Activation of phospholipase C-gamma by phosphatidylinositol 3,4,5-trisphosphate.

Signal transduction across cell membranes often involves the activation of both phosphatidylinositol (PI)-specific phospholipase C (PLC) and phosphoinositide 3-kinase (PI 3-kinase). Phosphatidylinositol 4,5-bisphosphate (PI(4,5)P2), a substrate for both enzymes, is converted to phosphatidylinositol 3,4, 5-trisphosphate (PI(3,4,5)P3) by the action of PI 3-kinase. Here, we show that PI(3,4,5)P3 activates purified PLC-gamma isozymes by interacting with their Src homology 2 domains. Furthermore, the expression of an activated catalytic subunit of PI 3-kinase in COS-7 cells resulted in an increase in inositol phosphate formation, whereas platelet-derived growth factor-induced PLC activation in NIH 3T3 cells was markedly inhibited by the specific PI 3-kinase inhibitor LY294002. These results suggest that receptors coupled to PI 3-kinase may activate PLC-gamma isozymes indirectly, in the absence of PLC-gamma tyrosine phosphorylation, through the generation of PI(3,4,5)P3.

The phosphoinositides PI(3,4)P 2 and PI(3,4,5)P 3 are not substrates of any known PLC (8) and are normally absent from resting cells; however, they appear within seconds to minutes of stimulation of cells with various growth factors or other cellular activators. In contrast, the concentration of phosphatidylinositol 3-phosphate does not change substantially in response to cell stimulation (4 -7). It has thus been suggested that PI(3,4)P 2 and PI(3,4,5)P 3 might function as intracellular messengers (4 -7). With regard to potential targets of these D-3-phosphorylated lipids, they have been shown to activate Ca 2ϩ -independent isoforms of protein kinase C (9, 10) as well as to bind the pleckstrin homology (PH) domain of the protein serine-threonine kinase Akt, thereby activating its kinase activity (11)(12)(13), and to the SH2 domains of the 85-kDa (p85) subunit of PI 3-kinase, thereby preventing its binding to tyrosine-phosphorylated proteins (14).
The 10 mammalian PLC isozymes identified to date are single polypeptides and can be divided into three types: PLC-␤, PLC-␥, and PLC-␦ (1). All contain a PH domain in their NH 2terminal region. The ␥ type isozymes differ from the other two types in that they contain two SH2 domains, one SH3 domain, and an additional PH domain that is split by the SH domains; these domains are arranged in the order PH(N)-SH2-SH2-SH3-PH(C), where N and C in parentheses denote NH 2 -and COOHterminal locations, respectively. Upon stimulation of cells with growth factors like platelet-derived growth factor (PDGF) and epidermal growth factor, the SH2 domain of PLC-␥ binds to the autophosphorylated tyrosine residues of growth factor receptors, leading to tyrosine phosphorylation and activation of PLC-␥ (1). PLC isozymes can also be activated at least in vitro by the presence of phosphatidic acid (15) or arachidonic acid (16). Therefore, activation of PLC-␥ isozymes may occur secondarily to receptor-mediated activation of phospholipase D and cytosolic phospholipase A 2 , which results in the production of phosphatidic acid and arachidonic acid, respectively. We now report that another lipid-derived messenger, PI(3,4,5)P 3 , activates PLC-␥ isozymes specifically by interacting with their SH2 domains.

EXPERIMENTAL PROCEDURES
Materials-PLC isozymes were purified from HeLa cells that had been transfected with recombinant vaccinia virus containing the entire coding sequence of the respective enzyme (17). Dipalmitoyl-PI(3,4)P 2 and dipalmitoyl-PI(3,4,5)P 3 were synthesized as described (18). An expression vector (pCMVp110-CAAX) that encodes a fusion protein consisting of Myc epitope and p110-CAAX was kindly provided by J. Downward (Imperial Cancer Fund, London, United Kingdom).
Preparation of GST Fusion Proteins-For each fusion protein (denoted PH(N)-SH2-SH2-SH3-PH(C), SH2-SH2-SH3, SH2-SH2, N-SH2, C-SH2, PH(NC), and SH3), the corresponding polymerase chain reaction product, flanked by BamHI and EcoRI linkers, was inserted into the BamHI and EcoRI sites of the glutathione S-transferase (GST) expression vector pGEX-2TK (Pharmacia Biotech Inc.). Amino acid sequences of PH(N)-SH2-SH2-SH3-PH(C), SH2-SH2-SH3, SH2-SH2, N-SH2, C-SH2, and SH3 correspond to residues 483-936, 533-851, 550 -745, 550 -657, 668 -745, and 758 -851 of PLC-␥1, respectively. Polymerase chain reaction products corresponding to residues 482-523 (PH(N)) and 865-936 (PH(C)) of PLC-␥1 were fused with a glycine codon insertion to yield the combined PH(NC) construct. The 5Ј and 3Ј primers for PH(N) contained BamHI and SmaI sites, respectively; thus, the PH(NC) construct contained sequential BamHI, SmaI, and EcoRI sites. Escherichia coli cells were transformed with the various expression vectors and cultured at 30°C. Expression of the GST fusion proteins was induced with 0.1 mM isopropyl-1-thio-␤-D-galactopyranoside, and the cells were subsequently collected by centrifugation at 2000 ϫ g for 15 min, sonicated in phosphate-buffered saline, and centrifuged at 5000 ϫ g for 15 min. The resulting supernatant was mixed and incubated at room temperature for 30 min with 2 ml of a 50% (v/v) slurry of glutathione-Sepharose 4B (Pharmacia) that had been equilibrated with phosphatebuffered saline. After centrifugation of the mixture at 5000 ϫ g for 15 min, the supernatant was removed and the pellet washed with 10 bed volumes of phosphate-buffered saline. Bound proteins were cleaved from GST by incubation of the beads with thrombin (10 g/ml) at room temperature for 6 h. The eluted proteins were further purified by high performance liquid chromatography on a Mono Q column and quantitated spectrophotometrically with extinction coefficients at 280 nm calculated on the basis of their amino acid composition.
To determine whether activation of PLC by PI(3,4,5)P 3 could be detected in intact cells, we transiently expressed in COS-7 cells the 110-kDa subunit (p110) of PI 3-kinase with c-Myc epitope and farnesylation signal (CAAX) sequences located at the NH 2 and COOH termini, respectively (21), and measured the release of inositol phosphates resulting from the hydrolysis of PI(4,5)P 2 . Expression of the Myc-tagged p110-CAAX protein was detected by immunoblot analysis with antibodies specific to the Myc sequence (Fig. 3A). The farnesylation signal se- quence causes the constitutive activation of p110 by targeting it to the cell membrane. As expected from the fact that most mammalian cells contain a relatively high concentration (Ͼ20 M) of inositol phosphates even before stimulation (22), a substantial amount of 3 H-labeled inositol phosphates was detected in COS-7 cells labeled with [ 3 H]inositol to equilibrium (Fig.  3B). Expression of Myc-p110-CAAX induced a 45% increase in the amount of inositol phosphates, and this effect was blocked by pretreatment of cells with LY294002, a specific inhibitor of PI 3-kinase (23). LY294002 had no effect on the amount of inositol phosphates in cells not expressing Myc-p110-CAAX. These results suggest that PI(3,4,5)P 3 generated by the activated p110 subunit was able to activate PLC. The increase in PLC activity was smaller than that in cells stimulated with PLC-activating agonists, probably because in the local area where PLC-␥ was activated by PI(3,4,5)P 3 , PI(4,5)P 2 (a common substrate for PLC and PI 3-kinase) had been depleted by Myc-p110-CAAX. It has been shown previously that availability of PI(4,5)P 2 is a limiting factor for PLC activity (24).
Further evidence that PI(3,4,5)P 3 activates PI(4,5)P 2 hydrolysis in intact cells is provided by previous studies with wortmannin, a potent (median inhibitory concentration (IC 50 ), 3 nM), irreversible inhibitor of PI 3-kinase (23). Wortmannin inhibits I(1,4,5)P 3 formation and Ca 2ϩ mobilization in bovine adrenal glomerulosa cells stimulated by angiotensin II (25), rat basophilic leukemia (RBL-2H3) cells stimulated by cross-linking of high affinity immunoglobulin E (IgE) receptors (26), and human neutrophils stimulated by fMLP (27). At the time of the studies with neutrophils and adrenal glomerulosa cells, wortmannin was known to inhibit myosin light chain kinase (MLCK), but its effect on PI 3-kinase was not known. The inhibition of I(1,4,5)P 3 and Ca 2ϩ responses by 20 nM wortmannin in adrenal glomerulosa cells was thus speculated to result from inhibition of MLCK. However, it is now known that the concentration of wortmannin required for inhibition of MLCK is 100 times that required for inhibition of PI 3-kinase (23), so that MLCK was likely not inhibited by 20 nM wortmannin. Our data indicate that they are likely attributable to prevention of PI(3,4,5)P 3 -induced activation of PLC-␥. PLC-␥1 is a widely expressed and abundant enzyme, whereas PLC-␥2 is abundant in cells of hematopoietic origin.
Recently, type III PI 4-kinase was shown to be inhibited by wortmannin (IC 50 , 50 nM) (23). Because PI 4-kinase is required for PI(4,5)P 2 synthesis, it is possible that the reduced activity of PLC observed in wortmannin-treated cells was attributable to the diminished supply of substrate. Thus, we studied the effect of the more specific inhibitor LY294002 (IC 50 : 2 and 100 M for PI 3-kinase and type III PI 4-kinase, respectively; Ref. 23) on PI(4,5)P 2 hydrolysis induced by PDGF in NIH 3T3 cells. Whereas PDGF induced an 8-fold increase in PLC activity in control cells, pretreatment of cells with 10 M LY294002 reduced this response by 40% (Fig. 4A). The presence of the inhibitor affected neither the basal amount of inositol phosphates (Fig. 4A) nor the tyrosine phosphorylation of PLC-␥1 and PDGF receptor (Fig. 4B). In a similar experiment with COS-7 cells, epidermal growth factor induced a 3-fold increase in PLC activity and pretreatment of cells with 10 M LY294002 reduced the response by 34% (data not shown).
More than 60 different receptors are known to stimulate PLC (1-3). Although only 30 receptors are currently known to stimulate PI 3-kinase (4 -7), most of these also activate PLC. Therefore, the amount of I(1,4,5)P 3 generated in a variety of cells in response to an agonist may reflect the sum of that produced as a result of direct activation of PLC and that produced by indirect activation of PLC-␥ through the PI 3-kinase pathway. Given that LY294002, at a concentration sufficient for inhibition of PI 3-kinase but not for inhibition of PI 4-kinase, inhibits 40% of PDGF-dependent PLC activity, the ratio of direct to indirect activation appears to be 3 to 2 in NIH 3T3 cells stimulated with PDGF. The indirect activation of PLC-␥ is likely not limited to PTK receptors such as the PDGF receptor and COS-7 cells were plated at a density of 3 ϫ 10 5 cells/well in six-well plates and allowed to recover for 24 h. The cells were then incubated for 6 h with 1.3 g of empty (pCMV) or Myc-p110-CAAX-encoding (pC-MVp110-CAAX) plasmid and 6.7 l of LipofectAMINE in 1 ml of Opti-MEM medium (Life Technologies, Inc.), after which 1 ml of Dulbecco's modified Eagle's medium (DMEM) supplemented with 20% fetal bovine serum was added to each well and the cells were incubated for an additional 24 h. The medium was then aspirated, and cells were incubated for 16 h in myo-inositol-free DMEM supplemented with [ 3 H]myoinositol (1 Ci/ml, 25 mCi/mmol) (DuPont Biotechnology Systems). The cells were then incubated in the absence or presence of 10 M LY294002 for 45 min, followed by additional incubation in DMEM supplemented with 20 mM LiCl for 60 min. The incubation was terminated by adding perchloric acid to a final concentration of 5% (w/v). Cells were scraped into Eppendorf tubes and subjected to centrifugation. A, the resulting pellets were subjected to immunoblot analysis with antibodies to Myc (Santa Cruz Biotechnology). Arrow indicates Myc-p110-CAAX protein. B, the amount of total [ 3 H]inositol phosphates in the cell supernatants was measured as described (34). Data are expressed as a percentage of the values for pCMV-transfected cells not treated with LY294002 and are means Ϯ S.E. from three independent experiments. Fig. 3, and then incubated first in the absence or presence of 10 M LY294002 for 45 min and then with or without PDGF (50 ng/ml) for 10 min in the presence of 20 mM LiCl. Total [ 3 H]inositol phosphates were then measured as described in Fig.  3. Data are expressed as a percentage of the value for cells not exposed to LY294002 or PDGF and are means Ϯ S.E. from three independent experiments. B, NIH 3T3 cells were incubated in the absence or presence of 10 M LY294002 for 45 min and then with or without PDGF (50 ng/ml) for 10 min. Cell lysates were subjected to immunoblot analysis with antibodies to PLC-␥1 (␣PLC-␥1, left panel) or with antibodies to phosphotyrosine (␣PY, right panel). The positions of PLC-␥1 as well as tyrosine-phosphorylated PLC-␥1 and PDGF receptor (PDGFR) are indicated.

FIG. 4. Effect of LY294002 on the PDGF-induced production of inositol phosphates in NIH 3T3 cells. A, NIH 3T3 cells were labeled with [ 3 H]myo-inositol as described in
PTK-coupled receptors such as the IgE receptor. The fact that wortmannin inhibits PLC activation induced by G proteincoupled receptors such as those for angiotensin II and fMLP suggests that PLC-␥ might be activated indirectly in response to the occupancy of such receptors. However, the activation of PI 3-kinase appears not always to result in activation of PLC-␥, because there is no evidence that insulin or colony-stimulating factor-1, both of which activate PI 3-kinase, elicits the production of I(1,4,5)P 3 (31,32). One possible explanation is that PLC-␥ isozymes may not be located at the sites where PI(3,4,5)P 3 is generated in the cells activated with colonystimulating factor or insulin.
Finally, our results may explain how engagement of the Fc⑀RIIB inhibitory receptor in mast cells reduces the IgEinduced increase in intracellular Ca 2ϩ concentration (33). Fc⑀RIIB binds SHIP, a phosphatase that dephosphorylates the D-5 position of PI(3,4,5)P 3 or I(1, 3,4,5)P 4 . A decrease in the concentration of PI(3,4,5)P 3 could reduce I(1,4,5)P 3 production by PLC-␥ isozymes and thereby reduce long term release of intracellular Ca 2ϩ and ultimately Ca 2ϩ influx.