The Role of the Pleckstrin Homology Domain in Membrane Targeting and Activation of Phospholipase C b 1 *

Current studies involve an investigation of the role of the pleckstrin homology (PH) domain in membrane targeting and activation of phospholipase C b 1 (PLC b 1 ). Here we report studies on the membrane localization of the isolated PH domain from the amino terminus of PLC b 1 (PLC b 1 -PH) using fluorescence microscopy of a green fluorescent protein fusion protein. Whereas PLC b 1 -PH does not localize to the plasma membrane in serum-starved cells, it undergoes a rapid but transient migration to the plasma membrane upon stimulation of cells with serum or lysophosphatidic acid (LPA). Regulation of the plasma membrane localization of PLC b 1 -PH by phosphoinositides was also investigated. PLC b 1 -PH was found to bind phosphatidylinositol 3-phosphate most strongly, whereas other phosphoinositides were bound with lower affinity. The plasma membrane localization of PLC b 1 -PH induced by serum and LPA was blocked by wortmannin pretreatment and by LY294002. In parallel, activation of PLC b by LPA was inhibited by wortmannin, by LY294002,

. This hydrolysis is catalyzed by the phosphoinositidespecific phospholipase C enzymes, which release two intracellular second messengers, inositol 1,4,5-trisphosphate (Ins(1,4,5)P 3 ) and diacylglycerol (3,4). Phosphoinositide-specific phospholipase Cs have been grouped by sequence into three major families, PLC␤, PLC␦, and PLC␥ (5). PLC␤ is stimulated by activation of heterotrimeric G proteins, whereas PLC␥ is activated by receptor-tyrosine protein kinases. The mechanism of PLC␦ regulation is presently not clear. All mammalian PLC␤s have a modular domain organization, containing an amino-terminal pleckstrin homology (PH) domain, four elongation factor hand motifs, a catalytic X/Y domain, a C2 domain, and a carboxyl-terminal extension that is unique to the PLC␤ family. PLCs in all three classes have a PH domain at their amino terminus. In the cases of PLC␦ and PLC␥, this PH domain appears to be essential for membrane association and for efficient activity in cellular membranes (6 -8), although it is not required for catalytic activity in vitro (6).
Phosphatidylinositol 3-kinases (PI 3-kinases) are involved in signaling pathways that lead to cell growth and transformation (9). Both growth factor activation and oncogenic transformation of cells leads to the formation of the lipid products of this enzyme, phosphatidylinositol 3,4-bisphosphate (PtdIns(3,4)P 2 ) and phosphatidylinositol 3,4,5-trisphosphate (PtdIns(3,4,5)P 3 ) (10 -12). PtdIns(3,4)P 2 and PtdIns(3,4,5)P 3 have recently been shown to act as "lipid second messengers" by binding to the PH domains of several different signaling proteins. For example, the PH domains of Bruton's tyrosine kinase (Btk), the Ras GTPase activating protein Gap1, and the ARF exchangers Grp1, cytohesin-1, and ARNO have all been shown to bind PtdIns(3,4,5)P 3 with high affinity (13)(14)(15)(16). The serine-threonine kinase Akt/PKB has been found to be activated and recruited to the plasma membrane by PtdIns(3,4,5)P 3 (17). Akt/ PKB activation in response to insulin or PDGF requires both its PH domain and PI 3-kinase activity (18,19). Growth factorinduced activation of PI 3-kinases leads to an increase in the levels of 3-phosphorylated phosphoinositides that bind specifically to its PH domain, consequently recruiting the protein to the plasma membrane (20). Similarly, we have previously shown that the amino-terminal PH domain from PLC␥ 1 is targeted to the plasma membrane in a manner dependent on PI 3-kinase activity (7). Another 3-phosphoinositide, phosphatidylinositol 3-phosphate (PtdIns(3)P), has recently been shown to have a potential role as a second messenger, specifically in vesicular transport (21)(22)(23). In fact, recent data have revealed that a conserved zinc binding domain, the FYVE finger (named after the first letter of four proteins containing it) binds specifically to PtdIns(3)P (21,22). FYVE domains are found in several proteins, some of which have been shown to be involved in membrane trafficking (23). In the present study, we investigate the role of the PH domain of PLC␤ 1 (PLC␤ 1 -PH) in plasma membrane binding and enzyme activation. We show that PLC␤ 1 -PH binds to PtdIns(3)P and that the G␤␥ subunit is involved in the signal-dependent membrane localization of PLC␤ 1 -PH. Therefore, we conclude that the interaction of these two ligands with its PH domain mediates the specific membrane association and activation of PLC␤ 1 .

EXPERIMENTAL PROCEDURES
Construction of Expression Plasmids-The cDNA fragments corresponding to the PH domains from rat PLC␤ 1 (residues 6 -154), rat PLC␥ 1 (residues 14 -150), and rat PLC␦ 1 (residues 11-140) were amplified using the polymerase chain reaction (PCR). The PCR products were digested with BamHI alone (for PLC␥ 1 -PH) or BamHI plus EcoRI (for PLC␤ 1 -PH and PLC␦ 1 -PH), depending on the sites present in the oligonucleotides used for PCR, and were ligated into the appropriately digested pGEX-2T or pGEX-2TK bacterial expression vector to direct expression of the GST-PH domain fusion proteins. The DNA sequence of each PH domain insert was verified by dideoxynucleotide sequencing.
GFP Expression Vectors-cDNA encoding the different PH domains was subcloned (in frame with GFP) into the GFP fusion protein expression vector pEGFP-C1 (CLONTECH), using the BseAI and EcoRI sites for PLC␤ 1 -PH, BglII, and BamHI sites for PLC␥ 1 -PH and the BglII and EcoRI sites for PLC␦ 1 -PH. The sequence of each insert was verified by dideoxynucleotide sequencing.
Expression and Purification of PH Domains-The pGEX-2TK plasmid containing the PH domains was transformed into Escherichia coli DH5␣. Cells were grown to an A 600 of 0.3-0.4 and induced with 1 mM isopropyl-1-thio-␤-D-galactopyranoside (Roche Molecular Biochemicals) for 3 h at 37°C (for PLC␦ 1 -PH) or for 4 -7 h at 25°C (for PLC␥ 1 -PH and PLC␤ 1 -PH). After pelleting, cells were lysed by sonication in 50 mM Tris, pH 8.0, 100 mM NaCl, 10% glycerol, 1 mM dithiothreitol, and protease inhibitors (1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, and 10 g/ml aprotinin). Triton X-100 was added to 1% in the lysate, and particulates were removed by centrifugation for 20 min at 16,000 rpm in an SS34 (Sorvall) rotor. The clarified lysate was incubated with glutathione-agarose beads (Sigma) for 1 h at 4°C, washed three times with ice-cold PBS containing 1 mM dithiothreitol plus protease inhibitors, and the protein bound to the glutathione-agarose beads was stored at 4°C. For microinjection studies, the protein was eluted from the beads by incubation with 15 mM reduced glutathione in PBS, and the eluted protein was buffer-exchanged into microinjection buffer (PBS) using a Centricon-10 (Amicon).

Analysis of [ 3 H]Inositol Phosphate and Glycerophosphoinositide
Binding to PH Domains-Mixtures of standard [ 3 H]inositol phosphates (NEN Life Science Products, 21 Ci/mmol; ARC Inc., St. Louis), used at the same concentrations, were added to the GST-PH fusion proteins (50 g) on glutathione-agarose beads. After incubation at 25°C for 10 min, the beads were washed 3 times with 10 mM Tris, pH 7.4, and the bound counts eluted with 0.1 N HCl. The eluted material was analyzed by strong anion exchange (SAX) HPLC as described previously (25), using a Partisil 10 SAX analytical column, eluting with a shallow gradient from 0 to 1 M (NH 4 ) 2 HPO 4 (pH 3.80). Radioactivity in the eluate was monitored with an on-line radioactivity flow detector (Packard FLO ONE A-525). Quantitative analysis was obtained by calculating the amount of radioactivity under the peaks, and using the same protein concentrations and specific activities of the different [ 3 H]inositols. This assay was initially tested by using several PH domains, and other putative lipid binding domains, derived from different proteins, and gave very reproducible results consistent with those obtained with other binding assays. In addition, with several PH domains, such as those from pleckstrin and dynamin, we did not find any binding at all (as for GST alone), indicating that only specific binding was detected.
Cell Culture and Transfection-HeLa and COS-7 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS). Cells were plated onto glass coverslips and starved in 0.5% FBS for 16 -24 h in DMEM. The transfection of GFP fusion protein vectors was done with cells grown on glass coverslips in 25-cm dishes using 2 mg of plasmid DNA and 5 ml of LipofectAMINE (Life Technologies, Inc.), according to the manufacturer's instructions. Following transfection, cells were serum-starved in DMEM with 0.5% FBS overnight.
Confocal Microscopy-GFP fusion proteins were visualized following fixation of cells with 4% paraformaldehyde in PBS for 10 min at room temperature using an upright SARASTRO 2000 CLSM confocal microscope (Molecular Dynamics). For living cells experiments, cells were plated onto 25-mm diameter circular glass coverslips. Cells were examined in an inverted microscope under a ϫ 63 oil immersion objective (Carl Zeiss, Jena GmbH D) and a Bio-Rad laser confocal microscope system (MRC-1000) connected to an Axiovert 100 microscope (Zeiss). Confocal images were taken at every 60 s using a Kalman filter (n ϭ 2). The laser potency, photomultiplier, and pin-hole size were kept constant during the whole experiment.
Analysis of PLC Activity-For inositol metabolite evaluation, cells grown in 6-well plates were labeled for 24 h in medium 199 containing myo-[ 3 H]inositol (6 Ci/ml). Following this labeling the cells were washed once with PBS and preincubated for 10 min in 2 ml of PBS containing 10 mM LiCl (pH 7.4) at 37°C, prior to the addition of the agonist. The cells were then extracted with methanol/HCl (1 N)/chloroform (1:1:1) and analyzed by HPLC as described previously (25).
Microinjection-NIH 3T3 cells were plated at a density of 3.5 ϫ 10 4 onto 13-mm glass coverslips. The following day, cells were serumstarved in 0.5% calf serum for 12 h in DMEM. The cDNAs of GFP-PLC␤ 1 -PH, and G␤ 1 and/or G␥ 2 subunits in PBS were then microinjected into the cells in HEPES-buffered culture medium, using a Zeiss (Oberkochen, Germany) Axiovert IM 35 microscope and an Eppendorf (Hamburg, Germany) micromanipulator. The cDNA of the G protein subunits (␤ 1 ␥ 2 ) were kindly provided by Prof. P. Gierschik (Ulm, Germany). The concentration of microinjected GST-PLC␤ 1 -PH fusion protein was 0.5-1.0 mg/ml. After microinjection NIH 3T3 cells were incubated in DMEM at 37°C. Microinjected cells were identified by using fluorescein isothiocyanate-dextran.

RESULTS
The PH Domain of PLC␤ 1 Is Localized to the Plasma Membrane-Previous studies have demonstrated that the binding of phosphoinositides to the PH domains of the PLC␥ and PLC␦ isozymes leads to their membrane association and plays a role in enzyme activation (6 -8). To investigate the function of the PH domain from the other family of PLC isozymes, PLC␤, the cellular localization of the amino-terminal PH domain from PLC␤ 1 (PLC␤ 1 -PH), was explored by fusing it to green fluorescent protein (GFP). The GFP-PLC␤ 1 -PH fusion protein was transiently overexpressed in COS-7 and HeLa cells, and the distribution of the fluorescent label was examined by confocal fluorescence microscopy. These studies showed that the protein is homogenously distributed in the cell cytoplasm in the absence of serum (Fig. 1A), whereas in the presence of serum it is concentrated at the plasma membrane, as shown in Fig. 1B for HeLa cells and in COS-7 cells (data not shown). This result provides evidence that PLC␤ 1 -PH localizes to the plasma membrane, at least in cells that have been subjected to constant serum stimulation. PLC␤ 1 -PH Membrane Localization Is Serum-and LPA-dependent-In order to understand the mechanism of PLC␤ 1 -PH binding to the plasma membrane, and to test its signal dependence, localization of PLC␤ 1 -PH was studied in serum-starved COS-7 and HeLa cells. As shown in Fig. 2A, the protein is homogenously distributed in the COS-7 cell cytoplasm in the absence of serum stimulation. However, upon stimulation of the cells with either LPA (Fig. 2B) or serum (data not shown), PLC␤ 1 -PH is rapidly localized to the plasma membrane. These findings argue that the plasma membrane localization of PLC␤ 1 -PH occurs only in activated cells, as previously reported for PLC␥ 1   Similar results were obtained in HeLa cells (data not shown).
The nuclear staining present in both unstimulated and stimulated cells was also found with the empty vector and therefore represents nonspecific staining due to the small size of the fusion protein that permits its diffusion through the nuclear pores. To characterize further whether the PLC␤ 1 -PH could localize to the nucleus, we microinjected a GST fusion protein of PLC␤ 1 -PH into NIH 3T3 cells. As shown in Fig. 3 no nuclear localization was observed either in serum-starved cells (Fig.  3A) or in serum-stimulated cells (Fig. 3B), whereas in serumstimulated cells a clear plasma membrane localization was observed (Fig. 3B).
As there can be questions about data selection or variability between controls and samples, we have also monitored the same cells before and after the addition of serum. As shown in Fig. 4, serum-starved COS-7 cells did not show any membrane localization. Serum addition induced a clear translocation of PLC␤ 1 -PH to the plasma membrane that appears after 2 min of stimulation and reaches its maximum after around 5 min; this localization then gradually disappeared and was totally redistributed into the cytoplasm after 10 min of serum addition.
Specific Binding of PLC␤ 1 -PH to PtdIns(3)P-To investigate the interactions of PLC␤ 1 -PH with phosphoinositides, a GST-PLC␤ 1 -PH fusion protein was generated and purified. We have recently developed a "lipid blotting" approach for screening lipid extracts and mixtures for potential PH domain ligands (24). Different lipids are spotted onto nitrocellulose, and the dried nitrocellulose is blocked as described (24), without detergent, and is probed with the 32 P-labeled GST fusion proteins of PH domains that have been labeled by protein kinase A phosphorylation of a protein kinase A site included in the construct. We used this approach to analyze phosphoinositide binding by PLC␦ 1 -PH and PLC␤ 1 -PH. As shown in Fig. 5, we loaded the different phosphoinositides onto nitrocellulose and probed with GST and GST fused to PLC␦ 1 -PH and PLC␤ 1 -PH. GST-PLC␦-PH, our positive control, binds most strongly to PtdIns(4,5)P 2 (Fig. 5A), whereas GST alone did not show any binding (data not shown). Interestingly, GST-PLC␤ 1 -PH binds specifically to PtdIns(3)P (Fig. 5B).
Since this assay does not allow a quantitative comparison, we further investigated the phosphoinositide binding to PLC␤ 1 -PH by using a different assay. We incubated a GST fusion of PLC␤ 1 -PH bound to glutathione beads with phospholipid extracts from cells labeled with [ 3 H]inositol. In order to eliminate artifacts due to unspecific micelle aggregation, and since binding of phosphoinositides to most PH domains occurs via the polar head group, we deacylated the cell extracts, obtaining the respective glycerophosphoinositols (gPtdIns(3)P, etc.). After a brief incubation and the washing off of nonspecifically bound counts, the radioactivity associated with the GST fusion proteins was analyzed by HPLC. With GST alone, no radioactivity was detected (data not shown). When PLC␤ 1 -PH was incubated with the cell extracts, the product that it associated with to the greatest extent was gPtdIns(3)P ( Fig. 6A). A much smaller degree of interaction was seen with the other glycerophosphoinositides. The results obtained were consistent with those obtained using the lipid dot-blot assay. These data suggest that the isolated PLC␤ 1 -PH specifically interacts with PtdIns(3)P and binds more weakly to PtdIns(4,5)P 2 and PtdIns(3,4,5)P 3 . Similar data were obtained when the GST fusion of PLC␤ 1 -PH was incubated with a mixture of [ 3 H]inositol phosphate standards ( Fig. 6B; see "Experimental Procedures") and by using a centrifugation assay for PH domain binding to lipid vesicles (data not shown; see "Experimental Procedures"). 1 -PH showed that stimulation of cells with serum or PDGF causes an immediate change in the intracellular localization of this PH domain, from the cytosol to the plasma membrane (7), and that this localization is induced by the formation of PtdIns(3,4,5)P 3 upon stimulation of PI 3-kinase. Similar findings have been reported for Btk, cytohesin-1, ARNO, Grp1, and Gap1 (16,24,26,27). To determine whether PI 3-kinase is important for membrane recruitment of PLC␤ 1 -PH, we incubated COS-7 and HeLa cells with the PI 3-kinase inhibitors wortmannin (100 nM) and LY294002 (25 M). In unstimulated cells transfected with GFP-PLC␤ 1 -PH, only cytosolic fluorescence was seen ( Fig.  2A). LPA stimulation resulted in a clear localization of PLC␤ 1 -PH to the plasma membrane of COS-7 cells (Fig. 7A). However, when cells were pretreated with wortmannin or LY294002 (data not shown) for 10 min, no membrane localization was observed upon LPA stimulation (Fig. 7B). These results strongly support the idea that the PLC␤ 1 -PH plasma membrane localization induced by LPA involves PI 3-kinase.

Effects of PI 3-Kinase Inhibitors on PLC␤ 1 -PH Plasma Membrane Localization-Our previous work with PLC␥
Effects of the G Protein ␤␥ Subunit on PLC␤ 1 -PH Plasma Membrane Localization-Previous studies have demonstrated that the PH domain of PLC␤ 1 binds to the G␤␥ subunit and may play a role in the anchoring of the protein to the membrane surface (28). In order to test whether the G␤␥ subunit interaction with PLC␤ 1 is involved in the signal-dependent recruitment of PLC␤ 1 to the plasma membrane, we microinjected DNA of the G␤ and the G␥ subunits together with GFP-PLC␤ 1 -PH into NIH 3T3 cells. As shown in Fig. 8A, no FIG. 5. Dot-blot assay comparing phosphoinositide-binding specificities of PH domains. 32 P-Labeled GST-PH domain fusion proteins were used to probe nitrocellulose filters that had been spotted with specific phosphoinositides as described under "Experimental Procedures." GST alone did not give any signal above background. The amount of lipid spotted in each row was 0.2, 0.4, 0.6, 0.8, and 1.0 g. specific localization was observed in serum-starved cells microinjected with GFP-PLC␤ 1 -PH together with the G␤ subunit alone, whereas when the G␤␥ subunits were microinjected GFP-PLC␤ 1 -PH was recruited to the plasma membrane (Fig.  8B). Interestingly, pretreatment of NIH 3T3 cells with wortmannin completely abolished the effect of the G␤␥ subunit on GFP-PLC␤ 1 -PH translocation (Fig. 8C).

Effect of the PH Domain on PLC␤ 1 Enzyme Activity-We
have recently shown that the PH domain of PLC␥ 1 binds most strongly to PtdIns(3,4,5)P 3 and that PI 3-kinase activity plays a role in PLC␥ 1 activation and membrane localization (7). To investigate how suppression of PI 3-kinase affects PLC␤ activity, Ins(1,4,5)P 3 formation was examined in COS-7 cells stim- . The beads were washed three times, and the glycerophosphoinositols or inositols that remained associated were eluted with 0.1 N HCl and analyzed by HPLC as described previously (25). The amount of radioactivity of each compound was calculated as the percentage of the total radioactivity present in the extract and then represented as the percent of the PtdIns(3)P bound. Under the same conditions, no association of radioactivity was detected when GST alone was used. The data shown are means of three independent experiments, with error bars indicating standard deviations.  ulated with LPA (25 M; 3 min). As shown in Fig. 9A, pretreatment of COS-7 cells with 100 nM wortmannin or 25 M LY294002 inhibited LPA-induced Ins(1,4,5)P 3 formation by more than 50%. In order to analyze further the role of the PH domain in PLC␤ activation, the effect of overexpression of PLC␤ 1 -PH on Ins(1,4,5)P 3 formation induced by LPA was also investigated. The data presented in Fig. 9B shows that LPAinduced Ins(1,4,5)P 3 production is reduced by around 50% in cells that overexpress PLC␤ 1 -PH. In contrast, no effect was seen when PLC␥ 1 -PH or PLC␦ 1 -PH were overexpressed under the same conditions, indicating that this effect is specific to PLC␤ 1 -PH. Taken together, these results indicate that by binding to PtdIns(3)P the PLC␤ 1 PH domain may play a role in signal-dependent recruitment of the enzyme to the cell surface.

LPA Induces PtdIns(3)P Production in COS-7 and HeLa
Cells-LPA stimulates PI 3-kinase activity in diverse cell types. To test the hypothesis that LPA regulates PLC-␤ 1 activity by inducing PtdIns(3)P formation, we determined the time course for the activation of PI 3-kinase in LPA-stimulated COS-7 and HeLa cells by measuring the accumulation of labeled 3-phosphorylated phosphoinositides.
As shown in Fig. 10, PtdIns(3)P increases by about 50% after stimulation of COS-7 cells with LPA; this peak value is reached within 2 min and decreases thereafter with time (data not shown). Although modest (about 50% over basal), this increase is significant and could account for a role in PLC-␤ 1 activation. DISCUSSION Recent studies have demonstrated the importance of PI 3-kinase activation in PH domain-mediated translocation of a number of signaling proteins, such as PLC␥ 1 , Btk, Akt, Grp1, etc. (7,(13)(14)(15)(16). Other findings have suggested that PH domains of some proteins can bind to membrane proteins such as the ␤␥ subunits of heterotrimeric G proteins (29). The data reported here demonstrate that the activity of PLC␤ 1 is also regulated by PI 3-kinase, possibly by its product PtdIns(3)P, and by the G␤␥ subunits. PI 3-kinase can phosphorylate the 3 position of inositol in PtdIns, PtdIns(4)P, or PtdIns(4,5)P 2 in vitro, although in vivo PtdIns(4,5)P 2 seems to be its preferred substrate (30). Agonistinduced PI 3-kinase activity leads to the rapid accumulation of intracellular PtdIns(3,4,5)P 3 , with the other 3-phosphorylated lipids arising primarily as breakdown products of PtdIns(3,4,5)P 3 . These data might be interpreted as meaning that PtdIns(3)P is a degradation product from the primary, agonist-sensitive PI 3-kinase activity and that it is unlikely to have any functional significance. However, recent results suggest that this is an over-simplification. First, a series of PtdInsspecific PI 3-kinases have been characterized in yeast and in mammalian cells (31). Second, in human platelets it has been recently reported that integrin-dependent signaling activates the transient formation of PtdIns(3)P, followed by the generation of PtdIns(3,4)P 2 , but not PtdIns(3,4,5)P 3 (32,33). This is the first evidence for stimulatory control of PtdIns(3)P formation in any cell. Finally, PtdIns(3)P has been reported to bind an important component of clathrin-coated structures, known as the adaptor complex AP-2, and leads to a significant enhancement in its capacity to recognize tyrosine-based endocytic signals (34). More recently, it was also shown that the human early-endosomal autoantigen EEA1 binds to PtdIns(3)P through a protein domain called the FYVE domain (21,22). Taken together, these data indicate that PtdIns(3)P might play a role in signal-induced membrane targeting of proteins, as already reported for PtdIns(3,4,5)P 3 . PLC␤ 1 binds with relatively high affinity to membranes, and it has been proposed that the membrane localization of PLC␤ isozymes is likely to be mediated by both the COOH-terminal domain and the PH domain that could bind PtdIns(4,5)P 2 as reported for PLC␦ 1 (35). Direct measurements of the affinities of the mammalian PLC␤ or PLC␥ for PtdIns(4,5)P 2 have shown that these isozymes did not bind the polar head group of PtdIns(4,5)P 2 with a specificity comparable to that of PLC␦ 1 (7,36,37). In fact, the amino acid residues responsible for PtdIns(4,5)P 2 binding to the PLC␦ 1 PH domain are not well conserved in other PLC isozymes. Therefore, the affinities of the other PLC isoforms for PtdIns(4,5)P 2 appear too weak to account for membrane translocation; hence, an additional membrane component is required. It is well known that the carboxyl-terminal portion of PLC␤ 1 is a critical region required for activation by G proteins (38,39). Our data suggest that the amino-terminal PH domain of PLC␤ 1 represents another region responsible for enzyme regulation. Thus, as reported for the PH domain of PLC␥ 1 , which is likely to cooperate with the SH2 domain in signal-dependent membrane recruitment of the intact molecule (7,40), a similar scenario could be depicted in the case of PLC␤ 1 .
Interestingly, a similar mechanism that we propose for the PLC␤ 1 PH domain has already been suggested for the ␤-adrenergic receptor kinase (41). In fact, the PH domain of ␤-adrenergic receptor kinase has been shown to interact with both G␤␥ and PtdIns(4,5)P 2 at the carboxyl and amino termini, respectively, of the PH domain (41). The interaction of these two ligands with the intact enzyme is required for its membrane association and activation. The evidence that phosphoinositides and G␤␥ bind to different parts of the PH domain suggests a cooperative, and not a competitive, role of their interaction (30). The potentially cooperative binding of multiple ligands to PH domains may give a significance to some weak interactions found in the case of some PH domains and phosphoinositides and membrane proteins that might display higher affinity in the presence of both. This cooperative mechanism could also reveal a functional role giving a specificity to the localization of specific proteins. The data presented demonstrate that the PH domain of PLC␤ 1 localizes to the plasma membrane, although we cannot rule out a specific localization to internal compartments under different circumstances. Thus, a mechanism already shown for EEA1 can be hypothesized, where PtdIns(3)P together with the small G protein Rab5 induces the localization of the EEA1 to the Golgi compartment (42).
The data that the PLC␤ 1 PH domain binds to PtdIns(3)P introduces the novelty that a PH domain can bind specifically to PtdIns(3)P and that this interaction occurs at the plasma membrane. The localization of the PLC␤ 1 PH domain to the plasma membrane and its displacement by pretreatment with wortmannin (known to reduce the level of PtdIns(3)P) suggest not only that a PH domain may bind specifically to this lipid but also that a portion of the PtdIns(3)P could be constitutively present on the plasma membrane or, alternatively, be produced by ligand stimulation as our data might suggest. The PI 3-kinase involved in PLC␤-PH translocation is probably an enzyme that specifically produces PtdIns(3)P and is distinct from the class I PI 3-kinase that produces mainly PtdIns(3,4)P 2 and PtdIns(3,4,5)P 3 . In fact, it is well known that PDGF induces increases in PtdIns(3,4)P 2 and PtdIns(3,4,5)P 3 levels but not that of PtdIns(3)P (30).
In summary, our studies using the PH domain of PLC-␤ 1 indicate that LPA-stimulated inositol lipid turnover is dependent on a cooperative mechanism involving PtdIns(3)P and G␤␥ activation.