Catenin to the Nucleolus by a Mechanism Dependent on the DF3/MUC1 Oncoprotein

Vol. 1, 765–775, August 2003 Molecular Cancer Research 765

Heregulin Targets ;-Catenin to the Nucleolus by a Mechanism Dependent on the DF3/MUC1 Oncoprotein

Yongqing Li,1,2 Wei-hsuan Yu,1 Jian Ren,1 Wen Chen,1 Lei Huang,1 Surender Kharbanda,1,2 Massimo Loda,1 and Donald Kufe1

1Dana-Farber Cancer Institute, Harvard Medical School and 2ILEX Products, Inc., Boston, MA

Abstract EGFR-ErbB2 heterodimers (4). Similarly, heregulin/neuregu- The DF3/MUC1 transmembrane oncoprotein is lin-1 (HRG) binds to the ErbB3 and ErbB4 receptors and aberrantly overexpressed in most human breast activates ErbB2 through heterodimerization and transphos- carcinomas and interacts with the Wnt effector phorylation (5). ErbB2 may thus function as a coreceptor ;-catenin. Here, we demonstrate that MUC1 associates that potentiates signaling of the other ErbB family members constitutively with ErbB2 in human breast cancer cells (6–8). Dimerization of the ErbB receptors results in activa- and that treatment with heregulin/neuregulin-1 (HRG) tion of the intrinsic kinase function and phosphorylation of increases the formation of MUC1-ErbB2 complexes. tyrosine residues that serve as binding sites for proteins that The importance of the MUC1-ErbB2 interaction is contain Src homology 2 or phosphotyrosine binding domains supported by the demonstration that HRG induces (9, 10). Activation of ErbB2 is also associated with disrup- binding of MUC1 and ;-catenin and targeting of the tion of epithelial cell polarity and initiation of proliferation MUC1-;-catenin complex to the nucleolus. Significantly, (11, 12). In normal polarized glandular epithelial cells, nucleolar localization of ;-catenin in response to HRG is effectors of the Wnt signaling pathway, h- and g-catenin, dependent on MUC1 expression. Moreover, mutation of a are localized to the adherens junction where they function RRK motif in the MUC1 cytoplasmic domain abrogates with E-cadherin in cell-cell interactions (13). Loss of polarity HRG-induced nucleolar localization of MUC1 and as found with ErbB2 activation (11), however, is associated ;-catenin. In concert with these results, we show with catenin translocation from the adherens junction to the nucleolar localization of MUC1 and ;-catenin in human cytoplasm and nucleus (14). A functional relationship between breast carcinomas but not in normal mammary ductal ErbB2 signaling and Wnt regulation of catenins is unknown, epithelium. These findings demonstrate that MUC1 although both ErbB2 and Wnt have been linked to the functions in cross talk between ErbB2 and Wnt pathways development of breast carcinomas. by acting as a shuttle for HRG-induced nucleolar Human DF3/MUC1 is a mucin-like transmembrane glyco- targeting of ;-catenin. protein, which is overexpressed by breast and other carcinomas (15). MUC1 expression is restricted to the apical borders of normal secretory epithelial cells and is aberrantly expressed Introduction by breast carcinoma cells at high levels over the entire cell surface (15). Importantly, overexpression of MUC1 is The ErbB family of receptor tyrosine kinases includes sufficient to induce transformation (16). The MUC1 protein ErbB1/epidermal growth factor receptor (EGFR), ErbB2/neu, consists of a NH -terminal (N-ter) ectodomain with variable ErbB3, and ErbB4. Activation of ErbB1, ErbB3, and ErbB4 is 2 numbers of conserved 20-amino acid tandem repeats that are conferred by direct binding of at least 10 different growth modified by O-glycosylation (17, 18). The f25-kd COOH- factors that induce receptor homodimerization and hetero- terminal (C-ter) subunit includes a transmembrane domain and dimerization (1). The ErbB2 receptor, which has no known a 72-amino acid cytoplasmic domain (CD). The extracellular ligand, is transactivated through heterodimerization with the >250-kd ectodomain associates with the C-ter subunit as a other ErbB family members (2, 3). Stimulation of EGFR with heterodimer. A SAGNGGSSL motif in the MUC1-CD the epidermal growth factor (EGF) induces the formation of functions as a binding site for h-catenin (19). The SAGNG- GSSL motif also serves as a binding site for g-catenin (plakoglobin) (19). Glycogen synthase kinase 3h (GSK3h) phosphorylates MUC1 on serine in a SPY site adjacent to that for h/g-catenin binding and decreases the interaction between Received 3/3/03; revised 6/20/03; accepted 6/24/03. The costs of publication of this article were defrayed in part by the payment of MUC1 and h-catenin (20). Conversely, EGFR- or c-Src- page charges. This article must therefore be hereby marked advertisement in mediated phosphorylation of MUC1 on tyrosine in the SPY accordance with 18 U.S.C. Section 1734 solely to indicate this fact. site up-regulates the formation of MUC1-h-catenin complexes Grant support: National Cancer Institute grant CA97098. Note: Y.L. and W.-h.Y. contributed equally to this work. (21, 22). The demonstration that MUC1 and E-cadherin, a Requests for reprints: Donald Kufe, Dana-Farber Cancer Institute, Harvard transmembrane protein that functions in Ca2+-dependent Medical School, Boston, MA 02115. Phone: (617) 632-3141; Fax: (617) 632-2934. E-mail: [email protected] epithelial cell-cell interactions (23), compete for binding to Copyright D 2003 American Association for Cancer Research. h-catenin (20) has supported a role for MUC1 in regulating

adherens junction function. Other studies have demonstrated there was no increased colocalization of MUC1 and ErbB2 that MUC1 also colocalizes with h-catenin in the nucleus in cells stimulated with EGF (Fig. 1E). These findings (16, 24). Less is known about the regulation of binding demonstrate that colocalization of MUC1 and ErbB2 at the between MUC1 and g-catenin. cell membrane is regulated by HRG stimulation. The present studies demonstrate that MUC1 interacts with ErbB2 and that HRG stimulation of human breast carcinoma cells is associated with increased binding of HRG Regulates Interaction of MUC1 and c-Catenin MUC1 and g-catenin. The functional significance of this To determine whether HRG affects the interaction be- signaling pathway is supported by the finding that HRG targets tween MUC1 and catenins, lysates from control and HRG- g-catenin to the nucleolus by a MUC1-dependent mechanism treated ZR-75-1 cells were subjected to immunoprecipitation and that a RRK motif in MUC1-CD is required for this with anti-MUC1. Immunoblot analysis of the precipitates with response. anti-h-catenin demonstrated that HRG has little effect on binding of MUC1 and h-catenin (Fig. 2A). By contrast, HRG treatment was associated with an increase in binding of MUC1 and g-catenin (Fig. 2A). For comparison, ZR-75-1 cells were Results stimulated with EGF. As shown previously, EGF induced HRG Induces the Association of MUC1 and ErbB2 binding of MUC1 and h-catenin (22) (Fig. 2B). Conversely, Previous studies have demonstrated that human ZR-75-1 EGF had little effect on the interaction of MUC1 with breast cancer cells express MUC1 and the four ErbB family g-catenin (Fig. 2B). To extend these findings, we used human members (EGFR and ErbB2–4) (20, 22, 25). To determine HCT116 carcinoma cells that are MUC1 negative as whether MUC1 associates with ErbB2, anti-MUC1 (DF3) determined by immunoblotting with anti-MUC1 antibodies N-ter immunoprecipitates from lysates of human ZR-75-1 and by reverse transcription-PCR for sequences encoding the cells were analyzed by immunoblotting with anti-ErbB2. The C-ter [(26) and data not shown]. Moreover, flow cytometric results demonstrate that ErbB2 coprecipitates with MUC1 analysis of HCT116 cells demonstrated that all four ErbB (Fig. 1A). Whereas HRG stimulates ErbB2 activity, lysates family members are expressed at the cell membrane and that were prepared from ZR-75-1 cells treated with HRG for ErbB2 is detectable at somewhat higher levels than these 5 min. Immunoblot analysis of anti-MUC1 immunoprecipi- found for EGFR, ErbB3, and ErbB4 (Fig. 2C). HCT116 cells tates with anti-ErbB2 demonstrated that HRG stimulates the that stably express an empty vector or MUC1 were treated formation of complexes containing MUC1 and ErbB2 with HRG. In concert with the findings in ZR-75-1 cells, (Fig. 1A). In the reciprocal experiment, immunoblot analysis immunoblot analysis of anti-MUC1 immunoprecipitates with of anti-ErbB2 immunoprecipitates with anti-MUC1 confirmed anti-g-catenin demonstrated that HRG induces binding of that HRG increases the basal association of MUC1 and MUC1 and g-catenin (Fig. 2D). These findings indicate that ErbB2 (Fig. 1A). Treatment of ZR-75-1 cells with EGF had HRG stimulates the formation of MUC1-g-catenin complexes. little (if any) effect on binding of MUC1 and EGFR (22). As a control and in contrast to the effects of HRG, treatment with EGF also had no apparent effect on binding of MUC1 Nucleolar Localization of MUC1-c-Catenin Complexes and ErbB2 (data not shown). HRG binds to ErbB3 and ErbB4 To define the subcellular localization of MUC1-g-catenin and induces their heterodimerization with ErbB2 (3). To complexes, ZR-75-1 cells were analyzed by confocal determine whether MUC1 associates with ErbB3 or ErbB4, microscopy after incubation with antibodies against MUC1 immunoprecipitates prepared with antibodies against these C-ter and g-catenin. The results show colocalization of MUC1 receptors were subjected to immunoblotting with anti-MUC1. C-ter and g-catenin at the cell membrane (Fig. 3A). By The results show that MUC1 associates with ErbB3 and contrast, HRG stimulation for 20 min was associated with ErbB4 (Fig. 1B). Moreover, HRG stimulated the association localization of MUC1 C-ter in the nucleus (Fig. 3B). A similar of MUC1 with ErbB3 and ErbB4, but to a much lesser extent pattern was observed for g-catenin, and overlay demonstrated than that found for MUC1 and ErbB2 (Fig. 1B). To define colocalization with MUC1 C-ter (Fig. 3B). The well-circum- the subcellular localization of MUC1 and ErbB2, confocal scribed colocalization of MUC1 and g-catenin in the nucleus microscopy was performed with mouse anti-MUC1 and rabbit suggested a nucleolar pattern (Fig. 3B). Indeed, staining with anti-ErbB2. In control ZR-75-1 cells, MUC1 was distributed an anti-nucleolin antibody confirmed HRG-induced redistri- uniformly over the cell membrane (Fig. 1C, left). A similar bution of MUC1 C-ter to the nucleolus (Fig. 3C). A similar pattern was obtained for the distribution of ErbB2 (Fig. 1C, pattern of nucleolar colocalization for MUC1 C-ter with second panel). Overlay of the signals supported some g-catenin was observed in the ErbB2-positive MCF-7 breast colocalization (red + green ! yellow) (Fig. 1C, right). cancer cells (data not shown). Notably, stimulation of ZR-75-1 Following HRG stimulation for 5 min, MUC1 was clustered cells with EGF was associated with localization of MUC1 in patches on the cell surface (Fig. 1D, left). Staining for C-ter in a diffuse pattern throughout the nucleus (Fig. 3D). ErbB2 revealed a similar pattern (Fig. 1D, second panel), and Moreover, the lack of colocalization with nucleolin indicated overlay of the signals showed increased colocalization of that EGF induces nuclear targeting of MUC1 C-ter to MUC1 and ErbB2 in clusters at the cell membrane (Fig. 1D, nonnucleolar sites (Fig. 3D). Following EGF stimulation, right). There was no apparent HRG-induced localization of nuclear MUC1 C-ter colocalizes with h-catenin and not MUC1 N-ter to the nucleus (Fig. 1D). Moreover, as a control, g-catenin (unpublished data).

FIGURE 1. HRG stimulates interaction of MUC1 and ErbB2. ZR-75-1 cells were left untreated or stimulated with 20-ng/ml HRG for 5 min. A. Lysates were subjected to immunoprecipitation (IP) with anti-MUC1 (DF3) N-ter (left panel) or anti-ErbB2 (right panel). Mouse IgG was used as a control. The immunoprecipitates were analyzed by immunoblotting (IB) with anti-ErbB2 and anti-MUC1 N-ter. Intensity of the signals was determined by densitometric scanning and compared with that obtained for untreated cells. B. Lysates from control and HRG-treated ZR-75-1 cells were subjected to immunoprecipitation with anti-ErbB3 (left panel) or anti-ErbB4 (right panel). The immunoprecipitates were analyzed by immunoblotting with the indicated antibodies. ZR-75-1 cells were grown to 60% confluence and incubated in medium with 0.1% serum for 24 h. The cells were left untreated (C), stimulated with 20-ng/ml HRG for 5 min (D), or stimulated with 10-ng/ml EGF for 5 min (E), fixed, and double stained with anti-MUC1 N-ter (green) and anti-ErbB2 (red).

Role of MUC1 in the Subcellular Distribution of c-Catenin was associated with colocalization of MUC1 C-ter and g-catenin To assess the functional role of MUC1 in g-catenin in discrete nuclear structures (Fig. 4B). As found in ZR-75-1 signaling, HCT116/vector and HCT116/MUC1 cells were cells, colocalization of MUC1 C-ter and nucleolin indicated analyzed for localization of g-catenin following HRG stimu- that MUC1 C-ter and g-catenin are targeted to the nucleolus lation. The confocal images show that g-catenin localizes to the (data not shown). cell membrane of HCT116/vector cells (Fig. 4A). Moreover, Whereas a RRK motif in MUC1-CD may contribute to treatment of the HCT116/vector cells with HRG for 20 min had nuclear localization, similar studies were performed on no apparent effect on the distribution of g-catenin (Fig. 4A). HCT116 cells stably expressing a MUC1(RRK ! AAA) In HCT116/MUC1 cells, MUC1 C-ter and g-catenin were mutant. Coimmunoprecipitation studies demonstrated that predominantly detectable at the cell membrane (Fig. 4B). By binding of MUC1 to g-catenin is not affected by the RRK contrast, HRG treatment of HCT116/MUC1 cells for 20 min ! AAA mutation (data not shown). In contrast to HCT116/

FIGURE 2. HRG stimulates the interaction between MUC1 and g-catenin. A. Lysates from ZR-75-1 cells left untreated or stimulated with HRG for 5 min were subjected to immunoprecipitation with anti-MUC1 N-ter or, as a control, IgG. The immunoprecipitates were analyzed by immunoblotting with the indicated antibodies. B. Lysates from ZR-75-1 cells left untreated or stimulated with 10-ng/ml EGF for 5 min were subjected to immunoprecipitation with anti-MUC1 or IgG. The immunoprecipitates were analyzed by immunoblotting with the indicated antibodies. C. HCT116 cells were incubated with antibodies against the indicated ErbB family members (open patterns) or control mouse IgG (solid patterns) and analyzed by flow cytometry. Similar results were obtained for HCT116/MUC1 cells. D. HCT116/vector (HCT116/V) and HCT116/MUC1 cells were left untreated or stimulated with HRG. Anti-MUC1 N-ter immunoprecipitates were subjected to immunoblotting with anti-g-catenin or anti-MUC1 N-ter.

vector cells (Fig. 5A), MUC1 C-ter staining was intense over (Fig. 6). Moreover, purity of the nuclear preparations was the cell membrane of HCT116/MUC1(RRK ! AAA) cells demonstrated with antibodies against the cytosolic InBa,the (Fig. 5B). Similar patterns were observed for g-catenin in both membrane-associated MUC1 N-ter subunit, and the endoplasmic HCT116/vector and HCT116/MUC1(RRK ! AAA) cells (Fig. reticulum protein, calreticulin (Fig. 6). These findings collec- 5, A and B). However, in contrast to HCT116/vector cells tively indicate that the RRK motif is important for nucleolar (Fig. 5A), stimulation of HCT116/MUC1(RRK ! AAA) cells localization of MUC1 C-ter and g-catenin in the response to with HRG for 20 min was associated with redistribution of HRG stimulation. both MUC1 C-ter and g-catenin to the cytoplasm (Fig. 5B). Moreover, there was no detectable HRG-induced targeting of Confocal Microscopy of Human Breast Carcinomas MUC1 C-ter and g-catenin to the nucleolus (Fig. 5B). To define the localization of MUC1 C-ter and g-catenin in To extend these observations, the localization of MUC1 mammary tissues, confocal microscopy was first performed on C-ter and g-catenin was assessed by subcellular fractionation normal ductal epithelium. The results show localization of of control and HRG-treated cells. Immunoblot analysis of the MUC1 C-ter along the apical borders of the epithelial cells nuclear fractions demonstrated that MUC1 C-ter is detectable lining the ducts (Fig. 7A). g-Catenin colocalized with MUC1 in the nuclei of HCT116/MUC1 cells but not of HCT116/ C-ter at the apical borders and was expressed at lateral borders vector or HCT116/MUC1(RRK ! AAA) cells (Fig. 6). The of the ductal epithelium (Fig. 7A). Little (if any) MUC1 C-ter results also demonstrate that HRG increases nuclear targeting or g-catenin was detectable in the nucleus (Fig. 7A). of MUC1 C-ter in the HCT116/MUC1 cells (Fig. 6). More- Significantly, sections from ErbB2-positive breast carcinomas over, HRG treatment of HCT116/MUC1, but not HCT116/vector showed immunoflourescence staining of MUC1 C-ter and or HCT116/MUC1(RRK ! AAA), was associated with an g-catenin as discrete nuclear clusters (Fig. 7B). Sections were increase in nuclear g-catenin (Fig. 6). Equal loading of the also stained with anti-MUC1 C-ter and antinucleolin. The nuclear fractions was confirmed by immunoblotting for lamin B results demonstrate prominent colocalization of MUC1 C-ter

and nucleolin in breast carcinoma cells (Fig. 7C). Similar protective mucous barrier. The function of the C-ter, which results were obtained for g-catenin and nucleolin (Fig. 7D). The consists of an extracellular domain of f58 amino acids, a results indicate that over 50% of the breast cancer cells within transmembrane domain, and a 72-amino acid cytoplasmic tail, invasive islands exhibit nucleolar localization of MUC1 C-ter is largely unknown. The finding that MUC1-CD binds directly and g-catenin. These findings in tissues and those in cultured to h- and g-catenin suggested that the C-ter might function in cells collectively demonstrate that MUC1-CD and g-catenin are transducing signals from the cell surface to the interior of the targeted to nucleolus. cell (19). Indeed, the demonstration that MUC1-CD functions as a substrate for GSK3h (20) and c-Src (21) has indicated that Discussion the MUC1 C-ter may function in integrating signals from the Interaction of MUC1 and ErbB2 Wnt and growth factor receptor pathways. In this context, The MUC1 mucin-like glycoprotein is expressed on the activation of the EGFR is associated with tyrosine phospho- apical borders of normal mammary epithelium and at rylation of MUC1-CD and regulation of the interaction between substantially increased levels over the entire cell surface of MUC1 and h-catenin (22, 27). breast carcinoma cells (15). Significantly, overexpression of Recent studies have shown that MUC1 associates with EGFR MUC1 is associated with transformation as evidenced by and ErbB2–4 in pregnant and lactating mouse mammary glands anchorage-independent growth and tumorigenicity (16). The (27). The present work has explored the interaction between shed MUC1 N-ter is believed to function in the generation of a MUC1 and ErbB2–4 in human breast cancer cells. The results

FIGURE 3. HRG induces nucleolar colocalization of MUC1 C-ter and g-catenin. ZR-75-1 cells were grown to 60% confluence and incubated in medium with 0.1% serum for 24 h. The cells were left untreated (A)or stimulated with 20-ng/ml HRG for 20 min (B), fixed, and double stained with anti-MUC1 C-ter (green) and anti-g-catenin (red). Nuclei were stained with SYN- TOX blue. High (100) (upper panels) and low (63) (lower panels) magnifications are shown. ZR-75-1 cells were stimulated with 20-ng/ml HRG for 20 min (C) or with 10-ng/ml EGF for 20 min (D), fixed, and stained with anti-MUC1 C-ter and anti-nucleolin.

FIGURE 4. MUC1 is neces- sary for HRG-induced targeting of MUC1 C-ter and g-catenin to the nucleolus. HCT116/vector (A) and HCT116/MUC1 (B) cells were left untreated or stimulated with HRG for 20 min. The cells were assessed for reactivity with anti- MUC1 C-ter and anti-g-catenin. Nuclei were stained with SYNTOX blue.

of coimmunoprecipitation studies demonstrate the association of (19). In this regard, other studies have indicated that MUC1 and MUC1 with ErbB2–4. Significantly, treatment with HRG is E-cadherin compete for the same pool of h-catenin (20). associated with increases in MUC1-ErbB2 complexes and Moreover, negative regulation of the MUC1-h-catenin interac- colocalization of these complexes in clusters at the cell tion by GSK3h is associated with increased binding of h- membrane (Fig. 8). Members of the ErbB family form both catenin to E-cadherin (20). In this model, down-regulation of homodimers and heterodimers in response to the diverse ligands GSK3h by Wnt signaling would subvert E-cadherin function that stimulate these receptors (1, 28). The available evidence in homotypic cell-cell interactions by titrating binding of h- suggests that ErbB2 functions as a coreceptor and is a preferred catenin to MUC1. MUC1 is expressed along the apical borders heterodimerization partner among the ErbB family members of normal ductal epithelial cells that are devoid of cell-cell (1, 28). In addition, ErbB2 is overexpressed in in situ and interactions. By contrast, aberrant expression of MUC1 over invasive ductal carcinomas of the breast (28). The finding that the entire surface of carcinoma cells may contribute to loss of HRG stimulates the association between ErbB2 and MUC1 may E-cadherin function by disrupting interactions with h- and/or therefore be of importance to ErbB2 signaling, particularly in g-catenin. tumors that overexpress both of these proteins. The present results show that the MUC1-ErbB2 interaction is associated with HRG-induced binding of MUC1 and g- Interaction of MUC1 and c-Catenin catenin (Fig. 8). HRG stimulation had less of an effect on h- and g-catenin bind directly to MUC1 at a SAGNGGSSL the interaction between MUC1 and h-catenin. Conversely, motif in the CD (19). These vertebrate homologues of EGFR signaling increases binding of MUC1 and h-catenin Drosophila armadillo are found in the adherens junction (22) but has little effect on the interaction between MUC1 and where they link E-cadherin to the actin cytoskeleton through g-catenin. EGFR signaling also increases phosphorylation of a-catenin (29). The finding that complexes between MUC1 MUC1 on tyrosine in the SPY site (22), while HRG and h-org-catenin contain little (if any) a-catenin has stimulation had no apparent effect on tyrosine phosphorylation supported a function distinct from their roles with E-cadherin of MUC1-CD (data not shown). Activation of ErbB2, but not

EGFR, in growth-arrested mammary acini results in reinitiation C-ter in the nucleolus. Similar results were obtained with g- of proliferation, disruption of tight junctions, loss of polarity, catenin, supporting the likelihood that the MUC1-g-catenin and filled lumina (11). These results indicate that ErbB2 complex is targeted to the nucleolus in response to HRG activation can selectively disrupt regulation of mammary epi- stimulation. In concert with these findings, MUC1 C-ter and thelial cell proliferation and organization. Other effectors, such g-catenin are detectable in nucleoli of ErbB2-positive primary as Rac, Cdc42, and PI3K, which induce invasiveness of breast carcinomas. The observation that over 50% of the mammary epithelial cells, may cooperate with ErbB2 in breast cancer cells exhibit nucleolar colocalization of MUC1 disrupting polarized epithelia (30). One report has also C-ter and g-catenin indicate that, as found in vitro, MUC1 indicated that ErbB2 suppresses E-cadherin expression in may interact with the ErbB2 signaling pathway in primary mammary epithelial cells (31), but such regulation was not breast carcinomas. The nucleolus is a membrane-free nuclear found in other studies (11). The present findings provide subdomain in which rRNAs are transcribed and processed evidence for the involvement of ErbB2 activation and the into ribosome subunits (32). Additional functions that may be regulation of g-catenin signaling as another potential mecha- attributable to the nucleolus include the processing of other nism for increasing invasiveness. Thus, HRG-induced increases ribonucleoproteins (33, 34) and export of mRNAs and in binding of g-catenin to MUC1 could decrease the availability tRNAs (35, 36). In addition, the nucleolus may function in of g-catenin for linking E-cadherin to the actin cytoskeleton sequestering specific regulatory factors (37). For example, and thereby disrupt homotypic cell-cell signaling. Mdm2 is sequestered in the nucleolus by an ARF-dependent mechanism (38–40). Disassembly of the nucleolus during Nucleolar Localization of MUC1 C-Ter and c-Catenin cell cycle progression can in turn release sequestered factors. The present results further indicate that HRG stimulation In the nucleus, g-catenin interacts with the T-cell factor/ is associated with nuclear targeting of MUC1 C-ter and g- lymphoid enhancer factor transcription factors and functions catenin (Fig. 8). The well-circumscribed nuclear distribution as a coactivator. Like h-catenin, g-catenin can contribute to of the MUC1 C-ter signal and colocalization with anti- cell transformation by a mechanism involving transactivation of nucleolin staining supported compartmentalization of MUC1 c-Myc expression (41).

FIGURE 5. Nucleolar localization of MUC1 C-ter and g- catenin is conferred by the MUC1 RRK motif. HCT116/vector (A) and HCT116/MUC1(RRK ! AAA) (B) cells were left untreated or stimulated with HRG for 20 min. Cells were analyzed for staining with anti-MUC1 C-ter and anti-g-catenin. Morphology of the cells was visualized by bright- field microscopy.

catenin (16, 24). These findings have indicated that MUC1 may function in the import and/or stabilization of nuclear h-catenin. Importantly, the nuclear colocalization of MUC1- h-catenin complexes is found outside the nucleolus (16, 24, and unpublished data). The present results in HCT116/vector and HCT116/MUC1 cells indicate that HRG-induced nucleolar localization of g-catenin is dependent on MUC1 expression. The MUC1-CD contains a RRK motif that may function as a monopartite nuclear localization signal (46). Studies of the c-Myc nuclear localization signal (PAAKRVKLD) have demonstrated the functional role of neutral amino acids and the dipeptide LD in nuclear targeting (47). The RRK basic cluster in the MUC1-CD is also flanked by neutral amino acids and the LD dipeptide (CQCRRKNYGQLD). Importantly, mutation of the MUC1 RRK motif to AAA abrogated HRG-induced nucleolar localization of MUC1 C-ter. In addition, targeting of g-catenin to the nucleolus in response to HRG was not found in cells expressing the MUC1(RRK ! AAA) mutant. These findings provide the first evidence that MUC1 functions in nuclear signaling and that g-catenin is transported to the nucleolus by a MUC1-dependent mechanism.

Materials and Methods Cell Culture Human ZR-75-1 and MCF-7 breast carcinoma cells (American Type Culture Collection, Manassas, VA) were cultured in RPMI 1640 high-glucose medium containing 10% heat-inactivated fetal bovine serum (HI-FBS), 100-U/ml penicillin, 100-Ag/ml streptomycin, and 2-mML-glutamine. HCT116 colon carcinoma cells (American Type Culture Collection) were grown in DMEM containing 10% HI-FBS and antibiotics. Cells were maintained in medium with 0.1% HI-FBS for 24 h and stimulated with 20-ng/ml HRG or 10-ng/ ml EGF (Calbiochem-Novabiochem, San Diego, CA) at 37jC.

Cell Transfections pIRESpuro2, pIRESpuro2-MUC1, and pIRESpuro2- MUC1(RRK ! AAA) were transfected into HCT116 cells by LipofectAMINE. Stable transfectants were selected in the presence of 0.4-Ag/ml puromycin (Calbiochem-Novabiochem).

FIGURE 6. HRG-induced nuclear localization of MUC1 and g-catenin. Immunoprecipitation and Immunoblotting Nuclear fractions were analyzed by immunoblotting with the indicated Lysates were prepared from subconfluent cells as described antibodies. Whole cell lysates (WCL) were used as a positive control. (20). Equal amounts of cell lysate protein were incubated with antibody DF3 (anti-MUC1) (15), anti-ErbB2 (Santa Cruz Activation of the Wnt signaling pathway is associated with Biotechnology, Santa Cruz, CA), anti-ErbB3 (Santa Cruz accumulation of h-andg-catenin in the nucleus. The Biotechnology), anti-ErbB4 (Santa Cruz Biotechnology), or mechanisms responsible for targeting h- and g-catenin to the mouse IgG. The immune complexes were prepared as described nucleus are not clear. Neither protein has a definitive nuclear (20), separated by SDS-PAGE, and transferred to nitrocellulose localization signal; however, h-catenin is imported into the membranes. The immunoblots were probed with anti-MUC1, nucleus by binding directly to the nuclear pore machinery anti-ErbB2, anti-ErbB3, anti-ErbB4, anti-h-catenin (Zymed, San (42). Moreover, binding to T-cell factor/lymphoid enhancer Francisco, CA), or anti-g-catenin (Zymed). Reactivity was factor transcription factors is probably not responsible for detected with horseradish peroxidase-conjugated second anti- nuclear localization of h-catenin (43). The adenomatous bodies and chemiluminescence (Perkin-Elmer Corp., Boston, polyposis coli protein can function as a h-catenin chaperone MA). in nuclear export but apparently not in nuclear import (44, 45). Recent studies have demonstrated that MUC1 colocalizes with Immunoflourescence Confocal Microscopy h-catenin in the nucleus and increases nuclear levels of h- Cultured cells were washed three times in PBS (containing

Mg2+ and Ca2+), fixed with 3.7% formaldehyde in buffer A The cells were then washed three times with PBS and (PBS containing 10-AM ZnCl2) for 10 min, permeabilized with incubated with blocking buffer (PBS containing 4% 0.25% Triton X-100/3.7% formaldehyde in buffer A for 5 min, protease-free BSA and 5% normal goat serum). Incubation and postfixed with 3.7% formaldehyde in buffer A for 5 min. with anti-MUC1, anti-ErbB2, anti-MUC1 C-ter (Neomarkers,

FIGURE 7. Colocalization of MUC1 C-ter and g-catenin to the nucleolus of human breast carcinoma cells. Sections of normal mammary ductal epithelium (A) and two anti-HER2/ErbB2-positive primary invasive ductal breast carcinomas (B, upper and lower panels) were assessed for reactivity with anti-MUC1 C-ter and anti-g-catenin. Morphology was visualized at high and low (inset) power by H&E staining. Breast carcinoma cells were stained with anti-MUC1 C-ter (C) or anti- g-catenin (D) and anti-nucleolin.

References 1. Olayioye, M. A., Neve, R. M., Lane, H. A., and Hynes, N. E. The ErbB signaling network: receptor heterodimerization in development and cancer. EMBO J., 19: 3159 – 3167, 2000. 2. Carraway, K. L., III and Cantley, L. C. A neu acquaintance for erbB3 and erbB4: a role for receptor heterodimerization in growth signaling. Cell, 78: 5–8, 1994. 3. Riese D. J., II and Stern, D. F. Specificity within the EGF family/ErbB receptor family signaling network. Bioessays, 20: 41 – 48, 1998. 4. Wada, T., Qian, X. L., and Greene, M. I. Intermolecular association of the p185neu protein and EGF receptor modulates EGF receptor function. Cell, 61: 1339 – 1347, 1990. 5. Plowman, G. D., Culouscou, J. M., Whitney, G. S., Green, J. M., Carlton, G. W., Foy, L., Neubauer, M. G., and Shoyab, M. Ligand-specific activation of HER4/p180erbB4, a fourth member of the epidermal growth factor receptor family. Proc. Natl. Acad. Sci. USA, 90: 1746 – 1750, 1993. 6. Tzahar, E., Waterman, H., Chen, X., Levkowitz, G., Karunagaran, D., Lavi, S., Ratzkin, B. J., and Yarden, Y. A hierarchical network of interreceptor interactions determines signal transduction by Neu differentiation factor/neuregulin and epidermal growth factor. Mol. Cell. Biol., 16: 5276 – 5287, 1996. 7. Graus-Porta, D., Beerli, R. R., Daly, J. M., and Hynes, N. E. ErbB-2, the preferred heterodimerization partner of all ErbB receptors, is a mediator of lateral signaling. EMBO J., 16: 1647 – 1655, 1997. 8. Olayioye, M. A., Graus-Porta, D., Beerli, R. R., Rohrer, J., Gay, B., and Hynes, N. E. ErbB-1 and ErbB-2 acquire distinct signaling properties dependent upon their dimerization partner. Mol. Cell. Biol., 18: 5042 – 5051, 1998. FIGURE 8. Schematic representation of the involvement of MUC1 in 9. Ricci, A., Lanfrancone, L., Chiari, R., Belardo, G., Pertica, C., Natali, P. G., HRG-induced targeting of g-catenin to the nucleolus. Pelicci, P. G., and Segatto, O. Analysis of protein-protein interactions involved in the activation of the Shc/Grb-2 pathway by the ErbB-2 kinase. Oncogene, 11: 1519 – 1529, 1995. 10. Zrihan-Licht, S., Deng, B., Yarden, Y., McShan, G., Keydar, I., and Avraham, Fremont, CA), anti-g-catenin, and anti-nucleolin (Research H. Csk homologous kinase, a novel signaling molecule, directly associates with Diagnostics, Flanders, NJ) in blocking buffer was performed the activated ErbB-2 receptor in breast cancer cells and inhibits their proliferation. overnight at 4jC. The cells were washed with PBS, incubated J. Biol. Chem., 273: 4065 – 4072, 1998. 11. Muthuswamy, S. K., Gilman, M., and Brugge, J. S. Controlled dimerization overnight with secondary FITC- or Texas Red-conjugated goat of ErbB receptors provides evidence for differential signaling by homo- and anti-hamster or anti-mouse IgG antibodies (Jackson Immuno- heterodimers. Mol. Cell. Biol., 19: 6845 – 6857, 1999. Research Laboratories, West Grove, PA) at 4jC, washed with 12. Janda, E., Litos, G., Grunert, S., Downward, J., and Beug, H. Oncogenic PBS, washed three times with buffer B (20-mM Tris, pH 7.5, Ras/Her-2 mediate hyperproliferation of polarized epithelial cells in 3D cultures and rapid tumor growth via the PI3K pathway. Oncogene, 21: 5148 – 0.15-M NaCl), and stained with 0.2-AM of SYNTOX Blue 5159, 2002. Nuclei C solution for 2 h. After washing again with buffer B, the 13. Polakis, P. Wnt signaling and cancer. Genes Dev., 14: 1837 – 1851, 2000. cells were mounted with Slowfade solution and analyzed by 14. Lin, S. Y., Xia, W., Wang, J. C., Kwong, K. Y., Spohn, B., Wen, Y., Pestell, confocal microscopy using an inverted Zeiss LSM510 scope R. G., and Hung, M. C. h-catenin, a novel prognostic marker for breast cancer: its (Carl Zeiss, Inc., Thornwood, NY). Images were captured at roles in cyclin D1 expression and cancer progression. Proc. Natl. Acad. Sci. USA, 97: 4262 – 4266, 2000. 0.6-nm increments along the Z axis and converted to composites 15. Kufe, D., Inghirami, G., Abe, M., Hayes, D., Justi-Wheeler, H., and by LSM510 software version 3.0. Schlom, J. Differential reactivity of a novel monoclonal antibody (DF3) with human malignant versus benign breast tumors. Hybridoma, 3: 223 – 232, 1984. Flow Cytometry 16. Li, Y., Liu, D., Chen, D., Kharbanda, S., and Kufe, D. Human DF3/MUC1 Cells were incubated with anti-EGFR, anti-ErbB2, anti- carcinoma-associated protein functions as an oncogene. Oncogene, 22: 6107 – ErbB3, anti-ErbB4, or mouse IgG for 30 min, washed, 6110, 2003. incubated with goat antimouse immunoglobulin-flourescein- 17. Gendler, S., Taylor-Papadimitriou, J., Duhig, T., Rothbard, J., and Burchell, J. A. A highly immunogenic region of a human polymorphic epithelial mucin conjugated antibody (Santa Cruz Biotechnology), and fixed expressed by carcinomas is made up of tandem repeats. J. Biol. Chem., 263: in 1% formaldehyde/PBS. Reactivity was detected by immu- 12820 – 12823, 1988. noflourescence FACScan. 18. Siddiqui, J., Abe, M., Hayes, D., Shani, E., Yunis, E., and Kufe, D. Isolation and sequencing of a cDNA coding for the human DF3 breast carcinoma-associated antigen. Proc. Natl. Acad. Sci. USA, 85: 2320 – 2323, Subcellular Fractionation 1988. Preparation of nuclear fractions was performed as described 19. Yamamoto, M., Bharti, A., Li, Y., and Kufe, D. Interaction of the DF3/MUC1 breast carcinoma-associated antigen and h-catenin in cell adhesion. J. Biol. (48). Purity of the fractionations was monitored by immunoblot Chem., 272: 12492 – 12494, 1997. analysis with anti-lamin B (Oncogene Science, Cambridge, 20. Li, Y., Bharti, A., Chen, D., Gong, J., and Kufe, D. Interaction of glycogen MA), anti-calreticulin (Santa Cruz Biotechnology) and anti- synthase kinase 3h with the DF3/MUC1 carcinoma-associated antigen and h InBa (Santa Cruz Biotechnology) antibodies. -catenin. Mol. Cell. Biol., 18: 7216 – 7224, 1998. 21. Li, Y., Kuwahara, H., Ren, J., Wen, G., and Kufe, D. The c-Src tyrosine kinase regulates signaling of the human DF3/MUC1 carcinoma-associated Acknowledgments antigen with GSK3h and h-catenin. J. Biol. Chem., 276: 6061 – 6064, 2001. The authors acknowledge Kamal Chauhan for excellent technical support. D.K. 22. Li, Y., Ren, J., Yu, W.-H., Li, G., Kuwahara, H., Yin, L., Carraway, K. L., has a financial interest in ILEX. and Kufe, D. The EGF receptor regulates interaction of the human DF3/MUC1

carcinoma antigen with c-Src and h-catenin. J. Biol. Chem., 276: 35239 – 36. Bertrand, E., Houser-Scott, F., Kendall, A., Singer, R. H., and Engelke, D. R. 35242, 2001. Nucleolar localization of early tRNA processing. Genes Dev., 12: 2463 – 2468, 23. Takeichi, M. Cadherins: a molecular family important in selective cell-cell 1998. adhesion. Annu. Rev. Biochem., 59: 237 – 252, 1990. 37. Visintin, R. and Amon, A. The nucleolus: the magician’s hat for cell cycle 24. Li, Y., Chen, W., Ren, J., Yu, W., Li, Q., Yoshida, K., and Kufe, D. DF3/ tricks. Curr. Opin. Cell Biol., 12: 372 – 377, 2000. MUC1 signaling in multiple myeloma cells is regulated by interleukin-7. Cancer 38. Zhang, Y. and Xiong, Y. Mutations in human ARF exon 2 disrupt its Biol. Ther., 2: 187 – 193, 2003. nucleolar localization and impair its ability to block nuclear export of MDM2 and 25. Shimizu, H., Koyama, N., Asada, M., and Yoshimatsu, K. Aberrant p53. Mol. Cell, 3: 579 – 591, 1999. expression of integrin and erbB subunits in breast cancer cell lines. Int. J. 39. Tao, W. and Levine, A. p19ARF stabilizes p53 by blocking nucleo- Oncol., 21: 1073 – 1079, 2002. cytoplasmic shuttling of mdm2. Proc. Natl. Acad. Sci. USA, 96: 6937 – 6941, 26. Ren, J., Li, Y., and Kufe, D. Protein kinase C y regulates function of the DF3/ 1999. MUC1 carcinoma antigen in h-catenin signaling. J. Biol. Chem., 277: 17616 – 40. Lohrum, M. A., Ashcroft, M., Kubbutat, M. H., and Vousden, K. H. 17622, 2002. Identification of a cryptic nucleolar-localization signal in MDM2. Nat. Cell Biol., 27. Schroeder, J., Thompson, M., Gardner, M., and Gendler, S. Transgenic 2: 179 – 181, 2000. MUC1 interacts with epidermal growth factor receptor and correlates with 41. Kolligs, F. T., Kolligs, B., Hajra, K. M., Hu, G., Tani, M., Cho, K. R., mitogen-activated protein kinase activation in the mouse mammary gland. J. Biol. and Fearon, E. R. g-catenin is regulated by the APC tumor suppressor and Chem., 276: 13057 – 13064, 2001. its oncogenic activity is distinct from that of h-catenin. Genes Dev., 14: 28. Harari, D. and Yarden, Y. Molecular mechanisms underlying ErbB2/HER2 1319 – 1331, 2000. action in breast cancer. Oncogene, 19: 6102 – 6114, 2000. 42. Fagotto, F., Gluck, U., and Gumbiner, B. M. Nuclear localization signal- 29. Hulsken, J., Birchmeier, W., and Behrens, J. E-cadherin and APC compete independent and importin/karyopherin-independent nuclear import of h-catenin. for the interaction with h-catenin and the cytoskeleton. J. Cell Biol., 127: 2061 – Curr. Biol., 8: 181 – 190, 1998. 2069, 1994. 43. Prieve, M. G. and Waterman, M. L. Nuclear localization and formation of 30. Keely, P. J., Westwick, J. K., Whitehead, I. P., Der, C. J., and Parise, L. V. h-catenin-lymphoid enhancer factor 1 complexes are not sufficient for activation Cdc42 and Rac1 induce integrin-mediated cell motility and invasiveness through of gene expression. Mol. Cell. Biol., 19: 4503 – 4515, 1999. PI(3)K. Nature, 390: 632 – 636, 1997. 44. Henderson, B. R. Nuclear-cytoplasmic shuttling of APC regulates h-catenin 31. D’Souza, B. and Taylor-Papadimitriou, J. Overexpression of ERBB2 subcellular localization and turnover. Nat. Cell Biol., 2: 653 – 660, 2000. in human mammary epithelial cells signals inhibition of transcription of the 45. Neufeld, K. L., Zhang, F., Cullen, B. R., and White, R. L. APC-mediated E-cadherin gene. Proc. Natl. Acad. Sci. USA, 91: 7202 – 7206, 1994. downregulation of h-catenin activity involves nuclear sequestration and nuclear 32. Shaw, P. J. and Jordan, E. G. The nucleolus. Annu. Rev. Cell Dev. Biol., 11: export. EMBO Rep., 1: 519 – 523, 2000. 93 – 121, 1995. 46. Dingwell, C. and Laskey, R. A. Nuclear targeting sequences—a consensus? 33. Politz, J. C., Yarovoi, S., Kilroy, S. M., Gowda, K., Zwieb, C., and Pederson, Trends Biochem. Sci., 16: 478 – 482, 1991. T. Signal recognition particle components in the nucleolus. Proc. Natl. Acad. Sci. 47. Makkerh, J. P., Dingwall, C., and Laskey, R. A. Comparative mutagenesis of USA, 97: 55 – 60, 2000. nuclear localization signals reveals the importance of neutral and acidic amino 34. Lange, T. S. and Gerbi, S. A. Transient nucleolar localization of U6 small acids. Curr. Biol., 6: 1025 – 1027, 1996. nuclear RNA in Xenopus laevis oocytes. Mol. Biol. Cell, 11: 2419 – 2428, 2000. 48. Kharbanda, S., Saleem, A., Yuan, Z.-M., Kraeft, S., Weichselbaum, R., Chen, 35. Schneiter, R., Kadowaki, T., and Tartakoff, A. M. mRNA transport in yeast: L. B., and Kufe, D. Nuclear signaling induced by ionizing radiation involves time to reinvestigate the functions of the nucleolus. Mol. Biol. Cell, 6: 357 – 370, colocalization of the activated p56/p53lyn tyrosine kinase with p34cdc2. Cancer 1995. Res., 56: 3617 – 3621, 1996.

Downloaded from mcr.aacrjournals.org on October 3, 2021. © 2003 American Association for Cancer Research. Heregulin Targets γ-Catenin to the Nucleolus by a Mechanism Dependent on the DF3/MUC1 Oncoprotein 1 1 National Cancer Institute grant CA97098. Note: Y.L. and W.-h.Y. contributed equally to this work.

Yongqing Li, Wei-hsuan Yu, Jian Ren, et al.

Mol Cancer Res 2003;1:765-775.

Updated version Access the most recent version of this article at: http://mcr.aacrjournals.org/content/1/10/765

Cited articles This article cites 48 articles, 28 of which you can access for free at: http://mcr.aacrjournals.org/content/1/10/765.full#ref-list-1

Citing articles This article has been cited by 21 HighWire-hosted articles. Access the articles at: http://mcr.aacrjournals.org/content/1/10/765.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mcr.aacrjournals.org/content/1/10/765. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.