The Protein-tyrosine Phosphatase SHP-2 Associates with Tyrosine-phosphorylated Adhesion Molecule PECAM-1 (CD31)*

Aggregation of many cell-surface receptors results in tyrosine phosphorylation of numerous proteins. We previously observed the tyrosine phosphorylation of the platelet/endothelial cell adhesion molecule, PECAM-1 (CD31), after FcεRI stimulation in rat basophilic leukemia RBL-2H3 cells. Here we found that PECAM-1 was also transiently tyrosine-phosphoryated after adherence of these cells to fibronectin. Similarly aggregation of the T cell receptor on Jurkat cells also induced this tyrosine phosphorylation. The protein-tyrosine phosphatase SHP-2 is a widely expressed cytosolic enzyme with two Src homology 2 (SH2) domains. SHP-2, but not the related protein-tyrosine phosphatase SHP-1, associated with PECAM-1. This association of the two proteins correlated with the extent of the tyrosine phosphorylation of PECAM-1. A fusion protein containing the two SH2 domains of SHP-2 precipitated PECAM-1 from cell lysates and also directly bound to phosphorylated PECAM-1. In immune precipitate phosphatase assays, there was tyrosine dephosphorylation of PECAM-1. Therefore, integrin and immune receptor activation results in tyrosine phosphorylation of PECAM-1 and the binding of the protein-tyrosine phosphatase SHP-2, which could regulate receptor-mediated signaling in cells.

Tyrosine phosphorylation and dephosphorylation of molecules regulated by both protein-tyrosine kinases and phosphatases are critical events in signal transduction for cell growth, differentiation, and metabolism. The phosphorylation state of these proteins controls protein-protein association or dissociation thereby propagating and regulating downstream signal transduction.
PECAM-1 is a member of the immunoglobulin superfamily of cell adhesion molecules expressed on platelets, endothelial cells, and cells of the myeloid lineage including leukocytes, monocytes, some T cell subsets, and basophils (8,9). It functions in inter-endothelial cell adhesion, leukocyte-endothelial interactions, and in the transendothelial migration of cells (8,10). PECAM-1 is a single chain integral membrane glycoprotein containing six extracellular Ig-like homology domains, a single transmembrane region, and a cytoplasmic tail of ϳ118 amino acids. The cytoplasmic domain has numerous serine, threonine, and tyrosine residues that could potentially become phosphorylated. Receptor activation results in phosphorylation of PECAM-1 on serine and tyrosine residues (7,(11)(12)(13)(14). Although the function of PECAM-1 as an adhesion molecule is regulated by its cytoplasmic domain, little is known about the significance of tyrosine phosphorylation of PECAM-1.
The protein-tyrosine phosphatase SHP-2 (previously called SH-PTP2, Syp, PTP1D, PTP2C, and SH-PTP3) is a ubiquitously expressed cytosolic protein that contains two aminoterminal tandem SH2 domains and a carboxyl-terminal catalytic domain (15)(16)(17). SHP-2 is the homologue of both the Drosophila csw gene product, Csw and SHP-1 (previously called SH-PTP-1, PTP1C, HCP and SHP). SHP-2 associates with tyrosine-phosphorylated epidermal growth factor receptor, the platelet-derived growth factor receptor, insulin receptor substrate-1, Fc⑀RI, and with the T and B cell receptors (18 -20). It becomes tyrosine-phosphorylated upon cell stimulation and may act either as a negative regulator for receptor function or as a positive effector for downstream signaling (21-24).
Previously we observed that the minimal tyrosine phosphorylation of PECAM-1 in nonstimulated cells was dramatically increased after Fc⑀RI aggregation (7). Here we report that PECAM-1 was tyrosine-phosphorylated after either cell adherence or activation of other immune receptors such as the T cell receptor. The protein-tyrosine phosphatase SHP-2 but not SHP-1 was associated with tyrosine-phosphorylated PECAM-1. The tyrosine phosphorylation of PECAM-1 and the recruitment of protein-tyrosine phosphatases may play an important role in regulating signaling from cell-surface receptors.

EXPERIMENTAL PROCEDURES
Materials and Antibodies-Glutathione-Sepharose 4B was purchased from Pharmacia Biotech, Inc. MOPS was obtained from Fluka (Ronkonkoma, NY). All other materials not indicated under "Experimental Procedures" were described previously (7,25,26).
Polyclonal rabbit anti-SHP-2 (N-16 and C-18), polyclonal goat anti-PECAM-1, and monoclonal anti-GST or anti-CD3 (21-L5) antibodies were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Monoclonal anti-SHP-1 and anti-SHP-2 antibodies were obtained from Transduction Laboratories (Lexington, KY). The monoclonal anti-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The Jurkat T cell line (clone E6 -1) was obtained from the ATCC (Rockville, MD) and maintained in suspension culture in RPMI 1640 medium. For T cell receptor stimulation, 2 ϫ 10 7 cells were washed twice with ice-cold RPMI 1640 medium containing 0.1% BSA and 10 mM Tris-HCl, pH 7.5. Cells were then incubated on ice with 1 g/ml monoclonal anti-CD3 antibody for 30 min followed by 10 g/ml F(abЈ) 2 fragment of goat anti-mouse IgG for 3 min at 37°C. In some experiments the cells were treated with pervanadate as described above.
Post-nuclear supernatants were prepared by centrifuging the cell lysates at 16,000 ϫ g for 30 min and then precleared with Sepharose 4B beads. Immunoprecipitation was with 20 l of the anti-PECAM-1 antibody R23 coupled to Sepharose 4B beads, 4 g of rabbit anti-SHP-2, or 10 g of rabbit anti-PECAM-1 antibodies prebound to protein A-Sepharose 4B beads at 4°C for 90 min. The beads were then washed five times with lysis buffer, and proteins were eluted from the beads by boiling in SDS-PAGE sample buffer.

Expression of GST Fusion Proteins and Precipitation of Proteins with GST Fusion Proteins-
The plasmids containing the two SH2 domains of SHP-2, the full length of SHP-1 and SHP-2 as GST fusion proteins were kindly provided by Dr. Gen-Sheng Feng (Indiana University) and Dr. Benjamin G. Neel (Beth Israel Hospital, Boston). The proteins were expressed in Escherichia coli, affinity purified on glutathione-Sepharose beads, and characterized as recommended by the manufacturer. Precipitation with the 100 pmol of the GST fusion protein was as described previously (20).
Cell Adhesion Assays-This was as described previously (7,30) with the modification that bacterial culture type Petri plates (100-mm diameter) were used after coating with either 15 ml of 10 g/ml fibronectin (Calbiochem) or 15 ml of phosphate-buffered saline. RBL-2H3 cells at 5 ϫ 10 5 cells/ml in Eagle's modified essential medium containing 10 mM Tris, pH 7.4, were added to each plate, and the plates were incubated at 37°C for the indicated times. Although the cells did not adhere to the BSA-coated dishes, more than 90% of the cells attached to fibronectincoated plates by 30 min. The nonadherent cells from the BSA-coated wells were recovered by centrifugation at 200 ϫ g for 5 min at 4°C. The cells were solubilized in 1 ml of Triton lysis buffer.
Immune Complex Phosphatase Assays-These assays were carried out as described previously (31). Briefly, the RBL-2H3 cells (5 ϫ 10 6 ) were solubilized with 1 ml of ice-cold Brij lysis buffer (3% Brij 96, 20 mM Tris, pH 7.5, 100 mM NaCl, 1 mM Na 3 VO 4 , 2 mM phenylmethanesulfonyl fluoride, 21 g/ml aprotinin). After immunoprecipitation with anti-PECAM-1, the beads were washed twice with ice-cold protein-tyrosine phosphatase assay buffer (80 mM MOPS, 10 mM EDTA, 10 mM dithiothreitol, pH 7.0) and resuspended with 50 l of this buffer in the presence or the absence of 1 mM Na 3 VO 4 . The beads were then incubated at the indicated temperatures. After the reaction, the beads were either added to 5 ϫ SDS-PAGE sample buffer or washed twice with Brij lysis buffer before the addition of sample buffer. The proteins were eluted from the beads by boiling for 5 min.
Blotting of Proteins-Total cell lysates and precipitated proteins were separated by SDS-PAGE (Novex, San Diego, CA) and then electrotransferred to polyvinylidene difluoride membranes (Immobilon, Millipore Corp., Bedford, MA). Immunoblotting or blotting with GST fusion proteins was as described previously (7,20).

RESULTS
Association of PECAM-1 with SHP-2-We recently observed that the adhesion molecule PECAM-1 (CD31) became tyrosinephosphorylated after Fc⑀RI aggregation. Therefore we investi-gated whether other signaling molecules associate with PECAM-1 after it becomes tyrosine-phosphorylated. For these experiments we utilized the following two cell lines: rat basophilic leukemia RBL-2H3 cells and the Jurkat T cells which are known to express PECAM-1 (7,32). In preliminary experiments cells were treated with pervanadate to maximally tyrosine phosphorylate PECAM-1, and the lysates were immunoprecipitated with monoclonal anti-PECAM-1 antibody coupled to beads. These precipitates were then analyzed by immunoblotting with anti-SHP-1 and anti-SHP-2 antibodies. There was strong association of SHP-2 with PECAM-1 after pervanadate treatment (data not shown). Experiments then studied the association of SHP-1 and SHP-2 when cells were stimulated with more physiological stimulants such as the calcium ionophore A23187 and activation of Fc⑀RI that induce the tyrosine phosphorylation of PECAM-1 (Fig. 1). The minimal amount of PECAM-1 in the SHP-2 immunoprecipitates from nonstimulated cells was dramatically increased after stimulation of the cells. In contrast, anti-SHP-1 coprecipitated very little, if any, PECAM-1. Therefore, although SHP-2 has homology to SHP-1, there was a dramatic difference in their association with PECAM-1. These data strongly suggested that PECAM-1 associates with SHP-2, and this interaction appeared to depend upon tyrosine phosphorylation of PECAM-1.
SHP-2 Association with PECAM-1(CD31) Is Regulated by Membrane Receptors-The next series of experiments examined whether the association of SHP-2 with PECAM-1 correlated with changes in the tyrosine phosphorylation of PECAM-1. As observed previously, the tyrosine phosphorylation of PECAM-1 in RBL-2H3 cells was detectable within 1 min after receptor aggregation and increased up to 30 min after Fc⑀RI stimulation (Fig. 2A). The amount of SHP-2 coimmunoprecipitated with anti-PECAM-1 antibody paralleled the tyrosine phosphorylation of PECAM-1. Similarly, changing the concentration of the anti-Fc⑀RI␣ antibody used for stimulation of the cells varied the extent of the tyrosine phosphorylation of PECAM-1 and in parallel the amount of coprecipitated SHP-2 (Fig. 2B). These data indicate a strong correlation between the tyrosine phosphorylation of PECAM-1 and its association with SHP-2.
The aggregation of the T cell receptor results in the activation of several protein-tyrosine kinases. When the cultured Jurkat T cell line was stimulated with anti-CD3, there was increased tyrosine phosphorylation of PECAM-1 and an increase in the coprecipitation of SHP-2 with PECAM-1 (Fig. 3). Therefore, the activation of immune receptors results in the tyrosine phosphorylation of PECAM-1 and its interaction with SHP-2.
Cell Adhesion Caused Tyrosine Phosphorylation of PECAM-1-Although PECAM-1 is an adhesion molecule, its aggregation with antibodies does not induce tyrosine phosphorylation of PECAM-1 itself or tyrosine phosphorylation of other cellular proteins (7). Here we investigated whether integrins could induce PECAM-1 tyrosine phosphorylation. When RBL-2H3 cells were added to either BSA or fibronectin-coated dishes, more than 90% of the cells adhered to fibronectin-coated surfaces, but none of them attached to BSA-coated dishes. After 30 min there was an increase in the tyrosine phosphorylation of PECAM-1 in adherent compared with nonadherent cells (Fig.  4A). The cells were still attached at 90 min, but the phosphorylation of PECAM-1 decreased to levels similar to nonadherent cells (data not shown). Although the tyrosine phosphorylation of PECAM-1 due to cell adherence was not as dramatic as that by Fc⑀RI stimulation, there was still increased association of SHP-2 with PECAM-1 after cell adhesion (Fig. 4, A and B). Interestingly, adhesion also increased the tyrosine phosphorylation of SHP-2.

PECAM-1 Was Precipitated by Fusion Proteins Containing
SH2 Domains of SHP-2-RBL-2H3 cells were either nonstimulated or stimulated with anti-Fc⑀RI␣ antibodies, and the cell lysates were precipitated with the different GST fusion proteins immobilized on glutathione coupled to beads (Fig. 5). PECAM-1 was detected in the precipitates with the SHP-2 but not the SHP-1 fusion proteins from anti-Fc⑀RI␣-stimulated cells. Although the two SH2 domains expressed separately bound PECAM-1, binding was better when both were expressed in tandem. After maximal tyrosine phosphorylation by pervanadate treatment, there was some precipitation of PECAM-1 by SHP-1 that was, however, still much less than with SHP-2 (data not shown). These data indicate that the association of PECAM-1 and SHP-2 is mediated by the SH2 domains of SHP-2 and requires the tyrosine phosphorylation of PECAM-1.

Direct Interaction of Fusion Protein Containing the Two SH2 Domains of SHP-2 with Tyrosine-phosphorylated PECAM-1-
Although there was association of PECAM-1 and SHP-2 in immunoprecipitates, it was still possible that this interaction was indirect and mediated by other proteins. Therefore, membrane binding studies were used to study the direct interaction of PECAM-1 with the SHP-2 fusion protein (Fig. 6). PECAM-1 was immunoprecipitated from both nonstimulated and stimulated cells and the proteins separated by SDS-PAGE and blotted with GST-SHP-2. As expected, the GST-SHP-2 fusion protein bound to PECAM-1 only when the lysates were from stimulated cells. These results indicate that the direct interaction of PECAM-1 with SHP-2 can be mediated by the SH2 domains of SHP-2.
Protein-tyrosine Phosphatase Activity of SHP-2 Coimmunoprecipitated with PECAM-1-We examined whether the SHP-2 coimmunoprecipitated with PECAM-1 was still enzymatically active and could dephosphorylate PECAM-1 in vitro. PECAM-1 was immunoprecipitated from stimulated RBL-2H3 cells and then subjected to immune complex phosphatase assays (Fig. 7). The extent of the tyrosine phosphorylation of PECAM-1 decreased during the 30 min of in vitro incubation. As previously observed, there was SHP-2 present in these immunoprecipitates. However, if the immunoprecipitates were washed after the in vitro phosphatase assay, there was dissociation of SHP-2 that correlated with the dephosphorylation of PECAM-1. These data strongly suggested that the SHP-2 in the immunoprecipitates still had catalytic activity in vitro and further show that the interaction of these molecules depends on the tyrosine phosphorylation state of PECAM-1.  (10 7 ) were stimulated with 1 g/ml monoclonal anti-CD3 antibody for 30 min on ice, transferred to 37°C, and treated with 10 g/ml goat anti-mouse IgG for 3 min. Lysates were immunoprecipitated (IP) with anti-PECAM-1 antibody and analyzed by immunoblotting with anti-phosphotyrosine (Anti-pTyr), anti-SHP-2, and anti-PECAM-1 antibodies.

DISCUSSION
Activation of cells by adherence to fibronectin or by aggregating immune receptors induced the tyrosine phosphorylation of PECAM-1. The Fc⑀RI-induced tyrosine phosphorylation of PECAM-1 is an early event, independent of Ca 2ϩ influx or of cell adhesion (7). Receptor aggregation is thought to result in activation of a protein-tyrosine kinase, probably Lyn, which results in tyrosine phosphorylation of the receptor subunits (33). Syk then binds to the tyrosine-phosphorylated receptor subunits and is activated to propagate downstream signals including the tyrosine phosphorylation of phospholipase C-␥ and the rise in intracellular calcium. There is minimal tyrosine phosphorylation of PECAM-1 in Syk-deficient cells (7). By analogy to those results the tyrosine phosphorylation of PECAM-1 in T cells must be downstream of the activation of the proteintyrosine kinase ZAP-70.
SHP-2 was associated with the tyrosine-phosphorylated PECAM-1. Recent studies suggest a role for the protein-tyrosine phosphatase SHP-2 in several signal transduction pathways (17). SHP-2 associates with the tyrosine-phosphorylated epidermal growth factor or platelet-derived growth factor receptors by its amino-terminal SH2 domain and functions as a negative regulator by dephosphorylating the autophosphorylated receptors (22,34). However, SHP-2 in some systems becomes tyrosine-phosphorylated and then functions as an adaptor protein with positive effects for downstream signaling. For Lysates from RBL-2H3 cells stimulated with anti-Fc⑀RI␣ antibody were immunoprecipitated with anti-PECAM-1 coupled to Sepharose 4B beads, and the beads were washed and assayed for protein-tyrosine phosphatase activity. At the indicated times the reaction was either stopped by the addition of 5 ϫ sample buffer or by washing twice with lysis buffer followed by the addition of 2 ϫ sample buffer. The precipitates were analyzed by immunoblotting with anti-phosphotyrosine (Anti-pTyr), anti-PECAM-1, or anti-SHP-2 antibodies. example, the Grb2-Sos complex binds to the tyrosine-phosphorylated SHP-2 bound to the platelet-derived growth factor receptors and this then activates mitogenic signaling by the Ras pathway (35). In activated RBL-2H3 cells Grb2 was coprecipitated with PECAM-1 and SHP-2 (data not shown). Therefore, the SHP-2 associated with PECAM-1 may function as a phosphatase and also as an adaptor molecule.
The SH2 domains of SHP-2 were critical for its association with PECAM-1. For such interactions the topology of the SH2 domains and of the phosphorylated tyrosines on the proteins is important (36). From studies of the crystal structure of SHP-2 the amino and carboxyl SH2 domains are widely separated with fixed and opposite orientation for their binding sites. Studies with synthetic peptides suggest that the amino-terminal SH2 domain of SHP-2 preferentially binds the Tyr(P)hydrophobic-X-hydrophobic amino acid sequence (37,38). The residue at pY-2 which is a valine at the site of SHP-2 binding to platelet-derived growth factor receptor may also contribute to binding (39). However, the binding preference of the carboxylterminal SH2 domain has not been reported. The deduced amino acid sequences of the cytoplasmic domains of PECAM-1 have four (human) or five (bovine and mouse) tyrosine residues. The sequences V 677 EYTEV or T 700 VYSEI of murine PECAM-1 could be potential sites for binding by the SH2 domains of SHP-2. The first of these sites is similar to the recently recognized tyrosine containing amino acid motif (I/V)XpYX 2 (L/V) which is important for down-regulating immune receptor function (40 -42). This immunoreceptor tyrosine-based inhibiting motif or ITIM is found on the cytoplasmic domain of several molecules including Fc␥RIIB in B cells and mast cells, CD22 in B cells, the killer-inhibitory receptors in NK cells, and gp49B1 in mast cells (43,44). Phosphorylation of the tyrosine in this ITIM recruits SHP-1, SHP-2, and the SH2 domain-containing inositol polyphosphate 5-phosphatase, SHIP, that down-regulates receptor-mediated signal transduction (45,46). Therefore, the cytoplasmic domain of PECAM-1 has similarities to these negative ITIM sequences and can recruit SHP-2.
The binding of tyrosine-phosphorylated peptides to the SH2 domain of SHP-2 increases its phosphatase activity (47,48). The enzymatic activity of protein-tyrosine phosphatases is also enhanced by phospholipids (49). Therefore, the recruitment of SHP-2 to the membrane by binding to the tyrosine-phosphorylated PECAM-1 could enhance its phosphatase activity by these two mechanisms. The recruitment of SHP-2 to PECAM-1 would bring it in close proximity to other potential substrates such as Fc⑀RI, the Src family tyrosine kinase Lyn, the proteintyrosine kinase Syk, the adaptor protein Shc, phospholipase C␥1, and of the SH2 domain-containing inositol polyphosphate 5-phosphatase, SHIP (26,50). Thus, SHP-2 could regulate the extent of the tyrosine phosphorylation of these molecules in receptor-activated cells thereby controlling the resulting downstream signals.
Adhesion receptor-mediated binding of cells to other cells or to the extracellular matrix results in aggregation of these receptors and the generation of intracellular signals (51). The cytoplasmic domain of adhesion molecules can control their function (52)(53)(54)(55)(56)(57). RBL-2H3 cells bind to fibronectin through integrin receptors resulting in cell spreading and a redistribution of the granules to the periphery of the cells (30,58). Aggregation of these adhesion receptors is accompanied by tyrosine phosphorylation of proteins such as pp125 FAK and of the cytoskeletal protein paxillin (59). Here we observed that adhesion to fibronectin induced the tyrosine phosphorylation of PECAM-1. Although this was not as dramatic as immune receptor stimulation, it still resulted in increased SHP-2 association with PECAM-1. Aggregation of PECAM-1 results in an increase in the adhesion mediated by integrins (32,60,61). The present results indicate that there are changes in PECAM-1 that are induced by binding to integrins. Therefore, there is cross-talk between PECAM-1 and integrins.
Many different PECAM-1 positive cells including monocytes, basophils, and T cells accumulate at sites of inflammation. These cells require multiple adhesive interactions with the endothelium to leave the circulation and migrate into sites of inflammation. PECAM-1 functions in the transmigration of cells across the endothelium. For example, anti-PECAM-1 antibodies by binding to leukocytes or to endothelial cells block the transmigration of leukocytes without inhibiting the adherence of the cells to the endothelium (62)(63)(64)(65). PECAM-1 also controls the function of other adhesion molecules; thus, aggregation of PECAM-1 with antibodies regulates the adhesive properties of ␤1 and ␤2 integrins (32,60,61,66). Therefore, SHP-2 by regulating the level of the tyrosine phosphorylation of PECAM-1 could control these functions of PECAM-1.