Search

Pleckstrin is a modular platelet protein consisting of N- and C-terminal pleckstrin homology (PH) domains, a central dishevelled egl10 and pleckstrin (DEP) domain and a phosphorylation region. Following agonist-induced platelet stimulation, dimeric pleckstrin translocates to the plasma membrane, is phosphorylated and then monomerizes. A recent study found that pleckstrin null platelets from a knockout mouse have a defect in granule secretion, actin polymerization and aggregation. However, the mechanism of pleckstrin signaling for this function is unknown. Our recent studies have led to the identification of a novel pleckstrin-binding protein, serum deprivation response protein (SDPR), by co-immunoprecipitation, GST-pulldowns and nanospray quadruple time of flight mass spectrometry. We show that this interaction occurs directly through N-terminal sequences of pleckstrin. Both pleckstrin and SDPR are phosphorylated by protein kinase C (PKC), but the interaction between pleckstrin and SDPR was shown to be independent of PKC inhibition or activation. These results suggest that SDPR may facilitate the translocation of nonphosphorylated pleckstrin to the plasma membrane in conjunction with phosphoinositides that bind to the C-terminal PH domain. After binding of pleckstrin to the plasma membrane, its phosphorylation by PKC exerts downstream effects on platelet aggregation/secretion.

Pleckstrin (plek)-null platelets from a knockout mouse have been shown to be defective in granule secretion, aggregation and actin polymerization. However, the mechanism of plek signaling is currently unknown. Therefore, we sought to identify plek-binding proteins in platelets by using GST pulldown assays and immunoprecipitation to isolate proteins from extracts of protein kinase C-activated or inhibited human platelets. Co-purified plek-binding proteins were resolved by SDS-PAGE and identified via nanospray quadruple TOF MS. Identified proteins may be involved in various cellular processes including cytoskeletal reorganization (moesin, radixin and alpha-actinin) and signal transduction (serum deprivation response protein, 17 beta-hydroxysteroid dehydrogenase 4 and factor XIIIA). Both platelet aggregation and/or secretion require actin polymerization. However, studies have shown no direct association between plek and actin. Based on our findings we propose indirect associations between plek and actin through 17 beta-hydroxysteroid dehydrogenase 4, alpha-actinin, moesin, radixin and factor XIIIA, which in turn suggest new roles for plek in platelet biology.

The mutant gunmetal mouse exhibits reduced rates of platelet synthesis, abnormalities of platelet alpha and dense granules and hypopigmentation. Several of these features resemble those of human alpha/delta platelet storage pool disease, grey platelet syndrome and Hermansky-Pudlak syndrome. Gunmetal mice have reduced levels of Rab geranylgeranyltransferase (RabGGTase), which adds lipophilic prenyl groups to the carboxyl terminus of Rab proteins. The degree of prenylation and the subcellular distribution of several Rab proteins were evaluated in mutant platelets, melanocytes and other tissues. Significant deficits in prenylation and membrane binding of most Rabs were observed in platelets and melanocytes. In contrast, minimal alterations in Rab prenylation were apparent in several other gunmetal tissues despite the fact that RabGGTase activity was equally diminished in these tissues. The mutant tissue-specific effects are probably due to increased concentrations of Rab proteins in platelets and melanocytes. These experiments show that Rab proteins are differentially sensitive to levels of RabGGTase activity and that normal platelet synthesis, platelet organelle function and normal pigmentation are highly sensitive to the degree of prenylation and membrane association of Rab proteins. Further, the tissue-specific effects of the gunmetal mutation suggest that RabGGTase is a potential target for therapy of thrombocytosis.

Pleckstrin is the major substrate of protein kinase C (PKC) in platelets. We sought to determine whether pleckstrin phosphorylation is sufficient to target the soluble protein to binding sites. Permeabilization of platelets by streptolysin O (SLO) was used to separate bound and soluble pleckstrin. Platelets were incubated with phorbol 12-myristate 13-acetate (PMA) and/or guanosine 5'-[gamma-thio]triphosphate (GTP[S]) in the presence of [gamma-(32)P]ATP and SLO. PMA stimulated pleckstrin phosphorylation, but this pleckstrin diffused from permeabilized platelets. Addition of GTP[S] with PMA caused up to 40-50% of pleckstrin to be retained within platelets and enhanced secretion of platelet 5-hydroxytryptamine. PKC alpha pseudosubstrate peptide inhibited pleckstrin phosphorylation, the binding of pleckstrin and secretion. After extraction of permeabilized platelets containing bound pleckstrin with Triton X-100, the protein was solubilized. Thus, phosphorylated pleckstrin was retained in platelets only after activation of GTP-binding proteins that stimulate the formation of membrane-bound pleckstrin ligands. Translocation of pleckstrin may facilitate the associated secretion.

GTP-binding proteins of the Rab family were cloned from human platelets using RT-PCR. Clones corresponding to two novel Rab proteins, Rab31 and Rab32, and to Rab11A, which had not been detected in platelets previously, were isolated. The coding sequence of Rab31 (GenBank accession no. U59877) corresponded to a 194 amino-acid protein of 21.6 kDa. The Rab32 sequence was extended to 1000 nucleotides including 630 nucleotides of coding sequence (GenBank accession no. U59878) but the 5' coding sequence was only completed later by others (GenBank accession no. U71127). Human Rab32 cDNA encodes a 225 amino-acid protein of 25.0 kDa with the unusual GTP-binding sequence DIAGQE in place of DTAGQE. Northern blots for Rab31 and Rab32 identified 4.4 kb and 1.35 kb mRNA species, respectively, in some human tissues and in human erythroleukemia (HEL) cells. Rabbit polyclonal anti-peptide antibodies to Rab31, Rab32 and Rab11A detected platelet proteins of 22 kDa, 28 kDa and 26 kDa, respectively. Human platelets were highly enriched in Rab11A (0.85 microg x mg of platelet protein(-1)) and contained substantial amounts of Rab32 (0.11 microg x mg protein(-1)). Little Rab31 was present (0.005 microg x mg protein(-1)). All three Rab proteins were found in both granule and membrane fractions from platelets. In rat platelets, the 28-kDa Rab32 was replaced by a 52-kDa immunoreactive protein. Rab31 and Rab32, expressed as glutathione S-transferase (GST)-fusion proteins, did not bind [alpha-(32)P]GTP on nitrocellulose blots but did bind [(35)S]GTP[S] in a Mg(2+)-dependent manner. Binding of [(35)S]GTP[S] was optimal with 5 microm Mg(2+)(free) and was markedly inhibited by higher Mg(2+) concentrations in the case of GST-Rab31 but not GST-Rab32. Both proteins displayed low steady-state GTPase activities, which were not inhibited by mutations (Rab31(Q64L) and Rab32(Q85L)) that abolish the GTPase activities of most low-M(r) GTP-binding proteins.

Hermansky Pudlak syndrome (HPS) is an autosomal recessive inherited disorder characterized by defects in synthesis and/or secretion of three related subcellular organelles: melanosomes, platelet-dense granules, and lysosomes. In the mouse, mutant forms of any of 14 separate genes result in an HPS-like phenotype. The mouse pearl and mocha genes encode subunits of the AP3 adaptor protein complex, confirming that HPS mutations involve proteins regulating intracellular vesicular trafficking. Therefore, expression of several additional proteins involved in vesicular transport was examined by immunoblotting of platelet extracts from HPS mutant and control mice. Platelet levels of SCAMPS (secretory carrier membrane proteins), Rab11, Rab31, NSF (N-ethylmaleimide-sensitive fusion protein), syntaxin 2, syntaxin 4, munc18c, and p115/TAP (p115/transcytosis-associated protein) were not significantly altered in several different HPS mutants. However, gunmetal (gm/gm) platelets contained decreased amounts of SNAP-23. The Snap23 gene was mapped to mouse chromosome 5, demonstrating it cannot encode the gm gene, which maps to chromosome 14. It is likely therefore that the gm gene functions upstream of SNAP-23 in vesicular trafficking., (Copyright 1999 Academic Press.)

Human platelets containing dense granules labelled with 5-hydroxy[14C]tryptamine ([14C]5-HT) were permeabilized by exposure to streptolysin O (SLO) in the presence of 4 mM [gamma-32P]ATP. Addition of either 100 nM phorbol 12-myristate 13-acetate (PMA) or of Ca2+ (pCa 5) at the same time as SLO induced secretion of dense-granule [14C]5-HT and the phosphorylation of pleckstrin by protein kinase C (PKC). Ca2+ also induced phosphorylation of myosin P-light chains. Guanosine 5'-[gamma-thio]triphosphate (GTP[S], 100 microM) did not stimulate secretion from SLO-permeabilized platelets in the absence of Ca2+ (pCa>9), but greatly potentiated secretion in the presence of low PMA (10 nM) or low Ca2+ (pCa 6). However, GTP[S] did stimulate myosin P-light-chain phosphorylation in the absence of Ca2+, an effect that was associated with morphological changes, including granule centralization. Inhibition of PKC and of pleckstrin phosphorylation by Ro 31-8220 blocked secretion induced by PMA or by GTP[S] and PMA in the absence of Ca2+, but did not prevent the GTP[S]-induced phosphorylation of myosin P-light chains or secretion induced by Ca2+ at pCa 5. When the time period between exposure of platelets to SLO and challenge at pCa>9 with PMA or with GTP[S] and PMA was increased, there were rapid and parallel decreases in the secretion and pleckstrin phosphorylation responses, which were lost after 3-5 min. In contrast, the responsiveness of secretion to Ca2+ (pCa 5) or to GTP[S] and Ca2+ (pCa 6) persisted for at least 10 min after exposure of platelets to SLO, although the ability of pleckstrin to undergo phosphorylation was still lost after 3-5 min. Both PKC and pleckstrin were undetectable within platelets after 5 min exposure to SLO. The results suggest that the loss of responsiveness to PMA or to GTP[S] and PMA is attributable to the leakage of PKC (and possibly pleckstrin) from the platelets, whereas secretion stimulated by Ca2+ or by GTP[S] and Ca2+ utilizes membrane-associated Ca2+- and GTP-binding proteins and occurs independently of PKC activation.

cGMP enhances cAMP accumulation in platelets via cGMP-inhibited phosphodiesterase (PDE3) [Maurice and Haslam (1990) Mol. Pharmacol. 37, 671-681]. However, cGMP might also limit cAMP accumulation by activating cGMP-stimulated phosphodiesterase (PDE2). We therefore evaluated the role of PDE2 in human platelets by using erythro-9-(2-hydroxy-3-nonyl)adenine (EHNA) to inhibit this enzyme selectively. IC50 values for the inhibition of platelet PDE2 by EHNA, with 10 microM cAMP as substrate in the absence and in the presence of 1 microM cGMP, were 15 and 3 microM respectively. Changes in platelet cyclic [3H]nucleotides were measured after prelabelling with [3H]adenine and [3H]guanine. Nitroprusside (NP) caused concentration-dependent increases in [3H]cGMP and a biphasic increase in [3H]cAMP, which was maximal at 10 microM (49+/-6%) and smaller at 100 microM (32+/-6%) (means+/-S.E.). In the presence of EHNA (20 microM), which had no effects alone, NP caused much larger increases in platelet [3H]cAMP (125+/-14% at 100 microM). EHNA also enhanced [3H]cGMP accumulation at high NP concentrations. In accord with these results, EHNA markedly potentiated the inhibition of thrombin-induced platelet aggregation by NP. The roles of cAMP and cGMP in this effect were investigated by using 2', 5'-dideoxyadenosine to inhibit adenylate cyclase. This compound decreased the accumulation of [3H]cAMP but not that of [3H]cGMP, and diminished the inhibition of platelet aggregation by NP with EHNA. We conclude that much of the effect of NP with EHNA is mediated by cAMP. Lixazinone (1 microM), a selective inhibitor of PDE3, increased platelet [3H]cAMP by 177+/-15%. This increase in [3H]cAMP was markedly inhibited by NP; EHNA blocked this effect of NP. Parallel studies showed that NP suppressed the inhibition of platelet aggregation by lixazinone. EHNA enhanced the large increases in [3H]cAMP seen with 20 nM prostacyclin (PGI2), but had no effect with 1 nM PGI2. NP and 1 nM PGI2 acted synergistically to increase [3H]cAMP, an effect attributable to the inhibition of PDE3 by cGMP; EHNA greatly potentiated this synergism. In contrast, NP decreased the [3H]cAMP accumulation seen with 20 nM PGI2, an effect that was blocked by EHNA. The results show that, provided that cGMP is present, PDE2 plays a major role in the hydrolysis of low cAMP concentrations and restricts any increases in cAMP concentration and decreases in platelet aggregation caused by the inhibition of PDE3. At high cAMP, PDE2 plays the major role in cAMP breakdown, whether cGMP is present or not.

We have used reverse transcriptase PCR, platelet mRNA and degenerate primers based on platelet peptide sequences, to amplify a fragment of platelet cGMP-inhibited phosphodiesterase (cGI-PDE; PDE3). Sequence analysis of this clone established that both the platelet and the cardiac forms of PDE3 were derived from the same gene (PDE3A). A RT-PCR product representing the C-terminal half of platelet PDE3 cDNA and corresponding to amino acid residues 560-1141 of the cardiac enzyme, was cloned and expressed in Escherichia coli cGI-PDEDelta1. Further deletion mutants were constructed by removing either an additional 100 amino acids from the N-terminus (cGI-PDEDelta2) or the 44-amino-acid insert characteristic of the PDE3 family, from the catalytic domain (cGI-PDEDelta1Deltai). In addition, site-directed mutagenesis was performed to explore the function of the 44-amino-acid insert. All mutants were evaluated for their ability to hydrolyse cAMP and cGMP, their ability to be photolabelled by [32P]cGMP and for the effects of PDE3 inhibitors. The Km values for hydrolysis of cAMP and cGMP by immunoprecipitates of cGI-PDEDelta1 (182+/-12 nM and 153+/-12 nM respectively) and cGI-PDEDelta2 (131+/-17 nM and 99+/-1 nM respectively) were significantly lower than those for immunoprecipitates of intact platelet PDE3 (398+/-50 nM and 252+/-16 nM respectively). Moreover, N-terminal truncations of platelet enzyme increased the ratio of Vmax for cGMP/Vmax for cAMP from 0.16+/-0.01 in intact platelet enzyme, to 0.37+/-0.05 in cGI-PDEDelta1 and to 0.49+/-0.04 in cGI-PDEDelta2. Thus deletion of the N-terminus enhanced hydrolysis of cGMP relative to cAMP, suggesting that N-terminal sequences may exert selective effects on enzyme activity. Removal of the 44-amino-acid insert generated a mutant with a catalytic domain closely resembling those of other PDE gene families but despite a limited ability to be photolabelled by [32P]cGMP, no cyclic nucleotide hydrolytic activities of the mutant were detectable. Mutation of amino acid residues in putative beta-turns at the beginning and end of the 44-amino-acid insert to alanine residues markedly reduced the ability of the enzyme to hydrolyse cyclic nucleotides. The PDE3 inhibitor, lixazinone, retained the ability to inhibit cAMP hydrolysis and [32P]cGMP binding by the N-terminal deletion mutants and the site-directed mutants, suggesting that PDE3 inhibitors may interact exclusively with the catalytic domain of the enzyme.

The major substrate of protein kinase C(PKC) in platelets is the 40 kDa protein, pleckstrin. Addition of the homobifunctional reagent, bis(sulphosuccinimidyl)suberate (BS3), to platelet lysate, cytosol fraction or to electropermeabilized platelets resulted in cross-linking of pleckstrin to give higher-molecular-mass complexes of 68 kDa, 90 kDa and 100-120 kDa respectively, which were visualized by immunoblotting with an anti-pleckstrin antibody. Higher levels of cross-linking were observed in permeabilized platelets than in platelet lysates. The yields of the cross-linked complexes were much reduced after dilution of platelet lysate or lysis of electropermeabilized platelets and, in the case of the 90 kDa and 100-120 kDa species, after activation of PKC by phorbol 12-myristate 13-acetate. Similar experiments with purified pleckstrin indicated that the 90 kDa and 100-120 kDa species consist, at least in part, of pleckstrin dimers and higher oligomers. After incubation of purified pleckstrin (0.45 mg/ml) for 1 h with 2 mM BS3, about 25% of the protein was present in cross-linked species. The results indicate that pleckstrin undergoes a reversible self-association that can be prevented by phosphorylation of the protein, and also interacts with an unidentified platelet protein of about 28 kDa.

In human platelets, thrombin not only stimulates the phosphorylation of pleckstrin (P47) and of myosin P-light chains, but also induces the dephosphorylation of an 18-19 kDa phosphoprotein (P18) [Imaoka, Lynham and Haslam (1983) J. Biol. Chem. 258, 11404-11414]. We have now studied this protein in detail. The thrombin-induced dephosphorylation reaction did not begin until the phosphorylation of myosin P-light chains and the secretion of dense-granule 5-hydroxytryptamine were nearly complete, but did parallel the later stages of platelet aggregation. Experiments with ionophore A23187 and phorbol 12-myristate 13-acetate indicated that dephosphorylation of P18 was stimulated by Ca2+, but not by protein kinase C. Two-dimensional analysis of platelet proteins, using non-equilibrium pH gradient electrophoresis followed by SDS/PAGE, showed that thrombin decreased the amount of phosphorylated P18 in platelets by up to 70% and slightly increased the amount of a more basic unlabelled protein that was present in 3-fold excess of P18 in unstimulated platelets. These two proteins were identified as the phosphorylated and non-phosphorylated forms of the pH-sensitive actin-depolymerizing protein, cofilin, by sequencing of peptide fragments and immunoblotting with a monoclonal antibody specific for cofilin. The molar concentration of cofilin in platelets was approx. 10% that of actin. Platelet cofilin was phosphorylated exclusively on serine. Experiments with electropermeabilized platelets showed that dephosphorylation of cofilin could be stimulated by guanosine 5'-[gamma-thio]triphosphate (GTP[S]) in the absence of Ca2+ or by a free Ca2+ concentration of 10 microM. This GTP[S]-induced dephosphorylation reaction was inhibited by 1-naphthyl phosphate, but not by okadaic acid. Our results add cofilin to the actin-binding proteins that may regulate the platelet cytoskeleton, and suggest that platelet cofilin can be activated by dephosphorylation reactions initiated either by a GTP-binding protein or Ca2+.

Ultraviolet irradiation of human platelet cytosol in the presence of 32P-labelled cyclic GMP (cGMP) can specifically label 110, 80, 55, 49 and 38 kDa proteins; the 110 kDa species is the subunit of cGMP-inhibited phosphodiesterase (PDE III) and the 80 kDa species that of cGMP-dependent protein kinase (Tang et al., 1993, Biochem. J. 294, 329). We have now shown that although photolabelling of platelet PDE III was inhibited by unlabelled cGMP, 8-bromo-cGMP and cyclic AMP (cAMP), it was not affected by phosphorothioate analogues of these cyclic nucleotides. Specific concentration-dependent inhibitions of the photolabelling of PDE III were observed with the following PDE inhibitors: trequinsin (IC50 = 13 +/- 2 nM), lixazinone (IC50 = 22 +/- 4 nM), milrinone (IC50 = 56 +/- 12 nM), cilostamide (IC50 = 70 +/- 9 nM), siguazodan (IC50 = 117 +/- 29 nM) and 3-isobutyl 1-methylxanthine (IBMX) (IC50 = 3950 +/- 22 nM). Thus, measurements of the inhibitory effects of compounds on the photolabelling of platelet PDE III provide a simple quantitative means of investigating their actions at a molecular level that avoids the need to purify the enzyme. Photolabelling of rat platelet lysate or rat heart homogenate by [32P]cGMP showed that the 110 kDa PDE III present in human material was replaced by a 115 kDa protein, labelling of which was also blocked by PDE III inhibitors. Heart and other rat tissues contained much less of this putative 115 kDa PDE III than rat platelets. In contrast, the 80 kDa protein was labelled much less in platelets than in many other rat tissue homogenates (e.g., heart, aorta, uterus and lung). Thus, comparison of the relative amounts of specific photolabelled proteins in different cells may provide an indication of different patterns of cyclic nucleotide action. We compared the abilities of phosphodiesterase inhibitors to block the photolabelling of PDE III in human platelet cytosol and to increase the iloprost-stimulated accumulation of cAMP in intact platelets. Whereas trequinsin (EC50 = 19 +/- 3 nM), lixazinone (EC50 = 122 +/- 8 nM), milrinone (EC50 = 5320 +/- 970 nM) and siguazodan (EC50 = 18880 +/- 3110 nM) all increased platelet cAMP to the same maximum extent, cilostamide and IBMX increased cAMP further, indicating that they inhibited a PDE isozyme in addition to PDE III.

Previous work has shown that guanosine 5'-[gamma-thio]triphosphate (GTP[S]) and GTP stimulate phospholipase D (PLD) in rabbit platelet membranes and that these effects are greatly enhanced by pretreatment of platelets with phorbol esters that activate protein kinase C [Van der Meulen and Haslam (1990), Biochem. J. 271, 693-700]. In the present study, the effects of Mg2+, various nucleoside triphosphates and phosphocreatine (PCr) were investigated. Platelet membranes containing phospholipids labelled with [3H]glycerol were assayed for PLD in the presence of an optimal Mg2+ concentration (10 mM) by measuring [3H]phosphatidylethanol formation in incubations that included 300 mM ethanol. In membranes from phorbolester-treated platelets, the same maximal increases in PLD activity (5-fold) were seen with 1 microM GTP[S]), and 100 microM GTP. Addition of adenosine 5'-[gamma-thio]triphosphate (ATP[S]), ITP, XTP, UTP and CTP had similar stimulatory effects, but only at > or = 1 mM. In contrast, ATP had a biphasic action, causing a maximal (2-fold) stimulation at 10 microM and smaller effects at higher concentrations; the inhibitory component of the action of ATP was blocked by 2 microM staurosporine. Guanosine 5'-[beta-thio]diphosphate decreased the stimulatory effects of ATP and ATP[S]. UDP, which can inhibit nucleoside diphosphate kinase (NDPK), decreased the activation of PLD by ATP[S], ATP, XTP, CTP and to a lesser extent ITP, but had no effect on the actions of GTP[S] and GTP. Rabbit platelet membranes contained NDPK and addition of [gamma-32P]ATP led to the formation of [32P]GTP in amounts sufficient to explain most or all of the activation of PLD; UDP prevented GTP formation. PCr (0.04-1 mM) also stimulated membrane PLD activity, an effect that was dependent on endogenous membrane-bound creatine kinase (CK). UDP and guanosine 5'-[beta-thio]diphosphate each inhibited this effect of PCr. The results show that in rabbit platelet membranes, CK, NDPK and the GTP-binding protein that activates PLD can be functionally coupled. However, assay of membrane preparations at increasing dilutions showed that stimulation of PLD by the compounds studied, with the partial exception of ATP[S], involved diffusible rather than protein-bound intermediates.

The photoaffinity labelling of platelet cyclic GMP (cGMP)-binding proteins by [32P]cGMP was studied; at least five labelled proteins (110, 80, 55, 49 and 38 kDa) were detected in platelet cytosol and four (80, 65, 49 and 38 kDa) in platelet membranes. The 110 kDa species was identified as cGMP-inhibited cyclic AMP (cAMP) phosphodiesterase (PDE III) by immunoprecipitation and by the inhibition of photolabelling by specific inhibitors of this enzyme. Similarly, the 80 kDa species was identified as cGMP-dependent protein kinase by immunoprecipitation and by the effects of cGMP analogues on photolabelling. Addition of cAMP greatly enhanced the labelling of this 80 kDa protein, implying the existence of a potentially important interaction between the effects of cGMP and cAMP. The 65 kDa photolabelled protein appears to be a novel platelet cyclic-nucleotide-binding protein. In contrast, the 49 and 55 kDa photolabelled species are probably the RI and RII regulatory subunits of cAMP-dependent protein kinase, and the 38 kDa protein(s) may be proteolytic fragment(s) of RI and/or RII.

We have tested the hypothesis that phospholipase D (PLD) is the effector of the unidentified G protein (GE) mediating Ca(2+)-independent exocytosis in platelets. Although GTP gamma S, and to a lesser extent phorbol 12-myristate 13-acetate (PMA), caused some secretion of 5-HT from electropermeabilized human platelets in the effective absence of Ca2+ (pCa > 9), these stimuli had much more potent synergistic effects when added together. In all cases, secretion of 5-HT was closely correlated to the stimulus-induced formation of [3H]phosphatidic acid ([3H]PA) from [3H]arachidonate-labelled phospholipids. Addition of ethanol inhibited both secretion and [3H]PA formation and led to the accumulation of [3H]phosphatidylethanol ([3H]PEt), indicating that [3H]PA was formed largely by activation of PLD. BAPTA and analogues caused dose-dependent inhibitions of both GTP gamma S-induced secretion and PLD activity in the permeabilized platelets. This action of BAPTA did not appear to be mediated by chelation of Ca2+ or by direct inhibition of protein kinase C (PKC). The results suggest that PLD is the target of GE in platelets and that BAPTA can block PLD activation.

Studies on electropermeabilized human platelets indicated that any two of three distinct factors must be present for marked secretion of dense or alpha-granule constituents to occur. These factors are Ca2+, activation of protein kinase C (PKC) and activation of an unidentified GTP-binding protein ('GE'). Thus, in the absence of Ca2+, phorbol ester and GTP[S] acted synergistically to promote secretion, whereas in the presence of Ca2+, either activation of PKC or addition of GTP[S] was sufficient. In all cases, secretion correlated with the activation of phospholipase D (PLD), as detected by the formation of [3H]phosphatidic acid (PA) in the absence of ethanol or of [3H]phosphatidylethanol (PEt) in the presence of ethanol. Secretion did not correlate with phospholipase C (PLC) activity or with the accumulation of 1,2-diacylglycerol (DAG), both of which required Ca2+ and were inhibited by phorbol ester. Ethanol partially inhibited secretion in the absence of Ca2+. BAPTA, a known inhibitor of Ca(2+)-independent secretion in permeabilized cells, caused parallel inhibitions of secretion and PLD activity. GTP[S] enhanced PKC activity, as indicated by pleckstrin phosphorylation, apparently by stimulating the formation of PA in the absence of Ca2+, as well as of DAG in the presence of Ca2+. PA and stable analogues, including PEt, stimulated the Ca(2+)-independent phosphorylation of pleckstrin and other proteins in platelet supernatant fraction. The results suggest that PA formed by activation of PLD may mediate secretion from permeabilized platelets by PKC-dependent and independent mechanisms. However, in intact platelets stimulated by thrombin, PLD accounted for only 10-20% of the total PA formed and can only play a major role in secretion if this PA fraction is distinct from that formed by the combined actions of PLC and DAG kinase.

Nitrovasodilators increase both cyclic GMP and cyclic AMP in isolated platelets (Maurice DH, Haslam RJ. Mol Pharmacol 1990;37:671-81). To determine whether this occurs in blood, platelet cyclic[3H]GMP and cyclic [3H]AMP were measured in prelabeled rabbit platelets resuspended in modified Tyrode's solution or citrated blood. In the former medium, increases in cyclic [3H]nucleotides in response to nitroprusside (NP) and 3-morpholinosydnonimine (SIN-1) were maximal by 1 min; in blood, maximal increases were observed only after 10 min and were much smaller. In blood, SIN-1 was more effective than the same concentration of NP. After 10 min, 100 microM SIN-1 increased platelet cyclic[3H )GMP by 475 +/- 58% and cyclic[3H]AMP by 29 +/- 7% (means +/- SEM, 18 experiments). Supraadditive increases in platelet cyclic [3H]AMP in blood were observed when SIN-1 was combined with prostaglandin E1 (PGE1). Thus, after 10 min, SIN-1 (100 microM), PGE1 (20 nM), and SIN-1 + PGE1 increased cyclic[3H]AMP by 25 +/- 7, 35 +/- 6, and 130 +/- 17%, respectively (four experiments). In the same experiments, release of platelet [14C]serotonin by platelet-activating factor (PAF) was inhibited by 22 +/- 5, 2 +/- 2, and 61 +/- 5%, respectively. Increases in platelet cyclic[3H]GMP with SIN-1 were unaffected by PGE1. These results suggest that although cyclic GMP may mediate the effects of SIN-1 alone on platelet function, cyclic AMP mediates the synergistic action of SIN-1 and PGE1. M&B 22,948 (a selective cyclic GMP phosphodiesterase inhibitor) enhanced the increases in platelet cyclic[3H]GMP and cyclic[3H]AMP caused by SIN-1 and also increased the associated inhibition of [14C]serotonin release. M&B 22,948 also augmented the synergistic increases in cyclic[3H]AMP and inhibition of platelet function caused by SIN-1 + PGE1. The results show that a selected nitrovasodilator (e.g., SIN-1), a prostaglandin and a cyclic GMP phosphodiesterase inhibitor can exert synergistic effects on platelets in blood. This may be relevant to the pharmacologic management of thromboembolic disease.

Electropermeabilized human platelets containing 5-hydroxy[14C]tryptamine ([14C]5-HT) were suspended in a glutamate medium containing ATP and incubated for 10 min with (in various combinations) Ca2+ buffers, phorbol 12-myristate 13-acetate (PMA), guanine nucleotides, and thrombin. Release of [14C]5-HT and beta-thromboglobulin (beta TG) were used to measure secretion from dense and alpha-granules, respectively. Ca2+ alone induced secretion from both granule types; half-maximal effects were seen at a -log [Ca2+ free] (pCa) of 5.5 and maximal secretion at a pCa of 4.5, when approximately 80% of 5-HT and approximately 50% of beta TG were released. Addition of PMA, guanosine 5'-O-(3-thiotriphosphate) (GTP gamma S), GTP, or thrombin shifted the Ca2+ dose-response curves for secretion of both 5-HT and beta TG to the left and caused small increases in the maximum secretion observed. These results suggested that secretion from alpha-granules, like that from dense granules, is a Ca(2+)-dependent process stimulated by the sequential activation of a G-protein, phospholipase C, and protein kinase C (PKC). However, high concentrations of PMA and GTP gamma S had distinct effects in the absence of Ca2+ (pCa greater than 9); 100 nM PMA released approximately 20% of platelet 5-HT but little beta TG, whereas 100 microM GTP gamma S stimulated secretion of approximately 25% of each. Simultaneous addition of PMA greatly enhanced these effects of GTP gamma S. Phosphorylation of pleckstrin in permeabilized platelets incubated with [gamma-32P]ATP was used as an index of the activation of PKC during secretion. In the absence of Ca2+, 100 nM PMA caused maximal phosphorylation of pleckstrin and 100 microM GTP gamma S was approximately 50% as effective as PMA; neither GTP gamma S nor Ca2+ enhanced the phosphorylation of pleckstrin caused by 100 nM PMA. These results indicate that, although activation of PKC promoted secretion, GTP gamma S exerted additional stimulatory effects on secretion from both dense and alpha-granules that were not mediated by PKC. Measurement of [3H]inositol phosphate formation in permeabilized platelets containing [3H]phosphoinositides showed that GTP gamma S did not stimulate phosphoinositide-specific phospholipase C in the absence of Ca2+. It follows that in permeabilized platelets, GTP gamma S can both stimulate PKC and enhance secretion via G-protein-linked effectors other than this phospholipase.

Rabbit platelets were labelled with [3H]glycerol and incubated with or without phorbol 12-myristate 13-acetate (PMA). Membranes were then isolated and assayed for phospholipase D (PLD) activity by monitoring [3H]phosphatidylethanol formation in the presence of 300 mM-ethanol. At a [Ca2+free] of 1 microM, PLD activity was detected in control membranes, but was 5.4 +/- 0.8-fold (mean +/- S.E.M.) greater in membranes from PMA-treated platelets. Under the same conditions, 10 microM-guanosine 5'-[gamma-thio]triphosphate (GTP[S]) stimulated PLD by 18 +/- 3-fold in control membranes, whereas PMA treatment and GTP[S] interacted synergistically to increase PLD activity by 62 +/- 12-fold. GTP[S]-stimulated PLD activity was observed in the absence of Ca2+, but was increased by 1 microM-Ca2+ (3.5 +/- 0.2-fold and 1.8 +/- 0.1-fold in membranes from control and PMA-treated platelets respectively). GTP exerted effects almost as great as those of GTP[S], but 20-30-fold higher concentrations were required. Guanosine 5'-[beta-thio]diphosphate inhibited the effects of GTP[S] or GTP, suggesting a role for a GTP-binding protein in activation of PLD. Thrombin (2 units/ml) stimulated the PLD activity of platelet membranes only very weakly and in a GTP-independent manner. The actions of PMA and analogues on PLD activity correlated with their ability to stimulate protein kinase C in intact platelets. Staurosporine, a potent protein kinase inhibitor, had both inhibitory and, at higher concentrations, stimulatory effects on the activation of PLD by PMA. The results suggest that PMA not only stimulates PLD via activation of protein kinase C but can also activate the enzyme by a phosphorylation-independent mechanism in the presence of staurosporine. However, under physiological conditions, full activation of platelet PLD may require the interplay of protein kinase C, increased Ca2+ and a GTP-binding protein, and may occur as a secondary effect of the activation of phospholipase C.

We investigated the roles of cyclic GMP and cyclic AMP in the inhibition of rabbit platelet aggregation and degranulation by two nitrovasodilators, sodium nitroprusside (SNP) and 3-morpholinosydnonimine (SIN-1; the active metabolite of molsidomine), with particular reference to the synergistic interaction of these drugs with prostaglandin E1 (PGE1). Changes in platelet cyclic [3H]GMP and cyclic [3H]AMP were measured by rapid and sensitive prelabeling techniques, the validity of which were confirmed by radioimmunoassays. Incubation of the platelets with 0.1 to 10 microM SNP alone for 0.5 min caused progressively greater inhibitions of platelet function associated with large dose-dependent increases in cyclic [3H]GMP and 1.4- to 3.0-fold increases in cyclic [3H]AMP. However, addition of SNP with the adenylate cyclase activator, PGE1, at a concentration of the latter that had little effect alone, caused much larger increases in cyclic [3H]AMP and greatly enhanced the inhibition of platelet aggregation. SIN-1 had effects similar to those of SNP, although it was less active. The adenylate cyclase inhibitor 2',5'-dideoxyadenosine (DDA) diminished the increases in cyclic [3H]AMP caused by SNP or SIN-1 in both the presence and absence of PGE1 but reduced the inhibition of platelet function caused by the nitrovasodilators only in the presence of PGE1. These results suggest that, although cyclic GMP may mediate the inhibition of rabbit platelet function by high concentrations of nitrovasodilators added alone, the synergistic interaction of lower concentrations with PGE1 depends on an enhanced accumulation of cyclic AMP. Synergistic effects on cyclic [3H]AMP accumulation were also observed on incubation of platelets with SNP and adenosine, another activator of adenylate cyclase. Hemoglobin, which binds nitric oxide, blocked or reversed the increases in both cyclic [3H]GMP and cyclic [3H]AMP in platelets caused by the nitrovasodilators added either alone or with PGE1. Cilostamide, a selective inhibitor of platelet low Km cyclic AMP phosphodiesterase, had effects on platelet cyclic [3H]AMP accumulation identical to those of SNP, suggesting that the action of the latter depends on inhibition of the same enzyme. M&B 22,948, a selective inhibitor of cyclic GMP phosphodiesterase, potentiated the increases in both cyclic [3H]GMP and cyclic [3H]AMP caused by SNP. A hyperbolic relationship was found between the increases in cyclic [3H]GMP and cyclic [3H]AMP caused by different concentrations of SNP; this relationship was not affected by addition of M&B 22,948. The results strongly suggest that the increases in platelet cyclic [3H]AMP caused by nitrovasodilators in the presence or absence of activators of adenylate cyclase are mediated by the inhibition by cyclic GMP of cyclic AMP breakdown.(ABSTRACT TRUNCATED AT 400 WORDS)

Polyclonal antibodies raised against specific recombinant low molecular mass GTP-binding proteins were tested for their ability to recognize partially purified human platelet membrane Gn-proteins (i.e. proteins that bind [alpha-32P]GTP on nitrocellulose blots of SDS/polyacrylamide gels). An antiserum against simian ralA protein recognized a 27 kDa human platelet protein with the same apparent molecular mass as the major platelet Gn-protein (Gn27). In further analysis by two-dimensional polyacrylamide gel electrophoresis, the isoelectric focusing step permitted resolution of 12 major Gn-protein forms, seven of 27 kDa (Gn27a-g), one of 26 kDa (Gn26) and four of 24 kDa (Gn24a-d). The ralA antibody reacted strongly with the five most basic Gn27 species (a-e), weakly with Gn26 and not at all with Gn27f, Gn27g or Gn24a-d. We conclude that ral gene products account for some but probably not for all of the platelet Gn-proteins.

Incubation of human platelets with myo-[3H]inositol in a low-glucose Tyrode's solution containing MnCl2 enhanced the labelling of phosphoinositides about sevenfold and greatly facilitated the measurement of [3H]inositol phosphates formed by the activation of phospholipase C. Labelled platelets were permeabilized by high-voltage electric discharges and equilibrated at 0 degree C with ATP, Ca2+ buffers and guanine nucleotides, before incubation in the absence or presence of thrombin. Incubation of these platelets with ATP in the presence or absence of Ca2+ ions led to the conversion of [3H]phosphatidylinositol to [3H]phosphatidylinositol 4-phosphate and [3H]phosphatidylinositol 4,5-bisphosphate ([3H]PtdInsP2). At a pCa of 6, addition of 100 microM GTP[gamma S] both prevented this accumulation of [3H]PtdInsP2 and stimulated its breakdown; the formation of [3H]inositol phosphates was increased ninefold. After 5 min these comprised 70% [3H]inositol monophosphate ([3H]InsP), 28% [3H]inositol bisphosphate ([3H]InsP2) and 2% [3H]inositol trisphosphate ([3H]InsP3). In shorter incubations higher percentages of [3H]InsP2 and [3H]InsP3 were found. In the absence of added Ca2+, the formation of [3H]inositol phosphates was decreased by over 90%. Incubation of permeabilized platelets with GTP[gamma S] in the presence of 10 mM Li+ decreased the accumulation of [3H]InsP and increased that of [3H]InsP2, without affecting [3H]InsP3 levels. Addition of unlabelled InsP3 decreased the intracellular hydrolysis of exogenous [32P]InsP3 but did not trap additional [3H]InsP3. These results and the time course of [3H]inositol phosphate formation suggest that GTP[gamma S] stimulated the action of phospholipase C on a pool of [3H]phosphatidylinositol 4-phosphate that was otherwise converted to [3H]PtdInsP2 and that much less hydrolysis of [3H]phosphatidylinositol to [3H]InsP or of [3H]PtdInsP2 to [3H]InsP3 occurred. At a pCa of 6, addition of thrombin (2 units/ml) to permeabilized platelets caused small increases in the formation of [3H]InsP and [3H]InsP2. This action of thrombin was enhanced twofold by 10-100 microM GTP and much more potently by 4-40 microM GTP[gamma S]. In the presence of the latter, thrombin also increased [3H]InsP3. The total formation of [3H]inositol phosphates by permeabilized platelets incubated with thrombin and GTP[gamma S] was comparable with that observed on addition of thrombin alone to intact platelets. However, HPLC of the [3H]inositol phosphates formed indicated that about 75% of the [3H]InsP accumulating in permeabilized platelets was the 4-phosphate, whereas in intact platelets stimulated by thrombin, up to 80% was the 1-phosphate.(ABSTRACT TRUNCATED AT 400 WORDS)

Jakobs, Bauer & Watanabe [(1985) Eur. J. Biochem. 151, 425-430] reported that treatment of platelets with phorbol 12-myristate 13-acetate (PMA) prevented GTP- and agonist-induced inhibition of adenylate cyclase in membranes from the platelets. This was attributed to the phosphorylation of the inhibitory guanine nucleotide-binding protein (Gi) by protein kinase C. In the present study, the effects of PMA on cyclic [3H]AMP formation and protein phosphorylation were studied in intact human platelets labelled with [3H]adenine and [32P]Pi. Incubation mixtures contained indomethacin to block prostaglandin synthesis, phosphocreatine and creatine kinase to remove ADP released from the platelets, and 3-isobutyl-1-methylxanthine to inhibit cyclic AMP phosphodiesterases. Under these conditions, PMA partially inhibited the initial formation of cyclic [3H]AMP induced by prostaglandin E1 (PGE1), but later enhanced cyclic [3H]AMP accumulation by blocking the slow decrease in activation of adenylate cyclase that follows addition of PGE1. PMA had more marked and exclusively inhibitory effects on cyclic [3H]AMP formation induced by prostaglandin D2 and also inhibited the action of forskolin. Adrenaline, high thrombin concentrations and, in the absence of phosphocreatine and creatine kinase, ADP inhibited cyclic [3H]AMP formation induced by PGE1. The actions of adrenaline and thrombin were attenuated by PMA, but that of ADP was little affected, suggesting differences in the mechanisms by which these agonists inhibit adenylate cyclase. sn-1,2-Dioctanoylglycerol (diC8) had effects similar to those of PMA. The actions of increasing concentrations of PMA or diC8 on the modulation of cyclic [3H]AMP formation by PGE1 or adrenaline correlated with intracellular protein kinase C activity, as determined by 32P incorporation into the 47 kDa substrate of the enzyme. Parallel increases in phosphorylation of 20 kDa and 39-41 kDa proteins were also observed. Platelet-activating factor, [Arg8]vasopressin and low thrombin concentrations, all of which inhibit adenylate cyclase in isolated platelet membranes, did not affect cyclic [3H]AMP formation in intact platelets. However, the activation of protein kinase C by these agonists was insufficient to account for their failure to inhibit cyclic [3H]AMP formation. Moreover, high thrombin concentrations simultaneously activated protein kinase C and inhibited cyclic [3H]AMP formation. The results show that, in the intact platelet, the predominant effects of activation of protein kinase C on adenylate cyclase activity are inhibitory, suggesting actions additional to inactivation of Gi.

Methods were developed for measuring changes in platelet sensitivity to a release-inducing stimulus and in platelet cyclic AMP in fresh whole blood samples from rabbits. These techniques permitted detection of the effects of exogenous and endogenous prostacyclin on circulating platelets. In these methods, rabbit platelets were labeled in vitro by incubation with [14C]serotonin and [3H]adenine and then transfused into other rabbits. Release of platelet [14C]serotonin by a standard dose of synthetic platelet-activating factor (40 pmol/ml) and the platelet cyclic [3H]AMP levels were then measured in citrated blood from the conscious animals within 2 min of arterial puncture. Bolus intravenous injections of prostacyclin (1-10 nmol/kg) caused concentration-dependent increases in platelet cyclic AMP after 2 min, which decreased approximately 75% by 5 min, and disappeared after 30 min. Significant inhibition of the platelet release reaction was detected 2 min but not 5 min after injection of 10 nmol of prostacyclin per kilogram. With lower doses, significant enhancement of the release of [14C]serotonin was observed after 5 min. Similar changes in platelet responsiveness and cyclic [3H]AMP were observed after release of endogenous prostacyclin by intravenous injection of angiotensin II (5 nmol/kg); inhibition of the release of [14C]serotonin after 2 min was followed by potentiation after 5 min, though platelet cyclic [3H]AMP remained above control values. In these experiments, the time course of the changes in platelet cyclic [3H]AMP correlated closely with values for blood prostacyclin obtained previously (Haslam, R.J., and M.D. McClenaghan, 1981, Nature [Lond.]., 292:364-366). Prostacyclin also had a biphasic effect on the release of [14C]serotonin when added to citrated blood in vitro, though both the increase in sensitivity to platelet-activating factor and the return of platelet cyclic [3H]AMP towards control values took place more slowly. At all times, addition of platelet-activating factor decreased platelet cyclic [3H]AMP towards but not below the control level observed in the absence of prostacyclin. Our results indicate that although transient increases in platelet cyclic AMP cause an immediate decrease in platelet responsiveness in vivo or in vitro, a period of enhanced platelet sensitivity follows as platelet cyclic AMP falls.

We have investigated the effects of NaCl and GTP on the inhibition of platelet adenylate cyclase by 1-O-octadecyl-2-O-acetyl-sn-glyceryl-3-phosphorylcholine (1-octadecyl-2-acetyl-G-3-PC), using particulate fractions from human and rabbit platelets that had been frozen and thawed in the presence of ethylene glycol bis(beta-aminoethyl ether)-N,N,N',N'-tetraacetate to prevent Ca2+-dependent proteolysis. When 10 microM GTP was present, 100 mM NaCl stimulated the activity of the rabbit enzyme 5.6-fold and that of the human enzyme 2.2-fold. Under these conditions, maximum inhibitions of 90% and 64% were obtained on addition of 100 nM 1-octadecyl-2-acetyl-G-3-PC to rabbit and human preparations, respectively. These inhibitions resulted partly from an NaCl-independent inhibition of basal enzyme activity and partly from reversal of the stimulatory effect of NaCl. The relative abilities of the chlorides of different monovalent cations to enhance inhibition of rabbit platelet adenylate cyclase were: NaCl greater than LiCl greater than KCl greater than choline chloride. NaCl also increased the concentrations of 1-octadecyl-2-acetyl-G-3-PC required for half-maximal inhibition of adenylate cyclase but this action of NaCl did not correlate with its stimulatory effect on enzyme activity. After particulate fractions from platelets of either species were washed, 10 microM GTP inhibited basal adenylate cyclase activity in the absence of NaCl but stimulated the enzyme in the presence of NaCl. Inhibition of adenylate cyclase by 1-octadecyl-2-acetyl-G-3-PC was then either enhanced by GTP (rabbit material) or completely dependent on added GTP (human material). Stimulation of the activity of the washed human preparations by NaCl required GTP, but concentrations lower than required for potentiation of the inhibitory effect of 1-octadecyl-2-acetyl-G-3-PC by NaCl were effective.

Secretion of platelet granule constituents is closely associated with the phosphorylation of a cytosol polypeptide of Mr = 47,000 that we have called P47 (Haslam, R. J., Lynham, J. A., and Fox, J. E. B. (1979) Biochem. J. 178, 397-406). This polypeptide is a substrate of Ca2+-activated phospholipid-dependent protein kinase (Kawahara, Y., Takai, Y., Minakuchi, R., Sano, K., and Nishizuka, Y. (1980) Biochem. Biophys. Res. Commun. 97, 309-317). Two-dimensional gel electrophoresis of protein from human platelets that had been preincubated with 32Pi demonstrated the presence under control conditions of 2-3 major forms of P47 that contained very little 32P (pI values, 6.6-6.8) and, after induction of secretion with thrombin, their replacement by 7-9 highly labeled phosphorylated forms of P47 (pI values, 6.1-6.5). Native phosphorylated P47 was purified from thrombin-stimulated 32P-labeled platelets by ammonium sulfate fractionation and column chromatography on DEAE-cellulose, phenyl-Sepharose, and hydroxylapatite. The final 32P-labeled product was obtained in a yield of 20-25% and was purified about 400-fold relative to platelet lysate. This material was homogeneous on sodium dodecyl sulfate-polyacrylamide gel electrophoresis but, like the starting material, contained 7-9 separable phosphorylated components with different pI values. Purified phosphorylated P47 had a sedimentation coefficient (s20,w) of 3.57 S and a Stokes radius of 3.33 nm from which an Mr = 49,000 and a frictional ratio (f/f0) of 1.4 were calculated. These findings and failure to detect multimers after treatment of the protein with dimethyl suberimidate indicate that P47 normally exists as a monomer. The 32P-labeled phosphate present in purified P47 had the chemical stability of serine or threonine phosphoesters and analysis indicated the presence of 83% phosphoserine and 17% phosphothreonine. Limited proteolysis of purified 32P-labeled P47 by Staphylococcus aureus V8 protease generated a major unlabeled fragment (Mr = 23,500) and up to six labeled fragments (Mr = 24,700-14,800), the relative amounts of the latter depending on the extent of proteolysis. The same labeled fragments were obtained after proteolysis of each of the major phosphorylated components of P47, suggesting that these represent different phosphorylation states of variants of the same protein and that most or all of the phosphorylation sites are on a single 14,800-Da segment of the protein. The availability of pure native phosphorylated P47 should facilitate investigation of the physiological role of this protein in platelets.

After human platelets were lysed by freezing and thawing in the presence of EDTA, about 35% of the total cyclic AMP-dependent protein kinase activity was specifically associated with the particulate fraction. In contrast, Ca2+-activated phospholipid-dependent protein kinase was found exclusively in the soluble fraction. Photoaffinity labelling of the regulatory subunits of cyclic AMP-dependent protein kinase with 8-azido-cyclic [32P]AMP indicated that platelet lysate contained a 4-fold excess of 49 000-Da RI subunits over 55 000-Da RII subunits. The RI and RII subunits were found almost entirely in the particulate and soluble fractions respectively. Chromatography of the soluble fraction on DEAE-cellulose demonstrated a single peak of cyclic AMP-dependent activity with the elution characteristics and regulatory subunits characteristic of the type-II enzyme. A major enzyme peak containing Ca2+-activated phospholipid-dependent protein kinase was eluted before the type-II enzyme, but no type-I cyclic AMP-dependent activity was normally observed in the soluble fraction. The particulate cyclic AMP-dependent protein kinase and associated RI subunits were solubilized by buffers containing 0.1 or 0.5% (w/v) Triton X-100, but not by extraction with 0.5 M-NaCl, indicating that this enzyme is firmly membrane-bound, either as an integral membrane protein or via an anchor protein. DEAE-cellulose chromatography of the Triton X-100 extracts demonstrated the presence of both type-I cyclic AMP-dependent holoenzyme and free RI subunits. These results show that platelets contain three main protein kinase activities detectable with histone substrates, namely a membrane-bound type-I cyclic AMP-dependent enzyme, a soluble type-II cyclic AMP-dependent enzyme and Ca2+-activated phospholipid-dependent protein kinase, which was soluble in lysates containing EDTA.

Rabbit platelets were labeled in vivo by intravenous injection of 35SO4=60 hours before the animals were exsanguinated. The labeled platelets were washed and resuspended in Tyrode's solution containing 0.35 per cent albumin. The aggregated normally upon the addition of ADP and during aggregation 14.0 +/- 2.7 per cent (mean +/- standard deviation) of the labeled material was lost into the suspending medium. The extent of loss was dependent on ADP concentration and was inhibited by inhibitors of ADP-induced aggregation (AMP, ATP, adenosine, prostaglandin E1, parachloromercuribenzensulfonate and lack of calcium). Since neither release of granule contents nor lysis occurs when rabbit platelets are aggregated by ADP, it was concluded that the labeled material that was lost had been part of the surface coat of the platelets. The labeled material has been identified in other studies as a proteoglycan in which the mucopolysaccharide is principally chondroitin sulfate A. Loss of this material did not make the platelets unresponsive to ADP. The platelets did not reincorporate labeled proteoglycan from plasma. Thrombin caused the release of 50 to 80 per cent of the total platelet-labeled material, presumably from both the granules and the surface. The functional significance of the loss of surface proteoglycan during platelet aggregation is unknown.

In platelets, agonists that stimulate phosphoinositide turnover cause the rapid phosphorylation of a protein of apparent relative molecular mass (Mr) 40-47,000, called P47, by protein kinase C (PKC). Diverse identities have been ascribed to P47 including lipocortin, inositol 1,4,5-trisphosphate 5-phosphomonoesterase, pyruvate dehydrogenase alpha subunit and an actin regulatory protein. We have isolated human P47 clones by immunological screening of a lambda gt11 complementary DNA library from HL-60 cells, a human promyelocytic leukaemia cell line. P47 recombinants thus identified hybridized to a 3.0 kilobase (kb) messenger RNA in mature white blood cell lines; the same mRNA was induced in HL-60 cells during differentiation. A 1,050 base pair (bp) open reading frame that could encode a protein of Mr40,087 was confirmed by comparison with peptide sequences from platelet P47, and by expression of the putative recombinant P47 in E. coli and in vitro. The P47 sequence appears to have been conserved throughout vertebrate evolution, and is not similar to any other known sequence including human lipocortin and the alpha subunit of pyruvate dehydrogenase. The P47 protein contains a potential Ca2+-binding 'EF-hand' structure and a region that strongly resembles known PKC phosphorylation sites.

Membrane proteins from rabbit and human platelets were separated by SDS/polyacrylamide-gel electrophoresis and the resolved polypeptides blotted on nitrocellulose. A family of GTP-binding proteins, termed Gn proteins, was detected by incubation of these blots with [alpha-32P]GTP in the presence of Mg2+. A major Gn protein with a molecular mass of 27 kDa (Gn27) and lesser amounts of 23, 24 and 25 kDa Gn proteins were observed in platelet membranes; much smaller amounts were in the platelet soluble fraction. Binding of [alpha-32P]GTP by platelet Gn proteins was blocked by GDP, GTP or guanosine 5'-[gamma-thio]triphosphate, but not by GMP or adenosine 5'-[beta gamma-imido]triphosphate. Rabbit and human red-cell membranes contained only Gn27. When rat tissues were analysed for Gn proteins, the largest amounts were found in brain, which contained two membrane-bound forms (Gn27 and Gn26) and a soluble form (Gn26).

Activation of protein kinase C (PKC) in platelets causes the immediate phosphorylation of pleckstrin, an apparent Mr 40-47,000 protein previously called 40K or P47. Pleckstrin presumably plays an important but as yet unknown role in mediating cellular responses evoked by agonist-induced phosphoinositide turnover. We have cloned the cDNA for pleckstrin from the HL-60 human promyelocytic leukemia cell line by immunological screening of a lambda gt11 expression library (Tyers et al.: Nature 333:470-473, 1988) and now report further analysis of the pleckstrin sequence. Pleckstrin has a deduced Mr of 40,087 and is encoded by a 1,050-bp open reading frame which is preceded by a short open reading frame that terminates before the correct initiator methionine. A single polymorphic site was found in the coding region. An unusual pattern of sequence heterogeneity occurred about a poly(A) tract in the 3' untranslated region. The 3.0-kb pleckstrin mRNA induced upon differentiation of HL-60 cells apparently has heterogeneous 5' ends which undergo differential regulation during HL-60 cell maturation. Analysis by multiple sequence alignment with known PKC substrates identified a strong candidate site for phosphorylation by PKC and a potential Ca2+-binding EF-hand motif. No other similarities to proteins in current databases were found.

Inhibition of adenylate cyclase in intact platelets by addition of compounds such as 2', 5' - dideoxyadenosine prevented the inhibition of platelet aggregation by PGE1 but did not affect the responses of platelets to aggregating agents in the absence of PGE1. This confirms that cyclic AMP mediates the effects of PGE1 but indicates that the level of cyclic AMP in unstimulated platelets is too low to affect the actions of aggregating agents. Studies on the phosphorylation of proteins in intact 32P-labelled platelets showed that PGE1 increased the phosphorylation of a membrane-bound polypeptide (P24) and prevented the increased phosphorylation of other polypeptides (P47 and P20) that occurred on addition of inducers of the release reaction. It is suggested that the cyclic AMP-dependent phosphorylation of P24 stimulates the active transport of Ca(2+) out of the platelet cytosol, so preventing phosphorylation of P47 and P20, reactions which may be involved in the release mechanism. As increases in platelet cyclic GMP could be dissociated from both platelet aggregation and the release reaction, it is proposed that the bidirectional regulation of platelet function is achieved primarily by the opposing actions of increases in the concentrations of Ca(2+) and cyclic AMP.

Arg8-vasopressin inhibited the adenylate cyclase activity of human platelet particulate fraction up to a maximum of 27% (IC50 = 1.2 nM). This inhibition required the presence of 10 microM GTP and was optimal with 100 mM NaCl. Orn8-vasopressin had similar effects. 1-Deamino-Val4, D-Arg8-vasopressin did not by itself affect adenylate cyclase activity but competitively inhibited the action of Arg8-vasopressin (pA2 = 7.74). Arg8-vasopressin did not inhibit adenylate cyclase in intact platelets but instead caused platelet aggregation, an effect that was also competitively inhibited by 1-deamino-Val4, D-Arg8-vasopressin (pA2 = 7.82). Thus, platelets possess vasopressin receptors of the V1 type that, under appropriate conditions, can mediate either inhibition of platelet adenylate cyclase or platelet aggregation.

The 27 kDa platelet membrane protein (Gn27) that binds [alpha-32P]GTP on nitrocellulose blots of SDS-polyacrylamide gels [(1987) Biochem. J. 245, 617-620] was compared with other low molecular mass GTP-binding proteins. Platelet membranes also contained 21 kDa proteins that bound anti-ras p21 antibody and 22-23 kDa proteins that could be ADP-ribosylated by botulinum neurotoxin type D. These groups of proteins were resolved electrophoretically from each other and from Gn27. A low molecular mass GTP-binding protein from bovine brain [(1987) Biochem. J. 246, 431-439] was also resolved from Gn27. At the levels normally present in cell membranes, only Gn-proteins bound significant amounts of [32P]GTP after transfer of protein from SDS-polyacrylamide gels to nitrocellulose.

Human platelets containing granule-bound [14C]serotonin were permeabilized, equilibrated at 0 degrees C with ATP and with various Ca2+ buffers and guanine nucleotides, and then incubated at 25 degrees C with or without a stimulatory agonist. Ca2+ alone induced the ATP-dependent secretion of [14C]serotonin (50% at a pCa of 5.1) but the sensitivity of secretion to Ca2+ was greatly enhanced by guanine nucleotides [6-fold by 100 microM GTP, 100-fold by 100 microM guanyl-5'-yl imidodiphosphate and greater than 500-fold by 100 microM guanosine 5'-O-(3-thiotriphosphate)] or by stimulatory agonists (10-fold by 2 units thrombin/ml and 4-fold by 1 microM 1-O-octadecyl-2-O-acetyl-sn-glyceryl-3-phosphorylcholine). When both GTP and a stimulatory agonist were added, they had synergistic effects on secretion. Cyclic GMP and GMP acted similarly to GTP. The effects of all these guanine nucleotides were inhibited by guanosine 5'-O-(2-thiodiphosphate), whereas those of stimulatory agonists were not. Our results demonstrate the presence in platelets of guanine nucleotide-dependent and independent mechanisms regulating the sensitivity of secretion to Ca2+.

1. Under optimal ionic conditions (4 mM-MnCl2) the specific activity of guanylate cyclase in fresh platelet lysates was about 10nmol of cyclic GMP formed/20 min per mg of protein at 30 degrees C. Activity was 15% of optimum with 10mM-MgCl2 and negligible with 4mM-CaCl2. Synergism between MnCl2 and MgCl2 or CaCl2 was observed when [MnCl2] less than or equal to [GPT]. 2. Lower than optimal specific activities were obtained in assays containing large volumes of platelet lysate, owing to the presence of inhibitory factors that could be removed by ultrafiltration. Adenine nucleotides accounted for less than 50% of the inhibitory activity. 3. Preincubation of lysate for 1 h at 30 degrees C increased the specific activity of platelet guanylate cyclase by about 2-fold. 4. Lubrol PX (1%, w/v) stimulated guanylate cyclase activity by 3--5-fold before preincubation and by about 2-fold after preincubation. Triton X-100 was much less effective. 5. Dithiothreitol inhibited the guanylate cyclase activity of untreated, preincubated and Lubrol PX-treated lysates and prevented activation by preincubation provided that it was added beforehand. 6. Oleate stimulated guanylate cyclase activity 3--4-fold and arachidonate 2--3-fold, whereas palmitate was almost inactive. Pretreatment of lysate with indomethacin did not inhibit this effect of arachidonate. Oleate and arachidonate caused marked stimulation of guanylate cyclase in preincubated lysate, but inhibited the enzyme in Lubrol PX-treated lysate. 7. NaN3 (10mM) increased guanylate cyclase activity by up to 7-fold; this effect was both time- and temperature-dependent. NaN3 did not further activate the enzyme in Lubrol PX-treated lysate. 8. The results indicated that preincubation, Lubrol PX, fatty acids and NaN3 activated platelet guanylate cyclase by different mechanisms. 9. Platelet particulate fractions contained no guanylate cyclase activity detectable in the presence or absence of Lubrol PX that could not be accounted for by contaminating soluble enzyme, suggesting that physiological aggregating agents may increase cyclic GMP in intact platelets through the effects of intermediary factors. The activated and inhibited states of the enzyme described in the present paper may be relevant to the actions of these factors.

Platelet-activating factor (PAF), which is thought to cause platelet aggregation and degranulation via a receptor-mediated activation of phospholipase C, had no direct action on PGE1-stimulated cyclic AMP formation in intact human platelets, although it caused a GTP and Na+-dependent inhibition of the adenylate cyclase activity of human platelet particulate fractions. Studies with PAF analogues indicated that the receptors mediating this inhibition of adenylate cyclase had structural specificity very similar or identical to that of the receptors mediating platelet aggregation. These results suggest that the PAF receptors linked to the activation of phospholipase C in intact platelets may, in membrane preparations, become coupled to the inhibition of adenylate cyclase via the guanine nucleotide-binding protein, Gi. Studies with permeabilized human platelets that secrete 5-HT on addition of low concentrations of Ca2+ showed that addition of either PAF or a guanine nucleotide decreased the Ca2+ required for secretion. When added together, PAF and GTP promoted secretion synergistically at low Ca2+ concentrations. Enhanced secretion of 5-HT was associated with increased formation of diacylglycerol. These results show that PAF can stimulate phospholipase C by both GTP-dependent and independent mechanisms. In intact human platelets, PAF receptors may interact preferentially with a guanine nucleotide-binding protein that promotes phosphoinositide breakdown by phospholipase C, rather than with Gi.

Rabbit platelets were labelled with [3H]inositol and a membrane fraction was isolated in the presence of ATP, MgCl2 and EGTA. Incubation of samples for 10 min with 0.1 microM-Ca2+free released [3H]inositol phosphates equivalent to about 2.0% of the membrane [3H]phosphoinositides. Addition of 10 microM-guanosine 5'-[gamma-thio]triphosphate (GTP[S]) caused an additional formation of [3H]inositol phosphates equivalent to 6.6% of the [3H]phosphoinositides. A half-maximal effect was observed with 0.4 microM-GTP[S]. The [3H]inositol phosphates that accumulated consisted of 10% [3H]inositol monophosphate, 88% [3H]inositol bisphosphate ([3H]IP2) and 2% [3H]inositol trisphosphate ([3H]IP3). Omission of ATP and MgCl2 led to depletion of membrane [3H]polyphosphoinositides and marked decreases in the formation of [3H]inositol phosphates. Thrombin (2 units/ml) or GTP (4-100 microM) alone weakly stimulated [3H]IP2 formation, but together they acted synergistically to exert an effect comparable with that of 10 microM-GTP[S]. The action of thrombin was also potentiated by 0.1 microM-GTP[S]. Guanosine 5'-[beta-thio]diphosphate not only inhibited the effects of GTP[S], GTP and GTP with thrombin, but also blocked the action of thrombin alone, suggesting that this depended on residual GTP. Incubation with either GTP[S] or thrombin and GTP decreased membrane [3H]phosphatidylinositol 4-phosphate ([H]PIP) and prevented an increase in [3H]phosphatidylinositol 4,5-bisphosphate ([3H]PIP2) observed in controls. Addition of unlabelled IP3 to trap [3H]IP3 before it was degraded to [3H]IP2 showed that only about 20% of the additional [3H]inositol phosphates that accumulated with GTP[S] or thrombin and GTP were derived from the action of phospholipase C on [3H]PIP2. The results provide further evidence that guanine-nucleotide-binding protein mediates signal transduction between the thrombin receptor and phospholipase C, and suggest that PIP may be a major substrate of this enzyme in the platelet.