The Protein-tyrosine Phosphatase TCPTP Regulates Epidermal Growth Factor Receptor-mediated and Phosphatidylinositol 3-Kinase-dependent Signaling*

In this study we have investigated the down-regulation of epidermal growth factor (EGF) receptor signaling by protein-tyrosine phosphatases (PTPs) in COS1 cells. The 45-kDa variant of the PTP TCPTP (TC45) exits the nucleus upon EGF receptor activation and recognizes the EGF receptor as a cellular substrate. We report that TC45 inhibits the EGF-dependent activation of the c-Jun N-terminal kinase, but does not alter the activation of extracellular signal-regulated kinase 2. These data demonstrate that TC45 can regulate selectively mitogen-activated protein kinase signaling pathways emanating from the EGF receptor. In EGF receptor-mediated signaling, the protein kinase PKB/Akt and the mitogen-activated protein kinase c-Jun N-terminal kinase, but not extracellular signal-regulated kinase 2, function downstream of phosphatidylinositol 3-kinase (PI 3-kinase). We have found that TC45 and the TC45-D182A mutant, which is capable of forming stable complexes with TC45 substrates, inhibit almost completely the EGF-dependent activation of PI 3-kinase and PKB/Akt. TC45 and TC45-D182A act upstream of PI 3-kinase, most likely by inhibiting the recruitment of the p85 regulatory subunit of PI 3-kinase by the EGF receptor. Recent studies have indicated that the EGF receptor can be activated in the absence of EGF following integrin ligation. We find that the integrin-mediated activation of PKB/Akt in COS1 cells is abrogated by the specific EGF receptor protein-tyrosine kinase inhibitor tyrphostin AG1478, and that TC45 and TC45-D182A can inhibit activation of PKB/Akt following the attachment of COS1 cells to fibronectin. Thus, TC45 may serve as a negative regulator of growth factor or integrin-induced, EGF receptor-mediated PI 3-kinase signaling.

turally diverse family of enzymes, characterized by the consensus sequence (I/V)HCXAGXXR. They are found in eukaryotes, prokaryotes and viruses and can either antagonize or potentiate protein-tyrosine kinase (PTK)-dependent signaling. PTPs have been shown to participate as either positive or negative regulators of signal transduction in a wide range of physiological processes, which include cellular growth and proliferation, migration, differentiation and survival (1)(2)(3). Despite their important roles in such fundamental physiological processes, the mechanism by which PTPs exert their effects is often poorly understood.
The human T-Cell PTP (TCPTP) is an intracellular nontransmembrane phosphatase that was originally cloned from a T-cell cDNA library but is now known to be expressed in many tissues. TCPTP contains a conserved catalytic domain and a non-catalytic C-terminal segment that varies in size and function as a result of alternative splicing. Two splice variants differing only in their extreme C termini are expressed. The 48-kDa form of human TCPTP (TC48) contains a 34-residue hydrophobic tail, which is replaced by a hydrophilic 6-residue sequence in the 45-kDa form (TC45). TC48 localizes to the endoplasmic reticulum (ER) (4,5), whereas under basal conditions TC45 is localized in the nucleus due to the presence of a bipartite nuclear localization sequence (5)(6)(7)(8)(9).
All PTPs contain an aspartic acid that is essential for catalysis. Mutation of this residue, Asp-182 in TCPTP, to alanine reduces the catalytic activity but maintains a high affinity for substrates, thereby generating a "substrate trapping" mutant, which can form stable complexes with tyrosine-phosphorylated proteins in vitro (10) and in vivo (11,12). Using the TCPTP D182A substrate trapping mutants, we have shown previously that TCPTP displays a restricted specificity in a cellular context, and that the EGF receptor is one of its substrates (12). Both TC48 and TC45 recognize the tyrosine-phosphorylated EGF receptor as a substrate in a cellular context; TC48 recognizes the receptor as it proceeds through the ER and may function to prevent inappropriate signaling by the nascent receptor during synthesis, whereas TC45 can exit the nucleus in response to EGF and gain access to signaling complexes containing the EGF receptor at the plasma membrane (12).
In the present study we have examined the effect of overexpression of TC45 on EGF receptor-induced signaling events. We show that TC45 can inhibit the EGF-induced activation of PKB/Akt and that this correlates with a reduced association of PI 3-kinase with the activated EGF receptor. In addition, we show that plating of COS1 cells on fibronectin leads to activation of PKB/Akt, in an EGF-independent but EGF receptor and PI 3-kinase-dependent manner, which is also inhibited by TC45. Thus, TC45 may serve as a negative regulator of specific signals from the EGF receptor that mediate PKB/Akt activation.

Cell Culture, Transfections, and Electroporations
COS1 cells were cultured at 37°C and 5% CO 2 in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum (FBS), 100 units/ml penicillin, and 100 g/ml streptomycin. Where indicated, COS1 cells were serum-starved for 24 h in DMEM containing 0.1% FBS, plus antibiotics. COS1 cells were transfected by the calcium phosphate precipitation method as described previously (12). Unless otherwise indicated cells were transfected using TCPTP pMT2 plasmid DNA at 20 g/10-cm dish. Cells were washed three times with phosphate buffered saline (PBS) at 5-6 h after transfection and supplemented with fresh DMEM containing 10% FBS. Where indicated approximately 1-2 ϫ 10 6 COS1 cells were electroporated in 250 l of medium with 20 g of TCPTP pMT2 plasmid at 200 V and 950 microfarads in 0.4-cm cuvettes and seeded into a 10-cm dish. Transfected or electroporated cells were collected at 36 -48 h after transfection, or washed once with PBS at 24 h after transfection, supplemented with DMEM containing 0.1% FBS and processed at 48 h after transfection. The efficiency of electroporation, as assessed by 5-bromo-4-chloro-3-indolyl ␤-D-galactosidase staining of pCMV-␤-galactosidase-electroporated COS1 cells, was routinely 50 -75%.

Immune Complex Kinase Reactions
ERK2 and JNK Assays-COS1 cells, in 10-cm dishes, were transfected with 2 g of HA-tagged ERK2 pJ3H plasmid or 5 g of FLAGtagged JNK pCMV plasmid and either 15 g of pMT2 plasmid or 15 g of the TC45 or TC45-D182A pMT2 plasmids. At 24 h after transfection, cells were washed once with PBS and serum-starved overnight for 24 h. Cells were then left unstimulated or stimulated with 100 ng/ml EGF for 15 min and processed for either ERK2 kinase assays as described previously (12) or JNK assays. For JNK assays, cells were lysed in 0.9 ml of immunoprecipitation (IP) lysis buffer (50 mM Tris, pH 7.5, 1% w/v Nonidet P-40, 150 mM NaCl, 50 mM NaF, 1 mM vanadate, 5 g/ml leupeptin, 5 g/ml aprotinin, 1 g/ml pepstatin A, 1 mM benzamidine, and 2 mM phenylmethylsulfonyl fluoride), centrifuged (12,000 ϫ g for 10 min at 4°C) and FLAG-tagged JNK immunoprecipitated from the supernatant with 5 g of anti-FLAG M2 antibody for 90 min at 4°C. Immune complexes were collected on protein G-Sepharose for 30 min at 4°C and JNK kinase activity measured using GST-Jun (GST fused to the N terminus of c-Jun) as substrate as described previously (13). Anti-FLAG immune complexes from these reactions were resolved by SDS-PAGE, immunoblotted with anti-FLAG antibodies, and quantitated by densitometry in order to normalize for FLAG-JNK activity.
PI 3-Kinase Assays-COS1 cells were electroporated with either pMT2 vector control, TC45 or TC45-D182A plasmids as indicated above. At 24 h after electroporation, cells were serum-starved for 24 h and then stimulated with EGF (100 ng/ml) for 15 min. Cells were lysed in 0.9 ml of IP lysis buffer, and lysates were precleared with 0.1 ml of Pansorbin (Cambridge, MA) for 60 min at 4°C. Precleared lysates were subsequently centrifuged (12,000 ϫ g for 10 min at 4°C), and the phosphotyrosine-containing proteins were immunoprecipitated over-night at 4°C under constant mixing, using a mixture of the antiphosphotyrosine antibodies G98 and G104 (20 l of G98 and 20 l of G104 ascites for every 10-cm dish of cells) (10,12). Immune complexes were collected on protein A-Sepharose CL-4B (Amersham Pharmacia Biotech, Uppsala, Sweden) for 60 min at 4°C, washed four times with IP lysis buffer; two times with 100 mM Tris, pH 7.5, buffer containing 100 mM NaCl and 1 mM EDTA; and three times with PI 3-kinase buffer (20 mM HEPES, pH 7.5, 5 mM MgCl 2 , 1 mM EGTA). PI 3-kinase assays were performed in 100 l of PI 3-kinase buffer containing 100 M ATP (plus 10 Ci of [␥-32 P]ATP) and 50 g of sonicated crude brain lipids (containing approximately 10% phosphatidylinositol) (14) for 20 min at room temperature. Reactions were stopped with 100 l of 1 M HCl, and the phospholipids were extracted and separated by thin layer chromatography on silica plates coated with potassium oxalate. [ 32 P]Phosphatidylinositol was quantitated on a PhosphorImager using Image-Quant software (Molecular Dynamics).
PKB/Akt Assays-COS1 cells, in 10-cm dishes, were transfected with 5 g of HA-tagged PKB/Akt pECE plasmid, with or without 5 g of p110K227E pSG5 plasmid, and either 15 g of the pMT2 plasmid or 15 g of the TC45 or TC45-D182A pMT2 plasmids. At 24 h after transfection, cells were serum-starved and either left unstimulated or stimulated with EGF (100 ng/ml) for 15 min. Cells were then lysed in 0.9 ml of IP lysis buffer, the lysates centrifuged (12,000 ϫ g for 10 min at 4°C) and HA-PKB/Akt immunoprecipitated from the supernatant with anti-HA antibody 12CA5 (2 l of ascites/10-cm dish of cells) for 2 h at 4°C. Immune complexes were collected on protein A-Sepharose CL-4B for 60 min at 4°C, washed three times with IP lysis buffer, and washed two times with PKB/Akt kinase buffer (50 mM Tris, pH 7.5, 10 mM MgCl 2 , 1 mM dithiothreitol). PKB/Akt kinase activity was assayed using the peptide substrate RPRAATF-NH 2 (15) by incubating the immunoprecipitated HA-PKB/Akt for 10 or 20 min at 30°C in 30 l of PKB/Akt kinase buffer containing 50 M ATP (plus 0.25 M [␥-32 P]ATP) and 50 M peptide substrate. The reaction was terminated by adding 15 l of 0.5 M EDTA and after brief centrifugation 20 l of the supernatant was spotted onto P-81 phosphocellulose paper. Unincorporated [␥-32 P]ATP was eliminated by three 5-min washes in 75 mM orthophosphoric acid and phosphorylated peptide bound to the paper counted. Anti-HA immune complexes from the kinase reactions were resolved by SDS-PAGE, immunoblotted with anti-HA antibodies, and quantitated by densitometry in order to normalize for HA-PKB/Akt activity. Transfections for PKB/Akt assays were conducted either in duplicate or triplicate.

Cell Stimulation with Fibronectin
COS1 cells were electroporated with pMT2 vector control, TC45, or TC45-D182A expression plasmids as indicated above. At 24 h after electroporation, cells were serum-starved for 24 h, washed in PBS, and then harvested by limited trypsin-EDTA treatment (1 ml of 0.025% trypsin plus 5 mM EDTA in DMEM minus phenol red/10-cm dish of cells). After trypsin inhibition by soybean trypsin inhibitor (1 mg of chromatographically purified type I-S trypsin inhibitor (Sigma)/10-cm dish of cells) in DMEM minus phenol red containing 0.25% (w/v) bovine serum albumin (BSA) (radioimmunoassay grade, fraction V from Sigma), the cells were pelleted by centrifugation and washed twice with DMEM minus phenol red containing 0.25% (w/v) BSA and resuspended in DMEM minus phenol red containing 0.1% (w/v) BSA. The cells were held in suspension at 37°C for 30 min prior to attachment for 1 h onto tissue culture dishes precoated with fibronectin (5 ml of 10 g/ml human plasma fibronectin/10-cm dish incubated overnight at 4°C and then washed once with DMEM minus phenol red and warmed to 37°C for 30 min). Attached cells were rinsed twice in DMEM minus phenol red containing 0.1% (w/v) BSA, once in ice-cold PBS, and then collected in hot 3ϫ Laemmli sample buffer containing 6% (v/v) ␤-mercaptoethanol. Proteins were resolved by SDS-PAGE and immunoblotted as indicated.

Selectivity of TC45 on EGF Receptor-mediated Mitogen-activated Protein Kinase (MAPK)
Signaling-In response to stimulation with EGF, the TC45-D182A substrate trapping mutant forms a stable complex with the EGF receptor (12). At early time points after EGF stimulation, the TC45-D182A mutant accumulates at the cell periphery and colocalizes with the EGF receptor ( Fig. 1), consistent with recognition of the tyrosinephosphorylated EGF receptor at the plasma membrane. To examine the role of TC45 in EGF receptor signaling, we have overexpressed wild type TC45 and the TC45-D182A mutant in COS1 cells and measured their effects on EGF-dependent activation of the MAPKs, extracellular signal-regulated kinase 2 (ERK2) and c-Jun N-terminal kinase (JNK) (Fig. 2). Consistent with our previous studies, we observed no apparent effect of TC45 on the activation of HA-tagged ERK2 (12). However, coexpression of TC45 inhibited the activation of FLAG-tagged JNK by 38.3 Ϯ 7.8%. These results illustrate the ability of TC45 to regulate selectively signaling events emanating from the EGF receptor. Recent studies have indicated that EGFinduced activation of JNK, but not ERK2, is mediated by PI 3-kinase (16,17). Consistent with these studies, we found that the PI 3-kinase inhibitor wortmannin had no effect on EGFinduced activation of ERK2 (Fig. 2, panel A, inset), but inhibited the activation of JNK by 57.1 Ϯ 9.4% (Fig. 2, panel B). These results suggest that PI 3-kinase may be a target by which TC45 mediates its effects on the activation of JNK.
TC45 Inhibits the EGF-induced Activation of the Protein Kinase PKB/Akt-Although the mechanism by which PI 3-kinase regulates the activation of JNK is not known, the ability of PI 3-kinase to regulate the activation of the protein kinase PKB/Akt, which mediates the effects of PI 3-kinase on cell proliferation and survival, is well documented (18 -24). The lipid product of PI 3-kinase, phosphatidylinositol-3,4,5-triphosphate, binds to the pleckstrin homology domain of PKB/Akt and recruits it to the plasma membrane, where it is further activated by phosphorylation of Thr-308 and Ser-473. To determine whether TC45 regulates PI 3-kinase-mediated signaling processes, we examined the effects of TC45 and TC45-D182A on the EGF-induced stimulation of PKB/Akt, by coexpressing HA-tagged PKB/Akt with either TC45 or TC45-D182A in COS1 cells. The activity of HA-immunoprecipitated PKB/Akt was measured in vitro using the specific peptide substrate RPRAATF (15). Both wild the type TC45 and the TC45-D182A mutant inhibited the EGF-induced activation of PKB/Akt by approximately 60 -70% (Fig. 3). Importantly, under similar conditions, expression of the ER-localized spliced variant TC48 had no apparent effect (Fig. 3, panel B). These results demonstrate that simple overexpression of a PTP is not sufficient to inhibit PKB/Akt activation and underscores the importance of proper subcellular localization for the control of TCPTP action.
TC45 Inhibits EGF-induced PI 3-Kinase Activity-As PKB/ Akt is not tyrosine-phosphorylated it cannot serve as a substrate of TC45. Consequently, it is likely that TC45 exerts its effects on PKB/Akt, and possibly JNK, at a step upstream in the signaling pathway by regulating negatively the EGF-induced activation of PI-3 kinase. PI 3-kinase is activated by associating with sites of tyrosine phosphorylation generated by activated receptor tyrosine kinases (25)(26)(27)(28)(29). Therefore, PI 3-kinase activity can be measured in anti-phosphotyrosine immunoprecipitates to determine the extent by which it is stimulated in response to EGF (29). To examine the effects of TC45 on PI 3-kinase activity, we electroporated COS1 cells with expression plasmids for either vector control, TC45, or TC45-D182A and measured PI 3-kinase activity in anti-phosphotyrosine immunoprecipitates. Under these conditions TC45 and TC45-D182A inhibited the EGF-induced PI 3-kinase activity by ϳ50% (Fig.  4, panel A). Furthermore, we observed that the phosphorylation of PKB/Akt on serine 473 was inhibited by a similar extent, as measured in lysates of these electroporated cells (Fig. 4, panel B). Since the electroporation efficiency in these experiments was also in the order of 50 -75% (data not shown), these results suggest that overexpression of TC45 can inhibit almost completely the activation of PI 3-kinase and PKB/Akt. These results are consistent with the ϳ70% inhibition of the EGF-induced activity of HA-tagged PKB/Akt (Fig. 3).
TC45 Acts Upstream of PI 3-Kinase and Inhibits the Recruitment of PI 3-Kinase to the EGF Receptor-To examine whether TC45 can inhibit the activation PKB/Akt independently of its effects on PI 3-kinase, we cotransfected COS1 cells with expression plasmids encoding HA-tagged PKB/Akt and a constitutively active form of the PI 3-kinase catalytic subunit, p110K227E (30), together with either vector control, TC45 or TC45-D182A and measured the effects on HA-PKB/Akt activity. Expression of TC45 and TC45-D182A did not inhibit the p110K227E-induced activation of PKB/Akt either in the absence or presence of EGF (Fig. 5). Therefore, these data illustrate that TC45 must act to suppress the activation of PI FIG. 1. TC45-D182A localizes to the cell periphery in response to EGF and colocalizes with the EGF receptor. COS1 cells were transfected by the calcium phosphate precipitation method with TC45-D182A (TC45D). At 24 h after transfection, cells were serum-starved and either left unstimulated or stimulated with EGF (100 ng/ml). A, serumstarved cells were stimulated for 5 min or 30 min and then processed for immunofluorescence using the TCPTP antibody CF4, as described previously (12). B, serum-starved cells were stimulated for 5 min and processed for confocal microscopy using TCPTP (CF4) and EGF receptor (EGFR) specific antibodies.
3-kinase, thereby inhibiting the downstream activation of PKB/Akt. These observations raise the issue of how TC45 suppresses the activation of PI 3-kinase. One possibility is that TC45 acts on the EGF receptor, to inhibit the recruitment of the PI 3-kinase p85 regulatory subunit, thereby preventing PI 3-kinase activation and the subsequent activation of PKB/Akt. Whereas the wild type TC45 has the capacity to dephosphorylate the EGF receptor, the TC45-D182A substrate trapping mutant would bind to phosphotyrosine sites on the EGF receptor, thus competing with SH2 domain-containing signaling molecules and thereby interfering with concomitant PI 3-kinase activation. We investigated whether expression of TC45 could inhibit the recruitment of the p85 PI 3-kinase subunit to the EGF receptor (Fig. 6). First, we examined the ability of TC45 to reduce the amount of p85 in anti-phosphotyrosine antibody immunoprecipitates from EGF-stimulated COS1 cells that had been electroporated with vector control, TC45 or TC45-D182A expression plasmids. Following EGF stimulation of control cells, the p85 regulatory subunit of PI 3-kinase could be detected in anti-phosphotyrosine immunoprecipitates. However, tyrosine-phosphorylated p85 was not detectable is cell lysates or p85 immunoprecipitates from EGF-stimulated cells (data not shown), suggesting that the p85 in anti-phosphotyrosine immunoprecipitates was associated with other phosphotyrosine-containing proteins. We found that wild type and mutant TC45 significantly decreased the amount of p85 in the anti-phosphotyrosine antibody immunoprecipitates (Fig. 6,  panel A) and the extent of this inhibition was similar to the observed decrease in PI 3-kinase activity in anti-phosphotyrosine antibody immunoprecipitates (Fig. 4, panel A). Since one

FIG. 2. Selective effects of TC45 on MAPK signaling pathways: TC45 inhibits the EGF-induced activation of JNK but not ERK2.
A, COS1 cells were cotransfected with plasmids expressing HA-tagged ERK2 and pMT2 vector control, TC45, or TC45-D182A (TC45D). Transfected cells were serum-starved and either left unstimulated or stimulated with EGF (100 ng/ml) for 15 min. Cells were then lysed, and HA-tagged ERK2 was precipitated and assayed using myelin basic protein as a substrate as described previously (12). Values shown are arbitrary units and are means Ϯ standard errors of three independent experiments involving duplicate transfections. Inset, COS1 cells were serum-starved, left untreated, or treated with 100 nM wortmannin (Wort) for 60 min and then stimulated with EGF (100 ng/ml) for 15 min. Cell lysates were probed with an anti-ERK2 antibody. B, COS1 cells were cotransfected with plasmids expressing FLAG-tagged JNK and either vector control, TC45, or TC45-D182A (TC45D). At 24 h after transfection cells were serum-starved in DMEM containing 0.1% FBS for 24 h, left untreated or treated with 100 nM wortmannin for 60 min, and then either left unstimulated or stimulated with EGF (100 ng/ml) for 15 min. Cells were then lysed and FLAG-tagged JNK was immunoprecipitated and assayed using GST-Jun. Anti-FLAG immune complexes from these reactions were resolved by SDS-PAGE, immunoblotted with anti-FLAG antibodies, and quantitated by densitometry in order to normalize for FLAG-JNK expression. Values shown are arbitrary units and are means Ϯ standard errors of three independent experiments with duplicate transfections .   FIG. 3. TC45, but not TC48, inhibits the EGF-induced activation of the protein kinase PKB/Akt. A, COS1 cells were cotransfected with plasmids expressing HA-tagged PKB/Akt and pMT2 vector control, TC45, or TC45-D182A (TC45D). At 24 h after transfection cells were serum-starved and either left unstimulated or stimulated with EGF (100 ng/ml) for 15 min. Cells were then lysed and HA-tagged PKB/Akt was precipitated and assayed using peptide substrate. Anti-HA immune complexes from these reactions were resolved by SDS-PAGE, immunoblotted with anti-HA antibodies, and quantitated by densitometry in order to normalize for HA-PKB/Akt activity. Values shown are arbitrary units and are means Ϯ standard errors of three independent experiments. B, COS1 cells in 10-cm dishes were cotransfected with 5 g of plasmid expressing HA-tagged PKB/Akt and 5 g of pMT2 vector control, the 48-kDa isoform of TCPTP (TC48), or TC45. Transfected cells were serum-starved and either left unstimulated or stimulated with EGF (100 ng/ml) for 15 min. Cells were then lysed, and HA-tagged PKB/Akt was precipitated and assayed using peptide substrate. Activity was normalized for HA-tagged PKB/Akt protein. Inset, representative lysates from cells transfected with pMT2 vector control, TC48, or TC45 were resolved by SDS-PAGE and immunoblotted with the TCPTP antibody CF4. of the major phosphotyrosine-containing protein in anti-phosphotyrosine antibody immunoprecipitates from EGF stimulated COS1 cells is the EGF receptor (12), the data infer that TC45 inhibits the recruitment of p85 to the EGF receptor. This was confirmed directly by immunoprecipitating the EGF receptor from COS1 cells that had been electroporated with either vector control, TC45 or TC45-D182A expression plasmids. Expression of either TC45 or TC45-D182A inhibited the EGFinduced association of the p85 regulatory subunit with the EGF receptor (Fig. 6, panel B).
TC45 Inhibits the Integrin-induced and EGF Receptor-mediated Activation of PKB/Akt but Not ERK2-Moro et al. (31) have shown recently that in human primary skin fibroblasts and in ECV304 endothelial cells, integrins can utilize the EGF receptor to transduce extracellular matrix-induced signaling pathways. This integrin-mediated activation of the EGF receptor is independent of EGF. In the present study, we have demonstrated that COS1 cells can activate the EGF receptor, PKB/Akt and ERK2 when plated on fibronectin (Fig. 7). The activation of PKB/Akt and ERK2 can be inhibited by tyrphostin AG1478 (Fig. 7), which is a specific inhibitor of the EGF receptor (32), suggesting that the activation of PKB/Akt and ERK2 following integrin ligation is mediated by the EGF receptor.
Moreover, as in the case of EGF-mediated signaling, activation of PKB/Akt, but not ERK2, is PI 3-kinase-dependent and is inhibited by wortmannin (Fig. 7). We examined whether TC45 could inhibit the EGF receptor-mediated activation of PKB/Akt following integrin ligation. COS1 cells electroporated with vector control, TC45, or TC45-D182A expression plasmids were serum-starved and then detached and replated onto fibronectin-coated dishes. First, we examined whether the EGF receptor could serve as a TC45 substrate following cell attachment, by comparing the state of tyrosine phosphorylation of the EGF receptor in cell lysates before and after plating on fibronectin. Compared with vector control, the EGF receptor was dephosphorylated by wild type TC45 but protected from dephosphorylation by the TC45-D182A substrate trapping mutant (Fig. 8,  panel A). These results are consistent with the EGF receptor being a direct substrate of TC45, as we have previously demonstrated for EGF-induced activation (12). We next assessed the effect of overexpressing TC45 or TC45-D182A on the extracellular matrix-induced activation of PKB/Akt, by immunoblot analysis of the cell lysates using antibodies specific for PKB/ Akt phosphorylated on Ser-473. Overexpression of either TC45 or TC45-D182A in COS1 cells significantly inhibited the activation of PKB/Akt induced by plating on fibronectin, but TC45 had no significant effect on the induction of ERK2 activity (Fig.  8, panel B). These data illustrate that TC45 can inhibit the integrin-induced, EGF receptor-mediated activation of PKB/ Akt in COS1 cells.

DISCUSSION
Activation of the EGF receptor PTK by ligand results in receptor dimerization and autophosphorylation on tyrosyl residues. The phosphotyrosyl residues that are produced serve as docking sites for SH2 domain-containing signaling molecules, leading to the assembly of multiprotein signaling complexes required for cell growth, proliferation, and survival. Although significant progress has been made in defining the tyrosine phosphorylation-dependent signaling pathways downstream of the EGF receptor, relatively little is known about which members of the PTP family serve to antagonize these signaling events. Through the use of substrate trapping mutants, we have demonstrated previously that the nuclear, 45-kDa form of TCPTP, TC45, can exit the nucleus in response to EGF and FIG. 4. TC45 inhibits the EGF-induced and phosphotyrosineassociated PI 3-kinase activity. COS1 cells were electroporated with either pMT2 vector control or plasmids expressing TC45 or TC45-D182A (TC45D). Electroporated cells were serum-starved and then stimulated with EGF (100 ng/ml) for 15 min. A, Cells were lysed and the phosphotyrosine-containing proteins were immunoprecipitated and assayed for PI 3-kinase activity. Labeled lipids were resolved by thin layer chromatography, and [ 32 P]phosphatidylinositol was quantitated on a PhosphorImager. Negligible [ 32 P]phosphatidylinositol was detected in immunoprecipitates from serum-starved cells not stimulated with EGF. Results from four independent experiments (mean Ϯ standard errors) were expressed as percentage of activity of vector control after EGF stimulation. B, lysates from electroporated cells were resolved by SDS-PAGE and immunoblotted with polyclonal antibodies specific for PKB/ Akt phosphorylated on Ser-473 (phospho-Akt). Phospho-Akt immmunoblots were stripped and reprobed with polyclonal antibodies specific for PKB/Akt. recognize tyrosine-phosphorylated substrates such as the EGF receptor and the 52-kDa isoform of Shc (12). In the case of Shc, TC45 can recognize preferentially Shc phosphorylated on tyrosine 239, compared to tyrosine 317, indicating that TC45 may be capable of regulating selectively Shc-dependent signaling events (12).
In this study we have characterized the ability of TC45 to regulate EGF receptor-induced signaling. TC45 inhibited the EGF receptor-mediated activation of PI 3-kinase and PKB/Akt and, to a lesser extent, JNK, but did not modulate the activation of ERK2. Thus, TC45 can regulate in a selective manner signaling processes which emanate from the EGF receptor. Our data indicate that TC45 may exert its effects on PI 3-kinase and PKB/Akt by inhibiting the recruitment of PI 3-kinase to the EGF receptor. Recruitment of the p85 regulatory subunit of PI 3-kinase to growth factor receptors is necessary for PI 3-kinase activation. The lipid products of PI 3-kinase then engage the pleckstrin homology domain of PKB/Akt and translocate it in the plasma membrane, where it is phosphorylated on Ser/ Thr residues for complete activation (18 -24). EGF receptormediated activation of PI 3-kinase can also contribute to the activation of JNK, although the mechanism involved is not defined, but ERK2 activation is PI 3-kinase-independent.
The p85 regulatory subunit of PI 3-kinase contains two Src homology 2 (SH2) domains that bind tyrosine-phosphorylated amino acids that have a consensus Tyr-X-X-Met motif (33).
Binding of p85 via its SH2 domain to tyrosine-phosphorylated receptors allows for the recruitment and activation of the PI 3-kinase p110 catalytic subunit. The five major autophosphorylation sites on the EGF receptor do not fit the Tyr-X-X-Met motif, but under certain circumstances Src can phosphorylate the EGF receptor on Tyr 920 which has the motif for p85 binding (34). Also, others have reported that the SH2 domains of p85 can interact directly with the tyrosine-phosphorylated EGF receptor (35). Alternatively, the EGF receptor can phosphorylate docking proteins such as p120 cbl and Gab1 (36,37), which in turn bind p85 and therefore recruit PI 3-kinase activity.
Regardless of the precise mechanism by which PI 3-kinase associates with the EGF receptor, tyrosine phosphorylation of the EGF receptor is necessary for this event. By dephosphorylating the EGF receptor, TC45 can inhibit the association of p85 and concomitant activation of PI 3-kinase and PKB/Akt.
The inhibition of p85 recruitment to the EGF receptor that we observed is consistent with the effects of TC45 and TC45-D182A on the activities of PI 3-kinase and PKB/Akt but not entirely consistent with their effects on JNK. Although both TC45 and TC45-D182A could equally inhibit the recruitment of p85 and the activation of PI 3-kinase and PKB/Akt, we observed that only wild type TC45 inhibited JNK. However, it is important to note that, unlike PKB/Akt, whose EGF-induced activation can be completely inhibited by the PI 3-kinase inhibitor wortmannin, JNK activity is only partially inhibited by this PI 3-kinase antagonist (this study and Ref. 16). In addition, dominant negative forms of the p85 regulatory subunit only partially inhibit the EGF-induced activation of JNK (16). Therefore, other signaling events in addition to PI 3-kinase would seem to be necessary for EGF-mediated activation of JNK. Until the exact nature of the PI 3-kinase-mediated activation of JNK has been defined, it will be difficult to speculate as to the mechanism of differential regulation of JNK and PKB/Akt by TCPTP.
FIG. 6. TC45 inhibits the EGF-induced association of the PI 3-kinase p85 regulatory subunit with the EGF receptor. COS1 cells were electroporated with either vector control or plasmids expressing TC45 or TC45-D182A (TC45D). Electroporated cells were serumstarved and then stimulated with EGF (100 ng/ml) for 15 min. A, cells from two 10-cm dishes were lysed in IP lysis buffer as described previously (12) and the phosphotyrosine (pTyr)-containing proteins were immunoprecipitated overnight using 10 l of the phosphotyrosine-specific G104 ascites, resolved by SDS-PAGE, and immunoblotted with antibodies specific for EGF receptor, or the p85 regulatory subunit of PI 3-kinase. EGF receptor (EGFR) blots were stripped and reprobed with the pTyr specific antibody G98. B, cells from two 10-cm dishes were lysed as described previously (12) and the EGF receptor immunoprecipitated with 5 g of anti-EGF receptor Ab1 antibody, resolved by SDS-PAGE, and immunoblotted with antibodies specific for the EGF receptor, or p85.

FIG. 7. The integrin-mediated activation of PKB/Akt and ERK2
in COS1 cells is EGF receptor-mediated. COS1 cells were serumstarved, detached by trypsinization, and kept in suspension (Sus) either in the absence or presence of 100 nM wortmannin (Wort) or 2 M tyrphostin AG1478 for 30 min. Cells were then plated onto plastic dishes coated with 10 g/ml fibronectin (Fib) and allowed to attach for 1 h. Cell lysates were resolved by SDS-PAGE and immunoblotted with polyclonal antibodies specific for (A) pTyr or the EGF receptor, or (B) PKB/Akt phosphorylated on Ser-473 (phospho-Akt) or ERK2. Phospho-Akt immmunoblots were stripped and reprobed with polyclonal antibodies specific for PKB/Akt. Prestained molecular weight standards (Bio-Rad) are indicated in the anti-pTyr blot and the position of the EGF receptor is shown by the arrow on the left. Moro et al. (31) have recently reported that adhesion of human primary skin fibroblasts or ECV304 endothelial cells to fibronectin results in EGF receptor activation in the absence of EGF and that this is necessary for the integrin-mediated activation of the MAPK ERK1. We found that, in COS1 fibroblast cells, the integrin-mediated activation of ERK2 as well as PKB/ Akt are also dependent on EGF receptor activation. Moreover, as in the case of EGF-induced signaling, the activation of PKB/ Akt, but not ERK2, is PI 3-kinase-dependent. In light of these observations, we examined whether TC45 could also regulate EGF receptor signaling processes following transactivation by integrins. Consistent with the effect we observe of TC45 on the EGF-induced activation of PKB/Akt, we have also demonstrated that TC45 can inhibit the activation of PKB/Akt, but not ERK2, following attachment of COS1 cells to fibronectin. Thus, it would be appear that TC45 may regulate EGF receptor signaling irrespective of whether activation of the receptor PTK is initiated by growth factor or integrins.
The EGF receptor is overexpressed or mutated in many human tumors including those derived from brain, lung, breast, and skin (38 -44). The high level of EGF receptormediated MAPK and PI 3-kinase signaling, which occurs as a consequence of this aberrant expression of the PTK, is believed to play an important role in the pathogenesis of these tumors (35,45). Certain PTPs such as SHP-2 and SHP-1 have been reported to exert positive or negative effects on the regulation of EGF receptor signaling to MAPK (46 -51). However, to our knowledge, our data represent the first occasion on which a PTP has been shown to act upstream of PI 3-kinase, most likely on the EGF receptor, to regulate negatively EGF receptormediated and PI 3-kinase-dependent signaling events. Thus, TC45 may serve as an important target for intervention in tumors where excessive EGF receptor-mediated PI 3-kinase signaling contributes to the disease.