The investigation of Growth-arrest-specific 2 functions in division and its cytoskeleton binding mechanisms

Tong Zhang

Department of Biology McGill University Montreal, Quebec, Canada December 2012

A thesis submitted to McGill University in partial fulfillment of the requirements of the degree of Doctor of Philosophy

© Tong Zhang 2012 Abstract

The eukaryotic cell cytoskeleton plays many important roles in cells and their coordination is crucial for these important biological processes. Their functions are partially achieved by the actin-microtubule (MT) cross-linking . Growth-arrest- specific (Gas) 2 was originally identified in murine fibroblasts under growth-arrest conditions. Gas2 protein contains putative actin- and MT-binding domains, which makes it a strong candidate to be an actin-MT cross-linking protein. However, the mechanisms of Gas2 cytoskeleton binding and its roles in cell division are still unclear. This thesis is focused on these questions.

In Chapter 2, we have determined the Gas2 protein functions in Xenopus laevis embryo cell division and oocyte wound healing process. We have found that the full-length (FL) Gas2 protein and its MT-binding Gas2 domain, but not the actin-binding calponin homology (CH) domain, inhibit cell division and result in multinucleated cells. The observation that Gas2 domain alone can arrest cell division suggests that FL-Gas2 function is mediated by binding MTs. To investigate the mechanism of Gas2-induced cell division arrest, we have used a wound-induced contractile array assay and showed that FL-Gas2 stabilizes MTs.

In Chapter 3, we have further investigated the FL-Gas2 protein cytoskeleton binding mechanisms in HeLa cells. The Gas2 domain itself significantly changes MTs morphology into a long loop-like bundled network. The FL-Gas2 facilitates MT polymerization and changes the dynamic behaviors of MTs. There are three protein kinase C (PKC) phosphorylation sites in the FL-Gas2 protein, and phosphomimetic point mutations reveal that all three are involved in the FL-Gas2 protein functional changes. Size-exclusive chromatography finally reveals that the FL-Gas2 protein forms a dynamic complex, which presumably has an important role in its cytoskeleton binding properties.

Taken together, our data suggest that the non-phosphorylated FL-Gas2 protein forms a dynamic complex to bundle MTs, but its ability to cross-link F-actin and MTs is achieved

2 by the phosphorylated FL-Gas2 protein. We propose that Gas2 function is mediated by binding and bundling MTs, leading to cell division arrest.

3 Résumé

Le cytosquelette participe à de nombreuses fonctions dans les cellules eucaryotes et sa bonne coordination est essentielle au bon fonctionnement de ces fonctions biologiques. Ces fonctions sont partiellement accomplies grâce aux protéines de liaisons des filaments actine-microtubule (MT). La protéine Gowth-arrest-specific (Gas) 2 a été originellement identifiée dans des fibroblastes murins en arrêt de croissance. Le fait que la protéine Gas2 contienne des domaines putatifs d’intéractions avec actine et les microtubules supporte l’hypothèse qu’elle agit comme protéine de liaison des filaments actine-microtubule. Cependant, les mécanismes que Gas2 utilise pour interagir avec le cytosquelette et participer à la division cellulaire ne sont pas connus. Cette thèse répond à ces questions.

Dans le deuxième chapitre, nous avons démontré que la protéine Gas2 participe à la division cellulaire de l’embryon et à la cicatrisation de la plaie de l’ovocyte chez le Xenopus laevis. Nous avons trouvé que la protéine entière (FL) Gas2 et son domaine d’intéraction avec les microtubules, mais pas son domaine d’intéraction avec actine homologue a calponine (CH), inhibent la division cellulaire et résultent en la formation de cellules multinucléées. Nous déduisons du fait que le domaine de Gas2 puisse causer un arrêt de la division cellulaire à lui seul que la protéine entière FL-Gas2 fonctionne grâce à l’intéraction avec les microtubules. Pour étudier comment Gas2 induit l’arrêt de la division cellulaire, nous avons utilisé un essai d’induction de contraction par cicatrisation et démontré que FL-Gas2 stabilise les microtubules.

Dans le troisième chapitre, nous avons étudié en détails comment la protéine FL-Gas2 interagit avec le cytosquelette dans les cellules HeLa. Le domaine de Gas2 change dramatiquement la morphologie des microtubules, celles-ci forme un réseau de boucles longues et regroupées. La protéine FL-Gas2 favorise la polymérisation des microtubules et change leur comportement dynamique. Il existe trois sites de phosphorylation par la protéine kinase C (PKC) dans FL-Gas2. Des points de mutations imitant la phosphorylation de ces trois sites ont démontré que les trois sites participent aux changements de fonctionnalité de FL-Gas2. La chromatographie par exclusion de taille a

4 démontré que FL-Gas2 forme un complexe dynamique qui doit jouer un rôle important dans l’intéraction de la protéine avec le cytosquelette.

En conclusion, nos résultats suggèrent que la protéine FL-Gas2 non-phosphorylée forme un complexe pour regrouper les microtubules, et qu’elle crée une liaison entre F-actine et les microtubules lorsqu’elle est phosphorylée. Nous formulons hypothèse que Gas2 agit en intéragissant et en regroupant les microtubules ce qui cause un arrêt de la division cellulaire.

5 Table of contents

Abstract ...... 2 Résumé ...... 4 Table of contents ...... 6 List of Figures ...... 8 List of Abbreviations ...... 9 Acknowledgements ...... 11 Preface ...... 12 Contribution of Authors ...... 13 Chapter 1 ...... 14 1. GENERAL INTRODUCTION ...... 14 1.1 Gas2 protein ...... 15 1.2 The eukaryotic cytoskeleton ...... 20 1.3 Rho family GTPase ...... 23 1.4 The animal cell division ...... 24 1.5 The single cell wound healing ...... 26 1.6 References ...... 28 Research rationale and objectives ...... 39 Chapter 2 ...... 44 2. GROWTH-ARREST-SPECIFIC PROTEIN 2 INHIBITS CELL DIVISION IN XENOPUS EMBRYOS ...... 44 2.1 Abstract ...... 46 2.2 Introduction ...... 47 2.3 Materials and Methods ...... 49 2.4 Results ...... 56 2.5 Discussion ...... 62 2.6 Acknowledgements ...... 65 2.7 References ...... 66 Connecting statement ...... 81 Chapter 3 ...... 82 3. GROWTH-ARREST-SPECIFIC 2 (GAS2) PROTEIN BUNDLES MICROTUBULES AND FACILITATES MICROTUBULE POLYMERIZATION ...... 82 3.1 Abstract ...... 84 3.2 Introduction ...... 85 3.3 Materials and Methods ...... 87 3.4 Results ...... 94 3.5 Discussion ...... 101 3.6 Acknowledgements ...... 106 3.7 References ...... 107 Chapter 4 ...... 126 4. DISCUSSION ...... 126 4.1 Gas2 protein regulations ...... 127 4.2 Gas2 protein phosphorylation ...... 128 4.3 Gas2 protein dynamic complex ...... 128 4.4 Gas2 facilitates MT polymerization ...... 129

6 4.5 Gas2 and taxol different effects on MTs ...... 129 4.6 Gas2 protein and Rho GTPase ...... 131 4.7 Gas2 knockdown study ...... 131 4.8 Gas2 protein structure ...... 132 4.9 Conclusion ...... 132 4.10 References ...... 134

7 List of Figures

Chapter1 Figure 1 ...... 40 Figure 2 ...... 42

Chapter2 Figure 1 ...... 70 Figure 2 ...... 72 Figure 3 ...... 74 Figure 4 ...... 76 Figure 5 ...... 78 Figure 6 ...... 80

Chapter3 Figure 1 ...... 111 Figure 2 ...... 113 Figure 3 ...... 115 Figure 4 ...... 117 Figure 5 ...... 119 Figure 6 ...... 121 Figure 7 ...... 123 Figure 8 ...... 125

8 List of Abbreviations

A alanine ADF actin depolymerizing factor ADP adenosine diphosphate APC adenomatous polyposis coli Arp2/3 actin relative protein 2 and 3 ATP adenosine triphosphate BSA bovine serum albumin CH calponin homology D aspartic acid DMSO dimethyl sulfoxide EB1 end binding protein 1 EB3 end binding protein 3 EGF epidermal growth factor eIF4E eukaryotic translation initiation factor 4E EM electron microscopy ER endoplasmic reticulum F-actin filamentous actin FCS fetal calf serum FL full-length FRAP fluorescence recovery after photobleaching G0 growth arrest or gap 0 G1 gap 1 G2 gap 2 G2L Gas2-like GAP GTPase activating protein GAR Gas2-related Gas growth-arrest-specific GDP guanosine diphosphate GDI guanine nucleotide dissociation inhibitor

9 GEF guanine nucleotide exchange factor GFP green fluorescent protein GTP guanosine triphosphate GTPase guanosine triphosphatase hGAR17 human Gas2-related on 17 hGAR22 human Gas2-related gene on chromosome 22 IF(s) intermediate filament(s) LC low complexity LPA lysophosphatidic acid M mitosis MARCKS myristoylated alanine-rich protein kinase C substrate Mdm2 murine double minute 2 MT(s) microtubule(s) MTOC microtubule organization center NDP nucleotide diphosphate NTP nucleotide triphosphate OR-2 oocyte ringer-2 P proline PBS phosphate buffered saline PDGF platelet derived growth factor Pigs pickled eggs PKC protein kinase C PMA phorbol myristate acetate PRC1 protein regulator of cytokinesis 1 RFP red fluorescent protein RNAi RNA interference S serine SDS-PAGE sodium dodecyl sulfate polyacrylamide gel electrophoresis T threonine TSB tubulin stabilization buffer

10 Acknowledgements

I would like to sincerely thank my dear supervisor, Prof. Craig Mandato, for his guidance and support during my graduate studies. Craig has not only helped me to develop independent thinking ability, but also always encouraged me to try new things. He has also helped me to improve my oral presentation and writing skills. I always feel very lucky to choose him as my supervisor and have my graduate studies in a warm and friendly lab. Craig is not only a good mentor, but also a great friend. He always takes care of my life in and out the lab, which makes me feel less lonely as a foreigner. The lab parties, Christmas and Chinese New Year dinners and hockey games; they all have become unforgettable moments in my life.

I would also like to thank the past and present Mandato lab members for their help and advice on my research projects. I would like to specially thank Ms. Bama Dayanandan, the lab former technician. She taught me the skills to do neat experiments.

Thanks to my past and present committee members, Profs. Martin Latterich, Jackie Vogel and Gary Brouhard for their advice on my research projects.

I would also like to thank to Prof. Jacque Paiement for providing the Gas2 antibody; Prof. Marko Horb for providing the experimental frogs; Dr. Wolfgang Reintsch for helping on the cyro technique; Prof. Bill Bement and Dr. Tadayuki Shimada for providing constructs; Dr. Dirk Hubmacher for helping on the chromatography; Mr. Ken McDonald for helping on the flow cytometry; Dr. Judith Lacoste and Dr. Guillaume Lesage for instruments support. Especially thank to Dr. Laetitia Sabatier for translating my thesis abstract into French.

Finally, I would like to take this opportunity to specially thank my dear parents for their love and support. They provided me an unusual opportunity to study at one of the best universities in the world. I have not only learnt the advanced knowledge, but also broadened my vision. It has changed the way of how I look at this world forever.

11 Preface

This is a manuscript-based thesis. It consists four chapters. The general introduction (Chapter 1) provides literature reviews on topics of Gas2 protein, cytoskeleton, Rho GTPase, cell division and single cell wound healing. Two research chapters (Chapter 2 and 3) discussed the Gas2 functions in cell division and its cytoskeleton binding mechanisms. Chapter 2 has been published in journal PLoS ONE. Chapter 3 has been written a manuscript format and it is ready to be submitted. Both Chapter 2 and 3 include their own Abstract, Introduction, Materials and Methods, Results and Discussion as a standard manuscript format. Each of these chapters is an independent paper, with their own Reference sections. And a final chapter (Chapter 4) is for the further discussion and conclusion. This thesis was entirely prepared by myself. All chapters including the first version of the manuscripts were written by me. The contributions of co-authors in Chapter 2 and 3 have been list in , and contributions of people who were not authors have been mentioned in .

12 Contribution of Authors

Chapter 2: “Growth-arrest-specific protein 2 inhibits cell division in Xenopus embryos”

Published in PLoS ONE. 2011 Sep. 6 (9): e24698.

Tong Zhang: Designed experiments, performed experiments, analyzed experimental data, wrote and edited the manuscript. Bama Dayanandan: Performed construct cloning. Isabelle Rouiller: Performed electron microscope experiments and edited the manuscript. Elizabeth J. Lawrence: Edited and wrote the manuscript. Craig A. Mandato: Designed experiments, performed experiments, edited and wrote the manuscript.

Chapter 3: “Growth-arrest-specific 2 (Gas2) Protein Bundles Microtubules and Facilitates Microtubule Polymerization”

This chapter has been written as a manuscript and it is ready to be submitted.

Tong Zhang: Designed experiments, performed experiments, analyzed experimental data, wrote and edited the manuscript. Bama Dayanandan: Performed construct cloning and part of the point mutation cloning. Michael T. Greenwood: Edited and wrote the manuscript. Craig A. Mandato: Designed experiments, edited and wrote the manuscript.

13 Chapter 1

1. GENERAL INTRODUCTION

14 The eukaryotic cell cytoskeleton system, especially actin and microtubules (MTs) are not only very dynamic, but also involve in many fundamental cellular processes, such as cell division, vesicle trafficking, cell movement and wound healing (Abreu-Blanco et al., 2012; Amano et al., 2010; Anitei and Hoflack, 2012; Dent et al., 2011; Heng and Koh, 2010). Together with many cytoskeleton binding proteins, motor proteins and cross- linking proteins, actin and MTs form a complex network to play those important roles. The coordination of actin and MTs are crucial for these important biological processes, which is also partially achieved by the actin-MT cross-linking proteins. Many actin-MT cross-linking proteins have been identified; however, their functions and regulation mechanisms are still not well understood (Rodriguez et al., 2003). Growth-arrest-specific (Gas) 2 protein was discovered in a genetic screen in murine fibroblasts under growth- arrest condition (Schneider et al., 1988). The Gas2 protein was identified as a potential actin-MT cross-linking protein (Brancolini et al., 1992; Sun et al., 2001). However, its cytoskeleton binding mechanisms and roles in cell division are still unclear. This thesis discussed my research on these questions by using modern cell biology techniques.

1.1 Gas2 protein

1.1.1 Gas2 protein expression Gas2 protein is widely expressed in human tissues and Western blot analysis has shown that the Gas2 protein expresses at highest level in liver, lung, kidney and brain (Collavin et al., 1998). The human Gas2 protein shares both structural and functional homology to murine Gas2 and they also have a similar pattern of tissue distribution and intracellular localization (Collavin et al., 1998). Mus musculus (house mouse) [P11862] Gas2 protein consists of 314 amino acids and bioinformatics analysis of its protein sequence reveals that mouse Gas2 protein shares very high amino acids homology to Homo sapiens (human) [O43903], Canis familiaris (Dog) [F1PBV2], Bos taurus (Bovine) [A8E4Q5], Gallus gallus (Chicken) [F1NSM4] and Xenopus tropicalis (Silurana tropicalis) [ENSXETP00000005555] (Zhang et al., 2011). This illustrates that Gas2 protein is highly conserved during evolution and may share the same important biological functions among different species.

15 1.1.2 Gas2 protein regulations Gas2 protein belongs to the growth-arrest-specific protein family, which was originally identified in a genetic screen in murine fibroblasts under growth-arrest condition or called G0 phase in cell cycle (Schneider et al., 1988). The half-life of Gas2 protein is longer than 12 hours and its intracellular level does not change dramatically during G0 to G1 transition; therefore, it is believed that there may be a regulated mechanism, such as phosphorylation in Gas2 protein (Brancolini et al., 1992). These are some evidences for this hypothesis. For instance, the phosphorylated serine (S) in Gas2 protein is detected by phosphoaminoacid analysis after serum addition and gas2 mRNA expression decreases during the first few hours and following by the down-regulation of Gas2 protein (Brancolini et al., 1992; Schneider et al., 1988).

The cell motility is restricted in growth-arrested cells; therefore, there may be some regulatory proteins in control of the cytoskeleton restrictions (Conrad et al., 1993). The Gas2 phosphorylation may be one of regulatory mechanism to relieve this restriction. It has been shown that Gas2 protein localizes at the cell border in serum starved NIH 3T3 cells. When it is hyper-phosphorylated, Gas2 re-localizes to the newly formed membrane ruffles after serum or growth factor addition. This suggests a relationship between the Gas2 protein hyper-phosphorylation and its subcellular location changes. The hyper- phosphorylated Gas2 protein can be induced by addition of fetal calf serum (FCS), platelet derived growth factor (PDGF) and phorbol myristate acetate (PMA), but not by lysophosphatidic acid (LPA). The phosphorylated Gas2 is detected to localize at the membrane ruffles, which is known to be caused by actin polymerization at the inner surface of the membrane (Stossel, 1993). LPA can only induce cells to form actin stress fiber, but not Gas2 protein hyper-phosphorylation (Brancolini and Schneider, 1994). Therefore, the hyper-phosphorylated Gas2 protein should only be found in the cell ruffles, but not in the stress fibers.

1.1.3 Gas2 protein and cytoskeleton

16 Gas2 protein contains two cytoskeleton binding domains: a putative actin-binding calponin homology (CH) domain near its N-terminus (Fig. 1, for mouse Gas2 [P11862] amino acids#: 36-158), and a MT-binding Gas2 domain, which is also called Gas2- related (GAR) domain, near its C-terminus (amino acids#: 201-274), which makes it as a good candidate for an actin-MT cross-linking protein. There are 2 low complexity (LC) domains (amino acids#: 167-178 and 181-198) between the CH and Gas2 domains, and the second LC domain contains 4 proline-serine (P-S) repeats (amino acids#: 183-190 shown in the boxed region in its sequence), which gives this region more structural flexibility (Zhang et al., 2011). The early immuno-fluorescence studies have demonstrated that the Gas2 protein co-localizes with F-actin in growth-arrest NIH 3T3 fibroblasts and the C-terminus of Gas2 protein is observed in co-localizing with MTs in COS-7 cells (Brancolini et al., 1992; Sun et al., 2001). When over-expressing the C- terminus of Gas2 protein, it can also induce a dramatic actin re-arrangement (Brancolini et al., 1995). However, no observations about Gas2 cross-linking actin and MT have been reported by date. We have shown that Gas2 co-localizes with MTs at the cell cortex of Gas2-injected Xenopus embryos using cryo-confocal microscopy and it is co-sedimented with MTs in cytoskeleton co-sedimentation assays (Zhang et al., 2011). Gas2 stabilizes MTs in a wound-induced contractile array assay. Our electron microscopy (EM) studies have demonstrated that Gas2 bundles MTs into higher-order structures (Zhang et al., 2011).

1.1.4 Gas2, and development Studies have shown that Gas2 protein is involved in the regulations of keratinocytes and muscle differentiation (Cowled et al., 1994; Manzow et al., 1996). The cartilage development in the limb is a complex process involving cell condensation, cell movement and apoptosis (Lewinson and Silbermann, 1992). All of these processes need the cell cytoskeleton arrangement. The limb has been used as a classic model for studying apoptosis for years (Lee et al., 1993). It has been shown that caspase 3 and 7 cleave at aspartic acid (D) 279 in Gas2 protein and the death substrate of Gas2 binds F-actin and changes actin morphology (Brancolini et al., 1995; Cotter et al., 1992; Sgorbissa et al., 1999). The gas2 gene has been shown to express in the interdigital tissues, chondrogenic

17 regions and myogenic regions during mouse embryo development. Gas2 is also involved in the execution of apoptosis in hindlimb interdigital tissues (Lee et al., 1999). The Gas2 protein has been found strongly expressed in the intervertebral tissues and the craniofacial mesenchyme regions that normally undergo extensive apoptosis. In the mouse cartilage of the limbs, cells are strongly expressed Gas2 and caspase 3. On the other hand, even heart has apoptosis during development, but no Gas2 expression is found by Western blot (Lee et al., 1999). The functions of Gas2 protein in development are not only in apoptosis, but also may be involved in its function in F-actin binding. Gas2 has been thought to be related to the recruitment of cells to form the chondrogenic nodules. Gas2 protein is expressed more strongly in the proliferative cells than in the differentiating cells. Together with its roles in apoptosis, increasing the level of Gas2 protein may help to control keratinocyte growth by avoiding apoptosis (Manzow et al., 1996).

1.1.5 Gas2 protein and cell division The gas2 gene and protein levels are up-regulated upon serum starvation in NIH 3T3 cells, which also results in cells arresting at G0 (Schneider et al., 1988). Conversely, Gas2 protein level is down-regulated with serum and growth factor stimuli, allowing for cell cycle progression (Brancolini et al., 1992). These two observations imply a relationship between cell cycle progression and Gas2 protein levels. It has also been shown that there are fewer mitotic cells in the chondrogenic regions, where are expressing gas2 gene (Lee et al., 1999). We have investigated this relationship by over- expressing Gas2 protein, CH and Gas2 domains in Xenopus embryos. Both the Gas2 protein and Gas2 domain, but not the CH domain, inhibit cell division and result in multinucleated cells. The observation that Gas2 domain alone can arrest cell division suggests that Gas2 function is mediated by MT binding (Zhang et al., 2011).

1.1.6 Gas2 protein and cancer Human gas2 gene localizes at Chromosome 11p14.3-p15.2 and re-arrangements or deletion of the short arm (the p region) of human are frequently found in tumors and variety of sporadic human cancers (Lee and Feinberg, 1997; Li et al., 1997;

18 Mitelman et al., 1997; Steiner et al., 1997). The analysis of oncogene transformed cells v- fos, v-myc, v-ras and v-src has showed that the Gas2 expression fails to increase in response to serum starvation (Brancolini et al., 1992). Gas2 protein inhibits activity to enhance dependent apoptosis in the murine double minute 2 (Mdm2) independent manner (Benetti et al., 2001). Proteomic studies identified the Gas2 protein in normal rat liver cells, but not in rat liver tumors (Mazzoni et al., 2005). The eukaryotic translation initiation factor 4E (eIF4E) binding proteins control p53-dependent senescence by regulating Gas2 translation (Petroulakis et al., 2009). The Gas2 protein expression levels are decreased in prostate cancer cells (Kondo et al., 2008). Overall, the relationship of Gas2 with many types of cancers implies that it may work as a cancer suppressor protein.

1.1.7 Gas2-like (G2L) protein family Three G2L proteins have been discovered in human and they all share the same core domains: the CH and Gas2 domains (Stroud et al., 2011). The first two G2L proteins were discovered by either tumor suppressor gene searching or by gene sequence similarity comparison, so they were also called the human Gas2-related gene on chromosome 22 (hGAR22) or G2L1 and the human Gas2-related gene on chromosome 17 (hGAR17) or G2L2 (Goriounov et al., 2003; Zucman-Rossi et al., 1996). The G2L1, which shares 35% identical amino acids sequence to both human and murine Gas2 protein, has been shown that it is a tumor suppressor gene on Chromosome 22q12, a region is frequently involved in many types of cancers (Zucman-Rossi et al., 1996). Both of G2L1 and G2L2 have 2 isoforms and they are named as G2L1/ and G2L2/. The  isoform has larger size comparing with  isoform by an extra long C-terminus amino acid sequence, which is also the main difference between Gas2 and G2L proteins (Fig. 2). This C-terminus tail may form another structure and directly impact on the G2L proteins cytoskeleton binding properties. It has been showed that the Gas2 domain in G2L3 protein is not essential for its localization to MTs and it does not directly interact with MTs in vitro (Stroud et al., 2011). This long C-terminus tail may directly impact on G2L3 protein functions including its ability of binding MTs. In a recent paper, G2L3 actin-MT crosslinking ability has also been confirmed by biochemical experiments and G2L3 has

19 been found at the midbody during cytokinesis (Wolter et al., 2012). When knocking down G2L3, it results abnormal chromatin oscillation during cytokinesis (Wolter et al., 2012). G2L3 presumably plays a regulating role in cytoskeleton during cell division. A Gas2-like protein called Pickled eggs (Pigs) in Drosophila co-localizes with both F-actin and MTs in the giant epithelial cells. Pigs is postulated to facilitate cytoskeleton rearrangements by binding and stabilizing F-actin and MTs (Pines et al., 2010).

1.2 The eukaryotic cytoskeleton In eukaryotes, cytoskeleton includes actin, MT and intermediate filament (IF). They are not only support cell structure, but also play many essential roles, such as in cell polarity, cell motility, cell division and wound healing (Abreu-Blanco et al., 2012; Amano et al., 2010; Anitei and Hoflack, 2012; Dent et al., 2011; Heng and Koh, 2010). Filamentous actin (F-actin) or actin filament, which is also called microfilament, is approximately 6 nm in diameter twisted cable filament. MT is approximately 25 nm in diameter hollow tubular structure, which is usually made with 13 parallel strands in vivo. Both actin monomer and tubulin dimer bind nucleotide triphosphate (NTP) to polymerize into polymers. They hydrolyze NTP to nucleotide diphosphate (NDP) after polymerization releasing a phosphate group. The NDP bound polymer is less stable than NTP bound form and de-polymerization will occur to release individual subunits for another round of polymerization and de-polymerization cycle. This cycle repeats in different places inside a cell to execute different tasks. IFs sizes are between F-actin and MTs. Both F-actin and MTs are polar polymers, but IFs are not. The polarity of F-actin and MTs are essential for their motor protein unidirectional movement. The three components of eukaryotic cytoskeleton have different yet connected roles in cells. For instance, F-actin provides the cell mechanical support and cell motility. MTs are very important for separation and intracellular vesicle transportation. IFs function as tendons to resist the mechanical force. Cytoskeleton cross-linking proteins contribute the roles to integrate all three cytoskeleton components together into a complex network. This network can sense the environment changes and even influence certain and cell differentiation (Discher et al., 2009).

20 1.2.1 Actin Actin together with myosin-2 were first discovered in muscle due to their abundance (Halliburton, 1887). It was later discovered in other types of cells. Actin is recognized as one of the most abundant globular protein in eukaryotic cells (Adelman and Taylor, 1969; Hatano and Oosawa, 1966). There are 6 actin isoforms in human and they are synthesized from multiple rather than from . The  and  actin isoforms are found in non-muscle cells. Four different  and  isoforms in various muscles. Only  isoform is in the red blood cells.  isoform is in the plasma membrane and  isoform in the stress fibers. In muscle cells,  isoform in the thin filament and  isoform around the mitochondria (Herman, 1993; Khaitlina, 2001).

F-actin is a polar polymer and its one end grows faster than the other one. Actin and F- actin turn over mechanism was first proposed by the treadmilling theory (Wegner, 1976). The driving power of this cycle is adenosine triphosphate (ATP) hydrolysis. Actin monomer can bind both of ATP or adenosine diphosphate (ADP). Once actin polymerizes into F-actin, it hydrolyzes ATP into ADP and slowly releases the phosphate group. Actin polymerization is the driving force of cell motility. ATP-actin adding to one end of polymer F-actin and the hydrolyzed ADP-actin disassociate from the other end of F-actin. This model has been confirmed by time lapse imaging (Fujiwara et al., 2002). The intracellular actin organization is carefully regulated to provide cell abilities in response to external stimuli and cell cycle response (Aderem, 1992; Ridley, 1994; Stossel, 1993).

The actin relative protein 2 and 3 (Arp2/3) initiates F-actin branch growth and this is the first step to change the cell shape and make the crawling protrusion for cell locomotion. Formin promotes F-actin growth at the fast growing end of F-actin by prevent other capping protein to bind on it, so that actin grows into long non-branched thin protrusion, which is called filopodia (Yang et al., 2007). The actin depolymerizing factor (ADF)– cofilin family promotes actin to turn over in eukaryotic cells (Maciver and Hussey, 2002). ADF-cofilin selectively binds ADP-actin and catalyze their de-polymerization by

21 changing the binding actin structure (Blanchoin and Pollard, 1999; Carlier et al., 1997; McGough et al., 1997).

1.2.2 Microtubule (MT) The - and -tubulin form a heterodimer, which is the building block of MTs. These heterodimers join together to form a linear structure called protofilament. Thirteen protofilaments finally form a 25nm hollow tubular structure in vivo—the MT (Mandelkow and Mandelkow, 1989).

MTs polymerization mechanism was first thought the same as actin treadmilling. Later, it has been shown that the dynamic instability is the dominant kinetic behavior for MTs. The guanosine triphosphate (GTP)-tubulin grows at the MT fast grow end, which is also called the plus end; and the guanosine diphosphate (GDP)-tubulin disassociates from the slow grow end, which is also called the minus end. The growth and shrinkage make MT dynamic instability (Mitchison and Kirschner, 1984). The GTP-tubulin also forms a cap to protect the MTs at the plus end from depolymerization. Loss the GTP cap by hydrolysis will trigger MT depolymerization. In cells, MTs have their minus ends anchored at the centrosome to  tubulin, which is usually positioned at microtubule organization center (MTOC) (Mandelkow and Mandelkow, 1990). The -tubulin localizes at MTOC, where -tubulin starts its nucleation of the growth of MTs (Oakley and Oakley, 1989).

MTs play many important roles in cells, such as long distance intracellular trafficking, cell division and neurite outgrowth in the differentiated neurons (Drubin and Nelson, 1996). MTs play a significant role in animal cell mitosis for chromosomes separation and establishing the cleavage furrow between two daughter cells. There are three classes of MTs in metaphase on the basis of their functions and dynamic behaviors. The kinetochore MTs have their plus end embedded in kinetochores and the minus end at or near the spindle poles. The interpolar MTs are through the spindles, but not attach to the kinetochore. Only a minority of interpolar MTs reach the pole and most of minus end MTs are near the pole but are not physically linked to it. So, they are free at both ends.

22 The central spindle contains a large number of anti-parallel MTs. The astral MTs project out the poles and interact with cell cortex. These astral MTs interact with the cell cortex and determine the cleavage furrow in cytokinesis (Glotzer, 2009; Straight and Field, 2000).

The end binding protein family was first discovered as a binding partner of adenomatous polyposis coli (APC) (Su et al., 1995). End binding protein 1 (EB1) binds to the plus end of MTs and it can also bind to the minus end in vitro (Tirnauer et al., 2002). EB1 facilitates MTs polymerization by stabilizing their protofilaments (Vitre et al., 2008).

1.2.3 Intermediate filament (IF) The IFs are relative larger family comparing with actin and MTs. They are the products of 73 genes in human and form 6 classes protein family (Szeverenyi et al., 2008). IFs are relatively stable and not sensitive to temperature, high concentration salts and even some detergents. IFs partially overlapping to form parellel molecular dimers (tetramers) and grow at both ends. The average diameter is approximately 10 nm. They are involved in cell differentiation and tissue development. IFs are regulated by their binding proteins including signaling molecules, apoptosis-related proteins, kinases and phosphatases (Green et al., 2005). IFs are degraded by caspases during apoptosis through for ubiquitination and proteosomal degradation (Ku and Omary, 2000; Marceau et al., 2007). IFs are very abundant in cells and they are also thought as stress proteins since they can be up-regulated by variety of stresses (Goldman et al., 2008; Zhong et al., 2004). It has been shown that IFs are also often altered in cancers (Petrak et al., 2008).

1.3 Rho family GTPase Twenty-two Rho family guanosine triphosphatase (GTPase) proteins have been discovered and the main differences among them are their subcellular locations and binding partners (Ridley, 2006). They are involved in many important cellular processes, such as cell transformation, cell cycle progression and cell polarity (Johnson, 1999; Olson et al., 1995; Qiu et al., 1995). RhoA, Cdc42 and Rac1 are the three best studied Rho family GTPase proteins. RhoA localizes at the plasma membrane and inside the cytosol.

23 Cdc42 localizes at plasma membrane and inside the Golgi. Rac1 localizes at plasma membrane (Ridley, 2006). The regulations are controlled by three classes of proteins, GTPase activating protein (GAP), guanine nucleotide exchange factor (GEF) and guanine nucleotide dissociation inhibitor (GDI) (Raftopoulou and Hall, 2004). Rho family GTPase proteins are inactive when they are bound to GDP in the complex with GDI. Dissociation of GDI will allow GEF to bind GTPase proteins. GEF releases GDP to exchange to GTP. The GTP bound active GTPase proteins binds and activates the downstream proteins. GAP binds the active GTPase and GTPase proteins will hydrolysis GTP to GDP and become inactive again (Jaffe and Hall, 2005).

The most dominant roles of these Rho family GTPase proteins are their ability of regulation of actin cytoskeleton. RhoA, Cdc42 and Rac1 can induce F-actin to form different characteristic structures in tissue cultured cells. For example, LPA binds to the G-protein-coupled receptor to activate RhoA and generate stress fibers (Ridley and Hall, 1992). The PDGF and epidermal growth factor (EGF) bind to the tyrosine kinase receptors to activate Rac1 to generate lamellipodia and membrane ruffling (Ridley et al., 1992). The bradykinin binding to its G-protein-coupled receptor activates Cdc42 to induce filopodia formation (Nobes and Hall, 1995).

1.4 The animal cell division The cell division or mitosis can be divided into prophase, prometaphase, metaphase, anaphase, telophase and cytokinesis. There are different events occurring in different stages. The chromosomes condensation happens in prophase and MTs become more dynamic, which also triggers the duplicated centrosomes separation. The cell nuclear envelope breaks down and the chromosomes are randomly attached to the MTs. Once both kinetochores on a chromosome are attached to the opposite sides of spindle poles, the chromosomes line up to the equator of the cell in prometaphase. All chromosomes are properly attached on their kinetochores by spindle MTs and ready to be separated into two daughter cells. The anaphase can be divided into A and B two stages. In anaphase A, chromosomes are separated by shortening kinetochore MTs. Both motor proteins and shortening the kinetochore MTs contribute to chromosomes movement in this process. In

24 anaphase B, the poles are separated by elongating of the spindle MTs. The mitotic spindle also starts to activate the cell cortex in preparation for cytokinesis. The nuclear envelope re-assembly starts in anaphase and will complete in telophase. The cytokinesis cleavage furrow is stimulated by the mitotic spindles, where they are perpendicular to the long axis of the spindles. At the end of cytokinesis, an intracellular bridge called the midbody is formed. It contains organized anti-parallel MT arrays. Cytokinesis is very important for a cell fate and any error will lead to aneuploid cells, which is the hallmark for many types of cancer cells (Scholey et al., 2003).

To succeed in cytokinesis, first of all, the cleavage furrow must be at the right position, so that the duplicated chromosomes can be equally separated into two daughter cells. Second, an actomysion based contractile ring forms and generates force to pitch the mother cell membrane. At the end, the two daughter cells are completely separated and sealed with new membrane (Gonzalez, 2003; Scholey et al., 2003). RhoA has been shown to be a key regulator for the cleavage furrow formation. Deletion of RhoA blocks cytokinesis (Bement et al., 2005; Canman et al., 2008; Miller et al., 2008). The spindle MTs are very important to define the Rho active zone since the location of the cleavage furrow are determined by MTs (Rappaport, 1971). A group of MTs are specifically stabilized at the cell cortex, where the cleavage furrow forms. These MTs are thought to be mark the site of the contractile ring formation (Canman et al., 2003; Foe and von Dassow, 2008). Anillin is an important component of the contractile ring. Anillin binds and bundles F-actin. Together with myosin-2, they build the contractile ring machinery in cytokinesis (Oegema et al., 2000; Sisson et al., 2000; Surka et al., 2002).

There are many evidences supporting actin-MT interactions during cell division. For example, actin is observed to be transported towards the astral MTs in Xenopus egg extracts (Sider et al., 1999; Waterman-Storer et al., 2000). MTs may transport actomyosin away from the MTOC in Drosophila embryos (Foe et al., 2000). Anillin and the septins are good candidates for actin-MT interactions since both of them are involved in cytokinesis and they also have both actin- and MT-binding affinity (Oegema et al., 2000; Sisson et al., 2000; Surka et al., 2002)

25 1.5 The single cell wound healing Many studies have shown that the single cell wound healing is a conserved process in metazoan (McNeil and Steinhardt, 2003; Schapire et al., 2009). When a cell is wounded, it must first repair the wound to prevent substance, especially calcium, from the outside environment getting into the cell and also prevent cell contents from lost (Sonnemann and Bement, 2011). To repair the wound, a cell first seals the wound with a membrane patch by exocytosis (McNeil et al., 2000). The origin of this new membrane patch is not well defined, but there are evidences showing that fused lysosomes and endoplasmic reticulum (ER) are the sources of this membrane patch (Jaiswal et al., 2002; Reddy et al., 2001) and some researchers also believe that this membrane patch is simply made by any membrane compartments near the wound (Bement et al., 2007; Sonnemann and Bement, 2011). The membrane fusion is thought to be controlled by several regulators including annexins and dysferlin (Bansal et al., 2003; McNeil et al., 2006). The MT dependent membrane trafficking may be involved in the wound repairing (Sonnemann and Bement, 2011). The calcium will trigger MT depolymerization at the wound site and then followed MT polymerization and transported to the wound via actomyosin contraction (Mandato and Bement, 2003). It has been shown that vesicles from Golgi are transported along the MTs to the wound site (Togo, 2006). F-actin and myosin-2 form the contractile machinery around the wound to heal it (Bement et al., 1999; Godin et al., 2011). This contractile machinery shares many key characters with the contractile ring in cytokinesis (Zhang et al., 2011).

The single cell wound healing mechanisms have been well studied in Xenopus oocytes, which are relatively large single cell comparing with a tissue cultured cell (1 mm in diameter vs. 50 μm on average); therefore, it provides an optimal condition for a high resolution imaging (Zhang and Mandato, 2007). The actomyosin contractile array is made by local actin polymerization and myosin-2 around the wound site, and cortical flow also brings nearby actin and MTs towards the wound (Mandato and Bement, 2001). MTs are also assembled into a radial array surrounding the wound. Actin-MT interaction is critical for wound healing process. MTs are observed to be transported with actin

26 towards the wound border, which is suggested there is a structural linkage between them (Mandato and Bement, 2003). RhoA and Cdc42 trigger the local F-actin and myosin-2 assembly. The active Rho directs myosin-2 localizing at the wound edge and the active Cdc42 controls actin dynamics around the wound (Benink and Bement, 2005).

27 1.6 References

Abreu-Blanco, M. T., Watts, J. J., Verboon, J. M. and Parkhurst, S. M. (2012). Cytoskeleton responses in wound repair. Cellular and molecular life sciences : CMLS 69, 2469-83. Adelman, M. R. and Taylor, E. W. (1969). Isolation of an Actomyosin- Like Protein Complex from Slime Mold Plasmodium and Separation of Complex into Actin- and Myosin-Like Fractions. Biochemistry 8, 4964-&. Aderem, A. (1992). Signal transduction and the actin cytoskeleton: the roles of MARCKS and profilin. Trends in biochemical sciences 17, 438-43. Amano, M., Nakayama, M. and Kaibuchi, K. (2010). Rho-kinase/ROCK: A key regulator of the cytoskeleton and cell polarity. Cytoskeleton 67, 545-54. Anitei, M. and Hoflack, B. (2012). Bridging membrane and cytoskeleton dynamics in the secretory and endocytic pathways. Nature cell biology 14, 11-9. Bansal, D., Miyake, K., Vogel, S. S., Groh, S., Chen, C. C., Williamson, R., McNeil, P. L. and Campbell, K. P. (2003). Defective membrane repair in dysferlin-deficient muscular dystrophy. Nature 423, 168-172. Bement, W. M., Benink, H. A. and von Dassow, G. (2005). A microtubule-dependent zone of active RhoA during cleavage plane specification. The Journal of cell biology 170, 91-101. Bement, W. M., Mandato, C. A. and Kirsch, M. N. (1999). Wound- induced assembly and closure of an actomyosin purse string in Xenopus oocytes. Current biology : CB 9, 579-87. Bement, W. M., Yu, H. Y., Burkel, B. M., Vaughan, E. M. and Clark, A. G. (2007). Rehabilitation and the single cell. Current opinion in cell biology 19, 95-100. Benetti, R., Del Sal, G., Monte, M., Paroni, G., Brancolini, C. and Schneider, C. (2001). The death substrate Gas2 binds m-calpain and increases susceptibility to p53-dependent apoptosis. The EMBO journal 20, 2702-14. Benink, H. A. and Bement, W. M. (2005). Concentric zones of active RhoA and Cdc42 around single cell wounds. The Journal of cell biology 168, 429-39.

28 Blanchoin, L. and Pollard, T. D. (1999). Mechanism of interaction of Acanthamoeba actophorin (ADF/cofilin) with actin filaments. Journal of Biological Chemistry 274, 15538-15546. Brancolini, C., Benedetti, M. and Schneider, C. (1995). Microfilament reorganization during apoptosis: the role of Gas2, a possible substrate for ICE-like proteases. The EMBO journal 14, 5179-90. Brancolini, C., Bottega, S. and Schneider, C. (1992). Gas2, a growth arrest-specific protein, is a component of the microfilament network system. The Journal of cell biology 117, 1251-61. Brancolini, C. and Schneider, C. (1994). Phosphorylation of the growth arrest-specific protein Gas2 is coupled to actin rearrangements during Go-->G1 transition in NIH 3T3 cells. The Journal of cell biology 124, 743-56. Canman, J. C., Cameron, L. A., Maddox, P. S., Straight, A., Tirnauer, J. S., Mitchison, T. J., Fang, G., Kapoor, T. M. and Salmon, E. D. (2003). Determining the position of the cell division plane. Nature 424, 1074-8. Canman, J. C., Lewellyn, L., Laband, K., Smerdon, S. J., Desai, A., Bowerman, B. and Oegema, K. (2008). Inhibition of Rac by the GAP Activity of Centralspindlin Is Essential for Cytokinesis. Science 322, 1543-1546. Carlier, M. F., Laurent, V., Santolini, J., Melki, R., Didry, D., Xia, G. X., Hong, Y., Chua, N. H. and Pantaloni, D. (1997). Actin depolymerizing factor (ADF/cofilin) enhances the rate of filament turnover: Implication in actin-based motility. Journal of Cell Biology 136, 1307-1322. Collavin, L., Buzzai, M., Saccone, S., Bernard, L., Federico, C., DellaValle, G., Brancolini, C. and Schneider, C. (1998). cDNA characterization and chromosome mapping of the human GAS2 gene. Genomics 48, 265-9. Conrad, P. A., Giuliano, K. A., Fisher, G., Collins, K., Matsudaira, P. T. and Taylor, D. L. (1993). Relative distribution of actin, myosin I, and myosin II during the wound healing response of fibroblasts. The Journal of cell biology 120, 1381- 91.

29 Cook, T. A., Nagasaki, T. and Gundersen, G. G. (1998). Rho guanosine triphosphatase mediates the selective stabilization of microtubules induced by lysophosphatidic acid. The Journal of cell biology 141, 175-85. Cotter, T. G., Lennon, S. V., Glynn, J. M. and Green, D. R. (1992). Microfilament-disrupting agents prevent the formation of apoptotic bodies in tumor cells undergoing apoptosis. Cancer Research 52, 997-1005. Cowled, P. A., Ciccarelli, C., Coccia, E., Philipson, L. and Sorrentino, V. (1994). Expression of growth arrest-specific (gas) genes in senescent murine cells. Experimental cell research 211, 197-202. Dent, E. W., Merriam, E. B. and Hu, X. (2011). The dynamic cytoskeleton: backbone of dendritic spine plasticity. Current opinion in neurobiology 21, 175-81. Discher, D. E., Mooney, D. J. and Zandstra, P. W. (2009). Growth factors, matrices, and forces combine and control stem cells. Science 324, 1673-7. Drubin, D. G. and Nelson, W. J. (1996). Origins of cell polarity. Cell 84, 335-44. Foe, V. E., Field, C. M. and Odell, G. M. (2000). Microtubules and mitotic cycle phase modulate spatiotemporal distributions of F-actin and myosin II in Drosophila syncytial blastoderm embryos. Development 127, 1767-87. Foe, V. E. and von Dassow, G. (2008). Stable and dynamic microtubules coordinately shape the myosin activation zone during cytokinetic furrow formation. The Journal of cell biology 183, 457-70. Fujiwara, I., Takahashi, S., Tadakuma, H., Funatsu, T. and Ishiwata, S. (2002). Microscopic analysis of polymerization dynamics with individual actin filaments. Nature cell biology 4, 666-73. Glotzer, M. (2009). The 3Ms of central spindle assembly: microtubules, motors and MAPs. Nature reviews. Molecular cell biology 10, 9-20. Godin, L. M., Vergen, J., Prakash, Y. S., Pagano, R. E. and Hubmayr, R. D. (2011). Spatiotemporal dynamics of actin remodeling and endomembrane trafficking in alveolar epithelial type I cell wound healing. American Journal of Physiology-Lung Cellular and Molecular Physiology 300, L615-L623.

30 Goldman, R. D., Grin, B., Mendez, M. G. and Kuczmarski, E. R. (2008). Intermediate filaments: versatile building blocks of cell structure. Current opinion in cell biology 20, 28-34. Gonzalez, C. (2003). Dispatch. Cell division: the place and time of cytokinesis. Current biology : CB 13, R363-5. Goriounov, D., Leung, C. L. and Liem, R. K. (2003). Protein products of human Gas2-related genes on chromosomes 17 and 22 (hGAR17 and hGAR22) associate with both microfilaments and microtubules. Journal of cell science 116, 1045-58. Green, K. J., Bohringer, M., Gocken, T. and Jones, J. C. (2005). Intermediate filament associated proteins. Advances in protein chemistry 70, 143-202. Halliburton, W. D. (1887). On Muscle-Plasma. The Journal of physiology 8, 133-202. Hatano, S. and Oosawa, F. (1966). Isolation and characterization of plasmodium actin. Biochimica et biophysica acta 127, 488-98. Heng, Y. W. and Koh, C. G. (2010). Actin cytoskeleton dynamics and the cell division cycle. The international journal of biochemistry & cell biology 42, 1622-33. Herman, I. M. (1993). Actin isoforms. Current opinion in cell biology 5, 48-55. Jaffe, A. B. and Hall, A. (2005). Rho GTPases: Biochemistry and biology. Annual review of cell and developmental biology 21, 247-269. Jaiswal, J. K., Andrews, N. W. and Simon, S. M. (2002). Membrane proximal lysosomes are the major vesicles responsible for calcium-dependent exocytosis in nonsecretory cells. The Journal of cell biology 159, 625-35. Johnson, D. I. (1999). Cdc42: An essential Rho-type GTPase controlling eukaryotic cell polarity. Microbiology and molecular biology reviews : MMBR 63, 54- 105. Khaitlina, S. Y. (2001). Functional specificity of actin isoforms. International review of cytology 202, 35-98. Kondo, Y., Shen, L., Cheng, A. S., Ahmed, S., Boumber, Y., Charo, C., Yamochi, T., Urano, T., Furukawa, K., Kwabi-Addo, B. et al. (2008). Gene silencing in cancer by histone H3 lysine 27 trimethylation independent of promoter DNA methylation. Nature genetics 40, 741-50.

31 Ku, N. O. and Omary, M. B. (2000). Keratins turn over by ubiquitination in a phosphorylation-modulated fashion. The Journal of cell biology 149, 547-52. Lee, K. K., Chan, W. Y. and Sze, L. Y. (1993). Histogenetic potential of rat hind-limb interdigital tissues prior to and during the onset of programmed cell death. The Anatomical record 236, 568-72. Lee, K. K. H., Tang, M. K., Yew, D. T. W., Chow, P. H., Yee, S. P., Schneider, C. and Brancolini, C. (1999). gas2 is a multifunctional gene involved in the regulation of apoptosis and chondrogenesis in the developing mouse limb. Developmental biology 207, 14-25. Lee, M. P. and Feinberg, A. P. (1997). Aberrant splicing but not mutations of TSG101 in human breast cancer. Cancer Research 57, 3131-3134. Lewinson, D. and Silbermann, M. (1992). Chondroclasts and endothelial cells collaborate in the process of cartilage resorption. The Anatomical record 233, 504- 14. Li, L., Li, X., Francke, U. and Cohen, S. N. (1997). The TSG101 tumor susceptibility gene is located in chromosome 11 band p15 and is mutated in human breast cancer. Cell 88, 143-154. Maciver, S. K. and Hussey, P. J. (2002). The ADF/cofilin family: actin- remodeling proteins. Genome Biology 3. Mandato, C. A. and Bement, W. M. (2001). Contraction and polymerization cooperate to assemble and close actomyosin rings around Xenopus oocyte wounds. The Journal of cell biology 154, 785-97. Mandato, C. A. and Bement, W. M. (2003). Actomyosin transports microtubules and microtubules control actomyosin recruitment during Xenopus oocyte wound healing. Current biology : CB 13, 1096-105. Mandelkow, E. and Mandelkow, E. M. (1989). Microtubular structure and tubulin polymerization. Current opinion in cell biology 1, 5-9. Mandelkow, E. and Mandelkow, E. M. (1990). Microtubular structure and tubulin polymerization. Current opinion in cell biology 2, 3-9.

32 Manzow, S., Brancolini, C., Marks, F. and Richter, K. H. (1996). Expression of growth arrest-specific (gas) genes in murine keratinocytes: Gas2 is specifically regulated. Experimental cell research 224, 200-203. Marceau, N., Schutte, B., Gilbert, S., Loranger, A., Henfling, M. E. R., Broers, J. L. V., Mathew, J. and Ramaekers, F. C. S. (2007). Dual roles of intermediate filaments in apoptosis. Experimental cell research 313, 2265-2281. Mazzoni, I., Zhang, T., Goetz, J., Wagner, J., Taheri, M., Munger, C., Sacher, M., Gilchrist, A., Au, C., Thibaul, G. et al. (2005). Involvement of Growth-arrest-specific protein 2, a smooth endoplasmic reticulum protein identified by proteomics, in cell cycle control. Canadian Proteomics Initiative Conference. McGough, A., Pope, B., Chiu, W. and Weeds, A. (1997). Cofilin changes the twist of F-actin: Implications for actin filament dynamics and cellular function. Journal of Cell Biology 138, 771-781. McNeil, A. K., Rescher, U., Gerke, V. and McNeil, P. L. (2006). Requirement for annexin A1 in plasma membrane repair. The Journal of biological chemistry 281, 35202-7. McNeil, P. L. and Steinhardt, R. A. (2003). Plasma membrane disruption: repair, prevention, adaptation. Annual review of cell and developmental biology 19, 697- 731. McNeil, P. L., Vogel, S. S., Miyake, K. and Terasaki, M. (2000). Patching plasma membrane disruptions with cytoplasmic membrane. Journal of cell science 113 ( Pt 11), 1891-902. Miller, A. L., von Dassow, G. and Bement, W. M. (2008). Control of the cytokinetic apparatus by flux of the Rho GTPases. Biochemical Society transactions 36, 378-80. Mitchison, T. and Kirschner, M. (1984). Dynamic instability of microtubule growth. Nature 312, 237-42. Mitelman, F., Mertens, F. and Johansson, B. (1997). A breakpoint map of recurrent chromosomal rearrangements in human neoplasia. Nature genetics 15 Spec No, 417-74.

33 Nobes, C. D. and Hall, A. (1995). Rho, rac, and cdc42 GTPases regulate the assembly of multimolecular focal complexes associated with actin stress fibers, lamellipodia, and filopodia. Cell 81, 53-62. Oakley, C. E. and Oakley, B. R. (1989). Identification of gamma-tubulin, a new member of the tubulin superfamily encoded by mipA gene of Aspergillus nidulans. Nature 338, 662-4. Oegema, K., Savoian, M. S., Mitchison, T. J. and Field, C. M. (2000). Functional analysis of a human homologue of the Drosophila actin binding protein anillin suggests a role in cytokinesis. The Journal of cell biology 150, 539-52. Olson, M. F., Ashworth, A. and Hall, A. (1995). An Essential Role for Rho, Rac, and Cdc42 Gtpases in Cell-Cycle Progression through G(1). Science 269, 1270-1272. Petrak, J., Ivanek, R., Toman, O., Cmejla, R., Cmejlova, J., Vyoral, D., Zivny, J. and Vulpe, C. D. (2008). Deja vu in proteomics. A hit parade of repeatedly identified differentially expressed proteins. Proteomics 8, 1744-9. Petroulakis, E., Parsyan, A., Dowling, R. J., LeBacquer, O., Martineau, Y., Bidinosti, M., Larsson, O., Alain, T., Rong, L., Mamane, Y. et al. (2009). p53-dependent translational control of senescence and transformation via 4E-BPs. Cancer cell 16, 439-46. Pines, M. K., Housden, B. E., Bernard, F., Bray, S. J. and Röper, K. (2010). The cytolinker Pigs is a direct target and a negative regulator of Notch signalling. Development 137, 913-22. Qiu, R. G., Chen, J., Kirn, D., McCormick, F. and Symons, M. (1995). An essential role for Rac in Ras transformation. Nature 374, 457-9. Raftopoulou, M. and Hall, A. (2004). Cell migration: Rho GTPases lead the way. Developmental biology 265, 23-32. Rappaport, R. (1971). Cytokinesis in animal cells. International review of cytology 31, 169-213. Reddy, A., Caler, E. V. and Andrews, N. W. (2001). Plasma membrane repair is mediated by Ca2+-regulated exocytosis of lysosomes. Cell 106, 157-169.

34 Ridley, A. J. (1994). Membrane Ruffling and Signal-Transduction. Bioessays 16, 321-327. Ridley, A. J. (2006). Rho GTPases and actin dynamics in membrane protrusions and vesicle trafficking. Trends in cell biology 16, 522-9. Ridley, A. J. and Hall, A. (1992). The small GTP-binding protein rho regulates the assembly of focal adhesions and actin stress fibers in response to growth factors. Cell 70, 389-99. Ridley, A. J., Paterson, H. F., Johnston, C. L., Diekmann, D. and Hall, A. (1992). The small GTP-binding protein rac regulates growth factor-induced membrane ruffling. Cell 70, 401-10. Rodriguez, O. C., Schaefer, A. W., Mandato, C. A., Forscher, P., Bement, W. M. and Waterman-Storer, C. M. (2003). Conserved microtubule- actin interactions in cell movement and morphogenesis. Nature cell biology 5, 599-609. Schapire, A. L., Valpuesta, V. and Botella, M. A. (2009). Plasma membrane repair in plants. Trends in plant science 14, 645-52. Schneider, C., King, R. M. and Philipson, L. (1988). Genes specifically expressed at growth arrest of mammalian cells. Cell 54, 787-93. Scholey, J. M., Brust-Mascher, I. and Mogilner, A. (2003). Cell division. Nature 422, 746-52. Sgorbissa, A., Benetti, R., Marzinotto, S., Schneider, C. and Brancolini, C. (1999). Caspase-3 and caspase-7 but not caspase-6 cleave Gas2 in vitro: implications for microfilament reorganization during apoptosis. Journal of cell science 112 ( Pt 23), 4475-82. Sider, J. R., Mandato, C. A., Weber, K. L., Zandy, A. J., Beach, D., Finst, R. J., Skoble, J. and Bement, W. M. (1999). Direct observation of microtubule-f-actin interaction in cell free lysates. Journal of cell science 112 ( Pt 12), 1947-56. Sisson, J. C., Field, C., Ventura, R., Royou, A. and Sullivan, W. (2000). Lava lamp, a novel peripheral Golgi protein, is required for Drosophila melanogaster cellularization. Journal of Cell Biology 151, 905-917.

35 Sonnemann, K. J. and Bement, W. M. (2011). Wound repair: toward understanding and integration of single-cell and multicellular wound responses. Annual review of cell and developmental biology 27, 237-63. Steiner, P., Barnes, D. M., Harris, W. H. and Weinberg, R. A. (1997). Absence of rearrangements in the tumour susceptibility gene TSG101 in human breast cancer. Nature genetics 16, 332-333. Stossel, T. P. (1993). On the crawling of animal cells. Science 260, 1086-94. Straight, A. F. and Field, C. M. (2000). Microtubules, membranes and cytokinesis. Current Biology 10, R760-R770. Stroud, M. J., Kammerer, R. A. and Ballestrem, C. (2011). Characterization of G2L3 (GAS2-like 3), a New Microtubule- and Actin-binding Protein Related to Spectraplakins. The Journal of biological chemistry 286, 24987-95. Su, L. K., Burrell, M., Hill, D. E., Gyuris, J., Brent, R., Wiltshire, R., Trent, J., Vogelstein, B. and Kinzler, K. W. (1995). APC binds to the novel protein EB1. Cancer Research 55, 2972-7. Sun, D., Leung, C. L. and Liem, R. K. (2001). Characterization of the microtubule binding domain of microtubule actin crosslinking factor (MACF): identification of a novel group of microtubule associated proteins. Journal of cell science 114, 161-172. Surka, M. C., Tsang, C. W. and Trimble, W. S. (2002). The mammalian septin MSF localizes with microtubules and is required for completion of cytokinesis. Molecular biology of the cell 13, 3532-3545. Szeverenyi, I., Cassidy, A. J., Chung, C. W., Lee, B. T., Common, J. E., Ogg, S. C., Chen, H., Sim, S. Y., Goh, W. L., Ng, K. W. et al. (2008). The Human Intermediate Filament Database: comprehensive information on a gene family involved in many human diseases. Human mutation 29, 351-60. Tirnauer, J. S., Grego, S., Salmon, E. D. and Mitchison, T. J. (2002). EB1-microtubule interactions in Xenopus egg extracts: role of EB1 in microtubule stabilization and mechanisms of targeting to microtubules. Molecular biology of the cell 13, 3614-26.

36 Togo, T. (2006). Disruption of the plasma membrane stimulates rearrangement of microtubules and lipid traffic toward the wound site. Journal of cell science 119, 2780-2786. Vitre, B., Coquelle, F. M., Heichette, C., Garnier, C., Chretien, D. and Arnal, I. (2008). EB1 regulates microtubule dynamics and tubulin sheet closure in vitro. Nature cell biology 10, 415-21. Waterman-Storer, C., Duey, D. Y., Weber, K. L., Keech, J., Cheney, R. E., Salmon, E. D. and Bement, W. M. (2000). Microtubules remodel actomyosin networks in Xenopus egg extracts via two mechanisms of F-actin transport. The Journal of cell biology 150, 361-76. Wegner, A. (1976). Head to Tail Polymerization of Actin. Journal of Molecular Biology 108, 139-150. Wolter, P., Schmitt, K., Fackler, M., Kremling, H., Probst, L., Hauser, S., Gruss, O. J. and Gaubatz, S. (2012). GAS2L3, a target gene of the DREAM complex, is required for proper cytokinesis and genomic stability. Journal of cell science 125, 2393-406. Yang, C., Czech, L., Gerboth, S., Kojima, S. I., Scita, G. and Svitkina, T. (2007). Novel roles of formin mDia2 in lamellipodia and filopodia formation in motile cells. Plos Biology 5, 2624-2645. Zhang, T., Dayanandan, B., Rouiller, I., Lawrence, E. J. and Mandato, C. A. (2011). Growth-arrest-specific protein 2 inhibits cell division in Xenopus embryos. PloS one 6, e24698. Zhang, T. and Mandato, C. A. (2007). Xenopus oocyte wound healing as a model system for analysis of microtubule-actin interactions. In Methods in Molecular Medicine: Microtubule Protocols, vol. 137, pp. 181-188: Humana Press. Zhong, B., Zhou, Q., Toivola, D. M., Tao, G. Z., Resurreccion, E. Z. and Omary, M. B. (2004). Organ-specific stress induces mouse pancreatic keratin overexpression in association with NF-kappaB activation. Journal of cell science 117, 1709-19.

37 Zucman-Rossi, J., Legoix, P. and Thomas, G. (1996). Identification of new members of the Gas2 and Ras families in the 22q12 chromosome region. Genomics 38, 247-54.

38 Chapter 1-Figure 1. Gas2 protein and its domains. The mouse Gas2 protein [P11862] and its domains information. The domains colors are matched with relative colored amino acid sequences. Gas2 protein contains an actin-binding calponin homology (CH) domain near its N- terminus (amino acids#: 36-158), and a MT-binding Gas2 domain near its C-terminus (amino acids#: 201-274). There are 2 low complexity (LC) domains (amino acids#: 167- 178 and 181-198) between the CH and Gas2 domains. The second LC domain contains 4 proline-serine (P-S) repeats (amino acids#: 183-190 shown in the boxed region in its sequence), which gives this region more structural flexibility (Zhang et al., 2011).

39 Figure 1

40 Chapter 1-Figure 2. Gas2 and G2L proteins comparison. Both of G2L1 and G2L2 have 2 isoforms and they are named as G2L1/ and G2L2/. The  isoform has larger size comparing with  isoform by an extra long C-terminus amino acid sequence, which is also the main difference between Gas2 and G2L proteins. This C-terminus tail may form another structure and directly impact on the G2L proteins cytoskeleton binding properties (Zhang et al., 2011).

41 Figure 2

42 Research rationale and objectives

Cytoskeleton and its cross-linking proteins play essential roles in cell division. Many cytoskeleton cross-linking proteins have been discovered; however, their functions and mechanism are not well understood. Gas2 protein is a potential actin-MT cross-linking protein and its expression levels change with cell cycle. We have followed the hypothesis that Gas2 protein plays a brake role during cell division and investigated Gas2 functions in cell division and its cytoskeleton binding mechanisms. To determine whether the Gas2 protein plays a brake role in cell division, in Chapter 2, we found that both Gas2 protein and its Gas2 domain, but not the CH domain, arrested Xenopus embryo cell division and also resulted in multinucleated cells. The observation that Gas2 domain alone can arrest cell division suggests that Gas2 function is mediated by MT binding. Followed this observation, we propose that Gas2 function is mediated by binding MTs, leading to cell division arrest. In Chapter 3, we showed the Gas2 protein cross-linked F-actin and MTs via its CH and Gas2 domains in HeLa cells. The Gas2 domain itself significantly changed MTs morphology into long loop-like bundled network. The phosphomimetic point mutations revealed that the phosphorylation regulated Gas2 protein cytoskeleton binding properties. Our experiments also showed that the Gas2 protein formed a dynamic complex and dynamically bound to MTs via its Gas2 domain. Taken together, our data suggest that the non-phosphorylated Gas2 protein forms a dynamic complex to bundle MTs, but its ability to cross-link F-actin and MTs is achieved by the phosphorylated Gas2 protein.

43 Chapter 2

2. GROWTH-ARREST-SPECIFIC PROTEIN 2 INHIBITS CELL DIVISION IN XENOPUS EMBRYOS

This chapter has been published in PLoS ONE. 2011 Sep. 6 (9): e24698.

44 Title: Growth-arrest-specific protein 2 inhibits cell division in Xenopus embryos

Authors and Affiliations: Tong Zhang1, Bama Dayanandan2, Isabelle Rouiller2, Elizabeth J. Lawrence2 and Craig A. Mandato1,2

1 Department of Biology, McGill University, Montreal, Quebec, Canada, 2 Department of Anatomy and Cell Biology, McGill University, Montreal, Quebec, Canada

Correspondence to Tong Zhang (e-mail: [email protected]) or Craig A. Mandato (e-mail: [email protected]).

Key words: Gas2, microtubules, actin-microtubule cross-linking, Xenopus, cell division

Abbreviations: CH, calponin homology; EM, electron microscopy; FL, full-length; Gas, growth-arrest-specific; MT(s), microtubule(s); P-S, proline-serine;

45 2.1 Abstract

2.1.1 Background: Growth-arrest-specific (Gas) 2 gene was originally identified in murine fibroblasts under growth arrest conditions. Furthermore, serum stimulation of quiescent, non-dividing cells leads to the down-regulation of gas2 and results in re-entry into the cell cycle. Cytoskeleton rearrangements are critical for cell cycle progression and cell division and the Gas2 protein has been shown to co-localize with actin and microtubules (MTs) in interphase mammalian cells. Despite these findings, direct evidence supporting a role for Gas2 in the mechanism of cell division has not been reported.

2.1.2 Methodology and Principal Findings: To determine whether the Gas2 protein plays a role in cell division, we over-expressed the full-length (FL) Gas2 protein and Gas2 truncations containing either the actin-binding CH domain or the tubulin- binding Gas2 domain in Xenopus laevis embryos. We found that both the FL-Gas2 protein and the Gas2 domain, but not the CH domain, inhibited cell division and resulted in multinucleated cells. The observation that Gas2 domain alone can arrest cell division suggests that Gas2 function is mediated by MT binding. Gas2 co-localized with MTs at the cell cortex of Gas2-injected Xenopus embryos using cryo-confocal microscopy and co-sedimented with MTs in cytoskeleton co-sedimentation assays. To investigate the mechanism of Gas2-induced cell division arrest, we showed, using a wound-induced contractile array assay, that Gas2 stabilized MTs. Finally, electron microscopy studies demonstrated that Gas2 bundled MTs into higher-order structures.

2.1.3 Conclusion and Significance: Our experiments show that Gas2 inhibits cell division in Xenopus embryos. We propose that Gas2 function is mediated by binding and bundling MTs, leading to cell division arrest.

46 2.2 Introduction

Cytoskeleton dynamics are essential for many fundamental cellular processes, including cell division, wound healing and cell motility (Grinnell and Petroll, 2010; Heng and Koh, 2010; Mandato and Bement, 2003). During cell division, for example, dramatic rearrangements of the actin and MT are required in order for the cell to change morphology, segregate its chromosomes and execute cytokinesis. The ability of the cytoskeleton to adapt to constant physiological changes is mediated, in part, by actin- and MT-binding proteins and cross-linking proteins that regulate cytoskeleton dynamics. Many actin-MT cross-linking proteins have been identified; however, their functions and mechanisms of regulation remain unclear (Rodriguez et al., 2003). One such potential cytoskeleton-interacting protein is the Gas2 protein.

The Gas2 protein belongs to the growth-arrest-specific protein family and is widely expressed in human tissues (Collavin et al., 1998). Although Gas2 has a putative N- terminal actin-binding calponin homology (CH) domain (Brancolini et al., 1992) and a C- terminal tubulin-binding Gas2 domain, no direct evidence for Gas2-cytoskeleton interactions has been reported. However, immunofluorescence studies demonstrated that the FL-Gas2 co-localizes with filamentous actin (F-actin) at the cell cortex and in stress fibers in growth-arrested NIH 3T3 fibroblasts (Brancolini et al., 1992) and the Gas2 domain co-localizes with MTs in COS-7 cells (Sun et al., 2001).

Although the majority of cells in an organism are quiescent, they are able to re-enter the cell cycle and proliferate after stimulation (Pardee et al., 1978). Several lines of evidences support a role for Gas2 in cell cycle progression. First, the gas2 gene was originally identified in a genetic screen of murine fibroblasts that were cultured under growth-arrest conditions (Schneider et al., 1988). Second, the gas2 gene is down-regulated upon serum and growth factor stimulation (Brancolini et al., 1992). Furthermore, the Gas2 protein is

phosphorylated on a serine residue at the G0 to G1 transition allowing quiescent cells to re-enter the cell cycle (Brancolini et al., 1992). However, whether Gas2 plays a direct

47 role in the mechanism of cell division and whether this function is mediated by its cytoskeleton binding properties are completely unknown.

In this study, Xenopus embryos and oocytes were used to study Gas2 functions in cell division. Xenopus embryo undergoes a time-regulated synchronized cell division in the early stages of its development and therefore is a useful in vivo model system for studying cell division. An established Xenopus oocyte wound-induced contractile array assay, which mimics cytokinesis, was used to study Gas2 interactions with the cytoskeleton in vivo (Bement et al., 1999). Furthermore, cytoskeleton co-sedimentation assays and electron microscopy were performed to study Gas2-cytoskeleton interactions in vitro. Our results suggest that the Gas2 protein plays a brake role in cell division and that its functions are mediated by binding and bundling MTs.

48 2.3 Materials and Methods

2.3.1 Experimental animals Xenopus laevis frogs were bought from Nasco (Fort Atkinson, WI, USA). The experiments were conducted by trained skilled personnel and were approved by McGill University Animal Care Committee (Protocol #: 4858 to C.A.M.).

2.3.2 Bioinformatics analysis Homo sapiens (human) [O43903], Canis familiaris (Dog) [F1PBV2], Mus musculus (house mouse) [P11862], Bos taurus (Bovine) [A8E4Q5], Gallus gallus (Chicken) [F1NSM4] and Xenopus tropicalis (Silurana tropicalis) [ENSXETP00000005555] Growth-arrest-specific protein 2 (Gas2) amino acid sequences were obtained from NCBI (http://www.ncbi.nlm.nih.gov), Uniprot (http://www.uniprot.org/) and Ensembl (http://www.ensembl.org/Xenopus_tropicalis) websites. They were compared the similarity with ClustalW (http://www.ebi.ac.uk/Tools/clustalw2/index.html). Phylogenetic tree was also derived by aligning the different species’ amino acids with ClustalW. Gas2 domain information was acquired from http://pfam.sanger.ac.uk/protein?acc=P11862.

2.3.3 Cloning DNA constructs were generated by PCR and cloned into pCS2-eGFP vector (a kind gift from Prof. Bill Bement, University of Wisconsin-Madison, USA). The FL-Gas2 (cDNA clone MGC:18565) was cloned with N-terminus GFP tag by using the forward primer 5’- ATGTGCACTGCCCTGA-3’ and reverse primer 5’- TCATTTAATCTCCTTCTTAGCCTTG-3’. Gas2 CH domain (with N-terminus GFP tag) was cloned with the forward primer 5’-TGCACTGCCCTGAGCC-3’ and reverse primer 5’-TCATCCAGTACTCTTCTTTCCTGAAG-3’; and its Gas2 domain (with N-terminus GFP tag) was cloned with the forward primer 5’-AGGTATGGTGTGGAGCCTCCT-3’ and reverse primer 5’-TCATTTAATCTCCTTCTTAGCCTTGT-3’. All clones were sequenced to ensure the sequence correctness.

49 2.3.4 Western blot GFP tagged Gas2 constructs were injected into Xenopus oocytes nuclei and oocytes were incubated for 48 hours to allow for relative protein expression. The oocytes were then sonicated in phosphate buffered saline (PBS: 137 mM NaCl, 2.7 mM KCl, 10 mM

Na2HPO4 and 2 mM KH2PO4 at pH 7.4) with ultrasound sonicator, and samples were centrifuged at 30,000 g for 10 minutes at 4°C. Supernatants were collected for Western blot analysis. Protein assays were performed using the BCA protein assay kit (Thermo). Protein samples were run on a 12% SDS-PAGE gel and transferred to 0.45 m Trans- Blot nitrocellulose paper (Bio-Rad), which was blocked with blotto buffer (5% non-fat milk in PBS) after transferring. After washing in 0.1% Tween-20 (Fisher) in PBS (PBS- T), monoclonal mouse GFP antibodies (Roche, Clone 7.1 and 13.1) were used at 1:3000 dilution in blotto buffer incubated at room temperature for 2 hours. The non-specific antibody binding was washed 3 times with PBS-T. Primary antibodies were detected using goat anti-mouse IgG antibodies conjugated to horse radish peroxidase (HRP) (Molecular Probes) at 1:3000 dilution in PBS at room temperature for 1 hour, and followed by 3 times washing with PBS-T. The signal detection was performed using the ECL method.

2.3.5 Protein expression and purification Gas2 (cDNA clone MGC:18565) was cloned into pTrcHis2B vector with 6X His tag at N-terminus using the forward primer 5’-ATGTGCACTGCCCTGA-3’ and reverse primer 5’-TTTAATCTCCTTCTTAGCCTTG-3’. Basically, transform pTrcHis2B-Gas2 construct into DH5 Escherichia coli to express FL-Gas2 protein. Protein was collected with Probond nickel-chelating resin (Invitrogen). The elution protein solution was kept in a dialysis bag (Spectra/Pov molecularporous membrane tubing, MWCO 12-14kDa, Fisher 08-667B) in PBS at 4°C overnight to remove imidazole. The protein sample was re-concentrated with Microcon centrifugal filter devices (Millipore: YM-10), and then was aliquoted into small volume and frozen in liquid nitrogen. Samples were stored at - 80°C for further use. The protein purity was tested by Coomassie-stained SDS-PAGE, and no significant additional bands were observed.

50 2.3.6 Xenopus laevis eggs collection and fertilization Human chorionic gonadotropin (hCG) was injected into 3 adult Xenopus females (1000 U/ml, 0.5 ml per frog) to induce ovulation. Eggs were harvested in glass Petri dish by gently squeezing the female. Eggs were fertilized by adding up a piece of a testis in the Petri dish and swirling gently. Ninety minutes later, 40 ml of 2% cysteine (Sigma) at pH 8.0 was added for 5 minutes to de-jelly embryos. Cysteine was poured of and replaced with Marc’s modified ringers (MMR: 100 mM NaCl, 2 mM KCl, 1 mM MgCl2, 2 mM

CaCl2 and 5 mM HEPES at pH 7.5). Embryos were incubated at 16°C. They develop more rapidly at room temperature.

2.3.7 Microinjection Microinjection needles (10 l Drummond microdispenser, Drummond Scientific Company) were pulled by Flaming/Brown micropipette puller (Model P-97, Sutter Instrument Company). Five nl of 12 g/l bovine serum albumin (BSA) in PBS solution or the same amount of purified FL-Gas2 protein in PBS solution was injected into one cell of 2-cell stage Xenopus embryos. Injections were performed using a Model PLI-100 Pico-Injector (Harvard Apparatus).

2.3.8 Live imaging movies The movies were recorded at a rate of 1 frame per minute with Zeiss Axioskop 2 microscope. Zeiss AxioCam MRm color camera and Plan-Neofluar 2.5X/0.075 N.A. objective lens were used. The images were modified with Volocity version 5.3.0 software (PerkinElmer) and played at 10 frames per second in Apple QuickTime (*.mov) format.

2.3.9 Microinjection experiments results statistics tabulation The post-injected embryos were examined for the arrested cell division phenotype under a stereo microscope. Statistics results were charted with Microsoft Office Excel software, and analysis was performed using Student’s t-test.

2.3.10 Cyro sample preparation and immunofluorescence

51 One hundred pg each of GFP, FL-GFP-Gas2, GFP-CH domain or GFP-Gas2 domain constructs were injected into one cell of 2-cell stage embryos cytoplasm. The injected embryos were incubated in MMR solution and developed until Stage 10 or 11 before cryo preparation. Embryos were fixed and embedded in 15%, then 25% fish gelatin (Norland). They were frozen on dry ice before cutting. Ten m thickness of cryo sections were cut in -20°C cyrotome (Leica), and collected on Superfrost plus slides (Fisher). A polyclonal rabbit Gas2 antibody (against peptide 78-99 KLH conjugation, provided by Prof. Jacque Paiement at University of Montreal) at 1:300 dilution, or a polyclonal rabbit GFP antibody (Molecular Probes) at 1:200 dilution, and a monoclonal alpha tubulin antibody (Sigma, DM1A) at 1:200 dilution were used. A goat-anti rabbit antibodies coupled to Alexa Fluor 488 (Molecular Probes) at 1:200 dilution, and a goat-anti mouse antibodies coupled to Alexa Fluor 546 (Molecular Probes) at 1:200 dilution were used for the staining. Nuclei were stained with DAPI (Molecular Probes) (0.5 g/ml) in PBS for 10 minutes. The yolk auto-fluorescence was quenched with 0.2% Eriochrome black (Sigma) in PBS for 10 minutes. Finally, the slides were mounted with ProLong Gold anti-fading reagent (Molecular Probes) and covered with cover slips (Fisher: 12-545M) for microscopy examination.

2.3.11 Xenopus laevis ooctyes collection and preparation Oocytes were surgically removed from an anesthetized adult Xenopus laevis female, as previously described (Zhang and Mandato, 2007) and stored at 16°C in oocyte ringer-2 solution (OR-2: 82.5 mM NaCl, 2.5 mM KCl, 1 mM CaCl2, 1 mM MgCl2, 1 mM

Na2HPO4, 5 mM HEPES at pH 7.4). Oocytes were treated with 0.2% collagenase (Type 2) OR-2 solution with gentle rotation at room temperature for 1 hour. Oocytes were washed after the treatment until the brown color collagenase was removed. The washed oocytes were kept at 16°C in OR-2 solution for at least 3 hours for recovery. Healthy oocytes have uniform animal and vegetal hemispheres. Stage VI oocytes were manually chosen and de-folliculated with Watchmaker’s forceps (Dumont #5) in a Petri dish containing OR-2 solution. This step is critical for cortex wound healing experiments.

2.3.12 Pharmacological perturbations and oocyte wounding

52 Oocytes were incubated in fresh 10 M taxol (Sigma) or 20 M nocodazole (Calbiochem) OR-2 solution for 1 hour at 16°C. The wounding experiments were performed in the same solutions. Oocytes were placed in a Petri dish fitted with a nylon grid to position the animal pole facing upwards. They were then wounded in the animal pole with a microinjection needle (cut to outside with a diameter of approximate 150 m). Cells were incubated for about 5 minutes in OR-2 solution for recovering before being placed in fix solution.

2.3.13 Xenopus laevis embryos and oocytes immunofluorescence The protocol was the same as previously described (Zhang and Mandato, 2007). Briefly, embryos or wounded oocytes were fixed and incubated with Alexa Fluor 568 phalloidin (6.6 M, at 1:200 dilution) (Molecular probes) and monoclonal alpha tubulin antibody (1:200 dilution) (Sigma, DM1A). Samples were then re-probed with Alexa Fluor 568 phalloidin for F-actin and Alexa Fluor 647 goat anti-mouse IgG antibody (1:200 dilution) (Molecular probes) for alpha tubulin. Finally, samples were carefully placed in the center of a high vacuum grease (Dow Corning Co.) mounted ring on a microscope slide plain (Fisher), and covered with a microscope cover glass #1.5 thickness (Fisher) for microscopy examination.

2.3.14 Confocal fluorescence microscopy All confocal images were collected using Zeiss 510 LSM Meta confocal microscope at Cell Imaging and Analysis Network at McGill University. The samples were examined with Plan-Neofluar 25X/0.8 N.A. Immersion Correction DIC or Plan-Apochromat 63X/1.4 N.A. oil DIC objective lens. DAPI (Molecular probes) staining was visualized with 405 nm blue diode laser and band pass (BP) 420-480 filter; GFP signal was visualized with 458 nm Ar ion laser and BP 505-530 filter; Alexa Fluor 568 was visualized with 543 nm HeNe green laser and BP 560-615 filter; and Alexa Fluor 647 was visualized with 633 nm HeNe red laser and long pass (LP) 650 filter. Z-sections of varying depths were stacked into projection images with maximum intensity to allow visualization of details.

53 2.3.15 Cytoskeleton co-sedimentation assays The GFP-FL-Gas2, GFP-CH domain and GFP-Gas2 domain constructs were injected into Xenopus oocytes and oocytes were incubated at 16°C for 48 hours to allow for relative proteins expression. The expression proteins were collected by sonication and following by centrifugation. Only the cytosolic layers were used for co-sedimentation assays. The samples were separately pre-treated with 10 M taxol (Sigma), 20 M nocodazole (Calbiochem); and 20 M phalloidin (Calbiochem), and 20 M latrunculin B (Calbiochem) to investigate Gas2 cytoskeleton binding properties. The mixture samples were incubated on ice for 1 hour to de-polymerize MTs, and then taxol was step wisely added to the final concentration 20 M to polymerize tubulin into MTs filaments at room temperature for one hour. Samples were loaded on the top layer of 30% sucrose in tubulin stabilization buffer (TSB: 1 mM EGTA, 5mM MgCl2, 80 mM K-PIPES at pH 7.0), and then they were centrifuged at 100,000g for 30 minutes at room temperature. The pellets were diluted into equal volume as the supernatant samples. Samples were run on 10% SDS-PAGE for Western blot analysis. The following antibodies were used for the Western blot analysis: the monoclonal mouse tubulin antibodies (Sigma, DM1A) were used at 1: 3000 dilution; the polyclonal rabbit actin antibodies (Biomedical Technologies Inc. BT-560) were used at 1: 3000 dilution; the monoclonal mouse GFP antibodies (Roche, Clone 7.1 and 13.1) were used at 1:3000 dilution. Primary antibodies were detected using IgG antibodies conjugated to HRP (Molecular Probes) at 1:3000 dilution by ECL method.

2.3.16 Electron microscopy Purified actin (AKL99-A) and tubulin (TL238-A) were bought from Cytoskeleton Inc. (Denver, CO.). Actin was first dissolved in the general actin buffer (5 mM Tris-HCl at pH 8.0, 0.2 mM CaCl2, 0.2 mM ATP), and then phalloidin (Calbiochem) was added up to final concentration 20 M to polymerize actin into F-actin. Tubulin was first dissolved into general tubulin buffer (80 mM PIPES at pH 6.9, 2 mM MgCl2, 0.5 mM EGTA, 1 mM GTP), and then 20 M final concentration taxol (Sigma) was used to polymerize tubulin into MTs. The purified FL-Gas2 protein in PBS was mixed with F-actin or MTs

54 or both for EM examination. A similar experiment was done by adding the FL-Gas2 to actin or tubulin samples before their polymerization. For each EM sample, 5 l protein sample was placed on a glow discharged pre-carbon coated copper grid for 1 minute, and then was replaced by 5 l 1% uranyl acetate for 1 minute treatment. Uranyl acetate was carefully removed and the treated grid was left by air-dried. Samples were examined in an FEI Tecnai 12 electron microscope, and digital images were taken by a Gatan 792 Bioscan wide angle multiscan CCD camera.

55 2.4 Results

2.4.1 Gas2 protein is conserved during evolution Bioinformatics comparisons of the Gas2 protein sequences among Homo sapiens (human) [O43903], Canis familiaris (Dog) [F1PBV2], Mus musculus (house mouse) [P11862], Bos taurus (Bovine) [A8E4Q5], Gallus gallus (Chicken) [F1NSM4] and Xenopus tropicalis (Silurana tropicalis) [ENSXETP00000005555] reveal that Gas2 is conserved during evolution (Fig. 1A and B), which suggests that Gas2 has a conserved biological function. The Gas2 protein contains two cytoskeleton binding domains: a putative actin-binding calponin homology (CH) domain near its N-terminus (Fig. 1C for mouse Gas2 [P11862] amino acids#: 36-158), and a tubulin-binding Gas2 domain near its C-terminus (amino acids#: 201-274). There are two low complexity domains (amino acids#: 167-178 and 181-198) between the CH and Gas2 domains, and the second low complexity domain contains 4 proline-serine (P-S) repeats (amino acids#: 183-190 shown in the boxed region in its sequence), which gives this region more structural flexibility. The N-terminus GFP-tagged mouse FL-Gas2 protein [P11862] and its CH and Gas2 domains were cloned to study their functions in Xenopus laevis (Fig. 1C), and the protein expression in Xenopus oocytes was verified by Western blot analysis (Fig. 1D).

2.4.2 Exogenous expression of the FL-Gas2 protein in Xenopus embryos inhibits cell division The Gas2 protein is up-regulated upon serum starvation in NIH 3T3 cells and it also results in cells arresting at G0 phase of cell cycle (Schneider et al., 1988). Conversely, the Gas2 protein level is down-regulated with serum and growth factor stimulation, allowing for cell cycle progression (Brancolini et al., 1992). These two observations imply a relationship between cell cycle progression and the Gas2 protein level. To test this hypothesis, Xenopus laevis embryos were used to investigate whether the over- expression of the Gas2 protein could inhibit cell division. The first cell division of Xenopus embryo occurs approximately 90 minutes post-fertilization at room temperature, and subsequent cell divisions occur at 30 minute intervals. Xenopus embryos are

56 relatively large, approximately 1 mm in diameter, facilitating the microinjection of the Gas2 protein and the observation of cell morphological changes.

The fertilized eggs were injected after they completed their first cell division. One cell of the 2-cell stage embryo was microinjected with either 25 ng bovine serum albumin (BSA) in phosphate buffered saline (PBS) solution or the bacterial purified mouse FL- Gas2 protein into the cytoplasm of a cell. The non-injected cell of the 2-cell embryo acts as an internal homo-genomic negative control (Fig. 2A). The BSA-injected embryos are shown in Fig. 2B-D (Video S1) and Gas2-injected embryos are shown in Fig. 2E-G’ (Video S2). BSA-injected embryos proceed through normal cell division and the number of cells increased 2-fold every 30 minutes. However, cells that were injected with 25 ng Gas2 protein divided once and then arrested in subsequent cell divisions while the non- injected control cells of the same embryo divided normally (Fig. 2H smaller cells on the left). In Gas2-injected cells, cell division arrested after approximately 30 minutes, consistent with the time required for Gas2 to diffuse in the cytoplasm of the injected cell (approximately 0.35 m/second under the injection needle pressure). Statistical analysis of the microinjection experiments showed that 7.2±4.2% BSA-injected embryos arrested in cell division, likely as a result of unsuccessful fertilization and/or microinjection damage to the embryos. In contrast, 79.2±5.5% Gas2-injected embryos arrested in cell division, which was a significantly higher percentage than the BSA control group (Fig. 2K). The non-100% arresting effect by Gas2 in cell division is thought to be caused by the Gas2 protein clotting in the injection needle. BSA-injected embryos succeed in developing into tadpoles, but Gas2-injected embryos all died within 24 hours, presumably due to the interference of the arrested cells with the embryo development. In 32-cell embryos, the cortical MTs were significantly longer in the Gas2-injected arrested cells than non-injected control cells (Fig. 2I). The two large Gas2 arrested cells were connected together with MTs (arrow in J). Two-cell stage Xenopus embryos were microinjected with increasing amounts of the Gas2 protein to investigate the minimum amount of Gas2 required to arrest their cell division. Dosage dependent analysis was done by microinjecting the same volume (5 nl) of a serial dilution of Gas2 protein, and

57 observing the cell division arresting phenotype. The calculated lethal dose 50% (LD50) was 5.5 ng or 31 M for one cell of the 2-cell stage Xenopus embryo (Fig. 2L).

2.4.3 Gas2 co-localizes with MTs in arrested cells and over-expression of either the FL-Gas2 protein or the Gas2 domain alone results in multinucleated cells To further investigate the mechanism of how Gas2 arrested Xenopus embryo cell division, cryo-confocal microscopy was performed with Gas2-injected embryos that were fixed and stained for Gas2, tubulin and DNA by DAPI (Fig. 3A-D). Gas2-injected cells that failed division were larger than non-injected control dividing cell. Gas2 localizes to the cortex in arrested cells and co-localizes with cortical MTs (arrow in A). The Gas2- injected cell has no DAPI staining presumably due to the fact that the injected cell arrested in the early embryo developmental stage when the cell is relatively large; therefore, it is difficult to maintain the integrity of all cellular organelles (such as the cell nucleus) during cryo-cutting (Fig. 3C shows no DAPI staining in the large arrested cell). The same observations were obtained with repeated experiments.

To examine the exogenous Gas2 during cell division, GFP control, GFP-FL-Gas2, GFP- CH domain and GFP-Gas2 domain constructs were injected into the cytoplasm of one cell of the 2-cell stage embryos, and embryos were fixed at either Stage 10 or 11. Cryo- confocal microscopy was performed to examine post-injected embryos. Embryos expressing the GFP control (Fig. 3E-H) and GFP-CH domain (Fig. 3I-L) are approximately the same size as neighboring non-expressing control cells, which is an indication of normal cell division. Some GFP and GFP-CH domain expressing cells can be recognized in anaphase of mitosis on the basis of their MT morphology (arrows in F and J) and their separating chromosomes (arrows in G and K). However, cells expressing both GFP-FL-Gas2 (Fig. 3M-P) and GFP-Gas2 domain (Fig. 3Q-T) are larger than control cells, and also have multiple nuclei, indicative of a failure in cell division (arrows in O and S). The GFP-FL-Gas2 (Fig. 3M) and GFP-Gas2 domain (Fig. 3Q) localize to the cell cortex and co-localize with MT spindles.

58 2.4.4 Gas2 stabilizes MTs via its Gas2 domain in Xenopus oocytes To test the hypothesis that Gas2 regulated MT dynamics/stability leading to cell division arrest, Xenopus laevis oocyte wound healing contractile array assay, which mimics cytokinesis, was used to study Gas2 interactions with the cytoskeleton in vivo (Bement et al., 1999). The relatively large size of Xenopus Stage VI oocytes (approximately 1 mm in diameter) provides a useful tool for studying cytoskeleton dynamics under the cell surface (Zhang and Mandato, 2007). The nuclei of Stage VI oocytes were injected with either GFP-FL-Gas2, GFP-CH domain or GFP-Gas2 domain constructs, and incubated for 48 hours to allow for protein expression. A micro-capillary glass tube pulled needle was used to wound the animal poles of the oocytes. After wounding, the oocytes were allowed to recover for approximately 5 minutes in oocyte ringer-2 (OR-2) solution, and were then fixed and stained for F-actin and tubulin (Fig. 4A). The non-injected control oocyte forms a single actin contractile ring around the wound; and MTs radially distribute around the wound (Fig. 4B-D). In oocytes pre-treated for one hour with 10 M taxol to stabilize MTs, actin assembles into two separate rings surrounding the wound (Fig. 4E-G). The internal ring is composed of the contractile F-actin, while the outer ring is formed by de novo actin polymerization (arrows in E) (Mandato and Bement, 2003). Please see the discussion for further details.

Oocytes expressing GFP-FL-Gas2 have a similar phenotype to taxol-treated oocytes. The GFP-Gas2 rings (arrows in 4H) co-localize with the two separate actin rings (arrows in 4I) during wound healing (Fig. 4H-K). This observation implies that Gas2 mimics taxol- treatment and therefore may stabilize MTs. When GFP-FL-Gas2-expressing oocytes were pre-treated with 20 M nocodazole to destabilize MTs for one hour prior to wounding, GFP-Gas2 co-localized with a single actin ring instead of double rings (Fig. 4L-O); therefore, MT stabilization by Gas2 was sensitive to nocodazole treatment. Gas2 has an actin-binding CH domain near its N-terminus, which may explain the observed co- localization of GFP-Gas2 and actin rings. Oocytes expressing GFP-CH domain have a single actin ring during wound healing (Fig. 4P-S). This result is also observed in the control (Fig. 4B-D) and nocodazole-treated oocytes expressing GFP-FL-Gas2 (Fig. 4L- O). GFP-CH domain co-localizes with the actin ring since it is the actin-binding domain

59 of Gas2. Oocytes expressing the GFP-Gas2 domain alone form double actin rings (arrows in 4U), and these double actin rings are similar to the taxol pre-treated (Fig. 4E) and GFP-FL-Gas2 expressing oocytes (Fig. 4I). GFP-Gas2 domain forms a ring-like structure and radially distributes around the wound (Fig. 4T), but it does not overlap with either actin ring. This series of experiments showed that the tubulin-binding Gas2 domain alone is sufficient for eliciting the double actin rings phenotype, suggesting that Gas2 binds MTs via its Gas2 domain and stabilizes them during oocyte wound healing.

2.4.5 Gas2 protein co-sediments with F-actin and MTs Cytoskeleton co-sedimentation assays were performed to investigate the cytoskeleton binding properties of Gas2 in vitro. GFP-FL-Gas2, GFP-CH domain and GFP-Gas2 domain constructs were injected into Xenopus oocytes and the expressed proteins were used for co-sedimentation assays (Fig. 5A). As expected, the CH domain co-sedimented with polymerized F-actin in the pellet (Fig. 5B). When F-actin was de-polymerized into actin monomers with latrunculin B, the CH domain was detected in the supernate (Fig.

5B: SL and PL). Similarly, the tubulin binding Gas2 domain co-sedimented with MTs in the pellet (Fig. 5C). When MTs were de-polymerized with nocodazole, more Gas2

domain was found in the supernate compared with taxol-treated samples (Fig. 5C: SN and

PN vs. ST and PT). FL-Gas2 co-sedimented with both F-actin and MTs (Fig. 5D). However, when F-actin was de-polymerized with latrunculin B, Gas2 remained in the supernate (Fig. 5D: SL and PL). It was surprising to note that Gas2 did not co-sediment with MTs in the pellet in the latrunculin B treated sample.

2.4.6 Gas2 protein bundles MTs in vitro Gas2 interactions with F-actin and MTs were studied at high resolution by electron microscopy (EM). The phalloidin-stabilized F-actin alone appears as non-organized long filaments (Fig. 6A). In the presence of the purified FL-Gas2 protein, F-actin remains as non-organized long filaments (Fig. 6B). Therefore, FL-Gas2 does not possess F-actin organizing properties in vitro. Similarly, the taxol-stabilized polymerized MTs appear as randomly non-organized long cables (Fig. 6C). However, in the presence of the purified FL-Gas2 protein, MTs appear as distinct, well-organized bundles (Fig. 6D). Therefore,

60 FL-Gas2 protein has MT-organizing ability. Since Gas2 protein has only one tubulin- binding domain, we believe Gas2 must form a protein complex in order to bundle MTs. When F-actin, MTs and FL-Gas2 are mixed together, Gas2 bundles MTs, but F-actin still appears as non-organized long filaments (Fig. 6E). Similar experimental results were obtained by mixing actin or tubulin with purified FL-Gas2 protein at the beginning of their polymerization.

61 2.5 Discussion

The data reported here identify Gas2 as a MT-bundling protein that inhibits cell division in Xenopus oocytes. Over-expression of FL-Gas2 protein arrested Xenopus embryo cell division and resulted in multinucleated cells, indicating a failure of cytokinesis. A similar phenotype was observed upon the over-expression of the tubulin-binding Gas2 domain alone, but not when the actin-binding CH domain was over-expressed. We propose, therefore, that Gas2 inhibits cell division via its C-terminal tubulin-binding Gas2 domain.

The cytoskeleton plays a key role in cell division and cytokinesis. To investigate the cytoskeleton-binding activity of Gas2, we used a cytoskeleton co-sedimentation assay and showed that the Gas2 protein indeed co-sedimented with both F-actin and MTs in vitro (Fig. 5D). Interestingly, when F-actin was de-polymerized, FL-Gas2 did not sediment with MT polymers in the pellet as would be expected considering the binding interaction between Gas2 and MTs (Fig. 5D: SL and PL). One explanation is the presence of a flexible linker region connecting the CH and Gas2 domains (Fig. 1C). This structural flexibility could permit masking of the Gas2 domain by the N-terminal region of the protein and would inhibit the MT-binding activity of FL-Gas2. Another possibility is that, due to the high affinity of Gas2 for F-actin, FL-Gas2 protein binds to small, incompletely de-polymerized F-actin filaments that are too light to be pelleted and thus remain in the supernate.

Dynamic MTs that are able to switch between polymerizing and de-polymerizing states are a necessary requirement for cell division (Kline-Smith and Walczak, 2004). Using electron microscopy, we have shown that Gas2 bundles purified MTs into higher-order structures in vitro, indicating a direct Gas2-MT interaction (Fig. 6D). The FL-Gas2 MT- bundling ability is preserved when F-actin is present in the same sample, but no F-actin- MT cross-linking has been observed in vitro (Fig. 6E). We hypothesize that Gas2-MT bundling ability inhibits MT dynamics and thereby causes cell division arrest and cytokinesis failure in Xenopus embryos.

62 To investigate the mechanism of Gas2-induced cell division arrest and cytokinesis failure, we used a wound-induced contractile array assay in Xenopus oocytes, which is also a model of cytokinesis (Bement et al., 1999). This assay permitted us to study Gas2- cytoskeleton interactions during MT-dependent actomyosin array formation and contraction in vivo (Bement et al., 1999). Oocyte wound healing and cytokinesis are highly dynamic processes that require coordination between the actin and MT cytoskeletons (Mandato and Bement, 2001). We have shown that Gas2 possesses both actin- and tubulin-binding properties in vitro; therefore, we propose that Gas2 functions as a structural cross-linking protein between these two cytoskeleton systems. When oocytes are treated with taxol to stabilize MTs prior to wounding, an abnormal double ring of actin forms around the wound border (Mandato and Bement, 2003). Staining with phalloidin for F-actin showed that the internal ring is composed of contractile F-actin, while staining for actin monomers showed that the outer ring is formed by de novo actin polymerization (Mandato and Bement, 2001; Mandato and Bement, 2003). Interestingly, we found that when FL-Gas2 or its Gas2 domain was over-expressed, oocytes developed the abnormal double actin ring phenotype upon wounding. Thus, Gas2 mimics the taxol- treatment phenotype, suggesting that Gas2 has MT-stabilizing activity. The Gas2-induced double rings can be rescued by de-polymerizing MTs with nocodazole (Fig. 4L-O) perhaps because the Gas2 protein binds dynamically to MTs. FL-Gas2 co-localizes with the actin rings, presumably by binding actin via its CH domain (Fig. 4H-K). It is interesting to note that the Gas2 domain alone also forms a ring structure and localizes between two actin rings during the wound healing (Fig. 4T-W), but it does not overlap with either actin ring since it has no actin-binding domain.

In summary, we have demonstrated that Gas2 can inhibit cell division in Xenopus embryos. We have also provided evidences that Gas2 has the ability to change MT dynamics and bundle MTs in vitro. We believe that the MT-bundling ability of Gas2 causes Xenopus embryo cell division arrest and the formation of the abnormal double actin rings in oocyte wound healing.

63 Drosophila contains a spectraplakin family protein, Short stop (Shot), that modulates MT dynamics via its Gas2 domain (Röper and Brown, 2004). Röper and Brown showed that Shot has a Gas2 domain at its C-terminus and is required for organizing MTs in a branched membrane structure called the fusome in meiotic cysts. They demonstrated that, in the absence of Shot, MTs did not assemble, and oocytes failed to become specified. The authors postulated that Shot elicits this effect via its Gas2 domain influencing MT dynamics. Similarly, Pines et al. showed that a Gas2-like protein, Pickled eggs (Pigs) co- localizes with both F-actin and MTs in the giant epithelial cells in Drosophila (Pines et al., 2010). Pigs is postulated to facilitate cytoskeleton rearrangements by binding and stabilizing F-actin and MTs. Consistent with our data, it is likely that Gas2 has conserved cytoskeleton-binding properties in different species. Although a recent paper by Stroud et al., showed that the GAR domain (the other name for Gas2 domain) of Gas2-like 3 protein is not essential for its localization to MTs and does not directly interact with MTs in vitro, this may be due to the large C-terminus domain which is present in Gas2-like 3 protein, but not in the Gas2 protein (Stroud et al., 2011). This domain may directly impact on protein functions including the ability of Gas2-like 3 to bind MTs.

Several studies have provided evidences suggesting that Gas2 activity is mis-regulated in cancer. Human gas2 is located on the short arm of Chromosome 11 and genetic rearrangements or deletions of this region are frequently found in tumors and sporadic human cancers (Lee and Feinberg, 1997; Li et al., 1997; Mitelman et al., 1997; Steiner et al., 1997). The analysis of oncogene transformed cells v-fos, v-myc, v-ras and v-src showed that the expression of Gas2 fails to increase in response to serum starvation (Brancolini et al., 1992). Furthermore, proteomic studies have identified Gas2 in normal rat liver cells, but not in the rat liver tumors (Mazzoni et al., 2005). Finally, translation initiation factor eIF4E binding proteins have been shown to control p53-dependent senescence by regulating Gas2 translation (Petroulakis et al., 2009). In light of this, a better understanding of the role of Gas2 functions in cell division and how Gas2 is regulated will aid the development of improved cancer therapies.

64 2.6 Acknowledgements

We would like to thank Prof. Jacque Paiement at University of Montreal for providing the Gas2 antibody; Prof. Marko Horb at Institut de recherches cliniques de Montreal for providing the experimental frogs; Dr. Wolfgang E. Reintsch for helping on the cyro technique; Cell Imaging and Analysis Network at McGill University for using the confocal microscope, and Facility for Electron Microscopy Research at McGill University for using the electron microscope.

65 2.7 References

Bement, W. M., Mandato, C. A. and Kirsch, M. N. (1999). Wound- induced assembly and closure of an actomyosin purse string in Xenopus oocytes. Current biology : CB 9, 579-587. Brancolini, C., Bottega, S. and Schneider, C. (1992). Gas2, a growth arrest-specific protein, is a component of the microfilament network system. The Journal of cell biology 117, 1251-61. Collavin, L., Buzzai, M., Saccone, S., Bernard, L., Federico, C., DellaValle, G., Brancolini, C. and Schneider, C. (1998). cDNA characterization and chromosome mapping of the human GAS2 gene. Genomics 48, 265-9. Conrad, P. A., Giuliano, K. A., Fisher, G., Collins, K., Matsudaira, P. T. and Taylor, D. L. (1993). Relative distribution of actin, myosin I, and myosin II during the wound healing response of fibroblasts. The Journal of cell biology 120, 1381- 91. Grinnell, F. and Petroll, W. M. (2010). Cell motility and mechanics in three-dimensional collagen matrices. Annual review of cell and developmental biology 26, 335-61. Heng, Y. W. and Koh, C. G. (2010). Actin cytoskeleton dynamics and the cell division cycle. The international journal of biochemistry & cell biology 42, 1622-33. Kline-Smith, S. L. and Walczak, C. E. (2004). Mitotic spindle assembly and chromosome segregation: refocusing on microtubule dynamics. Molecular cell 15, 317-27. Lee, M. P. and Feinberg, A. P. (1997). Aberrant splicing but not mutations of TSG101 in human breast cancer. Cancer Research 57, 3131-3134. Li, L., Li, X., Francke, U. and Cohen, S. N. (1997). The TSG101 tumor susceptibility gene is located in chromosome 11 band p15 and is mutated in human breast cancer. Cell 88, 143-154. Mandato, C. A. and Bement, W. M. (2001). Contraction and polymerization cooperate to assemble and close actomyosin rings around Xenopus oocyte wounds. The Journal of cell biology 154, 785-97.

66 Mandato, C. A. and Bement, W. M. (2003). Actomyosin transports microtubules and microtubules control actomyosin recruitment during Xenopus oocyte wound healing. Current biology : CB 13, 1096-105. Mazzoni, I., Zhang, T., Goetz, J., Wagner, J., Taheri, M., Munger, C., Sacher, M., Gilchrist, A., Au, C., Thibaul, G. et al. (2005). Involvement of Growth-arrest-specific protein 2, a smooth endoplasmic reticulum protein identified by proteomics, in cell cycle control. Canadian Proteomics Initiative Conference. Mitelman, F., Mertens, F. and Johansson, B. (1997). A breakpoint map of recurrent chromosomal rearrangements in human neoplasia. Nature genetics 15 Spec No, 417-74. Pardee, A. B., Dubrow, R., Hamlin, J. L. and Kletzien, R. F. (1978). Animal cell cycle. Annual review of biochemistry 47, 715-50. Petroulakis, E., Parsyan, A., Dowling, R. J., LeBacquer, O., Martineau, Y., Bidinosti, M., Larsson, O., Alain, T., Rong, L., Mamane, Y. et al. (2009). p53-dependent translational control of senescence and transformation via 4E-BPs. Cancer cell 16, 439-46. Pines, M. K., Housden, B. E., Bernard, F., Bray, S. J. and Röper, K. (2010). The cytolinker Pigs is a direct target and a negative regulator of Notch signalling. Development 137, 913-22. Rodriguez, O. C., Schaefer, A. W., Mandato, C. A., Forscher, P., Bement, W. M. and Waterman-Storer, C. M. (2003). Conserved microtubule- actin interactions in cell movement and morphogenesis. Nature cell biology 5, 599-609. Röper, K. and Brown, N. H. (2004). A spectraplakin is enriched on the fusome and organizes microtubules during oocyte specification in Drosophila. Current biology : CB 14, 99-110. Schneider, C., King, R. M. and Philipson, L. (1988). Genes specifically expressed at growth arrest of mammalian cells. Cell 54, 787-93. Steiner, P., Barnes, D. M., Harris, W. H. and Weinberg, R. A. (1997). Absence of rearrangements in the tumour susceptibility gene TSG101 in human breast cancer. Nature genetics 16, 332-333.

67 Stroud, M. J., Kammerer, R. A. and Ballestrem, C. (2011). Characterization of G2L3 (GAS2-like 3), a New Microtubule- and Actin-binding Protein Related to Spectraplakins. The Journal of biological chemistry 286, 24987-95. Sun, D., Leung, C. L. and Liem, R. K. (2001). Characterization of the microtubule binding domain of microtubule actin crosslinking factor (MACF): identification of a novel group of microtubule associated proteins. Journal of cell science 114, 161-172. Zhang, T. and Mandato, C. A. (2007). Xenopus oocyte wound healing as a model system for analysis of microtubule-actin interactions. In Methods in Molecular Medicine: Microtubule Protocols, vol. 137, pp. 181-188: Humana Press.

68 Chapter 2-Figure 1. Gas2 protein is conserved during evolution. (A) Phylogenetic tree of Gas2 protein. The phylogenetic relationship was derived by ClustalW program. The numbers represent the evolutionary distances. (B) Multiple sequences alignment of Gas2 amino acid sequences from different species. The alignment was generated using ClustalW. “*” indicates identical amino acids in all sequences in the alignment; “:” indicates that conserved substitutions have been observed; and “.” indicates that semi-conserved substitutions have been observed. (C) The mouse Gas2 protein [P11862] domain structure and amino acid sequence. N-terminal GFP-FL-Gas2, GFP-CH domain and GFP-Gas2 domain constructs were used in this study. The domains colors are matched with relative amino acid colored sequences. The boxed region indicates the 4 P-S repeats location, which gives this region more structural flexibility. (D) Western blot analysis of GFP Gas2 constructs expression. GFP=27 kDa, GFP-FL-Gas2=62 kDa, GFP-CH domain=49 kDa and GFP-Gas2 domain=44 kDa.

69 Figure 1

70 Chapter 2-Figure 2. Exogenous Gas2 inhibits cell division in Xenopus embryos. (A) The experimental flowchart and schematically represented results. The darker side of the egg represents the animal pole and the lighter bottom side represents the vegetal pole. The 90 degree angle rotation views show cell numbers in BSA-injected control vs. Gas2-injected embryos. (B-D) BSA-injected Xenopus embryos continue through normal cell divisions (Video S1). (E-G) Gas2-injected embryo cells divide once, then arrest in subsequent cell divisions. The non-injected control cell of the Gas2-injected embryo divides normally, and the cell number increases 2-fold every 30 minutes at room temperature (Video S2). The red circles in B and E indicate the needle injection sites. Time was set to 0 at 8-cell stage for demonstration purposes. (G’) The rotation view of Fig. 2G embryo shows one of the large arrested cells on the left. Bars, 0.3 mm. (H) Stereo microscopy examination of a Gas2-injected embryo at the 32-cell stage. The normal dividing cells are on the left and the two large arrested cells are on the right. Bar, 0.3 mm. (I and J) Confocal microscopy analysis of post-injected embryo cells with tubulin antibody staining. The white arrow in J indicates that the two large, arrested cells remain connected with MTs. Bars, 20 m. (K) Statistical analysis of cell division rate in BSA- and Gas2-injected embryos. Only 7.2±4.2% of BSA-injected embryos arrest in cell division; however, 79.2±5.5% of Gas2-injected embryos arrest in cell division (n=300 embryos and from 5 experiments, p<0.001). (L) Dosage dependent analysis of cell division rate in Gas2-

injected embryos. The calculated LD50 from the dosage dependent analysis graph is 5.5 ng or 31 M for one cell of the 2-cell stage Xenopus embryo. The x-axis of the graph is in logarithmic scale.

71 Figure 2

72 Chapter 2-Figure 3. Over-expression of either the FL-Gas2 protein or the Gas2 domain alone arrests cell division in Xenopus embryos. (A-D) Gas2 protein injected cell is relatively larger than non-injected normal dividing cells. Gas2 localizes to the injected cell cortex (arrow in A) and also co-localizes with MT network shown in yellow in the merged image D. Bar, 100 m. (E-H) Cells expressing GFP alone and (I-L) GFP-CH domain are similar in size to the neighboring non-expressing control cells. Cells in anaphase can be recognized by their MT morphology (arrows in F and J) and separating chromosomes (arrows in G and K). Bars in E-H, 10 m and bars in I-L, 20 m. (M-P) Cells expressing GFP-FL-Gas2 and (Q-T) GFP-Gas2 domain expressing cells are relatively larger than neighboring non-expressing cells and they also have multiple nuclei, indicating a failure in cell division (arrows in O and S). Bars, 20 m.

73 Figure 3

74 Chapter 2-Figure 4. The expression of either FL-Gas2 protein or the Gas2 domain alone results in abnormal double actin rings at the wound border. (A) The experimental flowchart and schematically represented results. Oocytes nuclei were injected with different GFP Gas2 constructs and incubated for 48 hours to allow for protein expression. The animal pole of an oocyte was then wounded, fixed and stained for F-actin and tubulin. The wound site was excised and examined by confocal microscopy. (B-D) Oocyte wound healing control experiment. F-actin forms a single ring and MTs radially distribute around the wound. Bars, 5 m. (E-G) Oocytes pre-treated with Taxol form abnormal double actin rings (arrows in E) during wound healing. Bars, 10 m. (H- K) Oocytes expressing the GFP-FL-Gas2 form double Gas2 rings (arrows in H), which co-localize with double actin rings (arrows in I) during wound healing. Bars, 20 m. (L- O) Oocytes expressing the GFP-FL-Gas2 and pre-treated with nocodazole form a single Gas2 ring, which co-localizes with single actin ring during wound healing. Bars, 10 m. (P-S) Oocytes expressing GFP-CH domain alone form a single actin ring during wound healing. Bar, 10 m. (T-W) Oocytes expressing GFP-Gas2 domain alone form a single Gas2 domain ring, which localizes between the double actin rings (arrows in U) during wound healing. The GFP-Gas2 domain does not co-localize with either actin ring. Bar, 10 m.

75 Figure 4

76 Chapter 2-Figure 5. Gas2 protein co-sediments with F-actin and MTs. (A) The experimental flowchart of the cytoskeleton co-sedimentation assays. The samples were pre-treated with either 10 M taxol, 20 M nocodazole, 20 M phalloidin, or 20 M latrunculin B to study the cytoskeleton binding properties of Gas2. The samples were incubated on ice for one hour to de-polymerize MTs. Taxol was added stepwise to a final concentration of 20 M to polymerize tubulin into MTs and samples were incubated at room temperature for one hour. Samples were loaded on the top layer of 30% sucrose solution in tubulin stabilization buffer (TSB), then centrifuged at 100,000g for 30 minutes at room temperature. The pellets were re-suspended into the same volume as the supernate. Samples were run on SDS-PAGE for Western blot analysis. (B) The GFP-CH domain co-sediments with F-actin in the pellet. When F-actin was de-polymerized into actin

monomers with latrunculin B, the CH domain was detected in the supernate (Fig. 5B: SL and PL). (C) The GFP-Gas2 domain co-sediments with MT in pellets. When MTs were de-polymerized with nocodazole, more Gas2 domain was found in the supernate

compared with taxol-treated samples (Fig. 5C: SN and PN vs. ST and PT). (D) The GFP- FL-Gas2 co-sediments with both F-actin and MTs. When F-actin is de-polymerized with

latrunculin B, the FL-Gas2 surprisingly remains in the supernate only. ST: Supernate of

Taxol treatment, PT: Pellet of Taxol treatment; SN: Supernate of Nocodazole treatment,

PN: Pellet of Nocodazole treatment; SP: Supernate of Phalloidin treatment, PP: Pellet of

Phalloidin treatment; SL: Supernate of Latrunculin B treatment, and PL: Pellet of Latrunculin B treatment.

77 Figure 5

78 Chapter 2-Figure 6. Gas2 protein bundles MTs in vitro. (A) F-actin alone appears as non-organized long filaments. (B) In the presence of Gas2, F-actin remains as non-organized long filaments as in A. Bars, 200 nm. (C) MTs alone appear as long, randomly distributed cables. (D) MTs bundle together when Gas2 is present in the sample. Bars, 0.2 m. (E) Gas2 bundles MTs, but does not organize F-actin when MTs, F-actin and Gas2 protein are mixed together. Bar, 200 nm. The order of Gas2 addition relative to the MTs or F- actin has no effect on the observed results.

79 Figure 6

80 Connecting statement

The observation that Gas2 domain alone can arrest cell division suggests that Gas2 function is mediated by MT binding and bundling. However, the mechanisms by which Gas2 binds to the cytoskeleton are still unclear. Xenopus embryo and oocyte are in rich of auto-fluorescent yolks, which are not ideal for live cell imaging of cytoskeleton dynamics. Therefore, we decided to continue investigating Gas2 cytoskeleton binding mechanisms using HeLa cells with both microscopy and in vitro experiments in Chapter 3.

81 Chapter 3

3. GROWTH-ARREST-SPECIFIC 2 (GAS2) PROTEIN BUNDLES MICROTUBULES AND FACILITATES MICROTUBULE POLYMERIZATION

This chapter has been written as a manuscript and it is ready to be submitted.

82 Title: Growth-arrest-specific 2 (Gas2) Protein Bundles Microtubules and Facilitates Microtubule Polymerization

Authors and Affiliations: Tong Zhang1, Bama Dayanandan2, Michael T. Greenwood3 and Craig A. Mandato1,2

1Department of Biology, McGill University, Montreal, Quebec, H3A 1B1, Canada 2Department of Anatomy and Cell Biology, McGill University, Montreal, Quebec, H3A 2B2, Canada, 3Department of Chemistry and Chemical Engineering, Royal Military College, Kingston, Ontario, K7K 7B4, Canada

Running title: Gas2 facilitates microtubule polymerization

Correspondence to Tong Zhang (e-mail: [email protected]) or Craig A. Mandato (e-mail: [email protected]).

Key words: Gas2; microtubules; actin-microtubule cross-linking; cell division; HeLa cells

Abbreviations: CH, calponin homology; FL, full-length; FRAP, fluorescence recovery after photobleaching; Gas, growth-arrest-specific; LC, low complexity; MT(s), microtubule(s); P-S, proline-serine;

83 3.1 Abstract

3.1.1 Background: Growth-arrest-specific (Gas) 2 protein was initially identified as a filamentous actin (F-actin) binding protein in murine fibroblasts under growth arrest condition; furthermore, the C-terminus of the Gas2 protein co-localized with microtubules (MTs). Gas2 contains putative actin and MT binding domains, which makes it a strong candidate to be an actin-MT cross- linking protein. However, the mechanisms by which Gas2 binds to the cytoskeleton are still unclear.

3.1.2 Methodology and Principal Findings: In this study, we showed the full-length (FL) Gas2 protein cross-linked F-actin and MTs via its actin-binding CH domain and MT-binding Gas2 domain, but the majority of the FL-Gas2 protein co-localized with F-actin when it was over-expressed in HeLa cells. The Gas2 domain itself significantly changed MTs morphology into long loop-like bundled network. Over-expression of the FL-Gas2, its CH or Gas2 domain resulted in a significant increase of multinucleated cells. The FL-Gas2 facilitated MT polymerization, which resembled parallel bundling MTs. In addition, it also changed the dynamic behaviors of the MTs. There were three predicted protein kinase C (PKC) phosphorylation sites in the FL-Gas2 protein, and phosphomimetic point mutations revealed that all three were involved in the FL- Gas2 protein function changes. Cytoskeleton co-sedimentation assays and fluorescence recovery after photobleaching (FRAP) experiments showed that the FL-Gas2 protein dynamically bound MTs via its Gas2 domain. Size-exclusive chromatography finally revealed that the FL-Gas2 protein formed a dynamic complex, which presumably had an important role in its cytoskeletal binding properties.

3.1.3 Conclusion and Significance: Taken together, our data suggest that the non-phosphorylated FL-Gas2 protein forms a dynamic complex to bundle MTs, but its ability to cross-link F-actin and MTs is achieved by the phosphorylated FL-Gas2 protein.

84 3.2 Introduction

The eukaryotic cytoskeleton, especially actin and MTs are not only very dynamic, but also involved in many fundamental cellular processes, such as cell division, vesicle trafficking, cell movement and wound healing (Fletcher and Mullins, 2010). Together with many cytoskeleton binding proteins, motor proteins and cross-linking proteins, actin and MTs form a complex network to play those important roles. The coordination of actin, MTs and their cross-linking proteins is crucial for these important biological processes (Rodriguez et al., 2003). Many actin-MT cross-linking proteins have been identified; however, their functions and mechanisms of action are still not well understood (Rodriguez et al., 2003). The Gas2 protein is suggested to be a potential actin-MT cross-linking protein, but its mechanisms involved in cytoskeleton binding are still unclear.

Gas2 protein belongs to the growth-arrest-specific protein family, which was originally identified in a genetic screen of murine fibroblasts under growth arrest conditions (Schneider et al., 1988). The FL-Gas2 contains a putative actin-binding CH domain at its N-terminus (Brancolini et al., 1992) and a MT-binding Gas2 domain (also called GAR domain) at its C-terminus. However, no evidence ever shows that FL-Gas2 directly cross-links actin and MTs. The early immunofluorescence studies demonstrated that FL-Gas2 protein co-localized with F-actin in growth-arrested NIH 3T3 fibroblasts (Brancolini et al., 1992). A separate study demonstrated that the Gas2 domain was observed co-localizing with MTs in COS-7 cells (Sun et al., 2001). However, whether FL-Gas2 protein can cross-link actin and MTs is still not known.

Gas2 protein is widely expressed in human tissues and Western blot analysis has shown that the Gas2 protein is expressed at highest level in liver, lung, kidney and brain (Collavin et al., 1998). Human Gas2 shares both structural and functional homology to the Gas2 protein from many other species, which implies that the Gas2 protein is conserved during evolution and presumably plays similarly important roles (Zhang et al., 2011). Human gas2 gene localizes to Chromosome

85 11p14.3-p15.2 (Collavin et al., 1998) and re-arrangements or deletion of short arm of human Chromosome 11 are frequently found in tumors and variety of sporadic human cancers (Lee and Feinberg, 1997; Li et al., 1997; Mitelman et al., 1997; Steiner et al., 1997). The analysis of cells transformed with oncogenes such as v- fos, v-myc, v-ras and v-src has showed that the Gas2 expression fails to increase in response to serum starvation in these cells (Brancolini et al., 1992). Proteomic studies identified the Gas2 protein in normal rat liver cells, but not in rat liver tumors (Mazzoni et al., 2005). The Gas2 protein expression level is also decreased in prostate cancer cells, which suggests that it may work as a prostate cancer suppressor protein (Kondo et al., 2008). Furthermore, the translation initiation factor eIF4E binding protein controls p53-dependent senescence at least in part by regulating Gas2 translation (Petroulakis et al., 2009). Together with its cytoskeleton binding properties, Gas2 protein may work as a cytoskeleton binding cancer suppressor protein.

In this study, we have used both in vivo models and in vitro methodologies to investigate the FL-Gas2 protein and its domain specific cytoskeleton binding mechanisms. Time-lapse microscopy was also used to observe the FL-Gas2 protein and its domains cytoskeleton binding behaviors and their effects on cell division. The phosphomimetic point mutations were done on the FL-Gas2 protein to study its biological functions. Our data suggest that the non-phosphorylated FL-Gas2 protein forms a dynamic complex to bundle MTs and facilitates MT polymerization, but cross-linking actin and MTs is achieved by phosphorylated FL-Gas2 protein.

86 3.3 Materials and Methods

3.3.1 Experimental use HeLa cell line The human HeLa cell line was purchased from ATCC (cat. #: CCL-2). Cells were grown in DMEM (Invitrogen) with 10% FBS (HyClone) media and kept in a

37°C incubator with 5% CO2.

3.3.2 Bioinformatics analysis Mus musculus (house mouse) [P11862] Gas2 amino acid sequence was obtained from NCBI (http://www.ncbi.nlm.nih.gov) website. Gas2 domain information was acquired from http://pfam.sanger.ac.uk/protein?acc=P11862. The Gas2 protein phosphorylation sites were predicted by NetPhosK (http://www.cbs.dtu.dk/services/NetPhosK/).

3.3.3 Cloning DNA constructs were generated by PCR and cloned into pCS2-eGFP vector (a kind gift from Prof. Bill Bement, University of Wisconsin-Madison, USA). The FL-Gas2 (cDNA clone MGC: 18565) was cloned with N-terminus GFP tag by using the forward primer 5’-atgtgcactgccctga-3’ and reverse primer 5’- tcatttaatctccttcttagccttg-3’. The CH domain (with N-terminus GFP tag) was cloned with the forward primer 5’-tgcactgccctgagcc-3’ and reverse primer 5’- tcatccagtactcttctttcctgaag-3’. The Gas2 domain (with N-terminus GFP tag) was cloned with the forward primer 5’-aggtatggtgtggagcctcct-3’ and reverse primer 5’- tcatttaatctccttcttagccttgt-3’. Point mutations were generated by PCR with Stratagene QuickChange II site-directed mutagenesis kit. S193D mutation was cloned with the forward primer 5’-cgccttctccttcatcaaaggattcaggaaagaagagtactg-3’ and reverse primer 5’-cagtactcttctttcctgaatcctttgatgaaggagaaggcg-3’. S194D mutation was cloned with the forward primer 5’- ccttcgccttctccttcatcaaagtctgatggaaagaagagtactg-3’ and reverse primer 5’- cagtactcttctttccatcagactttgatgaaggagaaggcgaagg-3’. T306D mutation was cloned with the forward primer 5’-acctggtggtctctgccgactacaaggctaagaagg-3’ and reverse primer 5’-ccttcttagccttgtagtcggcagagaccaccaggt-3’. S193D/S194D double point

87 mutation was generated with the S194D as the template and cloned with the forward primer 5’-cgccttctccttcatcaaaggatgatggaaagaagagtactg-3’ and reverse primer 5’-cagtactcttctttccatcatcctttgatgaaggagaaggcg-3’. S193A/S194A double point mutation was generated with the FL-Gas2 as the template and cloned with the forward primer 5’-gccttctccttcatcaaaggctgcaggaaagaagagtactgg-3’ and reverse primer 5’-ccagtactcttctttcctgcagcctttgatgaaggagaaggc-3’. S193D/S194D/T306D triple point mutation was generated with the T306D as the template and cloned with the forward primer 5’- tctccttcgccttctccttcatcaaaggatgatggaaagaagagtactggaaacttactg-3’ and reverse primer 5’-cagtaagtttccagtactcttctttccatcatcctttgatgaaggagaaggcgaaggaga-3’. All clones were sequenced to ensure the sequence correctness.

3.3.4 Immunofluorescence The transfected cells were grown on #1.5 thickness cover glass (Fisher) and incubated for 24 hours to allow the expression of the constructs. Cells were fixed in 4% paraformaldehyde (Fisher) phosphate buffered saline (PBS: 137 mM NaCl,

2.7 mM KCl, 10 mM Na2HPO4 and 2 mM KH2PO4 at pH 7.4) solution for 30 minutes at room temperature and followed by washing with PBS. The washed cells were blocked with 5% bovine serum albumin (BSA) (Fisher) and 0.2% Triton X-100 (Fisher) in PBS solution for 30 minutes at room temperature. The blocked cells were washed with 0.1% Triton X-100 PBS solution and then incubated with Alexa Fluor 568 phalloidin (6.6 M, at 1:200 dilution; Molecular probes) and monoclonal alpha tubulin antibody (1:200 dilution; Sigma, DM1A) in 0.5% BSA PBS solution for 2 hours at room temperature. Cells were then probed with Alexa Fluor 568 phalloidin (6.6 μM, at 1:200 dilution; Molecular probes) for F-actin and Alexa Fluor 647 goat anti-mouse IgG antibody (1:200 dilution; Molecular probes) for alpha tubulin for 1 hour at room temperature after washing with 0.1% Triton X-100 PBS solution. Finally, cells’ nuclei were stained with DAPI (1:5000 dilution; Molecular probes) for 2 minutes. The cover glass was carefully mounted on the microscope slide plain (Fisher) with GelTol aqueous mounting medium (Thermo) for immunofluorescence.

88

3.3.5 Confocal fluorescence microscopy All fixed cell confocal images were collected using a Zeiss 510 LSM Meta confocal microscope with Plan-Apochromat 63X/1.4 N.A. oil DIC objective lens and the microscope was operating with Zeiss AIM version 3.2 program. Cargille Type LDF immersion oil was used. DAPI (Molecular probes) staining was visualized with 405 nm blue diode laser and band pass (BP) 420-480 filter; GFP signal was visualized with 458 nm Ar ion laser and BP 505-530 filter; Alexa Fluor 568 was visualized with 543 nm HeNe green laser and BP 560-615 filter; and Alexa Fluor 647 was visualized with 633 nm HeNe red laser and long pass (LP) 650 filter. Z-sections of varying depths were stacked into projection images with maximum intensity to allow visualization of details. The final images were adjusted brightness and contrast in Adobe Photoshop.

3.3.6 Live cell time-lapse microscopy HeLa cells were cultured on 35 mm glass bottom dishes (MatTek) in DMEM (Invitrogen) with 10% FBS (HyClone) media. For the time-lapse microscopy, the cell culture dish was kept on a heated 37°C stage with 5% CO2. A spinning-disk confocal microscope (Leica DMI6000B) (Quorum Technologies) with an EM- CCD camera (Hamamatsu, C9100-12) and a 63X/1.4 N.A. HCX PL APO CS DIC oil objective lens or a 40X/1.25 N.A. HCX PL APO CS DIC oil objective lens were used for all time-lapse microscopy and the microscope was operated with Volocity software (PerkinElmer). Cargille Type 37 immersion oil was used. The 491 nm laser with BP 483/32 filter was used for GFP signal detection and the 561 nm laser with BP 692/40 filter was used for mCherry signal detection. The final videos were edited with Volocity software.

3.3.7 Flow cytometry Hela cells were transfected with GFP-Gas2 constructs in Effectene (Qiagen) reagent and the transfected cells were kept in a 37°C incubator for 32 hours to allow relative protein expression. Cells were dispersed with trypsin and fixed in

89 70% ethanol (-20°C) and re-suspended in propidium iodide (Sigma) solution for DNA labeling. Samples were scanned with an FACScan flow cytometry analyzer. Only GFP signal positive cells were counted in the cell population. Ten thousand cells were counted for each condition and experiments were repeated 3 times. The experimental results were analyzed with Students’ t-test.

3.3.8 Pharmacological treatments HeLa cells were incubated in 2 M taxol (Sigma) or DMSO (Fisher) media for 6 hours before fixation for immunofluorescence.

3.3.9 Protein expression and purification The FL-Gas2 (cDNA clone MGC: 18565) was cloned into pTrcHis2B vector with 6X His tag at N-terminus using the forward primer 5’-atgtgcactgccctga-3’ and reverse primer 5’-tttaatctccttcttagccttg-3’. The resultant plasmid pTrcHis2B-Gas2 was transformed into DH5 Escherichia coli to express FL-Gas2 protein. Protein was collected with Probond nickel-chelating resin (Invitrogen). The elution protein solution was kept in a dialysis bag (Spectra/Pov molecularporous membrane tubing, MWCO 12-14kDa, Fisher 08-667B) in PBS at 4°C overnight to remove imidazole. The protein sample was concentrated with Microcon centrifugal filter devices (Millipore: YM-10), and then was aliquoted into small volume and frozen in liquid nitrogen. Samples were stored at -80°C for further use. The protein purity was tested by Coomassie-stained SDS-PAGE, and no significant additional bands were observed.

3.3.10 Cytoskeleton co-sedimentation assays The purified tubulin (TL238-A, Cytoskeleton) was either first polymerized into MTs by adding 20 M taxol at 37°C for 30 minutes or mixed with the purified FL-Gas2 protein prior to polymerization. The purified FL-Gas2 protein was added by a series of half-concentration dilutions. Samples were loaded on the top layer of 30% sucrose in tubulin stabilization buffer (TSB: 1 mM EGTA, 5mM MgCl2, 80 mM K-PIPES at pH 7.0) and then centrifuged at 100,000g for 30 minutes at

90 room temperature. The supernant was carefully removed into a new tube and the pellet was resuspended into an equal volume as the supernant. Samples were separated on a 10% SDS-PAGE followed by Deep Purple (GE healthcare) staining. The SDS-PAGE gels were scanned with a Typhoon trio laser scanner (GE healthcare).

3.3.11 MT polymerization speed measurements Human EB3 in mCherry-pHSV virus was first used to transduct HeLa cells in a 35 mm glass bottom dishes (MatTek). The cell culture media was changed 5 hour later to reduce the toxicity. The transducted cells were then transfected with GFP- Gas2 domain construct using Effectene (Qiagen) reagent the next day and kept in the incubator for 18 hours. The cell culture dish was placed on a heated 37°C stage with 5% CO2. The mCherry-EB3 migration was recorded by a spinning-disk confocal microscope (Leica DMI6000B) (Quorum Technologies) with 63X/1.4 N.A. HCX PL APO CS DIC oil objective lens by 491 nm laser with BP 483/32 filter for GFP and 561 nm laser with BP 692/40 filter for mCherry. The EB3 migration speed was analyzed with the Imaris (Bitplane) program.

3.3.12 Electron microscopy The purified tubulin (TL238-A, Cytoskeleton) was first dissolved into general tubulin buffer (80 mM PIPES at pH 6.9, 2 mM MgCl2, 0.5 mM EGTA, 1 mM GTP), and then 20 M final concentration of taxol (Sigma) was added to polymerize tubulin into MTs. The purified FL-Gas2 protein in PBS was mixed with MTs for EM examination. For each EM sample, 5 l sample was placed on a glow discharged pre-carbon coated copper grid for 1 minute, and then was replaced by 5 l 1% uranyl acetate for a 1 minute treatment. Uranyl acetate was carefully removed and the treated grid was left by air-dried. Samples were examined in an FEI Tecnai 12 electron microscope, and digital images were taken by a Gatan 792 Bioscan wide angle multiscan CCD camera.

3.3.13 MT length measurement

91 MT length was measured with NIH ImageJ program and results were charted with Microsoft Office Excel software. Statistics analysis was performed using Student’s t-test.

3.3.14 FRAP experiments FRAP experiments were performed with a Zeiss 510 LSM Meta confocal microscope with Plan-Apochromat 63X/1.4 N.A. oil DIC objective lens. The cell culture was maintained on a heated 37°C stage with 5% CO2. The target areas were pre-scanned 3 times before bleaching. The system was set at zoom 1 with pinhole at 2 airy units and bleached with 488nm Ar ion laser at 100% power for 20 iterations. After bleaching, the recovery was recorded every 3 seconds with 488 nm Ar ion laser at 0.5% output power through a LP 505 filter with scan speed at 9 to avoid photobleaching. The experimental data were analyzed with Microsoft Office Excel software.

3.3.15 Size-exclusive chromatography An analytical Superose 12 300GL column (GE Healthcare) was equilibrated with 150 ml PBS overnight. The experiment was operated by ÄKTApurifier (GE Healthcare). The 250 g of purified FL-Gas2 protein PBS solution was loaded into the column and eluted at a flow rate of 1 ml per minute. Every ml fraction was collected. The data were plotted with Microsoft Office Excel software.

3.3.16 BS3 cross-linker Western blot The purified FL-Gas2 protein was centrifuged at 100,000 g for 10 minutes at 4°C for clarification. Supernant was collected and protein assays were performed using the BCA protein assay kit (Thermo). A 6 g of FL-Gas2 protein was treated with 1 mM BS3 cross-linker (Thermo) in PBS for 5 minutes at room temperature. Samples were separated on a 12% SDS-PAGE gel and transferred to 0.45 m Trans-Blot nitrocellulose paper (Bio-Rad), which was blocked with blotto buffer (5% non-fat milk in PBS) after transferring. After washing in 0.1% Tween-20 (Fisher) in PBS (PBS-T), polyclonal rabbit Gas2 antibodies were used at 1:3000

92 dilution in blotto buffer incubated at room temperature for 2 hours. The non- specific antibody binding was washed 3 times with PBS-T. Primary antibodies were detected using goat anti-rabbit IgG antibodies conjugated to horse radish peroxidase (HRP) (Molecular Probes) at 1:3000 dilution in PBS at room temperature for 1 hour, and followed by 3 times washing with PBS-T. Detection was performed using the ECL method.

93 3.4 Results

3.4.1 Gas2 protein contains both actin- and MT-binding domains Mus musculus (house mouse) Gas2 protein consists of 314 amino acids and bioinformatics analysis of its protein sequence reveals that mouse Gas2 protein contains an actin-binding CH domain near its N-terminus (amino acids#: 36-158) and a MT-binding Gas2 domain near its C-terminus (amino acids#: 201-274). There are two low complexity (LC) domains (amino acids#: 167-178 and 181- 198) between the CH and Gas2 domains, and the second LC domain contains 4 proline-serine (P-S) repeats (amino acids#: 183-190), which tends to give this region more structural flexibility (Zhang et al., 2011). The N-terminus GFP tagged FL-Gas2, its CH and Gas2 domains were cloned and used to investigate their functions in this study (Fig. 1A).

The FL-Gas2 protein was originally discovered as an F-actin binding protein in serum starved NIH 3T3 cells (Brancolini et al., 1992). Here, we showed that when over-expressing the GFP-FL-Gas2 protein in HeLa cells, it mainly co-localized with F-actin, which was recognized from its distribution patterns and the yellow color in the merged image (Fig. 1B). Bundled MTs were also observed in the same transfected cells (Fig. 1B arrows). In some cells, the GFP-FL-Gas2 protein cross-linked F-actin and MTs and it also altered the classical MT radial distribution into long-loop like structure (Fig. 1B’ arrow). GFP-CH domain co- localized with F-actin, as expected, since it was the actin-binding domain of the FL-Gas2 protein (Fig. 1C). The MT-binding, GFP-Gas2 domain, co-localized with MTs and also changed MTs’ radial distribution from MT organization center into long-loop like structure (Fig. 1D). GFP control group cells’ cytoskeleton morphology looked the same as non-transfected cells’ as expected (Fig. 1E). An auto-inhibition mechanism may also exist through a direct interaction between CH and Gas2 domains, which may also explain why the FL-Gas2 protein was originally discovered as an F-actin binding protein (Brancolini et al., 1992). To test this hypothesis, the RFP-CH domain and GFP-Gas2 domain were co- expressed in the same cell. If there was a direct interaction between the two

94 domains, the auto-inhibition mechanism would inhibit their binding to respective cytoskeleton components. However, we did not observe any significant co- localization between these two domains in the same cell (Fig. 1F), and they still bound relative cytoskeleton as usual. MTs’ morphology had also been changed by the Gas2 domain as expected on the basis of our previous experimental observation (Fig. 1D).

3.4.2 Over-expression of the FL-Gas2 protein, CH and Gas2 domains result in multinucleated cells The gas2 gene and protein levels are up-regulated upon serum starvation in NIH 3T3 cells, which also results in cells arresting at G0 (Schneider et al., 1988). Conversely, Gas2 protein level is down-regulated with serum and growth factor stimuli, allowing for cell cycle progression (Brancolini et al., 1992). These two observations imply a relationship between cell cycle progression and Gas2 protein levels. We have demonstrated that the FL-Gas2 protein, CH and Gas2 domains interact with the F-actin and MT cytoskeleton in Xenopus embryos and oocytes (Zhang et al., 2011). The cytoskeleton plays many key roles in a cell, such as in cell division (Heng and Koh, 2010). Here, we investigated the potential actin-MT cross-linking properties of the FL-Gas2 protein and its domains functions in cell division. The FL-Gas2 and its domains were expressed in HeLa cells to study their functions in cell division by time-lapse microscopy and flow cytometry.

GFP control cells divided normally (Fig. 2A). The GFP signal was evenly distributed in the entire cell and the cell divided into two daughter cells. Cells expressing GFP-CH domain completed normal cell division (Fig. 2B). The CH domain was recruited and observed localizing at the contractile ring during cell division (Fig. 2B arrows). Cells expressing GFP-FL-Gas2 protein had severely perturbed cell division attempts, followed by global bleb and presumably death through apoptosis (Fig. 2C). We observed a cell expressing GFP-FL-Gas2 had a problem with contractile ring formation (Fig. 2C arrow). The normal division cells were also observed with GFP-FL-Gas2 protein expression (data not shown).

95 We attribute the severity of phenotypes/effects due to the protein expression levels. Cells expressing GFP-Gas2 domain had the most striking phenotype (Fig. 2D). These cells had unusual long-loop like MTs, which had been also observed from our fixed cell images (Fig. 1D). The representative daughter cells shown in Fig. 2D never completed cytokinesis and they were connected with long midbody (arrow) following by global bleb.

Flow cytometry was performed to determine the cell cycle distribution of transfected cells. The analysis was done by assorting GFP signal positive cells to eliminate the non-expressing cells. The results showed that GFP-FL-Gas2, GFP- CH and GFP-Gas2 domains expressing cells all had significantly less G0/G1 cells, but more multinucleated cells (>4N) than GFP control group (Fig. 2E).

3.4.3 Gas2 domain has different effects on MT dynamic comparing with MT stabilization drug taxol We observed the long-loop like MTs in HeLa cells expressing GFP-Gas2 domain (Fig. 1D and Fig. 2D). The unusual long-loop like MTs’ properties had been changed by the Gas2 domain and MTs also formed a thick bundled network (Fig. 3A). We observed that MTs kept growing and shrinking in the limited space of a cell (Fig. 3B). The bundled MTs phenotype was also observed from a time-lapse movie, in which two thick MT filaments joined together (Fig. 3B arrows). MT stabilization drug taxol treated HeLa cells only raised short, linear and rigid- looking MTs (Fig. 3C), which shared no common phenotype when compared with the Gas2 domain expressing cells’ MTs. The taxol solvent dimethyl sulfoxide (DMSO) control group cells had morphologically normal MTs (Fig. 3D).

3.4.4 Gas2 domain facilitates MT polymerization We have observed that MTs in cells expressing Gas2 domain growing into long- loop like structure (Figs 1D, 2D and 3A). These results suggest that the Gas2 domain facilitates MT polymerization and presumably also inhibits their de- polymerization; therefore, a cell ends up containing very long MTs.

96

Cytoskeleton co-sedimentation assays were applied to study Gas2-MT biochemical properties. The same amount of purified tubulin and decreasing amount of FL-Gas2 protein were incubated together, and then MTs were polymerized and stabilized by adding taxol at room temperature. The analysis of the protein fractions after centrifugation by SDS-PAGE showed that there was more un-polymerized tubulin left in the supernant with the FL-Gas2 protein amount decreasing (Fig. 4A). The experimental results suggested that the FL- Gas2 protein facilitated MT polymerization. In order to confirm this result, MT plus end binding protein, EB3 and GFP-Gas2 domain were co-expressed in HeLa cells. The mCherry-EB3 (red color) migration speeds were carefully measured to monitor MTs polymerization changes in vivo (Fig. 4B). The analysis of EB3 migration speeds in Gas2 domain positive cells showed that EB3 migrated at 0.25±0.019 μm/s, which was 19% faster than control cells’, in which EB3 migrated at 0.21±0.020 μm/s on average (Fig. 4C). In other words, MTs polymerized 19% faster in Gas2 domain expression condition. This supported the results obtained with the cytoskeleton co-sedimentation assays (Fig. 4A), in which MTs polymerized much faster with more FL-Gas2 protein in the sample.

3.4.5 FL-Gas2 protein bundles MTs and significantly changes MT length We showed that cells expressing the Gas2 domain had long-loop like MTs (Figs 1D, 2D and 3A). To have a better understanding about how Gas2 interacts with MTs, the purified MTs and FL-Gas2 protein were examined by confocal and electron microscopy (EM). The average length of MTs labeled with fluorescent taxol was 3.3±0.6 μm, but the average length of the MTs incubated with the FL- Gas2 protein was 14.3±1.8 μm (Fig. 5A-C). The EM experiments confirmed this observation, and also revealed that the longer and brighter MTs seen from the confocal images were actually bundled different lengths MTs (Fig. 5D,E); therefore, we saw broader and brighter MTs from confocal images.

97 MTs not only have intrinsic polarity, but also have different biological and biochemical properties at their two ends (Baas and Lin, 2011). We have observed that bundled MTs by the FL-Gas2 protein in vitro (Fig. 5F,G), but we do not know the orientation of these bundled MTs. In an EB3 and Gas2 domain co- expressing cell, several EB3 migrated on the same bundled MT cluster towards the same direction (Fig. 5H arrows), where Gas2 domain parallel bundled MTs. Therefore, the FL-Gas2 protein also parallel bundles MTs in the same way as Gas2 domain does.

3.4.6 FL-Gas2 protein contains three functional PKC phosphorylation sites It has been shown that the FL-Gas2 protein is phosphorylated on serine amino acid during the G0 to G1 transition with the serum and growth factors stimulation allowing for cell cycle progression (Brancolini et al., 1992). The second LC domain of the FL-Gas2 protein contains 4 P-S repeats, which gives this region more structural flexibility (Zhang et al., 2011). The FL-Gas2 protein phosphorylation sites were predicted by NetPhosK. The results showed that there were three high potential phosphorylation sites by protein kinase C (PKC) (Fig. 6A). Two sites are in the second LC domain and they (S193 and S194) are also very close to the P-S repeats sequence (amino acids#: 183-190). A third site, T306, is at the C-terminal end of the FL-Gas2 protein. The phosphorylation- mimic and non-phosphorylation-mimic point mutations of GFP tagged FL-Gas2 constructs were made to further investigate these potential phosphorylation sites’ functions.

Expression of any of three phosphorylation-mimic single point mutation S193D, S194D or T306D led to actin-MTs cross-linking phenotypes and bundled MTs (Fig. 6B-D). Similar phenotypes were observed in the phosphorylation-mimic double point mutations S193D/S194D (Fig. 6E). However, these phenotypes were absent in the cells expressing the non-phosphorylation-mimic double point mutation S193A/S194A (Fig. 6F). The bundled MTs phenotype had been

98 observed in some S193A/S194A point mutation expressing cells, but there was no actin-MTs cross-linking evidence (Fig. 6F’). In the cells expressing the mutant with all three sites phosphorylation-mimic S193D/S194D/T306D point mutations, cells had multiple nuclei as a sign of failure in cytokinesis (Fig. 6G) and two daughter cells were still connected together (Fig. 6G’). These point mutation experiments have revealed that these potential phosphorylation sites have functional and possibly structural effects on the FL-Gas2 protein.

3.4.7 FL-Gas2 protein dynamically binds to MTs and it also forms an unstable complex Cytoskeleton co-sedimentation assays were once again used in order to carry out temporal analysis of the Gas2-MT biochemical binding properties. This time, the same amount of purified tubulin was first polymerized into MTs and then stabilized with taxol. Decreasing amounts of FL-Gas2 protein were added to a fixed amount of MTs in order to investigate the binding ratio between them. The analysis of the protein fractions after centrifugation by SDS-PAGE showed that there was always residual FL-Gas2 protein left in the supernant regardless the amount added to the MTs (Fig. 7A). One possibility is the FL-Gas2 protein dynamically binds to MTs; therefore, there is always a pool of mobile FL-Gas2 protein in the supernant. Fluorescence recovery after photobleaching (FRAP) experiment was used to test this hypothesis. The FRAP analysis graph showed that GFP-Gas2 domain dynamically bound to MTs with T1/2 about 2.5 seconds at both the edge and the center of a cell (Fig. 7B and C). The two different locations were chosen on the basis of MT different dynamic behaviors. The fluorescent signal recovery speeds were almost identical at these two locations (Fig. 7C).

Since FL-Gas2 protein only has one MT-binding domain, it needs at least two to bundle MTs. Therefore, we hypothesize that FL-Gas2 protein may form a complex to bundle MTs. Size-exclusive chromatography was used to test this hypothesis, and it resulted in three peaks as shown in the elution graph (Fig. 7D). The three peaks represent the appropriate sizes for the monomers, dimers and

99 trimers of FL-Gas2 protein by comparing with the loading standard. The monomers’ peak and area were much larger than the peaks and areas for the trimers’ and dimers’. The smallest dimers’ peak implied a transition between the monomers and trimers. Therefore, FL-Gas2 protein presumably forms an unstable dynamic complex in vitro. Bis sulfosuccinimidyl suberate (BS3) cross-linker with a spacer arm length 11.4 Å (8 atoms) was also applied to the same sample and it resulted 3 major bands by Western blot analysis (Fig. 7E). The sizes of these three bands match to the sizes of the monomers (35 kDa), dimers (70 kDa) and trimers (105 kDa) of the FL-Gas2 protein.

100 3.5 Discussion

In this study, we demonstrated that the FL-Gas2 protein not only co-localized with F-actin, but it also could cross-link F-actin and MTs in HeLa cells (Fig. 1B,B’). The over-expression of the FL-Gas2, its CH and Gas2 domains in HeLa cells resulted in a significant increase of multinucleated cells when compared to control cells (Fig. 2E). Furthermore, we demonstrated that the Gas2 domain bound MTs and changed MTs into long-loop bundled network (Figs 1D, 2D and 3A). The biochemical assays and size-exclusive chromatography revealed that the FL-Gas2 protein formed an unstable dynamic complex (Fig. 7D), which impacted its cytoskeleton binding properties. Together with the phosphorylation-mimic point mutation experiments data, we think that the non-phosphorylated FL-Gas2 protein forms a dynamic complex to bundle MTs and also facilitates MT polymerization, but cross-linking actin and MTs is achieved by the phosphorylated FL-Gas2 protein.

The MTs in cells expressing the Gas2 domain appeared as long-loop like structure and they also concentrated at the midbody during cell division (Fig. 2D). The midbody is composed of anti-parallel MTs, which are bundled by proteins such as the protein regulator of cytokinesis 1 (PRC1) (Jiang et al., 1998; Mollinari et al., 2002) and also include a group of parallel MTs (Elad et al., 2011). However, the confocal microscopy optical resolution cannot provide us with too much information about single MT orientation. We confirmed the FL-Gas2 protein MT bundling ability using EM, but it was also difficult to distinguish the MT orientation with purified MT. A MT plus end binding protein, EB3, helped to resolve the problem. We observed EB3 migrated towards the same direction on the Gas2 domain bundled MTs (Fig. 5H). Therefore, we believe that Gas2 protein parallel bundles MTs just like its Gas2 domain. We have also discovered that the FL-Gas2 protein facilitates MT polymerization via its Gas2 domain, but Gas2 domain effects on MTs are different than what is observed with the MT stabilization drug taxol. Taxol binds and stabilizes MTs; therefore, it raises short and rigid looking MTs (Fig. 3C). On the other hand, the FL-Gas2 protein or Gas2

101 domain dynamically binds to MTs and promotes MTs polymerization, which results in long-loop like MTs (Fig. 3A and B).

Our flow cytometry experiments showed that cells expressing GFP-Gas2 constructs had significantly less G0/G1 cells, but more multinucleated cells (>4N) than the GFP control group (Fig. 2E). The FL-Gas2 protein is up-regulated when a cell enters the G0 phase of cell cycle (Schneider et al., 1988); therefore, it may work like a “brake” to ensure the cell cycle arresting and prevents cell division. One may think when over-expression of the FL-Gas2 protein will result in more G0 phase cells, but in fact, more dividing cells were first directly affected by the FL-Gas2 protein “brake” ability and more none-completely divided cells (>4N) were accumulated with time in the cell population. The slightly less multinucleated cells in GFP-FL-Gas2 and GFP-Gas2 domain transfected cell population were presumably due to their lethal effects (Fig. 2C,D). The dead cells were floating on the top of the culture media and had been washed away during sample preparations.

There are three putative PKC phosphorylation sites in the mouse Gas2 protein by NetPhosK prediction. Interestingly, two sites localize at the second LC domain and they (S193 and S194) are also very close to the P-S repeats sequence (amino acids#: 183-190) (Zhang et al., 2011). If they were phosphorylated, they could directly impact on the flexible linker and even change the structure of the FL- Gas2 protein. One of the possibilities may expose the Gas2 domain to MTs and the FL-Gas2 protein gains the actin-MT cross-linking ability. We have tested this hypothesis by making phosphorylation-mimic point mutation on these sites. Both S193D and S194D phosphorylation-mimic single point mutation give rise to an actin-MT cross-linking phenotype (Fig. 6B,C). The non-phosphorylation-mimic double point mutation S193A/S194A only binds to F-actin or the MTs have the bundled phenotype (Fig. 6 F,F’). Therefore, we think that there is a structural change in the FL-Gas2 protein if either of the 2 sites gets phosphorylated. It also explained why the FL-Gas2 protein mainly co-localized with F-actin (Fig. 1B),

102 but sometimes it also cross-linked actin-MTs (Fig. 1B’). We think the cross- linking phenotype is achieved by the phosphorylated version of the FL-Gas2 protein. We also obtained other evidences that supported these conclusions. For example, the bundled MTs phenotype in cells expressing the S193A/S194A point mutation (Fig. 6F’) looks similar to the wild type FL-Gas2 protein bundled MTs (Fig. 1B). We purified the non-phosphorylation version of the FL-Gas2 protein for the EM experiments and observed its MT bundling ability (Fig. 5E,G). The Gas2 domain itself also has the ability to bundle MTs (Fig. 3B). Therefore, we think the FL-Gas2 protein MT bundling ability is via its Gas2 domain. Now, the question is why the non-phosphorylation version FL-Gas2 protein only binds to F- actin or bundles MTs, but cannot cross-link them. The size-exclusive chromatography experiments revealed that the FL-Gas2 protein formed an unstable dynamic complex, but the monomers were the most dominant conformation (Fig. 7D). The cytoskeleton co-sedimentation assays (Fig. 7A) and FRAP experiments (Fig. 7B and C) showed that the FL-Gas2 protein or Gas2 domain dynamically bound/bundled MTs. We hypothesize the Gas2 trimers have the ability to bundle MTs since the FL-Gas2 protein only has one MT-binding domain and it takes at least two to bundle MTs. However, the trimers are unstable since the size-exclusive chromatography has shown that there is an exchange between the large population of monomers and relative small population of trimers. The majority of the FL-Gas2 exists as monomers and this form is the one that binds to F-actin. Since the monomers are the most abundant form in the population; therefore, it is not unexpected to see that most of the wild type FL- Gas2 protein binds to F-actin. This could be also one of the reasons that it was first discovered as an actin-binding protein. Of course, the differences in the protein binding affinity may also have significant effects here.

From our experimental data, we propose that the three potential phosphorylation sites have functional and structural effects on the FL-Gas2 protein. The two sites, S193 and S194 are involved in the FL-Gas2 protein functional changes. We hypothesize that the FL-Gas2 protein is in a folded conformation and it has a

103 higher binding affinity to F-actin as monomers. This folded conformation Gas2 forms unstable trimers and binds/bundles MTs. When S193, S194 or both get phosphorylated by a kinase, PKC for instance, the FL-Gas2 protein becomes an open conformation. This open conformation FL-Gas2 protein also forms a complex and cross-links F-actin and MTs (Fig. 8A). A similar mechanism also applies to the T306 site, but this close-open conformation (Fig. 8B) is controlled by this end site, and the same structural and functional changes apply to FL-Gas2 protein just like what we proposed for the fold-open model. However, our experimental data cannot exclude that these phosphorylation sites may work in some different combination.

Three Gas2-like proteins have been identified in human and they all share the similar domain structures (Stroud et al., 2011). There is a long C-terminus domain following the Gas2 domain in Gas2-like 1, 2 and 3 proteins, which is also the main difference between the Gas2-like proteins and Gas2 protein. Some study has reported that the Gas2 domain in the Gas2-like 3 protein has weaker binding affinity to MTs than the C-terminus domain itself in protein truncation studies (Stroud et al., 2011). We think this long C-terminus domain in Gas2-like 3 protein may directly impact on protein functions including the ability to bind MTs.

There is an interesting relationship between the FL-Gas2 protein levels and cell division. It has been shown that the FL-Gas2 protein level is up-regulated by serum starvation to induce the cells to enter growth-arrest condition (Schneider et al., 1988). We have also demonstrated that over-expressing the FL-Gas2 in both Xenopus embryos (Zhang et al., 2011) and HeLa cells result more multinucleated cells (Fig. 2E). Therefore, the Gas2 protein can be thought as a cell division suppressor protein. A cell must carefully regulate the Gas2 protein levels to progress through normal cell division. Some cancers have been found to have very low levels or no Gas2 protein expression (Brancolini et al., 1992; Kondo et al., 2008; Mazzoni et al., 2005; Petroulakis et al., 2009). These evidences support the notion that there is also a relationship between Gas2 protein levels and certain

104 cancers. We think that the Gas2 protein functions as a cytoskeleton binding cancer suppressor protein.

105 3.6 Acknowledgements

We would like to thank Prof. Jacque Paiement at University of Montreal for providing the Gas2 antibody; Dr. Tadayuki Shimada at Montreal Neurological Institute for providing mCherry-EB3 construct; Dr. Dirk Hubmacher at McGill University for helping on the size-exclusive chromatography; Mr. Ken McDonald at Flow Cytometry Facility of McGill University for helping on the flow cytometry; Facility for Electron Microscopy Research of McGill University for using the electron microscope; Cell Imaging and Analysis Network of McGill University for using the confocal and spinning-disk microscopes and Typhoon trio laser scanner and the Imaging Facility of Life Sciences Complex of McGill University for using the image analysis software.

106 3.7 References

Baas, P. W. and Lin, S. (2011). Hooks and comets: The story of microtubule polarity orientation in the neuron. Developmental neurobiology 71, 403-18. Brancolini, C., Bottega, S. and Schneider, C. (1992). Gas2, a growth arrest-specific protein, is a component of the microfilament network system. The Journal of cell biology 117, 1251-61. Collavin, L., Buzzai, M., Saccone, S., Bernard, L., Federico, C., DellaValle, G., Brancolini, C. and Schneider, C. (1998). cDNA characterization and chromosome mapping of the human GAS2 gene. Genomics 48, 265-9. Conrad, P. A., Giuliano, K. A., Fisher, G., Collins, K., Matsudaira, P. T. and Taylor, D. L. (1993). Relative distribution of actin, myosin I, and myosin II during the wound healing response of fibroblasts. The Journal of cell biology 120, 1381-91. Elad, N., Abramovitch, S., Sabanay, H. and Medalia, O. (2011). Microtubule organization in the final stages of cytokinesis as revealed by cryo-electron tomography. Journal of cell science 124, 207-15. Fletcher, D. A. and Mullins, R. D. (2010). Cell mechanics and the cytoskeleton. Nature 463, 485-92. Heng, Y. W. and Koh, C. G. (2010). Actin cytoskeleton dynamics and the cell division cycle. The international journal of biochemistry & cell biology 42, 1622-33. Jiang, W., Jimenez, G., Wells, N. J., Hope, T. J., Wahl, G. M., Hunter, T. and Fukunaga, R. (1998). PRC1: a human mitotic spindle- associated CDK substrate protein required for cytokinesis. Molecular cell 2, 877- 85. Kondo, Y., Shen, L., Cheng, A. S., Ahmed, S., Boumber, Y., Charo, C., Yamochi, T., Urano, T., Furukawa, K., Kwabi-Addo, B. et al. (2008). Gene silencing in cancer by histone H3 lysine 27 trimethylation independent of promoter DNA methylation. Nature genetics 40, 741-50.

107 Lee, M. P. and Feinberg, A. P. (1997). Aberrant splicing but not mutations of TSG101 in human breast cancer. Cancer Research 57, 3131-3134. Li, L., Li, X., Francke, U. and Cohen, S. N. (1997). The TSG101 tumor susceptibility gene is located in chromosome 11 band p15 and is mutated in human breast cancer. Cell 88, 143-154. Mazzoni, I., Zhang, T., Goetz, J., Wagner, J., Taheri, M., Munger, C., Sacher, M., Gilchrist, A., Au, C., Thibaul, G. et al. (2005). Involvement of Growth-arrest-specific protein 2, a smooth endoplasmic reticulum protein identified by proteomics, in cell cycle control. Canadian Proteomics Initiative Conference. Mitelman, F., Mertens, F. and Johansson, B. (1997). A breakpoint map of recurrent chromosomal rearrangements in human neoplasia. Nature genetics 15 Spec No, 417-74. Mollinari, C., Kleman, J. P., Jiang, W., Schoehn, G., Hunter, T. and Margolis, R. L. (2002). PRC1 is a microtubule binding and bundling protein essential to maintain the mitotic spindle midzone. The Journal of cell biology 157, 1175-86. Petroulakis, E., Parsyan, A., Dowling, R. J., LeBacquer, O., Martineau, Y., Bidinosti, M., Larsson, O., Alain, T., Rong, L., Mamane, Y. et al. (2009). p53-dependent translational control of senescence and transformation via 4E-BPs. Cancer cell 16, 439-46. Rodriguez, O. C., Schaefer, A. W., Mandato, C. A., Forscher, P., Bement, W. M. and Waterman-Storer, C. M. (2003). Conserved microtubule-actin interactions in cell movement and morphogenesis. Nature cell biology 5, 599-609. Schneider, C., King, R. M. and Philipson, L. (1988). Genes specifically expressed at growth arrest of mammalian cells. Cell 54, 787-93. Steiner, P., Barnes, D. M., Harris, W. H. and Weinberg, R. A. (1997). Absence of rearrangements in the tumour susceptibility gene TSG101 in human breast cancer. Nature genetics 16, 332-333.

108 Stroud, M. J., Kammerer, R. A. and Ballestrem, C. (2011). Characterization of G2L3 (GAS2-like 3), a New Microtubule- and Actin-binding Protein Related to Spectraplakins. The Journal of biological chemistry 286, 24987-95. Sun, D., Leung, C. L. and Liem, R. K. (2001). Characterization of the microtubule binding domain of microtubule actin crosslinking factor (MACF): identification of a novel group of microtubule associated proteins. Journal of cell science 114, 161-172. Zhang, T., Dayanandan, B., Rouiller, I., Lawrence, E. J. and Mandato, C. A. (2011). Growth-arrest-specific protein 2 inhibits cell division in Xenopus embryos. PloS one 6, e24698.

109 Chapter 3-Figure 1. Gas2 protein cross-links F-actin and MTs via its actin-binding CH domain and MT-binding Gas2 domain. (A) The N-terminus GFP tagged FL-Gas2 protein, its CH and Gas2 domains constructs were used in this study. LC stands for the low complexity domain. (B-E) Hela cells were transfected with different GFP tagged Gas2 constructs and the cells were fixed and stained to visualize the cytoskeleton morphology by confocal miscroscopy. (B) The GFP- FL-Gas2 protein mainly co-localized with F-actin. MTs were observed to be bundled together in the same expressing cell (arrows). (B’) In some cells, GFP- FL-Gas2 protein cross-linked F-actin and MTs, and changed MTs’ classic radial distribution morphology into long-loop like structure (arrow). (C) The GFP-CH domain co-localized with F-actin as expected. (D) The GFP-Gas2 domain not only co-localized with MTs, but also changed the expressing cells’ MTs into long- loop like structure. (E) GFP control cells’ cytoskeleton appeared the same as non- expressing cells’ as expected. Bars, 20 μm. (F) Individual RFP-CH and GFP- Gas2 domains constructs were transfected into Hela cells to examine the possibility of any auto-inhibition/interaction between them. There was no significant co-localization between RFP-CH and GFP-Gas2 domains in the same cell. RFP-CH domain appeared as F-actin pattern and GFP-Gas2 domain co- localized with long-loop like MTs. Bar, 10 μm.

110 Figure 1

111 Chapter 3-Figure 2. Over-expression of the FL-Gas2 protein, CH or Gas2 domains result in multinucleated cells. The FL-Gas2 and its domains were expressed in HeLa cells to study their effects in cell division by time-lapse microscopy and flow cytometry. (A) The GFP control cells divided normally. (B) The cells expressing GFP-CH domain successfully completed normal cell division and it also localized at the contractile ring during cell division (arrows). (C) A cell expressing GFP-FL-Gas2 protein had side effects on its contractile ring formation (arrow), and it died from global bleb. (D) The MTs in cells expressing GFP-Gas2 domain appeared as long-loop like structure. Two daughter cells died from un- completed abscission (arrow). Bars, 20 μm. (E) Flow cytometry analysis of cells expressing GFP-Gas2 constructs showed that all GFP-Gas2 constructs resulted in less G0/G1 cells, but more multinucleated cells than GFP control cells. GFP control group has 60.02±0.92% G0/G1 cells, 13.11±0.83% S cells, 17.25±0.62% G2/M cells and 9.84±2.79% >4N cells. GFP-FL-Gas2 group has 40.09±1.38% G0/G1 cells, 17.36±2.72% S cells, 18.11±0.27% G2/M cells and 24.48±0.23*% >4N cells. GFP-CH domain group has 35.24±0.61% G0/G1 cells, 18.33±3.48% S cells, 18.76±0.06% G2/M cells and 27.95±2.27*% >4N cells. GFP-Gas2 domain group has 43.69±4.25% G0/G1 cells, 19.28±1.83% S cells, 19.92±0.18% G2/M cells and 17.69±0.37*% >4N cells. Mean values were reported as ± s.d. of the mean. The data was collected from 3 independent experiments. All GFP-FL-Gas2, GFP-CH and GFP-Gas2 domain >4N results were significant comparing to GFP control. *P<0.001 was compared by Students’ t-test.

112 Figure 2

113 Chapter 3-Figure 3.

The Gas2 domain raises long-loop like bundled MTs, which is different than taxol treated short and rigid MTs. HeLa cells were either transfected with GFP-Gas2 domain or treated with MT stabilization drug taxol to examine their effects on MTs. (A) The MTs in cells expressing GFP-Gas2 domain changed their classic radial distribution morphology and formed a thick bundled network. (B) The MTs in cells expressing GFP-Gas2 domain kept growing and shrinking in the limited space of a cell. MTs bundled into thick clusters (arrows). (C) MTs in taxol treated HeLa cells appeared as short, linear and rigid morphology. (D) MTs in DMSO control group cells had the normal morphology. Bars, 5 μm.

114 Figure 3

115 Chapter 3-Figure 4. The Gas2 domain facilitates MTs polymerization. (A) The equal amount of purified tubulin and decreasing amount of FL-Gas2 protein were incubated together. MTs were polymerized and stabilized by adding taxol at room temperature. Cytoskeleton co-sedimentation assays showed that there were more un-polymerized tubulin left in the supernant with the Gas2 protein amount decreasing in SDS-PAGE. (B) The mCherry-EB3 (red color) and GFP-Gas2 domain (green color) co-expressing cells were used to monitor MT polymerization speed comparing with the mCherry-EB3 alone control cells’. EB3 molecules bound at the fast growing end of MTs and they were present for MT polymerization. Bar, 20 μm. (C) EB3 migrated at 0.25±0.019* μm/s in cells expressing Gas2 domain, which was approximately 19% faster than control cells’, in which EB3 migrated at 0.21±0.020 μm/s on average. The data was collected from 3 independent experiments and 30 measurements for each condition were done. Mean values were reported as ± s.d. of the mean. *P<0.001 was compared by Students’ t-test.

116 Figure 4

117 Chapter 3-Figure 5. Gas2 protein bundles MTs and makes MTs appear longer in vitro. The purified MTs alone and in combination with FL-Gas2 protein were examined and analyzed by confocal and EM. (A,B) MTs appeared thicker and longer when the FL-Gas2 protein was in the sample examined by confocal microscopy. Bars, 20 μm. (C) When MTs were mixed with the FL-Gas2 protein, they appeared almost five times longer than MTs alone (14.3±1.8* μm vs. 3.3±0.6 μm) by confocal microscopy analysis. Mean values were reported as ± s.d. of the mean. The data was collected from 3 independent experiments and 30 measurements for each condition were done. *P<0.001 was compared by Students’ t-test. (D,F) MTs alone randomly distributed on the EM grid. (E,G) The FL-Gas2 protein bundled different lengths MTs; therefore, MTs appeared as thicker, longer and organized pattern. Bars in D and E, 2 μm; in E and G, 200 nm. (H) EB3 migrated towards the same direction on a Gas2 domain bundled MT cluster; therefore, Gas2 domain presumably parallel bundled MTs. Bar, 3.3 μm.

118 Figure 5

119 Chapter 3-Figure 6. The PKC phosphorylation sites have functional and presumably structural effects on the FL-Gas2 protein. (A) Schematic representation of the GFP-FL-Gas2 protein depicting the functional domains and the locations of the potential PKC phosphorylation sites by NetPhosK prediction. Two of them (S193 with score: 0.81 and S194 with score: 0.92) localized inside the second LC domain. The last one (T306 with score: 0.94) was at the end. (B-G’) GFP constructs with the single, double or triple point mutations of the FL-Gas2 protein were transfected in Hela cells and subsequently examined for their interactions with F-actin and MT. (B-D) The cells expressing S193D, S194D or T306D all had actin-MT cross-linking phenotypes, and MTs also bundled together. (E) Cells expressing S193D/S194D had similar phenotypes as the single point mutation. (F) Cells expressing S193A/S194A looked like the wild type FL-Gas2 expressing cells. (F’) Some cells expressing S193A/S194A also had bundled MTs phenotype. Bars, 20 μm. (G,G’) Cells expressing S193D/S194D/T306D triple point mutation had multiple nuclei as a sign of failure in cytokinesis. DAPI staining was used to show the multiple nuclei in these cells, but they did not appear in the merged images. Bars, G: 20 μm and G’ 10 μm.

120 Figure 6

121 Chapter 3-Figure 7. Gas2 protein forms a dynamic complex to bind and bundle MTs. (A) Cytoskeleton co-sedimentation assays were used to study the binding ratio between MTs and the FL-Gas2 protein. The equal amount of purified tubulin was first polymerized into MTs and then stabilized with taxol, and decreasing amounts of FL-Gas2 protein were added to each sample. The SDS-PAGE results showed that there was always a mobile pool of FL-Gas2 protein left in the supernant, which revealed the possibility of the FL-Gas2 protein dynamically bound to MTs. (B,C) FRAP experiments were performed to cells expressing GFP-Gas2 domain. Two different locations (at a cell’s edge and the center) of a cell were chosen on the basis of MT different dynamic properties. The experimental results further confirmed that Gas2 domain dynamically binds to MTs with a similar recovery rate at both the edge and the center of a cell. Three independent experiments and 30 measurements for each condition were done. Bar, 10 μm. (D) Size-exclusive chromatography experiments revealed that the FL-Gas2 protein formed a dynamic complex and the majority of FL-Gas2 protein existed as monomers in solution. (E) BS3 cross-linker treated FL-Gas2 protein sample also resulted three major bands present for Gas2 monomers, dimers and trimers by Western blot analysis.

122 Figure 7

123 Chapter 3-Figure 8. The hypothetical models of Gas2 protein. From our experimental data, we propose that the three potential phosphorylation sites play very important functional roles in the FL-Gas2 protein. (A) The fold-open model. S193 and S194 two sites may be involved in the conformation changes of the FL-Gas2 protein. When the FL-Gas2 protein is in a folded conformation, it has a higher binding affinity to F-actin. This folded conformation FL-Gas2 forms unstable trimers and bundles MTs. When S193, S194 or both sites get phosphorylated by a kinase, PKC for instance, the FL-Gas2 protein becomes an open conformation. This open conformation FL-Gas2 protein also forms a complex and cross-links F-actin and MTs. (B) The close-open model. A similar mechanism may also apply to T306 site, but the close-open conformation is controlled by this end site, and the same structural and functional changes apply to Gas2 protein just like what we proposed for the fold-open model above. However, our experimental data cannot exclude that there is a possibility that all of the three sites are involved in the structural and functional changes of the FL-Gas2 protein or in some order combinations. Arrows meanings: black double direction arrows: inter-changeable; green arrows: favorite binding partner or favorite direction; and red arrows: non- favorite direction. Pi: phosphorylation.

124 Figure 8

125

Chapter 4

4. DISCUSSION

126 4.1 Gas2 protein regulations A protein can be regulated by its expression or post-translation modification (Hodges et al., 1998; Narayanan and Jacobson, 2009). The regulations of gas2 gene and protein levels are similar to the myristoylated alanine-rich protein kinase C substrate (MARCKS), which is also up-regulated in growth-arrest condition and down-regulated during G0 to G1 transition (Brooks et al., 1992; Herget et al., 1993; Thelen et al., 1991). The ability of MARCKS cross-linking F-actin is negatively regulated by phosphorylation and calcium (Hartwig et al., 1992).

The regulations of Gas2 protein in G0 cells are done in two steps. The Gas2 protein is phosphorylated in the first few minutes and then Gas2 translation is decreased within in few hours after serum addition (Brancolini et al., 1992). The cells have to decrease the levels of Gas2 protein to re-entry regular cell cycle. We have investigated three high potential phosphorylation sites in Gas2 protein and confirmed that they all have impacts on Gas2 protein cytoskeleton binding ability (Chapter 3-Figure 6). The Gas2 protein may play a brake role for cell cycle progress and maintain them at G0. The C-terminus of Gas2 protein has been shown to be involved in apoptosis (Brancolini et al., 1995). The Gas2 protein not only maintains cells in G0, but also works as a functional protein for apoptosis. Both of these two functions are executed via the cell cytoskeleton system. The expression of Gas2 protein is also increased susceptibility to apoptosis (Brancolini et al., 1997). This suggests that Gas2 may play dual roles in apoptosis, as a caspase effecter and as an apoptosis modulator in response to specific signals (Collavin et al., 1998). Our flow cytometry experiments showed that cells expression Gas2 constructs had significantly less G0/G1 cells, but more multinucleated cells than the control group (Chapter 3-Figure 2E), but the multiple nuclei cells are slightly less than the over-expression CH domain condition. The less multinucleated cells in GFP-FL-Gas2 and GFP-Gas2 domain transfected cells were presumably due to their lethal effects to the cells (Chapter 3-Figure 2C,D). The dead cells were floating on the top of the culture media and lost during sample preparations.

127 4.2 Gas2 protein phosphorylation It has been shown that Gas2 protein localizes at the cell border in serum starved NIH 3T3 cells (Brancolini et al., 1995). When Gas2 protein is phosphorylated by serum or growth factors addition, it localizes at the cell membrane ruffles (Brancolini et al., 1995). The membrane ruffles are formed by actin polymerization at the inner surface of the membrane (Stossel, 1993). This suggests a relationship between the Gas2 protein phosphorylation and its subcellular locations. The Gas2 protein subcellular locations are presumably determined by its size and charge since the phosphorylation not only puts more mass on a protein, but also changes the net charge of a protein. F-actin, -actinin and together with myosin-2 form stress fibers in cultured cells, but F-actin together with filamin forms membrane ruffles (Matsudaira, 1994). The different F-actin morphological distributions decide if Gas2 or phosphorylated Gas2 can be fit among F-actin cables.

Our bioinformatics study showed that there were three high potential phosphorylation sites in Gas2 protein. The phosphorylation-mimic and non- phosphorylation-mimic point mutations of FL-Gas2 had different effects on its cytoskeleton binding properties (Chapter 3-Figure 6). We are currently studying these effects with the live cell imaging methods. The relative mutations are also expressed to purify their proteins for EM studies. We observed that FL-Gas2 had the ability to bundle MTs, but not F-actin in vitro (Chapter 2-Figure 6E). No actin-MT cross-linking phenotype was observed in this study. The protein used in the EM studies was the non-phosphorylated Gas2. We observed the phosphorylation-mimic Gas2 gave actin-MT cross-linking phenotype in vivo (Chapter 3-Figure 6). Therefore, we believe that the phosphorylated Gas2 has the ability to cross-link actin and MTs, which was proposed in our models (Chapter 3-Figure 8).

4.3 Gas2 protein dynamic complex We observed that Gas2 protein bundled MTs (Chapter 3-Figure 3A,B). It takes at least two MT-binding domains to bundle MTs, but Gas2 protein only has one.

128 Therefore, Gas2 protein presumably forms a complex. We have confirmed this hypothesis in the size-exclusive chromatography experiments. The Gas2 protein forms a dynamic complex with conformation changes from monomers to trimers via dimers. We believe the Gas2 trimers have the ability to bundle MTs. However, the trimers are unstable as shown that there is an exchange between the large monomers population and relative smaller trimers population via the dimers (Chapter 3-Figure 7D). The monomers are the most abundant conformation in the Gas2 protein population; therefore, it is not unexpected to see that most of the Gas2 protein binds to F-actin. This could be also one of the reasons that it was first discovered as an actin-binding protein (Brancolini et al., 1992). The differences in the protein binding affinity may also have significant effects. Two Gas2 proteins in the trimer complex bundle MTs and the third one may work as a scaffold to hold the dimer together since the dimer itself is not stable. The third Gas2 protein may also be a member of another trimer complex and bundles MTs with its dimer partner. This could explain that we saw bundled MTs overlapping with other MTs in the EM studies (Chapter 3-Figure 5G).

4.4 Gas2 facilitates MT polymerization We showed that Gas2 facilitated MT polymerization in both vivo and vitro. The exact mechanisms of how Gas2 facilitates MT polymerization are not known. Gas2 may help to bring tubulin dimers to the polymerizing MTs or delay the GTP hydrolysis in MTs. Gas2 dynamically binds and bundles MTs via its Gas2 domain. Therefore, Gas2 may help to increase the local tubulin concentration near the MT polymerization site to facilitate MT polymerization. Gas2 can also stabilize MTs at the same time. Gas2 also forms a dynamic complex to bundle MTs. All these effects by Gas2 lead to MT faster polymerization and more stable MTs.

4.5 Gas2 and taxol different effects on MTs When Xenopus oocytes were treated with MT stabilization drug taxol prior to wounding, an abnormal double actin ring formed around the wound border (Mandato and Bement, 2003). Staining with phalloidin for F-actin showed that the

129 internal ring was composed of contractile F-actin, while staining for actin monomers showed that the outer ring was formed by de novo actin polymerization (Mandato and Bement, 2001; Mandato and Bement, 2003). The abnormal double actin ring is a typical phenotype arising by taxol during oocytes wound healing. A similar phenotype will be expected when MTs are stabilized. Interestingly, we found that when FL-Gas2 or its Gas2 domain was over- expressed, oocytes resulted the same abnormal double actin ring phenotype after wounding (Chapter 2-Figure 4). This suggested that Gas2 had MT-stabilizing ability. We also showed that Gas2 domain dynamically bound MTs by FRAP experiments (Chapter 3-Figure 7). The Gas2 induced double rings could be rescued by de-polymerizing MTs with nocodazole. It was interesting to note that the Gas2 domain alone also formed a ring structure and localized between two actin rings during the wound healing, but it did not overlap with either actin ring since it had no actin-binding domain. It has been shown that both actin and MTs are moved together towards the wound border during oocyte wound healing (Mandato and Bement, 2003). From these facts, we could know the Gas2 domain, which localized between the double actin ring, was brought by MTs to the wound border.

The midbody is composed of anti-parallel MTs, which are bundled by proteins such as the protein regulator of cytokinesis 1 (PRC1) (Jiang et al., 1998; Mollinari et al., 2002) and also include a group of parallel MTs (Elad et al., 2011). We observed that MTs in HeLa cells expressing the Gas2 domain appeared as long- loop like structure and they also concentrated at the midbody during cell division (Chapter 3-Figure 2D). However, the confocal microscopy optical resolution cannot provide us with too much information about the bundled MT orientation. We also confirmed the FL-Gas2 protein MT bundling ability by EM, but it was also difficult to distinguish the MT orientation with purified MT. We observed EB3 migrated towards the same direction on the Gas2 domain bundled MTs (Chapter 3-Figure 5H). Therefore, the Gas2 protein presumably parallel bundles MTs just like its Gas2 domain.

130

Both Gas2 domain and taxol raise the double actin ring phenotype during Xenopus oocytes wound healing. By comparing their effects on HeLa cells’ MTs, we know their mechanisms on MTs are not identical. The common result is that they both change the MT dynamical properties. Gas2 domain dynamically binds, bundles MTs and also facilitates MT polymerization by an unknown mechanism. On the other hand, taxol binds -tubulin and stabilizes polymerized MT (Orr et al., 2003). So, we saw taxol stabilized MTs appeared as short and rigid morphology, but Gas2 domain bound MTs resulted as long-loop like network (Chapter 3-Figure 3).

4.6 Gas2 protein and Rho GTPase Gas2 protein contains both actin- and MT-binding domains. We observed that Gas2 protein co-localized with F-actin and also cross-linked actin and MTs in HeLa cells (Chapter 3-Figure 1B’). It has been shown that the Rho family GTPase not only regulate actin, but can be regulated by MTs. MT de-polymerization or reduce MT density promotes RhoA activation (Ren et al., 1999). RhoA is activated at low density MTs away from the cleavage furrow, which promotes contraction between the aster MTs (Mandato et al., 2000). MTs polymerization promotes Rac1 activation and Rac1 regulates both actin and MTs polymerization in forming lamellipodia (Small et al., 2002; Waterman-Storer et al., 1999; Wittmann et al., 2003). As a cytoskeleton binding protein, Gas2 protein could also be regulated directly or indirectly by Rho GTPase. The mechanisms of Rho family GTPase effects on Gas2 functions and cytoskeleton binding are important topics for future studies. Together with Rho family GTPase, Gas2 protein may work as a key player in many biological pathways and with varied functions in regulations of cytoskeleton, cell cycle, apoptosis and even many cancer diseases.

4.7 Gas2 knockdown study The RNA interference (RNAi) technique has been widely used for knockdown study. Knocking down certain gene will also deplete the coding protein at its expression level (Caplen and Mousses, 2003). The future Gas2 study should use

131 RNAi to knock down gas2 gene and investigate relative effects. The gas2 gene knockdown cells can be studied by live cell imaging for both cytoskeleton dynamics and cell division defects. A recent study by RNAi knockdown g2l3 showed that it resulted abnormal chromosomes oscillation, which was also observed in deletion of anillin and led to cytokinesis failure (Wolter et al., 2012). The depletion of G2L3 by RNAi also resulted multiple nuclear cells (Wolter et al., 2012). We had the same observations by over-expression of Gas2 in Xenopus embryos and HeLa cells (Chapter 2-Figure 3O and Chapter 3-Figure 2E). It should be noticed that Gas2 and G2L3 have different molecular sizes and expression levels. Gas2 expresses the highest level at G0/G1, but G2L3 highest level is at G2/M (Wolter et al., 2012). Even they share the same domains and are categorized in the same protein family, the different expression levels through cell cycle are implicated that they have functions in cells.

4.8 Gas2 protein structure Gas2 protein contains both actin-binding CH domain and MT-binding Gas2 domain. Both CH and Gas2 domains structures have been solved, but there is no FL-Gas2 protein structure available to date (). The FL-Gas2 structure should not be simply thought as the combination of the CH and Gas2 domains structures since there are two LC domains between the CH and Gas2 domains and the second LC domain contains 4 P-S repeats, which gives this region more structural flexibility (Zhang et al., 2011). The FL-Gas2 protein structure is important for understanding Gas2 protein regulations and conformation changes.

4.9 Conclusion In this thesis, I have investigated Gas2 protein functions in cell division and its cytoskeleton binding mechanisms, especially Gas2 domain affects MT dynamics. Gas2 protein can be thought as a cell division suppressor protein. A cell must carefully regulate the Gas2 protein levels to progress through cell cycle. Some cancers have been found with very low levels or no Gas2 protein expression (Brancolini et al., 1992; Mazzoni et al., 2005; Petroulakis et al., 2009). These

132 evidences support the notion that there is a relationship between Gas2 protein levels and certain cancers. Taken together, Gas2 protein functions are mediated cytoskeleton binding and may function as a cancer suppressor protein.

133 4.10 References

Brancolini, C., Benedetti, M. and Schneider, C. (1995). Microfilament reorganization during apoptosis: the role of Gas2, a possible substrate for ICE-like proteases. The EMBO journal 14, 5179-90. Brancolini, C., Bottega, S. and Schneider, C. (1992). Gas2, a growth arrest-specific protein, is a component of the microfilament network system. The Journal of cell biology 117, 1251-61. Brancolini, C., Marzinotto, S. and Schneider, C. (1997). Susceptibility to p53 dependent apoptosis correlates with increased levels of Gas2 and Gas3 proteins. Cell death and differentiation 4, 247-53. Brooks, S. F., Herget, T., Broad, S. and Rozengurt, E. (1992). The expression of 80K/MARCKS, a major substrate of protein kinase C (PKC), is down-regulated through both PKC-dependent and -independent pathways. Effects of bombesin, platelet-derived growth factor, and cAMP. The Journal of biological chemistry 267, 14212-8. Caplen, N. J. and Mousses, S. (2003). Short interfering RNA (siRNA)-mediated RNA interference (RNAi) in human cells. Annals of the New Y ork Academy of Sciences 1002, 56-62. Collavin, L., Buzzai, M., Saccone, S., Bernard, L., Federico, C., DellaValle, G., Brancolini, C. and Schneider, C. (1998). cDNA characterization and chromosome mapping of the human GAS2 gene. Genomics 48, 265-9. Elad, N., Abramovitch, S., Sabanay, H. and Medalia, O. (2011). Microtubule organization in the final stages of cytokinesis as revealed by cryo-electron tomography. Journal of cell science 124, 207-15. Hartwig, J. H., Thelen, M., Rosen, A., Janmey, P. A., Nairn, A. C. and Aderem, A. (1992). Marcks Is an Actin Filament Cross-Linking Protein Regulated by Protein-Kinase-C and Calcium Calmodulin. Nature 356, 618-622. Herget, T., Brooks, S. F., Broad, S. and Rozengurt, E. (1993). Expression of the Major Protein-Kinase-C Substrate, the Acidic 80-Kilodalton

134 Myristoylated Alanine-Rich C-Kinase Substrate, Increases Sharply When Swiss 3t3 Cells Move out of Cycle and Enter G(0). Proceedings of the National Academy of Sciences of the United States of America 90, 2945-2949. Hodges, M., Tissot, C. and Freemont, P. S. (1998). Protein regulation: tag wrestling with relatives of ubiquitin. Current biology : CB 8, R749-52. Jiang, W., Jimenez, G., Wells, N. J., Hope, T. J., Wahl, G. M., Hunter, T. and Fukunaga, R. (1998). PRC1: a human mitotic spindle- associated CDK substrate protein required for cytokinesis. Molecular cell 2, 877- 85. Mandato, C. A. and Bement, W. M. (2001). Contraction and polymerization cooperate to assemble and close actomyosin rings around Xenopus oocyte wounds. The Journal of cell biology 154, 785-97. Mandato, C. A. and Bement, W. M. (2003). Actomyosin transports microtubules and microtubules control actomyosin recruitment during Xenopus oocyte wound healing. Current biology : CB 13, 1096-105. Mandato, C. A., Benink, H. A. and Bement, W. M. (2000). Microtubule-actomyosin interactions in cortical flow and cytokinesis. Cell motility and the cytoskeleton 45, 87-92. Matsudaira, P. (1994). The fimbrin and alpha-actinin footprint on actin. The Journal of cell biology 126, 285-7. Mazzoni, I., Zhang, T., Goetz, J., Wagner, J., Taheri, M., Munger, C., Sacher, M., Gilchrist, A., Au, C., Thibaul, G. et al. (2005). Involvement of Growth-arrest-specific protein 2, a smooth endoplasmic reticulum protein identified by proteomics, in cell cycle control. Canadian Proteomics Initiative Conference. Mollinari, C., Kleman, J. P., Jiang, W., Schoehn, G., Hunter, T. and Margolis, R. L. (2002). PRC1 is a microtubule binding and bundling protein essential to maintain the mitotic spindle midzone. The Journal of cell biology 157, 1175-86.

135 Narayanan, A. and Jacobson, M. P. (2009). Computational studies of protein regulation by post-translational phosphorylation. Current opinion in structural biology 19, 156-63. Orr, G. A., Verdier-Pinard, P., McDaid, H. and Horwitz, S. B. (2003). Mechanisms of Taxol resistance related to microtubules. Oncogene 22, 7280-95. Petroulakis, E., Parsyan, A., Dowling, R. J., LeBacquer, O., Martineau, Y., Bidinosti, M., Larsson, O., Alain, T., Rong, L., Mamane, Y. et al. (2009). p53-dependent translational control of senescence and transformation via 4E-BPs. Cancer cell 16, 439-46. Ren, X. D., Kiosses, W. B. and Schwartz, M. A. (1999). Regulation of the small GTP-binding protein Rho by cell adhesion and the cytoskeleton. The EMBO journal 18, 578-85. Small, J. V., Stradal, T., Vignal, E. and Rottner, K. (2002). The lamellipodium: where motility begins. Trends in cell biology 12, 112-20. Stossel, T. P. (1993). On the crawling of animal cells. Science 260, 1086-94. Thelen, M., Rosen, A., Nairn, A. C. and Aderem, A. (1991). Regulation by phosphorylation of reversible association of a myristoylated protein kinase C substrate with the plasma membrane. Nature 351, 320-2. Waterman-Storer, C. M., Worthylake, R. A., Liu, B. P., Burridge, K. and Salmon, E. D. (1999). Microtubule growth activates Rac1 to promote lamellipodial protrusion in fibroblasts. Nature cell biology 1, 45-50. Wittmann, T., Bokoch, G. M. and Waterman-Storer, C. M. (2003). Regulation of leading edge microtubule and actin dynamics downstream of Rac1. The Journal of cell biology 161, 845-51. Wolter, P., Schmitt, K., Fackler, M., Kremling, H., Probst, L., Hauser, S., Gruss, O. J. and Gaubatz, S. (2012). GAS2L3, a target gene of the DREAM complex, is required for proper cytokinesis and genomic stability. Journal of cell science 125, 2393-406.

136 Zhang, T., Dayanandan, B., Rouiller, I., Lawrence, E. J. and Mandato, C. A. (2011). Growth-arrest-specific protein 2 inhibits cell division in Xenopus embryos. PloS one 6, e24698.

137 138