Phosphatidylinositol 4,5-bisphosphate mediates Ca2+-induced platelet alpha-granule secretion: evidence for type II phosphatidylinositol 5-phosphate 4-kinase function.

To understand the molecular basis of granule release from platelets, we examined the role of phosphatidylinositol 4,5-bisphosphate (PtdIns(4,5)P(2)) in alpha-granule secretion. Streptolysin O-permeabilized platelets synthesized PtdIns(4,5)P(2) when incubated in the presence of ATP. Incubation of streptolysin O-permeabilized platelets with phosphatidylinositol-specific phospholipase C reduced PtdIns(4,5)P(2) levels and resulted in a dose- and time-dependent inhibition of Ca(2+)-induced alpha-granule secretion. Exogenously added PtdIns(4,5)P(2) inhibited alpha-granule secretion, with 80% inhibition at 50 microm PtdIns(4,5)P(2). Nanomolar concentrations of wortmannin, 33.3 microm LY294002, and antibodies directed against PtdIns 3-kinase did not inhibit Ca(2+)-induced alpha-granule secretion, suggesting that PtdIns 3-kinase is not involved in alpha-granule secretion. However, micromolar concentrations of wortmannin inhibited both PtdIns(4,5)P(2) synthesis and alpha-granule secretion by approximately 50%. Antibodies directed against type II phosphatidylinositol-phosphate kinase (phosphatidylinositol 5-phosphate 4-kinase) also inhibited both PtdIns(4,5)P(2) synthesis and Ca(2+)-induced alpha-granule secretion by approximately 50%. These antibodies inhibited alpha-granule secretion only when added prior to ATP exposure and not when added following ATP exposure, prior to Ca(2+)-mediated triggering. The inhibitory effects of micromolar wortmannin and anti-type II phosphatidylinositol-phosphate kinase antibodies were additive. These results show that PtdIns(4,5)P(2) mediates platelet alpha-granule secretion and that PtdIns(4,5)P(2) synthesis required for Ca(2+)-induced alpha-granule secretion involves the type II phosphatidylinositol 5-phosphate 4-kinase-dependent pathway.

␣-Granules are the dominant platelet secretory granules. They contain many components that have been implicated in thrombosis and atherosclerosis, including adhesion molecules, coagulation factors, soluble mediators of inflammation, and growth factors (1). The molecular mechanisms that mediate the release of these components are not well defined. This secretory process is tightly controlled to prevent unregulated release of thrombogenic factors. In many respects, platelets represent a unique secretory system. They are anucleate, contain an open canalicular system that undergoes eversion following platelet activation (2), and demonstrate homotypic granule fusion (3). Ultrastructural studies have shown that the ␣-granules of platelets are secreted primarily via fusion with the surfaceconnected open canalicular system (1) following apparent centralization of granules and microtubule reorganization. Such observations have lead to speculation that cytoskeletal reorganization is responsible for ␣-granule release. However, inhibition of actin polymerization (4) or microtubule organization (5,6) does not inhibit granule secretion. Furthermore, granule secretion and shape change can be dissociated under several experimental conditions (7)(8)(9). More recently, SNARE 1 proteins have been demonstrated in platelets (10 -12) and shown to mediate platelet ␣-granule secretion (12)(13)(14)(15). Purified SNARE proteins are capable of fusing lipid membranes in vitro (16,17). However, permeabilized cell systems require molecules other than SNARE proteins to secrete granules. Furthermore, regulated secretion necessitates that the SNARE protein apparatus respond to activation-dependent signals (18). Thus, an understanding of the molecular mechanisms of platelet granule secretion requires the identification of molecules that are both modified upon platelet activation and capable of interacting directly with the secretory machinery.
Phospholipid metabolism has long been regarded as an essential component of regulated secretion in platelets. Initial studies performed in intact (19) and electropermeabilized (20) platelets indicated a role for the products of phosphatidylinositol-specific PLC, diacylglycerol and inositol 1,4,5-trisphosphate, in stimulating granule secretion (21). However, both dense and ␣-granule secretion can be stimulated in the absence of PLC activation, suggesting that PLC activity is not essential to platelet granule secretion (22)(23)(24). Subsequent studies in Ca 2ϩ -independent permeabilized secretory systems showed that the generation of phosphatidic acid by phospholipase D correlates closely with secretion (23,24). Inhibition of phospholipase D, however, fails to block Ca 2ϩ -independent dense granule secretion, and phosphatidic acid itself cannot stimulate dense granule secretion (25). These observations support a modulatory role of phospholipase D in platelet granule secretion. Thus, no product of phospholipid metabolism has been demonstrated to be fusogenic in platelets (25), and the role of phospholipid metabolism in platelet membrane fusion remains to be defined. Phosphoinositide phosphorylation has been invoked in vesicle fusion models and shown to mediate granule secretion. Initial observations in chromaffin cells demonstrated that the maintenance of polyphosphoinositides is crucial for vesicle secretion (26). Subsequent studies revealed a PtdIns 4-kinase activity associated with granules in chromaffin, mast, and pancreatic ␤-cells (27). In chromaffin cells, this PtdIns 4-kinase activity correlates with granule secretion (28). In PC12 cells, phosphatidylinositol transfer protein was determined to be one of three cytosolic factors required to restore Ca 2ϩ -induced secretion from semi-intact PC12 cells (29). Type I PIPK is a second cytosolic factor necessary to reconstitute secretion (30). Subsequent to these studies, phosphatidylinositol transfer protein was found to reconstitute Ca 2ϩ -induced, GTP␥S-dependent granule secretion from SL-O-permeabilized HL-60 cells (31). In RBL-2H3 mast cells, ADP-ribosylation factor-1 was demonstrated to stimulate exocytosis via PtdIns(4,5)P 2 synthesis (32). These data suggest a central role for PtdIns(4,5)P 2 in regulated secretion.

EXPERIMENTAL PROCEDURES
Chemicals and Reagents-All buffer constituents, solvents, ATP, CaCl 2 , PtdIns, PtdIns 4-phosphate, PtdIns(4,5)P 2 , and PtdIns-specific PLC (from Bacillus cereus) were purchased from Sigma. Sepharose 2B was obtained from Amersham Pharmacia Biotech. Reduced SL-O was purchased from Corgenix (Peterborough, England). Wortmannin and LY294002 were purchased from Calbiochem. Phycoerythrin-conjugated anti-P-selectin antibody AC1.2 was purchased from Becton Dickinson (San Jose, CA). [ 32 P]Orthophosphoric acid was obtained from PerkinElmer Life Sciences. Silica Gel G TLC plates were obtained from Whatman Ltd. (Kent, England). GST-tagged type II PIPK was generously provided by Dr. Lucia Rameh. All solutions were prepared using water purified by reverse-phase osmosis on a Millipore Milli-Q purification water system.
Antibodies-Affinity-purified mouse monoclonal antibodies to PtdIns 3-kinase p85␣ (clone 8 -2D-4D) were obtained from Lab Vision (Fremont, CA). Affinity-purified rabbit anti-peptide polyclonal antibodies to type II PIPK were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). The anti-amino-terminal end antibody was raised against a peptide consisting of the 19 amino-terminal amino acids of type II␣ PIPK. The anti-carboxyl-terminal end antibody was raised against a peptide consisting of the 18 carboxyl-terminal amino acids of type II␣ PIPK.
Platelet Preparation-Blood from healthy donors who had not ingested aspirin in the 2 weeks prior to donation was collected by veni-puncture into 0.4% sodium citrate. Citrate-anticoagulated blood was centrifuged at 200 ϫ g for 20 min to prepare platelet-rich plasma. Platelets were then purified from platelet-rich plasma by gel filtration using a Sepharose 2B column equilibrated in PIPES/EGTA buffer (25 mM PIPES, 2 mM EGTA, 137 mM KCl, 4 mM NaCl, and 0.1% glucose, pH 6.4). Final gel-filtered platelet concentrations were 1-2 ϫ 10 8 platelets/ml.
Permeabilization of Platelets-Platelets were permeabilized using reduced SL-O. The ability of each batch of SL-O to permeabilize platelets was tested by analyzing for incorporation of fluorescein isothiocyanate-dextran sulfates by flow cytometry as described previously (12,40). Briefly, gel-filtered platelets (20 l) were incubated with 25 M fluorescein isothiocyanate-dextran sulfate of various molecular masses in the presence or absence of SL-O for 15 min. Platelets were subsequently analyzed for fluorescence by flow cytometry. An increase in fluorescence in the samples exposed to SL-O compared with non-permeabilized samples confirmed permeabilization.
Analysis of P-selectin Surface Expression-For analysis of P-selectin surface expression from permeabilized platelets, 20 l of gel-filtered platelets (1-2 ϫ 10 8 /ml) in the indicated concentration of MgATP were incubated with the concentrations of reduced SL-O indicated in the figure legends. Samples were adjusted to pH 6.9 immediately following permeabilization. The timing of addition of inhibitors varied according to the inhibitor that was being evaluated and is indicated in the figure legends. Following the incubation with inhibitor, CaCl 2 was added to the reaction mixture. The amount of CaCl 2 required to give a free Ca 2ϩ concentration of 10 M in the presence of 2 mM EGTA at pH 6.9 was calculated for each condition using a computer program (gift from Dr. P. J. Padfield) based on the algorithms described by Fabiato and Fabiato (41). Following an additional incubation after the addition of Ca 2ϩ , 10 l of reaction mixture were transferred to 5 l of phycoerythrin-conjugated anti-P-selectin antibody AC1.2. Phosphate-buffered saline (500 l) was added to the sample after a 20-min incubation, and the platelets were analyzed immediately by flow cytometry as described below.
Analysis of Platelet Phosphoinositides-Gel-filtered platelets (350 l/sample) were incubated in the presence of 2 mCi/ml [ 32 P]orthophosphoric acid at 37°C for 2 h. Unbound 32 P was separated from platelets by sequential centrifugation. Platelets were then permeabilized with the indicated concentrations of SL-O in the presence or absence of ATP. Samples were adjusted to pH 6.9 immediately following permeabilization. The timing of addition of inhibitors varied according to the inhibitor and is indicated in the figure legends. Platelets were subsequently solubilized in 250 l of a 20:40:1 solution of chloroform, methanol, and 12 N HCl and mixed vigorously. Chloroform (75 l) was added to this solution. The solution was mixed vigorously, and its phases were separated at 1300 ϫ g for 10 min at 4°C. The upper phase was discarded, and the lower phase was washed three times with 625 l of the 20:40:1 solution of chloroform, methanol, and 12 N HCl. Each extract (5 l) was applied to one lane of a 20 ϫ 20-cm Silica Gel G plate previously soaked in 1% potassium oxalate for 30 min and then baked for 5 min at 100°C. Phospholipids were separated by chromatography in 1-propanol and 2 M acetic acid (65:35). The location of PtdIns 4-phosphate and PtdIns(4,5)P 2 was determined by applying 5 g of known standards to each lane. Standards were detected by staining in a saturated iodine tank. The radioactivity on the plates was detected using a Bio-Rad GS-525 molecular imaging system. Phosphoinositide levels were quantified both by calculating the pixel density of bands comigrating with PtdIns standards using Molecular Analyst software (Bio-Rad) and by extracting the radioactivity comigrating with the PtdIns standards from the plate and quantifying radioactivity using a Tri-Carb 2100TR liquid scintillation analyzer (Packard Instrument Co.). Under the conditions of our assay, a pixel density of 1 corresponds to 10 cpm.
Flow Cytometry-Flow cytometry was performed on gel-filtered platelet samples using a Becton Dickinson FACSCalibur flow cytometer. Fluorescent channels were set at logarithmic gain. 10,000 particles were acquired for each sample. A 530/30-nm band-pass filter was used for FL-2 fluorescence. Phycoerythrin was measured in the FL-2 channel. Data were analyzed using CellQuest software on a Macintosh Power PC. (4,5)P 2 -To study the molecular mechanisms of platelet ␣-granule secretion, we have developed and characterized an SL-O-permeabilized, Ca 2ϩ -induced ␣-granule secretory system (12). In this system, exposure of platelets to SL-O permits the entry of molecules as large as 260 kDa (12). In the presence of 5 mM ATP, 96 Ϯ 1% of platelets exposed to SL-O expressed P-selectin on their surface in response to 10 M Ca 2ϩ . In contrast, nonpermeabilized platelets expressed almost no P-selectin on their surface in response to ATP and Ca 2ϩ ( Fig. 1) (12). These observations demonstrate that Ͼ95% of platelets were permeabilized following incubation with SL-O. Multiple Ca 2ϩ -induced permeabilized secretory systems, including permeabilized platelet models (42,43), have demonstrated a requirement for ATP exposure prior to Ca 2ϩ -induced triggering of granule secretion. Many permeabilized secretory systems also require cytosol to reconstitute significant secretion (44). A role for PtdIns(4,5)P 2 synthesis in PC12 cell granule secretion was demonstrated when type I PIPK and PtdIns transfer protein were identified as proteins within cytosol responsible for reconstituting ATP-mediated priming of granule secretion (29,30). In contrast, cytosol is not required for Ca 2ϩ -mediated secretion of ␣-granules from SL-O-permeabilized platelets (12). However, ␣-granule secretion from SL-O-permeabilized platelets remained ATP-dependent ( Fig. 1). ATP did not support ␣-granule secretion from non-permeabilized platelets or under conditions in which the free Ca 2ϩ concentration was maintained at pCa 2ϩ Ͻ 8 with 2 mM EGTA (Fig. 1B). The ability of ATP to presence or absence of 5 mM ATP at pH 6.9 for 10 min. Platelets were then exposed to 10 M Ca 2ϩ for 10 min. Pselectin surface expression was assayed by incubating the sample (10 l) with phycoerythrin-conjugated anti-P-selectin antibody AC1.2 (5 l) for 20 min. Phosphate-buffered saline (500 l) was added to the sample, and the platelets were analyzed immediately by flow cytometry as described under "Experiment Procedures." Histograms represent the distribution of the relative fluorescence of 10,000 platelets. B, gelfiltered platelets in PIPES/EGTA buffer supplemented with 5 mM MgCl 2 were incubated in the presence or absence of 5 mM ATP as indicated. Platelets were then incubated in the presence or absence of 4 units/ml SL-O at pH 6.9 for 10 min and subsequently exposed to buffer or 10 M Ca 2ϩ for 10 min. Platelets were assayed for P-selectin expression by flow cytometry. Error bars represent the S.E. of three to seven independent experiments. C, gel-filtered platelets in PIPES/EGTA buffer supplemented with 5 mM MgCl 2 were permeabilized with 4 units/ml SL-O in the presence of the indicated concentrations of ATP at pH 6.9 for 10 min in the presence (E) or absence (q) of 500 M ADP. Platelets were then exposed to 10 M Ca 2ϩ for 10 min and assayed for P-selectin expression by flow cytometry. Error bars represent the S.E. of six independent experiments.

FIG. 2. Effect of ATP on PtdIns(4,5)P 2 synthesis in SL-O-permeabilized platelets.
A, gel-filtered platelets (350 l/sample) in PIPES/EGTA buffer supplemented with 5 mM MgCl 2 were labeled with 2 mCi/ml [ 32 P]orthophosphoric acid at 37°C for 2 h. 32 P-Labeled platelets were then permeabilized with 3 units/ml SL-O in the presence or absence of 5 mM ATP at pH 6.9 for 30 min. The lipids were extracted, and PtdIns(4,5)P 2 levels were analyzed by TLC as described under "Experimental Procedures." Error bars represent the S.E. of six samples. B, to assess the time course of PtdIns(4,5)P 2 synthesis, gel-filtered platelets were labeled with [ 32 P]orthophosphoric acid and permeabilized with 2 units/ml SL-O in the presence of 5 mM MgATP at pH 6.9 for the indicated amounts of time. The lipids were then extracted, and PtdIns(4,5)P 2 levels were analyzed by TLC (q). To assess the time course of P-selectin surface expression, gel-filtered platelets were permeabilized with 2 units/ml SL-O in the presence of 5 mM MgATP, 10 M Ca 2ϩ , and phycoerythrin-conjugated anti-P-selectin antibody (5 l) at pH 6.9 for the indicated amounts of time. Samples were then diluted in phosphate-buffered saline (500 l) and analyzed immediately by flow cytometry (छ). Data are expressed as the percentage of P-selectin expression compared with a sample permeabilized in the presence of 5 mM MgATP and 10 M Ca 2ϩ for 30 min and subsequently incubated with phycoerythrin-conjugated anti-P-selectin antibody for the indicated amounts of time. Error bars represent the S.E. of four independent experiments. support ␣-granule secretion was dose-dependent with an EC 50 of ϳ1 mM. ADP neither supported ␣-granule secretion from SL-O-permeabilized platelets nor augmented Ca 2ϩ -induced ␣-granule secretion in the presence of submaximal concentrations of ATP (Fig. 1C). These data demonstrate that Ca 2ϩinduced ␣-granule secretion in this system requires permeabilization with SL-O and exposure to ATP.
The observation that ATP, but not platelet cytosol, is required for Ca 2ϩ -induced ␣-granule secretion raises the question of whether SL-O-permeabilized platelets are able to synthesize PtdIns(4,5)P 2 upon exposure to ATP in the absence of added cytosol. To address this question, platelets were radiolabeled with [ 32 P]orthophosphate, incubated in the presence or absence of ATP, and analyzed for PtdIns(4,5)P 2 levels by TLC. PtdIns(4,5)P 2 levels were markedly increased in SL-O-permeabilized platelets incubated in the presence of ATP compared with levels in platelets incubated in the absence of ATP ( Fig.  2A). The contribution of radioactivity from the 4 Ϯ 1% of platelets that were not permeabilized upon exposure to SL-O was calculated to constitute ϳ1% of the total radioactivity in the PtdIns(4,5)P 2 band. We next performed a time course study to determine whether the ability to secrete ␣-granules in response to Ca 2ϩ exposure is temporally correlated with PtdIns(4,5)P 2 synthesis. PtdIns(4,5)P 2 synthesis and the degree of ␣-granule secretion in response to Ca 2ϩ increased in parallel in SL-O-permeabilized platelets over the first 30 min The lipids were extracted, and PtdIns(4,5)P 2 levels were analyzed by TLC. Data are expressed as percent PtdIns(4,5)P 2 levels compared with samples exposed to buffer alone. Error bars represent the S.E. of six samples. B, gel-filtered platelets were permeabilized with 3 units/ml SL-O in the presence of 5 mM MgATP at pH 6.9 for 15 min. Permeabilized platelets were then incubated for 15 min with buffer (No Addition), 5 units/ml PtdIns-specific PLC, or 5 units/ml heat-denatured PtdIns-specific PLC. Following this incubation, platelets were exposed to either buffer (No Ca 2ϩ ) or 10 M Ca 2ϩ for 5 min. Platelets were then assayed for P-selectin expression by flow cytometry as described under "Experiment Procedures." Histograms represent the distribution of the relative fluorescence of 10,000 platelets. C, gel-filtered platelets were permeabilized with 3 units/ml SL-O in the presence of 5 mM MgATP at pH 6.9 for 15 min. Permeabilized platelets were then incubated with the indicated concentrations of PtdIns-specific PLC for 15 min. Platelets were exposed to 10 M Ca 2ϩ for 5 min and assayed for P-selectin expression. Data are expressed as the percent inhibition of P-selectin expression in samples exposed to PtdIns-specific PLC compared with samples exposed to buffer alone. Error bars represent the S.E. of three independent experiments. D, gel-filtered platelets were permeabilized with 3 units/ml SL-O in the presence of 5 mM MgATP at pH 6.9 for 15 min. Permeabilized platelets were then incubated with 5 units/ml PtdIns-specific PLC for the indicated times. After 1 h following permeabilization, platelets were exposed to 10 M Ca 2ϩ for 5 min and assayed for P-selectin expression. Data are expressed as the percent inhibition of P-selectin expression in samples exposed to PtdIns-specific PLC compared with samples exposed to buffer alone. Error bars represent the S.E. of three independent experiments. PtdIns(4,5)P 2 Mediates ␣-Granule Secretion following exposure to ATP (Fig. 2B). Thus, SL-O-permeabilized platelets are capable of synthesizing PtdIns(4,5)P 2 in response to ATP without the addition of cytosol, and PtdIns(4,5)P 2 synthesis correlates temporally with the ability to secrete ␣-granules in response to Ca 2ϩ exposure.
Effect of PtdIns-specific PLC on Ca 2ϩ -induced ␣-Granule Secretion-To determine whether PtdIns(4,5)P 2 mediates Ca 2ϩ -induced ␣-granule secretion from SL-O-permeabilized platelets, we hydrolyzed endogenous platelet PtdIns(4,5)P 2 using PtdIns-specific PLC. PtdIns-specific PLC reduced endogenous PtdIns(4,5)P 2 levels in the SL-O-permeabilized platelets by ϳ75% (Fig. 3A). However, the use of PtdIns-specific PLC in secretory assays is complicated by the fact that the by-products of PtdIns-specific PLC activity, diacylglycerol and inositol 1,4,5-trisphosphate, are themselves capable of influencing secretion from permeabilized platelets (19). Diacylglycerol acts primarily via activation of protein kinase C (21). Sloan and Haslam (45) have shown that protein kinase C diffuses out of platelets following exposure to SL-O. Consistent with these observations, we found that stimulation of P-selectin surface expression by phorbol 12-myristate 13-acetate was reduced by Ͼ90% in SL-O-permeabilized platelets compared with nonpermeabilized platelets (data not shown). Thus, permeabilized platelets are relatively insensitive to diacylglycerol because of protein kinase C leakage. Inositol 1,4,5-trisphosphate is not likely to stimulate secretion in this system because Ca 2ϩ chelation by EGTA blocks the increase in [Ca 2ϩ ] i that is elicited by inositol 1,4,5-trisphosphate in unchelated systems. In the secretion experiments, platelets were permeabilized with SL-O for 15 min prior to exposure to PtdIns-specific PLC. Platelets were then incubated in the presence or absence of PtdInsspecific PLC for 15 min prior to Ca 2ϩ exposure. Ca 2ϩ -induced ␣-granule secretion was inhibited in permeabilized platelets exposed to PtdIns-specific PLC (Fig. 3B). Platelets incubated with PtdIns-specific PLC in the absence of Ca 2ϩ exposure demonstrated a small degree of secretion under these conditions (8 Ϯ 2% P-selectin expression compared with samples incubated with buffer and subsequently exposed to Ca 2ϩ ) that might result from the generation of diacylglycerol. Heat-denatured PtdIns-specific PLC had no effect on Ca 2ϩ -induced ␣-granule secretion. PtdIns-specific PLC had no effect on intact platelets since the amount of SFLLR-induced P-selectin surface expression from intact platelets exposed to 5 units/ml Following this incubation, platelets were exposed to either buffer (No Ca 2ϩ ) or 10 M Ca 2ϩ for 5 min. Platelets were then assayed for P-selectin expression by flow cytometry as described under "Experimental Procedures." Histograms represent the distribution of the relative fluorescence of 10,000 platelets. B, data are expressed as percent Pselectin surface expression compared with samples exposed to buffer alone. Error bars represent the S.E. of three independent experiments. MgATP at pH 6.9 for 25 min. Platelets were exposed to 10 M Ca 2ϩ for 5 min and analyzed for P-selectin expression by flow cytometry as described under "Experimental Procedures." Data are expressed as percent P-selectin surface expression compared with samples exposed to 0.3% dimethyl sulfoxide alone. B, gel-filtered platelets were incubated with MgATP in the presence of buffer (No addition), 40 g/ml anti-PtdIns 3-kinase p85␣ monoclonal antibody, or 40 g/ml nonimmune mouse IgG (nonimmune antibody). Platelets were subsequently permeabilized with 2 units/ml SL-O at pH 6.9 for 25 min. Platelets were then exposed to 10 M Ca 2ϩ for 5 min and analyzed for P-selectin expression by flow cytometry. Data are expressed as percent P-selectin surface expression compared with samples exposed to buffer alone.
PtdIns-specific PLC was 97 Ϯ 3% of that from intact platelets not exposed to PtdIns-specific PLC. Similarly, PtdIns-specific PLC failed to inhibit Ca 2ϩ -induced P-selectin expression from platelets permeabilized with ␣-toxin. ␣-Toxin creates pores that restrict the entry of molecules greater than ϳ4.4 kDa (46). Ca 2ϩ -induced P-selectin expression from platelets permeabilized with ␣-toxin and subsequently exposed to 5 units/ml PtdIns-specific PLC was 101 Ϯ 9% of that from ␣-toxin-permeabilized platelets not exposed to PtdIns-specific PLC. Thus, the effects of PtdIns-specific PLC on ␣-granule secretion require entry into the platelet cytosol. Inhibition of Ca 2ϩ -induced ␣-granule secretion by PtdIns-specific PLC occurred in a dosedependent manner with an IC 50 of ϳ0.5 units/ml (Fig. 3C). Inhibition of ␣-granule secretion by PtdIns-specific PLC was also time-dependent, with the amount of inhibition increasing to ϳ75% over a 45-min incubation time (Fig. 3D). These data demonstrate that incubation of SL-O-permeabilized platelets with PtdIns-specific PLC interferes with Ca 2ϩ -induced ␣-granule secretion.
Effect of Wortmannin, LY294002, and Anti-PtdIns 3-Kinase Antibody on Ca 2ϩ -induced ␣-Granule Secretion-PtdIns 3-kinase has previously been implicated in the process of regulated granule exocytosis from hematopoietic cells (49,50). We therefore sought to determine whether PtdIns 3-kinase is involved in Ca 2ϩ -induced ␣-granule secretion from SL-O-permeabilized platelets. To this end, experiments using inhibitors of PtdIns 3-kinase, wortmannin and LY294002, were performed. In these experiments, neither 250 nM wortmannin nor 33.3 M LY294002 inhibited Ca 2ϩ -induced ␣-granule secretion from SL-O-permeabilized platelets (Fig. 5A). The activity of 250 nM wortmannin and 33.3 M LY294002 was confirmed in platelet aggregation studies. Reversible platelet aggregation, which is dependent on PtdIns 3-kinase activity (51), was inhibited to 36 Ϯ 3% of base-line levels by 250 nM wortmannin and to 29 Ϯ 6% of base-line levels by 33.3 M LY294002. Thus, the concentrations of wortmannin and LY294002 used to assess the role of PtdIns 3-kinase in Ca 2ϩ -induced ␣-granule secretion are sufficient to inhibit PtdIns 3-kinase-dependent platelet functions. Antibodies directed at PtdIns 3-kinase p85␣ also failed to inhibit Ca 2ϩ -induced ␣-granule secretion from SL-O-permeabilized platelets (Fig. 5B). These results demonstrate that PtdIns 3-kinase is not involved in Ca 2ϩ -induced, ATP-dependent ␣-granule secretion.
Experiments performed in the presence of increasing concentrations of wortmannin demonstrated that wortmannin inhibited Ca 2ϩ -induced ␣-granule secretion to a maximum of ϳ50% at 3.125 M (Fig. 6A). Similarly, 3.125 M wortmannin inhibited the synthesis of PtdIns(4,5)P 2 in SL-O-permeabilized platelets by ϳ50% (Fig. 6B). To determine whether exposure to wortmannin inhibits ATP-or Ca 2ϩ -dependent processes, SL-Opermeabilized platelets were exposed to wortmannin either 1) prior to ATP exposure or 2) after ATP exposure, but prior to Ca 2ϩ exposure. Incubation of platelets with 3.125 M wortman-

FIG. 6. Effect of micromolar concentrations of wortmannin on Ca 2؉ -induced ␣-granule secretion from and PtdIns(4,5)P 2 synthesis in SL-O-permeabilized platelets.
A, gel-filtered platelets were incubated with the indicated concentrations of wortmannin for 30 min. Platelets were then permeabilized with 2 units/ml SL-O for 5 min and subsequently incubated with 5 mM MgATP at pH 6.9 for 25 min. Platelets were exposed to 10 M Ca 2ϩ for 5 min and analyzed for P-selectin expression by flow cytometry as described under "Experimental Procedures." Data are expressed as percent P-selectin surface expression compared with samples exposed to dimethyl sulfoxide alone. Error bars represent the S.E. of three to seven independent experiments. B, 32 P-labeled gel-filtered platelets (350 l/sample) in PIPES/ EGTA buffer were incubated with either 0.3% dimethyl sulfoxide (vehicle) or 3.125 M wortmannin for 30 min. Labeled platelets were then permeabilized with 2 units/ml SL-O in the presence of 5 mM MgATP at pH 6.9 for 30 min. The lipids were extracted, and PtdIns(4,5)P 2 levels were analyzed by TLC. Data are expressed as percent PtdIns(4,5)P 2 levels compared with samples exposed to vehicle alone. Error bars represent the S.E. of six samples. C, one set of platelets was incubated with either vehicle (white bars) or 3.125 M wortmannin (black bars) for 30 min. Platelets were subsequently permeabilized with 2 units/ml SL-O for 5 min and incubated with 5 mM MgATP at pH 6.9 for 15 min. A second set of platelets was permeabilized with 2 units/ml SL-O for 5 min and incubated with 5 mM MgATP at pH 6.9 for 15 min. Following incubation with MgATP, these samples were incubated with vehicle or 3.125 M wortmannin for 45 min. Following incubation with MgATP, all samples were exposed to 10 M Ca 2ϩ for 5 min and analyzed for P-selectin surface expression. Data are expressed as percent P-selectin surface expression compared with samples exposed to vehicle. Error bars represent the S.E. of three independent experiments. nin for 45 min inhibited Ca 2ϩ -induced ␣-granule secretion from SL-O-permeabilized platelets when wortmannin was added prior to ATP exposure (Fig. 6C). However, ␣-granule secretion was not inhibited in SL-O-permeabilized platelets exposed to ATP for 15 min and subsequently incubated with 3.125 M wortmannin for 45 min (Fig. 6C). Thus, wortmannin inhibits ATP-dependent events, but not Ca 2ϩ -mediated triggering of ␣-granule secretion. The correlation between the reduction of PtdIns(4,5)P 2 synthesis by micromolar concentrations of wortmannin and the inhibition of an ATP-dependent step of Ca 2ϩinduced ␣-granule secretion is consistent with a role for PtdIns(4,5)P 2 synthesis in this process.
Effect of Anti-type II PIPK Antibodies on Ca 2ϩ -induced ␣-Granule Secretion-PtdIns 5-phosphate levels increase in platelets following stimulation with thrombin (37). Furthermore, type II PIPK has previously been demonstrated to be abundant in platelets (38). We therefore determined the effect of anti-type II PIPK antibodies on PtdIns(4,5)P 2 synthesis in and ␣-granule secretion from SL-O-permeabilized platelets. Anti-type II PIPK antibodies directed against the amino-and carboxyl-terminal ends of the ␣-isoform of human type II PIPK were used in these studies. The peptides used to generate these antibodies have no homology to type I PIPK. Each antibody recognized a single band of 53 kDa in platelet lysates (Fig. 7A). In contrast, nonimmune IgG failed to recognize any bands in platelet lysates (data not shown). In addition, both antibodies recognized a recombinant GST-tagged type II PIPK fusion protein of 79 kDa in immunoblot analysis (Fig. 7B). When incubated with MgATP-exposed platelets prior to permeabilization with SL-O, both anti-amino-and anti-carboxyl-terminal type II PIPK antibodies inhibited PtdIns(4,5)P 2 levels by ϳ50% compared with permeabilized platelets incubated with nonimmune antibody (Fig. 7C). In contrast, neither anti-type II PIPK antibody inhibited PtdIns(4,5)P 2 levels when incubated with platelets for 15 min following permeabilization with SL-O. We next determined the effect of anti-type II PIPK antibodies on Ca 2ϩinduced ␣-granule secretion from SL-O-permeabilized platelets. In these experiments, platelets were incubated in the presence of nonimmune antibody or anti-type II PIPK antibody directed against either the amino-or carboxyl-terminal end of type II PIPK. Platelets were permeabilized with SL-O and then exposed to 10 M Ca 2ϩ . Both antibodies significantly inhibited Ca 2ϩ -induced ␣-granule secretion from SL-O-permeabilized platelets (Fig. 7D). In contrast, nonimmune antibody had no effect on Ca 2ϩ -induced ␣-granule secretion. As observed with wortmannin, anti-type II PIPK antibodies inhibited Ca 2ϩ -induced ␣-granule secretion when added prior to, but not after, ATP (Fig. 7D). The amount of SFLLR-induced P-selectin surface expression from intact platelets exposed to anti-amino-or anti-carboxyl-terminal type II PIPK antibody was 104 Ϯ 24 and 93 Ϯ 18%, respectively, of that from intact platelets not exposed to anti-type II PIPK antibodies. Similarly, Ca 2ϩ -induced P-selectin expression from platelets permeabilized with ␣-toxin in the presence of anti-amino-or anti-carboxyl-terminal type II PIPK antibody was 100 Ϯ 6 and 91 Ϯ 22%, respectively, of that from ␣-toxin-permeabilized platelets not exposed to anti-type II PIPK antibodies. Thus, the effects of exogenously added anti-FIG. 7. Effect of anti-type II PIPK antibodies on PtdIns(4,5)P 2 synthesis in and Ca 2؉ -induced ␣-granule secretion from SL-Opermeabilized platelets. A, proteins from human platelets were solubilized at 95°C in sample buffer, separated by SDS-polyacrylamide gel electrophoresis on a 12% gel, and electrophoretically transferred to polyvinylidene difluoride membranes. Immunoblotting was performed with antibodies directed against the amino-and carboxyl-terminal peptides of type II PIPK as indicated. Bands were visualized using enhanced chemiluminescence for detection. The positions of the molecular mass standards used are indicated on the left. B, recombinant GSTtagged type II PIPK was solubilized at 95°C in sample buffer, separated by SDS-polyacrylamide gel electrophoresis on a 12% gel, and electrophoretically transferred to polyvinylidene difluoride membranes. Immunoblotting was performed with antibodies directed against the amino-and carboxyl-terminal peptides as described for A. C, one set of 32 P-labeled gel-filtered platelets (black bars) was permeabilized with 4 units/ml SL-O in the presence of 40 g/ml nonimmune antibody, 40 g/ml anti-amino-terminal type II PIPK antibody (N-term), or 40 g/ml anti-carboxyl-terminal type II PIPK antibody (C-term) with 5 mM MgATP at pH 6.9 for 90 min. A second set of 32 P-labeled platelets (white bars) was permeabilized with 4 units/ml SL-O in the presence of 5 mM MgATP alone at pH 6.9 for 15 min. This set of platelets was then exposed to 40 g/ml nonimmune antibody, 40 g/ml anti-amino-terminal type II PIPK antibody, or 40 g/ml anti-carboxyl-terminal type II PIPK antibody for 75 min. The lipids were extracted, and PtdIns(4,5)P 2 levels were analyzed by TLC. Data are expressed as percent PtdIns(4,5)P 2 levels compared with samples exposed to nonimmune antibody. Error bars represent the S.E. of three to six samples. D, one set of gel-filtered platelets (black bars) was incubated in the presence of 40 g/ml nonimmune antibody, 40 g/ml anti-amino-terminal type II PIPK antibody, or 40 g/l anti-carboxyl-terminal type II PIPK antibody. Samples were then exposed to 10 mM MgATP and permeabilized with 5 units/ml SL-O. Following permeabilization, all samples were adjusted to pH 6.9 and subsequently exposed to 10 M Ca 2ϩ . A second set of gel-filtered platelets (white bars) was incubated in the presence of 10 mM MgATP at pH 6.9 with no addition. Samples were then perme-abilized with 5 units/ml SL-O. Following this MgATP exposure, samples from the second set were incubated with 40 g/ml nonimmune antibody, 40 g/ml anti-amino-terminal type II PIPK antibody, or 40 g/l of anti-carboxyl-terminal type II PIPK antibody, followed by 10 M Ca 2ϩ . Following a 15-min incubation in the presence of Ca 2ϩ , all samples were assayed for P-selectin expression by flow cytometry as described under "Experimental Procedures." Data are expressed as percent P-selectin expression compared with samples incubated in the presence of buffer alone. Error bars represent the S.E. of four to eight independent experiments. type II PIPK antibodies require entry into the platelet cytosol. We conclude that type II PIPK mediates ␣-granule secretion from Ca 2ϩ -induced SL-O-permeabilized platelets and acts prior to Ca 2ϩ exposure.
To assess the possibility that wortmannin and type II PIPK inhibit Ca 2ϩ -induced ␣-granule secretion by distinct mechanisms, we performed experiments using both inhibitors. In these experiments, we incubated platelets in the presence of 3.125 M wortmannin or vehicle for 30 min and exposed them to nonimmune, anti-amino-terminal end, or anti-carboxyl-terminal end antibodies prior to ATP exposure and SL-O permeabilization. Under these conditions, ϳ40% inhibition of Ca 2ϩinduced ␣-granule secretion was observed in platelets exposed to only wortmannin or either of the anti-type II PIPK antibodies compared with platelets treated with vehicle and nonimmune antibody (Fig. 8). Increasing the concentration of antitype II PIPK antibodies to 200 g/ml failed to increase the degree of inhibition of ␣-granule secretion. However, ␣-granule secretion from SL-O-permeabilized platelets exposed to both wortmannin and anti-type II PIPK antibodies was inhibited by ϳ70% compared with platelets exposed to vehicle and nonimmune antibody (Fig. 8). Thus, inhibition by wortmannin and anti-type II PIPK antibodies is additive. DISCUSSION One advantage of permeabilized secretory systems is that target-specific, cell-impermeable inhibitors such as antibodies and enzymes can be introduced intracellularly. This characteristic permits for more precise identification of molecules directly involved in the secretory process. The fact that Ca 2ϩinduced ␣-granule secretion from SL-O-permeabilized platelets is inhibited by PtdIns-specific PLC (which cleaves PtdIns(4,5)P 2 ), by exogenously added PtdIns(4,5)P 2 (which may act by displacing proteins bound to endogenous PtdIns(4,5)P 2 ), and by two different anti-type II PIPK antibodies (which inhibit PtdIns(4,5)P 2 synthesis) provides strong evidence that PtdIns(4,5)P 2 mediates ␣-granule secretion in this system. Since none of these inhibitors have inhibitory activity in intact platelets or in ␣-toxin-permeabilized platelets, access to the platelet cytosol is required for their inhibitory activity. The assertion that PtdIns(4,5)P 2 mediates ␣-granule secretion is also supported by the observations that the ability to secrete ␣-granules upon Ca 2ϩ exposure is temporally correlated with PtdIns(4,5)P 2 synthesis and that micromolar concentrations of wortmannin inhibit both PtdIns(4,5)P 2 synthesis and Ca 2ϩinduced ␣-granule secretion.
A second advantage of permeabilized secretory systems is that they allow for the dissection of the secretory process into component parts. Like other permeabilized cells, platelets will undergo Ca 2ϩ -mediated secretion only following exposure to ATP (42,43,46). SL-O-permeabilized platelet systems differ from many permeabilized cell systems, however, in that Ca 2ϩinduced granule secretion from platelet systems does not exhibit a requirement for exogenous cytosolic proteins (12,45,52). The observation that PtdIns(4,5)P 2 synthesis occurs in SL-O-permeabilized platelets in the absence of exogenous cytosol suggests that some PtdIns(4,5)P 2 synthetic machinery remains cell-associated in this permeabilized secretory system. In our studies, inhibitors of PtdIns(4,5)P 2 synthesis inhibited ␣-granule secretion when added prior to ATP exposure, but not when added following ATP exposure, but prior to Ca 2ϩ -mediated triggering of secretion. This finding raises the possibility that activation-dependent changes in localized PtdIns(4,5)P 2 concentrations mediate granule secretion. Indeed, activation of platelets is accompanied by both an increase in PtdIns kinase activity (33,35,36) and relocalization of PtdIns kinases (53). Future studies will determine whether activation-dependent increases in PtdIns(4,5)P 2 at zones of granule secretion within the platelet are necessary for ␣-granule secretion.
Previous studies of hematopoietic cells have implicated PtdIns 3-kinase in the process of regulated granule exocytosis. For example, nanomolar concentrations of wortmannin inhibit granule secretion from a basophilic leukemia cell line (RBL-2H3) (49) and from natural killer cells (50). However, the fact that nanomolar concentrations of wortmannin, 33.3 M LY294002, and antibodies directed at PtdIns 3-kinase p85␣ did not inhibit secretion from SL-O-permeabilized platelets (Fig. 5) demonstrates that D 3 phosphoinositides are not involved in ␣-granule secretion from platelets. The observation that wortmannin failed to inhibit ␣-granule secretion at concentrations that inhibit PtdIns 3-kinase is consistent with the data of Kovacsovics et al. (51). Their studies using intact platelets demonstrated that although wortmannin at nanomolar concentrations inhibits thrombin receptor agonist peptide-induced activation of glycoprotein IIb-IIIa, it fails to affect ␣-granule secretion from platelets. Similarly, wortmannin at nanomolar concentrations is a relatively poor inhibitor of neutrophil granule secretion, even though it inhibits the neutrophil oxidative burst with an IC 50 of Ͻ5 nM (54). Thus, granule secretion from various cells of hematopoietic origin differs in requirements for phosphoinositides, and platelet ␣-granule secretion demonstrates a requirement for D4, but not D3, phosphoinositides.
In addition to demonstrating that PtdIns(4,5)P 2 is synthesized in SL-O-permeabilized platelets, these studies demonstrate that the type II PIPK-dependent pathway is involved in the synthesis of PtdIns(4,5)P 2 required for Ca 2ϩ -induced ␣-granule secretion. In PC12 cells, type I, but not type II, PIPK reconstitutes Ca 2ϩ -mediated secretion (30). To our knowledge, type II PIPK has not previously been demonstrated to function in a cellular process (55). However, PtdIns 5-kinase levels are increased upon exposure of intact platelets to thrombin (37). In this study, we show that anti-type II PIPK antibodies partially inhibit Ca 2ϩ -induced ␣-granule secretion from SL-O-permeabi- Platelets were then exposed to 40 g/ml nonimmune antibody or 40 or 200 g/ml antibody directed against an amino-terminal (N-term) or carboxyl-terminal (C-term) peptide of type II PIPK as indicated and permeabilized with 4 units/ml SL-O in the presence of 5 mM MgATP. Following permeabilization, all samples were adjusted to pH 6.9. After a 30-min incubation, samples were exposed to 10 M Ca 2ϩ for an additional 10-min incubation. Samples were then analyzed for P-selectin expression by flow cytometry as described under "Experimental Procedures." Data are expressed as percent P-selectin expression compared with samples incubated in the presence of vehicle and nonimmune antibody. lized platelets. Three isoforms of type II PIPK (␣, ␤, and ␥) have been identified. The antibodies used in this study are directed to the amino-and carboxyl-terminal ends of the ␣-isoform. However, these antibodies cross-react with the ␤-isoform (data not shown); and thus, we are not able to determine which isoform(s) of type II PIPK is active in platelets. Anti-type II PIPK antibodies do not cause complete inhibition of Ca 2ϩinduced ␣-granule secretion from SL-O-permeabilized platelets. The degree of inhibition of secretion was no greater when the concentration of antibody was increased from 40 to 200 g/ml (Fig. 8). Neither the anti-amino-nor anti-carboxyl-terminal type II PIPK antibody is directed against the proposed active site of the kinase. Thus, it is possible that the antibodies cause partial inhibition of type II PIPK function by disrupting its interactions with other molecules (e.g. PtdIns kinases) required for PtdIns(4,5)P 2 production (56). Another possibility is that partial PtdIns(4,5)P 2 production in the presence of antitype II PIPK antibodies is secondary to the activity of type I PIPK isoforms, which are also found in platelets. 2 The fact that exposure of SL-O-permeabilized platelets to micromolar concentrations of wortmannin results in a 50% reduction of both PtdIns(4,5)P 2 synthesis and Ca 2ϩ -induced, ATP-dependent ␣-granule secretion when added prior to, but not after, ATP exposure raises the possibility that the PtdIns 4-kinase pathway is also involved in ␣-granule secretion. Three isoforms of PtdIns 4-kinase have been identified and can be distinguished based on their sensitivity to wortmannin. A 55-kDa PtdIns 4-kinase is resistant to wortmannin (57). A 230-kDa PtdIns 4-kinase ␣ with a 97-kDa splice variant is inhibited by wortmannin with an IC 50 of ϳ1 M and is nearly completely inhibited at 10 M (58). PtdIns 4-kinase ␤ is inhibited by wortmannin with an IC 50 of ϳ120 nM and is completely inhibited at 1 M. However, wortmannin at micromolar concentrations is not specific for PtdIns 4-kinases. Inhibition of the ATP-dependent step of Ca 2ϩ -induced ␣-granule secretion by micromolar concentrations of wortmannin may occur by interfering with molecules not involved in PtdIns(4,5)P 2 synthesis. The fact that PtdIns(4,5)P 2 synthesis is also inhibited by wortmannin may be secondary to inhibition of a PtdIns 5-kinase. Although the PtdIns 5-kinases in platelets have not been well characterized, a 235-kDa mammalian PtdIns kinase originally discovered in insulin-sensitive cells (59) was subsequently found to synthesize PtdIns 5-phosphate (60). Wortmannin inhibits this PtdIns 5-kinase with an IC 50 of ϳ600 nM and completely inhibits kinase activity at 4 M wortmannin (60). Thus, it is possible that the effect of wortmannin on PtdIns(4,5)P 2 synthesis and Ca 2ϩ -induced ␣-granule secretion is via inhibition of a PtdIns 5-kinase pathway rather than through a PtdIns 4-kinase pathway. However, the identification of the complete set of PtdIns kinase and PIPK isoforms used in the synthetic pathway of PtdIns(4,5)P 2 required for ␣-granule secretion remains to be determined.
How might PtdIns(4,5)P 2 mediate ␣-granule secretion? Platelet ␣-granule secretion is known to require the SNARE protein machinery (12)(13)(14)(15). In other secretory systems, it has been postulated that PtdIns(4,5)P 2 might act by interacting with proteins, such as synaptotagmin and Munc-18-interacting protein, that both bind PtdIns(4,5)P 2 and interact with SNARE proteins (61). In addition, PtdIns(4,5)P 2 interacts with ADPribosylation factors and GTPase-activating proteins to localize these molecules to sites of membrane fusion (61). However, the presence of ADP-ribosylation factors in platelets and the role of GTPase-activating proteins and ADP-ribosylation factors in platelet secretion remain to be explored. PtdIns(4,5)P 2 also stimulates phospholipase D (62), which generates phosphatidic acid and is thought to influence platelet granule secretion (24). Phosphatidic acid, in turn, stimulates type I PIPK activity (63). Thus, a feed-forward mechanism is established whereby cell activation leads to increased synthesis of phosphatidic acid and PtdIns(4,5)P 2 . Future studies will provide a more detailed understanding of how PtdIns(4,5)P 2 mediates interactions between platelet signaling events and SNARE protein rearrangements that result in membrane fusion.