Src Kinase Activity Is Regulated by the SHP-1 Protein-tyrosine Phosphatase*

Activation of the cellular Src tyrosine kinase depends upon dephosphorylation of the carboxyl-terminal inhibitory tyrosine phosphorylation site. Herein we show that Src isolated from human platelets and Jurkat T cells is preferentially dephosphorylated at its inhibitory phosphotyrosine site by the SHP-1 tyrosine phosphatase. The data also revealed association of Src with SHP-1 in both platelets and lymphocytes and the capacity of Src to phosphorylate SHP-1 and interact with the SHP-1 NH2-terminal SH2 domain in vitro. Analysis of Src activity in thymocytes from SHP-1-deficient motheaten and viable motheaten mice revealed this kinase activity to be substantially lower than that detected in wild-type thymocytes, but to be enhanced by in vitroexposure to SHP-1. Similarly, immunoblotting analysis of thymocyte Src expression before and after selective depletion of active Src protein indicated that the proportion of active relative to inactive Src protein is markedly reduced in motheaten compared with wild-type cells. These observations, together with the finding of reduced Src activity in HEY cells expressing a dominant negative form of SHP-1, provide compelling evidence that SHP-1 functions include the positive regulation of Src activation.

Activation of the cellular Src tyrosine kinase depends upon dephosphorylation of the carboxyl-terminal inhibitory tyrosine phosphorylation site. Herein we show that Src isolated from human platelets and Jurkat T cells is preferentially dephosphorylated at its inhibitory phosphotyrosine site by the SHP-1 tyrosine phosphatase. The data also revealed association of Src with SHP-1 in both platelets and lymphocytes and the capacity of Src to phosphorylate SHP-1 and interact with the SHP-1 NH 2 -terminal SH2 domain in vitro. Analysis of Src activity in thymocytes from SHP-1-deficient motheaten and viable motheaten mice revealed this kinase activity to be substantially lower than that detected in wild-type thymocytes, but to be enhanced by in vitro exposure to SHP-1. Similarly, immunoblotting analysis of thymocyte Src expression before and after selective depletion of active Src protein indicated that the proportion of active relative to inactive Src protein is markedly reduced in motheaten compared with wild-type cells. These observations, together with the finding of reduced Src activity in HEY cells expressing a dominant negative form of SHP-1, provide compelling evidence that SHP-1 functions include the positive regulation of Src activation.
Both the kinase and transforming activities of the pp60 c-src (Src) protein-tyrosine kinase (PTK) 1 are repressed by phosphorylation of a tyrosine residue within the Src carboxyl terminus (1)(2)(3)(4)(5). The negative regulatory effect of this phosphotyrosine (Tyr-530 and Tyr-529 in human and murine Src, respectively) has been ascribed to its association with the Src SH2 domain and consequent SH2, as well as SH3 domain-mediated intramolecular interactions that repress activity of the kinase domain (6 -8). The catalytic activities of the other Src family members are similarly inhibited by phosphorylation of this conserved C-tail tyrosine (9), and activation of each of these PTKs thus depends on dephosphorylation at this site. However, while several lines of evidence implicate the Csk tyrosine kinase (10 -12) as well as Src-mediated autophosphorylation (13) in the phosphorylation of the COOH-terminal regulatory tyrosine, relatively little is known about the mechanisms whereby this inhibitory phosphotyrosine residue is dephosphorylated and Src activation achieved. In contrast to the Src-related Lck and possibly Fyn PTKs, the Src COOH-terminal phosphotyrosine does not appear subject to dephosphorylation by the transmembrane protein-tyrosine phosphatase (PTP), CD45 (14 -16). Moreover, while Src activity has been found to be increased in rodent cell lines overexpressing the receptor-like protein-tyrosine phosphatase, RPTP␣ (17,18), a direct role for this PTP in dephosphorylating Tyr-529 in vivo has not been established. However, recent data from studies of the SHP-1 tyrosine phosphatase, a cytosolic, dual SH2 domain-containing protein expressed primarily in hemopoietic and epithelial cells (19 -22), have raised the possibility that this PTP plays a role in the dephosphorylation and activation of Src. Thus, for example, thrombin stimulation of platelets, which is associated with an enhancement of Src activity as well as its rapid translocation to the cytoskeletal fraction (23), has recently been shown to induce increases in SHP-1 tyrosine phosphorylation, association of SHP-1 with Src and redistribution of this PTP to the cytoskeleton with kinetics similar to those observed for Src (24,25). These observations, together with data implicating the viral transforming v-Src protein in the phosphorylation of SHP-1 (26), suggest that, in at least some cell lineages, SHP-1 interacts with Src and may serve to activate Src by dephosphorylating the COOH-terminal regulatory tyrosine.

MATERIALS AND METHODS
Antibodies and Proteins-The monoclonal anti-Src 327 (27), anti-Src GD11 (28), and anti-SHP-1 (recognizing the COOH-terminal portion of SHP-1) antibodies were obtained from Dr. J. Brugge (ARIAD Pharmaceuticals, Cambridge, MA), Upstate Biotechnology Inc. (UBI, Lake Placid, NY), and Transduction Laboratories (Lexington, KY), respectively. The monoclonal anti-Src antibody, clone 28, which selectively recognizes the active, carboxyl-terminal-dephosphorylated form of Src (29), was a gift from Drs. J. Yano and K. Owada (Kyota Pharmaceutical University, Kyoto, Japan). The SRC2 rabbit polyclonal anti-Src antibody, which recognizes a similar epitope as clone 28 and was demonstrated in comparative studies to also selectively immunoprecipitate the active form of Src (data not shown), was obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Rabbit polyclonal anti-SHP-1 antibody recognizing the tandem SH2 domains of SHP-1 was generated in our laboratory as described previously (30). Rabbit anti-p56 lck antibody was produced in our laboratory by immunizing rabbits with a bacterial TrpE fusion protein containing amino acids 7-144 of Lck (31). Rabbit muscle enolase (Boehringer Mannheim, Lavel, Canada) was acid-treated before use as described previously (32). Fyn and Lck enzyme purified from bovine thymus were obtained from UBI, while purified Src was pur-chased from UBI as a recombinant protein expressed in SF9 insect cells by baculovirus containing the human c-src gene.
Cells and Cell Lines-The human Jurkat leukemia and the MOLT 4 T cell lines were obtained from the American Type Culture Collection (ATCC, Rockville, MD). Cells were cultured in RPMI 1640 (Life Technologies, Inc.) supplemented with 10% (v/v) fetal bovine serum (Sigma) and 10 M gentamycin (Life Technologies, Inc.). Cells were grown at 37°C in a fully humidified incubator containing 5% CO 2 . Fresh human platelets and resting T cells were obtained from normal healthy peripheral blood buffy coats (Toronto Canadian Red Cross). Platelets were isolated from plasma after low speed centrifugation (PBL). T cells were isolated as peripheral blood lymphocytes (PBL) after purification of buffy coat blood on density gradients as described previously (33). Proliferating T cell lymphoblasts were prepared from PBL by stimulation with PHA (10 g/ml; Difco) for 2 days followed with 100 units/ml recombinant interleukin-2 (Cetus) for an additional 2 days. The HEY ovarian adenocarcinoma cell line was obtained from ATCC and maintained in complete medium.
Immunoprecipitation and Western Blotting-Cells were lysed in 1% Nonidet P-40 lysis buffer (1% Nonidet P-40, 50 mM HEPES, pH 7.23, 150 mM NaCl, 50 M NaF, 50 M O-phosphate, 50 M ZnCl 2 , 2 mM EDTA, 2 mM Na 3 VO 4 , 2 mM phenylmethylsulfonyl fluoride). Cell lysates were centrifuged at 14,000 ϫ g for 10 min at 4°C, and protein concentrations then determined by means of the bicinchoninic acid (BCA) assay (Pierce). Immunoprecipitation and Western immunoblots were performed as described previously (33). For protein detection, blots were blocked overnight in 5% nonfat milk for detection of Src, washed, incubated with the appropriate secondary antibody, and subjected to enhanced chemiluminescence (ECL) or incubation with 125 I-Protein A. Where indicated, the immunoblots were stripped and reprobed with anti-Src or anti-SHP-1 antibody.
In Vitro Kinase Assay-Tyrosine kinase activity of Src was assayed by immunoprecipitation of lysates as described above. The immunoprecipitates were then washed in kinase wash buffer (50 mM HEPES, pH 7.23, 150 mM NaCl, 1 mM MgCl 2 , 1 mM MnCl 2 , 0.5% Nonidet P-40). The precipitates were then incubated in 20 l of kinase buffer containing 5 Ci of [␥-32 P]ATP (ICN) with or without 2 g of acid-treated rabbit muscle enolase. The mixture was incubated for 10 min at 37°C, reduced SDS-gel sample buffer added, and the samples boiled for 10 min. The samples were centrifuged at 14,000 ϫ g for 10 min prior to loading supernatants onto SDS-PAGE gels containing 10 -12% polyacrylamide. The 32 P-labeled proteins were electrophoretically transferred to Immobilon-P membranes (Millipore Corp., Bedford, MA) and the labeled proteins visualized by autoradiography. Src quantitation was also performed by anti-Src immunoblotting of the membranes using ECL. For assays of Src kinase activity toward SHP-1, 10 ng of recombinant Src protein was incubated at 37°C for 10 min in 20 l of kinase buffer containing 5 Ci of [␥-32 P]ATP in the presence of acid-denatured enolase (2 g), GST (10 g), or GST-SHP-1 (2 g) fusion protein. Samples were then boiled for 10 min in reduced SDS-gel sample buffer, centrifuged at 14,000 ϫ g for 10 min, and the supernatant proteins fractionated by SDS-PAGE. After electrotransfer to Immobilon-P, Src was visualized by autoradiography.
Cyanogen Bromide (CNBr) Cleavage and Phosphoamino Acid Analysis-In vitro [␥-32 P]ATP-labeled or in vivo 32 PO 4 -labeled Src was immunoprecipitated with anti-pp60 c-src 327 and isolated by SDS-PAGE. After electrotransfer to nitrocellulose membranes, the 60-kDa Srccontaining band was excised from the membranes and subjected to CNBr cleavage as described previously (34). The excised Src protein was incubated with 60 mg/ml CNBr in 70% formic acid for at least 2 h at room temperature. Samples were then washed and dried and the CNBr-generated peptide fragments resuspended in Tricine SDS sample buffer, resolved by separation on 10 -20% gradient Tricine SDS-PAGE (Novex, San Diego, CA), transferred to Immobilon-P membranes, and visualized by autoradiography. For phosphoamino acid analysis, CNBrgenerated protein fragments were excised from Immobilon-P membranes, hydrolyzed in 6 M HCl at 110°C for 1 h and separated by one-dimensional phosphoamino acid analysis on a cellulose plate as described previously (34 -36). Labeled phosphoamino acids were detected by autoradiography and identified by comparison to ninhydrinstained standards.
In Vivo Metabolic Labeling-In vivo labeled Src for CNBr mapping studies was derived by incubating 4 ϫ 10 8 human platelets or 5 ϫ 10 7 Jurkat or MOLT 4 T cells for 2 h at 37°C with 32 PO 4 (1 mCi/ml) in phosphate-free RPMI as described previously (23,35,36) followed by lysis in Nonidet P-40 lysing buffer and subsequent isolation of the labeled Src proteins as described above.
GST Fusion Proteins-The glutathione S-transferase (GST) fusion proteins used in this study were derived by subcloning the following cDNA-or PCR-amplified fragments into pGEX2T: the full-length murine SHP-1 cDNA (GST-SHP-1), a full-length murine SHP-1 cDNA containing a Cys-453 3 Ser mutation (GST-SHP-1 (C453S)), the SHP-1 NH 2 -terminal SH2 domain (amino acids 1-95), the SHP-1 COOHterminal SH2 domain (amino acids 110 -205), and the SHP-1 NH 2 -and COOH-terminal SH2 domains (amino acids 1-221). These GST expression plasmids were transfected into Escherichia coli JM101, and the fusion proteins purified from isopropyl-1-thio-␤-D-galactopyranosideinduced bacteria with glutathione-conjugated Sepharose beads (Pharmacia Biotech Inc.). GST-SHP-1 fusion protein was also obtained from UBI. For binding studies, 1 g of each GST fusion protein and GST beads was incubated with 7 ng of in vitro 32 P-labeled recombinant Src protein at 4°C for 2 h. The beads were then washed and resuspended in SDS-sample buffer, and the proteins subjected to fractionation over 10% SDS-PAGE, transferred to nitrocellulose, and the 32 P-labeled Src protein visualized by autoradiography.
Phosphatase Reactions-For phosphatase reactions using GST fusion proteins, proteins were eluted from equal amounts of GST-SHP-1 and GST-SHP-1 (C453S) bead complexes into 300 l of elution buffer (10 mM reduced glutathione, 50 mM Tris-HCl, pH 8.0). Equal amounts of these eluted proteins were then incubated with immunoprecipitated 32 P-labeled Src (in vitro or in vivo labeled) at 37°C in 200 l of phosphatase buffer (10 mM Tris-HCl, 1.0 mM EDTA, 1 mg/ml bovine serum albumin, 0.1% 2-mercaptoethanol, 0.01% NaN 3 , pH 7.34). To confirm the activity of the eluted proteins, activities of the PTPs were assayed in parallel using 10 mM p-nitrophenyl phosphate as substrate (37).
Transfection Studies-For stable transfections of HEY cells, the plasmids pCMV4Neo, pCMV4Neo-SHP-1, and pCMV4Neo-SHP-1(C453S) were introduced by lipofection into HEY cells and the cells selected for Geneticin resistance. The resultant lines were assayed for expression of SHP-1 by immunoblotting analysis with rabbit anti-SHP-1 antibody and found to overexpress at similar levels either wild-type SHP-1 or phosphatase-inactive SHP-1 (C453S).

RESULTS AND DISCUSSION
To determine whether the COOH-terminal or other phosphotyrosine residues within Src are subject to dephosphorylation by SHP-1, the effects of this phosphatase on Src tyrosine phosphorylation were initially examined using CNBr cleavage analysis. As illustrated in Fig. 1A, CNBr treatment of 32 P-labeled human Src has been shown previously to yield phosphorylated cleavage fragments of about 31, 9.7, and 4.7 kDa, which, respectively, contain the Src NH 2 -terminal region encompassing the major sites for serine phosphorylation on Src, Ser-12 and Ser-17 (31-kDa fragment), the inhibitory tyrosine phosphorylation site, Tyr-530 (4.7-kDa fragment), and a key site for autophosphorylation on activated Src, Tyr-419 (9.7-kDa fragment; Refs. 14 -16, 23, and 38 -41). As is consistent with these data, Tricine SDS-PAGE and subsequent phosphoamino acid analysis of the cleavage fragments derived by CNBr hydrolysis of Src immunoprecipitates from 32 P-labeled human T cells (Molt 4 and Jurkat) and freshly-isolated platelets also revealed a predominantly phosphoserine-containing 31-kDa and predominantly phosphotyrosine-containing 9.7-and 4.7-kDa fragments (Fig. 1, B and C). However, in addition to these fragments, the CNBr cleavage products generated in this analysis also included a phosphorylated fragment of about 3.3 kDa (Fig.  1B), which was shown by phosphoamino acid analysis to contain primarily phosphotyrosine (Fig. 1C). Based on the apparent molecular weight of this fragment and the locations of methionine residues within Src (Fig. 1A), the 3.3-kDa cleavage fragment most likely corresponds to amino acids 315-341 within the Src catalytic domain, a possibility that suggests Tyr-338, a residue recently shown to become phosphorylated in the context of in vitro phosphate labeling of Src (42), represents another site of Src tyrosine phosphorylation in vivo and as such may also be of relevance to the expression of Src kinase activity. By contrast, although one other site of tyrosine phosphorylation (corresponding to Tyr-216 in human Src) has recently been identified within the 31-kDa CNBr fragment (43), this phosphotyrosine was not detected under the conditions used in this current analysis.
As CNBr cleavage analysis under the conditions used in this study distinguished the Src region containing Tyr-530 as well as the other major known sites for tyrosine phosphorylation on Src, this approach was next used to determine whether any of these sites represent targets for SHP-1-mediated dephosphorylation in vitro. To this end, 32 P-labeled recombinant Src was incubated with GST fusion proteins carrying SHP-1 or, as a control, a mutated (C453S), catalytically inert form of SHP-1, and the effects of the latter proteins on the profile of phosphorylated CNBr Src cleavage fragments then examined. As shown in Fig. 2A, the phosphorylated 9.7-, 4.7-, and 3.3-kDa CNBr fragments known to represent phosphotyrosine-containing regions of Src were again detected following CNBr hydrolysis of Src pretreated with inactive SHP-1 (C453S). By contrast, intensities of the bands representing each of these 32 Plabeled species, but most notably the 4.7-and 3.3-kDa phosphorylated fragments, were markedly reduced in the context of Src pretreatment with SHP-1, a result which suggests that the phosphotyrosine residues located in these fragments can be dephosphorylated in vitro by SHP-1. Along similar lines, SHP-1 treatment of in vitro 32 P-labeled Src immunoprecipitates from human platelets was also associated with dramatic decreases in the signal intensities and, by extension, phosphorylation states of the three Tyr-containing CNBr-generated cleavage fragments (Fig. 2B). As shown in Fig. 2B, phosphorylation of the Src regions represented by the 4.7-and 3.3-kDa cleavage fragments again appeared to be more sensitive to SHP-1 pretreatment than was phosphorylation of the 9.7-kDa fragment. A further assessment of these changes by densitometric analysis revealed the reduction in phosphorylation levels following a 30-min or 60-min incubation of Src with SHP-1 to be about 8 -10-fold and 2-fold, respectively, greater for the 4.7-and 3.3-than for the 9.7-kDa fragment (Fig. 2C). Thus, while these data suggest the capacity of SHP-1 to dephospho-rylate in vitro Tyr-419, Tyr-338, and Tyr-530, the major sites for tyrosine phosphorylation on the 9.7-, 3.3-, and 4.7-kDa cleavage fragments, respectively, these findings also indicate that SHP-1 exerts a more pronounced effect on the phosphorylation of Tyr-338 and Tyr-530 than on Tyr-419.
To determine whether SHP-1 also dephosphorylates Src phosphotyrosine residues labeled in vivo, additional CNBr cleavage studies were performed on SHP-1-treated Src immunoprecipitates derived from 32 P-biosynthetically labeled human platelets and Jurkat T cells. As shown in Fig. 2D, pretreatment of the in vivo labeled Src with catalytically inert SHP-1 (C453S) was again associated with the generation of the three phosphorylated CNBr cleavage fragments known to contain key residues for Src tyrosine phosphorylation. However, in contrast to the data obtained by analysis of in vitro labeled Src protein, SHP-1 treatment of the in vivo labeled Src immunoprecipitates was associated with a reduction in the phosphorylation of the 4.7-kDa CNBr-derived cleavage fragment, but had little effect on phosphorylation status of the 9.7-and 3.3-kDa fragments. As indicated by the phosphoamino acid analysis of these respective segments shown in Fig. 1C, this difference in the susceptibility of in vivo versus in vitro 32 P-labeled Src to SHP-1 catalytic activity cannot be ascribed to differences in the phospholabeling of serine/threonine residues and may therefore reflect in vivo labeling of additional tyrosine residues not efficiently phosphorylated under in vitro conditions. These residues might include, for example, tyrosines 329 and 343, which are also contained within the 3.3-kDa fragment and which are not susceptible to the action of SHP-1. While this latter issue requires further investigation, these data are consistent with the hypothesis that Src represents a substrate for SHP-1 and raises the possibility that SHP-1 mediates dephosphorylation of the COOH-terminal inhibitory phosphotyrosine site.
In view of the data suggesting SHP-1 involvement in the dephosphorylation and potentially regulation of Src, the possibility that these two enzymes physically associate with one another in activated lymphocytes was next investigated. As shown in Fig. 3A (upper panel), anti-Src immunoblotting analysis of SHP-1 immunoprecipitates prepared from interleukin- and Jurkat T cells were metabolically labeled with 32 PO 4 and Src then isolated by immunoprecipitation. Following purification on SDS-PAGE, the labeled Src was subjected to CNBr hydrolysis and the cleavage fragments analyzed by electrophoresis over 10 -20% gradient Tricine SDSpolyacrylamide, followed by transfer to Immobilon-P and autoradiography. C, phosphoamino acid analysis of the phosphorylated CNBr Src fragments indicating whether phosphorylation is on serine (S), threonine (T), or tyrosine (Y).
2-stimulated peripheral blood lymphocytes, revealed an association of SHP-1 with Src in these cells, although comparison of the amount of Src present in total cell lysates versus the amount coprecipitated with SHP-1 suggests that the fraction of total cellular Src associated with SHP-1 is very low. SHP-1-Src association was also observed in reciprocal experiments involving anti-SHP-1 immunoblotting analysis of anti-Src immunoprecipitates (Fig. 3A, lower panel) and in similar studies of Jurkat T lymphocytes and human platelets (data not shown), but was consistently not detected when cell lysates for these studies were prepared in buffer lacking orthovanadate, a potent tyrosine phosphatase inhibitor (Fig. 3A). Together these observations link the association of SHP-1 with Src to a phosphotyrosine-dependent interaction and, as is consistent with previous data showing that thrombin stimulation induces both association and increased tyrosine phosphorylation of platelet SHP-1 and Src (24), suggest that the interaction of these enzymes may be mediated through the binding of one or both of the SHP-1 SH2 domains with Src phosphotyrosine residues. To address this issue, GST fusion proteins carrying full-length SHP-1 and SHP-1 (C453S) or the tandem or single SHP-1 SH2 domains were evaluated for their capacities to interact with phosphorylated recombinant Src. The results of this in vitro analysis revealed specific binding of phosphorylated Src (Fig. 3B), but not phosphorylated Fyn or Lck to the SHP-1 (C453S) fusion protein. Furthermore, as shown in Fig.  3C, Src was also precipitated by the fusion proteins containing both SH2 or the NH 2 -terminal SHP-1 SH2 domain, but not by the fusion protein containing only the COOH-terminal SHP-1 SH2 domain. Thus, as has also been described in relation to SHP-1 interactions with other putative substrates, such as the erythropoietin and c-Kit receptors (44,45), these observations suggest that a single (in this instance, the NH 2 -terminal) SH2 domain mediates the binding of SHP-1 to phosphotyrosine(s) on Src. Conversely, GST fusion proteins containing the Src SH2 domain have been shown to precipitate phosphorylated SHP-1 from thrombin-activated platelets (24), and it is therefore possible that SHP-1 association with Src involves multiple sites of interaction on these respective proteins. However, in view of our results indicating that GST-SHP-1 (C453S) fusion proteins also precipitate Src from peripheral blood lymphoblast lysates containing orthovanadate, but not from lysates lacking orthovanadate (data not shown) as well as data showing SHP-1 activity to be substantially increased by the binding of its NH 2 -terminal, but not COOH-terminal SH2 domain to tyrosine-phosphorylated ligands (46), the interaction of the SHP-1 NH 2 -terminal SH2 domain with Src appears to represent a physiologically relevant molecular association, which may facilitate SHP-1 recruitment to and ultimately dephosphorylation of Src.
While the analysis of SHP-1 effects on Src tyrosine phosphorylation identify Src as a potential substrate for this PTP, previous data showing that SHP-1 can be tyrosine-phosphorylated by v-Src (26), the product of the transforming gene of Rous sarcoma virus, suggest that SHP-1 may also represent a substrate of the cellular Src homologue. To address this possibility, the capacity of recombinant Src to phosphorylate GST-SHP-1 fusion proteins was investigated using an in vitro kinase assay. As shown in Fig. 3D, the results of this analysis revealed the induction of SHP-1 tyrosine phosphorylation by Src, thereby identifying SHP-1 as a potential substrate for this kinase. SHP-1 has also been shown to be a substrate for Lck (47) and ZAP-70 (48), and the respective relevance of these PTKs to SHP-1 phosphorylation in vivo thus requires further analysis. However, in view of data showing SHP-1 activity to be enhanced by tyrosine phosphorylation (49), the current data raise the possibility that a reciprocal functional relationship exists between SHP-1 and Src, the physical association of these  Fig. 1. B, CNBr cleavage of SHP-1-treated in vitro labeled platelet Src. Src immunoprecipitates prepared from 4 ϫ 10 8 human platelets were 32 P-labeled in vitro, incubated for 30 or 60 min at 37°C with 30 ng of GST-SHP-1 fusion protein, and then purified and subjected to CNBr hydrolysis as above. C, phosphorylation of each phosphotyrosine-containing CNBr fragment shown in B was quantitated using a Molecular Dynamics Computing Densitometer and tyrosine phosphorylation (pY) of the 9.7-, 4.7-, and 3.3-kDa cleavage fragments derived from SHP-1-treated Src was compared with that of the respective fragments derived from SHP-1 (C453S)-treated Src and the difference expressed as the percentage decrease in tyrosine phosphorylation. D, CNBr cleavage of SHP-1-treated in vivo labeled Src. Human platelets and Jurkat T cells were metabolically labeled with 32 PO 4 , and Src was then isolated by immunoprecipitation. Following 30 min of treatment with 30 ng of GST-SHP-1 (C453S) or SHP-1 fusion proteins, the labeled Src was purified and subjected to CNBr cleavage analysis as above. The data shown in each panel are representative results of at least two independent experiments. proteins allowing each enzyme to activate the other in either a coordinate or sequential fashion. Along similar lines, binding of SHP-1 to another putative PTK substrate (ZAP-70) has been shown to induce changes in the catalytic activities of both enzymes, although in contrast to Src, ZAP-70 kinase activity appears to be diminished in conjunction with increases in SHP-1 phosphatase activity (48).
In addition to SHP-1, a number of other PTPs, including most recently the structurally similar SHP-2 protein, have been shown to dephosphorylate Src in vitro at the COOHterminal regulatory tyrosine (17,18,50,51), but the extent to which these respective enzymes contribute to Src activation in vivo remains unclear. To specifically ascertain the biologic relevance of SHP-1 to the regulation of Src, activity of this PTK was next investigated in unstimulated thymocytes from motheaten (me) and viable motheaten (me v ) mice, animals that express negligible SHP-1 catalytic activity consequent to lossof-function mutations in the SHP-1 gene (30). As is consistent with previous data from our group (33), the results of anti-Src immunoblotting analysis confirmed the presence of Src in thymocytes from both mutant and normal, congenic mice. However, while levels of Src protein appeared equivalent in the SHP-1-deficient and normal thymocytes, the in vitro kinase activity of Src as evaluated by autophosphorylation following immunoprecipitation was markedly lower in the me and me v cells compared with normal thymocytes (Fig. 4A). By contrast, levels of Lck autophosphorylation and protein after immuno-precipitation were no different in the me and wild-type cells (Fig. 4B). These findings are consistent with the in vitro data demonstrating the capacity of SHP-1 to dephosphorylate the Src COOH-terminal inhibitory tyrosine and, although the biologic functions of Src in thymocytes are currently unclear, these data strongly suggest a role for SHP-1 in the activation of Src in vivo. This contention is also supported by the finding that SHP-1 incubation with Src immunoprecipitates from me thymocytes substantially increased the extent of Src autophosphorylation in vitro (Fig. 4C), a result that again links SHP-1 to not only the dephosphorylation, but also the activation of Src.
To confirm the association of the me mutation with reduced Src activity, Src protein was immunoprecipitated from me and wild-type thymocytes using antibodies (clone 28 and SRC 2) raised against the tyrosine dephosphorylated Src COOH terminus and shown previously to selectively recognize the active form of the enzyme (29). As illustrated in Fig. 4D (panel II), the results of an in vitro kinase assay using these respective immunoprecipitates revealed Src activity to be detectable in the wild-type, but not in the me-derived immunoprecipitates. As immunoblotting analysis of the me cell lysates with the GD11 (Fig. 4D, panel I), or 327 (data not shown) anti-Src antibodies, which react with both the inactive and active forms of Src, demonstrated comparable levels of total Src in me and wildtype cells, these data indicate that me thymocyte Src protein is not immunoprecipitable by the clone 28 and SRC 2 antibodies and therefore imply that the Src present in me thymocytes is largely inactive. As is consistent with this contention, re-immunoprecipitation of Src from supernatants of clone 28 or SRC2 immunoprecipitates using the 327 anti-Src antibody revealed the me, but not the wild-type immunoprecipitates, to contain a substantial level of Src, presumably representing the inactive protein (Fig. 4D, panel III). These findings, together with the demonstration that the me mutation does not alter expression of other prominent thymocyte PTPs, such as CD45 and SHP-2 (Fig. 4E), provide strong evidence that SHP-1 deficiency impairs Src activation in vivo.
As SHP-1 is expressed in epithelial as well as hemopoietic lineages, the role for SHP-1 in Src activation was also investigated by evaluating Src activity in HEY ovarian cancer cells transfected with expression constructs containing cDNAs for either wild-type SHP-1 or SHP-1 (C453S), the latter of which has been shown to function in a dominant negative fashion in some cell types (48,52). Compared with HEY cells transfected with vector alone, levels of SHP-1 protein in the cells stably transfected with SHP-1 and SHP-1 (C453S) were found to be increased by about 2-fold (data not shown). As shown in Fig.  4F, analysis of the in vitro kinase activity of Src protein immunoprecipitated from these transfectants revealed the phosphorylation of both Src as well as exogenous enolase substrate to be relatively reduced in the SHP-1 (C453S) overexpressing cells compared with cells overexpressing wild-type SHP-1 or transfected with vector alone. Following epidermal growth factor stimulation of these cells, Src activity again appeared lower in the SHP-1 (C453S) expressing cells than in the empty vector transfectants, a comparison of enolase phosphorylation with Src protein levels suggesting about 1.5-2 fold reduction in Src kinase activity in the former relative to latter cells both before and after stimulation. By contrast, phosphorylation of a number of species coprecipitated with Src from the epidermal growth factor-treated cells was markedly increased in the wild- FIG. 4. SHP-1 is a positive regulator of Src. A, Src kinase activity is reduced in motheaten thymocytes. Lysates were prepared from thymocytes obtained from motheaten (me/me), viable motheaten (me v /me v ), and congenic wild-type (ϩ/ϩ) mice and the lysate proteins (1.5 mg) immunoprecipitated with anti-Src antibody. Aliquots of the precipitates were then subjected to an in vitro kinase reaction, fractionated over SDS-PAGE, and transferred to Immobilon-P membranes and the membranes subjected to autoradiography (upper panel). Equal aliquots of the remaining immunoprecipitates were fractionated by SDS-PAGE and immunoblotted with anti-Src antibody (lower panel). B, Lck activity is present in motheaten thymocytes. Lck was immunoprecipitated from motheaten (me/me) and wild-type (ϩ/ϩ) mice, and the immune complexes then subjected to an in vitro kinase assay as above (upper panel). Equal levels of Lck protein were confirmed by Western immunoblotting of total cell lysates with anti-Lck antibody (lower panel). C, activity of motheaten Src is reconstituted by exogenous SHP-1. Src immunoprecipitates derived from motheaten thymocyte lysates (1.5 mg) were incubated at 37°C for 2 h in phosphatase reaction buffer with 30 ng GST-SHP-1 fusion protein.
Following washing in kinase buffer containing 2 mM orthovanadate, the complexes were subjected to an in vitro kinase reaction, fractionated by SDS-PAGE, transferred to Immobilon-P, and assessed by autoradiography (upper panel). Membranes were then subjected to anti-Src immunoblotting using ECL (lower panel). D, motheaten thymocytes show marked reduction in proportion of active versus inactive Src protein. Lysates were prepared from motheaten (me/me) and congenic wild-type (ϩ/ϩ) thymocytes and the lysate proteins (1.0 mg) were subjected to immunoblotting analysis with anti-Src (GD11) antibody (panel I). Src was immunoprecipitated from motheaten and wild-type thymocyte lysates using an anti-Src antibody, which recognizes only the active, COOH-terminal dephosphorylated form of Src (29), and the immune complexes were then subjected to an in vitro kinase reaction (as above) and the reaction products assessed by autoradiography (panel II). Remaining supernatants were then subjected to re-immunoprecipitation with the 327 anti-Src antibody, which recognizes both the active and inactive forms of Src followed by immunoblotting with the GD11 anti-Src antibody (panel III). E, the motheaten mutation does not alter expression of SHP-2 or CD45. Cell lysate proteins were obtained from motheaten (me/me) and wild-type (ϩ/ϩ) thymocytes and from Jurkat T cells and immunoblotted with antibodies to SHP-1, SHP-2, and CD45. F, dominant negative SHP-1 expression results in decreased Src activity in HEY cells; lysates were prepared from unstimulated HEY cells or HEY cells stimulated with epidermal growth factor (100 ng/ml) and stably transfected with pCMV4Neo (vector) or pCMV4Neo containing either the SHP-1 or SHP-1 (C453S) cDNAs and the lysate proteins (1 mg), then subjected to immunoprecipitation with anti-Src 327 antibody. The immune complexes were then collected and a portion of the precipitates subjected to an in vitro kinase reaction with enolase added as an exogenous substrate. Following fractionation and transfer to Immobilon-P, the complexes were analyzed by autoradiography (upper panel). Remaining aliquots of the precipitates were subjected to immunoblotting analysis with anti-Src antibody (lower panel). Positions of Src, enolase, and immunoglobulin heavy chain (Ig(H)), are indicated by arrows, and molecular mass standards are shown on the left. Relative specific kinase activity (below lower panel) represents the ratio between the levels of Src-induced enolase phosphorylation and Src protein as determined by densitometric analysis of the relevant bands shown in the upper and lower panel, respectively. type SHP-1-expressing relative to empty vector-transfected cells, suggesting that Src kinase activity on endogenous substrates is enhanced in the context of SHP-1 overexpression. Together these findings strongly suggest that the effects of SHP-1 on Src tyrosine phosphorylation are relevant to the activation of Src not only in hemopoietic, but also in epithelial cell lineages.
The data presented herein, revealing the capacity of SHP-1 to dephosphorylate Src preferentially at the COOH-terminal negative regulatory tyrosine as well as a positive correlation between the expression of SHP-1 and Src activity in vivo, provide compelling evidence that SHP-1 plays a role in the activation of Src in vivo. The identification of Src as a target for positive modulation by SHP-1, a PTP previously implicated in the negative regulation of a spectrum of signaling effectors, including the ZAP-70 and JAK2 tyrosine kinases (48,53), indicates that SHP-1 effects on cell behavior may be realized through enhancement as well as inhibition of intracellular signaling cascades. Thus, while the structural basis for and physiologic sequelae of SHP-1-Src association require further investigation, these data suggest SHP-1 regulatory effects on Src may be relevant to the genesis of human epithelial and potentially other cancers associated with increases in Src activity.