A player of many parts: the spotlight falls on thrombin's structure

Structural examination of the influence of phosphorylation on the binding of fibrinopeptide A to bovine thrombin

Upon addition of thrombin, fibrinopeptides A and B are cleaved off from the N-termini of four chains of fibrinogen (Aalpha Bbeta gamma)2, and sites of polymerization are exposed, resulting in formation of a fibrin clot. For the fibrinogen Aalpha chain, cleavage occurs most prevalently at the Arg16-Gly17 peptide bond. About 25-30% of the human fibrinogen Aalpha chains are phosphorylated in nature at the position of Ser3, but the function for this modification is not understood. Previous NMR studies indicated that the N-terminal portion (1ADSGE5) of unphosphorylated fibrinopeptide A does not interact with the surface of bovine thrombin. Kinetic and NMR studies have now been carried out to assess whether phosphorylation at Ser3 allows the N-terminal segment (1ADSGEGDFLAEGGGVR16) to become anchored on the thrombin surface, leading to formation of a catalytically more efficient enzyme-substrate complex. Kinetic results indicate that phosphorylation leads to an approximately 65% increase in substrate specificity (kcat/Km) toward hydrolysis of fibrinogen Aalpha(1-20). 31P NMR studies reveal that the phosphorylated group does interact with thrombin, and 1H line broadening studies suggest that phosphorylation does promote binding of amino acids 1-5. Two-dimensional transferred nuclear Overhauser effect spectroscopy studies of bound fibrinopeptide A(1-16 Ser3P) indicate that phosphorylation allows new through-space interactions involving amino acid residues 1ADSGE5 to be observed. Computational docking of the peptide onto the X-ray structure of thrombin suggests that the phosphate may interact with basic residues at the rim of the heparin binding site of thrombin. As a result, the phosphate may serve as an anionic linker between the fibrinopeptide and the enzyme thrombin.

Differential transcriptional regulation of the human thrombin receptor gene by the Sp family of transcription factors in human endothelial cells

The mitogenic effects of thrombin are mediated by a G-protein-coupled receptor. Because the effects of thrombin are strongly influenced by the expression of its receptor, an understanding of its regulatory mechanisms is essential. To identify mechanisms of human thrombin receptor (HTR) gene regulation, a series of HTR-promoter-luciferase constructs were made and transfected into human microvascular endothelial cells for analysis. Deletion from bp -303 to -164 abolished reporter gene expression. Dimethyl sulphate treatment in vivo and DNase I footprinting in vitro demonstrated that a cluster of three GC box consensus sites was occupied, and electrophoretic mobility-shift assays established that Sp1 and Sp3 both bind to this 3' GC box cluster. We mutated each of the three GC boxes individually and all three collectively within this 3' cluster. Basal promoter activity was decreased to 46%, 78% and 29% of control for each of the GC boxes mutated individually, and to 6% when the three were mutated collectively. To test the individual abilities of Sp1 and Sp3 to activate or repress HTR transcription, we conducted co-transfection experiments with wild-type or mutated HTR-promoter-luciferase constructs. Co-transfection with Sp1 significantly augmented wild-type HTR promoter activity. Sp3 alone did not affect activity, and inhibited Sp1-mediated activation. Competition for shared binding sites by Sp1 and Sp3 might differentially regulate HTR expression in vascular endothelial cells.

Identification and initial structure-activity relationships of a novel class of nonpeptide inhibitors of blood coagulation factor Xa

The discovery and some of the basic structure-activity relationships of a series of novel nonpeptide inhibitors of blood coagulation Factor Xa is described. These inhibitors are functionalized beta-alanines, exemplified by 2a. Docking experiments placing 2a in the active site of Factor Xa implied that the most expeditious route to enhancing in vitro potency was to modify the group occupying the S3 site of the enzyme. Increasing the hydrophobic contacts between the inhibitor and the enzyme in this region led to 8, which has served as the prototype for this series. In addition, an enantioselective synthesis of these substituted beta-alanines was also developed.

Hirunorms are true hirudin mimetics. The crystal structure of human alpha-thrombin-hirunorm V complex

A novel class of synthetic, multisite-directed thrombin inhibitors, known as hirunorms, has been described recently. These compounds were designed to mimic the binding mode of hirudin, and they have been proven to be very strong and selective thrombin inhibitors. Here we report the crystal structure of the complex formed by human alpha-thrombin and hirunorm V, a 26-residue polypeptide containing non-natural amino acids, determined at 2.1 A resolution and refined to an R-factor of 0.176. The structure reveals that the inhibitor binding mode is distinctive of a true hirudin mimetic, and it highlights the molecular basis of the high inhibitory potency (Ki is in the picomolar range) and the strong selectivity of hirunorm V. Hirunorm V interacts through the N-terminal tetrapeptide with the thrombin active site in a nonsubstrate mode; at the same time, this inhibitor specifically binds through the C-terminal segment to the fibrinogen recognition exosite. The backbone of the N-terminal tetrapeptide Chg1"-Val2"-2-Nal3"-Thr4" (Chg, cyclohexyl-glycine; 2-Nal, beta-(2-naphthyl)-alanine) forms a short beta-strand parallel to thrombin main-chain residues Ser214-Gly219. The Chg1" side chain fills the S2 subsite, Val2" is located at the entrance of S1, whereas 2-Nal3" side chain occupies the aryl-binding site. Such backbone orientation is very close to that observed for the N-terminal residues of hirudin, and it is similar to that of the synthetic retro-binding peptide BMS-183507, but it is opposite to the proposed binding mode of fibrinogen and of small synthetic substrates. Hirunorm V C-terminal segment binds to the fibrinogen recognition exosite, similarly to what observed for hirudin C-termninal tail and related compounds. The linker polypeptide segment connecting hirunorm V N-and C-terminal regions is not observable in the electron density maps. The crystallographic analysis proves the correctness of the design and it provides a compelling proof on the interaction mechanism for this novel class of high potency multisite-directed synthetic thrombin inhibitors.

Reactivities of the S2 and S3 subsite residues of thrombin with the native and heparin-induced conformers of antithrombin

A pentasaccharide (PS) fragment of heparin capable of activating antithrombin (AT) markedly accelerates the inhibition of factor Xa by AT, but has insignificant effect on inhibition of thrombin. For inhibition of thrombin, the bridging function of a longer polysaccharide chain is required to accelerate the reaction. To study the basis for the similar reactivity of thrombin with the native or heparin-activated conformers of AT, several residues surrounding the active site pocket of thrombin were targeted for mutagenesis study. Leu99 and Glu192, the variant residues influencing the S2 and S3 subsite specificity of thrombin were replaced with Tyr and Gln. The Tyr60a, Pro60b, Pro60c, and Trp60d residues forming part of the S2 specificity pocket were deleted from the B-insertion loop of the wild-type and Leu99/Glu192 --> Tyr/Gln thrombins. Kinetic studies indicated that the reactivities of all mutants with AT were moderately or severely impaired. Although heparin largely corrected the defect in reactivities, it also markedly elevated the stoichiometries of inhibition with the mutants. Interestingly, PS also accelerated AT inhibition of the mutants 5-68-fold, suggesting that the mutants are able to discriminate between the native and activated conformers of AT. Based on these results and the recent crystal structure determination of AT in complex with PS, a model for thrombin-AT interaction is proposed in which the S2 and S3 subsite residues of thrombin are critical for recognition of the P2 and P3 residues of AT in the native conformation. In the activated conformation, other residues are made accessible for interaction with the protease, and the similar reactivity of thrombin with the native and heparin-activated conformers of AT may be coincidental. The results further suggest that the S2 and S3 subsite residues are crucial in controlling the partitioning of the thrombin-AT intermediate into the alternative inhibitory or substrate pathways of the reaction.

Thrombomodulin increases the rate of thrombin inhibition by BPTI

Thrombin undergoes allosteric modulation by thrombomodulin (TM) that results in a shift in macromolecular specificity, blocking fibrinogen clotting while enhancing protein C activation. The TM enhancement of protein C activation involves both an 8-fold decrease in Km and a 200-fold increase in kcat. Although TM-mediated conformational changes in thrombin have been detected by many techniques, the nature of these changes remains obscure. Access to the active center of thrombin is relatively restricted due to the presence of a large insertion loop at residue 60 (chymotrypsin numbering) that has been implicated in modeling studies as being responsible for poor inhibition by BPTI. Thrombin and the E192Q mutant, which binds BPTI much more tightly than thrombin, are both inhibited very slowly by BPTI. TM increases the rate of thrombin or thrombin E192Q inhibition by BPTI approximately 10-fold. When analyzed as slow tight binding inhibition, the TM effect on thrombin E192Q inhibition by BPTI is primarily on the first, reversible step in the reaction. Structural studies of the thrombin E192Q-BPTI complex have previously shown that the 60 loop lies over the BPTI, a position which requires 8 A movement at the apex of the 60 loop, and that BPTI is found in the same canonical orientation as in the trypsin complex. It follows that TM enhancement of the initial interaction of thrombin results in a conformation that favors interactions with BPTI, probably involving motion of the 60 loop.

Characterization of the reciprocal binding sites on human alpha-thrombin and factor XIII A-chain

Solution- and solid-phase techniques were used to probe Factor XIII A-chain-alpha-thrombin interactions. Alpha-thrombin activated Factor XIII more efficiently (Km = 0.83 +/- 0.08 x 10(-7) M; V/K = 14.90 +/- 3.20 x 10(-3) min(-1)) than beta-thrombin (Km = 6.14 +/- 1.26 x 10(-7) M; V/K = 3.30 +/- 1.00 x 10(-3) min(-1)) or gamma-thrombin (Km = 6.25 +/- 1.15 x 10(-7) M; V/K = 3.00 +/- 0.80 x 10(-3) min(-1)). Immobilized FPR-alpha-thrombin bound plasma Factor XIII (Kd = 0.17 +/- 0.04 x 10(-7) M) > Factor XIIIa (Kd = 0.69 +/- 0.18 x 10(-7) M) > liver transglutaminase (Kd = 4.73 +/- 1.01 x 10(-7) M) > Factor XIII A-chain (Kd = 49.00 +/- 9.40 x 10(-7) M). FPR-alpha-thrombin and alpha-thrombin also bound immobilized Factor XIII A-chain with affinities inversely related to protease activity: maximal binding at 1.36 x 10(-7) M and 13.6 x 10(-7) M, respectively. Plasma Factor XIII, transglutaminase, and dithiothreitol competitively inhibited Factor XIII A-chain binding to FPR-alpha-thrombin: IC50 = 1.0 x 10(-7) M, 3.0 x 10(-6) M and 1.52 x 10(-4) M, respectively. Transglutaminase also inhibited Factor XIII binding to alpha-thrombin (IC50 = 2.0 x 10(-6) M). Thrombin-binding site was localized to G38-M731 fragment of Factor XIII A-chain, probably within homologous regions (N72-A493) of transglutaminase. R320-E579 of alpha-thrombin was Factor XIII A-chain binding site. Intra-B-chain disulfides in alpha-thrombin were essential for binding but not catalytic H363 or residues R382-N394 and R443-G475. These studies propose a structural basis for Factor XIII activation, provide a regulatory mechanism for Factor XIIIa generation, and could eventually help in the development of new structure-based inhibitors of thrombin and Factor XIIIa.

New insights into the regulation of the blood clotting cascade derived from the X-ray crystal structure of bovine meizothrombin des F1 in complex with PPACK

BACKGROUND

The conversion of prothrombin to thrombin by factor Xa is the penultimate step in the blood clotting cascade. In vivo, where the conversion occurs primarily on activated platelets in association with factor Va and Ca2+ ions, meizothrombin is the major intermediate of the two step reaction. Meizothrombin rapidly loses the fragment 1 domain (F1) by autolysis to become meizothrombin des F1 (mzTBN-F1). The physiological properties of mzTBN-F1 differ dramatically from those of thrombin due to the presence of prothrombin fragment 2 (F2), which remains covalently attached to the activated thrombin domain in mzTBN-F1.

RESULTS

The crystal structure of mzTBN-F1 has been determined at 3.1 A resolution by molecular replacement, using only the thrombin domain, and refined to R and Rfree values of 0.205 and 0.242, respectively. The protease active site was inhibited with D-Phe-Pro-Arg-chloromethylketone (PPACK) to reduce autolysis. The mobile linker chain connecting the so-called kringle and thrombin domains and the first two N-acetylglucosamine residues attached to the latter were seen in electron-density maps improved with the program SQUASH. Previously these regions had only been modeled.

CONCLUSIONS

The F2 kringle domain in mzTBN-F1 is bound to the electropositive heparin-binding site on thrombin in an orientation that is systematically shifted and has significantly more interdomain contacts compared to a noncovalent complex of free F2 and free thrombin. F2 in mzTBN-F1 forms novel hydrogen bonds to the carbohydrate chain of thrombin and perhaps stabilizes a unique, rigid conformation of the gamma-autolysis loop through non-local effects. The F2 linker chain, which does not interfere with the active site or fibrinogen-recognition site, is arranged so that the two sites cleaved by factor Xa are separated by 36 A. The two mzTBN-F1 molecules in the asymmetric unit share a tight 'dimer' contact in which the active site of one molecule is partially blocked by the F2 kringle domain of its partner. This interaction suggests a new model for prothrombin organization.

Proteolysis of Alzheimer's disease beta-amyloid precursor protein by factor Xa

Amyloid beta-protein is a 4-kDa peptide which originates from proteolysis of a larger protein precursor (APP) and accumulates in senile plaques in brains of Alzheimer's disease (AD) patients. Since secreted APP inhibits factors IXa, Xa and XIa, and thrombin appears to play a role in APP secretion and proteolysis, a relationship between hemostasis system and APP metabolism seems to exist. In this work we investigate the susceptibility to proteolytic cleavage by factor Xa of a fusion construct containing full-length APP prepared in bacteria, and demonstrate that both APP695 and APP770 are substrates for this protease. Factor Xa was found to cleave APP after arginines 102, 268, 510, 573 and 601 (APP695 numeration); most of these sites appear to be common for different coagulation factors. In addition, APP incubation with factor Xa generates an array of six potentially amyloidogenic fragments. Comparative kinetic analysis of APP695 and APP770 cleavage by factor Xa suggests that Kunitz-type inhibitor-containing isoforms exert an inhibitory effect on the protease. However, this inhibition is far from complete even at a 5-fold molar excess of inhibitor. Our results raise the possibility that proteases from the coagulation cascade may contribute to APP proteolysis, and support the notion that these proteases play a role in AD pathogenesis.

Networking in the hemostatic system. Integrin alphaiibbeta3 binds prothrombin and influences its activation

Prothrombin activation is a pivotal event in thrombosis and hemostasis because thrombin can mediate fibrin formation and can activate and aggregate platelets. Platelet aggregation depends upon the binding of adhesive proteins to integrin alphaIIbbeta3 on the platelet surface. In the present study, a novel interface between the blood coagulation system and platelets is demonstrated by showing that 1) prothrombin binds to alphaIIbbeta3 and 2) this interaction accelerates prothrombin activation. Prothrombin bound to purified alphaIIbbeta3 in a specific, saturable, and divalent cation-dependent manner. This interaction was inhibited by certain monoclonal antibodies to alphaIIbbeta3, by the alphaIIbbeta3 ligands fibrinogen and RGD peptides, but not by thrombin or unrelated proteins. Prothrombin also interacted with alphaIIbbeta3 on resting and stimulated platelets as demonstrated by soluble ligand binding and platelet adhesion assays. Activation of prothrombin by Factor Xa alone or Factor Xa-Va was accelerated by alphaIIbbeta3, and this enhancement was blocked by a monoclonal antibody that inhibited prothrombin binding to the receptor. Taken together, these data identify a previously unrecognized linkage between platelets and the blood coagulation system that may have a significant regulatory consequence.

Evidence for allosteric linkage between exosites 1 and 2 of thrombin

Investigations to date have demonstrated that ligand binding to exosites 1 or 2 on thrombin produces conformational changes at the active site. In this study, we directly compared the effect of ligand binding to exosites 1 and 2 on the structure and function of the active site of thrombin and investigated functional linkage between the two exosites. Binding studies were performed in solution with fluorescein-Phe-Pro-Arg-CH2Cl (FPR)-thrombin. Hirudin-(54-65) and sF2, a synthetic peptide corresponding to residues 63-116 of prothrombin fragment 2, were used as ligands for exosites 1 and 2 of thrombin, respectively. The two ligands produce diametric changes in the fluorescence of fluorescein-FPR-thrombin and also have opposing effects on the rate of thrombin hydrolysis of a number of chromogenic substrates. These results indicate that sF2 and hirudin-(54-65) differentially affect the conformation of the active site. Experiments then were performed to investigate whether both ligands can bind to thrombin simultaneously. When thrombin-bound fluorescein-sF2 is titrated with hirudin-(54-65), complete displacement of fluorescein-sF2 is observed. Likewise, when thrombin-bound fluorescein-hirudin-(54-65) is titrated with sF2, complete displacement occurs. Additional support for reciprocal binding was obtained in fluorescence experiments where both probes were labeled and in experiments monitoring ligand binding to agarose-immobilized thrombin. This mutually exclusive binding of either ligand can be explained by reciprocal, allosteric modulation of ligand affinity between the two exosites. Thus, not only do the two exosites differentially influence the active site, they also affect the binding properties of the opposing exosite.

The role of thrombin-like (serine) proteases in the development, plasticity and pathology of the nervous system

There is increasing evidence suggesting that members of the serine protease family, including thrombin, chymotrypsin, urokinase plasminogen activator, and kallikrein, may play a role in normal development and/or pathology of the nervous system. Serine proteases and their cognate inhibitors have been shown to be increased in the neural parenchyma and cerebrospinal fluid following injury to the blood brain barrier. Zymogen precursors of thrombin and thrombin-like proteases as well as their receptors have also been localized in several distinct regions of the developing or adult brain. Thrombin-like proteases have been shown to exert deleterious effects, including neurite retraction and death, on different neuronal and non-neuronal cell populations in vitro. These effects appear to be mediated through cell surface receptors and can be prevented or reversed with specific serine protease inhibitors (serpins). Furthermore, we have recently shown that treatment with protease nexin-1 (a serpin that inhibits thrombin-like proteases) promotes the survival and growth of spinal motoneurons during the period of programmed cell death and following injury. Taken together, these observations suggest that thrombin-like proteases play a deleterious role, whereas serpins promote the development and maintenance of neuronal cells. Thus, changes in the balance between serine proteases and their cognate inhibitors may lead to pathological states similar to those associated with some neurodegenerative diseases such as Alzheimer's disease. The present review summarizes the current state of research involving such serine proteases and speculates on the possible role of these thrombin-like proteases in the development, plasticity and pathology of the nervous system.

Crystal structure of fibrinogen-Aalpha peptide 1-23 (F8Y) bound to bovine thrombin explains why the mutation of Phe-8 to tyrosine strongly inhibits normal cleavage at Arg-16

A peptide containing residues 1-50 of the Aalpha-chain of fibrinogen, expressed as a fusion peptide with beta-galactosidase, is rapidly cleaved by thrombin at Arg-16, similarly to whole fibrinogen. When Phe-8, which is highly conserved, is replaced with tyrosine (F8Y), the cleavage is slowed drastically [Lord, Byrd, Hede, Wei and Colby (1990) J. Biol. Chem. 265, 838-843]. To examine the structural basis for this result, we have determined the crystal structure of bovine thrombin complexed with a synthetic peptide containing residues 1-23 of fibrinogen Aalpha and the F8Y mutation. The crystals are in space group P43212, with unit-cell dimensions of a = 88.3 A (1 A = 0.1 nm), c = 195.5 A and two complexes in the asymmetric unit. The final R factor is 0.183 for 2sigma data from 7.0 to 2.5 A resolution. There is continuous density for the five residues in the P3, P2, P1, P1' and P2' positions of the peptide (Gly-14f to Pro-18f) at the active site of thrombin, and isolated but well-defined density for Tyr-8f at position P9 in the hydrophobic pocket of thrombin. The tyrosine residue is shifted relative to phenylalanine in the native peptide because the phenol side chain is larger and makes a novel, intrapeptide hydrogen bond with Gly-14f. Adjacent peptide residues cannot form the hydrogen bonds that stabilize the secondary structure of the native peptide. Consequently, the 'reaction'geometry at the scissile bond, eight residues from the mutation, is perturbed and the peptide is mostly uncleaved in the crystal structure.

Molecular cloning and expression of bothrojaracin, a potent thrombin inhibitor from snake venom

Bothrojaracin is a potent and selective thrombin inhibitor that has been isolated from the venom of Bothrops jararaca. It does not interact with the catalytic site of the enzyme but binds to both anion-binding exosites 1 and 2 resulting in a potent inhibition of thrombin activity towards fibrinogen and platelets [Zingali, R. B., Jandrot-Perrus, M., Guillin, M. C. & Bon, C. (1993) Biochemistry 32, 10794-108021. Bothrojaracin is a 27-kDa protein composed of two disulfide-linked polypeptide chains, A and B, of 15 kDa and 13 kDa, respectively. The sequences of A and B chains determined by molecular cloning exhibit a high degree of identity with other snake venom lectin-like proteins. In contrast to other ligands that interact with thrombin exosite 1, the amino acid sequence of bothrojaracin does not contain an acidic sequence similar to the C-terminal tail of hirudin. Expression of functional bothrojaracin was achieved in COS cells upon transfection with two pcDNA3 vectors containing the complete cDNAs. Recombinant bothrojaracin, which was secreted into the medium, was able to bind to and inhibit thrombin. When expressed alone, the B chain formed inactive dimers that were secreted into the culture medium. In contrast, no bothrojaracin-related protein was detected in conditioned media from cells transfected with the A chain.

Structure-based design of a potent chimeric thrombin inhibitor

Using the three-dimensional structures of thrombin and the leech-derived tryptase inhibitor (LDTI), which does not inhibit thrombin, we were able to construct three LDTI variants inhibiting thrombin. Trimming of the inhibitor reactive site loop to fit thrombin's narrow active site cleft resulted in inhibition constants (Ki) in the 10 nM concentration range; similar values were obtained by the addition of an acidic C-terminal peptide corresponding to hirudin's tail to LDTI. Combination of both modifications is additive, resulting in very strong inhibition of thrombin (Ki in the picomolar range). On the one hand, these results confirm the significance of the restricted active site cleft of thrombin in determining its high cleavage specificity; on the other, they demonstrate that sufficient binding energy at the fibrinogen recognition exosite can force thrombin to accept otherwise unfavorable residues in the active site cleft. The best inhibitor thus obtained is as effective as hirudin in plasma-based clotting assays.

Inactivation of thrombin by antithrombin is accompanied by inactivation of regulatory exosite I

Exosite I of the blood clotting proteinase, thrombin, mediates interactions of the enzyme with certain inhibitors, physiological substrates and regulatory proteins. Specific binding of a fluorescein-labeled derivative of the COOH-terminal dodecapeptide of hirudin ([5F] Hir54-65) to exosite I was used to probe changes in the function of the regulatory site accompanying inactivation of thrombin by its physiological serpin inhibitor, antithrombin. Fluorescence-monitored equilibrium binding studies showed that [5F]Hir54-65 and Hir54-65 bound to human alpha-thrombin with dissociation constants of 26 +/- 2 nM and 38 +/- 5 nM, respectively, while the affinity of the peptides for the stable thrombin-antithrombin complex was undetectable (>/=200-fold weaker). Kinetic studies showed that the loss of binding sites for [5F]Hir54-65 occurred with the same time-course as the loss of thrombin catalytic activity. Binding of [5F] Hir54-65 and Hir54-65 to thrombin was correlated quantitatively with partial inhibition of the rate of the thrombin-antithrombin reaction, maximally decreasing the bimolecular rate constants 1.7- and 2.1-fold, respectively. These results support a mechanism in which thrombin and the thrombin-Hir54-65 complex can associate with antithrombin and undergo formation of the covalent thrombin-antithrombin complex at modestly different rates, with inactivation of exosite I leading to dissociation of the peptide occurring subsequent to the rate-limiting inactivation of thrombin. This mechanism may function physiologically in localizing the activity of thrombin by allowing inactivation of thrombin that is bound in exosite I-mediated complexes with regulatory proteins, such as thrombomodulin and fibrin, without prior dissociation of these complexes. Concomitant with inactivation of thrombin, the thrombin-antithrombin complex may be irreversibly released due to exosite I inactivation.

Influence of Arginines 93, 97, and 101 of thrombin to its functional specificity

Mutation of three Arg residues, 93, 97, and 101, to Ala in thrombin (thrombin R93,97,101A) has previously been shown to eliminate most heparin acceleration of thrombin inhibition by antithrombin and most of the ability of chondroitin sulfate (CS) on thrombomodulin (TM) to enhance affinity for TM and to eliminate the characteristic high-affinity interaction with protein C observed with TM lacking CS. In this study we examined the relative impact of mutation of these Arg residues alone and in combination on the above reactions and, in addition, on the ability of rabbit TM to accelerate thrombin inhibition by antithrombin. The order of importance for heparin acceleration of inhibition by antithrombin was Arg 101, 93, and 97. In contrast, Arg 97 was the major residue required for TM-dependent acceleration of reactivity with antithrombin and for CS-dependent enhancement of TM affinity. Arg 101 and 93 were critical for TM-dependent, high-affinity protein C interaction at low Ca2+ concentrations, while Arg 97, which was critical for the other TM-dependent effects, played no detectable role in this metal dependence. These results illustrate that these Arg residues in anion binding exosite 2 contribute very differently to the diverse reactions dependent on that domain in thrombin.

Stable expression and purification of a secreted human recombinant prethrombin-2 and its activation to thrombin

A human prothrombin cDNA has been engineered to obtain a cDNA coding for a secreted form of human prethrombin-2. The secreted prethrombin-2 has been produced in a mammalian expression system using DXB11 cells, a mutant strain of CHO cells in which the dihydrofolate reductase gene has been deleted, and an expression vector carrying the dihydrofolate reductase cDNA. Methotrexate-induced gene amplification favored an efficient production of the recombinant protein which accumulated in the culture medium of the DXB11 cells. Growth in suspension of the stable transformants in an airlift fermenter resulted in the production of 25 mg/L recombinant prethrombin-2. The recombinant protein was purified using single-step affinity chromatography on a recombinant-hirudin column and activated by agarose gel-immobilized ecarin. All purified recombinant prethrombin-2 was activated and the generated recombinant thrombin showed catalytic properties identical to those of plasma-derived alpha-thrombin. This expression system can be used to prepare mutants of prethrombin-2 for structure-function studies investigating thrombin interactions with substrate proteins, inhibitors, and cell membranes.

Human alpha-thrombin inhibition by the highly selective compounds N-ethoxycarbonyl-D-Phe-Pro-alpha-azaLys p-nitrophenyl ester and N-carbobenzoxy-Pro-alpha-azaLys p-nitrophenyl ester: a kinetic, thermodynamic and X-ray crystallographic study

Kinetics, thermodynamics and structural aspects of human alpha-thrombin (thrombin) inhibition by newly synthesized low molecular weight derivatives of alpha-azalysine have been investigated. The thrombin catalyzed hydrolysis of N-ethoxycarbonyl-D-Phe-Pro-alpha-azaLys p-nitrophenyl ester (Eoc-D-Phe-Pro-azaLys-ONp) and N-carbobenzoxy-Pro-alpha-azaLys p-nitrophenyl ester (Cbz-Pro-azaLys-ONp) was investigated at pH 6.2 and 21.0 degrees C, and analyzed in parallel with that of N-alpha-(N,N-dimethylcarbamoyl)-alpha-azalysine p-nitrophenyl ester (Dmc-azaLys-ONp). Decarboxylation following the enzymatic hydrolysis of these p-nitrophenyl esters gave the corresponding 1-peptidyl-2(4-aminobutyl) hydrazines (peptidyl-Abh) showing properties of thrombin competitive inhibitors. Therefore, thermodynamics for the reversible binding of D-Phe-Pro-Abh, Cbz-Pro-Abh and Dmc-Abh to thrombin was examined. These results are consistent with the minimum four-step catalytic mechanism for product inhibition of serine proteinases. Eoc-D-Phe-Pro-azaLys-ONp and Eoc-D-Phe-Pro-Abh display a sub-micromolar affinity for thrombin together with a high selectivity versus homologous plasmatic and pancreatic serine proteinases acting on cationic substrates. The three-dimensional structures of the reversible non-covalent thrombin:Eoc-D-Phe-Pro-Abh and thrombin:Cbz-Pro-Abh complexes have been determined by X-ray crystallography at 2.0 A resolution (R-factor = 0.169 and 0.179, respectively), and analyzed in parallel with that of the thrombin:Dmc-azaLys acyl-enzyme adduct. Both Eoc-D-Phe-Pro-Abh and Cbz-Pro-Abh competitive inhibitors are accommodated in the thrombin active center, spanning the region between the aryl binding site and the S1 primary specificity subsite.

Role of Leu99 of thrombin in determining the P2 specificity of serpins

A recent study indicated that Tyr99 (chymotrypsin numbering) of factor Xa and Thr99 of activated protein C are S2 subsite residues that determine the P2 specificity of their substrates and inhibitors. To investigate the contribution of Leu99 to the P2 binding specificity of thrombin, three mutants of thrombin were prepared in which Leu99 was substituted with Tyr (L99Y), Thr (L99T), or Gly (L99G). Kinetic analysis indicated that antithrombin (AT with P2 Gly) inhibited thrombin L99Y, 14.1- and 5.5-fold slower than thrombin in the absence and presence of heparin, respectively. The L99Y mutation increased the stoichiometry of AT inhibition in the presence of heparin from approximately 1.6 to approximately 4.6, indicating that L99Y recognized AT as a substrate. The inhibition rates of L99T and L99G by AT, respectively, were 500.0- and 916.7-fold slower than thrombin in the absence of heparin but only 41.8- and 64.5-fold slower than thrombin in the presence of heparin. Resolution of the two-step reactions of AT with the mutant thrombins revealed that the impaired reactivities occurred in the second reaction step in which a non-covalent AT-thrombin encounter complex is converted to a stable, covalent complex. In reactions with protein C inhibitor (PCI with P2 Phe), L99Y was inhibited 3.5-fold slower than thrombin, L99T was inhibited at a similar or faster rate, and L99G was inhibited 23.9-fold faster than thrombin. The epidermal growth factor-like domains 4-6 of thrombomodulin (TM4-6) accelerated the PCI inhibition of wild-type and L99G thrombins 73.9- and 5.3-fold, respectively. Further studies indicated that the fibrinogen clotting and protein C activation rates by the mutants were impaired, but the cofactor function of TM was not affected as TM4-6 bound to wild-type [Kd(app) = 5.9 nM] and mutant thrombins with similar affinities [Kd(app) = 4.4-6.9 nM] and enhanced protein C activation rates by all mutants effectively. These results indicate that (1) Leu99 of thrombin is critical for determination of the P2 specificity of serpins, AT and PCI, (2) increasing the polarity of the S2 pocket of thrombin by introduction of a hydrophilic residue into this pocket is detrimental for reaction with AT, but it is tolerated in reaction with PCI, so that only the size of the S2 pocket of thrombin determines the P2 specificity of PCI, and (3) the thrombomodulin-induced conformational change that results in acceleration of thrombin inhibition by PCI involves Leu99.

The second Kunitz domain of human tissue factor pathway inhibitor: cloning, structure determination and interaction with factor Xa

Tissue Factor Pathway Inhibitor (TFPI) is a 36 kDa glycoprotein that helps maintain haemostasis by inhibiting Factor Xa and the Factor VIIa/Tissue Factor (TF) complex. TFPI contains three tandemly linked Kunitz inhibitor domains, of which the second inhibits factor Xa. We have undertaken a multidisciplinary approach to study the structure and function of the second Kunitz domain of TFPI, with a view towards the rational design of factor Xa inhibitors. Amino acid residues 93 to 154 of the mature TFPI protein, corresponding to the second Kunitz domain (TFPI-kII), were expressed in Escherichia coli. The protein was purified to near homogeneity by ion exchange, hydrophobic interaction, and size exclusion chromatography, respectively. TFPI-kII is a potent factor Xa inhibitor with a Ki of 1.5 x 10(-10) M, a value that does not differ significantly from that of intact TFPI. The three-dimensional structure of TFPI-kII in aqueous solution was determined by 1H nuclear magnetic resonance spectroscopy (NMR). A set of 30 conformers was calculated with the program DIANA using 906 distance constraints derived from nuclear Overhauser effects and 23 dihedral angle constraints. This set, representing the solution structure of TFPI-kII, has an average root-mean-square deviation of 0.78 A for the backbone atoms and 1.38 A for all heavy atoms of residues 1 to 58. The structure of TFPI-kII has also been determined in complex with porcine trypsin using X-ray crystallographic techniques. The complex has been solved to a resolution of 2.6 A, with a final R-factor of 16.2%. Comparison of the NMR derived structure with that of TFPI-kII in complex with trypsin reveals little divergence of the two structures, with the exception of residue Tyr17. Superposition of the trypsin:TFPI-kII complex on factor Xa provides insights into macromolecular determinants for the inhibition of factor Xa. Complexation would require a degree of reorganisation of factor Xa residues, in particular of TyrF99, but also perhaps of the F148-loop. The interaction was further investigated using restrained molecular dynamics. Electrostatic interactions would appear to play a major role. The reorganisation of factor Xa is in contrast to the proposed factor Xa:TAP interaction, where TAP would bind to the "ground state" structure of factor Xa.

The thrombin E192Q-BPTI complex reveals gross structural rearrangements: implications for the interaction with antithrombin and thrombomodulin

Previous crystal structures of thrombin indicate that the 60-insertion loop is a rigid moiety that partially occludes the active site, suggesting that this structural feature plays a decisive role in restricting thrombin's specificity. This restricted specificity is typified by the experimental observation that thrombin is not inhibited by micromolar concentrations of basic pancreatic trypsin inhibitor (BPTI). Surprisingly, a single atom mutation in thrombin (E192Q) results in a 10(-8) M affinity for BPTI. The crystal structure of human thrombin mutant E192Q has been solved in complex with BPTI at 2.3 A resolution. Binding of the Kunitz inhibitor is accompanied by gross structural rearrangements in thrombin. In particular, thrombin's 60-loop is found in a significantly different conformation. Concomitant reorganization of other surface loops that surround the active site, i.e. the 37-loop, the 148-loop and the 99-loop, is observed. Thrombin can therefore undergo major structural reorganization upon strong ligand binding. Implications for the interaction of thrombin with antithrombin and thrombomodulin are discussed.

Bradykinin antagonizes the effects of alpha-thrombin

alpha-Thrombin (AT) and bradykinin (BK) are endogenous mediators that are released during an inflammatory response, and could have a synergistic effect on endothelial permeability. Human umbilical vein endothelial cells (HUVEC) were grown on Transwell membranes and then tested for alterations in permeability to fluorescein isothiocyanate-labeled human serum albumin. Addition of 1 microM AT produced a significant increase in the permeability coefficient at 30 minutes from control levels of 1.59 x 10(-6) cm/sec to 4.92 x 10(-6) cm/sec. BK (1 microM) produced a similar increase to 4.46 x 10(-6) cm/sec. For both compounds, permeability remained elevated for 90 minutes. Pre-treatment of the HUVEC with the bradykinin receptor antagonist, Na-adamantaneacetyl-bradykinin (NA-BK) (1 microM), prior to addition of AT, reduced the AT permeability coefficient to 2.69 x 10(-6) cm/sec. Addition of NA-BK (1 microM) for 5 minutes, then BK (1 microM) for 5 minutes, inhibited the effect of BK and of AT (1 microM on permeability, decreasing the permeability coefficient of the endothelial monolayer to control levels (1.62 x 10(-6) cm/sec). AT (1 microM) increased HUVEC intracellular calcium mobilization, as monitored by FURA-2, to 245 nM from control (70 nM), however, pre-treatment with either BK or the bradykinin receptor antagonist decreased the AT induced intracellular calcium mobilization compared to AT alone. Pre-treatment of the HUVEC with bradykinin (1 microM) for 2 minutes also inhibited the effects of alpha-thrombin (1 microM) on f-actin distribution examined by BODIPY-phallodin staining and increased the clotting times for an alpha-thrombin dependent fibrinogen to fibrin clotting assay. However, incubation of bradykinin (1 microM) with alpha-thrombin (1 microM) for either 10 minutes or 100 minutes produced no detectable hydrolysis products. These data strongly suggest that the inflammatory mediators alpha-thrombin and bradykinin when released together, rather than being synergistic, are antagonistic.

Thrombin interaction with fibrin polymerization sites

Thrombin is central to hemostasis, and postclotting fibrinolysis and wound healing. During clotting, thrombin transforms plasma fibrinogen into polymerizing fibrin, which selectively adsorbs the enzyme into the clot. This protects thrombin from heparin-antithrombin inactivation, thus preserving the enzyme for postclotting events. To determine how the fibrin N-terminal polymerization sites of A alpha 17-23 (GPRVVER) and B beta 15-25 (GHRPLDKKREE) and their analogs may interact with thrombin, amidolysis vs. plasma- and fibrinogen-clotting assays were used to differentiate blockade of catalytic site vs. other thrombin domains. Amidolysis studies suggest GPRVVER inhibition of thrombin catalytic site through hydrophobic interaction, and GPRVVER inhibited clotting. Neither GPRP nor VVER nor the B beta 15-25 homologs inhibited amidolysis. Contrary to heparin, acyl-DKKREE promoted plasma-clotting, but inhibited fibrinogen-clotting. In addition, acyl-DKKREE reversed the anticoagulant effect of heparin (0.1 U/ml) in plasma. The results suggest fibrin B beta 15-25 interaction with thrombin, possibly by blocking the heparin-binding site. Together with the reported fibrin A alpha 27-50 binding to thrombin, polymerizing fibrin appears to initially bind to thrombin catalytic site and exosite-1 through A alpha 17-50, and to another thrombin site through B beta 15-25. As these fibrin sites are also involved in polymerization, competition of the polymerization process with thrombin-binding could subsequently dislodge thrombin from fibrin alpha-chain. This may re-expose the catalytic site and exosite-1, thus explaining the thrombogenicity of clot-bound thrombin. The implications of these findings in polymerization mechanism and anticoagulant design are discussed.

Gene Cloning of Rat and Mouse Platelet Glycoprotein V: Identification of Megakaryocyte-Specific Promoters and Demonstration of Functional Thrombin Cleavage

Platelet glycoprotein (GP) V is a major surface protein cleaved during thrombin-induced platelet activation. GPV associates noncovalently with the GPIb-IX complex to form GPIb-V–IX, a receptor for von Willebrand factor and thrombin. We describe the cloning of the genes coding for rat and mouse GPV and compare them with the human gene. The two rodent genes have a similar structure and resemble the human GPV gene with a coding sequence (≈1,700 nucleotides) entirely contained in one exon and a single intron (≈900 nucleotides) in the 5′ untranslated region. Both genes have megakaryocyte-type promoters with conserved tandem Ets and GATA recognition motifs and lack a TATA box. The mature rat and mouse proteins comprise 551 amino acids, have 70% sequence identity, and contain an additional 8–amino acid intracellular segment as compared with the human protein. As in human GPV, there is an NH2 -terminal leucine-rich region of 15 repeats and a thrombin cleavage recognition sequence. Whereas the rat and human thrombin cleavage sites are similar, the mouse cleavage site resembles that of the human thrombin receptor. Functionality of these sites was demonstrated by thrombin cleavage of synthetic peptides and analysis by high-performance liquid chromatography (HPLC) or mass spectrometry. Cleavage of native rat GPV was confirmed by means of a polyclonal antibody directed against the new NH2 -terminal peptide exposed after thrombin cleavage. This antibody specifically recognized thrombin-activated rat platelets by fluorescence-activated cell sorting (FACS) analysis. In addition, we raised monoclonal antibodies specific for rat GPV (88 kD), which recognized the NH2 -terminal soluble fragment (70 kD) liberated after thrombin cleavage. Knowledge of these rodent GPV genes and availability of species-specific peptides and antibodies will be essential to further studies aiming to define the exact in vivo function of platelet GPV using animal models of thrombosis and gene inactivation experiments.

Contribution of lysine 60f to S1' specificity of thrombin

Lys60f has been proposed to limit the S1' substrate binding site specificity of thrombin to small polar P1' residues by occluding the S1' binding pocket, based on the X-ray crystal structure of thrombin. To test this proposal, we prepared a Lys-->Ala (K60fA) mutant of recombinant thrombin and determined whether this mutation enhanced the reactivity of thrombin with a variant inhibitor [antithrombin (AT)-Denver] and a substrate (protein C) containing poorly recognized P1' Leu residues. AT-Denver in the presence of heparin inhibited K60fA thrombin with a second-order association rate constant [k = 4.2 +/- 0.1) x 10(5) M-1 s-1] that was 3.2-fold faster than thrombin [k = (1.3 +/- 0.1) x 10(5) M-1 s-1]. Wildtype AT (P1' Ser) under the same conditions inhibited K60fA thrombin with a 2.5-fold slower rate constant [k = (1.1 +/- 0.1) x 10(7) M-1 s-1] than thrombin [k = (2.8 +/- 0.1) x 10(7) M-1 s-1]. These results indicate an overall 8.3-fold improvement in the recognition of the P1' Leu of AT-Denver by K60fA thrombin over that of wild-type thrombin; i.e., the K60fA mutation partly overcomes the defect in thrombin inhibition produced by the P1' mutation in AT-Denver. Resolution of the two-step reactions of AT and AT-Denver with wild-type and mutant thrombins revealed that the enhanced recognition of P1' Leu in AT-Denver by K60fA thrombin occurs primarily in the second reaction step in which a noncovalent AT-thrombin encounter complex is converted to a stable, covalent complex. Thrombin K60fA activated Gla-domainless protein C (GDPC) approximately 2- and approximately 4-fold faster than thrombin in the presence and absence of thrombomodulin (TM), respectively, consistent with an improved interaction of the Leu P1' residue with the mutant S1' pocket. In contrast, the mutant thrombin clotted fibrinogen (P1' Gly) approximately 3-fold slower than thrombin. Kinetic analysis revealed that the improvement in the catalytic rate of activation of GDPC by K60fA thrombin in the presence of TM was localized in the second reaction step, as reflected by an approximately 2-fold increase in kcat. Direct binding studies showed that the K60fA mutation minimally affected the affinity of thrombin for Na+, indicating that the changes in S1' site-specificity of K60fA thrombin did not result from altering the allosteric transition induced by Na+. We conclude that Lys60f limits the P1' substrate and inhibitor specificity of thrombin by influencing the size and polarity of the S1' site which thereby affects the stability of the transition state for cleavage of the scissile bond in the second reaction step.

Molecular basis of residue 192 participation in determination of coagulation protease specificity

Residue 192 (chymotrypsin numbering system) in thrombin, activated protein C, and factor Xa contributes to the specificity of these enzymes toward their substrates and inhibitors. A Glu192-->Gln mutation in both thrombin and activated protein C yielded enzymes that reacted better with some, but not all, of their natural substrates and inhibitors. To determine whether the specificity change is due to productive interactions of Gln192 with substrates and inhibitors or elimination of repulsive electrostatic interactions, we prepared forms of thrombin, des-(1-45)-factor Xa and activated des-(1-45)-protein C with Glu, Gln, or Met at position 192 and compared their activities toward inhibitors and substrates. All mutants had nearly normal amidolytic activity. The Glu192-->Gln and Glu 192-->Met mutations of thrombin and activated des-(1-45)-protein C increased the second-order rate constant (k2) of inhibition by alpha 1-antitrypsin about 700-fold and 170-fold for thrombin, and 185-fold and 150-fold for activated des-(1-45)-protein C, respectively. [E192]faxtor Xa, but not [M192]factor Xa, was resistant to inhibition by alpha 1-antitrypsin. Glu-->Gln or Glu-->Met mutants of both thrombin and activated des-(1-45)-protein C were effectively inhibited by tissue factor pathway inhibitor (K1 < 200 nM) and, except for [M192]thrombin, by bovine pancreatic trypsin inhibitor (K1 < 60 nM). With respect to substrate cleavage, Glu192-->Gln and Glu192-->Met mutations of activated des-(1-45)-protein C both inactivated factor Va 2-3-fold faster than activated des-(1-45)-protein C. Thrombin and [M192]thrombin activated protein C at similar slow rates compared to rapid activation by [Q192]thrombin. The Gln192-->Met mutants of des-(1-45)-factor Xa activated prethrombin 1.8-11-fold slower than wild-type enzyme. With thrombomodulin or factor Va present, these differences in protein C and prethrombin 1 activation rates were decreased to about 2-fold. We conclude that residue 192 contribution to enzyme specificity is achieved by both productive and repulsive interactions and that the magnitude and nature of the participation varies among enzymes, substrates and inhibitors.

The shape of thrombomodulin and interactions with thrombin as determined by electron microscopy

Studies have been carried out to investigate aspects of the structure of thrombomodulin, an endothelial cell glycoprotein that binds thrombin and accelerates both the thrombin-dependent activation of protein C and the inhibition of antithrombin III. We have determined the shape of SolulinTM, a soluble recombinant form of human thrombomodulin missing the transmembrane and cytoplasmic domains, by electron microscopy of preparations rotary-shadowed with tungsten. Solulin appears to be an elongated molecule about 20 nm long that has a large nodule at one end and a smaller nodule near the other end from which extends a thin strand. About half of the molecules form bipolar dimers apparently via interactions between these thin strands. Electron microscopy of complexes formed between Solulin and human alpha-thrombin revealed that a single thrombin molecule appears to bind to the smaller nodule of Solulin, suggesting that this region contains the epidermal growth factor-like domains 5 and 6. Epidermal growth factor-like domains 1-4 comprise the connector between the small and large nodule, which is the lectin-like domain; the thin strand at the other end of the molecule is the carbohydrate-rich region. With chondroitin sulfate-containing soluble thrombomodulin produced from either human melanoma cells Bowes or Chinese hamster ovary cells, a higher percentage of molecules bound thrombin and, in some cases, two thrombin molecules were attached to one soluble thrombomodulin in approximately the same region. These structural studies provide insight into the structure of thrombomodulin and its interactions with thrombin as well as aspects of the mechanisms of its actions.

Crystal structures of thrombin with thiazole-containing inhibitors: probes of the S1' binding site

Structures of the blood clotting enzyme thrombin complexed with hirugen and two active site inhibitors, RWJ-50353 10080(N-methyl-D-phenylalanyl-N-[5-[(aminoiminomethyl)amino]-1- [[(2-benzothiazolyl)carbonyl]butyl]-L-prolinamide trifluoroacetate hydrate) and RWJ-50215 (N-[4-(aminoiminomethyl)amino-1-[2- (thiazol-2-ylcarbonylethyl)piperidin- 1-ylcarbonyl]butyl]-5-(dimethylamino)naphthalenesulfonamide trifluoroacetate hydrate), were determined by x-ray crystallography. The refinements converged at R values of 0.158 in the 7.0-2.3-A range for RWJ-50353 and 0.155 in the 7.0-1.8-A range for RWJ-50215. Interactions between the protein and the thiazole rings of the two inhibitors provide new valuable information about the S1' binding site of thrombin. The RWJ-50353 inhibitor consists of an S1'-binding benzothiazole group linked to the D-Phe-Pro-Arg chloromethyl ketone motif. Interactions with the S1-S3 sites are similar to the D-phenylalanyl-prolyl-arginyl chloromethylketone structure. In RWJ-50215, a S1'-binding 2-ketothiazole group was added to the thrombin inhibitor-like framework of dansylarginine N-(3-ethyl-1,5-pentanediyl)amide. The geometry at the S1-S3 sites here is also similar to that of the parent compound. The benzothiazole and 2-ketothiazole groups bind in a cavity surrounded by His57, Tyr60A, Trp60D, and Lys60F. This location of the S1' binding site is consistent with previous structures of thrombin complexes with hirulog-3, CVS-995, and hirutonin-2 and -6. The ring nitrogen of the RWJ-50353 benzothiazole forms a hydrogen bond with His57, and Lys60F reorients because of close contacts. The oxygen and nitrogen of the ketothiazole of RWJ-50215 hydrogen bond with the NZ atom of Lys60F.

Binding of fibrin monomer and heparin to thrombin in a ternary complex alters the environment of the thrombin catalytic site, reduces affinity for hirudin, and inhibits cleavage of fibrinogen

Interaction of the blood clotting proteinase, thrombin, with fibrin monomer and heparin to form a thrombin.fibrin monomer.heparin ternary complex is accompanied by a change in thrombin catalytic specificity. Equilibrium binding interactions in the assembly of the ternary complex were characterized quantitatively using thrombin labeled at the active site with a fluorescent probe and related to changes in thrombin specificity toward exosite I-dependent binding of hirudin and cleavage of fibrinogen. Changes in the active site environment accompanying binding of heparin or fibrin to thrombin in binary complexes were reported by fluorescence enhancements which contributed additively to the perturbation accompanying formation of the ternary complex. Quantitative analysis of the interactions supports a preferentially ordered path of ternary complex assembly, in which initial binding of heparin to thrombin facilitates binding of fibrin monomer with an approximately 40-fold increased affinity. Binding of fibrin monomer in the ternary complex decreased the affinity of native thrombin for hirudin by >100-fold and inhibited cleavage of fibrinogen, but this inhibition was overcome when fibrin(ogen)-fibrin interactions occurred. These results support a ternary complex model in which heparin binding through exosite II of thrombin facilitates fibrin monomer binding via exosite I, with accompanying changes in thrombin catalytic specificity resulting from perturbations in the active site and reduced accessibility of exosite I to hirudin and fibrinogen.

Bovine thrombin complexed with an uncleavable analog of residues 7-19 of fibrinogen A alpha: geometry of the catalytic triad and interactions of the P1', P2', and P3' substrate residues

The crystal structure of the noncovalent complex of bovine thrombin and a fibrinogen-A alpha tridecapeptide substrate analog, G17 psi, in which the scissile bond amide nitrogen of Gly-17f has been replaced by a methylene carbon, has been determined at 2.3 A resolution with an R factor of 17.1%. The geometry of the active site indicates that the crystal structure is a close model of the true Michaelis complex. The three independently determined thrombin/G17 psi complexes in the crystal asymmetric unit reveal novel interactions for the P2' and P3' residues-Pro-18f and Arg-19f, respectively-on the carboxyl-terminal side of the scissile bond and confirm previously observed interactions of the P1 (Arg-16f) through P10 (Asp-7f) positions on the amino-terminal side. The thrombin S2' binding site for Pro-18f, as observed in all three complexes, differs from that predicted by modeling studies and is notable for including two carbonyl oxygens of the thrombin main chain. Arg-19f occupies two binding sites on thrombin, S3'A and S3'B, which have dramatically different placements for the arginyl side chain and carboxyl terminus.

Role of residue 99 at the S2 subsite of factor Xa and activated protein C in enzyme specificity

It is thought that only a limited number of residues in the extended binding pocket of coagulation proteases are critical for substrate and inhibitor specificity. A candidate residue from the crystal structures of thrombin and factor Xa (FXa) that may be critical for specificity at the S2 subsite is residue 99. Residue 99 is Tyr in FXa and Thr in activated protein C (APC). To determine the role of residue 99 in S2 specificity, a Gla-domainless mutant of protein C (GDPC) was prepared in which Thr99 was replaced with Tyr of FXa. GDPC T99Y bound Ca2+ and was activated by the thrombin-thrombomodulin complex normally. The T99Y mutant, similar to FXa, hydrolyzed the chromogenic substrates with a Gly at the P2 positions. This mutant was also inhibited by antithrombin (AT) (k2 = 4.2 +/- 0.2 x 10(1) M-1 s-1), and heparin accelerated the reaction >350-fold (k2 = 1.5 +/- 0.1 x 10(4) M-1 s-1). The T99Y mutant, however, did not activate prothrombin but inactivated factor Va approximately 2-fold better than wild type. To try to switch the specificity of FXa, both Tyr99 and Gln192 of FXa were replaced with those of APC in the Gla-domainless factor X (GDFX Y99T/Q192E). This mutant was folded correctly as it bound Ca2+ with a similar affinity as GDFX and was also activated by the Russell's viper venom at similar rate, but it cleaved the chromogenic substrates with a Gly at the P2 positions poorly. The mutant, instead, cleaved the APC-specific chromogenic substrates efficiently. The Y99T/Q192E mutant became resistant to inhibition by AT in the absence of heparin but was inhibited by AT almost normally in the presence of heparin (k2 = 3.4 +/- 0.5 x 10(5) M-1 s-1). The Y99T/Q192E mutant did not inactivate factor Va, and prothrombin activation by this mutant was impaired. These results indicate that 1) residue 99 is critical for enzyme specificity at the S2 subsite, 2) a role for heparin in acceleration of FXa inhibition by AT may involve the S2-P2 modulation, and 3) the exchange of residues 99 and 192 in FXa and APC may switch the enzyme specificity with the chromogenic substrates and inhibitors but not with the natural substrates.

Identification and characterization of the thrombin binding sites on fibrin

Thrombin binds to fibrin at two classes of non-substrate sites, one of high affinity and the other of low affinity. We investigated the location of these thrombin binding sites by assessing the binding of thrombin to fibrin lacking or containing gamma' chains, which are fibrinogen gamma chain variants that contain a highly anionic carboxyl-terminal sequence. We found the high affinity thrombin binding site to be located exclusively in D domains on gamma' chains (Ka, 4.9 x 10(6) M-1; n, 1.05 per gamma' chain), whereas the low affinity thrombin binding site was in the fibrin E domain (Ka, 0.29 x 10(6) M-1; n, 1.69 per molecule). The amino-terminal beta15-42 fibrin sequence is an important constituent of low affinity binding, since thrombin binding at this site is greatly diminished in fibrin molecules lacking this sequence. The tyrosine-sulfated, thrombin exosite-binding hirudin peptide, S-Hir53-64 (hirugen), inhibited both low and high affinity thrombin binding to fibrin (IC50 1.4 and 3.0 microM respectively). The presence of the high affinity gamma' chain site on fibrinogen molecules did not inhibit fibrinogen conversion to fibrin as assessed by thrombin time measurements, and thrombin exosite binding to fibrin at either site did not inhibit its catalytic activity toward a small thrombin substrate, S-2238. We infer from these findings that there are two low affinity non-substrate thrombin binding sites, one in each half of the dimeric fibrin E domain, and that they may represent a residual aspect of thrombin binding and cleavage of its substrate fibrinogen. The high affinity thrombin binding site on gamma' chains is a constitutive feature of fibrin as well as fibrinogen.

Enhanced thrombin sensitivity of a factor VIII-heparin cofactor II hybrid

Generation of thrombin at a site of vascular injury is a key event in the arrest of bleeding. In addition to the conversion of fibrinogen into the insoluble fibrin, thrombin can initiate a number of positive and negative feedback mechanisms that either sustain or down-regulate clot formation. We have modulated the thrombin sensitivity of human blood coagulation factor VIII, an essential cofactor in the intrinsic pathway of blood coagulation. We have substituted an acidic region of factor VIII corresponding to amino acid sequence Asp712-Ala736 by amino acid sequence Ile51-Leu80 of the thrombin inhibitor heparin cofactor II. Functional analysis of the resulting factor VIII-heparin cofactor II hybrid, termed des-(868-1562)-factor VIII-HCII, revealed an increase in procoagulant activity as measured in a one-stage clotting assay. Incubation of purified des-(868-1562)-factor VIII-HCII with different amounts of thrombin showed that this protein was more readily activated by thrombin when compared with des-(868-1562)-factor VIII, a control protein lacking amino acid sequence Ile51-Leu80 of heparin cofactor II. This was manifested by an increase in the second order rate constant of activation by thrombin for des-(868-1562)-factor VIII-HCII (12.0 +/- 0.48 x 10(6) M-1 s-1) compared with des-(868-1562)-factor VIII (1.77 +/- 0.21 x 10(6) M-1 s-1). Our data suggest that amino acid sequence Ile51-Leu80 of heparin cofactor II endows factor VIII with increased sensitivity towards thrombin which results in accelerated clot formation.

Inhibition of thrombin generation by aspirin is blunted in hypercholesterolemia

Recent evidence indicates that aspirin inhibits thrombin generation in clotting blood. We noticed that this effect was less pronounced in patients with hypercholesterolemia. The aim of the study was to prove this observation. The effects of aspirin on thrombin generation were evaluated in (1) 46 healthy volunteers, 2 hours after ingestion of a single, 500-mg dose and (2) 28 survivors of myocardial infarction who took 300 mg aspirin/d for 2 weeks. In both populations, two well-matched subgroups were distinguished, using a serum cholesterol level of 6.2 mmol/L (240 mg/dL) and an LDL cholesterol level of 4.0 mmol/L (155 mg/dL) as borderline. Thrombin generation was monitored ex vivo in blood emerging from a skin microvasculature injury and additionally, in a single-dose study in vitro in recalcified plasma. Aspirin depressed thrombin generation in the group of subjects with serum cholesterol < 6.2 mmol/L and LDL cholesterol < 4.0 mmol/L but not in the group with high blood cholesterol levels. Inhibitory effects of aspirin were more pronounced after the 2-week treatment than after a single dose. There was a significant correlation between total serum cholesterol or LDL cholesterol and total amount of thrombin generated after aspirin treatment. In subjects with high blood cholesterol levels, thrombin generation was not affected by aspirin. Blunting of aspirin action in hypercholesterolemia might be explained by (1) alterations in platelet lipid-protein matrix that render their membrane proteins less accessible for acetylation by aspirin and (2) changes in composition and structure of plasma lipoproteins that diminish the chance of aspirin to interact with prothrombin.

Structure-based understanding of ligand affinity using human thrombin as a model system

Kinetic study of a series of compounds containing the thrombin-directed peptide D-Phe-ProboroArg-OH had indicated that the structure of the N-terminal blocking group may be correlated with binding [Kettner, C., Mersinger, L., & Knabb, R. (1990) J. Biol. Chem. 265, 18289-18297]. In order to further study this phenomenon, a second series of compounds that contains a C-terminal methyl ester in place of the boronic acid was synthesized, binding measured, and the three-dimensional structure in complex with human thrombin determined by X-ray crystallography. Incubation of Ac-D-Phe-Pro-Arg-OMe, Boc-D-Phe-Pro-Arg-OMe, and H-D-Phe-Pro-Arg-OMe resulted in the formation of thrombin-product complexes within the crystal. Ki values for the corresponding products (free carboxylic acids) were 60 +/- 12 microM, 7.8 +/- 0.1 microM, 0.58 +/- 0.02 microM, respectively, indicating that the nature of the N-terminal blocking group has a significant effect on affinity. Examination of the crystal structures indicated that the higher affinity of the H-D-Phe peptide is due to rearrangement of one residue comprising the S3 site (Glu192) in order to maximize electrostatic interactions with the "NH3(+)-" of H-D-Phe. The relative affinity of Boc-D-Phe-Pro-Arg-OH is due to favorable hydrophobic interactions between thrombin and the bulky butyl group. However, this results in less favorable binding of Arg-P1 in the oxyanion hole as shown by long hydrogen-bonding distances. This work gave rise to some general observations applicable to structure-based drug design: (1) altering the structure of an inhibitor at one site can affect binding at an unchanged distal site; (2) minor alteration of inhibitor structure can lead to small, but significant reorganization of neighboring protein structure; (3) these unexpected reorganizations can define alternate binding motifs.

Involvement of thrombin anion-binding exosites 1 and 2 in the activation of factor V and factor VIII

The role of anion-binding exosites of thrombin in the activation of factor V and factor VIII was studied using thrombin Arg93 --> Ala, Arg97 --> Ala, and Arg101 --> Ala (thrombin RA), a recombinant exosite 2 defective mutant, and a synthetic N-acetylated dodecapeptide, Ac-Asn-Gly-Asp-Phe-Glu-Glu-Ile-Pro-Glu-Glu-Tyr-O-SO4Leu (hirugen), which competitively inhibits binding of macromolecules to exosite 1. The catalytic efficiency of the activation of factor VIII or of the first step of factor V activation by thrombin RA was approximately 10% that of wild-type thrombin. The overall rate of conversion to factor Va was not influenced by the mutation. In contrast to factor V, the slow activation of factor VIII by thrombin RA was associated with a decreased rate of cleavage at all three proteolytic sites (Arg372, Arg740, and Arg1689). Hirugen inhibited factor V and factor VIII activation. These results indicate that both anion-binding exosites of thrombin are involved in the recognition of factor V and factor VIII.

Rational design of true hirudin mimetics: synthesis and characterization of multisite-directed alpha-thrombin inhibitors

We describe here the design, synthesis, and activity of a novel class of alpha-thrombin inhibitors named hirunorms. They were rationally designed to interact through their N-terminal end with the alpha-thrombin active site in a nonsubstrate mode and to specifically bind the fibrinogen recognition exosite. An appropriate spacer that is able to properly orient the N-terminal end in the active site was also selected. This spacer allowed the size of the inhibitors to be reduced to about one-third of the amino acid residues in the hirudin sequence. Hirunorms specifically inhibit the amidolytic action of human alpha-thrombin toward a small chromogenic substrate. The most active compounds of the series, hirunorms IV and V, inhibit alpha-thrombin catalyzed hydrolysis of Tos-Gly-Pro-Arg-p-nitroanilide with K(i) = 0.09 and K(i) = 0.21 nM, respectively. Comparison of the anticoagulant properties of hirunorms, natural hirudin from the European leech Hirudo medicinalis, and the synthetic analog hirulog-1 revealed that hirunorm IV is about 10-fold and 3-fold more active, on a molar base, than hirudin and hirulog-1 in increasing the aPTT, PT, and TT of normal human plasma. The peculiar structure of hirunorms makes them stable to the amidolytic action of thrombin without the introduction of any peptide bond modification. These molecules display long-lasting activity in human plasma, due to the presence of several unnatural amino acids in susceptible positions. Hirunorms are potential candidates for injectable anticoagulants, due to their potency, specificity of action, long-lasting activity, and safety profiles.

Comparison of the structures of the cyclotheonamide A complexes of human alpha-thrombin and bovine beta-trypsin

Thrombin, a trypsin-like serine protease present in blood, plays a central role in the regulation of thrombosis and hemostasis. A cyclic pentapeptide, cyclotheonamide A (CtA), isolated from sponges of the genus Theonella, inhibits thrombin, trypsin, and certain other serine proteases. Enzyme inhibition data for CtA indicate that it is a moderate inhibitor of alpha-thrombin (K(i) = 1.0 nM), but substantially more potent toward trypsin (K(i) = 0.2 nM). The comparative study of the crystal structures of the CtA complexes of alpha-thrombin and beta-trypsin reported here focuses on structure-function relationships in general and the enhanced specificity of trypsin, in particular. The crystal structures of the CtA complexes of thrombin and trypsin were solved and refined at 1.7 and 2.0 A resolution, respectively. The structures show that CtA occupies the active site with the Pro-Arg motif positioned in the S2 and S1 binding sites. The alpha-keto group of CtA is involved in a tetrahedral intermediate hemiketal structure with Ser 195 OG of the catalytic triad and is positioned within bonding distance from, and orthogonal to, the re-face of the carbonyl of the arginine of CtA. As in other productive binding modes of serine proteases, the Ser 214-Gly 216 segment runs in a twisted antiparallel beta-strand manner with respect to the diaminopropionic acid (Dpr)-Arg segment of CtA. The Tyr 60A-Thr 60I insertion loop of thrombin makes a weak aromatic stacking interaction with the v-Tyr of CtA through Trp 60D. The Glu 39 Tyr and Leu 41 Phe substitutions in trypsin produce an enhanced aromatic interaction with D-Phe of CtA, which also leads to different orientations of the side chains of D-Phe and the v-Tyr. The comparison of the CtA complexes of thrombin and trypsin shows that the gross structural features of both in the active site region are the same, whereas the differences observed are mainly due to minor insertions and substitutions. In trypsin, the substitution of Ile 174-Arg 175 by Gly 174-Gln 175 makes the S3 aryl site more polar because the Arg 175 side chain is directed away from thrombin and into the solvent, whereas Gln 175 is not. Because the site is occupied by the Dpr group of CtA, the occupancy of the S3 site is better in trypsin than in thrombin. In trypsin, the D-Phe side chain of CtA fits between Tyr 39 and Phe 41 in a favorable manner, whereas in thrombin, these residues are Glu 39 and Leu 41. The higher degree of specificity for trypsin is most likely the result of these substitutions and the absence of the fairly rigid Tyr 60A-Thr 60I insertion loop of thrombin, which narrows access to the active site and forces less favorable orientations for the D-Phe and v-Tyr residues.

Heparin enhances the catalytic activity of des-ETW-thrombin

The thrombin mutant, des-ETW-thrombin, lacking Glu(146), Thr(147), and Trp(148) within a unique insertion loop located at the extreme end of the primary specificity pocket, has been shown previously to exhibit reduced catalytic activity with respect to macromolecular and synthetic thrombin substrates and reduced or enhanced susceptibility to inhibition. Investigation of the hydrolysis of peptidyl p-nitroanilide substrates by des-ETW-thrombin showed increased activity in the presence of heparin and other sulphated glycosaminoglycans. No effect was observed upon the activity of wild-type thrombin. Heparin was found to decrease the K(m) for cleavage of four thrombin-specific substrates by des-ETW-thrombin by 3-4-fold. Similarly, pentosan polysulphate (PPS) decreased the K(m) with these substrates by 8-10-fold. Heparin also increased the rate of inhibition of des-ETW-thrombin by antithrombin III and D-phenylalanyl-prolyl-arginylchloromethane (PPACK). The inhibition of des-ETW-thrombin by a number of thrombin-specific peptide boronic acids also showed significant reduction in the final K(i) in the presence of heparin, due to reduction in the off-rate. A peptide analogue of a sequence of hirudin which binds thrombin tightly to exosite I (fibrinogen recognition site) potentiated the activity of des-ETW-thrombin against peptide p-nitroanilide substrates in a manner similar to heparin. The K(i) for the inhibition of des-ETW-thrombin by p-aminobenzamidine was decreased by these ligands from 9.7 mM to 7.5 mM, 5.1 mM, and 2.5 mM in the presence of heparin, hirudin peptide and PPS respectively, suggesting the increased catalytic activity is due to enhanced access to the primary specificity pocket. The positive influence of these ligands on des-ETW-thrombin was reversed in the presence of ATP or ADP; the latter has previously been shown to inhibit thrombin activity by blocking initial interaction with fibrinogen at exosite 1. Because the effect of heparin and PPS is similar to that of hirudin peptide, it is proposed that the most likely mechanism is that binding at the heparin-binding site (thrombin exosite 2) facilitates interaction at exosite 1 causing a conformational change which partially corrects the defective ground-state binding of the mutant thrombin. Although no effect was observed upon the activity of wild-type thrombin, our findings do provide further evidence of an allosteric property of thrombin which may control the geometry of, and access to, the primary specificity pocket.

Crystallographic evidence that the F2 kringle catalytic domain linker of prothrombin does not cover the fibrinogen recognition exosite

The 2.6-A x-ray crystal structure of bovine alpha-thrombin in complex with rhodniin, a protein inhibitor isolated from the bug Rhodnius prolixus, has been solved and refined. The structure has enabled us to trace the N-terminal part of the 49-residue A-chain of bovine alpha-thrombin for the first time, which is fixed in a U-shaped loop on the molecular surface opposite the active site canyon. Model building shows that the 25 amino acid residues that link the A-chain and F2 kringle cannot run through the fibrinogen recognition exosite. This demonstrates that this fibrinogen recognition exosite is available in prothrombin and meizothrombin.

Amino acid residues 721-729 are required for full factor VIII activity

Recombinant two-chain factor VIII, from which the B domain had been deleted, was expressed in Chinese hamster ovary cells. In addition to the major product, three minor factor VIII forms were isolated. The A2 domains generated by thrombin cleavage showed different electrophoretic mobilities. Peptide mapping of the A2 domains showed that two of the factor VIII forms had the expected C-terminus of the heavy chain at Arg740 [FVIII-(1-740)] and that the other factor VIII forms had C-termini at Tyr729 [FVIII-(1-729)] or Glu720 [FVIII-(1-720)]. The major FVIII-(1-740) form, FVIII-(1-729), and FVIII-(1-720) contained sulfated tyrosine residues at Tyr718, Tyr719 and Tyr723. The minor FVIII-(1-740) form was shown to lack sulfation at these positions. The specific clotting activity was approximately 1 x 10(4) U/mg for FVIII-(1-740) (both forms) and FVIII-(1-729), but twofold lower for FVIII-(1-720). A time study of thrombin activation showed that FVIII-(1-720) was activated slower than FVIII-(1-740), FVIII-(1-729) and plasma-derived factor VIII. Partially sulfated FVIII-(1-740) was activated at the same rate as the fully sulfated FVIII-(1-740). The equilibrium dissociation constant for binding of factor VIII to inactivated immobilized thrombin was the same for all factor VIII forms, showing that the slower activation of FVIII-(1-720) was not due to a lower affinity for the anion-binding exosite in thrombin.