Positive Effect of Overexpressed Protein-tyrosine Phosphatase PTP1C on Mitogen-activated Signaling in 293 Cells*

PTP1C, an SH2 domain-containing protein-tyrosine phosphatase, is predominantly expressed in hematopoietic cells, in which it negatively regulates cellular sig- naling. However, this enzyme is also expressed in many non-hematopoietic cells. We demonstrate here that in non-hematopoietic 293 cells, overexpression of a catalytically inactive mutant of PTP1C strongly suppressed the stimulatory effects of the epidermal growth factor or serum on cell proliferation, early gene transcription, and DNA synthesis. Similarly, the phosphorylation of the mitogen-activated protein kinase and mitogen-acti- vated protein kinase kinase activity was markedly inhibited by overexpression of mutant PTP1C. The inhib- itory effect of mutant PTP1C was overcome by cotransfection with wild-type PTP1C, but not with the structurally related PTP2C. Furthermore, expression of the mutant phosphatase resulted in hyperphosphoryla- tion on tyrosine of a 95-kDa protein that was co-immu-noprecipitated with the mutant, but not with the wild- type protein. These results suggest that, unlike in hematopoietic cells, PTP1C in 293 cells plays a positive role in epidermal growth factor- or serum-activated mitogenesis. Thus, and Amersham Corp., respectively. Construction of Mutant PTP1C and Transfection— A catalytically inactive mutant, PTP1C C455S, in which Cys-455 was mutated to Ser, was generated by the polymerase chain reaction method (24). The constructs with the cytomegalovirus promoter for expressing wild-type PTP1C and mutant PTP1C C455S and the procedures for selection of cell lines stably expressing PTP1C and mutant PTP1C C455S were previously described (24). For the stable expression of two constructs in

PTP1C, an SH2 domain-containing protein-tyrosine phosphatase, is predominantly expressed in hematopoietic cells, in which it negatively regulates cellular signaling. However, this enzyme is also expressed in many non-hematopoietic cells. We demonstrate here that in non-hematopoietic 293 cells, overexpression of a catalytically inactive mutant of PTP1C strongly suppressed the stimulatory effects of the epidermal growth factor or serum on cell proliferation, early gene transcription, and DNA synthesis. Similarly, the phosphorylation of the mitogen-activated protein kinase and mitogen-activated protein kinase kinase activity was markedly inhibited by overexpression of mutant PTP1C. The inhibitory effect of mutant PTP1C was overcome by cotransfection with wild-type PTP1C, but not with the structurally related PTP2C. Furthermore, expression of the mutant phosphatase resulted in hyperphosphorylation on tyrosine of a 95-kDa protein that was co-immunoprecipitated with the mutant, but not with the wildtype protein. These results suggest that, unlike in hematopoietic cells, PTP1C in 293 cells plays a positive role in epidermal growth factor-or serum-activated mitogenesis. Thus, PTP1C participates in multiple signaling pathways, where the enzyme, depending on its target molecules, may function as either a positive or negative mediator.
Upon binding of external ligands, several cell-surface receptors, such as the insulin receptor, epidermal growth factor (EGF) 1 receptor, and platelet-derived growth factor (PDGF) receptor, are activated via dimerization, resulting in autophosphorylation on multiple tyrosine residues within the intracellular domains of the receptors (1,2). The autophosphorylated domain of the receptor then provides high affinity binding sites for specific cellular SH2 domain-containing proteins (3). The interaction of these SH2-containing proteins with specific tyrosyl-phosphorylated sites of the receptor causes other signaling molecules to be recruited to the receptor, where these molecules, as well as most of the SH2-containing proteins, can be phosphorylated by the receptors and/or other recruited pro-teins (3,4). Formation of these signal transduction complexes coordinates the multiple intracellular programs that initiate various changes in cell proliferation.
Previously, we and others have identified two SH2 domaincontaining protein-tyrosine phosphatases, designated PTP1C (5) (also termed SH-PTP1/HCP/SHP) (6 -9) and PTP2C (9) (also known as SH-PTP2, Syp, and PTP1D) (10 -12). PTP2C, a ubiquitously expressed enzyme, is highly homologous to another member of this SH2-containing subfamily, the product of the Drosophila corkscrew (csw) gene (13). The SH2 domains of PTP2C interact with activated receptor protein-tyrosine kinases, such as the EGF and PDGF receptors (11,12), and with the substrates of these receptor protein-tyrosine kinases, such as insulin receptor substrate-1 (14). As a result of these interactions, PTP2C itself becomes tyrosyl-phosphorylated by the receptor protein-tyrosine kinases, and its activity can be modulated by such tyrosyl phosphorylation (12, 14 -16). In these mitogen-activated protein (MAP) kinase pathways, PTP2C appears to be a positive mediator (17)(18)(19)(20)(21). PTP1C is predominantly expressed in hematopoietic cells (6,7). Like PTP2C, PTP1C can be tyrosyl-phosphorylated in response to various ligands, such as colony-stimulating factor-1, stem cell factor, PDGF, insulin, thrombin, and PMA (22)(23)(24)(25)(26). In hematopoietic cells, PTP1C was found to associate with c-Kit, the receptor for the Fc region of IgG, and the erythropoietin receptor (22,27,28). The interaction of PTP1C with these receptors through the SH2 domains appears to play a role as a negative regulator in hematopoietic cells (22,(27)(28)(29). However, recent data revealed that PTP1C is also substantially expressed in a variety of non-hematopoietic cells, especially in some malignant epithelial cells (5,6,12,30), suggesting that this enzyme plays some roles in cellular signal transduction in these cells. In this report, we demonstrate that in human embryonic kidney 293 cells, the phosphatase activity of PTP1C, like that of PTP2C, is a positive mediator in serum-and EGF-activated mitogenic signaling.

EXPERIMENTAL PROCEDURES
Materials-Human recombinant EGF was obtained from Boehringer Mannheim. Human embryonic kidney 293 cells were obtained from American Type Culture Collection. Polyclonal anti-PTP1C (serum 237) and anti-MAP kinase (serum 1263) antibodies were produced by immunization of rabbits with the antigens as described previously (21). Antiphosphotyrosine antibody 4G10 was obtained from Upstate Biotechnology, Inc. Myelin basic protein and [␥-32 P]ATP were purchased from Sigma and Amersham Corp., respectively.
Construction of Mutant PTP1C and Transfection-A catalytically inactive mutant, PTP1C C455S, in which Cys-455 was mutated to Ser, was generated by the polymerase chain reaction method (24). The constructs with the cytomegalovirus promoter for expressing wild-type PTP1C and mutant PTP1C C455S and the procedures for selection of cell lines stably expressing PTP1C and mutant PTP1C C455S were previously described (24). For the stable expression of two constructs in the same cells, the second construct was transfected with an additional selective marker (hygromycin B) into cells that already stably expressed the mutant phosphatase encoded by the first construct, PTP1C C455S. With selection of the transfected clones, the expression levels of the recombinant proteins encoded by different forms of the proteintyrosine phosphatases were comparable to each other as confirmed by Western blotting. Phosphatase activity was determined as described previously (21).
Stimulation of Cells and Assay for MAP Kinase Kinase Activity-The stimulation of cells with EGF and the preparation of cell lysates were carried out as described (21). MAP kinase kinase activity was assayed as the ability to stimulate the myelin basic protein kinase activity of recombinant MAP kinase ERK2 in a coupled assay system as described previously (21). The recombinant Escherichia coli cells were generously provided by Dr. Melanie Cobb (University of Texas).
Immunoprecipitation and Western Blotting-Stimulated and unstimulated 293 cells were washed with ice-cold phosphate-buffered saline containing 2 mM Na 3 VO 4 and then lysed in 1.0 ml of lysis buffer (supplemented with 1% Triton X-100 and 0.1 M NaCl)/10-cm dish as described above. The extract was clarified by centrifugation at full speed in a microcentrifuge. The clarified extracts were subjected to immunoprecipitation as described (21). For Western blotting, the cell extracts were immunoblotted with anti-MAP kinase antibody or anti-PTP1C antibody as described previously (21,24).
Assay for DNA Synthesis-Cells (5000/well) were seeded in a 96-well plate and made quiescent by starvation in Dulbecco's modified Eagle's medium/Ham's F-12 medium (1:1) for 24 h. The quiescent cells were stimulated in Dulbecco's modified Eagle's medium/Ham's F-12 medium (1:1) containing [ 3 H]thymidine (1 Ci/well) with various concentrations of EGF and serum as indicated. After 12 h of incubation, the cells were washed three times with ice-cold phosphate-buffered saline and fixed with trichloroacetic acid (10%). Cells were then recovered with 0.1 N NaOH, and the incorporated radioactivity was counted.
Luciferase Activity Assay-The plasmid of the serum response element-driven luciferase reporter gene (SRE-Luc) was generously provided by Dr. Jeffery Pessin (University of Iowa). A plasmid for expression of green fluorescence protein (GFP) (CLONTECH) was used as a reference plasmid. Control cells and cells expressing PTP1C and PTP1C C455S were plated at 10 6 cells/90-mm dish and transiently transfected with 10 g of DNA from the SRE-Luc and GFP plasmids using the calcium phosphate method. 18 h after transfection, cells were split equally into 35-mm dishes for different treatments and serum-starved for 18 h. Then the cells were stimulated with EGF (10 ng/ml), serum (10%), or PMA (5 nM) or left in the medium for 6 h. A whole cell extract was prepared, and luciferase activity from 20 l of cell lysates was measured by the luciferase assay system (Promega) with a Luminometer (Lumat LB9501, Berthold). Cells in an additional dish were trypsinized and analyzed in a FCAScan cytometer (EPICS profile analyzer) for monitoring the GFP transfection efficiency, which was used to normalize the luciferase activity. All transfections were performed at least twice in triplicate.
Assay for the Growth Rate of 293 Cells-Cells (2 ϫ 10 4 ) were seeded in 35-mm dishes and cultured in Dulbecco's modified Eagle's medium supplemented with EGF or serum at various concentrations; thereafter, the cells were counted at the indicated intervals.

Overexpression of PTP1C and PTP1C C455S in 293 Cells-
The biological consequences of PTP1C overexpression were examined in cultured human embryonic kidney 293 cells, in which endogenous PTP1C was readily detectable by Western blot assay, although its level is relatively low (see the Western blot in Fig. 4, lower panel). A cysteine residue in the HCSA motif of the protein-tyrosine phosphatase domain is highly conserved among various protein-tyrosine phosphatases and is crucial for their catalytic activity. We therefore generated a mutant, PTP1C C455S, in which Cys-455 was mutated to Ser to inactivate the phosphatase activity. Constructs containing PTP1C, mutant PTP1C C455S, and the vector alone were used to transfect 293 cells. After transfection, the cloned cell lines were isolated by selection with G418. In three independent transfection experiments, it was found that the number of neomycin-resistant clones of mutant PTP1C C455S was consistently much lower than that of wild-type PTP1C or the vector alone. Furthermore, clones containing mutant PTP1C C455S grew much slower than those containing wild-type PTP1C or the vector alone (see below). These results suggest that expression of catalytically inactive PTP1C may have a negative effect on cell growth. Western blots confirmed the high expression levels of both exogenous PTP1C and PTP1C C455S. The phosphatase activity in immunoprecipitates from most PTP1C clones increased Ͼ6-fold when compared with controls (data not shown). In contrast, phosphatase activity precipitated from mutant PTP1C C455S was virtually the same as that from the control cells, confirming that the Cys-to-Ser mutation effectively abolished PTP1C-specific phosphatase activity. When mutant PTP1C C455S was cotransfected with wild-type PTP1C or PTP2C, the phosphatase activity in immunoprecipitates resumed to a level corresponding to those transfected with the wild type alone. Of various cell clones, two wild-type PTP1C (WT3 and WT5) and two mutant PTP1C C455S (C/S4 and C/S7) as well as two PTP1C C455S/PTP1C and PTP1C C455S/PTP2C clones, each obtained from independent transfections, were chosen for further characterization. These clones typically express a high but comparable level of PTP1C or PTP1C C455S or both the mutant and wild-type phosphatases based on immunoblotting and phosphatase activity. All experimental data described below were generated using at least two different clones.
Phosphorylation of PTP1C following EGF Stimulation-293 cells express a substantial amount of EGF receptor, which makes this cell line suitable for study of the EGF-activated signal cascade. Tyrosine phosphorylation of both PTP1C and the inactive mutant PTP1C C455S was readily observed following EGF stimulation. As shown in Fig. 1, PTP1C underwent rapid tyrosine phosphorylation upon stimulation with EGF. Phosphorylation reached a maximum after 3 min following EGF stimulation and declined after 10 min. The extent of tyrosine phosphorylation in mutant PTP1C C455S was usually higher than that observed in wild-type PTP1C in response to EGF (data not shown), due to the auto-dephosphorylation of the wild-type enzyme (24). The phosphatase activity of tyrosine-phosphorylated PTP1C induced by EGF stimulation was not significantly changed relative to unphosphorylated enzyme when assayed in immunoprecipitates prepared with and without EGF treatment (data not shown).
Overexpression of Mutant PTP1C C455S Reduces Cell Growth Rate-During the course of isolating PTP1C and PTP1C C455S clones from 293 cells, we observed that clones expressing the catalytically inactive mutant PTP1C C455S grew much slower than those expressing PTP1C or control cells. To ascertain whether the inactive mutant PTP1C truly had a negative effect on cell growth, we examined the growth rate of control cells as well as cells expressing PTP1C and PTP1C C455S. As shown in Fig. 2, the growth rate of 293 cells responded to either serum or EGF in a dose-dependent manner. At a low serum concentration (1%) or at a low concentra-

PTP1C in Mitogenic Pathways
tion of EGF (1 ng/ml), 293 cells expressing wild-type PTP1C grew approximately twice as fast as the control cells transfected with the vector alone. As the concentrations of the mitogens were increased, the differences between the growth rate diminished. Conversely, 293 cells expressing mutant PTP1C C455S displayed a much slower growth rate when compared with cells expressing wild-type PTP1C or even with control cells. The inhibition of cell growth exerted by expression of PTP1C C455S was observed at both low and high mitogen concentrations. These data indicate that overexpression of the catalytically inactive mutant PTP1C C455S suppresses the growth rate of 293 cells.
Effect of Overexpressed PTP1C and PTP1C C455S on DNA Synthesis-Since the expression of mutant PTP1C C455S inhibited cell growth, we speculated that expression of PTP1C C455S would antagonize the mitogenic potential of EGF and serum factors, thus leading to inhibition of DNA synthesis. Serum-or EGF-stimulated DNA synthesis was monitored by incorporation of [ 3 H]thymidine. DNA synthesis in 293 cells was highly dependent upon growth factor addition. We examined the dose responses of both serum and EGF on stimulation of DNA synthesis in 293 cells expressing the vector, PTP1C, and PTP1C C455S. Fig. 3 shows that the control and PTP1C-transfected cells displayed similar levels of [ 3 H]thymidine incorporation at a high concentration of EGF (10 g/ml) or serum (10%). However, at low concentrations of EGF (1.0 ng/ml) and serum (1%), only the PTP1C-transfected cells showed an appreciable increase in DNA synthesis in response to the mitogens. This response correlates to what was observed for cell growth (see Fig. 2). In contrast, G 0 -arrested cells transfected with PTP1C C455S responded very poorly to either EGF or serum. Even at a high concentration of EGF (10 ng/ml) or serum (10%), the DNA synthesis rate of these cells hardly reached 30% of that observed in control cells during a 12-h period. These experiments demonstrate that the inactive mutant PTP1C C455S exerts a strong inhibitory effect on EGFstimulated DNA synthesis.
Effects of Overexpressed PTP1C and PTP1C C455S on Phosphorylation of MAP Kinase and on MAP Kinase Kinase Activity-The MAP kinase pathway plays a central role in the regulation of cell proliferation. To study where the inactive mutant PTP1C C455S was exerting its inhibitory effect in the mitogenactivated signaling pathway, leading from EGF receptor to SRE stimulation, we further examined the effect of mutant PTP1C C455S on MAP kinase phosphorylation and on MAP kinase kinase activity. It is known that upon stimulation with mitogens, MAP kinase is activated by phosphorylation on both tyrosine and threonine residues, resulting in mobility shifts of MAP kinase (ERK1 and ERK2) on SDS-polyacrylamide gel. Cells expressing wild-type PTP1C and mutant PTP1C were treated with EGF. As shown in Fig. 4, in control cells, following a 5-min stimulation with EGF, a substantial proportion of both ERK1 and ERK2 displayed the typical mobility shift on the gel due to the phosphorylation of the enzymes. In cells transfected with wild-type PTP1C, the EGF-stimulated phosphorylation of both ERK1 and ERK2 was significantly increased under the same conditions. Conversely, expression of mutant PTP1C C455S dramatically reduced the mobility shifts of both ERK1 and ERK2 on the gel in response to EGF (Fig. 4).
The effects of different forms of PTP1C on the MAP kinase cascade were further examined by directly measuring MAP kinase kinase activity. The activity assay was performed 5 min after EGF stimulation. The time courses showed that the peak activity for all transfected cells occurred at approximately this point (data not shown). Expression of wild-type PTP1C slightly increased the MAP kinase kinase activity (ϳ15%) when compared with control cells. In contrast, overexpression of the catalytically inactive mutant PTP1C C455S decreased MAP kinase kinase activity by ϳ30% (Fig. 5).
Effect of Overexpressed PTP1C and PTP1C C455S on c-fos Promoter-dependent Luciferase Activity-To further determine the role of PTP1C in growth factor-stimulated mitogenic signal transduction, we assayed the effect of PTP1C on c-fos promoter- dependent luciferase activity. The luciferase gene driven by the human c-fos promoter represents a highly sensitive reporter of growth factor-induced transcriptional activity (31). This early gene promoter contains the well characterized SRE, whose activity is induced upon activation of the serum-responsive factor (32). In control cells, SRE-regulated gene expression was dramatically stimulated by either serum or EGF. The stimulatory effect of EGF was approximately three times greater than that of serum (Fig. 6). The stimulatory response to EGF was increased by ϳ30% with the expression of wild-type PTP1C. In contrast, expression of the catalytically inactive mutant PTP1C C455S strongly suppressed the stimulatory effect of both EGF and serum. As shown in Fig. 6, expression of mutant PTP1C C455S inhibited luciferase activity by ϳ50% in response to serum and by nearly 50% in response to EGF when compared with the control (Fig. 6). However, the inactive mutant PTP1C C455S did not affect the stimulatory effect of PMA on the SRE-Luc activity (Fig. 6, lower panel), suggesting that the inhibitory effect of mutant PTP1C C455S on SRE-Luc reporter gene activity is mitogen-specific. Thus, all experiments, as assayed by growth rate, SRE-Luc activity, MAP kinase, and MAP kinase kinase activity, clearly demonstrated that overexpression of the catalytically inactive mutant PTP1C C455S in 293 cells strongly suppressed the mitogenic responses.

Coexpression of Wild-type PTP1C Overcomes the Inhibitory Effect of PTP1C C455S on c-fos Promoter-dependent Luciferase
Activity-To confirm whether the inhibitory effect of PTP1C C455S on SRE-Luc activity was due to a specific inhibition of the endogenous PTP1C activity, rather than to an effect of the related phosphatase PTP2C, cells stably expressing PTP1C C455S were cotransfected with either wild-type PTP1C or PTP2C. As shown in Fig. 7, coexpression of wild-type PTP1C with PTP1C C455S greatly elevated the SRE-Luc activity to a level nearly comparable to that observed in cells overexpressing PTP1C alone (see the SRE-Luc activity with EGF stimulation in Fig. 6), totally abolishing the inhibitory effect of the mutant. However, coexpression of PTP2C with PTP1C C455S only partially reversed the inhibitory effect of PTP1C C455S. These results suggest that only PTP1C can specifically overcome the inhibitory effect of mutant PTP1C C455S. A moderate increase of SRE-Luc activity in cells cotransfected with PTP2C is very likely due to the intrinsic positive effect of PTP2C, rather than to any specific competition with mutant PTP1C C455S in signaling. In the presence of the dominant negative form of PTP1C, coexpression of PTP2C resulted in an ϳ20% increase of SRE-Luc activity (Fig. 7), similar to the increase of mitogenic activity in response to EGF stimulation observed in 293 cells expressing PTP2C alone as demonstrated previously (21). Since nearly the same increase in response to EGF was caused by PTP2C in the presence or absence of the dominant negative form of PTP1C C455S, PTP1C and PTP2C appear to operate via distinct signaling pathways.
Expression of the Inactive Mutant PTP1C C455S Induces the Association of the Mutant Protein with Specific Tyrosine-phosphorylated Proteins-Since the inactive mutant PTP1C C455S markedly blocked mitogenic signals, it is possible that tyrosinephosphorylated protein(s) might be targeted by PTP1C. To identify potential substrates for this enzyme, cell lysates were prepared from control cells, PTP1C-transfected cells, and mutant PTP1C C455S-transfected cells. The lysates were directly subjected to Western blot probing with anti-phosphotyrosine antibody. In parallel, the lysates were first precipitated by anti-PTP1C antibody, and the precipitates were then subjected to Western blot analysis with anti-phosphotyrosine antibody. As shown in Fig. 8 (left panel), many tyrosine-phosphorylated proteins were detected in the cell lysates. Comparison of the tyrosine phosphorylation pattern of the lysate of PTP1C cells (lane 2) with that of mutant PTP1C C455S (lane 3) showed that a strongly tyrosine-phosphorylated protein band of 95 kDa was present in the lysates of cells expressing mutant PTP1C C455S (lane 3). This protein was also detected in control cells (lane 1), but the extent of tyrosine phosphorylation of the protein was dramatically reduced. Co-immunoprecipitation further showed that the 95-kDa protein was associated with the catalytically inactive mutant protein (right panel, lane 3), but not with wild-type PTP1C (lane 2). A minor protein band of 100 kDa was also found in the immunoprecipitates. Similarly, the in vitro binding experiment also showed that the 95-kDa band bound to the glutathione S-transferase-SH2 fusion protein of PTP1C was detected only in the cell lysate of the PTP1C C455Stransfected cells, but not in that of the wild-type PTP1C-transfected cells (data not shown). These results suggest that PTP1C may initially interact with the 95-kDa protein through its SH2 domains and subsequently dephosphorylate the protein on tyrosine residues and dissociate from it. Interaction of these proteins with the inactive mutant enzyme did not result in tyrosine dephosphorylation, thus avoiding dissociation from the mutant phosphatase. DISCUSSION In our analysis of the PTP1C gene, we have recently shown that the expression of PTP1C in hematopoietic cells versus non-hematopoietic cells is regulated by two tissue-specific promoters. The hematopoietic form of the PTP1C transcript is initiated exclusively from a specific promoter (P2), whereas in non-hematopoietic cells, an upstream promoter (P1) is transcriptionally activated (33). Indeed, PTP1C has been detected in a variety of non-hematopoietic cells (5,12,30), from which PTP1C was initially identified (5). Recent studies have further established that in hematopoietic cells, PTP1C negatively regulates a number of major signaling pathways. For instance, PTP1C inhibited B cell signaling and turned off erythropoietinstimulated hematopoiesis (27,28). However, the function of PTP1C in non-hematopoietic cells is largely unknown. Since the EGF receptor is expressed in many non-hematopoietic cells, but is not found in hematopoietic cells, we used EGF as a model system to investigate the effect of PTP1C on mitogenic pathways in 293 cells.
The function of PTP1C in EGF-and serum-activated mitogenic pathways was investigated by expression of wild-type PTP1C and its inactive mutant form in 293 cells. Overexpression of a wild-type PTP1C slightly enhanced EGF-and serumactivated mitogenesis as compared with the control cells. The positive effect of PTP1C on cell growth and DNA synthesis was more apparent at low concentrations of serum and EGF (see Figs. 2 and 3). It is possible that at high concentrations of the mitogens, other elements may have been activated in multiple signal transduction pathways that might mask the positive function of PTP1C. Most important, overexpression of a catalytically inactive mutant of PTP1C strongly suppressed all mitogen-activated pathways, as measured by cell growth, DNA synthesis, early gene transcription, MAP kinase phosphorylation, and MAP kinase kinase activity. Taken together, the results suggest that, in contrast with its negative regulatory roles observed in hematopoietic cells, the intrinsic phosphatase activity of PTP1C in 293 cells likely functions as a positive regulator in EGF-or serum-stimulated mitogenesis. Interestingly, the catalytically inactive mutant of PTP2C, a structural homologue of PTP1C, displayed a similar positive effect on mitogenesis activated by several growth factors in non-hematopoietic cells (18 -21). However, although PTP1C and PTP2C share a great deal of sequence homology, two lines of evidence FIG. 7. Effect of mutant PTP1C C455S cotransfected with wildtype PTP1C or PTP2C on SRE-Luc activity. Cells stably expressing PTP1C C455S (PTP1C(C-S)) were cotransfected either with wild-type PTP1C or with PTP2C and selected by an additional marker, hygromycin B. Cells were transiently transfected with the SRE-Luc and GFP reporter plasmids and treated for EGF stimulation as described for Fig.  6. The SRE-Luc activity is presented as the intensity of fluorescent lights. Data are presented as means Ϯ S.E. from three independent experiments, each performed in triplicate. suggested that the positive effect of overexpressed PTP1C on mitogen-activated pathways in 293 cells was not due to competition with PTP2C for the same phosphotyrosine-binding site(s) on the target. First, in similar experiments, the catalytically inactive mutant of PTP2C in 293 cells bound to a 47-kDa protein (21), rather than to the 95-kDa protein shown to bind to the inactive PTP1C here, indicating that the two SH2 domaincontaining phosphatases target different molecules. Second, the inhibitory function of the overexpressed catalytically inactive mutant of PTP1C in EGF-stimulated SRE-Luc activity was nearly completely abolished by wild-type PTP1C. In this experiment, PTP2C caused an increase in response to EGF nearly identical to that observed in the absence of the dominant negative form of PTP1C (21). Together, these results suggest that in 293 cells, the positive effects of overexpressed PTP1C and PTP2C on EGF-stimulated mitogenesis arise through different signaling pathways.
Although our data suggest that PTP1C, like PTP2C, functions as a positive regulator in the mitogen-activated Ras-Raf-MAP kinase pathway in 293 cells, the mechanism by which the inactive mutant PTP1C attenuated the EGF-and serum-activated signaling is unknown. It is known that PMA, a potent activator of protein kinase C, induces MAP kinase activation in a Ras-independent manner (34). Since mutant PTP1C strongly suppressed the serum-and EGF-stimulated activity of the SRE-Luc reporter gene, but did not interfere with the PMA activation of the gene, it is very likely that PTP1C C455S targets an upstream molecule in the Ras-MAP kinase pathway, rather MAP kinase itself or its downstream molecules. This molecule may link Ras-Raf activation as suggested for the positive role of PTP2C in insulin-stimulated signaling (19,20). Alternatively, tyrosine-phosphorylated PTP1C, as with PTP2C (35), can serve as a docking site for Grb2 binding (24). Formation of the PTP1C-Grb2 complex may activate the Ras-Raf-MAP kinase pathway. However, expression of the catalytically inactive mutant PTP1C C455S appeared to have no significant effect on complex formation between Grb2 and the EGF receptor, and site-directed mutagenesis experiments also showed that eliminating the Grb2-binding site on PTP1C did not reduce the mitogenic responses. 2 Ideally, as a positive inducer in signaling, PTP1C (or PTP2C) may dephosphorylate a molecule on its phosphotyrosine site(s) that negatively regulate its function, such as the kinases of the Src family. In searching for potential targets for PTP1C, we have found a tyrosine-hyperphosphorylated 95-kDa protein in cells expressing the inactive mutant PTP1C. This protein was co-immunoprecipitated with mutant PTP1C C455S, but not with the active phosphatase. Since the 95-kDa protein bound to the glutathione S-transferase-SH2 fusion protein of PTP1C in vitro, it is likely that PTP1C associates with the 95-kDa protein through its SH2 domain(s). These results suggest that this tyrosine-phosphorylated 95-kDa protein might be one of the downstream targets of PTP1C in signaling.
The demonstration of the positive regulatory function for PTP1C in mitogenesis as described here suggests a more complicated role for these SH2-containing protein-tyrosine phosphatases than previously considered, such as the view that PTP1C plays only a negative role in hematopoietic cells and, conversely, that PTP2C, a mammalian homologue of corkscrew, the Drosophila csw gene product, functions as a positive regulator. In fact, we have observed that PTP2C, in contrast with its positive function in mitogen-activated pathways, plays a negative role in membrane ruffling signaling (36). The negative function of PTP2C in regulating EGF-dependent cell growth was also recently reported (37). A positive regulatory function for PTP1C has also been suggested for the thrombin-activated mitogen pathway in blood platelets (26). On the other hand, PTP1C was found to down-regulate the proliferation of v-Srctransformed rat fibroblast cells (38). Thus, the results presented here, together with previous data, imply that PTP1C participates in complicated and multiple signaling pathways, where, depending on cell types, its compartmentalization, and which molecules it interacts with, the enzyme can play either a positive or negative role in regulating individual signal transduction pathways.