Grb10 Identified as a Potential Regulator of Growth Hormone (GH) Signaling by Cloning of GH Receptor Target Proteins*

The cloning of receptor targets procedure, used so far to identify proteins associated with tyrosine kinase receptors was modified to clone SH2 proteins able to bind to the growth hormone receptor (GHR). The cytoplasmic region of GHR, a member of the cytokine receptor superfamily does not contain tyrosine kinase activity. It was thus phosphorylated in bacteria by the Elk tyrosine kinase and radiolabeled to screen a mouse expression library. With this probe, we identified Shc and the p85 subunit of phosphatidylinositol 3-kinase as direct targets of the receptor. The other proteins identified, Csk, Shb, Grb4, and Grb10 are new potential transducers for cytokine receptors. We show in Huh-7 hepatoma cells that Grb10 and GHR associate under GH stimulation. Co-transfections in 293 cells further show that Grb10 interacts with both the GHR and Jak2. Functional tests demonstrate that Grb10 inhibits transcription of two reporter genes containing, respectively, the serum response element of c-fos and the GH response element 2 of the Spi2.1 gene, whereas it has no effect on a reporter gene containing only Stat5 binding elements. Our results suggest that Grb10 is a new target for a member of the cytokine receptor family that down-regulates some GH signaling pathways downstream of Jak2 and independently of Stat5.

The growth hormone receptor (GHR) 1 is involved in the regulation of expression of a broad range of genes important for growth, metabolism, and cellular differentiation. GHR signaling involves ligand-induced receptor homodimerization and activation of the tyrosine kinase Jak2 (1,2). These events initiate a cascade of phosphorylation of cellular proteins including the kinase itself and the receptor, thereby providing binding sites for SH2-containing molecules (2,3). Among these molecules, the Stat proteins (signal transducers and activators of transcription) have been shown to participate in GHR signaling (4 -7). Stat1, Stat3, and the two Stat5 isoforms are activated by GHR. Shc has also been reported to associate with the GHR⅐Jak2 complex (8). The GHR cytoplasmic domain contains two regions. The membrane proximal part containing the proline-rich region (Box 1) seems sufficient for GH-induced proliferation (9,10) and activation of mitogen-activated protein kinase (11). In contrast, the distal part that contains most of the tyrosines phosphorylated by Jak2 is required for transcriptional activation of specific genes, such as those of Spi2.1 (12) and insulin (13); these effects seem at least partially mediated through Stat5 activation (6,14). However, the molecular mechanisms involved in the GH regulation of other genes such as insulin-like growth factor-1 are still unknown (15,16).
In order to identify additional transducer molecules interacting with the activated GHR, we designed a new cloning strategy inspired from the cloning of receptor targets (CORT) procedure (17,18). The screening of the EXlox expression library with the phosphorylated and radiolabeled GHR COOH-terminal domain allowed us to isolate several clones that all are new GHR interacting molecules. Among these proteins, Grb10 is a recently identified adaptor that was reported to associate with insulin receptor (19) as well as with other tyrosine kinase receptors (20 -23). We show in the Huh-7 hepatoma cell line which expresses both Grb10 (24) and GHR (25,26), that GH stimulation induces the association of the receptor with Grb10. In order to better characterize the interactions between Grb10 and the activated GHR complex, as well as to test the potential role of this new adaptor protein in GHR signaling, we used 293 cells co-transfected with GHR and Grb10 cDNA. We show that Grb10 associates with the GHR⅐Jak2 complex through several binding sites and regulates specific pathways of GHR signaling.

EXPERIMENTAL PROCEDURES
Plasmids, Constructs, and Protein Purification-A cDNA fragment encoding the domain between amino acids (aa) 454 -620 of the rabbit GHR was inserted into the bacterial expression vector pGEX-2TK as described previously (7). We expressed the fusion protein in TKB1 bacteria (Stratagene) which allow phosphorylation of tyrosine residues, as described in the instruction manual, and also in MH1 bacteria. The expressed proteins phosphorylated or not, were purified on glutathione-Sepharose beads and radiolabeled by protein kinase A (Sigma) and [␥-32 P]ATP (ICN, 150 mCi/ml, 6000 Ci/mmol) (27). The probes (ϳ10 7 dpm/g) were eluted with 10 mM glutathione and filtered on Millipore filter (SJHVOO4NS) prior to use for library screening. NH 2 -terminal Flag-tagged molecules were constructed by inserting first the polymerase chain reaction product corresponding to the NH 2terminal region of the rabbit GHR into NotI/ClaI restriction sites of the pFlag-CMV-1 vector (IB-Kodak). Then, the COOH-terminal regions of GHR wt or GHR 454 were added by subcloning directly into ClaI/XbaI sites of the previous construction, the fragments coding for aa (350 3 620) for the GHR wt and (350 3 454) for the mutant GHR 454 .
LHRE, GHRE2 (7), and SRE reporter genes were constructed using a pUC18 vector containing the thymidine kinase minimal promoter linked to luciferase reporter gene (7). Doubled-stranded oligonucleotide 5Ј-CTAGAGGATGTCCATATTAGGACATCTGGATCTAG-3Ј for SRE was inserted into the XbaI and BamHI sites of the vector. * 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 1 The abbreviations used are: GHR, growth hormone receptor; aa, amino acid; hGH, human growth hormone; CORT, cloning of receptor targets; Grb, growth factor receptor bound protein, SH2, Src homology 2; GST, glutathione S-transferase; tk, thymidine kinase; Tyr(P), phosphotyrosine; C/EBP, cAMP-response element binding protein.
A Grb10 cDNA fragment encoding the carboxyl-terminal 165 aa was digested and inserted into the EcoRI site of the pGEX5X-3 vector (Pharmacia). The expressed GST-Grb10 SH2 and control GST proteins were purified on glutathione-Sepharose.
EXlox Library Screening-A EXlox library from a 16-day mouse embryo (Novagen) was plated at 40,000 phages per plate in Escherichia coli strain BL21(DE3)pLysE according to the manufacturer's instructions. After growth for 8 -12 h, plates were covered overnight with nitrocellulose impregnated with 1 mM isopropyl ␤-D-thiogalactoside. Filters were then blocked as described (27) and incubated for 12 h at 4°C with the GHR 454 -620 probe at 2 ϫ 10 6 dpm/ml. After purification, phages were converted to plasmids by using bacterial strain BM25.8 (Novagen). These plasmids were used to transform bacterial strain DH5␣ and inserts were sequenced.
Transcription Assays-293 cells were plated in 6-well plates (4 ϫ 10 5 cells/well) and were transfected with Grb10 cDNA (100 ng to 1 g), the reporter gene (100 ng of SRE-tk-Luc, 1.5 g of GHRE2 or LHRE-tk-Luc), the GHR cDNA (50 ng in tests with SRE-tk-Luc, 200 ng for the other tests) and 50 ng of cytomegalovirus ␤-galactosidase cDNA. In experiments with the SRE-tk-Luc, 20 ng of Jak2 cDNA were also transfected. Cells were incubated for 16 h in serum-free medium containing 50 nM hGH and then were lysed as described (11). Cell extracts were used for determination of luciferase and ␤-galactosidase activities and luciferase activity was normalized to ␤-galactosidase activity.

CORT Cloning Methodology Applied to the Non-tyrosine Kinase GH Receptor-
The principle of the CORT cloning methodology involves the use of the tyrosine-phosphorylated COOH terminus of tyrosine kinase receptors in order to identify new SH2 domain containing molecules (17,18). The intrinsic tyrosine kinase activity of the cytoplasmic domain of EGF receptor allowed the phosphorylation and the radiolabeling with 32 P of the tyrosine residues in a single step. As the GHR cannot autophosphorylate, we prepared our probe in two steps: first we expressed the GHR COOH terminus from residue 454 to 620 downstream of the GST and a protein kinase A phosphorylation site in TKB1 bacteria, which display intrinsic Elk tyrosine kinase activity. This tyrosine kinase permitted very strong (non-radioactive) phosphorylation of the receptor (data not shown). Second, the tyrosine-phosphorylated GST-protein kinase A phosphorylation site, GHR 454 -620 , was purified and radiolabeled in vitro with protein kinase A and high specific activity [␥-32 P]ATP.
In order to check the specificity of our probe, we incubated it with different bacterially expressed proteins spotted on filters. The phosphorylated GST-GHR 454 -620 only bound to SH2 domains of certain signaling proteins such as p85␣ and phospholipase C␥ as well as SHP-2 (Fig. 1A). No binding was detected with SH2 domains of Grb2 and SHP-1. The integrity of these fusion proteins were verified by their ability to bind specifically to tyrosines contained in Shc or in the erythropoietin receptor, respectively, as described (8,28) (Fig. 1B). Control GST also did not bind to the probe (Fig. 1A). Moreover, when the fusion construct was expressed in MH1 bacteria and was thus not tyrosine phosphorylated, no interaction was observed, suggesting that the positive signals obtained are specific to the phosphorylated tyrosines. The tyrosine-phosphorylated and radiolabeled probe was then used to screen the EXlox embryonic mouse expression library. Among 1.6 million clones screened, eight positives were obtained. Sequencing revealed that they encoded six different proteins, all of which contained SH2 domains (Fig. 2). The clones obtained twice also gave the strongest signals. They encoded the adaptor Shc and the p85 subunit of phosphatidylinositol 3-kinase. Shc has been previously shown to interact with Jak2 upon GH stimulation (8), but our result demonstrates that Shc can also be a direct target of GHR. Moreover as the clone obtained does not contain the PTB coding sequence, we show that Shc can interact directly with GHR through its SH2 domain. The p85 clone is identical to the ␤ isoform of the p85 subunit of phosphatidylinositol 3-kinase. The protein Csk, which regulates the activity of c-Src family molecules (29) and the mouse homologue of the adaptor Shb (30) were also identified. DNA sequencing of the last two clones revealed that they are highly related to Grb4 and Grb10 (18,31). Our Grb4 clone is 61% identical at the protein level to human Nck, and 94% identical to the partial protein sequence of mouse Grb4 (18).
Identification of Grb10 as a New GHR Target-The Grb10 clone contains a 0.8-kilobase insert encoding a 160-aa polypeptide which is almost identical to the carboxyl-terminal part of the recently identified signaling mediator Grb10 (Fig. 3A) (31). The SH2 domain of our clone, which was named Grb10-GHR, is 100% identical to the SH2 domain of the previously published mouse Grb10 sequence (31). However, two contiguous aa substitutions (Asn 491 3 Lys, Gly 492 3 Arg) were observed in the region of the molecule upstream of the SH2 domain. In addition, there is a divergence located 70 base pairs after the stop codon, between the 3Ј-untranslated region from our clone which is terminated by a cluster of Ala, and the 3Ј-untranslated region of the published mouse Grb10 sequence (Fig. 3A).
Moreover interaction between the receptor and the Grb10-GHR clone expressed in the library was characterized using as a probe the 32 P-labeled COOH-terminal domain of the GHR expressed in MH1 bacteria. Under these conditions, there was no interaction of the clone with the probe as shown by the lack of radioactive signal on the right portion of Fig. 3B. This indicates that the interaction between Grb10-GHR and the receptor involves one or several of the phosphorylated tyrosine residue(s) located between amino acids 454 and 620 of the receptor. Under our experimental conditions, one or several of the tyrosine residues present in the fusion protein could be phosphorylated. For all the positive clones expressed in the library, the interaction was dependent on the tyrosine phosphorylation of the probe (data not shown).
Association between Grb10 and GHR in Huh-7 Human Hepatoma Cells-Human hepatoma cells and in particular Huh-7 cells were reported to express hGrb10 (24) and hGHR (25,26). We thus used this cell line to test the association between hGrb10 and the receptor (Fig. 4). In lysates from cells stimulated with GH, a broad 120-kDa tyrosine-phosphorylated band was co-precipitated specifically with Grb10 antibodies; this band was absent in immunoprecipitates with control antibody. Reprobing of the blot with an anti-GHR antibody identified this band as the receptor. A control blot with the anti-Grb10 antibody confirmed that similar amounts of hGrb10 were present in the two conditions of immunoprecipitation and that the association was induced by the hormonal stimulation.
Association between Grb10 and GHR in Transfected 293 Cells-As we were not able to efficiently immunoprecipitate the receptor with our anti-GHR antibodies, we decided to construct an epitope Flag-tagged GHR to study in more detail the interactions between the receptor complex and Grb10. Constructs were done with GHR wt and a carboxyl-terminal truncated mutant lacking the cytoplasmic domain that we used for the screening of GHR-associated molecules. 293 cells were cotransfected with expression vectors containing the cDNAs of mGrb10 and Jak2, GHR wt (Fig. 5A), or GHR 454 (Fig. 5B). Lysates from cells stimulated or not with hGH were immunoprecipitated with an antibody directed against the Flag epitope fused to the receptor. Immunoblotting with an anti-Grb10 antibody revealed the association of the three isoforms of mGrb10 with the GHR wt (Fig. 5A, top panel), as well as with the deleted mutant GHR 454 (Fig. 5B, top panel). The three bands could correspond to Grb10 isoforms generated by alternative use of initiation sites of translation as reported (31). When GHR cDNA was not transfected, a faint signal was detected suggesting that a low amount of Grb10 nonspecifically associated with the anti-Flag antibody or the beads. When Grb10 cDNA was not transfected, no band was seen, suggesting the absence of a detectable level of endogenous Grb10 in the 293 cells. Similar amounts of Flag-tagged GHRs were observed in each lane, as shown after membrane stripping and blotting with an anti-Flag antibody (Fig. 5, A and B, lower panels). Anti-phosphotyrosine immunoblots showed that Grb10 has no effect on the GH-induced tyrosine phosphorylation of the full-length receptor upon GH stimulation (Fig. 5A, middle panel). In contrast, GHR 454 was not significantly phosphorylated under hormonal stimulation as reported previously (11) (Fig. 5B, middle panel).

FIG. 3. Structure of Grb10-GHR and interaction with GHR 454 -620 .
A, comparison of DNA sequences between the partial Grb10-GHR cDNA and the published mouse Grb10 sequence (31). A, n represents the cluster of Ala, and **, the two aa substitutions described in the text. B, the COOH-terminal GHR (GHR 454 -620 ) was phosphorylated or not on tyrosine residues and incubated with a filter containing protein expressed by the purified Grb10-GHR clone, 32.

FIG. 4. Association between Grb10 and GHR in human hepatoma Huh-7 cells.
Huh-7 hepatoma cells were grown to confluence in P150 plates and were then stimulated or not with hGH for 10 min. Lysates were incubated overnight with anti-Grb10 or preimmune antibodies and then proteins transferred to membranes were analyzed with anti-Tyr(P), anti-GHR, or anti-Grb10 antibodies. Molecular mass markers (kDa) are shown at the right of panels.
We thus speculated that Grb10 association with the truncated receptor might be mediated through an interaction with the receptor associated kinase Jak2.
Association between Grb10 and the Tyrosine Kinase Jak2-293 cells were transfected with or without 2 g of Jak2 cDNA (under these conditions Jak2 is able to autophosphorylate) and with or without 2 g of Grb10 cDNA (Fig. 6, A-C). When cell lysates were immunoprecipitated with anti-Grb10 antibody and revealed with anti-Tyr(P) antibody (Fig. 6B), a 120-kDa tyrosine-phosphorylated band was present in Grb10 immunoprecipitates. This band was only present when Jak2 cDNA was co-transfected and was absent in control experiments with preimmune serum (Fig. 6A). When cell lysates were immunoprecipitated with anti-Jak2 antibody (Fig. 6C) and revealed with anti-Tyr(P) antibody, no significant difference in Jak2 tyrosine phosphorylation could be observed due to Grb10 overexpression. The data shown in Fig. 6, B and C, suggested that Grb10 is also able to associate with Jak2 without affecting its level of tyrosine phosphorylation. However, stripping and reprobing of this blot with an anti-Jak2 antibody failed to directly identify the phosphorylated band as Jak2. It is possible that only a low amount of Jak2 bound to Grb10 under these conditions. This hypothesis is consistent with the fact that the anti-Tyr(P) antibody exhibits stronger affinity than the anti-Jak2 antibody. In vitro experiments were then performed to confirm Grb10 interaction with Jak2 (Fig. 6D). Lysates from cells transfected with or without Jak2 cDNA were incubated with GST alone, GST-SH2Grb10, or immunoprecipitated with anti-Jak2 antibody. Specific binding of Jak2 to the Grb10 SH2 domain was detected by an anti-Jak2 immunoblot while only a very low amount of Jak2 associated with GST alone.
Increasing amounts of Grb10 cDNA strongly inhibited the 3-fold GH induction of SRE-tk-Luc reporter gene (70% inhibition) and also to a lesser extent the 9-fold GH induction of the GHRE2 promoter (30% inhibition) (Fig. 7). In contrast, no significant difference was observed in the 15-fold stimulation by GH of LHRE-tk-Luc. Moreover, basal levels remained unchanged (data not shown). Only the GH-stimulated fraction decreased significantly in the SRE and GHRE2 tests. Control transfections with the empty vector of Grb10 had no effect. Thus, Grb10 could be involved as a negative regulator of specific enhancer activities of GH-stimulated genes.
Grb10 Does Not Interfere with Stat5 Tyrosine Phosphorylation-Our results with the three functional tests further suggested that Grb10 interferes with GHR signaling downstream of Jak2 and independently of Stat5 activation as the three pathways are dependent of Jak2 activation and both LHRE and GHRE2 can bind Stat5. The data shown in Fig. 6C demonstrated that overexpression of Grb10 did not inhibit Jak2 tyrosine phosphorylation. In Fig. 8, we analyzed the effect of Grb10 overexpression on Stat5 tyrosine phosphorylation. 293 cells were transfected with GHR and Grb10 cDNAs in a relative ratio similar to those used in the functional test. We show that Grb10 overexpression does not inhibit endogenous Stat5 tyrosine phosphorylation. Hence, the effect of Grb10 on the GHRE2 promoter is probably independent of any interference with Stat5 activation. DISCUSSION We used the tyrosine-phosphorylated carboxyl-terminal part of the GHR to screen a EXlox expression library by a procedure derived from the CORT methodology (17,18). The specificity of our probe tyrosine phosphorylated in TKB1 bacteria was preliminarily tested in binding experiments with different SH2-containing proteins. The probe did not associate directly with SH2 domains of Grb2, or SHP-1, as previously suggested (8). In contrast, the probe bound efficiently to the two SH2 domains of p85␣, phospholipase C␥ and SHP-2. Taken together, these data demonstrate that the probe could discrimi- FIG. 5. Association between Grb10, GHR wt , and GHR 454 in 293 cells. 293 cells were co-transfected with expression plasmids containing the Flag-tagged GHR wt (panel A) or GHR 454 (panel B) (4 g), Jak2 (0.5 g), and Grb10 (4 g). Proteins were immunoprecipitated with an anti-Flag antibody and characterized by immunoblotting with an anti-Grb10 antibody (upper panels). Membranes were then stripped and incubated with anti-Tyr(P) antibody (middle panels) or with anti-Flag antibody (lower panels). Molecular mass markers (kDa) are shown at the left of panels.
FIG. 6. Grb10 associates with Jak2. 293 cells were co-transfected with 2 g of Grb10 cDNA and/or 2 g of Jak2 cDNA. Lysates were incubated with control preimmune serum (panel A), anti-Grb10 antibody (panel B), or with an anti-Jak2 antibody (panel C) and analyzed in Western blot with an anti-Tyr(P) (upper panels A-C). After stripping, proteins from panels A and B were incubated with an anti-Grb10 antibody and proteins from panel C with an anti-Jak2 antibody. D, 293 cell lysates transfected with (ϩ) or without (Ϫ) 2 g of Jak2 cDNA were incubated with GST alone or GST Grb10 SH2 fusion protein. As control, Jak2 was immunoprecipitated with an anti-Jak2 antibody. Proteins were analyzed by immunoblot with an anti-Tyr(P) (upper panel) or with an anti-Jak2 antibody (lower panel). Molecular mass markers (kDa) are shown at the left of each panel.
nate between various SH2 containing proteins. The efficiency of the radiolabeling of the probe combined with the high expression of the library clones under control of the T7 polymerase promoter (18) allowed detection of a direct and strong interaction between Shc or p85␤ and the receptor. Shc has been shown to be phosphorylated and to interact with Jak2 after GH stimulation (8). Although Tyr 469 of GHR is part of a consensus site for Shc binding, direct interaction of GHR and Shc could never be shown previously by other techniques. Our probe thus is a sensitive tool to visualize direct interactions. It seems that Shc can interact with both GHR 454 -620 and Jak2 and, hence that Shc-dependent pathways, such as the mitogen-activated protein kinase activation, could be linked to the proximal and the distal parts of the GHR cytoplasmic domain. Several bind-ing sites for Shc have also been reported for other receptors such as platelet-derived growth factor and epidermal growth factor receptors (38,39). GH also stimulates tyrosine phosphorylation of insulin receptor substrate-1/2 and, thereby the interaction of p85 with insulin receptor substrate-1/2 (40 -43). However, no evidence for a direct interaction between the receptor or Jak2 and insulin receptor substrate-1/2 or p85 has been reported. Our results show that p85␤ and p85␣ can bind directly and efficiently to the carboxyl-terminal part of GHR and that this interaction depends on the presence of the phosphorylated tyrosines. Identification of Csk among our positive clones suggests that molecules of the Src family (29) could be involved in GHR signaling. Two Src family molecules, c-Src and Fyn, have been reported to be involved in signal transduction of prolactin receptor (44,45) which exhibits strong structural and functional similarities with the GHR (46). The molecule Grb4, whose full-length cDNA has not yet been cloned (18), as well as the Shb adaptor (30), are newly identified molecules; their role in signaling is unclear. However, Shb has been suggested to transduce apoptotic signals from tyrosine kinase receptors under certain conditions (47). All these new molecules that specifically bind to GHR presumably via an SH2 domain could be involved potentially in new signal transduction pathways that could explain some of the multiple effects of GH.
One isolated clone exhibited a very high homology with the recently identified SH2 protein Grb10 (31). The slight differences observed in the coding region between Grb10-GHR and the mouse Grb10 cDNA could be due to genetic polymorphism, since a single gene seems to encode Grb10 in mouse and the two clones are derived from the same species, but from a different library. The divergence in the 3Ј-untranslated region of the molecule could be due to an alternative splicing. Four splice variants of Grb10 have been reported in humans; they are truncated in the pleckstrin homology domain or modified in the NH 2 -terminal region (24, 48 -50). One mouse isoform which is truncated by 75 base pairs between the proline-rich region and the pleckstrin homology domain coding sequences has also been identified (23).
hGrb10 mRNA appears to be ubiquitously expressed but was specially strongly found in muscle and adipose tissue (49,50). It was also found in various beast cancer cell lines and hepatoma cells (24). We thus tested the association of hGrb10 with the GHR in the Huh-7 hepatoma cell line which appeared the most enriched in Grb10 and which was also reported to express the GHR (25,26). We were able to detect a band corresponding to the GHR in Grb10 immunoprecipitates when cells were stimulated by GH. However, to characterize more easily the association of the receptor complex with Grb10 we used Flagtagged receptors co-transfected with mGrb10 in 293 cells.
Our study reports the first demonstration of an interaction of Grb10 with a nontyrosine kinase receptor. Different tyrosine kinase receptors such as Ret (19), insulin (56), insulin-like growth factor-1 (21), and ELK receptors (22) have been reported to associate with Grb10. In our screening procedure, we have shown that Grb10 has the ability to interact directly with the phosphorylated tyrosines of the region 454 -620 of GHR. However, in transfected cells, Grb10 interacts with the complex formed by the full-length or the truncated GHR and Jak2. So we conclude that Grb10-like Shc can bind to alternative sites on the GHR⅐Jak2 complex. A very recent study reports that two independent but cooperative binding sites to tyrosine-phosphorylated proteins are located in the carboxyl-terminal part of Grb10: the SH2 domain and an additional domain, the BPS domain, located between the pleckstrin homology and the SH2 domains (51). Future studies will test the relative importance of these domains in the association between Grb10 and the GHR. Very little is known about the functional role of Grb10 in signal transduction. We show that Grb10 inhibits GH stimulated luciferase activity under the control of both the SRE element of the c-fos gene and the GHRE2 element of the Spi2.1 gene. These effects do not seem to be related to a nonspecific toxic effect of Grb10, since overexpression of Grb10 does not affect the Jak2-Stat5 pathway. A human Grb10 isoform, hGrb-IR, which associates with the insulin receptor has been shown to inhibit tyrosine phosphorylation of insulin receptor substrates, p60 GTPase-activating protein-associated protein and phosphatidylinositol 3-kinase (48). Our results demonstrate that Grb10 does not inhibit GHR or Jak2 phosphorylation nor the tyrosine phosphorylation of Stat5. Thus, Grb10 could interact downstream of Jak2, inhibiting the activity of other molecules involved in the activation of SRE and GHRE2. One possibility is that the SH2 domain of Grb10 competes with the SH2 domain of other molecules. Results obtained with ELKR (22) and Ret (20) nevertheless suggest that Grb10 would not compete with the Grb2 molecule as binding sites of Grb2 and Grb10 to phosphorylated tyrosines of these receptors are distinct. Moreover, Grb10 does not compete with Grb2 for interaction with Shc (52). Another hypothesis is that Grb10 would be involved in the regulation of the C/EBP pathway. GHRE2 has been shown to bind Stat5 (6) and C/EBP molecules (37). C/EBP␤ can also bind to the SRE and activates c-fos (35,36). Furthermore, GH has been shown to stimulate C/EBP expression (53). This hypothesis of an effect of Grb10 on the C/EBP pathway is consistent with the absence of effect on LHRE reporter gene activation which contains only Stat5 binding elements.
Our data show that Grb10 is potentially involved in GHmediated c-fos down-regulation and thus could play a role in modulating negatively the proliferative effects induced by GH in some cell lines. However, the mechanisms potentially involved in Grb10 regulation of cell proliferation are still obscure. Conflicting results relative to the possible inhibition of insulin dependent mitogenic pathway by microinjection of Grb10 SH2 alone (49) or SH2 and BPS domains were reported (51). In addition, overexpression of Grb10 in P6 cells has also been reported to inhibit cell cycle transition from S to G 2 phase in insulin-like growth factor-1 receptor signaling (52).
Grb10 belongs to a new family of molecules including Grb7 (18) and Grb14 (54) and the Caenorhabditis elegans MIG-10 protein (55). This family is characterized by a common central structural motif called the GM region including the pleckstrin homology domain. MIG-10 is required for migration of embryonic neurons and proper development of excretory canals (55). Little is known about functions of Grb7 or Grb14, however, Grb7 and Grb14 mRNA are overexpressed in some breast cancer cell lines, suggesting also a role of these molecules in cell proliferation (54,56).
With the CORT methodology applied to the tyrosine-phosphorylated GHR, we successfully identified for the first time new substrates of a nontyrosine kinase receptor. Among these molecules Grb10 appears to be a new regulator of certain GHR transcriptional activities. A similar screening procedure has been initiated in order to directly identify the proper targets of Grb10 in GHR as well as in other receptor signaling pathways.