Cloning, Chromosome Localization, Expression, and Characterization of an Src Homology 2 and Pleckstrin Homology Domain-containing Insulin Receptor Binding Protein hGrb10γ*

hGrb10α (previously named Grb-IR) is a Src-homology 2 domain-containing protein that binds with high affinity to the tyrosine-phosphorylated insulin receptor and insulin-like growth factor-1 receptor. At least two isoforms of human Grb10, (hGrb10α and hGrb10β), which differ in the pleckstrin homology (PH) domain and the N-terminal sequence, have previously been identified in insulin target tissues such as human skeletal muscle and fat cells. Here we report the cloning of the third isoform of the hGrb10 family (hGrb10γ) from human skeletal muscle and its localization to human chromosome 7. We have also determined the human chromosome localization of Grb7 to 17q21-q22 and Grb14 to chromosome 2. hGrb10γ contains an intact PH domain and an N-terminal sequence that is present in hGrb10α but absent in hGrb10β. RNase protection assays and Western blot analysis showed that hGrb10α and hGrb10γ are differentially expressed in insulin target cells including skeletal muscle, liver, and adipocyte cells. hGrb10γ is also expressed in HeLa cells and various breast cancer cell lines. The protein bound with high affinity to the insulin receptor in cells, and the interaction was dependent on the tyrosine phosphorylation of the receptor. hGrb10γ also underwent insulin-stimulated membrane translocation and serine phosphorylation. hGrb10γ phosphorylation was inhibited by PD98059, a specific inhibitor of mitogen-activated protein kinase kinase, and wortmannin, a specific inhibitor of phosphatidylinositol 3-kinase. Taken together, our data suggest that hGrb10 isoforms are potential downstream signaling components of the insulin receptor tyrosine kinase and that the PH domain may play an important role in the involvement of these isoforms in signal transduction pathways initiated by insulin and other growth factors.

hGrb10␣ (previously named Grb-IR) is a Src-homology 2 domain-containing protein that binds with high affinity to the tyrosine-phosphorylated insulin receptor and insulin-like growth factor-1 receptor. At least two isoforms of human Grb10, (hGrb10␣ and hGrb10␤), which differ in the pleckstrin homology (PH) domain and the N-terminal sequence, have previously been identified in insulin target tissues such as human skeletal muscle and fat cells. Here we report the cloning of the third isoform of the hGrb10 family (hGrb10␥) from human skeletal muscle and its localization to human chromosome 7. We have also determined the human chromosome localization of Grb7 to 17q21-q22 and Grb14 to chromosome 2. hGrb10␥ contains an intact PH domain and an N-terminal sequence that is present in hGrb10␣ but absent in hGrb10␤. RNase protection assays and Western blot analysis showed that hGrb10␣ and hGrb10␥ are differentially expressed in insulin target cells including skeletal muscle, liver, and adipocyte cells. hGrb10␥ is also expressed in HeLa cells and various breast cancer cell lines. The protein bound with high affinity to the insulin receptor in cells, and the interaction was dependent on the tyrosine phosphorylation of the receptor. hGrb10␥ also underwent insulin-stimulated membrane translocation and serine phosphorylation. hGrb10␥ phosphorylation was inhibited by PD98059, a specific inhibitor of mitogen-activated protein kinase kinase, and wortmannin, a specific inhibitor of phosphatidylinositol 3kinase. Taken together, our data suggest that hGrb10 isoforms are potential downstream signaling components of the insulin receptor tyrosine kinase and that the PH domain may play an important role in the involvement of these isoforms in signal transduction pathways initiated by insulin and other growth factors.
Insulin stimulates and regulates cell growth and metabolism by binding to its receptor on the cell membrane. The binding of insulin to the insulin receptor (IR) 1 results in receptor autophosphorylation and receptor tyrosine kinase activation, followed by tyrosine phosphorylation of various cellular substrates including a 185-kDa protein called the IR substrate 1 or IRS-1 (1,2). Tyrosine phosphorylation of IRS-1 creates docking sites for multiple downstream signaling molecules with specific sequence motifs. One of these functional motifs found in signaling proteins is the Src homology 2 (SH2) domain. The SH2 domain is a sequence of approximately 100 amino acids that binds with high affinity to phosphotyrosine-containing proteins (3). Another functional domain is the pleckstrin homology (PH) domain that has been suggested to play important roles in protein-protein and protein-lipid interactions (4 -6).
Although numerous studies have shown that IRS-1 is critical in IR signal transduction, evidence does exist that other proteins may also be involved to transduce a signal from the IR to downstream targets (2). In a search for signaling molecules involved in IR signaling, we used the yeast two-hybrid system to find proteins that interact directly with the cytoplasmic domain of the human IR (7). We identified an SH2 domaincontaining protein (hGrb10␣, previously named Grb-IR) that binds with high affinity to the autophosphorylated IR. Sequence comparison of hGrb10␣ with several recently cloned proteins including mGrb7 (8), mGrb10 (9), and hGrb14 (10) suggests that they belong to a special family. All of these proteins contain an SH2 domain at their C termini and a PH domain in the central regions. The SH2 domain of hGrb10 isoforms is 99, 79, and 59% identical to that of mGrb10, mGrb7, and hGrb14, respectively. The PH domains of these proteins are also highly homologous. Although the N termini of these proteins are less conserved, all of them contain a highly conserved proline-rich sequence (P(S/A)IPNPFPEL) (see Fig. 1A), which has been shown to be the potential binding site for SH3 domain-containing proteins (11). In reverse transcription-PCR experiments we also showed that there are at least two isoforms of hGrb10, one that contains and one that lacks an intact PH domain. Expression of hGrb10␣, the isoform lacking an intact PH domain, inhibits insulin-stimulated substrate tyrosine phosphorylation and PI 3-kinase activation in cells (7). The mechanism of inhibition has not been characterized.
To further study the role of Grb10 proteins in signaling, we cloned and characterized the PH domain-containing isoform and named it hGrb10␥. We find that hGrb10 isoforms are differentially expressed in insulin target cells and various human breast cancer cell lines and are phosphorylated differently in response to insulin stimulation. In addition, we have shown that the insulin-stimulated hGrb10␥ phosphorylation was blocked by wortmannin, a relative specific inhibitor of PI 3kinase, and PD98059, an inhibitor of MAP kinase. Our findings suggest that Grb10 isoforms are potential signaling components of pathways mediated by insulin and other growth factors and that the PH domain may play an important role in the function of the protein.

EXPERIMENTAL PROCEDURES
cDNA Cloning, Reverse Transcription-PCR, and RNase Protection Assays-The cDNA encoding hGrb10␥ was identified by screening a human muscle cDNA library (Stratagene) using a radiolabeled 0.9-kb hGrb10␣ cDNA fragment (7) as a probe. The nucleotide sequence of this clone was determined by the dideoxy chain termination method using the Sequenase 2.0 sequencing kit (Amersham Corp.). Total RNAs from human breast cancer cell lines MCF-7 and MDA-435A were isolated using the Totally RNA kit from Ambion (Austin, TX), and the first strand cDNAs were synthesized using the Superscript kit from Life Technologies, Inc. First strand cDNA from HEK 293 cells was a gift of L. Ballou (University of Texas Health Science Center at San Antonio). The forward PCR primer (P1) was: 5Ј-TTGAAGAAGGCAGAAG-GAACCC-3Ј, which is in the 5Ј-untranslated region that is specific for hGrb10␣ and hGrb10␥. The reverse PCR primer (P2) was: 5Ј-CATTGC-CACGAGGGAGTTCTCGGA-3Ј, which is located downstream of the cDNA sequence encoding for the PH domain and is common in all three hGrb10 isoforms (Fig. 1B). These primers amplify fragments with predicted sizes of 1.17 and 1.3 kb for hGrb10␣ and hGrb10␥, respectively. To determine the expression of Grb10 isoforms in different human breast cancer cells and to assess the relative abundance of the isoform mRNAs, a bifunctional antisense [ 32 P]riboprobe was synthesized in the presence of [ 32 P]uridine triphosphate using a cDNA template consisting of a 40-nt fragment of hGrb10␥ cDNA encoding the PH domain inserted into the EcoRI and BamHI restriction sites of the pBluescript II SK(ϩ) vector and oriented antisense to the T3 RNA polymerase site. The probe protects a 260-nt sequence of hGrb10␣ and a 398-nt sequence of hGrb10␤ or hGrb10␥ mRNA. RNase protection assays were carried out by hybridizing the probe with 20 g of total RNA isolated from several breast cancer cell lines (MCF-7, ZR-75, T47D, MDA-468, MDA-231, MDA-231VC (12), MDA-435A) and a prostate cancer cell line (DU145). For 36B4 loading control (13), a 145 PstI-PstI fragment was cloned into pGEM4Z, linearized with EcoRI, and transcribed with T7 RNA polymerase. Single-stranded RNA was digested with RNase A, and samples were separated by 8 M urea 6% SDS-PAGE (14). tRNA was run as a negative control. RNA from Chinese hamster ovary (CHO) cells stably transfected with hGrb10␣ (CHO/IR/hGrb10␣) and hGrb10␥ (CHO/IR/ hGrb10␥) was run as a positive control.
Chromosome Mapping-The chromosome location of the hGrb10 gene was determined by PCR using hGrb10␣-specific primers against the Stanford Human Genome Center G3 radiation hybrid panel (Research Genetics, Inc.). The PCR primers were: 5Ј-CAGCAGCAAATCG-TCGTTTA-3Ј and 5Ј-GAACACTGGAGTGAAGAAGCG-3Ј. The PCR results were analyzed using the RH server Version 2.0. 2 Additional data were obtained from BLAST searches of the sequence-tagged site (STS) data base. 3 Phylogenetic Analysis-A multiple alignment of all members of the Grb10, Grb7, and Grb14 protein family was generated using Clustalw (15). The alignment was used in a maximum parsimony analysis using Phylogenetic Analysis Using Parsimony (PAUP) (16). The statistical significance of the resulting trees was evaluated using bootstrap analysis within PAUP.
Cell Lines and Anti-hGrb10 Antibody-CHO cells expressing both the human IR and hGrb10␣ have been described previously (7). To establish cell lines expressing the IR and hGrb10␥, we subcloned hGrb10␥ cDNA into the mammalian expression vector pBEX (17) inframe with a sequence encoding a 9-amino acid hemagglutinin-tag (YPYDVPDYA) to generate the recombinant plasmid pBEX/hGrb10␥. Transfection of CHO/IR cells with plasmids pBEX/hGrb10␥ and pBS pac⌬p (18) was carried out using the calcium phosphate method as described previously (19). Stable cell lines expressing hGrb10␥ (CHO/ IR/hGrb10␥) were selected with 8 g/ml puromycin. Positives were identified by Western blotting using the antibody to the hemagglutinin tag or to hGrb10 (7) and cloned by limiting dilution. The hepatoblastoma cell line Huh7 was from M. Adamo (20). All other cell lines used in this study were from ATCC. Anti-hGrb10 rabbit polyclonal antisera were raised against the C-terminal region of Grb-IR/hGrb10␣ (amino acids 369 -548) fused to the GST protein (GST/hGrb10c) as described previously (7). The antibody was purified by passing through a GSTimmobilized column to deplete antibodies to GST and then eluted from a GST/hGrb10c-immobilized column with a buffer containing 0.1 N glycine, pH 2.8, 0.1% Triton X-100, and 150 mM NaCl and neutralized with 1 M Tris-HCl, pH 9.0.
Potato Alkaline Phosphatase (PAP) Treatment-Insulin-treated or nontreated CHO/IR/hGrb10␣ and CHO/IR/hGrb10␥ cells were lysed in buffer containing 50 mM Hepes, pH 7.6, 150 mM NaCl, 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, and 10 g/ml leupeptin. After incubation at 0°C for 10 min, the lysates were centrifuged at 12,000 ϫ g for 10 min at 4°C. The supernatants were incubated at 30°C for 20 min with PAP in the presence or absence of phosphatase inhibitors (1 mM sodium orthovanadate, 20 mM sodium pyrophosphate, 10 mM sodium fluoride, and 20 mM ␤-glycerolphosphate). The reaction was terminated by adding SDS-polyacrylamide gel electrophoresis (PAGE) sample buffer and boiling at 95°C for 3 min.
Phosphoamino Acid Analysis-hGrb10␥ bound on the Immobilon P membrane was cut out, and the protein was hydrolyzed under vacuum for 1 h at 110°C in 1 ml of 6 M HCl (Pierce). The membrane was rehydrated by the addition of 1 drop of methanol, and the amino acids were eluted from the membrane with water. After lyophilization, the sample was resuspended in 10 l of phosphoamino acid standard containing phosphotyrosine, phosphoserine, and phosphothreonine. Approximately 150 cpm of each sample was spotted onto a TLC cellulose plate (EM Separation Technology). Two-dimensional thin-layer electrophoresis was carried out under the following conditions: first dimension, pH 1.9 buffer (formic acid/acetic acid/water 1:3.12:35.88), 1.5 kV for 20 min; second dimension, pH 3.5 buffer (acetic acid/pyridine/water 10:1:189), 1.3 kV for 16 min. The plate was dried and exposed to x-ray film.
Treatment of Cells with PD98059 and Wortmannin-Cells in a 6-well plate (80 -90% confluent) were incubated with serum-free medium for 1 h at 37°C. Wortmannin (Sigma) or PD98059 (Calbiochem) were added to wells at a final concentration of 50 nM and 50 M, respectively. After incubation for 1 h, the cells were stimulated with 10 Ϫ8 M insulin for 15 min. Cells were then washed with a buffer containing 50 mM phosphate, pH 7.4, 150 mM NaCl, lysed in SDS sample buffer, and subjected to SDS-PAGE.
Immunofluorescence Studies-CHO/IR/hGrb10␥ cells were grown on glass slides to a 70% confluence and treated with or without 10 Ϫ8 M insulin for 15 min. After rinsing with cold phosphate-buffered saline (PBS), the cells were fixed in 4% formaldehyde for 5 min, rinsed three times with cold phosphate-buffered saline, and incubated with acetone for 1 min. After blocking with 1% fetal bovine serum for 30 min at room temperature, the cells were incubated with a polyclonal anti-hGrb10 antibody (1:1,000 dilution) for 1 h at room temperature. The cells were then rinsed with cold phosphate-buffered saline and incubated with fluorescein-conjugated goat anti-rabbit IgG for 30 min at room temperature. The localization of hGrb10␥ was examined using a fluorescent microscope.
Immunoprecipitation, SDS-PAGE, and Western Blot Analysis-Coimmunoprecipitation of the IR and hGrb10 isoforms were performed as described previously (7). SDS-PAGE was carried out using 10% (w/v) polyacrylamide gels. After electrophoresis, proteins on the gel were transferred to nitrocellulose membranes for Western blot analysis. The immunoblots were blocked at room temperature with buffer containing 10 mM Tris-HCl, pH 7.5, 154 mM NaCl, 0.1% Tween 20, and 1% (w/v) dry milk and then incubated with antibody against hGrb10. hGrb10 isoforms were detected with alkaline phosphatase-conjugated anti-rabbit IgG secondary antibody and a chromatogenous substrate reaction.

RESULTS
Identification and Cloning of hGrb10␥-We have previously shown by reverse transcription-PCR that there are at least two hGrb10 isoforms that differ in the PH domain region (7). To clone the cDNA encoding for the isoform containing an intact PH domain, we screened a human muscle cDNA library using a 0.9-kb hGrb10␣ cDNA as a probe. The deduced amino acid sequence of clone 2, which had an insert of 1.7 kb, is shown in Fig. 1A. This cDNA sequence is identical to that encoding for hGrb10␣, except for a 138-nt addition in the PH domain encoding for an additional 46 amino acids (Fig. 1B). A search of GenBank TM revealed several other hGrb10-like sequences that include a newly deposited cDNA sequence identified from human myeloblast (KIA0207; sequence accession number D86962) and an N-terminal 58-amino acid shorter form, Grb10/ IR-SV1 or hGrb10␤ (11,22). The amino acid sequences of hGrb10␥, hGrb10␤, and KIA0207/hGrb10␦ are exactly the same except for the N-termini (Fig. 1C).
By reverse transcription-PCR, we also confirmed the presence of hGrb10 isoforms in other human cell lines. As shown in Fig. 1D, the isoform-specific primers amplified several DNA fragments from mRNAs isolated from the HEK 293 cells (lane 3) and two representative human breast cancer cell lines (lane 4 and 5). Comparison of the sizes of the PCR products amplified from these cells with those amplified using either hGrb10␣ (Fig. 1D, lane 1) or hGrb10␥ (Fig. 1D, lane 2) as template suggests that mRNAs of both isoforms were present in these cell lines and that the major isoform in these cells is the one containing an intact PH domain. The detection of an additional PCR product with a molecular mass greater than that of hGrb10␥ suggests that there may be another isoform in these cells, which is consistent with the observation by Frantz et al. (11), who found the presence of an additional PCR product in human skeletal muscle using hGrb10␣-specific primers.
Chromosome Localization-PCR analysis of the Stanford Human Genome Center G3 radiation hybrid panel places the gene for hGrb10 on human chromosome 7 between markers D7S506 and D7S499 (Fig. 2). BLAST searches of the STS data base using segments of the hGrb10 gene specific for the ␣, ␤, and ␥ isoforms identify three specific STS markers (SWSS3378, WI-14569, and WI-18551), which have also been mapped within this interval (23). These STS markers were originally derived from the hGrb10␣ cDNA. There are several other uncharacterized cDNA transcripts that have been mapped in the region, and the nearest other gene is the IGFBP1 gene (24). The murine Grb10 gene has been mapped to chromosome 11 at position 8.0 cm (9). This region of mouse chromosome 11 is syntenic with human p13-p11, based upon a survey of murine genes that have been mapped to chromosome 11. 4 BLAST searches of the STS data base using hGrb7 and hGrb14 have also identified STS markers that have been mapped to their respective chromosome localization. WI-9366 is 100% identical to hGrb7 and has been mapped to YAC 946 E 2, which has been localized to human chromosome 17q21-q22 near topoisomerase 2. Grb14 matches STS marker WI-11831, which has been mapped to human chromosome 2 at position 836.45 centirads using a radiation hybrid approach. 5 Expression of hGrb10 -It has been shown that Grb7 and Grb14, the other two members of the Grb7/10/14 gene family, were overexpressed in certain breast cancer cells (9,10). To investigate whether Grb10 was also expressed in these cells, several breast cancer cell lines were analyzed by RNase protection assays using a probe that spanned the PH splice domain of hGrb10␥ (Fig. 3A). As a positive control, total mRNA from CHO cells stably transfected with hGrb10␣ (Fig. 3A, ⌬PH) or hGrb10␥ (Fig. 3A, PH) were also analyzed. The majority of cells show a protected band specific for the PHϩ isoform (expected size 398 nt) with no significantly detectable bands of the splicevariant hGrb10␣ (expected size, 260 nt) (Fig. 3A).
Using an anti-hGrb10 polyclonal antibody directed against the C-terminal region of the protein, we also examined hGrb10 protein expression in several breast cancer cell lines. Unlike Grb7, which is expressed in only a limited number of breast cancer cell lines (25), an anti-hGrb10 antibody-reactive protein was detected in all breast cell lines studied (Fig. 3B). In ZR-75 and HS578T cells, multiple bands were detected by the antibody, probably due to a phosphorylation-induced gel mobility shift of the protein (see below). Comparison of the apparent molecular mass of this protein with hGrb10 isoforms expressed in the transfected CHO/IR cells suggests that they are the PH domain-containing isoform hGrb10␥. In HeLa cells, from which hGrb10␣ was originally identified, the antibody detected three major protein bands with a molecular mass of approximately 68, 62, and 50 kDa, respectively (Fig. 3C, lanes 1 and 2). The migration distances of the 68-and 62-kDa proteins in the HeLa cells suggests that these proteins are probably the hGrb10 isoforms with or without the deletion in the PH domain.
To evaluate hGrb10 protein expression in insulin target cells, Western blot experiments were carried out on two human hepatocyte cell lines, 3T3-L1 adipocyte cells and human skeletal muscle cells. In human skeletal muscle cells, the antibody detected hGrb10␣ and a major 50-kDa protein band with unknown identity (Fig. 3D, lanes 5-7). On the other hand, only hGrb10␥ was detected in human hepatocyte cells (Fig. 3C,  lanes 3 and 4). The 50-kDa anti-hGrb10 immunoreactive protein, which was detected in the HeLa cells (Fig. 3C, lanes 1   2), human liver cells (Fig. 3C, lanes 3 and 4), and human skeletal muscle cells (Fig. 3D, lanes 5-7) could either be the degradation product of hGrb10 proteins or another isoform of the product. Because its molecular mass is significantly smaller than that of hGrb10␤ (the calculated mass of which is approximately 61 kDa), it is unlikely that this protein band is the ␤ isoform of hGrb10. Western blot of cell lysates of the differentiated 3T3-L1 adipose cells revealed several protein bands with molecular masses ranging from 48 to 65 kDa (Fig.  3C, lane 5). These proteins are probably the mouse homologues of the hGrb10 isoforms.
Interaction of hGrb10␥ with the IR in Cells-To study the interaction between hGrb10␥ and the IR in cells, we established cell lines expressing both the IR and hGrb10␥. Lysates from insulin-treated or nontreated CHO/IR, CHO/IR/hGrb10␣, or CHO/IR/hGrb10␥ cells were incubated with antibody to the ␤-subunit of the IR (29B4, a gift from R. A. Roth, Stanford University, Stanford, CA) immobilized to protein G-Sepharose beads. After separation by SDS-PAGE and transfer to a nitrocellulose membrane, the IR and its associated hGrb10 isoforms were detected with antibodies to phosphotyrosine (Fig. 4A) or to hGrb10 (Fig. 4B), respectively. A comparable amount of the autophosphorylated IR was precipitated by the anti-IR antibody from the parent CHO/IR cells and the hGrb10 transfected cells lines (Fig. 4A). Both hGrb10␣ and hGrb10␥ could be coimmunoprecipitated by the anti-IR antibody in cells expressing these proteins (Fig. 4B, lanes 4 -6). The interaction between hGrb10 isoforms was dependent on the tyrosine phosphorylation of the receptor, and no hGrb10 isoforms were coimmunoprecipitated in the absence of insulin stimulation (Fig. 4, lanes  2-3).
To further characterize the interaction between the IR and hGrb10␥, we studied the localization of the protein in the presence or absence of insulin treatment. In unstimulated CHO/IR/hGrb10␥ cells, hGrb10␥ was detected in both the cytosol and membrane (Fig. 5A). After insulin stimulation, the fraction of the protein on the membrane was significantly increased (Fig. 5B). Insulin-stimulated translocation was also observed for hGrb10␣, which lacks an intact PH domain (not shown). These results are consistent with the recent finding that Grb10 moves from cytosol to membrane fractions in the IR-overexpressing Rat1 fibroblasts after insulin stimulation (11).
hGrb10␥ Undergoes Insulin-stimulated Serine Phosphorylation-Our preliminary studies showed that insulin stimulation resulted in a gel mobility shift of hGrb10␥, suggesting that multiple forms of this protein, for instance differentially phosphorylated species, may occur. To test this hypothesis, hGrb10 isoforms in cell extracts were subjected to phosphatase treatment. Fig. 6A shows an immunoblot of hGrb10␣ and hGrb10␥ from insulin-stimulated or nonstimulated cells after treatment with PAP in the presence or absence of phosphatase inhibitors. Multiple bands of hGrb10␥, which were upward-shifted after insulin stimulation, were detected in lysates when the PAP treatment was carried out in the presence of phosphatase inhibitors (Fig. 6A, lanes 3 and 4). Treatment of the lysates from insulin-stimulated CHO/IR/hGrb10␥ cells with PAP in the absence of phosphatase inhibitors converted the broad hGrb10␥ bands to a single band with a mobility similar to that of the protein from cells not treated with insulin (Fig. 6A, lanes 7 and  8). PAP had no effect on the gel mobility of hGrb10␣ (Fig. 6A,  lanes 1, 2, 5, and 6). It should be pointed out that the presence of multiple phosphorylated forms of Grb10 was not only in cells overexpressing the protein but also in cells expressing endogenous Grb10 (9). In addition, the Grb10 gel mobility shift was also stimulated by other growth factors such as the plateletderived growth factor and the fibroblast growth factor in NIH3T3 cells expressing endogenous receptors (9). However, in HeLa cells and 3T3-L1 adipose cells, no significant insulinstimulated Grb10 mobility shift was observed (data not shown). These data suggest that the growth factor-stimulated Grb10 phosphorylation may be cell type-dependent.
To further characterize the phosphorylation of hGrb10 isoforms, in vivo phosphorylation and phosphoamino acid analysis experiments were carried out. Fig. 6B shows an autoradiograph of 32 P-labeled hGrb10␥ precipitated from insulin-treated (ϩ) and nontreated (Ϫ) CHO/IR/hGrb10␥ cells using the antibody against hGrb10. hGrb10␥ was phosphorylated under ba-sal conditions, probably due to the overexpression of the receptor and the protein. Insulin stimulation resulted in a 40% increased in the phosphorylation as quantified by PhosphorImager analysis (average of two independent experiments, Fig.  6B). This seems to be an underestimated number, as control experiments showed that our antibody bound to the hyperphosphorylated protein with a much lower affinity (data not shown). The phosphorylation was also significantly decreased for hGrb10␣ under similar conditions (data not shown). Phosphoamino acid analysis of hGrb10␥ indicates that the protein phosphorylation occurred on serine residues (Fig. 6C). However, it should be mentioned that in some of the experiments, we also observed a small but not reproducible tyrosine phosphorylation of hGrb10 by anti-phosphotyrosine immunoblot (data not shown), suggesting that a transient tyrosine phosphorylation of hGrb10 might occur in cells.
hGrb10␥ Is a Common Target for Kinases Existing in Both the MAP Kinase and PI 3-Kinase Signaling Pathways-Several enzymes, including MAP kinase and PI 3-kinase, have been shown to be activated by insulin. To test whether kinases involved in these pathways play roles in the phosphorylation of hGrb10␥, we studied hGrb10␥ phosphorylation in the presence or absence of specific inhibitors. Wortmannin is a relatively specific inhibitor for PI 3-kinase, and it inhibits the enzyme at nM concentrations (26). PD98059 binds to the inactive form of MAP kinase kinase (MEK1) and prevents its activation by upstream kinases such as c-Raf (27). Fig. 7 shows an immunoblot of hGrb10␥ in the presence or absence of these inhibitors. Three different mobility forms of hGrb10␥ termed a, b, and c could be visualized in CHO/IR/hGrb10␥ cell extracts (Fig. 7,  lane 1). Stimulation of cells with insulin increased hGrb10␥ phosphorylation, which was evident by a decrease in intensity of the fastest migrating band c and an increase in intensity of bands b and a (Fig. 7, lanes 1 and 2). Pretreatment of cells with PD98059 inhibited the basal phosphorylation of hGrb10␥ (Fig.   FIG. 4. Interaction of hGrb10␥ with the IR in cells. Lysates from insulin-treated (ϩ) or nontreated (Ϫ) CHO/IR/hGrb10␣ or CHO/IR/ hGrb10␥ cells were precipitated with a monoclonal anti-IR antibody (29B4). The hGrb10 isoforms were detected by anti-hGrb10 antibody.

FIG. 5. Fluorescence microscopy of CHO/IR/hGrb10␥ cells.
Cells were treated without (A) or with (B) 10 Ϫ8 M insulin for 15 min, fixed, and permeabilized as described under "Experimental Procedures." After blotting the cells with a polyclonal anti-hGrb10 antibody and a fluorescein-conjugated anti-rabbit IgG second antibody, the localization of hGrb10␥ was examined by fluorescence microscopy. 7, lane 3). This inhibition, however, was partially overcome after stimulating the cells with insulin (Fig. 7, lane 4). By contrast, wortmannin inhibited both the basal and the insulinstimulated phosphorylation of hGrb10␥ (Fig. 7, lane 5 and 6). Treatment of the cells with phorbol 12-myristate 13-acetate, an activator of several protein kinase C isozymes, did not increase the phosphorylation of hGrb10␥ (data not shown), suggesting that these protein kinase C isozymes are not essential for the phosphorylation of hGrb10␥. Taken together, our data show that hGrb10␥ is a potential common target for kinases participating in both the MAP kinase and the PI 3-kinase signaling pathways. DISCUSSION We have identified the PH domain-containing hGrb10␣ isoform and named it hGrb10␥. We propose a nomenclature that is consistent with other current gene family nomenclatures (Fig. 8). There are three genes located on different chromosomes in human with Grb7 on human chromosome 17q21-q22, Grb10 on human chromosome 7p12-p11, and Grb14 on human chromosome 2.
Each of these genes has an orthologous gene that has been identified in mouse, with a partial cDNA for mouse Grb14 being identified in this study as Expressed Sequence Tag clone 726559 (GenBank TM accession number AA394102). This partial cDNA encodes a part of the PH domain of the murine and other regions of the murine cDNA.
Since each of these genes has an orthologue in the murine genome, we suggest that we refer to the genes as hGrb7/10/14 or mGrb7/10/14. Within the Grb7 and Grb14 genes, no alternatively spliced cDNAs have been reported, although the potential exists for these differences. Within the Grb10 gene in humans, four different isoforms that potentially encode different proteins have been identified. The first human form is human Grb-IR, which we refer to as hGrb10␣. Two reports followed with a hGrb10␣ isoform, which has a different amino terminus and a complete PH domain (11,22), which we refer to as hGrb10␤. In this report, we describe a hGrb10␣ isoform called Grb10␥ with a complete PH domain and the same amino terminus observed in hGrb10␣. Another isoform derived from a randomly cloned cDNA (clone KIA0207, GenBank TM accession number D86962) has a third alternative amino terminus and a complete PH domain and is termed hGrb10␦. The mouse Grb10 protein originally identified by screening a NIH3T3 cell cDNA expression library using a radiolabeled tyrosine-phosphorylated epidermal growth factor receptor contains an 80-amino acid region that is not observed in human cDNA clones (9). An isoform of this protein has been reported very recently that lacks the first 25 amino acids of this 80-amino acid insert (28).
The nomenclature for isoforms across species is a difficult issue in general but even more difficult for the Grb10 isoforms. The murine isoform differs in two substantial ways from the human isoforms. 1) It contains an amino terminus different from any of the four hGrb10 isoforms and 2) it contains a region of 55 or 80 amino acids inserted near the PH domain, which has not been observed in humans. The functional consequences of these differences, if any, have not been determined. For these reasons, at this time we propose a distinct nomenclature for the human and murine protein isoforms. As additional murine and human isoforms are reported and the functional consequences of these splicing variations are determined, a consistent nomenclature across species may be developed.
The data presented in this paper have shown that both hGrb10␣ and hGrb10␥ bind with high affinity to the IR in mammalian cells, and the interaction is dependent on receptor tyrosine phosphorylation (Fig. 4). In addition, both isoforms undergo insulin-stimulated translocation from the cytosol to the plasmic membrane ( Fig. 5 and data not shown). These results suggest that hGrb10␣ and hGrb10␥ are potential signaling molecules in the IR signaling process. However, there are also notable differences between these two isoforms. First, hGrb10␥ contains a 46-amino acid sequence contained in the PH domain, which is absent in hGrb10␣. The difference in this  1-4) or absence (lanes 5-8) of phosphatase inhibitors. hGrb10␣ or hGrb10␥ were resolved by SDS-PAGE, transferred to a nitrocellulose membrane, and detected with a polyclonal antibody to hGrb10 as described under "Experimental Procedures." B, immunoprecipitation of 32 P-labeled hGrb10␥. CHO/IR/hGrb10␥ cells were grown in 10-cm plates and labeled with 32 P orthophosphate. After treatment with (ϩ) or without (Ϫ) insulin (10 Ϫ8 M) for 15 min, cells were lysed, and hGrb10␥ was immunoprecipitated with the polyclonal antibody against hGrb10, resolved by SDS-PAGE, blotted to a membrane, and visualized by autoradiography. C, phosphoamino acid analysis of the in vivo labeled hGrb10␥. The 32 P-labeled hGrb10␥ bands were excised from the membrane, hydrolyzed, and analyzed by two-dimensional thin-layer electrophoresis and autoradiography. M.W., molecular mass. functional domain may thus suggest a different role for hGrb10 isoforms in receptor tyrosine kinase signaling. The natural occurrence of two isoforms of hGrb10 that differ in the PH domain may thus provide a mechanism for the regulation of insulin signaling.
The second difference between hGrb10␣ and hGrb10␥ is their cell expression. We have previously shown that although both the ⌬PH and the PHϩ isoforms of hGrb10 mRNAs are expressed in insulin target cells such as skeletal muscle and fat cells, the ⌬PH isoform mRNA is more abundantly expressed than that of the PHϩ isoform (7). In agreement with these studues, our present Western blot studies reveal that the ⌬PH isoform hGrb10␣ protein is expressed in human skeletal muscle cells (Fig. 3D). The failure to detect hGrb10␥ is probably due to a lower expression of this isoform in the cells. On the other hand, the major isoform in human breast cancer cell lines and human liver cells is the PH domain-containing isoform hGrb10␥ (Fig. 3, A-C). These data suggest that hGrb10 isoforms vary among different cells or tissues. The different expression of these isoforms could reflect the involvement of these proteins in specific signal transduction or regulation processes, which could depend on functional differences between these isoforms. The observation that only hGrb10␥, but not hGrb10␣, undergoes insulin-stimulated serine phosphorylation provides additional evidence that the two isoforms may function differently in the signaling process. Since protein phosphorylation has been shown to be involved in a variety of biological events such as signal transduction and regulation, the phosphorylation of hGrb10␥ may be a mechanism that regulates the protein activity in the signaling pathway. For example, phosphorylation may affect the binding of hGrb10 to the IR or to other interacting proteins to amplify or regulate the IR signal. The observation that hGrb10␥ phosphorylation is stimulated by insulin (Fig. 6) also suggests a potential role for the protein as a component in IR signal transduction. Another possible role for hGrb10 isoforms in the signaling process may be to provide a scaffolding function. As these isoforms contain several functional domains including the SH2 domain, the PH domain, and an N-terminal proline-rich sequence (P(A/S)IPNPFPEL), which is conserved among several proteins including Grb7 (8), Grb10 (9), and hGrb14 (10), they are capable of binding multiple protein molecules important for the signal transduction process. Recent studies from several laboratories have shown Grb10 binds not only to the insulin receptor but also to other receptors such as the insulin-like growth factor 1R (22,29,30), the ELK receptor (31), and Ret (32). These data suggest that the protein may have a more general role in receptor tyrosine kinase signal transduction and regulation.
The inhibition of hGrb10␥ phosphorylation by wortmannin and PD98059 suggests that the protein is a potential target of kinase(s) in both the MAP kinase and PI 3-kinase pathways. Whether these enzymes or enzymes downstream of these proteins catalyze the phosphorylation is currently unknown. Sequence analysis of hGrb10 isoforms reveals that the proteins contain three potential MAP kinase phosphorylation sites PX(S/T)P (33), suggesting that MAP kinase may directly phosphorylate the protein. It is also interesting to note that insulin can partially overcome the inhibition of hGrb10␥ phosphorylation caused by PD98059 but not that caused by wortmannin. These data suggest that there may be multiple phosphorylation sites on hGrb10␥ that are targets of different kinases. Identification and characterization of these phosphorylation sites and the hGrb10 kinase(s) should provide further insight into the role of Grb10 in pathways initiated by the IR and possibly other receptor tyrosine kinases. FIG. 8. Classification of Grb10 isoforms. On the left side, each of the Grb7/ Grb10/Grb14 genes were analyzed using maximum parsimony within PAUP. On the right is a table describing the proposed nomenclature, the functional differences within the human Grb10 isoforms, and the corresponding chromosome locations of each of the human genes. *These names are laboratory names for murine isoforms (28).