Differential requirement of Tyr1062 multidocking site by RET isoforms to promote neural cell scattering and epithelial cell branching

Abstract

The receptor tyrosine kinase RET is alternatively spliced to yield two main isoforms, RET9 and RET51, which differ in their carboxyl terminal. Activated RET induces different biological responses such as morphological transformation, neurite outgrowth, proliferation, cell migration and branching. The two isoforms have been suggested to have separate intracellular signaling pathways and different roles in mouse development. Here we show that both isoforms are able to induce cell scattering of SK-N-MC neuroepithelioma cell line and branching tubule formation in MDCK cell line. However, the Y1062F mutation, which abrogates the transforming activity of both activated RET isoforms in NIH3T3 cells, does not abolish scattering and branching morphogenesis of RET51, whereas impairs these biological effects of RET9. The GDNF-induced biological effects of RET51 are inhibited by the simultaneous abrogation of both Tyr1062 and Tyr1096 docking sites. Thus, Tyr1096 may substitute the functions of Tyr1062. GRB2 is the only known adaptor protein binding to Tyr1096. Dominant-negative GRB2 expressed in MDCK cells together with RET9 or RET51 significantly reduces branching. Therefore, GRB2 is necessary for RET-mediated branching of MDCK cells.

Introduction

The RET gene encodes a transmembrane receptor tyrosine kinase with two main isoforms differing in their C-terminal tails (Takahashi et al., 1985, 1988; Ishizaka et al., 1989). RET is the signaling component of the multiprotein receptor complexes for glial cell line-derived neurotrophic factor (GDNF) family ligands (GFLs). The interaction between RET and GFLs is mediated by glycosyl-phosphatidylinositol (GPI)-membrane-anchored molecules, named GFRα1–4. GDNF, neurturin, artemin and persephin use GFRα1, GFRα2, GFRα3, and GFRα4 as the preferred receptors, respectively (Airaksinen and Saarma, 2002). However, GFLs appear to be able to cross-talk through different GFRαs in the presence of RET (Baloh et al., 1997; Jing et al., 1997; Sanicola et al., 1997).

RET is widely expressed in mammalian embryos and plays diverse roles in development (Taraviras and Pachnis, 1999; Baloh et al., 2000; Schedl and Hastie, 2000). During embryogenesis, the main sites of RET expression are in the nervous and excretory systems (Pachnis et al., 1993; Tsuzuki et al., 1995). Mice homozygous for a specific mutation of RET display intestinal aganglionosis and severe hypodisplasia or aplasia of the kidney (Schuchardt et al., 1994, 1996; Durbec et al., 1996; Srinivas et al., 1999).

Different alterations of RET are implicated in a number of diseases: activating germline mutations of RET confer a predisposition to multiple endocrine neoplasia type 2 (MEN2) A or B and familial medullary thyroid carcinoma (FMTC), while inactivating germ-line mutations cause Hirschsprung's disease (HSCR) (Pasini et al., 1996; Jhiang, 2000). Somatic RET rearrangements, named RET/PTCs, are present in about 1/3 of human papillary thyroid carcinomas (Pierotti et al., 1996; Alberti et al., 2003).

The RET gene is alternatively spliced to yield two main protein isoforms of 1072 (RET9 or short) or 1114 (RET51 or long) amino acids differing at the C-terminal region, by displaying 9 or 51 unrelated amino acids (Ishizaka et al., 1989). The intracellular domain of RET contains 14 tyrosine residues in the long and 12 in the short isoform, the latter lacking two tyrosine residues in the C-terminus. The RET9 and RET51 isoforms are evolutionary highly conserved over a broad range of species, suggesting that the two isoforms may have different roles in the physiological functions of RET (Carter et al., 2001). RET9 is critically important in kidney morphogenesis and enteric nervous system development, whereas RET51 appears dispensable (de Graaff et al., 2001). It has been suggested that RET51 may be involved in differentiation events in later kidney organogenesis (Lee et al., 2002).

Ligand-stimulated wild-type RET, as well as constitutive active oncogenic RET mutants, are phosphorylated at specific cytoplasmic tyrosine residues (Liu et al., 1996; Coulpier et al., 2002). Tyrosine autophosphorylation is required for downstream RET signaling, and interactions of RET with a number of downstream targets have been identified. Phosphorylated tyrosine residues Tyr905, Tyr1015 and Tyr1096 (the latter long isoform-specific) were identified as docking sites for the adapter proteins Grb7/Grb10, Phospholipase-Cγ and GRB2, respectively (Pandey et al., 1995, 1996; Borrello et al., 1996; Liu et al., 1996; Alberti et al., 1998). Tyr1062 is a multidocking site interacting with a number of transduction molecules: SHC, FRS2, IRS1/2, DOK proteins, Enigma and PKCα (Arighi et al., 1997; Lorenzo et al., 1997; Hennige et al., 2000; Grimm et al., 2001; Kurokawa et al., 2001; Melillo et al., 2001; Andreozzi et al., 2003). These Tyr1062-associated adapter proteins contribute to activation of several downstream signaling pathways such as the RAS/ERK, PI3K/AKT, p38MAPK, JNK and ERK5 (Besset et al., 2000; Hayashi et al., 2000; Kurokawa et al., 2001; Segouffin-Cariou and Billaud, 2000). Activation of these pathways mediates RET-dependent biological responses, such as neoplastic transformation of NIH3T3 cells, survival and neurite outgrowth of pheochromocytoma PC12 cells, scattering of neuroepithelioma SK-N-MC cells and branching tubule formation of kidney epithelial MDCK cells (Asai et al., 1996; van Puijenbroek et al., 1997; Tang et al., 1998; De Vita et al., 2000). The Tyr1062 multidocking site is essential for mitogenesis, survival and transformation activities promoted by RET (Durick et al., 1995; Asai et al., 1996; De Vita et al., 2000; Segouffin-Cariou and Billaud, 2000; Mercalli et al., 2001). Although both the short and long isoforms of RET have been used in previous studies, an exhaustive comparison of biological effects induced by the two RET isoforms has yet to be performed.

Here we examined ligand-dependent cell scattering of the human neuroepithelioma SK-N-MC cell line induced by exogenously expressed RET isoforms. Cell scattering is characterized by loss of epithelial features and beginning of individual cell motion. The scattering effect in SK-N-MC cells is associated with a reduced growth rate and altered cell morphology resulting from the acquisition of a more differentiated phenotype (van Puijenbroek et al., 1997).

In this study, we showed that both RET isoforms induce GDNF-dependent cell scattering. Surprisingly though, we found that the Y1062F mutation did not abolish the ability of SK-N-MC cells expressing the long isoform of RET to scatter. In the kidney MDCK cell line, in which neither GFRα nor RET are endogenously expressed, the transfected short isoform of RET promotes branching morphogenesis when activated by GDNF and soluble GFRα1 (Tang et al., 1998). We showed that both the exogenously introduced RET isoforms induce cell branching of MDCK cells. The Y1062F mutation impairs the branching morphogenesis of RET9 but not of RET51. The long isoform-specific Tyr1096 of RET replaces Tyr1062 functions and allows GDNF-dependent cell scattering and branching morphogenesis. Nevertheless, the Tyr1062 is required by the long isoform of RET to promote cell transformation in NIH3T3 cells. Moreover, our results indicate GRB2 as the adapter which plays a significant role in RET-dependent cell differentiation effects.

Results

RET9 and RET51 isoforms mediate GDNF-induced SK-N-MC(RET) cell scattering

To compare the biological effects of the two main RET isoforms, RET9 and RET51 (Figure 1a), in a neural cell environment, we analysed RET-induced cell scattering in the human neuroectodermal tumor cell line SK-N-MC (herein referred as MC). The scattering response starts with cell dissociation followed by active migration. MC cells endogenously express GFRα co-receptors, but not RET. To examine the biological effects of RET isoforms, we generated MC cell lines stably expressing either the short or long isoform of RET, herein referred as MC(RET9) and MC(RET51), respectively.

Figure 1
figure1

Ligand-induced SK-N-MC(RET) cell scattering. (a) Schematic representation of the two RET isoforms, RET9 and RET51, with the signal peptide (SP), cadherin-like (CAD), cysteine-rich (CYS), transmembrane (TM) and tyrosine kinase (KD) domains. The Tyr1062 and Tyr1096 docking sites are marked. (b) Phase-contrast micrographs of SK-N-MC parental cells (referred to as MC) and MC cells expressing RET isoforms, MC(RET9) and MC(RET51), (magnification × 10). The cells were incubated in the absence or presence of 10 μ M RPI1/Cpd1, the RET inhibitor, for 1 h and then treated with 100 ng/ml of GDNF or NTN for 16 h, as indicated. The number of scattered cells was determined as described in ‘Materials and methods’. Average of the number of scattered cells per field of MC(RET9)+GDNF was set at 1.0. At least two different clones of each RET mutation were used for the analyses. MC parental cells were used as a control. Means±s.e.m. of three independent experiments are shown in the graph. MC(RET9) vs MC(RET9)+GDNF/NTN, P<0.005; MC(RET51) vs MC(RET51)+GDNF/NTN, P<0.005. Differences between MC(RET) cells+GDNF vs+NTN were not statistically significant. (c) MC(RET9) and MC(RET51) cells were serum starved, incubated for 1 h with 10 μ M RPI1/Cpd1 and then treated with 100 ng/ml GDNF or NTN for 10 min. In all, 1 mg of PLCLB cell extracts was immunoprecipitated with the anti-Ret common I antibody and immunoblotted with anti-Ret or anti-pTyr antiserum

MC(RET9) and MC(RET51) cells, but not parental MC cells, scattered when stimulated by either GDNF or neurturin (NTN) (Figure 1b). The expression of RET proteins in MC cells was comparable and their tyrosine phosphorylation was ligand-dependent (Figure 1c).

Treatment of RET-expressing MC cells with RPI1/Cpd1, a chemical inhibitor of the kinase activity of RET (Lanzi et al., 2000), completely blocked MC(RET9) and MC(RET51) scattering induced by GDNF (Figure 1b). The inhibition of cell scattering was due to the inhibition of RET autophosphorylation by RPI1/Cpd1 (Figure 1c). MC cells expressing a RET protein lacking kinase activity, MC(RET9-758R), were also unable to display GDNF-dependent scattering (data not shown).

These results demonstrated that both RET isoforms promote GDNF/NTN-dependent cell scattering of MC cells and that RET kinase activity is required for inducing this cellular response.

Y1062F mutation abrogates scattering of MC(RET9) but not MC(RET51) cells

To determine whether Tyr1062 of RET is required for MC cell scattering, we generated MC cell lines expressing either the short or long RET isoform devoid of the multidocking site, MC(RET9-1062F) and MC(RET51-1062F) cells, respectively. GDNF stimulation of MC(RET9-1062F) cells did not induce cell scattering (Figure 2), thus suggesting that Tyr1062 in the short isoform of RET is necessary for scattering response. Unexpectedly, MC(RET51-1062F) cells were able to scatter upon GDNF stimulation, although to a less extent than MC(RET51) cells. Tyr1096 is the only long isoform-specific phosphorylated tyrosine residue. To analyse its role in RET long isoform biological activity, we generated MC cell lines expressing RET51-1096F and the double mutant RET51-1062F/1096F. The Y1096F mutation barely affected GDNF-induced cell scattering. Cells expressing the RET51-1062F/1096F double mutant protein were unable to scatter (Figure 2). Quantification of GFRα mRNAs from at least two different SK-N-MC clones of each RET mutant protein ruled out that differences in GFRα expression levels were responsible for the different abilities to scatter (data not shown). Altogether, these results suggested that the Tyr1096 functionally substitutes the Tyr1062 multidocking site in RET-dependent cell scattering.

Figure 2
figure2

Role of the RET Tyr1062 and Tyr1096 docking sites in GDNF-induced MC(RET) cell scattering. MC clones stably expressing wild-type or a mutated short or long isoform of RET were left untreated or treated for 16 h with 100 ng/ml of GDNF. A representative MC(RET) cell line is shown. The number of scattered cells was determined as described in ‘Materials and methods’. Average of the number of scattered cells per field of MC(RET9)+GDNF was set at 1.0. At least two different clones of each RET mutation were used for the analyses. Means±s.e.m. of three independent experiments are shown in the graph. ***P<0.005; MC(RET9)+GDNF vs MC(RET9-1062F)+GDNF, P<0.005; MC(RET51)+GDNF vs MC(RET51-1062F)+GDNF, P<0.005; MC(RET51)+GDNF vs MC(RET51-1096F)+GDNF, P<0.05; MC(RET51)+GDNF vs MC(RET51-1062F/1096F)+GDNF, P<0.005; MC(RET51-1062F)+GDNF vs MC(RET51-1096F)+GDNF, P<0.005. The expression of RET proteins in these clones is shown in Figure 3

Activation of AKT, ERK1/2 and p38 signal transducers correlates with RET-dependent cell scattering

To define which biochemical pathways triggered by the RET multidocking site are required for MC cell scattering, the phosphorylation of different signal transducers was analysed in the above-described GDNF-stimulated MC(RET) cell lines (Figure 3). SHC proteins were phosphorylated in all MC cells expressing RET with a functional phosphorylated Tyr1062, in contrast to MC cells expressing RET proteins lacking Tyr1062 and parental MC cells. AKT, ERK1/2 and p38 signal transducers were activated in GDNF-stimulated MC cells expressing wild-type RET proteins. The presence of the Y1062F mutation in the short but not long isoform of RET blocked the activation of AKT, ERK and p38 by GDNF. The presence of the Y1096F mutation did not affect the activation of these transducers. On the contrary, the double mutant of the RET long isoform, RET51-1062F/1096F, was unable to activate these pathways. Taken together, our results showed a correlation between the scattering response and the activation of AKT, ERK1/2 and p38 pathways.

Figure 3
figure3

Analysis of the GDNF-activated signaling pathways in MC(RET) cells. Parental and MC clones stably expressing wild-type or mutated RET proteins were serum starved and left untreated or treated with 100 ng/ml GDNF for 10 min. In all, 1 mg of PLCLB cell lysates was immunoprecipitated with anti-Ret common I antiserum and blotted with anti-pTyr or anti-Ret common II. Also, 0.5 mg of the same PLCLB cell extracts were immunoprecipitated with anti-Shc antibodies, followed by immunoblotting with anti-pTyr or anti-Shc antibodies. In total, 50 μg of SDS total cell lysates were subjected to immunoblotting with anti-Akt or anti-phospho(Ser473)Akt and anti-Erk1/2 or anti-phospho(Thr202/Tyr204)Erk1/2 antibody. The p38 activity was assessed by in vitro kinase assay as the radioactivity incorporated into recombinant ATF2. Arrows point out the specific proteins recognized by the indicated antisera

Effect of chemical inhibitors on GDNF-induced MC(RET) cell scattering

To further analyse the involvement of ERK, AKT and p38 signaling pathways in cell scattering, GDNF-activated MC(RET) cells were treated with the specific chemical inhibitors: PD98059 (MEK1 inhibitor), Wortmannin (PI3 K inhibitor), SB203580 or SB202190 (p38 inhibitors), SB202474 (mock control of the p38 inhibitor, data not shown). All these inhibitors have been previously demonstrated to be effective and specific for their targets on MC cells (Iwahashi et al., 2002). None of these compounds were able to completely inhibit RET-induced cell scattering as the RPI1/Cpd1 RET-specific inhibitor (Figure 4). On the other hand, each specific inhibitor caused a partial but significant inhibition of cell scattering. No differences were observed between cells expressing either of the two RET isoforms. These data suggested that each analysed pathway, ERK1/2, AKT/PI3 K and p38, might be necessary but not sufficient for cell scattering.

Figure 4
figure4

Effects of chemical inhibitors on GDNF-induced MC(RET) cells scattering. MC clones stably expressing wild-type or mutated RET proteins were pre-incubated for 1 h with 40 μ M PD98059 (MEK1 inhibitor), 30 nM Wortmannin (PI3 K inhibitors), 10 μ M SB202190 (p38 inhibitor) or left untreated. GDNF (100 ng/ml) was added to the cells for 16 h and cell scattering was analysed. The number of scattered cells was determined as described in ‘Materials and methods’. Average of the number of scattered cells per field of MC(RET9)+GDNF was set at 1.0. At least two different clones of each RET mutation were used for the analyses. Means±s.e.m. of at least three independent experiments are shown in the graph. MC(RET9) vs MC(RET9)+GDNF, P<0.005; MC(RET51) vs MC(RET51)+GDNF, P<0.005. Differences between MC(RET)+GDNF vs MC(RET)+GDNF in the presence of any chemical inhibitor were statistically significant

Y1062F mutation abrogates RET9- but not RET51-mediated branching of MDCK cells

To determine whether the role of Tyr1062 and Tyr1096 in cell differentiation was cell type-specific, we established MDCK cell lines expressing wild-type RET isoforms and the different RET mutants previously described. Equal amounts of RET proteins were expressed by MDCK(RET) cell lines. RET proteins were tyrosine phosphorylated in the presence of soluble GFRα1 (sGFRα1) and GDNF (Figure 5b). MDCK(RET9) cells have been shown to form tubule-like structures in the presence of sGFRα1 and GDNF (Tang et al., 1998). We tested the ability of MDCK cells expressing either wild-type or mutant RET proteins to form branching tubules in three-dimensional collagen gel in response to GDNF/sGFRα1 (Figure 5a). Both RET9 and RET51 promoted GDNF/sGFRα1-dependent branching of MDCK cells.

Figure 5
figure5

GDNF induces branching of MDCK cells expressing RET51-1062F and RET51-1096F, but not RET51-1062F/1096F. (a) MDCK clones stably expressing wild-type or mutated RET9 or RET51 were put in collagen gel culture and grown with sGFRα1 (100 ng/ml) alone or GDNF (100 ng/ml) together with sGFRα1 (100 ng/ml). After 3 days, the cells were fixed and counted as described in ‘Materials and methods’. At least two different clones of each RET mutation were used for the analyses and no discrepancies were found between them. Mock-transfected MDCK cells were used as a control. The morphology of MDCK(RET) cells is shown (magnification × 40). Only the tubules longer than two cell body diameters were counted. The results are representative of at least three independent experiments. Means±s.e.m. of three to six independent repeats are shown in the graph, ** P<0.01, ***P<0.005. (b) MDCK clones stably expressing wild-type or mutated RET proteins were treated with GDNF (100 ng/ml) together with sGFRα1 (100 ng/ml) for 2 h. In all, 1 mg of PLCLB cell lysates was immunoprecipitated with anti-Ret common I antiserum and blotted with anti-pTyr or anti-Ret common II

The Y1062F mutation impaired branching morphogenesis of MDCK(RET9-1062F) cells. The presence of either the Y1062F or Y1096F mutation in the long isoform of RET did not affect the ability of MDCK(RET51) cells to branch, while branching activity was inhibited in MDCK(RET51-1062F/1096F) cells. Thus, in agreement with the results obtained using MC(RET) cells, Tyr1096 functionally substitutes the Tyr1062 multidocking site to induce differentiation events in MDCK cells.

RET-dependent branching of MDCK cells requires GRB2-coupled pathways

Since GRB2 is the only transducer known to bind Tyr1096 of RET, we examined the ability of the above-described RET mutants, transiently expressed in HEK293T cells, to bind GRB2 in pull-down experiments. In accord with previous results obtained with RET/PTC2 and RET-634R oncogenic proteins (Liu et al., 1996; Alberti et al., 1998), the Y1096F mutation strongly reduced while the simultaneous mutations of both Tyr1062 and Tyr1096 completely abolished GRB2 binding to RET51 (data not shown).

To verify whether GRB2 has a role in promoting cell branching, we used a dominant-negative GRB2 mutant (Tanaka et al., 1995), herein referred to as DN-GRB2, to interfere with signaling pathways triggered by Tyr1096. DN-GRB2 was transiently transfected in MDCK(RET) cells. Wild-type GRB2 and green fluorescent protein (GFP) were used as controls (Figure 6).

Figure 6
figure6

DN-GRB2 inhibits branching of RET-expressing MDCK cells. MDCK cells stably expressing RET9, RET51, RET51-1062F or RET51-1096F were transiently co-transfected with pEGFP-N1, encoding the GFP and wild-type GRB2 (WT-GRB2), dominant-negative GRB2 (DN-GRB2) or pEGFP-N1 alone. The day after transfection, cells were put in collagen gel culture and grown with sGFRα1 (100 ng/ml) alone, or GDNF (100 ng/ml) together with sGFRα1 (100 ng/ml). After 2 days, cultures were fixed and fluorescent cysts were counted under the fluorescent microscope. The percentage of branching cysts was calculated. Only the tubules with more than two cell body diameters were counted. The morphology of RET-expressing MDCK cells is shown. The results are representative of three independent experiments. Means±s.e.m. of three independent repeats are shown in the graph, **P<0.01, ***P<0.005

The presence of DN-GRB2 in MDCK(RET) cells expressing wild-type RET proteins caused a partial reduction of GDNF-dependent branching morphogenesis. Reduction, but not complete abolition, of cell branching was also observed in GDNF-stimulated MDCK(RET51-1062F) cells. DN-GRB2 completely abrogated the ability of MDCK(RET51-1096F) cells to form tubules, leading us to hypothesize that, besides GRB2, other unknown transducer proteins may be capable of binding Tyr1096. Taken together, these results indicated that GRB2 has a role in the differentiation of MDCK cells induced by activated RET.

Tyr1062 is required for transforming activity induced by both RET isoforms activated by MEN2A-associated C634R mutation

To further study the role of Tyr1062, we analysed the transforming potential of RET9 and RET51 genes carrying two mutations: C634R, associated with multiple endocrine neoplasia type 2A (MEN2A), and Y1062F. The constructs RET9-634R/1062F and RET51-634R/1062F (Lorenzo et al., 1997) were used in NIH3T3 cells focus formation assay. As positive control, the constructs RET9-634R and RET51-634R were used. RET51-634R/1062F and RET9-634R/1062F showed a strongly reduced transforming activity (0.09 and 0.01, respectively) when compared with the corresponding oncogenic forms with a functional Tyr1062 multidocking site (Figure 7). Therefore, both RET isoforms require a functional Tyr1062 for their transforming activity.

Figure 7
figure7

Tyr1062 is required for transforming activity induced by both RET isoforms activated by MEN2A-associated C634R mutation. NIH3T3 cells were stably transfected with RET9-634R, RET51-634R, RET9-634R/1062F or RET51-634R/1062F. NIH3T3 G418-resistant colonies and transformed foci were fixed and counted 14 and 21 days after transfection, respectively. Transforming activity was calculated by normalizing the number of transformed foci for that of G418-resistant colonies. Each bar is the average of three plates of G418R colonies and three plates of foci from two independent transfection experiments. The transforming activity of NIH3T3(RET-634R) was set at 1.0

Discussion

We show here that two cell differentiation effects induced by the RET receptor complex, namely cell scattering of SK-N-MC neuroepithelioma cell line and branching morphogenesis of MDCK epithelial cells, are promoted by both isoforms of RET. However, mutation of the RET Tyr1062 docking site, which abolishes the transforming activity in NIH3T3 cells of both activated RET isoforms, only impairs cell scattering and branching morphogenesis induced by the short, but not by the long, isoform.

The RET gene is crucial in the development of enteric nervous system and kidney (Schuchardt et al., 1994, 1996; Durbec et al., 1996; Srinivas et al., 1999). Each RET isoform plays a distinct role in mouse development. RET9 has been suggested to be essential for early development of the kidney and the enteric nervous system. RET51 seems to have a role in later kidney organogenesis (Lee et al., 2002). The different physiological roles between the two RET isoforms are in agreement with the high evolutionary conservation of RET splicing variants over a wide range of species (Carter et al., 2001).

It has been shown that the short isoform of RET promotes cell branching of MDCK cells (Tang et al., 1998) and in the chimeric form EGFR/RET induces cell scattering of SK-N-MC cell line (van Puijenbroek et al., 1997). RET-transfected SK-N-MC cells have been used to investigate GDNF-induced intracellular signals and biological effects such as lamellipodia formation and cell rescue from apoptosis (van Weering and Bos, 1997; Hayashi et al., 2000; Iwashita et al., 2001; Mograbi et al., 2001a, 2001b; Fukuda et al., 2002; Iwahashi et al., 2002). GFRα1 and GFRα2 were reported to be expressed in SK-N-MC cells (Hayashi et al., 2000). However, by Western Blotting, we could detect neither endogenous GFRα1 nor GFRα2 proteins in SK-N-MC cells. GFRα1 and GFRα2 mRNAs were present in parental SK-N-MC and derived clones by real-time PCR even if GFRα1 levels are nearly null in terms of absolute expression (data not shown). Nevertheless, the ability of GDNF to cause RET phosphorylation through both GFRα1 and GFRα2 (Sanicola et al., 1997; Trupp et al., 1998) makes its use suitable for RET stimulation in SK-N-MC cells.

An exhaustive analysis which compares the biological activity of each RET isoform in different cell lines has not been yet performed. We have analysed GDNF-induced cell branching and scattering activities using both RET isoforms and different mutant RET proteins lacking the docking tyrosine residues Tyr1062 and Tyr1096. The Tyr1096 is unique to the long isoform (Liu et al., 1996). Tyr1062, although common to both isoforms, is only two residues amino-terminal to the C-terminal RET splice site, which thus alters the context of this residue between the short and long isoforms. In addition, the two isoforms, although sharing identical extracellular domains, do not complex with each other and show different tyrosine-phosphorylated associated proteins in sympathetic neurons (Tsui-Pierchala et al., 2002).

Here we show that the multidocking site Tyr1062 of RET is required for inducing cell scattering and branching by the short but not by the long isoform. In fact, the long isoform-specific Tyr1096 replaces Tyr1062 functions for promoting both biological effects. In contrast, the Tyr1062 is essential in both RET isoforms to activate pathways leading to cell transformation.

It has been previously shown that various signaling pathways including RAS/ERK, PI3K/AKT and p38 are activated by the phosphorylation of Tyr1062 of RET (Besset et al., 2000; Hayashi et al., 2000, 2001; Segouffin-Cariou and Billaud, 2000). However, the mutation of Tyr1062 only partially affects the PI3K/AKT pathway, which can be activated via Tyr1096 (Besset et al., 2000). We show here that ERK, AKT and p38 signaling transducers were not activated by the short isoform of RET bearing the Y1062F mutation. Nevertheless, the same mutation in the long isoform did not affect the activation of these pathways. The presence of active ERK, AKT and p38 correlated with scattering of the GDNF-stimulated SK-N-MC(RET) cells. In addition, SK-N-MC cells expressing the long isoform RET double mutant, devoid of both Tyr1062 and Tyr1096, showed inactive ERK, AKT and p38 proteins, and were unable to scatter. These results indicate the existence of alternative pathways leading to AKT, ERK and p38 activation independently of phosphorylation of Tyr1062, suggesting that Tyr1062 and Tyr1096 docking sites may activate redundant signaling pathways leading to cell scattering. A barely increase of p52 SHC and ERK phosphorylation by the double mutant RET51 upon GDNF stimulation was observed. The residual phosphorylation might be the effect of the signaling triggered by a different phosphorylated RET tyrosine, for example, Tyr1090 (Kawamoto et al., 2004) or by other RET51-specific sequences able to activate SHC and other signaling transducers. Indeed, mock-transfected SK-N-MC cells did not show GDNF-dependent SHC phosphorylation, thus ruling out the possibility of a RET-independent effect. However, p52 SHC and ERK phosphorylation levels induced by the double mutant RET51 are likely not sufficient to pass the threshold for producing biological effects, such as cell scattering and branching morphogenesis.

To exclude that the capacity of Tyr1096 to replace the multidocking site Tyr1062 was cell-type specific, we analysed the RET-induced differentiation activity in epithelial cells. MDCK cells expressing the short isoform of RET have been shown to display tubule formation in the presence of soluble GFRα1 and GDNF (Tang et al., 1998). We demonstrated that both RET isoforms were able to promote GDNF-dependent branching activity in MDCK cells. Consistent with SK-N-MC(RET) cell results, the Tyr1062 was required for branching morphogenesis only by the short isoform of RET. The simultaneous lack of the two docking sites Tyr1062 and Tyr1096 completely impaired the GDNF-dependent branching activity of MDCK(RET51) cells.

GRB2 is the only transducer known to directly bind Tyr1096 of RET, while it is recruited via SHC or FRS2 by the Tyr1062 (Liu et al., 1996; Arighi et al., 1997; Alberti et al., 1998; Kurokawa et al., 2001). GRB2, which is a well-known link between receptor tyrosine kinases and RAS/ERK pathway, might recruit cell-specific transducers activating ERK and p38 pathways in SK-N-MC, and probably in MDCK cells, but not in other cell lines, such as NIH3T3 fibroblasts (Besset et al., 2000). By using the dominant-negative GRB2 protein (Tanaka et al., 1995), we were able to show that GRB2 has a role in RET-induced differentiation of MDCK cells. In fact, dominant-negative GRB2 partially inhibited cell branching promoted by both RET isoforms. Branching morphogenesis was entirely blocked in MDCK cells expressing RET51-1096F, the long isoform RET lacking the GRB2 docking site. The partial reduction of branching of MDCK(RET9) cells despite the absence of Tyr1096 docking site might result from the differences in the amino acids surrounding Tyr1062 between short and long isoforms. The Tyr1062 is indeed only two residues amino-terminal to the C-terminal RET splice site, which thus alters the context of this residue between the two isoforms. Consistent with this, Tyr1062 multidocking site in short and long isoforms does appear to have differential interactions with SHC and ENIGMA proteins (Lorenzo et al., 1997; Borrello et al., 2002). Consequently, the Tyr1062 of the short isoform might be able to selectively activate unidentified transducers, possibly different from GRB2, which are required to induce cell-branching activity. The reduction but not abolition of cell branching by DN-GRB2 in GDNF-stimulated MDCK(RET51-1062F) cells also suggests the existence of unknown transducer proteins, besides GRB2, that may be able to bind Tyr1096.

To further characterize the role of Tyr1062 in the context of each RET isoform, we examined the transforming activity of both oncogenic RET-MEN2A isoforms lacking the Tyr1062 multidocking site. We found that both RET isoforms required the Tyr1062 docking site for promoting transformation of NIH3T3 cells. Our findings extend previous reports on the transforming activity of the rearranged oncogene RET/PTC2 (Mercalli et al., 2001) and of the RET long isoform carrying the MEN2A- or MEN2B-associated mutation (Asai et al., 1996). Altogether, these results indicate that signaling pathways activated by Tyr1062 and Tyr1096 leading to the transforming activity promoted by oncogenic RET proteins are not redundant.

In summary, this study provides novel information on different roles of the two RET isoforms. In addition to the previously demonstrated distinct roles in mouse development, we have shown a difference in the ability of the two isoforms to recruit signaling transducers involved in differentiation pathways. Our results also suggest that GRB2 plays a crucial role in these pathways.

Materials and methods

Plasmid construction

The cloning of RET9 and RET51 in pCEP9β has been previously reported (Borrello et al., 1995; Lorenzo et al., 1997). The RET cDNA inserts were excised from the pCEP9β vector with HindIII and NotI and cloned into the mammalian expression vector pcDNA3 (Invitrogen) carrying the Geneticin resistance gene or in pCEP4 vector (Invitrogen) carrying the Hygromycin resistance gene. The mutants RET9-758R, RET9-1062F, RET51-1062F, RET51-1096F and RET51-1062F/1096F were generated by using the GeneEditor kit (Promega) and the oligonucleotides 5′-IndexTermACGGTGGCCGTGAGGATGCTGAA-3′ (RET9-758R), 5′-IndexTermACAAACTCTTTGGTAGAATTT-3′ (RET9-1062F), 5′-IndexTermGAAAACAAACTCTTTGGCATGTCAGAC-3′ (RET51-1062F) and 5′-IndexTermATGATAGTGTATTTGCTAACTGGATGC-3′ (RET51-1096F) as primers. The cDNA inserts were cloned into the mammalian expression vectors pCEP4 and pcDNA3. RET9-634R and RET51-634R were kindly provided by Bruce A Ponder. RET9-634R/1062F and RET51-634R/1062F were kindly provided by Darrin P Smith (Lorenzo et al., 1997).

Cell cultures and transfections

SK-N-MC human primitive neuroectodermal tumor cell line (van Weering et al., 1995) was cultured in RPMI 1640 supplemented with 15% fetal bovine serum (FBS). Early-passage MDCK, NIH3T3 and HEK293T cells were cultured in MEM medium containing 10% FBS. SK-N-MC cells were stably transfected using ExGen 500 (MBI Fermentas) with the following RET cDNAs: RET9, RET9-758R, RET9-1062F, RET51, RET51-1062F, RET51-1096F and RET51-1062F/1096F, cloned in pCEP4. The day before the transfection, 5 × 104 cells were seeded in a 6-cm diameter dish. In all, 3 μg of pCEP4 plasmid DNA and 17 μl of ExGen 500 were used for each transfection. The RET-expressing clones were selected in complete medium supplemented with 150 μg/ml hygromycin B (Roche) and then cultured in the presence of 50 μg/ml hygromycin B. RET-expressing clones were selected by Western blotting analysis. MDCK cells were stably transfected using Fugene 6™ reagent (Roche Diagnostics) with the following RET cDNAs: RET9, RET9-1062F, RET51, RET51-1062F, RET51-1096F and RET51-1062F/1096F, cloned in pcDNA3. Stable lines were selected in 400 μg/ml G418 (GibcoBRL, Life Technologies) and then cultured in 200 μg/ml G418. RET-expressing clones were selected by Western blotting analysis. In each experiment, at least two independent clones were tested. Transient transfections of MDCK cells were performed with Lipofectamine 2000 (Invitrogen Corp.) following the supplier's protocol. MDCK cells, stably expressing RET9, RET51, RET51-1062F or RET51-1096F, were co-transfected with the pEGFP-N1 vector (Clontech Laboratories, Inc.) and pcDNA3-GRB2, pcDNA3-GRB2-W36,193K (double SH3 mutant), or with pEGFP-N1 vector alone. The day after transfection (transfection efficiency was approximately 80%), the cells were put in collagen gel and cultured in media, supplemented with 100 ng/ml soluble GFRα1 (sGFRα1) or sGFRα1 together with 100 ng/ml GDNF. After 2 days, cultures were fixed and fluorescent cysts were counted under the fluorescent microscope. The percentage of branching cysts was calculated. For each MDCK cell line, two independent clones were co-transfected with GRB2 constructs and analysed. No significant discrepancies between clones were observed in GDNF-induced branching response. The pcDNA3-GRB2 (WT-GRB2) and pcDNA3-GRB2-W36,193K (DN-GRB2) plasmids were kindly provided by Bruce J Mayer and Mordechai Anafi (Tanaka et al., 1995).

For cell-scattering experiments, SK-N-MC(RET) cells were allowed to grow as discrete colonies in complete medium, incubated in the absence or presence of each chemical inhibitor for 1 h and then treated with 100 ng/ml of GDNF or NTN for 16 h. From the total number of counted cells in eight fields under a light microscope, the number of scattered cells per field was calculated. A cell was considered as a ‘scattered cell’ when it had lost contact with its neighbours and exhibited a fibroblast-like phenotype.

NIH3T3 (2 × 105 per 10-cm diameter dish) cells were stably transfected by calcium phosphate co-precipitation using 200 ng of plasmid DNA and 40 μg of mouse carrier DNA. G418-resistant colonies were selected in DMEM plus 10% calf serum and G418 antibiotic (0.5 mg/ml). Transformation foci were selected in DMEM containing 5% calf serum. NIH3T3 G418-resistant colonies and transformed foci were fixed and counted 14 and 21 days after transfection, respectively. HEK293 T cells were transiently transfected by calcium phosphate precipitation.

GDNF and hNTN were purchased from Alomone Labs. Recombinant Rat GFRα-1/Fc chimera was from R&D Systems. All the indicated chemical inhibitors but RPI1/Cpd1 were purchased from Calbiochem. RPI1 (formerly Cpd1) was kindly provided by Lanzi (Lanzi et al., 2000).

Branching tubule formation assays

Trypsinized MDCK cells, stably expressing RET9, RET51, RET9-1062F, RET51-1062F, RET51-1096F or RET51-1062F/1096F, were mixed 1 : 3 with collagen type I solution and plated in 24-well plates. MEM with 10% FBS was overlaid on the gels with both 100 ng/ml GDNF and sGFRα1 or 100 ng/ml sGFRα1 alone. Cells in collagen were cultured for 3 days, fixed by 2.5% glutaraldehyde in PBS and counted under the light or fluorescent microscope (indicated in the text). The percentage of branching cysts was calculated. Each condition was repeated at least three times. The average number and s.e.m. were calculated. Statistical significance was determined using the Student's t-test.

Antibodies

Anti-Ret common I and II rabbit polyclonal antibodies were developed, respectively, against the 1011–1027 and 1000–1014 amino acids of the C-terminal sequence common to the two RET isoforms, as previously described (Borrello et al., 2002). Anti-phosphoTyrosine (anti-pTyr) monoclonal antibody and anti-Shc polyclonal antibodies were purchased from Upstate Biotechnology, Anti-Akt, anti-phospho(Ser473)Akt, anti-Erk1/2, anti-phospho(Thr202/Tyr204)Erk1/2 and anti-p38 antibodies were purchased from Cell Signaling. ATF2 (1–96) was from Santa Cruz.

Immunoprecipitation and Western Blotting

Culture cells were lysed in PLCLB buffer (50 mM HEPES (pH 7.4), 150 mM NaCl, 10% glycerol, 1% Triton X-100, 1.5 mM MgCl2, 1 mM EGTA, 10 mM Na4P2O7, 10 mM NaF) supplemented with a protease inhibitor cocktail (Roche) and with 1 mM Na3VO4. For total cell lysates, cells were lysed in SDS lysis buffer (62.5 mM Tris–HCl (pH 6.8), 2% SDS). In all, 0.5–1 mg of PLCLB extracts were immunoprecipitated with the anti-Ret common I and anti-Shc antisera, as indicated. Total extracts and immunoprecipitates were resolved by SDS–PAGE and transferred onto nitrocellulose membranes. The filters were blocked with 2% BSA or 5% milk and immunoblotted. Immunoreactive bands were detected by using a horseradish peroxidase conjugate anti-rabbit or anti-mouse antiserum and ECL detection system (Amersham).

p38 kinase assay

Culture cells were lysed in lysis buffer (25 mM HEPES, 20 mM β-glycerol-phosphate, 0.1 mM Na3VO4, 0.1% Triton X-100, 0.3 M NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, 0.5 mM DTT) supplemented with a protease inhibitor cocktail (Roche). In total, 0.4 mg of cell extracts were mixed with 4 μl of recombinant ATF2. The immune complexes were washed three times with washing buffer (20 mM HEPES, 50 mM NaCl, 2.5 mM MgCl2, 0.1 mM EDTA, 0.05% Triton X-100). The kinase assay was performed in 40 μl of incubation buffer (50 mM β-glycerol-phosphate, 0.1 mM Na3VO4, 10 mM MgCl2, 10 μCi γ-32P-ATP) at 30°C for 30 min. The reaction was stopped by adding 500 μl of Washing Buffer. The samples were dried, eluted with Laemmli buffer and subjected to 15% SDS–PAGE. The polyacrylamide gels were dried and the radioactivity incorporated into GST-ATF2 was revealed by autoradiography.

In vitro binding experiments using GST-fused protein

Bacterial cultures expressing recombinant pGEX containing GRB2-SH2 domain were grown in LB medium containing 100 μg/ml ampicillin and induced with 1 mM isopropyl β-D-thiogalactopyrenoside for 3 h. The GST-fused protein was purified as previously described (Arighi et al., 1997). Culture cells were starved o/n and lysed in RIPA buffer (20 mM Tris–HCl (pH 7.5), 150 mM NaCl, 5 mM EDTA, 1% NP40); the cellular lysates were incubated with 2 μg of immobilized GST-fused protein for 60 min at 4°C. Protein complexes were resolved by SDS–PAGE and analysed by Western blotting.

References

  1. Airaksinen MS and Saarma M . (2002). Nat. Rev. Neurosci., 3, 383–394.

  2. Alberti L, Borrello MG, Ghizzoni S, Torriti F, Rizzetti MG and Pierotti MA . (1998). Oncogene, 17, 1079–1087.

  3. Alberti L, Carniti C, Miranda C, Roccato E and Pierotti MA . (2003). J. Cell. Physiol., 195, 168–186.

  4. Andreozzi F, Melillo RM, Carlomagno F, Oriente F, Miele C, Fiory F, Santopietro S, Castellone MD, Beguinot F, Santoro M and Formisano P . (2003). Oncogene, 22, 2942–2949.

  5. Arighi E, Alberti L, Torriti F, Ghizzoni S, Rizzetti MG, Pelicci G, Pasini B, Bongarzone I, Piutti C, Pierotti MA and Borrello MG . (1997). Oncogene, 14, 773–782.

  6. Asai N, Murakami H, Iwashita T and Takahashi M . (1996). J. Biol. Chem., 271, 17644–17649.

  7. Baloh RH, Tansey MG, Golden JP, Creedon DJ, Heuckeroth RO, Keck CL, Zimonjic DB, Popescu NC, Johnson Jr EM and Milbrandt J . (1997). Neuron, 18, 793–802.

  8. Baloh RH, Enomoto H, Johnson Jr EM and Milbrandt J . (2000). Curr. Opin. Neurobiol., 10, 103–110.

  9. Besset V, Scott RP and Ibanez CF . (2000). J. Biol. Chem., 275, 39159–39166.

  10. Borrello MG, Alberti L, Arighi E, Bongarzone I, Battistini C, Bardelli A, Pasini B, Piutti C, Rizzetti MG, Mondellini P, Radice MT and Pierotti MA . (1996). Mol. Cell. Biol., 16, 2151–2163.

  11. Borrello MG, Mercalli E, Perego C, Degl'Innocenti D., Ghizzoni S, Arighi E, Eroini B, Rizzetti MG and Pierotti MA . (2002). Biochem. Biophys. Res. Commun., 296, 515–522.

  12. Borrello MG, Smith DP, Pasini B, Bongarzone I, Greco A, Lorenzo MJ, Arighi E, Miranda C, Eng C, Alberti L, Bocciardi R, Mondellini P, Scopsi L, Romeo G, Ponder BAJ and Pierotti MA . (1995). Oncogene, 11, 2419–2427.

  13. Carter MT, Yome JL, Marcil MN, Martin CA, Vanhorne JB and Mulligan LM . (2001). Cytogenet. Cell. Genet., 95, 169–176.

  14. Coulpier M, Anders J and Ibanez CF . (2002). J. Biol. Chem., 277, 1991–1999.

  15. de Graaff E, Srinivas S, Kilkenny C, D'Agati V, Mankoo BS, Costantini F and Pachnis V . (2001). Genes Dev., 15, 2433–2444.

  16. De Vita G, Melillo RM, Carlomagno F, Visconti R, Castellone MD, Bellacosa A, Billaud M, Fusco A, Tsichlis PN and Santoro M . (2000). Cancer Res., 60, 3727–3731.

  17. Durbec PL, Larsson-Blomberg LB, Schuchardt A, Costantini F and Pachnis V . (1996). Development, 122, 349–358.

  18. Durick K, Yao VJ, Borrello MG, Bongarzone I, Pierotti MA and Taylor SS . (1995). J. Biol. Chem., 270, 24642–24645.

  19. Fukuda T, Kiuchi K and Takahashi M . (2002). J. Biol. Chem., 277, 19114–19121.

  20. Grimm J, Sachs M, Britsch S, Di Cesare S, Schwarz-Romond T, Alitalo K and Birchmeier W . (2001). J. Cell Biol., 154, 345–354.

  21. Hayashi H, Ichihara M, Iwashita T, Murakami H, Shimono Y, Kawai K, Kurokawa K, Murakumo Y, Imai T, Funahashi H, Nakao A and Takahashi M . (2000). Oncogene, 19, 4469–4475.

  22. Hayashi Y, Iwato M, Arakawa Y, Fujisawa H, Thoma Y, Hasegawa M, Tachibana O and Yamashita J . (2001). J. Neurooncol., 55, 51–58.

  23. Hennige AM, Lammers R, Arlt D, Hoppner W, Strack V, Niederfellner G, Seif FJ, Haring HU and Kellerer M . (2000). Mol. Cell. Endocrinol., 167, 69–76.

  24. Ishizaka Y, Ochiai M, Tahira T, Sugimura T and Nagao M . (1989). Oncogene, 4, 789–794.

  25. Iwahashi N, Murakami H, Nimura Y and Takahashi M . (2002). Biochem. Biophys. Res. Commun., 294, 642–649.

  26. Iwashita T, Kurokawa K, Qiao S, Murakami H, Asai N, Kawai K, Hashimoto M, Watanabe T, Ichihara M and Takahashi M . (2001). Gastroenterology, 121, 24–33.

  27. Jhiang SM . (2000). Oncogene, 19, 5590–5597.

  28. Jing S, Yu Y, Fang M, Hu Z, Holst PL, Boone T, Delaney J, Schultz H, Zhou R and Fox GM . (1997). J. Biol. Chem., 272, 33111–33117.

  29. Kawamoto Y, Takeda K, Okuno Y, Yamakawa Y, Ito Y, Taguchi R, Kato M, Suzuki H, Takahashi M and Nakashima I . (2004). J. Biol. Chem. (Jan 6 Epub ahead of print).

  30. Kurokawa K, Iwashita T, Murakami H, Hayashi H, Kawai K and Takahashi M . (2001). Oncogene, 20, 1929–1938.

  31. Lanzi C, Cassinelli G, Pensa T, Cassinis M, Gambetta RA, Borrello MG, Menta E, Pierotti MA and Zunino F . (2000). Int. J. Cancer, 85, 384–390.

  32. Lee DC, Chan KW and Chan SY . (2002). Oncogene, 21, 5582–5592.

  33. Liu X, Vega QC, Decker RA, Pandey A, Worby CA and Dixon JE . (1996). J. Biol. Chem., 271, 5309–5312.

  34. Lorenzo MJ, Gish GD, Houghton C, Stonehouse TJ, Pawson T, Ponder BAJ and Smith DP . (1997). Oncogene, 14, 763–771.

  35. Melillo RM, Carlomagno F, De Vita G, Formisano P, Vecchio G, Fusco A, Billaud M and Santoro M . (2001). Oncogene, 20, 209–218.

  36. Mercalli E, Ghizzoni S, Arighi E, Alberti L, Sangregorio R, Radice MT, Gishizky ML, Pierotti MA and Borrello MG . (2001). Oncogene, 20, 3475–3485.

  37. Mograbi B, Bocciardi R, Bourget I, Busca R, Rochet N, Farahi-Far D, Juhel T and Rossi B . (2001a). J. Biol. Chem., 276, 45307–45319.

  38. Mograbi B, Bocciardi R, Bourget I, Juhel T, Farahi-Far D, Romeo G, Ceccherini I and Rossi B . (2001b). Mol. Cell. Biol., 21, 6719–6730.

  39. Pachnis V, Mankoo B and Costantini F . (1993). Development, 119, 1005–1017.

  40. Pandey A, Duan H, Di Fiore PP and Dixit VM . (1995). J. Biol. Chem., 270, 21461–21463.

  41. Pandey A, Liu X, Dixon JE, Di Fiore PP and Dixit VM . (1996). J. Biol. Chem., 271, 10607–10610.

  42. Pasini B, Ceccherini I and Romeo G . (1996). Trends Genet., 12, 138–144.

  43. Pierotti MA, Bongarzone I, Borrello MG, Greco A, Pilotti S and Sozzi G . (1996). Genes Chromosomes Cancer, 16, 1–14.

  44. Sanicola M, Hession C, Worley D, Carmillo P, Ehrenfels C, Walus L, Robinson S, Jaworski G, Wei H, Tizard R, Whitty A, Pepinsky RB and Cate RL . (1997). Proc. Natl. Acad. Sci. USA, 94, 6238–6243.

  45. Schedl A and Hastie ND . (2000). Curr. Opin. Genet. Dev., 10, 543–549.

  46. Schuchardt A, D'Agati V, Larsson-Blomberg L, Costantini F and Pachnis V . (1994). Nature, 367, 380–383.

  47. Schuchardt A, D'Agati V, Pachnis V and Costantini F . (1996). Development, 122, 1919–1929.

  48. Segouffin-Cariou C and Billaud M . (2000). J. Biol. Chem., 275, 3568–3576.

  49. Srinivas S, Wu Z, Chen CM, D'Agati V and Costantini F . (1999). Development, 126, 1375–1386.

  50. Takahashi M, Buma Y, Iwamoto T, Inaguma Y, Ikeda H and Hiai H . (1988). Oncogene, 3, 571–578.

  51. Takahashi M, Ritz J and Cooper GM . (1985). Cell, 42, 581–588.

  52. Tanaka M, Gupta R and Mayer BJ . (1995). Mol. Cell. Biol., 15, 6829–6837.

  53. Tang MJ, Worley D, Sanicola M and Dressler GR . (1998). J. Cell Biol., 142, 1337–1345.

  54. Taraviras S and Pachnis V . (1999). Curr. Opin. Genet. Dev., 9, 321–327.

  55. Trupp M, Raynoschek C, Belluardo N and Ibanez CF . (1998). Mol. Cell. Neurosci., 11, 47–63.

  56. Tsui-Pierchala BA, Ahrens RC, Crowder RJ, Milbrandt J and Johnson Jr EM . (2002). J. Biol. Chem., 277, 34618–34625.

  57. Tsuzuki T, Takahashi M, Asai N, Iwashita T, Matsuyama M and Asai J . (1995). Oncogene, 10, 191–198.

  58. van Puijenbroek AA, van Weering DH, van den Brink CE, Bos JL, van der Saag PT, de Laat SW and den Hertog J . (1997). Oncogene, 14, 1147–1157.

  59. van Weering DH and Bos JL . (1997). J. Biol. Chem., 272, 249–254.

  60. van Weering DH, Medema JP, van Puijenbroek A, Burgering BM, Baas PD and Bos JL . (1995). Oncogene, 11, 2207–2214.

Download references

Acknowledgements

We gratefully acknowledge Mordechai Anafi and Bruce J Mayer for the GRB2 constructs, Darrin P Smith for RET-MEN2A mutant constructs and Cinzia Lanzi for RPI1 inhibitor. We thank Miss Maria Teresa Radice for excellent technical assistance and Miss Cristina Mazzadi for secretarial help. This study was supported by the Associazione Italiana per la Ricerca sul Cancro (AIRC) and Consiglio Nazionale Ricerche (CNR).

Author information

Correspondence to Marco A Pierotti or Maria Grazia Borrello.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Degl'Innocenti, D., Arighi, E., Popsueva, A. et al. Differential requirement of Tyr1062 multidocking site by RET isoforms to promote neural cell scattering and epithelial cell branching. Oncogene 23, 7297–7309 (2004). https://doi.org/10.1038/sj.onc.1207862

Download citation

Keywords

  • RET isoforms
  • Tyr1062
  • cell branching
  • cell scattering
  • cell transformation

Further reading