p130Cas Regulates the Activity of AND-34, a Novel Ral, Rap1, and R-Ras Guanine Nucleotide Exchange Factor*

We previously identified a novel murine protein, AND-34, with a carboxyl-terminal domain homologous to Ras family guanine nucleotide exchange factors (GEFs), which bound to the focal adhesion docking protein p130Cas. Work by others has implicated both the human homologue of AND-34, BCAR3, and human p130Cas, BCAR1, in the resistance of breast cancer cells to the anti-estrogen tamoxifen. Here we report that AND-34 displays GEF activity on RalA, Rap1A, and R-Ras but not Ha-Ras GTPases in cells. In contrast to several other Ral-GEFs, the Ral GEF activity of AND-34 is not augmented by constitutively active Ha-RasVal-12, consistent with the absence of a detectable Ras-binding domain. Efficient binding to AND-34 required both the Src-binding domain and a flanking carboxyl-terminal region of p130Cas. The p130Cas-binding site mapped to a carboxyl-terminal sequence within the AND-34 GEF domain. Overexpression of p130Cas, but not an AND-34-binding mutant of p130Cas, inhibited the Ral GEF activity of co-transfected AND-34. This work identifies a new potential function for p130Cas and a new regulatory pathway involved in the control of Ral, Rap, and R-Ras GTPases that may participate in the progression of breast cancer cells to tamoxifen resistance.

The Ras superfamily of GTPases are 20 -30-kDa guaninenucleotide binding proteins that play a central role in the regulation of cell proliferation, differentiation, and activation (1). The superfamily can be divided into many subfamilies including Ras, Rho, Rap, Arf, and Ran. All members of the superfamily cycle between inactive GDP-bound and active GTP-bound states. Guanine nucleotide exchange factors (GEFs) 1 with specificity for different Ras family members induce the dissociation of GDP, allowing the more abundant GTP to bind and activate these GTPases. In an effort to identify genes transcriptionally up-regulated following the induction of thymocyte apoptosis by cross-linking CD3⑀ in vivo, we recently cloned a novel murine cDNA, AND-34, which contains a carboxyl-terminal domain similar to GEFs for one or more GTPases of the Ras subfamily of GTPases, including Ras, Ral, Rap, R-Ras, and TC21 (2). In contrast to previously reported Ras subfamily GEFs, AND-34 contained an SH2 domain at its amino terminus.
We also observed that AND-34 forms a complex in cells with p130 Cas , a known c-Src substrate and a docking protein that binds to the Crk adaptor protein, focal adhesion kinase, PYK2/ RAFTK tyrosine kinases, and PTP-PEST or PTP1B phosphatases (3). Consistent with such interactions, AND-34 undergoes tyrosine phosphorylation following adhesion of trypsinized fibroblasts to fibronectin-coated plates (2). AND-34 is the murine homologue of NSP-2, a cDNA characterized as one of three closely related family members referred to as NSP1, NSP-2, and NSP-3 (4). NSP-1 was also shown to be tyrosine-phosphorylated following treatment of cells with EGF or insulin.
The human homologue of AND-34 was detected in a study designed to detect cDNAs whose expression converts tamoxifen-sensitive breast cancer cells to tamoxifen insensitivity (5).
Remarkably, the same screen has also recently detected p130 Cas (6), adding support to the notion that p130 Cas and AND-34 function in a common signaling pathway to regulate hormone-dependent proliferation of breast cancer cells. Additional support for the relevance of these findings is that AND-34 expression correlates with loss of estrogen receptors in breast cancer cell lines (5). Furthermore, high levels of p130 Cas in primary human breast carcinomas are associated with poor relapse-free survival, poor overall survival, and a reduced response rate to tamoxifen in patients with recurrent disease (7).
In the study described below, we have shown that AND-34 is an activator of Ral and to a lesser extent R-Ras and Rap1A GTPases. Our mapping studies and our in vivo GDP exchange assays suggest that p130 Cas inhibits the GEF activity of AND-34.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfections-HEK293 human embryonic kidney cells, a gift from Dr. Barbara Slack (Boston University School of Medicine) (8), were grown in DMEM supplemented with 10% fetal bovine serum (FBS), 2 mM L-glutamine, 100 units/ml penicillin, 100 g/ml streptomycin, 10 mM Hepes, pH 7.4, 10 M 2-mercaptoethanol. 293 cells were trypsinized 1 day prior to transfection and transferred to 6-well (35 mm) plates. Cells were 80% confluent on the day of transfection. For each well of the plate, 2.5 g of plasmid DNA was mixed with 5 l of DNA PLUS reagent (Life Technologies, Inc.) in 125 l of serum-free DMEM at room temperature for 15 min. 7.5 l of LipofectAMINE (Life Technologies, Inc.) in 125 l of serum-free DMEM was then added for a further 15 min. The 250-l final mixture was then added evenly to cells covered with 0.6 ml of fresh DMEM with 10% FBS and incubated at 37°C, 5% CO 2 , for 3 h. One ml of DMEM with 10% FBS was then added to each well.
Immunoprecipitation and Immunoblotting-24 h after transfection, cells were washed with PBS and lysed in ice for 1 h with 1 ml of lysis * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  buffer (50 mM Hepes, pH 7.4, 150 mM NaCl, 1% Triton X-100, 10% glycerol, 1 mM EGTA, 1.5 mM MgCl 2 , 20 mM NaF, 1 mM sodium vanadate, 1 mM phenylmethylsulfonyl fluoride, 4 g/ml aprotinin, 2 g/ml leupeptin, and 2 g/ml pepstatin). The protein supernatant was collected after centrifugation at 14,000 rpm (relative centrifugal force ϭ 16,000) at 4°C for 10 min. 0.5 mg of protein was mixed with the primary antibody for 2 h at 4°C. 25 l of protein A/G-agarose beads (Santa Cruz Biotechnology) were added and incubated overnight at 4°C. The beads were washed three times with 1 ml of lysis buffer, and protein was released from beads by boiling them with 20 l of 2ϫ sample buffer for 5 min. The supernatant was run on SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose membrane, and the immunoblot was developed by ECL as described previously (2).
Measurement of GEF Activity in Vivo-Cos7 cells (5 ϫ 10 5 per 60-mm dish on the day prior to transfection) were transfected by the DEAEdextran method with expression vectors containing GST-GTPases, GEFs, and/or full-length or SBD-deleted p130 Cas , as described in the figure legends. 48 h later, cells were labeled for 4 h in phosphate and serum-free media with 0.05 mCi/ml of 32 P i (ICN). GST-fused GTPases were collected by incubation of cell lysates with glutathione-agarose beads (Sigma) for 1 h and washed with lysis buffer. GTPase-bound 32 P-labeled guanine nucleotides were separated by thin layer chromatography (Andrich) and quantitated by PhosphorImager analysis (Molecular Dynamics) as described previously (9). The data are presented as a ratio of GTP/(GDP ϩ GTP).
Four hemagglutinin epitope-tagged p130 Cas Src-binding domain mutants were constructed using an identical cloning strategy to that described above for the AND-34 mutants, with the exception that the 5Ј oligonucleotides contained a flanking BamHI site. Two 5Ј oligonucleotides, 5PSBD-L and 5PSBD-S, contained 21 nucleotides distal to nucleotides 1693 and 1947 (residues 553 and 638), respectively, of rat p130 Cas (GenBank TM accession number D29766). The two 3Ј oligonucleotides, 3PSBD-L and 3PSBD-S, contained the inverse complement of 21 nucleotides proximal to nucleotides 2920 and 2706 (residues 968 and 890), respectively, of p130 Cas . Sequence encoding a core 253-residue region (residues 638 -890) previously shown to be required for Src binding to p130 Cas was amplified with 5PSBD-S and 3PSBD-S oligonucleotides and subcloned to generate SBD-Core (predicted size 27.9 kDa) (11). A 415-residue sequence encoding 84 residues NH 2 -terminal and 78 residues COOH-terminal to the core SBD region was amplified with 5PSBD-L and 3PSBD-L and subcloned to generate SBD-Long (residues 553-968, predicted size ϭ 46.8 kDa). Two additional sequences corresponding to the core SBD combined with either flanking NH 2 -terminal (NH 2 -SBD) or COOH-terminal (SBD-COOH) sequence were also amplified and subcloned using the oligonucleotides described above (predicted sizes of 38.4 and 37.8 kDa, respectively).
Myc Epitope-tagged AND-34 -Full-length AND-34 was amplified by PCR with Deep Vent DNA polymerase using oligonucleotides with flanking XhoI (5Ј oligonucleotide) and EcoRI (3Ј oligonucleotide) restriction sites. XhoI/EcoRI-digested PCR product was ligated to similarly digested pcDNA3.1/Myc-His B vector (Invitrogen). Myc-tagged AND-34 deletion mutants were similarly cloned using the same oligonucleotide sequences described for the HA-tagged mutants.

AND-34 Stimulates GDP Release from Ral, Rap, and R-
Ras-The existence of a CDC25-like domain at the COOH terminus of AND-34 suggested that the protein had GEF activity against members of the Ras subfamily of Ras superfamily GTPases. The ability of AND-34 to promote the active GTPbound state of a variety of Ras subfamily members in cells, including RasH, Rap1A, R-Ras, and RalA was tested by cotransfection of individual GST-GTPase fusion constructs along with AND-34 into Cos cells. After labeling cells with 32 P, the proportion of GTPase associated with GTP in cells was assessed by isolation of the GTPase on glutathione beads and quantitation of bound nucleotides by thin layer chromatography. In these assays, AND-34 increased the proportion of GTP RalA in cells from ϳ15% GTP to 30% GTP. In comparison, transfection of the same amount of DNA encoding Ral-GDS, a known GEF for Ral, elevated Ral-GTP levels to only ϳ22% (Fig.  1A). AND-34 was less effective at activating R-Ras, elevating its GTP levels from ϳ15 to ϳ23% (Fig. 1B). AND-34 was not as effective as Ras-GRF1, a known R-Ras GEF, which increased R-Ras GTP levels to ϳ29%. AND-34 also elevated Rap1A GTP levels from 5% GTP to ϳ11% GTP. It was not as effective as the known Rap1A GEF, C3G, which increased Rap-GTP levels to 22% (Fig. 1C). In contrast to RalA, R-Ras, and Rap1A, RasH was unresponsive to AND-34 expression even though RasH did become activated by the known Ras GEF, SOS, under the same conditions (Fig. 1D). Thus, AND-34 appears to be a multifunctional GEF with the ability to activate RalA, R-Ras, and Rap1A but not Ha-Ras.
Most fully characterized Ral-GEFs isolated to date have a Ras-binding domain that allows them to become activated by GTP-Ras. However, inspection of the amino acid sequence of AND-34 failed to reveal such a binding site. As expected, co- transfection of a constitutively activated Ras did not increase the GEF activity of AND-34 GEF as it did for Ral-GDS which contains a Ras-binding site. Interestingly, for reasons that are not yet clear, activated Ras appeared to decrease the activity of AND-34.
Association of the AND-34 Carboxyl Terminus with p130 Cas -We have previously shown that AND-34 associates with p130 Cas , a focal adhesion docking protein (2,12). In order to map the site of interaction between these two proteins, we constructed a series of NH 2 -terminal hemagglutinin (HA) epitope-tagged deletion mutants of AND-34 (Fig. 2). After transient expression of the mutants in HEK293 cells, anti-HA immunoprecipitates were immunoblotted with anti-p130 Cas antibody to detect complexes with endogenous p130 Cas . Deletion of the SH2 domain (⌬SH2) and proline-rich domain (⌬SH2/ Pro) did not significantly reduce binding of AND-34 to p130 Cas (Fig. 3). This result was surprising as the polyproline-rich region of AND-34 contains a motif that matched a consensus site (XPpϩP) for binding to the SH3 domain of p130 Cas (2).
In contrast, each of four constructs lacking the GEF domain (SH2, SH2/Pro, ⌬GEF, and ⌬SH2/GEF) failed to associate with p130 Cas (Fig. 3). Immunoblots with anti-HA antibody of the corresponding whole cell lysates demonstrated that the HAtagged truncation mutants were of appropriate size and were expressed with comparable efficiency (Fig. 3).
To delineate further the region of AND-34 responsible for binding to p130 Cas , we utilized the ⌬SH2 mutant to construct three further mutants in which 70, 140, or 210 amino acid portions of the GEF domain were removed from the carboxyl terminus of AND-34 (Fig. 2, ⌬SH2/GEF750, -680, or -610, respectively). In contrast to wild type AND-34, none of the proteins expressed by these constructs bound to p130 Cas , suggesting that at least the terminal 70 amino acids of the AND-34 GEF domain are required for this interaction (Fig. 4). However, an AND-34 mutant containing solely the GEF domain failed to bind to p130 Cas (Fig. 4)

in cells.
To confirm these findings, we synthesized an independent set of deletion mutants in which the Myc epitope tag was placed on the AND-34 carboxyl terminus (nomenclature identical to that of HA-tagged mutants). Once again, immunoprecipitation of three AND-34 mutants (SH2, SH2/Pro, and ⌬SH2/GEF) that lacked the GEF domain failed to associate with p130 Cas in these assays (data not shown). In aggregate, these experiments demonstrate that the minimum domain required for AND-34 association with p130 Cas in cells includes the GEF domain and at least 70 amino acids amino-terminal to it.
Both the Src-binding Domain (SBD) and the Flanking COterminal Region of p130 Cas Are Required for Association with AND-34 in Cells-In order to identify the region of p130 Cas required for association with AND-34, we transiently co-transfected a full-length Myc-tagged AND-34 construct in HEK293 cells with HA-tagged deletion mutants of p130 Cas (see Fig. 5) (11). Deletion of the SH3 domain (⌬SH3) failed to significantly reduce association of p130 Cas with AND-34 (Fig. 6). This finding is consistent with the AND-34 deletion analysis, which showed that its proline-rich region was not needed for binding to p130 Cas . Deletion of the substrate domain (⌬SD) of p130 Cas also failed to significantly reduce association of p130 Cas with AND-34 (Fig. 6). In contrast, deletion of the 253-residue Srcbinding domain (⌬SBD) eliminated the association of these two proteins (Fig. 6). Mutation of either an RPLP sequence or Tyr-762 within the p130 Cas SBD, both of which have previously been shown to reduce the association of p130 Cas with Src (11), failed to abrogate binding of AND-34 to p130 Cas (Fig. 6, RPLP  and Y762F). Thus, although the ⌬SBD failed to bind AND-34, Src itself did not tether AND-34 to p130 Cas .
To delineate better the AND-34-binding site on p130 Cas , we constructed four HA epitope-tagged forms of the SBD. SBD-Core consisted of a 253-residue peptide (638 -890, numbering according to Sakai et al. (3,11)) corresponding precisely to the region deleted in ⌬SBD (see Fig. 5). As shown in the top and middle panels of Fig. 7, immunoprecipitates of either Myctagged AND-34 or HA-tagged-SBD failed to demonstrate association of these two proteins. Whole cell lysates showed expression of the appropriate 27-kDa HA-tagged peptide (Fig. 7, lower  panel). Thus, whereas the SBD region is necessary for p130 Cas to bind to AND-34, it is not sufficient for such binding. Three additional HA epitope-tagged constructs were tested to assess whether 84 NH 2 -terminal residues or COOH-terminal 78 residues flanking the p130 Cas SBD contribute to AND-34 binding. Although all three of these constructs were expressed well in transfected 293 cells, only the constructs containing COOHterminal flanking sequences along with the SBD-Core efficiently bound to AND-34 (see SBD-Long and COOH-SBD versus SBD-NH 2 in Fig. 7).

Full-length but Not SBD-deleted p130 Cas Inhibits AND-34
Ral GEF Activity-The surprising observation that AND-34 requires its GEF domain for p130 Cas binding suggests that such binding modulates AND-34 GEF activity. To test this hypothesis, we examined the effect of overexpression of p130 Cas on the Ral GEF activity of AND-34 by the previously described in vivo GEF assay (Fig. 8). Transfection of full-length p130 Cas blocked the ability of AND-34 to promote Ral-GTP levels in cells. Importantly, overexpression of ⌬SBD-p130 Cas , which we previously demonstrated fails to bind AND-34, had no effect on AND-34-induced Ral GEF activity. The effect of p130 Cas was due to a change in the GEF activity of AND-34 since neither p130 Cas nor its AND-34-binding mutant had any effect on Ral GTP levels when AND-34 was not co-transfected (Fig. 8). Immunoblots show that p130 Cas had no detectable effect on the expression level of AND-34. These results suggest that p130 Cas is a negative regulator of AND-34 GEF activity. DISCUSSION In this work, we have identified AND-34 as a novel GDP exchange factor with activity in vivo on RalA, Rap1A, and R-Ras GTPases. The spectrum of GTPases activated by AND-34 is somewhat unusual. Previously identified Ral-GEFs have not been demonstrated to have multiple targets. Moreover, Rap and R-Ras are activated by C3G, and Ras and R-Ras are activated by Ras-GRF1, but no GEF identified to date has been able to activate all three (13)(14)(15). Our experiments suggest that AND-34 is most effective against RalA, although this conclusion is not firm since the conditions used involved overexpression of both exchange factor and GTPase. Whether the differences in exchange activity observed reflect properties intrinsic to AND-34 or the contribution of additional cellular components remains to be established. For example, the efficiency with which endogenous AND-34 interacts with these GTPases in cells could be affected dramatically by interacting proteins or by the subcellular localization of the endogenous GEF and GTPases. It is also likely that the AND-34 signaling specificity may vary with cell type and with the mechanism by which AND-34 is activated. All of these issues are clearly worthy of study in the future.
AND-34 is the murine homologue of the human protein, NSP2, one of three structurally related proteins termed NSP1-NSP3 by Lu and colleagues (4). SHEP1 (NSP3) was shown to bind to R-Ras and Rap1A but not RasH or RalA (16). However, its isolated CDC25 domain failed to promote nucleotide exchange on any GTPases in vitro. It is likely that AND-34 is one of a family of GEFs that are regulated by similar upstream signals but activate different sets of GTPases.
A further unusual feature of AND-34 is that, unlike most Ral-GEFs reported on to date, it does not have a detectable Ras-binding site, and its activity is not enhanced by co-expression with activated Ras. In this respect, AND-34 resembles Ral-GPS, a recently described Ral-GEF family that contains a pleckstrin homology domain but no Ras-binding site (17). Interestingly, we actually observed an inhibitory effect of Ras expression on AND-34. Presumably this is through an indirect mechanism mediated by some Ras effector system, possibly as part of a negative regulatory network.
AND-34 and other NSP family members have amino-terminal SH2 domains suggesting that they can be modulated by tyrosine kinase signaling. In fact, EGF and insulin promote the tyrosine phosphorylation of NSP1, whereas integrins promote the tyrosine phosphorylation of NSP1 and AND-34 (2,4). In addition, the SH2 domain of murine NSP3, SHEP1, binds to a phosphorylated tyrosine motif in the juxtamembrane region of the EphB2 receptor, and NSP-3 is heavily tyrosine-phosphorylated in cells expressing activated EphB2 (4,16). Whether such signals change the GEF activity of AND-34 remains to be determined. FIG. 4. Truncation of the AND-34 GEF domain eliminates association with p130 Cas . Lower panels, HA epitope-tagged ⌬SH2 (left), but not ⌬SH2 truncation mutants in which carboxyl-terminal AND-34 sequences were deleted following residue 610, 680 or 750 (center), binds to endogenous p130 Cas in HEK293 cells. The immunoreactive bands that migrate with faster mobility than p130 Cas are nonspecific. Myc epitope-tagged ⌬SH2 (right), but not a Myc-tagged construct containing only AND-34 GEF sequence binds to endogenous p130 Cas . Upper panels, Western blotting for the HA-epitope or Myc epitope of the AND-34 truncation mutants demonstrates comparable expression of immunoreactive proteins of the appropriate MW in the WCL utilized for the IP experiments. Another common feature of some members of this class of proteins is their ability to bind to the c-Src substrate and adaptor protein p130 Cas . When we set out to map the domains responsible for the interaction of AND-34 to p130 Cas , we suspected that it might be through the SH3 domain of p130 Cas , as AND-34 contains a polyproline-rich domain. This region in-cludes the conserved motif XPpϩP, a motif previously determined to mediate the association of PTP-PEST, PTP1B, focal adhesion kinase, and C3G with the p130 Cas SH3 domain (2). Alternatively, it was plausible that the amino-terminal SH2 domain of AND-34 bound to the p130 Cas substrate domain, a region containing nine YDXP sites that are close to the binding motif of the SH2 domain of v-Crk (3). To our surprise, the AND-34 SH2 domain and the proline-rich region were expendable for binding to p130 Cas , whereas the carboxyl terminus containing the GEF domain was not. In fact the nature of the binding between the two proteins may be rather complex, since binding required the entire GEF catalytic domain as well as additional amino-terminal sequence. Although our data do not rule out the possibility that AND-34 binds to p130 Cas through an intermediary protein, a recent report demonstrating that murine NSP-3 binds directly to the COOH terminus of p130 Cas suggests that this is unlikely (18). Along the same lines, our studies also show that AND-34 binds to the COOH terminus of p130 Cas , a region encompassing its Src-binding domain (SBD) and flanking COOH-terminal sequences but not its SH3 domain.
The fact that p130 Cas binds to a region encompassing the GEF domain of AND-34 motivated us to test whether its binding influences AND-34 GEF activity. When p130 Cas was coexpressed with AND-34, we did in fact detect a significant decrease in the ability of the exchange factor to activate Ral in cells. This effect was likely due to the binding between the two proteins since a p130 Cas mutant that did not bind to AND-34 under the same co-transfection conditions failed to suppress its GEF activity. This finding suggests that AND-34 is under negative regulatory control by p130 Cas binding, possibly by steric hindrance of GTPase binding to the catalytic domain. Alternatively, p130 Cas could target AND-34 to a location in the cell where it does not have access to its GTPase substrates. It should be noted, however, that in some reports, docking or Lower panels, expression of four HA-tagged p130 Cas SBD constructs was verified in whole cell lysates of 293 cells which were co-transfected with both full-length myc-tagged AND-34 and the indicated HA-p130 Cas SBD construct. Middle panels, Myc-tagged AND-34 was immunoprecipitated and immunoblots probed with anti-HA antibody to detect association with the p130 Cas SBD mutants. The upper band seen in all lanes corresponds to Ig heavy chain (IgH). The lower band seen on the left panel corresponds to Ig light chain (IgL). Immunoprecipitates from HA-SBD-Core-transfected cells was separated on 15% polyacrylamide gels while the remaining lysates were separated on 6% gels. Upper panels, HA-tagged p130 Cas mutants were immunoprecipitated and immunoblots probed with anti-Myc antibodies to detect association with AND-34.

FIG. 8. p130 Cas overexpression inhibits AND-34
Ral-GEF activity. Cos cells were transfected with GST-Ral with or without the addition of AND-34, p130 Cas and/or ⌬SBD (Cas-SBD), a p130 Cas mutant in which the Src-binding domain has been deleted. The percentage of GTP-bound Ral in cells was determined as described in Fig. 1. The lower two panels demonstrate expression levels of AND-34 (middle panel) or p130 Cas or ⌬SBD (bottom panel) in whole cell lysates derived from the transfections utilized in the GEF assays. AND-34 was detected using a polyclonal rabbit anti-AND-34 antisera and p130 Cas /⌬SBD using an anti-HA antibody (2). scaffolding proteins that have first been detected as inhibitors of their binding partners subsequently were shown to function in a positive fashion (19). Thus, p130 Cas could conceivably normally function to influence the specificity of AND-34 signaling rather than as a regulator of its catalytic activity.
AND-34 is the murine homologue of human BCAR3, a gene cloned on the basis of its ability to convert tamoxifen-sensitive breast cancer cell lines to tamoxifen resistance (5). This finding implies that signal transduction pathways altered by AND-34 overexpression mediate at least some aspects of this process. Many metastatic estrogen receptor-positive breast cancers are treatable with anti-estrogens such as tamoxifen, perhaps because cell proliferation or survival remains dependent upon estrogen. However, such tumors usually progress to a tamoxifen-resistant (estrogen-independent) state. Unfortunately, little is known about the molecular basis for the development of resistance to tamoxifen therapy. If the model system used to detect BCAR3 (AND-34) is reflective of the development of tamoxifen resistance in breast cancer patients then an understanding of the mechanism of AND-34 (BCAR3) action in this system may be quite important. The identifiable domains in AND-34 suggest at least four mechanisms of action could be involved, via the amino-terminal SH2 domain, the proline-rich region, the p130 Cas -binding site, and/or the GEF domain.
In this paper we show that the GEF domain can potentially activate RalA, Rap1A, and R-Ras GTPases. Thus, signaling pathways downstream of these GTPases are potential mediators. For example, Ral GTPases are known to interact with multiple signaling molecules including RalBP1, a GTPase-activating protein for Rac and CDC42 GTPases; filamin, an actinbinding protein; and phospholipase D (20 -24). Of particular interest is our recent discovery that Ral GTPases mediate EGF activation of c-Src (25). Elevated levels of EGF receptor family members and c-Src activity are commonly found in breast cancer samples (26). Furthermore, expression of EGF receptor is associated with a lack of response to hormonal therapy in patients with recurrent breast cancer (27). Rap GTPases can activate B-Raf and mediate ligand-induced cell adhesion, whereas R-Ras can activate phosphatidylinositol 3-kinase and influence inside-out integrin signaling and apoptosis (28 -31). R-Ras has also been shown to promote invasion of breast cancer cells (29). Which, if any, of these signaling pathways mediate AND-34 action are presently under investigation.
Remarkably, the same screen that detected BCAR3 (AND-34) as an inducer of tamoxifen resistance also recently found human p130 Cas (6). The fact that a random screen of genes detected two cDNAs that code for proteins that naturally bind to each other in cells argues strongly that the signaling pathways affected by these two proteins are particularly potent in promoting tamoxifen resistance, at least in breast cancer cell lines. The suggestion that p130 Cas expression may influence human breast cancer biology is supported by the finding that patients with breast tumors expressing a high level of p130 Cas appear to experience more rapid disease recurrence and are more resistant to tamoxifen therapy (7).
Our finding that overexpression of p130 Cas suppresses the GEF activity of AND-34 suggests that AND-34 functions through p130 Cas in the induction of tamoxifen resistance, rather than p130 Cas functioning through the GEF activity of AND-34. However, as described above, it is not clear that the natural function of p130 Cas is to suppress AND-34 GEF activity and so firm conclusions must await additional experiments. It is possible, for instance, that both p130 Cas and the GEF domain of AND-34 activate common signaling pathways when overexpressed. One obvious set of candidates are c-Src mediated events, because as we recently showed, Ral-GEFs can activate c-Src in cells and p130 Cas has been shown to participate in c-Src signaling (25,32).
In conclusion, we have identified a new GEF with novel signaling specificity. It can potentially activate multiple members of the Ras subfamily of GTPases, Ral, R-Ras and Rap. It also appears to be regulated by a well characterized adaptor protein, p130 Cas . Finally, understanding the interrelationships between AND-34 and p130 Cas may reveal important insights into human oncogenesis.