The interaction of Gα13 with integrin β1 mediates cell migration by dynamic regulation of RhoA.

Heterotrimeric G protein Gα13 is known to transmit G protein-coupled receptor (GPCR) signals leading to activation of RhoA and plays a role in cell migration. The mechanism underlying the role of Gα13 in cell migration, however, remains unclear. Recently we found that Gα13 interacts with the cytoplasmic domain of integrin β3 subunits in platelets via a conserved ExE motif. Here we show that a similar direct interaction between Gα13 and the cytoplasmic domain of the integrin β1 subunit plays a critical role in β1-dependent cell migration. Point mutation of either glutamic acid in the Gα13-binding (767)EKE motif in β1 or treatment with a peptide derived from the Gα13-binding sequence of β1 abolished Gα13-β1 interaction and inhibited β1 integrin-dependent cell spreading and migration. We further show that the Gα13-β1 interaction mediates β1 integrin-dependent Src activation and transient RhoA inhibition during initial cell adhesion, which is in contrast to the role of Gα13 in mediating GPCR-dependent RhoA activation. These data indicate that Gα13 plays dynamic roles in both stimulating RhoA via a GPCR pathway and inhibiting RhoA via an integrin signaling pathway. This dynamic regulation of RhoA activity is critical for cell migration on β1 integrin ligands.

INTRODUCTION

G protein–coupled receptors (GPCRs) transmit signals via the heterotrimeric G proteins and are important, working in concert with β1 integrins, for directed cell migration (Sakai , b, 1999). Gα13 is among the G proteins important in the directed cell migration and chemotaxis in various cell types (Goulimari ; Bian ; Rieken ). It has been believed that this role of Gα13 is mediated via the Gα13-dependent activation of RhoGEFs, including p115RhoGEF (ARHGEF1), which subsequently activates the small GTPase RhoA (Kozasa ; Suzuki ; Radhika ; Goulimari ; Kreutz ). RhoA activates actomyosin–mediated contraction through inhibition of myosin light chain phosphatases (Kimura ) and increases actin polymerization through activation of formin (mDia1), events that are important for cell migration (Goulimari ).

The integrin family of adhesion receptors mediates cell adhesion and migration on extracellular matrices and is important in many physiological and pathological processes, such as development, wound healing, immunity, thrombosis, and cancer metastasis (Hynes, 2002; Ley ; Li ; Margadant and Sonnenberg, 2010; Huttenlocher and Horwitz, 2011; Seguin ). In particular, the widely expressed β1 integrins are critical for regulating anchorage-dependent cell survival, proliferation, and migration on extracellular matrices. Integrins transmit signals bidirectionally: high-affinity ligand binding to integrins can be activated by “inside-out” signaling (Ginsberg ), whereas ligand binding induces “outside-in” signaling (Shattil, 2005). Outside-in signaling elicits a cascade of intracellular signaling events, resulting in two essential cellular responses to the extracellular matrix proteins: cell spreading (formation of protrusions, filopodia and lamellipodia) and cell retraction, both of which are important in driving cell migration on extracellular matrices (Huttenlocher and Horwitz, 2011; Shen ). It is unclear how integrins signal to coordinate cell spreading and retraction and mediate cell migration. It is also unclear whether there is a connection between the roles of Gα13 and integrins in cell migration.

Recently we found that the platelet integrin β3 subunit binds directly to Gα13, and the Gα13–β3 binding plays a critical role in integrin αIIbβ3–mediated outside-in signaling, cell spreading, and amplification of thrombus formation (Gong ; Shen ). Of interest, Gα13 binding to the integrin β3 subunit requires an ExE motif that is highly conserved among most integrin β subunits, including β1 (Shen ). Thus we hypothesized that Gα13 might also be important in mediating signaling of β1 integrins.

In this study, we demonstrate that the interaction of Gα13 with integrin β1 plays a critical role not only in β1 integrin–mediated cell spreading, but also in β1-dependent cell migration on the extracellular matrix. Furthermore, we show that, in contrast to the previously reported role of Gα13 in the activation of RhoGEF/RhoA pathway (Kozasa ; Suzuki ), the Gα13–β1 interaction mediates transient inhibition of RhoA signaling pathway through activation of Src (Arthur ). Thus our studies suggest a new mechanism in which the dynamic regulation of the RhoA signaling pathway by Gα13 occurs through its binding to RhoGEF GTPases and β1 integrins. This novel mechanism is critical for β1 integrin–dependent cell migration.

RESULTS

The importance of Gα13 in cell migration

To assess the role of Gα13 in cell migration, we developed two Gα13-specific short hairpin RNAs (shRNAs) and their corresponding nonspecific scrambled control shRNA using a lentiviral vector. Transfection of the Gα13-specific shRNA, but not the scrambled control, resulted in ∼90% knockdown of Gα13 in CHO cells (Figure 1A), which express endogenous β1 and β5 integrins but not β3 integrins (Ylanne ). We used a scratched wound–healing assay to assess the role of Gα13 in cell migration. As shown in Figure 1B, 20 h after the scratch, CHO cells and CHO cells transfected with control shRNA had migrated to almost seal the gap. In contrast, Gα13-knockdown cells were defective in migration (Figure 1, B and C).

FIGURE 1:

The importance of Gα13 in cell migration. (A) Western blot comparison of Gα13 expression levels in CHO cells and CHO cells transfected with Gα13-specific shRNA #1, shRNA #2, or scrambled control shRNA lentivirus. α-Tubulin was used as loading control. Quantitative data are shown as mean ± SD. (B) Phase contrast images of CHO cells before and after 20-h migration following wound scratches. (C) Quantification of data as shown in B; n = 6. (D) Total cell counts at 24-h time point (five experiments). Cell count at 0 h is 0.5 × 103.

The process of scratched wound healing consists of both cell proliferation and cell migration. To exclude the possibility that the phenotypes we observed were attributed to suppression of cell proliferation, we demonstrated that CHO cells with Gα13 knockdown were not defective in cell proliferation, but instead cell proliferation was slightly increased compared with control (Figure 1D). Thus these data indicate that Gα13 plays an important role in migration of CHO cells.

The role of Gα13 in β1 integrin–dependent cell migration

To determine specifically the role of Gα13 in β1 integrin–dependent cell migration, we analyzed cell migration in GD25 cells, an embryonic stem cell–derived fibroblast-like cell line originated from β1 integrin–knockout mice (Wennerberg ), and GD25 cells transfected with wild-type (Wt) or mutant integrin β1 subunits (Figure 2A).

FIGURE 2:

The role of Gα13 in integrin β1–dependent cell migration. (A) Flow cytometric analysis of β1 expression on GD25 cell surface after lentiviral transfection of various integrin β1 mutants. Similar expression levels of these mutants were achieved by cell sorting with FITC-conjugated anti-β1 antibody K20. β1− GD25 cells were used as negative control, and the quantification of the relative fold changes of K20 staining is shown (n = 3). (B) Phase contrast images of β1(Wt)GD25 cells, with or without Gα13-specific or scrambled control shRNA lentiviral transfection, before and after 20 h migration following wound scratches. (C) Western blot comparison of Gα13 expression levels in β1(Wt)GD25 cells and β1(Wt)GD25 cells transfected with Gα13-specific or scrambled control shRNA. α-Tubulin was used as loading control. Quantitative data are shown in the bar graph (mean ± SD, n = 3). (D) Quantification of data as shown in B (mean ± SD, n = 6). (E) Total cell count at 24-h time point (five experiments). Cell count at 0 h is 0.5 × 103.

The β1− GD25 cells displayed severely impaired migration in a scratched wound–healing assay after 20 h (Figure 2B). In contrast, Expression of β1 integrin in GD25 cells corrected the defective migration (Figure 2B). These data are consistent with a previous study showing that β1 expression in β1-knockout cell lines enhanced cell migration (Sakai , b, 1999; Gimond ), indicating that GD25 cell migration in the scratched wound–healing analysis is dependent upon integrin β1. Moreover, knockdown of Gα13 in the β1-expressing GD25 cells using either of the two Gα13 shRNAs caused significant defects in cell migration (Figure 2, B– D), indicating that Gα13 plays a critical role in β1-integrin–mediated cell migration. Similar to the aforementioned experiments using CHO cells, Gα13 knockdown did not inhibit, but instead mildly increased cell proliferation compared with control (Figure 2E), indicating that the defective wound healing in Gα13-knockdown cells was not due to inhibited cell growth but instead to suppressed cell migration.

The interaction of Gα13 with β1 integrins via the ExE motif

We previously showed that Gα13 directly interacts with integrin β subunits via a conserved ExE motif (Shen ). To determine whether Gα13 interacts with β1 via the 767EKE sequence, we incubated purified Gα13 with Sepharose beads conjugated with purified glutathione bead–bound glutathione S-transferase–Wt β1 integrin cytoplasmic domain fusion protein (GST-β1CD) or with GST-β1CD carrying mutations that change the 767EKE sequence to AKA or AAA. Gα13 binds directly to the GST-β1CD Sepharose beads, whereas Gα13 failed to bind to the AKA or AAA mutants (Figure 3A). Thus the ExE motif is crucial in the direct binding of Gα13 to integrin β1CD. To study whether the interaction between intact β1 subunit and Gα13 occurs in cells, we also introduced a set of mutations in the 767EKE sequences changing EKE to AKE, EKA, AKA, or AAA and expressed these mutants in GD25 cells at similar levels (Figure 2A). We then performed coimmunoprecipitation experiments using these cells. There was little integrin β1 associated with Gα13 in suspended cells (Figure 3B). However, after cell adhesion on fibronectin for 1 h, the interaction becomes prominent (Figure 3B). This suggests that integrin ligation is required for the binding of Gα13 to β1 in these cells. Furthermore, AKE or EKA mutants showed significantly reduced Gα13–β1 association, and mutation of the glutamic residues (AKA or AAA) almost totally abolished the interaction (Figure 3B). Thus the β1 integrin ExE motif is critical for Gα13–β1 interaction in cells. Consistent with previous results in β3 integrins (Shen ), AAA mutation in β1 did not inhibit talin association with β1 integrin, which was in fact slightly increased (Figure 3C), suggesting that AAA mutation selectively abolished Gα13–β1 interaction without inhibiting talin binding.

FIGURE 3:

The critical role of the β1 ExE motif in the Gα13–β1 interaction. (A) Binding of purified recombinant Gα13 to glutathione bead–bound glutathione S-transferase (GST), GST-Wt β1 cytoplasmic domain fusion protein (GST-β1Wt), GST-(AKA)β1, and GST-(AAA)β1 mutants. (B) β1GD25 cells expressing similar levels of wild-type (Wt) β1 and various ExE β1 mutants (AKE, EKA, AKA, or AAA) were allowed to adhere to fibronectin for 1 h. After 1 h, cells were lysed, and cell lysates were immunoprecipitated with anti-Gα13 antibody or an equal amount of control rabbit IgG. Immunoprecipitates were immunoblotted with anti-Gα13 and anti-β1 antibodies. (C) β1GD25 cells expressing similar levels of Wt and AAA mutant β1 were allowed to adhere to fibronectin. After 1 h, cells were lysed, and cell lysates were immunoprecipitated with integrin β1 antibody or IgG. Immunoprecipitates were immunoblotted with anti-talin and anti-β1 antibodies.

The interaction between Gα13 and β1 integrins mediates β1-dependent cell migration

To determine the importance of Gα13–β1 interaction in β1-dependent cell migration, we used the scratched wound–healing assay to examine the migration of GD25 cells expressing Wt β1 or various β1 ExE motif mutants. The β1/Gα13-dependent cell migration was partially abolished in cells with AKE mutation in integrin β1 and was totally abolished in cells expressing AKA, AAA, or EKA mutant (Figure 4, A and B). Mutations in the β1 ExE motif did not affect proliferation of these β1-expressing cells (Figure 4C). Thus it appears that Gα13–integrin interaction is responsible for the role of Gα13 in cell migration, and Gα13–β1 interaction is critically important for β1-dependent cell migration.

FIGURE 4:

The interaction between Gα13 and β1 integrins mediates β1-dependent migration. (A) Phase contrast images of β1GD25 cells expressing similar levels of wild-type (Wt) β1 or various ExE β1 mutants (AKE, EKA, AKA, or AAA), before and after 20 h migration following wound scratches. (B) Quantification of migration data in A (mean ± SD, n = 6). (C) Total cell count at 24-h time point (five experiments). Cell count at 0 h is 0.5 × 103.

To determine the role of Gα13–integrin interaction in cell migration under different conditions, we also used a transwell migration assay. The integrin β1 ligand fibronectin was coated on the bottom side of a transwell insert with 8-μm pores, which allow cells on the top side to migrate through the pores to the bottom side of the insert. As shown in Figure 5, transwell migration of GD25 cells requires β1 integrin expression. Gα13 knockdown abolished transwell migration almost completely. Similar to the scratched wound–healing assay, the AKE mutant of β1 only partially supported transwell migration, and the EKA, AKA, and AAA mutants had very little activity (Figure 5, A and B). Furthermore, AAA mutation or Gα13 knockdown significantly inhibited transwell migration when fibronectin coating was replaced with collagen coating (Figure 5C). On the basis of these results, we conclude that Gα13–β1 interaction is required for the β1-dependent cell migration.

FIGURE 5:

The interaction between Gα13 and β1 integrins mediates β1-dependent transwell migration. (A) GD25 cells without or with expression of similar levels of wild-type (Wt) β1 and various ExE β1 mutants (AKE, EKA, AKA, or AAA) or β1(Wt)GD25 cells transfected with scrambled or Gα13-specific shRNA were compared in a transwell migration assay with fibronectin coated on the bottom side of the insert well. Cells were fixed and stained with crystal violet after 6 h of migration. (B) Quantification of migrated cells in A (mean ± SD, four experiments). (C) Wild-type or AAA mutant β1GD25 cells or Wt β1GD25 cells transfected with scrambled or Gα13-specific shRNA were compared in the transwell migration assay as described in A, with the exception of collagen coating on the bottom side. Cells were fixed and stained with crystal violet after 6 h of migration. Numbers of migrated cells were compared (mean ± SD, three experiments).

The role of Gα13–β1 interaction in mediating β1 integrin outside-in signaling leading to cell spreading

Next we wanted to determine the mechanism responsible for the importance of Gα13–integrin interaction in β1-dependent cell migration. We previously showed that Gα13 binding to the integrin β3 is important in αIIbβ3 outside-in signaling and consequent cell spreading (Shen ). Hence we hypothesize that Gα13-mediated β1 integrin outside-in signaling and the consequent cell membrane movement (spreading and retraction) are responsible for their functions during cell migration. Thus we investigated how Gα13 knockdown affects integrin-dependent cell spreading. CHO cells with Gα13 knockdown were defective in spreading on immobilized integrin ligand fibronectin (Figure 6, A and B), suggesting the importance of Gα13 in integrin-dependent cell spreading. Previous studies showed that early-phase cell spreading on fibronectin was defective in β1− GD25 cells (Wennerberg ; Pankov ; Green ). Our data also show that early-phase cell spreading of GD25 cells on fibronectin is dependent on β1 integrins. GD25 cells expressing the β1 ExE motif mutants exhibited diminished early-phase cell spreading to a similar level as observed in β1-deficient GD25 cells (Figure 6C), suggesting that the Gα13–integrin interaction is important in β1-mediated outside-in signaling leading to β1-dependent cell spreading.

FIGURE 6:

The role of Gα13–β1 interaction in mediating β1 integrin outside-in signaling leading to cell spreading. (A) Fluorescence microscopy images of phalloidin-stained CHO cells spreading on fibronectin for 1 h. (B) Quantification of surface areas of individual cells (mean ± SE, n = 31, 24, 28, and 29 for CHO and CHO transfected with scrambled or Gα13-specific shRNAs). (C) Quantification of surface areas of individual GD25 cells spreading on fibronectin for 30 min (mean ± SE). n = 23, 25, 26, 24, 28, and 31 for Wt, AKE, EKA, AKA, AAA, and β1 knockout GD25 cells, respectively.

Gα13–β1 binding mediates outside-in signaling through activation of c-Src and transient inhibition of RhoA

To determine whether and how Gα13 binding to β1 mediates integrin outside-in signaling, we seeded Wt and mutant β1-expressing GD25 cells onto immobilized fibronectin and measured Src activation (indicated by phosphorylation at Y416) and RhoA activity (indicated by RhoA pull down with GST rhotekin Rho binding domain protein [RBD] (Ren ), which specifically binds to active RhoA) at different time points after adhesion. Adhesion of the β1(Wt)GD25 cells on fibronectin resulted in robust but transient RhoA inactivation (Figure 7, A and B). In contrast, this transient RhoA inactivation was not observed in β1(AAA)GD25 cells, which showed constant activation of RhoA (Figure 7, A and B). Furthermore, Src was activated in Wt β1–expressing GD25 cells after adhesion to fibronectin, and Src activation was inhibited in the AAA mutant β1–expressing GD25 cells (Figure 7, C and D). Because it is known that transient RhoA inhibition after β1 outside-in signaling is Src dependent (Arthur ), our data indicate that Gα13 binding to the ExE motif of β1 integrins mediates Src activation and Src-mediated inhibition of RhoA signaling. To explore further the role of the Gα13-binding β1767ExE motif in regulating RhoA activity after cell adhesion, GD25 cells adherent to immobilized fibronectin were stained with the GST-rhotekin RBD (Ren ; Figure 8). Indeed, Wt but not AAA mutant β1–expressing cells showed a transient reduction in GST-RBD staining (Figure 8A). The staining of the GST-RDB was specific, as it was abolished by Rho inhibitor C3-transferase, and GST protein alone failed to stain the cells (Figure 8A). Thus these data indicate that transient RhoA inhibition soon after cell adhesion (30 min) depended on Gα13–β1 interaction (Figure 8B).

FIGURE 7:

Gα13–β1 binding mediates activation of c-Src and transient inhibition of RhoA. GD25 cells expressing similar levels of Wt or AAA mutant β1 were allowed to adhere to immobilized fibronectin, solubilized at various time points, and analyzed for RhoA activation (A) and c-Src Tyr416 phosphorylation (C). (B, D) Quantification of three independent experiments.

FIGURE 8:

In situ RhoA activation analysis of the role Gα13 binding to β1. (A) GD25 cells were allowed to adhere to immobilized fibronectin, fixed, and stained with anti-β1 antibody (green) and GST-RBD proteins (red). β1(Wt)GD25 cells were also preincubated with 20 μg/ml C3-transferase and stained with anti-β1 (green) and GST-RBD (red). Cells stained with anti-β1 antibody (green) and GST proteins (red) served as additional negative controls. (B) Quantification of the active RhoA in A (mean ± SD, n = 3).

The effect of an inhibitor peptide based on the Gα13-binding sequence in β1

The foregoing results indicate that integrin β1–Gα13 binding plays an important role in β1 integrin–mediated activation of Src and transient inhibition of RhoA and in β1-dependent cell migration. Thus we hypothesize that an inhibitor of β1–Gα13 interaction should inhibit β1-dependent integrin outside-in signaling and cell migration. To test this hypothesis, we synthesized a myristoylated, cell-permeable peptide containing the Gα13-binding ExE motif, mβ1P6 (Myr-FEKEKM). Preincubation of this peptide with β1(Wt)GD25 cells abolished the interaction of Gα13 to integrin β1 in the coimmunoprecipitation assay (Figure 9A). In contrast, the scrambled control peptide mβ1P6Scr (Myr-EKMFEK) had no effect. These data indicate that mβ1P6 is effective in inhibiting Gα13–β1 interaction (Figure 9A). Preincubation of mβ1P6 but not the scrambled peptide also significantly inhibited transwell migration in β1(Wt)GD25 cells (Figure 9, B and C). Furthermore, c-Src activity and transient RhoA inactivation were both inhibited by this inhibitor (Figure 9, D– G), mirroring the results of ExE mutation (Figure 7). Thus this inhibitory peptide of β1–Gα13 interaction was an effective blocker of integrin β1–mediated outside-in signaling and cell migration.

FIGURE 9:

The inhibitory effects of a myristoylated peptide mβ1P6 on Gα13–β1 interaction, cell migration, c-Src activation, and transient RhoA inactivation. (A) β1(Wt)GD25 cells treated with 100 μM β1mP6 or scrambled control peptide β1mP6Src were allowed to adhere to fibronectin for 1 h. Cells were lysed, and cell lysates were immunoprecipitated with anti-Gα13 antibody or an equal amount of control rabbit IgG. Immunoprecipitates were immunoblotted with anti-Gα13 and anti-β1 antibodies. (B) Transwell migrated β1(Wt)GD25 cells with coated fibronectin 6 h after migration. Cells were treated with 100 μM β1mP6 or scrambled control peptide β1mP6Src before migration and stained with crystal violet after fixation. Number of migrated cells was compared (mean ± SD, n = 3). (C) Quantification of migrated cells using coated collagen (mean ± SD, n = 3). (D–F) β1(Wt)GD25 cells treated with 100 μM β1mP6 or scrambled control peptide β1mP6Src were allowed to adhere to immobilized fibronectin, solubilized at various time points, and analyzed for RhoA activation (D, E) and c-Src Tyr416 phosphorylation (F, G). (D, F) typical gels. (E, G) Quantification of the Western blots (mean ± SD, n = 3).

DISCUSSION

In this study, we demonstrated that Gα13 directly binds to the ExE motif of integrin β1 cytoplasmic domain, and this binding is required for β1 outside-in signaling and cell migration on the β1 integrin ligand fibronectin. Of importance, we showed that Gα13–integrin interaction mediates transient RhoA inhibition, which is in contrast to the well-known role of Gα13 in mediating RhoA activation by binding to RhoGEFs (Goulimari ; Chen ). Thus our data suggest that Gα13-mediated dynamic regulation of RhoA activity is a novel mechanism responsible for the role Gα13 during β1 integrin–dependent cell migration.

The importance of integrin β1 in cell migration has been reported in various cell types (White ; Liu ). In addition, evidence has also been given for a role for Gα13 in directed cell migration (Radhika ; Shan ; Tan ). However, it was not previously appreciated why and how β1 integrins and Gα13 play important roles in mediating cell migration and whether there is a connection. Here we show that integrin β1–dependent cell migration requires direct binding of Gα13 to the cytoplasmic domain of β1, establishing a direct connection between these two important molecules in cell migration. Recently we showed that Gα13 interacts with the platelet integrin αIIbβ3 and plays an important role in αIIbβ3 outside-in signaling (Shen ). Migration is not ordinarily considered or studied as an activity of platelets, and it is unclear whether Gα13–β3 interaction is a common mechanism of integrin signaling that is shared with β1 integrin signaling and important in cell migration. Furthermore, there are major differences in outside-in signaling mechanisms between platelet β3 integrins and β1 integrins in nucleated cells. In particular, β3 outside-in signaling requires Src binding to the C-terminal site of β3, which can be cleaved by calpain (Arias-Salgado ; Flevaris ). Calpain cleavage of β3 cytoplasmic domain abolishes c-Src binding and switches β3 integrin signaling from mediating cell spreading to retraction (Flevaris ). However, Src does not appear to bind to β1 C-terminus (Arias-Salgado ), and thus its interaction with β1 is likely to be regulated in a different way. Furthermore, β1 signaling was reported to involve complex formation of focal adhesion kinase (FAK) with Src and integrin and FAK-dependent Src activation (Xing ; Thomas ), which has not been shown in platelet αIIbβ3. We recently showed that platelet αIIbβ3-dependent activation of Src requires Gα13 binding to β3 (Gong ; Shen ). Here we further show that Gα13 directly binds to the ExE motif in β1. Disruption of Gα13–β1 binding through mutations in the ExE motif or use of an ExE motif peptide abolished β1-dependent Src activation and inhibited β1-dependent outside-in signaling and cell spreading. These data indicate that, despite the difference between β1 and β3 signaling, the ExE motif–containing integrins β1 and β3 share similar Gα13-dependent mechanisms of Src activation and outside-in signaling in platelets and migrating nucleated cells. Of importance, our data for the first time indicate that the interaction between Gα13 and β1 ExE motif plays a critical role in integrin-dependent cell migration.

Our data not only indicate the important role for the Gα13–β1 interaction in cell migration, but they also suggest a novel mechanism of Gα13-dependent dual regulation of RhoA activity in migrating cells (Figure 10). Cell migration on β1 integrin ligands involves coordinated integrin-dependent cell spreading and retraction. The alternate protrusion and retraction occur during cell spreading at the leading edge (Machacek ; Tkachenko ). Retraction in the rear of a cell pulls the cell forward (Lauffenburger and Horwitz, 1996; Ridley ). The Rho-family GTPase RhoA is a major regulator of cell retraction. RhoA activates Rho kinase. Rho kinase inhibits myosin light chain (MLC) phosphatase and increases MLC phosphorylation, resulting in actomyosin-mediated cell retraction (Kimura ), which drives inward movement of cell membranes. Thus RhoA activity inhibits cell spreading and stimulates cell retraction (Vega ). Consequently, RhoA-dependent retractile signaling needs to be dynamically activated and inhibited in order for cells to migrate. On activation by GPCRs, Gα13 directly stimulates RhoGEFs and activation of RhoA (Kozasa ), which is believed to be the reason that Gα13 is important in cell migration (Bian ; Patel ). Here we for the first time demonstrate the other aspect of this dynamic regulation. We show that Gα13 binding to β1 ExE motif mediates β1 integrin–dependent activation of Src and transient inhibition of RhoA in migrating cells. This finding is not only consistent with previous data suggesting that integrin β1 outside-in signaling transiently inhibits RhoA activity via Src during cell spreading (Arthur and Burridge, 2001), but it also provides a plausible mechanism for initiating Src-dependent transient RhoA inhibition. Thus our data, together with the previously reported role of Gα13 in activating RhoA, suggest a novel mechanism by which dynamic regulation of RhoA activation by Gα13 is achieved: Gα13 binding to ligand-bound integrins induces Src-dependent transient inhibition of RhoA, which is required for cell spreading. On the other hand, Gα13 binding to RhoGEFs stimulates RhoA and thus drives cell retraction together with RhoA activation induced by late-phase integrin signaling (Dubash ). Dynamic regulation of RhoA activation by Gα13 thus provides a novel mechanism explaining the importance of Gα13 in driving coordinated cell spreading and retraction, leading to cell migration.

FIGURE 10:

A new model for the Gα13-dependent dynamic regulation of RhoA and cell migration. GPCR-dependent activation of Gα13 stimulates the activation of RhoGEFs, leading to Rho activation, which has been suggested to be important for cell retraction in response to GPCR stimuli. Integrin ligation induces the interaction between the cytoplasmic domains of β subunits and the activated Gα13, which mediates Src-dependent transient inhibition of RhoA and activates the Rac1 and PI3K pathways. These events lead to spreading of cells (lamellipodia and filopodia) toward the direction of migration. Late-phase integrin signaling results in reactivation of RhoA and cell retraction, driving the cell movement toward the direction of migration. Thus Gα13-dependent dynamic regulation of RhoA results in coordinated cell spreading and retraction.

MATERIALS AND METHODS

Reagents

Human integrin β1 cDNA was cloned into plenti6/V5-DEST vector after digestion with EcoRI and XhoI. Integrin E-to-A mutants were generated using PCR and cloned into plenti6/V5-DEST vector by EcoRI and XhoI. GST-β1CD and recombinant Gα13 purification was described previously (Shen ). Anti-RhoA antibody was from Cytoskeleton (Denver, CO); anti-Gα13(sc410), anti–c-Src (sc18), anti–integrin β1 K20 (sc18887), and anti–integrin β1 JB1B (sc59829) antibodies were from Santa Cruz Biotechnology (Dallas, TX); anti-Gα13 (26004) was from NewEast Biosciences (King of Prussia, PA); anti–phospho-Src Y416 antibody was from Cell Signaling Technology (Danvers, MA); anti–GST tag antibody and Alexa Fluor 555 conjugate were from EMD Millipore (Billerica, MA); Lipofectamine 2000, ViraPower Lentiviral Expression System, and Alexa Fluor 546–conjugated phalloidin were from Invitrogen (Carlsbad, CA); fibronectin was from BD Biosciences (Franklin Lakes, NJ); the Active Rho Pull-Down and Detection Kit was from Pierce, Thermo Scientific (Waltham, MA).

Cell culture

CHO (Gu ; Xi ), 293FT, and GD25 cells were cultured in DMEM complete (Cellgro) supplemented with 10% fetal bovine serum (FBS; Corning), 2 mM l-glutamine (Corning), 100 U/ml penicillin plus 100 μg/ml streptomycin (Corning), 1 mM sodium pyruvate (Gibco), 0.1 mM nonessential amino acids (Gibco), and 10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid buffer (Gibco). Cells were serum starved in DMEM containing each of the foregoing components except the FBS for at least 4 h before experiments.

Lentiviral infection and integrin β1 reconstitution

Gα13 shRNA lentivirus was prepared as described previously (Gong ). Briefly, pLL3.7-scrambled shRNA or pLL3.7-Gα13 shRNA (#1 and #2) were cotransfected into subconfluent 293FT cells with pLP1, pLP2, and pLP/VSVG plasmids (Invitrogen) using Lipofectamine 2000. After 48 h, cell culture supernatant was collected, filtered, and used to infect CHO or GD25 cells. Similarly, expression of integrin β1 in GD25 cells was achieved by infecting GD25 cells with 293FT supernatant after plenti6-V5-DEST-β1 transfection using the ViraPower Lentiviral Expression System. Expression of Wt or mutant β1 integrin on the GD25 cell surface was assayed by flow cytometry using fluorescein isothiocyanate (FITC)–conjugated anti–integrin β1 antibody K20.

Purified Gα13 binding to integrin β1 cytoplasmic domains

GST-tagged integrin β1 cytoplasmic domain proteins were incubated with purified recombinant Gα13 at 4°C overnight in NP40 buffer (50 mM Tris, pH 7.4, 10 mM MgCl2, 150 mM NaCl, 1% NP-40, 1 mM sodium orthovanadate, 1 mM NaF) with complete protease inhibitor cocktail tablets (1 tablet/5 ml buffer; Roche). After three washes with the NP40 buffer, bound Gα13 was analyzed by SDS–PAGE and Western blots with antibodies against Gα13.

Cell-spreading assay

Cultured cells were detached with 0.053 mM EDTA and resuspended in serum-free DMEM (Flevaris ). Cells were allowed to spread on 10 μg/ml fibronectin–coated coverslips for different time points, fixed, permeabilized, stained with Alexa Fluor 546–conjugated phalloidin, and viewed with a Leica RMI RB microscope as previously described (O’Brien ; Shen ). Quantification of the surface area of spreading cells was performed using ImageJ (National Institutes of Health, Bethesda, MD).

Wound-healing assay

Similar to what was previously described (Liang ), cells were cultured until subconfluent and serum starved in serum-free DMEM for at least 6 h before the experiment. After starvation, scratches were made using p200 pipet tips, and DMEM complete was supplied to the cells. Cell migration and wound healing were monitored by taking pictures at selected time points using bright-field microscopy.

Transwell migration assay

Fibronectin or collagen (30 μg/ml) was coated on the outer surface of 6.5-mm trans­well inserts with 8-μm pore size (Corning) overnight at room temperature. The insert was washed with phosphate-buffered saline (PBS) and blocked with 2% bovine serum albumin in PBS for 30 min at 37°C. Serum-starved cells in DMEM with 0.5% FBS were seeded on the upper chamber of the insert and allowed to migrate for 6 h in a 37°C incubator. After that, cells on the outer surface were fixed with 3.7% paraformaldehyde, stained with crystal violet, and visualized with bright-field microscopy (Huttenlocher ).

Coimmunoprecipitation

As previously described (Shen ), cells expressing Wt or mutant human integrin β1 were solubilized in NP40 lysis buffer with complete protease inhibitor cocktail tablets (1 tablet/5 ml buffer; Roche). Lysis debris was cleared after centrifugation at 14,000 × g for 10 min. Lysates were then immunoprecipitated with rabbit anti-Gα13 immunoglobulin G (IgG) or an equal amount of control rabbit IgG for at least 6 h before protein A/G Sepharose beads were added. After incubation of protein A/G Sepharose beads for 45 min at 4°C, beads were centrifuged down and washed six times with NP40 lysis buffer. Immunoprecipitates were analyzed by immunoblotting using anti-β1 antibody JB1B or anti-Gα13 antibody (26004).

In situ RhoA immunofluorescence assay

GD25 cells were allowed to spread on immobilized fibronectin (10 μg/ml), fixed with 3.7% paraformaldehyde, permeabilized, and incubated with purified GST-RBD proteins as previously described (Flevaris ). Then cells were stained with anti–integrin β1 antibody conjugated with Alexa Fluor 488 and anti–GST tag antibody conjugated with Alexa Fluor 555 and viewed with a Zeiss LSM510 META confocal microscope. Images were analyzed using ImageJ, and the quantification of the total fluorescence per cell, which is associated with RhoA activity, was performed using the corrected total cell fluorescence.

Src phosphorylation and RhoA activity assay

Cells in modified Tyrode’s buffer or adherent on immobilized fibronectin were solubilized in cold NP40 lysis buffer at 4°C, and debris-cleared lysates were immunoblotted for phospho-Src Y418, total c-Src, or total RhoA. To measure the RhoA activity, debris-cleared lysates were incubated for 1 h with purified GST-RBD beads (Pierce, Thermo Scientific), washed, and then immunoblotted with an anti-RhoA monoclonal antibody (Ren ).

Peptide inhibitors

Myristoylated peptides mβ1P6 (Myr-FEKEKM) and mβ1P6Scr (Myr-EKMFEK) were synthesized and purified at the Research Resource Center at the University of Illinois at Chicago. The peptides were prepared in dimethyl sulfoxide for use in biochemistry experiments, dissolved in DMEM complete, and filtered for transwell migration experiments.

Statistics

Student’s t tests and one-way analysis of variance were used for comparison; Analyses were performed with GraphPad Prism 4 software. Unless otherwise specified, an asterisk indicates p < 0.005.