Cell Adhesion Regulates the Interaction between the Docking Protein p130Cas and the 14-3-3 Proteins*

Integrin ligand binding induces a signaling complex formation via the direct association of the docking protein p130Cas (Cas) with diverse molecules. We report here that the 14-3-3ζ protein interacts with Cas in the yeast two-hybrid assay. We also found that the two proteins associate in mammalian cells and that this interaction takes place in a phosphoserine-dependent manner, because treatment of Cas with a serine phosphatase greatly reduced its ability to bind 14-3-3ζ. Furthermore, the Cas-14-3-3ζ interaction was found to be regulated by integrin-mediated cell adhesion. Thus, when cells are detached from the extracellular matrix, the binding of Cas to 14-3-3ζ is greatly diminished, whereas replating the cells onto fibronectin rapidly induces the association. Consistent with these results, we found that the subcellular localization of Cas and 14-3-3 is also regulated by integrin ligand binding and that the two proteins display a significant co-localization during cell attachment to the extracellular matrix. In conclusion, our results demonstrate that 14-3-3 proteins participate in integrin-activated signaling pathways through their interaction with Cas, which, in turn, may contribute to important biological responses regulated by cell adhesion to the extracellular matrix.

Integrin-mediated cell-extracellular matrix (ECM) 1 interactions affect many aspects of cell behavior, including cell proliferation, differentiation, and migration (1,2). Upon ligand binding, integrins recruit a number of signaling proteins to sites of cell contact with the ECM known as focal adhesions (3). As a result, several focal adhesion proteins are rapidly tyrosinephosphorylated and engaged in protein-protein interactions with signaling proteins containing Src homology 2 (SH2) domains. One of the tyrosine-phosphorylated proteins that orchestrates the assembly of these signaling complexes is p130 Cas (Cas), a docking protein originally identified as the major tyrosine-phosphorylated substrate in v-Crk and v-Src transformed cells (4 -7). Cas has a unique structure composed of several domains capable of mediating protein-protein interac-tions (see Fig. 1). An amino-terminal SH3 domain has been shown to associate with proline-rich regions of the focal adhesion kinase (FAK) (8), and of the protein tyrosine phosphatases PTP1B (9) and PTP-PEST (10). Furthermore, the binding of Src family kinases toward the carboxyl terminus of Cas (11) results in the tyrosine phosphorylation of this docking protein in a region known as the substrate domain (SD) (see Fig. 1) (12)(13)(14). Nine YDV/TP motifs that conform to the binding motif for the Crk SH2 domain appear in this region and, indeed, Cas-Crk interaction is known to take place upon integrin-mediated cell adhesion (12,13,15). Cas also contains a serine-rich domain (designated here as SR) (see Fig. 1); however, despite the fact that Cas is known to undergo serine phosphorylation in response to integrin-mediated cell adhesion (Ref. 16, and data not shown), the function of this region or the significance of serine phosphorylation of Cas is unknown.
To get insight into the function of the SR region in Cas, we set forth to identify the proteins that directly bind Cas within this domain. We report here that among these molecules are the 14-3-3 proteins. Members of the 14-3-3 family form homoor heterodimeric complexes that mediate interactions between diverse components of signaling pathways, including the serine/threonine kinase c-Raf-1 (17)(18)(19)(20), the tyrosine phosphatase Cdc25 (21), the phosphatidylinositol 3Ј-kinase (22), the protooncogene product Cbl (23), the tumor suppressor gene p53 (24), and the Bcl-2 family member Bad (25) (for a review, see Refs. 26 and 27). In many cases, serine phosphorylation of the binding partner is crucial in regulating the interaction with 14-3-3 (28), and by using phosphoserine-oriented libraries, two different binding consensus motifs for 14-3-3 were identified (29). Through the various interactions, the 14-3-3 proteins have been proposed to regulate enzymatic activity of their binding partners, to serve as clustering proteins bringing together enzymes and substrates, and to function as "chaperone" molecules that stabilize structural conformations of the associated proteins (26). Our results reported here suggest a role for 14-3-3 proteins in integrin signaling pathways and provide new insights into their functional significance in intracellular signaling events.

EXPERIMENTAL PROCEDURES
DNA Constructs and Mutagenesis-The yeast two-hybrid cloning vectors pVP16 and pBTM116, the DNA construct pBTM116/lamin, and the yeast strain L40 have been described in (30). The "bait" construct used for the two-hybrid screening, pBTM116/Cas(SR), contains the SR domain of Cas encompassing the amino acids 520 to 712 ( Fig. 1) (numbering according to Sakai et al. (7)), and was generated by cloning the corresponding cDNA fragment in frame with the LexA coding sequence into the vector pBTM116. The DNA constructs encoding glutathione S-transferase (GST) fusion proteins of the 14-3-3 isoforms , ␤, , and have been described by Yaffe et al. (29). The mammalian expression plasmids for the GST-tagged Cas constructs GST/Cas (coding for the full-length wild-type Cas) and GST/Cas⌬(SD) (coding for a form of Cas in which the substrate domain, amino acids 213-514, has been deleted) are described by Mayer et al. (31) (see Fig. 1). The 14-3-3/Myc cDNA contains the c-Myc epitope at the amino-terminal domain of the 14-3-3 * This work was supported by National Institutes of Health Grant CA71560 (to K. V.). 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  protein and was cloned into the mammalian expression vector pcDNA3 (Invitrogen). Single point mutations and DNA deletions were performed using polymerase chain reaction techniques and the QuickChange sitedirected mutagenesis kit (Stratagene). The Cas⌬(SR) construct was prepared by deleting the DNA sequences encoding for the residues 520 to 712 of the full-length Cas cDNA (Fig. 1), followed by cloning into the mammalian expression vector pSSR␣.
Yeast Two-hybrid Analysis-The yeast two-hybrid interaction screen was performed essentially as described by Vojtek et al. (30). The plasmid pBTM116/Cas(SR) was introduced into the L40 yeast strain, total protein extracts were prepared from individual colonies (32), and the expression of LexA-Cas(SR) was assessed by immunoblotting (anti-LexA antibody) (Santa Cruz Biotechnology). A L40 yeast clone expressing LexA-Cas(SR) was transformed with a murine whole embryo library (embryonic day 9.5-10.5) constructed in the plasmid pVP16 (33). The resulted transformants (2 ϫ 10 6 ) were selected using histidine prototrophy and analyzed for ␤-galactosidase activity. Positive yeast clones were selected and expanded in selective medium for the expression of the gene Leu2 (pVP16-vector). The specificity of the protein-protein interactions was assessed by transforming the selected yeast clones with the plasmids pBTM116/Lamin or pBTM116/Cas(SR). Plasmid DNA from positive clones was isolated, sequenced, and compared with the nucleotide data base at the National Library of Medicine (www3.ncbi.nlm.nih.gov/). The yeast two-hybrid protein-protein interactions were also evaluated by quantifying the ␤-galactosidase activity in liquid cultures using o-nitrophenyl-␤-galactopyranoside (Sigma) (34).
Cell Culture, Adhesion, and Transfection-Mammalian cells used in the study were grown in Dulbecco's modified Eagle's medium containing glutamine Pen-Strep (Irvine Scientific) and 10% fetal calf serum (Tissue Culture Biologicals). For cell adhesion experiments, Rat-1 cells were grown to 90% confluency as a monolayer, serum-starved for 18 h, and detached with trypsin-EDTA treatment followed by treatment with 0.5 mg/ml of trypsin inhibitor. Cell suspensions were incubated in 0.5% bovine serum albumin, Dulbecco's modified Eagle's medium at 37°C with continuous rotation for the indicated time periods and plated onto dishes precoated with 10 g/ml of fibronectin (FN) or with 2 mg/ml of poly-L-lysine (2 mg/ml). Cell transfections were performed using the LipofectAMINE reagent (Life Technologies, Inc.) following the manufacturer's protocol.
Preparation of Cell Lysates, Immunoprecipitations, and Immunoblotting-Cell cultures were rinsed with ice-cold phosphate-buffered saline (PBS) and lysed on ice for 5 min (50 mM Hepes, pH 7.9, 150 mM NaCl, 1 mM EGTA, 10% glycerol, 1% Triton X-100, 1.5 mM MgCl 2 , 200 M Na 3 VO 4 , 50 mM NaF, 0.1 units/ml aprotinin, 10 g/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride). Cell lysates were cleared by centrifugation for 15 min at 4°C, normalized for the protein content and subjected to immunoprecipitation or affinity association experiments (described later). Immunoprecipitations were carried out with a commercial anti-Cas antibody (C-20, Santa Cruz Biotechnology) or with a rabbit polyclonal antibody generated in the laboratory against the SR domain of Cas and protein A-Sepharose, followed by SDS-polyacrylamide gel electrophoresis (PAGE). Immunoblotting was performed with an anti-Cas antibody (Transduction Laboratories), anti-14-3-3␤ antibody (Santa Cruz), or anti-c-Myc antibody (Calbiochem). Immunoreactive bands were detected with horseradish peroxidase-conjugated antimouse IgG or protein A and enhanced chemiluminescence (SuperSignal Chemiluminescent Substrate) (Pierce).
Affinity Association Using GST Fusion Proteins or Glutathione-Sepharose-In vitro association experiments were carried out with GST fusion proteins containing the full-length 14-3-3 molecule. The fusion proteins were expressed in Escherichia coli and purified as described (13). Cell lysates prepared as described above were normalized for the protein content and incubated with ϳ10 g of GST alone or GST fusion protein, which had been immobilized on glutathione-Sepharose beads. Protein complexes were allowed to associate for 4 h at 4°C, recovered by centrifugation, and washed with lysis buffer. Complexes were analyzed by standard SDS-PAGE and immunoblotting techniques. Affinity purification of GST-tagged Cas proteins from mammalian cell lysates with glutathione-Sepharose 4B (Amersham Pharmacia Biotech) was carried out by using a similar procedure, except that the lysis buffer contained 400 mM NaCl, and the washing buffer contained 500 mM NaCl.
Alkaline Phosphatase Treatment-Cas was immunoprecipitated from Rat-1 cells using an anti-Cas polyclonal antibody, protein complexes were recovered by centrifugation, equilibrated with 50 l of phosphatase reaction buffer (50 mM Hepes, pH 7.9, 2 mM MgCl 2 , 140 mM NaCl, 0.5 mM dithiothreitol) and treated with 30 units of calf alkaline phosphatase (New England Biolabs) for 30 min at 30°C. The protein complexes were then eluted with 1% SDS (5 min at 95°C), diluted 1:10 with lysis buffer, incubated with GST/14-3-3 fusion proteins immobilized on Sepharose and analyzed with GST/14-3-3 fusion proteins as above. Parallel control experiments omitted calf alkaline phosphatase.
Confocal Microscopy-Rat-1 cells were fixed with 4% paraformaldehyde and permeabilized with 0.5% Triton X-100 in PBS for 5 min at room temperature. The cells were incubated for 1 h at room temperature with commercial antibodies against 14-3-3␤ (Santa Cruz), vinculin (Bio-Rad) and Cas (Transduction Laboratories). Anti-14-3-3 and antivinculin antibodies were diluted in 5% fetal calf serum, PBS, and applied to cells that had been incubated with 5% fetal calf serum, PBS for 30 min at room temperature. The anti-Cas antibody was diluted in 0.5% bovine serum albumin, PBS and used without preblocking the cells. The immunocomplexes were detected using fluorescein isothiocyanate-or tetramethyl rhodamine-conjugated anti-rabbit/anti-mouse IgG (Jackson ImmunoResearch Laboratories, Inc.), FluoroGuard Antifade mounting reagent (Bio-Rad), and visualized by digital confocal immunofluorescence (Zeiss LSM-410 confocal microscope). Images were captured with a Zeiss Axiovert 405 M inverted microscope with an attached CCD camera (100x Plan-Neofluare).

The Serine-rich Domain of Cas Interacts with the 14-3-3 Protein in Yeast-
The amino acids 520 -712 of the docking protein Cas encompass a region that is characterized by the presence of multiple serine phosphorylation consensus motifs, the SR domain ( Fig. 1). Importantly, this region is significantly conserved in Cas-homologous proteins, including Efs/Sin (35,36) and HEF1/Cas-L (37,38). This information prompted us to investigate the possibility that the S domain of Cas might participate in intracellular signaling by mediating protein-protein interactions. The cDNA coding for the SR domain of Cas, LexA-Cas(SR), was used as a bait, pBTM116-LexA-Cas(SR) for two-hybrid screen of a mouse embryonic library (2 ϫ 10 6 transformants) subcloned into pVP16. Six of the total of 67 selected clones were found to interact in a specific manner with the LexA-Cas(SR) bait. Plasmid DNA from these clones was isolated and sequenced. DNA data base analysis demonstrated that clone number 11 contained a fragment corresponding to the amino acids 1 to 138 of the mouse 14-3-3 protein fused in frame with the VP16 protein. This clone was renamed as pVP16/14-3-3 (11). As shown in Fig. 2, yeast transformed with the pBTM116/Cas(SR) and pVP16/14-3-3(11) plasmids expressed significant ␤-galactosidase activity and grew in conditional media lacking leucine, tryptophan, and histidine. The clone pVP16/14-3-3(11) contains DNA sequences (ϳ60 base pairs) that correspond to the 5Ј-untranslated region of the mouse 14-3-3 cDNA that were deleted to generate the construct pVP16/14-3-3(DUT). Expression of this protein together with the SR domain of Cas, pBTM116/Cas(SR), resulted in a similar transactivation as compared with that obtained with , and Src binding site (SBS, this region contains a proline-rich sequence that associates with the SH3 domain, and a tyrosine residue, which upon phosphorylation, binds to the SH2 domain). The yeast two-hybrid screen bait construct coding for the SR domain of Cas and containing amino acids 520 -712 is shown (pBTM116/Cas(SR)). The Cas⌬(SD) construct contains a deletion of the entire substrate domain of Cas (amino acids 213-514; the deletion is indicated as Cas⌬(SD)). The Cas⌬(SR) construct lacks the SR domain of Cas (amino acids 520 -712; the corresponding deletion is indicated as Cas⌬(SR)). the original clone isolated by two-hybrid screen, pVP16/14-3-3(11) (Fig. 2), confirming a direct interaction between the 14-3-3 peptide and Cas(SR) domain. The specificity of the interaction between Cas(SR) and 14-3-3 in yeast was further demonstrated by the lack of interaction of pVP16/14-3-3 with other domains of Cas or lamin (Fig. 2).
Full-length Cas Interacts with 14-3-3 Proteins in Vitro and in Vivo in Mammalian Cells-To examine whether 14-3-3 interacts with the full-length Cas protein in mammalian cells, a GST fusion protein coding for 14-3-3 was used in affinity association, or pull-down, experiments. Analysis of lysates prepared from COS-1, HeLa, and 293 human embryonic kidney cell lines demonstrated that Cas interacts in vitro with 14-3-3 (Fig. 3A). The association was more prominent in HeLa and 293 cell lysates, suggesting that the interaction might be regulated by cellular processes, such as post-translational protein modifications. Pull-down studies using GST fusion proteins of various 14-3-3 isoforms showed that Cas efficiently associates with the 14-3-3␤ and isoforms as well (Fig. 3B, left panel). Inter-estingly, the interaction observed between Cas and 14-3-3 was consistently lower than that between the other isoforms, despite the fact that equal amounts of the different GST fusion proteins were used in the experiments (Fig. 3B, right panel).
We next investigated whether 14-3-3 proteins interact with Cas in vivo. To this end, GST-tagged forms of Cas, GST/Cas, or GST/Cas⌬(SD), or an untagged form of wild-type Cas were expressed alone or in combination with 14-3-3-Myc in HeLa (Fig. 4A) or COS-1 cells (Fig. 4B). Protein lysates were prepared 48 h after transfection and the GST-tagged Cas proteins were precipitated with glutathione-Sepharose (Fig. 4, A and B). The protein complexes were analyzed by immunoblotting with anti-14-3-3␤ antibodies that cross-react with all 14-3-3 isoforms (Fig. 4A), or with an anti-Myc antibody that recognizes the expressed Myc-tagged 14-3-3 protein (Fig. 4B). As shown in Fig. 4, the recombinant 14-3-3-Myc protein migrates with an apparent molecular mass of ϳ35 kDa, which is slightly larger than that of the endogenous 14-3-3 proteins (ϳ32-33 kDa). Both endogenous and recombinant 14-3-3 proteins were found to associate with the wild-type GST/Cas protein in vivo. A mutant form of Cas in which the substrate domain has been deleted, GST/Cas⌬(SD), associated with the 14-3-3 proteins to a similar extent (Fig. 4A). Because this mutant form does not become phosphorylated on tyrosine at a detectable level (39), our results suggest that the Cas-14-3-3 interaction is not dependent on or regulated by the tyrosine phosphorylation levels of Cas. To rule out the possibility that the association of 14-3-3 with the Cas constructs was mediated by binding to the GST tag and not to Cas, we performed co-immunoprecipitation experiments with anti-Cas antibodies. Endogenous 14-3-3 proteins readily co-immunoprecipitated with the endogenous Cas proteins from Rat-1 cell lysates thereby confirming that 14-3-3 is a component of the Cas protein complex in vivo (Fig. 4C).

FIG. 2. The serine-rich domain of Cas interacts with 14-3-3 in yeast.
A, liquid assays for ␤-galactosidase activity. The L40 yeast strain was cotransformed with LexA DNA-binding domain fusion plasmids expressing either the serine-rich domain of Cas (pBTM116/ Cas(SR)), the Cas proline-rich sequence P1 (pBTM116/Cas(P1)), or lamin (pBTM116/Lamin), together with either (i) the plasmid pVP16/ 14-3-3(11) containing the activation domain of VP16 (pVP16) fused with the first 138 amino acids of the mouse 14-3-3 protein, (ii) the plasmid pVP16/14-3-3(DUT), in which the sequences connecting the VP16 peptide to the 14-3-3 protein were deleted, or (iii) the plasmid pVP16/ Cas(SH3) expressing the SH3 domain of Cas. Three separated colonies per sample were analyzed, and the results are shown as the average units with the error bars denoting the standard deviation. B, colony-lift ␤-galactosidase assay of yeast growth in restrictive medium lacking tryptophan, leucine, and histidine. The numbers correspond to the L40 yeast transformants shown in A.

FIG. 3. Full-length Cas from different cell lines associates in vitro with 14-3-3 proteins.
A, protein lysates prepared from the indicated cell lines were incubated in vitro with Sepharose-immobilized GST/14-3-3, or GST alone as a negative control. The associated protein complexes and samples of the total protein preparations (total cell lysate) were analyzed by SDS-PAGE and immunoblotting using an anti-Cas antibody. B, protein lysates from HeLa cells were subjected to affinity association analysis using various isoforms (␤, , , ) of Sepharose-immobilized GST/14-3-3 proteins and analyzed as indicated in A (left panel). A Coomassie-stained SDS-PAGE gel analysis of the purified GST proteins is shown in the right panel.
teraction-A charge-reversal mutation of lysine-49 in the peptide binding groove within the 14-3-3 molecule is thought to disrupt the binding of 14-3-3 to the serine-phosphorylated target proteins (see "Discussion") and indeed, this mutation has been shown to greatly decrease the association of 14-3-3 with Raf-1 and exoenzyme S (40). We found that the association between Cas and 14-3-3 is also critically dependent on interactions mediated by this lysine residue. Parallel pull-down experiments using the wild-type GST/14-3-3 or the GST/14-3-3 K49E mutant demonstrated that Cas associates exclusively with the wild-type protein, despite the fact that similar amounts of bacterially expressed fusion proteins were used in the studies (Fig. 5A). These data prompted us to further inves-tigate the requirement of serine phosphorylation in Cas for the association with 14-3-3. To this end, Cas was immunoprecipitated from Rat-1 cell lysates, treated with calf intestinal alkaline phosphatase and subjected to an in vitro association analysis with GST/14-3-3 and control GST fusion proteins as described under "Experimental Procedures." Fig. 5B shows that phosphatase treatment of Cas completely abolished the interaction with the GST/14-3-3. The decrease in association was not because of proteolysis of Cas, as equivalent amounts of Cas protein were present in all the samples (data not shown). Therefore, we concluded that phosphoserine residue(s) of Cas

FIG. 4. Association of 14-3-3 proteins with Cas in vivo.
A, HeLa cells were transiently transfected with mammalian expression vectors containing the cDNAs for 14-3-3/Myc, for the full-length, untagged Cas (Cas), or for two GST-tagged forms of Cas: GST/Cas⌬(SD), in which the substrate domain of Cas had been deleted, and GST/Cas, in which the GST tag was fused with the full-length Cas cDNA. Total protein lysates were prepared and incubated with glutathione-Sepharose 4B (see "Experimental Procedures"). The protein complexes were resolved on SDS-PAGE gels and immunoblotted using an anti-14-3-3␤ antibody that recognizes all known 14-3-3 isoforms. B, COS-1 cells were transfected with the indicated cDNAs and assayed as in A but immunoblotting was carried out with an anti-c-Myc antibody that recognizes the exogenously expressed 14-3-3/Myc protein. C, endogenous Cas immunocomplexes were prepared from Rat-1 cells and analyzed by immunoblotting with an anti-14-3-3 antibody as in A. The band corresponding to endogenous 14-3-3 proteins is indicated in the figure. A protein band of ϳ60 -65 kDa was observed also in control experiments using irrelevant antibodies and therefore was determined to be nonspecific.
FIG. 5. 14-3-3 proteins and Cas associate in a phosphoserinedependent manner. A, the single point mutation 14-3-3 K49E disrupts the interaction in vitro with Cas. Protein lysates from 293 cells were incubated with Sepharose-immobilized GST/14-3-3 wild type (WT) or GST/14-3-3 K49E. The associated complexes and total cell lysates were analyzed by SDS-PAGE and immunoblotting with an anti-Cas antibody. The Coomassie-stained SDS-PAGE gel analysis of the purified GST fusion proteins is shown on the right panel. B, dephosphorylation of Cas disrupts its interaction with 14-3-3 proteins. Cas immunoprecipitated from Rat-1 cell lysates were treated (ϩ) or not (Ϫ) with calf alkaline phosphatase as described under "Experimental Procedures." The samples were then subjected to affinity association experiments using wild-type or K49E mutant GST/14-3-3 proteins, and the complexes were analyzed as in A. C, 293 cells were transiently transfected with the cDNA expressing Cas⌬(SD) or Cas⌬(SR) and protein lysates were analyzed as described in A. Note that GST/14-3-3 associates to a significantly higher extent with endogenous Cas and the mutant Cas⌬(SD) rather than Cas⌬(SR). The apparent SDS-PAGE molecular masses of Cas⌬(SR) and Cas⌬(SD) proteins are ϳ110 kDa and ϳ85 kDa, respectively. Immunoblot analysis of total cell lysate protein samples shows that the expression levels of the endogenous or recombinant proteins are comparable.
are an important determinant for the binding to 14-3-3. These results together with the yeast two-hybrid interaction of the serine-rich domain of Cas with 14-3-3 directed us to investigate whether the SR domain in Cas is the major binding site for 14-3-3 proteins in the context of full-length Cas. Therefore, we generated a Cas mutant with the entire SR domain deleted, Cas⌬(SR). Pull-down experiments showed that deletion of this domain of Cas greatly diminished, but did not completely eliminate, the interaction with GST/14-3-3. Despite the fact that the protein expression levels of the recombinant proteins and the endogenous Cas were comparable (Fig. 5C), GST/14-3-3 readily interacted with the endogenous full-length Cas and with the mutant Cas⌬(SD), but to a lesser degree with the recombinant Cas⌬(SR) protein (Fig. 5C). These results demonstrate that the SR domain of Cas is a major structural determinant mediating the association of Cas to 14-3-3 proteins and that additional, yet-to-be identified 14-3-3 binding domains appear to be present in Cas outside of this region.
Integrin-mediated Cell Adhesion Induces Cas-14-3-3 Interaction-Cas is known to become not only tyrosine-, but also serine-phosphorylated in response to integrin-mediated cell adhesion (16). We therefore investigated the possibility that the Cas-14-3-3 interaction would be regulated by cell adhesion to extracellular matrix proteins. Pull-down experiments demonstrated that Cas and 14-3-3 readily interacted in Rat-1 cells that had been cultured as a monolayer in 10% fetal calf serum and Dulbecco's modified Eagle's medium or serum-starved for 18 h (Fig. 6). Upon cell detachment, the Cas-14-3-3 interaction was remarkably reduced in a time-dependent manner. When the Rat-1 fibroblasts were replated onto FN substratum, to which the cells adhere in an ␣ 5 ␤ 1 integrin-dependent manner, the Cas-14-3-3 interaction was restored. A gradual increase in the association was observed as early as 20 min after replating of the cells, with a complete recovery to steady-state levels within 4 h after replating (Fig. 6A). In contrast, cell adhesion to poly-L-lysine, to which cells adhere in an integrin-independent manner, failed to stimulate Cas-14-3-3 interaction (Fig. 6B). Together, these results demonstrate the requirement for integrin ligand binding in the signaling pathway that induces the Cas-14-3-3 association.
Co-localization of Cas and 14-3-3 Proteins-Previous studies have demonstrated that Cas localizes in membrane ruffles (39) and focal adhesions (41) during the initial attachment of cells to matrix proteins, whereas, in stationary cells, the majority of Cas protein localizes in the cytosol (42). We examined the subcellular distribution of Cas and 14-3-3 proteins using digital confocal immunofluorescence (Fig. 7). Rat-1 cells were kept in suspension for 40 min followed by replating on FN for 20 min or 4 h. During the initial steps of cell adhesion (20 min), lamellipodia extensions were observed in the spreading cells. Simultaneous two-color confocal immunofluorescence showed that 14-3-3 proteins (Fig. 7A) accumulate at the edges of the membrane ruffles where Cas (Fig. 7D) was highly concentrated. Upon merging the images, a remarkable co-localization of 14-3-3 with Cas (Fig. 7G) was observed in the lamellipodia structures (yellow). After 4 h of attachment to FN, lamellipodia structures had disappeared, and focal adhesion contacts were well organized as defined by immunostaining with anti-vinculin antibodies, a marker for focal adhesions (43) (Fig. 7E). At this time point, the 14-3-3 proteins were localized throughout the cytosol; however, co-localization analysis with vinculin also demonstrated the presence of 14-3-3 proteins within some focal adhesion structures (Fig. 7H). Cytosolic 14-3-3 proteins showed a prominent punctuate distribution (Fig. 7, B and C) that was very similar to the one observed for Cas (Fig. 7F). Upon superimposing the two images, significant co-localization of Cas and 14-3-3 was detected in the cytosol (Fig. 7I). These results further support the notion that the two proteins interact in vivo and that this association may have a functional significance. DISCUSSION The docking protein Cas was originally identified as a target for tyrosine kinase activity in various transformed cells, and consequently the tyrosine phosphorylation levels of Cas were found to correlate well with the transforming activity of various oncogenes (6, 7, 44 -47). Significantly, recent results obtained with CasϪ/Ϫ fibroblasts demonstrate that Cas is essential for Src-induced morphological transformation and anchorage independence, suggesting a causal role for Cas in oncogene-induced malignant transformation (48). In nontransformed cells, the tyrosine phosphorylation levels of Cas are regulated by extracellular stimuli such as integrin-mediated cell adhesion (12, 13, 15, 49 -51). As a result, Cas associates with the SH2 domain containing adapter protein Crk. This interaction has been shown to lead to the activation of signaling pathways controlling cellular functions such as actin organization (52), cell migration (39), and c-Jun N-terminal kinase activation (63). The primary structure of Cas suggests that it may participate in additional signal transduction pathways through mul- Anti-Cas immunoblotting of the total cell lysate preparations was used to confirm that equal amounts of protein were used (lower panel). B, the experiment was performed following conditions similar to A, but the cells were allowed to adhere either onto poly-Llysine (PLL) or fibronectin (FN) for 50 min. tiple differentially regulated protein-protein interactions. Among the putative signaling domains in Cas is a SR region, which is significantly conserved (56% similar) between Cas and a close family member, HEF1/Cas-L (37,38). The SR region of Cas was found here to interact with the 14-3-3 protein in a yeast two-hybrid system. Importantly, this interaction was not limited to yeast as our results indicate that the full-length Cas interacts both in vitro and in vivo in mammalian cells with various isoforms of the 14-3-3 family.
Although much of the attention has been focused on the role of the tyrosine-phosphorylated form of Cas in intracellular signaling events, our data demonstrate that tyrosine phosphorylation levels of Cas do not correlate with the ability of Cas to bind 14-3-3. Thus, a mutant construct in which the major tyrosine-phosphorylated region (substrate domain) of Cas was deleted readily interacted with the 14-3-3 proteins in vivo. Moreover, we observed no correlation between the Cas-14-3-3 interaction and the tyrosine phosphorylation levels of Cas in transformed cells. Although Cas is heavily tyrosine-phosphorylated in both v-Crk and v-Src transformed cells (5, 7), we only observed Cas-14-3-3 association in cell lysates prepared from v-Crk, and not from v-Src-expressing cells (data not shown). Significantly, our results firmly support a model in which serine phosphorylation of Cas determines its binding to the 14-3-3 proteins. First, the charge-reversal mutation K49E in 14-3-3 completely abolished its ability to associate with Cas. Crystallographic analysis has demonstrated that this lysine residue is lining the amphipathic peptide-binding groove of the 14-3-3 proteins (53), which mediate electrostatic interactions with phosphoserine residues in the 14-3-3 target proteins (40,54). Second, the Cas-14-3-3 interaction was highly reduced when Cas immunoprecipitates were treated with alkaline serine phosphatase before affinity association experiments; a similar experimental approach has been used to demonstrate the serine-phosphorylation requirement for association between the docking protein c-Cbl and 14-3-3 (55). In addition, our results further demonstrate a role for cell adhesion in the regulation of 14-3-3 function, as we found that the Cas-14-3-3 interaction is highly enhanced upon integrin-mediated cell adhesion. It is also interesting to note that several focal adhesion proteins, including Cas (16), FAK (56), and paxillin (57,58), have been reported to undergo serine phosphorylation in response to integrin-mediated cell adhesion. Taken together, we propose a model in which integrin ligand binding connects to the multifunctional 14-3-3 proteins by regulating the serine phosphorylation levels of Cas. This interaction in turn may have an important functional role in integrin signaling pathways.
Our yeast two-hybrid results together with the in vitro interaction studies in which very stringent conditions were used to disrupt pre-existing Cas-protein complexes (1% SDS elution and heat denaturation) suggest that the interaction between Cas and 14-3-3 is likely to be direct and mediated by the SR domain of Cas. Moreover, deletion of this domain (mutant Cas⌬(SR)) did greatly reduce the ability to bind 14-3-3, further confirming the importance of structural 14-3-3 binding determinants located within the SR domain of Cas. Nevertheless, the binding of the Cas⌬(SR) mutant to 14-3-3 was still detectable. Detailed analysis of the primary structure of Cas revealed the presence of several sequences that conform to classical 14-3-3-consensus motifs (29). Only one of these sequences, RPLPSPP (amino acids 733-739), which overlaps with the Src-SH3 binding site, is located outside of the SR domain. Further mutation (S737Q mutation) of this motif in the context of the Cas⌬(SR) construct did not affect the interaction with GST/14-  A, B, and C), vinculin, used as a marker for focal adhesion contacts (panel E), and Cas (panels D and F). Rat-1 cells were detached from ECM and subsequently allowed to adhere to fibronectin-coated glass surfaces for different periods of time: 20 min (panels A and D) or 4 h (panels B, C, E, and F). The immunofluorescence analysis was performed as indicated under "Experimental Procedures." The lower panels (G, H, and I) depict the overlay of the corresponding upper images to show the co-localization of 14-3-3 and vinculin or Cas (yellow color). The images were obtained with a 100ϫ magnification. Bar ϭ 10 M.
3-3 (data not shown). These results indicate the presence of additional 14-3-3-recognition sites in Cas that do not conform to the canonical recognition sequences. The presence of active 14-3-3 binding sites, distinct from the established consensus motifs, has been described in other proteins, such as the docking molecule c-Cbl (55) and the intermediate filament protein keratin 18 (59).
During the initial steps of cell adhesion to ECM, the cells extend a leading edge that adheres to the substratum through integrin receptors. This process of protrusion of the lamellipodium involves actin polymerization (60) and the accumulation of components of focal adhesion structures at the edge of the membrane (61). Recent observations made with CasϪ/Ϫ fibroblasts suggest that Cas is a key molecule in mediating actin polymerization events (48). It has also been reported that Cas is markedly concentrated in membrane ruffles (39) and has an important role during cell migration (62) via interaction with signaling molecules such as c-Crk (39). Our subcellular localization studies demonstrate that 14-3-3 proteins also accumulate at the leading edge of the lamellipodia and co-localize with Cas in these structures. Once the cells stabilize the focal adhesion structures and become stationary, Cas and 14-3-3 redistribute through the cytosol displaying a punctuated pattern and extensive co-localization. Thus, the similar dynamic subcellular distribution shared by 14-3-3 and Cas suggests that these two proteins may cooperate in cellular functions, such as the orchestration and stabilization of cellular contact sites with the ECM. Further studies with e.g. dominant-interfering mutants of Cas or 14-3-3 will be required to fully explore the signaling roles that the Cas-14-3-3 complex may have in integrin signaling pathways.