Cyclic AMP-dependent Phosphorylation of Thromboxane A2 Receptor-associated Gα13 *

Although it is well established that cAMP inhibits platelet activation induced by all agonists, the thromboxane A2 signal transduction pathway was found to be particularly sensitive to such inhibition. Therefore, we examined whether cAMP-dependent kinase mediates phosphorylation of the thromboxane A2 receptor-G-protein complex. It was found that cAMP induces protein kinase A-dependent [γ-32P]ATP labeling of solubilized membrane proteins in the region of Gα subunits, i.e. 38–45 kDa. Moreover, ligand affinity chromatography purification of thromboxane A2 receptor-G-protein complexes from these membranes revealed that 38–45-kDa phosphoproteins co-purify with thromboxane A2 receptors. Immunoprecipitation of the affinity column eluate with a Gα13 antibody demonstrated that 8-Br-cAMP increased phosphorylation of thromboxane A2receptor-associated Gα13 by 87 ± 27%. In separate experiments, immunopurification of Gα13 on microtiter wells coated with a different Gα13 antibody revealed that 8-Br-cAMP increased Gα13 phosphorylation by 53 ± 19%. Finally, treatment of 32P-labeled whole platelets with prostacyclin resulted in a 90 ± 14% increase in phosphorylated Gα13 that was abolished by pretreatment with the adenylate cyclase inhibitor MDL-12. These results provide the first evidence that protein kinase A mediates phosphorylation of Gα13 both in vitro and in vivoand provides a basis for the preferential inhibition of thromboxane A2-mediated signaling in platelets by cAMP.

Thromboxane A 2 (TXA 2 ) 1 is a potent platelet agonist and vasoconstrictor (1, 2) that modulates not only the hemostatic process but also potentiates thrombotic events such as myocar-dial infarction and stroke (3)(4)(5)(6)(7). On this basis, investigators have sought to understand better TXA 2 -induced signal transduction processes in an attempt to develop new therapeutic strategies for managing TXA 2 receptor-mediated thrombotic events. However, despite these efforts, significant gaps remain in our understanding of signal transduction, and modulation of signal transduction, through this pathway. At present, it is known that at least one pool of platelet TXA 2 receptors is coupled to phospholipase C through the pertussis toxin-insensitive G-protein G␣ q (8 -14). Thus, TXA 2 receptor activation of phospholipase C leads to inositol triphosphate/diacylglycerol generation and a subsequent mobilization of intraplatelet calcium (8 -10). More recently, evidence has been provided that, in addition to G␣ q , TXA 2 receptors also couple to the G␣ 13 subunit (14 -17). Although the effector for this putative signaling protein has not yet been established in platelets, a variety of G 13 -mediated effects in other cell systems (18 -25) have been proposed.
Regarding modulation of TXA 2 receptor activity, recent studies have focused on cyclic nucleotide-mediated phosphorylation of specific platelet proteins (26 -29). Thus, it is well known that platelet activity is inhibited by elevated cAMP levels, and it is generally accepted that this inhibitory effect is mediated by protein kinase A (PKA)-dependent phosphorylation of different protein substrates. For example, cAMP-induced phosphorylation of certain downstream signaling components such as myosin light chain kinase and thrombolamban causes inhibition of contractile processes and calcium-dependent processes, respectively (30 -33). Furthermore, separate studies have established a direct relationship between platelet cAMP levels and the ability of TXA 2 receptors to mobilize intraplatelet calcium (34). This latter phenomenon, in turn, raises the possibility that the TXA 2 signaling components themselves may serve as potential PKA substrates.
In this connection, it is known that phosphorylation of receptors by second messenger kinases can modulate their function (35)(36)(37)(38). This phenomenon has been well studied in the ␤-adrenergic receptor system, in which PKA-induced receptor phosphorylation has been linked to heterologous receptor desensitization (39,40). On the other hand, the physiological consequence of PKA-mediated phosphorylation of TXA 2 receptors is unclear. Thus, even though it was found that TXA 2 receptor protein can be phosphorylated by PKA in receptor expression systems, the authors of this study (27) concluded that PKA had no functional significance in mediating agonistinduced TXA 2 receptor desensitization. In contrast to cAMP, however, other investigators (29) have recently reported that cGMP-dependent phosphorylation of TXA 2 receptors by protein kinase G (PKG) inhibits agonist-induced GTPase activity in platelet membranes. This finding may in part explain the inhibitory effects of cGMP on platelet function. * This work was supported in part by National Institutes of Health Grant HL-24530, the North Atlantic Treaty Organization Grant CRG-940595, and "Cellular Signaling in the Cardiovascular System" Training Grant T32 HL07692 and was conducted under the auspices of the Association for U. S.-French Biomedical Cooperation. A portion of this work was presented at the annual meeting of the American Society of Hematology, December 4 -8, 1998, Miami Beach, FL. 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  Finally, phosphorylation of G-proteins has also been shown to constitute a mechanism for signal modulation through seven transmembrane receptors. For example, protein kinase C (PKC) has been shown to induce phosphorylation of at least three G␣ subunits, i.e. G␣ i , G␣ z , and G␣ 12 (41)(42)(43)(44)(45)(46)(47). Results regarding PKC-mediated phosphorylation of G␣ 13 , however, are not definitive. Thus, whereas Offermanns et al. (47) found that G␣ 13 can be phosphorylated by PKC in platelets as well as in a reconstituted system, Kozasa and Gilman (46) reported that the purified G␣ 13 subunit was not a PKC substrate. In contrast to PKC, there have been fewer reports of G-protein phosphorylation induced by cyclic nucleotides. In this regard, PKA-and PKG-mediated phosphorylation of G ␣ subunits has only been described to occur in limited cell types and has not yet been demonstrated to occur in platelets (43,(47)(48)(49)(50)(51).
In summary, various mechanisms have been proposed for cyclic nucleotide-induced modulation of signaling either at the TXA 2 receptor level itself or at more distant effector sites. In order to provide additional information concerning this potentially important regulatory process, the present study investigated the ability of cAMP to induce phosphorylation of TXA 2 receptor-associated G-proteins. Since other investigators (43) have reported that platelet G␣ q is not a PKA substrate, we examined the ability of PKA to phosphorylate TXA 2 receptorcoupled G␣ 13 . In these studies, it was found in both platelet membranes and intact platelets that cAMP stimulates phosphorylation of G␣ 13 . These results provide the first evidence for PKA-mediated phosphorylation of an endogenous G-protein ␣ subunit in human platelets and identify one such ␣ subunit as being TXA 2 receptor-associated G␣ 13 . This ability of cAMP to cause G␣ 13 phosphorylation may, in part, explain our separate finding that TXA 2 -mediated signal transduction is preferentially sensitive to inhibition by prostacyclin (PGI 2 ).  [42][43][44][45][46][47] ) and was purchased from Research Genetics, (Huntsville, AL). U46619 and PGI 2 were purchased from Cayman Chemicals (Ann Arbor, MI); tissue solubilizer-1 was obtained from Research Products International, Corp. (Mount Prospect, IL); 4-chloro-1-naphthol (horseradish peroxidase color development reagent) was from Bio-Rad; biotinylated goat anti-rabbit IgG (H ϩ L) and the Vectastain ABC kits were purchased from Vector Laboratories (Burlingame, CA). The G␣ 13 -N IgG rabbit polyclonal antibody was purchased from Santa Cruz Laboratories (Santa Cruz, CA). The anti-peptide G-13 antibody, directed against an internal G␣ 13 sequence, (Table I) was raised in rabbits according to standard protocols, and the IgG fraction was purified by protein A-Sepharose chromatography as described (17).

Materials
Platelet Membrane Preparation and Solubilization-Typically, 12 units of outdated human platelet concentrate were pooled and incubated with 3 mM aspirin for 45 min. Solubilized platelet membranes were then prepared as described (52) and typically resulted in a final protein concentration of 4.5 mg/ml. Platelet membranes were prepared using a modification of this protocol. Specifically, platelet concentrate was centrifuged at 160 ϫ g for 15 min to remove red blood cells. The platelet-rich plasma was supplemented with 40 nM PGI 2 , and platelets were pelleted by centrifugation at 1600 ϫ g for 20 min. The platelet pellet was suspended in buffer A (25 mM Tris-HCl, 5 mM MgCl 2 , pH 7.4) plus PGI 2 (40 nM) and was re-centrifuged at 1600 ϫ g for 20 min. The resulting platelet pellet was resuspended in buffer A, was sonicated on ice (45-s bursts with 15-s rests) for 4 min, and was centrifuged at 1600 ϫ g for 5 min. The supernatant was placed into centrifuge tubes, and the pellet was again resuspended, sonicated, and centrifuged. The supernatants were then combined and centrifuged at 100,000 ϫ g for 30 min. The resulting membrane pellet was resuspended in buffer A to yield a final protein concentration of 2-4 mg/ml.
Polyacrylamide Gel Electrophoresis-SDS-PAGE was performed according to the method of Laemmli (53), and the gels were silver-stained as described by Morrissey (54), dried, and subjected to autoradiography. Silver-stained gels and autoradiographic films were analyzed by using a Protein Databases, Inc. densitometer (Huntington Station, NY).
Phosphorylation of Platelet Membranes and Solubilized Platelet Membranes-Solubilized platelet membranes were phosphorylated (2-4 mg of protein) in a reaction mixture containing 1 M cAMP, 1 mM 8-Br-cAMP or vehicle (deionized water), 100 -150 M [␥-32 P]ATP (4500 Ci/mmol), 60 M CaCl 2 , and incubation buffer (45 mM histidine HCl, 50 mM KH 2 PO 4 , 20 mM NaF, 120 mM KCl, pH 7.4) in a total volume of 1600 -2300 l. In some experiments, NaF was 40 mM, and 2 mM sodium orthovanadate was added as an additional phosphatase inhibitor. In other experiments, the specific PKA inhibitor, IP 20 (55), was added (3.6 M). The reaction mixture was allowed to incubate for 30 min at 25°C, after which time the phosphorylated proteins were either added to Laemmli sample buffer (62.5 mM Tris-HCl, 3% SDS, 10% glycerol, pH 6.8) and resolved by SDS-PAGE, purified by ligand affinity chromatography as described under "Ligand Affinity Purification of TXA 2 Receptor-G-Protein Complexes," or added to G-13 antibody-coated microtiter plates as described under "Immunological Techniques." In separate experiments, platelet membranes (1-2 mg of protein) were phosphorylated in a reaction mixture containing 1 M cAMP, 86 M [␥-32 P]ATP (4500 Ci/mmol), 1 mM EGTA, and incubation buffer in a volume of 2.25 ml. Following incubation for 30 min at 25°C, the membranes were centrifuged for 15 min at 100,0 ϫ g. The pellet was washed twice with buffer A and was solubilized in 2 ml of buffer B (10 mM CHAPS, 50 mM Tris, 5 mM MgCl 2, pH 7.4) that was supplemented with 20 mM NaF. The solubilized membranes were then centrifuged at 100,000 ϫ g for 30 min. The resulting supernatant was subjected to ligand affinity chromatography in order to purify the TXA 2 receptor-Gprotein complex, as described below.
Ligand Affinity Purification of TXA 2 Receptor-G-Protein Complexes-Platelet TXA 2 receptor-G-protein complexes were purified as described previously (52). Briefly, phosphorylated, solubilized platelet membranes were supplemented with 500 mM KCl, 0.5 mg/ml asolectin, 20% glycerol, 0.2 mM EGTA, and 10 mM CHAPS and were incubated overnight with the ligand affinity column matrix, which utilizes the TXA 2 receptor antagonist SQ31,491 coupled to Affi-Gel 102. Unbound proteins were washed away with buffer D (20 mM Tris base, 10 mM CHAPS, 20% glycerol, 500 mM KCl, 0.2 mM EGTA, 0.5 mg/ml asolectin, pH 7.4) that was supplemented with 20 mM NaF and 2 mM sodium orthovanadate. The TXA 2 receptor-G-protein complexes were eluted from the column using the TXA 2 antagonist 50 mM BM13.177 in buffer D at a flow rate of 1 ml/8 min. The first 1-ml elution fraction was collected, and the column was clamped for 30 min. Subsequently, two 1-ml elution fractions were collected, and the fractions were pooled, concentrated, and subjected to SDS-PAGE. To determine whether the same amount of receptor was eluted from each column, the pooled elution fraction was subjected to SDS-PAGE, and the gels were silverstained. Densitometric analysis of the stained receptor bands verified that the same amount of receptor protein was eluted in each set of experiments. In separate experiments, the pooled elution fraction was used for immunoprecipitation studies using the G␣ 13 -N IgG as described under "Immunological Techniques." Incubation of Platelets with [ 32 P]H 3 PO 4 -Human platelet-rich plasma (PRP) was prepared from acid/citrate/dextrose anticoagulated blood as described (34). Platelets were then 32 P-labeled by a modification of the method of Carlson et al. (41). Briefly, PRP was incubated with 3 mM aspirin for 45 min at room temperature to inhibit cyclooxygenase. Platelets were collected from PRP by centrifugation at 1000 ϫ g for 10 min and were washed once in a solution of modified Tyrode's buffer (20 mM HEPES, 138 mM NaCl, 2.9 mM KCl, 1 mM MgCl 2 , 1 mM glucose, 5 mM EDTA, 0.35% bovine serum albumin, 60 g/ml apyrase) to which 50 units/ml heparin was added. Platelets were pelleted by centrifugation at 1000 ϫ g for 10 min, resuspended in modified Tyrode's buffer (without heparin) to a concentration of 2 ϫ 10 9 cells/ml, and were incubated with 0.5 mCi/ml [ 32 P]H 3 PO 4 (9000 Ci/mmol) for 60 min at room temperature. Following incubation with 3.3 mM NaH 3 PO 4 for 15 min, portions of the platelet suspension (e.g. 2 ϫ 10 9 cells) were incu-bated with 500 M MDL-12 or vehicle (10 mM HEPES, 150 mM NaCl, pH 7.4) for 15 min and were subsequently incubated with 1 mM IBMX and 1 M PGI 2 or vehicle (ethanol and 0.05 M Tris buffer, pH 9, respectively) for 5 min. Samples were immediately cooled in an ice slurry, and the platelets were pelleted by centrifugation at 7000 ϫ g for 3 min and were washed twice in ice-cold buffer A that was supplemented with 2 M okadaic acid. Platelets were solubilized in buffer B containing 2 M okadaic acid and were sonicated (30-s bursts with 15-s rests) for 4 min and homogenized (10 strokes with a Teflon homogenizer). The solubilized mixture was centrifuged at 100,000 ϫ g for 30 min, and the sonication and centrifugation steps were repeated. The solubilized proteins were subjected to immunoprecipitation as described below. Precipitated 32 P counts were normalized for the amount of immunoprecipitated G␣ 13 by immunoblotting the precipitated protein fraction with the G␣ 13 -N antibody and analyzing the labeled protein by densitometry.
Immunological Techniques-For immunoprecipitation experiments, 1 ml of the pooled ligand affinity column eluate was incubated with 75 l of either G␣ 13 -N IgG (200 g/ml) or preimmune IgG (200 g/ml) and was mixed overnight at 4°C. Protein-A Sepharose beads were then added (55 l of a 10% (w/v) suspension). After mixing overnight at 4°C, the beads were isolated by centrifugation and were washed three times with 400 l of buffer containing 10 mM CHAPS in phosphate-buffered saline, pH 7.4. The final pellet was boiled for 5 min in the presence of Laemmli sample buffer (which was devoid of bromphenol blue) plus buffer B, and the mixture was centrifuged to pellet the beads. The resulting supernatant (60 l) was added to scintillation mixture, and the 32 P-labeled proteins were quantified using a liquid scintillation spectrometer (Beckman LS 6800). Specific phosphorylation was calculated as the difference between counts obtained with G␣ 13 -N IgG and preimmune IgG. For the whole cell experiments, G␣ 13 was precipitated using 150 l of G␣ 13 -N IgG and 55 l of the protein A-Sepharose bead suspension.
For immunoaffinity purification of G␣ 13 on microtiter plates, Immulon 2 microtiter wells were coated overnight at 4°C with 50 l of G-13 IgG (100 g/ml) or preimmune IgG (100 g/ml). The wells were then blocked with 3% bovine serum albumin in phosphate-buffered saline for 1 h at 25°C and were washed once with 300 l of phosphate-buffered saline. Solubilized platelet membranes that were phosphorylated in the presence or absence of 1 mM 8-Br-cAMP (50 l), as described previously, were incubated on the antibody-coated microtiter wells overnight at 4°C. The wells were washed three times with 300 l of buffer B, and the labeled protein was eluted from the wells using 50 l of tissue solubilizer. The eluted, 32 P-labeled proteins were added to scintillation mixture and were quantified by liquid scintillation counting. Specific phosphorylation was calculated as the difference between counts obtained with G-13 IgG and preimmune IgG.
Immunoblot Analysis-Proteins were transferred electrophoretically from SDS gels onto nitrocellulose membranes according to the method of Towbin et al. (56), and membranes were blocked with 3% gelatin in Tris-buffered saline. The membranes were incubated overnight at room temperature with the indicated dilution of G␣ 13 -N IgG or preimmune IgG. Membranes were then washed with 0.1% Tween in Tris-buffered saline and were treated with biotinylated goat anti-rabbit IgG (H ϩ L) as the secondary antibody. Immunoreactive proteins were conjugated with horseradish peroxidase-labeled avidin and were detected using 4-chloro-1-naphthol as the substrate (0.5 mg/ml).
Platelet Aggregation-Blood was drawn into acid/citrate/dextrose anticoagulant, and PRP was prepared as described above. Platelets were treated for 1 min with indomethacin (10 M) in order to prevent endogenous TXA 2 production. Aggregation was stimulated with either U46619, ADP, TRAP, or A23187 at a concentration that yielded approximately 85% of the maximal aggregation response (sub-maximal dose), and the aggregation response was measured by the turbidimetric method (57) using a model 400 Lumi-aggregometer (Chronolog Corp., Havertown, PA). The inhibitory effect of cAMP was measured by adding PGI 2 (0.25-1 nM) to the PRP 1 min prior to addition of agonist. The concentration of PGI 2 was titrated to a dose that substantially inhibited U46619-induced aggregation (at the sub-maximal concentration), and the effect of that same dose of PGI 2 was measured on aggregation induced by the sub-maximal concentrations of ADP, TRAP, and A23187. The dose-dependent inhibitory effect of PGI 2 was determined in indomethacin-treated platelets by incubating PRP for 1 min with various concentrations of PGI 2 , as indicated, and subsequently stimulating aggregation with the sub-maximal dose of each agonist.
Statistical Analysis-Data were analyzed according to the Student's two-tailed paired t test or one-way analysis of variance, as indicated, using GraphPad PRISM statistical software (San Diego, CA). Statistical significance is defined as p Ͻ 0.05 or p Ͻ 0.01, as indicated.

cAMP-dependent Phosphorylation of Solubilized Platelet
Membranes-In order to study the possibility that cAMP-dependent kinase can induce phosphorylation of G-protein ␣ subunits, we examined whether cAMP induces phosphorylation of a platelet membrane protein(s) at the molecular weight region in which G␣ proteins have been detected immunologically (58 -60). In these experiments, platelet membranes were solubilized in CHAPS and were phosphorylated by [␥-32 P]ATP in the presence or absence of cAMP (1 M). After a 30-min incubation, the reaction was terminated by adding samples to Laemmli buffer, and the phosphorylated proteins were separated by SDS-PAGE and were visualized by silver staining. Detection of the 32 Plabeled proteins by autoradiography revealed that cAMP induced substantial phosphorylation at several molecular masses, notably at 22 and 38 45 kDa (Fig. 1). The PKA dependence of this phosphorylation was next examined using the specific PKA inhibitor IP 20 to inhibit endogenous PKA activity. In these experiments, solubilized platelet membranes were incubated with [␥-32 P]ATP in the presence of cAMP (1 M), cAMP (1 M) plus IP 20 (3.6 M), or vehicle. After a 30-min incubation, the phosphorylated proteins were resolved by SDS-PAGE. It can be seen from the autoradiogram (Fig. 2) that the increase in phosphorylation induced by cAMP in both the 22-kDa and the 38 -45-kDa regions (lane 2) was inhibited to nearly control levels (lane 1) by IP 20 (lane 3). These results, therefore, establish that a significant amount of the observed phosphorylation is PKA-specific. Although the 22-kDa protein is thought to be PKA-phosphorylated thrombolamban (32,33), the identity of the 38 -45-kDa proteins is currently unknown. However, since G-protein ␣ subunits are known to migrate to this molecular weight region upon electrophoresis, the next series of experiments investigated whether the 38 -45-kDa phosphoproteins may represent PKA-phosphorylated G-proteins.
Purification of TXA 2 Receptor-G-protein Complexes-Previous studies in our laboratory have demonstrated that ligand affinity chromatography purification of TXA 2 receptors from solubilized platelet membranes results in co-purification of TXA 2 receptors with their associated G-proteins (13,17,52). Based on this consideration, experiments that employed this technique were conducted to determine whether 38 -45-kDa phosphoproteins are TXA 2 receptor-associated. In these studies, platelet membranes were phosphorylated by [␥- 32  bation, the membranes were solubilized in CHAPS and were incubated with the TXA 2 receptor affinity matrix. After washing the column to remove unbound proteins, TXA 2 receptor-Gprotein complexes were specifically eluted using the TXA 2 receptor antagonist BM13.177. The column eluate was subjected to SDS-PAGE to resolve the purified proteins, and the gel was silver-stained. It can be seen from the autoradiogram (Fig. 3) that phosphoproteins at approximately 44 kDa co-eluted with the TXA 2 receptors upon ligand affinity chromatography purification. Analysis of the autoradiogram by densitometry revealed that cAMP induced a 97 Ϯ 13% increase in phosphorylation of these TXA 2 receptor-associated proteins. To verify that the increase in phosphorylation was not due to an increase in the amount of receptor eluted from each column, the silverstained receptor bands were analyzed by densitometry. Thus, it was confirmed that the same amount (Ϯ4%) of receptor protein was eluted from each column.
Similar phosphorylation results were obtained using a preparation of solubilized platelet membranes. In these experiments, CHAPS-solubilized platelet membranes were phosphorylated by [␥-32 P]ATP in the presence or absence of cAMP (1 M). The phosphorylated, solubilized membranes were then purified by ligand affinity chromatography and were separated by SDS-PAGE. Autoradiograms from these experiments also revealed the presence of TXA 2 receptor-associated phosphoproteins in the 38 -45-kDa range (Fig. 4). Densitometric analysis of the 38 -45-kDa region demonstrated that cAMP induced a 64 Ϯ 9% increase in phosphorylation of the TXA 2 receptorassociated proteins. Therefore, in both intact membranes and in solubilized membranes, it appears that cAMP causes phosphorylation of proteins that are in the molecular weight region of ␣ subunits and that co-purify with TXA 2 receptors.
Immunopurification of Phosphorylated G␣ 13 from Solubilized Membranes-Experiments were next performed to identify these TXA 2 receptor-associated phosphoproteins. Recent results from our laboratory have demonstrated that G␣ 13 copurifies with platelet TXA 2 receptors by ligand affinity chromatography (17). Based on this observation, we probed the ligand affinity column eluate for the presence of cAMP-phosphorylated G␣ 13 . Specifically, solubilized platelet membranes, which were phosphorylated by [␥-32 P]ATP in the presence or absence of the hydrolysis-resistant cAMP analog 8-Br-cAMP (1 mM), were purified by ligand affinity chromatography, as described previously. The column eluate was then probed with G␣ 13 -N IgG, which is an anti-peptide antibody that is specific for the amino terminus of human G␣ 13 (Table I). Following incubation of the column eluate with this antibody, the 32 Plabeled immune complexes were bound to protein A-Sepharose beads and were precipitated by centrifugation. The washed, immunoprecipitated phosphoproteins were then quantified by scintillation spectrometry. Control experiments were performed in parallel using a preimmune IgG in order to determine nonspecific phosphorylation. Specific phosphorylation is represented as the difference between G␣ 13 antibody counts and preimmune antibody counts from the immunoprecipitated proteins. It was found in these experiments that treatment with 8-Br-cAMP caused an 87 Ϯ 27% increase in specific phosphorylation of G␣ 13 (Fig. 5A). Immunoblot analysis of proteins precipitated by the G␣ 13 -N antibody verified that G␣ 13 was precipitated from both control and phosphorylated solubilized membranes, indicating that this antibody recognizes both the non-phosphorylated and phosphorylated form of the native Gprotein (Fig. 5B). These results, therefore, provide evidence that cAMP causes phosphorylation of G 13 ␣ subunits and, furthermore, indicate that these phosphorylated ␣ subunits are in physical association with platelet TXA 2 receptors.
Further evidence demonstrating the ability of cAMP to phosphorylate G␣ 13 was provided using a different antibody against this G-protein as well as a separate technique to specifically isolate phosphorylated G 13 ␣ subunits. Briefly, a rabbit polyclonal, anti-peptide antibody was raised against an internal amino acid sequence that is unique to G␣ 13 (Table I) (17), and the G-13 IgG was coated onto microtiter wells. Solubilized platelet membranes were then phosphorylated by [␥-32 P]ATP in the presence or absence of 8-Br-cAMP (1 mM) and were subsequently incubated on the antibody-coated microtiter wells. After washing the wells to remove unbound proteins, specifically immobilized 32 P-labeled proteins were eluted and were quantified by liquid scintillation counting. In parallel experiments, a preimmune antibody control was used to determine nonspecific phosphorylation. Specific phosphorylation is represented as the difference between G␣ 13 antibody counts and preimmune antibody counts. In these experiments it was found that 8-Br-cAMP induced a 53 Ϯ 19% increase in specific phosphorylation of G␣ 13 (Fig. 6).
Immunopurification of G␣ 13 from 32 P-Labeled Whole Platelets-The ability of G␣ 13 to undergo cAMP-mediated phosphorylation in intact cells was next examined by treating platelets with the physiological inhibitor PGI 2 . In these experiments, platelets were labeled with [ 32 P]P i and were subsequently incubated for 5 min with either PGI 2 (1 M) in the presence of the phosphodiesterase inhibitor IBMX (1 mM) or with vehicle. The cells were then cooled in an ice slurry and were solubilized in CHAPS buffer. G␣ 13 was immunoprecipitated from the solubilized cells using the G␣ 13 -N IgG, and the precipitated 32 P counts were quantified by scintillation counting, and the counts were normalized for the amount of precipitated G␣ 13 protein. In these experiments, it was found that treatment of platelets with PGI 2 caused a 90 Ϯ 14% increase in phosphorylation of G␣ 13 as compared with control platelets (Fig. 7). Finally, when adenylate cyclase activity was blocked using the cell-permeable specific cyclase inhibitor MDL-12, PGI 2 -induced phosphorylation of G␣ 13 was completely abolished, i.e. Ϫ17 Ϯ 16% (n ϭ 2). These findings, therefore, establish that activation of platelet PGI 2 receptors leads to cAMP-dependent phosphorylation of G␣ 13 .
Inhibition of Platelet Aggregation by PGI 2 -The above results demonstrate cAMP-mediated phosphorylation of TXA 2 receptor-associated G␣ 13 . Although the consequence of such phosphorylation on TXA 2 -induced signal transduction is unknown, the following experiments demonstrate that platelet aggregation induced by TXA 2 is preferentially sensitive to inhibition by PGI 2 . In these studies, platelets were treated with indomethacin to inhibit endogenous TXA 2 production. Subsequently, platelets were pretreated with PGI 2 for 1 min to elevate cAMP levels, and aggregation was measured in response to sub-maximal concentrations of the TXA 2 mimetic U46619, ADP, TRAP, and A23187 (Fig. 8A). The submaximal dose represents an agonist concentration that yields the same biological response for each agonist in the absence of PGI 2 , i.e. 85% of the maximal response. As illustrated in Fig. 8B, a concentration of PGI 2 that caused significant inhibition of U46619-induced aggregation (62%) had only a minimal effect on aggregation induced by ADP (5%), TRAP (10%), or the calcium ionophore A23187 (15%). Furthermore, when various concentrations of PGI 2 were tested, U46619-induced aggregation was found to be far more sensitive to inhibition by PGI 2 than any of the other agonists tested (Fig. 8C).  13 . A, solubilized platelet membranes were phosphorylated by [␥-32 P]ATP in the presence (filled column) or absence (open column) of 1 mM 8-Br-cAMP, and TXA 2 receptor-G-protein complexes were then purified by ligand affinity chromatography as described. The column eluate was incubated with G␣ 13 -N or preimmune IgG, and immunoprecipitation was performed as described under "Experimental Procedures." The precipitated, labeled protein was quantified by liquid scintillation counting, and the specifically eluted counts were determined as described in the text. Results are the average of six separate experiments Ϯ the S.E. Statistical analysis was performed using the Student's paired t test. *, p Ͻ 0.05. B, immunoblot of phosphorylated and non-phosphorylated proteins precipitated by the G␣ 13 -N antibody. Solubilized platelet membranes were phosphorylated in the absence (lane 1) or presence (lane 2) of 8-Br-cAMP and were incubated with G␣ 13 -N or preimmune IgG (PI IgG). Immunoprecipitation was performed as described, and the precipitated proteins were detected by immunoblot analysis using a 1:100 dilution of the G␣ 13 -N antibody, as described under "Experimental Procedures." P-labeled protein was eluted and quantified by scintillation counting, and the specific counts were determined as described under "Experimental Procedures." Results are the average of five separate experiments Ϯ the S.E. Statistical analysis was performed using the Student's paired t test. *, p Ͻ 0.05.

DISCUSSION
This study demonstrates that cAMP induces PKA-dependent phosphorylation of solubilized platelet membrane proteins in the molecular weight region of G-protein ␣ subunits, i.e. 38 -45 kDa, which is consistent with other reports (61,62). Furthermore, it shows that platelet proteins that are phosphorylated by cAMP co-purify with TXA 2 receptors and are, therefore, closely associated with these receptors. Immunoaffinity purification experiments have identified, in both solubilized platelet membranes and intact platelets, that one of these phosphoproteins is TXA 2 receptor-associated G␣ 13 .
Previous reports have established that the G 13 ␣ subunit is a 43-44-kDa protein which is expressed ubiquitously among cells (14,15,60,63). Although the role of G␣ 13 in platelets is not understood, it has been linked to a number of cellular functions in other cell systems, including stimulation of the Na/H exchanger (18,20), activation of L-type Ca 2ϩ channels (23,24), activation of phospholipase D (25), and organization of the cytoskeleton through Rho-dependent stress fiber formation and focal adhesion assembly (21). Based on these considerations, modulation of G␣ 13 functional activity could potentially impact a number of different effector systems.
Classically, cellular activity is modulated by phosphorylation of signal transduction components. Thus, signaling through a receptor-G-protein complex can be modified by phosphorylation of either the receptor or the G-protein. In this regard, it is well established that phosphorylation of receptors by different protein kinases can mediate receptor desensitization (35)(36)(37)(38)(39)(40). Previous reports have indicated that platelet TXA 2 receptors can indeed serve as substrates for PKC, PKA, and PKG (26 -29). Specifically, Kinsella et al. (26) originally demonstrated that PKC and PKA caused phosphorylation of a fusion protein containing the latter third of the TXA 2 receptor. This finding suggested a possible mechanism by which phorbol 12-myristate 13-acetate (PMA) or cAMP may inhibit TXA 2 -induced platelet activation is through phosphorylation of the receptor protein. However, subsequent results by the same group, using a transfected HEK 293 cell line, led them to conclude that this receptor phosphorylation was of "trivial" physiological significance and was not a major mechanism by which PMA or cAMP modifies TXA 2 receptor signal transduction (27). Recently, this group has reported that upon stimulation of platelets with various agonists, PKC can indeed induce phosphorylation of TXA 2 receptors (28). Thus, differing results appear to be obtained depending on the cell system and experimental conditions employed.
Regarding cGMP, however, a separate group has recently provided evidence that PKG causes TXA 2 receptor phosphorylation (29). Thus, 8-Br-cGMP-induced activation of PKG resulted in increased phosphorylation of immunoaffinity purified TXA 2 receptor protein. In addition, these studies also demonstrated that 8-Br-cGMP caused a decrease in U46619-stimulated GTPase activity in platelet membranes. Based on these results, the authors (29) proposed that cGMP-induced phosphorylation of TXA 2 receptors modulates receptor-G-protein coupling to regulate signaling through these receptors.
In addition to receptor phosphorylation, it has been recently recognized that G-protein phosphorylation may also be an important regulatory mechanism. In this regard, it was found that PKC-induced phosphorylation of purified G z and G 12 ␣ subunits resulted in inhibition of their interaction with ␤␥ subunits (46,64). In addition, PKC-induced phosphorylation of G␣ i was found to correlate with attenuation of adenylate cyclase inhibition in S49 cells (44) and inhibition of agonistinduced calcium mobilization in platelets (45). Finally, separate studies have found that PGD 2 , which elevates cAMP, inhibits thrombin-induced dissociation of G␣ i from platelet thrombin receptors (65). Therefore, G-protein phosphorylation may functionally uncouple signal transduction by altering the stability of the heterotrimer or by tightening its association with the receptor.
In light of these findings, the present results demonstrating cAMP-mediated phosphorylation of G␣ 13 suggest a potential mechanism for modulation of TXA 2 signal transduction in platelets. While effectors for G␣ 13 in platelets are currently unknown, previous results have established that calcium mobilization in the TXA 2 receptor signaling pathway is modulated by cAMP (34,66). Specifically these studies have shown that there is a direct relationship between platelet cAMP levels and inhibition of U46619-induced calcium release (34). Moreover, inhibition of TXA 2 -induced calcium release from the plateletdense tubular system was found to be PKA-dependent (66). The above findings, coupled with the notion that G␣ 13 has been linked to calcium channels in other cells (23,24), raises the possibility that G␣ 13 phosphorylation may serve as one mechanism by which cAMP inhibits TXA 2 -mediated calcium mobilization in platelets.
In summary, the present results provide the first evidence that PGI 2 can cause PKA-induced phosphorylation of an endogenous platelet G-protein ␣ subunit and that phosphorylated G␣ 13 is associated with platelet TXA 2 receptors. Furthermore, our results also demonstrate a preferential sensitivity of TXA 2 receptor signaling to PGI 2 -stimulated increases in cAMP. Taken together, these results, therefore, suggest an alternative mechanism for cAMP inhibition of TXA 2 receptor signaling and may explain the preferential sensitivity of this signaling pathway to elevated cAMP.