Association of Phosphorylated Insulin-like Growth Factor-I Receptor with the SH2 Domains of Phosphatidylinositol 3-Kinase p85*

Insulin-like growth factor-I (IGF-I) stimulates the production of 3-inositides and markedly increases the phosphatidylinositol 3-kinase activity that is immunoprecipitated by anti-phosphotyrosine antibodies, a portion of which is also associated with the IGF-I receptor. In this study, recombinant p86, the regulatory subunit of phosphatidylinositol 3-kinase, and fusion proteins containing various subdomains were used to investigate the association of p86 with the IGF-I receptor and to demonstrate that p86 is a direct in vitro substrate of the IGF-I receptor kinase. IGF-I was lysate, phosphatidylinositol protein of and molecular of the subunits of 3-kinase. p86 not SH3 breakpoint A fusion protein the SH2 domains of auto- protein from Specificity of the SH2-mediated Inhibition-To After the associated Ptdlns kinase specificity of the SH2-containing molecule for the inhibition activity assayed, Ptdlns(3)P of the association between ptdIns 3-kinase activity and the IGF-I receptor, SH2-containing fusion proteins from GAP

Insulin-like growth factor-I (IGF-I) stimulates the production of 3-inositides and markedly increases the phosphatidylinositol 3-kinase activity that is immunoprecipitated by anti-phosphotyrosine antibodies, a portion of which is also associated with the IGF-I receptor. In this study, recombinant p86, the regulatory subunit of phosphatidylinositol 3-kinase, and fusion proteins containing various subdomains were used to investigate the association of p86 with the IGF-I receptor and to demonstrate that p86 is a direct in vitro substrate of the IGF-I receptor kinase.
Solubilized IGF-I receptor was immobilized on antireceptor antibody-agarose beads. Following in vitro receptor phosphorylation and incubation with cell lysate, immobilized receptor became associated with phosphatidylinositol 3-kinase activity and with protein bands with molecular masses of 86 and 110 kDa, which correspond to the known molecular masses of the subunits of phosphatidylinositol 3-kinase. These associations were inhibited by the addition of recombinant intact p86 or SH2-containing fusion proteins, but not by fusion proteins containing its SH3 domain or breakpoint cluster homology region. A fusion protein containing the SH2 domains of Ras GTPase-activating protein also inhibited the association of phosphatidylinositol 3-kinase activity with immobilized IGF-I receptor, although less effectively than p85, whereas a similar construct containing the SH2 domain of pp60'" was without effect. When immobilized phosphorylated IGF-I receptor was incubated with intact p86 or the SH2-containing fusion proteins, it became associated with and phosphorylated these proteins.
These results demonstrate that at least in vitro, a tight association occurs between phosphorylated IGF-I receptor and phosphatidylinositol 3-kinase, that the region of phosphatidylinositol 3-kinase that contains its SH2 domains is directly involved in this association, and that this region is a direct substrate for IGF-I receptor tyrosine kinase. Furthermore, these results suggest that Ras GTPase-activating protein can also interact with the IGF-I receptor and that different SH2 domain-containing proteins interact with the IGF-I receptor with widely differing affinities.
* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Phosphatidylinositol 3-kinase (PtdIns 3-kinase)' is thought to be an important component of the signaling pathway of several receptor and nonreceptor tyrosine kinases (1). The evidence for this, however, is circumstantial, and the functions of its products are unknown. PtdIns 3-kinase is a heterodimer composed of 85-and 110-kDa subunits (2)(3)(4). The 85-kDa subunit, p85, has been cloned (5)(6)(7). Its amino-terminal region bears considerable homology with the SH3 domain of pp60"". Its midregion has weak homology with BCR, the break point cluster involved in c-ab1 translocations (7,8). Its carboxylterminal half contains two regions with high homology to the SH2 domain of p~6 0 *~ (See Fig. lA). SH2 domains are found in a variety of proteins including the viral oncogene, v-crk, nonreceptor tyrosine kinases, and proteins important in signaling pathways including phospholipase C-y and Ras GTPase-activating protein (GAP). SH2 domains are known to bind to phosphotyrosine-containing proteins and therefore may play a role in the formation of multisubunit signaling complexes which form in response to tyrosine kinase activation and regulate intracellular signaling pathways (1,(9)(10)(11)(12)(13)(14).
Expression of recombinant p85 in different systems shows that this protein is not the PtdIns 3-kinase catalytic subunit (7). Instead, it appears to be a regulatory subunit and/or acts as a bridge between the tyrosine kinase and the catalytic subunit (2, 6). Recombinant p85 is able to bind to middle t-pp60"'" complex and to platelet-derived growth factor (PDGF) receptor and serves as a substrate for tyrosine phosphorylation (6,7). PDGF receptor mutant in the kinase insert domain is no longer able to associate to recombinant p85, and a tyrosine-phosphorylated peptide from the same kinase insert region is able to inhibit the binding of p85 to the receptor (6). Polyclonal antibodies against recombinant p85 immunoprecipitate PtdIns 3-kinase activity from bovine brain and p85 and p l l 0 were found in this immunoprecipitate (7). These data are consistent with the hypothesis that p85 functions as a regulatory subunit through which the catalytic subunit of PtdIns 3-kinase is modulated by tyrosine kinases.
IGF-I and insulin, which have as receptors closely related tyrosine kinases, stimulate the production of 3-inositides and cause a marked increase in the amount of PtdIns 3-kinase activity that can be immunoprecipitated by anti-phosphotyrosine antibodies (15)(16)(17)(18)(19)(20). There are several possible explanations for the latter finding: PtdIns 3-kinase may be precipitated by anti-phosphotyrosine antibodies because it is di-The abbreviations used are: PtdIns, phosphatidylinositol; BCR, breakpoint cluster region; IGF-I, insulin-like growth factor I; PDGF, platelet-derived growth factor; GAP, Ras GTPase-activating protein; PIP, phosphatidylinositol monophosphate; PtdIns(3)P, phosphati-dylinositol3-monophosphate; a-Ptyr, anti-phosphotyrosine antibody; Hepes, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid; SDS, sodium dodecyl sulfate; TLC, thin layer chromatography; HPLC, high performance liquid chromatography. rectly or indirectly tyrosine-phosphorylated in response to IGF-I or insulin. Currently, there is no direct evidence for this. PtdIns 3-kinase may be precipitated by anti-phosphotyrosine antibodies because it is associated with tyrosine phosphorylated IGF-I receptors. Following IGF-I stimulation, PtdIns 3-kinase activity is present in IGF-I receptor immunoprecipitates (17,18). However, this accounts for only a small percent of the PtdIns 3-kinase activity that is precipitated by anti-phosphotyrosine antibody (17,18). PtdIns 3kinase may be precipitated by anti-phosphotyrosine antibodies, because it is associated with other proteins that are phosphorylated in response to IGF-I. Following insulin stimulation, PtdIns 3-kinase activity is present in immunoprecipitates of IRS-1, a major substrate of insulin receptor kinase in most cells (21). IRS-1 is also a major substrate of IGF-I receptor kinase. Both IGF-I receptors and IRS-1 have tyrosine phosphorylation sites that are homologous to the site on middle T to which PtdIns 3-kinase is known to bind (1,21).
In order to further investigate the role of PtdIns 3-kinase in the IGF-I receptor-signaling pathway, we cloned and expressed full-length p85 and different portions of this molecule. We describe here that exogenous p85 inhibited the association of PtdIns 3-kinase activity with IGF-I receptor and that the carboxyl-terminal portion containing both SH2 domains was responsible for this inhibition. We also showed that an SH2containing fragment of GAP partially inhibited this association, whereas an SH2-containing fragment of pp60"" did not.

EXPERIMENTAL PROCEDURES
Materials-LISN C4 cells, a mouse fibroblast cell line that overexpresses human IGF-I receptors (22), were kindly provided by Michael Kaleko. The anti-phosphotyrosine monoclonal antibody (a-Ptyr) was purchased from Upstate Biotechnology, Inc. a-IR-3 is an anti-IGF-I receptor monoclonal antibody (23). Phosphatidylinositol and reduced glutathione were from Sigma. [-p3'P]ATP (30 Ci/mmol) was purchased from Du Pont-New England Nuclear. Alkaline phosphatase-linked goat anti-rabbit IgG was purchased from Jackson Immunoresearch. The pGEX2T vector and glutathione-Sepharose 4B columns were purchased from Pharmacia LKB Biotechnology Inc.
Plusmid Constructions-Cloning of the different constructs was performed by RNA-polymerase chain reaction, using RNA as template and different sets of oligonucleotides. The reactions were performed using the RNA-polymerase chain reaction kit from Perkin-Elmer Cetus following the manufacturer's recommended procedure. The amplified products were subcloned in the pGEX2T vector, using the appropriate restriction enzymes (underlined below). All the fragments were sequenced using Sequenase (United Stated Biochemical Corp.) according to manufacturer's protocols. Total human brain RNA, total human placenta RNA, and total RNA from mouse TG1 neuroblastoma cell line were used as templates for p85 constructs, SH2 of GAP, and SH2 of pp60"", respectively. Expression and Purification of Glutathione S-Transferase Fusion Protein-Glutathione S-transferase fusion protein expression was induced as described previously (24) with minor modifications. Overnight cultures grown in Luria broth containing ampicillin (0.1 mg/ ml) were diluted 1:lO in 500 ml of fresh medium and grown for 2 h at 37 "C. Isopropyl-1-thio-(3-D-galactopyranoside was then added to 0.5 mM, and cultures were grown for approximately 1 h. Cells were pelleted and resuspended with 10 ml of cold phosphate-buffered saline containing 50 mM EDTA, 10 pg/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride, and 1% aprotinin. The cell suspension was lysed by sonication, and Triton X-100 was added to 1%. The lysate was centrifuged at 10,000 X g for 10 min, and the supernatant was applied on a glutathione-Sepharose 4B column. After adsorption, the column was washed with 10 column volumes of phosphate-buffered saline. The glutathione S-transferase fusion proteins were eluted with 50 mM Tris-HC1 (pH 8.0) plus 10 mM reduced glutathione.
Cell Culture-LISN C4 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 5% bovine calf serum. The cells were washed once with 10 ml of Dulbecco's modified Eagle's medium and then detached by incubation in 10 ml of phosphate-buffered saline, 1 mM EDTA. The detached cells were collected by centrifugation and lysed in lysis buffer composed of 20 mM Tris (pH 7.4), 150 mM NaC1,5 mM EDTA, 1% (v/v) Nonidet P-40, 2 mM Na3V04, 0.1 mM phenylmethylsulfonyl fluoride, 50 pg leupeptin/ml for 30 min at 4 "C. The lysed cells were centrifuged at 10,000 X g for 10 min, and supernatants (lysates) were removed and saved.
Preparation of a-IR-3-Agarose Beads-a-IR-3 was coupled to Affi-Gel 15 at a concentration of 1 mg of antibody/ml of gel for 4 h at 4 "C in 0.1 M Hepes (pH 7.5). The resin was washed extensively with buffer, blocked with 0.1 M ethanolamine for 2 h at 22 "C, and extensively washed with 0.1 M Hepes (pH 7.5) before final suspension in 0.1 M Hepes (pH 7.5), 0.025% NaN3. The efficiency of immunoprecipitation of labeled IGF-I receptors by this gel-coupled antibody was judged to be identical to that of an equivalent amount of uncoupled a-IR-3 (data not shown).
Adsorption of IGF-I Receptor to a-IR-3-Agarose Beads and in Vitro Phosphorylation-Lysates (0.5 ml) were incubated with a-IR-3 coupled to Affi-Gel 15 for 2 h at 4 "C. The beads with adsorbed receptor were collected by centrifugation, and the supernatant was saved for further experiment. The beads were washed three times with lysis buffer and twice with 10 mM Tris (pH 7.4), 0.1 M NaCl, 1 mM EDTA (TNE). The beads were resuspended with TNE containing 50 mM ATP and 20 mM MgClz and incubated at 25 "C for 15 min. They were then washed three times with TNE and resuspended with TNE. This suspension was used either for further incubation with the saved lysates or for SDS-gel electrophoresis as described.
Binding of PtdIm 3-Kinase Activity to IGF-I Receptor-The protein concentration of the saved lysate (IGF-I receptor-depleted lysate) was determined by the method of Bradford (25), and lysates (0.5 ml) containing approximately 700 pg of protein were incubated with immobilized phosphorylated IGF-I receptor in the presence or absence of the fusion protein of the different part of p85 or a SH2containing fusion protein from GAP or pp6OoK for 2 h at 4 "C. After incubation, the immunoprecipitated IGF-I receptors were washed three times with lysis buffer, twice with 0.5 M LiCl, 0.1 M Tris (pH 7.6), and twice with TNE and resuspended in TNE. The final immunoprecipitates were used for PtdIns kinase assay.
PtdIns Kinase Assay-After indicated incubations, washed a-IR-3-agarose beads containing adsorbed proteins were incubated with 400 pg PtdIns liposome/ml, 20 mM MgC12,lO pCi of [y3*P]ATP at a final concentration of 50 p~ ATP in a volume of 0.025 ml for 15 min at 25 "C. The reaction was terminated by the addition of 0.1 ml cold TNE, and half of the solution was transferred to an Eppendorf tube containing 0.02 ml of 8 N HCl and 0.16 ml chloroform/methanol (1:l). Phospholipids were extracted, washed once with 0.1 ml of 1 M HCl/methanol (1:l) and separated by thin layer chromatography (TLC) as described previously (26). The identity of the PtdIns phosphate detected on TLC was confirmed by lipid extraction, deacylation, and analysis on high performance liquid chromatography (HPLC) as described previously (27).
Gel Electrophoresis-Purified p85 proteins, phosphorylated or un-phosphorylated IGF-I receptor immunoprecipitates, or the half of the incubation from the Ptdlns kinase assay were mixed with SDS sohbilizing solution consisting of 0.125 M Tris (pH 6.8), 0.1% SDS, 10% glycerol, 100 mM dithiothreitol and heated at 95 'C for 5 rnin. T h e solubilized proteins were subjected to gel electrophoresis in the presence of SDS on 10-20% gradient or 7.5% homogeneous polyacrylamide gels as described previously (28).
Imrnunofdotting-In some experiments, separated proteins were transferred electrophoretically to nitrocellulose blots as described previously (29). After blocking the nitrocellulose for 4 h, the blot was incubated with a-Ptyr at a 1:1000 dilution for 4-24 h a t 4 "C. The blot was washed extensively, and immunoreactive proteins were visualized by incubation with goat anti-mouse antibody coupled to alkaline phosphatase as described previously (29).

RESULTS
Expression and Purification of p85 Fusion Proteins-To investigate the association between the IGF-I receptor and PtdIns 3-kinase activity, different regions of p85, a subunit of PtdIns 3-kinase, were expressed as fusion proteins in E. coli (Fig. IA). These regions were: an amino-terminal fragment containing the SH3 region, a carboxyl-terminal fragment containing two SH2 regions, and a middle fragment shown previously to have homology with BCR (7, 8). All sequences were fused downstream of the glutathione S-transferase gene in the pGEX2T expression vector. The glutathione S-transferase fusion proteins were expressed to high levels in bacteria and were rapidly purified in a single step by affinity chromatography on glutathione-Sepharose (Fig. 1B).
Ptdlns 3-Kinase Activity Is Associated with the Autophosphorylated ZGF-I Receptor-IGF-I receptor was adsorbed to a-IR-3-agarose beads by incubating the beads with LISN C4 cell lysate. The washed beads were then incubated in the absence or presence of ATP, and phosphorylated receptor was detected by anti-phosphotyrosine Western blots. As shown in Fig. 2A, following incubation with ATP there are prominent phosphotyrosine-containing bands with molecular weights of 95,000 and 180,000, which correspond to the autophosphorylated 0 subunit of the IGF-I receptor and its unprocessed precursor. Thus IGF-I receptor was adsorbed to a-IR-3-agarose beads, and in the basal state the majority of the receptor was unphosphorylated, but became autophosphorylated following incubation with ATP.
An in vitro reconstitution system was used to study the association between PtdIns 3-kinase activity and the IGF-I receptor. Adsorbed IGF-I receptor, treated as described above, was incubated with the IGF-I receptor-depleted LISN C4 cell lysate, and the resulting immunocomplex was assayed for PtdIns kinase activity. PtdIns kinase activity was associated with IGF-I receptor which had been phosphorylated in vitro but not with unphosphorylated receptor (Fig. 2 R ) . The iden-

IGF-I Receptor
tity of the lipid moiety produced in these PtdIns kinase assays was confirmed as PtdIns 3-monophosphate (PtdIns(3)P) through the analysis of the deacylated lipid products on HPLC (data not shown). Thus, the PtdIns kinase that was associated with IGF-I receptor immunoprecipitates possessed PtdIns 3kinase activity. Recombimntp85 Inhibits the Association of PtdIns 3-Kinase Activity with IGF-I Receptor: Involvement of the SH2 Domuin-To test if p85 has an effect on the association between IGF-I receptor and PtdIns 3-kinase activity, the p85 fusion protein was incubated with LISN C4 cell lysates and the immunoprecipitated IGF-I receptor, and PtdIns 3-kinase activity was measured. The P85 fusion protein (100 nM) totally inhibited the association of PtdIns 3-kinase activity with the immunoprecipitated IGF-I receptor (Fig. 3A). This inhibition was dose-dependent with an IDso <10 nM (Fig. 3A). T o identify the region in the p85 molecule responsible for the inhibition, the constructs shown in Fig. 1 were tested in the same assay. Only the SH2-containing fusion protein was able t o inhibit the association of PtdIns 3-kinase activity with IGF-I receptor, whereas fusion proteins of p85 containing SH3 domain, BCR homologous domain, and glutathione Stransferase had no effect (Fig. 3B). No PtdIns 3-kinase activity was detected when the immunoprecipitated IGF-I receptor was incubated with buffer alone (Fig. 3R).
Beads, treated exactly as described in Fig. 3 4 were analyzed by SDS-polyacrylamide gel electrophoresis and autoradiography to identify "P-phosphorylated proteins (Fig. 3C). A prominent phosphorylated IGF-I receptor , 9 subunit is present in all lanes. When the SH2-containing fusion protein (lane 5 ) or intact p85 fusion protein (lane 6 ) was included in the incubation, prominent phosphorylated bands corresponding t o these proteins were present in addition. Two faint bands with molecular weights of 110,000 and 85,000, which correspond to the known molecular weights of the subunits of The prominent bands corresponding to the SH2-containing and p85 fusion proteins suggest that these proteins bind directly to IGF-I receptor, most probably through SH2-phosphotyrosine interaction. T o confirm this direct association, exogenous SH2-containing and p85 fusion proteins were incubated with immunoprecipitated IGF-I receptor. After an in vitro kinase reaction, the sample was subjected to SDS-polyacrylamide gel electrophoresis, and phosphorylated bands were analyzed by autoradiography. Fig. 4 clearly shows that the SH2-containing fusion protein was bound to IGF-I receptor and phosphorylated by the receptor tyrosine kinase. This occurred only if the receptor was activated by in vitro autophosphorylation prior to incubation with the SH2-containing fusion protein. The same result was obtained when p85 fusion protein was used (data not shown).
T o further establish that the SHP-containing fusion protein was inhibiting PtdIns 3-kinase activity by competing for binding to the IGF-I receptor and not by directly inhibiting PtdIns 3-kinase catalytic activity, IGF-I receptor was incubated with the fusion protein, washed, and then incubated with cell lysate before measuring PtdIns 3-kinase activity or After washing, the immunocomplexes were incuhated with lysis buffer ( l a n e 1 ) or with IGF-I receptor-depleted LISN C4 cell lysate (lanes [2][3][4][5][6][7][8][9][10]. Intact p8S fusion protein or glutathione S-transferase were added to these incubations at the concentrations indicated in the figure. After washing, the associated PtdIns kinase activity was assayed, and resultant PtdIns(3)P was separated hy TLC and visualized by autoradiography. These result.. are representative of two different experiments. l?, IGF-I receptor was immunoprecipitated from unstimulated LISN C4 cells and phosphorylated in uitro. After washing, the immunocomplexes were incuhated with lysis buffer (lane 1 ) or with IGF-I receptor-depleted LISN C4 cell lysate (lanes [2][3][4][5][6][7]. Different portions of the pR5 molecule at the concentration of 200 nM were included in these incubations as indicated in the fimre. After washing, the associated PtdIns kinase activity was assayed, and resultant PtdIns(3)P was separated by TLC and visualized by autoradiography. These results are representative of three different experiments. The radioactivity contained in the area comigrating with PIP was 31 cpm ( l a n e I ) , 28, 600 cpm (lane 2 ) , 27,500 cpm ( l a w 3 ) . 24, 600 cpm (lane 4 ) , 210 cpm (law 5). 160 cpm ( l a w 6 ) , and 26. S00   cpm ( l a n e 7). C, half of the incuhation of the assay in l? was mixed with SDS sample buffer and proteins were wparated on a 10-206 gradient gel. Phosphorylated hands were visualized by autoradiography. The positions of the exogenously added SH2-containing and pR5 fusion proteins are indicated hy the white arrow and arrowhmd. respectively. The positions of the LISN C4 cells pll0 and pR5 are indicated by the black arrow and arrowhead, respectively. The position of the 0 subunit of IGF-I receptor is also indicated. first incubated with cell lysate and then the fusion protein.
Preincubation of the immunoprecipitated IGF-I receptor with the SH2-containing fusion protein inhibited the association of PtdIns 3-kinase activity with the immunoprecipitated IGF-I receptor (Fig. 5 , lanes 1 and 2 ) . However the SH2-containing 1 2 the various SH2-containing fusion proteins. The SH2-con-

ik-:
taining fusion protein of p85 (100 nM) almost completely inhibited the association of PtdIns 3-kinase activity with immunoprecipitated IGF-I receptor (Fig. 6, lanes I and 2 ) , whereas glutathione S-transferase (300 nM) had no effect (Fig. 6, lanes 3 and 4 ) . The SH2-containing fusion protein from GAP partially inhibited the association of PtdIns 3kinase activity with immunoprecipitated IGF-I receptor (Fig.  6, lanes 7 and 8), whereas the SH2-containing fusion protein from pp60"" had no effect (Fig. 6, lanes 9 and IO). This study demonstrates that there is a high affinity association between p85 and the IGF-I receptor that is dependent upon the phosphorylation state of the receptor and occurs through the carboxyl-terminal half of p85 which contains both SH2 domains. Because of the well known ability of SH2 domains to tightly bind to phosphotyrosine-containing peptides (1,(9)(10)(11)(12)(13)(14), it is likely that one or both of these SH2 domains are responsible for this interaction. It is probable that through this interaction the catalytically active PtdIns 3-kinase heterodimer becomes associated with the IGF-I receptor, since this association is inhibited by p85 and its carboxyl-terminal SH2-containing fragment.
Furthermore, p85 and its carboxyl-terminal fragment are directly phosphorylated by IGF-I receptor adsorbed to antibody beads. This strongly suggests that this portion of p85 is a direct in uitro substrate for IGF-I receptor kinase. Also when IGF-I receptor adsorbed to antibody heads is incubated with cell lysate, and then incubated with ATP, 85-and 110-kDa bands, which probably represent subunits of PtdIns 3-kinase, become phosphorylated. Thus, under these in uitro conditions both subunits may be substrates of the IGF-I receptor kinase. Because of the manner in which these studies were performed, The proteins were added as follows: Iaws 3 and 4, glutathione Stransferase (300 nM): lanes.5 and 6. the SH2-containina fusion protein fusion protein did not inhibit PtdIns 3-kinase activity which containing fusion from GAP or 3w nM, respectively); from p85 (100 or 300 nM, respectively); lunes 7 and 8. the SH2was prebound to the IGF-1 receptor (Fig. 5, lanes 3 and 4 ) . lanes 9 and 10, the SH2-containing Fusion protein from p p 6 W (100 Specificity of the SH2-mediated Inhibition-To study the or 300 nM, respectively). After washing, the associated Ptdlns kinase specificity of the SH2-containing molecule for the inhibition activity was assayed, and resultant Ptdlns(3)P was separated by TLC of the association between ptdIns 3-kinase activity and the and visualized by autoradiography. These results are representative IGF-I receptor, SH2-containing fusion proteins from GAP of two different experiments. The radioactivity contained in the area comigrating with PIP was 7100 cpm (lane I ), 8100 cpm (lune 2 ) . 9200 and pp60"" were also and expressed. These the possibility that a tightly associated protein kinase other than IGF-I receptor is actually responsible for the observed phosphorylation cannot be excluded. How does the association of IGF-I receptor with PtdIns 3kinase and its phosphorylation observed in these in vitro studies relate to what happens in intact cells, and what role does it play in PtdIns 3-kinase activation? When intact cells are treated with IGF-I there is a large increase in the amount of PtdIns 3-kinase activity that is immunoprecipitated by anti-phosphotyrosine antibody (17,18). This result is consistent with a direct phosphorylation of PtdIns 3-kinase by IGF-I receptor tyrosine kinase. However, another possible explanation for these findings is that PtdIns 3-kinase becomes associated with a protein which is tyrosine phosphorylated in response to IGF-I and is co-immunoprecipitated. Indeed, PtdIns 3-kinase is co-precipitated with IRS-1, a major substrate for both insulin and IGF-I receptors, when cells are treated with insulin (21). The availability of anti-PtdIns 3kinase antibodies should make it possible to determine whether PtdIns 3-kinase is directly phosphorylated in intact cells exposed to IGF-I.
When intact cells are incubated with IGF-I, some PtdIns 3-kinase activity appears to associate with the IGF-I receptor (17,18). However, this represents only a small fraction of the PtdIns 3-kinase activity that can be precipitated with antiphosphotyrosine antibodies, even in LISN C4 cells which express a very large number of IGF-I receptors (17,18). This is surprising in view of the very high affinity of p85 for phosphorylated IGF-I receptor, and the very much greater amount of IGF-I receptor that associates with PtdIns 3-kinase when incubated with cell lysate. Similar results have been reported with PDGF receptors, which react with PtdIns 3kinase to a greater extent in cell lysates than in intact cells (30). There are several possible explanations for the small amount of p85 that is associated with IGF-I receptor in intact cells; other proteins, such as IRS-1, may become phosphorylated in intact cells and effectively compete with the receptor for binding PtdIns 3-kinase (21). In intact cells following IGF-I stimulation, PtdIns 3-kinase may become associated with the receptor transiently and dissociate after its phosphorylation. In case of epidermal growth factor receptor, it is proposed that tyrosine phosphorylation of phospholipase C--y may signal its dissociation from epidermal growth factor receptor (31). The tyrosine-containing IGF-I receptor sequence that is most homologous to the known PtdIns 3-kinase binding site in middle t antigen and the PDGF receptor is only a minor phosphorylation site in intact cells (1,32), but extrapolated from the case of the insulin receptor this site is more extensively phosphorylated in broken cells. Interestingly, a mutant insulin receptor in which this region was deleted was able to increase the activity of PtdIns 3-kinase in anti-phosphotyrosine immunoprecipitates (33). It will be interesting to determine if this mutation or a comparable mutation in the IGF-I receptor affects its ability to bind PtdIns 3-kinase or directly phosphorylate p85.
A fusion protein containing the SH2 domains of GAP, but not a fusion protein containing the SH2 domain of pp60"" inhibited the association of PtdIns 3-kinase activity with phosphorylated IGF-I receptor. The GAP fusion protein was less effective than the p85 fusion protein. It is not clear if this is because GAP has a lower affinity for IGF-I receptor or because it binds to a different site on the receptor and only partially inhibits p85 binding. Although GAP is known to associate with several tyrosine kinases (34), this is the first report of a direct association of GAP with IGF-I receptors. Whether the interaction is sufficiently strong to have any I 5 to IGF-I Receptor physiological significance is unclear.
It is possible that the binding of endogenous p85 to its target protein is regulated not only by tyrosine phosphorylation of the protein but also by competition with other SH2containing molecules. It is also likely that the SH2 domain of p85 possesses specificity for each phosphotyrosine containing motif and that it directs PtdIns 3-kinase activity to its target. Experiments are currently underway to define which part of the SH2 domain of p85 is responsible for the inhibitory effect on the association between IGF-I receptor and PtdIns 3kinase activity. Although the role of PtdIns 3-kinase in signal transduction is unknown, further studies on its association with other molecules will help to increase our understanding of the importance of this enzyme.