Protein Kinase C (cid:1) Regulates Function of the DF3/MUC1 Carcinoma Antigen in (cid:2) -Catenin Signaling*

The DF3/MUC1 mucin-like glycoprotein is aberrantly overexpressed in most human carcinomas. The MUC1 cytoplasmic domain interacts directly with (cid:2) -catenin, a component of the adherens junction of mammalian epithelial cells. The present results demonstrate that MUC1 associates with protein kinase C (cid:1) (PKC (cid:1) ). A TDR sequence adjacent to the (cid:2) -catenin binding motif in the MUC1 cytoplasmic domain functions as a site for PKC (cid:1) phosphorylation. We show that phosphorylation of MUC1 by PKC (cid:1) increases binding of MUC1 and (cid:2) -cate-nin in vitro and in vivo . The functional significance of the MUC1-PKC (cid:1) interaction is further supported by the demonstration that mutation of the PKC (cid:1) phosphorylation site abrogates MUC1-mediated decreases in binding of (cid:2) -catenin to E-cadherin. We also show that the stim-ulatory effects of MUC1 on anchorage-independent growth are abrogated by mutation of the PKC (cid:1) phosphorylation site. These findings support a novel role for PKC (cid:1) in regulating the interaction between MUC1 and the (cid:2) -catenin signaling pathway. The protein kinase C wild-type MUC1/CD and MUC1/CD(T41A) were incubated with 0.3 (cid:9) g of recombinant PKC (cid:4) (PanVera) for 1 h at 4 °C. Anti-PKC (cid:4) immunoprecipitates were analyzed by immunoblotting with anti-MUC1/CD (20) and anti-PKC (cid:4) . In other experiments, purified His- tagged wild-type and mutant MUC1/CD proteins were incubated with 0.3 (cid:9) g of PKC (cid:4) (PanVera) in the absence and presence of 200 (cid:9) M ATP for 20 min at 30 °C. GST- (cid:2) -catenin bound to glutathione beads was then added, and the reaction was incubated fo r 1 h at 4°C. The precipitated proteins were subjected to immunoblot analysis with anti-MUC1/CD (20). In Vitro Phosphorylation— Purified His-tagged, wild-type and mu- tant MUC1/CD proteins were incubated with 0.3 (cid:9) g of recombinant PKC (cid:4) (PanVera) in kinase buffer (20 m M Tris-HCl, pH 7.6, 10 m M MgCl 2 , 5 (cid:9) M dithiothreitol). The reactions were initiated by addition of 10 (cid:9) Ci of [ (cid:3) - 32 P]ATP and incubated for 20 min at 30 °C. Phosphorylated proteins were separated by SDS-PAGE and analyzed by autoradiography. Cells were in modified Eagle’s medium supplemented fetal serum antibiotics. The cell suspension was layered over 3.5 ml agar/Dulbecco’s modified Eagle’s medium in 60-mm dishes. One

The DF3/MUC1 mucin-like glycoprotein is aberrantly overexpressed in most human carcinomas. The MUC1 cytoplasmic domain interacts directly with ␤-catenin, a component of the adherens junction of mammalian epithelial cells. The present results demonstrate that MUC1 associates with protein kinase C␦ (PKC␦). A TDR sequence adjacent to the ␤-catenin binding motif in the MUC1 cytoplasmic domain functions as a site for PKC␦ phosphorylation. We show that phosphorylation of MUC1 by PKC␦ increases binding of MUC1 and ␤-catenin in vitro and in vivo. The functional significance of the MUC1-PKC␦ interaction is further supported by the demonstration that mutation of the PKC␦ phosphorylation site abrogates MUC1-mediated decreases in binding of ␤-catenin to E-cadherin. We also show that the stimulatory effects of MUC1 on anchorage-independent growth are abrogated by mutation of the PKC␦ phosphorylation site. These findings support a novel role for PKC␦ in regulating the interaction between MUC1 and the ␤-catenin signaling pathway.
The protein kinase C (PKC) 1 family of serine/threonine protein kinases is involved in intracellular signaling pathways that regulate growth, differentiation, and apoptosis. The PKC isoforms have been divided into: (i) the conventional PKCs (cPKCs; ␣, ␤, ␥) which are dependent on calcium and activated by diacylglycerol or 12-O-tetradecanoylphorbol-13-acetate; (ii) the novel PKCs (nPKCs; ␦, ⑀, ) which are calcium-independent and activated by diacylglycerol or 12-O-tetradecanoylphorbol-13-acetate; and (iii) the atypical PKCs (aPKCs; , ) which are calcium-independent and not activated by diacylglycerol or 12-O-tetradecanoylphorbol-13-acetate (1,2). Of the 12 known PKC isoforms, the ubiquitously expressed PKC␦ is unique as a substrate for tyrosine phosphorylation in response to activation of the epidermal growth factor receptor (EGFR) (3), the plateletderived growth factor receptor (4), or the insulin-like growth factor I receptor (4). Transformation by Ras (5) or v-Src (6) also results in phosphorylation of PKC␦ on tyrosine. Interactions between PKC␦ and EGFR or c-Src have supported a role for PKC␦ as a tumor suppressor (7,8). Other studies have provided support for involvement of PKC␦ in the apoptotic response of cells to genotoxic and oxidative stress (9 -12). Targeting of PKC␦ to mitochondria induces apoptosis through loss of the mitochondrial transmembrane potential, release of cytochrome c, and activation of caspase-3 (12)(13)(14)(15).
The human DF3/MUC1 mucin-like transmembrane glycoprotein is expressed on the apical borders of normal secretory epithelial cells and at high levels over the entire surface of carcinoma cells (16). The MUC1 protein consists of an N-terminal ectodomain with variable numbers of 20 amino acid tandem repeats that are extensively modified by O-glycosylation (17,18). The Ͼ250-kDa ectodomain associates with a ϳ25-kDa C-terminal proteolytic fragment as a heterodimer at the cell surface. The C-terminal subunit includes a transmembrane domain and a 72-amino acid cytoplasmic domain (CD). ␤-Catenin, a component of the adherens junction of mammalian epithelial cells, binds directly to the MUC1/CD at a serine-rich motif that is similar to ␤-catenin-binding sites on E-cadherin and the adenomatous polyposis coli (APC) tumor suppressor (19). MUC1 competes with E-cadherin, but not APC, for binding to ␤-catenin (20). The interaction between MUC1 and ␤-catenin is down-regulated by phosphorylation of the MUC1/CD by glycogen synthase kinase 3␤ (GSK3␤) (20). Conversely, phosphorylation of MUC1 by c-Src stimulates binding of MUC1 to ␤-catenin (21). The recent findings that EGFRmediated phosphorylation of MUC1 regulates the interaction of MUC1 with c-Src and ␤-catenin has suggested that aberrant overexpression of MUC1 in human carcinoma cells could contribute to the transformed phenotype by dysregulation of EGFR signaling (22,23).
The present studies demonstrate that PKC␦ interacts with MUC1. A T41DR motif in the MUC1/CD has been identified as a PKC␦ phosphorylation site. The results show that PKC␦ regulates binding of MUC1 to ␤-catenin and that the T41A mutant abrogates the effects of MUC1 on anchorage-independent cell growth.
Binding Studies-Glutathione S-transferase (GST) or GST-MUC1/CD bound to glutathione beads were incubated with 0.3 g of purified recombinant, kinase-active PKC␦ (PanVera Corp., Madison, WI). The adsorbates were analyzed by immunoblotting with anti-PKC␦. His-tagged wild-type MUC1/CD and MUC1/CD(T41A) were incubated with 0.3 g of recombinant PKC␦ (PanVera) for 1 h at 4°C. Anti-PKC␦ immunoprecipitates were analyzed by immunoblotting with anti-MUC1/CD (20) and anti-PKC␦. In other experiments, purified Histagged wild-type and mutant MUC1/CD proteins were incubated with 0.3 g of PKC␦ (PanVera) in the absence and presence of 200 M ATP for 20 min at 30°C. GST-␤-catenin bound to glutathione beads was then added, and the reaction was incubated for 1 h at 4°C. The precipitated proteins were subjected to immunoblot analysis with anti-MUC1/CD (20).
In Vitro Phosphorylation-Purified His-tagged, wild-type and mutant MUC1/CD proteins were incubated with 0.3 g of recombinant PKC␦ (PanVera) in kinase buffer (20 mM Tris-HCl, pH 7.6, 10 mM MgCl 2 , 5 M dithiothreitol). The reactions were initiated by addition of 10 Ci of [␥-32 P]ATP and incubated for 20 min at 30°C. Phosphorylated proteins were separated by SDS-PAGE and analyzed by autoradiography.
Anchorage-independent Growth-Cells (4 ϫ 10 4 ) were suspended in 1.5 ml of 0.33% Noble agar (DIFCO, Detroit, MI) in Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated fetal bovine serum and antibiotics. The cell suspension was layered over 3.5 ml of 0.5% agar/Dulbecco's modified Eagle's medium in 60-mm dishes. One ml of fresh culture medium was added at 2 weeks. Colonies composed of Ͼ10 cells were counted at 3 weeks.

MUC1 Binds
Directly to PKC␦-To determine whether MUC1 associates with PKC␦, lysates from human ZR-75-1 cells were subjected to immunoprecipitation with anti-MUC1 and, as a control, IgG. Immunoblot analysis of the precipitates with anti-PKC␦ demonstrated the detection of MUC1-PKC␦ complexes (Fig. 1A, left). In the reciprocal experiment, immunoblot analysis of anti-PKC␦ immunoprecipitates with anti-MUC1 confirmed that MUC1 associates with PKC␦ ( Fig. 1A, right). By contrast, there was no detectable interaction between MUC1 and PKC␤II, PKC or PKC (Fig. 1B). To extend these findings, 293 cells, which are negative for MUC1 (20), were transfected to express MUC1 or MUC1ϩPKC␦. Immunoblot analysis of anti-MUC1 immunoprecipitates with anti-PKC␦ demonstrated binding of MUC1 with endogenous PKC␦ (Fig. 1C). Moreover, coexpression of MUC1 and PKC␦ resulted in increased formation of MUC1-PKC␦ complexes (Fig. 1C). To assess whether binding is direct, purified GST or a GST fusion protein containing the MUC1/CD (GST-MUC1/CD) was incubated with recombinant PKC␦. Adsorbates to glutathione beads were subjected to immunoblot analysis with anti-PKC␦. The finding that PKC␦ binds to GST-MUC1/CD and not to GST alone supported a direct interaction (Fig. 1D).
PKC␦ Phosphorylates MUC1 on T41-To determine whether MUC1/CD is a substrate for PKC␦, we incubated purified His-MUC1/CD with recombinant PKC␦ and [␥-32 P]ATP. Analysis of the reaction products by SDS-PAGE and autoradiography demonstrated phosphorylation of MUC1/CD ( Fig. 2A). A ST41DRS site in MUC1/CD conforms to the preferred (S/T)X(K/R) motif for PKC␦ phosphorylation. To determine whether ST41DRS is phosphorylated by PKC␦, we mutated this site in MUC1/CD to SA41DRS (Fig. 2B). PKC␦-mediated phosphorylation of MUC1/ CD(T41A) was attenuated compared with that obtained with wild-type MUC1/CD (Fig. 2C, left). By contrast, phosphorylation of MUC1 by PKC␦ was unaffected by Ser 3 Ala mutations of either or both of the flanking serines (Fig. 2C, left). The results also demonstrate that autophosphorylation of PKC␦ is decreased in the presence of MUC1/CD(T41A) (Fig. 2C, left). To determine whether the T41A mutation affects binding of PKC␦, wild-type MUC1/CD and MUC1/CD(T41A) were incubated with recombinant PKC␦. Immunoblot analysis of anti-PKC␦ immunoprecipitates with anti-MUC1/CD demonstrated that PKC␦ binds equally to wild-type MUC1/CD and MUC1/ CD(T41A) (Fig. 2C, right). Previous studies have shown that phosphorylation of the MUC1/CD Ser 44 site by GSK3␤ decreases the interaction between MUC1/CD and ␤-catenin (20). To assess the effects of PKC␦-mediated phosphorylation of MUC1/CD, we incubated MUC1/CD with PKC␦ in the presence and absence of ATP. After phosphorylation of MUC1/CD, GST or GST-␤-catenin was added for 1 h at 4°C. Proteins precipitated with glutathione beads were analyzed by immunoblotting with anti-MUC1/CD. As shown previously, MUC1/CD binds to GST-␤-catenin and not GST (20). Preincubation of MUC1/CD with PKC␦ and ATP was associated a higher level of MUC1/CD binding to GST-␤-catenin than that obtained in the absence of PKC␦ or ATP (Fig. 2D). By contrast, preincubation of MUC1/ CD(T41A) with PKC␦ and ATP had no detectable effect on binding of MUC1/CD(T41A) to ␤-catenin (Fig. 2D). The finding that, in the absence of PKC␦ phosphorylation, less ␤-catenin binds to MUC1/CD(T41A) as compared with that for wild-type MUC1/CD may reflect conformational effects of the T41A mutation on the ␤-catenin-binding site (Fig. 2D). These findings demonstrate that PKC␦ phosphorylates MUC1/CD on Thr 41 and thereby increases binding of MUC1/CD and ␤-catenin.
PKC␦ Regulates Interaction of MUC1 with ␤-Catenin in Vivo-To determine whether PKC␦ regulates the interaction between MUC1 and ␤-catenin in vivo, transfection studies were performed in the MUC1-negative 293 cells. After transfection of vectors expressing MUC1 and PKC␦ or the kinase-inactive PKC␦(K-R) mutant, lysates were subjected to immunoprecipitation with anti-MUC1. Immunoblot analysis of the precipitates with anti-␤-catenin demonstrated that PKC␦ increases the interaction between MUC1 and ␤-catenin as compared with that in cells transfected with PKC␦(K-R) (Fig. 3A). In concert with these results and involvement of Thr 41 , PKC␦ had little if any effect on binding of ␤-catenin to the MUC1(T41A) mutant (Fig. 3A). To extend the analysis, we transfected MUC1-negative HCT116 cells to stably express the empty vector, wild-type MUC1, or MUC1(T41A) (Fig. 3B, left). Anti-MUC1 immunoprecipitates from HCT116/V, HCT116/MUC1, and HCT116/ MUC1(T41A) cells were subjected to immunoblotting with anti-␤-catenin. The results demonstrate that MUC1, but not MUC1(T41A), binds to ␤-catenin (Fig. 3B, right). When these cells were transfected to express GFP-PKC␦, immunoblot anal- ysis of anti-MUC1 immunoprecipitates with anti-␤-catenin demonstrated that PKC␦ induces binding of ␤-catenin to wildtype MUC1 and not the MUC1(T41A) mutant (Fig. 3B, right). Binding of MUC1 to endogenous PKC␦ and ectopically expressed GFP-PKC␦ was also decreased with the MUC1(T41A) mutant as compared with that with wild-type MUC1 (Fig. 3B,  right). Previous work has demonstrated that MUC1 and Ecadherin compete for binding to ␤-catenin (20). To determine whether expression of the MUC1(T41A) mutant affects binding of ␤-catenin to E-cadherin, anti-E-cadherin immunoprecipitates were analyzed by immunoblotting with anti-␤-catenin. In concert with previous findings (20), expression of wild-type MUC1 was associated with decreased binding of E-cadherin and ␤-catenin (Fig. 3C). By contrast, expression of MUC1(T41A) had less of an effect on the interaction of Ecadherin and ␤-catenin compared with that in HCT116/MUC1 cells (Fig. 3C). Overexpression of GFP-PKC␦ was associated with decreases in binding of ␤-catenin to E-cadherin in HCT116/V cells (Fig. 3C). Moreover, while GFP-PKC␦ had little additional effect on the interaction of E-cadherin and ␤-catenin in HCT116/MUC1 cells, expression of both GFP-PKC␦ and the MUC1(T41A) mutant resulted in an increase in E-cadherin-␤catenin complexes (Fig. 3C). These findings demonstrate that PKC␦ regulates the interaction between MUC1 and ␤-catenin in cells and thereby binding of E-cadherin with ␤-catenin.
Effects of MUC1 on Anchorage-independent Growth Are Abrogated by the T41A Mutation-To assess the functional significance of the interaction between MUC1 and PKC␦, HCT116/V, HCT116/MUC1, and HCT116/MUC1(T41A) cell lines were analyzed for anchorage-dependent and -independent growth. Expression of wild-type MUC1 or the MUC1(T41A) mutant had no apparent effect on growth in tissue culture compared with that for HCT116/V cells (Fig. 4A). Similar results were obtained with clones selected from separate transfections (Fig.  4A). In soft agar, the wild-type MUC1 transfectants formed colonies that were substantially larger than those obtained  2-4). Lysate not subjected to immunoprecipitation was used as a control (right, lane 1). Similar studies were performed on cells transfected to express GFP-PKC␦ (right, lanes 5-7). C, anti-E-cadherin immunoprecipitates from HCT116/V, HCT116/ MUC1, and HCT116/MUC1(T41A) were analyzed by immunoblotting with anti-␤catenin (upper panel) and anti-E-cadherin (lower panel) (lanes 1-3). Similar studies were performed on cells transfected to express GFP-PKC␦ (lanes 4-6).
with HCT116/V cells (Fig. 4B). By contrast, expression of MUC1(T41A) was associated with the formation of colonies that were similar to those found with HCT116/V cells (Fig. 4B). Similar results were obtained with the independently selected clones (Fig. 4B). The number of colonies obtained for HCT116/ MUC1 cells was also higher than those found for HCT116/V and HCT116/MUC1(T41A) cells (Fig. 4C). These findings demonstrate that expression of wild-type MUC1 contributes to anchorage-independent growth and that mutation of the PKC␦ phosphorylation site abrogates this effect. DISCUSSION PKC␦ has been implicated in the response of cells to activation of the EGFR, platelet-derived growth factor receptor, and insulin-like growth factor I receptors (3,4). Other findings have supported a role for PKC␦ in the apoptotic response of cells to stress (9 -15). Functional involvement of PKC␦ at the cell membrane has also been shown through regulation of phospholipid scramblase activity during both cell activation and apoptosis (25). The present studies demonstrate that PKC␦ interacts with the transmembrane MUC1 protein. The results show that PKC␦ phosphorylates the MUC1 cytoplasmic domain at Thr 41 (Fig. 4D). Previous work has shown that GSK3␤ phosphorylates Ser 44 in the TDRSPYE domain of MUC1/CD (Fig. 4D) (20). In addition, the Tyr 46 site is phosphorylated by EGFR (23) and c-Src (21) (Fig. 4D). Phosphorylation of Ser 44 by GSK3␤ decreases ␤-catenin binding, while phosphorylation of Tyr 46 increases the interaction of MUC1 and ␤-catenin. The present results demonstrate that PKC␦ also regulates the formation of MUC1-␤-catenin complexes. In vitro binding of PKC␦ was similar with wild-type MUC1/CD and the MUC1/CD(T41A) mutant. By contrast, binding of PKC␦ to MUC1(T41A) in vivo was decreased compared with that found for wild-type MUC1. These findings suggest that, in the presence of other binding proteins, such as c-Src and GSK3␤, the T41A mutation disrupts the interaction between MUC1 and PKC␦. Other work has shown that c-Src disrupts binding of MUC1 and GSK3␤ (20). Thus, given the proximity of the MUC1 sites for interactions with c-Src, GSK3␤, and PKC␦ (Fig. 4D), modification of Thr 41 , Ser 44 , and/or Tyr 46 by phosphorylation or mutation may affect the integration of MUC1 signaling with diverse pathways.
The available evidence indicates that MUC1 and E-cadherin compete for binding to the same pool of ␤-catenin (20). Ecadherin functions in homotypic recognition and the regulation of cell mobility (26). The interaction between E-cadherin and ␤-catenin is essential for cell adhesion by connecting E-cadherin to ␣-catenin and thereby the cytoskeleton (27)(28)(29)(30). Importantly, disruption of E-cadherin function and cell adhesion has been associated with tumor development (29,(31)(32)(33).
Other studies have demonstrated that MUC1 affects E-cadherin-mediated cell adhesion (34). The present results demonstrate that PKC␦ increases the interaction between MUC1 and ␤-catenin in vitro and in cells. Moreover, mutation of the MUC1 Thr 41 phosphorylation site abrogates the effects of PKC␦ on formation of MUC1-␤-catenin complexes. The in vitro binding studies indicate that the T41A mutation may also affect the MUC1/CD-␤-catenin interaction, perhaps by causing changes in the tertiary structure of the ␤-catenin-binding site. Nonetheless, while MUC1 decreased binding of E-cadherin and ␤-catenin in cells, expression of the MUC1(T41A) mutant in part reversed this effect. Whereas enforced expression of PKC␦ in HCT116/MUC1(T41A) cells was also associated with increased formation of E-cadherin-␤-catenin complexes, these experimental conditions could result in interactions not found with endogenous PKC␦. The findings thus support a signaling pathway in which the interaction between MUC1 and PKC␦ increases binding of MUC1 and ␤-catenin, while decreasing the formation of E-cadherin-␤-catenin complexes.
MUC1 is normally expressed at the apical borders of secretory epithelial cells (16). In carcinoma cells, polarization of MUC1 is lost with high levels of expression over the entire cell surface (16,35). The apical border of the normal glandular epithelium is devoid of cell-cell interactions at the surface lining secretory ducts. Overexpression of MUC1 by carcinoma cells, however, could confer an anti-adhesive function to the entire cell surface by disrupting E-cadherin-mediated homotypic recognition. To address the function of MUC1 on cell growth, MUC1-negative HCT116 cells were stably transfected to express wild-type MUC1 or the MUC1(T41A) mutant. Expression of MUC1 or MUC1(T41A) had no significant effect on adherent growth in tissue culture. By contrast, wild-type MUC1 conferred a substantial increase in anchorage-independent growth in soft agar. These effects could not be attributed to the heavily glycosylated ectodomain because anchorage-independent growth of HCT116 cells expressing similar levels of MUC1(T41A) was comparable with that found with control HCT116/V cells. Rather, the distinct patterns of anchorageindependent growth for HCT116 cells expressing wild-type MUC1 or MUC1(T41A) support direct involvement of the PKC␦ phosphorylation site. These findings and the demonstration that PKC␦-mediated phosphorylation of MUC1 regulates binding of MUC1 and ␤-catenin are in concert with a model in which PKC␦ subverts E-cadherin function by titrating binding of ␤-catenin to MUC1. Conversely, in a simplified model, mutation of MUC1 at Thr 41 restores binding of ␤-catenin to Ecadherin and decreases anchorage-independent growth.
In addition to interactions with E-cadherin and MUC1 at the cell membrane, ␤-catenin binds directly to the APC gene product in the cytosol (36,37). The interaction between adenomatous polyposis coli and ␤-catenin regulates ␤-catenin turnover (38) and alters cell adhesion (39). ␤-Catenin also forms nuclear complexes with members of the T-cell factor/lymphoid enhancing factor 1 family of transcription factors (40,41). Loss of APC-mediated regulation of ␤-catenin in transformed cells is associated with constitutive activation of ␤-catenin-T-cell factor/lymphoid enhancing factor 1 transcriptional complexes (42)(43)(44). The present findings contribute to a role for MUC1 in regulating ␤-catenin at the cell membrane. While regulation of ␤-catenin may also occur as a result of binding to MUC1 that accumulates in the cytosol of carcinoma cells (16), there are presently no known functions for MUC1 in control of ␤-catenin signaling in the nucleus.