Key role of Shc signaling in the transforming pathway ... - Nature

9 downloads 0 Views 419KB Size Report
unable to activate Shc: Ret/ptc2 long isoform (iso51)-. Y586F and both ... Received 23 November 2000; revised 7 March 2001; accepted 14. March 2001 ...
Oncogene (2001) 20, 3475 ± 3485 ã 2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $15.00 www.nature.com/onc

Key role of Shc signaling in the transforming pathway triggered by Ret/ptc2 oncoprotein Elena Mercalli1, Simona Ghizzoni1, Elena Arighi1, Luisella Alberti1, Romina Sangregorio1, Maria T Radice1, Mikhail L Gishizky2, Marco A Pierotti*,1 and Maria Grazia Borrello1 1 2

Department of Experimental Oncology, Research Unit #3, Istituto Nazionale Tumori, Via G. Venezian, 1 20133 Milan, Italy; SUGEN, 230 East Grand Ave, California, CA 94080 South San Francisco, USA

The RET/PTC oncogenes, generated by chromosomal rearrangements in papillary thyroid carcinomas, are constitutively activated versions of protoRET, a gene encoding two protein isoforms of a transmembrane tyrosine kinase receptor. By using Ret/ptc2 short isoform (iso9), we have previously demonstrated that Tyr586 (Tyr1062 of protoRet) is the docking site for both the PTB and the SH2 domains of Shc. To determine the relevance of this interaction for the transforming activity of Ret/ptc oncogenes, we have generated and characterized novel Ret/ptc mutants unable to activate Shc: Ret/ptc2 long isoform (iso51)Y586F and both isoforms of Ret/ptc2-N583A. These mutants neither activate Shc nor transform NIH3T3 cells. Since Tyr1062 shows features of a multifunctional docking site, we have used a Shc mutant (Shc Y317F) to directly assess Shc role. We have demonstrated that in our cell system Shc Y317F behaves like a dominant interfering mutant on the activation of the Grb2-Sos pathway by endogenous Shc triggered by Ret/ptc2. A strong reduction of the transforming activity of Ret/ptc2 in presence of this mutant was also demonstrated. Our data suggest that Shc activation play a key role in the transforming pathways triggered by Ret/ptc oncoproteins. Moreover, we have shown that coexpression of the Shc-Y317F mutant with Ret/ptc2 speci®cally causes apoptosis, and that the surviving cells lose the long-term expression of one of the two genes. Oncogene (2001) 20, 3475 ± 3485. Keywords: Ret; Shc; oncogene; thyroid carcinoma Introduction The RET gene encodes a transmembrane tyrosine kinase (Takahashi et al., 1988). RET is a component of a multiprotein complex that acts as a receptor for four closely related molecules: glial cell line-derived neurotrophic factor (GDNF), neurturin, persephin and artemin (Durbec et al., 1996; Trupp et al., 1996;

*Correspondence: MA Pierotti Received 23 November 2000; revised 7 March 2001; accepted 14 March 2001

Creedon et al., 1997; Milbrandt et al., 1998). RET was originally described as an oncogene (Takahashi et al., 1985) and it has been demonstrated to be involved, with distinct alterations, in di€erent pathologies. Activating mutations of RET are found in human thyroid carcinomas: chromosomal rearrangements in papillary thyroid carcinoma (PTC) (Pierotti et al., 1996) and point mutations in medullary thyroid carcinoma (MTC), in particular in almost all MTCs of familial origin, FMTC, MEN2A and MEN2B and in 20 ± 50% of the sporadic forms (Pasini et al., 1996). In contrast, inactivating mutations of RET cause Hirschsprung's disease (Pasini et al., 1995). RET rearrangements found in PTC, named RET/PTCs, as well as mutations in the extracellular part of RET found in MEN2A, lead to the formation of constitutively active homodimers (Bongarzone et al., 1993; Durick et al., 1995). Intracellular mutations found in MEN2B, instead, produce a kinase with changed substrate speci®city (Songyang et al., 1995). The RET gene encodes two protein isoforms of 1072 (short) or 1114 (long) aminoacids (Ishizaka et al., 1989) di€ering at the C-terminus region, by displaying nine or 51 unrelated aminoacids. Since for a long time Ret was an orphan oncogenic receptor tyrosine kinase, extensive studies on Ret signaling have been performed using chimeric and/or oncogenic forms of Ret (Borrello et al., 1994; Santoro et al., 1994; Pandey et al., 1995, 1996; Asai et al., 1996; Durick et al., 1996, 1998; Arighi et al., 1997; Lorenzo et al., 1997; van Weering and Bos, 1997; Alberti et al., 1998). The intracellular domain of Ret contains 14 tyrosine residues in the long isoform and 12 in the short one, the latter lacking two tyrosine residues at the Cterminus. Interactions of Ret with a variety of downstream targets have been identi®ed. Phosphorylated tyrosine residues Tyr905, Tyr1015, Tyr1062 and Tyr1096 were identi®ed as docking sites for the adapter proteins Grb7/Grb10, phospholipase Cg (PLCg), SHC/ Enigma/p85PI3K and Grb2, respectively (Durick et al., 1996; Liu et al., 1996; Borrello et al., 1996; Lorenzo et al., 1997; Arighi et al., 1997; Alberti et al., 1998; SegounCariou and Billaud, 2000). The SHC gene encodes three protein isoforms of 46, 52 and 66 kD arising from alternative splicing and translational start sites (Pelicci et al., 1992). Shc contains two phosphotyrosine interaction domains: an

Shc role on Ret/ptc2 signaling E Mercalli et al

3476

Oncogene

amino-terminal PTB domain and a carboxyl-terminal SH2 domain that ¯anks a proline/serine rich region, the collagen homology region (CH1). The Shc PTB domain speci®cally binds to a NXXpY motif, where p indicates phosphorylation (Kavanaugh and Williams, 1994; Trub et al., 1995), while the SH2 domain binds pYXXL/V/M sequences (Songyang et al., 1994). Shc proteins are ubiquitously expressed and are involved in signaling pathways triggered by di€erent classes of proteins: by tyrosine kinase (TK) receptors, by non receptor TKs such as Src and Fps, by surface receptors without intrinsic TK activity such as interleukin receptors and B-cell antigen receptor and by middle T antigen of Polyoma virus (see Arighi et al., 1997 and references therein). Upon phosphorylation, Shc binds to the Grb2-SOS complex leading to the activation of the Ras signaling pathway (Lowenstein et al., 1992). Recently the p66 isoform of Shc was demonstrated to control the oxidative stress response and life span in mammals (Migliaccio et al., 1999). We and others have shown that Shc is activated by Ret/ptc oncoproteins as well as by transmembrane forms of Ret: proto-Ret receptor activated by ligand and its oncogenic versions, RetMEN2A and RetMEN2B (Borrello et al., 1994; Arighi et al., 1997; van Weering and Bos, 1997; Lorenzo et al., 1997; Trupp et al., 1999). In addition, we have used Ret/ptc2 short isoform (iso9) (Bongarzone et al., 1993), a representative RET/PTC oncogene encoding a fusion protein in which Ret cytoplasmic domains are fused to the regulatory PKA subunit RIa, to identify the Shc docking site. We have demonstrated that Tyr586 (Tyr1062 of proto-Ret) is the docking site for both the PTB and the SH2 domains of Shc (Arighi et al., 1997; Lorenzo et al., 1997). Here we show, by using the same Ret/ptc chimeric protein, that Ret/ptc2-Y586F and Ret/ptc2-N583A mutants of both Ret isoforms, unable to activate Shc, also lack the capacity to transform NIH3T3 cells. Since Shc activation was shown in human tumors of di€erent origin, including a papillary thyroid carcinoma cell line (Pelicci et al., 1995) as well as in NIH3T3 cells expressing Ret/ptc2 (Borrello et al., 1994), Shc was supposed to be the obvious candidate responsible for the lack of transforming activity of Ret/ptc2 mutants. However, Tyr586 of Ret/ptc2 (Tyr1062 of protoRet) has been recently proposed also as a multifunctional docking site which binds additional signaling molecules such as Enigma and p85 of PI3K (Durick et al., 1996; Segoun-Cariou and Billaud, 2000). To precisely de®ne the role of Shc in cell transformation we have used a Shc mutant, Shc-Y317F, that we demonstrated able to interfere with Shc activation of Grb2-SOS triggered by Ret. This Shc mutant, coexpressed with Ret/ptc2, signi®cantly reduced the transforming activity of Ret/ptc2 for NIH3T3 cells. Moreover, we have shown that coexpression of the Shc-Y317F mutant with Ret/ptc2 speci®cally causes apoptosis. The surviving cells showed long term loss of expression of one of the two exogenous genes, more frequently of Ret/ptc2.

Results In vitro interaction of Ret/ptc2-Y586F and Ret/ptc2-N583A mutants with SH2 and PTB domains of Shc To characterize Ret-Shc interaction, we have generated, by site-directed mutagenesis of RET/PTC2 cDNAs encoding short or long isoform, the following mutants: RET/PTC2-Y586F long isoform, RET/PTC2Y586F short isoform (Arighi et al., 1997), RET/PTC2N583A short and long isoforms. All the mutations are illustrated in Figure 1a. The RET/PTC2 mutants were sequenced and cloned into the pRC-CMV eukaryotic expression vector carrying the G418 resistance gene. All the constructs were transiently expressed in COS7 cells in order to verify the synthesis of the correctly sized proteins and to compare biochemical properties of mutant and wt proteins. As shown in Figure 1b, all the constructs expressed the Ret/ptc2 tyrosine phosphorylated proteins of the expected size, p76 for the short isoform and p81 for the long one (upper panel). In addition, wt and mutant proteins showed a comparable autophosphorylation activity in immunokinase assay (lower panels).

Figure 1 Schematic representation of RET/PTC2-iso9 and iso51 isoforms carrying the N583A and the Y586F mutations and biochemical analysis of their products. (a) Schematic representation of the two isoforms of RET/PTC2 showing the mutations converting Asparagine 583 to Alanine residue and Tyrosine 586 to Phenylalanine residue. (b) Tyrosine phosphorylation of Ret/ ptc2wt- iso9 (p76) and iso51 (p81), Ret/ptc2-N583A (lane 583A) iso9 and iso51 and Ret/ptc2-Y586F (lane 586F) iso9 and iso51 proteins (upper panel). In vitro immunocomplex kinase assay of Ret/ptc2wt, Ret/ptc2-Y586F and Ret/ptc2-N583A iso9 and iso51 isoforms transiently expressed in COS7 cells. The autophosphorylation of Ret/ptc2 is shown in duplicate in the lower panels

Shc role on Ret/ptc2 signaling E Mercalli et al

To analyse the ability of Ret/ptc2 mutants to bind PTB and SH2 Shc domains, protein extracts from transiently transfected COS7 cells were used for pull down assays with PTB or SH2 domain Gst-fused proteins. As shown in Figure 2a, not only Y to F but also N to A substitutions completely abrogate the ability of Ret/ptc2 short and long isoforms to bind the PTB domain. In contrast, binding to the SH2 domain was abrogated only by Y586F mutations, in keeping with the known SH2 consensus sequence. Shc coimmunoprecipitation and phosphorylation in the presence of Y586F or N583A Ret/ptc2 mutants The ability of the mutants to bind and activate Shc in living cells was then analysed. Coimmunoprecipitation

experiments are shown in Figure 2b: protein extracts from 293T cells transiently transfected with each mutant or Ret ptc2-wt genes were immunoprecipitated with anti-Shc and blotted with anti-phosphotyrosine or anti-Ret antibodies to verify Shc phosphorylation and coimmunoprecipitation with Ret/ptc2 phosphorylated mutant isoforms. A strong reduction of coimmunoprecipitation is shown not only for Ret/ptc2-Y586F but also for Ret/ptc2-N583A mutants that show only faint residual Tyr-phosphorylated Ret/ptc2 bands compared with the corresponding Ret/ptc2-wt proteins. This result con®rms the major role of the PTB domain for Shc-Ret/ptc interaction (Arighi et al., 1997). The reduction was observed with both the short and the long isoform of Ret/ptc2. In addition, the same blot shows that Ret/ptc2 induced Shc phosphorylation on

3477

Figure 2 Interaction and activation of Shc by Ret/ptc2-Y586F and Ret/ptc2-N583A in vitro and in living cells. (a) Protein extracts from COS7 cells transiently transfected with Ret/ptc2wt, Y586F and N583A iso9 and iso51 isoforms (p76 and p81 respectively) were used in a pull down assay performed with Sepharose bound Gst-SH2 or Gst-PTB Shc fusion proteins. The same extracts were immunoprecipitated with anti-Ret antiserum and immunoblotted with anti-phosphotyrosine antiserum, as control. (b) Phosphorylation and coimmunoprecipitation of Shc with Ret/ptc2-Y586F and N583A mutants. 293T cells were transiently transfected with the indicated Ret/ptc2 mutants. About 1 mg of protein extracts were immunoprecipitated with anti-Shc and immunoblotted with anti-phosphotyrosine or anti-Ret antibodies. As control, the same extracts were immunoprecipitated and immunoblotted with anti-Shc or anti-Ret antibodies Oncogene

Shc role on Ret/ptc2 signaling E Mercalli et al

3478

tyrosine is strongly and similarly reduced in the presence of both Y586F and N583A mutations (Figure 2b). Transforming activity on NIH3T3 cells of Ret/ptc2-Y586F and Ret/ptc2-N583A mutants The above described Ret/ptc2 mutants, unable to interact with and to phosphorylate Shc, were next analysed by NIH3T3 focus formation assay, to evaluate their oncogenic activity. As a control, the previously characterized Ret/ptc2-iso9-K282R, a kinase negative Ret/ptc2 mutant, was used (Arighi et al., 1997). Figure 3 shows representative plates of focus formation experiments; the results are reported in Table 1. No transformation foci were obtained using Ret/ptc2-Y586F short or long isoform. The transforming activity, if any, was less than 1/100 of that displayed from the wt. Surprisingly, the same results

were obtained using Ret/ptc2-N583A mutants. All these mutants a€ecting the Tyr586 docking site in our experimental conditions behaved identically to the kinase-dead Ret/ptc2 mutant. Shc-Y317F mutant interferes in Ret/ptc2-dependent Grb2/SOS pathway activation To investigate the role of Shc in the transforming pathways triggered by Ret/ptc2, we have used Shc mutants that were previously demonstrated to have a dominant interfering e€ect on the activation of endogenous Shc (Hill et al., 1996). To this aim, we have used di€erent hemoagglutinin (HA)-tagged p52Shc genes: SHC-WT, SHC-Y317F mutant (the Tyr317 residue is involved in Grb2 interaction), SHCR401K (SH2 domain inactivating mutant), SHC-DPI (the PTB domain deletion) represented in Figure 4a. The genes, cloned into the pCGN vector harboring

Figure 3 N583A and Y586F mutations completely abolish the Ret/ptc2 transforming activity. Focus formation assay in NIH3T3 cells stably transfected with Ret/ptc2 short and long isoform wt and mutants (N583A and Y586F). As negative control, a Ret/ptc2 mutant devoid of kinase activity (K282R) (described in Arighi et al., 1997) was used. Representative plates ®xed and stained 14 days upon transfection are shown Oncogene

Shc role on Ret/ptc2 signaling E Mercalli et al

3479

Table 1 Biological activity of RET/PTC2-N583a and RET/PTC2-Y586f on NIH3T3 cells Transfected plasmid (500 ng)

Transformed foci (N/mg DNA)

Transformed foci/ G418R colonies

Relative transforming activity

PTC2wt-iso9 PTC2/N583A-iso9 PTC2/Y586F-iso9 PTC2/K282R-iso9

2.016102 51 51 51

2.861072 50.0161072 50.0161072 50.0161072

1 50.01 50.01 50.01

PTC2wt-iso51 PTC2/N583A-iso51 PTC2/Y586F-iso51

36102 51 51

3.261072 50.0161072 50.0161072

1 50.01 50.01

NIH3T3 G418 colonies were ®xed at day 14 and transformation foci at day 21 after transfection. The experiment was repeated three times, in triplicate

resistance to hygromycin, were used in transient transfections. As shown in the middle panel of Figure 4b they were all expressed when transiently transfected in 293T cells. All but the Shc-DPI mutant show a small electrophoretic shift, due to the HA epitope, compared with endogenous p52 Shc; Shc-DPI encodes a protein of about 38 kD. When cotransfected with Ret/ptc2 (lower panel, Figure 4b), Shc-wt and Shc-Y317F proteins are able to coimmunoprecipitate with the oncoprotein whereas Shc-R401K and Shc-DPI show, as expected, a reduced ability to coimmunoprecipitate with Ret/ptc2. This is shown in the upper panel of Figure 4b, where about a residual 40% (for ShcR401K) and 10% (for Shc-DPI) of the amount of the Shc-wt, respectively, was assessed to coimmunoprecipitate by PhosphorImager analysis. All the Shc-wt and Shc mutants reduce at variable degrees the phosphorylation of endogenous Shc (upper panel, Figure 4c). Noteworthy, the mutants, in spite of lacking PTB (ShcDPI) or SH2 (Shc-R401K) domains, show, when overexpressed, a signi®cant steady state phosphorylation. To determine which mutant is able to block the Shc pathway, the ability of all the Shc mutants to coimmunoprecipitate with Sos was analysed. They were all binding Sos (data not shown), except Shc-Y317F mutant, which showed a strongly reduced ability to coimmunoprecipitate with Sos compared to Shc-wt (upper panel, Figure 5a). In addition, the coimmunoprecipitation of endogenous Shc with Sos seems lowered by the presence of Shc-Y317F mutant, as shown in the middle panel of Figure 5a by blotting with anti-Shc the anti-Sos immunoprecipitates. Accordingly, Erk activation induced by Ret/ptc2 coexpressed with Shc-Y317F is signi®cantly reduced compared to Erk 1/2 activation induced by Ret/ptc2 and Shc-wt (upper panel, Figure 5b), thus suggesting that Shc signaling is partially blocked by overexpressing Shc-Y317F in 293T cells. Transforming activity on NIH3T3 cells of Ret/ptc2 in the presence of Shc-Y317F dominant interfering mutant To determine the transforming activity of when the Shc pathway is partially blocked, Y317F mutant, demonstrated above to dominant negative activity on endogenous

Ret/ptc2 the Shcexert a Shc, or

Shc-wt, were coexpressed with Ret/ptc2 in NIH3T3 cells. To this aim, ®rstly NIH3T3 cell lines expressing Shc mutants or containing empty pCGN vector were isolated and demonstrated to stably express the transfected SHC genes (data not shown). All the NIH3T3 clones named NIH (vector), containing empty pCGN vector, NIH (Shc-wt), expressing Shc-wt, and NIH (Shc-Y317F) expressing Shc-Y317F mutant, were demonstrated to be similarly transfection competent as the parental NIH3T3 cells by using a b-gal reporter plasmid (data not shown). Standard NIH3T3 as well as the above described Shc expressing and control vector NIH3T3 clones were then used as recipient for Ret/ ptc2wt focus formation assay stable transfections. The results of repeated transfections are shown in Figure 6. A reduction of 70 ± 80% of the transforming activity of Ret/ptc2 is displayed when Ret/ptc2 is transfected in NIH (Shc-Y317F) cells, compared to NIH or NIH (vector) cells. A minor reduction of Ret/ ptc2 transforming activity was observed using a recipient NIH3T3 expressing Shc-wt or Shc-DPI. Comparable results were obtained by scoring foci in standard conditions (see Materials and methods) or by selecting cells in G418R or hygromycin, the resistances carried by either of the two vectors used. Foci selected in hygromycin were isolated from NIH (vector), NIH (Shc-wt) and from NIH (Shc-Y317F). Whereas cell lines stably coexpressing Ret/ptc2 and Shc-wt or Ret/ptc2 expressing cells from NIH (vector) were easily obtained (nine positive out of 15 analysed), Ret/ptc2 NIH (Shc-Y317F) stably coexpressing cells were rare (one positive out of 13 analysed). For this reason, starting from early time of transfection (fourth week), we checked the following parameters in the isolated foci once a week: expression of exogenous genes by Western Blot, cell morphology and apoptosis, assayed by TUNEL. The results are summarized in Table 2. The six cell lines derived each from NIH (vector) and NIH (Shc-wt) all displayed a morphologically transformed phenotype, were able to maintain the expression of exogenous genes and showed no positivity in TUNEL assays during the 5 weeks time period of analysis. Instead, the cell lines derived from NIH (ShcY317F) transfected with Ret/ptc2 had a partially Oncogene

Shc role on Ret/ptc2 signaling E Mercalli et al

3480

Figure 5 Sch-Sos coimmunoprecipitation and Erk 1/2 activation in cells transiently expressing Ret/ptc2 and Shcwt or ShcY317F mutant. (a) 293T cells were transiently co-transfected with Ret/ ptc2-iso9 and the indicated Shc genes (as described in Figure 4 legend) or mock transfected with empty vectors (lanes mock). One mg of protein extracts were immunoprecipitated with anti-Sos and immunoblotted with anti-Shc or anti-HA antibodies. (b) 293T cells were transiently co-transfected and 50 mg of total lysates were immunoblotted with anti-phospho Erk1/2 (upper panel) and, as control, anti Erk1/2 (lower panel) antibodies

Figure 4 Coimmunoprecipitation of Ret/ptc2wt-iso9 with Shc mutants and endogenous Shc phosphorylation. (a) Schematic representation of SHC mutants. (b) Coimmunoprecipitation of Ret/ptc2wt-iso9 (ptc2wt-iso9) with Shc mutants. COS7 cells were transiently co-transfected with Ret/ptc2wt-iso9 (1 mg) and the indicated Shc genes (5 mg) or mock transfected with empty vectors (lanes mock). One mg of protein extracts were immunoprecipitated with anti-Shc antiserum and immunoblotted with anti-Ret or anti HA antibodies to detect exogenous HA-tagged Shc proteins. As control, the same extracts were analysed for the expression of Shc and Ret. Endogenous Shc proteins (p46 and p52), as well as exogenous Shc proteins (p53 and p38), are indicated. (c) Phosphorylation of the endogenous Shc in the presence of Shc mutants. 293T cells were transiently cotransfected (as described above) and 0.8 mg of protein extracts were immunoprecipitated with anti-Shc and immunoblotted with anti-phosphotyrosine antibodies. Direct anti-Shc Western blot and blot anti-Ret on anti-Ret immunoprecipitates were performed as controls

revertant ¯at morphology at the ®fth week, becoming ¯atter and growing in a more organized fashion in the following weeks. Two representative cell lines coexpressing Ret/ptc2 and Shc-Y317F proteins are shown in Figure 7. In the same Figure, parental cell lines, as well as Ret/ptc2 or Ret/ptc2 and Shc-wt coexpressing cells, are shown for comparison. In addition, from week 5 to 9 from the transfection, cells coexpressing Oncogene

Ret/ptc2 and Shc-Y317F proteins showed a reproducible TUNEL positive staining, ranging from 2 ± 5%, 10 ± 15% at their peak, while parental NIH (ShcY317F) and NIH (Ret/ptc2) cell lines, assayed as controls, were always negative by TUNEL assay (Table 2). The same cell lines coexpressing Ret/ptc2 with Shcwt or with Shc-Y317F or with empty vector were also early analysed (at week 6) to test the activation of Shc and, downstream, of Erk1/2. In Figure 8 it is shown that, as for the 293T cell context, also in NIH3T3 cells the Y317F mutation impaired the coimmunoprecipitation of Shc with Sos (upper panel), thus blocking the Shc pathway. Cell lines coexpressing Shc-Y317F mutant and Ret/ptc2 showed also a strong reduction in Erk 1 and 2 tyrosine phosphorylation compared with NIH (vector) cells expressing Ret/ptc2 or with cells coexpressing Ret/ptc2 and Shc-wt (Figure 8, direct Western blot with anti phospho-Erk 1/2 antibodies). Discussion RET/PTC oncogenes, generated by somatic rearrangements in a signi®cant fraction of human papillary thyroid carcinomas (Pierotti et al., 1996), are capable of transforming NIH3T3 cells (Bongarzone et al., 1989) and are thought to be associated with the pathogenetic

Shc role on Ret/ptc2 signaling E Mercalli et al

mechanisms a€ecting the original thyroid cells (Fischer et al., 1998). With the e€ort to understand the mechanism of transformation triggered by RET/PTC

Figure 6 Ret/ptc2 transforming activity on NIH3T3 cells expressing Shc mutants. Transforming activity of Ret/ptc2 on NIH3T3 cells expressing Shc mutants. NIH3T3 cells transfected with Ret/ptc2-iso9 were selected in Hygromycin (25 mg/ml) or in G418 (0.5 mg/ml) when indicated, and ®xed at day 21 after transfection. Each bar is the average of nine plates from three independent DNA precipitations. Standard deviations are indicated

oncogenes, we have investigated the role of Shc, previously demonstrated constitutively activated in cells expressing Ret/ptc oncoproteins (Borrello et al., 1994; Pelicci et al., 1995). We have generated mutants of a representative RET/PTC oncogene, RET/PTC2, and analysed their transforming activity. Our results show that Ret/ptc2Y586F, a mutant lacking the tyrosine docking site for SH2 and PTB domains of Shc and therefore unable to activate Shc (Arighi et al., 1997), completely loses the ability to transform NIH3T3 cells. Because of redundancy in signaling pathways, mutations of individual phosphorylation sites in a receptor tyrosine kinase do not usually abolish a speci®c response (Valius et al., 1993; Fambrough et al., 1999). Instead, a strong reduction of transforming activity of Tpr-Met due to a Y to F mutation at a multifunctional docking site was reported (Fixman et al., 1995). In the case of Ret, Tyr586 of Ret/ptc2 (Tyr1062 of protoRet) has been recently proposed to be a multifunctional docking site, binding Shc, Enigma and p85 of PI3K (Arighi et al., 1997; Durick et al., 1996; SegounCariou and Billaud, 2000). However, not even recent studies concerning the Tyr1062 of Ret and the transducer molecules binding this crucial docking site (Segoun-Cariou and Billaud, 2000; Besset et al., 2000; Hennige et al., 2000; Hayashi et al., 2000) have formally demonstrated that Shc is essential for transformation. To better understand the role of Shc, we have therefore constructed the RET/PTC2N583A mutant. We have shown that Ret/ptc2-N583A short and long isoform mutants are unable to bind Shc-PTB and to activate Shc. Moreover, these mutants have completely lost their transforming activity. The correlation between the lack of Shc activation and the absence of transforming activity on NIH3T3 cells of Y586F and N583A Ret/ptc2 mutants strongly suggests a key role of Shc in the transforming activity of Ret. In keeping with present ®ndings, Shc activation was found in human tumors of di€erent origin including a papillary thyroid carcinoma cell line (Pelicci et al., 1995). Moreover, Shc activation mediated by Ret/ptc2 has been demonstrated in NIH3T3 cells (Borrello et al., 1994; Arighi et al., 1997). Our data do not exclude however the role of other possible Tyr586-binding signal transducers having the same Asparagine requirement in position 73 to Tyr.

3481

Table 2 Biological and biochemical analysis of NIH3T3 cell lines coexpressing RET/PTC2 and SHC genes 48/58 week Ret expr.

Shc expr.

Morph.

NIH (ShcY317F)+Ret/ptc2S

9/10

10/10

NIH (Shcwt)+Ret/ptc2S NIH (vector)+Ret/ptc2S

6/6 6/6

6/6 ±

5/10 F 5/10 I 6/6 T 6/6 T

88/98 week

Tunel positivity

Ret expr.

Shc expr.

Morph.

Tunel positivity

8/9

5/9

8/9

9/9 F

7/9

0/6 0/6

6/6 6/6

6/6 ±

6/6 T 6/6 T

0/6 0/6

F=¯at. T=transformed. I=intermediate ¯at/transformed Oncogene

Shc role on Ret/ptc2 signaling E Mercalli et al

3482

Figure 7 Morphology of NIH3T3 cell lines expressing Shc mutants and Ret/ptc2wt-iso9. NIH3T3 containing empty pCGN (NIH3T3 (vector)), expressing Shc-Y317F mutant (NIH (Shc-Y317)) or Shc-wt (NIH (Shc-wt)) recipient cells were transfected with Ret/ptc2wt-iso9. The foci, all expressing the transfected genes, were photographed. Scale bar=50 mm

Figure 8 Shc-Sos coimmunoprecipitation and Erk 1/2 activation in NIH3T3 foci expressing Ret/ptc2wt-iso9 and Shc mutants. Foci were serum starved and 1 mg of protein extracts were immunoprecipitated with anti-SOS and immunoblotted with anti-HA antiserum (upper panel). As control, the same extracts were immunoprecipitated and analysed for the expression of Ret/ ptc2 and Shc mutants (2nd and 3rd panels). Fifty mg of the total lysates were resolved in SDS ± PAGE and immunoblotted with anti-phospho Erk1/2 antiserum (4th panel), or anti-Erk 1/2 antiserum, as control (lower panel) Oncogene

It was recently shown by Segoun-Cariou and Billaud (2000) that Tyr1062 is the docking site of p85 of PI3K and that the binding is probably mediated by an unidenti®ed protein. In addition, Segoun-Cariou and Billaud, 2000 proved that the PI3K pathway is essential for the transforming activity of Ret(MEN2A)1062F. Therefore, to discriminate between the contribution of Shc to the Ret/ptc2 transforming pathway and that of other possible signal transducers, we have used Shc mutants and shown that all the Shc proteins analysed are able to interact with Ret/ptc2 and to compete with endogenous Shc, when overexpressed. This was true for Shc-wt and Shc-Y317F having both phosphotyrosine binding domains, but unexpectedly also for Shc-R401K, a mutant with an inactive SH2 domain, and for Shc-DPI, lacking the PTB domain. The Shc mutants displaying only one phosphotyrosine binding domain are able to trigger Grb2-Sos pathway and to activate downstream Erk1 and Erk2 similarly or slightly less than endogenous Shc-wt in the presence of Ret/ptc2 (data not shown). However, when Shc-Y317F mutant was coexpressed with Ret/ptc2, Shc coimmunoprecipitation with Sos was almost completely abrogated. Accordingly, Ret/ ptc2-dependent Erk activation was lowered by ShcY317F expression. These results suggest that in our cell system, the Tyr317 of Shc has a major role in coupling Ret signaling to the Grb2-Sos pathway. Depending on the cell environment and on the proteins that trigger Shc activation, Shc Tyr317 and Shc Tyr239/240

Shc role on Ret/ptc2 signaling E Mercalli et al

mutants seem to have di€erent roles. For example, Nerve Growth Factor (NGF)-induced neurite outgrowth in pheochromocytoma PC12 cells requires phosphorylation on Tyr239/240 but not on Tyr317 (Thomas and Bradshaw, 1997). For Epidermal Growth Factor (EGF)-induced mitogenic signaling in NIH3T3 cells, the ShcY317F mutant has a much reduced Grb2 coimmunoprecipitation and MAPK activation in response to EGF whereas the Shc-239/240 mutant is able to fully activate MAPK pathway (Gotoh et al., 1997). Therefore, the Shc-Y317F mutant might represent a way to selectively block the Shc pathway triggered by Ret in NIH3T3 cells. Since the Tyr586 is a multifunctional docking site, overexpression of ShcY317F protein might compete not only with endogenous Shc but also with other signal transducers binding the same docking site. Consequently, we have used Shc-wt as control to dissect a genuine dominant negative e€ect of Shc-Y317F in Ret/ptc2-induced NIH3T3 transformation. Coexpression of Shc-wt with Ret/ptc2 produces about a 20 ± 25% of reduction in the number of foci whereas Shc-Y317F increases this value to about a 70 ± 80% of reduction. These results suggest that competition of exogenous Shc with other transducers occurs with a minor e€ect whereas the major e€ect is due to the block of Grb2-Sos pathway, also lowering Erk activation, induced by Shc-Y317F mutant. It is worth noting that, by using dominant interfering Shc-Y317F mutant, we have obtained a reduction of Ret/ptc2 transforming activity similar to the reduction found by Segoun-Cariou and Billaud', (2000) by coexpressing a dominant interfering PI3K mutant with Ret(MEN2A) (Segoun-Cariou and Billaud, 2000). This supports the notion that cellular transformation by di€erent receptor-tyrosine kinase (RTK) oncoproteins results from integration of multiple signaling pathways (Porter and Vaillancourt, 1998). On this view, dominant interfering mutants, unbalancing intracellular pathways, may cause e€ects that are not additive. Alternatively, di€erent pathways (Shc versus PI3K) could have a key role for di€erent oncogenic Ret proteins (Ret/ptc2 versus RetMEN2A). Interestingly, we have found that the morphology of foci, however arising in low number by transfecting Ret/ptc2 on NIH(Shc-Y317F) cells, was not comparable to that of typical Ret/ptc2 foci, the former ones being ¯atter and more organized. Moreover, they tended to lose Ret/ptc2 expression during cell propagation. Unexpectedly, we have found, in fact, a low but reproducible and speci®c apoptotic e€ect (as judged by TUNEL assay) in NIH(Shc-Y317F)-Ret/ptc2 cells, whereas no such e€ect was detected in parental NIH(Shc-Y317F) cell lines that were stable for months, and in typical Ret/ptc2 foci arising on NIH cells. Therefore, the apoptotic e€ect was most probably caused by the coexpression of SHC-Y317F and RET/ PTC2 genes. Although cell proliferation, essential for tumorigenic transformation, and cell death are opposite cellular processes, it is known that a strong oncogenic stimulus to proliferate is associated with an increased sensitivity to induction of apoptosis (re-

viewed in Guo and Hay, 1999). We can speculate that the unbalance in the signaling pathways arisen by Shc mutant overexpression causes the predominance of death signaling, usually masked by proliferating pathways. This hypothesis, with intriguing therapeutic implications for Ret expressing tumors, needs however to be carefully investigated, both to de®ne the biochemical mechanisms underlying this phenomenon and to study Shc dominant negative mutants e€ects in the appropriate tumor cell environment. Recently, it was reported that Ret in the absence of ligand or Ret carrying HSCR-associated mutations induces apoptosis (Bordeaux et al., 2000). However, the same authors have found that the constitutively activated RetC634R (the MEN2A-associated Ret mutant) does not show any pro-apoptotic activity. This result furthermore supports the hypothesis that the constitutively activated-ligand independent Ret/ptc2 may induce apoptosis in the presence of Shc-Y317F mutant with a di€erent mechanism. Reduction of transforming activity and apoptosis induction were also found by coexpressing SHC-Y317F mutant with the TRKT3 oncogene, a rearranged version of NGF receptor (A Greco, personal communication) activated, alternatively to Ret, in papillary thyroid carcinomas (Pierotti et al., 1996). We have also demonstrated that cells coexpressing Ret/ptc2 and Shc-Y317F mutant show a lower Erk1/ Erk2 activation compared with cells expressing Ret/ ptc2. This e€ect was not surprising since it is known that Shc, by recruiting Grb2-Sos complex, contributes to the activation of Erk pathways by tyrosine kinases (Schlessinger, 1993). In our case, a sustained Erk1 and Erk2 activation characterizes Ret/ptc2 induced NIH3T3 transformation, while a low level of ERK phosphorylation correlates with the `less transformed' phenotype of NIH(Shc-Y317F)-Ret/ptc2 cells. In conclusion, the present data strongly indicate that Shc activation plays a key role in the transforming pathway(s) triggered by Ret/ptc2 oncoproteins. Further investigation is required to better understand the mechanisms leading cells coexpressing Shc-Y317F and Ret/ptc2 to undergo apoptosis or to segregate one of the two exogenous genes.

3483

Materials and methods Site-directed mutagenesis and cloning of RET/PTC2 mutants The cloning of the RET/PTC2-iso9 and iso-51 cDNAs has been previously reported (Bongarzone et al., 1993; Arighi et al., 1997). Site directed mutagenesis was performed on pRCCMV carrying RET/PTC2 WT iso9 or iso51 cDNA, using an in vitro Oligonucleotide Mutagenesis System (Promega). The construction of the RET/PTC2-iso9Y586F mutant is described in Arighi et al. (1997). Oligonucleotide carrying Y586F for the long isoform was 5'-GAAAACAAACTCTTTGGCATGTCAGACC-3'. Oligonucleo-tide carrying N583A was 5'-ACATGGATTGAAGCCAA-ACTCTATGG3' for both isoforms. Oncogene

Shc role on Ret/ptc2 signaling E Mercalli et al

3484

Mutant clones were identi®ed by Allele Speci®c Oligonucleotide (ASO) technique using reverse complement oligonucleotide carrying wt (5'-ATTGAAAACAAACTCTAT-3') or N583A (5'-ATTGAAGCCAAACTCTAT-3') sequence or the corresponding wt (5'-AAACTCTATGGCATGTCA-3') or Y586F (5'-AAACTCTTTGGCATGTCA-3') sequences. Full length RET/PTC2-iso9, RET/PTC2-iso51N583A and RET/ PTC2-iso51Y586F cDNA clones (inserted in pRC-CMV plasmid) were entirely sequenced. Cell cultures and transfections NIH3T3, COS7 and 293T cells were grown in DMEM containing 10% calf serum. NIH3T3 (26105 per 10 cm diameter dish) cells were stably transfected by calcium phosphate coprecipitation using 500 ng of plasmid DNA and 40 mg of mouse carrier DNA. G418-resistant colonies were selected in DMEM plus 10% calf serum and G418 antibiotic (0.5 mg/ml); hygromycin resistant colonies were selected in hygromycin antibiotic (25 mg/ml). Transformation foci were selected in DMEM containing 5% calf serum and hygromycin (25 mg/ml) or G418 (0.5 mg/ ml) and ®xed and counted or isolated for further studies 3 weeks after transfections. NIH3T3 resistant colonies and transformation foci were ®xed and counted or isolated 14 ± 21 days respectively after transfection. NIH3T3 cells were starved (when indicated) in 0.2% calf serum medium for 20 h before the extraction. Transient transfections in COS7 cells were performed by DEAE-dextran method (Arighi et al., 1997). Transient transfections in 293T cells (16106 cells per 10 cm diameter dish) were performed by calcium phosphate coprecipitation method with 1 ± 5 mg of plasmid DNA. 293T cells were starved in serum-free medium for 12 h before the extraction. In vitro binding experiments using Gst-fused proteins Bacterial cultures expressing pGEX vector alone or recombinant pGEX containing the SH2 or the PTB domain of Shc were grown in LB containing 100 mg ml71 ampicillin and induced with 1 mM isopropyl b-D-thiogalactopyranoside for 3 ± 6 h. The GST-fused proteins were puri®ed as described in Arighi et al. (1997). The clari®ed cells lysates, obtained as described in Borrello et al. (1994) were incubated with 10 ± 20 mg immobilized Gst or Gst-fused proteins for 60 min at 48C. Protein complexes were resolved by SDS ± PAGE and transferred to nitrocellulose. Blots were probed as described in Borrello et al. (1996). In vitro immunocomplex kinase assay The anti-Ret immunoprecipitates, adsorbed on protein ASepharose beads, were washed twice with lysis bu€er, once with incubation bu€er (50 mM HEPES pH 7.2, 20 mM MgCl2, 5 mM MnCl2), and incubated for 15 min at 48C in 20 ml of the same bu€er containing 0.5 mM DTT, 4 mCi of [g-32P]ATP (5000 Ci mmol71, Amersham) diluted with

unlabeled ATP. The reaction was stopped by addition of concentrated Laemmli bu€er. Proteins were eluted and subjected to 15% SDS ± PAGE. 32P-labeled bands were revealed by autoradiography of the dried gel and analysed by PhosphorImager. Immunoprecipitation and Western blotting Protein samples were prepared as reported in Borrello et al., 1996 and immunoprecipitated with the speci®ed antibodies: anti-Ret common antiserum (Borrello et al., 1996), anti-Shc rabbit polyclonal antiserum (UBI), anti-SOS1 rabbit polyclonal antiserum (UBI), and anti-HA mouse monoclonal antiserum (BabCo). One to two milligrams of protein extracts was used for binding experiments. Immunoprecipitates were resolved by SDS ± PAGE, transferred onto nitrocellulose ®lters, and immunoblotted with the anti-Ret antiserum described in Borrello et al., 1996, anti-pTyr (UBI), anti-Shc (UBI), anti-HA (BabCo). For total cells lysates (used for the analysis of Erk1/2 phosphorylation), cells were lysed in SDS lysis bu€er (62.5 mM Tris-HCl, pH 6.8, 2% SDS). Total extracts were separated by SDS ± PAGE. The anti Erk1/2 antibody and the anti-phospho Erk1/2 antibody used for the immunoblotting were from New England Biolabs. Immunoreactive bands were followed by autoradiography, using a horseradish peroxidase conjugate anti-rabbit or antimouse antiserum and ECL detection system (Amersham) or visualized using 125I-labeled protA (Amersham). When [125I]protA was used, the ®lters were exposed to storage phosphor screen ®lms and analysed with PhosphorImager apparatus (Molecular Dynamics) in order to quantify the c.p.m. associated to the Ret-speci®c bands. Analysis of apoptosis Early detection of apoptosis in NIH3T3 cell clones was accomplished by utilizing the In Situ Cell Death Detection Kit (Boehringer). The terminal deoxynucleotidiyl transferasemediated dUTP nick end-labeling (TUNEL) process was used to incorporate ¯uorescein at free 3'OH DNA ends of apoptotic cells. TUNEL reactivity was examined with a ¯uorescence microscopy. The number of TUNEL-positive cells was estimated by counting at least 100 cells from six separate ®elds.

Acknowledgments We are grateful to Angela Greco and Domenico Delia for helpful discussions. We thank Maria Grazia Rizzetti for excellent technical assistance and Cristina Mazzadi for secretarial assistance. This study was supported by the Associazione Italiana per la Ricerca sul Cancro (AIRC), by Italy/USA A81A10 and by the CNR special project 900522.PF49. E Mercalli is a recipient of an AIRC fellowship.

References Alberti L, Borrello MG, Ghizzoni S, Torriti F, Rizzetti MG and Pierotti MA. (1998). Oncogene, 17, 1079 ± 1087. Arighi E, Alberti L, Torriti F, Ghizzoni S, Rizzetti MG, Pelicci G, Pasini B, Bongarzone I, Piutti C, Pierotti MA and Borrello MG. (1997). Oncogene, 14, 773 ± 782. Oncogene

Asai N, Murakami H, Iwashita T and Takahashi M. (1996). J. Biol. Chem., 271, 17644 ± 17649. Besset V, Scott RP and Ibanez CF. (2000). J. Biol. Chem., 275, 39159-39166.

Shc role on Ret/ptc2 signaling E Mercalli et al

Bongarzone I, Monzini N, Borrello MG, Carcano C, Ferraresi G, Arighi E, Mondellini P, Della Porta G and Pierotti MA. (1993). Mol. Cell. Biol., 13, 358 ± 366. Bongarzone I, Pierotti MA, Monzini N, Mondellini P, Manenti G, Donghi R, Pilotti S, Grieco M, Santoro M, Fusco A, Vecchio G and Della Porta G. (1989). Oncogene, 4, 1457 ± 1462. Bordeaux MC, Forcet C, Granger L, Corset V, Bidaud C, Billaud M, Bredesen DE, Edery P and Mehlen P. (2000). EMBO J., 19, 4056 ± 4063. Borrello MG, Pelicci G, Arighi E, De Filippis L, Greco A, Bongarzone I, Rizzetti MG, Pelicci PG and Pierotti MA. (1994). Oncogene, 9, 1661 ± 1668. Borrello MG, Alberti L, Arighi E, Bongarzone I, Battistini C, Bardelli A, Pasini B, Piutti C, Rizzetti MG, Mondellini P, Radice MT and Pierotti MA. (1996). Mol. Cell. Biol., 16, 2151 ± 2163. Creedon DJ, Tansey MG, Baloh RH, Osborne PA, Lampe PA, Fahrner TJ, Heuckeroth RO, Milbrandt J and Johnson Jr EM. (1997). Proc. Natl. Acad. Sci. USA, 94, 7018 ± 7023. Durbec P, Marcos-Gutierrez CV, KilKenny C, Grigoriou M, Wartiowaara K, Suvanto P, Smith D, Ponder B, Costantini F, Saarma M, Sariola H and Pachnis V. (1996). Nature, 381, 789 ± 793. Durick K, Yao VJ, Borrello MG, Bongarzone I, Pierotti MA and Taylor SS. (1995). J. Biol. Chem., 270, 24642 ± 24645. Durick K, Wu RY, Gill GN and Taylor SS. (1996). J. Biol. Chem., 271, 12691 ± 12694. Durick K, Gill GN and Taylor SS. (1998). Mol. Cell. Biol., 18, 2298 ± 2308. Fambrough D, McClure K, Kazlauskas A and Lander ES. (1999). Cell, 97, 727 ± 741. Fischer AH, Bond JA, Taysavang P, Battles OE and Wynford-Thomas D. (1998). Am. J. Pathol., 153, 1443 ± 1450. Fixman ED, Naujokas MA, Rodrigues GA, Moran MF and Park M. (1995). Oncogene, 10, 237 ± 249. Gotoh N, Toyoda M and Shibuya M. (1997). Mol. Cell. Biol., 17, 1824 ± 1831. Guo M and Hay BA. (1999). Curr. Opin. Cell. Biol., 11, 745 ± 752. Hayashi H, Ichihara M, Iwashita T, Murakami H, Shimono Y, Kawai K, Kurokawa K, Murakumo Y and Imai T. (2000). Oncogene, 14, 4469 ± 4475. Hennige AM, Lammers R, Arlt D, Hoppner W, Strack V, Niederfellner G, Seif FJ, Haring H and Kellerer M. (2000). Mol. Cell. Endocrinol., 167, 69 ± 76. Hill RJ, Zozulya S, Lu YL, Hollenbach PW, Joyce-Shaikh B, Bogenberger J and Gishizky ML. (1996). Cell Growth Di€., 7, 1125 ± 1134. Ishizaka Y, Ochiai M, Tahira T, Sugimura T and Nagao M. (1989). Oncogene, 4, 789 ± 794. Kavanaugh WM and Williams LJ. (1994). Science, 266, 1862 ± 1865. Liu X, Vega QC, Decker RA, Pandey A, Worby CA and Dixon JE. (1996). J. Biol. Chem., 271, 5309 ± 5312. Lorenzo MJ, Gish GD, Houghton C, Stonehouse TJ, Pawson T, Ponder BAJ and Smith DP. (1997). Oncogene, 14, 763 ± 771. Lowenstein EJ, Daly RJ, Batzer AG, Li W, Margolis B, Lammers R, Ullrich A, Skolnik EY, Bar-Sagi D and Schlessinger J. (1992). Cell, 70, 431 ± 442.

Migliaccio E, Giorgio M, Mele S, Pelicci G, Reboldi P, Pandol® PP, Lanfrancone L and Pelicci PG. (1999). Nature, 402, 309 ± 313. Milbrandt J, de Sauvage FJ, Fahrner TJ, Baloh RH, Leitner ML, Tansey MG, Lampe PA, Heuckeroth RO, Kotzbauer PT, Simburger KS, Golden JP, Davies JA, Vejsada R, Kato AC, Hynes M, Sherman D, Nishimura M, Wang LC, Vandlen R, Mo€at B, Klein RD, Poulsen K, Gray C, Garces A, Henderson CE, Phillips HS and Johnson Jr EM. (1998). Neuron, 20, 245 ± 253. Pandey A, Duan H, Di Fiore PP and Dixit VM. (1995). J. Biol. Chem., 270, 21461 ± 21463. Pandey A, Liu X, Dixon JE, Di Fiore PP and Dixit VM. (1996). J. Biol. Chem., 271, 10607 ± 10610. Pasini B, Borrello MG, Greco A, Bongarzone I, Luo Y, Mondellini P, Alberti L, Miranda C, Arighi E, Bocciardi R, Seri M, Barone V, Romeo G and Pierotti MA. (1995). Nature Genet., 10, 35 ± 40. Pasini B, Ceccherini I and Romeo G. (1996). Trends Genet., 12, 138 ± 144. Pelicci G, Lanfrancone L, Grignani F, McGlade J, Cavallo F, Forni G, Nicoletti I, Pawson T and Pelicci PG. (1992). Cell, 70, 93 ± 104. Pelicci G, Lanfrancone L, Salcini AE, Romano A, Mele S, Borrello MG, Segatto O, Di Fiore PP and Pelicci PG. (1995). Oncogene, 11, 899 ± 907. Pierotti MA, Bongarzone I, Borrello MG, Greco A, Pilotti S and Sozzi G. (1996). Genes Chrom. Cancer, 16, 1 ± 14. Porter AC and Vaillancourt RR. (1998). Oncogene, 17, 1343 ± 1352. Santoro M, Wong WT, Aroca P, Santos E, Matoskova B, Grieco M, Fusco A and Di Fiore PP. (1994). Mol. Cell. Biol., 14, 663 ± 675. Schlessinger J. (1993). Trends Biochem. Sci., 18, 273 ± 275. Segoun-Cariou C and Billaud M. (2000). J. Biol. Chem., 275, 356 ± 3576. Songyang Z, Carraway III KL, Eck MJ, Harrison SC, Feldman RA, Mohammadi M, Schlessinger J, Hubbard SR, Smith DP, Eng C, Lorenzo MJ, Ponder BAJ, Mayer BJ and Cantley LC. (1995). Nature, 373, 536 ± 539. Songyang Z, Shoelson SE, McGlade J, Olivier P, Pawson T, Bustelo XR, Barbacid M, Sabe H, Hanafusa H and Yi T. (1994). Mol. Cell. Biol., 14, 2777 ± 2785. Takahashi M, Buma Y, Iwamoto T, Inaguma Y, Ikeda H and Hiai H. (1988). Oncogene, 3, 571 ± 578. Takahashi M, Ritz J and Cooper GM. (1985). Cell, 42, 581 ± 588. Thomas D and Bradshaw RA. (1997). J. Biol. Chem., 272, 22293 ± 22299. Trub T, Choi WE, Wolf G, Ottinger E, Chen YJ, Weiss M and Shoelson SE. (1995). J. Biol. Chem., 270, 18205 ± 18208. Trupp M, Arenas E, Fainzilber M, Nilsson AS, Sieber BA, Grigoriou M, KilKenny C, Salazar-Grueso E, Pachnis V and Arumae U. (1996). Nature, 381, 785 ± 788. Trupp M, Scott R, Whittemore SR and Ibanez CF. (1999). J. Biol. Chem., 274, 20885 ± 20894. Valius M, Bazenet C and Kazlauskas A. (1993). Mol. Cell. Biol., 13, 133 ± 143. van Weering DHJ and Bos JL. (1997). J. Biol. Chem., 272, 249 ± 254.

3485

Oncogene