A mutation at tyrosine 1062 in MEN2A-Ret and MEN2B-Ret impairs their transforming activity and association with shc adaptor proteins.

Germ line mutations of the ret proto-oncogene are associated with the development of three dominantly inherited neoplastic disorders, multiple endocrine neoplasia (MEN) 2A, MEN 2B, and familial medullary thyroid carcinoma. It has been demonstrated that the mutations result in constitutive activation of the Ret protein, leading to transformation of NIH 3T3 cells. In the present study we investigated the role of tyrosine residues present in the carboxyl-terminal sequence for the transforming activity of Ret with the MEN 2A or MEN 2B mutation (MEN2A-Ret or MEN2B-Ret). Substitution of phenylalanine for tyrosine 1062 (designated Y1062F) markedly impaired the transforming activity of both MEN2A-Ret and MEN2B-Ret, whereas substitution or deletion for four other tyrosines (codons 981, 1015, 1090, and 1096) did not affect their activity. The Shc adaptor proteins bound to the MEN2A-Ret and MEN2B-Ret proteins and were phosphorylated on tyrosine in the transfectants. The binding of Shc to the Y1062F mutant proteins was reduced by approximately 80%, indicating that tyrosine 1062 is a major binding site for Shc. In addition, phosphopeptide analysis of MEN2A-Ret demonstrated that tyrosine 1062 represents an autophosphorylation site of the mutant Ret proteins.

The ret proto-oncogene encodes a receptor tyrosine kinase whose ligand has not been identified (1)(2)(3). It turned out that germ line mutations of ret are responsible for the development of four different neural crest disorders including multiple endocrine neoplasia (MEN) 1 2A, MEN 2B, familial medullary thyroid carcinoma (FMTC), and Hirschsprung's disease (4 -9). MEN 2A, MEN 2B, and FMTC are dominantly inherited neoplastic disorders, the former two of share the clinical feature of medullary thyroid carcinoma and pheochromocytoma. FMTC is characterized by the development of medullary thyroid carcinoma alone. Hirschsprung's disease represents a congenital disorder associated with the absence of intrinsic ganglion cells in the distal gastrointestinal tract. Recent studies demon-strated that MEN 2A, MEN 2B, and FMTC mutations represent gain-of-function mutations (10 -14) whereas Hirschsprung mutations result in inactivation of the Ret protein (15). MEN 2A mutations that involve cysteine residues present in the extracellular domain induce ligand-independent dimerization of the Ret protein, leading to its constitutive activation (10 -12). In contrast, the MEN 2B mutation detected in the kinase domain of Ret appears to activate the Ret protein without dimerization, probably due to a conformational change of its catalytic core region (12,14).
In addition, we found that tyrosine residues in the kinase domain essential for transforming activity are different between Ret with the MEN 2A mutation (MEN2A-Ret) and Ret with the MEN 2B mutation (MEN2B-Ret) (14). Substitution of phenylalanine for tyrosine 905 completely abolished the transforming activity of MEN2A-Ret but not MEN2B-Ret. This tyrosine corresponds to tyrosine 416 of the Src protein and is known to be conserved in all tyrosine kinases and play crucial roles in their catalytic and/or biological activities (16 -22). Since the activation of MEN2A-Ret could mimic that of other receptor tyrosine kinases caused by the ligand-dependent dimerization, phosphorylation of tyrosine 905 may be important for the activity of MEN2A-Ret. On the other hand, tyrosines 864 and 952, instead of tyrosine 905, were required for the activity of MEN2B-Ret but not MEN2A-Ret, supporting the view that the MEN 2B mutation induces a conformational change of the Ret kinase domain. Tyrosine 905 and tyrosines 864 and 952 appeared to regulate the tyrosine kinase activity of MEN2A-Ret and MEN2B-Ret, respectively (14).
To date, the intracellular signaling pathways via the Ret protein have not been well characterized, because a ligand for Ret is still unknown. Tyrosine residues present in the carboxylterminal sequence of receptor tyrosine kinases are known to be the sites recognized by various signaling molecules containing Src homology 2 (SH2) domains such as phospholipase C type ␥, Ras GTPase-activating protein, Grb2, and Shc (23)(24)(25). In the present study, to characterize the signaling pathways via the mutant Ret protein, we mutated tyrosines present mostly in the carboxyl-terminal sequence of Ret. Among five tyrosines examined (tyrosines 981, 1015, 1062, 1090, and 1096), replacement of tyrosine 1062 with phenylalanine drastically decreased the transforming activity of both MEN2A-Ret and MEN2B-Ret. In addition, we found that tyrosine 1062 is a major binding site for the Shc adaptor proteins that are phosphorylated on tyrosine in the transfectants.

EXPERIMENTAL PROCEDURES
Plasmid Construction-A cDNA clone containing the entire coding sequence (for 1072 or 1114 amino acids) of the human c-ret gene was inserted between HindIII and EcoRI sites of the Rc/CMV plasmid (Invitrogen, San Diego, CA). Each mutation was introduced by polymerase chain reaction. In brief, primers containing the mutations were * This work was supported in part by grants-in-aid for scientific research and for cancer research from the Ministry of Education, Science, and Culture of Japan and by grants from Otsuka Pharmaceutical Co., Ltd., and the Ichiro Kanehara Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Transfection-Each recombinant plasmid (0.05-0.2 g) was transfected into NIH 3T3 cells (5 ϫ 10 5 cells in a 60 mm-diameter dish) with 10 g of NIH 3T3 DNA as described previously (1). Transformed foci were scored on day 12 after transfection. Then foci were picked up and grown into cell lines.
Western Blotting-Total cell lysates were prepared from NIH 3T3 cells and transfectants as described previously (26). The lysates were subjected to sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride membranes (Nihon Millipore Kogyo KK, Tokyo, Japan). After membranes were reacted with anti-Ret antibody (26,27), anti-Shc antibody (Upstate Biotechnology Inc., Lake Placid, NY), or anti-phosphotyrosine antibody (Zymed Laboratories, Inc., South San Francisco, CA), the reaction was examined by the avidin-biotin complex immunoperoxidase method (26) or 125 I-protein A (ICN, Irvine, CA). Expression and Phosphorylation of Fusion Proteins-A cDNA fragment with or without the Y1062F mutation comprising nucleotides 3075-3714 (numbered according to the published sequence (2)) was inserted between StuI and PstI sites of pMAL-c expression vector (New England BioLabs, Beverly, MA). The recombinant plasmids were transformed into Epicurian Coli TKX1 competent cells (Stratagene, La Jolla, CA) that contain a plasmid-encoded, inducible tyrosine kinase gene. To express the fusion proteins, transformed cells were grown in 2 ϫ YTG broth (16 g of tryptone/liter, 10 g of NaCl/liter, and 10 g of yeast extract/liter) containing 2% (w/v) glucose, 50 g/ml ampicillin and 12.5 g/ml tetracycline, and the induction was performed with 0.3 mM isopropyl-␤-D-thiogalactopyranoside when the cultures reached A 600 ϭ 1-2. The cells were spun down at 2,000 ϫ g and resuspended in TK induction media (see manufacturer's (Stratagene) protocol) to an A 600 of 0.5. They were then grown for 2 h at 37°C to phosphorylate the fusion proteins. Phosphorylation of the fusion proteins was evaluated by Western blotting with anti-phosphotyrosine antibody.
Cell Labeling and Immunoprecipitation-Cells were labeled for 4 -5 h in phosphate-free RPMI medium containing [ 32 P]orthophosphate (1 mCi/ml; ICN) supplemented with 10% dialyzed fetal calf serum. After washing with phosphate-buffered saline, pH 7.2, cells were lysed in RIPA buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 2 mM EDTA, 1% Triton X-100) containing 1 mM phenylmethylsulfonyl fluoride and 0.5 mM sodium orthovanadate. The lysates were clarified by centrifugation (15,000 ϫ g) for 1 h, incubated with Sepharose beads conjugated with antibodies at 4°C overnight, and washed with RIPA buffer four times. The resulting antigen-antibody complex was suspended in SDS-sample buffer in the presence of 80 mM dithiothreitol and boiled for 3 min.
A peptide comprising amino acids 1058 -1078 of the long isoform of Ret (Fig. 4A) was synthesized. Five g of peptide was mixed with the product of the Ret intracellular domain (amino acids 682-1114) with the MEN 2B mutation expressed in baculovirus and incubated with [␥-32 P]ATP (10 Ci, 6000 Ci/mmol, Amersham, United Kingdom) in the kinase buffer (20 mM Tris-HCl, pH 7.5, 10 mM MnCl 2 , and 0.01% Triton X-100) for 20 min at 30°C. Phosphorylated peptide was separated on Tricine/SDS-16.5% polyacrylamide gels, and the peptide band was excised, digested with CNBr, and lyophilized as described above. Then peptides were digested with trypsin and separated in two dimensions.

RESULTS
Transforming Activity of Ret with Mutations of Tyrosine Residues-We introduced the MEN 2A (Cys 634 -Arg, C634R) or MEN 2B (Met 918 -Thr, M918T) mutation into two Ret isoforms of 1114 amino acids (long isoform) and 1072 amino acids (short isoform) (Fig. 1) and investigated their transforming activity. The 51 carboxyl-terminal amino acids of the long isoform were replaced by the 9 unrelated amino acids of the short isoform by alternative splicing in the 3Ј region (30). The former sequence contains an additional two tyrosines (tyrosines 1090 and 1096) as compared with the latter sequence (Fig. 1). The mutant ret cDNAs were inserted into the expression vector containing the cytomegalovirus promoter and transfected into NIH 3T3 cells. As shown in Table I, the transforming activity of the long isoform with the MEN 2A mutation was 1.5-fold higher than that of its short isoform. On the other hand, the activity of the long isoform with the MEN 2B mutation was approximately 10-fold higher than that of the short isoform with the MEN 2B mutation and 2-3-fold higher than that of both isoforms with the MEN 2A mutation. Since the transfection efficiencies of each construct (5-6 ϫ 10 6 G418-resistant colonies/g of DNA) were comparable, these results suggested that the carboxylterminal sequences of two isoforms differently regulate the activity of the mutant Ret proteins.
To investigate whether the length of the carboxyl-terminal sequence influences the activity of Ret with the MEN 2A or MEN 2B mutation (MEN2A-Ret or MEN2B-Ret), we truncated the long isoform after codon 1074 (designated DEL-1074). This truncation removed tyrosines 1090 and 1096 present in the long isoform. However, the truncation did not significantly affect the transforming activity of MEN2A-Ret and MEN2B-Ret (Table I). Since the length of the carboxyl-terminal tail in the DEL-1074 mutant protein was comparable with that of the short isoform, the low transforming activity of the short isoform of MEN2B-Ret appears to be caused by its specific carboxyl-terminal sequence. In addition, this result indicated that tyrosines 1090 and 1096 do not play a crucial role for the transforming activity of the long isoform of MEN2A-Ret and MEN2B-Ret.
We next replaced tyrosines 1015 and 1062 in the carboxylterminal tail of the long isoform as well as tyrosine 981 in the kinase domain with phenylalanine ( Fig. 1; designated Y1015F, Y1062F, and Y981F, respectively). Although another tyrosine (codon 1029) was present in the carboxyl-terminal sequence, substitution for this tyrosine has been unsuccessful in our experiments. Among these, replacement of tyrosine 981 or tyrosine 1015 did not affect the transforming activity of MEN2A-Ret and MEN2B-Ret. In contrast, substitution for tyrosine 1062 severely impaired the activity of both of them (Table I), suggesting that tyrosine 1062 is one of major sites recognized by signaling molecules important for their transforming activity.
Establishment of the Cell Lines Expressing the Mutant Ret Proteins-In order to analyze the role of tyrosine residues in the intracellular signaling via the Ret protein, we established the cell lines expressing each mutant Ret protein at high levels ( Fig. 2A). As expected, molecular mass of the short isoform and the DEL-1074 mutant protein (150 and 170 kDa) was approximately 5 kDa smaller than that of the long isoform (155 and 175 kDa). The cell lines expressing MEN2A-Ret with the Y1062F mutation or MEN2B-Ret with the Y1062F mutation (designated NIHret(C634R,Y1062F)L and NIHret-(M918T,Y1062F)L cells) showed a partially transformed phenotype, whereas other cell lines expressing the Y981F, Y1015F, or DEL-1074 mutant proteins were spindle-shaped and highly refractile. The cells expressing the short isoform of MEN2B-Ret also showed a fully transformed phenotype. When these cells were injected subcutaneously into Scid mice, all of them formed solid tumors, although the latency of NIHret(C634R,Y1062F)L and NIHret(M918T,Y1062F)L cells was longer (15-21 days) than that of the other cell lines (6 -7 days) ( Table I). Fig. 2B shows Western blot analysis with anti-phosphotyrosine antibody. The patterns of tyrosine phosphorylation are similar among the cell lines expressing each isoform of MEN2A-Ret or MEN2B-Ret or expressing the Y981F, Y1015F, or DEL-1074 mutant proteins. As we have already reported (11,14), the level of tyrosine phosphorylation of the 170 -175-kDa Ret proteins present on the cell surface was higher than that of the 150 -155-kDa Ret proteins present in the endoplasmic reticulum. In addition, several other proteins including 74-, 58-, and 50-kDa proteins were phosphorylated on tyrosine at variable levels in each transfectant. On the other hand, the level of tyrosine phosphorylation somewhat decreased in NIH(C634R,Y1062F)L and NIH(M918T,Y1062F)L cells (Fig.  2B), although the Y1062F mutation did not influence the autokinase activity of MEN2A-Ret and MEN2B-Ret in vitro (data not shown).
Shc Adaptor Proteins Bind to Tyrosine 1062-Since Borrello et al. (31) reported that two forms of rearranged Ret (Ret/ptc1 and Ret/ptc2) found in human papillary thyroid carcinoma bound the Shc proteins in vivo, we investigated whether the MEN2A-Ret and MEN2B-Ret proteins also bind Shc. After the lysates from NIHret(C634R)L, NIHret(M918T)L, NIHret-(C634R,Y1062F)L, and NIHret(M918T,Y1062F)L cells were immunoprecipitated with anti-Ret antibody, they were immunoblotted with the anti-Ret or anti-Shc antibody (Fig. 3A). As a result, it turned out that the 52-and 46-kDa Shc proteins were coprecipitated with the MEN2A-Ret or MEN2B-Ret protein.
Interestingly, the degree of binding of Shc to MEN2A-Ret and MEN2B-Ret with the Y1062F mutation markedly decreased (ϳ80%). Since the Y981F, Y1015F and DEL-1074 mutant proteins did not show a significant change of the binding ability for Shc as compared with that of the MEN2A-Ret and MEN2B-Ret proteins (data not shown), the results suggested that tyrosine 1062 of Ret represents a major binding site for Shc and that its binding is associated with the transforming activity of the mutant Ret proteins.  To confirm that tyrosine 1062 is a binding site for Shc, we synthesized a fusion protein consisting of maltose-binding protein (MBP) and carboxyl-terminal 139 amino acids of the long isoform which include tyrosines 981, 1015, 1029, 1062, 1090, and 1096. A cDNA fragment with or without the Y1062F mutation comprising nucleotides 3075-3714 (numbered according to the published sequence (2)) was inserted into pMAL-c expression vector. In order to obtain tyrosine-phosphorylated fusion proteins, Epicurian Coli TKX1 competent cells that contain a plasmid-encoded, inducible tyrosine kinase gene were used for transformation, and the induced phosphorylated fusion proteins (approximately 57 kDa) were purified by mixing with the amylose resin (Fig. 3C, left panel). Tyrosine phosphorylation of the fusion proteins was evaluated by Western blotting with anti-phosphotyrosine antibody (Fig. 3C, right panel). Then the fusion proteins with or without the Y1062F mutation bound to the amylose resin were incubated with the lysate from NIH 3T3 cells. After the resin was extensively washed, protein complexes were subjected to Western blotting with anti-Shc antibody. As shown in Fig. 3D, the phosphorylated wild type fusion protein bound the Shc proteins, whereas the content of Shc binding to the fusion protein with the Y1062F mutation markedly decreased. These results confirmed that tyrosine 1062 is required for binding of Shc to the mutant Ret protein.
Tyrosine 1062 of Ret Is an Autophosphorylation Site in Vivo-In vivo phosphorylation of the MEN2A-Ret protein was analyzed by immunoprecipitating the lysates of 32 P-metabolically labeled NIHret(C634R)L and NIHret(C634R,Y1062F)L cells with anti-Ret antibody. After the 175-kDa Ret protein was eluted from the gel and digested with cyanogen bromide (CNBr), it was separated on Tricine/SDS-16.5% polyacrylamide gels (Fig. 4B, lanes 1 and 2). A 6-kDa phosphorylated band that was able to be detected by anti-Ret antibody against the 19 amino acids with tyrosine 1062 (27) was recovered from the gel (Fig. 4B, lane 3), digested with trypsin, and analyzed in two dimensions on thin layer plates (Fig. 4, C and D). According to this protocol, it was expected to obtain a fragment with tyrosine 1062 but not with other tyrosines (Fig. 4A).
As shown in Fig. 4, C and D, spot b was specifically identified in the peptide sample derived from NIHret(C634R) cells but not from NIHret(C634R,Y1062F) cells. To verify that this spot represents a fragment containing tyrosine 1062, we synthesized a peptide of 21 amino acids containing tyrosine 1062 (amino acids 1058 -1078 of the long isoform) (Fig. 4A). The synthetic peptide was then phosphorylated in vitro by the product of the Ret intracellular domain with the MEN 2B mutation expressed in baculovirus (Fig. 4B, lane 4), digested with CNBr and trypsin, and separated in two dimensions (Fig.  4E). As a result, a spot produced by digests of this synthetic peptide overlapped with spot b (Fig. 4, D-F), suggesting that spot b represents a peptide containing tyrosine 1062 that was autophosphorylated in vivo.

DISCUSSION
In the present study, we first compared the transforming activity of two Ret isoforms that differ in their carboxyl-terminal sequence. The carboxyl-terminal 9 amino acids of the short isoform were replaced by the 51 amino acids of the long isoform that contains two additional tyrosine residues (Tyr 1090 and Tyr 1096 ) (1,2,30). Although the difference in the transforming activity between both isoforms of MEN2A-Ret was small, the activity of the short isoform of MEN2B-Ret was approximately 10-fold lower than that of its long isoform and 3-4-fold lower than that of both isoforms of MEN2A-Ret. These results were in agreement with the results reported by Borrello et al. (13) demonstrating the low transforming activity of the short isoform of MEN2B-Ret. Since most, but not all, transfectants that we isolated showed relatively low levels of expression of the short isoform of MEN2B-Ret in comparison with the expression levels of its long isoform (data not shown), the former protein may be more unstable than the latter. On the other hand, it is possible that the short isoform of MEN2A-Ret might be stabilized by its dimerization, resulting in higher transforming activity than that of the short isoform of MEN2B-Ret.
To examine whether the carboxyl-terminal sequence of the long isoform with tyrosines 1090 and 1096 plays a role in the transforming activity, we truncated its 41 carboxyl-terminal amino acids (DEL-1074). However, this truncation did not decrease the transforming activity of MEN2A-Ret and MEN2B-Ret. This result suggested that the low transforming activity of the short isoform of MEN2B-Ret was caused by its specific carboxyl-terminal sequence rather than by its short length.
Since deletion of tyrosines 1090 and 1096 did not affect the activity of the long isoform of MEN2A-Ret and MEN2B-Ret, we changed three other tyrosines (tyrosines 981, 1015, and 1062) to phenylalanine. Among these, replacement of tyrosine 1062 severely impaired the transforming activity of MEN2A-Ret and MEN2B-Ret, indicating that tyrosine 1062 represents a major binding site for signaling molecules responsible for cell transformation. Consistent with this view, the degree of binding of the Shc adaptor proteins to the MEN2A-Ret and MEN2B-Ret proteins decreased to about 20% in the presence of the Y1062F mutation in vivo as well as in vitro. Simultaneously, phosphopeptide analysis demonstrated that tyrosine 1062 represents an autophosphorylation site of the MEN2A-Ret protein.
Thus, it seems likely that the Shc proteins recognize phosphorylated tyrosine 1062 and transmit the signal of MEN2A-Ret and MEN2B-Ret. In this respect, it has recently been reported that substitution of phenylalanine for tyrosine 1062 abolished the mitogenic activity of Ret/ptc2, a rearranged form of Ret detected in human papillary thyroid carcinoma (32). On the other hand, replacement of tyrosine 1029, the biological role of which was not examined in our experiments, had no significant effect on the mitogenic activity. Furthermore, Borrello et al. (31) observed that two forms of rearranged Ret (Ret/ptc1 and Ret/ptc2) bound the Shc proteins in the transfectants. These also supported our results that tyrosine 1062 is a binding site for Shc that may play a crucial role in the transforming activity of MEN2A-Ret and MEN2B-Ret. Using the established cell lines expressing each mutant protein, we are currently investigating the signaling pathway via the Shc adaptor proteins responsible for cell transformation.
Shc contains two domains, SH2 domain and phosphotyrosine binding domain (33), both of which are known to recognize specific phosphotyrosine-containing sequences (34,35). The binding specificity of Shc SH2 and phosphotyrosine binding domains is determined by residues carboxyl-terminal and amino-terminal to phosphotyrosine, respectively. The sequence (NKL) amino-terminal to tyrosine 1062 of Ret matches the consensus sequence (NXXpY) for the binding of the Shc phosphotyrosine binding domain (34), whereas tyrosine 1062 is not embedded in the consensus sequence (pY(I/E/Y)X(I/L/M)) for the binding of the Shc SH2 domain (35). This finding suggested that tyrosine 1062 is the target to which the phosphotyrosine binding domain binds rather than the SH2 domain. In addition, since the sequence (GMS) carboxyl-terminal to this tyrosine does not match the consensus sequences for the SH2 domains of other signaling molecules including phospholipase C type ␥, Grb2, phosphatidylinositol-3 kinase, Syp and Src family (23,35), it seems unlikely that tyrosine 1062 is a major binding site for these molecules.
The fact that the Y1062F mutation did not completely abolish the transforming activity of MEN2A-Ret and MEN2B-Ret and their binding ability to Shc suggested that tyrosine residues other than tyrosine 1062 represent minor binding sites for Shc which may be necessary for low levels of their transforming activity. To elucidate other binding sites for Shc, further investigation including the introduction of double or triple mutations of the tyrosines present in the intracellular domain of Ret will be required.