Macrocyclic factor XIa inhibitors

A series of macrocyclic factor XIa (FXIa) inhibitors was designed based on an analysis of the crystal structures of the acyclic phenylimidazole compounds. Further optimization using structure-based design led to inhibitors with pM affinity for FXIa, excellent selectivity against a panel of relevant serine proteases, and good potency in the activated partial thromboplastin time (aPTT) clotting assay.     

Isolation and functional characterization of the active light chain of activated human blood coagulation factor XI

Human blood coagulation Factor XIa was reduced and alkylated under mild conditions. The mixture containing alkylated heavy and light chains was subjected to affinity chromatography on high Mr kininogen-Sepharose. Alkylation experiments using [14C]iodoacetamide showed that a single disulfide bridge between the light and heavy chains was broken to release the light chain. The alkylated light chain (Mr = 35,000) did not bind to high Mr kininogen-Sepharose while the heavy chain (Mr = 48,000), like Factors XI and XIa, bound with high affinity. The isolated light chain retained the specific amidolytic activity of native Factor XIa against the oligopeptide substrate, pyroGlu-Pro-Arg-p-nitroanilide. Km and kcat values for this substrate were 0.56 mM and 350 s-1 for both Factor XIa and its light chain, and the amidolytic assay was not affected by CaCl2. However, in clotting assays using Factor XI-deficient plasma in the presence of kaolin, the light chain was only 1% as active as native Factor XIa. Human coagulation Factor IX was purified and labeled with sodium [3H]borohydride on its carbohydrate moieties. When this radiolabeled Factor IX was mixed with Factor XIa, an excellent correlation was observed between the appearance of Factor IXa clotting activity and tritiated activation peptide that was soluble in cold trichloroacetic acid. Factor XIa in the presence of 5 mM CaCl2 activated 3H-Factor IX 600 times faster than Factor XIa in the presence of EDTA. In the absence of calcium, Factor XIa and its light chain were equally active in activating 3H-Factor IX. In contrast to Factor XIa, the light chain in this reaction was inhibited by calcium ions such that, in the presence of 5 mM CaCl2, Factor XIa was 2000 times more effective than its light chain. Neither phospholipid nor high Mr kininogen and kaolin affected the activity of Factor XIa or its light chain in the activation of 3H-Factor IX. These observations show that the light chain region of Factor XIa contains the entire enzymatic active site. The heavy chain region contains the high affinity binding site for high Mr kininogen. Furthermore the heavy chain region of Factor XIa plays a major role in the calcium-dependent mechanisms that contribute to the activation of Factor IX.     

Intrinsic versus extrinsic coagulation. Kinetic considerations.

A study to compare the kinetics of activation of factor IX by Factor XIa/Ca2+ and by Factor VIIa/tissue factor/Ca2+ has been undertaken. When purified human proteins, detergent-extracted brain tissue factor and tritiated-activation-peptide-release assays were utilized, the kinetic constants obtained were: Km = 310 nM, kcat. = 25 min-1 for Factor XIa and Km = 210 nM, kcat. = 15 min-1 for Factor VIIa. The kinetic constants for the activation of Factor X by Factor VIIa/brain tissue factor were: Km = 205 nM, kcat. = 70 min-1. Predicted rates for the generation of Factor IXa and Factor Xa were obtained when human monocytic tumour U937 cells (source of tissue factor) and Factor VIIa were used to form the activator. In other experiments, inclusion of high-Mr kininogen did not increase the activation rates of Factor IX by Factor XIa in the presence or absence of platelets and/or denuded rabbit aorta. These kinetic data strongly indicate that both Factor XIa and Factor VIIa play physiologically significant roles in the activation of Factor IX.     

Interaction of type 1 plasminogen activator inhibitor with the enzymes of the contact activation system

The interaction between type 1 plasminogen activator inhibitor (PAI-1), a serine protease inhibitor, and the three serine proteases generated during contact activation of plasma was studied using functional and immunologic approaches. Incubation of Factor XIIa, Factor XIa, and plasma kallikrein with either purified PAI-1 or platelet-derived PAI-1 resulted in the formation of sodium dodecyl sulfate-stable complexes as revealed by immunoblotting techniques. Functional assays indicated that Factor XIa and, to a lesser extent, Factor XIIa and plasma kallikrein neutralized the ability of purified PAI-1 to bind to immobilized tissue-type plasminogen activator (t-PA). Immunoblotting demonstrated that these enzymes also neutralized the ability of PAI-1 to form complexes with fluid-phase t-PA. Clot lysis assays employing purified proteins and 125I-fibrinogen were used to investigate the profibrinolytic effect of these contact activation enzymes. At enzyme concentrations that did not result in direct activation of plasminogen, only Factor XIa was capable of stimulating the lysis of clots supplemented with both t-PA and PAI-1. As a consequence of their interactions with PAI-1, the amidolytic activity of Factor XIIa, Factor XIa, and plasma kallikrein was neutralized by this inhibitor in a time-dependent and concentration-dependent manner. Minimum values estimated for the apparent second-order rate constant of inhibition were 1.6 x 10(4), 2.1 x 10(5), and 6.0 x 10(4) M-1 s-1 for Factor XIIa, Factor XIa, and plasma kallikrein, respectively. These data define new reactions between coagulation and fibrinolysis proteins and suggest that a major mechanism for stimulation of the intrinsic fibrinolytic pathway may involve neutralization of PAI-1 by Factor XIa.     

Characterization of Novel Forms of Coagulation Factor XIa: independence of factor XIa subunits in factor IX activation

Factor XI is the zymogen of a dimeric plasma protease, factor XIa, with two active sites. In solution, and during contact activation in plasma, conversion of factor XI to factor XIa proceeds through an intermediate with one active site (1/2-FXIa). Factor XIa and 1/2-FXIa activate the substrate factor IX, with similar kinetic parameters in purified and plasma systems. During hemostasis, factor IX is activated by factors XIa or VIIa, by cleavage of the peptide bonds after Arg145 and Arg180. Factor VIIa cleaves these bonds sequentially, with accumulation of factor IX alpha, an intermediate cleaved after Arg145. Factor XIa also cleaves factor IX preferentially after Arg145, but little intermediate is detected. It has been postulated that the two factor XIa active sites cleave both factor IX peptide bonds prior to releasing factor IX abeta. To test this, we examined cleavage of factor IX by four single active site factor XIa proteases. Little intermediate formation was detected with 1/2-FXIa, factor XIa with one inhibited active site, or a recombinant factor XIa monomer. However, factor IX alpha accumulated during activation by the factor XIa catalytic domain, demonstrating the importance of the factor XIa heavy chain. Fluorescence titration of active site-labeled factor XIa revealed a binding stoichiometry of 1.9 +/- 0.4 mol of factor IX/mol of factor XIa (Kd = 70 +/- 40 nm). The results indicate that two forms of activated factor XI are generated during coagulation, and that each half of a factor XIa dimer behaves as an independent enzyme with respect to factor IX.     

Human factor XIa cleaves fibrinogen: effects on structure and function

Factor XIa, the enzymatic form of the factor XI zymogen, is generated as a result of factor XII-dependent surface activation in plasma. Factor XIa degrades high molecular weight kininogen, its cofactor for activation (which binds factor XIa to the surface), as well as cleaves and activates coagulation factor IX. In this report, we present evidence that factor XIa can also cleave fibrinogen and decrease the thrombin-catalyzed formation of the fibrin clot. Furthermore, the products of factor XIa-digested fibrinogen markedly inhibited the rate of polymerization of fibrin monomers. Factor XIa initially cleaved the A alpha-chain of fibrinogen and subsequently degraded the B beta-chain. However, the cleavage sites on both chains were distinct from those susceptible to thrombin. The gamma-chain was degraded only after prolonged incubation with factor XIa. Furthermore, the profile of fibrinogen proteolysis by factor XIa was distinctly different from that of plasmin-catalyzed fibrinogenolysis. Unlike plasmin, factor XIa was not able to cleave the NH2-terminus of the B beta-chain of fibrinogen. Moreover, factor XIa, unlike plasmin, failed to hydrolyze fibrin. Further study of the proteolytic digests of fibrinogen produced by factor XIa may give additional insight into the mechanism of polymerization of this protein.     

The heparin-catalysed inhibition of human factor XIa by antithrombin III is dependent on the heparin type.

The effect of various well-characterized heparin preparations on the inactivation of human Factor XIa by human antithrombin III was studied. The heparin preparations used were unfractionated heparin and four heparin fractions obtained after anion-exchange chromatography. Inactivation of Factor XIa was monitored with S2366 as chromogenic substrate and followed pseudo-first-order reaction kinetics under all reaction conditions tested. Enhancement of the rate of inhibition of Factor XIa in the presence of unfractionated heparin correlated to the binding of antithrombin III to heparin. From the kinetic data a binding constant of 0.1 microM was inferred. The maximum rate enhancement, achieved at saturating heparin concentrations, was 30-fold. The rate enhancement achieved in the presence of each of the heparin fractions could also be correlated to the binding of antithrombin III to the heparin. The binding constant inferred from the kinetic data varied from 0.10 to 0.28 microM and the number of binding sites for antithrombin III varied from 0.06 to 0.74 site per heparin molecule. The maximum rate enhancements, achieved at saturating heparin concentrations, were strongly dependent on the type of heparin used and varied from 7-fold for fraction A to 41-fold for fraction D. Therefore, although the stimulation of Factor XIa inactivation by antithrombin III could be quantitatively correlated to the binding of antithrombin III to heparin, the heparin-catalysed inhibition of Factor XIa is dependent not only upon the degree of binding of antithrombin III to heparin but also upon the type of heparin to which antithrombin III is bound.     

The molecular-weight dependence of the rate-enhancing effect of heparin on the inhibition of thrombin, factor Xa, factor IXa, factor XIa, factor XIIa and kallikrein by antithrombin.

Heparin fractions of different molecular weight and with high affinity for antithrombin were studied with respect to their ability to potentiate the inhibition of activated clotting factors by antithrombin. Inhibition of thrombin, Factor IXa and Factor XIa showed similarities in the dependence on the molecular weight of heparin and was found to decrease with decreasing molecular weight. Inactivation of Factor Xa, Factor XIIa and kallikrein was, however, less dependent on the size of the polysaccharide and, to a great extent, was potentiated even by low-molecular-weight heparin fractions that had virtually no effect on the inhibition of thrombin, Factor IXa and Factor XIa.     

Assembly and expression of an intrinsic factor IX activator complex on the surface of cultured human endothelial cells.

Endothelial cells expose specific receptors for blood clotting factors and, upon perturbation, can initiate and propagate the reactions of the extrinsic pathway of blood coagulation leading to fibrin formation on the cell surface. The existence of an intrinsic mechanism of Factor IX activation on cultured human umbilical vein cells (HUVECs) was investigated by studies of the interaction between HUVECs and two proteins of the contact activation system, the cofactor high molecular weight kininogen (H-kininogen) and the zymogen Factor XI. In the presence of zinc ions (10-300 microM), 125I-labeled H-kininogen bound to HUVECs in a time-dependent, reversible, and saturable manner, with calcium ions exerting an inhibitory effect on the zinc-dependent binding. Analysis of the binding data by the LIGAND computer program indicated that HUVECs, in the presence of 2 mM CaCl2 and 100 microM ZnCl2 at 37 degrees C, bound 1.14 x 10(7) H-kininogen molecules per cell with an apparent dissociation constant of 55 nM. HUVEC-bound H-kininogen functions as the cell surface receptor for both 125I-labeled Factor XI and 125I-labeled Factor XIa, since HUVECs cultured in contact factor-depleted serum do not detectably bind either the zymogen or the enzyme in the absence of H-kininogen and zinc ions. In the presence of saturating concentrations of H-kininogen, 2 mM CaCl2 and 100 microM ZnCl2, the binding of 125I-labeled Factor XI and Factor XIa to HUVECs was time-dependent, reversible, and saturable, with apparent dissociation constants of 4.5 and 1.5 nM, respectively. HUVEC-bound complexes of H-kininogen and Factor XI generated Factor XIa activity only after the addition of purified Factor XIIa, and cell-bound Factor XIa in turn activated Factor IX, as documented by a 3H-labeled activation peptide release assay for 3H-Factor IX activation. The results indicate that cultured HUVECs provide a surface for the assembly and expression of an intrinsic Factor IX activator complex that may participate in the initiation of blood coagulation at sites of vascular injury.     

The factor IX phospholipid-binding site is required for calcium-dependent activation of factor IX by factor XIa

To determine the functional role of the metal-dependent conformational changes in Factor IX, two populations of conformation-specific anti-Factor IX antibodies were prepared. Anti-Factor IX X Mg(II) antibodies bind to Factor IX in the presence of Mg(II) and other metal ions, but not in the absence of metal ions. Anti-Factor IX X Ca(II)-specific antibodies bind to Factor IX in the presence of Ca(II) and Sr(II), but not in the presence of Mn(II), Mg(II), and Ba(II). In the presence of a metal ion that induces the conformational transition recognized by the anti-Factor IX X Mg(II) antibodies, the concentrations of CaCl2 and SrCl2 needed for the half-maximal binding of the anti-Factor IX X Ca(II)-specific antibodies to Factor IX were reduced 3- and 20-fold, respectively. Factor IX binding to phospholipid vesicles was inhibited by the Fab fragments of the anti-Factor IX X Ca(II)-specific antibodies, but was not inhibited by the Fab fragments of the anti-Factor IX X Mg(II) antibodies. Factor XIa activation of Factor IX was also inhibited by the Fab fragments of the anti-Factor IX X Ca(II)-specific antibodies, but not by the anti-Factor IX X Mg(II) antibodies. These results support the hypothesis that Factor IX undergoes two metal-dependent conformational transitions: FIX----FIX'----FIX*. The first transition (FIX----FIX') is metal-dependent but cation-nonselective; the second transition (FIX'----FIX*) is metal-selective for Ca(II) or Sr(II). The second transition results in the expression of conformational determinants necessary for membrane binding and the Ca(II)-dependent activation of Factor IX by Factor XIa. These results suggest chemical similarity between a surface of a domain of Factor XIa and phospholipid vesicles, both of which interact with Factor IX in the presence of Ca(II).     

Stoichiometry of binding of high molecular weight kininogen to factor XI/XIa

Factor XI is a dimeric protein and circulates in plasma complexed with high molecular weight kininogen (HMWK). We investigated the binding of HMWK to factor XIa utilizing two active site directed fluorescent probes: nitrobenzoxadiazole aminopentyl methylphosphonofluoridate for serine and dansyl-glu-gly-arg-chloromethyl ketone for histidine. In the presence of saturating amounts of HMWK, the fluorescence of factor XIa-fluorophore was quenched by approximately 28% for each probe. Titrations of the fluorescent factor XIa with HMWK revealed that each subunit of factor XIa binds one molecule of HMWK with a Kd approximately 3.4 X 10(-8)M.     

Reactivity of bovine blood coagulation factor IXa beta, factor Xa beta, and factor XIa toward fluorogenic peptides containing the activation site sequences of bovine factor IX and factor X

The published activation site sequences of bovine factors IX and X have been utilized to synthesize a number of peptides specifically designed respectively as substrates for bovine factors XIa and IXa beta. The substrates contain a fluorophore (2-aminobenzoyl group, Abz) and a quenching group (4-nitrobenzylamide, Nba) that are separated upon enzymatic hydrolysis with a resultant increase in fluorescence that was utilized to measure hydrolysis rates. Factor XIa cleaved all of the peptides bearing factor IX activation site sequences with Abz-Glu-Phe-Ser-Arg-Val-Val-Gly-Nba having the highest kcat/KM value. The kinetic behavior of factor XIa toward the synthetic peptide substrate indicates that it has a minimal extended substrate recognition site at least five residues long spanning S4 to S1' and has favorable interactions over seven subsites. The hexapeptide Abz-Glu-Phe-Ser-Arg-Val-Val-Nba was the most specific factor XIa substrate and was not hydrolyzed by factors IXa beta or Xa beta or thrombin. Factor IXa beta failed to hydrolyze any of the synthetic peptides bearing the activation site sequence of factor X. This enzyme slowly cleaved four hexa- and heptapeptide substrates with factor IX activation site sequences extending from P4 or P3 to P3'. Factor Xa beta poorly hydrolyzed all but one of the factor XIa substrates and failed to cleave any of the factor IXa beta substrates. Thrombin failed to hydrolyze any of the peptides examined while trypsin, as expected, was highly reactive and not very specific. Phospholipids had no effect on the reactivity of either factors IXa beta or Xa beta toward synthetic substrates. Both factor IXa beta and Xa beta cleaved the peptide substrates at similar rates to their natural substrates under comparable conditions. However the rates were substantially lower than optimum activation rates observed in the presence of Ca2+, phospholipids, and protein cofactors. In the future, it may be useful to investigate synthetic substrates that can bind to phospholipid vesicles in the same manner as the natural substrates for factors IXa beta and Xa beta.     

Macrocyclic ureas as potent and selective Chk1 inhibitors: an improved synthesis, kinome profiling, structure-activity relationships, and preliminary pharmacokinetics

A new series of potent macrocyclic urea-based Chk1 inhibitors are described. A detailed SAR study on the 4-position of the phenyl ring of the 14-member macrocyclic ureas 1a and d led to the identification of the potent Chk1 inhibitors 2, 5-7, 10, 13, 14, 19-21, 25, 27, and 31-34. These compounds significantly sensitize tumor cells to the DNA-damaging antitumor agent doxorubicin in a cell-based assay and efficiently abrogate the doxorubicin-induced G2/M and camptothecin-induced S checkpoints, indicating that the potent biological activities of these compounds are mechanism-based through Chk1 inhibition. Kinome profiling analysis of a representative macrocyclic urea 25 against a panel of 120 kinases indicates that these novel macrocyclic ureas are highly selective Chk1 inhibitors. Preliminary PK studies of 1a and b suggest that the 14-member macrocyclic inhibitors may possess better PK properties than their 15-member counterparts. An improved synthesis of 2 and 20 by using 2-(trimethylsilyl)ethoxycarbonyl (Teoc) to protect the amino group not only readily provided the desired compounds in pure form but also facilitated the scale up of potent compounds for various biological studies.     

Substrate-dependent modulation of the mechanism of factor XIa inhibition

Factor XIa is a serine protease which participates in both the extrinsic and intrinsic pathways of blood coagulation. In this work we used active site directed inhibitors to study the mechanism of factor IX activation by factor XIa. To this end, we developed a new sensitive method for the detection of factor IXa based on its affinity to antithrombin III. Using this assay, we found that the peptidic inhibitors, leupeptin and aprotinin, exhibited similar potencies in inhibiting factor IX activation and the cleavage of a tripeptidic chromogenic substrate by factor XIa. As expected, leupeptin and aprotinin were competitive with respect to the tripeptidic chromogenic substrate. However, the inhibition of factor IX activation was best described by mixed-type inhibition with the affinity of leupeptin and aprotinin to the factor XIa-factor IX complex only approximately 10-fold lower than their affinity toward factor XIa. These results, consistent with previous factor XI domain analyses, suggest that the active site of factor XIa does not contribute significantly to the affinity of factor XIa toward factor IX. The competitive component of the inhibition of factor IX activation suggests that binding of factor IX to factor XIa heavy chain affects the interactions of leupeptin and aprotinin with the active site.     

Role of charged groups in factor XI/XIa activity

To elucidate the role of charged groups in expression of factor XI coagulant activity, the charged groups of purified human blood coagulation factor XI/XIa containing 125I-XI/XIa were derivatized: free amino groups by succinylation, guanido groups of arginine by reaction with phenylglyoxal hydrate, and free carboxyl groups by reaction with ethylenediamine. The modified proteins were tested for: 1) ability to adsorb to glass, 2) ability to be cleaved by trypsin or factor XII-high molecular weight kininogen, 3) coagulant activity. The amino group-modified factor XI had a significantly decreased ability to bind to glass; modification of arginine or carboxyl groups did not affect adsorption. Trypsin cleaved factor XI with modified free amino, guanido, or carboxyl groups. Factor XII-high molecular weight kininogen could cleave only the arginine-modified factor XI. Amino group-modified factor XI and carboxyl group-modified factor XI lost all their factor XI assay activity, whereas arginine-modified factor XI retained 50% of the original activity. Amino group-modified factor XI could not be activated by trypsin, but arginine-modified and carboxyl group-modified factor XI could be activated by trypsin to 50% of the original activity. Succinylation of the amino groups of factor XIa destroyed all its factor XIa activity. Arginine-modified and carboxyl group-modified factor XIa retained 50% of their factor XIa activity. We conclude that epsilon-amino groups are essential for adsorption; activation by factor XII-high molecular weight kininogen requires free amino and carboxyl but not guanido groups; free amino, carboxyl, and guanido groups in factor XIa all appear to be critical for interaction of factor XIa with factor IX.     

Role of calcium ions and the heavy chain of factor XIa in the activation of human coagulation factor IX

Since optimal rates of factor IX activation by factor XIa require the presence of calcium ions and the heavy chain of the enzyme as well as the active-site-containing light chain, we have studied the effects of calcium ions and the heavy chain on the reaction kinetics. Whereas the amidolytic activities of factor XIa and of its active-site-containing light chain were almost indistinguishable, the two enzymes behaved quite differently when factor IX was the substrate. Factor XIa was 100-fold more potent in the presence of Ca2+ than in its absence. On the contrary, the presence or absence of Ca2+ made very little difference in the case of the isolated light chain of factor XIa. Moreover, the enzymatic activity of the light chain was almost identical with that of intact factor XIa when Ca2+ was absent. Using an optimal concentration of Ca2+, we studied the activation in the presence of various concentrations of two monoclonal antibodies, one (5F4) directed against the light chain of factor XIa and the other (3C1) against its heavy chain. Analysis of 1/V vs. 1/S plots showed that whereas inhibition by 5F4 was noncompetitive, 3C1 neutralized the enzyme in a classical competitive fashion. We conclude that in the calcium-dependent activation of factor IX by factor XIa the heavy chain of the enzyme is involved in the binding of the substrate and this is essential for optimal reaction rates.     

Bovine plasma protein C inhibitor with structural and functional homologous properties to human plasma protein C inhibitor

Bovine plasma protein C inhibitor was purified; it was then characterized in comparison with human protein C inhibitor. The specific inhibitory activity of the purified inhibitor for bovine activated protein C was 8,500 times that of the inhibitor in plasma. The purified inhibitor showed a single band with Mr 56,000 by SDS-PAGE at pH 7.0, and two bands at pH 8.8, a major one with Mr 56,000 and a minor one with Mr 105,000, under both unreduced and reduced conditions. The pI range of the inhibitor was between 4.4 and 6.1. The Mr of the inhibitor was reduced by treatment with neuraminidase, O-glycanase, and also with glycopeptidase-A, suggesting that the inhibitor has both Asn-linked and Ser/Thr-linked carbohydrate chains. Twenty-seven of the NH2-terminal 49 amino acid residues of the bovine inhibitor, which lacks the first 4 residues from the NH2-terminal amino acid sequence of human inhibitor, were identical to those of the human inhibitor. The bovine inhibitor inhibited bovine and human activated protein C, human thrombin, Factor Xa, Factor XIa, and plasma kallikrein with Ki = 1.0, 5.2, 2.6, 3.0, 1.3 X 10(-8) M, and 4.5 X 10(-9) M, respectively. The inhibitory rates for activated protein C and thrombin were accelerated significantly in the presence of heparin or negatively charged dextran sulfate. However, the acceleration by heparin or dextran sulfate for the inhibition of Factor Xa, Factor XIa, and plasma kallikrein was not significant. The bovine inhibitor did not inhibit human Factor XIIa or plasmin.(ABSTRACT TRUNCATED AT 250 WORDS)     

Design and SAR of macrocyclic Hsp90 inhibitors with increased metabolic stability and potent cell-proliferation activity

A novel series of macrocyclic ortho-aminobenzamide Hsp90 inhibitors is reported. A basic nitrogen within the tether linking the aniline nitrogen atom to a tetrahydroindolone moiety allowed access to compounds with good physical properties. Important structure-activity relationship information was obtained from this series which led to the discovery of a soluble and stable compound which is potent in an Hsp90 binding and cell-proliferation assay.     

Synthesis and in vitro biological evaluation of aryl boronic acids as potential inhibitors of factor XIa

A series of functionalized aryl boronic acids were synthesized and evaluated as potential inhibitors of factor XIa. Crystal structures of the protein-inhibitor complexes led to the design and synthesis of second generation compounds showing single digit micromolar inhibition against FXIa and selectivity against thrombin, trypsin, and FXa.     

Exosite-mediated substrate recognition of factor IX by factor XIa. The factor XIa heavy chain is required for initial recognition of factor IX

Studies of the mechanisms of blood coagulation zymogen activation demonstrate that exosites (sites on the activating complex distinct from the protease active site) play key roles in macromolecular substrate recognition. We investigated the importance of exosite interactions in recognition of factor IX by the protease factor XIa. Factor XIa cleavage of the tripeptide substrate S2366 was inhibited by the active site inhibitors p-aminobenzamidine (Ki 28 +/- 2 microM) and aprotinin (Ki 1.13 +/- 0.07 microM) in a classical competitive manner, indicating that substrate and inhibitor binding to the active site was mutually exclusive. In contrast, inhibition of factor XIa cleavage of S2366 by factor IX (Ki 224 +/- 32 nM) was characterized by hyperbolic mixed-type inhibition, indicating that factor IX binds to free and S2366-bound factor XIa at exosites. Consistent with this premise, inhibition of factor XIa activation of factor IX by aprotinin (Ki 0.89 +/- 0.52 microM) was non-competitive, whereas inhibition by active site-inhibited factor IXa beta was competitive (Ki 0.33 +/- 0.05 microM). S2366 cleavage by isolated factor XIa catalytic domain was competitively inhibited by p-aminobenzamidine (Ki 38 +/- 14 microM) but was not inhibited by factor IX, consistent with loss of factor IX-binding exosites on the non-catalytic factor XI heavy chain. The results support a model in which factor IX binds initially to exosites on the factor XIa heavy chain, followed by interaction at the active site with subsequent bond cleavage, and support a growing body of evidence that exosite interactions are critical determinants of substrate affinity and specificity in blood coagulation reactions.