Oncogene (1999) 18, 1975 ± 1982 ã 1999 Stockton Press All rights reserved 0950 ± 9232/99 $12.00 http://www.stockton-press.co.uk/onc Rho-dependent and -independent of kinase, paxillin and p130Cas mediated by Ret kinase

Hideki Murakami1, Toshihide Iwashita1, Naoya Asai1, Yosuke Iwata1, Shuh Narumiya2 and Masahide Takahashi*,1

1Department of Pathology, Nagoya University School of Medicine, 65 Tsurumai-cho, Showa-ku, Nagoya 466, Japan and 2Department of Pharmacology, Kyoto University Faculty of Medicine, Yoshida, Sakyo-ku, Kyoto 606, Japan

Glial cell line-derived neurotrophic factor (GDNF) ability to support the survival of superior cervical signals through a unique receptor system that includes ganglion neurons (Kotzbauer et al., 1996). GDNF and Ret receptor and a glycosyl-phosphati- NTN show approximately 20% identity to dylinositol-linked cell surface . In the present members of the transforming (TGF)-b study, we have identi®ed several in neuroblas- superfamily and contain seven conserved toma cells that are phosphorylated on tyrosine in residues, six of which are involved in the formation response to GDNF. The phosphorylated proteins include of three intramolecular disul®de bonds referred to as a focal adhesion kinase (FAK), paxillin and Crk-associated cysteine knot motif (McDonald and Hendrickson, substrate, p130Cas, all of which are known to be 1993). associated with focal adhesions. Of these, paxillin and Several reports have revealed that GDNF and NTN p130Cas interacted with Crk proteins in GDNF-treated represent the ligands for Ret receptor tyrosine kinase. neuroblastoma cells. GDNF also induced reorganization However, unlike other receptor tyrosine kinases, Ret of the actin cytoskelton. Tyrosine phosphorylation of activation was not induced via direct binding of FAK, paxillin and p130Cas was inhibited by cytochala- GDNF or NTN but glycosyl-phosphatidylinositol sin D or two speci®c inhibitors of phosphatidylinositol-3' (GPI)-linked cell surface proteins (called GFRa-1 and kinase (PI-3' kinase), wortmannin and LY294002, GFRa-2) were required as ligand-binding components indicating that their tyrosine phosphorylation depends (Jing et al., 1996; Treanor et al., 1996; Klein et al., on the formation of actin stress ®ber and activation of 1997; Buj-Bello et al., 1997; Baloh et al., 1997; Sanicola PI-3' kinase. In addition, phosphorylation of FAK but et al., 1997). Moreover, we found that Ca2+ ions are not of paxillin and p130Cas was markedly impaired by the necessary for the complex formation of Ret and these Clostridium botulinum C3 exoenzyme that speci®cally neurotrophic factors, suggesting the importance of the ADP-ribosylates and inactivates Rho. These results cadherin-like domain present in the Ret extracellular suggested the presence of Rho-dependent and -indepen- domain (Nozaki et al., 1998). Thus, it appears that dent signaling pathways downstream of PI-3' kinase that GDNF and NTN use a unique receptor system mediate tyrosine phosphorylation of FAK, paxillin and including several components. p130Cas through Ret kinase. We have recently reported that GDNF is able to induce Shc binding to Ret, resulting in Shc tyrosine Keywords: Ret; GDNF; focal adhesion kinase; paxillin; phosphorylation (Ohiwa et al., 1997). Shc phosphor- p130Cas ylation leads to the formation of signal transducing complex consisting of Ret, Shc and Grb2 adaptor proteins that might be responsible for activation of the Ras-mitogen activated protein kinase (MAPK) signal- Introduction ing pathway (Ohiwa et al., 1997). Activation of this pathway by GDNF and NTN may be critical for the Glial cell line-derived neurotrophic factor (GDNF) was growth or di€erentiation of neurons as observed for ®rst identi®ed as a potent neurotrophic factor to nerve growth factor (Yao and Cooper, 1995). Shc promote the survival of midbrain dopaminergic binding was also associated with the transforming neurons in culture (Lin et al., 1993). GDNF is now activity of Ret with a multiple endocrine neoplasia known to promote the survival of various central and (MEN) 2A or 2B mutation since a mutation at peripheral neurons including autonomic and sensory tyrosine 1062 in Ret that represents a Shc binding neurons and motoneurons (Beck et al., 1995; site severely impaired its transforming activity (Asai et Oppenheim et al., 1995; Tomac et al., 1995; Yan et al., 1996). al., 1995). In addition, a GDNF-related neurotrophic In the present study, we further investigated the factor, neurturin (NTN) was puri®ed from conditioned intracellular signaling pathway through Ret activated media of Chinese hamster ovary cells based on its by GDNF, using human neuroblastoma cells. Tyrosine phosphorylation of focal adhesion kinase (FAK), paxillin and p130Cas that are associated with focal adhesions (Schaller and Parsons, 1994) was induced by GDNF, accompanying the formation of actin stress ®ber. Our data also suggested the presence of divergent *Correspondence: M Takahashi Received 7 July 1998; revised 23 October 1998; accepted 23 October signaling pathways responsible for tyrosine phosphor- 1998 ylation of FAK, paxillin and p130Cas. Phosphorylation of focal adhesion proteins by Ret activation H Murakami et al 1976 with activated Ret proteins in vivo. Cell lysates from Results TGW-I-nu cells untreated or treated with GDNF (150 ng/ml) were immunoprecipitated with anti-Crk, Binding ability of intracellular signaling molecules with anti-p85 subunit of PI-3' kinase or anti-Nck antibody, Ret activated by GDNF followed by immunoblotting with anti-Ret antibody. In order to characterize the intracellular signaling As a result, autophosphorylated Ret was detected in pathway mediated by Ret tyrosine kinase, we ®rst Crk immunoprecipitates (Figure 2a and b) but not in investigated whether Src homology 2 (SH2) domains of PI-3' kinase or Nck immunoprecipitates (data not several signaling molecules including Shc, Grb2, Grb7, shown), suggesting that the binding anity of PI-3' Grb10, p85 subunit of PI-3' kinase, Crk and Nck bind kinase and Nck to Ret might be weak in vivo. Despite to Ret activated by GDNF in vitro. Human TGW-I-nu this fact, however, the activity of PI-3' kinase increased neuroblastoma cells expressing Ret were untreated or approximately fourfold in GDNF-treated TGW-I-nu treated with GDNF (Figure 1a) and their cell lysates cells as compared with that in untreated cells (Figure were incubated with the fusion proteins composed of 3). glutathione S transferase (GST) and an SH2 domain of each signaling molecule. As shown in Figure 1b, the Association of focal adhesion kinase, paxillin and p130Cas SH2 domains of Grb7, Shc, p85 subunit of PI-3' kinase with Crk and their tyrosine phosphorylation in and Crk bound to 170 kDa autophosphorylated Ret GDNF-treated cells whereas association of the SH2 domains of Grb2, Grb10 and Nck with Ret was weak or undetectable. When the Crk immunoprecipitates from GDNF- As we have already reported, Shc adaptor proteins treated neuroblastoma cells were immunoblotted with bind to tyrosine 1062 in Ret in vivo, leading to the anti-phosphotyrosine antibody, four tyrosine phos- activation of the Ras-MAPK pathway (Asai et al., phorylated proteins of 170, 130, 70 and 42 kDa were 1996; Ohiwa et al., 1997). On the other hand, the Grb2 detected (Figure 2a). The 170 and 42 kDa bands binding to Ret appeared to be very weak in intact cells, represented phosphorylated Ret and Crk proteins, although the carboxy-terminal tail of Ret contains the respectively (Figure 2b and data not shown). Since it consensus sequence for the binding of the Grb2 SH2 is known that Crk binds to paxillin (65 ± 70 kDa) and domain (Ohiwa et al., 1997; Borrello et al., 1993). In Crk-associated substrate of 130 kDa (p130Cas), the Crk the present study, we further investigated whether Crk, immunoprecipitates were immunoblotted with anti- Nck and the p85 subunit of PI-3' kinase are associated paxillin or anti-p130Cas antibody. As shown in Figure

Figure 1 In vitro binding ability of SH2 domains of signaling molecules with Ret. (a) Tyrosine phosphorylation of Ret by GDNF. TGW-I-nu human neuroblastoma cells were untreated or treated with GDNF (150 ng/ml) and their cell lysates were immunoprecipitated with anti-Ret antibody, followed by immunoblotting with anti-Ret or anti-phosphotyrosine antibody. The 170 kDa cell surface form of the Ret protein was phosphorylated. (b) In vitro binding of SH2 domains of signaling molecules to Ret. The fusion proteins composed of GST and SH2 domains of the designated signaling molecules were puri®ed and immobilized as described previously (Ohiwa et al., 1997). The immobilized GST-fusion proteins were incubated with the lysates from GDNF- untreated or treated TGW-I-nu cells and bound proteins were analysed by Western blotting with anti-Ret antibody Phosphorylation of focal adhesion proteins by Ret activation HMurakamiet al 1977

Figure 2 Detection of Crk-associated proteins in GDNF-treated neuroblastoma cells. (a) Tyrosine-phosphorylated proteins associated with Crk proteins. TGW-I-nu human neuroblastoma cells were treated with GDNF (150 ng/ml) for the indicated time and their lysate was immunoprecipitated with anti-Crk antibody, followed by anti-phosphotyrosine antibody. Exposures were 10 min for 170 and 42 kDa bands and 30 s for 130 and 70 kDa bands. (b) Detection of Crk-associated proteins. The lysate from GDNF-treated TGW-I-nu cells was immunoprecipitated with anti-Crk antibody, followed by immunoblotting with anti-Ret, anti- p130Cas, anti-paxillin or anti-Crk antibody

2b, p130Cas and paxillin were detected in Crk Seu€erlein and Rozengurt, 1994; Sasaki et al., 1996; immunoprecipitates, suggesting that phosphorylated Rankin et al., 1996). First, to investigate whether bands of 130 and 70 kDa represent p130Cas and disruption of actin stress ®ber interferes with the paxillin, respectively. To con®rm this, we further increase in GDNF or LPA-induced tyrosine phosphor- investigated whether paxillin and p130Cas were phos- ylation of focal adhesion proteins, the quiescent TGW- phorylated on tyrosine in TGW-I-nu cells stimulated I-nu cells were treated with 2 mM cytochalasin D for with GDNF. Tyrosine phosphoryation of paxillin was 15 min before addition of GDNF (150 ng/ml) or LPA detected at 5 min after treatment with GDNF and (1 mM). This treatment almost completely inhibited continued at least for 60 min (Figure 4). On the other tyrosine phosphorylation of FAK, paxillin and p130Cas hand, tyrosine phosphorylation of p130Cas was found at by GDNF or LPA (Figure 6a). maximal levels at 5 min after treatment and decreased We next assessed a role of PI-3' kinase in afterward. phosphorylation of these molecules. The TGW-I-nu Paxillin and p130Cas are present in focal adhesions cells were preincubated for 15 min with 30 nM and are believed to play a role in -mediated wortmannin or 15 mM LY294002 both of which are intracellular signaling (Burridge et al., 1992; Schaller known as speci®c inhibitors of PI-3' kinase and then and Parson, 1994). Another important signaling stimulated with GDNF for 5 min. As shown in Figure molecule in focal adhesions is focal adhesion kinase 6b, the increase of tyrosine phosphorylation of FAK, (FAK). FAK was also phosphorylated on tyrosine paxillin and p130Cas was markedly impaired by after GDNF treatment and the time course of its pretreatment of wortmannin or LY294002. These phosphorylation was similar to that of paxillin (Figure results thus suggested that tyrosine phosphorylation 4). As expected, treatment of TGW-I-nu cells with of FAK, paxillin and p130Cas depends on the formation GDNF for 15 min caused an increase in anti-paxillin of actin stress ®ber and PI-3' kinase activity. (Figure 5d) and anti-FAK (data not shown) immunos- taining of focal adhesions. Moreover, when GDNF- Di€erential e€ects of botulinum C3 exoenzyme on treated TGW-I-nu cells were stained with rhodamine- tyrosine phosphorylation of FAK, paxillin and p130Cas conjugated phalloidin, the formation of actin stress ®ber was observed (Figure 5d). The small GTP-binding protein, Rho, also regulates the assembly of focal adhesion proteins and the formation of actin stress ®bers in response to growth Tyrosine phosphorylation of FAK, paxillin and p130Cas factors or LPA (Kumagi et al., 1993; Jalink et al., requires integrity of actin cytoskelton and activation of 1993; Rankin et al., 1994). The Clostridium botulinum PI-3' kinase C3 exoenzyme speci®cally ADP-ribosylates and inacti- Integrity of the actin cytoskelton and PI-3' kinase vates Rho. To see the role of Rho in tyrosine activity was shown to be important for tyrosine phosphorylation of focal adhesion proteins, we phosphorylation of focal adhesion proteins induced examined the level of GDNF-induced tyrosine by growth factors or lysophosphatidic acid (LPA) phosphorylation of FAK, paxillin and p130Cas in the (Huang et al., 1993; Sinnett-Smith et al., 1993; C3 exoenzyme-pretreated neuroblastoma cells. Phosphorylation of focal adhesion proteins by Ret activation H Murakami et al 1978 Although tyrosine phosphorylation of FAK was observed (Figure 6c). Similarly, tyrosine phosphoryla- prevented by the C3 exoenzyme, the impairment of tion of FAK induced by LPA but not of paxillin was phosphorylation of paxillin and p130Cas was not impaired by the C3 exoenzyme.

Discussion

In the present study, we ®rst investigated signaling molecules bound to Ret activated by GDNF. We previously reported that Shc adaptor proteins were associated with Ret in vivo, leading to the activation of the Ras-MAPK pathway (Asai et al., 1996; Ohiwa et al., 1997). In vitro binding assay using the fusion proteins composed of GST and SH2 domains of several signaling molecules revealed that, in addition to the SH2 domains of Shc, the SH2 domains of Crk, p85 subunit of PI-3' kinase and Grb7 bound strongly to autophosphorylated Ret. On the other hand, the binding anity of the SH2 domains of Grb2, Grb10 and Nck to Ret appeared to be very weak. Of these molecules, we found that Crk was associated with Ret in vivo. Although the activation of PI-3' kinase was detected in GDNF-treated neuroblastoma cells, its binding to Ret was unclear in intact cells. As observed for c-Met-, EGF- and insulin receptor, multisite docking proteins such as insulin receptor substrate-1 (IRS-1) (Jian et al., 1991) and Gab1 (Holgado- Madruga et al., 1996; Weidner et al., 1996) may be necessary for activation of PI 3'-kinase through Ret activated by GDNF. Paxillin and p130Cas that localize to focal adhesions were found to be associated with Crk proteins and phosphorylated on tyrosine in GDNF-treated neuro- blastoma cells. Paxillin and p130Cas contain several pYXXP motifs that represent the binding site for the Crk SH2 domain (Schaller and Parsons, 1995). Since these two molecules were not detected in the Ret immunoprecipitates (data not shown), it seems likely that paxillin and p130Cas were associated with Crk Figure 3 Activation of PI-3' kinase by GDNF. TGW-I-nu cells proteins in focal adhesion rather than with Crk bound were treated with GDNF (150 ng/ml) for the indicated time and to Ret. On the other hand, another focal adhesion their lysate was immunoprecipitated with anti-p85 subunit of PI- protein, FAK, was not coprecipitated with Crk, 3' kinase monoclonal antibody. Activity of PI-3' kinase in the immunoprecipitate was analysed by thin-layer chromatography although it was also highly phosphorylated on (bottom panel). Relative activity of PI-3' kinase was shown in the tyrosine. Subcellular localization of FAK, paxillin upper panel. PIP, phosphatidylinositol phosphate and p130Cas suggests that they might participate in

Figure 4 Tyrosine phosphorylation of paxillin, p130Cas and FAK by GDNF. TGW-I-nu cells were treated with GDNF (150 ng/ml) for the indicated time and their lysate was immunoprecipitated with anti-paxillin, anti-p130Cas or anti-FAK antibody, followed by immunoblotting with the same antibody or anti-phosphotyrosine antibody Phosphorylation of focal adhesion proteins by Ret activation HMurakamiet al 1979

Figure 5 Anti-paxillin and tetramethylrhodamine-isothiocyanate (TRITC)-phalloidin staining of GDNF-treated neuroblastoma cells. (a, b) Phase-contrast micrographs of TGW-I-nu cells untreated (a) or treated (b) with GDNF (150 ng/ml) for 15 min. (c, d) Anti-paxillin (yellow or green) and TRITC-phalloidin (red) staining of GDNF-untreated (c) and -treated (d) cells. The formation of focal ashesions and actin stress ®ber was observed in GDNF-treated cells. The scale bars represent 50 mm

the regulation of cell adhesion and motility including induced by GDNF is Rho-dependent. Interestingly, the reorganization of the actin cytoskelton (Schaller pretreatment with the C. botulinum C3 exoenzyme that and Parsons, 1994; Guan, 1997). Consistent with this speci®cally ADP ribosylates and inactivates Rho notion, GDNF was able to induce the formation of markedly impaired tyrosine phosphorylation of FAK actin stress ®ber. Pretreatment of neuroblastoma cells but not of paxillin and p130Cas in GDNF-treated with cytochalasin D that disrupts the actin cytoskelton neuroblastoma cells, indicating that Rho might be abolished phosphorylation of FAK, paxillin and upstream of FAK tyrosine phosphorylation and p130Cas, indicating that the integrity of the cytoskelton downstream of PI-3' kinase. On the other hand, is essential for their phosphorylation. phosphorylation of paxillin and p130Cas appeared to It was previously reported that PI-3' kinase activity be regulated in a Rho-independent manner. These was required for tyrosine phosphorylation of FAK and results thus suggested the presence of divergent paxillin in PDGF-stimulated cells (Rankin et al., 1996). signaling pathways downstream of PI-3' kinase that Our present results also showed that tyrosine mediate tyrosine phosphorylation of FAK, paxillin and phosphorylation of FAK, paxillin and p130Cas induced p130Cas by GDNF (Figure 7). by GDNF was PI-3' kinase-dependent, because In several systems, it was shown that paxillin and treatment with wortmannin or LY294002, two p130Cas might be putative substrates for FAK (Schaller structurally unrelated inhibitors of PI-3' kinase, and Parsons, 1994, 1995; Guan, 1997; Harte et al., abolished their phosphorylation in GDNF-treated 1996). However, the present study revealed that even neuroblastoma cells. when FAK phosphorylation was inhibited by the C3 Despite this PI-3' kinase-dependency, the fact that exoenzyme, paxillin and p130Cas were still highly paxillin and p130Cas but not FAK were coprecipitated phosphorylated on tyrosine. In addition, the time with Crk suggests the possibility that tyrosine courses of phosphorylation were di€erent between phosphorylation of the former two proteins and FAK FAK and p130Cas. Taken together with these results, may be di€erentially regulated. Since it is known that it seems likely that phosphorylation of paxillin and Rho, a member of small GTP-binding proteins, is an p130Cas is regulated by tyrosine kinases other than upstream regulator of tyrosine phosphorylation of FAK such as Src or Pyk2 (Schaller and Parsons, 1994; FAK and paxillin induced by LPA, bombesin and Guan, 1997; Li and Earp, 1997; Schaller and Sasaki, endothelin (Kumagai et al., 1993; Rankin et al., 1994), 1997; Zheng et al., 1998; Brinson et al., 1998) a FAK- we further investigated whether their phosphorylation related protein tyrosine kinase, although it is still Phosphorylation of focal adhesion proteins by Ret activation H Murakami et al 1980

Figure 6 E€ects of cytochalasin D, wortmannin, LY294002 or C3 exoenzyme on tyrosine phosphorylation of FAK, paxillin and Cas p130 induced by GDNF. (a) Quiescent TGW-I-nu cells untreated or treated with 2 mM cytochalasin D for 15 min were incubated Cas with GDNF (150 ng/ml) or LPA (1 mM) for 5 min. Tyrosine phosphorylation of FAK, paxillin or p130 was analysed as described in Figure 4. (b) Quiescent TGW-I-nu cells untreated or treated with 30 nM wortmannin or 15 mM LY294002 for 15 min were incubated with GDNF and tyrosine phosphorylation of FAK, paxillin or p130Cas was analysed. (c) TGW-I-nu cells were cultured with or without 3 mg/ml C3 exoenzyme, ®rst for 16 h in the complete medium and then for 12 h in serum-free medium. The resulting cells were incubated with GDNF or LPA and tyrosine phosphorylation of FAK, paxillin or p130Cas was analysed

possible that FAK is responsible for basal levels of and p130Cas in these cells, this ®nding suggests that high their phosphorylation. levels of tyrosine phosphorylation of focal adhesion Germline mutations of the RET proto-oncogene are proteins may be responsible for biological properties of associated with the development of multiple endocrine transformed cells expressing the MEN 2B mutant neoplasia (MEN) types 2A and 2B. MEN 2A and proteins. MEN 2B are autosomal dominant cancer syndromes characterized by the development of medullary thyroid Materials and methods carcinoma and pheochromocytoma. MEN 2B is distinguished from MEN 2A by a more complex Cell line phenotype including mucosal neuroma, hyperganglio- nosis of the gastrointestinal tract and marfanoid TGW-I-nu cells were grown in RPMI medium supplemen- habitus. Recently, we and others demonstrated that ted with 10% fetal calf serum. MEN 2A and MEN 2B develop by di€erent molecular mechanisms based on their mutations (Asai et al., Antibodies 1995; Santoro et al., 1995; Borrello et al., 1995; Anti-Crk, anti-FAK, anti-paxillin and anti-phosphotyro- Iwashita et al., 1996; Rossel et al., 1997; Ito et al., sine monoclonal antibodies were purchased from Trans- 1997). In addition, Bocciardi et al. (1997) reported that duction Laboratories (Lexington, KY, USA). Anti-p130Cas paxillin was dramatically phosphorylated on tyrosine in polyclonal antibody was purchased from Upstate Biotech- cells expressing the MEN 2B mutant proteins. nology Inc. (Lake Placid, NY, USA). Anti-Ret antibody Although they did not show phosphorylation of FAK was produced as described previously (Asai et al., 1996). Phosphorylation of focal adhesion proteins by Ret activation HMurakamiet al 1981 viously (Ohiwa et al., 1997). For in vitro binding assay, TGW-I-nu cells stimulated or unstimulated with GDNF (150 ng/ml) were lysed in RIPA bu€er and clari®ed by centrifugation as described above. The supernatant was incubated with 5 mg of immobilized GST or GST-fusion proteins for 3 h at 48C and washed with RIPA bu€er four times. The proteins bound to GST-fusion proteins were eluted by boiling in SDS-sample bu€er, resolved by SDS ± polyacrylamide gel electrophoresis and immunoblotted with anti-Ret antibody.

Assay of PI-3' kinase TGW-I-nu cells were treated with GDNF (150 ng/ml) for 5 ± 10 min. Cells were washed once with PBS and lysed in the lysis bu€er (20 mM Tris-HCl, pH 7.4, 0.5 mM EDTA, 1% NP-40, 5% glycerol, 1 mM phenylmethylsulfonyl ¯uoride). The lysate was clari®ed by centrifugation at 15 000 g for 20 min and the resulting supernatant was incubated with 2 mg of anti-p85 subunit of PI-3' kinase monoclonal antibody for 3 h at 48C. The immunocom- plexes were collected with 30 ml of protein G-sepharose and Figure 7 Schematic illustration of intracellular signaling washed twice with each of the following bu€ers containing mediated by Ret kinase (I) 20 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1% NP-40 and 1mM dithiothreitol; (II) 100 mM Tris-HCl, pH 7.4,

500 mM LiCl2 and 1 mM dithiothreitol; and (III) 10 mM Tris-HCl, pH 7.4, 100 mM NaCl and 1 mM dithiothreitol. The immunoprecipitates were suspended in 50 mlof Immunoprecipitation and immunoblotting reaction bu€er (20 mM Tris-HCl, pH 7.4, 100 mM NaCl,

Cells were grown subcon¯uently in 60-mm dishes and 10 mM MgCl2,0.5mM EGTA) containing 100 mM phos- serum-starved for 6 h. Then the cells were stimulated with phatidylinositol, 100 mM phosphatidylserine, 120 mM ade- GDNF (150 ng/ml) for the indicated periods, washed once nosine and 10 mM [g-32P]ATP and incubated for 10 min at with ice-cold phosphate-bu€ered saline (PBS) and lysed in 308C. The reaction was terminated by adding 200 mlof1M radioimmunoprecipitation assay (RIPA) bu€er (20 mM HCl and 80 ml of chloroform/methanol (1 : 1, vol/vol). The Tris-HCl, pH 7.5, 150 mM NaCl, 2 mM EDTA, 1% products were separated in chloroform/methanol/25%

Nonidet P (NP)-40) containing 1 mM phenylmethylsulfo- NH4OH/water (43 : 38 : 5 : 7; vol/vol) by thin-layer chroma- nyl ¯uoride, 0.5 mM sodium orthovanadate and 0.1 mM tography. Radioactive phosphatidylinositol phosphate aprotinin. Cell lysates were centrifuged at 15 000 g for (PIP) was detected by autoradiography. 20mintoremovecelldebrisandincubatedwith2mgof antibodies for 3 h at 48C. The resulting immunocomplexes Staining of GDNF-treated cell with anti-paxillin antibody and were collected with Protein A sepharose (Sigma, St Louis, tetramethylrhodamine-isothiocyanate (TRITC)-phalloidin MO, USA) and washed four times with RIPA bu€er. The complexes were eluted in sodium dodecyl sulfate (SDS)- Cells were grown on glass coverslips and serum-starved for sample bu€er (20 mM Tris-HCl, pH 6.8, 2 mM EDTA, 2% 6 h. After stimulation with GDNF (150 ng/ml) for 15 min, SDS, 10% sucrose, 20 mg/ml bromophenol blue) by boiling the cells were washed twice with PBS and ®xed in 4% for5minandsubjectedtoSDS±polyacrylamidegel paraformaldehyde in PBS at room temperature for 10 min. electrophoresis. Separated proteins were transferred to Fixed cells were permealized with 0.15% Triton X-100, polyvinylidene di¯uoride membranes and reacted with incubated with anti-paxillin mouse monoclonal antibody antibodies. The reaction was examined by enhanced and stained with ¯uorescein-isothiocyanate (FITC)-con- chemiluminescence (Amersham Corp., UK). jugated goat anti-mouse IgG and tetramethylrhodamine- isothiocyanate (TRITC)-phalloidin (Sigma) in PBS for 30 min. Production of GST-fusion proteins and in vitro binding assay The SH2 domains of Grb2, Grb7, Grb10, Shc, Crk, Nck and p85 subunit of PI-3' kinase were cloned from their Acknowledgements cDNA or from mRNA of TGW-I-nu cells by reverse- We are grateful to K Imaizumi and J Aoki for technical transcriptase polymerase chain reaction, and inserted into assistance. This work was supported in part by grants-in- pGEX-3X vector (Pharmacia, Milwaukee, WI, USA). The aid for COE research, scienti®c research and cancer resulting recombinant plasmids were used to transform E. research from the Ministry of Education, Science, Sports coli, JM109. Puri®cation of the glutathione S transferase and Culture of Japan and by grants from the Cell Science (GST)-fusion proteins was performed as described pre- Research Foundation and the Mitsubishi Foundation.

References

Asai N, Iwashita T, Matsuyama M and Takahashi M. Beck KD, Valverde J, Alexi T, Poulsen K, Mo€at B, Vandlen (1995). Mol. Cell. Biol., 15, 1613 ± 1619. RA, Rosenthal A and Hefti F. (1995). Nature, 373, 339 ± Asai N, Murakami H, Iwashita T and Takahashi M. (1996). 341. J. Biol. Chem., 271, 17644 ± 17649. Bocciardi R, Mograbi B, Pasini B, Borrello MG, Pierotti Baloh RH, Tansey MG, Golden JP, Creedon DJ, Heuck- MA, Bourget I, Fischer S, Romeo G and Rossi B. (1997). eroth RO, Keck CL, Zimonjic DB, Popescu NC, Johnson Oncogene, 15, 2257 ± 2265. Jr EM and Milbrandt J. (1997). Neuron, 18, 793 ± 802. Phosphorylation of focal adhesion proteins by Ret activation H Murakami et al 1982 Borrello MA, Pelicci G Arighi E, De Filippis L, Greco A, OhiwaM,MurakamiH,IwashitaT,AsaiN,IwataY,Imai Bongarzone I, Rizzetti MG, Pelicci PG and Pierotti MA. T, Funahashi H, Takagi H and Takahashi M. (1997). (1993). Oncogene, 9, 1661 ± 1668. Biochem. Biophys. Res. Commun., 237, 747 ± 751. BorelloMA,SmithDP,PasiniB,BongarzoneI,GrecoA, OppenheimRW,HouenouLJ,JohnsonJE,LinL-FH,LiL, Lorenzo MJ, Arighi E, Miranda C, Eng C, Alberti L, Lo AC, Newsome AL, Prevette DM and Wang S. (1995). Bocciardi R, Mondellini P, Scopsi L, Romeo G, Ponder Nature, 373, 344 ± 346. BAJ and Pierotti MA. (1995). Oncogene, 11, 2419 ± 2427. Rankin S, Hooshmand-Rad R, Claesson-Welsh L and Brinson AE, Harding T, Diliberto PA, He Y, Li X, Hunter Rozengurt E. (1996). J. Biol. Chem., 271, 7829 ± 7834. D, Herman B, Earp HS and Graves LM. (1998). J. Biol. Rankin S, Morii N, Narumiya S and Rozengurt E. (1994). Chem., 73, 1711 ± 1718. FEBS Lett., 354, 315 ± 319. Buj-Bello A, Adu J, Pinon LGP, Horton A, Thompson J, RosselM,PasiniA,ChappuisS,GenesteO,FournierL, Rosenthal A, Chinchetru M, Buchman VL and Davies Schu€enecker I, Takahashi M, van Grunsven LA, AM. (1997). Nature, 387, 721 ± 724. Urdiales JL, Rudkin BB, Lenoir GM and Billaud M. Burridge K, Turner CE and Romer LH. (1992). J. Cell. Biol., (1997). Oncogene, 14, 265 ± 275. 119, 893 ± 903. Sanicola M, Hession C, Worley D, Carmillo P, Ehrenfels C, Guan J-L. (1997). Matrix Biology, 16, 195 ± 200. WalusL,RobinsonS,JaworskiG,WeiH,TizardR, Harte MT, Hildebrand JD, Burnham MR, Bouton AH and Whitty A, Pepinsky RB and Cate RL. (1997). Proc. Natl. Parsons JT. (1996). J. Biol. Chem., 271, 13649 ± 13655. Acad.Sci.USA,94, 6238 ± 6243. Holgado-Madruga M, Emlet DR, Moscatello DK, Godwin Santoro M, Carlomagno F, Romano A, Bottaro DP, Dathan AK and Wong AJ. (1996). Nature, 379, 560 ± 564. NA,GriecoM,FuscoA,VecchioG,MatoskovaB,Kraus Huang MM, Lipfert L, Cunningham M, Brugge JS, Ginsberg MH and Di Fiore PP. (1995). Science, 267, 381 ± 383. MH and Shattil S. (1993). J. Cell. Biol., 122, 473 ± 483. Sasaki T, Hazeki K, Hazeki O, Ui M and Katada T. (1996). ItoS,IwashitaT,AsaiN,MurakamiH,IwataY,SobueG Biochem. J., 315, 1035 ± 1040. and Takahashi M. (1997). Cancer Res., 57, 2870 ± 2872. Schaller MD and Parsons JT. (1994). Curr. Opin. Cell. Biol., Iwashita T, Asai N, Murakami H, Matsuyama M and 6, 705 ± 710. Takahashi M. (1996). Oncogene, 12, 481 ± 487. Schaller MD and Parsons JT. (1995). Mol. Cell. Biol., 15, JalinkK,vanCorvenEJ,HengeveldT,MoriiN,NarumiyaS 2635 ± 2645. and Moolenaar WH. (1993). J. Cell. Biol., 126, 801 ± 810. Schaller MD and Sasaki T. (1997). J. Biol. Chem., 272, Jian X, Rothenberg P, Kahn CR, Backer JM, Araki E, 25319 ± 25325. Wilden PA, Cahill DA, Goldstein BJ and White MF. Seu€erlein T and Rozengurt E. (1994). J. Biol. Chem., 269, (1991). Nature, 352, 73 ± 77. 9345 ± 9351. Jing S, Wen D, Yu Y, Holst PL, Luo Y, Fang M, Tamir R, Sinnett-Smith J, Zachary I, Valverde A and Rozengurt E. Antonio L, Hu Z, Cupples R, Louis JC, Hu S, Altrock BW (1993). J. Biol. Chem., 268, 14261 ± 14268. and Fox GM. (1996). Cell, 85, 1113 ± 1124. Tomac A, Lindqvist E, Lin L-FH, Ogren SO, Young D, Klein RD, Sherman D, Ho W-H, Stone D, Bennett GL, Ho€er BJ and Olson L. (1995). Nature, 373, 335 ± 339. Mo€at B, Vandlen R, Simmons L, Gu Q, Hongo J-A, Treanor JJS, Goodman L, de Sauvage F, Stone DM, Poulsen Devaux B, Poulsen K, Armanini M, Nozaki C, Asai N, KT, Beck CD, Gray C, Armanini MP, Pollock RA, Hefti Goddard A, Phillips H, Henderson CE, Takahashi M and F, Phillips HS, Goddard A, Moore MW, Buj-Bello A, Rosenthal A. (1997). Nature, 387, 717 ± 721. DaviesAM,AsaiN,TakahashiM,VandlenR,Henderson Kotzbauer PT, Lampe PA, Heuckeroth RO, Golden JP, CE and Rosenthal A. (1996). Nature, 382, 80 ± 83. Creedon DJ, Johnson Jr EM and Milbrandt J. (1996). Weidner KM, Di Cesare S, Sachs M, Brinkmann V, Behrens Nature, 384, 467 ± 470. J and Birchmeier W. (1996). Nature, 384, 173 ± 176. Kumagai N, Morri N, Fujisawa K, Nemoto Y and Narumiya Yan Q, Matheson C and Lopez OT. (1995). Nature, 373, S. (1993). J. Biol. Chem., 268, 24535 ± 24538. 341 ± 344. Li X and Earp HS. (1997). J. Biol. Chem., 272, 14341 ± 14348. Yao R and Cooper GM. (1995). Science, 267, 2003 ± 2006. Lin L-FH, Doherty DH, Lile JD, Bektesh S and Collins F. ZhengC,XingZ,BianZC,GuoC,AkbayA,WarnerLand (1993). Science, 260, 1130 ± 1132. Guan JL. (1998). J. Biol. Chem., 273, 2384 ± 2389. McDonald NQ and Hendrickson WA. (1993). Cell, 73, 421 ± 424. Nozaki C, Asai N, Murakami H, Iwashita T, Iwata Y, Horibe K, Klein RD, Rosenthal A and Takahashi M. (1998). Oncogene, 18, 293 ± 299.