Protein Kinase C Induces Actin Reorganization via a Src- and Rho-dependent Pathway*

We have investigated the mechanism of PKC-induced actin reorganization in A7r5 vascular smooth muscle cells. PKC activation by 12-O-tetradecanoylphorbol-13-acetate induces the disassembly of actin stress fibers concomitant with the appearance of membrane ruffles. PKC also induces rapid tyrosine phosphorylation in these cells. As we could show, utilizing the Src-specific inhibitor PP2 and a kinase-deficient c-Src mutant, actin reorganization is dependent on PKC-induced Src activation. Subsequently, the activity of the small G-protein RhoA is decreased, whereas Rac and Cdc42 activities remain unchanged. Disassembly of actin stress fibers could also be observed using the Rho kinase-specific inhibitor Y-27632, indicating that the decrease in RhoA activity on its own is responsible for actin reorganization. In addition, we show that tyrosine phosphorylation of p190RhoGAP is increased upon 12-O-tetradecanoylphorbol-13-acetate stimulation, directly linking Src activation to a decrease in RhoA activity. Our data provide substantial evidence for a model elucidating the molecular mechanisms of PKC-induced actin rearrangements.

We have investigated the mechanism of PKC-induced actin reorganization in A7r5 vascular smooth muscle cells. PKC activation by 12-O-tetradecanoylphorbol-13acetate induces the disassembly of actin stress fibers concomitant with the appearance of membrane ruffles. PKC also induces rapid tyrosine phosphorylation in these cells. As we could show, utilizing the Src-specific inhibitor PP2 and a kinase-deficient c-Src mutant, actin reorganization is dependent on PKC-induced Src activation. Subsequently, the activity of the small G-protein RhoA is decreased, whereas Rac and Cdc42 activities remain unchanged. Disassembly of actin stress fibers could also be observed using the Rho kinase-specific inhibitor Y-27632, indicating that the decrease in RhoA activity on its own is responsible for actin reorganization. In addition, we show that tyrosine phosphorylation of p190RhoGAP is increased upon 12-O-tetradecanoylphorbol-13-acetate stimulation, directly linking Src activation to a decrease in RhoA activity. Our data provide substantial evidence for a model elucidating the molecular mechanisms of PKC-induced actin rearrangements.
The living cell has a dynamic actin cytoskeleton. To provide cell motility and adhesion and to allow cell division, the cell needs a tightly regulated and highly coordinated system of actin polymerization and depolymerization. The Rho (Ras homology) family of small GTP-binding proteins plays a central role in regulating the dynamic modulation of the actin cytoskeleton; the activity of Rho is thought to be responsible for the assembly of actin stress fibers and focal adhesions, whereas Rac is thought to be involved in the generation of lamellipodia and focal complexes, and Cdc42 is thought to be involved in microspike formation (reviewed in Refs. [1][2][3]. Rho family members also stimulate other signaling pathways that are important for both normal cellular function and transformation, including cell cycle progression, activation of c-Jun N-terminal kinase and p38 mitogen-activated protein kinase pathways, and regulation of transcription (reviewed in Refs. [1][2][3].
Like Ras, Rho, Cdc42, and Rac cycle between a GDP-bound inactive state and a GTP-bound active state. Activation is accomplished by guanine nucleotide exchange factors that favor the release of bound GDP and subsequent GTP loading.
Inactivation is mediated by GTPase-activating proteins (GAPs) 1 that stimulate the intrinsic GTPase activity of small G-proteins. In general, multiple GAPs and guanine nucleotide exchange factors can regulate the activity of a single small G-protein (reviewed in Refs. [1][2][3]. Signals are transmitted further downstream via kinases like p21-activated kinase and Rho-activated kinase or scaffolding molecules like mDia or WASP that are directly regulated by these proteins (reviewed in Ref. 4).
Protein kinase C (PKC) is a family of phospholipid-dependent serine/threonine kinases comprising 10 isoenzymes differing in their molecular domain organization of up to four variable and three constant regions and in their functions. These PKC isoenzymes are subdivided into three classes: (i) the "conventional" PKCs PKC␣, PKC␤ (␤I and ␤II), and PKC␥, which can be activated by phosphatidylserine, diacyl glycerol (DAG), or phorbolesters through binding to the C-1 domain and Ca 2ϩ through binding to a Ca 2ϩ -binding site in their second constant region, C-2; (ii) the "novel" PKCs PKC␦, PKC⑀, PKC, and PKC lack the C-2 region and thus are calcium-independent but still DAG-, phosphatidylserine-, and phorbolester-responsive; and (iii) the "atypical" PKCs PKC/ and PKC also lack the C-2 region and, in addition, are devoid of a functional DAG-binding site; hence they are only responsive to phosphatidylserine but not to DAG or phorbolester (reviewed in Refs. 5 and 6).
PKCs reside in the cytosol in an inactive conformation and translocate to the membrane (or other subcellular sites) upon activation where they modify various cellular functions through phosphorylation of target substrates. Among other kinases, PKCs were found to be involved in intracellular signal transduction pathways that regulate cell growth, differentiation, and apoptosis, and they have been implicated in the rearrangement of the cytoskeleton and migration. PKCs were also described as important regulators of cytoskeletal function in numerous studies (reviewed in Ref. 7).
In endothelial cells, PKC was described to be essential for the development of actin stress fibers (8), and PKC␣ was reported to display a positive effect on endothelial cell migration (9). In lymphocytes, PKC␦ was reported to disrupt interleukin-3-induced ruffles downstream of Rac (10). PKC␣ was described to be involved in actin reorganization in murine megakaryocytes and subsequent proplatelet formation (11). Several actin binding and modulating proteins, e.g. MARCKs (12), SSeCS (13), fascin (14), vinculin (15), talin (16), ezrin (17), and moesin (18) were identified as PKC substrates in vitro and in vivo, and colocalization of some isozymes with cytoskeletal components was observed (19). The relevance of all these findings for PKCinduced actin reorganization, however, remains mostly unclear. In addition, several groups have described the involvement of atypical, TPA-unresponsive PKC isoforms in migration and cytoskeletal rearrangements (20 -22).
One of the most prominent effects of the TPA-responsive PKC isozymes is the disruption of actin stress fibers and the appearance of ruffles in a wide variety of cells (23)(24)(25). The molecular mechanism of this reorganization, however, has yet not been elucidated.
To investigate the molecular mechanism underlying the PKC-induced actin reorganization, we examined the dynamics of the actin cytoskeleton upon TPA treatment in the rat aortic smooth muscle cell line A7r5. Our results indicated that the actin reorganization was mediated through PKC-induced activation of Src kinase activity, leading to tyrosine phosphorylation of p190RhoGAP and subsequent down-regulation of Rho activity, whereas Rac and Cdc42 activity remained unaffected.

MATERIALS AND METHODS
Reagents and Antibodies-Monoclonal antibodies against Rho1, Rac1, Cdc42, and p190RhoGAP were from Transduction Laboratories. The monoclonal Src antibody (Ab-1) was from Oncogene, and the 2-17 monoclonal Src-specific antibody was a generous gift from Suzanne Simon (Salk Institute, La Jolla, CA). The phosphotyrosine-specific antibody (pTyr-100) was obtained from Cell Signaling. Horseradish peroxidase-conjugated secondary antibodies and the Cy-2-conjugated antimouse antibody were from Dianova. TPA, GF109203X, and PP2 were obtained from Calbiochem. Enolase, para-formaldehyde, Triton X-100, and proteinase and phosphatase inhibitors were obtained from Sigma. Alexa 488-and Alexa 568-conjugated phalloidin was obtained from Molecular Probes (Leiden, The Netherlands). Polyvinylidene difluoride and nitrocellulose membranes, ECL, and [␥-32 P]ATP were from Amersham Biosciences. All of the cell culture reagents were from Invitrogen.
Cell Culture and Transfections-The vascular smooth muscle cell line A7r5 was used for all experiments. The cells were grown in Dulbecco's modified Eagle's medium without phenol red containing 10% fetal calf serum and 2 mM glutamine. The medium was changed every 3 days.
For transfection, A7r5 cells were seeded freshly in glass bottom chamber slides (Nunc) or on glass coverslips (Marienfeld), and transfections were performed after about 24 h of culture or when cells were grown to 50 -70% confluence. Expression vectors for EGFP-␤-actin (26), EGFP-V14-Rho (a kind gift from Dr. Kranewitter, Salzburg), Src wild type and kinase inactive (a kind gift from Joeren den Hertog, Hubrecht Laboratories, Utrecht, The Netherlands) (27) and Src Y529F constitutive active (a kind gift from Albrecht Piiper, Frankfurt, Germany) (27) were dissolved in the above medium without serum, and Superfect transfection reagent (Qiagen) was added according to the manufacturer's recommendations. Transfected cells were examined 2 days after transfection by confocal laser scanning microscopy or fluorescence microscopy.
Phalloidin Staining and Immunofluorescence-The cells were seeded on glass coverslips overnight and stimulated with TPA for the indicated time points. The cells were fixed in 4% para-formaldehyde for 30 min, permeabilized with 0.3% Triton X-100 in PBS for 1 min and washed several times with PBS. The permeabilized cells were incubated with phalloidin for 1 h and washed several times with PBS before imaging.
Where indicated, the cells were transfected with Src expression constructs, treated with TPA, fixed, and permeabilized. Src was visualized using 2-17 monoclonal antibody diluted in PBS with 10% goat serum at 1:300 and actin filaments with phalloidin, followed by staining with Cy-2-conjugated secondary antibody diluted in PBS at 1:50.
Microscopic Imaging-The fluorescent images were recorded on a Zeiss Axioscope equipped with an Axiocam driven by the manufacturer's software package (Zeiss, Vienna, Austria) using a 63ϫ oil immersion lens. Confocal images were taken on a Zeiss Axiovert 35M equipped with a Bio-Rad MRC 600 Argon Laser. The cells were examined in chamber slides (Nunc) using 100ϫ magnification. The images were visualized using the "confocal assistant" software.
In Vitro Src Kinase Assay-c-Src was immunoprecipitated from lysates of stimulated or unstimulated cells prepared in RIPA buffer (see above) using an anti-Src antibody (Ab-1), coupled to protein G-Sepharose. The precipitates were washed three times in RIPA buffer and twice in kinase buffer containing 20 mM Tris, pH 7.5, 5 mM MgCl 2 , 0.1% Nonidet P-40, and 0.1 mM Na 3 VO 4 as described elsewhere (31).
The samples were subjected to in vitro kinase assay in 50 l of kinase buffer supplemented with 5 M ATP and 5 Ci of [␥-32 P]ATP (3000 Ci/mM; Amersham Biosciences) using 2 g of acid-denatured enolase (Sigma) as a exogenous substrate. The kinase reactions were performed for 10 min at 30°C.
The reaction was stopped with 10 l of 6ϫ SDS sample buffer and heated to 95°C for 5 min. The samples were fractionated by SDS-PAGE (10%) and exposed to x-ray film. Equal volumes of the sample were separated in parallel on an SDS-PAGE and subjected to immunoblot analysis using the anti-Src antibody (Ab-1) to quantitate the levels of c-Src protein.
Immunoprecipitated proteins were released from the beads by boiling in SDS sample buffer and electrophoresed in 7.5% polyacrylamide gels. The proteins were electrophoretically transferred to polyvinylidene difluoride (to detect p190RhoGAP-protein) or nitrocellulose (to detect tyrosine phosphorylation of p190RhoGAP) membranes. The blots were blocked and probed with anti-p190RhoGAP monoclonal antibody or pTyr-100 (Cell Signaling).

PKC Activation Induces Disassembly of Actin Stress Fibers
and Appearance of Ruffles-To investigate the role of PKC in the reorganization of the actin cytoskeleton, we utilized the rat aortic smooth muscle cell line A7r5. These cells express most of the TPA-responsive PKC isozymes (PKC␣, PKC␤, PKC␦, PKC, and PKC) (32). In addition, they show a well developed actin stress fiber network, crossing the whole cell body when resting (Fig 1A).
Upon PKC activation via TPA stimulation, a complete reorganization of the actin cytoskeleton could be observed. As shown in Fig. 1A, the visible disassembly of the actin stress fibers started as soon as 30 min after TPA stimulation. After 2 h, most of the cells had completely lost their stress fibers. As evident, actin was reorganized in ruffles, and reorganization was essentially completed after 2 h. It has to be noted that not all cells showed the same behavior. We could observe that after only 30 min of TPA stimulation some cells showed a complete loss of stress fibers, whereas a few cells (less than 5%) maintained some stress fibers even after 2 h of stimulation.
Phorbolesters do not activate only PKC but also other signaling molecules like chimaerin (33) and Ras-guanyl nucleotidereleasing protein (34). In addition, these proteins have been implied in the reorganization of the actin cytoskeleton. Hence, it was essential to prove that the PKC kinase activity was in fact responsible for the observed effects on actin. To this end, we utilized the specific PKC inhibitor GF 109203X (35). The cells were preincubated with 2 M GF109203X for 10 min before TPA stimulation. As evident from Fig. 1B, no significant reorganization of the actin cytoskeleton could be induced by TPA stimulation in the presence of the PKC inhibitor. Therefore, the PKC kinase activity is responsible for the observed TPA effects. To further confirm our findings, we also utilized cells expressing EGFP-␤-actin. As evident from Fig. 1C, EGFPactin revealed behavior identical to that of the endogenous protein: complete disintegration of the stress fibers and the appearance of ruffles within 2 h after TPA treatment. Again, the reorganization could be completely blocked by the use of GF109203X.
The relatively long time required for the TPA-induced actin reorganization suggested that not only post-translational modifications but also de novo protein synthesis might be involved. To examine whether the TPA-induced actin reorganization does in fact require translation of new proteins, the cells were preincubated with 10 g/ml cycloheximide. This did not influence actin reorganization (not shown), and it is obvious that this process occurred independently of protein synthesis. This is in agreement with the observation of Imamura et al. (24), who also described a translationally independent TPA-mediated loss of actin stress fibers in MDCK cells.
PKC-induced Actin Reorganization Is Transmitted via Inhibition of RhoA Activity-The dynamics of the cytoskeleton is regulated through the small G-proteins Rac, Rho, and Cdc42. To investigate the role of the small G-proteins in our experimental setting, we analyzed their activation state after TPA stimulation. The activity of these proteins was examined using activation state-specific binding proteins. Surprisingly, the activity of Rac1 and Cdc42 remained unchanged after TPA stimulation ( Fig. 2A). In contrast, the activity of RhoA showed a slight decrease after 30 min of TPA stimulation and a dramatic reduction after 60 min, indicating that the reduced RhoA activity might be the cause of the changes in the actin cytoskeleton.
To prove that PKC kinase activity was responsible for the reduction of Rho activity, we preincubated the cells with GF109203X before TPA treatment. As expected from our previous data, the TPA-induced decrease in RhoA activity could be inhibited by GF109203X (Fig. 2B).
To investigate whether the down-regulation of RhoA activity was responsible for the observed actin rearrangements, we utilized an expression vector encoding RhoV14 fused to EGFP. Cells expressing this constitutively active variant of Rho (as indicated by the green fluorescence) displayed massive actin stress fibers even after 2 h of TPA treatment, whereas the surrounding cells had completely lost their stress fibers (Fig.  2C).
The down-regulation of RhoA activity resulted in reduced activity of the downstream kinase Rho-activated kinase, and hence inhibition of Rho-activated kinase activity was expected to result in a phenotype similar to that of the PKC activation. To test this hypothesis, Rho-activated kinase activity was inhibited utilizing the specific inhibitor Y27632. As has been described by several others (36,37), treatment of the cells with 10 M Y27632 for 1 h resulted in a similar phenotype with respect to actin organization as PKC activation: a complete loss of stress fibers (not shown). These data indicate that the downregulation of Rho and hence the disturbance of the balance between small G-proteins on its own is sufficient to induce a Rac-like phenotype.
PKC-induced Src Activation Is Essential for Actin Reorganization-Because inactivation of Rho, but not activation of Rac or Cdc42, contributes to the observed phenotype after PKC activation, we hypothesized that p190RhoGAP might be involved in this process. P190RhoGAP is a GTPase-activating protein with high specificity for Rho (38). The activity of p190RhoGAP itself is regulated by tyrosine phosphorylation, and several studies have described Src as one of the kinases responsible for phosphorylation of this upstream regulator of GTPase signaling (39 -41).
This led us to investigate the role of tyrosine phosphorylation in our system. As evident from Fig. 3A, we were able to show that TPA stimulation led to an overall increase of tyrosine phosphorylation in whole cell lysates as detected by Western blot analysis. These data indicated that PKC could induce activation of a tyrosine kinase like Src.
To test this, Src was immunoprecipitated from serumstarved A7r5 cells, and the kinase activity was examined before and after TPA stimulation using acid denatured enolase as a substrate. As shown in Fig. 3B, Src activity is robustly induced upon TPA stimulation. To further test the hypothesis that Src activation was necessary for PKC-induced actin rearrangement, we utilized the inhibitor PP2, which is specific for the Src family kinases. A7r5 cells were preincubated with 50 M PP2 for 10 min and subsequently stimulated with TPA. As a control, cells that had not been treated with PP2 were used. Although the disassembly of stress fibers and the appearance of ruffles were easily visible in the control cells, only a small effect of the PKC activation could be observed when Src activity was inhibited (Fig. 3C).
We also showed that PKC-mediated Rho inactivation could be inhibited with the Src-specific inhibitor. The cells were pretreated with 50 M PP2 prior to TPA stimulation, and the Rho activity was determined. As shown in Fig. 3D, the decrease in Rho activity observed in control cells after 1 h could be inhibited in cells that were pretreated with PP2. After 2 h, we could detect a moderate yet reproducible and significant decrease of Rho activity in cells prestimulated with PP2, which is in agreement with our morphological studies, where a small number of dorsal microruffles could be observed. Hence, PKCinduced Src activation is essential for the TPA-induced cytoskeletal changes.
Kinase-inactive Src Interferes with PKC-induced Disassembly of Actin Stress Fibers-To confirm our results with the Src inhibitor PP2, we utilized an expression vector for kinaseinactive Src (Src K/M). Cells overexpressing Src K/M were visualized with the Src-specific antibody 2-17. Expression of Src K/M did not result in any significant differences in the actin cytoskeleton in the absence of TPA (data not shown). Upon stimulation with TPA for 2 h, the Src K/M-expressing cells retained thin stress fibers, whereas the surrounding untransfected cells had almost completely lost their actin stress fibers (Fig. 4). As is evident, the Src K/M-expressing cells reveal partial actin reorganization, most likely because of the activity of the endogenous c-Src. This further substantiated our hypothesis that the PKC-induced changes in actin cytoskeleton are dependent on Src activation.
As controls, expression vectors encoding wild type or constitutively active Src were used. As expected, overexpression of c-Src was without effect with regard to PKC mediated actin reorganization. As has been described by several others (41,42), the expression of the constitutively active variant led to a complete loss of stress fibers without TPA stimulation (not shown).
PKC Mediates Tyrosine Phosphorylation of p190RhoGAP via Src Activation-As noted above, the link between Src activity and Rho could be p190RhoGAP. To test our hypothesis, p190RhoGAP was immunoprecipitated from the cells before and after TPA stimulation, and the level of tyrosine phosphorylation was measured using the phosphotyrosine-specific antibody pTyr-100. As evident from Fig. 5A, PKC activation resulted in an increase of tyrosine phosphorylation in p190RhoGAP, indicating activation of this protein in the A7r5 cells. Moreover, using specific inhibitors, we were able to show that the TPA-mediated increase in tyrosine phosphorylation depended on both PKC and Src kinase activity (Fig. 5B).
A schematic summary of our results is shown in Fig. 6. TPA treatment of A7r5 cells leads to a disassembly of actin stress fibers and a redistribution of actin in ruffles within 2 h. This

FIG. 2. PKC-dependent down-regulation of Rho activity.
The activity of the small G-proteins Rac, Cdc42, and Rho during TPA stimulation was investigated. A, A7r5 cells were stimulated with TPA at 1 M for the time indicated, the active (GTP-bound) proteins were precipitated using activation state-specific proteins, and their abundance was investigated by Western blotting. As a control, whole cell lysates were used. Although Rac and Cdc42 activity remained unchanged, Rho activity was decreased severalfold. B, to test whether PKC is responsible for the down-regulation of Rho activity, the cells were preincubated for 10 min with 2 M GF109203X before being stimulated with TPA at 1 M for the indicated time points, and Rho activity was measured as described. The control cells were TPA-stimulated only. Whereas in the control cells a TPA-dependent decrease in Rho activity is visible, in cells that have been pretreated with GF109203X, Rho activity remains unaffected. C, to analyze whether Rho activity on its own is responsible for the observed effects, A7r5 cells were transfected with RhoV14 fused to EGFP. 2 days after transfection the cells where stimulated for 2 h with TPA at 1 M, fixed, and stained for actin using phalloidin. The cells with a green fluorescence (indicating EGFP-RhoV14 expression) were analyzed for their actin cytoskeleton. The cells expressing the constitutively active variant of Rho show a well developed stress fiber network, whereas the surrounding cells have completely lost their stress fibers.
process was dependent on the PKC-induced activation of Src, as shown by the use of the specific inhibitors for PKC (GF109203X) and Src (PP2) and with kinase-inactive Src. Activation of c-Src resulted in tyrosine phosphorylation and acti-vation of p190RhoGap, which in turn led to an increase in the Rho GTPase activity and diminished Rho activity. Subsequently, the actin stress fibers were disassembled. DISCUSSION We have elucidated the signaling pathway utilized by PKC to mediate the reorganization of the cytoskeleton in A7r5 rat aortic smooth muscle cells. Similar observations, the disassembly of actin stress fibers upon PKC activation, have also been made by several other groups with different cell lines (23)(24)(25). This strongly suggests that the PKC-induced actin rearrangement is a general phenomenon and not restricted to A7r5 cells. In all of these reports, a TPA-mediated loss of actin stress fibers was described, but the molecular mechanism of PKCinduced actin reorganization has not been elucidated. Our data show that a first step in the signaling cascade leading to the disassembly of stress fibers is the activation of Src and the induction of tyrosine phosphorylation. We could detect a consistent overall increase in tyrosine phosphorylation and Src kinase activity in a TPA-dependent manner. As our experiments with the Src-specific inhibitor PP2 and kinase-inactive Src have shown, this induction of the Src kinase activity is necessary for the PKC-mediated actin reorganization. Although PKC-induced tyrosine phosphorylation has been described in several different cellular systems (43)(44)(45)(46)(47)(48), its central role in actin reorganization has not been discovered.
TPA treatment has also been described by others to result in an increase of Src kinase activity in vivo (44,49,50). There are conflicting reports, however, about whether PKC could activate Src directly. Although Moyers et al. (51) reported two PKC phosphorylation sites in Src that are required for the enhanced As is evident, as soon as 15 min after stimulation an increase in tyrosine phosphorylation of proteins at ϳ40, 50, 60, and 120 kDa can be observed. B, Src was immunoprecipitated (IP) from cell lysates of A7r5 cells that had been treated with TPA for the time indicated. Src kinase activity was measured using acid-treated enolase. Kinase reactions were resolved on an SDS gel, blotted onto a polyvinylidene difluoride membrane, and examined by autoradiography. As a control for equal amounts of protein, the amount of Src was examined using a Src-specific antibody. C, A7r5 cells were stimulated with TPA at 1 M for 2 h in the presence or absence of 50 M PP2, fixed, and stained for actin using phalloidin. In the presence of the Src inhibitor, stress fibers are maintained even after 2 h of TPA treatment. Only a few dorsal microruffles can be observed. As a control, the cells were stimulated with TPA at 1 M in the absence of PP2, resulting in the remodeling of the actin architecture. D, Rho activity was measured as described in the presence or absence of PP2. The cells were preincubated with the Src-specific inhibitor PP2 at 50 M before being stimulated with TPA. Controls were stimulated with TPA only. Whereas in control cells a PKC-mediated down-regulation of Rho activity could be observed, in cells that have been prestimulated with PP2 only a slight decrease of Rho activity could be observed after only 2 h of TPA treatment. response to ␤-adrenergic agonists in cells overexpressing c-Src, earlier reports indicate that the kinase activity of Src is unaffected by the PKC phosphorylation in vitro (52). Shanmugam et al. (53) published that PKC␦ associates with Src and that Src kinase activity could be up-regulated in a TPA-dependent manner. Additional reports describe the association of RACK1 with Src. RACK1, a well defined association protein for activated PKCs, also associates with c-Src upon tyrosine phosphorylation and hence might serve as a mediator of the interaction of PKC␦ with Src (54). More recently, Cabodi et al. (55) reported that PKC associates with Fyn, another Src family member, in vivo, and this association has been shown to be necessary for growth arrest and subsequent differentiation of keratinocytes. Although activation of Fyn by recombinant PKC could be shown in vitro, the authors failed to propose a model for how PKC might activate Fyn in vivo.
Another possibility is the indirect activation of Src by PKC through activation of PTP1␣. PTP1␣ has been shown to be a physiological regulator of Src (56,57). It dephosphorylates Src at tyrosine 527, thereby allowing Src to undergo an activating autophosphorylation at tyrosine 417. In addition, PTP1␣ was described as a physiological substrate of PKC (58), and its phosphatase activity is increased through phosphorylation by PKC (59). Hence it is tempting to speculate that PKC activates Src through activation of PTP1␣.
The final step in the PKC/Src-induced rearrangement of the actin cytoskeleton is the organization of actin into ruffles. The appearance of ruffles is generally considered as a Rac-dependent phenomenon and hence would require activation of Rac. Surprisingly, we could only detect a TPA-dependent decrease of Rho activity, whereas the activation state of Rac and Cdc42 remained unaffected. As evident, RhoA activity is reduced 30 min after TPA stimulation, concomitant with a first appearance of ruffles and the reduction of stress fibers in some cells. At later time points, RhoA activity is reduced more than 80%. This massive reduction is reflected by the almost complete loss of stress fibers after 2 h. Our data reveal that even in the absence of Rac activation, these structures are formed. These findings favor a model where the regulation of the actin cytoskeleton is achieved by the ratio of the different small Gproteins rather than by their absolute activity. Hence, inhibition of Rho results in the activation of Rac-dependent structures, even if the Rac activity remains unchanged.
The central role of Rho in the PKC-dependent actin reorganization is further supported by our experiments utilizing the constitutively active RhoV14 mutant. The presence of this protein, which is no longer susceptible to regulation by p190RhoGAP, completely inhibits the TPA-induced disassembly of actin stress fibers as well as the formation of ruffles. Our data strongly suggest that the down-regulation of Rho activity is mediated through Src-dependent tyrosine phosphorylation/ activation of p190RhoGAP, as has been reported by several other groups (39 -41, 60).
Initial evidence for the involvement of p190RhoGAP in our studies came from two recent publications, both of which describe a TPA-mediated increase of phosphorylation in p190RhoGAP. Cheng et al. (61) have reported a TPA-dependent increase of tyrosine phosphorylation of p190RhoGAP, which is in agreement with our data. In addition, Brouns et al. (62) have reported a TPA-mediated overall increase in phosphorylation of p190RhoGAP in in vivo labeling experiments. Further, Vincent and Settleman (63) have shown that inhibition of the p190RhoGAP activity is sufficient for the induction of Rhomediated actin reorganization, indicating its important role in the regulation of Rho-dependent signal transduction pathways.
In addition to the p190RhoGAP tyrosine phosphorylation, we observed an increase of Ras activity using activation state- FIG. 5. Tyrosine phosphorylation of p190RhoGAP upon PKC activation. A, P190RhoGAP was immunoprecipitated from A7r5 cells using 5 g of monoclonal antibody against p190RhoGAP. Tyrosine phosphorylation of p190 was examined after 0, 15, 30, and 60 min of TPA stimulation using the phosphotyrosine-specific antibody pTyr-100. As evident, tyrosine phosphorylation of p190RhoGAP is increased upon TPA treatment. B, as a control, the cells were preincubated with either 2 M GF109203X or 50 M PP2 for 15 min before being stimulated with TPA. The TPA-mediated increase of tyrosine phosphorylation could be inhibited with both inhibitors.
FIG. 6. Schematic drawing of the PKC-induced signaling leading toward actin reorganization. Activation of PKC results in the induction of Src kinase activity, possibly indirectly (indicated by the dotted arrow). Src phosphorylates and activates p190RhoGAP, which in turn leads to increased activity of the Rho GTPase activity and hence reduction of Rho activity. specific proteins in preliminary experiments (data not shown), supporting the possible involvement of p190RhoGAP in our experiments. p190RhoGAP has been reported to bind p120RasGAP in a phosphotyrosine-dependent manner (60). The association between the two proteins results in a reduced RasGAP activity, which is linked to an increase of GTP-bound Ras (64,65).
Although we see a significant inhibition of the TPA-induced actin reorganization by PP2, it is also evident that the Src inhibitor does not completely block the TPA effects on actin organization especially in later stages of stimulation. Because we found no evidence for an incomplete inhibition of the Src activity, these observations suggest that there are additional mechanisms that regulate p190RhoGAP independent of Src activity. This is in good agreement with recently reported data on p190RhoGAP regulation. Haskell et al. (66) showed that the phosphorylation of Tyr 1105 in p190RhoGAP by Src is necessary but not sufficient on its own for epidermal growth factor-induced stress fiber disassembly. These authors suggest a model by which a second, as yet unidentified epidermal growth factorinduced signal is required for the activation of p190RhoGAP. In other studies, phosphorylation of p190RhoGAP on serine residues has been described (39,60), but the physiological relevance of this phosphorylation still needs further investigation. Roof et al. (60) have reported that this serine phosphorylation was without effect with regard to the interaction with p120RasGAP in in vitro studies, but no in vivo data concerning this serine phosphorylation sites are currently available. This phosphorylation could be the missing link in the regulation of p190RhoGAP and could explain why the Src-specific inhibitor does not completely block the TPA effect at later stages of stimulation. We thus propose a model where PKC, either directly or indirectly, is responsible for additional serine phosphorylation of p190RhoGAP, which serves as a fine tuning mechanism.
Basal levels of tyrosine phosphorylation in p190RhoGAP might explain why we could observe a slight decrease of Rho activity even in the presence of PP2 in our pull-down experiments and dorsal microruffles in our morphological studies after 2 h of stimulation. A more detailed model for the exact regulation of p190RhoGAP must be the aim of further studies.
Evidence for the involvement of PKC isoenzymes in actin dynamics has been reported by several groups. Uberall et al. (22) reported that PKC and PKC/ are involved in Ras-mediated disassembly of stress fibers. Although these data are consistent with our observations regarding the PKC-dependent disassembly of stress fibers, it is not clear whether the pathways are identical. It is important to note that the atypical PKCs do not serve as phorbolester receptors and are hence not activated by TPA treatment of the cells. In addition, these atypical PKCs cannot be inhibited by the concentration of GF109203X that was used in our experiments. Etienne-Manneville and Hall (21) describe the role of PKC in the activation of Cdc42. The fact that we did not see any change in Cdc42 activity in our experiments suggests that only atypical PKCs and not the TPA-responsive novel PKCs and conventional PKCs are capable of activating Cdc42.
The results presented here describe the molecular basis for PKC-induced actin reorganization. In addition, we show that an important physiological function of PKCs is in fact dependent on tyrosine phosphorylation.