The role of protein-tyrosine phosphatase 1B in integrin signaling.

Protein-tyrosine phosphatase 1B (PTP1B) is a key negative regulator of insulin and leptin signaling and a novel therapeutic target for the treatment of type 2 diabetes, obesity, and other associated metabolic syndromes. Because PTP1B regulates multiple signal pathways and it can both enhance and antagonize a cellular event, it is important to establish the physiological relevance of PTP1B in these processes. In this study, we utilize potent and selective PTP1B inhibitors to delineate the role of PTP1B in integrin signaling. We show that down-regulation of PTP1B activity with small molecule inhibitors suppresses cell spreading and migration to fibronectin, increases Tyr(527) phosphorylation in Src, and decreases phosphorylation of FAK, p130(Cas), and ERK1/2. In addition, PTP1B "substrate-trapping" mutants bind Tyr(527)-phosphorylated Src and protect it from dephosphorylation by endogenous PTP1B. These results establish that PTP1B promotes integrin-mediated responses in fibroblasts by dephosphorylating the inhibitory pTyr(527) and thereby activating the Src kinase. We also show that PTP1B forms a complex with Src and p130(Cas), and that the proline-rich motif PPRPPK (residues 309-314) in PTP1B is essential for the complex formation. We suggest that the specificity of PTP1B for Src pTyr(527) is mediated by protein-protein interactions involving the docking protein p130(Cas) with both Src and PTP1B in addition to the interactions between the PTP1B active site and the pTyr(527) motif.

Protein-tyrosine phosphatases (PTPs) 1 constitute a large family of signaling enzymes that together with protein-tyrosine kinases modulate the cellular level of protein-tyrosine phosphorylation (1,2). Genetic and biochemical studies indicate that PTPs are important for many cellular processes and are involved in a number of human diseases (3). For example, PTP1B received much attention because mice lacking PTP1B did not develop diabetes nor did they become obese when placed on a calorie-rich diet (4,5). The phenotype of PTP1B knock-out mice suggests that PTP1B inhibitors may address both obesity and insulin resistance, thereby presenting a unique therapeutic opportunity (6). Besides a role in insulin and leptin signal-ing, PTP1B is also implicated in several other signaling pathways such as growth factor-(7-9) and integrin-(10 -13) mediated processes. Because PTP1B may be a regulator of multiple signal pathways and can both enhance and antagonize a cellular event, the physiological relevance of PTP1B in these processes must be established. This is an important prerequisite for the development of PTP1B-based therapeutics for type 2 diabetes and obesity.
Previous studies suggested conflicting models for the role of PTP1B in integrin signaling. Integrins are transmembrane receptors used by cells to interact with the extracellular matrix such as fibronectin in order to transmit biochemical signals across the plasma membrane. Integrin-dependent adhesion modulates actin cytoskeleton organization, cell motility, cell growth, tumor metastasis, and the ability of cells to escape from apoptosis (14,15). In addition, integrin-mediated signaling can also cooperate with growth factor receptors by enhancing signaling pathways in response to insulin, platelet-derived growth factor, epidermal growth factor, fibroblast growth factor, and vascular endothelial growth factor (16). Upon integrin engagement, the Src family kinases and focal adhesion kinase (FAK) are activated. Together and independently, Src and FAK phosphorylate many other molecules including the adaptor protein p130 Cas (Crk-associated substrate), resulting in the activation of downstream signaling pathways such as the extracellular signal-regulated kinases (ERK) 1 and 2 (17,18).
Although PTP1B has been implicated in integrin signaling, its role in this process is poorly understood. Initial studies with PTP1B overexpression in Rat-1 fibroblasts suggested that PTP1B negatively regulates integrin signaling presumably through association and dephosphorylation of p130 Cas (10,11). In a separate study (12), PTP1B was proposed to positively regulate integrin signaling based on the observation that L fibroblast cells overexpressing a catalytically inactive mutant PTP1B exhibit reduced adhesion and migration on fibronectin, as well as decreased tyrosine phosphorylation in FAK and paxillin. In a more recent study (13), PTP1B was also suggested as a positive regulator of integrin-mediated responses, because loss of PTP1B in SV40 large T antigen-transformed embryonic fibroblasts leads to a dramatic reduction in p130 Cas tyrosine phosphorylation, ERK activation, and cell spreading. However, because little effect was observed on fibronectin signaling in PTP1B-deficient primary fibroblasts, the observed dependence on PTP1B in integrin signaling in SV40 large T antigen immortalized cells may reflect alterations in gene expression caused by transformation (19).
It is unclear what is responsible for the apparent discrepancies from previous studies. It is possible that the overexpression and gene knock-out approaches employed may contribute to some of the differences. Given the potential cross-talk between growth factor and integrin-mediated signaling pathways, and the importance of integrin signaling in various physiological processes including cell growth, invasion, and migration, it is crucial to define the role of PTP1B in integrin signaling. In this study we seek to provide a better understanding of the role of PTP1B in integrin-mediated processes by employing potent and selective small molecule PTP1B inhibitors. The use of small molecule inhibitors to study molecular components of signal transduction pathways offers an invaluable means of analysis complementary to genetic techniques (20,21). The main advantages of the small molecule approach include its simplicity, speed, tunability, and reversibility. In addition, small molecule inhibitors exert their effect on endogenous targets, avoiding the need for overexpression of dominant-negative or constitutively active mutants, which can cause artifacts and lead to erroneous conclusions. Indeed, the availability of agents with specificity for a number of protein kinases has greatly enhanced our ability to identify their substrates and physiological functions (22). Together with PTP "substrate-trapping" approach, our study with potent and selective PTP1B inhibitors generates strong evidence that PTP1B positively regulates fibronectin-mediated integrin signaling via activation of the Src kinase by dephosphorylating the inhibitory pTyr 527 .
Protein Expression and Purification-The catalytic domain of PTP1B (residues 1-321) was used for in vitro study. PTP1B mutants D181A, D181A/Q262A, D181A/P309A/P310A, or D181A/Q262A/P309A/P310A were generated by PCR reactions according to the standard procedure of the QuikChange site-directed mutagenesis kit (Stratagene). The recombinant wild-type and mutant PTP1Bs were expressed in Escherichia coli and purified to homogeneity as described (25). Protein concentration was determined by measuring absorbance at 280 nm using an absorbance coefficient of 1.24 for 1 mg/ml PTP1B. Purified PTP1B was coupled to Affi-Gel 10 agarose beads (Bio-Rad) according to the manufacturer's instructions for the in vitro binding assay.
Cell Spreading Assay-L cells were grown to ϳ80% confluency and serum-starved for 16 h in DMEM. Cells were then detached from the culture dishes by trypsinization, and trypsin was neutralized by addition of DMEM containing 125 g/ml soybean trypsin inhibitor (Sigma). Cells were washed twice with serum-free DMEM and resuspended in this medium at a concentration of 2 ϫ 10 5 cells/ml. 2.5 ml of cells were then applied to 1% bovine serum albumin precoated 6-well plates and incubated for 1 h with or without the presence of different concentrations of PTP1B inhibitors. After incubation, cells were transferred into fibronectin-coated 6-well plates and incubated for 30 min in the cell culture incubator. (The plates were coated with 10 g/ml fibronectin in PBS overnight at 4°C and then rinsed twice with PBS and once with serum-free DMEM before use.) Cells were then photographed under a microscope at ϫ200 magnification. Three random microscopic fields were counted per well. Spreading cells were defined as cells with extended processes, lacking a rounded morphology, and not phase-bright. The non-spreading cells were rounded and phase-bright under microscope.
Cell Migration Assay-L cells were starved in serum-free medium overnight, harvested, and incubated with different concentrations of PTP1B inhibitors as described above. Chambers for haptotaxis assays (Corning Costar, Cambridge, MA) were prepared by precoating the undersurface of the polycarbonate membrane with fibronectin (10 g/ml in PBS) for 2 h at 37°C. The chambers were then washed with PBS, and 1 ϫ 10 5 cells in 0.3 ml of DMEM were added to the upper chamber. The migration was allowed for 3 h at 37°C. Cells on the upper chamber surface were mechanically removed, and the migratory cells on the lower membrane surface were fixed by methanol and stained with Giemsa stain. The migrated cells were then counted under a microscope at ϫ200 magnification. Three random microscopic fields were counted per well.
Immunoprecipitation and Immunoblotting-Cells were grown to ϳ80% confluency, serum-starved, and applied to fibronectin-coated dishes as described above. 2 ml of lysis buffer (50 mM Tris-HCl, pH 7.5, 1% Triton X-100, 5 mM EGTA, 150 mM NaCl, 10 mM sodium phosphate, 10 mM sodium fluoride, 5 mM iodoacetic acid, 1 mM benzamidine, 10 g/ml leupeptin, and 10 g/ml aprotinin) were added to each 15-cm dish. Cells were scraped with a rubber policeman and lysed on ice for 30 min. 10 mM dithiothreitol was added to the lysate at this step. The cell lysate was precleared by centrifugation at 15,000 rpm for 15 min. Protein concentration of the lysate was estimated using the Bio-Rad protein assay reagent. For immunoprecipitation, 1 mg of cell lysate was immunoprecipitated with agarose-conjugated anti-HA antibody. The immunocomplexes were washed four times with the cell lysis buffer, and the immunoprecipitated proteins were eluted with lysis buffer containing 8 M urea. The eluted proteins were separated by SDS-PAGE and transferred electrophoretically to nitrocellulose membrane, which were then immunoblotted by appropriate antibodies followed by incubation with horseradish peroxidase-conjugated secondary antibodies. The blots were developed by the enhanced chemiluminescence technique (ECL kit, Amersham Biosciences) according to the manufacturer's instructions.
In Vitro Binding Assay-100 g of the catalytic domain of the wild type, D181A, D181A/Q262A, D181A/P309A/P310A, or D181A/Q262A/ P309A/P310A PTP1B was coupled to Affi-Gel 10 agarose beads according to the manufacturer's instruction. The protein-linked beads were then mixed with 500 g of L cell lysate, which was prepared as described above, for 3 h at 4°C on a rotating platform. The beads were washed four times with the lysis buffer, and the bound proteins were then subjected to SDS-PAGE and immunoblotting to detect the presence of Src or p130 Cas .

RESULTS AND DISCUSSION
The use of small molecule inhibitors that specifically modulate the activity of individual PTPs represents a powerful approach to determine the function and therapeutic potential of PTPs (27). We have recently identified the most potent and selective PTP1B inhibitor (Fig. 1, compound I) reported to date, which displays a K i value of 2.4 nM for PTP1B and exhibits several orders of magnitude of selectivity in favor of PTP1B against a panel of PTPs (28). The negatively charged difluorophosphonate functional groups participate in key active site and near active site interactions that are responsible for the high potency of compound I and selectivity toward PTP1B (29,30). Not surprisingly, the highly charged inhibitor I is not cell-permeable (23). To enhance its potential for biological studies, we prepared a cell-permeable derivative of compound I by coupling it to the cell-penetrating peptide (D)Arg 8 via a disulfide bridge to generate II ( Fig. 1) (24). Like Drosophila antennapedia homeodomain and HIV-1 TAT, polyarginine peptides can enable rapid transport of small molecule, peptides, small proteins, and oligonucleotides across the plasma membrane into living cells (31,32). Once inside the cell, the cytosolic reducing environment will cleave the disulfide moiety and thereby release I in a freely diffusible form.
We have determined that compound II is indeed cell permeable and enhances insulin-stimulated tyrosine phosphorylation of insulin receptor ␤ (IR␤) and insulin receptor substrate 1 (IRS-1) by nearly 2-fold at 5 nM concentration (twice its K i value) (24). Similar results were obtained with another cellpermeable PTP1B inhibitor, compound III ( Fig. 1), which was derived by attaching a highly lipophilic fatty acid to an analog of compound I (23). Treatment of a number of insulin-sensitive cell lines with III markedly enhanced IR␤ and IRS-1 phosphorylation, Akt and ERK1/2 activation, Glut4 translocation, glucose uptake, Elk1 transcriptional activation, and cell proliferation in insulin signaling (23). Most importantly, the observed cellular effects (e.g. ϳ2-fold increase in IR␤ and IRS-1 phosphorylation) with compounds II and III are similar in magnitude to those observed with antisense-mediated reduction of PTP1B (33) and PTP1B knock-out mice, which are resistant to diabetes and obesity (4). Collectively, the data demonstrate that II and III have the expected biological effects on insulin signaling and serve as both insulin mimetics and sensitizers, validating the notion that small molecule PTP1B inhibitors can be used in antidiabetes therapeutics. Moreover, the data also establish the utility of potent and selective PTP1B inhibitors in further studies of the roles of PTP1B in other signal pathways. In this study, we have employed both compounds II and III to delineate the role of PTP1B in integrin-mediated processes.
Compound II Inhibits L Cell Spreading on Fibronectin-Attachment of integrins to fibronectin induces focal adhesion formation and actin stress fiber assembly, which leads to spreading cell morphology. To assess the effect of PTP1B inhibition on cell spreading, we incubated the suspended cells with or without compound II for 1 h, followed by applying the cells to fibronectin-coated culture dishes. The cells were photographed under the microscope at ϫ200 magnification. As shown in Fig. 2, after 30 min on fibronectin, the number of spreading cells decreases in a dose-dependent manner when the cells were treated with a range of compound II concentrations. Compared with the control, we observed an 11-35% decrease in cell spreading in the presence of 5-50 nM compound II. These results indicate that efficient cell spreading upon integrin engagement requires the catalytic activity of PTP1B.
Compound II Impairs L Cells Migration toward Fibronectin-We next examined the effect of PTP1B inhibition on haptotactic migration of L cells toward fibronectin in a transwell chamber. In this experiment, L cells were starved in serum-free medium overnight, harvested, and incubated with different concentrations (5-50 nM) of compound II for 1 h. The undersurface of the chamber for haptotaxis assays was precoated with fibronectin. Cells were added to the upper chamber, and migration was allowed for 3 h in the CO 2 incubator. After removing the cells on the upper surface, the migratory cells on the lower membrane surface were fixed and enumerated. As shown in Fig. 3, cells treated with compound II showed a dose-dependent decrease in the rate of migration, consistent with a defect in fibronectin-stimulated response when PTP1B is inhibited. The maximal inhibition of cell migration was observed when 20 nM of compound II was applied, and the number of the migratory cells was decreased by 24% compared with the control. It is interesting to note that even at saturating concentrations of compound II, the fibronectin-mediated cell spreading and migration was decreased only by 20 -30%. This is perhaps not surprising because PTP1B may not be the only PTP involved in integrin-mediated processes, and several other PTPs such as SHP2, PTP-PEST, and PTP␣ have previously been implicated as either positively or negatively regulating integrin signaling by dephosphorylating key signaling components such as Src, FAK, and p130 Cas (34 -38).
Effects of Compounds II and III on Integrin Signaling-The effects of PTP1B inhibitor II on cell spreading and migration suggest that PTP1B is required for optimal integrin signaling. It has been shown that when integrin is engaged by the extracellular matrix, Tyr 397 in FAK is autophosphorylated, which creates a binding site for the SH2 domain of Src (39,40). The activated Src then phosphorylates multiple tyrosine residues in the adaptor protein p130 Cas and FAK itself (41)(42)(43), leading to full activation of many of the downstream pathways, most notably the activation of the ERK pathway (44,45). To further define the role of PTP1B in integrin-mediated processes, we examined the effect of PTP1B inhibition with compound II on the phosphorylation status of Src, FAK, p130 Cas , and ERK. L cells were exposed to compound II over a selected range of concentrations for 1 h, followed by adhesion onto fibronectin for 30 min. Cell lysates were either subjected directly to immunoblotting (Src and ERK) or immunoprecipitation (FAK and p130 Cas ). Total cell lysates or immunocomplexes were resolved by SDS-PAGE, electrotransferred to nitrocellulose membranes, and probed with specific antibodies. We first assessed the effect of compound II on Src phosphorylation, as Src activity is regulated by phosphorylation at two distinct tyrosine residues. Autophosphorylation of Tyr 416 in the kinase domain activates Src, whereas phosphorylation of Tyr 527 in the C-terminal tail by C-terminal Src kinase (CSK) inactivates Src. The inhibitory effect arises from the fact that Tyr 527 , when phosphorylated, interacts intramolecularly with the Src SH2 domain, which causes the Src molecule to assume a closed conformation that covers the kinase domain and reduces its potential for sub- The Role of PTP1B in Integrin Signaling strate interaction (46,47). Consequently, activation of Src requires the removal of the C-terminal phosphate by specific PTPs. As shown in Fig. 4, the phosphorylation levels at Tyr 416 and Tyr 527 in Src were determined using phosphospecific antibodies directed toward pSrc 416 and pSrc 527 , respectively. It is clear that treatment of the cells with compound II causes a significant increase in Tyr 527 phosphorylation without any obvious effect on Tyr 416 . When normalized with the Src protein level, treatment of cells with 20 nM compound II leads to ϳ140% increase in Tyr 527 phosphorylation. It is worth pointing out that a significant effect on Tyr 527 phosphorylation was observed at as low as 5 nM compound II concentration, which is only twice its K i value. It is highly unlikely that compound II inhibits any other PTPs at this concentration (28). The results indicate that PTP1B acts as a Src activator by directly and specifically dephosphorylating Tyr 527 . This is consistent with the observations that inhibition of PTP1B with compound II impairs integrin-mediated cell spreading and migration.
We next examined the effect of PTP1B inhibition on the phosphorylation status of FAK and p130 Cas . If FAK and/or p130 Cas serve as direct substrates for PTP1B, then we would observe an increase in tyrosine phosphorylation in the proteins when cells are treated with compound II. If these proteins are Chambers for haptotaxis assays were prepared by precoating the undersurface of the polycarbonate membrane with fibronectin. 1 ϫ 10 5 cells were added to the upper chamber, and migration was allowed for 3 h at 37°C. Migratory cells were counted under a microscope at ϫ200 magnification. Three random microscopic fields were counted per well. Compared with the control, cell migration decreased 13, 19, 24, and 23% when cells were treated with 5, 10, 20, and 50 nM of compound II, respectively.

FIG. 4. Effects of II on the integrin-mediated phosphorylation of Src, p130
Cas , FAK, and ERK. L cells were applied to fibronectincoated culture dishes after being incubated with different concentrations of II. Cell lysates were subjected to immunoblotting or immunoprecipitation. The activation status of Src and ERK1/2 and the total protein-tyrosine phosphorylation level of FAK and p130 Cas were detected using specific antibodies. Lane 1, suspended cells; lanes 2-6, cells were treated with 0, 5, 10, 20, 50 nM inhibitor II, respectively, before applying to fibronectin. not PTP1B substrates, two possible outcomes are expected. In the first, there will be no change in tyrosine phosphorylation associated with these proteins. In the second scenario, there will be a decrease in total tyrosine phosphorylation. This prediction is based on the following. Both FAK and p130 Cas are known to be phosphorylated by Src. Because Src is activated by PTP1B, inhibition of PTP1B by compound II would cause a decrease in Src kinase activity. This would result in a decrease in the phosphorylation of Src substrates. Indeed, treatment of the cells with compound II leads to a dose-dependent decrease in tyrosine phosphorylation of FAK and p130 Cas (Fig. 4). At 20 nM compound II, the phosphorylation levels of FAK and p130 Cas drop off by ϳ160 and 120%, respectively, in comparison with those in the absence of compound II. The results from this experiment demonstrate that neither FAK nor p130 Cas are direct substrates of PTP1B. Finally, we also examined the effect of PTP1B inhibition on the ERK kinase cascade, which is known to act downstream of Src in integrin signaling. If PTP1B activates Src kinase, then a decrease in phospho-ERK would be observed when cells are treated with compound II. This was indeed the case (Fig. 4).
To ensure that the observed results with compound II are caused by inhibition of PTP1B and not nonspecific effects, we also evaluated the effect of a different PTP1B inhibitor compound III on Src and p130 Cas phosphorylation (Fig. 5). Consistent with the observations made with compound II, inhibition of PTP1B with compound III (K i ϭ 26 nM) at equivalent concentrations (50, 100, 250, and 500 nM) also leads to a dose-dependent increase in Tyr 527 phosphorylation, whereas it exerts no obvious effect on Tyr 416 phosphorylation. In addition, decreased tyrosine phosphorylation of p130 Cas was also detected when the cells were treated with III. Collectively, the results with both compounds II and III suggest that the phosphorylated Tyr 527 residue of Src is directly dephosphorylated by PTP1B. Therefore, PTP1B positively regulates integrin signaling by activating Src via removal of the inhibitory phosphate moiety from pTyr 527 .
Further Evidence for Src as a PTP1B Substrate-There has been some confusion in the literature whether Src or p130 Cas serve as a PTP1B substrate in integrin signaling (10 -13, 48). To provide further evidence that Src is a substrate for PTP1B, we performed substrate-trapping experiments using PTP1B mutants that are capable of substrate binding but unable to carry out the dephosphorylation reaction (8,49). PTP1B substrates identified using this approach include insulin receptor, (50), epidermal growth factor receptor (8,49), and JAK2 (51).
We transiently transfected L cells with HA-tagged wild-type or PTP1B substrate-trapping mutants D181A and D181A/ Q262A. The D181A mutant has been used to trap the epidermal growth factor receptor as a PTP1B substrate (8). The D181A/Q262A mutant exhibits an affinity for substrates that is severalfold higher than that of D181A and is thus a more superior PTP1B substrate-trapping mutant (49). After transfection, cells were serum-starved and applied to fibronectincoated dishes. Cells were then lysed, and PTP1B was immunoprecipitated by anti-HA antibody-conjugated agarose beads. The antibody-bound immunocomplexes were denatured and eluted with 8 M urea (52). This is necessary to avoid interference of the otherwise eluted heavy chain of the antibody by SDS sample buffer, which has a similar size to Src and PTP1B. The eluent was then subjected to SDS-PAGE and immunoblotting to detect the presence of PTP1B, Src, Src/pTyr 527 , p130 Cas , and its tyrosine phosphorylation level. As shown in Fig. 6, both Src and p130 Cas are co-immunoprecipitated with the wild-type and PTP1B-trapping mutants. However, it is clear that the trapping mutants, especially D181A/Q262A, bind more Tyr 527phosphorylated Src, indicating that the trapping mutants protect pTyr 527 from dephosphorylation by endogenous PTP1B. This supports the conclusion from studies with PTP1B inhibitors II and III that the Tyr 527 -phosphorylated form of Src is a substrate of PTP1B.
Remarkably, although similar amounts of p130 Cas protein were immunoprecipitated by the wild-type and mutant PTP1Bs, the p130 Cas proteins associated with the trapping mutants are barely phosphorylated, whereas those bound to the wild-type PTP1B are heavily tyrosine-phosphorylated. A likely explanation for this phenomenon is that Src proteins bound to PTP1B-trapping mutants are catalytically inactive because Tyr 527 is phosphorylated, while Src proteins associated with the wild-type PTP1B are activated because of the removal of the inhibitory phosphate. The activated Src can then phosphorylate the associated p130 Cas . The data provide further evidence that p130 Cas is not a PTP1B substrate.
The Proline-rich Region in PTP1B Is Required for Association with Src and p130 Cas -We have shown that PTP1B can associate with Src and dephosphorylate pTyr 527 . Interestingly, although p130 Cas is not a PTP1B substrate, both Src and p130 Cas can be immunoprecipitated by PTP1B. This is perhaps not surprising because p130 Cas is a major Src substrate (53). In addition, PTP1B contains a proline rich sequence PPRPPK (residues 309 -314), which was previously shown to mediate binding to the SH3 domain of p130 Cas , which could be disrupted by replacing proline residues 309 and 310 with alanine (10,11). To determine if this proline-rich motif is required for the association of PTP1B with Src and p130 Cas , we first examined the ability of Affi-Gel 10 bead-conjugated recombinant wild-type and mutant PTP1Bs (residues 1-321) to bind Src and p130 Cas from L cell lysate. As shown in Fig. 7A, both the wild-type PTP1B and the trapping mutants (D181A and D181A/Q262A) are able to pull-down Src and p130 Cas from the cell lysate. When both Pro 309 and Pro 310 were replaced with Ala residues, the association of PTP1B with Src and p130 Cas was almost completely abolished, indicating that the proline-rich region of PTP1B is important for binding Src and p130 Cas .
We next examined whether the proline-rich region is required for association of PTP1B with Src and p130 Cas in vivo. We transiently transfected L cells with HA-tagged wild-type or mutant PTP1B constructs. After incubation on fibronectincoated culture dishes, the cells were lysed. Proteins bound to PTP1B were immunoprecipitated with agarose-conjugated anti-HA antibodies. The immunocomplexes were separated by SDS-PAGE, and the presence of Src and p130 Cas were detected with specific antibodies (Fig. 7B). Similar to the in vitro binding experiment described above, wild-type PTP1B and the D181Aand D181A/Q262A-trapping mutants were all capable of binding the Src and p130 Cas proteins, whereas little Src and p130 Cas were immunoprecipitated when Pro 309 and Pro 310 were replaced with an Ala. The data indicate again that the interaction between PTP1B and Src or p130 Cas is mediated through the proline-rich region.
A Model for Src Recognition by PTP1B-We have shown that PTP1B can associate with Src and p130 Cas both in vitro and in vivo, and that the proline-rich sequence in PTP1B is essential for the association. We have also presented evi-dence that although PTP1B does not dephosphorylate p130 Cas , it specifically activates Src by dephosphorylating the inhibitory site pTyr 527 and therefore plays a positive role in integrin signaling. What is the mechanism of Src recognition by PTP1B? Previous studies have established that p130 Cas binds both the SH2 and SH3 domains of Src via its C-terminal region including Tyr 762 and a proline-rich motif RPLPSPP (42,43,54), which matches the consensus class I SH3 domain binding motif RXLPXXP (55). Interestingly, the proline-rich sequence PPRPPK (residues 309 -314) in PTP1B conforms to the canonical class II SH3 domain binding motif PXXPX(R/K) (55) and is required for physical association with the SH3 domain of p130 Cas (10,11). Based on these findings and the results obtained from this study, we propose a model for Src recognition by PTP1B (Fig. 8). In this model, only the p130 Cas -associated Src can serve as a substrate for PTP1B, and p130 Cas serves as an intermediary protein linking PTP1B together with Src. The direct binding of p130 Cas to the SH2 and SH3 domains of Src displaces the intramolecular interactions that maintain the closed Src conformation resulting in an open Src conformation. This renders pTyr 527 accessible to the active site of PTP1B. The PTP1B-catalyzed pTyr 527 dephosphorylation is further enhanced by the interaction between the proline-rich region in PTP1B and the SH3 domain of p130 Cas , which serves to bring Src in proximity of PTP1B and to increase the local Src substrate concentration.
In summary, we have shown that inhibition of PTP1B with small molecule inhibitors results in 20 -30% suppression of cell spreading and migration to fibronectin. Inhibition of PTP1B with small molecule inhibitors also leads to an increase in Src Tyr 527 phosphorylation and a decrease in FAK, p130 Cas , and ERK phosphorylation. In addition, PTP1B substrate-trapping mutants protect pTyr 527 from dephosphorylation by endogenous PTP1B. Collectively, the results provide strong evidence that PTP1B plays a positive role in integrin signaling by spe- FIG. 7. The proline-rich region in PTP1B is required for association with Src and p130 Cas . For in vitro experiment (A), wild-type and mutant recombinant PTP1B proteins were conjugated to Affi-Gel 10 beads. L cell lysate was mixed with the PTP1B-conjugated beads, and the binding proteins were subjected to immunoblotting to detect the presence of Src and p130 Cas . For in vivo experiments (B), L cells were transiently transfected with wild-type or mutant PTP1B constructs. Cells were lysed after being applied to fibronectin-coated dishes. The HA-tagged PTP1B was immunoprecipitated, and the bound proteins were separated and immunoblotted to detect the presence of Src and p130 Cas .

FIG. 8. A model for Src recognition by PTP1B
. We propose a model in which only the p130 Cas -associated Src can serve as a substrate for PTP1B. The interaction between the proline-rich region in PTP1B and the SH3 domain of p130 Cas promotes PTP1B-catalyzed Src dephosphorylation by bringing Src in close proximity to PTP1B thus increasing the local substrate concentration, whereas the interaction between the C-terminal region of p130 Cas with both the SH2 and SH3 domains of Src induces Src to assume a more open conformation, thus making pTyr 527 accessible to the PTP1B active site. cifically dephosphorylating the inhibitory pTyr 527 and activating the Src kinase. We also show that PTP1B forms a complex with Src and p130 Cas and that the proline-rich sequence at the C terminus of PTP1B is critical for the association. These observations, together with results from previous studies, suggest that the specificity of PTP1B for Src pTyr 527 is controlled by not only classical interactions between PTP1B active site and the pTyr 527 motif, but also by protein-protein interactions involving p130 Cas with both Src and PTP1B. This provides the first example by which PTP substrate specificity is regulated by an intermediary protein interacting with both the substrate and the PTP.