Oncogene (1998) 16, 2425 ± 2434  1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00 http://www.stockton-press.co.uk/onc Crk protein binds to PDGF receptor and insulin receptor substrate-1 with di€erent modulating e€ects on PDGF- and insulin-dependent signaling pathways

Andrey Sorokin1, Eleanor Reed1, Naenna Nnkemere2, Nickolai O Dulin1 and Joseph Schlessinger3

1Department of Medicine, Medical College of Wisconsin, 8701 Watertown Plank Road, Milwaukee, Wisconsin 53226-0509; 2Department of Medicine, Case Western Reserve University, 10900 Euclid Avenue, Cleveland, Ohio 44106; 3Department of Pharmacology, New York University Medical Center, 550-First Avenue, New York, NY 10016, USA

We have studied the involvement of murine c-Crk, an oncogene (p120) have been reported (Birge et al., 1993; SH2/SH3 containing adaptor protein, in signaling Sakai et al., 1994; Buday et al., 1996; Fukazawa et al., pathways stimulated by di€erent receptor tyrosine 1996). Further, guanine-nucleotide exchange factors kinases. We show here that c-Crk is associated with C3G and Sos have been described as intracellular components of insulin- and PDGF-dependent signaling targets of c-Crk SH3 domains (Tanaka et al., 1994; pathways. Insulin treatment of murine myoblast cells Knudsen et al., 1994; Feller et al., 1994; Matsuda et al., induces the formation of stable complex of endogenous c- 1994). Recently, protein DOCK180, that was shown to Crk with insulin receptor substrate-1 (IRS-1) mediated be involved in the control of cell morphology, has been via the SH2 domain of Crk. The ligand dependent described as a major target for Crk SH3 domain physical association of c-Crk with IRS-1 is direct. (Hasegawa et al., 1996). Association of Crk SH3 However IRS-1 is also co-precipitated with c-Crk from domain with a substrate of EGF receptor EPS15 was quiescent L6 cells. The association of IRS-1 with c-Crk also reported (Matsuda and Kurata, 1996; Hansen et in quiescent cells is probably not direct since Far al., 1997). Western blot analysis did not reveal the binding of Src homology 2 and 3 domains (SH2 and SH3) are neither SH2 domain nor amino-terminal SH3 domain of small protein modules found in a variety of signaling c-Crk to IRS-1 from unstimulated cells. We also show proteins, which mediate protein-protein interactions in that PDGF treatment of murine myoblast cells induces signal transduction pathways activated by protein association of c-Crk with the PDGF receptor and tyrosine kinases (Pawson and Schlessinger, 1993; tyrosine phosphorylation of c-Crk. Overexpression of c- Cohen et al., 1995). SH2 domains bind to short Crk enhanced insulin- but not PDGF-induced activation phosphotyrosine containing sequences in growth of MAP kinases when compared to parental cell lines. factor receptors and other phosphoproteins. SH3 Thus, the formation of the direct IRS-1/Crk complex domains bind to target proteins through sequences appears to be crucial for Crk-mediated insulin-induced containing proline and hydrophobic amino acids. activation of MAP kinase, whereas Crk is probably Adaptor protein Crk has a simple structure consisting involved in other PDGF-induced responses. These data of one SH2 and two SH3 domains. With the SH2 provide support to the hypothesis that insulin and PDGF domain of adaptor protein bound to speci®c tyrosines employ di€erent mechanisms for activation of MAP in phosphorylated proteins, the SH3 domains are free kinase cascade. to interact with proline rich target proteins, resulting in the formation of multisubunit signaling complexes. Keywords: Crk; PDGF receptor; IRS-1; ERK cascade The actions of insulin are initiated by binding of the hormone to its heterotetrameric receptor (IR), which possesses intrinsic protein tyrosine kinase activity. The IR is composed of two a- and two b-subunits and Introduction insulin binding immediately activates the tyrosine kinase in the b-subunit, resulting in autophosphoryla- The oncogene v-Crk, originally isolated from avian tion of the receptor (Myers and White, 1993; White sarcoma viruses CT10 (Mayer et al., 1988), induces and Kahn, 1994). The autophosphorylation of the IR enhanced tyrosine phosphorylation of several proteins functions predominantly to up-regulate the activity of in v-Crk-transformed cells (Mayer and Hanafusa, the tyrosine kinase catalytic domain (Herrera and 1990) leading to cell transformation. Human Crk Rosen, 1986). The IR phosphorylates the widely cDNA has been subsequently isolated from a placenta expressed cytosolic protein IRS-1, a major target for cDNA library (Matsuda et al., 1992). To date little is both the IR and the insulin-like growth factor-1 known about intracellular function of c-Crk, although (IGF1) receptor (Myers et al., 1994a; White, 1994; phosphorylation of Crk by Abl tyrosine kinase (Ren et White et al., 1985). al., 1994), and association of Crk through its SH2 Phosphorylation of IRS-1 appears to be essential for domain with paxillin (p70), Cas (Crk-associated activation of many responses (White and Kahn, 1994), substrate, p130) and the product of the c-cbl proto- including insulin-stimulated mitogenesis (Wang et al., 1993; Sun et al., 1992; Waters et al., 1993; Rose et al., 1994), insulin stimulation of phosphatidylinositol 3'- Correspondence: A Sorokin kinase (PI 3'-kinase) (Myers et al., 1994b) and ERK, as Received 5 December 1997; revised 8 December 1997; accepted 9 well as S6 kinase (Chuang et al., 1993a,b; Myers et al., December 1997 1994b). Both IRS-1 and the recently discovered IRS-2 Role of Crk in PDGF- and insulin-dependent signaling pathways A Sorokin et al 2426 (Sun et al., 1995) play a major role in insulin-mediated signaling molecules, we overexpressed c-Crk stably in metabolic signaling and glucose transport (Rice and skeletal muscle myoblasts L6. Overexpression of c-Crk Garner, 1994; Quon et al., 1994). Many of the e€ects enhanced insulin- but not PDGF-induced activation of of insulin are thought to be mediated by the interaction ERKs when compared to parental cell lines. Since of IRS-1 and IRS-2 (Sun et al., 1991) with SH2- previously it has been speculated that insulin and containing enzymes and adaptor molecules (Sun et al., PDGF di€er in their Ras-GTP loading requirement for 1993; White and Kahn, 1994). IRS-2 is the alternative ERK activation these data provide evidence of a substrate of IR in IRS-1 de®cient mice and has further point of divergence between the two ligands signi®cant structural similarity to IRS-1 (Patti et al., and their mechanisms of Ras dependent signaling. 1995). Recently another insulin receptor substrate (termed IRS3) was described in rat adipocytes (Smith-Hall et al., 1997). Results Among signaling proteins PI 3'-kinase, the tyrosine phosphatase Syp, adaptors Grb2 and Nck were shown Interaction between Crk and IRS-1 to be associated with IRS-1 through their SH2 domains (Case et al., 1994; Lee et al., 1993; Skolnik The rapid induction of tyrosine phosphorylation of et al., 1993a; Sun et al., 1992). It has been suggested IRS-1 sets the stage for association of IRS-1 with SH2- that binding of IRS-1 to Grb2 might be one pathway containing adaptor proteins. To test the possibility that used by insulin for activation of Ras, since insulin- c-Crk associates with IRS-1 in vivo, we performed a co- induced activation of ERKs was found to be enhanced immunoprecipitation experiment with anti-Crk anti- by stable overexpression of Grb2 protein (Skolnik et bodies using insulin-stimulated L6 myoblasts (Figure al., 1993b). However, the formation of IRS-1/Grb2/Sos 1). The protein corresponding to the tyrosine complex was shown to be insucient for activation of phosphorylated IRS-1 was detected after immunoblot- ERKs (Pruett et al., 1995). Association of IRS-1 with ting with antiphosphotyrosine antibodies. However, the c-Crk was demonstrated in NIH3T3 and 293 cells major phosphotyrosine containing proteins associated stably transfected with an expression vector containing with c-Crk has an apparent molecular mass of p130. the Crk cDNA (Beitner-Johnson et al., 1996). Over- This protein is likely to be Cas (p130cas), an SH3- expression of Crk enhanced IGF-1 mitogenesis of containing signaling molecule with a cluster of SH2- NIH3T3 cells, but association of IRS-1 with Crk was binding motifs (Sakai et al., 1994). Cas is known to be decreased by IGF-1 treatment (Beitner-Johnson et al., highly phosphorylated on tyrosine residues in cells 1996). Clearly, current knowledge on the signaling transformed by v-Crk and v-Src (Sakai et al., 1994; pathways employed by insulin is far from complete. Kanner et al., 1991). Cas forms a stable complex with Platelet-derived growth factor (PDGF) is a potent v-Crk and c-Crk in a phosphorylation-dependent mitogen involved in the normal proliferation and manner and is a substrate for Src-family kinases di€erentiation of a variety of cell types (Heldin and (Vuori et al., 1996). That Cas was phosphorylated on Westermark, 1990). In contrast to insulin, interactions tyrosine in the absence of insulin can be explained by between PDGF receptor and signaling proteins are in adhesion-induced tyrosine phosphorylation of Cas many cases direct and do not involve the phosphoryla- (Nojima et al., 1995; Vuori and Ruoslahti, 1995). Cas tion of a docking protein (such as IRS-1 and IRS-2) with multiple phosphorylation sites serving as targets for SH2 domains. Ligand-induced autophosphoryla- tion of PDGF receptor creates binding sites for a set of signaling proteins, including phospholipase Cg, L 6 L6 C3 GTPase activating protein of Ras (GAP), PI 3'- P.I. 335P.I. 335 Lysate kinase, tyrosine phosphatase Syp, members of the Src family of protein kinases and the adaptor protein Nck INS –––––+++++ (Kazlauskas et al., 1991, 1992, 1993; Kashishian et al., 1992; Ronnstrand et al., 1992; Li et al., 1992; 190 — Lechleider et al., 1993). c-Crk is phosphorylated in response to PDGF BB in porcine aortic endothelial 125 — cells and is associated in ligand-dependent manner with 88 — 72 kDa component related to the adaptor molecule STAM (Hansen et al., 1997). 65 — 56 — In this report, we describe experiments designed to investigate the role of adaptor protein Crk in insulin- and PDGF-stimulated mitogenic signaling. We have 38 — 33.5 — cloned a murine c-Crk gene by screening of lEXlox expression library of 16-day mouse embryo with the Figure 1 Association of c-Crk protein with tyrosine phosphory- autophosphorylated region of the epidermal growth lated proteins in L6 cells. Lysates from insulin-stimulated (+) or unstimulated (7) L6 cells were subjected to immunoprecipitation factor receptor in a method designated CORT (Skolnik with pre-immune serum (PI) or anti-Crk antiserum (#335) before et al., 1991; Lowenstein et al., 1992; Margolis et al., separation. The resulting precipitates were analysed by Western 1992). Using antibodies against c-Crk and fusion blotting with polyclonal anti-P-tyr antibodies. The position of proteins of di€erent domains of crk we have detected immunoprecipitated phosphorylated band corresponding to IRS-1 the formation of IRS-1/Crk and PDGF receptor/Crk is indicated by arrow. No tyrosine phosphorylated proteins were immunoprecipitated by pre-immune serum. Positions of molecular complexes both in vivo and in vitro. In an attempt to weight markers (in kilodaltons) are shown on the left. Experiment elucidate the function of Crk when in complex with was repeated multiple times with similar results Role of Crk in PDGF- and insulin-dependent signaling pathways A Sorokin et al 2427 was found to be tyrosine-phosphorylated in some generated and used the fusion proteins of Crk continuously adherent cells (Harte et al., 1996; Ribon cytoplasmic signaling domains responsible for its and Saltiel, 1996). We observed a decreased association interaction with intracellular signaling molecules (the of Crk with Cas after insulin treatment. This is most SH2 domain and amino-terminal SH3 domain) likely due to insulin-induced redistribution of c-Crk (Matsuda and Kurata, 1996). IRS-1 was immunopre- from the docking protein Cas to newly phosphorylated cipitated from unstimulated and insulin-treated L6 cells proteins. As shown in Figure 1 a tyrosine phosphory- and the Far Western blot was performed (Figure 3a), lated protein with molecular mass of 175 kDa was co- along with control immunoblotting with anti-IRS-1 immunoprecipitated with anti-Crk antibodies from L6 antibodies (Figure 3b). To obtain the evidence that our cells, but not by pre-immune antiserum. Bovine fusion protein of Crk amino-terminal SH3 domain can pancreas and human recombinant insulin (both from bind to proteins which are known to interact with the Sigma) were used with similar results. Following amino-terminal SH3 domain of Crk we have per- insulin treatment of L6 cells, Crk dissociates from formed the control Far Western blot with C3G and p130, possibly due to a higher anity of Crk SH2 for DOCK180 proteins immunoprecipitated from unstimu- IRS-1 compared with p130 (presumably Cas). The lated L6 cells (Figure 3c). Both C3G and DOCK180 similar dynamics of interaction of Crk with p130 was were detected in a Far Western blot analysis with Crk reported following EGF stimulation (Khwaja et al., 1996), in which case Crk was shown to dissociate from p130 and bind to p120 (cbl). To prove that the co- precipitated protein with molecular mass of 175 kDa a represents IRS-1, we peformed immunodepletion Ins –––+++ experiments. As shown in Figure 2 (upper panel) 202 — when L6 cell lysates were immunodepleted using anti- IRS-1 Protein A Agarose a tyrosine phosphorylated protein with molecular mass of 175 kDa was not further co-immunoprecipitated with anti-Crk antibo- 109 — dies. The ®nal veri®cation of the identity of co- immunoprecipitated band was obtained by Western 78 — blot analysis with anti-IRS-1 antibodies (Figure 2 lower panel). Thus, a tyrosine phosphorylated band co-immunoprecipitated with anti-Crk antibodies repre- 46 — sents IRS-1. Moreover, it appears that nonphosphory- lated IRS-1 is co-precipitated with c-Crk from quiescent L6 cells. Although an association of Crk 34 — with IRS-1 in unstimulated cells was shown previously (Beitner-Johnson et al., 1996), the nature of this association is unclear. SH3 SH2 GST We sought to determine whether Crk and IRS-1 are able to interact directly with each other. To this end we b c studied the interaction between Crk and IRS-1 using the Far Western method utilizing fusion proteins of + Crk – C3G Ins IP DOCK180 di€erent domains of c-Crk (Figure 3a). We have 202 — 202 — DOCK180 109 — C3G 109 — 123456 78 — 78 — IB anti-P-Tyr 46 — 46 —

34 — 34 — 123456 IB IRS–1 SH3 IB anti-IRS-1 Figure 3 Binding of SH2 domain of c-Crk to IRS-1. Far Western blot analysis of insulin-stimulated (5 min) (+) or unstimulated (7) L6 cells. Positions of molecular weight markers Figure 2 Association of c-Crk protein with IRS-1 in L6 cells. (in kilodaltons) are shown on the left. The position of IRS-1 is Lysates from insulin-stimulated (5 min) (2,4,6) and unstimulated indicated. (a) The IRS-1 was immunoprecipitated and after (1,3,5) L6 cells were subjected either to immunoprecipitation with separation by SDS ± PAGE was probed either with GST-Crk anti-Crk antiserum (Santa Cruz) (lanes 1 and 2) or ®rst fusion protein containing SH2 domain (SH2) or n-terminal SH3 immunodepleted with anti-IRS-1 Protein A Agarose (lanes 5 domain (SH3). As a control IRS-1 immunoprecipitates were and 6) prior to immunoprecipitation with anti-Crk antiserum probed with glutathione-S-transferase (GST). (b) Control blot (lanes 3 and 4) antibodies. The resulting precipitates were with anti-IRS-1 antibodies. (c) The ability of used GST-Crk analysed by Western blotting with monoclonal anti-P-tyr fusion protein containing n-terminal SH3 domain to recognize antibodies (upper panel) and after stripping reblotted with anti- target proteins was veri®ed with immunoprecipitates of C3G and IRS-1 antibodies (lower panel). The position of IRS-1 is shown Dock180 proteins. Experiment was repeated twice with similar by arrow. Experiment was repeated twice with similar results results Role of Crk in PDGF- and insulin-dependent signaling pathways A Sorokin et al 2428 amino-terminal SH3 domain. Furthermore, the fusion endogenous c-Crk in a ligand-dependent manner protein of Crk amino-terminal SH3 domain bound (Figure 5a). As in the case of Crk-IRS-1 interaction predominately to C3G and DOCK180 proteins co- overexpression of Crk did not cause enhanced precipitated with Crk when Crk immunoprecipitate association of Crk with PDGF receptor. Although was subjected to a Far Western blot analysis (Figure the ability of SH2 domain of Crk to associate with 3c). These results demonstrate that the formation of PDGF receptor has been previously reported IRS-1 complexed with c-Crk in insulin-treated cells is (Anderson et al., 1990), our data is the ®rst evidence mediated through the c-Crk SH2 domain and not of association of PDGF receptor with Crk in lysates through its SH3 domain. Our results also suggest that from PDGF stimulated cells. The association between interaction of Crk with IRS-1 in unstimulated cells is Crk and PDGF receptor was also demonstrated in either not direct, or is not mediated by Crk cytoplasmic dog kidney epithelial cells (TRMP) expressing wild- signaling domains. type human PDGF receptor type b (Figure 6). A Stable expression of c-Crk in myoblast L6 cells was tyrosine phosphorylated protein with apparent mole- examined by immunoprecipitation of Crk with poly- cular mass about 180 kD was co-precipitated with Crk clonal anti-Crk antibodies followed by immunoblotting antibodies in PDGF-dependent manner only from with a di€erent polyclonal anti-Crk antibodies. Approximately 6 ± 8-fold overexpression of Crk was demonstrated in cell lines L6C3 (Figure 4) and L6C6 (Figure 7b). Surprisingly, overexpression of Crk did not enhance the association of Crk with IRS-1 (Figure a Lysates 335 1). However, Crk overexpression increased the associa- tion of Crk with tyrosine phosphorylated p130 protein L6L6 C3 L6 L6 C3 (presumably Cas) (Figure 1) as well as with guanine PDGF –+––+ – + – – – + – nucleotide exchange proteins (see below). It is likely that the amount of Crk-IRS-1 complex formed in L6 cells is primarily dependent upon the amount of IRS-1 available. b PDGFR ––+ +

Crk binds directly to PDGF receptor in PDGF-treated 190 — cells and is phosphorylated on tyrosine PDGFR We found that PDGF receptor is eciently co- immunoprecipitated from L6 myoblast cells with 125 —

L 6 L 6 C 3 88 —

323P.I. 323 P.I. 65 — 88 —

56 —

65 —

56 —

38 — SH2 GST Figure 5 Association of c-Crk protein with PDGF receptor in L6 cells. Lysates from PDGF-stimulated (5 min) (+) and 33.5 — unstimulated (7) cells were either (a) immunoprecipitated with anti-Crk antibodies, separated by SDS ± PAGE and analysed by Western blotting with polyclonal anti-P-Tyr antibodies or (b) Figure 4 Overexpression of c-Crk protein in L6 cells. Lysates of subjected to Far Western blot analysis and probed with GST-Crk parental L6 cells and cells overexpressing c-Crk (L6C3) were fusion protein containing SH2 domain (SH2). As a control the immunoprecipiated with pre-immune serum (PI) or anti-Crk lysates were probed with glutathione-S-transferase (GST). The antiserum (#323). The immunocomplexes were analysed by position of a phosphorylated band, corresponding to PDGF SDS ± PAGE and Western blotting with an anti-Crk antibodies receptor is indicated Role of Crk in PDGF- and insulin-dependent signaling pathways A Sorokin et al 2429

Lysates IP 335 a Lysates IP ControlPDGFR Control PDGFR PDGF –+++ – – – + ––––++++ — 56 IB anti-P-Tyr 175 — PDGF R — 38

83 — — 33.5

IB anti-Crk Crk

L6L6 C3 L6 L6 C3 Figure 6 Association of c-Crk protein with PDGF receptor in canine epithelial cells overexpressing PDGF b receptor. Lysates from PDGF-stimulated (5 min) (+) and unstimulated (7) canine b epithelial cells were either subjected to SDS ± PAGE (Lysates) or L6 L6 C6 were immunoprecipitated with anti-Crk antibodies (335), sepa- rated by SDS ± PAGE before the Western blotting with –++ – monoclonal anti-P-Tyr antibodies (upper panel) and with anti- Crk antibodies (lower panel). Mock transfected canine epithelial cells (Control) and cells overexpressing PDGF b receptor IB anti-P-Tyr (PDGFR) were analysed. The position of a phosphorylated band, corresponding to PDGF receptor and a position of Crk are indicated. Positions of molecular weight markers (in kilodaltons) are shown on the left. Experiment was repeated twice with similar results

IB anti-Crk

TRMP expressing PDGF receptor, but not from mock transfected TRMP. Far Western blot analysis using Figure 7 c-Crk is phosphorylated in response to PDGF. (a) Lysates from PDGF-stimulated (5 min) (+) or unstimulated (7) fusion protein of SH2 domains of c-Crk of lysates of L6 and L6 C3 cells were either separated by SDS ± PAGE PDGF-treated and untreated L6 cells demonstrated (Lysates) or subjected to immunoprecipitation with anti-Crk that Crk/PDGF receptor interaction is direct and is antiserum before separation by SDS ± PAGE. The resulting mediated by c-Crk SH2 domain (Figure 5b). Here we precipitates and lysates were analysed by Western blotting with show that PDGF receptor is a potent stimulator of c- an anti-Crk antibodies. Positions of molecular weight markers (in kilodaltons) are shown on the right. PDGF-induced redistribution Crk phosphorylation (Figure 7). c-Crk was detected as of Crk bands can be clearly seen both in cell lysates and in two bands (Figure 4) and PDGF treatment was found immunoprecipitates from cells overexpressing c-Crk (L6C3). (b) to induce a redistribution of c-Crk between these two Similar experiment has been performed with L6 and L6 C6 cells. bands (Figure 7). The induced redistribution was Immunoprecipitates with anti-Crk antiserum were analysed by obvious not only following Crk immunoprecipitation, Western blotting either with an anti-Crk antibodies, or with anti- P-tyr antibodies. The c-Crk band possessing less electrophoretic but was also detectable after analysis of cell lysates. In mobility corresponded to thyrosine phosphorylated c-Crk. order to show that the band characterized by reduced Experiment was repeated multiple times with similar results electrophoretic mobility corresponded to the tyrosine phosphorylated form of c-Crk, the Crk immunopreci- pitate was immunoblotted in parallel with anti-P-tyr and anti-Crk antibodies (Figure 7b). Only the Crk of insulin-stimulated ERK tyrosine phosphorylation band possessing reduced electrophoretic mobility was was further con®rmed by utilizing antibodies against shown to be phosphorylated on tyrosine residue. The phosphorylated forms of ERK1 and ERK2 (New phosphorylation of Crk in PDGF-stimulated cells was England Biolabs) (data not shown). The ability of recently reported in porcine aortic endothelial cells insulin to induce tyrosine phosphorylation of IRS-1, expressing PDGF b-receptors (Hansen et al., 1997). detected either by Western blot analysis with antipho- sphotyrosine antibodies, or by following the insulin- induced shift in electrophoretic mobility of IRS-1 Cellular responses of L6 and L6C6 to insulin and PDGF (Western with anti-IRS-1 antibodies) was identical in As expected, insulin and PDGF induced MAP kinase parental L6 cells and in cells overexpressing c-Crk activation in L6 cells. We compared the level of these (data not shown). It is noteworthy that the increased responses in parental L6 cells to L6 cells over- tyrosine phosphorylation of ERKs in cell lines expressing Crk. Overexpression of Crk (approxi- overexpressing Crk did not result simply from a mately sixfold) resulted in the signi®cant change in the kinetics of ERK phosphorylation. The enhancement of the ability of insulin to stimulate kinetics of insulin-induced tyrosine phosphorylation of tyrosine phosphorylation of ERK1 and ERK2 (Figure ERKs was similar in parental cells and Crk over- 8a). In contrast, the PDGF-induced stimulation of expressing cells. We have also checked whether the ERKs phosphorylation was not altered by Crk absence of the e€ect of Crk overexpression upon overexpression (Figure 8b). The increased activation tyrosine phosphorylation of ERKs in cells stimulated Role of Crk in PDGF- and insulin-dependent signaling pathways A Sorokin et al 2430

a L6 L6 C6 LysatesIP 335 Lysates IP 335

005510 10 L6 L6 L6 L6 ERK1 L6 C6 L6 C6 L6 C6 L6 C6 ERK2 200 — IB Sos1 IB C3G 113 —

b L6 L6 C6

Insulin ––++ –– IgG PDGF – + – –+ – Crk

ERK1 ERK2 IB Crk Figure 9 Association of c-Crk with Sos and C3G. Lysates from quiescent control cells (L6) and cells overexpressing Crk (L6 C6) were immunoprecipitated with anti-Crk antibodies. Lysates and c L6 L6 C3 the resultant immunoprecipitates were resolved by SDS ± PAGE and then immunoblotted either with anti-Sos1 (upper left panel) or anti-C3G (upper right panel). The lower panel represents 000.55.050.0 0.5 5.0 50.0 immunoblotting with anti-Crk antibodies. The position of Crk is indicated. Positions of molecular weight markers (in kilodaltons) are shown on the left. The data are representative of three independent experiments

by PDGF is obscured by maximal stimulation of d ERKs phosphorylation already in PDGF-treated control cells. Figure 8c shows that at any tested L6 L6 C6 concentration of PDGF overexpression of Crk had no Ins –++ –– – e€ect upon tyrosine phosphorylation of ERKs. Finally PDGF – –+ – – + we have performed direct assay of ERK activity using polyclonal antibody recognizing ERK2 (Figure 8d) MBP (Wang et al., 1992). In accordance with our phosphotyrosine blots we show that overexpression of Crk has an enhancing e€ect upon Insulin (but not PDGF) induced ERK activity.

Crk association with guanine nucleotide exchange proteins Sos and C3G C3G catalyzes a guanine nucleotide exchange reaction of Rap1 (Gotoh et al., 1995), whereas Sos induces the dissociation of GDP from Ras, allowing the formation of an activated GTP ± Ras complex (Bon®ni et al.,1992). Amino-terminal SH3 domain of Crk inter- acts with proline-rich sequences in C3G and Sos (Matsuda and Kurata, 1996). We have detected an increase in insulin-induced ERK activation in the cells overexpressing Crk in comparison with the parental L6 Figure 8 c-Crk overexpression enhances tyrosine phosphoryla- tion of ERKs by insulin, but not by PDGF. (a) Time course of cells. Since the ERK signaling cascade is under the insulin-induced tyrosine phosphorylation of ERK1 and ERK2 in control by Ras we determined the association of Crk cells overexpressing c-Crk (L6C6) and in control cell line (L6). with Sos and C3G in cells overexpressing Crk. Co- After stimulation with insulin for various time (top, in minutes), precipitation experiments revealed an increase in cell lysates were processed and immunoblotted with anti-P-Tyr association of Crk with both Sos and C3G in L6C6 antibodies. ERK1 and ERK2 are indicated. (b) After stimulation with insulin or PDGF for 5 min, cell lysates were processed and cells in comparison with L6 cells. It is of note that immunoblotted with anti-P-Tyr antibodies. ERK1 and ERK2 are indicated. Insulin-stimulated phosphorylation of ERKs is enhanced in cells overexpressing c-Crk. PDGF-stimulated phosphorylation of ERKs is not changed in these cells. The the ability of immunoprecipitated ERK to phosphorylate myelin di€erence in the level of manifestation of ERKs phosphorylation basic protein (MBP). The position of MBP is indicated. The lower between a and b is due to di€erent time of ®lm exposure. (c) After panel shows quanti®cation of MBP phosphorylation (closed stimulation with di€erent concentrations of PDGF (ng/ml) for columns ± unstimulated cells, open columns ± stimulation with 5 min, cell lysates of cells were processed and immunoblotted with insulin for 5 min, gray columns ± stimulation with PDGF for anti-P-Tyr antibodies. Overexpression of c-Crk (L6 C3 cells) does 5 min). Values are expressed as arbitrary units based upon the not result in enhanced PDGF-induced phosphorylation of ERKs amount of c.p.m. in the MBP band in each lane. Incubation time in comparison with control cell line (L6) at any concentration of with mitogens 5 min. Experiment was repeated twice with similar the ligand. (d) The upper panel shows ERK activity assayed by results and representative experiment is shown Role of Crk in PDGF- and insulin-dependent signaling pathways A Sorokin et al 2431 overexpression of Crk had more e€ect upon formation cascade. However the respective ecacy with which of the C3G-Crk complex, than upon the formation of EGF, PDGF and insulin stimulate levels of p21ras-GTP Sos1-Crk complex. may vary with the cell type, for example in 3T3-L1 cells, PDGF was a more e€ective than was insulin (Klarlund et al., 1995). Discussion One of the principal di€erences in signal transduction by insulin when compared to PDGF signaling is an An active participation of SH2/SH3 containing engagement by insulin receptor of IRS-1, which is adaptor proteins in signaling pathways is quite phosphorylated on tyrosine by the insulin receptor and established (Schlessinger, 1994). It is generally serves as a docking protein for SH2 domains of multiple accepted that the function of adaptor protein Grb2 signaling molecules (White and Kahn, 1994). No similar links protein tyrosine kinases to the Ras/MAPK substrate has been described for the PDGF signaling signaling pathway. This is achieved through recruit- network and it had become evident that tyrosine ment of the guanine nucleotide exchange factor Sos autophosphorylation of PDGF receptor itself creates (Simon et al., 1993; Egan et al., 1993; Rozakis-Adcock binding sites for SH2 containing intracellular proteins. et al., 1993; Li et al., 1993; Gale et al., 1993). The We show here that adaptor protein Crk, binds to understanding of Grb2 function was supported by tyrosine phosphorylated IRS-1 and to autophosphory- biochemical studies as well as by genetic analysis in lated PDGF receptor, and is involved in signaling Caenorbabditis elegans and Drosophila. To date genetic downstream of insulin and PDGF receptors. analysis of the role of Crk is not available. Because There are few demonstrations of ligand-induced guanine nucleotide exchange factor C3G was shown to association of c-Crk with a tyrosine kinase receptor be associated with Crk in the way similar to in vivo in cell lines where neither c-Crk nor the growth association of Grb2 with Sos (Tanaka et al., 1994), it factor receptor are overexpressed. Association of EGF is possible that adaptor proteins Grb2, Crk and Nck receptor with Crk is a subject of intense discussion. perform similar intracellular functions: recruit di€erent Both, the fact of in vitro EGF receptor/Crk interaction partners which share the ability to activate the Ras (Birge et al., 1992; Fajardo et al., 1993), and the signaling pathway. Assuming that c-Crk is linking observation of in vivo association of v-Crk with EGF insulin- and PDGF-induced signaling pathways to Ras receptor (Hempstead et al., 1994), suggest that c-Crk activation, our data demonstrate that insulin and may be interacting with phosphorylated EGF receptor PDGF dependent pathways di€er in the requirement in vivo. Interaction of c-Crk with EGF receptor in vivo for MAP kinase activation. was shown in highly transformed A431 cells, char- Work from several laboratories has indicated that acterized by overexpression of EGF receptors (Fajardo the level of GTP-bound Ras increases following PDGF et al., 1993), and in human mammary epithelial cells stimulation of cells (Gibbs et al., 1990; Satoh et al., (Fukazawa et al., 1996). We have detected in vivo 1990; Nanberg and Westermark, 1993; Maruta and association of endogenous c-Crk with endogenous Burgess, 1994). In cells expressing wild-type PDGF PDGF receptor in L6 myoblast utilizing co-immuno- receptor, PDGF stimulation resulted in an approx- precipitation methodology. imate twofold increase in the GTP content of Ras, with Recently, it was shown that IGF1 stimulation of only 5% of GTP versus 95% of GDP following cells overexpressing Crk resulted in a dissociation of immunoprecipitation with anti-Ras antibodies (Valius Crk protein from IRS-1 (Beitner-Johnson et al., 1996). and Kazlauskas, 1993). In contrast to PDGF, EGF These workers found that in clones of 3T3-NIH and converts up to 65% of cellular p21ras from the GDP 293 cells, overexpressed Crk was in stable association form into the active GTP-bound state (Osterop et al., with IRS-1 and IRS-2 and this link was disrupted by 1993). Although insulin is less potent in p21ras addition of IGF1. Experiments in the present study activation than EGF, it has been demonstrated that indicate that the binding of endogenous Crk mediated in cells expressing 7 ± 7.56105 cell71 insulin receptors by Crk SH2 domain to endogenous IRS-1 was the high degree of Ras-GTP formation (30 ± 60%) is consequence of the stimulation of L6 rat myoblasts observed following insulin stimulation (Osterop et al., with insulin. However in our experiments, in accor- 1993; Medema et al., 1993). Thus, PDGF appears to be dance with experiments described by Beitner-Johnson a less potent stimulator of GTP loading of p21ras than et al. (1996), IRS-1 was able to be co-precipitated with insulin and the fact that c-Crk overexpression enhances Crk from unstimulated L6 cells. A di€erence between the ability of insulin, but not of PDGF, to induce previously reported data (Beitner-Johnson et al., 1996) ERKs activation is indirect evidence that PDGF may and those reported here is the demonstration in the engage some alternative mechanisms to activate ERK current work of insulin-induced SH2-mediated interac- cascade. It is of note that although there is a striking tion of Crk with IRS-1. This di€erence is may be due di€erence in the respective capacities of insulin and to the di€erent stimuli (IGF1 versus insulin) or PDGF to stimulate an increase in Ras-GTP, the di€erent cell types used in two studies. IRS-1 co- overexpression of the interfering mutant of p21ras, precipitated from unstimulated cells was not phos- RasAsn17, inhibits both insulin- and PDGF-induced phorylated (as revealed by phosphotyrosine blots and activation of ERK2 completely (Vries-Smiths et al., by its electrophoretic mobility) and was not recognized 1992). Our present data, dealing with the role of Crk in by Crk cytoplasmic signaling domains in far Western PDGF signaling, together with previously published blot analysis. This may be due to the indirect observations of other groups may be considered as an association of Crk with IRS-1 in unstimulated cells. indication that a very small increase in Ras-GTP, The recruitment of IRS-1 by activated insulin receptor perhaps in conjunction with other PDGF-induced is mediated by phosphotyrosine binding domain (PTB) signals, is sucient for the activation of ERK kinase of IRS-1 (Eck et al., 1996). One possibility is that in Role of Crk in PDGF- and insulin-dependent signaling pathways A Sorokin et al 2432 the absence of autophosphorylated insulin receptor described previously (Chen and Okayama, 1987), using PTB mediates an interaction of IRS-1 with some co-transfection with 0.5 mg of pSV2neo and 40 mgofthe tyrosine phosphorylated protein which is recognized at expression vector encoding c-Crk, followed by isolation of the same time by Crk SH2 domain. The possibility of individual G418-resistant colonies. Cells expressing mouse existence of such docking protein for both Crk and c-Crk were identi®ed by immunoblotting of cell lysates using antibodies raised against the carboxyl terminus of IRS-1 should be analysed further. mouse c-Crk. Parental L6 cell line and cells overexpressing Our data indicate that c-Crk is an endogenous c-Crk were maintained in Dulbecco's modi®ed Eagle's substrate of the PDGF receptor in rat myoblasts. The medium (DMEM) supplemented with 10% fetal bovine fact that PDGF induces Crk phosphorylation is serum and antibiotics. Dog kidney epithelial cells (TRMP) intriguing, since the role of tyrosine phosphorylation mock transfected and expressing wild type human PDGF of adaptor proteins is unknown. It was previously receptor were maintained in Dulbecco's modi®ed Eagle's shown that members of Abl family of kinases bind to medium with 10% fetal bovine serum and antibiotics the ®rst Crk SH3 domain and phosphorylate c-Crk on (Kazlauskas et al., 1992). TRMP were the gift of Dr J tyrosine 221 (Y221) (Feller et al., 1994; Ren et al., Cooper. 1994). Crk is also an endogenous substrate of the IGF- I receptor kinase (Beitner-Johnson and LeRoith, 1995). Far Western blotting It was reported that phosphorylation of Crk is negatively regulated by its C-terminal SH3 domain For the production of GST-Crk fusion proteins, Escher- ichia coli expressing individual pGEX constructs was (Ogawa et al., 1994). It is of note that the recently induced by the addition of isopropyl-b-D-thiogalactopyr- isolated human crk-like gene CRKL (Hoeve et al., anoside and the fusion proteins were captured on 1993) is the major tyrosine-phosphorylated protein in glutathione-Sepharose 4B beads (Pharmacia) as described neutrophils from patients with chronic myelogenous previously (Smith and Johnson, 1988). For Far Western leukemia (Oda et al., 1994). blot analysis cell lysates were subjected either directly to Together, these data implicate c-Crk as a novel SDS ± PAGE or ®rst immunoprecipitated with appropriate component of two quite di€erent signaling networks: antibodies and after SDS ± PAGE and transferring to insulin-dependent and PDGF-dependent. Further nitrocellulose ®lters were incubated with 1 mg of GST or studies are necessary to determine the molecular fusion protein per ml in Tris-bu€ered saline-bovine serum mechanisms responsible for the di€erent e€ect of c- albumin for 1 h at room temperature and washed once with Tris-bu€ered saline, twice with Tris-bu€ered saline- Crk overexpression upon PDGF and insulin signaling. 0.05% Triton X-100, and once with Tris-bu€ered saline (10 min each) prior to incubation with anti-GST antibodies (Upstate Biotechnology). The blot was processed to completion as described below. Materials and methods

Cloning of mouse c-Crk and construction of expression Cell lysis, immunoprecipitation and immunoblotting plasmids Cells were washed twice with ice-cold PBS prior to lysis The method of expression cloning CORT (cloning of with 500 mloflysisbu€er(50mM HEPES [N-2-hydro- receptor targets) we used has been described elsewhere xyethylpiperazine-N'-2-ethanesulfonic acid; pH 7.5), (Margolis et al., 1992). This method was used to clone 150 mM NaCl, 10% glycerol, 1% Triton X-100, 1.5 mM

proteins that contain SH2 domains or Phosphotyrosine MgCl2,1mM EGTA [ethylene glycol-bis(b-aminoethyl Binding Domains (PI/PTB), relying on the fact that these ether)-N,N,N',N'-tetraacetic acid], 10 mg of aprotinin per proteins can bind to autophosphorylated growth factor ml, 10 mg of leupeptin per ml, 1 mM phenylmethylsulfonyl receptors. Bacterial expression lEXlox library of 16-day ¯uoride, 0.2 mM sodium orthovanadate per 10-cm dish. mouse embryo (Novagene) was screened with the tyrosine Lysates were incubated for 5 min on ice and centrifuged phosphorylated carboxyterminal tail of the EGF-receptor for 20 min at 48Cand16000g. Supernatants were and sequencing of positive clones was done with Sequenase subjected to further analysis. (US Biochemicals). The clone encoding the full sequence of Equal amounts of total protein, as measured by the Bio- mouse c-Crk (analog of human c-Crk II) was used for the Rad protein assay, were incubated with antibodies for 1.5 h experiments described below. at 48C prior to adding Protein A-Sepharose for 1 h. For The expression plasmid for generating stable cell line immunoblot analysis, immunoprecipitates were washed three overexpressing c-Crk has been constructed as follows. The times with ice-cold lysis bu€er, boiled in sodium dodecyl DNA encoding the entire coding region of c-Crk was sulfate-polytacrylamide gel electrophoresis (SDS ± PAGE) ampli®ed by PCR and subcloned into the cytomegalovirus- sample bu€er for 5 min, subjected to SDS ± PAGE, and based eucaryotic expression vector pRK5 (Koch et al., 1994). transferred to nitrocellulose. For expression of di€erent c-Crk domains as GST fusion Immunodepletion experiments were performed as follows: proteins sequences encoding the mouse c-Crk SH2 (amino cell lysates were subjected to immunoprecipitation with anti- acids 1 ± 131) and amino-terminal SH3 (amino acids 106 ± IRS-1 Protein A Agarose (Upstate Biotechnology) following 236) regions were generated by PCR ampli®cation with manufacturer protocol. Immunodepleted cell lysates were mouse c-Crk and the appropriate oligonucleotide primers. immunoprecipitated with anti-Crk antiserum (Santa Cruz). In The PCR products were cloned into EcoRI and HindIII sites parallel Crk immunoprecipitation was performed from not

of vector pGEXTAG. The PCR products ligated into depleted cell lysates. eucaryotic and procaryotic expression vectors were subjected For Western blotting the nitrocellulose was ®rst incubated to DNA sequence analysis. with TBS/BSA solution (20 mM Tris pH 7.8, 150 mM NaCl, 5 mg/ml BSA) for 1 h at 428C and was incubated in TBS/ Cell lines BSA with blotting antibodies for 2 h at room temperature. The blotting was performed either with iodinated protein A, L6 C3 and L6 C6 are cell lines overexpressing the mouse c- or by utilizing ECL methodology. For ECL the nitrocellulose Crk. These cells were established by calcium phosphate- was subjected to a set of washings (®ve times 5 min each with mediated DNA transfection of L6 rat myoblasts as TBS/Triton (TBS with 0.05% Triton X-100), two times 5 min Role of Crk in PDGF- and insulin-dependent signaling pathways A Sorokin et al 2433 each with TBS) and incubated for 1 h with HRP-protein A portion before immunization. Rabbit polyclonal anti-Crk (Bio-Rad) in TBS/5% non fat dry milk. After the similar set antibodies #335 and #336 are directed against a peptide in of washes nitrocellulose was incubated with ECL mixture carboxy terminus of mouse c-Crk. Peptide (Amersham) according to manufacturer protocols. To detect VRLLDQQNPDEDFS was coupled to keyhole limpet primary antibodies with iodinated protein A, the nitrocellu- hemocyanine by glutaraldehyde and was used for lose was washed once with TBS (10 min), twice with TBS/ immunization. Commercial rabbit polyclonal antibodies Triton (each for 10 min) and again with TBS (10 min). After against c-Crk (Santa Cruz, CA) were used when speci®ed. incubation with iodinated Protein A in TBS/BSA for 1 h, the Tyrosine phosphoproteins were blotted and precipitated nitrocellulose was subjected to similar set of washes, dried with rabbit polyclonal antiphosphotyrosine (anti-P-Tyr) and exposed. Monoclonal anti-P-tyr antibodies conjugated to antibodies (Batzer et al., 1994). Commercial monoclonal HRP were used according to manufacturer protocols anti-P-tyr antibodies conjugated to HRP (Oncogene (Oncogene Research Products). Research Products) were used when speci®ed. Rabbit polyclonal anti-GST antibodies (Upstate Biotechnology) were used to detect GST fusion proteins in Far Western ERK activity assay blotting. Bovine pancreas and human recombinant insulin ERK activity was performed as described previously were used in concentrations from 0.85 to 1.7 mM. PDGF (Bokemeyer et al., 1996). BB was from R&D Systems (Minneapolis, MN) and was used in concentration 50 ng/ml.

Antibodies Anti-IRS-1 and anti-IRS-2 antibodies were purchased from Acknowledgements Santa Cruz Biotechnology, Inc. (Santa Cruz, CA) and This work was partly supported by National Institutes of Transduction Laboratories (Lexington, KY). A panel of Health Grant HL22563 and DK41684 and by grants from rabbit polyclonal antibodies against c-Crk able to The Milheim Foundation for Cancer Research (95-54) and immunoprecipitate and immunodetect c-Crk was gener- by the American Cancer Society Institutional Research ated. Rabbit polyclonal anti-Crk antibodies #323 are Grant to AS. We thank Ben Margolis, who participated in directed against a full length Crk expressed as a fusion the early stages of this project and acknowledge the help of protein in bacterial cells and cleaved from the fusion Ann McGinty in the preparation of the manuscript.

References

Anderson D, Koch CA, Grey L, Ellis C, Moran MF and Gibbs JB, Marshall MS, Scolnick EM, Dixon RAF and Pawson T. (1990). Science, 250, 979 ± 982. Vogel US. (1990). J. Biol. Chem., 265, 20437 ± 20442. Batzer AG, Rotin D, Urena JM, Skolnik EY and Schles- Gotoh T, Hattori S, Nakamura S, Kitayama H, Noda M, singer J. (1994). Mol. Cell. Biol., 14, 5192 ± 5201. Takai Y, Kaibuchi K, Matsui H, Hatase O, Takahashi H, Beitner-Johnson D and LeRoith D. (1995). J. Biol. Chem., Kurata T and Matsuda M. (1995). Mol. Cell. Biol., 15, 270, 5187 ± 5190. 6746 ± 6753. Beitner-Johnson D, Blakesley VA, Shen-Orr Z, Jimenez M, Fajardo JE, Birge RB and Hanafusa H. (1993). Mol. Cell. Stannard B, Wang LM, Pierce J and LeRoith D. (1996). J. Biol., 13, 7295 ± 7302. Biol. Chem., 271, 9287 ± 9290. Feller SM, Knudsen B and Hanafusa H. (1994). EMBO J., Birge RB, Fajardo JE, Mayer BJ and Hanafusa H. (1992). J. 13, 2341 ± 2351. Biol. Chem., 275, 10588 ± 10595. Feller SM, Ren R, Hanafusa H and Baltimore D. (1994). Birge RB, Fajaedo JE, Reichman C, Shoelson SE, Songyang Trends Biochem. Sci., 19, 453 ± 458. Z, Cantley LC and Hanafusa H. (1993). Mol. Cell. Biol., Fukazawa T, Miyake S, Band V and Band H. (1996). J. Biol. 13, 4648 ± 4656. Chem., 271, 14554 ± 14559. Bokemeyer D, Sorokin A, Yan M, Ahn NG, Templeton DJ Hansen K, Ronnstrand L, Claesson-Welsh L and Heldin C- and Dunn MJ. (1996). J. Biol. Chem., 271, 639 ± 642. H. (1997). FEBS Lett., 409, 195 ± 200. Bon®ni L, Karlovich CA, Dasgupta C and Banerjee U. Harte MT, Hildebrand JD, Burnham MR, Bouton AH and (1992). Science, 255, 603 ± 606. Parsons JT. (1996). J. Biol. Chem., 271, 13649 ± 13655. Buday L, Khwaja A, Sipeki S, Farago A and Downward J. Hasegawa H, Kiyokawa E, Tanaka S, Nagashima K, Gotoh (1996). J. Biol. Chem., 271, 6159 ± 6163. N, Shibuya M, Kurata T and Matsuda M. (1996). Mol. CaseRD,PiccioneE,WolfG,BenettAM,LechleiderRJ, Cell. Biol., 16, 1770 ± 1776. Neel BG and Shoelson SE. (1994). J. Biol. Chem., 269, Heldin C-H and Westermark B. (1990). Cell Regul., 1, 555 ± 10467 ± 10474. 566. Chen C and Okayama H. (1987). Mol. Cell. Biol., 7, 2745 ± Hempstead BL, Birge RB, Fajardo JE, Glassman R, 2752. Mahadeo D, Kraemer R and Hanafusa H. (1994). Mol. Chuang LM, Myers MG, Seidner GA, Birnbaum MJ, White Cell. Biol., 14, 1964 ± 1971. MF and Kahn CR. (1993a). Proc. Natl. Acad. Sci. USA, Herrera R and Rosen OM. (1986). J. Biol. Chem., 261, 90, 5172 ± 5175. 11980 ± 11985. Chuang LM, Myers MG, Backer JM, Shoelson SE, White Hoeve J, Morris C, Heisterkamp N and Gro€en J. (1993). MF, Birnbaum MJ and Kahn CR. (1993b). Mol Cell. Oncogene, 8, 2469 ± 2474. Biol., 13, 6653 ± 6660. Kanner SB, Reynolds AB, Wang HC, Vines RR and Parsons Cohen GB, Ren R and Baltimore D. (1995). Cell, 80, 237 ± JT. (1991). EMBO J., 10, 1689 ± 1698. 248. Kashishian A, Kazlauskas A and Cooper JA. (1992). EMBO Eck MJ, Dhe-Paganon S, Trub T, Nolte RT and Shoelson J., 11, 1373 ± 1382. SE. (1996). Cell, 85, 695 ± 705. Kazlauskas A, Durden DL and Cooper JA. (1991). Cell Egan SE, Giddings B, Brooks MW, Buday L, Sizeland AM Regul., 2, 413 ± 425. and Weinberg RA. (1993). Nature, 363, 45 ± 51. Kazlauskas A, Kashishian A, Cooper JA and Valius M. Gale N, Kaplan S, Lowenstein EJ, Schlessinger J and Bar- (1992). Molec. Cell. Biol., 12, 2534 ± 2544. Sagi D. (1993). Nature, 363, 88 ± 92. Role of Crk in PDGF- and insulin-dependent signaling pathways A Sorokin et al 2434 Kazlauskas A, Feng G-S, Pawson T and Valius M. (1993). Ren R, Ye Z-S and Baltimor D. (1994). Genes Dev., 8, 783 ± Proc. Natl. Acad. Sci. USA, 90, 6939 ± 6943. 795. Khwaja A, Hallberg B, Warne PH and Downward J. (1996). Ribon V and Saltiel AR. (1996). J. Biol. Chem., 271, 7375 ± Oncogene, 12, 2491 ± 2498. 7380. Klarlund JK, Cherniack AD and Czech MP. (1995). J. Biol. Rice KM and Garner CW. (1994). Bioch. Bioph. Res. Chem., 270, 23421 ± 23428. Commun., 198, 523 ± 530. Knudsen BS, Feller SM and Hanafusa H. (1994). J. Biol. Ronnstrand L, Mori D, Arridsson A-K, Eriksson A, Chem., 269, 32781 ± 32787. Wernstedt C, Hellman U, Claesson-Welsh L and Heldin Koch WJ, Hawes BE, Inglese J, Luttrell LM and Lefkowitz CH. (1992). EMBO J., 11, 3911 ± 3919. RJ. (1994). J. Biol. Chem., 269, 6193 ± 6197. Rose DW, Saltiel AR, Majumdar M, Decker SJ and Olefsky Lechleider RJ, Sugimoto S, Bennet AM, Kashishian AS, JM. (1994). Proc. Natl. Acad. Sci. USA, 91, 797 ± 801. Cooper JA, Shoelson SE, Walsh CT and Neel BG. (1993). Rozakis-Adcock M, Fernley R, Wade J, Pawson T and J. Biol. Chem., 268, 21478 ± 21481. Bowtell D. (1993). Nature, 363, 83 ± 85. Lee C-H, Li W, Nishimura R, Zhou M, Batzer AG, Myers Sakai R, Iwamatsu A, Hirano N, Ogawa S, Tanaka T, Mano MG, White MF, Schlessinger J and Skolnik EY. (1993). H, Yazaki Y and Hirai H. (1994). EMBO J., 13, 3748 ± Proc. Natl. Acad. Sci. USA, 90, 11713 ± 11717. 3756. Li N, Batzer A, Daly R, Yajnik V, Skolnik E, Chardin P, Satoh T, Endo M, Nakafuku M, Nakamura S and Kaziro Y. Bar-Sagi D, Margolis B and Schlessinger J. (1993). Nature, (1990). Proc. Natl. Acad. Sci. USA, 87, 5993 ± 5997. 363, 85 ± 88. Schlessinger J. (1994). Current Biol., 7, 25 ± 30. Li W, Hu P, Skolnik EY, Ullrich A and Schlessinger J. Simon MA, Dodson GS and Rubin GM. (1993). Cell, 73, (1992). Mol. Cell. Biol., 12, 5824 ± 5833. 169 ± 177. Lowenstein EJ, Daly RJ, Batzer AG, Li W, Margolis B, Skolnik EY, Margolis B, Mohammadi M, Lowenstein E, Lammers R, Ullrich A and Schlessinger J. (1992). Cell, 70, Fisher R, Drepps A, Ullrich A and Schlessinger J. (1991). 431 ± 442. Cell, 65, 83 ± 90. Margolis B, Silvennoinen O, Comoglio F, Roonprapunt C, Skolnik EY, Lee CH, Batzer A, Vincenti LM, Zhou M, Daly Skolnik EY, Ullrich A and Schlessinger J. (1992). Proc. R, Myers MG, Backer JM, Ullrich A, White MF and Natl. Acad. Sci. USA, 89, 8894 ± 8898. Schlessinger J (1993a). EMBO J., 12, 1929 ± 1936. Maruta H and Burgess AW. (1994). Bioessays, 16, 489 ± 496. Skolnik EY, Batzer A, Li N, Lee C-H, Lowenstein E, Matsuda M and Kurata T. (1996). Cell. Signal., 8, 335 ± 340. Mohammadi M, Margolis B and Schlessinger J. (1993b). MatsudaM,TanakaS,NagataS,KojimaA,KurataTand Science, 260, 1953 ± 1955. Shibuya M. (1992). Mol. Cell. Biol., 12, 3482 ± 3489. Smith DB and Johnson KS. (1988). Gene, 67, 31 ± 40. Matsuda M, Hashimoto Y, Muroya K, Hasegawa H, Kurata Smith-HallJ,PonsS,PattiME,BurksDJ,YenishL,SunXJ, T, Tanaka S, Nakamura S and Hattori S. (1994). Mol. Kahn CR and White MF. (1997). Biochemistry, 36, 8304 ± Cell. Biol., 14, 5495 ± 5500. 8310. Mayer BJ and Hanafusa H. (1990). J. Virology, 64, 3581 ± Sun XJ, Rothenberg P, Kahn CR, Backer JM, Araki E, 3589. Wilden PA, Cahill DA, Goldstein BJ and White MF. Mayer BJ, Hamaguchi M and Hanafusa H. (1988). Nature, (1991). Nature, 352, 73 ± 77. 332, 272 ± 275. Sun XJ, Miralpeix M, Myers MG, Glasheen EM, Backer JM, Medema RH, de Vries-Smits AMM, van der Zon GCM, Kahn CR and White MF. (1992). J. Biol. Chem., 267, Maassen JA and Bos JL. (1993). Mol. Cell. Biol., 13, 155 ± 24058 ± 24063. 162. Sun XJ, Crimmins DL, Myers MG, Miralpeix M and White Myers MG and White MF. (1993). Diabetes, 42, 643 ± 650. MF. (1993). Mol. Cell. Biol., 13, 7418 ± 7428. Myers MG, Sun XJ and White MF. (1994a). TIBS, 19, 289 ± Sun XJ, Wang LM, Zhang Y, Yenush L, Myers MG, 293. Glasheen E, Lane WS, Pierce JH and White MF. (1995). Myers MG, Grammer TC, Wang L-M, Sun XJ, Pierce JH, Nature, 377, 173 ± 177. Blenis J and White MF. (1994b). J. Biol. Chem., 269, Tanaka S, Morishita T, Hashimoto Y, Hattori S, Nakamura 28783 ± 28789. S, Shibuya M, Matuoka K, Takenawa T, Kurata T, Nanberg E and Westermark B. (1993). J. Biol. Chem., 268, Nagashima K and Matsuda M. (1994). Proc. Natl. Acad. 18187 ± 18194. Sci. USA, 91, 3443 ± 3447. NojimaY,MorinoN,MimuraT,HamasakiK,FuruyaH, Valius M and Kazlauskas A. (1993). Cell, 73, 321 ± 334. Sakai R, Sato T, Tachibana K, Morimoto C, Yazaki Y Vries-Smits AMM, Burgering BMT, Leevers SJ, Marshall CJ and Hirai H. (1995). J. Biol. Chem., 270, 15398 ± 15402. and Bos JL. (1992). Nature, 357, 602 ± 604. Oda T, Heaney C, Hagopian JR, Okuda K, Grin JD and Vuori K and Ruoslahti E. (1995). J. Biol. Chem., 270, Druker BJ. (1994). J. Biol. Chem., 269, 22925 ± 22928. 22259 ± 22262. Ogawa S, Toyoshima H, Kozutsumi H, Hagiwara K, Sakai Vuori K, Hirai H, Aizawa S and Ruoslahti E. (1996). Molec. R, Tanaka T, Hirano N, Mano H, Yazaki Y and Hirai H. Cell. Biol., 16, 2606 ± 2613. (1994). Oncogene, 9, 1669 ± 1678. Wang Y, Simonson MS, Pouyssegur J and Dunn MJ. (1992). Osterop APRM, Medema RH, Zon GCM, Bos JL, Moller W Biochem. J., 287, 589 ± 594. and Maassen JA. (1993). Eur. J. Biochem., 212, 477 ± 482. Wang LM, Myers MG, Sun X-L, Aaronson SA, White M Patti M-E, Sun X-J, Bruening JC, Araki E, Lipes MA, White and Pierce JH. (1993). Science, 261, 1591 ± 1594. MF and Kahn CR. (1995). J. Biol. Chem., 270, 24670 ± Waters SB, Yamauchi K and Pessin JE. (1993). J. Biol. 24673. Chem., 268, 22231 ± 22234. Pawson T and Schlessinger J. (1993). Current Biol., 3, 434 ± White MF. (1994). Curr. Opin. Genet. Devel., 4, 47 ± 54. 442. White MF, Maron R and Kahn CR. (1985). Nature, 318, Pruett W, Yuan Y, Rose E, Batzer AG, Harada N and 183 ± 186. Skolnik EY. (1995). Mol. Cell. Biol., 15, 1778 ± 1785. White MF and Kahn CR. (1994). J. Biol. Chem., 269, 1±4. Quon MJ, Butte AJ, Zarnowski MJ, Sesti G, Cushman SW and Taylor SI. (1994). J. Biol. Chem., 269, 27920 ± 27924.