Development of an Efficient "Substrate-trapping" Mutant of Src Homology Phosphotyrosine Phosphatase 2 and Identification of the Epidermal Growth Factor Receptor, Gab1, and Three Other Proteins as Target Substrates*

Yehenew M. Agazie and Michael J. HaymanDagger

From the Department of Molecular Genetics and Microbiology, State University of New York at Stony Brook, Stony Brook, New York, 11794-5222

Received for publication, October 18, 2002, and in revised form, February 5, 2003

    ABSTRACT
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES

Src homology containing phosphotyrosine phosphatase 2 (SHP2) is a positive effector of growth factor, cytokine, and integrin signaling. However, neither its physiological substrate nor its mechanism of action in tyrosine kinase signaling has been demonstrated. We reasoned that the identification of physiological substrates of SHP2 would be a stepping stone in elucidating its mechanism of action, and, thus, we constructed a potent trapping mutant of SHP2. Surprisingly, the frequently used Asp to Ala substitution did not give rise to a trapping mutant. However, we were able to develop an efficient trapping mutant of SHP2 by introducing Asp to Ala and Cys to Ser double mutations. The double mutant (DM) protein identified the epidermal growth factor receptor (EGFR), the Grb2 binder 1, and three other, as yet unidentified, phosphotyrosyl proteins as candidate physiological substrates. Given that substrate trapping occurred in intact cells and that the interaction was very specific, it is highly likely that EGFR and Gab1 represent physiological SHP2 substrates. Therefore, the DM protein would serve as an important tool in future SHP2 studies, including identification of p190, p150, and p90.

    INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES

Tyrosine phosphorylation of proteins is central to receptor tyrosine kinase (RTK)1 signal transduction (1-5). Some of the most extensively studied RTKs include the epidermal growth factor receptor (6), the platelet-derived growth factor receptor (7), the fibroblast growth factor receptor (8), the insulin receptor (9), and the Met receptor. The signaling activity of RTKs is controlled by a distinct class of enzymes known as phosphotyrosyl phosphatases (PTP) (9-17). Recent advances in the area of PTPs indicate that tyrosine phosphorylation and dephosphorylation of signaling proteins must be integrated for normal cell growth, differentiation, migration, and development (18). Uncontrolled tyrosine phosphorylation of RTKs and/or their downstream targets results in abnormal cell growth and differentiation that lead to diseases such as cancer and developmental problems (6, 19, 20). Thus, the PTPs play major role in maintaining proper cellular signaling.

PTPs encompass a large family of transmembrane and non-transmembrane enzymes (14, 15, 18). Among the non-transmembrane mammalian PTPs, only two are Src homology 2 (9) domain-containing, namely SHP1 (also called PTP1B, SH-PTP1, PTP1C, HCP, or SHP) and SHP2 (also known as SH-PTP2, SH-PTP3, PTP2C, PTP1D, or SYP). SHP2 is a ubiquitously expressed mammalian PTP, whereas its close relative, SHP1, is primarily restricted to hematopoietic cells (11, 12, 14, 21). SHP1 and SHP2 show significant structural homology but play opposite physiological roles, the former being a negative and the latter being primarily a positive regulator of tyrosine kinase signaling (14, 16, 18). Both PTPs possess two tandemly arranged SH2 domains in their N-terminal region, a phosphatase domain in their C-terminal regions and tyrosine phosphorylation sites in their extreme C-terminal regions (7, 11, 22). A Drosophila PTP, Corkscrew, also has similar structure but shows functional similarity only to SHP2 (23-25).

SHP2 directly binds to some autophosphorylated RTKs such as the epidermal growth factor receptor (6), the platelet derived growth factor receptor (7), and the Met receptor (7, 14, 23, 26) or to tyrosine phosphorylated adaptor proteins such as insulin receptor substrate 1 and 2 (27, 28), Shc, fibroblast growth factor receptor (FGFR) substrate 1 and 2 (FRS1 and 2) (8, 29), and Grb2 adaptor binder 1 and 2 (30-34). These interactions are essential for growth factor- or cytokine-induced activation of the Ras-extracellular signal-regulated kinase (ERK) pathway, but the molecular mechanism by which SHP2 effects these functions is unknown. There has been a great deal of effort to identify SHP2 substrates. For instance, it was shown that SHP2 dephosphorylates SHP2 substrate 1 (SHPS1) under in vitro conditions (35-37), but the in vivo relevance of these findings is not clear. A stumbling block in defining the molecular mechanism of SHP2 could be that the physiological substrates (9) of SHP2 have not yet been identified. Thus, identification of such substrates would be a stepping stone toward elucidating the mechanism of action of SHP2 in RTK signaling.

Substrate-trapping mutants of PTPs have been shown to be ideal reagents for substrate identification. It was demonstrated that such mutants of PTPs can be produced by mutation of Asp to Ala in the conserved WPD loop (38). For instance, Asp to Ala mutants of PTP1B, TC-PTP, and PTP-PEST helped identify EGFR, p52Shc, and p130Cas as candidate substrates, respectively (38-40). It was also demonstrated that a Cys to Ser mutation in the signature motif (41) makes PTPs become substrate trapping (42-44), but with less efficiency. Here, we report the development of a new trapping mutant for SHP2 by introducing Cys to Ser and Asp to Ala double mutations that were superior to the single Cys to Ser mutant. Surprisingly, and in contrast to widely held views, an Asp to Ala mutant of SHP2 was non-trapping. To our knowledge, this is the first report on the development of an efficient trapping mutant of SHP2 and the first demonstration in intact cells that EGFR and Gab1 are target substrates.

    MATERIALS AND METHODS
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES

Cells, Cell Culture, Antibodies, et Cetera-- The cell types used in this study were COS-1, human embryonic kidney 293T, and NIH-3T3. All were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum and maintained at 37 °C with 7% CO2. The epidermal growth factor was purchased from Invitrogen. Anti-PTP1D, anti-RasGAP, anti-p130Cas, anti-FYB, and anti-p190 were purchased from BD Transduction Laboratories. Anti-phosphotyrosine monoclonal antibody (4G10) was a gift from Dr. Morrison, and anti-SOS1, anti-Gab1, and anti-PDGFR antibodies were purchased from Upstate Biotechnology. Polyclonal antibodies against SHP2 and EGFR were raised by immunizing rabbits with glutathione S-transferase (GST) fusion proteins of the amino-terminal region and the kinase domain, respectively. Horseradish peroxidase-conjugated secondary antibodies were purchased from Amersham Biosciences. For fluorescent quantitation of band intensities, alkaline phosphatase-conjugated anti-rabbit or anti-mouse secondary antibodies were purchased from Zymed Laboratories Inc., whereas the AttoPhos fluorescent substrate system was from Promega.

Plasmid Construction and Site-directed Mutagenesis-- Construction of WT-SHP2 or C459S-SHP2 mutant into the pMT2 vector was described previously (45). The D425A-SHP2 or D425A/C459S-SHP2 mutants were obtained by introducing the indicated point mutation to the WT or the C459S mutants using complementary primers that span the WPD loop region. The Stratagene site-directed mutagenesis kit and protocol were used to produce the mutants. Introduction of the point mutation was confirmed by sequencing the relevant region.

For purification of His-tagged WT or C459S proteins, corresponding cDNA was cut out from the pMT2 vector with EcoRI and then inserted in-frame into pET-28(a+) vector (Novagen) using the same restriction site. Orientation of inserts was analyzed by restriction digestion followed by protein expression.

Cell Transfection, Preparation of Lysates, Immunoprecipitation, and Immunoblotting-- Plasmid constructs were transfected into COS-1, 293T, or NIH-3T3 cells grown to ~60-70% confluency in 6-cm dishes using the FuGENE transfection regent as recommended by the manufacturer (Invitrogen). After transfection, cells were incubated for ~30 h, serum-starved for ~12 h, and then stimulated with 100 ng/ml EGF or 10% fetal calf serum. All immunoprecipitation and immunoblotting experiments were performed as described previously (46).

Fluorescent Quantitation of Protein Band Intensities-- Separated proteins on nitrocellulose membranes were stained with a primary antibody for 2 h at room temperature or overnight at 4 °C, washed three times, and then stained with alkaline phosphatase-conjugated secondary antibody for 1 h at room temperature. After washing three times, membranes were exposed to AttoPhos fluorescent substrate for 1 min, and band intensities were measured on a PhosphorImager using ImageQuant software (Storm 860 PhosphorImager, Amersham Biosciences).

In Vitro Phosphatase Assay-- Dephosphorylation of substrate phosphotyrosyl proteins was performed as described previously (47). Briefly, EGFR or Gab1 was immunoprecipitated from COS-1 or 3T3 cells with anti-EGFR or anti-Gab1 antibody, respectively, washed 2 times with lysis buffer and once with phosphatase assay buffer (50 mM sodium acetate, pH 5.5, 50 mM NaCl, 10 mM dithiothreitol, and 2 mM EDTA). Beads were resuspended in the phosphatase buffer, varying concentrations of WT-PTP were added, and then the beads were incubated for 30 min at 30 °C with shaking. The double mutant (DM) protein was used as a negative control in the phosphatase assay. Reactions were performed in a total volume of 100 µl and stopped by adding Laemmli sample loading buffer and boiling. The level of tyrosine phosphorylation of substrate proteins was analyzed by immunoblot with an anti-Tyr(P) antibody.

    RESULTS
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES

Despite intensive and extensive research to identify physiological substrates and define the site of action of SHP2, the problem has remained virtually unresolved. We reasoned that identification of the physiological substrates of SHP2 would be a logical step forward. For this, we produced an efficient trapping mutant of SHP2 that would be instrumental in identifying target phosphotyrosyl substrate proteins.

Production of a Trapping Mutant of SHP2-- Previous reports showed that trapping mutants of PTPs can be produced by mutation of Asp to Ala in the conserved WPD loop (38). We utilized the same principle and introduced an Asp to Ala point mutation into WT-SHP2. WT-SHP2, D425A-SHP2, or C459S-SHP2, herein after referred to as WT, D425A, and C459S, respectively, were transfected into COS-1 cells, and EGF-induced coprecipitation of phosphotyrosyl proteins was analyzed by immunoblot. In cells transfected with the C459S construct, three phosphotyrosyl protein bands identified as p70, p120, and p170 were observed. In cells transfected with the WT or D425A, relatively weaker bands of p70 and p170 were observed. Reblotting with anti-SHP2 and anti-EGFR antibodies showed that the p70 and the p170 were SHP2 and the EGFR, respectively (data not shown). These results showed that the D425A protein was non-trapping.

We hypothesized that the trapping ability of the C459S mutant might be improved by introducing an additional mutation. For this, we introduced a D425A mutation into the C459S mutant and produced a double mutant SHP2 (D425A/C459S-SHP2), herein after referred to as the DM protein. The DM protein, together with the above mentioned constructs, were transiently overexpressed in COS-1 cells, and expression was monitored by immunoblot; all were expressed at approximately five times that of the endogenous SHP2 (Fig. 1A, bottom). The pattern of tyrosine-phosphorylated proteins was similar in all cases except p70, which was significantly intense in the C459S and the DM cells (Fig. 1A, top). Several phosphotyrosyl proteins coimmunoprecipitated with the DM protein, which were identified as p170, p150, p120 and p90 (Fig. 1B, top). Lower amounts of p170 and p120 also coprecipitated with the C459S protein as well. In the case of the WT or the D425A protein, a significantly lower amount of p170 only was observed. Reblotting with anti-EGFR and ant-SHP2 antibodies reconfirmed that p170 and p70 were EGFR and SHP2, respectively (Fig. 1B, middle and bottom). Interestingly, the tyrosine phosphorylation level of the D425A protein was similar to the WT, suggesting that it could be enzymatically active. By visually comparing band intensities (both anti-Tyr(P) and anti-EGFR blots), it was apparent that significantly higher amounts of phosphotyrosyl proteins coprecipitated with the DM protein, suggesting that the DM protein was the most efficient substrate-trapping mutant.


View larger version (57K):
[in this window]
[in a new window]
 
Fig. 1.   The DM protein trapped a 170-kDa and several other phosphotyrosyl proteins in EGF-stimulated COS-1 cells. A, anti-Tyr(P) (pY) immunoblot (IB) of total cell lysates (TCL) show the presence of similar amount and pattern of tyrosine-phosphorylated proteins in all experimental groups and also to show expression of the different SHP2 proteins. B, immunoprecipitation (IP) of COS-1 cell lysates with anti-SHP2 and immunoblot with the indicated antibodies, i.e. anti-Tyr(P) (pY, top), anti-EGFR (EGFR, middle), and anti-SHP2 (SHP2, bottom). C, total cell lysates of 293T cells blotted with anti-Tyr(P) (top) and anti-SHP2 (bottom). D, immunoprecipitation of 293T cell lysates with anti-SHP2 and immunoblot with the indicated antibodies, i.e. anti-Tyr(P) (top), anti-EGFR (middle) and anti-SHP2 (bottom). Vec, vector; D/A, D425A-SHP2; C/S, C495S-SHP2.

Because COS-1 cells express higher amounts of the EGFR, which was the most tyrosine-phosphorylated protein upon EGF stimulation, it is possible that this might have influenced the interaction. To test this possibility, we conducted the same experiment in 293T cells, which express significantly lower amounts of the EGFR (Fig. 1C, top). As in COS-1 cells, four phosphotyrosyl proteins that correspond in molecular weight to p170, p150, p120, and p90 coimmunoprecipitated with the DM protein, and lower amounts of p170 and p120 coimmunoprecipitated with the C459S protein (Fig. 1D, top). Again, significantly lower amounts of p170 were detected with the vector, the WT, or the D425A proteins. Reprobing the membrane with anti-EGFR and anti-SHP2 antibodies confirmed that p170 and p70 were EGFR and SHP2, respectively (Fig. 1D, middle and bottom). Therefore, the trapping of EGFR by the DM protein was an inherent property of this mutant but not a concentration artifact. Furthermore, it is also important to note that the DM protein had no obvious effect on the overall tyrosine phosphorylation of the cellular proteins. The tyrosine-phosphorylated proteins trapped by the DM protein were a very specific subset of the overall tyrosine-phosphorylated proteins (compare Fig. 1, A with B and C with D). This indicated that the DM protein was reacting in a very specific manner with the cellular proteins.

Time Course Studies on the Trapping of EGFR by Mutant SHP2 Proteins-- We examined the dynamics of substrate trapping by the DM protein at various time points of stimulation. A higher amount of EGFR coprecipitated with the DM protein at all time points tested and seemed to accumulate over time (Fig. 2, top panel). Despite the presence of a high amount of tyrosine-phosphorylated EGFR and a comparable expression level of SHP2 constructs (Fig. 2, middle and bottom), a lower amount of EGFR coprecipitated with the C459S protein and did not show any increase over time. Similarly, a significantly lower amount of EGFR coprecipitated with the WT and D425A proteins. These findings suggested that the DM protein formed a stable complex with its target substrate EGFR, which would be a desirable characteristic to isolate and identify other SHP2 substrates.


View larger version (47K):
[in this window]
[in a new window]
 
Fig. 2.   EGFR trapped by the DM protein accumulated over time. Lysates from COS-1 cells transfected with the indicated plasmid constructs were immunoprecipitated (IP) with anti-SHP2 and immunoblotted (IB) with anti-EGFR (top). The same membrane was stripped and reblotted with anti-SHP2 to show immunoprecipitation of an equal amount of SHP2 in all lanes (bottom). Total cell lysate (TCL) was immunoblotted with anti-Tyr(P) (pY) to show the presence of an equal amount of tyrosine-phosphorylated EGFR in all stimulated cells (middle). Vec, vector; D/A, D425A-SHP2; C/S, C495S-SHP2.

SHP2 Partially Dephosphorylates the EGFR-- The trapping of the EGFR by the DM protein indicated that it could be a target substrate of SHP2. We thus determined the level of EGFR tyrosine phosphorylation in COS-1 cells transfected with the SHP2 constructs used in Fig. 1. Although the expression of all SHP2 proteins was comparable (Fig. 3A, bottom), the tyrosine phosphorylation level of the EGFR was moderately higher only in the DM cells (Fig. 3A, top). Interestingly, overexpression of the WT protein did not result in global dephosphorylation of the EGFR (compare vector (V) versus WT, Fig. 3A, top), suggesting that only a few phosphotyrosine (pY, Fig. 3A, top) residues were targeted. To better compare these observations, band intensities were quantified. In the vector, the WT and the D425A transfected cells, the level of EGFR tyrosine phosphorylation was virtually identical, ~50-fold over basal. However, in the C459S and the DM cells it was 60- and 90-fold over basal, respectively (Fig. 3B). When comparison was made between the vector, the WT, or the D425A and the C459S or the DM, the level of EGFR tyrosine phosphorylation in the C459S and the DM cells was 1.2 and 1.8-fold higher, respectively.


View larger version (46K):
[in this window]
[in a new window]
 
Fig. 3.   SHP2 partially dephosphorylated EGFR. A, immunoprecipitation (IP) with anti-EGFR and immunoblot (IB) with anti-Tyr(P) (pY, top) of COS-1 cells transfected with the indicated SHP2 constructs. The same membrane was stripped and blotted with anti-EGFR (middle) to show the presence of an equal amount of EGFR in all lanes and stripped with anti-SHP2 (bottom) to demonstrate expression of the different SHP2 proteins. B, quantitation of the level of EGFR tyrosine phosphorylation. C, immunoblot analysis of the level of EGFR tyrosine phosphorylation following expression of increasing amounts of the WT protein. The DM protein was used as a control. TCL, total cell lysate; V, vector plasmid without SHP2 insert; D/A, D425A-SHP2; C/S, C495S-SHP2

The dephosphorylation of EGFR by SHP2 was also tested by expression of an increasing amount of the WT protein. Maximum dephosphorylation was achieved at 5 µg of DNA per 6-cm plate of confluent COS-1 cells, and no further dephosphorylation was observed by raising the DNA concentration to 10 µg (Fig. 3C), indicating that SHP2 is not a global phosphatase. As expected, cells transfected with 10 µg of the DM construct did not show any dephosphorylation of the EGFR. These results were consistent with specific and partial dephosphorylation and suggested that the EGFR might be an in vivo substrate of SHP2.

SHP2 Dephosphorylates EGFR and Gab1 in Vitro-- To further confirm that phosphotyrosyl proteins trapped by SHP2 are target substrates, we carried out an immunocomplex in vitro phosphatase assay. We concentrated only on the EGFR and Gab1, because these are the only identified substrates (Figs. 1 and 6) against which specific antibodies are available. EGFR from COS-1 cells stimulated with EGF or Gab1 from 3T3 cells stimulated with serum was immunoprecipitated with a specific antibody and then subjected to a PTP assay using varying concentrations of purified WT protein (see "Materials and Methods" for details). The WT protein partially dephosphorylated both the EGFR and Gab1, whereas the DM protein did not (Fig. 4, A and B, top). In both cases, the addition of five times more WT protein did not result in global dephosphorylation, further confirming that whether it is inside or outside the cell, SHP2 is very specific. Reblotting with anti-EGFR (Fig. 4A, bottom) or anti-Gab1 (Fig. 4B, bottom) showed that a comparable amount of protein was present in all lanes.


View larger version (63K):
[in this window]
[in a new window]
 
Fig. 4.   SHP2 dephosphorylates EGFR and Gab1 in vitro. COS-1 or NIH-3T3 cells were grown to confluency, serum starved for 6 h, treated with 1 mM vanadate for 30 min, and then stimulated with EGF or serum for 10 min, respectively. Lysates prepared from these cells were subjected to immunoprecipitation (IP) with anti-EGFR or anti-Gab1 antibody and then to an immunocomplex phosphatase assay as described under "Materials and Methods." Note that the WT protein partially dephosphorylated EGFR (A) and Gab1 (B) but not the DM protein. The immunoblots (IB) presented were representatives of at least three independent experiments. pY, Tyr(P).

Subset of SHP2 Substrates Vary Depending on Stimulus and Cell Type-- In EGF-stimulated COS-1 cells, the EGFR and three unidentified phosphotyrosyl proteins were shown to be SHP2 substrates. We treated SHP2-transfected COS-1 cells with another stimulus, namely 10% fetal calf serum. Anti-Tyr(P) blot showed trapping by the DM protein of four phosphotyrosyl proteins that were identified as p190, p150, p120, and p90 (Fig. 5A, top). What was different from the case with EGF-stimulated cells was the presence of p190 and the absence of p170 (EGFR), suggesting that SHP2 substrates vary depending on the stimulus. Despite comparable expression levels (Fig. 5A, bottom), a relatively lower amount of p190 and p90 only were trapped by the C459S protein. Furthermore, a significantly lower amount of p190 coprecipitated with the endogenous or overexpressed WT and the D425A proteins. For the purpose of comparing the COS-1 cell results with other cell type, we conducted a similar experiment in NIH-3T3 cells using only three of the SHP2 proteins. Again, the DM protein trapped four phosphotyrosyl proteins identified as p190, p150, p120, and p90 (Fig. 5B, top). In case of the C459S protein, p190 and a significantly lower amount of p150 and p90 were trapped. Some amount of p190 also coprecipitated with the WT protein, which may represent an SH2 domain-mediated interaction. These results indicated that the subset of SHP2 target substrates varies depending on the stimulus and/or cell type.


View larger version (52K):
[in this window]
[in a new window]
 
Fig. 5.   The DM protein identified four phosphotyrosyl proteins as its target substrates in serum-stimulated cells and only p190 in PDGF-stimulated cells. COS-1 (A and C) or NIH-3T3 (B and D) cells were transfected with an expression plasmid containing the indicated SHP2 constructs and stimulated either with serum or PDGF. Lysates prepared from these cells were subjected to immunoprecipitation (IP) with anti-SHP2 and immunoblot (IB) with anti-Tyr(P) (pY) antibodies (top panel, in all cases). Corresponding membranes were stripped and reblotted with anti-SHP2 antibody to show expression of SHP2 proteins (bottom panel, in all cases). Time course study with serum stimulation in COS-1 cells is shown in C. Vec, vector; D/A, D425A-SHP2; C/S, C495S-SHP2.

Time course studies in COS-1 cells with serum stimulation revealed that only the DM protein was able to trap significantly higher amounts of target substrates at all time points tested, and the amounts seemed to slightly increase over time, at least in case of p190 and p90 (Fig. 5C). Both in COS-1 and NIH-3T3 cells, the trapping of p90 and p150 was serum stimulation-independent, although stimulation enhanced the amount. In contrast, the amount of target substrates trapped by the C459S protein was low and largely serum stimulation-dependent.

Previous studies demonstrated that SHP2 dephosphorylates the PDGFR under in vitro conditions (47). Because serum contains PDGF and the molecular weight of p190 was close to that of the PDGFR, we assumed that p190 could be the PDGFR. For this, we overexpressed the DM protein in NIH-3T3 cells and then stimulated them with PDGF for 10 min. Lysates prepared from these cells were subjected to immunoprecipitation with anti-SHP2 and immunoblot analysis first with anti-Tyr(P) and later with anti-PDGFR antibodies. An anti-Tyr(P) blot detected p190 in PDGF-stimulated cells (Fig. 5D, top). Interestingly, only p190 was trapped by the DM protein, although there were several phosphotyrosyl protein bands in the corresponding lysate (compare +PDGF IP with +PDGF TCL in Fig. 5D), suggesting that the interaction was highly specific. However, reprobing the membrane with several anti-PDGFR antibodies could not positively identify p190 as the PDGFR (data not shown). Thus, the identity of p190, p150, and p90 remain unknown. Further reprobing of the membrane with anti-SHP2 antibody showed the presence of equivalent amount of the DM protein in all lanes (Fig. 5D, bottom).

Identification of the Other Phosphotyrosyl Protein Bands-- We made efforts to identify the other phosphotyrosyl proteins that were trapped by the DM protein using a variety of antibodies against known proteins in the indicated molecular weight range. Analysis with anti-p130Cas, anti-FYB, and anti-SHP2 substrate 1 antibodies gave negative results (data not shown). Another potential candidate substrate is Gab1, which was shown to interact with SHP2 (33). Immunoblotting anti-SHP2 immunoprecipitates with anti-Gab1 showed the presence of a higher amount of Gab1 in the DM cells as compared with the WT or the C459S cells (Fig. 6A, top), although the expression level of all SHP2 proteins was similar (Fig. 6A, bottom). These experiments were repeated in a reciprocal way, i.e. immunoprecipitation with anti-Gab1 and immunoblotting with anti-SHP2. Significantly higher amount of the DM protein coprecipitated with Gab1 as compared with the WT or the C459S proteins (Fig. 6B, top panel). Reprobing with anti-Gab1 antibody showed that there was an equivalent amount of Gab1 in all lanes (Fig. 6B, middle). Immunoblot with anti-Tyr(P) of Gab1 immunoprecipitates showed that the tyrosine phosphorylation level of Gab1 was higher in the DM cells as compared with the WT or the C459S cells (Fig. 6B, bottom). Similar results were obtained in NIH-3T3 cells (data not shown). Thus, Gab1 could be a substrate of SHP2.


View larger version (58K):
[in this window]
[in a new window]
 
Fig. 6.   The DM protein identified Gab1 as its target substrate. COS-1 cells were transfected with the indicated SHP2 constructs, and lysates prepared from these cells were subjected to immunoprecipitation (IP) and immunoblot (IB) analysis. A, immunoprecipitation with anti-SHP2, and immunoblot with anti-Gab1 (top) and anti-SHP2 (bottom). B, immunoprecipitation with anti-Gab1 and immunoblot with anti-SHP2 (top), anti-Gab1 (middle), and anti-Tyr(P) (pY, bottom). C/S, C495S-SHP2.


    DISCUSSION
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES

Genetic and biochemical studies have demonstrated that SHP2 is primarily a positive effector of RTK signaling. Deletion of the N-SH2 or the phosphatase domain or mutation of the conserved Cys residue to Ser (C459S) abrogated the biological activity of SHP2 (48, 49). Mice homozygous for the N-SH2 deletion mutant die in the uterus before day E10.5 from multiple defects in mesoderm development (19). Also, microinjection of SH2 or phosphatase deletion mutant mRNAs into Xenopus eggs caused abnormal embryonic development (49, 50). However, the mechanism(s) by which SHP2 effects its biological activity is unknown. We reasoned that identification of physiological substrates might be a stepping stone in elucidating the molecular mechanism of SHP2 function. Here, we report the development of an efficient substrate-trapping mutant of SHP2 and also demonstrate for the first time in intact cells that EGFR, Gab1, and three other as yet unidentified phosphotyrosyl proteins are candidate physiological substrates.

Trapping mutants of PTPs can be produced by the mutation of Asp to Ala in the conserved WPD loop (38, 40, 51). We utilized the same principle and produced an Asp to Ala mutant of SHP2, but found that it was not trapping. In line with our findings, it was recently reported that the Asp to Ala mutant of RPTPalpha was also non-trapping (52). These findings are in contrast with previous reports wherein the introduction of an Asp to Ala mutation to PTP1B and TC-PTP resulted in substrate-trapping mutants (38-40). Thus, this principle might not apply to all PTPs. Structural studies on the PTP domain of SHP1 bound to a phosphopeptide derived from SHPS-1 showed that the WPD loop did not close as expected (53, 54). Because SHP1 is a close structural relative of SHP2, we speculate that the Asp in the WPD loop of SHP2 may not serve as an efficient proton donor/acceptor for the hydrolysis of the thiol-ester intermediate. This observation is further supported by the finding that the D425A protein autodephosphorylated itself comparable with the WT (Fig. 1A, top panel), suggesting that the mutation did not cause loss of enzymatic activity. As demonstrated for VHR (41), it could be that the hydroxy amino acid Thr, immediately C-terminal to the Arg in the signature motif of SHP2, might be responsible for the hydrolysis of the thiol-ester intermediate.

The above findings led us to resort to improving the trapping ability of the C459S mutant. We reasoned that, although the D425A mutant of SHP2 per se is non-trapping, it might reinforce the ability of the C459S mutant. Indeed, the DM protein showed a superior trapping ability. Thus, although the Asp in the WPD loop of SHP2 does not seem to have a role in the catalytic process, it could be close enough to cause electrostatic repulsion on the phosphate moiety of the substrate. Our results are consistent with this hypothesis, because the DM protein efficiently trapped the EGFR, Gab1, and three other, as yet unidentified, phosphotyrosyl proteins (Figs. 1, 5, and 6). Xie et al. recently reported that the Asp to Ala and Cys to Ser double mutant PTP1b was non-trapping (55). Although it was not clear from the presented data or explanations given, it may simply reflect the structural and functional differences between the two proteins.

SHP2 interacts with autophosphorylated RTKs or tyrosine phosphorylated adaptor proteins (7, 8, 23, 26, 29, 32, 56, 57). Because the SH2 domains in all constructs used were intact, it is possible that some amount of EGFR bound to the DM protein was SH2-mediated. Indeed, a low amount of EGFR was detected in the WT and D425A mutants. However, compared with the amount of EGFR that coprecipitated with the DM protein, the SH2 domain-mediated interaction was extremely low and thus cannot account for the observed difference.

It could also be argued that the interaction of the DM protein with EGFR was indirect via an unknown target substrate of SHP2 that binds to EGFR. However, the partial dephosphorylation of EGFR in the control and the WT cells as compared with the C459S and the DM suggested otherwise. If the interaction was indirect or if EGFR was not a substrate of SHP2, we would expect similar levels of EGFR tyrosine phosphorylation in all experimental groups. Thus, the most likely explanation is a direct trapping of the EGFR by the PTP domain of the DM protein. The time-dependent nature of the accumulation of enzyme-substrate complexes following stimulation (Figs. 2 and 5C) suggested that the complex, once formed, is stable. Given that the interaction of the DM protein with candidate substrates occurred in intact cells following ligand stimulation and that the trapping was very specific (Figs. 1, 5, and 6), it is highly likely that EGFR and Gab1 represent physiological SHP2 substrates. Thus, the DM protein is an ideal tool for the isolation and identification of physiological substrates of SHP2 and as well for identification and characterization of specific tyrosine phosphorylation sites on substrates.

Another point worth mentioning is that the double mutation may cause structural changes that affect the specificity of the DM protein. The specificity of the SH2 domain-containing PTPs is regulated at two levels, i.e. proper cellular translocation mediated by the SH2 domains and inherent substrate discrimination mediated by the PTP domain. Because both SH2 domains are intact in all of the SHP2 proteins used, including the DM protein, proper cellular translocation to a substrate microdomain has not been compromised by the double mutation. The crystal structure of SHP2 showed that its catalytic activity is controlled by an intramolecular interaction that involves the N-SH2 and the PTP domains (58). Critical residues that mediate this interaction are Asp-61 and Glu-76 on N-SH2, and Arg465 and main-chain amine groups of the "signature motif" on the PTP domain (58, 59). Furthermore, Arg-465 and main-chain amide groups of the signature motif on the PTP domain are critical for coordinating the phosphate moiety of substrate proteins upon binding to the active site. Again, these residues are intact in the DM protein, and, thus, the double mutation is unlikely to cause either disregulation or loss of substrate specificity. The demonstration that the DM protein selectively trapped only p190 in PDGF-stimulated samples (Fig. 5D) is an excellent example for this conclusion.

In EGF-stimulated cells, the level of tyrosine phosphorylation of the EGFR was moderately higher in DM-transfected cells. Furthermore, expression of increasing amounts of the WT protein did not result in global dephosphorylation of the EGFR. Interestingly, an in vitro immunocomplex phosphatase assay on EGFR and Gab1 (Fig. 4) also provided results that mirrored those in intact cells (Figs. 3 and 6). These findings were consistent with partial dephosphorylation of the EGFR and Gab1 by SHP2. Therefore, SHP2 is a very specific PTP that only targets certain phosphorylation sites on substrate proteins. Recent reports on the DEP-1 receptor-like PTP also show that it dephosphorylates the Met receptor partially (60). However, previous work that compared the chimeras of the SH2 and PTP domains of SHP1 and SHP2 toward EGFR as a substrate concluded that SHP2 does not dephosphorylate the EGFR (61). The basis for this conclusion could be that no comparison between the WT and the phosphatase-defective mutant of SHP2 was made and that complete but not partial dephosphorylation was probably the expectation.

Important questions that remain to be answered are which tyrosine residue(s) on the EGFR are targets of SHP2 PTP activity and why? Previous work has shown that SHP2 preferentially dephosphorylated a phosphopeptide derived from Tyr-992 of the EGFR under in vitro conditions (62). Similarly, Corkscrew, the Drosophila homologue of SHP2, dephosphorylated a single autophosphorylation site on the Torso RTK that serve as a binding site for RasGAP, the down-regulator of GTP-Ras (25). Thus, it could be that SHP2 dephosphorylates a few phosphotyrosine residues on the EGFR that may serve as binding sites for negative regulatory proteins.

We also found that Gab1 is a substrate of SHP2. Gab1 is an adaptor protein with several tyrosine residues that become phosphorylated upon induction by growth factors or cytokines (30, 32-34, 56). These phosphotyrosines serve as binding sites for SH2 domain-containing signaling proteins, the most noted being SHP2 and phosphatidylinositol-3 kinase (PI3K) (30, 34, 56). The same question as for the EGFR could be asked, i.e. why does SHP2 dephosphorylate Gab1, whose tyrosine phosphorylation is essential for signal transduction? Possible explanations would again be similar, i.e. that SHP2 might be dephosphorylating negative regulatory sites on Gab1. Again, a direct answer to this question must await mutational analysis of phosphorylation sites on Gab1.

The identities of p90, p150, and p190 are currently unknown. Although the PDGFR is a plausible candidate for p190, we were unable to positively identify this protein to be the PDGFR, given the limitations of the currently available antibodies. Thus, identification of p90, p150, and p190 must be addressed in future studies that utilize protein isolation and peptide sequencing and, if necessary, cDNA cloning, nucleotide sequencing, and characterization.

The findings that the subset of SHP2 substrates varied depending on the stimulus suggested that target proteins of SHP2 PTP activity vary, depending on the pathways that become activated and/or the cell type. The ability of SHP2 to act on a different set of substrates in different signaling pathways might determine its functional specificity in mediating cell growth, differentiation, or motility.

    FOOTNOTES

* This work was supported by National Institutes of Health Public Service Grants CA28146 and CA42573 (to M. J. H.).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.

Dagger To whom correspondence should be addressed. Tel.: 631-632-8792; Fax: 631-632-8891; E-mail: mhayman@ms.cc.sunysb.edu.

Published, JBC Papers in Press, February 11, 2003, DOI 10.1074/jbc.M210670200

    ABBREVIATIONS

The abbreviations used are: RTK, receptor tyrosine kinase; PTP, phosphotyrosyl phosphatase; SHP, Src homology protein; EGF, epidermal growth factor; EGFR, EGF receptor; WT, wild-type; DM, double mutant (D425A/C459S); PDGF, platelet-derived growth factor; PDGFR, PDGF receptor; N-SH2, N-terminal Src homology 2 domain.

    REFERENCES
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES

1. Pawson, T., and Gish, G. D. (1992) Cell 71, 359-362[Medline] [Order article via Infotrieve]
2. Schlessinger, J., Mohammadi, M., Margolis, B., and Ullrich, A. (1992) Cold Spring Harbor Symp. Quant. Biol. 57, 67-74[Medline] [Order article via Infotrieve]
3. Schlessinger, J., and Ullrich, A. (1992) Neuron 9, 383-391[Medline] [Order article via Infotrieve]
4. Schlessinger, J. (1993) Harvey Lect. 89, 105-123[Medline] [Order article via Infotrieve]
5. Odaka, M., Kohda, D., Lax, I., Schlessinger, J., and Inagaki, F. (1997) J Biochem. (Tokyo) 122, 116-121[Abstract]
6. Chen, B., Bronson, R. T., Klaman, L. D., Hampton, T. G., Wang, J. F., Green, P. J., Magnuson, T., Douglas, P. S., Morgan, J. P., and Neel, B. G. (2000) Nat. Genet. 24, 296-299[CrossRef][Medline] [Order article via Infotrieve]
7. Feng, G. S., Shen, R., Heng, H. H., Tsui, L. C., Kazlauskas, A., and Pawson, T. (1994) Oncogene 9, 1545-1550[Medline] [Order article via Infotrieve]
8. Ong, S. H., Guy, G. R., Hadari, Y. R., Laks, S., Gotoh, N., Schlessinger, J., and Lax, I. (2000) Mol. Cell. Biol. 20, 979-989[Abstract/Free Full Text]
9. Adachi, M., Fischer, E. H., Ihle, J., Imai, K., Jirik, F., Neel, B., Pawson, T., Shen, S., Thomas, M., Ullrich, A., and Zhao, Z. (1996) Cell 85, 15[Medline] [Order article via Infotrieve]
10. Barford, D., Jia, Z., and Tonks, N. K. (1995) Nat. Struct. Biol. 2, 1043-1053[Medline] [Order article via Infotrieve]
11. Feng, G. S., and Pawson, T. (1994) Trends Genet. 10, 54-58[CrossRef][Medline] [Order article via Infotrieve]
12. D'Ambrosio, D., Hippen, K. L., Minskoff, S. A., Mellman, I., Pani, G., Siminovitch, K. A., and Cambier, J. C. (1995) Science 268, 293-297[Medline] [Order article via Infotrieve]
13. Bennett, A. M., Hausdorff, S. F., O'Reilly, A. M., Freeman, R. M., and Neel, B. G. (1996) Mol. Cell. Biol. 16, 1189-1202[Abstract]
14. Neel, B. G., and Tonks, N. K. (1997) Curr. Opin. Cell Biol. 9, 193-204[CrossRef][Medline] [Order article via Infotrieve]
15. Tonks, N. K., and Neel, B. G. (1996) Cell 87, 365-368[Medline] [Order article via Infotrieve]
16. Keyse, S. M. (2000) Curr. Opin. Cell Biol. 12, 186-192[CrossRef][Medline] [Order article via Infotrieve]
17. Elchebly, M., Cheng, A., and Tremblay, M. L. (2000) J. Mol. Med. 78, 473-482[CrossRef][Medline] [Order article via Infotrieve]
18. Tonks, N. K., and Neel, B. G. (2001) Curr. Opin. Cell Biol. 13, 182-195[CrossRef][Medline] [Order article via Infotrieve]
19. Saxton, T. M., Henkemeyer, M., Gasca, S., Shen, R., Rossi, D. J., Shalaby, F., Feng, G. S., and Pawson, T. (1997) EMBO J. 16, 2352-2364[Abstract/Free Full Text]
20. Saxton, T. M., Ciruna, B. G., Holmyard, D., Kulkarni, S., Harpal, K., Rossant, J., and Pawson, T. (2000) Nat. Genet. 24, 420-423[CrossRef][Medline] [Order article via Infotrieve]
21. Krautwald, S., Buscher, D., Kummer, V., Buder, S., and Baccarini, M. (1996) Mol. Cell. Biol. 16, 5955-5963[Abstract]
22. Feng, G. S., Hui, C. C., and Pawson, T. (1993) Science 259, 1607-1611[Medline] [Order article via Infotrieve]
23. Lechleider, R. J., Sugimoto, S., Bennett, A. M., Kashishian, A. S., Cooper, J. A., Shoelson, S. E., Walsh, C. T., and Neel, B. G. (1993) J. Biol. Chem. 268, 21478-21481[Abstract/Free Full Text]
24. Allard, J. D., Chang, H. C., Herbst, R., McNeill, H., and Simon, M. A. (1996) Development 122, 1137-1146[Abstract/Free Full Text]
25. Cleghon, V., Feldmann, P., Ghiglione, C., Copeland, T. D., Perrimon, N., Hughes, D. A., and Morrison, D. K. (1998) Mol. Cell 2, 719-727[Medline] [Order article via Infotrieve]
26. Frearson, J. A., and Alexander, D. R. (1998) J. Exp. Med. 187, 1417-1426[Abstract/Free Full Text]
27. Kuhne, M. R., Pawson, T., Lienhard, G. E., and Feng, G. S. (1993) J. Biol. Chem. 268, 11479-11481[Abstract/Free Full Text]
28. Myers, M. G., Jr., Wang, L. M., Sun, X. J., Zhang, Y., Yenush, L., Schlessinger, J., Pierce, J. H., and White, M. F. (1994) Mol. Cell. Biol. 14, 3577-3587[Abstract]
29. Hadari, Y. R., Kouhara, H., Lax, I., and Schlessinger, J. (1998) Mol. Cell. Biol. 18, 3966-3973[Abstract/Free Full Text]
30. Bardelli, A., Longati, P., Gramaglia, D., Stella, M. C., and Comoglio, P. M. (1997) Oncogene 15, 3103-3111[CrossRef][Medline] [Order article via Infotrieve]
31. Gale, N. W., Kaplan, S., Lowenstein, E. J., Schlessinger, J., and Bar-Sagi, D. (1993) Nature 363, 88-92[CrossRef][Medline] [Order article via Infotrieve]
32. Carlberg, K., and Rohrschneider, L. R. (1997) J. Biol. Chem. 272, 15943-15950[Abstract/Free Full Text]
33. Nishida, K., Yoshida, Y., Itoh, M., Fukada, T., Ohtani, T., Shirogane, T., Atsumi, T., Takahashi-Tezuka, M., Ishihara, K., Hibi, M., and Hirano, T. (1999) Blood 93, 1809-1816[Abstract/Free Full Text]
34. Schaeper, U., Gehring, N. H., Fuchs, K. P., Sachs, M., Kempkes, B., and Birchmeier, W. (2000) J. Cell Biol. 149, 1419-1432[Abstract/Free Full Text]
35. Fujioka, Y., Matozaki, T., Noguchi, T., Iwamatsu, A., Yamao, T., Takahashi, N., Tsuda, M., Takada, T., and Kasuga, M. (1996) Mol. Cell. Biol. 16, 6887-6899[Abstract]
36. Takada, T., Matozaki, T., Takeda, H., Fukunaga, K., Noguchi, T., Fujioka, Y., Okazaki, I., Tsuda, M., Yamao, T., Ochi, F., and Kasuga, M. (1998) J. Biol. Chem. 273, 9234-9242[Abstract/Free Full Text]
37. Ochi, F., Matozaki, T., Noguchi, T., Fujioka, Y., Yamao, T., Takada, T., Tsuda, M., Takeda, H., Fukunaga, K., Okabayashi, Y., and Kasuga, M. (1997) Biochem. Biophys. Res. Commun. 239, 483-487[CrossRef][Medline] [Order article via Infotrieve]
38. Flint, A. J., Tiganis, T., Barford, D., and Tonks, N. K. (1997) Proc. Natl. Acad. Sci. U. S. A. 94, 1680-1685[Abstract/Free Full Text]
39. Garton, A. J., Flint, A. J., and Tonks, N. K. (1996) Mol. Cell. Biol. 16, 6408-6418[Abstract]
40. Tiganis, T., Bennett, A. M., Ravichandran, K. S., and Tonks, N. K. (1998) Mol. Cell. Biol. 18, 1622-1634[Abstract/Free Full Text]
41. Denu, J. M., and Dixon, J. E. (1995) Proc. Natl. Acad. Sci. U. S. A. 92, 5910-5914[Abstract/Free Full Text]
42. Zhou, G., Denu, J. M., Wu, L., and Dixon, J. E. (1994) J. Biol. Chem. 269, 28084-28090[Abstract/Free Full Text]
43. Zhang, Z. Y., Wang, Y., and Dixon, J. E. (1994) Proc. Natl. Acad. Sci. U. S. A. 91, 1624-1627[Abstract]
44. Jia, Z., Barford, D., Flint, A. J., and Tonks, N. K. (1995) Science 268, 1754-1758[Medline] [Order article via Infotrieve]
45. Park, C. Y., and Hayman, M. J. (1999) J. Biol. Chem. 274, 7583-7590[Abstract/Free Full Text]
46. Agazie, Y., Ischenko, I., and Hayman, M. (2002) Oncogene 21, 697-707[CrossRef][Medline] [Order article via Infotrieve]
47. Klinghoffer, R. A., and Kazlauskas, A. (1995) J. Biol. Chem. 270, 22208-22217[Abstract/Free Full Text]
48. Deb, T. B., Wong, L., Salomon, D. S., Zhou, G., Dixon, J. E., Gutkind, J. S., Thompson, S. A., and Johnson, G. R. (1998) J. Biol. Chem. 273, 16643-16646[Abstract/Free Full Text]
49. O'Reilly, A. M., and Neel, B. G. (1998) Mol. Cell. Biol. 18, 161-177[Abstract/Free Full Text]
50. Tang, T. L., Freeman, R. M., Jr., O'Reilly, A. M., Neel, B. G., and Sokol, S. Y. (1995) Cell 80, 473-483[Medline] [Order article via Infotrieve]
51. Tiganis, T., Flint, A. J., Adam, S. A., and Tonks, N. K. (1997) J. Biol. Chem. 272, 21548-21557[Abstract/Free Full Text]
52. Buist, A., Blanchetot, C., Tertoolen, L. G., and den Hertog, J. (2000) J. Biol. Chem. 275, 20754-20761[Abstract/Free Full Text]
53. Yang, J., Liang, X., Niu, T., Meng, W., Zhao, Z., and Zhou, G. W. (1998) J. Biol. Chem. 273, 28199-28207[Abstract/Free Full Text]
54. Yang, J., Cheng, Z., Niu, T., Liang, X., Zhao, Z. J., and Zhou, G. W. (2000) J. Biol. Chem. 275, 4066-4071[Abstract/Free Full Text]
55. Xie, L., Zhang, Y. L., and Zhang, Z. Y. (2002) Biochemistry 41, 4032-4039[CrossRef][Medline] [Order article via Infotrieve]
56. Cunnick, J. M., Dorsey, J. F., Munoz-Antonia, T., Mei, L., and Wu, J. (2000) J. Biol. Chem. 275, 13842-13848[Abstract/Free Full Text]
57. Rotin, D., Margolis, B., Mohammadi, M., Daly, R. J., Daum, G., Li, N., Fischer, E. H., Burgess, W. H., Ullrich, A., and Schlessinger, J. (1992) EMBO J. 11, 559-567[Abstract]
58. Hof, P., Pluskey, S., Dhe-Paganon, S., Eck, M. J., and Shoelson, S. E. (1998) Cell 92, 441-450[Medline] [Order article via Infotrieve]
59. O'Reilly, A. M., Pluskey, S., Shoelson, S. E., and Neel, B. G. (2000) Mol. Cell. Biol. 20, 299-311[Abstract/Free Full Text]
60. Palka, H. L., Park, M., and Tonks, N. K. (2002) J. Biol. Chem.
61. Tomic, S., Greiser, U., Lammers, R., Kharitonenkov, A., Imyanitov, E., Ullrich, A., and Bohmer, F. D. (1995) J. Biol. Chem. 270, 21277-21284[Abstract/Free Full Text]
62. Sugimoto, S., Lechleider, R. J., Shoelson, S. E., Neel, B. G., and Walsh, C. T. (1993) J. Biol. Chem. 268, 22771-22776[Abstract/Free Full Text]


Copyright © 2003 by The American Society for Biochemistry and Molecular Biology, Inc.