SH3BP1, an Exocyst-Associated RhoGAP, Inactivates Rac1 at the Front to Drive Cell Motility

The coordination of the several pathways involved in cell motility is poorly understood. Here, we identify SH3BP1, belonging to the RhoGAP family, as a partner of the exocyst complex and establish a physical and functional link between two motility-driving pathways, the Ral/exocyst and Rac signaling pathways. We show that SH3BP1 localizes together with the exocyst to the leading edge of motile cells and that SH3BP1 regulates cell migration via its GAP activity upon Rac1. SH3BP1 loss of function induces abnormally high Rac1 activity at the front, as visualized by in vivo biosensors, and disorganized and instable protrusions, as revealed by cell morphodynamics analysis. Consistently, constitutively active Rac1 mimics the phenotype of SH3BP1 depletion: slow migration and aberrant cell morphodynamics. Our finding that SH3BP1 downregulates Rac1 at the motile-cell front indicates that Rac1 inactivation in this location, as well as its activation by GEF proteins, is a fundamental requirement for cell motility. The coordination of the several pathways involved in cell motility is poorly understood. Here, we identify SH3BP1, belonging to the RhoGAP family, as a partner of the exocyst complex and establish a physical and functional link between two motility-driving pathways, the Ral/exocyst and Rac signaling pathways. We show that SH3BP1 localizes together with the exocyst to the leading edge of motile cells and that SH3BP1 regulates cell migration via its GAP activity upon Rac1. SH3BP1 loss of function induces abnormally high Rac1 activity at the front, as visualized by in vivo biosensors, and disorganized and instable protrusions, as revealed by cell morphodynamics analysis. Consistently, constitutively active Rac1 mimics the phenotype of SH3BP1 depletion: slow migration and aberrant cell morphodynamics. Our finding that SH3BP1 downregulates Rac1 at the motile-cell front indicates that Rac1 inactivation in this location, as well as its activation by GEF proteins, is a fundamental requirement for cell motility. The RhoGAP SH3BP1 interacts and colocalizes with the exocyst in motile cells SH3BP1 is required for cell migration because it inactivates Rac1-GTP at the front GDP/GTP cycling of Rac1 is needed to generate stable and organized protrusions SH3BP1 connects two motility-driving pathways: the Ral/exocyst and Rac pathways Cell motility is a highly coordinated cellular process that relies on the precise spatiotemporal integration of various pathways (Ridley et al., 2003Ridley A.J. Schwartz M.A. Burridge K. Firtel R.A. Ginsberg M.H. Borisy G. Parsons J.T. Horwitz A.R. Cell migration: integrating signals from front to back.Science. 2003; 302: 1704-1709Crossref PubMed Scopus (3607) Google Scholar), and understanding the connections among migration-regulating molecular machineries is a major challenge in cell biology. Recent studies have identified a migration-regulatory pathway that emanates from the RalB GTPase and its downstream effector complex known as the exocyst (Lim et al., 2006Lim K.H. O'Hayer K. Adam S.J. Kendall S.D. Campbell P.M. Der C.J. Counter C.M. Divergent roles for RalA and RalB in malignant growth of human pancreatic carcinoma cells.Curr. Biol. 2006; 16: 2385-2394Abstract Full Text Full Text PDF PubMed Scopus (166) Google Scholar, Oxford et al., 2005Oxford G. Owens C.R. Titus B.J. Foreman T.L. Herlevsen M.C. Smith S.C. Theodorescu D. RalA and RalB: antagonistic relatives in cancer cell migration.Cancer Res. 2005; 65: 7111-7120Crossref PubMed Scopus (99) Google Scholar, Rosse et al., 2006Rosse C. Hatzoglou A. Parrini M.C. White M.A. Chavrier P. Camonis J. RalB mobilizes the exocyst to drive cell migration.Mol. Cell. Biol. 2006; 26: 727-734Crossref PubMed Scopus (109) Google Scholar). The exocyst is comprised of eight subunits and tethers secretory vesicles to the plasma membrane (He and Guo, 2009He B. Guo W. The exocyst complex in polarized exocytosis.Curr. Opin. Cell Biol. 2009; 21: 537-542Crossref PubMed Scopus (298) Google Scholar). Although RalB is known to control the assembly and localization of exocyst subunits at the leading edge of motile cells (Rosse et al., 2006Rosse C. Hatzoglou A. Parrini M.C. White M.A. Chavrier P. Camonis J. RalB mobilizes the exocyst to drive cell migration.Mol. Cell. Biol. 2006; 26: 727-734Crossref PubMed Scopus (109) Google Scholar), how the exocyst in turn controls migration remains uncertain. One contributing mechanism was revealed recently when the exocyst complex was shown to regulate the dynamics of cell-matrix adhesion by coordinating the activities of atypical protein kinase C (aPKC) and Jun N-terminal kinase (JNK) (Rosse et al., 2009Rosse C. Formstecher E. Boeckeler K. Zhao Y. Kremerskothen J. White M.D. Camonis J.H. Parker P.J. An aPKC-exocyst complex controls paxillin phosphorylation and migration through localised JNK1 activation.PLoS Biol. 2009; 7: e1000235https://doi.org/10.1371/journal.pbio.1000235Crossref PubMed Scopus (86) Google Scholar). The exocyst is also thought to contribute to polarized delivery of regulatory molecules to the migration front, but clear experimental evidence for this is lacking. The small GTPases of the Rho family (Cdc42, Rac, Rho) regulate cell motility by controlling the dynamics of the actin cytoskeleton (Raftopoulou and Hall, 2004Raftopoulou M. Hall A. Cell migration: Rho GTPases lead the way.Dev. Biol. 2004; 265: 23-32Crossref PubMed Scopus (1099) Google Scholar). Specifically, Cdc42 is crucial for establishing front-rear polarity, Rac1 for producing networks of polymerized actin at protrusions, and RhoA for inducing actomyosin contractility. We reasoned that there should exist some regulatory mechanisms connecting the Ral/exocyst signaling pathway to the action driven by Rho family GTPases. In pursuing this possibility, we identified a molecular link for the coordination between Ral and Rac during migration: the RhoGAP SH3BP1, which partners with the exocyst complex to spatially restrict Rac1 activity. Specifically, SH3BP1 inhibits Rac1 activity by promoting the hydrolysis of bound GTP to GDP, and failure of this Rac1 inactivation leads to anarchic protrusions and ineffective migration. SH3BP1 (SH3-domain binding protein 1, also known as 3BP-1; NP_061830) was identified in a series of yeast two-hybrid screens aimed at identifying partners of the eight subunits of the exocyst complex; it was found to bind to both the Exo84 and Sec8 subunits. SH3BP1 contains an N-terminal BAR (Bin-Amphiphysin-Rvs) domain (putatively involved in protein-protein interactions and in binding to curved membranes), a central RhoGAP domain, and a C-terminal tail with several proline-rich sequences (Cicchetti et al., 1992Cicchetti P. Mayer B.J. Thiel G. Baltimore D. Identification of a protein that binds to the SH3 region of Abl and is similar to Bcr and GAP-rho.Science. 1992; 257: 803-806Crossref PubMed Scopus (416) Google Scholar). The smallest fragments of SH3BP1 recovered from the screens were amino acids 79–255 for the Exo84 interaction and amino acids 65–257 for the Sec8 interaction; thus the putative exocyst-interacting domain is in the N-terminal BAR-domain of SH3BP1 (Figure 1A ). The SH3BP1-exocyst interaction appears to be specific since (1) SH3BP1 was the only BAR-containing protein identified in the exocyst screens, in spite of the fact that ten other BAR-containing proteins have been identified in other screens of the same library, and (2) the same SH3BP1 region was found in only one other screen, although more than 1200 screens have been performed on the same library with unrelated bait proteins (data not shown). We demonstrated an association in vivo between SH3BP1 and the exocyst by showing that overexpressed full-length SH3BP1 coimmunoprecipitated with endogenous Sec8. When the BAR domain was deleted, this association was abolished, confirming that this domain is required for the interaction between SH3BP1 and the exocyst (Figure 1C). In these experiments, only a small fraction of SH3BP1 associated with the endogenous exocyst, possibly because of the large excess of overexpressed SH3BP1 or because of the highly regulated nature of the interaction. In addition, by overexpressing Exo84 or Sec8 along with various SH3BP1 forms (Figures 1B), we found that both of these exocyst components can interact with full-length SH3BP1 and with a SH3BP1 form lacking the C-terminal tail, but not with a SH3BP1 form lacking the N-terminal BAR-domain (Figures 1D and 1E). Thus, both yeast two-hybrid and coimmunoprecipitation studies indicate that SH3BP1 binds to Exo84 and Sec8, and that these interactions are mediated by the SH3BP1 BAR-domain. We did not succeed in coimmunoprecipitating endogenous exocyst subunits with endogenous SH3BP1. We reason that if the two proteins interact only very locally, such an association would be masked by the vast excess of nonassociated exocyst and SH3BP1. We studied the localization of SH3BP1 in migrating cells. In normal rat kidney (NRK) cells, endogenous SH3BP1 was most prominently localized at the leading edge of motile cells (Figure 2A ). This staining at the leading edge was specific, was reproducible using three different antibodies (see Figure S1 available online), and was strongly reduced in cells depleted of SH3BP1 by RNAi (Figure 2B and Figure S1). Confocal microscopy analysis of cells costained for SH3BP1 and Exo84 or Sec8 showed extensive regions of colocalization at the leading edges of migrating cells, but not elsewhere in the plasma membrane or in the cytoplasm (Figure 2C). We confirmed that SH3BP1 also colocalized with the exocyst (Sec15 subunit) at the leading front in a second model for cell motility: wound healing in cultures of the human PC-3 prostate tumor cell line (Figure 2D). Depletion of exocyst components (Exo84, Sec8, Sec5) by RNAi reduced recruitment of SH3BP1 to the leading edge (Figure 2B) and, conversely, depletion of SH3BP1 reduced recruitment of the exocyst (Sec8 subunit) to the leading edge (Figure 2E), indicating that localization is mutually dependent. Taken together, these results are consistent with SH3BP1 and the exocyst being associated at the leading front, where SH3BP1 might contribute to the molecular machinery that is responsible for migration. While exogenously expressed full-length SH3BP1 localized to the front as does endogenous protein, the SH3BP1 ΔBAR construct was not recruited to the leading edge (Figures 2F and 2G), suggesting that the BAR domain and potentially the interaction with exocyst are required to transport SH3BP1 to the front. We assessed the contribution of SH3BP1 to cell migration by selectively depleting the SH3BP1 product via an RNAi approach. We first carried out wound-healing assays using NRK cells, and found that SH3BP1 depletion (>85%, with two independent siRNAs) (Figure 3B ) strongly inhibited wound closure (Figures 3A and Movie S1). Cell-tracking analysis indicated that average speed of cell migration during wound closure was reduced by 30%–45% in this context, and also revealed a slight but significant (10%–15%) decrease in persistence of migration (Figure 3C). SH3BP1 depletion also resulted in robust inhibition of wound closure in other cell lines, including HEK-HT (human embryonic kidney) and RPE1 (human retinal pigment epithelial) cells (data not shown), suggesting that the role of SH3BP1 in regulating cell migration is general and conserved. Consistent with the wound-healing assays, Boyden chamber assays revealed that cells treated with a siRNA against SH3BP1 were defective in migration (Figure 3D). This defect was corrected by subsequent DNA transfection with a vector expressing a human form of SH3BP1 that is resistant to the SH3BP1 siRNA. In contrast to the expression of wild-type SH3BP1, expression of the R312A SH3BP1 mutant, whose GAP activity is impaired due to substitution of the critical “arginine finger” in the GAP domain (Bos et al., 2007Bos J.L. Rehmann H. Wittinghofer A. GEFs and GAPs: critical elements in the control of small G proteins.Cell. 2007; 129: 865-877Abstract Full Text Full Text PDF PubMed Scopus (1136) Google Scholar) (Figure S2A), did not lead to the rescue of normal motility (Figure 3D). We confirmed these results by tracking cells expressing cherry-fused alleles of SH3BP1 in the context of a wounded monolayer where endogenous SH3BP1 had been depleted or not by RNAi. Speeds of cells in the different genetic contexts are listed in Figure 3E. Cells where endogenous SH3BP1 was depleted, but which expressed a siRNA-resistant SH3BP1, migrated as fast as control cells. This was not the case when a siRNA-resistant SH3BP1 R312A mutant was expressed, confirming the requirement of the GAP activity of SH3BP1. When the SH3BP1 ΔBAR allele was expressed, SH3BP1-depleted cells did not recover normal speed, pointing out the importance of the BAR domain for SH3BP1 function during migration. The isolated RhoGAP domain of SH3BP1 displays GAP activity toward Cdc42 and Rac1, but not toward RhoA (Cicchetti et al., 1995Cicchetti P. Ridley A.J. Zheng Y. Cerione R.A. Baltimore D. 3BP-1, an SH3 domain binding protein, has GAP activity for Rac and inhibits growth factor-induced membrane ruffling in fibroblasts.EMBO J. 1995; 14: 3127-3135Crossref PubMed Scopus (62) Google Scholar), in vitro. We obtained similar results in vivo by expressing the full-length SH3BP1 protein together with FRET-based Raichu probes (Nakamura et al., 2006Nakamura T. Kurokawa K. Kiyokawa E. Matsuda M. Analysis of the spatiotemporal activation of rho GTPases using Raichu probes.Methods Enzymol. 2006; 406: 315-332Crossref PubMed Scopus (59) Google Scholar) (tools with which to monitor the activity of Cdc42, Rac1, and RhoA; Figures S3A and S3B). Consistent with the results of previous pull-down studies (Lu and Mayer, 1999Lu W. Mayer B.J. Mechanism of activation of Pak1 kinase by membrane localization.Oncogene. 1999; 18: 797-806Crossref PubMed Scopus (69) Google Scholar), we observed a small preference for Rac1 rather than Cdc42 as substrate. We questioned whether the physiological target of SH3BP1 in motile cells is Cdc42, Rac1, or both. Reorientation of the microtubule-organizing center (MTOC) and Golgi apparatus in front of the nucleus serves as a landmark of cell polarization during migration and is controlled by Cdc42 activity (Etienne-Manneville, 2004Etienne-Manneville S. Cdc42—the centre of polarity.J. Cell Sci. 2004; 117: 1291-1300Crossref PubMed Scopus (515) Google Scholar, Ridley et al., 2003Ridley A.J. Schwartz M.A. Burridge K. Firtel R.A. Ginsberg M.H. Borisy G. Parsons J.T. Horwitz A.R. Cell migration: integrating signals from front to back.Science. 2003; 302: 1704-1709Crossref PubMed Scopus (3607) Google Scholar). We confirmed that, as in previous studies (Osmani et al., 2006Osmani N. Vitale N. Borg J.P. Etienne-Manneville S. Scrib controls Cdc42 localization and activity to promote cell polarization during astrocyte migration.Curr. Biol. 2006; 16: 2395-2405Abstract Full Text Full Text PDF PubMed Scopus (162) Google Scholar), the depletion of RhoGEF β-PIX strongly inhibited MTOC orientation in our cells under our experimental conditions. MTOC repositioning in front of the nucleus was not affected by the depletion of RalB, exocyst components (Sec5, Sec8, Exo84), or SH3BP1 (Figure 4). Western blot analysis confirmed that each of the tested proteins was efficiently depleted (Figure S3C and Figure 3B). Moreover, expression of SH3BP1 ΔBAR mutant, which is defective in exocyst binding, did not perturb the MTOC orientation (Figure S3D), indicating that interaction with exocyst is not sufficient for SH3BP1 to select Rac over Cdc42 as substrate. We conclude that the RalB/exocyst/SH3BP1 pathway does not regulate cell migration by controlling MTOC reorientation, and that it does not control Cdc42 activity in motile cells. It remains unclear how SH3BP1 constrains its biochemical activity on Rac1 in the context of cell motility. We directly tested the hypothesis that Rac1 is the SH3BP1 target relevant to cell migration, using a FRET approach with Raichu biosensors to monitor the spatiotemporal activation of Rac1 in living motile cells. In analyzing the first row of motile NRK cells that express Raichu-Rac1 (KRasCter; targeted to the plasma membrane via the K-Ras C-terminal tail) (Itoh et al., 2002Itoh R.E. Kurokawa K. Ohba Y. Yoshizaki H. Mochizuki N. Matsuda M. Activation of rac and cdc42 video imaged by fluorescent resonance energy transfer-based single-molecule probes in the membrane of living cells.Mol. Cell. Biol. 2002; 22: 6582-6591Crossref PubMed Scopus (438) Google Scholar, Nakamura et al., 2006Nakamura T. Kurokawa K. Kiyokawa E. Matsuda M. Analysis of the spatiotemporal activation of rho GTPases using Raichu probes.Methods Enzymol. 2006; 406: 315-332Crossref PubMed Scopus (59) Google Scholar) (starting acquisition and analysis 3 hr postwounding), we observed that Rac1 was activated in a dynamic gradient that grew more intense toward the leading edge, as previously reported for other cell types (Itoh et al., 2002Itoh R.E. Kurokawa K. Ohba Y. Yoshizaki H. Mochizuki N. Matsuda M. Activation of rac and cdc42 video imaged by fluorescent resonance energy transfer-based single-molecule probes in the membrane of living cells.Mol. Cell. Biol. 2002; 22: 6582-6591Crossref PubMed Scopus (438) Google Scholar, Kraynov et al., 2000Kraynov V.S. Chamberlain C. Bokoch G.M. Schwartz M.A. Slabaugh S. Hahn K.M. Localized Rac activation dynamics visualized in living cells.Science. 2000; 290: 333-337Crossref PubMed Scopus (551) Google Scholar). Control nonfunctional Raichu probes did not present a FRET gradient pattern and did not show any variation upon SH3BP1 silencing (Figure S4A). In cells treated with siSH3BP1, Rac1 appeared to be activated more strongly and across a broader area (Figure 5A and Movie S2). We quantitatively compared FRET signals on the total cellular area in control cells (n = 33) and SH3BP1-depleted cells (n = 28). The average Rac1 activity was only slightly higher in SH3BP1-depleted cells than in control cells (p = 0.07, Student's t test) (Figure 5B), indicating that the loss of SH3BP1 did not significantly affect the global levels of active Rac1. However, when we compared FRET signals spatially along a line going from the nucleus to the leading edge, we found that the average Rac1 gradient was significantly upregulated in siSH3BP1-treated cells compared to control cells (Figure 5C). These measurements indicate that SH3BP1 loss induced a localized, rather than global, defect in Rac1 activity. We obtained very similar results with another probe, Raichu-Rac1 (Rac1Cter), which includes the C terminus of Rac1 and therefore better reflects the localization of endogenous Rac1 (Figures S4B–S4D). Importantly, the spatial defect in Rac1 activity of SH3BP1-depleted cells was rescued by expression of wild-type SH3BP1, but not of the R312A GAP-defective mutant (Figure 5D). The action of SH3BP1 at the leading edge appears rather specific, since silencing of two other GAP proteins, p190RhoGAP and RLIP76/RalBP1, acting on Rac1 (Cantor et al., 1995Cantor S.B. Urano T. Feig L.A. Identification and characterization of Ral-binding protein 1, a potential downstream target of Ral GTPases.Mol. Cell. Biol. 1995; 15: 4578-4584Crossref PubMed Scopus (252) Google Scholar, Jullien-Flores et al., 1995Jullien-Flores V. Dorseuil O. Romero F. Letourneur F. Saragosti S. Berger R. Tavitian A. Gacon G. Camonis J.H. Bridging Ral GTPase to Rho pathways. RLIP76, a Ral effector with CDC42/Rac GTPase-activating protein activity.J. Biol. Chem. 1995; 270: 22473-22477Crossref PubMed Scopus (279) Google Scholar, Ligeti et al., 2004Ligeti E. Dagher M.C. Hernandez S.E. Koleske A.J. Settleman J. Phospholipids can switch the GTPase substrate preference of a GTPase-activating protein.J. Biol. Chem. 2004; 279: 5055-5058Crossref PubMed Scopus (54) Google Scholar), did not perturb the Rac1 activity gradient (Figure 5E), despite their efficient depletion (Figure 5F). Taking these results together, we conclude that the GAP SH3BP1 is required to locally downregulate Rac1 at the leading front. Since Rac1 activity intimately regulates the formation of protrusions (Pankov et al., 2005Pankov R. Endo Y. Even-Ram S. Araki M. Clark K. Cukierman E. Matsumoto K. Yamada K.M. A Rac switch regulates random versus directionally persistent cell migration.J. Cell Biol. 2005; 170: 793-802Crossref PubMed Scopus (356) Google Scholar, Raftopoulou and Hall, 2004Raftopoulou M. Hall A. Cell migration: Rho GTPases lead the way.Dev. Biol. 2004; 265: 23-32Crossref PubMed Scopus (1099) Google Scholar, Ridley et al., 2003Ridley A.J. Schwartz M.A. Burridge K. Firtel R.A. Ginsberg M.H. Borisy G. Parsons J.T. Horwitz A.R. Cell migration: integrating signals from front to back.Science. 2003; 302: 1704-1709Crossref PubMed Scopus (3607) Google Scholar, Wu et al., 2009Wu Y.I. Frey D. Lungu O.I. Jaehrig A. Schlichting I. Kuhlman B. Hahn K.M. A genetically encoded photoactivatable Rac controls the motility of living cells.Nature. 2009; 461: 104-108Crossref PubMed Scopus (752) Google Scholar), we asked whether abnormally upregulated Rac1 activity has an impact on protrusion dynamics in SH3BP1-depleted cells. We addressed this question by using a membrane-targeted fluorescent RFP protein (RFP-CAAX) to visualize movement of the cell periphery, and comparing videos of migrating SH3BP1-depleted and control cells (Figure 6A and Movie S3). Visual inspection of several movies revealed that the large majority of control cells had only one dynamic, well-defined lamellipodium toward the front. The protruding front in siSH3BP1-transfected cells was longer than that in control cells (average distance between nucleus and leading edge, 27.5 μm, SEM = 1.6 μm versus 21.3 μm, SEM = 1.3 μm; p < 0.005, Student's t test) and was often discontinuous, forming multiple subprotrusions. As a consequence, the average number of active protrusions per cell was higher in SH3BP1-depleted cells than in control cells (1.59 versus 1.17; p < 0.005, Student's t test), and the percentage of ectopic protrusions (those not within the 120° angle facing the wound) increased from 12% in control cells to 22% in SH3BP1-depleted cells (Figure 6B). To objectively examine the role of SH3BP1 in plasma membrane morphodynamics, we performed a computer-assisted analysis of time-lapse images. We first measured velocities of movements at the cell edge, generating velocity maps for control and SH3BP1-depleted cells (Figures 6C and 6E). Edge fractions were classified into three categories: retracting, when velocity was lower than −2.5 μm/10 min; static, when velocity was between −2.5 μm/10 min and +2.5 μm/10 min; and protruding, when velocity was higher than +2.5 μm/10 min. We found significant increases in both the retracting and protruding edge fractions in siSH3BP1-treated cells, indicating that the leading edges of SH3BP1-depleted cells moved faster than those of control cells (Figure 6G). We also quantified the stability of membrane dynamics at the cell edge by using the mathematical tool of autocorrelation analysis as previously reported (Dobereiner et al., 2006Dobereiner H.G. Dubin-Thaler B.J. Hofman J.M. Xenias H.S. Sims T.N. Giannone G. Dustin M.L. Wiggins C.H. Sheetz M.P. Lateral membrane waves constitute a universal dynamic pattern of motile cells.Phys. Rev. Lett. 2006; 97: 038102Crossref Scopus (128) Google Scholar, Maeda et al., 2008Maeda Y.T. Inose J. Matsuo M.Y. Iwaya S. Sano M. Ordered patterns of cell shape and orientational correlation during spontaneous cell migration.PLoS ONE. 2008; 3: e3734https://doi.org/10.1371/journal.pone.0003734Crossref PubMed Scopus (91) Google Scholar). Autocorrelation is the spatiotemporal cross-correlation of the signal (the edge velocity in our case) with itself. We generated autocorrelation maps of the edge velocity as a function of time (minutes of the videos of the moving cells) and of space (along the cell periphery). Whereas control cells exhibited an ordered pattern as expected for typical directional migration (Figure 6D), SH3BP1-depleted cells showed a much less ordered pattern (Figure 6F), indicating that protrusion formation during migration of the latter cells is random and instable. Plotting temporal cuts of autocorrelation maps (ΔCell periphery = 0 degrees; broken white arrows in Figures 6D and 6F) revealed that, although autocorrelation coefficients were fairly stable over time in the case of control cells, they decayed rapidly in the case of SH3BP1-depleted cells (Figure 6H). Thus, the persistence of protrusion and retraction dynamics in SH3BP1-depleted cells was significantly reduced compared to those in control cells. We could correct the morphodynamics defect of SH3BP1-depleted cells by expressing wild-type SH3BP1, but not the R312A GAP-defective mutant, as shown by an analysis of the autocorrelation coefficients at Δt = 10 min (Figure 6I and Figure S5). This rescue result strongly supports a direct role of SH3BP1 GAP activity in the stability of membrane dynamics of motile cells. Since in comparison to protrusions in control cells those in SH3BP1-depleted cells are more numerous, partially delocalized, and developed more rapidly but with a shorter persistence, we conclude that SH3BP1 is necessary for the spatiotemporal organization of protrusions during cell migration. If inactivation of Rac1 at the front by SH3BP1 is required to organize protrusions and to drive efficient directional motility, cells expressing GTP-hydrolysis-deficient RacG12V should present defects in motility and morphodynamics. We tested this prediction by expressing RFP-fused wild-type or constitutively active mutant G12V Rac1 in motile NRK cells. Cells expressing RacG12V were severely impaired in motility, while cells expressing wild-type Rac1 migrated as control cells (untransfected or RFP-only transfected) (Figures 7A and 7B ). RacG12V-expressing cells displayed a much more rapid drop of the autocorrelation coefficients, as compared with wild-type Rac-expressing cells, indicating that GTP hydrolysis is necessary for the persistence of membrane dynamics (Figures 7C). The fast-cycling F28L mutant showed a similar but milder phenotype, with inhibition of motility (Figure 7B) and perturbation of morphodynamics (Figures 7C). Since RacG12V is always loaded with GTP, while RacF28L constitutively cycles between GDP- and GTP-bound states, their different phenotype strength could simply reflect their different GTP-loading levels. Moreover, since RacF28L-expressing cells presented abnormalities, we can conclude that GDP/GTP cycling per se is not sufficient to drive efficient protrusion organization. It seems that Rac not only needs to cycle but needs to cycle at the right place and with the right kinetics, as spatiotemporally dictated by GEF and GAP proteins. Thus, expression of constitutively active Rac1 mutants mimicked the phenotype of SH3BP1 depletion, clearly pointing out the importance of inactivating Rac1 during the motility process. We report here that SH3BP1 is a GAP that stimulates the GTPase activity of Rac1 during migration. However, several GEF proteins, including βPIX (ten Klooster et al., 2006ten Klooster J.P. Jaffer Z.M. Chernoff J. Hordijk P.L. Targeting and activation of Rac1 are mediated by the exchange factor beta-Pix.J. Cell Biol. 2006; 172: 759-769Crossref PubMed Scopus (202) Google Scholar), DOCK3 (Sanz-Moreno et al., 2008Sanz-Moreno V. Gadea G. Ahn J. Paterson H. Marra P. Pinner S. Sahai E. Marshall C.J. Rac activation and inactivation control plasticity of tumor cell movement.Cell. 2008; 135: 510-523Abstract Full Text Full Text PDF PubMed Scopus (680) Google Scholar), Asef (Itoh et al., 2008Itoh R.E. Kiyokawa E. Aoki K. Nishioka T. Akiyama T. Matsuda M. Phosphorylation and activation of the Rac1 and Cdc42 GEF Asef in A431 cells stimulated by EGF.J. Cell Sci. 2008; 121: 2635-2642Crossref PubMed Scopus (52) Google Scholar), and Tiam1 (Palamidessi et al., 2008Palamidessi A. Frittoli E. Garre M. Faretta M. Mione M. Testa I. Diaspro A. Lanzetti L. Scita G. Di Fiore P.P. Endocytic trafficking of Rac is required for the spatial restriction of signaling in cell migration.Cell. 2008; 134: 135-147Abstract Full Text Full Text PDF PubMed Scopus (334) Google Scholar), have been proposed to activate Rac1 at the leading edge by replacing bound GDP with GTP. Thus, the presence of a Rac1-specific GAP at the leading edge—where Rac1 needs to be active—may seem surprising. One might have expected to instead find a RacGAP at the back of the cell in order to maintain the back-to-front Rac1-GTP gradient. However, current evidence supports the concept that unlike the prototype Ras, which needs to be locked into the GTP-bound state to execute its biological functions, other small GTPases require GDP/GTP cycling (Barale et al., 2006Barale S. McCusker D. Arkowitz R.A. Cdc42p GDP/GTP cycling is necessary for efficient cell fusion during yeast mating.Mol. Biol. Cell. 2006; 17: 2824-2838Crossref PubMed Scopus (22) Google Scholar, Lin et al., 1997Lin R. Bagrodia S. Cerione R. Manor D. A novel Cdc42Hs mutant induces cellular transformation.Curr. Biol. 1997; 7: 794-797Abstract Full Text Full Text PDF PubMed Scopus (180) Google Scholar, Miller and Bement, 2009Miller A.L. Bement W.M. Regulation of cytokinesis by Rho GTPase flux.Nat. Cell Biol. 2009; 11: 71-77Crossref PubMed Scopus (176) Google Scholar). This could explain why transformation induced by GEFs specific to Rho family GTPases is much more efficient than transformation induced by constitutively active Rho family GTPases, and also why GTPase-defective Rho/Rac/Cdc42 mutants have never been found in human tumors (Karlsson et al., 2009Karlsson R. Pedersen E.D. Wang Z. Brakebusch C. Rho GTPase function in tumorigenesis.Biochim. Biophys. Acta. 2009; 1796: 91-98Crossref PubMed Scopus (266) Google Schola