The c-Src Tyrosine Kinase Regulates Signaling of the Human DF3/MUC1 Carcinoma-associated Antigen with GSK3 b and b -Catenin*

The DF3/MUC1 mucin-like glycoprotein is aberrantly overexpressed in most human carcinomas. The cytoplasmic domain of MUC1 interacts with glycogen synthase kinase 3 b (GSK3 b ) and thereby decreases binding of MUC1 and b -catenin. The present studies demonstrate that MUC1 associates with the c-Src tyrosine kinase. c-Src phosphorylates the MUC1 cytoplasmic domain at a YEKV motif located between sites involved in interactions with GSK3 b and b -catenin. The results demonstrate that the c-Src SH2 domain binds directly to pY-EKV and inhibits the interaction between MUC1 and GSK3 b . Moreover and in contrast to GSK3 b , in vitro and in vivo studies demonstrate that c-Src-mediated phosphorylation of MUC1 increases binding of MUC1 and b -catenin. The findings support a novel role for c-Src in regulating interactions of MUC1 with GSK3 b and b -catenin. b -catenin, a component of the adherens junctions of mam-malian epithelial cells, directly to the cytoplasmic domain of the transmembrane E-cadherin protein that functions in Ca -dependent epithelial cell-cell In turn, a -catenin to -catenin and the to glutathione c-Src ab- sence °Cbefore GSK3 b an additional precipitated with the were separated by SDS-PAGE and to immunoblot analysis with anti-GSK3 b ZR-75-1 cells were transiently transfected with pCMV or pCMV/c-Src by electropo- ration. 48 h, the cells were harvested, and lysates were subjected to immunoprecipitation ( IP ) with anti-MUC1. The immunoprecipitates were analyzed by immunoblotting with anti-c-Src and anti-GSK3 b

The finding that ␤-catenin and GSK3␤ interact with the cytoplasmic domain of the DF3/MUC1 mucin-like glycoprotein has supported the involvement of an additional pathway in ␤-catenin signaling (16,17). MUC1 is highly overexpressed by human carcinomas (18). In addition, whereas MUC1 expression is restricted to the apical borders of normal secretory epithelial cells, MUC1 is aberrantly expressed by carcinoma cells at high levels throughout the cytoplasm and over the entire cell surface (18 -20). The MUC1 protein consists of an N-terminal ectodomain with variable numbers of 20-amino acid tandem repeats that are subject to extensive O-glycosylation (21,22). The C-terminal region includes a transmembrane domain and a 72-amino acid cytoplasmic tail. MUC1 is subject to proteolytic cleavage and the large ectodomain containing the tandem repeats can remain complexed to the 25-kDa C-terminal subunit or undergo release from the cell surface (23). ␤-catenin binds directly to MUC1 at a SAGNGGSSL motif in the cytoplasmic domain (16). Similar SXXXXXSSL sites in E-cadherin and APC are responsible for ␤-catenin interactions (4 -6). GSK3␤ also binds directly to MUC1 and phosphorylates serine in a DRSPY site adjacent to that for the ␤-catenin interaction (17). GSK3␤-mediated phosphorylation of MUC1 decreases the association of MUC1 and ␤-catenin (17).
The present studies demonstrate that the c-Src tyrosine kinase interacts directly with MUC1. A YEKV motif in the MUC1 cytoplasmic domain (CD) has been identified as a site for c-Src phosphorylation. The results demonstrate that c-Src regulates the interactions of MUC1 with GSK3␤ and ␤-catenin.
Immunoprecipitation and Immunoblotting-Equal amounts of protein from cell lysates were incubated with normal mouse IgG, MAb DF3 (anti-MUC1) (18), anti-c-Src (Upstate Biotechnology, Lake Placid, NY), or the rabbit anti-DF3-E antibody prepared against a peptide derived from the MUC1 extracellular domain (HDVETQFNQYKTEAAS). After incubation for 2 h at 4°C, the immune complexes were precipitated with protein G-agarose. The immunoprecipitates were washed with lysis buffer, separated by SDS-PAGE, and transferred to nitrocellulose membranes. The immunoblots were probed with 500 ng/ml anti-MUC1 or 1 g/ml anti-c-Src. Reactivity was detected with horseradish peroxidase-conjugated second antibodies and chemiluminescence (ECL, Amersham Pharmacia Biotech).
In Vitro Phosphorylation-Purified wild-type and mutant MUC1/CD proteins were incubated with 1. , GST-Src-SH2, or GST-␤-catenin bound to glutathione beads was then added, and the reaction was incubated for 1 h at 4°C. After washing, the proteins were subjected to SDS-PAGE and immunoblot analysis with the anti-MUC1/CD antibody that was generated against the cytoplasmic domain (17). In other studies, GST-MUC1/CD bound to glutathione beads was incubated with 1.5 units of c-Src in the presence and absence of 200 M ATP for 30 min at 30°C before adding 0.1 mg of purified GSK3␤ (New England BioLabs) for an additional 1 h. Precipitated proteins were analyzed by immunoblotting with anti-GSK3␤.

RESULTS AND DISCUSSION
To determine whether DF3/MUC1 forms a complex with c-Src, anti-MUC1 immunoprecipitates from lysates of human ZR-75-1 cells were analyzed by immunoblotting with anti-c-Src. The results demonstrate that c-Src coprecipitates with MUC1 ( Fig. 1A, left). In the reciprocal experiment, analysis of anti-c-Src immunoprecipitates by immunoblotting with anti-MUC1 confirmed the association of MUC1 and c-Src (Fig. 1A, right). Similar results have been obtained in human HeLa cells (data not shown). To assess whether the binding is direct, we incubated purified His-tagged MUC1 cytoplasmic domain (His-MUC1/CD) with a GST fusion protein that contains the c-Src SH3 domain. Analysis of the adsorbate to glutathione beads by immunoblotting with anti-MUC1/CD demonstrated binding of   MUC1 Integrates c-Src, GSK3␤, and ␤-Catenin Signaling MUC1/CD to GST-Src SH3, and not GST or a GST-Src SH2 fusion protein (Fig. 1B). As an additional control, His-MUC1/CD was incubated with a GST fusion protein containing a mutated c-Src SH3 domain (GST-Src SH3De90/92) (26). The finding that MUC1/CD binds to wild-type c-Src SH3 but not the mutant supported a direct interaction between MUC1 and c-Src (Fig. 1C).
To determine whether MUC1/CD is a substrate for c-Src, we incubated MUC1/CD with purified c-Src and [␥-32 P]ATP. Analysis of the reaction products by SDS-PAGE and autoradiography demonstrated c-Src-mediated phosphorylation of MUC1/CD ( Fig. 2A). Previous studies have demonstrated that GSK3␤ phosphorylates MUC1/CD on Ser at a DRSPYEKV site (17). As the adjacent YEKV sequence represents a consensus for c-Src phosphorylation, MUC1/CD was generated with a FEKV mutation (Fig. 2B). Incubation of MUC1/CD(Y46F) with c-Src demonstrated a decrease in phosphorylation as compared with that found with wild-type MUC1/CD (Fig. 2C). These findings indicate that c-Src phosphorylates MUC1/CD predominantly but not exclusively at the YEKV site. As the c-Src SH2 domain interacts with a preferred pYEEI sequence (27), c-Srcmediated phosphorylation of YEKV in MUC1/CD provides a potential site for c-Src SH2 binding. To determine whether the c-Src SH2 domain binds to phosphorylated MUC1/CD, we incubated MUC1/CD with c-Src and ATP and then assessed binding to GST-Src SH2. The results demonstrate that GST-Src SH2 associates with phosphorylated but not unphosphorylated MUC1/CD (Fig. 2D). Moreover, compared with MUC1/CD, there was substantially less binding of GST-Src SH2 to the MUC1/CD(Y46F) mutant that had been incubated with c-Src and ATP (Fig. 2D). These results support c-Src-mediated phosphorylation of MUC1/CD and thereby a direct interaction of phosphorylated MUC1/CD with the c-Src SH2 domain.
As the c-Src phosphorylation site on MUC1/CD resides next to the binding and phosphorylation site for GSK3␤ (17), we asked if the interaction of MUC1/CD with c-Src affects that with GSK3␤. GST-MUC1/CD was incubated with c-Src and ATP before addition of GSK3␤. Analysis of proteins precipitated with glutathione beads demonstrated that c-Src-mediated phosphorylation of MUC1/CD is associated with a decrease in binding of MUC1/CD and GSK3␤ (Fig. 3A). To assess the effects of c-Src on the interaction of MUC1/CD and GSK3␤ in vivo, ZR-75-1 cells were transfected to express the empty vector or c-Src. Anti-MUC1 immunoprecipitates were analyzed by immunoblotting with anti-GSK3␤. The results demonstrate that c-Src also decreases the interaction of MUC1 and GSK3␤ in vivo (Fig. 3B). These findings indicate that GSK3␤ interacts with MUC1/CD by a c-Src-dependent mechanism.
Phosphorylation of MUC1 by GSK3␤ decreases binding of MUC1 to ␤-catenin in vitro and in cells (17). To determine whether c-Src-mediated phosphorylation of MUC1 affects the interaction of MUC1 with ␤-catenin, we incubated MUC1/CD with c-Src and ATP. Phosphorylated and unphosphorylated MUC1/CD were then incubated with GST or GST-␤-catenin. MUC1 Integrates c-Src, GSK3␤, and ␤-Catenin Signaling 6063 MUC1/CD to GST-␤-catenin (Fig. 4A). By contrast, there was no detectable binding of phosphorylated or unphosphorylated MUC1/CD to GST (Fig. 4A). Studies performed with MUC1/ CD(Y46F) demonstrated that c-Src-dependent phosphorylation of the YEKV site on MUC1/CD is necessary for the formation of MUC1/CD-␤-catenin complexes (Fig. 4A). To assess whether c-Src affects the interaction of MUC1 and ␤-catenin in vivo, MUC1-positive ZR-75-1 cells were transfected with pCMV or pCMV/c-Src. Anti-MUC1 immunoprecipitates prepared from the transfected cells were subjected to immunoblot analysis with anti-c-Src, anti-P-Tyr, and anti-␤-catenin. The results demonstrate that c-Src associates with MUC1 in cells and induces tyrosine phosphorylation of MUC1 (Fig. 4B, left panel).
In addition, c-Src expression induced the interaction of MUC1 and ␤-catenin (Fig. 4B, left panel). The finding that the MUC1 C-terminal subunit and not the large ectodomain is subject to tyrosine phosphorylation is consistent with an interaction between c-Src and MUC1/CD (Fig. 4B, left panel). To confirm these findings, we performed immunoprecipitation studies with the anti-DF3-E antibody that was generated against the extracellular region of the C-terminal subunit. Immunoblot analysis of the precipitates demonstrated c-Src-mediated phosphorylation of the MUC1 C-terminal subunit and increased binding of MUC1/CD to ␤-catenin (Fig. 4B, middle panel). By contrast, expression of a kinase-inactive c-Src(K295R) mutant resulted in less phosphorylation of MUC1 on tyrosine as compared with the control (Fig. 4B, middle panel). Moreover, expression of c-Src(K295R) was associated with a decreased interaction between MUC1/CD and ␤-catenin (Fig. 4B, middle panel). To extend these findings, MUC1-negative 293 cells (17) were transfected to express MUC1 or MUC1(Y46F) in which the CD YEKV site has been mutated to FEKV. There was a low but detectable level of MUC1 binding to endogenous c-Src (Fig.  4C, left panel). Moreover, cotransfection of MUC1 and c-Src was associated with increased formation of MUC1-c-Src complexes (Fig. 4C, middle panel). Cotransfection of MUC1 and c-Src was also associated with increased tyrosine phosphorylation of MUC1 and binding of MUC1 and c-Src (Fig. 4C, middle panel). By contrast, cotransfection of MUC1(Y46F) and c-Src resulted in little binding of these proteins (Fig. 4C, middle panel). Moreover, there was little if any tyrosine phosphorylation of MUC1(Y46F) (Fig. 4C, middle panel). Importantly, cotransfection of MUC1 but not MUC1(Y46F) with c-Src induced the binding of MUC1 and ␤-catenin (Fig. 4C, middle panel). These findings demonstrate that c-Src-mediated phosphorylation of the MUC1 YEKV site increases the interaction of MUC1 and ␤-catenin in cells.
The present findings thus demonstrate that signaling of ␤-catenin and the MUC1 carcinoma-associated protein is regulated by the c-Src tyrosine kinase. Previous studies have shown that ␤-catenin interacts with the cytoplasmic domain of MUC1 and that GSK3␤ inhibits the formation of MUC1/CD-␤catenin complexes (17). By contrast, the present work supports a model in which c-Src phosphorylates MUC1/CD and promotes the interaction of MUC1/CD and ␤-catenin. The c-Src kinase functions in signaling pathways activated by heterotrimeric G protein-coupled receptors (28) and neuronal ion channels (29 -31). c-Src also participates in the transduction of signals from the epidermal growth factor receptor (EGF-R), platelet-derived growth factor receptor (PDGF-R) and other receptor tyrosine kinases (32). The available evidence indicates that c-Src phosphorylates the EGF-R and thereby contributes to mitogenesis and transformation (33). Mitogenesis induced by PDGF is also positively regulated by c-Src-mediated phosphorylation of the PDGF-R (34). Other substrates of c-Src that include focal adhesion kinase, p130Cas, and cortactin have functional associations with the actin cytoskeleton (32). These findings have collectively provided support for the involvement of c-Src in the integration of mitogenic, cell adhesion, and cytoskeletal responses. The present studies extend these findings by demonstrating that the MUC1 carcinoma-associated antigen is also a substrate for c-Src and that interaction of MUC1 with GSK3␤ and ␤-catenin are regulated by c-Src-dependent signals.