Blocking the Function of Tyrosine Phosphatase SHP-2 by Targeting Its Src Homology 2 Domains*

SHP-2 is an Src homology 2 (SH2) domain-containing tyrosine phosphatase with crucial functions in cell signaling and major pathological implications. It stays inactive in the cytosol and is activated by binding through its SH2 domains to tyrosine-phosphorylated receptors on the cell surface. One such cell surface protein is PZR, which contains two tyrosine-based inhibition motifs responsible for binding of SHP-2. We have generated a glutathione S-transferase fusion protein carrying the tandem tyrosine-based inhibition motifs of PZR, and the protein was tyrosine-phosphorylated by co-expressing c-Src in Escherichia coli cells. The purified phosphoprotein displays a strong binding to SHP-2 and causes its activation in vitro. However, when introduced into NIH 3T3 cells by using a protein delivery reagent, it effectively inhibited the activation of ERK1/2 induced by growth factors and serum but not by phorbol ester, in reminiscence of the effects caused by expression of dominant negative SHP-2 mutants and deletion of functional SHP-2. The data suggest that the exogenously introduced PZR protein specifically binds SHP-2, blocks its translocation, and renders it functionally incompetent. This is further supported by the fact that the phosphorylated PZR protein had no inhibitory effects on fibroblasts derived from mice expressing only a mutant SHP-2 protein lacking most of the N-terminal SH2 domain. This study thus provides a novel and highly specific method to interrupt the function of SHP-2 in cells.

SHP-2 is a widely distributed intracellular tyrosine phosphatase that contains two SH2 1 domains (1-3). It has a crucial role in cell signaling. Earlier studies by overexpressing the catalytically inactive cysteine to serine mutant of SHP-2 in cell lines demonstrated that the enzyme plays a positive role in activation of ERK1/2 induced by growth factors (1)(2)(3). In Xenopus, a dominant negative mutant of SHP-2 blocks fibroblast growth factorand activin-mediated induction of mesoderm (4). In mice, disruption of the Shp-2 gene caused death of mouse embryos at midgestation (5). Further studies with cells derived from SHP-2deficient mice demonstrated impairment in erythropoiesis and cell migration (6 -8). SHP-2 also has major pathological implications. Mutation of Shp-2 causes Noonan syndrome (9), an autosomal dominant disorder characterized by dysmorphic facial features, proportionate short stature, and heart disease. Excessive SHP-2 activity caused by mutation in the N-terminal SH2 domain is considered to be responsible for the pathogenesis. Furthermore, SHP-2 is an intracellular target of the CagA protein in Helicobacter pylori (10). H. pylori are associated with severe gastritis and gastric cancers, and CagA is introduced from the attached H. pylori into host cells and undergoes tyrosine phosphorylation and thereby activates SHP-2.
Considering the important role of SHP-2 in cell signaling and its pathological implications, inhibition of the enzyme has great research and therapeutic values. However, a specific inhibitor of SHP-2 is still lacking. The catalytic domains of all classic protein-tyrosine phosphatases (PTPs) share extremely high similarity in ternary structures. In fact, the catalytic domains of SHP-2 and its closest homologue SHP-1 share 60% sequence identity, but the ternary structures of the two catalytic domains are almost superimposible (11)(12)(13). This makes it difficult to find highly selective inhibitors targeting the catalytic domains. On the other hand, PTPs are known to have highly specific functions, and this specificity is at least partly conferred by the non-catalytic domains or segments beyond the catalytic domains (3). SHP-1 and SHP-2 both contain SH2 domains that play important regulatory and targeting roles. In vitro, purified SHP-1 and SHP-2 display very low specific activity (14,15). This low level of activity is because of an internal suppression because the enzymes are significantly activated by removal of either C-terminal segments or SH2 domains (14,15), by interaction with anionic phospholipids (15)(16)(17), and by binding of SH2 domains to specific phosphopeptides (18,19). Furthermore, x-ray crystal structures of SHP-1 and SHP-2 reveal blocking of the catalytic center by the N-terminal SH2 domains (12,13). In vivo, SHP-1 and SHP-2 reside mainly in the cytosol where they are presumably inactive. It is generally believed that SHP-1 and SHP-2 are activated by binding to tyrosine-phosphorylated receptors and other proteins concomitant with translocation of the enzymes to the plasma membrane (1)(2)(3). Therefore, activation of SHP-1 and SHP-2 is controlled by membrane or membrane-associated proteins that recruit them. We reasoned that physiological functions of SHP-1 and SHP-2 could be efficiently blocked by exogenously introduced tyrosine-phosphorylated peptides or proteins that bind the enzymes and keep them in the cytosol or other cellular compartments where the enzymes are not functional.
SHP-1 and SHP-2 have been known to be recruited to many receptors. Recently, we have isolated a cell surface glycoprotein designated PZR that specifically recruits SHP-2 through its immunoreceptor tyrosine-based inhibition motifs (ITIMs) (20 -22). PZR are efficiently phosphorylated by the Src family tyrosine kinases in vitro and in vivo (23). We have prepared a glutathione S-transferase (GST) fusion protein carrying tyrosine-phosphorylated ITIMs of PZR from Escherichia coli cells coexpressing active c-Src. In this study, we introduced the GST fusion protein into cells by using the BioPorter reagent. This effectively inhibited growth factor-and serum-induced activation of ERK1/2 by blocking SHP-2 function. Our study thus provides a novel and highly specific method to interrupt the function of SHP-2 in intact cells.

EXPERIMENTAL PROCEDURES
Materials-NIH 3T3 cells and fibroblasts cells from wild type (SHP-2 ϩ/ϩ ) and functional SHP-2-deficient (Shp-2⌬ 46 -110 ) mice were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum and 50 units/ml each of penicillin and streptomycin. Human embryonic kidney 293 cells overexpressing SHP-2 were maintained in the same medium plus 0.25 mg/ml G418 sulfate (20). Polyclonal anti-SHP-2, monoclonal anti-SHP-2, polyclonal anti-pERK(Thr 202 /Tyr 204 ), monclonal anti-GST, and monoclonal anti-phosphotyrosine 4G10 were purchased from Santa Cruz Biotechnology, Signal Transduction Laboratories, Cell Signaling Technology, Sigma, and Upstate Biotechnology Inc., respectively. SHP-1, SHP-2, and ⌬SH2-SHP-2 were purified as previously described (14,15). Fig. 1A shows schematic structures of PZR and two recombinant fusion proteins. GST-⌬PZR represents a GST fusion protein carrying the intracellular domain of PZR corresponding to amino acid residues 192-269 of the molecule. The recombinant protein was expressed in E. coli cells and purified by using glutathione-Sepharose beads. GST-p⌬PZR is a tyrosine-phosphorylated form of GST-⌬PZR. The phosphorylation occurred only at the ITIMs of PZR and was achieved by co-expressing c-Src carried by a pET9a vector in E. coli cells as previously described (23). The protein was also purified from glutathione-Sepharose columns.

Expression of GST Fusion Proteins Carrying the Intracellular Domain of PZR-
SHP-2 Binding and Phosphatase Activity Assays-GST-p⌬PZR, GST-PZR, and GST-p⌬PZR immobilized on glutathione-Sepharose beads were incubated with extracts of 293 cells overexpressing SHP-2. The cell extracts were made in Buffer A containing 25 mM ␤-glycerolphosphate (pH 7.3), 0.1 M NaCl, 1% Triton X-100, 5 mM EDTA, 0.5 mM Na 3 VO 4 , 10 mM ␤-mercaptoethanol, and a protease inhibitor mixture (Roche Diagnostics). After washing three times with Buffer A, proteins bound to the beads were analyzed by SDS gels followed by immunoblotting or Coomassie Blue staining. To investigate the specific activation of SHP-2 by GST-p⌬PZR, PTP activities of SHP-1, SHP-2, and ⌬SH2-SHP-2 were measured with 2 mM p-nitrophenyl phosphate as substrate in the presence of GST-⌬PZR or GST-p⌬PZR. The assays were carried out at room temperature, and the assay system contained 25 mM HEPES-NaOH (pH 7.0), 1.0 mM EDTA, 0.5 mg/ml bovine serum albumin, and 1.0 mM dithiothreitol.
BioPorter-mediated Transfer of Proteins into Cells-We employed the BioPorter reagent (Gene Therapy Systems) to transfer GST-⌬PZR and GST-p⌬PZR into NIH 3T3 cells following the protocol provided by the manufacturer. Briefly, 5 l of BioPorter reagent was air-dried under a tissue culture hood for 2-3 h and then resuspended in 50 l of phosphate-buffered saline containing 3-6 g of GST-⌬PZR or GST-p⌬PZR. After 5 min incubation, the suspensions were mixed with 250 l of plain Dulbecco's modified Eagle's medium and were then added to NIH 3T3 cells (ϳ80 -90% confluency) cultured in 12-well cell culture plates. Prior to the addition of the protein-BioPorter complex, the cells were washed twice with phosphate-buffered saline. After addition of the complex, cells were further incubated at 37°C in 5% CO 2 culture incubator for 4 -6 h before further treatment.
Cell Stimulation, Extraction, and Western Blotting Analyses-After 3-6 h of incubation with GST-⌬PZR and GST-p⌬PZR carried by the BioPorter reagent, cells were treated with fetal bovine serum (FBS), epidermal growth factor, platelet-derived growth factor (PDGF), insulin, and phorbol 12-myristate 13-acetate (PMA) for various periods of time, and the stimulation was stopped by washing with ice-cold phosphatebuffered saline. The cells were then extracted in the Buffer A. Cell lysates were cleared by centrifugation, and clear cell extracts were then separated by 10% SDS gels for Western blot analyses with specified antibodies. Detection was made by enhanced chemiluminescence, and quantification of the gel bands was performed by using a gel scanner.

GST-p⌬PZR Specifically Binds and Activates SHP-2 in
Vitro-In our previous study (23), we demonstrated that Src family protein-tyrosine kinases are at least partly responsible for tyrosine phosphorylation of PZR in cells, and we further showed that co-expression of GST-⌬PZR with c-Src in E. coli cells causes strong phosphorylation of GST-⌬PZR. The phosphorylation occurs on the tyrosyl residues of PZR embedded in ITIMs because mutation of the residues diminished the phosphorylation (23). By using glutathione-Sepharose columns, we were able to purify a large amount of tyrosine-phosphorylated GST-p⌬PZR from E. coli cell extracts. To analyze the ability of GST-p⌬PZR to interact with SHP-2, we incubated glutathione-Sepharose beads carrying GST-p⌬PZR with cell extracts obtained from 293 cells overexpressing SHP-2 (20). As shown in Fig. 1B, Western blotting analysis with anti-SHP-2 antibodies revealed a strong binding of SHP-2 with GST-p⌬PZR but minimal binding with non-phosphorylated GST-⌬PZR. GST-p⌬PZR-associated SHP-2 was also clearly detected by Coomassie Blue staining. In a sense, GST-p⌬PZR provides a very effective method to enrich SHP-2 from cell extracts. Considering the fact that the GST-p⌬PZR protein sample may not be 100% phosphorylated, the data suggest a near stoichiometric association of GST-p⌬PZR with SHP-2. When similar experiments were conducted with cell extracts from 293 cells overexpressing SHP-1, a much smaller level of SHP-1 was detected by Western blotting, and no binding was seen in Coomassie Blue staining (data not shown). This suggests a much preferable interaction of GST-p⌬PZR with SHP-2 over SHP-1. These data agree with our previous results obtained with immunoprecipitation (20 -22). The specific interaction of GST-p⌬PZR with SHP-2 is also demonstrated by specific activation of SHP-2 but not SHP-1 or ⌬SH2-SHP-2 by GST-p⌬PZR. As shown in Fig. 2, the presence of GST-p⌬PZR in phosphatase assay reactions resulted in near 20-fold activation of SHP-2, but non-phosphorylated GST-⌬PZR had no stimulatory effect. In contrast, GST-p⌬PZR only caused marginal activation of SHP-1 and had no effects on ⌬SH2-SHP-2, an SH2 domain-truncated form of SHP-2. These results further verify that PZR specifically interacts with the SH2 domains of SHP-2. Because GST-p⌬PZR activate SHP-2 by binding to its SH2 domains, SHP-2 activation can be considered as a measure of the interaction between SH2 domains and GST-p⌬PZR.
The BioPorter Protein Delivery Reagent Efficiently Transfers GST-p⌬PZR into Cultured Cells and Initiates its Interaction with SHP-2-To deliver GST-p⌬PZR into live cells, we employed the BioPorter protein delivery reagent developed by Gene Therapy Systems, Inc. The BioPorter protein delivery reagent is a cationic lipid mixture that when complexed with  -2 (B). A, Y and pY represent non-phosphorylated and phosphorylated tyrosyl residues embedded in the intracellular ITIMs of PZR, respectively. B, glutathione beads carrying GST-⌬PZR or GST-p⌬PZR were incubated with cell extracts from 293 cells overexpressing SHP-2. The beads were washed 3 times with Buffer A, and proteins bound to the beads were resolved on SDS gels for Coomassie Blue staining or for transferring to polyvinylidene difluoride membranes followed by blotting with anti-SHP-2 antibody.
proteins and peptides allows for direct intracellular delivery. The complex formed is non-covalent and it therefore does not interfere with the biological activity of the protein. This technique is rapid and relatively uncomplicated. With the Bio-Porter reagent, we were able to effectively deliver GST-⌬PZR and GST-p⌬PZR into NIH 3T3 cells. As shown in Fig. 3, both tyrosine-phosphorylated GST-p⌬PZR and non-phosphorylated GST-⌬PZR were recovered in the cell extracts when the proteins complexed with the BioPorter reagent were incubated with the cells. In contrast, no incorporation of GST-p⌬PZR into the cells was observed in the absence of the BioPorter reagent. Furthermore, transferred GST-p⌬PZR was able effectively interact with SHP-2 as indicated by the "pull down" assays with glutathione-Sepharose beads. Therefore, the BioPorter reagent provides an effective way to deliver functional GST-p⌬PZR into 3T3 cells to target SHP-2. In fact, Western blotting analyses of cell extracts and culture medium revealed that 15-20% of the GST fusion proteins entered the cells, and immunofluorescent cell staining showed nearly 100% cells were uploaded with the exogenously introduced fusion proteins in the entire cytoplasma (data not shown). Note that the efficient transfer mediated by the BioPorter reagent was obtained in serum-free medium, which is also used as the serum starvation step for stimulation of cells.
GST-p⌬PZR Efficiently Blocks Activation of ERK1/2 Induced by Serum and Growth Factors-SHP-2 has been defined as a positive signal transducer in growth factor signaling. To find out whether introduction GST-p⌬PZR into cells block normal functions of SHP-2, we investigated activation of ERK1/2. NIH 3T3 cells were incubated with mixtures of the BioPorter reagent and GST-⌬PZR or GST-p⌬PZR in serum-free medium. After 3 and 6 h of incubation, the cells were treated with 15% FBS. Western blotting analyses of the cell extracts are shown Fig. 4. A 3-h incubation was sufficient to transfer the GST fusion proteins into the cells. A longer incubation period (6 h) did not significantly change the transfer efficiency or the stability of transferred proteins, but a higher concentration of GST-⌬PZR and GST-p⌬PZR clearly favored the transfer. We employed a phospho-specific anti-ERK1/2 antibody to determine the activation of ERK1/2. The antibody detects ERK1 and ERK2 (p42 and p44 MAP kinase) only when they phosphorylated at the threonyl and tyrosyl residues that are required for activation of the enzymes. Upon stimulation with FBS, the control cells and the cells treated with non-phosphorylated GST-⌬PZR showed a marked increase in the phosphorylation of ERK1/2 over the basal level. In contrast, the cells treated with GST-p⌬PZR at both lower and higher concentrations ex-hibited essentially no response. Mobility shifts of ERK1/2 detected by a regular anti-ERK1/2 antibody further verified the data. As shown in the bottom panel of Fig. 4, a clear up-shift of the ERK2 band was seen with control and GST-⌬PZR cells but not with the cell treated with GST-p⌬PZR. The mobility shift is also an indication of ERK1/2 phosphorylation. Therefore, the data indicate that introduction of GST-p⌬PZR into cells significantly inhibited FBS-induced activation of ERK1/2. We further analyzed the effects of GST-p⌬PZR on activation of ERK1/2 induced by growth factors and the phorbol ester PMA. As shown in Fig. 5 (top panel), whereas GST-p⌬PZR significantly reduced the activation of ERK1/2 induced by PDGF, epidermal growth factor, and insulin, it had no effect on that induced by PMA. This is reminiscent of the data obtained with cells overexpressing of the dominant negative mutant form of SHP-2 and cells derived from Shp-2⌬ 46 -110 mice (20,24,25). A quantitative representation of the effects of GST-p⌬PZR on the activation of ERK2 is shown in Fig. 6. The delivery of GST-p⌬PZR into NIH 3T3 cells reduced FBS-and growth factor-FIG. 2. Specific activation of SHP-2 by GST-p⌬PZR. Activities of SHP-1, SHP-2, and ⌬SH2-SHP-2 were analyzed with 2 mM para-nitrophenyl phosphate as substrate in the presence of 80 g/ml GST-⌬PZR or GST-p⌬PZR. The assay buffer contained 25 mM HEPES-NaOH (pH 7.0), 1 mM dithiothreitol, 0.5 mg/ml bovine serum albumin, and 1 mM EDTA. The final concentrations of enzymes were 20 g/ml SHP-1, 40 g/ml SHP-2, and 4 g/ml ⌬SH2-SHP-2. Control experiments were conducted without addition of GST-⌬PZR or GST-p⌬PZR. Data represent relative activity.

FIG. 3. BioPorter-mediated transfer of GST-⌬PZR and GST-p⌬PZR into NIH 3T3 cells.
GST-⌬PZR and GST-p⌬PZR proteins were added to cultured NIH 3T3 cells with or without preincubation with the BioPorter reagent. After 6 h of incubation, cells were washed three times with ice-cold phosphate-buffered saline and extracted in Buffer A. The cell extracts were then incubated with glutathione-Sepharose beads to pull down GST-⌬PZR and GST-p⌬PZR. Proteins bound to the beads were subjected to Western blotting analyses with antibodies against GST (for GST-⌬PZR and GST-p⌬PZR), phosphotyrosine, and SHP-2 as indicated.

FIG. 4. Inhibition of serum-induced activation of ERK1/2 by
GST-p⌬PZR. NIH 3T3 cells were incubated with 0 (for control), 3, or 6 g of GST-⌬PZR or GST-p⌬PZR premixed with the BioPorter reagent. After 3 or 6 h of incubation, cells were left unstimulated or stimulated with 15% FBS for 15 min followed by extraction in Buffer A. Samples with equal amounts of total proteins were subjected to Western blotting analyses with antibodies against phosphotyrosine, GST, phospho-ERK, and ERK as indicated.
induced ERK2 phosphorylation by 70 -90% but had no effect on that induced by PMA. Taken together, the data indicate that introduction of GST-p⌬PZR specifically inhibited growth factor-and serum-induced activation of ERK1/2 in a way similar to the expression of dominant negative SHP-2 or deletion of functional SHP-2.
Delivery of GST-p⌬PZR into NIH 3T3 Cells Causes Hyperphosphorylation of a 90-kDa Protein-By specifically targeting SHP-2, GST-p⌬PZR may affect tyrosine phosphorylation of specific cellular proteins. We performed Western blotting analyses of cell extracts with an anti-phosphotyrosine antibody to find out whether this is indeed the case. When the whole cell extracts of NIH 3T3 cells were separated on SDS gels, the PDGF receptor running at ϳ180 kDa was shown as a major tyrosine-phosphorylated band, and its tyrosine phosphorylation was markedly enhanced upon stimulation of cells with PDGF (Fig. 5). However, introduction of neither GST-⌬PZR nor GST-p⌬PZR in the cells had significant effects on its tyrosine phosphorylation at the basal level or under PGDF stimulation. Immunoprecipitation of the PDGF receptor followed by antiphosphotyrosine immunoblotting analyses further verified the results with the PDGF receptor, and similar data were obtained with the epidermal growth factor receptor and the insulin receptor (data not shown). This suggests that GST-p⌬PZR does not affect growth factor signaling at the receptor level. Interestingly, treatment of cells with GST-p⌬PZR but not with GST-⌬PZR resulted in appearance of a broad tyrosinephosphorylated protein band around 90 kDa (see Fig. 5). Tyrosine phosphorylation of this protein was not altered in response to stimulation with growth factors. This protein likely corresponds to the tyrosine-phosphorylated protein of similar molecular size observed in fibroblast cells derived from the Shp-2⌬ 46 -110 mice (26). This also suggests that GST-p⌬PZR effectively blocks growth factor-induced activation of ERK1/2 by blocking the function of SHP-2.  Fig. 5. The intensity of pERK2 bands was quantified by gel scanning and normalized against total proteins in cell extracts. Data represent relative intensity (mean Ϯ S.D., n Ն 3). The level of pERK2 at the basal level in control cells was defined as 1. EGF, epidermal growth factor. truncated enzyme (5). Fibroblast cells derived from wild type mice were used as control. As obtained with NIH 3T3 cells, the BioPorter reagent effectively delivered GST-p⌬PZR into both SHP-2 ϩ/ϩ and SHP-2⌬ 46 -110 cells. However, while wild type SHP-2 was found to be associated with GST-p⌬PZR, N-terminal SH2 domain-truncated SHP-2⌬ 46 -110 was not, suggesting that the remaining C-terminal SH2 domain in SHP-2⌬ 46 -110 is not sufficient to mediate efficient interactions with ITIMs of PZR (Fig. 7A). This is further supported by the fact that pervanadate failed to induce associations of SHP-2⌬ 46 -110 with endogenous PZR in contrast to the robust associations seen with native SHP-2 (Fig. 7B). We further treated the GST-p⌬PZR-loaded cells with FBS and PDGF. As shown in Fig. 7C, SHP-2 ϩ/ϩ cells displayed strong activation of ERK1/2, but the activation was significantly inhibited by treatment of cells with GST-p⌬PZR. In contrast, SHP-2⌬ 46 -110 cells exhibited a much weaker response to FBS and PDGF, but GST-p⌬PZR had no inhibitory effects. These data provide strong evidence that GST-p⌬PZR inhibits serum-and growth factor-induced activation of ERK1/2 by blocking the normal function of SHP-2.

DISCUSSION
In this study, we have demonstrated that a GST fusion protein carrying tyrosine-phosphorylated ITIMs of PZR effec-tively inhibited serum-and growth factor-induced activation of ERK1/2. We believe that the inhibitory effects of GST-p⌬PZR are mediated by blocking the normal function of tyrosine phosphatase SHP-2 in the cells. This is supported by the following facts. First, the inhibitory role is dependent on tyrosine phosphorylation of PZR, and tyrosine-phosphorylated ITIMs of PZR specifically interact with the SH2 domain of SHP-2 in vitro. Second, the inhibitory effects are specific for growth factor but not for phorbol ester PMA. This agrees with the expression of the dominant negative mutant of SHP-2 and targeted disruption of SHP-2 in mice (24 -32). Third, GST-p⌬PZR had no effect on functional SHP-2-deficient fibroblast cells derived from Shp-2⌬ 46 -110 mice. Therefore, this study provides a novel method and an effective reagent to block normal function of SHP-2 in intact cells. It should serve as a valuable tool for studying the biological function of SHP-2 and for developing therapeutic drugs targeted against SHP-2.
We have turned a powerful in vitro activator of SHP-2 into an effective blocker of the enzyme in vivo. This is because of the structural and regulatory nature of the enzyme. The model shown in Fig. 8 provides a schematic explanation. The activity of SHP-2 is tightly regulated. During the resting state of cells, SHP-2 stays mostly in the cytosol where it remains inactive because of its internal suppressive structure. Upon stimulation, specific tyrosine-phosphorylated anchor proteins recruit SHP-2 through the SH2 domains and results in translocation and activation of the enzyme. Translocation is the key in regulation of the enzyme. In the cytosol or other cellular compartments where SHP-2 substrates do not exist, SHP-2 should be non-functional even if it is in an enzymatically active state. Therefore, a peptide or protein that binds the SH2 domains of SHP-2 and thereby blocks its translocation to desired locations should specifically inhibit the function of SHP-2. This notion is fully supported by our current experimental data. This view also explains the loss of SHP-2 function in SHP-2⌬ 46 -110 mice in which exon 3 of the Shp-2 gene was deleted. Targeted dele- After extensive washing with Buffer A, proteins bound to the beads together with crude cell extracts were subjected to Western blotting analyses with anti-SHP-2 and anti-GST as indicated. B, pervanadatetreated and non-treated cells were extracted in Buffer A. The cell extracts were then immunoprecipitated with rabbit polyclonal anti-SHP-2, and the immunoprecipitates were subjected to Western blotting analyses with anti-PZR and monoclonal anti-SHP-2 antibodies. C, GST-p⌬PZR-loaded cells were stimulated with 15% FBS for 15 min or with 10 ng/ml PDGF for 5 min and then extracted in Buffer A. Cell extracts containing equal amounts of total proteins were analyzed by Western blotting with anti-phospho-ERK and anti-ERK antibodies.
FIG. 8. A model to explain how GST-p⌬PZR blocks the function of SHP-2. SHP-2 stays in an inactive state in the cytosol. It is activated by binding through its SH2 domains to tyrosine-phosphorylated proteins on the plasma membrane. Activated SHP-2 dephosphorylates membrane or membrane-associated substrates and thereby facilitates signaling transduction. By interacting with the SH2 domains of SHP-2, GST-p⌬PZR prevents the enzyme from translocating to the plasma membrane and thereby blocks its function. tion of exon 3 gave rise to an N-terminal SH2 domain-truncated enzyme (5). The truncated enzyme is expressed in mouse cells and is presumably more active. However, it is apparently nonfunctional in vivo because of its inability to be recruited to desired locations (5)(6)(7)(8). A recent study with a total knockout of SHP-2 showed similar embryonic lethality and thereby excluded any possible gain-of-function of the truncated SHP-2 (discussed in Ref. 3). Both introduction of GST-p⌬PZR in this study and the deletion of SHP-2 exon 3 in the earlier studies resulted in an enzymatically active but non-transferable enzyme, and the consequence is loss of SHP-2 function. It should be pointed out that because the affinity of SHP-2 to substrate is largely conferred by target domains and its catalytic domain possesses substrate specificity, this catalytically active but non-transferable enzyme per se should not cause nonspecific effects. On the other hand, if an exogenously introduced protein is localized to the plasma membrane and has the ability to recruit SHP-2, it should functionally activate SHP-2. This is probably the case with the CagA protein of H. pylori. CagA is localized to the plasma membrane of host cells and is constitutively phosphorylated on tyrosyl residues. Tyrosine-phosphorylated CagA recruits SHP-2 through SH2 domain binding to the plasma membrane and cause its functional activation. This is believed to be responsible for gastric tumors caused by bacteria. Recent studies have also linked mutations of Shp-2 with ϳ50% of Noonan syndrome (9,33,34). Many of these mutations involve residues that participate in the interaction of the N-SH2 and PTP domains that cause autoinhibition of PTP activity. It is generally thought that activation of SHP-2 is responsible for the pathogenesis. Based on current data, the activation of SHP-2 per se may not be important. Instead, the enhanced ability of SHP-2 to bind tyrosine-phosphorylated peptide motifs and consequent translocation may be crucial. This explains the pathogenic effectiveness of some mutations of SHP-2 in Noonan syndrome that are not related to the N-SH2 and PTP domain interaction responsible for autoinhibition of the enzyme (33,34).
It is apparent that inhibition of SHP-1 and SHP-2 function has important research and therapeutic values. Because the specific functions of SHP-1 and SHP-2 are defined by both catalytic and SH2 domains, inhibitors can target the SH2 domains as well as the catalytic domains. Determination of the crystal structures revealed very high structural similarity between SHP-1 and SHP-2, especially in the catalytic and N-SH2 domains (12,13). Both SHP-1 and SHP-2 use the N-SH2 domain to keep the enzymes in the inactive conformation. However, the C-SH2 domain in the SHP-1 structure is differently orientated from that in SHP-2, and its secondary structure elements are relatively more openly organized (13). Therefore, the C-SH2 domains may play a major role in determining the specificity of SHP-1 and SHP-2, and therefore, targeting the C-SH2 domain may be more likely to produce specific inhibitors. Main methodology of drug development in industry has centered on the screening of vast libraries of molecules for biological activity upon which variants are produced to maximize beneficial medical characteristics. Recent developments in the understanding of the biochemistry of intracellular processes and protein-protein interaction have given rise to an alternative method by using peptides and proteins targeted at specific protein domains. New methods of noninvasive delivery of functional peptides and proteins to cells have made this feasible (35)(36)(37). In this study, we employed the BioPorter reagent to deliver GST-⌬PZR into NIH 3T3 cells. This method worked very well for NIH 3T3 cells when the transfer is conducted in serum-free conditions. We believe other delivery methods including attachment of cell-permeable peptide se-quences can also be employed. We employed a tyrosine-phosphorylated GST fusion protein in our current study. In principle, synthetic diphosphotyrosyl peptides derived from the ITIM sequences should be equally effective. In addition, the strategy for functional inhibition of SHP-2 can be applied to SHP-1. SHP-1 and SHP-2 share over 50% sequence identity but their functions are often opposite. Our earlier studies demonstrated that SHP-1 preferentially binds LAIR-1, and the specificity is conferred by the interactions between the dual ITIMs and the dual SH2 domains (38). Therefore, tyrosine-phosphorylated ITIMs of LAIR-1 should serve as a selective blocker of SHP-1.