A Role for CAP, a Novel, Multifunctional Src Homology 3 Domain-containing Protein in Formation of Actin Stress Fibers and Focal Adhesions*

c-Cbl-associated protein, CAP, was originally cloned from a 3T3-L1 adipocyte cDNA expression library using full-length c-Cbl as a bait. CAP contains a unique structure, with three adjacent Src homology-3 (SH3) domains in the COOH terminus and a region sharing significant sequence similarity with the peptide hormone sorbin. Expression of CAP in NIH-3T3 cells overexpressing the insulin receptor induced the formation of stress fibers and focal adhesions. This effect of CAP expression on the organization of the actin-based cytoskeleton was independent of the type of integrin receptors engaged with extracellular matrix, whereas membrane ruffling and decreased actin stress fibers induced by insulin were not affected by expression of CAP. Immunofluorescence microscopy demonstrated that CAP colocalized with actin stress fibers. Moreover, CAP interacted with the focal adhesion kinase, p125FAK, both in vitro and in vivo through one of the SH3 domains of CAP. The increased formation of stress fibers and focal adhesions in CAP-expressing cells was correlated with decreased tyrosine phosphorylation of p125FAK in growing cells or upon integrin-mediated cell adhesion. These results suggest that CAP may mediate signals for the formation of stress fibers and focal adhesions. Src homology 2 first identified

Src homology 2 (SH2) 1 and Src homology 3 (SH3) domains, first identified as noncatalytic components of the Src family of protein kinases, are protein modules involved in protein-protein interactions that underlie the diversity and specificity of intracellular signaling (1). SH2 domains mediate protein-protein interactions by direct recognition of specific phosphotyrosine-containing motifs (2,3). SH3 domains are regions of 60 -70 amino acids that mediate protein interactions by selectively binding to proline-rich sequences present in their target molecules (4 -6). SH3 domains are found in tyrosine kinases, phosphatases, and phospholipases (7,8), as well as in various cytoskeletal proteins, including spectrin and cortactin (9,10). SH3 domains have also been shown to regulate the enzymatic activity of associated proteins (11,12), modulate GTPase activity (13), and mediate the localization of signaling proteins to specific sites within the cell (14,15).
We reported recently that insulin markedly stimulated the tyrosine phosphorylation of the c-Cbl proto-oncogene product in 3T3-L1 adipocytes, but not in 3T3-L1 fibroblasts or in any other cell lines expressing high levels of functional insulin receptors and c-Cbl (16). To explore this specificity, c-Cbl was used as a bait in search for adipose-specific signaling proteins. We isolated a novel protein that we termed CAP for c-Cbl-associated protein. CAP contains a unique structure with three adjacent SH3 domains in the COOH terminus and a region sharing significant sequence similarity with the porcine peptide hormone sorbin, termed the sorbin homology domain (17,18). Interestingly, both CAP mRNA and proteins were expressed in 3T3-L1 adipocytes but not in 3T3-L1 or NIH-3T3 fibroblasts. CAP interacts with c-Cbl and Sos in living cells via the COOHterminal SH3 domain of CAP. Furthermore, one major CAP isoform associated with the insulin receptor, and insulin stimulation resulted in the dissociation of CAP from the insulin receptor (18).
Here, we report that ectopic expression of one form of CAP cDNA in NIH-3T3 cells resulted in increased stress fibers and focal adhesion formation. These morphological changes correlate with reduction of adhesion-dependent tyrosine phosphorylation of the focal adhesion kinase, p125FAK. Our results suggest that CAP may play an important role in protein-protein associations involved in cell architecture changes.
with Ca/Mg-free phosphate-buffered saline (PBS) and then with DMEM containing 0.5% bovine serum albumin. Cell suspensions were incubated in DMEM, 0.5% bovine serum albumin at 37°C for 30 min, then plated onto dishes coated with fibronectin (20 g/ml), and incubated at 37°C for 15 min.
For calculating cell growth rate, the cell lines were seeded in six-well culture plates at a density of 5 ϫ 10 4 cells/well and maintained in a medium containing 10% fetal bovine serum. The culture medium was changed every day. Every 24 h, cells from individual wells were removed by incubation with trypsin and counted (in triplicate) in a Coulter counter.
Expression Vector Construct and Stable Expression of CAP in NIH-3T3-IR Cells-A Flag epitope-tagged full-length CAP cDNA in the pBabe retroviral vector (19) was constructed as follows. The cDNA of the amino-terminal region of CAP was amplified by polymerase chain reaction with CAP cDNA in pEXlox as the template (17). The primer 5Ј-CGCGGATCCGCCGCCACCATGGACTACAAGGACGACGATGAC-AAGAGTTCTGAATGTGAT-3Ј was designed to have a BamHI restriction site, followed by a coding sequence for a Flag epitope in frame with amino acid 2 and the primer 5Ј-AATGTCTGGAGTCGG-3Ј corresponding to amino acids 345-350. The amplified DNA fragment was digested with BamHI and SacI (present in the CAP gene), and ligated into the largest CAP cDNA clone, isolated by the yeast two-hybrid screen, in pGAD-GH vector digested with BamHI and SacI. The Flag-tagged full-length CAP cDNA was then liberated by digestion with BamHI and EcoRI and ligated into the pBabe vector at the same sites. The sequence of the Flag-tagged CAP cDNA (Flag-CAP) in this vector was confirmed by DNA sequencing. NIH-3T3-IR cells were infected with pBabe-Flag-CAP, and stable clones were selected with 2 g/ml puromycin (Sigma).
Immunofluorescence Microscopy-NIH-3T3-IR cells expressing CAP (NIH-3T3-IR-CAP) or NIH-3T3-IR cells expressing the vector alone were grown on glass Lab-Tex chamber slides (Nunc Inc.), or slides coated with fibronectin (20 g/ml) or laminin (50 g/ml). The cells were washed with PBS and fixed for 20 min in PBS containing 4% paraformaldehyde at 4°C. Fixed cells were rinsed with PBS and treated with 25 mM NH 4 Cl in PBS for 10 min to quench free aldehyde groups. Cells were then permeabilized with 0.2% Triton X-100 for 5 min, washed three times with PBS, and incubated for 30 min in blocking buffer containing 0.2% Nonidet P-40, 5% dry milk, 50 mM Tris-HCl (pH 7.6), 150 mM NaCl. Affinity-purified anti-CAP (1 g/ml), anti-Flag (10 g/ml), or anti-vinculin (1:400 dilution) antibodies were applied in blocking buffer for 2 h, followed by incubation with Oregon Green™-labeled goat antirabbit or goat anti-mouse (2 g/ml) antibodies for 1 h. For actin localization, the cells were stained with 0.1 g/ml tetramethylrhodamine isothiocyanate (TRITC)-labeled phalloidin (Sigma) for 30 min. For insulin stimulation, serum-starved cells were removed by trypsin treatment and washed as described for the cell adhesion experiments, plated on fibronectin-coated Lab-Tex chamber slides for 3 h at 37°C, and then stimulated with insulin (100 nM) at 37°C for 15 min. Cells were viewed on a Zeiss 135 Axiovert microscope. Fluorescence photographs were taken on Kodak 400 ASA film and digitally processed using Adobe Photoshop. In all cases, images presented here are representative of at least three separate experiments.
Immunoprecipitations and Immunoblotting-Cells were washed twice with ice-cold PBS and lysed with buffer containing 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 5 mM EDTA, 1% Nonidet P-40, 10 mM sodium pyrophosphate, 1 mM sodium orthovanadate, 40 mM NaF, 10 g/ml each aprotinin and leupeptin, and 1 mM phenylmethylsulfonyl fluoride on ice for 15 min. After centrifugation at 10,000 ϫ g for 15 min at 4°C, protein concentrations in the clarified supernatants were determined by the Bio-Rad method. Equal protein amounts were incubated with the indicated antibodies for 3 h at 4°C. The immune complexes were precipitated with protein G/protein A-agarose for 2 h, and washed extensively with lysis buffer before solubilization in Laemmli sample buffer. Bound proteins were separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to nitrocellulose membrane. Individual proteins were detected with the specified antibodies and visualized by blotting with horseradish peroxidase-linked secondary antibodies. To reprobe immunoblots, the nitrocellulose membranes were incubated for 30 min at 60°C with 62.5 mM Tris-HCl (pH 6.8), containing 2% SDS and 0.7% 2-mercaptoethanol and then washed extensively with 10 mM Tris-HCl (pH 8.0) and 150 mM NaCl.
In Vitro Binding Assays-The DNA fragments encoding the individual SH3 domains of CAP were generated by polymerase chain reaction and inserted in frame between the BamHI and EcoRI sites of the GST expression vector pGEX-2T (Pharmacia). All constructs were verified by DNA sequencing. In vitro association experiments were performed with equal amounts (5 g) of the immobilized GST fusion proteins as de-scribed previously (20). The bound proteins were analyzed as described above for immunoprecipitates.

Stable Expression of CAP in NIH-3T3-IR Cells-Full-length
c-Cbl was used as a target protein to screen a 3T3-L1 adipocyte cDNA library in the yeast two-hybrid system. Multiple, independent cDNA inserts that were cloned from this library encoded a novel protein, designated as CAP for c-cbl-associated protein (18). Analysis of the predicted amino acid sequence of these clones revealed three adjacent SH3 domains in the carboxyl terminus and distinct sequences in the NH 2 terminus, sharing significant sequence similarity with the peptide hormone sorbin, that we have termed the sorbin homology domain (Fig. 1A). The structure of CAP suggests that it may participate in multiple signaling cascades. DNA sequencing of the isolated cDNAs and Northern blot analysis indicated that there are multiple splice variants of CAP that may result in multiple isoforms with distinct biological functions. Both CAP mRNA and proteins are expressed predominately in 3T3-L1 adipocytes and not in 3T3-L1 fibroblasts (18).
To explore the function of CAP, a retroviral-based gene transfer system was used for stable expression of one isoform of CAP in NIH-3T3 cells overexpressing the insulin receptor (NIH-3T3-IR). To follow expression of the exogenous CAP molecule, it was Flag epitope-tagged at its NH 2 terminus (Flag-CAP). Puromycin-resistant clones were isolated, and two representative clones that express comparable amounts of Flag-CAP (NIH-3T3-IR-CAP2 and NIH-3T3-IR-CAP8) are depicted in Fig. 1. Anti-CAP antibodies detected a single protein with an apparent molecular mass of 88 kDa in lysates prepared from NIH-3T3-IR-CAP2 or NIH-3T3-IR-CAP8 cells, but not from NIH-3T3-IR cells (Fig. 1B). Similar results were obtained when lysates prepared from NIH-3T3-IR-CAP2 cells were immunoprecipitated with anti-CAP antibodies followed by immunoblotting with anti-Flag antibodies (Fig. 1C). The NH 2 -terminal Flag-tagged CAP was recognized by both the anti-Flag antibodies and an anti-CAP serum raised against the COOH-terminal domain of CAP. Both antibodies detected CAP only in whole cell lysates prepared from the CAP-infected NIH-3T3-IR cells and not in cells infected with the vector alone. NIH-3T3-IR cells express no detectable endogenous proteins recognized by the anti-CAP antibodies. The molecular mobility of CAP is consistent with the molecular mass of the isoform predicted from the cDNA.
During construction of the cell lines for stable expression, we observed that the growth rate of the cells expressing Flag-CAP was slower than that observed for NIH-3T3-IR cells. As shown in Fig. 1D, both cell lines expressing CAP exhibit a growth rate approximately one-third that of the parental NIH-3T3-IR cells.
CAP Expression Induces Stress Fibers and Focal Adhesion Formation in NIH-3T3 IR Cells-In addition to a slower growth rate, CAP-expressing cells exhibited marked differences in size and shape compared with their parental cells. Analysis of the actin-based cytoskeleton of NIH-3T3-IR-CAP cells by phalloidin staining revealed that CAP expression induced a dramatic increase in the number and density of F-actin stress fibers along both axes of cells ( Fig. 2A). Staining with antibodies against vinculin, a major component of focal adhesions (21), showed that CAP expression increased the number of elongated, arrowhead-shaped vinculin-containing complexes linked to the longitudinal stress fibers, indicating the formation of focal adhesions (Fig. 2AЈ). In contrast, NIH-3T3-IR cells grown under the same conditions had few organized actin filaments and little if any focal adhesions (Fig. 2, D and DЈ). Thus, the increased content of F-actin stress fibers correlates with an increased formation of adhesion complexes, suggesting that CAP may modulate signaling events associated with integrin-mediated adhesion.
Since the cytoskeletal structure of cells is dynamically regulated by extracellular matrix in contact with the cell surface (22,23), we analyzed the cell shape and organization of the cytoskeleton in cells grown on a fibronectin coated surface (Fig.  2, B and BЈ). NIH-3T3-IR-CAP cells grown on fibronectin contained a dense array of organized actin fibers, accompanied by an assembly of focal adhesions. As a consequence, the overall size of the cells bound to fibronectin was enlarged compared with NIH-3T3-IR cells grown on fibronectin (Fig. 2, E and EЈ). Similar phalloidin and vinculin staining patterns were observed when NIH-3T3-IR-CAP cells were plated on laminin (Fig. 2, C and CЈ). Interestingly, less actin organization and fewer focal adhesions were observed in NIH-3T3-IR cells plated on laminin than on fibronectin (Fig. 2, F and FЈ), suggesting a different integrin-mediated cytoskeleton reorganization in NIH-3T3-IR cells (24).
Insulin-induced Rearrangement of Actin Stress Fibers in NIH-3T3-IR-CAP Cells-We next studied whether the CAP-induced changes in the actin-based cytoskeleton could modulate the response of these cells to insulin. Previous studies have shown that insulin produced stress fiber breakdown and membrane ruffling in various cell types (25)(26)(27). Serum-starved NIH-3T3-IR-CAP or NIH-3T3-IR cells were plated on a fibronectin-coated surface and were either left untreated (Control), or stimulated with 100 nM insulin for 15 min. Treatment with insulin of NIH-3T3-IR-CAP or NIH-3T3-IR cells markedly reduced the number and length of actin stress fibers as compared with untreated cells (Fig. 3). Additionally, insulin treatment of both cell lines resulted in increased membrane ruffling around the cell margins, consistent with earlier observations (Fig. 3, B and D, and Refs. 25 and 26).
CAP Is Localized with Actin Stress Fibers in NIH-3T3-IR-CAP Cells-To correlate the formation of stress fibers and focal adhesions in NIH-3T3 cells expressing CAP with its subcellular localization, immunofluorescence labeling of NIH-3T3-IR or NIH-3T3-IR-CAP cells was carried out using affinity-purified anti-CAP antibodies. A modest, nonspecific staining was observed when 3T3-NIH-IR cells were incubated with anti-CAP antibodies (Fig. 4A, NIH-3T3-IR, Anti-CAP). Immunofluorescence of NIH-3T3-IR cells expressing Flag-tagged CAP with anti-CAP antibodies showed distinct, arrowhead-shaped structures (Fig. 4A, NIH-3T3-IR-CAP, Anti-CAP). Dual labeling with phalloidin (Fig. 4A, NIH-3T3-IR-CAP, Actin) and superimposition of anti-CAP and actin images revealed an overlap between the staining patterns observed with anti-CAP antibodies and the actin stress fibers (Fig. 4A, NIH-3T3-IR-CAP, Anti-CAP/Actin). These results suggest the colocalization of CAP with actin stress fibers. The specificity of the staining pattern observed with the anti-CAP antibodies was confirmed by incubation of NIH-3T3-IR-CAP cells with anti-Flag antibodies. As  Functional Association of CAP with the Focal Adhesion Kinase p125FAK-The results described above provide a functional linkage between CAP expression and cytoskeletal rearrangements. The presence of three divergent domains in CAP (Fig. 1A) suggests that some of the biological roles of CAP may arise from interactions of the SH3 domains with proteins that are involved in the organization of the actin cytoskeleton. One candidate is the focal adhesion kinase, p125FAK, a nonreceptor protein-tyrosine kinase that contains proline-rich motifs in the COOH terminus (28,29). To determine whether CAP may associate with p125FAK, each of the individual SH3 domains of CAP were expressed as GST fusion proteins. These GST fusion proteins were incubated with lysates prepared from NIH-3T3-IR-CAP cells, and the bound proteins were analyzed by immunoblotting with anti-p125FAK antibodies. As Shown in Fig. 5A, the middle SH3 domain of CAP (SH3B, Fig. 1A) most efficiently bound p125FAK, whereas modest binding was detected for the NH 2 -terminal (SH3A) and the carboxyl-terminal (SH3C) SH3 domains. Interestingly, the carboxyl-terminal SH3 domain of CAP (SH3C) was able to bind c-Cbl and Sos in cell lysates (18).
To gain further insight regarding the functional importance of CAP and p125FAK association, we examined whether these proteins interact in intact cells. Lysates prepared from NIH-3T3-IR-CAP or NIH-3T3-IR cells were immunoprecipitated with anti-p125FAK antibodies followed by immunoblotting with anti-CAP antibodies (Fig. 5B). CAP was readily coimmunoprecipitated with endogenous p125FAK in NIH-3T3-IR-CAP but not in NIH-3T3-IR cells, indicating that CAP and p125FAK form stable complexes in intact cells, possibly through the SH3 domains of CAP.
Protein-tyrosine Phosphorylation in NIH-3T3-IR-CAP Cells-Tyrosine phosphorylation of proteins involved in actin cytoskeletal rearrangements and interactions are critical events in integrin-mediated signal transduction (22,30,31). The association of CAP with p125FAK prompted us to examine the tyrosine phosphorylation state of p125FAK in growing NIH-3T3-IR-CAP cells. Equal protein amounts from NIH-3T3-IR-CAP or 3T3-NIH-IR cell lysates were incubated with anti-p125FAK antibodies. The resulting immunoprecipitates were analyzed by immunoblotting with anti-phosphotyrosine antibodies. As shown in Fig. 6A (Anti-pY Blot), p125FAK is prominently tyrosine-phosphorylated in attached, growing NIH-3T3-IR cells (30). Interestingly, p125FAK tyrosine phosphorylation in NIH-3T3-IR-CAP cells is only a fraction of that observed in NIH-3T3-IR cells. Reprobing of the membrane with anti-p125FAK antibodies showed that CAP did not affected the expression levels of p125FAK (Fig. 6A, Anti-p125FAK Blot).
The tyrosine phosphorylation of p125FAK is increased by integrin-mediated cell adhesion (28,30). To analyze the level of tyrosine phosphorylation of p125FAK in NIH-3T3-IR-CAP cells following integrin receptor activation, NIH-3T3-IR-CAP and NIH-3T3-IR cells were either kept in suspension or plated on dishes coated with fibronectin for 30 min. Cell lysates prepared from suspended or adherent cells were immunoprecipitated with anti-p125FAK antibodies, followed by immunoblotting with anti-phosphotyrosine antibodies (Fig. 6B, Anti-pY Blot). p125FAK immunoprecipitated from suspended NIH-3T3-IR-CAP or NIH-3T3-IR cells contained little if any phosphotyrosine. Replating of 3T3-NIH-IR cells on fibronectin resulted in enhanced tyrosine phosphorylation of p125FAK, as reported previously (28,30). In contrast, p125FAK immunoprecipitated from NIH-3T3-IR-CAP cells plated on fibronectin contained significantly reduced levels of tyrosine phosphorylation compared with levels detected in NIH-3T3-IR cells under the same conditions. To confirm that p125FAK was precipitated equally from all samples, the blot was completely stripped of the antiphosphotyrosine antibodies and reprobed with anti-p125FAK antibodies (Fig. 6B, Anti-p125FAK Blot). Equal amounts of p125FAK were precipitated from all samples. These results indicate that the lower basal tyrosine phosphorylation state of p125FAK in CAP-expressing cells is due in part to an impaired adhesion-mediated p125FAK tyrosine phosphorylation.
The tyrosine phosphorylation of p130 CAS and paxillin has also been shown to coincide with cell adhesion (32)(33)(34). We examined the tyrosine phosphorylation status of p130 CAS and paxillin in NIH-3T3-IR-CAP cells upon adhesion to fibronectin (Fig. 6C). Samples from suspended or adherent cells prepared as described in panel B were immunoprecipitated with anti-p130 CAS or anti-paxillin antibodies and phosphorylation was analyzed by immunoblotting with anti-phosphotyrosine antibodies (Fig. 6C, Anti-pY Blot). As reported previously (33), both p130 CAS and paxillin immunoprecipitated from suspended NIH-3T3-IR-CAP or NIH-3T3-IR cells contained low but detectable tyrosine phosphorylation, which increased dramatically upon adhesion to fibronectin. This level of tyrosine phosphorylation was similar in NIH-3T3-IR-CAP or NIH-3T3-IR cells. DISCUSSION CAP contains three adjacent SH3 domains in the carboxyl terminus and a region with homology to the peptide sorbin at the NH 2 terminus. DNA sequencing and Northern blot analysis indicated that there are multiple splice variants of CAP that may result in a family of different isoforms expressed in a tissue-specific manner with distinct physiological functions (18). These transcripts were preferentially highly expressed in heart, liver, skeletal muscle, and kidney. Although CAP mRNA and proteins were not detected in 3T3-L1 or NIH-3T3 fibroblasts, both CAP mRNA and protein isoforms accumulate following differentiation of 3T3-L1 cells into adipocytes (18).
To explore the function of the CAP proteins, we ectopically expressed one isoform in NIH-3T3 cells overexpressing the insulin receptor (NIH-3T3-IR-CAP). Expression of this protein induced profound morphological changes in these cells, producing larger cells that were prone to spreading. These changes were accompanied by a reduction in growth rate. The morphological changes were substantiated by examining the organization of the actin cytoskeleton and dynamic regulation of signaling proteins thought to be involved in cytoskeleton reorganization. The abnormally enlarged cell shape of NIH-3T3-IR cells expressing CAP compared with normal cells was consistent with increased number and thickness of actin stress fibers along with the formation of vinculin-FIG. 6. Analysis of protein-tyrosine phosphorylation in NIH-3T3-IR-CAP cells. A, growing NIH-3T3-IR-CAP or NIH-3T3-IR cells on cell culture plates were lysed and equal amounts of proteins were immunoprecipitated (IP) with anti-p125FAK antibodies. The immune complexes were then immunoblotted with anti-phosphotyrosine antibodies (Anti-pY Blot). The blot was stripped of the anti-phosphotyrosine antibodies and reprobed with anti-p125FAK antibodies (Anti-p125FAK Blot). B, NIH-3T3-IR-CAP or NIH-3T3-IR cells were either plated on dishes coated with fibronectin (FN) for 15 min at 37°C or held in suspension (S). Cell lysates were prepared, and equal amounts of proteins were immunoprecipitated with anti-p125FAK antibodies followed by immunoblotting with anti-phosphotyrosine antibodies (Anti-pY Blot). The blot was stripped of the anti-phosphotyrosine antibodies and reprobed with anti-p125FAK antibodies (Anti-p125FAK Blot). C, anti-p130 CAS or anti-paxillin antibodies immunoprecipitates from cells prepared as in B were analyzed with anti-phosphotyrosine antibodies (Anti-pY Blot). The positions of molecular mass markers (in kDa) are indicated on the left. containing focal adhesions. Moreover, the induction of actin stress fibers and focal adhesion formation in NIH-3T3-IR-CAP cells was independent of the type of integrin receptors engaged with extracellular matrix. This was evident when NIH-3T3-IR-CAP or NIH-3T3-IR cells were plated onto laminin-coated dishes. Laminin less efficiently induced organization of actin and focal adhesion formation in NIH-3T3-IR cells, but it supported the actin stress fibers and focal adhesions in CAP-expressing cells as well as fibronectin. These results suggest that ectopic expression of CAP in fibroblasts may reproduce adhesion-mediated reorganization of the actin cytoskeleton.
Recent studies have shown that the regulation of the actin cytoskeleton is under the control of the small GTP-binding proteins Rho, Rac, and Cdc42 (35). Activation of Rho in fibroblasts results in stress fibers and focal adhesion formation through the bundling of actin filaments, while Rac has been shown to induce actin polymerization at the plasma membrane to form membrane ruffles and lamellipodia (25,(35)(36)(37). Cdc42 activation triggers the formation of filopodia (37). However, the biochemical mechanisms that account for the Rho-, Rac-, and Cdc42-mediated changes in actin cytoskeleton remain poorly defined. The induction of dense, highly organized actin stress fibers traversing the cells with focal adhesion formation by CAP expression mimics the effect of Rho on actin organization and focal adhesions (37,38), suggesting a role for CAP in the Rho-dependent pathways. Rac has been shown to be involved in insulin-induced membrane ruffling (25). The observation that CAP expression failed to inhibit the effect of insulin on the actin reorganization in NIH-3T3-IR-CAP cells suggests the involvement of CAP in Rho-mediated stress fiber formation, with only a minor effect on the Rac-induced membrane ruffling. It remains to be determined whether CAP is an upstream or downstream effector of Rho.
The subcellular localization of CAP with actin stress fibers suggests that its effect on actin stress fibers and focal adhesion formation in NIH-3T3-IR-CAP cells may be due to interactions with proteins associated with cytoskeletal signaling. Indeed, CAP forms stable complexes with the nonreceptor focal adhesion tyrosine kinase p125FAK via its middle SH3 domain. Recently, the SH3 domains of p130 CAS and the GTPase regulatory protein Graf, shown to be components of cytoskeletal signaling (39 -41), were found to bind p125FAK. On the other hand, the COOH-terminal SH3 domain of CAP binds c-Cbl. The latter protein has recently been suggested as an intermediate signaling protein in integrin-mediated cell adhesion (18,42). The binding specificity of the three SH3 domains of CAP suggests that coordinate, yet distinct functions of these domains may contribute to the biological role of CAP in modulating actin cytoskeleton organization.
A further molecular link between the effect of CAP expression and the cytoskeletal changes is the observation of impaired tyrosine phosphorylation of p125FAK both in growing cells or following integrin-mediated adhesion. The tyrosine phosphorylation of paxillin and p130 CAS coincides with the tyrosine phosphorylation of p125FAK during integrin-mediated cell adhesion (32)(33)(34). However, despite decreased tyrosine phosphorylation of p125FAK in NIH-3T3-IR-CAP cells, there was no parallel decrease in the tyrosine phosphorylation of paxillin or p130 CAS in response to adhesion of these cells to extracellular matrix. These results are reminiscent of those observed in cells deficient of p125FAK. FAK-deficient cells possess increased number of focal adhesions, with normal adhesion-induced tyrosine phosphorylation of p130 CAS , paxillin, tensin, and cortactin (43,44), suggesting that p125FAK may be involved in the turnover of focal adhesions rather than in their assembly. This, in turn, would induce tighter adhesion to sub-strate. The association of CAP with p125FAK, along with the subsequent reduction in the tyrosine phosphorylation state of p125FAK, may contribute to the observed altered morphology by promoting inappropriate assembly of an active signaling complex. The inhibitory effect of CAP expression on cell growth is consistent with results obtained in cells microinjected with the carboxyl terminus of p125FAK (45).
We show here for the first time that expression of a SH3 domain-containing protein resulted in increased assembly of actin stress fibers and focal adhesion formation. Our results suggest that CAP may play an important role in a signal transduction network regulating cytoskeleton organization and serve as a framework for further studies to elucidate the mechanism and specific functions of this protein.