Stimulation by serotonin of 40 kDa and 20 kDa protein phosphorylation in human platelets

In human platelets, serotonin is known to induce a shape change followed by (reversible) aggregation. Recently, it was found that the amine triggers the elevation of cytosolic free calcium and activates phospholipase C. On stimulation of human platelets with serotonin we found an immediate increase in protein kinase C activity, phosphorylating its 40 kDa substrate protein. A 20 kDa protein, most likely the myosin light chain, was phosphorylated to the same extent. Ketanserin, a highly selective serotonin-S2 antagonist inhibited both phosphorylation processes at subnanomolar concentrations.     

Localization of protein kinase A and vitronectin in resting platelets and their translocation onto fibrin fibers during clot formation

Physiological stimulation of platelets with thrombin brings about the release of protein kinase A (PKA) into the plasma. In human blood, this kinase singles out and phosphorylates vitronectin (Vn), a multifunctional regulatory protein, which was proposed to play an important role in the control of fibrinolysis. Here we present immuno-cytochemical evidence to show: (i) that intact platelets possess on their surface an ecto-PKA which can preferentially phosphorylate Vn; (ii) that in the resting platelet, both the catalytic and the regulatory subunits of PKA are present on the platelet surface, in the surface-connected canalicular system, and within the alpha-granules of the platelets; (iii) that the process initiated upon platelet activation, which leads to the formation of fibrin fibers and consequently forms the fibrin net, is accompanied by a translocation of PKA, of Vn, and of PAI-1 onto the fibrin fibers. We propose that the localization and the translocation of these proteins in the fibrin net, together with our finding that PKA phosphorylation of Vn reduces its grip of PAI-1, can unleash PAI-1 in its free form. The free PAI-1 can then assume its latent (non inhibitory) conformation, allow plasminogen activators to trigger the formation of active plasmin, and to initiate fibrinolysis.     

Purification of calcium-sensitive regulatory protein of platelets which inhibits the gelation of actin

Ca²⁺-sensitive regulatory protein of human platelets which inhibits the gelation of actin was purified by DEAE-Sepharose and an affinity column using actin as a ligand. The protein was a single polypeptide chain with an average molecular weight of 90,000 and it bound to actin and inhibited its gelation at concentration from 10^?6–10^?7M of free calcium. Since the protein existed in the form of a complex with actin even though at concentration lower than 10^?7M of free calcium, binding and dissociation of actin and the protein appeared to be dependent on the concentration of free calcium, and complete dissociation was not seen.     

Calcium sequestration in human platelets: is it stimulated by protein kinase C?

Sequestration of calcium into an intracellular storage site is an important mechanism in helping to maintain a low cytoplasmic Ca2+ level in many cells. In platelets, increasing cytoplasmic cAMP lowers the free calcium level in correlation with the phosphorylation of a 22 kD protein. This protein has been thought to enhance uptake of calcium into a platelet membrane bound storage site by activating a calcium-ATPase activity by analogy with phospholamban in cardiac muscle. The evidence for an analogue of phospholamban in platelets is unclear. A pathway involving cAMP dependent kinase also seems unlikely to account for the transience of the calcium signal following agonists in platelets, some of which inhibit the cAMP dependent kinase. Here we discuss the issue of whether activation of protein kinase C, which follows agonist action, leads to enhanced calcium sequestration in platelets and if so, what indications there are for a mechanism. The evidence from our experiments with phorbol myristate acetate treated platelets shows that such an enhancement can be produced by activating protein kinase C. Phosphorylation studies suggest the involvement of a polypeptide or polypeptides distinct from the 22 kD polypeptide. Further work to test this idea is necessary. A brief overview of research on the role of phosphoproteins in calcium regulation in platelets and comparison with their role in cardiac muscle is also presented.     

Binding of anticoagulant vitamin K-dependent protein S to platelet-derived microparticles.

Vitamin K-dependent protein S is an anticoagulant plasma protein serving as cofactor to activated protein C in degradation of coagulation factors Va and VIIIa on membrane surfaces. In addition, it forms a noncovalent complex with complement regulatory protein C4b-binding protein (C4BP), a reaction which inhibits its anticoagulant function. Both forms of protein S have affinity for negatively charged phospholipids, and the purpose of the present study was to elucidate whether they bind to the surface of activated platelets or to platelet-derived microparticles. Binding of protein S to human platelets stimulated with various agonists was examined with FITC-labeled monoclonal antibodies and fluorescence-gated flow cytometry. Protein S was found to bind to membrane microparticles which formed during platelet activation but not to the remnant activated platelets. Binding to microparticles was saturable and maximum binding was seen at approximately 0.4 microM protein S. It was calcium-dependent and reversed after the addition of EDTA. Inhibition experiments with monoclonal antibodies suggested the gamma-carboxyglutamic acid containing module of protein S to be involved in the binding reaction. An intact thrombin-sensitive region of protein S was not required for binding. The protein S-C4BP complex did not bind to microparticles or activated platelets even though it bound to negatively charged phospholipid vesicles. Intact protein S supported binding of both protein C and activated protein C to microparticles. Protein S-dependent binding of protein C/activated protein C was blocked by those monoclonal antibodies against protein S that inhibited its cofactor function. In conclusion, we have found that free protein S binds to platelet-derived microparticles and stimulates binding of protein C/activated protein C.(ABSTRACT TRUNCATED AT 250 WORDS)     

Phorbol ester stimulates calcium sequestration in saponized human platelets

When platelets are activated by agonists, calcium (Ca2+) is released from an intracellular storage site. Recent studies using fura-2 show that, after thrombin stimulation, the rise in free calcium is transient and returns to base-line levels in 2-3 min, while the transient following ADP stimulation lasts only 15-20 s. We reported previously that the phorbol ester 12,13-phorbol myristate acetate (PMA), added at nanomolar levels after thrombin, immediately accelerated the rate of return of calcium to the base line severalfold (Yoshida, K., Dubyak, G., and Nachmias, V.T. (1986) FEBS Lett. 206, 273-278). In the present study, we used both intact and saponized platelets to determine whether this is due to stimulation of calcium sequestration. Using fura-2 and intact platelets, we found 1) that PMA stimulated the restoration of free Ca2+ levels after ADP as well as after thrombin, and 2) that H-7, an inhibitor of protein kinase C (Ca2+/phospholipid-dependent enzyme), slowed the return of Ca2+ to baseline levels. Using saponized platelets, we also found 3) that pretreatment of platelets with PMA before saponin treatment increased the ATP-dependent 45Ca2+ uptake 2-fold, with a half-maximal effect at 5 nm; 4) that most of the Ca2+ released by ionomycin or by myoinositol 1,4,5-trisphosphate; and 5) that a GTP-binding protein inhibitor, guanosine 5'-O-(2-thiodiphosphate), decreased basal or PMA-stimulated 45Ca2+ uptake in saponin-treated platelets. Our data suggest that activation of protein kinase C stimulates the sequestration of Ca2+ independently of cAMP or myoinositol 1,4,5-trisphosphate.     

Identification of ecto-PKC on surface of human platelets: role in maintenance of latent fibrinogen receptors

Human platelets express a protein phosphorylation system on their surface. A specific protein kinase C (PKC) antibody, monoclonal antibody (MAb) 1.9, which binds to the catalytic domain of PKC and inhibits its activity, causes the aggregation of intact platelets while inhibiting the phosphorylation of platelet surface proteins. Photoaffinity labeling with 100 nM 8-azido-[alpha(32)P]ATP identified this ecto-PKC as a single surface protein of 43 kDa sensitive to proteolysis by extracellular 0.0005% trypsin. Inhibition of the binding of 8-azido-[alpha(32)P]ATP to the 43-kDa surface protein by MAb 1.9 identified this site as the active domain of ecto-PKC. Covalent binding of the azido-ATP molecule to the 43-kDa surface protein inhibited the phosphorylative activity of the platelet ecto-PKC. Furthermore, PKC pseudosubstrate inhibitory peptides directly induced the aggregation of platelets and inhibited azido-ATP binding to the 43-kDa protein. Platelet aggregation induced by MAb 1.9 and by PKC inhibitory peptides required the presence of fibrinogen and resulted in an increase in the level of intracellular free calcium concentration. This increase in intracellular free calcium concentration induced by MAb 1.9 was found to be dependent on the binding of fibrinogen to activated GPIIb/IIIa integrins, suggesting that MAb 1.9 causes Ca(2+) flux through the fibrinogen receptor complex. We conclude that a decrease in the state of phosphorylation of platelet surface proteins caused by inhibition of ecto-PKC results in membrane rearrangements that can induce the activation of latent fibrinogen receptors, leading to platelet aggregation. Accordingly, the maintenance of a physiological steady state of phosphorylation of proteins on the platelet surface by ecto-PKC activity appears to be one of the homeostatic mechanisms that maintain fibrinogen receptors of circulating platelets in a latent state that cannot bind fibrinogen.     

Selective interaction of cytoskeletal proteins with liposomes

A protein kinase activity analogous to that found in interferon-treated HeLa cells is detectable in human plasma rich in platelets. This kinase activity is manifested by the phosphorylation of an endogenous Mr 72000 protein which could be conveniently assayed after partial purification on poly(G)-Sepharose. Here, we show that the protein kinase system in the plasma consists of at least 2 components. The protein kinase is found to be localised in the platelets whereas most of the substrate (the Mr 72000 protein) is found free in the plasma and a fraction of it associated with the surface of platelets.     

Phosphatidylinositol turnover in platelet activation; calcium mobilization and protein phosphorylation

Ca2+-activated, phospholipid-dependent protein kinase (C-kinase) in platelets is normally activated by diacylglycerol, which is derived from phosphatidylinositol through its receptor-linked breakdown. Under appropriate conditions this enzyme can also be activated by synthetic diacylglycerol which is directly added to intact platelets. C-Kinase thus activated preferentially phosphorylates an endogenous platelet protein having a molecular weight of approximately 40,000. This protein phosphorylation is merely a prerequisite but not a sufficient requirement for the release of serotonin. Evidence is presented suggesting that Ca2+ mobilization and C-kinase activation are synergistically involved in the physiological response of platelets to extracellular messengers, such as thrombin, collagen and platelet-activating factor.     

Collagen-induced exposure of anionic phospholipid in platelets and platelet-derived microparticles.

We have shown recently that the calcium-dependent phospholipid-binding protein annexin V (placental anticoagulant protein I) can be used to study the exposure of anionic phospholipid after platelet activation. In this study we have further examined the mechanism of this process. Collagen-induced exposure of annexin V binding sites correlated directly with increased ability to support activity of the reconstituted prothrombinase complex. The potency of annexin V as an inhibitor of platelet prothrombinase was the same as its Kd for platelets. Prior incubation of platelets with 5'-p-fluorosulfonylbenzoyladenosine or p-chloromercuribenzenesulfonate had no significant effect on annexin V binding. Similarly, inhibition of platelet cyclic endoperoxide synthesis by acetylsalicylic acid or indomethacin did not inhibit annexin V binding. Staurosporine inhibited collagen-induced, but not A23187-induced, annexin V binding. Agents that increase intraplatelet cyclic nucleotides partially inhibited collagen-induced annexin V binding. Thus, collagen-induced exposure of anionic phospholipid appears to depend primarily on increases in intraplatelet free calcium and may be independent of ADP- or endoperoxide-mediated pathways. Binding sites for annexin V on microparticles derived from collagen-stimulated platelets were demonstrated by flow cytometry and gel filtration. In addition, prior incubation of platelets with 100 nM annexin V inhibited factor Va binding to both platelets and platelet-derived microparticles. These results support the concept that the procoagulant effect of platelets and platelet-derived microparticles is mediated by calcium-induced exposure of anionic phospholipids.     

Exogenous sn-1,2-diacylglycerols containing saturated fatty acids function as bioregulators of protein kinase C in human platelets

The ability of exogenous sn-1,2-diacylglycerols and analogs to function as bioregulators of protein kinase C in human platelets was investigated. The activation of protein kinase C in platelets is indicated by specific phosphorylation of a 40,000-dalton protein. Dihexanoylglycerol, dioctanoylglycerol (diC8), didecanoylglycerol, and sn-1-oleoyl-2-acetylglycerol were active in stimulating 40,000-dalton protein phosphorylation. Only a trace of phosphorylation was elicited by dibutyrylglycerol. Phosphorylation was not induced by analogs of diC8 in which an -H, -SH, or -Cl group replaced the free -OH, nor by monoacylglycerols or long chain diacylglycerols. Maximum phosphorylation was induced by dihexanoylglycerol, diC8, and didecanoylglycerol at concentrations from 5 to 20 microM and between 5 and 30 S after exposure of platelets to these diacylglycerols. Under conditions of maximal phosphorylation of the 40,000-dalton protein, these diacylglycerols did not induce phosphatidylinositol turnover, or platelet aggregation, or stimulate release of ATP or serotonin. A small degree of aggregation was evident with platelets isolated in the absence of prostacyclin, and release of serotonin was observed when 1 mM Ca2+ or submaximal concentrations of ionophore A23187 were included. These results are consistent with a model in which platelet activation requires the simultaneous formation of two intracellular signals, diacylglycerols and Ca2+. These diacylglycerols and diacylglycerol analogs provide useful tools to investigate the function of diacylglycerols as bioregulators in intact cells.     

Phorbol esters modulate thrombin-operated calcium mobilisation and dense granule release in human platelets

Human alpha-thrombin-induced elevation of cytosolic free calcium ([Ca2+]i) and dense granule release was examined in platelets preincubated with either activators or an inhibitor of protein kinase C. 12-O-Tetradecanoylphorbol 13-acetate (TPA) or two 12-deoxy analogues of TPA, when added alone to platelets, did not elevate [Ca2+]i, as monitored by quin2 fluorescence, though small amounts of dense granule release were detected. Preincubation of the platelets with either TPA or 12-deoxyphorbol 13-phenylacetate, but not the parent, 4-beta-phorbol, produced a dose-dependent inhibition of the elevation of [Ca2+]i and 5-hydroxytryptamine release induced by human alpha-thrombin. Furthermore, this phorbol ester-mediated inhibition of human alpha-thrombin-induced activation could be prevented by H7 (1-[5-isoquinolinesulphonyl]-2-methylpiperazine), the recently described inhibitor of protein kinase C. These results suggest a role for protein kinase C as a modulator of receptor-operated calcium fluxes in human platelets.     

Primary stimuli of icosanoid release inhibit arachidonoyl-CoA synthetase and lysophospholipid acyltransferase. Mechanism of action of hydrogen peroxide and methyl mercury in platelets

Icosanoid formation in platelets depends on the concentration of free arachidonate that is mainly liberated from membrane phospholipids by phospholipase A2. The concentration of free arachidonate is also controlled by the activities of the reacylating enzymes arachidonoyl-CoA synthetase and lysophospholipid acyltransferase. In human platelet microsomes we determined the high enzyme activities of 5.9 nmol.min-1.(10(9) platelets)-1 for the arachidonoyl-CoA synthetase and 37 nmol.min-1.(10(9) platelets)-1 for the lysophospholipid acyltransferase. The activities of these reacylating enzymes were strongly reduced by hydrogen peroxide (H2O2) and methyl mercury that are primary stimuli of arachidonate release in intact platelets. H2O2 inhibited the arachidonoyl-CoA synthetase with an IC50 of 3.3 mmol/l without affecting the lysophospholipid acyltransferase. Sulfhydryl group protection by 3-mercapto-1,2-propanediol did not overcome the inhibition but glutathione prevented the inhibition of the arachidonoyl-CoA synthetase by H2O2. This suggests that glutathione by virtue of the glutathione peroxidase reduces H2O2 rather than that it protects free sulfhydryl groups of the arachidonoyl-CoA synthetase. Methyl mercury left the arachidonoyl-CoA synthetase activity unaffected but inhibited the lysophospholipid acyltransferase activity with an IC50 of 3.4 mumol/l. The inhibition is probably evoked by the blockade of sulfhydryl groups of the lysophospholipid acyltransferase because it disappeared when 3-mercapto-1,2-propanediol was added at a concentration higher than that of methyl mercury. Thrombin as a physiological full agonist, Ca2+ less than or equal to 1 mmol/l, the calcium ionophore A23187 and phorbol 12-myristate 13-acetate (TPA) and 1-oleoyl-2-acetylglycerol as model stimuli of protein kinase C neither influenced arachidonoyl-CoA synthetase nor lysophospholipid acyltransferase. It is concluded that the inhibitory effect of H2O2 and methyl mercury on the arachidonate-reacylating enzymes arachidonoyl-CoA synthetase or lysophospholipid acyltransferase, respectively, are responsible for their capacity to stimulate icosanoid release in intact cells. Thrombin and its intracellular messengers Ca2+ and diacylglycerol do not directly affect arachidonoyl-CoA synthetase and lysophospholipid acyltransferase.     

CalDAG-GEFI integrates signaling for platelet aggregation and thrombus formation

Signaling through the second messengers calcium and diacylglycerol (DAG) is a critical element in many biological systems. Integration of calcium and DAG signals has been suggested to occur primarily through protein kinase C family members, which bind both calcium and DAG. However, an alternative pathway may involve members of the CalDAG-GEF/RasGRP protein family, which have structural features (calcium-binding EF hands and DAG-binding C1 domains) that suggest they can function in calcium and DAG signal integration. To gain insight into the signaling systems that may be regulated by CalDAG-GEF/RasGRP family members, we have focused on CalDAG-GEFI, which is expressed preferentially in the brain and blood. Through genetic ablation in the mouse, we have found that CalDAG-GEFI is crucial for signal integration in platelets. Mouse platelets that lack CalDAG-GEFI are severely compromised in integrin-dependent aggregation as a consequence of their inability to signal through CalDAG-GEFI to its target, the small GTPase Rap1. These results suggest that analogous signaling defects are likely to occur in the central nervous system when CalDAG-GEFI is absent or compromised in function.     

Activation-dependent redistribution of the adhesion plaque protein, talin, in intact human platelets [published erratum appears in J Cell Biol 1990 Mar;110(3):865]

Talin is a high molecular weight protein localized at adhesion plaques in fibroblasts. It binds vinculin and integrin and appears to participate in generating a transmembrane connection between the extracellular matrix and the cytoskeleton. We have recently shown that talin is an abundant protein in platelets, cells highly specialized for regulated adhesion. Although talin constitutes greater than 3% of the total protein in intact human platelets, its location within the cells had not been defined. In the work reported here, we have investigated the distribution of talin in resting and activated human platelets by immunofluorescence and immunoelectron microscopy. We have found that talin undergoes an activation-dependent change in its subcellular location. In resting platelets, which are nonadhesive, talin is uniformly distributed throughout the cytoplasm. In contrast, in thrombin- and glass-activated, substratum-adherent platelets, talin is concentrated at the cytoplasmic face of the plasma membrane. This dramatic, regulated redistribution of talin raises the possibility that talin plays a role in the controlled development of platelet adhesion.     

Regulation of the phosphoinositide hydrolysis pathway in thrombin-stimulated platelets by a pertussis toxin-sensitive guanine nucleotide-binding protein. Evaluation of its contribution to platelet activation and comparisons with the adenylate cyclase inhibitory protein, Gi

In platelets activated by thrombin, the hydrolysis of phosphatidylinositol 4,5-bisphosphate by phospholipase C produces inositol 1,4,5-triphosphate (IP3) and diacylglycerol, metabolites which are known to cause Ca2+ release from the platelet dense tubular system and granule secretion. Previous studies suggest that phospholipase C activation is coupled to platelet thrombin receptors by a guanine nucleotide-binding protein or G protein. The present studies examine the contribution of this protein to thrombin-induced platelet activation and compare its properties with those of Gi, the G protein which mediates inhibition of adenylate cyclase by thrombin. In platelets permeabilized with saponin, nonhydrolyzable GTP analogs reproduced the effects of thrombin by causing diacylglycerol formation, Ca2+ release from the dense tubular system and serotonin secretion. In intact platelets, fluoride, which by-passes the thrombin receptor and directly activates G proteins, caused phosphoinositide hydrolysis and secretion. Fluoride also caused an increase in the platelet cytosolic free Ca2+ concentration that appeared to be due to a combination of Ca2+ release from the dense tubular system and increased Ca2+ influx across the platelet plasma membrane. Guanosine 5'-O-(2-thiodiphosphate) (GDP beta S), which inhibits G protein function, inhibited the ability of thrombin to cause IP3 and diacylglycerol formation, granule secretion, and Ca2+ release from the dense tubular system in saponin-treated platelets. Increasing the thrombin concentration overcame the effects of GDP beta S on secretion without restoring diacylglycerol formation. The effects of GDP beta S on platelet responses to thrombin which had been subjected to partial proteolysis (gamma-thrombin) were similar to those obtained with native alpha-thrombin despite the fact that gamma-thrombin is a less potent inhibitor of adenylate cyclase than is alpha-thrombin. Thrombin-induced diacylglycerol formation and 45Ca release were also inhibited when the saponin-treated platelets were preincubated with pertussis toxin, an event that was associated with the ADP-ribosylation of a protein with Mr = 41.7 kDa. At each concentration tested, the inhibition of thrombin-induced diacylglycerol formation by pertussis toxin paralleled the inhibition of thrombin's ability to suppress PGI2-stimulated cAMP formation.(ABSTRACT TRUNCATED AT 400 WORDS)     

Fibrinogen and platelet aggregation. Role of the glycopeptidic part and of the fibrinopeptide B: Description of a new technique of fibrinoglycopeptide isolation

In order to determine the active groups of the fibrinogen molecule in ADP induced aggregation, various cleavage fragments of fibrinogen were tested on plasma protein-free platelets. An original technique is described for the isolation of fibrinogen glycopeptides. The glycopeptides thus obtained exert an inhibition on platelet aggregation by ADP in the presence of fibrinogen, when incubated previously with the plasma protein free platelets. The carbohydrate fraction seems thus to have an important role on ADP platelet aggregation.The N. DSK and E fragments are inactive as cofactors of ADP induced aggregation.It is suggested that the N-terminal part of the Bβ chain does not have an important role in the cofactor activity of fibrinogen. Moreover, the importance of an intact fibrinogen molecule is underlined.     

Properties of protein kinase C subspecies in human platelets

Protein kinase C (PKC) from human platelets was resolved into two fractions by hydroxyapatite column chromatography. One of the enzymes was indistinguishable from the brain type III PKC having alpha-sequence in its kinetic and immunological properties. The other enzyme was kinetically different from any of the brain PKC subspecies so far isolated, although it resembled the brain type II PKC having beta-sequence. With H1 histone as substrate, this platelet enzyme was not very sensitive to Ca2+, and activated partly by phosphatidylserine plus diacylglycerol or by free arachidonic acid. Both platelet enzymes could phosphorylate the P47 protein in vitro, but the enzyme physiologically responsible for the P47 protein phosphorylation in the activated platelets remains to be identified.     

Formation of a stable complex of thrombin and the secreted platelet protein glycoprotein G (thrombin-sensitive protein, thrombospondin) by thiol-disulfide exchange

When 125I-labeled thrombin was incubated with washed human platelets or with the supernatant solution of activated platelets, it formed a NaDodSO4-stable complex of apparent mass greater than 450 000 daltons. Formation of the complex was temperature dependent; with 20 nM thrombin incubated with the supernatant solution of ionophore-activated platelets, the initial rate of formation of the stable complex was 1 nM thrombin/min at 37 degrees C, 50 times the rate at 22 degrees C. Thrombin with all free amino groups methylated was still reactive. Active-site-blocked thrombin formed the complex only slowly. The complex that formed with active thrombin was not dissociated by hydroxylamine in urea. Reduction with 2-mercaptoethanol dissociated the complex, and its formation was blocked by the sulfhydryl-blocking agents iodoacetamide and 4,4'-dithiodipyridine. The complex was thus unlike those of thrombin and alpha 2-macroglobulin or antithrombin III, but it had characteristics of a disulfide-linked complex. Of the secreted proteins, albumin and glycoprotein G adhered to an activated thiol-Sepharose column, indicating that they contained free thiol groups. Purified glycoprotein G and thrombin formed a complex similar to the complex formed when thrombin was incubated with the supernatant solution of activated platelets. The purified glycoprotein bound 2.6 mol of radioactive N-ethylmaleimide/mol of protein, indicating three sulfhydryl groups per mole. After reacting with purified glycoprotein G, thrombin developed a new sulfhydryl group. It is concluded that glycoprotein G (thrombin-sensitive protein, thrombospondin) and thrombin form a dissociable complex that leads to a covalent complex by thiol-disulfide exchange of a thiol group on glycoprotein G and a disulfide on thrombin.     

Structural and biosynthetic studies of the granule membrane protein, GMP-140, from human platelets and endothelial cells

GMP-140 is an integral membrane glycoprotein of apparent Mr = 140,000 located in secretory storage granules of platelets and vascular endothelial cells. When these cells are activated, GMP-140 redistributes from the membrane of the granules to the plasma membrane. To gain insight into the potential function of GMP-140, we examined aspects of its structure and biosynthesis. The amino acid composition of platelet GMP-140 revealed elevated numbers of cystinyl (6.1%), prolinyl (7.2%), and tryptophanyl (2.1%) residues. GMP-140 contained 28.8% carbohydrate by weight, distributed among N-acetylneuraminic acid, neutral sugar, and N-acetylglucosamine residues. Enzymatic removal of N-linked oligosaccarides reduced the protein's apparent Mr by more than 50,000. The biosynthesis of GMP-140 in HEL cells, which share biochemical features with megakaryocytes, was studied by pulse-chase labeling with [35S]cysteine followed by immunoprecipitation. HEL cells synthesized a heterogeneous GMP-140 precursor of 98-125 kDa which converted to a mature 140-kDa form within 40-60 min. Removal of high mannose oligosaccarides by endo-beta-N-acetylglucosaminidase H treatment reduced the apparent Mr of the precursor but not the mature protein. Tunicamycin-treated HEL cells synthesized three to four precursors of 80-92 kDa, suggesting the possibility of heterogeneity of GMP-140 at the protein level. Exposure of activated platelets to proteases followed by Western blotting indicated that most of the mass of GMP-140 was located on the extracytoplasmic side of the membrane. Our studies indicate that GMP-140 is a cysteine-rich, heavily glycosylated protein with a large extracytoplasmic domain. These features are compatible with a receptor function for the molecule when it is exposed on the surface of activated platelets and endothelial cells.