Noonan Syndrome-associated SHP-2/Ptpn11 Mutants Enhance SIRPα and PZR Tyrosyl Phosphorylation and Promote Adhesion-mediated ERK Activation*

Noonan syndrome (NS) is an autosomal dominant disorder that is associated with multiple developmental abnormalities. Activated mutations of the protein-tyrosine phosphatase, SHP-2/PTPN11, have been reported in ∼50% of NS cases. Despite being activated, NS-associated SHP-2 mutants require plasma membrane proximity to evoke disease-associated signaling. Here we show that NS-associated SHP-2 mutants induce hypertyrosyl phosphorylation of the transmembrane glycoproteins, SIRPα (signal-regulatory protein α) and PZR (protein zero-related), resulting in their increased association with NS-associated SHP-2 mutants. NS-associated SHP-2 mutants enhanced SIRPα and PZR tyrosyl phosphorylation either by impairing SIRPα dephosphorylation or by promoting PZR tyrosyl phosphorylation. Importantly, during embryogenesis in a mouse model of NS, SIRPα and PZR were hypertyrosyl-phosphorylated and bound increased levels of the NS-associated SHP-2 mutant. SIRPα and PZR have been implicated in extracellular matrix-dependent signaling. Mouse embryonic fibroblasts derived from a mouse model of NS displayed enhanced ERK activation in response to fibronectin plating. Knockdown of SIRPα and PZR in these cells attenuated the enhanced activation of ERK following fibronectin plating. Thus, SIRPα and PZR serve as scaffolds that facilitate plasma membrane recruitment and signaling of NS-associated SHP-2 mutants.

The Src homology 2 (SH2) 2 domain-containing protein-tyrosine phosphatase (PTP), SHP-2, is a ubiquitously expressed cytoplasmic PTP with two NH 2 -terminal SH2 domains and a COOH-terminal PTP domain (1,2). SHP-2, in most cases, is a positive transducer of growth factor, cytokine, integrin, and hormone signaling pathways regulating processes such as cell proliferation, differentiation, adhesion, migration, and apoptosis (1,2). In virtually all cases, the catalytic activity of SHP-2 is required for its positive signaling effects that are mediated through the dephosphorylation of substrates that are negatively regulated by tyrosine phosphorylation (1)(2)(3). Either mutation or disruption of SHP-2 results in embryonic lethality in mice (4,5). During embryogenesis, SHP-2 is required for mammalian limb development (6) and heart valvulogenesis (7). Hence, SHP-2 plays an essential role in mammalian development.
SHP-2 is the protein product of the PTPN11 gene that is mutated in ϳ50% of Noonan syndrome (NS) cases (8,9). NS is an autosomal dominant disorder that is estimated to occur in 1:1,000 to 1:2,500 live births in the United States and worldwide (9 -11). NS patients are characterized by proportionate short stature, facial dysmorphisms, mental retardation, bleeding diathesis, and cardiovascular defects such as pulmonary stenosis (9 -11). In addition, NS patients display hematologic abnormalities, including myeloid disorders and juvenile myelomonocytic leukemia (9). Most of the disease-associated mutations of SHP-2 have been mapped to the NH 2 -SH2 domain, whereas other mutations map to either the COOH-SH2 or the PTP domain (8,12,13). SHP-2 is maintained in an "inactive" autoinhibited conformation that is maintained by interactions between the NH 2 -SH2 domain and the PTP domain (14). Mutations within the SH2-PTP domain interface disrupt intramolecular interactions between these two domains to relieve the autoinhibited conformation resulting in constitutive SHP-2 activation (15). Several NS/leukemia-associated SHP-2 mutants are activating, hence conferring gain-of-function properties to SHP-2 (8, 12, 16 -20). These observations support the notion that enhanced SHP-2 catalysis is causative to NS and leukemia.
It is now established that in response to a variety of growth factors and cytokines SHP-2 is required for the activation of Src (21)(22)(23), Ras/ERK (5, 24 -26), and phosphatidylinositol 3Ј-kinase/Akt pathways (27)(28)(29). In addition, SHP-2 is also responsible for mediating the activation of the Src and Ras/ERK pathways following integrin engagement (21,30,31). In response to growth factors and cytokines, the gain-of-function properties of the NS/leukemia-associated SHP-2 mutants enhance ERK activation (17)(18)(19)(32)(33)(34). However, in some cases the NS/leukemia-associated SHP-2 mutants fail to enhance ERK activation (18,35,36). Such differences might reflect cell type-specific and possibly stimulation-dependent variations in which the NS/leukemia-associated SHP-2 mutants operate. Activated and NS-associated SHP-2 mutants have also been shown to increase the calcium oscillatory frequency in fibroblasts and cardiomyocytes, and this correlates with the inhibition of the nuclear factor of activated T-cells in cardiomyocytes (37). In addition, the leukemia-associated mutants of SHP-2 also exhibit enhanced Akt activity (19,33).
Despite a growing number of signaling pathways that are subject to altered regulation by NS/leukemia-associated SHP-2 mutants, the mechanism through which these activated SHP-2 mutations enhance downstream signaling remains to be fully defined. One mechanism through which activated SHP-2 mutants might enhance signaling is by increasing the levels of tyrosyl phosphorylation of their upstream interacting proteins and hence their association and recruitment to locales where signaling is initiated. We first showed that an activated SHP-2 mutant increases tyrosyl phosphorylation of the fibroblast growth factor receptor substrate-2 in response to FGF-2 (38). Similar observations with activated SHP-2 mutants were also found to occur with the Grb2-associated binder-1 (Gab-1), which becomes hypertyrosyl-phosphorylated in response EGF (17). As predicted, enhanced tyrosyl phosphorylation of SHP-2-binding proteins results in their increased association with SHP-2 (17,18,33). These observations suggest that binding and thus localization of the NS/leukemia-associated SHP-2 mutants play an important role in disease-mediated signaling. This notion is supported by the observation that disruption of Gab-2 binding by mutation of the critical arginine within the SH2 domain of the leukemogenic E76K SHP-2 mutant abrogates cellular transformation (19).
In this study, we show that NS/leukemia-associated SHP-2 mutants interact with, and induce hypertyrosyl phosphorylation of, the transmembrane glycoproteins SIRP␣ (signal regulatory protein ␣) and PZR (protein zero-related). We show that in mice harboring an NS-associated SHP-2 mutant, both SIRP␣ and PZR are hypertyrosyl-phosphorylated. Fibroblasts derived from these mice also exhibited increased SIRP␣ and PZR tyrosyl phosphorylation. The enhanced ERK activation in response to adhesion was predominantly mediated by SIRP␣, with PZR playing less of a role. These results raise the possibility that SIRP␣ and PZR serve as scaffold proteins for NS/leukemiaassociated mutants and may thus play a role in the pathogenesis of these diseases.
Adenoviral Infections-Replication-defective recombinant WT and activated E76A SHP-2 adenoviruses have been described previously (38). Adenoviruses were amplified in 293 cells and purified according to the manufacturer's instructions (Quantum-Appligene).
Transient Transfections-C2C12 cells were transfected with the indicated plasmids using Lipofectamine 2000 (Invitrogen), and SYF cells were transfected using FuGENE transfection reagent according to the manufacturer's instructions (Roche Applied Science). For siRNA transfections, 100,000 MEFs derived from the D61G NS knock-in mice (18) were plated onto 60-mm tissue culture plates 1 day prior to transfection and were transfected with either 50 nM SIRP␣ siRNA (Santa Cruz Biotechnology), 25 nM PZR siRNA (Dharmacon), or control nontargeting siRNA (Dharmacon and/or Santa Cruz Biotechnology) using Lipofectamine RNAiMAX reagent (Invitrogen) according to the manufacturer's instructions. 24 h after transfection cells were serum-deprived (0.1% FBS/DMEM) overnight, and the cells were trypsinized for re-plating onto fibronectin-coated Petri dishes.
For immunoblotting, lysates or immune complexes were resolved by SDS-PAGE and transferred onto Immobilon-P membranes (Millipore). All primary antibodies were used by first blocking membranes with 5% nonfat dry milk or 5% BSA in Tris-buffered saline/Tween 20 (TBS-T) for 1 h at room temperature or overnight at 4°C. Primary antibodies were diluted in 2.5% nonfat dried milk in TBS-T or 2.5% BSA in TBS-T. Mouse monoclonal SHP-2 antibody (BD Transduction Laboratories) was used at a 1:1,000 dilution. Anti-PZR antibody (105 (6)) was used at a 1:5,000 dilution. SIRP␣ polyclonal antibody was used at a 1:5,000 dilution; monoclonal anti-phosphotyrosine (4G10) antibody was used at a 1:50 dilution; rabbit polyclonal phospho-p44/42 MAPK antibody (Cell Signaling Technology) was used at a 1:1,000 dilution; and rabbit polyclonal ERK1/2 (Santa Cruz Biotechnology) antibody was used at a 1:4,000 dilution. Monoclonal c-Src (GD11, Upstate Biotechnology, Inc.) antibody was used at a 1:1,000 dilution, and polyclonal phospho-Src Tyr-416 antibody (Cell Signaling Technology) was used at a 1:500 dilution. Primary antibodies were either used at room temperature for 2 h or overnight at 4°C. After primary antibody incubations, membranes were washed in TBS-T and incubated in secondary anti-rabbit Ig horseradish peroxidaselinked donkey antibody or anti-mouse Ig horseradish peroxidase-linked sheep antibody (GE Healthcare) for 1 h at room temperature. After secondary antibody incubations, membranes were washed in TBS-T and visualized by using enhanced chemiluminescence reagents (Amersham Biosciences).
Immune Complex Phosphatase Assays-C2C12 cells transfected with the indicated expression plasmids were lysed in 1% Nonidet P-40 lysis buffer 24 h after transfection, and cell lysates were subjected to immunoprecipitation with anti-SHP-2 polyclonal antibody overnight at 4°C. Immune complexes were captured by incubation with protein A-Sepharose and were washed several times with lysis buffer containing 10 mM NaF, once with STE buffer, and once with phosphatase assay buffer (24 mM HEPES, pH 7.4, 120 mM NaCl). Phosphatase assay buffer containing 10 mM para-nitrophenyl phosphate and 5 mM dithiothreitol was added to the immune complexes and incubated at 37°C for 30 min. Reactions were stopped by addition of 1.45 ml of 0.2 N NaOH, and the absorbance at 405 nm was measured.
Cell Adhesion Assays-Petri dishes were coated with fibronectin (5 g/ml) diluted in PBS for 1-2 h at 37°C and were blocked with 1% BSA in PBS for 30 min at 37°C. WT and D61G MEFs or D61G MEFs treated with siRNA were serum-deprived overnight and were detached by 0.05% trypsin/EDTA and neutralized with 0.5 mg/ml soybean trypsin inhibitor solution in DMEM. Cells were washed in serum-free DMEM containing 0.1% BSA, held in suspension for 30 min at 37°C under a humidified atmosphere containing 5% CO 2 , and were re-plated (575,000 cells per 60-mm dish) onto either 1% BSA blocked Petri dishes (suspension) or fibronectin-coated Petri dishes. Cells were harvested at the indicated times by washing twice in ice-cold PBS and lysing as indicated above.
For quantification of adhesion, wells of a 96-well plate (Linbro-Titertek) were coated with the indicated concentrations of fibronectin in triplicate. Serum-deprived cells were trypsinized and washed in serum-free DMEM containing 0.1% BSA. 20,000 cells/well were plated and allowed to attach for 30 min at 37°C. The attached cells were washed once with PBS, fixed in 4% paraformaldehyde solution in PBS for 20 min at room temperature, and stained with 0.5% crystal violet solution in 20% methanol for at least 1 h. Crystal violet was extracted with 10% acetic acid and was measured in a multiple well plate reader at 562 nm.
Cell Migration Assay-Migration of WT and D61G MEFs was quantified using a transwell migration assay. The bottom side of the transwell insert membrane (8 m pore size, Costar) was coated with 5 g/ml of fibronectin in triplicate at 37°C for 1 h and blocked with 1% BSA solution. Cells were trypsinized and washed as described above, plated at a density of 5,000 cells per well, and allowed to migrate for 5 h in a humidified incubator at 37°C. The cells that had migrated were washed once in PBS, fixed with 4% paraformaldehyde/PBS, and stained with 0.5% crystal violet solution. After the cells on top of the transwell insert membrane were scraped with cotton tips, the cells remaining on the bottom of the membrane were photographed in four fields/well using a 5ϫ objective and cells were counted using NIH Image J software.

RESULTS
Enhanced Association of Tyrosyl-phosphorylated Proteins with Activated SHP-2 Mutants-SHP-2 engages in a complex complement of protein-protein interactions mediated by its SH2 domains that serve to both activate and target it to specific subcellular compartments (1). Despite being constitutively active, mutants of SHP-2 likely require appropriate subcellular localization, mediated through its SH2 domains, to promote disease-associated signaling (19). We therefore sought to identify tyrosyl-phosphorylated proteins that might serve as signaling scaffolds for activated SHP-2 mutants in an attempt to explore the signaling mechanisms involved in the pathogenesis of NS/leukemia. We first began by generating four distinct activated SHP-2 mutants; the N308D mutation represents the most common mutation in NS (39); E69K and E76A mutations are associated with leukemia (39), and D61A is a molecular gain-of-function mutation (15) (Fig. 1A). These mutants were expressed transiently in C2C12 cells, and immune complex phosphatase assays were performed (Fig. 1B). Cells expressing these activated SHP-2 mutants, as expected (15,16,20,35,40), exhibited levels of phosphatase activity greater than that of WT overexpressing cells (Fig. 1B). The activated SHP-2 mutant expressing cells displayed varying levels of enhanced phosphatase activity with a rank order of E69K Ն E76A Ͼ N308D Ն D61A (Fig. 1B). Next, C2C12 cells cultured either under 10% FBS or 0.1% FBS conditions were transiently transfected with WT, D61A, N308D, or E76A activated SHP-2 mutants, and SHP-2 immune complexes derived from these cell lysates were immunoblotted with anti-phosphotyrosine antibodies (Fig. 1C). Cells expressing the activated SHP-2 mutants were found under proliferative conditions to interact with two tyrosyl-phosphorylated proteins that migrated with apparent molecular masses of 120 and 40 kDa, referred to herein as p120 and p40 proteins, respectively. In contrast, WT expressing cells failed to interact appreciably with either p120 or p40 tyrosyl-phosphorylated proteins (Fig. 1C). WT and SHP-2 activated mutants were comparably expressed (Fig. 1C), indicating that these activated SHP-2 mutants either enhance tyrosyl phosphorylation levels and/or interaction with the p120 and p40 tyrosyl-phosphorylated proteins.
SIRP␣ and PZR Are Targets for Hypertyrosyl Phosphorylation by Activated SHP-2 Mutants-We next set out to identify the SHP-2-associated p120 and p40 tyrosyl-phosphorylated proteins. Based on our previous observations, as well as those of others, we noted that the broad nature of the SHP-2-associated phosphotyrosyl p120 protein was reminiscent of the SHP-2binding protein known as SIRP␣, which also migrates with an apparent molecular mass between 110 and 120 kDa (41)(42)(43)(44). SIRP␣ is a transmembrane glycoprotein containing three immunoglobulin-like domains extracellularly, and four tyrosine residues intracellularly, two of which exist in context of an immunoreceptor tyrosine-based inhibitory motif (45,46). SIRP␣ was initially identified as a substrate and binding protein for SHP-2 (42,43,47), and it has been implicated in a number of signaling events, including the regulation of the extracellularregulated kinases (ERKs) 1 and 2, cell motility, and adhesionmediated integrin signaling (31,46,48,49). We tested whether SIRP␣ constitutes a component of the SHP-2-associated tyrosyl-phosphorylated p120 protein upon expression of the activated SHP-2 mutants.
The activated SHP-2 mutant enhanced the association of SHP-2 with the p120 and p40 tyrosyl-phosphorylated proteins using adenovirus-mediated delivery to express WT and E76A ( Fig. 2A). When immunoprecipitated, SIRP␣ was tyrosyl-phosphorylated to a greater extent in E76A as compared with WT expressing cells (Fig. 2B). We next determined whether SIRP␣ is, or constitutes a component of, the SHP-2-associated p120 tyrosyl-phosphorylated protein. To test this, we immunodepleted SIRP␣ from the lysates of cells infected with adenoviruses expressing GFP as a control, WT, or E76A followed by immunoprecipitation of SHP-2. When SIRP␣ was immunodepleted from these cell lysates followed by immunoprecipitation of SHP-2, the p120 tyrosyl-phosphorylated protein was depleted from these complexes as compared with the preimmune control immunodepletion (Fig. 2C). Both E76A and N308D SHP-2 mutants not only induced hypertyrosyl phosphorylation of SIRP␣, but also this was accompanied by increased levels of associated SHP-2 (Fig. 2D). These data identify SIRP␣ as the major component of the SHP-2-associated tyrosyl-phosphorylated p120 protein in cells expressing an activated SHP-2 mutant.
Next, we sought to identify the SHP-2-associated p40 tyrosyl-phosphorylated protein. Although a number of tyrosylphosphorylated SHP-2-interacting proteins have been identified, one of these proteins migrates with a molecular mass of ϳ40 kDa and is known as PZR. Like SIRP␣, PZR is also a transmembrane glycoprotein containing a single immunoglobulinlike domain extracellularly and two immunoreceptor tyrosinebased inhibitory motifs intracellularly (50). PZR was also shown to interact directly with SHP-2 and was suggested to be an SHP-2 substrate (51). To determine whether PZR constituted a component of the SHP-2-associated p40 tyrosyl-phosphorylated protein complex, C2C12 cells were infected with adenoviruses expressing GFP, WT, or E76A, and tyrosyl phosphorylation of PZR was determined. These experiments revealed that PZR was hypertyrosyl-phosphorylated in E76A as compared with WT expressing cells (Fig. 3A). We attempted to perform immunodepletion experiments to determine to what extent  PZR contributed to the SHP-2-associated p40 complex. However, PZR antibodies were incapable of immunodepleting cell lysates of PZR effectively. Regardless, these results still suggest that at least a component of the SHP-2-associated p40 phosphotyrosyl protein is PZR. The ability of PZR to undergo enhanced association with SHP-2 upon expression of the activated SHP-2 mutants was evaluated. Transient expression of either N308D or E76A resulted in increased association of SHP-2 with PZR (Fig. 3B). Unlike SIRP␣, where a number of reports have shown that it becomes tyrosyl-phosphorylated in response to growth factors, cytokines, and cell adhesion, the physiological stimuli that regulate PZR tyrosyl phosphorylation are unknown. When PZR was immunoprecipitated from cells that were stimulated with EGF, FGF-2, insulin-like growth factor-1, and serum, we found that it was not inducibly tyrosyl-phospho-rylated, although a slight increase in PZR tyrosyl phosphorylation was observed in response to PDGF (Fig. 3C). In contrast, PZR was inducibly tyrosyl-phosphorylated upon fibronectin plating (Fig. 3D). These results demonstrate that PZR, like SIRP␣,becomestyrosyl-phosphorylatedinresponsetoadhesiondependent signals.
SIRP␣ and PZR Are Targets for Enhanced Tyrosyl Phosphorylation and SHP-2 Recruitment in D61G Knock-in Mice-To determine whether SIRP␣ and/or PZR hypertyrosyl phosphorylation by activated SHP-2 mutants might have pathophysiological relevance to NS, we asked whether these transmembrane glycoproteins were hypertyrosyl-phosphorylated during embryogenesis in mice harboring a knock-in mutation of the NS-associated SHP-2 D61G mutation (18). Whole embryos were isolated from pregnant females between embryonic days 12.5 and 14.5. Extracts from WT and heterozygous (Ptpn11 D61G/ϩ ) embryos were subjected to immunoprecipitation for SHP-2, and the phosphatase activity was determined. The phosphatase activity of SHP-2 was ϳ3-fold more active in Ptpn11 D61G/ϩ embryos as compared with WT embryos (data not shown). Remarkably, SHP-2 coimmunoprecipitated with an identical profile of hypertyrosyl-phosphorylated associated proteins from Ptpn11 D61G/ϩ embryos as in transient cell culture experiments (Fig. 4A).    C. Moreover, SIRP␣ and PZR were hypertyrosyl-phosphorylated, and this correlated with enhanced SHP-2 binding in Ptpn11 D61G/ϩ embryos as compared with WT embryos (Fig.  4, B and C). We quantitated the levels of SIRP␣ and PZR tyrosyl phosphorylation in individual WT and Ptpn11 D61G/ϩ embryos (Fig. 4D). This analysis showed an increase in the distribution of Ptpn11 D61G/ϩ embryos expressing elevated levels of tyrosyl-phosphorylated SIRP␣ (Fig. 4D, left panel) and PZR (Fig. 4D, right panel) as compared with WT embryos. These results demonstrate that SIRP␣ and PZR are hypertyrosyl-phosphorylated in vivo and constitute prefer-ential SHP-2-interacting proteins during embryogenesis in Ptpn11 D61G/ϩ knock-in mice.

Activated SHP-2 Mutants Induce SIRP␣ and PZR Hypertyrosyl Phosphorylation in a c-Src-independent Manner-SHP-2
has been shown to be upstream of the Src family kinases (SFKs) (5,(21)(22)(23). In addition, both SIRP␣ and PZR contain immunoreceptor tyrosine inhibitory motifs, which are established SFK substrate motifs. Therefore, the activated SHP-2 mutants might induce SIRP␣ and PZR tyrosyl phosphorylation by activating the SFKs. Therefore, we determined whether c-Src activity is increased in Ptpn11 D61G/ϩ embryos. Whole embryo lysates were prepared from WT and Ptpn11 D61G/ϩ embryos, and c-Src Tyr-416 phosphorylation, which is representative of the active form of c-Src, was determined. We found that c-Src phosphorylation levels were unaltered in Ptpn11 D61G/ϩ embryos as compared with WT embryos (Fig. 5A). These results demonstrate that c-Src is unlikely to be a major target of the D61G NS mutant during embryogenesis that is responsible for either SIRP␣ or PZR tyrosyl phosphorylation. However, this result does not exclude the possibility that more discrete tissuespecific activation of the SFKs in Ptpn11 D61G/ϩ mice occurs. Another possibility is that SFKs other than c-Src might be tar-  gets for activated SHP-2 mutants that mediate SIRP␣ and PZR tyrosyl phosphorylation. To test other SFK members, fibroblasts lacking Src, Yes, and Fyn (SYFs) were transiently transfected with vector, WT, and three activated SHP-2 mutants. We found that all activated SHP-2 mutants were still capable of enhancing tyrosyl phosphorylation of the SHP-2-associated p120/SIRP␣ and p40/PZR proteins (Fig. 5B). Hence, the activated SHP-2 mutants appear to induce SIRP␣ and PZR hypertyrosyl phosphorylation in a c-Src-independent manner. Distinct Mechanisms of SIRP␣ and PZR Hypertyrosyl Phosphorylation by Activated SHP-2 Mutants-One possible explanation for why the activated SHP-2 mutants induced hypertyrosyl phosphorylation of SIRP␣ and PZR could be related to the open conformation of these mutants. When activated, the open conformation of SHP-2 is presumed to have increased availability to bind through its SH2 domains to tyrosyl-phosphorylated proteins resulting in their protection from dephosphorylation. To examine this possibility, we engineered a catalytically inactive/nonsubstrate-trapping SHP-2 mutant (R465M) in context of the E76A mutation to generate a compound SHP-2 mutant (E76A/R465M) that is catalytically inactive but resides in an open conformation. C2C12 cells were transfected with either vector, WT, E76A, or E76A/R465M along with plasmids expressing either SIRP␣ or PZR. As expected, expression of E76A resulted in the hypertyrosyl phosphorylation of both SIRP␣ and PZR (Fig. 6, A and B). Interestingly, expression of E76A/R465M also induced hypertyrosyl phosphorylation of SIRP␣ and PZR (Fig. 6, A and B). There are two possible interpretations of these data. The first is that hypertyrosyl phosphorylation of SIRP␣ and PZR might occur because of a protective effect of the SHP-2 SH2 domains. Hence, the catalytic activity of the activated SHP-2 mutant might be dispensable for SIRP␣ and PZR hypertyrosyl phosphorylation. The second possibility is that because both SIRP␣ and PZR are putative SHP-2 substrates, overexpression of a catalytically inactive mutant of SHP-2 could also result in their hypertyrosyl phosphorylation (47,51).
If the SH2 domains of the activated SHP-2 mutants simply protected SIRP␣ and PZR from dephosphorylation, expressing the SH2 domains of SHP-2 alone would result in the hypertyrosyl phosphorylation of SIRP␣ and PZR. However, if SHP-2 catalysis and/or adaptor function were required for SIRP␣ and PZR hypertyrosyl phosphorylation, the SH2 domains when overexpressed should function as a dominant-negative mutant of SHP-2 (26,52) and suppress SIRP␣ and PZR hypertyrosyl phosphorylation. To distinguish between these possibilities, C2C12 cells were transiently co-transfected with GST fusions of WT, N308D, E76A, and the SH2 domains alone of SHP-2 (SH2) along with either full-length SIRP␣ or PZR. When the tandem SH2 domains of SHP-2 were expressed in cells and the status of SIRP␣ tyrosyl phosphorylation was examined, we found that SIRP␣ tyrosyl phosphorylation was induced to levels equivalent to that of cells expressing either E76A or N308D (Fig. 6C). These results imply that SIRP␣ hypertyrosyl phosphorylation is likely caused by the SH2 domains of SHP-2 binding to, and subsequently protecting, SIRP␣ from dephosphorylation. In contrast, we found that PZR tyrosyl phosphorylation was suppressed by the SH2 domains of SHP-2 (Fig. 6D). The SH2 domains were expressed to levels that were comparable with WT and activated SHP-2 mutants (Fig. 6, C and D). These results indicate that, unlike SIRP␣, hypertyrosyl phosphorylation of PZR induced by the activated SHP-2 mutants is not a result of a protection effect by the SH2 domains. Therefore, in context of full-length activated SHP-2 mutants, PZR hypertyrosyl phosphorylation is presumably promoted either through the PTP domain of SHP-2 and/or an adaptor mechanism through its COOH terminus. Collectively, these observations suggest two distinct mechanisms for how the activated SHP-2 mutants induce SIRP␣ and PZR hypertyrosyl phosphorylation.
Enhanced ERK Activation in Response to Adhesion of Ptpn11(D61G) Fibroblasts-SIRP␣ plays a role in cytoskeletal reorganization and migration (49,53). Similarly, PZR has been shown to be involved in fibronectin-mediated cell migration (54). Because we have provided evidence that SIRP␣ and PZR are the two major hypertyrosyl-phosphorylated proteins associated with SHP-2 in Ptpn11 D61G/ϩ embryos, we hypothesized that MEFs derived from Ptpn11 D61G/D61G embryos (D61G MEFs) might display defects in cell migration, adhesion, and/or adhesion-mediated signaling. To test these possibilities we used MEFs from Ptpn11 D61G/D61G knock-in mice (18). We first determined whether SIRP␣ and PZR were hypertyrosyl-phosphorylated and bound increased levels of SHP-2 (D61G). As shown previously, SHP-2 (D61G) co-precipitated with tyrosyl-phosphorylated p120/SIRP␣ and p40/PZR proteins (Fig. 7A), and SIRP␣ was hypertyrosyl-phosphorylated in D61G MEFs as compared with WT MEFs (Fig. 7B). Similarly, PZR was hypertyrosyl-phosphorylated in D61G MEFs as compared with WT MEFs (Fig. 7C). Next, we determined whether there are differences in extracellular matrix-mediated adhesion and migration between D61G and WT MEFs. We detected a slight reduction in fibronectin-mediated adhesion in D61G MEFs as compared with WT MEFs (Fig. 7D). Using fibronectin as an extracellular matrix substrate, we tested whether the D61G MEFs exhibited migratory defects. We found no detectable differences in migration as quantified in a tran-swell migration assay between WT and D61G MEFs (Fig. 7E). SIRP␣ is involved in both positive and negative signaling effects to ERK in response to growth factors (42,49,55), although in response to adhesion-mediated integrin engagement SIRP␣ exerts a positive signaling effect on ERK (31). With regard to PZR, its role in ERK signaling has yet to be fully defined. Because NS-mediated signaling is proposed to involve enhanced Ras-ERK activation, we tested whether fibroblasts derived from the Ptpn11 D61G/D61G knock-in mice might exhibit enhanced adhesion-dependent activation of ERK and whether SIRP␣ and/or PZR contribute to this effect.
First, we determined whether fibronectin-mediated ERK activation was enhanced in D61G MEFs in response to adhesion. D61G MEFs when re-plated onto fibronectin displayed a sustained activation of ERK, whereas ERK activation was more transient in WT MEFs (Fig.  8A). These results demonstrated that the activated D61G SHP-2 mutant enhances adhesion-mediated ERK activation. Our working hypothesis predicts that the increased SIRP␣-D61G SHP-2 and/or PZR-D61G SHP-2 complex formation contributes to the sustained activation of ERK in response to adhesion-dependent signaling. Therefore, we tested whether down-regulating SIRP␣ expression attenuates the sustained activation of ERK in response to adhesion in D61G MEFs. We used siRNA to SIRP␣ to knock down its expression in D61G MEFs. As compared with nontargeting siRNA, transfection of SIRP␣ siRNA into D61G cells resulted in a substantial decrease in SIRP␣ expression (Fig. 8B). When we compared the kinetics of ERK activation in D61G MEFs following plating onto fibronectin, we found that siRNA SIRP␣-treated D61G MEFs exhibited an attenuation of ERK activation as compared with nontargeting siRNA-treated controls (Fig. 8B). These results suggest that SIRP␣ and consequently increased plasma membrane association of SHP-2 D61G with SIRP␣ contribute to the enhanced activation of ERK in response to adhesion. To determine whether PZR plays a role in enhanced ERK activation in D61G MEFs in response to adhesion, we used siRNA to PZR to knock down its expression in D61G MEFs. As shown in Fig. 8C, we were able to eliminate its expression by more than 90%. When  we determined the kinetics of ERK activation in siRNA PZRtreated D61G MEFs following plating onto fibronectin, there was a moderate but consistent reduction of adhesion-dependent ERK activation as compared with the control treated cells (Fig. 8C). Together, these data demonstrate that increased association of SIRP␣ with SHP-2 D61G, and to a lesser extent with PZR, contributes to enhanced adhesionmediated ERK signaling.

DISCUSSION
SHP-2 directs downstream signaling through its interactions with receptors, adaptor molecules, and transmembrane glycoproteins (1). In this regard, the effects of disease-associated SHP-2 signaling are likely to invoke similar requirements to promote aberrant signaling. Whereas the identity of some of these interacting proteins such as Gab-1 (17,18) and Gab-2 (19,33) has been elucidated, the pathophysiological targets of SHP-2 in vivo remain unknown. Here we identify the transmembrane glycoproteins SIRP␣ and PZR as hypertyrosyl-phosphorylated NS-associated SHP-2 mutant-interacting proteins during embryogenesis in a mouse model of NS. We show that SIRP␣ contributes to the enhanced ERK activation in D61G MEFs in response to adhesion. These data suggest that enhanced recruitment of NS-associated SHP-2 mutants to SIRP␣ and PZR during embryogenesis might contribute to enhanced ERK activation in response to extracellular matrix signals. Many developmental processes such as heart morphogenesis, neurogenesis, and vasculogenesis require the precise regulation of adhesion-dependent signals (56). Because NS patients exhibit complex cardiovascular defects and mental retardation, it is therefore conceivable that these defects might arise, at least in part, through promiscuous adhesion-dependent signaling mediated by activated SHP-2 mutants.
We identified PZR as a hypertyrosyl-phosphorylated protein in cells expressing activated SHP-2 mutants and during embryogenesis in the NS knock-in mouse model. The function of PZR still remains unclear, although our data suggest a potential role for PZR in adhesiondependent signaling because PZR becomes inducibly tyrosyl-phosphorylated upon plating cells onto fibronectin, yet it fails to do so in response to growth factors. In support of these observations, it has been proposed that PZR is involved in the regulation of cell motility (54). However, we have been unable to observe such effects upon overexpression of PZR in either myoblasts or endothelial cells (data not shown). It is clear that additional work is needed before the function of PZR is fully elucidated. Nevertheless, our experiments are consistent with the idea that PZR might serve as an adhesion-regulated signaling molecule. It is not unreasonable to speculate that PZR tyrosyl phosphorylation and association with SHP-2 may play a role in adhesion-mediated signaling and that hypertyrosyl phosphorylation of PZR by NS/leukemia-associated SHP-2 mutants results in dysregulated extracellular matrix-dependent signaling. Consistent with this, we observe a moderate attenuation of adhesion-dependent ERK activation in PZR siRNA-treated D61G cells. These results suggest that PZR might only play a minor role in enhanced adhesion-dependent ERK activation. Further work will be required to identify the major pathway through which PZR signals.
Unlike PZR, the function and mechanisms of action of SIRP␣ are more defined. SIRP␣ becomes tyrosyl-phosphorylated and associates with SHP-2 in response to a variety of stimuli, including growth factors and extracellular matrix engagement (31,43). In most cases it appears that tyrosyl phosphorylation of the ITIMs within the cytoplasmic tail of SIRP␣ are required for signaling. For example, overexpression of wild-type SIRP␣ leads to increased cell migration, whereas overexpression of the tyrosine phosphorylation site mutant has no effect (53 over, a role for the cytoplasmic domain in general, has been provided using fibroblasts lacking the SIRP␣ cytoplasmic domain. These cells display increased focal adhesions, actin stress fibers, and migration defects (49). Therefore, our observation that SIRP␣ becomes hypertyrosyl-phosphorylated upon expression of the activated SHP-2 mutants suggests that downstream signaling from SIRP␣ is likely to be altered. Although we were unable to detect major differences in either cell adhesion and migration in D61G MEFs, we did observe that siRNA-mediated down-regulation of SIRP␣ expression decreases ERK activation in D61G MEFs in response to adhesion. These observations suggest that SIRP␣ contributes to SHP-2 D61G mutant-induced enhanced ERK activation in response to adhesion. However, we cannot exclude the possibility that other SIRP␣-interacting proteins also participate in ERK activation. How do activated SHP-2 mutants lead to hypertyrosyl phosphorylation of these target proteins? We initially hypothesized that the activated SHP-2 mutants stimulate a tyrosine kinase that subsequently phosphorylates SIRP␣ and PZR. Several studies have shown that SHP-2 functions upstream of the SFKs; moreover, both SIRP␣ and PZR are targets for phosphorylation by c-Src (30,31,57). Thus, c-Src is an attractive tyrosine kinase that could mediate both SIRP␣ and PZR tyrosyl phosphorylation by the activated SHP-2 mutants. When we examined c-Src activation, no major differences were detected between WT and Ptpn11 D61G/ϩ embryos under conditions in which both SIRP␣ and PZR were hypertyrosyl-phosphorylated. These results suggest that c-Src is unlikely to be a major target of SHP-2 that mediates SIRP␣ and PZR hypertyrosyl phosphorylation in this mouse model of NS at this stage of embryogenesis. However, we cannot exclude the possibility that c-Src becomes activated by the NS-associated SHP-2 mutant in a tissue-specific manner similar to that observed for ERK (18).
To further examine the mechanism by which activated SHP-2 mutants lead to hypertyrosyl phosphorylation of SIRP␣ and PZR, we determined whether the catalytic activity of the activated SHP-2 mutant is required for hypertyrosyl phosphorylation of these target proteins. Both SIRP␣ and PZR are hypertyrosyl-phosphorylated in the presence of the E76A/ R465M SHP-2 mutant that represents an open conformation/ catalytically inactive nonsubstrate-trapping mutant of SHP-2. The activated SHP-2 mutants are predicted to be in an open conformation even in the absence of any ligand (14,15,20). Therefore, it is conceivable that the SH2 domains of the activated SHP-2 mutants bind and protect their target proteins from dephosphorylation. Indeed, we found that hypertyrosyl phosphorylation of SIRP␣ occurred upon expression of the SH2 domains alone. The simplest interpretation of these results is that the SH2 domains of SHP-2 protect SIRP␣ from dephosphorylation. Although there are other possibilities, such as the SH2 domains of SHP-2 activating a tyrosine kinase that phosphorylates SIRP␣, there is no evidence to support a positive signaling effect of the SH2 domains of SHP-2 alone. Hence, we favor the model that the SH2 domains of the activated SHP-2 mutants protect select phosphotyrosyl sites from dephosphorylation. In contrast, expression of the SH2 domains of SHP-2 alone blocked PZR tyrosyl phosphorylation. These results demonstrate that PZR hypertyrosyl phosphorylation does not occur as a result of a protection mechanism by the SH2 domains of SHP-2. We propose that the SH2 domains of SHP-2 acts as a dominant-negative (26,52), thereby preventing activation and/or recruitment of a tyrosine kinase that phosphorylates PZR. Even though the identity of this putative tyrosine kinase is unknown, our data suggest that it is unlikely to be either Src, Yes, or Fyn. The possibility that other Src family kinase members are involved in PZR tyrosyl phosphorylation has yet to be formerly ruled out. Altogether, our results demonstrate that the activated mutants of SHP-2 lead to hypertyrosyl phosphorylation of SIRP␣ and PZR by distinct mechanisms.
In summary, we have identified that activated SHP-2 mutants in cultured cells and in a mouse model of NS cause hypertyrosyl phosphorylation of SIRP␣ and PZR and increased SHP-2 association. Our results imply that NS/leukemia-associated SHP-2 mutants engage extracellular matrix signals resulting in their increased recruitment to the plasma membrane via SIRP␣ and PZR, which may then function to enhance ERK signaling.