J Biol Chem, Vol. 275, Issue 8, 5453-5459, February 25, 2000

Dissecting the Interaction of SHP-2 with PZR, an Immunoglobulin Family Protein Containing Immunoreceptor Tyrosine-based Inhibitory Motifs^*

Runxiang Zhao and Zhizhuang Joe Zhao

From the Division of Hematology/Oncology, Department of Medicine, Vanderbilt-Ingram Cancer Center, Vanderbilt University, Nashville, Tennessee 37232-6305

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Tyrosine phosphorylation of membrane proteins plays a crucial role in cell signaling by recruiting Src homology 2 (SH2) domain-containing signaling molecules. Recently, we have isolated a transmembrane protein designated PZR that specifically binds tyrosine phosphatase SHP-2, which has two SH2 domains (Zhao, Z. J., and Zhao, R. (1998) J. Biol. Chem. 273, 29367-29372). PZR belongs to the immunoglobulin superfamily. Its intracellular segment contains four putative sites of tyrosine phosphorylation. By site-specific mutagenesis, we found that the tyrosine 241 and 263 embedded in the consensus immunoreceptor tyrosine-based inhibitory motifs VIYAQL and VVYADI, respectively, accounted for the entire tyrosine phosphorylation of PZR. The interaction between PZR and SHP-2 requires involvement of both tyrosyl residues of the former and both SH2 domains of the latter, since its was disrupted by mutating a single tyrosyl residue or an SH2 domain. Overexpression of catalytically inactive but not active forms of SHP-2 bearing intact SH2 domains in cells caused hyperphosphorylation of PZR. In vitro, tyrosine-phosphorylated PZR was efficiently dephosphorylated by the full-length form of SHP-2 but not by its SH2 domain-truncated form. Together, the data indicate that PZR serves not only as a specific anchor protein of SHP-2 on the plasma membrane but also as a physiological substrate of the enzyme. The coexisting binding and dephosphorylation of PZR by SHP-2 may function to terminate signal transduction initiated by PZR and SHP-2 and to set a threshold for the signal transduction to be initiated.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Protein tyrosine phosphorylation plays a pivotal role in cell proliferation, differentiation, and transformation (1). Tyrosine phosphorylation of membrane proteins occurs after stimulation of cells with extracellular stimuli that include growth factors, cytokines, and extracellular matrix (2). Tyrosine-phosphorylated membrane proteins serve as anchors for SH2¹ domain-containing molecules that can either potentiate or inhibit signal transduction (3). Since the SH2 domains bind specifically with tyrosine-phosphorylated motifs, tyrosine kinase signaling pathways gain specificity from the intrinsic binding preferences of SH2 domains for short sequences that flank phosphotyrosine. Recently, we have purified and subsequently cloned a tyrosine-phosphorylated membrane protein that we designated PZR (4). PZR is a member of the immunoglobulin superfamily. Its extracellular segment has significant sequence homology to myelin P0, while its intracellular portion has two tyrosine phosphorylation sites resembling immunoreceptor tyrosine-based inhibitory motifs (ITIMs). With a (V/I)XYXX(L/V) consensus sequence, the ITIM was initially defined in FCRIIB (5) and later was found in many other hematopoietic cell inhibitory proteins including KIR (6), LAIR-1 (7), CD22 (8), PIR-B (9), PECAM-1 (10), and widely distributed receptor-like protein SIRP/SHPS-1 (11, 12). ITIMs are believed to play an inhibitory role in cell signaling by recruiting terminating enzymes including protein-tyrosine phosphatases SHP-1 and SHP-2 and inositol phosphatase SHIP (13-17). Like SIRP/SHPS-1, PZR is widely expressed. However, it specifically interacts with SHP-2 but not with SHP-1 (4).

SHP-2 is a widely distributed intracellular protein-tyrosine phosphatase containing SH2 domains (18). It shares high homology with Drosophila corkscrew, which plays a positive role in the transduction of the torso signal (19). Studies have shown that SHP-2 plays a similar positive role in mammalian cell signaling. Microinjection of anti-SHP-2 antibodies or glutathione S-transferase fusion protein encoding the SH2 domains of SHP-2 blocks insulin-stimulated DNA synthesis (20). Expression of a catalytically inactive cysteine-to-serine mutant of SHP-2 inhibits activation of MAP kinase induced by insulin, epidermal growth factor, platelet-derived growth factor, and -thrombin; decreases activation of STAT transcription factors by interferon /; and suppresses early gene transcription, DNA synthesis, and cell proliferation (21-27). In Xenopus embryogenesis, a dominant negative mutant of SHP-2 blocks fibroblast growth factor- and activin-mediated induction of mesoderm and MAP kinase activation induced by fibroblast growth factor (28). Most recently, it was shown that disruption of the mouse SHP-2 gene caused the death of mouse embryos at midgestation (29). Further studies with cells derived from SHP-2-deficient mice demonstrated impairment in FGF-induced MAP kinase activation (29), erythropoiesis (30), and cell migration (31). Although SHP-2 plays a crucial role in intracellular signaling elicited by various growth factors and hormones, the mechanism by which this occurs and the direct targets of SHP-2 are not well understand.

In the present study, we have characterized the interaction of SHP-2 with PZR by expressing various mutant forms of the proteins in cells. Our results demonstrate that tyrosine 241 and 263 embedded in the ITIMs of PZR account for the entire tyrosine phosphorylation of PZR and that both tyrosine residues are required for binding of SHP-2 through its SH2 domains. Furthermore, PZR serves not only as a specific anchor protein of SHP-2, but it is also a physiological substrate of SHP-2. The physiological meaning of the interaction will be discussed.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Materials-- Jurkat and 293 cells were obtained from the American Type Culture Collection. Polyclonal anti-SHP-2 serum 1263 and anti-PZR serum 105 were raised in rabbits against the SH2 domain-truncated form of SHP-2 and a GST fusion protein of the intracellular domain of PZR, respectively, as described (4, 32). Monoclonal anti-phosphotyrosine 4G10 and anti-SHP-2 were purchased from Upstate Biotechnology, Inc. and Transduction Laboratories, respectively. Pervanadate was made by mixing 0.1 M sodium vanadate and 0.1 M H₂O₂ and incubating at room temperature for 20 min before adding to cells (33). The SH2 domains of SHP-2, designated 2SH2, N-SH2, and C-SH2, were expressed as GST fusion proteins and purified from Escherichia coli by using glutathione-Sepharose beads. These proteins corresponded to amino acid residues 3-223, 3-109, and 97-223 of the SHP-2 molecule, respectively

cDNA Constructs of PZR and SHP-2-- Figs. 1 and 2 list the cDNA constructs of PZR and SHP-2 used in this study, respectively. The PZR constructs were made with the pCDNA3 vector, and the SHP-2 constructs were built with the pRC/CMV vector, an earlier version of the pCDNA3 vector. Both expression vectors contain the cytomegalovirus promoter for high expression in mammalian cells and the neomycin resistance gene (neo) for selection with G418 sulfate. Construction of PZR, SHP-2, and catalytically inactive Cys-to-Ser mutant SHP-2 (C-S) have been previously described (4, 23). Mutations of Tyr to Phe in PZR and of Arg to Lys and Cys to Ser in SHP-2 were carried out by polymerase chain reaction, and the mutageneses were confirmed by DNA sequencing. SHP-2 represents an SH2 domain-truncated form of SHP-2 and corresponds to amino acid residues 200-593, while 2SH2, the protein-tyrosine phosphatase domain-truncated protein, contains amino acid residues 1-210. Both were made by truncation of cDNA at convenient restriction sites and religating with appropriate linkers with an initiation codon or termination codon as required. Myr-SHP-2, which has a myristoylation tag for membrane targeting, was made by adding a consensus myristoylation sequence corresponding to the 15 N-terminal amino acid residues (MGSNKSKPKDASQRR) of human c-Src (34, 35).

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Schematic diagram of PZR constructs. One-letter amino acid codes are used.

View larger version (37K): [in this window] [in a new window]   Fig. 2.   Schematic diagram of SHP-2 constructs. One-letter amino acid codes are used. Myr represents the consensus myristoylation sequence MGSNKSKPKDASQRR.

Transient Expression of PZR in Jurkat Cells and SHP-2 in 293 Cells-- Transfection of Jurkat cells with various forms of PZR constructs was performed by electroporation as described previously (4). The cells were grown to ~2 × 10⁶/ml in RPMI 1640 medium supplemented with 10% fetal calf serum and 50 µg/ml each of streptomycin and penicillin. Cells (1 × 10⁷) were collected by centrifugation, washed with plain medium without serum, and then resuspended in 300 µl of the same plain medium. The cDNA plasmid (20 µg) in 100 µl of water was added to the cells. The electroporation was performed under 950 microfarads, 250 V, and 72 ohms with 4-mm cuvettes by using the ECM 600 electroporation system (BTX Inc.). After sitting on ice for 15 min, the cells were transferred to 5 ml of complete medium and continued in culture for 72 h before further treatment. Transfection of 293 cells was carried out according to a calcium phosphate co-precipitation protocol (36). Briefly, 293 cells were grown to confluency in DMEM supplemented with 10% fetal calf serum and 50 µg/ml each of streptomycin and penicillin and then were split 1:8 and cultured overnight to ~25% confluency with 4 ml of medium in 6-cm plates. This was followed by the addition of calcium-DNA precipitates made by mixing 10 µg of total DNAs and 0.25 M CaCl₂ in BES-buffered saline containing 25 mM BES-NaOH, pH 7.3, 0.14 M NaCl, Na₂HPO4. After a 24-h incubation, cells were washed with phosphate-buffered saline and then cultured in fresh medium for another 24 h before harvesting.

Cell Stimulation, Extraction, Immunoprecipitation, and Western Blotting Analyses-- Transfected Jurkat and 293 cells were left untreated or treated with 0.1 mM pervanadate for 30 min. After washing with ice-cold phosphate-buffered saline, the cells were lysed in buffer A containing 50 mM -glycerophosphate (pH 7.3), 0.1 M NaCl, 5 mM EDTA, 1 mM EGTA, 5 mM -mercaptoethanol, 1% Triton X-100, 0.2 mM Na₃VO₄, 0.1 µM microcystin, 1.0 mM benzamidine, 0.1 mM phenylmethylsulfonyl fluoride, 20 µg/ml leupeptin, 1 µM pepstatin A, and 1 µg/ml aprotinin. Extracts were cleared by centrifugation. For immunoprecipitation, cell extracts were incubated overnight with the anti-PZR and anti-SHP-2 antibodies prebound to protein A-Sepharose. The beads were washed three times with an IP washing buffer containing 50 mM -glycerophosphate (pH 7.3), 0.15 M NaCl, 2 mM EDTA, 1 mM EGTA, 5 mM -mercaptoethanol, 0.1% Triton X-100, and 0.2 mM Na₃VO₄. For Western blot analyses, samples were separated by 10% SDS-PAGE and transferred to polyvinylidene difluoride membranes. The membranes were probed with various primary antibodies and were detected by using the ECL system with horseradish peroxidase-conjugated secondary antibodies (Amersham Pharmacia Biotech).

In Vitro Dephosphorylation of PZR by SHP-2-- One plate (150 mm) of 293 cells overexpressing PZR, PZR(F241), or PZR(F263) was treated with 0.1 mM pervanadate for 30 min, and cells were extracted as described above. The cell extract was subjected to immunoprecipitation with anti-PZR serum as described above. After washing three times with the aforementioned immunoprecipitation washing buffer, beads were washed with protein-tyrosine phosphatase assay buffer containing 25 mM Tris-HCl (pH 7.0), 1.0 mM EDTA, 2 mM dithiothreitol, and 0.1% Triton X-100. The beads were suspended in 0.8 ml of the protein-tyrosine phosphatase assay buffer. Aliquots of 40 µl of the suspension were used for dephosphorylation reactions that were started by the addition of 0.6 µg of purified full-length recombinant SHP-2 or its SH2 domain-truncated form, SHP-2 (32). The reactions were allowed to proceed at room temperature for up to 1 h before termination with SDS gel sample buffer. Dephosphorylation of PZR was analyzed by Western blot with anti-phosphotyrosine antibody.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Tyr²⁴¹ and Tyr²⁶³ of PZR Are Phosphorylated and Responsible for Binding of SHP-2-- Our previous studies have shown that tyrosine-phosphorylated PZR specifically recruits SHP-2 (4). Among the four tyrosyl residues in the intracellular portion of the protein, Tyr²⁴¹ and Tyr²⁶³ embedded in the ITIMs are most likely phosphorylated and responsible for the binding of SHP-2. To verify this, we constructed three Tyr-to-Phe mutant forms of PZR, namely PZR(F241), PZR(F263), and PZR(F241,263), as shown in Fig. 1. These mutant constructs together with the pCDNA3 vector and the native PZR construct were used to transfect Jurkat cells by electroporation, and the transfected cells were stimulated with pervanadate. The cell extracts were subjected to immunoprecipitation with anti-PZR and anti-SHP-2 antibodies. This was followed by Western blot analyses with anti-phosphotyrosine as shown in Fig. 3A. In comparison with the marked tyrosine phosphorylation of the native form of PZR, mutation of either Tyr²⁴¹ or Tyr²⁶³ caused a significant decrease in tyrosine phosphorylation, whereas mutation of both tyrosine residues to phenylalanine resulted in a total loss of tyrosine phosphorylation. Western blot analyses with anti-PZR antibody revealed essentially equal expressions of PZR and its mutant in Jurkat cells. These data thus indicate that Tyr²⁴¹ and Tyr²⁶³ are responsible for tyrosine phosphorylation of PZR. Furthermore, as shown by the tyrosine-phosphorylated SHP-2 co-immunoprecipitated with PZR, binding of SHP-2 with PZR was abolished by mutation of a single site (Fig. 3A), suggesting that simultaneous phosphorylation of both sites is required for recruitment of SHP-2 to PZR. Immunoprecipitation with anti-SHP-2 further verified the results as shown in Fig. 3B. Both anti-phosphotyrosine and anti-PZR blots revealed strong binding of SHP-2 with the native form of PZR and minimal binding with the mutant forms of PZR. Note that the tyrosine-phosphorylated protein of about 95 kDa that co-immunoprecipitated with SHP-2 in vector control cells was absent in cells overexpressing the native form of PZR. This is probably due to a competition of PZR with the protein for binding to SHP-2, presumably through a similar interaction mechanism. We further transfected 293 cells with the PZR constructs to confirm the data obtained with Jurkat cells. As shown in Fig. 3C, mutation of either Tyr²⁴¹ or Tyr²⁶³ markedly reduced the tyrosine phosphorylation of PZR, while mutation of both residues totally diminished the phosphorylation. In all cases, the mutation caused a total loss of association of PZR with SHP-2. It should be noted that the exogenously introduced PZR constructs were overexpressed 20-40-fold, and thus the endogenous PZR in the cells did not have a major interfering effect on the results.

View larger version (31K): [in this window] [in a new window]   Fig. 3.   Tyr241 and Tyr263 of PZR are phosphorylated, and both are required for binding of SHP-2. Jurkat cells (A and B) and 293 cells (C) were transfected with pCDNA3 vector or constructs containing PZR, PZR(F241), PZR(F263), or PZR(F241,263) as described in the legend to Fig. 1. Cells were treated with 0.1 mM pervanadate for 30 min. Cell extracts were immunoprecipitated with anti-PZR or polyclonal anti-SHP-2 antibodies, and the immunoprecipitates were subjected to Western blot analyses with anti-phosphotyrosine and anti-PZR as indicated.

SH2 Domains of SHP-2 Are Required for Association with PZR-- The binding SHP-2 with PZR is presumably mediated by the interaction between SH2 domains of SHP-2 and the ITIMs of PZR. To confirm this, we performed site-specific mutagenesis of the SH2 domain of SHP-2. The crystal structure of SH2 domains revealed that residues Arg^A2 and Arg^B5 have a crucial role in binding by chelating the phosphotyrosine phosphate (3). The latter is within the conserved FLVRES sequence and corresponds to Arg³² and Arg¹³⁸ of the N-terminal and C-terminal SH2 domains of SHP-2, respectively. The Arg-to-Lys mutant forms of SHP-2 and SH2 domain-truncated SHP-2 are shown by the schematic diagram in Fig. 2. These constructs, including SHP-2, SHP-2(R³²-K), SHP-2(R¹³⁸-K), and the SH2 domain-truncated form, SHP-2, were used to transfect 293 cells. To increase the level of PZR, cells were co-transfected with the native form of PZR as described above. Overexpression of PZR is necessary to overcome the effects of endogenous SHP-2 that is more abundant than endogenous PZR in 293 cells. The reason for using 293 cells instead of Jurkat cells as described above is that we had difficulty expressing a high level of SHP-2 in Jurkat cells. The transfected 293 cells were treated with 0.1 mM pervanadate to induce tyrosine phosphorylation, and cell extracts were immunoprecipitated with anti-SHP-2 antibodies. Western blot analysis of the immunoprecipitates with anti-PZR is shown in Fig. 4. Compared with the native form of SHP-2, mutation of either arginyl residue caused marked decrease in binding of PZR with SHP-2. Moreover, some of the binding seen can be attributable to co-immunoprecipitation with endogenous SHP-2 as found in cells transfected with the SH2 domain-truncated SHP-2. This study indicates that the tandem SH2 domains of SHP-2 are responsible for binding with PZR and that both are required.

View larger version (39K): [in this window] [in a new window]   Fig. 4.   Both SH2 domains of SHP-2 are required for association with PZR. A, human 293 cells were co-transfected with PZR and SHP-2, SHP-2 (R32-K), SHP-2(R138-K), or SHP-2, as illustrated in Fig. 2. Cells were treated with 0.1 mM pervanadate for 30 min. Cell extracts were immunoprecipitated with anti-SHP-2 antibodies, and the immunoprecipitates were subjected to Western blot analyses with anti-PZR and anti-SHP-2 as indicated. IgG denotes the heavy chain of immunoglobulin G. B, cell extracts from pervanadate-treated 293 cells that were transfected with PZR, PZR(F241), or PZR(263) were incubated with glutathione-Sepharose beads carrying ~1 µg of GST fusion proteins corresponding to the 2SH2, N-SH2, C-SH2 domains of SHP-2. After overnight incubation at 4 °C followed by washing with the aforementioned IP washing buffer, PZR bound to the beads was analyzed by Western blotting with anti-PZR.

To further verify the results, we performed in vitro GST fusion protein "pull-down" experiments. GST fusion proteins, 2SH2, N-SH2, and C-SH2, corresponding to the tandem SH2 domains, N-terminal SH2 domain, and C-terminal SH2 domain, respectively, were immobilized on the glutathione-Sepharose beads. Aliquots of beads containing ~1 µg of each protein were incubated with cell extracts obtained from pervanadate-treated 293 cells, which were transfected with PZR, PZR(F241), or PZR(263). After overnight incubation at 4 °C followed by washing with the aforementioned IP washing buffer, PZR bound to the fusion protein-carrying beads was analyzed by Western blotting with anti-PZR. As shown in Fig. 4B, only the tandem SH2 domain fusion protein was able to pull down native PZR, and neither N-SH2 nor C-SH2 domain showed any significant binding to PZR. Furthermore, the interaction of the tandem SH2 domains with PZR was disrupted by mutation of either Tyr²⁴¹ or Tyr²⁶³ of PZR. This provides further evidence that both ITIMs of PZR and both SH2 domains of SHP-2 are required to mediate the interaction of the two molecules.

Expression of Catalytically Mutant Forms of SHP-2 Causes Tyrosine Phosphorylation of PZR-- Our previous studies demonstrated that overexpression of catalytically inactive Cys-to-Ser mutant but not the native form of SHP-2 caused hyperphosphorylation of PZR, suggesting that PZR is a putative substrate of SHP-2 (23). One possible mechanism by which the catalytically inactive SHP-2 prevents dephosphorylation of PZR is by binding the phosphotyrosyl motif through the catalytic domain directly. However, our study described above suggests that interaction between PZR and SHP-2 is mediated by interaction between tyrosine-phosphorylated ITIMs of PZR and SH2 domains of SHP-2, implying that PZR primarily serves as an anchor for SHP-2. Binding of SH2 domains to the ITIMs of PZR would also prevent the latter from dephosphorylation by SHP-2 or other protein-tyrosine phosphatases. To clarify this, we transfected 293 cells with a variety of catalytically inactive forms of SHP-2 as illustrated in Fig. 2. This was followed by analysis of tyrosine phosphorylation of intracellular proteins in nonstimulated cells. The expression of various forms of SHP-2 was determined by Western blotting with anti-SHP-2 serum, while tyrosine phosphorylation of PZR was analyzed by anti-phosphotyrosine immunoblotting of whole cell extracts and anti-PZR immunoprecipitates (Fig. 5). As expected, overexpression of SHP-2(C-S), which has intact SH2 domains, caused strong phosphorylation of PZR, which was also associated with SHP-2(C-S). When either one of the SH2 domains was mutated, phosphorylation of PZR had a marked decrease but was still visible. However, essentially no association of the mutant SHP-2 with PZR was found. When both SH2 domains were removed, no phosphorylation of PZR was observed, even when the truncated Cys-to-Ser mutant was targeted to the plasma membrane by attaching a myristoylation tag. This indicates that the catalytic domain alone is not sufficient to induce tyrosine phosphorylation of PZR. This may be attributable to a low affinity of the SH2 domain-truncated Cys-to-Ser mutation to the ITIMs of PZR. These results also suggest that the SH2 domain of SHP-2 is responsible for preventing dephosphorylation of PZR. This is further supported by the fact that expression of two SH2 domains of SHP-2 alone in 293 cells caused tyrosine phosphorylation of PZR (Fig. 5, last lane in each panel). Nonetheless, the fact that overexpression of the native form of SHP-2 does not enhance tyrosine phosphorylation of PZR (see Ref. 23) suggests that native SHP-2 is able to dephosphorylate PZR. In this regard, pervanadate-induced tyrosine phosphorylation of PZR and its association with SHP-2 can also be attributed to inactivation of SHP-2. It should also be pointed out that although protecting dephosphorylation of PZR by the SH2 domains of SHP-2 has a major role in causing basal tyrosine phosphorylation of PZR, activation of certain protein-tyrosine kinases cannot be ruled out. In fact, the latter might be responsible for the lower level phosphorylation of PZR induced by the Arg-to-Lys mutants of SHP-2 that showed no binding with PZR.

View larger version (35K): [in this window] [in a new window]   Fig. 5.   Tyrosine phosphorylation of PZR in cells overexpressing the catalytically inactive mutants of SHP-2. Human 293 cells were transfected with SHP-2, SHP-2(C-S), SHP-2(C-S,R32-K), SHP-2(C-S,R138-K), SHP-2(C-S), Myr-SHP-2, or 2SH2 as shown in Fig. 2. Cells were extracted with buffer A, and the extracts were immunoprecipitated with anti-PZR and rabbit polyclonal anti-SHP-2. The extracts and the immunoprecipitates were subjected to Western blot analyses with anti-phosphotyrosine, anti-PZR, polyclonal anti-SHP-2 (upper right panel), and monoclonal anti-SHP-2 (lower left panel) as indicated. The positions of SHP-2, SHP-2, 2SH2, and PZR are indicated. It should be pointed out that the polyclonal anti-SHP-2 antibody could detect 2SH2 by Western blot but failed to precipitate it.

To further reveal the tyrosine phosphorylation of PZR accompanying expression of catalytically inactive mutant SHP-2, we co-expressed SHP-2(C-S) with PZR and its Tyr-to-Phe mutant forms in 293 cells. As shown in Fig. 6, in light of equal expression levels of PZR, PZR(F241), PZR(F263), and PZR(F241,263), only native PZR was tyrosine-phosphorylated and was associated with SHP-2. The marginal phosphorylation observed with the PZR mutants may be attributable to the endogenous PZR in 293 cells. These results further verified the data obtained from pervanadate treatment of cells as described in the legend to Fig. 3. The fact that PZR(F241) and PZR(F263) did not show significant phosphorylation as observed in pervanadate-treated cells might be due to a lower level of tyrosine kinase activation.

View larger version (31K): [in this window] [in a new window]   Fig. 6.   Tyr241 and Tyr263 of PZR are responsible for tyrosine phosphorylation of PZR in cells overexpressing the catalytically inactive Cys-to-Ser mutant of SHP-2. PZR, PZR(F241), or PZR(F263) were co-transfected with SHP-2(C-S) into 293 cells. Cells were extracted with buffer A, and the extracts were immunoprecipitated with anti-PZR. The immunoprecipitates were subjected to Western blot analyses with anti-phosphotyrosine followed by reblotting with monoclonal anti-SHP-2 and rabbit polyclonal anti-PZR as indicated. The positions of SHP-2, IgG heavy chain, and PZR are indicated.

PZR Is Efficiently Dephosphorylated by Full-length SHP-2 but Not by Its SH2 Domain-truncated Form-- To further verify the specific dephosphorylation of PZR by SHP-2, we performed in vitro dephosphorylation of PZR. Tyrosine-phosphorylated PZR was immunopurified from pervanadate-treated 293 cells and incubated with full-length SHP-2 and its SH2 domain-truncated form, SHP-2. Previous studies have shown that truncation of the SH2 domains causes up to 50-fold activation of the enzyme (32). For the samples used in this particular study, the specific activities toward 10 mM para-nitrophenylphosphate analyzed at pH 5.0 were 1,800 and 33,000 units/ml for the full-length SHP-2 and the truncated enzyme, respectively. However, when equal protein amounts of the enzymes were used to treat tyrosine-phosphorylated PZR, full-length SHP-2 caused rapid dephosphorylation, while the truncated enzyme had essentially no effect. These data indicate that binding of SHP-2 to PZR through its SH2 domains greatly enhanced activity to PZR, while the SH2 domain-truncated SHP-2, despite its high activity toward low molecular weight artificial substrate para-nitrophenylphosphate, does not have sufficient affinity to bind and thereby to dephosphorylate PZR. This possibility is further supported by the fact that the Tyr-to-Phe mutants of PZR that do not bind SHP-2 could not be efficiently dephosphorylated by SHP-2 (Fig. 7, lower panels). The high affinity of the full-length SHP-2 to PZR is presumably conferred by the high affinity, specific SH2 domain-ITIMs interaction. The dephosphorylation probably occurs through conformational changes (which may be slow) within a PZR-SHP-2 complex that render the catalytic domain of SHP-2 capable of attacking the phosphotyrosyl residues of PZR. Nonetheless, one cannot rule out the possibility of an intercomplex reaction in which one PZR-complexed SHP-2 molecule attacks a different PZR molecule in another complex, although the efficiency might be lower in comparison with the intracomplex reaction. It should be noted that dephosphorylation of PZR by SHP-2 was not complete. This might be due to the loss of binding of PZR to SHP-2 after dephosphorylation of one of its tyrosine residues. In addition, a competition from pervanadate-inactivated SHP-2 that was co-immunoprecipitated with tyrosine-phosphorylated PZR may also cause the incomplete dephosphorylation. In any case, a higher concentration of SHP-2 (up to 5 µg/ml) helped to push the dephosphorylation to near completion (data not shown). Together, the data suggest that PZR is a substrate as well as an anchoring protein of SHP-2 and that efficient dephosphorylation requires binding of SHP-2 to PZR through the interaction between ITIMs and SH2 domains.

View larger version (27K): [in this window] [in a new window]   Fig. 7.   Dephosphorylation of PZR by SHP-2. Human 293 cells transfected with PZR, PZR(F241), or PZR(F263) were treated with 0.1 mM pervanadate for 30 min. The cell extracts were subjected to immunoprecipitation with anti-PZR and washed as described under "Experimental Procedures." Aliquots of immunoprecipitates were incubated with 0.6 µg of purified full-length SHP-2 or SH2 domain-truncated SHP-2 as indicated. The reaction was terminated by the addition of SDS gel sample buffer after the indicated periods of time (0-60 min). Dephosphorylation of PZR was analyzed by Western blot with anti-phosphotyrosine antibody.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

By specifically mutating tyrosyl residues of PZR and the SH2 domain of SHP-2, the present study demonstrated that Tyr²⁴¹ and Tyr²⁶³ embedded in ITIMs are responsible for phosphorylation of PZR and that both are required for binding with SHP-2 through its SH2 domains, implying that PZR serves as an anchor protein of SHP-2 on the plasma membrane. SH2 domain proteins transmit intracellular signals initiated by activated tyrosine kinase-linked receptors. Three-dimensional structures suggest mechanisms by which tandem SH2 domains might confer higher specificity than individual SH2 domains (37, 38). In vitro studies with phosphopeptides revealed that tandem SH2 domains bind bisphosphotyrosyl peptides 20-50-fold more strongly than individual SH2 domains (39). It was thus hypothesized that high biological specificity is conferred by the simultaneous interaction of two SH2 domains in a signaling enzyme with diphosphorylated motifs in activated receptors or their substrates. By showing that efficient co-immunoprecipitation of PZR with SHP-2 requires interaction of the tandem SH2 domains of SH2 and both ITIMs of PZR, our study thus provides evidence at the cellular level. SHP-2 has been shown to bind to a number of growth factor receptors (17, 18). However, in many cases, the interactions seem to be medicated by a single SH2 domain, and only a small fraction of SHP-2 and receptors were found associated. The physiological meaning of this binding might be different. Many signaling molecules with tandem SH2 domains interact with bisphosphotyrosyl motifs. These motifs include immunoreceptor tyrosine-based activation motifs (ITAMs) and the aforementioned ITIMs. Studies have shown that the space between the phosphotyrosine residues of the motifs is crucial for binding. In ITAMs, the tyrosyl residues are usually separated by 9-11 amino acid residues (13). Crystal structure of the tandem SH2 domains of ZAP-70 suggests that such a space would be optimal to bind its correspondent ITAMs (40). For ITIMs, it appears that more amino acid residues are required to fill the space. The ITIMs in PZR are separated by 21 amino acids, whereas those in KIR are separated by 29, those in LAIR-1 are separated by 29, those in PIR-B are separated by 29, those in PECAM are separated by 22, those in CD22 are separated by 33 and 19, and those in SHPS-1/SIRP are separated by 23 and 25. This long stretch can be explained by the fact that the correspondent tandem SH2 domains (e.g. SHP-2) in the intact enzyme are oriented differently, spaced widely and perpendicular to one another, so that they require the bisphosphotyrosyl ITAMs sequences to change direction to bind both sites (37, 38). It should be noted that while the ITIMs found in other proteins have one or more proline residues separating the tandem ITIMs, the 21 amino acid residues between the two ITIMs of PZR have two consecutive glycyl residues instead, which might also facilitate a turn. In addition, this stretch contains 4 seryl residues, each surrounded by charged amino acid residues (3 His, 2 Lys, 2 Asp, and 1 Glu) and may provide phosphorylation sites, thereby regulating interaction of tandem SH2 domains and the ITIMs. Above all, the presence of unique spacing amino acid residues between the two ITIM tyrosyl residues of PZR is a distinct feature of the molecule.

By showing that catalytically inactive but not active forms of SHP-2 caused hyperphosphorylation of PZR in vivo and that PZR can be efficiently dephosphorylated by full-length but not SH2 domain-truncated SHP-2 in vitro, our study also suggests that PZR is a physiological substrate of SHP-2. Overexpression of catalytically inactive mutants of SHP-2 causes hyperphosphorylation of PZR through occupation of phosphorylation sites by the SH2 domain. However, overexpression of the native enzyme did not have such an effect (23). This suggests that the SH2 domain of SHP-2 is able to prevent dephosphorylation of PZR by other enzymes but not by itself. Therefore, tyrosine 247 and 263 serve as binding site for SH2 domains of SHP-2 but can also be dephosphorylated by its catalytic domain. Binding of SHP-2 to tyrosine-phosphorylated PZR brings SHP-2 to the plasma membrane and causes its activation. Activated SHP-2 in turn dephosphorylates certain proteins in the vicinity and thereby initiates signal transduction. On the other hand, SHP-2 can also dephosphorylate PZR and thereby terminate the signal transduction initiated by phosphorylation of PZR. Furthermore, the binding and dephosphorylation process that form a futile cycle driven by hydrolysis of ATP enables phosphorylation of PZR and the activity of SHP-2 to stay at relatively high basal levels and thus set a threshold for signal transduction to be initiated.

The ITIMs were defined as inhibitory motifs because they were initially found in inhibitory immunoreceptors like FcRIIB and KIR and they mediate the inhibitory effects of these proteins on signal transduction (13-16). Furthermore, since phosphorylation of the tyrosyl residue in the ITIMs triggers binding and activation of SH2 domain-containing phosphatases like the tyrosine phosphatases SHP-1 and SHP-2 and the inositol phosphatase SHIP (13-16), the inhibitory function of ITIMs is thought to be executed by these phosphatases. It should noted, however, that that not all dephosphorylation means down-regulation of signal transduction. On the contrary, in many cases, dephosphorylation results in initiation of signal transduction (1). For example, SHP-2 has been largely considered as a positive signal transducer (17, 18). In fact, the positive role of SHPS-1 in growth factor-induced MAP kinase activation is believed to be mediated by SHP-2 (41). After all, since the ITIMs are found in more and more diverse signaling molecules, their functions may also be diversified. PZR is unique in the way that it interacts specifically with SHP-2 but not with SHP-1 (4), while most other ITIM-containing proteins, including KIR, PECAM, PIR-B, SIRP/SHPS-1, and gp49, bind both SHP-1 and SHP-2 (42-48). Our study suggests that PZR is a binding protein as well as a physiological substrate of SHP-2. This may represent a general phenomenon and thus should help to define the function of other ITIM-containing proteins.

	FOOTNOTES

^* This work was supported by National Institutes of Health Grants HL-57393, CA75218 (to Z. J. Z), and CA-68485 (to Vanderbilt-Ingram Cancer Center).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

To whom correspondence and reprint requests should be addressed: 547 MRB II, 2220 Pierce Ave., Nashville, TN 37232-6305. Tel.: 615-936-1797; Fax: 615-936-3853; E-mail: joe.zhao@mcmail.vanderbilt.edu.

	ABBREVIATIONS

The abbreviations used are: SH2, Src homology 2; ITIM, immunoreceptor tyrosine-based inhibitory motif; ITAM, immunoreceptor tyrosine-based activation motifs; GST, glutathione S-transferase; BES, N,N-bis(2-hydroxyethyl)-2-aminoethanesulfonic acid.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES