Synergistic promotion of c-Src activation and cell migration by Cas and AND-34/BCAR3.

The adapter molecule p130Cas (Cas) plays a role in cellular processes such as proliferation, survival, cell adhesion, and migration. The ability of Cas to promote migration has been shown to be dependent upon its carboxyl terminus, which contains a bipartite binding site for the protein tyrosine kinase c-Src (Src). The association between Src and Cas enhances Src kinase activity, and like Cas, Src plays an important role in cell proliferation and migration. In this study, we show that Src and Cas function cooperatively to promote cell migration in a manner that depends upon kinase-active Src. Another carboxyl-terminal binding partner of Cas, AND-34/BCAR3 (AND-34), functions synergistically with Cas to enhance Src activation and cell migration. The carboxyl-terminal guanine nucleotide exchange factor domain of AND-34, as well as the activity of its putative target Rap1, contribute to these events. A mechanism through which AND-34 may regulate Cas-dependent cell migration is suggested by the finding that Cas becomes redistributed from focal adhesions to lamellipodia located at the leading edge of AND-34 overexpressing cells. These data thus provide insight into how Cas and AND-34 may function together to stimulate Src signaling pathways and promote cell migration.

p130 Cas (Cas) has been implicated in a wide range of cellular processes, including proliferation, survival, cell adhesion and migration, oncogenic transformation, and potentially metastasis (for review see Refs. 1 and 2). Cas was initially identified as a phosphotyrosine (pTyr) 1 -containing protein in cells transformed by the v-src and v-crk oncogenes (3,4). It was subsequently shown to be comprised of multiple protein-protein interaction modules, consistent with its predicted role as an adapter molecule (5). Cas is a component of focal adhesions, which are molecular complexes that form at sites of cell attachment to the extracellular matrix (ECM). The presence of Cas in focal adhesions is consistent with a role for Cas in regulating the actin cytoskeleton and cell migration (6 -9).
One of the major initiating events for cell migration is engagement of integrin receptors by ECM ligands. In this way, integrins directly link ECM components to the actin cytoskeleton (for review see Ref. 10). Integrins also recruit cellular proteins, including kinases and adapter molecules such as Cas, to newly formed focal complexes and mature focal adhesions (for review see Ref. 11). Ultimately, many of the signaling molecules that participate in cell adhesion and migration impact the Rho family of small GTPases, which modify the actin cytoskeleton by regulating actin polymerization and the molecular composition of adhesion sites (12). Elucidating the functions of focal adhesion proteins that participate in the regulation of these small GTPases is critical for understanding mechanisms that govern cell motility.
Several pieces of evidence suggest that Cas plays a critical role in cell migration. Consistent with its presence in focal adhesions, Cas becomes tyrosine phosphorylated in response to integrin ligation by numerous ECM components, including fibronectin (FN), collagen, laminin, and vitronectin (8,(13)(14)(15). In addition, overexpression of Cas has been shown to enhance migration of FG pancreatic carcinoma cells and COS cells (16). Conversely, Cas Ϫ/Ϫ mouse embryo fibroblasts do not migrate as efficiently as their wild-type counterparts unless they are engineered to express ectopic Cas (17,18). Recent studies (19) have shown that the carboxyl terminus of Cas is required to fully rescue this defect in cell migration.
The Cas carboxyl terminus contains a bipartite binding site for the nonreceptor protein tyrosine kinase (PTK) c-Src (Src) (20). This PTK is not only a mediator of cell proliferation and survival (for review see Refs. 21 and 22), but it also plays an essential role in cell migration. Fibroblasts derived from mice lacking three Src family members (Src, Yes, and Fyn) are deficient in migration, and this defect can be rescued by reexpression of Src alone (23). In addition, the kinase activity of Src has been shown to be essential for integrin-mediated cell adhesion, spreading, and migration (24,25). Evidence suggests that Src family kinases mediate integrin-dependent Cas phosphorylation (23, 26 -28) and that Src is the primary PTK responsible for phosphorylation of Cas (29).
Interestingly, Cas is not only a substrate of Src, but also a potent enhancer of Src kinase activity (30,31). Src is typically found in a closed or kinase-inactive conformation, held in place by repressive intramolecular interactions involving its Src homology-2 (SH2) and -3 (SH3) domains (for review see Ref. 32). Because Cas contains motifs that interact with the SH2 and SH3 domains, it can prevent the adoption of this autoinhibitory conformation by directly binding to Src. We and others (30,31,33) have shown that the formation of Src/Cas complexes leads to a significant increase in PTK activity, as well as serum-and anchorage-independent growth. Interactions between Src and Cas may therefore initiate signaling cascades that lead to a number of cellular processes, including cell migration.
In addition to Src, members of a second group of proteins called NSPs (novel SH2-containing proteins) have also been shown to bind to the carboxyl terminus of Cas (34). These proteins contain an amino-terminal SH2 domain and a carboxyl-terminal domain with sequence homology to the Cdc25 family of guanine nucleotide exchange factors (GEFs) for Ras family GTPases. The association between Cas and one of these proteins, Chat (Nsp3), has been shown recently (35) to be involved in the regulation of cell adhesion. A second family member, AND-34/BCAR3 (Nsp2), hereafter referred to as AND-34, also binds to the carboxyl terminus of Cas (36,37). Murine AND-34 has been shown to exhibit GEF activity toward the small GTPases RalA, Rap1, and R-Ras (37). Because all three of these GTPases have been reported to function in various aspects of cell adhesion, migration, and/or proliferation (for review see Refs. 38 -40), it is interesting to speculate that AND-34 may contribute to some of the functions attributed to the carboxyl terminus of Cas-and Src/Cas-dependent signaling pathways.
In this study, we focused on determining whether Cas and AND-34 function cooperatively to regulate cell signaling pathways that lead to Src activation and cell migration. The use of strategies involving protein overexpression is supported for these studies by the fact that both Src and Cas are highly expressed in a number of naturally occurring cellular contexts, including human breast tumors (41)(42)(43). Moreover, the human homologues of both Cas (BCAR1) and AND-34 (BCAR3) were identified in a screen for genes that promoted resistance to the antiestrogen tamoxifen when overexpressed in breast cancer cells (43,44). We found that Cas promotes haptotactic and chemotactic migration in a manner that depends upon kinaseactive Src and that AND-34 functions synergistically with Cas to enhance both cell migration and Src activation. The carboxyl-terminal GEF domain of AND-34 was found to contribute to these activities, as was one of the reported targets of this GEF activity, Rap1. Immunolocalization studies suggest that AND-34 may participate in these processes by relocalizing Cas to lamellipodia located at the leading edge of migrating cells. Together, these data indicate that Cas and AND-34 can function together to stimulate Src signaling pathways and promote cell migration.
Plasmid Construction and Mutagenesis-pRK5 constructs encoding the genes for Myc-tagged full-length Cas, deletion of the minimal carboxyl terminus (⌬mCT; deletion of amino acids 693-874), and the minimal carboxyl terminus (mCT; amino acids 693-874) have been described previously (9,49). Single point mutations substituting a proline residue for leucine 791 (L791P) in full-length Cas and mCT, as well as mutation of leucine 762 (L762P) in mCT, were generated using QuikChange site-directed mutagenesis (Stratagene). The mCT double mutant (L762/791P) was generated by subjecting mCT-L791P to a second round of mutagenesis. All mutations were confirmed by automated DNA sequence analysis. Yellow fluorescent protein (YFP)-tagged Cas was constructed by inserting the complete Cas cDNA as a BamHI/ XbaI fragment into pLP-EYFP-C1 using the Creator TM cloning and expression system (Clontech). pCDNA-c-Src was generously supplied by S. J. Parsons (University of Virginia) (50), and pCDNA3FLAG2ABpaxillin has been described previously (30). FLAG-tagged AND-34 was constructed in pcDNA3 by replacing the polylinker of pcDNA3 with that of pFLAG-CMV2 (pFLAG3). The expressed sequence tag AI115884 was found to contain the entire coding sequence of the mouse BCAR3 homologue, AND-34. The first 24 amino acids of AND-34 were deleted to allow use of an NcoI site to blunt into the newly created pFLAG3 vector, resulting in pFLAG3-AND-34. AND-34 deleted for the GEF-containing carboxyl terminus (⌬GEF) was created from wild-type cDNA by synthesizing a PCR product using a 5Ј primer that contained a unique EcoRI site just upstream of codon 25 and a 3Ј primer containing a unique XbaI site just downstream of codon 593. The PCR product was digested with EcoRI and XbaI and ligated in-frame into pCDNA3FLAG2AB in place of the analogous restriction fragment. A constitutively active RalA construct (pCAG-Ral72L) was generously provided by L. Feig (Tufts University) (51), and constitutively active R-Ras (pCGN-R-Ras87L) was a gift from A. Cox (University of North Carolina) (52). Dominant-negative and constitutively active Rap1B constructs (pCMV-Rap1N17 and pCMV-Rap1V12) were kindly provided by P. Stork (Oregon Health Sciences Center) (53).
Transfection and Protein Expression-Transient transfections in COS-1 and REF-52 cells were performed according to manufacturer specifications using Superfect TM purchased from Qiagen, and C3H10T 1 ⁄2 cell lines were transfected using LipofectAMINE PLUS (Invitrogen). In those cases where cells were co-transfected with two or more constructs, immunofluorescence was performed to confirm that Ͼ90% of transfected cells expressed both proteins. Cells were lysed 24 h post-transfection in modified radioimmune precipitation assay buffer (150 mM NaCl, 50 mM Tris, pH 7.5, 1% Igepal CA-630, 0.5% deoxycholate) supplemented with protease and phosphatase inhibitors (100 M leupeptin, 1 mM phenylmethylsulfonyl fluoride, 0.15 unit/ml aprotinin, 1 mM sodium orthovanadate) as described previously (54). Protein concentrations were determined using the BCA assay from Pierce.
Immunoprecipitation and Immunoblotting-For Myc immunoprecipitates, cell lysate was incubated with mAb 9E10 overnight at 4°C with rotation, and the immune complexes were recovered by a 1-h incubation with protein A-Sepharose beads that had been preincubated with rabbit anti-mouse immunoglobulin. For FLAG immunoprecipitates, cell lysates were incubated with M2 resin (25 l/mg protein) for 1 h at 4°C with rotation. Complexes were washed twice in modified radioimmune precipitation assay buffer and twice in Tris-buffered saline, resuspended in 2ϫ Laemmli sample buffer, and boiled for 5 min. Proteins were resolved by SDS-PAGE, transferred to nitrocellulose, and incubated with antibodies as indicated. Proteins were detected by horseradish peroxidase-conjugated anti-mouse or anti-rabbit immunoglobulin followed by enhanced chemiluminescence (PerkinElmer Life Sciences).
Cell Migration-C3H10T 1 ⁄2 cell lines were transfected with the indicated constructs and cultured overnight in low-serum migration medium (DMEM containing 0.05% FCS). The following day, cells were trypsinized, counted, and a total of 10 5 cells in 500 l of migration medium were loaded into the top of a modified Boyden chamber (24well, 8.0-m BioCoat insert; BD Biosciences). For migration toward FN (Sigma), the underside of the insert membrane was preincubated with FN overnight at 4°C (20 g/ml in phosphate-buffered saline), and the bottom chamber was filled with migration medium. For migration toward serum, uncoated inserts were placed in wells filled with DMEM containing 10% FCS. Cells were allowed to migrate for 2 h at 37°C, and the nonmigratory cells were then removed from the top of the membrane with cotton swabs. The underside of the membrane was fixed in 3% paraformaldehyde for 20 min at room temperature and mounted on glass slides using VectaShield (Vector Laboratories Inc., Burlingame, CA). Migrated cells were counted using a Nikon fluorescence microscope, and the ratio of fluorescent cells to total cells (visualized by phase microscopy) was used to determine percent migration. This was divided by the percent transfection for each condition, resulting in the migratory index (55). The log values for each group of data were analyzed by single factor analysis of variance. When significant differences were found between groups at the 5% level, they were compared using the Student's t test assuming unequal variance. The data are shown as the mean relative migratory index (normalized to vector-transfected cells) Ϯ S.D.
Immunofluorescence-For immunolocalization of AND-34 wild-type and mutant constructs, REF-52 cells were transfected as described above. The following day, cells were replated on FN-coated coverslips (20 g/ml) for 4 -5 h. The cells were then fixed in 3% paraformaldehyde for 20 min at room temperature and permeabilized using 0.4% Triton X-100 in phosphate-buffered saline for 5 min at room temperature. Cells were incubated with FLAG M5 mAb in 2% bovine serum albumin/ phosphate-buffered saline for 1 h, followed by fluorescein isothiocyanate-conjugated goat anti-mouse and TR-conjugated phalloidin for 1 h. For Cas/AND-34 colocalization, transfected REF-52 cells were plated on FN-coated coverslips for 4 -5 h and then fixed, permeabilized, incubated with FLAG M5 and Cas B polyclonal antisera, and subsequently incubated with fluorescein isothiocyanate-conjugated goat anti-rabbit and TR-conjugated goat anti-mouse antibodies. Cells were visualized through a Nikon fluorescence microscope and photographed using a cooled CCD camera controlled by Inovision Isee software.

Cell Migration Is Induced by Overexpression of Src and
Cas-We have shown previously (30) that Src and Cas form stable molecular complexes under conditions in which they are both overexpressed. Using a cell system derived from C3H10T 1 ⁄2-5H murine fibroblasts, which stably overexpress wild-type (WT) Src at levels 16-fold greater than endogenous Src (45), we demonstrated that one functional consequence of these interactions is serum-and anchorage-independent proliferation (30). Because both Src and Cas have been shown to also play a role in cell adhesion and migration, we used the C3H10T 1 ⁄2-5H model to determine whether these molecules functioned cooperatively in these processes. Src-overexpressing C3H10T 1 ⁄2-5H cells were transiently transfected with plasmids encoding YFP or YFP-Cas and analyzed 24 h later for their ability to migrate toward FN or serum (Fig. 1, WT). The relative migratory index represents the percent of migrated fluorescent cells corrected for transfection efficiency and normal-ized to vector-transfected cells. Cas was found to promote a significant 2-fold increase in migration toward FN and serum in the presence of overexpressed WT Src. To determine whether this augmentation of cell migration by Cas was dependent on the kinase activity of overexpressed Src, a similar experiment was performed using C3H10T 1 ⁄2-KD fibroblasts stably overexpressing kinase-dead (KD) Src (46). Cas was unable to promote migration toward either FN or serum in cells expressing KD Src. This deficiency did not appear to be because of an inability of Cas to associate with KD Src, because KD Src-Cas complexes were readily detectable in these cell lysates (data not shown). Thus, it appears from these data that Casdependent cell migration requires Src kinase activity.
AND-34 Synergizes with Cas to Promote Src Activation and Cell Migration-Coincident with their effect on cell migration, interactions between Src and Cas result in the activation of Src, leading to tyrosine phosphorylation of several Src substrates and the promotion of adhesion-and serum-independent growth (30). A recently identified Cas binding partner, AND-34, also demonstrates pro-proliferative effects when overexpressed in breast cancer cell lines grown in the presence of the antiestrogen tamoxifen (44). To determine whether AND-34 participates in Cas-mediated Src activation and substrate phosphorylation, COS-1 cells were transiently co-transfected with plasmids encoding WT Src, Myc-tagged Cas, and/or FLAG-tagged AND-34. PTK activity was determined 24 h posttransfection by measuring the level of pTyr on the Src substrate paxillin, which was expressed as a FLAG-tagged construct, together with the other proteins. Consistent with our previous report (30), co-expression of Src and Myc-tagged Cas resulted in enhanced phosphorylation of paxillin ( Fig. 2A, top  panel, lane 2). Interestingly, co-expression of AND-34 with Cas and Src dramatically increased paxillin phosphorylation above the level observed when Cas and Src were expressed alone (compare lanes 2 and 3). This effect of AND-34 on Src activity was shown to require Cas, as co-expression of AND-34 and Src in the absence of Cas did not stimulate phosphorylation of paxillin (lane 4). Control immunoblots indicated that roughly equal amounts of FLAG-paxillin were immunoprecipitated from cell extracts (second panel) and that AND-34 and Cas were expressed appropriately (bottom panels). AND-34 also enhanced the phosphorylation of other Src substrates in the presence of Cas, including Shc, cortactin, and Src itself at the autophosphorylation site (Tyr-416) (data not shown).
One potential mechanism by which AND-34 may enhance Cas-dependent Src activation is through the stabilization of Src/Cas interactions. To address this possibility, Src immunoblots were performed on Myc immune complexes isolated from the same cell extracts used for the experiment shown in Fig.  2A. The amount of Src associated with Myc-Cas was not affected by AND-34 expression (Fig. 2B, middle panel, compare  lanes 2 and 3). AND-34 expression also had no effect on the absolute expression level of Cas (top panel) or Src (bottom panel) in these cells. These data indicate that the enhancement of Cas-mediated Src activation by AND-34 does not involve a mechanism that alters Src/Cas interactions per se.
Given that AND-34 appeared to synergize with Cas to promote Src substrate phosphorylation, we next investigated whether AND-34 expression also stimulated Cas-mediated migration in cells overexpressing Src. C3H10T 1 ⁄2-5H cells were transfected with plasmids encoding YFP, YFP-Cas, and/or FLAG-AND-34. As was the case in Fig. 1, expression of Cas in these cells increased migration toward both FN and serum (Fig. 2C, black bars). Expression of AND-34 in the absence of Cas had no such stimulatory effect (gray bars). When both Cas and AND-34 were co-expressed, however, a significant increase in migration toward FN and serum was observed relative to both vector-and Cas-transfected cells (hatched bars). Taken together, these data indicate that AND-34 is a potent enhancer of both Cas-mediated Src activation and cell migration.

Cas/AND-34 Interactions Are Not Required for the Promotion of Src Activation and Cell Migration-AND-34
is a member of the NSP family of proteins, which contain an amino-terminal SH2 domain and a carboxyl-terminal region with homology to the Ras GEF Cdc25 (34). To further investigate the functional synergy between Cas and AND-34, we sought to specifically define the region of Cas required for AND-34 binding. Data from other investigators indicated that AND-34 bound to sequences within the extreme carboxyl terminus of Cas, which contains a divergent helix-loop-helix (dHLH) domain (37,56). To determine whether this region, hereafter referred to as mCT (9), was required for Cas/AND-34 binding, COS-1 cells were transiently co-transfected with plasmids encoding FLAGtagged AND-34 and Myc-tagged WT Cas, a Myc-tagged Cas molecule deleted for amino acids 693-874 (⌬mCT), or the in-  ). B, AND-34 expression does not affect the association of Cas and Src. Myc immune complexes were generated from 800 g of the same cell lysate described in A and immunoblotted with Myc or Src antibodies (top two panels, respectively). 40 g of total cell lysate was also immunoblotted for Src (bottom panel). C, AND-34 promotes cell migration only in the presence of co-expressed Cas. C3H10T 1 ⁄2-5H cells were transfected with plasmids encoding YFP, YFP-Cas, FLAG-AND-34, or YFP-Cas, together with FLAG-AND-34. The relative migratory index was determined as described for Fig. 1. The data represent the mean Ϯ S.D. for four to five independent experiments. * indicates p Ͻ 0.05 relative to YFP, and # indicates p Ͻ 0.05 relative to Cas. dependently expressed Myc-tagged mCT (Fig. 3A). Myc immune complexes were isolated from cell extracts and immunoblotted for Myc and FLAG (Fig. 3B). AND-34 was observed to associate with WT Cas under these conditions (third panel, lane 2) but not with ⌬mCT (lane 3). AND-34 was also able to associate with the independently expressed mCT molecule (lane 4). The reduced recovery of AND-34 in mCT immune complexes relative to full-length Cas was likely caused by less AND-34 expression in this experiment (bottom panel, compare  lanes 2 and 4). However, the fact that equal amounts of AND-34 were present in cells expressing ⌬mCT, which did not bind to AND-34, and mCT, which was found to associate with AND- 34 (bottom panel, lanes 3 and 4), is consistent with the carboxyl-terminal mCT region of Cas being both necessary and sufficient for association with AND-34.
To address whether the helices of the dHLH motif were themselves critical for AND-34 binding to Cas, single amino acid substitutions were engineered into the mCT construct, replacing conserved leucine residues at positions 762 of the first helix and 791 of the second helix with prolines. Mutation of the analogous leucines in the Cas family member HEF1 has been shown to disrupt dHLH functions associated with proteinprotein interactions and subcellular localization (56,57). WT and mutant Myc-tagged mCT molecules were expressed together with FLAG-tagged AND-34 in COS-1 cells, and Myc immune complexes were then examined for the presence of FLAG-AND-34 (Fig. 3C). Mutation of leucines 762 and 791, either singly or in combination, completely abrogated AND-34 binding to mCT (middle panel, lanes 2-4). These data suggest either that leucines 762 and 791 are responsible for directly binding to AND-34 or that structures governed by the dHLH motif play an essential role in the association of Cas and AND-34.
Having delineated specific residues of Cas required for AND-34 binding, we next addressed whether the ability of AND-34 to enhance Cas-dependent Src activation required the association of Cas and AND-34. COS-1 cells were transfected with plasmids encoding Src, FLAG-paxillin, FLAG-AND-34, and either WT Myc-Cas or a full-length Myc-Cas molecule containing the L791P mutation. PTK activity, as measured by phosphorylation of paxillin, was found to be identical in the presence of WT Cas or Cas-L791P (Fig. 4A, top panel, compare  lanes 1 and 2), despite the fact that AND-34 binding to Cas-L791P was severely compromised (Fig. 4B, middle panel, lane  2). Importantly, Src/Cas interactions were maintained in the presence of the L791P mutation (bottom panel). These data indicate that synergistic enhancement of Cas-dependent Src activity by AND-34 does not require a stable association between Cas and AND-34.
Function of the AND-34 Carboxyl Terminus-The finding that AND-34 does not need to stably associate with Cas to augment Cas-dependent Src activation suggested that its role in this process may be related to functions of the molecule that are independent of Cas binding. Among its putative functional domains, AND-34 contains sequences within its carboxyl terminus that have homology to GEF domains, and expression of this molecule has been reported to coincide with activation of the small GTPases R-Ras, RalA, and Rap1 (36,37). To address whether this domain was required for promotion of Src activity by Cas, the carboxyl terminus of AND-34 was deleted to produce a mutant lacking putative GEF activity (⌬GEF). WT and mutant FLAG-tagged AND-34 molecules were expressed together with myc-Cas and Src in COS-1 cells, and PTK activity was measured. Endogenous cortactin was used to measure Src activity in this case instead of paxillin, because FLAG-paxillin and the AND-34 ⌬GEF molecule co-migrated on SDS-PAGE.
As was the case for paxillin, cortactin phosphorylation was elevated in the presence of overexpressed Cas and Src ( Because Src activation by Cas and AND-34 correlated with promotion of cell migration (see Fig. 2), we investigated whether the inability of the AND-34 ⌬GEF mutant to augment Cas-dependent Src activity was coincident with a similar inability to promote cell migration. Whereas AND-34 expression was again found to significantly enhance the effect of Cas on cell migration toward FN (Fig. 5B) and serum (data not shown), expression of ⌬GEF had no such effect. Together, these data indicate that the carboxyl terminus of AND-34 is functionally important for the promotion of both Src activity and cell migration by Cas.
Functions Associated with the GEF Domain of AND-34 -Because the GEF domain located at the carboxyl terminus of AND-34 has been reported to have activity toward RalA, Rap1, and R-Ras (37), we hypothesized that enhancement of Src activation and Src/Cas-dependent cell migration by AND-34 may occur through a mechanism that involves activation of one or more of these GTPases. If this is the case, then activated forms of these GTPases would be expected to function like AND-34 to augment the positive effect of Cas on Src activity and cell migration. To test this hypothesis, COS-1 cells were co-transfected with plasmids encoding Src, Cas, paxillin, and activated forms of R-Ras (R-Ras 87L), Rap1 (Rap1 V12), or Ral (Ral 72L). None of these constitutively active GTPases had any effect on paxillin phosphorylation in the absence of Cas (Fig. 6,  A (top panel, lanes 2 and 3) and B (top panel, lane 2)). However, in the presence of Cas, Rap1 V12 dramatically enhanced the level of paxillin phosphorylation above the level induced by Cas alone (Fig. 6A, top panel, compare lanes 4 and 6). Active Ral had no such stimulatory effect (Fig. 6B, top panel, compare  lanes 3 and 4), and active R-Ras appeared to actually inhibit Cas-induced paxillin phosphorylation (Fig. 6A, top panel, compare lanes 4 and 5). From these data, it appears that constitutively active Rap1, but not Ral or R-Ras, can stimulate Cas-dependent Src activation in the absence of AND-34.
If Rap1 activation by AND-34 is required for augmentation of Src activity by Cas, then expression of a dominant-inhibitory Rap1 molecule would be expected to inhibit the positive effect of AND-34 on Src activation. To determine whether this was the case, Cas, Src, paxillin, and AND-34 were expressed in COS-1 cells in the absence or presence of dominant-inhibitory Rap1 N17. FLAG-paxillin immunoprecipitated from these cell extracts was heavily tyrosine phosphorylated in the presence of Cas and AND-34 (Fig. 6C, top panel, lane 4). Expression of Rap1 N17 reduced this phosphorylation (lane 5), albeit not to the level observed when Cas was expressed in the absence of AND-34 (lane 2). Rap1 N17 also caused a slight decrease in paxillin phosphorylation in the presence of Cas alone (compare lanes 2 and 3). Based on these data, it appears that Rap1 can function downstream of AND-34 and Cas in the promotion of Cas-mediated Src activation. The incomplete inhibition of Src activation in the presence of dominant-negative Rap1 could be the result of either technical issues that prevented complete inhibition of endogenous Rap1 activity or the contribution of additional signals emanating from Rap1-independent pathways.
We next investigated whether the stimulatory effect of AND-34 on Cas/Src-dependent cell migration was dependent on Rap1 activity. As before, expression of YFP-Cas in Srcoverexpressing C3H10T 1 ⁄2-5H cells promoted cell migration toward serum, and co-expression of AND-34 caused a further enhancement in cell migration (Fig. 6D, white bars). Rap1 N17 blocked the up-regulation of Cas-promoted migration by AND-34 but had no statistically significant effect on migration in the absence of AND-34 (black bars). These data, together with the data on Src activation shown in Fig. 6, A-C, implicate Rap1 as a key mediator of AND-34 signaling through Src/Cas.

Subcellular Localization of Cas and AND-34 -
The ability of AND-34 to promote cell migration though Src/Cas signaling pathways raises the question of whether the function of these molecules is coordinated at specific sites within the cell. To examine the subcellular localization of AND-34 in adherent cells, REF-52 cells were transiently transfected with plasmids encoding FLAG-tagged AND-34 or ⌬GEF and immunostained with FLAG antibodies to visualize AND-34, as well as phalloidin to visualize polymerized actin (Fig. 7). WT AND-34 was strongly localized to membrane ruffles and lamellipodia in areas where significant cortical actin was detected (panels A-C). Interestingly, the carboxyl terminus of AND-34 did not appear to be required for this localization, because ⌬GEF was also present at the cell periphery (panels D-F). To investigate whether Cas localization was altered in the presence of overexpressed AND-34, REF-52 cells were transfected with plasmids encoding Cas alone or in combination with WT or mutant AND-34 molecules (Fig. 8). In the absence of AND-34, Cas was localized predominantly to discrete focal adhesions and perinuclear regions of the cell (panel A). When co-expressed with WT AND-34, Cas showed a dramatic relocalization to membrane ruffles and lamellipodia, coincident with AND-34 localization (panels B-D). Approximately 80% of co-expressing cells also underwent a change in morphology, marked by the adoption of a phenotype that exhibited classical leading and trailing edges, with Cas and AND-34 localized primarily at the leading edge. Cells expressing ⌬GEF did not exhibit this polarized morphology, although Cas still co-localized with AND-34 at membrane ruffles (panels E-G). Thus, the enhancement of cell migration exhibited by cells co-expressing Cas and WT AND-34 coincided with both the relocalization of Cas to lamellipodia and the acquisition of a polarized phenotype. DISCUSSION This study focuses on elucidating mechanisms by which two molecules, Cas and AND-34, modulate Src kinase activity and promote cell migration. Throughout these studies, we observed a correlation between conditions that resulted in the activation of Src, as determined by the phosphorylation of several Src substrates, and conditions that promoted cell migration. For example, Cas was shown to serve as a potent enhancer of Src activity, as well as cell migration through a pathway that

FIG. 5. The carboxyl terminus of AND-34 is required for enhancement of Cas-mediated Src activation and cell migration.
A, deletion of the carboxyl terminus of AND-34 ablates the enhancement of Cas-mediated Src activation. COS-1 cells were transfected with plasmids encoding Src, Myc-Cas, and either FLAG-tagged WT AND-34 or ⌬GEF. 40 g of total cell lysate was immunoblotted for pY421 (a phospho-specific antibody for cortactin), cortactin, Myc, Src, or FLAG as indicated. B, deletion of the carboxyl terminus does not allow promotion of Cas-mediated cell migration. C3H10T 1 ⁄2-5H cells were transfected with plasmids encoding YFP, YFP-Cas, or YFP-Cas, together with either FLAG-tagged WT AND-34 or ⌬GEF. The relative migratory index was determined as described for Fig. 1. The data presented are the mean Ϯ S.D. for four independent experiments. * indicates p Ͻ 0.05 relative to YFP, # indicates p Ͻ 0.05 relative to Cas, andˆindicates p Ͻ 0.05 relative to Cas ϩ AND-34. required kinase-active Src. Similarly, AND-34 expression augmented both Cas-mediated Src activation and cell migration. This close correlation among Src activation, substrate phosphorylation, and cell migration raises the possibility that these events are functionally related, as opposed to being distinct activities that are independently controlled by Src, Cas, and AND-34. The increased Src activity promoted by Cas and AND-34 may directly impact cell motility, because Src expression and kinase activity play an important role in cell migration (23)(24)(25). Moreover, the profound redistribution of Cas from focal adhesions and focal complexes to lamellipodia at the leading edge of cells overexpressing AND-34 may contribute to the enhancement of cell migration observed when Cas, Src, and AND-34 are overexpressed.
Regulation of Src Activity by Cas, AND-34, and Rap1-The ability of AND-34 to enhance Src activation and cell migration was found to be absolutely dependent on co-expression of Cas. AND-34 expression in the absence of Cas was unable to stimulate Src-PTK activity or the migration of Src-overexpressing C3H10T 1 ⁄2-5H cells, suggesting that AND-34 does not directly regulate Src PTK activity. This is further supported by our findings that the relative steady-state level of Src/Cas complexes was not affected by AND-34 expression and that there was no requirement for direct binding of AND-34 to Cas. These data thus argue against a direct effect of AND-34 on the PTK activity of Src/Cas complexes.
One mechanism through which AND-34 may coordinate with Src and Cas in these processes involves phosphorylation of AND-34 by Src. AND-34 is a phosphoprotein that becomes tyrosine phosphorylated upon stimulation by serum and adhesion to FN (36). Interestingly, in the present study, AND-34 was consistently observed to be phosphorylated in a manner similar to other Src substrates when Cas and Src were overexpressed (data not shown). Although the effect of tyrosine phosphorylation on AND-34 function is currently not known, there is precedent for the regulation of GEF activity in other molecules by tyrosine phosphorylation. For example, the GEF activities of RasGRF1, Vav, and C3G are stimulated by tyrosine phosphorylation, and RasGRF1 in particular is activated following tyrosine phosphorylation by Src (58 -60). These reports raise the important question of whether AND-34 is a substrate of Src whose GEF activity can be regulated by phosphorylation. We are currently testing this possibility.
FIG. 6. Rap1 participates in the enhancement of Cas-mediated Src activation and migration by AND-34. A, constitutively active Rap1, but not R-Ras, enhances paxillin phosphorylation in the presence of Src and Cas. Plasmids encoding Myc-Cas, Src, FLAG-paxillin, and either Myc-R-Ras 87L or Rap1 V12 were transiently transfected into COS-1 cells. FLAG immune complexes were isolated from 400 g of cell lysate, divided in half, and immunoblotted with pTyr (top panel) or FLAG (second panel) antibodies. 40 g of total cell lysate was also immunoblotted for Myc, Src, or Rap1 to confirm appropriate expression of these constructs (lower panels). B, activated Ral does not promote paxillin phosphorylation. COS-1 cells were transfected with plasmids encoding Myc-Cas, Src, FLAG-paxillin, and Ral 72L. 400 g of cell lysate was immunoprecipitated and immunoblotted as described for A (top two panels), and 40 g of total cell lysate was immunoblotted for Myc, Src, and Ral to confirm appropriate expression of these constructs (lower panels). C, dominant-negative Rap1 inhibits AND-34-promoted paxillin phosphorylation by Src/Cas. Plasmids encoding Src, FLAGpaxillin, Myc-Cas, FLAG-AND-34, and/or Rap1 N17 were transfected into COS-1 cells. 400 g of cell lysate was immunoprecipitated and immunoblotted as described for A (top two panels), and 40 g of total cell lysate was immunoblotted for Myc, Src, FLAG, and Rap1 to confirm appropriate expression of these constructs (lower panels). D, dominantnegative Rap1 inhibits AND-34 enhancement of Src/Cas-dependent migration. C3H10T 1 ⁄2-5H cells were transfected with plasmids encoding YFP, YFP-Cas, or YFP-Cas, together with FLAG-AND-34 in the presence or absence of Rap1 N17. The relative migratory index was determined as described for Fig. 1. The data presented are the mean Ϯ S.D. for four independent experiments. * indicates p Ͻ 0.05 relative to YFP, # indicates p Ͻ 0.05 relative to Cas, andˆindicates p Ͻ 0.05 relative to Cas ϩ AND-34.
The data presented in this report suggest that the carboxyl terminus of AND-34, which contains the putative GEF domain, is important for its ability to synergize with Cas and Src. First, AND-34 molecules lacking the carboxyl terminus were not capable of enhancing Cas-mediated Src activation or cell migration. Second, constitutive activation of one of the reported targets of AND-34 GEF activity, Rap1, was shown to mimic the effect of AND-34 on Cas-mediated Src activation. Finally, a dominant-negative form of Rap1 inhibited AND-34-dependent enhancement of Cas-mediated Src activation and cell migration. Thus, the ability of AND-34 to synergize with Cas appeared to be at least partly dependent on the physical presence of the AND-34 carboxyl terminus and the activity of Rap1. Together, these data are consistent with a model in which the initial activation of Src that is caused by Cas binding is sufficient to phosphorylate AND-34, resulting in stimulation of GEF activity and subsequent activation of Rap1. This in turn may result in a further activation of Src PTK activity and more robust promotion of cell migration.
There are several recent reports that document a relationship between Rap1 activities and Src/Cas functions. Rap1 was shown by Xing et al. (31) to function downstream of Src/Cas complexes to activate transcription through serum response elements and AP-1 sites. In this study, we provide evidence that Rap1 can also function upstream of Src/Cas to induce Src substrate phosphorylation. To the best of our knowledge, this is the first time that Rap1 has been shown to regulate what is typically considered to be an upstream kinase. The mechanism by which Rap1 is able to enhance Src activation and substrate phosphorylation is not understood. Rap1 shares several of the same effectors as Ras, including B-Raf, Ral-specific GEFs, and other molecules such as AF-6, phosphatidylinositol 3-kinase, and RasGAP (reviewed in Ref. 39). Whether any of these downstream targets of Rap1 play a role in its enhancement of Src activation is not known and represents an intriguing area for further study.
An alternative process by which Rap1 could affect Src activation is through modulation of integrin affinity and/or avidity via inside-out signaling. A number of studies (39,(61)(62)(63)(64) in several cell types have shown that Rap1 can control integrinmediated adhesion and potentially induce changes in integrin affinity/avidity for ECM ligands. In turn, Src activity becomes elevated upon integrin activation and ligand binding (15,65,66). Interestingly, in cells expressing ␣ IIb ␤ 3 integrins, Rap1 can augment ligand binding and affinity/avidity. Src has been shown to be constitutively associated with this receptor and becomes strongly activated in response to ␣ IIb ␤ 3 -mediated adhesion (67,68). Although Rap1 has not been shown previously to function upstream of Src, a second putative target of AND-34 GEF activity, Ral, has been reported to promote Src-dependent phosphorylation of the substrates Stat3 and cortactin in response to epidermal growth factor stimulation (69). Moreover, dominant-negative Ral was shown to inhibit epidermal growth factor stimulation of Src activity. In contrast to these findings, we did not observe increased phosphorylation of the Src substrate paxillin in the presence of activated Ral. This suggests that the potential function of Ral as an activator of Src is likely to be dependent on cell context, as well as the mechanism of activation.
Regulation of Cell Migration by Cas, AND-34, and Rap1-A role for Rap1 in cell migration is somewhat better established. In several cases, Rap1 has been shown to be a positive regulator of cell migration (70,71), and one of its targets, the extracellular signal-regulated kinase/mitogen-activated protein kinase signaling cascade, is known to be important for cell migration and invasion (18,72). However, there is also evidence to suggest that Rap1 activation may be inhibitory for cell migration (73). For example, Ohba et al. (74) showed that fibroblasts lacking the Rap1 GEF C3G exhibited reduced cell adhesion and increased cell migration as compared with wildtype mouse embryo fibroblasts and that expression of activated Rap1 in the C3G-null cells resulted in decreased migration and greater cell adhesion. Ultimately, the effects of Rap1 on cell migration are also likely to be dependent on cell context and the nature and concentration of the migratory stimulus.
It is important to point out that there is currently some debate concerning the relevant targets of the AND-34 GEF domain and whether this molecule is in fact capable of functioning as an authentic GEF (39,75). We have not specifically tested the requirement for direct activation of Rap1 by AND-34 in this study. Consequently, it is formally possible that Rap1 may function downstream of AND-34 through a mechanism that is independent of AND-34 GEF activity per se. Sakakibara et al. (35) showed in a recent report that the AND-34 family member Chat can induce Rap1 activation through a pathway that involves Cas signaling to C3G. Phosphorylation of the YXXP motifs in the substrate binding domain of Cas can result in its association with the adapter protein Crk (5), which is constitutively bound to the Rap1 GEF C3G (reviewed in Ref. 76). The possibility that AND-34 may function in a similar manner could help to reconcile conflicting data regarding the ability of AND-34 to directly activate Rap1 and other GTPases (37,39,75).
Finally, AND-34 may contribute to Cas-induced cell migration through its ability to relocalize Cas to membrane ruffles. Localized assembly of Cas/Crk complexes occurs at the cell periphery and can lead to activation of Rac1 through the recruitment of a second Crk-binding protein, DOCK180 (18,77,78). In this way, relocalization of Cas by AND-34 may activate Rac1 at the leading edge, resulting in the extension of lamellipodia and the induction of a polarized phenotype marked by discrete leading and trailing edges that are typically exhibited by migrating cells (79). This is consistent with previous data from a number of other investigators who characterize Cas/Crk as a "molecular switch" that promotes cytoskeletal reorganization and cell migration (16,18,72).
The mechanisms by which AND-34 recruits Cas to the cell periphery remain to be determined, although some clues can be gained from the observation that Cas was efficiently localized to the cell periphery in cells expressing both the WT and ⌬GEF proteins. The putative GEF activity and Cas binding functions of AND-34 are both missing from the ⌬GEF construct (36,37). Therefore, these activities appear to be dispensable for Cas re-localization by AND-34. However, the AND-34 carboxyl terminus was found to be important for adoption of the polarized phenotype exhibited by cells co-expressing Cas and AND-34, because polarization was not observed in ⌬GEF-expressing cells. Determination of the specific features of the AND-34 carboxyl terminus that are required for the induction of this phenotype will greatly enhance our understanding of the mechanisms governing cell polarity and motility.