Interaction of the Adaptor Protein Shc and the Adhesion Molecule Cadherin*

In mitogenic signaling pathways, Shc participates in the growth factor activation of Ras by interacting with activated receptors and/or the Grb-2 (cid:122) Sos complex. Using several experimental approaches we demonstrate that Shc, through its SH2 domain, forms a complex with the cytoplasmic domain of cadherin, a transmembrane pro- tein involved in the Ca 2 (cid:49) -dependent regulation of cell-cell adhesion. This interaction is demonstrated in a yeast two-hybrid assay, by co-precipitation from mam-malian cells, and by direct biochemical analysis in vitro . The Shc-cadherin association is phosphotyrosine-dependent and is abrogated by addition of epidermal growth factor to A-431 cells maintained in Ca 2 (cid:49) -free medium, a condition that promotes changes in cell shape. Shc may therefore participate in the control of cell-cell adhesion as well as mitogenic signaling through Ras. Shc (1, 2) The SH2 domain Shc Ras


EXPERIMENTAL PROCEDURES
Materials-The antibodies used were rabbit IgG fractions to phosphotyrosine (Transduction Laboratories, horseradish peroxidase-coupled), to cadherin (pan-cadherin, ICOS Corp.), and to Shc for Western blotting (Transduction Laboratories). For the immunoprecipitation of Shc, antiserum to recombinant p52 Shc was produced. The coding sequence for the 52-kDa form of Shc was cloned into the pAcHLT-B baculovirus transfer vector (Pharmingen, Corp.) and transferred into sf9 insect cells. The His-tagged Shc protein was overexpressed in High 5 insect cells and purified via Ni 2ϩ affinity chromatography (Qiagen). 150 g of p52Shc from the 150 mM imidazole elution fraction was used as immunogen to subcutaneously inject a rabbit. Following three booster injections, immune serum was harvested. Sodium orthovanadate was purchased from Fisher, and hydrogen peroxide was from Sigma.
Yeast Two-hybrid Screen-The SH2 domain (residues 373-469) of Shc (1) was fused to the LexA DNA-binding domain and used as bait to screen a mouse 10-day embryo library fused to VP16 transcription activation domain. The yeast two-hybrid assay with this library was carried out essentially as described elsewhere (13,14) except that to tyrosine phosphorylate library proteins a constitutive active form of c-Src under the control of ADH1 promoter (15) was cloned into the NaeI site of the same bait plasmid containing the LexA-Shc-SH2 construct (pBTM116 Shc-SH2ϩSrc).
Immunoprecipitation-A-431 and NIH 3T3 cells were grown in Dulbecco's modified Eagle's medium with 10% calf serum at 37°C under 5% CO 2 . Upon reaching confluence and growth factor treatment (where indicated), cells were lysed in TGH buffer (1% Triton X-100, 10% glycerol, 50 mM Hepes, pH 7.2, and 100 mM NaCl) supplemented with 10 ng/ml leupeptin, 10 ng/ml aprotinin, 544 M iodacetamide, 1 mM phenylmethylsulfonyl fluoride, and 1 mM sodium vanadate. 5 l preimmune serum or serum containing p52Shc antibody were added to 500 g of cell lysate, and after incubation at 4°C for 2 h, immunocomplexes were collected by addition of protein A conjugated to Sepharose beads (Sigma). After washing with immunoprecipitation washing buffer (20 mM Hepes, pH 7.2, 100 mM NaCl, 10% glycerol, and 0.1% Triton X-100), bound proteins were eluted with SDS sample buffer, subjected to SDS-PAGE and transferred to nitrocellulose filters. The filters were blocked with TBSTB buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Tween 20, and 3% bovine serum albumin) and incubated with primary antibody in TBSTB buffer for 2 h at room temperature. The filters were washed with TBST buffer without 3% bovine serum albumin, followed by incubation with protein A-horseradish peroxidase (Zymed) in TBSTB buffer for 1 h. The filters were then washed with TBST buffer, incubated with ECL working solution (Amersham Corp.) for 1 min, and exposed to x-ray film.
In Vitro Tyrosine Phosphorylation of GST-N-Cadherin C Terminus-Clone S24 corresponding to the N-cadherin intracellular domain (residue 792-906) was fused in-frame to GST, expressed in Escherichia coli, and purified on glutathione-Sepharose 4B beads according to the manufacturer's manual (Pharmacia Biotech Inc., pGEX-5X-1 vector). Purified c-Src (Upstate Biotechnology Inc.) was used in kinase assays according to the manufacturer's instructions.
Gel Overlay Assay-The nitrocellulose filter was treated with 6 M guanidine hydrochloride to denature proteins at 4°C for 10 min in Hyb buffer (20 mM Hepes, pH 7.6, 75 mM KCl, 0.l mM EDTA, 2.5 mM MgCl 2 , 1 mM dithiothreitol, and 0.05% Nonidet P-40), and proteins were renatured at 4°C by five successive dilutions (40 min each) of the guanidine HCl to a final concentration of 0.185 M in the same buffer. Following two 30-min washes with Hyb buffer, the filter was blocked by 30-min incubations in 5 and 1% milk in Hyb buffer. The filter was incubated with purified p52Shc overnight at 4°C, washed with Hyb buffer, and blotted with anti-Shc (Transduction Laboratory).

RESULTS AND DISCUSSION
Two-hybrid Assay-To identify tyrosine phosphorylated molecules that recognize the SH2 domain of Shc, a modified yeast two-hybrid screen was performed in a system that included the Shc SH2 domain fused to the LexA DNA-binding domain, a mouse embryo library (13,14) fused to the VP16 transactivation domain, and a constitutively active form of the tyrosine kinase c-Src regulated by the ADH1 promoter (15), because SH2 interacting molecules are expected to contain phosphotyrosine. One positive clone, clone S24 (Fig. 1B), contained sequences that when translated correspond to residues 792-906 within the cytoplasmic domain of mouse N-cadherin (Fig. 1B). Cadherins are transmembrane proteins that regulate cell-cell adhesion in a Ca 2ϩ -dependent manner (16). The cytoplasmic domains of the three major cadherins are relatively conserved in sequence, particularly within the region corresponding to clone S24 (Fig. 1C). Clone S24 was subsequently retested in the two-hybrid assay in the presence and the absence of c-Src. No interaction of S24 with the Shc SH2 motif was detected in the absence of c-Src. Also, the central region of PLC-␥1, which contains two SH2 domains, did not, when substituted for the Shc SH2 domain, interact with clone S24 in the presence of c-Src. 2 These results indicate a putative recognition of cadherin by the SH2 domain of Shc.
Shc-Cadherin Association in Vivo-To determine whether the native Shc protein interacts with cadherin, co-immunoprecipitation assays were performed with A-431 and NIH 3T3 cells. The results shown in Fig. 2 demonstrate the specific co-precipitation of cadherin in Shc immunoprecipitates ob-tained from both cell types and in the absence of exogenous growth factor stimulation. Therefore, under typical cell culture conditions the association of Shc and cadherin is constitutive and likely dependent on the basal activity of tyrosine kinases.
The extracellular domain of cadherins binds Ca 2ϩ and mediates Ca 2ϩ -dependent cell-cell association. This recognition event involves the lateral dimerization of cadherin molecules (17)(18)(19) and the homophilic association of cadherin extracellular domains between adjacent cells (16). Cell-cell interaction then transmits biochemical signals through the cadherin cytoplasmic domain to effector molecules, such as the catenins, that bring about changes in actin cytoskeletal structure.
When placed in Ca 2ϩ -free medium, adherent and spread-out A-431 cells undergo a rapid morphological change to a round morphology following the addition of epidermal growth factor (EGF) (20). Given the Ca 2ϩ dependence of cadherin function in cell-cell association, we examined the state of Shc association with cadherin in the presence or the absence of extracellular Ca 2ϩ and EGF. The results presented in Fig. 3A demonstrate that the addition of EGF to A-431 cells in Ca 2ϩ -containing medium has no significant influence on cadherin co-precipitation with Shc (lanes 1 and 2). However, when the cells were placed in a Ca 2ϩ -free medium and EGF was added, a large decrease in cadherin association with Shc is detected (lanes 3 and 4). Because the incubation period for this experiment was 30 min and cell rounding occurs within this time, the observed loss of cadherin association with Shc could be a consequence of the change in cell shape that occurs when EGF is added to cells in the Ca 2ϩ -free medium. Therefore, under the same conditions A-431 cells were analyzed for Shc-cadherin association at much earlier times (1-30 min). As shown in Fig. 3B, cadherin association with Shc was significantly decreased 1 min (lane 2) after the addition of EGF to A-431 cells in this Ca 2ϩ -free medium. Hence, cadherin interaction with Shc is disrupted prior to observable changes in cell morphology. However, the biochemical mechanism underlying dissociation of the Shccadherin complex under these experimental conditions is not known.
In this assay system, prolonged incubation in Ca 2ϩ -free medium without EGF does decrease Shc-cadherin association to a moderate extent. However, the addition of EGF dramatically enhances the rapidity and extent of complex dissociation. The low level of cadherin that remains detectable in Shc immunoprecipitates obtained from cells treated with EGF in Ca 2ϩ -free medium (Fig. 3B, lanes 4 and 6), is due, at least in part, to cadherin present nonspecifically in precipitates obtained from B, depiction of cadherin structural organization and the region encompassed by clone S24 obtained as prey in the two-hybrid assay. C, comparison of the intracellular domain sequences of mouse N-, P-, and E-cadherin that correspond to the sequence of clone S24. The black dots denote the six tyrosine residues that may serve as phosphorylation sites. Cad, cadherin.

FIG. 2. Association of Shc and cadherin in intact cells.
Confluent A-431 or NIH 3T3 cells were lysed in a TGH buffer, and equal aliquots were used for precipitation (IP) with 5 l of either preimmune or immune serum obtained from a rabbit immunized with a baculovirus-expressed, purified, p52 isoform of Shc. Precipitates were collected with protein A-Sepharose and washed with buffer, and proteins were separated by SDS-PAGE. After transfer to nitrocellulose, standard procedures were employed for Western blotting (WB) with anti-pancadherin and detection of bound antibody with protein A-coupled horseradish peroxidase and ECL. The arrows indicate the position of cadherin.
The Src dependence of the two-hybrid assay results and the known properties of SH2 domains would predict that the SH2 domain of Shc recognizes phosphotyrosine within the cytoplasmic domain of cadherin. Although none of the six tyrosine residues in the cadherin cytoplasmic domain (Fig. 1C) have been identified as phosphorylation sites, Tyr 851 and Tyr 883 conform to the consensus recognition sequence of the Shc SH2 domain (pY with L/I/M at the ϩ3 position), as deduced from random peptide libraries (21) and the T cell receptor chain peptide used to solve the solution structure of a liganded Shc SH2 domain (22). Others have reported the presence of phosphotyrosine at low levels on cadherin (23)(24)(25), although cadherin is not generally cited as a tyrosine phosphorylated protein. Perhaps complicating phosphotyrosine detection is the reported association of phosphotyrosine phosphatases with cadherin (25)(26)(27). We have assayed A-431 cells for the presence of phosphotyrosine on cadherin (Fig. 4A). In the presence but not the absence of pervanadate, an inhibitor of phosphotyrosine phosphatases, tyrosine phosphorylation of cadherin is more readily detected (lane 2).
If pervanadate enhances cadherin tyrosine phosphorylation and Shc association with cadherin is phosphotyrosine-dependent, then it is expected that pervanadate will increase the level of Shc-cadherin complexes. The data in Fig. 4B show that pervanadate substantially increases the amount of cadherin present in Shc immunoprecipitates (lanes 1 and 2) and the level of Shc, particularly the p52 isoform, detectable in cadherin immunoprecipitates (lanes 3 and 4). These data, therefore, are consistent with the association of Shc and cadherin in a phosphotyrosine-dependent manner. Because EGF stimulates Shc tyrosine phosphorylation (1) but does not enhance Shc association with cadherin (Fig. 3B), the pervanadate influence on this association is likely due to the increased phospho-tyrosine on cadherin.
Shc Association with Tyrosine Phosphorylated Cadherin in Vitro-To determine whether Shc interacts directly with cadherin and to resolve the issue of whether cadherin must be tyrosine phosphorylated to affect this association, the in vitro experiments described in Fig. 5 were performed. The clone S24 sequence corresponding to residues 702-906 of N-cadherin was expressed as a GST fusion protein and purified by absorption on glutathione-Sepharose. As a control, GST was absorbed to the glutathione matrix. Aliquots of GST and GST-cadherin were then incubated with c-Src in the presence or the absence of ATP, eluted from the column with glutathione, separated by SDS-PAGE, and, following transfer to nitrocellulose, blotted with anti-phosphotyrosine (Fig. 5A) or anti-GST (Fig. 5B). The results show clearly that GST-cadherin is tyrosine phosphorylated in the presence of c-Src and ATP.
In a parallel experiment, following incubation with c-Src and/or ATP, GST and GST-cadherin were transferred to filters, denatured in 6 M guanidine HCl, and then gradually renatured. The filters were subsequently incubated with a baculovirus expressed, purified p52 form of Shc. 3 After washing, the filters were incubated with anti-Shc. As shown in Fig. 5C, Shc association with GST-cadherin was detected only in those samples where GST-cadherin had been previously incubated with c-Src and ATP. This demonstrates a direct interaction between Shc and the tyrosine phosphorylated cytoplasmic domain of cadherin. As shown in Fig. 5D, this direct interaction was also observed when isolated p52 Shc was added to Sepharose beads coupled to GST or GST-cadherin, which had been preincubated with c-Src and/or ATP. Following washing of the beads, elution 3 Y. Xu and G. Carpenter, unpublished results.

Shc-Cadherin Association 13465
with glutathione, and SDS-PAGE, Western blotting showed that Shc associated only with tyrosine phosphorylated GST-cadherin. Physiological requirements for cell proliferation, particularly within tissues, include the coordinated modulation of intracellular functions, such as nuclear transcription and cytoskeletal structure, with changes in the relationship of cells to their immediate extracellular environment, such as neighboring cells and the extracellular matrix. Cadherins represent a major molecular system by which cell-cell adhesion occurs. The results described in this manuscript demonstrate a Shc-cadherin association that is modulated by extracellular Ca 2ϩ and EGF. This raises the possibility that Shc, a tyrosine kinase substrate, may participate in the control of cadherin function in addition to its known role in the mitogenic activation of Ras and thereby nuclear signaling. Interaction of cells with the extracellular matrix is mediated by receptors termed integrins. Recently, Shc association with the tyrosine phosphorylated ␤ 4 integrin subunit has been reported (28). The ␤ 4 -Shc complex has been shown to be dissociated by EGF treatment of cells (29) and the loss of Shc association capacity by integrins results in aberrant cell cycle progression (30). Hence Shc may function to coordinate multiple alterations in cell physiology necessary for proliferation.
FIG. 5. Phosphotyrosine-dependent association of Shc and cadherin in vitro. Purified GST or GST-cadherin (GST-Cad; N-cadherin residues 792-906) were coupled to glutathione-Sepharose 4B and subjected to in vitro phosphorylation with c-Src in the presence or the absence of ATP as indicated. After washing with PBS, proteins were eluted with SDS sample buffer, divided into three equal aliquots, separated on different gels by SDS-PAGE, and transferred to separate nitrocellulose filters. One filter (A) was incubated with anti-phosphotyrosine and a second parallel filter (B) incubated with anti-GST (Santa Cruz). The third filter (C) from the above experiment was treated with 6 M guanidinium HCl to completely denature protein, and the proteins were renatured by gradual reduction of the guanidinium concentration. This filter was then incubated with purified p52 Shc, washed, and then incubated with anti-Shc. D, Sepharose-bound GST and GST-cadherin were phosphorylated with c-Src in the presence or the absence of ATP. After washing, p52 Shc was incubated with the beads for 2 h at 4°C. The beads were then washed to remove unbound Shc. Glutathione was used to elute proteins from the Sepharose, and the eluted proteins were separated by SDS-PAGE prior to Western blotting (WB) with anti-Shc.