Portal de Periódicos da CAPES

Signaling through Gi Family Members in Platelets

2002; Elsevier BV; Volume: 277; Issue: 48 Linguagem: Inglês

10.1074/jbc.m208519200

1083-351X

Jing Yang, Jie Wu, Hong Jiang, Richard M. Mortensen, Sandra Austin, David R. Manning, Donna S. Woulfe, Lawrence F. Brass,

Platelet Disorders and Treatments

Platelet responses at sites of vascular injury are regulated by intracellular cAMP levels, which rise rapidly when prostacyclin (PGI2) is released from endothelial cells. Platelet agonists such as ADP and epinephrine suppress PGI2-stimulated cAMP formation by activating receptors coupled to Gi family members, four of which are present in platelets. To address questions about the specificity of receptor:G protein coupling, the regulation of cAMP formation in vivo and the contribution of Gi-mediated pathways that do not involve adenylyl cyclase, we studied platelets from mice that lacked the α subunits of one or more of the three most abundantly expressed Gi family members and compared the results with platelets from mice that lacked the PGI2receptor, IP. As reported previously, loss of Gi2α or Gzαinhibited aggregation in response to ADP and epinephrine, respectively, producing defects that could not be reversed by adding an adenylyl cyclase inhibitor. Platelets that lacked both Gi2α and Gzα showed impaired responses to both agonists, but the impairment was no greater than in the individual knockouts. Loss of Gi3α had no effect either alone or in combination with Gzα. Loss of either Gzα or Gi2αimpaired the ability of ADP and epinephrine to inhibit PGI2-stimulated adenylyl cyclase activity and caused a 40%-50% rise in basal cAMP levels, whereas loss of Gi3α did not. Conversely, deletion of IP abolished responses to PGI2 and caused cAMP levels to fall by 30%, effects that did not translate into enhanced responsiveness to agonists ex vivo. From these results we conclude that 1) cAMP levels in circulating platelets reflect ongoing signaling through Gi2, Gz, and IP, but not Gi3; 2) platelet epinephrine (α2A-adrenergic) and ADP (P2Y12) receptors display strong preferences among Gi family members with little evidence of redundancy; and 3) these receptor preferences do not extend to Gi3. Finally, the failure of ADP and epinephrine to inhibit basal, as opposed to PGI2-stimulated, cAMP formation highlights the need during platelet activation for Gi signaling pathways that involve effectors other than adenylyl cyclase. Platelet responses at sites of vascular injury are regulated by intracellular cAMP levels, which rise rapidly when prostacyclin (PGI2) is released from endothelial cells. Platelet agonists such as ADP and epinephrine suppress PGI2-stimulated cAMP formation by activating receptors coupled to Gi family members, four of which are present in platelets. To address questions about the specificity of receptor:G protein coupling, the regulation of cAMP formation in vivo and the contribution of Gi-mediated pathways that do not involve adenylyl cyclase, we studied platelets from mice that lacked the α subunits of one or more of the three most abundantly expressed Gi family members and compared the results with platelets from mice that lacked the PGI2receptor, IP. As reported previously, loss of Gi2α or Gzαinhibited aggregation in response to ADP and epinephrine, respectively, producing defects that could not be reversed by adding an adenylyl cyclase inhibitor. Platelets that lacked both Gi2α and Gzα showed impaired responses to both agonists, but the impairment was no greater than in the individual knockouts. Loss of Gi3α had no effect either alone or in combination with Gzα. Loss of either Gzα or Gi2αimpaired the ability of ADP and epinephrine to inhibit PGI2-stimulated adenylyl cyclase activity and caused a 40%-50% rise in basal cAMP levels, whereas loss of Gi3α did not. Conversely, deletion of IP abolished responses to PGI2 and caused cAMP levels to fall by 30%, effects that did not translate into enhanced responsiveness to agonists ex vivo. From these results we conclude that 1) cAMP levels in circulating platelets reflect ongoing signaling through Gi2, Gz, and IP, but not Gi3; 2) platelet epinephrine (α2A-adrenergic) and ADP (P2Y12) receptors display strong preferences among Gi family members with little evidence of redundancy; and 3) these receptor preferences do not extend to Gi3. Finally, the failure of ADP and epinephrine to inhibit basal, as opposed to PGI2-stimulated, cAMP formation highlights the need during platelet activation for Gi signaling pathways that involve effectors other than adenylyl cyclase. Platelet activation at sites of vascular injury typically begins with the exposure of collagen within the subendothelial connective tissue matrix and then continues as additional platelets are recruited into the growing platelet mass by soluble agonists such as thrombin, ADP, epinephrine, and thromboxane A2. To balance the ability of platelets to be rapidly activated when needed, a number of regulators exist to prevent unwarranted platelet activation. Two such regulators are PGI2 1The abbreviations used for: PGI2, prostacyclin; IP, prostacyclin receptor; PAR4, proteinase-activated receptor-4; TBS, Tris-buffered saline. and NO, both of which are released from activated endothelial cells. PGI2 suppresses intracellular signaling by stimulating adenylyl cyclase and causing an increase in intracellular cAMP (1Moncada S. Ryglewski R. Bunting S. Vane J.R. Nature. 1976; 263: 663-665Crossref PubMed Scopus (2940) Google Scholar, 2Miller O.V. Gorman R.R. J. Cyclic Nucleotide Res. 1976; 2: 79-87PubMed Google Scholar, 3Gorman R.R. Fitzpatrick F.A. Miller O.V. Adv. Cyclic Nucleotide Res. 1978; 9: 597-609PubMed Google Scholar). NO causes an increase in cGMP that, among other effects, inhibits cAMP phosphodiesterase (4Haslam R.J. Dickinson N.T. Jang E.K. Thromb. Haemostasis. 1999; 82: 412-423Crossref PubMed Scopus (150) Google Scholar). Previous investigators have shown that even a modest increase in cAMP levels (far less than the maximum that can be produced by incubating platelets with PGI2 or PGE1) will inhibit platelet responses to thrombin and ADPin vitro (5Mills D.C.B. Smith J.B. Biochem. J. 1971; 121: 185-196Crossref PubMed Scopus (316) Google Scholar, 6Eigenthaler M. Nolte C. Halbrugge M. Walter U. Eur. J. Biochem. 1992; 205: 471-481Crossref PubMed Scopus (146) Google Scholar, 7Keularts I.M.L.W. Van Gorp R.M.A. Feijge M.A.H. Vuist W.M.J. Heemskerk J.W.M. J. Biol. Chem. 2000; 275: 1763-1772Abstract Full Text Full Text PDF PubMed Scopus (85) Google Scholar). Experiments performed in vivosuggest the same conclusion, because drugs that stimulate PGI2 receptors, inhibit cAMP phosphodiesterases, or block ADP receptors can be effective anti-platelet agents (8Schwarz U.R. Walter U. Eigenthaler M. Biochem. Pharmacol. 2001; 62: 1153-1161Crossref PubMed Scopus (286) Google Scholar, 9Brass S. Nature. 2001; 409: 145-147Crossref PubMed Scopus (15) Google Scholar). Conversely, deletion of the gene that encodes the platelet PGI2receptor (IP) enhances thrombosis after arterial injury in mice (10Murata T. Ushikubi F. Matsuoka T. Hirata M. Yamasaki A. Sugimoto Y. Ichikawa A. Aze Y. Tanaka T. Yoshida N. Ueno A. Oh-Ishi S. Narumiya S. Nature. 1997; 388: 678-682Crossref PubMed Scopus (689) Google Scholar), presumably because the inability to respond to PGI2enhances platelet responsiveness, although that point has not been fully addressed. Observations such as these contribute to the current view that maintenance of the intracellular cAMP concentration within narrowly defined limits is essential for normal platelet function in vivo. As in other types of cells, cAMP formation in platelets is regulated bimodally by activation of G protein-coupled receptors that stimulate or inhibit adenylyl cyclase. PGI2 increases cAMP formation via Gs-coupled IP receptors, whereas ADP and epinephrine inhibit cAMP formation via receptors coupled to Gi family members. Four Gi family members are expressed in platelets: Gi1, Gi2, Gi3, and Gz. Of these, Gi2, Gz, and Gi3 are the most abundant (11Williams A. Woolkalis M.J. Poncz M. Manning D.R. Gewirtz A. Brass L.F. Blood. 1990; 76: 721-730Crossref PubMed Google Scholar, 12Gagnon A.W. Manning D.R. Catani L. Gewirtz A. Poncz M. Brass L.F. Blood. 1991; 78: 1247-1253Crossref PubMed Google Scholar). Previous studies with receptor antagonists and with genetically engineered mice show that ADP inhibits adenylyl cyclase via P2Y12 purinergic receptors primarily coupled to Gi2 (13Jantzen H.M. Milstone D.S. Gousset L. Conley P.B. Mortensen R.M. J. Clin. Invest. 2001; 108: 477-483Crossref PubMed Scopus (135) Google Scholar), whereas epinephrine inhibits adenylyl cyclase via α2A-adrenergic receptors primarily coupled to Gz (14Yang J. Wu J. Kowalska M.A. Dalvi A. Prevost N. O'Brien P.J. Manning D. Poncz M. Lucki I. Blendy J.A. Brass L.F. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 9984-9989Crossref PubMed Scopus (171) Google Scholar). However, a number of fundamental questions remain unanswered. How is the basal cAMP concentration established in circulating platelets? Will relatively small chronic changes in the cAMP concentration produce the same global changes in platelet responsiveness that are seen after acute changes? How critical is the ability of Gi-coupled agonists to inhibit adenylyl cyclase when platelets are not exposed to PGI2? How much redundancy and selectivity exist among the Gi family members in platelets, three of which (Gi1, Gi2, and Gi3) have α subunits that are nearly identical? If selectivity exists, how much of it occurs at the level of receptor:G protein coupling and how much occurs at the level of G protein:effector interactions? To what extent is the well established requirement for Gi-mediated signaling during platelet activation caused by the need to suppress cAMP formation, and to what extent is it caused by the activation of other downstream effectors such as phosphatidylinositol 3-kinase γ? To answer these questions we examined the regulation of cAMP formation and agonist responsiveness in platelets from mice that lack the α subunit of one or more Gi family members or the platelet PGI2 receptor, IP. The results provide further evidence for functional differences among the Gi family members in platelets and show that maintenance of a normal basal cAMP concentration reflects ongoing signaling through Gi2, Gz, and IP but not Gi3. The results also show that a decrease or increase in the basal cAMP concentration does not by itself make platelets more or less sensitive to agonists, and that unless PGI2 is present, platelet agonists have little or no effect on platelet cAMP levels. Finally, the results emphasize the importance of Gi-coupled effectors other than adenylyl cyclase. U46619 was obtained from Calbiochem. Collagen was obtained from Chrono-log (Havertown, PA). PGI2, PGE1, and SQ22536 were obtained from Sigma. The PAR4 agonist peptide, AYPGQV, was synthesized by core facilities at the University of Pennsylvania School of Medicine. Gzα(−/−) (14Yang J. Wu J. Kowalska M.A. Dalvi A. Prevost N. O'Brien P.J. Manning D. Poncz M. Lucki I. Blendy J.A. Brass L.F. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 9984-9989Crossref PubMed Scopus (171) Google Scholar), Gi2α(−/−) (13Jantzen H.M. Milstone D.S. Gousset L. Conley P.B. Mortensen R.M. J. Clin. Invest. 2001; 108: 477-483Crossref PubMed Scopus (135) Google Scholar, 15Jain M. Lim C.C. Nagata K. Davis V.M. Milstone D.S. Liao R. Mortensen R.M. Am. J. Physiol. 2001; 280: H569-H575Crossref PubMed Google Scholar, 16Nagata K. Ye C. Jain M. Milstone D.S. Liao R. Mortensen R.M. Circ. Res. 2000; 87: 903-909Crossref PubMed Scopus (58) Google Scholar), Gi3α(−/−) (15Jain M. Lim C.C. Nagata K. Davis V.M. Milstone D.S. Liao R. Mortensen R.M. Am. J. Physiol. 2001; 280: H569-H575Crossref PubMed Google Scholar, 16Nagata K. Ye C. Jain M. Milstone D.S. Liao R. Mortensen R.M. Circ. Res. 2000; 87: 903-909Crossref PubMed Scopus (58) Google Scholar) and IP(−/−) (17Austin S.C. Stalker T.J. Scalia R. Lefer A.M. FitzGerald G.A. Arterioscler. Thromb. Vasc. Biol. 2001; 21 (abstr.): 709Google Scholar, 18Cheng Y. Austin S.C. Rocca B. Koller B.H. Coffman T.M. Grosser T. Lawson J.A. FitzGerald G.A. Science. 2002; 296: 539-541Crossref PubMed Scopus (732) Google Scholar) mice were generated as described previously. Double knockout mice were produced by breeding single knock-outs with heterozygous mice. Blood was collected from the inferior vena cava of anesthetized mice (100 mg/kg pentobarbital) by using a heparinized syringe (15 units/ml blood). Samples from four mice were pooled and diluted with 3 ml of HEPES/Tyrode buffer (129 mmNaCl, 8.9 mm NaHCO3, 2.8 mm KCl, 0.8 mm KH2PO4, 10 mmHEPES, 0.8 mm MgCl2, 5.6 mmdextrose, pH 7.4). Red cells were removed by centrifugation, and the final platelet count was adjusted to 2–3 × 108/ml by using HEPES/Tyrode buffer. Aggregation was measured in a lumi-aggregometer (Chrono-log). Washed platelets were incubated as noted in each experiment in the absence of a phosphodiesterase inhibitor (unless otherwise indicated). The reaction was then stopped by adding ice-cold, 10% trichloroacetic acid. cAMP was measured by radioimmunoassay (PerkinElmer Life Sciences). Platelet lysates in hypotonic buffer (25 mm HEPES, pH 7.5, 1 mm EDTA, and 1 mm dithiothreitol) plus a protease inhibitor mixture (Sigma) were centrifuged at 660 × g for 10 min at 4 °C to remove cell debris and then frozen and thawed three times. Membranes were pelleted at 10,000 × g for 30 min at 4 °C and resuspended in radioimmune precipitation assay buffer (1% Nonidet P-40, 0.5% sodium deoxycholate, and 0.1% SDS in phosphate-buffered saline). Protein samples (20 μg) were loaded on 12% SDS-PAGE gels, separated, and then transferred to polyvinylidene difluoride membranes. The membranes were blocked with 5% nonfat dry milk in TBS for 1 h at room temperature and then incubated with the appropriate antisera (1:200 dilution) in TBS with 1% gelatin for 2 h at room temperature. Afterward, the polyvinylidene difluoride membranes were washed three times with ×0.2% Tween 20 in TBS for 10 min and then incubated with horseradish peroxidase-conjugated anti-rabbit immunoglobulin (1:2000 × dilution, Amersham Biosciences) for 1 h at room temperature. The blot was then washed three times with Tris-buffered saline/Tween and analyzed by using an ECL Western blot detection kit (AmershamBiosciences). ERK (extracellular signal-regulated kinase)-1 was detected by using a rabbit polyclonal antiserum purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Basal cAMP concentration is determined by the balance between cAMP formation and hydrolysis and potentially sets the tone for platelet responsiveness to agonists. Therefore, as a first step we measured the basal cAMP concentration in platelets isolated from mice that lacked Gi2α, Gzα, Gi3α, or the PGI2 receptor (IP) and compared them with their wild type littermates. "Basal" was defined as the intracellular cAMP concentration in the absence of either a phosphodiesterase inhibitor or an adenylyl cyclase activator. The basal cAMP concentration in the wild type mice was 4.7 ± 0.7 pmol/108 platelets (mean ± S.E., n = 16), which is very similar to values reported previously for human platelets (6Eigenthaler M. Nolte C. Halbrugge M. Walter U. Eur. J. Biochem. 1992; 205: 471-481Crossref PubMed Scopus (146) Google Scholar, 7Keularts I.M.L.W. Van Gorp R.M.A. Feijge M.A.H. Vuist W.M.J. Heemskerk J.W.M. J. Biol. Chem. 2000; 275: 1763-1772Abstract Full Text Full Text PDF PubMed Scopus (85) Google Scholar, 19Zahavi M. Zahavi J. Kakkar V.V. Thromb. Haemostasis. 1984; 52: 205-209Crossref PubMed Scopus (14) Google Scholar). Deletion of the genes that encode Gi2α or Gzα caused a 40%-50% increase in the basal cAMP concentration, whereas loss of IP caused the basal cAMP concentration to fall by 30% (Fig. 1). Loss of Gi3α had no effect (not shown). These results suggest that the set point for the basal cAMP concentration in platelets is determined in part by exposure to endothelium-derived PGI2 and in part by signaling through Gi family members. This fact is presumably the result of ongoing exposure of platelets to agonists having receptors that are coupled to Gi2 or Gz but not to Gi3. On the basis of studies in which the cAMP concentration was manipulatedin vitro by incubation of platelets with PGE1(6Eigenthaler M. Nolte C. Halbrugge M. Walter U. Eur. J. Biochem. 1992; 205: 471-481Crossref PubMed Scopus (146) Google Scholar, 7Keularts I.M.L.W. Van Gorp R.M.A. Feijge M.A.H. Vuist W.M.J. Heemskerk J.W.M. J. Biol. Chem. 2000; 275: 1763-1772Abstract Full Text Full Text PDF PubMed Scopus (85) Google Scholar), in theory even the small increase in basal cAMP levels caused by deletion of Gi2α or Gzα could be sufficient to inhibit platelet responses to agonists. If so, this loss of responsiveness should extend to agonists whose receptors are not directly coupled to the missing G protein as well as those that are. Defects in platelet aggregation for mice that lack either Gi2α or Gzα have been described previously (13Jantzen H.M. Milstone D.S. Gousset L. Conley P.B. Mortensen R.M. J. Clin. Invest. 2001; 108: 477-483Crossref PubMed Scopus (135) Google Scholar, 14Yang J. Wu J. Kowalska M.A. Dalvi A. Prevost N. O'Brien P.J. Manning D. Poncz M. Lucki I. Blendy J.A. Brass L.F. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 9984-9989Crossref PubMed Scopus (171) Google Scholar). In the absence of Gzα, epinephrine is unable to potentiate the effects of other platelet agonists except when added at supraphysiologic concentrations (14Yang J. Wu J. Kowalska M.A. Dalvi A. Prevost N. O'Brien P.J. Manning D. Poncz M. Lucki I. Blendy J.A. Brass L.F. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 9984-9989Crossref PubMed Scopus (171) Google Scholar). Responses to other agonists are normal. In the absence of Gi2α, platelet responses to ADP are reduced, whereas responses to other agonists are normal except to the extent that they require reinforcement by secreted ADP (Ref. 13Jantzen H.M. Milstone D.S. Gousset L. Conley P.B. Mortensen R.M. J. Clin. Invest. 2001; 108: 477-483Crossref PubMed Scopus (135) Google Scholar and Fig. 2 A). Notably, deletion of the genes that encode Gi2α or Gzα appears to have no effect on the level of expression of other Giα family members or Gβ in platelets (Refs. 13Jantzen H.M. Milstone D.S. Gousset L. Conley P.B. Mortensen R.M. J. Clin. Invest. 2001; 108: 477-483Crossref PubMed Scopus (135) Google Scholar and 14Yang J. Wu J. Kowalska M.A. Dalvi A. Prevost N. O'Brien P.J. Manning D. Poncz M. Lucki I. Blendy J.A. Brass L.F. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 9984-9989Crossref PubMed Scopus (171) Google Scholar and Fig. 2 B). In general, we found that the decrease in ADP-induced aggregation and suppression of cAMP formation (Fig. 2) caused by loss of Gi2α was less striking than the loss of epinephrine responses in Gzα(−/−) platelets (14Yang J. Wu J. Kowalska M.A. Dalvi A. Prevost N. O'Brien P.J. Manning D. Poncz M. Lucki I. Blendy J.A. Brass L.F. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 9984-9989Crossref PubMed Scopus (171) Google Scholar). In contrast, loss of Gi3α had no apparent effect on platelet activation by ADP, epinephrine, or AYPGQV, a peptide agonist for the murine PAR4 thrombin receptor (Fig.3). Western blotting confirmed that Gi3α was no longer expressed in the Gi3(−/−) platelets and that no compensatory changes occurred in the levels of expression of other G proteins (not shown).Figure 3Analysis of platelets from Gi3α(−/−) mice. A, platelet aggregation in response to (top) 10 μm ADP or (middle) 0.2 μm ADP plus 1 μm epinephrine or (bottom) 200 μm AYPGQF, a PAR4 agonist peptide. The results shown are representative of three experiments.B, inhibition of PGI2-stimulated cAMP formation. The final PGI2 concentration was 20 μm. The results shown are the mean from two experiments. WT, wild type.View Large Image Figure ViewerDownload (PPT) Although loss of Gzimpaired responses to epinephrine and loss of Gi2 impaired responses to ADP, neither deletion abolished aggregation. To determine whether the remaining responses in each case are mediated by the other Gi family members, double knockouts were produced by crossing the respective single knockouts. Platelets from mice that lacked both Gi2α and Gzα showed a pattern of aggregation abnormalities that was approximately the sum of those seen in the absence of Gi2α and Gzα individually (Fig.4 A). That is, aggregation of the double knockout platelets was diminished in response to both ADP and epinephrine. However, just as in the Gzα(−/−) platelets, the impaired ability of epinephrine to potentiate responses to other agonists in the Gi2α(−/−)/Gzα(−/−) platelets can be partially overcome by increasing the epinephrine concentration. As illustrated in the experiment shown in Fig.4 A, this increase is achieved by adding epinephrine to platelets along with a suboptimal concentration of the thromboxane A2 receptor agonist, U46619. The response to these agonists could mean that one of the two remaining Gifamily members (Gi1 and Gi3) is able to substitute for Gz when Gi2 is absent. Mice that lacked Gi1 were not available. Gi3α(−/−)/Gzα(−/−) mice were produced by crossing mice carrying the individual knock-outs. The mice themselves were viable and had no evident defects. In general, platelet aggregation by the Gi3α(−/−)/ Gzα(−/−) platelets was indistinguishable from Gzα(−/−) platelets. Reduced potentiation by epinephrine was observed, but responses to other agonists were normal (Fig. 4 B and data not shown). As already noted, deletion of the gene that encodes platelet PGI2receptors caused a 30% decrease in basal cAMP levels (Fig.1 C). It also abolished the increase in cAMP that would otherwise be caused by either PGI2 or PGE1(Fig. 5 A). On the basis of these observations and the sensitivity of platelet responses to small changes in cAMP concentrations, it might be predicted that IP(−/−) platelets would have increased responses to platelet agonists. Using an independently derived strain of IP(−/−) mice, Murata et al. (10Murata T. Ushikubi F. Matsuoka T. Hirata M. Yamasaki A. Sugimoto Y. Ichikawa A. Aze Y. Tanaka T. Yoshida N. Ueno A. Oh-Ishi S. Narumiya S. Nature. 1997; 388: 678-682Crossref PubMed Scopus (689) Google Scholar) observed increased thrombosis after carotid artery injury. However, they found no differences between IP(−/−) and wild type platelets in the rate or extent of aggregation in response to 5 μm ADP (10Murata T. Ushikubi F. Matsuoka T. Hirata M. Yamasaki A. Sugimoto Y. Ichikawa A. Aze Y. Tanaka T. Yoshida N. Ueno A. Oh-Ishi S. Narumiya S. Nature. 1997; 388: 678-682Crossref PubMed Scopus (689) Google Scholar). To see whether such differences could be elicited, we measured aggregation at suboptimal concentrations of ADP,U46619, and collagen. No differences were observed under any of the conditions tested (Fig. 5 B). Therefore, it is apparently not the change in basal cAMP levels that predisposes IP(−/−) mice to thrombosis. We have shown previously that the impaired ability of epinephrine to potentiate platelet aggregation in Gzα(−/−) mice is accompanied by a reduced ability of epinephrine, but not ADP, to inhibit PGI2-stimulated cAMP formation (14Yang J. Wu J. Kowalska M.A. Dalvi A. Prevost N. O'Brien P.J. Manning D. Poncz M. Lucki I. Blendy J.A. Brass L.F. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 9984-9989Crossref PubMed Scopus (171) Google Scholar). Conversely, inhibition of PGI2-stimulated cAMP formation by ADP is reduced in Gi2α(−/−) platelets, whereas inhibition by epinephrine is intact (Ref. 13Jantzen H.M. Milstone D.S. Gousset L. Conley P.B. Mortensen R.M. J. Clin. Invest. 2001; 108: 477-483Crossref PubMed Scopus (135) Google Scholar and Fig. 2 C). Deletion of Gi3α had no effect on this response to either agonist (Fig. 3 B). Given these results, we asked whether ADP or epinephrine would decrease the basal cAMP concentration in wild type platelets and whether a direct inhibitor of adenylyl cyclase would restore the defects in platelet aggregation caused by deletion of a Gi family member. Representative results are shown in Fig.6. In contrast to their ability to inhibit PGI2-stimulated adenylyl cyclase activity, neither ADP nor epinephrine caused a decrease in the cAMP concentration in wild type platelets in the absence of PGI2. Conversely, preincubation of Gzα(−/−) platelets with the membrane-permeable adenylyl cyclase inhibitor, SQ22536, at a concentration sufficient to inhibit PGI2-stimulated cAMP formation to the same extent as ADP or epinephrine in wild type platelets (data not shown) failed to restore a normal response to epinephrine (Fig. 6 B). Thus, under basal conditions the ability to inhibit cAMP formation appears to be neither necessary nor sufficient for platelet aggregation to occur. Human platelets express members of the Gs, Gi, Gq, and G12 families of heterotrimeric G proteins. Current models of platelet activation suggest that each family has a defined role: Gs stimulates adenylyl cyclase, Gq activates phospholipase Cβ, and G12 helps to initiate reorganization of the actin cytoskeleton (13Jantzen H.M. Milstone D.S. Gousset L. Conley P.B. Mortensen R.M. J. Clin. Invest. 2001; 108: 477-483Crossref PubMed Scopus (135) Google Scholar, 14Yang J. Wu J. Kowalska M.A. Dalvi A. Prevost N. O'Brien P.J. Manning D. Poncz M. Lucki I. Blendy J.A. Brass L.F. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 9984-9989Crossref PubMed Scopus (171) Google Scholar, 20Watson S.P. McConnell R.T. Lapetina E.G. J. Biol. Chem. 1984; 259: 13199-13203Abstract Full Text PDF PubMed Google Scholar, 21Ryningen A. Jensen O. Holmsen H. Biochim. Biophys. Acta. 1998; 1394: 235-2348Crossref PubMed Scopus (32) Google Scholar). The four Gi family members that are expressed in platelets are something of a conundrum. Three of them have α subunits that are nearly identical at the protein level. The α subunit of the fourth, Gz, is only 60% identical to the others. It is also not a substrate for pertussis toxin and has a slower intrinsic rate of GTP hydrolysis than the others. Even with these differences and similarities, it is not entirely clear what benefit derives from the expression of so many Gi family members. Recent studies on platelets from mice that lacked Gzα or Gi2α suggest part of the answer. Gz is the preferred partner for α2A-adrenergic receptors (14Yang J. Wu J. Kowalska M.A. Dalvi A. Prevost N. O'Brien P.J. Manning D. Poncz M. Lucki I. Blendy J.A. Brass L.F. Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 9984-9989Crossref PubMed Scopus (171) Google Scholar), whereas Gi2 is the preferred partner for the P2Y12 ADP receptor (Ref. 13Jantzen H.M. Milstone D.S. Gousset L. Conley P.B. Mortensen R.M. J. Clin. Invest. 2001; 108: 477-483Crossref PubMed Scopus (135) Google Scholar and the present study). As a result, loss of Gi2 inhibits ADP responses and loss of Gz inhibits epinephrine responses. Issues left unanswered include the role of the other two Gi family members, Gi3 and Gi1, the basis for the residual response to ADP in the Gi2(−/−) platelets and to epinephrine in the Gz(−/−) platelets, the role of Gi family members in the maintenance of the basal cAMP concentration, and the contribution of effectors other than adenylyl cyclase to the requirement for Gi-dependent signaling during platelet activation. These and related issues have been addressed in the present study. We propose that several conclusions can be drawn from the results. The preference of platelet ADP (P2Y12) receptors and epinephrine (α2A-adrenergic) receptors for Gi2 versus Gz does not appear to extend to Gi3. Deletion of Gi3α had no detectable effect on platelet responses to a variety of agonists. On the other hand, unless it is caused by the small amount of Gi1 expressed in platelets no single Gi family member appears to account for the residual responses to epinephrine in Gz(−/−) platelets or the remaining ADP response in Gi2(−/−) platelets. The double knockout of Gi2 and Gz produced a pattern of impaired agonist responses that was approximately the sum of the individual knockouts, not a more profound defect. The remaining responses to ADP and epinephrine in the Gi2(−/−)/Gz(−/−) platelets were nearly the same as in the individual knockouts. Likewise, deletion of both Gi3 and Gz had no greater effect than loss of Gz alone. Evidently a strongly preferred partner exists, but in the absence of the preferred partner other Gi family members can substitute for the missing G protein in a fairly promiscuous, if less efficient, manner. Perhaps because of their greater similarity in Gα sequence, Gi3 or Gi1 may be able to replace Gi2 to a greater extent than all three of these can replace Gz. This ability might account in part for the observation that P2Y12-mediated ADP responses are affected by deletion of Gi2α to a lesser extent than are α2A-adrenergic receptor-mediated responses to epinephrine in the absence of Gzα. Regardless, it now seems clear that the two receptors most commonly associated with Gi-dependent events in platelets (P2Y12 and α2A-adrenergic) have strong preferences for particular Gi family members. It also appears that, at least at one level, platelets express more than one Gi family member to accommodate those preferences. A second conclusion from these studies concerns the regulation and impact of small changes in the basal cAMP concentration in platelets. Ample evidence exists that small, acute changes in cAMP concentration can have large effects on platelet responses to agonists (5Mills D.C.B. Smith J.B. Biochem. J. 1971; 121: 185-196Crossref PubMed Scopus (316) Google Scholar, 6Eigenthaler M. Nolte C. Halbrugge M. Walter U. Eur. J. Biochem. 1992; 205: 471-481Crossref PubMed Scopus (146) Google Scholar, 7Keularts I.M.L.W. Van Gorp R.M.A. Feijge M.A.H. Vuist W.M.J. Heemskerk J.W.M. J. Biol. Chem. 2000; 275: 1763-1772Abstract Full Text Full Text PDF PubMed Scopus (85) Google Scholar). Among other consequences, an increase in cAMP concentration essentially shuts down agonist-induced phosphoinositide hydrolysis (20Watson S.P. McConnell R.T. Lapetina E.G. J. Biol. Chem. 1984; 259: 13199-13203Abstract Full Text PDF PubMed Google Scholar), impairs resynthesis of phosphatidylinositol-4,5-P2 (21Ryningen A. Jensen O. Holmsen H. Biochim. Biophys. Acta. 1998; 1394: 235-2348Crossref PubMed Scopus (32) Google Scholar), and accelerates the uptake of Ca2+ into the dense tubular system (22Kaser-Glanzmann R. Jakabova M. George J.N. Luscher E.F. Biochim. Biophys. Acta. 1977; 466: 429-440Crossref PubMed Scopus (197) Google Scholar, 23Kaser-Glanzmann R. Gerber E. Luscher E.F. Biochim. Biophys. Acta. 1979; 558: 34-347Google Scholar). Known substrates for protein kinase A in platelets include glycoprotein Ibβ (24Fox J.E.B. Reynolds C.C. Johnson M.M. J. Biol. Chem. 1987; 262: 12627-12631Abstract Full Text PDF PubMed Google Scholar, 25Wardell M.R. Reynolds C.C. Berndt M.C. Wallace R.W. Fox J.E.B. J. Biol. Chem. 1989; 264: 15656-15661Abstract Full Text PDF PubMed Google Scholar), actin-binding protein (26Wallach D. Davies P.J.A. Pastan I. J. Biol. Chem. 1978; 253: 4739-4745Abstract Full Text PDF PubMed Google Scholar), myosin light chain kinase (27Hathway D.R. Adelstein R.S. Proc. Natl. Acad. Sci. U. S. A. 1979; 76: 1653-1657Crossref PubMed Scopus (286) Google Scholar, 28Hallam T.J. Daniel J.L. Kendrick Jones J. Rink T.J. Biochem. J. 1985; 232: 373-377Crossref PubMed Scopus (42) Google Scholar), Rap1B (29Fischer T.H. White G.C. Biochem. Biophys. Res. Commun. 1989; 159: 644-650Crossref PubMed Scopus (8) Google Scholar), and G13α (30Manganello J.M. Djellas Y. Borg C. Antonaskis K. Le Breton G.C. J. Biol. Chem. 1999; 274: 28003-28010Abstract Full Text Full Text PDF PubMed Scopus (47) Google Scholar). The cAMP concentration in mouse platelets (4–5 pmol/108 platelets) is very similar to the basal cAMP concentration in human platelets (6Eigenthaler M. Nolte C. Halbrugge M. Walter U. Eur. J. Biochem. 1992; 205: 471-481Crossref PubMed Scopus (146) Google Scholar, 7Keularts I.M.L.W. Van Gorp R.M.A. Feijge M.A.H. Vuist W.M.J. Heemskerk J.W.M. J. Biol. Chem. 2000; 275: 1763-1772Abstract Full Text Full Text PDF PubMed Scopus (85) Google Scholar, 19Zahavi M. Zahavi J. Kakkar V.V. Thromb. Haemostasis. 1984; 52: 205-209Crossref PubMed Scopus (14) Google Scholar). Deletion of Gi2 or Gz caused this value to increase, whereas deletion of the only known PGI2 receptor in platelets, IP, caused it to fall. The magnitude of the changes was sufficient to affect responses to thrombin when imposed acutelyin vitro (7Keularts I.M.L.W. Van Gorp R.M.A. Feijge M.A.H. Vuist W.M.J. Heemskerk J.W.M. J. Biol. Chem. 2000; 275: 1763-1772Abstract Full Text Full Text PDF PubMed Scopus (85) Google Scholar). Nonetheless, in the present study such changes in cAMP levels appeared to have no effect on agonist responses ex vivo or at least no effect that could not be attributed to the loss of a preferred Gi family member when ADP or epinephrine responses were measured. If nothing else, this fact suggests that the basal cAMP concentration in circulating platelets reflects ongoing signaling through PGI2 receptors and through Gi2 and Gz as well as the balance that is maintained between ongoing cAMP synthesis and hydrolysis. This finding may also suggest that small, chronic changes in basal cAMP concentration have less of an impact than acute changes of similar magnitude, although this particular point was not addressed directly. A third conclusion from the present observations regards the different roles that are played by Gi family members during platelet activation. Recent studies, including those by Kunapuli and co-workers (31Jin J.G. Kunapuli S.P. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 8070-8074Crossref PubMed Scopus (481) Google Scholar, 32Kim S. Foster C. Lecchi A. Quinton T.M. Prosser D.M. Jin J.G. Cattaneo M. Kunapuli S.P. Blood. 2002; 99: 3629-3636Crossref PubMed Scopus (155) Google Scholar), have shown that optimal platelet activation requires signaling through Gi-coupled receptors as well as those that couple to Gq and G12 family members. The present study shows that the requirement for Gi-dependent signaling cannot be solely the result of the ability of these proteins to inhibit adenylyl cyclase. First, in the absence of PGI2 neither of the prototypical Gi-coupled agonists (ADP and epinephrine) had a measurable effect on cAMP levels and second, the defect in Gzα(−/−) platelets could not be reversed by the simple addition of a membrane-permeable inhibitor of adenylyl cyclase, even though that inhibitor was found to suppress PGI2-stimulated cAMP formation as effectively as ADP or epinephrine. Therefore, although Gi family members can clearly inhibit PGI2-stimulated cAMP formation in platelets, their ability to regulate other effectors presumably accounts for their essential role in promotion of platelet activation when PGI2 is absent. This conclusion is consistent with observations on human platelets with the use of receptor antagonists and adenylyl cyclase inhibitors (33Daniel J.L. Dangelmaier C. Jin J.G. Kim Y.B. Kunapuli S.P. Thromb. Haemostasis. 1999; 82: 1322-1326Crossref PubMed Scopus (124) Google Scholar). The full range of effectors for Gi family members other than adenylyl cyclase is still being explored, but recent studies show that at least one of those pathways involves phosphatidylinositol 3-kinase γ and the Ras family member, Rap1B, possibly via Gi-derived Gβγ (34Hirsch E. Bosco O. Tropel P. Laffargue M. Calvez R. Altruda F. Wymann M.P. Montrucchio G. FASEB J. 2001; 15: NIL307-NIL326Crossref Scopus (207) Google Scholar, 35Woulfe D. Jiang H. Mortensen R. Yang J. Brass L.F. J. Biol. Chem. 2002; 277: 23382-23390Abstract Full Text Full Text PDF PubMed Scopus (175) Google Scholar). The particular relevance of this pathway is suggested by the recent demonstration that Rap1B supports αIIbβ3activation (36Bertoni A. Tadokoro S. Eto K. Pampori N. Parisi L.V. White G.C. Shattil S.J. J. Biol. Chem. 2002; 277: 25715-25721Abstract Full Text Full Text PDF PubMed Scopus (164) Google Scholar). Similar to the inhibition of adenylyl cyclase, Rap1B activation can be triggered in platelets by ADP (via Gi2) and epinephrine (via Gz) (35Woulfe D. Jiang H. Mortensen R. Yang J. Brass L.F. J. Biol. Chem. 2002; 277: 23382-23390Abstract Full Text Full Text PDF PubMed Scopus (175) Google Scholar). The fact that more than one Gi family member can do so suggests again that the specificity in the roles of the Gi family members in platelets lies at the level of receptor coupling and not at the level of effector interactions. Finally, the results of the present study also have implications for the role of PGI2 in the regulation of platelet responsiveness. In theory the tendency toward injury-induced thrombosis in IP(−/−) mice could be the result of either the observed fall in basal cAMP levels or of a failure to respond to local accumulations of PGI2. Our inability to detect enhanced sensitivity of IP(−/−) platelets to low concentrations of agonists calls into question the impact of the decrease in basal cAMP. PGI2 is synthesized in endothelial cells and released in response to endothelial cell agonists, including thrombin (37Weksler B.B. Marcus A.J. Jaffe E.A. Proc. Natl. Acad. Sci. U. S. A. 1977; 74: 3922-3926Crossref PubMed Scopus (486) Google Scholar, 38Weksler B.B. Ley C.W. Jaffe E.A. J. Clin. Invest. 1978; 62: 923-930Crossref PubMed Scopus (504) Google Scholar). The fact that thrombin is produced locally in response to vascular injury suggests that part of the response to injury is to generate an inhibitor of platelet activation (PGI2) whose effects on platelet function must then be overcome to permit the formation of a platelet plug. Although this process seems somewhat cumbersome, it is likely that the system exists to place a threshold on platelet responsiveness when the extent of injury does not warrant extensive platelet activation. Only when the impetus for platelet plug formation is sufficiently great does the ability of agonists to inhibit adenylyl cyclase via Gi family members become essential. It also suggests that the increased tendency toward thrombosis in IP(−/−) mice after arterial injury is not caused by a generalized increase of agonist responsiveness in the circulating platelet population but rather by a loss of the regulatory effects of PGI2, specifically at the site of injury.