Csk homologous kinase, a novel signaling molecule, directly associates with the activated ErbB-2 receptor in breast cancer cells and inhibits their proliferation.

Substantial evidence exists supporting direct roles for ErbB-2/neu and Src kinase activation in breast cancer. The Csk homologous kinase (CHK) is a recently identified tyrosine kinase which, like Csk, phosphorylates the C-terminal tyrosine of Src kinases, resulting in inactivation of these enzymes. Recently, we observed that CHK is associated with the ErbB-2/neu receptor upon heregulin stimulation of breast cancer cells. Here, we report that CHK expression was observed in 70 out of 80 primary breast cancer specimens but not in normal breast tissues (0/19). Confocal microscopy analysis revealed co-localization of CHK with ErbB-2 in these primary specimens (6/6). In addition, we observed that the cytoplasmic domain of the ErbB-2/neu receptor is sufficient for its interaction with the CHKSH2 domain. Phosphopeptide inhibition of the in vitro interaction of CHKSH2 or native CHK with ErbB-2/neu, as well as site-directed mutagenesis of ErbB-2/neu, indicated that CHKSH2 binds to Tyr1253 of ErbB-2/neu. Interestingly, autophosphorylation at this site confers oncogenicity to this receptor. Moreover, CHK was able to down-regulate ErbB-2/neu-activated Src kinases. Overexpression of CHK in MCF-7 breast cancer cells markedly inhibited cell growth and proliferative response to heregulin as well as decreased colony formation in soft agar. These studies indicate that CHK binds, via its SH2 domain, to Tyr1253 of the activated ErbB-2/neu and down-regulates the ErbB-2/neu-mediated activation of Src kinases, thereby inhibiting breast cancer cell growth. These data strongly suggest that CHK is a novel negative growth regulator in human breast cancer.

Breast cancer is the second leading cause of cancer death among women in the United States and is the leading cause of death among women aged 30 to 70 (1)(2)(3). The majority of breast carcinomas appear to be sporadic and have a complex accumulation of molecular and cellular abnormalities that constitute the malignant phenotype (4 -5). In many cases, random onset of breast cancer has been correlated with increased ErbB-2/neu receptor expression and Src tyrosine kinase activity (6 -12). Substantial evidence indicates that the c-Src proto-oncogene and ErbB-2/neu play important roles in breast cancer (7,13). Src kinase activity is elevated in ErbB-2/neu (Neu) induced mammary tumors, and this elevated activity correlates with its capacity to physically associate with ErbB-2/neu (14 -15). A common pathway linking the activation mechanisms in ErbB-2/neu amplification in breast cancer is increased tyrosine kinase activity, which leads to cellular transformation (16).
Four members of the ErbB (HER) family are presently known: p170 ErbB-1 (epidermal growth factor receptor (EGF-R) 1 ), p185 ErbB-2 , p180 ErbB-3 , and p180   (3,17,18). In particular, the overexpression of the p185 ErbB-2 correlates with a poor clinical prognosis in breast cancer (9). ErbB-2/neu undergoes autophosphorylation on five tyrosine residues that are located on its non-catalytic C terminus (19,20). The autophosphorylated tyrosine residues function as docking sites for proteins that contain SH2 domains (19,20). The sequence homology between the human and rodent ErbB-2/neu is high, particularly in the C terminus (ϳ95%). The autophosphorylated tyrosine residue Tyr 1253 of rodent neu (20,21) or the human homologue Tyr 1248 of ErbB-2 (21) is the most critical residue for oncogenicity and the transforming potential of ErbB-2/neu. Although overexpression of the ErbB-2/neu gene products contributes to the aggressive behavior of various human adenocarcinomas, including breast cancer (12,22,23), the precise molecular mechanisms explaining this phenomenon are unknown.
The Csk homologous kinase (CHK), originally referred to as the megakaryocyte-associated tyrosine kinase, was identified in our laboratory (29 -31). The kinase was also independently identified as Lsk (32), Hyl (33), Ctk (34), and Batk (35). CHK shares ϳ55% identity with Csk tyrosine kinase (29,30) and consists of SH3, SH2, and tyrosine kinase domains. In contrast to Csk, which is widely expressed, CHK is highly restricted in its expression to brain and hematopoietic cells. Like Csk, CHK was also shown to phosphorylate pp60src in vitro on its Cterminal tyrosine (29,36,37). Studies in which CHK or Csk was expressed in Csk-deficient mouse embryo fibroblasts showed that murine p52 CHK was comparable to Csk in its ability to reduce the activity of the Src kinases Fyn and pp60src (37). Recently, we found a specific interaction between CHK and the HRG-activated ErbB-2/neu receptor in MCF-7 and T47D breast cancer cell lines. The CHK SH2 and CHK SH3-SH2 domains precipitated the tyrosine-phosphorylated ErbB-2/neu receptor upon stimulation with HRG. In vivo association of the tyrosinephosphorylated ErbB-2/neu with CHK was also observed in co-immunoprecipitation studies using anti-CHK antibodies. This association of CHK with ErbB-2/neu occurred via the CHK-SH2 domain and appeared to be receptor-specific (e.g. ErbB-2/neu) and ligand-specific (e.g. HRG) (38).
In this report, we characterize the binding of the SH2 domain of CHK to specific tyrosine-phosphorylated sites on ErbB-2/neu. Our results indicate that the CHK SH2 domain binds directly to ErbB-2/neu at the Tyr 1253 site, which has been shown to confer oncogenicity to this receptor. Furthermore, we demonstrate that CHK down-regulates ErbB-2/neu-activated Src kinases and is associated with anti-proliferative activities. Taken together, these results suggest a novel mechanism by which CHK alters ErbB-2/neu and Src kinases in breast cancer and thus may modulate growth of breast cancer cells.

Materials
Recombinant human HRG (␥HRG-␤1, 177-244), rabbit polyclonal anti-ErbB-2/neu antibodies, and 3E8 monoclonal anti-ErbB-2/neu antibody were generously provided by Dr. Mark A. Sliwkowski, Genentech (San Francisco, CA). Ab-4 anti-neu oncogene antibodies were purchased from Oncogene Science. Monoclonal anti-phosphotyrosine antibody (4G10) was kindly provided by Dr. Brian J. Druker (Division of Hematology and Medical Oncology, Oregon Health Sciences University, Portland). Monoclonal and polyclonal antibodies for EGF-R and polyclonal anti-CHK antibodies were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-GST monoclonal antibody was purchased from Pharmacia Biotech Inc. Rhodamine-conjugated anti-rabbit IgG antibody was obtained from Sigma. EGF was purchased from Collaborative Biomedical Products (Bedford, MA). Tyrosine-phosphorylated and nonphosphorylated synthetic peptides were obtained from the Dana Farber Cancer Institute Molecular Biology Core Facility (Boston, MA). Peptides were analyzed for purity by high pressure liquid chromatography, mass spectroscopy, and amino acid analysis. T 7 polymerase vaccinia recombinant virus, the vaccinia wild-type virus, and the PTM-1 vector were generously provided by Dr. Bernard Moss (National Institutes of Health, Bethesda). CHK SH2 GST fusion protein was prepared as described previously (31,38). The primers for the polymerase chain reaction were synthesized by an automated DNA synthesizer (Applied Biosystem model 394). Reagents for electrophoresis were obtained from Bio-Rad. Enhanced chemiluminescence (ECL) reagents were purchased from Amersham Corp. All other reagents were purchased from Sigma.
Immunohistochemistry and Confocal Laser Scanning Microscopy Analysis-Paraffin blocks from human primary breast carcinomas were serially sectioned at 5-6 m. Tissue samples were stained with hematoxylin and eosin to identify "normal" as well as tumorous areas. Parallel sections were deparaffinized, washed in PBS, blocked (1% normal donkey serum and 0.1% bovine serum albumin in PBS) for 1 h, and double-stained with rabbit anti-human CHK polyclonal antibody and mouse monoclonal antibody against ErbB-2 (Oncogene Science). Following washing, the cells were incubated with fluorescein isothiocyanate-conjugated donkey anti-rabbit (1:50) and rhodamine-conjugated donkey anti-mouse antibodies (1:100) (The Jackson Laboratory, Bar Harbor, ME). Stained cells were analyzed using a Zeiss confocal laser scanning microscope. The Zeiss (Oberkochen, Germany) confocal laser scanning microscope 410 is equipped with a 25-megawatt kryptonargon laser and a 10-megawatt HeNe laser (488, 543, and 633 maximum lines); co-localization analysis was performed on simultaneously labeled samples using the Zeiss co-localization procedure. This procedure results in a graphic representation of the distribution of green (y axis) and red (x axis) fluorescence for each pixel. Images were printed using a Codonics NP 1600 printer (Codonics, Middleburg Heights, OH).
For confocal microscopy, CHK-Flag-MCF-7 cells were fixed and prepared as described previously (39). Fixed and permeabilized cells were incubated with CHK antibodies (1:100 dilution) (Santa Cruz, CA) for 1 h. Cells were washed in PBS, incubated with rhodamine-conjugated (rabbit IgG) antibodies for 1 h (Sigma), and examined using a Sarastro 2000 confocal laser scanning microscope (confocal laser scanning microscope, Molecular Dynamics, Sunnyvale, CA).

Expression Vectors and Transfection
NEC and TEC chimeric plasmids containing the extracellular domain of EGF-R and the transmembrane domain of the ErbB-2/neu oncogene were previously described (19,26). The P 1 and Y1253F mutated constructs as well as the neu* construct were previously described (19 -20). Transfections of NIH3T3 and COS cells were performed using LipofectAMINE reagents (Life Technologies, Inc.) according to the manufacturer's protocol.

Precipitation with GST Fusion Proteins and Immunoprecipitation
Approximately 5 ϫ 10 6 cells/plate TEC or NEC stable transfectants were starved for 4 h and then stimulated with 100 ng/ml EGF for 5 min at room temperature and analyzed as described previously (38). Lysates were precleared and then precipitated with GST fusion proteins coupled to glutathione-Sepharose beads (40). For the immunoprecipitations, polyclonal CHK antibodies (10 l), monoclonal EGF-R (10 g/ml) (Santa Cruz, CA), or ErbB-2/neu (Ab-4) (Oncogene Science) (10 g/ml) were used. The samples were analyzed by 7 or 10% SDS-PAGE as described (38). Bound proteins were immunoblotted with anti-phosphotyrosine (4G10), polyclonal EGF-R antibodies, or polyclonal ErbB-2/neu antibodies. The blots were developed using the ECL system (Amersham Corp.). Blots were stripped for 30 min at 55°C in stripping buffer according to the manufacturer's protocol (Amersham Corp.).

Peptide Inhibition of the CHK SH2 GST Fusion Protein-ErbB-2/neu Interaction
Cell lysates from COS cells transiently transfected with the neu*, P 1 , or Y1253F plasmids were added to the CHK SH2 GST fusion proteins (10 g/incubation) that were preincubated with synthetic peptides as described previously (40). Lysates were incubated for 1 h at 4°C and then precipitated by the addition of glutathione 4B for 30 min at 4°C. The washed precipitates were separated on 7.5% SDS-PAGE.

Association of CHK SH2 GST or Native CHK with Peptide Beads
Phosphorylated and non-phosphorylated peptides were linked to Affi-Gel 15 as described previously (40). Peptide beads (15-l bead volume) were incubated with 10 g of CHK SH2 for 1.5 h at 4°C. The washed samples were separated on 10% SDS-PAGE. To test the binding of native CHK to the peptide beads, 15 l of beads were incubated with 1.5 ml of MCF-7, CHK-Flag-MCF-7, or neo-1 cell extracts for 1.5 h at 4°C. After three washes, the precipitates were subjected to SDS-PAGE and Western blotting with anti-CHK antibodies.

T 7 Polymerase Vaccinia Expression System
The CHK cDNA (1.6 kb) was cloned into EcoRI sites in the PTM-1 vector. The NcoI site that includes the ATG initiation codon was added to the CHK sequence by polymerase chain reaction using a 5Ј sense primer from nucleotides 269 to 296 and a 3Ј antisense primer from nucleotides 510 to 481 (30). The polymerase chain reaction product was introduced to the vector at the NcoI-BstEII sites (New England Biolabs, Beverly, MA). The construct was sequenced, and its expression was examined by transfection assay followed by Western blot analysis.
Generation of recombinant CHK vaccinia virus was performed using the method described previously (41). The CHK vaccinia recombinant virus was used along with the T 7 vaccinia recombinant virus for the co-infection assay, as follows: approximately 5 ϫ 10 5 cells/plate were seeded. One day later, the cells were infected with trypsinized recombinant viruses 10 ϫ multiplicity of infection each, for 1-2 h in 2.5% fetal calf serum/DMEM at 37°C. Next, 5 ml of 10% DMEM were added, and the plates were incubated overnight. The infected cells were harvested the day after infection. Infections of transiently transfected COS cells were performed 2 days after transfections, as described above.

Transformation Assays
Transformation of the cells was assessed by their ability to demonstrate anchorage-independent growth (42). MCF-7, neo-1, or CHK-Flag cells (1 ϫ 10 5 in 6-well dishes) were grown in medium containing 0.4% agar. After 2 weeks of growth, the colonies were visualized by staining with 0.33% iodonitrotetrazolium violet. All assays were performed in duplicate.
In Vitro Kinase Assay MCF-7 cells were co-infected with the T 7 polymerase vaccinia virus and CHK recombinant virus as described above. One day post-infection, the cells were starved for 4 h in serum-free media and then stimulated with 10 nM HRG as described previously (38). Cell lysates were immunoprecipitated using anti-Src antibodies, and the washed precipitated proteins were submitted to kinase assay or analyzed by SDS-PAGE. In vitro kinase assays were performed by incubating washed Src immunoprecipitates with lysates from CHK-expressing or non-expressing cells in kinase buffer for 20 min at room temperature as described (14 -15, 29). Proteins were separated by 7.5% SDS-PAGE under reducing conditions. The gels were dried and subjected to autoradiography.

Expression and Localization of CHK in Breast Cancer-Im-
munohistochemical studies demonstrated CHK protein expression in 70 of 80 breast adenocarcinoma specimens (stage I, 32/41; stage II, 34/35; stage III, 4/4) but not in normal or benign breast tissues (normal breast and fibroadenoma, 0/19). Confocal microscopic imaging in primary breast tumors 6/6 ( Fig. 1A) and activated T47D cells (not shown) immunostained for ErbB-2 and CHK indicated a similar pattern of distribution and co-localization of CHK and ErbB-2 (Fig. 1A). Our previous study indicated that upon HRG stimulation, CHK associates via its SH2 domain with the ErbB-2/neu receptor (38). To examine further this interaction, confocal microscopy studies were performed in MCF-7 cells stably transfected with the CHK-Flag construct. We observed that CHK was localized to the cytosolic compartment (Fig. 1B). However, upon HRG stimulation, CHK was redistributed to the subplasma membrane, away from the perinuclear regions (Fig. 1B). These results, along with our previous results, suggest that upon ligand stimulation, CHK is translocated from the cytosol to the membrane vicinity and associates with ErbB-2/neu receptors.
Association of the CHK SH2 Domain with NEC and TEC EGF-ErbB-2/neu Chimeric Receptors-To verify that the association of CHK with ErbB-2/neu is mediated exclusively by ErbB-2/neu and not by other members of the ErbB family, we used chimeric molecules composed of the extracellular domain of the EGF-R and the transmembrane and cytoplasmic domains of ErbB-2/ neu. Two different EGF-R/ErbB-2/neu chimeric plasmids were used: the TEC construct which contains a point mutation (Val 664 3 Glu) in the ErbB-2/neu transmembrane domain causing constitutive activation of the molecule, and the NEC construct that contains the wild-type sequence of the ErbB-2/ neu cytoplasmic domain and therefore can be stimulated by EGF (19). NIH3T3 cells were stably transfected with either the NEC or TEC construct. The transfected cells were analyzed for construct expression by immunoprecipitation followed by Western blot analysis using anti-EGF-R antibodies. Upon EGF stimulation, CHK SH2 was associated with NEC (Fig. 2, B-II), whereas its association with TEC was constitutive and not dependent on EGF stimulation (Fig. 2, A-II). Therefore, the cytoplasmic domain of the ErbB-2/neu receptor appears to be sufficient for its interaction with the CHK SH2 domain.
Inhibition of the CHK SH2 -ErbB-2/neu Interaction by Tyrosine-phosphorylated Peptides-To identify the binding site of CHK SH2 within the ErbB-2/neu receptor, we synthesized a series of tyrosine-phosphorylated peptides derived from the five autophosphorylated tyrosine residue sites of the cytoplasmic domain of the ErbB-2/neu receptor (Tyr 1028 , Tyr 1144 , Tyr 1286/7 and Tyr 1253 ). These tyrosine-phosphorylated peptides were used to inhibit the interaction between the CHK SH2 domain and the activated ErbB-2/neu receptor. COS cells were transiently transfected with the transformed ErbB-2/neu (neu*) plasmid that codes for the constitutively phosphorylated receptor (point mutation Val 664 3 Glu), as described previously (19 -20). Complexes of ErbB-2/neu and CHK SH2 were indicated by the presence of ErbB-2/neu in the washed CHK SH2 GST fusion protein precipitates (Fig. 3, A and B). Of the four peptides, peptide P 1 (ENPEY*LGLDV, where * indicates the phosphorylated tyrosine residue) most significantly inhibited complex formation (Fig. 3A). We also found inhibition by peptide P 5 (AEEY*LVPQQ). To compare the relative abilities of the P 1 and P 5 peptides to inhibit the CHK SH2 -ErbB-2/neu interaction, various concentrations of peptides from 5 to 100 M were tested.
The results indicate that inhibition by the P 1 peptide is much more significant throughout all the tested concentrations as compared with other peptides (Fig. 3C) and suggest that binding of CHK is primarily at the P 1 (ENPEY*LGLDV) site of the ErbB-2/neu receptor. Moreover, inhibition by the P 1 peptide was phosphorylation-dependent, since the P 1 non-phosphorylated peptide had no inhibitory effect (data not shown).
Binding of the CHK SH2 Domain and Native CHK to Immobilized Peptides-To test further the binding of CHK to the phosphorylated P 1 site, we linked the tyrosine-phosphorylated P 1 peptide or the non-phosphorylated P 1 peptide to Affi-Gel 15 beads, and the association of either CHK SH2 GST fusion protein or native CHK to the beads was analyzed. As shown in Fig.  4A, CHK SH2 GST was associated in a phosphotyrosine-dependent manner to the phosphorylated P 1 peptide. Similar specificity was observed when we tested the association of native CHK to the peptide beads. The (P 1 *) peptide was able to associate with native CHK from extracts of CHK-Flag-MCF-7-transfected cells (Fig. 4B). No binding was observed by the MCF-7 or neo-1 lysates, which do not express CHK. Similar results were demonstrated using CHK obtained from the vaccinia expression system (Fig. 4C). This specificity was in agreement with the peptide inhibition experiments, indicating a direct association between ErbB-2/neu and CHK SH2 mediated by the P 1 tyrosine-phosphorylated site.
CHK Is Associated with the P 1 Phosphotyrosine as Shown by Site-directed Mutagenesis-To confirm the significance of the phosphorylated peptide studies, we analyzed whether the CHK SH2 GST fusion protein could bind to the ErbB-2/neu receptor bearing the P 1 -phosphorylated site alone (Tyr 1253 ). Two constructs of the activated ErbB-2/neu receptor (neu*-mutated Val 664 3 Glu) (19,20) were used to transfect the COS cells as

FIG. 1. Confocal micrographs of breast tumors and CHK-Flag-MCF-7 cells.
A, co-localization analysis of ErbB-2 and CHK in primary breast tumors. Confocal micrographs of breast cancer tissues immunolabeled with ErbB-2 staining (dilution 1:500), followed by rhodamine-labeled secondary antibody (red), CHK staining (dilution 1:100), followed by fluorescein isothiocyanate-labeled secondary antibody (green) are shown. For overlay images, yellow represents co-localized staining. Using co-localization analysis, the co-localized pixels were depicted (blue) and overlaid on the red-green image. A graphic representation of the distribution of green (y axis) and red (x axis) fluorescence for each pixel is shown. The number of pixels is color coded where dark blue indicates one to few, yellow is high, and dark red is the highest co-localization. Similar results were obtained with an additional five breast cancer specimens. B, CHK-Flag-MCF-7 transfected cells. CHK-Flag-MCF-7 cells were stimulated with HRG (10 nM follows: 1) the P 1 construct that contains the extracellular and transmembrane domains of ErbB-2/neu and the P 1 binding site, and 2) the Y1253F construct that contains the full sequence of the constitutively activated ErbB-2/neu, including a point mutation at the P 1 site (Tyr 1253 3 Phe). Both ErbB-2/neu constructs were tyrosine-phosphorylated (Fig. 5A). Cell extracts from the same experiment were incubated with the CHK SH2 GST fusion protein, separated by SDS-PAGE, and analyzed by Western blotting using ErbB-2 antibody. The CH-K SH2 GST precipitated the ErbB-2/neu in the COS cells transfected with the P 1 construct (Fig. 5B), whereas no association was found with the ErbB-2/neu carrying the point mutation on the P 1 site (Y1253F). Therefore, the CHK SH2 exclusively bound to the P 1 -Tyr 1253 site of ErbB-2/neu.
To examine for the in vivo association of the CHK molecule with the P 1 -Tyr 1253 site, we used the T 7 polymerase vaccinia system to overexpress CHK. COS cells were first transiently transfected with the P 1 plasmid, and then 2 days post-transfection the cells were co-infected with the T 7 polymerase virus alone or with the T 7 polymerase and CHK recombinant viruses. CHK was expressed in cells infected with the CHK recombinant virus but not in cells that were infected with the T 7 polymerase virus alone (Fig. 5C). Lysates from the same experiment were tested for the expression of the P 1 -ErbB-2/neu molecule using anti-ErbB-2/neu antibody (Fig. 5D).
To demonstrate the in vivo association between P 1 -ErbB-2/ neu and CHK, the same cell extracts were immunoprecipitated using CHK antibody and then immunoblotted with ErbB-2/neu antibody. P 1 -ErbB-2/neu was present only in precipitates from the CHK-expressing cells and not in those of the T 7 -infected cells (Fig. 5D). Taken together, these results indicate an in vivo association between CHK and the P 1 binding site (Tyr 1253 ) of ErbB-2/neu.

Investigation of ErbB-2/neu Signaling Mechanisms by Vaccinia-driven Overexpression of CHK in MCF-7
Cells-To elucidate the involvement of CHK in the regulation of Src kinase activity, we again overexpressed CHK using the T 7 polymerase-vaccinia expression system. MCF-7 cells were co-infected either with a CHK vaccinia recombinant virus (CHK-vacc) and T 7 polymerase virus (T 7 ) or with the T 7 virus alone as a control. One day after co-infection, the cells were starved for 4 h and then stimulated with heregulin (10 nM). Cell extracts were immunoprecipitated using Src antibody, and the enzymatic activity of Src was determined using poly(Glu/Tyr) (4:1) as a substrate. In CHK-expressing cells, poly(Glu/Tyr) phosphorylation was decreased about 4-fold compared with the control T 7 -infected cells upon stimulation with HRG (Fig. 6). Therefore, CHK expression resulted in a significant reduction in Src at room temperature) for 0 and 15 min (a and b, respectively). After stimulation, the cells were washed with PBS and fixed with 2% paraformaldehyde. Cells were immunostained for 1 h with anti-CHK antibodies (1:100), followed by rhodamine-labeled secondary antibody. c shows the secondary antibody alone, and d shows staining of untransfected MCF-7 cells with anti-CHK antibodies as controls. The slides were visualized by confocal microscopy. Extracts from neu* transiently transfected COS cells were added to CHK SH2 which was preincubated individually with each of four different ErbB-2/neu phosphopeptides as indicated below. The final concentration of peptide was 100 M for each incubation. The samples were washed and separated on 7% SDS-PAGE. A shows immunoblotting with ErbB-2/neu polyclonal antibody. B, immunoblotting with 4G10 anti-phosphotyrosine antibody. The peptides are as follows: P 1 , ENPEY*LGLDV; P 2,3 , DNLY*Y*WDQNS; P 4 , QPEY*VNQSE; P 5 , AEEY*LVPQQ; where Y* refers to the phosphotyrosine residue within each peptide as listed above. C, concentration dependence of inhibition of CHK SH2 ErbB-2/neu interaction by ErbB-2/neu phosphopeptides. Extracts from neu* transiently transfected COS cells were added to CHK SH2 GST fusion protein which was preincubated with increasing concentrations (0 -100 M) of different ErbB-2/neu phosphopeptides, P 1 and P 5 peptides. The samples were separated on 7% SDS-PAGE and immunoblotted with 4G10 anti-phosphotyrosine antibody.
kinase activity upon HRG stimulation, indicating that CHK may regulate the ErbB-2/neu-activated Src kinases.
Suppression of Cell Growth by CHK-To elucidate whether CHK might affect the growth of MCF-7 cells, the proliferation rate of CHK-MCF-7 clones was analyzed using the 3-(4,5-dimethyl thiazol-2-yl)-2,5-diphenyl tetrazolium bromide method (Fig. 7). The proliferation rate of the CHK-expressing cells (CHK-Flag-MCF-7) was significantly reduced (p Ͻ 0.001) com-pared with the control untransfected MCF-7 cells or cells transfected with vector alone (neo-1) (Fig. 7A). Furthermore, when the cells expressing CHK were stimulated with HRG, we observed a significant reduction in their proliferative response to HRG (Fig. 7A). These data suggest that CHK can reduce the proliferative activity of breast cancer cells and cause desensitization to the growth-promoting effects of HRG.
The anti-transforming potential of CHK was evaluated in MCF-7 clones by examining the ability of CHK-transfected cells to escape contact inhibition when grown on tissue culture plastic and to support anchorage-independent growth in soft agar. MCF-7 cells, as well as neo-1 cells, grew in culture plates to a higher density, displayed a tendency to pile up (data not shown), and acquired the ability to form colonies in soft agar (Fig. 7B). CHK-expressing cells did not show anchorage-independent growth, and the number of colonies formed in soft agar decreased approximately 4-fold compared with control MCF-7 or neo-1 cells (Fig. 7B). Taken together, these results demonstrate that CHK expression is associated with anti-proliferative activity and can reduce the transformation ability of breast cancer cells. FIG. 4. Association of CHK to phosphopeptides linked to beads. A, CHK SH2 GST fusion protein (SH2) or the GST proteins alone (GST) were incubated with phosphorylated (P 1 *) or non-phosphorylated (P 1 -non) P 1 peptides linked to Affi-Gel 15 beads. Washed precipitates were separated on 10% SDS-PAGE and immunoblotted with CHK antibodies (P 1 *, ENPEY*LGLDV). B, extracts from CHK-Flag-MCF-7 cells (CHK-Flag), control MCF-7, or MCF-7 cells transfected with vector alone (neo-1) were incubated with phosphorylated (P 1 *) or non-phosphorylated (P 1 -non) P 1 peptides linked to Affi-Gel 15 beads. Washed precipitates were separated on 10% SDS-PAGE and immunoblotted with CHK antibodies. C, extracts from MCF-7 cells co-infected with CHK vaccinia recombinant virus (CHK-vacc) and T 7 polymerase-vaccinia recombinants or controls infected with the T 7 -recombinant virus alone (T 7 ) were incubated with phosphorylated (P 1 *) P 1 peptide linked to Affi-Gel 15 beads. The precipitates were separated on 10% SDS-PAGE and immunoblotted with anti-CHK antibodies.
FIG. 5. Association in vitro and in vivo of CHK with the P 1 binding site (Tyr 1253 ) of ErbB-2/neu receptor. COS cells were transiently transfected either with plasmid coding for the phosphorylated P 1 site (P 1 ) or with a Y1253F construct that codes for the four wild-type autophosphorylated sites of ErbB-2/neu with a point mutation (Tyr 1253 3 Phe) in the P 1 site (Y1253F). The extracts were divided into two portions. One portion (A) was immunoprecipitated using Ab-4 (anti-neu monoclonal antibody), separated on 7% SDS-PAGE, and blotted either with polyclonal anti-ErbB-2 antibody or with 4G10 anti-phosphotyrosine antibody. The second portion (B) was precipitated with CHK SH2 GST fusion protein. The washed precipitates were separated on 10% SDS-PAGE, and the blot was immunoblotted with anti-ErbB-2 antibody. COS cells were transiently transfected with the P 1 ErbB-2/neu* binding site construct (C). Two days post-transfection, the cells were either co-infected with CHK-recombinant vaccinia virus (CHK-vacc) or with the T 7 recombinant virus alone (T 7 ). Extracts were prepared and analyzed for expression of CHK in the infected cells by Western blot analysis using anti-CHK antibody (C). Extracts were immunoprecipitated with anti-CHK antibody or with Ab-4 (anti-neu monoclonal antibody). The precipitates were separated on 10% SDS-PAGE and immunoblotted with polyclonal anti-ErbB-2 antibody (D).
FIG. 6. Immune complex kinase reactions. MCF-7 cells (5 ϫ 10 6 ) were infected with either wild-type vaccinia (T 7 ) or CHK recombinant virus (CHK). One day after infection, cells were stimulated with heregulin for 5 min and lysed. Lysates were immunoprecipitated with a control MoAb or Src MoAb. Anti-Src immunoprecipitates were analyzed by immune complex kinase assay. The activity of pp60src was determined in an immune complex kinase reaction, using poly(Glu/Tyr) as a model substrate.

DISCUSSION
In the present study, we demonstrated direct binding of CHK to the cytoplasmic domain of activated ErbB-2/neu using the chimeric EGFR-ErbB-2/neu receptor constructs, NEC and TEC (Fig. 2). Moreover, we identified the binding site of CHK SH2 to the activated ErbB-2/neu receptor as the P 1 -Tyr 1253 . This site is known to confer oncogenicity and transforming abilities to ErbB-2/neu. To demonstrate that the interaction of CHK SH2 at the Tyr 1253 site is direct and not mediated by other proteins, we showed that native CHK from cell extracts as well as a bacterial CHK SH2 GST fusion protein could bind to phosphorylated P 1 -Tyr 1253 peptide linked to beads. In addition, we demonstrated in vivo binding of intact CHK to an ErbB-2 molecule bearing the P 1 site alone. Using vaccinia virus-driven CHK protein expression, we observed that CHK was able to downregulate the ErbB-2/neu-activated Src kinases. Furthermore, overexpression of CHK in MCF-7 breast cancer cells markedly inhibited cell growth and reduced their proliferative response to HRG stimulation. These results indicate for the first time that CHK associates with the Tyr 1253 site of the ErbB-2/neu receptor and lead us to hypothesize that CHK could antagonize growth promoting signaling mediated by ErbB-2/neu-activa-tion of downstream Src kinases.
Overexpression and amplification of ErbB-2/neu play a major role in the genesis of mammary tumors (9), although the molecular mechanism of this process is poorly understood. Once activated, receptor tyrosine kinases undergo autophosphorylation and recruit SH2 domain-containing proteins to their phosphorylated tyrosine residues. The Tyr 1253 of ErbB-2/ neu is the most critical residue for oncogenicity of this receptor (19,20), and complete inhibition of the transforming action of human ErbB-2/neu is seen with the analogous Y1248F mutant (21). This site (ENPEY*LGLDV) is completely conserved between human and rodent ErbB-2/neu (20,21). Recently, it was reported that autophosphorylation of tyrosine residues within ErbB-2/neu is involved in both the negative and positive regulation of ErbB-2/neu-mediated transformation (43). The identification of CHK as a signaling molecule that directly interacts with the P 1 -Tyr 1253 site implies that CHK might modulate ErbB-2/neu activity and oncogenicity, since this C-terminal sequence enables coupling of ErbB-2/neu signaling to downstream pathways that include Ras, mitogen-activated protein kinase, and transactivation of c-Jun (19). The results presented here further suggest that CHK exerts anti-mitogenic activity, since its overexpression in MCF-7 cells decreased their proliferative response to HRG stimulation. In addition, MCF-7 cells overexpressing CHK showed a reduction in colony size and number when assayed for growth in soft agar (Fig. 7).
Activation of not only ErbB-2/neu but downstream Src tyrosine kinases plays an important role in mammary tumorigenesis (44 -47). Human breast cancer specimens possess more than a 4-fold increase in c-Src tyrosine kinase activity when compared with normal breast tissues (16), and neu-induced mouse mammary tumors possess a 6 -7-fold increase in c-Src tyrosine kinase activity as compared with their normal counterparts (14,15). Transgenic mice expressing the middle T antigen in the mammary epithelium (44) have 4 -5-fold increases in the tyrosine kinase activities of both c-Src and the Src family kinase c-Yes (45). Furthermore, expression in transgenic mice of a constitutively active form of c-Src (Y527F) under the control of the murine mammary tumor virus promoter/ enhancer results in epithelial hyperplasia and mammary tumors (23). The increased Src kinase activity observed in neuinduced tumors also results from the ability of the Src SH2 domain to directly interact with ErbB-2/neu in a phosphotyrosinedependent manner. Both Src and Yes were found to bind to the same site on ErbB-2/neu (14 -15, 25, 43). However, the site of interaction between Src family members and ErbB-2/neu is not yet known. In the present study, we demonstrate that upon HRG stimulation of breast cancer cells, there is extensive elevation of Src kinase activity and that expression of CHK completely inhibits this activity. CHK down-regulates Src kinase activity probably by phosphorylation of its C-terminal regulatory tyrosine. The results presented here further suggest that CHK can exert anti-proliferative activity. Overexpression of CHK in MCF-7 cells resulted in decreased cell proliferation and a significant decline in the number and size of the colonies formed in soft agar (Fig. 7).
The observations that activation of Src in ErbB-2/neu-expressing cells occurs through tyrosine-phosphorylated ErbB-2/ neu and that CHK can down-regulate ErbB-2/neu-activated Src kinases have important implications in understanding the biological properties of the molecules. It is conceivable that CHK association to ErbB-2/neu upon HRG stimulation is followed by the down-regulation of ErbB-2/neu-activated Src kinases. We suggest a model for this regulation of ErbB-2/neuactivated Src kinases by CHK based on our data (Fig. 8). In this model, stimulation of ErbB-2/neu by HRG, presumably via FIG. 7. Suppression of cell growth by CHK. A, mitogenic effects of heregulin on MCF-7 clones. MCF-7, neo-1, and CHK-Flag-MCF-7 cells were grown in 1% fetal calf serum untreated (a) (Ϫ) and treated (b) (ϩ) with heregulin (10 nM) for the indicated times. Each data point is the mean of three readings within the same experiment. This is a representative experiment out of three. B, formation of colonies in soft agar by MCF-7 clones. Cells were seeded in soft agar at a density of 2.5 ϫ 10 4 cells/dish and allowed to grow for 2 weeks before counting viable colonies (3 cells or more per colony). Each experiment was performed in triplicate.
ligand-driven heterodimerization and cross-phosphorylation, leads to autophosphorylation of ErbB-2/neu and an initial association of Src kinase to one of the phosphorylated tyrosine sites, thereby resulting in the activation of the Src kinase (Fig.  8A). Association of CHK with P 1 -Tyr 1253 on activated ErbB-2/ neu is likely to facilitate phosphorylation of Src at its C-terminal phosphotyrosine (Fig. 8B). These interactions lead to the inactivation and dissociation of the Src kinase due to the selfassociation of the C-terminal phosphotyrosine to its own SH2 domain (Fig. 8C). Further studies will address this model, particularly whether Src kinase bound to ErbB-2/neu is the primary substrate for CHK. Investigation of the role of CHK in breast cancer may contribute to an understanding of the mechanisms of oncogenic signal transduction and provide a basis for utilizing this tyrosine kinase to oppose the malignant process.