Direct Association between the Ret Receptor Tyrosine Kinase and the Src Homology 2-containing Adapter Protein Grb7*

Adapter proteins containing Src homology 2 (SH2) do- mains link transmembrane receptor protein-tyrosine ki-nases to downstream signal transducing molecules. A family of SH2 containing adapter proteins including Grb7 and Grb10 has been recently identified. We had previously shown that Grb10 associates with Ret via its SH2 domain in an activation-dependent manner (Pan- dey, A., Duan, H., Di Fiore, P. P., and Dixit, V. M. (1995) J. Biol. Chem. 270, 21461–21463). We now demonstrate that the related adapter molecule Grb7 also associates with Ret in vitro and in vivo , and that the binding of the SH2 domain of Grb7 to Ret is direct. This binding is dependent upon Ret autophosphorylation since Grb7 is incapa- ble of binding a kinase-defective mutant of Ret. Thus two members of the Grb family, Grb7 and Grb10, likely relay signals emanating from Ret to other, as yet, uni-dentified targets within the cell. The ret protooncogene encodes a transmembrane receptor protein-tyrosine kinase (RPTK). 1 The extracellular domain of Ret is unrelated to any other subfamily of RPTKs and contains multiple cadherin-like repeats and a cysteine-rich region (1). Ret plays a critical role in renal development as well as that of endocrine organs derived from the neural crest including the constructs were generated by cloning the SH2 domains of Grb7 (amino acids 434–535 of mouse Grb7) and Grb10 (amino acids 503–622 of mouse Grb10) into the pET15b vector, and COOH-terminal SH2 domain of PLC (cid:103) (amino acids 668–756 of bovine PLC (cid:103) ) into the pQE-30 vector. The His-tagged fusion proteins were purified using Ni 2 (cid:49) -NTA agarose beads (Qiagen) and eluted with imid-azole according to manufacturer’s instructions. GST Binding and Competition Assays— Direct binding of the SH2 domain of Grb7 to activated EGFR/Ret cells transfected with EGFR/Ret, and after 48 h, starved overnight in 1% BSA. Anti-EGFR immunoprecipitates from EGF-treated or untreated lysates were resolved by SDS-PAGE and trans- ferred onto nitrocellulose. The nitrocellulose membrane was then incubated with 5 (cid:109) g/ml of SH2 Grb7 GST in TBS-Tween containing 1% BSA for 1 h at room temperature. After extensive washing, the membrane was immunoblotted with anti-GST monoclonal antibody to visualize the bound SH2 Grb7 GST. For in vitro competition assays, 293 cells were transfected with EGFR/Ret, the EGFR/ ret receptor immunoprecipitated using anti-EGFR antibody (Upstate Biotechnology, Inc.), and the immunoprecipi- tates phosphorylated in vitro in the presence of 45 (cid:109) M ATP. The samples were then divided into equal parts, and preincubated with the indicated amounts of soluble His-tagged SH2 domains for 30 min at room temperature. The samples were then incubated with 5 (cid:109) g of soluble GST fusion of SH2 domains for 1 h at room temperature followed by washing and SDS-PAGE. The gels were transferred onto nitrocellulose, and immunoblotted with an anti-GST monoclonal antibody (Santa Cruz Biotechnology, Inc., CA) to detect the bound SH2 domain GST fusion protein. Coimmunoprecipitation Assays— For in vivo binding studies, 293 cells cotransfected EGFR/Ret and either Grb7 or CD40bp Grb7 BSA 100 ng/ml EGF lysed lysis cell lysates

The ret protooncogene encodes a transmembrane receptor protein-tyrosine kinase (RPTK). 1 The extracellular domain of Ret is unrelated to any other subfamily of RPTKs and contains multiple cadherin-like repeats and a cysteine-rich region (1). Ret plays a critical role in renal development as well as that of endocrine organs derived from the neural crest including the adrenal medulla and the thyroid gland. Mutations in the ret protooncogene have been shown to result in a variety of disorders including Hirschsprung's disease, multiple endocrine neoplasias 2A and 2B, and familial medullary thyroid carcinoma (2)(3)(4)(5)(6)(7)(8). More recently, ret has been shown to be a dominant acting oncogene capable of germline transmission. Knockout studies have demonstrated that mice lacking ret have renal agenesis or severe dysgenesis and lack enteric neurons (9). Despite the fact that the ligand for Ret has yet to be identified, several signaling molecules have been identified that interact with Ret including PLC␥, rasGAP, Shc, Grb2, and paxillin (10 -13).
In an effort to identify additional signaling intermediates capable of transducing signals originating from Ret, we had previously used the cytoplasmic domain of Ret as bait in a yeast two-hybrid screen. This resulted in the identification of Grb10 as a molecule that specifically bound the activated form of Ret (14). Grb10 is a recently described SH2 domain containing adapter protein that belongs to an emerging family of adapter proteins. This family includes the previously described protein Grb7 and is characterized by a carboxyl-terminal SH2 domain, a central domain, and a proline-rich region (15)(16)(17). The central domain (also termed GM domain), is similar to the Caenorhabditis elegans gene mig-10 that is crucial for the migration of a subset of neuronal cells and contains a pleckstrin homology domain. The central domains of Grb7 and Grb10 are 54% identical to each other (16). Since Grb10 interacted with the Ret receptor protein-tyrosine kinase in an activation dependent manner, we asked if Grb7 might also associate with Ret in a similar fashion. This reasoning was based on the fact that since the SH2 domains of Grb7 and Grb10 are 64% identical to each other, it was possible that they engaged similar receptor protein-tyrosine kinases. Grb7 has previously been shown to associate with HER2/neu, a receptor closely related to the epidermal growth factor receptor (EGFR) (18). Activation of the HER2 receptor led to association and tyrosine phosphorylation of Grb7 in cells overexpressing the chimeric EGFR/HER2 receptor (18). In this study we demonstrate that Grb7 directly associates with Ret via its SH2 domain, and unlike Grb10, it undergoes tyrosine phosphorylated in response to Ret activation.

MATERIALS AND METHODS
Expression Vectors-Construction of hemagglutinin tagged Ret cytoplasmic domain (HA-Ret) plasmid has been described previously (14). Construction and characterization of the EGFR/Ret construct has also been described earlier (13). For these studies, EGFR/Ret was subcloned into the mammalian expression vector, pZeoSV (Invitrogen) where its expression was driven by the SV40 promoter. The Grb7 expression construct was a gift from Dr. Ben Margolis (University of Michigan). The expression vector for Flag-epitope tagged CD40 binding protein (CD40bp-Flag) has been described earlier (19). Construction of hemagglutinin (HA)-tagged Ret cDNA and the kinase defective mutant encoding methionine instead of lysine at position 758 (K758M) have been described elsewhere (20).
Production of Fusion Proteins-The SH2 domain of Grb7 expressed as a GST fusion was a gift from Dr. Ben Margolis. Construction of GST fusions expressing the NH 2 -terminal and COOH-terminal domains of PLC␥ have been described earlier (21). GST fusion proteins were prepared using standard procedures and the recombinant proteins immobilized onto glutathione-Agarose beads (Sigma). Soluble GST fusions of SH2 Grb7 and Grb10 were generated by eluting the beads with 10 mM reduced glutathione followed by dialyzing the eluate against phosphatebuffered saline. His-tagged constructs were generated by cloning the SH2 domains of Grb7 (amino acids 434 -535 of mouse Grb7) and Grb10 (amino acids 503-622 of mouse Grb10) into the pET15b vector, and COOH-terminal SH2 domain of PLC␥ (amino acids 668 -756 of bovine PLC␥) into the pQE-30 vector. The His-tagged fusion proteins were purified using Ni 2ϩ -NTA agarose beads (Qiagen) and eluted with imidazole according to manufacturer's instructions. ʈ Postdoctoral fellow of the Cancer Foundation of America. 1 The abbreviations used are: RPTK, receptor protein tyrosine kinase; GST, glutathione S-transferase; SH, Src homology; EGF, epidermal growth factor; EGFR, epidermal growth factor receptor; HA, hemagglutinin; PLC␥, phospholipase C␥; PAGE, polyacrylamide gel electrophoresis; BSA, bovine serum albumin.
GST Binding and Competition Assays-Direct binding of the SH2 domain of Grb7 to activated EGFR/Ret was evaluated as follows. First, 293 cells were transfected with EGFR/Ret, and after 48 h, starved overnight in 1% BSA. Anti-EGFR immunoprecipitates from EGFtreated or untreated lysates were resolved by SDS-PAGE and transferred onto nitrocellulose. The nitrocellulose membrane was then incubated with 5 g/ml of SH2 Grb7 GST in TBS-Tween containing 1% BSA for 1 h at room temperature. After extensive washing, the membrane was immunoblotted with anti-GST monoclonal antibody to visualize the bound SH2 Grb7 GST.
For in vitro competition assays, 293 cells were transfected with EGFR/Ret, the EGFR/ret receptor immunoprecipitated using anti-EGFR antibody (Upstate Biotechnology, Inc.), and the immunoprecipitates phosphorylated in vitro in the presence of 45 M ATP. The samples were then divided into equal parts, and preincubated with the indicated amounts of soluble His-tagged SH2 domains for 30 min at room temperature. The samples were then incubated with 5 g of soluble GST fusion of SH2 domains for 1 h at room temperature followed by washing and SDS-PAGE. The gels were transferred onto nitrocellulose, and immunoblotted with an anti-GST monoclonal antibody (Santa Cruz Biotechnology, Inc., CA) to detect the bound SH2 domain GST fusion protein.
Coimmunoprecipitation Assays-For in vivo binding studies, 293 cells were cotransfected with EGFR/Ret and either Grb7 or CD40bp expression constructs, or Grb7 alone, starved in 1% BSA overnight, and either not treated or treated with 100 ng/ml EGF for 5 min and lysed in lysis buffer. Cleared cell lysates were incubated with 5 g of anti-Grb7 (N-20) (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) or anti-Flag monoclonal antibody for 2 h at 4°C, followed by incubation with protein A/G, and washed and bound material was eluted by boiling in sample buffer. The eluates were subjected to Western blot analysis as described below.
For in vivo binding studies using SKBR-3 cells, these cells were transiently transfected with EGFR/Ret using the calcium phosphate method. 48 h after transfection, cells were starved in 1% BSA overnight and left untreated or treated with EGF followed by lysis as described above. The lysates were immunoprecipitated with anti-Grb7 (N-20) (Santa Cruz Biotechnology, Inc.) followed by Western blot analysis with anti-phosphotyrosine as described below.
Western Blot Analysis-Precipitated immune complexes were resolved by SDS-PAGE, transferred to nitrocellulose by electroblotting, blocked with 1% BSA in Tris-buffered saline containing 0.1% Tween (TBS-T) overnight at 4°C and then incubated with 4G10 anti-phosphotyrosine monoclonal antibody (Upstate Biotechnology, Inc), followed by incubation with horseradish peroxidase-conjugated goat anti-mouse or IgG (Bio-Rad). After extensive washing, membranes were developed using a chemiluminescent reaction (Boehringer Mannheim) according to the manufacturer's instructions. Detection of EGFR/Ret protein was done by immunoprecipitating parallel samples with anti-human EGFR monoclonal antibody (Upstate Biotechnology, Inc.) followed by immunoblotting with the same antibody.

RESULTS AND DISCUSSION
Given the importance of Ret in mammalian development and oncogenesis, characterization of molecules that participate in its signaling is of paramount importance. Since the ligand for Ret is unidentified, a strategy was utilized to exploit the finding that the cytoplasmic domain of Ret mimics the activated form of the receptor as it possesses constitutive tyrosine kinase activity that results in autophosphorylation. The phosphorylated tyrosines in turn, should be capable of engaging downstream signaling molecules. Using the Ret cytoplasmic domain fused to the GAL4 DNA binding domain as bait in the yeast two-hybrid screen, we had previously identified Grb10, an SH2-containing adapter protein, as a downstream target of Ret (14). Grb10 belongs to an emerging family of SH2 containing adapter proteins with Grb7 as the prototypical member (15)(16)(17). Both of these proteins contain a single SH2 domain at the COOH terminus, a central region containing a pleckstrin homology domain, and a proline-rich region at the NH 2 terminus.
Grb7 was initially identified by screening a gt11 expression library with a phosphorylated 32 P-labeled carboxyl-terminal segment of EGFR (17). Despite this, it has not been shown to bind to EGFR in vivo. It does, however, associate with HER2/ neu, a receptor closely related to the EGF receptor (18). As a result of this binding, Grb7 is tyrosine-phosphorylated which may be sufficient to initiate further downstream signaling. Since the SH2 domain of Grb7 is 64% identical to Grb10, we asked whether Grb7 could associate with Ret in a manner similar to Grb10. To directly ascertain if the SH2 domain of Grb7 could bind the cytoplasmic domain of Ret, human embryonic kidney 293 cells were transfected with an expression vector encoding hemagglutinin-tagged Ret cytoplasmic domain (HA-Ret). The lysates were incubated with Grb7 SH2-GST, NH 2 -and COOH-terminal SH2 domains of PLC␥, or GST alone. As shown in Fig. 1, the SH2 domain of Grb7 bound the Ret cytoplasmic domain as did the COOH-terminal SH2 domain of PLC␥. The COOH-terminal SH2 domain of PLC␥ has been previously shown to bind Ret, and served as an important positive control. This interaction was specific since the NH 2terminal SH2 domain of PLC␥ or GST by itself did not bind.
To confirm that the binding was due to the ability of the Ret kinase to autophosphorylate itself, a kinase-defective mutant of Ret (K758M) was generated by replacing lysine in the ATP binding pocket of the kinase domain with methionine (20). 293 cells were transfected either with wild type or kinase-defective Ret receptor, and the receptors were immunoprecipitated, phosphorylated in vitro, and incubated with a soluble GST fusion protein of the SH2 domain of Grb7. Although Grb7 efficiently bound the wild type receptor, it showed no detectable binding to the kinase defective mutant of Ret (data not shown).
Since the ligand for Ret is not known, it is difficult to activate the native receptor in vivo and determine if the tyrosine-phosphorylated cytoplasmic domain has engaged Grb7. To circumvent this problem, we have used an EGFR/Ret chimera in which the extracellular domain of EGFR is fused to transmembrane and cytoplasmic regions derived from Ret. When such a chimeric receptor is tested, the tyrosine kinase activity of Ret is increased substantially upon addition of EGF. 293 cells were either transfected with Grb7 alone, EGFR/Ret and Grb7, or EGFR/Ret and an unrelated gene product CD40bp-Flag to serve as a control. Cells were then either left untreated or treated with EGF, and anti-Grb7 or anti-Flag immunoprecipitates were subjected to immunoblotting with anti-phosphotyrosine antibody to detect coprecipitating EGFR/Ret. As shown in Fig. 2A, Grb7 associated with EGFR/Ret in an activationdependent manner, and as expected, the negative control, CD40bp, did not show any detectable binding. The bottom panel shows equal expression of EGFR/Ret in the various samples. Fig. 2B shows that Grb7 underwent an increase in tyrosine phosphorylation upon addition of EGF. This is analogous to the situation with cells expressing the EGFR/HER2 receptor chimera where stimulation of HER2 by EGF leads to increased tyrosine phosphorylation of Grb7. This suggests that Grb7, besides binding receptor kinases, also serves as a substrate being phosphorylated on tyrosine residues. In order to eliminate the possibility that the observed association of Grb7 with Ret was due to overexpression of Grb7, the SKBR-3 cell line FIG. 1. SH2 domain of Grb7 specifically associates with Ret in vitro. 293 cells were transfected with an expression vector containing the cytoplasmic domain of Ret that was HA epitope-tagged (HA-Ret). Cells were then lysed and the lysates incubated with the indicated GST fusion proteins. Bound material was dissociated by boiling in sample buffer, subjected to SDS electrophoresis, transferred onto nitrocellulose, and immunoblotted with anti-HA antibody.
which expresses endogenous Grb7 was used for analysis. This cell line was transiently transfected with the EGFR/Ret plasmid, and left untreated or treated with EGF. Anti-Grb7 immunoprecipitates were then immunoblotted with an anti-phosphotyrosine antibody to detect co-precipitated EGFR/Ret. As shown in Fig. 2C, the endogenous Grb7 again associated with the activated chimeric Ret receptor but not with the unactivated receptor. There was equal expression of EGFR/Ret in the unactivated and activated samples (data not shown).
The SH2 domain of Grb7 could either bind the autophosphorylated Ret receptor directly, or this binding could be mediated by adapter proteins such as Shc which are phosphorylated by Ret. In order to distinguish between these two possibilities, a direct binding assay using a soluble GST fusion of the SH2 domain of Grb7 was carried out. EGFR/Ret was immunoprecipitated from transfected cells that were either untreated or treated with EGF and the immunoprecipitates were resolved by SDS-PAGE followed by transfer onto a nitrocellulose membrane. This membrane was then incubated with soluble SH2 GST Grb7 followed by conventional immunoblotting with an anti-GST monoclonal antibody to detect the GST fusion bound to immunoprecipitated EGFR/Ret receptor. As shown in Untreated or EGF-treated cells were lysed, and the lysates were immunoprecipitated with anti-Grb7 antibody followed by anti-phosphotyrosine antibody to detect coprecipitating EGFR/Ret. tagged fusion protein of the corresponding SH2 domain. As shown in Fig. 3B, a 50-fold molar excess of competing Grb10 SH2 domain was able to completely abrogate the binding of Grb7 to Ret suggesting that they bind to the same site. However, a 10-fold excess of His-tagged Grb10 was not as effective as a 2-fold excess of His-tagged Grb7 which totally abolished binding to Ret implying a higher affinity of the Grb7 SH2 domain for Ret. Importantly, a 20-fold excess of His-tagged COOH-terminal SH2 PLC␥ had no effect on binding of Grb7 suggesting that the observed inhibition was not due to nonspecific binding. Additionally, when the experiment was repeated in reverse, a 10-fold molar excess of Grb7 SH2 domain was sufficient to completely abolish binding of Grb10 to Ret (Fig.  3C).
In summary, we have shown that Grb7 associates with the Ret receptor tyrosine kinase through its SH2 domain and that activation of Ret results in tyrosine phosphorylation of Grb7. Given the importance of Ret during development and in carcinogenesis, signaling molecules downstream of Ret are likely to be important as they may mediate some of these actions of Ret. Clearly, determining the binding sites of Grb7 and Grb10 will be important to further dissection of the signaling pathways initiated by Ret. However, it should be noted that it has been exceedingly difficult to map the Grb7 binding sites on the HER2/neu receptor (18). Nevertheless, studies are ongoing in our laboratories to delineate the site(s) on the Ret receptor protein-tyrosine kinase responsible for binding to these adapter molecules. Further, it should be noted that though Grb7 is tyrosine phosphorylated in response to Ret activation, we have failed to see a similar phosphorylation in the case of Grb10. 2 However, the significance of this phosphorylation is not clear at the present time. One explanation may be that this difference in tyrosine phosphorylation may serve to recruit distinct downstream effectors. Identification of these downstream effectors will be important in dissecting Ret signaling pathways.