Phosphoinositide 3-Kinase Is Activated by Phosphopeptides That Bind to the SH2 Domains of the 85-kDa Subunit*

, Tyrosine-phosphorylated peptides based on the regions of polyoma virus middle t antigen and the plate- let-derived growth factor receptor that bind phosphoinositide 3-kinase are shown to activate this enzyme 2-3-fold in vitro. The concentrations of the peptides required to activate the enzyme are at least 10-1000- fold higher than the dissociation constants of these peptides for the individual SH2 domains of the 85-kDa subunit (KO < 100 nM). Doubly phosphorylated peptides are more effective than singly phosphorylated peptides. The results suggest that a fraction of the cellular phosphoinositide 3-kinase has SH2 domains with relatively low affinity for phosphopeptides and that binding of phosphopeptides to these enzymes causes activation. Thus, SH2 domains may be involved not only in recruiting the enzyme but also in regulating activity.

Tyrosine-phosphorylated peptides based on the regions of polyoma virus middle t antigen and the platelet-derived growth factor receptor that bind phosphoinositide 3-kinase are shown to activate this enzyme 2-3-fold in vitro. The concentrations of the peptides required to activate the enzyme are at least 10-1000fold higher than the dissociation constants of these peptides for the individual SH2 domains of the 85-kDa subunit (KO < 100 nM). Doubly phosphorylated peptides are more effective than singly phosphorylated peptides. The results suggest that a fraction of the cellular phosphoinositide 3-kinase has SH2 domains with relatively low affinity for phosphopeptides and that binding of phosphopeptides to these enzymes causes activation. Thus, SH2 domains may be involved not only in recruiting the enzyme but also in regulating activity.
Phosphoinositide (PtdIns)' 3-kinase is the critical enzyme in a recently discovered intracellular signaling pathway that is activated by a wide range of growth factors, oncoproteins, and nonmitogenic stimuli (1). PtdIns 3-kinase was discovered because of its association with the oncogene products pp60""" and middle T . pp60"'"" complex (2)(3)(4)(5). Studies of polyoma virus middle t antigen (mT) mutants indicate that m T must associate with both pp60"" (or a close relative of pp60""") and with PtdIns 3-kinase to transform cells (2,(4)(5)(6)(7)(8)(9)(10)(11)(12). The association of m T with ~~6 0 "~" results in activation of the protein-tyrosine kinase activity of pp6OC-" and the phosphorylation of m T (11, 13). The lipid products of PtdIns 3-kinase are elevated in cells transformed by mT, indicating a correlation between the association of this enzyme with m T and increased activity in intact cells (14,15). Similarly, PtdIns 3-kinase directly associates with and is activated by a variety * This work was supported by National Institutes of Health Grants GM 41890 and GM 36624 (to L. C. C.) and CA 34722 (to B. S.), an American Heart Association-Parke-Davis Clinician Scientist Award (to C. L. C.), and United States Public Health Service Grant IP30DK39428 (to K. R. A.). The first two authors have contributed equally to this work. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore he hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

278-3033.
of growth factor receptors. The mechanism of activation of PtdIns 3-kinase is not known.
PtdIns 3-kinase has been purified from rat liver, bovine brain, mouse fibroblasts, and bovine thymus (16)(17)(18)(19)(20). PtdIns 3-kinase purified from rat liver is a heterodimer of an 85-kDa protein (p85) and one of two 110-kDa proteins (~1 1 0 ) (16). The cDNA that encodes the 85-kDa subunit of PtdIns 3kinase has been cloned (17,18,21). The protein encoded by this cDNA contains three regions of homology to pp6OC-"", two SH2 domains and an SH3 domain. This subunit also contains a region of homology to rho-GTPase-activating protein (rho-GAP) and the C terminus of the breakpoint cluster region gene product. The cDNA does not predict a nucleotide binding domain, and the expressed proteins do not have PtdIns kinase activity. The cDNA for the 110-kDa subunit of PtdIns 3-kinase has also been recently cloned, and the expressed protein has PtdIns kinase activity (22). The cDNA has homology only to a yeast gene VPS 34. The SH2 domains of p85 bind to tyrosine kinases suggesting that p85 mediates the binding of PtdIns 3-kinase to tyrosine-phosphorylated proteins (18,(23)(24)(25).
Tyrosine 315 of mouse m T is implicated as the site which is important for PtdIns 3-kinase binding to mT. Tyrosine 315 has also been identified as a primary site of m T phosphorylation (26). A mutant of m T in which tyrosine 315 is replaced by phenylalanine has little associated PtdIns 3-kinase activity in uiuo and is defective in cell transformation (4, 27). The SH2 domains of p85 bind in solution and on nitrocellulose blots to the region of mT that contains tyrosine 315, when that tyrosine is phosphorylated (28). Phosphorylation of m T is necessary for its association with PtdIns 3-kinase (29) and synthetic tyrosine-phosphorylated peptides, based on the region of tyrosine 315 of mT, block the association of baculovirus-expressed m T with purified PtdIns 3-kinase (30). The sequence surrounding tyrosine 315 of m T is similar to sequences of two regions of the kinase insert domain of the platelet-derived growth factor (PDGF) receptor, which also contain phosphotyrosine residues, that are implicated as PtdIns 3-kinase binding sites (31, 32). The sequences in the PDGF receptor are tyrosines 740 and 751 in the human P PDGF receptor. Similar sequences have also been identified in other proteins known to bind PtdIns 3-kinase (1).
PtdIns 3-kinase activity has usually been measured in immunoprecipitates from lysates of quiescent or stimulated cells using anti-receptor, anti-phosphotyrosine, or anti-p85 immunoprecipitates. These assays have not provided a reliable estimate of specific activity or been able to dissociate the effects of protein-protein interaction and phosphorylation of PtdIns 3-kinase. However, using purified PtdIns 3-kinase we have been able to address the regulation of this enzyme.
In this paper we have investigated the effects of association of PtdIns 3-kinase with mT and the PDGF receptor using phosphopeptides based on the binding site of PtdIns 3-kinase in these proteins. We show that doubly phosphorylated peptides that associate with the SH2 domains of p85 activate both purified PtdIns 3-kinase and the crude cytosolic enzyme. These results indicate that association with a tyrosine-phosphorylated protein alone can activate this enzyme and indicate that the SH2 domains regulate enzymatic activity.

EXPERIMENTAL PROCEDURES
Materials-Phosphorylated peptides were synthesized using an N*%moc synthetic strategy as described elsewhere (30,44). Phosphotyrosine residues were incorporated using N"-Fmoc-L-phosphotyrosine and extended coupling times. Peptides were cleaved from the resin and the more labile side chain-protecting groups were simultaneously removed using trifluoroacetic acid, thioanisole, ethanedithiol, and anisole (90:5:3:2). Following ether precipitation, methyl esterprotecting groups on the tyrosine phosphate side chain were removed during a second stage of deprotection with trimethylsilyl bromide (44). All phosphopeptides had the expected amino acid composition and were greater than 80% homogeneous by reversed-phase high performance liquid chromatography analysis. The glutathione transferase fusion genes were expressed in Escherichia coli as described elsewhere (28). These genes were all derived from a p85a clone isolated from rat liver? The N-terminal SH2 domain included amino acids 321 to 470 (based on the sequence of Skolnik et al. (21)). The C-terminal SH2 domain included amino acids 562 to the C terminus.
Antibodies-The mT antiserum is a polyclonal rabbit antiserum kindly provided by D. Pallas and T. Roberts of the Dana-Farber Cancer Institute. The antiserum to p85 was raised against glutathione-S-transferase fusions of the full length and C-terminal half of p85. Protein Purification-PtdIns 3-kinase was purified from rat liver to homogeneity as previously described (16). The rat liver cytosol was prepared as previously described for the purification of PtdIns 3kinase. pp60""" and polyoma mT genes were expressed in Sf9 cells using baculovirus as previously described (30).
Ptdlns 3-Xime Assays-PtdIns 3-kinase assays were done as previously described (16) except that the assays contained 200 p~ sodium vanadate. Peptides were preincubated with PtdIns 3-kinase for 20 min in a volume of 12 pl. Lipids and ATP/MgCl, were then added to start the reaction which had a final volume of 20 pl. The final concentration of PtdIns 3-kinase in these reactions was about 100 nM. The initial peptide concentrations during the preincubation were higher than the final concentration after adding the lipids and ATP/MgCIZ. The final peptide concentrations are shown in the figure legends. PtdIns 3-kinase activity was quantitated by liquid scintillation counting of the chloroform extract of the reaction. Rat liver * C. L. Carpenter, J. Cremins, and L. C. Cantley, manuscript in preparation.
cytosol was assayed as previously described (16) after preincubation with peptides for 20 min at room temperature. These assays, however, also contained 200 p~ sodium vanadate. The products were separated by thin layer chromatography using 1-propanol and acetic acid as solvents (30).
Zn Vitro Association of Purified PtdZns 3-Kinase with Baculouirusexpressed mT.pp6P."" Complex-To determine the relative ability of the peptides to block the association of PtdIns 3-kinase and mT, we preincubated PtdIns 3-kinase with the peptides at 4 "C for 1 h and then added lysates of Sf9 cells which were doubly infected with baculoviruses containing mT and pp6OC-"". The incubation was then continued for another hour. Immunoprecipitates with antiserum raised against mT were done as previously described (30). PtdIns kinase assays were done on the washed immunoprecipitates, as previously described (30).
Association of mT with the SH2 Domains of p85"Constructs containing the individual SH2 domains of p85 were expressed in the pGEX3 vector. The proteins were purified from bacterial sonicates with glutathione-agarose beads. The proteins bound to the glutathione-agarose beads were washed extensively with a buffer of 150 mM NaCI, 10 mM Tris, pH 7.5, 1 mM EDTA, and 1% Nonidet P-40. The peptides were then added and incubated with the SH2 domains for 20 min. A mT lysate was then added and the incubation continued for another hour. The beads were then washed as for an immunoprecipitate, and the samples run on an 8.5% SDS-PAGE gel. The proteins were then transferred to nitrocellulose or Immobilon-P paper, blocked with milk, probed with the mT antiserum, and visualized by the alkaline phosphatase reaction.
Cell Culture and Immunoprecipitates-Rat-l cells were grown and immunoprecipitated with the p85 antibody as described elsewhere (45). Immunoprecipitates were incubated at 4 "C for 1 h with peptides before the addition of antibody. PtdIns 3-kinase activity was assayed on the immunoprecipitates as previously described (45). The products were separated by thin layer chromatography, visualized by autoradiography, and quantitated by liquid scintillation counting.
Sucrose Gradients-Purified PtdIns 3-kinase was incubated with 100 PM peptide G and then layered over a 5-25% sucrose gradient containing 100 p~ peptide G. Protein standards were also included. PtdIns 3-kinase without phosphopeptide was run on a similar gradient as a control. The gradients were centrifuged for 6 h in a Beckman SW 55 rotor at 55,000 rpm. Fractions were removed and assayed for PtdIns kinase activity and Western blotted for p85. Samples were also separated by SDS-PAGE and Coomassie Bluestained to detect the standards (bovine serum albumin, yeast alcohol dehydrogenase, and horse apoferritin).

Submicromolar Concentrations of Tyrosine-phosphorylated
Peptides Block the Association of PtdIns 3-Kinase with mT.
pp6~~""Recombinant mT -pp60c~"" tightly associates with purified PtdIns 3-kinase and phosphorylates both subunits on tyrosine (30). We wanted to further investigate the mechanism by which mT.pp60"-"" activates PtdIns 3-kinase. To test the possibility that association with mT activates PtdIns 3-kinase, we synthesized a series of tyrosine-phosphorylated peptides based on the regions of mT (30) and the PDGF receptor (31) that bind PtdIns 3-kinase (Table I).
To be certain that the peptides mediated their effect by binding to PtdIns 3-kinase we investigated the ability of the peptides to block association of purified PtdIns 3-kinase with baculovirus-expressed mT .pp60""" complex. Both a singly phosphorylated peptide (peptide C) and a doubly phosphorylated peptide (peptide G) block association of purified PtdIns 3-kinase with mT~pp60""" complex ( Fig. 1). Peptide C inhibited 50% of maximal binding at a concentration of approximately 40 nM. Peptide G, which is phosphorylated at both tyrosine 315 and tyrosine 322, caused 50% maximal inhibition at a concentration of approximately 25 nM. Tyrosine 322 is phosphorylated in mT immunoprecipitates and may be phosphorylated in vivo (46,47). Nonphosphorylated peptides had no effect in the concentration range investigated (not shown, see also Ref. 30 Table I), at the indicated concentrations, were added to a lysate of Sf9 cells coinfected with baculovirus containing mouse polyoma mT antigen and chicken pp60'.sx. A mT immunoprecipitate was then done and assayed for PtdIns kinase activity. The products were separated by thin-layer chromatography and quantitated by scintillation counting. The data curves (solid line, peptide G; dotted line, peptide C) are weighted nonlinear least squares fits.
by Binding to the SH2 Domains of p85"Since p85 associates with mT and the association of PtdIns 3-kinase with mT is dependent on phosphorylation of mT, it is very likely that the association of PtdIns 3-kinase with mT is mediated by binding of tyrosine-phosphorylated residues of mT with the SH2 domains of p85 (28). It is therefore also likely that the synthetic phosphopeptides based on the mT sequence bind to the SH2 domains of p85 and mediate their effects through that interaction. We used constructs of the SH2 domains of p85 expressed in pGEX 3 vector to show that phosphopeptide D prevents the association of mT with these constructs ( Tyrosine-phosphorylated Peptides Based on mT and PDGF Receptor Domains Sequence Actiuate Cytosolic PtdIns 3-Kime-We investigated the effect of these peptides on PtdIns 3-kinase activity in cell lysates. The enzymatic activity of PtdIns 3-kinase in cytosol when PtdIns-4,5-Pz is used as a substrate appears to be suppressed compared to the activity of purified enzyme (16). We added tyrosine-phosphorylated peptide D or G or nonphosphorylated peptide E to the cytosolic fraction of liver in order to test the possibility that association with a phosphopeptide alone could activate the cytosolic enzyme. Since there are other phosphoinositide kinases in rat liver cytosol, we assayed PtdIns 3-kinase activity utilizing the substrate PtdIns-4,5-P*. No kinases other than PtdIns 3-kinase are known to phosphorylate this lipid. As shown in Fig. 3, both of the tyrosine-phosphorylated peptides stimulated the activity of the enzyme in crude cytosol. The nonphosphorylated peptide (E) had no significant effect on activity. The doubly phosphorylated peptide (G) activated the enzyme more than the singly phosphorylated peptide (D). Much higher concentrations of the peptides were required to elicit activation than were required to block binding of purified PtdIns 3-kinase to mT.pp60""". Addition of higher concentrations of vanadate (up to 1 mM) to inhibit phosphatases did not significantly reduce the concentration of peptide necessary to activate the enzyme.
We have also examined the effect on PtdIns 3-kinase activity of a doubly phosphorylated peptide based on the human p PDGF receptor sequence (peptide I). This peptide is phos- phorylated at sites equivalent to tyrosines 740 and 751 in the PDGF receptor sequence. In rat liver cytosol this peptide results in a 1.7-fold increase in PtdIns 3-kinase activity a t a concentration of 10 pM and a 2.4-fold increase in activity at a concentration of 100 pM (Fig. 4). Higher concentrations of the peptide do not result in further activation.
We also investigated the effect of the doubly phosphorylated peptide (G) on PtdIns 3-kinase activity in anti-p85 immunoprecipitates of fibroblast and rat liver cytosol. The antibodies we have raised against recombinant p85 quantitatively immunoprecipitate the purified enzyme and precipitate activity from cell lysates. However, immunoprecipitation of the purified enzyme results in loss of 8 0 4 0 % of the activity. This may be due to steric hindrance of access of the substrate to the active site in the immune complex or to conformational effects of the antibodies. The phosphopeptides had no significant effect on the activity of PtdIns 3-kinase in anti-p85 immunoprecipitates from rat liver or fibroblasts (Fig. 4). However, if peptide G is preincubated with the cytosol or lysate before the addition of antibody, activation of PtdIns 3kinase activity is seen after immunoprecipitation (Fig. 4). Preincubation with the singly phosphorylated peptide D resulted in a slight inhibition of activity (Fig. 5).
Since the antibody is partially inhibitory, the doubly phosphorylated peptide is likely relieving inhibition rather than activating in this experiment. A possible interpretation is that the doubly phosphorylated peptide locks the enzyme into an active conformation prior to immunoprecipitation so that some of the inhibitory antibodies in the polyclonal serum fail to bind. However, if the enzyme is precipitated with these antibodies first then the peptide is not effective a t removing the inhibitory antibodies. The singly phosphorylated peptide was obviously not effective at blocking inhibition even when added prior to immunoprecipitation, perhaps do to a lower affinity for the SH2 domains.
Tyrosine-phosphorylated Peptides Based on mT and PDGF Receptor Domains Activate Purified PtdIns 3-Kinase-We also investigated the ability of the phosphopeptides to activate purified PtdIns 3-kinase. There was no loss of enzyme activity during incubation with buffer alone (data not shown). The unphosphorylated peptide E had no effect on the activity of purified PtdIns 3-kinase at a concentration of 300 p~ (Fig.  5 ) . A phosphopeptide based on a sequence from pp60"."", which has a low affinity for the SH2 domains of PtdIns 3kinase (51) also had no effect on activity at a concentration of 300 p~. However, incubation of purified PtdIns 3-kinase with the doubly phosphorylated peptide (peptide G) caused a 2-fold increase in enzymatic activity. The 50% maximal effect was approximately 1 pM. There is some variability among  Table I). The peptide concentration during the preincubation was 1.67-fold higher than the concentration during the PtdIns 3-kinase assay. The concentration during the assay is shown. PtdIns 3-kinase assays were then done with PtdIns as the substrate and quantitated by liquid scintillation counting of the chloroform extract. Each point represents the mean of three to six experiments _t S.E.

Ptdlns 3-Kinase
Activation by Phosphopeptides different preparations of PtdIns 3-kinase. In some preparations of PtdIns 3-kinase, a 2-fold activation was seen at a concentration of peptide G of 50 nM and a %fold activation at 500 nM (data not shown). Considerable variation in response to the PDGF receptor phosphopeptide was also found in some experiments a 3-fold activation was seen at 100 p~, whereas in other experiments no significant activation was seen (data not shown). The singly phosphorylated peptides containing phosphotyrosine 315 (peptide D) and phosphotyrosine 322 (peptide F) were less effective in activating PtdIns 3-kinase. A 100-fold higher concentration was necessary to see any effect. The maximal activation by these peptides was also less than that caused by the doubly phosphorylated peptide. Similar activation curves were observed using all three substrates of the PtdIns 3-kinase (PtdIns, PtdIns-4-P, and PtdIns-4,5-Pz; not shown). The concentration dependence for activation of the enzyme by both the doubly phosphorylated peptide and singly phosphorylated peptides was considerably higher than the concentration dependence for blocking binding to mT (Fig.   1). Singly phosphorylated peptides A and C activated PtdIns 3-kinase at similar concentrations and to the same degree as singly phosphorylated peptides D and F. To determine whether doubly phosphorylated peptide G causes a change in the oligomerization state of PtdIns 3kinase the rate of sedimentation in a sucrose gradient was investigated. Control enzyme and enzyme preincubated and sediment with 100 /*M peptide G migrated identically, as determined by PtdIns 3-kinase activity and Western blotting for p85 with an apparent molecular mass of about 200 kDa (not shown). Thus, activation does not involve a change in the oligomerization of PtdIns 3-kinase.

DISCUSSION
Association of PtdIns 3-kinase with membrane-bound tyrosine-phosphorylated proteins is correlated with the appearance of 4,15,[36][37][38][39]. The increase in activity of PtdIns 3-kinase likely involves several processes. Recruitment of the enzyme from the cytosol to the membrane provides proximity to the lipid substrates. Phosphorylation or dephosphorylation reactions involving tyrosine and/or serine residues of the 85-kDa subunit are also likely to be important. The fraction of PtdIns 3-kinase that is associated with mT is phosphorylated on both tyrosine and serine residues in vivo (48). Phosphorylation of p85 on serine by a tightly associated serinelthreonine kinase inhibits the PtdIns 3-kinase activity (45), and dephosphorylation at these sites could cause activation in vivo.
Here we present an additional mechanism for regulation of PtdIns 3-kinase. We used tyrosine-phosphorylated peptides, containing either a single phosphotyrosine or two phosphotyrosines, based on mouse mT Tyr-315. This region of mT has been implicated in cell transformation and association with PtdIns 3-kinase. A mutant (d123) in which these residues are deleted is transformation defective and has dramatically reduced PtdIns 3-kinase activity (5). A point mutation, converting tyrosine 315 to phenylalanine, results in a reduced frequency and different spectrum of tumors (27) and loss of the majority of associated PtdIns 3-kinase activity (4). Antibodies against this region immunoprecipitate a fraction of mT that does not associate with PtdIns 3-kinase activity (27). In addition to tyrosine 315, tyrosine 322 is phosphorylated in immunoprecipitates (46,47). The tyrosine 315 to phenylalanine mutant has some associated PtdIns 3-kinase activity which suggests that other tyrosine-phosphorylated sites might also be involved in this association. The ability of the doubly phosphorylated peptide (analogous to phosphotyrosine 315 and phosphotyrosine 322) to activate purified PtdIns 3-kinase and the lesser ability of singly phosphorylated peptides (analogous to phosphotyrosine 315 or phosphotyrosine 322) to activate in spite of comparable binding to PtdIns 3-kinase suggests that other phosphotyrosine residues downstream of tyrosine 315 could be important in regulation of activity. Besides tyrosine 322, other potential tyrosine-phosphorylation sites in this region might also be involved and need to be further investigated.
In the three conditions in which we have looked at the effect of phosphopeptides on PtdIns 3-kinase activity (purified enzyme, cell lysates, and immunoprecipitates) the doubly phosphorylated peptide is much more effective at activating PtdIns 3-kinase. The more marked effect of the doubly phosphorylated peptide in activating purified PtdIns 3-kinase suggests that the activation is a result of the peptide binding either to a region that best recognizes two phosphotyrosines or to two distinct sites that each recognize phosphotyrosine. The equal ability of both singly and doubly phosphorylated peptides to block association of PtdIns 3-kinase with mT at concentrations lower than those required for activation suggests that the activation is due to binding at a separate site or to the same sites on a pool of PtdIns 3-kinase with SH2 domains that have low affinities for the phosphopeptides. The doubly phosphorylated peptide appears to have a much higher affinity for the sites that activate. The PtdIns 3-kinase assays were done at substrate concentrations much above the K, values, indicating the the activation is due to an increase the V, , , of the reaction. Backer et al. (49) have reported the activation of PtdIns 3kinase by insulin receptor substrate (IRS-1) and tyrosinephosphorylated peptides (phosphorylated at a single position) in immunoprecipitates of p85. Unlike Backer et al. we found no activation of PtdIns by phosphopeptides in immunoprecipitates. We did find greater activity in immunoprecipitates from cell lysates preiincubated with phosphopeptides, suggesting that the activation in immunoprecipitates may be due to prevention of inhibition by the antibody. We used a different antibody than Backer et al. so that direct comparison of these results is not possible.
Several models would explain these data. Although the individually expressed SH2 domains have similar affinities for the phosphopeptide (see Fig. 2) the affinities of the SH2 domains in the native protein may differ markedly. In this model one SH2 domain may recognize singly and doubly phosphorylated peptide equally well, but binding to this SH2 domain would not cause activation. The other SH2 domain would have a much higher affinity for doubly phosphorylated peptide and binding to this SH2 domain would cause activation. At high concentrations the singly phosphorylated peptides would also bind to this SH2 domain and cause activation.
An alternative explanation is that two phosphotyrosine recognition sites must be occupied for activation: a high and a low affinity site. Only the high affinity site would be involved in binding to mT. Peptide competition for binding of PtdIns 3-kinase to mT then would occur at low peptide concentrations. The low affinity of the second site would be overcome by the doubly phosphorylatedpeptide because it would already be bound to the high affinity site and would have a significant entropic advantage in binding to the low affinity site. This second site could either be the other SH2 domain or a second phosphotyrosine recognition site within the same SH2 domain. At high concentrations the singly phosphorylated peptides would bind to both sites and cause activation.
A third possibility is that a fractions of the PtdIns 3-kinase is modified in such a way that the SH2 domains have low affinities for phosphopeptides. This fraction would not bind tightly to mT. pp60"-"" and therefore would not be detected in the competition assay shown in Fig. 1. Binding of the phosphopeptides to the SH2 domains of this population of PtdIns 3-kinase could force the SH2 domains into a normal conformation and in so doing activate the catalytic subunit. We favor this model since there is already evidence that in PDGFstimulated cells a significant fraction of p85 is modified so that it has a much reduced affinity for the PDGF receptor and measured by a ligand blotting assay (31). In addition a significant portion of PtdIns 3-kinase in cell lysates fails to associate mT.pp6O""" in the presence of excess mT~pp60'~"".3 Finally, the variability in the activation of PtdIns 3-kinase by the phosphopeptides could be explained by variation in the amount of modified enzyme. The data of Escobedo et al. (31) suggest that the affinities of the SH2 domains are modified by phosphorylation. It is also possible that association with another protein could alter the affinities.
Higher concentrations of the phosphopeptides were required to see significant activation in liver cytosol compared to purified enzyme. There are several possible explanations for this finding. Phosphatases may have dephosphorylated some of the peptide but activation is not seen at lower peptide concentrations even with verv high concentrations of vanadate. The peptide could be competing with another protein bound to p85 or the peptide could be binding to other proteins (lowering the effective concentration).
Many proteins that activate PtdIns 3-kinase have clusters of multiple tyrosine-phosphorylated peptides in close proximity. As discussed above tyrosines 708 and 719 of the mouse PDGF p receptor (and 740 and 751 of the human receptor) are known to be involved in PtdIns 3-kinase binding (31). IRS-1 also binds PtdIns 3-kinase. It is phosphorylated on tyrosine and has several closely spaced clusters of sequences that satisfy the requirements for being PtdIns 3-kinase binding sites (50). Interestingly, in addition to the 85-kDa subunit " of PtdIns 3-kinase, several other enzymes are known to have two SH2 domains (PLC-7 and ras-GAP): doubly phosphorylated peptides might also exert regulatory affects on these enzymes by a similar mechanism.
In summary, the results presented here suggest a new model for regulation of PtdIns 3-kinase that may be applicable for activation of this enzyme by a wide range of growth factors and oncogene products. This model also implies a new function for SH2 domains. These domains could act as direct switches to regulate enzymes by interacting with tyrosine phosphorylated proteins.