Phosphoinositide 3-Kinase Regulates Phospholipase Cγ-mediated Calcium Signaling*

It has been demonstrated that the lipid products of the phosphoinositide 3-kinase (PI3K) can associate with the Src homology 2 (SH2) domains of specific signaling molecules and modify their actions. In the current experiments, phosphatidylinositol 3,4,5-trisphosphate (PtdIns-3,4,5-P3) was found to bind to the C-terminal SH2 domain of phospholipase Cγ (PLCγ) with an apparent K d of 2.4 μm and to displace the C-terminal SH2 domain from the activated platelet-derived growth factor receptor (PDGFR). To investigate the in vivorelevance of this observation, intracellular inositol trisphosphate (IP3) generation and calcium release were examined in HepG2 cells expressing a series of PDGFR mutants that activate PLCγ with or without receptor association with PI3K. Coactivation of PLCγ and PI3K resulted in an ∼40% increase in both intracellular IP3generation and intracellular calcium release as compared with selective activation of PLCγ. Similarly, the addition of wortmannin or LY294002 to cells expressing the wild-type PDGFR inhibited the release of intracellular calcium. Thus, generation of PtdIns-3,4,5-P3by receptor-associated PI3K causes an increase in IP3production and intracellular calcium release, potentially via enhanced PtdIns-4,5-P2 substrate availability due to PtdIns-3,4,5-P3-mediated recruitment of PLCγ to the lipid bilayer.

Activation of the receptor-associated phosphoinositide 3-kinase (PI3K) 1 has been shown to cause mitogenesis and enhanced cell motility, although the exact mechanism by which PI3K mediates cell signaling during these events has been difficult to elucidate (1)(2)(3). The lipid products of PI3K have now been found to activate certain calcium-independent protein kinases C and to bind to a subset of Src homology 2 (SH2) domains (4,5). In addition, PtdIns-3,4-P 2 and/or PtdIns-3,4,5-P 3 has been found to bind to and/or stimulate several pleckstrin homology (PH) domain-containing proteins, including the Akt/PKB serine/threonine protein kinase (6,7), the PDK serine/threonine protein kinase (8), and the Grp1 exchange factor for Arf1 (9). Falasca et al. (10) have demonstrated that the PH domain of phospholipase C␥ will bind to PtdIns-3,4,5-P 3 , targeting PLC␥ to the membrane.
Recently, Bae et al. (11) have found that the addition of PtdIns-3,4,5-P 3 can enhance phospholipase C␥-mediated Pt-dIns-4,5-P 2 hydrolysis in vitro and that overexpression of a constitutively active form of the p110 catalytic subunit of PI3K increases intracellular IP 3 levels, raising the possibility that PtdIns-3,4,5-P 3 may regulate calcium signaling as well. This possibility is supported by the observation that wortmannin, an inhibitor of the catalytic site of PI3K, as well as several related enzymes, diminishes the intracellular calcium transient seen in adrenal glomerulosa cells, neutrophils, and rat leukemia cells following stimulation (12)(13)(14)(15).
To determine whether phospholipase C␥ might interact with the lipid products of PI3K in a manner capable of modifying ligand-dependent IP 3 generation and calcium signaling, we examined PtdIns-3,4,5-P 3 binding to the SH2 domains of PLC␥ and found that PtdIns-3,4,5-P 3 associates with high affinity with the C-terminal SH2 (CT-SH2) domain of PLC␥ and is capable of displacing the CT-SH2 domain of PLC␥ from the activated PDGFR. We demonstrate that activation of a PDGFR mutant that selectively activates PLC␥ (Y1021) results in a substantially diminished intracellular calcium release when compared with coactivation of PI3K and PLC␥ by the wild-type PDGFR. In addition, inhibition of PtdIns-3,4,5-P 3 generation by either wortmannin or LY294002 partially inhibits PDGFmediated calcium release by the wild-type PDGFR, but has no effect on calcium release initiated by the Y1021 PDGFR.

PtdIns-3,4,5-P 3 Direct Binding Assay-Glutathione S-transferase
(GST) fusion proteins were expressed in Escherichia coli and extracted as described (16). Synthetic water-soluble 3 H-labeled dioctanoyl (C 8 )-PtdIns-3,4,5-P 3 (5 ϫ 10 7 cpm/mol) (17) was incubated with Sepharose beads containing ϳ0.3 nmol of GST-CT-SH2 domain fusion protein or GST alone in HNE buffer (30 mM Hepes, pH 7.0, 100 mM NaCl, and 1 mM EDTA) containing 0.02% Nonidet P-40. After 1 h at room temperature, the beads were separated from the supernatant by centrifugation. The supernatant was then collected, mixed with scintillation liquid, and counted on a Beckman counter. * This work was supported by National Institutes of Health Grants DK48871 (to L. G. C.), GM48339 (to A. K.), and GM41890 (to L. C. C.) and by the Medical Foundation of Massachusetts (to L. E. R.). 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.
PDGFR Binding Assay-GST-CT-SH2 domain-containing beads were incubated with the appropriate concentration of PtdIns-4,5-P 2 or PtdIns-3,4,5-P 3 (sonicated in HNE buffer) for 1 h at room temperature. NIH 3T3 cells were stimulated with PDGF for 5 min and lysed in HNE buffer containing 0.5% Nonidet P-40. Lysates were centrifuged for 10  min at 14,000 ϫ g, and supernatants were added to the GST-CT-SH2 domain/lipids for 10 min. Beads were washed three times with HNE buffer containing 0.5% Nonidet P-40, resuspended in an equal volume of 2ϫ Laemmli buffer, and incubated for 5 min at 100°C. Proteins were resolved by 7.5% SDS-polyacrylamide gel electrophoresis, transferred to nitrocellulose membrane, blotted using anti-Tyr(P) antibody (4G10), and visualized using chemiluminescence detection (DuPont). The autoradiograms were quantified using the NIH Image program.
PDGFR Mutants-These clones are as described previously (18 -21). Briefly, HepG2 cells were transfected with PDGFR constructs encoding the wild-type PDGFR; a mutant that selectively excludes PI3K binding (F740/751); a mutant that excludes PI3K, RasGAP, and SHP-2 binding, but associates with PLC␥ (Y1021); a mutant that selectively associates with PI3K (Y740/751); or a mutant that excludes binding to all four signaling molecules (F5). Cell lines were fluorescence-activated cellsorted to normalize receptor expression and maintained in Dulbecco's modified Eagle's medium with 10% fetal calf serum.
Immunoprecipitation and Western Analysis-Subconfluent cells were washed twice with phosphate-buffered saline and refed with Dulbecco's modified Eagle's medium in the absence of fetal bovine serum (Life Technologies, Inc.). Following 24 h of serum deprivation, cells were treated with either PDGF (100 ng/ml; Upstate Biotechnology, Inc.) or vehicle control for 2 min, washed twice with ice-cold phosphate-buffered saline, and lysed in ice-cold lysis buffer (see above). Cells were vortexed vigorously and centrifuged for 10 min at 12,000 ϫ g. One-mg aliquots were immunoprecipitated overnight at 4°C with anti-phosphotyrosine antibody (Upstate Biotechnology, Inc.) using protein A-Sepharose beads (Sigma). Beads were washed three times with phosphate-buffered saline, 1% Nonidet P-40, and 2 mM sodium vanadate; two times with 0.1 M Tris, 0.5 M LiCl, and 2 mM sodium vanadate, pH 7.5; and two times with 10 mM Tris, 100 mM NaCl, 1 mM EDTA, and 2 mM sodium vanadate, pH 7.5.
The final pellet was boiled for 5 min in the presence of ␤-mercaptoethanol and separated by SDS-polyacrylamide gel electrophoresis. Proteins were electrophoretically transferred to Immobilon-P membranes (Millipore Corp.). Blots were probed for anti-phosphotyrosine, anti-p85 (generous gift of Brian Schaffhausen), or anti-PLC␥ (Santa Cruz Biotechnology) antibody. The blots were exposed to the appropriate horseradish peroxidase-linked secondary antibody prior to detection of proteins using a chemiluminescence system (ECL, Amersham International).
Quantitation of Intracellular IP 3 Levels-Cells were plated at a concentration of 400,000 cells/well on 12-well plates and incubated for 48 h in inositol-free medium supplemented with 5 Ci/ml [ 3 H]inositol (Amersham Pharmacia Biotech). Cells were then washed twice with phosphate-buffered saline and stimulated for 30 s with PDGF (100 ng/ml), and the reaction was stopped by the addition of ice-cold 10% trichloroacetic acid. Cells were lysed by sonication for 30 s, extracted three times FIG. 3. Ability of PtdIns-4,5-P 2 and PtdIns-3,4,5-P 3 to block binding of the phosphorylated PDGFR to the PLC␥ C-terminal SH2 domain. PLC␥1 GST-CT-SH2 domain-containing beads were incubated with increasing concentrations of PtdIns-4,5-P 2 (PI4, 5P 2 ) or PtdIns-3,4,5-P 3 (PI3, 4, 5P 3 ), followed by incubation with lysates from PDGF-stimulated cells. The amount of PDGFR bound to the CT-SH2 domain was analyzed by Western blotting using anti-Tyr(P) antibody (top panel) and quantified as described under "Materials and Methods" (bottom panel). The lines drawn represent a best fit to the data assuming a single class of binding sites (see "Materials and Methods").
with water-saturated ether, and filtered, and [ 3 H]inositol polyphosphates were analyzed by anion-exchange HPLC, using a Partisphere SAX column (Whatman), as described previously (2). All experiments were performed in triplicate.
Intracellular Calcium Determination-HepG2 cells were maintained in Dulbecco's modified Eagle's medium/nutrient mixture F-12 supplemented with 10% FCS. Cells were loaded with Fura-2/AM (Molecular Probes, Inc.) as follows. Nearly confluent 150-mm 2 plates of cells were incubated for 1 min in 5 ml of HBBS/EDTA buffer (HBBS buffer ϭ 118 mM NaCl, 4.6 mM KCl, 10 mM glucose, and 20 mM Hepes, pH 7.2, prepared in calcium-free water; HBBS/EDTA buffer ϭ HBBS ϩ 0.02% EDTA). HBBS/EDTA buffer was then aspirated, and cells were incubated for 1 min at 37°C, followed by scraping in 25 ml of HBBS/CaCl 2 buffer (HBBS buffer ϩ 1 mM CaCl 2 ). Cells were pelleted at 1000 ϫ g for 10 min and resuspended in 5 ml of HBBS/CaCl 2 buffer. Fura-2/AM was then added at a final concentration of 1 mM, and cells were incubated in the dark for 40 min at 37°C. Cells were then washed three times with HBBS/CaCl 2 buffer and resuspended in HBBS/CaCl 2 buffer at a concentration of 1 ϫ 10 6 cells/ml. Intracellular free calcium measurements were performed in a constantly stirred 37°C cuvette using a dual-beam fluorescence spectrophotometer (Spex AR-CM). For experiments performed in the absence of extracellular calcium, 10 mM EGTA was added at the beginning of the experiment, followed by a 60-s equilibration period. Controls were performed for each condition with addition of vehicle alone. Minimal fluorescence and maximal fluorescence were determined for each experiment by adding digitonin (final concentration ϭ 50 mM) and CaCl 2 (final concentration ϭ 11 mM), respectively. Curves shown were generated by determining 5-s [Ca 2ϩ ] i values using a best fit of the raw data. Results are expressed as mean Ϯ S.E.

PtdIns-3,4,5-P 3 Can Bind to the CT-SH2 Domain of PLC␥ and Mediate
Dissociation from the PDGFR-To determine whether PtdIns-3,4,5-P 3 might interact with PLC␥ in vivo, binding of the SH2 domains of PLC␥ to PtdIns-3,4,5-P 3 was examined in vitro. Water-soluble C 8 -PtdIns-3,4,5-P 3 bound to the recombinant C-terminal SH2 domain of PLC␥1 with a K d of 2.4 M and a stoichiometry of ϳ1 mol/mol (Fig. 1). In contrast, the N-terminal SH2 domain did not significantly associate with C 8 -PtdIns-3,4,5-P 3 under these conditions, whereas a construct including both SH2 domains had an affinity for PtdIns-3,4,5-P 3 that was similar to that of the C-terminal SH2 domain alone (data not shown).
Since SH2 domains mediate the interaction of PLC␥ with phosphotyrosine residues, we postulated that PtdIns-3,4,5-P 3 might compete for PLC␥ binding to the phosphorylated PDGFR. To examine this, the CT-SH2 domain of PLC␥1, immobilized on beads, was preincubated with increasing concentrations of either PtdIns-3,4,5-P 3 or PtdIns-4,5-P 2 . PDGF-stimulated NIH 3T3 cell lysates were added to the beads, and the ability of the SH2 domain to associate with the PDGFR was determined. PtdIns-3,4,5-P 3 blocked the association of PLC␥ with the phosphorylated PDGFR in a dose-dependent fashion with a K i(app) of ϳ9 M (Fig. 3). The higher K i(app) for PtdIns-3,4,5-P 3 binding in this circumstance as compared with competition with [ 3 H]dioctanoyl-PtdIns-3,4,5-P 3 (Fig. 2) is likely due to the high affinity of the phosphorylated PDGFR for the CT-SH2 domain of PLC␥. In agreement with the relative affinities observed in Fig. 2, PtdIns-4,5-P 2 competed with the PDGFR for SH2 domain binding with a 3.6 times lower affinity than did PtdIns-3,4,5-P 3 .
Activation of Receptor Mutants with PDGF-To examine the in vivo effect of the lipid products of PI3K on PLC␥ signaling, we utilized PDGFR mutants stably expressed in HepG2 cells. To ensure that the PDGFR mutants were similarly expressed and recruiting the appropriate downstream signaling proteins, anti-phosphotyrosine immunoprecipitates of PDGF-stimulated HepG2 cells were examined for co-immunoprecipitation of phospholipase C␥ and PI3K (Fig. 4). Phosphorylation of the PDGFR in response to added PDGF was nearly equivalent for all cell lines (top panel). As expected from previous studies, PLC␥ associated with the F740/751, Y1021, and wild-type receptors in a ligand-dependent fashion, but did not associate with the F5 receptor (middle panel). Interestingly, the association of PLC␥ with the Y1021 mutant was increased as compared with the wild-type receptor (see "Discussion"). The p85 subunit of PI3K associated only with the wild-type receptor (bottom panel) and with the Y740/751 mutant (data not shown). Thus, activation of only the wild-type receptor resulted in recruitment of both signaling molecules.

Intracellular IP 3 Generation in Response to PDGF Is Enhanced by Coactivation of PLC␥ and PI3K-HepG2 cells ex-
pressing either the wild-type PDGFR or the F740/751 receptor were loaded with [ 3 H]inositol for 48 h, and total IP 3 generation in response to PDGF was determined by HPLC analysis. A time course of IP 3 generation in cells expressing the wild-type PDGFR revealed that total IP 3 began increasing 15 s following PDGF stimulation and peaked at 30 s (Fig. 5A). The addition of PDGF to cells expressing the F740/751 mutant receptor resulted in a lesser increase in intracellular total IP 3 as compared with the wild-type receptor (Fig. 5B, 1762 Ϯ 201 dpm for the F740/751 mutant versus 3028 Ϯ 148 dpm for the wild-type receptor at 30 s following stimulation; p Ͻ 0.01), with the increase in IP 3 diminished by ϳ40%.
Intracellular Calcium Release in Response to PDGF Is Enhanced by Coactivation of PLC␥ and PI3K-HepG2 cells were loaded with Fura-2 as described, and extracellular calcium was removed by the addition of excess extracellular EGTA. Resting intracellular calcium in all clones in the absence of extracellu- WT, wild-type PDGFR. F740/751 is a receptor that selectively excludes PI3K binding; Y1021 is a receptor that selectively binds PLC␥; and F5 is a mutant that excludes binding to PI3K, PLC␥, RasGAP, and SHP-2 (see "Materials and Methods"). lar calcium was ϳ15 nM (Fig. 6) and remained stable for Ն5 min. The addition of PDGF to HepG2 cells expressing the wild-type PDGFR resulted in a peak intracellular free calcium concentration of 93 Ϯ 6 nM ϳ75 s following stimulation (Fig. 6,  closed circles). In cells expressing either the Y1021 mutant (which selectively activates PLC␥) or the F740/751 mutant (which activates all of the signaling pathways except PI3K), the peak intracellular calcium response to PDGF was significantly diminished (51 Ϯ 7 and 58 Ϯ 7 nM, respectively) and delayed (105 and 90 s, respectively). Cells expressing the Y740/751 mutant (which selectively binds PI3K, but not PLC␥, RasGAP, or SH-PTP2) or the F5 mutant (which binds none of these signaling enzymes) demonstrated no release of intracellular calcium in response to PDGF. Thus, coactivation of PLC␥ and PI3K resulted in a 1.5-2 fold increase in intracellular free calcium release as compared with selective activation of PLC␥.
Lipid Products of PI3K Mediate Enhanced Calcium Signaling-The in vivo effect of the interaction between PtdIns-3,4,5-P 3 and PLC␥ was examined by inhibition of PI3K-dependent PtdIns-3,4,5-P 3 generation by the HepG2 mutant receptor cell line. The addition of 100 nM wortmannin to HepG2 cells (which results in inhibition of 80% of the immunoprecipitable PI3K activity in these cells) expressing the wild-type PDGFR resulted in a 24% inhibition of the calcium transient (Fig. 7), whereas wortmannin had no effect on the calcium transient in either the Y1021 or F740/751 mutant. The specificity of this effect of wortmannin for tyrosine kinase receptor-mediated coactivation of PI3K and PLC␥ was confirmed by comparison with the calcium transient induced following fetal calf serum addition (Fig. 8). Fetal calf serum contains lysophosphatidic acid, which causes a rapid calcium transient (peak [Ca 2ϩ ] i at ϳ10 s) secondary to release of intracellular calcium stores following GTP-dependent phosphoinositide hydrolysis (22,23). Wortmannin inhibited the intracellular calcium release following PDGF addition while having no effect on calcium release in response to FCS.
Based on our observation that 100 nM wortmannin only partially inhibited PI3K activity in our cells, we also examined the effects of the alternative PI3K inhibitor LY294002 on HepG2 cells expressing the wild-type PDGFR (Fig. 9). In these experiments, wortmannin again inhibited the peak [Ca 2ϩ ] i by 25%, whereas 50 M LY294002 inhibited the maximal intracellular calcium release by 47%, consistent with the 40 -50% lower calcium release detected in cells expressing the Y1021 PDGFR mutant (Figs. 6 and 7). Similarly, in NIH 3T3 fibroblasts expressing the wild-type PDGFR, LY294002 inhibited the intracellular calcium transient following PDGF stimulation better than did wortmannin (⌬[Ca 2ϩ ] i ϭ 39.0 Ϯ 1.9 nM following  7. Wortmannin effects on PDGF-stimulated calcium release in PDGFR mutants. Peak intracellular calcium release was determined following stimulation of PDGFR mutant clones with PDGF in the presence of wortmannin (100 nM) or vehicle (Me 2 SO). Wortmannin had no effect on the stimulated calcium release in Y1021-or F740/751expressing cells, but inhibited the calcium release in cells expressing the wild-type PDGFR by 24% (n ϭ 6; p Ͻ 0.05%). f, base line (ϩEGTA); _, PDGF; Ⅺ, PDGF ϩ wortmannin. PDGF, 31.0 Ϯ 4.7 nM following PDGF ϩ wortmannin, and 18.7 Ϯ 5.0 nM following PDGF ϩ LY294002). Thus, inhibition of PtdIns-3,4,5-P 3 generation using PI3K inhibitors prevented the enhanced release of intracellular calcium following PDGF stimulation of cells that coactivate PI3K and PLC␥ and had no effect on the lesser calcium response in the Y1021 and/or F740/ 751 cells, which activate PLC␥ without coactivation of PI3K.

DISCUSSION
The present data demonstrate that the PDGF-dependent coactivation of phospholipase C␥ and the phosphoinositide 3-kinase results in enhanced intracellular generation of IP 3 and an augmented intracellular calcium transient as compared with those observed with selective activation of PLC␥, suggesting that recruitment of active PI3K to the PDGFR results in increased hydrolytic activity of PLC␥. The importance of PtdIns-3,4,5-P 3 in this PI3K-dependent increase in intracellular calcium release is demonstrated by two observations. First, the inhibition of intracellular calcium release following the addition of wortmannin or LY294002 (Figs. 7-9) (10, 12-15) argues that PtdIns-3,4,5-P 3 is upstream of PLC␥ activation since wortmannin and LY294002 inhibit PtdIns-3,4,5-P 3 production without preventing PI3K association with the receptor. 2 Second, the observation by Bae et al. (11)that constitutive expression of the p110 catalytic subunit of PI3K can cause sustained increases in cellular IP 3 levels, without the necessity of receptor activation, also supports the importance of local generation 2 The degree of inhibition of calcium release and/or IP 3 generation by wortmannin appears to be cell type-dependent, with marked inhibition of calcium release in COS-1 cells expressing the PDGFR, granulosa cells, and leukemia cells (10,12,15) and essentially no inhibition of A31 cells (L. G. Cantley, personal observation). This could be due to differences in the importance of PtdIns-3,4,5-P 3 for PLC␥ activation by different receptors or to different cell sensitivities to wortmannin inhibition.

FIG. 8. Wortmannin effects on PDGF-versus FCS-stimulated calcium release in HepG2 cells expressing the wild-type PDGFR.
Peak intracellular calcium release was determined following stimulation of HepG2 cells expressing the wild-type PDGFR with either 100 ng/ml PDGF or 1% fetal calf serum in the absence of extracellular calcium. The peak calcium release occurred at ϳ75 s following PDGF addition (see Fig. 6) and ϳ10 s following FCS addition. The PDGF-dependent intracellular calcium increase was inhibited by 38% in the presence of wortmannin, whereas the FCS-mediated calcium transient was not affected (n ϭ 7; p ϭ 0.02 for PDGF versus PDGF ϩ wortmannin and p ϭ 0.002 for PDGF ϩ wortmannin versus FCS ϩ wortmannin). of PtdIns-3,4,5-P 3 in enhancing PLC activity.
Recent data demonstrating that PtdIns-3,4,5-P 3 can directly associate with SH2 domain-containing proteins and regulate their activity therefore led us to examine PtdIns-3,4,5-P 3 association with the SH2 domains of PLC␥1. The in vitro data presented demonstrate that PtdIns-3,4,5-P 3 binds to the Cterminal SH2 domain of PLC␥1 with an affinity similar to that which we have observed with PtdIns-3,4,5-P 3 binding to the C-terminal SH2 domain of p85 PI3K and ϳ3-fold weaker than its binding to the PH domain of BTK (24) and Grp1 (25).
In a related series of experiments, Bae et al. (11) have found that PtdIns-3,4,5-P 3 can directly activate PLC␥ hydrolytic activity in vitro and that this activation can be blocked by the addition of isolated SH2 domains of PLC␥. In addition, the PH domain of PLC␥ may be involved in the association of PtdIns-3,4,5-P 3 with the enzyme. Data from the laboratory of Dr. Joseph Schlessinger (10) demonstrate that the N-terminal PH domain of PLC␥ can bind PtdIns-3,4,5-P 3 and that mutations of this domain diminish recruitment of PLC␥ to the cell membrane, raising the possibility that simultaneous binding of the PH and SH2 domains to PtdIns-3,4,5-P 3 may be involved in enzyme activation. In contrast, the split PH domain of PLC␥ does not appear to bind PtdIns-3,4,5-P 3 (11).
We postulate that the mechanism by which the PI3K mediates enhanced activation of PLC␥ may directly relate to the SH2 domain/PtdIns-3,4,5-P 3 interaction. As seen in Fig. 3, the association of PtdIns-3,4,5-P 3 with the C-terminal SH2 domain of PLC␥ prevents association of the SH2 domain with the tyrosine-phosphorylated PDGFR. Dissociation of PLC␥ from the receptor due to local production of PtdIns-3,4,5-P 3 could thus allow translocation of the enzyme to the lipid bilayer, increasing its substrate availability. Since PtdIns-3,4,5-P 3 itself is not hydrolyzed by PLC␥, its presence in the lipid bilayer would also serve to stabilize PLC␥ at that site, via association with the SH2 domain, the amino-terminal PH domain, or both domains simultaneously. Coactivation of PLC␥ and PI3K by the wild-type PDGFR, with subsequent dissociation of PLC␥ from the activated receptor due to local production of PtdIns-3,4,5-P 3 , might also explain the observation of greater association of PLC␥ with the Y1021 PDGFR mutant as compared with the wild-type receptor (Fig. 4).
In our experiments, PtdIns-3,4,5-P 3 prevented binding of the PLC␥ CT-SH2 domain to the phosphorylated PDGFR with only a 3.6-fold higher affinity than did PtdIns-4,5-P 2 . The low total cellular concentration of PtdIns-3,4,5-P 3 as compared with Pt-dIns-4,5-P 2 thus raises the question of the physiologic relevance of this observation. It is likely, we believe, that recruitment of active PI3K to the receptor results in the local conversion of much of the PtdIns-4,5-P 2 into PtdIns-3,4,5-P 3 , resulting in a reversal of these concentrations in the immediate vicinity of the receptor. We postulate that the generation of PtdIns-3,4,5-P 3 as a high affinity "anchor" for PLC␥ at the membrane is therefore important for stabilization of the enzyme in the vicinity of its substrate.
It is conceivable that PtdIns-3,4,5-P 3 production at the membrane could recruit non-phosphorylated PLC␥ to this site (via association with the PLC␥ SH2 and/or PH domain) and thereby trigger PtdIns-4,5-P 2 hydrolysis in the absence of PLC␥ recruitment to receptors. This is not the case in all circumstances, however, since activation of neither the Y740/751 PDGFR mutant (Fig. 6) nor the insulin receptor (26), which activate PI3K but not PLC␥, will cause an intracellular calcium transient.
The present results, along with the recent observations that PtdIns-3,4-P 2 and/or PtdIns-3,4,5-P 3 can activate the serine/ threonine protein kinases Akt/PKB and PDK1 (3) and that PtdIns-3,4,5-P 3 can stimulate the Arf1 nucleotide exchange protein, Grp1 (25), suggest that PI3K may regulate multiple signaling pathways via PtdIns-3,4-P 2 -and PtdIns-3,4,5-P 3 -dependent interactions with the SH2 and/or PH domains of proteins within those pathways. We have previously shown that either selective activation of PI3K or the direct addition of PtdIns-3,4,5-P 3 to epithelial cells is capable of initiating cell motility, at least in part by activation of calcium-independent protein kinases C (4,27). Our present data suggest that receptor-mediated activation of PI3K could result in enhanced activation of calcium-dependent protein kinase C isoforms by enhancing PLC␥ substrate hydrolysis and therefore IP 3 production and intracellular calcium release. The possibility that the magnitude and/or duration of intracellular calcium transients can moderate cell fate has recently been confirmed by the work of Dolmetsch et al. (28), in which it was shown that activation of NF-B and c-Jun N-terminal kinase occurred with rapid high calcium transients, whereas NFAT was preferentially activated by a low, more sustained calcium release.
The present data demonstrate that coactivation of PI3K and PLC␥ results in an increase in IP 3 generation and intracellular calcium release. The ability of PtdIns-3,4,5-P 3 to bind to the CT-SH2 domain of PLC␥ and to compete for its association with the phosphorylated PDGFR suggests that the mechanism of the increased calcium signaling involves PtdIns-3,4,5-P 3 -mediated recruitment and/or stabilization of PLC␥ at the lipid bilayer.