The activation of bovine coagulation factor X

Activation of human protein C by blood coagulation factor Xa in the presence of anionic phospholipids. Enhancement by sulphated polysaccharides

The activation of protein C by thrombin is thought to occur at the endothelial cell surface in the presence of an essential membrane glycoprotein cofactor, thrombomodulin. In the present study it is demonstrated that, in the presence of hirudin, the most potent known inhibitor of thrombin, human protein C can be activated by human factor Xa (20 nM), but by a thrombomodulin-independent mechanism requiring only the presence of Ca2+ and phospholipid vesicles bearing a high proportion of negative charges (30-75% phosphatidylserine, depending on the conditions). At an optimal concentration of phosphatidylserine/phosphatidylcholine (1:1, w/w) of 75 microM, the apparent Km was 1 microM with a kcat. of 1 min-1. At 25 microM-phospholipid the Km was unchanged and the kcat. was 0.67 min-1. At either lipid concentration, increasing the density of negative charges by the adjunction of sulphated polysaccharides, like pentosan polysulphate or standard heparin at optimal concentrations of 2-5 micrograms/ml and 5-10 micrograms/ml respectively, resulted in a 4-fold increase of the kcat. without affecting the Km. Sulphated polysaccharides alone were poor promoters of protein C activation by factor Xa. In any case the presence of Ca2+ was essential, the dependence being sigmoidal with Hill coefficients ranging from 1.4 to 2.0. No significant activation of 4-carboxyglutamic acid-domainless protein C, a chymotrypic derivative lacking the phospholipid-binding domain, could be detected in the presence of phospholipids and Ca2+, with or without pentosan polysulphate. In a large molar excess, other phospholipid-binding entities like prothrombin fragments F1 or F1+2 could inhibit protein C activation by factor Xa, but pentosan polysulphate exerted a clear protective effect. Factor Xa irreversibly inhibited at its active centre, but not di-isopropyl phosphoro-thrombin, behaved as an inhibitor but in a more complex manner than simple Michaelis-Menten kinetics. Among several derivatives of pentosan polysulphate or of heparin which were tested, those having the higher degree of sulphation and/or molecular mass were the most efficient in enhancing the rate of activation of protein C by factor Xa in the presence of phospholipids. These results suggest that human factor Xa, at physiological concentrations, could activate human protein C in the presence of anionic phospholipids and that this activation could be potentiated by therapeutic concentrations of sulphated polysaccharides.

Inhibition of serine proteinases by p-carbethoxyphenyl esters of epsilon-guanidino- and epsilon-amino caproic acid: thermodynamic and molecular modeling study

The inhibitory effect of the clinically used p-carbethoxyphenyl ester of epsilon-guanidino-caproic acid methanesulphonate (epsilon-GCA-CEP) on the catalytic properties of human LYS77-plasmin (EC 3.4.21.7), bovine factor Xa (EC 3.4.21.6), bovine alpha-thrombin (EC 3.4.21.5), ancrod (EC 3.4.21.28), crotalase (EC 3.4.21.30), bovine beta-trypsin (EC 3.4.21.4), porcine pancreatic beta-kallikrein-B (EC 3.4.21.35), human urinary kallikrein (EC 3.4.21.35) and the Mr 54,000 species of human urokinase (EC 3.4.21.31) was investigated (between pH 2.0 and 8.5, I = 0.1 M; T = 21 +/- 0.5 degrees C), and analyzed in parallel with that of the homologous derivative p-carbethoxyphenyl epsilon-amino-caproate hydro chloride (epsilon-ACA-CEP). On lowering the pH from 5.5 to 3.0, values of the apparent dissociation inhibition constant (Ki) for epsilon-GCA-CEP and epsilon-ACA-CEP interaction with the serine proteinases considered increase, reflecting the acidic pK-shift upon inhibitor binding of a single ionizing group. Over the whole pH range explored, (i) epsilon-GCA-CEP interacts with bovine factor Xa and bovine alpha-thrombin with an higher affinity than that observed for epsilon-ACA-CEP binding; (ii) both inhibitors associate to bovine beta-trypsin with the same affinity; and (iii) epsilon-ACA-CEP inhibits human Lys77-plasmin and the Mr 54,000 species of human urokinase with an higher affinity than that reported for epsilon-GCA-CEP association, thus reflecting the known enzyme primary specificity properties. However, the affinity of epsilon-ACA-CEP for ancrod, crotalase, porcine pancreatic beta-kallikrein-B and human urinary kallikrein, all of which preferably bind arginyl rather than lysyl side chains at the primary position of substrates and/or inhibitors, is paradoxically higher than that displayed by epsilon-GCA-CEP. By considering the amino acid sequences, the X-ray three-dimensional structures and/or the computer-generated molecular models of serine proteinase: inhibitor adducts, the observed binding behaviour of epsilon-GCA-CEP and epsilon-ACA-CEP to the enzymes considered has been related to the inferred stereochemistry of proteinase: inhibitor contact region(s).

Effects of lipid-binding proteins apo A-I, apo A-IL, beta 2-glycoprotein I, and C-reactive protein on activation of factor X by tissue factor--factor VIIa

Tissue factor is the membrane-associated protein which mediates activation of factors IX and X by factor VII. In a purified, reconstituted bovine system, factor X activation by the tissue factor-factor VIIa complex is inhibited by the mixed apoproteins from human high density lipoprotein (HDL) and by isolated apolipo-protein A-II (apo A-II). Other proteins found associated with plasma lipoproteins, apolipoprotein A-I (apo A-I), C-reactive protein (CRP), and beta 2-glycoprotein I (beta 2 GPI), have been examined for effects on the activation of factor X by tissue factor-factor VIIa. In these experiments, bovine tissue factor, reconstituted into phosphatidylserine-phosphatidylcholine (PS/PC; 30/70) vesicles, was used at a single concentration while factor X (the substrate), factor VIIa (the enzyme), and the potentially inhibitory proteins were varied in a continuous chromogenic assay. Apo A-II and CRP clearly inhibit tissue factor-factor VIIa activation of factor X, while apo A-I and beta 2 GPI have little or no effect. These results demonstrate that different lipid binding proteins vary in their effects on tissue factor activity.

Interaction of polylysine with bovine factor Xa: effect of divalent cations

Poly-L-lysine has been demonstrated to partially replace biological cofactors in the activation of prothrombin by factor Xa. The present study was initiated to determine if poly-L-lysine has an effect on the enzymatic activity of factor Xa in the absence of prothrombin. At low ionic strength (50 mM Tris-Cl, pH 8.0, ambient temperature), poly-L-lysine inhibits amidase activity (S-2222) of bovine factor Xa with high affinity (Ki = 7 nM). The inhibition was readily reversed by 100 mM NaCl. The inhibition was also markedly reduced by the addition of 1.0 mM CaCl2 but not by MnCl2 or MgCl2. All three metal ions enhance amidase activity in the absence of poly-L-lysine. Poly-L-lysine also inhibits the amidase activity of factor Xa from which the gamma-carboxyglutamic acid domain has been removed by limited proteolysis with chymotrypsin (factor Xa-GD) but with somewhat lower avidity (Ki = 35 nM). As with native factor Xa, calcium ions reduce the observed inhibition while either manganese or magnesium ions are much less effective. The amidase activity of factor Xa-GD is enhanced with any one of the three divalent cations. These results provide additional support for the existence of a functionally significant binding site for calcium ions outside of the gamma-carboxyglutamic domain of factor Xa.

Continuous chromogenic tissue factor assay: comparison to clot-based assays and sensitivity established using pure tissue factor

The continuous chromogenic tissue factor assay has been compared to one-stage and two-stage coagulation assays, and its sensitivity has been established using purified human placental tissue factor. Log-log plots of the clot-times versus A405/min2 are linear over the mutually useful ranges of tissue factor concentration. Using human placental tissue factor reconstituted into phosphatidylserine-phosphatidylcholine (30:70) vesicles, the chromogenic assay quantitatively detected tissue factor in dilutions containing from 3 femtomoles to less than 10 attomoles of protein.

Inhibition of human vitamin-K-dependent protein-S-cofactor activity by a monoclonal antibody specific for a Ca2+-dependent epitope

Protein S is an anticoagulant vitamin-K-dependent plasma protein functioning as a cofactor to activated protein C in the degradation of factors Va and VIIIa. A murine monoclonal antibody, HPS 7, specific for a calcium-stabilized epitope in human protein S, is described. The epitope was available in intact protein S, both in its free form and when protein S was bound to C4b-binding protein. It disappeared upon reduction of disulfide bridges and also after thrombin of chymotrypsin cleavage of protein S. Thrombin cleaves protein S close to the calcium-binding region containing gamma-carboxyglutamic acid (Gla). The cleaved protein still contains the Gla region, linked by a disulfide bridge, but it has a lower affinity for calcium and no protein C cofactor activity. The thrombin-mediated cleavage of protein S could be inhibited by HPS 7. The Ka for the interaction between protein S and the monoclonal was estimated to be approximately 0.7 X 10(8) M-1. Half-maximal binding between HPS 7 and protein S was observed at a calcium concentration of 0.50 mM, indicating that saturation of the Gla region with calcium was required for the interaction. The recently reported Gla-independent high-affinity calcium binding did not induce the epitope. The calcium-dependent binding of protein S to phospholipid vesicles as well as the protein C cofactor activity was inhibited by HPS 7. The data suggests that the epitope for HPS 7 is located in the Gla region of protein S or in the closely positioned thrombin-sensitive region.

Inhibition of serine proteinases by tetra-p-amidinophenoxy-neo-pentane: thermodynamic and molecular modeling study

The inhibitory effect of the aromatic tetra-benzamidine derivative tetra-p-amidinophenoxy-neo-pentane (TAPP) on the catalytic properties of beta-trypsin (EC 3.4.21.4), alpha-thrombin (EC 3.4.21.5), factor Xa (EC 3.4.21.6), Lys77-plasmin (EC 3.4.21.7) and beta-kallikrein-B (EC 3.4.21.35) was investigated (between pH 2 and 8, I = 0.1 M; T = 37 +/- 0.5 degrees C), and analyzed in parallel with that of benzamidine, commonly taken as a molecular inhibitor model of serine proteinases. Over the whole pH range explored, TAPP and benzamidine show the same values of the dissociation inhibition constant (Ki) for beta-trypsin; at variance with the affinity of TAPP for alpha-thrombin, factor Xa, Lys77-plasmin and beta-kallikrein-B which is higher than that found for benzamidine association around neutrality, but tends to converge in the acidic pH limb. On lowering the pH from 5.5 to 3.0, values of Ki for TAPP binding to beta-trypsin as well as for benzamidine association to all the enzymes investigated decreased thus reflecting the pK-shift, upon inhibitor binding, of a single ionizing group. Over the same pH range, values of Ki for TAPP binding to alpha-thrombin, factor Xa, Lys77-plasmin and beta-kallikrein-B may be described as depending on the pK-shift, upon inhibitor association, of two equivalent proton-binding amino acid residues. Considering the X-ray three-dimensional structures and the computer-generated molecular models of serine proteinases: TAPP and :benzamidine adducts, the observed binding behaviour of TAPP and benzamidine to the enzymes considered has been related to the inferred stereochemistry of proteinase: inhibitor contact region(s).

Effect of limited modification of amino groups on the reactivity of human factor Xa

The basis of the specificity of human coagulation factor Xa has been probed with a reagent that reacts with nucleophiles, N-succinimidylpropionate. At pH 8.0 and 0.25 mM N-succinimidylpropionate, 0.4 microM factor Xa lost approx. 90% of its activity toward prothrombin in 4 min. The decay was first-order, k = 0.64 min-1, which increased to 0.98 min-1 in 1 mM Ca2+, and the dependence of k upon pH was consistent with primary amines being the target. The rate of modification was unaffected by the presence of a tetrapeptide substrate during modification; likewise, activity toward a tripeptide p-nitroanilide was unaltered during exposure of factor Xa to N-succinimidylpropionate with or without Ca2+. In addition, inhibition by antithrombin III was retained with a somewhat enhanced rate after modification; however, the acceleration of this by heparin was significantly less. Kinetic determination of the number of residues modified gave a reaction order of 2.0, while reaction with N-succinimidyl[3H]propionate yielded labeled factor Xa containing 1.0 mol N-succinimidylpropionate/mol factor Xa and 50% normal clotting activity, or 2.0 mol N-succinimidylpropionate/mol and 1% activity, respectively. Thus, one nucleophilic group is required for the reaction of factor Xa with prothrombin but not for the hydrolysis of peptides or recognition of antithrombin III. The decay of clotting activity of the factor X zymogen in N-succinimidylpropionate was much slower though still Ca2+-dependent. Conversely, the reaction of a related compound--N-succinimidyl(4-hydroxyphenyl)propionate or Bolton-Hunter reagent--with factor Xa broadly resembled that of N-succinimidylpropionate but the decay curves indicated more complex kinetics. Therefore, the target groups vary in their accessibility to modification according to the structural characteristics of both the protein and the reagent.

Gabexate mesylate inhibition of serine proteases: thermodynamic and computer-graphics analysis

The inhibitory effect of gabexate mesylate, which is used therapeutically in the treatment of pancreatitis and disseminated intravascular coagulation, and as a regional anticoagulant agent for hemodialysis, has been measured on bovine factor Xa, bovine alpha-thrombin, human Lys77-plasmin, human urinary kallikrein, human urokinase, porcine pancreatic beta-kallikrein-B, and bovine beta-trypsin catalyzed hydrolysis of p-nitrophenyl esters of N-alpha-carbobenzoxy-L-arginine and N-alpha-carbobenzoxy-L-lysine. On the basis of enzyme:gabexate mesylate affinities, the serine proteases can be arranged as follows: human urinary kallikrein approximately porcine pancreatic beta-kallikrein-B much less than bovine beta-trypsin approximately bovine factor Xa approximately human Lys77-plasmin approximately human urokinase approximately bovine alpha-thrombin. The mode of binding of gabexate mesylate to the serine proteases conforms to the active-reactive site geometries observed in their complexes with natural and synthetic inhibitors. Differences in gabexate mesylate affinities for these proteases reflect structural differences at their primary specificity subsite, which have been investigated by comparative analysis of amino acid sequences and by computer-graphics techniques.

An ordered addition, essential activation model of the tissue factor pathway of coagulation: evidence for a conformational cage

One way in which coagulation may be initiated is by the action of factor VIIa (a plasma serine protease) and tissue factor (a membrane-bound lipid-dependent glycoprotein). We show that in the absence of either factor VIIa or tissue factor, the activation of the natural coagulation substrates, factors IX and X, is not detectable; i.e., tissue factor is an essential activator. We propose that the reaction is fully ordered; that is, the enzyme-activator complex picks up substrate to form a ternary product forming species. Our model precludes the formation of enzyme-substrate and activator-substrate complexes. We have derived equations for the two possible variations of this model: one in which product formation is accompanied by the release of the enzyme-activator complex and the other in which product, free enzyme, and free activator are formed with each catalytic cycle. Our data support only the former which is consistent with both steady-state and rapid equilibrium assumptions. The model is supported by experiments using a monoclonal anti-tissue factor antibody, which affects only the Km app, and a modified form of factor VIIa, which, depending on the sequence in which reagents are added to the reaction, either decreases the Vmax or increases the Km app. We present equations describing the initial velocity of these reactions. Utilizing dilution-jump experiments, we show that the system is hysteretic and suggest that this phenomenon is due to a slow release of enzyme from activator. However, the kinetically determined dissociation constant of enzyme and activator, previously found to be 4.5 nM under equilibrium conditions, was estimated to be 0.04-0.09 nM. Accordingly, we examined other essential activation models in which the product-forming species consists of a complex of enzyme, activator, and substrate at a molar ratio of 1:1:1; none could account for the apparent tight binding of enzyme and activator. We therefore postulate an ordered addition, essential activation model in which the enzyme undergoes two conformational transformations: one as a consequence of binding to tissue factor, resulting in a species which binds to and hydrolyzes its natural substrates. The other conformational change in the enzyme is induced by substrate, resulting in a species which binds more tightly to its activator. Thus, we hypothesize a "conformational cage" which precludes the dissociation of enzyme from activator while significant concentrations of substrate are present.

Factor VII binding to tissue factor in reconstituted phospholipid vesicles: induction of cooperativity by phosphatidylserine

The binding of factor VII and tissue factor produces a membrane-associated proteolytic complex which may be the primary biological initiator of coagulation. Homogeneous tissue factor, a glycoprotein purified from bovine brain, was reconstituted into phospholipid vesicles ranging from neutral (100% phosphatidylcholine) to highly charged (40% phosphatidylserine) with octyl glucoside. The vesicles were characterized with respect to size and tissue factor content and orientation. Employing data from protease digestion, we deduced that tissue factor is randomly oriented; thus, its effective concentration in these vesicles was half its total concentration. In all binding experiments, 1 mol of enzyme was bound per mole of available activator at saturation. This stoichiometry was not affected by the form of the enzyme employed or the phospholipid composition of the vesicles. With tissue factor incorporated into phosphatidylcholine vesicles, the Kd was 13.2 +/- 0.72 nM for factor VII and 4.54 +/- 1.37 nM for factor VIIa. Thus, the one-chain zymogen binds to the activator with only slightly less affinity than the more active two-chain enzyme. Active-site modification of factor VII and factor VIIa with diisopropyl fluorophosphate resulted in tighter binding of the derivatized molecules. Inclusion of phosphatidylserine in the vesicles altered the binding both quantitatively and qualitatively. With increasing acidic phospholipid, the concentration of enzyme required to occupy half the activator sites was decreased. In addition, positive cooperativity was observed, the degree of which depended on the vesicle charge and the form of the enzyme. An explicit two-site cooperative binding model is presented which fits these complex data. In this model, tissue factor is at least a dimer with two interacting enzyme binding sites.

Membrane-dependent coagulation reaction is independent of the concentration of phospholipid-bound substrate: fluid phase factor X regulates the extrinsic system

Negatively charged phospholipids accelerate blood coagulation; one suggestion is that the binding of the gamma-carboxyglutamic acid-containing zymogens to these lipid surfaces increases their effective concentration as substrates. Alternatively, the charged phospholipids could enhance the direct interaction of substrate with the catalytic complex, which is localized at the membrane surface. We distinguished the alternatives by using prothrombin fragment 1 to compete with the substrate for membrane binding sites without interfering with the direct enzyme-substrate interaction. In a tissue factor-factor VIIa system containing neutral phospholipids (to which the substrate does not bind), prothrombin fragment 1 had no significant effect on factor X activation (Km, 877 +/- 111 nM and 791 +/- 103 nM, with and without prothrombin fragment 1, respectively). In contrast, in a system containing 30% phosphatidylserine, prothrombin fragment 1 displaced phospholipid-bound factor X, increasing the free factor X concentration and the reaction velocity in all 22 determinations. As the velocity increases correlated only with the free factor X concentration, we conclude that the free factor X concentration controls this reaction velocity. The Km in the 30% phosphatidylserine system, calculated using free factor X concentration, was 41 +/- 5 nM and 63 +/- 9 nM, with and without prothrombin fragment 1, respectively. Thus, the negatively charged lipids decreased the intrinsic Km by over 90%. The methodology employed should be applicable to ligand-receptor systems in which ligand binds nonspecifically to the membrane surface.

Comparison of the abilities of synthetic and platelet-derived membranes to enhance thrombin formation

The relative abilities of platelet-derived membranes and synthetic phospholipid vesicles to enhance the prothrombinase-catalyzed conversion of prothrombin to thrombin have been determined. For each type of membrane, the maximum amount of thrombin formed as a function of amount of available lipid was measured using a chromogenic substrate assay. The lipid concentration at which the amount of thrombin formed began to exceed that formed in the absence of lipid (critical phospholipid concentration) was used to compare the surfaces' abilities to support thrombin formation. For platelet-derived membranes and for equimolar, charged-lipid/phosphatidylcholine (PC) vesicles, the critical concentrations increased in the following order: platelet-derived membranes approximately equal to phosphatidylserine (PS) approximately equal to phosphatidic acid (PA) less than monomethyl PA and monoethyl PA much less than phosphatidylinositol and phosphatidylglycerol. For mixed anionic/neutral lipid vesicles above their phase transitions, measured critical concentrations were relatively insensitive to changes in lipid acyl chains, the neutral lipid component, and membrane curvature but were sensitive to changes in the anionic lipid content of the mixtures. Comparison of these data suggested that equimolar PS/PC and PA/PC vesicles can emulate reasonably well the thrombin-generating ability of platelet-derived membranes.

Tissue factor gene localized to human chromosome 1 (1pter----1p21)

Tissue factor (tissue thromboplastin, coagulation factor III), a protein component of cell membranes, is an essential cofactor for factor VII-dependent initiation of blood coagulation. Since no tissue factor-deficient condition has been described, it is one of only a few proteins of the coagulation system for which the pattern of inheritance has not been ascertained. Because of the species-specificity of tissue factor activity and the availability of a very sensitive chromogenic assay, it was possible in the present study to use somatic cell hybrids to assign the chromosomal location of the tissue factor structural gene (F3) to human chromosome 1 (1pter----1p21).

Transition-state affinity chromatography of trypsin-like proteinases with dipeptidyl argininal ligands

Dipeptidyl argininal (arginine aldehyde) affinity resins of general formula R-(X-Y-argininal) (where R = resin matrix and X, Y = amino acids of varied structure) are synthesized in a solid-phase procedure in which the dipeptide (-X-Y-) is first attached to the resin, followed by the joining of the Y amino acid to argininal semicarbazone, and decomposition of the semicarbazone in a methanol/acetic acid/formaldehyde reagent. An R-(Gly-Gly-argininal) resin binds urokinase tightly, but does not bind thrombin. However, thrombin binds strongly to R-(Phe-Pro-argininal), whereas urokinase does not bind. Accordingly, the X-Y-argininal ligands selectively bind proteinases of identical primary binding site specificity to arginine, but different secondary site specificity in -X-Y-. The selectivity is due to an amplification of peptide binding specificity caused by the transition-state analog properties of the ligands. While the affinity constants between peptide aldehyde and proteinase approach those of antibody-antigen interactions, the elution with semicarbazide (aldehyde-trapping reagent) buffers easily remove tightly bound proteinases without proteinase inhibitors or denaturation. Conditions for the binding and elution of proteinases, methods of regeneration and other characteristics of the resins are described.

Purification of human vitamin K-dependent protein S and its limited proteolysis by thrombin

Vitamin K-dependent protein S exists in two forms in human plasma, namely as the free protein and in complex with C4b-binding protein [Dahlbäck & Stenflo (1981) Proc. Natl. Acad. Sci. U.S.A. 78, 2512-2516]. Now reported is a simple purification procedure for human protein S that includes barium citrate adsorption, DEAE-Sephacel chromatography and chromatography on Blue Sepharose. The yield was approx. 30% relative to the concentration of free protein S in plasma, which was found to be approx. 10 mg/l. Purified protein S migrated as a single-chain band on sodium dodecyl sulphate/polyacrylamide-gel electrophoresis under non-reducing conditions and as a doublet of Mr approx. 85 000 and 75 000 on reduction. A third band of Mr 16 000 was observed after electrophoresis of 125I-labelled protein S and radioautography of reduced samples. This band appears to be disulphide-linked to the 75 000-Mr chain before reduction. Thrombin converted the 85 000-Mr chain of protein S into a 75 000-Mr chain and an 8000-Mr fragment, the latter again being detectable only by radioautography of reduced samples. The 16 000-Mr fragment was not observed, suggesting its degradation by thrombin. Under non-reducing conditions, no change in apparent molecular weight of thrombin-treated protein S was observed, indicating disulphide linkage of the fragments. Thrombin also affected the mobility of protein S on agarose-gel electrophoresis in the presence of Ca2+, suggesting a decreased affinity to Ca2+ of the cleaved form of protein S as compared with the undegraded molecule. After activation of the complement system in human serum, protein S was found to be a constituent part of the complex formed by C4b-binding protein and component C4b.

Inhibitory effect of activated protein C on activation of prothrombin by platelet-bound factor Xa

Factor Xa binds to a receptor available on the platelet surface after the release reaction. The receptor consists of phospholipid and factor V. Factor Xa bound to the receptor catalyses the activation of prothrombin effectively. The effects of bovine protein C, a vitamin-K-dependent zymogen of a serine protease, on prothrombin activation by platelet-bound bovine factor Xa has been studied. Protein C was found to be activated (protein Ca) by thrombin formed in the prothrombin-platelet-factor Xa incubation. Protein Ca in contrast to the zymogen, protein C, or protein Ca inactivated with diisopropylphosphofluoridate inhibited prothrombin activation by factor Xa in the presence of platelets. protein Ca was found to destroy the receptor by proteolysis whereas direct binding of protein Ca to the receptor could not be demonstrated. The inhibition by protein Ca could be monitored as a parallel decrease in factor Xa binding and prothrombin activation. The receptor was protected by factor Xa from proteolysis by protein Ca. Protein Ca was also found to inhibit the interaction between prothrombin and the factor Xa platelet receptor. These results indicate that protein C after activation may have a role as a regulator of prothrombin activation in vivo.

Binding of bovine coagulation factor Xa to platelets

The binding of highly purified bovine coagulation factor Xa to washed bovine platelets was studied. 125I-labeled factor Xa underwent binding to a platelet receptor that became accessible only after induction of the platelet release reaction by thrombin or by the calcium ionophore A 23187. The zymogen factor X did not bind to platelets. The factor Xa binding was saturable, reversible, and correlated with the rate of thrombin formation. The number of factor Xa binding sites per platelet was 290--420 and the apparent association constant was estimated to be 2.8 x 109 to 1.0 x 1010 M-1. Diisoprophyl fluorophosphate-factor Xa bound to the same platelet receptor as factor Xa indicating that limited proteolysis of a receptor protein was not required for binding. The rate of factor Xa binding was rapid (2.1 x 10(6) to 2.9 x 10(6) M-1 s-1) and similar to that preveiously found for the rate of binding of polypeptide hormones to their receptors. Displacement of factor Xa from the platelet receptor by diisopropyl fluorophosphate-factor Xa effectively blocked thrombin formation. Low concentrations of factor Xa catalyze prothrombin activation more effectively in the presence of platelets than in the presence of phospholipid and factor V.