A ThesisThesis

entitled

Identification and characterization of RhoGAPs involved in the regulation of invadopodia

by

Kyle Lee Snyder

Submitted to the Graduate Faculty as partial fulfillment of the requirements for the

Master of Science Degree in Cellular and Molecular Biology

______Dr. Rafael Garcia-Mata, Committee Chair

______Dr. Deborah Chadee, Committee Member

______Dr. Song-Tao Liu, Committee Member

______Dr. Patricia R. Komuniecki, Dean College of Graduate Studies

The University of Toledo

April, 2016

Copyright 2016, Kyle Lee Snyder

This document is copyrighted material. Under copyright law, no parts of this document may be reproduced without the expressed permission of the author. An Abstract of

Identification and characterization of RhoGAPs involved in the regulation of invadopodia

by

Kyle Lee Snyder

Submitted to the Graduate Faculty as partial fulfillment of the requirements for the Master of Science Degree in Cellular and Molecular Biology

The University of Toledo April, 2016

Invadopodia are actin rich structures that enhance a cancer cells ability to degrade the extracellular matrix (ECM) and promote metastasis. Formation of invadopodia is regulated by Rho , a family of small G that regulate actin rearrangement, cellular migration, and invasion. These proteins exist in two states, inactive GDP-bound, and active GTP-bound conformations. Activation is regulated by

GEFs (guanine nucleotide exchange factors), whereas inactivation is modulated by GAPs

(GTPase activating proteins). In our preliminary studies we screened 18 members of the

RhoGAP family to identify if any were involved in signaling events contributing to invadopodia formation. We identified three candidates, TCGAP, CHN1 and

ARHGAP12. We have confirmed that the knockdown of each of these genes is sufficient to increase invadopodia formation, and validated these results with over expression studies. We have also been able to identify associated Rho proteins for CHN1 and

ARHGAP12. Future work will include further characterization of the activity of these

RhoGAPs. This work adds to the current knowledge available regarding invadopodia and will contribute to future intervention strategies targeting metastatic events.

iii

This body of work is dedicated to those that were lost along the way.

Table of Contents

Abstract iii

Table of Contents v

List of Tables vii

List of Figures viii

List of Abbreviations ix

List of Symbols x

1. Introduction 1

1.1 Cancer and Metastasis 2

1.2 Invadopodia 4

1.3 Rho Family GTPases 6

1.4 RhoGAPs 9

1.5 TCGAP 13

1.6 CHN1 15

1.7 ARHGAP12 18

1.8 Sum159 Cells 21

2. Materials and Methods 23

2.1 General Cell Maintenance 23

2.2 Antibodies and other reagents 23

2.3 Cell lysis and immunoblotting 23

2.4 Plasmids and other reagents 24

2.5 shRNA mediated knockdown 24

2.6 Immunofluorescence microscopy 25

v 2.7 High throughput image acquisition and analysis 25

2.8 Immunofluorescence invadopodia quantification 26

2.9 Preparation of cDNA 26

2.10 qRT-PCR 27

2.11 Cell transfection 29

2.12 Rho GTPase activity assay 29

2.13 RhoGAP-GTPase binding assay 30

3. Results 32

3.1 RhoGAP Screen 32

3.2 TCGAP 39

3.3 CHN1 43

3.4 ARHGAP12 50

4. Discussion 54

4.1 RhoGAP Screen 54

4.1.1 Future Aims 55

4.2 TCGAP 56

4.2.1 Future Aims 58

4.3 CHN1 60

4.3.1 Future Aims 62

4.4 ARHGAP12 63

4.4.1 Future Aims 65

References 66

vi List of Tables

Table 1 Primers used for qRT-PCR analysis of gene knockdown. Forward (sense)

and Reverse (anti-sense) used to amplify a 100-150 bp region of the targeted

gene...... 28

Table 2 A list representing the RhoGAPs assessed in the initial screening performed.

18 GAPs were chosen, numbered in the left hand column. The middle

column contains a reference number from the GAP shRNA library used and

the right column identifies the RhoGAP tested by the most common

reference name...... 33

vii List of Figures

Figure 1 Rho GTPase Cycle ...... 7

Figure 2 The RhoGAP Family ...... 11

Figure 3 TCGAP domain structure ...... 14

Figure 4 CHN1 domain structure ...... 17

Figure 5 ARHGAP12 domain structure ...... 20

Figure 6 96 well plate layout ...... 35

Figure 7 GAP screen ...... 37

Figure 8 RhoGAP screen (images) ...... 38

Figure 9 TCGAP is involved in invadopodia formation ...... 40

Figure 10 TCGAP knock down increases RhoG activity...... 42

Figure 11 CHN1 is involved in invadopodia formation ...... 44

Figure 12 Punctate invadopodia form in CHN1 knock down ...... 45

Figure 13 CHN1 suppresses invadopodia and binds Rac1, RhoG ...... 47

Figure 14 CHN1 knock down increases GTPase activity ...... 49

Figure 15 ARHGAP12 is involved in invadopodia formation ...... 51

Figure 16 ARHGAP12 suppresses invadopodia and binds Rac1...... 53

viii List of Abbreviations aa ...... amino acid bp(s) ...... base pairs cDNA ...... Complementary DNA DAG ...... Diacylglycerol DAPI ...... 4',6-Diamidino-2-Phenylindole, Dihydrochloride DMEM ...... Dulbecco’s Modified Eagle Medium DNA ...... deoxyribonucleic acid ECM ...... Extracellular Matrix FBS ...... Fetal Bovine Serum FITC ...... Fluorescein isothiocyanate GAPs ...... GTPase Activating Proteins GDI ...... GDP-dissociation inhibitor GEFs ...... Guanine Nucleotide Exchange Factor GST ...... Glutathione S-Transferase HGF...... Hepatocyte Growth Factor HRP ...... Horseradish Peroxidase KD ...... Knock Down kDa ...... kilodalton MMPs ...... Matrix Metalloproteases mRNA ...... Messenger RNA N-WASP ...... Neuronal Wiskott-Aldrich Syndrome P/S ...... Penicillin/Streptomycin PBD ...... PAK Binding Domain PBS ...... Phosphate Buffered Saline PDBu ...... Phorbol 12,13– dibutyrate PH ...... Pleckstrin Homology PRR ...... Proline rich region(s) PX ...... Phox Homology Q61L ...... glutamine to leucine substitution at amino acid residue 61 Q63L ...... glutamine to leucine substitution at amino acid residue 63 qRT-PCR...... quantitative real time polymerase chain reaction Rho ...... Ras Homology RNA ...... Ribonucleic Acid ROCK ...... Rho-Associated Protein Kinase ROS ...... Reactive Oxygen Species RSV-CEF ...... Rous Sarcoma Virus Transformed Chicken Embryo Fibroblasts SH2 ...... Src Homology 2 SH3 ...... Src Homology 3 shRNA...... Short Hairpin RNA TKS4 ...... Tyrosine Kinase Substrate with four Src homology 3 domains TKS5 ...... Tyrosine Kinase Substrate with five Src homology 3 domains TNBC ...... Triple Negative Breast Cancer(s) TRITC ...... Tetramethylrhodamine WW ...... Domain containing two conserved Tryptophan (W) amino acids

ix List of Symbols

º ...... degree C ...... Celsius

α ...... alpha β ...... beta

µ ...... Mu [micro]

x Chapter One

1. Introduction

Understanding the molecular mechanisms controlling carcinogenic cell invasion has the potential to reveal novel targets for therapeutic intervention. Cancer cells invade other tissues and enter the bloodstream by forming actin rich membrane protrusions called invadopodia. Invadopodia enhance the cells’ ability to degrade the extracellular matrix (ECM) and promote metastasis (Baldassarre et al., 2006). Formation of invadopodia is regulated by Rho GTPases, a family of small G proteins that has roles involving actin rearrangement, cellular migration, and invasion (Ridley, 2015). These proteins exist in two states; inactive GDP-bound, and active GTP-bound conformations.

Activation is regulated by RhoGEFs (Guanine nucleotide exchange factors) and inactivation is modulated by RhoGAPs (GTPase activating proteins) (Bishop and Hall,

2000). Interestingly, only a small number of GAPs for the Rho GTPases have been linked to cell invasion (Campa and Randazzo, 2008; Kusama et al., 2006). The RhoGAP family is comprised of 67 proteins, and since Rho GTPases are widely involved in cell invasion it is likely that there are many more invasion-related RhoGAPs. The goal of this project is to identify novel RhoGAPs that have a functional role in this cellular process. Providing a clear picture as to how cells are fundamentally able to regulate invasion and associated structures has the potential to reveal additional therapeutic approaches or targets for metastatic prevention therapies.

1 1.1 Cancer and Metastasis

Cancer is a prevalent disease in western society. Roughly one half of males in the united states will be diagnosed with some form of cancer in their lifetime, and of those, one in four will die as a result. For women, one in three will be affected and roughly one in five will succumb to the disease (Howlader N, 2015). According to data collected and analyzed between 1977 and 2012, breast cancer accounts for nearly one third of all female cancer diagnosis in the United States (Howlader N, 2015).

The majority of cancer deaths are not caused by the primary tumor in a patient but by the formation of secondary tumor sites throughout the body. The process of cancer cells leaving the primary tumor site, entering the blood stream (intravasation), subsequently exiting the blood stream (extravasation) and colonizing secondary tumor sites is known as metastasis (Fidler and Kripke, 1977). Physical removal of the tumor is only an effective treatment if spreading has not occurred, thus treatment modalities are beginning to focus on prevention at the intravasation and extravasation steps in the metastatic process. There are still numerous unknowns surrounding cancer as a disease process which demonstrates the need for further basic research as well as continuing to explore therapeutic options.

As an example of what remains to be fully characterized, it is uncertain how or when tumor cells are able to realize their metastatic potential and begin breaking/moving away from the primary tumor. One long standing and widely held theory put forth by

Fidler & Kripke suggests that within the population of cells comprising a tumor mass resides a small subset of cells which have acquired (through genetic alterations over time) an increased ability to metastasize (Fidler and Kripke, 1977). The implication is that in

2 early stages of tumor growth there is cellular uniformity throughout the tumor. However, over time (in the later stages of tumorigenesis) a small and highly metastatic subset of tumor cells arise and the tumor shifts from a homogenous to a more heterogeneous state

(Fidler and Kripke, 1977). More recently, however, this model of thinking has been challenged. Deoxyribonucleic Acid (DNA)-microarray studies indicated that metastasized cancer cells could be distinguished from the cells in the primary site by their gene-expression profile alone. This suggests that cellular diversity is seen in the early stages of tumorigenesis as opposed to genetic alterations occurring slowly over long periods of time (Bernards and Weinberg, 2002). This second theory implies that some cells have gain of function genetic mutations selecting for increased metastasis and increased growth ability around the time of tumorigenesis as opposed to acquiring metastatic potential at a later time. It remains unknown whether the metastatic ability of a tumor is an innate feature of the cells comprising the primary tumor site or if it is an evolutionarily acquired trait (Weigelt et al., 2005). Further basic research is necessary to provide additional insight into this disease process and allow researchers to attack the disease in new ways.

There are many difficulties when it comes to treating cancer or developing new treatment therapies. One such difficulty lies in the lack of information on regulatory signaling pathways surrounding this cellular event. Our work is centered on identifying previously unexplored signaling pathways and uncovering the roles they have in regulating cellular structures involved in metastatic events. The following research begins to explore the RhoGAP family and its involvement in the regulation of invadopodia formation. Elucidating the cellular signaling processes that have a role in metastasis will

3 provide a clearer picture to the scientific community on how cancer cells are able to spread throughout the body.

1.2 Invadopodia

Invadopodia are actin-rich structures present in carcinogenic cells which are capable of degrading the ECM to promote metastasis (Stoletov and Lewis, 2015). ECM degradation provides a route for the cell to invade other cellular layers, hence the term invadopodia. In the early years of invadopodia research some confusion was created as to the proper nomenclature for this cell structure. Invadopodia as we know them today were first seen in 1980 as a cluster of cell structures, containing α-actinin and vinculin, in a ring on the cell and given the name ‘rosettes’ (David-Pfeuty and Singer, 1980). In 1985 it was noted that these cellular structures protruded from the ventral side of the cell and were sites of adhesion to the ECM (Tarone et al., 1985). Finally, in 1989 Chen discovered that these contact sites were locations of ECM degradation and coined the term invadopodia. However, the term podosome is still used today and it is currently accepted that in non-carcinogenic cells the degrading structures present are referred to as podosomes and in cancer cells they are referred to as invadopodia (Murphy and

Courtneidge, 2011).

Substrate degrading structures were first described by Chen and colleagues in

1984 in Rous sarcoma virus-transformed cells (Chen et al., 1984). These structures formed on the ventral side in Rous sarcoma virus transformed chicken embryonic fibroblasts (RSV-CEF) and were later named “invadopodia” (Chen, 1989). Chen noted in his observations that “rosette shaped structures” formed that were actin-rich and contained numerous actin-associated proteins (Chen, 1989). These structures were

4 capable of degrading fibronectin, a function that was key to setting the new structures

(invadopodia) apart from other previously identified adhesion structures (Chen, 1989). It was also determined that matrix metalloproteases (MMPs) were responsible for the degradation associated with invadopodia (Chen, 1989). After the initial discovery of these degradative structures it was determined that β1 integrins localized to invadopodia

(Mueller and Chen, 1991). The following year the same group determined that cellular sites positive for invadopodia also demonstrated a drastic increase in the amount of tyrosine-phosphorylated proteins (Mueller et al., 1992). Over time several additional proteins were found to be associated with invadopodia including cortactin, paxillin and

PKCµ among others (Ayala et al., 2006; Bowden et al., 1999). In the past 15 years, a plethora information involving the structure and function of invadopodia has been revealed. It is now known that actin regulators are crucial to invadopodia function, namely cortactin and neural Wiscott Aldrich syndrome protein (N-WASP) (Murphy and

Courtneidge, 2011). It has also been established that other adaptor proteins are necessary, including Tyr kinase substrate with four SH3 domains (TKS4) and Tyr kinase substrate with five SH3 domains (TKS5) (Murphy and Courtneidge, 2011).

As mentioned previously, cancer is a pervasive disease that becomes extremely difficult to treat if metastasis occurs. As such, a plethora of research is being done in an attempt to find ways to slow or even prevent this from happening. It has even been shown that invadopodia are necessary for cancer cell extravasation, a process required for the formation of secondary tumor sites (Leong et al., 2014). Currently there are no treatments that specifically target cancer at the metastatic level (Stoletov and Lewis, 2015). The reason for this is twofold: Difficulty in treating cancer and only affecting carcinogenic

5 cells while leaving normal human tissues unharmed, and secondly finding cellular targets that are suitable intervention therapies. This research is centered on identifying previously unknown regulatory proteins that are involved in the formation or regulation of invadopodia.

1.3 Rho Family GTPases

The Rho GTPase family was first identified in 1985 (Madaule and Axel, 1985). It comprises a subfamily of the larger Ras superfamily of proteins. These proteins, along with other subfamilies of the Ras superfamily, are identified as ‘small’ GTPases due to their size. Small GTPases are known for being between 20 and 30 kilodaltons (kDa)

(Bishop and Hall, 2000). The name Rho is derived from this larger family and is short for

Ras homolog(y) due to the similarities between the two groups (Madaule and Axel,

1985). Today it is known that the Rho GTPase family of proteins consists of 22 closely related small G proteins.

Rho GTPases function as molecular switches, alternating between an “ON” and an “OFF” state (Hall and Nobes, 2000). A Rho GTPase is considered “ON” when bound to GTP, and “OFF” when bound to GDP (Hall and Nobes, 2000). In the GTP-bound state this family of proteins is able to bind to downstream effector proteins and elicit a cellular response. The conversion from GDP to GTP is facilitated by a family of proteins known as GEFs (Hall and Nobes, 2000). Conversely GAPs are involved in the hydrolysis of bound GTP to GDP (Hall and Nobes, 2000). There is a third type of protein involved in this cycle which is capable of binding Rho proteins and sequestering them in their GDP- bound state, known as the Rho protein GDP dissociation inhibitor (Rho GDI) (Olofsson,

1999). This cycle is represented visually in Figure 1.

6

One of the primary roles of Rho GTPases in the cell is in the regulation of the actin cytoskeleton, a cellular component that has numerous functions within the cell including cytokinesis, membrane trafficking, cell cycle progression and transcriptional regulation (Hall, 1990; Olson et al., 1995). The actin cytoskeleton is the framework within a cell that allows it to hold its shape and move in response to internal/external stimuli (Tapon and Hall, 1997). Also, the actin cytoskeleton is ever changing, either

7 expanding or retracting in response to cellular signals via protein-protein interactions and regulated actin polymerization (Ozaki and Hatano, 1984).

Of the Rho family GTPases, the three best characterized members are RhoA,

Rac1 and Cdc42. Each of these proteins have been linked to pivotal functions involving the cytoskeleton. Rac1 is important for cellular movement as well as membrane ruffling,

RhoA is capable of regulating the assembly of focal adhesions and actin stress fibers, and

Cdc42 as well as regulating filopodia and cellular polarity, but also has lesser known functions involving invadopodia formation (Furmaniak-Kazmierczak et al., 2007;

Nakahara et al., 2003; Ridley and Hall, 1992). As stated previously, invadopodia formation and regulation is centered on actin-cytoskeletal rearrangement and the importance of the Rho GTPase family in this process makes them and their regulatory proteins (GEFs, GAPs and RhoGDI) ideal candidates for study.

Much of the current knowledge and research being performed is centered on Rho

GTPase activation and their downstream effector proteins, yet comparatively less is known about GTPase inactivation (Wertheimer et al., 2012). For Cdc42 alone, a positive regulator of invadopodia formation, “more than twenty downstream effectors [have been identified], including protein kinases, lipid kinases, scaffolding proteins, and cytoskeletal interacting proteins, resulting in changes in cellular processes including cell polarity, adhesion, migration, proliferation, actin cytoskeleton remodeling, membrane trafficking and transcription” (Arias-Romero and Chernoff, 2013). As many as twenty three GEFs for Cdc42 are identified and only seven GAPs (Arias-Romero and Chernoff, 2013). In order to further our understanding of how the Rho GTPases are involved in carcinogenic cellular processes it is important to have a full picture of the proteins involved in both

8 activation and inactivation of Rho GTPase signaling events. Not only could a clearer picture provide additional therapeutic targets for carcinogenic intervention, it may also reveal entirely new methods of approaching the problem of metastasis.

1.4 RhoGAPs

The RhoGAP gene family is able to promote the hydrolysis of Rho-bound GTP to

GDP, thus inactivating Rho proteins. This is because RhoGAPs stabilize the Rho-GTP structure in a conformation that promotes GTP hydrolysis (Scheffzek et al., 1997). There is also evidence that shows GAPs can induce a GDP-like charge distribution once bound to Rho-GTP molecules, an effect that is likely responsible for inducing conformational stability (Kötting and Gerwert, 2004). The particular protein binding domain that is responsible for this event was discovered in 1993 and subsequently identified as the

RhoGAP domain (Zheng et al., 1993). This domain comprises 170 amino acids and is conserved across species, from yeast to mammals (Tcherkezian and Lamarche-Vane,

2007). Evolutionarily conserved cellular signaling control mechanisms are ideal candidates for study as dysregulation generally results in pronounced aberrant cellular behaviors such as carcinogenic growth or cell death.

Interestingly, there are more than three times as many RhoGAPs as there are actual Rho proteins (Tcherkezian and Lamarche-Vane, 2007). There are sixty-seven

RhoGAPs in the human genome and although they all share a RhoGAP domain there is a great deal of variety within the family (Tcherkezian and Lamarche-Vane, 2007). No single theory has been able to explain the reason for the excess of RhoGAPs compared to

Rho GTPases, however multiple explanations have been proposed.

9 Firstly, it has been shown that not all RhoGAPs are ubiquitously expressed. The

RhoGAP Grit, for example, has been found to have only brain tissue expression

(Nakamura et al., 2002). It is also speculated that expression could vary over the life span of a cell: i.e. RhoGAPs necessary during development (neurogenesis/angiogenesis for example) may not continue to be expressed in later stages of the life span (Tcherkezian and Lamarche-Vane, 2007). Thus, selective gene expression and varying roles over the life time of a cell may contribute to the excess of RhoGAPs compared to Rho GTPases. It is also possible that RhoGAPs are able to selectively repress Rho GTPase activity in the cell depending on their subcellular localization, in a concentration dependent manner or in some other fashion (Tcherkezian and Lamarche-Vane, 2007). This level of regulation would require a high degree of specificity resulting in a wide variety of RhoGAPs. One last point is made that the GAP domain of some of the RhoGAP family proteins may not be active, meaning it would not cause inactivation but can still bind and serve other functions, such as scaffolding, localization or preventing function by sequestering Rho

GTPases (Chiang et al., 2003; Kozma et al., 1996). The implication is that RhoGAPs may have a multitude of roles within the cell and in order to properly modulate Rho GTPase function a high degree of specificity is required, thus necessitating a wide array of

RhoGAPs.

10

11 When beginning to look into the cellular signaling events that lead to Rho GTPase inactivation things quickly become convoluted. For example, FilGAP is a RhoGAP that is capable of inactivating Rac1 (Ohta et al., 2006). However, FilGAP itself is activated via phosphorylation by Rho kinase (ROCK), which is an effector for RhoA (Ohta et al.,

2006). Thus, when RhoA is active ROCK is stimulated to phosphorylate FilGAP leading to Rac1 inactivation. Conversely, Rac1 activity is able to affect RhoA activity in a similar fashion (Nimnual et al., 2003). Through generation of reactive oxygen species (ROS)

Rac1 is able to stimulate the GAP ability of p190RhoGAP via tyrosine phosphorylation

(Nimnual et al., 2003). This GAP activation results in a decrease in the amount of active

RhoA (Nimnual et al., 2003). Thus, when Rac1 is active RhoA is being suppressed and vice versa. This high degree of crosstalk between Rho GTPases, effectors, GEFs and

GAPs make Rho GTPase signaling events seem less like directional occurrences and more like a web of signaling nodes that can be activated or inactivated depending on cell type, subcellular location or other proteins involved.

As if intertwined signaling events were not complicated enough, at least one

RhoGAP is capable of altering its Rho GTPase specificity at different points in the cell cycle. MgcRacGAP was originally identified as a RhoGAP with specificity for Rac1 and

Cdc42 (Toure et al., 1998). It was also demonstrated by several groups that this RhoGAP did not have GAP activity toward RhoA (Kawashima et al., 2000; Raymond et al., 2001).

Despite having no GAP activity toward RhoA, it was shown that during cytokinesis,

MgcRacGAP co-localized with RhoA and a protein called Aurora B (Minoshima et al.,

2003). Furthermore, it was suggested that all three proteins were physically interacting with each other. Evidently, Aurora B can phosphorylate MgcRacGAP at Ser387 which

12 then allows for latent GAP activity toward RhoA, simultaneously suppressing the

“normal” Rac1 and Cdc42 GAP specificity of MgcRacGAP specifically during cytokinesis and not at other points in the cell cycle (Minoshima et al., 2003).

It is rather apparent that the study of the RhoGAP family is not as straight forward as originally believed. There are GAPs containing inactive GAP domains, GAPs that can change their function and specificity based on interaction with other proteins and GAPs that are intertwined in auto-regulatory signaling pathways which are modulated by closely related Rho GTPases. Perhaps the most challenging aspect is that most of these functions and specialty GAP actions can vary from cell type to cell type with differing protein expression patterns.

1.5 TCGAP

TCGAP is a RhoGAP that was first characterized in 2003 as having a role in insulin regulation in adipocytes (Chiang et al., 2003). There are multiple names which refer to TCGAP including Neurite Outgrowth Multi-Adaptor-GAP (NOMA-GAP),

ARHGAP33 and Sortin nexin 26 (SNX26) (Kim et al., 2013). For the purpose of this work TCGAP will be used as this is the name given upon initial characterization (Chiang et al., 2003). This differentiation in nomenclature is partly due to the fact that not much research has been performed on this particular RhoGAP. As of this writing there are only eleven scientific journal articles which discuss TCGAP.

In neuronal cells TCGAP is expressed at a higher level than in the rest of the body

(Liu et al., 2006). Despite being characterized in adipocytes, the brain enrichment of this protein has focused the majority of the research performed on TCGAP to be involved

13 with neurite outgrowth, dendritic growth, synapse development and other neuron specific functions (Rosario et al., 2007; Schuster et al., 2015; Shen et al., 2011; Simo and Cooper,

2012).

As noted in Figure 3, TCGAP is a 137 kDa protein. It has been shown that

TCGAP has GAP activity specifically directed toward TC10β, a lesser studied Rho

GTPase, and Cdc42, although it is possible other Rho GTPases interact as not all of the

22 Rho GTPases have been tested (Chiang et al., 2003). In fact, TC10β and Cdc42 are the namesake of the ‘TC’ in TCGAP. The N-terminal half of TCGAP contains three binding domains including a Phox homology (PX) domain, a SRC Homology 3 (SH3) domain and a RhoGAP domain listed sequentially from the first amino acid (Chiang et al., 2003).

The C-terminal portion of this RhoGAP is devoid of specific binding domains yet contains as many as nine proline-rich regions (PRRs) (Chiang et al., 2003).

The PX domain, found at the N-terminus of TCGAP, is also found in the Sortin nexin family of proteins, which is the reason SNX26 is an alternate name (Kim et al.,

2013). The PX domain was characterized in 1996 and it was later discovered that the

14 primary role for this domain is subcellular targeting of proteins to the plasma membrane

(Ponting, 1996; Wishart et al., 2001). The SH3 domain, located between the PX and

RhoGAP domains, contributes to the specificity of protein-protein interactions (Pawson and Schlessingert, 1993). It is known that PRRs, which in this case are located on the C- terminal portion of the protein, are capable of binding SH3 domains (Alexandropoulos et al., 1995). These three binding regions on TCGAP are responsible for its localization, protein-protein interactions and the folded structure of the protein, all of which directly affect its function within the cell.

As mentioned previously, it is known that the Rho GTPase Cdc42 is a positive regulator of invadopodia formation: thus, with an increase in Cdc42 activity there is a corresponding increase in invadopodia formation (Qadir et al., 2015; Yamaguchi et al.,

2005). Based on our preliminary results, we hypothesized that TCGAP has a suppressive role during invadopodia formation or total invadopodia lifetime by acting through Cdc42.

Inactivation of Cdc42 by TCGAP at invadopodia would result in invadopodia disassembly, and shorter invadopodia lifetimes would result in decreased ECM degradation. Currently, there is little information available on TCGAP function and no roles have been established in carcinogenic cells. Our research provides evidence that

TCGAP does have a functional role in breast cancer.

1.6 CHN1

The chimaerin family of RhoGAPs has been shown to be enriched in the brain and testes (Hall et al., 1993; Toure et al., 1998). This family was first characterized as a

RhoGAP in 1990, the first member being called η-chimaerin or neuronal chimaerin

(Ahmed et al., 1990; Hall et al., 1990). Since then four closely related family members

15 have been identified: two isoforms of chimaerin1 (CHN1) known as α1-chimaerin and

α2-chimaerin, and two isoforms of chimaerin2 (CHN2) called β1-chimaerin and β2- chimaerin (Yang and Kazanietz, 2007). Expression of the α2 and β2 isoforms have been detected in a wide range of cell types, including T-lymphocytes and COS-1 cells

(fibroblast-like cells derived from monkey ), whereas the α1 and β1 forms are currently known to only be expressed in the brain and testes (Caloca et al., 2003; Siliceo et al., 2006; Yang and Kazanietz, 2007).

The four chimaerin isoforms originate from two distinct gene loci. CHN1 isoforms (α1 and α2) are at chromosome locus 2q31-32.1 and the CHN2 isoforms (β1 and β2) at chromosome locus 7p15.3 (Hall et al., 1993). Each of these locations contain two promoter sites that can be selectively activated to produce similar but distinct protein products (Hall et al., 1993). This means the cell can either begin transcription at the first site, resulting in a longer protein product, or the second promoter site resulting in a shorter protein product. The α1 and β1 variants contain only two domains, the C1 and

RhoGAP domains (Yang and Kazanietz, 2007). Conversely, the α2 and β2 variants are longer than their counter parts and contain an SH2 domain in addition to the C1 and

RhoGAP domains (Yang and Kazanietz, 2007). A representation of α2-chimaerin domain structure, the chimaerin of interest in our research, can be seen in figure 4.

16

Current knowledge of the chimaerin RhoGAP family provides evidence that the

GAP domain is specific for Rac (Diekmann et al., 1991; Hall et al., 1990; Hall et al.,

1993). However, only the more common GTPases (Rho, Rac, and Cdc42) have been assessed thus far and our research suggests a second interacting GTPase (Unpublished

Data). The chimaerin family is unique in that it is the only known RhoGAP that is capable of binding phorbol esters and diacylglycerol (DAG), which is accomplished via interaction with the C1 domain (Kazanietz, 2000; Kazanietz, 2002; Yang and Kazanietz,

2003). Interaction with tumor promotors like phorbol esters could imply an involvement in tumor progression and invadopodia formation. The SH2 domain (Src Homology 2) is similar in function to the aforementioned SH3 domain. As such, it is involved in specifying protein-protein interaction and is also capable of binding to platelet-derived growth factor (PDGF), epidermal growth factor (EGF) receptors, as well as receptor auto- phosphorylation sites within the cell (Pawson and Schlessingert, 1993).

It is known that Rac1 function is essential for proper cell motility. Previous work has shown that Rac1 is involved in invadopodia regulation in several cell lines, including

MDA-MB231 and MCF10A lines which are classified as triple negative breast cancer

17 (TNBC) cells (Chavez et al., 2010; Harper et al., 2010; Lin et al., 2014; Pignatelli et al.,

2012). Recently, work from our lab has also identified a role for Rac1 involvement in the formation/regulation of invadopodia in SUM159PT breast cancer cells, another TNBC cell line (Unpublished Data). We have also implicated CHN1 in this process. We hypothesize that CHN1 is capable of regulating the activity state of Rac1, decreasing the amount of GTP-Rac1 thus not allowing invadopodia to form. This work expands upon the known role of Rac1 in cancer and provides insight into a signaling event contributing to invadopodia regulation.

1.7 ARHGAP12

We initially used a candidate approach to determine the role of three ARHGAP proteins during invadopodia formation, ARHGAP1 and ARHGAP10 in addition to

ARHGAP12. However, neither of the first two produced a notable phenotype when assayed, thus ARHGAP12 is the third and final RhoGAP studied in this project. This

RhoGAP has been shown to be ubiquitously expressed in a variety of tumor cell lines as well as normal tissues (Tcherkezian and Lamarche-Vane, 2007).

Little work has been produced on ARHGAP12 since it was first characterized in

2002 (Zhang et al., 2002). However, six years after its initial characterization evidence showed ARHGAP12 is a GAP toward Rac1 with weak GAP activity toward Cdc42 and no activity towards RhoA (Gentile et al., 2008). This same work revealed that hepatocyte growth factor (HGF) is capable of suppressing both the mRNA (messenger Ribonucleic

Acid) and protein expression levels of ARHGAP12 in MLP-29 cells (Gentile et al.,

2008). It was also recently discovered that ARHGAP12 has functions in the nervous system and is capable of “coordinate[ing] dendritic spine morphology and synaptic

18 strength via its GAP activity and interaction with CIP4” (Ba et al., 2016). ARHGAP12 was also shown to suppress Rac/ Cdc42 activity in RAW 264.7 macrophages (Schlam et al., 2015). ARHGAP12 localizes to the phagocytic cup and its presumed GAP activity toward Rac/Cdc42 at that location allows the process of phagocytosis to complete

(Schlam et al., 2015).

ARHGAP12 has two similar isoforms known as ARHGAP12a and ARHGAP12b

(Zhang et al., 2002). The ‘a’ isoform contains eighteen exons while ‘b’ contains twenty

(Zhang et al., 2002). The nucleotide sequence of the two isoforms are identical with the exception that ARHGAP12b has an extra 90 bp making up exons eight and nine. This

‘extra region’ starts at bp 1569 and is located between the second WW domain and the

PH domain (Zhang et al., 2002).

Figure 5 shows the domain structure of ARHGAP12b (Zhang et al., 2002). This

RhoGAP contains five domains; from the N-terminus these are the SH3 domain, followed by two WW domains, a PH and a RhoGAP domain at the C-terminus (Zhang et al., 2002). The general characteristics of the SH3 domain and the RhoGAP domain

(ARHGAP12 being specific for Rac1 with possible binding toward Cdc42) were described previously for TCGAP.

19

The WW domain, initially discovered in 1994, is made up of 40 amino acids (aa) and is named after the consensus motif that makes up the domain: two conserved tryptophans (W) spaced 20-22 aa apart (Bork and Sudol, 1994). It was then noted that the

WW domain is capable of modulating protein-protein interactions, similar to but distinct from the SH3 domain, through interactions with PRRs (Sudol et al., 1995). The

Pleckstrin Homology (PH) domain was first characterized in 1993 and, like the WW domain, was found to have similar to but separate functions from the SH2 and SH3 domains (Haslam et al., 1993; Shaw, 1993). PH domains are best known for binding to phosphoinositides phosphorylated at different sites (Saraste and Hyvonen, 1995).

In this project, the significance of ARHGAP12 is similar to that of CHN1. Since previous work has shown ARHGAP12 also having GAP activity toward Rac, it is likely that these two RhoGAPs are working in a similar manner to control Rac activity during invadopodia formation or regulation. It is unclear whether these two RhoGAPs function in conjunction with one another or act independently. As with CHN1, we hypothesize

20 that ARHGAP12 is capable of suppressing Rac activity which in turn suppresses invadopodia formation in SUM159 cells.

1.8 SUM159 Cells

There are many different breast cancer cell lines that are used for research purposes today. These cell lines are considered ‘immortalized’ as they can grow and divide indefinitely (Fillmore and Kuperwasser, 2008). They differ from one another based on the presence of cell surface receptors and overall metastatic capability (Fillmore and Kuperwasser, 2008). It has also been shown that different breast cancer cell lines have differing protein expression patterns (aside from cell surface receptors), determined by analysis of the transcription profiles of 51 unique breast cancer cell lines (Neve et al.,

2006).

The SUM159 (SUM-159PT) cell line is an epithelial breast cancer cell line that was first characterized in 1999 (Flanagan et al., 1999). There are three main classifications of breast cancers. Most luminal breast cancers express estrogen and progesterone receptors (Chevalier et al., 2015). HER2-positive breast cancers express

ERBB2, a tyrosine kinase receptor (Chevalier et al., 2015). SUM159 cells are classified as TNBCs, meaning that this cell line lacks three proteins; the aforementioned estrogen and progesterone receptors as well as the Her2/Neu receptor

(Chevalier et al., 2015). SUM159 cells are highly metastatic, which is a notable characteristic of all triple negative breast cancers (Chevalier et al., 2015; Flanagan et al.,

1999). As these cells are estrogen independent (noted by their lack of the estrogen receptor) they are considered a good model system for late-stage invasive breast cancers

(Flanagan et al., 1999).

21 Because our studies focus on cellular structures involved in metastatic events, we chose SUM159 cells as our model system. We also chose this cell line because little information is available on the treatment of TN breast cancers and because “TNBCs are the most aggressive and have the worst prognosis due to the lack of specific therapies”

(Chevalier et al., 2015). The overall goal of this line of research is to identify and characterize the cellular signaling pathways active in breast cancer that promote or enhance cellular metastasis. GAPs have previously been identified as potential targets for therapeutic intervention; as such, we will provide evidence which identifies three new target RhoGAPs for use in treatment modalities targeting metastatic signaling pathways in breast cancer (Vigil et al., 2010).

22 Chapter Two

2. Materials and Methods

2.1 General Cell Maintenance

All cells used during this body of work were cultured at 37° Celsius (C) with 5% carbon dioxide (CO2) supplied. The SUM159 cells (a gift from Dr. Carol Otey, UNC

Chapel Hill) were cultured in Ham's F12 with 10% fetal bovine serum (FBS), 5 μg/ml insulin, 1 μg/ml hydrocortisone and antibiotics (penicillin/streptomycin, (P/S)).

HEK293FT cells were cultured in Dulbecco’s Modification of Eagle’s Medium (DMEM) with 4.5 g/L glucose, L-glutamine and sodium pyruvate supplemented with 10% FBS and

P/S.

2.2. Antibodies and other reagents

The following antibodies were used: myc (9E10) (Santa Cruz, Santa Cruz, CA); tubulin alpha (Sigma, St. Louis, MO); cortactin (Santa Cruz, Santa Cruz, CA, USA);

Rac1 (BD Biosciences, San Jose, CA). Alexa fluor 488- and Alexa fluor 594-conjugated anti-mouse IgG and anti-rabbit IgG secondary antibodies and Alexa 488-conjugated

Phalloidin (Life Technologies, Carlsbad, CA). Horseradish peroxidase (HRP)-conjugated anti-mouse and anti-rabbit secondary antibodies (Jackson Immunoresearch, West Grove,

PA). Phorbol-12,13-dibutyrate (PDBu) (Sigma, St. Louis, MO).

2.3. Cell lysis and immunoblotting

Cells cultured on 100 mm tissue culture dishes were briefly rinsed with phosphate buffered saline (PBS) and then scraped into a lysis buffer containing 50 mM Tris/HCl, pH 7.4, 10 mM MgCl2, 150 mM NaCl, 1% Triton X-100, and EZBlock Protease

23 Inhibitor Cocktail (BioVision, Mipitas, CA). The supernatant was collected after centrifugation at 14,000 rpm for 10 min. For immunoblotting, lysates were boiled in 2X

Laemmli buffer, and 20 μg of protein were resolved by sodium dodecyl sulfate– polyacrylamide gel electrophoresis in each lane of a 13% gel. The proteins were transferred to nitrocellulose and immuno-blotted. Immunocomplexes were visualized using the Immobilon Western Millipore Chemiluminescence HRP substrate (Millipore,

Billerica, MA).

2.4 Plasmids and other reagents

The glutathione S-transferase (GST) fusion constructs encoding Q61L-Rac1,

Q61L-Cdc42, Q63L-RhoA, and Q61L-RhoG have been described elsewhere (Ellerbroek et al., 2004). Generation of eukaryotic expression vectors of the myc-tagged RhoGAPs

(TCGAP, ARHGAP12 and Chimaerin 1) was done using Gateway recombination technology following the manufacturer’s instructions (Life Technologies. Carlsbad, CA).

2.5 shRNA mediated knock down

pLKO lentiviral short hairpin RNA (shRNA)mir (control non-silencing, human

Chimaerin 1, human TCGAP, and human ARHGAP12 shRNA) vectors were obtained from Open Biosystems (Huntsville, AL, USA). Lentiviruses were prepared at the Lenti- shRNA Core Facility, University of North Carolina, Chapel Hill, North Carolina. Cells were infected with lentivirus particles overnight. The following day, the infection media was removed and replaced with complete medium containing puromycin (2.5 µg/ml) to select for shRNA expressing cells and total cell lysates were subjected to Western blot analysis for protein expression as described.

24 2.6 Immunofluorescence microscopy

Cells grown on coverslips were washed in phosphate-buffered saline (PBS), fixed in 3.7% paraformaldehyde for 10 min, and quenched with 10 mM ammonium chloride.

Cells were permeabilized with 0.1% Triton X-100 in PBS. The coverslips were then washed with PBS and blocked in PBS, 2.5% goat serum, 0.2% Tween 20 for 5 min followed by blocking in PBS, 0.4% fish skin gelatin, and 0.2% Tween 20. Cells were incubated with primary antibody for 1h at room temperature. Coverslips were washed with PBS, 0.2% Tween 20 and incubated with secondary antibodies for 45 min.

Coverslips were washed as described above and mounted on glass slides using MOWIOL mounting media.

2.7 High throughput image acquisition and analysis

ImageXpress® Micro system from Molecular Devises (Sunnyvale, CA) was used to perform the screening for RhoGAPs involved in the regulation of invadopodia. This microscope is capable of high throughput automated immunofluorescent imaging. The immunofluorescence protocol described previously was used with the only change being that the cells were treated/fixed/etc. within a 96 well plate instead of on glass coverslips.

Four sets of images were captured per well using a 40X ELWD objective. The image sets consisted of a fluorescein isothiocyanate (FITC), tetramethylrhodamine (TRITC) and

4',6-Diamidino-2-Phenylindole, Dihydrochloride (DAPI) color channels which were then merged into one image. The number of cells within the image as well as the number of cells identified as positive for invadopodia was recorded manually.

25 2.8 Immunofluorescence Invadopodia quantification

Co-localization of actin (green) and cortactin (red) resulting in a yellow color not overlapping with the location of a lamellipodia was noted as a positive indicator for invadopodia. Total cells counted was recorded as well as the number of cells expressing invadopodia. From these numbers the percentage of cells showing invadopodia was calculated.

Outside of the screening process cells were either viewed and counted at the microscope or images were obtained. An Olympus IX81 fluorescent microscope, an

Olympus UPNFLN 40x/1.30 oil objective and a XM10 camera (Olympus, Tokyo, Japan) was used for viewing/image acquisition. The number of cells in each image ranged from approximately 2-12 with an approximate cellular confluency of 30-40%. The total number of cells counted per condition (control and separate knock downs) ranged from

150-200. cellSens software, provided by Olympus, was used to rapidly capture and overlay three separate images at different wavelengths, representing FITC, TRITC and

DAPI color channels.

2.9 Preparation of cDNA

Ribonucleic acid (RNA) was isolated from cells using Trizol®. In brief, 1mL of

Trizol® was added to a 100mm culture dish containing desired cells at approximately

80% confluency. Once cells detached (via pipetting) this volume was transferred into an autoclaved Eppendorf tube. 200 µL of Chloroform was added, the mixture was vortexed and then incubated at room temperature for 10 minutes. This was then centrifuged at

12,000 RPM at 4ºC for 15 minutes. The upper (clear) portion of the separated mixture

26 (~550 µL) was transferred to a new Eppendorf tube and the rest was discarded. An equal volume of Isopropanol was added, the mixture was inverted gently 3 times and incubated at room temperature for 30 minutes. This was then centrifuged at 12,000 RPM at 4ºC for

10 minutes, supernatant was removed via pipetting and 1 mL of 70% EtOH was added.

This was then centrifuged at 12,000 RPM at 4ºC for 10 minutes, supernatant was removed and the sample was allowed to air dry for approximately 15 minutes. The pellet was re-dissolved in 40 µL of nuclease free water for 1 hour at room temperature. RNA concentration was measured using a Nano Drop 2000 spectrophotometer from Thermo

Fisher Scientific (Waltham, MA).

After RNA isolation was complete, 2 µg of RNA was treated with DNase I from

Invitrogen (Carlsbad, CA) according to the manufacturer’s recommendations.

Superscript™ VILO® master mix reverse transcriptase from Invitrogen™ (Carlsbad,

CA) was then used to prepare cDNA for qRT-PCR (quantitative real time polymerase chain reaction) analysis The cDNA concentration was then measured on the Nano Drop spectrophotometer and diluted to 100ng for use in qRT-PCR analysis.

2.10 qRT-PCR

qRT-PCR was used to quantify the amount of gene knock down obtained with each shRNA construct used. The following table (Table 1) shows the primer pairs that were used to amplify each gene of interest.

27

qRT-PCR was performed using PowerUp™ SYBR® Green Master Mix from

Applied Biosystems (Foster City, CA) following the manufacturer’s instructions. Each reaction mixture consisted of 10 µl of SYBR® Green, forward and reverse primers at 500 nM final concentration, 100 ng of cDNA, and enough nuclease free water to equal a 20 ul reaction volume. GAPDH was used as a reference gene and the 2(-ΔΔCT) mathematical formula was utilized to determine the relative amount of a RhoGAP in control and knock down cell lines. qRT-PCR run information: denaturation, 10 seconds at 95º C, annealing and elongation, 30 seconds at 56 º C repeated for 50 cycles.

28 2.11 Cell Transfection

We used the TransIT-X2® Dynamic Delivery System by Mirus to transfect DNA into SUM159 cells following the manufactures instruction. In brief, cells were plated in a six-well plate so that the confluency at the time of transfection would reach approximately 90%. First, the X2 reagent was warmed to room temperature. Next, 7.5 µl of X2 and 2.5 µg of DNA was added to 250 µl of OptiMEM® media. This was incubated at room temperature for 20 minutes then gently dropped onto the aforementioned 6-well plate. 24 hours after addition, the cells were detached with trypsin and seeded onto coverslips for treatment and fixation prior to performing immunofluorescence.

For RhoGAP binding assays, HEK293 cells were transfected using calcium phosphate. In brief, 450 µl of sterile H2O was combined with 50 µl of CaCl2 (2.5 M filtered) and 10 µg of DNA. Next, 500 µl of HBSP 2x buffer was added to the bottom of the tube (HBSP buffer: 280 mM NaCl, 10 mM KCl, 1.5 mM Na2HPO4 - 12 H2O, Hepes

50 mM pH 7.05, Glucose 12 mM). Immediately after addition of HBSP 2x buffer the contents were mixed using bubbles from the bottom (by pipetting air). The DNA complex was incubated at room temperature for 30 seconds and added drop-wise to the cells. The following day the media was replaced with fresh and on the second day the cells were lysed.

2.12 Rho GTPase Activity Assay

Active RhoG and Rac1 pulldown assays were performed as described previously

(van Buul et al., 2007). In brief, 10 cm cell culture dishes with control and experimental

SUM159 cell lines were lysed in 50 mM Tris/HCl, pH 7.4, 10mM MgCl2, 150 mM NaCl,

29 1% Triton X-100 and EZBlock Protease inhibitor cocktail. The lysates were then centrifuged at 4º C at 14,000 x G for 5 minutes. Protein concentrations of the supernatants were determined and equal amounts of total protein were used to measure active (GTP bound) Rho GTPases. This was accomplished by rotating proteins with glutathione-Sepharose beads (GE Healthcare, PA) which were loaded with either 50 µg of glutathione transferase (GST)-ELMO to precipitate active RhoG or GST-Pak1binding domain (PBD) for Rac1. After rotating for 30 minutes the beads were washed four times with lysis buffer. Both the lysates and pulldowns were then immunoblotted with RhoG

(Santa Cruz, Santa Cruz) or Rac1 (BD Biosciences) antibodies.

2.13 RhoGAP-GTPase Binding Assay

For these expeiments, HEK293 cells were transfected with the RhoGAP of interest using calcium phosphate. The RhoGAP binding assay was performed as previously described (Garcia-Mata et al., 2006). Briefly, the HEK293 expressing the

RhoGAP of interest were washed, lysed and centrifuged at 13,000 rpm for 5 min at 4°C.

Once spun down, the supernatant was transferred to a separate tube (if multiple plates were used to increase protein production/concentration, the lysates were combined into a single tube and mixed to ensure homogeneity). Equal volumes of the supernatant were then added to tubes containing GST beads loaded with constitutively active Rho proteins including RhoA, Rac1, Cdc42 and RhoG. The constitutively active Rho mutants are unable to hydrolyze of GTP, so they form a stable structure enabling GAPs capable of binding Rho proteins to interact but unable to release the Rho GTPase due to an inability to convert to GDP-Rho. For Cdc42, Rac1 and RhoG the modification is a glutamine to leucine substitution at amino acid residue 61 (Q61L) and for RhoA it is a glutamine to

30 leucine substitution at amino acid residue 63 (Q63L). After addition of the supernatant to the various GST-fusion beads the tubes were rotated for 60 minutes at 4º C. The beads were then washed with lysis buffer 3-5 times and both the lysate and pulldowns were immunoblotted using a Myc (9E10) (Santa Cruz) antibody to detect the RhoGAP protein of interest.

31 Chapter Three

3. Results

3.1 RhoGAP screen

It has been previously shown that Rho GTPases have a functional role in invadopodia formation and cancer metastasis (Ayala et al., 2006; Nakahara et al., 1998;

Nakahara et al., 2003; Yamaguchi et al., 2005). However, the signaling pathways that regulate Rho GTPase signaling during these processes are not well characterized. The goal of this study was to identify novel RhoGAPs that have a role during invadopodia formation. By reviewing the available RhoGAP literature we compiled a list of 18 candidate RhoGAPs that could have an effect on invadopodia formation (Table 2). We then tested their role by silencing their expression and monitoring changes in invadopodia formation.

The RhoGAPs chosen for our screen were selected either by their known Rho

GTPase binding or because they are closely related to GAPs with association to invadopodia formation/function. For example, CHN1 is known to bind to Rac1, and Rac1 is known to have a role in invadopodia regulation (Colomba et al., 2011; Moshfegh et al.,

2014). We then speculated that knocking down a Rac1 GAP would affect Rac1 activity, which, in turn, would impact invadopodia formation. CHN2 is a RhoGAP which is closely related to CHN1 (described previously). It is possible that these two RhoGAPs work in conjunction with one another to regulate Rho GTPase activity, thus we chose to knock down CHN2 as well.

32

33 In our screen we utilized a RhoGAP targeted lentiviral shRNA library to knock down the candidate RhoGAPs in SUM159 cells. The virus conferred puromycin resistance along with the targeted shRNA, thus puromycin was used to select for only the cells that were infected. The infection/selection process was performed in a 96 well plate, and included five separate shRNA sequences for each gene targeted, a scrambled shRNA and non-infected cells as controls (Figure 6). Following 72 hours of puromycin selection, the cells were treated with PDBu, a phorbol ester which promotes invadopodia formation, fixed and stained for cortactin and actin so that invadopodia could be identified and quantified.

The 96 well plate was used in order to utilize a method of automated, high through-put data acquisition. We used the ImageXpress® microscope from Molecular

Devices (Sunnyvale, CA) to automatically capture images from 4 separate locations within each well in the 96 well plate, for a total of 12 images per well (DAPI, FITC, and

TRITC channels). This automation guaranteed a non-biased approach to data collection.

Once the data was collected images were exported and analyzed individually using Image

J software (Schneider et al., 2012).

34

35 The results of the screening process are shown in figure 7. The graph represents the percentage of cells that displayed invadopodia, using actin and cortactin as markers.

The average number of cells examined per condition ranged from ~20 to ~170, with an average of ~75-85 analyzed per knock down and per RhoGAP. The separate colors represent the RhoGAPs tested whereas the separate smaller bars represent the five individual shRNA sequences used to knock down each RhoGAP. The horizontal red line shows the average percentage of control SUM159 cells containing invadopodia. Of special interest are the TCGAP, CHN1 and ARHGAP12 genes. Each of these showed a marked increase in cells with invadopodia involving two or more of the shRNA constructs used. Also, and notably, aside from the increase in invadopodia, the cells themselves appeared similar to the control cells with no drastic changes in morphology.

36

Images captured from the RhoGAP screening process can be seen in figure 8.

These are representative images that show one shRNA sequence for each RhoGAP tested, as indicated by the number after the gene name along the top row. Sequence 5 for

TCGAP, sequence 5 for CHN1 and sequence 3 for ARHGAP12 were selected due to a pronounced phenotype. Individual pictures of cortactin (row 1) and actin (row 2) were

37 procured from the microscope used. The images were then combined into an overlay image and colored using Image J software (Schneider et al., 2012). The white arrows indicate locations of invadopodia.

38 3.2 TCGAP

TCGAP knock down had a reproducible impact on invadopodia formation. We repeated the knock down experiment in order to confirm what was seen in the RhoGAP screen. Our first step was to make stable knock down cell lines as described previously.

As these cell lines are not derived from a single cell (clonal), it is possible that amount of knock down present varied from cell to cell. Next, we used qRT-PCR to confirm that

TCGAP was successfully knocked down in the SUM159 cell lines. Figure 9-A shows a significant knock down of TCGAP with all three shRNA sequences used based on the screening results, sequences 1, 2, and 5. Approximately 60-70% knock down was identified in the cell lines tested.

After knock down confirmation we stained the cells to detect and quantify invadopodia. The results can be seen in figure 9-B. As seen in the initial RhoGAP screen, shRNA sequence 5 had a significant impact on invadopodia formation. These cells showed approximately twice the amount of cells with invadopodia than the control cells, an increase from ~30% to over 60%. Although not as drastic, shRNA sequences 1 and 2 also showed a significant increase in cells with invadopodia, rising to ~45% for each knock down.

Representative pictures of SUM159 cells after 30 minute PDBu treatment can be seen in Figure 9-C. TCGAP knock down cells form invadopodia that appear similar to the invadopodia seen in control cells, which are primarily in clumps or clusters located underneath the nuclear area, but do so more abundantly.

39

40 After confirming TCGAP knock down and performing invadopodia assays, we next attempted to determine whether TCGAP knock down had an impact on the active pool of Rac1 and RhoG, two Rho GTPases that our lab is interested in due to their ability to impact invadopodia formation. Figure 10-A shows that total Rac1 is unaffected by

TCGAP knock down and that there is no significant difference in the active pool of Rac1 between the control and TCGAP knock down. Interestingly, however, when TCGAP is knocked down a significant increase in GTP-RhoG is noted, especially with shRNA sequence 5, with no change in the total amount of RhoG in the cell.

41

42 3.3 CHN1

Much like TCGAP, CHN1 had a profound impact upon cells after PDBu treatment. Our first task was to establish that the lentiviral shRNA had successfully knocked down CHN1. Figure 11-A shows qRT-PCR data that suggests we were able to obtain approximately 70% knock down of CHN1 using the shRNA sequence 5 after 5 independent experiments. shRNA sequences 2 and 4 show the presence of a knock down, however only two experimental repeats were performed and statistical significance was not obtained. Currently P=.0522 for sequence 2 and P=0.1055 for sequence 4. It is likely that additional experiments will achieve statistical significance.

Next we repeated the invadopodia assays to corroborate the results obtained from the RhoGAP screen. These results can be viewed in Figure 11-B. CHN1 shRNA sequence 5 showed a statistically significant increase with roughly ~55% of cells forming invadopodia, compared to ~30% in the control cells. The other two sequences tested, 2 and 4, had roughly 40% and 45% of cells forming invadopodia. Three independent experiments were performed and over 150 cells were counted per experiment and per condition.

Figure 11-C shows a representative image of invadopodia formed in control and

CHN1 knock down cells. CHN1 knock down cells have more abundant invadopodia structures which appear punctate, while invadopodia in control cells are clustered. This phenotypic difference can be seen in figure 12.

43

44

45 After characterizing the effects of silencing CHN1 expression, we analyzed the effects of overexpressing myc-tagged CHN1 in SUM159 cells. Results from these experiments are shown in Figure 13-A and 13-B. In Figure 13-A, Myc is shown in green and cortactin (red) was used as an invadopodia marker. The cells that were successfully transfected showed a marked reduction in the number of cells forming invadopodia, around 15% or roughly half of what the control non-transfected cells exhibited. The

CHN1 overexpressing cells also appear to have fewer membrane extensions, such as lamellipodia and filopodia, and less membrane ruffling. This aspect of CHN1 overexpression was not fully explored as it is outside of the scope of this project.

Previous literature indicated that CHN1 is a RhoGAP specific for Rac1 with no binding/activity toward Cdc42 or RhoA. RhoG, however, was not tested and since Rac1 and RhoG are closely related we next assessed the binding of CHN1 to these various Rho

GTPases. In this assay a plasmid containing Myc-CHN1 DNA was transfected into

HEK293FT cells. Following this, equal volumes of CHN1 expressing lysate were incubated with GST bound constitutively active Rho proteins (RhoA, Rac1, Cdc42 and

RhoG) as well as empty GST beads as a control. After bead incubation and washing, 2% of the total lysate along with the total amount of protein captured from the beads were then separated using a 13% polyacrylamide gel and immunoblotted for myc to detect the precipitated CHN1. Figure 13-C shows that CHN1 clearly binds Rac1. The bottom portion of figure 13-C shows the membrane stained with amido black which shows the amount of protein-bead complex used to pull-down myc-CHN1.

46

After knocking down and overexpressing CHN1, we next assessed the impact of

CHN1 knock down on Rho GTPase activity within the cell. Two shRNA sequences,

47 numbers 4 and 5, we used for this purpose. Because CHN1 binds to Rac1 (as shown previously) were expected an increase in Rac1 activity levels when CHN1 was silenced.

Also, an increase in GTP-RhoG would further support interaction between CHN1 and

RhoG. As can be seen in 14-A, shRNA sequence 5 has a 2.5 fold increase in GTP-Rac1 levels when compared to the control. CHN1 knock down also increased the amount of

GTP-RhoG detected by approximately 2.5 fold despite a weaker interaction between

CHN1 and RhoG. This result further supports that CHN1 is capable of interacting with

RhoG. These results can be seen in Figure 14-B.

48

3.4 ARHGAP12

49 Our approach for ARHGAP12 is identical to how we began researching the previous two RhoGAPs involved in our study. First we established the knock down of

ARHGAP 12. As can be seen in figure 15-A, ARHGAP12 mRNA was significantly reduced using shRNA sequences 3 and 4. This reduction was calculated to be between

~55% and ~70% compared to ARHGAP12 expression levels in control cells.

ARHGAP12 sequence 5 appears to have reduced mRNA expression, but after 3 independent experiments a P value of 0.0525 was obtained which is outside of statistical significance.

We next repeated the invadopodia assays to validate results of the initial screen.

Once again, we were able to confirm the results obtained from the RhoGAP screen, showing that ARHGAP12 knock down increases the numbers of cells that form invadopodia in response to PDBu. As can be seen in Figure 15-B, shRNA knock down sequences 3 and 4 showed over 50% of cells forming invadopodia. ARHGAP12 sequence 5 only showed a slight increase in invadopodia over the control which was also consistent with what was seen in the screen.

Representative images of ARHGAP12 knock down cells after 30 minutes of

PDBu treatment can be seen in Figure 15-C. The knock down cells showed a strong phenotype consisting of tightly clustered groups of invadopodia located primarily at the cell interior underneath the nucleus. This structure and localization of the invadopodia is similar to what is seen in the control cells, with invadopodia being more prevalent in the

ARHGAP12 knock down cell lines.

50

51 The next step, similar to our approach when looking at CHN1, was to overexpress

ARHGAP12 and assess the impact on invadopodia formation. A representation of

ARHGAP12 over expressed calls and the quantification of the results can be seen in

Figures 16-A and B. When overexpressed, ARHGAP12 appears to concentrate in the peri-nuclear region. Additional experiments would be required to accurately determine its subcellular localization.

As predicted, the number of cells with invadopodia in ARHGAP12 overexpressed cells was decreased when compared to the control cells. However, a statistically significant difference was not achieved. After two independent experiments the decrease in cells with invadopodia seen has a P value of 0.0677. It is likely that an additional experiment will result in significance. However, the amount of overexpression is not quantifiable via western blot or qRT-PCR due to low transfection efficiency. It is possible that the amount of myc-ARHGAP12 transfected is only slightly above basal expression levels. In knock down cell lines 3 and 4, over 50% of cells counted showed invadopodia, whereas in the overexpression experiments only 20% of the cells formed invadopodia.

Previous literature has stated that ARHGAP12 binds specifically to Rac1. Our research supports this assertion. As can be seen in Figure 16-C, ARHGAP12 is unable to bind to either RhoA, Cdc42 or RhoG, while binding strongly to Rac1.

52

53 Chapter Four

4. Discussion

This body of work began with a screening of 18 RhoGAPs in an attempt to identify one or more that are involved in the formation of invadopodia. After the screening was complete, three GAPs were chosen for additional study. We have since began to assess how these RhoGAPs are functioning within the cell by characterizing their effect on invadopodia formation and the Rho GTPases they are capable of binding.

We have also looked into the effect of RhoGAP knockdown on the active pool of select

Rho GTPases within the cell. Our goal in researching the RhoGAPs involved in invadopodia formation is to expand upon the knowledge currently available to researchers on cellular metastasis and to provide novel gene targets for the development of novel therapeutic approaches in cancer treatment.

4.1 RhoGAP screen

The initial goal of our RhoGAP screen was to identify a RhoGAP that functioned as a promotor of invadopodia formation. If such a RhoGAP were knocked down within a cell the formation or assembly of invadopodia would be repressed, theoretically through an increase in the concentration of a GTP bound Rho GTPase. However, in our screen of

18 RhoGAPs there were no notable knock down cell lines that showed that particular phenotype. Interestingly, three RhoGAPs showed the opposite effect. When TCGAP,

CHN1 or ARHGAP12 were knocked down a sharp increase in cells containing invadopodia was noted, suggesting that these RhoGAPs normally act as suppressors of

54 invadopodia formation and when absent invadopodia are able to assemble at an increased rate.

It is possible, however, that one or more of these GAPs are not acting on invadopodia formation and that invadopodia are forming at the same rate in the knock down cells lines as the invadopodia forming in the control cells. One or all of these

RhoGAPs may be involved, instead, in the disassembly rather than the assembly of invadopodia. If this is the case, these RhoGAPs would normally act within cellular signaling pathways that promote the turnover of invadopodia, and when expression levels are diminished in the cell, invadopodia are more persistent and have a longer lifetime.

Alternate roles for these RhoGAPs (not related to GAP activity) within the cell is a possible third explanation for the noted increase in the number of cells forming invadopodia. It is possible that these RhoGAPs are part of a scaffolding network or involved in the translocation of a protein (or protein complex) that facilitates the dissolution of the invadopodia structure.

4.1.1 Future aims

It is clear that there are numerous questions remaining in regards to the RhoGAP screening project. One of of which revolves around the remaining RhoGAPs that have yet to be screened for their involvement in the formation or regulation of invadopodia. This research has examined 18 RhoGAPs, leaving 49 RhoGAP family members that could play a role in this cellular process. Future work will continue on through these remaining

GAPs and identify others that are capable of impacting invadopodia formation.

55 Aside from screening additional RhoGAPs there are also other candidates found in our RhoGAP screen that warrant further study. These RhoGAPs include RICS, p190A and srGAP1. Although the phenotypes observed in these knock down cell lines were not as pronounced or consistent as with the three RhoGAPs discussed in this research, it may be worth additional experimentation to determine if these proteins are also involved in invadopodia formation or disassembly.

4.2 TCGAP

Not much information is currently available in regards to TCGAP which makes it difficult to speculate as to the actions or interactions of this gene. Adding to this is the fact that existing research was performed in adipocytes or neuronal cells with nothing produced in tumor cell lines to date. Nonetheless, combining our research with available information suggests a relatively straightforward method through which TCGAP is able to impact invadopodia formation.

This work has demonstrated that TCGAP knock down results in an increase in cells containing invadopodia in SUM159 breast cancer cells. It has also suggested that

TCGAP knockdown increases the active pool of RhoG within the cell, yet does not impact the active amount of Rac1. We have not yet determined the effect of TCGAP knockdown on Cdc42 activity in SUM159 cells, but we speculate that TCGAP is acting as a regulator of Cdc42 activity based on previous reports that TCGAP interacts with this

Rho protein (Chiang et al., 2003).

It has been shown that TCGAP has in vitro GAP activity not only toward Cdc42 but also toward TC10β, a lesser known Rho GTPase (Chiang et al., 2003). The same

56 work also determined that TCGAP does not act on Rac1 or RhoA (Chiang et al., 2003). It is also known that Cdc42 promotes invadopodia formation and matrix degradation

(Razidlo et al., 2014). It is likely that TCGAP functions to regulate Cdc42 activity within

SUM159 cells, maintaining a balance between GTP and GDP bound Cdc42. The lack of

TCGAP in our knockdown cell lines likely causes an increase of GTP-Cdc42 and this surplus Cdc42 activity would act to increase invadopodia formation, thus increasing the number of cells expressing invadopodia. This explanation, however, does not take into consideration the increase in active RhoG in the TCGAP knock down cells.

Our research shows that RhoG knockdown dramatically increases invadopodia formation in SUM159 cells, with as many as ~75-80% of cells examined containing invadopodia (control: ~30%) (Unpublished Data). We have also determined that in RhoG knockdown cells invadopodia lifetime is significantly increased (Unpublished Data). This suggests that RhoG acts on invadopodia disassembly. Even though TCGAP knockdown increases the active pool of RhoG, it is possible that invadopodia are forming at such an increased rate that even the increased RhoG is incapable of keeping up with the number of invadopodia forming within the cell. It is also possible that the increase in RhoG is negligible compared to the increase in Cdc42, however as mentioned previously this has yet to be determined. As we have yet to explore interaction partners of TCGAP, we are unable to speculate as to whether increased active RhoG is a direct response to TCGAP knockdown or if it is related to an imbalance of the Rho GTPases within the cell.

Alternatively, it is possible that TCGAP acts as a scaffold protein allowing for disassembly machinery to assemble at, or localize to, sites where invadopodia are

57 present. In either case the number of cells showing invadopodia would increase due to increased invadopodia lifetimes.

4.2.1 Future aims

It is apparent that a lot of work remains to be done on this project. First, it is important to create or obtain appropriate TCGAP clones so that overexpression and binding experiments can be performed. We expect that TCGAP over expression will suppress invadopodia formation and that TCGAP will bind to Cdc42, as shown in previous work, and possibly RhoG. We next want to perform activity assays for Cdc42 and RhoA. As we suspect that TCGAP is acting through Cdc42, we anticipate an increase in the active pool of Cdc42. Previous literature states that TCGAP does not interact with

RhoA, thus it is expected that RhoA activity levels will remain the same in TCGAP knock down and control cells which would act as an internal experimental control

(Chiang et al., 2003).

Previous literature has shown that TCGAP has GAP activity toward Cdc42 in vitro (Chiang et al., 2003). As we have detected an increase in RhoG activity in TCGAP knock down cells we will perform a GAP activity assay to determine whether TCGAP is

GAP active toward RhoG. Next, we will transfect a GAP inactive mutant TCGAP into

TCGAP knockdown cells. This would determine if the GAP activity of TCGAP is required for the impact on invadopodia formation or if localization/scaffold roles are responsible. If GAP inactive TCGAP reverts the cells to a control phenotype, GAP activity would be deemed unnecessary for its role in invadopodia formation/disassembly.

58 Confocal microscopy will then be used in control, TCGAP knockdown and GFP-

TCGAP overexpressed cells. This will provide information as to the localization of

TCGAP within the cell and total invadopodia lifetime under these conditions. This will also determine whether or not TCGAP is found at sites of invadopodia, however a different reporter construct would be necessary in order to determine if GAP activity is taking place specifically at these sites. We speculate that TCGAP acts on Cdc42 at invadopodia resulting in Cdc42 inactivation and promoting invadopodia disassembly. We could further confirm our findings by transfecting constitutively active Cdc42 into

SUM159 cells. If TCGAP is acting through Cdc42, and increased GTP-Cdc42 is responsible for the effects on invadopodia formation/disassembly seen in TCGAP knock down cell lines, cells overexpressing constitutively active Cdc42 should resemble

TCGAP knockdown in invadopodia number and overall invadopodia lifetime.

The final experiments proposed for TCGAP are matrix degradation and cell invasion assays. An increase in the number of invadopodia in TCGAP knockdown cell lines should, in theory, result in increased matrix degradation. This can be determined using FITC-labeled gelatin coated coverslips. After seeding cells onto the coverslips and allowing for incubation time, total fluorescence can be obtained. Areas of matrix degradation will not fluoresce and are associated with invadopodia localization. Thus, cells with an increase in the number of matrix degrading invadopodia will result in less total fluorescence. Since TCGAP knockdown cell lines show an increase in invadopodia, there will likely be a decrease in total fluorescence. Increased degradation will likely result in increased cellular invasion, and if this is the case the trans-well migration experiments will show more migrated cells in the TCGAP knock down cell lines.

59 These proposed experiments will provide a clear picture as to how TCGAP is involved in invadopodia formation, or possibly disassembly, in SUM159 cells. After this is accomplished, distant future directions could include assessing whether or not TCGAP has differing activation states. GAPs have been shown to be activated via phosphorylation (Morishita et al., 2015) or by lipid binding (Erlmann et al., 2009) and either of these may contribute to TCGAP activation or inactivation.

4.3 CHN1

The chimaerin gene was first characterized in 1990 and the current pool of knowledge surrounding this protein is fairly extensive (Ahmed et al., 1990). However, similar to TCGAP, the focus of the majority of Chimaerin research has been in brain tissue. Most notably for this work is that CHN1 is characterized as a Rac1 specific

RhoGAP and the only one to bind to phorbol esters and DAG (Yang and Kazanietz,

2007). Here we present a novel role for CHN1 in the regulation of invadopodia in

SUM159 breast cancer cells.

We have identified that CHN1 knock down increases the number of cells that have invadopodia after 30 minutes of PDBu treatment and that CHN1 overexpression decreases this number. In our screening, and in later invadopodia assays (data not shown), CHN2 was also knocked down and no impact was noted on invadopodia formation indicating that this result is CHN1 specific. We have also confirmed previous reports that CHN1 binds to Rac1 and not RhoA or Cdc42. Interestingly, our RhoGAP-

GTPase binding experiments suggest that CHN1 may interact with RhoG, which is supported with the finding that CHN1 knock down increases the active pool of RhoG as well as Rac1 within the cell. Data generated in our lab shows that over expression of

60 Rac1 increases the number of cells with invadopodia in a dose dependent manner and that

Rac1 knock out abolishes invadopodia formation (Unpublished Data). However, we have also demonstrated that RhoG knock down drastically increases the number of cells with invadopodia as mentioned previously (Unpublished Data). This suggests that RhoG and

Rac1 are both able to influence invadopodia, yet do so in different ways.

We propose that in the absence of CHN1 the active pool of Rac1 is increased which, in turn, promotes the formation of invadopodia within the cell. Also, as mentioned previously, RhoG knockdown increases invadopodia lifetime suggesting a role in invadopodia disassembly (Unpublished Data). However, in CHN1 knock down cells an increase in RhoG as well as Rac1 activity is seen. Thus, while there is an increase in invadopodia formation there is likely also an increase in invadopodia disassembly. This would result in an increase in invadopodia activity within the cell which is the likely reason for the overall increase in invadopodia seen in CHN1 knock down cells. This increase in invadopodia activity (formation and disassembly) is most likely sustained over time due to the sustained increase in active Rac1 and RhoG which results in the overall increase in cells with invadopodia noted in CHN1 knock down cell lines.

It is also possible that PDBu treatment is capable altering the activation state of

CHN1 via its C1 domain (phorbol ester/DAG binding) resulting in increased GAP activity. It has been speculated previously that phorbol esters may cause a conformational change in CHN1 resulting in increased GAP activity (Caloca et al., 2003). This increase in GAP activity would then result in a decrease in active Rac1 and, as mentioned previously and supported by the literature, Rac1 activity promotes invadopodia formation

(Harper et al., 2010; Nascimento et al., 2011; Pignatelli et al., 2012). In the absence of

61 CHN1 Rac1 activity is not repressed causing an increase in the number of cells forming invadopodia.

4.3.1 Future aims

The future goals for the CHN1 project are in parallel with the TCGAP project.

First, we will perform RhoA and Cdc42 activity assays. Although CHN1 is shown to not bind to these Rho GTPases, it should be established that the increase in Rac1 and RhoG activity levels in CHN1 knock down cell lines does not affect the activity of RhoA or

Cdc42. It should also be determined that the increase in invadopodia in CHN1 knock down cells is not related to an increase in Cdc42 activity.

Next, a GAP inactive mutant of CHN1 will be used in an attempt to rescue the

CHN1 knock down phenotype. If the lack of CHN1 GAP activity is responsible for the observed phenotype in the knock down cells, the presence of GAP inactive CHN1 should not result in a return to control levels of invadopodia expression. If a phenotypic rescue is noted and the GAP activity is not responsible for the increase in cells with invadopodia, other actions such as scaffolding or protein recruitment will be implicated in the process and explored as described in the previous future aims section. Additionally, a CHN1 mutant will be created in order to determine if the C1 domain, responsible for phorbol ester/DAG binding, is responsible for the increase in cells with invadopodia. If expression of the C1 domain mutant is incapable of repressing invadopodia in control cells or phenotypically rescuing the CHN1 knock down cells, a PDBu-CHN1 signaling event will be implicated in this process.

62 Confocal microscopy will be used to determine the lifetime of invadopodia and the time of increased invadopodia activity after the addition of PDBu. We anticipate that invadopodia lifetime will be unaffected, yet the time of invadopodia activity will be increased related to the increase in active Rac1 and RhoG levels within CHN1 knockdown cells. If noted, this could be replicated by co-transfecting constitutively active RhoG and Rac1, which would produce a phenotype similar to CHN1 knock down and providing further evidence to support our hypothesis. Conversely, dominant negative

Rac1 and RhoG could be co-transfected into CHN1 knock down cells which would presumably rescue the phenotype by out-competing the active forms within the cell. The difficulty in these experiments lies in the immunofluorescent detection, with separately tagged Rac1 and RhoG, F-actin, and cortactin needing to be stained for simultaneously.

Finally, the impact of CHN1 knockdown on matrix degradation and cell invasion will be determined using previously described assays. As an increase in invadopodia is noted in CHN1 knock down cells, as with TCGAP knock down, it is likely that an increase in matrix degradation and cell invasion will be seen. However, it is possible that increased invadopodia turnover could result in immature or non-functional invadopodia which would decrease/abolish the degrading capability of the cell and reduce the invasive potential of a cell. In order to make concrete predictions the results of either the matrix degradation assays or the live imaging experiments are necessary.

4.4 ARHGAP12

Although ARHGAP12 was first characterized in 2002 not much information is available in regards to this RhoGAP (Zhang et al., 2002). In this work we propose a role for ARHGAP12 in invadopodia formation in breast cancer cells. During the initial

63 characterization it was shown that ARHGAP12 is ubiquitously expressed in musculoskeletal, placental, , , kidney, liver, brain and pancreatic cells (Zhang et al., 2002). It was further shown through northern blot analysis that ARHGAP12 was expressed strongly in breast cancer cell lines as well as seven other tumor cell lines

(Zhang et al., 2002). This evidence supports a role for ARHGAP12 in carcinogenic cell lines.

As seen with TCGAP and CHN1, when ARHGAP12 is knocked down within

SUM159 breast cancer cells an increase in the number of cells with invadopodia is noted after PDBu treatment. This suggests that ARHGAP12 has a basal function in the cell that has either a repressive role in invadopodia formation or promotes the disassembly of invadopodia, similar to what was previously discussed in regards to TCGAP and CHN1 knock down cells. Our preliminary data showed that ARHGAP12 over expression decreases the number of cells with invadopodia, although This needs to be firmly established in the future.

Much like CHN1, ARHGAP12 is able to physically interact with Rac1. However, we have seen no interaction with the other three Rho GTPases tested. It is possible that

ARHGAP12 and CHN1 have overlapping roles within the cell and that these signaling pathways are redundant, meaning that they perform the same function but using different signaling pathways/events.

We propose that ARHGAP12 is a regulator of Rac1 activity and, much like in

CHN1 knock down, the absence of ARHGAP12 results in an increased pool of active

Rac1 within the cell. This increase in Rac1 activity would drive invadopodia assembly and result in an increase in cells with invadopodia.

64 4.4.1 Future aims

The proposed future aims for ARHGAP12 mirror those of TCGAP and CHN1, and as such will be kept brief. Activity assays for ARHGAP12 knock down cells will be performed (Cdc42, Rac1, and RhoG) as well as GAP inactive mutant assays as described previously. Dominant active Rac1 will be expressed in ARHGAP12 knock down cells which should out compete the GTP-Rac1, thus decreasing the invadopodia number in

ARHGAP12 knock down cell lines. Finally, matrix degradation and cell invasion assays will be performed.

References

65 Ahmed, S., R. Kozma, C. Monfries, C. Hall, H.H. Lim, P. Smith, and L. Lim. 1990.

Human brain n-chimaerin cDNA encodes a novel phorbol ester receptor. Biochem

J. 272:767-773.

Alexandropoulos, K., G. Cheng, and D. Baltimore. 1995. Proline-rich sequences that bind

to Src homology 3 domains with individual specificities. Proc Natl Acad Sci U S

A. 92:3110-3114.

Arias-Romero, L.E., and J. Chernoff. 2013. Targeting Cdc42 in cancer. Expert Opin Ther

Targets. 17:1263-1273.

Ayala, I., M. Baldassarre, G. Caldieri, and R. Buccione. 2006. Invadopodia: a guided

tour. Eur J Cell Biol. 85:159-164.

Ba, W., M.M. Selten, J. van der Raadt, H. van Veen, L.L. Li, M. Benevento, A.R.

Oudakker, R.S. Lasabuda, S.J. Letteboer, R. Roepman, R.J. van Wezel, M.J.

Courtney, H. van Bokhoven, and N. Nadif Kasri. 2016. ARHGAP12 Functions as

a Developmental Brake on Excitatory Synapse Function. Cell Rep.

Baldassarre, M., I. Ayala, G. Beznoussenko, G. Giacchetti, L.M. Machesky, A. Luini,

and R. Buccione. 2006. Actin dynamics at sites of extracellular matrix

degradation. Eur J Cell Biol. 85:1217-1231.

Bernards, R., and R.A. Weinberg. 2002. A progression puzzle. Nature. 418:823.

Bishop, A.L., and A. Hall. 2000. Rho GTPases and their effector proteins. Biochem J.

348 Pt 2:241-255.

Bork, P., and M. Sudol. 1994. The WW domain: a signalling site in dystrophin? Trends

Biochem Sci. 19:531-533.

66 Bowden, E.T., M. Barth, D. Thomas, R.I. Glazer, and S.C. Mueller. 1999. An invasion-

related complex of cortactin, paxillin and PKCmu associates with invadopodia at

sites of extracellular matrix degradation. Oncogene. 18:4440-4449.

Caloca, M.J., H. Wang, and M.G. Kazanietz. 2003. Characterization of the Rac-GAP

(Rac-GTPase-activating protein) activity of beta2-chimaerin, a 'non-protein

kinase C' phorbol ester receptor. Biochem J. 375:313-321.

Campa, F., and P.A. Randazzo. 2008. Arf GTPase-activating proteins and their potential

role in cell migration and invasion. Cell Adh Migr. 2:258-262.

Chavez, K.J., S.V. Garimella, and S. Lipkowitz. 2010. Triple negative breast cancer cell

lines: one tool in the search for better treatment of triple negative breast cancer.

Breast Dis. 32:35-48.

Chen, W.T. 1989. Proteolytic activity of specialized surface protrusions formed at rosette

contact sites of transformed cells. J Exp Zool. 251:167-185.

Chen, W.T., K. Olden, B.A. Bernard, and F.F. Chu. 1984. Expression of transformation-

associated protease(s) that degrade fibronectin at cell contact sites. J Cell Biol.

98:1546-1555.

Chevalier, C., A. Cannet, S. Descamps, A. Sirvent, V. Simon, S. Roche, and C. Benistant.

2015. ABL tyrosine kinase inhibition variable effects on the invasive properties of

different triple negative breast cancer cell lines. PLoS One. 10:e0118854.

Chiang, S.H., J. Hwang, M. Legendre, M. Zhang, A. Kimura, and A.R. Saltiel. 2003.

TCGAP, a multidomain Rho GTPase-activating protein involved in insulin-

stimulated glucose transport. EMBO J. 22:2679-2691.

67 Colomba, A., S. Giuriato, E. Dejean, K. Thornber, G. Delsol, H. Tronchere, F. Meggetto,

B. Payrastre, and F. Gaits-Iacovoni. 2011. Inhibition of Rac controls NPM-ALK-

dependent lymphoma development and dissemination. Blood Cancer J. 1:e21.

David-Pfeuty, T., and S.J. Singer. 1980. Altered distributions of the cytoskeletal proteins

vinculin and alpha-actinin in cultured fibroblasts transformed by Rous sarcoma

virus. Proc Natl Acad Sci U S A. 77:6687-6691.

Diekmann, D., S. Brill, M.D. Garrett, N. Totty, J. Hsuan, C. Monfries, C. Hall, L. Lim,

and A. Hall. 1991. Bcr encodes a GTPase-activating protein for p21rac. Nature.

351:400-402.

Ellerbroek, S.M., K. Wennerberg, W.T. Arthur, J.M. Dunty, D.R. Bowman, K.A.

DeMali, C. Der, and K. Burridge. 2004. SGEF, a RhoG guanine nucleotide

exchange factor that stimulates macropinocytosis. Mol Biol Cell. 15:3309-3319.

Erlmann, P., S. Schmid, F.A. Horenkamp, M. Geyer, T.G. Pomorski, and M.A. Olayioye.

2009. DLC1 activation requires lipid interaction through a polybasic region

preceding the RhoGAP domain. Mol Biol Cell. 20:4400-4411.

Fidler, I.J., and M.L. Kripke. 1977. Metastasis results from preexisting variant cells

within a malignant tumor. Science. 197:893-895.

Fillmore, C.M., and C. Kuperwasser. 2008. Human breast cancer cell lines contain stem-

like cells that self-renew, give rise to phenotypically diverse progeny and survive

chemotherapy. Breast Cancer Res. 10:R25.

Flanagan, L., K. Van Weelden, C. Ammerman, S.P. Ethier, and J. Welsh. 1999. SUM-

159PT cells: a novel estrogen independent human breast cancer model system.

Breast Cancer Res Treat. 58:193-204.

68 Furmaniak-Kazmierczak, E., S.W. Crawley, R.L. Carter, D.H. Maurice, and G.P. Cote.

2007. Formation of extracellular matrix-digesting invadopodia by primary aortic

smooth muscle cells. Circ Res. 100:1328-1336.

Garcia-Mata, R., K. Wennerberg, W.T. Arthur, N.K. Noren, S.M. Ellerbroek, and K.

Burridge. 2006. Analysis of activated GAPs and GEFs in cell lysates. Methods

Enzymol. 406:425-437.

Gentile, A., L. D'Alessandro, L. Lazzari, B. Martinoglio, A. Bertotti, A. Mira, L.

Lanzetti, P.M. Comoglio, and E. Medico. 2008. Met-driven invasive growth

involves transcriptional regulation of Arhgap12. Oncogene. 27:5590-5598.

Hall, A. 1990. The cellular functions of small GTP-binding proteins. Science. 249:635-

640.

Hall, A., and C.D. Nobes. 2000. Rho GTPases: molecular switches that control the

organization and dynamics of the actin cytoskeleton. Philos Trans R Soc Lond B

Biol Sci. 355:965-970.

Hall, C., C. Monfries, P. Smith, H.H. Lim, R. Kozma, S. Ahmed, V. Vanniasingham, T.

Leung, and L. Lim. 1990. Novel human brain cDNA encoding a 34,000 Mr

protein n-chimaerin, related to both the regulatory domain of protein kinase C and

BCR, the product of the breakpoint cluster region gene. J Mol Biol. 211:11-16.

Hall, C., W.C. Sin, M. Teo, G.J. Michael, P. Smith, J.M. Dong, H.H. Lim, E. Manser,

N.K. Spurr, T.A. Jones, and et al. 1993. Alpha 2-chimerin, an SH2-containing

GTPase-activating protein for the ras-related protein p21rac derived by alternate

splicing of the human n-chimerin gene, is selectively expressed in brain regions

and testes. Mol Cell Biol. 13:4986-4998.

69 Harper, K., D. Arsenault, S. Boulay-Jean, A. Lauzier, F. Lucien, and C.M. Dubois. 2010.

Autotaxin promotes cancer invasion via the lysophosphatidic acid receptor 4:

participation of the cyclic AMP/EPAC/Rac1 signaling pathway in invadopodia

formation. Cancer Res. 70:4634-4643.

Haslam, R.J., H.B. Koide, and B.A. Hemmings. 1993. Pleckstrin domain homology.

Nature. 363:309-310.

Howlader N, N.A., Krapcho M, Garshell J, Miller D, Altekruse SF, Kosary CL, Yu M,

Ruhl J, Tatalovich Z, Mariotto A, Lewis DR, Chen HS, Feuer EJ, Cronin KA

(eds). 2015. SEER Cancer Statistics Review, 1975-2012. Vol. 2016. National

Cancer Institute.

Kawashima, T., K. Hirose, T. Satoh, A. Kaneko, Y. Ikeda, Y. Kaziro, T. Nosaka, and T.

Kitamura. 2000. MgcRacGAP is involved in the control of growth and

differentiation of hematopoietic cells. Blood. 96:2116-2124.

Kazanietz, M.G. 2000. Eyes wide shut: protein kinase C isozymes are not the only

receptors for the phorbol ester tumor promoters. Mol Carcinog. 28:5-11.

Kazanietz, M.G. 2002. Novel "nonkinase" phorbol ester receptors: the C1 domain

connection. Mol Pharmacol. 61:759-767.

Kim, Y., C.M. Ha, and S. Chang. 2013. SNX26, a GTPase-activating protein for Cdc42,

interacts with PSD-95 protein and is involved in activity-dependent dendritic

spine formation in mature neurons. J Biol Chem. 288:29453-29466.

Kötting, C., and K. Gerwert. 2004. Time-resolved FTIR studies provide activation free

energy, activation enthalpy and activation entropy for GTPase reactions.

Chemical Physics. 307:227-232.

70 Kozma, R., S. Ahmed, A. Best, and L. Lim. 1996. The GTPase-activating protein n-

chimaerin cooperates with Rac1 and Cdc42Hs to induce the formation of

lamellipodia and filopodia. Mol Cell Biol. 16:5069-5080.

Kusama, T., M. Mukai, H. Endo, O. Ishikawa, M. Tatsuta, H. Nakamura, and M. Inoue.

2006. Inactivation of Rho GTPases by p190 RhoGAP reduces human pancreatic

cancer cell invasion and metastasis. Cancer Sci. 97:848-853.

Leong, H.S., A.E. Robertson, K. Stoletov, S.J. Leith, C.A. Chin, A.E. Chien, M.N.

Hague, A. Ablack, K. Carmine-Simmen, V.A. McPherson, C.O. Postenka, E.A.

Turley, S.A. Courtneidge, A.F. Chambers, and J.D. Lewis. 2014. Invadopodia are

required for cancer cell extravasation and are a therapeutic target for metastasis.

Cell Rep. 8:1558-1570.

Lin, C.W., M.S. Sun, M.Y. Liao, C.H. Chung, Y.H. Chi, L.T. Chiou, J. Yu, K.L. Lou,

and H.C. Wu. 2014. Podocalyxin-like 1 promotes invadopodia formation and

metastasis through activation of Rac1/Cdc42/cortactin signaling in breast cancer

cells. Carcinogenesis. 35:2425-2435.

Liu, H., T. Nakazawa, T. Tezuka, and T. Yamamoto. 2006. Physical and functional

interaction of Fyn tyrosine kinase with a brain-enriched Rho GTPase-activating

protein TCGAP. J Biol Chem. 281:23611-23619.

Madaule, P., and R. Axel. 1985. A novel ras-related gene family. Cell. 41:31-40.

Minoshima, Y., T. Kawashima, K. Hirose, Y. Tonozuka, A. Kawajiri, Y.C. Bao, X.

Deng, M. Tatsuka, S. Narumiya, W.S. May, Jr., T. Nosaka, K. Semba, T. Inoue,

T. Satoh, M. Inagaki, and T. Kitamura. 2003. Phosphorylation by aurora B

converts MgcRacGAP to a RhoGAP during cytokinesis. Dev Cell. 4:549-560.

71 Morishita, Y., K. Tsutsumi, and Y. Ohta. 2015. Phosphorylation of Serine 402 Regulates

RacGAP Activity of FilGAP. J Biol Chem.

Moshfegh, Y., J.J. Bravo-Cordero, V. Miskolci, J. Condeelis, and L. Hodgson. 2014. A

Trio-Rac1-Pak1 signalling axis drives invadopodia disassembly. Nat Cell Biol.

16:574-586.

Mueller, S.C., and W.T. Chen. 1991. Cellular invasion into matrix beads: localization of

beta 1 integrins and fibronectin to the invadopodia. J Cell Sci. 99 ( Pt 2):213-225.

Mueller, S.C., Y. Yeh, and W.T. Chen. 1992. Tyrosine phosphorylation of membrane

proteins mediates cellular invasion by transformed cells. J Cell Biol. 119:1309-

1325.

Murphy, D.A., and S.A. Courtneidge. 2011. The 'ins' and 'outs' of podosomes and

invadopodia: characteristics, formation and function. Nat Rev Mol Cell Biol.

12:413-426.

Nakahara, H., S.C. Mueller, M. Nomizu, Y. Yamada, Y. Yeh, and W.T. Chen. 1998.

Activation of beta1 integrin signaling stimulates tyrosine phosphorylation of

p190RhoGAP and membrane-protrusive activities at invadopodia. J Biol Chem.

273:9-12.

Nakahara, H., T. Otani, T. Sasaki, Y. Miura, Y. Takai, and M. Kogo. 2003. Involvement

of Cdc42 and Rac small G proteins in invadopodia formation of RPMI7951 cells.

Genes Cells. 8:1019-1027.

Nakamura, T., M. Komiya, K. Sone, E. Hirose, N. Gotoh, H. Morii, Y. Ohta, and N.

Mori. 2002. Grit, a GTPase-activating protein for the Rho family, regulates

72 neurite extension through association with the TrkA receptor and N-Shc and

CrkL/Crk adapter molecules. Mol Cell Biol. 22:8721-8734.

Nascimento, C.F., A.S. de Siqueira, J.J. Pinheiro, V.M. Freitas, and R.G. Jaeger. 2011.

Laminin-111 derived peptides AG73 and C16 regulate invadopodia activity of a

human adenoid cystic carcinoma cell line. Exp Cell Res. 317:2562-2572.

Neve, R.M., K. Chin, J. Fridlyand, J. Yeh, F.L. Baehner, T. Fevr, L. Clark, N. Bayani,

J.P. Coppe, F. Tong, T. Speed, P.T. Spellman, S. DeVries, A. Lapuk, N.J. Wang,

W.L. Kuo, J.L. Stilwell, D. Pinkel, D.G. Albertson, F.M. Waldman, F.

McCormick, R.B. Dickson, M.D. Johnson, M. Lippman, S. Ethier, A. Gazdar, and

J.W. Gray. 2006. A collection of breast cancer cell lines for the study of

functionally distinct cancer subtypes. Cancer Cell. 10:515-527.

Nimnual, A.S., L.J. Taylor, and D. Bar-Sagi. 2003. Redox-dependent downregulation of

Rho by Rac. Nat Cell Biol. 5:236-241.

Ohta, Y., J.H. Hartwig, and T.P. Stossel. 2006. FilGAP, a Rho- and ROCK-regulated

GAP for Rac binds filamin A to control actin remodelling. Nat Cell Biol. 8:803-

814.

Olofsson, B. 1999. Rho guanine dissociation inhibitors: pivotal molecules in cellular

signalling. Cell Signal. 11:545-554.

Olson, M.F., A. Ashworth, and A. Hall. 1995. An essential role for Rho, Rac, and Cdc42

GTPases in cell cycle progression through G1. Science. 269:1270-1272.

Ozaki, K., and S. Hatano. 1984. Mechanism of regulation of actin polymerization by

Physarum profilin. J Cell Biol. 98:1919-1925.

Pawson, T., and J. Schlessingert. 1993. SH2 and SH3 domains. Curr Biol. 3:434-442.

73 Pignatelli, J., D.A. Tumbarello, R.P. Schmidt, and C.E. Turner. 2012. Hic-5 promotes

invadopodia formation and invasion during TGF-beta-induced epithelial-

mesenchymal transition. J Cell Biol. 197:421-437.

Ponting, C.P. 1996. Novel domains in NADPH oxidase subunits, sorting nexins, and

PtdIns 3-kinases: binding partners of SH3 domains? Protein Sci. 5:2353-2357.

Qadir, M.I., A. Parveen, and M. Ali. 2015. Cdc42: Role in Cancer Management. Chem

Biol Drug Des. 86:432-439.

Raymond, K., E. Bergeret, M.C. Dagher, R. Breton, R. Griffin-Shea, and M.O.

Fauvarque. 2001. The Rac GTPase-activating protein RotundRacGAP interferes

with Drac1 and Dcdc42 signalling in Drosophila melanogaster. J Biol Chem.

276:35909-35916.

Razidlo, G.L., B. Schroeder, J. Chen, D.D. Billadeau, and M.A. McNiven. 2014. Vav1 as

a central regulator of invadopodia assembly. Curr Biol. 24:86-93.

Ridley, A.J. 2015. Rho GTPase signalling in cell migration. Curr Opin Cell Biol. 36:103-

112.

Ridley, A.J., and A. Hall. 1992. The small GTP-binding protein rho regulates the

assembly of focal adhesions and actin stress fibers in response to growth factors.

Cell. 70:389-399.

Rosario, M., R. Franke, C. Bednarski, and W. Birchmeier. 2007. The neurite outgrowth

multiadaptor RhoGAP, NOMA-GAP, regulates neurite extension through SHP2

and Cdc42. J Cell Biol. 178:503-516.

Saraste, M., and M. Hyvonen. 1995. Pleckstrin homology domains: a fact file. Curr Opin

Struct Biol. 5:403-408.

74 Scheffzek, K., M.R. Ahmadian, W. Kabsch, L. Wiesmuller, A. Lautwein, F. Schmitz, and

A. Wittinghofer. 1997. The Ras-RasGAP complex: structural basis for GTPase

activation and its loss in oncogenic Ras mutants. Science. 277:333-338.

Schlam, D., R.D. Bagshaw, S.A. Freeman, R.F. Collins, T. Pawson, G.D. Fairn, and S.

Grinstein. 2015. Phosphoinositide 3-kinase enables phagocytosis of large particles

by terminating actin assembly through Rac/Cdc42 GTPase-activating proteins.

Nat Commun. 6:8623.

Schneider, C.A., W.S. Rasband, and K.W. Eliceiri. 2012. NIH Image to ImageJ: 25 years

of image analysis. Nat Methods. 9:671-675.

Schuster, S., M. Rivalan, U. Strauss, L. Stoenica, T. Trimbuch, N. Rademacher, S.

Parthasarathy, D. Lajko, C. Rosenmund, S.A. Shoichet, Y. Winter, V. Tarabykin,

and M. Rosario. 2015. NOMA-GAP/ARHGAP33 regulates synapse development

and autistic-like behavior in the mouse. Mol Psychiatry. 20:1120-1131.

Shaw, G. 1993. Identification of novel pleckstrin homology (PH) domains provides a

hypothesis for PH domain function. Biochem Biophys Res Commun. 195:1145-

1151.

Shen, P.C., D.F. Xu, J.W. Liu, K. Li, M. Lin, H.T. Wang, R. Wang, and J. Zheng. 2011.

TC10beta/CDC42 GTPase activating protein is required for the growth of cortical

neuron dendrites. Neuroscience. 199:589-597.

Siliceo, M., D. Garcia-Bernal, S. Carrasco, E. Diaz-Flores, F. Coluccio Leskow, J.

Teixido, M.G. Kazanietz, and I. Merida. 2006. Beta2-chimaerin provides a

diacylglycerol-dependent mechanism for regulation of adhesion and chemotaxis

of T cells. J Cell Sci. 119:141-152.

75 Simo, S., and J.A. Cooper. 2012. Regulation of dendritic branching by Cdc42 GAPs.

Genes Dev. 26:1653-1658.

Stoletov, K., and J.D. Lewis. 2015. Invadopodia: a new therapeutic target to block cancer

metastasis. Expert Rev Anticancer Ther. 15:733-735.

Sudol, M., H.I. Chen, C. Bougeret, A. Einbond, and P. Bork. 1995. Characterization of a

novel protein-binding module--the WW domain. FEBS Lett. 369:67-71.

Tapon, N., and A. Hall. 1997. Rho, Rac and Cdc42 GTPases regulate the organization of

the actin cytoskeleton. Curr Opin Cell Biol. 9:86-92.

Tarone, G., D. Cirillo, F.G. Giancotti, P.M. Comoglio, and P.C. Marchisio. 1985. Rous

sarcoma virus-transformed fibroblasts adhere primarily at discrete protrusions of

the ventral membrane called podosomes. Exp Cell Res. 159:141-157.

Tcherkezian, J., and N. Lamarche-Vane. 2007. Current knowledge of the large RhoGAP

family of proteins. Biol Cell. 99:67-86.

Toure, A., O. Dorseuil, L. Morin, P. Timmons, B. Jegou, L. Reibel, and G. Gacon. 1998.

MgcRacGAP, a new human GTPase-activating protein for Rac and Cdc42 similar

to Drosophila rotundRacGAP gene product, is expressed in male germ cells. J

Biol Chem. 273:6019-6023. van Buul, J.D., M.J. Allingham, T. Samson, J. Meller, E. Boulter, R. Garcia-Mata, and K.

Burridge. 2007. RhoG regulates endothelial apical cup assembly downstream

from ICAM1 engagement and is involved in leukocyte trans-endothelial

migration. J Cell Biol. 178:1279-1293.

76 Vigil, D., J. Cherfils, K.L. Rossman, and C.J. Der. 2010. Ras superfamily GEFs and

GAPs: validated and tractable targets for cancer therapy? Nat Rev Cancer.

10:842-857.

Weigelt, B., J.L. Peterse, and L.J. van 't Veer. 2005. Breast cancer metastasis: markers

and models. Nat Rev Cancer. 5:591-602.

Wertheimer, E., A. Gutierrez-Uzquiza, C. Rosemblit, C. Lopez-Haber, M.S. Sosa, and

M.G. Kazanietz. 2012. Rac signaling in breast cancer: a tale of GEFs and GAPs.

Cell Signal. 24:353-362.

Wishart, M.J., G.S. Taylor, and J.E. Dixon. 2001. Phoxy lipids: revealing PX domains as

phosphoinositide binding modules. Cell. 105:817-820.

Yamaguchi, H., M. Lorenz, S. Kempiak, C. Sarmiento, S. Coniglio, M. Symons, J.

Segall, R. Eddy, H. Miki, T. Takenawa, and J. Condeelis. 2005. Molecular

mechanisms of invadopodium formation: the role of the N-WASP-Arp2/3

complex pathway and cofilin. J Cell Biol. 168:441-452.

Yang, C., and M.G. Kazanietz. 2003. Divergence and complexities in DAG signaling:

looking beyond PKC. Trends Pharmacol Sci. 24:602-608.

Yang, C., and M.G. Kazanietz. 2007. Chimaerins: GAPs that bridge diacylglycerol

signalling and the small G-protein Rac. Biochem J. 403:1-12.

Zhang, Z., C. Wu, S. Wang, W. Huang, Z. Zhou, K. Ying, Y. Xie, and Y. Mao. 2002.

Cloning and characterization of ARHGAP12, a novel human rhoGAP gene. Int J

Biochem Cell Biol. 34:325-331.

Zheng, Y., M.J. Hart, K. Shinjo, T. Evans, A. Bender, and R.A. Cerione. 1993.

Biochemical comparisons of the Saccharomyces cerevisiae Bem2 and Bem3

77 proteins. Delineation of a limit Cdc42 GTPase-activating . J Biol

Chem. 268:24629-24634.

78