When a cell migrates, the RhoA–ROCK-mediated contractile signal is suppressed in the leading edge to allow dynamic adhesions for protrusion. However, several studies have reported that RhoA is indeed active in the leading edge of a migrating cell during serum stimulation. Here, we present evidence that regulation of ROCKII phosphorylation at the Y722 site in peripheral focal contacts is crucial for controlling the turnover of the focal adhesion (FA) complex uncoupled from RhoA activation during serum-stimulated migration. However, this phosphorylation control is dispensable for migration when RhoA is downregulated in cells treated with platelet-derived growth factor (PDGF). We further present evidence that ROCKII is phosphorylated by Src in FAs and this phosphorylation event decreases RhoA binding activity of ROCKII. Lack of this regulatory control leads to sustained myosin-mediated contractility and FA elongation during lysophosphatidic acid (LPA) stimulation. Altogether, our data suggest that Src-dependent ROCKII phosphorylation provides a means of tuning contractility required for FAs dynamics when RhoA is active.
Cell migration is a highly coordinated process involving protrusion and retraction in order to allow movement of the cell body in the required direction (Lauffenburger and Horwitz, 1996; Raftopoulou and Hall, 2004; Ridley et al., 2003). In this process, actin polarization is acquired by the concerted formation of filopodia, lamellipodia and actomyosin contraction induced by the Rho GTPases Cdc42, Rac and RhoA, respectively (Hall, 1998; Heasman and Ridley, 2008; Jaffe and Hall, 2005; Ridley, 2001). At the rear of the cell body, RhoA activates ROCKI and II (Rho-associated protein kinase 1 and 2), which phosphorylate myosin light chain (MLC) and inactivate myosin light chain phosphatase to generate the contraction force needed for maturation of focal adhesion (FA) for retraction (Amano et al., 1997; Amano et al., 2000; Riento and Ridley, 2003). In the protruding front, Rac mediates the actin filaments to form lamellipodia, which requires activation of focal adhesion kinase (FAK) and is coordinated with the continuous formation of new small contact sites around the periphery (Mitra et al., 2005; Nobes and Hall, 1995; Tilghman et al., 2005). In general, too much adhesion tension decreases migration speed (Gupton and Waterman-Storer, 2006). The RhoA–ROCK-mediated contractile signal needs to be suppressed in the leading edge for the formation of small FA complexes that are rapidly turned over in the front of the migrating cell, and Rac activation for lamellipodia formation (Burridge and Wennerberg, 2004; Cox et al., 2001; Small et al., 2002). However, several studies have reported that RhoA is indeed active in the leading edge (Kurokawa and Matsuda, 2005; Machacek et al., 2009; Pertz et al., 2006; Wong et al., 2006), eliciting a question of how ROCK-mediated contractility is suppressed when RhoA is active.
During cell adhesion and migration, activation of Cdc42, Rac and RhoA is required for Src recruitment to the FA complex (Fincham et al., 1996; Timpson et al., 2001), where integrin-mediated activation of FAK creates a binding site for Src activation (Carragher and Frame, 2004; Moissoglu and Schwartz, 2006). The FAK–Src signaling complex binds to and phosphorylates various adaptor proteins, such as p130Cas and Crk, which in turn activate Rac-specific GEF (guanine nucleotide exchange factor), to stimulate actin polarization signaling and promote focal contact turnover during cell migration (Frame, 2004; Mitra and Schlaepfer, 2006; Webb et al., 2004). It has long been thought that activation of Src downregulates RhoA via phosphorylation of p190RhoGAP, thereby relaxing cytoskeleton tension for membrane extension (Arthur et al., 2000; Chang et al., 1995). However, the high RhoA activity observed in the protruding leading edge of the migrating cell after serum stimulation (Kurokawa and Matsuda, 2005; Machacek et al., 2009; Pertz et al., 2006; Wong et al., 2006) does not fit this theory. Moreover, several reports have shown that activated Src does not always decrease RhoA activity (Berdeaux et al., 2004; Pawlak and Helfman, 2002). The importance of Src activation in the motility of fibroblasts has been linked to the regulation of FA turnover, since cells lacking Src family members are still able to assemble FAs, but these are enlarged and have a slower turnover (Fincham and Frame, 1998; Webb et al., 2004).
Our laboratory has previously shown that phosphorylation of ROCKII at Y722 during cell adhesion results in reduced RhoA binding and kinase activity (Lee and Chang, 2008). During cell de-adhesion, Shp2-mediated dephosphorylation increases ROCK activation to generate sufficient contraction force for cell rounding and detachment. Since RhoA–ROCK-mediated contractility is essential for FA maturation, we are interested in the role of ROCKII Y722 phosphorylation in the control of FA turnover during cell migration. Our investigation demonstrated that regulation of ROCKII phosphorylation at the Y722 position is necessary for the control of FA turnover and directional migration in serum-stimulated cells. Furthermore, we found that Src kinase is responsible for ROCKII phosphorylation in the focal adhesion complex, by which ROCK-mediated contraction is uncoupled from RhoA activation. We showed that ROCK phosphorylation by Src is required for preventing too much FA elongation in the leading edge when RhoA is activated by LPA. Accordingly, it is possible that the loss of ROCKII phosphorylation control might contribute to the enlarged FAs found in cells defective of Src catalytic activity. Thus, this study provides new insights into control of FAs by Src through ROCK phosphorylation in migrating cells when RhoA is not suppressed in the leading edge.
ROCKII Y722 phosphorylation control is required for FA turnover and cell migration during serum stimulation
Given the importance of RhoA–ROCK-mediated contractility in regulation of FA maturation and turnover, we analyzed FAs in cells expressing similar level of wild-type (WT) and unphosphorylatable Y722F myc–ROCKII. Cells were transfected with pEGFP–paxillin and placed on fibronectin (FN)-coated glass-bottomed dish for image recording by time-lapse microscopy in serum-containing medium. The expression levels of GFP–paxillin in these cells were similar as determined by quantification of total intensity at the starting time (data not shown). We then measured the lifetime of newly formed adhesions in the protrusion regions of cells. As shown in Fig. 1A, Y722F cells contained elongated FAs, with a lifetime longer than 20 minutes as compared to that of less than 5 minutes in WT cells (see supplementary material Movies 1 and 2). In addition, we found that the fluctuation of paxillin Y31 phosphorylation in WT cells during serum stimulation was significantly increased at 30 minutes and dropped latter. By contrast, serum-stimulated Y722F cells showed a consistent level of Y31 paxillin phosphorylation for at least 60 minutes (Fig. 1B). Upon addition of blebbistatin, an inhibitor of myosin ATPase, FAs in Y722F cells disassembled immediately (supplementary material Fig. S1), confirming the requirement of myosin-II-mediated contractility in FA stabilization. Since the RhoA activity was similar in these two cell clones in the serum-containing condition (Fig. 1C) and higher level of MLC phosphorylation was observed in Y722F cells (Lee and Chang, 2008), it is probable that stabilization of FA in Y722F cells was due to gain-of-kinase function. These result led us to propose that loss of Y722 phosphorylation of ROCKII impairs FA dynamics because of the increase in RhoA–ROCK-mediated contraction force.
Next, we compared the migration abilities of WT and Y722F NIH3T3 cells by wound-healing assay in serum-containing medium. Y722F cells showed a marked delay in wound closure (Fig. 2A), indicating the importance of Y722 phosphorylation of ROCKII in migration. To substantiate our finding in the general physiological significance of ROCKII phosphorylation in migration and avoid possible bias from NIH3T3 clones, we examined the role of ROCKII Y722 phosphorylation in migration in Madin-Darby canine kidney (MDCK) epithelial cells. The conservation of the sequence surrounding Y722 of ROCKII let us to investigate the phosphorylation of endogenous ROCKII in MDCK cells. By immunofluorescence staining using an antibody against phosphorylated Y722 ROCKII peptide, ROCKII phosphorylation was detected in small focal complex and mature adhesions as well in MDCK cells (Fig. 2B). The staining signal was abolished by competition with phospho-Y722 peptide, indicating the specificity of the immunofluorescence staining (supplementary material Fig. S2). We then established MDCK epithelial clones expressing WT or Y722F myc–ROCKII under the control of a tetracycline-repressible promoter. Western blot analysis showed similar levels of myc–ROCKII induced by the removal of doxycycline in WT and Y722F MDCK clones (Fig. 2C).
Wounds were introduced in confluent monolayers of WT and Y722F MDCK cells for morphological observation during directional migration by time-lapse video recording. Cells with WT myc–ROCKII induction had a smooth edge in the protrusion front during migration, whereas those expressing Y722F myc–ROCKII had a crinkled edge with vigorous retraction in the front of the cell (Fig. 2D; Movie 3 in supplementary material). Furthermore, induction of Y722F but not WT myc–ROCKII completely abolished serum-stimulated directional migration of MDCK cells in a Transwell assay (Fig. 2E). Cells were also transfected with GFP–paxillin for adhesion turnover analysis. The results show that the rate constant for disassembly of paxillin was unaffected by the induction of WT ROCKII, but was reduced to ~30% after induction of Y722F ROCKII in MDCK cells (supplementary material Table S1). In sum, Y722 phosphorylation of ROCKII is crucial for FA dynamics and for directional migration in NIH3T3 and MDCK cells during serum stimulation.
Decrease of RhoA activity by PDGF treatment overrides the requirement of ROCK II phosphorylation for FA dynamics and directional migration
It has been shown that platelet-derived growth factor (PDGF) promotes cell migration through activating FAK-Src signaling (Sieg et al., 2000), and RhoA activity is suppressed in the leading edge of migrating cells with PDGF stimulation (Pertz et al., 2006). We then considered whether downregulation of RhoA by PDGF treatment could override the requirement of ROCKII phosphorylation for FA turnover and control of cell migration. By GST–RBD pull-down analysis, we found that treatment of NIH3T3 fibroblasts with LPA, as expected, upregulated endogenous RhoA activity, whereas addition of a high dose of PDGF (50 ng/ml) to the medium completely abolished RhoA activation (Fig. 3A). Notably, a low dose of PDGF (5 ng/ml) did not cause a significant reduction in LPA-induced RhoA activation but still stimulated cell migration. Western blot analysis of total lysates showed that a low dose of PDGF moderately increased Src activation as indicated by Y416 phosphorylation (Fig. 3A). To verify the effect of ROCKII Y722 phosphorylation on FA dynamics in the PDGF-stimulated cells, NIH3T3 clones were transfected with pEGFP–paxillin, serum-starved, and plated onto FN-coated coverslips for FA turnover analysis. FA formation was seen in these cells after LPA stimulation for 15 minutes, and these FAs were stable, with lifetime of more than 1 hour in both WT and Y722F cells (Fig. 3B). Addition of a low dose of PDGF reduced the lifetime of FAs to 6 minutes in WT cells and 16 minutes in Y722F cells (Fig. 3C). The extent of MLC phosphorylation was lower in WT than Y722F cells when treated with LPA and a low dose of PDGF, and Src inhibition increased MLC phosphorylation in WT but not Y722F cells (supplementary material Fig. S3). Given that Y27632 treatment abolished MLC phosphorylation in both WT and Y722F cells (supplementary material Fig. S3), these results indicated that the sustained tension force in Y722F cells is due to a resistance to Src-mediated downregulation of ROCK activity. In accordance with the myosin II-mediated tension force in FA maturation (Balaban et al., 2001; Geiger and Bershadsky, 2001; Pasapera et al., 2010; Webb et al., 2002), higher tension in Y722F cells prolongs the lifetime of FAs when a low dose of PDGF is present. However, a high dose of PDGF stimulation decreased the lifetime of FAs in Y722F cells, to a range close to what was observed in WT cells (Fig. 3B,C). A wound-healing assay showed that Y722F cells were much slower in wound closure with a low dose of PDGF. However, when treated with a high dose of PDGF, WT and Y722F cells exhibited similar migratory rates (Fig. 3D,E). It is known that PDGF potently activates Rac (Ridley and Hall, 1992) and downregulates RhoA activity in fibroblasts (Sander et al., 1999). We then expressed dominant-active RacV12 and found that in serum-containing medium, expression of RacV12 restored FA turnover in Y722F fibroblasts (supplementary material Fig. S4) and lamellipodial protrusion in the leading edge of Y722F MDCK cells (supplementary material Fig. S5). These results indicate that under the condition of high RhoA activity, negative control of ROCKII by Y722 phosphorylation is required for FA turnover and cell migration. However, in the condition of low RhoA and high Rac activity, phosphorylation control of ROCKII is no longer needed for migration.
ROCKII phosphorylation in focal adhesions is Src-kinase dependent
Since Src is one of the important kinase in FAs, we then tested the role of Src in ROCKII Y722 phosphorylation in murine embryonic fibroblasts (MEFs), by immunostaining. As shown in Fig. 4A, in control MEFs there was a clear punctate staining signal of Y722 phosphorylation colocalized with paxillin staining at the termini of stress fibers. Silencing of endogenous ROCKII by siRNA in control MEFs abolished the immunofluorescence staining (supplementary material Fig. S6), confirming the antibody specificity under our experimental conditions. Treatment of cells with the Src family kinase (SFK) inhibitor, SU6656, abolished the Y722 phosphorylation signal. In addition, little Y722 phosphorylation signal was detected in SYF MEFs, which are deficient in Src, Yes and Fyn kinases (Klinghoffer et al., 1999). Restoring expression of wild-type myc-tagged Src recapitulated Y722 phosphorylation in the cell periphery, and inhibition of Src kinase activity again abolished Y722 phosphorylation (Fig. 4B). These data suggest that endogenous Src kinase is crucial for ROCKII Y722 phosphorylation in FAs.
Src phosphorylates ROCKII and reduces its RhoA-GTP binding activity
We then co-expressed dominant active Src (Src527F) with WT or Y722F myc–ROCKII in HEK293T cells and examined the phosphorylation of ROCKII by immunoprecipitation combined with western blot analysis. As shown in Fig. 5A, the extent of Y722 phosphorylation of WT myc–ROCKII was increased markedly with increasing amounts of activated Src, as indicated by Y416 phosphorylation, whereas almost no phosphorylation was detected in Y722F myc–ROCKII immunoprecipitates. Western blot analysis of the anti-myc-antibody-precipitated immunocomplexes showed co-immunoprecipitation of Src with WT or Y722F myc–ROCKII (supplementary material Fig. S7). By in vitro kinase assay with recombinant Src kinase, the purified WT myc–ROCKII but not the Y722F mutant, showed a dose-dependent increase in Y722 phosphorylation (Fig. 5B). Next, we determined whether Src-dependent phosphorylation was able to repress the RhoA binding activity of ROCKII. GST-RhoA-GTPγS pull-down analysis showed that coexpression of Src527F reduced the amount of WT myc-ROCKII bound to GST-RhoA-GTPγS but had no repression effect on Y722F myc–ROCKII (Fig. 5C). Similar to ROCKII, ROCKI was also tyrosine phosphorylated by coexpression of Src527F (supplementary material Fig. S8), and its GTP-RhoA binding activity was also decreased by tyrosine phosphorylation (Lee and Chang, 2008). Altogether, these data indicate that Src-dependent phosphorylation of ROCKII negatively modulates RhoA-dependent ROCK activation.
LPA-induced FA maturation is limited by Src-dependent ROCKII phosphorylation
Application of a RhoA–ROCK-mediated contractile force induces FA growth, whereas inhibition of contractility decreases FA size (Amano et al., 1997; Chrzanowska-Wodnicka and Burridge, 1996; Ridley and Hall, 1992). We then tested whether Src-mediated ROCKII phosphorylation plays a role in regulating FA size. To this end, we treated the serum-starved NIH3T3 fibroblasts with LPA to induce formation of stress fibers and FAs, where Src is also activated (Timpson et al., 2001). LPA-induced RhoA activation caused FA growth with very little FA turnover. By immunofluorescence staining, the activated Src was detected at the termini of stress fibers during LPA stimulation colocalized with ROCKII Y722 phosphorylation in a Src kinase-dependent manner (Fig. 6A). As indicated by anti-paxillin staining, LPA stimulation induced FA lengthening to an average length of about 3.6 μm at 60 minutes. Addition of SU6656 to these cells for another 60 minutes caused FA elongation to 5.6 μm, whereas FAs in cells without Src kinase inhibition were 2.9 μm. Addition of the ROCK inhibitor, Y27632, during LPA stimulation abolished FA formation (Fig. 6B). These results indicate that inhibition of Src activity promotes FA elongation through a ROCK-mediated mechanism during LPA stimulation. Thus, Src activity modulates the extent of ROCK-dependent FA maturation.
We then compared the effect of Src inhibition on FA size during LPA stimulation in WT and Y722F cells. FAs were stained with anti-paxillin and measured. Src inhibition consistently caused FA elongation in WT cells, whereas in Y722F cells, the FAs were longer and were relatively unresponsive to Src inhibition (Fig. 6C). Endogenous LPA-induced RhoA activity, detected by the GST–RBD pull-down assays, was of a similar magnitude in WT and Y722F cells, and SU6656 treatment had little effect on RhoA activity (Fig. 6D). Western blot analysis of MLC phosphorylation indicated that contractility was transiently increased by LPA stimulation in WT cells in contrast to the sustained activation in Y722 cells, whereas inhibition of Src was able to restore MLC phosphorylation in WT but not Y722 cells (Fig. 6E). Since MLC and MLC phosphatase are substrates of ROCKII, the elevation of MLC phosphorylation in Y722F cells is probably a result of both the lack of dephosphorylation and the increase in MLC phosphorylation. In conclusion, Src-mediated ROCKII Y722 phosphorylation is involved in the regulation of contractility, thereby controlling the modulation of FA maturation when RhoA is active.
During cell migration, newly forming adhesions can either turn over or mature into more organized structures (Petit and Thiery, 2000; Webb et al., 2002). Spatiotemporal control of RhoA–ROCK-mediated contractility has a key role in regulating this process, being reduced for FA disassembly and enhanced for FA stabilization and maturation (Balaban et al., 2001; Geiger and Bershadsky, 2001; Pasapera et al., 2010; Webb et al., 2002). High RhoA activity observed in the leading edge of migrating cells in the serum-containing medium (Dohn et al., 2009; Kurokawa and Matsuda, 2005; Pertz et al., 2006; Wong et al., 2006) indicates the existence of a control that suppresses ROCK-mediated contraction for migration. In this report, we highlight ROCKII as a molecular target of Src in the regulation of FA dynamics required for migration. First, cells expressing unphosphorylatable Y722F ROCKII have stabilized FAs and are defective in directional migration in response to serum stimulation. Second, we showed that Src kinase is essential for ROCKII phosphorylation in FAs, and that Src-dependent phosphorylation renders ROCKII less susceptible to RhoA binding. Third, Src-dependent phosphorylation of ROCKII is involved in preventing too much elongation of FAs when RhoA is activated. Collectively, these data suggest that in the leading edge with high RhoA activity, Src-dependent phosphorylation of ROCKII provides a means to decrease RhoA-dependent activation of ROCK, allowing appropriate contractility for FA turnover in migration.
It has been reported that activation of Src by PDGF stimulation is able to inactivate RhoA via phosphorylation of p190RhoGAP, thereby relaxing cytoskeleton tension for membrane extension and migration (Arthur et al., 2000; Chiarugi et al., 2000). We show, in this study, that expression of Y722F ROCKII had no effect on cell migration and FA turnover when cells were treated with a high dose of PDGF that suppressed RhoA. Therefore, the control of ROCKII by Src-dependent phosphorylation is no longer important when RhoA is substantially downregulated. However, under the condition when the amount of active RhoA is sufficient to bind ROCKII, negative modulation of this binding process by phosphorylation would become important in suppressing RhoA–ROCK-mediated contractility. In support of this hypothesis, Y722-phosphorylated ROCKII isolated from cells expressing dominant active Src527F exhibited a marked decrease in binding to RhoA–GTP in vitro, whereas Y722F ROCKII had a consistently higher binding ability irrespective of Src527F expression. In accordance, we found sustained MLC phosphorylation in Y722F but not in WT cells when treated with LPA to induce RhoA–ROCK activation for stable FA formation. Inhibition of Src during LPA stimulation in WT cells was able to sustain MLC phosphorylation. Clearly, a negative control of ROCKII by Src is occurring in WT but not Y722F cells. As WT and Y722F cells have similar level of RhoA activity, we conclude that Src-mediated phosphorylation is important in limiting RhoA-dependent activation of ROCK, optimizing the tension force. Without this control, FAs lengthen during LPA stimulation, indicating the extent of FA maturation is finely tuned by Src-mediated restriction of ROCK activation. Unlike FA stabilization in LPA-treated cells in the serum-deprived condition, FA turnover in serum-containing medium is acquired through growth factor-mediated signaling, which involves Rac activation and downregulation of ROCK-mediated contractility. Presumably, when the lamellipodium moves past nascent adhesions, some adhesions disassemble, whereas others mature through a process dependent on the contraction force (Bershadsky et al., 2006; Choi et al., 2008; Giannone et al., 2007). Since expression of Y722F ROCKII severely impairs FA turnover and directional migration in the serum-containing condition, it is clear that Rac-signaling for lamellipodium is also affected by the loss of ROCKII Y722 phosphorylation control.
It has been reported that expression of a kinase-defective mutant of Src results in enlarged peripheral FAs (Fincham and Frame, 1998). Analyses of FA disassembly in SYF MEFs and normal cells expressing kinase-defective Src showed that Src kinase activity is required for FA turnover (Webb et al., 2004). According to our findings, SYF MEFs should have a higher tension force. In fact, more intense F-actin bundling in these cells and parental MEFs treated with Src kinase inhibitor did support this idea. In serum-deprived cells, we also found that LPA-induced FA elongation was enhanced by Src inhibition. Given a role of Src in activation of p190RhoGAP, it is possible that lack of negative control of RhoA activation and ROCK phosphorylation generates too much contractility in Src-deficient cells, resulting in defective FA turnover.
In summary, we proposed that the regulation of RhoA–ROCK by Src-dependent phosphorylation of ROCKII in peripheral regions offers an additional contractility control, so that FA turnover still can take place when local RhoA activity remains high. Having this control would be beneficial for the dichotomous role of RhoA in cell migration such as RhoA–Dia-mediated actin reorganization (Gupton et al., 2007; Watanabe et al., 1999) and microtubule stabilization (Palazzo et al., 2004) in the leading edge. Thus, Src affects cell migration during serum stimulation in two ways: it stimulates FA turnover by activating FAK, a number of adaptors and Rac-GEFs, but also restricts the ROCK-generated contraction force that would otherwise suppress the new FA-mediated activation of Rac needed for protrusion during migration.
Materials and Methods
Plasmids and reagents
The plasmids encoding GFP–paxillin, wild-type (WT) and Y722F mutant of myc–ROCKII and the production of the anti-phospho-Y722 ROCKII antibody have been described previously (Lee and Chang, 2008). The expression constructs of myc–Src and Src527F were kindly provided by R. H. Chen (Institute of Biochemistry, Academia Sinica, Taipei, Taiwan). RacV12 cDNA was subcloned into the pcDNA3 or pEGFP vector. The siRNA against mouse ROCKII was obtained from Dharmacon, Y-27632, blebbistatin, and SU6656 were obtained from EMD, recombinant Src protein and anti-Src antibody from Upstate, PDGF-BB from Millipore, PP2 from Biomol, anti-phospho-Y416 Src and anti-phospho-Y31 paxillin antibodies from Invitrogen, anti-paxillin antibody from BD Biosciences, LPA, anti-β-actin, anti-β-tubulin, and anti-MLC antibodies and Rhodamine–phalloidin from Sigma-Aldrich, anti-phospho-MLC2 (T18/S19) antibody from Cell Signaling Technology, and anti-ROCKII and anti-RhoA antibodies from Santa Cruz Biotechnology Inc. Anti-myc antibody was purified from hybridoma clone 9E10 (ATCC).
Cell culture and transient transfection
SYF and control MEFs were provided by R. H. Chen (Institute of Biological Chemistry, Academia Sinica, Taipei, Taiwan). The NIH3T3 clones stably expressing WT or Y722F mutant myc–ROCKII have been described elsewhere (Lee and Chang, 2008). All cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% (v/v) fetal bovine serum (FBS) or calf serum (CS) in a humidified atmosphere of 5% CO2, 95% air at 37°C. To generate MDCK clones stably expressing WT or Y722F ROCKII under the control of a tetracycline-repressible promoter, WT or Y722F myc–ROCKII cDNA was subcloned into the pUHD10.3 vector (Hermann Bujard, University of Heidelberg, Germany). These constructs and the hygromycin-resistant plasmid pHMR272 (DNA ratio 20:1) were used to stably co-transfect the T23 MDCK cell line (Barth et al., 1997), which expresses the tetracycline-regulated transactivator, using lipofectamine, and stable transfectants were selected in 200 μg/ml hygromycin B and 20 ng/ml doxycycline (Jou and Nelson, 1998).
Endogenous RhoA activity assay and RhoA-ROCK interaction assay
Endogenous RhoA activity was determined by affinity precipitation with GST–RBD as described by Ren and Schwartz (Ren and Schwartz, 2000). The RhoA–ROCKII interaction was measured as described previously (Lee and Chang, 2008). Briefly, cells expressing myc–ROCKII were harvested and lysed with a buffer (0.2% NP-40, 20 mM Tris–HCl, pH 7.4, 20 mM NaCl, 10 mM MgCl2, 1 mM EDTA, 1 mM DTT, 50 mM NaF, 2 mM Na3VO4, 1 mM PMSF, 10% glycerol, and protease inhibitor cocktail). The pre-cleared supernatants were incubated with different amount (0–5 μg) of GTPγS-loaded GST–RhoA protein for 15 minutes, followed by incubation with glutathione beads for pull-down at 4°C. After extensive washing, myc–ROCKII pulled-down by GTPγS–GST–RhoA was detected by western blotting with anti-myc antibody.
Focal adhesion analysis
Cells were transiently transfected with pEGFP-paxillin and plated on a fibronectin (FN)-coated (10 μg/ml) glass-bottom dish and incubated with Phenol-Red-free DMEM supplemented with serum (10%) and Trolox (0.5 mM) in a temperature- and CO2-controlled chamber. A series of FA dynamics images was acquired using an inverted fluorescence microscope (Carl Zeiss, AxioObserver A1) with a ×63 oil-immersion lens and collected using a cooled charge-coupled device video camera operated by AxioVision (Zeiss) image software. Individual FAs in the protrusion regions were monitored in GFP–paxillin-containing complexes by time-lapse image recording. The lifetime of an FA was taken as the time from appearance to disappearance. The fluorescence intensities of individual adhesions from cells expressing GFP–paxillin were measured over time using ImageJ (NIH). For an adhesion, the background-subtracted fluorescence intensity for paxillin was normalized to the total intensity of whole cell at each time points. The decrease in fluorescence intensity of adhesion was linear on a semilogarithmic plot of the fluorescence intensity as a function of time. From these plots, rate constants for the disassembly of paxillin could be determined from the slope.
Cell migration analysis
For the wound-healing assay, confluent monolayers of NIH3T3 fibroblasts or MDCK cells were scratched and incubated in different conditions for migration. Images were captured at the indicated times after wounding. The percentage of wound area closed at 24 hours after wounding was measured using AxioVision (Zeiss) image software. For Transwell assays, 105 serum-starved (1%) cells were trypsinized and re-plated onto an FN-coated upper chamber membrane (8 μm pore filter; Corning Costar) of the Transwell. The lower chamber of the Transwell was filled with 20% serum-containing DMEM. After 12 hours incubation, the filters were removed and the cells on the membrane were fixed with 4% paraformaldehyde. The migrated cells on the underside of the membrane and the total cells on both sides of the membrane were stained with 0.1% of Crystal Violet. The dye was extracted using 1% sodium deoxycholate and quantified by measuring the optical density at 590 nm.
Unless stated otherwise, all procedures were at room temperature. Cells were fixed with 4% paraformaldehyde in phosphate-buffered saline, pH 7.4, for 30 minutes, followed by permeabilization with Tris-buffered saline containing 0.3% Triton X-100 for 5 minutes. After blocking with 5% normal goat serum for 30 minutes, they were incubated overnight at 4°C with antibodies against phospho-Y722 ROCKII (1:100), paxillin (1:200), or phospho-Y416 Src (1:100), then for 1 hour with FITC- or TRITC-conjugated secondary antibodies, Rhodamine–phalloidin, and Hoechst, washed, mounted and examined on a fluorescence microscope (Carl Zeiss, AxioObserver A1) with a ×63 oil-immersion lens. Images were captured using a cooled CDD camera operated by AxioVision (Zeiss) image software and arranged using Photoshop (Adobe) software.
We thank the staff of the Second Core Lab, Department of Medical Research, National Taiwan University Hospital for technical support and R. H. Chen (Institute of Biological Chemistry, Academia Sinica, Taipei, Taiwan, ROC) for the Src and SYF MEF constructs. We acknowledge support from the National Science Council (NSC98-3112-B-010-024) and a grant from the Ministry of Education in National Yang-Ming University, Aim for the Top University Plan.