Oncogene (2000) 19, 2895 ± 2903 ã 2000 Macmillan Publishers Ltd All rights reserved 0950 ± 9232/00 $15.00 www.nature.com/onc Identi®cation of Grb10 as a direct substrate for members of the Src tyrosine kinase family

Paul Langlais1, Lily Q Dong1, Derong Hu1 and Feng Liu*,1

1Department of Pharmacology and Biochemistry, The University of Texas Health Science Center at San Antonio, San Antonio, Texas, TX 78229, USA

Treatment of cells with insulin and protein tyrosine such as the IRS proteins (Cheatham and Kahn, 1995; phosphatase inhibitors such as vanadate and pervanadate White, 1998). Tyrosine phosphorylation of the IR also resulted in the tyrosine phosphorylation of Grb10, a Src generates binding sites for SH2-domain-containing homology 2 (SH2) and pleckstrin homology domain- proteins such as Shc (Cheatham and Kahn, 1995; containing adaptor protein which binds to a number of White, 1998) and Grb10 (Liu and Roth, 1998). receptor tyrosine kinases including the insulin receptor Grb10 is a SH2 and PH domain-containing protein (IR). Although Grb10 binds directly to the kinase that binds to autophosphorylated receptor tyrosine domain of the IR, our data show that Grb10 is not a kinases including the IR (Frantz et al., 1997; Hansen et direct substrate for the IR tyrosine kinase. Consistent al., 1996; Laviola et al., 1997; Liu and Roth, 1995; with this ®nding, Grb10 tyrosine phosphorylation in cells O'Neill et al., 1996), IGF-1R (Frantz et al., 1997; was inhibited by herbimycin A, a relatively speci®c Morrione et al., 1996; O'Neill et al., 1996), ELK (Stein inhibitor for members of the Src tyrosine kinase family, et al., 1996), and Ret (Durick et al., 1996; Pandey et and by the expression of dominant negative Src or Fyn. al., 1995). Grb10 also binds to receptor downstream In addition, Grb10 tyrosine phosphorylation was stimu- signaling components such as BCR ± ABL (Bai et al., lated by expression of constitutively active Src or Fyn in 1998), Raf-1 and MEK (Nantel et al., 1998), and cells and by incubation with puri®ed Src or Fyn in vitro. Nedd4 (Morrione et al., 1999). Overexpression of the The insulin stimulated or Src/Fyn-mediated tyrosine full-length Grb10 stimulates PDGF, IGF-1, and phosphorylation in vivo was signi®cantly reduced when insulin-induced DNA synthesis, suggesting Grb10 Grb10 tyrosine 67 was changed to glycine. This mutant may play an important role in growth factor-mediated form of Grb10 bound with higher anity to the IR in mitogenesis (Wang et al., 1999). The ®ndings that cells than that of the wild-type protein, suggesting that Grb10 interacts with multiple signaling molecules and tyrosine phosphorylation of Grb10 may normally that the protein undergoes tetramerization in cells negatively regulate its binding to the IR. Our data show (Dong et al., 1998) also suggest that the protein may that Grb10 is a new substrate for members of the Src function as an adaptor/sca€olding protein in various tyrosine kinase family and that the tyrosine phosphor- cellular signaling processes. Several isoforms of Grb10 ylation of the protein may play a potential role in cell have recently been identi®ed from human tissues which signaling processes mediated by these kinases. Oncogene include hGrb10a, hGrb10b, and hGrb10g. HGrb10a, (2000) 19, 2895 ± 2903. originally named Grb-IR, was identi®ed by screening a yeast two-hybrid library derived from HeLa cells using Keywords: Grb10; Src; IR; SH2 the cytoplasmic domain of the IR as a bait (Liu and Roth, 1995). This isoform di€ers from hGrb10g by a 46-amino acid deletion in the PH domain, possibly due to alternative splicing events. Both isoforms of Grb10 bind to the autophosphorylated IR and the interaction Introduction is mediated by the SH2 domain of Grb10. We (Dong et al., 1997b) and others (O'Neill et al., 1996) have shown Protein tyrosine phosphorylation plays a crucial role in that, unlike other SH2 domain-containing adaptor cell regulation and signal transduction. In cells, the proteins, Grb10 binds to the autophosphorylated extent of tyrosine phosphorylation of a protein is tyrosine residues in the activation loop of the IR. We controlled by the steady state of activities of both have also found that Grb10 is serine phosphorylated in protein tyrosine kinases and protein tyrosine phospha- cells and that this serine phosphorylation can be tases (PTPase). The shift of this steady state by growth stimulated by insulin treatment (Dong et al., 1997a). factors, cytokines, and enzyme activators/inhibitors However, whether Grb10 also undergoes growth enhances or decreases the phosphorylation state of factor-stimulated tyrosine phosphorylation remains to key signaling molecules and initiates or regulates be established. Phosphoamino acid analysis revealed various biological processes (Hunter, 1998). that mouse Grb10 was phosphorylated exclusively on The IR tyrosine kinase is activated upon insulin serine residues before and after EGF treatment (Ooi et stimulation, and the activated tyrosine kinase phos- al., 1995). On the other hand, we (Liu and Roth, 1995) phorylates various downstream signaling molecules and others (Frantz et al., 1997) have found that a phosphotyrosine-containing protein, which was immu- noprecipitated by the anti-Grb10 antibody, comigrated *Correspondence: F Liu with Grb10. Although this ®nding suggests that Grb10 Received 2 September 1999; revised 30 March 2000; accepted may be tyrosine phosphorylated in cells, it is also 4 April 2000 possible that another tyrosine phosphorylated protein, Tyrosine phosphorylation of Grb10 P Langlais et al 2896 rather than Grb10 itself, coimmunoprecipitated with Grb10. To further understand the potential roles of Grb10, we investigated the tyrosine phosphorylation of two human Grb10 isoforms, hGrb10a (Liu and Roth, 1995) and hGrb10g (Dong et al., 1997a). In this report, we present evidence that Grb10 is tyrosine phosphorylated in cells in response to the treatment of insulin and PTPase inhibitors. However, we have found that Grb10 is not a direct substrate for the IR tyrosine kinase and the tyrosine phosphorylation of Grb10 is independent of its binding to the IR in cells. We provide evidence that Grb10 is a substrate for Src and Fyn in vitro and is a potential physiological target for these kinases in cells. The identi®cation of Grb10 as a new target for Src tyrosine kinase family members may suggest a potential role for the involvement of Grb10 in various biological events mediated by these tyrosine kinases.

Results

Grb10 isoforms underwent insulin and pervanadate-stimulated tyrosine phosphorylation in cells Figure 1 Tyrosine phosphorylation of Grb10. (a) The e€ect of In our previous studies, we detected a phosphotyr- PTPase inhibitors on the tyrosine phosphorylation of cellular osine-containing protein in the immunoprecipitates proteins in CHO/IR cells. Cells were grown in 60 mm plates to from insulin-stimulated CHO/IR/Grb10a cell lysates con¯uence. After serum starvation for 1 h, cells were left using antibody against Grb10 (Liu and Roth, 1995). untreated (lane 1) or treated with 1 mM Na3VO4 (lane 2), 1 mM 78 However, whether this immunoprecipitated protein was H2O2 (lane 3), 10 M insulin (lane 4), 0.1 mM pervanadate (lane 5), and 1078 M insulin plus 0.1 mM pervanadate (lane 6). Grb10 itself, or another tyrosine phosphorylated Treatment lasted for 10 min with insulin and pervanadate and protein which might co-immunoprecipitate with 30 min with the rest. Cells were lysed in 150 ml SDS ± PAGE gel Grb10, was unclear. To test whether Grb10 was loading bu€er and total cell lysates (1.5 ml for lysate from tyrosine phosphorylated in cells, we carried out studies pervanadate-treated cells and 15 ml with the rest) were separated by SDS ± PAGE, transferred onto a nitrocellulose membrane, and in the presence of sodium orthovanadate and perva- blotted with anti-phosphotyrosine antibody. (b) Tyrosine phos- nadate, inhibitors of protein tyrosine phosphatases. As phorylation of HA-tagged human Grb10. CHO/IR/hGrb10a expected, treatment of CHO/IR cells with insulin led to (lanes 1 ± 6) or CHO/IR/hGrb10g (lanes 7 ± 12) cells grown in the tyrosine phosphorylation of several proteins, 60 mm plates were left untreated (lanes 1 and 7) or treated with 1mM Na VO (lanes 2 and 8), 1 mM H O (lanes 3 and 9), including the 95 kDa IR b-subunit and the 185 kDa 3 4 2 2 1078 M insulin (lanes 4 and 10), 1078 M insulin plus 1 mM IRS-1 (Figure 1a, lane 4). The tyrosine phosphoryla- Na3VO4 (lanes 5 and 11), or 0.1 mM pervanadate (lanes 6 and tion of cellular proteins was substantially increased in 12). Grb10 isoforms were immunoprecipitated with antibody to the presence of pervanadate, an irreversible inhibitor of the protein, separated by SDS ± PAGE, and the tyrosine PTPases (Huyer et al., 1997) (Figure 1a, lanes 5 and 6). phosphorylation of the protein was visualized with antibody to phosphotyrosine (upper panel). The same membrane was stripped However, little tyrosine phosphorylation was detected and reblotted with antibody to the HA-tag to ensure equal in lysates from non-treated cells (Figure 1a, lane 1) or loading of the protein in each lane (lower panel). MM, molecular from cells treated with vanadate or H2O2 alone (Figure mass 1a, lanes 2 and 3). To examine whether Grb10 was tyrosine phosphory- lated in cells, we treated CHO/IR/hGrb10a and CHO/ the cells were treated with insulin alone (Figure 1b, IR/hGrb10g cells with vanadate or pervanadate. In upper panels, lanes 4 and 10). On the other hand, insulin and vanadate-treated CHO/IR/hGrb10a cells, vanadate and H2O2 alone had essentially no e€ect on the anti-Grb10 antibody precipitated two phosphotyr- Grb10 tyrosine phosphorylation (Figure 1b, upper osine-containing proteins with molecular masses of 62 panel, lanes 2, 3, 8, and 9). Similar results were also and 64 kDa, respectively (Figure 1b, upper panel, lane observed in CHO, HEK293, Rat1-IR, and NIH3T3-IR 5). Stripping and reblotting of the same membrane cells expressing endogenous Grb10 (data not shown). with antibody to the HA tag revealed that the 62 kDa protein was hGrb10a (Figure 1b, lower panel, lanes 1 ± Grb10 is not a direct substrate for the IR tyrosine kinase 6). The 64 kDa protein, which was not recognized by the anti-HA antibody, co-migrated with hGrb10g To determine the mechanism of the tyrosine phosphor- (Figure 1b, upper panel, lanes 7 ± 12), suggesting that ylation, we followed the time course of insulin and it was probably the endogenous Grb10. The tyrosine vanadate-stimulated tyrosine phosphorylation of the phosphorylation of the Grb10 isoforms was further Grb10 isoforms. Treatment of CHO/IR cells with stimulated by treating cells with pervanadate (Figure insulin and vanadate (Figure 2a) or pervanadate 1b, upper panel, lanes 6 and 12). A small amount of (Figure 2b) markedly increased the phosphotyrosine tyrosine phosphorylation of Grb10 was observed when content of Grb10. Substantial insulin-stimulated tyr-

Oncogene Tyrosine phosphorylation of Grb10 P Langlais et al 2897

Figure 2 Time course of endogenous Grb10 tyrosine phosphor- ylation. (a) Insulin and Na3VO4 stimulated tyrosine phosphoryla- tion of Grb10. CHO/IR cells grown in 100 mm tissue culture plates were serum starved for 1 h at 378C and treated with 1 mM 78 Na3VO4 for 30 min. Cells were then stimulated with 10 M insulin for the indicated times. Grb10 was immunoprecipitated Figure 3 Tyrosine phosphorylation of Grb10 is independent of with antibody to the protein, separated by SDS ± PAGE, and its binding to the IR in cells. Cells expressing the IR and hGrb10a blotted to a nitrocellulose membrane. Tyrosine phosphorylation (lanes 1 ± 4), the IR and hGrb10g (lanes 5 ± 8), or the IR and R520E of Grb10 was visualized by Western blot with antibody to hGrb10g (lanes 9 ± 12) were grown in 100 mm plates to sub- phosphotyrosine (upper panel). The same membrane was stripped con¯uence, serum starved for 1 h, and left untreated or treated 78 78 and reblotted with antibody to Grb10 to document equal loading. with 10 M insulin, 1 mM Na3VO4 plus 10 M insulin, and (b) Pervanadate stimulated tyrosine phosphorylation of Grb10 0.1 mM pervanadate as indicated. Grb10 proteins were immuno- isoforms. CHO/IR cells grown in 100 mm tissue culture plates precipitated with antibody to the HA-tag, separated by SDS ± were serum starved for 1 h at 378C and treated with pervanadate PAGE, and transferred to a nitrocellulose membrane. The for the indicated times. The tyrosine phosphorylated Grb10 was tyrosine phosphorylation of Grb10 was visualized by Western immunoprecipitated by antibody to Grb10, separated by SDS ± blot with antibody to phosphotyrosine (a). The same membrane PAGE, blotted to a nitrocellulose membrane, and detected with was re-blotted with antibody to the b-subunit of the insulin antibody to phosphotyrosine (upper panel) or to Grb10 (lower receptor (b) or to Grb10 (c). MM, molecular mass panel)

osine phosphorylation could be detected as early as ylation of the IR and the Grb10 isoforms (Figure 3a, 5 min. Similar ®ndings were also observed for hGrb10 lanes 2 and 6). The tyrosine phosphorylation of isoforms overexpressed in CHO/IR cells (data not Grb10 isoforms was further enhanced in the presence shown). of the PTPase inhibitors vanadate (Figure 3a, lanes 3, Because Grb10 binds to the kinase domain of the 7 and 11) or pervanadate (Figure 3a, lanes 4, 8 and IR (Dong et al., 1997b; O'Neill et al., 1996), we were 12). In addition to the 95 kDa b-subunit of the IR, interested in determining whether Grb10 was a direct we also detected several other phosphotyrosine- substrate of the receptor tyrosine kinase. First we containing proteins of 60, 135, 185, and 200 kDa, examined if Grb10 could be phosphorylated by respectively, in anti-HA immunoprecipitates from puri®ed IR in vitro. The IR was puri®ed from lysates of the pervanadate-treated cells (Figure 3a, CHO/IR cells by wheat germ agglutinin-gel absorp- lanes 4, and 8). These tyrosine phosphorylated tion and activated in vitro according to a previously proteins could be pulled-down by the SH2 domain described protocol (Zhang et al., 1991). Phosphoryla- of Grb10 fused with GST (GST/Grb10-SH2) (data tion of the His-tagged hGrb10g by the activated IR not shown), suggesting that the interaction may be was performed in the presence of divalent metal ions mediated by the phosphotyrosine residues on these and 32P-g-ATP. As a control, we also tested the proteins and the SH2 domain of Grb10. Western blot phosphorylation of His-tagged p62dok (Carpino et al., revealed that the 185 kDa protein was IRS-1 and the 1997; Yamanashi and Baltimore, 1997), a p21ras GAP- 95 kDa protein was the b-subunit of the IR (data not associated protein that is tyrosine phosphorylated shown and Figure 3b), respectively. However, the after insulin stimulation (Hosomi et al., 1994). While identities of the other tyrosine phosphorylated incubation of the activated IR with p62dok bound to proteins are currently unknown. Nevertheless, the Ni-NTA-agarose beads resulted in a marked increase mutant hGrb10g, which could no longer interact with in the tyrosine phosphorylation of the protein, no the IR, was still tyrosine phosphorylated in response notable tyrosine phosphorylation of Grb10 was to treatment with vanadate and insulin (Figure 3a, observed under the same condition (data not shown). lane 11) or pervanadate alone (Figure 3a, lane 12). To further characterize Grb10 tyrosine phosphoryla- Taken together, these ®ndings suggest that Grb10 tion, we expressed a mutant hGrb10g in which a tyrosine phosphorylation was probably mediated by critical arginine residue in the SH2 domain was an IR downstream kinase(s) rather than by the IR changed to glutamate (hGrb10gR520E). This mutant of tyrosine kinase itself. It is interesting to note that a hGrb10 did not bind the IR in the yeast two-hybrid smaller amount of the IR was co-immunoprecipitated system (Dong et al., 1997b) or in mammalian cells in the pervanadate-treated cells (Figure 3b, lanes 4 (Figure 3b, lanes 9 ± 12). Treatment of cells expressing and 8), suggesting that either tyrosine phosphoryla- the IR and the wild-type hGrb10a (Figure 3, lanes tion of Grb10 or excessive tyrosine phosphorylation 1 ± 4) or hGrb10g (Figure 3, lanes 5 ± 8) with insulin of the IR may negatively regulate the interaction resulted in a notable increase in tyrosine phosphor- between these two proteins.

Oncogene Tyrosine phosphorylation of Grb10 P Langlais et al 2898 Tyrosine phosphorylation of Grb10 by members of the Src tyrosine kinase family, Src and Fyn To identify the tyrosine kinase(s) that phosphorylated Grb10, we examined whether herbimycin A, a selective inhibitor for the Src tyrosine kinase family (Uehara et al., 1985), blocked insulin-stimulated Grb10 tyrosine phosphorylation. As shown in Figure 4a, in the absence of herbimycin A, insulin and vanadate stimulation resulted in the tyrosine phosphorylation of Grb10 (lane 1). However, the insulin and vanadate- stimulated tyrosine phosphorylation was signi®cantly inhibited when cells were pretreated with herbimycin A for 3 h (Figure 4a, lane 2) and was completely blocked after pretreatment of the cells for 24 h (Figure 4a, lane 3). Herbimycin treatment also led to a decrease in pervanadate-stimulated tyrosine phosphorylation of Grb10g in CHO/IR cells (Figure 4b). To further test whether Src was involved in Grb10 tyrosine phosphorylation, we transiently co-transfected CHO/IR cells with hGrb10g and constitutively active (SrcY527F) or kinase defective (SrcK7) mutants of Src. Consistent with our earlier ®ndings (Figures 1 ± 3), expressing hGrb10g with insulin and Na3VO4 resulted in an increase in the tyrosine phosphorylation of the protein (Figure 4c, lane 2). Overexpression of a constitutively active form of Src kinase resulted in a marked increase in Grb10 tyrosine phosphorylation, which was independent of insulin treatment (Figure 4c, lanes 3 and 4). On the other hand, the insulin- stimulated tyrosine phosphorylation of Grb10 was almost completely blocked by overexpression of the dominant negative form of Src kinase (SrcK7) (Figure

cells. CHO/IR cells were transiently transfected with cDNA encoding hGrb10g (lanes 1 and 2) or cotransfected with cDNAs encoding both hGrb10g and constitutively active (SrcY527F) (lanes 3 and 4) or dominant negative (SrcK7) (lanes 5 and 6) mutants of Src. Two days after transfection, cells were serum-starved for 1 h, treated with 1 mM Na3VO4 for 20 min and then stimulated with (+) or without (7)1078 M insulin for an additional 10 min. After immunoprecipitation with antibody to the HA-tag, the tyrosine phosphorylated Grb10 was detected by Western blot with anti-phosphotyrosine antibody (upper panel). The expression of Figure 4 Identi®cation of Grb10 as a potential substrate for Src Grb10 or Src in these cells was detected by Western blot using tyrosine kinase in cells. (a) Inhibition of insulin and vanadate- antibodies to the HA-tag (middle panel) or to Src (bottom panel), stimulated Grb10 tyrosine phosphorylation by herbimycin A. respectively. (d) The e€ect of constitutively active or dominant CHO/IR/hGrb10g cells were grown in 100 mm plates to sub- negative Fyn on Grb10 tyrosine phosphorylation in cells. CHO/ con¯uence, serum-starved, and left untreated (lane 1) or treated IR cells were transiently transfected with cDNA encoding with 3 mM herbimycin A for 3 h (lane 2) or 24 h (lane 3). Cells hGrb10g (lanes 1 and 2) or with cDNAs encoding both hGrb10g Y531F were then treated with 1 mM Na3VO4 for 20 min and then and constitutively active (Fyn ) (lanes 3 and 4) or dominant stimulated with 1078 M insulin for an additional 10 min. hGrb10g negative (FynK7) (lanes 5 and 6) mutants of Fyn. Cells were was immunoprecipitated with antibody to the HA-tag, separated treated with 1 mM Na3VO4 and with (+) or without (7)10nM by SDS ± PAGE, and transferred to a nitrocellulose membrane. insulin as described in (c). The tyrosine phosphorylated Grb10 The tyrosine phosphorylated hGrb10g was visualized by Western was detected by Western blot with anti-phosphotyrosine antibody blot with antibody to phosphotyrosine. The same membrane was (upper panel). The expression of Grb10 or Fyn in these cells was re-blotted with antibody to the HA-tag to ensure equal loading of detected by Western blot using antibodies to the HA-tag (middle the protein (lower panel). (b) Inhibition of pervanadate stimulated panel) or to Fyn (bottom panel), respectively. (e) Inhibition of tyrosine phosphorylation of hGrb10 by herbimycin A. CHO/IR/ SrcY527F and FynY531F induced tyrosine phosphorylation of hGrb10g cells were grown in 100 mm plates to sub-con¯uence, hGrb10g by herbimycin A. CHO/IR cells were transiently serum starved, and left untreated (lane 1) or treated with 3 mM cotransfected with cDNAs encoding hGrb10g and either SrcY527F herbimycin A for 3 h (lane 2). Cells were then treated with or FynY531F. Forty-eight hours after transfection, cells were serum 0.1 mM pervanadate for 10 min. hGrb10g was immunoprecipi- starved for 3 h in the presence or absence of 3 mM herbimycin A. tated with antibody to the HA-tag, separated by SDS ± PAGE, After treatment, cells were lysed in Bu€er A and hGrb10g was and transferred to a nitrocellulose membrane. The tyrosine immunoprecipitated with antibody to the HA-tag. One-third of phosphorylated hGrb10g was visualized by Western blot with the immunoprecipitates were separated by SDS ± PAGE and antibody to phosphotyrosine (upper panel). The same membrane analysed by Western blot with anti-phosphotyrosine antibody was re-blotted with antibody to the HA-tag to ensure equal (top panel). One-third of the immunoprecipitates were used for loading (lower panel). (c) The e€ect of constitutively active or Western blot with a monoclonal antibody to the HA-tag (bottom dominant negative Src on Grb10 tyrosine phosphorylation in panel)

Oncogene Tyrosine phosphorylation of Grb10 P Langlais et al 2899

Figure 5 Identi®cation of Grb10 as a direct substrate of Src. (a) Time course study of Grb10 phosphorylation by Src in vitro. Figure 6 Phosphorylation of Grb10 on Tyrosine 67. (a) GST-hGrb10g fusion protein bound to glutathione agarose beads Schematic of hGrb10g showing the proline-rich sequence (residues was washed once with 50 mM HEPES, pH 7.6, and twice with 136 ± 143, indicated as a hatched box), the pH and the SH2 Bu€er C. In vitro kinase assays were initiated by adding 30 ml domains. (b) Tyrosine phosphorylation of His-hGrb10g and His- 32 Bu€er C containing 25 mM ATP, 2 mCi P-g-ATP, and 1 unit of hGrb10gDC by Src in vitro. His-tagged hGrb10g and hGrb10gDC Src. Reactions were carried out at 308C for times indicated. proteins conjugated to Ni-NTA-agarose beads were washed twice Reactions were stopped by adding SDS loading bu€er and boiling with Bu€er B and twice with Bu€er C. In vitro kinase assays were at 958C for 3 min. Samples were separated by SDS ± PAGE, carried out as described in the legend of Figure 5a. After transferred to a nitrocellulose membrane, and phosphorylation separation by SDS ± PAGE, the phosphorylated Grb10 proteins was visualized by autoradiography (insert). The incorporation of were visualized by autoradiography (top panel) and the protein 32 P into the GST-hGrb10g fusion protein was quanti®ed by level was examined by Coomassie blue staining (bottom panel). PhosphorImager analysis. (b) Phosphoamino analysis of in vitro (c) Tyrosine phosphorylation of GST-hGrb10g and GST- 32 phosphorylated Grb10. The P-labeled hGrb10g bands were hGrb10gY67G by Src and Fyn in vitro. hGrb10g-or excised from the membrane, hydrolyzed and analysed by two- hGrb10gY67G-GST fusion proteins conjugated to glutathione dimensional thin-layer electrophoresis and autoradiography beads were phosphorylated as described in the legends of Figure 5a. A similar protocol was employed for Fyn-catalyzed Grb10 phosphorylation, except Bu€er D was used. After separation by SDS ± PAGE, the gels were stained with Coomassie blue (bottom 4c, lanes 5 and 6). Similar results were also obtained panel) and the phosphorylation of Grb10 was visualized by autoradiography (top panel) for cells overexpressing a constitutively active (Figure 4d, lanes 3 and 4) or a dominant negative Fyn mutant (Figure 4d, lanes 5 and 6). The Src-or Fyn-stimulated tyrosine phosphorylation of Grb10g was also reduced by treatment of cells with herbimycin A (Figure 4e). phosphorylation site(s) on the protein. To narrow These ®ndings provide further evidence that the down the region of the tyrosine phosphorylation, we insulin-stimulated tyrosine phosphorylation of Grb10 compared the phosphorylation of the full-length and a is probably mediated by members of the Src tyrosine carboxyl terminal truncated form of Grb10. This kinase family such as Src and Fyn. mutant form of Grb10 had a 180 amino acid-deletion To determine whether Grb10 was a direct substrate in the carboxyl terminus, including the SH2 domain for members of the Src kinase family, we carried out in (Figure 6a). We found that Src phosphorylated this vitro phosphorylation studies using puri®ed Src and mutant form of Grb10 (hGrb10gDC) as eciently as Grb10. As shown in Figure 5a, incubation of GST- the full-length protein (Figure 6b), suggesting that the hGrb10g bound to glutathione agarose beads with Src tyrosine phosphorylation site(s) may be located at the resulted in a time-dependent tyrosine phosphorylation amino terminus. Analysis of the amino acid sequence of hGrb10g. Phosphoamino acid analysis indicated that of hGrb10gDC revealed the presence of 12 tyrosine the phosphorylation was entirely on tyrosine residues residues and one of them, Tyr67, appeared to be a (Figure 5b). Similar results were obtained when the in potential candidate for tyrosine kinases. This residue is vitro phosphorylation was carried out using puri®ed preceded by a hydrophobic leucine at the 71 and a Fyn tyrosine kinase (data not shown). negatively charged glutamate at the 73 positions, respectively (Figure 6a), which constitute an optimal substrate sequence for members of the Src tyrosine Identification of tyrosine 67 in Grb10 as the kinase family (Ruzzene et al., 1997; Songyang and Src/Fyn-catalyzed phosphorylation site Cantley, 1995). Quite recently, this residue was shown Having found that Grb10 was a direct substrate for Src to be tyrosine phosphorylated by Tec tyrosine kinase and Fyn, we then attempted to identify the tyrosine (Mano et al., 1998). To test whether Tyr67 in hGrb10g

Oncogene Tyrosine phosphorylation of Grb10 P Langlais et al 2900 was the site of phosphorylation by Src or Fyn, we ®ndings provide further evidence that tyrosine phos- replaced it with a glycine by site-directed mutagenesis. phorylation of Grb10 inhibits the binding of the The mutant form of Grb10 was expressed in bacterial protein to the IR. cells as a GST-fusion protein, puri®ed by glutathione agarose beads, and tested for phosphorylation by Src and Fyn kinases in vitro. As shown in Figure 6c, while the wild-type Grb10 was readily phosphorylated by both Src and Fyn tyrosine kinases, no tyrosine phosphorylation could be detected for the GST- Grb10Y67G mutant, suggesting that Tyr67 was the Src/ Fyn-catalyzed phosphorylation site in Grb10. To con®rm that Tyr67 was the tyrosine phosphoryla- tion site in vivo, we subcloned the cDNA encoding the mutant of hGrb10g into a mammalian expression vector (pBEX) (Bram et al., 1993) and transiently expressed the HA-tagged wild-type or the mutant proteins in CHO/IR cells. Replacing tyrosine at position 67 with glycine did not signi®cantly alter the overall structure of the protein since the insulin- stimulated gel mobility shift, which was caused mainly by wortmannin and PD98059-sensitive serine phos- phorylation (Dong et al., 1997a), was similar to that of the wild-type protein (data not shown). Cells expres- sing the wild-type or the mutant hGrb10g (hGrb10gY67G) were treated with or without insulin, vanadate, or pervanadate and the tyrosine phosphory- lated Grb10 was examined by anti-phosphotyrosine blot of anti-HA immunoprecipitates. As shown in Figure 7a, treatment of cells with insulin and vanadate (lane 5) or pervanadate (lane 7) resulted in a signi®cant increase in tyrosine phosphorylation of the wild-type Grb10. The insulin/vanadate- (Figure 7a, lanes 5 and 6), Src- (Figure 7b, lanes 1 and 2), or Fyn- (Figure 7b, lanes 3 and 4) stimulated tyrosine phosphorylation, however, was notably decreased for the hGrb10gY67G mutant. The pervanadate-induced tyrosine phosphor- ylation of hGrb10gY67G was also decreased signi®cantly (Figure 7a, lane 8). However, residual tyrosine phosphorylation could still be observed for this Figure 7 Tyrosine phosphorylation of Grb10 inhibits the mutant, indicating that additional site(s) may be binding of the protein to the IR. (a) Tyrosine Phosphorylation tyrosine phosphorylated in response to pervanadate- of hGrb10g and hGrb10gY67G in vivo. CHO/IR cells were transiently transfected with plasmids encoding hGrb10g or treatment. hGrb10gY67G. Forty-eight hours after transfection, cells were left untreated (lanes 1 and 2) or treated with 1078 M insulin (lanes 3 78 and 4), 1 mM Na3VO4 plus 10 M insulin (lanes 5 and 6), or Tyrosine phosphorylation of Grb10 inhibits the binding of 0.1 mM pervanadate (lanes 7 and 8). The treatment was 30 min the protein to the IR for vanadate and 10 min for the rest. After treatment, cells were lysed in Bu€er A and both hGrb10g and hGrb10gY67G were As mentioned earlier, pervanadate-stimulated tyrosine immunoprecipitated with antibody to the HA-tag. The proteins phosphorylation resulted in a decrease in the amount were separated by SDS ± PAGE and transferred to a nitrocellulose of IR co-immunoprecipitated with Grb10 (Figure 3b, membrane. Tyrosine phosphorylation of hGrb10g or hGrb10gY67G lanes 4 and 8). This ®nding suggested that either was visualized with antibody to phosphotyrosine (upper panel). The same membrane was stripped and reblotted with antibody to tyrosine phosphorylation of Grb10 or excessive tyr- the HA-tag to ensure equal loading of the protein (lower panel). osine phosphorylation of the IR may a€ect the binding (b) Mutation of tyrosine 67 decreased SrcY527F or FynY531F- of the adaptor protein to the receptor. To further test induced tyrosine phosphorylation of hGrb10g. CHO/IR cells were this idea, we examined the interaction between the IR transiently cotransfected with hGrb10g or hGrb10gY67G together with either SrcY527F or FynY531F. Forty-eight hours after and hGrb10g or the hGrb10gY67G mutant. Consistent transfection, cells were serum starved for 1 h. Cells were then with our previous ®ndings, insulin-stimulation of lysed in Bu€er A and both hGrb10g and hGrb10gY67G were CHO/IR cells expressing the wild-type hGrb10g immunoprecipitated with antibody to the HA-tag. One-third of resulted in a signi®cant increase of the tyrosine the immunoprecipitates were separated by SDS ± PAGE and phosphorylated IR, which could be co-immunopreci- analysed by Western blot with anti-phosphotyrosine antibody (top panel). One-third of the immunoprecipitates were used for pitated with hGrb10g (Figure 7c, lane 3). The insulin- Western blot with a monoclonal antibody to the HA-tag (bottom stimulated IR/Grb10 interaction was enhanced in cells panel). (c) Association of the IR with the wild-type or the tyrosine overexpressing hGrb10gY67G (Figure 7c, lane 4). The phosphorylation site mutant of Grb10. In the same experiment enhanced interaction between the IR and the mutant described in (a), the tyrosine phosphorylated IR co-immunopre- Y67G Grb10 was also observed in cells treated with insulin cipitated with hGrb10g (lanes 3, 5 and 7) or hGrb10g (lanes 4, 6 and 8) was detected with antibody to phosphotyrosine (upper plus vanadate (Figure 7c, lanes 5 and 6) or panel). The same membrane was stripped and reblotted with pervanadate alone (Figure 7c, lanes 7 and 8). These antibody to the b-subunit of the IR (lower panel)

Oncogene Tyrosine phosphorylation of Grb10 P Langlais et al 2901 Discussion (McGlade et al., 1992) and the p85 subunit of PI- 3kinase (Fukui and Hanafusa, 1991; Haefner et al., In this report we demonstrate that Grb10 is tyrosine 1995; Liu et al., 1993). It has also been found that Shc phosphorylated in cells in response to insulin stimula- is able to interact with the SH3 domain of Src through tion. Although the adapter protein binds directly to its proline rich sequence (Weng et al., 1994). Recently, the kinase domain of the IR (Dong et al., 1997b), the one of the members of the Src tyrosine kinase family, present results suggest that tyrosine phosphorylation Fyn, has been shown to interact with IRS-1 (Sun et al., of Grb10 is mediated by a downstream kinase(s), 1996). The involvement of Src family kinases in insulin rather than the IR tyrosine kinase itself. We present and IGF-1 signaling have also been implicated by the evidence that members of the Src tyrosine kinase ®nding that the b-subunit of the IGF-1 receptor is family, such as Src and Fyn, may be potential kinases tyrosine phosphorylated by Src (Peterson et al., 1996). that phosphorylate Grb10. First, Grb10 tyrosine These ®ndings suggest that the Src tyrosine kinase phosphorylation was stimulated by overexpression of family members may play roles in modulating a constitutively active Src or Fyn in cells (Figure molecules involved in the IR signal transduction 4c,d). In addition, the phosphorylation was inhibited pathway. by herbimycin A (Figure 4a,b,e), a relatively speci®c The binding of Grb2-Sos to tyrosine phosphorylated inhibitor of Src tyrosine kinase family members Shc is thought to couple the IR to Ras activation (Uehara et al., 1985), and by overexpression of a (Pronk et al., 1993; Skolnik et al., 1993). In our study, dominant negative mutant of Src or Fyn (Figure we detected no interaction between tyrosine phos- 4c,d). Furthermore, Grb10 was directly phosphory- phorylated Grb10 and Grb2 (data not shown). In lated by the puri®ed Src and Fyn in vitro (Figure 5). addition, we found that the tyrosine phosphorylation Finally, replacement of Tyr67 with glycine in Grb10, was independent of the binding of Grb10 to the IR and which abrogated the Src/Fyn-catalyzed phosphoryla- that Grb10 was not a direct substrate of the IR tion of Grb10 in vitro (Figure 6c), also reduced the tyrosine kinase. Although Grb10 contains a proline- insulin-stimulated tyrosine phosphorylation in cells rich sequence that is conserved among the Grb7/ (Figure 7a,b). However, while our data suggest that Grb10/Grb14 family members (Liu and Roth, 1998), Grb10 is a potential target for Src and Fyn, these this sequence does not bind to the SH3 domain of Fyn, ®ndings do not exclude the possibility that other p85, or Grb2 (Frantz et al., 1997). members of the Src tyrosine kinase family may also In summary, we have shown that the adaptor phosphorylate Grb10 in cells. protein Grb10 is tyrosine phosphorylated in cells and In the present study, we provide further evidence identi®ed Src and/or Fyn as two potential kinases that that the interaction between Grb10 and the IR is may catalyze the phosphorylation. The identi®cation of dependent on the SH2 domain of the protein and the Grb10 as a new substrate for members of the Src autophosphorylated tyrosine residues in the receptor tyrosine kinase family may suggest a potential role for (Figure 3). Although we have previously shown that the involvement of Grb10 in these kinase-mediated the carboxyl terminus of Grb10, which contains the downstream biological events. However, since all these SH2 domain, was sucient to bind to the autopho- studies were carried out in CHO cells overexpressing sphorylated IR (Liu and Roth, 1995), there is evidence the insulin receptor, it remains to be established suggesting that the insert between the SH2 and the PH whether insulin is a physiological activator of Grb10 domains (BPS) may also be involved in the binding of tyrosine phosphorylation in insulin target cells such as the protein to the receptor (He et al., 1998). However, skeletal muscle and adipocytes that express normal it is unknown whether this region is sucient to bind levels of the endogenous insulin receptor. the IR. In this study, we have found that a single point mutation in the SH2 domain was sucient to abolish the interaction between Grb10 and the IR (Figure 3), suggesting that the speci®city between Grb10 and the Materials and methods IR is mainly determined by the interaction between the SH2 domain in Grb10 and the phosphotyrosine Materials residues in the IR. However, it is possible that an additional binding site in Grb10 may be induced after All chemicals were purchased from Sigma unless otherwise indicated. Pervanadate was freshly prepared by mixing equal the SH2 domain binds to the receptor. Thus, the volumes of 0.1 M H2O2 and 0.1 M Na3VO4 and incubating binding of the second region of Grb10 to the IR may for 10 min at room temperature. Monoclonal antibody to the provide additional speci®city for the interaction. HA-tag was from BABCO. Monoclonal antibody to Fyn, The Src tyrosine kinase family consists of at least polyclonal antibodies to Src and the b-subunit of the IR were nine members and have been shown to undergo growth from Santa Cruz Biotechnology, Inc., (Santa Cruz, CA, factor stimulated activation in cells (Brown and USA). Antibody to phosphotyrosine was from Transduction Cooper, 1996). This group of related tyrosine kinases Laboratories (Lexington, KY, USA). Polyclonal antibody to play important roles in a variety of cellular processes, Grb10 was raised in rabbits against the C-terminus of Grb10 including cellular DNA synthesis, proliferation, di€er- (amino acids 447 to 548 in hGrb10a) fused to GST and entiation, and neuronal signaling (Brown and Cooper, anity puri®ed as described previously (Dong et al., 1997a; 1996). Most of the family members are expressed Liu and Roth, 1995). Mammalian expression vectors contain- ing the constitutively active Src and Fyn (SrcY527F, FynY531F) predominantly in hematopoietic cells and several of or dominant negative Src and Fyn (SrcK7, FynK7) kinases them, including Src and Fyn, are more ubiquitously were generous gifts of Dr SA Courtneidge (Sugen, Inc, CA, expressed (Courtneidge et al., 1993). Several compo- USA) and have been described previously (Twamley-Sein et nents of the IR signaling pathway have been shown to al., 1993). Puri®ed Src and Fyn tyrosine kinases were from be substrates of Src, including the adaptor protein Shc Upstate Biotechnology, Inc. (Lake Placid, NY, USA).

Oncogene Tyrosine phosphorylation of Grb10 P Langlais et al 2902 extracts were centrifuged 14 000 g for 10 min at 48C. Buffers Epitope-tagged or non-tagged proteins were immunoprecipi- Bu€er A consisted of 50 mM HEPES, pH 7.6, 1% Triton X- tated by incubation for 4 h at 48C with antibodies to the HA- 100, 150 mM NaCl, 1 mM phenylmethanesulfonyl ¯uoride, tag or to the proteins bound to protein G/A Sepharose beads 1mM Na3VO4,1mM NaF, 1 mM sodium pyrophosphate, (Life Technologies, Inc.). The immunoprecipitates were 20 mM b-glycerophosphate, 10 mg/ml aprotinin and 10 mg/ml washed three times with bu€er B. Proteins were eluted by leupeptin. Bu€er B consisted 50 mM HEPES, pH 7.6, heating at 958C for 3 min in SDS ± PAGE sample loading 150 mM NaCl and 0.1% Triton X-100. Bu€er C consisted bu€er, separated by SDS ± PAGE, transferred to a nitrocel- of 50 mM Tris-HCl, pH 7.4, 25 mM MgCl2,5mM MnCl2, lulose membrane, and probed with antibodies to either 1mM EGTA, 0.5 mM Na3VO4 and 0.5 mM dithiothreitol. phosphotyrosine, the tag, or the proteins. Immunocomplexes Bu€er D consisted of 50 mM Tris-HCl, pH 7.2, 25 mM were detected using chemiluminescence or alkaline phospha- MgCl2,5mM MnCl2,1mM EGTA, 0.5 mM Na3VO4, and tase conjugated reactions. All immunoprecipitation experi- 0.5 mM dithiothreitol. ments were repeated independently three to ®ve times to ensure the consistency of the results. Cell lines In Vitro phosphorylation of Grb10 by Src and Fyn tyrosine Chinese hamster ovary (CHO) cells expressing the human IR kinases (CHO/IR) and di€erent isoforms of hGrb10 (CHO/IR/ hGrb10a and CHO/IR/hGrb10g) have been described Phosphorylation of Grb10 by Src was initiated by addition of previously (Dong et al., 1997a; Liu and Roth, 1995) and the puri®ed Src tyrosine kinase (1 unit) to a tube containing were maintained in Ham's F-12 medium supplemented with 30 ml Bu€er C, 25 mM ATP, 2 mCi of [g-32P]ATP and 5 mgof 10% new born calf serum. To establish cell lines expressing GST, GST-hGrb10g, or GST-hGrb10gY67G coupled to the IR and an SH2-domain mutant Grb10 (hGrb10gR520E), we glutathione agarose beads (Sigma). After incubation at subcloned the cDNA encoding this mutant (Dong et al., 308C for 30 min, the beads were washed extensively with 1997b) into the mammalian expression vector pBEX (Bram et Bu€er B and the phosphorylated hGrb10g was separated by al., 1993) in-frame with a sequence encoding a 9-amino acid SDS ± PAGE. The phosphorylation of Grb10 was then hemaglutinin (HA)-tag (YPYDVPDYA) to generate the examined by autoradiography and by PhosphorImager recombinant plasmid pBEX/hGrb10gR520E. Transfection of analysis. Phosphorylation of Grb10 by Fyn kinase in vitro CHO/IR cells with plasmids pBEX/hGrb10gR520E and was performed using a similar protocol described above, pBSpacDp (de la Luna et al., 1988) were carried out by except that bu€er D was used. Phosphoamino acid analysis electroporation. Stable cell lines expressing the SH2 domain was carried out according to a protocol described previously mutant hGrb10g (CHO/IR/hGrb10gR520E) were selected with (Dong et al., 1997a). All experiments were repeated 8 mg/ml puromycin. Positives were identi®ed by Western independently two to four times to ensure the consistency blotting using antibody to the HA-tag or to Grb10 (Liu and of the results. Roth, 1995) and cloned by limiting dilution.

Site-directed mutagenesis and plasmid construction Single-stranded DNA site-directed mutagenesis was carried out to replace tyrosine 67 of Grb10 with glycine, using the full-length hGrb10g cDNA as a template (Dong et al., Abbreviations 1997a). The cDNA encoding the mutant hGrb10g was then The abbreviations used are: CHO, Chinese hamster ovary; ampli®ed by PCR and subcloned into the plasmid pGEX-4T- GST, glutathione S-transferase; IGF-I, insulin-like growth 1 (Pharmacia) or into pBEX (Bram et al., 1993). The factor-I; IR, insulin receptor; IRS-1, IR substrate-1; expression and puri®cation of GST-Grb10 fusion proteins PAGE, polyacrylamide gel electrophoresis; PH, plekstrin- were carried out according to a protocol described previously homology; PI 3-kinase, phosphatidylinositol 3-kinase; (Dong et al., 1998). The expression and puri®cation of His- PTPase, protein tyrosine phosphatase; SH2, Src homology tagged full-length or a carboxyl terminal truncated form 2. (hGrb10gDC) of hGrb10g have been described previously (Dong et al., 1998). Expression of the protein in DH5a cells was induced by the addition of 1 mM isopropyl-b-D- thiogalactoside and the protein was anity-puri®ed by Ni- Acknowledgments NTA agarose beads (Qiagen, CA, USA). We thank Dr SA Courtneidge for mammalian expression vectors encoding constitutively active and dominant nega- tive Src and Fyn. This research was supported by National Immunoprecipitation and Western blot Institute of Health Grant DK52933 (to F Liu and LQ Cells grown in tissue culture plates were treated with or Dong) and a research grant from the American Heart without various stimuli and solubilized with bu€er A. The Association, Texas Aliate (To LQ Dong).

References

Bai RY, Jahn T, Schrem S, Munzert G, Weidner KM, Wang Courtneidge SA, Fumagalli S, Koegl M, Superti-Furga S and JYJ and Duyster J. (1998). Oncogene, 17, 941 ± 948. Twamley-Stein GM. (1993). Develop. (suppl.), 57 ± 64. Bram RJ, Hung DT, Martin PK, Schreiber SL and Crabtree de la Luna S, Soria I, Pulido D, Ortin J and Jimenez A. GR. (1993). Mol. Cell. Biol., 13, 4760 ± 4769. (1988). Gene, 62, 121 ± 126. Brown MT and Cooper JA. (1996). Biochem. Biophys. Acta, Dong LQ, Du H-Y, Porter S, Kolakowski J, LF, Lee AV, 1287, 121 ± 149. Mandarino LJ, Fan JB, Yee D and Liu F. (1997a). J. Biol. Carpino N, Wisniewski D, Strife A, Marshak D, Kobayashi Chem., 272, 29104 ± 29112. R, Stillman B and Clarkson B. (1997). Cell, 88, 197 ± 204. DongLQ,FarrisS,ChristalJandLiuF.(1997b).Mol. Cheatham B and Kahn CR. (1995). Endocrine Rev., 16, 117 ± Endocrinol., 11, 1757 ± 1765. 142.

Oncogene Tyrosine phosphorylation of Grb10 P Langlais et al 2903 Dong LQ, Porter S, Hu D and Liu F. (1998). J. Biol. Chem., Nantel A, Mohammad-Ali K, Sherk J, Posner BI and 273, 17720 ± 17725. Thomas DY. (1998). J. Biol. Chem., 273, 10475 ± 10484. Durick K, Wu R-Y, Gill GN and Taylor SS. (1996). J. Biol. O'Neill TJ, Rose TW, Pillay TS, Hotta K, Olefsky JM and Chem., 271, 12691 ± 12694. Gustafson TA. (1996). J. Biol. Chem., 271, 22506 ± 22513. Frantz JD, Giorgetti-Peraldi S, Ottinger EA and Schoelson OoiJ,YajnikV,ImmanuelD,GordonM,MoskowJJ, SE. (1997). J. Biol. Chem., 272, 2659 ± 2667. Buchberg AM and Margolis B. (1995). Oncogene, 10, Fukui Y and Hanafusa H. (1991). Mol. Cell. Biol., 11, 1972 ± 1610 ± 1630. 1979. Pandey A, Duan H, Di Fiore PP and Dixit VM. (1995). J. Haefner B, Baxter R, Fincham VJ, Downes CP and Frame Biol. Chem., 270, 21461 ± 21463. MC. (1995). J. Biol. Chem., 270, 7937 ± 7943. Peterson JE, Kulik G, Jelinek T, Reuter CWM, Shannon JA Hansen H, Svensson U, Zhu J, Laviola L, Giorgino F, Wolf and Weber MJ. (1996). J. Biol. Chem., 271, 31562 ± 31571. G, Smith RJ and Riedel H. (1996). J. Biol. Chem., 271, Pronk GJ, McGlade I, Pelicci G, Pawson T and Bos JL. 8882 ± 8886. (1993). J. Biol. Chem., 268, 5748 ± 5753. He W, Rose DW, Olefsky JM and Gustafson TA. (1998). J. Ruzzene M, Songyang Z, Marin O, Donella-Deana A, Biol. Chem., 273, 6860 ± 6867. Brunati AM, Guerra B, Agostinis P, Cantley LC and Hosomi H, Shii K, Ogawa W, Matsuba H, Yoshida H, Pinna LA. (1997). Eur. J. Biochem., 246, 433 ± 439. Okada Y, Yokono K, Kasuga M, Baba S and Roth RA. Skolnik EY, Batzer A, Li N, Lee CH, Lowenstein E, (1994). J. Biol. Chem., 269, 11498 ± 11502. Mohammadi M, Margolis B and Schlessinger J. (1993). Hunter T. (1998). Biol. Sci., 353, 583 ± 605. Science, 260, 1953 ± 1955. Huyer G, Liu S, Kelly J, Mo€at J, Payette P, Kennedy B, Songyang Z and Cantley LC. (1995). Trends Biochem. Sci., Tsaprailis G, Gresser MJ and Ramachandran C. (1997). J. 20, 470 ± 475. Biol. Chem., 272, 843 ± 851. Stein E, Cerretti DP and Daniel TO. (1996). J. Biol. Chem., Laviola L, Giorgino F, Chow JC, Baquero JA, Hansen H, 271, 23588 ± 23593. Ooi J, Zhu J, Riedel H and Smith RJ. (1997). J. Clin. SunXJ,PonsS,AsanoT,MyersJrMG,GlasheenEand Invest., 99, 830 ± 837. White MF. (1996). J. Biol. Chem., 271, 10583 ± 10587. Liu F and Roth RA. (1995). Proc. Natl. Acad. Sci. USA, 92, Twamley-Stein GM, Pepperkok R, Ansorge W and Court- 10287 ± 10291. neidge SA. (1993). Proc. Natl. Acad. Sci. USA, 90, 7696 ± Liu F and Roth RA. (1998). Mol. Cell. Biochem., 182, 73 ± 7700. 78. Uehara Y, Hori M, Takeuchi T and Umezawa H. (1985). Jpn Liu X, Marengere LE, Koch CA and Pawson T. (1993). Mol. J. Cancer Res. (Gann), 76, 672 ± 675. Cell. Biol., 13, 5225 ± 5232. Wang J, Dai H, Yousaf N, Moussaif M, Deng Y, Boufelliga Mano H, Ohya K, Miyazato A, Yamashita Y, Ogawa W, A, Swamy OR, Leone ME and Riedel H. (1999). Mol. Cell. Inazawa J, Ikdea U, Shimada K, Hatake K, Kasuga M, Biol., 19, 6217 ± 6228. Ozawa K and Kajigaya S. (1998). Genes Cells, 3, 431 ± 441. Weng Z, Thomas SM, Rickles RJ, Taylor JA, Brauer AW, McGlade J, Cheng A, Pelicci G, Pelicci PG and Pawson T. Seidel-Dugan C, Michael WM, Dreyfuss G and Brugge JS. (1992). Proc. Natl. Acad. Sci. USA, 89, 8869 ± 8873. (1994). Mol. Cell. Biol., 14, 509 ± 521. MorrioneA,PlantP,ValentinisB,StaubO,KumarS,Rotin White MF. (1998). Mol. Cell. Biochem., 182, 3 ± 11. D and Baserga R. (1999). J. Biol. Chem., 274, 24094 ± Yamanashi Y and Baltimore D. (1997). Cell, 88, 205 ± 211. 24099. Zhang B, Tavare JM, Ellis L and Roth RA. (1991). J. Biol. Morrione A, Valentinis B, Li S, Ooi JYT, Margolis B and Chem., 266, 990 ± 996. Baserga R. (1996). Cancer Res., 56, 3165 ± 3167.

Oncogene