Vinexin β Regulates the Anchorage Dependence of ERK2 Activation Stimulated by Epidermal Growth Factor*

ERK is activated by soluble growth factors in adherent cells. However, activation of ERK is barely detectable and not sufficient for cell proliferation in non-adherent cells. Here, we show that exogenous expression of vinexin β, a novel focal adhesion protein, allows anchorage-independent ERK2 activation stimulated by epidermal growth factor. In contrast, expression of vinexin β had no effect on ERK2 activation in adherent cells, suggesting that vinexin β regulates the anchorage dependence of ERK2 activation. Analyses using deletion mutants demonstrated that a linker region between the second and third SH3 domains of vinexin β, but not the SH3 domains, is required for this function of vinexin β. To evaluate the pathway regulating the anchorage dependence of ERK2 activation, we used a dominant-negative mutant of p21-activated kinase (PAK) and a specific inhibitor (H89) of cAMP-dependent protein kinase (PKA) because PAK and PKA are known to regulate the anchorage dependence of ERK2 activation. The dominant-negative mutant of PAK suppressed the anchorage-independent ERK2 activation induced by expression of vinexin β. The dominant-negative mutant of vinexin β inhibited the anchorage-independent ERK2 activation induced by the PKA inhibitor. Together, these observations indicate that vinexin β plays a key role in regulating the anchorage dependence of ERK2 activation through PKA-PAK signaling.

ERK is activated by soluble growth factors in adherent cells. However, activation of ERK is barely detectable and not sufficient for cell proliferation in non-adherent cells. Here, we show that exogenous expression of vinexin ␤, a novel focal adhesion protein, allows anchorage-independent ERK2 activation stimulated by epidermal growth factor. In contrast, expression of vinexin ␤ had no effect on ERK2 activation in adherent cells, suggesting that vinexin ␤ regulates the anchorage dependence of ERK2 activation. Analyses using deletion mutants demonstrated that a linker region between the second and third SH3 domains of vinexin ␤, but not the SH3 domains, is required for this function of vinexin ␤. To evaluate the pathway regulating the anchorage dependence of ERK2 activation, we used a dominant-negative mutant of p21-activated kinase (PAK) and a specific inhibitor (H89) of cAMP-dependent protein kinase (PKA) because PAK and PKA are known to regulate the anchorage dependence of ERK2 activation. The dominant-negative mutant of PAK suppressed the anchorage-independent ERK2 activation induced by expression of vinexin ␤. The dominant-negative mutant of vinexin ␤ inhibited the anchorage-independent ERK2 activation induced by the PKA inhibitor. Together, these observations indicate that vinexin ␤ plays a key role in regulating the anchorage dependence of ERK2 activation through PKA-PAK signaling.
Cell attachment to the extracellular matrix and stimulation by soluble growth factors coordinately regulate numerous cellular events, including cell proliferation. Normal cells require both cell adhesion to the appropriate extracellular matrix and stimulation by growth factors for cell proliferation. Loss of this anchorage dependence of cell proliferation leads to cell malignant transformation (1)(2)(3)(4). Extracellular signal-regulated kinase (ERK), 1 also known as MAPK, is likely to be the most important molecule for regulating the anchorage dependence of cell proliferation. Activation of ERK by growth factor stimuli is a prerequisite for cell proliferation and is also anchorage-dependent (5)(6)(7)(8)(9)(10)(11). In adherent cells, soluble growth factors such as epidermal growth factor (EGF) stimulate their receptor tyrosine kinase, leading to GTP loading of Ras, which in turn activates the Raf-MEK-ERK kinase cascade. Activated ERK then stimulates the transcription of growth-related genes such as c-fos through the activation of transcription factors. However, in non-adherent cells, the activation of ERK by growth factor stimuli is barely detectable and is not sufficient for cell proliferation.
Integrins and downstream molecules have been reported to be involved in the anchorage dependence of ERK activation. Antibody-mediated engagement and clustering of integrin ␣ 5 ␤ 1 on suspended cells mimic anchorage-dependent signals (7). Other integrin ␣-chains also transmit anchorage-dependent signals (12). Recently, the constitutively active forms of focal adhesion kinase and p21-activated kinase (PAK), which are localized at focal adhesions or focal complexes and function as downstream mediators of integrin signals (13)(14)(15)(16), have been reported to allow the activation of ERK in response to growth factor stimuli even in non-adherent cells (17,18). Although focal adhesion and its components are likely to play important roles in regulating the anchorage dependence of ERK activation, the detailed mechanisms of this regulation remain to be determined.
Vinexin is localized at focal adhesions and cell-cell junctions (19) and belongs to a novel vinculin-interacting protein family (20 -26). Vinexin is transcribed into at least two isoforms, vinexins ␣ and ␤, both of which contain a common carboxylterminal sequence containing three SH3 domains. The first and second SH3 domains bind to vinculin (19), and the third SH3 domain binds to Sos, a guanine nucleotide exchange factor for Ras and Rac (27). Expression of vinexin enhances actin cytoskeletal organization and cell spreading as well as growth factor-induced c-Jun N-terminal kinase (JNK) activation, whereas it does not affect ERK activation in adherent cells (27). These observations suggest that vinexin ␤ plays roles in regulating both cell adhesion and growth factor signaling, although whether vinexin ␤ is involved in the coordinated regulation of cell adhesion and growth factor signaling or the anchorage dependence of signal transduction is unknown.
Vinculin, the binding partner of vinexin, has also been reported to be involved in the anchorage dependence of cell proliferation (28 -30). This raises the possibility that vinexin may function at a convergence point of cell adhesion and growth factor signaling and contribute to regulation of the anchorage dependence of signal transduction. Here, we examined the function of vinexin ␤ in regulating the anchorage dependence of ERK activation. Our results indicate that vinexin ␤ plays a role in PAK-mediated anchorage-dependent signaling.
Antibodies-Anti-FLAG epitope monoclonal antibody M2, anti-phospho-ERK MAPK antibody, and anti-ERK2 antibody were obtained from Sigma, New England Biolabs Inc., and Santa Cruz Biotechnology, respectively.
Cell Culture and Transfection-NIH3T3 cells were cultured with Dulbecco's modified Eagle's medium supplemented with 10% calf serum. NIH3T3 cells were transfected with GFP-ERK2 and other expression plasmids using LipofectAMINE Plus (Invitrogen). Transfected cells were then serum-starved by incubation with medium containing 0.5% calf serum for 20 h. For experiments in cell suspension, cells were detached by trypsinization, washed twice with Dulbecco's modified Eagle's medium containing 2 mg/ml soybean trypsin inhibitor, and resuspended in OPTI-MEMI reduced serum medium (Invitrogen) containing 0.2% bovine serum albumin. In some experiments, the PKA inhibitor H89 (Sigma) was added to the media. Cells were incubated in suspension with gentle rocking for 3 h. Adherent and suspended cells were stimulated by the addition of 100 ng/ml EGF for 5 min.
To determine ERK2 activity directly, an immune complex kinase assay (34) was used with minor modifications. Briefly, cell lysates were prepared as described above. GFP-tagged ERK2 were immunoprecipitated with anti-GFP antibody (Santa Cruz Biotechnology). Immune complexes were then resuspended in 100 l of kinase reaction mixture containing 25 mM HEPES (pH 8.0), 10 mM MgCl 2 , 1 mM dithiothreitol, 1 mM benzamidine, 30 g of myelin basic protein, 50 M unlabeled ATP, and 1 Ci of [␥-32 P]ATP. After incubation at 30°C for 20 min, kinase reaction products were analyzed by SDS-PAGE and autoradiography. Quantitation was performed using a BAS2000 image analyzer (Fuji Photo Film Co.).

Expression of Vinexin ␤ Induces Anchorage-independent
ERK2 Activation-To determine whether vinexin ␤ is involved in the anchorage dependence of ERK2 activation, we first generated GFP-tagged ERK2 to allow detection of activated ERK2 in cotransfection experiments. Activated GFP-ERK2 can be separated from endogenous ERK2 by SDS-PAGE without immunoprecipitation and detected using anti-active ERK antibody because GFP-ERK2 migrates slower on SDS-polyacrylamide gel than endogenous ERK2. Expressed GFP-ERK2 was strongly activated by EGF stimulation in adherent NIH3T3 cells, but was barely activated in suspended cells, similar to endogenous or hemagglutinin-tagged ERK2 ( Fig. 1, A, C, and D; and data not shown). An immune complex kinase assay for GFP-ERK2 verified the results obtained by Western blotting using anti-active ERK antibody (see Fig. 2A and data not shown). These observations indicate that both GFP-ERK2 and endogenous ERK2 respond to EGF in an anchorage-dependent manner in NIH3T3 cells and that GFP-ERK2 behaves naturally.
Using GFP-ERK2, we determined the effects of vinexin ␤ on the anchorage dependence of ERK2 activation in NIH3T3 cells. GFP-ERK2 was efficiently activated by EGF stimulation in adherent cells, but was barely activated in suspended cells, as described above. However, expression of vinexin ␤ permitted ERK2 activation in suspended cells to 50% of the levels in adherent cells (Fig. 1, A and B). Interestingly, expression of vinexin ␤ did not activate ERK2 without EGF stimulation or alter ERK2 activity in adherent cells (Fig. 1, A and B), consistent with our previous report (27). To confirm the results obtained with anti-active ERK antibody, we performed an immune complex kinase assay for detecting GFP-ERK2 activity (Fig. 2). Again, expression of vinexin ␤ increased the EGFinduced ERK2 activity in suspended cells (p Ͻ 0.05), but had no effect in adherent cells (Fig. 2). These observations suggest that vinexin ␤ is involved in the anchorage dependence of ERK2 activation. To verify that vinexin ␤ does not activate ERK2 directly, various concentrations of EGF were used to stimulate NIH3T3 cells transfected with vector alone or FLAG-tagged vinexin ␤ under adherent conditions (Fig. 1C). Vinexin ␤-transfected cells showed the same ability to activate ERK2 as vector-FIG. 1. Vinexin induces the anchorage-independent activation of ERK2. A, GFP-ERK2 was cotransfected with vector alone or FLAGtagged vinexin ␤ into NIH3T3 cells. The cells were then serum-starved and maintained on tissue culture plates (adherent (Adh)) or were placed in suspension (Sus) for 3 h before stimulation with 100 ng/ml EGF for 5 min as indicated. Cell lysates (7.5 g) were immunoblotted (IB) using anti-active ERK (phospho-ERK (p-ERK)) antibody to detect the activated (phosphorylated) forms of GFP-ERK2 and anti-ERK2 antibody to detect total GFP-ERK2 protein. B, values of GFP-ERK2 activation are expressed as -fold increase with respect to vector-transfected adherent cells. Values are normalized to GFP-ERK2 protein levels and represent the means Ϯ S.D. from five independent experiments. C, NIH3T3 cells cotransfected with GFP-ERK2 and vector or vinexin ␤ were serumstarved and stimulated with the indicated concentrations of EGF for 5 min. The activity of GFP-ERK2 was measured as described for A. D, NIH3T3 cells cotransfected with GFP-ERK2 and vector or vinexin ␤ were serum-starved and placed in suspension for the indicated times. After stimulation with 100 ng/ml EGF for 5 min, the activity of GFP-ERK2 was measured as described for A.
transfected cells at any concentration of EGF under adherent conditions. These results suggest that vinexin ␤ is involved in the anchorage dependence of ERK2 activation rather than in the direct activation of ERK2.
We examined the duration of the anchorage-independent ERK2 activation induced by expression of vinexin ␤. NIH3T3 cells transfected with FLAG-tagged vinexin ␤ were incubated in suspension for various times and then stimulated with EGF. As shown in Fig. 1D, ERK2 could not be activated in the vector-transfected cells after incubation in suspension for 3 or 5 h. In contrast, ERK2 was activated to some degree in vinexin ␤-transfected cells even after incubation in suspension for 5 h. This observation indicates that expression of vinexin ␤ can sustain the capacity for EGF-induced ERK2 activation in suspension. Furthermore, to confirm that cells incubated in suspension for 3 h were still alive and not dead due to anoikis, cells incubated in suspension for 3 h were allowed to re-adhere to tissue culture plates and then stimulated with EGF. These cells showed EGF-induced ERK2 activation equivalent to cells continuously maintained in adherent culture (data not shown), suggesting that vinexin ␤ is involved in the anchorage dependence of ERK2 activation and not in preventing anoikis under these conditions. Altogether, these observations show that expression of vinexin ␤ can substitute for cell adhesion at least partially and that it allows anchorage-independent ERK2 activation in response to EGF.
A Linker Region between the Second and Third SH3 Domains of Vinexin ␤ Is Necessary for Anchorage-independent ERK2 Activation-The first and second SH3 domains of vinexin ␤ bind to vinculin, and the third SH3 domain binds to Sos. Therefore, we first used mutants of each SH3 domain of vinexin ␤ to test whether any of these domains might be involved in the anchorage independence of ERK2 activation (Fig. 3). All mutants were expressed at similar levels compared with wild-type vinexin ␤ in NIH3T3 cells (data not shown). As shown in Fig. 3, expression of vinexin ␤ with a mutation in the third SH3 domain (3rdSH3WF) allowed ERK2 activation induced by EGF stimulation in suspension. We next tested mutants of the first SH3 domain (1stSH3WF), the second SH3 domain (2nd SH3WF), and both domains (1st2ndSH3WF), which show re-duced vinculin-binding ability. 2 All of these mutants also permitted ERK2 activation in an anchorage-independent manner (Fig. 3). These observations suggest that the functions of SH3 domains are dispensable for the anchorage-independent ERK2 activation induced by expression of vinexin ␤.
To identify the functional domain of vinexin ␤ involved in anchorage-independent ERK2 activation, we constructed various deletion mutants of vinexin ␤ (Fig. 4A). The mutants were expressed at comparable levels in NIH3T3 cells, except the C-Half mutant, which was significantly expressed at low levels (data not shown). The ⌬3SH3 and C-Half mutants lack the third SH3 domain or the first and second SH3 domains, respectively. Both mutants contain a linker region between the second and third SH3 domains (Fig. 4A). These mutants allowed ERK2 activation in suspended cells (Fig. 4B). The effect of ⌬3SH3 on the anchorage independence of ERK2 activation was significantly greater than that of wild-type vinexin ␤. In contrast, the N-Half and ⌬Linker mutants, which lack the linker region, failed to activate ERK2 in suspended cells. Furthermore, these deletion mutants slightly but significantly reduced the efficient activation of ERK2 even in adherent cells (Figs. 4B and 5C), possibly by dominant-negative blockade of cell adhesion signals, whereas the mutants containing the linker region (⌬3SH3 and C-Half) permitted increased ERK2 activation under the same conditions (Fig. 4B). These results indicate that the linker region between the second and third SH3 domains plays important roles in supporting anchorage-independent ERK2 activation in response to EGF stimulation and that ⌬3SH3 and ⌬Linker function as an active and a negative mutant, respectively.
Vinexin ␤ Functions in the PKA-PAK Signaling Pathway Regulating Anchorage-dependent ERK2 Activation-We next searched for proteins that may link vinexin ␤ and the anchorage dependence of ERK2 activation. We first determined the effect of vinexin ␤ or ⌬3SH3 on focal adhesion kinase, which has been implicated in the anchorage dependence of ERK activation (18). However, expression of vinexin ␤ did not affect the protein levels and tyrosine phosphorylation of GFP-tagged 2 S. Aizawa and N. Kioka, unpublished data.

FIG. 2. Vinexin induces the anchorage-independent activation of GFP-ERK2 activity.
A, GFP-ERK2 was cotransfected with vector alone or FLAG-tagged vinexin ␤ into NIH3T3 cells. Cells were then serum-starved and maintained on tissue culture plates (adherent (Adh)) or were placed in suspension (Sus) for 3 h before stimulation with 100 ng/ml EGF for 5 min as indicated. Two g of lysates were directly immunoblotted (IB) with anti-ERK2 antibody (lower panel) to verify expression of GFP-ERK2. The remaining lysates (200 g) were immunoprecipitated with anti-GFP antibody, and the immune complexes were subjected to ERK2 kinase assay using myelin basic protein as a substrate. The phosphorylated substrates were resolved by SDS-PAGE and visualized by autoradiography (upper panel). B, quantitation of the GFP-ERK2 kinase assays is shown graphically. Values are expressed as -fold increase with respect to EGF-stimulated vector-transfected adherent cells and represent the means Ϯ S.D. from four independent experiments.
FIG. 3. Mutations in each SH3 domain do not affect the ability of vinexin ␤ to activate ERK2 in an anchorage-independent manner. GFP-ERK2 was cotransfected with vector alone or FLAGtagged mutant vinexin ␤ (1stSH3WF (1stWF), 2ndSH3WF (2ndWF), 3rdSH3WF (3rdWF), or 1st2ndSH3WF (1st2ndWF)) into NIH3T3 cells. Cells were then serum-starved and maintained on tissue culture plates (adherent (Adh)) or were placed in suspension (Sus) for 3 h before stimulation with 100 ng/ml EGF for 5 min as indicated. Cell lysates were immunoblotted (IB) using anti-active ERK (phospho-ERK (p-ERK)) antibody to detect the activated forms of GFP-ERK2 and anti-ERK2 antibody to detect total GFP-ERK2 protein. Results representative of at least three independent experiments are shown.
focal adhesion kinase both in adherent and suspended cells (data not shown). The active form of Cdc42 has been demonstrated to enhance EGF-induced ERK activation in suspended cells and to control the anchorage requirement for cell proliferation (35,36). Under our experimental conditions, expression of a dominant-negative mutant of Cdc42 (17NCdc42) significantly blocked EGF-induced ERK2 activation in adherent cells, possibly by blocking cell adhesion signals, but did not block the anchorage-independent ERK2 activation induced by expression of vinexin ␤ (data not shown). These observations suggest that neither focal adhesion kinase nor Cdc42 mediates the function of vinexin ␤ in regulating the anchorage dependence of ERK2 activation.
PAK is a direct effector of Rac and Cdc42 (37,38) and was recently reported to regulate the anchorage dependence of ERK2 activation induced by soluble growth factors (17). We hypothesized that PAK, rather than focal adhesion kinase or Cdc42, may function downstream of vinexin ␤ in anchorage-dependent ERK2 activation. To test this hypothesis, we cotransfected the dominant-inhibitory mutant of PAK (PAK/RD) (32) with active vinexin ␤ (⌬3SH3). As shown in Fig. 5A, expression of PAK/RD alone significantly blocked the efficient activation of ERK2 in response to EGF stimulation both in adherent and suspended cells, indicating that PAK/RD blocks the signals from cell adhesion, as reported previously (17). PAK/RD also substantially inhibited the anchorage-independent activation of ERK2 induced by expression of ⌬3SH3. This observation suggests that vinexin ␤ functions upstream of PAK activation.
Recently, PKA has been shown to be activated in cells in suspension and that activated PKA negatively regulates PAK activity, leading to inhibition of ERK2 activation in suspended cells (17). On the other hand, inhibition of PKA allows PAK and ERK2 activation in an anchorage-independent manner (17). To determine whether vinexin ␤ is involved in the PKA-PAK signaling pathway leading to anchorage-dependent ERK2 activation, NIH3T3 cells were treated with H89, a specific inhibitor of PKA. H89-treated NIH3T3 cells allowed ERK2 activation even in suspension in a dose-dependent manner (Fig. 5B), indicating that PKA also regulates the anchorage dependence of ERK2 activation under our experimental conditions. Expression of a negative mutant of vinexin ␤ (⌬Linker) blocked the anchorageindependent activation of ERK2 induced by H89 treatment (Fig. 5C). These results indicate that vinexin ␤ functions downstream of PKA and is necessary for regulation of the anchorage dependence of ERK2 activation. DISCUSSION We have recently identified vinexin ␤ as a novel focal adhesion protein that binds to vinculin and regulates cell adhesion and growth factor-mediated signaling (19,27). Here, we have shown that vinexin ␤ is involved in the anchorage dependence FIG. 4

. The linker region between the second and third SH3 domains is necessary for activating ERK2 in suspended cells. A,
shown is a schematic diagram of deletion mutants of vinexin ␤. B, GFP-ERK2 was cotransfected with vector alone, FLAG-tagged wildtype vinexin ␤ (w.t.), or deletion mutants (⌬3SH3, C-Half, N-Half, and ⌬Linker) into NIH3T3 cells. Cells were then serum-starved and maintained on tissue culture plates or were placed in suspension for 3 h before stimulation with 100 ng/ml EGF for 5 min. Cell lysates were immunoblotted (IB) using anti-active ERK (phospho-ERK (p-ERK)) antibody to detect the activated forms of GFP-ERK2 and anti-ERK2 antibody to detect total GFP-ERK2 protein. Results representative of at least three independent experiments are shown.

FIG. 5. Vinexin ␤ is involved in anchorage-dependent ERK2
activation regulated by the PKA-PAK pathway. A, GFP-ERK2 was cotransfected with vector alone, the active mutant of vinexin ␤ (⌬3SH3), or the dominant-inhibitory mutant of PAK (PAK/RD) into NIH3T3. Cells were incubated for 3 h on tissue culture plates (adherent (Adh)) or in suspension (Sus). Cell lysates were immunoblotted using anti-active ERK (phospho-ERK (p-ERK)) antibody to detect the activated forms of GFP-ERK2 and anti-ERK2 antibody to detect total GFP-ERK2 protein. B, shown are the dose-dependent effects of the PKA-specific inhibitor H89. NIH3T3 cells were incubated for 3 h on tissue culture plates or in suspension with or without the indicated concentrations of H89. Cell lysates were immunoblotted using anti-active ERK and anti-ERK2 antibodies. Note that two bands were detected using anti-active ERK antibody because endogenously expressed ERK was detected in this experiment. C, NIH3T3 cells were cotransfected with GFP-ERK2 and vector or FLAG-tagged mutant vinexin ␤ (⌬Linker). Cells were incubated with or without H89 (25 M) and then treated as described for A. Cell lysates were immunoblotted (IB) using anti-active ERK and anti-ERK2 antibodies. Results representative of at least three independent experiments are shown. of ERK2 activation induced by EGF. This conclusion was supported by three lines of evidence. First, we found that ERK2 was activated efficiently in vinexin ␤-transfected cells even in suspension, but minimally in non-transfected cells in suspension. Interestingly, when cells were attached, expression of vinexin ␤ did not affect ERK2 activation stimulated by any concentration of EGF tested. Second, a negative mutant of vinexin ␤ (⌬Linker) suppressed the anchorage-independent ERK2 activation elicited by inhibition of PKA as well as ERK2 activation in adherent cells. Third, the anchorage-independent ERK2 activation induced by vinexin ␤ was inhibited by a dominant-negative mutant of PAK, which is known to play a role in the anchorage dependence of ERK2 activation (17,32). These observations indicate that vinexin ␤ is involved in the anchorage dependence of ERK2 activation.
In this study, we demonstrated that vinexin ␤ functions upstream of PAK for regulating the anchorage dependence of ERK2 activation. PAK activation itself has been shown to be anchorage-dependent (17,39), whereas constitutively active PAK allows anchorage-independent ERK2 activation (17). Therefore, regulation of PAK activity is likely to play a role in the anchorage dependence of ERK2 activation. PAK is activated by binding with GTP-bound Cdc42 or Rac (38,40) and by binding with other PAK-binding proteins (41,42). PAK-binding proteins, including Nck and PIX, have been shown to activate PAK through physical interaction and/or recruitment of PAK to the membrane (41,42). Human vinexin (GenBank TM /EBI accession number AF037261) was also reported to bind to PAK in the GenBank TM /EBI Data Bank. Therefore, one possible mechanism for regulating the anchorage dependence of ERK2 activation by vinexin ␤ is that vinexin ␤ may support the efficient activation of PAK by binding or recruiting it to the membrane under adherent conditions. We showed that the linker region between the second and third SH3 domains plays a role in regulating the anchorage dependence of ERK2 activation. The function of this region is unclear so far. There are three PXXP sequences, core sequences for SH3 domain binding, in the linker region. Interestingly, CAP (c-Cbl-associated protein)/ponsin, another member of the vinexin family, contains PQQP sequence in the linker region between the second and third SH3 domains (22,24). This region actually binds to the SH3 domains of Grb4/Nck␤, an adaptor molecule that binds to PAK (43). Therefore, it is possible that the linker region of vinexin ␤ functions as a proteininteracting domain and regulates PAK activity indirectly. Indeed, we have isolated several proteins that bind to the linker region using yeast two-hybrid screening. 3 Vinexin ␤ binds to Sos, a guanine nucleotide exchange factor for Ras and Rac, and enhances EGF-induced JNK activation in NIH3T3 cells (27). However, the vinexin ␤-Sos interaction is not likely to be involved in the anchorage dependence of ERK2 activation. Vinexin ␤ binds to Sos via its third SH3 domain (27). Mutation of the third SH3 domain disrupts binding to Sos and the ability to enhance JNK (27). Moreover, vinexin ␤ with a mutation in the third SH3 domain inhibits EGF-induced JNK activation in a dominant-negative fashion (27). In contrast, the same mutation of the third SH3 domain did not affect the function of vinexin ␤ in its regulation of the anchorage dependence of ERK2 activation. The vinexin ␤ mutant ⌬3SH3, which lacks the binding site for Sos, still supported anchorage-independent ERK2 activation in response to EGF stimulation. These observations suggest that vinexin ␤ regulates ERK2 and JNK by different mechanisms.
Vinculin, a vinexin-binding protein, is also involved in the anchorage dependence of cell proliferation (28 -30). One hypothesis regarding the role of vinexin-vinculin interaction is that vinexin ␤ is localized at focal adhesions by binding to vinculin and recruits signaling molecules such as PAK under adherent conditions, whereas in suspension culture, vinexin ␤ does not bind to vinculin and does not recruit signaling molecules to the membrane. Indeed, vinculin binding of vinexin is regulated by cell adhesion. 4 This model is compatible with the finding that the 1stSH3WF and 2ndSH3WF mutants, which do not bind to vinculin, still induced adhesion signaling. Loss of the membrane-localizing and PAK-activating abilities of Rac1 is overcome by overexpression of Rac1 (39). In our system, overexpression of vinexin mutants may overcome the loss of vinculin-binding and membrane-localizing abilities.
In conclusion, we have demonstrated that vinexin ␤ can regulate the anchorage dependence of ERK2 activation through PAK. Expression of vinexin ␤ allows ERK2 activation stimulated by EGF without cell anchorage.