Oncogene (1997) 14, 773 ± 782  1997 Stockton Press All rights reserved 0950 ± 9232/97 $12.00

Identi®cation of Shc docking site on Ret tyrosine kinase

Elena Arighi1, Luisella Alberti1, Francesca Torriti1, Simona Ghizzoni1, Maria Grazia Rizzetti1, Giuliana Pelicci2, Barbara Pasini3, Italia Bongarzone1, Claudia Piutti4, Marco A Pierotti1 and Maria Grazia Borrello1

1Division of Experimental Oncology A, Istituto Nazionale Tumori, 20133 Milan; 2European Institute of Oncology, 20141 Milan; 3Scienti®c Direction, Istituto Nazionale Tumori, 20133 Milan; 4Oncology and Immunology Preclinical Research, Pharmacia and Upjohn Research, 20014 Nerviano, Italy

The RET proto-oncogene encodes two isoforms of a mutations of proto-RET have been identi®ed in the receptor type tyrosine kinase which plays a role in inherited cancer syndromes Multiple Endocrine neural crest and kidney development. Distinct germ-line Neoplasia (MEN) type 2A and 2B and Familial mutations of RET have been associated with the Medullary Thyroid Carcinoma (FMTC) as well as in inherited cancer syndromes MEN2A, MEN2B and a congenital malformation, the Hirschsprung disease FMTC as well as with the congenital disorder (HSCR) (reviewed in van Heyningen, 1994 and see Hirschsprung disease (HSCR), whereas somatic rear- references herein). All MEN2A mutations involve rangements (RET/PTCs) have been frequently detected cysteine residues present in the extracellular domain in the papillary thyroid carcinoma. Despite these of proto-RET, whereas a single point mutation in the ®ndings, suggesting a relevant role for RET product TK domain, changing the methionine 918 to in development and neoplastic processes, little is known threonine, has been found in almost all cases of about the signalling triggered by this receptor. In this MEN2B (reviewed in Ponder, 1995 and references study, we have demonstrated that the transducing herein). On the contrary, somatic rearrangements of adaptor molecule Shc is recruited and activated by RET, generating RET/PTC oncogenes, have been both Ret isoforms and by the rearranged cytoplasmatic detected in about 35% of surgical specimens from Ret/ptc2 oncoproteins as well as by the membrane papillary histotype of thyroid carcinomas (Pierotti et bound receptor activated by MEN2A or by MEN2B al., 1996). RET/PTCs encode fusion proteins in which associated mutations. Moreover, our analysis has the tyrosine kinase (TK) and the C-terminal domains identi®ed the Ret tyrosine residue and the Shc domains of RET are fused to di€erent donor genes (Grieco et involved in the interaction. In fact, here we show that al., 1990; Bongarzone et al., 1993, 1994). All the both the phosphotyrosine binding domains of Shc, PTB donor genes, being constitutively expressed in all and SH2, interact with Ret/ptc2 in vitro. However, tissues, are responsible for the ectopic expression of PTB domain binds 20 folds higher amount of Ret/ptc2 the oncogenes in thyroid cells and, as suggested by than SH2. The putative binding site for either SH2 and sequence analysis and coimmunoprecipitation experi- PTB domains has been identi®ed as Tyr586 of Ret/ptc2 ments, their products are endowed with a constitutive (Tyr1062 on proto-Ret). In keeping with this ®nding, by dimerization capability, which is then conferred to the using RET/PTC2-Y586F mutant, we have demonstrated chimeric proteins (Bongarzone et al., 1993; Pierotti et that this tyrosine residue, the last amino acid but one al., 1996). As a consequence, the Ret/ptc oncoproteins before the divergence of the two Ret isoforms, is the are constitutively phosphorylated on tyrosine and docking site for Shc. display a constitutive tyrosine kinase activity (Lanzi et al., 1992; Borrello et al., 1994). In addition, we Keywords: Ret; Shc; receptor tyrosine kinase; oncogene have recently demonstrated that Ret/ptc proteins activate the signal transducer PLCg (Borrello et al., 1996), as do di€erent RTKs triggered by the respective ligands. Introduction By transfection assays, RET/PTCs transform NIH3T3 cells and di€erentiate the phechromocytoma Proto-RET gene encodes two isoforms of a receptor cell line PC12 (Grieco et al., 1990; Bongarzone et al., type tyrosine kinase (RTK) (Takahashi et al., 1988; 1993, 1994; Pasini et al., 1995). Proto-RET genes, Tahira et al., 1990). Only very recently a Ret ligand activated by point mutations associated with MEN2A has been identi®ed as the glial derived neurotrophic or MEN2B, are also able to transform NIH3T3 and to factor (GDNF) (Trupp et al., 1996; Durbec et al., induce di€erentiation of PC12 cells (Borrello et al., 1996). Several studies also involving transgenic mice 1995). have demonstrated a critical function for this receptor Shc is a highly conserved gene, isolated as an in kidney and enteric nervous system development SH2-containing protein (Pelicci et al., 1992). It (Schuchardt et al., 1994; Pachnis et al., 1993). encodes three overlapping proteins of 46, 52 and Alterations of RET gene have been found associated 66 kDa, that share three distinct domains: a single with ®ve di€erent diseases. In fact, distinct germ-line SH2 domain in the C-terminal portion of the protein; a glycine- and proline-rich region containing Correspondence: MA Pierotti and MG Borrello a tyrosine phosphorylation site at position 317 Received 17 July 1996; revised 21 October 1996; accepted 22 October responsible for binding the adapter protein Grb2, 1996 and the PTB domain, a newly recognized phospho- Shc docking site on Ret EArighiet al 774 tyrosine-binding domain unrelated to SH2 domain Results (Blaikie et al., 1994; Kavanaugh and Williams, 1994). Shc proteins are ubiquitously expressed and are In NIH3T3 cell lines expressing di€erent oncogenic involved in the signalling pathways triggered by forms of Ret, Shc is tyrosine phosphorylated and binds di€erent classes of proteins. Several tyrosine kinase Ret oncoproteins receptors (Pelicci et al., 1992, 1995; Yokote et al., 1994; Segatto et al., 1993; Giorgetti et al., 1994; We have previously demonstrated that Shc is Vainikka et al., 1994; Borrello et al., 1994; Stephens constitutively activated in cells expressing the Ret et al., 1994) as well as non-receptor kinase such as oncoproteins Ret/ptc1 and Ret/ptc2 (Borrello et al., Src, Fps (McGlade et al., 1992) and Lck (Baldari et 1994; Pelicci et al., 1995). Here we show in Figure 1 al., 1995) and surface receptors without intrinsic TK that Shc proteins are tyrosine phosphorylated in activity (e.g. interleukin receptors, the B-cell antigen NIH3T3 cell lines expressing either isoforms p76 or receptor and the erythropoietin receptor) are all able p81 encoded by the chimeric oncogene RET/PTC2 to interact with Shc (Damen et al., 1993; Matsuguchi (lane PTC2iso9 and PTC2iso51 of panel a), in et al., 1994 and references herein). Shc has also been NIH3T3 cells expressing Ret-C634R protein, a demonstrated to bind to middle T antigen of MEN2A-associated Ret mutant, and in a cell line Polyoma virus and is involved in middle T antigen- expressing Ret-M918T, the MEN2B-associated Ret induced oncogenic transformation (Dilworth et al., mutant (lane Ret2A and Ret2B of panel d). In 1993). Upon phosphorylation, Shc binds to the addition, the same panels show that a signi®cant Grb2-Sos complex leading to the activation of the fraction of all the Ret oncoproteins coimmunopreci- Ras signaling pathway (Lowenstein et al., 1992). pitate with Shc. On the contrary, untransfected Recently, it has been reported that the direct NIH3T3 cells and an NIH3T3 cell line expressing interactions of Shc with a number of TK receptor the wild type unstimulated proto-Ret product show a such as IR (Gustafson et al., 1995), TRKA (Dikic et very low level of phosphorylated Shc proteins (lanes al., 1995), EGFR and ErbB-3 (Prigent et al., 1995) NIH and Retwt of panel d). All the cell lines in are primarily mediated by its PTB domain. The PTB Figure 1 are representative of a number of cell lines domain, at variance with all the SH2 domains, derived from di€erent NIH3T3 transfection experi- interacts with sequences amino-terminal to the ments (Bongarzone et al., 1993; Borrello et al., 1995) phosphotyrosine, speci®cally recognizing NXXpY and are designated according to the name of the motif, where p indicates phosphorylation (Kava- expressed Ret protein. Figure 1b,c,e and f show Shc naugh and Williams, 1994; Trub et al., 1995). and Ret expression on the analysed cell lines. All the We have previously demonstrated that the chimeric oncoproteins, but not proto-Ret product, are consti- oncogenic proteins Ret/ptc1 and Ret/ptc2 activate Shc tutively phosphorylated on tyrosine (data not shown, proteins. In fact, in cells expressing Ret/ptc oncopro- see also Borrello et al., 1995, 1996). teins, Shc is constitutively phosphorylated and We have previously shown that the SH2 domain of coimmunoprecipitates with Ret/ptc oncoproteins and Shc binds in vitro Ret/ptc oncoproteins (Borrello et with Grb2 (Borrello et al., 1994). al., 1994). To investigate whether both the phospho- A number of studies have been performed in the tyrosine-binding domains of Shc, SH2 and PTB last few years in order to identify speci®c phosphotyr- domains, could directly bind Ret oncoproteins, an osine residues in TK receptors as major or unique in vitro binding experiment was performed using the binding sites for each one of the known SH2 Gst-fused SH2 or PTB proteins and a Ret/ptc2 containing proteins (Pawson and Schlessinger, 1993). protein extract pretreated with SDS at 1008Cto In the case of Ret, the lack of ligand has so far dissociate preexisting protein-protein interactions. The hampered the analysis of its signal transduction results are shown in Figure 2a: both the Shc domains pathway that is relevant for the comprehension of are still able to bind Ret/ptc2 after the treatment, the molecular basis of the di€erent Ret-associated thus demonstrating that the binding could be direct. diseases. However, it is possible to study the signaling To control our experimental conditions, we used in of Ret using the Ret oncogenic forms, derived from the same experiment Gst-Grb2 and an extract the di€erent diseases, all displaying a constitutively containing activated PDGF receptor, known to activated tyrosine kinase. interact through Syp (Li et al., 1994) and we have In this report, we demonstrate that the transducing con®rmed that this binding is not direct (Figure 2a, molecule Shc is recruited and activated by all the last panel). activated forms of Ret: the membrane bound receptor To estimate the relative anities of the SH2 and activated by MEN2B or MEN2A associated muta- PTB Shc domains for Ret, binding experiments were tions, as well as the rearranged cytoplasmic Ret/ptc2 performed in the presence of equal amount of the oncoproteins but not the unstimulated proto-Ret two domains and increasing amounts of the Ret/ product. Moreover, by in vitro inhibition experiments ptc2-containing extract. As shown in Figure 2b, the using synthetic phosphopeptides modeled on Ret PTB domain appears to bind about 20 times more cytoplasmic amino acidic sequence and the SH2 and Ret/ptc2 protein than the SH2 domain, as estimated PTB Shc domains, we have identi®ed the putative by PhosphorImager analysis from three independent docking sites for Shc on Ret C-terminal. By site- experiments. In addition, we have estimated that directed mutagenesis, we have demonstrated that a about 15 ± 20% of the Ret oncoproteins immunopre- Ret/ptc2-Y586F (Y-to-F change at position 586) cipitated with anti-Ret antiserum binds PTB in vitro, mutant oncoprotein, while maintaining its TK activ- with no di€erence between the two Ret isoforms, ity, completely loses its ability to coimmunoprecipitate whereas 0.5 ± 1% Ret oncoprotein binds SH2 with Shc transducer protein. domain. Shc docking site on Ret EArighiet al 775

a b c

NIH PTC2 PTC2iso9 PTC2iso51 NIHPTC2 PTC2iso9 PTC2iso51 NIH PTC2 PTC2iso9 PTC2iso51 L

81 L L L 76 – 66

52 — – 46 — –

IP: αShc αRet BLOT: αPTyr αShc αRet

de f Ret2B Retwt Ret2B Ret2A Ret2A Retwt Retwt Ret2A

NIH NIH NIH Ret2B L 160 L L L 140

66 — –

52 — – 46 — –

IP: αShc αRet BLOT: αPTyr αShc αRet Figure 1 Shc is tyrosine phosphorylated and binds Ret oncoproteins in NIH3T3 cell lines expressing di€erent oncogenic forms of Ret. Western blot analysis with the indicated antisera of immunoprecipitates (IP) from NIH3T3 cells (lanes NIH) or from NIH3T3 cell lines expressing the following proteins: Ret/ptc2 (lanes PTC2), Ret/ptc2 short isoform (lanes PTC2iso9), Ret/ptc2 long isoform (lanes PTC2iso51), protoRet (lanes Retwt), Ret-M918T (lanes Ret2B), Ret-C634R (lanes Ret2A). 4 mg of protein extracts was used for immunoprecipitations shown in (a) and (d), 0.5 mg for (c) and 1 mg for (f). (b) and (e) are the same blots of (a) and (d) rehybridized. The cells were all serum starved 40 h and treated 3 h with sodium orthovanadate. Arrowheads indicate the di€erent Ret proteins while lines indicate Shc isoforms. Numbers are molecular weights in kD

the last amino acid but one before the divergence of Identi®cation of putative docking sites for Shc the two isoforms (Takahashi et al., 1988, see also To determine the potential Shc binding sites for either peptide sequence in Figure 5). The sequence of both domains on Ret protein, a series of synthetic tyrosine peptides is in agreement with the consensus sequence phosphorylated decapeptides, corresponding to all the for binding to the PTB, NXXpY (Kavanaugh and tyrosine residues in the tyrosine kinase domain and in Williams, 1994). However, since the binding sequence the C-terminal region of Ret, were used (Table 1A). of PTB domain lays in the sequence amino-terminal to When 100 mM of each of these peptides were examined pTyr, and the amino acid in 75 position is required for their ability to prevent the in vitro association for an high anity binding (Trub et al., 1995), a new between the autophosphorylated Ret/ptc2 oncopro- phosphopeptide, 76pY586+1, was synthesized, dis- teins and the PTB domain fused to Gst, only peptides playing a sequence common to both the isoforms pY586(iso9) and pY586(iso51) had a slight but (Table 1B). As shown in Figure 3a, 76pY586+1 consistent e€ect (data not shown). These peptides peptide, but not its unphosphorylated version, encompass the Tyr586 of Ret/ptc2 (Tyr1062 of proto- signi®cantly inhibits the binding between Ret/ptc2 Ret) in the two Ret isoforms; in fact this tyrosine is and the PTB domain. The 76pY586+1 phosphopep- Shc docking site on Ret EArighiet al 776

a PTC2 NIH + PDGF C + SDS C + SDS C + SDS ° ° ° 100 100 – ––100 –

180 ¨

81 L 76 L

IP/BP: αRet Gst–SH2 Gst–PTB Gst αPDGFR Gst –SH2 (Shc) (Shc) (Grb2) Blot: αRet αPDGFR

b PTC2 protein extract (mg): 0.5 123123

81 L 76 L

IP/BP: αRet Gst–SH2 Gst–PTB (Shc) (Shc) Blot: αRet Figure 2 In vitro binding of Ret/ptc2 oncoproteins to SH2 and PTB domains of Shc. Western blot analysis with anti-Ret or anti- PDGFR antisera using protein extracts from NIH3T3 cell line expressing Ret/ptc2 oncoprotein (PTC2), or from NIH3T3 treated 15' with 50 ng/ml PDGF (NIH+PDGF), in a binding experiment using 10 mg of Sepharose-bound Gst-fusion protein (BP) bearing the indicated domains (SH2 or PTB from Shc and SH2 from Grb2) or the Sepharose-bound Gst (Gst). In (a) 1 mg amount of protein extract in PLC-LB bu€er was used. The samples were left untreated (lanes7) or boiled in the presence of 1% SDS (lanes 1008C+SDS) before binding 0.3 mg of the same cell extract was immunoprecipitated (IP) with anti-Ret antiserum. Immunoreactive bands were visualized using 125I-labeled protein A. In (b) 0.5, 1, 2 or 3 mg of Ret/ptc2 expressing protein extract was used. Immunoreactive bands were visualized using peroxidase conjugate secondary antibodies and ECL detection reagents. Arrowheads indicate Ret/ptc2 protein, arrow indicates PDGF receptor

tide shows, however, a lower competing activity, In vitro phosphorylation of Ret synthetic peptides Y586 estimated about one tenth, than pY(MT) (Table 1B, by Ret/ptc2 Zhou et al., 1996), a peptide corresponding to the sequence 76pY+1 around the docking site of middle To determine whether Tyr586 of Ret/ptc2 could be T antigen of Polyoma virus for Shc PTB (Figure 3a phosphorylated by Ret kinase, the unphosphorylated last panel). decapeptides 74Y586(iso9)+5 and 74Y586(iso51)+5 The same series of phosphopeptides listed in Table were tested as substrates in an in vitro kinase reaction 1A was used to inhibit the binding between Ret/ptc2 using Ret/ptc2 (Figure 4). In previous works 74pY+5 and the SH2 domain of Shc fused to Gst. As shown in peptides were demonstrated to be good substrates in Figure 3b, two phosphopeptides, pY539 and vitro for PDGFR and Ret kinases (Turck and pY586(iso9) were able to prevent in vitro association Edenson, 1994; Borrello et al., 1996). The products of between autophosphorylated Ret/ptc2 oncoproteins the immunokinase reaction on Y586 peptides, partially and Shc SH2 domain, while pY586(iso51) showed a puri®ed by reverse-phase HPLC to eliminate ATP, minor inhibiting e€ect. All the other Ret peptides, were subjected to TLC followed by autoradiography. listed in Table 1A, have no inhibiting e€ect (data not As shown in Figure 4, spots comigrating with the shown). respective cold phosphorylated peptides, used as Shc docking site on Ret EArighiet al 777

a b ¨ Y586 pY586

pY586 ¨ pY(MT) pY586(51) pY535 pY539 – – peptides – pY586(9) L

L 81 L

L 76

Gst–PTB αRet Gst–PTB αRet ¨ IP/BP ¨ Gst–SH2 αRet αRet ¨ BLOT ¨ αRet Figure 3 Identi®cation of putative docking sites on Ret/ptc2 by phosphopeptide competition of in vitro binding. Western blot analysis with anti-Ret antiserum on ant-Ret immunoprecipitates (IP), as control, or on sepharose-bound Gst-fused proteins (BP) using 1 mg of cell extract from NIH3T3 cell line expressing Ret/ptc2, 10 mg of the indicated Gst-fusion protein and 100 mM of unphosphorylated (Y) or tyrosine-phosphorylated (pY) synthetic peptides. pY586 indicates the 76pY586+1 phosphopeptide that displays a sequence commmon to both the isoforms. pY586(9) and pY586(51) indicate the isoform-speci®c 74pY586+5

phosphopeptides (see Table 1). 7, no peptide ƒequen™es of phosphopeptides

„—˜le I a b

@eA ± Rp‰CS ‚et phosphopeptides for ƒrP dom—in

p‰EPUT q‚eqp‰„„†e†

¨ p‰EQIS

†suvp‰qegƒ ¨ p‰EQQHCQQQ vs†ip‰eup‰qƒv‚q

p‰EQSH †q€qp‰vqƒqq

p‰EQVV qwp‰veiwu

p‰ERPR ƒ‚h†p‰iihƒ‰

p‰ERPW iihƒp‰†u‚ƒ ATP ATP

p‰ERSP phrsp‰„„ƒh

p‰ERUT qqx€p‰€qs€€

p‰ESHS ƒiiwp‰‚vwv

p‰ESQW u‚‚hp‰vhvee

p‰ESSQ

hƒvsp‰hhqvƒ pY586 t0 t60 pY586 t0 t60 p‰ESVT@WA

ixuvp‰q‚sƒr (iso51) (iso9) p‰ESVT@SIA ixuvp‰qwƒh€

p‰ETIR

qp€‚p‰€xhƒ† Figure 4 In vitro phosphorylation of Y586 containing decapep-

p‰ETPH

xhƒ†p‰ex‡wv tides by Ret/ptc2 tyrosine kinase. An in vitro kinase reaction was @fA ± Tp‰CI phosphopeptides for €„f dom—in carried out with immobilized Ret/ptc2 oncoprotein to phosphor-

ylate the synthetic decapeptides Y586iso9 (a) and Y586iso51 (b). ‡sixuvp‰q

p‰ESVT The peptide containing fractions, derived from the kinase reaction vvƒx€„p‰ƒ

p‰@w„A partially puri®ed by HPLC, stopped immediately after the addition of the peptide (to) or subjected to 60' of kinase reaction (t60) were runned on TLC. The same plate was stained with ninhydrin to reveal the reference cold phosphorylated peptides references, were present when both the Y586 peptides (lanes pY586(iso9) and pY586(iso51), indicated by arrowheads); modeled on Ret/ptc2 either short or long isoform unreacted ATP is indicated sequences, were used.

Construction of RET/PTC2-Y586F mutant order to verify the synthesis of the correctly sized Because of its ability to bind in vitro both the (76 kDa) protein and to compare the biochemical phosphotyrosine binding domains of Shc, Tyr586 was characteristics of the two proteins. chosen as the target for mutagenesis to eliminate the As shown in the left panel of Figure 5b, both putative docking site for Shc in the Ret/ptc2 oncopro- constructs express the expected Ret/ptc2 protein tein. Regarding the second putative docking site, Tyr539, reacting with anti-Ret antiserum, while untransfected we had already demonstrated that a RET/PTC2-Y539F COS7 cells fail to reveal anti-Ret reactive bands. Wt and mutant is still able to activate Shc (Borrello et al., 1996). mutant Ret proteins display comparable steady state The Tyr586?Phe mutation (Y586F) was introduced tyrosine phosphorylation level (Figure 5b, right panel). by site-directed mutagenesis in the RET/PTC2 cDNA Moreover, analysed by in vitro immunokinase assay, wt encoding the short isoform (Figure 5a and see and mutant proteins showed comparable autopho- Materials and methods). A full-length RET/PTC2- sphorylation and substrate-phosphorylation activities Y586F cDNA clone was sequenced in its entirety and on MBP (Figure 5c). An additional band of slightly was demonstrated to di€er from the RET/PTC2-wt higher molecular weight was also variably present only for the introduced missense mutation (data not (Figure 5b) in anti-Ret immunoprecipitates because of shown). The RET/PTC2-Y586F was then cloned in translation starting from the ATG present in the pRC- pRC-CMV eukaryotic expression vector, carrying the CMV vector 6 bp upstream of the cloning site and in neomycin resistance gene. The pRC-CMV construct frame with the RET/PTC2 sequence. This band was and the corresponding construct containing RET/ absent when pMAMNeoblue vector was used (Pasini et PTC2-wt were transiently expressed in COS7 cells in al., 1995) and was previously shown not to a€ect on the Shc docking site on Ret EArighiet al 778

a

b c

– Y586F Y586F wt Y586F wt – wt

L Ret/ptc2 L

L Ret/ptc2 ¨ MBP

IP: αRet Immunokinase (αRet) assay BLOT: αRet αPTyr Figure 5 Schematic representation of RET/PTC2-Y586F and ®rst analysis of its product. (a): Schematic representation of Ret/ptc2 and proto-Ret showing the introduced mutation converting tyrosine 586 of Ret/ptc2 to phenylalanine. (b): Expression and tyrosine phosphorylation of Ret/ptc2-wt (lanes wt) and Ret/ptc2-Y586F (lanes Y586F) proteins, indicated by arrowheads, transiently expressed in COS7 cells. (c): In vitro immunocomplex kinase assay of Ret/ptc2-wt and Ret/ptc2-Y586F, transiently expressed in COS7 cells. Shown are autophosphorylation (Ret/ptc2 proteins indicated by arrowhead) and substrate myelin basic protein (MBP) phosphorylation (indicated by arrow)

results of the comparison between RET/PTC2-wt and whereas Ret/ptc2-Y539F, like Ret/ptc2-wt, was still other RET/PTC2 mutants (Pasini et al., 1995; Borrello et able to coimmunoprecipitate with Shc proteins, both al., 1996). Ret/ptc2-Y586F, which lacks the putative docking site for Shc, and Ret/ptc2-K282R, which is devoid of tyrosine kinase activity, were not (Figure 6a), clearly RET/PTC2-Y586F mutant is not able to interact with indicating that the presence of the phosphorylated Shc proteins Tyr586 is essential for the interaction between Ret and To analyse the ability of Ret/ptc2-Y586F mutant to Shc. Figure 6b, c and d are control Western blot bind and activate Shc proteins, COS7 cells were analyses, showing that all the constructs were similarly transfected with wt and mutant Ret/ptc2. As con- expressed (Figure 6c) and that the Y539F and Y586F, trols, COS7 cells, either untransfected or transfected but not the K282R mutants display tyrosine phosphor- with Ret/ptc2-Y539F, or with the kinase negative ylation similar to the wt (Figure 6d). The in vitro mutant Ret/ptc2-K282R were used. Ret/ptc2-K282R binding ability of wt and Y586F mutant Ret/ptc2 for was obtained by site-directed mutagenesis and cloned the two separate domains of Shc was also analysed. in the pRC-CMV vector (data not shown). According The Y586F mutation clearly abolished the ability of to studies on other TK receptors, replacement of the Ret/ptc2 to bind both the PTB and the SH2 domain of corresponding lysine residue is predicted to inactivate Shc (data not shown). the RET kinase by disrupting the ATP binding site (Chen et al., 1987). This Ret/ptc2 mutant, obtained by in vitro mutagenesis and cloned on pRC-CMV vector, Discussion was in fact demonstrated unable to autophosphorylate and to phosphorylate the exogenous substrate MBP Distinct RET gene mutations are associated with (data not shown). several forms of human diseases. Both the point Since even after 40 h serum starvation, untransfected mutated RET genes found in multiple endocrine COS7 cells showed phosphorylated Shc proteins, we neoplasias MEN2A and MEN2B, and somatically could not use this test to analyse the ability of the rearranged RET gene of papillary thyroid carcinoma various RET/PTC2 mutants to activate Shc by tyrosine (RET/PTCs), encode oncogenic forms of Ret, showing phosphorylation. However, we could demonstrate that, a constitutive activation of their tyrosine kinase activity Shc docking site on Ret EArighiet al 779

a b c d Y539F – Y586F Y586F K282R Y586F Y539F Y539F Y586F K282R – K282R – – Y539F wt wt wt K282R wt L L p76 L 66 — –

52 — – 46 — –

IP: αShc αRet

BLOT: αPTyr αShc αRet αPTyr Figure 6 Coimmunoprecipitation of di€erent Ret/ptc2 mutated proteins, transiently expressed in COS7 cells, with Shc proteins. Western blot analysis with the indicated antisera of immunoprecipitates (IP) from COS7 cells untransfected (7) or transfected with RET/PTC2-wt (lanes wt), with RET/PTC2-Y586F (lanes Y586F), RET/PTC2-Y539F (lanes Y539F), RET/PTC2-K282R (lanes K282R). 2 mg amount of cell extract was used for immunoprecipitations shown in (a) and (b) 0.5 mg for (c) and (d) and, upon transfection, the ability to transform photyrosine with high anity, the ability of Ret/ptc2 NIH3T3 cells and to di€erentiate PC12 pheochromo- to directly bind these domains was tested. We have cytoma cell line (Pasini et al., 1995; Borrello et al., shown that both PTB and SH2 Shc domains are able 1995; Pierotti et al., 1996). We have previously shown to directly bind in vitro Ret/ptc2, although PTB shows that the transducing molecule Shc is constitutively a 20 fold higher anity, suggesting that it could be activated by Ret/ptcs oncoproteins (Borrello et al., the major or the unique binding site for Ret. By in 1994; Pelicci et al., 1995). In this paper, we have vitro binding inhibition experiments, we have deter- analysed Shc activation in stable cell lines expressing mined the putative docking site for both SH2 and the Ret gene carrying a MEN2A (Ret-C634R) or the PTB domains. MEN2B mutation (Ret-M918T) and we have demon- Using a series of phosphorylated decapeptides, strated that all Ret oncoproteins are able to recruit and corresponding to all the cytoplasmatic tyrosine activate Shc. Given that Ret/ptc oncoproteins localize residues of Ret, we have shown that three peptides, in the cytoplasm, whereas Ret-C634R and Ret-M918T one encompassing Tyr539 and two Tyr586, with its maintain the transmembrane localization as proto-Ret, isoform 9 or isoform 51 environments, inhibited the these ®ndings indicate that, irrespectively of the binding between Ret/ptc2 and SH2(Shc). Whereas the subcellular localizations, the cytoplasmic domain of role of Tyr586 (Tyr1062 for proto-Ret) was not Ret is suited to recruit Shc. In addition, although the previously investigated, we had previously shown, M?T mutation associated with MEN2B, as well as the using stable cell lines expressing Ret/ptc2-Y539F, that M?T mutation at the corresponding amino acid in the presence of Tyr539, the docking site for PLCg,is EGFR, was demonstrated responsible for changing the not necessary to activate Shc (Borrello et al., 1996), substrate speci®city of the kinase (Songyang et al., thus ruling out that Tyr539 (Tyr1015 for proto-Ret) 1995; Pandit et al., 1996), Shc appears a common could have a crucial role as docking site for Shc. substrate for wild type and MEN2B Ret kinase. This The quanti®cation of the binding experiments ®nding is in agreement with the evidence that both the showed, however, that less than 1% of the Ret/ptc2 receptor-type tyrosine kinases, harboring a methionine, present in the extract was able to bind in vitro the Gst- and the Src-like tyrosine kinases, harboring a threonine SH2 protein, whereas about 20% of the protein in the corresponding position (Hanks et al., 1994), are interacts with Gst-PTB fusion protein in the same able to activate Shc (e.g. Pelicci et al., 1992; McGlade experimental conditions. The candidate docking site for et al., 1992). PTB domain was again Tyr586. In fact, the Moreover, we have analysed the ability of either Ret 74pY586+5 phosphopeptide was the only one able isoforms, separately expressed, to activate Shc. Ret to inhibit the binding between Ret/ptc2 and PTB(Shc) gene codes for two isoforms di€ering in their C- and, in addition, the amino acid sequence around this terminal (Tahira et al., 1990), the longest isoform tyrosine residue was in good agreement with the displaying two tyrosines absent in the short one. consensus for the PTB domain. The sequence NXXpY Interestingly, a common tyrosine, the last amino acid was strictly required for PTB binding (Kavanaugh and but one before the divergence of the two isoforms Williams, 1994; Van der Geer et al., 1996; Laminet et (Tyr586 on Ret/ptc2), could have di€erent SH2 al., 1996) and Tyr586 was the only one cytoplasmic Ret domain-speci®city because of the di€erence in the two tyrosine with Asn in 73 position. Although the isoforms C-terminal amino acid sequence. However, by majority of the identi®ed PTB docking sites share the using cell lines expressing Ret/ptc2-iso9 or Ret/ptc2- sequence NPXpY (Zhou et al., 1996), the presence of iso51 only, we have demonstrated a comparable Shc Pro in 72, absent in 72 position of Tyr 586, was recruitment, thus suggesting that the major docking demonstrated not essential (Dikic et al., 1995). In site for Shc lays in the aminoacidic sequence common addition, PTB domain, at variance with SH2 domains, to both isoforms. needs at least ®ve residues N-terminal to pTyr for its Since Shc proteins harbor a classical SH2 domain binding, with a strong requirement for a non-polar and a structurally unrelated domain, recently char- residue in 75 position (Van der Geer et al., 1996; acterized and designated PTB (Pelicci et al., 1992; Kavanaugh and Williams, 1994; Songyang et al., 1995; Kavanaugh and Williams, 1994), which bind phos- Dikic et al., 1995). Accordingly, Tyr586 shows an Shc docking site on Ret EArighiet al 780 isoleucine residue in 75 position. A 76pY586+1 Materials and methods phosphopeptide, corresponding to the last common Ret amino acids and thus representative of both Ret Site-directed mutagenesis and cloning of RET/PTC2 mutants isoforms, was used to block Ret/ptc2-PTB(Shc) The cloning of the RET/PTC2 cDNA has been previously binding. The peptide was indeed e€ective although reported (Bongarzone et al., 1993) (EMBL data bank about tenfold less ecient than the 76pY+1 peptide accession number L03357). The 2 kb XbaI cDNA insert corresponding to the PTB docking site of polyoma was cloned into pAlter vector (Promega). Site-directed virus middle T antigen (Laminet et al., 1996). Studies mutagenesis was performed using an in vitro oligonucleo- of Zhou et al. (1996), using NMR and surface plasmon tide mutagenesis system (Altered Sites in vitro mutagenesis resonance techniques with phosphorylated peptides, system from Promega). Oligonucleotide carrying the demonstrated that binding anity of PTB(Shc) domain Tyr586?Phe was 5'-GAAAACAAACTCTTTGGTAGA- ATTTCCC-3', oligonucleotide carrying the Lys282?Arg for its docking sites (K 0.03 ± 5.3 mM) is generally d was 5'-GGTGGCCGTGAGGATGCTGAAAGAG-3' syn- higher than that of SH2(Shc) domain for its own sites thesized with an Applied Biosystem 391 apparatus. The (Kd 50 ± 130 mM) and, in particular, the Kd for PTB second mutation was screened for the creation of a new and middle T was 0.14. These data are in keeping with FokI restriction site; since the introduced mutation did not those here reported showing that Ret/ptc2 binds PTB destroy nor create restriction sites for the ®rst mutation, domain one order of magnitude more than SH2 mutant clones were identi®ed by Allele Speci®c Oligonu- domain. Moreover, following similar experiments cleotide (ASO) technique: a 193 bp fragment was ampli®ed using Trk oncoproteins (Borrello et al., 1994), we by PCR around codon 586 and hybridized with the speci®c found that the latter bind at least 100 times more PTB wt and mutated oligonucleotides 5'-AAACTCTATGG- than SH2 domain of Shc (data not shown), according CATGTCA-3' or 5'-AAACTCTTTGGCATGTCA-3' at stringent conditions to identify mutant clones. Full-length with the results of Dikic et al. (1995) obtained with RET/PTC2 cDNA clones carrying the appropriate muta- activated proto-Trk receptor. tions were entirely sequenced. wt and mutant cDNAs were In conclusion, we have de®ned a major role for the then inserted into the XbaI site of the eukaryotic PTB domain and a minor role, if any, for the SH2 expression vectors pRC-CMV (Invitrogen) under the domain in driving Ret-Shc interaction. Since the same control of the human cytomegalovirus (CMV) promoter. tyrosine residue is the putative docking site for both PTB and SH2 domains, Shc could interact with two Cell culture and transient transfection Ret molecules by using both domains, thus playing a role in stabilizing dimers' formation, as previously COS7 and mouse NIH3T3 cells were grown in Dulbecco's hypothesized for EGFR and ERBB3 (Prigent et al., modi®ed Eagle's medium (DMEM) containing 10% fetal calf serum and 10% calf serum, respectively. Before protein 1995). extraction, NIH3T3 cells and NIH3T3 derived cell lines At variance with most of TK receptors showing were serum starved for 48 h in DMEM plus 0.1% bovine multiple docking sites for Shc (Pelicci et al., 1992; serum albumin (BSA) and 10 mM HEPES bu€er; in some Segatto et al., 1993), we have demonstrated that a cases, the cells were pretreated with 50 nM sodium

unique tyrosine residue in the C-terminal of Ret is orthovanadate (Na3VO4)for3h. responsible for binding to Shc. In fact, the Ret/ptc2- For transient transfection COS7 cells were seeded at the Y586F mutant was not only unable to bind in vitro density of 16106 cells/10 cm plate in complete medium and PTB, but also completely inecient in recruiting Shc in grown for 24 h prior to a 30' incubation in serum free living cells. medium with 500 mg/ml DEAE-destran, 1 mg/plate of speci®c These results will allow in the near future to study DNA and 20 mg/plate of carrier plasmid DNA. After 30', the medium was removed and cells were incubated 3 h in more precisely the role of Shc activation in the context complete medium plus 100 mM chloroquine. Cells were then of the di€erent biological e€ects, transformation and grown in DMEM, 10% FCS, for 72 h before protein di€erentiation, induced by the two Ret physiological extraction. isoforms and by its various pathological versions. The latter include RET/PTC oncogenes as well as the Immunoprecipitation and Western blotting MEN2-associated RET-C634R and RET-M918T mu- tations which are all able to transform NIH3T3 cells Protein samples were prepared as previously reported (Pierotti et al., 1996; Borrello et al., 1995). Since the (Borrello et al., 1996) and immunoprecipitated with the adaptor protein Shc is oncogenic by itself when speci®ed antisera: anti-Ret common antiserum (Borrello et overexpressed (Pelicci et al., 1992) and is involved in al., 1996), anti-phosphotyrosine (anti-PTyr), a commer- cially available monoclonal IgG2b antiserum (Upstate T-antigen-mediated oncogenic transformation (Dil- Biotechnology Incorporated), anti-Shc monoclonal anti- worth et al., 1993), it will be of interest to study body (from PG Pelicci already described (Borrello et al., whether the di€erent oncogenic forms of Ret carrying 1996)) and a commercially available rabbit polyclonal anti- Y?F mutation at residue 1062, and thus completely Shc (Upstate Biotechnology Incorporated). Di€erent unable to recruit Shc, are still able to transform amounts of protein extracts were used, as speci®ed in the NIH3T3 cells. The transforming ability of YF mutants ®gures: for binding experiments 1 mg, for coimmunopreci- should be analysed for either RET isoforms, since the pitation experiments about 3 ± 4 mg, for other immunopre- long one alone is able to directly bind with high anity cipitations experiments about 0.5 mg. Immunoprecipitates Grb2, whereas Grb2 binding to the short isoform could were resolved by electrophoresis on 7.5% or 8.5% SDS- be Shc-mediated (manuscript in preparation). In polyacrylamide gels (PAGE). Proteins were transferred onto nitrocellulose ®lters, blocked with 5% bovine serum addition, the present results will contribute to the albumin (BSA) or 0.5% gelatin in Tris-bu€ered saline analysis of the role of Shc in the physiological (TBS) pH 7.6 and immunoblotted with the same anti-Ret activation of proto-Ret receptor by the recently antiserum described above or with another anity puri®ed discovered GDNF ligand (Trupp et al., 1996; Durbec anti-Ret common polyclonal antiserum (Borrello et al., et al., 1996). 1996) both designated anti-Ret (3.5 mg ml71), with anti- Shc docking site on Ret EArighiet al 781 PTyr (0.2 mg ml71) or anti-Shc antisera (1 mgml71). (60 : 50 : 75 : 15), air dried and exposed to X-ray ®lm Immunoreactive bands were visualized using 125I-labeled (Hyper®lm-MP, Amersham). In addition, the same plate protein A (Amersham) followed by autoradiography or was stained with ninhydrin to reveal the reference peptides. using a horseradish peroxidase conjugate anti-rabbit or anti-mouse antiserum and ECL detection reagents (Amer- In vitro binding experiments using Gst-fused proteins and sham). When 125I-protein A was used, the ®lters were phosphopeptides competition exposed to storage phosphor screen ®lms and analysed with PhosphorImager apparatus (Molecular Dynamics) in Bacterial cultures expressing pGEX vector alone (Phar- order to quantify the c.p.m. associated to the Ret-speci®c macia) or recombinant pGEX containing the SH2 or the bands reacting with anti-Ret antisera. PTB domain of Shc or that one of Grb2 were grown in LB containing 100 mgml71 ampicillin and induced with 1 mM isopropyl b-D-thiogalactopyranoside for 3 to 6 h. The In vitro immunocomplex kinase assay induced bacteria were lysed by sonication in PBS contain- The anti-Ret immunoprecipitates, absorbed on protein A- ing 10 mgml71 of Aprotinin, Leupeptin and Pepstatin. Gst Sepharose beads, were washed twice with lysis bu€er, once or Gst-fused proteins were puri®ed using glutathione- with incubation bu€er (50 mM HEPESpH7.2,20mM Sepharose (Pharmacia) and retained in the immobilized

MnCl2,5mMPMSF) and incubated for 15 min at 48Cin form bound to glutathione-Sepharose. Cells were lysed as 20 ml of the same bu€er containing 0.5 mM DTT, 4 mCi of described (Borrello et al., 1994). The clari®ed lysates were [g32P]ATP (5000 Ci mmol71, Amersham) diluted with incubated with 5 ± 10 mg of immobilized Gst or Gst-fused unlabeled ATP to the ®nal concentration of 26 pmol proteins for 90 min at 48C. When indicated, 1% SDS was ATP per sample and 50 mM of myelin basic protein added and protein samples were treated 5' at 1008Cand (MBP). The reaction was stopped by addition of subsequently diluted to 0.1% SDS concentration, before concentrated Laemmli bu€er. Proteins were eluted and the incubation with the Gst proteins. Protein complexes subjected to 15% SDS ± PAGE. 32P-labeled bands were were resolved by SDS ± PAGE and transferred to nitro- revealed by autoradiography of the dried gel and analysed cellulose. Blots were blocked and probed as described by PhosphorImager. above. The peptide competition was performed by preincubating the Gst-fused proteins bound to the glutathione-Sepharose In vitro kinase assay using synthetic peptides with the indicated concentration of peptides and phospho- In vitro kinase assay was performed as described in the peptides dissolved in PLC-LB bu€er, before adding the cell previous paragraph with the following modi®cations: lysates. The phosphopeptides were synthesized as previously 250 mM of synthetic peptides instead of MBP were reported (Borrello et al., 1996). added, the reaction was carried out at 378Cfor1hand was stopped by adding 30% acetic acid. The peptides contained in the mixture of the immunokinase reaction were dried, resuspended in 25 ml water and separated by reverse phase HPLC on a Waters DELTA PAK 18C-300AÊ column, 15m, 7.8 mm6300 mm. Elution was performed at 1mlmin71 with linear gradient from 15% to 85% solvent B in 30 min (solvent A was 0.05% tri¯uoroacetic acid in Acknowledgements water/acetonitrile 95 : 5 and solvent B was 0.05% trifluor- We are grateful to Darrin Smith and Bruce Ponder for the oacetic acid in water/acetonitrile 40 : 60). The eluant was availability of the constructs RET-M918T and RET- monitored by a u.v. detector (Kontron Instruments) at C634R and to Pier Giuseppe Pelicci and Carlo Battistini 220 nm. The fractions were manually collected. The for the helpful discussion. We thank Maria Teresa Radice peptide-containing fractions (deduced by adding the and Mario Azzini for excellent technical assistance and synthetic cold phosphorylated reference peptide to the Anna Grassi for competent secretarial work. This study mixture, in a separate run) were dried, resuspended in 5 ml was supported by the Associazione and Fondazione 30% acetic acid and spotted in 1 ml aliquots onto a Italiana per la Ricerca sul Cancro (AIRC/FIRC), by the 10620 cm2 cellulose TLC plate (Merck). TLC plates were National Council for Research (CNR), Special Project eluted with water/pyridine/acetic acid/1-butanol ACRO, and by the Italian Ministry of Health.

References

Baldari C, Pelicci G, Di Somma MM, Milia E, Giuli S, Borrello MG, Alberti L, Arighi E, Bongarzone I, Battistini Pelicci PG and Telford JL. (1995). Oncogene, 10, 1141 ± C, Bardelli A, Pasini B, Piutti C, Rizzetti MG, Mondellini 1147. P, Radice MT and Pierotti MA. (1996). Mol. Cell. Biol., Blaikie P, Immanuel D, Wu J, Li N, Yajnik V and Margolis 16, 2151 ± 2163. B. (1994). J. Biol. Chem., 269, 32031 ± 32034. ChenWS,LazarCS,PoenieM,TsienRY,GillGNand Bongarzone I, Monzini N, Borrello MG, Carcano C, Rosenfeld MG. (1987). Nature, 328, 820 ± 823. Ferraresi G, Arighi E, Mondellini P, Della Porta G and Damen JE, Liu L, Cutler RL and Krystal G. (1993). Blood, Pierotti MA. (1993). Mol. Cell. Biol., 13, 358 ± 366. 82, 2296 ± 2303. Bongarzone I, Butti MG, Coronelli S, Borrello MG, Santoro Dikic I, Batzer AG, Blaikie P, Obermeier A, Ullrich A, M, Mondellini P, Pilotti S, Fusco A, Della Porta G and Schlessinger J and Margolis B. (1995). J. Biol. Chemistry, Pierotti MA. (1994). Cancer Res., 54, 2979 ± 2985. 270, 15125 ± 15129. Borrello MG, Pelicci G, Arighi E, De Filippis L, Greco A, Dilworth S, Brewster C, Jones M, Lanfrancone L, Pelicci G Bongarzone I, Rizzetti MG, Pelicci PG and Pierotti MA. and Pelicci PG. (1993). Nature, 367, 87 ± 89. (1994). Oncogene, 9, 1661 ± 1668. Durbec P, Marcos-Gutierrez CV, KilKenny C, Grigoriou M, Borrello MG, Smith DP, Pasini B, Bongarzone I, Greco A, Wartiowaara K, Suvanto P, Smith D, Ponder B, LorenzoMJ,ArighiE,MirandaC,EngC,AlbertiL, Costantini F, Saarma M, Sariola H and Pachnis V. Bocciardi R, Mondellini P, Scopsi L, Romeo G, Ponder (1996). Nature, 381, 789 ± 793. BAJ and Pierotti MA. (1995). Oncogene, 11, 2419 ± 2427. Shc docking site on Ret EArighiet al 782 Giorgetti S, Pelicci G, Pelicci PG and Van Obberghen E. Pierotti MA, Bongarzone I, Borrello MG, Greco A, Pilotti S (1994). Eur. J. Biochem., 223, 195 ± 202. and Sozzi G. (1996). Genes Chrom. Cancer, 16, 1 ± 14. Grieco M, Santoro M, Berlingieri MT, Melillo RM, Donghi Ponder BAJ. (1995). Cancer Surv., 25, 195 ± 205. R, Bongarzone I, Pierotti MA, Della Porta G, Fusco A Prigent SA, Pillay TS, Ravichandran KS and Gullick WJ. and Vecchio G. (1990). Cell, 60, 557 ± 563. (1995). J. Biol. Chem., 270, 22097 ± 22100. Gustafson TA, He W, Craparo A, Schaub CD and O'Neill Schuchardt A, D'Agati V, Larsson-Blomberg L, Costantini TJ. (1995). Mol. Cell. Biol., 15, 2500 ± 2508. F and Pachnis V. (1994). Nature, 367, 380 ± 383. Hanks SK, Quinn AM and Hunter T. (1994). Science, 241, Segatto O, Pelicci G, Giuli S, Digiesi G, Di Fiore PP, 42 ± 52. McGlade J, Pawson T and Pelicci PG. (1993). Oncogene, 8, Kavanaugh WM and Williams LJ. (1994). Science, 266, 2105 ± 2112. 1862 ± 1865. Songyang Z, Carraway KL III, Eck MJ, Harrison SC, Laminet AA, Apell G, Conroy L and Kavanaugh WM. Feldman RA, Mohammadi M, Schlessinger J, Hubbard (1996). J. Biol. Chem., 271, 264 ± 269. SR, Smith DP, Eng C, Lorenzo MJ, Ponder BAJ, Mayer Lanzi C, Borrello MG, Bongarzone I, Migliazza A, Fusco A, BJ and Cantley LC. (1995). Nature, 373, 536 ± 539. Grieco M, Santoro M, Gambetta RA, Zunino F, Della Stephens RM, Loeb DM, Copeland TD, Pawson T, Greene Porta G and Pierotti MA. (1992). Oncogene, 7, 2189 ± LA and Kaplan DR. (1994). Neuron, 12, 691 ± 705. 2194. Tahira T, Ishizaka Y, Itoh F, Sugimura T and Nagao M. Li W, Nishimura R, Kashishian A, Batzer AG, Kim WJH, (1990). Oncogene, 5, 97 ± 102. Cooper JA and Schlessinger J. (1994). Mol. Cell. Biol., 14, Takahashi M, Buma Y, Iwamoto T, Inaguma Y, Ikeda H 509 ± 517. and Hiai H. (1988). Oncogene, 3, 571 ± 578. Lowenstein EJ, Daly RJ, Batzer AG, Li W, Margolis B, Trub T, Choi WE, Wolf G, Ottinger E, Chen YJ, Weiss M Lammers R, Ullrich A, Skolnik EY, Bar-Sagi D and and Shoelson SE. (1995). J. Biol. Chem., 270, 18205 ± Schlessinger J. (1992). Cell, 70, 431 ± 442. 18208. Matsuguchi T, Salgia R, Hallek M, Eder M, Druker B, Ernst Trupp E, Arenas E, Fainzilber M, Nilsson AS, Sieber BA, T and Grin J. (1994). J. Biol. Chem., 269, 5016 ± 5021. Grigoriou M, KilKenny C, Salazar-Grueso E, Pachnis V, McGlade J, Cheng A, Pelicci G, Pelicci PG and Pawson T. Arumae U, Sariola H, Saarma M and Ibanez CF. (1996). (1992). Proc. Natl. Acad. Sci. USA, 89, 8869 ± 8873. Nature, 381, 785 ± 789. Pachnis V, Mankoo B and Costantini F. (1993). Develop- Turck CW and Edenson SP. (1994). Peptide Research, 7, ment, 119, 1005 ± 1017. 140 ± 145. Pandit SD, Donis-Keller H, Iwamoto T, Tomich JM and Vainikka S, Joukov V, Wenstrom S, Bergman M, Pelicci PG Pike LJ. (1996). J. Biol. Chem., 271, 5850 ± 5858. and Alitalo K. (1994). J. Biol. Chem., 269, 18320 ± 18326. Pasini B, Borrello MG, Greco A, Bongarzone I, Luo Y, Van der Geer P, Wiley S, Gish GD, Lai VKM, Stephens R, Mondellini P, Alberti L, Miranda C, Arighi E, Bocciardi White MF, Kaplan D and Pawson T. (1996). Proc. Natl. R, Seri M, Barone V, Romeo G and Pierotti MA. (1995). Acad. Sci. USA, 93, 963 ± 968. Nature Genetics, 10, 35 ± 40. van Heyningen V. (1994). Nature, 367, 319 ± 320. Pawson T and Schlessinger J. (1993). Curr. Biol., 3, 434 ± 442. Yokote K, Mori S, Hansen K, McGlade J, Pawson T, Heldin Pelicci G, Lanfrancone L, Grignani F, McGlade J, Cavallo CH and Claesson-Welsh L. (1994). J. Biol. Chem., 269, F, Forni G, Nicoletti I, Pawson T and Pelicci PG. (1992). 15337 ± 15343. Cell, 70, 93 ± 104. Zhou MM, Harlan JE, Wade WS, Crosby S, Ravichandran Pelicci G, Lanfrancone L, Salcini AE, Romano A, Mele S, KS, Burako€ SJ and Fesik SW. (1996). J. Biol. Chem., 271, Borrello MG, Segatto O, Di Fiore PP and Pelicci PG. 31119 ± 31123. (1995). Oncogene, 11, 899 ± 907.