Oncogene (2001) 20, 156 ± 166 ã 2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $15.00 www.nature.com/onc Gab1 phosphorylation: a novel mechanism for negative regulation of HGF receptor signaling

P Gual*,1, S Giordano1, S Anguissola1, PJ Parker2 and PM Comoglio1

1Institute for Cancer Research and Treatment (IRCC), University of Torino Medical School, Str. Prov. 142, Km 3.95, 10060 Candiolo, Italy; 2Protein Phosphorylation Laboratory, Imperial Cancer Research Fund, 44 Lincoln's Inn Fields, London WC2A 3PX, UK

Signal transduction by HGF receptor, the tyrosine formation and acquisition of an invasive-metastatic kinase encoded by the MET oncogene, switches on a phenotype (Giordano et al., 1993; Bellusci et al., 1994; genetic program called `invasive growth' inducing Rong et al., 1994). HGF receptor is a cell surface epithelial cell dissociation, migration, growth, and glycoprotein composed of one extracellular a-subunit ultimately leading to di€erentiation into branched and one transmembrane b-subunit. Binding of the tubular structures. Sustained tyrosine phosphorylation ligand to the receptor stimulates the tyrosine kinase of the downstream adaptor protein Gab1 is required for activity of the b-subunit that leads to receptor the HGF response. Here we show that serine/threonine autophosphorylation, membrane recruitment and tyr- phosphorylation of Gab1 provides a control mechanism osine phosphorylation of intracellular signal transdu- for negative regulation. Treatment with okadaic acid, a cers. Two-phosphorylated tyrosines (Y1349 and potent inhibitor of the serine/threonine protein phospha- Y1356) in the C-terminal tail of the HGF receptor b- tases PP1 and PP2A, was followed by activation of a subunit bind multiple SH2-containing transducers such number of serine/threonine kinases, hyper-phosphoryla- as phosphatidylinositol 3-kinase (PI 3-kinase) (Graziani tion in serine and threonine of Gab1 and severe inhibition et al., 1991), the adaptor proteins GRB2 (Ponzetto et of the HGF-induced biological responses. Under these al., 1994), SHC (Pelicci et al., 1995), the transcriptional conditions, Gab1 was found to be concomitantly hypo- factor STAT (Boccaccio et al., 1998) and the docking phosphorylated in tyrosine, and thus endowed with protein Gab1 (Weidner et al., 1996). The latter reduced ability to recruit SH2 containing signal contains several sites for the above transducers and transducers such as PI3 kinase. Among the serine- provides an essential anchorage for PLCg1 (Gual et al., threonine kinases activated by PP1 and PP2A inhibition, 2000). Phosphorylation of the two tail tyrosines of we found that PKC-a and PKC-b1 are required for HGF receptor is essential for its biological activity negative regulation of Gab1. These data provide a novel (Ponzetto et al., 1994; Bardelli et al., 1997b, 1998). negative mechanism for the HGF receptor signaling Gab1 (GRB2-associated binder 1) belongs to the pathways and highlight a potentially useful target for Insulin Receptor Substrate-1 like family of adaptor inhibitors of invasive growth. Oncogene (2001) 20, molecules (Herbst et al., 1996; Raabe et al., 1996; 156±166. Yamanashi and Baltimore 1997; Carpino et al., 1997; Gu et al., 1998; Jones and Dumont 1998; Kouhara et Keywords: HGF; Gab1; Ser/Thr phosphorylation; al., 1997; Yenush and White 1997) and it is tyrosine negative regulation; invasive growth; PKC phosphorylated in response to insulin (Holgado- Madruga et al., 1996), a number of growth factors (EGF, HGF, NGF, PDGF) (Weidner et al., 1996; Introduction Holgado-Madruga et al., 1996, 1997; Rakhit et al., 2000), cytokines (IL 3 and 6, Interferon a and b, Biological responses to Hepatocyte Growth Factor erythropoietin, thrombopoietin, granulocyte-macro- (HGF; also known as PRGF-1 or `Scatter factor') are phage colony-stimulating factor, stem cell factor) mediated by the tyrosine kinase receptor encoded by (Takahashi-Tezuka et al., 1998; Lecoq-Lafon et al., the Met oncogene. HGF receptor (Met) triggers a 1999; Nishida et al., 1999) and T and B cell antigens multi-step genetic program including cell-dissociation, (Nishida et al., 1999). In spite of this lack of speci®city, invasion of extracellular matrices and growth, which ± Gab1 plays a critical role in HGF signaling. Its in physiological conditions ± results in polarization and overexpression mimics some of the HGF responses formation of three-dimensional branched tubular (Weidner et al., 1996) while Gab1-de®cient mice structures (Montesano et al., 1991). Inappropriate display phenotypes similar to those observed in HGF activation of HGF receptor induces neoplastic trans- null mice (Itoh et al., 2000). Gab1 contains a Pleckstrin Homology (PH) domain, two carboxy-terminal GRB2 SH3 binding sites (GBS), a Met binding site (MBS), 16 potential tyrosine- *Correspondence: P Gual Received 13 September 2000; revised 20 October 2000; accepted 23 phosphorylation sites and 47 potential serine/threonine October 2000 phosphorylation sites (Holgado-Madruga et al., 1996; Negative regulation of HGF signaling P Gual et al 157 Lock et al., 2000; Schaeper et al., 2000). The PH domain plays an important role in Gab1 membrane localization and is required for HGF-mediated branch- ing tubulogenesis (Maroun et al., 1999a,b). In cells, recruitment of Gab1 to the receptor is mediated via the formation of a complex with the SH3 domains of GRB2, which in turn binds one of the two phosphotyrosines (Y1356) of the receptor tail by its SH2 domain (Bardelli et al., 1997b; Nguyen et al., 1997). However, it has also been reported that Gab1 binds directly to the HGF receptor via its MBS (Lock et al., 2000; Schaeper et al., 2000). The GRB2 binding may increase the strength of this interaction. In response to HGF, Gab1 becomes tyrosine phosphory- lated and recruits a number of SH2 containing signal transducers, behaving as an adaptor. These include, SHPTP-2, PI3 kinase, CRK-II, PLC-g1, CRK-like and SHC (Weidner et al., 1996; Bardelli et al., 1997; Garcia-Guzman et al., 1999; Gual et al., 2000; Sakkab et al., 2000; Schaeper et al., 2000). While the role of Gab1 tyrosine phosphorylation has been largely elucidated, the role of serine/threonine phosphorylation has not yet been determined. The presence of 47 potential serine/threonine phosphoryla- tion sites suggests that Gab1 may be subjected to functional regulation by phosphorylation (Holgado- Madruga et al., 1996). This study shows that this is Figure 1 E€ect of okadaic acid on Gab1 electrophoretic indeed the case. mobility. (a) Clari®ed lysates from MLP29 cells treated with either okadaic acid (decreasing doses from 120 nM to 15 nM)or okadaic acid 7, 10, 24, 28 tetraacetate (120 nM) were incubated with antibodies to Gab1 preadsorbed on protein G-Sepharose. The pellets were then washed twice and samples were analyzed by Results Western blotting with anti-Gab1 antibodies. (b) Clari®ed lysates from MLP29 cells incubated or not with okadaic acid (120 nM) Gab1 is hyper-Ser/Thr phosphorylated after were incubated with antibodies to Gab1 preadsorbed on protein- okadaic acid treatment G Sepharose. After immunopuri®cation of Gab1, alkaline phosphatase treatment was carried out as described in Materials Upon tyrosine phosphorylation by the HGF receptor, and methods. The samples were analysed by Western blotting the docking protein Gab1 recruits a number of SH2- with anti-Gab1 antibodies containing proteins, which enhance the multiple intracellular signals required for HGF-mediated re- sponses. However, Gab1 contains also 47 potential serine/threonine phosphorylation sites, whose role in the HGF-mediated response has not yet been eluci- dated. It is known that inhibition of the serine/ threonine protein phosphatases PP1 and PP2A by okadaic acid leads to activation of a number of serine/ threonine kinases, with subsequent phosphorylation of their substrates (Schonthal, 1998; Ricciarelli and Azzi, Figure 2 Gab1 is hyper-ser/thr phosphorylated in response to 1998). As shown in Figure 1a, Gab1 is one of these okadaic acid. After serum-depletion, 293 cells transfected with substrates, since its electrophoretic mobility was Myc-tagged Gab1 were incubated with (1) ethanol, (2) okadaic progressively reduced with increasing concentrations acid 7, 10, 24, 28 tetraacetate (120 nM) and (3 ± 4) okadaic acid of okadaic acid. As shown in Figure 1b, the (120 nM and 60 nM, respectively). Gab1 was then immunopreci- electrophoretic mobility of Gab1 immunopuri®ed from pitated from clari®ed lysates with anti-Tag antibodies. Samples were analysed by Western blotting with the indicated antibodies treated cells returned to that observed in Gab1 from control cells after incubation with alkaline phospha- tase. This indicates that the observed decrease of electrophoretic migration was due to hyper-phosphor- lated in response to okadaic acid. As expected, the ylation. Moreover, Gab1 immunoprecipitated from treatment of cells with the inactive form of okadaic transfected cells, treated with or without okadaic acid, acid (okadaic acid 7, 10, 24, 28 tetraacetate) had no was analysed by anti-phosphothreonine and anti- e€ect. Stripping the membrane and reprobing with phosphoserine Western blotting. As shown in Figure speci®c antibodies determined the total level of protein 2, Gab1 was hyper-serine and -threonine phosphory- analysed.

Oncogene Negative regulation of HGF signaling P Gual et al 158 Serine/threonine phosphorylation of Gab1 correlates with (Schonthal, 1998), we had to rule out that abrogation of HGF mediated responses after okadaic acid the abrogation of HGF-mediated responses treatment was not due to induction of cell apoptosis. Mouse Liver Progenitor cells (MLP29) in culture, form We observed that in presence of okadaic acid the cells tight epithelial monolayers with junctional complexes. were still alive and able to grow in an anchorage The addition of HGF induces disruption of inter- independent manner (data not shown). In agreement cellular junctions and stimulates cell motility (`scatter- with our observation, a recent study has shown that ing'). Moreover, in tridimensional collagen gels, the the treatment of fetal rat lung explant cultures with cells proliferate and migrate forming long and okadaic acid reduces branching morphogenesis without branched tubules (Medico et al., 1996). Since over- inducing apoptosis (Taylor et al., 2000). expression of Gab1 mimics the HGF-mediated cell scattering and branching tubulogenesis (Weidner et al., 1996), we investigated the consequences of the hyper- Serine/threonine phosphorylation prevents tyrosine phosphorylation of Gab1 serine/threonine phosphorylation of Gab1. As shown in Figure 3, cell treatment with okadaic acid severely Gab1 immunoprecipitated from cells treated with or impaired cell scattering (Figure 3a) and tubule without okadaic acid was analysed by anti-phospho- formation (Figure 3b) in response to HGF. Incubation tyrosine Western blotting. The total level of protein of cells with the inactive form of okadaic acid (okadaic analysed was determined by reprobing the membrane acid 7, 10, 24, 28 tetraacetate), as control, was with speci®c antibodies. As shown in Figure 4a, the ine€ective. This indicates that serine/threonine phos- tyrosine phosphorylation was severely impaired in phorylation of Gab1 negatively modulates the signals Gab1 serine-threonine phosphorylated. Accordingly, transduced by HGF. Gab1 immunopuri®ed from cells treated with HGF Since okadaic acid is able to induce apoptosis in and okadaic acid demonstrated a reduced ability to various primary as well as transformed cells coprecipitate SH2 containing transducers, such as the

Figure 3 Cell treatment with okadaic acid abrogates HGF mediated responses. (a) MLP29 cells were plated in DMEM containing 10% serum in 96-well plates. The medium was changed after 5 h and HGF (100 units/ml), okadaic acid (30 nM) and okadaic acid 7, 10, 24, 28 tetraacetate (30 nM) were added as indicated. After 12 h, the cells were ®xed with glutheraldeyde and stained with crystal violet. (b) Tubulogenesis assays with MLP29 cells were performed as described in the Materials and methods section. (a) and (b) Show a representative experiment that has been performed in duplicate three times

Oncogene Negative regulation of HGF signaling P Gual et al 159 insulin receptor in response to okadaic acid is due to the disruption of the IRS-1/insulin receptor complex by the binding of 14-3-3 to serine phosphorylated IRS-1 (Ogihara et al., 1997). First, using an antibody broadly reactive with the seven di€erent 14-3-3 isotypes, we investigated whether 14-3-3 proteins are potential partners for Gab1 in the receptor system studied. As shown in Figure 5b, 14-3-3 proteins were co-immuno- precipitated with Gab1. Furthermore, the cell-treat- ment with okadaic acid led to the association of one additional isotype of the 14-3-3 protein family, indicating that the binding between this 14-3-3 isotype and Gab1 is serine phosphorylation dependent (14-3-3 proteins bind to phosphorylated serine residues within the consensus sequence Arg-Ser-X-pSer-X-Pro (Muslin et al., 1996); mouse Gab1 has two similar amino acid sequences, Arg-Ser-Tyr-Ser266 and Arg-Ser-Ser-Ser420). However, while 14-3-3 binds to serine phosphorylated Gab1, cell treatment with okadaic acid does not prevent the recruitment of Gab1 by the HGF receptor (Figure 5c). The possibility of involvement of a tyrosine phosphatase has also been investigated. The pretreatment of 293 cells transfected with Gab1 with a tyrosine phosphatase inhibitor (sodium orthovana- date), before addition of okadaic acid, largely pre- Figure 4 Ser/thr phosphorylation of Gab1 prevents its tyrosine phosphorylation following HGF treatment. (a) After serum- vented the decrease of Gab1 electrophoretic migration. depletion, 293 cells transfected with Myc-tagged Gab1 were (Figure 5d). This correlated with the increase of Gab1 incubated with (1) ethanol, (2) okadaic acid 7, 10, 24, 28 tyrosine phosphorylation. Since the cell treatment with tetraacetate (120 nM) and (3 ± 4) okadaic acid (120 nM and 60 nM, okadaic acid and sodium orthovanadate led to a respectively) before addition of HGF (400 units/ml). From the decrease of serine/threonine phosphorylation of Gab1, clari®ed lysates, Gab1 was immunoprecipitated with anti-Tag antibodies. The pellets were washed twice and samples were we could not conclude that the e€ect of okadaic acid is analysed by Western blotting with indicated antibodies. (b) due to activation of a tyrosine phosphatase. Clari®ed lysates from MLP29 cells untreated or treated with The Gab1 PH domain plays an important role in okadaic acid (120 nM) for 15 h before addition of HGF (as Gab1 membrane localization and is required for HGF- indicated) were incubated with antibodies to Gab1. The pellets were washed twice and samples were analysed by Western mediated branching tubulogenesis (Weidner et al., blotting with indicated antibodies. (a) and (b) Show a 1996; Maroun et al., 1999a,b). Another possible representative experiment mechanism could be that okadaic acid treatment impairs the localization of Gab1 to the cell membrane. However, this does not seem to be the case since the regulatory subunit of the PI3 kinase (Figure 4b). This immuno¯uorescence microscopy assay with a speci®c indicates that hyper-serine/threonine phosphorylation anti-Gab1 antibody revealed that the majority of Gab1 of Gab1 impairs its tyrosine phosphorylation by the was localized at the cell periphery in the absence or the HGF receptor and reduces its ability to recruit SH2 presence of okadaic acid (Figure 6). Gab1 was containing signal transducers. localized in the cytoplasm only when cells were starved for 48 h. Cell treatment with okadaic acid does not prevent the activation of the HGF receptor or its ability to recruit Activation of PKCs leads to Gab1 phosphorylation Gab1 Since inhibition of the serine/threonine protein phos- A hypothetical mechanism to explain the decrease of phatases PP1 and PP2A by okadaic acid leads to tyrosine phosphorylation of Gab1 in cells treated with activation of MAP Kinases (ERKs) and PKCs okadaic acid could be that the drug prevents activation (Schonthal, 1998; Ricciarelli and Azzi, 1998), we tested of the HGF receptor tyrosine kinase in response to the whether these serine/threonine kinases are responsible ligand. However, this is not the case, since tyrosine for Gab1 phosphorylation. Inhibition of ERKs activity phosphorylation of the HGF receptor was not altered (by MEK1 inhibitor PD-98059) did not prevent Gab1 in the presence of okadaic acid (Figure 5a). serine/threonine phosphorylation mediated by okadaic Another explanation could be that the HGF receptor acid (data not shown). On the contrary, pretreatment loses its ability to recruit Gab1 after okadaic acid of cells with a PKC inhibitor (staurosporine), before treatment. Indeed, in the insulin signaling system, it addition of okadaic acid, slightly prevented the has been shown that the inhibition of tyrosine decrease of Gab1 electrophoretic migration (Figure phosphorylation of the docking protein IRS1 by 7a) indicating that PKCs are among the di€erent

Oncogene Negative regulation of HGF signaling P Gual et al 160

Figure 5 Cell treatment with okadaic acid does not prevent HGF receptor activation (a) or its ability to recruit Gab1 (c). Clari®ed lysates from either (a) MLP29 cells or (c) 293 cells transfected with Myc-tagged Gab1, untreated or treated with okadaic acid (120 nM) before addition of HGF (400 units/ml), as indicated, were incubated with antibodies to HGF receptor (HGF-R). Samples were analysed by Western blotting with the appropriate antibodies (as indicated). (b) Cell treatment with okadaic acid increases the 14.3.3 recruitment with Gab1. Clari®ed lysates from 293 cells transfected with Myc-tagged Gab1, untreated or treated with okadaic acid (120 nM), were incubated with anti-Myc antibodies. Samples were analysed by Western blotting with the anti-Gab1 antibodies and anti-14.3.3 antibodies. (d) The increase of tyrosine phosphorylation of Gab1 correlates with the decrease of its serine/threonine phosphorylation. Clari®ed lysates from 293 cells transfected with Myc-tagged Gab1, untreated or treated with okadaic acid (120 nM) and sodium orthovanadate (500 mM), were incubated with anti-Myc antibodies. Samples were analysed by Western blotting with anti-Gab1 antibodies and then with anti-phosphotyrosine antibodies

serine/threonine kinases activated by okadaic acid and are involved in Gab1 phosphorylation. In agreement with this observation, activation of PKCs by 12-o- tetradecanoyl-13-acetate (TPA) decreased the gel elec- trophoretic mobility of Gab1 (Figure 7a) and its tyrosine phosphorylation (Figure 7b). Next, to identify the PKCs involved, total lysates of cells transfected with Gab1 and the constitutively active forms of PKC a or PKC b1 were analysed by anti-protein Western blotting. As shown in Figure 7c the presence of active Figure 6 Cell treatment with okadaic acid does not prevent the PKC a or PKC b1 decreased the electrophoretic localization of Gab1 to sites of cell-cell contact. MLP 29 cells were cultured in DMEM supplemented with 5% FCS for 72 h. mobility of Gab1, indicating their involvement in For serum starvation experiments, cells were grown for 48 h in serine-threonine phosphorylation. To determine 5% FCS and then transferred for 24 h to medium lacking FCS. whether Gab1 can be a direct substrate of PKC a or Cells were ®xed in 4% paraformaldehyde, and the localization of PKC b1, we performed kinase assays in vitro, using Gab1 was determined by indirect immuno¯uorescence labeling with anti-Gab1 antibodies, followed by rhodamine-tagged immunopuri®ed enzymes and GST-Gab1 (Figure 8). secondary antibody. Photographs were taken at a magni®cation These assays were performed in the presence of of6630 inhibitors of protein kinases A and Calmodulin-

Oncogene Negative regulation of HGF signaling P Gual et al 161

Figure 7 PKCa and PKC b1 phosphorylate in vivo Gab1. (a) Clari®ed lysates from MLP29 cells treated with TPA or okadaic acid or staurosporine (performed as described in the Materials and methods section) were incubated with antibodies to Gab1 preadsorbed on protein G-Sepharose. The pellets were then washed twice and samples were analysed by Western blotting with the anti-Gab1 antibodies. (b) Clari®ed lysates from COS cells expressing Myc-tagged Gab-1, treated or not with TPA, were incubated with antibodies to Gab1 preadsorbed on protein G-Sepharose. The pellets were then washed twice and samples were analysed by Western blotting with anti-Gab1 antibodies and then with anti-phosphotyrosine antibodies. (c) COS-7 cells transfected with Myc- tagged Gab-1 and a constitutively active form of PKC b1 or PKCa were starved for 15 h. An aliquot (100 mg of total protein) of each total cell lysate was resuspended in Laemmli sample bu€er and samples were analysed by Western blotting with anti-Gab1 and anti-PKCs antibodies dependent kinases. Both PKC a and PKC b1 scored Discussion positive. In epithelial cells, activation of the HGF receptor induces cell scattering, migration, proliferation, protec- Inactivation of PKC prevents the inhibition of tion from apoptosis and branching tubulogenesis HGF-mediated cell scattering by okadaic acid (Stella and Comoglio 1999; Bardelli et al., 1997a). Cell Finally, to demonstrate that the chain of events scattering is mostly dependent on PI3 kinase and Rac triggered by okadaic acid and the inhibition of the activation (Graziani et al., 1991; Ridley et al., 1995; HGF responses are mediated by PKCs, we measured Royal and Park 1995; Khwaja et al., 1998). Cell cell scattering in cells co-incubated with the PKCs growth requires stimulation of the ras-ERK cascade inhibitor (staurosporine) and okadaic acid. As shown (Ponzetto et al., 1996). Cell protection from apoptosis in Figure 9, the inhibitor rescued the scattering. As involves the PI3 kinase signaling pathway (Franke et expected, cell treatment with either okadaic acid or al., 1997; Kau€mann-Zeh et al., 1997; Dudek et al., staurosporine alone does not mediate cell-cell dissocia- 1997; Holgado-Madruga et al., 1997) and the anti- tion (data not shown). This result shows that PKCs apoptotic protein BAG1 (Bardelli et al., 1996). The full play a major role in serine phosphorylation of Gab1 program, leading to the acquisition of cell polarity and and, thus, in the negative regulation of the HGF ending with the di€erentiation of sprouting tubular response. structures, is dependent on PLCg1 and ERK signaling

Oncogene Negative regulation of HGF signaling P Gual et al 162

Figure 8 GST-Gab1 is a direct substrate for PKCa and PKCb1. Clari®ed lysates from COS-7 cells transfected with either Myc- tagged PKCa or PKCb1 treated with TPA (160 nM) were incubated with antibodies to Tag preadsorbed on protein G-Sepharose. Pellets were washed twice and GST-Gab1 or GST alone and the phosphorylation bu€er were added. The phosphorylation reactions were stopped after 45 min at 48C by addition of Laemmli sample bu€er. The samples were analysed by SDS ± PAGE. The gel was dried and subjected to autoradiography

Figure 9 Inactivation of PKC prevents the inhibition of HGF-mediated cell scattering by okadaic acid. MLP29 cells were plated in DMEM containing 10% serum in 96-well plates. The medium was changed after 5 h and HGF (100 units/ml), okadaic acid (30 nM) and staurosporine (17 nM) were added as indicated. After 12 h, the cells were ®xed with glutheraldeyde and stained with crystal violet. We show a representative experiment that had been performed in triplicate three times

pathways (Gual et al., 2000; Schaeper et al., 2000) and (Weidner et al., 1996). Furthermore, disruption of the requires activation of the transcriptional factor STAT- association between the receptor and Gab1 impairs 3 (Boccaccio et al., 1998). HGF-mediated branching tubulogenesis (Weidner et In response to HGF, the signal ampli®er Gab1 al., 1996; Maroun et al., 1999a). Finally, it has recently becomes tyrosine phosphorylated and plays an im- been shown that Gab1 de®cient mice display pheno- portant role in the activation of PLCg1 (Gual et al., types similar to those observed in mice lacking HGF 2000) and in potentiation of the above pathways signaling (Itoh et al., 2000). (Weidner et al., 1996; Bardelli et al., 1997b; Maroun While the role of Gab1 tyrosine phosphorylation is et al., 1999a; Garcia-Guzman et al., 1999; Sakkab et well studied, the possible serine/threonine phosphoryla- al., 2000). In some experimental conditions, over- tion has not yet been investigated in spite of the expression of Gab1 in epithelial cells can induce presence of 47 potential phosphate acceptor sites constitutive cell scattering and branching tubulogenesis (Holgado-Madruga et al., 1996). Here, we show that

Oncogene Negative regulation of HGF signaling P Gual et al 163 activation of serine/threonine kinases by okadaic acid, IRS1 after PKCs activation prevents insulin-mediated a speci®c inhibitor of the serine/threonine protein responses (Chin et al., 1993; De Fea and Roth 1997). phosphatases PP1 and PP2A, is followed by the Since PKCs are activated in response to HGF hyper-serine/threonine phosphorylation of Gab1. Inter- (Okano et al., 1993; Machide et al., 1998), the serine/ estingly, serine/threonine phosphorylation concomi- threonine phosphorylation of Gab1 by these kinases tantly inhibits HGF receptor-mediated Gab1 tyrosine suggests the existence of a negative feedback loop for phosphorylation and, thus, impairs Gab1 ability to HGF-signaling. HGF receptor-triggered invasive recruit SH2 containing transducers. The ultimate growth is required for embryonic development and its consequence is the inhibition of the HGF-mediated inappropriate activation confers invasive and meta- invasive growth. static properties to cancer cells (Giordano et al., 1993; In okadaic acid-treated cells, inhibition of Gab1 Bellusci et al., 1994; Rong et al., 1994). Selective tyrosine phosphorylation is not due to the inhibition of inhibitors of this process could thus be useful both in the receptor kinase nor to its ability to recruit Gab1 understanding the HGF biology and in targetting the itself. A possible explanation is that the conformation invasive metastatic potential of cells overexpressing the of Gab1 is modi®ed as a consequence of its serine/ oncogenic receptor (HGF receptor). threonine phosphorylation, which hides tyrosine phos- phorylation sites. Accordingly, it is known that this post-translational modi®cation can indeed in¯uence the tridimensional structure of proteins (Baron et al., 1992; Materials and methods Gual et al., 1995; Bies et al., 2000). For example, the docking protein IRS1 immunopuri®ed from okadaic Materials and antibodies acid-treated cells is a poorer substrate for both the All reagents used were from Fluka (FlukaChemie, AG). insulin receptor and the cytoplasmic kinase Jak1 than Okadaic acid, okadaic acid 7,10,24,28 tetraacetate, stauros- IRS1 (Tanti et al., 1994; Cengel and Freund 1999). porine and 12-o-tetradecanoyl-13-acetate (TPA) were from Here, we show that PKC isoforms (PKCa and PKC Sigma (Sigma Chemicals Co.). Reagents for SDS ± PAGE 32 b1, in particular) phosphorylate Gab1 in vitro and in were from Bio-Rad (Biorad Laboratories) [g- P]ATP was purchased from Armersham Italia. Recombinant HGF was vivo. Moreover, in epithelial cells, the inhibition of the obtained from baculovirus-infected SF-9 cells. The mono- PKCs prevents the inhibition of cell scattering induced clonal anti-human HGF receptor antibodies DQ 13 were by okadaic acid. However, it is known that PKC puri®ed from ascite ¯uids of hybridoma from mice, and isoforms are involved in the regulation of gene kindly provided by M Prat and R Albano. Anti-phospho- expression, cell protection from apoptosis, cell growth threonine and anti-phosphoserine antibodies were purchased and di€erentiation (Nishizuka, 1992; Whelan and from Zymed (Zymed laboratories Inc, CA, USA). Anti c-myc Parker 1998; Parekh et al., 2000) and also in the antibody was purchased from Oncogene (Oncogene Research negative regulation of tyrosine kinase receptors Products, UK). Anti-PKC, anti-mouse-HGF receptor and (Takayama et al., 1984, 1988; Gandino et al., 1990). anti-14-3-3 antibodies were purchased from Santa Cruz We have previously shown that the phosphorylation of (Santa Cruz Biotechnology). Anti-phosphotyrosine, anti-Pl3 kinase p85, and anti-Gab1 antibodies were purchased from serine 985 by the TPA-activated PKC negatively UBI (Lake Placid, NY, USA). regulates kinase activity of HGF receptor (Gandino et al., 1994). On the other hand, we now observe that okadaic acid fails to trigger the PKC-induced inhibi- Plasmid construction tion of the receptor tyrosine phosphorylation. This The mouse Gab1 cDNA subcloned into pBluescript-SK was di€erence can be explained by the presence of a gift from Dr W Birchmeier. The Gab1 Not-1-fragment was 10 potential PKC phosphorylation sites in Gab1 inserted in-frame into pGEX-4T1 vector (Pharmacia Biotech (Holgado-Madruga et al., 1996) versus only one in Inc). To construct the expression vector, a Myc epitope the HGF receptor. Another explanation could be that sequence was fused to the 5' end of Gab1 cDNA in Gab1, but not the HGF receptor, is a substrate for pBluescript KS, and the Not-1 fragment containing epitope- tagged Gab1 was subcloned into the pcDNA3 expression PKC-dependent serine kinases. This hypothesis is vector (Invitrogen, CA, USA). The Myc tagged PKCa and strengthened by the ®nding that PKCs activates the PKC b1 cDNA were subcloned into the pEFLINK ERKs (Blumer and Johnson 1994; Seger and Krebs expression vector. Constitutively activate forms of PKCa 1995; Robinson and Cobb 1997; Schonwasser et al., and PKC b1 were subcloned into pCO2 expression vector 1998) and Gab1 is a good substrate in vitro and in vivo (Pears et al., 1990). for these kinases (Roshan et al., 1999). In conclusion, PKCs negatively regulate HGF signaling by inhibiting Cell culture both the receptor kinase activity and the Gab1- dependent ampli®cation machinery. The serine phos- 293 cells and COS-7 cells were purchased from ATCC phorylation of Gab1 severely inhibits its ability to (American Type Culture Collection). MLP29 is an epithelial cell line derived from mouse liver oval cells (Medico et al., recruit SH2-containing transducers, which is directly 1996). Cells were cultured in Dulbecco's modi®ed Eagle's correlated to the abrogation of HGF mediated medium supplemented with 5% fetal calf serum (Sigma responses. The importance of PKCs in the negative Chemical Co., St. Louis, USA) and maintained at 378Cina regulation of docking proteins seems to be a general 5% CO2-humidi®ed atmosphere. COS-7 or 293 cells were mechanism. For example, serine phosphorylation of transfected using the calcium phosphate precipitation techni-

Oncogene Negative regulation of HGF signaling P Gual et al 164 que (CellPhect, Pharmacia, Uppsala, Sweden). After 2 days, the cells were ®xed with glutheraldeyde and stained with the expression of the respective recombinant proteins was crystal violet. The morphogenic activity of the MLP29 tested by Western blot analysis. cultures was analysed as described previously (Medico et al., 1996). Brie¯y, the cells were harvested from cultures using trypsin-EDTA and suspended at a concentration of 105 cells Phosphorylation of Gab1 and HGF receptor in intact cells per ml in collagen solution containing type I collagen gel Con¯uent MLP29 cells growing in 145 mm culture dishes (3 mg/ml) (Collaborative Biomedical Products; Becton Dick- were starved in DMEM supplemented with 10 mM glutamine. inson Labware), 106DMEM (Gibco ± BRL), and 0.5 M Okadaic acid (120 nM) and okadaic acid 7, 10, 24, 28 HEPES (pH 7.4). Aliquots (100 ml) of the cell suspension tetraacetate (120 nM) were added as indicated for 15 h before were dispensed in 96-well microlitter plates and allowed to gel addition of HGF (400 units/ml) for 10 min. Cells were for 15 min at 378C before adding 200 ml of DMEM with subsequently washed with ice-cold Stop bu€er (20 mM TRIS, 20% FCS. After 12 h, HGF (100 units/ml), okadaic acid 7, 150 mM NaCl, 5 mM EDTA, 150 mM NaF and 2 mM sodium 10, 24, 28 tetraacetate (30 nM) and okadaic acid (30 nM) were orthovanadate, pH 7.4) before solubilization for 30 min at added as indicated. The medium was changed every 2 days. 48C in EB bu€er (20 mM TRIS, 150 mM NaCl, 5 mM EDTA, After 5 ± 7 days the branching tubulogenesis responses were 150 mM NaF, 2 mM sodium orthovanadate, 0.5 mM phenyl- evaluated. methylsulfonyl ¯uoride, 100 U/ml aprotinin, 20 mM leupeptin, 10% glycerol and 1% (vol/vol) Triton X-100, pH 7.4) for Co-immunoprecipitation assays Gab1 or in RIPA bu€er (EB bu€er with sodium dodecyl sulfate (SDS) 0.1% and deoxycolate 1%, pH 7.4) for HGF After serum-depletion, con¯uent MLP29 cells or 293 cells receptor. Clari®ed lysates obtained after centrifugation transfected with Myc-tagged Gab1 cDNA were incubated in (15 min at 15 000 g at 48C) were incubated for 3 h at 48C the presence or absence, as indicated, of okadaic acid with appropriate antibodies preadsorbed on protein-G (120 nM) (15 and 4 h, respectively) before addition of HGF Sepharose (4 mg of antibodies/sample). Pellets washed with (400 units/ml) for 10 min. Clari®ed lysates were incubated for lysis bu€er were resuspended in Laemmli sample bu€er and 3 h at 48C with the appropriate antibodies, preadsorbed on separated by SDS ± PAGE using an 8% resolving gel, before protein-G Sepharose. Pellets washed with lysis bu€er were transfer to a nitrocellulose membrane. resuspended in Laemmli sample bu€er and separated by After serum-depletion, 293 cells transfected with Myc- SDS ± PAGE using an 8 or 12% resolving gel, before transfer tagged Gab1 cDNA were incubated with okadaic acid to nitrocellulose ®lters for probing with appropriate anti- (120 or 60 nM) for 4 h before addition of HGF (400 units/ bodies (rabbit anti-hHGF receptor antibodies and mouse ml) for 10 min. Gab1 was immunoprecipitated from clari®ed anti-14-3-3 antibodies at 0.4 mg/ml, rabbit anti-P13 kinase lysates with anti-Tag antibodies and samples were separated p85 antibodies at a 1 : 2000 dilution). by SDS ± PAGE using an 8 or 12% resolving gel, before transfer to a nitrocellulose membrane. In all cases, the Indirect immunofluorescence microscopy membrane was blocked with saline bu€er (10 mM Tris, 140 mM NaCl, pH 7.4) containing 5% (wt/vol) BSA for 2 h MLP29 cells were plated onto 24-well plates (Costar) at 228C and probed with the appropriate antibodies (Mouse containing 1.4 cm2 glass coverslips. Cells were cultured in anti-phosphotyrosine antibodies and rabbit anti-Gab1 anti- DMEM supplemented with 5% of FCS for 72 h or for 48 h bodies at 1 mg/ml, mouse anti-phosphothreonine antibodies in 5% of FCS and then transferred for 24 h to medium and rabbit antiphosphoserine antibodies at 5 mg/ml). Speci®c lacking FCS. Cells were ®xed in a freshly prepared solution binding was detected by an Enhanced Chemiluminescence containing 4% paraformaldehyde in PBS, pH 7.6, for 20 min System (ECL, Amersham). In some cases, the membrane was at 48C. Cells were permeabilized by soaking coverslips in PBS stripped for 30 min at 508Cin62mM Tris, 100 mM 2- containing 0.2% Triton X-100 for 4 min at 48C. After mercaptoethanol and 2% (wt/vol) SDS, and reprobed with saturation with PBS-8% BSA for 1 h at room temperature, the indicated antibodies. cells were washed twice with PBS. The primary antibodies (anti-Gab1 antibodies, 10 mg/ml in PBS-1 %BSA) were layered onto cells and incubated in a moist chamber for Alkaline phosphatase treatment 1 h. After rinsing in PBS, coverslips were incubated with Clari®ed lysates from MLP29 cells treated or not with appropriate rhodamine-tagged secondary antibodies (1 : 50 okadaic acid (120 nM for 15 h) were incubated for 3 h at 48C dilution) in a moist chamber for 1 h. Specimens were with antibodies to Gab1 preadsorbed on protein-G Sephar- observed with inverted photomicroscope (model DM IRB ose. Pellets were washed with phosphatase bu€er (100 mM HC; Leica Microsystems, Wetzlar, Germany) equipped with TRIS, 100 mM NaCl, 10 mM MgCl2,1mM ZnCl2, pH 8) and mercury short arc epi¯uorescence lamp. The appropriate incubated with alkaline phosphatase (alkaline phosphatase combination of ®lters was used to separate excitation and calf intestinal 5 m/sample). After 60 min at 378C, the reaction emission spectra. The objective used was an immersion oil PL was stopped by addition of Laemmli sample bu€er. The APO 636/1.4. Fluorescence images of 102461024 pixels samples were analysed by SDS ± PAGE using an 8% were captured using a cooled digital CCD Hamamatsu resolving gel and transferred to nitrocellulose ®lters for ORCA camera (Hamamatsu Photonics Italia, Arese, Italy), probing with anti-Gab1 antibodies. digitally recorded with ImageProPlus 4.0 imaging software (Media Cybernetics, Silver Spring, MD, USA), and then printed with a Tetronix printer. Scattering and branching tubulogenesis assays MLP29 cells (26103 cells per well) were plated in DMEM Phosphorylation of Gab1 by PKCs in intact cells containing 10% serum in 96-well plates. The medium was changed after 5 h and HGF (100 units/ml), okadaic acid Con¯uent MLP29 cells growing in 145 mm culture dishes (30 nM), okadaic acid 7, 10, 24, 28 tetraacetate (30 nM) and were treated with staurosporine (107 nM) for 2 h, or with staurosporine (17 nM) were added as indicated. After 12 h, TPA (160 nM) for 30 min, or with okadaic acid (2 mM) for

Oncogene Negative regulation of HGF signaling P Gual et al 165 1 h or pre-incubated with staurosporine (107 nM) for 1 h twice with 20 mM HEPES, 100 mM NaCl containing 1% before addition of okadaic acid (2 mM) for 1 h. Clari®ed (vol/vol) of Triton X-100. GST-Gab1, assay dilution bu€er, lysates were incubated for 3 h at 48C with antibodies to Gab1 inhibitor cocktail (330 nM protein kinase A inhibitor peptide, preadsorbed on protein-G Sepharose. Pellets washed with 3.33 mM compound R24571), protein kinase C liquid lysis bu€er were resuspended in Laemmli sample bu€er and activator were mixed with pellets containing the immunopur- separated by SDS ± PAGE using an 8% resolving gel, before i®ed PKCs. The phosphorylation reaction was initiated by transfer to nitrocellulose ®lters. addition of 20 mM MOPS, 25 mM b-glycerol phosphate, COS-7 cells transfected with Myc-tagged Gab-1 and a 1mM Sodium orthovanadate, 1 mM dithiothreitol, 1 mM 32 constitutively active form of PKC a or PKC b1, were starved CaCl2, 12.5 mM MgCl2,83mM [g- P] (2.5 Ci/mmol) ATP. in DMEM supplemented with 10 mM glutamine for 15 h. An The phosphorylation reaction was stopped after 45 min at aliquot of each clari®ed lysate (100 mg of total protein) was 48C by addition of Laemmli sample bu€er. The samples were resuspended in Laemmli sample bu€er and separated by analysed by SDS ± PAGE using a 12% resolving gel. The gel SDS-PAGE using an 8% resolving gel, before transfer to was dried and subjected to autoradiography. We show a nitrocellulose ®lters. In all cases, the membranes were probed representative experiment that has been performed three with the appropriate antibodies (mouse PKC antibodies at times (Figure 8). 0.4 mg/ml).

Production of fusion proteins The pGEX-Gab1 was expressed in Escherichia coli. Expres- Abbreviations sion of recombinant proteins was induced with 0.1 mM The abbreviations used are: PKC, Protein Kinase C; HGF, isopropyl-b-D-thiogalactopyranoside for 3 h. Bacteria were hepatocyte growth factor; GRB-2, growth factor receptor- lysed with 20 mM Tris, 1 M NaCl, 0.2 mM EDTA, 0.2 mM bound protein 2; Gab1, GRB2 associated binder-1; PI 3 EGTA, 1 mg/ml lysozyme, 0.5 mM phenyl-methylsulfonyl kinase, phosphoinositide 3 kinase; STAT, signal transducer ¯uoride, 100 U/ml aprotinin, 20 mM leupeptin (pH 7.4) for and activator of the transcription; SHPTP, SH2 containing 30 min in ice, sonicated three times, and frozen in liquid Phosphotyrosine phosphatase; SHC, Src homology/col- nitrogen. Lysates were centrifuged (30 000 g for 30 min at lagen; SH2 Src homology 2; p85, the 85 kDa regulatory 48C), and the supernatants were incubated with glutathione- subunit of PI 3 kinase; pr, precursor; GST, glutathione S- Sepharose (Pharmacia) for 1 h at 48C. The fusion proteins transferase. were eluted with 50 mM glutathione and 100 mM HEPES (pH 8). Acknowledgments We are grateful to Dr W Birchmeier for mouse Gab1 PKC assay cDNA. The excellent technical assistance of L Palmas, R All PKC assays were performed using the Protein Kinase C Albano and G Petruccelli is gratefully acknowledged. We assay Kit (UBI, Lake Placid, NY, USA). Brie¯y, COS-7 cells thank Antonella Cignetto for secretarial help and Elaine transfected with Myc-tagged PKCa or PKC b1 were Wright for editing the manuscript. This work was incubated with TPA (160 nM) for 30 min at 378C. Clari®ed supported by BIOMED EC grant no. BMH4-CT983852 lysates were incubated with antibodies to Myc preadsorbed to PM Comoglio. P Gual is supported by Marie Curie on protein-G Sepharose for 3 h. The pellets were washed TMR fellowships.

References

Bardelli A and Comoglio PM. (1997a). Ciba Found. Symp., Carpino N, Wisniewski D, Strife A, Marshak D, Kobayashi 212, 133 ± 144; discussion 144 ± 147. R, Stillman B and Clarkson B. (1997). Cell, 88, 197 ± 204. Bardelli A, Longati P, Albero D, Goruppi G, Schneider C, Cengel KA and Freund GG. (1999). J. Biol. Chem., 274, Ponzetto C and Comoglio PM. (1996). EMBO J., 15, 27969 ± 27974. 6205 ± 6212. Chin J, Dickens M, Tavare JM and Roth RA. (1993). J. Biol. Bardelli A, Longati P, Gramaglia D, Basilico C, Tamagnone Chem., 268, 6338 ± 6347. L, Giordano S, Ballinari D, Michieli P and Comoglio PM. De Fea K and Roth RA. (1997). Biochemistry, 36, 12939 ± (1998). Proc. Natl. Acad. Sci. USA, 95, 14379 ± 14383. 12947. Bardelli A, Longati P, Gramaglia D, Stella MC and Dudek H, Datta SR, Franke TF, Birnbaum MJ, Yao R, Comoglio PM. (1997b). Oncogene, 18, 3103 ± 3111. Cooper GM, Segal RA, Kaplan DR and Greenberg ME. Baron V, Kaliman P, Gautier N and Van Obberghen E. (1997). Science, 275, 661 ± 665. (1992). J. Biol. Chem., 267, 23290 ± 23294. Franke TF, Kaplan DR and Cantley LC. (1997). Cell, 88, Bellusci S, Moens G, Gaudino G, Comoglio P, Nakamura T, 435 ± 437. Thiery JP and Jouanneau J. (1994). Oncogene, 9, 1091 ± Gandino L, Di Renzo MF, Giordano S, Bussolino F and 1099. Comoglio PM. (1990). Oncogene, 5, 721 ± 725. Bies J, Feikova S, Bottaro DP and Wol€ L. (2000). Gandino L, Longati P, Medico E, Prat M and Comoglio PM. Oncogene, 19, 2846 ± 2854. (1994). J. Biol. Chem., 269, 1815 ± 1820. Blumer KJ and Johnson GL. (1994). Trends Biochem. Sci., Garcia-GuzmanM,Dol®F,ZehKandVuoriK.(1999). 19, 236 ± 240. Oncogene, 18, 7775 ± 7786. Boccaccio C, Ando M, Tamagnone L, Bardelli A, Michieli P, Giordano S, Zhen Z, Medico E, Gaudino G, Galimi F and Battistini C and Comoglio PM. (1998). Nature, 391, 285 ± Comoglio PM. (1993). Proc. Natl. Acad. Sci. USA, 90, 288. 649 ± 653.

Oncogene Negative regulation of HGF signaling P Gual et al 166 Graziani A, Gramaglia D, Cantley LC and Comoglio PM. Parekh DB, Ziegler W and Parker PJ. (2000). EMBO J., 19, (1991). J. Biol. Chem., 266, 22087 ± 22090. 496 ± 503. Gu H, Pratt JC, Burako€ SJ and Neel BG. (1998). Mol. Cell., Pears CJ, Kour G, House C, Kemp BE and Parker PJ. 2, 729 ± 740. (1990). Eur. J. Biochem., 194, 89 ± 94. Gual P, Baron V, Alengrin F and Van Obberghen E. (1995). Pelicci G, Giordano S, Zhen Z, Salcini AE, Lanfrancone L, Endocrinology, 136, 5298 ± 5304. Bardelli A, Panayotou G, Water®eld MD, Ponzetto C, Gual P, Giordano S, Williams TA, Rocchi S, Van Obberghen Pelicci PG and Comoglio PM. (1995). Oncogene, 10, E and Comoglio PM. (2000). Oncogene, 19, 1509 ± 1518. 1631 ± 1638. Herbst R, Carroll PM, Allard JD, Schilling J, Raabe T and Ponzetto C, Bardelli A, Zhen Z, Maina F, dalla Zonca P, Simon MA. (1996). Cell, 85, 899 ± 909. Giordano S, Graziani A, Panayotou G and Comoglio PM. Holgado-Madruga M, Emlet DR, Moscatello DK, Godwin (1994). Cell, 77, 261 ± 271. AK and Wong AJ. (1996). Nature, 379, 560 ± 564. Ponzetto C, Zhen Z, Audero E, Maina F, Bardelli A, Basile Holgado-Madruga M, Moscatello DK, Emlet DR, Dieterich ML, Giordano S, Narsimhan R and Comoglio PM. R and Wong AJ. (1997). Proc. Natl. Acad. Sci. USA, 94, (1996). J. Biol. Chem., 271, 14119 ± 14123. 12419 ± 12424. Raabe T, Riesgo-Escovar J, Liu X, Bausenwein BS, Deak P, Itoh M, Yoshida Y, Nishida K, Narimatsu M, Hibi M and Maroy P and Hafen E. (1996). Cell, 85, 911 ± 920. Hirano T. (2000). Mol. Cell. Biol., 20, 3695 ± 3704. Rakhit S, Pyne S and Pyne NJ. (2000). Mol. Pharmacol., 58, Jones N and Dumont DJ. (1998). Oncogene, 17, 1097 ± 1108. 413 ± 420. Kau€mann-Zeh A, Rodriguez-Viciana P, Ulrich E, Gilbert Ricciarelli R and Azzi A. (1998). Arch. Biochem. Biophys., C, Co€er P, Downward J and Evan G. (1997). Nature, 385, 355, 197 ± 200. 544 ± 548. Ridley AJ, Comoglio PM and Hall A. (1995). Mol. Cell. Khwaja A, Lehmann K, Marte BM and Downward J. (1998). Biol., 15, 1110 ± 1122. J. Biol. Chem., 273, 18793 ± 18801. Robinson MJ and Cobb MH. (1997). Curr. Opin. Cell. Biol., Kouhara H, Hadari YR, Spivak-Kroizman T, Schilling J, 9, 180 ± 186. Bar-Sagi D, Lax I and Schlessinger J. (1997). Cell, 89, Rong S, Segal S, Anver M, Resau JH and Vande Woude GF. 693 ± 702. (1994). Proc. Natl. Acad. Sci. USA., 91, 4731 ± 4735. Lecoq-Lafon C, Verdier F, Fichelson S, Chretien S, Roshan B, Kjelsberg C, Spokes K, Eldred A, Crovello CS Gisselbrecht S, Lacombe C and Mayeux P. (1999). Blood, and Cantley LG. (1999). J. Biol. Chem., 274, 36362 ± 93, 2578 ± 2585. 36368. Lock LS, Royal I, Naujokas MA and Park M. (2000). J. Biol. Royal I and Park M. (1995). J. Biol. Chem., 270, 27780 ± Chem., 275, 31536 ± 31545. 27787. Machide M, Kamitori K, Nakamura Y and Kohsaka S. Sakkab D, Lewitzky M, Posern G, Schaeper U, Sachs M, (1998). J. Neurochem., 71, 592 ± 602. Birchmeier W and Feller SM. (2000). J. Biol. Chem., 275, Maroun CR, Holgado-Madruga M, Royal I, Naujokas MA, 10772 ± 10778. Fournier TM, Wong AJ and Park M. (1999a). Mol. Cell. Schaeper U, Gehring NH, Fuchs KP, Sachs M, Kempkes B Biol., 19, 1784 ± 1799. and Birchmeier W. (2000). J. Cell. Biol., 149, 1419 ± 1432. Maroun CR, Moscatello DK, Naujokas MA, Holgado- Schonthal AH. (1998). Front Biosci., 3, D1262 ± D1273. Madruga M, Wong AJ and Park M. (1999b). J. Biol. Schonwasser DC, Marais RM, Marshall CJ and Parker PJ. Chem., 274, 31719 ± 31726. (1998). Mol. Cell. Biol., 18, 790 ± 798. Medico E, Mongiovi AM, Hu€ J, Jelinek MA, Follenzi A, Seger R and Krebs EG. (1995). FASEB J., 9, 726 ± 735. Gaudino G, Parsons JT and Comoglio PM. (1996). Mol. Stella MC and Comoglio PM. (1999). Int. J. Biochem. Cell. Biol. Cell., 7, 495 ± 504. Biol., 31, 1357 ± 1362. Montesano R, Matsumoto K, Nakamura T and Orci L. Takahashi-Tezuka M, Yoshida Y, Fukada T, Ohtani T, (1991). Cell, 67, 901 ± 908. Yamanaka Y, Nishida K, Nakajima K, Hibi M and Muslin AJ, Tanner JW, Allen PM and Shaw AS. (1996). Cell, Hirano T. (1998). Mol. Cell. Biol., 18, 4109 ± 4117. 84, 889 ± 897. Takayama S, White MF and Kahn CR. (1988). J. Biol. Nguyen L, Holgado-Madruga M, Maroun C, Fixman ED, Chem., 263, 3440 ± 3447. Kamikura D, Fournier T, Charest A, Tremblay ML, Takayama S, White MF, Lauris V and Kahn CR. (1984). Wong AJ and Park M. (1997). J. Biol. Chem., 272, 20811 ± Proc. Natl. Acad. Sci. USA, 81, 7797 ± 7801. 20819. Tanti JF, Gremeaux T, van Obberghen E and Le Marchand- Nishida K, Yoshida Y, Itoh M, Fukada T, Ohtani T, Brustel Y. (1994). J. Biol. Chem., 269, 6051 ± 6057. Shirogane T, Atsumi T, Takahashi-Tezuka M, Ishihara K, Taylor BK, Stoops TD and Everett AD. (2000). Am. J. Hibi M and Hirano T. (1999). Blood, 93, 1809 ± 1816. Physiol. Lung Cell. Mol. Physiol., 278, L1062 ± L1070. Nishizuka Y. (1992). Science, 258, 607 ± 614. Weidner KM, DiCesare S, Sachs M, Brinkmann V, Behrens J Ogihara T, Isobe T, Ichimura T, Taoka M, Funaki M, and Birchmeier W. (1996). Nature., 384, 173 ± 176. Sakoda H, Onishi Y, Inukai K, Anai M, Fukushima Y, Whelan RD and Parker PJ. (1998). Oncogene., 16, 1939 ± Kikuchi M, Yazaki Y, Oka Y and Asano T. (1997). J. Biol. 1944. Chem., 272, 25267 ± 25274. Yamanashi Y and Baltimore D. (1997). Cell., 88, 205 ± 211. Okano Y, Mizuno K, Osada S, Nakamura T and Nozawa Y. Yenush L and White MF. (1997). Bioessays., 19, 491 ± 500. (1993). Biochem. Biophys. Res. Commun., 190, 842 ± 848.

Oncogene