Identification of Protein-tyrosine Phosphatase 1B as the Major Tyrosine Phosphatase Activity Capable of Dephosphorylating and Activating c-Src in Several Human Breast Cancer Cell Lines*

c-Src tyrosine kinase activity is elevated in several types of human cancer, and this has been attributed to elevated c-Src expression levels, increased c-Src specific activity, and activating mutations in c-Src. We have found a number of human breast cancer cell lines with elevated c-Src specific activity that also possess elevated phosphatase activity directed against the carboxyl-terminal negative regulatory domain of Src family kinases. To identify this phosphatase, cell extracts from MDA-MB-435S cells were chromatographed and the fractions were assayed for phosphatase activity. Four peaks of phosphatase activity directed against the nonspecific substrate poly(Glu/Tyr) were detected. One peak also dephosphorylated a peptide modeled against the c-Src carboxyl-terminal negative regulatory domain and intact human c-Src. Immunoblotting and immunodepletion experiments identified the phosphatase as protein-tyrosine phosphatase 1B (PTP1B). Examination of several human breast cancer cell lines with increased c-Src activity showed elevated levels of PTP1B protein relative to normal control breast cells.In vitro c-Src reactivation experiments confirmed the ability of PTP1B to dephosphorylate and activate c-Src. In vivo overexpression of PTP1B in 293 cells caused a 2-fold increase of endogenous c-Src kinase activity. Our findings indicate that PTP1B is the primary protein-tyrosine phosphatase capable of dephosphorylating c-Src in several human breast cancer cell lines and suggests a regulatory role for PTP1B in the control of c-Src kinase activity.

c-Src is a membrane-associated non-receptor tyrosine kinase that plays an important role in relaying signals received by receptors on the cell surface to the cell interior (for a recent review, see Refs. 1 and 67). When activated, c-Src is capable of phosphorylating a variety of intracellular substrates, with consequent effects on downstream signaling pathways affecting cell division, cell differentiation, and cell mobility.
Because of its potent ability to influence cellular functions, c-Src is normally maintained in an inactive state. Within the cell, CSK 1 and CSK-homologous kinase phosphorylates Tyr-530 within the carboxyl-terminal tail of human c-Src (Tyr-530 of human c-Src is equivalent to Tyr-527 of chicken c-Src) (2,3). This results in an intramolecular interaction between this phosphorylated residue and the c-Src SH2 domain (4 -6). This interaction is largely responsible for keeping c-Src in an inactive state and is augmented by SH3 binding interactions with the central "linker" and kinase domains of c-Src (7).
Activation of c-Src normally occurs transiently during cellular events such as growth factor receptor activation and during mitosis (8 -10). c-Src becomes activated as a result of molecular processes that interrupt its negative regulatory interactions. These processes include binding of the c-Src SH2 or SH3 domains to a binding protein such as a tyrosine phosphorylated growth factor receptor (9) or a protein containing proline-rich sequences (11,12), phosphorylation of residues within c-Src (8,10,(13)(14)(15), and dephosphorylation of the carboxyl-terminal negative regulatory site (16). Activation of c-Src can also occur as a result of mutation, and naturally occurring (i.e. v-Src) and laboratory-derived mutants have been identified (17)(18)(19). These mutants have typically been identified by their ability to cause cellular transformation, and it appears that constitutive activation of the tyrosine kinase activity of c-Src is a potent transforming signal.
Because c-Src has the potential to transform cells when activated, investigators have examined whether c-Src might contribute to the transformed phenotype of human cancers. Elevations in c-Src kinase activity have been noted in a variety of human cancers, particularly in colon and breast cancers (20 -29). Several mechanisms have been implicated in causing these elevations in c-Src kinase activity in tumors, including elevations in Src protein levels, elevations in Src specific activity, and mutations within c-Src that disrupt normal regulation (20, 24 -29).
We have recently reported elevations in Src kinase activity in 30 -40% of human breast cancer cell lines tested (29). These increases in activity were attributable, in most cases, to increases in both Src specific activity and protein levels. Interestingly, the breast cancer cell lines exhibiting elevated c-Src activity were also found to possess high levels of PTPase activity capable of dephosphorylating a synthetic peptide (FCP) modeled against the negative regulatory region of Src family members. When the FCP phosphatase activity was quantitated, we found 4.8 -9.6-fold higher activity in the transformed cells as compared with normal control cells. Because this tyrosine phosphatase activity might play an important role in c-Src regulation and activation in breast cancer, we have extended these studies by partially purifying and identifying the phosphatase responsible for the Src dephosphorylation activity as PTP1B. We demonstrate that PTP1B has the potential to activate c-Src by dephosphorylation of its negative regulatory site and suggest that PTP1B and c-Src might act together to contribute to the development and/or progression of human breast cancer.

EXPERIMENTAL PROCEDURES
Materials-pJ3H vector containing full-length human PTP1B was a kind gift of Dr. J. Chernoff. The pFLAG-CMV-2 mammalian expression vector was from Eastman Kodak Co. pBluescript vector was from Stratagene. Anti-PTP1B antibody was from Upstate Biotechnologies, Inc. 2-17 and 327 anti-Src monoclonal antibodies were from Quality Biotech and from a hybridoma provided by Dr. Joan Brugge, respectively. Recombinant human c-Src and CSK were expressed and purified as described previously (30,31). Recombinant GST-PTP1B consisting of glutathione S-transferase fused to the first 321 amino acids of human PTP1B was expressed and purified as described previously (32). Poly-(Glu/Tyr) was from Sigma. FCP and cdc2 synthetic peptides were synthesized by Fmoc (N-(9-fluorenyl)methoxycarbonyl) and have been previously described (30,33). Chymotrypsin was from Roche Molecular Biochemicals. Phosphate-free Dulbecco's modified Eagle's medium was from Sigma. Thin layer cellulose chromatography plates were from Kodak. The human breast cell lines and 293 cells were obtained from the American Type Culture Collection.
Cell Culture-All the breast cancer cell lines and the 293 cells were maintained in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and antibiotics (100 units/ml penicillin, 100 g/ml streptomycin, and 0.25 g/ml amphotericin) at 37°C in 5% CO 2 .
Labeling of Phosphatase Substrates-Both the peptide substrates poly(Glu/Tyr) (348 g/ml) and FCP (830 M) were labeled in 115 l of reaction buffer containing 50 mM Hepes, pH 7.0, 2.2 M MnCl 2 , 150 mM NaCl, 1 mM DTT, 20 M ATP, and 200 Ci of [␥-32 P]ATP (3000 Ci/ mmol). For labeling poly(Glu/Tyr), 2.5 g of purified recombinant c-Src, and for labeling FCP, 1.0 g of purified recombinant CSK, were added to the tubes containing reaction buffer and the corresponding peptide and incubated for 2 h at 30°C. Following the phosphorylation reaction, the labeled peptides were separated from unincorporated ATP by chromatography on 5-ml gel filtration columns. The column running buffer consisted of 50 mM Hepes, pH 7.0, and 150 mM NaCl. Poly(Glu/Tyr) was chromatographed on Sephadex G-50 and FCP on Sephadex G-25 (set up in disposable plastic 5-ml pipettes). The peak fractions containing the labeled peptides were pooled and stored frozen at Ϫ20°C until use.
Recombinant c-Src for use as a phosphatase substrate was prepared as follows; 2 g of purified recombinant c-Src was incubated for 30 min at 30°C with 2 g of purified recombinant CSK in 100 l of buffer containing 50 mM Hepes, pH 7.4, 5 mM MgCl 2 , 150 mM NaCl, 1 mM DTT, 5 M ATP, and 100 Ci of [␥-32 P]ATP (3000 Ci/mmol). After the phosphorylation reaction, free ATP was separated from the labeled c-Src by chromatography on a 5.0-ml plastic disposable column containing Sephadex G-50 in buffer composed of 50 mM Hepes, pH 7.0, 150 mM NaCl, and 0.1% Nonidet P-40.
Q Sepharose Ion-exchange Chromatography-Batches of 10 15-cm plastic tissue culture plates of cells at 75% confluence were washed two times with PBS. The cells were then scraped into 40 ml of ice-cold PTPase lysis buffer (10 mM Tris, pH 8.0, 10% glycerol, 1.0% Nonidet P-40, 1.0 mM DTT, 5 g/ml leupeptin, and 1 g/ml aprotinin). The scraped cell extract was stirred at 4°C for 30 min and clarified by centrifugation for 20 min at 8,000 ϫ g. The extract was then filtered through a 0.45-m syringe filter prior to loading on the Q-Sepharose column. Cell extracts were chromatographed on a 2.6 ϫ 9-cm Q-Sepharose column equilibrated in 20 mM Tris, pH 8.0, and 10% glycerol. After sample loading, the column was washed with 10 column volumes of low salt buffer (20 mM Tris, pH 8.0, 10% glycerol, 0.05% Nonidet P-40, 2 g/ml leupeptin, 1 g/ml aprotinin, and 1 mM DTT). The bound proteins were eluted with a 700-ml 0 -0.5 M NaCl gradient, followed by 100 ml of 0.5-1.0 M NaCl (NaCl solutions were made up in low salt buffer). 40 fractions consisting of 20-ml volumes each were collected and assayed for protein content, phosphatase activity, and the presence of various specific tyrosine phosphatases by immunoblot analysis.
Dephosphorylation Assays-20-l samples to be tested for phospha-tase activity were combined with buffer containing 50 mM Hepes, pH 7.2, 0.1% Triton X-100, 5.0 mM EDTA, 5.0 mM EGTA, 20 mM NaF, 5.0 mM NaK-tartrate, 1.0 mM tetramisole, 1 mM DTT, 2.0 g/ml leupeptin, 1.0 g/ml aprotinin, 1 mM benzamidine-HCl, and 0.1 mM phenylmethylsulfonyl fluoride (50 l final reaction volume). Phosphatase substrates (FCP, poly(Glu/Tyr), or recombinant c-Src) were added and incubated for 15 min at 37°C. The reactions were stopped by the addition of 40 l of 50% (v/v) acetic acid. 50 l from each reaction tube was spotted onto 1.5 ϫ 2.0-cm pieces of either 3MM filter paper (poly-(Glu/Tyr) substrate) or P-81 phosphocellulose paper (FCP or recombinant c-Src substrates). The filter papers were then washed five times with either 10% trichloroacetic acid (poly(Glu/Tyr) substrate) or 0.43% (v/v) phosphoric acid (FCP or recombinant c-Src substrates), rinsed once with acetone, and air-dried before quantitation by scintillation counting. Immunodepletion Experiments-PTP1B or control rabbit antibody was preimmobilized on protein G-agarose by first incubating the antibody with protein G-agarose. 3 g of PTP1B or control antibody/20-l portion of packed protein G-agarose beads was incubated for 1 h at 20°C on a rotator. The antibody beads were then washed two times with buffer containing 20 mM Tris, pH 8.0, 10% glycerol, 150 mM NaCl, 0.1% Nonidet P-40, 1 mM DTT, and 2 g/ml leupeptin. The beads were then aliquoted into microcentrifuge tubes such that each immunodepletion tube contained 3 g of antibody bound to 20-l beads, the beads were pelleted by centrifugation at 10,000 ϫ g for 5 s, and all supernatant was removed. Cell extracts or column fractions to be tested for the presence of PTP1B were initially quantitated by the FCP dephosphorylation assay. The samples were then diluted with Q-Sepharose low salt buffer supplemented to contain 150 mM NaCl such that (a) their activities fell within a quantifiable region of the assay and (b) their activities were approximately equal. The diluted samples were then added to tubes containing PTP1B or control rabbit IgG antibody beads and incubated for 90 min at 4°C on a rotator. The beads were pelleted by centrifugation, the supernatant removed, a portion of the supernatant saved for the FCP phosphatase assay, and the remainder transferred to a tube containing a new antibody bead pellet. The incubation, centrifugation, and transfer of supernatant to a fresh tube containing immobilized antibody was repeated twice more for a total of three cycles over the antibody beads. Three cycles of immunodepletion under these conditions was found to quantitatively remove greater than 95% of the PTP1B in test samples.
c-Src Inactivation/Reactivation Assay-For the inactivation step of the assay, 40 ng of purified recombinant c-Src was incubated with 1 g of purified recombinant CSK in a 35-l final volume of kinase buffer containing 50 mM Hepes, pH 7.4, 5 mM MgCl 2 , 150 mM NaCl, 1.0 mM DTT, and 50 M ATP. In experiments where the phosphorylation status of c-Src was examined, 35 Ci of [␥-32 P]ATP (3000 Ci/mmol) was also added to label the c-Src. This reaction was allowed to proceed for 30 min at 30°C. The reaction was stopped and the c-Src immunoprecipitated by the addition of 35 l of SDB buffer (50 mM Hepes, pH 7.4, 100 mM NaCl, 1% glycerol, 1 mM DTT, 0.1% Nonidet P-40) that also contained 20 mM EDTA and 1.5 g of 2-17 anti-Src antibody. After incubation for 1 h at 4°C, 15 l of protein G-agarose was added and further incubated for 1 h at 4°C on a rotator. The beads containing the immune complexes were then washed three times with SDB buffer and all the supernatant aspirated.
For the reactivation step of the assay, GST-PTP1B was diluted in SDB buffer, added to the bead pellets containing inactivated c-Src, and incubated 1 h at 4°C. The beads were then washed three times with SDB buffer containing 200 M sodium orthovanadate and 4 mg/ml p-nitrophenolphosphate, and c-Src kinase activity was measured.
Immunoprecipitation of c-Src from Cell Extracts-Cells in tissue culture dishes were rinsed twice with phosphate-buffered saline at 4°C. They were then lysed in 0.80 ml of RIPA buffer (50 mM Tris-HCl, pH 7.2, 0.15 M NaCl, 1.0 mM EDTA, 0.1% SDS, 1.0% Triton X-100, 1.0% sodium deoxycholate), supplemented with phosphatase and protease inhibitors (1 mM sodium orthovanadate, 3 mg/ml p-nitrophenolphosphate, 50 g/ml leupeptin, and 10 g/ml aprotinin). c-Src was immunoprecipitated by the addition of 1 g of 327 anti-Src antibody to 300 g of cell extract for 1 h at 4°C, followed by incubation with 15 l of protein G-agarose on a rotator at 4°C for 1 h. The immune complexes were washed four times with RIPA buffer (supplemented with phosphatase and protease inhibitors) and once with SDB buffer.
Measurement of c-Src Kinase Activity-c-Src kinase activity was measured in the immmunoprecipitates by the addition of 50 l of kinase assay buffer containing 50 mM Hepes, pH 7.4, 150 mM NaCl, 1 mM DTT, mg/ml p-nitrophenolphosphate. Following incubation for 15 min at 30°C, 25 l of 50% (v/v) acetic acid was added to each tube, after which 50 l was spotted onto a 1.5 ϫ 2.0-cm square of P-81 phosphocelluose paper. The filter papers were then washed with dilute phosphoric acid as described in the dephosphorylation assay.
Phosphopeptide Mapping-Samples to undergo phosphopeptide mapping were initially resolved on 8% SDS-PAGE. The gel was fixed and washed on a orbital shaker for 1 h each in three solutions containing, respectively, 250 ml of 40% methanol and 10% acetic acid, 250 ml of 20% methanol and 5% acetic acid, and 250 ml of 10% methanol and 2.5% acetic acid. The gel was then dried between two sheets of clear cellulose, exposed to x-ray film, and the radioactive bands excised from the dried gel. The gel slices were rehydrated in 10% methanol, tweezers were used to remove the cellulose, and the 10% methanol was pipetted from the acrylamide slices and discarded. The slices were then dried on a SpeedVac. To generate chymotryptic fragments of the c-Src protein, 500 l of 0.05 M NH 4 HCO 3 containing 3 g of chymotrypsin was added to each tube and the tubes were incubated at 35°C for 4 h in a shaking water bath. 100 l of 0.05 M NH 4 HCO 3 containing 1 g of chymotrypsin was further added to each tube, incubated for 2 h, and repeated once. The supernatant containing the majority of the chymotryptic peptide fragments was then collected. Complete elution of the digested peptides from the gel slices was facilitated by further incubation of the gel slices for 2 h with 0.5 ml of 0.05 M NH 4 HCO 3 , followed by two incubations with 0.75 ml of water for 2 h each at 35°C in a shaking water bath. After each incubation, the supernatants containing eluted peptides were pooled with the original supernatant and dried on the SpeedVac. This procedure ensured complete recovery of the chymotryptic protein fragments (recovery, as measured by transfer of radioactivity from the gel slices to the pooled supernatants was typically greater than 95%). Each digested sample was then resuspended in 500 l of water, and redried. This was repeated three times to ensure efficient removal of the NH 4 HCO 3 .
The samples were resuspended in 5-10 l of water and were applied to a 20 ϫ 20-cm cellulose thin layer chromatography plate. The peptides were separated in the first dimension by electrophoresis at pH 4.72 in a buffer containing n-butanol/pyridine/acetic acid/water (2/1/1/36) for 1 h at 800 V using a Savant thin layer electrophoresis apparatus. The peptides were then separated in the second dimension by chromatography in buffer containing n-butanol/pyridine/acetic acid/water (15/10/ 3/12). The plates were then dried in a fume hood overnight and subjected to autoradiography. In situations where quantitation of the phosphopeptides was necessary, imaging of the thin layer plates was carried out with a Storm 860 PhosphorImager (Molecular Dynamics). The identification of the phosphopeptide spots was made by comparison with phosphorylated synthetic peptides containing Tyr-419 and Tyr-530, by comparison with recombinant c-Src that had been subjected to either autophosphorylation or CSK phosphorylation (Tyr-419 and Tyr-530, respectively), and by phosphoamino acid analysis and mobility during electrophoresis/chromatography (Ser-17) (results not shown).
Preparation of the Expression Constructs and Cell Transfections-Mammalian expression vectors expressing FLAG-tagged human PTP1B were constructed. pJ3H containing either full-length wild-type human PTP1B or a catalytically inactive mutant form of full-length PTP1B possessing a point mutation (C215S) (34) were digested with BamHI to release a 1.4-kilobase pair fragment containing the entire human PTP1B coding region. The BamHI fragment was ligated into BamHI-digested pFLAG-CMV-2 mammalian expression vector. This expression vector produces amino-terminally FLAG-tagged full-length PTP1B under the control of the CMV promoter.
293 human embryonal kidney cells in 3.5-cm dishes at approximately 50% confluence were transiently transfected with 4 g of purified plasmid containing of the FLAG-tagged constructs using Lipo-fectAMINE 2000 (Life Technologies, Inc.) transfection reagent as per the manufacturer's instructions. All plasmids used in transfection protocols were purified using a Qiagen endotoxin-free plasmid purification kit according to the manufacturer's instructions. The cell culture medium was changed once 24 h after transfection, and the cells were lysed in RIPA buffer 48 h after transfection.
In Vivo Orthophosphate Labeling of 293 Cells-293 cells to be labeled with [ 32 P]orthophosphate were grown to approximately 75% confluence in 6-cm tissue culture dishes. The cells were rinsed once with phosphate-free Dulbecco's modified Eagle's medium and then incubated for 8 h in 3 ml of phosphate-free Dulbecco's modified Eagle's medium containing 1% fetal bovine serum, 9% dialyzed fetal bovine serum (dialyzed against 0.9% NaCl), and 1 mCi/ml [ 32 P]orthophosphate. The cells were then lysed in RIPA buffer (supplemented with phosphatase and protease inhibitors), c-Src was immunoprecipitated, and the im-mune complex resuspended in sample buffer and subjected to SDS-PAGE.

RESULTS
Partial Purification of a Protein-tyrosine Phosphatase Activity from a Human Breast Cancer Cell Line-Our laboratory has previously reported elevated levels of Src tyrosine kinase activity in approximately 30 -40% of the human breast cancer cell lines examined (29). The increase in kinase activity was due to both elevations in Src specific activity as well as protein levels. These increases in c-Src activity were correlated with elevations in the activity of an unknown phosphatase that was capable of dephosphorylating a synthetic peptide modeled against the negative regulatory carboxyl terminus of Src family members. We have now extended these investigations by carrying out experiments to purify and identify from human breast cancer cell lines the phosphatase responsible for this activity.
Chromatography of crude cell extracts from MDA-MB-435S cells, a human breast cancer cell line that exhibited elevated c-Src kinase activity, on a Q-Sepharose ion-exchange column revealed that the PTPase activity capable of dephosphorylating the nonspecific phosphatase substrate poly(Glu/Tyr) was broadly spread over four peaks (Fig. 1, upper panel, ࡗ). The first (fractions 11 and 12) and second (fractions 14 -17) peaks possessed the majority of the phosphatase activity, whereas the third (fractions 19 -21) and fourth (fraction 23) peaks possessed lower amounts of activity. The same column fractions were then assayed for their ability to dephosphorylate a synthetic peptide, FCP, modeled against the carboxyl-terminal regulatory region of Src-family members (Fig. 1, lower panel, OE). Virtually all of the PTPase activity capable of dephosphorylating the FCP peptide was found to elute in fractions corresponding to the second peak of poly(Glu/Tyr) phosphatase activity (fractions 14 -17). Next, we wanted to confirm that the second peak of phosphatase activity would also be capable of dephosphorylating the carboxyl-terminal regulatory tyrosine (Tyr-530) in the intact c-Src molecule. The c-Src protein, when phosphorylated on Tyr-530, is folded such that its carboxyl terminus interacts with its SH2 domain, and as a result, a peptide might behave differently than the intact c-Src protein. Recombinant c-Src was phosphorylated in vitro in the presence of radioactive ATP and an excess amount of recombinant CSK, a tyrosine kinase that specifically targets Tyr-530 in c-Src. This phosphorylated c-Src product was then used as a substrate for the phosphatase activity in the column fractions (Fig. 1, lower panel, q). In accordance with the previous result demonstrating activity against the FCP peptide, the major peak of phosphatase activity against intact c-Src corresponded to the second peak of poly(Glu/Tyr(P)) phosphatase activity (fractions 14 -17). Although some minor peaks of activity against intact c-Src were also observed corresponding to the three additional peaks of activity detected against poly(Glu/Tyr), most of the activity against intact c-Src appeared localized in the major peak (fractions 14 -17).
Identification of the Protein-tyrosine Phosphatase Activity in Human Breast Cancer Cells-In order to simplify the identification of the protein-tyrosine phosphatase activity present in the major peak of FCP dephosphorylation activity from the Q-Sepharose column, we initially wished to rule out known protein-tyrosine phosphatases. To do this, we examined the column fractions by immunoblot analysis utilizing commercially available antibodies against several known protein-tyrosine phosphatases. We initially examined a panel of six antibodies, including antibodies against RPTP-␣ and SHP-1, two phosphatases that are capable of dephosphorylating and activating c-Src (16,(35)(36)(37). Five of the antibodies, including antibodies against cdc25b, SHP-1, SHP-2, RPTP-␣, and LAR, either did not detect any immunoreactivity in the fractions or identified protein bands in fractions other than the fractions containing the FCP dephosphorylation activity (results not shown). In contrast, the anti-PTP1B antibody recognized a 50-kDa PTP1B band in the same fractions (fractions 14 -17) containing the majority of the FCP dephosphorylating activities (Fig. 2). Fraction 17 also contained a pair of closely spaced bands at approximately 42 kDa that were recognized by the antibody against PTP1B.
To further confirm that PTP1B was the phosphatase responsible for the FCP dephosphorylation activity and to determine what proportion of the FCP dephosphorylation activity might be attributable to PTP1B, we carried out immunodepletion experiments. The Q-Sepharose peak fraction 15, a pool of Q-Sepharose fractions 14 -17, and the whole cell extract of MDA-MB-435S cells prior to Q-Sepharose chromatography were diluted in buffer such that the three samples possessed similar FCP dephosphorylation activities. The fractions were then subjected to three cycles of immunodepletion with an excess amount of anti-PTP1B antibody, and the FCP dephosphorylation activity remaining in the supernatant was measured. The difference in activity between the FCP dephosphorylation activity before and after immunodepletion was attributed to PTP1B, and was expressed as a percentage of the total activity in each test sample before immunodepletion (Fig. 3). 93% and 88%, respectively, of the FCP dephosphorylation activity in the peak fraction and pooled peak fractions could be immunodepleted and thus, could be attributed to PTP1B. In the whole cell extract, 70% of the FCP dephosphorylation activity was attributable to PTP1B. Immunodepletion with normal rabbit IgG did not remove significant amounts of FCP phosphatase activity from any samples (results not shown). We also depleted whole cell extract from several additional human breast cancer cell lines with elevated c-Src activity, including BT-483, Hs578T, and SK-BR-3. We found that, similar to the above result with MDA-MB-435S cells, more than 50% of the FCP dephosphorylation activity within these other cell lines was attributable to PTP1B (results not shown).
PTP1B Protein Levels in Human Breast Cancer Cell Lines-Because PTP1B appeared to be the major activity responsible for the c-Src dephosphorylation activity in a number of breast cancer cell lines, we were interested in examining the relative levels of PTP1B protein present in human breast cancer cell lines as compared with normal control cells. Immunoblots of whole cell extracts from four breast cancer cell lines (SK-BR-3, BT-483, MDA-MB-435S, and MDA-MB-468; lanes 1, 2, 4, and 5, respectively) were compared with normal breast cells (HS-578-Bst; lane 3). As shown in Fig. 4, the highest levels of PTP1B protein were found in the BT-483 cell line, followed in order of abundance by the SK-BR-3, MDA-MB-435S, and the MDA-MB-468 breast cancer cell lines. The least amount of PTP1B was found in the normal control HS-578-Bst. All the whole cell extracts also contained a lower molecular weight form of PTP1B (approximately 42 kDa). This 42-kDa form of PTP1B was most prominent in the breast cancer cell lines, and less so in the normal control cells.
Dephosphorylation and Activation of c-Src by PTP1B in Vitro-We demonstrated in Fig. 1 that column fractions containing PTP1B could dephosphorylate intact c-Src, but we also wanted to examine whether PTP1B-mediated dephosphorylation of c-Src could activate its kinase activity. To do this, we utilized an in vitro inactivation/reactivation assay. For the inactivation step, purified recombinant c-Src was incubated with excess CSK in the presence of ATP. This results in the   FIG. 2. Anti-PTP1B immunoblot of Q-Sepharose column fractions. 100 l of each column fraction was subjected to 10% SDS-PAGE and transfer to nitrocellulose. The nitrocellulose was then blotted with anti-PTP1B antibody, followed by detection using horseradish peroxidase-conjugated anti-rabbit antibody and ECL chemiluminescence detection reagents. phosphorylation of the negative regulatory site, Tyr-530, within the carboxyl terminus of c-Src. The inactivated c-Src was then immunoprecipitated and incubated with GST-PTP1B for the reactivation step. The immunoprecipitates were washed to remove the phosphatase, after which the Src kinase activity measured. In Fig. 5A, c-Src is shown to be inactivated by incubation with CSK. This is accompanied by phosphorylation of c-Src by CSK (Fig. 5B, lane 1), a phosphorylation that occurs primarily on Tyr-530, as demonstrated by two-dimensional phosphopeptide maps (Fig. 5C, panel 1). A small amount of autophosphorylation on the major c-Src autophosphorylation site, Tyr-419, can also be observed. Incubation of the inactivated c-Src with PTP1B results in a dose-dependent reactivation of c-Src tyrosine kinase activity. Incubation of c-Src with 2 or 10 g of PTP1B results in partial or full restoration of c-Src kinase activity, respectively (Fig. 5A). The activation is accompanied by dephosphorylation of c-Src (Fig. 5B, lanes 2 and 3), both of Tyr-419 and of Tyr-530 (Fig. 5C, panels 2 and 3). In the presence of the phosphatase inhibitor sodium orthovanadate, both the reactivation (Fig. 5A, ϩvanadate) and dephosphorylation (Fig. 5B, ϩvanadate, lane 4; Fig. 5C, panel 4) are inhibited.
Activation of c-Src by PTP1B in Vivo-In addition to showing in vitro activation of c-Src by PTP1B, we wanted to examine whether PTP1B was capable of c-Src activation in vivo. To do this, we overexpressed wild-type PTP1B in 293 human embryonal kidney cells and looked for changes in c-Src kinase activity. In Fig. 6A, overexpression of PTP1B is shown to result in a 2-fold increase in c-Src activity over that observed in both untransfected cells and cells transfected with the vector alone. In addition, overexpression of a mutant form of PTP1B lacking FIG. 3. Immunodepletion of PTP1B from column fractions and whole cell extracts. Equal amounts of FCP phosphatase activity derived from either Q-Sepharose peak fraction 15 (see Fig. 1), pooled Q-Sepharose peak fractions 14 -17, or whole cell extract from MDA-MB-435S cells were immunodepleted three times using anti-PTP1B antibody as described under "Experimental Procedures." The activity in each sample following immunodepletion as well as prior to immunodepletion was quantitated by the FCP dephosphorylation assay. The percentage of the total activity removed by the immunodepletion procedure is indicated in the bar graph and is representative of duplicate experiments. breast cancer cells were subjected to 10% SDS-PAGE and transferred to nitrocellulose. The nitrocellulose was then blotted with anti-PTP1B antibody, followed by detection using horseradish peroxidase-conjugated anti-rabbit antibody and ECL chemiluminescence detection reagents. The intact 50-kDa PTP1B band is marked with an arrow.

FIG. 5. Activation of c-Src by PTP1B in vitro.
c-Src that had been inactivated by CSK, was immunoprecipitated, and then treated with 0, 2, or 10 g of PTP1B for 1 h at 4°C. 1 mM sodium orthovanadate was added to one tube containing 10 g of PTP1B. The c-Src immune complex was washed extensively to remove the added phosphatase, and the c-Src kinase activity was measured using the cdc2 synthetic peptide substrate. The results are expressed as counts per minute incorporated into the cdc2 peptide and are shown in A, where the bars represent the mean of duplicate measurements (Ϯ 1 S.D.) and the experiment is representative of duplicate experiments. In a parallel set of tubes, [␥-32 P]ATP was included in the first incubation with CSK, and the phosphorylation status of c-Src was monitored during subsequent treatments. The samples were subjected to 8% SDS-PAGE. The fixed and dried gel was exposed to X-Omat AR film and an autoradiogram is shown in B. The radioactive bands from B were excised and digested with chymotrypsin. The digested samples were subjected to two-dimensional phosphopeptide mapping and autoradiograms are shown in C. The sample in panel 1 was treated with CSK alone. In panels 2 and 3, the samples were treated with CSK and either 2 or 10 g of PTP1B. Finally, in panel 4, the sample was treated with CSK and with 10 g of PTP1B in the presence of vanadate. The origin of application is indicated in the lower right-hand corner of each panel, with electrophoresis carried out in the horizontal dimension (anode on left) and chromatography in the vertical dimension. phosphatase activity appeared unable to modify c-Src activity from that observed in untransfected cells. The expression levels of both the wild-type and catalytically inactive PTP1B appeared approximately equal in the transfected cells, as shown in the anti-PTP1B immunoblot in Fig. 6B. The changes in c-Src tyrosine kinase activity appeared to be due to changes in specific activity rather that changes in expression level, as the untransfected and transfected cells contained equal amounts of c-Src protein as shown in the anti-Src immunoblot in Fig. 6C.
We next examined whether these changes in c-Src specific activity were accompanied by changes in its phosphorylation status, in particular, whether there were any changes in the phosphorylation status of Tyr-530. 293 cells were either left untreated or transfected with pFLAG-CMV-2 vector or vector expressing wild-type PTP1B. 48 h after transfection, the cells were metabolically labeled with [ 32 P]orthophosphate for 8 h, the cells were lysed, and c-Src was immunoprecipitated and subjected to phosphopeptide mapping. As shown in Fig. 7, two major in vivo sites of phosphorylation were detected, Tyr-530 and Ser-17. Transfection of 293 cells with vector alone appeared to have little effect upon either phosphorylation site as compared with untreated cells, whereas in cells transfected with vector expressing wild-type PTP1B, there was a reduction in the intensity of the Tyr-530 spot, with also a slight reduction in the Ser-17 spot. To confirm these visual observations, the phosphopeptide spots were quantitated using a PhosphorImager. As shown in Table I, quantitation of Tyr-530 indicated there was a 56% reduction in phosphorylation in the cells expressing wild-type PTP1B as compared with the untreated cells and cell transfected with vector alone, and this confirmed the visual data in Fig. 7.
To further confirm the significance of the reduction of phosphorylation of Tyr-530, we also compared the level of Tyr-530 phosphorylation to phosphorylation at Ser-17, a c-Src phosphorylation site that is normally phosphorylated in vivo in many cells types (38,39), but is not a substrate of PTP1B. The intensity of the Tyr-530 phosphopeptide spot was found to be higher than the intensity of the Ser-17 phosphopeptide spot in both the untreated as well as the vector control cells, having a ratio of 1.22 in both cases (see Table I). In cells transfected with vector expressing wild-type PTP1B, this ratio shifted to 0.68. This shift in the Tyr-530/Ser-17 ratio was not due to an increase in the total amount of the Ser-17 spot (which actually was reduced by approximately 23%) in the wild-type PTP1Btransfected sample, but was largely due to the decrease in Tyr-530 phosphorylation. DISCUSSION In this report, we provide evidence indicating that a high level of the PTPase PTP1B contributes to the activation of c-Src kinase activity in human breast cancer cells. We have partially  Table I for quantitation of these results.

TABLE I Quantitation of the c-Src phosphopeptides from transfected 293 cells
The radioactivity in the phosphopeptide spots, corresponding to Tyr-530 and Ser-17 in Figure 7, was quantitated using a PhosphorImager and expressed in arbitrary units (first two columns). The ratio of the intensities of Tyr-520 to Ser-17 was calculated and is shown in the last column. purified a protein-tyrosine phosphatase activity from a human breast cancer cell line that is capable of dephosphorylating the carboxyl-terminal regulatory tyrosine of c-Src and of activating c-Src in an in vitro inactivation/reactivation assay. This activity has been identified as PTP1B by immunoblotting and immunodepletion experiments. Overexpression of PTP1B in human embryo kidney 293 cells by transient transfection resulted in reduced Tyr-530 phosphorylation and activation of endogenous c-Src activity, demonstrating the ability of PTP1B to activate c-Src in vivo. This may represent a very high level of, or complete dephosphorylation of, Tyr-530 in cells in the population that were transfected by PTP1B, since transfection efficiency is less than 100%. Our identification of PTP1B as the primary phosphatase in human breast cancer cell lines that is capable of dephosphorylating the negative regulatory site of c-Src was initially somewhat unexpected. PTP1B is a rather ubiquitously expressed endoplasmic reticulum-associated protein-tyrosine phosphatase (40) that has been implicated in the regulation of a variety of cellular processes including cell growth, differentiation, signaling, and transformation. In particular, PTP1B has been demonstrated to antagonize insulin (41,42) and insulin-like growth factor-I (42) signaling, antagonize signaling by the oncoprotein p210 bcr-abl (43), suppress transformation by Neu (44), v-Crk (45), v-Src (45), and v-Ras (45), and regulate integrin signaling (46,47). Although PTP1B has been implicated in playing a role in many cellular functions, definitive studies addressing whether PTP1B has specific targets or whether it cooperates with other phosphatases in the dephosphorylation of its targets have been lacking.
Identification of potential intracellular substrates of PTP1B has been complicated by the rather nonspecific nature of protein-tyrosine phosphatases. Investigators utilizing peptide substrates have noted some specificity of PTP1B, in particular, its preference for acidic residues at the Ϫ2 and Ϫ3 position (48) and aromatic residues at the Ϫ1 position (49). c-Src and other members of the Src family possess a highly conserved acidic residue, Glu, at position Ϫ3 in relation to the carboxyl-terminal regulatory Tyr-530 that might play an important role in substrate recognition of this region by PTP1B. Higher order structures within proteins have also been demonstrated to affect the ability of PTP1B to dephosphorylate target proteins (48). This would be of particular relevance to c-Src as well as the Src family of tyrosine kinases, where the carboxyl-terminal tyrosine interacts with the SH2 domain in an intramolecular manner such that the kinase activity is inhibited. In fact, association of SH2 domains with Src family carboxyl-terminal regulatory tyrosine residues have previously been shown to protect the Src family members from dephosphorylation and activation (50), and this interaction might also affect which phosphatases are capable of carrying out the dephosphorylation event. It is interesting to note that PTP1B possesses two proline-rich regions containing the consensus for class II SH3 binding ligands (PXXPX(R/K)) (51) and is capable of binding to the SH3 of p130 Cas (52). It is possible that one or both these proline-rich regions might interact with the SH3 domain of c-Src. This might assist in unfolding c-Src such that Tyr-530 becomes accessible for dephosphorylation, and we are currently investigating this possibility. In our paper, we note that the major phosphatase activity present in the human breast cancer cells was capable of dephosphorylating both a synthetic peptide modeled against the carboxyl terminus of Src family members as well as the intact c-Src protein. This suggests that PTP1B is likely the major activity within the breast cancer cells that is capable of performing this task.
Evidence consistent with our finding that PTP1B can de-phosphorylate and regulate c-Src was reported recently in studies addressing the ability of PTP1B to modify integrinmediated adhesion and signaling in mouse L cell fibroblasts (46). Expression of a dominant-negative form of PTP1B was found to cause reduced cell adhesion and almost a complete loss of focal adhesions and stress fibers. These changes were attributed to the accompanying reductions in tyrosine phosphorylation of focal adhesion kinase and paxillin, two targets of Src family kinases. Additional experiments demonstrated reductions in c-Src activity, and the reduction in activity was attributable to increased phosphorylation of the carboxyl-terminal regulatory tyrosine. In the same set of studies, expression of wild-type PTP1B increased Src kinase activity (expected if c-Src is a PTP1B substrate), but did not modify integrin-mediated adhesion and signaling. Because both our laboratory and others have now implicated PTPT1-B in the modification of c-Src activity by dephosphorylation of the carboxyl-terminal regulatory tyrosine residue (46), it will be of interest to determine where the two molecules interact within the breast cancer cells. PTP1B has previously been shown to possess a carboxyl-terminal domain that targets it for endoplasmic reticulum localization (40). More recently, it has been demonstrated that PTP1B also associates with proteins at sites other than the endoplasmic reticulum, including focal adhesions in mouse L cell fibroblasts (46) and adherens junctions in chick retina (53,54). Importantly, c-Src has also been localized to both focal adhesions (55) and adherens junctions (56), and PTP1B appears to play a role in c-Src activation at focal adhesions (46). At focal adhesions, activated c-Src has been shown to phosphorylate and activate focal adhesion kinase (57), whereas, at adherens junctions, c-Src has been implicated in the phosphorylation of ␤-catenin (56,58). Thus, current evidence suggests that c-Src and PTP1B do colocalize and can interact in different cell types, and we are currently examining if the same situation exists in human breast cancer cells.
We have examined the PTP1B protein levels in several human breast cancer cell lines that previously demonstrated elevations in c-Src activity, and have found levels of PTP1B higher than those found in a normal control breast cell line. PTP1B overexpression has also been reported by several other groups in the cases of human breast and ovarian cancer (59,60), both confirming our findings, and suggesting its overexpression might not be limited to breast cancer alone. In some cases, increased PTP1B expression has been associated with overexpression of c-erbB-2 (c-neu) (60, 61), a protooncogene found overexpressed in 25-30% of human breast cancers (62), and that is known to contribute to the malignant phenotype. In addition to increased PTP1B protein expression in human breast cancer cell lines, we have preliminary evidence suggesting that PTP1B specific activity is also elevated in these cells. We have previously reported 4.8 -9.6-fold elevations in FCP dephosphorylation activity (29) in the same set of breast cancer cell lines examined in the current study. The increased levels of activity exceed the increased levels of protein expression we have noted in PTP1B immunoblots. Careful quantitation will need to be carried out in the future, but these two pieces of data suggest that PTP1B specific activity is also elevated. Interestingly, we noted in the immunoblots of the breast cancer cell lines, the presence of a lower molecular weight form of PTP1B. A similar sized fragment of PTP1B has been reported as an in vivo proteolyzed form of PTP1B that is generated upon platelet activation (63). This form of PTP1B was found to have 2-fold higher enzymatic activity than intact PTP1B (63). The presence of this "activated" form of PTP1B, in combination with increased expression levels of PTP1B, may enhance PTP1B's ability to dephosphorylate substrates, and in this particular situation, activate c-Src.
Several protein-tyrosine phosphatases have now been implicated in the dephosphorylation and activation of c-Src, including receptor-like protein-tyrosine phosphatase ␣ (16,35,36), protein-tyrosine phosphatase (64), SHP-1 (37), and SHP-2 (65). These phosphatases dephosphorylate the carboxyl-terminal regulatory tyrosine of c-Src resulting in c-Src activation, with SHP-2 also capable of a non-enzymatic mode of activation (66). Our identification of PTP1B as a protein-tyrosine phosphatase that can dephosphorylate and activate c-Src suggests that, depending upon their expression patterns and cellular distributions, one or more phosphatases may contribute to the phosphorylation status of the carboxyl-terminal regulatory tyrosine of c-Src and thus, also contribute to the resulting enzymatic activity of c-Src.