Ozdil B & Pesen-Okvur D (2014) Step-by-step quantitative analysis of focal adhesions. MethodsX 1: 56–59

Innocenti M, Zucconi A, Disanza A, Frittoli E, Areces LB, Steffen A, Stradal TEB, Di Fiore PP, Carlier M-F & Scita G (2004) Abi1 is essential for the formation and activation of a WAVE2 signalling complex. Nat Cell Biol 6: 319–327

Iwaya K, Norio K & Mukai K (2007) Coexpression of Arp2 and WAVE2 predicts poor outcome in invasive breast carcinoma. Mod Pathol 20: 339–343

Jia D, Gomez TS, Billadeau DD & Rosen MK (2012) Multiple repeat elements within the FAM21 tail link the WASH actin regulatory complex to the retromer. Mol Biol Cell 23: 2352–2361

Jia D, Gomez TS, Metlagel Z, Umetani J, Otwinowski Z, Rosen MK & Billadeau DD (2010) WASH

23 and WAVE actin regulators of the Wiskott-Aldrich syndrome protein (WASP) family are controlled by analogous structurally related complexes. Proc Natl Acad Sci U S A 107: 10442– 10447

Johnston JA, Ward CL & Kopito RR (1998) Aggresomes: a cellular response to misfolded proteins. J Cell Biol 143: 1883–1898

Kunda P, Craig G, Dominguez V & Baum B (2003) Abi, Sra1, and Kette Control the Stability and Localization of SCAR/WAVE to Regulate the Formation of Actin-Based Protrusions. Curr Biol 13: 1867–1875

Linkner J, Witte G, Stradal T, Curth U & Faix J (2011) High-resolution X-ray structure of the trimeric Scar/WAVE-complex precursor Brk1. PLoS ONE 6: e21327

Liu X, Xu L, Liu Y, Tong X, Zhu G, Zhang XC, Li X & Rao Z (2009) Crystal structure of the hexamer of human heat shock factor binding protein 1. Proteins 75: 1–11

Lomakin AJ, Lee K-C, Han SJ, Bui DA, Davidson M, Mogilner A & Danuser G (2015) Competition for actin between two distinct F-actin networks defines a bistable switch for cell polarization. Nat Cell Biol 17: 1435–1445

Menzl I, Lebeau L, Pandey R, Hassounah NB, Li FW, Nagle R, Weihs K & McDermott KM (2014) Loss of primary cilia occurs early in breast cancer development. Cilia 3: 7

Molinie N & Gautreau A (2017) The Arp2/3 regulatory system and its deregulation in cancer. Physiological Reviews In press:

Monteiro P, Rossé C, Castro-Castro A, Irondelle M, Lagoutte E, Paul-Gilloteaux P, Desnos C, Formstecher E, Darchen F, Perrais D, Gautreau A, Hertzog M & Chavrier P (2013) Endosomal WASH and exocyst complexes control exocytosis of MT1-MMP at invadopodia. J Cell Biol 203: 1063–1079

Nagel BM, Bechtold M, Rodriguez LG & Bogdan S (2017) Drosophila WASH is required for integrin-mediated cell adhesion, cell motility and lysosomal neutralization. J Cell Sci 130: 344– 359

Obino D, Farina F, Malbec O, Sáez PJ, Maurin M, Gaillard J, Dingli F, Loew D, Gautreau A, Yuseff M-I, Blanchoin L, Théry M & Lennon-Duménil A-M (2016) Actin nucleation at the centrosome controls lymphocyte polarity. Nat Commun 7: 10969

Park L, Thomason PA, Zech T, King JS, Veltman DM, Carnell M, Ura S, Machesky LM & Insall RH (2013) Cyclical action of the WASH complex: FAM21 and capping protein drive WASH recycling, not initial recruitment. Dev Cell 24: 169–181

Piotrowski JT, Gomez TS, Schoon RA, Mangalam AK & Billadeau DD (2013) WASH Knockout T Cells Demonstrate Defective Receptor Trafficking, Proliferation, and Effector Function. Mol Cell Biol 33: 958–973

Pollard TD (2007) Regulation of actin filament assembly by Arp2/3 complex and formins. Annu. Rev. Biophys. Biomol. Struct. 36: 451–477

Rotty JD, Wu C & Bear JE (2013) New insights into the regulation and cellular functions of the ARP2/3 complex. Nat Rev Mol Cell Biol 14: 7–12

Rouiller I, Xu X-P, 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

24 Biol 180: 887–895

Rousseau A & Bertolotti A (2016) An evolutionarily conserved pathway controls proteasome homeostasis. Nature 536: 184–189

Roy L & Case MA (2009) Electrostatic determinants of stability in parallel 3-stranded coiled coils. Chem. Commun. (Camb.): 192–194

Sahara K, Kogleck L, Yashiroda H & Murata S (2014) The mechanism for molecular assembly of the proteasome. Adv Biol Regul 54: 51–58

Satyal SH, Chen D, Fox SG, Kramer JM & Morimoto RI (1998) Negative regulation of the heat shock transcriptional response by HSBP1. Genes Dev 12: 1962–1974

Schlüter K, Waschbüsch D, Anft M, Hügging D, Kind S, Hänisch J, Lakisic G, Gautreau A, Barnekow A & Stradal TEB (2014) JMY is involved in anterograde vesicle trafficking from the trans-Golgi network. Eur J Cell Biol 93: 194–204

Schubert U, Anton LC, Gibbs J, Norbury CC, Yewdell JW & Bennink JR (2000) Rapid degradation of a large fraction of newly synthesized proteins by proteasomes. Nature 404: 770–774

Semba S, Iwaya K, Matsubayashi J, Serizawa H, Kataba H, Hirano T, Kato H, Matsuoka T & Mukai K (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: 2449– 2454

Shafaq-Zadah M, Gomes-Santos CS, Bardin S, Maiuri P, Maurin M, Iranzo J, Gautreau A, Lamaze C, Caswell P, Goud B & Johannes L (2016) Persistent cell migration and adhesion rely on retrograde transport of β(1) integrin. Nat Cell Biol 18: 54–64

Song Y, DiMaio F, Wang RY-R, Kim D, Miles C, Brunette T, Thompson J & Baker D (2013) High- resolution comparative modeling with RosettaCM. Structure 21: 1735–1742

Söding J (2005) Protein homology detection by HMM–HMM comparison. Bioinformatics 7: 951–960

Steffen A, Faix J, Resch GP, Linkner J, Wehland J, Small JV, Rottner K & Stradal TEB (2006) Filopodia formation in the absence of functional WAVE- and Arp2/3-complexes. Mol Biol Cell 17: 2581–2591

Temkin P, Lauffer B, Jäger S, Cimermancic P, Krogan NJ & Zastrow von M (2011) SNX27 mediates retromer tubule entry and endosome-to-plasma membrane trafficking of signalling receptors. Nat Cell Biol 13: 715–721

Théry M, Racine V, Piel M, Pépin A, Dimitrov A, Chen Y, Sibarita J-B & Bornens M (2006) Anisotropy of cell adhesive microenvironment governs cell internal organization and orientation of polarity. Proc Natl Acad Sci U S A 103: 19771–19776

Ueda M, Gräf R, MacWilliams HK, Schliwa M & Euteneuer U (1997) Centrosome positioning and directionality of cell movements. Proc Natl Acad Sci U S A 94: 9674–9678

Wakida NM, Botvinick EL, Lin J & Berns MW (2010) An intact centrosome is required for the maintenance of polarization during directional cell migration. PLoS ONE 5: e15462

Wan Y, Yang Z, Guo J, Zhang Q, Zeng L, Song W, Xiao Y & Zhu X (2012) Misfolded Gβ is recruited to cytoplasmic dynein by Nudel for efficient clearance. Cell Res 22: 1140–1154

25 White DP, Caswell PT & Norman JC (2007) alpha v beta3 and alpha5beta1 integrin recycling pathways dictate downstream Rho kinase signaling to regulate persistent cell migration. J Cell Biol 177: 515–525

Woolfson DN (2005) The design of coiled-coil structures and assemblies. Adv. Protein Chem. 70: 79– 112

Wu S, Ma L, Wu Y, Zeng R & Zhu X (2012) Nudel is crucial for the WAVE complex assembly in vivo by selectively promoting subcomplex stability and formation through direct interactions. Cell Res 22: 1270–1284

Zech T, Calaminus SDJ, Caswell P, Spence HJ, Carnell M, Insall RH, Insall RH, Norman J, Machesky LM & Machesky LM (2011) The Arp2/3 activator WASH regulates α5β1-integrin-mediated invasive migration. J Cell Sci 124: 3753–3759

Zhang L, Luga V, Armitage SK, Musiol M, Won A, Yip CM, Plotnikov SV & Wrana JL (2016) A lateral signalling pathway coordinates shape volatility during cell migration. Nat Commun 7: 11714

26 Figure legends

Figure 1. HSBP1 is a specific partner of the CCDC53 subunit of the WASH complex. (A) Identification of HSBP1. 293 stable cell lines expressing PC-mCherry tagged subunits of the WASH complex were subjected to PC immunoprecipitations. Immunoprecipitates were resolved by SDS- PAGE and stained with colloidal coomassie. * indicates the precipitated bait. ° indicates the position of a faint protein partner, which appears to co-immunoprecipitate with CCDC53, but not with the other WASH subunits. This protein was identified as HSBP1 by mass spectrometry. M: molecular weight markers in kDa. (B) 293T cells were transiently transfected with plasmids expressing FLAG- HSBP1 and PC-GFP tagged subunits of the WASH complex as indicated. GFP precipitation of WASH complex subunits confirms the specific co-immunoprecipitation of HSBP1 with the CCDC53 subunit. (C) His tagged full-length CCDC53 and HSBP1 were purified from E. coli. Purity was assessed by coomassie staining. (D) SEC–MALS analysis of CCDC53, HSBP1 or of a complex of the two proteins. CCDC53 and HSBP1 proteins are both trimeric. When mixed, His tagged CCDC53 and untagged HSBP1 spontaneously form a heterotrimer that contains a single molecule of the CCDC53 subunit of the WASH complex.

Figure 2. Structural model of the CCDC53-HSBP1 heterotrimer. (A) Ribbon representation of trimeric coiled-coil assemblies formed by HSBP1, CCDC53 or the mixed heterotrimer. The HSBP1 homotrimer comes from its X-ray structure (PDB code 3CI9). The CCDC53 homotrimer is modeled based on the X-ray structure of Brk1, the distantly related subunit in the WAVE complex. The CCDC53-HSBP1 heterotrimer with 1:2 stoichiometry is modeled based on the two previous structures. (B) Focus on the coiled-coil heptads of the CCDC53-HSBP1 heterotrimer. Residues buried in the trimer core of each heptad (labeled from h1 to h5) and residues that form salt-bridges are shown as sticks. The E-K/R salt-bridges that were shown in model systems to stabilize trimeric coiled-coils, are highlighted by a green asterisk in heptads 4 and 5.

Figure 3. HSBP1 promotes the assembly of the WASH complex. (A) MDA-MB-231 clones stably transfected with plasmids expressing control or HSBP1 targeting shRNAs were obtained. HSBP1 depleted cells (shHSBP1) display significant down-regulation of the CCDC53 and WASH subunits compared to control cells (shCtrl). Mean ± s.e.m. of densitometric signals; 3 independent experiments; ANOVA, *P<0.05, **P<0.01. (B) The stable MDA-MB-231 cell line expressing FLAG-HaloTag- WASH (FHT-WASH) was analyzed by western blot. Overexpression of tagged WASH induces up- regulation of CCDC53, whereas FAM21 and Strumpellin levels are not affected. Tagged WASH replaces the endogenous WASH in the WASH complex. Stable MDA-MB-231 cells expressing FHT- WASH were treated for 24 h with HaloPROTAC3, a small molecule that degrades HaloTag tagged proteins (Buckley et al, 2015). FHT-WASH degradation induced CCDC53 down-regulation and

27 reappearance of endogenous WASH. FAM21 and Strumpellin levels were not affected. (C) HaloPROTAC3 wash-out for the indicated time allows to monitor subunit build-up around FHT- WASH by FLAG immunoprecipitations. Upon HSBP1 depletion, the association of WASH with CCDC53 is delayed. (D) Structural model of WASH complex assembly. HSBP1 promotes WASH complex assembly by dissociating the CCDC53 homotrimeric precursor and delivering a single CCDC53 molecule to the assembling WASH complex. Based on the analogous structure of the WAVE complex, a new heterotrimeric coiled coil within the WASH complex, where CCDC53, WASH and FAM21 contribute one strand each is proposed.

Figure 4. Reduced adhesions of HSBP1 depleted MDA-MB-231 cells. (A) HSBP1 depleted clones or control were immunostained for α5, β1 integrin (green) and DNA (blue). HSBP1 depleted clones display a less developed array of focal adhesions. Scale bar: 10 µm. (B) Reduced number of focal adhesions per cell and reduced size of focal adhesions in HSBP1 depleted clones (at least 10 cells per condition, 2 independent experiments). (C) FACS analysis reveals reduced levels of α5β1 integrins at the surface of HSBP1-depleted clones (2 independent experiments). (D) HSBP1 depleted cells adhere less efficiently to fibronectin (FN) and collagen type I. 3 independent experiments, mean ± s.e.m.; ANOVA **P<0.01, ***P<0.001, ****P<0.0001.

Figure 5. Impaired migration and invasion in HSBP1 depleted MDA-MB-231 cells. (A) Cell translocation through pores of transwell filters covered either with fibronectin (migration) or with a thick collagen gel supplemented with fibronectin (invasion). 4 independent experiments; data are mean ± s.e.m., ANOVA (B) Single cell trajectories, mean square displacement (MSD), cell speed and directional persistence are displayed. HSBP1 depleted clones are significantly different from the control. At least 10 cells were quantified per condition, 3 independent experiments, 2-way ANOVA. ****P<0.0001.

Figure 6. Decreased polarity of HSBP1 depleted MDA-MB-231 cells. (A) Videomicroscopy of cells plated on a fibronectin-coated glass slide reveals that HSBP1 depleted cells are able to form a lamellipodium, but the lamellipodium they form poorly polarizes. Scale bar: 10 μm. (B) At all time points, HSBP1 depleted cells remain more circular than control cells, as illustrated by the circularity index or the aspect ratio, i.e. the ratio of the cell long axis over the short one. The aspect ratio of each control cell is highly irregular, as a consequence of transient loss of polarity, when the cell stops migrating. Hence the 'volatility' of the aspect ratio is significantly lower in HSBP1 depleted cells, which remain unpolarized. 3 independent experiments were performed with similar qualitative results. Results from the different experiments were pooled to reach at least 20 cells per conditions. Data are mean ± s.e.m.; ANOVA ***P<0.001, ****P<0.0001.

28 Figure 7. The role of HSBP1 in assembling functional WASH complexes is conserved in Dictyostelium discoideum. (A) WASH is downregulated in the HSBP1 knock-out (KO) amoeba. This effect is rescued by HSBP1 re-expression in the KO background. The WASH KO amoeba demonstrates the specificity of the WASH antibody. (B) The indicated subunits were detected in wild- type (WT) and HSBP1 KO amoebae and their levels compared using mass spectrometry after differential labeling of tryptic peptides. WASH is specifically down-regulated, when compared to the other subunits of the WASH complex, FAM21, Strumpellin and SWIP. (C) The recruitment of the Arp2/3 complex at surface of intracellular vesicles, as observed by confocal microscopy of GFP- ARPC4, is severely impaired in HSBP1 KO amoebae, as in WASH KO amoebae. At least, 80 vesicles were quantified per condition. (D) Exocytosis of fluorescent dextran is impaired in HSBP1 KO amoebae, as in WASH KO amoebae, and is rescued by HSBP1 re-expression in the KO background. The time course of fluorescence decay indicates a profound exocytosis defect, not a mere delay. Pictures were acquired by fluorescence and DIC microscopy after 5 h of fluorescence dextran wash- out. 4 independent experiments, data are mean ± s.e.m., ANOVA ****P<0.0001. Scale bars: 10 µm.

Figure 8. HSBP1 localizes to the centrosome. (A) MDA-MB-231 cells were stained with HSBP1 and γ-tubulin antibodies and DAPI to stain nuclei. HSBP1 is associated with the pericentriolar material stained by γ-tubulin. (B) HSBP1 is also associated with the pericentriolar material in tissue sections from human breast. The HSBP1 staining is more elongated shape in normal tissue than in mammary carcinomas. All 7 biopsies tested of breast tumors displaying both the carcinoma and adjacent normal tissue gave similar stainings. Scale bars: 10 µm.

Figure 9. HSBP1 over-expression in breast carcinomas is associated with increased levels of WASH complex and poor prognosis. (A) HSBP1 expression in 446 breast tumors was measured by qRT-PCR and compared to its expression in normal tissue. High HSBP1 expression is associated with poor prognosis. Metastasis-Free Survival (MFS) in the retrospective cohort is plotted using the Kaplan-Meier representation (optimal cut-off of 1.07-fold the normal expression, 287 patients in the High group, 159 patients in the Low group, P=0.020 using the log-rank test). (B) Extracts were prepared from 13 paired samples corresponding to breast tumors (T) and adjacent normal tissue (N) from the same patients. Extracts were normalized for total protein levels, analyzed by western blot using the indicated antibodies and the signal quantified by densitometry. In tumors exhibiting HSBP1 overexpression, the WASH complex subunits CCDC53, WASH and Strumpellin, but not FAM21, were also significantly overexpressed (paired t-test; **P<0.01, ***P<0.001). p150Glued is a loading control.

Supplementary information

Table of content

Supplementary Figure 1. Structural models of the HSBP1 homotrimer and of the HSBP1-CCDC53 heterotrimer. 2

Supplementary Figure 2. CCDC53 and WASH are down-regulated in HeLa cells upon HSBP1 depletion. 3

Supplementary Figure 3. Impaired 3D migration of HSBP1 depleted MDA-MB-231 cells. 4

Supplementary Figure 4. Isolation HSBP1 KO amoebae and characterization of HSBP1 in the amoeba. 5

Supplementary Figure 5. Specificity of HSBP1 centrosomal staining. 7

Supplementary Figure 6. Mammary carcinoma cells co-express HSBP1, CCDC53 and WASH 8

Legend to Supplementary Movies 10

Supplementary Table 1: Relationship between HSBP1 transcripts levels and clinical parameters in a series of 446 breast cancer patients. 11

Supplementary methods 12

Supplementary references 14

Figure S1. Structural models of the HSBP1 homotrimer and of the HSBP1-CCDC53 heterotrimer. (A) Superimposition of the HSBP1 homotrimers that crystallized as a dimer of homotrimers (PDB code 3CI9). HSBP1 homotrimeric structures exhibit a deviation from the parallel orientation of the helices at the level of a buried and invariant phenylalanine leading to different orientations of helix termini. (B) Sequence logo representing the conservation of different positions in the alignments of HSBP1 and CCDC53 orthologs. F27 is the only invariant position among all HSBP1 orthologs. The HSBP1-CCDC53 heterotrimer is more relaxed due to a cysteine residue instead of the bulky F27. The two E(e)-K/R(g') salt-bridges that stabilize trimeric coiled coils positions are indicated by green lines and asterisks. The residues involved in these salt bridges are also among the most conserved in the different orthologs.

Figure S2. CCDC53 and WASH are down-regulated in HeLa cells upon HSBP1 depletion. HeLa cells, transfected with non-targeting siRNAs or siRNAs targeting HSBP1, were analyzed by western blot using the indicated antibodies and signals were quantified by densitometry. Mean ± s.e.m of 3 independent experiments; ANOVA **P<0.01, ***P<0.001, ****P<0.0001.

Figure S3. Impaired 3D migration of HSBP1 depleted MDA-MB-231 cells. Cells were embedded in a collagen gel containing fibronectin. Top panels display single-cell trajectories. Mean Square Displacement (MSD) analysis shows that HSBP1 depleted cells explore a smaller territory than control cells. This effect might be accounted for by decreased speed and directional persistence. Globally, however, cell migration is less persistent in 3D than in 2D. At least 10 cells were tracked per condition, in 3 independent experiments; data are mean ± s.e.m; 2-way ANOVA **P<0.01; ****P < 0.0001.

Figure S4. Isolation HSBP1 KO amoebae and characterization of HSBP1 in the amoeba. (A) The whole ORF of HSBP1 was replaced by the gene encoding Blasticidin resistance (BSR) upon homologous recombination. Approximately 1.2kb of genomic DNA to the 5’ and 3’ sides of the HSBP1 gene were amplified by PCR and ligated on either side of a blasticidin- resistance cassette in a Dictyostelium cloning vector to create a knockout vector. The knockout cassette was cut out using appropriate restriction enzymes and linear DNA was transformed into Dictyostelium cells by electroporation. Diagnostic PCR of recombination on genomic DNA of two isolated blasticidin resistant clones. (B) Growth of HSBP1 KO is impaired in a medium containing 20% dextran. After 5 days, some HSBP1 KO amoebae accumulate multiple dense vesicles and are enlarged. This phenotype, which was previously described for WASH KO amoebae, is never observed in the parental strain. DIC microscopy, scale bar: 10 µm. (C) Incorporation of fluorescent dextran reaches a steady state, where

5 exocytosis compensates endocytosis, after 2 h in WT amoebae, but a plateau is not yet reached after 5 h in HSBP1 or WASH KO amoebae. (D) Localization of HSBP1-GFP and GFP-HSBP1 in Dictyostelium amoeba. In both cases, HSBP1 localizes to central dot-like structures, which correspond to centrosomes, as indicated by -tubulin-mRFP colocalization.

Figure S5. Specificity of HSBP1 centrosomal staining. (A) Centrosomal staining of HSBP1 is lost in the HSBP1 depleted MDA-MB-231 clones. Scale bar: 10 µm. (B) Quantification of the HSBP1 intensity around the pericentriolar material stained by γ-tubulin. At least 15 cells per condition, 3 independent experiments. ****P<0.0001 (C) HSBP1 and γ-tubulin are associated in the same z planes. Confocal planes were distant of 0.25 µm.

Figure S6. Mammary carcinoma cells co-express HSBP1, CCDC53 and WASH. Tissue sections of a mammary carcinoma overexpressing HSBP1 (from patient #4 in Fig.9) is co- stained with DAPI, cytokeratin 7 (CK7), a marker of carcinoma cells, HSBP1 and with CCDC53 (A) or with WASH (B). Centrosomal staining of HSBP1 is not visible in these immunofluorescences performed from formalin fixed paraffin embedded tumors. The centrosomal staining requires cryosections and methanol/acetone fixation. Scale bar: 20 m.

Legend to Supplementary Movies

Movie S1. Impaired 2D migration of HSBP1 depleted MDA-MB-231 cells. Cells were allowed to spread on fibronectin-coated µ-slide (Ibidi) for 4-5 h, before being imaged for 10 h with pictures taken every 10 min using phase contrast and a Plan-Apochromat 20x/0.80 air objective. Scale bar: 20 µm.

Movie S2. Impaired 3D migration of HSBP1 depleted MDA-MB-231 cells. Cells were sandwiched between two collagen gels in order to have most cells in a single plane of view at the beginning of movie acquisition. Cells were imaged for 35 h with pictures taken every 20 min using phase contrast and a Plan-Apochromat 20x/0.80 air objective. Scale bar: 20 µm

Table S1 : Relationship between HSBP1 transcript levels and clinical parameters in breast cancer patients

Total population HSBP1 transcript levels p (%) relative to normal breast

Total 446 (100.0) 1.33 (0.40-21.3)

Age 50 94 (21.1) 1.37 (0.69-21.3) 0.25 (NS) a >50 352 (78.9) 1.30 (0.40-6.75)

SBR histological grade c,d I 57 (13.0) 1.10 (0.63-5.51) 0.0081 b II 222 (50.8) 1.31 (0.40-21.3) III 158 (36.2) 1.48 (0.55-5.21)

Lymph node status e 0 117 (26.3) 1.16 (0.40-5.51) 0.082 (NS) b 1-3 231 (51.9) 1.32 (0.53-6.75) >3 97 (21.8) 1.39 (0.61-21.3)

Macroscopic tumor size f 25mm 218 (49.8) 1.24 (0.40-5.55) 0.16 (NS) a >25mm 220 (50.2) 1.37 (0.55-21.3)

ER status Negative 115 (25.8) 1.37 (0.55-5.55) 0.38 (NS) a Positive 331 (74.2) 1.31 (0.40-21.3)

PR status Negative 191 (42.8) 1.37 (0.40-6.75) 0.22 (NS) a Positive 255 (57.2) 1.28 (0.53-21.3)

ERBB2 status Negative 353 (79.1) 1.23 (0.40-21.3) <0.0001 a Positive 93 (20.9) 1.62 (0.55-5.55)

Molecular subtypes RH- ERBB2- 68 (15.2) 1.24 (0.57-4.66) 0.0005 b 42 (9.4) 1.48 (0.55-5.55) RH- ERBB2+ 285 (63.9) 1.23 (0.40-21.3) RH+ ERBB2- 51 (11.4) 1.63 (0.64-5.51) RH+ ERBB2+

The median and range in parenthesis is given for HSBP1 transcript levels. a Mann Whitney’s Test. b Kruskal Wallis’s Test. c Scarff Bloom Richardson classification. d Information available for 437 patients. e for 445 patients. f for 438 patients.

Supplementary methods

Patient cohort for mRNA analysis Samples of 446 unilateral invasive primary breast tumors excised from women managed at Institut Curie – Hôpital René Huguenin (Saint-Cloud, France) from 1978 to 2008 have been analyzed. All patients cared in our institution before 2007 were informed that their tumor samples might be used for scientific purposes and had the opportunity to decline. Since 2007, patients treated in our institution have given their approval by signed inform consent. This study was approved by the local ethics committee (Breast Group of René Huguenin Hospital). Samples were immediately stored in liquid nitrogen until RNA extraction. A tumor sample was considered suitable for our study if the proportion of tumor cells exceeded 70%. All patients (mean age 61.8 years, range 31-91 years) met the following criteria: primary unilateral nonmetastatic breast carcinoma for which complete clinical, histological and biological data were available; no radiotherapy or chemotherapy before surgery; and full follow-up at Institut Curie - Hospital René Huguenin.

Treatment (information available for 438 patients) consisted of modified radical mastectomy in 278 cases (63.9%) and breast-conserving surgery plus locoregional radiotherapy in 160 cases (36.1%). The patients had a physical examination and routine chest radiotherapy every 3 months for 2 years, then annually. Mammograms were done annually. Adjuvant therapy was administered to 360 patients, consisting of chemotherapy alone in 87 cases, hormone therapy alone in 172 cases and both treatments in 101 cases. The histological type and the number of positive axillary nodes were established at the time of surgery. The malignancy of infiltrating carcinomas was scored according to Scarff Bloom Richardson’s (SBR) histoprognostic system.

Hormone receptor (HR) (i.e. estrogen receptor alpha (ER) and progesterone receptor (PR) and human epidermal growth factor receptor 2 (ERBB2) statuses were determined at the protein level by using biochemical methods (dextran-coated charcoal method, enzyme immunoassay or immunohistochemistry) and confirmed by real-time quantitative RT-PCR assays.

The population was divided into four groups according to HRs (ERα and PR) and ERBB2 status, as follows: two luminal subtypes [HR+/ERBB2+ (n=45)] and [HR+/ERBB2- (n=195)]; an ERBB2+ subtype [HR-/ERBB2+ (n=46)] and a triple-negative subtype [HR- /ERBB2- (n=64)]. During a median follow-up of 9.1 years (range 4.3 months to 33.2 years),

12 176 patients developed metastasis. Ten specimens of adjacent normal breast tissue from breast cancer patients and normal breast tissue from women undergoing cosmetic breast surgery were used as sources of normal RNA.

Patient biopsies for western blot and immunofluorescence analyses

20 patients with invasive breast carcinoma of no special type (T1-3N0-3M0-1; grade 1-2; luminal A, B, and HER2+), between 46 and 68 years of age (mean age: 55.2±9.0), and treated in the Cancer Research Institute, Tomsk NRMC (Tomsk, Russia) between 2013 and 2015 were included. All cases were without any preoperative therapy and underwent mastectomy, radical or sectoral resection. The procedures followed in this study were in accordance with the Helsinki Declaration (1964, amended in 1975 and 1983). This study was approved by the institutional review board; all patients signed an informed consent for voluntary participation.

Breast tumors for centrosome immunofluorescence were luminal carcinomas from patients, who had undergone a mastectomy without preliminary therapy, at the N.N. Blokhin Russian Cancer Research Center (Moscow, Russia). The Institutional Review Board of N.N. Blokhin Russian Cancer Research Center approved the project. All patients, who were involved in the study gave written informed consents that their samples could be used for investigational purposes. Data were analyzed anonymously. All potential participants, who declined to participate or otherwise did not participate, were eligible for treatment (if applicable) and were not disadvantaged in any other way by not participating in the study.

qRT-PCR Total RNA was extracted from breast tissue samples by using acid-phenol guanidium. RNA quality was determined by electrophoresis through agarose gels, staining with ethidium bromide and visualization of the 18S and 28S RNA bands under ultraviolet light. Quantitative values were obtained from the cycle number (Ct value) at which the increase in the fluorescence signal associated with exponential growth of PCR products started to be detected by the laser detector of the ABI Prism 7900 sequence detection system (Perkin-Elmer Applied Biosystems, Foster City, CA), using PE biosystems analysis software according to the manufacturer’s manuals.

13 The precise amount of total RNA added to each reaction mix (based on optical density) and its quality (i.e., lack of extensive degradation) were both assessed. Transcripts of the TBP gene (Genbank accession NM_003194) encoding the TATA box-binding protein were also quantified as an endogenous RNA control. Each sample was normalized on the basis of its TBP content. Results, expressed as N-fold differences in HSBP1 gene expression relative to the TBP gene and termed “NHSBP1”, were determined as NHSBP1 = 2ΔCtsample, where the ΔCt value of the sample was determined by subtracting the average Ct value of the HSBP1 gene from the average Ct value of the TBP gene. The HSBP1 values of the breast tumor samples were subsequently normalized such that the median of the HSBP1 values for the 10 normal breast tissues was 1.

The primers for HSBP1 and TBP were chosen with the assistance of the Oligo 6.0 program (National Biosciences, Plymouth, MN). We scanned the dbEST and nr databases to confirm the total gene specificity of the nucleotide sequences chosen for the primers and the absence of single nucleotide polymorphisms. The nucleotide sequences of the primers used were as follows: HSBP1-U (5’-GGTCTGAGACATCACCGCCAAG-3’) and HSBP1-L (5’- CAGGAGTGTCTGCACCACCGA-3’) for HSBP1 gene (PCR product of 97 bp), and TBP-U (5’-TGC ACA GGA GCC AAG AGT GAA-3’) and TBP-L (5’-CAC ATC ACA GCT CCC CAC CA-3’) for the TBP gene (PCR product of 132 bp). To avoid amplification of contaminating genomic DNA, one of the two primers was placed at the junction between two exons or on two different exons. Agarose gel electrophoresis was used to verify the specificity of PCR amplicons. The conditions of cDNA synthesis and PCR were previously described (Bieche et al, 2001).

The distribution of HSBP1 mRNA levels was characterized by median values and ranges. Relationships between HSBP1 mRNA levels and clinical parameters were identified using non parametric tests, namely the Mann-Whitney U test and the Kruskal Wallis test. Differences were considered significant at confidence levels greater than 95% (p < 0.05). Metastasis-free survival (MFS) was determined as the interval between initial diagnosis and detection of the first metastasis.

Supplementary references

Bieche I, Parfait B, Laurendeau I, Girault I, Vidaud M & Lidereau R (2001) Quantification of estrogen receptor alpha and beta expression in sporadic breast cancer. Oncogene 20: 8109– 8115

14 This page intentionally left blank

85 This page intentionally left blank

86 Discussion

87 This page intentionally left blank

88 1/ HSBP1 dissociate CCDC53 trimers to promote WASH complex assembly

The WASH complex is a distant relative of the WAVE complex. It is also a stable multi-subunit molecular machine that acts as an obligate complex (Jia et al. 2010). Usually, even if one subunit is destabilized, the whole complex losses its function and gets degraded by the proteasome machinery (Derivery et al. 2009b, Jia et al. 2010). The WASH complex just like the WAVE complex is built in such a way that, its Arp2/3 activating VCA domain is masked by its subunits. Such masking of the VCA domain is mandatory for cells to avoid dominant negative effect of the NPFs. Arp2/3 complex activation in the whole cytosol by exposed VCA domain of the NPF is deleterious to the cells, as it prevents the normal Arp2/3 recruitment and activation (Machesky and Insall 1998). Such mechanism of VCA domain masking within the complex would be possible by the means of regulated molecular assembly and not by the spontaneous association of subunits. Moreover, if the stable multi-protein complexes are assembled in such a way that only the ‘law of mass action’ was at play, then these complexes should dissociate during their purification as there are dilutions induced by gel filtration columns or sucrose gradients, and but they doesn’t appear to dissociate (Gautreau et al. 2004, Derivery and Gautreau 2010a). Thus, these molecular machines should undergo regulated assemblies that are driven by extrinsic assembly factors as seen in big molecular machines like proteasomes (Sahara et al. 2014). Recently, it has been observed that the WAVE complex assembly is assisted by such assembly factor called Nudel (Wu et al. 2012). However, for the WASH complex, such assembly factors have not been found.

In search of WASH complexes assembly factor, we did proteomic screening and identified that HSBP1 is a specific partner to CCDC53 subunit of the WASH complex and not to other subunits. Its interaction with CCDC53 subunit rather than the whole complex suggested its role as an assembly factor. If it has interacted with the whole complex, it would be rather a subunit of the complex and not an assembly factor. Subsequently, we have demonstrated that HSBP1 is required for the assembly of WASH complex, hence making it the first assembly factor identified for the WASH complex.

89 Figure19. HSBP1 dissociates active HSF1 trimer to attenuate heat shock response. Inspired from Pockley, (2003); Satyal et al., (1998). As a heat shock response, inactive HSF1 monomers are translocated from cytoplasm to nucleus. In nucleus, HSF1 monomers get trimerized to become active HSF1 that binds to heat shock element in the genome. This binding results in transcription of heat shock proteins. For the attenuation of transcription process, HSBP1 dissociates the trimeric form of HSF1.

90 Heat Shock Factor Binding Protein 1, simply called HSBP1 is a small protein of 8.5 kDa mass. As the name suggests, HSBP1 binds to HSF1 (Heat Shock Factor 1) which is a transcriptional activator of the HSP proteins (heat shock proteins)(Satyal et al. 1998). During the heat shock response, HSF1 protein which is inactive as monomers are localized at the cytoplasm of the cell translocate to the nucleus. Inside the nucleus, it undergoes oligomerization and becomes a trimer through the coiled-coil region. This trimeric form of HSF1 is the active confirmation which binds to HSE (heat shock element) throughout the genome to activate the transcription of the HSP proteins (Figure 19). Once the heat shock response has to be stopped by the cell, HSBP1 comes into play where it terminates the HSF1 transcription in nucleus by dissociating the active HSF trimer.

Here, in our studies, we report a cytosolic function of the HSBP1 protein. With the help of in vitro and in vivo methods, we demonstrate that HSBP1 can dissociate the precursor trimeric form of CCDC53 to favor the WASH complex assembly. Thus, HSBP1 protein functions that are observed in the nucleus and cytosol involve a structural transition of its effector proteins which are trimers. The in silico analysis of HSBP1 crystal structure revealed that, due the presence of bulky F27 residue, by nature, HSBP1 remains as a marginally stable trimeric coiled coil. Also, this F27 residue is highly conserved and by design due to this residue, HSBP1 readily dissociates the trimeric coiled-coil proteins.

2/ HSBP1 function is conserved and is necessary for CCDC53- WASH sub-complex formation during WASH assembly pathway

HSBP1 inactivation in human cells and Dictyostelium amoeba renders the WASH complex inactive and thus, the function of HSBP1 is conserved across evolution. In both organisms, HSBP1 inactivation destabilizes few subunits but not all. More specifically it destabilizes CCDC53 and WASH but not Strumpellin, SWIP and FAM21. This is reminiscent of the fact that only CCDC53 is destabilized when the WASH is depleted and the vice versa (Jia et al. 2010). Explanation for such an observation can be that, CCDC53 and WASH subunits are direct partners in the WASH complex and they require each

91 other for their stability which is acquired by the assembly. Furthermore, we found that, by overexpressing Halotag tagged WASH or selectively destabilizing it through HaloPROTAC3 treatment, only CCDC53 subunit mirrors the amount of WASH. Also in the absence of HSBP1, CCDC53 loses its ability to get associated with the WASH subunit during the assembly. From these observations, it is evident that CCDC53-WASH forms a sub-complex in cells driven by the assembly of HSBP1 where HSBP1 dissociates the precursor CCDC53 trimer and helps in delivering a single CCDC53 molecule that eventually gets associated with the WASH subunit by an unknown mechanism.

Evidently, from these observations it is clear that, there is an existence of another sub-complex composed of Strumpellin-SWIP-FAM21 which is stable in cells. The stable presence of Strumpellin-SWIP-FAM21 sub-complex in cytosol as compared to CCDC53-WASH sub-complex might be due to the fact that CCDC53-WASH sub- complex could be deleterious to the cell which in turn is due to unmasked VCA domain of WASH protein. Thus CCDC53-WASH sub-complex has to undergo immediate consecutive assembly to mask the VCA domain or else get degraded. For such immediate assembly, Strumpellin-SWIP-FAM21 sub-complex is required to be readily available in cytosol, for that its stable existence might play a role. It is tempting to speculate that, the assembly of Strumpellin-SWIP-FAM21 sub-complex can also be executed by other extrinsic assembly factors in the cell. Hence, whether this partial complex simply represents a stable assembly intermediate or whether it has its own function and thus a partner of its own are interesting questions for further studies.

3/ HSBP1 inactivation phenocopies WASH complex inactivation in the cells

Inactivation of HSBP1 produced phenotypes very similar to the WASH complex depletion. In Dictyostelium amoeba, we created HSBP1 knock-out strains which lost the WASH complex. Similar to the WASH knock-out strain, HSBP1 knock-out also exhibited a defect in recruitment of Arp2/3 at the lysosomal vesicles surface. As a result, in both cases amoeba lost the capability to perform exocytosis of non-digestible dextran. In human cells, WASH complex and filamentous actin localize to the subdomain of early and recycling endosomes (Derivery et al. 2009b, Gomez and Billadeau 2009, Duleh and Welch 2010) that are compartments through which integrins are trafficked (Caswell et

92 al., 2009; Pellinen and Ivaska, 2006). Hence, it is well observed that, WASH depletion leads to reduced surface level of the 5 1 integrins and a decreased number of focal adhesions (Zech et al. 2011, Duleh andα Welchβ 2012) due to defective integrin trafficking. WASH depletion also reduced the invasive ability of tumor cells (Zech et al. 2011). We have observed that in all these WASH phenotypes of HSBP1 depleted human cells and Dictyostelium amoebae thus confirming that HSBP1 is required to assemble functional WASH complexes.

Strikingly, HSBP1 depleted cells displayed a defect in polarity establishment that prevented their migration even if they were able to generate membrane protrusions. Morphologically, HSBP1 depleted cells maintained a constant circularity over time which is a morphodynamic signature of unpolarized cells whereas control cells exhibited low circularity which signifies the polarization for migration. This phenotype was not previously reported upon WASH complex inactivation.

But it is known, integrin-mediated adhesions that are dynamically formed and turned over are necessary for cell dissemination and migration. As well, this polarized assembly and disassembly plays a crucial role in polarized cellular migration (Huttenlocher and Horwitz 2011, Bridgewater et al. 2012). One of the primary concepts behind cellular polarity and directional migrating is the directed transport of proteins and membranes to front of the cell (Bretscher and Aguado-Velasco 1998, Wehrle-haller 2006). Most importantly, internalized at the cells rearα5β1-end integrinsand is transported which are to recycled the front by (Figure WASH 20) complex (Laukaitis get et al. 2001).

Recent studies shows that, recycling of 5 1 integrins through their association with Rab25 (Caswell et al. 2007b) and theirα βindirect route through the Golgi by following retrograde pathway is critical for migration persistence (Shafaq-Zadah et al. 2015). Establishment and maintenance of cell polarity also relies on Rho-mediated contractility (Lomakin et al. 2015) which is regulated by the recycling of integrins (White et al. 2007). Hence, even if polarity defect has been reported for WASH inactivation, this phenotype of HSBP1 depletion is consistent with the polarity defect observed due to defective integrin recycling.

Figure20. Proposed model of WASH complex role in α5β1 focal adhesion turnover. During migration, mature -

α5β1 focal adhesion gets internalizeds from(EE). rear The endWASH of complex the cell. that This is internalized assembled at α5β1 the centrosome integrins reaches is recruited early to endosome the early endosome surface. Recruited WASH complex generates branched actin networks through Arp2/3 activation and thereby facilitates membrane scission of the early endosomes containing

α5β1front byintegrins. recycling From endosome there, sα5β1 (RE), integrinsto promote are the sorted nascent and focal delivered adhesions to the formation. migrating

94 Furthermore, we found the association of HSBP1 with centrosome in both human cells and Dictyostelium amoebae. Centrosome is primarily characterised as a major MTOC and has been implicated in cell migration, vesicle trafficking and cell polarity (Bornens 2012). Especially, centrosome represents a polarity marker oriented towards the migrating front of the cell. In many migrating cells, including fibroblasts and macrophages, the centrosome localizes between the nucleus and the leading edge (Kupfer et al. 1982, Nemere et al. 1985), suggesting that the position of the centrosome may determine the direction of the cell polarization. This inference is further supported by a study in epithelial cells, in which laser ablation of the centrosome caused a block in cell migration and cell polarization (Wakida et al. 2010). However, such observations of centrosomal positioning during cell migration do not appear to be universal and perhaps, it is cell type specific. In a study comparing centrosome position in migrating CHO and Ptk cells, the centrosome was localized towards the front of the nucleus in CHO, but not in Ptk (Yvon et al. 2002). Hence, the exact role of centrosome alone in cell polarity is not clear whereas, its association with Golgi sheds light on its role in the establishment of cell polarity.

It has been shown that the Golgi and the centrosome move together toward the leading edge of migrating cells (Bisel et al. 2008). The Golgi apparatus represents the second major mammalian MTOC which has a role to play in both microtubule nucleation and anchoring (Chabin-Brion et al. 2001, Zhu and Kaverina 2013, Rios 2014, Wu and Akhmanova 2017). In contrast to the centrosome which forms a symmetric array, Golgi- derived microtubule networks are polarized and can thus drive asymmetric vesicle transport and promote overall cell polarity (Vinogradova et al. 2009, 2012, Hurtado et al. 2011). A recent study revealed that in mesenchymal cells, the presence of Golgi- derived microtubules accelerates reorientation of the whole microtubule network, including the centrosome, in the direction of migration (Wu et al. 2016).

Hence, the centrosome potentially works in synergy with the Golgi in dictating their localization and as such, they can reorient integrin delivery to the leading edge and contribute migration persistence (Théry et al. 2006, Wakida et al. 2010, Shafaq-Zadah et al. 2015).

95 4/ HSBP1 putative role at the centrosome

We found that HSBP1 localizes at the centrosome and its localization partially -tubulin, a marker for the peri-centriolar material. Especially, one can appreciateoverlaps with this γpartial overlap in breast tissue sections. Indeed, HSBP1 staining reveals an elongated structure from the peri-centriolar material in normal breast tissue whereas this structure is not anymore elongated in tumor sections. The explanation of such observation is that, centrosome is the base of cilia in normal breast cells whereas these structures regress during cancer progression (Menzl et al. 2014). HSBP1 staining was always partially overlapping with -tubulin but appeared slightly more elongated in the presence of cilia in normal breastγ cells. Such elongated structures were not observed in breast cancer cell lines in culture (Nobutani et al. 2014), similarly, HSBP1 staining looks the same for the one described for MDA-MB-231 cells. However, it is not clear whether HSBP1 also specifically associates with the cilia and has some associated physiological role to play in cells. These remain as open questions for further investigation. Furthermore, HSBP1 also localize at the centrosome of Dictyostelium amoeba, emphasizing its conservation across the evolution.

Interestingly, WASH complex was recently reported to recruit Arp2/3 complex and generate branched actin networks at the centrosome, in addition to its main localization at the endosomes surface (Farina et al. 2016). These centrosomal branched actin networks by the WASH complex were implicated in anchoring centrosomes to the nucleus (Obino et al. 2016). Hence, in the context of HSBP1 being an assembly factor of WASH complex, its localization at centrosome suggests that centrosomes are sites of WASH complex assembly (Figure 20). The centrosome might also be a privileged site for the assembly of other multi-protein complexes such as the assembly of WAVE complex. This is because the WAVE complex assembly factor, Nudel protein also accumulates at the centrosome, in line with its regulatory activity towards dynein, the MT minus-end directed motor (Guo et al. 2006).

The centrosome is also a privileged site for degradation which is often associated with multi-protein complex assembly (Derivery and Gautreau 2010a, Wu et al. 2012). Indeed, mis-folded proteins accumulate at the centrosome in form of specific structures called aggresome, when the proteasomes are overwhelmed by the substrate to degrade

96 (Johnston et al. 1998). Aggresomes were shown to be highly regulated structures containing proteasomes and chaperones that are assembled at the centrosome in response to mis-folded or damaged proteins through dynein-mediated microtubule- dependent transport (Johnston et al. 2002, Corboy et al. 2005, Fisk 2012). Many cellular proteins destined for degradation have been found at centrosomes, including p53 and Hsp70 in cells expressing adenoviral E1A/E1B proteins (Brown et al. 1994), and misfolded nucleoprotein from the influenza virus (Antón et al. 1999). As well misfolded cystic fibrosis trans-membrane conductance regulator accumulated at centrosomes, together with the molecular chaperones thought to be responsible for its presentation to the proteasome (Johnston et al. 1998, Loo et al. 1998, Fisk 2012).

These investigations on aggresomes focused on the degradation of damaged or misfolded proteins, and contributed to a view of centrosomes as command centers where regulators mingle but do not reside (Doxsey 2001, Schatten 2008, Fisk 2012). However, with respect to proteasomal degradation, the centrosome is a critical point of execution for degradation. Interestingly, the 20S proteasome core, 19S and 11S regulatory subunits, ubiquitin, and several molecular chaperones were shown to concentrate at centrosomes and co-fractionate with -tubulin independently of aggresome formation (Wigley et al. 1999, Fisk 2012), suggestingγ that centrosome is a gathering point for proteins and acts as an assembly point for molecular machines like proteasomes.

Thus, centrosome potentially acts as a center that accumulates the subunits and assembly factors to facilitate multi-protein complex assemblies. The complex assembly that takes place at the centrosome also has a chance to be unidirectional and irreversible by restricting the intermolecular collisions to a small set of possibilities. Hence, centrosome as a site of action for the multi-protein assembly would improve the robustness of molecular assembly. The proteasome machinery that works in synergy might impose a quality control and regulate assembly intermediates within the centrosome surface which is potently deleterious to the cell if they are diffused freely in the cytosol, example –WASH or WAVE assembly intermediates with unmasked VCA domain.

97 5/ HSBP1 putatively drive cancer progression through WASH complex assembly

We found that HSBP1 is overexpressed in breast cancers and its overexpression is associated with the poor prognosis for patients. As a next step, we sought to verify whether HSBP1 modulates the WASH complex at protein levels in breast tumors. Subsequently, we documented HSBP1 overexpression in tumors compared to normal tissue in about half the patients. Fully consistent with HSBP1 role in assembling WASH complexes, this overexpression was associated with the increased levels of WASH complex subunits. This is the first time to report the WASH complex overexpression, even if its role in tumor invasion was established (Zech et al. 2011, Monteiro et al. 2013). Only the FAM21 level didn’t correlate with HSBP1 level and other subunits of the complex. Explanation for FAM21 behavior can be due to its moonlighting functional characteristics. FAM21 is not only functional within the WASH complex but also functional apart from it where it was found to regulate transcription by NF- nucleus independently (Deng et al. 2015). In Dictyostelium, FAM21 knockκB in- outthe phenotype is also quite different from the one observed in WASH knock-out amoeba (Park et al. 2013). Hence, HSBP1 is overexpressed in cancers and its overexpression is associated with the up-regulation of the WASH complex. Thus, HSBP1 might potentially drive metastasis through WASH complex assembly and this explains the overexpression of HSBP1 for being a poor prognostic factor for breast cancers.

In conclusion, during my thesis we identified HSBP1 as the first assembly factor of the WASH complex. We dissected HSBP1 mediated WASH complex assembly mechanism and characterised the consequence of HSBP1 inactivation in cell physiology, which rises due to defective WASH assembly. We have demonstrated that in vitro migration and invasion ability of breast carcinoma cells could be controlled through inactivation of HSBP1. We are first to illustrate that WASH complex is up-regulated in breast tumors and its up-regulation is associated with its assembly factor overexpression. This study reveals a mechanism that cancer cells might employ to augment molecular machines which is by the means of overexpressing assembly factors. As well, we show HSBP1 overexpression is a poor prognostic factor for breast cancer patients. Hence, these observations provide a new perspective for the assembly factors

98 as potential therapeutic targets, to control the levels of the up-regulated protein complexes that drive the tumor progression.

This page intentionally left blank

100

References

101 Abou-Kheir, W., B. Isaac, H. Yamaguchi, and D. Cox. 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. 121:379–90.

Adighibe, O., H. Turley, R. Leek, A. Harris, A. S. Coutts, N. La Thangue, K. Gatter, and F. Pezzella. 2014. JMY Protein, a Regulator of P53 and Cytoplasmic Actin Filaments, Is Expressed in Normal and Neoplastic Tissues. Virchows Arch. 465:715–722.

Alberts, B. 1998. The Cell as a Collection Overview of Protein Machines: Preparing theNext Generation of Molecular Biologists. Cell 92:1–4.

Allan, V. J. 2011. Cytoplasmic Dynein. Biochem. Soc. Trans. 39.

Allison, R., J. R. Edgar, G. Pearson, T. Rizo, T. Newton, S. Günther, F. Berner, J. Hague, J. W. Connell, J. Winkler, J. Lippincott-Schwartz, C. Beetz, B. Winner, and E. Reid. 2017. Defects in ER–endosome Contacts Impact Lysosome Function in Hereditary Spastic Paraplegia. J. Cell Biol. 216:1337–1355.

Anitei, M., and B. Hoflack. 2011. Bridging Membrane and Cytoskeleton Dynamics in the Secretory and Endocytic Pathways. Nat. Cell Biol. 14:11–19.

Antón, L. C., U. Schubert, I. Bacík, M. F. Princiotta, P. A. Wearsch, J. Gibbs, P. M. Day, C. Realini, M. C. Rechsteiner, J. R. Bennink, and J. W. Yewdell. 1999. Intracellular Localization of Proteasomal Degradation of a Viral Antigen. J. Cell Biol. 146:113–24.

Aspenström, P. 2002. The WASP-Binding Protein WIRE Has a Role in the Regulation of the Actin Filament System Downstream of the Platelet-Derived Growth Factor Receptor. Exp. Cell Res. 279:21–33.

Aspenström, P. 2004. The Mammalian Verprolin Homologue WIRE Participates in Receptor- Mediated Endocytosis and Regulation of the Actin Filament System by Distinct Mechanisms. Exp. Cell Res. 298:485–98.

Beltzner, C. C., and T. D. Pollard. 2007. Pathway of Actin Filament Branch Formation by Arp2/3 Complex. J. Biol. Chem. 283:7135–44.

Benesch, S., S. Lommel, A. Steffen, T. E. B. 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. 277:37771–6.

Benesch, S., S. Polo, F. P. L. Lai, K. I. Anderson, T. E. B. Stradal, J. Wehland, and K. Rottner. 2005. N-WASP Deficiency Impairs EGF Internalization and Actin Assembly at Clathrin-Coated Pits. J. Cell Sci. 118.

Bisel, B., Y. Wang, J.-H. Wei, Y. Xiang, D. Tang, M. Miron-Mendoza, S. Yoshimura, N. Nakamura, and J. Seemann. 2008. ERK Regulates Golgi and Centrosome Orientation towards the Leading Edge through GRASP65. J. Cell Biol. 182:837–43.

Blagg, S. L., M. Stewart, C. Sambles, and R. H. Insall. 2003. PIR121 Regulates Pseudopod Dynamics and SCAR Activity in Dictyostelium. Curr. Biol. 13:1480–1487.

Blanchoin, L., K. J. Amann, H. Higgs, J.-B. Marchand, D. A. Kaiser, and T. D. Pollard. 2000. Direct Observation of Dendritic Actin Filament Networks Nucleated by Arp2/3 Complex and WASP/Scar Proteins. Nature 404:1007–1011.

Blanchoin, L., and T. D. Pollard. 1999. Mechanism of Interaction of Acanthamoeba Actophorin (ADF/Cofilin) with Actin Filaments. J. Biol. Chem. 274:15538–46.

102 Bochtler, M., L. Ditzel, M. Groll, C. Hartmann, and R. Huber. 1999. The Proteasome. Annu. Rev. Biophys. Biomol. Struct. 28:295–317.

Borissenko, L., and M. Groll. 2006. 20S Proteasome and Its Inhibitors: Crystallographic Knowledge for Drug Development. Chem. Rev.

Bornens, M. 2012. The Centrosome in Cells and Organisms. Science (80-. ). 335:422–6.

Bretscher, M. S., and C. Aguado-Velasco. 1998. Membrane Traffic during Cell Locomotion. Curr. Opin. Cell Biol. 10:537–541.

Bridgewater, R. E., J. C. Norman, and P. T. Caswell. 2012. Integrin Trafficking at a Glance. J. Cell Sci. 125:3695–3701.

Brown, C. R., S. J. Doxsey, E. White, and W. J. Welch. 1994. Both Viral (Adenovirus E1B) and Cellular (Hsp 70, p53) Components Interact with Centrosomes. J. Cell. Physiol. 160:47–60.

Buck, M., W. Xu, and M. K. Rosen. 2001. Global Disruption of the WASP Autoinhibited Structure on Cdc42 Binding. Ligand Displacement as a Novel Method for Monitoring Amide Hydrogen Exchange. Biochemistry 40:14115–22.

Budenholzer, L., C. L. Cheng, Y. Li, and M. Hochstrasser. 2017. Proteasome Structure and Assembly. J. Mol. Biol.

Cold Spring Harb. Perspect. Biol.:1–14. Burd, C., and P. J. Cullen. 2014. Retromer : A Master Conductor of Endosome Sorting. Cai, L., A. M. Makhov, and J. E. Bear. 2007. F-Actin Binding Is Essential for Coronin 1B Function in Vivo. J. Cell Sci. 120.

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.

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 134:148–161.

Campellone, K. G., and M. D. Welch. 2010. A Nucleator Arms Race: Cellular Control of Actin Assembly. Nat. Rev. Mol. Cell Biol. 11:237–51.

Cao, W., J. P. Goodarzi, and E. M. De La Cruz. 2006. Energetics and Kinetics of Cooperative Cofilin–Actin Filament Interactions. J. Mol. Biol. 361:257–267.

Carlier, M. F., P. Nioche, I. Broutin-L’Hermite, R. Boujemaa, C. Le Clainche, C. Egile, C. Garbay, A. Ducruix, P. Sansonetti, and D. Pantaloni. 2000. GRB2 Links Signaling to Actin Assembly by Enhancing Interaction of Neural Wiskott-Aldrich Syndrome Protein (N-WASp) with Actin- Related Protein (ARP2/3) Complex. J. Biol. Chem. 275:21946–52.

Carnell, M., T. Zech, S. D. Calaminus, S. Ura, M. Hagedorn, S. A. Johnston, R. C. May, T. Soldati, L. M. Machesky, and R. H. Insall. 2011. Actin Polymerization Driven by WASH Causes V-ATPase Retrieval and Vesicle Neutralization before Exocytosis. J. Cell Biol. 193:831–839.

Caswell, P. T., H. J. Spence, M. Parsons, D. P. White, K. Clark, K. W. Cheng, G. B. Mills, M. J. Humphries, A. J. Messent, K. I. Anderson, M. W. McCaffrey, B. W. Ozanne, and J. C. Norman.

Microenvironments. Dev. Cell 13:496–510. 2007a. Rab25 Associates with α5β1 Integrin to Promote Invasive Migration in 3D

103 Caswell, P. T., H. J. Spence, M. Parsons, D. P. White, K. Clark, K. W. Cheng, G. B. Mills, M. J. Humphries, A. J. Messent, K. I. Anderson, M. W. McCaffrey, B. W. Ozanne, and J. C. Norman.

Microenvironments. Dev. Cell 13:496–510. 2007b. Rab25 Associates with α5β1 Integrin to Promote Invasive Migration in 3D Caswell, P. T., S. Vadrevu, and J. C. Norman. 2009. Integrins: Masters and Slaves of Endocytic Transport. Nat. Rev. Mol. Cell Biol. 10:843–853.

Chabin-Brion, K., J. Marceiller, F. Perez, C. Settegrana, A. Drechou, G. Durand, and C. Poüs. 2001. The Golgi Complex Is a Microtubule-Organizing Organelle. Mol. Biol. Cell 12:2047–60.

Chan, C., C. C. Beltzner, and T. D. Pollard. 2009. Article Cofilin Dissociates Arp2/3 Complex and Branches from Actin Filaments. Curr. Biol. 19:537–545.

Chen, B., S. B. Padrick, L. Henry, and M. K. Rosen. 2014. Biochemical Reconstitution of the WAVE Regulatory Complex. Methods Enzymol. 540:55–72.

Chen, Z., D. Borek, S. B. Padrick, T. S. Gomez, Z. Metlagel, A. M. Ismail, J. Umetani, D. D. Billadeau, Z. Otwinowski, and M. K. Rosen. 2010. Structure and Control of the Actin Regulatory WAVE Complex. Nature 468:533–538.

Chesarone, M. A., A. Grace DuPage, and B. L. Goode. 2009a. Unleashing Formins to Remodel the Actin and Microtubule Cytoskeletons. Nat. Rev. Mol. Cell Biol.

Chesarone, M., C. J. Gould, J. B. Moseley, and B. L. Goode. 2009b. Displacement of Formins from Growing Barbed Ends by Bud14 Is Critical for Actin Cable Architecture and Function. Dev. Cell.

Le Clainche, C., D. Pantaloni, and M.-F. Carlier. 2003. ATP Hydrolysis on Actin-Related Protein 2/3 Complex Causes Debranching of Dendritic Actin Arrays. Proc. Natl. Acad. Sci. U. S. A. 100:6337–42.

Clairfeuille, T., C. Mas, A. S. M. Chan, Z. Yang, M. Tello-Lafoz, M. Chandra, J. Widagdo, M. C. Kerr, B. Paul, I. Mérida, R. D. Teasdale, N. J. Pavlos, V. Anggono, and B. M. Collins. 2016. A Molecular Code for Endosomal Recycling of Phosphorylated Cargos by the SNX27-Retromer Complex. Nat. Struct. Mol. Biol. 23:921–932.

Collins, A., A. Warrington, K. Taylor, and T. Svitkina. 2011. Structural Organization of the Actin Cytoskeleton at Sites of Clathrin-Mediated Endocytosis. Curr. Biol. 21:1167–1175.

Corboy, M. J., P. J. Thomas, and W. C. Wigley. 2005. Aggresome Formation. Methods Mol. Biol. 301:305–27.

Cory, G. O. C., R. Cramer, L. Blanchoin, and A. J. Ridley. 2003. Phosphorylation of the WASP-VCA Domain Increases Its Affinity for the Arp2/3 Complex and Enhances Actin Polymerization by WASP. Mol. Cell 11:1229–39.

D A Begg, R Rodewald, L. I. R. 1978. The Visualization of Actin Filament Polarity in Thin Sections. Evidence for the Uniform Polarity of Membrane-Associated Filaments. J. Cell Biol.

Dang, I., R. Gorelik, C. Sousa-Blin, E. Derivery, C. Guérin, J. Linkner, M. Nemethova, J. G. Dumortier, F. A. Giger, T. A. Chipysheva, V. D. Ermilova, S. Vacher, V. Campanacci, I. Herrada, A.-G. Planson, S. Fetics, V. Henriot, V. David, K. Oguievetskaia, G. Lakisic, F. Pierre, A. Steffen, A. Boyreau, N. Peyriéras, K. Rottner, S. Zinn-Justin, J. Cherfils, I. Bièche, A. Y. Alexandrova, N. B. David, J. V. Small, J. Faix, L. Blanchoin, and A. Gautreau. 2013. Inhibitory Signalling to the Arp2/3 Complex Steers Cell Migration. Nature 503.

104 Deng, Z.-H., T. S. Gomez, D. G. Osborne, C. A. Phillips-Krawczak, J.-S. Zhang, and D. D. Billadeau. 2015. Nuclear FAM21 Participates in NF- B-Dependent Gene Regulation in Pancreatic Cancer Cells. J. Cell Sci. 128:373–384.

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.

Derivery, E., and A. Gautreau. 2010a. Generation of Branched Actin Networks: Assembly and Regulation of the N-WASP and WAVE Molecular Machines. BioEssays 32:119–131.

Derivery, E., and A. Gautreau. 2010b. Evolutionary Conservation of the WASH Complex, an Actin Polymerization Machine Involved in Endosomal Fission. Commun. Integr. Biol. 3:227–230.

Derivery, E., E. Helfer, V. Henriot, A. Gautreau, and L. Berland. 2012. Actin Polymerization Controls the Organization of WASH Domains at the Surface of Endosomes. PLoS One 7:e39774.

Derivery, E., B. Lombard, D. Loew, and A. Gautreau. 2009a. The Wave Complex Is Intrinsically Inactive. Cell Motil. Cytoskeleton 66:777–790.

Derivery, E., C. Sousa, J. J. Gautier, B. Lombard, D. Loew, and A. Gautreau. 2009b. The Arp2/3 Activator WASH Controls the Fission of Endosomes through a Large Multiprotein Complex. Dev. Cell 17:712–723.

Derry, J. M. J., H. D. Ochs, and U. Francke. 1994. Isolation of a Novel Gene Mutated in Wiskott- Aldrich Syndrome. Cell 76:635–644.

Dion, P. A., H. Daoud, and G. A. Rouleau. 2009. Genetics of Motor Neuron Disorders: New Insights into Pathogenic Mechanisms. Nat. Rev. Genet. 10.

Doherty, G. J., and H. T. Mcmahon. 2009. Mechanisms of Endocytosis. Annu. Rev. Biochem.

Dong, R., Y. Saheki, S. Swarup, L. Lucast, J. W. Harper, and P. De Camilli. 2016. Endosome-ER Contacts Control Actin Nucleation and Retromer Function through VAP-Dependent Regulation of PI4P. Cell 166:408–423.

Doxsey, S. J. 2001. Centrosomes as Command Centres for Cellular Control. Nat. Cell Biol. 3:E105– E108.

Drenckhahn, D., and T. D. Pollard. 1986. Elongation of Actin Filaments Is a Diffusion-Limited Reaction at the Barbed End and Is Accelerated by Inert Macromolecules. J. Biol. Chem. 261:12754–8.

Drengk, A., J. Fritsch, C. Schmauch, H. R?hling, and M. Maniak. 2003. A Coat of Filamentous Actin Prevents Clustering of Late-Endosomal Vacuoles in Vivo. Curr. Biol. 13:1814–1819.

Duleh, S. N., and M. D. Welch. 2010. WASH and the Arp2/3 Complex Regulate Endosome Shape and Trafficking. Cytoskeleton 67:193–206.

Duleh, S. N., and M. D. Welch. 2012. Regulation of Integrin Trafficking, Cell Adhesion, and Cell Migration by WASH and the Arp2/3 Complex. Cytoskeleton 69:1047–1058.

Eden, S., R. Rohatgi, A. V. Podtelejnikov, M. Mann, and M. W. Kirschner. 2002. Mechanism of Regulation of WAVE1-Induced Actin Nucleation by Rac1 and Nck. Nature 418:790–793.

Edwards, M., A. Zwolak, D. A. Schafer, D. Sept, R. Dominguez, and J. A. Cooper. 2014. Capping

105 Protein Regulators Fine-Tune Actin Assembly Dynamics. Nat. Rev. Mol. Cell Biol. 15:677–89.

Egile, C., I. Rouiller, X.-P. Xu, N. Volkmann, R. Li, and D. Hanein. 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.

Euteneuer, U., and M. Schliwa. 1992. Mechanism of Centrosome Positioning during the Wound Response in BSC-1 Cells. J. Cell Biol. 116:1157–66.

Farina, F., J. Gaillard, C. Guérin, Y. Couté, J. Sillibourne, L. Blanchoin, and M. Théry. 2016. The Centrosome Is an Actin-Organizing Centre. Nat. Cell Biol. 18:65–75.

Firat-Karalar, E. N., P. P. Hsiue, and M. D. Welch. 2011. The Actin Nucleation Factor JMY Is a Negative Regulator of Neuritogenesis. Mol. Biol. Cell 22:4563–74.

Fisk, H. A. 2012. Many Pathways to Destruction: The Role of the Centrosome In, and Its Control by Regulated Proteolysis. Pages 133–155The Centrosome. Humana Press, Totowa, NJ.

Flores-Rodriguez, N., S. S. Rogers, D. A. Kenwright, T. A. Waigh, P. G. Woodman, and V. J. Allan. 2011. Roles of Dynein and Dynactin in Early Endosome Dynamics Revealed Using Automated Tracking and Global Analysis. PLoS One 6:e24479.

Follett, J., S. J. Norwood, N. A. Hamilton, M. Mohan, O. Kovtun, S. Tay, Y. Zhe, S. A. Wood, G. D. Mellick, P. A. Silburn, B. M. Collins, A. Bugarcic, and R. D. Teasdale. 2014. The Vps35 D620N Mutation Linked to Parkinson’s Disease Disrupts the Cargo Sorting Function of Retromer. Traffic 15:230–244.

Freeman, C. L., G. Hesketh, and M. N. J. Seaman. 2014. RME-8 Coordinates the Activity of the WASH Complex with the Function of the Retromer SNX Dimer to Control Endosomal Tubulation. J. Cell Sci. 127.

Freeman, C., M. N. J. Seaman, and E. Reid. 2013. The Hereditary Spastic Paraplegia Protein Strumpellin: Characterisation in Neurons and of the Effect of Disease Mutations on WASH Complex Assembly and Function. Biochim. Biophys. Acta - Mol. Basis Dis. 1832:160–173.

Funakoshi, M., R. J. Tomko, H. Kobayashi, and M. Hochstrasser. 2009. Multiple Assembly Chaperones Govern Biogenesis of the Proteasome Regulatory Particle Base. Cell 137:887– 899.

Gallon, M., T. Clairfeuille, F. Steinberg, C. Mas, R. Ghai, R. B. Sessions, R. D. Teasdale, B. M. Collins, and P. J. Cullen. 2014. A Unique PDZ Domain and Arrestin-like Fold Interaction Reveals Mechanistic Details of Endocytic Recycling by SNX27-Retromer. Proc. Natl. Acad. Sci. U. S. A. 111:E3604-13.

Galvez, T., J. Gilleron, M. Zerial, and G. A. O. 2012. SnapShot: Mammalian Rab Proteins in Endocytic Trafficking. Cell 151:234–234.e2.

Gandhi, M., B. A. Smith, M. Bovellan, V. Paavilainen, K. Daugherty-Clarke, J. Gelles, P. Lappalainen, and B. L. Goode. 2010. GMF Is a Cofilin Homolog That Binds Arp2/3 Complex to Stimulate Filament Debranching and Inhibit Actin Nucleation. Page Current Biology.

Gautier, J. J., M. E. Lomakina, L. Bouslama-Oueghlani, E. Derivery, H. Beilinson, W. Faigle, D. Loew, D. Louvard, A. Echard, A. Y. Alexandrova, B. Baum, and A. Gautreau. 2011. Clathrin Is Required for Scar/Wave-Mediated Lamellipodium Formation. J. Cell Sci. 124:3414–3427.

Gautreau, A., H. H. Ho, J. Li, H. Steen, S. P. Gygi, and M. W. Kirschner. 2004. Purification and Architecture of the Ubiquitous Wave Complex. Proc. Natl. Acad. Sci. U. S. A. 101:4379–83.

106 Gautreau, A., K. Oguievetskaia, and C. Ungermann. 2014. Function and Regulation of the Endosomal Fusion and Fission Machineries. Cold Spring Harb. Perspect. Biol. 6:1–16.

Glickman, M. H., D. M. Rubin, O. Coux, I. Wefes, G. Nter Pfeifer, Z. Cjeka, W. Baumeister, V. A. Fried, and D. Finley. 1998. A Subcomplex of the Proteasome Regulatory Particle Required for Ubiquitin-Conjugate Degradation and Related to the COP9-Signalosome and eIF3. Cell 94:615–623.

Godinho, S. A., R. Picone, M. Burute, R. Dagher, Y. Su, C. T. Leung, K. Polyak, J. S. Brugge, M. Théry, and D. Pellman. 2014. Oncogene-like Induction of Cellular Invasion from Centrosome Amplification. Nature 510:167–71.

Gomez, T. S., and D. D. Billadeau. 2009. A FAM21-Containing WASH Complex Regulates Retromer-Dependent Sorting. Dev. Cell 17:699–711.

Gomez, T. S., J. a. Gorman, a. Artal-Martinez de Narvajas, a. O. Koenig, and D. D. Billadeau. 2012. Trafficking Defects in WASH-Knockout Fibroblasts Originate from Collapsed Endosomal and Lysosomal Networks. Mol. Biol. Cell 23:3215–3228.

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.

Granger, E., G. Mcnee, V. Allan, and P. Woodman. 2014. The Role of the Cytoskeleton and Molecular Motors in Endosomal Dynamics. Semin. Cell Dev. Biol. 31:20–29.

Gu, Z. C., and C. Enenkel. 2014. Proteasome Assembly. Cell. Mol. Life Sci. 71:4729–4745.

Gunning, P. W., U. Ghoshdastider, S. Whitaker, D. Popp, and R. C. Robinson. 2015. The Evolution of Compositionally and Functionally Distinct Actin Filaments. J. Cell Sci.

Guo, J., Z. Yang, W. Song, Q. Chen, F. Wang, Q. Zhang, and X. Zhu. 2006. Nudel Contributes to Microtubule Anchoring at the Mother Centriole and Is Involved in Both Dynein-Dependent and -Independent Centrosomal Protein Assembly. Mol. Biol. Cell 17:680–9.

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. 492:215–220.

Hao, Y. H., J. M. Doyle, S. Ramanathan, T. S. Gomez, D. Jia, M. Xu, Z. J. Chen, D. D. Billadeau, M. K. Rosen, and P. R. Potts. 2013. Regulation of WASH-Dependent Actin Polymerization and Protein Trafficking by Ubiquitination. Cell 152.

Harbour, M. E., S. Y. A. Breusegem, R. Antrobus, C. Freeman, E. Reid, and M. N. J. Seaman. 2010. The Cargo-Selective Retromer Complex Is a Recruiting Hub for Protein Complexes That Regulate Endosomal Tubule Dynamics. J. Cell Sci. 123:3703–17.

Harbour, M. E., S. Y. Breusegem, and M. N. J. Seaman. 2012. Recruitment of the Endosomal WASH Complex Is Mediated by the Extended “Tail” of Fam21 Binding to the Retromer Protein Vps35. Biochem. J 442:209–220.

Helfer, E., M. E. Harbour, V. Henriot, G. Lakisic, C. Sousa-Blin, L. Volceanov, M. N. J. Seaman, and A. Gautreau. 2013. Endosomal Recruitment of the WASH Complex: Active Sequences and Mutations Impairing Interaction with the Retromer. Biol. Cell 105:191–207.

Hernandez-Valladares, M., T. Kim, B. Kannan, A. Tung, A. H. Aguda, M. Larsson, J. A. Cooper, and R. C. Robinson. 2010. Structural Characterization of a Capping Protein Interaction Motif Defines a Family of Actin Filament Regulators. Nat. Struct. Mol. Biol. 17:497–503.

107 Higgs, H. N., L. Blanchoin, and T. D. Pollard. 1999. Influence of the C Terminus of Wiskott-Aldrich Syndrome Protein ( WASp ) and the Arp2 / 3 Complex on Actin Polymerization. Biochemistry 38:15212–15222.

Higgs, H. N., and T. D. Pollard. 2000. Activation by Cdc42 and Pip2 of Wiskott-Aldrich Syndrome Protein (Wasp) Stimulates Actin Nucleation by Arp2/3 Complex. J. Cell Biol. 150.

Hirano, Y., H. Hayashi, S.-I. Iemura, K. B. Hendil, S.-I. Niwa, T. Kishimoto, M. Kasahara, T. Natsume, K. Tanaka, and S. Murata. 2006. Cooperation of Multiple Chaperones Required for the Assembly of Mammalian 20S Proteasomes. Mol. Cell 24:977–84.

Hirano, Y., K. B. Hendil, H. Yashiroda, S. Iemura, R. Nagane, Y. Hioki, T. Natsume, K. Tanaka, and S. Murata. 2005. A Heterodimeric Complex That Promotes the Assembly of Mammalian 20S Proteasomes. Nature 437:1381–5.

Hirano, Y., T. Kaneko, K. Okamoto, M. Bai, H. Yashiroda, K. Furuyama, K. Kato, K. Tanaka, and S. Murata. 2008. Dissecting Beta-Ring Assembly Pathway of the Mammalian 20S Proteasome. EMBO J. 27:2204–13.

Ho, H.-Y. H., R. Rohatgi, A. M. Lebensohn, Le Ma, J. Li, S. P. Gygi, and M. W. Kirschner. 2004. Toca- 1 Mediates Cdc42-Dependent Actin Nucleation by Activating the N-WASP-WIP Complex. Cell 118:203–216.

Ho, H.-Y. H., R. Rohatgi, L. Ma, and M. W. Kirschner. 2001. CR16 Forms a Complex with N-WASP in Brain and Is a Novel Member of a Conserved Proline-Rich Actin-Binding Protein Family. Proc. Natl. Acad. Sci. 98:11306–11311.

Hoyt, M. A., S. McDonough, S. A. Pimpl, H. Scheel, K. Hofmann, and P. Coffino. 2008. A Genetic Screen forSaccharomyces Cerevisiae Mutants Affecting Proteasome Function, Using a Ubiquitin-Independent Substrate. Yeast 25:199–217.

Huotari, J., and A. Helenius. 2011. Focus Review Endosome Maturation. EMBO J. 30:3481–3500.

Hurtado, L., C. Caballero, M. P. Gavilan, J. Cardenas, M. Bornens, and R. M. Rios. 2011. Disconnecting the Golgi Ribbon from the Centrosome Prevents Directional Cell Migration and Ciliogenesis. J. Cell Biol. 193:917–933.

Huttenlocher, A., and A. R. Horwitz. 2011. Integrins in Cell Migration TL - 3. Cold Spring Harb. Perspect Biol 3 VN-re:1–16.

Huxley, H. E. 1963. Electron Microscope Studies on the Structure of Natural and Synthetic Protein Filaments from Striated Muscle. J. Mol. Biol. 7:281–IN30.

Imai, K., S. Nonoyama, and H. D. Ochs. 2003. WASP (Wiskott-Aldrich Syndrome Protein) Gene Mutations and Phenotype. Curr. Opin. Allergy Clin. Immunol. 3:427–36.

Innocenti, M., A. Zucconi, A. Disanza, E. Frittoli, L. B. Areces, A. Steffen, T. E. B. 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. 6:319–327.

Iwaya, K., K. Norio, and K. Mukai. 2007. Coexpression of Arp2 and WAVE2 Predicts Poor Outcome in Invasive Breast Carcinoma. Mod. Pathol. 20:339–43.

Jia, D., T. S. Gomez, D. D. Billadeau, and M. K. Rosen. 2012. Multiple Repeat Elements within the FAM21 Tail Link the WASH Actin Regulatory Complex to the Retromer. Mol. Biol. Cell.

Jia, D., T. S. Gomez, Z. Metlagel, J. Umetani, Z. Otwinowski, M. K. Rosen, and D. D. Billadeau. 2010.

108 WASH and WAVE Actin Regulators of the Wiskott-Aldrich Syndrome Protein (WASP) Family Are Controlled by Analogous Structurally Related Complexes. Proc. Natl. Acad. Sci. 107:10442–10447.

Johnston, J. A., M. E. Illing, and R. R. Kopito. 2002. Cytoplasmic Dynein/dynactin Mediates the Assembly of Aggresomes. Cell Motil. Cytoskeleton 53:26–38.

Johnston, J. A., C. L. Ward, and R. R. Kopito. 1998. Aggresomes: A Cellular Response to Misfolded Proteins. J. Cell Biol. 143:1883–1898.

Kaneko, T., J. Hamazaki, S. Iemura, K. Sasaki, K. Furuyama, T. Natsume, K. Tanaka, and S. Murata. 2009. Assembly Pathway of the Mammalian Proteasome Base Subcomplex Is Mediated by Multiple Specific Chaperones. Cell 137:914–925.

Kato, M., H. Miki, S. Kurita, T. Endo, H. Nakagawa, S. Miyamoto, and T. Takenawa. 2002. WICH, a Novel Verprolin Homology Domain-Containing Protein That Functions Cooperatively with N-WASP in Actin-Microspike Formation. Biochem. Biophys. Res. Commun. 291:41–7.

Kim, A. S., L. T. Kakalis, N. Abdul-Manan, G. A. Liu, and M. K. Rosen. 2000. Autoinhibition and Activation Mechanisms of the Wiskott–Aldrich Syndrome Protein. Nature 404:151–158.

King, J. S., D. M. Veltman, M. Georgiou, B. Baum, and R. H. Insall. 2010. SCAR/WAVE Is Activated at Mitosis and Drives Myosin-Independent Cytokinesis. J. Cell Sci. 123:2246–55.

Kobayashi, K., S. Kuroda, M. Fukata, T. Nakamura, T. Nagase, N. Nomura, Y. Matsuura, N. Yoshida-Kubomura, A. Iwamatsu, and K. Kaibuchi. 1998. p140Sra-1 (Specifically Rac1- Associated Protein) Is a Novel Specific Target for Rac1 Small GTPase. J. Biol. Chem. 273:291–5.

Kowalski, J. R., C. Egile, S. Gil, S. B. Snapper, R. Li, and S. M. Thomas. 2005. Cortactin Regulates Cell Migration through Activation of N-WASP. J. Cell Sci. 118:79–87.

Krause, M., and A. Gautreau. 2014. Steering Cell Migration: Lamellipodium Dynamics and the Regulation of Directional Persistence. Nat. Rev. Mol. Cell Biol. 15:577–90.

Kriegenburg, F., M. Seeger, Y. Saeki, K. Tanaka, A.-M. B. Lauridsen, R. Hartmann-Petersen, and K. B. Hendil. 2008. Mammalian 26S Proteasomes Remain Intact during Protein Degradation. Cell 135:355–365.

Krzewski, K., X. Chen, J. S. Orange, and J. L. Strominger. 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.

Kull, F. J., and S. A. Endow. 2013. Force Generation by Kinesin and Myosin Cytoskeletal Motor Proteins. J. Cell Sci. 126.

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. 13:1867–1875.

Kupfer, A., D. Louvard, and S. J. Singer. 1982. Polarization of the Golgi Apparatus and the Microtubule-Organizing Center in Cultured Fibroblasts at the Edge of an Experimental Wound. Proc. Natl. Acad. Sci. U. S. A. 79:2603–7.

Kusmierczyk, A. R., M. J. Kunjappu, M. Funakoshi, and M. Hochstrasser. 2008. A Multimeric Assembly Factor Controls the Formation of Alternative 20S Proteasomes. Nat. Struct. Mol. Biol. 15:237–44.

109 Laukaitis, C. M., D. J. Webb, K. Donais, and A. F. Horwitz. 2001. Differential Dynamics of Alpha-5 Integrin, Paxillin, and Alpha-Actinin during Formation and Disassembly of Adhesions in Migrating Cells. J. Cell Biol. 153:1427–1440.

Le, J., E. L. Mallery, C. Zhang, S. Brankle, and D. B. Szymanski. 2006. Arabidopsis BRICK1/HSPC300 Is an Essential WAVE-Complex Subunit That Selectively Stabilizes the Arp2/3 Activator SCAR2. Curr. Biol. 16:895–901.

Lee, S., J. Chang, C. Blackstone, L. Cunjiang, and J. Jianping. 2016. FAM21 Directs SNX27– retromer Cargoes to the Plasma Membrane by Preventing Transport to the Golgi Apparatus. Nat. Commun. 7:10939.

Li, Y.-H., N. Zhang, Y.-N. Wang, Y. Shen, and Y. Wang. 2016. Multiple Faces of Protein Interacting with C Kinase 1 (PICK1): Structure, Function, and Diseases. Neurochem. Int. 98:115–121.

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.

Linkner, J., G. Witte, T. Stradal, U. Curth, and J. Faix. 2011. High-Resolution X-Ray Structure of the Trimeric Scar/WAVE-Complex Precursor Brk1. PLoS One 6:e21327.

Liu, A. P., and D. A. Fletcher. 2006. Actin Polymerization Serves as a Membrane Domain Switch in Model Lipid Bilayers. Biophys. J. 91:4064–4070.

Liu, R., M. T. Abreu-Blanco, K. C. Barry, E. V. Linardopoulou, G. E. Osborn, and S. M. Parkhurst. 2009. Wash Functions Downstream of Rho and Links Linear and Branched Actin Nucleation Factors. Development 136:2849–2860.

Lomakin, A. J., K.-C. Lee, S. J. Han, D. A. Bui, M. Davidson, A. Mogilner, and G. Danuser. 2015. Competition for Actin between Two Distinct F-Actin Networks Defines a Bistable Switch for Cell Polarization. Nat. Cell Biol.

Loo, M. A., T. J. Jensen, L. Cui, Y. Hou, X. B. Chang, and J. R. Riordan. 1998. Perturbation of Hsp90 Interaction with Nascent CFTR Prevents Its Maturation and Accelerates Its Degradation by the Proteasome. EMBO J. 17:6879–87.

Machesky, L. M., S. J. Atkinson, C. Ampe, J. Vandekercldaove, and T. D. Pollard. 1994. Purification of a Cortical Complex Containing Two Unconventional Actins from Acanthamoeba by Affinity Chromatography on Profilin-Agarose. J. Cell Biol.

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. 8:1347–1356.

Machesky, L. M., R. D. Mullins, H. N. Higgs, D. A. Kaiser, L. Blanchoin, R. C. May, M. E. Hall, and T. D. Pollard. 1999. Scar, a WASp-Related Protein, Activates Nucleation of Actin Filaments by the Arp2/3 Complex. Proc. Natl. Acad. Sci. U. S. A. 96:3739–44.

Machesky, L. M., E. Reeves, F. Wientjes, F. J. Mattheyse, A. Grogan, N. F. Totty, A. L. Burlingame, J. J. Hsuan, and A. W. Segal. 1997. Mammalian Actin-Related Protein 2/3 Complex Localizes to Regions of Lamellipodial Protrusion and Is Composed of Evolutionarily Conserved Proteins. Biochem. J 328:105–112.

Maritzen, T., T. Zech, M. R. Schmidt, E. Krause, L. M. Machesky, and V. Haucke. 2012. Gadkin Negatively Regulates Cell Spreading and Motility via Sequestration of the Actin-Nucleating ARP2/3 Complex. Proc. Natl. Acad. Sci. U. S. A. 109:10382–7.

110 Martin, A. C., M. D. Welch, and D. G. Drubin. 2006. Arp2/3 ATP Hydrolysis-Catalysed Branch Dissociation Is Critical for Endocytic Force Generation. Nat. Cell Biol. 8:826–833.

McGough, I. J., F. Steinberg, D. Jia, P. A. Barbuti, K. J. McMillan, K. J. Heesom, A. L. Whone, M. A. Caldwell, D. D. Billadeau, M. K. Rosen, and P. J. Cullen. 2014. Retromer Binding to FAM21 and the WASH Complex Is Perturbed by the Parkinson Disease-Linked VPS35(D620N) Mutation. Curr. Biol. 24:1670–1676.

McMillan, K. J., M. Gallon, A. P. Jellett, T. Clairfeuille, F. C. Tilley, I. McGough, C. M. Danson, K. J. Heesom, K. A. Wilkinson, B. M. Collins, and P. J. Cullen. 2016. Atypical Parkinsonism– associated Retromer Mutant Alters Endosomal Sorting of Specific Cargo Proteins. J. Cell Biol. 214.

McMillan, K. J., H. C. Korswagen, and P. J. Cullen. 2017. The Emerging Role of Retromer in Neuroprotection. Curr. Opin. Cell Biol. 47:72–82.

McNiven, M. A., L. Kim, E. W. Krueger, J. D. Orth, H. Cao, and T. W. Wong. 2000. Regulated Interactions between Dynamin and the Actin-Binding Protein Cortactin Modulate Cell Shape. J. Cell Biol. 151.

Menzl, I., L. Lebeau, R. Pandey, N. B. Hassounah, F. W. Li, R. Nagle, K. Weihs, and K. M. McDermott. 2014. Loss of Primary Cilia Occurs Early in Breast Cancer Development. Cilia 3:7.

Merrifield, C. J., and M. Kaksonen. 2014. Endocytic Accessory Factors and Regulation of Clathrin- Mediated Endocytosis. Cold Spring Harb Perspect Biol.

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 Clathrin- Mediated Endocytosis in Cultured Fibroblasts. Eur. J. Cell Biol 83.

Mestres, I., and C.-H. Sung. 2017. Nervous System Development Relies on Endosomal Trafficking. Neurogenesis 4:e1316887.

Miki, H., K. Miura, and T. Takenawa. 1996. N-WASP, a Novel Actin-Depolymerizing Protein, Regulates the Cortical Cytoskeletal Rearrangement in a PIP2-Dependent Manner Downstream of Tyrosine Kinases. EMBO J. 15:5326–35.

Miki, H., T. Sasaki, Y. Takai, and T. Takenawa. 1998a. Induction of Filopodium Formation by a WASP-Related Actin-Depolymerizing Protein N-WASP. Nature 391:93–96.

Miki, H., S. Suetsugu, and T. Takenawa. 1998b. WAVE, a Novel WASP-Family Protein Involved in Actin Reorganization Induced by Rac. EMBO J. 17:6932–6941.

Miki, H., and T. Takenawa. 2002. WAVE2 Serves a Functional Partner of IRSp53 by Regulating Its Interaction with Rac. Biochem. Biophys. Res. Commun. 293:93–99.

Mizutani, K., H. Miki, H. He, H. Maruta, and T. Takenawa. 2002. Essential Role of Neural Wiskott- Aldrich Syndrome Protein in Podosome Formation and Degradation of Extracellular Matrix in Src-Transformed Fibroblasts. Cancer Res. 62.

Molinie, N., and A. Gautreau. 2017. The Arp2/3 Regulatory System and Its Deregulation in Cancer. Physiol. Rev. In press.

Monteiro, P., C. Rossé, A. Castro-Castro, M. Irondelle, E. Lagoutte, P. Paul-Gilloteaux, C. Desnos, E. Formstecher, F. Darchen, D. Perrais, A. Gautreau, M. Hertzog, and P. Chavrier. 2013. Endosomal WASH and Exocyst Complexes Control Exocytosis of MT1-MMP at Invadopodia.

111 J. Cell Biol. 203:1063–1079.

Mooren, O. L., B. J. Galletta, and J. A. Cooper. 2012. Roles for Actin Assembly in Endocytosis. Annu. Rev. Biochem. 81:661–686.

Muller, J., Y. Oma, L. Vallar, E. Friederich, O. Poch, and B. Winsor. 2005. Sequence and Comparative Genomic Analysis of Actin-Related Proteins. Mol. Biol. Cell 16:5736–5748.

Mullins, D. R., 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. Cell Biol. 95:6181–6186.

Munsie, L. N., A. J. Milnerwood, P. Seibler, D. A. Beccano-Kelly, I. Tatarnikov, J. Khinda, M. Volta, C. Kadgien, L. P. Cao, L. Tapia, C. Klein, and M. J. Farrer. 2015. Retromer-Dependent Neurotransmitter Receptor Trafficking to Synapses Is Altered by the Parkinson’s Disease VPS35 Mutation p.D620N. Hum. Mol. Genet. 24:1691–703.

Murata, S., H. Yashiroda, and K. Tanaka. 2009. Molecular Mechanisms of Proteasome Assembly. Nat. Rev. Mol. Cell Biol. 10:104–115.

Nagel, B. M., M. Bechtold, L. G. Rodriguez, and S. Bogdan. 2016. Drosophila WASH Is Required for Integrin-Mediated Cell Adhesion, Cell Motility and Lysosomal Neutralization. J. Cell Sci.:jcs.193086.

Nemere, I., A. Kupfer, and S. J. Singer. 1985. Reorientation of the Golgi Apparatus and the Microtubule-Organizing Center inside Macrophages Subjected to a Chemotactic Gradient. Cell Motil. 5:17–29.

Nerurkar, P., M. Altvater, S. Gerhardy, S. Sch€ Utz, U. Fischer, C. Weirich, and V. Govind Panse. 2015. Eukaryotic Ribosome Assembly and Nuclear Export. Int. Rev. Cell Mol. Biol. 319:107– 140.

Nobutani, K., Y. Shimono, M. Yoshida, K. Mizutani, A. Minami, S. Kono, T. Mukohara, T. Yamasaki, T. Itoh, S. Takao, H. Minami, T. Azuma, and Y. Takai. 2014. Absence of Primary Cilia in Cell Cycle-Arrested Human Breast Cancer Cells. Genes to Cells 19:141–152.

Obino, D., F. Farina, O. Malbec, P. J. Sáez, M. Maurin, J. Gaillard, F. Dingli, D. Loew, A. Gautreau, M.- I. Yuseff, L. Blanchoin, M. Théry, and A.-M. Lennon-Duménil. 2016. Actin Nucleation at the Centrosome Controls Lymphocyte Polarity. Nat. Commun. 7:10969.

Ohashi, E., K. Tanabe, Y. Henmi, K. Mesaki, Y. Kobayashi, and K. Takei. 2011. Receptor Sorting within Endosomal Trafficking Pathway Is Facilitated by Dynamic Actin Filaments. PLoS One 6:e19942.

Oikawa, T., H. Yamaguchi, T. Itoh, M. Kato, T. Ijuin, D. Yamazaki, S. Suetsugu, and T. Takenawa. 2004. PtdIns(3,4,5)P3 Binding Is Necessary for WAVE2-Induced Formation of Lamellipodia. Nat. Cell Biol. 6:420–6.

Pantaloni, D., and M. F. Carlier. 1993. How Profilin Promotes Actin Filament Assembly in the Presence of Thymosin Beta 4. Cell 75:1007–14.

Park, L., P. A. Thomason, T. Zech, J. S. King, D. M. Veltman, M. Carnell, S. Ura, L. M. Machesky, and R. H. Insall. 2013. Cyclical Action of the WASH Complex: FAM21 and Capping Protein Drive WASH Recycling, Not Initial Recruitment. Dev. Cell 24:169–181.

Park, S., J. Roelofs, W. Kim, J. Robert, M. Schmidt, S. P. Gygi, and D. Finley. 2009. Hexameric Assembly of the Proteasomal ATPases Is Templated through Their C Termini. Nature

112 459:866–870.

Pellinen, T., and J. Ivaska. 2006. Integrin Traffic. J. Cell Sci. 119.

Piotrowski, J. T., T. S. Gomez, R. A. Schoon, A. K. Mangalam, and D. D. Billadeau. 2013. WASH Knockout T Cells Demonstrate Defective Receptor Trafficking, Proliferation, and Effector Function. Mol. Cell. Biol. 33:958–73.

Pockley, A. G. 2003. Heat Shock Proteins as Regulators of the Immune Response. Lancet:1–8.

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. 2016. Actin and Actin-Binding Proteins. Cold Spring Harb. Perspect. Biol. 8:a018226.

Pollard, T. D., and J. A. Cooper. 1984. Quantitative Analysis of the Effect of Acanthamoeba Profilin on Actin Filament Nucleation and Elongation. Biochemistry 23:6631–41.

Pollard, T. D., and J. A. Cooper. 2009. Actin, a Central Player in Cell Shape and Movement. Science 326:1208–12.

Porkka, K. P., T. L. J. Tammela, R. L. Vessella, and T. Visakorpi. 2004. RAD21 and KIAA0196 at 8q24 Are Amplified and Overexpressed in Prostate Cancer. Genes. Chromosomes Cancer 39:1–10.

Prehoda, K. E., J. A. Scott, R. D. Mullins, and W. A. Lim. 2000. Integration of Multiple Signals through Cooperative Regulation of the N-WASP-Arp2/3 Complex. Science 290:801–6.

Pring, M., M. Evangelista, C. Boone, C. Yang, and S. H. Zigmond. 2003. Mechanism of Formin- Induced Nucleation of Actin Filaments. Biochemistry.

Puthenveedu, M. A., B. Lauffer, P. Temkin, R. Vistein, P. Carlton, K. Thorn, J. Taunton, O. D. Weiner, R. G. Parton, and M. von Zastrow. 2010. Sequence-Dependent Sorting of Recycling Proteins by Actin-Stabilized Endosomal Microdomains. Cell 143:761–73.

Qian, S.-B., M. F. Princiotta, J. R. Bennink, and J. W. Yewdell. 2006. Characterization of Rapidly Degraded Polypeptides in Mammalian Cells Reveals a Novel Layer of Nascent Protein Quality Control. J. Biol. Chem. 281:392–400.

Ramos, P. C., J. Höckendorff, E. S. Johnson, A. Varshavsky, and R. J. Dohmen. 1998. Ump1p Is Required for Proper Maturation of the 20S Proteasome and Becomes Its Substrate upon Completion of the Assembly. Cell 92:489–499.

Rios, R. M. 2014. The Centrosome-Golgi Apparatus Nexus. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 369:20130462–20130462.

Robinson, R. C., K. Turbedsky, D. A. Kaiser, J.-B. Marchand, H. N. Higgs, S. Choe, and T. D. Pollard . 2001. Crystal Structure of Arp2/3 Complex. Science.

Rocca, D. L., S.࿣ Martin, E. L. Jenkins, and J. G. Hanley. 2008. Inhibition of Arp2/3-Mediated Actin Polymerization by PICK1 Regulates Neuronal Morphology and AMPA Receptor Endocytosis. Nat. Cell Biol. 10:259–271.

Roelofs, J., S. Park, W. Haas, G. Tian, F. E. McAllister, Y. Huo, B.-H. Lee, F. Zhang, Y. Shi, S. P. Gygi, and D. Finley. 2009. Chaperone-Mediated Pathway of Proteasome Regulatory Particle Assembly. Nature 459:861–865.

113 Rohatgi, R., H. Y. Ho, and M. W. Kirschner. 2000. Mechanism of N-WASP Activation by CDC42 and Phosphatidylinositol 4, 5-Bisphosphate. J. Cell Biol. 150:1299–310.

Rohatgi, R., L. Ma, H. Miki, M. Lopez, T. Kirchhausen, T. Takenawa, and M. W. Kirschner. 1999. The Interaction between N-WASP and the Arp2/3 Complex Links Cdc42-Dependent Signals to Actin Assembly. Cell 97:221–31.

Rohatgi, R., P. Nollau, H.-Y. H. Ho, M. W. Kirschner, and B. J. Mayer. 2001. Nck and Phosphatidylinositol 4,5-Bisphosphate Synergistically Activate Actin Polymerization through the N-WASP-Arp2/3 Pathway. J. Biol. Chem. 276:26448–26452.

Ropers, F., E. Derivery, H. Hu, M. Garshasbi, M. Karbasiyan, M. Herold, G. Nü Rnberg, R. Ullmann, A. Gautreau, K. Sperling, R. Varon, and A. Rajab. 2011. Identification of a Novel Candidate Gene for Non-Syndromic Autosomal Recessive Intellectual Disability: The WASH Complex Member SWIP. Hum. Mol. Genet.

Rotty, J. D., C. Wu, and J. E. Bear. 2013a. New Insights into the Regulation and Cellular Functions of the ARP2/3 Complex. Nat. Rev. Mol. Cell Biol. 14:7–12.

Rotty, J. D., C. Wu, and J. E. Bear. 2013b. New Insights into the Regulation and Cellular Functions of the ARP2/3 Complex. Nat. Rev. Mol. Cell Biol. 14.

Rouiller, I., X.-P. Xu, K. J. Amann, C. Egile, S. Nickell, D. Nicastro, R. Li, T. D. Pollard, N. Volkmann, and D. Hanein. 2008. The Structural Basis of Actin Filament Branching by the Arp2/3 Complex. J. Cell Biol. 180:887–95.

Rousseau, A., and A. Bertolotti. 2016. An Evolutionarily Conserved Pathway Controls Proteasome Homeostasis. Nature 536:184–189.

Roux, A., D. Cuvelier, P. Nassoy, J. Prost, P. Bassereau, and B. Goud. 2005. Role of Curvature and Phase Transition in Lipid Sorting and Fission of Membrane Tubules. EMBO J. 24:1537–45.

Roux, A., K. Uyhazi, A. Frost, and P. De Camilli. 2006. GTP-Dependent Twisting of Dynamin Implicates Constriction and Tension in Membrane Fission. Nature.

Saeki, Y., A. Toh-e, T. Kudo, H. Kawamura, and K. Tanaka. 2009. Multiple Proteasome-Interacting Proteins Assist the Assembly of the Yeast 19S Regulatory Particle. Cell 137:900–913.

Sahara, K., L. Kogleck, H. Yashiroda, and S. Murata. 2014. The Mechanism for Molecular Assembly of the Proteasome. Adv. Biol. Regul. 54:51–58.

Satyal, S. H., D. Chen, S. G. Fox, S. H. Satyal, D. Chen, S. G. Fox, J. M. Kramer, and R. I. Morimoto. 1998. Negative Regulation of the Heat Shock Transcriptional Response by HSBP1. Genes Dev.:1962–1974.

Sawa, M., and T. Takenawa. 2006. Caenorhabditis Elegans WASP-Interacting Protein Homologue WIP-1 Is Involved in Morphogenesis through Maintenance of WSP-1 Protein Levels. Biochem. Biophys. Res. Commun. 340:709–17.

Schachtner, H., M. Weimershaus, V. Stache, N. Plewa, D. F. Legler, U. E. Höpken, and T. Maritzen. 2015. Loss of Gadkin Affects Dendritic Cell Migration In Vitro. PLoS One 10:e0143883.

Schatten, H. 2008. The Mammalian Centrosome and Its Functional Significance. Histochem. Cell Biol. 129:667–86.

Schlüter, K., D. Waschbüsch, M. Anft, D. Hügging, S. Kind, J. Hänisch, G. Lakisic, A. Gautreau, A. Barnekow, and T. E. B. Stradal. 2014. JMY Is Involved in Anterograde Vesicle Trafficking

114 from the Trans-Golgi Network. Eur. J. Cell Biol. 93:194–204.

Schmidt, M. R., T. Maritzen, V. Kukhtina, V. A. Higman, L. Doglio, N. N. Barak, H. Strauss, H. Oschkinat, C. G. Dotti, and V. Haucke. 2009. Regulation of Endosomal Membrane Traffic by a Gadkin/AP-1/kinesin KIF5 Complex. Proc. Natl. Acad. Sci. U. S. A. 106:15344–9.

Scita, G., S. Confalonieri, P. Lappalainen, and S. Suetsugu. 2008. IRSp53: Crossing the Road of Membrane and Actin Dynamics in the Formation of Membrane Protrusions. Trends Cell Biol. 18:52–60.

Seaman, M. N. J., A. Gautreau, and D. D. Billadeau. 2013. Retromer-Mediated Endosomal Protein Sorting: All WASHed Up!

Semba, S., K. Iwaya, J. Matsubayashi, H. Serizawa, H. Kataba, T. Hirano, H. Kato, T. Matsuoka, and K. Mukai. 2006. Coexpression of Actin-Related Protein 2 and Wiskott-Aldrich Syndrome Family Verproline-Homologous Protein 2 in Adenocarcinoma of the Lung. Clin. Cancer Res. 12:2449–2454.

Sept, D., and J. A. McCammon. 2001. Thermodynamics and Kinetics of Actin Filament Nucleation. Biophys. J. 81:667–674.

Shafaq-Zadah, M., C. S. Gomes-Santos, S. Bardin, P. Maiuri, M. Maurin, J. Iranzo, A. Gautreau, C. Lamaze, P. Caswell, B. Goud, and L. Johannes. 2015. Persistent Cell Migration and Adhesion Nat. Cell Biol.

Shen,Rely Q.-T., on P. Retrograde P. Hsiue, C. TransportV. Sindelar, of M. β D.1 Integrin. Welch, K. G. Campellone, and H.-W. Wang. 2012. Structural Insights into WHAMM-Mediated Cytoskeletal Coordination during Membrane Remodeling. J. Cell Biol. 199:111–124.

Soldati, T., and M. Schliwa. 2006. Powering Membrane Traffic in Endocytosis and Recycling. Nat. Rev. Mol. Cell Biol.

Soppina, V., A. K. Rai, A. J. Ramaiya, P. Barak, and R. Mallik. 2009. Tug-of-War between Dissimilar Teams of Microtubule Motors Regulates Transport and Fission of Endosomes. Proc. Natl. Acad. Sci. U. S. A. 106:19381–6.

Sossey-Alaoui, K., K. Head, N. Nowak, and J. K. Cowell. 2003. Genomic Organization and Expression Profile of the Human and Mouse WAVE Gene Family. Mamm. Genome 14:314– 22.

Sossey-Alaoui, K., A. Safina, X. Li, M. M. Vaughan, D. G. Hicks, A. V Bakin, and J. K. Cowell. 2007. Down-Regulation of WAVE3, a Metastasis Promoter Gene, Inhibits Invasion and Metastasis of Breast Cancer Cells. Am. J. Pathol. 170:2112–21.

Stadtmueller, B. M., E. Kish-Trier, K. Ferrell, C. N. Petersen, H. Robinson, D. G. Myszka, D. M. Eckert, T. Formosa, and C. P. Hill. 2012. Structure of a Proteasome Pba1-Pba2 Complex: Implications for Proteasome Assembly, Activation, and Biological Function. J. Biol. Chem. 287:37371–82.

Steffen, A., J. Faix, G. P. Resch, J. Linkner, J. Wehland, J. V. Small, K. Rottner, and T. E. B. Stradal. 2006. Filopodia Formation in the Absence of Functional WAVE- and Arp2/3-Complexes. Mol. Biol. Cell 17:2581–91.

Steinberg, F., M. Gallon, M. Winfield, E. C. Thomas, A. J. Bell, K. J. Heesom, J. M. Tavaré, and P. J. Cullen. 2013. A Global Analysis of SNX27-Retromer Assembly and Cargo Specificity Reveals a Function in Glucose and Metal Ion Transport. Nat. Cell Biol. 15:461–71.

115 Stewart, D. M., L. Tian, and D. L. Nelson. 1999. Mutations That Cause the Wiskott-Aldrich Syndrome Impair the Interaction of Wiskott-Aldrich Syndrome Protein (WASP) with WASP Interacting Protein. J. Immunol. 162:5019–24.

Stradal, T. E. B., 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. 14:303–311.

Suetsugu, S., and A. Gautreau. 2012. Synergistic BAR?NPF Interactions in Actin-Driven Membrane Remodeling. Trends Cell Biol. 22:141–150.

Suetsugu, S., H. Miki, and T. Takenawa. 1999. Identification of Two Human WAVE/SCAR Homologues as General Actin Regulatory Molecules Which Associate with the Arp2/3 Complex. Biochem. Biophys. Res. Commun. 260:296–302.

Suetsugu, S., K. Murayama, A. Sakamoto, K. Hanawa-Suetsugu, A. Seto, T. Oikawa, C. Mishima, M. Shirouzu, T. Takenawa, and S. Yokoyama. 2006. The RAC Binding Domain/IRSp53-MIM Homology Domain of IRSp53 Induces RAC-Dependent Membrane Deformation. J. Biol. Chem. 281:35347–35358.

Sun, L., and Z. J. Chen. 2004. The Novel Functions of Ubiquitination in Signaling. Curr. Opin. Cell Biol. 16:119–126.

Takenawa, T., H. Miki, H. Yamaguchi, and S. Suetsugu. 2000. IRSp53 Is an Essential Intermediate between Rac and WAVE in the Regulation of Membrane Ruffling. Nature 408:732–735.

Le Tallec, B., M.-B. Barrault, R. Courbeyrette, R. Guérois, M.-C. Marsolier-Kergoat, and A. Peyroche. 2007. 20S Proteasome Assembly Is Orchestrated by Two Distinct Pairs of Chaperones in Yeast and in Mammals. Mol. Cell 27:660–74.

Le Tallec, B., M.-B. Barrault, R. Guérois, T. Carré, and A. Peyroche. 2009. Hsm3/S5b Participates in the Assembly Pathway of the 19S Regulatory Particle of the Proteasome. Mol. Cell 33:389–399.

Temkin, P., B. Lauffer, S. Jäger, P. Cimermancic, N. J. Krogan, and M. von Zastrow. 2011. SNX27 Mediates Retromer Tubule Entry and Endosome-to-Plasma Membrane Trafficking of Signalling Receptors. Nat. Cell Biol. 13:715–21.

Théry, M., V. Racine, M. Piel, A. Pépin, A. Dimitrov, Y. Chen, J.-B. Sibarita, and M. Bornens. 2006. Anisotropy of Cell Adhesive Microenvironment Governs Cell Internal Organization and Orientation of Polarity. Proc. Natl. Acad. Sci. U. S. A. 103:19771–6.

Torres, E., and M. K. Rosen. 2006. Protein-Tyrosine Kinase and GTPase Signals Cooperate to Phosphorylate and Activate Wiskott-Aldrich Syndrome Protein (WASP)/neuronal WASP. J. Biol. Chem. 281:3513–20.

Ueda, M., R. Gräf, H. K. MacWilliams, M. Schliwa, and U. Euteneuer. 1997. Centrosome Positioning and Directionality of Cell Movements. Proc. Natl. Acad. Sci. U. S. A. 94:9674–8.

Veltman, D. M., and R. H. Insall. 2010. WASP Family Proteins: Their Evolution and Its Physiological Implications. Mol. Biol. Cell 21:2880–93.

Verboon, J. M., T. K. Rahe, E. Rodriguez-Mesa, and S. M. Parkhurst. 2015a. Wash Functions Downstream of Rho1 GTPase in a Subset of Drosophila Immune Cell Developmental Migrations. Mol. Biol. Cell 26:1665–1674.

Verboon, J. M., H. Rincon-Arano, T. R. Werwie, J. J. Delrow, D. Scalzo, V. Nandakumar, M.

116 Groudine, and S. M. Parkhurst. 2015b. Wash Interacts with Lamin and Affects Global Nuclear Organization. Curr. Biol. 25:804–810.

Vilariño-Güell, C., C. Wider, O. A. Ross, J. C. Dachsel, J. M. Kachergus, S. J. Lincoln, A. I. Soto- Ortolaza, S. A. Cobb, G. J. Wilhoite, J. A. Bacon, B. Behrouz, H. L. Melrose, E. Hentati, A. Puschmann, D. M. Evans, E. Conibear, W. W. Wasserman, J. O. Aasly, P. R. Burkhard, R. Djaldetti, J. Ghika, F. Hentati, A. Krygowska-Wajs, T. Lynch, E. Melamed, A. Rajput, A. H. Rajput, A. Solida, R.-M. Wu, R. J. Uitti, Z. K. Wszolek, F. Vingerhoets, and M. J. Farrer. 2011. VPS35 Mutations in Parkinson Disease. Am. J. Hum. Genet. 89:162–167.

Vinogradova, T., P. M. Miller, and I. Kaverina. 2009. Microtubule Network Asymmetry in Motile Cells: Role of Golgi-Derived Array. Cell Cycle 8:2168–74.

Vinogradova, T., R. Paul, A. D. Grimaldi, J. Loncarek, P. M. Miller, D. Yampolsky, V. Magidson, A. Khodjakov, A. Mogilner, and I. Kaverina. 2012. Concerted Effort of Centrosomal and Golgi- Derived Microtubules Is Required for Proper Golgi Complex Assembly but Not for Maintenance. Mol. Biol. Cell 23:820–33.

Volkmann, N., K. J. Amann, S. Stoilova-McPhie, C. Egile, D. C. Winter, L. Hazelwood, J. E. Heuser, R. Li, T. D. Pollard, and D. Hanein. 2001. Structure of Arp2/3 Complex in Its Activated State and in Actin Filament Branch Junctions. Science (80-. ). 293.

Wakida, N. M., E. L. Botvinick, J. Lin, and M. W. Berns. 2010. An Intact Centrosome Is Required for the Maintenance of Polarization during Directional Cell Migration. PLoS One 5:1–12.

Wan, Y., Z. Yang, J. Guo, Q. Zhang, L. Zeng, W. Song, Y. Xiao, and X. Zh Recruited to Cytoplasmic Dynein by Nudel for Efficient Clearance. Cell Res. 22. u. 2012. Misfolded Gβ Is Wang, C., M. Niu, Z. Zhou, X. Zheng, L. Zhang, Y. Tian, X. Yu, G. Bu, H. Xu, Q. Ma, and Y.-W. Zhang. 2016. VPS35 Regulates Cell Surface Recycling and Signaling of Dopamine Receptor D1. Neurobiol. Aging 46:22–31.

Wehrle-haller, B. 2006. The Role of Integrins in Cell Migration. Integrins Dev.:1–24.

Welch, M. D., A. H. Depace, S. Verma, A. Iwamatsu, and T. J. Mitchison. 1997a. The Human Arp2/3 Complex Is Composed of Evolutionarily Conserved Subunits and Is Localized to Cellular Regions of Dynamic Actin Filament Assembly. J. Cell Biol. 138:375–384.

Welch, M. D., A. Iwamatsu, and T. J. Mitchison. 1997b. Actin Polymerization Is Induced by Arp2/3 Protein Complex at the Surface of Listeria Monocytogenes. Nature 385:265–9.

Welch, M. D., J. Rosenblatt, J. Skoble, D. A. Portnoy, and T. J. Mitchison. 1998. Interaction of Human Arp2/3 Complex and the Listeria Monocytogenes ActA Protein in Actin Filament Nucleation. Science (80-. ). 281.

Dictate Downstream Rho Kinase Signaling to Regulate Persistent Cell Migration. J. Cell Biol. White,177. D. P., P. T. Caswell, and J. C. Norman. 2007. αvβ3 and α5β1 Integrin Recycling Pathways

Wigley, W. C., R. P. Fabunmi, M. G. Lee, C. R. Marino, S. Muallem, G. N. DeMartino, and P. J. Thomas. 1999. Dynamic Association of Proteasomal Machinery with the Centrosome. J. Cell Biol. 145:481–90.

Winter, D., A. V. Podtelejnikov, M. Mann, and R. Li. 1997. The Complex Containing Actin-Related Proteins Arp2 and Arp3 Is Required for the Motility and Integrity of Yeast Actin Patches.

Wu, J., and A. Akhmanova. 2017. Microtubule-Organizing Centers. Annu. Rev. Cell Dev. Biol.:1–25.

117 Wu, J., C. de Heus, Q. Liu, B. P. Bouchet, I. Noordstra, K. Jiang, S. Hua, M. Martin, C. Yang, I. Grigoriev, E. A. Katrukha, A. F. M. Altelaar, C. C. Hoogenraad, R. Z. Qi, J. Klumperman, and A. Akhmanova. 2016. Molecular Pathway of Microtubule Organization at the Golgi Apparatus. Dev. Cell 39:44–60.

Wu, S., L. Ma, Y. Wu, R. Zeng, and X. Zhu. 2012. Nudel Is Crucial for the WAVE Complex Assembly in Vivo by Selectively Promoting Subcomplex Stability and Formation through Direct Interactions. Cell Res. 22.

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.

Yamazaki, D., T. Fujiwara, S. Suetsugu, and T. Takenawa. 2005. A Novel Function of WAVE in Lamellipodia: WAVE1 Is Required for Stabilization of Lamellipodial Protrusions during Cell Spreading. Genes to Cells 10:381–392.

Yamazaki, D., T. Oikawa, and T. Takenawa. 2007. Rac-WAVE-Mediated Actin Reorganization Is Required for Organization and Maintenance of Cell-Cell Adhesion. J. Cell Sci. 120:86–100.

Yamazaki, D., S. Suetsugu, H. Miki, Y. Kataoka, S.-I. Nishikawa, T. Fujiwara, N. Yoshida, and T. Takenawa. 2003. WAVE2 Is Required for Directed Cell Migration and Cardiovascular Development. Nature 424:452–6.

Yan, C., N. Martinez-Quiles, S. Eden, T. Shibata, F. Takeshima, R. Shinkura, Y. Fujiwara, R. Bronson, S. B. Snapper, M. W. Kirschner, R. Geha, F. S. Rosen, and F. W. Alt. 2003. WAVE2 Deficiency Reveals Distinct Roles in Embryogenesis and Rac-Mediated Actin-Based Motility. EMBO J. 22:3602–12.

Yewdell, J. W., U. Schubert, L. C. Antón, J. Gibbs, C. C. Norbury, and J. R. Bennink. 2000. Rapid Degradation of a Large Fraction of Newly Synthesized Proteins by Proteasomes. Nature 404:770–774.

Yi, J. Y., K. M. Ori-McKenney, R. J. McKenney, M. Vershinin, S. P. Gross, and R. B. Vallee. 2011. High-Resolution Imaging Reveals Indirect Coordination of Opposite Motors and a Role for LIS1 in High-Load Axonal Transport. J. Cell Biol. 195.

Yvon, A.-M. C., J. W. Walker, B. Danowski, C. Fagerstrom, A. Khodjakov, and P. Wadsworth. 2002. Centrosome Reorientation in Wound-Edge Cells Is Cell Type Specific. Mol. Biol. Cell 13:1871–80.

013. Local Cytoskeletal and Organelle Interactions Impact Molecular-Motor-Driven Early Endosomal Trafficking. Curr. Zajac,Biol. A., Y. 23:1173 Goldman,–1180. E. F. Holzbaur, and E. ?Michae. Ostap. 2

Zavodszky, E., M. N. J. Seaman, K. Moreau, M. Jimenez-Sanchez, S. Y. Breusegem, M. E. Harbour, and D. C. Rubinsztein. 2014. Mutation in VPS35 Associated with Parkinson’s Disease Impairs WASH Complex Association and Inhibits Autophagy. Nat. Commun. 5:1–16.

Zech, T., S. D. J. Calaminus, P. Caswell, H. J. Spence, M. Carnell, R. H. Insall, J. Norman, and L. M. Machesky. 2011. T -Integrin-Mediated Invasive Migration. J. Cell Sci. 124:1–7. he Arp2/3 Activator WASH Regulates α5β1 Zhu, X., and I. Kaverina. 2013. Golgi as an MTOC: Making Microtubules for Its Own Good. Histochem. Cell Biol. 140:361–7.

118 Zimprich, A., A. Benet-Pag?s, W. Struhal, E. Graf, S. Eck, M. Offman, D. Haubenberger, S. Spielberger, E. Schulte, P. Lichtner, S. Rossle, N. Klopp, E. Wolf, K. Seppi, W. Pirker, S. Presslauer, B. Mollenhauer, R. Katzenschlager, T. Foki, C. Hotzy, E. Reinthaler, A. Harutyunyan, R. Kralovics, A. Peters, F. Zimprich, T. Br?cke, W. Poewe, E. Auff, C. Trenkwalder, B. Rost, G. Ransmayr, J. Winkelmann, T. Meitinger, and T. Strom. 2011. A Mutation in VPS35, Encoding a Subunit of the Retromer Complex, Causes Late-Onset Parkinson Disease. Am. J. Hum. Genet. 89:168–175.

Zuchero, J. B., A. S. Coutts, M. E. Quinlan, N. B. La Thangue, and R. D. Mullins. 2009. p53-Cofactor JMY Is a Multifunctional Actin Nucleation Factor. Nat. Cell Biol. 11:451–459.

119

This page intentionally left blank

120

Acknowledgements

121

This page intentionally left blank

122 First and Foremost, I thank my advisor Prof. Alexis Gautreau for inviting me to France in the first place as an intern during my bachelor’s degree. Only because of his motivation and trust, I was inspired to pursue research studies, which led me in obtaining a research oriented master degree and transcending to pursue doctoral study under his supervision. The association that I had with Prof. Alexis Gautreau has influenced me intellectually in a very positive way. He has been a source of motivation during the tough times from the beginning. He being my supervisor, always supported me in my dull times and helped to tackle the hardships that one faces in the business of doing science. Most importantly I have imbibed technical and analytical skills from him. I have also improved my writing skills in wide aspects (a still lot to improve, but definitively better than the beginning) which I am sure will be a great asset in the future. I appreciate all his contributions of time and ideas to make my experience more productive and stimulating. Despite few ruffles, I am grateful to him for the whole enriching experience both professionally and personally.

Thanks to Dr.Raphael Guerois, Dr.Marc Mirande and Dr.Evgeny Denisov to have gracefully accepted to be a part of my jury. I would like to thank Dr.Emmanuel Derivery and Dr.Philippe Chavrier to have accepted to play the role of reviewers, for their time, helpful comments and insights.

Big thanks to all the team members Dominique Lallemand, Anna Polesskaya, Nathalie Rocques, Artem Fokin, Nicolas Molinie and Angelina Chemeris for all the help, guidance and suggestions. Special mentioning –Roman Gorelik, Shashank Shekhar, Shoeb Ahmad, Violane David, Irène Dang, Fabienne Pierre, Maria Lomakina and Xenia Naj. I am very much thankful for Artem Fokin for his critical reading and suggestions in my thesis manuscript. I also thank especially Nicolas Molinie for being very supportive in the lab and helping me with the French written documents.

Finally, I dedicate my successful journey of PhD to my father. He is greatly missed. He has always been very supportive of me in pursuing my dreams. Without his support, all this wouldn’t have been possible. He saw my dreams through his eyes, for that, I am lucky and grateful. I am also dedicating this equally to my mother, she is the reason I am academically inclined. She is a great support and always motivates me.

123 I would also like to thank my wife for her big support and positive role that she played during my PhD journey. She was kind enough to understand and be supportive towards me during my stressful days and as well as forgiving me for being late to home due to long hours of work. Only because of her support and care I could give my best at all times which led to my PhD accomplishment, so I share this success with my wife Aarthi Swaminathan as I joyfully share my life with her. Last but not least, my best friend, Srinivas Manchalu who has always been there for me during my good, bad and worst days and as well for inspiring me about life, research and many more stuff… in the first place.

124

7LWUH5{OHG¶XQHSURWpLQHWULPpULTXHjVXSHUKpOLFHGDQVO¶DVVHPEODJHGXFRPSOH[H:$6+

0RWVFOpV&RPSOH[H:$6+&\WRVTXHOHWWHG DFWLQH$VVHPEODJHGHVFRPSOH[HV0LJUDWLRQGHVFHOOXOHV ,QYDVLRQGHVFHOOXOHV&HQWURVRPH&RPSOH[H$US

5pVXPp/HFRPSOH[H$USJpQqUHGHVUpVHDX[G¶DFWLQHEUDQFKpVTXLSURGXLVHQWXQHIRUFpHGHSRXVVpHSHUPHWWDQWjOD FHOOXOHGHUHPRGHOHUVHVPHPEUDQHV/HFRPSOH[H:$6+DFWLYHOHFRPSOH[H$USjODVXUIDFHGHVHQGRVRPHVHWIDFLOLWH DLQVLODVFLVVLRQPHPEUDQDLUHGHVLQWHUPpGLDLUHVGHWUDQVSRUWVFRQWHQDQWVGHVUpFHSWHXUVLQWHUQDOLVpVWHOVTXHOHVLQWpJULQHV Įȕ'HFHIDLWOHFRPSOH[H:$6+HQIDYRULVDQWOHUHF\FODJHGHVLQWpJULQHVMRXHXQU{OHFUXFLDOGDQVO¶LQYDVLRQGHV FHOOXOHV WXPRUDOHV GXUDQW OD SURJUHVVLRQ WXPRUDOH &HSHQGDQW OH PpFDQLVPH G¶DVVHPEODJH GX FRPSOH[H :$6+ HVW LQFRQQX'DQVFHWWHpWXGHQRXVUDSSRUWRQVO¶LGHQWLILFDWLRQGXSUHPLHUIDFWHXUG¶DVVHPEODJHGXFRPSOH[H:$6+1RXV DYRQVLGHQWLILpODSURWpLQH+6%3JUkFHjXQFULEOHGHVSURWpLQHVTXLVHOLHQWDX[IRUPHVSUpFXUVHXUVGHVVRXVXQLWpVPDLV SOXV DX FRPSOH[H XQH IRLV DVVHPEOp /D UHFRQVWLWXWLRQ ELRFKLPLTXH HW OD PRGpOLVDWLRQ PROpFXODLUH QRXV D SHUPLV GH PRQWUHU TXH +6%3 HVW DVVRFLp DYHF OH SUpFXUVHXU WULPpULTXH &&'& OH GLVVRFLH HW IRUPH XQ KpWpURWULPqUH TXL YD pYHQWXHOOHPHQWOLEpUHUXQHIRUPHPRQRPpULTXHGH&&'&SRXUO¶DVVHPEODJHGXFRPSOH[H:$6+/HU{OHGH+6%3 GDQVO¶DVVHPEODJHGXFRPSOH[H:$6+HVWFRQVHUYp(QHIIHW:$6+HVWGpVWDELOLVpGDQVGHVFHOOXOHVPDPPDLUHVSDUOH NQRFNGRZQGH+6%3HWGDQVO¶DPLEH'LFW\RVWHOLXPSDUOHNQRFNRXWGH+6%3/DGpVWDELOLVDWLRQGXFRPSOH[H:$6+ SDUOHNQRFNRXWGH+6%3SKpQRFRSLHODGpSOpWLRQGH:$6+GDQVO¶DPLEH'LFW\RVWHOLXP'DQVGHVFHOOXOHVKXPDLQHVGH FDUFLQRPHV PDPPDLUHV O¶LQKLELWLRQ GH O¶H[SUHVVLRQ GH +6%3 DOWqUH OH UHF\FODJH GHV LQWpJULQHV j OD PHPEUDQH SODVPLGLTXH ,O HQ UpVXOWH GHV DGKpVLRQV IRFDOHV GpIHFWXHXVHV HW GHV FDSDFLWpV LQYDVLYHV UpGXLWHV 'H SOXV +6%3 HVW ORFDOLVpDX[FHQWURVRPHVHWHVWUHTXLVSRXUODSRODULWpGHVFHOOXOHVORUVGHODPLJUDWLRQ(QILQQRXVDYRQVWURXYpTXHOD VXUH[SUHVVLRQGH+6%3GDQVGHVWXPHXUVPDPPDLUHVHVWDVVRFLpHjXQHDXJPHQWDWLRQGHVQLYHDX[GXFRPSOH[H:$6+ HWjXQPDXYDLVSURQRVWLFSRXUOHVSDWLHQWHVDWWHLQWHVGHFDQFHUGXVHLQ(QFRQFOXVLRQ+6%3HVWXQIDFWHXUG¶DVVHPEODJH FRQVHUYpTXLFRQWU{OHOHVQLYHDX[GXFRPSOH[H:$6+

7LWOH7KHUROHRIDWULPHULFFRLOHGFRLOSURWHLQLQ:$6+FRPSOH[DVVHPEO\

.H\ZRUGV:$6+FRPSOH[$FWLQF\WRVNHOHWRQ0XOWLSURWHLQFRPSOH[DVVHPEO\&HOOPLJUDWLRQ&HOO LQYDVLRQ&HQWURVRPH$USFRPSOH[

$EVWUDFW7KH$USFRPSOH[JHQHUDWHVEUDQFKHGDFWLQQHWZRUNVZKLFKSURGXFHVDSXVKLQJIRUFHWKDWKHOSVWKHFHOOWR UHPRGHO LWV PHPEUDQHV 7KH :$6+ FRPSOH[ DFWLYDWHV WKH $US FRPSOH[ DW WKH VXUIDFH RI HQGRVRPHV DQG WKHUHE\ IDFLOLWDWHVWKHPHPEUDQHVFLVVLRQRIWKHWUDQVSRUWLQWHUPHGLDWHVFRQWDLQLQJLQWHUQDOL]HGUHFHSWRUVVXFKDVĮȕ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

8QLYHUVLWp3DULV6DFOD\ (VSDFH7HFKQRORJLTXH,PPHXEOH'LVFRYHU\ 5RXWHGHO¶2UPHDX[0HULVLHUV5'6DLQW$XELQ)UDQFH