Insulin-like growth factor-I receptor and insulin receptor association with a Src homology-2 domain-containing putative adapter.

Insulin receptor (IR) and the related insulin-like growth factor-I (IGF-I) receptor (IGF-IR) mediate a variety of metabolic and mitogenic cellular responses, some of which may involve unidentified receptor targets. A Src homology-2 (SH2) domain-coding region of a mouse protein was cloned based on its interaction with IR. It was designated mSH2-B based on its high similarity to an earlier reported rat sequence SH2-B. A role of mSH2-B in IGF-I and insulin action was suggested by the interaction of the SH2 domain with activated IGF-IR and IR catalytic fragments but not with an inactive IR catalytic fragment in the yeast two-hybrid system in vivo and by the hormone-dependent association of a glutathione S-transferase (GST) SH2 domain fusion protein of mSH2-B with both receptors in cell extracts. A comparison of IGF-IR and IR mutants lacking individual Tyr autophosphorylation sites for association with GST mSH2-B showed that homologous juxtamembrane (IR960/IGF-IR950) and C-terminal (IR1322/IGF-IR1316) receptor motifs were required. Synthetic phosphopeptides representing IR960 and IR1322 competed for GST mSH2-B binding to the receptor, suggesting that both motifs participate in the association with mSH2-B. Antibodies raised against GST mSH2-B identified a cellular protein of 92 kDa that was not found to be phosphorylated on Tyr. It co-immunoprecipitated with IGF-IR or IR, which was strictly dependent on receptor activation. IR and IGF-IR Tyr phosphorylation motifs were not identified in the complete SH2-B primary structure, suggesting that it may participate as an adapter rather than a substrate in the IGF-I and insulin signaling pathways.

Insulin-like growth factor-I (IGF-I) 1 has been implicated particularly in mitogenic functions and of insulin more in metabolic cellular functions (1,2). The insulin receptor (IR) and the closely related IGF-I receptor (IGF-IR) have served as a model system for the elucidation of receptor Tyr kinase-mediated signaling mechanisms (3). The nature for the physiologic differences between the IGF-I and insulin signals is unclear and may involve cellular targets such as Grb10 and pp120 that exhibit a preference for one of the receptors (4,5). Major cellular targets that are shared by both receptors include the insulin receptor substrates IRS-1, IRS-2, IRS-3, and Shc, most of which also play established roles in other signaling pathways (6 -10).
Some of the metabolic responses to insulin must be mediated by mechanisms that parallel PI 3Ј-kinase, which suggests a role of an undefined alternative IR signaling pathway (11). A role of IRS-1 and PI 3Ј-kinase has been proposed in the insulinstimulated glucose uptake by GLUT4 glucose transporter translocation (12,13); however, an IRS-1-and PI 3Ј-kinaseindependent, unidentified pathway has been described in the insulin stimulation of glucose uptake in independent studies (14,15). Insulin activates glycogen synthase in CHO cells by an unidentified, Ras-and PI 3Ј-kinase-independent (wortmanninindependent) mechanism (16). Despite normal Ras binding to Raf-1, truncation of the IR C terminus impairs Raf-1, mitogenactivated protein kinase kinase, and MAPK activities, glucose transport, glycogen synthesis, PI 3Ј-kinase, and phosphoprotein phosphatase-1 activities, whereas mitogenic responses remain largely unimpaired (17,18). Overexpression of IRS-1 was shown to restore the mitogenic response of a defective IR (Y960A, which fails to activate IRS-1) including activation of PI 3Ј-kinase but restored only partial glycogen synthesis and failed to restore MAPK activation (19). IR mutants in the tyrosine kinase region (R1174Q and L1178P) suggest a role of unidentified signaling mediators in the impaired glycogen synthesis, DNA synthesis, and MAPK activation in response to insulin in CHO cells, because IRS-1 appears to be normally activated (20). Mitogen-activated protein kinase and PI 3Јkinase signaling pathways are not sufficient for IGF-I-induced mitogenesis and tumorigenesis, suggesting a role of unidentified signaling mediators in these important IGF-I actions (21). Such mediators may act in parallel to IRS-1 and IRS-2 and may resemble other adapters known to associate with other receptor tyrosine kinases (22). In this study we have characterized the association of a newly identified protein target of 92 kDa, designated mSH2-B, with IGF-IR and IR that is strictly dependent on ligand stimulation for both receptors and was found in a yeast two-hybrid screen (23). We find that the SH2 domain of mSH2-B as well as the juxtamembrane regions and the C-terminal regions of both receptors are required for the association and that synthetic peptides representing these receptor motifs compete with receptor binding to mSH2-B. These characteristics suggest a role of mSH2-B as a Pro-rich and SH2 domain-containing mediator in IGF-I and insulin action.
Yeast Two-hybrid Screen and Interaction Specificity-Yeast two-hybrid plasmid pGBT9 (25) containing the C terminus of the IR ␤ subunit (26) was used as a bait to screen a 9.5-10.5-day post conception mouse embryo cDNA library in plasmid pVP16 (27). Library (prey) plasmids were isolated from His prototroph and ␤-galactosidase-positive yeast colonies (24). Plasmids were introduced into Escherichia coli strain DH5␣, sequenced, and re-introduced into Saccharomyces cerevisiae strain L40. Mating with strains (AMR70) carrying various test baits followed by the ␤-galactosidase color assay evaluated the specificity of the underlying bait-prey interactions (27).
GST Fusion Protein-The cloned cDNA insert encoding the SH2 domain of mSH2-B and surrounding coding sequences was released from the pVP16 library plasmid at its NotI cloning site and inserted into the NotI cloning site of pGEX-4T-1 (Pharmacia Biotech Inc.). GST mSH2-B fusion and control GST protein were expressed in E. coli strain DH5␣ and purified on a glutathione-agarose column (Pharmacia), eluted in 10 mM reduced glutathione in 50 mM Tris-HCl, pH 8.0, and stored after addition of 10 mM dithiothreitol, 1 mM EDTA as described by the manufacturer.
Cell Culture, Immunoprecipitation, and Immunoblotting-Subconfluent wild-type or mutant IR overexpressing CHO fibroblasts (kindly provided by Drs. Yousuke Ebina, Takashi Kadowaki, and Morris F. White) or wild-type or mutant IGF-IR overexpressing R Ϫ fibroblasts (kindly provided by Dr. Renato Baserga) were employed in various experiments (28 -33). Cells were incubated in serum-free medium for 16 h and stimulated with 100 ng/ml insulin or 100 ng/ml IGF-I for 15 min. Cells were washed twice with phosphate-buffered saline and harvested in ice-cold lysis buffer containing 50 mM HEPES, pH 7.4, 1% Triton X-100, 10% glycerol, 137 mM NaCl, 2 mM EDTA, 10 mM NaF, 100 mM Na 3 VO 4 , 10 mM sodium pyrophosphate, 10 g/ml leupeptin, 10 g/ml aprotinin, and 1 mM phenylmethylsulfonyl fluoride. Proteins were directly subjected to SDS-PAGE or were first mixed with GST mSH2-B fusion protein or antibodies directed against mSH2-B, IR, or IGF-IR and co-precipitated with glutathione-agarose or protein A-Sepharose beads, respectively. Precipitates were washed with lysis buffer, separated by SDS-PAGE, and analyzed by immunoblotting with specific antibodies using the ECL (Amersham Corp.) detection system (24).

Two-hybrid Analysis of IGF-IR and IR Association with the SH2 Domain of mSH2-B in Vivo-
We employed the cytoplasmic domains of the IR and IGF-IR as baits in a yeast twohybrid screen to identify unknown signaling mediators in both receptor pathways from mouse libraries (24,4). Most identified sequences represented SH2 domains of known IR and IGF-IR targets ( Fig. 1) including the C-terminal SH2 domain of the regulatory subunit p85 of PI 3Ј-kinase, the SH2 domain of the signaling mediator Grb10/IR, and the SH2 domain of the transforming protein Vav (4, 34 -39). In addition, we identified an SH2 domain-encoding sequence with the two-hybrid system (23) that was scanned against the sequence data base. It was designated mSH2-B because one sequence with high similarity was identified that had earlier been reported in rat as SH2-B (40). SH2-B (GenBank™ data base accession number U57391) had been found based on its association with the Fc epsilon RI receptor in a modified yeast two-hybrid screen, and its complete protein-coding region was isolated and shown to contain several Pro-rich sequences (40). The C-terminal fragment encoding the SH2 domain and adjacent sequences of mSH2-B that was identified in our two-hybrid screen displayed only few nucleotide replacements C1900T, C1918T, G1945A, C1981T, A1982G, A2023G, T2065C, T2077C, T2098A, T2257C, and A2261G, which resulted in a single aa change S547G when compared with the rat SH2-B sequence, which is represented by the first letter and its position number (40). These changes are likely explained by species-specific differences. A second related human sequence (GSDB, DDBJ, EMBL, and NCBI data base accession number AB000520) was identified that encodes a putative adapter designated APS that appears to represent a related but distinct gene based on a considerable number of aa replacements when compared with mSH2-B and rat SH2-B (41).
When introduced into yeast strains carrying different test bait plasmids, all identified sequences including mSH2-B specifically associated with the activated IGF-IR and IR cytoplasmic domains but not with control baits including the Ser/Thr protein kinase C␦, the insulin receptor substrate IRS-1, and the commonly used test protein lamin C (Fig. 1). The kinase-inactivated IR mutant K1018A scored negative, demonstrating the requirement of an activated receptor kinase for any of the SH2 domain associations (Fig. 1).
Biochemical Analysis of IGF-IR and IR Association with the SH2 Domain of mSH2-B in Vitro-We investigated this association biochemically by testing a GST mSH2-B fusion protein for its association with IGF-IR and IR in cell extracts. A GST mSH2-B fusion construct was prepared, and the protein was expressed in and partially purified from E. coli and mixed with detergent extracts of IGF-IR or IR overexpressing cells. IGF-IR and IR co-precipitated with glutathione-Sepharose beads directed against GST as detected in immunoblots with receptor-specific antibodies (Fig. 2). The association was strictly dependent on receptor activation by the respective ligand. Similarly, the association of the SH2 domain of mSH2-B in the two-hybrid screen had been observed with an active but not with an inactivated IR cytoplasmic domain under in vivo conditions in the yeast nucleus. If the observed associations were indirect the same adaptor protein would need to be present and active in the yeast nucleus and in mammalian detergent cell extracts under very different experimental conditions. Because this is unlikely our data suggest that mSH2-B binds directly to the receptors without the involvement of an intermediate protein.   (Fig. 3, A and B). To test whether these motifs are in fact able to bind to mSH2-B, synthetic phosphopeptides representing these regions were tested for competition with normal IR binding to GST mSH2-B.
In accordance with the receptor mutant analysis, peptides representing Tyr 960 and Tyr 1322 abolished IR binding in contrast to peptides representing other receptor regions (Fig. 3C). These data suggest that homologous sites of both receptors at IGF-IR Tyr 950 /IR Tyr 960 in the juxtamembrane region and at IGF-IR Tyr 1316 /IR Tyr 1322 at the C terminus are involved in the interaction with the SH2 domain of mSH2-B.

Ligand-dependent Association of Cellular SH2-B with IGF-IR and IR-
To learn more about the cellular SH2-B protein, a rabbit antiserum was raised against GST mSH2-B that identified a 92-kDa protein in immunoblots of CHO and mouse fibroblast cell extracts (Fig. 4, A and B). To evaluate the association of this protein with the IR and IGF-IR, fibroblasts overexpressing either receptor were ligand stimulated, and detergent cell extracts were immunoprecipitated with receptorspecific antibodies followed by SDS-PAGE and immunoblotting with mSH2-B antiserum. A 92-kDa protein of identical migration was specifically co-precipitated with either receptor, strictly dependent on ligand activation (Fig. 4, A and B). When cell extracts were precipitated with mSH2-B antiserum in complementary experiments, IR (Fig. 4A) and IGF-IR (Fig. 4B) were co-precipitated from the respective cell lysates in a liganddependent fashion. Our data indicate that mSH2-B associates with IR and IGF-IR upon receptor activation in three independent approaches based on 1) two-hybrid constructs, 2) GST mSH2-B fusion protein, and 3) authentic cellular SH2-B. These findings strongly suggest that mSH2-B plays an as yet unidentified role in insulin and IGF-I action as a direct receptor target.
The molecular size of the detected 92-kDa protein is in line with the predicted size for an expression product of 768 aa for the complete rat SH2-B coding sequence based on the observation that proteins frequently migrate more slowly in SDS-PAGE than predicted from their primary structure (40). We cannot exclude that other forms of SH2-B may exist that have not been detected. The 92-kDa protein was consistently observed in several cell lines, specifically recognized by mSH2-B antiserum, and co-precipitated with IGF-IR and IR in a strictly ligand-dependent fashion. It represents a promising candidate for a Pro-rich and SH2 domain-containing mediator involved in the IGF-IR and IR signaling pathways that appears to be shared by both receptors similar to most known IGF-IR and IR targets.
The cloned C-terminal SH2 domain coding fragment of mouse mSH2-B displays high sequence homology to the published rat sequence, designated SH2-B (40), suggesting that the observed changes are explained by species-specific differences. In our efforts to clone the complete mouse mSH2-B coding sequence we have isolated a number of sequence variants that suggest the existence of a family of SH2-B-related mediators including APS, some of which may be generated by differential splicing (41,42).
Because many IR and IGF-IR targets including IRS-1 (6), IRS-2 (8), and Shc (7) are receptor substrates that are phosphorylated on Tyr by the ligand-activated receptor Tyr kinase, we began to investigate whether mSH2-B may be a new member of this group. The analysis of its primary structure did not reveal any consensus target sites YMXM or YXXM for the IGF-IR or IR Tyr kinases (6), which strongly suggested against a role of mSH2-B as an IGF-IR and IR substrate. This was confirmed in biochemical experiments to evaluate whether SH2-B was phosphorylated on Tyr upon IGF-I or insulin stimulation and recognized by phosphotyrosine-specific antibodies in immunoblots. With these antibodies we consistently failed to detect SH2-B directly in cell lysates or after prior immunoprecipitation with mSH2-B antiserum. Similarly, immunoprecipitation with phosphotyrosine antibodies failed to bring down SH2-B based on subsequent immunoblots with mSH2-B antiserum. IR and IGF-IR autophosphorylation in these experiments confirmed appropriate experimental conditions and ligand stimulation (not shown). Given that an insulinstimulated IR-associated Ser kinase has been implicated in insulin action (43) we also addressed the putative phosphorylation of SH2-B on Ser or Thr. However, preliminary experiments failed to identify SH2-B as a phosphoprotein after 32 P labeling of various fibroblasts (not shown), independent of ligand stimulation (44). We consider that SH2-B may not be phosphorylated in response to IGF-I or insulin stimulation or that a weaker SH2-B phosphorylation may not have been detected due to limitations in the sensitivity of the assay (42). mSH2-B may function like other signaling adapter proteins such as Grb2 by forming complexes with other proteins and by directing their intracellular localization (22). It may undergo phosphorylation in response to other stimuli that remain to be defined (42).
Possible functions of mSH2-B may include a positive role in any of the pleiotropic responses to IGF-I and insulin. Alternatively, mSH2-B may be a negative regulator by counter balancing IRS-1, IRS-2, IRS-3, Shc-mediated, or other signals. There are few precedents for such a function, which may, however, provide important regulatory controls in analogy to the opposing actions of Tyr kinases and phosphatases (45). A negative regulatory role in insulin action has recently been proposed for the IR target Grb10/IR as one example for such a mechanism (37). Given that other IR and IGF-IR targets have been mapped to associate with the IR Tyr 960 /IGF-IR Tyr 950 motif including GTPase-activating protein and Grb10 or with the IR Tyr 1322 / IGF-IR Tyr 1316 motif including p85 PI 3Ј-kinase, Syp, SHP-2, and Grb10, mSH2-B is expected to compete with some of these mediators for receptor association and activation (24,35,39,46,47). Competition for a limited pool of a signaling mediator has been described in insulin action (48). This may contribute to the cellular role of mSH2-B in insulin and IGF-I action.